Transparent and flexible Sb-doped SnO2 films with a nanoscale AgTi alloyed interlayer for heat generation and shielding applications

Transparent and flexible Sb-doped SnO2 (ATO) films with a nanoscale AgTi alloyed interlayer were fabricated for use as plasma damage-free, indium-free, thermally stable electrodes for high performance heat generating films and shielding films in smart windows. The AgTi alloy-inserted ATO film on a PET substrate showed a low sheet resistance of 6.91 ohm per square and a high optical transmittance of 90.24% without thermal annealing or intentional substrate heating. Even after deformation using an outer bending radius of 4 mm, the ATO film with a AgTi interlayer showed a constant sheet resistance due to the mechanical robustness of the AgTi interlayer. Furthermore, the AgTi-inserted ATO film showed a constant resistance even after annealing at 500 °C, unlike the Ag-inserted ATO films. Furthermore, we demonstrated the feasibility of the AgTi-inserted ATO films as transparent heat generating films and shielding films for smart windows. The effective heat generation and shield performance of the ATO/Ag–Ti/ATO multilayer suggests that the multi-functional ATO/Ag–Ti/ATO films can be used to create energy-efficient smart windows for building energy management systems and automobiles.


Introduction
The rapid development of building energy management systems (BEMSs) requires high performance heat generating lms and heat shielding lm-based smart windows. [1][2][3][4][5] In addition, automobiles require high performance heat generating and heat shielding windows for removing ice or frost on the window and blocking sunlight from outside. Metal-based semi-transparent lms coated on exible substrates or glass substrates have been employed to fabricate simultaneous heat generating and heat shielding lms. 6,7 Sputtered or evaporated Ag lms on exible substrates have mainly been used because Ag-based lms generate high temperatures through resistive heating by obtaining power from heat absorbed at near infrared (NIR) wavelengths. However, Ag-based heat generating and shielding lms have very low optical transmittance due to the inherent opacity of the metal lm. 7,8 Therefore, it is important to develop highly transparent and exible metalbased electrodes for heat generating and shielding lms to replace conventional Ag-based semitransparent electrodes. Due to the absence of metal-based transparent electrodes, Agbased semi-transparent electrodes were applied in heat shielding lms, and a Sn-doped In 2 O 3 (ITO) lm was applied in a heat generating lm. 7,[9][10][11] Metallic transparent electrodes with metallic conductivity, high optical transmittance comparable to ITO lms, and outstanding mechanical exibility are required to realize heat generating and shielding lms for smart windows. Although various transparent electrode materials lms have been reported as transparent electrodes for thin lm heaters, such as conducting polymers, carbon (carbon nanotubes, graphene, graphene oxide), Ag nanowires, Cu nanowire, metal meshes, metal network, and oxide-metal-oxide, there are no reports of metallic transparent electrodes that can simultaneously be used in thin lm heat generating and heat shielding lms.  Recently, we reported semi-transparent Ag-Cu electrodes that could be applicable in thin lm heaters and heat shielding lms simultaneously, but the low optical transmittance of Ag-Cu alloy lm remained a critical problem. 7 In this study, we investigated the electrical, optical, morphological, and mechanical properties of ATO lms on PET substrates with a thermally evaporated Ag-Ti alloy layer for use as transparent, exible, thermally stable electrodes for heat generating and shielding lms in smart windows. Based on gure of merit values, the optimal ATO thickness was determined for highly transparent and conductive electrodes. In addition, the mechanical exibility of the Ag-Ti-inserted ATO lms were tested using specially designed bending testers. Furthermore, we demonstrated the feasibility of thermally evaporated ATO/Ag-Ti/ATO lms for high performance simultaneous heat generating lms and heat shielding lms.

Thermal evaporation of ATO/Ag-Ti/ATO lms
Highly transparent and exible ATO/Ag-Ti/ATO lms were thermally evaporated on PET substrates at room temperature without intentional substrate heating using a thermal evaporation system (NNS Vacuum 15NNS005). The ATO powder source and 95 wt% Ag 5 wt% Ti alloy source were loaded in a tungsten boat, and the evaporation chamber was evacuated to 1 Â 10 À6 torr. First, the bottom ATO layer was thermally evaporated on a PET substrate at a voltage of 0.48 V, a current of 53 A, a Z-factor of 0.724, and a tool factor of 128%. During thermal evaporation of the ATO lm, the PET substrate was constantly rotated at a speed of 10 rpm. Then, the metallic Ag-Ti interlayer was continuously evaporated onto the bottom ATO layer at a voltage of 0.34 V, a current of 55 A, a Z-factor of 0.529, and a tool factor of 155%. Aer evaporation of the Ag-Ti alloy interlayer, the top ATO layer was nally evaporated on the Ag-Ti interlayer at the same evaporation conditions as the bottom ATO layer. The continuous evaporation process is illustrated in Fig. 1a, and a picture of the optimized transparent and exible ATO/Ag-Ti/ATO sample is shown in Fig. 1b. The thickness of the ATO and Ag-Ti layers was precisely controlled using a thickness monitor in the thermal evaporation system.

Characterization of the ATO/Ag-Ti/ATO lms
The electrical properties of the thermal evaporated ATO/Ag-Ti/ ATO lms were examined using Hall measurements (HL5500PC, Accent Optical Technology) as a function of ATO thickness. The optical transmittance of the ATO/Ag-Ti/ATO multilayer was measured using a UV/visible spectrometer (UV 540, Unicam). The thermal stability of Ag-Ti and Ag interlayers in ATO lms was evaluated by comparing the sheet resistance aer rapid thermal annealing from 300 to 600 C. The resistance change in the ATO/Ag-Ti/ATO multilayer was measured during inner/outer bending and twisting tests to demonstrate the exibility of the multilayer. In addition, dynamic fatigue bending at a xed bending radius of 10 mm and twisting tests at a xed twisting angle of 15 were performed using a lab-made cyclic bending and twist test machine operated at a frequency of 1 Hz for 100 000 cycles. Field emission scanning electron microscopy (FESEM) was employed to investigate the surface morphology of the ATO/Ag-Ti/ATO multilayers before and aer bending tests. In addition, microstructure and interface of the ATO/Ag-Ti/ATO multilayer were investigated by transmission electron microscopy (TEM) examination. The TEM sample was prepared by cutting of sample, polishing, and ion milling with liquid nitrogen cooling.

Fabrication and evaluation of heat generating lms and heat shielding lms
To demonstrate that thermally evaporated ATO/Ag-Ti/ATO multilayers can be used for exible and transparent heat generating and shielding lms, we fabricated a typical thin lm heater with a size of 25 Â 25 mm 2 and a heat shielding lm with a size of 80 Â 80 mm 2 (Fig. 1b). Two terminal Ag contact electrodes were sputtered on the edge of the ATO/Ag-Ti/ATO multilayer for ATO/Ag-Ti/ATO-based thin lm heaters. A DC voltage was supplied by a power supply (OPS 3010, ODA technologies) to the ATO/Ag-Ti/ATO lms through an Ag contact electrode at the lm edge. The temperature of ATO/Ag-Ti/ATObased heater was measured using a thermocouple mounted on the surfaces of the transparent electrode and an IR thermal imager (A35sc, FLIR). The heat shielding performance of the ATO/Ag-Ti/ATO multilayer was examined using a halogen lamp that irradiated the ATO/Ag-Ti/ATO multilayer while the temperature inside an automobile (shown in Fig. 1b) was measured. An IR thermometer (62-MAX, Fluke) was used to measure the inside temperature of a model automobile with a bare PET window and an ATO/Ag-Ti/ATO/PET window. The ATO/Ag-Ti/ATO multilayers used as a thin lm heater and a heat shielding lm were identical samples prepared at the same evaporation conditions.    Fig. 2a. The increased resistivity could be attributed to the low carrier mobility regardless of ATO thickness as shown in Fig. 2b. The carrier concentration injected from the metal Ag-Ti interlayer decreased due to an increase in the top and bottom ATO layer volume at a constant Ag-Ti thickness of 12 nm. This increased the resistivity of the ATO/Ag-Ti/ATO multilayer lm. Fig. 2c shows the optical transmittance of the ATO/Ag-Ti (12 nm)/ATO multilayer lm with increasing top and bottom ATO layer thicknesses. The ATO/Ag-Ti/ATO multilayer lm with 40 nm-thick top and bottom ATO layers had the highest optical transmittance of 90.24% at a wavelength of 550 nm. The absorption edge of the ATO/Ag-Ti/ATO multilayer shied to longer wavelengths as the top and bottom ATO layer thicknesses increased. At a xed Ag-Ti thickness of 12 nm, the optical transmittance and absorption edge of the multilayer are inuenced by the top and bottom oxide thicknesses. [34][35][36] Although the ATO/Ag-Ti/ATO multilayer showed a high optical transmittance in the visible wavelength region (400-800 nm), the optical transmittance of the ATO/Ag-Ti/ATO multilayer in the near IR region was also low due to the existence of a metallic Ag-Ti interlayer. The low near IR transmittance of the ATO/Ag-Ti/ATO multilayer makes it suitable for use as a heat shielding lm to prevent heat ow through a window. Although the ATO lm with 14 nm thick Ag-Ti interlayer showed lower sheet resistance (4.82-6.93 ohm per square) regardless of the ATO thickness, an increase in Ag-Ti interlayer thickness led to reduction of optical transmittance (87.83%). To apply ATO/Ag-Ti/ATO multilayer as transparent electrode, higher optical transmittance is very important. Therefore, we selected the optimal thickness of Ag-Ti interlayer as 12 nm. As shown in Fig. 2d, gure of merit (FOM) values were calculated as a function of the thickness of the top and bottom ATO layers based on the sheet resistance and optical transmittance at a wavelength of 550 nm for the ATO/Ag-Ti/ATO multilayer. The ATO/Ag-Ti/ ATO multilayer lm with a 40 nm-thick ATO layer exhibited the highest FOM value of 51.83 Â 10 À3 ohm À1 due to a low sheet resistance of 6.91 ohm per square and a high optical transmittance of 90.24% at a wavelength of 550 nm. Therefore, we determined that the optimal thicknesses of ATO and the Ag-Ti layer were 40 and 12 nm, respectively.
To investigate microstructure of thermal evaporated ATO and Ag-Ti interlayer, TEM examination was carried out. Fig. 3a showed a cross-sectional TEM image of the ATO/Ag-Ti/ATO multilayer grown on PET substrate. Due to exact thickness control during thermal evaporation, the multilayer showed symmetric structure with clearly distinguished layers. It was clearly shown that the metallic Ag-Ti interlayer connected the top ATO and bottom ATO layer and acts as main conduction path. Because the thermal evaporation process was carried out at room temperature, the multilayer had well-distinguished interlayer without interfacial layer. The enlarged TEM image in Fig. 3b indicates the bottom ATO obtained from "A" region in the cross-sectional TEM image. The thermal evaporated ATO layer showed a typical amorphous structure with short-rangeorder. In addition, diffuse fast Fourier transform (FFT) pattern in the inset of Fig. 3b indicated that thermal evaporated ATO layer had a typical amorphous structure unlike conventional FTO or ATO lms grown by chemical vapor deposition. Fig. 3c showed enlarged Ag-Ti interlayer obtained from "B" region in the cross-sectional TEM image. As indicated FFT pattern in the inset of Fig. 3c, the thermal evaporated Ag-Ti layer showed a crystalline structure even though it was prepared at room temperature. Fig. 4 showed enlarged TEM images obtained from the interface region between amorphous ATO layer and crystalline Ag-Ti interlayer. Because the thermal evaporation process of bottom ATO, Ag-Ti interlayer, and top ATO layer was continuously conducted without breaking vacuum, there is no interfacial layer between ATO and Ag-Ti interlayer. Both interface (Ag-Ti/bottom ATO and top ATO/Ag-Ti) showed a discrete and well-dened interface between bright amorphous ATO and dark crystalline Ag-Ti interlayer. Unlike sputtered ITO top layer, which is easily crystallized on the crystalline metal interlayer, in sputtered ITO/Ag/ITO multilayer, the evaporated top ATO layer in Fig. 4a showed identical amorphous structure to the bottom ATO layer in Fig. 4b due to low energy of evaporated ATO atoms. 37 The mechanical exibility of the thermally evaporated optimal ATO/Ag-Ti/ATO multilayer lm was investigated using a lab-designed inner and outer bending test system. Fig. 5a shows the results of the outer/inner bending tests for the ATO/ Ag-Ti/ATO multilayer electrodes with decreasing outer/inner bending radius. We measured the resistance change (DR) of the ATO/Ag-Ti/ATO multilayer during substrate outer/inner bending to determine the critical bending radius. R 0 represents the initial resistance, and R indicates the measured resistance. The outer/inner bending test result in Fig. 5a shows that the evaporated ATO/Ag-Ti/ATO (40/12/40 nm) multilayer had a constant resistance until the bending radius reached 5 mm. The following equation can be used to calculate the peak strain for a curved ATO/Ag-Ti/ATO multilayer with decreasing bending radius: 38 here, d f and d PET are the thicknesses of the ATO/Ag-Ti/ATO multilayer and the PET substrate, respectively. The ATO/Ag-Ti/ATO multilayer lm (40/12/40 nm) on a 125 mm-thick PET substrate experienced a peak strain of 1.25% at a bending radius of 5 mm. The resistance of the ATO/Ag-Ti/ATO multilayer rapidly increased beyond the critical bending radius due to separation of cracked lms that experienced tensile stress. In the inner bending tests, the ATO/Ag-Ti/ATO multilayer showed no change in resistance even though cracks formed on the sample. Under the compressive stress illustrated in the inset picture of Fig. 5a, the cracked ATO/Ag-Ti/ATO lms overlapped  or physically touched. Fig. 5b shows the dynamic outer and inner bending fatigue test results for the ATO/Ag-Ti/ATO multilayer with increasing bending cycles at a xed bending radius of 10 mm and a repeating rate of 1 Hz. Both dynamic outer and inner bending fatigue tests showed no change in resistance (DR) during 100 000 bending cycles, demonstrating the good exibility of the ATO/Ag-Ti/ATO multilayer lm. We employed a twisting test as another mechanical exibility test. The inset pictures of Fig. 5c show images of the lab-designed twisting test steps. Twisting the sample allowed us to investigate the mechanical stability of the transparent electrode. Fig. 5c exhibits the twisting bending test results of the ATO/Ag-Ti/ATO multilayer at a constant twisting angle of 15 . Over 100 000 repeated twisting tests, the ATO/Ag-Ti/ATO multilayer showed no change in resistance due to the outstanding exibility of ATO/Ag-Ti/ATO multilayer. Overall, the electrical and optical analysis results and the mechanical tests showed that insertion of a metallic Ag-Ti layer into the ATO layer is an effective way to reduce the resistivity and improve the mechanical exibility of the evaporated ATO lms. Fig. 6 shows surface FESEM images of ATO/Ag-Ti/ATO (40/ 12/40 nm) multilayers before and aer outer/inner bending, dynamic fatigue, and twisting tests. Surface FESEM images (Fig. 6a) of as-evaporated ATO/Ag-Ti/ATO lms showed a smooth and featureless morphology because the evaporation occurred at room temperature. Fig. 6b shows surface FESEM images of the cracked ATO/Ag-Ti/ATO multilayer aer outer and inner bending at a bending radius of 1 mm, which is beyond the critical bending radius. Several cracks were observed on the ATO/Ag-Ti/ATO multilayer because the outer bending test applied severe tensile stress to the lms. The enlarged FESEM image shows that the cracks on the sample physically separated the ATO/Ag-Ti/ATO multilayer lm and increased the measured resistance change during outer bending. Even though the sample showed no resistance change during inner bending in Fig. 5a, there were also several cracks in the multilayer. Unlike the outer bending test, the cracks of inner bending test have a 'hill' near the edge due to overlaps in the cracked region during the inner bending test. This physically overlapped region could conduct current during the inner bending test, so the inner bent sample showed a constant resistance change even though the sample was severely cracked. There were no cracks in the multilayer aer the dynamic fatigue tests shown in Fig. 6c and d because they experienced repeated stress at a constant bending radius and angle below the critical bending angle and radius. The ATO/Ag-Ti/ATO multilayer showed a similar surface FESEM image to the as-deposited sample even aer 100 000 dynamic outer/inner bending and twisting cycles, as shown in Fig. 6c and d. The outstanding mechanical exibility of the ATO/Ag-Ti/ATO multilayer indicates that the thermally evaporated ATO/Ag-Ti/ATO multilayer lm is a promising exible and transparent electrode for exible heat generating lms and shielding lms used in curved or shaped smart windows.
To illustrate the thermal stability of the Ag-Ti inserted ATO lm, we compared the resistance change of Ag-inserted ATO and Ag-Ti-inserted ATO lms as shown in Fig. 7. The resistance of the Ag-inserted ATO lms abruptly increased aer rapid thermal annealing at 500 C. This increased resistance was due to agglomeration and out-diffusion of Ag though the grain boundaries of crystalline ATO layers, as we reported in our previous work. 37 The surface FESEM image in Fig. 7a showed a crystallized ATO top layer and agglomeration of Ag on the top of the ATO layer. However, the Ag-Ti-inserted ATO lm showed a smaller resistance change aer rapid thermal annealing at 500 C due to the thermal stability of the Ag-Ti interlayer, as shown in Fig. 7b. Surface FESEM images also showed only crystallized ATO top layers without agglomeration of the metal layer. The thermal stability of the Ag-Ti-based ATO electrode indicates that the Ag-Ti-inserted ATO is an appropriate electrode for heat generating lms.
ATO/Ag-Ti/ATO multilayer lm-based thin lm heaters were fabricated to investigate the feasibility of the thermal evaporated ATO/Ag-Ti/ATO multilayer as a exible and transparent electrode for heat generating lms in smart windows. The fabrication process is illustrated in Fig. 8a. As shown, the ATO/ Ag-Ti/ATO-based thin lm heater had a two-terminal contact electrode at the edges of the devices. To generate heat in the lms, a DC voltage was applied to the thin lm heaters by a power supply through an Ag metal contact electrode at the lm edge. The temperature of the ATO/Ag-Ti/ATO-based thin lm heaters was measured in situ using a thermocouple in direct contact with the surface of the thin lm heaters. Fig. 8b shows the temperature proles of the ATO/Ag-Ti/ATO-based thin lm heaters at different input DC voltages as a function of ATO layer thickness. When a DC voltage was supplied to the ATO/Ag-Ti/ATO-based thin lm heaters, the temperature of the thin lm heater rapidly increased and reached a saturation  temperature. It is noteworthy that the increase in the ATO layer thickness of the thin lm heater increased the DC voltage required to obtain an identical temperature. At an identical input voltage, the ATO/Ag-Ti/ATO-based thin lm heater with a thinner ATO layer showed a higher saturation temperature. Fig. 8c shows the DC voltage of the thin lm heater required to obtain a temperature of 100 C with increasing ATO thickness. An increase in ATO layer thickness led to an increase in the DC voltage required to reach the saturation temperature of 100 C because the thinner ATO-based electrode had a lower sheet resistance. The higher saturation temperature of the thin lm heater with a thinner ATO layer indicates efficient transduction of electric energy into Joule heating in ATO/Ag-Ti/ATO electrodes. As we described by Huang et al., the saturation temperature can be can be expressed as: 39 here, V is the applied voltage, R is the resistance of thin lm heater, Q d is heat dissipation, C is heat capacity of the lm, m is lm mass, and T s and T room are the saturation and room temperature, respectively. Because main heat dissipation path is air convection, temperature of thin lm heater can be saturated at T s when Joule heating and air convection reached a dynamic balance. 7,36 It is apparent from eqn (2a) that the saturation temperature of the thin lm heater increases with increasing input DC voltage (V) and decreasing resistance (R). Therefore, a lower sheet resistance in the ATO/Ag-Ti/ATO electrode with a thinner ATO layer is imperative for fabrication of high performance thin lm heaters with a lower DC input voltage. Fig. 8d shows measured temperature of thin lm heaters with different ATO thickness as a function of input power. Regardless of the top and bottom ATO thickness, all thin lm heaters showed linearly increased temperature with increasing input power. Fig. 9a shows the repeated heating and cooling prole of the ATO/Ag-Ti/ATO (40/12/40 nm)-based thin lm heaters over 10 cycles. The ATO/Ag-Ti/ATO-based thin lm heater showed identical heating-cooling proles and rapidly reached a saturation temperature of 100 C when a DC voltage of 4.6 V was applied. In addition, when a DC voltage of 4.6 V was supplied to the ATO/Ag-Ti/ATO-based thin lm heater for 1 hour, the thin lm heater maintained a saturation temperature of 100 C without temperature modulation, as shown in Fig. 9b. The similar surface FESEM images of the ATO/Ag-Ti/ATO-based thin lm heater before and aer heating also indicate the durability of the ATO/Ag-Ti/ATO multilayer as an electrode for exible and transparent thin lm heaters.
A water droplet test was performed to demonstrate the use of ATO/Ag-Ti/ATO-based thin lm heaters in smart windows and automobiles. Fig. 10a shows pictures and IR images of the water droplet test of the ATO/Ag-Ti/ATO-based exible and transparent thin lm heater with a saturation temperature of 102.8 C. When a DC input voltage of 4.6 V was supplied to the thin lm heater, a saturation temperature of 102.8 C was instantly achieved due to a dynamic balance between Joule heating and convection. The IR image indicates the uniformly heated ATO/Ag-Ti/ATO-based thin lm heater surface. At the saturation temperature, the water droplet disappeared immediately due to the high temperature of the ATO/Ag-Ti/ ATO-based thin lm heater. This result demonstrated that ATO/Ag-Ti/ATO-based thin lm heaters can be used as selfwarming windows. Fig. 10b shows the defrost test of the ATO/Ag-Ti/ATO-based thin lm heater before and aer frost formation. At an operating DC voltage of 4.6 V, the frost on the surface of the ATO/Ag-Ti/ATO electrode completely disappeared. Therefore, the ATO/Ag-Ti/ATO-based thin lm heater could be a transparent and exible defrost lm for automobiles.
Because solar shading and heat shielding windows are need in smart buildings and automobiles, smart windows must transmit visible light and shield IR radiation. To enable smart windows to balance the transmittance and sunlight-energyshielding, it is necessary to form cut-off lters that transmit visible light and reect near-infrared (NIR) light, as shown in Fig. 11a. Moreover, because heat shielding on the indoor side of the window is produced by reecting far-infrared light, which is a warming radiant heat energy, the reection of far-infrared light is also required. Fig. 11b shows the transmission and reection spectrum of the ATO/Ag-Ti/ATO multilayer in the visible to the near-infrared light region. The multilayer shows high optical transmittance in the visible wavelength region and high reection in the NIR wavelength region, which is appropriate for heat shielding lms. The ATO/Ag-Ti/ATO multilayer ensures good lighting with 90% visible light transmittance and high solar shading. The total energy consumption in a building or automobile can be effectively controlled by effective shielding of heat though a ATO/Ag-Ti/ATO lm attached to a smart window. Fig. 12a and b shows a heat shielding test of a bare PET substrate and an ATO/Ag-Ti/ATO multilayer-coated PET, respectively. Irradiating glass with light from a halogen lamp resulted in an increase in temperature. The surface temperature of the glass was measured in situ using a portable IR thermometer. The surface of the glass without the heat shielding lm showed a temperature of 55.9 C under halogen lamp irradiation as shown in Fig. 12a. When the light was blocked by a bare PET substrate, there was a small decrease in temperature (53.3 C). However, when the radiation was blocked by a transparent ATO/Ag-Ti/ATO multilayer lm, the surface temperature of the glass substrate was signicantly reduced from 55.9 to   37.1 C. Inside the automobile model, the temperature was 47.1 C under halogen lamp irradiation without heat shielding lm (shown in Fig. 12b). When the light was blocked by bare PET in a front window, there was a small decrease in temperature (45.8 C). However, when the radiation was blocked by a transparent ATO/Ag-Ti/ATO multilayer in the front window, the indoor temperature of the car was signicantly reduced from 47.1 C to 34.6 C, indicating that the ATO/Ag-Ti/ATO multilayer was effective in heat shielding due to its high reection in the NIR region. Therefore, the thermally evaporated ATO/Ag-Ti/ATO multilayer can be used as a heat generating and heat shielding lm simultaneously in automobile windows or smart windows for BEMS.

Conclusions
We developed thermally evaporated ATO/Ag-Ti/ATO multilayers for use as exible and transparent lms for transparent and exible heat generating lms and heat shielding lms. To optimize the thickness of ATO lm in the multilayer, the effect of ATO thickness on the sheet resistance and optical transmittance of the ATO/Ag-Ti/ATO lms was examined. Based on FOM values calculated from the sheet resistance and optical transmittance of the ATO/Ag-Ti/ATO multilayer, we obtained an optimized ATO/Ag-Ti/ATO multilayer lm with a sheet resistance of 6.91 ohm per square and an optical transmittance of 90.24%. In addition, the thermally evaporated ATO/Ag-Ti/ATO electrode showed a small critical outer/inner bending radius and outstanding exibility. Moreover, the thermal stability of the Ag-Ti interlayer was compared with a typical Ag interlayer. The time-dependent temperature prole of the thin lm heater with the ATO/Ag-Ti/ATO electrodes demonstrated that the thermally evaporated multilayer is a promising transparent electrode for high performance thin lm heaters for smart windows and automobiles. Furthermore, we investigated the heat shielding performance of the ATO/Ag-Ti/ATO multilayer for use as an energy efficient solar shading lm on smart window and automobiles. The simultaneous heat generation and heat shield performance of the optimized ATO/Ag-Ti/ATO multilayer indicated that it is feasible to use this ATO/Ag-Ti/ ATO multilayer to create energy-efficient automobile windows and smart windows for BEMS.

Conflicts of interest
There are no conicts to declare.