RESEARCH ARTICLE www.advancedscience.com Highly-Scattering Cellulose-Based Films for Radiative Cooling Juliana Jaramillo-Fernandez,* Han Yang, Lukas Schertel, Guy L. Whitworth, Pedro D. Garcia,* Silvia Vignolini,* and Clivia M. Sotomayor-Torres the peak wavelength (9.7 μm) of black-body Passive radiative cooling (RC) enables the cooling of objects below ambient radiation at average ambient temperature temperature during daytime without consuming energy, promising to be a (≈300 K), enabling the maximization of ra- game changer in terms of energy savings and CO2 reduction. However, so far diative heat transfer. The same principle ap- most RC surfaces are obtained by energy-intensive nanofabrication processes plies to all the objects on the Earth’s surface. When exposed to the sky, objects radiate or make use of unsustainable materials. These limitations are overcome by the excess heat to outer space through the developing cellulose films with unprecedentedly low absorption of solar IRAW[1–3] and exchange heat with the atmo- irradiance and strong mid-infrared (mid-IR) emittance. In particular, a sphere at wavelengths where both the ob- cellulose-derivative (cellulose acetate) is exploited to produce porous jects and atmosphere are opaque. To some scattering films of two different thicknesses, L ≈ 30 μm (thin) and L ≈ 300 μm extent, the release of heat is damped due to the absorption of the outgoing long-wave ra- (thick), making them adaptable to above and below-ambient cooling diation by the atmospheric gases at wave- applications. The thin and thick films absorb only ≈5% of the solar irradiance, lengths outside the IRAW. The absorption which represents a net cooling power gain of at least 17 W m−2 , compared to by the atmosphere and the consequent ra- state-of-the-art cellulose-based radiative-cooling materials. Field tests show diation of this energy in the infrared (by that the films can reach up to ≈5 °C below ambient temperature, when solar Kirchhoff’s law of thermal radiation)[4] re- absorption and conductive/convective losses are minimized. Under dryer sults in further heating the object at the Earth’s surface. Thus, its net radiative heat conditions (water column = 1 mm), it is estimated that the films can reach flux to outer space is limited by the incident average minimum temperatures of ≈7–8 °C below the ambient. The work thermal radiation from the atmosphere and presents an alternative cellulose-based material for efficient radiative cooling the absorbed sunlight. Radiative cooling that is simple to fabricate, cost-efficient and avoids the use of polluting technologies exploit this phenomenon to materials. achieve passive sub-ambient cooling during daytime, and have emerged recently as a promising solution in renewable-energy re- search. 1. Introduction To achieve efficient radiative cooling, the optical properties The Earth keeps its thermal equilibrium by dissipating the excess of materials must be engineered to i) minimize the absorption heat from absorbed solar energy to the cold outer space (≈3 K) of sunlight and atmospheric thermal radiation and ii) enhance via thermal radiation. This heat exchange between the planet’s their emittance within the atmospheric transparency window. surface and outer space takes place because the atmospheric Many such materials have now been demonstrated using nano- gases barely absorb the outgoing long-wave radiation between and micro-structured technologies and surface engineering, 8 and 13 μm. This spectral band, known as the first infra-red promising a viable cooling technology for intensive global en- (IR) atmospheric transparency window (IRAW), coincides with ergy savings that could help mitigating climate change.[5–10] For J. Jaramillo-Fernandez, G. L. Whitworth[+] , P. D. Garcia, The ORCID identification number(s) for the author(s) of this article C. M. Sotomayor-Torres can be found under https://doi.org/10.1002/advs.202104758 Catalan Institute of Nanoscience and Nanotechnology (ICN2) [+] Presentaddress: The Institute of Photonic Sciences (ICFO), Mediter- CSIC and BIST ranean Technology Park, Avinguda Carl Friedrich Gauss, 3, Castelldefels, Campus UAB, Bellaterra, Barcelona 08193, Spain Barcelona 08860, Spain E-mail: juliana.jaramillo@icn2.cat david.garcia@icn2.cat [++] Present address: School of Chemical Engineering, University of Chi- H. Yang[++] , L. Schertel, S. Vignolini nese Academy of Sciences, Beijing 100049, China Department of Chemistry University of Cambridge © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH. Lensfield Road, Cambridge CB2 1EW, UK This is an open access article under the terms of the Creative Commons E-mail: sv319@cam.ac.uk Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. C. M. Sotomayor-Torres ICREA—Institució Catalana de Recerca i Estudis Avançats DOI: 10.1002/advs.202104758 Barcelona 08010, Spain Adv. Sci. 2022, 9, 2104758 2104758 (1 of 8) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com example, remarkable designs include microsphere-based pho- The thickness of the porous PDRC material is an important tonic random media,[11] thin-film multi-layer structures,[12–16] design parameter because it not only affects the optical proper- microsphere-periodic arrays,[17,18] metal–dielectric nanophotonic ties but also the thermal performance. Films with larger thick- structures,[19] double-layer nanoparticle-based coatings,[20,21] ness L can be advantageous to easily achieve strong emittance aerogels,[22,23] porous synthetic polymer-based coatings,[24,25] in the infrared and, at the same time, to provide thermal in- and hybrid dielectric–polymer materials.[26–30] However, several sulation to minimize parasitic conduction losses. Such charac- of these technologies come at the cost of demanding engineer- teristics are useful for applications that require cooling below ing and fabrication processes, making difficult their large scale the ambient-temperature (below-ambient cooling). On the other production and therefore their use in real applications. Further- hand, thin structures can be more beneficial for above-ambient more, one would ideally exploit sustainable materials to produce cooling, where heat conduction is favorable, for example, when such coatings and avoid the use of plastics or heavy metals, the substrate emits a lot of heat, such as in a generator (Figure 1a). to have a meaningful positive impact on the environment. In Here, only using one single approach we produce, via phase- this context, cellulose derivatives are ideal candidates as they separation, films of cellulose (acetate) with different thicknesses can be produced from cellulose in industrial scale by various and controlled scattering morphology. We show that the pro- chemical modifications.[31] Specifically, cellulose acetate (AC) duced thick films (≈300 μm, filling fraction =48%) strongly re- is an insoluble cellulose derivative considered as a nontoxic flect the sunlight and emit in the mid-infrared (mid-IR) while and biodegradable material.[32,33] Here, we report on producing can also provide thermal insulation to the underlying substrate. highly scattering porous cellulose films for efficient radiative On the other hand, the thin films (≈30 μm, filling fraction = 36%) cooling with a scalable production process that does not require allow partial transmittance and moderate emittance in the IRAW any particular or challenging nanostructuring or engineering. while still strongly reflecting in the solar spectrum. Figure 1b,c show scanning electron micrographs of the films from which the thickness and morphology were characterized (see Experimental 2. Results and Discussion Section for details). Thin films of cellulose (acetate) with a disordered network mor- phology were produced by a phase separation method. The achieved porosity enables strong light scattering down to UV 2.2. Optical Characterization wavelengths and across the visible solar spectrum (Figure 2b), while showing thermal emission in the mid-IR and the IR atmo- To maximize their cooling effect, we need to engineer the opti- spheric transparency window. We characterize the network mor- cal properties of radiative cooling materials over a large spectral phology, optical transmittance and reflectance spectra as well as range from the UV to the mid-IR. In Figure 2, the optical prop- their thermal properties to show their applicability and perfor- erties of a thin (black) and a thick (red) film are shown and their mance for radiative daytime cooling and discuss their potential emittance is plotted in Figure 2a. For comparison, we also plot for sustainable large scale coating applications. the normalized solar irradiance spectrum at air mass (AM) 1.5 (yellow) and the atmospheric transmittance (blue). The features observed in both thick and thin films are very similar over the en- 2.1. Tunable Thermal and Optical Properties of Radiative Cooling tire spectral range. A major difference lays in the lower emittance Films of the thin film around 12 μm which is directly associated to a higher transmittance. This could enable the use of our materials Passive radiative cooling materials must emit heat to the not only for protection and cooling in below-ambient applications ultracold-outer space through the IRAW (𝜆 ≈ 8–13μm). Only but also for above-ambient cooling, that is, heat generating sub- when the emitted power exceeds the power absorbed by the ma- strates. The optical properties of the thick and thin films in the terial over all other wavelengths, the material is able to cool below ultraviolet (UV)/visible (vis) spectra are plotted in Figure 2b,c, re- ambient temperatures. During daytime, achieving temperatures spectively. Both films show a low absorption down to 300 nm. It below ambient can be challenging due to heating of the struc- is noteworthy that both films have a remarkable high reflectance tures by incident sunlight. The performance of passive-daytime over the entire visible range and absorb only about 5% of the so- radiative cooling (PDRC) materials with high emittance in the lar irradiance, even in the case of the thin film (Table 1). The high IRAW can be therefore enhanced by designing their morphology reflectance in the visible is mainly caused by an optimal pore size to minimize absorption of solar irradiance (i.e., reflecting it op- distribution for Mie scattering with dominating sizes in the hun- timally) over a wide spectral range (𝜆 ≈ 0.3–2.5μm). Such broad- dreds of nanometers (Figure S1, Supporting Information). The band high reflectance is usually easily achieved in thick films of low absorptance combined with the high-broadband reflectance random-network structures through multiple light scattering.[34] over the whole solar spectrum makes these structures ideal for However, this is not trivial in thin films because the light trans- passive daytime radiative cooling. port mean free path (MFP), which is the length scale on which the orientation of the incoming light is randomized by scattering, is comparable or greater that the film thickness L. On a microscopic 2.3. Radiative Cooling Performance length scale, the scattering is limited by the ratio of the film thick- ness L and the MFP. Thus, achieving high broadband reflectance The radiative cooling of the cellulose-based films was evaluated in thin-films is difficult as the MFP in nonoptimized scattering by measuring the temperature reached by both films and compar- materials is usually several microns long.[35] ing that with the ambient temperature. The films were attached Adv. Sci. 2022, 9, 2104758 2104758 (2 of 8) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com Figure 1. Highly scattering cellulose-based radiative cooling films. a) Illustration of radiative below-ambient (right) and above-ambient (left) cooling applications for a thick and a thin cellulose-based film. The inset is a schematic representation of the disordered network scattering structure. Sunlight (yellow) is reflected by the network structure of the film while thermal radiation (red) is emitted. The thin film enables improved heat conduction, favorable for above-ambient cooling applications while the thick one provides thermal insulation from conduction and convection. Electron microscopy images of b) a thick (≈300 μm, filling fraction =48%) film and c) a thin (≈30 μm„ filling fraction =36%) film made of cellulose acetate (AC). Top: Cross-section image. Scale-bars are 200 and 20 μm. Bottom: Zoom in. Scale-bars are 5 μm Inset: Macroscopic image of the films. Scale-bars are 1 cm. to PT100-thermometers and the ambient temperature was mea- temperature difference relative to the ambient are plotted in sured using a Bresser PC weather station probe. A schematic il- Figure 3e,f, respectively. During daytime, both films maintained lustration of the experimental test set-up and a cross-section de- their temperatures below the ambient (Figure 3b). Average picting the heat exchange processes are shown in Figure 3a,b. below-ambient temperatures of 4.6 and 4.1 °C were measured Details are given in the Experimental Section. The thermal mea- for the thick and thin films, respectively, from 11:00 to 16:00, as surements were performed by exposing the experimental setup plotted in Figure 3e. During this period, the sky was clear, with a to a clear sky throughout a 24-h day–night cycle at the rooftop of mean solar irradiance of 780Wm−2 , a wind speed of 2.4 km h−1 a four-story building in Barcelona (Spain) in Autumn. (Figure 3d) and an average relative humidity of 40% (Figure 3c). In Figure 3c–f, we plot the 24-h cycle of the measured temper- The maximum temperature difference with respect to the am- ature of both the thin and the thick cellulose-based films along bient was 5.3 and 4.9 °C around 11:30, for the thick and the with the environmental conditions during the measurement. For thin film, respectively. These results indicate that the produced reference, Figure 3c shows the relative humidity (RH) (yellow cellulose films are capable of both preventing heating from curve, left axis) and the dew point (green curve, right axis), while diffuse solar irradiance and efficiently emitting heat as IR radi- Figure 3d shows the solar irradiance (light-blue curve, left axis) ation. Finally, the maximal absolute temperature reduction from and the wind speed (dark-blue curve, right axis). The time evolu- ambient was 5.5 and 5.2 °C measured just before noon (at 11:28) tion of the temperature from both films and their corresponding and at peak irradiance of 840 Wm−2 , for the thick and thin film, Adv. Sci. 2022, 9, 2104758 2104758 (3 of 8) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com a) b) 100 AM 1.5 Atmospheric 𝜖 AC thin film 𝜖 AC thick film (norm) transmittance (norm) (30 µm) (300 µm) 80 R T R,T,A (%) 1 1 A 60 Emittance, 𝜖 = A = 1 - R - T Atmospheric transmittance 40 0.8 0.8 20 0.6 0.6 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 c) 100 0.4 0.4 80 R R,T,A (%) T A 60 0.2 0.2 40 20 0 0 0.3 0.5 0.7 1 2 5 10 15 20 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 Wavelength (µm) Wavelength (µm) Figure 2. Optical properties of the cellulose-based radiative cooling films in the visible and mid-IR. a) Comparison of the spectral emittance, ϵ, of the thin (black) and thick (red) cellulose based films measured from 0.25 to 25 μm. The spectra superimposed depict the normalized mid-latitude solar irradiance with an air mass coefficient (AM) of 1.5[36] (orange) and atmospheric transmittance from MODTRAN[37] (blue). Reflectance, transmittance and absorptance of b) thick and c) thin cellulose-based films in the UV–visible, from 0.2 to 0.9 μm Table 1. Comparison of the absorptance in the UV–vis or mid-IR spectral ranges for recently reported cellulose-based materials for radiative cooling. The symbol “-” refers to data not available. Thickness Solar abs. Solar abs. ϵIRAW TIRAW ϵmid-IR [μm] power [W m−2 ] [%] [%] [%] [%] (0.3–2.5 μm) (0.3–2.5 μm) (8–13 μm) (8–13 μm) (2.5–25 μm) AC thin film 30.0 46.7 5.2 87.9 7.5 78.2 AC thick film 300.0 44.8 5.0 93.6 1.3 87.1 Hybrid structural material [ 41 ] 4000.0 64.2 7.2 91.1 – 85.9 Structural material [ 38 ] ⩾10 000.0 70.4 7.8 91.7 – 86.1 respectively. During the night, both films keep their tempera- convection and conduction coefficient of h = 11 W m−2 K−1 and tures below that of the ambient air, yet the temperature reduction Tamb = 300 K. It is important to consider that to estimate the compared to daytime is smaller, which suggest a decreased radia- solar absorption, we only take into account diffuse sunlight, as tive cooling power. Average below-ambient temperatures values the samples were characterized using a radiation shield. The of 2.9 and 2.5 °C were recorded for the thick and the thin film, atmospheric transmittance used to calculate the net cooling respectively (from 20:00 to 8:00, with average relative humidity power was extracted using MODTRAN,[37] for vertical water of 86% and wind speed of 0.5 km h−1 ). The reduced capability vapor column conditions typical of Barcelona (estimated water of both films to keep temperatures below that of the ambient column =10 mm), that is, high content of total gaseous water air during the night is attributed to the higher relative humidity. in a vertical column of atmosphere, compared to dry conditions The maximal temperature below ambient was 3.5 and 3.0 °C (more details in Experimental Section). The results are plotted measured at 21:30, before the dew point temperature started to in Figure 3h, as black (thin film) and red (think film) continuous drop from 12 °C. lines. For the sake of comparison, we also report the theoretical To understand the measured thermal performance of the net cooling power that can be achieved under dry atmospheric produced cellulose films, we calculate their daytime net cooling conditions (water column =1 mm), as dashed lines in Fig- power as a function of the temperature difference between the ure 3h. Finally, the average minimum temperature of the film radiative coolers and the ambient. For this purpose, their spectral achieved experimentally (in the steady-state regime) is shown emittances measured by Fourier-transform infrared (FT-IR) are in Figure 3h, as open black (thin film) and red pentagons (thick extrapolated for angles from 0° to 90°, assuming that they behave film), respectively. This equilibrium temperature is reached as Lambertian emitters. In Figure 3g, we plot the directional when the net cooling power becomes equal to zero, Tr|Pcool =0 . We spectral emittance of thin (top) and thick (bottom) films, over use the difference between the emitter’s temperature and the the IR transparency atmospheric window (from 8–14 μm). In ambient at this point to characterize the below-ambient cooling Figure 3h, we report the net cooling power estimated using the performance of the surfaces (more details in Experimental emittance values over the full mid-IR spectrum, considering a Section). Adv. Sci. 2022, 9, 2104758 2104758 (4 of 8) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com Figure 3. Testing the radiative cooling performance of cellulose-based films. a) Cross section of the cellulose-based films in the test set-up insulated from the ground for continuous temperature measurements. b) Schematics of the test setup. c–g) Continuous temperature measurement for the thin and the thick AC films, along with the environmental conditions. c) Relative humidity (RH) (yellow curve, left axis) and dew point (green curve, right axis). d) Solar irradiance (blue curve, left axis) and wind speed (purple curve, right axis). e) Measured cycle of temperature during 24 h in a clear-sky day for the thin and the thick AC films. The ambient temperature is plotted in blue for reference. f) Temperature difference between the ambient and the AC films. g) Spectral directional emittance of thin (top) and thick (bottom) cellulose-acetate nanostructured films, assuming that they behave as Lambertian emitters. h) Predicted net cooling power of the thin and thick cellulose-based films, for a conduction/convection coefficient of 11 Wm−2 K−1 , high atmospheric water vapor content (water column =10 mm), representative of a Mediterranean location, such as Barcelona (continuous lines) and dry conditions (water column =1 mm) (dashed lines). The black and red stars in Figure 3h pinpoint the average which is reduced due to the absorption of the heat emitted from minimum temperatures measured below ambient in the thick the film by the water vapor molecules in the lower atmosphere. and thin films, 4.1 and 4.6 °C, respectively, which match our We expect our films to reach average minimum temperatures of predictions. The location of the experiment, close to Barcelona, is 7 (thin film) and 8 °C (thick film) below the ambient under dryer characterized by a Mediterranean climate, which exhibits higher conditions (water column =1 mm), as indicated in Figure 3h water vapor content compared to other arid or high altitude loca- by including a correction with dry conditions (dashed curves). tions under which several daytime radiative cooling experiments For negligible convection and conduction (coefficient h = 0 have been reported in the literature.[5,22,24,38] The water vapor W m−2 K−1 ), our model predicts average minimum cooling content in the atmosphere affects the radiative cooling efficiency temperatures of 37 (thin film) and 39 °C (thick film) relative to Adv. Sci. 2022, 9, 2104758 2104758 (5 of 8) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com the ambient under dry atmospheric conditions. This requires to Our scalable-production process is based on air-drying, which completely eliminate the non-radiative losses, as in previously does not require any sophisticated nanostructuring neither the reported works, for example, refs. [39, 40]. incorporation of dielectric particles that act as optical scatterers. The cellulose-acetate films have a disordered network morphol- 2.4. Toward Scalable and Sustainable Radiative-Cooling Coatings ogy enhancing the reflectance across the whole solar spectrum (𝜆 ≈ 0.3–2.5μm), which results in minimal absorption of ≈ 5% Hybrid cellulose-based materials have been designed recently to of the solar irradiance (Table 1). At the same time, the produced optimize their optical and thermal performance for radiative cool- films exhibit strong mid-IR emittance, in particular within the ing in either the vis or mid-IR spectral regimes.[27,41–44] In these IR atmospheric transparency window. In addition, we demon- materials, cellulose is serving as a matrix and/or scattering ma- strated that by tuning the thickness, different optical and ther- terial to optimize the optical properties in the visible but the IR mal properties can be achieved: The thick films have stronger emission is obtained by adding thermal-emitting polar materi- emittance in the IR atmospheric transparency window (𝜖IRAW ≈ als such as Al2 O3 , TiO2 , and SiO2 nano or microspheres. Here, 93.6%) and are suitable for sub-ambient applications, where con- we engineer cellulose as an all-in-one radiative-cooling material duction and convection minimization is required to cool down by obtaining thick and thin films via phase separation with opti- an underlying substrate. The thin films exhibit slightly lower mized optical and thermal properties simultaneously. Our mate- 𝜖IRAW ≈ 87.9% and modest transmittance TIRAW ≈ 7.5% in the rial has been additionally tested for radiation durability using a IR atmospheric transparency window. Their reduced thickness Bandol wheel instrument in which the films were irradiated for (lower thermal resistance), makes them ideal for above-ambient 100 h, which is equivalent to 50–100 years of indoor illumination applications, where the underlying material produces heat and conditions, or about 6 months of direct outdoor sun exposure. convective/conductive heat transfer become beneficial for the net The color change (△E00 ) is a useful number for assessing subtle cooling (see the Supporting Information). All the thermal per- changes. For reference, humans in general would not visually no- formances of the cellulose-based radiative coolers were assessed tice a difference until △E00 reaches ≈2. The △E00 of the thin and during field tests. The produced thin and thick cellulose films thick AC films after Bandol wheel test is only 0.88 and 0.16, re- reached up to ≈5.0°C and ≈5.2°C below ambient temperature, spectively. when convection/conduction losses and direct solar irradiance A comparison of the physical properties relevant to radiative where minimized. Under dryer conditions (water column = 1 cooling between recently reported cellulose-based materials and mm), we estimate that our films can reach average minimum our films is shown in Table 1. Chen et al.[41] and Li et al.[38] were temperatures of 7 (thin film) and 8 °C (thick film) below the am- the first to report on cellulose-based structural materials for ra- bient. We, therefore, believe that our work provides a novel all- diative cooling with excellent mechanical properties for architec- in-one large-scale cost-efficient radiative cooling material fabri- tural applications. These materials were developed for building cated using solely sustainable cellulose-based compounds, that applications and therefore their dimensions are, at least, one or- are adaptable to different types of cooling applications. der of magnitude greater than the films here reported. Interest- ingly, the highly scattering AC films absorb a much lower fraction of solar irradiance, as cellulose acetate has better performance in the UV spectral range compared to pure cellulose. The heat- 4. Experimental Section ing power resulting from solar absorption is estimated to be ≈47 Materials: Cellulose acetate—Acetylcellulose—(AC) with rela- and 45 W m−2 for the thin and the thick films, respectively, com- tive molecular mass ≃ 29 000) and acetone were purchased from pared to 64 and 70 W m−2 for the thicker materials from refs. Sigma Aldrich. [41] and [38]. This reduction represents a gain in net cooling Preparation of Cellulose Acetate Thin Films: The films were prepared power of at least 17 W m−2 , despite the fact that our films are via air drying in a petri dish from the same cellulose acetate solutions con- taining acetone, water, and cellulose acetate powder. About 3% wt. cel- two and one order of magnitude thinner than the reported struc- lulose acetate was added to an acetone–water mixture solution with 5% tural materials. Very recently, Gamage et al.,[45] demonstrated wt. of water. The suspension was stirred vigorously to completely dissolve reflective and transparent radiative coolers based on cellulose the cellulose acetate powder. The final solution contained 3% cellulose ac- derivatives manufactured via electrospinning and casting. Our etate. Thin white films were formed by slowly drying the cellulose acetate– 300 μm-thick films exhibit comparable optical properties in the acetone–water mixture solution in a glass petri dish. visible and mid-IR with respect to their 275 μm-thick film. In Preparation of Cellulose Acetate Thick Films: Cellulose acetate powder contrast, our 30 μm-thin films achieve high reflectance across was added into pure acetone under vigorously stirring till its weight per- centage was 24% wt. The glue-like solution was transferred into a glass the whole solar spectrum and strong average IRAW emittance petri dish which was soaked in water to form a white film and dried at 60 (Table 1), despite their small thickness, while their thinnest re- °C in a vacuum oven. The thickness of the film was determined by the total ported film (60 μm) is partially transparent in both wavelength volume of the solution poured into a petri dish with a diameter of 4.7 cm. ranges. This highlights the efficient broadband light scattering Typically, 1 and 2 mL solutions were used for thin and thick films, respec- that we achieve in such thin films with the network scattering tively. nanostructure. Characterization of Cellulose-Based White Films by scanning Electron Mi- croscope: The cross-section of each film was measured by scanning electron microscope (SEM) with a Mira3 system (TESCAN) operated at 3. Conclusion 5 kV and a working distance of about 6 mm. To prepare specimens, the films were frozen in liquid nitrogen and then cracked. The samples In conclusion, we have obtained efficient radiative-cooling films were mounted on aluminum stubs using conductive carbon tape and using only an abundant and biodegradable cellulose-derivative. coated with a layer of platinum (10 nm in thickness) by a sputter coater Adv. Sci. 2022, 9, 2104758 2104758 (6 of 8) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com (Quorum Q150T ES). The thickness of each film was determined from the PGC2018-101743-B-I00 (SIP) and RTI2018-093921-A-C44 (SMOOTH), as SEM images of their cross-sections. well as by the CERCA Program/Generalitat de Catalunya. L.S. acknowl- Optical Characterization of Cellulose-Based White Films by UV–Vis Spec- edges the support of the Swiss National Science Foundation under project troscopy: The optical properties of the films in the UV/vis spectra where 40B1-0 198708. P.D.G gratefully acknowledges the support the Ramon y measured from 300 to 800 nm using a Cary 4000 spectrometer equipped Cajal fellowship (RYC-2015-18124). L.S is grateful to Johannes S. Haataja with an integrating sphere. for providing the 3D network graphic in Fig. 1. J.J.F. is grateful to A. El- Characterization of Cellulose-Based White Films by FT-IR: The emittance Sachat and P. Xiao for thermal conductivity measurements and gold de- of the films in the mid-IR was measured by FT-IR spectroscopy, using a position, respectively. gold integrating sphere to collect the near-normal reflected, transmitted, and scattered light over a solid angle of 2𝜋 sr, from wavelength of 0.9 to 25 μm. A direct approach to obtain the emittance was by measuring the Conflict of Interest absorptance A, the fraction of radiation absorbed by the films. It was as- sumed that the directional spectral emittance was equal to the directional The authors declare no conflict of interest. spectral absorptance, ϵ(𝜆, 𝜃, ϕ) = A(𝜆, 𝜃, ϕ) under thermodynamic equi- librium, according to the Kirchhoff’s law of thermal radiation.[46] The ab- sorptance was given directly by taking the difference of the measured re- flectance and transmittance of the samples in the infrared spectrum from Author Contributions unity (A = 1 − R − T). J.J.F, H.Y., and L.S. contributed equally to this work. S.V. and P.D.G con- Atmospheric Transmittance and Net Cooling Power Calculations: The ceived the idea and the project, and supervised the work. H.Y. fabricated atmospheric transmittance used in the net cooling power calculation the samples and performed optical measurements for material choice. was obtained using moderate resolution atmospheric transsmission L.S. analyzed the optical, the thickness, and morphology results. J.J.F. per- (MODTRAN).[37] MODTRAN is a computer code used to predict atmo- formed the UV–vis and FT-IR spectroscopy, thermal and radiative cooling spheric spectral transmittances and radiances over the ultraviolet through measurements, and analyzed of the optical and thermal results. G.L.W. long wavelength infrared spectral regime. The atmosphere model consid- and J.J.F calculated the net cooling power. J.J.F., P.D.G., L.S, and H.Y wrote ered a mid lattitude region during summer, which corresponded to a ver- the manuscript with all author’s contributions. tical water vapor column conditions typical of Barcelona, that is, high con- tent of total gaseous water in a vertical column of atmosphere, compared to dry conditions. The net cooling power calculations were performed fol- lowing the procedure described in ref. [17]. Data Availability Statement Continuous Passive Radiative Cooling Measurements of Cellulose-Based Dataset available here: https://doi.org/10.17863/CAM.78840. Films: During the experiment, the parasitic conduction and convection were minimized. To reduce these heat losses, the films were placed on a low-thermal conductivity membrane suspended inside a poly- methacrylate box used as a sample holder and covered with aluminum Keywords foil to minimize heat gain due to solar absorption. The top of the poly- cellulose, high mid-infrared emittance, low solar absorption, radiative methacrylate box was covered with a low-density polyethylene film that cooling, scattering was transparent in the relevant wavelengths, in order to reduce convec- tion through the top. Further details are given in Supporting Information. The poly-methacrylate boxes used as sample holders were placed on an Received: October 26, 2021 acrylic structure, of 70 cm of height, that thermally insulated the system Revised: November 18, 2021 from the ground. The structure was also covered with aluminum foil to Published online: January 17, 2022 reflect the visible light and minimize parasitic heating (Figure 3b). Pol- ished aluminum sheets were placed as solar radiation shields, that also helped to minimize the lateral convection and reflect outgoing IR radia- tion. The polished aluminum sheets were mainly used to avoid parasitic [1] E. A. Goldstein, A. P. Raman, S. Fan, Nat. Energy 2017, 2, 17143. heating due to direct solar irradiance, as radiation shields, used following [2] M. M. Hossain, M. Gu, Adv. Sci. 2016, 3, 1500360. the experimental procedure reported by Chen et al.[40] [3] X. Sun, Y. Sun, Z. Zhou, M. A. Alam, P. Bermel, Nanophotonics 2017, 6, 997. [4] C. G. Granqvist, Appl. Opt. 1981, 20, 2606. Supporting Information [5] Y. Zhai, Y. Ma, S. N. David, D. Zhao, R. Lou, G. Tan, R. Yang, X. Yin, Science 2017, 335, 1062. Supporting Information is available from the Wiley Online Library or from [6] X. G. Xu, B. G. Ghamsari, J. H. Jiang, L. Gilburd, G. O. Andreev, C. Zhi, the author. Y. Bando, D. Golberg, P. Berini, G. C. Walker, Nat. Commun. 2014, 5, 4782. [7] E. Tervo, Z. Zhang, , B. Cola, Phys. Rev. Mater. 2017, 1, 015201. Acknowledgements [8] J. long Kou, Z. Jurado, Z. Chen, S. Fan, A. J. Minnich, ACS Photonics 2017, 4, 626. This work was supported by the BBSRC David Phillips fellowship [9] A. R. Gentle, G. B. Smith, Nano Lett. 2010, 10, 373. (BB/K014617/1) and the Horizon 2020 Framework Programme Marie [10] T. Wang, Y. Wu, L. Shi, X. Hu, M. Chen, L. Wu, Nat. Commun. 2021, Curie Individual Fellowship (793643-MFCPF), ERC SeSaME ERC2014STG H2020 639088, the PoC 963872, Cellunan, and the Isaac Newton Trust 12, 365. (SNSF3) and the Philip Leverhulme Prize (PLP-2019-271). The European [11] S. Atiganyanun, J. B. Plumley, S. J. Han, K. Hsu, J. Cytrynbaum, T. L. Union Horizon 2020 Research and Innovation Program partly funded this Peng, S. M. Han, S. E. Han, ACS Photonics 2018, 5, 1181. research (Marie Sklodoswa-Curie Actions Grant No. 665919 that sup- [12] A. P. Raman, M. A. Anoma, L. Zhu, E. Rephaeli, S. Fan, Nature 2014, ported J.J.F and the project FLEXPOL Grant. No. 721062 that supported 515, 540. G.L.W). Moreover, the research received funds from the Spanish Min- [13] Y. Zhu, D. Wang, C. Fang, P. He, Y. H. Ye, Polymers 2019, 11, 1203. ister of Science, Innovation and Universities via the Severo Ochoa Pro- [14] A. Hervé, J. Drévillon, Y. Ezzahri, K. Joulain, J. Quant. Spectrosc. Ra- gram (Grant No. SEV-2017-0706) that supports ICN2, and the projects diat. Transfer 2018, 221, 155. Adv. Sci. 2022, 9, 2104758 2104758 (7 of 8) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com [15] S. Meng, L. Long, Z. Wu, N. Denisuk, Y. Yang, L. Wang, F. Cao, Y. Zhu, [31] H. Seddiqi, E. Oliaei, H. Honarkar, J. Jin, L. C. Geonzon, R. G. Ba- Solar Energy Mater. Solar Cells 2020, 208, 110393. cabac, J. Klein-Nulend, Cellulose 2021, 28, 1893. [16] E. Blandre, R. A. Yalçin, K. Joulain, J. Drévillon, Opt. Express 2020, 28, [32] J. Ganster, H.-P. Fink, in Bio-Based Plastics: Materials and Applications, 29703. Wiley, New York 2013, pp. 35–62. [17] G. L. Whitworth, J. Jaramillo-Fernandez, J. A. Pariente, P. D. Garcia, [33] D. Klemm, B. Heublein, H.-P. Fink, A. Bohn, Angew. Chem., Int. Ed. A. Blanco, C. Lopez, C. M. Sotomayor-Torres, Opt. Express 2021, 29, 2005, 44, 3358. 16857. [34] J. Syurik, G. Jacucci, O. D. Onelli, H. Hölscher, S. Vignolini, Adv. Funct. [18] J. Jaramillo-Fernandez, G. L. Whitworth, J. A. Pariente, A. Blanco, P. D. Mater. 2018, 28, 1706901. Garcia, C. Lopez, C. M. Sotomayor-Torres, Small 2019, 15, 1905290. [35] G. Jacucci, L. Schertel, Y. Zhang, H. Yang, S. Vignolini, Adv. Mater. [19] E. Rephaeli, A. Raman, S. Fan, Nano Lett. 2013, 13, 1457. 2021, 33, 2001215. [20] Z. Huang, X. Ruan, Int. J. Heat Mass Transfer 2017, 104, 890. [36] ASTM G173-03(2012), Standard Tables for Reference Solar Spectral [21] C. Yijun, M. Jyotirmoy, L. Wenxi, S.-W. Ajani, T. Cheng-Chia, H. Wen- Irradiances: Direct Normal and Hemispherical on 37°Tilted Surface, long, S. Sajan, Y. Nanfang, H. R. P. S. C. Anyuan, Y. Yuan, Sci. Adv. ASTM International, West Conshohocken, PA 2012. 2021, 6, eaaz5413. [37] A. Berk, P. Conforti, R. Kennett, T. Perkins, F. Hawes, J. Van Den Bosch, [22] A. Leroy, B. Bhatia, C. C. Kelsall, A. Castillejo-Cuberos, M. Di Capua in Workshop on Hyperspectral Image and Signal Processing, Evolution in H., L. Zhao, L. Zhang, A. M. Guzman, E. N. Wang, Sci. Adv. 2019, 5, Remote Sensing, Vol. 9088, SPIE, Bellingham, WA 2014, p. 90880H. eaat9480. [38] T. Li, Y. Zhai, S. He, W. Gan, Z. Wei, M. Heidarinejad, D. Dalgo, R. [23] M. Yang, W. Zou, J. Guo, Z. Qian, H. Luo, S. Yang, N. Zhao, L. Pattelli, Mi, X. Zhao, J. Song, J. Dai, C. Chen, A. Aili, A. Vellore, A. Martini, R. J. Xu, D. S. Wiersma, ACS Appl. Mater. Interfaces 2020, 12, 25286. Yang, J. Srebric, X. Yin, L. Hu, Science 2019, 364, 760. [24] J. Mandal, Y. Fu, A. C. Overvig, M. Jia, K. Sun, N. N. Shi, H. Zhou, X. [39] F. Trombe, Rev. Gen. Therm. 1967, 6, 1285. Xiao, N. Yu, Y. Yang, Science 2018, 362, 315. [40] Z. Chen, L. Zhu, A. Raman, S. Fan, Nat. Commun. 2016, 7, 13729. [25] D. Li, X. Liu, W. Li, Z. Lin, B. Zhu, Z. Li, J. Li, B. Li, S. Fan, J. Xie, J. [41] Y. Chen, B. Dang, J. Fu, C. Wang, C. Li, Q. Sun, H. Li, Nano Lett. 2021, Zhu, Nat. Nanotechnol. 2021, 16, 153. 21, 397. [26] H. Zhang, K. C. Ly, X. Liu, Z. Chen, M. Yan, Z. Wu, X. Wang, Y. Zheng, [42] S. Gamage, E. S. H. Kang, C. Kerlind, S. Sardar, J. Edberg, H. Kariis, T. H. Zhou, T. Fan, Proc. Natl. Acad. Sci. U. S. A. 2020, 117, 14657. Ederth, M. Berggren, M. P. Jonsson, J. Mater. Chem. C 2020, 8, 11687. [27] B. Xiang, R. Zhang, Y. Luo, S. Zhang, L. Xu, H. Min, S. Tang, X. Meng, [43] W. Wei, Y. Zhu, Q. Li, Z. Cheng, Y. Yao, Q. Zhao, P. Zhang, X. Liu, Z. Nano Energy 2021, 81, 105600. Chen, F. Xu, Y. Gao, Solar Energy Mater. Solar Cells 2020, 211, 110525. [28] X. Li, J. Peoples, P. Yao, X. Ruan, ACS Appl. Mater. Interfaces 2021, 13, [44] X. Chen, M. He, S. Feng, Z. Xu, H. Peng, S. Shi, C. Liu, Y. Zhou, Opt. 21733. Mater. 2021, 120, 111431. [29] X. Wang, X. Liu, Z. Li, H. Zhang, Z. Yang, H. Zhou, T. Fan, Adv. Funct. [45] S. Gamage, D. Banerjee, M. M. Alam, T. Hallberg, C. Åkerlind, A. Sul- Mater. 2020, 30, 1907562. tana, R. Shanker, M. Berggren, X. Crispin, H. Kariis, D. Zhao, M. P. [30] W. Huang, Y. Chen, Y. Luo, J. Mandal, W. Li, M. Chen, C.-C. Tsai, Z. Jonsson, Cellulose 2021, 28, 9383. Shan, N. Yu, Y. Yang, Adv. Funct. Mater. 2021, 31, 2010334. [46] G. Kirchhoff, Ann. Phys. 1860, 185, 275. Adv. Sci. 2022, 9, 2104758 2104758 (8 of 8) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH