h2s sensing properties of wo3 based gas sensor

Gardner* School of Engineering, University of Warwick, Coventry, UK Abstract Semiconductor gas sensors based upon n-type WO3 nanoparticle powder were fabricated and their response fro

Procedia Engineering 168 ( 2016 ) 255 – 258

1877-7058 © 2016 Published by Elsevier Ltd This is an open access article under the CC BY-NC-ND license

( http://creativecommons.org/licenses/by-nc-nd/4.0/ ).

Peer-review under responsibility of the organizing committee of the 30th Eurosensors Conference

doi: 10.1016/j.proeng.2016.11.181

ScienceDirect

30th Eurosensors Conference, EUROSENSORS 2016

B Urasinska-Wojcik, T.A Vincent, J.W Gardner*

School of Engineering, University of Warwick, Coventry, UK

Abstract

Semiconductor gas sensors based upon n-type WO3 nanoparticle powder were fabricated and their response from 100 ppb to ppm levels of H 2 S in air has been investigated In this study, a low power MEMS based micro-hotplate was employed as the substrate and operated at a constant temperature of 350qC Experimental data demonstrate repeatable results across the range of concentrations The devices have good response and recovery times and importantly excellent stability Cross-sensitivities towards other reducing gases/VOCs showed that the WO 3 -based sensors are less sensitive to CO, acetone, H 2 and CH 4 even at much higher ppm levels

© 2016 The Authors Published by Elsevier Ltd

Peer-review under responsibility of the organizing committee of the 30th Eurosensors Conference

Keywords: gas sensor, H2 S detection; metal oxide; sensing layer

1 Introduction

H2S sensors are generally designed to save the lives of people in the workplace by warning them of excessive H2S concentrations in ambient air According to the UK Health and Safety Executive, the maximum peak exposure is 10 ppm for 15 min [1] Therefore, there is increasing demand for sensing devices that monitor low ppm H2S concentrations Nanostructured metal oxide semiconductors are widely used for the fabrication of sensors for the detection of both oxidising (such as NO2)and reducing gases (such as H2S) Among metal oxide semiconductors, specifically those based on WO3 are of particular interest due to their structural simplicity, high sensitivity, low cost and potential durability

In this work, we have studied experimentally the response of WO3-based CMOS gas sensors to sub ppm levels of

H2S in air and we have also investigated the cross-sensitivity to other gases at much higher levels Previously work has been reported on sensors based upon WO3 nanostructures, however by considering the advantages of thick film technology, we present a more sensitive sensor that has excellent properties comparable to those based on thin or

* Corresponding author Tel.: +44 24765 23695; fax: +0-000-000-0000

E-mail address: J.W.Gardner@warwick.ac

© 2016 Published by Elsevier Ltd This is an open access article under the CC BY-NC-ND license

( http://creativecommons.org/licenses/by-nc-nd/4.0/ ).

Peer-review under responsibility of the organizing committee of the 30th Eurosensors Conference

functionalised WO3 films [2]

In this study, a low power MEMS based micro-hotplate gas sensor was used, and the operating temperature was controlled by an adjustable constant current circuit The micro-hotplate is shown in Fig 1a (CCS09C, Cambridge CMOS Sensors Ltd) The MEMS structure was fabricated in a commercial foundry and is based upon silicon-on-insulator (SOI) technology In the membrane structure, a tungsten resistive micro-heater is embedded within a 5 Pm thick metal/oxide stack ensuring a low DC power consumption (e.g 65 mW at 600qC) The membrane is fabricated via a post CMOS deep reactive ion etch (DRIE) and both mechanically supports and thermally isolates the heater from the sidewalls The MEMS micro-hotplate can reach temperatures well in excess of 500qC and has a sub-5V controlled temperature ramp capability of 30 ms heating time and 60 ms cooling time from ambient to 500qC

WO3 powder (New Metals and Chemicals Ltd.) was mixed with an organic dispersant ESL 400 to obtain a paste The weight ratio of the powder and the organic dispersant was 1:2 The paste was drop caste onto the 1 mm × 1 mm silicon die, which comprised gold interdigitated electrodes on top of the membrane as a single-chip solution (Figs 1a and 1b) After deposition of the WO3 paste the substrate was left to dry in air at room temperature for ~12 h followed

by annealing at 450qC for 1 h, and then at 350qC for about 23 h under ambient air to obtain the sensor element consisting of n-type WO3 Finally, the silicon die was wire-bonded onto a standard TO-46 header (Fig 1c) Fig 1d presents a scanning electron microscopy (SEM) image of the annealed paste with a porous microstructure showing a widespread grain size distribution with average particle size in hundreds of nm

Fig 1 (a) Optical micrograph of a bare micro-hotplate; (b) a typical device after deposition of the WO 3 paste; (c) a gold wire-bonded micro-hotplate on a TO46 package and (d) SEM (250× magnification) of the surface of the deposited WO3 material

The gas sensing measurements were performed at the Microsensors & Bioelectronics Laboratory at the University

of Warwick using fully-automated custom rig (Figs 2a and 2b) The CMOS micro-hotplate substrates mounted on TO46 packages were connected to a custom made printed circuit board Both the micro-heater and chemoresistor were driven/measured using National Instruments DAQ hardware and software The gas sensing properties of the sensor element were characterised using a flow type sensing measurements apparatus The gas sensor was placed inside an aluminium sample chamber equipped with standard SwagelockTM gas inlet and outlet connectors Synthetic air was introduced into the sample chamber for 5 min and then a gas mixture of H2S in air was injected for 5 min in steps of varying concentrations of 5, 4, 3, 2, 1, 0.5, 0.25 and 0.1 ppm Precise gas mixtures in the ppb to ppm range were generated using three digital mass flow controllers The total gas flow rate was 0.5 slpm and the measurements were performed at room temperature in dry conditions and then at 25% relative humidity (RH) A LabView (National Instruments) interface allowed fully automated control of the digital mass flow controllers of the gas testing system

Fig 2 (a) Block diagram of the gas sensor setup; (b) Photograph of the fully-automated custom designed gas testing rig.

2 Results and Discussion

The time-dependent resistance changes to H2S pulses at concentrations ranging from 1 to 5 ppm and from 100 ppb

to 1 ppm in dry and humid air are shown in Figs 3a and 3b, respectively When H2S gas was introduced, the resistance

of the WO3 sensor element decreased with increase in the concentration of H2S This is a typical response of an n-type

oxide towards a reducing gas, leading to Rair > RH2S Initial response of sensors in 25% RH air was significantly

suppressed compared to the results obtained in dry air and this requires further analysis in the future

Fig 3 Dynamic response of WO 3 -based sensors in presence of H 2 S (a) 1-5ppm and (b) 0.1-1 ppm in dry and 25% RH air at 350qC

The response (S) of semiconductor metal oxide gas sensor is empirically represented by the following power law:

where Ag is a pre-factor that depends on the type of the sensing material, the operating temperature, and the type of

gas interacting with the sensor C g is the gas concentration and E is the exponent factor, and its ideal value of 0.5 or

1 depends on the charge state of surface oxygen species and the stoichiometry of the elementary reactions on the surface [3] According to the above power law equation, the value of E from the experimentally measured response

versus concentration plot (Fig 4a) was 0.88 This value of E exponent is close to theoretical value of 1 suggesting that

the chemisorbed surface oxygen species are nearly all in the O state It is also well known that reducing gases react with oxygen to produce water molecules and at typical sensor operating temperatures above 300qC, the dominant species on oxide surface is O Upon interaction of adsorbing H2S gas molecules with surface oxygen ions the following reaction is likely to take place:

This reaction donates an electron to the WO3 conduction band thereby increasing the conductivity of the tungsten oxide The sensor response and recovery time remained almost unchanged when exposed to H2S in dry air and ranged from 16-43 s and 43-48 s, respectively (Fig 4b) The response time did not significantly change when the concentration was lowered to 100 ppb

Fig 4 (a) Response of WO 3 based sensor as a function of concentration plot; (b) Response and recovery time vs concentration of H2 S in dry air

Average response (Rair/RH2S) values with standard deviations (STD) of WO3-based sensors to H2S in dry air are presented in Table 1 Thesesensors are known to be sensitive towards reducing gases The sensors were tested for cross-sensitivity in the presence of carbon monoxide, acetone, methane and hydrogen at various concentration ranges

in dry air Table 2 summarises the typical response of the sensing element to these gases The values show that the

WO3 material is not so sensitive to CO, acetone, CH4 and H2

Table 1 Average response values with standard deviations from 4 sensors Table 2 Response of WO 3 based sensors to various reducing gases

3 Conclusions

In this study we report upon the response of a tungsten oxide based MEMS gas sensor to sub ppm levels of H2S

The response was found to be excellent of ca 4.0 to a 5 ppm pulse of H2S in dry air at 350°C It was found that these

sensors with the active layer prepared by drop casting and annealing work reliably and showed superior stability throughout the gas concentration range studied in dry air We also showed that there was a significant response change

in humidity but a negligible cross-sensitivity to CO, H2, CH4 and acetone at much higher ppm levels In conclusion,

we believe that this MEMS based semiconducting gas sensor could find application in, for example, personal protective equipment in potential risk areas across oil and gas production

Acknowledgements

This research was funded by the EPSRC project: EP/L018330/1 The authors thank CCS Ltd for their support

References

[1] M.G Costigan, Hydrogen sulfide: UK occupational exposure limits, Occup Environ Med 60 (2003) 308–312

[2] V Kruefu, A Wisitsoraat, A Tuantranont, S Phanichphant, Ultra-sensitive H2S sensors based on hydrothermal/impregnation-made Ru-functionalized WO 3 nanorods, Sensor Actuat B-Chem 215 (2015) 630-636

Conc [ppm] 0.1 0.25 0.5 1 2 3 4 5

Response 1.4 1.3 1.3

1.

7 2.5 3.0 3.5 3.9

STD 0.3 0.2 0.1

0.

3 0.6 0.6 0.7 0.8

Gas CH4 H2

C

O C3H6O H2S Conc [ppm] 15000 20000 50 300 5

Response 1.0 1.0 1.6 2.3 3.9