Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Electrochromic switching of tungsten oxide films grown by reactive ion-beam sputter deposition

Electrochromic switching of tungsten oxide films grown by reactive ion-beam sputter deposition J Mater Sci (2021) 56:615–628 ELECTRONIC MATERIALS Electronic materials Electrochromic switching of tungsten oxide films grown by reactive ion-beam sputter deposition 1,3, 1,3 2,3 2,3 1,3 Mario Gies * , Fabian Michel , Christian Lupo´ , Derck Schlettwein , Martin Becker , and 1,3 Angelika Polity Institute for Exp. Physics I, Justus-Liebig-University Giessen, 35392 Giessen, Germany Institute of Applied Physics, Justus-Liebig-University Giessen, 35392 Giessen, Germany Center for Materials Research (LaMa), Justus-Liebig-University Giessen, 35392 Giessen, Germany Received: 26 June 2020 ABSTRACT Accepted: 8 September 2020 Chromogenic thin films are crucial building blocks in smart windows to modulate Published online: the flux of visible light and heat radiation into buildings. Electrochromic 6 October 2020 materials such as tungsten oxide are well established in those devices. Sputter deposition offers a well-suited method for the production of such layers, which The Author(s) 2020 can also be used on an industrial scale. Tungsten oxide films were prepared by means of reactive ion-beam sputter deposition. The choice of distinct gas mix- tures as well as the growth temperature during the sputtering process allows to tune the properties of the resulting layers. Especially, the variation in the growth temperatures was found to have an impact on the structure of the resulting samples and, as a consequence, on their optical and electrochemical properties. By specific choice of the reactive gas, the deposition of colorless transparent as well as blue films of different composition is possible. The optical transmittance in the visible spectral range was up to 75% for as-deposited oxygen-rich layers. Additionally, hydrogen-doped tungsten oxide samples were grown. Superior electrochromic switching was observed for H -doped layers, probably by some kind of preconditioning. This resulted in a value for the standardized optical coloration efficiency of 26.5 cm /C. technologies for the conversion, storage and use of Introduction renewable energies are constantly growing [1]. The building sector plays an important role in terms of Because of the global climate change, energy-saving energy saving potential. For example, this sector and sustainable technologies are becoming more and consumes 23% of global primary energy and 30% of more important. Therefore, the demands on global electricity demand [2]. The further Handling Editor: Kevin Jones. Address correspondence to E-mail: Mario.Gies@exp1.physik.uni-giessen.de https://doi.org/10.1007/s10853-020-05321-y J Mater Sci (2021) 56:615–628 development of technologies that enable energy to be properties result from the different level of oxygen used efficiently and economically is, therefore, of deficit in phases of WO , which were investigated by ´ ´ foremost interest. In particular, the class of so-called Magneli and are thus known as Magneli phases smart windows offers an approach to save energy in [23, 24]. In addition to their EC characteristics, tung- the building sector by efficiently regulating incident sten oxides can be successfully applied in, e.g., gas light [3]. This can be achieved by the use of elec- sensors, catalysts or electrode materials in lithium-ion trochromic (EC) thin films. Electrochromism denotes batteries [25–29]. the reversible change of optical absorbance driven by The EC properties of tungsten oxide layers depend an externally applied voltage; thus, EC materials on the composition, the crystal structure and the allow for the modulation of the incident light. morphology [8, 30–32]. Compared to crystalline films, The mostly studied EC materials include conju- amorphous WO exhibits more rapid coloration and gated conducting polymers, transition metal oxides improved efficiency but inferior stability [30, 31]. as well as metal coordination complexes [4–9]. Due to Hence, using low-dimensional nanostructures has their diversity in composition and structure as well as been proven effective to achieve faster switching, their superior performance, electrochromism based increased reversibility and enhanced durability on thin film transition metal oxides has become [30, 32]. This is due to the higher surface accessibility increasingly important in the last decade. Exemplar- and shorter diffusion path length versus the bulk ily, there has been significant progress in terms of counterpart [30]. The film characteristics are strongly new materials [10–15] as well as novel designs and dependent on the growth technique. A multitude of concepts [14, 16–19]. Among the various transition different techniques were used, metal organic chem- metal oxides, tungsten oxide is the most intensively ical vapor deposition [33], the sol–gel process [34, 35], studied material and has been established as material pulsed laser deposition [36–38] or sputter deposition of choice in this field, especially with regard to its [39, 40], among others. Especially, sputter-deposited extraordinary EC properties like high color efficien- coatings show some intrinsic advantages, such as a cies, high cyclability and high environmental stabil- strong layer adhesion, a constant film thickness of ity, among others [20, 21]. The underlying process of homogeneous structure, high growth rates and a high electrochromism is widely accepted as a result of degree of reproducibility, which trigger their domi- simultaneous injection/extraction of electrons and nance in industrial scale. Nevertheless, there are cations [22] and complies with the following reaction drawbacks in conventional setups such as the contact þ  between the substrate and the plasma during depo- ½WO þ xM þ xe  ,½M WO  ; ð1Þ 3 x 3 bleached colored sition. Therefore, the temperature of the substrate where x describes the number of incorporated ions surface can rise significantly. In this article, ion-beam and electrons, respectively. The type of ions, such as sputter deposition (IBSD) was used for the synthesis þ þ þ þ H ,Li or Na , is represented by M . The interca- of tungsten oxide films to achieve a more precise lation reaction of electrons and ions into the film can control of the sample temperature and, thus, the layer 6þ 5þ 4þ cause a reduction of the W -states to W -orW - properties [41]. This favors the deposition of amor- states [6]. This leads to a modification of the elec- phous tungsten oxide films, since otherwise plasma tronic structure of the material, whereby photons in irradiation of the substrate surface raises the tem- the visible range of the electromagnetic spectrum can perature, promotes atom migration and, as a conse- be absorbed. Simultaneously, the color impression of quence, can lead to crystallization of the material. For the tungsten oxide layer changes. The untreated the same reason, films already present on the sub- tungsten oxide layer appears colorless transparent, strate, e.g., for stack preparation, can be spared the whereas a blue coloration occurs upon electrochem- harsh conditions of a contact to the plasma. Besides, ical reduction with simultaneous incorporation of smooth and compact surfaces can be realized, which charge-neutralizing ions. A reversal of this process, is in contrast to conventionally sputtered thin films i.e., the removal of ions (deintercalation), results in a being rather rough and porous. The basic character- decolorization of the layer. istics of ion-beam sputtered tungsten oxide coatings Tungsten and its oxides are, further, of great as well as the possibilities to improve the coating interest because of their high availability, low price properties due to variation in synthesis parameters factor and high chemical stability. The special like gas mixture and growth temperature will be J Mater Sci (2021) 56:615–628 617 described. As-grown samples are investigated by spectroscopy was carried out with a 633 nm laser in a means of their optical, compositional and EC range between 100 and 1500 cm . Additionally, a characteristics. colored sample was examined for analysis by Raman spectroscopy. This sample was intercalated with Li - ions by means of a cyclic voltammetry measurement. Experimental As electrolyte, a 1 M solved lithium perchlorate (LiClO ) in propylene carbonate was used. The Tungsten oxide (WO ) films were prepared by ion- potential was varied with a rate of 10 mV/s until beam sputter deposition. The sputtering setup con- - 0.7 V were reached. Further details about the sists of the vessel of radio frequency (RF) ion source, electrochemical measurement setup are described which incorporates the plasma, and the main process below. To gain more precise information about the chamber in which the sputter target and the substrate composition of the film and the chemical bonds, X- are mounted. The ion beam extracted from the ion ray photoelectron spectroscopy (XPS) was conducted. source is directed onto the target, where the material The PHI VersaProbe system utilizes an Al anode of the target is atomized. In comparison with con- (Al-K = 1486.6 eV). Measurements were taken with ventional sputtering systems such as direct current a source angle of 45 and with charge neutralization (DC) or RF sputtering, in IBSD systems obviously on the sample surface. All resulting spectra were there is no direct contact between plasma and sub- referenced to the carbon signal (C 1s) at 284.8 eV. strate and, as a consequence, the temperature of the Depth profiles of the films were studied via in situ substrate surface will not rise significantly. Further argon ion etching with an acceleration voltage of 0.5 details concerning the IBSD setup can be found in or 1 kV. Electron paramagnetic resonance (EPR) reference [41]. spectra were measured at 4 K with a microwave The films were deposited from a metallic tungsten power of 201,17 mW and a microwave frequency of target of Kurt J. Lesker Company (purity of 99.95%) 9.49 GHz. For these investigations, tungsten oxide by reactive sputtering in an argon–oxygen mixture. was deposited on polytetrafluoroethylene (PTFE) foil. For some layers, hydrogen was used as additional The use of PTFE foil allows to remove the deposited reactive gas to deposit hydrogen-doped films. All layer from the substrate so that the remaining pow- gases got a purity of 99.999%. A fixed argon flux of der can be examined in a quartz ampule. Selected 2 sccm was used for all samples. The ion source was samples were examined for their EC properties. operated at an RF power of 220 W. To influence the Cyclic voltammetry was conducted in an IviumStat degree of crystallization of the layers, films were potentiostat between - 0.7 and 1.5 V with potential deposited under ambient as well as elevated growth sweeps of 10 mV/s. The electrochemical half-cell TM temperatures. K Glass coated with fluorine-doped consisted of a platinum wire (counter electrode), a tin oxide (FTO) with a thickness of 300 nm was used leak-free Ag/AgCl reference electrode and a tungsten as substrate in order to establish an electrical contact oxide layer (working electrode). The electrolyte used for electrochemical measurements. was 0.1 M sulfuric acid. During the cyclic voltam- X-ray diffraction (XRD) was carried out with a metry measurements, UV–Vis spectroscopy was diffractometer type D5000 of Siemens company using simultaneously performed with a TEC5 spectrometer the Cu-K emission line. The measurements were at the reversal points of the potential. These reversal taken in Bragg–Brentano geometry with a rate of 2 / points can be considered as the state of the interca- min. A Lambda 900 spectrometer from PerkinElmer lated and deintercalated layer. Due to the sweep rate Instruments was used to measure the optical prop- of the potential, this would result in a switching time erties of the layers. Scanning electron microscopy of 220 s. For the classification of the optical switching (SEM) and atomic force microscopy (AFM) were performance of the investigated layers, the elec- conducted to analyze the crystallite surface structure. trochromic characteristics listed in the following were The SEM measurements were taken using a Zeiss– used. For building glazing, the European standard Merlin setup. For the AFM investigations in air, a EN 4102 [42] serves as the basis for calculating the Smart SPM 1000 (AIST-NT) was used, utilizing optical characteristics. The standardized transmission NanoWorld Pointprobe SEIHR-20 AFM probes T is of special importance with respect to the use as vis designed for non-contact mode imaging. Raman J Mater Sci (2021) 56:615–628 window coating. This takes into consideration the Results and discussion spectral sensitivity of the human eye V(k). The cor- Figure 1a shows the transmittance in dependence on responding values can be found in reference [43]. wavelength for pristine tungsten oxide films depos- Here the spectral range from 380 to 780 nm is ited at room temperature and varied O flux up to examined, with the strongest weighting at about 2 10 sccm. The corresponding photographs are shown 555 nm. Furthermore, the intensity distribution of the as well with the O flux during the deposition incident light is taken into account. For the calcula- decreasing from left to right. Obviously, layers pro- tion, the CIE (International Commission on Illumi- duced under high O flux are highly transparent. nation) Standard Illuminant D65 is used as the Below a certain threshold in the O flux, however, a reference light source. This can be found in the ISO blue coloration of the layers occurs. This points at 11664-2 standard [44]. From these values, the stan- 5þ 4þ dardized transmission T is calculated as follows: partial presence of W - and W -states and, hence, vis an oxygen deficit of the samples. 780 nm D65ðkÞ VðkÞ TðkÞ k¼380 nm T ¼ : ð2Þ P Taking a look at the optical transmission of the vis 780 nm D65ðkÞ VðkÞ k¼380 nm layers in the spectral range between 300 and 650 nm, this tendency is also well established. As opposed a This results in the standardized optical coloration measurement of the substrate (FTO layer on glass) for efficiency: which absorption occurs up to 310 nm, the films of T ðbÞ vis log tungsten oxide absorb up to about 370 nm, well in T ðcÞ vis CE ¼ ; ð3Þ vis line with their colored appearance. Oxygen-rich DQ tungsten oxide samples deposited with an O flux of in which the bleached (b) and colored (c) state of the 6 sccm and above reach a transmittance in the visible standardized transmission and the charge involved is spectral range of up to 75%, close to the 80% of the taken into account. substrate. The general transmission profile of these Tungsten oxide thin films were grown by ion-beam samples is comparable. However, for samples pro- sputter deposition, a less common sputtering variant. duced under lower O flux a decrease in the mea- We showed the possibility of influencing technolog- sured transmittance can be seen. In particular, an O ically relevant samples characteristics by using dif- flux of 5.15 sccm was found to be the lowest O flux ferent preparation parameters (e.g., gas mixture or which reproducibly yields widely colorless trans- growth temperature). This allows to tune the ele- parent samples. Additionally, the sample shows the mental composition, optical properties or to influence highest transmittance for the violet and blue spectral the structure and the degree of crystallization in the range. Samples grown with higher oxygen deficit resulting thin films. Variation in these properties (deposited at an O flux of 5 sccm and less) show a allows for a positive selection of parameters guaran- bluish color impression accompanied by a shift in the teeing layers with beneficial EC characteristics. onset of absorption toward higher wavelengths and Exemplarily, layers of a significantly altered mor- increased absorption in the longer wavelength range. phology and a much more compact structure can be Figure 1b depicts the transmittance series for films produced. Additional in operando doping with prepared at a constant substrate temperature of hydrogen allows to further optimize the cycling 400 C. Again, the upper part of the figure shows behavior. The high reproducibility as well as the high photographs of the samples. In general, the trans- purity of IBSD-grown layers render ion-beam sputter mission measurements of these samples also show deposition a suitable candidate for growth of tung- the decrease in the transmittance for oxygen-poor sten oxide and, most likely, other chromogenic samples as well as the shift of the onset of absorption. materials. Compared to the previous series, however, the O flux below which blue coloring occurs shifts to higher values. This is clearly visible for the sample deposited under an O flux of 5.15 sccm. Here, the sample grown at a growth temperature of 400 C exhibits a dark blue color impression, while the one deposited J Mater Sci (2021) 56:615–628 619 (a) (b) (c) 10 sccm 8 sccm 6 sccm 5.15 sccm 5 sccm 4 sccm 10 sccm 8 sccm 6 sccm 5.15 sccm 5 sccm intercalation 2.5 cm 2.5 cm @ RT @ 400°C 80 Ref. FTO 80 Ref. FTO O flux O flux 2 2 60 60 10.00 sccm 10.00 sccm 8.00 sccm 8.00 sccm 40 6.00 sccm 40 6.00 sccm 5.15 sccm 5.15 sccm 5.00 sccm 5.00 sccm 20 20 1 cm 4.00 sccm 300 400 500 600 300 400 500 600 deintercalation λ (nm) λ (nm) Figure 1 Optical transmittance curves and photographs of coloration of tungsten oxide when reduced in an electrochemical tungsten oxide samples IBSD-grown at room temperature (a) cell at a potential of - 0.7 V versus Ag/AgCl (intercalated state) and 400 C(b) under variation of the O flux superimposed to the or 1.5 V versus Ag/AgCl (deintercalated state) in contact to 0.1 M fixed 2 sccm of Ar. The bluish coloration of some samples stems sulfuric acid (c). from tungsten excess and is not to be confused with the bluish at RT was colorless. Thus, different compositions can (002), (020) and (200) reflexes might be present [47]. It be obtained for samples deposited under the same O should be noted that these signals might be as well flux but at different growth temperatures. Conse- assigned to the presence of a triclinic phase [48]. quently, the different color impression allows an Furthermore, Fig. 2a comprises an additional marked assessment of the composition of the films. This is in area at 45 –55 with two broad signals for which a accordance with the literature, where a ratio of O/ clear assignment is not possible and a superposition W [ 2.5 is reported to show a transparent widely of different reflexes is most likely. Additionally, colorless aspect, while for O/W ’ 2.5 a bluish color around the (101) reflection of FTO further low-in- and for O/W \ 2.5 a metallic character of the films tensity signals seem to be overlayed, apparent in a was observed [45]. It should be noted that the col- significant broadening of the FTO-related signal. oration discussed in the previous paragraph is not to Again, reflections of the monoclinic tungsten oxide be mistaken for the coloration which results from de- are the obvious cause. In conclusion, it is seen that and intercalation processes, cf. Fig. 1c. polycrystalline tungsten oxide films were prepared in Figure 2a shows X-ray diffractograms of samples the 400 C-series and an influence of the chosen O deposited under different growth temperatures and flux during the deposition can be seen. The highest with a varying composition. Herein, the films grown degree of crystallinity is observed for the sample at RT without additional heating of the substrate do sputtered at an O flux of 8 sccm. not reveal any reflexes besides those to be assigned of The morphology of the layers was investigated in SnO (marked in the upper scale [46]), which belong SEM and AFM measurements. Smooth and compact to the FTO substrate. Thus, no crystalline WO was films of WO were realized. In Fig. 2b–e, SEM images 3 3 grown at room temperature. In order to be able to of films deposited under different O flux and with- produce crystalline samples, an increase in the out additional heating of the substrate during depo- growth temperature is necessary. Accordingly, the sition are shown. For O flux up to 8 sccm, compact samples deposited at a growth temperature of 400 C films without any pronounced grain structure are show reflexes of WO phases. A dominant signal formed. The sample deposited at an O flux of 3 2 arises in the region of 20 –25 . However, since the 10 sccm shows a similar morphology with some reflection appears rather broad, a clear assignment to protrusions of 0.2–0.3 lm size without, however, any a single lattice plane is not possible. A comparison distinct orientation, in a line with the findings of X- with the predicted reflexes for monoclinic WO (blue ray diffraction. In Fig. 2f, the influence of the growth squares in Fig. 2a) indicates that a superposition of temperature on the surface structure at a fixed O T(%) T (%) J Mater Sci (2021) 56:615–628 (a) (110) (101) (200) (211) (310) (301) SnO 4 sccm 6 sccm 5.15 sccm (b) (c) (f) (g) T = 25°C T = 400°C growth growth f = 5 sccm f = 5 sccm (i) O2 O2 f = 8 sccm f = 8 sccm O2 O2 f = 10 sccm f = 10 sccm O2 O2 8 sccm 10 sccm (d) (e) (h) WO (m) 1 µm 20 30 40 50 60 70 80 2Θ (°) Figure 2 X-ray diffractograms of samples of varying cf. (f) and AFM images of samples, deposited at room temperature composition, deposited at RT (green) or at 400 C (red) under a moderate O flux of 5.15 sccm (g) and under oxygen-poor compared to reference values of SnO (red bars on top) and conditions (h). Compared to the surface of a sample grown at monoclinic WO (shown by blue squares) (a). SEM images of 400 C(i), the surface roughness is significantly smoother. films deposited under different O flux at RT (b)–(e) or at 400 C, flux of 5.15 sccm during deposition is shown. A T = 25°C T = 400°C growth growth morphology different from the samples shown in f = 5 sccm f = 5 sccm O2 O2 f = 5.15 sccm Fig. 2b–e should be emphasized. Unlike those sam- O2 f = 5.15 sccm, int. O2 ples, the film in Fig. 2f shows well-separated grains f = 10 sccm f = 10 sccm O2 O2 with a size up to 0.3 lm but less pronounced crystal 5+ 6+ W -O 6+ W -O W =O facets. In Fig. 2g, the surface of a sample deposited at 5+ RT and a moderate O flux of 5.15 sccm is shown as W =O analyzed by AFM. Individual grains of about 0.2 lm size appear interconnected without sharply defined grain boundaries. The root-mean-square surface roughness was determined to be around 9 nm. In Lee 1998 [43] comparison, Fig. 2h shows the morphology of a Salje 1975 [41] sample synthesized at RT under oxygen-poor condi- tions. Again, no sharply defined grains are recog- nizable. However, the grains seem to be a bit more extended. The determined roughness of the surface is approximately 7 nm. At an increased deposition 200 400 600 800 1000 1200 1400 -1 temperature of 400 C, larger round-shaped grains of Raman-shift (cm ) about 0.5 lm lateral expansion were obtained, cf. Figure 3 Raman spectra of tungsten oxide thin films deposited at Fig. 2i, leading to an increased roughness of around different growth temperatures and O fluxes as well as a spectrum 20 nm, much higher than for the unheated samples. of a Li -intercalated sample (blue) with band assignments Figure 3 shows Raman spectra obtained for sam- indicated as vertical lines. Bluish WO layers display a higher ples with different optical impression (colorless intensity especially in the range of about 330 and 450 cm transparent or blue) as well as for samples synthe- 5þ 5þ attributed to W –O and W =O bonds, whereas in colorless sized at different growth temperature. The amor- 6þ transparent samples W -related Raman modes dominate. phous samples deposited under an O flux 5.15 sccm at room temperature show two broad 808 cm [50]. The second peak resulting from the signals of higher intensity in the region of 770 and 6þ W =O stretching mode of terminal oxygen atoms 950 cm . According to the literature, the first one is [51]. With respect to amorphous tungsten oxide, the 6þ due to the W –O-bond [49] and is caused by a basic structure can be described by the formation of superposition of the two strongest peaks at 719 and 6þ WO -octahedra. These consist of short W =O-bonds intensity (arb. units) intensity (arb. units) J Mater Sci (2021) 56:615–628 621 6þ signal consisting of two contributions at binding and longer W –O-bonds leading to clusters of 6þ energies of 530.3 and 531.6 eV, cf. Fig. 4b. The former deformed octahedra, where the W =O-bonds are 2 6þ has been associated with O and W ions, whereas expected at the outermost surface of the clusters the latter arises from contamination at the surface, [51, 52]. The bluish samples (oxygen poor and Li - such as hydroxyl groups or oxygen as part of carbon- intercalated) yield spectra with a weaker intensity in 6þ related impurities. the region of the W -states. The colorless transpar- Figure 4c shows the O 1s signal after an etching ent sample (deposited under an O flux of 5.15 sccm) time of 300 s upon which additional signals repre- shows a significant increase in intensity in the range 5þ 1 þ senting oxygen bound to W (signal at 530.8 eV) and of about 330–450 cm after the intercalation of Li - 4þ 5þ W (signal at 530.6 eV). A detailed spectrum of the ions. These signals are assigned to the W –O and 5þ W4f core level at the surface is given in Fig. 4d W =O bonds [52]. Following the model of small 5þ consisting of W 4f (at 35.4 eV) and W 4f (at 7=2 5=2 polaron transitions by Schirmer et al. [53], the W 37.5 eV), both associated with the oxidation state states contribute to the mechanism that leads to 6þ W . After an etching time of 300 s (Fig. 4e) further optical absorption. The Raman measurements of the contributions have to be considered pointing at the films therefore give an indication of the presence of 5þ 4þ these states in the stained samples. presence of W and W oxidation states, which Crystalline WO films showed three Raman signals confirm the assignment in the corresponding O 1s 1 1 region. Obviously, the surface of the sample only with a high intensity at about 270 cm , 690 cm 1 6þ contains W oxidation states. With increasing depth and 810 cm , slightly shifted relative to the reported 5þ 4þ 1 1 1 of layer, however, increasingly W and W oxida- positions 275 cm , 719 cm and 808 cm [49, 50]. tion states are present. Due to the fact that the sample Beside those modes, both samples show a higher considered is an oxygen-rich, colorless transparent intensity in the signal around 330 cm . At this wave 5þ tungsten oxide layer, a high proportion of W and number, Ozkan et al. mention the existence of a mode 4þ which occurs for crystalline samples [54]. In contrast W states, however, is unlikely. Rather, a preferen- tial oxygen etching during the depth profiling is to the amorphous films, the mode at 950 cm only assumed, which does not allow an exact determina- appears as a weak signal and predominantly for the tion of the concentration ratio by means of XPS. film prepared at low O flux. Tentatively, we assign 6þ Hence, we discuss the data in terms of a trend of the this to a removal of doubly bonded W for crys- W/O-ratio within the series. Figure 4f shows the W/ talline tungsten oxide. Thus, the intensity of the O-ratio of an oxygen-poor (blue film), a slightly 950 cm Raman mode might serve as a measure of substoichiometric (deposited under 5.15 sccm and at the degree of crystallization. In conclusion, a better RT or with a growth temperature of 400 C) and an crystallization is observed for the oxygen-rich oxygen-rich sample (colorless transparent film), pre- sample. sented in dependence of the O flux during deposi- Figure 4a presents the EPR spectra of a colorless tion. In each case, the measurement of the surface transparent as well as a bluish sample compared to a (squares) and after an etching step (triangles) is PTFE film as reference. For all examined samples, a shown. The surface of the sample deposited under signal can be seen at approximately H = 3370 G and the lowest O flux provides a nearly stoichiometric ´ 2 H = 3388 G. For this, a Lande factor of g ¼ 2:00 W/O ratio of about 0.31. However, due to the low results which is characteristic for free electrons. Two oxygen content during the deposition process, an more resonances are seen for the bluish sample at oxygen deficit can be assumed and the resulting ratio H = 3830 G and H = 4412 G. For those signals, 3 4 is due to surface contamination of the layer. In con- Lande´ factors of g ¼ 1:77 and g ¼ 1:54 are calculated. trast, an etching time for 300 s with 1 kV acceleration These values are in good agreement with those given 5þ voltages results in a W/O ratio of about 0.58. How- in the literature [51], assigned to W ions in colored ever, such a strong substoichiometry of oxygen is 5þ tungsten oxide films confirming the presence of W more than doubtful since for this ratio, a metallic states in our bluish samples. character would be expected. Comparable trends are XPS measurements on an oxygen-rich sample (de- evident for the two samples prepared without posited with an O flux of 10 sccm) show an O 1s growth temperatures and with an O flux of 5.15 and 2 J Mater Sci (2021) 56:615–628 Figure 4 EPR spectra of a (a) transparent (deposited with an H as-deposited O flux of 5.5 sccm at RT) as intercalated PTFE well as a bluish sample (deposited with an O flux of 5þ 3 sccm at RT) (a). W oxidation states are only present for blue samples. O 1s (b, c) and W 4f core level spectra (d, e) measured by XPS at the surface of a film 2000 3000 4000 5000 6000 prepared at RT and with an O H(G) flux of 10 sccm (b, d) as well as after etching the surface for (b) (c) 300 s at 1 kV (c, e). Features 1.0x10 attributed to oxidation states 1.4x10 6+ O(W ) 4þ 5þ W and W arise after the 1.2x10 6+ O (W ) etching procedure. Calculated 8.0x10 5+ O (W ) W/O-ratio (f) of the film 4 1.0x10 surface or during etching at 3 4+ 8.0x10 6.0x10 acceleration voltages of O (W ) OH 0.5 kV (open triangles) or 3 OH 6.0x10 1 kV (solid triangles) for 4.0x10 samples prepared at different 4.0x10 O flux at RT compared to a 540 535 530 525 540 535 530 film prepared at 400 C. binding energy (eV) binding energy (eV) 6þ Corresponding W / 5þ 4þ (W ? W ) ratios (d) (e) (g) resulting from the depth 6+ 4 4 5+ W 4f 6+ 1.0x10 1.0x10 7/2 W 4f W 4f 7/2 7/2 profile of the oxygen-rich and 6+ 5+ 3 W 4f 3 W 4f crystalline samples. 5/2 8.0x10 8.0x10 5/2 6+ 4+ W 4f W 4f 3 3 5/2 7/2 6.0x10 6.0x10 4+ W 4f 5/2 3 3 4.0x10 4.0x10 3 3 2.0x10 2.0x10 0.0 0.0 45 40 35 30 25 45 40 35 30 25 binding energy (eV) binding energy (eV) surface surface (f) (g) etching 1.0 kV for 300 s etching 1.0 kV for 300 s T 70 etching 0.5 kV etching 0.5 kV 0.6 growth surface T surface T growth growth etching T etching T growth growth 0.5 etching 0.4 etching etching 0.3 1 growth 0.2 0 2468 10 2.5 5.0 7.5 10.0 O flux [sccm] O flux (sccm) 2 2 intensity (cps) intensity (cps) ratio W/O intensity (arb. units) intensity (cps) intensity (cps) 6+ 4+ 5+ W /(W +W ) J Mater Sci (2021) 56:615–628 623 10 sccm. As expected, the examination of the surfaces influence of preferential oxygen sputtering is less shows that a higher selected O flux during the pronounced here due to the less altered W/O ratio of deposition process counteracts the deficient incorpo- 0.62 for depth measurement (red triangle). ration of O . However, on the basis of the series Figure 4g illustrates the ratio of the oxidation states 6þ 5þ 4þ shown, only a tendency of the decreasing ratio can be W /(W ? W ) for the measurements already seen. It remains to be mentioned that this results in a discussed in Fig. 4f. Only the thin film produced slight superstoichiometry of oxygen. Again, it is very under the highest O flux and the crystalline sample likely that this ratio is due to an oxygen-rich con- are shown. For both samples, the measurement of the tamination of the sample surface from ambient air. surface and the depth measurement after an etching Compared to the measurements of the strongly time of 300 s (with 1 kV acceleration voltages) is oxygen-deficient sample, the depth measurements of depicted. For the oxygen-rich sample, the additional the other two amorphous layers show comparable depth profile was measured with the lower configu- trends. Once more, conditions which would corre- ration of the acceleration voltage of 0.5 kV and after spond to a metallic character have to be determined. time steps of 180, 360 and 540 s. The corresponding 6þ 5þ 4þ A comparison of the trend of the conditions within W /(W ? W ) ratio decreases during depth the data points for the depth profiles provides an profiling. This is due to the preferential removal of analogy in the development of the surface measure- oxygen. With increasing depth, the ratio saturates. ments of the samples. For the depth measurements, We tentatively explain this as follows: At the begin- the W/O ratios are displaced to a higher W/O ratio. ning of the measurement (within the first 180 s), the Although a direct influence of the oxygen content 6þ 5þ W oxidation states are reduced to W oxidation depending on the chosen O flux during deposition 6þ states. Thereafter, in addition to the reduction of W can be seen, a quantitative analysis is hindered. To 5þ oxidation levels also W oxidation states are determine the influence of preferential etching dur- 4þ reduced and thus the presence of W oxidation ing depth profiling, detailed XPS-measurements were levels increases. Thus, compared to the first 180 s of taken under lower acceleration voltages of 0.5 kV for 6þ the etching time, the change of the W / the argon ions. Measurements were taken after an 5þ 4þ (W ? W ) ratio decreases again. In contrast, it is etching time of 180, 360 and 540 s for an oxygen-rich clear for the crystalline sample that the determined sample and are represented in Fig. 4f as open trian- oxidation states during the measurement result in a gles. After an etching time of 180 s, only a slight nearly non-changing ratio. Thus, the crystalline layer substoichiometry with an W/O ratio of 0.34 is found. shows, compared to the results of the amorphous Further etching resulted in a ratio of 0.4. Despite samples, a generally deviating behavior with respect enhanced substoichiometry of oxygen in the sample to the XPS measurements. The reason for this could for further etching steps, the W/O ratio remains be the presence of more stable bonds to oxygen or the below 0.47, which was measured after an etching concentration ratios being maintained despite etching time of 300 s at an acceleration voltage of 1 kV. The during the data acquisition. measurements suggest a less pronounced preferential The IBSD grown amorphous as well as crystalline etching of oxygen at this lower acceleration voltage of tungsten oxide layers with varying ratios of tungsten 0.5 kV, however, presumably also reduced erosion oxidation states were examined by means of cyclic rates of the layer. voltammetry regarding their EC properties. Figure 5a Furthermore, the figure shows the results of XPS shows nine cycles of the amorphous, colorless measurements on a crystalline film with blue color transparent sample deposited under an oxygen flux (deposited under an O flux of 5.15 sccm and at a of 5.15 sccm. For the reduction in the film, accom- growth temperature of 400 C, red symbols). Com- panied by H -intercalation (eq. 1), the current den- pared to the above mentioned samples, a clear devi- sities become increasingly negative starting at a ation of the O/W ratio of 0.54 at the surface is voltage of about 0.1 V and rapidly becoming more obvious (red square). Although understoichiometry negative at about - 0.48 V. At the reversal point of can be expected due to the coloration, the result - 0.7 V, the recorded current densities finally reach a appears to be too strongly understoichiometric, as the maximum negative value of about 0.14 mA=cm . metallic character of the layer can be assumed for this Peaks for individual reduction reactions (such as ratio. Simultaneously, it can be observed that the J Mater Sci (2021) 56:615–628 (a) (b) (c) (d) -5 t -5 t 5x10 5x10 -0.7 V / 1.5 V -0.7 V / 1.5 V 0 0 +1.5 V cycle 1 cycle 1 +1.5 V cycle 2 cycle 2 cycle 3 cycle 1 cycle 3 cycle 1 -5 -5 cycle 2 -5x10 -5x10 cycle 4 cycle 2 cycle 4 cycle 3 cycle 3 cycle 5 cycle 4 60 cycle 5 cycle 4 cycle 5 cycle 6 -0.7 V 60 cycle 5 cycle 6 cycle 6 -4 -4 cycle 6 cycle 7 -1x10 -0.7 V -1x10 cycle 7 cycle 7 cycle 7 cycle 8 cycle 8 cycle 8 cycle 9 cycle 8 cycle 9 cycle 9 40 -4 cycle 9 -4 -2x10 -2x10 -0.75 0.00 0.75 1.50 500 750 1000 -0.75 0.00 0.75 1.50 500 750 1000 potential (V) potential (V) λ (nm) λ (nm) Figure 5 Cyclic voltammetry (cycles proceeding from lighter to (b) and (d), respectively, measured at the reversal points - 0.7 to darker color in the plots) for an amorphous (a) and a crystalline 1.5 V. The calculated optical impression at the reversal points [55] sample (c) with the corresponding optical transmission spectra is shown above the transmission spectra (b, d). 6þ 5þ 5þ 4þ more steeply than for the remaining cycles and two W ! W or W ! W ) cannot be assigned. distinct features at approximately - 0.3 and - 0.1 V Upon reversal of the voltage sweep, only a broad occur re-oxidation/deintercalation. As for the amor- oxidation peak is seen. Accordingly, the assignment 4þ 5þ phous film, the charge increases upon successive of single oxidation steps (like W ! W or 5þ 6þ cycling. W ! W ) is not possible. The maximum current 2 2 Figure 5d shows the transmittance and the optical densities of about 2:4  10 mA=cm occur around a color impression for subsequent cycles. Obviously, voltage of - 0.45 V. Toward higher voltages, the the switching process is not completely reversible. current densities decay quite quickly, indicating Particularly in the near infrared range, a proceeding complete oxidation of the film. Based on the contin- drop in transmittance occurs from cycle to cycle. uously increasing current densities upon successive Both types of samples can be reduced/intercalated cycling of the potential, the enclosed area of the and re-oxidized/ deintercalated. For the crystalline voltammogram slightly increases with the increasing sample, a certain irreversibility reveals caused by number of cycles, which suggests that an increasing incomplete deintercalation. The irreversible switch- charge is involved in the process. Consequently, the ing behavior could be caused by a sweep rate of the film is reduced and re-oxidized to an increasing potential which was too fast. The more densely extent. packed atomic structure in crystalline samples could The optical transmission of the sample at each hinder the ion diffusion. As already mentioned in reversal point of the potential, cf. Fig. 5b, represents Sect. 1, an extended response time of the coloration in the bleached (at ? 1.5 V) and colored (at - 0.7 V) crystalline samples is a well-known phenomenon states. Rather small differences of transmission of the [30]. Thus, inserting protons gradually change the as- bleached and colored state can be seen, however, grown crystalline samples. reversibly within every single cycle and increases To suppress such behavior, an additional series of with the increasing number of cycles, caused by samples was grown in which hydrogen was used as increasing coloration. This is underlined by the additional reactive gas in the sputtering process. It is optical color impression of the film in the different expected that hydrogen is incorporated into the (de)intercalated conditions at the reversal points deposited layers and influences the structure and, (- 0.7 V or 1.5 V) shown in Fig. 5b. consequently, to show an influence on the EC Figure 5c shows nine cycles of cyclic voltammetry properties. on a crystalline WO sample. The characteristics of Figure 6a illustrates photographs of the samples the first cycle differ slightly from those of the subse- (above) as well as the corresponding optical trans- quent cycles, in that the current density during mission measurements (below). For this series, the reduction/intercalation of the first cycle increases current density (A/cm ) T(%) current density (A/cm ) T(%) J Mater Sci (2021) 56:615–628 625 (a) (b) (c) 0 sccm 0.5 sccm 0.65 sccm 0.8 sccm 1.5 sccm 2 sccm 2.5 cm 0 -0.7 V / 1.5 V cycle 1 +1.5 V 80 Ref. FTO Cycle 1 cycle 2 -4 Cycle 2 -1x10 H flux Cycle 3 2 cycle 3 Cycle 4 0.00 sccm cycle 4 0.50 sccm 0.65 sccm -4 -2x10 0.80 sccm -0.7 V 1.50 sccm 2.00 sccm -4 -3x10 300 400 500 600 -1.0 -0.5 0.0 0.5 1.0 1.5 500 750 1000 λ (nm) potential (V) λ (nm) Figure 6 Optical transmission spectra and photographs (top) of points (c) measured at a colorless transparent hydrogen-doped tungsten oxide films deposited at RT with fluxes of 2 sccm Ar and WO film (fluxes of 2 sccm Ar, 7 sccm O and 10 sccm H ) with x 2 2 5.15 sccm O , and increasing additional flux of H (a). Cyclic the color impression of the film in the current switching state 2 2 voltammetry (b) as well as optical transmission at the reversal (above the graph). additional H flux was varied between 0.5 and To compare the electrochromic characteristics of 2 sccm, whereas the O flux of 5.15 sccm and the IBSD grown films, Table 1 summarizes the results of growth temperature (RT) were fixed. Layers which an undoped tungsten oxide sample (measurement is were produced with a hydrogen flux of 0.8 sccm or shown in Fig. 5a, b) as well as the high doped sample above show a blue coloration indicative of partial from Fig. 6b, c. A comparison shows a clear increase reduction. Increasing the H flux even further in the change of the standardized transmission at strengthens the coloring. The absorption edge of all visible range. For the H -doped sample, this change samples starts at about 320 nm. In order to prevent is about twice as large. At the same time, compared to coloration of samples due to the deficient oxygen the undoped sample, a larger charge quantity of content, the O flux was increased to 7 sccm. At the 2.3 mC/cm is involved within the cycle under con- same time, an H flux of 10 sccm was chosen to 2 sideration. An improved value of 26.5 cm /C is also achieve a high doping. This resulted in a colorless obtained for the standardized optical coloration effi- transparent layer. ciency of the doped sample. Figure 6b shows four cyclic voltammetry cycles of such a hydrogen-doped sample. Essentially, the voltammograms resemble those shown in Fig. 5a. Conclusion However, the coloration current density is more Tungsten oxide thin films were grown by ion-beam negative. Quite constant current density and charge sputter deposition, a less common sputtering variant. are observed for all cycles. The transmittance spectra We showed the possibility of influencing technolog- recorded simultaneously and the optical color ically relevant samples characteristics by using dif- impression of the layer, cf. Fig. 6c, show a widely ferent preparation parameters (e.g., gas mixture or reversible switching process of high contrast in growth temperature). This allows to tune the ele- transmission. For the reduced/intercalated layer, the mental composition, optical properties or to influence transmittance reversibly drops to values as low as the structure and the degree of crystallization in the 50% at 800 nm as opposed to the irreversible char- resulting thin films. Variation in these properties acteristics of the samples reported in Fig. 5. It can, allows for a positive selection of parameters guaran- therefore, be proven that hydrogen, used as an teeing layers with beneficial EC characteristics. additional reactive gas in the manufacturing process, Exemplarily, layers of a significantly altered mor- yields samples which have significantly improved EC phology and a much more compact structure can be switching properties. T (%) current density (A/cm ) T (%) J Mater Sci (2021) 56:615–628 Table 1 EC characteristics of EC characteristics an undoped tungsten oxide layer (fluxes of 2 sccm Ar and 2 2 Sample DT (%) DQ (mC/cm)CE (cm /C) vis vis 5.15 sccm O ) and an H - 2 2 doped film (fluxes of 2 sccm Amorphous-undoped 5.3 1.3 22.2 Ar, 7 sccm O and 10 sccm 2 H -doped 10.7 2.3 26.5 H ), deposited under ambient For both the characteristics after the fourth cycle is given growth temperatures produced. Additional in operando doping with References hydrogen allows to further optimize the cycling [1] Ahmadi MH, Ghazvini M, Nazari MA, Ahmadi MA, Pour- behavior. The high reproducibility as well as the high fayaz F, Lorenzini G, Ming T (2019) Renewable energy purity of IBSD-grown layers render ion-beam sputter harvesting with the application of nanotechnology: a review. deposition a suitable candidate for growth of tung- Int J Energy Res 43:1387–1410 sten oxide and, most likely, other chromogenic [2] Urge-Vorsatz D, Cabeza LF, Serrano S, Barreneche C, Pet- materials. richenko K (2015) Heating and cooling energy trends and drivers in buildings. Renew Sustain Energy Rev 41:85–98 [3] Ta¨llberg R, Jelle BP, Loonen R, Gao T, Hamdy M (2019) Acknowledgements Comparison of the energy saving potential of adaptive and We thank P. J. Klar and D. M. Hofmann for fruitful controllable smart windows: a state-of-the-art review and discussions and their constant support. The work of simulation studies of thermochromic, photochromic and M. Gies and F. Michel was funded by the DFG via the electrochromic technologies. Sol Energy Mater Sol C 200: RTG (Research Training Group) 2204 ‘‘Substitute 109828. https://doi.org/10.1016/J.SOLMAT.2019.02.041 Materials for Sustainable Energy Technologies’’. [4] Green S, Backholm J, Geore´n P, Granqvist CG, Niklasson GA (2009) Electrochromism in nickel oxide and tungsten oxide thin films: ion intercalation from different electrolytes. Funding Sol Energy Mater Sol Cells 93:2050–2055 [5] Mortimer RJ (2011) Electrochromic materials. Annu Rev Open Access funding enabled and organized by Mater Res 41:241–268 Projekt DEAL. [6] Granqvist CG (2014) Electrochromics for smart windows: oxide-based thin films and devices. Thin Solid Films Open Access This article is licensed under a Crea- 564:1–38 tive Commons Attribution 4.0 International License, [7] Lahav M, Boom ME (2018) Polypyridyl metallo-organic which permits use, sharing, adaptation, distribution assemblies for electrochromic applications. Adv Mater 30: and reproduction in any medium or format, as long 1706641. https://doi.org/10.1002/adma.201706641 as you give appropriate credit to the original [8] Wu W, Wang M, Ma J, Cao Y, Deng Y (2018) Elec- author(s) and the source, provide a link to the Crea- trochromic metal oxides: recent progress and prospect. Adv tive Commons licence, and indicate if changes were Electron Mater 4:1800185 made. The images or other third party material in this [9] Sadeghi K, Yoon J-Y, Seo J (2019) Chromogenic polymers article are included in the article’s Creative Commons and their packaging applications: a review. Polym Rev licence, unless indicated otherwise in a credit line to 60:442–492 the material. If material is not included in the article’s [10] Cai GF, Tu JP, Zhang J, Mai YJ, Lu Y, Gu CD, Wang XL Creative Commons licence and your intended use is (2012) An efficient route to a porous NiO/reduced graphene not permitted by statutory regulation or exceeds the oxide hybrid film with highly improved electrochromic permitted use, you will need to obtain permission properties. Nanoscale 4:5724–5730 directly from the copyright holder. To view a copy of [11] Cai G, Tu J, Zhou D, Li L, Zhang J, Wang X, Gu C (2014) this licence, visit http://creativecommons.org/licen Constructed TiO /NiO core/shell nanorod array for efficient ses/by/4.0/. electrochromic application. J Phys Chem C 118:6690–6696 J Mater Sci (2021) 56:615–628 627 [12] Zhang JH, Cai GF, Zhou D, Tang H, Wang XI, Gu CD, Tu photothermal catalytic performance. J Mater Chem A JP (2014) Co-doped NiO nanoflake array films with 4:5314–5322 enhanced electrochromic properties. J Mater Chem C [27] Solis JL, Saukko S, Kish L, Granqvist CG, Lantto V (2001) 2:7013–7021 Semiconductor gas sensors based on nanostructured tungsten [13] Zhang JH, Tu JP, Zhou D, Tang H, Li L, Wang XI, Gu CD oxide. Thin Solid Films 391:255–260 (2014) Hierarchical SnO @NiO core/shell nanoflake arrays [28] Yoon S, Jo C, Noh SY, Lee CW, Songc JH, Lee J (2011) as energy-saving electrochromic materials. J Mater Chem C Development of a high-performance anode for lithium ion 2:10409–10417 batteries using novel ordered mesoporous tungsten oxide [14] Kim J, Ong GK, Wang Y, LeBlanc G, Williams TE, Mattox materials with high electrical conductivity. Phys Chem Chem TM, Helms BA, Milliron DJ (2015) Nanocomposite archi- Phys 13:11060–11066 tecture for rapid, spectrally-selective electrochromic modu- [29] Yoon S, Woo S-G, Jung K-N, Song H (2014) Conductive lation of solar transmittance. Nano Lett 15:5574–5579 surface modification of cauliflower-like WO and its elec- [15] Zhou D, Shi F, Xie D, Wang DH, Xia XH, Wang XL, Gu trochemical properties for lithium-ion batteries. J Alloys CD, Tu JP (2016) Bi-functional Mo-doped WO nanowire Compd 613:187–192 array electrochromism-plus electrochemical energy storage. [30] Cong S, Geng F, Zhao Z (2016) Tungsten oxide materials for J Colloid Interface Sci 465:112–120 optoelectronic applications. Adv Mater 28:10518–10528 [16] Wei D, Scherer MR, Bower C, Andrew P, Ryhanen T, [31] Wang Y, Runnerstrom EL, Milliron DJ (2016) Switchable Steiner U (2012) A nanostructured electrochromic superca- materials for smart windows. Annu Rev Chem Biomol Eng pacitor. Nano Lett 12:1857–1862 7:283–304 [17] Llorde´s A, Garcia G, Gazquez J, Milliron DJ (2013) Tunable [32] Zhou D, Xie D, Xia X, Wang X, Gu C, Tu J (2017) All- near-infrared and visible-light transmittance in nanocrystal- solid-state electrochromic devices based on WO k NiO in-glass composites. Nature 500:323–326 films: material developments and future applications. Sci. [18] Cong S, Tian Y, Li Q, Zhao Z, Geng F (2014) Single-crys- China Chem 60:3–12 talline tungsten oxide quantum dots for fast pseudocapacitor [33] Cruz-Leal M, Goiz O, Cha´vez F, Pe´rez-Sa´nchez GF, Her- and electrochromic applications. Adv Mater 26:4260–4267 na´ndez-Como N, Santes V, Felipe C (2019) Study of the [19] Wang JL, Lu YR, Li HH, Liu JW, Yu SH (2017) Large area thermal annealing on structural and morphological properties co-assembly of nanowires for flexible transparent smart of high-porosity A-WO films synthesized by HFCVD. windows. J Am Chem Soc 139:9921–9926 Nanomaterials 9:1298. https://doi.org/10.3390/nano9091298 [20] Niklasson GA, Granqvist CG (2007) Electrochromics for [34] Cronin JP, Tarico DJ, Tonazzi JCL, Agrawal A, Kennedy SR smart windows: thin films of tungsten oxide and nickel (1993) Microstructure and properties of sol–gel deposited oxide, and devices based on these. J Mater Chem WO coatings for large area electrochromic windows. Sol 17:127–156 Energy Mater Sol C 29:371–386 [21] Mardare CC, Hassel AW (2019) Review on the versatility of [35] Wen-Cheun AB, Chan K-Y, Knipp V (2019) Effect of film tungsten oxide coatings. Phys Status Solidi A 216:190004- thickness on electrochromic performance of sol–gel depos- 7. https://doi.org/10.1002/pssa.201900047 ited tungsten oxide (WO ). Opt Mater 94:387–392 [22] Granqvist CG (2000) Electrochromic tungsten oxide films: [36] Rougier A, Portemer F, Que´de´ A, El Marssi M (1999) review of progress 1993–1998. Sol Energy Mater Sol C Characterization of pulsed laser deposited WO thin films for 60:201–262 electrochromic devices Appl. Appl Surf Sci 153:1–9 [23] Magne´li AA (1950) Structure of b-tungsten oxide. Nature [37] Dinca V, Suchea M (2019) Functional nanostructured inter- 165:356–357 faces for environmental and biomedical applications, vol [24] Magne´li G, Andersson B, Blomberg L, Kihlborg L (1952) 363. Elsevier, Amsterdam Identification of molybdenum and tungsten oxides by X-ray [38] Parshina LS, Novodvorsky OA, Khramova OD, Lotin AA, powder patterns. Anal Chem 24:1998–2000 Shchur PA (2019) PLD of thin WO films for solid-state [25] Zhang L, Wang H, Liu J, Zhang Q, Yan H (2020) Nonsto- electrochromic cells. J Phys Conf Ser 1164:012003. https://d ichiometric tungsten oxide: structure, synthesis, and appli- oi.org/10.1088/1742-6596/1164/1/012003 cations. J Mater Sci Mater Electron 31:861–873 [39] Berggren L, Niklasson GA (2003) Optical absorption and [26] Wang L, Wang Y, Cheng Y, Liu Z, Guo Q, Ha MN, Zhao Z durability of sputtered amorphous tungsten oxide films. (2016) Hydrogen-treated mesoporous WO as a reducing Solid State Ionics 165:51–58 agent of CO to fuels (CH and CH OH) with enhanced [40] Kumar KU, Bhat SD, Subrahmanyam A (2019) Elec- 2 4 3 trochromic device with magnetron sputtered tungsten oxide J Mater Sci (2021) 56:615–628 (WO ) and nafion membrane: performance with varying Raman and FTIR spectroscopies. Appl Surf Sci tungsten oxide thickness—a report. Mater Res Express 6:- 33–34:804–811 045513. https://doi.org/10.1088/2053-1591/aafef1 [50] Salje E (1975) Lattice dynamics of WO . Acta Crystallogr [41] Becker M, Gies M, Polity A, Chatterjee S, Klar PJ (2019) 31:360–363 Materials processing using radio-frequency ion-sources: ion- [51] Gabrusenoks JV, Cikmach PD, Lusis AR, Kleperis JJ, beam sputter-deposition and surface treatment. Rev Sci Ramans GM (1984) Electrochromic colour centres in Instrum 90:023901. https://doi.org/10.1063/1.5063976 amorphous tungsten trioxide thin films. Solid State Ionics [42] DIN EN 410:2011-04, Glas im Bauwesen - Bestimmung der 14:25–30 lichttechnischen und strahlungsphysikalischen Kenngro¨ßen [52] Lee SH, Cheong HM, Tracy CE, Mascarenhas A, Benson von Verglasungen DK, Deb SK (1999) Raman spectroscopic studies of elec- [43] DIN 5031-1:1982-03, Strahlungsphysik im optischen Bere- trochromic A-WO . Electrochim Acta 44:3111–3115 ich und Lichttechnik; Gro¨ßen, Formelzeichen und Einheiten [53] Schirmer OF (1980) Small polaron aspects of defects in der Strahlungsphysik oxide materials. J Phys Coll 41:479–484 [44] ISO 11664-2:2008-02, Colorimetry—part 2: CIE standard [54] Ozkan E, Lee S-H, Tracy CE, Pitts JR, Deb SK (2003) illuminants Comparison of electrochromic amorphous and crystalline [45] Gerard P, Deneuville A, Hollinger G, Minh Duc T (1977) tungsten oxide films. Sol Energy Mater Sol C 79:439–448 Color in ‘‘tungsten trioxide’’ thin films. J Appl Phys [55] Lupo´ C, Eberheim F, Schlettwein D (2020) Facile low- 48:4252–4255 temperature synthesis of nickel oxide by an internal com- [46] McCarthy G, Welton J (1989) X-ray diffraction data for bustion reaction for applications in electrochromic devices. SnO . An illustration of the new powder data evaluation J Mater Sci 55:14401–14414. https://doi.org/10.1007/s1085 methods. Powder Diffract 4:156–159 3-020-04995-8 [47] Grier D, McCarthy G (1991) North Dakota State University. Publisher’s Note Springer Nature remains neutral with ICDD Grant-in-Aid, Fargo, North Dakota, USA regard to jurisdictional claims in published maps and [48] Diehl R, Brandt G, Salje E (1978) The crystal structure of triclinic WO institutional affiliations. . Acta Crystallogr B 34:1105–1111 [49] Shigesato Y, Murayama A, Kamimori T, Matsuhiro K (1988) Characterization of evaporated amorphous WO films by http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Materials Science Springer Journals

Electrochromic switching of tungsten oxide films grown by reactive ion-beam sputter deposition

Loading next page...
 
/lp/springer-journals/electrochromic-switching-of-tungsten-oxide-films-grown-by-reactive-ion-GuEzTRxtc7

References (57)

Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2020
ISSN
0022-2461
eISSN
1573-4803
DOI
10.1007/s10853-020-05321-y
Publisher site
See Article on Publisher Site

Abstract

J Mater Sci (2021) 56:615–628 ELECTRONIC MATERIALS Electronic materials Electrochromic switching of tungsten oxide films grown by reactive ion-beam sputter deposition 1,3, 1,3 2,3 2,3 1,3 Mario Gies * , Fabian Michel , Christian Lupo´ , Derck Schlettwein , Martin Becker , and 1,3 Angelika Polity Institute for Exp. Physics I, Justus-Liebig-University Giessen, 35392 Giessen, Germany Institute of Applied Physics, Justus-Liebig-University Giessen, 35392 Giessen, Germany Center for Materials Research (LaMa), Justus-Liebig-University Giessen, 35392 Giessen, Germany Received: 26 June 2020 ABSTRACT Accepted: 8 September 2020 Chromogenic thin films are crucial building blocks in smart windows to modulate Published online: the flux of visible light and heat radiation into buildings. Electrochromic 6 October 2020 materials such as tungsten oxide are well established in those devices. Sputter deposition offers a well-suited method for the production of such layers, which The Author(s) 2020 can also be used on an industrial scale. Tungsten oxide films were prepared by means of reactive ion-beam sputter deposition. The choice of distinct gas mix- tures as well as the growth temperature during the sputtering process allows to tune the properties of the resulting layers. Especially, the variation in the growth temperatures was found to have an impact on the structure of the resulting samples and, as a consequence, on their optical and electrochemical properties. By specific choice of the reactive gas, the deposition of colorless transparent as well as blue films of different composition is possible. The optical transmittance in the visible spectral range was up to 75% for as-deposited oxygen-rich layers. Additionally, hydrogen-doped tungsten oxide samples were grown. Superior electrochromic switching was observed for H -doped layers, probably by some kind of preconditioning. This resulted in a value for the standardized optical coloration efficiency of 26.5 cm /C. technologies for the conversion, storage and use of Introduction renewable energies are constantly growing [1]. The building sector plays an important role in terms of Because of the global climate change, energy-saving energy saving potential. For example, this sector and sustainable technologies are becoming more and consumes 23% of global primary energy and 30% of more important. Therefore, the demands on global electricity demand [2]. The further Handling Editor: Kevin Jones. Address correspondence to E-mail: Mario.Gies@exp1.physik.uni-giessen.de https://doi.org/10.1007/s10853-020-05321-y J Mater Sci (2021) 56:615–628 development of technologies that enable energy to be properties result from the different level of oxygen used efficiently and economically is, therefore, of deficit in phases of WO , which were investigated by ´ ´ foremost interest. In particular, the class of so-called Magneli and are thus known as Magneli phases smart windows offers an approach to save energy in [23, 24]. In addition to their EC characteristics, tung- the building sector by efficiently regulating incident sten oxides can be successfully applied in, e.g., gas light [3]. This can be achieved by the use of elec- sensors, catalysts or electrode materials in lithium-ion trochromic (EC) thin films. Electrochromism denotes batteries [25–29]. the reversible change of optical absorbance driven by The EC properties of tungsten oxide layers depend an externally applied voltage; thus, EC materials on the composition, the crystal structure and the allow for the modulation of the incident light. morphology [8, 30–32]. Compared to crystalline films, The mostly studied EC materials include conju- amorphous WO exhibits more rapid coloration and gated conducting polymers, transition metal oxides improved efficiency but inferior stability [30, 31]. as well as metal coordination complexes [4–9]. Due to Hence, using low-dimensional nanostructures has their diversity in composition and structure as well as been proven effective to achieve faster switching, their superior performance, electrochromism based increased reversibility and enhanced durability on thin film transition metal oxides has become [30, 32]. This is due to the higher surface accessibility increasingly important in the last decade. Exemplar- and shorter diffusion path length versus the bulk ily, there has been significant progress in terms of counterpart [30]. The film characteristics are strongly new materials [10–15] as well as novel designs and dependent on the growth technique. A multitude of concepts [14, 16–19]. Among the various transition different techniques were used, metal organic chem- metal oxides, tungsten oxide is the most intensively ical vapor deposition [33], the sol–gel process [34, 35], studied material and has been established as material pulsed laser deposition [36–38] or sputter deposition of choice in this field, especially with regard to its [39, 40], among others. Especially, sputter-deposited extraordinary EC properties like high color efficien- coatings show some intrinsic advantages, such as a cies, high cyclability and high environmental stabil- strong layer adhesion, a constant film thickness of ity, among others [20, 21]. The underlying process of homogeneous structure, high growth rates and a high electrochromism is widely accepted as a result of degree of reproducibility, which trigger their domi- simultaneous injection/extraction of electrons and nance in industrial scale. Nevertheless, there are cations [22] and complies with the following reaction drawbacks in conventional setups such as the contact þ  between the substrate and the plasma during depo- ½WO þ xM þ xe  ,½M WO  ; ð1Þ 3 x 3 bleached colored sition. Therefore, the temperature of the substrate where x describes the number of incorporated ions surface can rise significantly. In this article, ion-beam and electrons, respectively. The type of ions, such as sputter deposition (IBSD) was used for the synthesis þ þ þ þ H ,Li or Na , is represented by M . The interca- of tungsten oxide films to achieve a more precise lation reaction of electrons and ions into the film can control of the sample temperature and, thus, the layer 6þ 5þ 4þ cause a reduction of the W -states to W -orW - properties [41]. This favors the deposition of amor- states [6]. This leads to a modification of the elec- phous tungsten oxide films, since otherwise plasma tronic structure of the material, whereby photons in irradiation of the substrate surface raises the tem- the visible range of the electromagnetic spectrum can perature, promotes atom migration and, as a conse- be absorbed. Simultaneously, the color impression of quence, can lead to crystallization of the material. For the tungsten oxide layer changes. The untreated the same reason, films already present on the sub- tungsten oxide layer appears colorless transparent, strate, e.g., for stack preparation, can be spared the whereas a blue coloration occurs upon electrochem- harsh conditions of a contact to the plasma. Besides, ical reduction with simultaneous incorporation of smooth and compact surfaces can be realized, which charge-neutralizing ions. A reversal of this process, is in contrast to conventionally sputtered thin films i.e., the removal of ions (deintercalation), results in a being rather rough and porous. The basic character- decolorization of the layer. istics of ion-beam sputtered tungsten oxide coatings Tungsten and its oxides are, further, of great as well as the possibilities to improve the coating interest because of their high availability, low price properties due to variation in synthesis parameters factor and high chemical stability. The special like gas mixture and growth temperature will be J Mater Sci (2021) 56:615–628 617 described. As-grown samples are investigated by spectroscopy was carried out with a 633 nm laser in a means of their optical, compositional and EC range between 100 and 1500 cm . Additionally, a characteristics. colored sample was examined for analysis by Raman spectroscopy. This sample was intercalated with Li - ions by means of a cyclic voltammetry measurement. Experimental As electrolyte, a 1 M solved lithium perchlorate (LiClO ) in propylene carbonate was used. The Tungsten oxide (WO ) films were prepared by ion- potential was varied with a rate of 10 mV/s until beam sputter deposition. The sputtering setup con- - 0.7 V were reached. Further details about the sists of the vessel of radio frequency (RF) ion source, electrochemical measurement setup are described which incorporates the plasma, and the main process below. To gain more precise information about the chamber in which the sputter target and the substrate composition of the film and the chemical bonds, X- are mounted. The ion beam extracted from the ion ray photoelectron spectroscopy (XPS) was conducted. source is directed onto the target, where the material The PHI VersaProbe system utilizes an Al anode of the target is atomized. In comparison with con- (Al-K = 1486.6 eV). Measurements were taken with ventional sputtering systems such as direct current a source angle of 45 and with charge neutralization (DC) or RF sputtering, in IBSD systems obviously on the sample surface. All resulting spectra were there is no direct contact between plasma and sub- referenced to the carbon signal (C 1s) at 284.8 eV. strate and, as a consequence, the temperature of the Depth profiles of the films were studied via in situ substrate surface will not rise significantly. Further argon ion etching with an acceleration voltage of 0.5 details concerning the IBSD setup can be found in or 1 kV. Electron paramagnetic resonance (EPR) reference [41]. spectra were measured at 4 K with a microwave The films were deposited from a metallic tungsten power of 201,17 mW and a microwave frequency of target of Kurt J. Lesker Company (purity of 99.95%) 9.49 GHz. For these investigations, tungsten oxide by reactive sputtering in an argon–oxygen mixture. was deposited on polytetrafluoroethylene (PTFE) foil. For some layers, hydrogen was used as additional The use of PTFE foil allows to remove the deposited reactive gas to deposit hydrogen-doped films. All layer from the substrate so that the remaining pow- gases got a purity of 99.999%. A fixed argon flux of der can be examined in a quartz ampule. Selected 2 sccm was used for all samples. The ion source was samples were examined for their EC properties. operated at an RF power of 220 W. To influence the Cyclic voltammetry was conducted in an IviumStat degree of crystallization of the layers, films were potentiostat between - 0.7 and 1.5 V with potential deposited under ambient as well as elevated growth sweeps of 10 mV/s. The electrochemical half-cell TM temperatures. K Glass coated with fluorine-doped consisted of a platinum wire (counter electrode), a tin oxide (FTO) with a thickness of 300 nm was used leak-free Ag/AgCl reference electrode and a tungsten as substrate in order to establish an electrical contact oxide layer (working electrode). The electrolyte used for electrochemical measurements. was 0.1 M sulfuric acid. During the cyclic voltam- X-ray diffraction (XRD) was carried out with a metry measurements, UV–Vis spectroscopy was diffractometer type D5000 of Siemens company using simultaneously performed with a TEC5 spectrometer the Cu-K emission line. The measurements were at the reversal points of the potential. These reversal taken in Bragg–Brentano geometry with a rate of 2 / points can be considered as the state of the interca- min. A Lambda 900 spectrometer from PerkinElmer lated and deintercalated layer. Due to the sweep rate Instruments was used to measure the optical prop- of the potential, this would result in a switching time erties of the layers. Scanning electron microscopy of 220 s. For the classification of the optical switching (SEM) and atomic force microscopy (AFM) were performance of the investigated layers, the elec- conducted to analyze the crystallite surface structure. trochromic characteristics listed in the following were The SEM measurements were taken using a Zeiss– used. For building glazing, the European standard Merlin setup. For the AFM investigations in air, a EN 4102 [42] serves as the basis for calculating the Smart SPM 1000 (AIST-NT) was used, utilizing optical characteristics. The standardized transmission NanoWorld Pointprobe SEIHR-20 AFM probes T is of special importance with respect to the use as vis designed for non-contact mode imaging. Raman J Mater Sci (2021) 56:615–628 window coating. This takes into consideration the Results and discussion spectral sensitivity of the human eye V(k). The cor- Figure 1a shows the transmittance in dependence on responding values can be found in reference [43]. wavelength for pristine tungsten oxide films depos- Here the spectral range from 380 to 780 nm is ited at room temperature and varied O flux up to examined, with the strongest weighting at about 2 10 sccm. The corresponding photographs are shown 555 nm. Furthermore, the intensity distribution of the as well with the O flux during the deposition incident light is taken into account. For the calcula- decreasing from left to right. Obviously, layers pro- tion, the CIE (International Commission on Illumi- duced under high O flux are highly transparent. nation) Standard Illuminant D65 is used as the Below a certain threshold in the O flux, however, a reference light source. This can be found in the ISO blue coloration of the layers occurs. This points at 11664-2 standard [44]. From these values, the stan- 5þ 4þ dardized transmission T is calculated as follows: partial presence of W - and W -states and, hence, vis an oxygen deficit of the samples. 780 nm D65ðkÞ VðkÞ TðkÞ k¼380 nm T ¼ : ð2Þ P Taking a look at the optical transmission of the vis 780 nm D65ðkÞ VðkÞ k¼380 nm layers in the spectral range between 300 and 650 nm, this tendency is also well established. As opposed a This results in the standardized optical coloration measurement of the substrate (FTO layer on glass) for efficiency: which absorption occurs up to 310 nm, the films of T ðbÞ vis log tungsten oxide absorb up to about 370 nm, well in T ðcÞ vis CE ¼ ; ð3Þ vis line with their colored appearance. Oxygen-rich DQ tungsten oxide samples deposited with an O flux of in which the bleached (b) and colored (c) state of the 6 sccm and above reach a transmittance in the visible standardized transmission and the charge involved is spectral range of up to 75%, close to the 80% of the taken into account. substrate. The general transmission profile of these Tungsten oxide thin films were grown by ion-beam samples is comparable. However, for samples pro- sputter deposition, a less common sputtering variant. duced under lower O flux a decrease in the mea- We showed the possibility of influencing technolog- sured transmittance can be seen. In particular, an O ically relevant samples characteristics by using dif- flux of 5.15 sccm was found to be the lowest O flux ferent preparation parameters (e.g., gas mixture or which reproducibly yields widely colorless trans- growth temperature). This allows to tune the ele- parent samples. Additionally, the sample shows the mental composition, optical properties or to influence highest transmittance for the violet and blue spectral the structure and the degree of crystallization in the range. Samples grown with higher oxygen deficit resulting thin films. Variation in these properties (deposited at an O flux of 5 sccm and less) show a allows for a positive selection of parameters guaran- bluish color impression accompanied by a shift in the teeing layers with beneficial EC characteristics. onset of absorption toward higher wavelengths and Exemplarily, layers of a significantly altered mor- increased absorption in the longer wavelength range. phology and a much more compact structure can be Figure 1b depicts the transmittance series for films produced. Additional in operando doping with prepared at a constant substrate temperature of hydrogen allows to further optimize the cycling 400 C. Again, the upper part of the figure shows behavior. The high reproducibility as well as the high photographs of the samples. In general, the trans- purity of IBSD-grown layers render ion-beam sputter mission measurements of these samples also show deposition a suitable candidate for growth of tung- the decrease in the transmittance for oxygen-poor sten oxide and, most likely, other chromogenic samples as well as the shift of the onset of absorption. materials. Compared to the previous series, however, the O flux below which blue coloring occurs shifts to higher values. This is clearly visible for the sample deposited under an O flux of 5.15 sccm. Here, the sample grown at a growth temperature of 400 C exhibits a dark blue color impression, while the one deposited J Mater Sci (2021) 56:615–628 619 (a) (b) (c) 10 sccm 8 sccm 6 sccm 5.15 sccm 5 sccm 4 sccm 10 sccm 8 sccm 6 sccm 5.15 sccm 5 sccm intercalation 2.5 cm 2.5 cm @ RT @ 400°C 80 Ref. FTO 80 Ref. FTO O flux O flux 2 2 60 60 10.00 sccm 10.00 sccm 8.00 sccm 8.00 sccm 40 6.00 sccm 40 6.00 sccm 5.15 sccm 5.15 sccm 5.00 sccm 5.00 sccm 20 20 1 cm 4.00 sccm 300 400 500 600 300 400 500 600 deintercalation λ (nm) λ (nm) Figure 1 Optical transmittance curves and photographs of coloration of tungsten oxide when reduced in an electrochemical tungsten oxide samples IBSD-grown at room temperature (a) cell at a potential of - 0.7 V versus Ag/AgCl (intercalated state) and 400 C(b) under variation of the O flux superimposed to the or 1.5 V versus Ag/AgCl (deintercalated state) in contact to 0.1 M fixed 2 sccm of Ar. The bluish coloration of some samples stems sulfuric acid (c). from tungsten excess and is not to be confused with the bluish at RT was colorless. Thus, different compositions can (002), (020) and (200) reflexes might be present [47]. It be obtained for samples deposited under the same O should be noted that these signals might be as well flux but at different growth temperatures. Conse- assigned to the presence of a triclinic phase [48]. quently, the different color impression allows an Furthermore, Fig. 2a comprises an additional marked assessment of the composition of the films. This is in area at 45 –55 with two broad signals for which a accordance with the literature, where a ratio of O/ clear assignment is not possible and a superposition W [ 2.5 is reported to show a transparent widely of different reflexes is most likely. Additionally, colorless aspect, while for O/W ’ 2.5 a bluish color around the (101) reflection of FTO further low-in- and for O/W \ 2.5 a metallic character of the films tensity signals seem to be overlayed, apparent in a was observed [45]. It should be noted that the col- significant broadening of the FTO-related signal. oration discussed in the previous paragraph is not to Again, reflections of the monoclinic tungsten oxide be mistaken for the coloration which results from de- are the obvious cause. In conclusion, it is seen that and intercalation processes, cf. Fig. 1c. polycrystalline tungsten oxide films were prepared in Figure 2a shows X-ray diffractograms of samples the 400 C-series and an influence of the chosen O deposited under different growth temperatures and flux during the deposition can be seen. The highest with a varying composition. Herein, the films grown degree of crystallinity is observed for the sample at RT without additional heating of the substrate do sputtered at an O flux of 8 sccm. not reveal any reflexes besides those to be assigned of The morphology of the layers was investigated in SnO (marked in the upper scale [46]), which belong SEM and AFM measurements. Smooth and compact to the FTO substrate. Thus, no crystalline WO was films of WO were realized. In Fig. 2b–e, SEM images 3 3 grown at room temperature. In order to be able to of films deposited under different O flux and with- produce crystalline samples, an increase in the out additional heating of the substrate during depo- growth temperature is necessary. Accordingly, the sition are shown. For O flux up to 8 sccm, compact samples deposited at a growth temperature of 400 C films without any pronounced grain structure are show reflexes of WO phases. A dominant signal formed. The sample deposited at an O flux of 3 2 arises in the region of 20 –25 . However, since the 10 sccm shows a similar morphology with some reflection appears rather broad, a clear assignment to protrusions of 0.2–0.3 lm size without, however, any a single lattice plane is not possible. A comparison distinct orientation, in a line with the findings of X- with the predicted reflexes for monoclinic WO (blue ray diffraction. In Fig. 2f, the influence of the growth squares in Fig. 2a) indicates that a superposition of temperature on the surface structure at a fixed O T(%) T (%) J Mater Sci (2021) 56:615–628 (a) (110) (101) (200) (211) (310) (301) SnO 4 sccm 6 sccm 5.15 sccm (b) (c) (f) (g) T = 25°C T = 400°C growth growth f = 5 sccm f = 5 sccm (i) O2 O2 f = 8 sccm f = 8 sccm O2 O2 f = 10 sccm f = 10 sccm O2 O2 8 sccm 10 sccm (d) (e) (h) WO (m) 1 µm 20 30 40 50 60 70 80 2Θ (°) Figure 2 X-ray diffractograms of samples of varying cf. (f) and AFM images of samples, deposited at room temperature composition, deposited at RT (green) or at 400 C (red) under a moderate O flux of 5.15 sccm (g) and under oxygen-poor compared to reference values of SnO (red bars on top) and conditions (h). Compared to the surface of a sample grown at monoclinic WO (shown by blue squares) (a). SEM images of 400 C(i), the surface roughness is significantly smoother. films deposited under different O flux at RT (b)–(e) or at 400 C, flux of 5.15 sccm during deposition is shown. A T = 25°C T = 400°C growth growth morphology different from the samples shown in f = 5 sccm f = 5 sccm O2 O2 f = 5.15 sccm Fig. 2b–e should be emphasized. Unlike those sam- O2 f = 5.15 sccm, int. O2 ples, the film in Fig. 2f shows well-separated grains f = 10 sccm f = 10 sccm O2 O2 with a size up to 0.3 lm but less pronounced crystal 5+ 6+ W -O 6+ W -O W =O facets. In Fig. 2g, the surface of a sample deposited at 5+ RT and a moderate O flux of 5.15 sccm is shown as W =O analyzed by AFM. Individual grains of about 0.2 lm size appear interconnected without sharply defined grain boundaries. The root-mean-square surface roughness was determined to be around 9 nm. In Lee 1998 [43] comparison, Fig. 2h shows the morphology of a Salje 1975 [41] sample synthesized at RT under oxygen-poor condi- tions. Again, no sharply defined grains are recog- nizable. However, the grains seem to be a bit more extended. The determined roughness of the surface is approximately 7 nm. At an increased deposition 200 400 600 800 1000 1200 1400 -1 temperature of 400 C, larger round-shaped grains of Raman-shift (cm ) about 0.5 lm lateral expansion were obtained, cf. Figure 3 Raman spectra of tungsten oxide thin films deposited at Fig. 2i, leading to an increased roughness of around different growth temperatures and O fluxes as well as a spectrum 20 nm, much higher than for the unheated samples. of a Li -intercalated sample (blue) with band assignments Figure 3 shows Raman spectra obtained for sam- indicated as vertical lines. Bluish WO layers display a higher ples with different optical impression (colorless intensity especially in the range of about 330 and 450 cm transparent or blue) as well as for samples synthe- 5þ 5þ attributed to W –O and W =O bonds, whereas in colorless sized at different growth temperature. The amor- 6þ transparent samples W -related Raman modes dominate. phous samples deposited under an O flux 5.15 sccm at room temperature show two broad 808 cm [50]. The second peak resulting from the signals of higher intensity in the region of 770 and 6þ W =O stretching mode of terminal oxygen atoms 950 cm . According to the literature, the first one is [51]. With respect to amorphous tungsten oxide, the 6þ due to the W –O-bond [49] and is caused by a basic structure can be described by the formation of superposition of the two strongest peaks at 719 and 6þ WO -octahedra. These consist of short W =O-bonds intensity (arb. units) intensity (arb. units) J Mater Sci (2021) 56:615–628 621 6þ signal consisting of two contributions at binding and longer W –O-bonds leading to clusters of 6þ energies of 530.3 and 531.6 eV, cf. Fig. 4b. The former deformed octahedra, where the W =O-bonds are 2 6þ has been associated with O and W ions, whereas expected at the outermost surface of the clusters the latter arises from contamination at the surface, [51, 52]. The bluish samples (oxygen poor and Li - such as hydroxyl groups or oxygen as part of carbon- intercalated) yield spectra with a weaker intensity in 6þ related impurities. the region of the W -states. The colorless transpar- Figure 4c shows the O 1s signal after an etching ent sample (deposited under an O flux of 5.15 sccm) time of 300 s upon which additional signals repre- shows a significant increase in intensity in the range 5þ 1 þ senting oxygen bound to W (signal at 530.8 eV) and of about 330–450 cm after the intercalation of Li - 4þ 5þ W (signal at 530.6 eV). A detailed spectrum of the ions. These signals are assigned to the W –O and 5þ W4f core level at the surface is given in Fig. 4d W =O bonds [52]. Following the model of small 5þ consisting of W 4f (at 35.4 eV) and W 4f (at 7=2 5=2 polaron transitions by Schirmer et al. [53], the W 37.5 eV), both associated with the oxidation state states contribute to the mechanism that leads to 6þ W . After an etching time of 300 s (Fig. 4e) further optical absorption. The Raman measurements of the contributions have to be considered pointing at the films therefore give an indication of the presence of 5þ 4þ these states in the stained samples. presence of W and W oxidation states, which Crystalline WO films showed three Raman signals confirm the assignment in the corresponding O 1s 1 1 region. Obviously, the surface of the sample only with a high intensity at about 270 cm , 690 cm 1 6þ contains W oxidation states. With increasing depth and 810 cm , slightly shifted relative to the reported 5þ 4þ 1 1 1 of layer, however, increasingly W and W oxida- positions 275 cm , 719 cm and 808 cm [49, 50]. tion states are present. Due to the fact that the sample Beside those modes, both samples show a higher considered is an oxygen-rich, colorless transparent intensity in the signal around 330 cm . At this wave 5þ tungsten oxide layer, a high proportion of W and number, Ozkan et al. mention the existence of a mode 4þ which occurs for crystalline samples [54]. In contrast W states, however, is unlikely. Rather, a preferen- tial oxygen etching during the depth profiling is to the amorphous films, the mode at 950 cm only assumed, which does not allow an exact determina- appears as a weak signal and predominantly for the tion of the concentration ratio by means of XPS. film prepared at low O flux. Tentatively, we assign 6þ Hence, we discuss the data in terms of a trend of the this to a removal of doubly bonded W for crys- W/O-ratio within the series. Figure 4f shows the W/ talline tungsten oxide. Thus, the intensity of the O-ratio of an oxygen-poor (blue film), a slightly 950 cm Raman mode might serve as a measure of substoichiometric (deposited under 5.15 sccm and at the degree of crystallization. In conclusion, a better RT or with a growth temperature of 400 C) and an crystallization is observed for the oxygen-rich oxygen-rich sample (colorless transparent film), pre- sample. sented in dependence of the O flux during deposi- Figure 4a presents the EPR spectra of a colorless tion. In each case, the measurement of the surface transparent as well as a bluish sample compared to a (squares) and after an etching step (triangles) is PTFE film as reference. For all examined samples, a shown. The surface of the sample deposited under signal can be seen at approximately H = 3370 G and the lowest O flux provides a nearly stoichiometric ´ 2 H = 3388 G. For this, a Lande factor of g ¼ 2:00 W/O ratio of about 0.31. However, due to the low results which is characteristic for free electrons. Two oxygen content during the deposition process, an more resonances are seen for the bluish sample at oxygen deficit can be assumed and the resulting ratio H = 3830 G and H = 4412 G. For those signals, 3 4 is due to surface contamination of the layer. In con- Lande´ factors of g ¼ 1:77 and g ¼ 1:54 are calculated. trast, an etching time for 300 s with 1 kV acceleration These values are in good agreement with those given 5þ voltages results in a W/O ratio of about 0.58. How- in the literature [51], assigned to W ions in colored ever, such a strong substoichiometry of oxygen is 5þ tungsten oxide films confirming the presence of W more than doubtful since for this ratio, a metallic states in our bluish samples. character would be expected. Comparable trends are XPS measurements on an oxygen-rich sample (de- evident for the two samples prepared without posited with an O flux of 10 sccm) show an O 1s growth temperatures and with an O flux of 5.15 and 2 J Mater Sci (2021) 56:615–628 Figure 4 EPR spectra of a (a) transparent (deposited with an H as-deposited O flux of 5.5 sccm at RT) as intercalated PTFE well as a bluish sample (deposited with an O flux of 5þ 3 sccm at RT) (a). W oxidation states are only present for blue samples. O 1s (b, c) and W 4f core level spectra (d, e) measured by XPS at the surface of a film 2000 3000 4000 5000 6000 prepared at RT and with an O H(G) flux of 10 sccm (b, d) as well as after etching the surface for (b) (c) 300 s at 1 kV (c, e). Features 1.0x10 attributed to oxidation states 1.4x10 6+ O(W ) 4þ 5þ W and W arise after the 1.2x10 6+ O (W ) etching procedure. Calculated 8.0x10 5+ O (W ) W/O-ratio (f) of the film 4 1.0x10 surface or during etching at 3 4+ 8.0x10 6.0x10 acceleration voltages of O (W ) OH 0.5 kV (open triangles) or 3 OH 6.0x10 1 kV (solid triangles) for 4.0x10 samples prepared at different 4.0x10 O flux at RT compared to a 540 535 530 525 540 535 530 film prepared at 400 C. binding energy (eV) binding energy (eV) 6þ Corresponding W / 5þ 4þ (W ? W ) ratios (d) (e) (g) resulting from the depth 6+ 4 4 5+ W 4f 6+ 1.0x10 1.0x10 7/2 W 4f W 4f 7/2 7/2 profile of the oxygen-rich and 6+ 5+ 3 W 4f 3 W 4f crystalline samples. 5/2 8.0x10 8.0x10 5/2 6+ 4+ W 4f W 4f 3 3 5/2 7/2 6.0x10 6.0x10 4+ W 4f 5/2 3 3 4.0x10 4.0x10 3 3 2.0x10 2.0x10 0.0 0.0 45 40 35 30 25 45 40 35 30 25 binding energy (eV) binding energy (eV) surface surface (f) (g) etching 1.0 kV for 300 s etching 1.0 kV for 300 s T 70 etching 0.5 kV etching 0.5 kV 0.6 growth surface T surface T growth growth etching T etching T growth growth 0.5 etching 0.4 etching etching 0.3 1 growth 0.2 0 2468 10 2.5 5.0 7.5 10.0 O flux [sccm] O flux (sccm) 2 2 intensity (cps) intensity (cps) ratio W/O intensity (arb. units) intensity (cps) intensity (cps) 6+ 4+ 5+ W /(W +W ) J Mater Sci (2021) 56:615–628 623 10 sccm. As expected, the examination of the surfaces influence of preferential oxygen sputtering is less shows that a higher selected O flux during the pronounced here due to the less altered W/O ratio of deposition process counteracts the deficient incorpo- 0.62 for depth measurement (red triangle). ration of O . However, on the basis of the series Figure 4g illustrates the ratio of the oxidation states 6þ 5þ 4þ shown, only a tendency of the decreasing ratio can be W /(W ? W ) for the measurements already seen. It remains to be mentioned that this results in a discussed in Fig. 4f. Only the thin film produced slight superstoichiometry of oxygen. Again, it is very under the highest O flux and the crystalline sample likely that this ratio is due to an oxygen-rich con- are shown. For both samples, the measurement of the tamination of the sample surface from ambient air. surface and the depth measurement after an etching Compared to the measurements of the strongly time of 300 s (with 1 kV acceleration voltages) is oxygen-deficient sample, the depth measurements of depicted. For the oxygen-rich sample, the additional the other two amorphous layers show comparable depth profile was measured with the lower configu- trends. Once more, conditions which would corre- ration of the acceleration voltage of 0.5 kV and after spond to a metallic character have to be determined. time steps of 180, 360 and 540 s. The corresponding 6þ 5þ 4þ A comparison of the trend of the conditions within W /(W ? W ) ratio decreases during depth the data points for the depth profiles provides an profiling. This is due to the preferential removal of analogy in the development of the surface measure- oxygen. With increasing depth, the ratio saturates. ments of the samples. For the depth measurements, We tentatively explain this as follows: At the begin- the W/O ratios are displaced to a higher W/O ratio. ning of the measurement (within the first 180 s), the Although a direct influence of the oxygen content 6þ 5þ W oxidation states are reduced to W oxidation depending on the chosen O flux during deposition 6þ states. Thereafter, in addition to the reduction of W can be seen, a quantitative analysis is hindered. To 5þ oxidation levels also W oxidation states are determine the influence of preferential etching dur- 4þ reduced and thus the presence of W oxidation ing depth profiling, detailed XPS-measurements were levels increases. Thus, compared to the first 180 s of taken under lower acceleration voltages of 0.5 kV for 6þ the etching time, the change of the W / the argon ions. Measurements were taken after an 5þ 4þ (W ? W ) ratio decreases again. In contrast, it is etching time of 180, 360 and 540 s for an oxygen-rich clear for the crystalline sample that the determined sample and are represented in Fig. 4f as open trian- oxidation states during the measurement result in a gles. After an etching time of 180 s, only a slight nearly non-changing ratio. Thus, the crystalline layer substoichiometry with an W/O ratio of 0.34 is found. shows, compared to the results of the amorphous Further etching resulted in a ratio of 0.4. Despite samples, a generally deviating behavior with respect enhanced substoichiometry of oxygen in the sample to the XPS measurements. The reason for this could for further etching steps, the W/O ratio remains be the presence of more stable bonds to oxygen or the below 0.47, which was measured after an etching concentration ratios being maintained despite etching time of 300 s at an acceleration voltage of 1 kV. The during the data acquisition. measurements suggest a less pronounced preferential The IBSD grown amorphous as well as crystalline etching of oxygen at this lower acceleration voltage of tungsten oxide layers with varying ratios of tungsten 0.5 kV, however, presumably also reduced erosion oxidation states were examined by means of cyclic rates of the layer. voltammetry regarding their EC properties. Figure 5a Furthermore, the figure shows the results of XPS shows nine cycles of the amorphous, colorless measurements on a crystalline film with blue color transparent sample deposited under an oxygen flux (deposited under an O flux of 5.15 sccm and at a of 5.15 sccm. For the reduction in the film, accom- growth temperature of 400 C, red symbols). Com- panied by H -intercalation (eq. 1), the current den- pared to the above mentioned samples, a clear devi- sities become increasingly negative starting at a ation of the O/W ratio of 0.54 at the surface is voltage of about 0.1 V and rapidly becoming more obvious (red square). Although understoichiometry negative at about - 0.48 V. At the reversal point of can be expected due to the coloration, the result - 0.7 V, the recorded current densities finally reach a appears to be too strongly understoichiometric, as the maximum negative value of about 0.14 mA=cm . metallic character of the layer can be assumed for this Peaks for individual reduction reactions (such as ratio. Simultaneously, it can be observed that the J Mater Sci (2021) 56:615–628 (a) (b) (c) (d) -5 t -5 t 5x10 5x10 -0.7 V / 1.5 V -0.7 V / 1.5 V 0 0 +1.5 V cycle 1 cycle 1 +1.5 V cycle 2 cycle 2 cycle 3 cycle 1 cycle 3 cycle 1 -5 -5 cycle 2 -5x10 -5x10 cycle 4 cycle 2 cycle 4 cycle 3 cycle 3 cycle 5 cycle 4 60 cycle 5 cycle 4 cycle 5 cycle 6 -0.7 V 60 cycle 5 cycle 6 cycle 6 -4 -4 cycle 6 cycle 7 -1x10 -0.7 V -1x10 cycle 7 cycle 7 cycle 7 cycle 8 cycle 8 cycle 8 cycle 9 cycle 8 cycle 9 cycle 9 40 -4 cycle 9 -4 -2x10 -2x10 -0.75 0.00 0.75 1.50 500 750 1000 -0.75 0.00 0.75 1.50 500 750 1000 potential (V) potential (V) λ (nm) λ (nm) Figure 5 Cyclic voltammetry (cycles proceeding from lighter to (b) and (d), respectively, measured at the reversal points - 0.7 to darker color in the plots) for an amorphous (a) and a crystalline 1.5 V. The calculated optical impression at the reversal points [55] sample (c) with the corresponding optical transmission spectra is shown above the transmission spectra (b, d). 6þ 5þ 5þ 4þ more steeply than for the remaining cycles and two W ! W or W ! W ) cannot be assigned. distinct features at approximately - 0.3 and - 0.1 V Upon reversal of the voltage sweep, only a broad occur re-oxidation/deintercalation. As for the amor- oxidation peak is seen. Accordingly, the assignment 4þ 5þ phous film, the charge increases upon successive of single oxidation steps (like W ! W or 5þ 6þ cycling. W ! W ) is not possible. The maximum current 2 2 Figure 5d shows the transmittance and the optical densities of about 2:4  10 mA=cm occur around a color impression for subsequent cycles. Obviously, voltage of - 0.45 V. Toward higher voltages, the the switching process is not completely reversible. current densities decay quite quickly, indicating Particularly in the near infrared range, a proceeding complete oxidation of the film. Based on the contin- drop in transmittance occurs from cycle to cycle. uously increasing current densities upon successive Both types of samples can be reduced/intercalated cycling of the potential, the enclosed area of the and re-oxidized/ deintercalated. For the crystalline voltammogram slightly increases with the increasing sample, a certain irreversibility reveals caused by number of cycles, which suggests that an increasing incomplete deintercalation. The irreversible switch- charge is involved in the process. Consequently, the ing behavior could be caused by a sweep rate of the film is reduced and re-oxidized to an increasing potential which was too fast. The more densely extent. packed atomic structure in crystalline samples could The optical transmission of the sample at each hinder the ion diffusion. As already mentioned in reversal point of the potential, cf. Fig. 5b, represents Sect. 1, an extended response time of the coloration in the bleached (at ? 1.5 V) and colored (at - 0.7 V) crystalline samples is a well-known phenomenon states. Rather small differences of transmission of the [30]. Thus, inserting protons gradually change the as- bleached and colored state can be seen, however, grown crystalline samples. reversibly within every single cycle and increases To suppress such behavior, an additional series of with the increasing number of cycles, caused by samples was grown in which hydrogen was used as increasing coloration. This is underlined by the additional reactive gas in the sputtering process. It is optical color impression of the film in the different expected that hydrogen is incorporated into the (de)intercalated conditions at the reversal points deposited layers and influences the structure and, (- 0.7 V or 1.5 V) shown in Fig. 5b. consequently, to show an influence on the EC Figure 5c shows nine cycles of cyclic voltammetry properties. on a crystalline WO sample. The characteristics of Figure 6a illustrates photographs of the samples the first cycle differ slightly from those of the subse- (above) as well as the corresponding optical trans- quent cycles, in that the current density during mission measurements (below). For this series, the reduction/intercalation of the first cycle increases current density (A/cm ) T(%) current density (A/cm ) T(%) J Mater Sci (2021) 56:615–628 625 (a) (b) (c) 0 sccm 0.5 sccm 0.65 sccm 0.8 sccm 1.5 sccm 2 sccm 2.5 cm 0 -0.7 V / 1.5 V cycle 1 +1.5 V 80 Ref. FTO Cycle 1 cycle 2 -4 Cycle 2 -1x10 H flux Cycle 3 2 cycle 3 Cycle 4 0.00 sccm cycle 4 0.50 sccm 0.65 sccm -4 -2x10 0.80 sccm -0.7 V 1.50 sccm 2.00 sccm -4 -3x10 300 400 500 600 -1.0 -0.5 0.0 0.5 1.0 1.5 500 750 1000 λ (nm) potential (V) λ (nm) Figure 6 Optical transmission spectra and photographs (top) of points (c) measured at a colorless transparent hydrogen-doped tungsten oxide films deposited at RT with fluxes of 2 sccm Ar and WO film (fluxes of 2 sccm Ar, 7 sccm O and 10 sccm H ) with x 2 2 5.15 sccm O , and increasing additional flux of H (a). Cyclic the color impression of the film in the current switching state 2 2 voltammetry (b) as well as optical transmission at the reversal (above the graph). additional H flux was varied between 0.5 and To compare the electrochromic characteristics of 2 sccm, whereas the O flux of 5.15 sccm and the IBSD grown films, Table 1 summarizes the results of growth temperature (RT) were fixed. Layers which an undoped tungsten oxide sample (measurement is were produced with a hydrogen flux of 0.8 sccm or shown in Fig. 5a, b) as well as the high doped sample above show a blue coloration indicative of partial from Fig. 6b, c. A comparison shows a clear increase reduction. Increasing the H flux even further in the change of the standardized transmission at strengthens the coloring. The absorption edge of all visible range. For the H -doped sample, this change samples starts at about 320 nm. In order to prevent is about twice as large. At the same time, compared to coloration of samples due to the deficient oxygen the undoped sample, a larger charge quantity of content, the O flux was increased to 7 sccm. At the 2.3 mC/cm is involved within the cycle under con- same time, an H flux of 10 sccm was chosen to 2 sideration. An improved value of 26.5 cm /C is also achieve a high doping. This resulted in a colorless obtained for the standardized optical coloration effi- transparent layer. ciency of the doped sample. Figure 6b shows four cyclic voltammetry cycles of such a hydrogen-doped sample. Essentially, the voltammograms resemble those shown in Fig. 5a. Conclusion However, the coloration current density is more Tungsten oxide thin films were grown by ion-beam negative. Quite constant current density and charge sputter deposition, a less common sputtering variant. are observed for all cycles. The transmittance spectra We showed the possibility of influencing technolog- recorded simultaneously and the optical color ically relevant samples characteristics by using dif- impression of the layer, cf. Fig. 6c, show a widely ferent preparation parameters (e.g., gas mixture or reversible switching process of high contrast in growth temperature). This allows to tune the ele- transmission. For the reduced/intercalated layer, the mental composition, optical properties or to influence transmittance reversibly drops to values as low as the structure and the degree of crystallization in the 50% at 800 nm as opposed to the irreversible char- resulting thin films. Variation in these properties acteristics of the samples reported in Fig. 5. It can, allows for a positive selection of parameters guaran- therefore, be proven that hydrogen, used as an teeing layers with beneficial EC characteristics. additional reactive gas in the manufacturing process, Exemplarily, layers of a significantly altered mor- yields samples which have significantly improved EC phology and a much more compact structure can be switching properties. T (%) current density (A/cm ) T (%) J Mater Sci (2021) 56:615–628 Table 1 EC characteristics of EC characteristics an undoped tungsten oxide layer (fluxes of 2 sccm Ar and 2 2 Sample DT (%) DQ (mC/cm)CE (cm /C) vis vis 5.15 sccm O ) and an H - 2 2 doped film (fluxes of 2 sccm Amorphous-undoped 5.3 1.3 22.2 Ar, 7 sccm O and 10 sccm 2 H -doped 10.7 2.3 26.5 H ), deposited under ambient For both the characteristics after the fourth cycle is given growth temperatures produced. Additional in operando doping with References hydrogen allows to further optimize the cycling [1] Ahmadi MH, Ghazvini M, Nazari MA, Ahmadi MA, Pour- behavior. The high reproducibility as well as the high fayaz F, Lorenzini G, Ming T (2019) Renewable energy purity of IBSD-grown layers render ion-beam sputter harvesting with the application of nanotechnology: a review. deposition a suitable candidate for growth of tung- Int J Energy Res 43:1387–1410 sten oxide and, most likely, other chromogenic [2] Urge-Vorsatz D, Cabeza LF, Serrano S, Barreneche C, Pet- materials. richenko K (2015) Heating and cooling energy trends and drivers in buildings. Renew Sustain Energy Rev 41:85–98 [3] Ta¨llberg R, Jelle BP, Loonen R, Gao T, Hamdy M (2019) Acknowledgements Comparison of the energy saving potential of adaptive and We thank P. J. Klar and D. M. Hofmann for fruitful controllable smart windows: a state-of-the-art review and discussions and their constant support. The work of simulation studies of thermochromic, photochromic and M. Gies and F. Michel was funded by the DFG via the electrochromic technologies. Sol Energy Mater Sol C 200: RTG (Research Training Group) 2204 ‘‘Substitute 109828. https://doi.org/10.1016/J.SOLMAT.2019.02.041 Materials for Sustainable Energy Technologies’’. [4] Green S, Backholm J, Geore´n P, Granqvist CG, Niklasson GA (2009) Electrochromism in nickel oxide and tungsten oxide thin films: ion intercalation from different electrolytes. Funding Sol Energy Mater Sol Cells 93:2050–2055 [5] Mortimer RJ (2011) Electrochromic materials. Annu Rev Open Access funding enabled and organized by Mater Res 41:241–268 Projekt DEAL. [6] Granqvist CG (2014) Electrochromics for smart windows: oxide-based thin films and devices. Thin Solid Films Open Access This article is licensed under a Crea- 564:1–38 tive Commons Attribution 4.0 International License, [7] Lahav M, Boom ME (2018) Polypyridyl metallo-organic which permits use, sharing, adaptation, distribution assemblies for electrochromic applications. Adv Mater 30: and reproduction in any medium or format, as long 1706641. https://doi.org/10.1002/adma.201706641 as you give appropriate credit to the original [8] Wu W, Wang M, Ma J, Cao Y, Deng Y (2018) Elec- author(s) and the source, provide a link to the Crea- trochromic metal oxides: recent progress and prospect. Adv tive Commons licence, and indicate if changes were Electron Mater 4:1800185 made. The images or other third party material in this [9] Sadeghi K, Yoon J-Y, Seo J (2019) Chromogenic polymers article are included in the article’s Creative Commons and their packaging applications: a review. Polym Rev licence, unless indicated otherwise in a credit line to 60:442–492 the material. If material is not included in the article’s [10] Cai GF, Tu JP, Zhang J, Mai YJ, Lu Y, Gu CD, Wang XL Creative Commons licence and your intended use is (2012) An efficient route to a porous NiO/reduced graphene not permitted by statutory regulation or exceeds the oxide hybrid film with highly improved electrochromic permitted use, you will need to obtain permission properties. Nanoscale 4:5724–5730 directly from the copyright holder. To view a copy of [11] Cai G, Tu J, Zhou D, Li L, Zhang J, Wang X, Gu C (2014) this licence, visit http://creativecommons.org/licen Constructed TiO /NiO core/shell nanorod array for efficient ses/by/4.0/. electrochromic application. J Phys Chem C 118:6690–6696 J Mater Sci (2021) 56:615–628 627 [12] Zhang JH, Cai GF, Zhou D, Tang H, Wang XI, Gu CD, Tu photothermal catalytic performance. J Mater Chem A JP (2014) Co-doped NiO nanoflake array films with 4:5314–5322 enhanced electrochromic properties. J Mater Chem C [27] Solis JL, Saukko S, Kish L, Granqvist CG, Lantto V (2001) 2:7013–7021 Semiconductor gas sensors based on nanostructured tungsten [13] Zhang JH, Tu JP, Zhou D, Tang H, Li L, Wang XI, Gu CD oxide. Thin Solid Films 391:255–260 (2014) Hierarchical SnO @NiO core/shell nanoflake arrays [28] Yoon S, Jo C, Noh SY, Lee CW, Songc JH, Lee J (2011) as energy-saving electrochromic materials. J Mater Chem C Development of a high-performance anode for lithium ion 2:10409–10417 batteries using novel ordered mesoporous tungsten oxide [14] Kim J, Ong GK, Wang Y, LeBlanc G, Williams TE, Mattox materials with high electrical conductivity. Phys Chem Chem TM, Helms BA, Milliron DJ (2015) Nanocomposite archi- Phys 13:11060–11066 tecture for rapid, spectrally-selective electrochromic modu- [29] Yoon S, Woo S-G, Jung K-N, Song H (2014) Conductive lation of solar transmittance. Nano Lett 15:5574–5579 surface modification of cauliflower-like WO and its elec- [15] Zhou D, Shi F, Xie D, Wang DH, Xia XH, Wang XL, Gu trochemical properties for lithium-ion batteries. J Alloys CD, Tu JP (2016) Bi-functional Mo-doped WO nanowire Compd 613:187–192 array electrochromism-plus electrochemical energy storage. [30] Cong S, Geng F, Zhao Z (2016) Tungsten oxide materials for J Colloid Interface Sci 465:112–120 optoelectronic applications. Adv Mater 28:10518–10528 [16] Wei D, Scherer MR, Bower C, Andrew P, Ryhanen T, [31] Wang Y, Runnerstrom EL, Milliron DJ (2016) Switchable Steiner U (2012) A nanostructured electrochromic superca- materials for smart windows. Annu Rev Chem Biomol Eng pacitor. Nano Lett 12:1857–1862 7:283–304 [17] Llorde´s A, Garcia G, Gazquez J, Milliron DJ (2013) Tunable [32] Zhou D, Xie D, Xia X, Wang X, Gu C, Tu J (2017) All- near-infrared and visible-light transmittance in nanocrystal- solid-state electrochromic devices based on WO k NiO in-glass composites. Nature 500:323–326 films: material developments and future applications. Sci. [18] Cong S, Tian Y, Li Q, Zhao Z, Geng F (2014) Single-crys- China Chem 60:3–12 talline tungsten oxide quantum dots for fast pseudocapacitor [33] Cruz-Leal M, Goiz O, Cha´vez F, Pe´rez-Sa´nchez GF, Her- and electrochromic applications. Adv Mater 26:4260–4267 na´ndez-Como N, Santes V, Felipe C (2019) Study of the [19] Wang JL, Lu YR, Li HH, Liu JW, Yu SH (2017) Large area thermal annealing on structural and morphological properties co-assembly of nanowires for flexible transparent smart of high-porosity A-WO films synthesized by HFCVD. windows. J Am Chem Soc 139:9921–9926 Nanomaterials 9:1298. https://doi.org/10.3390/nano9091298 [20] Niklasson GA, Granqvist CG (2007) Electrochromics for [34] Cronin JP, Tarico DJ, Tonazzi JCL, Agrawal A, Kennedy SR smart windows: thin films of tungsten oxide and nickel (1993) Microstructure and properties of sol–gel deposited oxide, and devices based on these. J Mater Chem WO coatings for large area electrochromic windows. Sol 17:127–156 Energy Mater Sol C 29:371–386 [21] Mardare CC, Hassel AW (2019) Review on the versatility of [35] Wen-Cheun AB, Chan K-Y, Knipp V (2019) Effect of film tungsten oxide coatings. Phys Status Solidi A 216:190004- thickness on electrochromic performance of sol–gel depos- 7. https://doi.org/10.1002/pssa.201900047 ited tungsten oxide (WO ). Opt Mater 94:387–392 [22] Granqvist CG (2000) Electrochromic tungsten oxide films: [36] Rougier A, Portemer F, Que´de´ A, El Marssi M (1999) review of progress 1993–1998. Sol Energy Mater Sol C Characterization of pulsed laser deposited WO thin films for 60:201–262 electrochromic devices Appl. Appl Surf Sci 153:1–9 [23] Magne´li AA (1950) Structure of b-tungsten oxide. Nature [37] Dinca V, Suchea M (2019) Functional nanostructured inter- 165:356–357 faces for environmental and biomedical applications, vol [24] Magne´li G, Andersson B, Blomberg L, Kihlborg L (1952) 363. Elsevier, Amsterdam Identification of molybdenum and tungsten oxides by X-ray [38] Parshina LS, Novodvorsky OA, Khramova OD, Lotin AA, powder patterns. Anal Chem 24:1998–2000 Shchur PA (2019) PLD of thin WO films for solid-state [25] Zhang L, Wang H, Liu J, Zhang Q, Yan H (2020) Nonsto- electrochromic cells. J Phys Conf Ser 1164:012003. https://d ichiometric tungsten oxide: structure, synthesis, and appli- oi.org/10.1088/1742-6596/1164/1/012003 cations. J Mater Sci Mater Electron 31:861–873 [39] Berggren L, Niklasson GA (2003) Optical absorption and [26] Wang L, Wang Y, Cheng Y, Liu Z, Guo Q, Ha MN, Zhao Z durability of sputtered amorphous tungsten oxide films. (2016) Hydrogen-treated mesoporous WO as a reducing Solid State Ionics 165:51–58 agent of CO to fuels (CH and CH OH) with enhanced [40] Kumar KU, Bhat SD, Subrahmanyam A (2019) Elec- 2 4 3 trochromic device with magnetron sputtered tungsten oxide J Mater Sci (2021) 56:615–628 (WO ) and nafion membrane: performance with varying Raman and FTIR spectroscopies. Appl Surf Sci tungsten oxide thickness—a report. Mater Res Express 6:- 33–34:804–811 045513. https://doi.org/10.1088/2053-1591/aafef1 [50] Salje E (1975) Lattice dynamics of WO . Acta Crystallogr [41] Becker M, Gies M, Polity A, Chatterjee S, Klar PJ (2019) 31:360–363 Materials processing using radio-frequency ion-sources: ion- [51] Gabrusenoks JV, Cikmach PD, Lusis AR, Kleperis JJ, beam sputter-deposition and surface treatment. Rev Sci Ramans GM (1984) Electrochromic colour centres in Instrum 90:023901. https://doi.org/10.1063/1.5063976 amorphous tungsten trioxide thin films. Solid State Ionics [42] DIN EN 410:2011-04, Glas im Bauwesen - Bestimmung der 14:25–30 lichttechnischen und strahlungsphysikalischen Kenngro¨ßen [52] Lee SH, Cheong HM, Tracy CE, Mascarenhas A, Benson von Verglasungen DK, Deb SK (1999) Raman spectroscopic studies of elec- [43] DIN 5031-1:1982-03, Strahlungsphysik im optischen Bere- trochromic A-WO . Electrochim Acta 44:3111–3115 ich und Lichttechnik; Gro¨ßen, Formelzeichen und Einheiten [53] Schirmer OF (1980) Small polaron aspects of defects in der Strahlungsphysik oxide materials. J Phys Coll 41:479–484 [44] ISO 11664-2:2008-02, Colorimetry—part 2: CIE standard [54] Ozkan E, Lee S-H, Tracy CE, Pitts JR, Deb SK (2003) illuminants Comparison of electrochromic amorphous and crystalline [45] Gerard P, Deneuville A, Hollinger G, Minh Duc T (1977) tungsten oxide films. Sol Energy Mater Sol C 79:439–448 Color in ‘‘tungsten trioxide’’ thin films. J Appl Phys [55] Lupo´ C, Eberheim F, Schlettwein D (2020) Facile low- 48:4252–4255 temperature synthesis of nickel oxide by an internal com- [46] McCarthy G, Welton J (1989) X-ray diffraction data for bustion reaction for applications in electrochromic devices. SnO . An illustration of the new powder data evaluation J Mater Sci 55:14401–14414. https://doi.org/10.1007/s1085 methods. Powder Diffract 4:156–159 3-020-04995-8 [47] Grier D, McCarthy G (1991) North Dakota State University. Publisher’s Note Springer Nature remains neutral with ICDD Grant-in-Aid, Fargo, North Dakota, USA regard to jurisdictional claims in published maps and [48] Diehl R, Brandt G, Salje E (1978) The crystal structure of triclinic WO institutional affiliations. . Acta Crystallogr B 34:1105–1111 [49] Shigesato Y, Murayama A, Kamimori T, Matsuhiro K (1988) Characterization of evaporated amorphous WO films by

Journal

Journal of Materials ScienceSpringer Journals

Published: Jan 6, 2021

There are no references for this article.