Hybrid nanoparticles based on tungsten-doped vanadium dioxide (W_xVO₂) nanoparticles and poly(N-isopropylacrylamide) (PNIPAm) for thermochromic smart window applications were synthesized via silica layer formation using (3-mercaptopropyl)trimethoxysilane (MPTMS) followed by the surface-mediated free radical polymerization of PNIPAm, resulting in the aqueous dispersion of the hybrid nanoparticles. The metal-insulator transition of the W_xVO₂ core at around 28 °C and the lower critical solution temperature of PNIPAm at around 32 °C enable the aqueous dispersion of the hybrid nanoparticles to present an efficient thermochromicity depending on the temperature over a broad spectral range from visible to near-infrared light. The difference in the solar modulation was over 25%, while the luminous transmittance (T_lum) significantly decreased from 85% to 15% with increasing W_xVO₂ composition in the hybrids. Our results suggest that the hybridization of W_xVO₂ and PNIPAm is a potential methodology for preparing efficient thermochromic materials

열변색 스마트 창호 소재로서 텅스텐이 도핑된 바나듐이산화물(W_xVO₂) 나노입자와 폴리이소프로필아크릴아미드(PNIPAm)의 하이브리드 나노입자가 실리카 쉘 형성 및 PNIPAm의 표면라디칼 중합반응에 의해 합성되었다. 하이브리드 나노입자의 수분산액은 섭씨 28도의 금속-절연체 전이온도를 보유한 W_xVO₂ 코어와 섭씨 32도의 하한 임계 용액온도를 보유한 PNIPAm 쉘의 특성으로 인해 가시광선에서 근적외선까지의 넓은 스펙트럼 범위에서 효과적인 열변색 특성을 보였다. 하이브리드 나노입자에서 PNIPAm의 함량을 고정시키고 W_xVO₂의 함량을 변화시키며 특성을 측정한 결과 섭씨 20도에서 40도의 온도 변화에서 25% 이상의 태양광 변조 변화를 구현하였으며 W_xVO₂의 함량증가에 따라 입자의 증가된 뭉침현상에 의해 투과도는 85%에서 15%까지 급격하게 감소하였다. 본 연구결과는 W_xVO₂와 PNIPAm의 하이브리드화가 효율적인 열변색 소재를 제조하기 위한 주요 기술이 될 수 있음을 보여준다

Hybrid nanoparticles based on tungsten-doped vanadium dioxide (W_xVO₂) nanoparticles and poly(N-isopropylacrylamide) (PNIPAm) for thermochromic smart window applications were synthesized via silica layer formation using (3-mercaptopropyl)trimethoxysilane(MPTMS) followed by the surface-mediated free radical polymerization of PNIPAm, resulting in the aqueous dispersion of the hybrid nanoparticles.

Keywords: vanadium oxide, tungsten doping, poly(N-isopropylacrylamide), thermochromic, smart windows

This research was supported by the Chung-Ang University Graduate Research Scholarship in 2019 and the Korea Institute of Energy Technology Evaluation and Planning, funded by the Ministry of Trade, Industry and Energy (Grant No. 20182020109500), Republic of Korea.

Nowadays, technologies for saving energy are receiving widespread attention, with many materials and methods having been studied. Among them, thermochromic smart windows that can control the transmission of sunlight are popular energy-saving devices in buildings because energy consumption used for buildings accounts for 30-40% of the total energy consumption.^1,2 Energy can be saved by efficiently manipulating the amount of sunlight transmitted through the smart window.³ For example, when the outside temperature is high, the smart window can reduce the amount of sunlight entering the building, thereby decreasing the cost of cooling the interior of the building. On the other hand, when the outside temperature is low, the smart window can transmit the sunlight, thus maximizing energy-saving efficiency through internal temperature control.⁴
A representative thermochromic material used for smart windows is monoclinic vanadium dioxide (VO₂(M)). It presents a metal-insulator transition (MIT) at around 68 ℃, above which its structure changes from monoclinic to tetragonal and is accompanied by a change in the electrical property of VO₂ from semiconducting to metallic.^5,6 The resulting reflection of near-infrared light (NIR) at a temperature above 68 ℃ enables smart windows utilizing VO₂ to present the highest solar modulation ability (T_sol) of 22.3% and an average luminous transmission (T_{lum, average}) of 42.8%.^7,8 However, the MIT temperature of pristine VO₂(M) is too high for practical smart window applications that require an MIT at around room temperature and also the thermochromic property is limited to the NIR range.^9,10 Thus, a new technology based on VO₂ with an MIT near room temperature and solar modulation capability in the visible-to-NIR range is strongly required.
A promising approach to resolve this issue is to combine tungsten-doped VO₂ (W_xVO₂) with a thermo-responsive polymer that can modulate sunlight transmission in the visible light range. When W⁶⁺ or Ti⁴⁺ is incorporated into the crystalline structure of VO₂, the donation of electrons and distortion of the crystalline structure results in a decrease in the MIT temperature of pristine VO₂(M) down to room temperature.^11-13 On the other hand, poly(N-isopropylacrylamide) (PNIPAm), an organic thermo-responsive polymer, has a lower critical solution temperature (LCST) at around 32 ℃ and has received wide attention in smart window applications due to their change of transparency from a transparent state to an opaque state when the temperature changes above the LCST.^14-19 Because of the aggregation and precipitation of polymer chains above the LCST, the aqueous solution of PNIPAm becomes cloudy, thereby blocking the transmission of visible light, which is useful for smart window applications. For example, a PNIPAm hydrogel film shows a high T_lum of 70.7% and DT_sol of 25.5% in the visible range.¹⁷ Thus, one can expect that the hybridization of W_xVO₂ with PNIPAm will provide solar modulation capability in the visible-to-NIR range at around 30 ℃, which should be useful for practical applications of smart windows. The hybridization of VO₂ and PNIPAm has already been reported.^7,14,17 However, those studies focused on microgel and film forms that are different from the aqueous solution systems in our study, and the hybridization of W_xVO₂ with PNIPAm has not yet been reported.
In this study, we demonstrate the fabrication and thermochromicity of hybrid nanoparticles comprising W_xVO₂ and PNIPAm prepared via the surface polymerization of PNIPAm on W_xVO₂ nanoparticles in Scheme 1. Firstly, the W_xVO₂ nanoparticles was synthesized by hydrothermal synthesis with sodium tungstate. And then, a silica layer with surface thiol groups was first formed on the W_xVO₂ nanoparticles via the functionalization with (3-mercaptopropyl)trimethoxysilane (MPTMS) to provide surface thiol functionalities for polymerization and chemical stability²⁰ to the core W_xVO₂ under exposure to aqueous media. PNIPAm was then synthesized on the silica shell surface via free radical polymerization. The resulting hybrid organic-inorganic nanoparticles (W_xVO₂@ MPTMS@PNIPAm) dispersed in an aqueous medium presented a reversible phase-transition property from a transparent (hydrophilic) to an opaque (hydrophobic) state at its LCST temperature, which is characteristic of PNIPAm.¹⁴ Moreover, they presented an efficient solar modulation ability over the NIR range at around room temperature, which is characteristic of W_xVO₂. The aqueous dispersion of W_xVO₂@MPTMS@ PNIPAm hybrid nanoparticles suggests the facile fabrication of smart windows by injecting the solution into an empty narrow space between two glass windows, which should be beneficial for simple, cost-effective fabrication.


Scheme 1. Synthesis and coating procedure for WxVO2, WxVO2@ MPTMS, and WxVO2@MPTMS@PNIPAm.

Materials.Oxalic acid dihydrate (C₂H₂O₄∙2H₂O), vanadium pentoxide (V₂O₅), sodium tungsten oxide dihydrate (Na₂WO₄∙ 2H₂O) were purchased from Alfa Aesar (Heysham, UK). N-isopropylacrylamide (NIPAm) and 2,2'-azobis[2-(2-imidazolin-2-yl)propane] dihydrate (VA044, C₁₂H₂₂N₆∙2HCl) were purchased from the Tokyo Chemical Industry (TCI, Tokyo, Japan). Polyvinylpyrrolidone (PVP, M_W~1300000 Da) and MPTMS (SH(CH₂)₃Si(OCH₃)₃) were purchased from Sigma-Aldrich (St. Louis, MO, USA). All of the reagents were used as received.
W_xVO₂ Nanoparticles Synthesis.Oxalic acid (10.08 g, 0.112 mol) and V₂O₅ (7.27 g, 0.04 mol) were dissolved in deionized (DI) water (100 mL) in a three-neck round-bottomed flask equipped with a condenser for reflux. Na₂WO₄ (0.29 g; 2 mol% of W to V) was dissolved in 10 mL of DI water. The mixture was stirred at 400 rpm for 24 h at 60 °C. Afterward, the mixture was transferred to a Teflon-lined autoclave and the hydrothermal reaction was conducted at 185 °C for 12 h. After completing the reaction, the product was washed by centrifuging with ethanol three times and dried in a vacuum oven at room temperature for 24 h. The dried black-blue powder was calcinated at 700 °C for 2 h under a nitrogen atmosphere of a furnace.
W_xVO₂@MPTMS Coating. W_x-VO₂ (0.05 g) dispersed in ethanol (5 mL) via sonication for 30 min and PVP (0.128 g, 1.2 mmol on a repeat unit basis) dissolved in ethanol (10 mL) were added to a round-bottomed flask and stirred at room temperature for 72 h. Next, MPTMS solution (0.5 mL dissolved in 10 mL of ethanol) and ammonium solution (10 mL, 35 wt%) were added and stirred at room temperature for 12 h, resulting in a gray mixture solution. The precipitate was collected by centrifuging the solution at 8000 rpm for 15 min with excess ethanol to remove any uncoated W_xVO₂. The powder was dried in a vacuum at room temperature for 24 h.
W_xVO₂@MPTMS@PNIPAm Synthesis.In a three-neck round flask, W_xVO₂@MPTMS powder (0.05 g) was dispersed in DI water (10 mL) via sonication for 30 min, followed by the addition of VA044 (0.05 g, 0.155 mmol) in DI water (10 mL) and NIPAm (1.13 g; 10 mmol) in DI water (10 mL). The resulting solution was stirred at 60 °C for 4 h with nitrogen bubbling. Afterward, the solution mixture was precipitated in excess methanol (100 mL) to obtain a sticky solid product comprising black W_xVO₂ coated with white PNIPAm synthesized on the nanoparticle surfaces. This product was dried in a vacuum at room temperature for 24 h, which resulted in a hard product.
Characterization.Morphological observations of the nanoparticles shell were conducted by using a field-emission transmission electron microscope (FE-TEM; JEM-F200, JEOL, Japan). The crystalline structures of W_xVO₂(M) and VO₂(M) were verified via X-ray diffraction (XRD; Bruker-AXS NEW D8 Advance, Billerica, MA, USA). The MIT behaviors of W_XVO₂(M) and VO₂(M) were investigated using differential scanning calorimetry (DSC; Q-600, TA Instruments, Newcastle, DE, USA) analysis in the range of 20-130 °C at a heating rate of 10 °C/min. The thermochromic properties of W_xVO₂(M) and W_xVO₂@MPTMS@PNIPAm were measured with a UV-Vis-NIR spectrometer (V-770, JASCO, Tokyo, Japan) in the range of 300-2500 nm. Thermal decomposition behaviors were analyzed by thermogravimetric analysis (TGA; TA SDT Q600, TA Instruments, Newcastle, DE, USA) in the range of 20-600 °C at a heating rate of 10 °C/min.

The structures and phase-transition behavior of VO₂ and W_xVO₂ were analyzed via measurements from XRD patterns and DSC thermograms. The diffraction peaks in the observed XRD patterns of both VO₂ and W_xVO₂ were accurately matched with the monoclinic VO₂ crystalline structure (JCPDS NO.43-1051).²¹ In the XRD patterns, sharp diffraction peaks for VO₂(M) at 26.8°, 27.8°, 33.4°, 37.0°, 39.8°, 42.1°, 42.2°, 55.5°, and 57.4°, corresponding to the (110), (011), (-102), (-202; -211; 200), (002), (-212;210), (-213; 220; 211), and (022) planes, show the high crystallinity of VO₂ and W_xVO₂ prepared in this study (Figure 1). The phase-transition behavior observed on the DSC thermograms clearly indicates the tungsten doping effect. When 2 mol% of tungsten to vanadium was added, the phase-transition temperature decreased from 66.9 to 27.9 ℃, as shown in Figure 2(b). These results clearly show that although the small amount of tungsten incorporated into the pristine crystalline structure of VO₂ did not significantly modify the original structure, it did nevertheless decrease the MIT temperature to around room temperature. The shoulder peak appeared at around 80 °C in the DSC thermogram of VO₂ is ascribed to phase transition of agglomerate of particles. It should be noted that the shoulder peak is significantly weakened after multiple heating and cooling cycles as shown in Figure 2(b) and that the phase transition temperatures are slightly shifted to 29.7 and 67.9 °C for W_xVO₂ and VO₂, respectively. These results suggest that the agglomerates might be loosened and that the particles have stable phase transition behaviors over the multiple thermo-cycles.
To functionalize the surface of W_xVO₂ with thiol groups for the free radical polymerization of PNIPAm, W_xVO₂ nanoparticles were first coated with a silica layer. We used a literature procedure²¹ except that we directly used MPTMS to form the silica shell with the thiol groups instead of using a conventional alkyl siloxane followed by MPTMS coating. In the process, PVP was first dissolved in ethanol and then mixed with an ethanol dispersion of W_xVO₂ nanoparticles, after which the resulting mixture solution was stirred for 72 h to stabilize the nanoparticles via the adsorption of PVP chains on the nanoparticle surfaces.^22,23 Following this, the formation of the silica shell using MPTMS was carried out by adding it and an ammonium solution to hydrolyze and condense the siloxanes, which was followed by stirring for 12 h.^24-26 Gray nanoparticles were collected and the successful introduction of surface thiol groups was confirmed by measuring an FTIR spectrum and a TEM image of the nanoparticles. As shown in Figure 3, the FTIR spectrum of W_xVO₂ nanoparticles coated with a silica layer containing thiol groups show characteristic bands of C-H stretching of aliphatic alkyls (3000-2800 cm^-1), S-H stretching (2560.5 cm^-1), -CH₂- bending (1407.8 cm^-1), Si-C bending (1258.8 cm^-1), and Si-O-Si asymmetric stretching (1092.5 & 1001.4 cm^-1).²⁷ The splitting of peaks at 1092.5 and 1001.4 cm^-1 originates from the longitudinal and transverse optical vibrational modes of Si-O-Si asymmetric stretching vibration.²⁸ These very strong split peaks characteristic of Si-O-Si bonds suggest that the silica skeleton formed by hydrolysis and condensation of MPTMS on the surfaces of the W_xVO₂ nanoparticles was sufficiently ordered. Furthermore, the TEM image shown in Figure 4(a) clearly demonstrates the shell layer of the W_xVO₂ nanoparticle with a thickness of 40-50 nm, further confirming the successful synthesis of the silica shell. In addition, C-H and S-H stretching bands in the FTIR spectrum suggest the existence of thiol groups in the silica shell that are available for further surface reactions.
Preparation of hybrid nanoparticles of W_xVO₂ and PNIPAm was carried out by polymerizing PNIPAm on the surfaces of the W_xVO₂ nanoparticles. NIPAm was polymerized via a free radical mechanism by utilizing a water-soluble azo-based radical initiator (VA044) and thiol groups as reacting sites on the surfaces of the W_xVO₂ nanoparticles.²⁹ The successful polymerization of PNIPAm on the surfaces of the W_xVO₂@MPTMS nanoparticles was confirmed via measurements on an FTIR spectrum (Figure 3) and a TEM image (Figure 4(b)). The FTIR spectrum of W_xVO₂@MPTMS@PNIPAm shows that the S-H peak at 2560.5 cm^-1 disappears and that characteristic bands for PNIPAm,³⁰ including O-H stretching (~3433.2 cm^-1), NH stretching (3280.3 cm^-1), and C(=O)-N-H (1639.7 cm^-1, amide I; 1540.4 cm^-1, amide II), clearly appear. Moreover, PNIPAm on the W_xVO₂ nanoparticles was clearly observed as a light shell with a thickness of 100-200 nm surrounding the nanoparticles due to its low electron density in comparison to W_xVO₂, as shown in Figure 4(b) and 4(c). All of these results confirm that the hybrid structure of a W_xVO₂ core and a PNIPAm shell was successfully synthesized.
The compositions of these organic-inorganic hybrids were analyzed by measuring TGA pyrograms (Figure 5). Three distinct weight loss at around 100, 400 and 500 °C are allocated to the moisture evaporation, degradation of PNIPAm and inorganic components, respectively. We prepared three hybrid nanoparticles by using three different amounts of W_xVO₂(M) particles (0.02, 0.05, and 0.1 g) while the amount of NIPAm was fixed at 1.13 g, which means 1.7, 4.4, and 8.8 wt% of W_xVO₂ to PNIPAm in the hybrid nanoparticles. The residual weight percentages were -4.5, 2.9 and 7.2 wt%, respectively. Although the results show experimental errors, it was confirmed that W_xVO₂ nanoparticles were increasingly incorporated as designed.
The thermochromicity of the hybrid nanoparticles was examined in the wavelength range of 300-2500 nm using UV-Vis-spectroscopy below and above the phase-transition temperatures of W_xVO₂ and PNIPAm (20 and 40 °C, respectively). The hybrid nanoparticles were dispersed in DI water at 0.5 wt%; the resulting solutions are denoted as dispersions I, II, and III in the order of increasing W_xVO₂ weight percentage. The transmittance spectra of the three dispersion amounts measured at 20 and 40 °C were measured (Figure 6) and used to estimate their solar modulation (T_sol), solar modulation difference (ΔT_sol), and luminous transmittance (T_lum).
T_sol at each temperature was calculated as



where φ(λ) and Tr(λ) are the solar irradiation spectrum for an air mass of 1.5, corresponding to the sun standing 37° above the horizontal³¹ and transmittance at specific wavelength λ, respectively.
ΔT_sol was estimated by applying³²



where φ_lum(λ) is the spectral sensitivity of the light-adapted eye.³³
The estimated ΔT_sol values utilizing the spectral variation depending on the temperature in Figure 5 were 5.4%, 13.9%, and 25.5% for the dispersions I, II, and III, respectively. These values show that the solar modulation ability of the hybrid nanoparticles was enhanced with increasing W_xVO₂ content. It is notable that the solar modulation of dispersion I with the lowest W_xVO₂ content can be mainly ascribed to spectral variations in the visible light range (Figure 6(a)) while the spectral variations in both the visible and NIR regions contribute to the enhanced ΔT_sol values of dispersions II and III (Figure 6(b) and 6(c)). Thus, it seems that W_xVO₂ at over 8 wt% in the hybrid nanoparticles with PNIPAm is necessary to observe a distinct contribution from the W_xVO₂ toward the thermochromicity over the visible-to-NIR range.
On the other hand, the T_lum values of dispersions I, II, and III at 20 °C decreased with increasing W_xVO₂ content (84.8, 42.4, and 14.7%, respectively). It should be mentioned that even though the ΔT_sol value of dispersion III is the highest, its T_lum value of dispersion III is too low for practical applications. This can be ascribed to the size of hybrid nanoparticles being over a few hundred nanometers and their aggregation, which can severely scatter light and decrease light transmission. It should be mentioned that the polymerization of PNIPAm via the free radical mechanism can occur both in the medium and on the surfaces of W_xVO₂ nanoparticles. As a result, the particles can be even connected by PNIPAm chains as shown in the TEM image (Figure 4(c)). We presume that these agglomeration of the W_xVO₂ nanoparticles and PNIPAm chains cause such a severe light scattering in the NIR region and the non-linear dependence of the NIR transmittance on the weight fraction of W_xVO₂ nanoparticles. However, our results suggest that the hybrid nanoparticles based on W_xVO₂ and PNIPAm with a particle size of less than 100 nm can present excellent thermochromicity in both the visible and NIR ranges.

In this study, we synthesized hybrid nanoparticles of tungsten-doped VO₂ core with a PNIPAm shell that had efficient thermochromicity for smart window applications over a broad spectral range (from the visible to the NIR region). Unlike conventional gel- or film-coating type thermochromic devices, our hybrid nanoparticles were designed to be applied as an aqueous dispersion that could simply be injected in between glass plates to simply achieve low-cost smart windows. W_xVO₂ nanoparticles were coated with a silica shell offering numerous thiol groups via the hydrolysis and condensation of MPTMS and finally, an outermost shell of PNIPAm synthesized via free radical polymerization utilizing the surface thiol groups. The resulting aqueous dispersions of the hybrid nanoparticles presented increasing ΔT_sol until the W_xVO₂ content reached 25%. However, it was noted that T_lum was drastically decreased down below 20% with the increase in W_xVO₂ in the hybrid nanoparticles, thereby making them unsuitable for practical applications. These issues suggest that further investigation should be carried out to determine the best route for synthesizing ultrafine tungsten-doped VO₂ with a diameter of less than 100 nm via a process that avoids aggregation. We believe that a novel synthetic route such as a modified polyol process³⁴ can provide W_xVO₂ nanoparticles with an MIT temperature at around RT and with a size less than 100 nm, enabling the formation of hybrids with negligible light scattering in the NIR region.