solid state electrical conductivity and humidity sensing properties of wo3–y2o3 composites

phys. stat. sol. (a) 201, No. 3, 529–535 (2004) / DOI 10.1002/pssa.200306754

© 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Solid state electrical conductivity and humidity sensing properties of WO3–Y2O3 composites

R. Sundaram and K. S. Nagaraja*

Department of Chemistry, Loyola Institute of Frontier Energy (LIFE), Loyola College, Chennai 600 034. India

Received 10 May 2003, revised 27 September 2003, accepted 1 December 2003 Published online 22 January 2004

PACS 61.10.Nz, 72.80.Tm, 81.05.Mn, 82.47.Rs

Experimental results on the composites made from WO3 (WO) and Y2O3 (YO) for electrical and humidity sensing properties are described. The compound and composites of WO3 and Y2O3 in mole ratios (1 :1, 1 :2, 2 :1, 1 :3, 3 :1 and 3 :2%) were sintered in the form of a disc of 10 mm diameter were subjected to dc conductance measurements over the temperature range 373–673 K from which the activation energies were determined. The composites were identified by powder XRD data. The scanning electron micros-copy (SEM) studies were carried out to study the surface and pores structure of the sensor materials. The Brunauer–Emmett–Teller (BET) surface adsorption studies showed that the radius of the pore sizes were found to be distributed from 10–45 Å. The pore specific volume was calculated to be 0.01 cm3 gm–1. As the composites having micropores are preferred for humidity sensing properties, the composites were sub-jected to dc resistance measurements as a function of relative humidity in the range of 5–98% RH, achieved by different water vapour buffers thermostated at room temperature. The sensitivity factor S

f

(R5%/R98%) measured at 25 °C revealed that WOYO-31 (WO3 and Y2O3 in 3 :1mole ratio) has the highest humidity sensitivity factor of 1535 (±80). The response and recovery time for this humidity sensing com-posite was also studied.


1 Introduction

Humidity is a significant environment factor, which has to be monitored and controlled for a comfortable living [1–3]. Ceramics have some attractive properties like thermal stability, water resistance, corrosion resistance and microstructure, which renders the materials to be highly promising [4] as humidity sen-sors. WO3 and Y2O3 are known to be potential materials from both the fundamental and technical point of view [5–7]. Gas sensing devices are one of the major applications of the ion conducting materials [8]. WO3, Y2O3 and SnO2 have been reported [9] for humidity sensing metal oxides and ceramics. Several sensors based on the ion conducting materials are reported such as [10, 11] for a variety of gases SO2, SO3, NO, NO2, CO, CO2 [12, 13], O2 [14] and H2O (humidity) [15, 16]. The humidity sensing devices can be classified [15, 16] into two types according to their working principle is impedance type in which resistance or conductance or capacitance of the sensors changes and electrochemical cell type in which potential or current changes due to sorption of water molecules. Several hydrates have been used [17] to develop the impedance type humidity sensors such as oxides of antimony operative at temperature 40–100 °C. Iwahara et. al. [18] reported an electrochemical or galvanic cell type humidity sensor using high temperature proton conducting ceramic sintered oxide SrCeO3 – doped with yttrium or ytterbium oxide working at 400 °C and Miyazaki et. al. [19] used electrolytic MnO2 and its derivatives as solid electrolytes. We hereby report the electrical and humidity sensing properties of WO3–Y2O3 composites. * Corresponding author: e-mail: [email protected], [email protected], Tel.: +91-44-28175660,

Fax: +91-44-28175566

530 R. Sundaram and K. S. Nagaraja: Solid state electrical conductivity an humidity sensing properties


2 Experimental

The different mole ratios of WO3 (Aldrich, >99%) and Y2O3 (Koch-Light, 99.9%) were mixed together for the fabrication of WO3–Y2O3 composite. The mixture was milled for 12 h for homogeneity in a vi-bromill and was subsequently ground under absolute ethanol for 2 h in an agate mortar. After drying, the mixtures were compacted into cylindrical disks of about 10 mm diameter and 5 mm thickness in a hy-draulic press at a pressure of 100 Mpa. These pellets were then heated in a high purity alumina support in the uniform temperature zone of a tubular furnace in ambient air. Different heating rates were employed for better sintering and tensile strength. The samples were heated at a rate of 10 K min–1 up to 673 K, 2 K min–1 up to 973 K, and followed by 1 K min–1 up to the target temperature 973 K at which the sam-ple was maintained for 12 h. The composites were characterized by powder XRD method. The phases present in the sintered samples were ascertained by a powder X-ray diffractometer (Rigaku Rotaflex, Japan) using Cu–Kα radiation with in a 5 mass percent limit of its detection of impurity phases. The samples were compacted into cylindrical discs of 10 mm diameter and 5 mm thickness in a hy-draulic press at a pressure of 100 Mpa and subsequently heated to 473 K for 3 h and cooled to room temperature. The electrical contacts were made on the surface of the pellet by the means of two thin copper wires affixed with silver paint. The pellet was inserted in the middle of the pyrex tube of 5 cm diameter on which the kanthal wire was uniformly wounded externally (Fig. 1). The kanthal wire ends were connected to a varian to vary the temperature and a copper-constantine thermocouple kept at the pellet was used to measure the temperature of the sample. The electrodes were connected to a dc power supply and a picoammeter in series. The temperature dependent (373 – 673 K) electrical conductivity of the composites at a constant potential of 20 V has been measured. The sampling for the scanning electron microscopy (SEM) was carried out by distributing the com-posite on a very thin transparent double-sided sticky plastic foil, which was placed on an aluminum stub of 15-mm diameter. The aluminum stub with the sample was then kept in a chamber of very low pressure where the entire plastic foil containing the sample were coated with gold (60 µ thickness) for 5 min using EIKO IB2 ion coater in order to remove electrical charging during SEM observations. The gold-coated specimens were then observed on a Hitachi SEM S415A operating at an accelerating voltage of 25 kV. The surface area of the composite, WO3–Y2O3 was determined employing BET equations in a Carlo Erba Sorptometer using N2 adsorption at 77 K. All the samples were preheated at 473 K at 20 µ vacuum before measurements. The controlled approximate relative humidity (RH%) environments were achieved [20] using anhy-drous P2O5 (5) and saturated aqueous solutions of CH3COOK (20), CaCl2 ⋅ 6H2O (31), Zn(NO3)2 ⋅ 6H2O

Fig. 1 Experimental set up for the tem-perature variation conductivity measure-ments.

phys. stat. sol. (a) 201, No. 3 (2004) / www.pss-a.com 531


(42), Ca(NO3)2 ⋅ 4H2O (51), NaNO2 (66), NH4Cl (79), BaCl2 ⋅ 2H2O (88) and CuSO4 ⋅ 5H2O (98%) in a closed glass vessel at an ambient temperature (25 °C), which were monitored, by a Barigo hygrometer. Heat cleaning of the samples was found to be a must for better reproducibility and recovery. Hence, the samples were heated to 473 K, followed by cooling in a humidity-free atmosphere before and after the sensitivity measurements especially when the sensors were operated at higher RH. An evacuated glass chamber (200 cm3) was made use of for evaluating the response and recovery time characteristics. This chamber has a provision for a two-way inlet, one for transpiring dry air and the other for transpiring moist air from a wet candle. The air-drying was accomplished by transpiring the air stream through dry-ing columns packed with anhydrous CaCl2 and dry P2O5 connected in series. The resistance measure-ments in dry air as well as in moist air alternatively helped to establish the recovery and response time characteristics for moisture sensing.

3 Results and discussion

The powder XRD patterns (Fig. 2) of the composites correspond to WO3 and Y2O3 only implying that there are no impurity peaks. The samples showed the linear current–voltage curves and thus the electrical conductivity was calculated from the slope by curve fitting using the least square method. Since dc mode is used for resistance measurements at various relative humidities, the activation energy for electrical conduction was determined in air atmosphere in the temperature range 373–673 K by using a dc two-probe method [21]. The potential inaccuracy due to contact resistance could be assumed to be negligible

owing to the high resistivity of the materials under investigation. The dc measurements were carried out to measure chiefly the electronic conduction which could be used in humidity sensing measurements. Due to variation in elec-trical conductance because of the donation of electrons by water molecules either into the con-ductance band or positive holes of the composite materials, the sensitivity of the material varies. A polarization effect at the electrodes, if any, may be overlooked as the measurements were made uni-formly and graphical methods of evaluation and comparison have been resorted. The activation energies in Figs. 3a–f reproducible within 5–8% error limit for the samples in the temperature range of 373 – 673 K were obtained (Table 1) using the expression I = I0 e–Ea/kT, where I is the current, Ea the activation energy, k the Boltzmann constant, and T the temperature Since

Fig. 2 Powder XRD patterns of WO3–Y2O3 compos-ites; a) WOYO-12, b) WOYO-21 and c) WOYO-31.



Table 1 Activation energy and sensitivity factor for the composites and terminal phases.

no. of moles

WO3Y2O3

sample code activation energy Ea (eV)

resistancea) R5% (Ω)

resistancea) R98% (Ω)

Sf

(R5%/R98%)

1 1 1 2 1 3 3 0

0 1 2 1 3 1 2 1

WOYO-10 WOYO-11 WOYO-12 WOYO-21 WOYO-13 WOYO-31 WOYO-32 WOYO-01

0.69 0.42 0.56 0.88 0.48 0.69 0.75 0.75

1.63 (± 0.12) × 108 3.31 (± 0.14) × 109 1.11 (± 0.22) × 1010 7.86 (± 0.21) × 109 8.66 (± 0.31) × 109 7.77 (± 0.14) × 109 7.61 (± 0.22) × 109 2.35 (± 0.17) × 109

1.42 (± 0.18) × 107 1.42 (± 0.13) × 107 2.41 (± 0.23) × 107 1.47 (± 0.16) × 107 2.28 (± 0.15) × 108 5.12 (± 0.31) × 106 5.36 (± 0.25) × 108 6.29 (± 0.19) × 109

11 (± 1) 230 (± 15) 459 (± 50) 535 (± 40) 38 (± 2) 1535 (± 80)142 (± 5) 38 (± 3)

a) R5% and R98% are dc resistances measured at 25 °C corresponding to 5 and 98% RH, respectively.

ln I = ln I0 –Ea/kT, the plot of ln I versus T –1 gives a linear curve. The activation energy was calculated from the slope (–Ea/kT). The drop in dc resistance (Fig. 4) with increase in RH is smaller in the terminal phases WOYO-10 and WOYO-01 than that of the composites. The sensitivity factor Sf = R5%/R98%, where R5% and R98% are the dc resistances (Table 1) at 5% and 98% RH is used for a better appreciation of the material characteris-tics towards moisture. The greater the value of Sf, the higher the sensitivity of the material towards mois

Fig. 3 Activation energy plots of WO3–Y2O3 composites; a) WOYO-11, b) WOYO-12, c) WOYO-21, d) WOYO-13, e) WOYO-31 and f) WOYO-32.



a) b)

Fig. 5 Scanning electron microscopic photographs of WO3–Y2O3 composites; a) WOYO-21 and b) WOYO-31.

Fig. 4 Variation of dc resistance with relative humidity at 25 °C for WO3–Y2O3 composites.

Fig. 6 Pore size and pore volume distribution of WOYO-31 composite.



ture [17]. The maximum sensitivity occurs in WOYO-31 where the resistance drops by more than three orders of magnitude and there is a non-monotonic trend in Sf as a function of composition. The Bru- nauer–Emmett–Teller (BET) surface adsorption studies revealed that the pore size of the samples were distributed (Fig. 5) between 10 and 45 Å in radius, and the specific volume of the pore was 0.01 cm3 gm–1 which can easily trap the water molecules into it. The sensitivity factor of the composition should be indicative of the extent of moisture condensation in the pores. The pore size, grain size of the two phases, and the distribution of the pores should in turn govern the extent of moisture sorption. SEM photograph of the sensor materials sintered at 973 K for 5 h indicated that the porosity and grain size of the materials significantly increases. The scanning electron microscopy (SEM) revealed qualita-tively that WOYO-31 (Fig. 6) composition has greater and larger number of pores in situ level compared to the other composites. As the composites are placed in the higher humid environment, the pores in the composite will be saturated and the condensation of the moisture will take place. As a result, the conduc-tivity increases where as the resistance decreases. Thus this could explain the variation of Sf with differ-ent composition of the composite, but no regular trend could be observed. It should be mentioned that the sensitivity of neither WO3 nor Y2O3 is appreciable towards moisture. The earlier studies on ZnO–Y2O3 composites indicated that [22] the sensitivity factor of pure Y2O3 and WO3 are found to be 10 and 38 respectively. The present study also further indicated that an optimum blend on WO3 and Y2O3 is neces-sary for maximizing the sensitivity. Further, all the measurements were carried out in air ambient in the absence of any oxidizing/reducing gases. In the presence of such gases, cross-sensitivity measurements should be made which, however, is beyond the scope of the present investigation. Charging of the electrodes was not observed when the dc measurements were made on the samples while measuring in closed chambers [22]. WOYO-31 composite was chosen (Fig. 7) to evaluate the recovery and response time of the sensor. Within approximately 2 min of purging with moist air, the dc resistance of the material (109 ohm) under dry condition drops to 106 ohm. However, when dry air was again introduced to monitor the recovery characteristics, the recovery time was around 6 min. When on line measurements were carried out to check the response and recovery behaviour, the charging effect on the electrodes was found to be minimum. Most of the ceramic materials devised for humidity sensing applications require constant heat cleaning. Hence for better response and recovery characteristics, the sensors were repeatedly heat-refreshed at 353 K before and after the measurements.

4 Conclusion

Composites having different mole ratios of WO3–Y2O3 were fabricated and studied for humidity sensing applications. The scanning electron microscope revealed that the WOYO-31 composite has larger and greater number of microscopic pores hence is a good candidate for humidity sensor, which was further evidenced by the surface studies and its sensitivity factor higher than 1.5 × 103. The good response and recovery characteristics even at 298 K were another proof for a good humidity sensor and could be used either in situ or on line applications.

Fig. 7 Variation of dc resistance of WOYO-31 com-posite when exposed to dry and moist air alternatively at room temperature.



Acknowledgements We thank the CSIR, New Delhi for financial support (Ref: 8/293(27)/2002 – EMR – 1). Thanks are also due to Rev. Dr. John Pragasam, Director and Dr. K. Swaminathan of LIFE for their support.

References

[1] Y. C. Yeh, T. Y. Tseng, and D. A. Chang, J. Am. Ceram. Soc. 73, 1992 (1990). [2] E. Traversa, Sens Actuators B 23, 135 (1995). [3] L. Wu, C. C. Wu, C. Y. Chen, and C. T. Huang, J. Electron. Mater. 19, 197 (1990). [4] J. Arndt, Ceramics and Oxides, in: W. Gopel, J. Hesse, and J. N. Zemel (Eds.), Sensors: A Comprehensive

Survey, Vol. 1 (VCH Germany, 1989). [5] Y. Zhang and R. J. Puddephatt, Chem. Mater. 11, 148 (1999). [6] I. Stambolova, K. Konstantinov, S. Vassilev, P. Peshev, and T. Tsacheva, Mater. Chem. Phys. 63, 104 (2000). [7] K. H. Yoon, J. Cho, and D. H. Kang, Mater. Res. Bull. 34, 1451 (1999). [8] K. S. Goto, in: Solid State Electrochemistry and its applications to sensors and electronic devices (Elsevier,

Amsterdam, 1998). [9] J. Schoonman and H. S. Kilian, Solid State Ion 9/10, 1087 (1983). [10] N. Miura, H. Kato, N. Yamazoe, and T. Seiyama, Chem. Lett. 10, 1573 (1983). [11] M. Gauthier and A. Chamberland, J. Electrochem. Soc. 124, 1579 (1977). [12] G. Lu, N. Miura and N. Yamazoe, Sens. Actuators B 65, 125 (2000). [13] M. Baammer and W. C. Maskell, in: Sensors VI Technology system and applications, edited by K. T. V. Grat-

tan and A. T. Augousti (IOP publisher, London 1993) p. 21. [14] S. Deniard-Courant, Y. Pifford, P. Barbonx, and J. Liva, Solid State Ion 27, 189 (1988). [15] S. Chandra and S. A. Hashmi, Solid State Ion 40, 460 (1990). [16] A. Menne and W. Weppner, Electrochim. Acta. 36, 1823 (1999). [17] C. N. Xu, K. Miyazaki, and T. Watanabe, Sens. Actuators B 46, 87 (1998). [18] K. Miyazaki, Chao-Nan Xu, and M. Haeda, J. Electrochem. Soc. 141, L35 (1994). [19] H. Iwahara, H. Uchida, and J. Kondo, J. Appl. Electrochem. Soc. 13, 365 (1983). [20] A. M. E. Suresh Raj, C. Mallika, O. M. Sreedharan, and K. S. Nagaraja, Sens. Actuators B 81, 229 (2002). [21] T. Taka, Synth. Met. 57, 5014 (1993). [22] A. M. E. Suresh Raj, C. M. Magdalane, and K. S. Nagaraja, phys. stat. sol. (a) 191, 230 (2002).

solid state electrical conductivity and humidity sensing properties of wo3–y2o3 composites

Documents