Development of Y3+ and Mg2+-doped zirconia thick film humidity sensors
Meiying Su,Jing Wang,Yuwen Hao
DOI: https://doi.org/10.1016/j.matchemphys.2010.12.034
IF: 4.778
2011-01-01
Materials Chemistry and Physics
Abstract:Y 3+ -doped and Mg 2+ -doped zirconia thick film humidity sensors were investigated. The humidity sensors exhibited good sensing characteristics. The sensor got a high sensitivity with the impendence changed five orders of magnitude from 10 8 to 10 3 Ω in the relative humidity (RH) range of 11–98% at 20 °C. The response time is about 30 s for Y 3+ doped sensor, and 5 s for Mg 2+ doped one, and recovery time is 5 s for both sensors. Small humidity hysteresis is about 3% RH and 4% RH for Y 3+ and Mg 2+ doped ZrO 2 sensors, respectively. Moreover, good repeatability, linearity and temperature properties of the both sensors were also exhibited. X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS) and defect structure were used to analyze the influence of dopant on humidity sensitive properties. And the mechanism of humidity sensing properties was also discussed. Keywords Humidity sensor Y 3+ and Mg 2+ doped zirconia Thick films 1 Introduction Monitoring and controlling environmental humidity is highly necessary for domestic demands and industrial processes [1] . Many kinds of materials have been studied including ceramic [2–6] , organic polymers [7,8] , semiconductor type [9] , and solid electrolytes [10] . Due to its unique properties, such as thermally and chemically stability, high toughness, high strength and catalytic characterization, ZrO 2 is widely used to detect gases, such as O 2 [11–14] , NO x [15–18] , CO [19–21] , H 2 S [22] , H 2 [23,24] , SO 2 [25] and so forth. In Ref. [17] , a closed-one-end of YSZ tube (8 mol% Y 2 O 3 doped, Japan) was used to fabricate the NO x sensor sintered at 1200 °C for 2 h for monitoring car exhausts by Miura. In 2009, Brito-Chaparro et al. investigated 9.25 mol% MgO (preheated at 1200 °C) as a stabilizer in monoclinic ZrO 2 had great influence on fracture toughness, the ZrO 2 composition was sintered at 1720 °C during 1 h [26] . Some humidity sensors based on zirconia were also developed using various techniques for different application environments [2,27–33] . A limiting-current-type zirconia humidity sensor was studied in the range of 0–75% RH with high input voltage about A.C. 100 V/200 V by Yagi and Ichikawa [27] . The zirconia thin film humidity sensors were investigated at room temperature by Niranjan et al. [29] and Biju and Jain [33] with more than one order changes in resistance. Yttria stabilized zirconia and In 2 O 3 composite material were reported for detection of low concentration of water vapor at high temperature in the range of 500–900 °C [31] . Earlier study on bare monoclinic zirconia thick film humidity sensor was reported by our research group [32] . However, the response behavior of the bare zirconia humidity sensor took more than one hundred seconds, and the humidity hysteresis loop was over 8% RH in our previous work [32] . So the properties of zirconia humidity sensor need to be improved. In this report, the Y 3+ -doped and Mg 2+ -doped zirconia thick film humidity sensors were fabricated. XRD and XPS patterns were used to characterize the sensing materials. The humidity sensing properties of the doped ZrO 2 humidity sensors were investigated and the humidity sensing mechanism was also discussed. 2 Experimental 2.1 Humidity sensor fabrication ZrO 2 (≥99.9% purity, 20 nm) and MgO (≥99% purity, 40 nm) were purchased from Haitai Nanometer Materials Co. in Nanjing. Y 2 O 3 (4 N purity, 45 nm) was purchased from China National Pharmaceutical Industry Corporation Ltd. 3 mol% Y 2 O 3 was added to nanometer ZrO 2 to form a mixture powder. The mixture powder was ground with deionized water for 2 h to make a homogenous paste. The paste was screen printed on ceramic substrate with eight pairs of gold interdigitated electrodes to form a sensing film, and then the film was dried in air at room temperature for 2 h to obtain an Y 3+ -doped ZrO 2 thick film humidity sensor. Finally, this sensor was annealed at 450 °C for 20 min. As above process, a mixture of ZrO 2 and MgO with molar ratio of 9:1 was ground and fabricated a 10 mol% Mg 2+ -doped zirconia thick film humidity sensor. A schematic representation of the humidity sensors of bare or doped zirconia thick film humidity sensors is similar to our previous work [32] . Improving the stability and durability, the humidity sensors were aged for 24 h at 100% RH with A.C. 1 V. The crystal phases of the bare and doped zirconia powder were investigated by X-ray diffraction (XRD-6000, Shimadzu, Japan). The X-ray photoelectron spectroscopy (XPS) was characterized by a VG ESCALAB MK2 (UK) with Al Kα radiation and a powder of 250 W. Some experiments about Y 3+ and Mg 2+ doped into ZrO 2 were carried out. The doping concentration ranges were all 1–15 mol%. The results indicated that the humidity sensing properties were good when the doping concentration was 3 mol% Y 3+ and 10 mol% Mg 2+ , respectively. For increasing the formation of Schottky defect, thermal excitation was used via annealing the doped ZrO 2 materials at 450 °C. 2.2 Sensor measurement The electrical properties of the both sensors in various humidity environments were measured on a LCR intelligent test meter (ZL5, Shanghai, China). The voltage applied in testing process was A.C. 1 V, and the measurement frequency varied from 20 Hz to 100 kHz. The sensor was successively put into several chambers with different relative humidity levels at temperature of 20 °C. The different RHs were obtained with saturated aqueous solutions of salts: LiCl (11%), MgCl 2 (33%), KCO 3 (42%), Mg(NO 3 ) 2 (54%), NaCl (75%), KCl (85%) and K 2 SO 4 (98%). 3 Results and discussion 3.1 Materials characterization Fig. 1 shows the XRD patterns of the bare monoclinic zirconia (curve (a)), 3 mol% Y 3+ -doped (curve (b)), and 10 mol% Mg 2+ -doped zirconia powders (curve (c)), respectively. We observe from Fig. 1 that in curve (b), the peaks of cubic Y 2 O 3 were observed at 20.499°, 29.149°,and 33.784° corresponding to the standard card of the powder diffraction file (PDF-ICSD, File Id 79-1257). But there was not obvious peak of MgO in curve (c). According to the Scherrer equation, the average sizes of particles were 20 nm, 25 nm, and 25 nm for bare monoclinic zirconia, 3 mol% Y 3+ -doped and 10 mol% Mg 2+ -doped zirconia powders, respectively. Table 1 lists the main peak positions data of XPS spectrum O 1s and Zr 3d 5/2 of bare ZrO 2 , Y 3+ -doped and Mg 2+ -doped ZrO 2 , respectively. We can see from Table 1 that the peaks of O 1s for the doped ZrO 2 shift 0.1–0.2 eV to the low combination energy direction, and the peaks of Zr 3d 5/2 shift 0.3 eV to the high combination energy direction, compared with the peaks of bare ZrO 2 . The variation of the combination energy might be due to the following reasons: the lattice structure of ZrO 2 is disordered not the regular monoclinic after being mixed with Y 3+ and Mg 2+ , and small distortions are formed when the dopants come into the ZrO 2 crystalline. Fig. 2 (a)–(c) shows the XPS patterns of oxygen O 1s spectra of bare ZrO 2 , Y 3+ -doped ZrO 2 , and Mg 2+ -doped ZrO 2 , respectively. Each O 1s spectrum (solid line) includes two peaks, one with combination energy 529 eV corresponding to the lattice oxygen O l and the other with combination energy 531 eV corresponding to the adsorption oxygen O a [34] . Table 2 gives the proportions of the lattice oxygen and adsorption oxygen, and the relative ratio O a /O l . We notice that the proportions of the lattice oxygen are larger than that of the adsorption oxygen for these three kinds of materials. This is because that in the materials with ZrO 2 as the main component, lattice oxygen remains to be the bulk despite the existence of many crystalline interfaces and some adsorption centers formed by the hanging bonds. We can also see from the table that the ratio of the adsorption oxygen to the lattice oxygen (O a /O l ) in the doped materials, compared with that in undoped ZrO 2 , increases significantly, indicating the increase of the adsorption oxygen. Being mixed with dopants, the perfection of the zirconia lattice structure and the bond situation on the grain boundary of ZrO 2 become worse, resulting in the increases of the adsorption oxygen and the ratio O a /O l [34] . 3.2 Humidity sensing behavior 3.2.1 Sensitivity Sensitivity of humidity sensor can be defined as the following equation: (1) S ( k Ω % R H − 1 ) = Z 98 − Z 11 98 − 11 where Z 98 is the steady-state impedance after the sensors were exposed to 98% RH level, and Z 11 is the impedance value at 11% RH. Fig. 3 shows the dependence of impedance on RH for bare zirconia [32] and doped sensors at the same temperature (20 °C). From the results, it can be found that the sensitivity of the doped zirconia thick film humidity sensors is better than the bare one. The impedance of the doped sensors varied five orders of magnitude from 10 8 to 10 3 Ω when the relative humidity changed from 11 to 98% RH. At the same RH, the impedances of the doped zirconia sensors are larger than the undoped one. It may be attributed to the theory of depletion-type adsorption of semiconductor materials. It is known that the n-type electronic conduction is available for the bare ZrO 2 humidity sensor [32] , electrons as the majority carrier play the main role in conduction process. Based on the valency control principle, the addition of low valent cation dopants results in the concentration of electrons falling down [35] , indicating conductivity declining and impedance becoming larger. In this experiment, we can define the linear equation of the sensor as follows: (2) log y = A x + B Here y stands for impedance Z of the sensors, x stands for linear range of relative humidity level, A is the slope of curve standing for sensitivity, and B is the optimal exponential on the ordinate axis at 0% RH. The fitted values of the parameters are tabulated in Table 3 according to the fitting curve by using MATLAB software, wherein, R 2 is generally the best indicator of the fit quality. Comparing the column A , the sensitivities of the doped zirconia sensors are better than the bare one. Fig. 4 gives the dependence of impedance on RH at different measurement frequencies for Y 3+ -doped zirconia humidity sensors. It can be noted that the measurement frequency obviously influences the impedance properties of the sensor. The Mg 2+ -doped one has the same frequency properties. The log-linearity and sensitivity for both sensors at 20 Hz and 100 Hz were better than the ones at other frequency in the RH range of 33–98%. So the performance of humidity sensing of the both sensors was measured at the low frequency of 20 Hz. 3.2.2 Hysteresis Hysteresis curves exhibit humidification and desiccation processes of humidity sensors. The hysteresis of Y 3+ -doped and Mg 2+ -doped ZrO 2 humidity sensors at 20 Hz is shown in Fig. 5 (a) and (b) , respectively. Hysteresis of humidity sensor is defined as follows [36] : (3) H ( % RH ) = Z H − Z D S where Z H and Z D are the impedances obtained at the same RH in the humidification and desiccation process, respectively and S is sensitivity of the sensor obtained via Eq. (1) . The maximum of hysteresis is about 4% RH for the Y 3+ -doped sensor, and 3% RH for the Mg 2+ -doped ZrO 2 humidity sensor. 3.2.3 Response and recovery behavior Response and recovery behavior is one of the important features for evaluating the performance of the humidity sensors. Response time in the case of adsorption process and recovery time in the case of desorption are defined as the time taken by the sensor to achieve 90% variable quantity. Fig. 6 (a) and (b) gives the response and recovery properties and repeatability of the Y 3+ -doped and the Mg 2+ -doped humidity sensors at 20 Hz for four cycles when RH changed between 11% and 98%RH, respectively. The results indicate the doped ZrO 2 humidity sensors take on rapid response and recovery characteristics and good repeatability. Table 4 lists the response and recovery times, and the hysteresis values of the bare [32] , Y 3+ -doped and Mg 2+ -doped ZrO 2 humidity sensors. We can see from Table 4 that the response and recovery times of the doped sensor are much less than undoped ZrO 2 , and the humidity hysteresis of Y 3+ and Mg 2+ -doped ZrO 2 sensors is less than the one of bare ZrO 2 humidity sensor. The data in Table 4 demonstrate that some properties of humidity sensor are much improved by doping Y 3+ and Mg 2+ ions into ZrO 2 . 3.2.4 Temperature properties Fig. 7 shows the effect of temperature on the Mg 2+ -doped ZrO 2 humidity sensor at 20 °C, 30 °C, 40 °C and 50 °C. We can see that as the temperature increases, the impedance of the sensor decreases, and the sensitivities of the curves are nearly same. It is known that as temperature increases, the carriers of the doped ZrO 2 increase [35] . Meanwhile, the adsorbed water molecules make the H 3 O + hydrate more quickly into H 2 O and H + , resulting in the charged ions increase. So the magnitude of impedance decreases obviously as temperature increases. And the temperature properties for the Y 3+ -doped ZrO 2 sensor are similar to the Mg 2+ -doped ZrO 2 humidity sensor. 3.3 Humidity sensing mechanism Chemisorption and physisorption of water from a humid atmosphere on different oxides have been investigated, such as TiO 2 [37] , SiO 2 [38] , α-Fe 2 O 3 [39] . Chemisorption of water molecule took place during the sensors fabrication process because of its high adsorption enthalpy for ZrO 2 materials [40] . When a water molecule approaches the Zr–O components, it dissolves into H + and OH − ions which boned to the O 2− and the metal ion on the oxide surface, respectively, to form two hydroxyl (OH) groups bonded to the metal ions (Zr–OH) [40] . Once the chemisorbed layer formed, it is hard affected by the humid atmosphere which exposed to. The humidity sensing behaviors of the sensors measured at room temperature are due to the physisorption on top of the chemisorbed one commenced at low temperature. When RH is low, a few of physisorbed water molecules attached to two adjacent Zr–OH components to form H 3 O + at low concentration, the hydronium (H 3 O + ) proton hopping conduction mechanism is dominated. As RH increases, more and more water molecules physisorbed to form continuous water layers, a large amount of H + decomposed from water molecules become the dominant charged carriers. Besides the ionic conductivity mechanism, polarization of the sensing material took place in the electric field. As mentioned above, some humidity sensing properties of the doped zirconia thick film sensors are better than those of the bare one. We try to explain the improvements as follows: firstly, the adsorption oxygen content of the doped ZrO 2 humidity sensors is higher than the undoped one ( Fig. 2 and Table 2 ), indicating the adsorption capacity of the doped ZrO 2 to water molecular increases. Secondly, when Y 3+ or Mg 2+ doped into ZrO 2 , Y 3+ or Mg 2+ will substitute the site of Zr 4+ , which will deplete some electrons of n -type sensing material. This process could be explained by the following equations: (4) Z r Zr + MgO + 2 e ′ + 1 2 O 2 → M g ″ Zr + ZrO 2 (5) 2 Z r Zr + Y 2 O 3 + 2 e ′ + 1 2 O 2 → 2 Y ′ Zr + 2 ZrO 2 where M g ″ Zr and y ′ Zr stands for a substituted Mg 2+ and Y 3+ ion in the Zr 4+ site, and the signs ′ stands for a negative charge. The electronic conductivity of the sensing material decreases. 4 Conclusions Y 3+ -doped and Mg 2+ -doped zirconia thick film humidity sensors were fabricated. Good sensing characteristics were exhibited, such as high sensitivity with impedance varying five orders of magnitude from 10 8 to 10 3 Ω in a wide humidity range of 11–98%, small humidity hysteresis, rapid response and recovery behavior and good repeatability. Linearity and temperature properties were also reported in the present paper. The improved performance was analyzed by means of the XPS data and defect structure of materials. The improvement of the ZrO 2 humidity sensing properties may be attributed to the increase in the adsorption capacity and the decease of electronic conductivity by doping Y 3+ and Mg 2+ . Acknowledgements The subject was supported by The National Natural Science Foundation of China ( 60474052 , 90607002 ) and 973 Projects (2011CB302101 and 2011CB302105). References [1] E. Traversa Sens. Actuators B 23 1995 135 [2] M.K. Jain M.C. Bhatnagar G.L. Sharma Sens. Actuators B 55 1999 180 [3] P.G. Su S.C. Huang Sens. Actuators B 105 2005 170 [4] M.K. Jain K.P. Biju Thin Solid Films 516 2008 2175 [5] R. Sundaram Sens. Actuators B 124 2007 429 [6] T.H. Yeh C.C. Chou J. Phys. Chem. Solids 69 2008 386 [7] J.H. Kim S.M. Hong J.S. Lee B.M. Moon K. Kim 2009 4th IEEE International Conference on NEMS Shenzhen, China, 5–8 January 2009 2009 703 [8] P. Li Y. Li B.Y. Ying M.J. Yang Sen. Actuators B 141 2009 390 [9] L. Juhász J. Mizsei Thin Solid Films 517 2009 6198 [10] M. Zhou A. Ahmad Sens. Actuators B 129 2008 285 [11] N. Rajabbeigi B. Elyassi A.A. Khodadadi S. Mohajerzadeh Y. Mortazavi M. Sahimi Sens. Actuators B 108 2005 341 [12] H. Kaneko T. Okamura H. Taimatsu Y. Matsuki H. Nishida Sens. Actuators B 108 2005 331 [13] N. Izu W. Shin I. Matsubara N. Murayama N. Oh-hori M. Itou Sens. Actuators B 108 2005 216 [14] G.L. Tan X.J. Wu Thin Solid Films 330 1998 59 [15] V.V. Plashnitsa T. Ueda P. Elumalai N. Miura Sens. Actuators B 130 2008 231 [16] S.P. Oncley M. Buhr D.H. Lenschow D. Davis S.R. Semmer Atmos. Environ. 38 2004 5389 [17] N. Miura M. Nakatou S. Zhuiykov Ceram. Int. 30 2004 1135 [18] T. Nakamura Y. Sakamoto K. Saji J. Sakata Sens. Actuators B 93 2003 214 [19] K. Mochizuki R. Sorita H. Takashima K. Nakamura G. Lu Sens. Actuators B 77 2001 190 [20] N. Miura T. Raisen G. Lu N. Yamazoe Sens. Actuators B 47 1998 84 [21] H. Okamoto H. Obayashi T. Kudo Solid State Ionics 1 1980 319 [22] N. Miura Y. Yan G. Lu N. Yamazoe Sens. Actuators B 34 1996 367 [23] K.I. Shimizu I. Chinzei H. Nishiyama S. Kakimoto S. Sugaya H. Yokoi A. Satsuma Sens. Actuators B 134 2008 618 [24] G. Lu N. Miura N. Yamazoe Sens. Actuators B 35 1996 130 [25] S. Suganuma M. Watanabe T. Kobayashi S.I. Wakabayashi Solid State Ionics 126 1999 175 [26] J.A. Brito-Chaparro A. Aguilar-Elguezabal J. Echeberria M.H. Bocanegra-Bernal Mater. Chem. Phys. 114 2009 407 [27] H. Yagi K. Ichikawa Sens. Actuators B 25 1995 701 [28] I.S. Mulla S.D. Pradhan K. Vijayamohanan Sens. Actuators A 57 1996 217 [29] R.S. Niranjan S.D. Sathaye I.S. Mulla Sens. Actuators B 81 2001 64 [30] I.C. Cosentino E.N.S. Muccillo R. Muccillo Mater. Chem. Phys. 103 2007 407 [31] M. Nakatou N. Miura Electrochem. Commun. 6 2004 995 [32] J. Wang M.Y. Su J.Q. Qi L.Q. Chang Sens. Actuators B 139 2009 418 [33] K.P. Biju M.K. Jain Sens. Actuators B 128 2008 407 [34] J. Wang W.P. Yan J.C. Zhang F.B. Qiu T. Zhang G.F. Liu B.K. Xu Mater. Chem. Phys. 69 2001 288 [35] E.C. Akubuiro X.E. Verykios J. Phys. Chem. Solids 50 1989 17 [36] M. Matsuguch T. Kuroiwa T. Miyagishi S. Suzuki T. Ogura Y. Sakai Sens. Actuators B 52 1998 53 [37] T. Morimoto T. Iwaki J. Chem. Soc. Faraday Trans. 1 83 1987 943 [38] T. Morimoto M. Nagao J. Imai Bull. Chem. Soc. Jpn. 44 1971 1282 [39] R. Kuwabara T. Iwaki T. Morimoto Langmuir 3 1987 1059 [40] S. Raz K. Sasaki J. Maier I. Riess Solid State Ionics 143 2001 181