Enhanced Color‐Preserving Radiative Coolers for Versatile Architectural Applications

Global climate crises are the most significant challenges to be solved these days. As one of the technological endeavors to tackle the issue, radiative cooling is amongst the most attractive approaches for sustainable heat energy regulation, which involves maximizing solar heat reflection and thermal heat emission. These green technologies inevitably require architectural applicability, considering that building facades take a large proportion of the heat‐radiating surfaces. For mass‐production suitability and durability, radiative coolers (RCs) fabricated in a fully ceramic context are recently suggested, featuring scalable, thermally insulative, and non‐shrinking advantages. However, the visual effects are also imperative for architectural instances but are seldom accounted for. In this context, this article suggests the enhanced color‐preserving radiative cooling (ECRC) structure for practical architectural applications of glass‐infiltrated ceramic RCs. By simply blending ceramic pigment into the uppermost porous alumina layer, the ECRC structure can maintain the physical, and thermal features of all‐ceramic RC, while exhibiting color by visible reflectance adjustment. ECRCs exhibit an additional cooling performance of up to ≈17.3 °C depending on their color, compared to their conventional counterparts. With additional chromatic features, ECRC can further enhance the availability of radiative cooling technology for practically realizing the energy‐saving structures in real‐world architectural circumstances.


Introduction
A solution for global energy/climate crises is one of the most anticipated subjects in contemporary society.Provided that the DOI: 10.1002/adom.202400144thermal regulation in architectural instances accounts for a significant proportion of the entire energy consumption (≈40%), [1] a sustainable method of heat energy management is crucial for alleviating environmental issues.In the context of developing green materials for commercial and residential architectural constructions, passive radiative coolers (RCs) are in the limelight for their promising advantages in dealing with heat piles without consuming additional energy.Unlike conventional heat regulation methods that entail an immense amount of collateral energy consumption in countering the natural heat flow, radiative cooling technology provides a viable option for alleviating heat accumulation in a passive manner.Specifically, the technology involves the engineering of optical properties: spectral reflectance and emissivity.The designed structure can minimize heat absorption by maximizing reflectance at the solar spectrum (0.3-2.5 μm) while maintaining sufficient thermal emission toward the cold outer space (≈3K) by focusing emissivity at long-wave infrared (LWIR) spectral region where the atmosphere exhibits high transmissivity (i.e., atmospheric window). [2]With this mechanism, RCs can block radiative heat gain from solar radiation and dissipate the thermal energy of an object in a radiative form.9][10] Moreover, material engineering, [11][12][13][14] affordability, [5,15] and large-scale producibility [16][17][18] enhancements have been examined for successful plug-in of radiative cooling technology onto the architectural application.
From several factors facilitating the implementation of radiative cooling technology, porous polymer structures, and ceramic materials were suggested as strong candidates.However, even though the former has great advantages in thermal insulation and outstanding spectral characteristics for radiative cooling, the chemical/mechanical instability made the sole porous polymer structure prone to degradation.On the other hand, ceramic materials have strengths in their intrinsic flame endurance, corrosion resistance, and physical robustness, [18][19][20][21][22] but still have shrinkage problems during the fabrication process [23] which could be detrimental to spectral characteristics required for radiative cooling. [24]Recently, it has been reported that these two structures can be employed together to complement each other by enabling microscopic clutch: alternately stacking the porous alumina layer and porous borosilicate glass layer and infusing the glass particles into the porous alumina layer through a co-firing process.The whole-ceramic porous RC successfully maintains the advantages of each separated layer, while compensating for the drawbacks. [25]till, for the further practical applicability of the radiative cooling structures for architectural purposes, the visual aspects should also be considered. Table S1 (Supporting Information) summarizes the recent progress in colored radiative coolers.In this sense, we suggest an enhanced color-preserving radiative cooler (ECRC), which effectively adjusts the visible optical property for obtaining various colors while maintaining cooling performance to a proper extent, without significantly complicating the fabrication process of the whole-ceramic porous RCs.The resultant structure exhibits up to 17.3 °C lower equilibrium temperature compared to the conventional structure fabricated only with ceramic pigments and polymer additives (i.e., green sheet), depending on its color.To top it off, along with the excellent adiabatic property, [25] the robust physical feature and suppressed shrinkage of ECRC during the ceramic fabrication process further enhance the practical applicability of the radiative cooling technology in architectural circumstances.

Results and Discussion
The method of implementing color presentation of fully ceramic porous RCs with a subtle modification in the fabrication process, and the cooling performance analysis of the resultant structure with various visible reflectance management is demonstrated in this study.The entire structure comprises multiple stacks of porous ceramic layers; the porous Al 2 O 3 layers and the porous borosilicate layers.][30] On top of that, the sintering process enables the borosilicate particle agglomeration onto adjacent porous Al 2 O 3 layers, which reinforces the physical strength and non-shrinkable feature of the ECRC. [25]From here, the color presentation can be achieved by blending ceramic pigment particles within the uppermost porous Al 2 O 3 layer, in the batch/powder preparation stage in the fabrication process.
Radiative cooling exploits the radiative heat exchange mechanism, by minimizing heat gain and intensifying the heat emission.Generally, reflecting most of the solar radiation prevents the heat absorption of the structure, which is optimal for enhancing cooling performance.However, since the visible spectrum region is included within the solar radiation region, fine-tuning the visible spectral reflectance is crucial for achieving both sufficient cooling performance and color presentation.In other words, it is necessary to find the optimal condition in the trade-off relationship between these two aspects.Hence, to sufficiently preserve radiative cooling performance, ECRC should ideally maintain a high reflectance at the ultraviolet and near-infrared (UV/NIR) spectral region and high emissivity at the LWIR spectrum as much as possible.Figure 1a depicts the relationship between radiative heat energies and visible light, where the solar radiation is mostly distributed in the ultraviolet and visible (UV-vis; 0.28-4 μm) spectral region and the thermal radiation from the panel surface takes LWIR spectral range (8-13 μm).The UV-vis-NIR inset shows the ECRC panel ideally reflecting entire UV/NIR light and selectively reflecting visible light, from incoming solar radiation.The panel with a certain color partially absorbs visible light for the instance, hence it is optimal for the panel to minimize this heat absorption and increase the amount of reflected light for better cooling performance.The LWIR inset shows the thermal emission focused at LWIR successfully being conveyed toward outer space, through atmospheric windows.The sample should ideally exhibit focused emissivity at this spectral region to avoid heat being dissipated within the atmosphere.
Figure 1b shows the simplified structure of ECRC and the brief method of endowing color scheme.The base, or lower part of the entire ECRC is constructed with a form of alternating porous Al 2 O 3 and borosilicate layers bound by the infiltration of borosilicate particles into the void area of adjacent porous Al 2 O 3 layers.On top of the base structure, an additional porous alumina layer blended with ceramic pigments is stacked and sintered for desired color exhibition.Three colors are demonstrated in this research, which are brown, green, and yellow.These color-dyed porous Al 2 O 3 layers mostly regulate the visible reflection, as shown in Figure 1c.The upper inset shows the overall spectral reflection profiles of three different color cases, with a ceramic pigment content ratio of 75%.These ECRCs commonly exhibit low reflectance (and high emissivity) at the LWIR region and mostly differ at the solar spectrum.The visible and NIR reflectance spectra for each of the cases are compared with that of corresponding conventional color samples in the lower inset, per color.The green sheets, selected as the conventional counterparts, refer to a tape-cast layer of ceramic material that has not been fired (i.e., without sintering at the last step).The geometries of conventional samples are displayed in Figure S1 (Supporting Information).These sheets include a polymer binder, which results in a low refractive index contrast between Al 2 O 3 and binder materials.In contrast, the ECRC samples undergo an additional sintering process that burns off and removes all polymer binders.This process enables a high refractive index contrast between air voids and Al 2 O 3 particles, which triggers strong multiple Mie-scattering effects.In addition, the multiple Al 2 O 3 -borosilicate (BS) stack layers in ECRC enhance NIR reflectance more owing to the gradually changed particle sizes (Figure S2, Supporting Information).Owing to them, the ECRC samples show enhanced NIR reflectance in a range over ≈1.5 μm, which greatly reduces solar heat absorption.With this optical improvement, brown, green, and yellow ECRC samples exhibit less solar absorption than their conventional counterparts, respectively.Also, due to the enhanced emissivity at LWIR facilitated by the porous configuration, ECRCs can perform better than the conventional samples in terms of the thermal emission perspective.Figure 1d shows the calculated solar absorption and thermal emission for each color occasion.For 75% dye ratio samples of brown, green, and yellow cases, the ECRC takes 22.8%, 11%, and 47% less solar absorption while sufficiently maintaining LWIR thermal emission, compared to their color equivalent conventional samples, each.Figure 1e shows the fabricated ECRC samples, for demonstrated three color cases.
Along with the adjustment of the content ratio between Al 2 O 3 particles and ceramic pigment, the annealing temperature in the fabrication process is vital for cooling performance, according to the former research. [25]Specifically, the different annealing temperatures cause.The comparison of spectral reflectance spectra and calculated thermal equilibrium temperatures with different ceramic pigment ratios and annealing temperatures for each color scheme are demonstrated in Figure 2. Detailed overall reflectance data of samples are provided in Figure S3 (Supporting Information).The notable change in spectral reflectance with higher annealing temperature is the reflection enhancement at the NIR range.This result is attributable to the borosilicate infiltration toward the nearby porous alumina layer further facilitated at a higher annealing temperature of 1000 °C.Moreover, the UVvis reflectance of samples with 25% ceramic pigment proportion exhibits higher than that of counterparts, for all color cases.These tendencies of reflectance change with different annealing temperatures and ceramic pigment ratio mostly focus spectral reflectance enhancement at wavelength region where the solar radiation intensity peaks (≈0.5 μm), to reduce the solar absorption, which is the most crucial factor for radiative cooling.As a result, even though the optimal cases of cooling behavior slightly differ with color schemes, the samples annealed at 1000 °C exhibit low equilibrium temperature, in general.Besides the cooling performance, maintaining an annealing temperature of 1000 °C in fabricating the sample provides collateral advantages including enhanced physical strength (i.e., specific strength) and thermal insulation, which is also attributable to the sufficient borosilicate particle infiltration. [25]Consequently, choosing a 1000 °C annealing temperature in the manufacturing process has been proven to be optimal for infiltrated ceramics.By inducing the softening of borosilicate glass above 980 °C (Figure S4, Supporting Information), a durable sheet is formed without the occurrence of multi-layer separation, achieving both comprehensive stability and excellent cooling performance simultaneously.Also, the mechanical and thermal durability of ECRC is excellent owing to the infiltration of BS into Al 2 O 3 layers (Figure S5, Supporting Information).
From the measured reference reflectance spectra data of the ECRC samples (annealed at 1000 °C) with the ceramic pigment content ratio of 25% and 75% as well as the conventional green sheet samples, the correlation between color hue and cooling performance is demonstrated in Figure 3. Since the solar irradiance is primarily concentrated in two spectral regions such as visible (400-800 nm) and NIR (800-2500 nm) regions, it is crucial to maintain high reflectance in the NIR spectrum for ECRC. Figure S6 (Supporting Information) is provided to illustrate the spectral reflectance and solar irradiance within the visible spectrum, highlighting how the reflected visible light from the radiative cooler defines the observable colors.
To carry out the gradual color hue change along with the progressive mutation of spectral reflectance between the two ECRC samples with different ceramic pigment content, the interpolated intermediate reflectance spectrum steps are depicted in Figure 3a-c for brown, green, and yellow, respectively.The dotted line shows the conventional counterparts for each color case.The group of reflectance interpolated from that of the original ECRC mostly exhibits higher reflectance in the UV-vis-NIR spectral region, compared to that of conventional samples.Within the intermediate reflectance group, the visible reflectance spectrum tends to get higher as the color brightness increases or the ceramic pigment content ratio decreases.Color of the ECRC samples is analyzed based on the visible spectral reflectance by extracting RGB code using the CIE 1931 color space model, [31] and the resultant RGB codes are presented in Figure 3d-f, also for 25% and 75% color pigment portion samples in each of color cases.The color similarity examination between samples with different annealing temperatures and ceramic pigment content ratio within the same color scheme is demonstrated in Figure S7 (Supporting Information).
Under simulation configurations with the ambient temperature set as 300K and non-radiative heat exchange coefficient value as 7 W m −2 •K, the calculated equivalent temperature of samples corresponding to each interpolated reflectance spectrum are shown in Figure 3g-i for brown, green, and yellow cases, in turn.The detailed process for calculating equivalent temperature is described in the Experimental Section.The color of the bars represents the equivalent color from the RGB code extracted from the visible reflectance.From the simulation result, ECRCs of 25% and 75% color proportion and the interpolated models between two ECRCs exhibit better cooling performance compared with the conventional sample.As the reflectance approaches that of 25% pigment sample, the UV and visible reflectance increases and blocks the solar absorption more efficiently, considering that the intensity of the solar radiation is mostly concentrated on 0.28-2.5 μm, which covers the UV and visible spectral region.For brown color, the temperature of the conventional sample at the thermal equilibrium state is 63 °C, and the brown ECRC samples exhibit 11-17.3°C lower equilibrium temperature, which is largest when the ceramic pigment proportion of the ECRC sample is 25%.This tendency also similarly holds for green and yellow cases, where the green sample benefits from 4.5-10.1 °C (conventional green sample temperature as 50.7 °C) and the yellow sample for 4.3-9.9°C (conventional yellow sample temperature as 30.8 °C), and the maximum is achieved for 25% ceramic pigment content ratio ECRC samples for both cases, likewise.On the other hand, between the three color schemes, the best cooling performance is expected for yellow samples, since the visible reflectance spectrum is significantly higher compared with other color cases.Figure S8 (Supporting Information) illustrates the correlation between cooling and coloration in the three ECRC samples.Since the solar irradiance level is maintained at >60% of its peak value within most of this visible spectral region, it is crucial for a sample to exhibit high reflectance value for wider wavelength bandwidth.Hence, the saturation and value in the HSV code are expected to play a significant role in determining the capability of reducing solar radiative heat absorption and achieving the radiative cooling performance of the sample.
The overall fabrication process of ECRC samples along with the structure deformation examination is demonstrated in Figure 4.As from Figure 4a, the ceramic materials used in ECRC are processed with ceramic slurry preparation and ball milling stage as specified in Figure S9 (Supporting Information), especially mixing the ceramic powder (i.e., Al 2 O 3 and borosilicate) with a binder for the tape casting process.Specifically, Figure 4a provides an overview of the processes involved in the tape-casting technique for producing ceramic panels with a multi-layered ECRC structure.The manufacturing process follows a sequence of ball-milling, tape casting, stacking, dicing, and subsequent thermal annealing.Initially, individual preparation of each type of powder is carried out, including colored ceramic (from Foshan Pampascolors Co., Ltd, China), Al 2 O 3 (from Sumitomo Chemical Co., Japan), and BS.A ceramic pigment is typically described as a complex metal transition oxide acquired through the calcination process, known for its thermal and chemical stability features. [32]hree commercially available ceramic-based pigments are used for each color: yellow (Pr-Zr-Si), brown (Fe-Cr-Zn), and green (Cr-Ti).
These powders are then arranged in alternating order, with a colored ceramic sheet on top and a BS-Al 2 O 3 alternating layer at the bottom.Lamination is employed to bond these layers before undergoing the firing process, resulting in the formation of the complete ceramic structure.During the batching phase, a plasticizer-to-binder ratio of 0.4 is maintained, and the batch undergoes two milling cycles to stabilize the slip.The ball milling process occurs over 24 h at 120 rpm per cycle.Following this, a deformation process is applied to achieve a high-quality ceramic slip.In the ECRC structure, each layer is formed using the tape casting method, with colored ceramic comprising the topmost layer.The subsequent layers follow a repeating sequence, alternating between BS and Al 2 O 3 , such as BS-Al 2 O 3 -BS-Al 2 O 3 .This stacking arrangement is repeated five times.The final sample undergoes sintering at temperatures exceeding 800 °C, enabling BS particles to permeate the gaps between neighboring colored ceramic and alumina particles.Throughout the co-firing procedure, BS particles infiltrate the interlayer spaces, contributing to the strengthening of the structural integrity.Also, this sintering temperature controls the amount of infiltration of BS into Al 2 O 3 , leading to the spectral difference as shown in Figure 2.
For the uppermost color porous Al 2 O 3 layer, the ratio between ceramic pigment and Al 2 O 3 powder can be adjusted for the resultant sample to exhibit different color hues and saturation.After separately generating each layer, the whole structures including multiple porous Al 2 O 3 layers, borosilicate layers, and a color-dyed ceramic layer, are stacked and laminated for physical contact.The resultant sheet is then prepared by dicing into sufficient size (i.e., 3 × 3 cm 2 ) and finally sintered in a furnace.The thermal decomposition process occurs in this sintering process at the minimum annealing temperature of 600 °C, and changing the annealing temperature yields a different amount of borosilicate particle infiltration.Configuring annealing temperature between 800 and 1000 °C assures the meltdown of borosilicate while the Al 2 O 3 particle maintains its porous structure.The exploded view of the resultant structure is depicted in Figure 4b, which comprises multiple porous Al 2 O 3 layers and porous borosilicate layers, and the colored ceramic sheet layer is additionally stacked at the uppermost region.The righthand inset shows the 3D  For both samples with the Al 2 O 3 particle ratio of the uppermost layer prepared as 25% and 75%, the hole size is maintained even when the sample is annealed at 800 and 1000 °C, without significant change from the green sheet (i.e., generated sheet before co-firing process), as shown in Figure 4d.Especially, for the case of annealing temperature at 1000 °C, the initial and final hole size is shown in Figure 4e, for samples of both Al 2 O 3 content ratios.These non-shrinkable features further facilitate the scalability of ECRC sample fabrication.
Moreover, for the versatile usage of the structure in the curvature of the building surfaces, ECRC exhibits additional advantages in its fabrication process, which readily avails the features, as shown in Figure 4f-g.The green sheet: the sample fabricated without the annealing process, has flexible and bendable features, which enables the structure to be modified into a curved shape.Annealing the shaped green sheet sample then yields the fixed structure with the desired curvature.Again, the nonshrinkable feature of the multiple Al 2 O 3 -BS stack facilitates the generation of curved ECRC samples, only with the additional fabrication step of bending the green sheet.
To experimentally confirm the cooling performances of ECRC samples with different colors and ceramic pigment ratios, outdoor measurements of temperature were conducted on the rooftop of PNU (Pusan National University, 35.23 335°N, 129.08 419°E) over the daytime from 12:00 to 14:00.The overall measurement setup is shown in Figure 5a, where the ambient temperature, solar irradiance, and temperatures of six samples (i.e., samples with ceramic pigment content ratio of 25% and 75% for three color schemes) are measured side by side.As shown in the lower inset, samples are physically supported on the acrylic exterior through the acrylic connector, and thermocouples are attached on the bottom side of the samples to measure the sample temperature over time.Polystyrene material blocks the thermal conduction between thermocouples and lower interfaces.The entire acrylic chamber is tilted at ≈48.5°and its upper surface is directed toward the north sky per the AM1.5G reference.To examine the estimated temperatures of conventional colored samples (i.e., green sheet samples) with ECRCs, solar irradiation, and ambient temperature are also gauged in the same time interval.A pyranometer (CMP6, Kipp & Zonen) is installed next to the acrylic chamber with samples, maintaining identical slope angles and facing direction for measuring solar irradiance reaching the ECRC samples.To quantify the ambient temperature, a thermal air sensor is inserted in a paper box coated with Aluminum.The ambient air sensor setup can sufficiently forestall the sensor directly being overheated by solar irradiance and obtain the temperature value of air naturally flowing within the air chamber.
Figure 5b shows the measured temperature result of six representative ECRC samples for each annealing temperature case composed of 75% and 25% ceramic pigment ratio samples per brown, green, and yellow color instances, along with the measured solar irradiance value with time elapsed from 12:00 to 14:00.The measurement for three annealing temperature groups of 800, 900, and 1000 °C were conducted on September 3rd, July 27th, and September 9th of 2023, respectively.From the measured solar irradiance and ambient temperature, equilibrium temperatures of green sheets are estimated to compare with measured sample temperatures averaged in terms of time for each color and annealing temperature cases and examine the cooling performances of the ECRCs (the value of non-radiative heat exchange coefficient used in the estimation is 7 W m −2 •K).As shown in Figure 5c, ECRC samples show up to 17.3 °C lower temperature than their counterparts, depending on their color schemes and sample annealing temperature.The reason for such a large temperature difference mainly originates from a significant NIR reflection enhancement of ECRC compared to that of the conventional one as shown in Figure 1c.The ECRCs show decent cooling performances compared to their corresponding conventional counterparts.Also, the temperature variation can result from the amount of non-radiative heat exchange such as convection and conduction (Figure S10, Supporting Information). [33]his underscores h c as a pivotal factor, with solar irradiance being the subsequent influential element.However, h c alone cannot convert the ECRC from heating to cooling mode or vice versa; rather, it helps the ECRC's temperature stabilize near ambient levels.

Conclusion
With maintaining excellent thermally insulative features and physical strength of multiple Al 2 O 3 -BS stacks, we devised an ECRC as a color-preserving radiative cooler by regulating visible spectral reflectance by adjusting the content ratio of ceramic pigment blended in the uppermost porous Al 2 O 3 layer.We have examined cooling performances for various instances of color exhibition using three representative color cases-brown, green, and yellow-with different ceramic pigment content ratios.The features of the ECRC can be summarized into two major perspectives: maintaining sufficient radiative cooling performance while developing a variety of color exhibitions, and the availability of a practical fabrication process.ECRCs can sufficiently maintain their passive cooling performances by effectively reflecting solar irradiance and reducing solar heat absorption compared to the green sheet conventional counterparts while focusing thermal emission at the LWIR spectral region.These performances are evaluated by comparing the equilibrium temperature with that of the conventional samples.
The fabricated ECRC samples exhibited an equilibrium temperature of 4.3-17.3°C lower than that of the conventional color samples, depending on their color schemes and annealing temperature configurations.Moreover, with the virtue of the borosilicate infiltration, fabricated ECRC samples exhibited excellent shrinkage-preventing characteristics, whereas the punch-hole arrays showed minimal size deformation after the annealing process.By utilizing an identical tape-casting fabrication method, the color presentation can be achieved by additionally conducting a color-dye blending process in the powder preparation stage, without excessively complicating the entire process.
From various design choices of different ceramic pigment content proportions and annealing temperatures, we have conducted cooling performance comparisons for different ceramic pigment content ratios.For samples with a ceramic pigment ratio of 25% compared with 75% cases, even though the amount of LWIR heat emission is slightly dropped due to the lower emissivity at the LWIR spectral region, the sample with a lower ceramic pigment ratio exhibits significant solar absorption blockage and shows better cooling performance.We also have examined the optimal annealing temperature considering both cooling performance and physical durability along with the additional possibility of practical application by implementing curvature of the sample within the same fabrication context.To sum up, ECRC can further provide the availability of fully ceramic radiative cooling technology in various architectural instances.

Experimental Section
Calculation of Cooling Energy Saving and Equilibrium Temperature: The equilibrium temperatures of samples are calculated based on the heat exchange mechanism between the emitter, atmosphere, and the sun. [34,35]he radiative cooling strategy involves regulation of the amount of head radiation from the sample (E emit ) and heat absorption from the surrounding atmosphere (E amb ) and the sun (E sun ).Besides the exchange of radiative heat components, the heat energy movement caused by convection and conduction also plays a crucial role in determining the equilibrium sample temperature.These non-radiative heat components (E loss ) are determined by the ambient temperature and non-radiative heat exchange coefficient (h c ), which is affected by humidity and wind speed, according to the convection and conduction heat transfer model. [35]Each heat component can be expressed as following Equations (1-4), where I BB means spectral radiance of a black body and I AM1.5G represents spectral solar irradiance in accordance with AM 1.5G reference.
The net cooling power of the sample is determined as the heat emitted from the sample subtracted by the heat components the sample absorbs from surrounding environments, which can also be expressed as Equation (5).The equilibrium temperature of the sample is the sample temperature where the net cooling power is zero.If the sample temperature is lower than the equilibrium temperature, the net cooling power is lower than zero, hence the sample absorbs heat from the surroundings and results in the sample temperature approaching the equilibrium temperature.Likewise, if the sample temperature is higher than the equilibrium temperature, the net cooling power is higher than zero and the sample emits excessive heat, which also provokes the sample temperature to approach the equilibrium temperature.
RGB and HSV Color Code Extraction: One of the major points this study reveals is the relevance between the color of samples and the cooling performance for each of them, which is mostly caused by the spectral reflectance of the sample (Figure S8, Supporting Information).To determine the color exhibition, the RGB color codes were calculated for visual color evaluation, and HSV values for finding a correlation between cooling performance and color exhibition.The color code extraction method is based on CIE 1931 color analysis, [36] where each portion of red, green, and blue light perceived by the human eye is considered with color matching functions for each basic color; red, green, and blue.The spectral reflectance value is multiplied by color matching functions of red, green, and blue, and integrated over the visible spectral region as expressed in Equation ( 6), which respectively returns the portion of each basis color.
Specifically, for a method of embedding the color recognition of reflected light in the human eye, the color coordinates were first obtained of each light reflected from samples for in CIE 1931 color space, using the tristimulus color matching functions corresponding to red, green, and blue, which are x(), ȳ(), and z(), respectively.These tristimulus colormatching functions-which can be seen as spectral filters-involve the color reaction of cone cells in the eye, for the standard observer's perspective angle within an angular region of 2°.The measured reflectance data are used as spectral radiance I().The resultant values X, Y, and Z in each integral are CIE color coordinates.
0.4124 0.3576 0.1805 0.2126 0.7152 0.0722 0.0193 0.1192 0.9505 There are many approaches available in converting these coordinate values into conventionally used RGB codes, and sRGB representation of colors in the work is used, with a standard observer perspective of 2°angular range and reference white point of D65; where the RGB values can be obtained through linear transformation as expressed in Equation (7).These RGB codes are significant in the context of evaluating the color exhibition since these directly correlate with how colors are displayed.
Besides the RGB codes for color exhibition, also used another color analyzing method: the HSV code, can also be derived within the context of obtained RGB codes. [37]Each value of the HSV code represents the hue, saturation, and value of a color, which intuitively depicts the color in the sense of color perception of the human eye.First, color hue intuitively signifies the aspect of color, and is closely related to the wavelength in the visible spectrum, or the concentration amount of reflected visible light at a certain wavelength.On the other hand, the saturation and value respectively represent the vibrancy (purity) of the color and the brightness (intensity) of the color, respectively.These two values are deeply related to the overall level of reflectance in visible spectral regions.
The choice of additionally employing color analysis in terms of HSV code eases the qualitative analysis of visible solar reflectance and color exhibition, since these values better show how the spectral reflectance of a sample is characterized within visible wavelength region compared to traditional RGB values and enables analysis the impact on radiative cooling performance due to the change in solar irradiance absorption.For minimal solar absorption in the visible spectral region, the sample should exhibit high reflectance for wider wavelength bandwidth.Since saturation is related to the focusing of the spectral reflectance at a certain wavelength point and value as the overall reflectance level, the tendency of higher value and lower saturation yields higher overall solar reflection in the visible region and hence lower absorption of solar irradiance.
Modeling Non-Radiative Heat Exchange Coefficient for Equilibrium Temperature Estimation: For estimating the radiative cooling performance of the conventional samples in Figure 3, the thermal equilibrium temperature of each sample was calculated by using heat exchange equations, where the spectral reflectance data of the conventional samples and measured ambient temperature are plugged in along with the solar irradiance.Although the radiative heat is the main context focused on, the non-radiative heat exchange is also significant, and is strongly dependent on the climate circumstances.(i.e., wind speed, humidity, atmospheric pressure, and ambient temperature) As from the heat exchange equations, the non-radiative heat (which is denoted as E loss ) can be expressed as the following equation: Here, the term h c represents the non-radiative heat exchange coefficient, which also includes the climate data where the heat exchange mechanism between a sample and the surrounding environment holds.Hence, from the obtained climate data, the value of h c can be approximately obtained, with several assumptions for simplifying the modeling process while not significantly deviating from the exact value.In other words, the purpose of modeling the h c value is to properly estimate the actual nonradiative heat exchange characteristics appearing in time intervals when the measurements of ECRC temperatures are conducted, by utilizing the real-time climate data including ambient temperature, humidity, atmospheric pressure, and wind speed.Herein, the estimation is continued by using the empirical approach of approximating the value of h c , which utilizes the Prandtl number, Reynolds number, and Nusselt number. [38,39]irst, it is assumed that the properties of air (i.e., thermal conductivity, Prandtl number, and kinematic viscosity) used in calculations maintain their constant value within the temperature and pressure range.The thermal conductivity of air k air is approximately 0.026 W m•K −1 , which is nearly constant in the temperature range between 300 K and 900 K. [40] Also, the Prandtl number P r , which is defined as the ratio of momentum diffusivity to thermal diffusivity, is assumed to maintain its value at 0.707.Lastly, the kinematic viscosity of air  air is considered constant as 1.48 × 10 5 m 2 s −1 , which is its value at ambient temperature of 15 °C and 1 atm.With determining the reference length L into 3 cm (= 0.03 m) as a geometrical heat exchanging setup of a sample, the Reynolds number can first calculated with the following equation: It is then determined which correlation method to use, depending on the value of the calculated Reynolds number.The scaling factor C and exponent m obtained by empirical analysis are then chosen.If R < 5 × 10 5 , the corresponding C and m values are 0.664 and 0.8, respectively.On the other hand, if 5 × 10 5 ≤ R, the alternative C and m are set as 0.037 and 0.8, in turn.Lastly, by calculating values in the previous process, the Nusselt number can calculated, and finally, the convective heat transfer coefficient h c , using the following equations.Fabrication Method for Glass Frits: As the characteristics of glass in the glass infiltration method significantly impact ceramic sheet manufacturing, the desired properties at this stage include a low coefficient of thermal expansion, mechanical strength, and amorphous structure.Therefore, to meet these specifications, borosilicate glass (SiO 2 -B 2 O 3 -R 2 O: R = Li, Na, K, Ca) was manufactured with high refractoriness, thermal shock resistance, and thermal conductivity.In the manufacturing process (See Figure S7, Supporting Information), a BS glass frit was optimized by adjusting the composition to SiO 2 63 wt.%, B 2 O 3 24 wt.%, adding 1 wt.% of Al 2 O 3 to increase the operating temperature range, lower the melting point of glass, and enhance viscosity by increasing the composition of alkali materials.Subsequently, the dry mixing of these materials was performed using a powder mixer at 700 rpm for 15 min, followed by melting in a high-temperature lift-bottom furnace at 1500 °C for 30 min.Water quenching was then conducted.The quenched glass was initially crushed using a disk mill and further ground through ball mill and attrition mill processes for 24 h, ultimately producing glass frit with a particle size of 1.6 μm.
Fabrication of Borosilicate Glass: The raw materials for borosilicate glass comprised 81.0 wt.% SiO 2 , 12.5 wt.%B 2 O 3 , 4.0 wt.% Na 2 O, and 2.5 wt.% Al 2 O 3 , dry-mixed in a powder mixer at 700 rpm for 15 min.The mixture underwent a melting process at 1500 °C for 30 minutes in a hightemperature lift-bottom furnace, followed by water quenching to obtain the glass melt.The quenched glass frit underwent size reduction through dry grinding using a disk mill and wet grinding using a ball mill, followed by fine grinding with an attrition mill to produce BS frit.
Fabrication of Green Sheet: The solvent mixture for each ceramic green sheet consists of 60 vol.% toluene and 40 vol.%ethanol, and it includes a dispersant (BYK-111, BYK Chemical, Germany) at a 1 wt.% concentration to ensure slurry dispersion.The binder system employed polyvinyl butyral (PVB, Sekisui, Japan) as a binder and dibutyl phthalate (DBP, Dejung, Korea) as a plasticizer, maintaining a fixed plasticizer/binder ratio of 0.4.The binder content included 30 vol.% ceramic powder in the slurry.Subsequently, the ceramic powder, dispersant, and solvent underwent primary milling using a 10 mm diameter zirconia ball in a ball mill for 24 hours.Additional solvents, binder, and plasticizer were added, and secondary milling took place for another 24 hours.After a 15-m defoaming process to remove gas bubbles, the slurry was stabilized through 24-hour ball milling at 25 rpm, resulting in a slurry viscosity of ≈2000 cps.The ceramic sheets were shaped using a tape caster based on the doctor blade method of tape casting, with casting conditions including a speed of 2 m min −1 and a three-zone drying temperature profile of 35-60-75 °C.The tape width was 150 mm, and the green sheet thickness was set to 100 μmcolored sheets, 100 μm-Al 2 O 3 , and 80 μm-BS, creating a sandwich structure.The stacking conditions were 5 MPa, 60 °C, and 1 min using a manual stacker.The laminated structure was cut in the x-y direction using a blade cutter with a length of 30 mm.To eliminate excess organic materials within the laminate, it was heated at a rate of 3 °C min −1 up to 600 °C and held for 2 hours for binder burnout, followed by cooling.The sintering process involved heating the binder-burnt specimen on a zirconia substrate up to 1000 °C at a rate of 2 °C min −1 and holding it for 2 hours.

Figure 1 .
Figure 1.Color-preserved passive radiative coolers for practical applications for architectural instances.a) A concept illustration of an enhanced colorpreserving radiative cooler (ECRC) exhibiting UV and near-infrared (NIR) reflection, selective visible reflection, and thermal insulative feature.b) Exploded view of multiple porous Al 2 O 3 and borosilicate stack covered with colored porous Al 2 O 3 layer selectively reflecting visible light from incoming solar radiation.c) Spectral reflectance data of ECRCs and magnified view of visible-NIR reflectance spectra for brown, green, and yellow color schemes.d) Calculated long-wave infrared (LWIR) heat emission and solar absorption bar plots comparing conventional green sheet color samples and ECRCs.e) Pictures of fabricated ECRC samples of three different color schemes with ceramic pigment content ratios of 25% and 75%.

Figure 2 .
Figure 2. Annealing temperature and cooling performance.a-c) Spectral reflectance data of samples with ceramic pigment content ratio of 25% and 75%, for different sintering temperature cases of 800, 900, and 1000 °C in brown, green, and yellow color schemes.Insets show the amount of solar irradiance absorption for each sample.d-f) Calculated equilibrium temperatures for corresponding spectral reflectance data.The bars at each group represent the equilibrium temperature of 25%, 75%, and the averaged value, respectively.

Figure 3 .
Figure 3. Correlation between color hue, saturation, and value (HSV) code and cooling performance.a-c) Gradually changing spectral reflectance data interpolated from measured reflectance of 25% and 75% ceramic pigment content ratio samples and conventional green sheet samples for brown, green, and yellow cases.d-f) HSV color code of 25% and 75% ceramic pigment ratio samples extracted from visible spectral reflectance.g-i) Calculated equilibrium temperatures from interpolated reflectance spectra for brown, green, and yellow color schemes.Corresponding conventional sample equilibrium temperatures are depicted as dashed lines.

Figure 4 .
Figure 4. Overall fabrication process and demonstration of shrinkage blocking feature of ECRC sample and curvature implementation a) Tape casting fabrication process of ECRC samples.b) Exploded view of resultant ECRC sample comprising multiple porous Al 2 O 3 and borosilicate layers.The rightmost inset depicts the borosilicate particle infiltration toward the adjacent porous Al 2 O 3 layer."NP" indicates nanoparticle.c) A picture of ECRC patterned with holes punched in an array for four colors: white, green, brown, and yellow.d) Demonstration of overall hole sizes changing with three different sintering/annealing configurations for Al 2 O 3 content ratio of 25% and 75%: Green (not sintered; representing conventional sample), annealed at 800 and 1000 °C.e) Microscope images of ceramic hole pattern for ECRC sheet shrinkage at high annealing temperature.f) Shape modifiable green sheet before sintering and g) curved ECRC after sintering.

Figure 5 .
Figure 5. Daytime cooling performances of ECRCs in outdoor experiments.a) Depiction of an outdoor experiment setup comprising an ambient temperature sensing chamber, acrylic table with samples, and pyranometer.b) Measured data of sample temperatures, ambient air temperature, and solar irradiance in a time interval of 12:00 to 14:00.c) Estimated conventional sample equilibrium temperature calculated from measured ambient air temperature and solar irradiance data.The bars corresponding to ECRC samples are the averaged value of measured temperature data from the outdoor experiment.