Approaching the Theoretical Maximum Performance of Highly Transparent Thermochromic Windows

Abstract

Thermochromic window coatings represent a promising technology to improve the energy efficiency of buildings in intermediate climates. With the technology approaching market introduction it is important to investigate its performance limits within smart windows and to identify existing development challenges. Here we analyze the theoretical maximum performance of thermochromic window coatings that modulate IR transmission whilst retaining high visible transparency. The set limitations lead to a theoretical maximum solar modulation of 39.1%. Within an insulated glazing unit (IGU), where at least 2 glass panes and a conventional low-e coating are required, this value is further reduced to 12.9%. We show that by carefully selecting a low-e coating with the highest compatibility to a thermochromic coated glass and by allowing 10% of modulation in the visible spectral range, the theoretical maximum can be increased to 23.1%, illustrating the importance to codesign and match both coatings within a smart window to reach optimum performance. Furthermore, we compared our current best-performing VO₂:SiO₂ composite coating within an IGU to the theoretical maximum. The analysis shows that with a solar modulation of 13.4%, the coating is currently at 59% of the theoretical maximum. Finally, we propose and discuss several strategies to proceed further toward the theoretical maximum.

1. Introduction

In the process of increasing the energy efficiency of buildings, most building components require optimization to minimize interaction with the outside environment. Buildings are made air-tight, and the insulation of walls, roofs, and floors is further increased to realize a static indoor climate [1]. The part of the building skin for which interaction with the outside is still desired, are windows. For windows, capturing natural light and solar heat is desired to a certain extent. It has been demonstrated that sunlight and solar heat have a positive influence on people’s health and well-being [2] and all solar energy that can be directly used in a building for lighting and heating reduces the overall energy costs [3]. However, when certain irradiance levels are surpassed, glare becomes an issue, and shading is desired. In addition, solar heat transmission through windows accounts for 60% of the total solar heat gain in buildings. As a result, overheating of buildings and air-conditioning demand to reduce indoor temperatures are largely based on solar transmission through windows [4]. Therefore, the requirements for visible (T_vis) and total solar (T_sol) transmission vary depending on the climate zone and the amount of window surface [3,5]. The glass industry is approaching these different demands by offering static glazing with different T_vis and g values (the g value equals the total solar energy transmitted through a window, radiative and conductive) [6]. A wide range of different products are available which have been optimized over the past decades. However, since in many climate regions a building’s demands on sunlight and solar heat transmission change during the day and between the seasons, static glazing can never capture the full potential of sunlight.

Traditionally, the inability of windows to adapt to the changing requirement for sunlight and solar heat transmission has been accommodated by using external or internal shading devices [7]. Since all these shading devices themselves have certain limitations, like not being able to distinguish between visible light and solar heat transmission, recent development has been focused on smart windows that can adapt their interaction with sunlight directly upon an external stimulus [8]. Electrochromic windows are the most advanced of all smart window technologies, with several products already available on the market [9,10]. Due to their complexity and the necessity for wiring and special frames, these technologies are expensive and therefore still limited to niche applications. Since the costs cannot be earned back via energy savings in a reasonable amount of time, electrochromic smart windows are focused on increasing comfort via actively adjusting the shading of the glass and therefore providing privacy, as well as reducing glare and solar heat gain [11]. Another smart window technology that is being developed is photochromic windows. Photochromic glazing technology is already commercially applied in sunglasses that increase shading at high solar irradiance levels. In the same way, photochromic windows can be used to increase shading and reduce glare [8]. As already inherent to the technology, photochromic materials are foremost applicable for glare control. It is difficult to optimize photochromic windows for minimum energy demand, due to the fact that irradiance levels are not the best trigger to determine when to transmit or block solar heat. For example, on sunny winter days, the high irradiance levels can induce tinting, although low ambient temperatures make solar heat gain still beneficial for the energy demand. Thermochromic materials are best suited to be optimized for energy efficiency alone [12,13]. They can be designed to minimize the impact on T_vis, whilst modulating infrared (IR) transmission and therefore solar heat gain. Temperature as the external trigger is best suited to couple with a building’s thermostat by adjusting the transition temperature in the thermochromic material. Additionally, the material does not require special equipment to trigger the switch and can be installed like regular windows. This way, thermochromic windows can be designed for use in regular insulated glazing units (IGUs) enabling optimum use of solar heat whilst retaining high visible transparency.

Because of their high application potential, thermochromic materials—mainly vanadium dioxide (VO₂)—have been a research focus in the last years [12,13,14]. Materials for thermochromic smart windows have been optimized by increasing solar modulation (ΔT_sol) values at high T_vis [15,16,17,18]. Switching temperatures have been optimized to cover a wide range between 0 and 68 °C [19,20], increasing the potential to tune the properties to requirements in various climate zones. Furthermore, processes are being scaled up and production costs are being reduced [17,21]. The potential of thermochromic smart windows has been analyzed in several simulation studies [22], including the impact of switching temperature and hysteresis [23], and it has been shown that these new technologies surpass commercial glazing in energy efficiency in intermediate climates [3]. Nevertheless, most of the studies optimizing the performance of thermochromic windows, as well as many simulation studies, are studying the optical properties of the thermochromic material isolated from the real application in smart windows. But windows are not only important for their solar heat gain, but they are also the most vulnerable part of the building skin with respect to radiator heat loss. To increase the insulation value of windows, the number and thickness of glass panes as well as the number, width and atmosphere of gas spaces between the glass panes can be adjusted. In addition, the glass panes can be coated with a low emissivity (low-e) coating, significantly reducing the thermal conductivity of an IGU [24]. Since these low-e coatings not only impact infrared radiation between 5–25 µm, and therefore the emissivity, but also near-IR transmission, they have to be accounted for when analyzing the impact of thermochromic glazing. To realize insulation values (u values) of thermochromic smart windows in agreement with current standards, a combination of thermochromic glass and low-e coated glass is required. We have shown in a recent simulation study that only with such a combination the positive effect of changing solar heat gain can be realized. However, the low-e coating properties of partly blocking solar near-IR light reduces the solar modulation of the complete IGU [3]. This shows the necessity of codesigning all components of the IGU and investigating a perfect combination of thermochromic and low-e properties.

Since thermochromic smart windows are approaching market introduction, it is important to investigate the performance limit of smart windows, to identify existing development challenges [25]. For that purpose, we study practical combinations of our thermochromic glass with best-in-class optical performance (combination of visible transmission and solar modulation) [15] with selected commercially available low-e glasses. Here we report a detailed study on the maximum theoretical solar modulation when focusing solely on the IR part of the solar spectrum. Furthermore, we investigate how the theoretical maximum changes when applied in an IGU with different low-e coatings, and when including 10% of visible transmission in the modulated range. Additionally, we compare one of our high-performance thermochromic smart window coatings to these theoretical maximum values to investigate which major challenges are to overcome, and we give an outlook on how to optimize these coatings to come closer to their theoretical maximum performance.

2. Methods

2.1. Theoretical Investigation

For investigation of potential transmission modulation via thermochromic smart windows for several combinations of different glazing options, the UV-vis-NIR transmission spectrum of a commercial glass or combination of two glasses in an IGU was compared to a manipulated version of the respective spectrum, where transmission was set to 0 in wavelength regions > 780 nm. The two spectra were compared by calculating the total solar transmission (T_sol) and the solar modulation (ΔT_sol). T_vis, T_sol and ΔT_sol were calculated using the following formulas:

$$T_{vis,sol}=\frac{\int {\varphi }_{vis,sol}\left(\lambda \right)T\left(\lambda \right)d\lambda }{\int {\varphi }_{vis,sol}\left(\lambda \right)d\lambda }$$

(1)

$$\mathsf{\Delta }{T}_{sol}=T_{sol,cold}-T_{sol,hot}$$

(2)

with T(λ) as the transmittance at wavelength λ, Φ_vis(λ) as the photopic spectral sensitivity of human vision, and Φ_sol(λ) as the AM 1.5 solar irradiance spectrum. For IGUs also modulations of g values (Δg) were calculated. Here g values were calculated according to EN410 [26]. For each IGU the following assumptions were made (Figure S1): (i) double glazing with 2 glass panes and 1 gas space; (ii) glass thickness 4 mm; (iii) gas space 12 mm; (iv) gas space atmosphere argon; (v) adaptive glass as outer glass pane with coating on surface 2; (vi) coating of inner glass pane on surface 3.

The following glasses and glazing systems have been used in the study (

Table 1. Glasses and glazing systems used in this study.

Glass				Glazing System (Outer Glass Pane × Inner Glass Pane)
Product Name	Manufacturer	Type of Glass	Thickness
Optifloat Clear	Pilkington, Lathom, UK	Monolithic	4 mm
ECLAZ	Saint-Gobain, Courbevoie, France	Coated	4 mm	Optifloat × ECLAZ
Isolide Superplus	Scheuten Glas, Venlo, The Netherlands	Coated	4 mm	Optifloat × Isolide
SunCool 70/40Pro	Pilkington, Lathom, UK	Coated	6 mm *	Optifloat × SunCool
SunSmart 67/23	TNO, Geleen, The Netherlands	Coated/thermochromic	4 mm	SunSmart × ECLAZ

* The SunCool series was only available in 6 and 12 mm thickness in the database. However the deviation in glass thickness is not expected to have a major impact on the conducted analysis.

).

Optical properties of coated (Scheuten Isolide and TNO SunSmart) and uncoated (Pilkington Optifloat) glass plates have been measured via UV-vis-NIR spectrophotometry using a Perkin Elmer Lambda 750 (Perkin Elmer, Inc., Waltham, MA, USA) with integrating sphere and a customized module comprising a heating element equipped with a module for measuring transmission and reflection in the wavelength regions between 300 nm and 2400 nm. Optical data of commercial low-e coated glass (Saint-Gobain ECLAZ and Pilkington SunCool) have been used from the International Glazing Database (IGDB) within the WINDOW program (version 7.6) by Berkeley Labs [27]. The transmission spectra of all glasses used in this study are presented in Figure 1. T_vis and T_sol of glazing systems have been calculated using the Optics program by Berkeley Labs [27] and transmission spectra have been calculated using the following formula:

$$T_{glazing system}\left(\lambda \right)=T_{glass pane 1}\left(\lambda \right)\times T_{glass pane 2}\left(\lambda \right)$$

(3)

with T(λ) as the transmittance at wavelength λ.

2.2. Coating Preparation

The thermochromic coating used in this study was prepared in analogy to a previously published procedure [15]. One side of a barrier coated glass substrate was masked using d-c-fix^® self-adhesive foil. Subsequently, the thermochromic coating was applied to the non-covered, tin side using dip coating. The applied liquid coating formulation consisted of a mixture of vanadyl oxalate solution and silica at a SiO₂ to VO₂ ratio of 2.2:1 w/w. The xerogel coating was applied to the glass substrates at a withdrawal speed in the range of 8.0 mm·s⁻¹ to achieve the desired thickness. After drying for 5 min under ambient conditions, the adhesive foil was removed. The non-coated side was then placed onto a 6 inch silica wafer, and annealed in a Jipelec Jetfirst PV Rapid Thermal Processor (ECM, Grenoble, France), with the coated side facing the IR radiators. The coated glass plates were first heated to 270 °C in air for 480 s, and subsequently in N₂ at 450 °C for 30 s.

3. Results and Discussion

3.1. Investigation of Theoretical Maximum Performance

To investigate the theoretical maximum solar modulation for adaptive glazing, we started by analyzing the wavelength dependent light transmission of clear glass. Here we measured the UV-vis-NIR transmission of a 4 mm Pilkington Optifloat Clear glass and calculated the solar properties. The wavelength dependent transmission spectrum can be seen in Figure 1. The glass plate showed a T_vis of 90.3% and a T_sol of 85.4% (

Table 2. Solar properties of measured and theoretical glass and glazing systems with potential solar modulation for various scenarios.

Glass/Glazing System	Scenario	Tvis [%]	Tsol [%]	ΔTsol [%]
Optifloat Clear	Full transmission	90.3	85.4	39.1
	blocking > 780 nm	90.3	46.3
Optifloat × ECLAZ	Full transmission	83.2	60.3	19.0
	blocking > 780 nm	83.2	41.3
	blocking > 780 + 10% vis	74.9	37.2	23.1
Optifloat × Isolide	Full transmission	79.8	51.3	12.9
	blocking > 780 nm	79.8	38.4
Optifloat × SunCool	Full transmission	73.7	35.8	3.1
	blocking > 780 nm	73.7	32.7
SunSmart	Cold state	67.1	67.7	22.9
	Hot state	57.4	44.8
SunSmart × ECLAZ	Cold state	61.7	45.7	13.4
	Hot state	52.8	32.3

). According to EN 410 [22], visible transmission is defined as the region between 380–780 nm. Therefore, if the solar modulation should be limited to the IR region, all light transmitted >780 nm could theoretically be blocked. To simulate this theoretical blocking state, the light transmission between 780–2500 nm was set to 0, whilst retaining the wavelength dependent transmission between 300–780 nm of the Optifloat Clear glass. The thereby obtained theoretical transmission spectrum of a total IR blocking state (blocking >780 nm) was used to calculate the theoretical solar properties. Since T_vis was not influenced by the data manipulation, the same value as for the Optifloat Clear glass was obtained, whereas the T_sol value dropped to 46.3% (Table 2). Directly comparing the transmission spectrum of Optifloat Clear glass with the theoretical IR blocking state shows that a large part of the transmission can be modulated even by solely focusing on the IR (Figure 2). Since solar intensity drops significantly in higher wavelength regions and the IR only accounts for approximately 45% of the total solar radiation, the maximum theoretical solar modulation ΔT_sol between these two states is 39.1% (Table 2).

The analyzed optical properties of Optifloat Clear glass and the potential solar modulation of 39.1% for an adaptive glazing would be representative of single glazing. But this type of glazing has a very low insulation value of 5.8 W/m² K and therefore leads to significant radiator heat loss. Current standard glazing for newly built homes in the Netherlands is so-called HR++ glazing with an insulation value of 1.1 W/m² K, with similar standards for windows in most European countries. These are insulated glazing units (IGUs) made of 2 glass panes of which one is coated with a low-e coating and a gas space commonly filled with a low thermal conductivity gas, such as argon. This type of glazing has different optical properties from a single Optifloat Clear glass, especially in the IR. Therefore, to assess the theoretical maximum solar modulation of an adaptive glazing within a current state-of-the-art IGU, we investigated the optical properties of an IGU combining an Optifloat Clear glass with a Scheuten Isolide Superplus glass. In the UV-vis-NIR transmission spectrum, it can be seen that the Isolide Superplus glass still shows a high visible transmission, but the transmission in the IR drops significantly, especially in longer wavelength regions (Figure 1). In the IGU the transmission over the whole wavelength region drops slightly due to the transmission of the Optifloat glass between 80–90% for all wavelengths between 350–2500 nm (Figure 3). Overall, the representative HR++ glazing shows solar properties of T_vis = 79.8% and T_sol = 51.3% (Table 2). This means that the combination Optifloat x Isolide already blocks 48.7% of solar radiation from entering the building, which significantly reduces the amount of sunlight that can be modulated. To investigate the amount of IR light that can still be modulated within such an IGU, we simulated a theoretical IR blocking state with 0% IR transmission. To calculate the optical properties of such a glazing system, we combined the Isolide glass with the theoretical blocking > 780 nm glass. Such a combination showed again unchanged T_vis of 79.8% and a reduced T_sol of 38.4%. Comparing the UV-vis-NIR transmission spectrum of the Optifloat × Isolide combination with the theoretical blocking > 780 nm × Isolide state, it can be seen that the amount of IR light that can be modulated is reduced significantly in comparison to the Optifloat single glass (Figure 2 and Figure 3). Overall the theoretical maximum ΔT_sol drops to 12.9% for the HR++ glass.

Even though the Scheuten Isolide is a representative of the most common low-e coated glasses used in IGUs for residential buildings, there are other low-e coatings on the market that block more or less IR light. Since adaptive glazing can in general be combined with any low-e coated glass to reach an insulation value of 1.1 W/m² K, it is important to select the best suited low-e coated glass. The suitability here can be equated to the maximum possible solar modulation. To investigate the theoretical maximum solar modulation of IGUs with different low-e coated glasses and their suitability to be used in adaptive glazing systems, we calculated the solar properties of 2 additional IGUs with a highly transparent low-e coating and a so-called solar control coating, that blocks a large part of IR light. As highly transparent low-e coating we selected the Saint-Gobain ECLAZ series. Combined with an Optifloat clear glass in an IGU, a T_vis of 83.2% and T_sol of 60.3% were obtained (Table 2). The solar control glazing, where we selected Pilkington’s SunCool series as a representative case study, resulted in T_vis values of 73.7% and T_sol of 35.8% (Table 2). Here it could already be observed that the solar control coating blocks a large part of IR light and therefore leaves little room for modulation via adaptive glazing (Figure 1). To calculate ΔT_sol values and to compare the different low-e coatings, we again calculated transmission spectra where we combined the individual low-e coated glasses with the theoretical blocking > 780 nm spectrum. By comparing the IGUs with different low-e coatings with each other as well as with the respective theoretical IR blocking states, it was observed that the combination Optifloat × ECLAZ has the highest potential modulation range in the IR (Figure 4). Here, for the blocking state, a T_sol of 41.3% was calculated, which results in a ΔT_sol of 19.0% (Table 2). For the solar control glazing the blocking state had a T_sol of 32.7% which is not much lower than for the regular combination Optifloat × SunCool and therefore leads only to a ΔT_sol of 3.1%. The potential solar modulation of the HR++ glazing lies at 12.9% between those two. Since the highly transparent low-e coating (Saint-Gobain ECLAZ) transmits the largest part of IR light for all 3 studied low-e coatings it has the highest potential and best suitability to be combined with adaptive glazing.

Even though VO₂ based thermochromic coatings are aimed at staying highly transparent in the visible and solely modulating the IR region of sunlight, especially high-performing nanocomposite coatings [15] or coatings within anti-reflective coating stacks [28] also show a slight modulation in the visible range. Here, it has been shown that adding modulation of a few percent in the visible to a large modulation in the IR can significantly increase the overall ΔT_sol of a thermochromic coating. Furthermore, if the change in visible transmission is kept at a few percent, the change is barely visible to the human eye and the coated glass still appears highly transparent. To investigate the potential solar modulation of an IGU with an ECLAZ low-e coated glass, where full IR modulation is combined with partial modulation of visible light, we simulated a second theoretical blocking transmission spectrum. In accordance with the blocking > 780 nm spectrum, we started from the Optifloat Clear transmission spectrum whilst setting transmission > 780 nm to 0. In addition, we reduced the transmission of the Optifloat transmission spectrum ≤ 780 nm by 10%. This theoretical blocking > 780 + 10% vis spectrum was combined with the ECLAZ spectrum to simulate a theoretical blocking state of the IGU (Figure 5). For this blocking state, a T_vis of 74.9% and T_sol of 37.2% were calculated (Table 2). Comparing this with the transmission of the Optifloat × ECLAZ IGU, it can be seen that T_vis is reduced by 8.3% and a ΔT_sol of 23.1% is obtained. Since a larger modulation in the visible is not desired and the ECLAZ low-e coating shows the highest potential for solar modulation, it can be concluded that within these limitations a ΔT_sol of 23.1% is the theoretical maximum that can be achieved with VO₂ based thermochromic coatings within an IGU.

3.2. Current Record Performance

After assessing in detail the theoretical maximum solar modulation of a VO₂ based thermochromic coating within an IGU, we compared the properties of our current best-performing coating to these values. The coating we selected is a VO₂:SiO₂ composite coating, where the composition and coating thickness is optimized to realize an optimum combination of high T_vis of 67.1% and large ΔT_sol of 22.9%. A high T_vis is obtained via the integration of VO₂ domains in an SiO₂ matrix, which increases transparency and simultaneously reduces reflection via a reduced refractive index of the composite layer. In addition, the composition is optimized to create small evenly distributed VO₂ domains with an average size < 30 nm, which leads to increased light absorption in the rutile state (hot state) in wavelength regions between 800–1300 nm. This effect for small VO₂ domains within a coating matrix has been attributed in literature to an occurring plasmonic absorption of the metallic VO₂ phase [29,30]. The increased absorption in this wavelength region has a big impact on the overall T_sol in the hot state, since the solar IR intensity is highest between 780–1360 nm (Figure S2). Since solar IR intensity decreases >1400 nm significantly, the higher transmission of our SunSmart coating in the hot state in these wavelength regions does not negatively impact the overall ΔT_sol. By carefully designing the transmission of the thermochromic coating in both cold and hot states, the large ΔT_sol of 22.9% can be reached. To investigate the optical properties of a Smart Window IGU, we calculated the transmission spectra by combining our thermochromic coated glass with an ECLAZ low-e coated glass (Figure 6). For the cold state, we obtained T_vis values of 61.7% and T_sol of 45.7%, whereas for the hot state, we obtained a reduced T_vis of 52.8% and T_sol of 32.3% (Table 2). Overall we obtained a reduction in T_vis of 8.9%, which is similar to the blocking > 780 +10% vis × ECLAZ combination, and a ΔT_sol of 13.4%. A significant reduction in ΔT_sol in comparison to the single thermochromic coated glass can be observed, which is due to the IR-blocking properties of the ECLAZ coating. However, here also the designed optical properties with an increased modulation in the region between 800–1300 nm are beneficial, since in this region the ECLAZ coating still has a relatively high transparency, whereas >1400 nm the transparency drops significantly. Comparing the performance of the SunSmart IGU with the theoretical maximum performance, it can be concluded that the ΔT_sol of 13.4% is at 59% of the theoretical maximum.

3.3. Optimization Strategies to Approach Theoretical Maximum

With 59% of theoretical maximum performance, the current SunSmart coating already shows very good properties. However, optimization of the coating and the active material can further improve the performance to bring the resulting solar modulation closer to its theoretical maximum. By careful analysis of the composite coating, as well as optimization strategies in literature, we have identified four strategies to further optimize the solar modulation that will be elaborated in more detail.

3.3.1. Purity of Active Material

VO₂ can exist in an amorphous, as well as in various crystalline phases, of which only the monoclinic phase shows thermochromic properties [31]. Furthermore, vanadium can exist in several oxidation states ranging from +2 up to +5 [32]. Therefore, VO₂ in the oxidation state +4, can be easily oxidized or reduced. Since all vanadium species show distinct absorption in the visible, but only VO₂(M) has thermochromic properties, the T_vis of a coating is impacted by the total vanadium content, whereas ΔT_sol is only impacted by the VO₂(M) content. Therefore, to optimize the combination of high T_vis with high ΔT_sol it is important to increase the share of active VO₂(M) in the total vanadium content of a thermochromic coating.

In an earlier study, we investigated the purity of VO₂(M) in a VO₂:SiO₂ composite coating made from liquid precursors [15]. Here we could show that only 54% of the total vanadium content was crystalline VO₂(M). Since no other crystal phases or oxidation states were found, the remaining content was most probably amorphous VO₂. Therefore, further improving the purity of the VO₂(M) content via careful adjustment of the coating formulation and thermal anneal process can further improve the thermochromic performance.

Another possibility to produce thermochromic VO₂ coatings is by synthesizing VO₂ pigments and integrating those in a coating matrix. Here the purity of the VO₂(M) content can be more easily controlled, since the crystallization of VO₂(M) is decoupled from the coating formation. However, also the synthesis of VO₂ nanoparticles brings many challenges to achieving phase pure monoclinic material with a high switching enthalpy. Additionally, doping, which is required to reduce the switching temperature to application-oriented values, may also lead to more complex systems and other phases as side products which reduce the performance of the coating. Synthesizing VO₂ nanoparticles via a top-down approach using wet bead milling, we obtained particles with a switching enthalpy of 28 J/g and an amorphous fraction between 30–40% [33]. Therefore, more than one-third of the nanoparticles are non-switching VO₂. This will lead to a decrease in both the T_vis and ΔT_sol. Using a bottom-up hydrothermal approach to synthesize VO₂(M) nanoparticles, a switching enthalpy of 31 J/g with VO₂(A) as a side fraction was obtained [33]. In top-down synthesis, the purity of the VO₂(M) content can be improved via the synthesis of the VO₂ powder via calcination. Furthermore, the bead milling can be further improved to reduce the loss in crystallinity that occurs during size reduction. Here, every percentage improvement in the purity of VO₂(M) content within the nanoparticles will lead to improved thermochromic performance in the final coating.

3.3.2. Size and Distribution of Active Domains within the Matrix (Including Layer Thickness)

Particle size plays a key role in improving the optical properties of VO₂ films and coatings. To avoid Mie scattering the particles are required to be smaller than 100 nm [34]. With bead milling experiments we have increased both the T_vis and ΔT_sol by decreasing the particle size below 120 ± 10 nm [33]. Furthermore, the distribution of nanoparticles or active domains within the matrix of the coating or films is important to achieve an additional localized surface plasmon resonance (LSPR) in the metallic state of VO₂, which leads to a pronounced minimum at λ ≈ 1200 nm in the transmission spectrum [30,35]. Simulation work from Zhu et al. [30] illustrates the pronounced transmission minimum when 40 nm sized VO₂ particles are homogeneously dispersed in a coating matrix, whereas agglomeration leads to the disappearance of the LSPR absorption peak and optical properties that can be associated with a continuous VO₂ layer. Due to the high solar IR intensity between 780–1360 nm, the additional absorption associated with the LSPR leads to a significant increase in ΔT_sol.

VO₂ content within a film or coating as well as layer thickness also play a crucial role in optimization of thermochromic properties. We have shown that changing the ratio between VO₂ and SiO₂ in a VO₂:SiO₂ composite coating at similar thickness, leads to an increased T_vis with reduced VO₂ content, whilst retaining a similar ΔT_sol [15]. Here the increased T_vis was associated with a change in vanadium rich domain size and morphology in TEM/EDX analysis, as well as with a lower overall VO₂ content per surface area. Surprisingly, this lower VO₂ content per surface area did not result in a significant reduction in ΔT_sol, which indicates that the VO₂ domain size and morphology have a significant impact on the thermochromic performance, even though the dispersion in these coatings was not good enough to result in the occurrence of an LSPR. Furthermore, we showed that by increasing the layer thickness at a constant VO₂:SiO₂ ratio, the ΔT_sol could be increased significantly without leading to a major reduction in T_vis. This was already predicted by Schläfer et al. in a previous theoretical study on VO₂:SiO₂ composite coatings [36].

3.3.3. Integration into Antireflective Coating Stack

A common strategy within the literature to improve the thermochromic performance of VO₂ coatings is to integrate a continuous VO₂ layer into a multi-layer antireflective coating stack [12]. VO₂ coatings in their monoclinic phase have a refractive index around 2.5, which leads to light reflection at the interfaces around 10% over the whole solar wavelength region between 300–2500 nm. In addition, the refractive index of rutile VO₂ after phase transition is reduced to around 2 [37]. This leads, especially in the visible, to lower reflection and in continuous VO₂ coatings to a higher visible transmission in the hot state than in the cold state [38]. Since solar intensity per wavelength is much higher in the visible than in the IR, such an increase in T_vis upon switching leads to a significant reduction in ΔT_sol. By integrating a VO₂ layer into an anti-reflective coating stack, the refractive indices and thicknesses of the individual layers within the stack can be matched to significantly increase the transmission in the monoclinic state. Since only exact matching leads to destructive interference and reduced reflection, the same effect does not occur when the refractive index of the VO₂ layer changes upon switching. Therefore, the transmission in the hot state remains low, whilst the transmission in the cold state is increased. This simultaneously increases T_vis and ΔT_sol and can lead to high solar modulations of up to 21.8% [28]. Even though in a composite coating, the refractive index of the layer is already adjusted via the coating matrix to get rid of the effect of increasing T_vis upon switch, a further reduction in reflection via integration into an anti-reflective coating stack can also improve the thermochromic properties.

3.3.4. Other Strategies

Recently, more advanced strategies to improve the thermochromic performance of VO₂ coatings have been reported. Here a strategy is to combine VO₂ with another thermochromic material that changes transmission in the visible [39]. This way the overall solar modulation can be improved significantly. However, with tinting in the visible an additional effect is added to the coating, which changes the application area and is therefore not directly comparable.

Another strategy is to take advantage of the changing solar angle of incidence in summer and winter. Here, periodically patterned VO₂ coatings are prepared that enable transmission of sunlight at low angles of incidence without blocking by the VO₂ layer, whereas at higher angles of incidence, the light hits the patterned surface at the VO₂ domains and is therefore blocked [40]. Theoretically very high ΔT_sol. values of up to 25.8% can be achieved with this strategy. However, the concept still has to be validated in real-life experiments.

Furthermore, optimization of thermochromic properties within an IGU can be achieved by focusing on optimizing the optical properties of the low-e coating. In this study, we have already selected the most suitable commercially available low-e coating to be combined with a thermochromic coated glass. However, still, a significant reduction in solar modulation is observed due to the IR-blocking properties of the low-e coating. To further improve the thermochromic properties of the complete IGU, it is also possible to further increase the transmission of the low-e coating. Here it is crucial to retain the low emissivity properties, to not reduce the insulation value of the glazing whilst increasing the thermochromic performance.

4. Conclusions

In conclusion, we have investigated the theoretical maximum performance of thermochromic VO₂ coatings within a state-of-the-art IGU in detail. We have demonstrated that by focusing on the IR region of sunlight alone, the potential modulation range is limited to 39.1% for a single clear glass pane. We have shown that this range is further reduced by the low-e coated glass in HR++ glazing, which is necessary to reach the market standard insulation properties of IGUs. A careful selection of the most suitable low-e coating can limit this reduction and still enables a potential solar modulation of 19%, which can be increased to 23.1% when a small modulation of 10% in the visible is allowed. Larger modulation of visible light is not desired since the appearance of the IGU should remain (nearly) unchanged. In addition, we have demonstrated that our current best-performing VO₂ coating shows a solar modulation of 13.4% in combination with a low-e coated glass within an IGU. This represents a performance of 59% of the theoretical maximum. Lastly, we have discussed various strategies to further improve the thermochromic performance of such an IGU to proceed toward the theoretical maximum performance. Here especially increasing the purity of active material, and optimizing active domain size, size distribution, and layer thickness seem to be the best strategies to reach this goal.