A roll-to-roll process for multi-responsive soft-matter composite films containing Cs_xWO₃ nanorods for energy-efficient smart window applications

Abstract

This work provides a roll-to-roll processed flexible multi-responsive smart film containing tungsten bronze nanorods and a phase-separated liquid crystal—polymer composite, which can reversibly control the passage of visible light in response to temperature, an electric field and near infrared light, and also screen the heat rays from 800 nm to 2500 nm.

---

Conceptual insights

Smart windows that have functionalities of shielding near infrared (NIR) rays and controllable visible light transmittance can not only conserve electric energy as a result of the conditioner operation, but also protect privacy and prevent indoor members from intensive light stimuli. Herein we present a novel roll-to-roll production of a multi-responsive soft-matter composite smart film that can not only control the passage of visible light by temperature, an electric field and NIR light, but also screen more than 95% of the NIR irradiation from 800 nm to 2500 nm. These excellent properties of the film were achieved by creating a compatible poly(vinylpyrrolidone) (PVP) tuning layer between tungsten bronze (Cs_xWO₃) nanorods (NRs) and a polymeric syrup containing liquid crystals with a smectic A (SmA) to chiral nematic (N*) phase transition, followed by forming an elaborated–designed polymer structure within the film. Besides, the film also exhibits a wide working temperature range, excellent flexibility, high mechanical strength, long-term stability and large-area processability. The as-prepared film shows great potential for application as a smart energy-saving window, and the proposed strategy provides new insights into engineering novel hybrid inorganic–organic functional materials.

Buildings account for at least 40% of energy use in most countries, and this energy consumption is still increasing rapidly, as the construction industry booms, especially in developing countries.^1,2 Among all the components of a building, windows are often regarded as a less energy efficient building envelope with a larger maintenance requirement.³ The calculations show that more than 50% of the energy used in lighting, heating and cooling can be saved by deploying better control systems such as smart windows to control the influx of radiant heat transfer.⁴ Therefore, great efforts have been devoted to developing smart windows to replace the traditional way of sun-shading.^5–15

Several aspects must be considered for smart windows to be used as an energy-efficient building envelope in areas where human inhabitants employ them. Of the utmost importance is their optical modulation range in the whole solar spectrum ranging from visible to NIR light (800–2500 nm).^3,16,17 The invisible NIR light, especially the shorter wavelength range from 800 to 1500 nm, carries about 50% of the whole solar energy,¹⁸ and therefore the shielding performance of a smart window in this wavelength range is crucial for the purpose of improving the energy efficiency.^19,20 Equally, the regulation of visible light transmittance is also desirable for the purposes of blocking the intensive light on scorching days from outside and protecting privacy.^1,21 However, although our recent work demonstrated a broadband optical modulation within a flexible smart film,²² it is still a great challenge to combine the regulation of visible light transmittance and the shielding performance of NIR light from 800 nm to 2500 nm within a single material. Apart from optical modulation, other key factors including large-scale processability,²² long-term stability,²³ good flexibility,²⁴ high mechanical strength²⁵ and functionality over a range of exterior temperatures² should also be taken into account in material design for practical applications. Unfortunately, materials for smart windows integrated with all the above features have not been reported to the best of our knowledge.

Herein, we present a facile roll-to-roll production of an energy-efficient flexible smart film that can not only control the transmittance of visible light by temperature, electric fields and NIR light stimuli, but also screen more than 95% of NIR irradiation from 800 nm to 2500 nm. Our material for the smart film has a unique phase-separated polymer structure with a thickness of about 20 μm in which LCs with a smectic (SmA) to chiral nematic (N*) phase transition (PT-LCs) and tungsten bronze (Cs_xWO₃) nanorods (NRs) are homogeneously dispersed. It is essentially a LC/polymer composite; unlike traditional polymer dispersed liquid crystal (PDLC) materials, which exhibit an opaque appearance in the off state,^26–28 our smart film is transparent under normal conditions. Also, the high content of polymer in our material provides excellent mechanical strength, much higher than traditional polymer stabilized liquid crystal (PSLC) materials.^29,30 Moreover, the polymeric syrup containing highly compatible organic and inorganic components allows our material to be facilely processed via roll-to-roll production and thus realize large-scale manufacturing. Finally, a simulated system was used to characterize the indoor temperature regulation performance of this multi-responsive smart composite material, showing that the as-made film was suitable for use as an energy-efficient smart window.

The procedures for making the smart film are illustrated in Scheme 1. First, a homogeneous polymeric syrup containing PT-LCs, liquid-crystalline polymerizable monomers (LPMs), non-liquid-crystalline polymerizable monomers (NLPMs) and Cs_xWO₃ NRs were sandwiched between two plastic substrates via a roll-to-roll process, as shown in Scheme 1a and b. Then, in Scheme 1c, the film was irradiated under UV light for a short time to initiate a partial polymerization of the NLPMs and LPMs, leading to the formation of the phase-separated composite with a preliminary porous structure (Scheme 1d). In this step, the curing time should be precisely controlled within 90 seconds; otherwise, most of the LPMs would be consumed. After that, an electric field was applied across the film to homeotropically align the PT-LCs and LPMs (Scheme 1e and f). Meanwhile, a second-step of UV irradiation was carried out (Scheme 1g) to make the rest of the NLPMs and LPMs fully polymerized, forming oriented liquid-crystalline polymer networks (OLPNs) within the porous polymer matrix (Scheme 1h). After this two-step UV polymerization, a flexible multi-responsive energy-efficient smart film was prepared. When the environmental temperature is lower than the phase transition temperature of the PT-LCs, the PT-LCs in the SmA phase in the porous polymer are perpendicularly oriented to the substrate and stabilized by the OLPNs within the LC domains. The film exhibits a transparent state due to the well-matched refractive index between the PT-LCs and polymer. When the temperature surpasses the phase transition temperature of the PT-LCs, a focal-conic texture of the heat-induced N* phase of the PT-LCs is spontaneously formed, and the film becomes opaque due to the spatially varied refractive index between the PT-LCs and polymer. Also, this opaque state can be switched back into the transparent state either by decreasing the temperature into the SmA range (homeotropically aligned SmA phases would form again induced by the OLPNs) or by actively applying an electric field to homeotropically align the PT-LCs in the N* phase. Moreover, more than 95% of the NIR irradiation from 800 nm to 2500 nm in the solar spectrum can be effectively screened by the film due to the strong and broadband LSPR absorption of Cs_xWO₃ NRs. Meanwhile, the absorbed NIR energy can be immediately transformed into heat to initiate the SmA to N* phase transition, meaning that the passage of visible light in our film can not only be controlled by temperature and electric fields, but also NIR light stimuli.

Scheme 1 Schematic illustration of the procedures for making the smart film. (a and b) A homogenous polymeric syrup is sandwiched between two conductive substrates via a roll-to-roll process. (c and d) Then, the first-step of UV polymerization is carried out to initiate a partial phase separation between LCs and polymer. (e and f) After that, the LCs are homeotropically aligned by applying an electric field. (g and h) Meanwhile, the second-step of UV polymerization is carried out to complete the crosslinking among LPMs within LC domains. The as-prepared film can reversibly change between transparent (i) and opaque states (j) according to temperature. Also, the opaque state can be switched to transparency via an electric field (k).

Cs_xWO₃ NRs were synthesized using a previously reported method.³¹ They exhibit a rod-like morphology. The diameter and length of these NRs are 12 ± 2 nm and 68 ± 5 nm, respectively (Fig. 1a). Energy dispersive X-ray spectroscopy (EDS) mapping (Fig. 1b) demonstrates the existence of Cs, W, and O elements in the products, and the Cs/W atomic ratio is determined to be 0.32. The X-ray diffraction (XRD) pattern in Fig. 1c shows the hexagonal structure of Cs_0.32WO₃ (JCPDS No. 831334), without impurity peaks. XPS results suggest the mixed bonding state of tungsten element. The main peaks with a W 4f 5/2 at 37.3 eV and a W 4f 7/2 at 35.2 eV can be ascribed to W⁺⁶, while the second doublet with a lower binding energy at 34.4 and 36.5 eV is attributed to the W 4f 5/2 and W 4f 7/2 core levels from W⁺⁵. This mixed W⁺⁵ and W⁺⁶ is agreement with that of Cs_xWO₃ tungsten bronze, expressed as Cs_xW_x⁺⁵W_1−x⁺⁶O₃.²⁶

Fig. 1 (a) TEM image of the as-prepared Cs_xWO₃ NRs. Inset: HRTEM image of a single Cs_xWO₃ NR. (b) EDS mapping images, (c) XRD pattern and (d) fitted XPS spectra at the W 4f core level of the as-made Cs_xWO₃ NRs. (e) Scheme of PVP coating procedures for the as-made Cs_xWO₃ NRs. (f) Tyndall effect of the Cs_xWO₃/PVP NRs dispersed in the polymeric syrup. (g) FT-IR spectra of PVP, Cs_xWO₃ NRs and Cs_xWO₃/PVP NRs. (h) Photographs of the Cs_xWO₃ NRs and Cs_xWO₃/PVP NRs dispersed in different media. (i) Absorption spectra of the Cs_xWO₃/PVP NRs dispersed in ethanol.

The as-made Cs_xWO₃ NRs were found to be well-dispersed in water because of their high zeta-potential value of −67.55 ± 0.92 mV; however, they aggregate spontaneously after being introduced into the polymeric syrup containing LCs and polymerizable monomers likely because of the compression of the surface electric double layer, which not only results in a loss of transparency in the visible region, but also detrimentally affects the NIR shielding performance of the Cs_xWO₃ NRs, which resulted from a significant reduction in the surface electron density in large aggregates. Thus, to render these NRs well incorporated into this organic polymeric system, we promote an facile strategy by coating the surface of the Cs_xWO₃ NRs with a layer of poly(vinylpyrrolidone) (PVP) as a tuning agent. The polar functional groups such as the amide of the coated PVP layer can make the interface between Cs_xWO₃ NRs and the polymeric syrup highly compatible since the syrup containing LCs and polymeric monomers can be viewed as a solvent with high polarity.

As shown in Fig. 1e, the pristine Cs_xWO₃ NR powders were first dispersed in deionized water under ultrasonication to form a homogeneous dispersion. After that, a trace amount of PVP aqueous solution was introduced into the dispersion under continuous stirring for 24 h to allow a full absorption of PVP onto the Cs_xWO₃ NRs. Then, after being centrifuged and rinsed with deionized water, these PVP functionalized Cs_xWO₃ (Cs_xWO₃/PVP) NRs can be readily dispersed in the polymeric syrup and other organic solvents such as ethanol and acetone. The Tyndall effect in Fig. 1f indicates the existence of Cs_xWO₃ NRs in the polymeric syrup.

To demonstrate the success of this PVP coating process, an FT-IR spectrum was used to measure the surface chemistry of the Cs_xWO₃/PVP NRs. In Fig. 1g, five distinct peaks of PVP are observed, attributed to the stretch vibrational bands of ν_(O–H) at around 3490 cm⁻¹, the stretch vibrational bands of ν_(C–H) at 2938–2854 cm⁻¹, the stretch vibrational bands of ν_(C=O) at 1655 cm⁻¹, the flexural vibrational bands of δ_(C–H) at 1423 cm⁻¹ and the stretch vibrational bands of ν_(C–N) at 1290 cm⁻¹. In contrast, the unmodified Cs_xWO₃ NRs exhibit no obvious peaks except the stretch vibrational bands of ν_(W–O) at around 784 cm⁻¹. After PVP coating, the distinct peaks of PVP were all detected for the Cs_xWO₃/PVP NRs, indicating that the PVP coating process was successful. These Cs_xWO₃/PVP NRs can be stable for at least one month in the polymeric syrup, while the unmodified Cs_xWO₃ NRs would form large aggregates and be fully precipitated in polar solvents within half an hour (Fig. 1h). Most importantly, those Cs_xWO₃/PVP NRs show intensive and broadband absorption in the NIR wavelength range from 800 nm to 2500 nm, but only a slight absorption in the visible light range 400 nm to 800 nm, exhibiting promising performance to be used as solar filters, as confirmed by absorption spectra in Fig. 1i.

Obviously, the temperature range (TR) of PT-LCs is a crucial parameter in determining whether or not the as-made film is applicable for use over a range of exterior temperatures, since the TR of most PT-LC monomers is quite narrow. For example, the liquid-crystalline TR of 8CB (Cr 21.5 °C SmA 33.5 °C N 40.5 °C I) is only 19 °C. If used, the film would lose its functionality either below 21.5 °C or above 40.5 °C, which cannot be used in rigid environments. For practical applications, an ultra-wide liquid-crystalline TR is absolutely desirable. Here, our strategy for widening the TR was by mixing different LC monomers together to form a PT-LC mixture. The chemical structures of the LC monomers we used are provided in Scheme S1 (ESI†). Here, three PT-LC mixtures with different compositions are given. Differential scanning calorimetry (DSC) measurements in Fig. 2a indicate that they all exhibit an ultra-wide TR of more than 95 °C with a crystallization point below −30 °C and a clear point higher than 67 °C. Moreover, the SmA to N* phase transition temperature can be tuned from below 10 °C to above 40 °C (Fig. 2a), making the film much more designable to meet different preferences or requirements. Also, no impurity peaks were detected, indicating the good miscibility among the LC components. If not specified, we used PT-LC-2 (Cr < −30 °C SmA 31.5 °C N* 67.5 °C I) for next characterization and preparation. Fig. 2b and c show the typical textures of the SmA phase and the N* phase for PT-LC-2 at 25 °C and 35 °C, respectively. Fig. 2d and e show coexistent textures of SmA and N* phases during the heating and cooling processes, respectively. The inset in Fig. 2b indicates the high viscosity of the SmA phase, while the inset in Fig. 2c shows a drastic decrease of the viscosity of the N* phase compared with the SmA phase.

Fig. 2 (a) DSC thermographs of the PT-LCs with different phase transition temperatures. POM images of the optical texture of PT-LCs-2 (b) in the SmA phase, (c) in the N* phase, (d) during the phase transition from SmA to N* and (e) during the phase transition from N* to SmA.

Smart films containing Cs_xWO₃/PVP NRs were prepared by the step-wise polymerization in Scheme 1. Fig. 3a shows the temperature-dependent transmittance for each film with different amounts of Cs_xWO₃/PVP from 0% to 5 wt% at 550 nm. All the as-made films exhibit a transmittance transition from highly transparent to opaque as the temperature surpasses the phase transition temperature of PT-LC-2. We further tested the thermo-optical properties of the films with 0% and 5 wt% Cs_xWO₃/PVP NRs during the heating and cooling processes, respectively. As shown in Fig. 3b and c, the transmittance of visible light in both of the as-made films can be reversibly changed by controlling the temperature. The decreasing trend of transmittance in the transparent state of the films with more Cs_xWO₃/PVP NRs (Fig. 3a and b) can be attributed to a slight absorption of Cs_xWO₃/PVP NRs in the visible range, but it was acceptable as 5 wt% Cs_xWO₃/PVP NRs were incorporated. Furthermore, to test the stability and reversibility of the as-made film, the transmittance of our film at 550 nm was tested during a heating–cooling cycle 200 times. As shown in Fig. 3d, the transmittance of the transparent and opaque states almost reached the same values as their previous ones, demonstrating the total reversibility and stability of our film.

Fig. 3 (a) Transmittance of the smart films containing different amounts of Cs_xWO₃/PVP NRs during the heating process. (b) Transmittance of the smart films with 0 wt% Cs_xWO₃/PVP NRs and 5 wt% Cs_xWO₃/PVP NRs during the heating and cooling processes. (c) Photographs of the smart films containing 0 wt% Cs_xWO₃/PVP NRs and 5 wt% Cs_xWO₃/PVP NRs at 25 °C, partially heated by finger touching and at 40 °C, respectively. (d) Transmittance of the film with 5 wt% Cs_xWO₃/PVP NRs under the heating–cooling cycles 200 times.

This reversible change in transmittance can be attributed to the formation of OLPNs crosslinked between LCMs during the second-step UV polymerization within the porous polymer matrix, as confirmed by scanning electron microscopy (SEM) from both top (Fig. 4c and d) and side views (Fig. 4b). Note that before SEM observation, all the as-prepared films were dipped in cyclohexane for 15 days to fully extract the LC molecules. During the SmA to N* phase transition, the OLPNs within the LC domains induced the orientations of the SmA phases to be homeotropically aligned, making the film become transparent. Otherwise, the SmA phase would form a focal-conic texture, and the film would still be opaque after phase transition. Moreover, the high content of the porous polymer matrix provides both excellent mechanical strength and flexibility to the film. We tested the shearing strength of the as-made films containing 0% and 5 wt% Cs_xWO₃/PVP NRs (Fig. 4e). The shearing force of the as-made film with 5 wt% Cs_xWO₃/PVP NRs reached as high as 306.9 N, comparable to the as-made film without Cs_xWO₃/PVP NRs of the same size, and ∼255 times higher than the traditional PSLC films. Additionally, we also tested the flexibility of the film. In Fig. 4f, the excellent thermo-optical properties of the film were well preserved under folding or bending. Besides, using roll-to-roll production, the film can be manufactured on a large-scale with a size of 1.0 m × 1.5 m (Fig. 4g).

Fig. 4 (a) Schematic of the as-made film and (b and c) SEM photographs of the morphologies of the polymer structure within the films observed from (b) side and (c) top views. (d) Enlarged SEM image of the morphology of the polymer structure observed from a top view. (e) Shearing force–displacement curves of the as-made films with 0% (black) and 5 wt% Cs_xWO₃/PVP NRs (red) and the PSLC film (blue). Inset: The enlarged shearing force–displacement curve of the PSLC film. Digital photographs demonstrating (f) the flexibility of the as-made smart film and (g and h) the large-scale manufacturing of the film.

Due to the strong and broadband LSPR absorption of Cs_xWO₃/PVP NRs in the NIR region from 800 nm to 2500 nm, the NIR irradiation from the solar spectrum can be effectively rejected. We systematically investigated the Vis-NIR transmittance spectra of our film containing different amounts of Cs_xWO₃/PVP NRs in both transparent and opaque states. Obviously, as shown in Fig. 5a and b, the NIR shielding performance of the as-made films was enhanced as more Cs_xWO₃/PVP NRs were introduced. In particular, more than 95% of the NIR irradiation could be screened by incorporating 5 wt% Cs_xWO₃/PVP NRs in both transparent and opaque states. The phase transition from LCs mainly altered the transmittance of the visible light in the film, while the Cs_xWO₃/PVP NRs acted as solar filters to reject NIR light.

Fig. 5 The Vis-NIR transmittance spectra of the films with different concentrations of Cs_xWO₃/PVP NRs measured at 300 K (a) and 308 K (b). (c) Digital photographs of the film containing 5 wt% Cs_xWO₃/PVP NRs with various voltages applied. (d) The voltage dependence of the transmittance for the smart film containing 5 wt% Cs_xWO₃/PVP NRs. (e) Vis-NIR transmittance spectra of the film containing 5 wt% Cs_xWO₃/PVP NRs with various voltages applied.

The electro-optical properties of the film were also investigated since LCs are electric field responsive materials. As shown in Fig. 5c, the film containing 5 wt% Cs_xWO₃/PVP NRs could be actively and dynamically switched from the light scattering to transparent state under an applied voltage of ∼34 V. Fig. 5d shows the voltage dependence of the transmittance of the film. The result reveals that the threshold voltage (V_th) and saturation voltage (V_sat) of the film are 22.3 V and 34.2 V, respectively. Here, the V_th and V_sat values are defined as the voltages required for the transmittance to reach 10% and 90%, respectively. The inset in Fig. 5d shows that the response time of the film is ∼100 ms. Vis-NIR spectra of the film (Fig. 5e) indicates that the applied voltage mainly increased the transmittance in the visible region, but the NIR light could still be effectively rejected due to the absorption of Cs_xWO₃/PVP NRs in the film.

Apart from strong LSPR absorption in the NIR region, the Cs_xWO₃/PVP NRs in the as-made film could also act as nano-heaters to transform the absorbed NIR energy into heat, and initiate the SmA to N* phase transition. In this way, the passage of visible light can also be controlled by NIR light stimuli. To confirm the NIR light-responsive property, the films with different contents (0–5 wt%) of Cs_xWO₃/PVP NRs were irradiated using a NIR lamp (PHILIPS, R125) at the same power density. The temperature variations of the films were tracked using a thermographic camera (Fig. 6a). Thermographic images were taken at different time intervals of 0 s, 5 s, 15 s, 35 s and 55 s. As shown in Fig. S2 (ESI†) and Fig. 6b, the temperature increased more rapidly with a higher content of Cs_xWO₃/PVP NRs, indicating that the Cs_xWO₃/PVP NRs within the films could rapidly and effectively convert NIR energy into heat under NIR light exposure. Furthermore, we also measured the transmittance (550 nm) of the films at different exposure time intervals (Fig. 6c). The film without Cs_xWO₃/PVP NRs shows no obvious optical change under NIR light exposure. However, the transmittance of the films containing Cs_xWO₃/PVP NRs underwent a sharp decrease during the measurements. In particular, the film with 5 wt% Cs_xWO₃/PVP NRs could change from transparent to light-scattering within 5 seconds, exhibiting a fast response time under NIR exposure.

Fig. 6 (a) Schematic illustration of the simulated experiment. (b) Temperature and (c) transmittance variations of the films with different amounts of Cs_xWO₃/PVP NRs under NIR light exposure.

A comparative study was carried out to compare the thermoregulation effects of the as-made smart films containing 5 wt% Cs_xWO₃/PVP NRs and pure PET films (Fig. 7a–d). After a continuous simulated solar irradiation for 15 min, the temperature in the left model house attached with a pure PET film significantly increased from 22.8 °C to 35.7 °C, while the temperature in the right house attached with an as-made smart film containing 5 wt% Cs_xWO₃/PVP NRs changed from 22.8 °C to only 26.7 °C, indicating that the as-made smart film containing 5 wt% Cs_xWO₃/PVP NRs exhibited excellent performance in keeping the indoor temperature balanced and was applicable as an energy-saving smart window. Furthermore, we also compared the thermoregulation effects of the as-made smart film containing 5 wt% Cs_xWO₃/PVP NRs with the smart film without Cs_xWO₃/PVP NRs. As shown in Fig. 7e–h, the indoor temperature of the left house attached with an as-made smart film containing 0% Cs_xWO₃/PVP NRs noticeably went up from 18.3 °C to 32.3 °C after simulated solar radiation for 15 min, while that of the right house attached with an as-made smart film containing 5 wt% Cs_xWO₃/PVP NRs increased to only 24 °C during the same period. This result indicates that the Cs_xWO₃/PVP NRs within the as-made films play a key role in the shielding performance for heat rays.

Fig. 7 Photographs of a model house showing the temperature variation after continuous simulated solar irradiation: the model house attached with a PET film on the left and an as-made film containing 5 wt% Cs_xWO₃/PVP NRs on the right after (a) 0 min, (b) 3 min, (c) 9 min and (d) 15 min; and the model house attached with an as-made film containing 0% Cs_xWO₃/PVP NRs on the left and 5 wt% Cs_xWO₃/PVP NRs on the right after (e) 0 min, (f) 3 min, (g) 9 min and (h) 15 min.

Conclusions

In summary, we present a novel roll-to-roll production of a multi-responsive soft-matter composite smart film that can not only control the passage of visible light by temperature, electric fields and NIR light, but also screen 95% of the NIR solar irradiation from 800 nm to 2500 nm. These excellent properties of the film were achieved by creating a compatible PVP tuning layer between Cs_xWO₃ NRs and a polymeric syrup containing PT-LCs, followed by forming an elaborated–designed polymer structure within the film. The transmittance of the film in the visible region can be not only reversibly changed between 67% and 1.5% in response to temperature and NIR light stimuli accompanied by a SmA–N* phase transition, but also actively and dynamically regulated by applying an electric field to homeotropically align the LCs. Moreover, more than 95% of the NIR irradiation can also be effectively blocked due to the strong LSPR absorption of Cs_xWO₃ NRs well-dispersed within the composite. Besides, the film also exhibits a wide working temperature range, excellent flexibility, high mechanical strength, long-term stability and large-area processability. The as-prepared film shows great potential for application as a smart energy-saving window, and the proposed strategy provides new insights into engineering novel hybrid inorganic–organic functional materials.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (NSFC) (Grant No. 51333001, 51573006, 51561135014, 51671003, 51572059 and 51372029), the Major Project of Beijing Science and Technology Program (Grant No. Z121100006512002), National Basic Research Program of China (No. 2016YFB0100201), and the Sino-American Cooperative Project of the Chinese Ministry of Science and Technology (Grant No. 2013DFB50340).

Notes and references

1. D. Ge, E. Lee, L. Yang, Y. Cho, M. Li, D. S. Gianola and S. Yang, Adv. Mater., 2015, 27, 2489.
2. H. Khandelwal, A. P. H. J. Schenning and M. G. Debije, Adv. Energy Mater., 2017, 7, 1602209.
3. R. Baetens, B. P. Jelle and A. Gustavsen, Sol. Energy Mater. Sol. Cells, 2010, 94, 87.
4. N. Deforest, A. Shehabi, J. O'Donnell, G. Garcia, J. Greenblatt, E. S. Lee, S. Selkowitz and D. J. Milliron, Build. Environ., 2015, 89, 107.
5. H. Shin, S. Seo, C. Park, J. Na, M. Han and E. Kim, Energy Environ. Sci., 2016, 9, 117.
6. Y. Gao, C. Cao, L. Dai, H. Luo, M. Kanehira, Y. Ding and Z. L. Wang, Energy Environ. Sci., 2012, 5, 8708.
7. M. J. Powell, R. Quesada-Cabrera, A. Taylor, D. Teixeira, I. Papakonstantinou, R. G. Palgrave, G. Sankar and I. V. Parkin, Chem. Mater., 2016, 28, 1369.
8. M. J. Debije, Adv. Funct. Mater., 2010, 20, 1498.
9. H. Y. Lee, Y. Cai, S. Bi, Y. N. Liang, Y. Song and X. M. Hu, ACS Appl. Mater. Interfaces, 2017, 9, 6054.
10. Z. Xie, X. Jin, G. Chen, J. Xu, D. Chen and G. Shen, Chem. Commun., 2014, 50, 608.
11. D. K. Yang and Q. Li, Polymer stabilized cholesteric liquid crystal for switchable windows, Liquid Crystals Beyond Displays, John Wiley & Sons, Hoboken, 2012, p. 505.
12. K. G. Gutierrez-Cuevas, L. Wang, Z. Zheng, H. K. Bisoyi, G. Li, L. S. Tan, R. A. Vaia and Q. Li, Angew. Chem., Int. Ed., 2016, 55, 13090.
13. L. Wang, H. K. Bisoyi, Z. Zheng, K. G. Gutierrez-Cuevas, G. Singh, S. Kumar, T. J. Bunning and Q. Li, Mater. Today, 2017, 20, 230.
14. X. Liang, S. Guo, M. Chen, C. Li, Q. Wang, C. Zou, C. Zhang, L. Zhang, S. Guo and H. Yang, Mater. Horiz., 2017 10.1039/C7MH00224F.
15. Y. Gao, S. Wang, H. Luo, L. Dai, C. Cao, Y. Liu, Z. Chen and M. Kanehira, Energy Environ. Sci., 2012, 5, 6104.
16. A. Llordés, G. Garcia, J. Gazquez and D. J. Million, Nature, 2013, 500, 323.
17. Y. Zhou, Y. Cai, X. Hu and Y. Long, J. Mater. Chem. A, 2014, 2, 13550.
18. Y. Sang, Z. Zhao, M. Zhao, P. Hao, Y. Leng and H. Liu, Adv. Mater., 2015, 27, 363.
19. Z. Yu, Y. Yao, J. Yao, L. Zhang, Z. Chen, Y. Gao and H. Luo, J. Mater. Chem. A, 2017, 5, 6019.
20. Y. Gao, W. Yao, J. Sun, H. Zhang, Z. Wang, L. Wang, D. Yang, L. Zhang and H. Yang, J. Mater. Chem. A, 2015, 3, 10738.
21. J. Murray, D. Ma and J. N. Munday, ACS Photonics, 2016, 4, 1.
22. G. Cai, P. Darmawan, X. Cheng and P. S. Lee, Adv. Energy Mater., 2017, 7, 1602598.
23. Y. Wang, E. L. Runnerstrom and D. J. Milliron, Annu. Rev. Chem. Biomol. Eng., 2016, 7, 283.
24. K. Wang, H. Wu, Y. Meng, Y. Zhang and Z. Wei, Energy Environ. Sci., 2012, 5, 8384.
25. R. B. Goldner, T. Haas, K. K. Wong and G. Seward, US Pat., 4832463, 1989.
26. D. A. Higgins, Adv. Mater., 2000, 12, 251.
27. F. Kuschel, L. Hartmann and M. Bauer, Liq. Cryst., 2011, 38, 325.
28. J. K. Srivastava, R. K. Singh, R. Dhar and S. Singh, Liq. Cryst., 2011, 38, 849.
29. H. Kikuchi, M. Yokota, Y. Hisakado, H. Yang and T. Kajiyama, Nat. Mater., 2002, 1, 64.
30. I. Dierking, Adv. Mater., 2000, 12, 167.
31. C. Guo, S. Yin, L. Huang, L. Yang and T. Sato, Chem. Commun., 2011, 47, 8853.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c7nh00105c

‡ These authors contributed equally to this work.

---