Scalable, Patternable Glass‐Infiltrated Ceramic Radiative Coolers for Energy‐Saving Architectural Applications

Abstract A huge concern on global climate/energy crises has triggered intense development of radiative coolers (RCs), which are promising green‐cooling technologies. The continuous efforts on RCs have fast‐tracked notable energy‐savings by minimizing solar absorption and maximizing thermal emission. Recently, in addition to spectral optimization, ceramic‐based thermally insulative RCs are reported to improve thermoregulation by suppressing heat gain from the surroundings. However, a high temperature co‐firing process of ceramic‐based thick film inevitably results in a large mismatch of structural parameters between designed and fabricated components, thereby breaking spectral optimization. Here, this article proposes a scalable, non‐shrinkable, patternable, and thermally insulative ceramic RC (SNPT‐RC) using a roll‐to‐roll process, which can fill a vital niche in the field of radiative cooling. A stand‐alone SNPT‐RC exhibits excellent thermal insulation (≈0.251 W m−1 K−1) with flame‐resistivity and high solar reflectance/long‐wave emissivity (≈96% and 92%, respectively). Alternate stacks of intermediate porous alumina/borosilicate (Al2O3‐BS) layers not only result in outstanding thermal and spectral characteristics, causing excellent sub‐ambient cooling (i.e., 7.05 °C cooling), but also non‐shrinkable feature. Moreover, a perforated SNPT‐RC demonstrates its versatility as a breathable radiative cooling shade and as a semi‐transparent window, making it a highly promising technology for practical deployment in energy‐saving architecture.

3 shaped using a tape caster based on the doctor blade method of tape casting.The casting conditions included a casting speed of 2 m/min and a three-zone drying temperature profile of 35-60-75°C.The tape width was 150 mm.The green sheet thickness was set to 100 μm for Al2O3 and 80 μm or 40 μm for BS, creating a sandwich structure.The stacking conditions were set to 5 MPa, 60°C, and 1 minute 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 remove excess organic materials within the laminate, it was heated at a rate of 3°C/min 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 and holding it for 2 hours.

SI Note 2. Long-wave infrared region simulation
Figure S3 presents detailed demonstration of simulation-aided analysis for optical behavior underlying the low reflectance and high emissivity at the LWIR spectral range (8 ~ 13 μm).The wave response drastically changes within the LWIR wavelength region, since the refractive index of Al2O3 exhibits swift transition in the spectral realm.In the case of 8 μm wavelength, the real value of refractive index () maintains sufficiently high value while the extinction coefficient (κ) is nearly at zero.However, as the wavelength gets larger towards 13 μm at LWIR region,  value is suppressed to zero, while κ is gradually increased.In the case of low κ value, the incident electromagnetic wave successfully propagates into the deeper region of our model, due to the antireflective effect from its porous nature and without noticeable decadence.As the extinction coefficient increases, the amount of wave absorption at distributed Al2O3 particles accordingly increases, thus the wave is mostly dissipated at region near the air-structure interface.

Figure S3c
and Figure S3d portraits calculated spatial reflected power at the air-structure interface of porous Al2O3 structure and dense Al2O3 film.Conventionally, general Al2O3 exhibits poor emissivity, and high reflection at the surface.However, owing to the porous structure, Al2O3 structure holds abundant void area between particles, thus mostly eludes reflection at structure surface and successfully absorb penetrated wave through particles surrounding pores.Hence, the porous structure exhibits low reflected power at the air-structure boundary.Measured spectral reflectance of the porous Al2O3 structure and dense Al2O3 film depicted in Figure S3e show significant correspondence to the preceding numerical analysis; spectral reflectance of the powder Al2O3 at wavelength region of 11 ~ 13 μm (where the κ drastically increases while the  swiftly decays) is way lower than that of the counterpart.
Additional field distributions in both model at wavelength of 8 μm, 11 μm, and 13 μm presented in Figure S3f and Figure S3g reinforces the analysis of wave behavior at different configurations.Wave propagates towards deeper region in both structure at 8 μm.However, due to the antireflective effect occurring at porous structure, the field in the structure penetrates even deeper than the dense Al2O3 filled film, with less amount of dissipation.On the other hand, at wavelength of 13 μm, since the incident wave cannot infiltrate into the Al2O3 film structure, the wave is mostly scattered and absorbed at the surface, and exhibits high reflected power near the interface between air and the film.However, in the porous structure, the incident wave can slightly propagate into the void area between the Al2O3 particles, thus the wave is mostly scattered and absorbed at the pore area.Hence, the porous Al2O3 structure shows almost zero amount of reflected power at the air-structure interface.
The spectral thermal emittance is represented by E(, ).Furthermore,   (, ) signifies the blackbody spectral radiance at temperature T, assumed to be 25˚C.simulation is performed by calculating reflectance with the expansion of each alumina particle's radius; a factor p multiplied to the radius of the base particles is increased from 1 to 4, with an interval of 1.The spectra of reflectance are inclined to decrease as the size of particles gets larger.From this perspective, comparing the spectral reflectance of three different temperatures (i.e., 1000 ℃, 1300 ℃, and 1500 ℃), the annealing temperature of 1000 ℃ is optimal for the alumina structure to exhibit efficient spectral reflectance.Step 2).In the previous section, overlapped particles were presented as a single particle; thus, an additional water-shedding algorithm is implemented to separate discrete particles and resolve ambiguity (Step 4).By approximating the shapes of each captured particle as circles, the diameter of each can be defined from its perimeter (i.e., r = l/2π, where r and l represent the perimeter and radius of the particle, respectively) (Step 5).
As a result of particle analysis, the data of particles that appeared on the original SEM image can be extracted (i.e., particle counts, average, and standard deviation of the approximated Three plots present the statistics of projected particles, outermost particles, and pores, respectively.From these statistical values obtained from a partial area restricted to the SEM image region, we can estimate the structural features of the entire cross-section.(c) A logical diagram of the pore accumulation process and the resultant pore film is depicted.Every pore in the film must not overlap while maintaining the statistical characteristics we have previously obtained, namely the mean value and standard deviation of pore radius and the number of pores.
For each iteration, the position of the additive pore is set randomly, and the radius is configured 20 with a Gaussian distribution, which agrees with the measured statistics from the image analyzing process.If the pore overlaps with other pores, the distance between the centers of the two circles is smaller than the sum of the two radii, and we skip assigning and start a new iteration if this is the case.The repetitive process terminates when the pore count reaches the desired number.

Figure S1 .
Figure S1.(a) Illustrations for the additional Al2O3-BS interfaces cascaded to the sandwiched

Figure S2 .
Figure S2.(a) LWIR emittance for Al2O3 green, Al2O3, and Al2O3-BS samples (left) and optical Where, ℎ is the Plank's constant,  0 is the velocity of light, and  is the wavelength,   is the Boltzmann's constant, T is the absolute temperature of blackbody.(b) In-plane and throughplane thermal conductivity for Al2O3 (left) and Al2O3-BS (right).

Figure S3 .
Figure S3.Simulation-aided analysis for optical behavior underlying the low reflectance and

Figure S4 .
Figure S4.Examination on performance degradation of the SNPT-RC sample under pragmatic

Figure S5 .
Figure S5.Global climatic cooling performance simulation.(a) Cooling energy savings of

Figure S6 .
Figure S6.Fabrication process of SNPT radiative cooler.The entire fabrication processes of

Figure S7 .
Figure S7.(a) Measured thermal characteristics of several glass materials for borosilicate (BS),

Figure S8 .
Figure S8.Optical characteristics of alumina and glass materials, and reflectance improvement

Figure S9 .
Figure S9.(a) The refractive indices of Al2O3 and BS, and (b) the radius distribution of alumina

Figure S10 .
Figure S10.Detailed image processing procedure for modeling the Al2O3 layer.(a) The

Figure S11 .
Figure S11.Analysis of pore distribution from SEM image and generation of pore filter.(a)

Figure S12 .
Figure S12.Comprehensive description of generating 2D simulation plane.(a) A region of the

Figure S13 .
Figure S13.Supportive simulation results.(a) Borosilicate deposition is realized in the

Figure S14 .
Figure S14.The thermocouple (ST-50, K-type) is verified to obtain the reliable sample

Figure S15 .
Figure S15.Outdoor measurement without PE film.To ensure the cooling performance of the

Figure S16 .
Figure S16.Through-hole areal ratio calculations with size configurations of areal ratio 5.6%