Detection of Nitric Oxide and Nitrogen Dioxide with Photoluminescent

Anal. Chem. 1996, 68, 3713-3717

Detection of Nitric Oxide and Nitrogen Dioxide with Photoluminescent Porous Silicon Jessica Harper and Michael J. Sailor*

Department of Chemistry and Biochemistry, The University of California at San Diego, La Jolla, California 92093-0358

The visible photoluminescence of porous Si is quenched by nitric oxide and nitrogen dioxide to detection limits of 1.4 × 10-3 and 5.3 × 10-5 Torr, respectively (corresponding to 2 ppm and 70 ppb). At analyte partial pressures in the low milliTorr range, the photoluminescence quenching is partially reversible; recovery from nitrogen oxide exposure occurs on a time scale of minutes. For both NO and NO2, the reversible photoluminescence quenching response fits a Stern-Volmer kinetic model. At higher partial pressures, quenching deviates from Stern-Volmer kinetics and some permanent loss of photoluminescence intensity occurs due to oxidation of the porous Si surface. Photoluminescence from porous Si is not quenched by nitrous oxide or carbon dioxide and only slightly quenched by carbon monoxide and oxygen.

A sensor specific for nitric oxide is currently a highly desirable device. For decades, monitoring of NO concentrations in industrial processes and pollution control has been accomplished by monitoring a conductivity change in SnO2-based commercially available sensors. However, its high operating temperature (150550 °C)1 and response time on the order of minutes makes SnO2 undesirable for some applications.2 For example, increased therapeutic administration of NO has led to a need for real-time monitoring of low levels of nitric oxide added to breathing gases in medical respirators.3-5 Researchers have explored modified SnO2,6-8 other metal oxides,1,9-11 high-Tc superconductors,12,13 conductive polystyrene,14 and specialized electrodes15-17 as alter(1) Wiegleb, G.; Heitbaum, J. Sens. Actuators B 1994, 17, 93-99. (2) Janata, J. Principles of Chemical Sensors; Plenum Press: New York, 1989. (3) Feldman, P. L.; Griffith, O. W.; Stuehr, D. J. Chem. Eng. News 1993, 71 (Dec 20), 26-38. (4) Speziale, G.; De Biase, L.; De Vincentis, G.; Ierardi, M.; Ruvolo, G.; La Francesca, S.; Scopinaro, F.; Marino, B. Thorac. Cardiovasc. Surg. 1996, 44, 35-39. (5) Rossaint, R.; Faulke, K. J.; Lopez, F.; et al. N. Engl. J. Med. 1993, 328, 399-405. (6) Schmitte, F.; Wiegleb, G. Sens. Actuators B 1991, 4, 473-4777. (7) Sberveglieri, G.; Groppelli, S.; Nelli, P. Sens. Actuators B 1991, 4, 457461. (8) Sberveglieri, G.; Groppelli, S.; Nelli, P.; Lantto, V.; Torvela, H.; Romppainen, P.; Leppavuori, S. Sens. Actuators B 1990, 1, 79-82. (9) Akiyama, M.; Tamaki, J.; Miura, N.; Yamazoe, N. Chem. Lett. 1991, 16111614. (10) Ishihara, T.; Shiokawa, K.; Eguchi, K.; Arai, H. Sens. Actuators 1989, 19, 259-265. (11) Ishihara, T.; Shiokawa, K.; Eguchi, K.; Arai, H. Chem. Lett. 1988, 9971000. (12) Huang, X. J.; Schooner, J.; Chen, L. Q. Sens. Actuators B 1994, 22, 211218. (13) Huang, X. J.; Schooner, J.; Chen, L. Q. Sens. Actuators B 1994, 22, 219226. S0003-2700(96)00642-7 CCC: $12.00

© 1996 American Chemical Society

native sensor materials. A highly specific, sensitive, and stable sensor has not yet been developed. Porous silicon is a high surface area network of silicon nanocrystallites which can be synthesized directly from the same type of silicon wafers that are used in the fabrication of microchips. The nanocrystallites in porous silicon photoluminesce visible light with an external quantum efficiency of up to 5%.18-21 A variety of reports have indicated that the visible photoluminescence from porous Si can be reversibly quenched by gases or liquids.22-26 Methods to chemically alter porous silicon’s photoluminescent properties have also been reported, and some of these reactions can tune the sensitivity of this material toward specific adsorbates.27-30 The photoluminescence quenching phenomenon thus forms the basis for a simple and inexpensive chemical sensor device.31 In this work we report that NO and NO2 can cause quenching of photoluminescence from nanocrystalline porous silicon and investigate this phenomenon for its potential application to NOx sensing. EXPERIMENTAL SECTION Sample Preparation and Treatment. Luminescent porous Si samples were prepared by an electrochemical etch of n-type Si [phosphorous doped, (100) orientation, International Wafer Service] of resistivity between 1 and 6 Ω cm. The etching solution was prepared by adding an equal volume of pure ethanol (Quantum Chemicals) to an aqueous solution of HF (48% by (14) Christensen, W. H.; Sinha, D. N.; Agnew, S. F. Sens. Actuators B 1993, 10, 149-153. (15) Yao, S.; Shimizu, Y.; Miura, N.; Yamazoe, N. Chem. Lett. 1992, 587-590. (16) Tierney, M. J.; Kim, H. L.; Madou, M.; Otagawa, T. Sens. Actuators B 1993, 13-14, 408-411. (17) Bediouis, F.; Trevin, S.; Devynck, J. J. Electroanal. Chem. 1994, 377, 295298. (18) Brus, L. Adv. Mater. 1993, 5, 286-288. (19) Brus, L. J. Phys. Chem. 1994, 98, 3575-3581. (20) Canham, L. T. Appl. Phys. Lett. 1990, 57, 1046-1048. (21) Cullis, A. G.; Canham, L. T. Nature 1991, 353, 335-338. (22) Fisher, D. L.; Harper, J.; Sailor, M. J. J. Am. Chem. Soc. 1995, 117, 78467847. (23) Fisher, D. L.; Gamboa, A.; Harper, J.; Lauerhaas, J. M.; Sailor, M. J. Mat. Res. Soc. Symp. Proc. 1995, 358, 507-518. (24) Lauerhaas, J. M.; Credo, G. M.; Heinrich, J. L.; Sailor, M. J. J. Am. Chem. Soc. 1992, 114, 1911-1912. (25) Kelly, M. T.; Chun, J. K. M.; Bocarsly, A. B. Nature 1996, 382, 214-215. (26) Chun, J. K. M.; Bocarsly, A. B.; Cottrell, T. R.; Benziger, J. B.; Yee, J. C. J. Am. Chem. Soc. 1993, 115, 3024-3025. (27) Lee, E. J.; Ha, J. S.; Sailor, M. J. Mat. Res. Soc. Symp. Proc. 1995, 358, 387-392. (28) Lee, E. J.; Ha, J. S.; Sailor, M. J. J. Am. Chem. Soc. 1995, 117, 8295-8296. (29) Lee, E. J.; Bitner, T. W.; Ha, J. S.; Shane, M. J.; Sailor, M. J. J. Am. Chem. Soc. 1996, 118, 5375-5382. (30) Lauerhaas, J. M.; Sailor, M. J. Science 1993, 261, 1567-1568. (31) Sailor, M. J.; Credo, G.; Heinrich, J.; Lauerhaas, J. M. Method for Detection of Chemicals by Reversible Quenching of Silicon Photoluminescence. U.S. Patent 5,338,415, issued August 16, 1994.

Analytical Chemistry, Vol. 68, No. 21, November 1, 1996 3713

weight; Fisher Scientific). The etching cell was constructed of Teflon and was open to air. Si wafers were cut into squares with a diamond scribe and mounted in the bottom of the Teflon cell with a Viton O-ring seal, exposing 0.3 cm2 of the Si surface. Electrical contact was made to the back side of the Si wafer with a strip of heavy Al foil. A loop of Pt wire was used as a counter electrode. The exposed Si face was illuminated with ∼120 mW/ cm2 of white light from a 300 W tungsten lamp for the duration of the etch. Etching was carried out as a two-electrode galvanostatic experiment at an anodic current density of 5 mA/cm2 for 33 min. After etching, the samples were rinsed in ethanol, dried under a stream of N2(g), and placed in an evacuable glass chamber. The glass dosing chamber was connected to a glass-and-Teflon Schlenk line32 via an O-ring joint. Pressure inside the manifold was monitored using an MKS Baratron pressure gauge, containing two separate pressure transducers calibrated for the pressure ranges from 10-5 to 1 Torr and from 1 to 1000 Torr. The chamber was pumped to 0.2 Torr) of NO, the response of porous Si is much less reversible. We never observed complete quenching of photoluminescence by NO; even at the highest pressures studied (70 Torr) there was a small amount of photoluminescence remaining and this increased in intensity when the gas was evacuated. Thus, although the sensitivity is drastically reduced, some porous Si luminescence is still recovered even after exposure to higher NO pressures. Quenching of Photoluminescence by Nitrogen Dioxide. Exposure of porous Si to NO2 results in behavior similar to that observed for exposure to NO. Photoluminescence quenching is reversible at NO2 partial pressures below 4 × 10-4 Torr, and quenching follows the Stern-Volmer model (Figure 4). Porous Si exhibits greater sensitivity for NO2 than it does for NO; the smallest amount of NO2 detectable is less than 5.3 × 10-5 Torr, or 70 ppb. Because NO is known to react with O2 to form NO2, it is possible that measurements of NO quenching are interfered with 3716 Analytical Chemistry, Vol. 68, No. 21, November 1, 1996

by the presence of oxygen in the vacuum system. A small impurity of NO2 could greatly skew the response. However, when air was purposely introduced to a flask containing porous Si and low pressures of nitric oxide, the air caused no further quenching. The reaction mechanism to form NO2 from NO is termolecular; two molecules of NO must collide simultaneously with a molecule of O2.37 Thus at the low partial pressures of NO being examined, the rate of formation of significant quantities of NO2 was probably slower than the time scale of the measurements. Test for Photolysis of Nitrogen Dioxide. Photolysis of NO2 is known to occur at wavelengths below 430 nm to produce NO and O atoms.37 Although the quantum yield for this reaction is ∼0.8 at 397 nm, it drops to 0.023 by 420 nm.37 To test whether photolysis of NO2 is important under our experimental conditions, comparison was made between a porous Si sample exposed to 5 Torr of NO2 in light (435 nm, 5 mW/cm2) and in the dark. A freshly etched porous Si sample was exposed to NO2 in the presence of light for 10 s and then the chamber was evacuated for 30 min. The intensity of the photoluminescence spectrum recorded at this point was reduced to 3.5% of the original spectrum. A second fresh porous Si sample was treated in the same manner, except that the sample was not illuminated during NO2 exposure. The intensity of the photoluminescence spectrum was 9.5% of the initial vacuum spectrum. Thus both samples displayed significant losses in photoluminescence intensity, although the sample exposed to the 435 nm light exhibited a slightly greater relative loss. The infrared spectra of both samples taken after NO2 exposure showed large ν(OSi-Hx) ν(Si-O) bands in the 21902260 and 1040-1200 cm-1 regions, respectively, indicative of significant sample oxidation. The irradiated sample appeared to have slightly larger oxide peaks. Aside from this small difference, the spectra appear identical. Therefore we conclude that thermal oxidation by NO2 is the primary degradation pathway in this system, with a minor contribution from NO2 photolysis. Mechanism of Reaction: Comparison to NO Adsorption on Single Crystal Si. A controversy exists regarding the mechanism of adsorption of NO on Si.38-40 A mechanism similar to that for O2 has been proposed, in which NO adsorbs as a stable molecular species on the surface and then either desorbs or dissociates into surface-bound N and O molecules.41-44 However, the temperature at which NO can remain a stable adsorbed species, the likelihood for desorption, and the identity of the dissociated products are unclear. Some workers detect significant amounts of N2O as a dissociation product.41 Many reports show that NO must adsorb dissociatively at room temperature,38,45-47 but some observe only molecular species as the adsorbate.48,49 (37) Finlayson-Pitts, B. J.; Pitts, J. N. Atmospheric Chemistry: Fundamentals and Experimental Techniques; John Wiley and Sons: New York, 1986. (38) Taguchi, Y.; Fujisawa, M.; Kuwahara, Y.; Onchi, M.; Nishijima, M. Surf. Sci. 1989, 217, L413-L416. (39) Taguchi, Y.; Fujisawa, M.; Nishijima, M. Surf. Sci. 1990, 233, L251-L252. (40) Wormeester, H.; van Silfhout, A.; Keim, E. G.; Sasse, A. G. B. M. Surf. Sci. Lett. 1990, 233, L249-L250. (41) Ying, Z. C.; Ho, W. J. Chem. Phys. 1989, 91, 2689-2705. (42) Avouris, P.; Bozso, F.; Hamers, R. J. J. Vac. Sci. Technol. B 1987, 5, 13871392. (43) Richter, L. J.; Buntin, S. A.; King, D. S.; Cavanagh, R. R. J. Electron Spectrosc. Relat. Phenom. 1990, 54/55, 181-190. (44) Richter, L. J.; Buntin, S. A.; King, D. S.; Cavanagh, R. R. Phys. Rev. Lett. 1990, 65, 1957-1960. (45) Song-Bao, F.; Ru-Hong, Z.; Pei-Lin, C.; Jing-Chang, T. Surf. Sci. Lett. 1991, 247, L224-L228. (46) Rangelov, G.; Stober, J.; Eisenhut, B.; Fauster, T. Phys. Rev. B 1991, 44, 1954-1957.

Although there is agreement that dangling bonds are involved in adsorption, some researchers have found that the dangling bonds are capped by both N and O atoms,45,46 while others find dangling bonds exclusively occupied by O.41,42 Ying and Ho reported that irradiation of Si(111)-(7 × 7) at 90 K with UV, visible, or IR light enhances dissociation of NO.41 This leads to a surface covered with O and N atoms via an N2O intermediate. While they recognize that photogenerated charge carriers must play an important role, the mechanism of the photoreaction is not known. Because quenching of porous Si by low pressures of NO and NO2 follows reversible Stern-Volmer behavior, it is probable that an adsorption mechanism is operative. Subsequent dissociation and reaction with the porous Si surface would then account for the oxidation reaction that is observed at higher NOx pressures or longer exposure times. The oxidation reaction presumably introduces efficient nonradiative surface recombination centers that account for the lower photoluminescence intensity and reduced sensitivity of oxidized porous Si. Although surface traps introduced by surface oxidation reactions can account for the irreversible quenching of porous Si photoluminescence, the photophysical mechanism responsible for reversible quenching is not known. Photoluminescence from porous Si has been shown to be quenched by both energy transfer and charge transfer mechanisms.22,50,51 Presumably one or both of these mechanisms is operative in the NOx system. Further work involving characterization of the transient photoproducts is required in order to address this question. Other Gases. No other gas tested on porous Si evokes a reversible quenching response at pressures as low as those demonstrated for the nitrogen oxides. With 760 Torr of CO2 or N2O, no change in the photoluminescence intensity was perceptible. Because CO2 and N2O do not quench photoluminescence from porous Si, they would not be interferants to NO or NO2 detection. CO displays a minimal amount of photoluminescence quenching, ∼5%, at 760 Torr. Greater than 10 Torr of O2 is required to quench porous Si photoluminescence measurably. Air quenches photoluminescence by ∼20%, slightly more than predicted based on the partial pressure of O2 it contains. Water vapor, which quenches photoluminescence from oxidized porous Si, is the probable cause of the discrepancy.30 When mixtures of NO (47) Bhat, M.; Kamath, A.; Kwong, D. L. Appl. Phys. Lett. 1994, 65, 13141316. (48) Sasse, A. G. B. M.; van Silfhout, A. Phys. Rev. B 1989, 40, 1773-1782. (49) Ekwelundu, E.; Ignatiev, A. Surf. Sci. 1987, 179, 119-131. (50) Rehm, J. M.; McLendon, G. L.; Fauchet, P. M. J. Am. Chem. Soc. 1996, 118, 4490-4491. (51) Yan, F.; Bao, X.; Wu, X.; Chen, H. Appl. Phys. Lett. 1995, 67, 1-3. (52) Lauerhaas, J. M.; Sailor, M. J. Mat. Res. Soc. Symp. Proc. 1993, 298, 259263. (53) Chandler-Henderson, R. R.; Sweryda-Krawiec, B.; Coffer, J. L. J. Phys. Chem. 1995, 99, 8851-8855. (54) Coffer, J. L.; Lilley, S. C.; Mariin, R. A.; Files-Sesler, L. A. J. Appl. Phys. 1993, 74, 2094-2096. (55) McConnell, H. M.; Owicki, J. C.; Parce, J. W.; Miller, D. L.; Baxter, G. T.; Wada, H. G.; Pitchford, S. Science 1992, 257, 1906-1912. (56) Canham, L. T. Adv. Mater. 1995, 7, 1033-1037.

or NO2 with air were tested, there did not appear to be a significant deviation from the Stern-Volmer curves derived from NOx/N2 mixtures. Previous work has shown that porous Si photoluminescence is efficiently quenched by halogen gases.23,30,52 The halogens react with the porous Si surface, which leads to irreversible photoluminescence quenching. Aqueous and gasphase acids and bases have been shown to reversibly affect the intensity of photoluminescence from porous Si.26,53,54 In particular, Bocarsly and co-workers have recently shown that SO2 gas can quench luminescence from oxidized porous Si at levels as low as 440 ppb.25 It is interesting to note that the authors of the SO2 quenching study reported that porous Si luminescence was completely insensitive to NOx gases. The discrepancy probably arises from the chemical nature of the porous Si samples used; Bocarsly and co-workers used highly oxidized porous Si, while the current study used only mildly oxidized samples. Vapors of alcohols, aromatic and aliphatic hydrocarbons, ethers, and dichloromethane have also been shown to quench porous Si photoluminescence reversibly, although higher pressures are required to induce measurable quenching (typically >1 Torr).24,30 CONCLUSIONS Porous Si photoluminescence is quenched very efficiently by NO and NO2. At low pressures quenching is reversible on the time scale of minutes, but at pressures higher than ∼10-3 Torr, irreversible oxidation interferes with the reversible quenching process and leads to a reduction in sensitivity. Reversible quenching follows Stern-Volmer kinetics and so provides a means to develop an optical NOx sensor. A porous Si-based NOx sensor could be used to detect both small amounts (ppm or ppb) of NOx and the larger amounts of NOx that can overload conventional sensors.2 The lower detection limit is competitive with currently available sensors. Silicon has been incorporated in a variety of biological sensor devices,55 and porous Si has been shown to be compatible with biochemical media.56 Luminescent porous Si provides several advantages for application as a detector of NO: it is compatible with silicon microfabrication technology, its photoluminescence is relatively insensitive to other common gases, it operates at room temperature, and detection can be made immediately at the source of NOx. The drift in sensitivity that accompanies oxidation of porous Si places a limit on the use of this material to either short-term or single-use applications. ACKNOWLEDGMENT This work was supported by the National Science Foundation (DMR-9220367). M.J.S. acknowledges support from Camille Dreyfus and A. P. Sloan Fellowships. The authors thank Professor William C. Trogler and Dr. Joseph MacNeil for helpful discussions and experimental assistance. Received for review July 1, 1996. Accepted September 9, 1996.X AC960642Y X

Abstract published in Advance ACS Abstracts, October 1, 1996.

Analytical Chemistry, Vol. 68, No. 21, November 1, 1996

3717