Procedia Engineering 5 (2010) 143–146
Procedia www.elsevier.com/locate/procedia Engineering
Procedia Engineering 00 (2010) 1–4
Low Power Hydrogen Sensors Using Electrodeposited PdNi-Si Schottky Diodes. A. R. Usgaocara , C.H. de Groota , Cedric Boulartb , Alain Castillob , Val´erie Chavagnacb a School
of Electronics and Computer Science, University of Southampton, Highfield, Southampton SO17 1BJ, U.K b LMTG UMR 5563, CNRS, Universit´ e de Toulouse, Toulouse, France
Abstract The use of electrodeposited PdNi-Si Schottky barriers as low power Hydrogen sensors is investigated. The Palladium content of the film causes the Hydrogen molecules to dissociate and be absorbed by the film, changing the metal work function and Schottky barrier current. Electrodeposited PdNi-Si Schottky barriers exhibit very low reverse bias current and in a back to back configuration form a device that draws extremely low power when idle. The Schottky diodes were fabricated on 0.5-1.5 Ω.cm 100 n-type Si by electrodeposition of PdNi followed by evaporation of Aluminium contact pads. Electrical measurements at different Hydrogen pressures were performed on back to back Schottky diodes in a vacuum chamber using pure Nitrogen and a 5% Hydrogen-Nitrogen mixture. Very low currents of ≈ 1nA were measured in the absence of Hydrogen. Large increases in the currents, upto a factor of 100, were observed upon exposure to different Hydrogen partial pressures. The highest sensitivity was estimated to be 17.27 nA/mbar. The low idle current, simplicity of the fabrication process and ability to easily integrate with conventional electronics proves the suitability of electrodeposited PdNi-Si Schottky barriers as low power Hydrogen sensors. c 2010 Published by Elsevier Ltd.
Keywords:
1. Introduction Hydrogen sensors are crucial to enhancing safety in Hydrogen production, storage, transport and use. Leak detection is of important in any transport and storage process of large quantities of Hydrogen, especially as Hydrogen tends to embrittle it’s storage containers and can lead to potentially explosive leaks. Hydrogen is also actively being considered as an energy source of the future, with fuel cells being investigated for use in electric vehicles and possibly houses. Widespread acceptance of Hydrogen for such personal applications is dependent on adequate safety features, of which leak detection is of paramount importance. Other applications for Hydrogen sensors are in furthering scientific research. A large area of research is focussed on the study of deep sea Hydrothermal vents and one problem in such studies is the lack of chemical sensors for continuous monitoring of venting fluid chemistry, especially for gases like Hydrogen and Hydrogen sulphide [1]. In situ sensors for other dissolved gases like Methane have been demonstrated [2] but the lack of a robust, low power, high sensitivity Hydrogen sensor capable of functioning in adverse deep sea conditions limits the monitoring capabilities of such research projects. Solid state gas sensors are of interest due to their simplicity and their compatibility with allied electronic systems. Most solid state Hydrogen sensors generally measure the resistance change in Pd/Pd-alloy [3] films or the current changes in Pt/ZnO [4] or Pt/GaN Schottky barriers [5] on exposure to Hydrogen. A survey shows that most of these devices draw currents in the microampere to milliampere range, and are therefore not very efficient in terms of power consumption. In this paper we present an extremely low power, highly selective Hydrogen sensor based on electrodeposited PdNi-Si Schottky barriers fabricated using standard and inexpensive semiconductor processes.
c 2010 Published by Elsevier Ltd. 1877-7058 doi:10.1016/j.proeng.2010.09.068
144
A.R. Usgaocar et al. / Procedia Engineering 5 (2010) 143–146
A.R. Usgaocar et al. / Procedia Engineering 00 (2010) 1–4
2
2. Fabrication The critical step in the fabrication of this sensor was the use of electrodeposition as a metallisation technique. Electrodeposition is a process in which metal is deposited onto a substrate by passing an electrical current through a solution of the metal salt (electrochemical bath) with the substrate acting as the cathode. The applied electric field causes positively charged metal ions to migrate to the cathode, where they get reduced and are deposited on the cathode’s surface. Alloy depositions are accomplished by using an electrochemical bath of two metal salts, whose concentrations are adjusted so as to obtain a potential range where both metals will get deposited [6]. The electrodeposition of PdNi alloys was done by using the recipe for an electrochemical bath from Ref. [7]. Three baths (Ni-only, Pd-only and PdNi) were prepared as shown in Table. 1. Table 1: Concentrations of components in the three solutions prepared for each combination of Ni and Pd salts.
Components NiSO4 .6H2 O Pd(en)Cl2 (NH4 )2 SO4 NH3 (35%)
Ni bath(g/L) 44.40 0 16.70 45ml/L
Pd bath(g/L) 0 11.1 16.70 45ml/L
PdNi bath(g/L) 44.40 11.1 16.70 45ml/L
Cyclic voltammetry (CV) curves were measured for each of the three baths by cycling the deposition potential from 0V to -3V and back to study the dependence of the current in each bath as a function of the deposition potential. A high potential pretreatment pulse was applied with the Silicon chip as a cathode and the PdNi alloy was electrodeposited using a potential such that the deposition current density was in the 3 − 5mA/cm2 range. The pretreatment pulse promotes nucleation on the Silicon surface and improves film adhesion by providing a large number of anchor points on the Silicon. The subsequent deposition at lower potentials tends not to form islands but grow the film by depositing metal onto previously existing islands. The current density of 3 − 5mA/cm2 was chosen because densities outside this range adversely affected the film quality, yielding discontinuous and brittle films. A mechanism to control the composition of the alloy was devised by assuming that the individual chemical reactions involved in the PdNi bath are independent. The nominal value for the film composition can be calculated using the formula INi (Vd ) (1) INom = INi (Vd ) + IPd (Vd ) Where INi (Vd ) and IPd (Vd ) are the currents measured in the Ni-only and Pd-only baths at a deposition potential Vd . The nominal value was compared to the film composition as measured using energy dispersive X-Ray (EDX) spectroscopy. A plot showing the results of this comparison between nominal and EDX measured film compositions for different deposition potentials using the same PdNi bath is shown in Fig. 1(a). Good agreement between the nominal and experimental values of the atomic Ni fraction is observed. The accuracy of the prediction is higher for potentials more negative than -1.2V because both metals start to deposit in this potential range. The plot illustrates that a wide range of Ni concentrations can be obtained from a single electrochemical bath simply by varying the deposition potentials. The sensors were fabricated on a 2-inch, 0.5-1.5 Ω.cm, 100 n-type Silicon wafer. The wafer was cleaned by a 2 minute dip in fuming Nitric acid (FNA) followed by rinsing with deionised (DI) water. A 50nm thermal oxide layer was grown to act as the insulating layer for the second metal layer in the device. The oxide layer was patterned using photolithography and the exposed oxide was etched by a 20:1 Hydrofluoric acid (HF) etch. The entire wafer was immersed in the PdNi electrochemical bath and the PdNi alloy was electrodeposited on the exposed Silicon surface to form the Schottky diodes. A second photolithography step was used to pattern the substrate for contact pads and wires, which were formed by evaporating a 100nm Aluminium layer. An image of the finished device as seen through an optical microscope is shown in Fig. 1(b). The PdNi-Si Schottky diode and Aluminium wires and contact pads are marked in Fig. 1(b). Any two such Schottky diodes in Fig. 1(b) can be used in a back to back configuration for sensing Hydrogen. In this study, pairs of PdNi films opposite each other were chosen as a sensor device. These pairs were fabricated to have the same area, but with varying separations and overlaps with the Aluminium wiring, thus leaving different areas of the film exposed to gaseous Hydrogen. Electrodeposited PdNi alloys were found to form
145
A.R. Usgaocar et al. / Procedia Engineering 5 (2010) 143–146
3
A.R. Usgaocar et al. / Procedia Engineering 00 (2010) 1–4
0.8
Ni atomic fraction
0.7
Experimental Nominal
0.6 0.5 0.4 0.3 0.2 0.1 0 -0.8
-1
-1.2 -1.4 -1.6 -1.8
-2
Potential (V) (a)
(b)
Figure 1: (a)Nominal and EDX measured Ni atomic fractions in films deposited at different potentials from a 44.4g/L NiSO4 .6H2 O and 11.1 g/L Pd(en)Cl2 bath. (b)Structure of the chip with PdNi-Si Schottky barriers. The Schottky barrier is marked with a red circle.
Schottky barriers with extremely low reverse bias currents [8] and the back to back configuration during measurement ensures that only the low reverse bias current will be flowing through the device. 3. Results and Discussions To measure the response to gaseous Hydrogen, the chip was placed inside the vacuum chamber of a Lakeshore EMTTP4 probe station and the chamber was evacuated to a pressure of ≈ 1.5 × 10−2 mbar. Nitrogen was introduced till a base pressure of 0.3 Bar and the sensor was allowed to settle for 5 mins. A mixture of 5% Hydrogen in Nitrogen was then introduced in steps of 0.1 bar till the chamber pressure was a total of 1 bar. This is equivalent to introducing 5 mbar gaseous Hydrogen as the Hydrogen-Nitrogen mixture has 5% Hydrogen by volume. The back to back Schottky barrier response was measured by applying a bias of 1V and measuring the change in current with time at room temperature using an Agilent 4155C semiconductor parameter analyser. The sensor current increased rapidly on introduction of gaseous Hydrogen and took a long time (≈ 40min) to saturate. All the devices on the chip were measured at each pressure and the devices were allowed to rest 40 minutes before measurement so that they achieve their steady state currents. Fig. 2(a) shows the measured currents in the sensor device as a function of the exposed length of the PdNi film and for different pressures. It can be seen that the back to back Schottky barrier currents are very low in the absence of gaseous Hydrogen and were measured to be between 1-2nA and shows that the sensor draws ultra low power in the absence of Hydrogen. Exposure of the sensing element to 10 mBar of Hydrogen causes a large increase in the measured current of the device, with some samples showing increases greater by a factor of 100. Subsequent increases in Hydrogen pressure cause further increases in the measured current, but the current saturates at higher Hydrogen pressures as shown in Fig. 2(b). It is seen however, that the current always remains lower than 1μA and the sensor therefore draws low power even in the presence of Hydrogen, this low power consumption over the entire range of measured Hydrogen makes the sensor useful for systems with power constraints. The sensitivity of the sensor can be estimated by the slope of the initial linear part of the graph. For the three devices shown in Fig. 2(b), the highest sensitivity is found to be 17.27 nA/mbar and the lowest was 6.75 nA/mbar. It is intuitive to expect that the change in current in the device will be positively correlated with the exposed length of the sensing element, but Fig. 2(a) shows an inverse relationship between the two. This can be explained by the fact that the devices are allowed to rest for 40 minutes so that they achieve their steady state currents. Using the value
146
A.R. Usgaocar et al. / Procedia Engineering 5 (2010) 143–146
4
A.R. Usgaocar et al. / Procedia Engineering 00 (2010) 1–4
300
350 0 bar 10 mbar 20 mbar
L = 50 μm L = 45 μm L = 10 μm
300 250
200
Current (nA)
Current (nA)
250
150 100
200 150 100
50
50
0
0 0
10
20
30
40
Exposed Length(μm) (a)
50
60
0
5
10 15 20 H2 pressure (mbar)
25
(b)
Figure 2: (a) Dependence of device current on the exposed length of the Pd0.71 Ni0.29 sensing element. (b) Dependence of the current measured in three different devices with Hydrogen gas pressures. The variable “L” is the exposed length of the sensing element.
for the Hydrogen diffusion constant as measured by Flanagan and Oates [9], it can be seen that this time is sufficient for Hydrogen to diffuse through the entire film. The effect of the different exposed lengths is therefore negated and the different responses are probably due to the individual microstructures of the devices. This is an avenue for further investigation. 4. Conclusions The fabrication and characterisation of a Hydrogen sensor based on back to back Pd0.71 Ni0.29 electrodeposited PdNi-Si Schottky diodes is presented. The sensor is small in size, has an extremely low current when idle and exhibits a large change in current values on exposure to gaseous Hydrogen. Despite this large change, the current was observed to be less than 500nA over the measured Hydrogen pressure range. The change in the electrical characteristics is due to dissociation of Hydrogen molecules on Pd atoms on the surface and their absorption into the PdNi film, making the sensor highly selective. The use of electrodeposition in the fabrication of the Schottky diodes considerably reduces the complexity and cost of the PdNi alloy fabrication as compared to standard metallisation processes like evaporation and sputtering. Fabrication of the sensor on a Silicon substrate also makes it simpler to integrate the sensor with allied electronics. These characteristics make the Hydrogen sensor very useful for measuring low values of Hydrogen pressure and in systems where power consumption of the sensor is a primary concern. [1] K. Ding, W. E. Seyfried, M. K. Tivey, A. M. Bradley, In situ measurement of dissolved H2 and H2S in high-temperature hydrothermal vent fluids at the main endeavour field, Juan de Fuca ridge, Earth and Planetary Science Letters 186 (3-4) (2001) 417–425. [2] C. Boulart, M. C. Mowlem, D. P. Connelly, J.-P. Dutasta, C. R. German, A novel, low-cost, high performance dissolved Methane sensor for aqueous environments, Opt. Express 16 (17) (2008) 12607–12617. [3] Y.-T. Cheng, Y. Li, D. Lisi, W. M. Wang, Preparation and characterization of Pd/Ni thin films for Hydrogen sensing, Sensors and Actuators B: Chemical 30 (1) (1996) 11–16. [4] S. Kim, B. S. Kang, F. Ren, K. Ip, Y. W. Heo, D. P. Norton, S. J. Pearton, Sensitivity of Pt/ZnO schottky diode characteristics to hydrogen, Applied Physics Letters 84 (10) (2004) 1698–1700. [5] J. Schalwig, G. M¨uller, U. Karrer, M. Eickhoff, O. Ambacher, M. Stutzmann, L. G¨orgens, G. Dollinger, Hydrogen response mechanism of Pt–GaN schottky diodes, Applied Physics Letters 80 (7) (2002) 1222–1224. [6] M. Schlesinger, M. Paunovic, Modern Electroplating, 4th Edition, John Wiley and Sons, 2000. [7] S.-E. Nam, K.-H. Lee, A study on the Palladium/Nickel composite membrane by vacuum electrodeposition, Journal of Membrane Science 170 (1) (2000) 91–99. [8] A. R. Usgaocar, C. H. de Groot, Electrodeposited PdNi as possible ferromagnetic contacts for carbon nanotubes, Physica Status Solidi (b) 247 (4) (2010) 888–891. [9] T. B. Flanagan, W. A. Oates, The Palladium-Hydrogen system, Annual Review of Materials Science 21 (1) (1991) 269–304.