Solar Energy Materials & Solar Cells 212 (2020) 110584

Contents lists available at ScienceDirect

Solar Energy Materials and Solar Cells

journal homepage: http://www.elsevier.com/locate/solmat

Multilayered SiO2/Si3N4 photonic emitter to achieve high-performance

all-day radiative cooling
Hongchen Ma a, b, 1, Kaiqiang Yao a, b, 1, Shuliang Dou c, Meng Xiao d, Mingguang Dai e,
Liang Wang f, Haipeng Zhao a, b, Jiupeng Zhao g, Yao Li c, Yaohui Zhan a, b, *
a
School of Optoelectronic Science and Engineering & Collaborative Innovation Center of Suzhou Nano Science and Technology, Soochow University, Suzhou, 215006,
Jiangsu, China
b
Key Lab of Advanced Optical Manufacturing Technologies of Jiangsu Province & Key Lab of Modern Optical Technologies of Education Ministry of China, Soochow
University, Suzhou, 215006, China
c
National Key Laboratory of Science and Technology on Advanced Composites, Harbin Institute of Technology, Harbin Institute of Technology, Harbin, 150001, China
d
Key Laboratory of Artificial Micro- and Nano-Structures of Ministry of Education and School of Physics and Technology, Wuhan University, Wuhan, 430072, China
e
Suzhou Rongrui Nano Composites Technology Co., LTD, Suzhou, 215000, China
f
State Key Laboratory of High Performance Ceramics and Superfine Microstructure, Shanghai Institute of Ceramics, Chinese Academy of Sciences, Shanghai, 200050, PR
China
g
MIIT Key Laboratory of Critical Materials Technology for New Energy Conversion and Storage, School of Chemistry and Chemical Engineering, Harbin Institute of
Technology, Harbin, 150001, China

A R T I C L E I N F O A B S T R A C T

Keywords: Radiative cooling has gained considerable attention recently, mainly due to its spontaneous thermodynamic
Photonic emitter process of dissipating surplus heat from the terrestrial object to deep space without any external energy. How­
Radiative cooling ever, challenges still remain to meet better the stringent requirement on spectral selectivity for improving the
Passive cooling
current unoptimized cooling performance, especially for the multilayer-type radiative coolers. In this work, we
Thermal emission
proposed a radiative cooler with seven layers of SiO2 and Si3N4 optimized by using the evolutionary algorithm,
Mid-infrared
which firstly facilitates a high reflection of solar irradiation utilizing the optical impedance mismatch, secondly
enables a broadband emissivity merely within the atmospheric window by complementary phonon resonances,
and thirdly suppresses the emissivity outside the atmospheric window (not only the solar spectral range). The
continuous rooftop measurement and the comprehensive calculation collectively demonstrated the excellent
capability of the radiative cooler, e.g., a measured temperature drop of 8 � C and a simulated cooling power of 87
W/m2 under the sunlight of ~900 W/m2. Besides, the cooling performance can be further enhanced if exposed to
a dry climate due to the high humidity during the test. Our work contributes an alternative candidate to
multilayered radiative coolers with cooling performance comparable to the polymer counterparts, and as well as
providing a rational design approach exploiting both the structural and material aspects for similar photothermal
devices.

1. Introduction
Radiative cooling has attracted increasing attention from both the
scientific and industrial communities [1–5], which opens up a practical
approach to exploiting the coldness of outer space as a renewable
thermodynamic reservoir. During the process of radiative cooling, the
redundant heat from an object on earth can be taken away into deep
space via thermal emission without the need for external energy sup­
plies. Therefore, radiative cooling can be widely used as an alternative to
electrical cooling in off-grid situations, such as seawater desalination
during long voyages, food preservation in desert areas, and dew
condensation in arid mountains [6–8]. However, the application of
radiative cooling had been limited to nighttime for a few decays since
the heating effect of sunlight during the daytime is much stronger than

* Corresponding author. School of Optoelectronic Science and Engineering & Collaborative Innovation Center of Suzhou Nano Science and Technology, Soochow
University, Suzhou, 215006, Jiangsu, China.
E-mail address: yhzhan@suda.edu.cn (Y. Zhan).
1
The authors contributed equally.

https://doi.org/10.1016/j.solmat.2020.110584
Received 1 March 2020; Received in revised form 25 April 2020; Accepted 27 April 2020
Available online 6 May 2020
0927-0248/© 2020 Elsevier B.V. All rights reserved.
H. Ma et al. Solar Energy Materials and Solar Cells 212 (2020) 110584

the achievable cooling effect [9]. Until recently, the first passive cooler coolers have significantly advanced just recently, the advantage of
for daytime was experimentally demonstrated by Fan’s group, which spectral selectivity has not been fully exerted yet.
was consisted of multiple dielectric layers and yielded a cooling power In this work, to give full play to the advantage of multilayer con­
of 40.1 W/m2 and a temperature drop of 4.9 � C below the ambient figurations and improve the current cooling performance, we proposed
temperature under direct sunlight [10]. After that, the daytime radiative and fabricated a multilayered radiative cooler based on the optimization
cooler has been extensively investigated, which can be roughly divided of the evolutionary algorithm [26,27], achieving a substantial
into three categories based on materials, including polymer coatings, improvement in the performance of temperature drop (e.g., 8 � C) and
organic-inorganic composite materials, and the previously mentioned cooling power (e.g., 87 W/m2) even under a hot and humid climate. The
dielectric multilayer films. The radiative cooler based on polymer boost of cooling performance is ascribed to the combination of silica and
nanostructures is represented by the hierarchically porous P silicon nitride layers, which facilitates a high reflection of solar irradi­
(VdF-HFP)HP coating consisting of ~2- to 10-μm micropores, and allows ation utilizing the optical impedance mismatch, enables a broadband
for a sub-ambient temperature drop of ~6 � C under solar irradiance of emissivity merely within the atmospheric window by complementary
890 W/m2 [11]. The simplicity of the phase-inversion technique makes phonon resonances [12,28], and suppresses the emissivity outside the
the porous polymer coatings be applied to various surfaces by conven­ atmospheric window (not only the solar spectral range). In particular,
tional approaches like painting and spraying. However, the pore size benefited from the emissivity selectivity in the mid-infrared range, the
distribution of porous polymers is not that controllable, and in this cooling effect of our proposed cooler surpasses or is on par with that of
respect, the organic-inorganic composite provides a more straightfor­ the typical polymeric coolers, in spite of the slightly inferior averaged
ward way. The representative of the latter approach is the glass-polymer emissivity. The investigation provides an alternative candidate for the
hybrid metamaterial scalable-manufactured by the roll-to-roll process. high-performance radiative coolers and demonstrates well the desig­
Such a metamaterial embeds resonant polar dielectric microspheres nability of the multilayer films for the practical photonic and thermal
randomly in a polymeric matrix, resulting in an infrared emissivity devices.
greater than 0.93 across the atmospheric window [12]. To explore the
fundamental limit of radiative cooling, many ingenious strategies have 2. Methods and materials
been devised from different aspects. Bhatia et al. decoupled the direct
solar irradiation from the diffuse component, enabling sub-ambient 2.1. Design and numerical methods
cooling at solar noon regardless of the emitter properties [13]. Zhou
et al. developed a beaming radiative cooling system that used a solar The conceptual schematic is given in Fig. 1. As shown in Fig. 1(a),
shelter component to suppress the solar input and a tapered waveguide our proposed radiative cooler is composed of seven alternating layers of
to control the thermal emission, which realized a temperature reduction SiO2 and Si3N4 covered on a silver back-reflector. The multilayer of
of ~2–9 � C even though the emitter was surrounded by tall buildings SiO2/Si3N4 functions as a selective emitter, which is transparent in the
[14]. Chen et al. fabricated a sophisticated vacuum system with a solar spectrum but opaque in the atmospheric window. The silver layer
rigorous heat shielding, by which the selective thermal emission of deposited on the silica substrate is employed to reflect the solar irradi­
radiative coolers was taken to an extreme with a temperature reduction ation and also the thermal emission towards the sky. The combination of
beyond 29 � C [15,16]. Li and Zhu et al. designed advanced photonic SiO2 and Si3N4 layers in the design is expected to manifest its optical
crystals for strongly dissipating heat through thermal emission but
simultaneously preserving the sunlight absorption of the solar absorber,
demonstrating well the designability of dielectric structures to regulate
light in multiple wavebands [17–19]. Besides, driven by the strong de­
mand for flexible and wearable products, some innovative textiles were
developed by using the nanoporous polyethylene or the blend of fibers
with different responses to infrared waves, promoting the practical
application of radiative cooling on personal thermal management
[20–22]. To ease the burden of fabrication, the radiative coolers with
several layers were also widely investigated theoretically or experi­
mentally [15,16,23–25]. Jeong et al. proposed an optimized TiO2-SiO2
multilayer, yielding a temperature reduction of 7.2 � C under a sub­
tropical climate [23]. Chae et al. proposed an inorganic-based multi­
layer emitter, showing an emissivity of 0.87 and an average reflection of
nearly 0.95 [24]. Kecebas et al. designed a broadband TiO2-SiO2 filter
for radiative cooling and demonstrated the suppression of solar ab­
sorption by replacing the metallic layer with dielectric reflectors [25].
As mentioned above, each type of radiative cooler exhibits its char­
acteristics and advantages to be suitable for specific applications.
Whereas the polymer-based radiative cooler is low-cost, flexible, and
convenient to be integrated onto non-planar surfaces, the dielectric
multilayer has excellent mechanical properties, high corrosion resis­
tance, and favorable tolerance to high temperatures. Moreover, the
dielectric multilayers can be rigorously designed with the aid of Maxwell
equation solvers for attaining the spectral selectivity inside and beyond Fig. 1. (a) Conceptual sketch of the proposed radiative cooler with random
the atmospheric window, which is strongly desired but scarcely satisfied layers designed based on the genetic algorithm for maximizing the radiated
power (Prad) and simultaneously suppressing the absorbed solar power (Psun).
by the polymer-based coolers whose emissivity is always non-selective
(b) Normalized complex refractive indexes (~ n) of SiO2 (green lines) and Si3N4
within the mid-infrared. Despite so, the cooling performances of the
(red lines), manifesting potential advantages of selected materials in terms of
typical dielectric multilayer coolers are unexpectedly much lower than both n (dashed lines) and k (solid lines). The normalized AM 1.5 solar spectrum
those of the polymeric counterparts [10–13], which is mainly due to and atmospheric transmittance at the mid-infrared are included for a clearer
reason that the infrared emissivity spectrum of the former is narrow and illustration. (For interpretation of the references to colour in this figure legend,
weak. Although, the apparent performance of multilayer radiative the reader is referred to the Web version of this article.)

2
H. Ma et al.
advantage in two aspects. Firstly, as shown in Fig. 1(b), the real part
refractive index (n) of Si3N4 is nearly 1.5 times that of SiO2 in the
spectral range where solar irradiation accumulates, which would facil­
itate the silver reflector achieving a high solar reflection in virtue of the
large impedance mismatch at Si3N4/SiO2 interfaces. Secondly, the
extinction coefficient (k) spectra of SiO2 and Si3N4 peak at distinct
wavelengths within the atmospheric window where the thermal emis­
sion can escape unimpededly to deep space, which indicates the possi­
bility of their complementary contribution to the broadband emissivity.
For extracting the maximum output from the proposed configura­
tion, the layer thickness is designed by using the differential evolu­
tionary algorithm (See methods). The objective optimization function is
defined as the net cooling power density (Pcool), which can be obtained
from Eq. (1) as follows [10].
Fdef Pcool ðTÞ ¼ Prad ðTÞ Patm ðTamb Þ Psun ðθsun Þ Pcondþconv ðT; Tamb Þ
Z π =2 Z ∞
where Prad ðTÞ ¼ 2π sinθcosθIBB ðT; λÞεðλ; θÞdλdθ indicates the
0 0
Z π=2 Z ∞
power density radiated out by the cooler; Patm ¼ 2π
sinθcosθIBB ðTamb ; λÞεðλ; θÞεatm ðλ; θÞdλdθ stands for the absorbed power
density of the cooler due to atmospheric thermal radiation; Pcondþconv ¼
hc ðTamb TÞ is the power loss due to thermal conduction and convection;
IBB is the spectral emissivity of the blackbody. In the calculation, the
values of hc and Tamb are adopted to be 6.9 W m-2K-1 and 308 K,
respectively. The formula of Patm, Psun, and Pcondþconv are listed briefly as
above, and more details can be acquired by consulting the literature
[10]. The key parameters for the optimally designed structure are listed
in Table 1.
The full-wave finite element method was employed by using Comsol
Multiphysics for calculating the field distribution. The routine procedure
of differential evolutionary algorithm was used in the optimization,
whose flowchart generally includes the operations of mutation, cross­
over, calculation of fitness, selection, and iteration. The population size
and the maximal generation were set as 60 and 100, respectively. The
selection rate and the mutation rate were employed as 0.9 and 0.1,
respectively. The calculation of the optical response in the optimization
(including reflectivity, transmissivity, and absorptivity) was performed
by using the standard feature matrix method [29]. The calculations were
performed across the waveband from 0.3 um to 20 um with an interval
of 5 nm. The complex refractive index of Ag, SiO2, and Si3N4 came from
the experiment data [28,30]. The unpolarized solar illumination was
considered by averaging the cases of TE and TM incidences as a ritual.
2.2. Fabrication and measurements
The fused quartz substrates of 1 mm thick have been used, which
were pre-processed by ultrasonic cleaning with organic solvents (ethyl
alcohol, acetone, and isopropanol) and de-mineralized water in a
cleanroom. After drying with nitrogen, the substrates were previously
coated with chromium of ~20 nm thick and then silver (purity of
99.999%) of ~200 nm thick. The layers of SiO2 and Si3N4 were depos­
ited by Plasma Enhanced Chemical Vapor Deposition using PlasmaPro
System 100, consisting of a vacuum chamber with base pressure 2 � 10-
Table 1
Key structural parameters of the radiative cooler.
Layer No. Material
VII SiO2
VI Si3N4
V SiO2
IV Si3N4
III SiO2
II Si3N4
I SiO2
Solar Energy Materials and Solar Cells 212 (2020) 110584
4
Torr and RF generator 13.56 MHz. The process was performed by using
an RF power of 300 W and a temperature of 300 � C. In the deposition
process, the gas flows of SiH4, N2O, and NH3 were 40 sccm, 1600 sccm,
and 80 sccm, respectively. The cross-section of samples was observed in
SEM (Phenom-World BV, Eindhoven, the Netherlands), and character­
ized using the appendant EDS. The reflectivity of the samples was
measured using a UV–vis–NIR spectrophotometer (Lambda-950, Perki­
nElmer) with a wavelength range of 250–2500 nm. The infrared emis­
sivity over the wavelength of 2.5–25 μm was measured by using a
VERTEX 70 (Bruker) FT-IR spectrometer with an A562 integrating
sphere.
The rooftop measurement was performed in Soochow University,
Suzhou city (N31� 180 3.8400 , E120� 340 52) on some sunny days from
October to November 2019. The test apparatus was assembled from
convenient components, which has been proved to be reliable by pre­
(1)
vious work [31]. The cross-sectional view of the setup is displayed in
Fig. 2(b), showing detailedly the specific measures complying with the
basic principle. The acrylic Petri dish is standing in the air with the
support of glass rods and covered by a low-density polyethylene film,
which together creates a well-sealed and suspended air pocket. The
0 0 polyethylene film serves as an infrared-transparent windscreen for
suppressing the heat coefficient of convection and conduction. The
radiative cooler placed on the foam board has mounted a thermocouple
at the back surface. The temperatures beneath the sample and in the
ambient are monitored simultaneously. The fabricated radiative coolers
were qualitatively characterized using Energy Dispersive Spectrometer
(EDS), which shows a relatively clear morphology based on the
two-dimension distribution of chemical elements. It is noted that the
interfaces between SiO2 and Si3N4 layers cannot be distinguished by
Scanning Electron Microscopy (SEM) due to the same averaged atomic
weights (~20/atom). As shown in Fig. 2(c), the bright stripe in EDS
images represents the region that is rich in the labeled element; on the
contrary, the dark stripe (or area) in the same image indicates that the
detected element content is negligible. Based on the spatial distribution
of elements Si, Ag, N and O, the material composition of each layer is
ascertained as listed on the rightmost side. It can be seen from Fig. 2(c)
that the fabricated layer thickness is in good agreement with the
designed parameters, which can also be attested by comparing the
simulated and measured optical spectra as displayed in Fig. 3(a) and (b).
3. Results and discussion
The simulated results in Fig. 3(a) and (b) show that the reflectivity
and emissivity of the designed radiative cooler have excellent perfor­
mance, which is coincident well with the measured results of the
experimental sample. Fig. 3(a) shows the simulated and measured
reflection spectra from 300 nm to 2500 nm, which is nearly perfect
(>98%) across all the solar spectrum except the near ultra-violet range.
The oscillation peaks in the reflectivity spectra are ascribed to the
interference from SiO2/Si3N4 interfaces, which enhances the reflectance
of the silver back-reflector even further. Benefited from the optical
impedance mismatch, the simulated solar reflection is optimized in a
manner akin to that achievable using classic one-dimensional photonic
crystals. Since the measured reflectivity is a relative quantity with
respect to the reference whiteboard, it is slightly greater than 1 in the
range of 500 nm–700 nm, indicating the reflectance of the cooler is
better than that of the standard sample. As the wavelength is smaller
than 450 nm, the measured reflectivity begins to be reduced dramati­
cally, which is mainly due to the interband absorption of silver layer
Thickness (nm)
near the ultraviolet. In addition, even though the measured reflectivity
257
at the ultraviolet is substantially lower than the calculated values due to
217
421 the unoptimized preparation process, the performance is still compa­
292 rable to the state of the current [32]. Fig. 3(b) displays the emissivity
418 spectra covering the wavelength range from 2.5 μm to 20 μm, where an
257 excellent spectral selectivity can be observed. The emissivity plateau is
56
located rightly within the atmospheric windows (8–14 μm), and as also
3
H. Ma et al. Solar Energy Materials and Solar Cells 212 (2020) 110584

Fig. 2. (a) Photo of the apparatus and passive cooler on the test rooftop in Soochow University, Suzhou. (b) Schematic of the setup for testing all-day performance.
(c) Qualitative characterization of the multiple layers of the fabricated radiative cooler using Energy disperse spectroscopy. The corresponding chemical element is
labeled for the bright stripe, and the final material composition corresponding to each layer is listed on the rightmost side.

Fig. 3. (a) Measured (black line) and calculated (red

line) reflectivity covering the full solar spectrum. (b)
Measured (black line) and calculated (red line)
emissivity in the mid-infrared range. (c) Absorption
distribution in each layer of the cooler at the typical
wavelengths of 8.4 μm ( ) and 10.6 μm ( ) as shown
in (b). The odd-numbered (I, III, V, VII) and even-
numbered layers (II, IV, VI) correspond to SiO2 and
Si3N4, respectively. The position origin (z ¼ 0) starts
from the bottom of the silver layer and the maximum
of z represents the top surface of the cooler. (d)
Spectral contribution of SiO2 (green region) and Si3N4
(pink region) layers to the total emissivity over all the
concerned mid-infrared band. (For interpretation of
the references to colour in this figure legend, the
reader is referred to the Web version of this article.)

4
H. Ma et al.
expected the emissivity drops gradually outside the transparent window
(<8 μm & >14 μm). Such a unique spectral selectivity seems hard to be
realized by the polymer-based radiative coolers, the emissivity of which
preserves high throughout the concerned mid-infrared range (e.g.,
2.5–20 μm) non-selectively due to the infrared-absorbent feature of the
functional groups in polymers.
Interestingly, as shown in Fig. 3(b), two prominent peaks occur at
8.4 μm ( ) and 10.6 μm ( ), which is precisely predicted by the simu­
lation. To examine the field distribution at these typical wavelengths,
the spatial absorption distribution along the thickness-direction is
plotted in Fig. 3(c). As shown in Fig. 3(c), the absorptions of the entire
multilayered system at the wavelengths of 8.4 μm ( ) and 10.6 μm ( )
are mainly contributed from the SiO2 layers (III, V, and VII) and Si3N4
layers (II, IV and VI), respectively. The layers (from bottom to top) are
denoted as the numbers from I to VII, where the odd and even numbers
correspond to the SiO2 and Si3N4 layers, respectively. For overviewing
the contribution of SiO2 and Si3N4 layers across the full waveband, the
emissivity of two media is extracted separately as a function of wave­
length. As shown in Fig. 3(d), the emissivity of SiO2 layers absolutely
dominates in the wavelength range of 8–10 μm; as the wavelength in­
creases, the emissivity offered by Si3N4 layers sharply increases in the
range of 10–14 μm; as the wavelength is greater than 14 μm, although
the emissivity of SiO2 gradually increases, that of Si3N4 drops more
significantly, which eventually results in an drop of the total emissivity.
The delicate spectral engineering is accomplished not only by utilizing
the complementary spectral contribution of different materials for
improving the emissivity within the atmospheric window, but also by
handling well the tradeoff between these contributions for balancing the
emissivities within and beyond the atmospheric window.
To fully evaluate the detailed characteristics of the radiative cooler,
the azimuthal spectral response and the power balance are studied.
Fig. 4(a) shows the emissivity spectra as a function of incident angle,
covering all the concerned wavelengths from 0.3 to 25 μm. The angle
5
Solar Energy Materials and Solar Cells 212 (2020) 110584
ranges from 0 to 89� , not including 90� , due to numerical singularities.
In the solar spectral range, the emissivity is nearly zero in most cover­
ages and independent on the incident angles. In the mid-infrared range,
the spectral positions of the emissivity peaks remain almost unchanged,
but the intensity is gradually weakened at lager angles. To see the angle-
dependent properties quantitatively, the averaged reflectivity RðθÞ is
obtained by integrating the spectral reflectivity Rðλ; θÞ over the wave­
length range of 0.3–2.5 μm and the averaged emissivity εðθÞ is acquired
by integrating εðλ; θÞ over the atmospheric window of 8–14 μm, which
are numerically defined as Eqs. (2) and (3), respectively [33].
R 2:5μm
​ Rðλ; θÞIAM1:5 ðλÞdλ
(2)
0:3μm
RðθÞ ¼ R 2:5μm
​ IAM1:5 ðλÞdλ
0:3μm
R 14μm
​ εðλ; θÞIBB ðλÞdλ
εðθÞ ¼
8μm
R 20μm (3)
0:3μm
​ IBB ðλÞdλ
where IAM1.5 is the standard solar spectrum [34]; the reflectivity and
emissivity are both obtained by averaging the cases for TE and TM po­
larizations in the calculation, i.e., R ¼ ðRTE þRTM Þ=2 and ε ¼ ðεTE þ
εTM Þ=2. As shown in Fig. 4(b), the averaged reflectivity of the cooler is
~97% over all the incident angles, exhibiting an excellent omnidirec­
tional property. As for the averaged emissivity, it is approximately 0.75
at the normal angle and increases gradually at small angles (e.g.,
0� –30� ), reaching a maximal value of 0.8 around the angle 30� . Then,
the emissivity drops relatively quickly as the angle further increases.
Nevertheless, the averaged emissivity still remains as high as 0.6 at a
large angle of 75� . Fig. 4(c) shows the calculated powers of Prad, Patm,
and Pcool as a function of the cooler’s temperature. The solar irradiance
used in the calculation was obtained by integrating the standard AM1.5
spectrum, the total power density of which is ~900 W/m2 numerically.
During the calculation procedure of cooling powers, both the reflectance
Fig. 4. (a) Contour map of the emissivity versus the
wavelength and incident angle. (b) Averaged reflec­
tance (top panel) and averaged emissivity (bottom
panel) as a function of azimuthal angle. (c) Power
densities of Prad (black square), Patm (blue circle), and
Pcool (red triangle) as a function of temperature for the
cooler. (d) Simulated cooling power that is required
to reach a given temperature by the cooler below the
ambient air temperature (i.e., Tatm ‒ T), and the
corresponding dissipated power due to conduction
and convection as a function of (Tatm ‒ T). (For
interpretation of the references to colour in this figure
legend, the reader is referred to the Web version of
this article.)
H. Ma et al.
and emissivity were obtained from the electromagnetic simulation. The
angular dependence of emissivity had been accounted for in the rigorous
calculations as routine. As the temperature of radiative cooler drops
from 308 K (i.e., the ambient temperature) to 298 K, Prad decreases ~ 31
W/m2 (from 231 W/m2 to 200 W/m2), and Patm ~16 W/m2 (from 126
W/m2 to 110 W/m2). Meanwhile, Pcool has a dramatic cutdown of ~100
W/m2, showing a strong inverse dependence on the temperature of the
cooler. In the case of 298 K, the value of Pcool is negative (i.e., 10
W/m2), which indicates the cooling process can not work spontaneously
under such a condition and needs external injected energy to achieve the
equilibrium state. To display the cooling performance clearer, the
cooling power of the cooler at a temperature below the ambient (i.e.,
Tatm ‒ T) is given in Fig. 4(d), with the accompanying Pcondþconv
included for comparison. As shown in Fig. 4(d), the cooling power is
~87 W/m2 as the cooler is at the ambient temperature and the maximal
temperature drop is ~8.3 � C, corresponding to the thermal equilibrium
of Pcool ¼ 0. Different from the cooling power, the dissipation power due
to conduction and convection is positively proportional to the temper­
ature difference and acting as a major constraint on the net cooling
power. For instance, as Tatm ‒ T ¼ 8 � C, Pcondþconv ¼ hc (Tatm ‒ T) ¼ 56
W/m2 with the coefficient of hc ¼ 6.9, which is almost seven times that
of the corresponding Pcool.
The typical results of rooftop measurements are presented in Fig. 5,
including the primary meteorological conditions. Fig. 5(a) shows the
solar intensity across the test days from Oct 29, 2019 to Oct 30, 2019,
which reaches ~900 W/m2 at solar noon. According to solar irradiance,
we define nighttime as the period from 18:00 to 6:00 while daytime the
remained, which are labeled throughout subfigures for an intuitive
illustration. Fig. 5(b) gives the corresponding humidity data, which
shows the humidity can reach as high as 90% at a few hours before the
sunrise. The continuously measured temperatures in the ambient and
the cooler are provided in Fig. 5(c). As shown in Fig. 5(c), Tamb and
Tcooler rise and go down synchronously, along with the solar rising and
falling. The temperature drop is given in Fig. 5(d) by subtracting Tcooler
from Tamb. Once the cooler is exposed to the sky, the temperature of the
cooler drops immediately, and then the cooling amplitude varies with
the combined influence of the solar irradiation and humidity. As shown
in Fig. 5(d), ΔT dropped firstly at noon on the first day and arrived at the
lowest at 12:30 due to the increase of solar intensity. Then, ΔT increased
gradually and reached the highest at 15:00 with a peak cooling
Fig. 5. (a) Solar intensity and (b) humidity as a function of sample time across a whole day. The nighttime is defined by 6:00 p.m. to 6:00 a.m. (shadow area) and the
rest is daytime. (c) Ambient temperature and the temperature of the radiative cooler during the time beyond 24 h. (d) Temperature drop of the cooler relative to the
ambient circumstance.
6
Solar Energy Materials and Solar Cells 212 (2020) 110584
amplitude of 8 � C, which is ascribed to the nearly lowest humidity
(~20%) and then the slightly decreased solar irradiance (~700 W/m2).
As nighttime was approaching, the solar intensity continued reducing;
however, the humidity began to rise and eventually led to a drop in ΔT.
The minimum ΔT of ~2 � C occurred at dawn (~6:00) when humidity
accumulated to the highest (i.e., 90%), due to the atmospheric window
was mostly closed by the infrared-absorbent water vapor. The mea­
surements suggest that, despite the significant adverse effects of high
humidity, the proposed radiative cooler still maintains an all-day cool­
ing performance ranging from ~2 � C ‒ 8 � C. The performance of the
cooler is expected to be further enhanced if exposed to the arid desert
with extremely low humidity (e.g., 5%).
4. Conclusions
In summary, we have designed a seven-layered radiative cooler
based on the differential evolutionary algorithm and demonstrated an
all-day cooling capability under a humid meteorological condition in the
experiment. By exploiting the intrinsic optical properties of silica and
silicon nitride, a challenging selective spectral feature was achieved
with a high reflection across the solar spectrum and a broadband
emissivity merely within the atmospheric window simultaneously. The
comprehensive calculations accurately predicted the optical response
and energy exchange, which assisted in uncovering the underlying
physical mechanism quantitatively. Benefited from the spectral engi­
neering and the layer optimization, the radiative cooler possesses a
cooling range up to 8 � C and a predicted cooling power of 87 W/m2 with
respect to the ambient temperature. The maximal temperature drop
occurs when the solar irradiation stays at a high level (~700 W/m2) and
the humidity (~20%) tends to be the lowest, which reveals the crucial
role of humidity in restraining the cooling performance and suggests
that the radiative cooler is especially suitable for the hot and dry district
such as the vast desert areas. The investigation demonstrates the pos­
sibility of multilayered-type coolers possessing a superior or an equiv­
alent cooling performance as compared to the polymeric and even the
advanced multilayer counterparts. It is noted that, although the tem­
perature reductions of 7.2 � C and 8.3 � C have been recently reported for
the multilayer coolers [23,24], the measurements were performed in a
hot cavity formed by the solar-reflective walls, which inevitably in­
creases the temperature drop due to a higher reference ambient
H. Ma et al.
temperature than the transparent apparatus as in this work. Further­
more, the spectral emissivity is either nonselective in the mid-infrared or
unoptimized purposefully. The work also provides an alternative radi­
ative cooler or a practical selective-emitter with widely available ma­
terial and manufacture process for integrating to other applications such
as the solar cells, thermal photovoltaics, and mid-infrared sensors.
Declaration of competing interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
the work reported in this paper.
CRediT authorship contribution statement
Hongchen Ma: Investigation, Writing - original draft. Kaiqiang
Yao: Methodology, Software, Visualization. Shuliang Dou: Investiga­
tion. Meng Xiao: Writing - review & editing. Mingguang Dai: Investi­
gation. Liang Wang: Investigation. Haipeng Zhao: Validation. Jiupeng
Zhao: Visualization. Yao Li: Resources, Writing - review & editing.
Yaohui Zhan: Conceptualization, Supervision, Writing - review &
editing.
Acknowledgments
The work is supported by Natural Science Foundation of Jiangsu
Province [grant number BK20181167], National Natural Science
Foundation of China [grant number 61405132], Opening Project of
State Key Laboratory of High Performance Ceramics and Superfine
Microstructure [grant number SKL201912SIC], and Priority Academic
Program Development (PAPD) of Jiangsu Higher Education Institutions.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.solmat.2020.110584.
References
[1] P.-C. Hsu, A.Y. Song, P.B. Catrysse, C. Liu, Y. Peng, J. Xie, S. Fan, Y. Cui, Radiative
human body cooling by nanoporous polyethylene textile, Science 353 (2016)
1019–1023, https://doi.org/10.1126/science.aaf5471.
[2] W. Li, Y. Shi, Z. Chen, S. Fan, Photonic thermal management of coloured objects,
Nat. Commun. 9 (2018) 1–8, https://doi.org/10.1038/s41467-018-06535-0.
[3] L. Cai, A.Y. Song, W. Li, P.C. Hsu, D. Lin, P.B. Catrysse, Y. Liu, Y. Peng, J. Chen,
H. Wang, J. Xu, A. Yang, S. Fan, Y. Cui, Spectrally selective nanocomposite textile
for outdoor personal cooling, Adv. Mater. 30 (2018) 1–7, https://doi.org/10.1002/
adma.201802152.
[4] A.P. Raman, W. Li, S. Fan, Generating light from darkness, Joule 3 (2019) 1–8,
https://doi.org/10.1016/j.joule.2019.08.009.
[5] L. Cai, Y. Peng, J. Xu, C. Zhou, C. Zhou, P. Wu, D. Lin, S. Fan, Y. Cui, Temperature
regulation in colored infrared-transparent polyethylene textiles, Joule 3 (2019)
1478–1486, https://doi.org/10.1016/j.joule.2019.03.015.
[6] M.M. Hossain, M. Gu, Radiative cooling: principles, progress, and potentials, Adv.
Sci. 3 (2016) 1–10, https://doi.org/10.1002/advs.201500360.
[7] B. Zhao, M. Hu, X. Ao, N. Chen, G. Pei, Radiative cooling: a review of
fundamentals, materials, applications, and prospects, Appl. Energy 236 (2019)
489–513, https://doi.org/10.1016/j.apenergy.2018.12.018.
[8] M. Zeyghami, D.Y. Goswami, E. Stefanakos, A review of clear sky radiative cooling
developments and applications in renewable power systems and passive building
cooling, Sol. Energy Mater. Sol. Cells 178 (2018) 115–128, https://doi.org/
10.1016/j.solmat.2018.01.015.
[9] X. Lu, P. Xu, H. Wang, T. Yang, J. Hou, Cooling potential and applications
prospects of passive radiative cooling in buildings: the current state-of-the-art,
Renew. Sustain. Energy Rev. 65 (2016) 1079–1097, https://doi.org/10.1016/j.
rser.2016.07.058.
[10] A.P. Raman, M.A. Anoma, L. Zhu, E. Rephaeli, S. Fan, Passive radiative cooling
below ambient air temperature under direct sunlight, Nature 515 (2014) 540–544,
https://doi.org/10.1038/nature13883.
Solar Energy Materials and Solar Cells 212 (2020) 110584
[11] J. Mandal, Y. Fu, A.C. Overvig, M. Jia, K. Sun, N.N. Shi, H. Zhou, X. Xiao, N. Yu,
Y. Yang, Hierarchically porous polymer coatings for highly efficient passive
daytime radiative cooling, Science 362 (2018) 315–319, https://doi.org/10.1126/
science.aat9513.
[12] Y. Zhai, Y. Ma, S.N. David, D. Zhao, R. Lou, G. Tan, R. Yang, X. Yin, Scalable-
manufactured randomized glass-polymer hybrid metamaterial for daytime
radiative cooling, Science 355 (2017) 1062–1066, https://doi.org/10.1126/
science.aai7899.
[13] B. Bhatia, A. Leroy, Y. Shen, L. Zhao, M. Gianello, D. Li, T. Gu, J. Hu, M. Solja�ci�c, E.
N. Wang, Passive directional sub-ambient daytime radiative cooling, Nat.
Commun. 9 (2018) 1–8, https://doi.org/10.1038/s41467-018-07293-9.
[14] L. Zhou, H. Song, J. Liang, M. Singer, M. Zhou, E. Stegenburgs, N. Zhang, C. Xu,
T. Ng, Z. Yu, B. Ooi, Q. Gan, A polydimethylsiloxane-coated metal structure for all-
day radiative cooling, Nat. Sustain. 2 (2019) 718–724, https://doi.org/10.1038/
s41893-019-0348-5.
[15] Z. Chen, L. Zhu, A. Raman, S. Fan, Radiative cooling to deep sub-freezing
temperatures through a 24-h day-night cycle, Nat. Commun. 7 (2016) 1–5, https://
doi.org/10.1038/ncomms13729.
[16] Z. Chen, L. Zhu, W. Li, S. Fan, Simultaneously and synergistically harvest energy
from the sun and outer space, Joule 3 (2019) 101–110, https://doi.org/10.1016/j.
joule.2018.10.009.
[17] L. Zhu, A.P. Raman, S. Fan, Radiative cooling of solar absorbers using a visibly
transparent photonic crystal thermal blackbody, Proc. Natl. Acad. Sci. U.S.A. 112
(2015) 12282–12287, https://doi.org/10.1073/pnas.1509453112.
[18] W. Li, Y. Shi, K. Chen, L. Zhu, S. Fan, A comprehensive photonic approach for solar
cell cooling, ACS Photonics 4 (2017) 774–782, https://doi.org/10.1021/
acsphotonics.7b00089.
[19] E. Rephaeli, A. Raman, S. Fan, Ultrabroadband photonic structures to achieve high-
performance daytime radiative cooling, Nano Lett. 13 (2013) 1457–1461, https://
doi.org/10.1021/nl4004283.
[20] P. Hsu, C. Liu, A.Y. Song, Z. Zhang, Y. Peng, J. Xie, K. Liu, C. Wu, P.B. Catrysse,
L. Cai, S. Zhai, A. Majumdar, S. Fan, Y. Cui, A dual-mode textile for human body
radiative heating and cooling, Sci. Adv. 3 (2017), e1700895, https://doi.org/
10.1126/sciadv.1700895.
[21] P.B. Catrysse, A.Y. Song, S. Fan, Photonic structure textile design for localized
thermal cooling based on a fiber blending scheme, ACS Photonics 3 (2016)
2420–2426, https://doi.org/10.1021/acsphotonics.6b00644.
[22] L. Cai, A.Y. Song, P. Wu, P.C. Hsu, Y. Peng, J. Chen, C. Liu, P.B. Catrysse, Y. Liu,
A. Yang, C. Zhou, C. Zhou, S. Fan, Y. Cui, Warming up human body by nanoporous
metallized polyethylene textile, Nat. Commun. 8 (2017) 496, https://doi.org/
10.1038/s41467-017-00614-4.
[23] S.Y. Jeong, C.Y. Tso, J. Ha, Y.M. Wong, C.Y.H. Chao, B. Huang, H. Qiu, Field
investigation of a photonic multi-layered TiO2 passive radiative cooler in sub-
tropical climate, Renew. Energy 146 (2020) 44–55, https://doi.org/10.1016/j.
renene.2019.06.119.
[24] D. Chae, M. Kim, P.-H. Jung, S. Son, J. Seo, Y. Liu, B.J. Lee, H. Lee, Spectrally
selective inorganic-based multilayer emitter for daytime radiative cooling, ACS
Appl. Mater. Interfaces 12 (2020) 8073–8081, https://doi.org/10.1021/
acsami.9b16742.
[25] M.A. Kecebas, M.P. Menguc, A. Kosar, K. Sendur, Passive radiative cooling design
with broadband optical thin-film filters, J. Quant. Spectrosc. Radiat. Transf. 198
(2017) 179–186, https://doi.org/10.1016/j.jqsrt.2017.03.046.
[26] Y. Shi, W. Li, A. Raman, S. Fan, Optimization of multilayer optical films with a
memetic algorithm and mixed integer programming, ACS Photonics 5 (2018)
684–691, https://doi.org/10.1021/acsphotonics.7b01136.
[27] X. Wang, W. Cai, X. Yin, A global optimized operation strategy for energy savings
in liquid desiccant air conditioning using self-adaptive differential evolutionary
algorithm, Appl. Energy 187 (2017) 410–423, https://doi.org/10.1016/j.
apenergy.2016.11.073.
[28] J. Kischkat, S. Peters, B. Gruska, M. Semtsiv, M. Chashnikova, M. Klinkmüller,
O. Fedosenko, S. MacHulik, A. Aleksandrova, G. Monastyrskyi, Y. Flores, W.
T. Masselink, Mid-infrared optical properties of thin films of aluminum oxide,
titanium dioxide, silicon dioxide, aluminum nitride, and silicon nitride, Appl.
Optic. 51 (2012) 6789–6798, https://doi.org/10.1364/AO.51.006789.
[29] H. Shi, Y. Zhang, H. Wang, W. Liu, Matrix formalism for radiating polarization
sheets in multilayer structures of arbitrary composition, Chin. Optic Lett. 15
(2017), 081901, https://doi.org/10.3788/col201715.081901.
[30] E.D. Palik, Handbook of Optical Constants of Solids, Elsevier, Orlando, 1985,
https://doi.org/10.1016/C2009-0-20920-2.
[31] J.L. Kou, Z. Jurado, Z. Chen, S. Fan, A.J. Minnich, Daytime radiative cooling using
near-black infrared emitters, ACS Photonics 4 (2017) 626–630, https://doi.org/
10.1021/acsphotonics.6b00991.
[32] K. Xu, L. Hao, M. Du, J. Mi, Q. Yu, S. Li, J. Wang, S. Li, Thermal emittance of Ag
films deposited by magnetron sputtering, Vacuum 174 (2020), https://doi.org/
10.1016/j.vacuum.2020.109200, 109200.
[33] K. Yao, H. Ma, M. Huang, H. Zhao, J. Zhao, Y. Li, S. Dou, Y. Zhan, Near-perfect
selective photonic crystal emitter with nanoscale layers for daytime radiative
cooling, ACS Appl. Nano Mater. 2 (2019) 5512–5519, https://doi.org/10.1021/
acsanm.9b01097.
[34] Astm, Standard tables for reference solar spectral irradiances: direct normal and
hemispherical on 37� tilted surface, ASTM G173-03, www.astm.org, 2012.