Optical Sensors for Vapors, Liquids, and Biological ... - IEEE Xplore

IEEE TRANSACTIONS ON NANOTECHNOLOGY, VOL. 3, NO. 1, MARCH 2004

49

Optical Sensors for Vapors, Liquids, and Biological Molecules Based on Porous Silicon Technology Luca De Stefano, Luigi Moretti, Annalisa Lamberti, Olimpia Longo, Massimiliano Rocchia, Andrea M. Rossi, Paolo Arcari, and Ivo Rendina, Member, IEEE

Abstract—The sensing of chemicals and biochemical molecules using several porous silicon optical microsensors, based both on single-layer interferometers and resonant-cavity-enhanced microstructures, is reported. The operation of both families of sensors is based on the variation of the average refractive index of the porous silicon region, due to the interaction with chemical substances either in vapor or liquid state, which results in marked shifts of the device reflectivity spectra. The well established single-layer configuration has been used to test a new chemical approach based on Si-C bonds for covalent immobilization of biological molecules, as probe, in a stable way on the porous silicon surface. Preliminary results on complementary oligonucleotide recognition, based on this technique, are also presented and discussed. Porous silicon optical microcavities, based on multilayered resonating structures, have been used to detect chemical substances and, in particular, flammable and toxic organic solvents, and some hydrocarbons. The results put in evidence the high sensitivity, the reusability, and the low response time of the resonant-cavity-enhanced sensing technique. The possibility of operating at room temperature, of remote interrogation, and the absence of electrical contacts are further advantages characterizing the sensing technique. Index Terms—DNA biosensors, microcavities, nanostructures, optical sensor, porous silicon (PSi).

I. INTRODUCTION

P

OROUS SILICON (PSi) can be used as smart transducer material in sensing application and, in particular, in the detection of vapors, liquids, and biochemical molecules. In fact, on exposure at chemical substances, several physical quantities, such as refractive index, photoluminescence, and electrical conductivity, change drastically. A key feature of a physical transducer, being sensitive to organic and biological molecules, either in vapor or liquid state, is a large surface area: PSi has a porous sponge-like structure with a specific m cm , so that it can assure area of the order of a very effective interaction with several adsorbates. Moreover, Manuscript received June 16, 2003; revised October 17, 2003. This work was supported in part by Regione Campania “Centro di Competenza Nuove Tecnologie per le Attività Produttive,” Napoli, Italy. This paper was presented in part at the Symposium of Microtechnologies for the New Millenium, Nanotechnology Conference, Gran-Canaria, Spain, May 2003. L. De Stefano and I. Rendina are with the Istituto per la Microelettronica e Microsistemi, Consiglio Nazionale delle Ricerche, 80131 Napoli, Italy (e-mail: [email protected]). L. Moretti is with the Dipartimenti Elettronica Informatica Sistemistica, Università della Calabria, 87036 Rende, Italy. A. Lamberti, O. Longo, and P. Arcari are with the Dipartimenti Biochimica e Biotecnologie Mediche Università di Napoli “Federico II” and CEINGE-Biotecnologie Avanzate Scarl, 80131 Napoli, Italy. M. Rocchia and A. M. Rossi are with the Istituto Elettrotecnico Nazionale “G. Ferraris,” Torino 10135, Italy. Digital Object Identifier 10.1109/TNANO.2004.824019

PSi is an available, low-cost material, compatible with standard IC processes, so that it could usefully be employed in the realization of smart sensors and microsystems. In particular, in the last years, a lot of experimental work has been reported concerning its use as optical sensors in chemical and biological sensing [1]–[5]. In environmental monitoring, optical readout techniques are of particular interest mainly because they do not require electric contacts that may cause explosions or fire in dangerous environments, they are not affected by electromagnetic interference, and, moreover, allow wireless remote interrogation. Also in these kinds of applications, the exploitation of PSi seems to be very promising. As a matter of fact, the possibilities of modulating the PSi porosity, as well as good control of the interface between layers of different porosity, permit to fabricate not only singlelayer optical interferometers but also multilayer structures with high-optical contrast such as high-reflectivity Bragg reflectors, optical waveguides, and high-quality Fabry–Perot filters. All these structures are showing noticeable capability in optical sensing applications. In this paper, we present new results about sensing of chemical substances, flammable, and toxic organic solvents, and also some hydrocarbons, based on the exploitation of PSi optical microsensors. The sensors are based on multilayer resonant-cavity-enhanced microstructures. The technique does not need any sample preparation, the microsensors are remoteoptically-interrogated, reusable, operate at room temperature, and show very fast response times. Furthermore, PSi technology has shown great capability in detecting biological molecules with high selectivity, using specific linker agent and probe molecules. Although multilayer microcavities, due to their higher quality factor, are more sensible than single layer interferometers, these last ones make it easier to study the immobilization chemical process of probe molecules. For this reason, we have extensively exploited single-layer interferometric PSi structures to setup a new DNA immobilization and recognition technique. This is based on the functionalization of the hydrogen-terminated porous silicon surface by means of Si-C bonds, allowing the creation of a stable organic layer covalently attached to the PSi surface. Experimental results on this original technique are presented and discussed. Its capability in detecting DNA strands has also been experimentally demonstrated. II. DESIGN AND FABRICATION OF PSI OPTICAL DEVICES Both single-layer interferometers and multilayer microcavities were produced by electrochemical etching on p-type ( m cm) standard silicon wafers. PSi single-layer devices were prepared by electrochemical etching in an

1536-125X/04$20.00 © 2004 IEEE

50


Fig. 1. Plan view and cross section of a real device. In the inset is shown a schematic of PSi microcavity.

HF-Ethanol solution, 50% by volume; an etching current density of 550 mA/cm for 7.6 s results in a porosity of 93%. nm. Average pores diameters vary in the range The optical microcavities are made of two distributed Bragg -thick cavity in the reflectors (DBRs) with a Fabry–Pèrot middle. The scheme of the porous silicon microcavity (PSM) is reported in Fig. 1, here are also shown a surface plan view and a cross section of the real device, captured by SEM-FEG. Several alternating pairs of PSi layers, having different refractive indexes, obtained modulating the porosity, constitute the DBRs. The optical thickness of each single-layer is , where is the physical thickness of layer, its refractive index and is the Bragg wavelength, thus, the whole stack is a resonating structure at the Bragg wavelength . The optical characteristic is a high-reflectivity passband, having a narrow peak of transmittance approximately in the center. Its width is controlled by a proper design of the layers stack. Both the distributed Bragg reflectors, up and under the optical cavity, are constituted by five periods of alternated high- and low-refractive index -thick layers. An etching current density of 80 mA/cm for 1.989 s, resulting in a porosity of 55%, was used for high-refractive index layers, while an etching current density of 250 mA/cm for 1.1 s, resulting in a porosity of 68%, for low ones. Average pores diameters vary in the range nm. The -thick cavity is a low–index PSi layer. This structure is characterized by a single resonance peak at 1313 nm in a transmittivity stopband between 1200 and 1500 nm. In Fig. 2, the experimental measured reflectivity spectrum of the multilayered device is reported together with a numerical simulation. The reflectivity spectrum is reproduced by a transfer matrix method [6] including the wavelength dispersion of silicon. III. VAPORS AND LIQUID SENSING When the PSM is exposed to vapors of organic solvent, a repeatable and completely reversible change in the reflectivity spectrum is observed; the substitution of air with organic species in the pores determines an increasing of the average refractive index of the microcavity, resulting in a marked red shift of its characteristic peak. The same results, within the experimental error, are carried out by dipping the sensor in the liquid phase of the same substances. A quantitative analysis of this net red shift of the peak cavity is realized by applying the Bruggeman effective medium approximation theory [7]. In this way, we can describe the change of the average refractive index and calculate the expected peak shift as a function of the pores filling. In this approach, it has been assumed that at equilibrium, the filled

Fig. 2. Experimental and simulated spectrum of the PSMC. The solid line is the experimental registered spectrum; the dot line is the numerical simulated one.

layer liquid volume fraction is constant across the whole stack unless limited in some layers where all air has been displaced. This point of view is justified by the capillarity condensation phenomena. In Fig. 3(a), is reported the calculated peak shift as a function of the layer liquid fraction for exposure at different solvents, while in Fig. 3(b) the reflectivity spectra, before and after exposure, are shown. In Table I, some properties of the used chemicals are reported: as it can be noted even for substances with very close values of refractive index (acetone, 1.359, and ethanol, 1.360), the peak is well resolved. shift Time-resolved reflectivity measurements (see Fig. 4), carried out at the cavity resonance wavelength, show, in the case of iso-propanol, that identification of the solvent occurs in less then s), while the signal returns to its original value 10 s ( in still shorter time ( s). We expect that time constants of the same order would be found for all the other analyses, due to their similar chemical properties (see in Table I, STC and VP columns). The values of response and recover times depend not only on the physical phenomena involved (i.e., adsorption and desorption in the PSi layers, respectively) but also on the geometry of the test chamber and on the measurement procedure, i.e., saturated atmosphere or continuous flow mode. In saturation condition, the identification time is mainly determined by the diffusion of the gas into the chamber volume: in fact, when vapor is in contact with the porous silicon surface, the capillary condensation takes place instantaneously [8].

DE STEFANO et al.: OPTICAL SENSORS FOR VAPORS, LIQUIDS, AND BIOLOGICAL MOLECULES BASED ON POROUS SILICON TECHNOLOGY

51

Fig. 3. (a) Calculated wavelength shift of the PSM peak as a function of the organic solvent liquid fraction; (b) reflectivity spectra of the PSM after the exposure to different flammable organic substances. TABLE I CHEMICAL ORGANICS USED IN SENSING EXPERIMENTS AND SOME RELEVANT PHYSICAL-CHEMICAL PROPERTIES: n IS THE LIQUID REFRACTIVE INDEX; IS THE DENSITY (at 25 C); STC IS THE SURFACE TENSION COEFFICIENT (at 25 C); VP IS THE VAPOUR PRESSURE; BP IS THE BOILING POINT; THE PEAK SHIFT AND LLF IS THE LIQUID LAYER FRACTION

1

Therefore, due to high resolution (less than 1 nm) of our experimental setup, we expect to obtain detection limits of the order of few ppm. IV. DNA SENSING

Fig. 4.

Time-resolved measurement in the case of iso-propanol.

Preliminary experimental results on device performances indicate that iso-propanol vapor concentrations of 50 ppm induce nm). a complete shift of the resonant cavity peak (

PSi monolayers were used as trial substrates in order to attach synthetic oligonucleotides for biosensing purposes. The wide literature available on this subject shows sometimes discordant and nonreproducible results. The critical issue in DNA recognition is the chemistry of probe-oligonucleotides immobilization on the PSi surface, which seems to depend on the stable and affordable sensor operation. Many of the results carried out to date, derived from the standard DNA immobilization chemistry exploited on silicon-oxide rich substrates (such as glass), involve the silanization of the oxidized PSi surface [9]–[12]. A promising recently proposed alternative is that exploiting the reaction of acids molecules with the hydrogen-terminated porous silicon surface in order to obtain a more stable organic layer covalently attached to the PSi surface through Si-C bonds [13]. In particular, we functionalized the porous silicon surface

52


Fig. 6. Fig. 5. Infrared spectra of a PS sample: Fresh etched (dotted line) and modified with undecylenic acid (full line).

by treatment with undecylenic acid (Sigma Aldrich) [14]. PSi wafers, freshly prepared as described in Section II, were placed under a stream of nitrogen in a small glass container in the presence of 5 ml of melted undecylenic acid (m.p. 25 C–27 C) and incubated at 60 C, in a water bath for 16 h. After the reaction, the excess of reagent was removed and the wafers were extensively washed at room temperature with chloroform, ethanol, and then dried under nitrogen stream. Fig. 5 reports the FT-IR spectra of a PSi sample before (fresh etched) and after chemical modification with undecylenic acid. In the spectrum obtained from the modified PSi surface, new absorption bands, related both to the alkyl chain and to the acid function, apand 2865 cm can be assigned pear. The bands at 2922 cm to the asymmetric and symmetric stretching vibration of -CH -species, respectively, and the strong absorption at 1716 cm is due to the stretching vibration of the double bond. The absence of alkenes CH stretching above 3000 cm and double bond around 1650 cm support an anchoring mechadouble bond. Morenism which involves the breaking of over, the total consumption of the SiH species (centered at 2100 cm ) is consistent with a hydrosilylation reaction which passivates the surface [14]. The attachment of the DNA probe to the functionalized PSi was performed through the conversion of the acidic function to amide, according to the following protocol: 100 l of aqueous solution containing 50 M of modified oligonucleotide 3’-aminolink (5’-GGACTTGCCCGAATCTACGTGTCC-3’-O-PO H-O-CH CHOHCH NH ), in the following referred as the probe, and 200 M of 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) [15] were spotted on the wafer and incubated overnight at room temperature in a small plastic container filled at the bottom with water in order to reduce the evaporation of the reagent solution. The concentration of the probe used was identical to that used by Chan et al. [10] in the fabrication of a photoluminescent DNA biosensor. Wafers were then washed with deionized water and dried under a nitrogen flow before optical measurements. Hybridization was performed with a complementary single strand DNA (DNA C, 5’-GGACACGTAGATTCGGGCAAGTCC-3’), forming a

Reflectometric interference spectra of PSi monolayer for DNA sensing.

C, as highly stable duplex with melting point [16]. evaluated by the formula Hybridization with non complementary single stand DNA (DNA NC) was instead performed with the oligonucleotide 5’-CACTGTACGTGCGAATTAGGTGAA-3’. The hybridization protocol was carried out as follows: 50 M of a solution containing the DNA C or the DNA NC was spotted on the wafer surface and incubated at 37 C overnight in a plastic container as described above. The samples were then washed with deionized water and dried under nitrogen. Reflectivity spectra of the same sample, registered at any step of the sensing process, are shown in Fig. 6(a). While the shift due to chip interaction with noncomplementary DNA is less than 1 nm, a well defined blue-shift of 8 nm is detected after exposure of the PSi to complementary DNA. The signal difference is shown in Fig. 6(b). The 8-nm blue shift observed corresponds to a change nm, where is the in effective optical thickness change in the average refractive index of the porous silicon layer due to interaction with the biological molecules, and is its thickness. Even if is very difficult to hypothesize a quantitative model accounting for the size of the blue shift recorded, due to the complexity of the multistep sensing process, this result suggests that the molecular complexation generates in our system a new stack made of organic and inorganic matter with an average refractive index lower than the original porous layer one. At the moment, different mechanisms are under study to explain the nature and the size of the blue shift: free charge carrier transfer between organic and inorganic layers or mechanical stress induced by intermolecular interaction could cause variation in the refractive index value or in the thickness of the layer, thus explaining the change in the effective optical thickness. V. CONCLUSION The realization and characterization of porous silicon-based optical microsensors for chemicals and biochemical molecules is reported. The devices have been realized by standard microelectronic techniques in two main configurations: single-layer interferometers and high-quality multilayer microcavities. Both base their operation on the variation of the average refractive

DE STEFANO et al.: OPTICAL SENSORS FOR VAPORS, LIQUIDS, AND BIOLOGICAL MOLECULES BASED ON POROUS SILICON TECHNOLOGY

index of the active porous silicon region due to the interaction with chemical substances either in vapor in or liquid state. Depending on the possible interaction mechanism of the porous silicon surface with the analyzed chemicals, i.e., surface chemi-sorption or capillary condensation in the pores, marked red or blue shifts of the device reflectivity spectrum are obtained. The well-established single-layer configuration has been used to test a new chemical approach for covalently immobilizing probe biological molecules in a stable way on the porous silicon surface through Si-C bonds. Preliminary successful results on complementary oligonucleotides recognition, based on this technique, have been presented and discussed. Multilayered resonating structures have been instead used to detect chemical substances, such as flammable and toxic organic solvents, and hydrocarbons. A quantitative analysis of the peak cavity red shift observed has been carried out by applying the Bruggeman effective medium approximation theory. In this way, the change of the average refractive index and a calculation of the expected peak shift as a function of the pores filling by capillarity condensation phenomena have been obtained. The results have pointed out the high sensitivity, reusability, and low response time (never exceeding 10 s) of the resonant-cavity-enhanced sensing technique. ACKNOWLEDGMENT The authors gratefully thank Prof. F. Salvatore for the encouragement received. REFERENCES [1] H. F. Arrand, A. Loni, R. Arens-Fischer, M. G. Kruger, M. Thoenissen, H. Luth, S. Kershaw, and T. M. Benson, “Novel liquid sensor based on porous silicon optical waveguides,” IEEE Photon. Technol. Lett., vol. 10, pp. 1467–1469, Oct. 1998. [2] S. Zangooie, R. Bjorklund, and H. Arwin, “Ellipsometric characterization of anisotropic porous silicon Fabry–Pérot filters and investigation of temperature effects on capillary condensation efficiency,” J. Appl. Phys., vol. 86, p. 850, 1999. [3] J. Gao, T. Gao, and M. J. Sailor, “Porous-silicon vapor sensor based on laser interferometry,” Appl. Phys. Lett., vol. 77, p. 901, 2000. [4] V. Mulloni and L. Pavesi, “Porous silicon microcavities as optical chemical sensors,” Appl. Phys. Lett., vol. 76, p. 2523, 2000. [5] P. A. Snow, E. K. Squire, P. St. J. Russel, and L. T. Canaham, “Vapor sensing using the optical properties of porous silicon Bragg mirrors,” J. Appl. Phys., vol. 86, p. 1781, 1999. [6] M. A. Muriel and A. Carballar, “Internal field distributions in fiber Bragg gratings,” IEEE Photon. Technol. Lett., vol. 9, pp. 955–957, July 1997. [7] J. E. Spanier and I. P. Herman, “Use of hybrid phenomenological and statistical effective-medium theories of dielectric functions to model the infrared reflectance of porous SiC films,” Phys. Rev., vol. 61, p. 10437, 2000. [8] P. Allcock and P. A. Snow, “Time-resolved sensing of organic vapors in low modulating porous silicon dielectric mirrors,” J. Appl. Phys., vol. 90, pp. 5052–5057, 2001. [9] V. S.-Y. Lin, K. Motesharei, K.-P. S. Dancil, M. J. Sailor, and M. R. Ghadiri, “A porous silicon-based optical interferometric biosensor,” Science, vol. 278, p. 840, 1997. [10] S. Chan, P. M. Fauchet, Y. Li, L. J. Rothberg, and B. L. Miller, “Porous silicon microcavities for biosensing applications,” Phys. Stat. Sol. A., vol. 182, p. 541, 2000. [11] L. De Stefano, I. Rendina, L. Moretti, A. M. Rossi, A. Lamberti, and P. Arcari, “Porous silicon microcavities in biochemical sensing,” in Sensor for Environmental Control, P. Siciliano, Ed, Singapore: World Scientific, 2003, pp. 46–50. [12] K.-P. S. Dancil, D. P. Greiner, and M. J. Sailor, “A porous silicon optical biosensor: Detection of reversible binding of IgG to a protein A-modified surface,” J. Amer. Chem. Soc., vol. 121, pp. 7925–7930, 1999.

53

[13] M. P. Stewart and J. M. Buriak, “Exciton-mediated hydrosilylation on photoluminescent nanocrystalline silicon,” J. Amer. Chem. Soc., vol. 123, p. 7821, 2001. [14] R. Boukherroub, J. T. C. Wojtyk, D. D. M. Wayner, and D. J. Lockwood, “Thermal hydrosilylation of undecylenic acid with porous silicon,” J. Electrochem. Soc., vol. 149, p. H59, 2002. [15] M. C. Desai and L. M. S. Stramiello, “Polymer bound EDC (P-EDC): A convenient reagent for formation of an amide bond,” Tetrahedron Lett., vol. 43, p. 7685, 1993. [16] W. Rychlik, W. J. Spencer, and R. E. Rhoads, “Optimization of the annealing temperature for DNA amplification in vitro,” Nucl. Acids. Res., vol. 18, p. 6409, 1990.

Luca De Stefano received the physics degree, working on light propagation in liquid crystal slab waveguides under external fields, and the Ph.D. degree in optical integrated hybrid devices based on pure and doped liquid crystals, from the University of Naples “Federico II” Naples, Italy, in 1992 and 1996, respectively. From 1996 to 2000, he was with the Italian Electric Power Company (ENEL), Research Center of Brindisi, Brindisi, Italy, involved mainly on material characterization and optical monitoring of environmental pollution. In 2001, he joined the Institute for Microelectronics and Microsystems, National Council of Research, Naples. His current research interests include opto-electronic devices and sensor systems.

Luigi Moretti received the Bachelor’s degree in physics, (cum laude) from the “Federico II” University of Naples, Italy, in 1999. He is currently working toward the Ph.D. degree in engineering from the University of Cosenza, Cosenza, Italy, under the supervision of Prof. G. Cocorullo. In July 1999, he received a fellowship from the Institute of Research for Electromagnetics and Electronic Components (IRECE-CNR). His research interests include the design and characterization of resonant optical microcavities fabricated exploiting the microelectronic silicon technology.

Annalisa Lamberti is a Postdoctoral Fellow in the Department of Biochemistry, University of Naples, Federico II, Naples, Italy. Her research interests are in the field of cell cycle regulation and apoptosis, genes expression, cloning of genes, use of expression vectors, purification and biochemical characterization of recombinant proteins, and more recently, in the field of biosensing.

Olimpia Longo is working toward the Ph.D. degree in the Department of Biochemistry, University of Naples, Federico II, Naples, Italy. Her research interests include the field of cell biology, cloning of genes, expression, purification and biochemical characterization of recombinant proteins, use of expression vectors and lately in the field of biosensing.

Massimiliano Rocchia was born in Turin, Italy, in 1972. He received the degree in chemistry from the University of Turin in 1999, under the direction of Prof. A. Zecchina, and received the Ph.D. degree under the direction of Prof. E. Garrone at the Politecnico of Turin in 2002. In 2002, he was a Postdoctoral Researcher at the Istituto Elettrotecnico Nazionale “Galileo Ferraris,” Turin. He is currently working on the surface chemistry of porous silicon materials and its application in the sensor field.

54


Andrea M. Rossi received the M.Sc. degree in physics from the University of Turin, Turin, Italy. He is currently a Researcher with the Istituto Elettrotecnico Nazionale “G. Ferraris” (IEN) Nanotechnology and Microsystems Department, Turin, Italy. He was the Marie Curie Research Fellow in 1998 at the Center of Research CIEMAT, Madrid, Spain, a Guest Researcher in 2000 at CNR IRECE, Naples, Italy, and a Guest Researcher in 2002 at the Center of Research, Julich, Germany. His current research interests are in Porous Silicon technology, sensors, and photonic crystals.

Paolo Arcari is a Full Professor of Chemistry at the University of Naples Medical School, Naples, Italy. His research experience is mainly in the field of biochemistry and molecular biology of enzymes. In particular, his interest is focused in the study of the structure/function relationship of proteins isolated from extremophilic organisms. Other recent research interests include the study of the expression of specific protein profiles in different type of eukaryotic cells and the use of DNA in biosensing.

Ivo Rendina (M’00) is a Senior Scientist with the National Council of Research (CNR), Naples, Italy, where he has been the Director of the Institute for Electromagnetism and Electronic Components and is currently responsible for the Department of Napoli, Institute for Microelectronics and Microsystems, Naples, Italy. He also teaches Electronics and Optoelectronics at the University of Calabria, Cosenza, Italy. His present research interests are in the field of silicon optoelectronics and microsystems. He is a Member of the “Consiglio di Presidenza” of the Società Italiana di Ottica e Fotonica and SPIE.