All-Optical Hydrogen-Sensing Materials Based on Tailored Palladium

Cite this:Anal. ... The response time peaks in the α → β phase transition region (1625 s at 0.4% H2 for ..... Fibre optics in palladium-based hydr...
0 downloads 0 Views 105KB Size
Download PDF
Anal. Chem. 2004, 76, 6321-6326

All-Optical Hydrogen-Sensing Materials Based on Tailored Palladium Alloy Thin Films Z. Zhao,† Y. Sevryugina,‡ M. A. Carpenter,*,† D. Welch,§ and H. Xia§,|

College of NanoScale Science and Engineering and Department of Chemistry, University at AlbanysSUNY, Albany, New York 12203, and MTI-Instruments, 325 Washington Avenue Extension, Albany, New York 12205

Optical reflectance measurements were performed to determine the hydrogen response characteristics of 20nm-thick Pd-Au (Ag) films. The response characteristics displayed a strong dependence on r, mixed r/β, and β Pd-hydride phases formed in the films. The response time peaks in the r f β phase transition region (1625 s at 0.4% H2 for Pd0.94Ag0.06 and 405 s at 1% H2 for Pd0.94Au0.06), consistent with critical slowing down phenomena. The r f β phase transition region was shifted and inhibited by changing the alloy element to Au and increasing its corresponding content to 40 atom %, respectively. Initial hydrogen uptake rate measurements determined that, due to the adsorption of ambient background gases, the rate-limiting step for r or β phase PdH formation is dissociative chemisorption of hydrogen for each palladium alloy film. By tuning the alloy content and composition of the palladium films, the surface properties of the film become more receptive toward the rapid detection of hydrogen and a novel hydrogen-sensing material using Pd alloyed with 40 atom % Au is presented. The demand for using hydrogen as a next-generation, clean, and renewable energy source has stimulated considerable efforts toward developing sensitive, reliable, and cost-effective hydrogen sensors for the fast detection of hydrogen leaks below the lower explosive limit (LEL) of 4.65 vol % ratio of hydrogen to air.1-5 Currently, Pd or Pd-based alloys are commonly used as the sensing material, due to Pd’s high sensitivity and selectivity toward hydrogen. However, the reported sensor material response times are often not short enough, ranging, for example, from tens of seconds to several minutes at hydrogen concentrations of 1 vol %.6-8 * To whom correspondence should be addressed. E-mail: mcarpenter@ uamail.albany.edu. † College of NanoScale Science and Engineering, University at Albanys SUNY. ‡ Department of Chemistry, University at AlbanysSUNY. § MTI-Instruments. | Current address: Advanced Photonics & Electronics Laboratory, General Electric Global Research Center, One Research Circle, Niskayuna, NY 12309. (1) Bevenot, X.; Trouillet, A.; Veillas, C.; Gagnarie, H.; Clement, M. Sens. Actuators, B 2000, 67, 57. (2) Tabib-Azar, M.; Sutapun, B.; Petrick, R.; Kazemi, A. SPIE 1998, 3513, 80. (3) DiMeo, F., Jr.; Chen, I.; Chen, P.; Neuner, J.; Stawasz, M.; Welch, J. Proc. 2002 DOE Hydrogen Program Rev., U. S. Department of Energy, 2002. (4) Dwivedi, D.; Dwivedi, R.; Srivastava, S. K. Sens. Actuators, B 2000, 71, 161. (5) Chen, Y.; Li, Y.; Lisi, D.; Wang, W. Sens. Actuators, B 1996, 30, 11. 10.1021/ac0494883 CCC: $27.50 Published on Web 10/02/2004

© 2004 American Chemical Society

Response times of these materials are determined by the dynamic equilibrium process between the hydrogen-containing gas and the Pd-H system through steps involving hydrogen surface adsorption/desorption, surface dissociation/association, bulk diffusion, and Pd-hydride formation. The equilibration process is dependent on both the quality of the Pd material and the PdH phase equilibria. Such an equilibration process on pure Pd (UHV prepared/cleaned and characterized) involves the barrierless dissociation of hydrogen on the Pd surface, followed by diffusion of the surface H atoms into the Pd film. Such a process is limited by the diffusion time of the H atoms through the Pd sample with the corresponding diffusion coefficients DR ) 10-7 cm2/s and Dβ ) 10-6 cm2/s for R and β phase PdH. Efforts to enhance the response time of Pd-based hydrogen sensors have focused on using thinner Pd or Pd alloy films.2,5,8-11 While thinner films have enhanced the response time, nanometer-scale films that have been used to date still do not achieve response times predicted by a diffusion-limited process (∼40 µs, for a 20-nm-thick R-phase PdH sample). Therefore, a key step in the design process of these sensors is to develop an understanding of the surfacelimiting interactions that will undoubtedly play a role in the response time dynamics for sensor devices exposed to ambient conditions. We believe that what has been overlooked in the development process of these sensors is an understanding of the surface-limiting interactions that will undoubtedly play a role in response time dynamics for sensor devices exposed to ambient conditions. Evidence for such surface perturbations in determining physical interaction properties of surfaces is common in the scientific literature, and it is well known that binding energies and adsorption kinetics of UHV-cleaned materials are dramatically different from those exposed to ambient or various gaseous environmental conditions.12,13 Therefore, unless special care is taken to preserve an atomically clean surface, it is reasonable to assume that a field-deployed sensing system will have adsorbed (6) Jakubil, W. P.; Urbanczyk, M. W.; Kochowski, S.; Bodzenta, J. Sens. Actuators, B 2002, 4179, 1. (7) Baselt, D. R.; Fruhberger, B.; Klaassen, E.; Cemalovic, S. C. L. Britton Jr., Patel, S. V.; Mlsna, T. E.; McCorkle, D.; Warmack, B. Sens. Actuators, B 2002, 6810, 1. (8) Bevenot, X.; Trouillet, A.; Veillas, C.; Gagnarie, H.; Clement, M. Meas. Sci. Technol. 2002, 13, 118. (9) Walter, E. C.; Favier, F.; Penner, R. M. Anal. Chem. 2000, 74, 1546. (10) Hughes, R. C.; Schubert, W. K. J. Appl. Phys. 1992, 71, 542. (11) Hunter, G. W.; Bickford, R. L.; Jansa, E. D.; Makel, D. B.; Liu, C. C.; Wu, Q. H.; Powers, W. T. SPIE 1994, 2270, 77. (12) Attard, G.; Barnes, C. Surfaces; Oxford University Press: New York, 1998. (13) Gravil, P. A.; Toulhoat, H. Surf. Sci. 1999, 430, 176.

Analytical Chemistry, Vol. 76, No. 21, November 1, 2004 6321

large quantities of water, oxygen, carbon monoxide, hydrocarbons, and other compounds specific to the sensing environment. Of the above specific compounds, many are known to interfere with the adsorption of hydrogen in catalytic or UHV surface reaction studies, thus providing a further indication that surface interactions will lead to longer response times.14 The material transport dynamics that occur at the R f β PdH phase transition are another rate-limiting step that directly affects the response time of Pdbased hydrogen sensors. The phase transition has been observed by many research groups, and specifically a response time as long as 3600 s has been observed for an optical-based sensor and a factor of 100 increase (50 ms as compared to 5 s at the phase transition) has been observed for a Pd nanoparticle electrical-based hydrogen sensor.9,15 The PdH R f β phase transition occurs at a specific hydrogen concentration, which is dependent on the palladium film thickness, alloying element, and quantity. It is well known that upon exposure to hydrogen, palladium and palladium alloy films absorb hydrogen and form PdH. This process causes an increase in volume and a subsequent decrease in the free electron density of the material. Formation of PdH leads to an increase in the Fermi level with increasing H concentration and a subsequent decrease in both the real and imaginary parts of the dielectric function, resulting in a decrease in the reflectivity of the thin films.16,19,20 This paper demonstrates optical reflectance response characteristics of 20-nm-thick Pd-Au and Pd-Ag alloy films toward hydrogen exposure. Specifically, we have performed experiments that have demonstrated the dependence of both the response time and the signal change as a function of hydrogen concentration for Pd0.94Ag0.06, Pd0.94Au0.06, and Pd0.6Au0.4 films. The low alloy content films have response time characteristics that are dominated by critical phase effects between 0.3 and 1% H2 concentrations. An analysis of the initial kinetics of the Pd alloy and hydrogen system determined that, in general, surface effects and not H atom diffusion limited the initial response characteristics of these films. The initial hydrogen uptake kinetics display a dependence on the specific alloy film that was investigated, indicating that the surface composition of the adsorbed background gases is a function of changes in the Fermi level of the alloy films. The Pd0.6Au0.4 films have an alloy composition that is above the critical isotherm threshold and thus do not experience an R f β phase transition. These films have a response time limited only by surface interface effects. We likewise introduce the Pd0.6Au0.4 film as a novel hydrogen-sensing material developed for hydrogen detection below the LEL, whose reduction in response time at the hydrogen levels of interest was possible by controlling and selecting an appropriate Pd alloy and the corresponding alloy concentration. EXPERIMENTAL SECTION Film Deposition. The 20-nm Pd0.94Ag0.06, Pd0.94Au0.06, and Pd0.6Au0.4 thin films were deposited on glass substrates using a Denton (14) Peden, C. H. F.; Kay, B. D.; Goodman, D. W. Surf. Sci. 1986, 175, 215. (15) Kalli, K.; Othonos, A.; Christofides, C. J. Appl. Phys. 2002, 91, 3829. (16) Fortunato, G.; Bearzotti, A.; Caliendo, C.; D’Amico, A. Sens. Actuators 1989, 16, 43. (17) Okuhara, Y.; Takata, M. Bull. Mater. Sci. 1999, 22, 85. (18) Schwarz, R. B.; Khachaturyan, A. G. Phys. Rev. Lett. 1995, 74, 2523. (19) Tobiska, P.; Hugon, O.; Trouillet, A.; Gagnaire, H. Sens. Actuators, B 2001, 74, 168. (20) Mandelis, A.; Garcia, J. A. Sens, Actuators, B 1998, 49, 258.

6322 Analytical Chemistry, Vol. 76, No. 21, November 1, 2004

Figure 1. Schematic representation of the hydrogen sensor test bench.

Desk II sputtering system, with the resulting nanocrystallites ranging in size from ∼ 6 to 10 nm, measured using XRD analysis techniques. The Pd0.6Au0.4 film was annealed at 200 °C in an argon atmosphere for 1 h, which resulted in a factor of 4 enhancement of the reflectance signal change upon hydrogen exposure, with no affect on the corresponding response time. The Pd0.94Ag0.06 and Pd0.94Au0.06 data presented here are not the result of thermally annealed films, as the above annealing process results in an undesirable increase in response time at all of the analyzed hydrogen concentrations. Sensor Test Station. A schematic of the sensor testing station is displayed in Figure 1. The reflectance change of the Pd alloy film was monitored as a function of hydrogen (99.999% purity) concentration in a balance of nitrogen (99.999 purity). A 20-cm3 stainless steel chamber was held at room temperature and a total pressure of 760 Torr for the hydrogen exposure experiments. The 1200 sccm of N2 or a H2/N2 mixture was sequentially admitted to the chamber at precisely controlled concentrations by an Environics S4000 gas-mixing system. The response characteristics of the Pd alloy films were measured by an intensity modulation-based fiber-optic sensor, MTI Fotonics Sensor model 1000, which utilizes a bifurcated optical fiber bundle to direct a white light source to the sample and collect the reflected light for detection using a silicon photodiode. To prevent measured signal drift, the Fotonics 1000 system utilizes a control feedback circuit for a constant white light power source. After each test, the chamber was opened to air to aid the recovery of the sample reflectance signal to the initial level observed in N2. RESULTS AND DISCUSSION Sensor Response Characteristics. Figure 2 displays the reflectance response of Pd0.94Ag0.06, Pd0.94Au0.06, and Pd0.6Au0.4 films as a function of gas exposure time at hydrogen concentrations ([H2]) ranging from 0.1 to 4% in nitrogen. The reflectance change from the baseline value in nitrogen to the saturation value in hydrogen increases with an increase in [H2], while the response time variation is quite different for the three samples, where the response time is defined as that required for 90% of the full signal change. For Pd0.94Ag0.06 (Figure 2a), the response time has a maximum of 1625 s at 0.4% H2, while for Pd0.94Au0.06 (Figure 2b), the response time maximizes at 405 s at 1% H2. It is clear from Figure 2a and b that the hydrogen adsorption and desorption processes that lead to a reduction and an increase, respectively, in the optical reflectance from these films proceeds through processes with very different time scales. Such a hysteresis effect

Figure 3. Hydrogen concentration dependence of reflectance signal change (a) and response time (b) for ∼20-nm thin films of Pd0.94Ag0.06, Pd0.94Au0.06, and Pd0.6Au0.4.

Figure 2. Optical reflectance response of ∼20-nm thin films of Pd0.94Ag0.06 (a), Pd0.94Au0.06 (b), and Pd0.6Au0.4 (c) to exposure of different hydrogen gas concentrations. Note that (a) has a larger time scale while (b) and (c) share the same time scale.

has been seen previously with Pd and low-content Pd alloy films and has been attributed to differences in the thermodynamics of the adsorption and desorption processes arising from coherent strains induced by the R f β PdH phase transformation.17,18 In contrast, the signal for Pd0.6Au0.4 saturates quickly at each of the investigated [H2] (e.g., only 48 s at 0.1% H2) as shown in Figure 2c. The Pd0.6Au0.4 film also exhibits an almost complete reversion from Pd hydride to Pd in less than 2 min. These results illustrate that the Pd0.6Au0.4 sensing film is a very appropriate material for designing hydrogen safety sensors (for detection below LEL) as its response time is only 8 s as compared to 370 and 405 s at 1% H2 for Pd0.94Ag0.06 and Pd0.94Au0.06 films, respectively. The overall response time of this material is also faster than other current optical-based sensing materials reported in the literature; for example, at 1% H2, its response time (8 s) is 1 order of magnitude shorter than the 90 s reported for Pd0.9Ni0.1 alloy films.7,19-21 The sensitivity of the Pd0.6Au0.4 material is slightly reduced to ∼100300 ppm compared to ∼20-50 ppm for the high-content Pd (>90%) alloy films. Likewise, the signal-to-noise levels in Figure 2c are lower by a factor of ∼2 for the Pd0.6Au0.4 alloy film, which is attributed to the lower palladium content of this film. Optimization of the signal-to-noise ratio through changes in the optical source and detection electronics is underway and clearly is important for the future development of prototypes. Furthermore, (21) Chtanov, A.; Gal, M. Sens, Actuators, B 2001, 79, 196.

a comprehensive study of the sensing properties of the Pd0.6Au0.4 film as a function of temperature, humidity, carbon monoxide, and the corresponding reliability testing is underway and will be detailed in an upcoming publication. The signal change and response time characteristics of the investigated materials are compared and presented in panels a and b of Figure 3, respectively. The signal changes of both Pd0.94Ag0.06 and Pd0.94Au0.06 films display a nonlinear dependence on ln[H2] involving three regions, a small variation region below ∼0.3% H2, a saturation region above ∼1 or ∼1.5% H2, and a rapid increase region between these limits for the two materials, respectively. If one compares the data with pressure-composition isotherms of the H-Pd alloy systems,22 the three regions can be recognized as R, mixed R/β, and β crystallographic phases (all fcc). However, the signal change from Pd0.6Au0.4 films presents a linear dependence on ln[H2], which is typical of pressure-composition isotherms of H-Pd alloy systems with only one phase of palladium hydride. The response time versus [H2] shown in Figure 3b is seen to increase with a decrease in [H2] in both the R and β phase regions of Pd0.94Ag0.06 and Pd0.94Au0.06 films. This dependence is believed to result from a surface-related rate-limiting step, as discussed later. It is interesting to note that, in the mixed R/β phase region, the response time as a function of [H2] is characterized with a sharp peak, which is consistent with critical slowing down phenomena, as slight deviations of [H2] from that point (0.4% H2 for Pd0.94Ag0.06 and 1% H2 for Pd0.94Au0.06) leads to a substantial decrease in the response time. It is apparent that in this region the response time is rate-limited by a phase transition-related bulk film property. Kalli et al. reported a similar behavior of pure Pd (22) Lewis, F. A. The Palladium Hydrogen System; Academic: New York, 1967.

Analytical Chemistry, Vol. 76, No. 21, November 1, 2004

6323

with a maximum response time of 3600 s at a hydrogen concentration of ∼1.8%.23 They preliminarily attributed the peak in response time as mirroring the maximum expansion of the crystal lattice in the R and mixed R/β phase region, with the increase in equilibration time coinciding with the increase in time required for the expansion of the crystal lattice. However, this is not consistent with the decrease in response time they observed in the β phase region, where the lattice expansion is as high as 3.5%.22 By utilization of 10% Au rather than 10% Ag as the alloying component, the phase transition shifts, as shown in Figure 3, to higher hydrogen concentrations. Thus, by tuning the alloy composition, it is possible to shift the phase transition to different hydrogen concentrations, thus avoiding long response times due to critical slowing down type behavior. This alloy compositioninduced shifting of the phase transition is further exemplified by the Pd0.6Au0.4 data in Figure 3. The response time curve for the Pd0.6Au0.4 film contains no maximum and has reduced response times at all of the investigated hydrogen concentrations. The absence of a peak in response time is due to the inhibition of the R f β phase transition caused by a 40 atom % Au alloy content, as alloying more than 17% of Au in Pd will lead to a critical isotherm, where the R and β phases merge to one phase.22 The response time shows a negative power-law dependence on [H2], and we ascribe this dependence to a surface-related ratelimiting step. A close analysis of Figure 3b shows that the Pd0.94Ag0.06 and Pd0.94Au0.06 materials also show a similar power law dependence of the response time in both the R and β PdH phase regions. However, these characteristics are perturbed by the dominant time dynamics associated with the R f β phase transition, which leads to the nonlinear dependence of signal change on ln(H2). As the signal change and response time curves are characteristic of the PdH phase transformation, a deeper understanding of the dynamics of the response curves in Figure 2a and b is warranted. At hydrogen concentrations well below and above the critical hydrogen concentration, the optical response curves, representing the adsorption of hydrogen, reach equilibrium with apparently a single dynamical process as there are no breaks in the slope of the signal change versus time plots. However, at (and near) the critical concentration (e.g., 0.4% for Pd0.94Ag0.06 and 1.0% for Pd0.94Au0.06), the hydrogen response curves representing the adsorption of hydrogen consist of a primary decrease roughly within ∼700 or 300 s of hydrogen exposure, respectively, followed by a small and long-time secondary decrease in reflectance as the palladium alloy film approaches equilibrium with the hydrogen/ nitrogen mixture. The details of the primary and secondary reflectance changes as a function of [H2], internal stress, film thickness, alloy, and alloy concentration are currently under further investigation. We speculate however, that during the secondary decrease, after the formation of the initial mixed R/β phase, the driving force for further R f β phase transformation decreases. A reduction in the free energy required to form PdH may be attributed to the formation and release of locally strained mixed grain boundaries, with the sluggish response time at (and near) the critical point marking a slow advance of the R f β phase transition. (23) Kalli, K.; Othonos, A.; Christofides, C. J. Appl. Phys. 2002, 91, 3829.

6324

Analytical Chemistry, Vol. 76, No. 21, November 1, 2004

Initial Hydrogen Uptake Analysis. While the response time characteristics clearly have a long-time component due to the phase transition of the low alloy content films, an analysis of the initial kinetics of the system can provide further insight toward the design of the sensing material. The interpretation of the initial kinetics of the Pd-H system can be understood by considering the following scheme:

H2(gas) T 2H(adsorbed) H(adsorbed) T H(bulk) where

[H]adsorbed R [H2]gas1/2

H2 molecules impinging on the Pd surface must first undergo dissociative chemisorption, which is then followed by atomic hydrogen diffusion into the Pd film. The pressure dependence of the initial hydrogen uptake rate depends on which of these processes is rate limiting. If the dissociative chemisorption process is the rate-limiting step, then a plot of the initial uptake rate versus the gas-phase hydrogen concentration would produce a line linearly dependent on [H2]. However, if diffusion into the Pd film is rate limiting, then within the limits of Sievert’s law, a plot of the initial hydrogen uptake rate versus H2 concentration would be linear with respect to [H2]1/2. The kinetics of hydrogen uptake on UHV-cleaned Pd(110) bulk crystals was studied by Kay et al., and they determined that the initial hydrogen uptake rate was indeed limited by H atom diffusion; likewise, the half-life of the reaction was determined to be independent of [H2], and therefore the reaction was first order with respect to [H2].24 These experiments correlate well with the fundamental properties of a wellcharacterized system, in that H2 dissociates with near-unit probability on atomically clean Pd through a barrierless dissociative chemisorption interaction and the formation of PdH is limited by H atom diffusion into the bulk material. The reflectance experiments we have performed detect changes in the optical properties of the films as they undergo a transformation from metallic characteristics to the more insulating PdH structure. Therefore, changes in reflectance can be qualitatively compared to the amount of metallic Pd or insulating PdH present in the film. Initial rates of hydrogen uptake were determined by measuring the rate of reflectance change (V/s) during the initial stages of hydrogen exposure, where the reflectance is changing in a linear fashion. The reaction order, n, of the initial rate dependence for hydrogen uptake was determined from the slope of the log(initial rate) versus log[H2] as shown in Figure 4.25 Error bars represent experimental reproducibility of the corresponding initial rate for hydrogen uptake. As shown in Figure 4, each of the Pd alloys are approximately first order with respect to [H2], and qualitatively, the Pd0.94Ag0.06 alloy has the smallest overall rate constant as indicated by the y-intercepts. The near-first-order dependence clearly indicates that the process is limited by surface interactions and not by atomic hydrogen diffusion into the 20(24) Kay, B. D.; Peden, C. H.; Goodman, D. W. Phys. Rev. B 1986, 34, 817. (25) Atkins, P. W.; Physical Chemistry, 4th ed.; W. H. Freeman and Co: New York, 1990.

Figure 4. Log(initial rate) vs Log([H2]) for 20-nm-thick Pd0.94Au0.06, Pd0.6Au0.4, and Pd0.94Ag0.06 films.

Figure 5. Log(t1/2) vs Log([H2]) for 20-nm-thick Pd0.94Au0.06, Pd0.6Au0.4, and Pd0.94Ag0.06 films.

nm-thick film. Intuitively, the surface interaction limitation is not surprising as we have taken no special measures to clean the surface before or during the exposure experiments. Furthermore, these experiments more appropriately mimic a field-deployed environment and thus lead to an understanding of the limitations of the sensing layer. An analysis of the half-life of the hydrogen uptake rate and its dependence on [H2] is useful for gaining an understanding of the reaction order of the system at longer interaction times. The halflife in this experiment is defined as the time required for the reflectance to decrease to half of its baseline value prior to hydrogen introduction. Figure 5 shows a bilogarithmic plot of t1/2 versus [H2] with the resulting slope equal to 1 - n for the three palladium alloy films and the y-intercept being inversely related to the reaction rate constant. Error bars represent experimental reproducibility of the corresponding half-life measurement. As can be seen, the half-life dependence on the reaction order is equal to 1.6 for each of the three alloys. The increased dependence on [H2] for the half-lives as compared to an atomically clean Pd surface, where n ) 1, is further evidence that adsorbed background gases such as O2, CO, and hydrocarbons interfere with the kinetics of the Pd-hydrogen system. Ultimately, these types of surface limitations on the kinetics of sensing systems will

predominate for sensing layers that are prone to surface adsorption of background gases within the sensing environment. It is interesting to note that in Figures 4 and 5 the Pd0.94Ag0.06 alloy film has the smallest initial rate of hydrogen uptake and also the longest half-life for PdH formation. While granted the hydrogen solubility is highest for the PdAg alloy, this does not account for the differences observed in the initial uptake rate, which is dependent on the surface properties of the Pd alloy film, not on its corresponding hydrogen solubility. Likewise the difference between the two gold alloy films, with a 0.34 fractional increase in the alloy composition and a significant H2 solubility difference, shows only subtle differences in the initial hydrogen uptake rate and t1/2. As we have indicated, the quality and composition of the Pd alloy film surface is reflected in the kinetics of the initial hydrogen uptake rates. Therefore, from the kinetics analysis it appears that by changing the alloy composition from Ag to Au, while keeping the atomic concentrations the same, has a profound affect on the surface quality of the film. Experimentally each of the alloy films were treated in an identical fashion and there were no special measures to clean or modify the alloy film after deposition. Even though the films did not receive any special treatments, the films were all exposed to the same laboratory environment and therefore comparisons in the kinetics can still be made within the errors indicated in Figures 4 and 5. Such a lack of surface pretreatment (which mimics actual sensing conditions in the field) leads to the 30% error in the determination of the initial rate and t1/2 measurements from the experimental data. We propose that the bimetallic nature of these films may be the driving force behind the differences in the initial hydrogen uptake rates and the t1/2 values. In particular, it is well known that, by changing from a pure Pd film to a Pd alloy, the surface adsorption probability for common gases such as CO, water, and hydrocarbons is affected. This has been determined to result from shifts in the Fermi level from pure Pd and the Pd alloy. A weaker physisorbed interaction between the background gases and the alloyed film results, and thus, the surface composition of the adsorbed gases for a pure Pd and a Pd alloy film will not be the same.26,27 We intend to study the change in optical properties and the corresponding kinetics of the Pd alloy films as a function of controlled changes in the background CO, water, and hydrocarbon concentrations. However, our current studies appear to show that, by tuning the alloy composition and quantity, the film surface properties can be tuned to decrease the adsorbed gas quantity and composition, thus enhancing the sensor response properties under realistic environmental conditions. CONCLUSIONS We have demonstrated the phase-dependent hydrogen time response characteristics of Pd-Au (Ag) alloy films via optical reflectance measurements. The critical slowing down point is absent or shifts with a variation in the Pd alloying type and its corresponding content. The response times of the Pd0.94Ag0.06 and Pd0.94Au0.06 alloy films have a strong dependence on the hydrogen concentration, which is dominated by the long-time dynamics at [H2] concentrations at and near the R f β phase transition. The addition of a large amount of Au (40%) to Pd yields an alloy (26) Rodriguez, J. A. Surf. Sci. Rep. 1996, 24, 223. (27) Johanek, V.; Tsud, N.; Matolin, V.; Stara, I. Vacuum 2001, 63, 15.

Analytical Chemistry, Vol. 76, No. 21, November 1, 2004

6325

material that is above the critical isotherm threshold for undergoing a phase transition and overall has response times of less than 50 s for all of the investigated hydrogen concentrations. The kinetics of the three alloy systems have been qualitatively interpreted and due to uncharacterized surface impurities, largely found in a realistic sensing environment, each of the materials is limited by the first-order H2 dissociative chemisorption process at the alloy surface, followed by an increased n ) 1.6 dependence on the half-life of the reaction. We have shown that selection of the Pd alloy composition and content not only prevents Pd from undergoing the rate limiting R f β phase transition but appears to alter the surface impurity composition adsorbed on the sensing layer and thus leads to a more practical chemical sensor.

6326

Analytical Chemistry, Vol. 76, No. 21, November 1, 2004

ACKNOWLEDGMENT This work was supported by NYSTAR-TTIP and MTI-I (Contract 2000089). We gratefully acknowledge the support of the School of NanoSciences and NanoEngineering technical support staff for their assistance in the design and assembly of the sensor testing station. Jian Dai, Matthew Sleasman, and Peter Natale of MTI-I are acknowledged for their assistance in the development and optimization of the sensor testing station hardware and software.

Received for review April 2, 2004. Accepted August 17, 2004. AC0494883