Structurally Colored Radiative Cooling Cellulosic Films

Abstract Daytime radiative cooling (DRC) materials offer a sustainable approach to thermal management by exploiting net positive heat transfer to deep space. While such materials typically have a white or mirror‐like appearance to maximize solar reflection, extending the palette of available colors is required to promote their real‐world utilization. However, the incorporation of conventional absorption‐based colorants inevitably leads to solar heating, which counteracts any radiative cooling effect. In this work, efficient sub‐ambient DRC (Day: −4 °C, Night: −11 °C) from a vibrant, structurally colored film prepared from naturally derived cellulose nanocrystals (CNCs), is instead demonstrated. Arising from the underlying photonic nanostructure, the film selectively reflects visible light resulting in intense, fade‐resistant coloration, while maintaining a low solar absorption (≈3%). Additionally, a high emission within the mid‐infrared atmospheric window (>90%) allows for significant radiative heat loss. By coating such CNC films onto a highly scattering, porous ethylcellulose (EC) base layer, any sunlight that penetrates the CNC layer is backscattered by the EC layer below, achieving broadband solar reflection and vibrant structural color simultaneously. Finally, scalable manufacturing using a commercially relevant roll‐to‐roll process validates the potential to produce such colored radiative cooling materials at a large scale from a low‐cost and sustainable feedstock.

. Cross-section SEM images of CNC films (top) and CNC-EC bilayer films (bottom). These images were used for the thickness measurement reported in Fig. S2.   Fig. S13. Results of cross hatch adhesion test of (a) blue CNC-EC on a copper sheet with peeled tape on its left and (b) a free-standing blue CNC-EC bilayer films with peeled tape on its right. The structurally colored parts are the samples that underwent the cross hatch test. Fig. S14. Schematic of shear stress test for blue CNC-EC on a copper sheet. Fig. S15. Shear stress vs. displacement for blue CNC-EC on a copper sheet.   Fig. S18. Schematic of the illumination and collection procedures for angle-resolved spectroscopy (also referred to as goniometry), shown relative to a fixed frame of reference for the film. Fig. S19. Specular scans of (a) red, (b) green, and (c) blue CNC-EC bilayer films. Fig. S20. Off-specular scans of (a) an EC film, and (b) red, (c) green, (d) blue CNC-EC bilayer films with the illumination angle fixed at 30° relative to the film normal. Fig. S21. Photograph of the blue CNC-EC bilayer film showing a yellowish color at off-specular angles.

Supplementary Discussions
Note S1. Free-standing CNC chromaticity To confirm the color response to human eyes, the reflectance spectra in the visible range (360-830 nm) of free-standing CNC films are converted to tristimulus value X, Y, and Z by the CIE 1964 color-matching functions ( x , y , and z in Eq. S1-3).
represents the illumination source spectrum and the solar spectrum is used as a natural light condition. The colors can be located in CIE 1946 color space with normalized chromaticity parameters x and y as: The emittance spectrum can be derived from the transmittance and reflectance spectra by the thermal equilibrium principle (Eq. S6) and Kirchhoff's Law, where emittance is equivalent to absorptance.
Theoretical cooling power estimation The net cooling power (P cooling ) of the sample can be estimated from where, P rad : the power density of thermal radiation emitted by the cellulose sample, P sun : the heating power density from solar irradiation, P atm : the power density of downward thermal radiation from the atmosphere, P conv : the effective power density loss including convection and conduction from the cellulose sample, P rad can be derived from the measured emittance spectrum of the sample by BB 0 where ε s (λ) is the sample emittance and I BB is blackbody radiation intensity as a function of emitter temperature according to Planck's law. P sun from solar irradiation is calculated by integrating sample emittance over the Air Mass 1.5 (AM1.5) solar spectrum as, P atm is the amount of power emitted from the atmosphere and absorbed by the sample. The emittance spectra of the atmosphere and the cellulose sample are used (Eq. S10).
where the angular part of the atmospheric emittance can be obtained by where τ(θ) is the angular atmospheric transmittance. [1] The P conv can be evaluated via the sample temperature, ambient temperature, and the effective heat transfer coefficient h as Here, we evaluate and compare the isothermal theoretical cooling power of the samples assuming the same temperature for the sample and the ambient such that the convection loss term can be neglected. Fig. S1. Cross-section SEM images of CNC films (top) and CNC-EC bilayer films (bottom). These images were used for the thickness measurement reported in Fig. S2.  As such the theoretical maximum measured reflectivity at normal incidence is 50% of unpolarized light. Note the baseline arises from the specular reflection of the air-film interface.               , (b) green, and (c) blue CNC-EC bilayer films. The specular reflection from the CNC-EC bilayer film was measured by symmetrically increasing the angles of illumination and collection revealing a single peak at each angle, with a blue-shift at higher angles that is well described by Fergason's law (a modified version of Bragg's law of diffraction). [2] Fig. S20. Off-specular scans of (a) an EC film, and (b) red, (c) green, (d) blue CNC-EC bilayer films with the illumination angle fixed at 30° relative to the film normal. Color bars above each CNC-EC scan indicate the RGB color expected for the spectrum at each angle. At the specular angle (60°) the CNC reflection peak is observed (blue-shifted slightly due to Fergason's law). At most off-specular angles, the complementary subtractive color is observed. Additional spectral features are explained in the schematic in Fig. S22.   [3] (f) Back-reflection from the CNC cholesteric structure after initial diffusive reflection off the EC layer. (g) Off-specular angle-resolved spectroscopy scan for a green EC-CNC film (reproduced from Fig. S20 for ease of comparison). (h) Same as (g), but with dotted lines overlaid, indicating the effects due to the mechanisms shown in (d-f).

Fig. S23.
Polarized optical microscope images and corresponding micro-spectra, recorded through left-and right-circularity polarized filters (denoted LCP and RCP), for the laboratoryscale blade-cast CNC film and the R2R-cast CNC-EC bilayer film. For the blade-cast CNC film, the reported spectrum is the average of at least 10 measurements while for the R2R-cast film, the reported spectrum is the average of at least 10 measurements on every 4 different locations along the R2R sample. The peak exceeding 1 is due to the back scattering from the EC layer adding to the reflection peak from the CNC layer.