# Fileset

[132_23121.pdf](https://mdr.nims.go.jp/filesets/d2d065f4-26aa-4ccb-b6dc-bd4d17d6c069/download)

## Creator

[Yutaka Adachi](https://orcid.org/0000-0003-2666-5521)

## Rights

[Creative Commons BY Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/)

## Other metadata

[Effect of Si addition on epitaxial growth and gas sensing properties of tungsten oxide films](https://mdr.nims.go.jp/datasets/0d5ea082-4a25-42ae-83d0-3a1a77ea6e71)

## Fulltext

Effect of Si addition on epitaxial growth and gas sensing properties of tungsten oxide filmsFULL PAPEREffect of Si addition on epitaxial growthand gas sensing properties of tungsten oxide filmsYutaka Adachi1,³1Affiliation Electro-ceramics Group, Research Center for Electronic and Optical Materials,National Institute for Materials Science (NIMS), 1–1 Namiki, Tsukuba, Ibaraki 305–0044, JapanWO3 epitaxial films with the same thickness and in-plane crystal grain size were prepared using pulsed laserdeposition with various Si contents in the target to clarify the effect of Si addition to WO3 on the gas sensingproperties. X-ray diffraction measurements indicated that the films grown on the (1�102) face of sapphire had a(001) orientation with in-plane epitaxial relationships of [110]WO3//[01�11] or [110]WO3//[2�110]Al2O3, regard-less of the Si content. Scanning probe microscopy observations revealed that particles with a diameter of severaltens of nanometers grow on the surface of the Si-doped WO3 film. Measurements of the gas response to ethanoland acetone showed superior gas selectivity towards acetone gas at low temperatures, which is due to thecatalytic effect of SiOx particles on the film surface.Key-words : Tungsten trioxide, Epitaxial film, Si doping, Acetone, Ethanol, Gas selectivity[Received June 23, 2023; Accepted February 25, 2024]1. IntroductionN-type oxide semiconductors such as zinc oxide and tinoxide have been studied for application as gas sensors andare already in practical use.1) Oxide semiconductor gassensors have advantages such as relatively high sensitiv-ity, ease of miniaturization, and relatively low cost, whichmakes them suitable for use in mobile devices and sensornetwork applications; however, they have the disadvantageof poor gas selectivity.2) This drawback must be addressedfor applications such as health monitoring devices thatwould allow users to easily check their health at home bybreath analysis, and sensor networks that would monitorair pollution gases such as SOx and NOx.Gas detection with oxide semiconductor gas sensorsuses the change in the electrical resistance of the semicon-ductor caused by the reaction between oxygen adsorbed onthe sensor surface and the target gas to be detected, suchas hydrogen or ethanol. Therefore, similar reducing gasesin the atmosphere cause similar changes in the electricalresistance of the sensor, and this is one of the reasons forthe poor gas selectivity of oxide semiconductor gas sen-sors. As such, there have been many attempts to increasethe response to only specific gases by the addition ofimpurities.3)–5)Tungsten oxide (WO3) is also known as a matrix mate-rial for oxide semiconductor gas sensors, and it has beenreported to respond to NOx,6)–9) O2,10)–12) H2S,13)–15) andvolatile organic compound gases16),17) when doped withvarious impurities. However, the role of impurity dopingon the improvement of gas selectivity has yet to be clar-ified, especially with respect to the selection of additive im-purity species to increase the response to a particular gas.The reason why the effect of impurity addition is un-clear is probably because the sensor characteristics areaffected not only by the addition of impurities, but also bychanges such as the particle size, shape, and the crystalfacet exposed on sensor surfaces.18)–24) Oxide powders ornanostructures are typically employed for semiconduc-tor gas sensors. In the case of powders and nanostructures,the addition of impurities to improve sensor performancetends to change other parameters such as the shape andsize of the sensor matrix particles.25) It is important to cor-relate material parameters such as the grain size and shapeto obtain clear guidelines for improvement of the gas sens-ing properties. Therefore, to elucidate the true effect ofimpurity addition, it is necessary to prepare and comparesamples in which only the impurity addition concentrationdiffers, while other parameters, such as the particle sizeand shape, are the same.In this study, we have attempted to use epitaxial thinfilms as gas sensors to clearly capture the effect of impurityaddition. The thickness of epitaxial films can be easilycontrolled, and it is possible to fabricate samples with uni-form particle sizes. In addition, only specific crystal planesof these thin films are exposed on the topmost surface;therefore, the influence of differences in the crystal planeson the sensor properties can be suppressed. Furthermore,the same in-plane crystal orientation can be produced;therefore, the influence of the bonding state between crys-tal grains can be expected to be suppressed. The author hassuccessfully clarified the effects of spontaneous polariza-tion and Mg addition on the gas sensing properties of zinc³ Corresponding author: Y. Adachi; E-mail: adachi.yutaka@nims.go.jpJournal of the Ceramic Society of Japan 132 [5] 227-231 2024DOI https://doi.org/10.2109/jcersj2.23121 JCS-Japan©2024 The Ceramic Society of Japan 227This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/),which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.https://doi.org/10.2109/jcersj2.23121https://creativecommons.org/licenses/by/4.0/oxide (ZnO) epitaxial thin films used as gas sensors.23),24)In this study, epitaxial WO3 thin films doped with Si wereprepared on sapphire substrates to investigate how thesensor properties depend on the Si concentration. Sensorsthat exhibit excellent gas selectivity for acetone are re-quired for the development of hand-held breath analysisinstruments. Righettoni et al. reported that Si-doped WO3exhibits excellent acetone gas selectivity;26) however, themechanism that underlies this effect is still unclear. Clar-ification of the mechanism for the effect of Si addition isthus expected to contribute to further improvement ofacetone gas sensing selectivity.2. Experimental detailsSi-doped WO3 films were prepared by pulsed laserdeposition using a KrF excimer laser (­ = 246 nm) with apulse width of 20 ns, a repetition rate of 5Hz, and a laserfluence of approximately 1 J/cm2. The background pres-sure in the growth chamber was less than 4 © 10¹7 Pa. Si-doped WO3 ceramics (Toshima Co. Ltd.) were used as atarget. The films were grown at oxygen (O2) pressures of1.3 © 10¹3 to 33.3 Pa, at substrate temperatures (Tg) in therange of 350–650 °C. The films were deposited on the(10�12) face of sapphire to obtain epitaxial films.The film thickness was determined by X-ray reflectivity(PANalytical X’Pert MRD) measurements with CuK¡radiation. The orientation and crystallinity of the filmswere determined by X-ray diffraction analysis (XRD;PANalytical X’Pert MRD, CuK¡ radiation and a hybridmonochromator consisting of a coupled X-ray mirror anda 2-bounce Ge 220 monochromator). The film morphol-ogy was investigated with tapping mode scanning probemicroscopy (SPM; SII SPA-400). An in-house-built gassensing test system was used to measure the gas sensingproperties of the Si-doped WO3 films. The details of thistest system have been described elsewhere.22) The gasresponse to synthetic air and the target gas (40 ppm H2,ethanol, acetone) were measured at operating temperaturesof 150–350 °C. The working voltage of the measurementcircuit was set to 1Vand the film resistance was measured.The sensor response S, was determined viaS ¼ Ra=Rg ð1Þwhere Ra and Rg are the sample resistances in synthetic airand the target gas, respectively.3. Results and discussionThe effect of the oxygen partial pressure introducedduring film deposition was investigated to determine thegrowth conditions for epitaxial films. Figure 1 showsXRD measurement results for films fabricated at variousoxygen partial pressures in the range from 1.3 © 10¹3 to33.3 Pa. When the oxygen partial pressure was as low as1.3 © 10¹3 Pa, only a diffraction peak due to WO2 wasobserved, and no diffraction peaks associated with WO3appeared. When the oxygen partial pressure was increasedto 5.6 Pa, diffraction peaks due to WO3 002 and 004 beganto appear. When the oxygen partial pressure was furtherincreased, two more peaks appeared near the 002 and 004diffraction peaks. Figure 1(b) shows a magnified view ofthe 2ª range from 22.0 to 26.5 degrees, and these peakscan be indexed as 020 and 200 diffractions from mono-clinic WO3. These results indicate that when the oxygenpartial pressure is low during film growth, WO3 does notgrow because the oxygen supply is insufficient, and in-stead, the growth of oxygen-deficient WO2 occurs. Whenoxygen is supplied at around 5.6 Pa, (001) oriented WO3films are then grown. If the oxygen partial pressure isfurther increased, then polycrystalline films with (100) and(010) oriented crystal grains are formed.Next, the temperature for epitaxial growth was opti-mized. Figure 2 shows the XRD measurement results forfilms fabricated at various growth temperatures while theoxygen partial pressure was kept constant at 6.7 Pa. Theintensity of the diffraction peaks associated with WO3became weaker as the growth temperature decreased, andthe 002 diffraction peak tended to shift toward lowerangles, as shown in Fig. 2(b). This is due to a decrease incrystallinity and the formation of defects in the films.Therefore, this result indicates that film growth at temper-atures higher than 550 °C is necessary to grow films withthe desired degree of crystallinity.Fig. 1. (a) XRD profiles for WO3 films grown on r-plane sap-phire substrates at oxygen partial pressures in the range from1.3 © 10¹3 to 33.3 Pa. (b) Magnified view of the 2ª range from22.0 to 26.5 degrees.Fig. 2. (a) XRD profiles for WO3 films grown on r-plane sap-phire substrates at temperatures in the range from 350 to 650 °C.(b) Magnified view of the 2ª range from 22.0 to 26.5 degrees.Adachi: Effect of Si addition on epitaxial growth and gas sensing properties of tungsten oxide filmsJCS-Japan228To determine the in-plane orientation of the preparedfilms, ¤ scans of the WO3 222 diffraction peaks wereperformed. Table 1 shows the space groups and latticeconstants of monoclinic and tetragonal WO3. Only fourdiffraction peaks were observed in Fig. 3, which indicatesthat the films have 4-fold symmetry and that tetragonalcrystals were grown. However, tetragonal WO3 is a high-temperature phase observed above 720 °C; therefore, it isunlikely that tetragonal crystals were formed. MonoclinicWO3 is stable at room temperature; therefore, there is ahigh possibility that monoclinic crystal was formed. How-ever, if a monoclinic WO3 film was formed, then only twodiffraction peaks should be observed in the 222 ¤ scandue to the symmetry of a (001) oriented monoclinic WO3film. The reason why four diffraction peaks were observedcould be that (001) oriented monoclinic WO3 grew with in-plane 90-degree rotational domains. The full width at halfmaximum (FWHM) of the WO3 222 diffraction peak forundoped WO3 was unnaturally broad at 3.83 degrees. Thereason for this unnaturally broad FWHM is that the 222 and2�22 diffraction peaks, which should be separately observeddue to the rectangular ab-plane, were overlapped due to theinsufficient crystallinity of the film. Only four peaks wereobserved in the ¤ scan, which indicates there were no 45-degree rotational domains, and the in-plane orientation ofthe film is the substrate 10�11� �direction and the ©110ªdirection for WO3 are parallel to each other. These resultsindicate that (001) oriented epitaxial WO3 films with 90-degree in-plane rotation domains can be grown at an oxy-gen pressure of around 6.7 Pa and a temperature of 550 °C.Films were prepared using Si-doped WO3 targets underthe described growth conditions. Figure 4 shows the XRDresults for the prepared films. All the films prepared fromthe 1, 2.5 and 10% Si-doped targets showed (001) orien-tation. Figure 4(b) shows an enlarged view in the vicinityof the 002 diffraction peak; the 002 peak tends to shifttoward lower angles as the Si content increases. This maybe due to the presence of Si in the crystal lattice or theformation of crystal defects in the films.Figure 5 shows ¤ scan measurement results for the WO3film prepared with the 10% Si-doped target. As with theundoped WO3 film only four diffraction peaks were ob-served, which indicates that the in-plane orientation of thefilm is the same as that of the undoped film and the Si-doped films also have 90-degree in-plane rotation domains.Epitaxial films were obtained for both the undoped WO3and Si-doped WO3 targets, and the gas sensing propertiesof these films were measured. Table 2 gives the thick-nesses of these films. The gas sensing response of theundoped WO3 and 1% Si doped films are shown inFigs. 6(a) and 6(b). Two points should be noted regardingTable 1. Structural information for the WO3 phases27)Structure SymbolSpacegroupLatticeparametersMonoclinic £ P21/na = 0.7301 nmb = 0.7538 nmc = 0.7689 nm¢ = 90.893°Tetragonal ¡ Pmnba = 0.5250 nmc = 0.3915 nmFig. 3. (a) ¤ scans of the WO3 222 diffraction peaks forundoped WO3 film. A ¤ scan of the Al2O3 00012 diffraction peakis also shown as a reference. (b) Magnified view of the Al2O300012 and WO3 222 diffraction peaks.Fig. 4. (a) XRD profiles for Si-doped WO3 films grown on r-plane sapphire substrates at Si concentrations in the range from 0to 10 at%. (b) Magnified view of the 2ª range from 22.0 to 26.5degrees.Fig. 5. (a) ¤ scans of the WO3 222 diffraction peaks for Si10 at.% doped WO3 film. A ¤ scan of the Al2O3 00012 diffractionpeak is also shown as a reference. (b) Magnified view of theAl2O3 00012 and WO3 222 diffraction peaks.Journal of the Ceramic Society of Japan 132 [5] 227-231 2024 JCS-Japan229these measurements. First, the responses for acetone andethanol are almost the same at 350 °C. The maximumresponse for ethanol occurs around 300 °C, while that foracetone is larger and occurs above 350 °C. Measurementswere made only up to 350 °C to prevent degradation of thegold electrode fabricated on the film surface. However, it isexpected that measurements at higher temperatures wouldshow a higher response for acetone than for ethanol. Theother point is that the response in the range of 200 to 300 °Ctended to increase in the order of ethanol > acetone > hy-drogen, although the response of the Si-doped sampletoward ethanol at 150 °C was almost the same as that foracetone.As the Si concentration was further increased, the re-sponse toward acetone at 150 °C exceeded that of ethanol.Figures 6(c) and 6(d) show the results of WO3 sensingmeasurements with the addition of 2.5% and 10% Si,respectively. The addition of 10% Si reduced the differ-ence between the responses toward acetone and ethanol,although the response toward acetone still exceeded thatfor ethanol. This trend was not observed for the undopedWO3 film, which suggests that this is due to Si addition.On the other hand, the behavior at high temperaturesshowed the same trend as that without Si addition and with1% Si addition; therefore, this response could be tungstenoxide-derived rather than an effect of Si addition.The reason for these changes in the sensor properties isas follows. There are two different effects that improvesensor properties through impurity doping or addition:electrical and chemical sensitization effects.28) One is theeffect of a residual electron concentration. If the reactionshown in Eq. (2) occurs on the surface of the sensor, thenthe presence or absence of the target gas will cause achange in the electron concentration in the film.Oads þX ! XOþ 2e�: ð2ÞAssuming that the change of the electron concentration inthe film due to the absorption and desorption of oxygen onthe surface is ¤· and that the electron mobility ® does notchange by the atmospheric gas, the sensor response S isexpressed as:S ¼ RaRg¼ ·n g·n air¼ ·n air þ ¤·n·n air¼ 1þ ¤·n·n air; ð3Þwhere Ra and Rg are the sensor resistance in air and in thetarget gas, respectively. ·n_g and ·n_air are the electron con-centrations in the film when exposed to the target gas andair, respectively. If the concentration of the target gas isconstant, i.e., ¤·n is constant, then this equation indicatesthat a smaller residual electron concentration in air, i.e., alarger resistance value, will give a larger sensor response.However, this equation does not explain the difference inthe response with respect to the gas type.Another mechanism to improve the sensor responseproperties is the chemical sensitization effect. This is aneffect in which impurities on the surface of the semicon-ductor sensor matrix promote adsorption and dissociationof the target gas and enhance the reaction of adsorbed gaswith adsorbed oxygen. As shown in Fig. 7, SPM obser-vations indicated Si or SiOx particles with a diameter ofapproximately several tens nm on the top surface of thefilms prepared with the Si-doped target; therefore, theseparticles may have enhanced the response to acetone at150 °C.A shift in the maximum response at high temperatureswas observed for all the thin films prepared in this study. Itis assumed that there are many acetone adsorption sites thatinvolve Watoms on the film surface. A maximum responseshift at high temperature was observed for all the sensorsTable 2. Thickness of undoped and Si-doped WO3 films used inthis studySample Film thickness [nm]Undoped WO3 251 at.% Si-doped WO3 342.5 at.% Si-doped WO3 2310 at.% Si-doped WO3 22Fig. 6. Temperature-dependent sensor response to 40 ppm H2,acetone and ethanol for (a) 0 at.%, (b) 1 at.%, (c) 2.5 at.%, and(d) 10 at.% Si-doped WO3 films.Fig. 7. SPM images of (a) WO3 and (b) 10 at.% Si-doped WO3films grown on r-plane sapphire substrates.Adachi: Effect of Si addition on epitaxial growth and gas sensing properties of tungsten oxide filmsJCS-Japan230regardless of whether Si was doped or not. Therefore, theamount of doped Si has no relation to the maximumresponse temperature shift. In the case of MgZnO filmsensors, the maximum response to acetone is at approx-imately 300 °C.24) On the other hand, that for the WO3 filmsis over 350 °C, which indicates that the maximum responsetemperature shift for WO3 films due to the characteristics ofthe adsorption sites that involve W atoms.The temperature dependence of the sensor responsevalues is determined by the chemical reaction rate at ad-sorption sites at a certain temperature. If the catalytic effectof Si on the sensor response is not significant at high tem-peratures, then the temperature dependence of the responsevalue is determined by the chemical reaction rate at theadsorption sites and is not affected by the number ofadsorption sites. As the Si concentration increases, thenumber of adsorption sites decreases; therefore, the re-sponse value at each temperature should decrease as the Siaddition concentration increases. However, if the numberof adsorption sites on the surface is sufficiently greaterthan the number of acetone gas molecules present on thesurface of the sensor body under equilibrium conditions ata certain temperature, then the sensor response value willnot be affected by the Si dopant concentration but will beaffected by the characteristic of the adsorption sites thatinvolve W atoms. Therefore, the sensor response charac-teristics at high temperatures are be determined by thechemical reaction rate of adsorption sites that involve Watoms present on the thin film surface, and the catalyticaction of Si or SiOx has no significant influence.4. ConclusionEpitaxial WO3 films were fabricated without any varia-tion in particle size, shape and inter-particle bonding state,but with differing Si concentrations to clarify the effect ofSi addition to WO3. All of the samples prepared in thisstudy showed the highest response to ethanol at 300 °Cregardless of the Si content, but the highest response toacetone was observed at 350 °C. This result indicates thattungsten atoms on the surface result in a shift of the maxi-mum sensing response temperature. On the other hand, theaddition of Si resulted in a larger response to acetone thanto ethanol at 150 °C or lower. Particles of several tens nmin diameter were observed on the surface of the Si-dopedWO3 films; therefore, the improvement of the acetoneresponse at low temperatures is due to the catalytic effectof these particles. The results of this study are expected tocontribute to the development for sensors with excellentgas selectivity.Acknowledgments This work was partly supported bya Kakenhi Grant-in-Aid (No. 21K04647) from the JapanSociety for the Promotion of Science (JSPS).References1) H. Ji, W. Zeng and Y. Li, Nanoscale, 11, 22664–22684(2019).2) Q. Ren, Y.-Q. Cao, D. Arulraj, C. Liu, D. Wu, W.-M. Liand A.-D. Li, J. Electrochem. Soc., 167, 067528 (2020).3) K. Vijayalakshmi and A. Renitta, Ceram. Int., 41,14315–14325 (2015).4) A. S. M. I. Uddin and G. S. Chung, Sensor. Actuat. B-Chem., 231, 601–608 (2016).5) F. Fan, J. Zhang, J. Li, N. Zhang, R. R. Hong, X. Deng,P. Tang and D. Li, Sensor. Actuat. B-Chem., 241, 895–903 (2017).6) E. Rossinyol, A. Prim, E. Pellicer, J. Rodríguez, F.Peiró, A. Cornet, J. R. Morante, B. Tian, T. Bo and D.Zhao, Sensor. Actuat. B-Chem., 126, 18–23 (2007).7) M. Penza, C. Martucci and G. Cassano, Sensor. Actuat.B-Chem., 50, 52–59 (1998).8) M. Akiyama, J. Tamaki, N. Miura and N. Yamazoe,Chem. Lett., 9, 1611–1614 (1991).9) P. Ivanov, E. Llobet, F. Blanco, A. Vergara, X. Vilanova,I. Gracia, C. Cané and X. Correig, Sensor. Actuat. B-Chem., 118, 311–317 (2006).10) M. Gillet, K. Aguir, M. Bendahan and P. Mennini, ThinSolid Films, 484, 358–363 (2005).11) A. Labidi, E. Gillet, R. Delamare, M. Maaref and K.Aguir, Sensor. Actuat. B-Chem., 120, 338–345 (2006).12) S. Vallejos, V. Khatko, K. Aguir, K. A. Ngo, J. Calderer,I. Gràcia, C. Cané, E. Llobet and X. Correig, Sensor.Actuat. B-Chem., 126, 573–578 (2007).13) A. Ponzoni, E. Comini, G. Sberveglieri, J. Zhou, S. Z.Deng, N. S. Xu, Y. Ding and Z. L. Wang, Appl. Phys.Lett., 88, 203101 (2006).14) R. Ionescu, A. Hoel, C. G. Granqvist, E. Llobet and P.Heszler, Sensor. Actuat. B-Chem., 104, 132–139 (2005).15) C. S. Rout, M. Hegde and C. N. R. Rao, Sensor. Actuat.B-Chem., 128, 488–493 (2008).16) K. Kanda and T. Maekawa, Sensor. Actuat. B-Chem.,108, 97–101 (2005).17) J. Hubálek, K. Malysz, J. Prášek, X. Vilanova, P.Ivanov, E. Llobet, J. Brezmes, X. Correig and Z. Svěrák,Sensor. Actuat. B-Chem., 101, 277–283 (2004).18) G. K. Mani and J. B. B. Rayappan, Appl. Surf. Sci., 311,405–417 (2014).19) S. Ilican, M. Caglar and Y. Caglar, Appl. Surf. Sci., 256,7204–7210 (2010).20) G. Korotcenkov, Mat. Sci. Eng. B-Solid, 139, 1–23(2007).21) S. Fukami, M. Taguchi, Y. Adachi, I. Sakaguchi, K.Watanabe, T. Kinoshita, T. Muro, T. Matsushita, F.Matsui, H. Daimon and T. T. Suzuki, Phys. Rev. Appl.,7, 064029 (2017).22) Y. Adachi, K. Watanabe, N. Saito, I. Sakaguchi and T. T.Suzuki, J. Ceram. Soc. Jpn., 124, 668–672 (2016).23) Y. Adachi, N. Saito, I. Sakaguchi and T. T. Suzuki, ThinSolid Films, 685, 238–244 (2019).24) Y. Adachi, Surf. Interfaces, 28, 101597 (2022).25) S. Mridha and D. Basak, J. Phys. D Appl. Phys., 40,6902–6907 (2007).26) M. Righettoni, A. Tricoli and S. E. Pratsinis, Chem.Mater., 22, 3152–3157 (2010).27) P. M. Woodward, A. W. Sleight and T. Vogt, Solid StateChem., 131, 9–17 (1997).28) A. Šutka and K. A. Gross, Sensor. Actuat. B-Chem.,222, 95–105 (2016).Journal of the Ceramic Society of Japan 132 [5] 227-231 2024 JCS-Japan231