Thermochromic multilayer films of WO₃/VO₂/WO₃ sandwich structure with enhanced luminous transmittance and durability

Abstract

A novel thermochromic WO₃/VO₂/WO₃ sandwich structure was deliberately designed and deposited by a reactive magnetron sputtering technique. The double layer of WO₃ not only functions as an antireflection (AR) layer to enhance the luminous transmittance (T_lum) of VO₂, but also performs as a good protective layer for thermochromic VO₂. Basically, the bottom WO₃ layer functions as a buffer beneficial for the formation of the intermediate VO₂ layer and serves as an AR layer while the intermediate VO₂ layer with primary monoclinic phase acts as an automatic solar/heat control for energy saving. The top WO₃ layer acts as another AR layer, and provides protection in a complex environment. An obvious increase in T_lum by 49% (from 37.2% to 55.4%) is found for VO₂ films after introducing double-layer WO₃ AR coating. The VO₂ deposited on glass exhibits good thermochromism with an optical transition at 54.5 °C, which decreases to 52 °C in WO₃/VO₂/WO₃ sandwich structure, and the hysteresis loop is sharper around the transition temperature, which may be ascribed to the strain and interfacial diffusion. In comparison with single-layer VO₂, the durability in automatic solar/heat control of sandwich-structure VO₂ films is improved nearly 4 times in high temperature and humidity conditions. This multilayer will open a new avenue for the design and integration of advanced thermochromic heterostructures with controllable functionalities for intelligent window and sensing system applications.

---

1 Introduction

In recent years, global warming has attracted increasing concerns due to the excessive use of fossil energy and the rapid rise of the CO₂ level in the atmosphere.¹ And nowadays, we human beings pay more attention to renewable sources of energy and energy savings in different fields. Specifically, building energy consumption shares as much as 30–40% of the primary energy.² To reach the goal of energy saving and reducing the energy consumption of air-conditioning, an available proposal is to adjust the energy exchange between interior and exterior of buildings by windows, especially “intelligent windows” consisting of thermochromic materials.

Vanadium dioxide (VO₂), as a typical semiconductor to metal transition (SMT) compound, is a widely known and extensively used thermochromic material. Below the critical temperature (T_c), VO₂ behaves as a semiconductor, and reflects barely any energy, while it shows reflection behavior in a wide solar wavelength range when it behaves as a semi-metal after exceeding T_c. Bulk VO₂ initiates its phase transition from monoclinic to tetragonal at about 340 K (T_c ∼ 68 °C),³ which is obviously too high for buildings-related energy saving. Fortunately, T_c can be modulated to a suitable temperature via doping, which is dependent on many factors such as charge⁴ and size⁵ of the dopant ion as well as the change of electron carrier density.⁶ For instance, W⁶⁺ doping can increase the electron concentration (e⁻) so as to decrease T_c by ∼20–26 °C per at%, and Fe³⁺ doping deceases T_c by adding more vacancies (h⁺), while Ce³⁺ doping changes T_c on account of its larger radius.⁷ Among all the reported doping elements, tungsten^8–13 has become one of the most effective candidates for manipulating the SMT of VO₂ by doping, while magnesium^14,15 works efficiently in increasing T_lum, and thin films can also benefit from both the reducing property of W and the T_lum-increasing property of Mg via Mg/W co-doping.⁷ In addition, the thermochromic transition temperature of VO₂ can also be influenced by film thickness: VO₂ with a film thickness of less than 300 nm reveals a thermochromic transition temperature of 50 °C, which is ascribed to the strain effect.¹⁶

However, the unstable durability and low visible-light transparency limit the utilization of VO₂ in thermochromic applications. Strong absorptions of intra-band and inter-band exist in the short-wavelength range for both metallic and semiconductive states, which leads to low visible transmittance.¹⁷ Although previous studies have reported generally enhanced visible-light transparency achieved via various dopants like Mg,¹⁸ F,^18,19 and Zn,²⁰ the deterioration of switching behavior and durability accompanied with the enhancement via doping^18–20 cannot be neglected. In this context, a novel and effective way to increase luminous transmittance is to introduce an antireflection (AR) layer,^21–26 which may be promising to balance the requirements. Our previous work²⁷ has reported that the optimal constant of refractive index of an AR layer for VO₂ film should be in the range between 2.0 and 2.4, and conventional studies of AR layers have involved SiO₂,^21,22 TiO₂ (ref. 23 and 24) etc. These AR materials are in accordance with the optimal range of refractive index. AR layers have often been planar because of the simple route of preparation. However, another novel method to produce AR coating is to modify the surface morphology. For instance, in a previous study,²⁸ moth-eye nanostructure was fabricated to enhance the thermochromic properties of vanadium dioxide, for which different periodicity (d) was used to achieve AR. Nevertheless, increase of T_lum is not as good as solar modulation (ΔT_sol). Tungsten oxide is a promising electrochromic material for its prominent coloration efficiency and better electrochemical stability as compared with others,^29,30 and it is also used as a buffer layer in devices³¹ or even as an AR layer itself in WO₃–TiO₂ films.³² Although the refractive index of thin film is enough to meet requirements,^33,34 using WO₃ as the AR layer for VO₂ is rarely reported.

This paper reports recent results concerning the design, formation and characterization of thermochromic VO₂-based multilayer film with WO₃ AR coating. WO₃ layer was chosen for the following reasons. First, WO₃ has an appropriate refractive index and therefore it can be an effective AR coating for VO₂. Second, WO₃ can be a protective layer for VO₂ to ensure the stability of thermochromic property. Lastly, WO₃ has high transmittance in near-infrared and visible light regions because of its wide band gap of 2.62 eV,³⁵ which may slightly decrease transmittance in near-infrared light as an AR layer. In this work, a novel multilayer window combining thermochromic VO₂ with multifunctional WO₃ was prepared and we demonstrate that the luminous and infrared transmittance in conversion of temperature can be optimized depending on the design of the sandwich-structure multilayer stack.

2 Experimental methodology

2.1 Deposition process

BK7 flat glasses and Si (100) with an area of 45 × 45 mm² were applied as the substrates. Substrate/WO₃/VO₂/WO₃ (labeled as sample WVW) multilayer films were prepared with a medium-frequency reactive magnetron sputtering (MFRMS) system, while substrate/WO₃/VO₂ (labeled as sample WV) films and VO₂ (labeled as sample V) single layer were prepared under the same deposition conditions as references. The targets were metallic W (99.9% pure) and V₂O₃ (99.5%) with a rectangular area of 610 × 85 mm² and the deposition was carried out using lock-load system. After an initial pump-down process, the original pressure of the deposition chamber was down to 5.0 × 10⁻⁴ Pa. Afterwards, Ar and O₂ (both 99.99% purity) with various Ar/O₂ proportions were introduced to the deposition chamber. Sputtering then took place at a current of I onto substrates of flat glasses and Si (100). The Ar/O₂ ratio was kept at X and the pressure of chamber was P. The temperature of the deposition process was a constant below 100 °C. VO₂ films were made at I = 3 A, X = 60 and P = 0.8 while the conditions were kept at I = 5 A, X = 1 and P = 0.8 when depositing WO₃ films. Finally, both of these films were annealed at 450 °C at ∼5 Torr for 5 min in a vacuum furnace.

2.2 Film characterization

Normal incidence spectral transmittance (T%) of prepared films was obtained at different temperatures using a UV-vis-NIR spectrophotometer (Hitachi Corp., model UV-4100) equipped with an attachment to control the temperature of the films. The temperature was measured precisely with a temperature sensor in contact with the surface of films, and it was controlled via a temperature controlling unit. The crystalline structure was characterized by X-ray diffraction (XRD) measurements using a Rigaku Ultima IV diffractometer with grazing angle mode using Cu Kα radiation (λ = 0.15418 nm). The root-mean-square (RMS) roughness as well as the surface topography of the films were determined by atomic force microscopy (AFM, SII Nano Technology Ltd, Nanonavi Π) apparatus in tapping mode. AFM images were acquired in ambient atmosphere and the RMS roughness was calculated at different spots with an area of 2 × 2 μm². In addition, to measure the thickness and determine the microstructure of the samples, films were observed with field emission scanning electron microscopy (SEM, Hitachi SU8220). High-resolution transmission electron microscopy (HRTEM, JEOL-2100F) was used to observe the crystallinity, surface, and cross section of the multilayer film. Furthermore, X-ray photoelectron spectroscopy (XPS, Thermo Fisher Scientific Co. Inc., ESCA lab250) was employed to analyze the chemical states and electronic structures of the multilayer. The transition temperature hysteresis loops were measured by collecting the transmittance spectra of samples at a fixed wavelength (2500 nm) with a heating rate of ∼4 °C min⁻¹, and the transition temperature was defined as the temperature at the average value of the two differential extremum operators on the heating and cooling curves, as described in previous work.³⁶ Moreover, to estimate the stability of multilayer films, the samples were placed in a constant-temperature humidity chamber with the accelerated experimental conditions of T_h = 60 °C and 90% relative humidity, and optical transmittance measurements were carried out once a day.

2.3 Luminous and solar transmittance

The integrated luminous (T_lum, 380–780 nm) and solar (T_sol, 350–2600 nm) transmittance of the multilayer films can be obtained based on the calculated spectra using the following equations:andrespectively, where T(λ) denotes the spectral transmittance at a wavelength of λ, Φ_lum(λ) is the standard luminous efficiency for photopic vision, and Φ_sol(λ) is the solar irradiance spectrum for an air mass of 1.5 (corresponding to the sun standing 37° above the horizon).

3 Results and discussion

3.1 Design of sandwich structures

Sandwich structure is often used in optical design to obtain the preconceived spectrum, whether as the AR layer in thermochromic devices or other functional layers in electrochromic devices.^23,24,37,38 In this study, the VO₂ layer plays an important role in the proposed structure, which switches automatically with respect to transmittance to control the solar/heat radiation with changing environment temperature. Generally, T_lum increases with decreasing film thickness due to reduced absorption, but a deteriorated sharpness for optical switching in VO₂ occurs at the same time. Thus, as a thermochromic layer, an appropriate thickness should be chosen for VO₂ to satisfy the requirements for both a high T_lum and a correspondingly sharp switching condition. In this study, thermochromic VO₂ films with thicknesses of ∼50 nm were considered for deposition according to a preliminary calculation and other experience of VO₂ films.^23,27 As for the WO₃ layer, calculation was further progressed using a double-layer AR design for 50 nm VO₂, and the appropriate thickness could be ∼30 nm for each AR layer, which may lead to a noteworthy increase in luminous transmittance. In order to make a comparison, both a single AR layer and a double AR layer were designed and prepared.

3.2 Characterization of the multilayer films

Microstructure. The surface morphologies of deposited films on glasses are exhibited in Fig. 1a–c. As for VO₂ film, the primary particles are worm-like with a grain size of 50–100 nm. One can see that the smaller particles disperse among the worm-like particles, which take the shape of micropores with size of 30–50 nm. The RMS roughness of the surface is 5.5 nm, indicating that the VO₂ film is smooth and compact. Moreover, it is worth mentioning that the porous morphology of film may influence the relevant durableness to a certain extent, which will be discussed later. With the additional bottom WO₃ layer, the upper VO₂ film exhibits a uniform distribution of crystallites and grain size ranges from 20 to 50 nm. Isometric crystal particles occupy a wide range when some columnar crystals implant between them, forming a compact surface of VO₂ layer. As for the sandwich-structure sample, the top WO₃ layer consists of closely packed particles, with diameters of 30–60 nm. The RMS roughness of the latter two samples acquired from the AFM images is 1.1 nm and 1.0 nm, respectively. The rather small RMS roughness values of samples WV and WVW prove that the VO₂ film growing on the bottom WO₃ layer as well as the top WO₃ layer are extremely smooth and homogeneous, which may be attributed to the Frank–van der Merwe growth mode of VO₂ and WO₃ nanofilms, since the annealing process offers sufficient energy for film growth.

Fig. 1 AFM images of samples (a) VO₂, (b) WO₃/VO₂, (c) WO₃/VO₂/WO₃. The insets in the bottom right exhibit the corresponding SEM images. (d) XRD patterns of WO₃/VO₂/WO₃, WO₃/VO₂ and VO₂ single-layer samples.

XRD patterns of as-deposited samples measured at ambient temperature are shown in Fig. 1d. For the single VO₂ layer, the pattern has an intense XRD peak around 2θ = 28°, which corresponds to the signal of (011) monoclinic VO₂ phase (M) and all the peak positions were in accord with the standard reference data (JCPDS no. 43-1051), indicating the exclusive formation of single-phase VO₂ films. For WV as well as WVW films, the XRD peaks corresponding to monoclinic-phase VO₂ can also be observed and the intensity of each VO₂ peak was increased which means that the bottom WO₃ layer and the top one promote the crystallization of VO₂ phase. For the WO₃ coating, peaks corresponding to the primitive orthorhombic structure of WO₃ (JCPDS no. 20-1324) can be observed. In addition, other peaks located at 23.1° and 28.8°, corresponding to (001) plane and (111) plane of WO₃, were found, indicating that WO₃ and VO₂ can reciprocally improve the crystallinity.

To further confirm the crystalline structure of each layer, HRTEM study was carried out on samples V and WVW. For sample WVW, great attention has been paid to the interface between the bottom WO₃ layer and intermediate VO₂ layer as well as the interface between the top WO₃ layer and intermediate VO₂ layer. As shown in Fig. 2a and b, the as-deposited thickness of single VO₂ layer is about 50 nm, and VO₂ film is crystalline. The spacing of the lattice fringes at around 0.329 nm could be indexed as the (011) plane of monoclinic-phase VO₂ thin film. The WO₃/VO₂/WO₃ sandwich structure can be clearly observed in Fig. 2c. The film thickness of both the top and underlying WO₃ is around 30 nm and that of the VO₂ layer is 50 nm. Fig. 2d presents clear lattice fringes for middle VO₂ (M) layer and bottom WO₃ layer. The spacing of the lattice fringes at 0.328 nm and 0.269 nm could be indexed as the (011) plane of M-phase VO₂ and (220) plane of orthorhombic-phase WO₃, respectively. The interface between top layer and bottom layer was also identified clearly from Fig. 2e. The spacing of the lattice fringes was around 0.267 nm in the WO₃ layer and 0.243 nm in the VO₂ layer, which could be indexed as the (220) plane of orthorhombic WO₃ and (200) plane of M-phase VO₂, respectively. Herein, the preferential growth of WO₃ lattice can be considered as the extension of the crystal structure of VO₂ grains. In other words, the top WO₃ layer locally epitaxially grows on the VO₂ film. The localized epitaxy may present a growth relationship with [200] VO₂ (M)//[220] WO₃. They promote each other and interact during the annealing process. Nevertheless, the epitaxy was limited because of the large difference in atomic radius and clear lattice mismatch between tungsten and vanadium. Thus, the crystallinity of the above planes is poor compared with main lattice planes, as shown by XRD.

Fig. 2 HRTEM images of the single-layer VO₂ (a and b) and sample WO₃/VO₂/WO₃ (c–e). (c) HRTEM image of the cross section of sample WO₃/VO₂/WO₃. (d) HRTEM image of interface between middle VO₂ (M) layer and bottom WO₃ layer (the insets correspond to the red rings in the image with high magnification). (e) HRTEM image of interface between top WO₃ layer and middle VO₂ (M) layer (the inset corresponds to the red ring in the image with high magnification).

Optical properties. Fig. 3a depicts the optical transmittance of VO₂, WO₃/VO₂, and WO₃/VO₂/WO₃ films in response to different temperatures (20 °C and 90 °C). The corresponding calculated optical parameters are listed in Table 1. Compared with the VO₂ films, the WO₃/VO₂/WO₃ multilayer films on the BK7 glasses exhibit higher luminous transmittance. As presented in Table 1, the luminous transmittance in the low-temperature semiconductor phase (T_lum-L) varies from 37.2% (V sample) to 55.4% (WVW sample), which is a sharp increase of 49%. The transmittance difference in the visible light region between the two samples can be directly observed from the inset photograph in Fig. 3a, where the right WO₃/VO₂/WO₃ sample is brighter than the left dark VO₂ sample. Similarly, the luminous transmittance in the high-temperature metallic phase (T_lum-H) also changes from 37.3% (V sample) to 53.9% (WVW sample), and the increase range reaches 45% as well. Hence, the WO₃ layer can effectively enhance the transmittance in the visible light region and boost the integrated luminous transmittance at the same time. In addition, solar transmittance of each sample was also calculated to demonstrate the solar modulation. As shown in Table 1, the ΔT_sol values of WVW sample and WV sample exhibit a slight reduction in comparison with V sample, which may be ascribed to the diffusion from WO₃ layer and/or the additional thermal stress, which agrees with previous works.^22,39 Besides, the damage or defect of the VO₂ layer induced by the ion bombardment in the deposition process is also a potential factor. Furthermore, the three samples exhibit different room temperature transmittances in the near-infrared region. The V sample has the highest near-infrared transmittance while the WVW sample exhibits the lowest. The near-infrared transmittance of sample WV is between the other two, which is in agreement with other reports.^23,24 This phenomenon that the addition of WO₃ can decrease the near-infrared transmittance at room temperature may be caused by the absorption of additional AR layer. On the contrary, the near-infrared transmittance curve of WVW sample measured at high temperature is close to or even lower than that of VO₂ film. In other words, the near-infrared blocking ability of metallic-phase VO₂ is maintained after the addition of double-layer AR WO₃ and the adjustment ability of solar transmittance would be slightly decreased to guarantee the application of thermochromic VO₂.

Fig. 3 (a) Optical transmittance spectra of VO₂, WO₃/VO₂, and WO₃/VO₂/WO₃ films (the inset photograph shows (left) sample VO₂ and (right) WO₃/VO₂/WO₃). (b) Thermal hysteresis loops of samples VO₂, WO₃/VO₂ and WO₃/VO₂/WO₃ measured at a fixed wavelength of 2500 nm. First-order differential of hysteresis loops for (c) VO₂ and (d) WO₃/VO₂/WO₃ films. Optical transmittance spectra of (e) VO₂ and (f) WO₃/VO₂/WO₃ after being placed in hot and humid conditions for 20 days.

Table 1 The luminous transmittance and solar modulation properties of single VO₂ film, WO₃/VO₂ films, and WO₃/VO₂/WO₃ sandwich structure films

Film	Tlum-L (%)	Tlum-H (%)	ΔT2000nm (%)	Tsol-L (%)	Tsol-H (%)	ΔTsol
VO2	37.2	37.3	44.4	40.7	34.6	6.1
WO3/VO2	43.6	42.2	38.2	42.9	37.5	5.4
WO3/VO2/WO3	55.4	53.9	34.7	47.2	42.9	4.3

---

Thermal hysteresis loops of the optical transmittance of the three samples at a fixed wavelength of 2500 nm are presented in Fig. 3b. Fig. 3c and d exhibit the first-order differential of thermal hysteresis loops for WO₃/VO₂/WO₃ and VO₂ films, respectively, and the SMT temperature of samples V, WV and WVW is around 54.5 °C, 53.8 °C and 52 °C, respectively. All the transition temperatures are lower than 68 °C (bulk single crystal of VO₂).^3,40 Also, the difference of transition temperature for heating and cooling processes decreases from 18.8 °C to 13.4 °C after using double-layer WO₃ films, which makes the loops sharper. We consider that the T_c reduction occurs for the following reasons. First, a slight deviation from stoichiometry is a factor for T_c reduction.⁴¹ In addition, the crystal imperfection may cause the T_c reduction due to the damage of zigzag V–V pair chains by defects, which affects the characteristic of the low-temperature phase.²³ Finally, the stress in VO₂ film may act as a mainspring.⁴² For WO₃/VO₂/WO₃ and WO₃/VO₂ films in this work, the T_c reduction phenomenon can also be observed and the shapes of hysteresis are sharper than that of the single VO₂ film. We speculate that the tensile stress of multilayer films arising from the ion bombardment during the sputtering process and also from the cooling process of annealing treatment results in the decrease of T_c. Similar to the doping effect, moreover, tungsten atom diffusion from WO₃ to VO₂ in the interface between WO₃ and VO₂ films may also give rise to the reduction of transition temperature.

Durability. To estimate the weather resistance of prepared films, WVW, WV and V samples were placed in a constant-temperature (60 °C) humidity (90%) chamber for 5, 10, 15 and 20 days. The optical transmittances of sample WVW and sample V after each interval are shown in Fig. 3e and f, respectively. For single-layer VO₂ film, the transmittance contrast between the low-temperature semiconductor phase and high-temperature phase begins to decrease after 5 days, especially in the near-infrared region. The transmittance measured at low temperature drifts from bottom to top, and finally the thermochromism nearly vanishes after 20 days' treatment in the chamber. For WV films (not shown here), the transmittance contrast at different temperatures also begins to change after 5 days. After 20 days' treatment, the thermochromism drops a lot, but the deterioration degree of WV film is weaker than that of single-layer VO₂ film. As a striking contrast shown in Fig. 3f, the optical transmittance of WVW multilayer films remains almost intact, still maintaining the initial curve after 20 days' testing, while a reported Al₂O₃ protective coating only can maintain the thermochromism for 7 days in similar testing condition.⁴³ These phenomena indicate that the covering WO₃ coating not only serves as an AR layer to boost the energy-saving efficiency of VO₂-based smart window, but also can effectively protect the thermochromic VO₂ film from oxidation in hot and wet conditions to promote its durability. Hence, WO₃ functions as a novel candidate for protection layer for VO₂ film.

Component and chemical state. XPS was employed to analyze the component and chemical state of the prepared WO₃/VO₂ and WO₃/VO₂/WO₃ films. In this XPS study, Ar⁺ bombardment was used to strip surface atoms. In order to obtain relatively accurate values, the stripping time was under 10 s to avoid a change of chemical states caused by long-time Ar⁺ bombardment. The resulting spectra were fitted on the basis of the parameters in Table 2^44–47 and are shown in Fig. 4a for WO₃/VO₂ sample and Fig. 4b for WO₃/VO₂/WO₃ sample. For the WV sample, as shown in Fig. 4a, we found that tetravalent vanadium (V⁴⁺) and pentavalent vanadium (V⁵⁺) coexist in the surface of the VO₂ layer. The intensity of V⁵⁺ peak is much weaker than that of V⁴⁺. It seems that a small amount of V₂O₅ exists in the VO₂ layer, which may slightly affect the thermochromic properties of multilayer. On the one hand, the pentavalent vanadium may be caused by the oxidation reaction when the film is exposed to ambient atmosphere. On the other hand, the V⁵⁺ may be produced in the annealing process. For WO₃/VO₂/WO₃ sample, the chemical states of W in the top WO₃ layer are shown in Fig. 4b, and the characteristic peak is well symmetrical, commendably matching to the W⁶⁺ peak. This means that point defects (e.g. oxygen vacancy , interstitial tungsten atom , etc.) can hardly exist in the crystal lattice. If exists in the WO₃ film, a new band would be produced to match the polaron resonance absorption, leading to low transmittance in the infrared range,⁴⁸ which is a disadvantage for room temperature optical transmittance for intelligent windows. Thus, W⁶⁺ atoms in the WO₃ layer may act in a positive role for thermochromic films.

Table 2 Overview of XPS fitted parameters for the V 2p and W 4f signals for different oxygen ratios

Core line	BE (eV)	FWHM (eV)	% L-G	ΔBE (V 2p3/2 − V 2p1/2) (W 4f7/2 − W 4f5/2) (eV)
V^5+ 2p3/2	517.7 ± 0.5	1.4	20	6.8 ± 0.3
V^4+ 2p3/2	516.3 ± 0.5	1.5	20	6.8 ± 0.3
W^6+ 4f7/2	35.7 ± 0.1	1.0	20	2.1 ± 0.1
W^5+ 4f7/2	34.6 ± 0.2	0.7	20	2.1 ± 0.1

---

Fig. 4 Representative fitted (a) V 2p core level spectra and (b) W 4f core level spectra of the surface of WO₃/VO₂ and WO₃/VO₂/WO₃ after etching for 10 s. (c) The cross section of WO₃/VO₂/WO₃ multilayer for EDX measurement and corresponding EDX spectra.

As an addition to HRTEM, EDX (energy dispersive X-ray spectroscopy) was used to investigate the distribution of elements. As shown in Fig. 4c, EDX line scanning was executed from air to glass substrate at the cross section of WVW films. EDX spectra indicate that vanadium of the middle layer is well located and slightly diffuses sideways into WO₃, and the bottom WO₃ layer obstructs the interactive diffusion between VO₂ layer and glass. Even so, a small number of tungsten atoms from the top WO₃ layer diffuse to the middle VO₂ layer and this process may cause the reduction of T_c, which is in agreement with the hysteresis loop results discussed above. Besides, distribution of vanadium and tungsten seems uniform corresponding to each layer and the top WO₃ layer efficiently isolates the oxygen, maintaining the oxygen content of the middle layer, making the thermochromic VO₂ more stable and providing long-term durability as mentioned above.

4 Conclusion

An AR concept was suggested as a solution to enhance luminous transmittance of VO₂ film in the visible region. This paper has presented some novel results on design, formation and characterization of a sandwich-structure VO₂-based thermochromic film with WO₃ AR layer. WO₃/VO₂/WO₃ sandwich structure was deposited by reactive magnetron sputtering using V₂O₃ target for VO₂ and W target for WO₃ in Ar + O₂. A prominent increase in T_lum by 49% (from 37.2% to 55.4%) has been obtained with the WO₃/VO₂/WO₃ multilayer structure. The transition temperature of films varies from 54.5 °C to 52 °C for single-layer and multilayer films, respectively, and the hysteresis loop distinctly becomes sharper, which may be attributed to the interfacial diffusion and strain effect. Besides, near-infrared transmittance of the multilayer slightly decreases at room temperature and is almost maintained at high temperature, which affects thermochromism less. Lastly, the durability of multilayer films in hot and wet atmosphere improves a lot in comparison with single VO₂ film due to the protection from the WO₃ layer as well as the compact surface of the VO₂ layer. This described VO₂-based sandwich-structure film with enhanced luminous transmittance and durability may serve as a potential candidate in applications for energy-efficient buildings and automobiles.

Acknowledgements

This study was financially supported by the high-tech project of MOST (Ministry of Science and Technology of the People's Republic of China, No. 2014AA032802) and the Key Research Program of the Chinese Academy of Sciences (No. KFZD-SW-403).

Notes and references

1. D. S. Lee, D. W. Fahey, P. M. Forster, P. J. Newton, R. C. N. Wit, L. L. Lim, B. Owen and R. Sausen, Atmos. Environ., 2009, 43, 3520–3537.
2. I. P. Parkin and T. D. Manning, J. Chem. Educ., 2006, 83, 393–400.
3. F. J. Morin, Phys. Rev. Lett., 1959, 3, 34–36.
4. T. E. Phillips, R. A. Murphy and T. O. Poehler, Mater. Res. Bull., 1987, 22, 1113–1123.
5. J. b. Macchesn and H. j. Guggenhe, J. Phys. Chem. Solids, 1969, 30, 225–234.
6. J. W. Pierce and J. b. Goodenou, Phys. Rev. B: Solid State, 1972, 5, 4104–4111.
7. N. Wang, S. Y. Liu, X. T. Zeng, S. Magdassi and Y. Long, J. Mater. Chem. C, 2015, 3, 6771–6777.
8. C. B. Greenberg, Thin Solid Films, 1983, 110, 73–82.
9. M. A. Sobhan, R. T. Kivaisi, B. Stjerna and C. G. Granqvist, Sol. Energy Mater. Sol. Cells, 1996, 44, 451–455.
10. T. D. Manning, I. P. Parkin, M. E. Pemble, D. Sheel and D. Vernardou, Chem. Mater., 2004, 16, 744–749.
11. A. Romanyuk, R. Steiner, L. Marot and P. Oelhafen, Sol. Energy Mater. Sol. Cells, 2007, 91, 1831–1835.
12. C. Piccirillo, R. Binions and I. P. Parkin, Thin Solid Films, 2008, 516, 1992–1997.
13. S. Y. Li, G. A. Niklasson and C. G. Granqvist, Thin Solid Films, 2012, 520, 3823–3828.
14. S. L. Hu, S. Y. Li, R. Ahuja, C. G. Granqvist, K. Hermansson, G. A. Niklasson and R. H. Scheicher, Appl. Phys. Lett., 2012, 101, 4.
15. J. D. Zhou, Y. F. Gao, X. L. Liu, Z. Chen, L. Dai, C. X. Cao, H. J. Luo, M. Kanahira, C. Sun and L. M. Yan, Phys. Chem. Chem. Phys., 2013, 15, 7505–7511.
16. T. Maruyama and Y. Ikuta, J. Mater. Sci., 1993, 28, 5073–5078.
17. M. M. Qazilbash, A. A. Schafgans, K. S. Burch, S. J. Yun, B. G. Chae, B. J. Kim, H. T. Kim and D. N. Basov, Phys. Rev. B: Condens. Matter Mater. Phys., 2008, 77, 10.
18. I. Abdellaoui, G. Merad, M. Maaza and H. S. Abdelkader, J. Alloys Compd., 2016, 658, 569–575.
19. L. Dai, S. Chen, J. J. Liu, Y. F. Gao, J. D. Zhou, Z. Chen, C. X. Cao, H. J. Luo and M. Kanehira, Phys. Chem. Chem. Phys., 2013, 15, 11723–11729.
20. M. Jiang, S. H. Bao, X. Cao, Y. M. Li, S. T. Li, H. J. Zhou, H. J. Luo and P. Jin, Ceram. Int., 2014, 40, 6331–6334.
21. S. M. Babulanam, T. S. Eriksson, G. A. Niklasson and C. G. Granqvist, Sol. Energy Mater., 1987, 16, 347–363.
22. M. H. Lee and J. S. Cho, Thin Solid Films, 2000, 365, 5–6.
23. P. Jin, G. Xu, M. Tazawa and K. Yoshimura, Appl. Phys. A: Mater. Sci. Process., 2003, 77, 455–459.
24. N. R. Mlyuka, G. A. Niklasson and C. G. Granqvist, Sol. Energy Mater. Sol. Cells, 2009, 93, 1685–1687.
25. S. Y. Li, G. A. Niklasson and C. G. Granqvist, J. Appl. Phys., 2010, 108, 8.
26. L. Yao and J. H. He, Prog. Mater. Sci., 2014, 61, 94–143.
27. G. Xu, P. Jin, M. Tazawa and K. Yoshimura, Sol. Energy Mater. Sol. Cells, 2004, 83, 29–37.
28. X. K. Qian, N. Wang, Y. F. Li, J. H. Zhang, Z. C. Xu and Y. Long, Langmuir, 2014, 30, 10766–10771.
29. C. G. Granqvist, Sol. Energy Mater. Sol. Cells, 2000, 60, 201–262.
30. J. Livage and D. Ganguli, Sol. Energy Mater. Sol. Cells, 2001, 68, 365–381.
31. J. Meyer, T. Winkler, S. Hamwi, S. Schmale, H. H. Johannes, T. Weimann, P. Hinze, W. Kowlasky and T. Riedl, Adv. Mater., 2008, 20, 3839–3843.
32. E. O. Zayim, Sol. Energy Mater. Sol. Cells, 2005, 87, 695–703.
33. J. S. Hale, M. DeVries, B. Dworak and J. A. Woollam, Thin Solid Films, 1998, 313, 205–209.
34. Y. G. Mo, R. O. Dillon and P. G. Snyder, J. Vac. Sci. Technol., A, 1999, 17, 2933–2938.
35. F. P. Koffyberg, K. Dwight and A. Wold, Solid State Commun., 1979, 30, 433–437.
36. P. Jin and S. Tanemura, Jpn. J. Appl. Phys., Part 1, 1995, 34, 2459–2460.
37. H. Kakiuchida, P. Jin and M. Tazawa, Sol. Energy Mater. Sol. Cells, 2008, 92, 1279–1284.
38. Y. Yin, C. Y. Lan, H. Y. Guo and C. Li, ACS Appl. Mater. Interfaces, 2016, 8, 3861–3867.
39. W. Burkhardt, T. Christmann, B. K. Meyer, W. Niessner, D. Schalch and A. Scharmann, Thin Solid Films, 1999, 345, 229–235.
40. Y. F. Gao, S. B. Wang, H. J. Luo, L. Dai, C. X. Cao, Y. L. Liu, Z. Chen and M. Kanehira, Energy Environ. Sci., 2012, 5, 6104–6110.
41. C. h. Griffith and H. K. Eastwood, J. Appl. Phys., 1974, 45, 2201–2206.
42. K. Okimura, T. Watanabe and J. Sakai, J. Appl. Phys., 2012, 111, 6.
43. Y. X. Ji, S. Y. Li, G. A. Niklasson and C. G. Granqvist, Thin Solid Films, 2014, 562, 568–573.
44. M. Demeter, M. Neumann and W. Reichelt, Surf. Sci., 2000, 454, 41–44.
45. G. Silversmit, D. Depla, H. Poelman, G. B. Marin and R. De Gryse, J. Electron Spectrosc. Relat. Phenom., 2004, 135, 167–175.
46. O. Y. Khyzhun, J. Alloys Compd., 2000, 305, 1–6.
47. F. Y. Xie, L. Gong, X. Liu, Y. T. Tao, W. H. Zhang, S. H. Chen, H. Meng and J. Chen, J. Electron Spectrosc. Relat. Phenom., 2012, 185, 112–118.
48. Y. C. Xin, H. J. Zhou, X. J. Ni, Y. Pan, X. L. Zhang, J. Y. Zheng, S. H. Bao and P. Jin, RSC Adv., 2015, 5, 57757–57763.

---