A Study on the Aging of Ultra-Thin Palladium Films on SAW Hydrogen Gas Sensors B.H. Fisher and D.C. Malocha School of Electrical Engineering and Computer Science University of Central Florida, Orlando, FL. 32816-2450 [email protected], [email protected] Abstract—Traditionally, low-powered, room temperature sensing of gaseous hydrogen (H2) is difficult. With renewed interest in H2 as a source of energy, there is a need for reliable, energyefficient sensors. A potential solution can be found in using surface acoustic wave (SAW) devices, which have been implemented as passive, wireless RFID tag-sensors. Thus, in concept, it is advantageous to develop a SAW device with H2 sensing capabilities. Prior experiments have successfully demonstrated a passive SAW-based H2 gas sensor by placing an ultra-thin Palladium (Pd) film (<50Å) in the propagation path [1-3]. These sensors have an instantaneous response and a significant fractional change in SAW propagation loss; however, the lifetime of these sensors are still unknown. Hence, the objective of this study was to examine the influence of aging of ultra-thin Pd films on the usable life of passive SAW H2 gas sensors. The aging behavior of a thin film is highly dependent on its morphology and the environment in which it is used. Particularly, during growth, ultra-thin Pd films may form a discontinuous network of nano-sized atomic islands—nano-clusters. These nano-clusters may be amorphous or have a highly defective crystalline structure. When used in ambient air (1atm, 79% N2, 21% O2, 44% humidity, 21oC), these films are susceptible to oxygen adsorption (i.e., blocking H2 sorption and decreasing the H2 reactivity of the film), which reduces the useable life of the sensor. This paper presents data demonstrating that ultra-thin Pd films suffer from oxygen adsorption when exposed to ambient air. The results of the study provide promising solutions to the aging problem, such as encapsulation and film annealing. These solutions may accelerate the practical implementation of passive, wireless, SAW H2 gas sensors in various environments.

I. INTRODUCTION The long term goal of this work is to develop a room temperature, passive, wireless hydrogen gas sensor which can detect hydrogen gas in concentrations ranging from a few parts per million up to the lower explosive limit (LEL) of 4%. Furthermore, this sensor will be a node in a wireless sensing network, so it must be individually identifiable and thus have the ability to indicate the location and concentration of hydrogen gas in the event of a gas leak. Currently, there are a several methods by which the sensing of hydrogen gas at room temperature, and at concentrations below the lower explosive limit of 4%, may be achieved. These devices primarily utilize thin films of palladium (Pd), platinum (Pt) or alloys thereof, as a catalyst for hydrogen reactions which changes the electrical characteristics of the films thereby creating a measureable sensing event. The majority of devices which utilize these techniques are field effect (FET)[4, 5], resistive[6], optical[7] or acoustic wave devices [8]. The This work was supported through NASA-KSC STTR contract NNX09CB69C, and the National Institute of Aerospace contract C082638-UCF and the McKnight Doctoral Fellowship

FETs, resistive and acoustic wave devices have been implemented as a wireless, RFID tag sensor networks[5, 6, 9, 10]. The FETs were powered using a wired connection, solar or vibration based power sources with are not passive. A SAWbased hydrogen sensor which may be implemented passively, and wirelessly has been successfully demonstrated first by Yamanaka et al [11], then Fisher et al [1-3] and recently by Huang et al [10]. The device developed by Yamanaka showed good response times and sensitivity to trace amounts (10ppm) of H2 gas, but is difficult to fabricate due to use of a quartz ball as the piezoelectric substrate. The device developed by Huang et al is based on using a H2 sensitive resistor to modulate the electrical load on a SAW interdigitated transducer (IDT) consequently modulating the fraction of energy that is reflected by the IDT when the resistor is exposed to H2 gas. Huang’s device has a response time of approximately 15 minutes, which may be hazardous in situations where H2 gas leaks above the LEL are not detected in time to avoid an accident. The device developed by Fisher et al is based on SAW acoustoelectric effect and is implemented by placing an ultrathin film (UTF) of Pd in the SAW propagation path (Fig.1a). The sensor has an instantaneous response which consists of a large fractional change in SAW propagation characteristics when exposed to 2% H2 gas. This paper is an extension of the work done by Fisher et al by examining the effect of Pd film’s aging on the useable lifetime of the SAW H2 gas sensor. Palladium (Pd) UTFs were first used for room temperature, resistive H2 gas sensors [12-14]. These films have a nanoclustered morphology (i.e. a discontinuous network of metal islands (Fig. 1b)), and exhibit a very fast (<1second) and high fractional change in sheet resistance due to exposure to H2 gas (Fig. 2). Unfortunately, the ultra-thin Pd films suffer from oxygen adsorption when exposed to ambient air. Oxygen adsorption was observed in this study to cause a gradual increase in the film’s resistivity when exposed to O2 gas. A consequence of the O2 adsorption is the gradual desensitivity of the Pd film to H2 gas, ultimately shortening the sensor’s lifetime. To conduct the aging study, the sheet resistances of several ultra-thin Pd films were observed during film growth, then during exposure to different gases. The purpose of this study is to observe the aging behavior of ultra-thin Pd films in ambient air (1atm, 79% N2, 21% O2, 44% humidity, 21oC), and investigate methods to slow or stop this process, thereby extending the usable life of the SAW-H2 gas sensor.

The paper to be presented is structured as follows: section 2 covers the background of the work, section 3 details the experimental procedure, section 4 the experimental results and discussion of said results and finally section 5 concludes the report and comments on future research directions.

OFC-SAW H2 GAS SENSOR

2.5mm

II.

BACKGROUND

Film morphology and conduction mechanisms In order to understand the aging mechanisms of Pd UTFs, the film’s morphology and electrical conduction behavior must be understood. During early stage of nucleation, metal films are constituted of a discontinuous network of nano-sized atomic clusters called nanoclusters. Since the film is discontinuous, electrical conduction between islands is the primary achieved by quantum mechanical tunneling through the substrate[15]. The electrical conductivity is given by the Arrhenius equation 1 as: (1)

A

7.5 mm

where σo the intrinsic conductivity determined by the material and film geometry, T is the absolute temperature in Kelvin, kb is Boltzmann’s constant and δE is the activation energy required to transport an electron from one island to another and is given by the equation:

AFM Image of Nanoclustered Pd Film

(2)

B

= 250 nm

Figure 1. A: Orthogonal Frequency Coded (OFC) SAW RFID tagsensor with Pd film in the delay path for H2 gas sensing capability. B: Atomic force microscopy (AFM) image of the border between the Pd ultra-thin film and the YZ lithium niobate (YZ-LiNbO3) substrate. The AFM image shows that the film is constructed of a discontinuous network of atomic islands—nanoclusters. 1 0.9

-∆R/R0

Initial Response 10 Months later

2%H 2 Off

0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 − 0.1

0

30

2%H 2 On

60

90

120

150

180

210

240

270

300

Time (s)

Figure 2. 22% H2 gas exposure of an ultra-thin film Pd resistor. The abrupt resistance decrease when exposed to H2 gas is assumed due to a change in the cluster size and consequently the activation energy. After ten months, the fractional change in resistance due H2 gas exposure has decreased due to aging of the film.

where q is the electron charge, εr is the relative permittivity of the substrate, εo the permittivity of free space, r is the radius of the nano-clusters and s is the space between the nanoclusters. When exposed to H2 gas Pd clusters expand causing an increase in the Pd nanocluster’s radius, r, and a decrease in the separation, s, between the clusters. This results in a very fast and exponential decrease in the film’s resistivity consequently making the film an excellent sensor for H2 gas (Fig. 2). III.

EXPERIMENTATION

A. Substrate Preparation Prior to film growth the lithium niobate (LiNbO3) substrate was cleaned with an ultra-sonic bath in acetone, then rinsed in methanol, then de-ionized water, and finally plasma irradiated with O2 gas at 150W for a minute. B. Film Growth All films in this study were deposited using electron beam evaporation, at a growth rate of 0.01nm/s, in a bell jar which was evacuated to 10-6 Torr, and the substrate temperature controlled by the ambient (20-50oC). The film thickness was monitored using dual quartz microbalances, and the film resistance was monitored during growth (in-situ) with the use of a thin film resistor and a computer controlled multimeter. C. Resistive Pd H2 Gas Sensor Pd films were deposited on an interdigitated aluminum (Al) structure on a lithium niobate (LiNbO3) substrate. The structure was initially open-circuited and then became a resistor as an electrically-conductive Pd film was grown on top of it. A 2-3nm Pd film was grown atop the Al structure; when ex-

D. SAW Pd H2 Gas Sensor The OFC-SAW RFID tag-sensor was chosen as the platform for the hydrogen sensor due to its proven ability to make passive, wireless, RFID tag-senor networks [9]. SAW H2 gas sensor operation is governed by the acoustoelectric effect which predicts SAW propagation loss and velocity when an electrically conductive film is placed on the piezoelectric substrate. The SAW propagation loss coefficient per centimeter (cm) is given by equation 3 and is plotted as a function of film electrical resistivity in Fig.3. (3) where α, is the attenuation constant per unit distance given by:

the range where there is a strong acoustoelectric interaction between the film and the SAW (i.e. where there are measureable changes in propagation loss), then the SAW H2 gas sensor becomes insensitive to H2 gas. 0.1dB/cm was chosen as the boundary of SAW sensitivity to H2 gas as it is difficult measure any changes in propagation loss below this value. A curve fit of the measured data (Fig. 4) was performed which projected that the SAW propagation loss would decrease to 0.1dB/cm in approximately 34 hours. The acoustoelectric effect suggests this decrease in propagation loss from 75dB/cm to 0.1dB/cm is caused by an increase in the film’s resistivity (by approximately two orders of magnitude), which is believed attributable to the exposure the Pd UTF to ambient air. Stopping or slowing this gradual increase in film resistivity over time is important for extending the sensor’s lifetime. 100 90 80

Prop. Loss (dB/cm)

posed to 2% H2 gas, the film resistivity decreased by 60% in 1.5 seconds (Fig. 2). The sensor was stored in ambient air (1atm, 79% N2, 21% O2, 44% humidity, 21oC) for ten months. When exposed to hydrogen gas, the film resistivity decreased by 30% in 5 seconds when exposed to 2% H2 gas (Fig. 2). This decrease in the film’s sensitivity (i.e. fractional change and increase in the sensor’s response time) is due to the aging on the Pd film in ambient air. Equivalent films were grown on SAW devices and a similar aging behavior is expected in sensor’s response, but has yet to be investigated.

70 60 50 40 30 20 10

(4)

50MHz SAW on YZ-LiNbO3

0 1

where vsc is the SAW short circuit velocity, f is the SAW frequency, k2 is the piezoelectric electromechanical coupling coefficient, and fr is the dielectric relaxation frequency given by:

10

1 ×10

100

3

1×10

4

1×10

5

1×10

6

Resistivity (ohm-cm)

Figure 3. The SAW acoustoelectric (AE) theory predicts SAW propagation loss and velocity when a film is placed on the piezoelectric substrate.

(5)

75

where ρ is the electrical resistivity of film, εo is the permittivity of free space and εr is the relative permittivity of the substrate-film interface. The attenuation is maximized when the relaxation and SAW frequencies are equal. The relaxation frequency is dependent on the film’s resistivity and relative permittivity as shown is equation 5. The SAW device used in this study was a 1 port delay line with a center frequency of 50MHz, on a YZ-LiNbO3 substrate, with 1mm wide and 3-4nm thick Pd film in the SAW propagation path. A test fixture was created in order to obtain real time, in-situ measurements of the SAW propagation loss during film growth and/or during exposure to any gases. At the conclusion of the film deposition the SAW propagation loss was approximately, 75dB/cm, on the right-side of the peak attenuation, in the curve shown in Fig. 3. When exposed to ambient air (1atm, 79% N2, 21% O2, 44% humidity, 21oC), the propagation loss began an abrupt decrease which totaled approximately 5dB/cm in 20 minutes (Fig. 4). The sensing scheme employed for this SAW H2 gas sensor is based on the modulation of the SAW propagation loss, caused by the modulation of the Pd UTF’s resistivity when exposed to H2 gas. Thus, if the propagation loss continues to decrease beyond

73

Data Data Curve Fit

SAW Prop. Loss (dB/cm)

74

72 71 70 69 68 67 66 65 64

0

20

40

60

80

Time (min)

Figure 4. SAW propagation loss versus time is plotted for a sample film. When exposed to ambient air (1atm, 79% N2, 21% O2, 44% humidity, 21oC), the SAW propagation loss began an abrupt decrease which was observed for 20 minutes. A curve fit of the measured data was performed which projected that the SAW propagation loss as a function of time

E. O2 Adsorption The Pd films which are grown under the conditions previously discussed, (i.e. electron beam evaporation, at 10-6 Torr, and substrate temperature between 30 to 50oC) are likely to have an amorphous or highly defective crystalline structure

100

A2 Chamber Pressurized with N2 gas

A1 B1 0.7 0.6

∆R/R0

0.5

C

0.4

B2

Chamber Evacuated to 10-6 Torr

0.3 0.2

Chamber Open to Ambient Atmosphere 0.79 N2, 0.21 O2 44% Humidity

0.1 0 − 0.1

0

2.5

5

7.5

10

12.5

15

17.5

20

22.5

25

Time (min)

Figure 5. A1: Chamber Pressurized with N2 gas. B1: Upon initial exposure to ambient air, the resistance increased approximately 18% within 30 seconds. C: When the vacuum chamber was resealed and pumped down to 10-6 Torr, the reaction stopped. A2: Chamber re-pressurized to 1atm. B2: The chamber was re-opened and the resistance increased to 60% above its initial value.

F. Film Annealing Annealing (when done at the appropriate temperatures), heals the surface defects in the film [16], thus providing less nucleation sites for adsorption. Through a careful design

process, a high vacuum compatible, thermally controlled planetary, an in-situ resistance monitoring and an electrical feedthrough fixtures were constructed. Pd UTFs were grown in a 10-6 Torr environment at a rate of 0.01nm/s to a final thickness between 1 to 2nm. The films were then annealed at temperatures up to 290oC for two hours in a 10-4 Torr environment, while the resistance and temperature was monitored (Fig. 6). Annealing caused the film’s resistivity to decrease by an order of magnitude (Fig. 6); this suggests that a change in film morphology occurred. If the nanoclusters are merging due to added energy from the annealing process, then the activation energy changes because it is dependent on cluster size and separation distance is shown in equation 2. This change in activation energy causes an exponential change in resistance as observed in Fig. 6. R/Ro Temperature(°C)

0.1

1×10

100

0.01 1

10

100

3

Temperature oC

1

R/R0

[16]. In physical adsorption, O2 molecules stick to crystalline defect sites on the Pd film’s surface . The O2 molecules steal conduction carriers from the film which causes a dramatic increase in the electrical resistivity. A thin film resistor structure was used to measure the film’s resistivity during growth and the subsequent exposure of said film to various gases. The film was grown at a rate of 0.01nm/s to a final thickness between 1 to 2nm. A significant increase in electrical resistivity was observed when the Pd film was exposed to ambient air (Fig. 5). N2 was observed to have little effect on the film’s resistivity. The humidity effect was investigated by annealing the film in open air which should evaporate any condensed water molecules thus causing a reversible change in resistivity. The result of annealing the film in open air only served to increase the resistance more rapidly and changed the appearance of the film to a light white shade. The reaction appears to be accelerated by added energy (annealing), forms a whitish coat when annealed in air and causes increased film resistivity in O2 air. The result of this experiment suggests that oxygen adsorption may be the primary mechanism in the observed resistivity increase. Similar observations were reported by Yen [17] in 1974 for oxygen adsorption on zinc oxide thin films. There are three main techniques which may be employed to slow the adsorption process: (1) use of a Pd alloys, (2) annealing and (3) film encapsulation. Pd-Fe and Pd-Ni alloys were reported by Pitt et al [7] to be ineffective in slowing film aging, thus it was not considered in this study.

10 3 1×10

Time (min)

Figure 6. Observation of the resistivity of an ultra-thin Pd film during annealing process. The non-reversible change in resistivity suggests a change in film morphology occurred.

G. Film Encapsulation Film encapsulation is accomplished by coating the Pd film with a dielectric film which traps or repels O2 molecules but allows H2 molecules to diffuse through rapidly. Oxides are commonly used for thin film encapsulation coatings due to their ability to dramatically slow the diffusion of O2. Pitts et al [7, 18] reported that use the tungsten trioxide (WO3) as an encapsulation layer for Pd films extended the life of their fiber-optic H2 gas sensors by over a year. Aside from oxides, Lith et al. [13] reported using a photoresist layer to encapsulate and stabilize Pd UTFs with only a negligible effect of sensitivity of the film to H2 gas. Photoresist and other soft polymer-type films cannot be used on a Z-propagating SAW on YZ-LiNbO3 substrate as it severely damps the SAW. It is necessary to grow a dielectric encapsulation film atop the Pd film without breaking vacuum. Since the Pd films were grown using an electron beam evaporator, it was crucial to find an oxide film which may be evaporated with an electron beam. WO3 is not easily evaporated with an ebeam without dissociation into W and O2 gas and furthermore, was not available at the time of this experimentation. Aluminum oxide (Al2O3) and silicon monoxide (SiO) may be evaporated in an ebeam with minimal risk of the molecules dissociating [19]. Since SiO was not available at the time of this experimentation, ultra-thin Pd films were coated with 50nm and 150nm of Al2O3.

100

Ratio of Sample Aging Rates

O2 is Repelled or Trapped

H2 Diffuses Through Dielectric Film & Reacts with Pd

O2

Pd Nanoclusters

Dielectric Encapsulation Film

Ratio of Control to Annealed Samples Ratio of Control to Film Encapsulated with 50nm Al2O3 Ratio of Control to Film Encapsulated with 150nm Al2O3

10

1

Piezoelectric Substrate

IV. RESULTS AND DISCUSSION The annealed and the encapsulated samples were exposed to ambient air while the factional change in film resistivity (∆R/Ro) as a function of time was observed. A comparison of ∆R/Ro for the control sample (i.e. un-annealed or encapsulated), the annealed, and the encapsulated samples, is presented in Fig. 8. The results show that the ∆R/Ro per unit time for encapsulated films is dramatically lower than the unannealed and annealed samples. A comparison of the ratios the film’s aging rates per unit time (i.e. the ratio of ∆R/Ro for sample 1 to ∆R/Ro for sample 2) shows that when a sample is encapsulated with 150nm of Al2O3, ∆R/Ro is decreased by a factor of approximately 20 when compared to the control sample (Fig. 9). This data suggests that a combination of film annealing at 290oC and then encapsulating the film with 150nm of Al2O3 may be the most effective method of maximizing lifetime of the SAW ultra-thin Pd H2 gas sensor.

∆R/R0

0.5

5

6

7

8

V. CONCLUSIONS A study on the aging of SAW H2 gas sensors which use ultrathin Pd films for H2 gas sensing was performed. It is believed O2 molecules adhere to the surface of ultra-thin Pd films consequently increasing the film’s electrical resistivity and lowering its sensitivity to H2 gas. The effect limits the usable life of the sensor. In order to minimize the aging problem, films were annealed and encapsulated to determine which method was most effective. The Pd film’s aging rate was decreased by a factor of 2 when the film was annealed and a factor of 20 when it was encapsulated with 150nm of Al2O3. The experimental data suggests that encapsulating the Pd ultra-thin film with Al2O3 is best method to extend the life on the H2 gas sensor from this initial work. Future work will focus on exploring the effect of other dielectric films such as silicon monoxide (SiO) and tungsten trioxide (WO3) on the aging of ultra-thin Pd films; observing the effect of humidity on the response and aging of ultra-thin Pd films and developing a mathematical model which approximates the aging of ultra-thin Pd films in different environments. The authors are most grateful to Dr. Robert Youngquist, NASA-KSC for his continuing support discussions and suggestions. The authors acknowledge support through NASAKSC STTR contract NNX09CB69C, and the National Institute of Aerospace contract C08-2638-UCF and the McKnight Doctoral Fellowship Program.

0.45 0.4 0.35 0.3 0.25 0.2

REFERENCES

0.15 0.1

[1]

0.05 0

4

ACKNOWLEDGEMENTS

Control: Typical Response w/o Encapsulaton Film Annealed @ 290 for 2 hrs Film Encapsulated with 50nm Al2O3 Film Encapsulated with 150nm Al2O3

0.55

3

Figure 9. A comparison of the ratios the film’s aging rates per unit time (i.e. the ratio of ∆R/Ro for sample 1 to ∆R/Ro for sample 2) shows that when a sample is encapsulated with 150nm of Al2O3, ∆R/Ro is decreased by a factor of approximately 20 when compared to the control sample.

0.7 0.6

2

Time (min)

Figure 7. Schematic demonstrating the benefits of film encapsulation, by coating the ultra-thin Pd film with a dielectric film which traps or repels O2 molecule but allows H2 molecules to diffuse through rapidly, and react with palladium.

0.65

1

0

1

2

3

4

5

6

7

8

Time (min)

Figure 8. The control sample (i.e. un-annealed or encapsulated), annealed and the encapsulated samples were exposed to ambient air while the factional change in film resistivity (∆R/Ro) as a function of time was observed

[2] [3] [4] [5]

B. Fisher and D. C. Malocha, "Room Temperature Hydrogen Gas Sensing Using SAW Devices," in Proceedings of the Joint Japan-USA International Meeting on Acoustic Wave Devices. Orlando, FL, 2008, pp. 125-131. D. C. Malocha and B. H. Fisher, "Acoustoelectric Effect Study for SAW Sensors," presented at IEEE International Frequency Control Symposium, Besançon, France, 2009, pp. 254-259. B. H. Fisher and D. C. Malocha, "Study of the Acoustoelectric Effect for SAW Sensors," IEEE Transactions on Ultrasonics, Ferroelectrics and Frequency Control, vol. 57, no. 3, pp. 698-706, 2010. I. Lundström, S. Shivaraman, C. Svensson, and L. Lundkvist, "A hydrogen−sensitive MOS field−effect transistor," Applied Physics Letters, vol. 26, no. 2, pp. 55-57, 1974. X. Yu, C. Li, Z. N. Low, J. Lin, T. J. Anderson, H. T. Wang, F. Ren, Y. L. Wang, C. Y. Chang, S. J. Pearton, C. H. Hsu, A. Osinsky, A. Da-

biran, P. Chow, C. Balaban, and J. Painter, "Wireless hydrogen sensor network using AlGaN/GaN high electron mobility transistor differential diode sensors," Sensors and Actuators B: Chemical, vol. 135, no. 1, pp. 188-194, 2008. [6] J. Jun, B. Chou, J. Lin, A. Phipps, X. Shengwen, K. Ngo, D. Johnson, A. Kasyap, T. Nishida, H. T. Wang, B. S. Kang, F. Ren, L. C. Tien, P. W. Sadik, D. P. Norton, L. F. Voss, and S. J. Pearton, "A hydrogen leakage detection system using self-powered wireless hydrogen sensor nodes," Solid-State Electronics, vol. 51, no. 7, pp. 1018-1022, 2007. [7] R. Pitts, P. Liu, S.-H. Lee, E. Tracy, R. D. Smith, and C. Salter, "Interfacial Stability of Thin Film Hydrogen Sensors," in Proceedings of the DOE Hydrogen Program Review, 2001. [8] A. D'Amico, A. Palma, and E. Verona, "Palladium-surface acoustic wave interaction for hydrogen detection," Applied Physics Letters, vol. 41, no. 3, pp. 300-301, 1982. [9] N. Y. Kozlovski and D. C. Malocha, "SAW passive wireless multi sensor system," presented at Ultrasonics Symposium (IUS), 2009 IEEE International, 2009, pp. 1541-1544. [10] Y.-S. Huang, Y.-Y. Chen, and T.-T. Wu, "A passive wireless hydrogen surface acoustic wave sensor based on Pt-coated ZnO nanorods," Nanotechnology, vol. 21, no. 9, 2010. [11] K. Yamanaka, S. Ishikawa, N. Nakaso, N. Takeda, T. Mihara, and Y. Tsukahara, "Ball SAW Device For Hydrogen Gas Sensor," in IEEE Ultrasonics Symposium, 2003, pp. 299-302.

[12] T. Xu, M. P. Zach, Z. L. Xiao, D. Rosenmann, U. Welp, W. K. Kwok, and G. W. Crabtree, "Self-assembled monolayer-enhanced hydrogen sensing with ultrathin palladium films," Applied Physics Letters, vol. 86, no. 20, pp. 203104, 2005. [13] J. v. Lith, A. Lassesson, S. A. Brown, M. Schulze, J. G. Partridge, and A. Ayesh, "Hydrogen Sensor Based on Tunneling Between Palladium Clusters," Applied Physics Letters, vol. 91, 2007. [14] O. Dankert and A. Pundt, "Hydrogen-induced percolation in discontinuous films," Applied Physics Letters, vol. 81, no. 9, pp. 1618-1920, 2002. [15] R. M. Hill, "Electrical Conduction in Discontinous Metal Films," Contemporary Physics, vol. 10, no. 3, pp. 221-240, 1969. [16] M. Ohring, The Material Science on Thin Films. Boston: Academic Press Inc., 1992. [17] J. C. Yen, "An Investigation of the Electrical properties of Zinc Oxide Thin Films Influenced by Oxygen Adsorption," Journal of Vacuum Science, vol. 10, no. 1, pp. 47-51, 1974. [18] R. D. Smith, P. Liu, S.-H. Lee, E. Tracy, and R. Pitts, "Interfacial Stability of Thin Film Fiber-Optic Hydrogen Sensors," in Proceedings of the 2002 U.S. DOE Hydrogen Review., 2002. [19] R. W. Berry, P. M. Hall, and M. T. Harris, Thin Film Technology. New York: Van Nostrand Reinhold Company, 1968.