PDF (tesi dottorato ROTIROTI) - FedOA - Università degli Studi di ...
PDF (tesi dottorato ROTIROTI) - FedOA - Università degli Studi di ...
PDF (tesi dottorato ROTIROTI) - FedOA - Università degli Studi di ...
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
BIOCHIP AND LAB-ON-CHIP FOR<br />
GENOMIC AND PROTEOMIC<br />
APPLICATIONS<br />
Lucia Rotiroti<br />
Dottorato in Scienze Biotecnologiche – XXI ciclo<br />
In<strong>di</strong>rizzo Biotecnologie Industriali<br />
<strong>Università</strong> <strong>di</strong> Napoli Federico II
Dottorato in Scienze Biotecnologiche – XXI ciclo<br />
In<strong>di</strong>rizzo Biotecnologie Industriali<br />
<strong>Università</strong> <strong>di</strong> Napoli Federico II<br />
BIOCHIP AND LAB-ON-CHIP FOR<br />
GENOMIC AND PROTEOMIC<br />
APPLICATIONS<br />
Lucia Rotiroti<br />
Dottoranda: Rotiroti Lucia<br />
Relatore: Dr. De Stefano Luca<br />
Coor<strong>di</strong>natore: Prof. Giovanni Sannia
The most beautiful thing we can experience is the mysterious.<br />
It is the source of all true art and science.<br />
Albert Einstein<br />
A TE<br />
…che hai sempre confidato in me e nelle mie capacità…<br />
…che mi hai dato forza, sempre…<br />
…che hai sostenuto ogni mia scelta <strong>di</strong> vita…<br />
…illuminandone la via con straor<strong>di</strong>naria <strong>di</strong>screzione…<br />
…a te…grazie!
INDICE<br />
SUMMARY pag. i<br />
RIASSUNTO<br />
Chapter 1:Porous Silicon (PSi)<br />
pag. ii<br />
Introduction<br />
pag. 1<br />
Materials and methods<br />
pag. 10<br />
Thue-Morse: aperio<strong>di</strong>c multilayers made of PSi<br />
pag. 16<br />
Optical sensing of chemical compounds using porous silicon devices<br />
Chapter 2:Chemical and Biological functionalization of PSi surface<br />
pag. 22<br />
Introduction: PSi surface chemical functionalization<br />
pag. 29<br />
DNA Optical Detection Based on PSi Technology<br />
pag. 33<br />
Optical biosensors for the detection of analytes of clinical and industrial interest. pag. 40<br />
Optical evidence of protein structural changes in PSi devices<br />
pag. 55<br />
Oligonucleotides <strong>di</strong>rect synthesis on PSi chip<br />
pag. 61<br />
PSi-based optical biosensors and biochips<br />
Chapter 3:Hybrid systems based on PSi and biocompatible materials<br />
pag. 63<br />
Introduction: PSi hybrid systems<br />
pag. 67<br />
PSi-polymer composite matrix for an innovative class of optical biosensors pag. 70<br />
A polymer mo<strong>di</strong>fied PSi optical device for biochemical sensing<br />
pag. 73<br />
Bio/NON Bio Interfaces for a new class of proteins Microarrays<br />
Chapter 4: …towards the Lab-on-Chip<br />
pag. 77<br />
Introduction to Lab-on-chip technologies<br />
pag. 82<br />
Microsystems Based on PSi-Glass Ano<strong>di</strong>c Bon<strong>di</strong>ng technologies<br />
pag. 87<br />
Laser oxidation micropatterning of a PSi based biosensor for multianalytes<br />
microarrays<br />
Appen<strong>di</strong>x<br />
pag. 95
SUMMARY<br />
There is no other inorganic material that brings together such much intriguing features like<br />
porous silicon (PSi) for people working in the optical sensing field. Porous silicon was<br />
unexpectedly <strong>di</strong>scovered in the late fifties when chemists attempted to electropolish silicon<br />
wafers with an electrolyte containing hydrofluoric acid to get better electrical contacts in<br />
the first integrated electronic circuit fabrications. The acid <strong>di</strong>ssolution left a sponge-like<br />
nanocrystalline silicon structure on the wafer. Today, the electrochemical etching is a<br />
standard method to fabricate nanostructured porous silicon: a proper choice of the applied<br />
current density, the electrolyte composition, and the silicon doping allow precise control<br />
over the morphology and, consequently, on the physical and chemical properties of the<br />
porous silicon structure. Computer controlled electrochemical etching processes are<br />
exploited for the realization of porous silicon films of controlled thickness and porosity<br />
(defined as the percentage of void in the silicon volume). Nanoporous, mesoporous and<br />
macroporous structures can be achieved, with pore size ranging from few nanometers up<br />
to microns. Moreover, since the etching process is self-stopping, it is possible to fabricate<br />
with a single run process multilayer stacks made of single layers of <strong>di</strong>fferent porosity. The<br />
<strong>di</strong>electric properties of each PSi layer, and in particular its refractive index n, can be<br />
namely modulated between those of crystalline silicon (n = 3.54, porosity = 0) and air (n=<br />
1, porosity = 100 %); so that alternating high and low porosity layers, lot of photonic<br />
structures, such as Fabry-Perot interferometers, omni-<strong>di</strong>rectional Bragg reflectors, optical<br />
filters based on microcavities, and even complicated quasi-perio<strong>di</strong>c sequences can be<br />
simply realized.<br />
The other key feature for a transducer material is the large area and the chemistry of its<br />
surface: the porous silicon exhibits a very reactive hydrogenated specific area of the order<br />
of 200 – 500 m 2 cm -3 , which assures an effective interaction with several adsorbates. The<br />
porous silicon optical sensing features are based on the changes of its photonic<br />
properties, such as photoluminescence or reflectance, on exposure to the gaseous or<br />
liquid substances. Of course, these interactions are not specific. Therefore, the porous<br />
silicon hydrogenated surface has to be chemically or physically mo<strong>di</strong>fied in order to<br />
achieve the sensing selectivity through specific biochemical interactions. The biosensor<br />
reliability strongly depends on the functionalization process: how simple, homogenous and<br />
repeatable it can be. The substitution of the superficial Si-H bonds with Si-C ones or Si-O-<br />
C guarantees a much more stable surface interface from the thermodynamic point of view.<br />
Several functionalization strategies were investigated. Testing and demonstrating the<br />
porous silicon capabilities as a useful functional material in the immobilization of biological<br />
probes and in the optical transduction of biochemical interactions is only the first action in<br />
the realisation of an optical biochip based on this nanostructured material. In this case, all<br />
the fabrication processes should be compatible with the utilisation of biological probes and<br />
the feasibility of such devices must be proven. This means that the standard integrated<br />
circuit micro-technologies should be mo<strong>di</strong>fied and adapted to this new field of application.<br />
The porous silicon low cost technology could thus provide a link between the conventional<br />
CMOS technology and the photonic devices in the realization of the so-called smart<br />
sensors and biochips. A very strong inter<strong>di</strong>sciplinary approach is required to match and<br />
resolve all the technological problems. The purpose of this thesis was to design and<br />
fabricate <strong>di</strong>fferent resonant optical structures, integrate in simple lab-on-chip devices<br />
based on the porous silicon nanotechnology, and it was focused on optical biosensing<br />
applications of social interest, such as environmental monitoring and biome<strong>di</strong>cal<br />
<strong>di</strong>agnostics.<br />
i
RIASSUNTO<br />
Una delle più importanti scoperte scientifiche <strong>di</strong> tutti i tempi è sicuramente la deco<strong>di</strong>fica del<br />
genoma umano portata a termine nel 2000 dal progetto internazionale Human Genome<br />
Organization. Lo stu<strong>di</strong>o del genoma consiste nel descrivere la struttura, la posizione e la<br />
funzione dei circa 25.000 geni che caratterizzano la specie umana. A tal fine i genetisti<br />
hanno dovuto sequenziare i circa due metri <strong>di</strong> DNA contenuto in ogni cellula umana e<br />
quin<strong>di</strong> identificare la sequenza <strong>di</strong> approssimativamente tre miliar<strong>di</strong> <strong>di</strong> coppie <strong>di</strong> basi<br />
azotate che ne compongono la molecola. Come in ogni grande impresa scientifica, lo<br />
sviluppo tecnologico ha fortemente influenzato i tempi <strong>di</strong> esecuzione e termine <strong>di</strong> questo<br />
lavoro: basti pensare che il primo miliardo <strong>di</strong> basi è stato deco<strong>di</strong>ficato in quattro anni,<br />
mentre il secondo miliardo in solo quattro mesi. Oggigiorno il ritmo <strong>di</strong> sequenziamento<br />
delle basi è dell’or<strong>di</strong>ne <strong>di</strong> <strong>di</strong>verse decine <strong>di</strong> migliaia <strong>di</strong> basi al minuto, a seconda della<br />
tecnologia utilizzata. Lo stu<strong>di</strong>o dei geni permette l’identificazione non solo del loro ruolo<br />
ma soprattutto dei loro malfunzionamenti che sono alla base delle malattie: le prospettive<br />
della me<strong>di</strong>cina genomica sono talmente vaste da non esser ancora state tutte catalogate e<br />
pienamente comprese. Il successo del progetto genoma umano è dovuto principalmente<br />
ad un fattore chiave: la miniaturizzazione della strumentazione <strong>di</strong> laboratorio, in particolare<br />
<strong>di</strong> quella de<strong>di</strong>cata alla PCR (Polymerase Chain Reaction) e delle matrici a DNA, i<br />
cosiddetti “microarray” o “biochip”, che permettono <strong>di</strong> esaminare il profilo d’espressione <strong>di</strong><br />
un gene o <strong>di</strong> identificare la presenza <strong>di</strong> un gene specifico con elevatissima precisione e<br />
ripetibilità.<br />
I biochip sono una particolare categoria <strong>di</strong> <strong>di</strong>spositivi più complessi comunemente in<strong>di</strong>cati<br />
come BioMEMS (Sistemi Micro-Elettro-Meccanici Biologici) che, negli ultimi anni, hanno<br />
trovato un numero sempre crescente <strong>di</strong> applicazioni biologiche e biome<strong>di</strong>che. Lo sviluppo<br />
<strong>di</strong> questi <strong>di</strong>spositivi è parallelo a quello dei Lab-on-chip (LOC) che rappresentano<br />
l’evoluzione tecnologica più spinta del laboratorio analitico tra<strong>di</strong>zionale: i LOC sono dei<br />
microsistemi complessi che coniugano tutte le funzioni generalmente svolte da<br />
strumentazione da banco in un unico <strong>di</strong>spositivo fabbricato su <strong>di</strong> un supporto monolitico,<br />
tipicamente <strong>di</strong> silicio cristallino, ed assemblato con altri materiali compatibili, come il silicio<br />
poroso ed il vetro o i polimeri, aventi ciascuno una propria specifica funzionalità. Entrambe<br />
le due categorie <strong>di</strong> <strong>di</strong>spositivi, i biochip ed i LOC, beneficano <strong>degli</strong> sviluppi <strong>di</strong> quello che è<br />
sicuramente il campo più avanzato della tecnologia umana e cioè la microelettronica<br />
integrata: la produzione <strong>di</strong> computer sempre più potenti è legata alla capacità <strong>di</strong> fabbricare<br />
<strong>di</strong>spositivi su scala micrometrica e nanometrica con una capacità <strong>di</strong> controllo della materia<br />
solida prossima al limite atomico. La possibilità <strong>di</strong> miniaturizzare la componentistica <strong>di</strong><br />
laboratorio (separatori, concentratori, analizzatori, sensori, e così via) fa sì che il campo <strong>di</strong><br />
applicazione dei LOC abbia una evoluzione esponenziale trovando impiego dalla<br />
<strong>di</strong>agnostica al monitoraggio ambientale alla rivelazione <strong>di</strong> agenti pericolosi nella sicurezza<br />
civile e militare, fino alle applicazioni industriali in chimica sintetica (per esempio in prove<br />
rapi<strong>di</strong> in microreattori per la farmaceutica).<br />
A partire dai biochip a DNA, è stata sviluppata una serie <strong>di</strong> <strong>di</strong>spositivi che utilizzano altri<br />
organismi biologici quali anticorpi, enzimi, proteine, cellule. Data la capacità <strong>di</strong> queste<br />
biomolecole <strong>di</strong> agire come sonde altamente specifiche rispetto ai rispettivi ligan<strong>di</strong>, la<br />
principale applicazione dei <strong>di</strong>spositivi è come biosensori e cioè nel riconoscimento della<br />
presenza dei ligan<strong>di</strong> in miscele complesse quale può essere il sangue umano. Un<br />
biosensore, il cui schema è rappresentato in Figura 1, lavora con un elemento biologico,<br />
detto biorecettore, fissato su un substrato ove avviene una reazione biologica specifica tra<br />
il biorecettore e l’elemento ricercato. L’evento <strong>di</strong> riconoscimento biomolecolare è rilevato<br />
da un trasduttore che lo converte in un segnale elettrico, ottico o chimico, che può essere<br />
osservato e misurato. La sensibilità e la selettività dei biosensori è tale da poter applicare<br />
questi strumenti anche nello stu<strong>di</strong>o delle interazioni biomolecolari: un chiaro successo<br />
ii
commerciale <strong>di</strong> questi <strong>di</strong>spositivi è il Biacore ® , <strong>di</strong>spositivo ottico comunemente utilizzato<br />
nella caratterizzazione della cinetica molecolare e nel calcolo delle costanti <strong>di</strong><br />
<strong>di</strong>ssociazione e <strong>di</strong> affinità tra biomolecole.<br />
Internal<br />
Membrane<br />
Electrolyte<br />
Signal<br />
Transducer<br />
Biorecognition<br />
Element<br />
Semipermeable<br />
Membrane<br />
Figura 1 Schema generale <strong>di</strong> un biosensore.<br />
Sulla base del metodo <strong>di</strong> trasduzione impiegato, i biosensori possono essere raggruppati,<br />
in quattro gruppi fondamentali:<br />
1. Elettrochimici;<br />
2. Termici;<br />
3. Ottici;<br />
4. Acustici;<br />
Le prime due tecniche <strong>di</strong> trasduzione vengono spesso impiegate nella realizzazione <strong>di</strong><br />
<strong>di</strong>spositivi il cui biorecettore è un enzima; mentre i trasduttori più comunemente utilizzati<br />
per la realizzazione <strong>di</strong> biosensori <strong>di</strong> affinità (anticorpo/antigene) sono <strong>di</strong> tipo ottico ed<br />
acustico. I biosensori ottici sono <strong>di</strong> gran lunga i più <strong>di</strong>ffusi grazie ad alcuni vantaggi legati<br />
alla natura della luce: innanzitutto la possibilità <strong>di</strong> vedere a livello micrometrico e sub<br />
micrometrico la <strong>di</strong>stribuzione delle sonde molecolari con meccanismi <strong>di</strong> contrasto che<br />
vanno dalla fluorescenza al contrasto <strong>di</strong> fase; la non invasività dovuta alla non interazione<br />
<strong>di</strong> alcune lunghezze d’onda con le biomolecole; l’elevata sensibilità e così via. Se le<br />
tecniche <strong>di</strong> visualizzazione (“imaging”) e spettroscopia in fluorescenza rappresentano la<br />
storia dei saggi biologici e lo stato dell’arte nei microarray per applicazioni genomiche, vi è<br />
un crescente impulso scientifico e tecnologico, testimoniato da numerosi lavori scientifici<br />
negli ultimi cinque anni, per la realizzazione <strong>di</strong> <strong>di</strong>spositivi ottici che permettono <strong>di</strong> misurare<br />
<strong>di</strong>rettamente le reazioni <strong>di</strong> riconoscimento biomolecolare senza dover ricorrere a dei<br />
marcatori fluorescenti. Queste tecniche “label free” offrono vantaggi non secondari<br />
derivanti proprio dall’evitare l’operazione <strong>di</strong> marcatura con fluorofori della molecola<br />
interessata: benché quest’operazione non sia particolarmente complicata, non sempre è<br />
possibile <strong>di</strong>sporre <strong>di</strong> un saggio che prevede la marcatura del ligando, inoltre legare un<br />
fluoroforo ad una molecola significa perturbare in maniera non banale la struttura della<br />
stessa cambiandone così il comportamento biologico.<br />
Il progetto <strong>di</strong> ricerca oggetto della presente <strong>tesi</strong> <strong>di</strong> <strong>dottorato</strong> si inserisce in questo filone <strong>di</strong><br />
ricerca biotecnologica internazionale ed è stato finalizzato in particolare alla realizzazione<br />
<strong>di</strong> un microsistema ottico per lo stu<strong>di</strong>o delle interazioni biomelocolari tra singole eliche <strong>di</strong><br />
DNA oppure tra proteine e ligan<strong>di</strong>, sia per applicazioni genomiche e proteomiche, ma<br />
anche capace <strong>di</strong> configurarsi come biosensore abile all’in<strong>di</strong>viduazione <strong>di</strong> analiti bersaglio<br />
in miscele eterogenee. La realizzazione <strong>di</strong> un <strong>di</strong>spositivo del genere è un progetto<br />
complesso che può essere schematizzato nelle sue varie fasi come in<strong>di</strong>cato in Figura 2: si<br />
parte dall’in<strong>di</strong>viduazione <strong>di</strong> un materiale <strong>di</strong> supporto che abbia opportune caratteristiche<br />
ottiche <strong>di</strong> trasduzione, che possa essere opportunamente funzionalizzato con sonde<br />
molecolari e che possa essere integrato in un microsistema completo <strong>di</strong> microflui<strong>di</strong>ca. La<br />
realizzazione <strong>di</strong> questo microsistema ha richiesto un approccio fortemente multi<strong>di</strong>sciplinare<br />
nonché l’interazione con figure professionali specializzate in ambiti scientifici<br />
complementari ma molto <strong>di</strong>stanti tra loro come la fisica dei semiconduttori, la biologia<br />
Analyte<br />
iii
molecolare, la chimica delle superfici, l’ingegneria microelettronica. Il materiale scelto<br />
quale supporto ed elemento trasduttore è il silicio poroso: la sua caratteristica morfologica<br />
è quella <strong>di</strong> una struttura spugnosa interconnessa da nanocristalli <strong>di</strong> silicio che possono<br />
avere <strong>di</strong>sposizione or<strong>di</strong>nata o quasi caotica a seconda delle procedure <strong>di</strong> produzione<br />
adottate. Questa morfologia tipica gli conferisce un’elevata superficie specifica che<br />
assicura un’efficace interazione con agenti esterni. Inoltre varia fortemente le sue<br />
caratteristiche chimico-fisiche quando esposto ad agenti biochimici.<br />
Le problematiche tecnico-scientifiche affrontate nel corso dei tre anni del <strong>dottorato</strong>, sono<br />
state quin<strong>di</strong> in particolare:<br />
1. la fabbricazione del silicio poroso, quale elemento <strong>di</strong> supporto del biosensore e<br />
sistema <strong>di</strong> trasduzione del segnale ottico;<br />
2. lo stu<strong>di</strong>o <strong>di</strong> una procedura affidabile <strong>di</strong> funzionalizzazione della superficie <strong>di</strong> silicio<br />
poroso in grado <strong>di</strong> ancorare in modo covalente le molecole sonda al materiale<br />
supporto;<br />
3. l’in<strong>di</strong>viduazione <strong>di</strong> sonde molecolari particolarmente stabili e compatibili con la<br />
realizzazione <strong>di</strong> biosensori riutilizzabili, quali ad esempio le proteine purificate da<br />
organismi estremofili;<br />
4. la messa a punto <strong>di</strong> opportuni protocolli sperimentali per la definizione dell’attività<br />
delle sonde una volta agganciate al supporto trasduttore;<br />
5. lo sviluppo <strong>di</strong> tecniche <strong>di</strong> microlavorazione del silicio, del vetro e <strong>di</strong> altri materiali<br />
eventualmente impiegati per la realizzazione <strong>di</strong> un prototipo <strong>di</strong> “lab-on-chip”,<br />
compatibili con l’utilizzo <strong>di</strong> sonde molecolari biologiche.<br />
1<br />
2<br />
3<br />
Reflectivity (a.u.)<br />
1,0<br />
0,8<br />
0,6<br />
0,4<br />
0,2<br />
0,0<br />
600 800 1000 1200 1400 1600<br />
Wavelength (nm)<br />
Figura 2: schematizzazione grafica del progetto <strong>di</strong> <strong>tesi</strong><br />
L’integrazione <strong>di</strong> un biosensore con le tecnologie tipiche della microelettronica e dei circuiti<br />
integrati infatti non è mai una evoluzione <strong>di</strong>retta del <strong>di</strong>spositivo sensibile in un<br />
microsistema, perché l’impiego <strong>di</strong> sonde biologiche introduce una <strong>di</strong>fficoltà in più legata<br />
alla stabilità delle stesse in con<strong>di</strong>zioni <strong>di</strong> temperatura <strong>di</strong>verse da quelle ambientali.<br />
Il microsistema realizzato è stato utilizzato per rivelare l’ibridazione tra eliche <strong>di</strong> DNA<br />
complementari e per stu<strong>di</strong>are gli eventi <strong>di</strong> riconoscimento molecolare tra proteina e<br />
ligando. In quanto segue si riportano le metodologie ed i principali risultati ottenuti durante<br />
il lavoro <strong>di</strong> <strong>tesi</strong>.<br />
Produzione e caratterizzazione del materiale Silicio Poroso (J3, J11)<br />
Il Silicio Poroso (PSi) è prodotto per <strong>di</strong>ssoluzione elettrochimica <strong>di</strong> silicio cristallino<br />
drogato, ed ha una morfologia spugnosa che gli conferisce una superficie specifica molto<br />
elevata, dell’or<strong>di</strong>ne <strong>di</strong> alcune centinaia <strong>di</strong> metri quadri per centimetro cubo. La <strong>di</strong>ssoluzione<br />
4<br />
5<br />
iv
del silicio coinvolge due portatori liberi <strong>di</strong> carica (elettroni e e lacune h) come riportato nello<br />
schema <strong>di</strong> reazione riportato in Figura 3.<br />
Si + 2 F + 2 h<br />
SiF2 + 4 HF H2SiF6 + H2 + SiF2 SiF 2 + 4F SiF 6 2 + 2e<br />
2H + + 2e H 2<br />
Figura 3: semireazioni <strong>di</strong> <strong>di</strong>ssoluzione elettrochimica del silicio cristallino<br />
Le <strong>di</strong>mensioni dei pori della struttura sono variabili da qualche nanometro fino a più <strong>di</strong> un<br />
micron a seconda del tipo <strong>di</strong> substrato <strong>di</strong> silicio cristallino utilizzato, della composizione<br />
della soluzione chimica e della corrente erogata nella cella durante il processo <strong>di</strong><br />
fabbricazione.<br />
La porosità P è definita come la percentuale <strong>di</strong> aria presente in uno strato <strong>di</strong> materiale (ad<br />
esempio uno strato <strong>di</strong> silicio poroso, con porosità del 60 %, sarà composto da silicio<br />
nanocristallino solo per il 40 % in massa), e può essere misurata con una tecnica <strong>di</strong> tipo<br />
gravimetrico a partire dal rapporto tra la massa <strong>di</strong> silicio <strong>di</strong>ssolto M<strong>di</strong>s, durante il processo<br />
d’ano<strong>di</strong>zzazione, e la massa totale <strong>di</strong> silicio sottoposto all’attacco MSi. La determinazione<br />
dello spessore e della porosità <strong>di</strong> uno strato prodotto con un singolo processo può essere<br />
effettuata alternativamente con un metodo <strong>di</strong> caratterizzazione ottica quale l’ellissometria<br />
spettroscopica ad angolo variabile. Per ottenere una <strong>di</strong>fferenziazione tra le varie tipologie<br />
<strong>di</strong> PSi, può essere utilizzata una classificazione in base al <strong>di</strong>ametro me<strong>di</strong>o d dei pori<br />
(standard IUPAC):<br />
1. Silicio microporoso (<strong>di</strong>ametro me<strong>di</strong>o del poro d < 2nm);<br />
2. Silicio mesoporoso (<strong>di</strong>ametro me<strong>di</strong>o del poro d compreso nell’intervallo 2 – 50 nm);<br />
3. Silicio macroporoso (<strong>di</strong>ametro me<strong>di</strong>o del poro d > 50nm).<br />
Spesso la classificazione fatta utilizzando soltanto la <strong>di</strong>mensione del <strong>di</strong>ametro d del poro è<br />
riduttiva ed insufficiente per descrivere interamente la struttura dello strato poroso<br />
ottenuto. La <strong>di</strong>mensione dei pori, infatti, non contiene informazioni riguardo la morfologia.<br />
Con questo termine s’intende la forma, l’orientazione e l’interazione tra pori. Le proprietà<br />
termiche, elettriche, ottiche, del silicio poroso <strong>di</strong>pendono fortemente sia dal valore della<br />
porosità P che dalla morfologia della matrice spugnosa. Il parametro che influenza<br />
maggiormente la morfologia del PSi è il drogaggio del silicio cristallino <strong>di</strong> partenza:<br />
campioni poco drogati danno pori piccoli e solitamente allineati lungo una <strong>di</strong>rezione<br />
privilegiata, mentre campioni molto drogati una volta attaccati si ramificano in una matrice<br />
caotica <strong>di</strong> tipo spugnoso. Alcune morfologie <strong>di</strong> silicio poroso sono mostrate come esempio<br />
in Figura 4.<br />
Figura 4: Possibili morfologie <strong>di</strong> film <strong>di</strong> silicio poroso su substrati <strong>di</strong> tipo p e n.<br />
Per ottimizzare l’interazione del silicio poroso con le soluzioni biologiche si è scelto <strong>di</strong><br />
lavorare con un materiale mesoporoso e cioè costituito da pori con <strong>di</strong>mensioni me<strong>di</strong>e<br />
v
dell’or<strong>di</strong>ne dei 10-100 nm. Questo si ottiene dalla parziale <strong>di</strong>ssoluzione <strong>di</strong> silicio <strong>di</strong> tipo p + ,<br />
con elevato numero <strong>di</strong> portatori liberi <strong>di</strong> carica dunque, in una soluzione <strong>di</strong> acido fluoridrico<br />
ed etanolo in rapporto 1:1. Per ottenere strati <strong>di</strong> silicio poroso con porosità e spessore<br />
definiti è stato necessario, una volta fissata la composizione della soluzione elettrolitica ed<br />
il tipo e livello <strong>di</strong> drogaggio del substrato <strong>di</strong> silicio cristallino, calibrare il sistema <strong>di</strong><br />
produzione e cioè costruire delle curve sperimentali che determinano la velocità <strong>di</strong> attacco<br />
(“etch rate” [nm/s] ) e la porosità, ottenuta al variare della densità <strong>di</strong> corrente applicata alla<br />
cella.<br />
Una delle caratteristiche peculiari del processo <strong>di</strong> produzione del silicio poroso è <strong>di</strong> essere<br />
“self-stopping”: se si interrompe l’attacco elettrochimico fermando l’erogazione <strong>di</strong> corrente,<br />
lo strato <strong>di</strong> poroso formatosi è svuotato dai portatori liberi <strong>di</strong> carica che hanno partecipato<br />
alla <strong>di</strong>ssoluzione del silicio cristallino, per cui riprendendo ad erogare corrente il processo<br />
riparte dall’interfaccia silicio poroso-silicio cristallino, ove sono presenti i portatori <strong>di</strong> carica.<br />
In questo modo è possibile creare, durante un unico processo <strong>di</strong> attacco, strati successivi<br />
e a<strong>di</strong>acenti <strong>di</strong> silicio poroso con spessori e porosità <strong>di</strong>verse l’uno dall’altro partendo<br />
dall’alto (prima superficie esposta) verso il basso (fondo del wafer <strong>di</strong> silicio cristallino). Il<br />
buon controllo del processo <strong>di</strong> attacco permette la progettazione e la realizzazione <strong>di</strong><br />
strutture in PSi con specifiche caratteristiche ottiche. È possibile realizzare <strong>di</strong>spositivi a<br />
singolo strato, questi si comportano dal punto <strong>di</strong> vista ottico come <strong>degli</strong> interferometri<br />
(Fabry-Perot) il cui spettro in riflessione è riportato in Figura 5 (a) o progettare <strong>di</strong>spositivi<br />
multistrato a specifiche lunghezze d’onda <strong>di</strong> risonanza, quali microcavità ottiche (b),<br />
specchi <strong>di</strong> Bragg (c) e sequenze Thue-Morse (d).<br />
Reflectivity (a.u.)<br />
1.0<br />
0.8<br />
0.6<br />
0.4<br />
0.2<br />
0.0<br />
a<br />
Reflectivity (a.u)<br />
1.0<br />
0.8<br />
0.6<br />
0.4<br />
0.2<br />
monostrato<br />
600 800 1000 1200 1400 1600<br />
Wavelength (nm)<br />
specchio <strong>di</strong> Bragg<br />
600 800 1000 1200 1400 1600<br />
b<br />
Reflectivity (a.u)<br />
2.0<br />
1.5<br />
1.0<br />
0.5<br />
0.0<br />
microcavità ottica<br />
800 1000 1200 1400<br />
Wavelength (nm)<br />
600 800 1000 1200 1400 1600<br />
Wavelength (nm)<br />
Wavelength (nm)<br />
c<br />
d<br />
Figura 5: schematizzazioni e spettri in riflessione dei <strong>di</strong>spositivi in silicio poroso ottenuti illuminando con luce<br />
bianca.<br />
Lo spettro in riflessione si ottiene facendo incidere sul <strong>di</strong>spositivo un fascio <strong>di</strong> luce<br />
collimato, proveniente da una sorgente ad ampio spettro (400-1600 nm), ed inviando la<br />
luce riflessa ad un analizzatore ottico <strong>di</strong> spettro secondo lo schema riportato in Figura 6.<br />
Reflectivity (a.u)<br />
1.0<br />
0.8<br />
0.6<br />
0.4<br />
0.2<br />
0.0<br />
Thue Morse S6<br />
Figura 6: apparato sperimentale utilizzato per misurare gli spettri ottici <strong>di</strong> riflessione dei campioni in PSi<br />
vi
I sensori ottici basati sul silicio poroso sfruttano le variazioni delle proprietà <strong>di</strong>elettriche <strong>di</strong><br />
questo materiale quando esso è esposto all’analita che penetra nei suoi pori. (Figura 7)<br />
Reflectivity (a. u.)<br />
1.1<br />
1.0<br />
0.9<br />
0.8<br />
0.7<br />
0.6<br />
0.5<br />
0.4<br />
cavità<br />
imperturbata<br />
Clorobenzene<br />
1275 1300 1325 1400 1425 1450 1475 1500<br />
Wavelength (nm)<br />
A<br />
Peak shift (nm)<br />
12<br />
10<br />
8<br />
6<br />
4<br />
2<br />
0<br />
Limit of solubility in water<br />
Atrazine in H 2 O<br />
S = 1.19 (3) nm/ppm<br />
LOD = 0.17 ppm<br />
Atrazine in HA<br />
S = 0.51 (2) nm/ppm<br />
LOD = 0.2 ppm<br />
0 5 10 15 20 25 30 35<br />
C (ppm)<br />
B<br />
Figura 7: schematizzazioni del meccanismo <strong>di</strong> riconoscimento; A) spettri in riflessione <strong>di</strong> una micro cavità<br />
ottica sottoposta a vapori <strong>di</strong> sostanze organiche; B) curva dose-risposta del sensore sottoposto a<br />
concentrazioni <strong>di</strong>verse <strong>di</strong> pesticida in soluzione acquosa.<br />
L’elevato rapporto superficie-volume del silicio poroso e la risposta ottica dei <strong>di</strong>spositivi,<br />
permettono <strong>di</strong> monitorare gli effetti dei cambiamenti in in<strong>di</strong>ce <strong>di</strong> rifrazione dello strato <strong>di</strong><br />
poroso, dovuti alla presenza <strong>di</strong> sostanze liquide o gassose.La sostituzione dell’aria<br />
presente nei pori con le molecole delle sostanze organiche, provoca uno spostamento<br />
verso lunghezze d’onda maggiori come in<strong>di</strong>cato nella Figura 7A.<br />
Dal momento che il meccanismo <strong>di</strong> rivelazione non è selettivo e non riesce a <strong>di</strong>stinguere<br />
analiti presenti in miscele complesse è necessario implementare la selettività del silicio<br />
poroso utilizzando sonde molecolari ad alta selettività che possono essere legate<br />
covalentemente alla sua superficie con opportune tecniche <strong>di</strong> funzionalizzazione chimica.<br />
Funzionalizzazione chimica della superficie del Silicio Poroso (J6, J12, J21)<br />
La superficie del silicio poroso appena prodotto è fortemente reattiva in quanto idrogenata,<br />
presenta cioè legami Si-H che sono termo<strong>di</strong>namicamente instabili: tendono infatti a reagire<br />
con l’ossigeno presente in atmosfera per creare legami tipo Si-O-Si molto più stabili.<br />
Inoltre una superficie idrogenata è altamente idrofobica, pertanto una passivazione<br />
chimica o fisica ha la funzione non solo <strong>di</strong> indurre specifici legami tra gli atomi <strong>di</strong> silicio ed<br />
altre specie chimiche in modo da essere capace <strong>di</strong> legare selettivamente composti chimici<br />
o molecole <strong>di</strong> origine biologica, ma anche <strong>di</strong> rendere la superficie idrofilica in modo da<br />
permettere una facile penetrazione del materiale biologico, spesso in soluzione acquosa.<br />
In Figura 8 sono riportate le misure <strong>di</strong> angolo <strong>di</strong> contatto della superficie del PSi appena<br />
prodotto (A), superficie altamente idrofobica (θ=131°), e dopo funzionalizzazione chimica<br />
(B) da cui si ottiene una superficie idrofilica (θ=42°).<br />
A B<br />
Figura 8: caratterizzazione della superficie del PSi tramite misure <strong>di</strong> angolo <strong>di</strong> contatto. A) la superficie del<br />
PSi appena prodotto si presenta altamente idrofobica. B) dopo il processo <strong>di</strong> passivazione la superficie del<br />
PSi <strong>di</strong>venta idrofilica permettendo così la penetrazione delle soluzioni biologiche.<br />
vii
Un caso particolarmente interessante è quello in cui i legami con l’idrogeno sono sostituiti<br />
con legami Si-C i quali sono molto stabili sia rispetto alle variazioni <strong>di</strong> temperatura sia<br />
rispetto alle interazioni con gruppi nucleofili che hanno grande potere ossidante.<br />
Per in<strong>di</strong>viduare la procedura <strong>di</strong> mo<strong>di</strong>fica chimica della superficie <strong>di</strong> silicio poroso in grado<br />
<strong>di</strong> ancorare in modo covalente le molecole sonda, sono state sperimentate quattro<br />
tecniche <strong>di</strong> funzionalizzazione dei <strong>di</strong>spositivi ottici:<br />
1. una funzionalizzazione chimica ad opera dei reattivi <strong>di</strong> Grignard;<br />
2. una funzionalizzazione fotochimica, indotta da ra<strong>di</strong>azione ultravioletta e condotta<br />
in atmosfera inerte immergendo il campione in una soluzione <strong>di</strong> estere<br />
succinimmi<strong>di</strong>co <strong>di</strong> acido undecenoico;<br />
3. una funzionalizzazione in situ, ottenuta introducendo il linker organico, sotto<br />
forma <strong>di</strong> acido carbossilico, <strong>di</strong>rettamente nella soluzione <strong>di</strong> attacco<br />
elettrochimico;<br />
4. una funzionalizzazione della superficie del silicio poroso ossidata ad opera <strong>di</strong><br />
alcossi silani.<br />
Per analizzare la composizione superficiale del PSi si utilizza la tecnica della spettroscopia<br />
infrarossi a trasformata <strong>di</strong> Fourier (FTIR), che permette <strong>di</strong> acquisire informazioni sia<br />
quantitative sia qualitative riguardo alla chimica della superficie. In Figura 9 è mostrato un<br />
esempio <strong>di</strong> caratterizzazione prima e dopo il processo <strong>di</strong> funzionalizzazione, si monitora la<br />
scomparsa dei picchi caratteristici del PSi appena prodotto (SiH a 2100cm -1 ) a favore della<br />
comparsa <strong>di</strong> quelli caratteristici del linker chimico utilizzato.<br />
% Riflettività<br />
140<br />
120<br />
100<br />
80<br />
60<br />
40<br />
20<br />
0<br />
Bragg appena prodotto<br />
Bragg dopo funzionalizzazione con UANHS<br />
2855-2929 cm -1 CH 2<br />
1723 cm -1<br />
O-C=O<br />
1641-1563 cm -1<br />
succinimmide<br />
1288 cm<br />
-20<br />
4000 3500 3000 2500 2000 1500 1000<br />
-1 1038 cm<br />
CN; C=O<br />
-1<br />
N-O<br />
numero d'onda (cm -1 )<br />
Figura 9: spettro FTIR <strong>di</strong> un <strong>di</strong>spositivo in PSi prima (nero) e dopo (rosso) la funzionalizzazione chimica.<br />
Sulla base dei gruppi funzionali <strong>di</strong>sponibili sulla sonda da immobilizzare (DNA singola<br />
eliche, proteine, enzimi), è possibile in<strong>di</strong>viduare una procedura <strong>di</strong> passivazione chimica<br />
adatta al processo <strong>di</strong> funzionalizzazione: per esempio, è possibile funzionalizzare la<br />
superficie del silicio poroso con un silossano -NH2 terminale in modo da consentire<br />
l’ancoraggio <strong>di</strong> biomolecole o crosslinker carbossi terminali. In Figura 10 si riporta uno<br />
schema dei passaggi <strong>di</strong> funzionalizzazione a partire dal PSi appena prodotto fino<br />
all’immobilizzazione delle sonde molecolari.<br />
2100 cm -1 SiH<br />
viii
Si H<br />
OH<br />
O2 5% APTES<br />
Si H<br />
SiO2 OH<br />
15' @900°C 20' R.T<br />
Si H<br />
OH<br />
O 2 Si<br />
O2Si O<br />
O<br />
O<br />
O<br />
Si<br />
O<br />
O<br />
O<br />
Si<br />
O<br />
C<br />
N<br />
NH 2<br />
2.5% GA<br />
30' R.T<br />
or PROBE-COH<br />
C O<br />
H<br />
PROBE-NH 2<br />
Figura 10: schema <strong>di</strong> funzionalizzazione chimica della superficie del PSi<br />
Un’alternativa ai meto<strong>di</strong> classici <strong>di</strong> funzionalizzazione è rappresentata dalla possibilità <strong>di</strong><br />
infiltrare un polimero biocompatibile contenente pendagli funzionali regolarmente spaziati<br />
lungo la catena. In questo modo i gruppi funzionali sono reattivi e <strong>di</strong>sponibili per<br />
l’ancoraggio delle sonde biologiche.<br />
Il polimero scelto per gli esperimenti in questo lavoro <strong>di</strong> <strong>tesi</strong>, è il Poli(ε-caprolattone) (PCL),<br />
polimero alifatico idrofobico il cui impiego è stato ampiamente stu<strong>di</strong>ato negli ultimi anni, in<br />
campi <strong>di</strong> applicazione che vanno dalla realizzazione <strong>di</strong> tessuti ingegnerizzati al rilascio<br />
controllato <strong>di</strong> farmaci.<br />
Sono state messe a punto le con<strong>di</strong>zioni <strong>di</strong> deposizione per rotazione veloce (“spin<br />
coating”). Questa tecnica è utilizzata per depositare un film sottile ed uniforme su un<br />
substrato solido piano. La quantità in eccesso <strong>di</strong> una soluzione molto <strong>di</strong>luita della specie<br />
che si vuole depositare viene depositata sul substrato, che è successivamente messo in<br />
rapida rotazione, al fine <strong>di</strong> <strong>di</strong>stribuire il fluido sul substrato per effetto della forza centrifuga.<br />
È stata quin<strong>di</strong> determinata la concentrazione ottimale del polimero per avere un film con<br />
uno spessore confrontabile con la <strong>di</strong>mensione dei pori, dell’or<strong>di</strong>ne dei 10 nanometri. In<br />
Figura 11 sono riportati lo schema <strong>di</strong> deposizione per spin coating (a) ed il grafico della<br />
variazione dello spessore del film al variare della concentrazione <strong>di</strong> polimero (b).<br />
thickness (nm)<br />
30<br />
25<br />
20<br />
15<br />
10<br />
5<br />
0 1 2 3 4 5<br />
PCL concentration (mg/ml)<br />
a b<br />
Figura 11: a) Schema <strong>di</strong> deposizione <strong>di</strong> film sottili per Spin coating; b) variazione dello spessore in funzione<br />
della concentrazione <strong>di</strong> PCL.<br />
La caratterizzazione in spettroscopia ellissometrica ad angolo variabile mostra che<br />
l’infiltrazione <strong>di</strong> PCL porta ad una riduzione <strong>di</strong> porosità da 79% fino ad un valore <strong>di</strong> 75.2 %<br />
con una <strong>di</strong>somogeneità verticale nella <strong>di</strong>stribuzione del polimero. Il polimero infatti penetra<br />
nell’intera struttura ma si accumula sul fondo dove la presenza <strong>di</strong> idrogeno è maggiore, a<br />
causa dell’interazione idrofobica tra polimero e superficie idrogenata del PSi. Attraverso<br />
misure <strong>di</strong> angolo <strong>di</strong> contatto, variando la concentrazione del polimero infiltrato, si definisce<br />
come varia la bagnabilità della superficie <strong>di</strong> PSi appena prodotto: all’aumentare della<br />
ix
concentrazione aumenta la bagnabilità della superficie passando così da una superficie<br />
idrofobica ad una idrofilica.<br />
Essendo l’idrossido <strong>di</strong> so<strong>di</strong>o in soluzione acquosa un agente altamente corrosivo per il<br />
silicio poroso, è stata condotta una prova <strong>di</strong> resistenza ad una soluzione <strong>di</strong> [NaOH]=0,1M;<br />
sono stati realizzati due campioni <strong>di</strong> PSi gemelli, in uno è stato infiltrato il PCL mentre nel<br />
controllo no. Il campione infiltrato ha resistito per 18 minuti alla corrosione dovuta<br />
all’NaOH, raggiungendo poi una stazionarietà. È stata infiltrata anche una struttura<br />
multistrato in PSi, ed in particolare un Bragg costituito da 15 coppie <strong>di</strong> monostrati a due<br />
porosità <strong>di</strong>verse. Il <strong>di</strong>spositivo ha resistito 15 minuti (Figura 12), mantenendo la classica<br />
risposta ottica della struttura risonante.<br />
Reflectivity (a.u.)<br />
Reflectivity (a.u.)<br />
1<br />
0<br />
1.5<br />
1.0<br />
0.5<br />
PSi/PCL<br />
dopo 15 min in NaOH<br />
600 800 1000 1200 1400 1600<br />
PSi<br />
dopo 80s in NaOH<br />
0.0<br />
600 800 1000 1200 1400 1600<br />
Wavelength (nm)<br />
Figura 12: Confronto tra un <strong>di</strong>spositvo <strong>di</strong> silicio poroso infiltrato (A) ed uno no (B) prima (nero) e dopo<br />
l’immersione in una soluzione <strong>di</strong> NaOH 0.1 M (rosso).<br />
Sul campione è stato infine condotto un esperimento <strong>di</strong> sensing <strong>di</strong> solventi organici, per<br />
verificare che il trattamento alcalino non avesse alterato la struttura porosa del <strong>di</strong>spositivo.<br />
Il Bragg funziona ancora come sensore <strong>di</strong> composti chimici volatili, poiché sottoponendo la<br />
struttura ad atmosfera satura <strong>di</strong> solventi organici questi hanno provocato uno spostamento<br />
dello spettro <strong>di</strong> riflessione verso lunghezze d’onda maggiori, chiaro segno <strong>di</strong> una<br />
sostituzione dell’aria all’interno dei pori con una sostanza ad in<strong>di</strong>ce <strong>di</strong> rifrazione maggiore.<br />
Lo spettro si sposta verso il rosso in maniera proporzionale all’aumentare dell’in<strong>di</strong>ce <strong>di</strong><br />
rifrazione delle sostanze, con una sensibilità <strong>di</strong> 585 nm/RIU (RIU: Unità <strong>di</strong> In<strong>di</strong>ce <strong>di</strong><br />
Rifrazione) ed un LOD (Limite <strong>di</strong> rilevabilità) <strong>di</strong> 5*10 -4 RIU.<br />
Sonde biologiche utilizzate<br />
Sono state stu<strong>di</strong>ate, in collaborazione con il gruppo <strong>di</strong> ricerca del Prof. P. Arcari del<br />
Dipartimento <strong>di</strong> Biochimica e Biotecnologie Me<strong>di</strong>che dell’<strong>Università</strong> <strong>di</strong> Napoli “Federico II”,<br />
le interazioni ssDNA-cDNA e DNA/PNA<br />
DNA: è stata utilizzata una sequenza singola <strong>di</strong> DNA (ssDNA:<br />
5’GGACTTGCCCGAATCTACGTGTCCA3’,Primm), questa per gli esperimenti <strong>di</strong><br />
fluorescenza è stata marcata con Fluoresceina CY3.5 (con picco <strong>di</strong> assorbimento a<br />
581 nm e <strong>di</strong> emissione a 596 nm).<br />
ed in collaborazione col gruppo del dr. S. D’Auria dell’Istituto <strong>di</strong> Biochimica delle Proteine<br />
del Consiglio Nazionale delle Ricerche le interazioni proteina-proteina quali: GlnBP-<br />
Glutamina, GlnBP-Glia<strong>di</strong>na, e TMBP-Glucosio.<br />
Glutamine-bin<strong>di</strong>ng protein (GlnBP): la GlnBP, purificata da E. coli è una proteina<br />
monomerica composta da 224 residui amminoaci<strong>di</strong>ci (26 kDa) responsabile del<br />
primo step nel trasporto attivo della L-glutamina (Gln) attraverso la membrana<br />
citoplasmatica. La Gln è la risorsa principale <strong>di</strong> azoto e carbonio nel mezzo <strong>di</strong><br />
coltura cellulare: il suo monitoraggio è quin<strong>di</strong> fondamentale nel controllo dei<br />
bioprocessi. Tra tutti gli amminoaci<strong>di</strong> presenti in natura la GlnBP lega unicamente la<br />
Gln con una costante <strong>di</strong> <strong>di</strong>ssociazione Kd <strong>di</strong> 5 x 10 -9 M. È stata utilizzata anche per<br />
x
verificare la sua capacità <strong>di</strong> legare con alta affinità gli amminoaci<strong>di</strong> presenti nella<br />
glia<strong>di</strong>na (proteina presente in quasi tutti i cereali derivata dal glutine e ritenuta<br />
responsabile del morbo celiaco) costituita da una sequenza polimerica <strong>di</strong> gln.<br />
D-trehalose/D-maltose-bin<strong>di</strong>ng protein (TMBP): la TMBP è una componente del<br />
sistema rifornimento del trealosio (Tre) e del maltosio (Mal) nell’ipertermofilo T.<br />
litoralis. La TMBP estratta da T. litoralis è una macromolecola (48 kDa) costituita da<br />
due domini e contenente 12 residui <strong>di</strong> triptofano. La struttura tri<strong>di</strong>mensionale della<br />
TMBP è comune ad un gran numero <strong>di</strong> altre proteine atti a legare gli zuccheri. È<br />
stato recentemente stu<strong>di</strong>ato che la TMBP è capace <strong>di</strong> legare anche le molecole <strong>di</strong><br />
glucosio oltre a quelle dei due zuccheri, <strong>di</strong>meri del glucosio. Poiché il sangue<br />
umano non contiene né Tre né Mal, si è pensato <strong>di</strong> realizzare un biochip non<br />
invasivo per il monitoraggio del glucosio utilizzando questa proteina estratta da<br />
estremofili e per questo più stabile rispetto ad altre.<br />
<strong>Stu<strong>di</strong></strong>o delle interazioni biomolecolari con <strong>di</strong>spositivi ottici (J2, J4, J5, J8, J14, J16,<br />
J17, J19, J20, J23)<br />
Una volta messa a punto la procedura <strong>di</strong> mo<strong>di</strong>fica chimica della superficie del silicio<br />
poroso sono stati condotti esperimenti <strong>di</strong> immobilizzazione delle sonde biologiche (DNA<br />
singola eliche, proteine, enzimi). Inizialmente sono stati utilizzati campioni marcati con<br />
fluorofori, al fine <strong>di</strong> verificare, con l’ausilio <strong>di</strong> un macroscopio a fluorescenza, l’efficienza e<br />
la stabilità della funzionalizzazione della superficie <strong>di</strong> silicio poroso.<br />
L’obiettivo finale del progetto è la realizzazione <strong>di</strong> un sistema ottico <strong>di</strong> analisi delle<br />
interazioni biologiche che non richieda l’uso <strong>di</strong> marcatori: le interazioni specifiche sondaligando<br />
vengono stu<strong>di</strong>ate me<strong>di</strong>ante riflettometria in luce bianca in strutture fotoniche con<br />
precisa risposta ottica, funzionalizzate con sonde non marcate. L’incremento dell’in<strong>di</strong>ce <strong>di</strong><br />
rifrazione me<strong>di</strong>o del <strong>di</strong>spositivo, dovuto all’introduzione <strong>di</strong> materiale organico nei pori del<br />
<strong>di</strong>spositivo, che ivi permane anche dopo lavaggi in quanto agganciato alla rispettiva<br />
sonda, è rilevabile sperimentalmente da uno spostamento verso lunghezze d’onda<br />
maggiori dello spettro in riflessione ottenuto facendo incidere sul <strong>di</strong>spositivo un fascio <strong>di</strong><br />
luce collimato, proveniente da una sorgente ad ampio spettro.<br />
La prima parte della sperimentazione è consistita nel determinare la concentrazione<br />
ottimale delle sonde per ottenere un ricoprimento uniforme della superficie <strong>di</strong> silicio<br />
poroso. A tal fine sono state utilizzate sonde marcate con un opportuno gruppo cromoforo<br />
(Fluoresceina CY3.5, o Rhodamina), tramite la misura dell’intensità <strong>di</strong> fluorescenza in<br />
funzione della concentrazione <strong>di</strong> sonda, prima e dopo tre lavaggi nelle soluzioni buffer e<br />
due <strong>di</strong>alisi della durata <strong>di</strong> una notte. In questo modo, sono state stabilite le concentrazioni<br />
<strong>di</strong> saturazione. Passo successivo è, dopo aver legato covalentemente la sonda non<br />
marcata alla superficie del silicio poroso, la rivelazione dell’interazione molecolare con il<br />
rispettivo ligando misurando lo spostamento dello spettro <strong>di</strong> riflessione del <strong>di</strong>spositivo dopo<br />
il processo <strong>di</strong> interazione che avviene <strong>di</strong>rettamente su chip a temperatura ambiente.<br />
Gli esperimenti <strong>di</strong> riconoscimento molecolare sono stati ripetuti con <strong>di</strong>verse concentrazioni<br />
<strong>di</strong> ligando. È stato così possibile ottenere curve dose-riposta (Figura 13b) calcolando, per<br />
ogni concentrazione <strong>di</strong> ligando, la <strong>di</strong>fferenza <strong>di</strong> cammino ottico causata dalla variazione<br />
dell’in<strong>di</strong>ce <strong>di</strong> rifrazione me<strong>di</strong>o del silicio poroso dovuta all’introduzione del materiale<br />
biologico (Figura 13a). Per verificare che la variazione dello spettro <strong>di</strong> riflessione non fosse<br />
dovuta ad interazioni aspecifiche dei ligan<strong>di</strong> nella struttura porosa, è stata ripetuta la<br />
procedura sperimentale utilizzando dei controlli negativi, nei quali non si registrava alcuna<br />
variazione dello spettro <strong>di</strong> riflessione.<br />
xi
Reflectivity (a.u.)<br />
1.4<br />
1.2<br />
1.0<br />
0.8<br />
0.6<br />
0.4<br />
0.2<br />
ss-DNA<br />
ss-DNA+c-DNA 6μM<br />
ss-DNA+c-DNA 15μM<br />
ss-DNA+c-DNA 60μM<br />
0.0<br />
1150 1200 1250 1300 1350<br />
Wavelength (nm)<br />
Δλ (nm)<br />
35<br />
30<br />
25<br />
20<br />
15<br />
10<br />
5<br />
LOD cDNA = 350nM<br />
S cDNA =0.85 (4) nm/μM<br />
0<br />
0 20 40 60 80<br />
c-DNA concentration (μM)<br />
A B<br />
Figura 13: A) spostamento dello spettro <strong>di</strong> riflessione dovuto all’interazione ssDNA-cDNA. B) curva Doseresposta<br />
in funzione della concentrazione <strong>di</strong> cDNA.<br />
È stato possibile determinare per ogni coppia sonda-ligando i limiti <strong>di</strong> rivelazione (LOD) del<br />
nostro <strong>di</strong>spositivo e la sensibilità (S) <strong>di</strong> misura dello stesso.<br />
Di seguito è riportata la tabella con i valori <strong>di</strong> concentrazione per le sonde ed i valori <strong>di</strong><br />
LOD ed S calcolati secondo la metodologia IUPAC (IUPAC Compen<strong>di</strong>um of Analytical<br />
Nomenclature, Definitive Rules, 3d E<strong>di</strong>tion, Section 10.3.3.3.14 “Limit of Detection”,1997).<br />
Sonda Concentrazione Ligando Controllo<br />
negativo<br />
LOD S<br />
DNA 60 μM cDNA ncDNA 350nM 0.85 nm/μM<br />
GlnBP 1mM Gln Glucosio 9nM 3*10 -2 nm/nM<br />
GlnBP 1mM Glia<strong>di</strong>na Prolamina del riso 670pM 448 nm/μM<br />
TMBP 7.5 μM Glucosio Gln 10 μM 0.03 nm/μM<br />
Integrazione del silicio poroso in un lab-on-chip (J1, J7, J9, J13, J15)<br />
La realizzazione <strong>di</strong> sistemi multifunzionali complessi ove circuiti e componenti elettronici<br />
ed ottici, ed elementi <strong>di</strong> sensing coesistono e operano sinergicamente, è <strong>di</strong> importanza<br />
strategica in <strong>di</strong>versi settori (produzione-automazione, trasporti, me<strong>di</strong>cina, biochimica,<br />
sicurezza e ambiente). La <strong>di</strong>ffusione <strong>di</strong> tali sistemi multifunzionali, in grado <strong>di</strong> operare nei<br />
più svariati ambienti, anche potenzialmente pericolosi, e che svolgono molteplici funzioni<br />
<strong>di</strong> monitoraggio e <strong>di</strong> <strong>di</strong>agnostica, presuppone una drastica riduzione dei costi ed un<br />
significativo incremento della loro affidabilità.<br />
L’attività <strong>di</strong> ricerca relativa a quest’ultima fase del lavoro <strong>di</strong> <strong>tesi</strong> è stata volta alla<br />
progettazione, realizzazione e caratterizzazione <strong>di</strong> microsistemi dotati <strong>di</strong> funzionalità <strong>di</strong><br />
trasduzione ottica, basati sulle tecnologie <strong>di</strong> fabbricazione del silicio e dei materiali<br />
tecnologicamente compatibili. Poiché la produzione <strong>di</strong> microsensori ottici, integrabili su<br />
chip con altri componenti optoelettronici e microflui<strong>di</strong>ci, necessita della compatibilità con i<br />
processi ed i circuiti microelettronici, è stato affrontato il problema dell’integrazione<br />
dell’elemento sensibile in silicio poroso in un microsistema ottico in cui tutte le funzioni<br />
(alimentazione, analisi, spurgo, e così via) per il funzionamento del trasduttore sono<br />
miniaturizzate in un micro<strong>di</strong>spositivo.<br />
Per la realizzazione del <strong>di</strong>spositivo sono state ottimizzate alcune delle tecniche classiche<br />
della microelettronica e delle microlavorazioni, quali gli attacchi anisotropi in KOH, la<br />
fotolitografia e gli attacchi aci<strong>di</strong> del vetro.<br />
È stato inoltre definito un processo <strong>di</strong> incollaggio ano<strong>di</strong>co “ano<strong>di</strong>c bon<strong>di</strong>ng” per sigillare il<br />
silicio con vetro trasparente compatibilmente con le caratteristiche del silicio poroso.<br />
Il silicio poroso infatti per la sua morfologia spugnosa è abbastanza fragile dal punto <strong>di</strong><br />
vista meccanico; inoltre la sua superficie idrogenata tende ad ossidarsi facilmente alle<br />
temperature tipiche dell’“ano<strong>di</strong>c bon<strong>di</strong>ng” (400-800 °C) e poiché l’ossido cresce per un<br />
xii
60% al <strong>di</strong> sopra della superficie occupa parte dei pori con conseguente per<strong>di</strong>ta <strong>di</strong><br />
superficie specifica e quin<strong>di</strong> <strong>di</strong> sensibilità.<br />
Una sperimentazione de<strong>di</strong>cata ha permesso <strong>di</strong> stabilire la temperatura minima (250 °C) ed<br />
il tempo necessario (2 min.) per ottenere un processo tale da sigillare efficacemente il<br />
<strong>di</strong>spositivo con il vetro. Lo schema del <strong>di</strong>spositivo realizzato, insieme con il flusso <strong>di</strong><br />
processo per la fabbricazione, è mostrato in Figura 14a, mentre in Figura 14b è riportata<br />
una fotografia del microsistema su banco ottico.<br />
Figura 14: a) Schema del microsistema ottico e flusso <strong>di</strong> processo per la fabbricazione; b) fotografia del<br />
microsistema su banco ottico<br />
L’utilizzo <strong>di</strong> un microsistema <strong>di</strong> questo tipo riduce enormemente i tempi <strong>di</strong> risposta del<br />
<strong>di</strong>spositivo la cui durata è essenzialmente dovuta alla <strong>di</strong>ffusione delle specie biochimiche<br />
da analizzare all’interno della camera <strong>di</strong> test dove si trova l’elemento sensibile. Risultati<br />
ancora migliori possono essere ottenuti alimentando il sensore con un attuatore<br />
pneumatico: il fluido da analizzare, trasportato da un flusso inerte, arriva sul trasduttore<br />
solo in determinati intervalli <strong>di</strong> tempo che possono essere resi piccoli a piacere.<br />
La realizzazione <strong>di</strong> una matrice <strong>di</strong> elementi trasduttori richiede la messa a punto <strong>di</strong> un<br />
processo fotolitografico ed uno <strong>di</strong> attacco per la definizione delle strutture micrometriche.<br />
La fotolitografia (come si vede in Figura 15) è quel processo che consente il trasferimento<br />
del <strong>di</strong>segno da realizzare sul silicio da una maschera ad un sottile strato <strong>di</strong> materiale<br />
sensibile alla ra<strong>di</strong>azione (resist), che ricopre la superficie del substrato.<br />
Figura 15: Schema del processo fotolitografico<br />
Una delle tecniche base per la microlavorazione del silicio è quella che utilizza soluzioni<br />
acquose alcaline. Comunemente l’attacco umido anisotropo include soluzioni acquose <strong>di</strong><br />
KOH, miscele <strong>di</strong> KOH e IPA (alcol isopropilico), oppure EDP (Etilen<strong>di</strong>ammina e<br />
Pirocatecolo). Tali soluzioni rendono possibile l’etching del silicio, il quale viene attaccato<br />
con <strong>di</strong>fferenti velocità nelle varie <strong>di</strong>rezioni cristallografiche.<br />
xiii
Di fondamentale importanza è la scelta della maschera fotolitografica da utilizzare durante<br />
l’attacco, tenendo conto anche della velocità <strong>di</strong> incisione da parte della soluzione alcalina<br />
sullo strato <strong>di</strong> mascheramento. Per gli attacchi in KOH si usano generalmente maschere <strong>di</strong><br />
nitruro <strong>di</strong> silicio, che non sono attaccabili, <strong>di</strong>ossido <strong>di</strong> silicio che sono incise con una<br />
velocità <strong>di</strong> 1.4 nm/min e silicio <strong>di</strong> tipo p che offrono una riduzione sul silicio lievemente<br />
drogato della velocità <strong>di</strong> incisione tra 10:1 e 100:1, a seconda delle concentrazioni delle<br />
soluzioni e dalla temperatura del processo.<br />
Per la realizzazione <strong>di</strong> array <strong>di</strong> microvaschette con un volume <strong>di</strong> 10 μl, si è utilizzato una<br />
maschera <strong>di</strong> nitruro <strong>di</strong> silicio spessa 535 nm ed una soluzione acquose <strong>di</strong> KOH al 33% in<br />
peso alla temperatura <strong>di</strong> 80 ° C per 160 min.<br />
Il patterning, tuttavia, <strong>di</strong> campioni <strong>di</strong> PSi è fortemente limitato. La realizzazione del poroso<br />
richiede il contatto del campione con una soluzione HF in flusso <strong>di</strong> corrente, con<strong>di</strong>zioni<br />
molto aggressive per le maschere litografiche comunemente utilizzate. È stato, pertanto,<br />
messo a punto e caratterizzato un processo fotolitografico atto a produrre una maschera<br />
protettiva da potersi utilizzare per un attacco elettrochimico. La definizione dei parametri <strong>di</strong><br />
processo finalizzati alla realizzazione della maschera, dalla fotolitografia alla deposizione<br />
<strong>di</strong> film sottili, sono stati <strong>di</strong> estrema importanza, essendo questi dei processi tecnologici<br />
estremamente sensibili all’ambiente nel quale sono effettuati (variazioni anche minime<br />
nelle con<strong>di</strong>zioni <strong>di</strong> temperatura, pressione ed umi<strong>di</strong>tà e la presenza <strong>di</strong> particelle volatili non<br />
desiderate minano la riuscita dell’operazione), si è scelto <strong>di</strong> lavorare in clean room, un<br />
laboratorio ad ambiente controllato e depolverizzato.<br />
Partendo da lavori riportati in letteratura è stata deposta e poi fotolitografata una maschera<br />
<strong>di</strong> nitruro <strong>di</strong> silicio. Le maschere composte da specie al nitruro sono state scelte poiché<br />
offrono una buona resistenza all’attacco a fronte <strong>di</strong> un basso costo <strong>di</strong> produzione ed<br />
un’elevata risoluzione per quel che riguarda il design del substrato. Il campione è stato<br />
ricoperto da uno strato <strong>di</strong> nitruro <strong>di</strong> silicio, depositato tramite deposizione chimica in fase<br />
vapore assistita dal plasma (PECVD), come maschera per l’attacco elettrochimico. Lo<br />
stesso è stato poi processato al fine <strong>di</strong> ottenere il design desiderato. Sono stati messi a<br />
punto i parametri <strong>di</strong> processo per la deposizione regolando opportunamente flussi <strong>di</strong> silano<br />
e ammoniaca.<br />
Le maschere realizzate avevano 400 fori <strong>di</strong> 200μm <strong>di</strong> <strong>di</strong>ametro, spaziati tra <strong>di</strong> loro in<br />
entrambe le <strong>di</strong>rezioni <strong>di</strong> 600μm. È stato quin<strong>di</strong> condotto l’attacco elettrochimico per<br />
ottenere monostrati da 500 nm (spessore minimo per un’analisi all’ellissometro) con<br />
porosità del 50% utilizzando una J pari a 50mA/cm 2 . Una foto dell’array realizzato,<br />
acquisita al microscopio ottico è riportata in Figura 16.<br />
Figura 16: Immagine dell’array <strong>di</strong> monostrati in PSi acquisita al microscopio ottico: ingran<strong>di</strong>mento 1.25X<br />
L’analisi riflettometrica, come già descritto, prevede l’utilizzo <strong>di</strong> una sorgente <strong>di</strong> luce<br />
bianca, da far incidere perpen<strong>di</strong>colarmente sul campione, ed un analizzatore ottico <strong>di</strong><br />
spettro, al quale inviare la luce riflessa dal campione stesso. Questa connessione al<br />
campione avviene grazie ad una fibra ottica detta ‘ad Y’ l’unica che strutturalmente ci<br />
xiv
consente <strong>di</strong> inviare luce e convogliare quella riflessa, apparato e spettro ottico <strong>di</strong><br />
riflessione in Figura 17. La luce, in questo caso, è stata focalizzata sul campione<br />
utilizzando un obiettivo 20x, capace <strong>di</strong> <strong>di</strong>minuire la focale e restringere lo spot per poter<br />
analizzare campioni <strong>di</strong> poroso <strong>di</strong> <strong>di</strong>mensioni micrometriche.<br />
A B<br />
Figura 17: A)apparato sperimentale utilizzato per misurare gli spettri ottici <strong>di</strong> riflessione dei campioni<br />
dell’array; B) confronto tra lo spettro <strong>di</strong> riflessione simulato (rosso) e quello sperimentale (nero).<br />
Il <strong>di</strong>spositivo ottico <strong>di</strong> array in PSi così realizzato potrà essere applicato in <strong>di</strong>versi settori,<br />
dalla genomica, la post-genomica, la proteomica e le analisi molecolari finalizzate alla<br />
<strong>di</strong>agnostica ad alta automazione per la me<strong>di</strong>cina, la farmacogenomica e l'industria agroalimentare.<br />
Le soluzioni previste porteranno ad un miglioramento significativo del<br />
<strong>di</strong>spositivo in termini <strong>di</strong> sensibilità, riduzioni dei costi, velocità, pre<strong>di</strong>sposizione<br />
all'automazione, affidabilità e ripetibilità, in modo da poter monitorare<br />
contemporaneamente <strong>di</strong>versi analiti in miscele complesse.<br />
xv
Chapter 1<br />
Porous Silicon
Introduction<br />
The development of biochips is a major thrust of the rapidly growing biotechnology<br />
industry, which encompasses a very <strong>di</strong>verse range of research efforts inclu<strong>di</strong>ng<br />
genomics, proteomics, computational biology, and pharmaceuticals, among other<br />
activities. Advances in these areas are giving scientists new methods for unraveling<br />
the complex biochemical processes occurring inside cells, with the larger goal of<br />
understan<strong>di</strong>ng and treating human <strong>di</strong>seases. At the same time, the semiconductor<br />
industry has been stea<strong>di</strong>ly perfecting the science of microminiaturization. The<br />
merging of these two fields in recent years has enabled biotechnologists to begin<br />
packing their tra<strong>di</strong>tionally bulky sensing tools into smaller and smaller spaces, onto<br />
so-called biochips. These chips are essentially miniaturized laboratories that can<br />
perform hundreds or thousands of simultaneous biochemical reactions. Biochips<br />
enable researchers to quickly screen large numbers of biological analytes for a<br />
variety of purposes, from <strong>di</strong>sease <strong>di</strong>agnosis to detection of bioterrorism agents.<br />
The development of biochips has a long history, starting with early work on the<br />
underlying sensor technology. One of the first portable, chemistry-based sensors was<br />
the glass pH electrode, invented in 1922 by Hughes [1]. Measurement of pH was<br />
accomplished by detecting the potential <strong>di</strong>fference developed across a thin glass<br />
membrane selective to the permeation of hydrogen ions; this selectivity was achieved<br />
by exchanges between H + and SiO sites in the glass. The basic concept of using<br />
exchange sites to create permselective membranes was used to develop other ion<br />
sensors in subsequent years. For example, a K + sensor was produced by<br />
incorporating valinomycin into a thin membrane [2]. Over thirty years elapsed before<br />
the first true biosensor (i.e. a sensor utilizing biological molecules) emerged. In 1956,<br />
Leland Clark published a paper on an oxygen sensing electrode [3]. This device<br />
became the basis for a glucose sensor developed in 1962 by Clark and colleague<br />
Lyons which utilized glucose oxidase molecules embedded in a <strong>di</strong>alysis membrane<br />
[4]. The enzyme functioned in the presence of glucose to decrease the amount of<br />
oxygen available to the oxygen electrode, thereby relating oxygen levels to glucose<br />
concentration. This and similar biosensors became known as enzyme electrodes,<br />
and are still in use today.<br />
In 1953, Watson and Crick announced their <strong>di</strong>scovery of the now familiar double helix<br />
structure of DNA molecules and set the stage for genetics research that continues to<br />
the present day [5]. The development of sequencing techniques in 1977 by Gilbert<br />
and Sanger [6] (working separately) enabled researchers to <strong>di</strong>rectly read the genetic<br />
codes that provide instructions for protein synthesis. This research showed how<br />
hybri<strong>di</strong>zation of complementary single oligonucleotide strands could be used as a<br />
basis for DNA sensing. Two ad<strong>di</strong>tional developments enabled the technology used in<br />
modern DNA-based biosensors. First, in 1983 Kary Mullis invented the polymerase<br />
chain reaction (PCR) technique[5], a method for amplifying DNA concentrations. This<br />
<strong>di</strong>scovery made possible the detection of extremely small quantities of DNA in<br />
samples. Second, in 1986 Hood and coworkers devised a method to label DNA<br />
molecules with fluorescent tags instead of ra<strong>di</strong>olabels [7], thus enabling hybri<strong>di</strong>zation<br />
experiments to be observed optically.<br />
The rapid technological advances of the biochemistry and semiconductor fields in the<br />
1980s led to the large scale development of biochips in the 1990s. At this time, it<br />
became clear that biochips were largely a "platform" technology which consisted of<br />
several separate, yet integrated components. Figure 1 shows the makeup of a typical<br />
biochip platform. The actual sensing component (or "chip") is just one piece of a<br />
complete analysis system.<br />
1
Figure 1: schematic of a biosensor<br />
Transduction must be done to translate the actual sensing event (DNA bin<strong>di</strong>ng,<br />
oxidation/reduction, etc.) into a format understandable by a computer (voltage, light<br />
intensity, mass, etc.), which then enables ad<strong>di</strong>tional analysis and processing to<br />
produce a final, human-readable output. The multiple technologies needed to make a<br />
successful biochip — from sensing chemistry, to microarraying, to signal processing<br />
— require a true multi<strong>di</strong>sciplinary approach, making the barrier to entry steep. One of<br />
the first commercial biochips was introduced by Affymetrix. Their "GeneChip"<br />
products contain thousands of in<strong>di</strong>vidual DNA sensors for use in sensing defects, or<br />
single nucleotide polymorphisms (SNPs), in genes such as p53 (a tumor suppressor)<br />
and BRCA1 and BRCA2 (related to breast cancer) [8]. The chips are produced using<br />
microlithography techniques tra<strong>di</strong>tionally used to fabricate integrated circuits (see<br />
below).<br />
Biochips are a platform that require, in ad<strong>di</strong>tion to microarray technology, transduction and signal<br />
processing technologies to output the results of sensing experiments<br />
Today, a large variety of biochip technologies are either in development or being<br />
commercialized. Numerous advancements continue to be made in sensing research<br />
that enable new platforms to be developed for new applications. Cancer <strong>di</strong>agnosis<br />
through DNA typing is just one market opportunity. A variety of industries currently<br />
desire the ability to simultaneously screen for a wide range of chemical and biological<br />
agents, with purposes ranging from testing public water systems for <strong>di</strong>sease agents<br />
to screening airline cargo for explosives. Pharmaceutical companies wish to<br />
combinatorially screen drug can<strong>di</strong>dates against target enzymes. To achieve these<br />
ends, DNA, RNA, proteins, and even living cells are being employed as sensing
me<strong>di</strong>ators on biochips. Numerous transduction methods can be employed inclu<strong>di</strong>ng<br />
surface plasmon resonance, fluorescence, and chemiluminescence. The particular<br />
sensing and transduction techniques chosen depend on factors such as price,<br />
sensitivity, and reusability.<br />
Optical transduction is more and more used since photonic devices could be small,<br />
lightweight and thus portable due to the integrability of all optical components.<br />
Furthermore, optical devices do not require electric contacts. Fluorescence is by far<br />
the most used optical signalling method but a wide-use sensor can not be limited by<br />
the labelling of the probe nor the analyte, since this step is not always possible.<br />
Reagentless optical biosensors are monitoring devices which can detect a target<br />
analyte in a heterogeneous solution without the ad<strong>di</strong>tion of anything than the sample.<br />
In the fields of genomics and proteomics this is a straightforward advantage since it<br />
allows real-time readouts and, thus, very high throughoutputs analysis. A label free<br />
optical biosensor can be realized by integrating the biological probe with a signalling<br />
material which <strong>di</strong>rectly transduces the molecular recognition event into an optical<br />
signal.<br />
There are many advantages of optical sensing. Optical sensing schemes are<br />
sensitive. By measuring <strong>di</strong>fferences in wavelengths or times, optical signals can be<br />
rea<strong>di</strong>ly multiplexed. Some optical techniques, such as fluorescence, have intrinsic<br />
amplification in which a single label can lead to a million photons. In ad<strong>di</strong>tion, some<br />
optical techniques are zero or “black” background in which the only source of signal<br />
is due to the presence of the species being measured, thereby enabling high<br />
sensitivity measurements. Finally, optical signals travel in an open path—no wires or<br />
other transmitting conduits are necessary. This feature enables remote<br />
measurements to be made.<br />
There are a variety of optical sensing transduction mechanisms inclu<strong>di</strong>ng:<br />
• Luminescence<br />
• Fluorescence—intensity, lifetime, polarization<br />
• Phosphorescence<br />
• Fluorescence resonance energy transfer<br />
• Absorbance<br />
• Scattering<br />
• Raman-surface enhanced, resonance<br />
• Surface plasmon resonance<br />
• Interference<br />
Each of these techniques may have several variations. Another rubric for classifying<br />
optical sensors revolves around the categories of materials, surfaces, and arrays. In<br />
the materials field, there is a burgeoning effort in developing new materials for<br />
performing optical sensing. Such materials encompass work in the fields of polymers,<br />
ceramics, and semiconductors as well as more conventional inorganic and organic<br />
materials. These materials can be used as new recognition agents, as supports for<br />
immobilizing sensing materials, as materials for concentrating analytes, and as<br />
intrinsic optical sensors with integrated functionality inclu<strong>di</strong>ng bin<strong>di</strong>ng and signal<br />
transduction. This latter goal of creating materials that integrate bin<strong>di</strong>ng with optical<br />
signal transduction is a major area of interest and investigation.<br />
In the surfaces area, a host of new phenomena are being <strong>di</strong>scovered and employed<br />
for biosensing.<br />
It should be noted that there is a growing trend aimed at creating nanoscale devices.<br />
3
In the past two decades, the biological and me<strong>di</strong>cal fields have seen great advances<br />
in the development of biosensors and biochips for characterizing and quantifying<br />
biomolecules. The biochemistry-based issues may include the characteristics of the<br />
recognition molecules in terms of specificity, bin<strong>di</strong>ng avi<strong>di</strong>ty, reversibility of bin<strong>di</strong>ng,<br />
stability under con<strong>di</strong>tions of use and storage, functionality after immobilization, signal<br />
generation, biochemical signal amplification, <strong>di</strong>scrimination of specific from<br />
nonspecific bin<strong>di</strong>ng events, reaction time, availability and cost.<br />
Using bioreceptors from biological organisms or receptors that have been patterned<br />
after biological systems, scientists have developed new methods of biochemical<br />
analysis that exploit the high selectivity of the biological recognition systems [9].<br />
Arrays are being employed for an ever-increasing number of analytes. Ten years<br />
ago, the main focus on sensors was specificity and sensitivity. The need to measure<br />
multiple parameters was solved by bundling several sensors together in order to<br />
multiplex them. Today, arrays with tens of thousands, and even hundreds of<br />
thousands of features are commonplace in the DNA microarray field. Most such<br />
microarrays use fluorescence signals as the transduction mechanism. The ability to<br />
measure so many analytes simultaneously has revolutionized the thinking in sensing<br />
in general and optical sensing in particular. It is no longer sufficient in most<br />
applications to measure a single analyte as such a measurement captures only part<br />
of the information about a complex real-world sample. A variety of interesting and<br />
novel methods for preparing arrays have been developed for a wide variety of<br />
applications. Arrays also offer the ability to make replicate measurements to minimize<br />
false responses and to employ less selective sensing schemes coupled with<br />
intelligent processing to remove the requirement for absolute specificity.<br />
Silicon technology is already pervasive in our everyday lives. Silicon 'chips' often help<br />
us shop, cook and relax. They are involved in many forms of transport and control<br />
most forms of communication at the workplace.<br />
The choice of this thesis for the support material and contemporary for the transducer<br />
element fallen on porous silicon.<br />
• FABRICATION AND CHARACTERIZATION OF POROUS SILICON (PSi)<br />
MATERIALS<br />
The current interest in porous silicon (PSi) results from the demonstration of efficient<br />
visible photoluminescence of this material, first reported by Prof. Canham in 1990.<br />
However, PSi is not a new material: it was first reported over 40 years ago by Uhlir.<br />
During stu<strong>di</strong>es of the electropolishing of silicon in aqueous hydrofluoric acid (HF), he<br />
observed that the surface often became black, brown or red. More detailed stu<strong>di</strong>es<br />
were performed by Turner and Archer, but these films were not recognized as being<br />
PSi. It was Watanabe et al. who first reported their porous nature. Porous silicon,<br />
then, has been investigated for applications in microelectronics, optoelectronics,<br />
chemical and biological sensors, and biome<strong>di</strong>cal devices.<br />
The in vivo use of porous silicon was first promoted by Leigh Canham, who<br />
demonstrated its resorbability and biocompatibility in the mid 1990s. Subsequently,<br />
PSi or porous SiO2 (prepared from PSi by oxidation) host matrices have been<br />
employed to demonstrate in vitro release of the steroid dexamethasone, ibuprofen<br />
and many other drugs. The first report of drug delivery from PSi across a cellular<br />
barrier was performed with insulin, delivered across monolayers of Caco-2 cells. An<br />
excellent review of the potential for use of PSi in various drug delivery applications<br />
has recently appeared.
The basic method to fabricate PSi is electrochemical <strong>di</strong>ssolution of single crystalline<br />
silicon wafer in a hydrofluori<strong>di</strong>c acid electrolyte solution. This is obtained by<br />
monitoring either the ano<strong>di</strong>c current (galvanostatic) or voltage (potentiostatic).<br />
In general, constant current is preferable as it allows better control of the porosity and<br />
thickness. It also provides better reproducibility from sample to sample.<br />
Pore morphology and pore size can be varied by controlling the current density, the<br />
type and the concentration of dopant, the crystalline orientation of the wafer, and the<br />
electrolyte concentration in order to form macro-, meso-, and micropores. Pore sizes<br />
ranging from 1 nm to a few microns can be prepared.<br />
In the simplest setup to fabricate PSi, plates of silicon and Pt are <strong>di</strong>pped into HF<br />
solution and an etching current is applied between these electrodes. The porous<br />
layer is formed in the surface of the silicon wafer, which is used as a positive anode.<br />
Usually a cathode is made of platinum and the fabrication cell has to be made of HFresistant<br />
material, for example Teflon. Dilute HF solution or ethanolic HF are<br />
generally used as electrolytes. Ethanol is added to reduce formation of hydrogen<br />
bubbles and to improve the electrolyte penetration in the pores and, thus, uniformity<br />
of the PSi layer.<br />
The mechanism of pore formation is not generally agreed upon, but it is thought to<br />
involve a combination of electronic and chemical factors. The type of dopant in the<br />
original silicon wafer is important because it determines the availability of valence<br />
band holes that are the key oxi<strong>di</strong>zing equivalents in the reaction shown in Figure 2.<br />
Figure 2: schematic of the etch cell used to prepare porous silicon. The electrochemical half-reactions<br />
are shown.<br />
In general the relationships of dopant to morphology can be segregated into four<br />
groups based on the type and concentration of the dopant: n-type, p-type, highly<br />
doped n-type, and highly doped p-type. By “highly doped”, the dopant levels at which<br />
the conductivity behavior of the material is more metallic than semiconducting.<br />
The n-type silicon substrate has to be illuminated during ano<strong>di</strong>zation to generate<br />
photo-excited holes on the surface. For this silicon type wafers, with a relatively<br />
moderate doping level, exclusion of valence band holes from the space charge<br />
region determines the pore <strong>di</strong>ameter. Quantum confinement effects are thought to<br />
limit pore size in moderately p-doped material. For both dopant types the reaction is<br />
crystal face selective, with the pores propagating primarily in the <strong>di</strong>rection of<br />
the single crystal.<br />
5
A simplified mechanism for the chemical reaction in reported in Figure 3 [10].<br />
Figure 3: mechanism of silicon oxidation during the formation of porous silicon.<br />
Application of ano<strong>di</strong>c current oxi<strong>di</strong>zes a surface silicon atom, which is then attacked<br />
by fluoride. The net process is a 4 electron oxidation, but only two equivalents are<br />
supplied by the current source. The other two equivalents come from reduction of<br />
protons in the solution by surface SiF2 species. Pore formation occurs as Si atoms<br />
are removed in the form of SiF4, which reacts with two equivalents of F − in solution to<br />
form SiF6 2− .<br />
The porosity and the thickness of the PSi layers are among the most important<br />
parameters which characterize porous silicon. The porosity is defined as the fraction<br />
of void within the PSi layer and can be determined easily by weight measurements.<br />
The virgin wafer is first weighted before ano<strong>di</strong>zation (m1), then just after ano<strong>di</strong>zation<br />
(m2) and finally after <strong>di</strong>ssolution of the whole porous layer in a molar NaOH aqueous<br />
solution (m3). Uniform and rapid stripping in the NaOH solution is obtained when the<br />
PSi layer is covered with a small amount of ethanol which improves the infiltration of<br />
aqueous NaOH in the pores. The porosity is given simply by the following equation:<br />
% <br />
<br />
1 2<br />
1 3<br />
From these measured masses, it is also possible to determine the thickness of the<br />
layer accor<strong>di</strong>ng to the following formula:<br />
1 3<br />
<br />
where d is the density of bulk silicon and S the wafer area exposed to HF during<br />
ano<strong>di</strong>zation.<br />
The porosity of a growing porous Si layer is proportional to the current density being<br />
applied, and it typically ranges between 40 and 80%. Pores form only at the<br />
Si/porous Si interface, and once formed, the morphology of the pores does not<br />
change significantly for the remainder of the etching process.<br />
However, the porosity of a growing layer can be altered by changing the applied<br />
current. The film will continue to grow with this new porosity until the current<br />
changes. This feature allows the construction of layered nanostructures simply by
modulating the applied current during an etch. For example, one <strong>di</strong>mensional<br />
photonic crystals consisting of a stack of layers with alternating refractive index can<br />
be prepared by perio<strong>di</strong>cally modulating the current during an etch.<br />
The ability to easily tune the pore sizes and volumes during the electrochemical etch<br />
is a unique property of PSi that is very useful for biological applications. Other porous<br />
materials generally require a more complicated design protocol to control pore size,<br />
and even then, the available pore sizes tend to span a limited range. With<br />
electrochemically prepared PSi, control over porosity and pore size is obtained by<br />
adjusting the current settings during the etch. Typically, larger current density<br />
produces larger pores. Large pores are desirable when incorporating sizable<br />
molecules or drugs within the pores.(Figure 4)<br />
Figure 4: SEM images of typical porous silicon morphologies.<br />
• Stain etching<br />
Although the electrochemical ano<strong>di</strong>zation is commonly used in the fabrication of PSi,<br />
several other fabrication methods have been introduced. One of these methods in a<br />
so-called stain etching. The term stain etching refers to the brownish or red<strong>di</strong>sh color<br />
of the film of porous Si that is generated on a crystalline silicon material subjected to<br />
the process [11]. The stain films are produced by immersion of silicon substrate in HF<br />
solution without any electrical bias. This method is even simpler than the previously<br />
presented ano<strong>di</strong>zation. However, the control of the porosity, layer thickness and pore<br />
size of PSi is quite limited. Stain etching in fact is less reproducible than the<br />
electrochemical process, although recent advances have improved the reliability of<br />
the process substantially [12]. Furthermore, stain etching cannot be used to prepare<br />
stratified structures such as double layers or multilayered photonic crystals. However,<br />
PSi powders prepared by stain etch are now commercially available<br />
(http://vestaceramics.net), and a few ad<strong>di</strong>tional vendors are poised to enter the<br />
market. For the biome<strong>di</strong>cally inclined researcher this eliminates the need to set up a<br />
complicated and hazardous electrochemical etching system, and it should stimulate<br />
the growth of the field.<br />
• PSi Photosynthesis<br />
Another zero-current etching method is photosynthesis. In this approach, silicon<br />
substrate is merely illuminated with intense visible light in a HF solution, without<br />
electrochemical bias. This is a very interesting method since it allows the fabricating<br />
of PSi microstructure without a mask, which is usually used in lithographic methods.<br />
7
After the etching, the next step to be considered is drying of the PSi material, if a<br />
liquid electrolyte has been used. Drying, if the electrolyte in the pores is allowed to<br />
evaporate at atmospheric pressure and temperature, may induce cracking and<br />
shrinkage. Especially the higher, brittle porosity structures (>75%) often do not have<br />
mechanical strength enough to survive the electrolyte evaporation without significant<br />
degradation in the structure. This is because the formed liquid vapour interface can<br />
generate large capillary stresses. In most cases (porosity
PSi optical sensors are based on changes of photoluminescence or reflectivity when<br />
exposed to the target analytes which substitute the air into the PSi pores. The effect<br />
depends on the chemical and physical properties of each analyte, so that the sensor<br />
can be used to recognize the pure substances. Due to the sensing mechanism, these<br />
kinds of devices are not able to identify the components of a complex mixture. In<br />
order to enhance the sensor selectivity through specific interactions, we have to<br />
chemically or physically mo<strong>di</strong>fy the PSi surface. The common approach is to create a<br />
covalent bond between the porous silicon surface and the biomolecules which<br />
specifically recognize the target analytes [29].<br />
The reliability of the biosensor strongly depends on the functionalization process:<br />
how fast, simple, homogenous and repeatable it is. This step is also very important<br />
for the stability of the sensor.<br />
Imme<strong>di</strong>ately after silicon electrochemically etched, the surface is covered with<br />
reactive hydride species. These chemical functionalities provide a versatile starting<br />
point for various reactions that determine the <strong>di</strong>ssolution rate in aqueous me<strong>di</strong>a,<br />
allow the attachment of biological probes. The two most important mo<strong>di</strong>fication<br />
reactions are chemical oxidation and grafting of Si-C species. In this way the<br />
substitution of the Si-H bonds with these ones guarantees a much more stable<br />
surface from the thermodynamic point of view.<br />
The integration of sensitive elements into complex microsystems, i.e. lab on chip or<br />
micro-total-analysis systems, is of straightforward interest, especially for the<br />
miniaturisation of each component. This step is never a trivial one: the fabrication<br />
process often requires high temperatures and mechanical stresses which can<br />
damage or even destroy the bio- or chemo- sensitive transducer structures.<br />
The ano<strong>di</strong>c bon<strong>di</strong>ng (AB) is a standard IC fabrication technique which is widely used<br />
in microflui<strong>di</strong>c due to a wide spectrum of advantages among which the hermetic<br />
sealing [30]. Due to the good bon<strong>di</strong>ng quality, glass transparency, technological<br />
cleanness and high passivity to most of chemicals and biological substances, AB is<br />
also commonly exploited in the fabrication of lab-on-chip devices. The compatibility of<br />
this primary integration technique with highly deman<strong>di</strong>ng sensing micro-components<br />
plays a basic role in view of the realization of lab-on-chip devices.<br />
Testing and demonstrating the porous silicon capabilities as a useful functional<br />
material in the optical transduction of biochemical interactions is only the first action<br />
in the realisation of an optical biochip based on this nanostructured material. In this<br />
case, all the fabrication processes should be compatible with the utilisation of<br />
biological probes and the feasibility of such devices must be proven. This means that<br />
the standard integrated circuit micro-technologies should be mo<strong>di</strong>fied and adapted to<br />
this new field of application in order to preserve the stability of the transducer<br />
element and all the biological components (probes and targets). The porous silicon<br />
low cost technology could thus provide a link between the conventional CMOS<br />
technology and the photonic devices in the realization of the so-called smart sensors<br />
and biochips. A very strong inter<strong>di</strong>sciplinary approach is required to match and<br />
resolve all the technological problems.<br />
The final objective of this thesis is to design and fabricate <strong>di</strong>fferent resonant optical<br />
structures, integrate them into a simple lab-on-chip devices based on the porous<br />
silicon nanotechnology. Some porous silicon surface mo<strong>di</strong>fication strategies in order<br />
to realize an optical biosensor were exploited; the target was the fabrication of<br />
sensitive label-free biosensors, which are highly requested for applications in high<br />
throughput drug monitoring and <strong>di</strong>seases <strong>di</strong>agnostics, unlabeled analytes require in<br />
fact easy sample preparation.<br />
9
Materials and Methods: Porous Silicon’ production calibration<br />
Today, the electrochemical etching is a standard method to fabricate nanostructured<br />
porous silicon: a proper choice of the applied current density, the electrolyte<br />
composition, and the silicon doping allow precise control over the morphology and,<br />
consequently, on the physical and chemical properties of the porous silicon structure.<br />
Computer controlled electrochemical etching processes are exploited for the<br />
realization of porous silicon films of controlled thickness and porosity (defined as the<br />
percentage of void in the silicon volume).<br />
Nanoporous, mesoporous and macroporous structures can be achieved, with pore<br />
size ranging from few nanometers up to microns. Moreover, since the etching<br />
process is self-stopping, it is possible to fabricate with a single run process multilayer<br />
stacks made of single layers of <strong>di</strong>fferent porosity. The <strong>di</strong>electric properties of each<br />
PSi layer, and in particular its refractive index n, can be namely modulated between<br />
those of crystalline silicon (n = 3.54, porosity = 0) and air (n= 1, porosity = 100 %); so<br />
that alternating high and low porosity layers, lot of photonic structures, such as<br />
Fabry-Perot interferometers, omni-<strong>di</strong>rectional Bragg reflectors, optical filters based on<br />
microcavities, and even complicated quasi-perio<strong>di</strong>c sequences (Thue-Morse) can be<br />
simply realized, as it is shown in Figure 5.<br />
Reflectivity (a.u)<br />
Reflectivity (a.u)<br />
1,0<br />
0,8<br />
0,6<br />
0,4<br />
0,2<br />
2,0<br />
1,5<br />
1,0<br />
0,5<br />
0,0<br />
600 800 1000 1200 1400 1600<br />
Wavelength (nm)<br />
monolayer<br />
optical microcavity<br />
800 1000 1200 1400<br />
Wavelength (nm)<br />
Reflectivity (a.u.)<br />
600 800 1000 1200 1400 1600<br />
Figure 5. Experimental reflectivity spectra of <strong>di</strong>fferent PSi optical structures.<br />
In this study Fabry-Perot single layer, Bragg mirrors, Thue Morse and optical<br />
microcavities have been used as sensors.<br />
The porous silicon layer was obtained by electrochemical etching in a HF-based<br />
solution at room temperature. The substrate used was a highly doped p + -silicon,<br />
oriented, 0.01 Ω cm resistivity, 400 μm thick. Before ano<strong>di</strong>zation the substrate<br />
was placed in HF solution to remove the native oxide. With this substrate we can<br />
obtain mesoporous layers, with a pore average <strong>di</strong>mension of about 50 nm. In figure 6<br />
SEM images are reported.<br />
1,0<br />
0,8<br />
0,6<br />
0,4<br />
0,2<br />
0,0<br />
Reflectivity (a.u)<br />
1.0<br />
0.8<br />
0.6<br />
0.4<br />
0.2<br />
0.0<br />
Bragg<br />
Wavelength (nm)<br />
Thue Morse S6<br />
600 800 1000 1200 1400 1600<br />
Wavelength (nm)
Figure 6. SEM images of <strong>di</strong>fferent PSi morphologies. The substrate p+ type is the chosen for this<br />
work.<br />
The first step of this thesis work was to set the parameters to realize and control the<br />
porous silicon fabrication process.<br />
It was checked for the best HF concentration in ethanol solution (HF : EtOH 1:1 and<br />
3:7). The calibration curve for establish the etch rate for each current density and the<br />
layer porosity were performed for each HF solution. In Figure 7 the two calibrations<br />
curves are reported to clarify.<br />
P (%)<br />
P (%)<br />
64<br />
62<br />
60<br />
58<br />
56<br />
54<br />
52<br />
50<br />
48<br />
46<br />
44<br />
42<br />
50 100 150 200 250 300 350 400 450 500 550<br />
76<br />
74<br />
72<br />
70<br />
68<br />
66<br />
64<br />
62<br />
60<br />
58<br />
56<br />
54<br />
n +<br />
J (mA/cm 2 )<br />
0 20 40 60 80 100 120 140 160 180<br />
J<br />
p +<br />
A<br />
B<br />
E. R. (nm/s)<br />
E.R. (nm/s)<br />
500<br />
450<br />
400<br />
350<br />
300<br />
250<br />
280<br />
260<br />
240<br />
220<br />
200<br />
180<br />
160<br />
100 200 300 400 500<br />
J (mA/cm 2 )<br />
0 20 40 60 80 100 120 140 160 180<br />
Figure 7. Silicon wafer: p+ type, orientation, 8-12 mW cm resistivity. A) Etching solution:<br />
HF/EtOH=50/50; B) Etching solution: HF/EtOH=30/70.<br />
n<br />
J<br />
p<br />
11
The morphology of the porous silicon layers was investigated by variable angle<br />
spectroscopic ellipsometer (UVISEL, Horiba-Jobin-Yvon) and scanning electron<br />
microscope (SEM-FEG Gemini 300, Carl Zeiss).<br />
Both the analytical techniques have showed the presence of a top layer on the<br />
porous silicon structure of thickness between 20 nm and 100 nm and porosity of 20-<br />
40 % due to hydrogen contamination of the silicon wafer [31]. Such film prevents not<br />
only the pores from filling but also any biochemical interaction with the hydrogenated<br />
porous silicon surface after the etching process. For sensing purposes it is therefore<br />
mandatory to avoid the formation of the parasitic film by thermal treating the wafer at<br />
300 °C in nitrogen atmosphere before the electrochemical etch [32].<br />
In figure 8 a cross section SEM image of the PSi top surface is reported. The thin low<br />
porosity layer on the top of the porous silicon device is well evident. In the thermal<br />
treated silicon wafer, this top layer is absent. The same result is confirmed by<br />
ellipsometric measurements.<br />
20 nm<br />
Figure 8: Cross section scanning electron microscopy image of a porous silicon layer.<br />
Due to the nanostructured nature of PSi, it is essential to improve the pore infiltration<br />
of biomolecular probes: to this aim, the device was optimized employing a strong<br />
base post etch process. By immersing the device in an aqueous ethanol solution,<br />
containing millimolar concentration of KOH, for 15 min. This treatment produces an<br />
increase of about 15-20% the porosity without affecting the optical quality of the<br />
device [33]. The presence of Si-H bonds on the porous silicon surface has been<br />
monitored by means of infrared spectroscopy with a Fourier transform spectrometer<br />
(FT-IR Nicolet Nexus) [33].<br />
The fresh etched porous silicon device shows the characteristic peaks of Si-H bonds<br />
at 910 cm -1 and 2100 cm -1 . The KOH post-etch treatment used to increase the<br />
porosity and to improve the pore infiltration by biological solutions, unfortunately<br />
removes most of these bonds and oxi<strong>di</strong>zes the PSi. To restore the Si-H bonds the<br />
porous silicon device were rinsed in a low concentration HF-based solution (5 mM)<br />
for 30 s.<br />
Figure 9 shows the FT-IR spectrum of a porous silicon microcavity after KOH and HF<br />
treatments. The Si-H bonds, removed by KOH, are present again and the silicon<br />
<strong>di</strong>oxide fingerprint (at 1050-1100 cm -1 ) <strong>di</strong>sappears.
Reflectivity (a.u.)<br />
160<br />
140<br />
120<br />
100<br />
80<br />
60<br />
40<br />
20<br />
2100 cm -1<br />
Si-H<br />
after KOH treatment<br />
after HF treatment<br />
silicon oxide<br />
913 cm -1<br />
Si-H<br />
3500 3000 2500 2000 1500 1000<br />
Wavenumber (cm -1 )<br />
Figure 9: FT-IR spectrum of a porous silicon microcavity after KOH and HF post-etch treatments.<br />
The experimental set-up (see Figure 10) to characterize the PSi devices from an<br />
optical point of view is very simple: a tungsten lamp (400 nm < λ < 1800 nm)<br />
inquired, through an optical fiber and a collimator, the sensor closed in a glass vial<br />
which can be fluxed by liquids or gases. The reflected beam is collected by an<br />
objective, coupled into a multimode fiber, and then <strong>di</strong>rected in an optical spectrum<br />
analyser (Ando, AQ6315A). The reflectivity spectra had been measured with a<br />
resolution of 0.2 nm.<br />
Figure 10: optical set-up<br />
In Figure 11 the optical reflectivity (as explicative example) spectra of a porous<br />
silicon layer as-etched and post KOH and HF treatments are reported, in order to<br />
show that the optical response of the devices is not affected by the strong chemical<br />
treatments.<br />
13
Reflectivity (u.a)<br />
1.0<br />
0.8<br />
0.6<br />
0.4<br />
0.2<br />
0.0<br />
sample as etched<br />
after KOH treatment<br />
after refresh in HF solution<br />
600 800 1000 1200 1400<br />
Wavelength (nm)<br />
Figure 11: Reflectivity optical spectra of the porous silicon layer after KOH and HF treatments.<br />
In this chapter, are presented some experimental results about sensing of chemical<br />
substances, flammable and toxic organic solvents and also some hydrocarbons,<br />
using several porous silicon optical structures with <strong>di</strong>fferent pores structures.<br />
References<br />
1. Hughes 1922. J. Amer. Chem. Soc. 44, 2860.<br />
2. J.B. Schulz, M. Weller and T. Klockgether, J. Neurosci. 1996, 16 : 4696–4706<br />
3. L.C. Clark, Trans. Am. Soc. Art. Int. Org. 1956 2, 41–52.<br />
4. L.C. Clark Jr. and C. Lyons, Ann. NY Acad. Sci. 1962, 102: 29–45.<br />
5. Kubisiak, T.L., Nelson, C.D., Nowak, J., and Friend, A.L. J. Sustain. For. 2000, 10, 69-78.<br />
6. Sanger F, Air GM, Barrell BG, Brown NL, Coulson AR, Fiddes CA, Hutchison CA, Slocombe<br />
PM, Smith M. Nature. 1977 Feb 24;265(5596):687–695<br />
7. L.M. Smith, J.Z. Sanders, R.J. Kaiser, P. Hughes, C. Dodd, C.R. Connell, C. Heiner, S.B.H.<br />
Kent and L.E. Hood. Nature (London) 321 1986, pp. 674–679<br />
8. Ward I.M. and Chen,J. 2001 J. Biol. Chem., 276:, 47759–47762<br />
9. Sadana A 2002 Engineering Biosensors (San Diego, CA: Academic)<br />
10. V. Lehmann, U. Gosele, Appl. Phys. Lett. 1991 58: 856–858.<br />
11. R.J. Archer, J. Phys. Chem. Solids 1960 14: 104–110.<br />
12. E.A. de Vasconcelos, E.F. da Silva, B. dos Santos, W.M. de Azevedo, J.A.K. Freire, Phys.<br />
Status Soli<strong>di</strong> A—Appl. Mat. 2005 202 : 1539–1542.<br />
13. Beckmann KH. 1965. Surf Sci 3:314–332.<br />
14. Canham LT, Aston R. 2001. Phys World 14: 27–31.<br />
15. Canham LT. 1995. Bioactive silicon structure fabrication through nanoetching techniques. Adv<br />
Mater 7:1033–1037.<br />
16. Canham LT. 1997. Biome<strong>di</strong>cal applications of porous silicon. In: Canham LT, e<strong>di</strong>tor.<br />
Properties of porous silicon. EMIS Datareview series No 18. London: INSPEC. pp 371–376.<br />
17. Canham LT, Reeves CL, King DO, Branfield PJ, Crabb JG, Ward MCL. 1996. Bioactive<br />
polycrystalline silicon. Adv Mater 8:850–852.<br />
18. Bayliss SC, Buckberry LD, Harris PJ, Tobin M. 2000. Nature of the silicon-animal cell<br />
interface. J Porous Mater 7:191–195.<br />
19. Bayliss SC, Heald R, Flecther R, Buckberry LD. 1999. The culture of mammalian cells on<br />
nanostructured silicon. Adv Mater 11:318–321.<br />
20. Karlsson LM, Tengvall P, Lundstrom I, Arwin H. 2003. Adsorption of human serum albumin in<br />
porous silicon gra<strong>di</strong>ents. Phys Stat Sol (a) 197: 326–330.<br />
21. Karlsson LM, Tengvall P, Lundstrom I, Arwin H. 2003. Penetration and loa<strong>di</strong>ng of human<br />
serum albumin in porous silicon layers with <strong>di</strong>fferent size and thicknesses. J Colloid Interface<br />
Sci 266:40–47.<br />
22. Prestidge CA, Barnes TJ, Mierczynska-Vasilev A, Kempson I, Ped<strong>di</strong>e F, Lau CH, Barnett C.<br />
2007. Peptide and protein loa<strong>di</strong>ng into porous silicon wafers. Phys Stat Sol (a) In press.
23. Anderson SHC, Elliot H, Wallis DJ, Canham LT, Powell JJ. 2003. Dissolution of <strong>di</strong>fferent forms<br />
of partially porous silicon wafers under simulated physiological con<strong>di</strong>tions. Phys Stat Sol (a)<br />
197: 331–335.<br />
24. Canham LT, Reeves CL, Newey JP, Houlton MR, Cox TI, Buriak JM, Stewart MP. 1999.<br />
Derivatized mesoporous silicon with dramatically improved stability in simulated human blood<br />
plasma. Adv Mater 11:1505–1507.<br />
25. Bow<strong>di</strong>tch AP, Waters K, Gale H, Rice P, Scott EAM, Canham LT, Reeves CL, Loni A, Cox TI.<br />
1999. In-vivo assessment of tissue compatibility and calcification of bulk and porous silicon.<br />
Mat Res Soc Symp Proc 536:149–154.<br />
26. De Stefano L., Moretti L., Ren<strong>di</strong>na I., Rossi A. M., Tundo S. (2004), “Smart optical sensors for<br />
chemical substances based on porous silicon technology”, Appl. Opt. 43, No. 1, 167-172.<br />
27. De Stefano L., Moretti L., Rossi A. M., Rocchia M., Lamberti A., Longo O., Arcari P., Ren<strong>di</strong>na<br />
I., (2004), “Optical sensors for vapors, liquids, and biological molecules based on porous<br />
silicon technology”, IEEE Trans. Nanotech. Vol. 3 n. 1, 49-54.<br />
28. Dancil K.-P. S., Greiner D. P., and Sailor M. J. (1999), “A Porous Silicon Optical Biosensor:<br />
Detection of Reversible Bin<strong>di</strong>ng of IgG to a Protein A-Mo<strong>di</strong>fied Surface”, J. Am. Chem. Soc.<br />
121, 7925-7930.<br />
29. Hart B. R., Letant S. E., Kane S. R., Ha<strong>di</strong> M. Z., Shields S. J. and Reynolds J. G. (2003), “New<br />
method for attachment of biomolecules to porous silicon”, Chem. Comm., 322-323.<br />
30. Ljibisa Ristic, Sensors Technology and Devices, pp. 207, 1994.<br />
31. Gaillet M., Guendouz M., Ben Salah M., Le Jeune B., Le Brun G., Thin Solid Film 455-456,<br />
410 (2004).<br />
32. V. Chamard, G. Dolino and F. Muller, J. Appl. Phys. 84, 6659 (1998).<br />
33. Socrates G., Infrared and Raman characteristic Group Frequencies (Wiley, England 2001).<br />
15
Thue-Morse: aperio<strong>di</strong>c multilayers made of porous silicon<br />
This section is aimed at the optimization of the fabrication process and the<br />
characterization of the Thue-Morse S0-S7 sequences by using the PSi technology.
phys. stat. sol. (c) 4, No. 6, 1966–1970 (2007) / DOI 10.1002/pssc.200674342<br />
Optical properties of porous silicon Thue-Morse structures<br />
Luca De Stefano *, 1 , Ilaria Rea 1 , Lucia Rotiroti 1 , Luigi Moretti 2 , and Ivo Ren<strong>di</strong>na 1<br />
1<br />
Institute for Microelectronics and Microsystems, National Council of Research, Via P. Castellino 111,<br />
80131 Naples, Italy<br />
2<br />
DIMET “Me<strong>di</strong>terranea” University of Reggio Calabria, Località Feo <strong>di</strong> Vito, 89060 Reggio Calabria,<br />
Italy<br />
Received 17 March 2006, revised 15 September 2006, accepted 15 November 2006<br />
Published online 9 May 2007<br />
PACS 42.70.Qs, 78.66.Db, 78.67.–n, 81.05.Rm<br />
Dielectric aperio<strong>di</strong>c Thue-Morse structures up to 128 layers have been realized by the porous silicon technology.<br />
Normal incidence reflectivity measurements have been performed to investigate the photonic<br />
properties of the devices. A partial photonic band gap region, centered at 1100 nm and 70 nm wide has<br />
been observed for the S 6 and S 7 Thue-Morse structures. The S 6 multilayer has been stu<strong>di</strong>ed as sensor device<br />
on exposure to several chemical substances.<br />
1 Introduction<br />
© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim<br />
In recent years, <strong>di</strong>electric photonic band gap (PBG) structures have attracted great attention for their<br />
property to forbid the propagation of the light at fixed wavelengths and the possibility to realize an alloptical<br />
integrated circuit which is the basic element of the optical computer [1, 2]. The simplest, one<br />
<strong>di</strong>mensional PBG structure is the Bragg mirror which is made of alternating layers of high (nH) and low<br />
(nL) refractive index, whose thicknesses satisfy the Bragg relation 2(nHdH+nLdL) = λBm where m is the<br />
order of the Bragg con<strong>di</strong>tion. Lot of theoretical and experimental stu<strong>di</strong>es on these perio<strong>di</strong>c structures<br />
have shown that they are characterized by the presence of only one degenerate PBG in a period of the<br />
reciprocal space, even if it can be omni<strong>di</strong>rectional [3]. On the contrary, multiple PBGs are typical of<br />
aperio<strong>di</strong>c structures. The main example of 1D aperio<strong>di</strong>c multilayer is given by the Thue-Morse sequence<br />
which can be obtained alternating high (A) and low (B) refractive index layers following the substitution<br />
rules A→AB and B→BA [4]. The lower-orders of the Thue-Morse sequence are: S0=A, S1=AB,<br />
S2=ABBA, S3=ABBABAAB, S4=ABBABAABBAABABBA, and so on. Dielectric Thue-Morse multilayers<br />
have been already demonstrated using Si/SiO2 and TiO2/SiO2 as high/low refractive index materials<br />
[5, 6] and recently also in porous silicon (PSi) [7]. In this study we have fabricated and characterized<br />
the Thue-Morse S0-S7 sequences by using the PSi technology. The PSi is a very attractive material due to<br />
the possibility of realize in only one process multilayered structures that exploit a high quality optical<br />
response. PSi is obtained by the electrochemical etching of crystalline silicon in a hydrofluoridric acid<br />
(HF) based solution. The porosity and the thickness of a single layer are linear functions of the current<br />
density and the ano<strong>di</strong>zation time for a fixed doping level of the silicon wafer and HF concentration. The<br />
refractive index of the PSi film depends on its porosity, and can be calculated in the frame of several<br />
effective me<strong>di</strong>um approximations, like the Bruggemann or Maxwell-Garnett models [8]. Moreover, the<br />
PSi is an ideal material for sensing applications because it is characterized by a very large specific surface<br />
area (200-500 m 2 cm -3 ), so that an effective interaction with several substances is assured.<br />
* Correspon<strong>di</strong>ng author: e-mail: luca.destefano@na.imm.cnr.it, Phone: +390816132375, Fax: +390816132598<br />
© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
phys. stat. sol. (c) 4, No. 6 (2007) 1967<br />
2 Materials and methods<br />
A highly doped p + -silicon, oriented, 0.01 Ω cm resistivity, 400 µm thick was used as substrate to<br />
realize the Thue-Morse structures. Samples were fabricated in dark light at room temperature using a<br />
solution of 30% volumetric fraction of aqueous HF (50% wt) and 70% of Ethanol. Before ano<strong>di</strong>zation,<br />
we have removed the thin film of native oxide from the silicon wafer by rapid rinsing it in a <strong>di</strong>luted HF<br />
solution. High porosity layers, with an average refractive index nH ≅ 1.3 and a thickness d ≅ 135 nm,<br />
were obtained applying a current density of 150 mA/cm 2 for 0.88 s. Low porosity layers, with an effective<br />
refractive index nL ≅ 1.96 and a thickness of d ≅ 90 nm, were obtained with a current density of 5<br />
mA/cm 2 for 0.53 s. The thickness d of each layer was designed to satisfy the Bragg con<strong>di</strong>tion dn= λB/4<br />
where n is the average refractive index and λB=700 nm. Thicknesses and porosities have been estimated<br />
by variable angle spectroscopic ellipsometry measurements on the single PSi layers.<br />
An Y optical reflection probe (Avantes), connected to a white light source and to an optical spectrum<br />
analyzer (Ando, AQ6315A), has been used for the reflectivity measurements at normal incidence. The<br />
reflectivity spectra were measured between 600 and 1600 nm with a resolution of 0.2 nm.<br />
Reflectivity measurements on exposure to volatile substances have been performed in a steel test chamber<br />
equipped with an optical access through a quartz window and in/out channels for gas fee<strong>di</strong>ng.<br />
3 Results and <strong>di</strong>scussion<br />
A scheme of the porous silicon Thue-Morse sequences realized is reported in Fig. 1: the number of the<br />
layers increases, as 2 n where n is the Thue-Morse order, while the thickness of the devices is dSn = 2dSn-1<br />
for n > 1. The realized samples S0-S7 have thicknesses spanning the range between 0.135 µm and 14.4<br />
µm.<br />
Fig. 1 Scheme of the porous silicon Thue-Morse sequences realized and characterized in the present work.<br />
In Figs. 2 and 3 the experimental (solid line) and calculated (dash line) normal incidence reflectivity<br />
spectra are reported for S3 (2-a), S4 (2-b), S5 (2-c), S6 (3-d), and S7 (3-e) Thue-Morse structures which<br />
are constituted by 8, 16, 32, 64 and 128 layers, respectively. The reflectivity spectra have been reproduced<br />
by a transfer matrix method [9] inclu<strong>di</strong>ng the wavelength <strong>di</strong>spersion of silicon. At the increasing<br />
of the layers number we can observe a worsening of the agreement between the experimental and calcu-<br />
www.pss-c.com © 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1968 L. De Stefano et al.: Optical properties of porous Si Thue-Morse structures<br />
Reflectivity<br />
1.0<br />
0.5<br />
0.0<br />
600<br />
1.0<br />
800 1000 1200 1400 1600<br />
0.5<br />
(b)<br />
0.0<br />
600<br />
1.0<br />
800 1000 1200 1400 1600<br />
0.5<br />
0.0<br />
600 800 1000 1200 1400 1600<br />
Wavelength (nm)<br />
Experimental<br />
Simulation<br />
(a)<br />
(c)<br />
Fig. 2 Experimental (solid line) and calculated<br />
(dashed line) normal incidence reflectivity spectra for<br />
S 3 (a), S4 (b) and S5 (c) Thue-Morse structures.<br />
0.0<br />
600 800 1000 1200 1400 1600<br />
© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.pss-c.com<br />
Reflectivity<br />
1.0<br />
0.5<br />
0.0<br />
600 800 1000 1200 1400 1600<br />
1.0<br />
0.5<br />
Wavelength (nm)<br />
Experimental<br />
Simulation<br />
lated spectra. This effect is due to small porosity and thickness <strong>di</strong>shomogeneities of the deeper layers,<br />
and to scattering losses, caused by interfaces roughness. The spectrum of S3 Thue-Morse is characterized<br />
by two band gaps separated by a transmission peak at 1000 nm. At the increasing of the order of Thue-<br />
Morse sequence, the photonic band gap splits and very narrow transmission peaks appear. This is a well<br />
known effect due to the random nature of the aperio<strong>di</strong>c Thue-Morse: the eigenstates, i.e. the stationary<br />
configurations of the electromagnetic field, are more localized respect to the perio<strong>di</strong>c and quasi-perio<strong>di</strong>c<br />
geometrical structures, such as the Bragg and the Fibonacci geometries, respectively [10].<br />
Transmission peak FWHM (nm)<br />
200<br />
175<br />
150<br />
125<br />
100<br />
75<br />
50<br />
25<br />
0<br />
0 25 50 75 100 125 150<br />
Layers number<br />
Fig. 4 Full width at half maximum of the transmission peaks in the experimental reflectivity spectra of S3-S7 sequences.<br />
(d)<br />
(e)<br />
Fig. 3 Experimental (solid line) and calculated<br />
(dashed line) normal incidence reflectivity spectra for<br />
S 6 (d) and S 7 (e) Thue-Morse structures.
phys. stat. sol. (c) 4, No. 6 (2007) 1969<br />
The Fig. 4 reports the full width at half maximum (FWHM) of the transmission peaks present in the<br />
reflectivity stop bands spectra of the S3-S7 Thue-Morse structures: on increasing the number of the layers<br />
the peak width decreases exactly as an exponential function (R 2 = 0.998).<br />
The presence of resonant peaks in the reflectivity spectrum suggests the use of such porous structures as<br />
optical transducers for gas and liquid monitoring. The S6 Thue-Morse structure has been investigated as<br />
sensor device on exposure to several vapors [11, 12]. The gaseous substances penetrate into the nanometric<br />
pores, substituting the air. The average refractive indexes of the layers change, so that the reflectivity<br />
spectrum of the device shifts towards higher wavelengths. In Figure 5 the red-shift of the S6 structure<br />
transmission sharp peak at 1000 nm is reported as a function of the refractive index of each volatile substance.<br />
A linear response is observed and a sensitivity of 530 (50) nm/RIU (Refractive Index Unit) can<br />
be estimated.<br />
Table 1 Central wavelength and width of the photonic and fractal theoretical stop band.<br />
TM<br />
PBG I FBG II FBG<br />
Structure λ (nm) Width (nm) λ (nm) Width(nm) λ (nm) Width (nm)<br />
S3 766 161 1307 452 0 0<br />
S4 880 149 1138 324 0 0<br />
S5 840 111 1060 210 1240 146<br />
S6 870 114 1084 218 1250 112<br />
S7 813 119 1029 230 1216 125<br />
Red Shift (nm)<br />
220<br />
210<br />
200<br />
190<br />
180<br />
Refractive Index Unit (RIU) Sensitivity<br />
Methanol 1.328<br />
Pentane 1.358<br />
Isopropanol 1.377<br />
Isobutanol 1.396<br />
1.32 1.34 1.36<br />
Refractive Index<br />
1.38 1.40<br />
Fig. 5 Red-shift of the S6 structure transmission peak at 1000 nm versus the refractive index of vapor substances.<br />
4 Conclusions<br />
In conclusion, in this work we have reported the fabrication and the optical characterization of aperio<strong>di</strong>c<br />
Thue-Morse structures up to 128 layers based on PSi technology. The multilayers are relatively easy to<br />
realize and exploit a high quality optical response. A splitting of photonic band gap is observed at the<br />
increasing of the Thue-Morse order. A partial photonic band gap region, centered at 1100 nm and 70 nm<br />
wide was observed for the S6 and S7 Thue-Morse structures. The PSi Thue-Morse structures can be also<br />
successfully used as sensor devices due to their characteristic properties.<br />
www.pss-c.com © 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1970 L. De Stefano et al.: Optical properties of porous Si Thue-Morse structures<br />
References<br />
[1] E. Yablonovitch, Phys. Rev. Lett. 58, 2059 (1987).<br />
[2] S. John, Phys. Rev. Lett. 58, 2486 (1987).<br />
[3] V. Agarwal and J. A. del Rio, Appl. Phys. Lett. 82, 1512 (2003).<br />
[4] M. Dulea, M. Severin, and R. Riklund, Phys. Rev. B 42, 3680 (1990).<br />
[5] L. Dal Negro, M. Stolfi, Y. Yi, J. Michel, X. Duan, L. C. Kimerling, J. LeBlanc, and J. Haavisto, Appl. Phys.<br />
Lett. 84, 5186 (2004).<br />
[6] F. Qui, R. W. Peng, X. Q. Huang, X. F. Hu, Mu Wang, A. Hu, S. S. Jiang, and D. Feng, Europhys. Lett. 68,<br />
658 (2004).<br />
[7] M. E. Mora-Ramos, V. Agarwal, and J. A. Soto Urueta, Microelectron. J. 36, 413 (2005).<br />
[8] R. W. Hardeman, M. I. J. Beale, D. B. Gasson, J. M. Keen, C. Pickering, and D. J. Robbins, Surf. Sci. 152/153,<br />
1051 (1985).<br />
[9] M. A. Muriel and A. Carballar, IEEE Photon. Technol. Lett. 9, 955 (1997).<br />
[10] X. Jiang, Y. Zhang, S. Feng, K. Huang, Y. Yi, and J. D. Joannopoulos, Appl. Phys. Lett. 86, 201110 (2006).<br />
[11] P. A. Snow, E. K. Squire, P. St. J. Russell, and L. T. Canham, J. Appl. Phys. 86, 1781 (1999).<br />
[12] L. De Stefano, I. Ren<strong>di</strong>na, L. Moretti, S. Tundo, and A. M. Rossi, Appl. Opt. 43, 167 (2004).<br />
© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.pss-c.com
Optical sensing of chemical compounds using porous silicon<br />
devices<br />
In this section are presented some very interesting results about the detection of chemical<br />
compounds by using porous silicon optical devices. Some porous silicon structures, were<br />
exposed to several substances of environmental interest: the detection in aqueous and<br />
humic solutions of a common pesticide and organic compounds. Very large red shifts of<br />
the reflectivity spectra, due to changes in the average refractive index, have been<br />
observed.
Introduction to the sensing mechanism<br />
In this section, are presented optical sensors, based on porous silicon (PSi) technology for<br />
liquid and vapour detection. Optical sensors for chemical contaminants can adopt a<br />
spectroscopic method (usually in the IR region where gas absorption spectra have specific<br />
signatures), or measure the changes in some physical properties (colour, refractive index,<br />
photoluminescence, fluorescence and so on) due to the interaction between the<br />
substances under investigation and the sensor itself [1, 2]. From this point of view, PSi is a<br />
very interesting material due to its large specific surface area (on the order of 500 m 2 /cm 3 )<br />
which is a great advantage in gas sensing applications, so that this technology has been<br />
extensively stu<strong>di</strong>ed in this field. When the PSi is exposed to vapours, or <strong>di</strong>p in liquid<br />
substances, a repeatable and completely reversible change in the reflectivity spectrum is<br />
observed. In fact, the substitution of air in the pores by the organic molecules causes an<br />
increasing of the average refractive index of the device, shifting towards longer<br />
wavelengths its characteristic reflectivity spectrum.<br />
In a first optical characterization, it was followed the approach of Sailor et al. [3] choosing<br />
the simple case of a single layer of porous silicon as transducer device. Figure 1 shows a<br />
typical reflectivity spectrum (before and after chemical infiltration) from a PSi layer under<br />
white light illumination.<br />
Reflectance (a.u.)<br />
1.00<br />
0.75<br />
0.50<br />
0.25<br />
1.00<br />
0.75<br />
0.50<br />
after infiltration<br />
0.25<br />
11000 12000 13000 14000 15000 16000 17000<br />
Wavenumber (cm -1 )<br />
Maximum order ( Δm)<br />
16<br />
14<br />
12<br />
10<br />
8<br />
6<br />
4<br />
2<br />
11000 12000 13000 14000 15000 16000 17000<br />
PSi as etched<br />
Wavenumber (cm -1 )<br />
n xyl d=27352 (92) nm<br />
n unp d=24560 (110) nm<br />
Figure 1. Reflectivity spectrum of PSi layer. In the inset the m-order peak are plotted as function of the<br />
wavenumber.<br />
From the optical point of view, this structure is an optical Fabry-Perot interferometer, so<br />
that the maxima in the reflectivity spectrum appear at wavelengths λm which satisfy the<br />
fallowing relationship:<br />
m = 2 nd / λ<br />
(1)<br />
m<br />
where m is an integer, d is the film thickness and n is the average refractive index of the<br />
layer [4, 5].<br />
If it’s assumed that the refractive index is independent on the wavelength over the<br />
considered range, the maxima are equally spaced in the wavenumber (1/λm). When the<br />
maximum order m is plotted as a function of the wavenumber, each point lies on a straight<br />
22
line which slope is the optical path nd, as it is shown in the inset of Figure 1. When the PSi<br />
layer is exposed to vapors, or <strong>di</strong>p in the liquid phase of the same substance, the<br />
substitution of air into the pores by its molecules causes a fringes shift in wavelength,<br />
which corresponds to a change in the optical path nd. Since the thickness d is fixed by the<br />
physical <strong>di</strong>mension of the PSi matrix, the variation is clearly due to changes in the average<br />
refractive index. In the case of vapors and gases, the filling liquid phase is due to the<br />
capillary condensation phenomenon, which is regulated from the Kelvin equation:<br />
K sat<br />
BTρ ln( p / p)<br />
= 2γ<br />
cos / R<br />
l lg θ<br />
(2)<br />
where ρl is the density of the liquid phase, γlg is the liquid-gas surface tension coefficient at<br />
room temperature T, R is the ra<strong>di</strong>us of the pores, p/psat is the relative vapour pressure into<br />
the pores, and θ is the contact angle [6].<br />
A quantitative model taking into account the optical path increase can be realized by<br />
applying the Bruggemann effective me<strong>di</strong>um approximation theory. The relation between<br />
the volume fraction of each me<strong>di</strong>um and its <strong>di</strong>electric constant can be written as:<br />
⎛ ε ⎞ ⎛ ⎞ ⎛ ⎞<br />
Si −εp<br />
εair<br />
−εp<br />
εch<br />
−εp<br />
( 1 − p)<br />
⎜ ⎟+<br />
( p−V)<br />
⎜ ⎟+<br />
V⎜<br />
⎟ = 0<br />
(3)<br />
⎜ 2 ⎟ ⎜ 2 ⎟ ⎜ 2 ⎟<br />
⎝εSi<br />
+ εp<br />
⎠ ⎝εair<br />
+ εp<br />
⎠ ⎝εch<br />
+ εp<br />
⎠<br />
where p, V, εsi, εair, εch and εp are the layer porosity, the layer liquid fraction (LLF), the<br />
<strong>di</strong>electric constants of silicon, air, chemical substance and porous silicon, respectively [7].<br />
References<br />
1. P. A. Snow, E. K. Squire, P. St. J. Russel, and L. T. Canham, Vapor sensing using the optical<br />
properties of porous silicon Bragg mirrors, J. Appl. Phys. 86, (1999) 1781-1784.<br />
2. V. S. Y. Lin, K. Motesharei, K. P. S. Dancil, M. J. Sailor, and M. R. Gha<strong>di</strong>ri, A porous silicon-based<br />
optical interferometric biosensor, Science 278, (1997) 840-843.<br />
3. Gao, J.; Gao, T.; Sailor, M.J. Porous-silicon vapor sensor based on laser interferometry. Appl. Phys.<br />
Lett. 2000, 77, 901-903.<br />
4. Lin, V. S.-Y.; Motesharei, K.; Dancil, K.-P. S.; Sailor, M. J.; Gha<strong>di</strong>ri, M. R. A porous silicon-based<br />
optical interferometric biosensor. Science 1997, 278, 840-843.<br />
5. Anderson, M. A.; Tinsley-Brown, A.; Allcock, P.; Perkins, E. A.; Snow, P.; Hollings, M.; Smith, R. G.;<br />
Reeves, C.; Squirrell, D. J.; Nicklin, S.; Cox; T. I. Sensitivity of the optical properties of porous silicon<br />
layers to the refractive index of liquid in the pores. Phys. Stat. Sol. A 2003, 197, 528-533.<br />
6. Neimark, V.; Ravikovitch, P. I. Capillary condensation in MMS and pore structure characterization.<br />
Microporous Mesoporous Mater. 2001, 44-45, 697-707.<br />
7. Spanier, J. E.; Herman, I. P. Use of hybrid phenomenological and statistical effective-me<strong>di</strong>um<br />
theories of <strong>di</strong>electric functions to model the infrared reflectance of porous SiC films. Phys. Rev. B<br />
2000, 61, 10437-10450.<br />
23
phys. stat. sol. (c) 4, No. 6, 1941–1945 (2007) / DOI 10.1002/pssc.200674336<br />
Quantitative measurements of hydro-alcoholic binary mixtures<br />
by porous silicon optical microsensors<br />
Luca De Stefano *1 , Lucia Rotiroti 1 , Ilaria Rea 1 , Luigi Moretti 2 , and Ivo Ren<strong>di</strong>na 1<br />
1<br />
Institute for Microelectronics and Microsystems, National Council of Research, Via P. Castellino 111,<br />
80131 Naples, Italy<br />
2<br />
DIMET – “Me<strong>di</strong>terranea” University of Reggio Calabria, Località Feo <strong>di</strong> Vito, 89060 Reggio Calabria,<br />
Italy<br />
Received 17 March 2006, revised 15 September 2006, accepted 15 November 2006<br />
Published online 9 May 2007<br />
PACS 07.07.Df, 81.05.Rm, 81.70.Jb<br />
In this communication, we present a simple, completely reversible and well performing optical microsensor<br />
for the quantitative determination in liquid hydro-alcoholic binary mixtures, based on the simple and<br />
low cost porous silicon (PSi) nanotechnology. In our sensor device, the transducer element is a PSi optical<br />
microcavity (PSMC) which is constituted by a Fabry-Pérot layer between two <strong>di</strong>stributed Bragg reflectors.<br />
When the device is exposed to the mixture, the partial substitution of air in the pores changes the average<br />
refractive index of PSMC, causing a red shift of its reflectivity spectrum, characteristic of the concentration<br />
of each component. We have characterized, in the 0-100% composition range, three hydroalcoholic<br />
binary mixtures (water-Ethanol, water- Methanol and water- Isopropanol) which exhibit very<br />
<strong>di</strong>fferent behaviour.<br />
1 Introduction<br />
© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim<br />
Hydro-alcoholic mixtures are largely used in the food industry: for example, the monitoring of the DI<br />
water-Ethanol compound is of extreme importance since it determines the alcoholic volume content of<br />
the most common beverages such as wines, beers, liquors and spirits. The standard methods, prescribed<br />
for the quantitative determination of the alcohol concentration, suffer some drawbacks since they require<br />
professional laboratories with specialized personnel, and expensive equipment, such as the HPLC, and<br />
sometimes also the sample pre-treatment. These procedures cannot be applied on line and are useless for<br />
fast routine analyses. The optical sensors offer a very attractive solution in the realization of analytical<br />
devices: they could exhibit high sensitivity, fast response, small <strong>di</strong>mensions, low cost and also the possibility<br />
of remote sensing.<br />
PSi based optical sensors have recently gained a well established place among the optical monitoring<br />
devices due to the peculiar physical properties of this material. PSi is very cheap to produce, fully integrable<br />
with the standard silicon technology, moreover its high surface area assures a strong and fast interaction<br />
with liquid or vapor substances. On these bases, several applications of PSi monolayer and<br />
multilayer structures have been recently considered in biological and chemical sensing [1–3, 9]. A step<br />
beyond the optical identification of a pure substance is the quantitative determination of the components<br />
concentrations in a mixture. When the PSi device is exposed to a liquid or gaseous solution, a change in<br />
the optical reflectivity spectrum can be observed due to the partial substitution of air in the pores by a<br />
liquid phase, due to the <strong>di</strong>rect infiltration or to the capillary condensation, which increases the average<br />
* Correspon<strong>di</strong>ng author: e-mail: luca.destefano@na.imm.cnr.it, Phone +390816132375, Fax +390816132598<br />
© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1942 L. De Stefano et al.: Quantitative measurements of hydro-alcoholic binary mixtures<br />
refractive index. This phenomenon depends on both the refractive index of the liquid phase but also on<br />
the filling factor, i.e. how it penetrates into the spongy structure of the PSi. The dependence of the result<br />
on two parameters permits to resolve a binary mixture with only one optical measurement [4]. Our group<br />
has already published some quantitative results on gaseous and liquid binary mixtures, obtained by using<br />
PSi based resonant optical sensors. In particular, we have characterized the sensor response on exposure<br />
to vapors of several volatile compounds (Iso-propanol/Ethanol, Iso-propanol/Methanol, Xylene/Methanol),<br />
and after the interaction with liquid solutions of pesticides in water and humic acid [5–<br />
7]. We present here an all-optical PSi based chemical sensor for quantitative determination in liquid<br />
hydro-alcoholic binary mixtures.<br />
2 Material and methods<br />
The optical transducer we used in our experiments is a λ/2 Fabry-Pérot layer enclosed between two<br />
Bragg reflectors; each one fabricated by alternating seven layers with high (low) and low (high) refractive<br />
in<strong>di</strong>ces (porosities). This PSi multilayer structure is an optical microcavity (PSMC), electrochemically<br />
etched by a HF-based solution on a highly doped p + -silicon, oriented, 0.01 Ω cm resistivity,<br />
and 400 µm thick. A current density of 350 mA/cm 2 for 0.6 sec was applied to obtain the high refractive<br />
index layer, with a porosity of 62 %, while one of 50 mA/cm 2 was applied for 0.8 sec for the low index<br />
layer, with a porosity of 42 %. The Fabry-Pérot is a low refractive index layer. This structure has a characteristic<br />
resonance peak at 1110 nm in the middle of a 271 nm stop band. The infiltration of the aqueous<br />
solution into the spongy structure could be problematic since the PSi surface as etched is highly hydrophobic<br />
due to the presence of Si-H bonds. In order to assure an efficient interaction it is necessary to<br />
make this surface hydrophilic by a thermal oxidation process (at 900 °C for 45 minutes).<br />
White light<br />
source<br />
Collimator<br />
Optical fiber<br />
Objective<br />
DBR #1<br />
DBR #2<br />
Fig. 1 The experimental optical set-up: the sample was illuminated by a collimated white source; the output signal<br />
was <strong>di</strong>rected in an optical spectrum analyzer and it was measured over the range 1000-1050 nm with a resolution of<br />
0.2 nm.<br />
We stu<strong>di</strong>ed the resonance peak shift on exposure to the liquid mixtures as a function of the alcohols<br />
concentration in the de-ionized water (DI). To avoid some lens effect, we placed a drop (20 µL) of the<br />
solution on the PSi surface and covered it by a thin transparent glass. In Fig. 1 the experimental set-up<br />
used to measure the reflectivity spectrum is reported. When the liquid penetrated into the PSMC, a<br />
change in the cavity reflectivity spectrum was observed: the partial substitution of air by liquid in the<br />
© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.pss-c.com<br />
OSA<br />
d=λ/2nl<br />
X7<br />
Low P<br />
High P
phys. stat. sol. (c) 4, No. 6 (2007) 1943<br />
pores of each layer increases the average refractive index of the microcavity and, consequentially, generates<br />
a red-shift of the reflectivity spectrum. Measurements are stored when equilibrium con<strong>di</strong>tions are<br />
reached, i.e. when the signal does not change any more in time.<br />
Fig. 2 The optical spectrum of the PSMC before and after the thermal oxidation and the infrared spectrum after<br />
thermal treatment.<br />
3 Experimental results<br />
The optical and infrared spectra of the device before and after the oxidation are reported in Fig. 2. The<br />
microcavity has a FWHM of 12 nm and the thermal oxidation causes a blue shift of the reflectivity spectrum<br />
of 110 nm. The optical quality of the PSi response is preserved after the thermal treatment. In the<br />
infrared spectrum is well evident the presence of the characteristic peaks of Si-O-Si bonds at about 1000-<br />
1100 cm –1 and the absence of the Si-H bonds at 2100 cm –1 .<br />
In a recent paper 5 , we have already used a very simple method, common in the analytical chemistry, for<br />
the quantitative determinations of binary mixtures: the External Standard Method. The calibration curves<br />
can be obtained by measuring the physical parameter under monitoring, such as the infrared absorption at<br />
∆λ (nm)<br />
Reflectivity (u.a)<br />
14<br />
12<br />
10<br />
8<br />
6<br />
4<br />
2<br />
0<br />
1.0<br />
0.5<br />
PSMC as etched<br />
PSMC after thermal oxidation<br />
Isopropanol-DI<br />
EtOH-DI<br />
∆λ= 110 nm<br />
0.0<br />
600 800 1000 1200 1400 1600<br />
Wavelength (nm)<br />
0 20 40 60 80 100<br />
Alcohol concentration in DI (% V/V)<br />
0 20 40 60 80 100<br />
Fig. 3 The calibration curves in case of three binary mixtures (Iso-propanol/DI, Ethanol/DI and Methanol/DI).<br />
www.pss-c.com © 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim<br />
Reflectance (a.u.)<br />
∆λ (nm)<br />
4.5<br />
4.0<br />
3.5<br />
3.0<br />
2.5<br />
2.0<br />
1.5<br />
1.0<br />
0.5<br />
0.0<br />
-0.5<br />
140<br />
120<br />
100<br />
80<br />
60<br />
40<br />
20<br />
0<br />
2400 2200 2000 1800 1600 1400 1200 1000 800<br />
MeOH-DI<br />
polinomial fit<br />
Wavenumber (cm -1 )<br />
Alcohol concentration in DI (% V/V)<br />
SiO 2
n<br />
1944 L. De Stefano et al.: Quantitative measurements of hydro-alcoholic binary mixtures<br />
a fixed wavelength or the optical intensity, while the content of one component varies between 0 and<br />
100%, volume by volume (V/V). Also in the case under study, the calibration curves have been obtained<br />
by plotting the peak shifts measured as a function of the volume percentage of one component in the<br />
binary mixtures.<br />
The results are shown in Fig. 3. The errors on the experimental points are the standard deviations on<br />
three to five measurements, using the same chip and the same hydro-alcoholic solution. There is a well<br />
evident <strong>di</strong>ssimilarity respect to the linear behaviour of the optical response obtained on exposure to the<br />
vapours of the volatile binary compounds. Of course, even if the calibration curves are not linear, the<br />
unknown concentrations of the characterized mixtures can be still quantitatively determined with a single<br />
measurement of the peak shift: the volume concentrations of both the components can be picked out<br />
from the interpolation of the experimental data. Moreover, while the Isopropanol and Ethanol-DI mixtures<br />
shifts follow a slow parabolic behaviour, the Methanol-DI mixture exhibits a very <strong>di</strong>fferent response.<br />
The main <strong>di</strong>fference is that a 50 % solution of Methanol in DI causes a larger shift than both the<br />
pure substances. An explanation to the shape of the ∆λ curves can be found in the behaviour of the mixtures<br />
refractive indexes as a function of the alcohol concentration [8], shown in Fig. 4. In all the cases,<br />
the peak shift strictly follows the behaviour of the refractive index binary mixture. The anomalous shape<br />
of the Methanol-DI mixture can be ascribed to the higher <strong>di</strong>pole moment of its constituent: due to a<br />
strong <strong>di</strong>polar interaction this hydro-alcoholic mixture could exhibit a denser optical phase respect to the<br />
two pure substances [8].<br />
The time-resolved measurements have shown that sensor dynamic is of order of few seconds, also depen<strong>di</strong>ng<br />
on geometrical characteristic of the experimental set-up [3]. Moreover, the phenomenon is completely<br />
reversible and reproducible, so that the transducer PSi element can be reusable several times.<br />
As a practical application of this study, we have tried to use the PSMC to determine the alcohol content<br />
of a red wine (Aglianico del Taburno, DOC). To this aim, we have <strong>di</strong>luted the wine, with a declared<br />
alcohol content of 12% V/V, in DI water and we have measured the calibration curve up to this value.<br />
From the graph in Fig. 3, we expected a linear behavior, due to the small range of Ethanol concentration<br />
in the wine. In Fig. 5, we have reported the wine calibration curve together with the first order approximation<br />
of the Ethanol-DI calibration curve of Fig. 3. The two curves are both linear but they have <strong>di</strong>fferent<br />
slopes: in particular, the Ethanol-DI curve underestimates the wine percentage content. This is also a<br />
1.38<br />
1.37<br />
1.36<br />
1.35<br />
1.34<br />
1.33<br />
IPA-DI water refractive index<br />
EtOH-DI water refractive index<br />
0 20 40 60 80 100<br />
Alcohol concentration in DI (% V/V)<br />
Fig. 4 The mixtures refractive indexes as a function of the alcohol concentration in DI.<br />
n<br />
0 20 40 60 80 100<br />
somewhat expected result: in fact the wine is a very complex mixture where Ethanol and water are the<br />
main constituents but many other volatile organic compounds are present. Nevertheless, we believe that<br />
© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.pss-c.com<br />
1.344<br />
1.342<br />
1.340<br />
1.338<br />
1.336<br />
1.334<br />
1.332<br />
1.330<br />
1.328<br />
Methanol concentration in DI (% V/V)
phys. stat. sol. (c) 4, No. 6 (2007) 1945<br />
Fig. 5 The calibration curve in case of wine-DI water solution.<br />
the measurement technique could be improved, perhaps by time-resolved procedures, in order to give<br />
more precise results.<br />
4 Conclusion<br />
We have presented a simple and well performing optical microsensor for quantitative determination in<br />
liquid mixtures, based on the simple and low cost porous silicon nanotechnology. The device is a microcavity<br />
structure that is <strong>di</strong>rectly realized on the silicon chip by electrochemical etching. The fabrication<br />
process is compatible with silicon microelectronic and micro machining technologies, so that the sensors<br />
could be monolithically integrated with microelectronic circuits, or be part of a micro-optical-system. In<br />
terms of the measurement resolution, the resonant cavity enhanced operation of the proposed sensor<br />
offers very good performances: the interaction between the porous silicon microcavity and the external<br />
mixtures detunes the optical microcavity inducing large shifts of its resonant peak wavelength.<br />
References<br />
∆λ (nm)<br />
5<br />
4<br />
3<br />
2<br />
1<br />
0<br />
Wine-DI calibration curve<br />
First order approximation of<br />
the Ethanol-DI calibration curve<br />
0 2 4 6 8 10 12 14<br />
Wine concentration in DI (% V/V)<br />
[1] P. A. Snow, E. K. Squire, P. St. J. Russel, and L. T. Canaham, J. Appl. Phys. 86, 1781 (1999).<br />
[2] V. S. Y. Lin, K. Motesharei, K. P. S. Dancil, M. J. Sailor, and M. R. Gha<strong>di</strong>ri, Science 278, 840 (1997).<br />
[3] L. De Stefano, I. Ren<strong>di</strong>na, L. Moretti, A.M. Rossi, Sens. Actuators B 100, 168 (2004).<br />
[4] C.K. Ho, M.T. Itamura, M. Kelley, and R.G. Hughes, Review of Chemical Sensors for In-situ Monitoring of<br />
Volatile Contaminants, San<strong>di</strong>a Report SAND2001-0643, San<strong>di</strong>a National Laboratories, 2001.<br />
[5] L. De Stefano, I. Ren<strong>di</strong>na, L. Moretti, and A.M. Rossi, phys. stat. sol. (a) 201, 1011 (2004).<br />
[6] L. De Stefano, L. Moretti, I. Ren<strong>di</strong>na, and L. Rotiroti, Sens. Actuators B 111/112, 522 (2005).<br />
[7] L. Rotiroti, L. De Stefano, L. Moretti, A. Piccolo, I. Ren<strong>di</strong>na, and A. M. Rossi, Biosens. Bioelectron. 20, 2136<br />
(2005).<br />
[8] Handbook of Chemistry and Physics, 68th ed. (CRC Press, Boca Raton, 1987/1988).<br />
[9] L. De Stefano, I. Rea, I. Ren<strong>di</strong>na, L. Rotiroti, M. Rossi, and S. D'Auria phys. stat. sol. (a) 203, 886 (2006).<br />
www.pss-c.com © 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Chapter 2<br />
Chemical and Biological<br />
functionalization<br />
of Porous Silicon surface
Introduction: Porous Silicon surface chemical functionalization<br />
Biosensors are nowadays technological hot topics due to the possible applications in fields<br />
of social interest such as me<strong>di</strong>cal <strong>di</strong>agnostic and health care, monitoring of environmental<br />
pollutants, homeland and defence security [1].<br />
Reagentless optical biosensors are monitoring devices which can detect a target analyte in<br />
a heterogeneous solution without the ad<strong>di</strong>tion of anything than the sample. In the fields of<br />
genomics and proteomics this is a straightforward advantage since it allows real-time<br />
readouts and, thus, very high throughoutputs analysis. A label free optical biosensor can<br />
be realized by integrating the biological probe with a signalling material which <strong>di</strong>rectly<br />
transduces the molecular recognition event into an optical signal.<br />
PSi is also very used as photonic material due to the possibility of fabricating high quality<br />
optical structures, either as single layers, like Fabry-Perot interferometers [2], or<br />
multilayers, such as Bragg [3] or rugate filters [4]. A key feature for a recognition<br />
transducer is a large surface area: PSi has a porous structure with a specific area up to<br />
200 – 500 m 2 cm -3 , so that it can be very sensitive to the presence of biochemical species<br />
which penetrate inside the pores. Unfortunately, the surface of the “as etched” PSi is<br />
highly hydrophobic due to the Si-H bonds so that aqueous solution can not infiltrate the<br />
sponge like matrix. A proper passivation process must be applied to stabilise the surface<br />
and to covalently link the bioprobe [5].<br />
• Hydrosilylation to produce Si-C bonds<br />
Interesting results on the chemical derivatization of PSi were reported before 1998, but the<br />
problem was the low treatment efficiency and, thus, the remaining instability. Especially,<br />
the groups of Mike Sailor (Department of Chemistry and Biochemistry University of<br />
California, San Diego), Jean-Noel Chazalviel (Palaiseau Polytechnical Institute, France),<br />
and Jillian Buriak (Department of Chemistry, University of Alberta, and the National<br />
Institute for Nanotechnology) stu<strong>di</strong>ed <strong>di</strong>fferent possible treatments for the PSi surface<br />
mo<strong>di</strong>fications with Si–C bonds.<br />
Their work finally led to the derivatized PSi surface, in which the Si–H bonds were<br />
replaced with Si–C bonds using hydrosilylation of alkenes or alkynes on the PSi surface.<br />
The group of Buriak introduced three <strong>di</strong>fferent approaches to obtain the chemically<br />
derivatized PSi surface: Lewis acid me<strong>di</strong>ated hydrosilylation [6,7] white light-promoted<br />
hydrosilylation [8,9] and catho<strong>di</strong>c electrografting [10].<br />
Carbon <strong>di</strong>rectly bonded to silicon yields a very stable surface species. First recognized by<br />
Chidsey and coworkers [11], Si–C bonded species possess greater kinetic stability relative<br />
to Si–O due to the low electronegativity of carbon. Silicon can rea<strong>di</strong>ly form 5- and 6-<br />
coor<strong>di</strong>nate interme<strong>di</strong>ates, and an electronegative element such as oxygen enhances the<br />
tendency of silicon to be attacked by nucleophiles. Si–C bonds are usually formed on<br />
hydride-terminated porous Si surfaces by hydrosilylation (Figure 1). Hydrosilylation<br />
involves reaction of an alkene (usually terminal) or alkyne with a Si–H bond. On porous Si,<br />
the reaction can be thermal [12], photochemical [13], or Lewis acid catalyzed [14].<br />
Figure 1: mechanism of PSi functionalization by nucleophilic substitution.<br />
29
Thermal hydrosilylation provides a means to place a wide variety of organic functional<br />
groups on a crystalline Si or porous Si surface. The main requirement of the reaction is<br />
that the silicon surface contains Si–H species so they can react with a terminal alkene on<br />
the organic fragment. Thus it is important to use freshly etched porous Si and to exclude<br />
oxygen and water from the reaction mixture.<br />
• Chemical grafting of Si-C bonds<br />
As an alternative to hydrosilylation, covalently attached layers can be formed on porous Si<br />
surfaces using Grignard (figure 2) and alkyl- or aryllithium reagents [15]. Because of their<br />
ease of application and dramatic improvements in stability, hydrosilylation grafting of alkyl<br />
halides are useful reactions for the preparation of biointerfaces.<br />
Figure 2: mechanism of PSi functionalization by Grignard reactive.<br />
It is important to note that porous Si mo<strong>di</strong>fication reactions do not provide 100% surface<br />
coverage; they merely decorate the surface with the functional group. Thus infrared<br />
spectra show a large amount of surface Si–H groups remaining after hydrosilylation or<br />
grafting reaction.<br />
It is still an open question as to why the mo<strong>di</strong>fication reactions impart such stability to<br />
the material. The stability probably derives from a combination of two factors: the Si–C<br />
bond is kinetically inert due to the low electro negativity of carbon and the attached organic<br />
species (typically a hydrocarbon chain 8 or more carbons long) is sufficiently hydrophobic<br />
that aqueous nucleophiles are excluded from the vicinity of their Si atom target. Reaction<br />
of porous Si with gas phase acetylene generates highly carbonized porous Si that is<br />
possibly the most stable form of Si–C mo<strong>di</strong>fied porous Si [16]. This material is referred to<br />
as thermally carbonized PSi. It has been extensively investigated by Salonen at<br />
Department of Physics of University of Turku, (Turku, Finland) and coworkers, with many<br />
publications of relevance to drug loa<strong>di</strong>ng and delivery [17].<br />
• Conjugation of bio molecules to mo<strong>di</strong>fied PSi<br />
Carbon grafting stabilizes porous Si against <strong>di</strong>ssolution in aqueous me<strong>di</strong>a, but the surface<br />
must still avoid the non-specific bin<strong>di</strong>ng of proteins and other biological species.<br />
The oxides of porous Si are easy to functionalize using conventional silanol chemistries.<br />
When small pores are present (as with p-type samples), trialkoxysilanes [(RO)3Si–R′] can<br />
be employed as surface linkers [18].<br />
Whereas Si–C chemistries are robust and versatile, chemistries involving Si–O bonds<br />
represent an attractive alternative for two reasons.<br />
First, the timescale in which highly porous SiO2 is stable in aqueous me<strong>di</strong>a is consistent<br />
with many short-term biomolecular recognitions applications: typically 20 min to a few<br />
hours. Second, a porous SiO2 sample that contains no ad<strong>di</strong>tional stabilizing chemistries is<br />
less likely to produce toxic or antigenic side effects.<br />
30
• Oxidation of porous silicon<br />
With its high surface area, porous silicon is particularly susceptible to air or water<br />
oxidation. The simplest way to stabilize PSi is obviously a partial oxidation. A few hours<br />
even at quite mild con<strong>di</strong>tions, around 300°C, cause mainly the so called back-bond<br />
oxidation of PSi [19]. The oxygen atoms selectively attack the back-bonds of the surface Si<br />
atoms instead of replacing hydrogen atoms. The oxygen bridges formed between the<br />
surface Si atoms and the second atomic Si layer expand the local atomic structure by<br />
30%, causing a slight decrease in the pore <strong>di</strong>ameter. In ad<strong>di</strong>tion to the increased stability,<br />
the oxidation at 300°C also changes the surface from hydrophobic to hydrophilic, which is<br />
important in many biological applications under physiological con<strong>di</strong>tions. Increased<br />
temperature also increases the extent of oxidation lea<strong>di</strong>ng to completely oxi<strong>di</strong>zed PSi at<br />
around 900–1000°C [20]. Due to the structural expansion, the pore <strong>di</strong>ameter and porosity<br />
are dependent on the extent of oxidation, and a drastic drop in the specific surface area<br />
has been observed in PSi oxi<strong>di</strong>zed around 600°C [21].<br />
Once oxi<strong>di</strong>zed, Si-O bonds are easy to prepare on porous silicon by oxidation, and a<br />
variety of chemical or electrochemical oxidants can be used. There are also many other<br />
techniques to oxi<strong>di</strong>ze PSi, such as ano<strong>di</strong>c oxidation, [22,23] photo-oxidation, [24-26] and<br />
chemical oxidation. Like the thermal oxidation, the chemical oxidation is quite simple, in<br />
which PSi is oxi<strong>di</strong>zed with inorganic or organic agents. [27-31]<br />
Milder chemical oxidants, such as <strong>di</strong>methyl sulfoxide (DMSO), or piranha solutions<br />
(H2SO4/H2O2 7:3) can also be used for this reaction. Mild oxidants are sometimes<br />
preferred because they can improve the mechanical stability of highly porous Si films,<br />
which are typically quite fragile. The mechanical instability of PSi is <strong>di</strong>rectly related to the<br />
strain that is induced in the film as it is produced in the electrochemical etching process,<br />
and the volume expansion that accompanies thermal oxidation can also introduce strain.<br />
Electrochemical oxidation, in which a porous Si sample is ano<strong>di</strong>zed in the presence of a<br />
mineral acid such as H2SO4, yields a fairly stable oxide. Mild chemical oxidants<br />
presumably attack PSi preferentially at Si–Si bonds that are the most strained, and hence<br />
most reactive.<br />
Aqueous solutions of bases such as KOH can also be used to enlarge the pores after<br />
etching [32]. Oxidation imparts hydrophilicity to the porous structure, enabling the bin<strong>di</strong>ng<br />
and adsorption of hydrophilic biomolecules within the pores.<br />
References<br />
1. Optical Biosensors, Eds. F.S. Ligler and C.A. Rowe Taitt, Elsevier, Amsterdam, The Netherlands,<br />
2004.<br />
2. Dancil, K. -P. S.; Greiner, D. P.; Sailor, M. J., A Porous Silicon Optical Biosensor: Detection of<br />
Reversible Bin<strong>di</strong>ng of IgG to a Protein A-Mo<strong>di</strong>fied Surface, J. Am. Chem. Soc. 1999, 121, 7925-<br />
7930.<br />
3. Snow, P. A.; Squire, E. K.; Russel, P. St. J.; Canham, L. T., Vapor sensing using the optical<br />
properties of porous silicon Bragg mirrors, J. Appl. Phys. 1999, 86, 1781-1784.<br />
4. Lorenzo, E.; Oton, C. J.; Capuj, N. E.; Ghulinyan, M.; Navarro-Urrios, D.; Gaburro, Z.; Pavesi, L.,<br />
Porous silicon-based rugate filters, Appl. Opt. 2005, 44, 5415-5421.<br />
5. Ouyang, H.; Chrtistophersen, M.; Viard, R.; Miller, B. L.; Fauchet, P. M., Macroporous silicon<br />
microcavities for macromolecule detection, Adv. Funct. Mater. 2005, 15, 1851-1859.<br />
6. Buriak JM, Allen MJ. 1998. Lewis acid me<strong>di</strong>ated functionalization of porous silicon with substituted<br />
alkenes and alkynes. J Am Chem Soc 120: 1339–1340.<br />
7. Buriak JM, Sterward MP, Geders TW, Allen MJ, Choi HC, Smith J, Raftery D, Canham LT. 1999.<br />
Lewis acid me<strong>di</strong>ated hydrosilylation porous silicon surfaces. J AmChem Soc 121:11491–11502.<br />
8. Stewart MP, Buriak JM. 1998. Photopatterned hydrosilylation on porous silicon. Angew Chem Int Ed<br />
37:3257–3260.<br />
9. Stewart MP, Buriak JM. 2001. Exciton-me<strong>di</strong>ated hydrosilylation on photoluminescent nanocrystalline<br />
silicon. J Am Chem Soc 123:7821–7830.<br />
31
10. Robins EG, Stewart MP, Buriak JM. 1999. Ano<strong>di</strong>c and catho<strong>di</strong>c electrografting of alkynes on porous<br />
silicon. Chem Comm 24:2479–2480.<br />
11. M.R. Linford, C.E.D. Chidsey, Alkyl monolayers covalently attached to silicon surfaces, J. Am. Chem.<br />
Soc. 115 (1993) 12631–12632.<br />
12. R. Boukherroub, J.T.C. Wojtyk, D.D.M. Wayner, D.J. Lockwood, Thermal hydrosilylation of<br />
undecylenic acid with porous silicon, J. Electrochem. Soc. 149 (2002) 59–63.<br />
13. M.P. Stewart, J.M. Buriak, Photopatterned hydrosilylation on porous silicon, Angew. Chem. Int. Ed.<br />
Engl. 37 (1998) 3257–3260.<br />
14. J.M. Buriak, Organometallic chemistry on silicon and germanium surfaces, Chem. Rev. 102 (2002)<br />
1272–1308.<br />
15. J. Terry, M.R. Linford, C. Wigren, R.Y. Cao, P. Pianetta, C.E.D. Chidsey, Alkylterminated Si(111)<br />
surfaces: a high-resolution, core level photoelectron spectroscopy study, J. Appl. Phys. 85 (1999)<br />
213–221.<br />
16. M. Bjorkqvist, J. Salonen, E. Laine, L. Niinisto, Comparison of stabilizing treatments on porous<br />
silicon for sensor applications, Phys. Status Soli<strong>di</strong> A—Appl. Res. 197 (2003) 374–377.<br />
17. A.M. Kaukonen, L. Laitinen, J. Salonen, et al., Enhanced in vitro permeation of furosemide loaded<br />
into thermally carbonized mesoporous silicon (TCPSi) microparticles, Eur. J. Pharm. Biopharm. 66<br />
(2007) 348–356<br />
18. A. Janshoff, K.P.S. Dancil, C. Steinem, et al., Macroporous p-type silicon Fabry– Perot layers.<br />
Fabrication, characterization, and applications in biosensing, J. Am. Chem. Soc. 120 (1998) 12108–<br />
12116.<br />
19. Salonen J, Lehto VP, Laine E. 1997. Thermal oxidation of free-stan<strong>di</strong>ng porous silicon films. Appl<br />
Phys Lett 70:637–639.<br />
20. Canham LT. 1997. Chemical composition of intentionally oxi<strong>di</strong>zed porous silicon. In: Canham LT,<br />
e<strong>di</strong>tor. Properties of porous silicon. EMIS Datareview series No 18. London: INSPEC. pp 158– 162.<br />
21. Halimaoui A. 1995. Porous silicon: Material processing, properties and applications. In: Vial JC,<br />
Derrien J, e<strong>di</strong>tors. Porous silicon science and technology. France: Springer-Verlag. pp 33–52.<br />
22. Herino R. 1995. Luminescence of porous silicon after electrochemical oxidation. In: Vial JC, Derrien<br />
J, e<strong>di</strong>tors. Porous silicon science and technology. France: Springer-Verlag. pp 53–66.<br />
23. Mulloni V, Pavesi L. 2000. Electrochemically oxi<strong>di</strong>zed porous silicon microcavities. Mat Sci Eng B<br />
69–70:59–65.<br />
24. Tischler MA, Collin RT, Stathis JH, Tsang JC. 1992. Luminescence degradation in porous silicon.<br />
Appl Phys Lett 60:639–641.<br />
25. Salonen J, Lehto VP, Laine E. 1999. Photo-oxidation stu<strong>di</strong>es of porous silicon using<br />
microcalorimetric method. J Appl Phys 86:5888–5893.<br />
26. Gole JL, Dixon DA. 1998. Evidence for oxide formation of the single and multiphoton excitation of a<br />
porous silicon surface or silicon ‘‘nanoparticles’’. J Appl Phys 83:5985–5991.<br />
27. Song JH, Sailor MJ. 1999. Chemical mo<strong>di</strong>fication of crystalline porous silicon surfaces. Comments<br />
Inorg Chem 1-3:69–84.<br />
28. Nakajima A, Itakura T, Watanabe S, Nakayama N. 1992. Photoluminescence of porous Si, oxi<strong>di</strong>zed<br />
and then deoxi<strong>di</strong>zed chemically. Appl Phys Lett 61:46–48.<br />
29. Rao BVRM, Basu PK, Biswas JC, Lahiri SK, Ghosh S, Bose DN. 1996. Large enhancement of<br />
photoluminescence from porous silicon films by post-ano<strong>di</strong>zation treatment in boiling hydrogen<br />
peroxide. Solid State Comm 97:417–418.<br />
30. Yin F, Li XP, Zhang ZZ, Xiao XR. 1997. Investigations on the surface reactivity of luminescent<br />
porous silicon. Appl Surf Sci 119:310–312.<br />
31. Salonen J, Lehto VP, Laine E. 1997. The room temperature oxidation of porous silicon. Appl Surf Sci<br />
120:191–198.<br />
32. L. A. De Louise and B. L. Miller, Mat. Res. Soc. Symp. Proc. 782, A5.3.1 (2004).<br />
32
DNA Optical Detection Based on Porous Silicon Technology<br />
This section is aimed at the optimization of the functionalization process investigating also<br />
the role of thickness and porosity of the PSi chip, the time exposure to ultraviolet (UV) light<br />
and also the concentration of the ssDNA solution.<br />
In collaboration with Prof. P. Arcari of University of Naples “Federico II”, Italy
Sensors 2007, 7, 214-221<br />
Full Research Paper<br />
sensors<br />
ISSN 1424-8220<br />
© 2007 by MDPI<br />
www.mdpi.org/sensors<br />
DNA Optical Detection Based on Porous Silicon Technology:<br />
from Biosensors to Biochips<br />
Luca De Stefano 1, *, Paolo Arcari 2 , Annalisa Lamberti 2 , Carmen Sanges 2 , Lucia Rotiroti 1,3 ,<br />
Ilaria Rea 1, 4 and Ivo Ren<strong>di</strong>na 1<br />
1<br />
Institute for Microelectronics and Microsystems – Unit of Naples – National Council of Research, Via<br />
P. Castellino 111, 80131 Napoli, Italy (E-mail: luca.destefano@na.imm.cnr.it)<br />
2<br />
Department of Biochemistry and Me<strong>di</strong>cal Biotechnologies, University of Naples “Federico II”, Via S.<br />
Pansini 5, 80131 Napoli, Italy.<br />
3<br />
Department of Organic Chemistry and Biochemistry, University of Naples “Federico II”, Via Cinthia,<br />
80128 Napoli, Italy<br />
4<br />
Department of Physical Sciences, University of Naples “Federico II”, Via Cinthia, 80126 Naples, Italy<br />
* Author to whom correspondence should be addressed. E-mail: luca.destefano@na.imm.cnr.it<br />
Received: 15 February 2007 / Accepted: 26 February 2007 / Published: 28 February 2007<br />
Abstract: A photochemical functionalization process which passivates the porous silicon<br />
surface of optical biosensors has been optimized as a function of the thickness and the<br />
porosity of the devices. The surface mo<strong>di</strong>fication has been characterized by contact angle<br />
measurements. Fluorescence measurements have been used to investigate the stability of the<br />
DNA single strands bound to the nanostructured material. A dose-response curve for an<br />
optical label-free biosensor in the 6-80 μM range has been realized.<br />
Keywords: Optical Biosensors, Porous Silicon, Biochip.<br />
1. Introduction<br />
Biosensors are nowadays technological hot topics due to the possible applications in social interest<br />
fields such as me<strong>di</strong>cal <strong>di</strong>agnostic and health care, monitoring of environmental pollutants, home and<br />
defence security [1]. Besides the signal generated by the sensing device, the biosensor is constituted by<br />
the molecular recognition element and the transducer material. The molecular recognition element can<br />
be a biological molecule, such as DNA single strand, proteins, enzymes, or a biological system, such as
Sensors 2007, 6<br />
membrane, cell, and tissues: in this way, the sensing mechanism takes advantage of the natural<br />
sensitivity and specificity of the biomolecular interactions. Optical transduction is more and more used<br />
since photonic devices could be small, lightweight and thus portable due to the integrability of all optical<br />
components. Furthermore, optical devices do not require electric contacts. Fluorescence is by far the<br />
most used optical signalling method but a wide-use sensor can not be limited by the labelling of the<br />
probe nor the analyte, since this step is not always possible [2]. Reagentless optical biosensors are<br />
monitoring devices which can detect a target analyte in a heterogeneous solution without the ad<strong>di</strong>tion of<br />
anything than the sample. In the fields of genomics and proteomics this is a straightforward advantage<br />
since it allows real-time readouts and, thus, very high throughoutputs analysis. A label free optical<br />
biosensor can be realized by integrating the biological probe with a signalling material which <strong>di</strong>rectly<br />
transduces the molecular recognition event into an optical signal. Recently, lot of theoretical and<br />
experimental work, concerning the worth noting properties of nanostructured porous silicon (PSi) in<br />
chemical and biological sensing, has been reported, showing that, due to its morphological and physical<br />
properties, PSi is a very versatile sensing platform [3, 4]. PSi is an available, low cost material,<br />
completely compatible with VLSI and micromachining technologies, so that it could usefully be<br />
employed in the fabrication of micro-opto-electric-mechanical system and smart sensors. PSi is also very<br />
used as photonic material due to the possibility of fabricating high quality optical structures, either as<br />
single layers, like Fabry-Perot interferometers [5], or multilayers, such as Bragg [6] or rugate filters [7].<br />
A key feature for a recognition transducer is a large surface area: PSi has a porous structure with a<br />
specific area up to 200 – 500 m 2 cm -3 , so that it can be very sensitive to the presence of biochemical<br />
species which penetrate inside the pores. Unfortunately, the surface of the “as etched” PSi is highly<br />
hydrophobic so that aqueous solution can not infiltrate the sponge like matrix. A proper passivation<br />
process must be applied to stabilise the surface and to covalently link the bioprobe [8]. The<br />
development of PSi based biosensor arrays critically depends on the surface functionalization process<br />
and how it is compatible with the microfabrication technologies. In a recently published article, we have<br />
exploited a photochemical functionalization process of the PSi surface to covalently bind DNA single<br />
strands (ssDNA) and we have also demonstrate that the device works as an all optical biosensor for<br />
ssDNA-cDNA interactions [9]. In the present work, we have optimised the functionalization process by<br />
investigating the role of thickness and porosity of the PSi chip, the time exposure to ultraviolet (UV)<br />
light and also the concentration of the ssDNA solution. Fluorescent measurements have been used to<br />
test the chip stability against the washing in aqueous solutions.<br />
2. Experimental Section<br />
In this study we used as optical transducer a porous silicon layer of fixed thickness and porosity:<br />
from an optical point of view, this structure acts as a Fabry-Perot interferometer. To study the influence<br />
of porous silicon physical parameters, we have fabricated several layers, obtained by electrochemical<br />
etch in a HF/EtOH (3:7) solution, at room temperature and dark light. Highly doped p + -silicon, <br />
oriented, 0.005 Ω cm resistivity, 400 µm thick was used. The PSi samples were characterised by<br />
variable angle spectroscopic ellipsometry [10]. Etching times and ano<strong>di</strong>c current densities have been<br />
reported in Table 1.<br />
215
Sensors 2007, 6<br />
Table 1. PSi layers physical characteristics and fabrication parameters.<br />
Current Density<br />
(mA/cm 2 )<br />
Porosity<br />
(%)<br />
Etch Rate<br />
(nm/s)<br />
50 60 193<br />
125 70 247<br />
150 80 137<br />
Thickness Etch Time<br />
(μm) (s)<br />
2.0 10.4<br />
4.0 20.7<br />
6.0 31.1<br />
2.0 8.1<br />
4.0 16.2<br />
6.0 24.3<br />
2.0 14.6<br />
4.0 29.2<br />
6.0 43.8<br />
The photo-activated chemical mo<strong>di</strong>fication of PSi surface was based on the UV exposure of a<br />
solution of alkenes which bring some carboxylic acid groups. The PSi chip has been pre-cleaned in an<br />
ultrasonic acetone bath then washed in deionized water. After dried in N2 stream, it has been<br />
imme<strong>di</strong>ately covered with 10% N-hydroxysuccinimide ester (UANHS) solution in CH2Cl2. The UANHS<br />
was house synthesised as described in literature [11]. This treatment results in covalent attachment of<br />
UANHS to the porous silicon surface. The chip was then washed in <strong>di</strong>chloromethane in an ultrasonic<br />
bath for 10 min to remove any adsorbed alkene from the surface. The carboxyl-terminated monolayer<br />
covering the PSi surface works as a reactive substrate for the chemistry of the subsequent attachment of<br />
the DNA sequences. DNA single strands in a HEPES solution 10mM (pH=7.5) have been incubated<br />
overnight.<br />
FT-IR spectroscopy (Thermo - Nicholet NEXUS) has been used to verify the efficiency of the<br />
reaction. After the chemical functionalization we have also quantitatively measured the efficiency of the<br />
bin<strong>di</strong>ng between the DNA and the porous silicon surface using a fluorescent DNA probe<br />
(5'GGACTTGCCCGAATCTACGTGTCCA3', Primm) labelled with a proper chromophore group<br />
(Fluorescein CY3.5, the absorption peak is at 581 nm and the emission is at 596 nm). Fluorescence<br />
images were recorded by a Leica Z16 APO fluorescence macroscopy system. The fluorescent chips<br />
have been <strong>di</strong>alyzed overnight, first in water and then in a HEPES solution, at room temperature to<br />
assess the bin<strong>di</strong>ng between the bioprobe and the PSi surface.<br />
Contact angle measurements have been performed by using a KSV Instruments LTD CAM 200<br />
Optical Contact Angle Meter.<br />
The reflectivity measurements have been performed by a very simple experimental set-up: a tungsten<br />
lamp (400 nm < λ < 1800 nm) illuminates, through an optical fiber and a collimator, the sensor and the<br />
reflected beam is collected by an objective, coupled into a multimode fiber, and then <strong>di</strong>rected in an<br />
optical spectrum analyser (Ando, AQ6315A). The reflectivity spectra have been measured with a<br />
resolution of 0.5 nm.<br />
216
Sensors 2007, 6<br />
3. Results and Discussion<br />
Infrared spectroscopy is a powerful tool in surfaces characterisation: it is fast, accurate and could<br />
also performe quantitative determinations. We have investigated all the PSi monolayers before and after<br />
the photochemical passivation process since we were worried about the functionalization efficiency of<br />
thicker samples. A thicker layer can adsorb more bioprobes than a thinner one but the UV exposure<br />
could also be less effective, due to light absorption by PSi at those frequencies. Our measurements<br />
demonstrate that up to 6 μm thick PSi monolayer, a complete passivation and functionalization of the<br />
surface can be obtained by exposing for a sufficiently long time the sample: in Figure 1A are shown the<br />
FT-IR spectra of the PSi monolayer as etched and after three <strong>di</strong>fferent times of exposure to UV light.<br />
% Reflectance<br />
140<br />
120<br />
100<br />
80<br />
60<br />
40<br />
20<br />
0<br />
C-H 2<br />
6 μm 80% porosity<br />
after 4'20" exposure to UANHS<br />
after 8'40" exposure to UANHS<br />
after 13' exposure to UANHS<br />
Si-H<br />
O-C=O<br />
Succinimmide<br />
C-N<br />
C=O<br />
3000 2500 2000 1500 1000<br />
Wavenumber (cm -1 )<br />
A<br />
Peaks Area (%)<br />
100<br />
80<br />
60<br />
40<br />
20<br />
0<br />
Si-C<br />
20<br />
802300<br />
2200 2100 2000<br />
900 850 800 750 700<br />
Figure 1. A) FT-IR spectra of the PSi monolayer as etched and after three <strong>di</strong>fferent times of exposure<br />
to UV light. B) Particulars of the Si-H bond and Si-C bond peaks at 2100 and 880 cm -1 , respectively. C)<br />
Peaks area as function of the exposure time: the reaction yield increases monotonically with the<br />
exposure time.<br />
% Reflectance<br />
100<br />
80<br />
60<br />
40<br />
60<br />
40<br />
20<br />
Si-C<br />
0 2 4 6 8 10 12 14<br />
C<br />
UV exposure tim e (min)<br />
SiH<br />
SiC<br />
Si-H<br />
W avenumber (cm -1 )<br />
B<br />
217
Sensors 2007, 6<br />
The characteristic peaks of the Si-H bonds (at 2100 cm -1 ) progressively <strong>di</strong>sappear while the amide I<br />
band (at 1634 cm -1 ) and the Si-C peak (at 880 cm -1 ) become more and more evident. Figure 1B is an<br />
enlargement of the characteristic peaks of these resonances. The reaction yield increases quite<br />
monotonically with the exposure time, as it can be seen in Figure 1C. Thinner samples can be<br />
functionalised in faster exposure times, but the overall volume available to adsorb bioprobes is<br />
drastically reduced.<br />
A B<br />
Figure 2. Characterisation of the PSi surface by contact angles measurements. A) The PSi as etched<br />
shows a clearly hydrophobic behaviour. B) After the passivation process the PSi surface is hydrophilic<br />
thus allowing the penetration of biological solution.<br />
The surface passivation is also well evident by contact angle measurements: the hydrogenated PSi<br />
surface is highly hydrophobic, due to the presence of the Si-H bonds, so that a drop of deionised water<br />
exploits a high surface tension which prevents its <strong>di</strong>ffusion in the pores. This behaviour is well depicted<br />
in Figure 2A: the contact angle is 137±1 degree, averaging the left and right values of three <strong>di</strong>fferent<br />
drops having the same volumes, and the estimated surface tension is 179 mN/m. After the photoreaction<br />
with the UANHS, the Si-H bonds are replaced by the Si-C bonds and the PSi surface shows a<br />
hydrophilic behaviour, as shown in Figure 2B: the drop spreads on the surface and the contact angle<br />
drastically decreases to 43±3 degree and also the surface tension is lowered to 67 mN/m. In this<br />
con<strong>di</strong>tion, an aqueous solution can be infiltrated in the nanostructured layer.<br />
To test the stability of the covalent bon<strong>di</strong>ng between the organic linker layers, which homogeneously<br />
cover the PSi surface, and the biological probes we have used a fluorescent DNA single strand as an<br />
optical tracer. After the chemical bon<strong>di</strong>ng of the labelled ssDNA, the chip was observed by the<br />
fluorescence macroscopy system. Under the light of the 100W high-pressure mercury source, we have<br />
found a high and homogeneous fluorescence on the whole chip surface which still remains bright even<br />
after two overnight <strong>di</strong>alysis washings in a HEPES solution and in deionised water, as it can be seen in<br />
Figures 3 A, B, and C. We have also stu<strong>di</strong>ed the yield of the chemical functionalization by spotting<br />
<strong>di</strong>fferent concentrations of the fluorescent ssDNA and measuring the fluorescence intensities of the<br />
images before and after the washings. The results reported in Figure 4 confirm the qualitative fin<strong>di</strong>ngs of<br />
Figure 3: the fluorescent intensities decrease but remain of the same order of magnitude. From this<br />
218
Sensors 2007, 6<br />
graph we can also estimate the concentration of the DNA probe which saturates the bin<strong>di</strong>ng sites<br />
available.<br />
The PSi optical biosensors measure the change in the average refractive index of the device: when a<br />
bioconjugation event takes place, the refractive index of the molecular complex changes and the<br />
interference pattern on output is thus mo<strong>di</strong>fied. The label free optical monitoring of the ssDNA-cDNA<br />
hybri<strong>di</strong>zation is simply the comparison between the optical spectra of the porous silicon layer after the<br />
UANHS and probe immobilization on the chip surface and after its hybri<strong>di</strong>zation with the cDNA. Each<br />
step of the chip preparation increases the optical path in the reflectivity spectrum recorded, due to the<br />
substitution of the air into the pores by the organic and biological compounds. The interaction of the<br />
ssDNA with its complementary sequence has been detected as a fringes shift in the wavelengths, which<br />
corresponds to a change in the optical path. Since the thickness d is fixed by the physical <strong>di</strong>mension of<br />
the PSi matrix, the variation is clearly due to changes in the average refractive index.<br />
A B C<br />
Figure 3: A) Fluorescence of the chip surface after the bin<strong>di</strong>ng of the labelled ssDNA; B) after the<br />
overnight <strong>di</strong>alysis in HEPES solution; C) after the overnight <strong>di</strong>alysis in deionised water.<br />
In Figure 5A the reflectivity spectra of the PSi layer for <strong>di</strong>fferent cDNA concentration are reported,<br />
while in Figure 5B a dose-response curve is reported. A control measurement has been made using a<br />
ncDNA sequence: a very small shift (less than 2 nm) has been recorded in the reflectivity spectrum<br />
respect to the one obtained after the probe linking.<br />
Intensity (a.u.)<br />
225<br />
200<br />
175<br />
150<br />
125<br />
Chip as spotted with labelled DNA<br />
Chip after HEPES washing<br />
Chip after water washing<br />
75 150 225 300 375 450 525 600<br />
DNA Concentration (μM)<br />
Figure 4. Fluorescence intensities of the chip surface after the bin<strong>di</strong>ng of the labelled ssDNA and the<br />
two overnight <strong>di</strong>alysis as a function of the cDNA concentration.<br />
219
Sensors 2007, 6<br />
The sensor response has been fitted by a monoexponential growth model, y=A-Be -kx , where A is the<br />
offset, B the amplitude, k the rate and y’=Bk the limiting sensitivity, i.e. the sensitivity in the limit of<br />
zero ligand concentration. In this case, we obtained for this parameter the value of 1.1 (0.1) nm/μM,<br />
which corresponds to a limit of detection of 90 nM for a system able to detect a wavelength shift of 0.1<br />
nm.<br />
Reflectivity (a.u.)<br />
1.4<br />
1.2<br />
1.0<br />
0.8<br />
0.6<br />
0.4<br />
0.2<br />
0.0<br />
1150 1200 1250 1300 1350<br />
Figure 5. A) Fringes shifts due to ssDNA-cDNA interaction. B) Dose-response curve as a function of<br />
the cDNA concentration.<br />
In conclusions, we have optimized the PSi surface functionalization process by investigating the role<br />
of chip thickness and porosity, on exposure to ultraviolet (UV) light for <strong>di</strong>fferent times interval and also<br />
for <strong>di</strong>fferent concentration of the ssDNA probe. Fluorescent measurements have confirmed the chip<br />
stability against the washing in aqueous solutions. These results will be very useful in the design and<br />
realisation of a PSi based label free optical biochip.<br />
Acknowledgements<br />
Authors wish to acknowledge dr. Andrea M. Rossi of National Institute of Metrological Research,<br />
Turin, Italy, for helpful <strong>di</strong>scussions and Prof. N. Scaramuzza and dr. M. Giocondo of University of<br />
Calabria, Cosenza, Italy, for contact angle measurements. This work is partially supported by the FIRB<br />
Italian project RBLA033WJX_005.<br />
References and Notes<br />
A<br />
ss-DNA<br />
ss-DNA+c-DNA 6μM<br />
ss-DNA+c-DNA 15μM<br />
ss-DNA+c-DNA 60μM<br />
Wavelength (nm)<br />
1. Optical Biosensors, Eds. F.S. Ligler and C.A. Rowe Taitt, Elsevier, Amsterdam, The<br />
Netherlands, 2004.<br />
2. Salitermkan, S. S. Fundamentals of BioMEMS and Me<strong>di</strong>cal Microdevices; Wiley-Interscience,<br />
SPIE PRESS (Bellingham, Washington USA), 2006.<br />
3. Lin, V. S. Y.; Motesharei, K.; Dancil, K. P. S.; Sailor, M. J.; Gha<strong>di</strong>ri, M. R., A porous siliconbased<br />
optical interferometric biosensor, Science 1997, 278, 840-843.<br />
Δλ Δλ (nm)<br />
35<br />
30<br />
25<br />
20<br />
15<br />
10<br />
5<br />
0<br />
B<br />
0 20 40 60 80<br />
c-DNA concentration (μM)<br />
220
Sensors 2007, 6<br />
4. De Stefano, L.; Moretti, L.; Ren<strong>di</strong>na, I.; Rossi, A. M., Time-resolved sensing of chemical species<br />
in porous silicon optical microcavity, Sensor and Actuators B 2004, 100, 168-172.<br />
5. Dancil, K. -P. S.; Greiner, D. P.; Sailor, M. J., A Porous Silicon Optical Biosensor: Detection of<br />
Reversible Bin<strong>di</strong>ng of IgG to a Protein A-Mo<strong>di</strong>fied Surface, J. Am. Chem. Soc. 1999, 121, 7925-<br />
7930.<br />
6. Snow, P. A.; Squire, E. K.; Russel, P. St. J.; Canham, L. T., Vapor sensing using the optical<br />
properties of porous silicon Bragg mirrors, J. Appl. Phys. 1999, 86, 1781-1784.<br />
7. Lorenzo, E.; Oton, C. J.; Capuj, N. E.; Ghulinyan, M.; Navarro-Urrios, D.; Gaburro, Z.; Pavesi,<br />
L., Porous silicon-based rugate filters, Appl. Opt. 2005, 44, 5415-5421.<br />
8. Ouyang, H.; Chrtistophersen, M.; Viard, R.; Miller, B. L.; Fauchet, P. M., Macroporous silicon<br />
microcavities for macromolecule detection, Adv. Funct. Mater. 2005, 15, 1851-1859.<br />
9. De Stefano, L.; Rotiroti, L.; Rea, I.; Ren<strong>di</strong>na, I.; Moretti, L.; Di Francia, G.; Massera, E.;<br />
Lamberti, A.; Arcari, P.; Sangez, C., Porous silicon-based optical biochips, Journal of Optics A:<br />
Pure and Applied Optics 2006, 8, S540-S544.<br />
10. Wongmanerod, C.; Zangooie, S.; Arwin, H., Determination of pore size <strong>di</strong>stribution and surface<br />
area of thin porous silicon layers by spectroscopic ellipsometry, Applied Surface Science 2001,<br />
172, 117-125.<br />
11. Yin, H.B.; Brown, T.; Gref, R.; Wilkinson, J.S.; Melvin, T., Chemical mo<strong>di</strong>fication and<br />
micropatterning of Si(100) with oligonucleotides, Microelectronic Engineering 2004, 73-74, 830-<br />
836.<br />
© 2007 by MDPI (http://www.mdpi.org). Reproduction is permitted for noncommercial purposes.<br />
221
Examples of optical biosensors for the detection of analytes of<br />
clinical, industrial, and home security interest.<br />
This section is aimed at the characterization of a PSi device functionalized with some<br />
proteins and enzymes.<br />
It was stu<strong>di</strong>ed that Glutamine-bin<strong>di</strong>ng protein (GlnBP) from Escherichia coli specifically<br />
binds Glutamine and a wheat glia<strong>di</strong>n peptide (mainly responsible of inducing celiac<br />
<strong>di</strong>sease), and that acetylcholinesterase (AChE) activity can be inhibited with<br />
organophosphorous pesticides, nerve agents and nitro-aromatic compounds, very<br />
dangerous substances.<br />
In collaboration with Dr. S. D’Auria from Institute for Protein Biochemistry, National Council<br />
of Research, Italy
Glutamine-Bin<strong>di</strong>ng Protein from Escherichia coli Specifically Binds a<br />
Wheat Glia<strong>di</strong>n Peptide Allowing the Design of a New Porous<br />
Silicon-Based Optical Biosensor †<br />
Luca De Stefano, ‡ Mauro Rossi, § Maria Staiano, | Gianfranco Mamone, § Antonoietta Parracino, |<br />
Lucia Rotiroti, ‡ Ivo Ren<strong>di</strong>na, ‡ Mosè Rossi, | and Sabato D’Auria* ,|<br />
Institute for Microelectronics and Microsystems, CNR, Naples, Italy, Institute of Food Sciences, CNR, Avellino,<br />
Italy, and Institute of Protein Biochemistry, CNR, Naples, Italy<br />
Received January 20, 2006<br />
In this work, the bin<strong>di</strong>ng of the recombinant glutamine-bin<strong>di</strong>ng protein (GlnBP) from Escherichia coli<br />
to glia<strong>di</strong>n peptides, toxic for celiac patients, was investigated by mass spectrometry experiments and<br />
optical techniques. Mass spectrometry experiments demonstrated that GlnBP binds the following amino<br />
acid sequence: XXQPQPQQQQQQQQQQQQL, present only into the toxic prolamines. The bin<strong>di</strong>ng of<br />
GlnBP to glia<strong>di</strong>n suggested us to design a new optical biosensor based on nanostructured porous<br />
silicon (PSi) for the detection of trace amounts of glia<strong>di</strong>n in food. The GlnBP, which acts as a molecular<br />
probe for the glia<strong>di</strong>n, was covalently linked to the surface of the PSi wafer by a proper passivation<br />
process. The GlnBP-glia<strong>di</strong>n interaction was revealed as a shift in wavelength of the fringes in the<br />
reflectivity spectrum of the PSi layer. The GlnBP, covalently bonded to the PSi chip, selectively<br />
recognized the toxic peptide. Finally, the sensor response to the protein concentration was measured<br />
in the range 2.0-40.0 µg/L and the sensitivity of the sensor was determined.<br />
Keywords: glia<strong>di</strong>n • optical biosensors • porous silicon • celiac <strong>di</strong>sease<br />
There is a strong need for integrated and automated bioanalytical<br />
systems in me<strong>di</strong>cal, food, agricultural, environmental,<br />
and defense testing. Now, the main market is shared between<br />
blood glucose, pregnancy, antibody-based infectious <strong>di</strong>seases,<br />
and biological warfare agent detection. 1 The interaction between<br />
an analyte and a biological recognition system is<br />
normally detected in biosensors by the transducer element<br />
which converts the molecular event into a measurable effect,<br />
such as an electrical or optical signal. Because of its spongelike<br />
structure, porous silicon (PSi) is an almost ideal material as a<br />
transducer: its surface has a specific area of the order of 200-<br />
500 m 2 cm -3 , so that a very effective interaction with liquid or<br />
gaseous substances is assured. 2,3 PSi optical sensors are based<br />
on changes of photoluminescence or reflectivity when exposed<br />
to the target analytes which substitute the air into the PSi<br />
pores. 4 The effect depends on the chemical and physical<br />
properties of each analyte, so that the sensor can be used to<br />
recognize the pure substances. Because of the sensing mechanism,<br />
these kinds of devices are not able to identify the<br />
components of a complex mixture. To enhance the sensor<br />
selectivity through specific interactions, some researchers have<br />
† The authors wish to de<strong>di</strong>cate this work to the memory of Prof. Arturo<br />
Leone, Director of the Institute of Food Sciences, CNR, Avellino, Italy.<br />
* Addess correspondence to Dr. Sabato D’Auria, Institute of Protein<br />
Biochemistry, Consiglio Nazionale delle Ricerche, Via Pietro Castellino 111,<br />
80131 Naples, Italy. Fax, +39-0816132277; e-mail, s.dauria@ibp.cnr.it.<br />
‡ Institute for Microelectronics and Microsystems, CNR.<br />
§ Institute of Food Sciences, CNR.<br />
| Institute of Protein Biochemistry, CNR.<br />
proposed to chemically or physically mo<strong>di</strong>fy the PSi surface;<br />
the common approach is to create a covalent bond between<br />
the porous silicon surface and the biomolecules which specifically<br />
recognize the unknown analytes. 5-7 Among <strong>di</strong>fferent<br />
probes of biological nature, ligand-bin<strong>di</strong>ng proteins are particularly<br />
good can<strong>di</strong>dates in designing highly specific biosensors<br />
for small analytes; in particular, the glutamine-bin<strong>di</strong>ng protein<br />
(GlnBP) from Escherichia coli is a monomeric protein composed<br />
of 224 amino acid residues (26 kDa) responsible for the<br />
first step in the active transport of L-glutamine across the<br />
cytoplasm membrane. The GlnBP consists of two similar<br />
globular domains, the large domain (residues 1-84 and 186-<br />
224) and the small domain (residues 90-180), linked by two<br />
peptides. The deep cleft formed between the two domains<br />
contains the ligand-bin<strong>di</strong>ng site. The GlnBP binds L-glutamine<br />
with a <strong>di</strong>ssociation constant Kd of 5 × 10 -9 M 8 as well as polyglutamine<br />
residues. In this study, we show that GlnBP is able<br />
to bind with high affinity the amino acid sequences present in<br />
the glia<strong>di</strong>n allowing the design of a new reagentless microsensor<br />
for the optical interferometric detection of glia<strong>di</strong>n based<br />
on a chemically mo<strong>di</strong>fied porous silicon nanostructured surface<br />
covalently linking the GlnBP. The glia<strong>di</strong>n by the wheat gluten<br />
is mainly responsible of inducing celiac <strong>di</strong>sease, an inflammatory<br />
<strong>di</strong>sease of the small intestine affecting genetically susceptible<br />
people. 9 Presently, a strict gluten-free lifelong <strong>di</strong>et is<br />
mandatory for celiac patients for both intestinal mucosal<br />
recovery and prevention of complicating con<strong>di</strong>tions such as<br />
lymphoma. However, <strong>di</strong>etary compliance has been shown to<br />
10.1021/pr0600226 CCC: $33.50 © 2006 American Chemical Society Journal of Proteome Research 2006, 5, 1241-1245 1241<br />
Published on Web 03/21/2006
esearch articles De Stefano et al.<br />
be poor in most patients mainly because of inadvertent gluten<br />
consumption. From this point of view, a very sensitive assay<br />
for detection of toxic components of prolamins in food<br />
represents a perspective worth pursuing. Furthermore, the<br />
glia<strong>di</strong>n is a mixture of many proteins. 10 This observation<br />
highlighted the <strong>di</strong>fficulties in identifying sequences useful for<br />
developing an immunoassay addressed toward the toxic portions<br />
of glia<strong>di</strong>n. The obtained results show that, even if the<br />
protein is blocked on the PSi surface, the GlnBP can strongly<br />
bind the substrate and work as a molecular probe for the<br />
detection of glia<strong>di</strong>n.<br />
Materials and Methods<br />
Materials. All the chemicals used were commercial samples<br />
of the purest quality.<br />
Protein Purification. The recombinant glutamine-bin<strong>di</strong>ng<br />
protein from E. coli was prepared and purified accor<strong>di</strong>ng to<br />
D’Auria et al. 7 The protein concentration was determined by<br />
the method of Bradford 11 with bovine serum albumin as<br />
standard on a double beam Cary 1E spectrophotometer (Varian,<br />
Mulgrade, Victoria, Australia).<br />
Mass Spectrometry Analysis. MALDI-TOF experiments were<br />
carried out on a PerSeptive Biosystems (Framingham, MA)<br />
Voyager DE-PRO instrument equipped with a N2 laser (337 nm,<br />
3 ns pulse width). Each spectrum was taken with the following<br />
procedure: 1 µL-aliquot of sample was loaded on a stainless<br />
steel plate together with 1 µL of matrix R-cyano-4-hydroxycinnamic<br />
acid (10 mg in 1 mL of aqueous 50% acetonitrile). Mass<br />
spectrum acquisition was performed in both positive linear and<br />
reflectron mode by accumulating 200 laser pulses. The accelerating<br />
voltage was 20 kV. External mass calibration was<br />
performed with mass peptide standards (SIGMA). Tandem<br />
mass spectrometry (MS/MS) data for the extracted peptides<br />
were obtained using a Q-STAR mass spectrometer (Applied<br />
Biosystems) equipped with a nanospray interface (Protana,<br />
Odense, Denmark). Dried samples were resuspended in 0.1%<br />
TFA, desalted using Zip-Tip C18 microcolumns (Millipore), and<br />
sprayed from gold-coated, ‘me<strong>di</strong>um length’ borosilicate capillaries<br />
(Protana). The capillary voltage used was 800 V. Doubly<br />
charged ion isotopic cluster was selected by the quadrupole<br />
mass filter (MS1) and then induced to fragment by collision.<br />
The collision energy was 30-40 eV, depen<strong>di</strong>ng on the size of<br />
the peptide. The collision-induced <strong>di</strong>ssociation was processed<br />
using Analyst 5 software (Applied Biosystems). The deconvolution<br />
MS/MS was manually interpreted with the help of the<br />
Analyst 5 software.<br />
Porous Silicon Chip Preparation and GlnBP Immobilization.<br />
Proof of PSi in biosensing has been experimentally<br />
demonstrated in a variety of sensors devices, inclu<strong>di</strong>ng rugate<br />
filters, Bragg filters, single-layer films, and protein and DNA<br />
microarray, due to the large internal surface area which<br />
depen<strong>di</strong>ng upon depth can easily be over 1000 times that of a<br />
planar surface of equal <strong>di</strong>ameter, thus, enabling the immobilization<br />
of large amounts of selective receptors. 12-14<br />
However, for the device to function effectively, the porous<br />
morphology should enable homogeneous <strong>di</strong>ffusion everywhere<br />
in the matrix. In the case of porous silicon, the porosity,<br />
determined by pore size and <strong>di</strong>stribution, is tunable over a<br />
rather wide range by properly setting the fabrication parameters,<br />
that is, current density, etching time, etching solution,<br />
and wafer type. It is therefore possible to optimize the porous<br />
matrix structure to achieve maximum <strong>di</strong>ffusion and maintain<br />
high level optical features. Previous ellipsometric characteriza-<br />
1242 Journal of Proteome Research • Vol. 5, No. 5, 2006<br />
tion has shown that the surface of the porous silicon layer as<br />
prepared could be covered by a 100-200 nm thin parasitic film<br />
of very low porosity (
GlnBP Binds Wheat Glia<strong>di</strong>n Toxic Sequences research articles<br />
Figure 1. Photoinduced reaction used to link GlnBP to the PSi chip surface.<br />
Figure 2. MALDI-TOF spectrum of glia<strong>di</strong>n <strong>di</strong>gest bound to GlnBP. The amino acid sequence of MH + 2351.11 was determined by nanoESI-<br />
MS/MS.<br />
37 °C, the reaction was stopped by heating for 5 min; the<br />
peptic-tryptic (PT) <strong>di</strong>gest was freeze-dried and stored at -20<br />
°C.<br />
A solution of homogeneous 2.0 mg/mL GlnBP in 1.0 mL of<br />
0.1 M bicarbonate buffer, pH 9.0, was mixed with 10 µL of<br />
rodhamine (Molecular Probes) solution in N,N-<strong>di</strong>methylformamide<br />
(DMF) (1.0 µg of rodhamine/100 µL of DMF). The<br />
reaction mixture was incubated for 1.0 h at 30 °C, and the<br />
labeled protein was separated from the unreacted probe by<br />
passing over a Sephadex G-25 column equilibrated in 50 mM<br />
phosphate buffer and 100 mM NaCl, pH 7.0.<br />
The immobilization of GlnBP was carried out by applying<br />
50 µL of a 1 mg/mL solution prepared in 100 mM PBE buffer<br />
at pH 6.5 to each chip. The chips were incubated at -4 °C for<br />
1.0 h. After the reaction, the excess of reagent was removed<br />
and the wafers were extensively washed at room temperature<br />
by PBE buffer.<br />
To assess the protein penetration into the pores, we spotted<br />
20 µL of 1.0 mM so<strong>di</strong>um bicarbonate buffer containing the dyelabeled<br />
protein onto the porous silicon chip. This chip was<br />
observed by a confocal microscopy system.<br />
Results and Discussion<br />
GlnBP is a monomeric protein that binds glutamine (Gln)<br />
and poly-Gln residues with high affinity. 7,8 Since gluten proteins<br />
are rich in Gln residues, we questioned if GlnBP was able to<br />
bind amino acid sequences present in gluten proteins such as<br />
glia<strong>di</strong>n, a protein toxic for celiac patients. To check it, GlnBP<br />
was covalently bound to a CNBr-activated Sepharose 4B resin<br />
accor<strong>di</strong>ng to the manufacturer’s instructions (Amersham Biosciences<br />
Europe GmbH, Cologno Monzese, Italy) and a glia<strong>di</strong>n<br />
PT <strong>di</strong>gest was passed through the column. The column was<br />
washed with 3 vol of phosphate buffer saline, and the peptide-<br />
(s) bound to GlnBP were eluted with 0.2 M glycine/HCl, pH<br />
3.0. The eluted fractions were utilized for mass spectrometric<br />
experiments.<br />
MALDI-TOF analysis on the isolated glia<strong>di</strong>n <strong>di</strong>gest gave a<br />
complex peptide profile, which was <strong>di</strong>fficult to rationalize due<br />
to the high sequence variability among the components of the<br />
glia<strong>di</strong>n fraction (Figure 2). In fact, the main <strong>di</strong>fficulty in glia<strong>di</strong>n<br />
structural analysis lies in high protein heterogeneity owed to<br />
numerous duplications, and subsequent <strong>di</strong>vergences of the<br />
glia<strong>di</strong>n multigene family enco<strong>di</strong>ng a polymorphic set of polypeptides.<br />
To identify the derived PT peptides, the glia<strong>di</strong>n<br />
sequences in the eluted fraction were analyzed by MS/MS<br />
experiments. Figure 3 shows the nanoESI-MS/MS spectrum<br />
of the triply charged ion at m/z 784.37 (MH + ) 2351.11 Da),<br />
correspon<strong>di</strong>ng to the sequence XXQPQPQQQQQQQQQQQQL<br />
(X in<strong>di</strong>cates an unidentified residue) correspon<strong>di</strong>ng to that of<br />
wheat R-glia<strong>di</strong>n (Swiss-Prot, accession number Q9ZP09) with<br />
only the last 13 amino acid residues of peptide.<br />
The two amino acid residues located at the peptide Nteminal,<br />
correspon<strong>di</strong>ng to the “b2” fragment, were not exactly<br />
identified (Figure 3). In particular, since the “b2” fragment is<br />
232.08 Da, the first two amino acid residues could be either<br />
Met-Thr or Thr-Met. For this reason, we preferred to in<strong>di</strong>cate<br />
them as XX.<br />
This result prompts us to develop a new methodology for<br />
sensing toxic sequences for celiac patients. First, we immobilized<br />
GlnBP to a porous silicon wafer and tested the<br />
strength of the covalent bond between the labeled GlnBP and<br />
the porous silicon surface by washing the chip in a demi-water<br />
flux. In Figure 4, a confocal microscopy image of the PSi chip<br />
under a He-Ne laser beam irra<strong>di</strong>ation is shown. We found that<br />
the fluorescence is very high and homogeneous on the whole<br />
surface. We also checked, by Fourier transformed infrared (FT-<br />
IR) spectroscopy, that the PSi surface covered by the protein<br />
shows a good stability toward aqueous oxidation: after the<br />
water treatment, the characteristic peak of Si-O-Si bonds <strong>di</strong>d<br />
not appear in the FT-IR spectrum. In the presence of glia<strong>di</strong>n<br />
PT, GlnBP undergoes a large conformational change in its<br />
global structure to accommodate the ligand inside the bin<strong>di</strong>ng<br />
Journal of Proteome Research • Vol. 5, No. 5, 2006 1243
esearch articles De Stefano et al.<br />
Figure 3. MS/MS spectra of triple ion charged 784.37 (M ) 2351.11 Da).<br />
Figure 4. Confocal microscopy image of the PSi chip after<br />
immobilization of rodhamine-labeled GlnBP.<br />
site. The ligand-bin<strong>di</strong>ng event was detected as a fringe shift in<br />
wavelength, which corresponds to a change in the optical path<br />
nd. Since the thickness d is fixed by the physical <strong>di</strong>mension of<br />
the PSi matrix, the variation is clearly due to changes in the<br />
average refractive index.<br />
The experimental measurement to detect the bin<strong>di</strong>ng of<br />
glia<strong>di</strong>n to GlnBP is a two-step procedure: first, we registered<br />
the optical spectrum of the porous silicon layer after the GlnBP<br />
absorption on the chip surface and, then, after the PT-glia<strong>di</strong>n<br />
solution has been spotted on it. The organic material excess<br />
was removed by further rinses in the buffer solution. In Figure<br />
5, the shifts induced in the fringes of the reflectivity spectrum<br />
are reported. A well-defined red-shift of 2.6 ( 0.1 nm all over<br />
a wide range of wavelengths, due to protein-ligand interaction,<br />
was registered. Control experiments were made by using two<br />
<strong>di</strong>fferent solutions containing nontoxic peptides produced<br />
following a peptic-tryptic <strong>di</strong>gestion of zein, the correspon<strong>di</strong>ng<br />
prolamines from corn, and rice prolamines, that represent safe<br />
1244 Journal of Proteome Research • Vol. 5, No. 5, 2006<br />
Figure 5. Fringes’ shift due to protein-protein interaction.<br />
Continuous curve, porous silicon and GlnBP; dashed curve,<br />
porous silicon, GlnBP, and PT-glia<strong>di</strong>n. Wavelength shift between<br />
the continuous curve and the dashed curve is 2.6 (0.1) nm.<br />
cereals for celiac patients; the fringes’ shift in the reflectivity<br />
spectrum, with respect to the one obtained after the GlnBP<br />
absorption on the chip surface, was undetectable (∆λ < 0.1<br />
nm) in each single measurement.<br />
We also measured the signal response to the protein<br />
concentration after the ligand interaction. Figure 6 shows the<br />
dose-response curve as a function of the PT-glia<strong>di</strong>n concentration.<br />
Since the sensor <strong>di</strong>splays a linear response between<br />
2.0 and 8.0 µM, it is possible to calculate the sensitivity of the<br />
optical interferometer to the concentration of PT-glia<strong>di</strong>n by<br />
estimating the slope of the curve: sGli ) 421 (13) nm/µM. The<br />
sensor response saturates approximately at 35 µM, which<br />
means that about 45% of the spotted proteins have bound the<br />
respective peptide.<br />
The rationale of the proposed strategy is that gluten proteins<br />
(glia<strong>di</strong>ns and glutenins) are characterized by a high content of<br />
glutamine (Gln) residues and, consequently, they represent a
GlnBP Binds Wheat Glia<strong>di</strong>n Toxic Sequences research articles<br />
Figure 6. Dose-response curve. Optical variations as a function<br />
of the PT-glia<strong>di</strong>n concentration.<br />
substrate for the GlnBP. As a consequence, the relevance of<br />
these results for the topic of gluten determination has been<br />
addressed. The detection of residual gluten in food is fundamental<br />
for celiacs patients. Among toxic prolamins, glia<strong>di</strong>n has<br />
been very well characterized. Glia<strong>di</strong>n is a mixture of many<br />
proteins; 10 biochemical analysis revealed the existence of a<br />
relationship among the various constituents, so glia<strong>di</strong>n fractions<br />
have been grouped in three classes named R, γ, and ω<br />
glia<strong>di</strong>ns. 20 To date, several immuno-active glia<strong>di</strong>n peptides in<br />
the celiac population have been identified. 21,22 Some of these<br />
epitopes are encompassed in a 33-mer glia<strong>di</strong>n peptide found<br />
to be resistant to <strong>di</strong>gestion by gastric and pancreatic enzymes. 22<br />
It is noteworthy that most of these T-cell epitopes were<br />
recognized following deamidation catalyzed by tissue transglutaminase<br />
(tTG), in which some specific glutamine residues<br />
were converted to glutamic acid. 23 Moreover, T-cell epitopes<br />
cluster in regions that are rich in proline residues. 24 Taken<br />
together, these observations highlighted the <strong>di</strong>fficulties in<br />
identifying sequences useful for developing immunoassays<br />
addressed toward the toxic portions of glia<strong>di</strong>n. Ideally, a<br />
method to determine glia<strong>di</strong>n should be applicable in a wide<br />
range of food, irrespective of processing, and should be <strong>di</strong>rectly<br />
related with toxicity. Until now, none of the produced methods<br />
were considered to be fully satisfactory. 25,26 On the other hand,<br />
the use of reducing agents, that are present in the proposed<br />
solvent, can improve the extraction of prolamines but affect<br />
their immunochemical quantification. 27 From this point of<br />
view, our method can potentially work also in reducing<br />
con<strong>di</strong>tions, thus, overcoming the problem related to prolamin<br />
extraction. 28<br />
In conclusion, in this work, we have presented useful data<br />
for the development of an optical protein microsensor, based<br />
on porous silicon nanotechnology, for an easy and rapid<br />
detection of glia<strong>di</strong>n. The porous silicon matrix is demonstrated<br />
to be very useful as transducer material in this kind of<br />
biosensor: its highly specific area ensures good sensitivities<br />
in organic molecules sensing, and at the same time, it supplies<br />
a fast and easy readable optical response. Even if the GlnBP is<br />
covalently bonded to the PSi surface, due to a proper functionalization<br />
process, it still selectively recognizes the target<br />
analyte with an overall efficiency of 45%. Genetic manipulation<br />
experiments for improving the protein affinity to PT-glia<strong>di</strong>n are<br />
in progress.<br />
Abbreviations: GlnBP, glutamine-bin<strong>di</strong>ng protein; Gln,<br />
glutamine; PSi, porous silicon; MS/MS, tandem mass spectrometry;<br />
nanoESI, nanospray electrospray ionization.<br />
Acknowledgment. This project was realized in the<br />
frame of the CRdC-ATIBB and CRdc-NTP POR UE-Campania<br />
Mis 3.16 activities and in the frame of the CNR Commessa<br />
“Diagnostica avanzata ed alimentazione”.<br />
References<br />
(1) Lee, T. M.; Hsing, I. Anal. Chim. Acta 2005, 530, 1439153.<br />
(2) Arrand, H. F.; Loni, A.; Arens-Fischer, R.; Kruger, M. G.; Thoenissen,<br />
M.; Lueth, H.; Kershaw, S.; Vorazov, N. N. J. Lumin. 1999,<br />
80, 119-123.<br />
(3) Chamard, V.; Dolino, G.; Muller, F. J. Appl. Phys. 1998, 84, 6659-<br />
6666.<br />
(4) Properties of Porous Silicon; Canham, L., Ed.; IEE Inspec: London,<br />
1997.<br />
(5) Dancil, K.-P. S.; Greiner, D. P.; Sailor, M. J. J. Am. Chem. Soc. 1999,<br />
121, 7925-7930.<br />
(6) Dattelbaum, J. D.; Lakowicz, J. R. Anal. Biochem. 2001, 291, 89-<br />
95.<br />
(7) D’Auria, S.; Scirè, A.; Varriale, A.; Scognamiglio, V.; Staiano, M.;<br />
Ausili, A.; Rossi, M.; Tanfani, F. Proteins 2005; 58, 80-87.<br />
(8) Kuznetsova, I. M.; Stepanenko, O. V.; Turoverov, K. K.; Staiano,<br />
M.; Scognamiglio, V.; Rossi, M.; D’Auria, S. J. Proteome Res. 2005,<br />
4, 417-423.<br />
(9) Maki, M.; Collin, P. Lancet 1997, 349, 1755-1759.<br />
(10) Platt, S. G.; Kasarda, D. D. Biochim. Biophys. Acta 1971, 243, 407-<br />
415.<br />
(11) Bradford, M. M. Anal. Biochem. 1976, 72, 248-254.<br />
(12) De Stefano, L.; Moretti, L.; Ren<strong>di</strong>na, I.; Rossi, A. M.; Tundo, S.<br />
Appl. Opt. 2004, 43, 167-172.<br />
(13) De Stefano, L.; Moretti, L.; Rossi, A. M.; Rocchia, M.; Lamberti,<br />
A.; Longo, O.; Arcari, P.; Ren<strong>di</strong>na, I. IEEE Trans. Nanotech. 2004,<br />
3, 49-54.<br />
(14) Gao, J.; Gao, T.; Sailor, M. J. Appl. Phys. Lett. 2000, 77, 901-903.<br />
(15) Gaillet, M.; Guendouz, M.; Ben Salah, M.; Le Jeune, B.; Le Brun,<br />
G. Thin Solid Films 2004, 455-456, 410-416.<br />
(16) Chamard, V.; Dolino, G.; Muller, F. J. Appl. Phys. 1998, 84, 6659.<br />
(17) DeLouise, L. A.; Miller, B. L. Mater. Res. Soc. Symp. Proc. 2004,<br />
782, A5.3.1.<br />
(18) Hart, B. R.; Letant, S. E.; Kane, S. R.; Ha<strong>di</strong>, M. Z.; Shields, S. J.;<br />
Reynolds, J. G. Chem. Commun. 2003, 322-323.<br />
(19) Yin, H. B.; Brown, T.; Gref, R.; Wilkinson, J. S.; Melvin, T.<br />
Microelectron. Eng. 2004, 73-74, 830-836.<br />
(20) Wieser, H.; Mödl, A.; Seilmeier, W.; Belitz, H. D. Z. Lebens. -Unters.<br />
Forsch. 1987, 185, 371-378.<br />
(21) Vader, L. W.; de Ru, A.; van der Wal, Y.; Kooy, Y. M. C.;<br />
Benckhuijsen, W.; Merin, M. L.; Wouter Drijfhout, J. W.; van<br />
Veelen, P.; Koning, F. J. Exp. Med. 2002, 195, 643-649.<br />
(22) Van de Wal, Y.; Kooy, Y. M.; van Veelen, P. A.; Pena, S. A.; Mearin,<br />
L. M.; Molberg, O.; Lun<strong>di</strong>n, K. E.; Sollid, L. M.; Mutis, T.;<br />
Benckhuijsen, W. E.; Drijfhout, J. W.; Koning, F. Proc. Natl. Acad.<br />
Sci. U.S.A. 1998, 95, 10050-10054.<br />
(23) Shan, L.; Molberg, O. Science 2002, 297, 2275-2279.<br />
(24) Molberg, O.; McAdam, S. N.; Korner, R.; Quarsten, H.; Kristiansen,<br />
C.; Madsen, L.; Fugger, L.; Scott, H.; Noren, O.; Roepstorff, P.;<br />
Lun<strong>di</strong>n, K. E.; Sjostrom, H.; Sollid, L. M. Nat. Med. 1998, 4, 713-<br />
717.<br />
(25) Arentz-Hansen, H.; McAdam, S. N.; Molberg, O.; Fleckenstein,<br />
B.; Lun<strong>di</strong>n, K. E.; Jorgensen, T. J.; Jung, G.; Roepstorff, P.; Sollid,<br />
L. M. Gastroenterology 2002, 123, 803-809.<br />
(26) Skerrit, J. H.; Hill, A. S. J. Agric. Food Chem. 1990, 38, 1771-1778.<br />
(27) Sorell, L.; Lopez, J. A.; Valdes, I.; Alfonso, P.; Camafeita, E.;<br />
Acevedo, B.; Chirdo, F.; Gavilondo, J.; Mendez, E. FEBS Lett. 1998,<br />
439, 46-50.<br />
(28) Margheritis, A. I.; Dona, V. V.; Fossati, C. A.; Chirdo, F. G.<br />
Procee<strong>di</strong>ngs of the 17th Meeting of the Working Group on<br />
Prolamin Analysis and Toxicity, London, U.K., October 3-6, 2002;<br />
Verlag Wissenschaftliche Scripten, Zwickau.<br />
PR0600226<br />
Journal of Proteome Research • Vol. 5, No. 5, 2006 1245
INSTITUTE OF PHYSICS PUBLISHING JOURNAL OF PHYSICS: CONDENSED MATTER<br />
J. Phys.: Condens. Matter 18 (2006) S2019–S2028 doi:10.1088/0953-8984/18/33/S17<br />
Nanostructured silicon-based biosensors for the<br />
selective identification of analytes of social interest<br />
1. Introduction<br />
Sabato D’Auria 1,4 ,Marcella de Champdoré 1 ,Vincenzo Aurilia 1 ,<br />
Antonietta Parracino 1 ,MariaStaiano 1 ,Annalisa Vitale 1 ,MosèRossi 1 ,<br />
Ilaria Rea 2 ,LuciaRotiroti 2 ,Andrea M Rossi 3 ,Stefano Borini 3 ,<br />
Ivo Ren<strong>di</strong>na 2 and Luca De Stefano 2<br />
1 Institute of Protein Biochemistry, CNR, ViaPietro Castellino, 111 80131 Naples, Italy<br />
2 Institute for Microelectronics and Microsystems, CNR—Department of Naples, Via Pietro<br />
Castellino 111, 80131 Naples, Italy<br />
3 Istituto Nazionale <strong>di</strong> Ricerca Metrologica—INRIM, Via Strada delle Cacce 91, 10100 Turin,<br />
Italy<br />
E-mail: s.dauria@ibp.cnr.it<br />
Received 13 January 2006<br />
Published 4 August 2006<br />
Online at stacks.iop.org/JPhysCM/18/S2019<br />
Abstract<br />
Small analytes such as glucose, L-glutamine (Gln), and ammonium nitrate<br />
are detected by means of optical biosensors based on a very common<br />
nanostructured material, porous silicon (PSi). Specific recognition elements,<br />
such as protein receptors and enzymes, were immobilized on hydrogenated<br />
PSi wafers and used as probes in optical sensing systems. The bin<strong>di</strong>ng<br />
events were optically transduced as wavelength shifts of the porous silicon<br />
reflectivity spectrum or were monitored via changes of the fluorescence<br />
emission. The biosensors described in this article suggest a general approach<br />
for the development of new sensing systems for a wide range of analytes of<br />
high social interest.<br />
(Some figures in this article are in colour only in the electronic version)<br />
For many decades, scientists have recognized the power of incorporating biological principles<br />
and molecules into the design of artificial devices. Biosensors, an amalgamation of signal<br />
transducers and biocomponents, play a prominent role in me<strong>di</strong>cine, food, and processing<br />
technologies. Compactness, portability, high specificity, and sensitivity represent some reasons<br />
why biosensors are considered as hol<strong>di</strong>ng promise, with high potential for replacing current<br />
analytical practice [1].<br />
4 Address for correspondence: Institute of Protein Biochemistry, Italian National Research Council, Via Pietro<br />
Castellino, 111-80131 Naples, Italy.<br />
0953-8984/06/332019+10$30.00 © 2006 IOP Publishing Ltd Printed in the UK S2019
S2020 SD’Auriaet al<br />
Fluorescence detection is the dominant analytical approach in me<strong>di</strong>cal testing,<br />
biotechnology, and drug <strong>di</strong>scovery. Starting in the 1980s the first chemically synthesized<br />
fluorescence probes for specific analytes became available. However, the development of<br />
probes for complex analytes present in biological specimens cannot be approached via chemical<br />
synthesis alone [2–4]. In fact, the number of potential ligands specifically recognized by<br />
<strong>di</strong>fferent proteins is very large and ranges from small molecules to macromolecules (inclu<strong>di</strong>ng<br />
proteins themselves). The advantages of using proteins as components of biosensors are many<br />
and include relatively low costs in design and synthesis, the fact that proteins are, at least in<br />
general, soluble in water, and finally, with the progress of molecular genetics, the possibility of<br />
improving/changing some of the propertiesoftheproteins by genetic manipulation [5, 6].<br />
Recently, a lot of experimental work exploiting the properties of porous silicon (PSi) in<br />
chemical and biological sensing has been reported [7, 8]. PSi is an almost ideal material as a<br />
transducer due to its porous structure, with a hydrogen terminated surface, having a specific<br />
area of the order of 200–500 m 2 cm −3 ,soaveryeffective interaction with several adsorbates<br />
is ensured. Moreover, PSi is an available and low cost material, completely compatible<br />
with standard IC processes, and a useful component of so-called smart sensors [9]. The<br />
PSi is fabricated by the electrochemical etching of a silicon (Si) wafer in a hydrofluori<strong>di</strong>c<br />
acid solution. It is well known that the porous silicon ‘as etched’ has an Si–H terminated<br />
surface due to the Si <strong>di</strong>ssolution process [10]. PSi optical sensors are based on changes of<br />
photoluminescence or reflectivity upon exposure to the target analytes [11, 12], which replace<br />
the air in the PSi pores. The optical changes depend on the chemical and physical properties<br />
of the bound analyte. As a consequence, these devices cannot be used for the identification<br />
of analytes in a complex mixture. In order to enhance the sensor selectivity, it is possible to<br />
utilize PSi wafer as substrate for the covalent immobilization of biomolecules, such as proteins,<br />
enzymes, and antibo<strong>di</strong>es, that are able to recognize target analytes in a complex matrix [13–20].<br />
In this paper, we report new fin<strong>di</strong>ngs of relevance to the design of advanced biosensors<br />
obtained by a multi<strong>di</strong>sciplinary group of scientists, for the optical detection of analytes of social<br />
interest such as glucose, ammonium nitrate, and glutamine.<br />
2. Glucose detection<br />
Close control of blood glucose is essential to avoid the long term adverse consequences of<br />
elevated blood glucose, inclu<strong>di</strong>ng neuropathies, blindness, and other consequences [21]. Noninvasive<br />
measurements of blood glucose have been a long-stan<strong>di</strong>ng research target because<br />
they would allow the development of a variety of devices for <strong>di</strong>abetic health care, inclu<strong>di</strong>ng<br />
continuous painless glucose monitoring, control of an insulin pump, and warning systems for<br />
hyperglycaemic and hypoglycaemic con<strong>di</strong>tions. Hypoglycaemia is a frequent occurrence in<br />
<strong>di</strong>abetics, and can result in coma or death. The acute and chronic problems of <strong>di</strong>abetics and<br />
hypoglycaemia can be ameliorated by continuous monitoring of blood glucose. At present,<br />
the only reliable method for measuring blood glucose is via a finger prick and subsequent<br />
glucose measurement, typically using glucose oxidase. This procedure is painful and even<br />
the most compliant in<strong>di</strong>viduals, with good understan<strong>di</strong>ng and motivation for glucose control,<br />
are not willing to prick themselves more than several times per day. These me<strong>di</strong>cal needs<br />
have fostered intensive efforts to develop sensors for glucose [22, 23]. The absence of a<br />
suitable non-invasive glucose measurement has fostered decades of research, little of which<br />
has however resulted in simpler and/or improved glucose monitoring. Included in this effort is<br />
the development of fluorescence probes specific for glucose, typically based on boronic acid<br />
chemistry [24, 25]. An alternative approach to glucose sensing using fluorescence is based<br />
on proteins which bind glucose. Optical detection of glucose appears to have had its origin
Nanostructured biosensors S2021<br />
in the promising stu<strong>di</strong>es of Schultz and co-workers [26, 27], who developed a competitive<br />
glucose assay which does not require substrates and does not consume glucose. This assay used<br />
fluorescence resonance energy transfer (FRET) between a fluorescence donor and an acceptor,<br />
each covalently linked to concanavalin A (ConA) or dextran. In the absence of glucose the<br />
bin<strong>di</strong>ng between ConA and dextran resulted in a high FRET efficiency. The ad<strong>di</strong>tion of glucose<br />
resulted in its competitive bin<strong>di</strong>ng to ConA, <strong>di</strong>splacement of ConA from the labelled dextran,<br />
and a decrease in FRET efficiency. These results generated considerable enthusiasm for<br />
fluorescence sensing of glucose [28]. The glucose–ConA system was also stu<strong>di</strong>ed by other labs.<br />
They identified several problems, namely that the system was only partially reversible upon<br />
ad<strong>di</strong>tion of glucose, and that it became less reversible with time and showed aggregation [29].<br />
For this reason, the use of other glucose-bin<strong>di</strong>ng proteins as sensors was explored by several<br />
research groups [30–35]. If a reliable fluorescence assay for glucose could be developed, then<br />
the robustness of lifetime-based sensing [36]could allow development of a minimally invasive<br />
implantable glucose sensor, or of a sensor which uses extracted interstitial fluid. The lifetime<br />
sensor could measure glucose concentration through the skin [37] usingaredlaser <strong>di</strong>ode or<br />
light emitting <strong>di</strong>ode (LED) device as the light source. These devices are easily powered with<br />
batteries and can be engineered into a portable device. An implantable sensor can be expected<br />
to report on blood glucose because tissue glucose closely tracks blood glucose with a 15 min<br />
time lag [38].<br />
Glucose oxidase (GO) (EC 1.1.3.4) from Aspergillus niger catalyses the conversion of β-<br />
D-glucose and oxygen to D-glucono-1,5-lactone and hydrogen peroxide. It is a flavoprotein,<br />
highly specific for β-D-glucose, and is widely used to estimate glucose concentration in blood<br />
or urine samples through the formation of coloured dyes [39]. Because glucose is consumed,<br />
this enzyme cannot be used as a reversible sensor. We extended the use of GO under con<strong>di</strong>tions<br />
where no reaction occurs. In particular, in order to prevent glucose oxidation, we removed the<br />
FAD cofactor which isstrictly required for the reaction.<br />
The absorbance spectrum of apo-GO shows the characteristic shape of the coenzyme-free<br />
proteins, with an absorbance maximum at 278 nm due to the aromatic amino acid residues.<br />
The absence of absorption at wavelengths above 300 nm in<strong>di</strong>cates that the FAD has been<br />
completely removed. The fluorescence emission spectrum of apo-GO at room temperature<br />
upon excitation at 298 nm <strong>di</strong>splays an emission maximum at 340 nm, which is characteristic of<br />
partially shielded tryptophan residues. The ad<strong>di</strong>tion of 20 mM glucose to the enzyme solution<br />
resulted in a quenching of about 18% of the tryptophanyl fluorescence emission (data not<br />
shown). This result in<strong>di</strong>cates that the apo-GO is still able to bind glucose. The observed<br />
fluorescence quenching may be mainly ascribed to the tryptophanyl residue 426. In fact, as<br />
shown by x-ray analysis and molecular dynamics simulations, the glucose-bin<strong>di</strong>ng site of GO<br />
is formed by Asp 584, Tyr 515, His 559 and His 516. Moreover, Phe 414, Trp 426 and Asn 514<br />
are in locations where they might form ad<strong>di</strong>tional contacts with glucose [40].<br />
The intrinsic fluorescence of proteins is usually not useful for clinical sensing because of<br />
the need for complex or bulky light sources and because of the presence of numerous proteins<br />
in most biological samples. ANS is known to be a polarity-sensitive fluorophore which <strong>di</strong>splays<br />
an increased quantum yield in low polarity environments. Ad<strong>di</strong>tionally, ANS frequently binds<br />
to proteins with an increase in intensity. We examined the effects of GO on the emission<br />
intensity of ANS. Ad<strong>di</strong>tion of apo-GO to an ANS solution resulted in an approximate 30-fold<br />
increase in the ANS fluorescence intensity. Importantly, the intensity of the ANS emission was<br />
sensitive to glucose, decreasing by approximately 25% upon glucose ad<strong>di</strong>tion. The ANS was<br />
not covalently bound to the protein. That ad<strong>di</strong>tion of glucose results in a progressive decrease<br />
in the ANS fluorescence intensity suggests that the ANS is being <strong>di</strong>splaced into a more polar<br />
environment upon glucose bin<strong>di</strong>ng. The decreased ANS intensity occurred with a glucose-
S2022 SD’Auriaet al<br />
Phase angle (deg) or modulation (%)<br />
Deviations<br />
Moduli (%) Phase<br />
100<br />
80<br />
60<br />
40<br />
20<br />
0<br />
2<br />
0<br />
–2<br />
2<br />
0<br />
0 glucose<br />
20 mM glucose<br />
40 mM glucose<br />
60 mM glucose<br />
T = 25°C<br />
Apo - GO - 1,8 - ANS<br />
–2<br />
1 10 100<br />
Frequency (MHz)<br />
Figure 1. Frequency domain intensity decays of 1,8-ANS-glucose oxidase at <strong>di</strong>fferent<br />
concentrations of glucose. Excitation was at 335 nm while the emission was observed through<br />
interference filter 535/50 nm with two Corning 3–71 cut-off filters.<br />
bin<strong>di</strong>ng constant near 10 mM, which is comparable to the KD of the holo-enzyme. The fact<br />
that the bin<strong>di</strong>ng affinity has not changed significantly may imply that the bin<strong>di</strong>ng is still specific<br />
to glucose. Intensity decays of ANS-labelled apo-GO in the presence of glucose were stu<strong>di</strong>ed<br />
by frequency domain fluorometry. Ad<strong>di</strong>tion of glucose shifts the frequency responses to higher<br />
frequencies, which is due to a decreased ANS lifetime (figure 1). The shorter lifetime of ANS<br />
apo-GO in the presence of glucose is again consistent with the suggestion that glucose <strong>di</strong>splaces<br />
the ANS to a more polar environment. The mean lifetime decreases by over 40% upon ad<strong>di</strong>tion<br />
of glucose. These results demonstrate that apo-glucose oxidase, when labelled with suitable<br />
fluorophores, can serve as a protein sensor for glucose.<br />
3. Explosivedetection<br />
Organophosphorous pesticides, nerve agents and nitro-aromatic compounds are very dangerous<br />
substances which can be used against civil or military objectives. Recently, Bourne and<br />
colleagues have demonstrated the mechanism of inhibition, at molecular level, of the<br />
acetylcholinesterase (AChE) with several inhibitor molecules [41]. The peripheral anionic site<br />
on AChE, located at the active centre gorge entry (figure 2), encompasses overlapping bin<strong>di</strong>ng<br />
sites for allosteric activators and inhibitors; yet, the molecular mechanisms coupling this site<br />
to the active centre at the gorge base to modulate catalysis remain unclear. The peripheral site
Nanostructured biosensors S2023<br />
A<br />
B<br />
C<br />
Figure 2. Structures of the DI–mAChE and PI–mAChE complexes. (A) Ribbon <strong>di</strong>agram of the<br />
mAChE <strong>di</strong>mer (cyan, with the four-helix bundle in magenta) bound to DI (orange bonds, blue<br />
nitrogen and red oxygen atoms). The carbohydrate moieties linked to residues Asn 350 and Asn<br />
464 in both subunits are <strong>di</strong>splayed as grey bonds and coloured spheres. The side chains of the<br />
catalytic triad residues, Ser 203, Glu 334, and His 447, are shown as white bonds in the two <strong>di</strong>mer<br />
subunits. The PEG molecule bound at the centre of the four-helix bundle and the carbonate molecule<br />
bound to Ser 203 are shown as grey (green) and light grey (yellow) bonds, respectively. (B) Closeup<br />
stereo view of the DI molecule (coloured as in A) bound to the PAS, with the 2.35 ˚A resolution<br />
omit Fo − Fc electron density map contoured at 3.5σ (cyan) and 7.5σ (blue); the coor<strong>di</strong>nates of<br />
this region were omitted and the protein coor<strong>di</strong>nates were refined by simulated annealing before the<br />
phase calculation. The interacting side chains of mAChE residues His 287 and Leu 289, located<br />
in the loop region connecting helices α3 6,7 and α4 6,7 ,andof residues Trp 286 and Tyr 341 at the<br />
gorge entrance, are <strong>di</strong>splayed as grey (green) bonds; those of mAChE residues Tyr 72 and Glu<br />
292, whose respective mutations as methionine and lysine in BgAChE abolish PI bin<strong>di</strong>ng (see<br />
figure 6), are highlighted in light grey (orange). The chloride ions and solvent molecules are<br />
shown aspink and red spheres, respectively. The catalytic triad residues, Ser 203, Glu 334, and<br />
His 447, and the carbonate molecule (bottom) are shown as white and light grey (orange) bonds,<br />
respectively. Hydrogen bonds between mAChE and the DI molecule are shown as white dotted<br />
lines. (C) Close-up stereo view of the PI molecule (coloured as for DI) bound to the PAS, with the<br />
2.35 ˚A resolution omit Fo − Fc electron density map contoured at 3.5σ (cyan). The PI phenyl and<br />
<strong>di</strong>ethylmethylammonio moieties, which show alternative positions in the structure, are <strong>di</strong>splayed as<br />
red and orange bonds. The mAChE side chains interacting with the PI molecule are <strong>di</strong>splayed as<br />
grey (green) and light grey (orange) bonds as in (B). The mAChE molecular surfaces buried at the<br />
DI–mAChE (B) and PI–mAChE (C) complex interfaces are <strong>di</strong>splayed in transparency.
S2024 SD’Auriaet al<br />
Fluorescence Intensity (au)<br />
300000<br />
250000<br />
200000<br />
150000<br />
100000<br />
50000<br />
0<br />
580<br />
Acetylcholine (2mM)<br />
Acetylcholine (2mM) + Ammonium Nitrate (20mM)<br />
590 600 610 620 630<br />
Wavelength (nm)<br />
640 650<br />
Figure 3. Fluorescence emission spectra of AChE immobilized on Psi, in the absence and in the<br />
presence of ammonium nitrate.<br />
has also been proposed to be involved in heterologous protein associations occurring during<br />
synaptogenesis or upon neurodegeneration [42].<br />
A crystal form of mouse AChE, combined with spectrophotometric analyses of the<br />
crystals, allowed us to obtain unique structures of AChE with a free peripheral site, and<br />
as three complexes with peripheral site inhibitors: the phenylphenanthri<strong>di</strong>nium ligands,<br />
deci<strong>di</strong>um and propi<strong>di</strong>um, and the pyrogallol ligand, gallamine, at 2.20–2.35 ˚A resolution [43].<br />
Comparison with structures of AChE complexes with the peptide fasciculin or with organic<br />
bifunctional inhibitors unveiled new structural determinants contributing to ligand interactions<br />
at the peripheral site, and permitted a detailed topographic delineation of this site [44].<br />
These structures provide templates for designing compounds <strong>di</strong>rected to the enzyme surface<br />
that modulate specific surface interactions controlling catalytic activity and non-catalytic<br />
heterologous protein associations. Since several dangerous substances inhibit AChE activity,<br />
first responders and personnel at risk for exposure to these compounds need sensor systems that<br />
operate in real time and are reliable, cost-effective, compact, portable, and sensitive [45].<br />
AChE was immobilized on a porous silicon chip by applying 50 µl ofa1mgml −1<br />
solution prepared in 100 mM PBE buffer at pH 6.5 to each chip. The chip was incubated<br />
at 4 ◦ C overnight. After the reaction, the excess of reagent was removed and the wafers were<br />
extensively washed at room temperature with PBE buffer. The AChE activity was monitored<br />
by using the Amplex ® Red Acetylcholine/Acetylcholinesterase Assay Kit, by following the<br />
fluorescence variations at 590 nm.<br />
In figure 3, we report the fluorescence emission in the absence and in presence of<br />
ammonium nitrate. As we can see, the presence ofammonium nitrate results in a quenching of<br />
the fluorescence emission of about the 39%, in<strong>di</strong>cating the effectiveness of the optical biosensor<br />
for the detection of the basic compound of common explosives.<br />
4. Glutamine detection<br />
Glutamine (Gln) is an important energy and nitrogen source used in the culturing of eukaryotic<br />
cells [46, 47]. Since the catabolism of glutamine leads to the build-up of ammonia which may
Nanostructured biosensors S2025<br />
Figure 4. Optical microscope image of a PSi chip spotted by a labelled protein after electronic<br />
beam irra<strong>di</strong>ation under laser light excitation. The bright area due to the protein fluorescence is<br />
clearly evident. The inset shows the nanometric resolution of the pattern produced.<br />
become toxic to the growing culture and thus inhibit growth [48, 49], monitoring glutamine<br />
concentrations throughout the growth cycle is an important process. Biosensors available for<br />
glutamine are based on immobilized glutaminase [50, 51]. Complications in the determination<br />
of glutamine using this type of enzymatic biosensor arise due to the high degree of chemical<br />
similarity to glutamate as well as other amino acids. Consequently, ad<strong>di</strong>tional proteins and<br />
electrodes must be added to take into account the interfering signals.<br />
The Escherichia coli periplasmic space contains a <strong>di</strong>verse group of bin<strong>di</strong>ng proteins whose<br />
main function is to present various molecules, such as sugars, amino acids, peptides, and<br />
inorganic ions, for transport into the cell [52]. Glutamine-bin<strong>di</strong>ng protein (GlnBP) from E.<br />
coli is a monomeric protein composed of 224 amino acid residues (26 kDa) responsible for the<br />
first step in the active transport of L-glutamine across the cytoplasmic membrane [53, 54].<br />
Of the naturally occurring amino acids, only glutamine is bound by GlnBP, with a Kd of<br />
3 × 10 −7 M[54]. The protein consists of two similar globular domains linked by two peptide<br />
hinges, and x-ray crystallographic data in<strong>di</strong>cate that the two domains undergo large movements<br />
upon ligand bin<strong>di</strong>ng. The changes in global structure of GlnBP following the Gln bin<strong>di</strong>ng make<br />
it a good can<strong>di</strong>date for serving as the basis for new biosensor design.<br />
We detected the molecular bin<strong>di</strong>ng between the GlnBP from Escherichia coli and Gln by<br />
means of an interferometric microsensor based on the utilization of unmo<strong>di</strong>fied porous silicon<br />
as solid support for the chip preparation. The sensor operates by measurement of the Fabry–<br />
Perot fringes in the wavelength reflection spectrum from the porous silicon layer. The ligand<br />
bin<strong>di</strong>ng was detected as a fringe shift in wavelength, correspon<strong>di</strong>ng to a refractive index change.<br />
We developed a three-step procedure for the Gln detection. In particular, firstly we registered<br />
the optical spectrum of the porous silicon layer as etched; then we repeated the measurement<br />
after the GlnBP absorption on the chip surface, and, finally, a Gln solution was spotted on it.<br />
Figure 3 showsthe shifts induced by each step in the fringes, due to protein–ligand interaction.<br />
Figure 4 shows how effective the linking between the protein and the PSi surface is after<br />
electron beam irra<strong>di</strong>ation: the bright area is emitted as a spot on an irra<strong>di</strong>ated PSi area and<br />
photographed by an optical microscope under laser (Ar + ) excitation after several rinses in<br />
demi-water. The real nanometric resolution of this functionalizing method is clearly visible.<br />
The graft effect is due to the desorption of surface hydrogen atoms which is provoked by the
S2026 SD’Auriaet al<br />
(a)<br />
(b)<br />
∆nd (nm)<br />
160<br />
140<br />
120<br />
100<br />
80<br />
60<br />
Protein<br />
Protein-ligand complex<br />
10 15 20 25 30 35 40<br />
Concentration (µg/L)<br />
Figure 5. (a) Confocal microscopy image of the PSi chip infiltrated with the labelled GlnBP after<br />
<strong>di</strong>alysis in demi-water. (The fluorescence emittedbyrhodamine-labelled GlnBP covalently bounded<br />
to the photochemically mo<strong>di</strong>fied PSi surface is registered by a confocal microscope under laser beam<br />
(He–Ne) exposure.) (b) Dose–response curve as a function of the concentration of the protein and<br />
the Gln–GlnBP complex.<br />
electron beam. The naked surface is thus covered by silicon ra<strong>di</strong>cals which strongly react<br />
with the proteins. In figures 5(a) and (b) the features of an optical biosensor are presented.<br />
In particular, the fluorescence emitted by rhodamine-labelled GlnBP immobilized on the<br />
photochemically mo<strong>di</strong>fied PSi surface is shown in figure 5(a). We also measured the change in<br />
the optical path of the PSi layer due to the protein concentration before and after the interaction<br />
of the ligand. Figure 5(b) shows the dose–response curve as a function of the concentration<br />
of GlnBP and the complex GlnBP–Gln. In this curve, each point represents the average of<br />
three independent measurements and the error bars represent the standard deviations. In the<br />
protein concentration range investigated, the sensor exhibits a linear response (R 2 = 0.99) and<br />
it is possible to determine the sensitivity of the optical interferometer to the organic matter by<br />
estimating the slopes of the two curves: sprot = 21(1) nm µg −1 landscomp = 23(1) nm µg −1 l<br />
for the protein and complex, respectively. As expected, the relative <strong>di</strong>stance between the two<br />
curves is constant, the protein–complex ratio being always 1:1.
Nanostructured biosensors S2027<br />
5. Conclusions<br />
This paper reviews examples of optical biosensors for the detection of analytes of clinical,<br />
industrial, and home security interest. The biosensors described, based on porous silicon<br />
nanotechnology, may represent a model for designing advanced optical arrays (lab-on-a-chip)<br />
for simultaneously determining analytes of social interest.<br />
Acknowledgments<br />
This work was supported by CRdC-ATIBB POR UE-Campania Mis 3.16 activities (SD, MR)<br />
and bythe CNR Commessa Diagnostica avanzata ed alimentazione (SD).<br />
References<br />
[1] Wolfbeis O S 2000 Anal. Chem. 72 81R–89R<br />
[2] Lakowicz J R 1995 Proc. SPIE 2388 598<br />
[3] Spichiger-Keller U E 1998 Chemical Sensors and Biosensors for Me<strong>di</strong>cal and Biological Applications<br />
(New York: Wiley–VCH) p 313<br />
[4] Verkman A S, Sellers M C, Chao A C, Leung T and Ketcham R 1989 Anal. Biochem. 178 355–61<br />
[5] Tsien R Y, Rink T J and Poenie M 1985 Cell Calcium 6 145–57<br />
[6] Romoser V A, Hinkle P M and Persechini A 1997 J. Biol. Chem.272 13270–4<br />
[7] Dancil K-P S, Greiner D P and Sailor M J 1999 J. Am. Chem. Soc. 121 7925–30<br />
[8] Arrand H F, Loni A, Arens-Fischer R, Kruger M G, Thoenissen M, Lueth H, Kershaw S and Vorazov N N 1999<br />
J. Lumin. 80 119–23<br />
[9] Gao J, Gao T and Sailor M J 2000 Appl. Phys. Lett. 77 901–3<br />
[10] De Stefano L, Moretti L, Rossi A M, Rocchia M, Lamberti A, Longo O, Arcari P and Ren<strong>di</strong>na I 2004 IEEE<br />
Trans. Nanotech. 3 49–54<br />
[11] De Stefano L, Moretti L, Ren<strong>di</strong>na I, Rossi A M and Tundo S 2004 Appl. Opt. 43 167–72<br />
[12] Mulloni V and Pavesi L 2000 Appl. Phys. Lett. 76 2523<br />
[13] D’Auria S, Di Cesare N, Staiano M, Gryczynski Z, Rossi M and Lakowicz J R 2001 Anal. Biochem. 303 138–44<br />
[14] D’Auria S and Lakowicz J R 2001 Curr. Opin. Biotechnol. 12 99–104<br />
[15] D’Auria S, Rossi M, Gryczynski I, Gryczynski Z and Lakowicz J R 2000 Anal. Biochem. 283 83–8<br />
[16] D’Auria S, Di Cesare N, Gryczynski Z, Gryczynski I, Rossi M and Lakowicz J R 2000 Biochem. Biophys. Res.<br />
Commun. 274 727–31<br />
[17] D’Auria S, Rossi M, Hermann P and Lakowicz J R 2000 Biophys. Chem. 84 167–76<br />
[18] Staiano M, Bazzicalupo P, Rossi M and D’Auria S 2005 Mol. BioSyst. 1 354–62<br />
[19] Borini S, D’Auria S, Rossi M and Rossi A M 2005 Lab Chip. 5 1048–52<br />
[20] Scognamiglio V, Staiano M, Rossi M and D’Auria S 2004 J. Fluoresc. 14 491–8<br />
[21] The Diabetes Control and Complications Trial Research Group 1993 N. Engl. J. Med. 329 977–86<br />
[22] Roe J N and Smoller R 1998 Crit. Rev. Ther. Drug Carrier Syst. 15 199–241<br />
[23] Wang J, Chatrathi M P, Tian B and Polsky R 2000 Anal. Chem. 72 2514–8<br />
[24] James T D, Limaane P and Shinkai S 1996 Chem. Commun. 281–8<br />
[25] Tong A-J, Yamauchi A, Hayashita T, Zhang Z-Y, Smith B D and Teramae N 2001 Anal. Chem. 73 1530–6<br />
[26] Schultz J, Mansouri S and Goldstein I J 1982 Diabetes Care 5 245–53<br />
[27] Meadows D and Schultz J S 1988 Talanta 35 145–50<br />
[28] Russell R J and Pishko M V 1999 Anal. Chem. 71 3126–32<br />
[29] Lakowicz J R and Maliwal B P 1993 Anal. Chim. Acta 271 155–64<br />
[30] Tolosa L, Gryczynski I, Eichhorn L R, Dattelbaum J D, Castellano F N, Rao G and Lakowicz J R 1999 Anal.<br />
Biochem. 267 114–20<br />
[31] D’Auria S, Herman P, Rossi M and Lakowicz J R 1999 Biochem. Biophys. Res. Commun. 263 550–3<br />
[32] Marvin J S and Hellinga H W 1998 J. Am. Chem. Soc. 120 7–10<br />
[33] Giuliano K A and Taylor D L 1998 TIB Tech. 16 135–40<br />
[34] Wilson G S and Hu Y 2000 Chem. Rev. 100 2693–704<br />
[35] Schalkhammer T, Lobmaier C, Ecker B, Wakolbinger W, Kynclova E, Hawa G and Pittner F 1994 Sensors<br />
Actuators B 18/19 587–91
S2028 SD’Auriaet al<br />
[36] Szmacinski H and Lakowicz J R 1995 Sensors Actuators B 29 15–24<br />
[37] Bambot S B, Rao G, Romauld M, Carter G M and Lakowicz J R 1995 Biosens. Bioelectron. 10 643–52<br />
[38] Velho G, Froguel P, Thevenot D R and Reach G 1988 Diab. Nutr. Metab. 3 227–33<br />
[39] Menstein D J, Pai E F, Schopfer L M and Massey V 1986 Biochemistry 25 6807–16<br />
[40] Hecht H J, Kalisz H M, Hendle J, Schmid R D and Schmburg D 1993 J. Mol. Biol. 229 153–72<br />
[41] Bourne Y, Kolb H C, Ra<strong>di</strong>c Z, Sharpless K B, Taylor P and Marchot P 2004 Proc. Natl Acad. Sci. USA<br />
101 1449–54<br />
[42] Bourne Y, Ra<strong>di</strong>c Z, Kolb H C, Sharpless K B, Taylor P and Marchot P 2005 Chem. Biol. Interact.<br />
157/158 159–65<br />
[43] Bourne Y, Taylor P, Ra<strong>di</strong>c Z and Marchot P 2003 EMBO J. 22 1–12<br />
[44] Tai K, Shen T, Henchman R H, Bourne Y, Marchot P and McCammon J A 2002 J. Am. Chem. Soc. 124 6153–61<br />
[45] White B J, Legako J A and Harmon H J 2003 Biosens. Bioelectron. 18 729–34<br />
[46] Thomas J N 1986 Mammalian CellTechnology ed W G Thilly (Boston, MA: Butterworth) pp 109–30<br />
[47] Zielke H R, Zielke C L and Ozand P T 1984 Fed. Proc. 43 121–5<br />
[48] Ozturk S S and Palsson B O 1990 Biotechnol. Prog. 6 121–8<br />
[49] Oh G S, Izuishi T, Inoue T W and Yoshida T J 1996 J. Ferment. Bioeng. 81 329–36<br />
[50] Sorochinskii V V and Kurganov B I 1997 Appl. Biochem. Microbiol. 33 515–29<br />
[51] Madaras M B, Spokane R B, Johnson J M and Wookward J R 1997 Anal. Chem. 69 3674–8<br />
[52] Higgins C F 2001 Res. Microbiol. 152 205–10<br />
[53] Nohno T, Saito T and Hong J-S 1986 Mol. Gen. Genet. 205 260–9<br />
[54] Weiner J H and Heppel L A 1971 J. Biol. Chem.246 6933–41
Optical evidence of protein structural changes in porous<br />
silicon devices<br />
The protein-ligand molecular interactions imply strong geometrical and structural<br />
rearrangements of the biological complex which are normally detected by high sensitivity<br />
optical techniques such as time-resolved fluorescence microscopy. In this section are<br />
reported the measurements, made by optical spectroscopic reflectometry in the visiblenear<br />
infrared region, about the interaction between a sugars bin<strong>di</strong>ng protein (TMBP),<br />
covalently bound on the surface of a porous silicon, and the glucose, at <strong>di</strong>fferent<br />
concentrations and temperatures.<br />
I’m grateful to Dr. S. D’Auria from Institute for Protein Biochemistry, National Council of<br />
Research, Italy for supplying the TMBP protein and for helpful <strong>di</strong>scussion about the<br />
experiments.
Introduction: proteins from extremophiles<br />
The proteins purified from thermophilic organisms are characterized by high stability in<br />
harsh environment such as high temperature, high ionic strength, extreme pH values, even<br />
in presence of elevated concentration of detergents and chaotropic agents so that they<br />
can be successfully used as very effective bio-probes in sensors design[1].<br />
In particular, it has been recently demonstrated that the D-trehalose/D-maltose-bin<strong>di</strong>ng<br />
protein (TMBP), which is part of the sugars uptake system in the hyperthermophilic<br />
organism archaeon T. litoralis, is also able to bind glucose molecules. This result suggests<br />
the possibility of using TMBP as a stable probe within a biological recognition system for<br />
glucose monitoring [1]. TMBP is a monomeric 48 kDa macromolecule constituted by twodomains<br />
and containing twelve tryptophan residues. The presence of these fluorescent<br />
residues has been used to study the interaction between TMBP and glucose at <strong>di</strong>fferent<br />
temperatures by means of time-resolved fluorescence spectroscopy by Herman et al. [1]:<br />
they found that the highest affinity between the protein and this ligand was at a<br />
temperature around 60 °C, with a <strong>di</strong>ssociation constant (Kd) of about 40 μM. At room<br />
temperature TMBP still binds substrates while the activity of the TMBP based transport<br />
system becomes negligible [3,4], and this behaviour can be ascribed to an increased<br />
rigi<strong>di</strong>ty of TMBP structure at room temperature [5]. As other sugar-bin<strong>di</strong>ng proteins, TMBP<br />
also consists of two globular lobes linked by a hinge region made of a few polypeptide<br />
chains. The deep cleft formed between the two lobes, which present similarly tertiary<br />
structure, contains the ligand-bin<strong>di</strong>ng site [6].<br />
In a recent published paper [7], we have found that, at room temperature, the reflectivity<br />
spectra of a PSi <strong>di</strong>stributed Bragg reflector (DBR), properly functionalized with TMBP,<br />
undergo a red shift on exposure to glucose solutions. In particular, we observed a red shift<br />
of about 1.2 nm after the interaction with a 150 μM glucose solution. The estimated<br />
sensitivity of the monitoring method was 0.03 nm/μM.<br />
On the other hand, some preliminary experiments conducted at 60 °C [8] have shown that<br />
the reflectivity spectra of <strong>di</strong>fferent PSi-based structures, functionalized with TMBP,<br />
undergo blue-shifts as a consequence of the interaction with glucose.<br />
Than these apparently conflicting results was explained by combining our investigation on<br />
two PSi devices, both TMBP functionalized: an optical microcavity (PSMC), characterized<br />
by optical spectroscopic reflectometry, and a thin PSi monolayer, as a model system for<br />
the ellipsometric characterization in order to correlate the observed blue shifts of the<br />
PSMC spectra to the typical conformational changes of TMBP at 60 °C.<br />
Materials and methods<br />
D-Glucose and all the other chemicals used in this experiment were from Sigma. All<br />
commercial samples were of the best available quality. The TMBP was purified and<br />
supplied by dr. S. D’Auria of the Institute for Protein Biochemistry, National Council of<br />
Research, Napoli, Italy. Proteins were expressed, purified and quantified by his laboratory<br />
as described in reference 5, and references therein.<br />
In this experiment were designed and fabricated two PSi-based structures: a monolayer,<br />
that is a porous layer with fixed porosity, and a microcavity, which is constituted by a lowporosity<br />
(high refractive index) layer between two Bragg reflectors, each one obtained by<br />
alternating low and high porosity layers for 7 times. Both the samples were produced by<br />
electrochemical etching of a very highly doped n<br />
55<br />
++ -silicon wafers (Siltronix Inc., USA),<br />
oriented, 0.001 Ω cm resistivity, 400 μm thick, in a HF-based solution. The silicon<br />
was etched using a 50 wt.% HF/ethanol solution with halogen lamp illumination and at<br />
room temperature. Before ano<strong>di</strong>zation the substrate was placed in HF solution to remove<br />
the native oxide. The monolayer, characterized by variable angle spectroscopic
ellipsometry (Horiba-Jobin-Yvon, mod. UV-VISEL) in the wavelength range between 400<br />
and 1700 nm, was 503±1 nm thick with a porosity of 77.5±0.3 %. The porous silicon<br />
microcavity (PSMC) was electrochemically etched applying a current density of 400<br />
mA/cm 2 for 0.3 s to obtain the low refractive index layers, with a porosity of 59±1 %, while<br />
one of 50 mA/cm 2 was applied for 0.5 s for the high index layers, with a porosity of 50±1<br />
%. This optical structure has a characteristic resonance peak at 1017 nm, centered in a<br />
105 nm wide stop band.<br />
The covalent bind of the TMBP on the porous silicon surface is based on a three steps<br />
functionalization process constituted by a chemical passivation of the PSi surface after<br />
oxidation, as it is shown in Figure 1. The PSi optical structures have been thermally<br />
oxi<strong>di</strong>zed in O2 atmosphere, at 900 °C for 15 min; then, we have treated the surfaces with a<br />
proper chemical linker, the aminopropyltriethoxysilane (APTES). Samples have been<br />
rinsed by <strong>di</strong>pping in a 5 % solution of APTES and a hydroalcoholic mixture of water and<br />
methanol (1:1) for 20 min at room temperature. The chips were then washed by deionized<br />
water, and methanol, and dried in N2 stream. The silanized devices were then baked at<br />
100°C for 10 min. To create a surface able to link the amino group of the proteins, we have<br />
immersed the chips in a 2.5% glutaraldehyde (GA) solution in 20 mM HEPES buffer (pH<br />
7.4) for 30 min, and then rinsed it in deionized water and finally dried in N2 stream. The<br />
glutaraldehyde reacts with the amino groups on the silanized surface and coats the<br />
internal surface of the pores.<br />
O 2 Si<br />
Si H<br />
Si H<br />
Si H<br />
O 2 Si<br />
O<br />
OH<br />
O2 5% APTES<br />
SiO2 OH<br />
15' @900°C 20' R.T<br />
OH<br />
O<br />
O<br />
O<br />
Si<br />
O<br />
O<br />
O<br />
Si<br />
O<br />
C<br />
N<br />
NH 2<br />
2.5% GA<br />
30' R.T<br />
C N TMBP<br />
Figure 1. Porous silicon functionalization scheme: from the as-etched material up to the organic-inorganic<br />
chip.<br />
The experimental set-up (Figure 2) used to measure the reflectivity spectra is constituted<br />
by a white light, as source, <strong>di</strong>rected on the porous silicon chip through a Y fiber. The same<br />
fiber was used to guide the output signal to an optical spectrum analyzer (OSA, Ando,<br />
Japan). The spectrum was measured over the range 600-1400 nm with a resolution of 0.2<br />
nm. A hot plate, driven by a temperature controller, has been used to perform<br />
measurements at 60±1 °C, verified by a thermocouple placed <strong>di</strong>rectly on the chip.<br />
56
Figure 2. Experimental setup used to measure the optical reflectivity spectra of porous silicon microcavity.<br />
The functionalized PSi samples work as active substrates for the TMBP protein: on each<br />
PSi optical structures were spotted 20 μl of a 173 μM TMBP solution in 2mM phosphate<br />
buffer solution (pH 7.3) and incubated the system at 4 °C over night. After incubation and<br />
three washing steps of five minutes each, used to remove the excess of biological matter<br />
non covalently linked to the PSi surface, the presence of the protein in the spongy<br />
structure was optically verified as a red shift of about 60 nm in the reflectivity spectrum in<br />
the case of PSMC.<br />
Experimental results and <strong>di</strong>scussion<br />
When recorded at the temperature of 25 °C, the PSMC characteristic resonance peak<br />
undergoes only very small red shifts, of the order of 1 nm, on exposure to glucose<br />
solutions with increasing concentrations, up to 200 μM, accor<strong>di</strong>ng to the low affinity of<br />
TMBP with glucose at room temperature [7]: at this temperature few proteins can rearrange<br />
their conformational structure and efficiently bind the ligand. The results of this<br />
biomolecular interaction change dramatically if the same measurements are performed at<br />
60 °C: in Figure 3 is reported the peak shift of the functionalized microcavity on exposure<br />
to a 200 μM glucose concentration. The optical resonance undergoes a 4.5 nm shift<br />
toward lower wavelengths.<br />
Reflectivity (a.u.)<br />
1.0<br />
0.8<br />
0.6<br />
0.4<br />
0.2<br />
Unperturbed<br />
After exposure to a 200 μM glucose solution<br />
960 980 1000 1020 1040 1060 1080<br />
Δλ<br />
Wavelength (nm)<br />
Figure 3. PSMC resonance blue shift on exposure to a glucose solution of 200 μM concentration.<br />
The dose-response curve is shown in Figure 4: the curve is linear in the range of glucose<br />
concentrations between 0 and 200 μM, with an estimated sensitivity to solute<br />
57
concentration variations of 0.022 (4) nm/μM. For higher glucose concentrations, the blue<br />
shift <strong>di</strong>d not increased any more, which means that all the active proteins have bound their<br />
ligands.<br />
Δλ (nm)<br />
1<br />
0<br />
-1<br />
-2<br />
-3<br />
-4<br />
-5<br />
0 50 100 150 200 250 300<br />
Glucose concentration (μM)<br />
Figure 4: Dose-response curve for functionalized PSMC exposed to <strong>di</strong>fferent concentrations of glucose.<br />
The large blue-shift recorded corresponds to a decrease of the optical path that the light<br />
exploits when reflected by the porous structure. Since the optical path is defined as the<br />
thickness times the average refractive index, and there is not any variation of the<br />
multilayer thickness, the blue-shift can be attributed to a decrease in the value of the<br />
average refractive index. The chemical passivated surface of the PSi is highly stable<br />
respect to the biological chemical buffer solutions, so that there can not be any further<br />
oxidation nor any corrosion of the silicon matrix: the apparent increase of the porosity is<br />
due to the proteins volume reduction after the interaction with the glucose. Most of the keylock<br />
recognition mechanisms, which are at the basis of the biological interactions such as<br />
antibody-antigen or protein-ligand, require large rearrangement of one bio-molecule to<br />
host the other. The bin<strong>di</strong>ng of glucose to TMBP is strongly related to a movement of the<br />
protein lobes as well as to profound conformational changes in the hinge region. The lobes<br />
envelope on to the bin<strong>di</strong>ng site and cause a net reduction of protein’s volume, as<br />
confirmed by fluorescence measurements [1]. In order to confirm these results, a very<br />
sensitive optical technique was used, the variable angle spectroscopic ellipsometry<br />
(VASE), to quantify the porosity changes of a thin PSi layer after TMBP functionalization<br />
and on exposure to increasing concentration of glucose solutions. In Table 1 the porosities<br />
estimated by ellipsometry after each functionalization step of the monolayer, are reported.<br />
Porosity (%)<br />
As etched 77.5 ± 0.3<br />
After oxidation 49.1 ± 0.3<br />
After APTES treatment 41.7 ± 0.5<br />
After GA treatment 35.5 ± 0.4<br />
After TMBP incubation 31.9 ± 0.5<br />
Table 1: Porosity decrease of the PSi monolayer due to the functionalization steps.<br />
58
Chemical and biological functionalization steps considerably reduce the average porosity<br />
of the PSi layer. Ellipsometric spectra have been also acquired after each ad<strong>di</strong>tion of<br />
glucose solutions with increasing concentrations. In Figure 5 the changes in porosity of the<br />
PSi layer, ΔP, as estimated from the ellipsometric data, are shown as function of glucose<br />
concentration. On exposure to ligand buffer solutions, the VASE characterization gives a<br />
<strong>di</strong>rect measurement of the variation of each layer component. Using these experimental<br />
data, we have estimated, after the ad<strong>di</strong>tion of 250 μM of glucose, an increase of more than<br />
the 3.5 % in the monolayer porosity.<br />
ΔP(%)<br />
4.5<br />
4.0<br />
3.5<br />
3.0<br />
2.5<br />
2.0<br />
1.5<br />
1.0<br />
0.5<br />
0.0<br />
0 50 100 150 200 250<br />
Glucose concentration (μM)<br />
Figure 5: The porosity increase of a functionalized PSi monolayer as a function of glucose concentrations.<br />
On exposure to a concentration of 250 μM of glucose, an enhancement of more than 3.5 % in porosity is<br />
obtained.<br />
On the other hand, it was also numerically calculate which is the effect of such a porosity<br />
increment on the reflectivity spectrum of the PSi microcavity. The reflectivity spectrum of a<br />
microcavity with the same porosities and thicknesses of the one realized was simulated.<br />
The organic phase of the system has been modeled by supposing a Gaussian <strong>di</strong>stribution<br />
for both the TMBP molecules presence throughout the upper layers of the microcavity and<br />
also for the correspondent increment of porosity, ΔP, due to the glucose interaction [9]. In<br />
particular, the ΔP <strong>di</strong>stribution was peaked at the top of the structure (with a maximum<br />
value of 3.5 %) and was characterized by a HWHM of about 3 μm, which equals the halfwidth<br />
of the first Bragg reflector in the microcavity structure. The reflectivity spectra have<br />
been simulated by standard transfer matrix method [10] and results are reported in Figure<br />
6.<br />
Reflectivity (a.u.)<br />
0.8<br />
0.6<br />
0.4<br />
0.2<br />
0.0<br />
950 975 1000 1025 1050 1075 1100<br />
Wavelength (nm)<br />
(a)<br />
(b)<br />
59
Figure 6: (a) Simulated reflectivity spectrum of the PSi microcavity used in the experiment. (b) Simulated<br />
reflectivity spectrum of the same structure after a non-homogeneous increase of porosity.<br />
The result of such simulations is a blue shift of the resonant peak of 6.0±1.0 nm, which is<br />
consistent with the experimental one of 4.6±0.5 nm, observed after the ad<strong>di</strong>tion of a<br />
glucose solution, 250 μM.<br />
Even if highly specific, the protein-ligand interaction is in general reversible, so that<br />
proteins release completely the target molecules, especially when in buffer solutions. But<br />
when bound on the surface of a nanostructured material, proteins could be limited in their<br />
activity: after an overnight <strong>di</strong>alysis in deionized water, it was found a 2 nm red-shift of the<br />
resonant peak, which means that the ligand has been released only by some proteins.<br />
Anyway, when the measurements were repeated, the same blue-shift always about 5 nm<br />
with a 250 μM concentration of glucose was obtained, so that can be concluded that the<br />
effect was reproducible, at least for five replica made in the same experimental con<strong>di</strong>tions,<br />
but was not perfectly reversible.<br />
References<br />
1. P. Herman, I. Barvik, M. Staiano, A. Vitale, J. Vecer, M. Rossi and S. D’Auria, Biochimica et<br />
Biophysica Acta 2007, 1774, 540-544.<br />
2. J. Diez, K. Diederichs, G. Greller, R. Horlacher, W. Boos, W. Welte, J. Mol. Biol. 2001;305:905-915.<br />
3. K. B. Xavier, L. O. Martins, R. Peist, M. Kossmann, W. Boos, H. Santos, J. Bacteriol. 1996, 178,<br />
4773-4777.<br />
4. R. Horlacher, K. B. Xavier, H. Santos, J. Di Ruggiero, M. Kossmann, W. Boos, J. Bacteriol. 1998,<br />
180, 680-689.<br />
5. P. Herman, M. Staiano, A. Marabutti, A. Variale, A. Scire, F. Tanfani, J. Vecer, M. Rossi, S. D’Auria,<br />
Proteins 2006, 63, 754-767.<br />
6. Y. J. Sun, J. Rose, B. C. Wang, C. D. Hsiao, J. Mol. Biol. 1998, 278, 219-222.<br />
7. L. De Stefano, A. Vitale, I. Rea, M. Staiano, L. Rotiroti, T. Labella, I. Ren<strong>di</strong>na, V. Aurilia, M. Rossi, S.<br />
D’Auria, Extremophiles, DOI 10.1007/s00792-006-0058-6.<br />
8. L. De Stefano, L. Rotiroti, E. De Tommasi, A. Vitale, M. Rossi, I. Ren<strong>di</strong>na and S. D’Auria,<br />
Procee<strong>di</strong>ngs of the SPIE 2007 Volume 6585, pp. 658517 .<br />
9. L. De Stefano and S. D’Auria, Journal of Physics Condensed Matter 2007, 19, 395009 (7pp).<br />
10. M. A. Muriel and A. Carballar, IEEE P. Technol. Lett. 9, 955 (1997).<br />
60
Oligonucleotides <strong>di</strong>rect synthesis on porous silicon chip<br />
In this section is report the design, fabrication and characterization of a oligonucleotide<br />
(ON) biochip based on the PSi nanotechnology for <strong>di</strong>rect DNA solid phase synthesis.<br />
In collaboration with Prof. G. Piccialli of University of Naples “Federico II”, Italy
Oligonucleotides <strong>di</strong>rect synthesis on porous silicon chip<br />
Luca De Stefano 1 , Edoardo De Tommasi 1 , Ilaria Rea 1 , Lucia Rotiroti 1 , Luca Giangrande 1 , Giorgia<br />
Oliviero 2 , Nicola Borbone 2 , Aldo Galeone 2 and Gennaro Piccialli 2 *<br />
1<br />
Istituto per la Microelettronica e Microsistemi - Unità <strong>di</strong> Napoli, Consiglio Nazionale delle Ricerche, via P.<br />
Castellino 111, I-80131 Napoli, Italy and 2 Dipartimento <strong>di</strong> Chimica delle Sostanze Naturali, <strong>Università</strong> <strong>degli</strong><br />
<strong>Stu<strong>di</strong></strong> <strong>di</strong> Napoli “Federico II”, Via D. Montesano 49, I-80131 Napoli, Italy<br />
ABSTRACT<br />
A solid phase oligonucleotide (ON) synthesis on<br />
porous silicon (PSi) chip is presented. The prepared Si-<br />
OH surface were analyzed by FT-IR and the OH<br />
functions were quantified by reaction with 3’phosphorami<strong>di</strong>te<br />
nucleotide buil<strong>di</strong>ng block. Short ONs<br />
were synthesized on the chip surface and the coupling<br />
yields evaluated.<br />
INTRODUCTION<br />
The interest in DNA-based <strong>di</strong>agnostic tests has recently<br />
increased since they can be used in a lot of important<br />
applications such as gene analysis, fast detection of<br />
biological warfare agents, and also in forensic cases.<br />
Numerous DNA detection systems, both optical and<br />
electrochemical, based on the hybri<strong>di</strong>zation between a<br />
DNA target and its complementary probe fixed on a solid<br />
support have been described. 1 DNA biochips, due to<br />
integration and miniaturization of all components, allow<br />
continuous, fast and sensitive detection of DNA<br />
interactions. The fabrication of a DNA biochip requires the<br />
immobilization of the DNA probe on the support surface, 1,2<br />
or, as an alternative, the synthesis of the ON strand <strong>di</strong>rectly<br />
on the device surface. 3-5<br />
On the other hand, Porous silicon (PSi) structures are<br />
quite ideal support material for chemical and biological<br />
sensing, mainly due to their high specific area, up to 600<br />
m 2 cm -3 , low cost and compatibility with standard integrated<br />
circuit processes. PSi is fabricated by the electrochemical<br />
etching of a doped silicon wafer in a hydrofluori<strong>di</strong>c<br />
Glass<br />
Silicon<br />
Stainless steel capillary tubes<br />
Resin glue<br />
Liquid<br />
input/output<br />
channels<br />
Reaction chamber Porous silicon transducer<br />
Fig. 1. Schematic and real view of the μ-chamber for DNA<br />
biochip applications.<br />
aqueous solution. Moreover, its surface can be properly<br />
functionalised to covalently link the DNA probe.<br />
In this work, we report the design, fabrication and<br />
characterization of a ON biochip based on the PSi<br />
nanotechnology for <strong>di</strong>rect DNA solid phase synthesis.<br />
RESULTS AND DISCUSSION<br />
Porous silicon formation and chip assembly. We have<br />
designed and fabricated a DNA biochip, just integrating<br />
PSi, as the transducer material, and a glass slide, which<br />
ensures sealing and the interconnections for fluids inlet and<br />
outlet. The silicon wafer used in this work is a p + type,<br />
crystal orientation, with a resistivity of 8-12<br />
mΩcm. 6 The glass is a Borofloat 33 type, 1 mm thick. The<br />
reaction microchamber has been realised by a two-step<br />
electrochemical etching. The first step is a high current<br />
density (800 mA/cm 2 for about 300 s) electrochemical etch<br />
in hydrofluoric acid and ethanol (HF/EtOH 1:1,v/v)<br />
solution which creates a μ-well (100 μm deep and a volume<br />
of 10 μL) in the bulk silicon. The second step is a<br />
consecutive electrochemical etch which is used to fabricate<br />
the PSi layer at the bottom of the μ-chamber. The process<br />
parameters for this last step were: current density of 400<br />
mA/cm 2 and etching time of 3.2 s. After the mechanical<br />
drilling of the flow channels, the glass slide has been<br />
cleaned and activated for the ano<strong>di</strong>c bon<strong>di</strong>ng process<br />
following standard cleaning procedures. The silicon chip<br />
has also been carefully rinsed in deionized water for several<br />
minutes. Silicon etched wafer and glass top prefabricated<br />
components were ano<strong>di</strong>cally bonded together with mutual<br />
alignment: a successful bon<strong>di</strong>ng, in terms of mechanical<br />
strength and bond quality, has been obtained at a<br />
temperature of 200 o C, voltage of 2.5 kV, with a process<br />
time of 1.5 minutes. The cross section of the chip structure<br />
and the general-view of the real device are shown in Figure<br />
1.<br />
Functionalization of porous silicon layer and ON<br />
syntheses. The successive treatment of porous silicon layer<br />
with a solution of conc. H2SO4/H2O2 (8:2, v/v, 15 min R.T.)<br />
assure the formation of Si-OH groups on the solid matrix.<br />
The chip was, then, rinsed in deionized water for several<br />
minutes and dried under reduced pressure. The FT-IR<br />
spectrum (Figure 2) of the chip surface showed the<br />
appearance of the characteristic broad bands around 1050
cm -1 (Si-O-Si) and 1065, 880 cm -1 attributable to Si-OH<br />
bonds. To determine the surface coverage of Si-OH groups,<br />
we reacted the chip 1 with tetrazole activated 5’-DMT-3’phosporami<strong>di</strong>te-timi<strong>di</strong>ne<br />
nucleotide (2 Scheme 1) which<br />
assure the presence of a very reactive phosphorous(III) and<br />
T<br />
Si-OH + 2<br />
O<br />
Si-O P O ODMT<br />
i<br />
O<br />
Si-O P O P P OH<br />
1<br />
OCE<br />
ii<br />
OH<br />
2<br />
4<br />
5<br />
T T T<br />
Si-OH + 3<br />
O<br />
Si-O P O<br />
O O<br />
i<br />
S<br />
ODMT<br />
O<br />
Si-O P O R O<br />
O<br />
P O P P OH<br />
1<br />
OCE<br />
6<br />
OCE 7 OCE<br />
ii<br />
T<br />
(pi) 2N<br />
P O ODMT<br />
CEO<br />
O O<br />
(pi) 2N S<br />
P O ODMT<br />
CEO<br />
T<br />
P P<br />
T<br />
P<br />
T<br />
OH<br />
2 3<br />
8<br />
Scheme 1. Solid phase synthesis on PSi-OH chip 1; P:<br />
phosphotriester function; R: -(CH 2) 2SO 2(CH 2) 2; P:<br />
phospho<strong>di</strong>ester function; CE: 2-cyanoethyl; i: standard ON chain<br />
elongation; ii; treatment with methylamine.<br />
a 5’-<strong>di</strong>methoxytrytil group (DMT).<br />
The reaction cycle inclu<strong>di</strong>ng the standard phosphorous<br />
oxidation and the capping step, afforded the PSi-3’thymi<strong>di</strong>ne<br />
4 which was quantified by UV/VIS spectroscopy<br />
monitoring 5’-<strong>di</strong>methoxytrytil cation released by DCA<br />
treatment (ε =71700, λ = 498 nm). The Si-OH<br />
functionalization resulted to be in the range of 35-40<br />
nmol/chip and no significative <strong>di</strong>fferences were observed<br />
varying the concentration of the nucleotide/activator<br />
solution (30-60 mg/mL) or prolonging the coupling<br />
reaction time (30-180-min). Alternatively, PSi chip 1 was<br />
reacted with the phosphorami<strong>di</strong>te 3 thus obtaining the PSi<br />
support 6 (30-40 nmol/chip). The linker 3 ensures a more<br />
long and flexible arm for the reaction of terminal alcoholic<br />
OH groups and allows the release of the synthesized ON<br />
chains by a mild basic treatment, as well. On the supports 4<br />
and 6 three coupling cycles were performed with<br />
phosphorami<strong>di</strong>te 2 thus obtaining PSi supports 5 and 7<br />
respectively. The DMT measurements after each cycle<br />
in<strong>di</strong>cated a coupling efficiency of 87-90 % for support 4<br />
and 92-95% for support 6. These fin<strong>di</strong>ngs are in agreement<br />
with previous reported ON synthetic procedures on PSi<br />
chips which demonstrated the important role of the spacer<br />
linker attached on porous silicon surface. 3-5 The detachment<br />
of the synthesized short thymi<strong>di</strong>ne oligomer from support 7<br />
was obtained by treatment with anhydrous methylamine<br />
(20 min at room temperature). The combined amine<br />
solution and washings (methanol and then water), dried<br />
under reduced pressure, were analyzed by ESI-MS<br />
spectroscopy. The mass data confirmed the presence of<br />
expected 3’-phosphate-T3-5’-OH product (m/z 931.2 MH + ).<br />
The treatment of 5 with methylamine <strong>di</strong>d not release<br />
significative amount of nucleotide material from the chip,<br />
also confirmed by DMT deprotection and measurement<br />
T<br />
T<br />
T<br />
(upper)<br />
Fig. 2. FT-IR profiles of the PSi-surfaces 1 (bottom) and 5<br />
(upper).<br />
performed after the amine treatment. FT-IR spectrum<br />
(Figure 2) of the PSi-ON chip showed the appearance of<br />
the characteristic band around 1100 cm -1 and 860 cm -1<br />
attributable, respectively, at P=O and 5’-CH2-OH.<br />
CONCLUSION<br />
In this preliminary experiments of ON synthesis we have<br />
tested the <strong>di</strong>rect reactivity of the freshly prepared PSi OH<br />
functionalized chips with standard phosphorami<strong>di</strong>te<br />
buil<strong>di</strong>ng block and confirmed the compatibility of the PSi-<br />
3’-bonded thymi<strong>di</strong>ne with the chemical treatments required<br />
by standard ON solid phase synthesis. The synthesis of<br />
support 4 allowed as to estimate the amount of accessible<br />
OH groups on the silicon surface and to obtain a<br />
reproducible nucleotide functionalized chip. The data were<br />
also confirmed by the synthesis of the PSi-support 6. We<br />
hypothesize that, after coupling yield optimization, the<br />
support types 4 and 6 could be exploited both to obtain<br />
deprotected ON bonded to a PSi chip and to achieve a<br />
standard ON solid phase synthesis on PSi matrix.<br />
REFERENCES<br />
1. Sassolas, A., Leca-Bouvier, B.D., Blum, L.J. (2008)<br />
Chem. Rev., 108, 109-139.<br />
2. Di Francia, G., La Ferrara, V., Manzo, S., Chiavarini,<br />
S. (2005) Biosensors and Bioelectronics, 21, 661-665.<br />
3. Pike, A.R., Lie, L.H., Eagling, R.A., Ryder, L.C.,<br />
Patole, S.N., Connolly, B.A., Horrocks, B.R., Houlton,<br />
A. (2002) Angew. Chem. Int. Ed., 41, 615-617.<br />
4. Lie, L.H., Patole, S.N., Pike, A.R., Ryder, L.C.,<br />
Connolly, B.A., Ward, A.D., Tuite, E.M., Houlton, A.,<br />
Horrocks, B.R. (2004) Faraday Discuss., 125, 235-249.<br />
5. Bessueille, F., Dugas, V., Vikulov, V., Cloarec, J.P.,<br />
Souteyrand, E., Martin, J.R. (2005) Biosensors and<br />
Bioelectronics, 21, 908-916.<br />
6. De Stefano, L., Arcari, P., Lamberti, A., Sanges, C.,<br />
Rotiroti, L., Rea, I., Ren<strong>di</strong>na, I. (2007) Sensors, 7, 214-<br />
221.<br />
*Correspon<strong>di</strong>ng Author. E-mail: picciall@unina.it
Porous silicon-based optical biosensors and biochips<br />
This section is aimed at the demonstration that a PSi surface functionalized, allow the<br />
attachment of linkers through stable covalent Si–C bonds, and also demonstrated to be<br />
compatible to the high temperature ano<strong>di</strong>c bon<strong>di</strong>ng processes used for the realization of<br />
silicon-glass biochips.
Abstract<br />
Physica E 38 (2007) 188–192<br />
Porous silicon-based optical biosensors and biochips<br />
Ivo Ren<strong>di</strong>na , Ilaria Rea, Lucia Rotiroti, Luca De Stefano<br />
Institute for Microelectronics and Microsystems, National Council of Research, Via P. Castellino 111, 80131 Napoli, Italy<br />
Available online 31 December 2006<br />
Porous silicon multilayered microstructures have unique optical and morphological properties that can be exploited in chemical and<br />
biological sensing. The large specific surface of nanostructured porous silicon can be chemically mo<strong>di</strong>fied to link <strong>di</strong>fferent molecular<br />
probes (DNA strands, enzymes, proteins and so on), which recognize the target analytes, in order to enhance the selectivity and<br />
specificity of the sensor device. We designed fabricated and characterized several photonic porous silicon-based structures, which were<br />
used in sensing some specific molecular interactions. The next step is the integration of the porous silicon-based optical transducer in<br />
biochip devices: at this aim, we have tested an innovative ano<strong>di</strong>c bon<strong>di</strong>ng process between porous silicon and glass, and its compatibility<br />
with the biological probes.<br />
r 2007 Elsevier B.V. All rights reserved.<br />
PACS: 87.80. y; 07.07.Df; 42.79.Pw<br />
Keywords: Optical biosensors; Porous silicon; Biochip<br />
1. Introduction<br />
Recently, lot of experimental work, concerning the<br />
worth noting properties of nanostructured porous silicon<br />
(PSi) in chemical and biological sensing has been reported,<br />
showing that, due to its morphological and physical<br />
properties, PSi is a very versatile sensing platform [1–5].<br />
A key feature for a recognition transducer is a large surface<br />
area: PSi has a porous structure with a specific area of the<br />
order of 200–500 m 2 /cm 3 , so that it is very sensitive to the<br />
presence of biochemical species which penetrate inside the<br />
pores. Moreover, PSi is an available, low-cost material,<br />
completely compatible with VLSI and micromachining<br />
technologies, so that it could usefully be employed in the<br />
fabrication of MEMS, MOEMS and smart sensors.<br />
The PSi optical sensing features are based on the changes<br />
of its photonic properties, such as photoluminescence or<br />
reflectance, on exposure to the gaseous or liquid substances.<br />
Unfortunately, these interactions are not specific,<br />
so that the PSi cannot be used as a selective optical<br />
Correspon<strong>di</strong>ng author.<br />
E-mail address: ivo.ren<strong>di</strong>na@na.imm.cnr.it (I. Ren<strong>di</strong>na).<br />
1386-9477/$ - see front matter r 2007 Elsevier B.V. All rights reserved.<br />
doi:10.1016/j.physe.2006.12.050<br />
ARTICLE IN PRESS<br />
www.elsevier.com/locate/physe<br />
transducer material. One way to overcome this limit is to<br />
chemically or physically mo<strong>di</strong>fy the PSi hydrogenated<br />
surface in order to enhance the sensor selectivity through<br />
specific biochemical interactions. The reliability of the<br />
biosensor strongly depends on the functionalization<br />
process: how simple, homogenous and repeatable it can<br />
be. This fabrication step is also crucial for the stability of<br />
the sensor: it is well known that as-etched porous silicon<br />
has a very reactive Si–H terminated surface due to the Si<br />
<strong>di</strong>ssolution process [6]. The substitution of the Si–H bonds<br />
with Si–C ones guarantees a much more stable surface<br />
from the thermodynamic point of view.<br />
Testing and demonstrating the PSi capabilities as a<br />
useful functional material in the optical transduction of<br />
biochemical interactions is only the first action in the<br />
realization of an optical biochip based on this nanostructured<br />
material. In this case, all the fabrication processes<br />
should be compatible with the utilization of biological<br />
probes and the feasibility of such devices must be proven.<br />
This means that the standard integrated circuit microtechnologies<br />
should be mo<strong>di</strong>fied and adapted to this new<br />
field of application. A very strong inter<strong>di</strong>sciplinary is<br />
required to match and resolve all these problems.
In this work, we review our recent experimental results in<br />
optical biosensing using PSi-based photonic devices and<br />
report our newest results about the design and fabrication<br />
of optical biochip by exploiting the PSi nanotechnology<br />
[7–12].<br />
2. Materials and methods<br />
Following the pioneering work of Mike Sailor’s group<br />
[13], we started using the PSi monolayers, which optically<br />
act as Fabry–Perot interferometers, as basic transducer in<br />
biosensing experiments. This is a very convenient choice<br />
since PSi monolayers are very easy to fabricate and to<br />
characterize. These devices have also a high sensitivity to<br />
changes in refractive index, up to 70 ppm [14]. By<br />
exploiting the optical features of PSi Fabry–Perot interferometers,<br />
we stu<strong>di</strong>ed several bio-molecular interactions:<br />
DNA single strands with their complementary strands,<br />
Glutamine bin<strong>di</strong>ng protein (GlnBP) from Escherichia coli<br />
with Glutamine; GlnBP with Glia<strong>di</strong>ne [8,11,12]. The highquality<br />
optical response of PSi multilayers allows the<br />
fabrication of resonant optical devices and photonic crystals<br />
that are very attractive for sensing applications [15].<br />
In biochemical sensors based on 1-D photonic bandgap<br />
(PBG) structures, usually there is both a reduced group<br />
velocity of light within the sample, which increases the<br />
interaction with the analyte and thus the sensitivity, and<br />
the presence of a resonant characteristic wavelength. In this<br />
case, sensors whose operation is based on the measurement<br />
of the shift of a very narrow resonance (for instance, a<br />
transmission or reflection peak having a full-width at halfmaximum<br />
of few tens of microns) are characterized by<br />
increased sensitivity and resolution. We have used a PSi<br />
optical microcavity to study the molecular bin<strong>di</strong>ng of<br />
GlnBP and glutamine with a higher sensitivity with respect<br />
to the interferometric characterization [10].<br />
Another typical PBG structure used in optical sensing is<br />
the <strong>di</strong>stributed Bragg mirror which is obtained by<br />
alternating high (A) refractive index layers (low porosity)<br />
and low (B) refractive index layers (high porosity). The<br />
thicknesses of each layer satisfy the following relationship:<br />
nAdA+nBdB ¼ ml/2, where m is an integer and l is the<br />
Bragg resonant wavelength. The main optical feature of a<br />
Bragg mirror is the presence of a single, degenerate<br />
photonic band gap in a period of the reciprocal space. As<br />
a consequence, the reflectivity spectrum is characterized by<br />
a principal (m ¼ 1) stop band, centred around l, whose<br />
width depends on the refractive indexes contrast (nA/nB),<br />
while its height, depends on the number of the layers. The<br />
higher orders of the resonance (m ¼ 2, 3, 4,y) are<br />
characterized by a narrow width and a lower reflectivity<br />
[16].<br />
We have designed, simulated and realized a low contrast,<br />
20-layer-Bragg mirror, with the first-order (m ¼ 1) resonant<br />
wavelength at 4800 nm. A highly doped p + -silicon,<br />
/100S oriented, 0.01 O cm resistivity, 400 mm thick was<br />
used as substrate to realize the structures. The Bragg low-<br />
ARTICLE IN PRESS<br />
I. Ren<strong>di</strong>na et al. / Physica E 38 (2007) 188–192 189<br />
porosity layers (effective refractive index nLffi1.56, thickness<br />
dLffi641 nm) were produced by an etching current<br />
density of 125 mA/cm 2 for 2.58 s, while an etching current<br />
density of 150 mA/cm 2 for 3.62 s has been used for the<br />
high-porosity layers (n Hffi1.50, d Hffi933 nm).<br />
For the biosensing experiments, a strong base post-etch<br />
treatment was used to increase the pore size and remove the<br />
superficial nano-residue due to the electrochemical etch<br />
process, so to improve the infiltration of biomolecular<br />
probes into the pores [17]. Unfortunately, this process<br />
removes most of the native Si–H bonds, required by the<br />
subsequent functionalization treatment, from the porous<br />
silicon surface: to restore these bonds, we rinsed the porous<br />
silicon device in a very <strong>di</strong>luted HF-based solution for 30 s.<br />
The reflectivity spectra were measured over the range<br />
800–1600 nm with a resolution of 0.2 nm by using an Y<br />
optical fibre connected to a tungsten lamp, as a light<br />
source, and to an optical spectrum analyzer.<br />
We have estimated the refractive index sensitivity, i.e.,<br />
the response of the sensor to the changes of the average<br />
refractive index, by measuring the peak shift of a higher<br />
order Bragg resonance on exposure to several organic<br />
volatile substances with <strong>di</strong>fferent and increasing refractive<br />
indexes. Each measurement starts after that a small<br />
amount of volatile substance is added to reaction chamber<br />
and saturates with its vapours the atmosphere surroun<strong>di</strong>ng<br />
the sensor. The vapours penetrate into the nanometric<br />
pores, substituting the air. The average refractive index of<br />
the layers changes, so that the reflectivity spectrum of the<br />
device shifts towards higher wavelengths.<br />
Procedures for the immobilization of biomolecules on<br />
porous silicon are usually based on a chemistry that<br />
involves the silanization of the oxi<strong>di</strong>zed PSi surface [18]. A<br />
promising recently proposed alternative is that exploiting<br />
the reaction of acids molecules with the hydrogenterminated<br />
porous silicon surface in order to obtain a<br />
more stable organic layer covalently attached to the PSi<br />
Reflectivity (a.u.)<br />
0.6<br />
0.5<br />
0.4<br />
0.2<br />
0.1<br />
8 nm<br />
700 750 800 850<br />
Wavelength (nm)<br />
Fig. 1. Reflectivity spectra of a porous silicon Bragg mirror.
190<br />
surface through Si–C bonds. We have exploited a photoactivated<br />
mo<strong>di</strong>fication of PSi surface based on the UV<br />
exposure of an N-hydroxysuccinimide ester (UANHS)<br />
solution [19]. This kind of treatment is very fast since the<br />
time required for the complete surface passivation is of the<br />
order of few minutes. Moreover, after UV exposition the<br />
Red Shift (nm)<br />
200<br />
190<br />
180<br />
170<br />
Methanol 1.328<br />
Pentane 1.358<br />
Isopropanol 1.377<br />
Isobutanol 1.396<br />
Refractive Index Unit (RIU) Sensitivity<br />
Bragg sensitivity<br />
470±40 nm/RIU<br />
160<br />
1.32 1.34 1.36<br />
Refractive Index<br />
1.38 1.40<br />
Fig. 2. Red shift of the Bragg peak at 750 nm versus the refractive index of<br />
vapor substances.<br />
% Reflectivity<br />
140<br />
120<br />
100<br />
80<br />
60<br />
40<br />
20<br />
0<br />
ARTICLE IN PRESS<br />
I. Ren<strong>di</strong>na et al. / Physica E 38 (2007) 188–192<br />
chip can be analysed by FT–IR microscopy to verify the<br />
state of the process. In view of a possible integration of<br />
the PSi-based biosensor into a lab-on-chip, we have verified<br />
the thermal stability of the Si–C bonds up to the<br />
temperature adopted in the packaging process of the chip.<br />
Once functionalized, the PSi optical transducer was bonded<br />
to a glass slide by an innovative ano<strong>di</strong>c bon<strong>di</strong>ng process<br />
that takes into account the very peculiar characteristic of<br />
this material. Details of the process and of the microflui<strong>di</strong>c<br />
solutions adopted are elsewhere described [7].<br />
Even if our aim is the realization of a label-free optical<br />
biosensor based on the PSi nanotechnology, we have used a<br />
fluorescent protein to control the <strong>di</strong>stribution of the<br />
biological matter on the chip surface and to preliminary<br />
test the chemical stability of the covalent link between the<br />
protein and the PSi surface.<br />
3. Experimental results and <strong>di</strong>scussion<br />
-20<br />
4000 3500 3000 2500 2000 1500 1000<br />
Fig. 1 shows the reflectivity spectrum of the PSi Bragg<br />
mirror in the range 700–850 nm: we have found that despite<br />
the high order of the resonance (m ¼ 6), the resonant peak<br />
at l ¼ 750 nm has a reflectivity which is the 60% of the first<br />
order and the width is less than 10 nm. Note that due to<br />
refractive index and thickness fabrication uncertainties the<br />
m ¼ 6 Bragg wavelength is about 750 nm instead of<br />
800 nm.<br />
Wavenumber (cm -1 )<br />
Bragg before functionalization<br />
Bragg after UANHS treatment<br />
Fig. 3. FT–IR spectra of the realized porous silicon Bragg structure before and after the photoinduced functionalization process based on UV exposure,<br />
together with the reaction scheme.
In Fig. 2, the red shifts of the m ¼ 6 Bragg resonance<br />
peak at 750 nm are reported as function of the refractive<br />
index of each volatile substance employed. A well linear<br />
response is observed and a sensitivity of 470 (40) nm/RIU<br />
can be estimated, where RIU is the acronym for refractive<br />
index unit. Since the resolution of our spectrometer is<br />
0.05 nm, a limit of detection for the refractive index change<br />
of 1.06 10 4 can be estimated for this kind of optical<br />
device.<br />
The UV-induced surface passivation process results in<br />
covalent attachment of UANHS to the porous silicon<br />
surface, as clearly shown in the FT–IR spectrum reported<br />
in Fig. 3 together with the reaction scheme: all the<br />
characteristic absorption peaks due to the organic compounds<br />
are present in the sample after the functionalization<br />
process. The chip was also optically characterized: a red<br />
shift of about 54 nm in the reflectivity spectrum is observed.<br />
Reflectance %<br />
50<br />
40<br />
30<br />
20<br />
10<br />
2 micron 70% porosity with UANHS<br />
after treatment @250°C 2'<br />
after treatment @250°C 4'<br />
after treatment @250°C 6'<br />
Si-C<br />
C-N-C<br />
1100 1000 900 800 700<br />
Wavenumber (cm -1 )<br />
Fig. 4. FT–IR spectra of a PSi chip after a thermal treatment at <strong>di</strong>fferent<br />
times.<br />
ARTICLE IN PRESS<br />
I. Ren<strong>di</strong>na et al. / Physica E 38 (2007) 188–192 191<br />
Then the device was incubated over night with the<br />
biological probe. The selective recognition of the biomolecular<br />
target is optically verified by a further red shift<br />
(data will be published elsewhere).<br />
The FT–IR spectrum of a PSi chip after functionalization<br />
at <strong>di</strong>fferent temperatures is reported in Fig. 4: itis<br />
well evident that the link between the inorganic and<br />
organic phases is stable up to 250 1C. The Si–C and C–N–C<br />
bonds have high thermodynamic strength so that the<br />
passivated chip is still working after the ano<strong>di</strong>c bon<strong>di</strong>ng<br />
process.<br />
In Fig. 5, a real view of the hybrid porous silicon-glass<br />
chip (A) is reported, while in the insets (B) and (C) the<br />
fluorescent images observed by a Leica Z16 APO<br />
fluorescence macroscopy system after incubation of the<br />
labelled bioprobes bonded to the chip, are shown. By<br />
illuminating the chip spotted with labelled proteins<br />
(fluorescent glutamine bin<strong>di</strong>ng protein, purified by E. coli<br />
[8]) by a 100 W high-pressure mercury source, we found<br />
that the fluorescence is very high and homogeneous on the<br />
whole surface (see Fig. 5(B)). In Fig. 5(C) a detail of a<br />
scratch on the chip surface can be seen, showing that the<br />
fluorescent probes also penetrate into the inner pores of the<br />
chip. After several flushing run with demineralized water,<br />
injected by a micro-syringe, we found that the fluorescence<br />
is still bright.<br />
4. Conclusions<br />
In this work, our recent experimental results in optical<br />
biosensing using PSi-based photonic devices are reviewed.<br />
Newest results about the fabrication and characterization<br />
of PSi sensors based on the exploitation of <strong>di</strong>stributed<br />
Bragg reflectors operating at higher order harmonics are<br />
also presented. The PSi surface functionalization of such<br />
structures by UV stimulated processes, allowing the<br />
attachment of linkers through stable covalent Si–C bonds,<br />
is also demonstrated to be compatible to the hightemperature<br />
ano<strong>di</strong>c bon<strong>di</strong>ng processes used for the<br />
realization of silicon-glass biochips.<br />
Fig. 5. Images of the basic element of a lab-on-chip based on a PSi optical transducer: (A) a real view of the porous hybrid silicon-glass chip; (B)<br />
fluorescent images of labelled proteins; (C) particular of a scratch on the chip surface.
192<br />
Acknowledgments<br />
The authors are grateful to Prof. Mose` Rossi and Prof.<br />
Paolo Arcari for the helpful <strong>di</strong>scussions. They also wish to<br />
thank Dr. Sabato D’Auria and Dr. Annalisa Lamberti for<br />
all the contributes given in preparing and making the<br />
experiments with the biological matter.<br />
References<br />
[1] H. Ouyang, M. Christophersen, R. Viard, B.L. Miller, P.M. Fauchet,<br />
Adv. Funct. Mater. 15 (2005) 1851.<br />
[2] S.M. Weiss, H. Ouyang, J. Zhang, P.M. Fauchet, Opt. Express 13<br />
(2005) 1090.<br />
[3] F.P. Mathew, E.C. Alocilja, Biosensors Bioelectron. 20 (2005) 1656.<br />
[4] F. Besseueille, V. Dugas, V. Vikulov, J.P. Cloarec, E. Souteyrand,<br />
Biosensors Bioelectron. 21 (2005) 908.<br />
[5] C. Pacholski, M. Sartor, M.J. Sailor, F. Cunin, G.M. Miskelly,<br />
J. Am. Chem. Soc. 127 (2005) 11636.<br />
[6] L. Canham (Ed.), Properties of Porous Silicon, IEE Inspec, London,<br />
UK, 1997.<br />
[7] L. De Stefano, K. Malecki, F.G. Della Corte, L. Moretti, L. Rotiroti,<br />
I. Ren<strong>di</strong>na, Sensor Actuat B 114 (2006) 625.<br />
[8] L. De Stefano, I. Ren<strong>di</strong>na, L. Rotiroti, L. Moretti, V. Scognamiglio,<br />
M. Rossi, S. D’Auria, Biosensors Bioelectron. 21/8 (2006) 1664.<br />
ARTICLE IN PRESS<br />
I. Ren<strong>di</strong>na et al. / Physica E 38 (2007) 188–192<br />
[9] L. De Stefano, I. Rea, I. Ren<strong>di</strong>na, L. Rotiroti, M. Rossi, S. D’Auria,<br />
Phys. Stat. Sol. (a) 203 (2006) 886.<br />
[10] S. D’Auria, S. Boarino, A. Vitale, A.M. Rossi, I. Rea, I. Ren<strong>di</strong>na,<br />
L. Rotiroti, M. Staiano, M. de Champdorè, V. Aurilia, A. Parracino,<br />
M. Rossi, L. De Stefano, J. Phys.: Condens. Matter 18 (2006)<br />
S2019.<br />
[11] L. De Stefano, M. Rossi, M. Staiano, G. Mamone, A. Parracino, L.<br />
Rotiroti, I. Ren<strong>di</strong>na, M. Rossi, S. D’Auria, J. Proteome Res. 5 (5)<br />
(2006) 1241.<br />
[12] L. De Stefano, L. Rotiroti, I. Rea, I. Ren<strong>di</strong>na, L. Moretti, G. Di<br />
Francia, E. Massera, P. Arcari, A. Lamberti, C. Sangez, J. Opt. A 8<br />
(2006) S540.<br />
[13] V.S.-Y. Lin, K. Motesharei, K.-P.S. Dancil, M.J. Sailor, M.R.<br />
Gha<strong>di</strong>ri, Science 278 (1997) 840.<br />
[14] M.A. Anderson, A. Tinsley-Brown, P. Allcock, E.A. Perkins, P.<br />
Snow, M. Hollings, R.G. Smith, C. Reeves, D.J. Squirrell, S. Nicklin,<br />
T.I. Cox, Phys. Stat. Sol. (A) 197 (2) (2003) 528.<br />
[15] S. Chan, Y. Li, L.J. Rothberg, B.L. Miller, P.M. Fauchet, Mater. Sci.<br />
Eng. C 15 (2001) 277.<br />
[16] L. Pavesi, RIV. Nuovo Cimento 20 (1997) 1.<br />
[17] L.A. DeLouise, B.L. Miller, Mater. Res. Soc. Symp. Proc. 782 (2004)<br />
A5.3.1.<br />
[18] B.R. Hart, S.E. Letant, S.R. Kane, M.Z. Ha<strong>di</strong>, S.J. Shields, J.G.<br />
Reynolds, Chem. Commun. (2003) 322.<br />
[19] H.B. Yin, T. Brown, R. Gref, J.S. Wilkinson, T. Melvin, Microelectron.<br />
Eng. 73–74 (2004) 830.
Chapter 3<br />
Hybrid systems based on Porous Silicon<br />
and biocompatible materials
Introduction: PSi hybrid systems<br />
Much attention has been attracted in the field of porous silicon (PSi) for its <strong>di</strong>verse<br />
applications in bio- and chemical sensing [1–3]. Unique properties for theses applications<br />
include the significantly increased surface interaction area, simplicity and repeatability of<br />
fabrication, and compatibility with the well-established silicon microfabrication technology.<br />
One reason of this clear success is the easy fabrication of sophisticated optical<br />
multilayers, such as the one-<strong>di</strong>mensional photonic crystals, by a simple, but not trivial,<br />
electrochemical etch process, computer controlled [4, 5]. The reflectivity spectrum of the<br />
PSi based photonic crystals can show single or multiple resonances [6], band-gap and<br />
other characteristic shapes, depen<strong>di</strong>ng on the spatial arrangement of the porous layers [7],<br />
which are very useful in lot of applications from biochemical sensing to me<strong>di</strong>cal imaging.<br />
But the major drawback preventing a wide <strong>di</strong>ffusion of PSi devices is the instability of its<br />
native interface with a metastable Si–Hx termination. The metastable hydrosilicon can<br />
undergo spontaneous oxidation in ambient atmosphere and results in the degradation of<br />
surface structures.<br />
Moreover, it has been shown that porous silicon wafer can be <strong>di</strong>ssolved under the<br />
physiological con<strong>di</strong>tions which are very often used in biological experiments [8].<br />
Therefore, surface passivation is crucial for the technological success of this material.<br />
Many organic molecules such as alkenes or alkynes had been covalently grafted onto the<br />
PSi surface by wet chemical approaches to passivate its surface [Chapter 2].<br />
A very promising approach to overcome these <strong>di</strong>fficulties is in fabricate hybrid systems<br />
based on the infiltration of polymers or biofilm resulting from the self-assembling of<br />
proteins in the nanometric pores of the PSi matrix. This is an alternative to the<br />
conventional chemical surface mo<strong>di</strong>fication strategies.<br />
Recently, beside hydrocarbon monomer monolayers, polymer films, are of special interest<br />
because they provide a strong shield to protect the surface nanostructure.<br />
An example of hybrid PSi-polymer systems are the mono<strong>di</strong>sperse microparticles<br />
containing a tunable nanostructure, these devices are desired for various high-throughput<br />
screening and targeted drug delivery applications [9,10]. The applicability of porous siliconbased<br />
particles and films has been demonstrated [11-21].<br />
Moreover the field of tissue engineering entails the creation of a degradable three<br />
<strong>di</strong>mensional scaffold structure that has the appropriate physical, chemical, and mechanical<br />
properties to enable cell penetration and subsequent tissue formation for the purpose of<br />
<strong>di</strong>sease or trauma repair at the desired site [22]. Bone engineering specifically attempts to<br />
replace or increase current graft approaches by using porous scaffolds that are designed<br />
to support the migration, proliferation, and <strong>di</strong>fferentiation of osteoprogenitor cells and aid in<br />
the organization of these cells in the proper <strong>di</strong>mensions. While these scaffolds may be<br />
made from a wide variety of both natural and synthetic materials, many of the existing<br />
porous biodegradable polymeric systems have been found to have limitations for use as<br />
orthope<strong>di</strong>c scaffolds [23].<br />
Mesoporous Silicon, also referred to by its proprietary name BioSilicon , possesses<br />
several unique features highlighting its utility as a biomaterial. Previous stu<strong>di</strong>es by<br />
Bow<strong>di</strong>tch et al have demonstrated its ability to be resorbed in soft tissue with a negligible<br />
inflammatory response in vivo [24]. When coupled with an established biopolymer such as<br />
polycaprolactone (PCL), it can be processed into a broad range of self-assembling<br />
structures useful to tissue engineering [25].<br />
67
Schematic of a PSi device as etched and after polymer infiltration.<br />
As was pointed out in the previous chapter, biomolecules can be linked via Si–C bonds or<br />
Si–O. Covalent attachment provides a convenient means to link a biomolecular capture<br />
probe to the inner pore walls of PSi for biosensor applications.<br />
Hybrid materials, in which the functional material consists of an organic polymer or a<br />
biopolymer, forms an ad<strong>di</strong>tional class of devices. Composites are attractive can<strong>di</strong>dates for<br />
biological devices because they can <strong>di</strong>splay a combination of advantageous chemical and<br />
physical characteristics not exhibited by the in<strong>di</strong>vidual constituents. Advances in polymer<br />
and materials chemistries have greatly expanded the design options for nanomaterial<br />
composites in the past few years, and synthesis of materials using nanostructured<br />
templates has emerged as a versatile technique to generate ordered nanostructures [26].<br />
Templates consisting of micro or mesoporous membranes, zeolites, and crystalline<br />
colloidal arrays have been used, and many elaborate electronic, mechanical, or optical<br />
structures have resulted.<br />
In this chapter the mo<strong>di</strong>fication of PSi structures by infiltration of chemical and biological<br />
polymers is describes. In particular, it was used a biocompatible multi-block copolymer<br />
based on PCL and a high stable and resistant biofilm resulting from the self-assembling of<br />
the hydrophobins (HFBs) [27-29].<br />
The polymer contains pendant functional groups regularly spaced along the chain. The<br />
reactive groups are available for the attachment of chemical substances and their<br />
<strong>di</strong>stribution, can be tuned accor<strong>di</strong>ng to the molecular weight of both segments. In this way,<br />
the physical and chemical properties of the porous silicon surface were changed and<br />
modulated.<br />
On the other hand HFBs are a family of small and cysteine-rich fungal proteins produced<br />
in the hyphal cell walls and that can be purified by the culture me<strong>di</strong>um. The HFBs selfassemble<br />
into thin and amphiphilic membranes at hydrophilic-hydrophobic interfaces. In<br />
this way the wettability of a surface can be controlled and the hydrophobic behaviour can<br />
be converted in hydrophilic, and vice versa [30-32].<br />
A general classification of HFBs is made on the basis of the biofilm resistance: the class I<br />
HFBs form high insoluble assemblies, which can be <strong>di</strong>ssolved in strong acids, whereas<br />
class II biofilms can be <strong>di</strong>ssolved in ethanol or in so<strong>di</strong>um dodecyl sulphate [27].<br />
Solutions with HFB or PCL have been used at <strong>di</strong>fferent concentration and the infiltration<br />
process has been monitored by variable angle spectroscopic ellipsometry (VASE)<br />
measurements.<br />
References<br />
1. V.S. Lin, K. Motesharei, K.S. Dancil, M.J. Sailor, M.R. Gha<strong>di</strong>ri, Science 278 (1997) 840.<br />
2. J. Gao, T. Gao, M.J. Sailor, Appl. Phys. Lett. 77 (2000) 901.<br />
3. S. Le´tant, M.J. Sailor, Adv. Mater. 12 (2000) 355.<br />
4. F. Müller, A. Birner, U. Gösele, V. Lehmann, S. Ottow, H. Föll, Journal of Porous Materials 2000,<br />
7, 201.<br />
5. H. F. Arrand, T.M. Benson, A. Loni, R. Arens-Fischer, M. Kruger, M. Thonissen, H. Luth, S.<br />
Kershaw, IEEE Phot. Techn. Lett. 1998, 10, 1467.<br />
6. V. Mulloni, L. Pavesi, Appl. Phys. Lett. 2000, 76, 2523.<br />
7. L. Moretti, L. De Stefano, I. Rea, I. Ren<strong>di</strong>na, Appl. Phys. Lett. 2007, 90, 191112.<br />
68
8. S.H.C. Anderson, H. Elliot, D.J. Wallis, L.T. Canham, J.J. Powell, Phys. Stat. Sol. (a) 2003, 197,<br />
331.<br />
9. R. Duncan, Natl. Rev. Drug Discov. 126, 1470 (2003).<br />
10. New Types of Polymer Particles, AnalytiX: Advances in Analytical Chemistry, No. 5 (Sigma-<br />
Aldrich/Fluka Chemicals, 2001).<br />
11. J. R. Link and M. J. Sailor, Proc. Natl. Acad. Sci. USA 100, 607 (2003).<br />
12. V. S.-Y. Lin, K. Motesharei, K. S. Dancil, M. J. Sailor, and M. R. Gha<strong>di</strong>ri, Science 278, 840<br />
(1997).<br />
13. K.-P. S. Dancil, D. P. Greiner, and M. J. Sailor, J. Am. Chem. Soc. 121, 7925 (1999).<br />
14. P. A. Snow, E. K. Squire, P. S. J. Russell, and L. T. Canham, J. Appl. Phys. 86, 1781 (1999).<br />
15. S. Chan, P. M. Fauchet, Y. Li, L. J. Rothberg, and B. L. Miller, phys. stat. sol. (a) 182, 541<br />
(2000).<br />
16. S. Chan, S. R. Horner, B. L. Miller, and P. M. Fauchet, J. Am. Chem. Soc. 123, 797 (2001).<br />
17. T. A. Schmedake, F. Cunin, J. R. Link, and M. J. Sailor, Adv. Mater. 14, 1270 (2002).<br />
18. F. Cunin, T. A. Schmedake, J. R. Link, Y. Y. Li, J. Koh, S. N. Bhatia, and M. J. Sailor, Nature<br />
Mater. 1, 39 (2002).<br />
19. S. O. Meade, M. S. Yoon, K. H. Ahn, and M. J. Sailor, Adv. Mater. 16, 1811 (2004).<br />
20. L. T. Canham, Adv. Mater. 7, 1033 (1995).<br />
21. L. T. Chanham, M. P. Stewart, J. M. Buriak, C. L. Reeves, M. Anderson, E. K. Squire, P. Allcock,<br />
and P. A. Snow, phys. stat. sol. (a) 182, 521 (2000).<br />
22. Karp, J.M.; Dalton, P.D.; Shoichet, M.S., Mater. Res. Bull. 2003, 28,301.<br />
23. Hutmacher, D.W.. Biomaterials, 2000, 21, 2529.<br />
24. Bow<strong>di</strong>tch, A.; Waters, K.; Gale, H.; Rice, P.; Scott, E.; Canham, L.T.;Reeves, C.; Loni, A.; Cox,<br />
T. Mater. Res. Soc. Symp. Proc.,1999, 536,149.<br />
25. Coffer, J.; Whitehead, M.A.; Nagesha, D.; Mukherjee, P.; Akkaraju,G.;Totolici, M.;Saffie, R.;<br />
Canham, L.. Phys. Stat. Sol (a), 2005, 202,1451.<br />
26. K.S. Soppimath, T.M. Aminabhavi, A.R. Kulkarni, W.E. Rudzinski, Biodegradable polymeric<br />
nanoparticles as drug delivery devices, J. Control Release 70 (2001) 1–20.<br />
27. H. J. Hektor, K. Scholtmeijer, Current Opinion in Biotechnology 2005, 16, 434.<br />
28. H.A.B. Wosten, M. L. de Vocht, Biochimica and Biophysica Acta 2000, 1469, 79.<br />
29. G.R. Szilvay, A. Paananen, K. Laurikainen, E. Vuorimaa, H. Lemmetyinen, J. Peltonen, M.B.<br />
Linder, Biochemistry 2007, 46, 2345.<br />
30. J. T. Han, S. Kim, A. Karim, Langmuir 2007, 23, 2608.<br />
31. R. Wang, Y. Yang, M. Qin, L. Wang, L. Yu, B. Shao, M. Qiao, C. Wang, X. Feng, Chem. Mater.<br />
2007, 19, 3227.<br />
32. R. Wang, Y. Yang, M. Qin, L. Wang, L. Yu, B. Shao, M. Qiao, C. Wang, X. Feng, Langmuir<br />
2007, 23, 4465.<br />
69
Porous silicon-polymer composite matrix for an innovative<br />
class of optical biosensors<br />
This section is aimed at the characterization of PSi surface mo<strong>di</strong>fication, based on the<br />
infiltration of Poly(ε-caprolactone) (PCL) macro-monomers in the nanometric pores of its<br />
matrix.<br />
In collaboration with Prof. R. Palumbo of University of Naples “Federico II”, Italy
A polymer mo<strong>di</strong>fied PSi optical device for biochemical sensing
Introduction<br />
Due to their widespread utilisation, there are all over the world tens of thousand of<br />
hydrocarbons contaminated sites, where the pollution is caused by leaking of storage<br />
tanks, by chemical wastes dumps and also by extraction and transport operations. In<br />
ad<strong>di</strong>tion, there is the real problem of vapours leaks from <strong>di</strong>stributed gas pipelines in<br />
metropolitan areas. Therefore, an accurate characterization and long-term monitoring is<br />
required to reduce and control health risk and public safety. Four general groups of<br />
chemical sensors can be categorized on the base of physics and operating mechanisms:<br />
chromatography and spectrometry, electrochemical sensors, mass sensors and optical<br />
sensors [1]. Although a number of these devices are commercially available, there is a<br />
strong need of low-cost, robust and on field sensors of chemical species, especially for<br />
long, real-time monitoring [1].<br />
In particular, PSi optical sensor based on reflectivity measurements in Bragg grating<br />
mirrors [2] or on photoluminescence experiments in optical microcavities operating in the<br />
visible wavelength region [3, 4] have been already reported.<br />
A typical photonic band gap (PBG) structure used in optical sensing is the <strong>di</strong>stributed<br />
Bragg mirror which is obtained by alternating high (A) refractive index layers (low-porosity)<br />
and low (B) refractive index layers (high-porosity). The thicknesses of each layer satisfy<br />
the following relationship: nAdA+nBdB=m λ/2, where m is an integer and λ is the Bragg<br />
resonant wavelength.<br />
When the PSi Bragg mirror is exposed to vapors, or <strong>di</strong>p in liquid hydrocarbons, a<br />
repeatable and completely reversible change in the reflectivity spectrum is observed. In<br />
fact, the substitution of air in the pores by the organic molecules causes an increasing of<br />
the average refractive index of the microcavity, shifting towards longer wavelengths its<br />
characteristic peaks.<br />
• PSi functionalization by polymer infiltration<br />
Covalent attachment provides a convenient means to link a biomolecular capture probe to<br />
the inner pore walls of PSi for biosensor applications. As was pointed out in the previous<br />
section, biomolecules can be linked via Si–C bonds or Si–O.<br />
Advances in polymer [5] and materials chemistries have greatly expanded the design<br />
options for nanomaterial composites in the past few years, and synthesis of materials<br />
using nanostructured templates has emerged as a versatile technique to generate ordered<br />
nanostructures. Templates consisting of micro or mesoporous membranes, zeolites, and<br />
crystalline colloidal arrays have been used, and many elaborate electronic, mechanical, or<br />
optical structures have resulted.<br />
Materials and methods<br />
Bragg multilayers were constituted by 15 couples of 58 % and 69 % porosities and 72 nm<br />
and 117 nm thicknesses, respectively.<br />
A highly doped p+-silicon, oriented, 0.01 Ω cm resistivity, 400 μm thick was used as<br />
substrate to realize the structures. The Bragg low-porosity layers (effective refractive index<br />
nL≅1.903) have been produced by an etching current density of 100 mA/cm<br />
73<br />
2 for 2.2 s,<br />
while an etching current density of 200 mA/cm 2 for 2.3 s has been used for the highporosity<br />
layers (nH ≅ 1.632).<br />
The electrochemical <strong>di</strong>ssolution process leaves nano<strong>di</strong>mensional residues into the<br />
channels which partially fill the pores in the layers and prevent a homogeneous <strong>di</strong>ffusion of<br />
biological solutions. To remove the nanostructures and improve pore infiltration of polymer,<br />
a KOH post-etch process elsewhere described [6] was used. It was verified that the<br />
alkaline treatment does not degrade the optical response of our devices. In order to leave<br />
an hydrophobic surface the Si-H bonds were replaced on the surface, by immersing the<br />
samples in a very <strong>di</strong>luted HF solution for few seconds.
The polymer solutions were prepared by solving the polymer in <strong>di</strong>chloromethane (CH2Cl2).<br />
A solution volume of 100 μl was <strong>di</strong>rectly spotted on the PSi surface, spun at 2000 rpm for 5<br />
seconds, and baked at 100°C for 5 minutes.<br />
In Figure 1 the red shift induced in the Bragg reflectivity spectrum by the presence of PCL<br />
is reported.<br />
Reflectivity (a.u.)<br />
1.0<br />
0.8<br />
0.6<br />
0.4<br />
0.2<br />
0.0<br />
800 1000 1200<br />
Wavelength (nm)<br />
Bragg pre PCL<br />
post PCL<br />
Figure 1: PSi reflectivity spectra before and after PCL infiltration.<br />
• PSi/PCL hybrid system: alkaline stability<br />
In order to test the new properties of the PSi-polymer composite device, two twins chips,<br />
one polymer infiltrated and the other bare, were exposed to a 0.1 M aqueous solution of<br />
so<strong>di</strong>um hydroxide. The <strong>di</strong>ssolution time evolution of both the samples were compared:<br />
initially both undergo a shift through minor wavelength, due to the removal by NaOH<br />
<strong>di</strong>ssolution of some unprotected silicon nanocrystals, but while after 80 s the uncoated<br />
sample is almost completely <strong>di</strong>ssolved, while the PCL coated PSi Bragg still resist for 15<br />
minutes. In Figure 2 are reported the optical spectra of both samples for <strong>di</strong>rect<br />
comparison.<br />
Reflectivity (a.u.)<br />
Reflectivity (a.u.)<br />
1<br />
0<br />
1.5<br />
1.0<br />
0.5<br />
PSi/PCL<br />
after 15 min in NaOH<br />
600 800 1000 1200 1400 1600<br />
PSi<br />
after 80s in NaOH<br />
0.0<br />
600 800 1000 1200 1400 1600<br />
Wavelength (nm)<br />
Figure 2: Comparison between a coated (A) and an uncoated (B) PSi Bragg mirrors before (black) and after<br />
immersion in 0.1 M NaOH solution (red).<br />
Even if blue-shifted the optical spectrum (A) yet shows a reflectivity stop band.<br />
• PSi/PCL hybrid system: chemical sensing<br />
After the basic etch process the hybrid organic-inorganic device is still working as chemical<br />
optical transducer: it was tested its ability in sensing the vapours of <strong>di</strong>fferent volatile<br />
substances. In Figure 3-A, are reported the characteristic red-shift due to the capillary<br />
condensation of the vapours inside the nanometric pores of the PSi Bragg. The shift is<br />
completely reversible and the polymer coated device can be used with <strong>di</strong>fferent<br />
substances (Isopropanol, ethanol and methanol) giving high reproducible results.<br />
74
Organic compound Refractive index<br />
Isopropanol 1.375<br />
Ethanol 1.361<br />
Methanol 1.329<br />
It was also calculated the sensitivity of this optical transducer to the refractive index<br />
changes by exposing it to substances having <strong>di</strong>fferent refractive index. Assuming that the<br />
three solvents equally penetrates the nanostructured spongy multilayer, was estimated a<br />
sensitivity of 585 (3) nm expressed in refractive index units [7].<br />
Reflectivity (a.u.)<br />
1.0<br />
0.8<br />
0.6<br />
0.4<br />
0.2<br />
0.0<br />
600 800 1000 1200 1400 1600<br />
Wavelength (nm)<br />
A<br />
unperturbed Bragg<br />
Methanol<br />
Isopropanol<br />
Ethanol<br />
Δλ (nm)<br />
90<br />
80<br />
70<br />
60<br />
Refractive Index Unit (RIU) Sensitivity<br />
S=585 (3) nm/RIU<br />
1.32 1.33 1.34 1.35 1.36 1.37 1.38<br />
Refractive index<br />
B<br />
Figure 3: The PSi polymer mo<strong>di</strong>fied Bragg still works as optical transducer for vapour and liquid detection. A:<br />
the characteristic red shift of the optical spectrum on exposure to methanol, isopropanol and ethanol. B:<br />
determination of the sensitivity to refractive index changes of the optical transducer.<br />
It was demonstrated that the passivation of some optical devices based on the porous<br />
silicon technology when infiltrated by polymer not only protects the nanocrystalline material<br />
from basic <strong>di</strong>ssolution in NaOH, but also leaves unaltered the sensing ability of such<br />
optical transducers ad<strong>di</strong>ng the chemical stability which can be a key feature in<br />
biomolecular experiments.<br />
It was moreover experimentally stu<strong>di</strong>ed a feasible optical sensor for hydrocarbons<br />
detecting in the gaseous state. The optical experimental setup adopted and the structure<br />
of the sensing element do not prevent from the scaling down of each component in a<br />
compact sensor system with minimum effort. In fact, the white source can be replaced by<br />
small, wide bandwidth LEDs and the OSA bench version by an integrated fiber optic<br />
spectrometer, already available on the market. Moreover, the use of a resonant device<br />
allows “on/off” <strong>di</strong>gital rea<strong>di</strong>ng of the sensor element by a single-wavelength source. A net<br />
red-shift of the transmittance peak in the reflectivity spectrum has been repeatable<br />
observed when the sensor is exposed to the hydrocarbons species. The shift values, due<br />
to capillary condensation of the liquid into the pores, are characteristic of each species and<br />
depend on their physical and chemical properties. The sensing process is completely<br />
reversible.<br />
• PSi/PCL hybrid system:biochemical sensing<br />
The hybrid interface of silicon and PCL should act as an active substrate for covalently<br />
bin<strong>di</strong>ng of <strong>di</strong>fferent bioprobes, so that the quality and the characteristic of this polymer<br />
layer are key features for the entire device. This represents an alternative to the<br />
conventional chemical surface mo<strong>di</strong>fication strategies. The reactive groups, introduced by<br />
infiltration of biocompatible polymer into PSi matrix, are available for the attachment of<br />
chemical substances and biological probes, in order to verify this feature a methanol<br />
75
solution of 5(6)-Carboxyfluorescein N-hydroxysuccinimide ester (FITC) 300 μM was used<br />
as ligand. A schematic of immobilization chemistry is reported in figure 4.<br />
A B C<br />
Figure 4: The PSi polymer mo<strong>di</strong>fied device. A: the characteristic reactive groups on the PSi surface after spin<br />
coating. B: 5(6)-Carboxyfluorescein N-hydroxysuccinimide ester (FITC) molecular formula. C: Fluorescence<br />
image after incubation.<br />
The immobilization was carried in methanol solution at room temperature over night. After<br />
incubation, the sample was rinsed three times in methanol. By illuminating the fluorescent<br />
FITC samples, a bright fluorescence signal was recorded, and it was quite homogeneous<br />
on the whole surface. It was also qualitatively tested the strength of the bin<strong>di</strong>ng between<br />
the PCL and the FITC, by washing the chip in a <strong>di</strong>alysis membrane overnight in methanol.<br />
Since the fluorescent intensities before and after the <strong>di</strong>alysis do not <strong>di</strong>ffer, can be<br />
concluded that the PCL-FITC system is very stable.<br />
References<br />
1. C. K. Ho, M. T. Itamura, M. Kelley and R. G. Hughes, San<strong>di</strong>a Report SAND2001-0643, San<strong>di</strong>a<br />
National Laboratories, (2001) 1-27.<br />
2. P. A. Snow, E. K. Squire, P. St. J. Russel, and L. T. Canham, J. Appl. Phys. 86, (1999) 1781-<br />
1784.<br />
3. V. Mulloni and L. Pavesi, Appl. Phys. Lett. 76, (2000) 2523-2525.<br />
4. M. Ben-Chorin, A. Kux, and I. Schechter, Appl. Phys. Lett. 64, (1994) 481-483.<br />
5. P. A. Snow, E. K. Squire, P. St. J. Russel, and L. T. Canham, J. Appl. Phys. 86, (1999) 1781-<br />
1784.<br />
6. L. A. De Louise and B. L. Miller (2004), Mat. Res. Soc. Symp. Proc. Vol. 782, A5.3.1- A5.3.7.<br />
7. L. De Stefano, I. Ren<strong>di</strong>na, L. Moretti, A.M. Rossi, Materials Science and Engineering B 2003,<br />
100/3, 271.<br />
76
Bio/NON Bio Interfaces for a new class of proteins Microarrays<br />
This section is aimed at the characterization of PSi surface biological passivation, based<br />
on the infiltration by the fungal proteins called hydrophobins. The self-assembled biofilm<br />
changes the wettability of the PSi structures and protects the PSi structures from alcaline<br />
agent. Moreover these is a new active substrate as a functional one for proteins<br />
microarrays development.<br />
In collaboration with Prof.ssa P. Giar<strong>di</strong>na of University of Naples “Federico II”, Italy
5<br />
CREATED USING THE RSC ARTICLE TEMPLATE (VER. 3.1) - SEE WWW.RSC.ORG/ELECTRONICFILES FOR DETAILS<br />
ARTICLE TYPE www.rsc.org/xxxxxx | XXXXXXXX<br />
Bioactive Mo<strong>di</strong>fication of Silicon Surface using Self-assembled<br />
Hydrophobins from Pleurotus ostreatus.<br />
L. De Stefano* a , I. Rea a,b , E. De Tommasi a , I. Ren<strong>di</strong>na a , L. Rotiroti a,c , M. Giocondo d , S. Longobar<strong>di</strong> e , A.<br />
Armenante e , and P. Giar<strong>di</strong>na e .<br />
Received (in XXX, XXX) Xth XXXXXXXXX 200X, Accepted Xth XXXXXXXXX 200X<br />
First published on the web Xth XXXXXXXXX 200X<br />
DOI: 10.1039/b000000x<br />
A crystalline silicon surface can be made biocompatible and chemically stable by a self-assembled<br />
biofilm of proteins, the hydrophobins (HFBs) purified from the fungus Pleurotus ostreatus. The<br />
10 protein mo<strong>di</strong>fied silicon surface shows an improvement in wettability and is suitable for proteins<br />
immobilization. Two <strong>di</strong>fferent proteins were successfully immobilized on the HFBs coated chips:<br />
the bovine serum albumin and an enzyme, a laccase, which retains its catalytic activity even when<br />
bound on the chip. Variable angle spectroscopic ellipsometry (VASE), water contact angle (WCA),<br />
and fluorescence measurements demonstrated that the proposed approach in silicon surface<br />
15 bioactivation is a feasible strategy for the fabrication of a new class of hybrid biodevices.<br />
20<br />
25<br />
30<br />
35<br />
40<br />
45<br />
50<br />
Introduction<br />
Biological molecules are more and more used in a large class<br />
of research and commercial applications such as biosensors,<br />
DNA and protein microarrays, cell culturing, immunological<br />
assays, and so on. The fabrication of a new generation of<br />
hybrid biodevices, where a biological counterpart is integrated<br />
in a micro or a nano-electronic platform, is strictly related to<br />
the bio-compatibilization treatments of the surfaces involved.<br />
1<br />
The design ad the realization of bio/non-bio interfaces with<br />
specific properties of chemical stability, wettability, and<br />
biomolecules immobilization are key features in the<br />
miniaturization of devices needed in the development of<br />
genomic and proteomic research. 2 In particular, protein<br />
immobilization is a hot topic in biotechnology since<br />
commercial solutions such as the DNA array are not still<br />
available. Proteins are, due to their composition, a class of<br />
very heterogeneous macromolecules with variable properties.<br />
For these reasons, it is extremely complex to find a common<br />
surface suitable for <strong>di</strong>fferent proteins with a broad range in<br />
molecular weight and physical–chemical properties such as<br />
charge and hydrophobicity. A further aspect is the orientation<br />
of spotted compounds, that becomes of crucial relevance for<br />
some applications, like quantitative analysis, enzymatic<br />
reactions, interaction stu<strong>di</strong>es. Besides tethering the proteins to<br />
the surface by adhesion or non-oriented covalent attachment,<br />
recent developments for the <strong>di</strong>rected immobilization of<br />
proteins are emerging. These efforts are addressing challenges<br />
such as loss of enzymatic activity due to unfavourable<br />
orientation of the immobilized enzyme. 3 Despite the<br />
development of many <strong>di</strong>fferent surfaces in the last five years,<br />
notably only few systematic investigations have been<br />
conducted and yet, no universal surface ideal for all<br />
applications could be identified 4-8 . Among others, silicon is a<br />
very interesting platform due to its wide use in all the micro<br />
and nanotechnologies developed for the integrated circuits<br />
industry. Lot of stu<strong>di</strong>es about the chemical functionalization<br />
of silicon in order to make it compatible with the organic<br />
chemistry have been published in the last ten year. 9 A<br />
completely <strong>di</strong>fferent approach is to use a biological substrate<br />
to passivate the silicon surface: recently, a nanostructured<br />
self-assembled biofilm of amphiphilic proteins, the<br />
hydrophobins, was deposited by solution deposition on<br />
crystalline silicon and proved to be efficient as masking<br />
material in the KOH wet etch of the crystalline silicon. 10<br />
HFBs are a group of very surface active proteins. 11 They are<br />
small (around 10 kDa) proteins that originate from<br />
filamentous fungi, where they coat the spores and aerial<br />
structures, and they me<strong>di</strong>ate the attachment of fungal<br />
structures to hydrophobic surfaces and affect the cell wall<br />
composition. 12-13 HFBs self-assemble at the air/water interface<br />
and lower the surface tension of water. Furthermore, they<br />
shown self-assembling properties also at interfaces between<br />
oil and water, and between water and a hydrophobic solid. 14<br />
The primary structure of HFBs is characterized by a<br />
conserved pattern of eight cysteine residues, forming four<br />
intramolecular <strong>di</strong>sulfide bridges. 15 HFBs are further <strong>di</strong>vided<br />
into classes I and II based on their hydropathy patterns,<br />
although the amino acid sequence similarity both within and<br />
between the classes is small. Class I HB assemblies are<br />
remarkable due to their insolubility. They are insoluble even<br />
in hot solutions of so<strong>di</strong>um dodecyl sulfate (SDS) and can only<br />
be <strong>di</strong>ssolved in some strong acids, such as trifluoroacetic<br />
acid. 11<br />
Some very recent applications of class II hydrophobins in<br />
protein immobilization have been reported, nevertheless only<br />
one example of enzyme immobilization on a Class I<br />
hydrophobin (SC3) layer formed on a glassy carbon electrode<br />
has been investigated. 12 Inspired by these very promising<br />
results, we have stu<strong>di</strong>ed the immobilization of a fluorescent<br />
protein and a laccase enzyme on a Class I hydrophobin layer<br />
when self-assembled on silicon surface.<br />
This journal is © The Royal Society of Chemistry [year] Journal Name, [year], [vol], 00–00 | 1<br />
55<br />
60<br />
65<br />
70<br />
75<br />
80<br />
85<br />
Matherials and methods
5<br />
10<br />
15<br />
20<br />
25<br />
30<br />
35<br />
40<br />
45<br />
50<br />
55<br />
Protein purification. The HFB and the laccase POXC were<br />
isolated from the culture broth of P. ostreatus (type: Florida<br />
ATCC no. MYA-2306). For HFBs production, mycelia were<br />
grown at 28 °C in static cultures in 2 L flasks containing 500<br />
mL of potato dextrose (24 g/l) broth with 0.5% yeast extract.<br />
After 10 days of fungal growth, the HFBs aggregates were<br />
obtained by air bubbling of the culture broth by using a<br />
Waring blender. Foam was then collected by centrifugation at<br />
4000 g. The precipitate was freeze dried, treated with<br />
trifluoroacetic acid (TFA 100%) for 2h, and sonicated for 30<br />
min. The sample was dried again in a stream of air and then<br />
<strong>di</strong>ssolved in 60% ethanol. The precipitate formed was<br />
separated by centrifugation at 4000g. the purity of the HFB<br />
was ascertained by SDS-PAGE, by using a silver staining<br />
method. Mycelium growth and laccase purification were<br />
carried out following the procedures described by Palmieri et<br />
al.. 16<br />
Hydrophobin self-assembly on silicon chip. Silicon samples,<br />
single side polished, oriented (chip size: 1cm×1cm),<br />
after standard cleaning procedure, 18 have been washed in<br />
hydrofluori<strong>di</strong>c acid solution for three minutes to remove the<br />
native oxide thin layer (1-2 nm) due to silicon oxidation. Then<br />
samples have been incubated in HFBs solutions (0.1 mg/mL<br />
in 80% ethanol) for 1 h, dried for 10 min on the hot plate<br />
(80°C), and then washed by the solvent solution (80% ethanol<br />
in water).<br />
Laccase immobilization method. The laccase immobilization<br />
procedure on HFB mo<strong>di</strong>fied silicon surface was carried at 4<br />
°C for 1 hour, followed by a over night rinsing in 50 mM<br />
phosphate buffer, pH 7. After 16 hours, the silicon biochip<br />
was washed in the same buffer at 25°C till to no laccase<br />
activity was found in the washing buffer. Then each chip was<br />
dust-protected and kept at constant humi<strong>di</strong>ty con<strong>di</strong>tions when<br />
not in use.<br />
BSA immobilization method. Bovine serum albumin (BSA)<br />
was labelled by rhodamine, following the Molecular Probes ©<br />
procedure MP06161, and solubilised in water at three<br />
<strong>di</strong>fferent concentration (3, 6, 12 μM). To assess the protein<br />
bin<strong>di</strong>ng on the chip surface, we have spotted on the chips<br />
covered by the HFBs biofilm 50 μl of water solution<br />
containing the labeled protein. The immobilization was<br />
carried in water solution at 4°C over night. After incubation,<br />
the samples were rinsed three times in deionized water at<br />
room temperature.<br />
Enzyme assay Laccase activity was assayed at 25°C using 2,6<strong>di</strong>methoxyphenol<br />
(DMP) 10mM in McIlvaine citrate–<br />
phosphate buffer, pH 5. Oxidation of DMP was followed by<br />
the absorption increment at 477 nm (ε477=14.8×10 3 M -1 cm -1 )<br />
using Beckman DU 7500 spectrophotometer (Beckman<br />
Instruments). Enzyme activity was expressed in International<br />
Units (IU). Immobilized enzyme was assayed by silicon <strong>di</strong>p in<br />
10 ml of McIlvaine citrate–phosphate buffer, pH 5. The<br />
absorption increment at 477 nm was followed withdrawing<br />
200 μl of reaction mixture each 30s for 10 min. The total<br />
immobilised activity per unit of silicon (chip) were calculated<br />
as U/chip = ΔA min -1 /ε × 10 4 .<br />
Optical Techniques. Ellipsometric characterization of the<br />
hydrophobin biofilm deposited on the silicon substrate has<br />
been performed by a variable-angle spectroscopic<br />
ellipsometry model (UVISEL, Horiba-Jobin-Yvon).<br />
Ellipsometric parameters Δ and Φ have been measured at an<br />
angle of incidence of 65° over the range of 360-1600 nm with<br />
a resolution of 5 nm. Fluorescence images were recorded by a<br />
Leica Z16 APO fluorescence macroscopy system. Contact<br />
angle measurements have been performed by using a KSV<br />
Instruments LTD CAM 200 Optical Contact Angle Meter:<br />
each contact angle has been calculated as the average between<br />
the values obtained from three drops having the same volume<br />
(about 3 μL), spotted on <strong>di</strong>fferent points of the chip.You can<br />
also put lists into the text.<br />
2 | Journal Name, [year], [vol], 00–00 This journal is © The Royal Society of Chemistry [year]<br />
60<br />
65<br />
70<br />
75<br />
80<br />
85<br />
90<br />
95<br />
100<br />
Results and Discussion<br />
The immobilization of proteins and enzymes on the silicon<br />
mo<strong>di</strong>fied surface is schematized in Figure 1.<br />
HFB<br />
BSA<br />
POXC<br />
cSi cSi<br />
Figure 1. Schematic of immobilization of enzyme (POXC)<br />
and protein (rhodamine fluorescent BSA) on a crystalline<br />
silicon wafer biologically mo<strong>di</strong>fied by a self-assembled<br />
biofilm of hydrophobin.<br />
The hybrid interface of silicon and HFB should act as an<br />
active substrate for non-covalently bin<strong>di</strong>ng of <strong>di</strong>fferent<br />
bioprobes, so that the quality and the characteristic of this<br />
proteins layer are key features for the entire device. The<br />
coating of a solid surface by chemical solution deposition is<br />
not a finely controlled process, even if the covering substance<br />
has self-assembling properties, as it is the case of HFBs. The<br />
result is that the HFB biofilms self-assembled on the silicon<br />
surface by deposition of equal concentration solutions could<br />
have thicknesses ranging between 10 and 30 nanometers. This<br />
behaviour could be ascribed to <strong>di</strong>fferent local aggregation of<br />
HFB in the solution covering the silicon chip, due to the<br />
strong hydrophobic interactions among the proteins. In fact,<br />
during HFB deposition, solvent evaporation can determine<br />
local increases of protein concentration, thus favouring<br />
multimers and aggregate formation that can deposit on the<br />
layer. In this step there is no way to control the clustering and<br />
stacking of the proteins which are constituting the selfassembled<br />
biofilm. The association of hydrophobins in<br />
solution and their self assembly have both been attributed to<br />
the amphiphilic structure of the HFB monomer. 19-20 Before<br />
using the HFB biofilm as a nanostructured template for<br />
protein bin<strong>di</strong>ng, we have tested its stability and some features<br />
by optical methods such as VASE, spectrofometry, and WCA.
5<br />
10<br />
15<br />
20<br />
25<br />
30<br />
35<br />
The HFB biofilm self-assembled on the hydrophobic silicon<br />
shows a strong adhesion to the surface: the protein layers<br />
always exhibit great stability to alkaline solutions even at high<br />
temperature. After 10 minutes washing in so<strong>di</strong>um hydroxide<br />
and SDS at 100 °C, a residual biofilm, which has on average a<br />
thickness of about 3.1±0.7 nm, is still present. The biofilm<br />
characterization has been performed by VASE on 12 samples<br />
realized in the same experimental con<strong>di</strong>tions; the optical<br />
model for experimental data fitting has been developed in our<br />
previous work. 10 We believe that this is the thickness of a<br />
monolayer of HFBs when self-assembled on hydrophobic<br />
silicon: this value is also consistent with a typical molecular<br />
size and comparable to atomic force microscopy<br />
measurements. 21 Accor<strong>di</strong>ng to the above described model, the<br />
washing step of the chip is strong enough to remove the<br />
aggregates deposited on the HFB monolayer that <strong>di</strong>rectly<br />
interacts with the hydrophobic silicon surface. This behaviour<br />
points out the stronger interactions between the silicon surface<br />
and the HFB monolayer with respect to those between the<br />
HFB aggregates and the HFB monolayer.<br />
In a recent article, we have already demonstrated that the<br />
nanometric HFB biofilm uniformly covers the silicon surface:<br />
10<br />
on exposure to the potassium hydroxide (KOH), which<br />
anisotropically <strong>di</strong>ssolves the silicon, a silicon wafer coated by<br />
HFB is completely shielded by the etching agent, at least for<br />
several minutes. The protection against KOH can be<br />
prolonged at any time by increasing the thickness of the HFB<br />
biofilm self-assembled on the silicon surface. This is possible<br />
since successive depositions of proteins form a biofilm having<br />
thickness proportional to the number of depositions, as it is<br />
shown in Figure 2: we have reported the thicknesses of the<br />
HFB films after each one of three consecutive depositions and<br />
plotted them as function of the deposition cycle.<br />
Figure 2. Growth of a thick HFB biofilm by repeated cycles<br />
of deposition on a hydrophobic silicon substrate<br />
Moreover, the presence of the biofilms does not alter the<br />
optical behaviour of the silicon surface as it is shown in<br />
40 Figure 3, where the transmission curves of two <strong>di</strong>fferent<br />
biofilms on silicon are reported. In both cases the selfassembled<br />
layer is transparent in a very large interval of<br />
wavelengths: the transmittance is still the 80 % at 290 nm.<br />
The improvement in the silicon wettability is also a key<br />
45 features in biodevices realization: the silicon surface, after<br />
standard cleaning process and washing in HF, is hydrophobic,<br />
this occurrence can constitute a severe limit in the fabrication<br />
of microchannels for microflui<strong>di</strong>cs applications. Since HFBs<br />
self-assemble in stable biofilm on a variety of <strong>di</strong>fferent<br />
surfaces, such as mica, Teflon, and polystyrene 6-7 due to the<br />
amphiphilic interaction with hydrophobic or hydrophilic<br />
surfaces, we can mo<strong>di</strong>fy the silicon behaviour with respect to<br />
the water interaction by covering its surface by HFBs.<br />
This journal is © The Royal Society of Chemistry [year] Journal Name, [year], [vol], 00–00 | 3<br />
50<br />
Figure 3. Figure caption. Transmission curves of two<br />
55 <strong>di</strong>fferent samples of HFB self-assembled biofilm on silicon.<br />
60<br />
65<br />
The changing in silicon wettability has been monitored by<br />
water contact angle measurements. In Figure 4 A and B are<br />
reported the WCA results in case of the bare silicon surface<br />
and after the deposition of the HFB biofilm. The dramatic<br />
increase in wettability of the silicon surface is well evident: in<br />
the first case, the WCA results in 90°±1°, so that the surface<br />
can be classified as hydrophobic, while after the HFB<br />
deposition the WCA falls down to 25°±2° so that the surface<br />
is clearly hydrophilic. The HFB mo<strong>di</strong>fied silicon surface is<br />
highly stable and the WCA value does not change after<br />
months, even if stored at room temperature without particular<br />
saving con<strong>di</strong>tion.<br />
70 Figure 4. Changing the wettability of c-Si: the hydrophobin<br />
nanolayer turns the hydrophobic surface of c-Si into a<br />
hydrophilic one. A: a water drop on the c-Si after washing in<br />
hydrofluori<strong>di</strong>c acid forms a contact angle of 90°±1°. B: a<br />
water drop after the hydrophobin deposition forms a contact<br />
75 angle of 25°±2°.<br />
80<br />
Transmittance (%)<br />
90<br />
85<br />
80<br />
75<br />
70<br />
22 nm biofilm<br />
64 nm biofilm<br />
400 600 800 1000 1200 1400<br />
Wavelength (nm)<br />
A B<br />
The HFB mo<strong>di</strong>fied silicon surface has been tested by BSA<br />
immobilization: solutions containing the rhodamine labelled<br />
protein, at <strong>di</strong>fferent concentration, between 3 and 12 μM, have<br />
been spotted on the HFB film. The labelled bioprobes have<br />
been spotted also on some surface which have not been treated<br />
with the hydrophobins, as a negative control on the possible
5<br />
10<br />
15<br />
20<br />
25<br />
30<br />
35<br />
aspecific bin<strong>di</strong>ng between the silicon and the probe. All the<br />
samples have been washed in deionized water to remove the<br />
excess of biological matter and observed by the fluorescence<br />
macroscopy system.<br />
A<br />
B<br />
Fluorescence intensity<br />
fluorescence intensity (a.u.)<br />
18<br />
16<br />
14<br />
12<br />
10<br />
8<br />
6<br />
4<br />
2<br />
0<br />
4,8<br />
4,0<br />
3,2<br />
2,4<br />
1,6<br />
0,8<br />
0,0<br />
0 5 10 15 20<br />
Distance (μm)<br />
[BSA]=300 fmol/cm 2<br />
0 2 4 6 8 10 12<br />
BSA concentration (μM)<br />
Figure 5. A: Fluorescence image of labelled BSA spotted on a<br />
silicon chip covered by HFBs. B: Negative control sample:<br />
BSA spotted <strong>di</strong>rectly on silicon. C: Intensity profile of<br />
fluorescencent BSA on HFB biofilm inside and outside the<br />
drop. D: The dose-response curve of fluorescence intensity as<br />
function of the BSA concentration.<br />
By illuminating the fluorescent BSA samples, we found that<br />
the fluorescence is brighter than the negative control as it can<br />
be seen in Figure 5A and 5B. The fluorescence signal is also<br />
quite homogeneous on the whole surface, as confirmed by the<br />
intensity profile on a treated surface reported in Figure 5C.<br />
We have qualitatively tested the strength of the affinity bond<br />
between the HFB and the BSA, by washing the chip in a<br />
<strong>di</strong>alysis membrane overnight in deionized water. Since the<br />
fluorescent intensities before and after the <strong>di</strong>alysis do not<br />
<strong>di</strong>ffer, we can conclude that the HFB-BSA system is very<br />
stable. We have also realized a dose-response curve of<br />
fluorescence intensity as function of BSA concentration,<br />
shown in Figure 5D, and we have estimated a saturation<br />
concentration equal to 3.12 μM correspon<strong>di</strong>ng to 300<br />
fmol/cm 2 .<br />
Protein immobilization on the HFB film has been also verified<br />
and analyzed using an enzyme, POXC laccase, produced by<br />
the same fungus P. ostreatus 16 , easily available in our<br />
laboratory and whose activity assay is simple and sensitive.<br />
The enzymatic assay on the immobilized enzyme has been<br />
performed <strong>di</strong>pping the chip into the buffer containing the<br />
substrate and following its absorbance during several minutes.<br />
We chose to use a pH 5 buffer and DMP as a substrate, a good<br />
settlement between stability and activity of the enzyme. A<br />
volume of 30 µl of the enzyme solution (about 700 U/ml)<br />
D<br />
C<br />
have been deposited on the chips (10mm x 10mm) and, after<br />
several washing, an activity of 0.1÷0.2 U has been determined<br />
on each chip (Figure 6), with an immobilization yield of<br />
0.5÷1%.<br />
4 | Journal Name, [year], [vol], 00–00 This journal is © The Royal Society of Chemistry [year]<br />
40<br />
45<br />
50<br />
55<br />
60<br />
Figure 6. Enzymatic activity determination of POXC laccase<br />
immobilized on a chip.<br />
This value is comparable to that one (7%) obtained in the<br />
optimized con<strong>di</strong>tions for laccase immobilization on<br />
EUPERGIT C 250L © . 22 Taking into account the specific<br />
activity of the free enzyme (430 Umg -1 ) and its molecular<br />
mass (59 kDa), 0.5µg of laccase correspon<strong>di</strong>ng to about 8<br />
pmol (5 x 10 12 molecules) have been immobilized on each<br />
chip. A reasonable evaluation of the surface occupied by a<br />
single protein molecule can be based on crystal structures of<br />
laccases 23-24 and homology among these proteins. This surface<br />
should be 28 x 10 -12 mm 2 , considering the protein as a sphere<br />
with ra<strong>di</strong>us of 3 x 10 -6 mm. On this basis the maximal number<br />
of laccase molecule on each chip should be 3 x 10 12 . These<br />
data can in<strong>di</strong>cate that the number of active immobilized<br />
laccase molecules on each chip is of the same order of<br />
magnitude than the expected maximum. Laccase assays have<br />
been repeated on the same chip after 24 and 48 hours in the<br />
same con<strong>di</strong>tions. Figure 7 shows that about one half of the<br />
activity was lost after one day, but no variation of the residual<br />
activity was observed after the second day.<br />
ΔA/min<br />
Ads (477 nm)<br />
0.32<br />
0.30<br />
0.28<br />
0.26<br />
0.24<br />
0.22<br />
0.20<br />
0.18<br />
0.16<br />
0.7<br />
0.6<br />
0.5<br />
0.4<br />
0.3<br />
0.2<br />
0.1<br />
0.0<br />
-0.1<br />
ΔA/min=0,317± 0,006<br />
0.0 0.5 1.0 1.5 2.0<br />
Time (min)<br />
0 10 20 30 40 50<br />
Time (h)<br />
Figure 7. Time resolved profile of laccase activity of the<br />
65 immobilized enzyme.
5<br />
10<br />
15<br />
20<br />
25<br />
Conclusions<br />
We have demonstrated the bio-mo<strong>di</strong>fication of the silicon<br />
surface by using the self-assembling features of the HFB, an<br />
amphiphilic protein purified by the fungus P. ostreatus. The<br />
nanometric biofilm of proteins, casted on the silicon surface<br />
by solution deposition, is very stable from the chemical point<br />
of view, since it is still present after strong denaturing<br />
washing in NaOH and SDS at 100 °C. The biofilm turns the<br />
hydrophobic silicon surface into a hydrophilic one, which can<br />
be very useful in microflui<strong>di</strong>c integrated application.<br />
Moreover, the biofilm can arrange both in a monolayer or in a<br />
multilayer stack of proteins depen<strong>di</strong>ng on deposition<br />
con<strong>di</strong>tions. The HFB monolayer acts as a bioactive substrate<br />
to bind other proteins, such as BSA and laccase. These results<br />
can be the starting point in the fabrication of a new generation<br />
of hybrid devices for proteomic applications.<br />
ACKNOWLEDGMENT. The authors gratefully acknowledge<br />
dr. N. Malagnino of ST. Microelectronics in Portici (NA),<br />
Italy, for water contact angle measurements, Prof. P. Arcari<br />
(Dept. of Biochemistry and Me<strong>di</strong>cal Biotechnologies,<br />
University of Naples “Federico II”) for fluorescent BSA, dr.<br />
F. Autore for the purification of POXC laccase and G.<br />
Armenante for technical support.<br />
Notes and references<br />
a<br />
Unit of Naples-Institute for Microelectronics and Microsystems,<br />
National Council of Research, Via P. Castellino 111, 80131, Naples, Italy<br />
Fax: +390816132598; Tel: +390816132375; E-mail:<br />
luca.destefano@na.imm.cnr.it<br />
b<br />
30 Dept. of Physical Sciences, University of Naples “Federico II”, Via<br />
Cintia 4, 80126, Naples, Italy.<br />
c<br />
Dept. of Organic Chemistry and Biochemistry, University of Naples<br />
“Federico II”, Via Cintia 4, 80126, Naples, Italy.<br />
d<br />
LYCRIL – INFM-CNR, Via P. Bucci, Cubo 33/B, 87036 Arcavacata <strong>di</strong><br />
35 Rende, Cosenza, Italy.<br />
40<br />
45<br />
50<br />
55<br />
60<br />
1. 1 Yam, C. M.; Mayeux, A.; Milenkovic, A.; Cai, C.<br />
Z. Langmuir 2002, 18, 10274-10278.<br />
2. W. Xing and J. Cheng , Eds., Frontiers in Biochip<br />
Technology, 2006, Springer.<br />
3. Cha T, Guo A, Zhu XY. Proteomics 2005, 5, 416-419.<br />
4. Angenendt P, Glokler J, Murphy D, Lehrach H, Cahill DJ.<br />
Anal. Biochem. 2002, 309, 253-260.<br />
5. Angenendt P, Glokler J, Sobek J, Lehrach H, Cahill DJ. J.<br />
Chromatogr. A 2003, 1009, 97-104.<br />
6. Qin, M.; Wang, L. K.; Feng, X. Z.; Yang, Y. L.; Wang, R.;<br />
Wang, C.; Yu, L.; Shao, B.; Qiao, M.Q. Langmuir 2007,<br />
23, 4465-4471.<br />
7. Wang, R.; Yang, Y. L.; Qin, M.; Wang, L. K.; .Yu, L;<br />
Shao, B.; Qiao, M. Q.; Wang, C.; Feng, X.Z. Chem. Mater.<br />
2007, 19, 3227-3231.<br />
8. Gutmann O, Kuehlewein R, Reinbold S, Niekrawietz R,<br />
Steinert CP, de Heij B, Zengerle R, Daub M. Lab Chip<br />
2005, 5, 675-681.<br />
9. Buriak, J. M. Advanced Materials, 1999, 11, 265-268.<br />
10. L. De Stefano, I. Rea, P. Giar<strong>di</strong>na, A. Armenante, M.<br />
Giocondo, I. Ren<strong>di</strong>na. Langmuir 2007, 23, 7920,.<br />
11. Hektor H.J. and Scholtmeijer K. Current Opinion in<br />
Biotechnology 2005, 16, 434-439.<br />
12. Sunde M., Kwan A.H.Y., Templeton M.D., Beever R.E.,<br />
Mackay JP. Micron 2007<br />
doi:10.1016/j.micron.2007.08.003.<br />
13. Linder MB, Szilvay GR, Nakari-Setälä T, Penttilä ME.<br />
FEMS Microbiol Rev. 2005, 29(5), 877-96.<br />
14. Lumsdon SO, Green J, Stieglitz B. Colloids Surf B<br />
This journal is © The Royal Society of Chemistry [year] Journal Name, [year], [vol], 00–00 | 5<br />
65<br />
70<br />
75<br />
80<br />
85<br />
90<br />
Biointerfaces 2005, 44(4), 172-8.<br />
15. Kershaw M. J., Thornton C.R., Gavin E. Wakley and<br />
Talbot N. Molecular Microbiology 2005, 56(1), 117-125.<br />
16. Palmieri G., Giar<strong>di</strong>na P., Marzullo L., Desiderio B., Nitti G.<br />
and Sannia G. Appl. Microbiol. Biotechnol., 1993, 39, 632-<br />
636<br />
17. Corvis Y, Walcarius A, Rink R, Mrabet NT, Rogalska E.<br />
Anal Chem. 2005, 77(6), 1622-30.<br />
18. Kern W 1993, Ed., Handbook of Semiconductor Cleaning<br />
Technology, Noyes Publishing: Park Ridge, NJ, Ch 1.<br />
19. Kisko K, Szilvay GR, Vainio U, Linder MB, Serimaa R.<br />
Biophys J. 2008, 94(1), 198-206.<br />
20. Kwan AH, Winefield RD, Sunde M, Matthews JM,<br />
Haverkamp RG, Templeton MD, Mackay JP. Proc Nat<br />
Acad Sci USA 2006, 103(10), 3621-6.<br />
21. Yu L, Zhang B, Szilvay GR, Sun R, Jänis J, Wang Z, Feng<br />
S, Xu H, Linder MB, Qiao M. Microbiology, 2008,<br />
1541677-85.<br />
22. Russo M.E., Giar<strong>di</strong>na P., Marzocchella A., Salatino P.,<br />
Sannia G. Enz. Microbiol. Biotechnol., 2008, 42, 521–530<br />
23. Piontek K, Antorini M, Choinowski T. J Biol Chem. 2002;<br />
77, 37663-9.<br />
24. Ferraroni M, Myasoedova NM, Schmatchenko V,<br />
Leontievsky AA, Golovleva LA, Scozzafava A, Briganti F.<br />
BMC Struct Biol. 2007, 7, 60.
Chapter 4<br />
…towards the Lab-on-Chip
Introduction to Lab-on-chip technologies<br />
A heart attack is a typical situation for which a fast <strong>di</strong>agnostic is of paramount importance.<br />
In ad<strong>di</strong>tion to an electrocar<strong>di</strong>ogram, a blood test is often necessary, that is conventionally<br />
done in a laboratory. With the portable system marked by the Biosite company, this<br />
<strong>di</strong>agnostic can be made in only 10 min, <strong>di</strong>rectly with the patience [1]. The drop of blood is<br />
simply deposited in a reservoir in the system that then performs the entire test: minute<br />
quantities of blood are <strong>di</strong>splaced and injected into several tiny chambers in which <strong>di</strong>fferent<br />
reactions take place.<br />
This example clearly illustrates the principle of a Lab-on-Chip: integrating all the functions<br />
of a human-scale test laboratory inclu<strong>di</strong>ng transferring samples, drawing off a precise<br />
volume of a chemical product, mixing with reagents, heating, titration, etc, on a system of a<br />
few square centimeters.<br />
A LOC includes many functions depen<strong>di</strong>ng on the application such as pumps, valves,<br />
sensors, electronics, etc. Therefore, it can be considered as a complex microsystem<br />
inclu<strong>di</strong>ng mechanical, electronic, fluid functions, etc.. It also needs a network of<br />
microchannels.<br />
Since the ‘planar’ technology was introduced into microelectronics at the end of the 1950s,<br />
micromanufacturing techniques have never stopped improving, continuously exten<strong>di</strong>ng the<br />
limits of miniaturization. Silicon was initially chosen for its semi-conducting properties and<br />
the excellent quality of its oxide, and was also used to develop mechanical functions<br />
inclu<strong>di</strong>ng resonant beams, micro-motors, etc [2-4].<br />
The concept of microsystems such as (MEMS—Micro-Electro-Mechanical Systems)<br />
emerged during the 1980s at the University of Berkeley as the integration of the <strong>di</strong>fferent<br />
functions of a complete system on a single chip [5]. The first example of a microsystem<br />
marketed by Analog Devices in 1993 was the ADXL50 accelerometer that consisted of a<br />
capacitive sensor accompanied by its associated electronics monolithically integrated onto<br />
a 10 mm 2 substrate [6]. MEMS technology takes advantages from microelectronic<br />
production tools that allow for the manufacturing of thousands of miniaturized objects in<br />
parallel and thus reduces drastically fabrication costs.<br />
Several examples of microsystems are now used in daily life: accelerometers that are<br />
present in modern cars (airbag trigger system) [7,8], and also in Nike shoes (integrated<br />
pedometer), Sony Playstation games consoles (measurement of the inclination of the<br />
joystick) and GPS systems [9,10], inkjet print heads [11-14], arrays of micromirrors in<br />
video projectors [15,16]. Many sectors are now concerned such as optical and ra<strong>di</strong>o<br />
frequency telecommunications (switches, variable capacitances and inductances) [17-19].<br />
Finally, the interest in microsystems in the chemistry and biome<strong>di</strong>cal fields has never<br />
ceased increasing, particularly since the development of the miniaturized total chemical<br />
analysis system (μTAS) concept by A Manz at the beginning of the 1990s and then since<br />
the development of LOC [20].<br />
Over the last 10 years, this popularity has been a driving force for the development of new<br />
types of microsystems combining electrical and mechanical functions with microflui<strong>di</strong>c<br />
functions (microwells, microchannels, valves, pumps). It also encouraged the introduction<br />
of new processes and materials in microtechnologies. The glass technology widely used at<br />
first, combining wet etching and thermal bon<strong>di</strong>ng, is limited by <strong>di</strong>fficulties in terms of<br />
integration and aspect ratio, while etching of silicon is much more developed.<br />
Silicon and glass technologies<br />
• Bulk micromachining<br />
Bulk micromachining refers to the situation in which patterns are defined in the substrate<br />
itself. The example of wet etching of glass is shown in figure 1.<br />
90
Figure 1: (1) An example of bulk micromachining: wet etching of glass; (2) microchannels etched by<br />
wet etching in glass [IMT Neuchˆatel] [21] (© 2003 IEEE).<br />
It clearly illustrates the generic approach to bulk machining. After the substrate has been<br />
cleaned, the material that will be used for protection during the etching step is deposited<br />
(a). A photoresist is coated using a spin coater (b). This resist is exposed to ultraviolet<br />
ra<strong>di</strong>ation through a mask with transparent and opaque areas (c). The resist is then<br />
developed (d) and exposed areas are eliminated (positive resist, as illustrated in figure 1)<br />
or remain (negative resist), depen<strong>di</strong>ng on its polarity. The protective material is etched (e)<br />
and the resist is removed from the substrate (f ). The next step is the glass etching step<br />
(g). Finally, the mask material is removed from the substrate (h).<br />
• Wet etching<br />
The etching step is fundamental in bulk micromachining [22, 23]. Etching as presented in<br />
figure 2 is isotropic, i.e. the etching rate is the same in all <strong>di</strong>rections. Typical examples are<br />
wet etching of silicon in a mixture of hydrofluoric acid, nitric acid and ethanol [24], or wet<br />
etching of glass in hydrofluoric acid [25, 26]. Isotropic wet etching has several<br />
<strong>di</strong>sadvantages, inclu<strong>di</strong>ng <strong>di</strong>fficulty in controlling the profile due to the strong influence of<br />
stirring and isotropy that severely limits the possible resolution and depth, as illustrated<br />
in figure 2.<br />
Figure 2: Under-etching phenomenon during isotropic wet etching<br />
Anisotropic wet etching of silicon [27-30] in solutions such as potassium hydroxide (KOH)<br />
or tetramethyl ammonium hydroxide (TMAH) [31-33] can be used to create patterns<br />
defined by crystallographic planes of silicon as illustrated in figure 3. This process,<br />
together with etching-stop techniques (geometric, electrochemical, implantation of boron)<br />
[34-36] makes it easy for manufacturing very thin membranes or suspended structures.<br />
However, some simple shapes such as a circular hole cannot be made, the etching profile<br />
limits the integration density and the chemicals used are relatively aggressive.<br />
Figure 3: (a) Different anisotropic wet etching profiles in single-crystalline silicon; (b) examples of microwells<br />
made by anisotropic wet etching [GESIM] [37], © 1999, with permission from Elsevier.<br />
91
• Glass etching<br />
The technology for the microfabrication of glass is gaining importance because more and<br />
more glass substrates are currently being used to fabricate MEMS devices. Glass has<br />
many advantages as a material for MEMS applications, such as good mechanical<br />
properties, good optical properties, high electrical insulation, and it can be easily bonded to<br />
silicon substrates at temperatures lower than for fusion bon<strong>di</strong>ng.<br />
Glass is particularly useful for fabrication of microflui<strong>di</strong>c devices used for bioanalysis due<br />
to its high chemical resistance. Microflui<strong>di</strong>c <strong>di</strong>spensing and controlling devices, such as<br />
micro pumps, micro pressure/flow sensors, monolithic membrane valve/<strong>di</strong>aphragm pumps,<br />
have been developed with integrated glass components. DNA related microflui<strong>di</strong>c devices,<br />
such as micro flow cells for DNA single molecule handling, micro injectors for DNA mass<br />
spectrometry, and micro polymerase chain reaction (PCR) devices for DNA amplification,<br />
have also been presented [38]. Being transparent under a wide wavelength range ( 300–<br />
2000 nm for Pyrex), glass is a prime can<strong>di</strong>date for incorporation into micro bioanalytical<br />
devices where optical detection of the bioanalytes is used, for example, micro capillary<br />
electrophoresis (μCE) devices [39].<br />
Extensive investigation of glass microfabrication technologies is paving the way for<br />
increasing utilization of glass substrates in MEMS. These technologies include wet<br />
chemical etching using various concentrations of hydrofluoric acid (HF) [40, 41], dry etch<br />
techniques like deep reactive ion etching (DRIE) using SF6 plasma [42], laser<br />
micromachining [43] and ultrasonic drilling [44]. Laser drilling and ultrasonic drilling are<br />
suitable for fabricating the small numbers of through holes in the glass.<br />
Wet glass etching with various concentrations of HF is the most widely used method<br />
because the etching rate is fast and a large quantity of glass wafers can be processed<br />
simultaneously. Masks for glass etching in HF are normally photoresist [40] and Cr/Au<br />
photoresist combinations [45]. However, the formation of pinholes through defects within<br />
the metal mask is a notorious problem. This becomes especially severe when deep<br />
etching is required. A second problem found for HF etching is rapid undercutting of the Cr<br />
mask, which leads to a more rapid lateral etching of the glass than vertical etching. This<br />
leads to a poor aspect ratio generally smaller than one for etched cavities. Although the<br />
application of metal masks, for deep etching with HF, has the problem of pinholes and<br />
large lateral undercutting as mentioned above, the approach is simple and compatible with<br />
standard microfabrication facilities.<br />
• Assembly<br />
An assembly step between two machined substrates is usually necessary to seal<br />
structures. Several types of bon<strong>di</strong>ng techniques have been developed for <strong>di</strong>fferent<br />
applications. Wafer bon<strong>di</strong>ng is a cost-effective method for zero-level MEMS packaging and<br />
has increasingly become a key technology for materials integration in various areas of<br />
MEMS, microelectronics and optoelectronics. Different wafer bon<strong>di</strong>ng approaches are<br />
currently used in the MEMS industry: fusion, adhesive, eutectic, ano<strong>di</strong>c, solder bon<strong>di</strong>ng<br />
[46, 47].<br />
These bon<strong>di</strong>ng techniques are summarized in figure 4. One of the main concerns is the<br />
thermal stress, mainly in the case of heterogeneous assembly, i.e. assembly of materials<br />
with <strong>di</strong>fferent thermal expansion coefficients. The required level of cleanness may also be<br />
an obstacle in a process such as thermal bon<strong>di</strong>ng. Adhesive bon<strong>di</strong>ng is a low-temperature<br />
process and does not require complex cleaning procedures but introduces a<br />
supplementary material into the structure which may be undesirable in microflui<strong>di</strong>cs in<br />
which the control of surface properties is critical. It is often estimated that assembly is<br />
frequently the most expensive step in microsystem processes.<br />
92
Figure 4: Different types of bon<strong>di</strong>ng between substrates<br />
As fusion bon<strong>di</strong>ng processes require a high temperature annealing which is not always<br />
suitable for the devices with aluminum or copper integrated circuits; adhesive bon<strong>di</strong>ng is<br />
non-hermetic, more and more interests are focused on low temperature bon<strong>di</strong>ng<br />
(maximum process temperature below 300°), which can not only reduce process cost and<br />
time, but also minimize bon<strong>di</strong>ng-induced stress and warpage after cooling. Moreover, the<br />
wafers with large <strong>di</strong>fference in coefficient of thermal expansion can also be bonded. The<br />
hybrid integration approach appears better adapted to a lab-on-chip integration. The<br />
challenge is then to couple the <strong>di</strong>fferent elements, for example a sensor made using the<br />
silicon technology and a microflui<strong>di</strong>c network replicated in a thermoplastic, with good<br />
alignment and without introducing ad<strong>di</strong>tional dead volumes, and obviously at the lowest<br />
cost.<br />
Ano<strong>di</strong>c bon<strong>di</strong>ng is one of the most used wafer level packaging procedures. In ano<strong>di</strong>c<br />
bon<strong>di</strong>ng, the substrates are heated to a typical temperature above 400° and a typical<br />
voltage above 600V is applied to the wafer pair to be bonded, electrostatic force and the<br />
migration of ions lead to an irreversible chemical bond at the boundary layer between the<br />
in<strong>di</strong>vidual wafers. Conventionally, this is done in the furnace or on the hotplate, which<br />
require long ramp times, consume large amounts of power, and have significant<br />
manufacturing footprints. Currently, many efforts are focusing on improving ano<strong>di</strong>c<br />
bon<strong>di</strong>ng quality [48, 49].<br />
However, the bond quality is reduced with decrease in the bon<strong>di</strong>ng temperature, low<br />
bon<strong>di</strong>ng strength will result in the bubbles or cavities at the interface, only few literature<br />
reports the success in low temperature ano<strong>di</strong>c bon<strong>di</strong>ng with high bond strength and<br />
bubble-free interface [49].<br />
The concept of μTAS, later extended to the lab-on-a-chip concept, was introduced in 1990<br />
to mitigate the deficiencies of chemical sensors, particularly in terms of selectivity and life<br />
time [20]. Integration onto a chip can also very strongly reduce the consumption of<br />
chemicals that may be rare, expensive and/or polluting. Similarly, it reduces the production<br />
of waste. Automation of procedures illustrated by the DNA analysis station marketed by<br />
the Caliper Company, increases analysis rates and reduces the risk of error due to the<br />
human factor [50].<br />
Furthermore, these products enjoy the advantages of prices for mass production<br />
processes. Miniaturization also provides a means of making these systems portable.<br />
Cost and size reductions are not the only advantages; performance gain is also important.<br />
For example, an immunoassay was carried out in less than 25s while more than 10 min<br />
are necessary under classical con<strong>di</strong>tions [51]. Therefore miniaturization of volumes<br />
provides a means of reducing the time for the operation.<br />
93
Many lab-on-chip applications have been developed in genomics, under the emphasis of<br />
the human genome project and sequencing needs [52]. Moreover genetic analyses<br />
became increasingly routine operations, efforts have been made and are still being made<br />
on analysis of proteins. Its tra<strong>di</strong>tional approach includes extraction from cells, separation<br />
by electrophoresis on gel, <strong>di</strong>gestion and analysis by a mass spectrometer. These<br />
operations are slow and time consuming. A system was made to perform the <strong>di</strong>gestion,<br />
separation and analysis steps [53]. The operation that normally takes hours is done in a<br />
few minutes.<br />
References<br />
1. Biosite, Inc http://www.biosite.com/<br />
2. Nathanson H C, Newell W E, Wickstrom R A and Davis J R 1967 The resonant gate transistor IEEE Trans.<br />
Electron Devices 14 117<br />
3. Jolly R D and Muller R S 1980 Miniature cantilever beams fabricated by anisotropic etching of silicon J.<br />
Electrochem.Soc. 127 2750<br />
4. Tai Y C and Muller R S 1989 IC-processed electrostatic synchronous micromotors Sensors Actuators 20<br />
49<br />
5. Chen P L, Muller R S and Andrews A P 1984 Integrated silicon PI-FET accelerometer with proof mass<br />
Sensors Actuators 5 119<br />
6. Chau K H L, Lewis S R, Zhao Y, Howe R T, Bart S F and Marcheselli R G 1996 An integrated forcebalanced<br />
capacitive accelerometer for low-g applications Sensors Actuators A 54 472<br />
7. Zimmerman L, Ebershol J P, Lehung F, Berry J P, Bailleu F, Rey P, Diem B, Renart S and Caillat P 1995<br />
Airbag application—a microsystem inclu<strong>di</strong>ng a silicon capacitive accelerometer, cmos switched-capacitor<br />
electronics and true self-test capability Sensors Actuators A 46 190–5<br />
8. Kim K H, Ko J S, Cho Y H, Lee K, Kwak B M and Park K 1995 A skew-symmetric cantilever accelerometer<br />
for automotive airbag applications Sensors Actuators A 50 121–6<br />
9. Hanse J G 2004 Honeywell MEMS inertial technology & product status Proc. IEEE Position Location and<br />
Navigation Symp. (Monterey, CA) pp 43–48<br />
10. Fujino Y, Kitagawa K, Furukawa T and Ishii H 2005 Development of vehicle intelligent monitoring system<br />
(VIMS) smart structures and materials 2005: sensors and smart structures technologies for civil,<br />
mechanical, and aerospace Proc. SPIE 5765 148–57<br />
11. Allen R R, Meyer J D and Knight W R 1985 Thermodynamics and hydrodynamics of thermal ink jets<br />
Hewlett-Packard J. (May) 21–7<br />
12. O’Horo M, O’Neill J, Peters E and Vanderbroek S 1996 Micro electro mechanical system technology for<br />
commercial thermal inkjet document output products Proc. Eurosensors X (Leuven, Belgium) pp 431–5<br />
13. Lee J-D, Yoon J-B, Kim J-K, Chung H-J, Lee C-S, Lee H-D, Lee H-J, Kim C-K and Han C-H 1999 A thermal<br />
inkjet printhead with monolithically fabricated nozzle plate and self-aligned ink feed hole J.<br />
Microelectromech. Syst. 8 229–35<br />
14. Krause P, Obermeier E and Whel W 1996 A micromachined single chip inkjet printhead Sensors Actuators<br />
A 21–23 198–202<br />
15. Van Kessel P F, Hornbeck L J, Meier R E and Douglas Hornbeck M R 1998 Proc. IEEE 86 1687–704<br />
16. Yoder L 1996 Texas Instruments-DLPTM product literature<br />
17. Rebeiz G M 2006 RF MEMS for wireless-bands tunable networks Topical Meeting on Silicon Monolithic<br />
Integrated Circuits in RF Systems, Dig. of Papers pp 23–23<br />
18. Dubuc D et al 2004 3D MEMS circuits integration for RF and millimeter wave communications Proc. 2004<br />
Int. Semiconductor Conf. vols 1 and 2 pp 119–28<br />
19. Tentzeris M and Laskar J 2003 RF system-on-package (SOP) development for compact low cost wireless<br />
front-end systems Adv. Electron. Packag. 2 19–24<br />
20. Manz A, Graber N and Widmer H M 1990 Miniaturized total chemical analysis systems: a novel concept for<br />
chemical sensing Sensors Actuators B 1 244<br />
21. Verpoorte E and De Rooij N F 2003 Microflui<strong>di</strong>cs meets MEMS Proc. IEEE 91 930<br />
22. Petersen K E 1998 Silicon as a mechanical material Proc.IEEE 70 420–57<br />
23. Kendall D L and Shoultz R A 1997 Wet chemical etching of silicon and SiO2, and ten challenges for<br />
micromachiners Handbook of Microlithography, Micromachining, and Microfabrication vol 2 ed P Rai-<br />
Choudury (Bellingham,WA: SPIE Optical Engineering Press)<br />
24. Gardeniers J G E, Tjerkstra R W and Van den Berg A 2000 Fabrication and application of silicon-based<br />
microchannels Microreaction Technologies: Industrial Prospects pp 36–44<br />
25. Iliescu C and Tay F 2001 Wet etching of glass CAS 2005: Int.Semiconductor Conf. vols 1 and 2 pp 35–44<br />
26. Ilie M, Foglietti V and Cianci E 2002 Deep wet etching of borosilicate glass for microflui<strong>di</strong>c devices Sensors<br />
Microsyst., Proc. 6th Italian Conf. pp 55–9<br />
27. Seidel H, Csepregi L, Heuberger A and Baumgartel H 1990 Anisotropic etching of crystalline silicon in<br />
alkaline solutions J. Electrochem. Soc. 137 3612<br />
28. Kendall D L 1975 On etching very narrow grooves in silicon Appl. Phys. Lett. 26 195–8<br />
29. Kendall D L 1990 A new theory for anisotropic etching of Si and some underdeveloped chemical<br />
micromachining concepts J. Vac. Sci. Technol. A 8 3598<br />
94
30. Camon H and Mokta<strong>di</strong>r Z 1997 Simulation of silicon etching with HOH Microelectron. J. 28 142–5<br />
31. Tabata O, Asahi R, Funabashi H, Shimaoka K and Sugiyama S 1992 Anisotropic etching of silicon in TMAH<br />
solutions Sensors Actuators A 34 51–7<br />
32. Sato K, Shikida M, Yamashiro T, Asaumi K, Iriye Y and Yamamoto M 2000 Anisotropic etching rates of<br />
single-crystal silicon for TMAH water solution as a function of crystalline orientation Sensors Actuators A 73<br />
131–7<br />
33. Biswas K, Bas S, Maurya D K, Kal S and Lahiri S K 2006 Bulk micromachining of silicon in TMAH-based<br />
etchants for aluminium passivation and smooth surface Microelectron. J. 37 321–7<br />
34. French P J, Nagao M and Esashi M 1996 Electrochemical etch-stop in TMAH without externally applied<br />
bias Sensors Actuators A 56 279–80<br />
35. Wallman L, Bengtsson J, Danielsen N and Laurell T 2002 Electrochemical etch-stop technique for silicon<br />
membranes with p- and n-type regions and its application to neural sieve electrodes J. Micromech.<br />
Microeng. 12 265–70<br />
36. Gianchandani Y B and Najafi K 1992 A bulk silicon <strong>di</strong>ssolved wafer process for microelectromechanical<br />
devices IEEE J. Microelectromech. Syst. 1 77–85<br />
37. Richter K, Orfert M, Howitz S and Thierbach S 1999 Deep plasma silicon etch for microflui<strong>di</strong>c applications<br />
Surf. Coat. Technol. 461 116–9.<br />
38. Yokokawa R, Yumi Y, Takeuchi S, Kon T, Fujita H (2006) Nanotechnology 17:289–294<br />
39. Berthold, A and Laugere, F and Schellevis, H and de Boer, CR and Laros, M and Guijt, RM and Sarro, PM<br />
and Vellekoop, MJ (2002) Electrophoresis, 23 (20). pp. 3511-3519.<br />
40. C. Lin, G. Lee, Y. Lin, G. Chang, A fast prototyping process for fabrication of microflui<strong>di</strong>c systems on sodalime<br />
glass, J. Micromech. Microeng. 11 (2001) 726–732<br />
41. T. Corman, P. Enoksson, G. Stemme, Deep wet etching of borosilicate glass using an ano<strong>di</strong>cally bonded<br />
silicon substrate as mask, J.Micromech. Microeng. 8 (1998) 84–87.<br />
42. X. Li, T. Abe, M. Esashi, Deep reactive ion etching of Pyrex glass using SF6 plasma, Sens. Actuatators A<br />
87 (2001) 139–145.<br />
43. T. Abe, X. Li, M. Esashi, Endpoint detectable plating through femtosecond laser drilled glass wafers for<br />
electrical interconnctions, Sens. Actuatators A 108 (2003) 234–238.<br />
44. J.A. Plaza, M.J. Lopez, A. Moreno, M. Duch, C. Cane, Definition of high aspect ratio glass columns, Sens.<br />
Actuatators A 105 (2003) 305–310.<br />
45. I. Schneega_, R. Brautigam, J.M. Kohler, Miniaturized flow-through PCR with <strong>di</strong>fferent template types in a<br />
silicon thermocycler, Lab Chip 1 (2001) 42–49.<br />
46. Martin A. Schmidt 1998 Procee<strong>di</strong>ngs of the IEEE 86 1575<br />
47. Wen H. Ko 1995 Materials chemistry and physics 42 169<br />
48. T. Rogers and J. Kowal 1995 Sensors Actuators A 46–47 113<br />
49. J.S. Go, Y.H. Cho 1995 Sens. Actuators A 73 52<br />
50. J. Wei, H. Xie, M. L. Nai, C. K. Wong and L. C. Lee 2003 J. Micromech. Microeng. 13 217<br />
51. Caliper Life Sciences http://www.caliperls.com/<br />
52. Hatch A, Kamholz A E, Hawkins K R, Munson M S, Schilling E A, Weigl B H and Yager P 2001 A rapid<br />
<strong>di</strong>ffusion immunoassay in a T-sensor Nature Biotechnol. 19 46.<br />
53. Khandurina J and Guttman A 2002 Bioanalysis in microflui<strong>di</strong>c devices J. Chromatogr. A 943 159.
Microsystems Based on Porous Silicon-Glass Ano<strong>di</strong>c Bon<strong>di</strong>ng<br />
technologies<br />
This section is aimed to the demonstration of the compatibility of porous silicon and ano<strong>di</strong>c<br />
bon<strong>di</strong>ng technologies for the realization of sensing micro-components in lab-on-chip<br />
applications. In particular, it was developed and tested an innovative ano<strong>di</strong>c bon<strong>di</strong>ng<br />
process: voltage, time and temperature have been properly matched in order not to affect<br />
the properties of the transducing material, which can be strongly mo<strong>di</strong>fied by thermal<br />
treatments through oxidation.
Sensors 2006, 6, 680-687<br />
Full Research Paper<br />
sensors<br />
ISSN 1424-8220<br />
© 2006 by MDPI<br />
http://www.mdpi.org/sensors<br />
A Microsystem Based on Porous Silicon-Glass Ano<strong>di</strong>c Bon<strong>di</strong>ng<br />
for Gas and Liquid Optical Sensing<br />
Luca De Stefano 1,* , Krzysztof Malecki 1 , Francesco G. Della Corte 2 , Luigi Moretti 2 ,<br />
Ilaria Rea 1 , Lucia Rotiroti 1 and Ivo Ren<strong>di</strong>na 1<br />
1 Institute for Microelectronics and Microsystems, National Council of Research, Via P. Castellino<br />
111, 80131 Naples, Italy (E-mail for Ilaria Rea: ilaria.rea@na.imm.cnr.it)<br />
2 DIMET “Me<strong>di</strong>terranea” University of Reggio Calabria, Località Feo <strong>di</strong> Vito, 89060 Reggio Calabria,<br />
Italy<br />
* Author to whom correspondence should be addressed. E-mail: luca.destefano@na.imm.cnr.it<br />
Received: 15 March 2006 / Accepted: 22 June 2006 / Published: 23 June 2006<br />
Abstract: We have recently presented an integrated silicon-glass opto-chemical sensor for<br />
lab-on-chip applications, based on porous silicon and ano<strong>di</strong>c bon<strong>di</strong>ng technologies. In this<br />
work, we have optically characterized the sensor response on exposure to vapors of several<br />
organic compounds by means of reflectivity measurements. The interaction between the<br />
porous silicon, which acts as transducer layer, and the organic vapors fluxed into the glass<br />
sealed microchamber, is preserved by the fabrication process, resulting in optical path<br />
increase, due to the capillary condensation of the vapors into the pores. Using the<br />
Bruggemann theory, we have calculated the filled pores volume for each substance. The<br />
sensor dynamic has been described by time-resolved measurements: due to the analysis<br />
chamber miniaturization, the response time is only of 2 s. All these results have been<br />
compared with data acquired on the same PSi structure before the ano<strong>di</strong>c bon<strong>di</strong>ng process.<br />
Keywords: Porous Silicon, Ano<strong>di</strong>c Bon<strong>di</strong>ng, Microchamber, Lab-on-Chip.<br />
Introduction<br />
The physical and structural properties of porous silicon (PSi), first of all its high surface area<br />
matrix, have led many scientific researchers to investigate this material, using it as a transducer in<br />
sensing systems. In particular, high sensitivity results have been obtained by monitoring changes in<br />
optical properties, such as photoluminescence [1], reflectivity [2-3] and ellipsometry [4]. Both<br />
monolayers, which act as Fabry-Perot interferometers, and multilayer devices, like Bragg filters and
Sensors 2006, 6 681<br />
optical microcavities, can be easily fabricated and exploited in optical sensing experiments. The<br />
integration of sensitive elements into complex microsystems, i.e. lab on chip or micro-total-analysis<br />
systems, is of straightforward interest, especially for the miniaturisation of each component. This step<br />
is never a trivial one: the fabrication process often requires high temperatures and mechanical stresses<br />
which can damage or even destroy the bio- or chemo- sensitive transducer structures.<br />
The ano<strong>di</strong>c bon<strong>di</strong>ng (AB) is a standard IC fabrication technique which is widely used in<br />
microflui<strong>di</strong>c due to a wide spectrum of advantages among which the hermetic sealing [5-6]. Due to the<br />
good bon<strong>di</strong>ng quality, glass transparency, technological cleanness and high passivity to most of<br />
chemicals and biological substances, AB is also commonly exploited in the fabrication of lab-on-chip<br />
devices. The compatibility of this primary integration technique with highly deman<strong>di</strong>ng sensing microcomponents<br />
plays a basic role in view of the realization of lab-on-chip devices. We have recently<br />
reported on the optimisation of the standard silicon-glass AB parameters, searching for low<br />
temperature, low voltage and short time, taking into account the electrode type and thickness of glass<br />
wafers [7].<br />
In the present work, the above mentioned results have been exploited and combined with new<br />
bon<strong>di</strong>ng fin<strong>di</strong>ngs, to ensure AB compatibility with the features of porous silicon (PSi) transducer. We<br />
have stu<strong>di</strong>ed the static and dynamic sensor response on exposure to vapours of several organic<br />
compounds. The results have been compared with those obtained for the same PSi layer before the<br />
integration process.<br />
Theory<br />
Even if we have already realised µ-chambers with porous silicon multilayer optical structures, such<br />
as Bragg filters and optical microcavities which are more performing in terms of sensitivity and<br />
resolution, in this first optical characterisation, we have followed the approach of Sailor et al. [2]<br />
choosing the simple case of a single layer of porous silicon as our transducer device. Figure 1 shows a<br />
typical reflectivity spectrum from a PSi layer under white light illumination.<br />
Intensity (a. u.)<br />
1.8<br />
1.6<br />
1.4<br />
1.2<br />
1.0<br />
0.8<br />
0.6<br />
0.4<br />
Peak Order m<br />
14<br />
12<br />
10<br />
8<br />
6<br />
4<br />
2<br />
0<br />
11000 12000 13000 14000 15000 16000 17000<br />
Wavenumber (cm -1 )<br />
600 650 700 750 800 850 900<br />
Wavelength (nm)<br />
Figure 1. Reflectivity spectrum of PSi layer. In the inset the m-order peak are plotted as function of<br />
the wavenumber.
Sensors 2006, 6 682<br />
From the optical point of view, this structure is an optical Fabry-Perot interferometer, so that the<br />
maxima in the reflectivity spectrum appear at wavelengths λm which satisfy:<br />
m = 2 nd / λ m<br />
(1)<br />
where m is an integer, d is the film thickness and n is the average refractive index of the layer [8-9].<br />
If we assume that the refractive index is independent on the wavelength over the considered range, the<br />
maxima are equally spaced in the wavenumber (1/λm). When the maximum order m is plotted as a<br />
function of the wavenumber, each point lies on a straight line which slope is the optical path, as it is<br />
shown in the inset of Fig. 1. When the PSi layer is exposed to vapors, or <strong>di</strong>p in the liquid phase of the<br />
same substance, the substitution of air in the pores by its molecules causes a fringes shift in<br />
wavelength, which corresponds to a change in the optical path nd. Since the thickness d is fixed by the<br />
physical <strong>di</strong>mension of the PSi matrix, the variation is clearly due to changes in the average refractive<br />
index. In the case of vapors and gases, the filling liquid phase is due to the capillary condensation<br />
phenomenon, which is regulated from the Kelvin equation:<br />
K BTρ ln( psat<br />
/ p)<br />
= 2γ<br />
cos / R<br />
l lg θ<br />
(2)<br />
where ρl is the density of the liquid phase, γlg is the liquid-gas surface tension coefficient at room<br />
temperature T, R is the ra<strong>di</strong>us of the pores, p/psat is the relative vapor pressure into the pores, and θ is<br />
the contact angle [10].<br />
A quantitative model taking into account the optical path increase can be realized by applying the<br />
Bruggemann effective me<strong>di</strong>um approximation theory. The relation between the volume fraction of<br />
each me<strong>di</strong>um and its <strong>di</strong>electric constant can be written as:<br />
⎛ εSi<br />
−εp<br />
⎞ ⎛ εair<br />
−εp<br />
⎞ ⎛ εch<br />
−εp<br />
⎞<br />
( 1 − p)<br />
⎜ ⎟+<br />
( p−V)<br />
⎜ ⎟+<br />
V⎜<br />
⎟=<br />
0<br />
⎜ 2 ⎟ ⎜ 2 ⎟ ⎜ 2 ⎟<br />
⎝εSi<br />
+ εp<br />
⎠ ⎝εair<br />
+ εp<br />
⎠ ⎝εch<br />
+ εp<br />
⎠<br />
where p, V, εsi, εair, εch and εp are the layer porosity, the layer liquid fraction (LLF), the <strong>di</strong>electric<br />
constants of silicon, air, chemical substance and porous silicon, respectively [11].<br />
Experimental Section<br />
We have designed and fabricated a sensor device which can be considered the basic element of a<br />
lab-on-chip, just integrating PSi, as the transducer material, and a glass slide, which ensures sealing<br />
and the interconnections for fluids inlet and outlet. The silicon wafer used in this work is a p+ type,<br />
crystal orientation, with a resistivity of 8-12 mΩcm. The glass is a Borofloat 33 type, 1 mm<br />
thick. The reaction microchamber has been realized by a two-step electrochemical etching. The first<br />
step is a high current density (800 mA/cm 2 for about 300 s) electrochemical etch in hydrofluoric acid<br />
and ethanol (HF:EtOH/50:50) solution which creates a µ-well (100 µm deep) in the bulk silicon. The<br />
second step is a consecutive electrochemical etch which is used to fabricate the PSi layer at the bottom<br />
of the µ-chamber. The process parameters for this last step were: current density of 400 mA/cm 2 and<br />
etching time of 3.2 s. More details on µ-chamber fabrication can be found elsewhere [12]. The<br />
thickness and porosity of the PSi layer was measured by variable angle spectroscopic ellipsometry<br />
using a Horiba-Jobin-Yvon UVISEL ellipsometer. After the mechanical drilling of the flow channels,<br />
the glass slide has been cleaned and activated for the AB process following standard RCA and H2O2<br />
procedures. Because of the highly reactive PSi nature, the standard cleaning procedures had to be<br />
(3)
Sensors 2006, 6 683<br />
changed with a soft cleaning procedure based on trichloroethylene, acetone and ethanol. The silicon<br />
chip has also been carefully rinsed in deionized water for several minutes. Silicon etched wafer and<br />
glass top prefabricated components have been ano<strong>di</strong>cally bonded together with mutual alignment: a<br />
successful bon<strong>di</strong>ng, in terms of mechanical strength and bond quality, has been obtained at a<br />
temperature of 200 o C, voltage of 2.5 kV, with a process time of 1.5 minutes. The cross section of the<br />
chip structure and the general-view of the real device are shown in Figure 2. The reflectivity spectra in<br />
the VIS-IR wavelength region have been recorded with a very simple experimental set-up: a white<br />
source illuminates, through an optical fiber and a collimator, the porous silicon at nearly normal<br />
incidence. The reflected light is collected by an objective and coupled into a multimode fiber. The<br />
signal is <strong>di</strong>rected in an optical spectrum analyzer (Ando, Mod. AQ-6315B) and measured with a 0.2<br />
nm resolution. We have also performed time-resolved measurement in order to characterize the sensor<br />
dynamic: using the laser beam from an IR source, we have measured, as a function of time, the signal<br />
of a receiving photodetector before, during and after the exposure to acetone.<br />
Results and Discussion<br />
From the ellipsometric measurements we have estimated the thickness and the porosity of the<br />
porous silicon sensitive layer, resulting in about 9.1 µm and 79 %, respectively.<br />
Figure 2. Schematic and real view of the µ-chamber for lab-on-chip applications.<br />
Using the Bruggemann relation in Eq. (2), the refractive index is calculated as 1.34. As it can be<br />
noted in the reflectivity spectra reported in Figure 3, on exposure to several volatile substances in nonsaturated<br />
atmosphere, due to the phenomenon of capillary condensation, the average refractive index<br />
of the layer increases, and, as a consequence the optical thickness of the porous silicon layer also<br />
increases.
Sensors 2006, 6 684<br />
Intensity (a.u.)<br />
1.3<br />
1.2<br />
1.1<br />
1.0<br />
0.9<br />
0.8<br />
0.7<br />
0.6<br />
0.5<br />
0.4<br />
0.3<br />
840 850 860 870 880 890 900<br />
Wavelength (nm)<br />
Unperturbed<br />
On exposure to:<br />
Xylene<br />
Acetone<br />
Ethanol<br />
Iso-propanol<br />
Figure 3. Reflectivity spectra of the unperturbed sensor and on exposure to vapours of organic<br />
compounds.<br />
In Figure 4 the maxima of order m as function of wavenumber are shown; the slopes of the lines<br />
give the optical thickness of the layer for each substance. In the wavelength range considered, the<br />
assumption of a refractive index independent on the wavelength is satisfied since the PSi refractive<br />
index changes less then 2 % in this interval. Using the Eq. (3), the LLF, which is the filled volume by<br />
the condensed liquid, has been calculated for each substance obtaining the values reported in Table 1.<br />
The LLF values are about 30 % lower than ones obtained in the case of the PSi transducer before of<br />
the AB process [13]. These <strong>di</strong>fferences can be ascribed to a slight oxidation of the porous silicon layer<br />
due to the AB temperature process.<br />
maximum order (∆m)<br />
16<br />
14<br />
12<br />
10<br />
8<br />
6<br />
4<br />
2<br />
0<br />
Unperturbed<br />
On exposure to:<br />
Iso-propanol<br />
Acetone<br />
Ethanol<br />
Xylene<br />
12000 13000 14000 15000 16000 17000<br />
wavenumber (cm -1 )<br />
Figure 4. Maximum m-order as function of wavenumber (1/λm). The slopes of the straight lines are the<br />
optical thicknesses of the unperturbed layer and on exposure to the vapours.<br />
The silicon oxide can fill or obstruct the very small pores of the sponge-like structure, preventing<br />
the liquid to condense into them. Moreover, due to the air present during the AB process, there
Sensors 2006, 6 685<br />
probably exists an extra pressure into the µ-chamber because some air has been entrapped in the PSi<br />
matrix. As a consequence, accor<strong>di</strong>ng to the Kelvin equation (2), the capillary condensation of the<br />
liquid phase decreases.<br />
Table 1. Chemical organics substances used in sensing experiments and some relevant physicalchemical<br />
properties a .<br />
Chemical<br />
compounds<br />
n ρ<br />
(g/cm 3 )<br />
STC<br />
(mN/m)<br />
VP<br />
(kPa)<br />
BP<br />
(°C)<br />
∆<br />
(nm)<br />
LLF<br />
Iso-propanol 1.377 0.785 20.93 6.8 82.4 1000 0.22<br />
Ethanol 1.360 0.785 22.8 5.8 78 950 0.23<br />
Xylene 1.501 1.454 38.8 0.046 214 1400 0.22<br />
Acetone 1.359 0.791 23.46 30.8 56 998 0.23<br />
a<br />
n is the liquid refractive index; ρ is the density (@ 25 °C); STC is the surface tension<br />
coefficient (@ 25 °C); VP is the vapor pressure; BP is the boiling point; ∆ is the<br />
average optical path increase and LLF the layer liquid fraction.<br />
The result of time-resolved measurement is compared in Figure 5 to the data acquired on the same<br />
PSi layer before the AB integration process: it is well evident that, due to chamber miniaturization, the<br />
identification time (τid=2 s) is significantly shorter.<br />
Intensity (a. u.)<br />
1.00<br />
0.75<br />
0.50<br />
0.25<br />
0.00<br />
gas in<br />
τ id =8.7 s<br />
τ id =2 s<br />
τ rec =1.7 s<br />
τ rec =8 s<br />
gas out<br />
0 30 60 90 120 150 180 210 240 270 300<br />
Time (s)<br />
Figure 5. Time-resolved measurements of porous silicon layer during the monitoring of acetone in 0.4<br />
l test chamber (solid line) and in the integrated 14 µl µ-chamber (dashed line).<br />
The values of response time depends not only on the physical phenomena involved (i.e.,<br />
equilibrium between adsorption and desorption in the PSi layer) but also on the geometry of the test
Sensors 2006, 6 686<br />
chamber and on the measurement procedure, i.e. static or continuous flow mode. In static con<strong>di</strong>tion,<br />
the identification time is mainly determined by the <strong>di</strong>ffusion of the gas into the chamber volume: in<br />
fact, when vapor is in contact with the porous silicon surface, the capillary condensation takes place<br />
instantaneously [14-15]. For the same reason, the recovery time is longer (τrec=8 s): as soon as<br />
Nitrogen is introduced into the µ-chamber, the con<strong>di</strong>tions for capillary condensation are not still valid<br />
so that the liquid phase <strong>di</strong>sappears, depen<strong>di</strong>ng on atmosphere rate exchange. As it is shown in this<br />
graphic, the sensor response is completely reversible.<br />
Conclusions<br />
In this work, we have optically characterized an original hybrid silicon-glass sensor system, based<br />
on PSi and AB technologies, which can be thought as the reaction chamber of a more complex lab-onchip.<br />
Reflectivity spectrum of the device has been stu<strong>di</strong>ed on exposure to vapors of organic volatile<br />
solvents. An increase of the optical path, due to capillary condensation of the vapor into the pores, has<br />
been observed. Using the Bruggemann effective me<strong>di</strong>um approximation we have calculated the liquid<br />
volume fraction adsorbed into the layer for each substance, obtaining values between 0.22 and 0.23.<br />
We have characterized the sensor dynamic by time-resolved measurement during exposure to acetone.<br />
The data assure fast analysis time and the complete reversibility of the sensing mechanism. The<br />
comparison with data obtained from the same PSi transducer before the AB, shows that the integration<br />
process <strong>di</strong>d not degrade the sensor features.<br />
Acknowledgements<br />
This work has been supported by Marie Currie Fellowship of the European Community programme<br />
“Technologies for micro-opto-mechanical systems” under contract number CO-IMM-NA-01-2002.<br />
References<br />
1. Mulloni, V.; Pavesi, L. Porous silicon microcavities as optical chemical sensors. Appl. Phys. Lett.<br />
2000, 76, 2523-2525.<br />
2. Gao, J.; Gao, T.; Sailor, M.J. Porous-silicon vapor sensor based on laser interferometry. Appl.<br />
Phys. Lett. 2000, 77, 901-903.<br />
3. Benson, T. M.; Arrand, H. F.; Sewell, P.; Niemeyer, D.; Loni, A.; Bozeat, R. J.; Kruger, M.;<br />
Arens-Fischer, R.; Thoenissen, M.; Luth, H. Progress towards achieving integrated circuit<br />
functionality using porous silicon optoelectronic components. Mat. Sci. Eng. B 1999, 69-70, 92-<br />
95.<br />
4. Wongmanerod, C.; Zangooie, S.; Arwin, H. Determination of pore size <strong>di</strong>stribution and surface<br />
area of thin porous silicon layers by spectroscopic ellipsometry. Applied Surface Science 2001,<br />
172, 117-125.<br />
5. Puers, B.; Peeters, E.; Van Den Bossche, A.; Sansen, W. A capacitive pressure sensor with low<br />
impedance output and active suppression of parasitic effects. Sens. Actuat. A 1990, 21-23, 108-<br />
114.
Sensors 2006, 6 687<br />
6. Obermeier, E. Proc. Semiconductor Wafer Bon<strong>di</strong>ng, Science, Technology and Applications.<br />
Electrochem. Soc, 1995, 212-220.<br />
7. Malecki, K.; Della Corte, F.G.; Silicon-glass ano<strong>di</strong>c bon<strong>di</strong>ng at low temperature. Proc. of<br />
Micromachining and Microfabrication Process Technology X, SPIE Vol. 5715 2005, 180-189.<br />
8. Lin, V. S.-Y.; Motesharei, K.; Dancil, K.-P. S.; Sailor, M. J.; Gha<strong>di</strong>ri, M. R. A porous silicon-based<br />
optical interferometric biosensor. Science 1997, 278, 840-843.<br />
9. Anderson, M. A.; Tinsley-Brown, A.; Allcock, P.; Perkins, E. A.; Snow, P.; Hollings, M.; Smith,<br />
R. G.; Reeves, C.; Squirrell, D. J.; Nicklin, S.; Cox; T. I. Sensitivity of the optical properties of<br />
porous silicon layers to the refractive index of liquid in the pores. Phys. Stat. Sol. A 2003, 197,<br />
528-533.<br />
10. Neimark, V.; Ravikovitch, P. I. Capillary condensation in MMS and pore structure<br />
characterization. Microporous Mesoporous Mater. 2001, 44-45, 697-707.<br />
11. Spanier, J. E.; Herman, I. P. Use of hybrid phenomenological and statistical effective-me<strong>di</strong>um<br />
theories of <strong>di</strong>electric functions to model the infrared reflectance of porous SiC films. Phys. Rev.<br />
B 2000, 61, 10437-10450.<br />
12. De Stefano, L.; Malecki, K.; Moretti, L.; Ren<strong>di</strong>na, I. Ano<strong>di</strong>cally bonded silicon-glass optical chip for<br />
biochemical sensing applications. Proc. SPIE Int. Soc. Opt. Eng. 2005, 5718-7, 38-47.<br />
13. De Stefano, L.; Ren<strong>di</strong>na, I.; Moretti, L.; Rossi, A.M.; Tundo, S. Smart optical sensors for<br />
chemical substances based on porous silicon technology. Applied Optics 2004, 43/1, 167-172.<br />
14. Allcock, P.; Snow, P.A. Time-resolved sensing of organic vapors in low modulating porous<br />
silicon <strong>di</strong>electric mirrors. J. Appl. Phys. 2001, 90, 50525057.<br />
15. De Stefano, L.; Moretti, L.; Ren<strong>di</strong>na, I.; Rossi, A. M. Time-resolved sensing of chemical species<br />
in porous silicon optical microcavity. Sensor and Actuators B 2004, 100, 168-172.<br />
© 2006 by MDPI (http://www.mdpi.org). Reproduction is permitted for noncommercial purposes.
Laser oxidation micropatterning of a porous silicon based<br />
biosensor for multianalytes microarrays<br />
The aim of this section is to present the fabrication and characterization of oxi<strong>di</strong>zed<br />
micropatterned biosensors based on PSi, obtained by <strong>di</strong>rect laser writing and their<br />
chemical functionalization for biological applications. The laser local oxidation technique is<br />
a good alternative to the tra<strong>di</strong>tional photolithographic process, in fact the standard masking<br />
of PSi by means of photoresist presents remarkable <strong>di</strong>fficulties since the developer, being<br />
alkaline, causes the <strong>di</strong>ssolution of the PSi.
LASER OXIDATION MICROPATTERNING OF A POROUS<br />
SILICON BASED BIOSENSOR FOR MULTIANALYTES<br />
MICROARRAYS<br />
L. DE STEFANO 1 , L. <strong>ROTIROTI</strong> 1,3 , I. REA 1,2 , E. DE TOMMASI 1 , M. A. NIGRO 4 , F.<br />
G. DELLA CORTE 4 , AND I. RENDINA 1<br />
1 IMM-CNR - Sezione <strong>di</strong> Napoli, Via P. Castellino 111, 80131 Napoli, Italy<br />
2 Dept. of Physics, “Federico II” University of Naples, Monte S. Angelo, 80126 Naples,<br />
Italy<br />
3 Dept. of Organic Chemistry and Biochemistry, “Federico II” University of Naples,<br />
Monte S. Angelo, 80126 Naples, Italy<br />
4 “Me<strong>di</strong>terranea” University of Reggio Calabria Department of Computer Science,<br />
Mathematics, Electronics and Transports, Località Vito <strong>di</strong> Feo, Reggio Calabria, Italy<br />
Abstract<br />
In this communication, we present the fabrication and characterization of oxi<strong>di</strong>zed porous silicon<br />
(PSi) micropatterns, obtained by <strong>di</strong>rect laser writing, and chemically functionalized for biological<br />
applications. The laser local oxidation technique is a good alternative to the tra<strong>di</strong>tional<br />
photolithographic process since this approach allows the development of micropatterned<br />
biosensors. In fact, from a technological point of view, the standard masking of PSi by means of<br />
photoresist presents remarkable <strong>di</strong>fficulties because of the low resistance to the electrochemical<br />
process. The micropatterned PSi oxi<strong>di</strong>zed surfaces can be properly functionalized by a specific<br />
chemical linker. The mo<strong>di</strong>fied surface is able to link the carboxylic or aminic groups of biological<br />
probes. The bin<strong>di</strong>ng between the functionalized surface and the biological matter has been tested<br />
by <strong>di</strong>rectly spotting on the PSi local oxide a fluorescent labelled antibody. The device has been<br />
characterized by FT-IR measurements and fluorescence macroscopy.<br />
Keywords<br />
Porous silicon, laser writing, biosensors, microarrays.<br />
INTRODUCTION<br />
Porous silicon based optical biosensors offer several advantages in label-free<br />
detection and identification of organic molecules due to the well known<br />
properties of this material [1, 2]. PSi has a spongy structure with a large<br />
specific surface of the order of 100–500 m 2 cm -3 [3], so that a very effective<br />
interaction with several adsorbates is assured. When a biological solution<br />
penetrates into the PSi pores, it substitutes the air, so that the average <strong>di</strong>electric<br />
properties of the heterogeneous silicon-air-liquid system change inducing a<br />
red-shift in the reflectivity spectrum of the device. The effect depends on the<br />
refractive index value of the solution but also on how it penetrates into the<br />
pores. Due to this sensing mechanism, PSi optical devices cannot identify the<br />
382
383<br />
single components of a complex mixture. In order to enhance the sensor<br />
selectivity, several biological probes, which exploit very specific interactions<br />
only with selected biochemical molecules, can be linked to the PSi surface:<br />
most common examples are DNA strands, proteins, enzymes, antibo<strong>di</strong>es and<br />
so on. In this work, the fabrication and characterization of oxi<strong>di</strong>zed porous<br />
silicon micropatterns obtained by <strong>di</strong>rect laser writing [5] and their chemical<br />
functionalization for biological applications are reported. In particular we have<br />
employed a murine monoclonal antibody (IgG) UN1 previously selected for<br />
the specific reactivity with human thymocytes as compared to peripheral blood<br />
cells [4]. The antigen recognized by UN1 is a 100–120 kDa transmembrane<br />
glycoprotein showing biochemical features of cell membrane-associated<br />
mucin-like glycoproteins, a class of macromolecules that are involved in cellto-cell<br />
interactions and cancer progression [6].<br />
EXPERIMENTAL AND RESULTS<br />
The device was a PSi monolayer, 1 μm thick and with a porosity of 70%,<br />
produced by electrochemical etch in a HF-based solution, starting from a<br />
highly doped p+-silicon, oriented, 0.04 Ω cm resistivity, 400 μm thick.<br />
The silicon was etched using a 30 wt. % HF/ethanol solution in dark and at<br />
room temperature. After the etching the samples were rinsed in ethanol and<br />
dried under a stream of N2.<br />
Laser <strong>di</strong>ode<br />
M2<br />
Pinhole<br />
M1<br />
XY stage<br />
Objective X40<br />
PS sample<br />
M3<br />
Fig. 1: Scheme of the setup used to write the channel waveguides.<br />
The localized oxidation is obtained by exposing the sample to the beam of a<br />
laser <strong>di</strong>ode (μLS Micro Laser Sistem) at 408 nm with an output power of
384<br />
48mW, using a setup composed of a microscope objective and a motorized xy<br />
micrometric stage as shown in Figure 1. The beam <strong>di</strong>mension is about 1.4/0.32<br />
mm. We have <strong>di</strong>rectly written on the sample some oxide strips, 10 μm wide,<br />
with a laser fluence of 38 mJ/mm 2 .<br />
Starting from our previous experience on chemical mo<strong>di</strong>fication of the<br />
oxi<strong>di</strong>sed porous silicon, we have used an organosilane compound such as the<br />
aminopropyltriethoxysilane (APTES) and a functional crosslinker such as<br />
Glutaraldehyde (GA) which reacts with the amino groups on the silanized<br />
surface and coats the micropatterned PSi oxi<strong>di</strong>zed surfaces. [7]. The obtained<br />
mo<strong>di</strong>fied surface works as an active substrate for the chemistry of the<br />
following attachment of other biological probes, such as DNA single strand,<br />
proteins, enzymes, and antibo<strong>di</strong>es.<br />
The functionalization process takes place <strong>di</strong>rectly on the chip surface, simply<br />
covering it with the chemical solutions. In Fig. 2 is reported the scheme of the<br />
functionalisation process. The PSi device has been rinsed by immersion in a<br />
5% solution of APTES and an hydroalcoholic mixture of water and methanol<br />
(1:1), for 20 min at room temperature. After the reaction time, we have washed<br />
the sample with demi-water and methanol and dried in N2 stream. The<br />
silanized surface was then baked at 100°C for 10 min.<br />
The APTES mo<strong>di</strong>fied surface is able to link the carboxylic group of a<br />
biological probe such as IgG or other antibo<strong>di</strong>es. A further step of<br />
functionalization is required to bind the amino group always present in<br />
biological matter through the aldehy<strong>di</strong>c group of opportune chemical linker. To<br />
this aim, we have treated the chip in a 2.5% glutaraldehyde solution for 30<br />
min. The glutaraldehyde reacts with the amino groups on the silanized surface<br />
and coats the internal surface of the pores with another thin layer of molecules.<br />
Infrared spectra were collected by using a Fourier transform spectrometer<br />
equipped with a microscope (Nicolet Continuμm XL, Thermo Scientific) and<br />
are reported in Fig. 3, where it is well evident that, after the functionalization<br />
processes, the Si-OH peaks are completely substituted by the organic linkers<br />
characteristics ones.
Fig. 2: functionalization steps of PSi surface by means of APTES and Glutaraldehyde and<br />
subsequent bin<strong>di</strong>ng of the biological probe.<br />
Fig. 3: FT-IR spectra of the micropatterned PSi oxi<strong>di</strong>zed surfaces before and after the chemical<br />
surface mo<strong>di</strong>fication treatments.<br />
385<br />
The bin<strong>di</strong>ng between the surface and the biological matter has been tested by<br />
using fluorescent labelled bioprobes <strong>di</strong>rectly spotted on the PSi chemical<br />
mo<strong>di</strong>fied structures. We have spotted on the porous silicon chip a solution<br />
containing the rhodamine labelled IgG antibody, 6.8 μM, and incubated<br />
overnight at room temperature. The device has been observed by a Leica Z16
386<br />
APO fluorescence macroscopy system. By 100W high-pressure mercury<br />
source we found that the fluorescence is very high and homogeneous on the<br />
region locally oxi<strong>di</strong>zed by the laser as shown in Fig. 4.<br />
Fig. 4: Post-<strong>di</strong>alysis fluorescence image of the chip after rhodamine labeled IgG bin<strong>di</strong>ng<br />
CONCLUSIONS<br />
In this work, we have exploited the possibility to locally functionalize the PSi<br />
surface.<br />
A micropattern has been <strong>di</strong>rectly written on the porous silicon layer by a laser<br />
oxidation process. The pattern, constituted by strips 2 μm wide, has been<br />
functionalized by specific chemical linkers such as APTES and glutaraldehyde;<br />
then we have spotted on the PSi sample a fluorescent labelled antibody. The<br />
interaction between the mo<strong>di</strong>fied surface and the biological matter has been<br />
verified by fluorescence macroscopy; a high fluorescence signal has been<br />
observed only in the oxi<strong>di</strong>zed region.<br />
REFERENCES<br />
[1] H. Ouyang, C.C. Striemer, P.M. Fauchet, APL 88, 163108, 2006.<br />
[2] C. Pacholski, M. Sartor, M.J. Sailor, F. Cunin, G.M. Miskelly, J. Am.<br />
Chem. Soc. 127, 11636-11645, 2005.
387<br />
[3] R. Herino, G. Bomchil, K. Barla, C. Bertrand, J.L. Ginoux, J.<br />
Electrochem. Soc. 134-8, 1994, 1987.<br />
[4] Hilkens J., M.J. Ligtenberg, H.L. Vos, S.G. Litinov, Cell membraneassociated<br />
mucins and their adhesion-modulating property, Trends<br />
Biochem. Sci. 1992, 17, 359.<br />
[5] A. M. Rossi, S. Borini, L. Boarino, and G. Amato, Phys. Stat. Sol. (a) 197,<br />
No. 1, 284–287 (2003).<br />
[6] Cecco L., H.M. Bond, P. Bonelli, et al., Tissue Antigens 1998, 51, 23.<br />
[7] De Stefano L., A. Vitale, I. Rea, M. Staiano, L. Rotiroti, T. Labella, I.<br />
Ren<strong>di</strong>na, V. Aurilia, M. Rossi, S. D’Auria, Extremophiles (2007), DOI<br />
10.1007/s00792-006-0058-6.
PUBLICATIONS:<br />
J1. Luca De Stefano, F. G. Della Corte, K. Malecki, Ivo Ren<strong>di</strong>na, Lucia Rotiroti, Luigi Moretti,<br />
Andrea M. Rossi, “Integrated silicon-glass opto-chemical sensors for lab-on-chip application”,<br />
Sensor and Actuators B,114, 625-630, (2006).<br />
J2. L. De Stefano, I. Ren<strong>di</strong>na, L. Rotiroti, L. Moretti, V. Scognamiglio, M. Rossi, S. D’Auria,<br />
“Porous silicon-based optical microsensor for the detection of L-Glutamine”, Biosensors and<br />
Bioelectronics 21/8, 1664-1667, 2006.<br />
J3. L. Moretti, I. Rea, L. Rotiroti, I. Ren<strong>di</strong>na, G. Abbate, A. Marino, L. De Stefano “Photonic band<br />
gaps analysis of Thue-Morse multilayers made of porous silicon” Optics Express, 14 (2006)<br />
6264-6272.<br />
J4. L. De Stefano, I. Rea, I. Ren<strong>di</strong>na, L. Rotiroti, M. Rossi, and S. D’Auria, “Resonant cavity<br />
enhanced optical microsensor for molecular interactions based on porous silicon”, Phys Stat<br />
Sol (A) , 203 (2006) 886-891.<br />
J5. S. D’Auria, S. Borino, A. Vitale, A.M. Rossi, I. Rea, I. Ren<strong>di</strong>na, L. Rotiroti, M. Staiano, M. de<br />
Champdorè, V. Aurilia, A. Parracino, M. Rossi, L. De Stefano, “Nanostructured silicon-based<br />
biosensors for the selective identification of analytes of social interest”, Journal of Physics:<br />
Condensed Matter, 18 S2019-S2028, (2006).<br />
J6. L. De Stefano, L. Rotiroti, I. Rea, I. Ren<strong>di</strong>na, L. Moretti, G. Di Francia, E. Massera, A.<br />
Lamberti, P. Arcari, C. Sangez, “Porous silicon-based optical biochips” Journal of Optics A:<br />
Pure and Applied Optics, S540-S544, 8 (2006).<br />
J7. L. De Stefano, K.Malecki, F. G. Della Corte , L. Moretti , I. Rea , L. Rotiroti, I. Ren<strong>di</strong>na “A<br />
Microsystem Based on Porous Silicon-Glass Ano<strong>di</strong>c Bon<strong>di</strong>ng for Gas and Liquid Optical<br />
Sensing” Sensors, 6 (2006) 680-687.<br />
J8. L. De Stefano, M. Rossi, M. Staiano, G. Mamone, A. Parracino, L. Rotiroti, I. Ren<strong>di</strong>na, M.<br />
Rossi, and S. D’Auria, “Glutamine-bin<strong>di</strong>ng protein from Escherichia coli specifically binds a<br />
wheat glia<strong>di</strong>n peptide allowing the design of a new porous silicon-based optical biosensor”,<br />
Journal of Proteome Research 5,5, 1241-1245, 2006<br />
J9. L. De Stefano, L. Rotiroti, I. Rea, F. Della Corte, D. Alfieri, L. Moretti, I. Ren<strong>di</strong>na, “An<br />
integrated pressure-driven microsystem based on porous silicon for optical monitoring of<br />
gaseous and liquid substances”; Phys. Stat. Sol. (a) 204, No. 5, 1459-1463 (2007).<br />
J10. L. De Stefano, L. Rotiroti, I. Rea, I. Ren<strong>di</strong>na, L. Moretti, "Quantitative determinations<br />
in hydro-alcoholic binary mixtures by porous silicon optical microsensors", Phys. Stat. Sol. (c),<br />
4, No. 6, 1941– 1945 (2007).<br />
J11. L. De Stefano, I. Rea, L. Moretti, L. Rotiroti, I. Ren<strong>di</strong>na, “Optical properties of porous<br />
silicon Thue-Morse structures”, Phys. Stat. Sol. (c) 4, No. 6, 1966-1970 (2007).<br />
J12. I. Ren<strong>di</strong>na, I. Rea, L. Rotiroti, L. De Stefano, “Porous Silicon Based Optical<br />
Biosensors and Biochips”, Physica E, 38 (2007) 188-192 invited paper.<br />
J13. L. De Stefano, I. Rea, L. Rotiroti, I. Ren<strong>di</strong>na, L. Fragomeni, F.G. Della Corte, “An<br />
integrated hybrid optical device for sensing applications”, Phys. Stat. Sol. (c), 4, No. 6, 1946–<br />
1950 (2007)<br />
J14. L. De Stefano, L. Rotiroti, I. Rea, I. Ren<strong>di</strong>na, M. Rossi, S. D’Auria, “Assessment of a<br />
porous silicon transducer for the development of optical biochips”, Rivista <strong>di</strong> Materiali<br />
Nanocompositi e Nanotecnologie 3, 39-41, 2006<br />
J15. L. De Stefano, I. Rea, L. Rotiroti, M. Io<strong>di</strong>ce, I. Ren<strong>di</strong>na, “Optical microsystems based<br />
on a nanomaterial technology”, J. Phys.: Condens. Matter 19 395008, doi:10.1088/0953-<br />
8984/19/39/395008, (2007).<br />
J16. L. De Stefano, L. Rotiroti, I. Ren<strong>di</strong>na, A.M. Rossi, M. Rossi, S. D’Auria, “Biochips at<br />
work: porous silicon microbiosensor for proteomic <strong>di</strong>agnostic”, Phys.: Condens. Matter 19<br />
395007, doi:10.1088/0953-8984/19/39/395007 (2007).<br />
J17. L. De Stefano, P. Arcari, A. Lamberti, C. Sanchez, L. Rotiroti, I. Rea, I. Ren<strong>di</strong>na,<br />
“DNA optical detection based on porous silicon technology: from biosensors to biochips”<br />
Sensors, 7 (2007) 214-221.<br />
J18. L. De Stefano, A. Lamberti, L. Rotiroti, M. De Stefano, “Interfacing the nanostructured<br />
biosilica microshells of the marine <strong>di</strong>atom Coscino<strong>di</strong>scus wailesii with biological matter”, Acta<br />
Biomaterialia Volume 4, Issue 1, January 2008, Pages 126-130.<br />
J19. Ivo Ren<strong>di</strong>na, Edoardo De Tommasi, Ilaria Rea, Lucia Rotiroti, and Luca De Stefano,<br />
“Porous silicon optical transducers offer versatile platforms for biosensors” SPIE NEWSROOM<br />
DOI: 10.1117/2.1200801.0982 invited paper
J20. L. De Stefano, A. Vitale, I. Rea, M. Staiano, L. Rotiroti, T. Labella, I. Ren<strong>di</strong>na, V.<br />
Aurilia, M. Rossi, S. D’Auria, “Enzymes and proteins from extremophiles as hyperstable<br />
probes in nanotechnology: the use of D-trehalose/D-maltose-bin<strong>di</strong>ng protein from the<br />
hyperthermophilic archaeon Thermococcus litoralis for sugars monitoring”, Extremophiles, 12,<br />
1, 69-73 (2008).<br />
J21. L. De Stefano, E. De Tommasi, I. Rea, L. Rotiroti, L. Giangrande, G. Oliviero, N.<br />
Borbone, A. Galeone and G. Piccialli, “Oligonucleotides <strong>di</strong>rect synthesis on porous silicon<br />
chip”, Nucleic Acids Symposium Series No. 52 721–722<br />
J22. L. De Stefano, L. Rotiroti, M. De Stefano, A. Lamberti, S. Lettieri, A. Setaro, P.<br />
Maddalena, “Marine Diatoms as Optical Biosensors" has been accepted for publication in<br />
Biosensors and Bioelectronics, 2008.<br />
J23. E. De Tommasi, L. De Stefano, I. Rea, V. Di Sarno, L. Rotiroti, P. Arcari, A. Lamberti,<br />
C. Sanges, and I. Ren<strong>di</strong>na, “Porous Silicon Based Resonant Mirrors for Biochemical Sensing”,<br />
Sensors 2008, Volume 8 (10), Page 6549-6556.<br />
PROCEEDINGS:<br />
P1. L. De Stefano, L. Rotiroti, I. Rea, L. Moretti, I. Ren<strong>di</strong>na, “Quantitative determinations in<br />
liquid and gaseous binary misture by porous silicon optical microsensors”, Advances in<br />
Sensors and Interfaces, Proc. of IWASI 2005, Eds. D. De Venuto and B. Courtois, 178-180,<br />
(2005).<br />
P2. L. De Stefano, I. Ren<strong>di</strong>na, L. Rotiroti, L. Moretti, V. Scognamiglio, M. Rossi, S. D’Auria,<br />
“Protein-ligand interaction detection by a porous silicon optical sensor”, Procee<strong>di</strong>ngs of the<br />
10 th Italian Conference Sensors and Microsystems 2005, World Scientific, 89.<br />
P3. Luca De Stefano, F. G. Della Corte, K. Malecki, Ivo Ren<strong>di</strong>na, Lucia Rotiroti, Luigi Moretti,<br />
Andrea M. Rossi, “A microchamber for lab-on-chip applications based on silicon-glass ano<strong>di</strong>c<br />
bon<strong>di</strong>ng”, Procee<strong>di</strong>ngs of the 10 th Italian Conference Sensors and Microsystems 2005, World<br />
Scientific, 450, 2005.<br />
P4. L. De Stefano, K. Malecki, F. Della Corte, L. Moretti, I. Rea, L. Rotiroti, I. Ren<strong>di</strong>na,<br />
“Silicon/glass integrated optical sensor based on porous silicon for gas and liquid inspection”<br />
Procee<strong>di</strong>ngs EUROSENSORS XIX, Settembre 11-14, 2005, Barcellona, Spagna.<br />
P5. L. Rotiroti, L. De Stefano, I. Rea, L. Moretti, G. Di Francia, V. La Ferrara, A. Lamberti, P.<br />
Arcari, C. Sanges, I. Ren<strong>di</strong>na; “Porous silicon based optical biochips” Book of Abstract,<br />
European Optical Society Topical Meeting on Optical Microsystems, Settembre 15-18, 2005,<br />
Capri, Italia.<br />
P6. L. De Stefano, I. Rea, L. Rotiroti, I. Ren<strong>di</strong>na, A. Lamberti, P. Arcari, C. Sanges, A. M.<br />
Rossi, “Resonant cavity enhanced optical microsensor for molecular interaction based on<br />
porous silicon”; Book of Abstract, The 11th National Conference on Sensor and Microsystems,<br />
p. 139-140, Febbraio 8-10, 2006, Lecce, Italia.<br />
P7. L. De Stefano, L. Rotiroti, I. Rea, I. Ren<strong>di</strong>na, L. Moretti, “Optical quantitative determination<br />
of the alcohol fraction in binary mixtures”; Book of Abstract, The 11th National Conference on<br />
Sensor and Microsystems, p. 141-142, Febbraio 8-10, 2006, Lecce, Italia.<br />
P8. L. De Stefano, I. Rea, L. Rotiroti, I. Ren<strong>di</strong>na, L. Fragomeni, F. G. Della Corte, “A hybrid<br />
integrated optical microsystem for sensing application”; Book of Abstract, The 11th National<br />
Conference on Sensor and Microsystems, p. 245-246, Febbraio 8-10, 2006, Lecce, Italia.<br />
P9. L. De Stefano, I. Rea, L. Rotiroti, K. Malecki, L. Moretti, F. G. Della Corte, I. Ren<strong>di</strong>na,<br />
Sensors and Microsystems, 227-231 (2006).<br />
P10. L. De Stefano, I. Rea, L. Rotiroti, L. Moretti, I. Ren<strong>di</strong>na, “Optical properties of porous<br />
silicon Thue-Morse structures”; Extended Abstracts of 5th International Conference on Porous<br />
Semiconductuctors-Science and Technology, p. 73-74, Marzo 12-17, 2006, Sitges-Barcellona,<br />
Spagna.<br />
P11. L. De Stefano, L. Rotiroti, I. Rea, F. Della Corte, D. Alfieri, L. Moretti, I. Ren<strong>di</strong>na, “An<br />
integrated pressure-driven microsystem based on porous silicon for optical monitoring of<br />
gaseous and liquid substances”; Extended Abstracts of 5th International Conference on<br />
Porous Semiconductuctors-Science and Technology, p. 126-127, Marzo 12-17, 2006, Sitges-<br />
Barcellona, Spagna.
P12. L. De Stefano, L. Rotiroti, I. Rea, L. Moretti, I. Ren<strong>di</strong>na, “Quantitative determinations in<br />
hydro-alcoholic binary mixtures by porous silicon optical microsensors”; Extended Abstracts of<br />
5th International Conference on Porous Semiconductuctors-Science and Technology, p. 248-<br />
249, Marzo 12-17, 2006, Sitges-Barcellona, Spagna.<br />
P13. L. De Stefano, I. Rea, L. Rotiroti, L. Fragomeni, F. Della Corte, I. Ren<strong>di</strong>na, ”An integrated<br />
hybrid optical device for sensing applications”; Extended Abstracts of 5th International<br />
Conference on Porous Semiconductuctors-Science and Technology, p. 250-251, Marzo 12-<br />
17, 2006, Sitges-Barcellona, Spagna.<br />
P14. L. De Stefano, P. Maddalena, L. Moretti, I. Rea, I. Ren<strong>di</strong>na, “Porous biosilica from marine<br />
<strong>di</strong>atoms: a new brand porous material for photonic applications”; Extended Abstracts of 5th<br />
International Conference on Porous Semiconductuctors-Science and Technology, p. 252-253,<br />
Marzo 12-17, 2006, Sitges-Barcellona, Spagna.<br />
P15. L. De Stefano, I. Rea, L. Rotiroti, L. Moretti, I. Ren<strong>di</strong>na “Resonant photonic devices<br />
based on porous silicon: a perfect tool for optical sensing”; Procee<strong>di</strong>ngs of XX<br />
EUROSENSORS Conference, p. 294-295, Settembre 17-20, 2006, Göteborg, Svezia.<br />
P16. L. De Stefano, L. Rotiroti, I. Rea, E. De Tommasi, M. A. Nigro, F. G. Della Corte, and I.<br />
Ren<strong>di</strong>na, Laser oxidation micropatterning of a porous silicon based biosensor for multianalytes<br />
microarray, Procee<strong>di</strong>ngs of the 12 th Italian Conference Sensors and Microsystems 2007,<br />
World Scientific, 382-387.<br />
P17. L, De Stefano, L. Rotiroti, I. Rea, E. De Tommasi, A. Vitale, M. Rossi, I. Ren<strong>di</strong>na and S.<br />
D’Auria, "Design and realization of highly stable porous silicon optical biosensor based on<br />
proteins from extremophiles" Proc. of SPIE (2007) Vol. 6585 658517-1<br />
P18. L. De Stefano, I. Ren<strong>di</strong>na, I. Rea, L. Rotiroti, E. De Tommasi, and G. Barillaro, " An<br />
optical microsystem based on silicon-air Bragg mirror for the continuous monitoring of liquid<br />
substances" Proc. of SPIE (2007) Vol. 6585 65850N-1<br />
P19. Ivo Ren<strong>di</strong>na, Edoardo De Tommasi, Ilaria Rea, Lucia Rotiroti, and Luca De Stefano,<br />
“Porous silicon optical transducers offer versatile platforms for biosensors” DOI:<br />
10.1117/2.1200801.0982<br />
P20. L. De Stefano, I. Ren<strong>di</strong>na, I. Rea, L. Moretti, L. Rotiroti, “Aperio<strong>di</strong>c photonic bandgap<br />
devices based on nanostructured porous silicon”, Procee<strong>di</strong>ngs of SPIE, Volume 6593,<br />
Photonic Materials, Devices, and Applications II, Ali Serpengüzel, Gonçal Badenes, Giancarlo<br />
C. Righini, E<strong>di</strong>tors, 65931A (Jun. 5, 2007).<br />
P21. I. Ren<strong>di</strong>na, I. Rea, L. Rotiroti, E. De Tommasi, L. De Stefano, “Integrated optical<br />
biosensors and biochips based on porous silicon technology”, Procee<strong>di</strong>ngs of SPIE -- Volume<br />
6898, Silicon Photonics III, Joel A. Kubby, Graham T. Reed, E<strong>di</strong>tors, 68981D (Feb. 13, 2008)<br />
P22. Lucia Rotiroti, Paolo Arcari, Annalisa Lamberti, Carmen Sanges, Edoardo De Tommasi,<br />
Ilaria Rea, Ivo Ren<strong>di</strong>na, and Luca De Stefano, “Optical detection of PNA/DNA hybri<strong>di</strong>zation in<br />
resonant porous silicon-based devices” Proc. SPIE 6991, 699120 (2008).<br />
P23. L. Rotiroti, E. De Tommasi, I. Ren<strong>di</strong>na, M. Canciello, G. Maglio, R. Palumbo, and L. De<br />
Stefano,” BIOACTIVE POLIMERS MODIFIED POROUS SILICON NANOSTRUCTURES”, 6th<br />
International Conference on Porous Semiconductuctors-Science and Technology, p. 348-349,<br />
Marzo 10-14, 2008, Mallorca, Spagna.<br />
P24. L. Rotiroti, I. Ren<strong>di</strong>na, E. De Tommasi, M. Canciello, G. Maglio, R. Palumbo, and L. De<br />
Stefano, A Nanostructured Hybrid Device Based On Polymers Infiltrated Porous Silicon<br />
Layers For Biotechnological Applications, Procee<strong>di</strong>ngs of the Polymer Processing Society<br />
24th Annual Meeting ~ PPS-24 ~ June 15-19, 2008 Salerno (Italy).<br />
P25. L. Rotiroti, I. Ren<strong>di</strong>na, E. De Tommasi, M. Canciello, G. Maglio, R. Palumbo, and L. De<br />
Stefano, A Hybrid Optical Biosensor Based on Polymer Infiltrated Porous Silicon Device, IEEE<br />
SENSORS 2008, Lecce (Italy).<br />
NATIONAL AND INTERNATIONAL CONGRESS<br />
C1. L. De Stefano, I. Ren<strong>di</strong>na, L. Rotiroti, L. Moretti, V. Scognamiglio, M. Rossi, S. D’Auria,<br />
“Protein-ligand interaction detection by a porous silicon optical sensor”, Oral, AISEM 2005,<br />
Firenze, 15-17 Febbraio, 2005.
C2. Luca De Stefano, F. G. Della Corte, K. Malecki, Ivo Ren<strong>di</strong>na, Lucia Rotiroti, Luigi Moretti,<br />
Andrea M. Rossi, “A microchamber for lab-on-chip applications based on silicon-glass<br />
ano<strong>di</strong>c bon<strong>di</strong>ng”, Poster, AISEM 2005, Firenze, 15-17 Febbraio, 2005.<br />
C3. L. De Stefano, L. Rotiroti, I. Rea, L. Moretti, I. Ren<strong>di</strong>na, “Quantitative determinations in<br />
liquid and gaseous binary mixtures by porous silicon optical microsensors”, Oral, IWASI<br />
2005, Bari, 19-20 April, 2005.<br />
C4. L. De Stefano, L. Rotiroti, I. Rea, L. Moretti, I. Ren<strong>di</strong>na, “Microsensori ottici per applicazioni<br />
biochimiche basati sulla nanotecnologia del silicio poroso”, Oral, GE2005, Giar<strong>di</strong>ni Naxos<br />
(CT), 29 giugno – 2 luglio, 2005.<br />
C5. M. A. Ferrara, L. Sirleto, L. Rotiroti, L. Moretti, E. Santamato, I. Ren<strong>di</strong>na “EMISSIONE<br />
RAMAN IN SILICIO POROSO A 1.5 μm”, Oral, GE2005, Giar<strong>di</strong>ni Naxos (CT), 29 giugno –<br />
2 luglio, 2005.<br />
C6. L. De Stefano, K. Malecki, F. G. Della Corte, L. Moretti, I. Rea, L. Rotiroti, I. Ren<strong>di</strong>na, Oral,<br />
“Silicon/glass integrated optical sensor based on porous silicon for gas and liquid<br />
inspection”, Eurosensors XIX, Barcelona, Spain, 9-14 September, 2005.<br />
C7. L. Rotiroti, L. De Stefano, I. Rea, L. Moretti, G. Di Francia, V. La Ferrara, A. Lamberti, P.<br />
Arcari, C. Sanges, I. Ren<strong>di</strong>na, “Porous silicon based optical biochips”; Oral, European<br />
Optical Society Topical Meeting on Optical Microsystems, Settembre 15-18, 2005, Capri,<br />
Italia.<br />
C8. M. A. Ferrara, L. Sirleto, G. Messina, M.G. Donato, L. Rotiroti and I. Ren<strong>di</strong>na, “STRAIN<br />
MEASUREMENTS IN POROUS SILICON BY RAMAN SCATTERING”, AISEM 2006,<br />
Lecce, 8-10 Febbraio, 2006.<br />
C9. M. A. Ferrara, L. Sirleto, L. Moretti, L. Rotiroti, B. Jalali and I. Ren<strong>di</strong>na, “RAMAN<br />
EMISSION IN POROUS SILICON: PROSPECT FOR AN AMPLIFIER”, SPIE Silicon<br />
Photonics.<br />
C10. L. De Stefano, I. Rea, L. Rotiroti, I. Ren<strong>di</strong>na, A. Lamberti, P. Arcari, C. Sanges, A. M.<br />
Rossi, “Resonant cavity enhanced optical microsensor for molecular interaction based on<br />
porous silicon”; The 11th National Conference on Sensor and Microsystems, Febbraio 8-10,<br />
2006, Lecce, Italia.<br />
C11. L. De Stefano, L. Rotiroti, I. Rea, I. Ren<strong>di</strong>na, L. Moretti, “Optical quantitative<br />
determination of the alcohol fraction in binary mixtures”; The 11th National Conference on<br />
Sensor and Microsystems, Febbraio 8-10, 2006, Lecce, Italia.<br />
C12. L. De Stefano, I. Rea, L. Rotiroti, I. Ren<strong>di</strong>na, L. Fragomeni, F. G. Della Corte, “A hybrid<br />
integrated optical microsystem for sensing application”; The 11th National Conference on<br />
Sensor and Microsystems, Febbraio 8-10, 2006, Lecce, Italia.<br />
C13. I. Rea, D. Alfieri, K. Malecki, I. Ren<strong>di</strong>na, L. Rotiroti, L. De Stefano, L. Moretti, F.G. Della<br />
Corte, “Fast optical detection of chemical substances in integrated silicon-glass chip”; The<br />
11th National Conference on Sensor and Microsystems, Febbraio 8-10, 2006, Lecce, Italia.<br />
C14. L. De Stefano, I. Rea, L. Rotiroti, L. Moretti, I. Ren<strong>di</strong>na, “Optical properties of porous<br />
silicon Thue-Morse structures”; of 5th International Conference on Porous<br />
Semiconductuctors-Science and Technology, Marzo 12-17, 2006, Sitges-Barcellona,<br />
Spagna.<br />
C15. L. De Stefano, L. Rotiroti, I. Rea, F. Della Corte, D. Alfieri, L. Moretti, I. Ren<strong>di</strong>na, “An<br />
integrated pressure-driven microsystem based on porous silicon for optical monitoring of<br />
gaseous and liquid substances”; 5th International Conference on Porous<br />
Semiconductuctors-Science and Technology, Marzo 12-17, 2006, Sitges-Barcellona,<br />
Spagna.<br />
C16. L. De Stefano, L. Rotiroti, I. Rea, L. Moretti, I. Ren<strong>di</strong>na, “Quantitative determinations in<br />
hydro-alcoholic binary mixtures by porous silicon optical microsensors”; 5th International<br />
Conference on Porous Semiconductuctors-Science and Technology, Marzo 12-17, 2006,<br />
Sitges-Barcellona, Spagna.<br />
C17. L. De Stefano, I. Rea, L. Rotiroti, L. Fragomeni, F. Della Corte, I. Ren<strong>di</strong>na, “An integrated<br />
hybrid optical device for sensing applications”; 5th International Conference on Porous<br />
Semiconductuctors-Science and Technology, Marzo 12-17, 2006, Sitges-Barcellona,<br />
Spagna.
C18. L. De Stefano, I. Rea, L. Rotiroti, I. Ren<strong>di</strong>na, “Porous silicon biosensors and biochips”,<br />
Invited, EMRS 2006, May 29-June 2, Nice, France, 2006.<br />
C19. M. Angeloni, L. Moretti, V. Mocella, L. De Stefano, L. Rotiroti, I. Rea, S. Bernstorff, “In<br />
situ investigation of capillary condensation phenomena in porous silicon nanostructured by<br />
means of GISAX and SAXS measurements”, Poster, EMRS 2006, May 29-June 2, Nice,<br />
France, 2006.<br />
C20. L. De Stefano, I. Rea, L. Rotiroti, L. Moretti, F. G. Della Corte, I. Ren<strong>di</strong>na, “Integrated<br />
optical microsensors based on porous silicon technology for liquid and gas detection”,<br />
Poster, EMRS 2006, May 29-June 2, Nice, France, 2006.<br />
C21. M. A. Ferrara, L. Sirleto, M. Angeloni, L. Rotiroti, B. Jalalii, I. Ren<strong>di</strong>na, “Raman emission<br />
in porous silicon: prospects fora n amplifier”, Poster, EMRS 2006, May 29-June 2, Nice,<br />
France, 2006.<br />
C22. C. Sanges, A. Lamberti, L. Rotiroti, L. De Stefano, I. Ren<strong>di</strong>na, P. Arcari, “Biochip ottici in<br />
Silicio Poroso”, Poster, XII GIORNATE SCIENTIFICHE - Facoltà <strong>di</strong> Me<strong>di</strong>cina e Chirurgia,<br />
Agraria, Me<strong>di</strong>cina veterinaria, Farmacia, Scienze MM.FF.NN., Scienze Biotecnologiche, 15-<br />
16 June, Napoli, Italy, 2006.<br />
C23. I. Rea, L. Rotiroti, I. Ren<strong>di</strong>na. L. Moretti, L. De Stefano; “Proprietà ottiche <strong>di</strong> multistrati<br />
thue-morse realizzati con la tecnologia del silicio poroso” Riunione Annuale del Gruppo<br />
Elettronica (GE 2006), Giugno 21-23, Ischia (NA), Italia.<br />
C24. L. Rotiroti, I. Rea, D. Alfieri, L. Moretti I. Ren<strong>di</strong>na, F. Della Corte, L. De Stefano;<br />
“Microsensori ottici integrati basati sulla tecnologia del silicio poroso per il riconoscimento <strong>di</strong><br />
sostanze liquide e gassose” Riunione Annuale del Gruppo Elettronica (GE 2006), Giugno<br />
21-23, Ischia (NA), Italia.<br />
C25. L. De Stefano, L. Rotiroti, I. Rea, S. D’Auria, A. Vitale, M. Rossi, I. Ren<strong>di</strong>na, “A biosensor<br />
for selective identification of analites of social interest based on porous silicon technology”,<br />
IMCS11, July 16-19, Brescia, Italy, 2006.<br />
C26. L. De Stefano, I. Rea, L. Rotiroti, L. Moretti, I. Ren<strong>di</strong>na “Resonant photonic devices<br />
based on porous silicon: a perfect tool for optical sensing”; XX EUROSENSORS,<br />
Settembre 17-20, 2006, Göteborg, Svezia.<br />
C27. I. Rea, L. De Stefano, L. Rotiroti, M. Io<strong>di</strong>ce, I. Ren<strong>di</strong>na; “Optical microsystems based on a<br />
nanomaterial technology” Oral, Nanoscience & Nanotechnology 2006, Novembre 6-9,<br />
2006, Monte Porzio Catone (Rome), Italy.<br />
C28. I. Rea, L. De Stefano, L. Rotiroti, E. De Tommasi, A. Nigro, F.G. Della Corte, I. Ren<strong>di</strong>na,<br />
“Laser oxidation micropatterning of a porous silicon based biosensor for multianalytes<br />
microarrays”, Poster, AISEM 2007, Febbraio 12-14 2007, Napoli, Italy.<br />
C29. L. De Stefano, L. Rotiroti, I. Rea, L. Moretti, I. Ren<strong>di</strong>na, “High-sensitivity optical sensors<br />
based on aperio<strong>di</strong>c multilayer structures”, Poster, AISEM 2007, Febbraio 12-14 2007,<br />
Napoli, Italy.<br />
C30. L. De Stefano, L. Rotiroti, I. Rea, E. De Tommasi, A, Vitale, M. Rossi. I. Ren<strong>di</strong>na and S.<br />
D’Auria “Design and realization of highly stable porous silicon optical biosensor based on<br />
proteins from extremophiles” – Oral presentation. SPIE Europe, Optics and<br />
Optoelectronics, 2007, Prague, Czech Republic.<br />
C31. L. D Stefano, I. Ren<strong>di</strong>na, I. Rea, L. Rotiroti, E. De Tommasi and G. Barillaro “An optical<br />
microsystem based on vertical silicon-air Bragg mirror for liquid substances monitoring” –<br />
Oral presentation. SPIE Europe, Optics and Optoelectronics, 2007, Prague, Czech<br />
Republic.<br />
C32. L. Rotiroti, E. De Tommasi, I. Rea, A. Lamberti, C. Sanges, I. Ren<strong>di</strong>na, P. Arcari and L.<br />
De Stefano “Optical detection of PNA-DNA hybri<strong>di</strong>zation in resonant porous silicon based<br />
devices” – Poster presentation. Europtrode 2008, Dublin, Ireland.<br />
C33. E. De Tommasi, I. Rea. V. Di Sarno, L. Rotiroti, I. Ren<strong>di</strong>na and L. De Stefano “Porous<br />
Silicon Resonant Mirrors Biosensors” – Poster presentation. Porous Semiconductors<br />
Science & Technology, 2008, Mallorca, Spain.<br />
C34. E. De Tommasi, I. Rea. V. Di Sarno, L. Rotiroti, I. Ren<strong>di</strong>na and L. De Stefano “Porous<br />
Silicon Resonant Mirrors for Optical Biosensing” – Oral presentation. 13th National<br />
Conference AISEM, 2008, Rome.
C35. L. Rotiroti, E. De Tommasi, I. Ren<strong>di</strong>na, M. Canciello, G. Maglio, R. Palumbo, and L. De<br />
Stefano,” BIOACTIVE POLIMERS MODIFIED POROUS SILICON NANOSTRUCTURES”,<br />
6th International Conference on Porous Semiconductuctors-Science and Technology,<br />
Poster presentation., Marzo 10-14, 2008, Mallorca, Spagna.<br />
C36. L. Rotiroti, P. Arcari, A. Lamberti, C. Sanges, E. De Tommasi, I. Rea, I. Ren<strong>di</strong>na and L.<br />
De Stefano “Optical detection of PNA-DNA hybri<strong>di</strong>zation in resonant porous silicon based<br />
devices” – Poster presentation. SPIE Photonics Europe, 2008, Strasbourg, France.<br />
C37. L. De Stefano, E. De Tommasi, I. Rea, L. Rotiroti, I. Ren<strong>di</strong>na “Porous Silicon resonant<br />
devices as optical transducers for nanostructured hybrid biosensors”, Invited, New<br />
Frontiers in Micro and Nano Photonics, 23-26 Aprile 2008, Firenze, Italia.<br />
C38. L. Rotiroti, I. Ren<strong>di</strong>na, E. De Tommasi, M. Canciello, G. Maglio, R. Palumbo, and L. De<br />
Stefano, “A nanostructured hybrid device based on polymers infiltrated porous silicon layers<br />
for biotechnological applications”, oral presentation, PPS-24 Meeting The Polymer<br />
Processing Society 24th Annual Meeting June 15-19, 2008, Salerno, ITALY.<br />
C39. E. De Tommasi, I. Rea, V. Di Sarno, L. Rotiroti, I. Ren<strong>di</strong>na and L. De Stefano “Porous<br />
silicon resonant mirrors in the high sensitive monitoring of biological interactions” – Oral<br />
presentation. First Me<strong>di</strong>terranean Photonics Conference of the European Optical Society,<br />
2008, Ischia (Naples), Italy.<br />
C40. L. Rotiroti, I. Ren<strong>di</strong>na, E. De Tommasi, M. Canciello, G. Maglio, R. Palumbo, and L. De<br />
Stefano, A Hybrid Optical Biosensor Based on Polymer Infiltrated Porous Silicon Device, –<br />
Oral presentation, IEEE SENSORS 2008, Lecce (Italy).<br />
PATENTS<br />
Patent n. RM 2005 A 0002280 " An integrated device for the optical determination of the<br />
alcoholic content, in particular for wines and liquors, main procedures of fabrication, and main<br />
procedures of measurements"<br />
SCHOOLS AND WORKSHOPS:<br />
1. 11 giugno 2008 SPETTROSCOPY SEMPLIFIED – THERMO FISCHER SCIENTIFIC, Hotel<br />
Selene, Pomezia<br />
2. 6 febbraio 2008 LA BUONA PRATICA DI PESATA - GWP®, METTLER TOLEDO, Holiday<br />
Inn, Napoli.<br />
3. 17 luglio 2007 MICROSCOPIA A SONDA NELLA SCIENZA DEI MATERIALI, seminario<br />
indetto da 2M STRUMENTI “Veeco”, Portici (Napoli).<br />
4. 26 febbraio-2 marzo 2007 ADVANCED COURSE in design, synthesis, analysis and<br />
evaluation of bioactive molecules, Napoli.<br />
5. 5-6 Luglio 2006 Ciclo <strong>di</strong> seminari sulla Luce <strong>di</strong> Sincrotrone, Napoli.<br />
6. 26-27 giugno 2006 corso “ENZIMI DA IPERTERMOFILI E NANOTECNOLOGIE: ASPETTI<br />
APPLICATIVI”, Napoli.<br />
7. 19-21 Giugno 2006 scuola <strong>di</strong> <strong>dottorato</strong> del Gruppo Elettronica (GE 2006), Benevento (BN),<br />
Italia.<br />
8. 15-17 maggio 2006 corso teorico/pratico “FUNZIONALIZZAZIONE DI SUPERFICI VIA<br />
PLASMA PER APPLICAZIONI BIOTECNOLOGICHE”, Napoli.<br />
9. 12 marzo 2006 Short Course “5 th International Conference on Porous Semiconductuctors-<br />
Science and Technology”, Sitges-Barcellona, Spagna.
Appen<strong>di</strong>x
INSTITUTE OF PHYSICS PUBLISHING JOURNAL OF OPTICS A: PURE AND APPLIED OPTICS<br />
J. Opt. A: Pure Appl. Opt. 8 (2006) S540–S544 doi:10.1088/1464-4258/8/7/S37<br />
Porous silicon-based optical biochips<br />
Luca De Stefano 1 ,LuciaRotiroti 1 ,IlariaRea 1 ,LuigiMoretti 2 ,<br />
Girolamo Di Francia 3 ,Ettore Massera 3 ,Annalisa Lamberti 4 ,<br />
Paolo Arcari 4 ,CarmenSanges 4 and Ivo Ren<strong>di</strong>na 1<br />
1 Institute for Microelectronics and Microsystems, CNR—Department of Naples,<br />
Via PCastellino 111, 80131 Naples, Italy<br />
2 DIMET, University ‘Me<strong>di</strong>terranea’ of Reggio Calabria, Località Feo <strong>di</strong> Vito, 89060 Reggio<br />
Calabria, Italy<br />
3 ENEA—Centro Ricerche Portici, 80055 Portici (NA), Italy<br />
4 DBBM—<strong>Università</strong> <strong>di</strong>Napoli ‘Federico II’, Via S Pansini 5, 80131 Napoli, Italy<br />
Received 28 December 2005, accepted for publication 17 March 2006<br />
Published 12 June 2006<br />
Online at stacks.iop.org/JOptA/8/S540<br />
Abstract<br />
In this paper, we present our work on an optical biosensor for the detection of<br />
the interaction between a DNA single strand and its complementary<br />
oligonucleotide, based on the porous silicon (PSi) microtechnology. The<br />
crucial point in this sensing device is how to make a stable and repeatable<br />
link between the DNA probe and the PSi surface. We have experimentally<br />
compared some functionalization processes which mo<strong>di</strong>fy the PSi surface in<br />
order to covalently fix the DNA probe on it: a pure chemical passivation<br />
procedure, a photochemical functionalization process, and a chemical<br />
mo<strong>di</strong>fication during the electrochemical etching of the PSi. We have<br />
quantitatively measured the efficiency of the chemical bond between the<br />
DNA and the porous silicon surface using Fourier transform infrared<br />
spectroscopy (FT-IR) and light induced photoluminescence emission. From<br />
the results and for its intrinsic simplicity, photochemical passivation seems to<br />
be the most promising method.<br />
The interaction between a label-free 50 µMDNAprobe with<br />
complementary and non-complementary oligonucleotides sequences has been<br />
also successfully monitored by means of optical reflectivity measurements.<br />
Keywords: porous silicon, optical biosensor, surface functionalization,<br />
DNA analysis<br />
(Some figures in this article are in colour only in the electronic version)<br />
1. Introduction<br />
In the past two decades, the biological and me<strong>di</strong>cal fields<br />
have seen great advances in the development of biosensors and<br />
biochips for characterizing and quantifying biomolecules.<br />
Abiosensor can be generally defined as a device that<br />
consists of a biological recognition system, often called<br />
a bioreceptor, and a transducer. The interaction of the<br />
bioreceptor with the analyte is designed to produce an effect<br />
measured by the transducer, which converts the information<br />
into a measurable effect, such as an electrical signal.<br />
Using bioreceptors from biological organisms or receptors<br />
that have been patterned after biological systems, scientists<br />
have developed new methods of biochemical analysis that<br />
exploit the high selectivity of the biological recognition<br />
systems [1].<br />
Porous silicon (PSi) is an almost ideal material as<br />
a transducer due to its porous structure, with hydrogen<br />
terminated surface, having a specific area of the order of 200–<br />
500 m 2 cm −3 ,sothataveryeffective interaction with several<br />
adsorbates is assured. Moreover, PSi is an available and low<br />
cost material, completely compatible with standard integrated<br />
circuit processes. Therefore, it could usefully be employed in<br />
the so-called smart sensors [2]. Recently, much experimental<br />
work, exploiting the noteworthy properties of PSi in chemical<br />
and biological sensing, has been reported [3–6].<br />
PSi optical sensing devices are based on changes of its<br />
physical properties, such as photoluminescence or reflectance,<br />
on exposure to the surroun<strong>di</strong>ng environment. Unfortunately,<br />
this interaction is not specific, so a PSi sensor cannot<br />
<strong>di</strong>scriminate the components of a complex mixture. Some<br />
researchers have chemically or physically mo<strong>di</strong>fied the Si–H<br />
1464-4258/06/070540+05$30.00 © 2006 IOP Publishing Ltd Printed in the UK S540
White light<br />
source<br />
Gas channels<br />
Sealed<br />
or fluxed<br />
reaction<br />
chamber<br />
Glass<br />
Silicon<br />
Optical fibres<br />
Collimator<br />
Objective<br />
Optical<br />
Spectrum<br />
Analyser<br />
(OSA)<br />
Figure 1. The experimental optical set-up.<br />
Porous silicon<br />
microcavities<br />
surface sites in order to enhancethe sensor selectivity through<br />
specific interactions. The common approach is to create a<br />
covalent bond between the PSi surface and the biomolecules<br />
which specifically recognize the target analytes [7, 8].<br />
The reliability of a biosensor strongly depends on the<br />
functionalization process: how fast, simple, homogenous and<br />
repeatable it is. This step is also very important for the stability<br />
of the sensor: it is well known that ‘as-etched’ PSi has a Si–<br />
Hterminatedsurface due to the Si <strong>di</strong>ssolution process which<br />
is very reactive [9]. The substitution of the Si–H bonds with<br />
Si–C ones guarantees a much more stable surface from the<br />
thermodynamic point of view.<br />
In this work, we report some PSi surface mo<strong>di</strong>fication<br />
strategies in order to realize an optical biosensor: the target<br />
is the fabrication of sensitive label-free biosensors, which<br />
are highly requested for applications in high throughput<br />
drug monitoring and <strong>di</strong>sease <strong>di</strong>agnostics; unlabelled analytes<br />
require in fact easier and faster analytical procedures.<br />
2. Materials and methods<br />
Porous silicon is a very attractive material due to the possibility<br />
of fabricating high quality optical structures, either as single<br />
layers, like Fabry–Perot interferometers, or multilayers, such<br />
as Bragg or rugate filters [12]. In this study we used as<br />
a sensor a Fabry–Perot single layer. The PSi layer was<br />
obtained by electrochemical etch in a HF-based solution at<br />
room temperature. A highly doped p + -silicon substrate, 〈100〉<br />
oriented, 0.01 cm resistivity, 400 µmthickwasused. Before<br />
ano<strong>di</strong>zation, the substrate was placed in HF solution to remove<br />
the native oxide.<br />
The <strong>di</strong>electric and physical properties of the PSi layer<br />
were investigated by spectroscopic ellipsometry and scanning<br />
electron microscopy. Both the analysis methods showed the<br />
presence of a top layer on the PSi structure of about 50 nm<br />
thickness, characterized by a very low porosity. The top layer<br />
is due to a hydrogen contamination which passivates boron<br />
and changes the local resistivity. The presence of the top<br />
layer reduces pore infiltration. To remove the contamination<br />
we pre-treated the crystalline silicon substrate before the<br />
electrochemical etching by heating the wafer at 300 ◦ CinN2<br />
atmosphere for 30 min [10]. The ellipsometric measurements<br />
on the pre-treated PSi layer show that the top layer was<br />
eliminated.<br />
Reflectivity (a.u)<br />
1.0<br />
0.8<br />
0.6<br />
0.4<br />
0.2<br />
PSi as etched<br />
After KOH treatment<br />
Ater HF treatment<br />
Porous silicon-based optical biochips<br />
0.0<br />
600 800 1000 1200<br />
Wavelength (nm)<br />
Figure 2. Reflectivity optical spectra of the porous silicon layer after<br />
KOH and HF treatments.<br />
Due to the nanostructured nature of PSi, it is necessary<br />
to improve the pore infiltration of biomolecular probes: with<br />
this aim, we optimized our device by employing a strong base<br />
post-etch process. We immersed the device in an aqueous<br />
ethanol solution, containing millimolar concentration of KOH,<br />
for 15 min. This treatment produces an increase of about 15–<br />
20% in the porosity without affecting the optical quality of the<br />
device [11].<br />
We used a very simple experimental set-up (see figure 1)<br />
to characterize the PSi device from an optical point of view:<br />
atungsten lamp (400 nm
LDeStefanoet al<br />
Reflectance<br />
40<br />
35<br />
30<br />
25<br />
20<br />
15<br />
10<br />
5<br />
CH 2<br />
Si-H<br />
Si-C<br />
0<br />
4000 3500 3000 2500 2000 1500 1000 500<br />
Wavenumber (cm -1 )<br />
with EtMgBr<br />
After post-etch treatments<br />
PSi H<br />
CH3CH2 Mg Br<br />
@ 85°C<br />
PSi CH 2CH 3<br />
Figure 3. FT-IR spectra of the PSi monolayer before and after the pure chemical functionalization process based on EtMgBr, together with<br />
the reaction scheme.<br />
% Reflectance<br />
160<br />
140<br />
120<br />
100<br />
80<br />
60<br />
40<br />
20<br />
2854.7 CH 2<br />
2926 CH 2<br />
2100 Si-H<br />
1563 succinimmide<br />
1038 N-O<br />
1641 succinimmide<br />
1723 O-C=O<br />
0<br />
-20<br />
1288 C-N C=O<br />
4000 3000 2000 1000<br />
Wavenumbers (cm -1 after post-etch treatments<br />
after UV exposition with UANHS<br />
)<br />
PS i H<br />
UANHS<br />
UV<br />
O<br />
H<br />
O<br />
PSi C<br />
H<br />
(CH2) n C O N<br />
O<br />
DNA<br />
PSi<br />
H<br />
C (CH2 ) n<br />
H<br />
O<br />
C<br />
H<br />
N DNA<br />
Figure 4. FT-IR spectra of the PSi monolayer before and after the photoinduced functionalization process based on UV exposure, together<br />
with the reaction scheme.<br />
2.1. Chemical functionalization by Grignard reactives<br />
The chemical functionalization is based on ethyl magnesium<br />
bromide (CH3CH2MgBr) as a nucleophilic agent which<br />
substitutes the Si–H bonds with the Si–C. The reaction was<br />
performed at 85 ◦ Cinaninert atmosphere (argon) to avoid the<br />
deactivation of the reactive, for 8 h. The chip was then washed<br />
with a 1% solution of CF3COOH in <strong>di</strong>ethyl ether, and then with<br />
deionized water and pure <strong>di</strong>ethyl ether. The mo<strong>di</strong>fied surface<br />
chip was characterized by infrared spectroscopy to verify the<br />
efficiency of the method. The FT-IR spectrum and the reaction<br />
scheme are reported in figure 3.<br />
2.2. Photochemical functionalization<br />
The photo-activated chemical mo<strong>di</strong>fication of the PSi surface<br />
was based on the UV exposure of a solution of alkenes which<br />
bring some carboxylic acid groups. The PSi chip was precleaned<br />
in an ultrasonic acetone bath for 10 min then washed<br />
in deionized water. After being dried in a N2 stream, it was<br />
imme<strong>di</strong>ately covered with 10% N-hydroxysuccinimide ester<br />
(UANHS) solution in CH2Cl2. The UANHS was synthesized<br />
in house as described in [7]. This treatment results in covalent<br />
attachment of UANHS to the PSi surface, clearly shown in the<br />
FT-IR spectrum, reported in figure 4 together with the reaction<br />
scheme. The chip was then washed in <strong>di</strong>chloromethane in an<br />
S542<br />
ultrasonic bath for 10 min and rinsed in acetone to remove any<br />
adsorbed alkene from the surface.<br />
2.3. Functionalization during the etch process<br />
We also tried to chemically mo<strong>di</strong>fy the surface of the PSi<br />
by <strong>di</strong>rectly using a functionalizing agent during the etching<br />
process. We introduced some organic acids (eptinoic and<br />
pentenoic acid with concentrations from 0.4 M to 3 M) in an<br />
electrochemical cell in the presence of a <strong>di</strong>lute HF solution<br />
(HF:EtOH=1:2). Inthiscase a current density of 60 mA cm −2<br />
was applied to etch an area of 0.07 cm 2 .TheFT-IR spectrum<br />
is reported in figure 5.<br />
3. Experimental results<br />
We stu<strong>di</strong>ed the infrared spectrum in each case of functionalization<br />
method in order to determine the best reaction con<strong>di</strong>tions<br />
and the maximum yield in the Si–H/Si–C substitution.<br />
The post-etch KOH treatment, which we used to increase the<br />
porosity and to improve the pore infiltration, removes most of<br />
Si–H bonds on the fresh PSi surface and also induces its oxidation.<br />
To assure the formation of the Si–C bonds, we had<br />
first to restore the Si–H bonds by rinsing the PSi in a very <strong>di</strong>lute<br />
HF-based solution for 30 s. FT-IR spectroscopy confirmed<br />
the presence of Si–H bonds on the PSi surface after all pretreatments.
% Reflectance<br />
110<br />
100<br />
90<br />
80<br />
70<br />
60<br />
50<br />
40<br />
30<br />
20<br />
10<br />
before functionalization<br />
after functionalization<br />
2915.4CH2<br />
2383.7 CO2<br />
2348.1 CO2<br />
3500 3000 2500 2000 1500 1000<br />
Wavenumber (cm –1 )<br />
2104SiH<br />
2064.2SiH<br />
1709.4COOH<br />
1075 SiO<br />
1260 CH=CH<br />
1032 SiO<br />
903<br />
867.4<br />
Figure 5. FT-IR spectra of the PSi monolayer before (solid line) and<br />
after (dashed line) functionalization during the etching process.<br />
For each method we stu<strong>di</strong>ed, we report the infrared<br />
spectrum (figures 3–5). In all of them, the characteristic peaks<br />
of the Si–C bonds can be observed, even if with <strong>di</strong>fferent<br />
intensities. In the case of the pure chemical functionalization<br />
method, the FT-IR spectrum (figure 3) showsanincomplete<br />
substitution of the Si–H bonds, since their characteristic peaks<br />
(at 2100 cm −1 )are still present.<br />
In figure 4 we can also easily <strong>di</strong>stinguish all the<br />
characteristic peaks of the reactive (UANHS) immobilized on<br />
the PSi surface. The infrared analysis of the chip prepared<br />
by <strong>di</strong>rect functionalization during the etch process, shows<br />
a–COOH peak, whose intensity increases on increasing the<br />
acid concentration, as reported in figure 5.<br />
Among the three procedures experimented, we believe<br />
that the photoinduced method is the best one for several<br />
reasons: the relaxed reaction con<strong>di</strong>tions (atmospheric pressure<br />
and room temperature); the shorter reaction time; and the best<br />
reaction yield (from FT-IR measurements). This last result is<br />
somewhat expected because the reactive considered has a socalled<br />
‘outgoing group’ which promotes its substitution with<br />
the amine group of the DNA probe.<br />
In view of these considerations, on each chip with a<br />
photochemical mo<strong>di</strong>fied surface we incubated, overnight, a<br />
luminescent DNA probe. After washing, we measured the<br />
photoluminescence emission with a microscope equipped with<br />
a CCD. As a reference sample, we used the signal coming from<br />
chips not washed after labelled DNA incubation.<br />
The PL measurements showed that the photochemical<br />
mo<strong>di</strong>fication of the PSi surface is very efficient: we registered<br />
52 ± 8counts from treated chips, 5 ± 1counts from the<br />
unfunctionalized chips, and 63 ± 6counts from the reference<br />
samples. The count values are the arithmetic mean over<br />
three <strong>di</strong>fferent chip samples and the errors are the standard<br />
deviations. A total yield of about 83% can be thus estimated.<br />
The optical monitoring of DNA–cDNA hybri<strong>di</strong>zation is a<br />
two-step procedure: first, we registered the optical spectrum<br />
of the PSi layer after the UANHS and probe immobilization<br />
on the chip surface and, then, after the hybri<strong>di</strong>zation with<br />
the cDNA. Each step of the chip preparation increases the<br />
optical path in the reflectivity spectrum recorded, due to the<br />
substitution of the air into the pores by organic and biological<br />
compounds, so that several red-shifts can be observed after the<br />
Reflectivity (a.u)<br />
1.0<br />
0.8<br />
0.6<br />
0.4<br />
0.2<br />
Porous silicon-based optical biochips<br />
PSi with DNA probe<br />
PSi with c-DNA<br />
600 800 1000<br />
Wavelength (nm)<br />
1200<br />
Figure 6. Reflectivity optical spectra of the PSi device after DNA<br />
probe immobilization and after cDNA hybri<strong>di</strong>zation.<br />
functionalization (10 nm), and after the covalent bin<strong>di</strong>ng of the<br />
DNAsingle strand (probe) with the linker on the porous silicon<br />
layer (31 nm). The interaction with the cDNA is also detected<br />
as a shift in wavelength of the optical path (33 nm). In figure 6<br />
the reflectivity spectra of only these last two steps are reported<br />
for the sake of clarity. A control measurement was made using<br />
an ncDNA sequence: a very small shift (less than 2 nm) was<br />
recorded in the reflectivity spectrum with respect to the one<br />
obtained after probe linking.<br />
4. Conclusion<br />
In conclusion, we have presented our experimental results<br />
of the study of several procedures of porous silicon surface<br />
functionalization that are a fundamental step in realizing an<br />
optical porous silicon microsensor for DNA–cDNA interaction<br />
detection. The pore infiltration by the DNA single-strand<br />
solution was improved by introducing two more steps during<br />
the realization process: a pre-etch cleaning of the crystal<br />
substrate to remove hydrogen contamination; and an alkaline<br />
post-etching process that increases the device porosity. Among<br />
several surface functionalization methods used to immobilize<br />
in the PSi matrix a DNA probe which selectively recognizes<br />
its complementary sequence, we chose a photoinduced one<br />
because of the simplicity of the reaction con<strong>di</strong>tions and<br />
the well performing results. We also demonstrated that<br />
an optical Fabry–Perot interferometer based on a PSi layer<br />
can be effectively used as a transducer of the DNA–cDNA<br />
hybri<strong>di</strong>zation interaction.<br />
Acknowledgments<br />
This work was realised in the frame of the CRdC-NTAP and<br />
in the frame of the MIUR FIRB Project 2003 Costituzione<br />
<strong>di</strong> un laboratorio de<strong>di</strong>cato all’analisi delle interazioni ligandorecettore<br />
me<strong>di</strong>ante biochip <strong>di</strong> silicio.<br />
References<br />
[1] Sadana A 2002 Engineering Biosensors (San Diego, CA:<br />
Academic)<br />
S543
LDeStefanoet al<br />
[2] De Stefano L, Moretti L, Ren<strong>di</strong>na I, Rossi A M and<br />
Tundo S 2004 Smart optical sensors for chemical substances<br />
based on porous silicon technology Appl. Opt. 43 167–72<br />
[3] Dancil K-P S, Greiner D P and Sailor M J 1999 A porous<br />
silicon optical biosensor: detection of reversible bin<strong>di</strong>ng of<br />
IgG to a protein A-mo<strong>di</strong>fied surface J. Am. Chem. Soc.<br />
121 7925–30<br />
[4] De Stefano L, Moretti L, Rossi A M, Rocchia M, Lamberti A,<br />
Longo O, Arcari P and Ren<strong>di</strong>na I 2004 Optical sensors for<br />
vapors, liquids, and biological molecules based on porous<br />
silicon technology IEEE Trans. Nanotech. 3 49–54<br />
[5] Dattelbaum J D and Lakowicz J R 2001 Optical determination<br />
of glutamine using a genetically engineered protein Anal.<br />
Biochem. 291 89–95<br />
[6] Bradford M M 1976 A rapid and sensitive method for the<br />
quantitation of microgram quantities of protein utilizing the<br />
principle of protein-dye bin<strong>di</strong>ng’ Anal. Biochem. 72 248–54<br />
S544<br />
[7] Yin H B, Brown T, Gref R, Wilkinson J S and Melvin T 2004<br />
Chemical mo<strong>di</strong>fication and micropatterning of Si(100) with<br />
oligonucleotides Microelectron. Eng. 73/74 830–6<br />
[8] Hart B R, Letant S E, Kane S R, Ha<strong>di</strong> M Z, Shields S J and<br />
Reynolds J G 2003 New method for attachment of<br />
biomolecules to porous silicon Chem. Commun. 322–3<br />
[9] Canham L (ed) 1997 Properties of Porous Silicon (London:<br />
IEE INSPEC)<br />
[10] Chamard V, Dolino G and Muller F 1998 Origin of a parasitic<br />
surface film on p+ type porous silicon J. Appl. Phys.<br />
84 6659–66<br />
[11] DeLouise L A and Miller B L 2004 Optimization of<br />
mesoporous silicon microcavities for proteomic sensing<br />
Mater. Res. Soc. Symp. Proc. 782 A5.3.1–A5.3.7<br />
[12] Theiss W 1997 Optical properties of porous silicon Surf. Sci.<br />
Rep. 29 91–192
IOP PUBLISHING JOURNAL OF PHYSICS: CONDENSED MATTER<br />
J. Phys.: Condens. Matter 19 (2007) 395008 (7pp) doi:10.1088/0953-8984/19/39/395008<br />
Optical microsystems based on a nanomaterial<br />
technology<br />
1. Introduction<br />
LDeStefano 1 ,LRotiroti 1,2 ,IRea 1,3 ,MIo<strong>di</strong>ce 1 and I Ren<strong>di</strong>na 1<br />
1 National Council of Research-Institute for Microelectronic and Microsystems-Department of<br />
Naples, Via P Castellino 111, 80131 Naples, Italy<br />
2 Department of Organic Chemistry and Biochemistry, ‘Federico II’ University of Naples, Via<br />
Cinthia, 4-80126 Naples, Italy<br />
3 Department of Physics, ‘Federico II’ University of Naples, Via Cinthia, 4-80126 Naples, Italy<br />
E-mail: luca.destefano@na.imm.cnr.it<br />
Received 13 February 2007<br />
Published 30 August 2007<br />
Online at stacks.iop.org/JPhysCM/19/395008<br />
Abstract<br />
In this work, we present an optical sensor for quantitative determination of the<br />
alcohol content in hydro-alcohol mixtures, realized by using porous silicon<br />
(PSi) nanotechnology. The device is an oxi<strong>di</strong>zed PSi micro-cavity (PSMC)<br />
constituted by a Fabry–Perot layer between two <strong>di</strong>stributed Bragg reflectors.<br />
Due to the capillary condensation, a red shift of the PSMC reflectivity spectrum<br />
is observed on exposure to vapour mixtures. The phenomenon is completely<br />
reversible. Moreover, to reduce the analysis time, we have designed the<br />
integration of the sensor in a thermally controlled lab-on-chip, by merging PSi<br />
and ano<strong>di</strong>c bon<strong>di</strong>ng technologies. Numerical calculations have been performed<br />
to study the thermal behaviour of the integrated device.<br />
(Some figures in this article are in colour only in the electronic version)<br />
The fast and intensive development of the lab-on-chip (LOC) for sensing applications is due<br />
to several factors that are essential for routine analytical determinations, such as very limited<br />
sample consumption and short analysis time, which can be obtained by measuring a transient<br />
signal in a flow-through detector. Due to their intrinsic features, in particular contactless<br />
monitoring capability, optical sensors are very attractive for many application fields.<br />
In recent years, many devices based on porous silicon (PSi) technology have been realized<br />
for several sensing applications [1, 2]. PSi, fabricated by electrochemical etching of a silicon<br />
(Si) wafer in HF-based solution, is an ideal material for an optical transducer, due to its spongelike<br />
nanostructure (specific surfaces up to 500 m 2 cm −3 )[3], that assures an effective interaction<br />
with several chemicals and biological molecules. On exposure to chemical substances, several<br />
physical parameters, such as refractive index, photoluminescence, and electrical conductivity,<br />
0953-8984/07/395008+07$30.00 © 2007 IOP Publishing Ltd Printed in the UK 1
J. Phys.: Condens. Matter 19 (2007) 395008 L De Stefano et al<br />
Figure 1. Experimental optical set-up used for the sensing experiments.<br />
change drastically. It is also possible to realize in a single process multilayered structures that<br />
exhibit a high selective optical response, which is very attractive for sensing [4].<br />
Silicon–glass ano<strong>di</strong>c bon<strong>di</strong>ng (AB) is one of the most promising techniques in the<br />
micromachining field; it easily allows the encapsulation and sealing on silicon chips of three<strong>di</strong>mensional<br />
microflui<strong>di</strong>c structures, such as channels, chambers, cavities and other complex<br />
gas and liquid routes. Ad<strong>di</strong>tionally, glass transparency at optical wavelengths enables simple,<br />
but highly accurate, alignment of pre-patterned or structured glass and silicon wafers.<br />
We have recently demonstrated the compatibility of PSi and AB technologies for the<br />
realization of sensing microcomponents for LOC applications [5, 6]. The integration was<br />
not easy or straightforward: the microfabrication techniques are typically optimized for the<br />
microelectronic industry and are not well appropriated or compatible with the non-standard<br />
materials such as the ones used in biological or chemical sensing.<br />
We present here a simple optical sensor for the analysis of hydro-alcohol mixtures, based<br />
on PSi nanotechnology. Faster time analysis can be obtained by integrating the silicon–glass<br />
sensor for LOC applications by merging the PSi and AB technologies.<br />
2. Materials and methods<br />
For this study, we fabricated a λ/2 Fabry–Perot optical microcavity sandwiched between<br />
two Bragg reflectors of seven periods each, realized by alternating layers with high and low<br />
refractive in<strong>di</strong>ces. The porous silicon microcavity (PSMC) was obtained by electrochemical<br />
etch in a HF-based solution (50 wt% HF:ethanol = 1:1) in dark light and at room temperature.<br />
A current density of 400 mA cm −2 for 0.8 s was applied to a highly doped p + -type standard<br />
silicon wafer, 〈100〉 oriented, with 10 m cm resistivity, 400 μm thick, producing the highrefractive<br />
index layer (with a porosity of 75%), while a current density of 80 mA cm −2 was<br />
applied for 1.9 s in the case of the low-index layer (with a porosity of 67%). The λ/2 layer<br />
is a low-refractive-index one. This structure has a characteristic resonance peak at 1110 nm<br />
in the middle of a 280 nm wide stop band. To ensure the infiltration of aqueous solution, the<br />
device, which (as etched) is hydrophobic due to the presence of SiH bonds on its surface, was<br />
made hydrophilic by thermal oxidation at 1000 ◦ C for 30 min. The optical set-up required<br />
for our sensing experiments was very simple (see figure 1). The PSMC was placed in a test<br />
chamber equipped with a glass window. A white light source interrogated the sensor through an<br />
optical fibre and a collimator. The reflected beam was collected by an objective, coupled into<br />
2
J. Phys.: Condens. Matter 19 (2007) 395008 L De Stefano et al<br />
Figure 2. Optical spectra of the PSMC as etched and after thermal oxidation.<br />
a multimode fibre, and then <strong>di</strong>rected in an optical spectrum analyser (Ando model AQ6315A).<br />
The reflectivity spectra were measured over the range 600–1600 nm with a resolution of 0.2 nm.<br />
We stu<strong>di</strong>ed the resonance peak shift on exposure to an ethanol/deionized water (DI) binary<br />
mixture at <strong>di</strong>fferent concentrations, placing a drop of solution in the test chamber. Due to<br />
the capillary condensation of the solution vapours into the pores of the PSMC, a change in<br />
its reflectivity spectrum was observed: the partial substitution of air by liquid in each layer<br />
determines an increase of the average refractive index of the microcavity and, consequentially,<br />
a red shift of the spectrum. Measurements are stored when equilibrium con<strong>di</strong>tions are reached,<br />
i.e. when the signal does not change any more in time. The phenomenon is completely<br />
reversible and reproducible.<br />
3. Experimental results and <strong>di</strong>scussion<br />
In figure 2, the reflectivity spectra of the PSMC as etched and after thermal oxidation are<br />
reported. The oxidation process causes a blue shift in the spectrum of 110 nm due to the<br />
lower value of the SiO2 refractive index (nSiO2 ≈ 1.45) with respect to the Si refractive index<br />
(nSi ≈ 3.6). We have not observed degradation in cavity quality (Q = λ/λ ≈ 70 in both<br />
cases).<br />
A very simple method, commonly used in analytical chemistry, in quantitative<br />
determinations of binary mixtures is the external standard method: calibration curves can be<br />
obtained by measuring the parameter under monitoring, such as an optical intensity, while<br />
the content of one component varies between 0 and 100%. In our case, the curve has been<br />
obtained by plotting the peak shift measured as a function of the volume percentage of ethanol<br />
in the binary mixture, as shown in figure 3. The errors on the experimental points are the<br />
standard deviations on five measurements, using the same chip and the same hydro-alcohol<br />
solution. After the calibration, unknown concentrations of the same mixture can be quantitative<br />
determined with a single measurement of the peak shift: the volume concentrations of both<br />
components can be picked out from the curve obtained by interpolating the experimental<br />
calibration data. In order to limit the wide error bars and the time necessary for the PSMC<br />
to reach equilibrium with the mixture (by limiting the volume), integration in a single chip of<br />
the device is proposed.<br />
The compatibility of the porous silicon technology with the standard integrated circuit<br />
fabrication processes allows the design of complex microsystems in which the porous silicon<br />
3
J. Phys.: Condens. Matter 19 (2007) 395008 L De Stefano et al<br />
Figure 3. The calibration curve in case of an ethanol/DI binary mixture.<br />
Figure 4. Sketch of the alcohol microsensor.<br />
optical device is the transduction element. In figure 4, we report a rough and somehow intuitive<br />
sketch of the alcohol microsensor. The PSMC can be integrated in an LOC consisting of two<br />
μ-chambers thermally controlled by integrated heaters, and sealed by a cover glass after the<br />
ano<strong>di</strong>c bon<strong>di</strong>ng process. The mixture is injected in the left μ-chamber by means of an<br />
appropriate microsyringe and here the integrated heater induces the hydro-alcohol mixture<br />
evaporation. The vapours produced <strong>di</strong>ffuse and reach the second μ-chamber, cooled by another<br />
integrated element, where the PSMC is and the capillary condensation happens.<br />
4. Numerical calculations<br />
Before procee<strong>di</strong>ng towards the real fabrication steps of the whole device depicted in figure 4,<br />
it is mandatory to simulate its thermal behaviour since this is one of the key features in<br />
the measurement process. The precision and the accuracy in the quantitative determination<br />
of the mixture concentration strongly depend on the thermodynamic parameters such as the<br />
equilibrium time, i.e. the time required to have the same evaporation and condensation into<br />
the PSMC, and the temperature <strong>di</strong>stribution in each layer used to realize the LOC. In order<br />
to simulate the thermal response of the device, a finite element model has been implemented<br />
with the commercial FEM code ANSYS Multiphysics 9.0 [7]. We have restricted the analysis<br />
to a two-<strong>di</strong>mensional (2D) model correspon<strong>di</strong>ng to the device middle cross section sketched<br />
(not to scale) in figure 4. Ignoring the minimal border effects at the device ends, a temperature<br />
4
J. Phys.: Condens. Matter 19 (2007) 395008 L De Stefano et al<br />
Figure 5. Temperature <strong>di</strong>stribution in the device when the tank μ-chamber is heated with a power<br />
density of 1.0Wcm −2 .<br />
Table 1. Physical parameters for lab-on-chip materials used in thermal simulations.<br />
Thermal conductivity Density Specific heat<br />
Material (Wm −1 K −1 ) (kg m −3 ) (Jkg −1 K −1 )<br />
Silicon 156 2640 917<br />
Glass 1.05 2500 840<br />
Water 0.58 1000 4186<br />
Air 0.025 1.225 1005<br />
gra<strong>di</strong>ent in fact exists only in this plane. This hypothesis can be justified by the observation that,<br />
for the proposed geometry, the aspect ratio between its width (46 mm) and thickness (1.5 mm)<br />
is about 1/30. Of course, a full 3D model could be more accurate, though it is very time<br />
consuming during the FEM analysis. The structure has been meshed with 2242 isoparametric<br />
linear elements of the type plane 55, with a mesh refinement in the active regions of the device.<br />
In table 1 we have reported the main physical parameters of the materials used for ANSYS<br />
thermal simulations.<br />
The results of the numerical simulations depend on the boundary con<strong>di</strong>tions fixed: we have<br />
supposed a natural convective heat exchange on the upper face of the glass slide which is in<br />
contact with air at 20 ◦ C. We have assumed an exchange coefficient of 27 W m −2 K −1 [8].<br />
Lateral and lower surfaces have been considered a<strong>di</strong>abatic; this model could be a good<br />
approximation of a device mounted on a plastic case. The lower surface of the left μ-chamber,<br />
which is our tank for the liquid substance, is kept hot at constant temperature by a heater<br />
element. The right μ-chamber is cooled by a Peltier μ-cell with a maximum heat flux equal<br />
to 0.4 Wcm −2 (i.e. micro modules from TE Technology Inc., USA). In figure 5 we report the<br />
2D steady-state thermal <strong>di</strong>stribution, when the tank chamber is heated with a power density of<br />
1.0 Wcm −2 , correspon<strong>di</strong>ng to a maximum temperature on the hot side of about 100 ◦ C. From<br />
this calculation it is clear that the PSMC region reaches a temperature of about 50 ◦ C, which<br />
is a little bit higher of the evaporation temperature of ethanol; in this case the alcohol cannot<br />
condense inside the pores of the PSMC.<br />
5
J. Phys.: Condens. Matter 19 (2007) 395008 L De Stefano et al<br />
Figure 6. Temperature <strong>di</strong>stribution versus the lateral <strong>di</strong>mension of the device for two <strong>di</strong>fferent<br />
heating power densities.<br />
Figure 7. Temperature dynamic in the condensation chamber for two <strong>di</strong>fferent heating power<br />
densities.<br />
We have repeated the simulation considering a lower heating power density, 0.4 Wcm −2 ,<br />
correspon<strong>di</strong>ng to heating the tank μ-chamber to 35 ◦ C. In this case, the PSMC is cooled,<br />
thanks to the Peltier module, at about 5 ◦ C: the analysis chamber temperature decreases quite<br />
linearly with the heat source temperature, as expected. These results are shown schematically in<br />
figure 6, where the 1D lateral temperature <strong>di</strong>stribution, calculated at the silicon–glass interface,<br />
is reported versus the lateral <strong>di</strong>mension of the device for the two <strong>di</strong>fferent heating power<br />
densities. We have also calculated the PSMC average temperature as a function of time for<br />
the two <strong>di</strong>fferent heating power densities. The transient solver Antype4 wasusedtoperform<br />
6
J. Phys.: Condens. Matter 19 (2007) 395008 L De Stefano et al<br />
the analysis, using the multi load-step method provided in the code. From the plot in figure 7,<br />
it is well evident that, in the case of a heating power of 0.4Wcm −2 , the porous silicon–vapour<br />
system is at thermal equilibrium after about 50 s. In contrast, if we apply a heating power of<br />
1.0Wcm −2 , a much longer time is necessary.<br />
5. Conclusions<br />
We have presented a simple and well performing optical sensor for ethanol/DI binary mixture<br />
analysis, based on PSMC. Due to capillary condensation, the reflectivity spectrum of the device<br />
shifts towards higher wavelengths on exposure to vapours. The shift is characteristic of the<br />
concentration of each component of the mixture. The sensing process is completely reversible.<br />
In order to limit the time needed for the PSMC to reach equilibrium with the mixture, we have<br />
proposed the integration of the device in a single chip that is thermally controlled. The thermal<br />
behaviour of the designed LOC has been stu<strong>di</strong>ed by proper FEM numerical calculations.<br />
References<br />
[1] Dancil K-P S, Greiner D P and Sailor M J 1999 J. Am. Chem. Soc. 121 7925–30<br />
[2] De Stefano L, Moretti L, Rossi A M, Rocchia M, Lamberti A, Longo O, Arcari P and Ren<strong>di</strong>na I 2004 IEEE Trans.<br />
Nanotech. 3 49–54<br />
[3] Herino R, Bomchil G, Barla K, Bertrand C and Ginoux J L 1987 J. Electrochem. Soc. 134 1994–2000<br />
[4] Theiss W 1997 Surf. Sci. Rep. 29 95–192<br />
[5] Malecki K and Della Corte F G 2005 Proc. Micromachining and Microfabrication Process Technology X; SPIE<br />
5715 180–9<br />
[6] De Stefano L, Malecki K, Della Corte F G, Moretti L, Rea I, Rotiroti L and Ren<strong>di</strong>na I 2006 Sensors 6 680–7<br />
[7] Ansys 9.0 user manual see http://www.ansys.com<br />
[8] Sturm J C, Wilson W and Io<strong>di</strong>ce M 1998 IEEE J. Sel. Top. Quantum Electron. 4 75–82<br />
7
Extremophiles<br />
DOI 10.1007/s00792-006-0058-6<br />
ORIGINAL PAPER<br />
Enzymes and proteins from extremophiles as hyperstable probes<br />
in nanotechnology: the use of D-trehalose/D-maltose-bin<strong>di</strong>ng<br />
protein from the hyperthermophilic archaeon Thermococcus<br />
litoralis for sugars monitoring<br />
Luca De Stefano Æ Annalisa Vitale Æ Ilaria Rea Æ Maria Staiano Æ<br />
Lucia Rotiroti Æ Tullio Labella Æ Ivo Ren<strong>di</strong>na Æ Vincenzo Aurilia Æ<br />
Mose’ Rossi Æ Sabato D’Auria<br />
Received: 30 October 2006 / Accepted: 29 November 2006<br />
Ó Springer 2007<br />
Abstract The D-trehalose/D-maltose-bin<strong>di</strong>ng protein<br />
(TMBP), a monomeric protein of 48 kDa, is one<br />
component of the trehalose and maltose (Mal) uptake<br />
system. In the hyperthermophilic archaeon Thermococcus<br />
litoralis, this is me<strong>di</strong>ated by a protein-dependent<br />
ATP-bin<strong>di</strong>ng cassette system transporter. The gene<br />
co<strong>di</strong>ng for a thermostable TMBP from the archaeon T.<br />
litoralis has been cloned, and the recombinant protein<br />
has been expressed in E. coli. The recombinant TMBP<br />
has been purified to homogeneity and characterized. It<br />
exhibits the same functional and structural properties<br />
as the native one. In fact, it is highly thermostable and<br />
binds sugars, such as maltose, trehalose and glucose,<br />
with high affinity. In this work, we have immobilized<br />
TMBP on a porous silicon wafer. The immobilization<br />
of TMBP to the chip was monitored by reflectivity and<br />
Fourier Transformed Infrared spectroscopy. Furthermore,<br />
we have tested the optical response of the protein-Chip<br />
complex to glucose bin<strong>di</strong>ng. The obtained<br />
Communicated by D. A. Cowan.<br />
The authors wish to de<strong>di</strong>cate this work to Prof. Ignacy<br />
Gryczynski, University of North Texas, TX, USA, for his<br />
outstan<strong>di</strong>ng contribution to the development of new sensing<br />
methodologies.<br />
L. De Stefano I. Rea L. Rotiroti I. Ren<strong>di</strong>na<br />
Istituto <strong>di</strong> Microelettrica e Microsistemi, CNR, Napoli, Italy<br />
A. Vitale M. Staiano T. Labella V. Aurilia<br />
M. Rossi S. D’Auria (&)<br />
Istituto <strong>di</strong> Biochimica delle Proteine, CNR,<br />
Via Pietro Castellino 111, 80131 Napoli, Italy<br />
e-mail: s.dauria@ibp.cnr.it<br />
data suggest the use of this protein for the design of<br />
advanced optical non-consuming analyte biosensors for<br />
glucose detection.<br />
Keywords Trehalose/maltose-bin<strong>di</strong>ng protein<br />
Biosensors Porous silicon Glucose Diabetes<br />
Archaeon<br />
Introduction<br />
The general interest in biomolecules, isolated from<br />
thermophilic organisms was originally due to the biotechnological<br />
advantages offered by the utilization of<br />
these highly stable molecules in industrial processes<br />
(Brock 1985). In fact, enzymes and proteins isolated<br />
from thermophilic microorganisms exhibit a high stability<br />
in con<strong>di</strong>tions usually used to denature proteins:<br />
high temperature, ionic strength, extreme pH-values,<br />
elevated concentration of detergents and chaotropic<br />
agents (Jaenicke 1994). In ad<strong>di</strong>tion to the potential<br />
industrial applications, it is important to highlight that<br />
proteins and enzymes that are stable and active over<br />
100°C represent good models to shed light on the<br />
molecular adaptation of life at high temperature<br />
(D’Auria et al. 1998).<br />
The D-trehalose/D-maltose-bin<strong>di</strong>ng protein<br />
(TMBP) is one component of the trehalose (Tre) and<br />
maltose (Mal) uptake system, which, in the hyperthermophilic<br />
archaeon Thermococcus litoralis, is me<strong>di</strong>ated<br />
by a protein-dependent ATP-bin<strong>di</strong>ng cassette<br />
(ABC) system transporter (Xavier 1996). TMBP from<br />
T. litoralis is a monomeric 48 kDa two-domain macromolecule<br />
containing 12 tryptophan residues (Diez<br />
et al. 2001). TMBP shares common structural motifs<br />
123
with a number of other sugar-bin<strong>di</strong>ng proteins. This<br />
class of biomolecules is composed of proteins, whose<br />
structure consists of two globular domains connected<br />
by a hinge region made of two or three short peptide<br />
segments. The two domains are formed by non-contiguous<br />
polypeptide stretches and exhibit similar tertiary<br />
structure. The ligand-bin<strong>di</strong>ng site is located in the<br />
deep cleft between the two domains and the bin<strong>di</strong>ng is<br />
accompanied by a movement of the two lobes as well<br />
as by conformational changes in the hinge region (Sun<br />
et al. 1998).<br />
The gene co<strong>di</strong>ng for the thermostable TMBP from<br />
the archaeon T. litoralis has been cloned, and the recombinant<br />
protein has been expressed in E. coli<br />
(Horlacher et al. 1998). The recombinant TMBP has<br />
been purified to homogeneity and characterized. It<br />
exhibits the same functional and structural properties<br />
as the native one (Horlacher et al. 1998).<br />
In a recent work (Herman et al. 2006), we have<br />
investigated the bin<strong>di</strong>ng of several sugars to TMBP by<br />
time-resolved fluorescence spectroscopy, and molecular<br />
dynamics methods. We found that TMBP is also<br />
able to bind glucose molecules. Since human blood<br />
does not contain Tre and Mal, it is not outrageous to<br />
envisage the use of the TMBP as a probe for the design<br />
of a minimally invasive biochip for glucose detection.<br />
In this work, we have extended our investigation on<br />
the utilization of TMBP for sensing glucose by immobilizing<br />
the protein on a porous silicon chip. The<br />
TMBP-based chip has been characterized and tested as<br />
regards the response to glucose. The obtained results<br />
are <strong>di</strong>scussed.<br />
Materials and methods<br />
D-Glucose and all the other chemicals used in the<br />
present study were from Sigma. All commercial samples<br />
were of the best available quality.<br />
Purification of TMBP and protein concentration<br />
determination<br />
An mount of 20 g wet weight of the E. coli cell pellet<br />
containing the expressed TMBP were resuspended in<br />
50 mM Tris–HCl, pH 7.5, 500 mM NaCl (100 ml),<br />
ruptured by a French pressure cell at 16,000 psi, and<br />
centrifuged for 15 min at 19,000 g. The supernatant<br />
was heated to 90°C for 20 min and centrifuged for<br />
15 min at 19,000 g. The supernatant was collected and<br />
<strong>di</strong>alyzed for 12 h against 10 mM Tris–HCl, pH 7.5 at<br />
4°C. The solution was loaded onto a DEAE column<br />
previously equilibrated in 10 mM Tris–HCl, pH 7.5.<br />
123<br />
After washing the column with 10 mM Tris–HCl, pH<br />
7.5, TMBP was eluted by a NaCl gra<strong>di</strong>ent (0.0–1.0 M<br />
NaCl). Centricon 30 concentrator (Amicon, MA,<br />
USA) was used to concentrate and to <strong>di</strong>alyze TMBP<br />
against 10 mM Tris–HCl, pH 7.5. The purified protein<br />
was passed through a Blue A affinity column (Amersham,<br />
NJ, USA) in order to obtain DNA-free TMBP.<br />
The purity of TMBP was verified by SDS-PAGE and<br />
absorption spectra.<br />
The protein concentration was determined by the<br />
method of Bradford (Bradford 1976) with bovine<br />
serum albumin as standard on a double beam Cary<br />
1E spectrophotometer (Varian, Mulgrade, Victoria,<br />
Australia).<br />
Chip preparation and spectroscopic<br />
characterization<br />
In this study, we used as sensor an apo<strong>di</strong>zed Bragg<br />
reflector obtained by alternating high (A) refractive<br />
index layers (low-porosity) and low (B) refractive index<br />
layers (high-porosity) whose thicknesses satisfy the<br />
following relationship: n AdA +nBdB =mk/2, where m<br />
is an integer and k is the Bragg resonant wavelength.<br />
The <strong>di</strong>stributed Bragg reflector (DBR) was produced<br />
by electrochemical etch of a highly doped p + -silicon,<br />
Æ100æoriented, 0.04 W cm resistivity, 400 lm thick, in a<br />
HF-based solution. The silicon was etched using a<br />
30 wt.%HF/ethanol solution in dark and at room<br />
temperature. Before ano<strong>di</strong>zation the substrate was<br />
placed in HF solution to remove the native oxide. A<br />
current density of 148 mA/cm 2 for 2.31 s was applied<br />
to obtain the low refractive index layer (effective<br />
refractive index n L @ 1.505, thickness d L @ 598 nm)<br />
with a porosity of 72%, while one of 110 mA/cm 2 was<br />
applied for 2.34 s for the high index layer (nH @ 1.585,<br />
thickness d H @ 568 nm) with a porosity of 69%. This<br />
structure is characterized by a first order (m =1)<br />
resonance at 3,600 nm. In Fig.1, the experimental<br />
White Light<br />
Source<br />
OSA<br />
Y fiber<br />
fiber<br />
Porous Silicon Chip<br />
Extremophiles<br />
Fig. 1 Experimental setup used to measure the optical reflectivity<br />
spectrum of porous silicon chip
Extremophiles<br />
setup, used to measure the reflectivity spectrum, is<br />
reported. We have used as source a white light <strong>di</strong>rected<br />
on the porous silicon chip through a Y-fiber. The same<br />
fiber was used to guide the output signal to the optical<br />
spectrum analyzer. The spectrum was measured over<br />
the range 800–1,600 nm with a resolution of 0.2 nm.<br />
The porous silicon surface after electrochemical etching<br />
is hydrogenated and for this reason, very reactive<br />
and thermodynamically instable; in order to realize a<br />
sensor more stable and able to link the biological<br />
probes it is necessary to substitute the Si–H bonds with<br />
the Si–C or Si–O–C ones.<br />
To promote a covalent link between the porous<br />
silicon and the TMBP, we have exploited a three-step<br />
functionalization process, based on the chemical passivation<br />
of the PSi surface after oxidation. We have<br />
first thermally treated the PSi DBR, in O 2 atmosphere,<br />
at 1,000°C for 30 min, to remove all the Si–H bonds<br />
and create an oxide layer on the pores’ surface to assure<br />
the covalent attachment with a proper chemical<br />
linker, the minopropyltriethoxysilane (APTES). To<br />
this aim, we have rinsed by immersion the DBR in a<br />
5% solution of APTES and a hydroalcoholic mixture<br />
of water and methanol (1:1), for 20 min at room temperature.<br />
After the reaction time, we have washed the<br />
chip with DI-water and methanol and dried in N 2<br />
stream. The silanized device was then baked at 100°C<br />
for 10 min. The next step consists in creating a surface<br />
able to link the carboxylic group of the proteins: we<br />
have thus immersed the DBR in a 2.5% glutaraldehyde<br />
solution in 20 mM HEPES buffer (pH 7.4) for 30 min,<br />
and then rinsed it in DI-water and finally dried in N 2<br />
stream. The glutaraldehyde reacts with the amino<br />
Fig. 2 Optical reflectivity<br />
spectrum of the Bragg<br />
reflector as-etched (continues<br />
curve), after oxidation<br />
(dashed curve), after APTES<br />
treatment (dot curve) and<br />
after glutaraldehyde<br />
treatment (dot-dashed curve)<br />
Reflectivity (u.a.)<br />
1.2<br />
1.1<br />
1.0<br />
0.9<br />
0.8<br />
0.7<br />
0.6<br />
0.5<br />
0.4<br />
0.3<br />
0.2<br />
0.1<br />
groups on the silanized surface and coats the internal<br />
surface of the pores with another thin layer of molecules.<br />
We have monitored all the reaction steps by Fourier<br />
Transformed Infrared (FT-IR) Spectroscopy and the<br />
consequent mo<strong>di</strong>fication of the optical response of the<br />
PSi device by reflectivity spectroscopy.<br />
The mo<strong>di</strong>fied surface obtained works as an active<br />
substrate for the chemistry of the following attachment<br />
of the protein: we have spotted on the PSi chip 20 ll of<br />
7.5 lM so<strong>di</strong>um bicarbonate buffer (pH 7.35) containing<br />
a rhodamine labeled TMBP and incubated the<br />
system at –4°C overnight. Even if our aim is the realization<br />
of a label-free optical biosensor based on the<br />
PSi nanotechnology, we have used a fluorescent protein<br />
to control the <strong>di</strong>stribution of the biological matter<br />
on the chip surface and to test the chemical stability<br />
of the covalent link between the TMBP and the PSi<br />
surface.<br />
After this assessment phase, we have also optically<br />
detected the ligand-bin<strong>di</strong>ng interaction by following<br />
the wavelength shift of the reflectivity spectrum. The<br />
experimental measurement of the TMBP–glucose<br />
bin<strong>di</strong>ng is a two step procedure: first, we have registered<br />
the optical spectrum of the porous silicon layer<br />
after the TMBP immobilization on the DBR surface<br />
and after the Glucose solution has been spotted on it.<br />
Results<br />
In Fig. 2, we have reported the optical spectra of our<br />
device as etched and after each step of the chemical<br />
0.0<br />
600 800<br />
Wavelength (nm)<br />
Bragg as etched<br />
after oxidation<br />
after APTES treatment<br />
after Glutaraldehyde treatment<br />
123
treatment in the range 600–800 nm, where the m=5<br />
Bragg resonance peak at about 720 nm is present. The<br />
oxidation process causes a blue shift of the reflectivity<br />
spectrum of 99.5 nm due to the lower value of the<br />
SiO2 refractive index (noxi @ 1.5)with respect to the<br />
Si refractive index (nSi @ 3.2). On the contrary, the<br />
silanisation steps by APTES and glutaraldehyde produce<br />
red shifts of the reflectivity spectrum of 28 and<br />
17 nm, respectively, correspon<strong>di</strong>ng to an increasing of<br />
the average refractive index of the layers due to a<br />
filling of the pores by the organic layers.<br />
Since the protein <strong>di</strong>stributes uniformly on the pores<br />
surface, the TMBP attachment causes a new detectable<br />
red shift of only 9 nm in the reflectivity spectrum.<br />
The FT-IR spectra of the oxide PSi sample and after<br />
the silanization process are reported in Fig. 3: the main<br />
characteristic peaks of silicon <strong>di</strong>oxide (at 1,124 cm –1 ),<br />
of the APTES amino groups (at 3,300 and 3,352 cm –1 )<br />
and of gluteraldehyde cyano group (at 1,404 cm –1 ) are<br />
easily recognized.<br />
In Fig. 4a, is shown the DBR observed by a Leica<br />
Z16 APO fluorescence macroscopy system after incu-<br />
Fig. 3 FT-IR spectra of the<br />
Bragg reflector after<br />
oxidation and APTES/<br />
glutaraldehyde treatment<br />
Fig. 4 a Porous silicon chip<br />
after incubation with the<br />
labeled-TMBP. b Porous<br />
silicon chip after washings in<br />
demi-water<br />
123<br />
%R<br />
320<br />
300<br />
280<br />
260<br />
240<br />
220<br />
200<br />
180<br />
160<br />
140<br />
120<br />
100<br />
80<br />
60<br />
40<br />
20<br />
4000<br />
3500<br />
bation. By illuminating the chip spotted with the labeled<br />
protein by a 100 W high-pressure mercury<br />
source, we found that the fluorescence is very high and<br />
homogeneous on the whole surface. We have also<br />
qualitatively tested the strength of covalent bond between<br />
the protein and the porous silicon surface by<br />
washing the device in a <strong>di</strong>alysis membrane overnight in<br />
DI-water. Since the fluorescent intensities <strong>di</strong>ffer of<br />
only few counts, we can conclude that the PSi-TMBP<br />
double layer is very stable.<br />
We have also measured the signal response to the<br />
glucose concentration after the interaction with the<br />
protein in a range between 10 and 150 lM. The<br />
maximum shift of the Bragg wavelength is 1.2 nm.<br />
Figure 5 shows the dose–response curve to glucose<br />
ad<strong>di</strong>tions. The estimated sensitivity of the TMBP-<br />
Chip is 0.03 nm/lM. Interestingly, the concentration<br />
of glucose that induces a optical response of the<br />
protein is very close to the amount of the sugar<br />
present in the human interstitial fluids. This result<br />
suggests the use of this protein in designing of a nonconsuming<br />
and minimally invasive biosensor for the<br />
3352.17 NH2<br />
3300.00 NH2<br />
DBR post oxidation process<br />
DBR post APTES functionalization<br />
DBR post Glutaraldehyde functionalization<br />
3000<br />
2930.43CH3<br />
2500 2000<br />
cm-1<br />
1124.94 SiO2<br />
1656.52 COH<br />
1500<br />
1404.35 CN<br />
Extremophiles<br />
1000<br />
500
Extremophiles<br />
∆λ (nm)<br />
1.4<br />
1.2<br />
1.0<br />
0.8<br />
0.6<br />
0.4<br />
0.2<br />
0.0<br />
70 80 90 100 110 120 130 140 150<br />
Glucose concentration (µM)<br />
Fig. 5 Dose–response curve for PSi DBR optical sensor exposed<br />
to several concentration of glucose<br />
continuous detection of the level of glucose in <strong>di</strong>abetic<br />
patients.<br />
In conclusion, we have reported an effective methodology<br />
for the covalent immobilization of proteins<br />
from extremophiles on a nanostructured material, such<br />
as porous silicon wafer, that can be utilized as general<br />
platform for the development of very stable proteinbased<br />
arrays for analyses of high social interest.<br />
Acknowledgments The authors thank Dr. Fabienne Chevance,<br />
University of Utah, Salt Lake City, USA, and Prof. Winfried<br />
Boos, University of Konstanz, Germany, for supplying a partially<br />
purified protein sample. This work is in the frame of the project<br />
‘‘Diagnostica Avanzata ed Alimentazione’’ CNR Commessa of<br />
the Agro-Food Department (SD). This work was also supported<br />
by the ASI project MoMa No. 1/014/06/0 (SD) and by a grant<br />
from the Ministero <strong>degli</strong> Affari Esteri, Direzione Generale per la<br />
Promozione e la Cooperazione Culturale (SD).<br />
References<br />
Bradford MM (1976) A rapid and sensitive method for the<br />
quantitation of microgram quantities of protein utilizing the<br />
principle of protein-dye bin<strong>di</strong>ng. Anal Biochem 72:248–254<br />
Brock TD (1985) Life at high temperatures. Science 239:132–138<br />
D’Auria S, Moracci M, Febbraio F, Tanfani F, Nucci R, Rossi M<br />
(1998) Structure-function stu<strong>di</strong>es on b-glycosidase from<br />
Sulfolobus solfataricus. Molecular bases of thermostability.<br />
Biochemie 80:949–957<br />
Diez J, Diederichs K, Greller G, Horlacher R, Boos W, Welte W<br />
(2001) The crystal structure of a liganded trehalose/maltosebin<strong>di</strong>ng<br />
protein from the hyperthermophilic archaeon<br />
Thermococcus litoralis at 1.85 A. J Mol Biol 305:905–915<br />
Herman P, Staiano M, Marabotti A, Varriale A, Scirè A, Tanfani<br />
F, Vecer J, Rossi M, D’Auria S (2006) D-Trehalose/<br />
D-maltose-bin<strong>di</strong>ng protein from the hyperthermophilic archaeon<br />
Thermococcus litoralis: The bin<strong>di</strong>ng of trehalose and<br />
maltose results in <strong>di</strong>fferent protein conformational states.<br />
Proteins Struct Func Bioinform 63:754–767<br />
Horlacher R, Xavier KB, Santos H, DiRuggiero J, Kossmann M,<br />
Boos W (1998) Archaeal bin<strong>di</strong>ng protein-dependent ABC<br />
transporter: molecular and biochemical analysis of the<br />
trehalose/maltose transport system of the hyperthermophilic<br />
archaeon Thermococcus litoralis. J Bacteriol 180:680–689<br />
Jaenicke R (1994) <strong>Stu<strong>di</strong></strong>es on hyperstable proteins: crystallins<br />
from the eye-lens and enzymes from thermophilic bacteria.<br />
In: Doniach S (ed) Statistical mechanics, protein structure,<br />
and protein substrate interactions. Plenum, New York, pp<br />
49–62<br />
Sun YJ, Rose J, Wang BC, Hsiao CD (1998) The structure of<br />
glutamine-bin<strong>di</strong>ng protein complexed with glutamine at 1.94<br />
angstrom resolution: comparisons with other amino acid<br />
bin<strong>di</strong>ng proteins. J Mol Biol 278:219–222<br />
Xavier KB, Martins LO, Peist R, Kossmann M, Boos W, Santos<br />
H (1996) High-affinity maltose/trehalose transport system in<br />
the hyperthermophilic archaeon Thermococcus litoralis.<br />
J Bacteriol 178:4773–4777<br />
123
Sensors 2008, 8, 1-x manuscripts; DOI: 10.3390/sensors<br />
Article<br />
sensors<br />
ISSN 1424-8220<br />
www.mdpi.com/journal/sensors<br />
Porous Silicon Based Resonant Mirrors for Biochemical Sensing<br />
Edoardo De Tommasi 1,* , Luca De Stefano 1 , Ilaria Rea 1, 2 , Valentina Di Sarno 1, 2 , Lucia<br />
Rotiroti 1, 3 , Paolo Arcari 4 , Annalisa Lamberti 4 , Carmen Sanges 4 and Ivo Ren<strong>di</strong>na 1<br />
1<br />
Institute for Microelectronics and Microsystems – Unit of Naples – National Council of Research,<br />
Via P. Castellino 111, 80131 Napoli, Italy.<br />
2<br />
Department of Physical Sciences, University of Naples “Federico II”, Via Cinthia, 80126 Naples,<br />
Italy<br />
3<br />
Department of Organic Chemistry and Biochemistry, University of Naples “Federico II”, Via<br />
Cinthia, 80128 Napoli, Italy<br />
4<br />
Department of Biochemistry and Me<strong>di</strong>cal Biotechnologies, University of Naples “Federico II”, Via<br />
S. Pansini 5, 80131 Napoli, Italy.<br />
* Author to whom correspondence should be addressed: edetommasi@na.imm.cnr.it<br />
Received: 14 Oct 2008; in revised form: 6 Oct 2008 / Accepted: 21 Oct 2008 / Published:<br />
Abstract: We report on our preliminary results in the realization and characterization of a<br />
porous silicon (PSi) resonant mirror (RM) for optical biosensing. We have numerically and<br />
experimentally stu<strong>di</strong>ed the coupling between the electromagnetic field, totally reflected at<br />
the base of a high refractive index prism, and the optical modes of a PSi waveguide. This<br />
configuration is very sensitive to changes in the refractive index and/or in thickness of the<br />
sensor surface. Due to the high specific area of the PSi waveguide, very low DNA<br />
concentrations can be detected confirming that the RM could be a very sensitive and labelfree<br />
optical biosensor.<br />
Keywords: DNA Optical Biosensors, Porous Silicon, Resonant Mirrors.<br />
1. Introduction<br />
Optical label-free biosensors are ideal can<strong>di</strong>dates for high throughput screening in the prevention of<br />
bio-threat agents or social illnesses: all the <strong>di</strong>fferent configurations proposed in the literature, such as<br />
optical microcantilever [1], interferometric devices [2, 3], and also the recent photonic crystals<br />
geometries [4], show very high sensitivities in the recognition of specific molecules. The detection of
Sensors 2008 2<br />
analytes present in complex mixtures in ultra-low concentration, such as viruses in human blood, is<br />
possible due to two key factors: the utilization of bioprobres, properly linked on the sensor surface,<br />
and an optimized design to maximize the interaction between the biological matter and the<br />
electromagnetic field.<br />
Resonant Mirrors (RMs) are optical sensors which exploit the evanescent fields features to probe<br />
changes in refractive index and thickness taking place on their surface after exposure to gaseous or<br />
liquid substances. From the optical point of view, RMs are refractometric sensing devices similar to<br />
grating coupler sensors: they are characterized by the high sensitivity typical of wavegui<strong>di</strong>ng devices,<br />
and, on the other hand, by the simple scheme of Surface Plasmon Resonance (SPR) sensors [5]. In<br />
RMs, in fact, incident light gives rise to an evanescent wave at the interface between a prism and an<br />
optical waveguide; the coupling of light into the waveguide occurs only at those specific incident<br />
angles for which the propagation constant of the evanescent wave matches that of a waveguide mode.<br />
This matching produces a <strong>di</strong>p in the angular spectrum of the reflected light, so that each change in the<br />
refractive index and-or in thickness at the sensor surface produces a correspon<strong>di</strong>ng shift in the position<br />
of the <strong>di</strong>p [5, 6].<br />
In this work, we report on the realization of a porous silicon (PSi) based RM prototype on chip,<br />
whose modal properties and characteristic angular resonances have been both numerically computed<br />
and experimentally measured.<br />
PSi is by far one of the most intriguing material in optical sensing: the refractive index is widely<br />
tuneable, namely between the silicon refractive index and that of the air, and its specific surface is very<br />
large, up to 500 m 2 cm -3 . Due to these characteristics, lot of optical structures, such as rugate filters [7]<br />
and microcavities [8] have been proposed in literature for biosensing. Recently, a PSi waveguide<br />
biosensor have been theoretically and experimentally demonstrated [9]. We have stu<strong>di</strong>ed and reported<br />
in this paper the feasibility of such biosensor for DNA-DNA hybri<strong>di</strong>zation experiment.<br />
Figure 1. Experimental setup. DL: <strong>di</strong>ode laser emitting s-polarized ra<strong>di</strong>ation at 785 nm; L:<br />
lens with focal length f=7.5 cm; RM: resonant mirror; RS: rotation stage; PM: power meter<br />
on an independent rotation stage.
Sensors 2008 3<br />
2. Experimental<br />
A highly doped p+-silicon, oriented, 0.01 Ω·cm resistivity, 400 μm thick was used as<br />
substrate in the waveguide fabrication. The structure was obtained by electrochemical etching of<br />
crystalline silicon in a HF-based solution (50 wt. % HF:ethanol = 3:7) at room temperature. A current<br />
density of 10 mA/cm 2 was applied for 18.9 s to produce the core layer of thickness 2.5 μm and porosity<br />
65 %, while a current density of 109 mA/cm 2 was applied for 13.6 s in the case of the clad<strong>di</strong>ng layer<br />
with thickness 2.5 μm and porosity 78 %. These porosities correspond to a core and clad<strong>di</strong>ng refractive<br />
indexes of 1.749 and 1.384, respectively, calculated by the Bruggeman model [10] at a wavelength of<br />
785 nm. The device was then fully oxi<strong>di</strong>zed in pure O2 by a two step thermal treatment (400 °C for 30<br />
min and 900 °C for 15 min).<br />
In Figure 1 a wide view of our experimental set-up is shown. RM is mounted on a rotation stage<br />
(Melles Griot, model 07 TRS 501); the light source is a <strong>di</strong>ode laser emitting s-polarized ra<strong>di</strong>ation at<br />
λ=785 nm (CrystaLaser, model RCL-080-785-5); the laser beam exiting the prism is collected by a<br />
photometer (Newport, model 1830-C). The waveguide has been brought in close contact with one side<br />
of an SF6 prism (n=1.784), still allowing the presence of a sub-micrometric air layer acting as a<br />
coupling zone between the prism and the waveguide.<br />
The covalent bond of the DNA single strand (5’-GGACTTGCCCGAATCTACGTGTCC-3’) on<br />
the PSi surface is based on a three steps functionalization process constituted by a chemical<br />
passivation of the surfaces, as it is shown in Figure 2 (all the chemicals used in the present study were<br />
from Sigma). We have treated the surfaces with a proper chemical linker, the<br />
aminopropyltriethoxysilane (APTES). Samples have been rinsed by <strong>di</strong>pping in a 5 % solution of<br />
APTES and a hydroalcoholic mixture of water and methanol (1:1) for 20 min at room temperature.<br />
After the reaction time, we have washed the chips by deionized water and methanol, and dried in N2<br />
stream. The silanized devices were then baked at 100 °C for 10 min. To create a surface able to link<br />
the amino group of the biological probes, we have thus immersed the chips in a 2.5% glutaraldehyde<br />
(GA) solution in 20 mM HEPES buffer (pH 7.4) for 30 min, and then rinsed it in deionized water and<br />
finally dried in N2 stream. The GA reacts with the amino groups on the silanized surface and coats the<br />
internal surface of the pores with another thin layer of molecules. All these reaction steps have been<br />
monitored by means of FT-IR spectroscopy to verify the presence of the characteristic peaks of the<br />
organic linkers. The main characteristic peaks of Si-O-Si (1040 cm -1 ) and Si-OH bonds (3740 and<br />
1646 cm -1 ) are present, after oxidation, on the PSi. After the silanization process, the APTES<br />
characteristic peaks of the ethylic (at 1626, 1529 and 1379 cm -1 ), amino (at 1060 cm -1 ) and also the –<br />
CH (at 2908 and 2851 cm -1 ) groups are well evident. Finally, after the GA treatment, the characteristic<br />
imine group (at 1627 cm-1) due to the reaction with APTES is easily recognized.<br />
The PSi surface was covered overnight at room temperature with a 50 μM DNA probe solution (30<br />
μL). For each measurement, complementary (5’-GGACACGTAGATTCGGGCAAGTCC-3’) and noncomplementary<br />
(5’-CACTGTACGTGCGAATTAGGTGAA-3’) DNA (10 μL) were spotted on the<br />
chip and incubated for two hours. Before optical measurements, all samples have been extensively<br />
rinsed in deionized water to remove the excess of biological matter.
Sensors 2008 4<br />
Figure 2. Schematic of the functionalization process, from the oxi<strong>di</strong>zed PSi chip to the<br />
covalent attachment of the DNA single strands.<br />
SiO 2<br />
OH<br />
OH<br />
OH<br />
O 2Si<br />
O 2Si<br />
O<br />
5 % APT ES<br />
20' R.T<br />
O 2Si<br />
O<br />
O<br />
O<br />
Si<br />
O<br />
NH 2<br />
2.5% GA<br />
30' R.T<br />
O<br />
3. Results and Discussion<br />
O<br />
O<br />
Si<br />
O<br />
O<br />
Si<br />
O<br />
O<br />
N<br />
N<br />
C<br />
C<br />
C O<br />
C N<br />
ssDNA<br />
ss DN A-NH 2<br />
over night<br />
R.T<br />
The PSi waveguide, once oxi<strong>di</strong>zed in order to reduce the scattering losses [11] and to allow the<br />
bond with biological linkers, has been characterized by means of “m lines” technique, which<br />
determines the modal behavior of the waveguide at <strong>di</strong>fferent wavelengths and estimates the refractive<br />
in<strong>di</strong>ces nc and nb for the core an the buffer layers [12]. In Figure 3, is reported a standard m-lines<br />
spectrum due to the excitation of four adjacent modes in transverse electric (TE) polarization of the<br />
electromagnetic field at 785 nm: by measuring the coupling angles position is possible to quantify the<br />
core and clad<strong>di</strong>ng thicknesses and refractive indexes. We have obtained the refractive indexes of core<br />
and buffer and the thickness of the core equal to nc = 1.370 ± 0.001, nb = 1.18 ± 0.01, and dc = 2.71 ±<br />
0.01μm respectively. It is important to notice the strong reduction of the refractive indexes after the<br />
thermal oxidation process, mainly due to the substitution of silicon with silicon <strong>di</strong>oxide.<br />
The close contact between the prism and the waveguide still allows the presence of a sub-micrometric<br />
air layer where the evanescent fields coming from the prism and the waveguide can overlap, thus<br />
transferring the light energy from the prism to the waveguide. The thickness of this air layer strongly<br />
affects the efficiency of the coupling of the incoming light into the waveguide. In Figure 4, some<br />
simulations of RM spectra correspon<strong>di</strong>ng to <strong>di</strong>fferent air thicknesses are shown: the lower the<br />
thickness, the stronger the coupling of the evanescent wave with the waveguide. The <strong>di</strong>ps present in<br />
the total reflection zone (i.e. for incident angle θ > 34°<br />
), correspond to the characteristic modes<br />
propagating in the waveguide. The number of these modes can be obtained from the well known<br />
2d 2 2<br />
relation M = nc<br />
− nb<br />
≈ 5,<br />
where d stands for the thickness of the wavegui<strong>di</strong>ng layer.<br />
λ<br />
i
Sensors 2008 5<br />
Figure 3. Resonant peaks of the waveguide, correspon<strong>di</strong>ng to four guided modes in TE<br />
polarization at 785 nm, after functionalization with APTES and GA.<br />
Output power (μW)<br />
65<br />
60<br />
55<br />
50<br />
45<br />
40<br />
46 48 50 52<br />
Resonance Angle (deg)<br />
Figure 4. Numerical simulations of RM spectra as a function of the air gap thickness<br />
Normalized Output Intensity<br />
1.0<br />
0.5<br />
0.0<br />
30 35 40<br />
Resonance Angle (deg)<br />
Air thickness: 1 μm<br />
Air thickness: 500 nm<br />
Air thickness: 200 nm<br />
.<br />
In Figure 5, the results of angular measurements after incubation with probe DNA and after two<br />
steps of the DNA hybri<strong>di</strong>zation process at two <strong>di</strong>fferent concentrations are reported: the resonance<br />
shifts confirm the effectiveness of the DNA bin<strong>di</strong>ng.
Sensors 2008 6<br />
Figure 5. PSi RM resonances after functionalization with probe DNA and after two DNA<br />
hybri<strong>di</strong>zation steps: the coupling angle shifts demonstrate the molecular interaction<br />
between the probe DNA and its complementary target at <strong>di</strong>fferent concentration.<br />
Normalized Output Power<br />
1.0<br />
0.9<br />
0.8<br />
0.7<br />
0.6<br />
After incubation with 50 μM of DNA<br />
After incubation with 25 μM of target cDNA<br />
After incubation with 50 μM of target cDNA<br />
Δθ<br />
45.5 46.0 46.5 47.0 47.5<br />
Resonance angle (deg)<br />
The coupling angle shifts as a function of the concentration of the complementary DNA, as it can be<br />
seen in Figure 6. The high efficiency of hybri<strong>di</strong>zation is confirmed by the fact that, after exposure to<br />
50 μM of complementary DNA, the sensor response curve reaches a plateau, correspon<strong>di</strong>ng to a<br />
saturation of almost all of the active sites. The experimental points can be well fitted (R 2 =0.99, χ 2
Sensors 2008 7<br />
Figure 6. Resonance shifts of the waveguide mode coupling angle as a function of the<br />
DNA concentration.<br />
4. Conclusions<br />
Δθ (deg)<br />
1,8<br />
1,6<br />
1,4<br />
1,2<br />
1,0<br />
0,8<br />
0,6<br />
0,4<br />
0,2<br />
0,0<br />
0 10 20 30 40 50<br />
Target Concentration (μM)<br />
We have experimentally confirmed the ability of a PSi RM to detect a DNA-DNA hybri<strong>di</strong>zation<br />
process with very high sensitivity. The use of PSi technology in the fabrication of the waveguide,<br />
which is the heart of a RM sensor, allows the presence of a high specific-surface available for<br />
interaction with great amounts of analytes. We have demonstrated that our measuring system is<br />
characterized by a technical limit of about 80 nM of DNA, in good agreement with previously<br />
published theoretical calculations [13]. Next efforts will be aimed at the fabrication of a free-stan<strong>di</strong>ng<br />
waveguide (no crystalline silicon substrate), which, in conjunction with a flow delivery system, could<br />
allow to reduce considerably the time of incubation of the DNA probe and targets over the<br />
functionalized surface.<br />
Acknowledgements<br />
This work is partially supported by the FIRB Italian project RBLA033WJX_005.<br />
References and Notes<br />
1. Lechuga, L.M.; Tamayo, J.; Álvarez, M.; Carrascosa, L.G.; Yufera, A.; Doldán, R.; Peralías, E.;<br />
Rueda, A.; Plaza, A.J.; Zinoviev, K.; Domínguez, C.; Zaballos, A.; Moreno, M.; Martínez-A, C.;<br />
Wenn, D.; Harris, N.; Bringer, C.; Bar<strong>di</strong>nal, V.; Camps, T.; Vergenègre, C.; Fontaine, C.; Díaz, V.;<br />
Bernad, A. A highly sensitive microsystem based on nanomechanical biosensors for genomics<br />
applications. Sensor. Actuat. B-Chem. 2006, 118, 2-10.
Sensors 2008 8<br />
2. Pacholski C.; Sartor M.; Sailor M. J.; Cunin F.; Miskelly G. M. Biosensing using porous silicon<br />
double-layer interferometers: Reflective interferometric fourier transform spectroscopy. J. Am.<br />
Chem. Soc. 2005, 127, 11636-11645.<br />
3. De Stefano, L.; Rotiroti, L.; Rea, I.; Ren<strong>di</strong>na, I.; Moretti, L.; Di Francia, G.; Massera, E.; Lamberti,<br />
A.; Arcari, P.; Sangez, C. Porous silicon-based optical biochips. J. Opt. A-Pure Appl. Op. 2006, 8,<br />
S540-S544.<br />
4. Lee, M.R.; Fauchet, P.M. Two-<strong>di</strong>mensional silicon photonic crystal based biosensing platform for<br />
protein detection. Opt. Express 2007, 15, 4530-4535.<br />
5. Cush, R.; Cronin, J.M.; Stewart, W.J.; Maule, C.H.; Molloy, J.; Goddard, N.J. The resonant mirror:<br />
a novel optical biosensor for <strong>di</strong>rect sensing of biomolecular interactions. I: Principle of operation<br />
and associated instrumentation. Biosens Bioelectron. 1993, 8, 347-353.<br />
6. Bal<strong>di</strong>ni, F.; Chester, A.N.; Homola, J.; Martellucci, S. Optical Chemical Sensors. Springer:<br />
Dordrecht, The Netherlands, 2006.<br />
7. Chapron, J.; Alekseev, S.A.; Lysenko, V.; Zaitsev, V.N.; Barbier, D. Analysis of interaction<br />
between chemical agents and porous Si nanostructures using optical sensing properties of infra-red<br />
Rugate filters. Sensor. Actuat. B-Chem. 2007, 120, 706-711.<br />
8. Ouyang, H.; Chrtistophersen, M.; Viard, R.; Miller, B.L.; Fauchet, P.M. Macroporous silicon<br />
microcavities for macromolecule detection. Adv. Funct. Mater. 2005, 15, 1851-1859.<br />
9. Saarinen, J.J.; Weiss, S.M.; Fauchet, P.M.; Sipe, J.E. Optical sensor based on resonant porous<br />
silicon structures. Opt. Express 2005, 13, 3754-3764.<br />
10. Khardani, M.; Bouaïcha, M.; Bessaïs, B. Bruggeman effective me<strong>di</strong>um approach for modelling<br />
optical properties of porous silicon: comparison with experiment. Phys. Stat. Sol. 2007, 4, 1986-<br />
1990.<br />
11. Pirasteh, P.; Charrier, J.; Dumeige, Y.; Haesaert, S.; Joubert P. Optical loss of porous silicon and<br />
oxi<strong>di</strong>zed porous silicon planar waveguides. J. Appl. Phys. 2007, 101, 083110.<br />
12. Ulrich, R.; Torge, R. Measurement of thin film parameters with a prism coupler. Appl. Optics<br />
1973, 12, 2901-2908.<br />
13. Rong, G.; Najmaie, A.; Sipe, J. E.; Weiss, S. M. Nanoscale porous silicon waveguide for label-free<br />
DNA sensing. Biosens. Bioelectron. 2008, 23, 1572-1576.<br />
14. Descrovi, E.; Frascella, F.; Sciacca, B.; Geobaldo, F.; Dominici, L.; Michelotti, F. Coupling of<br />
surface waves in highly defined one-<strong>di</strong>mensional porous silicon photonic crystals for gas sensing<br />
applications. Appl. Phys. Lett. 2007, 91, 241109.<br />
© 2008 by the authors; licensee Molecular Diversity Preservation International, Basel, Switzerland.<br />
This article is an open-access article <strong>di</strong>stributed under the terms and con<strong>di</strong>tions of the Creative<br />
Commons Attribution license (http://creativecommons.org/licenses/by/3.0/).