Applied Thermal Engineering 225 (2023) 120162 Contents lists available at ScienceDirect Applied Thermal Engineering journal homepage: www.elsevier.com/locate/apthermeng Research Paper Enhanced behaviour of a passive thermoelectric generator with phase change heat exchangers and radiative cooling David Astrain a, *, Juliana Jaramillo-Fernandez b, Miguel Araiz a, Achille Francone b, Leyre Catalán a, Alejandra Jacobo-Martín b, Patricia Alegría a, Clivia M. Sotomayor-Torres b, c a Engineering Department, Smart Cities Institute, Public University of Navarre, Campus Arrosadía s/n, 31006 Pamplona, Spain b Catalan Institute of Nanoscience and Nanotechnology (ICN2), CSIC and BIST, Campus Universitat Autonoma de Barcelona (UAB), Bellaterra, Barcelona 08193, Spain c ICREA—Institució Catalana de Recerca i Estudis Avançats, Barcelona 08010, Spain A R T I C L E I N F O A B S T R A C T Keywords: Heat exchangers are essential to optimize the efficiency of Thermoelectric Generators (TEGs), and heat pipes Radiative cooling without fans have proven to be an advantageous design as it maintains the characteristic robustness of ther­ Heat-pipe moelectricity, low maintenance and lack of moving parts. However, the efficiency of these heat exchangers Thermoelectric generator decreases under natural convection conditions, reducing their heat transfer capacity and thus thermoelectric Thermal resistance power production. This work reports on a novel heat exchanger that combines for the first time, phase change and radiative cooling in a thermoelectric generator to improve its efficiency and increase the production of electrical energy, specially under natural convection. For this, two thermoelectric generators with heat-pipes on their cold sides have been tested: one with the radiative coating and the other without it. Their thermal re­ sistances have been determined and the electric power output was compared under different working conditions, namely, natural convection and forced convection indoors and outdoors. The experimental tests show a clear reduction of the heat exchanger thermal resistance thanks to the radiative coating and consequently, an increase of electric production 8.3 % with outdoor wind velocities of 1 m/s, and up to 54.8 % under free convection conditions. The application of the radiative surface treatment is shown to result in a more stable electrical energy production, suppressing the drastic decrease in the generated electric power that occurs in thermoelectric gen­ erators when they work under free convection. water and electricity. Mohammadnia et al. [6] investigated the combi­ 1. Introduction nation of thermoelectricity in solar concentration systems with a Stirling engine, and Karri et al. [7] achieved savings of up to 3 % in the fuel Thermoelectric generators (TEG) are devices capable of converting consumption of a SUV-type vehicle, thanks to the TEG placed in the thermal energy into electricity without moving parts, which leads to exhaust pipes. important advantages such as robustness and reliability, lower price, However, the main drawback of TEGs is their low efficiency in en­ absence of noise and vibrations, and scalability [1,2]. These character­ ergy conversion, obtaining values close to 3 % for a temperature dif­ istics, together with the need to increase power generation with ference of 120 ◦ C, while for the same temperature difference, systems renewable sources, has led to an important research activity in these based on Organic Rankine Cycles achieve an efficiency of 10 % [8]. This fields, developing, as a consequence, interesting and innovative appli­ is why there are important efforts in the improvement of the thermo­ cations for TEGs using different thermal sources for the production of electric materials [9], but also in the optimization of the heat ex­ electricity [3]. Thus, for example, Montecuco et al. [4] developed a changers, at both the hot and cold sides of the device. As demonstrated thermoelectric stove capable of generating 600 W of domestic hot water, by Astrain et al. [10], a 10 % decrease in the thermal resistance of the producing at the same time 27 W of electrical power, by means of a heat exchangers causes an 8 % increase in the generation of electric combustion TEG equipped with a water pumping system. Zheng et al. energy from a TEG. As a consequence, water heat exchange systems are [5] developed a prototype of a domestic thermoelectric co-generator, used in many applications of TEGs, such as in Niu et al. [11] which harnessing the waste heat present in the exhaust pipe to produce hot analyzed the power generation of a TEG using water heat exchangers * Corresponding author. E-mail addresses: david.astrain@unavarra.es (D. Astrain), juliana.jaramillo@icn2.cat (J. Jaramillo-Fernandez). https://doi.org/10.1016/j.applthermaleng.2023.120162 Received 8 September 2022; Received in revised form 16 January 2023; Accepted 31 January 2023 Available online 3 February 2023 1359-4311/© 2023 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by- nc-nd/4.0/). D. Astrain et al. Applied Thermal Engineering 225 (2023) 120162 Nomenclature Greek symbols α Seebeck’s coefficient, V/K Et Electromotive force, V ∊ Efficiency bx Uncertainty of parameter x I Electric current, A Abbreviations K Thermal conductance, W/K IPA Isopropyl alcohol IR Infrared Q̇C Heat released from the cold side, W HP Heat-pipe Q̇H Heat absorbed in the hot side, W NIL Nanoimprint Lithography R0 Electric resistance of the TE modules, Ω PDMS Poly dimethyl siloxane RC Cold side thermal resistance, K/W TEG Thermoelectric generator RH Hot side thermal resistance, K/W TEG_RC Thermoelectric generator with radiative treatment on the RL Load resistance, Ω heat-pipes TC Cold side temperature of the modules, ◦ C TEG_w/o_RC Thermoelectric generator without any treatment on TH Hot side temperature of the modules, ◦ C the heat-pipes Tsource Heat source temperature, ◦ C TEM Thermoelectric module T∞ Environment temperature, ◦ C UV Ultraviolet V Output voltage, V VIS Visible Ẇe Electric power output, W where the pressure losses, different water flows and temperature sources employing biphasic thermosyphon heat exchangers, capable of working were taken into account. The prototype developed by Kim et al. [12], under natural convection conditions. which achieved a maximum power output of 119 W with 40 thermo­ This type of TEGs without moving parts are widely used in remote electric modules, cooled with a channel heat exchanger through which sensor applications, such as the one reported in the work of Huang et al. water circulated at an inlet temperature of 20 ◦ C. However, these sys­ [19], who developed a thermoelectric micro generator to power sur­ tems require a pump that moves the liquid through the device, in veillance sensors in forests. Catalan et al. [20,21] developed a TEG using addition to dissipating the heat absorbed by the fluid to the environment HPs without fans, capable of feeding a remote volcanic monitoring through another heat exchanger, usually with a fan. Aranguren et al. station installed on Teide’s volcano. In this work, the importance of the [13] took these factors into account, studying the influence of the mass lack of moving parts was highlighted since it provides the necessary flow rate of the water in the channel heat exchanger, the air flow in the robustness for this application. However, the authors also reported how water-to-ambient heat sink, as well as the pressure losses in the system. the electric generation of this application decreased by 25 % when the Their most important conclusion was that in many cases, the electrical wind speed dropped to values below 1 m/s. Similarly, Alegría et al. [22] consumption of the pump and the fan necessary to dissipate the heat, did built a TEG prototype to harvest geothermal energy, with fan-less heat- not compensate for the increase in the power output of the TEG by the pipes. Their results showed the loss of efficiency when there is no wind, reduction in thermal resistances. due to the increase in the thermal resistance of the HPs, going from 35 W A significant advance to improve the efficiency of the heat ex­ of power output with 1.5 m/s of wind, to 20 W of electric generation changers in thermoelectric applications is the use of heat pipes (HP), without wind, equivalent to a reduction of 33 %. where, thanks to the phase change, the entire convective area in contact In order to solve these cited problems of TEǴ s and to increase their with the air is used, as it is able to distribute the heat flow, from the efficiency, maintaining the ability of working without moving parts, this thermoelectric modules throughout the area. Yan Liu et al. [14] devel­ work presents, for the first time, the combination of heat pipes with oped a TEG prototype with HP as heatsinks, working with air and water. radiative cooling, applied to the TEGs. Thanks to the low thermal resistances of the heat exchangers, they ob­ The goal is to improve the thermal resistance values of heat pipes tained up to 6 W of electric power per thermoelectric assembly. Ara­ heat exchangers that work without fans, especially when the wind ve­ nguren et al. [15] showed that a TEG that uses HP with fans, to harvest locity is very low or zero. In these circumstances, the heat transfer from waste heat from a chimney, produces 43 % more electric power than a the outer surface of the HP to the air happens under natural convection, TEG using finned heat sinks under the same working conditions. with very low convection heat transfer coefficient values (between 3 and All these heat exchange systems mentioned obtain thermal resistance 10 W/m2K), so that thermal radiation mechanisms, which are always values very suitable for thermoelectric applications. However, they have present, become more important [23]. The focus of this study is on the disadvantage of using devices with moving parts, such as fans or enhancing the overall heat transfer from the heat pipes to the air, under pumps, which, in addition to complicating the installation, consume natural convection, by improving the HP’s radiative properties at ther­ electricity. In all these applications, the intrinsic advantages of ther­ mal wavelengths (from 2.5 up to 25 µm). This is achieved by coating the moelectricity of being a technology without moving parts is lost. In copper envelope surface of the HPs with a thin broadband thermal order to study applications of TEGs free of moving parts, and recover this emitter that consist of a micro-structured resist. It enables strong emit­ advantage of thermoelectricity, heat exchangers without fans or pumps tance in the mid-infrared (mid-IR), and high transparency in the visible to drive the fluids have been developed. Thus, for example, Date et al. (VIS) and near-infrared (near-IR). This preserves the high reflectance of [16] compared the use of finned heatsinks of different geometries with cooper in the visible spectral range, while maximizing the cooling passive HPs as the heat exchanger of the cold side of a TEG. The electric capability of naturally cooled HPs, by changing the radiative properties power output produced with the HP-TEG system was 1.8 W, which is 4 of the copper envelope surface in the mid-IR, from highly reflective to times higher than the electric generation of the TEG that used the finned highly emitting (See Supplementary Information). In this manner, the device. Orr et al. [17] employed a fan-less HP in their experimental work overall heat transfer of the naturally cooled HPs is enhanced via radia­ with TEGs, obtaining 6 W per thermoelectric module and an efficiency tion, which has a critical role under these conditions. Furthermore, the in the conversion of 2.3 %. Araiz et al. [18] developed a TEG that gen­ advantages of TEGs that operate without moving parts are maintained, erates 259 W/m2 of net power from waste heat in an industry, as discussed above. This work is based on the emerging technique of 2 D. Astrain et al. Applied Thermal Engineering 225 (2023) 120162 radiative cooling, which enables the cooling of objects that are facing the sky even during daytime, without consuming energy. This passive process generally consists on maximising the radiated power while minimizing the heat gain due to solar and atmospheric absorption. This can be achieved by engineering the optical properties of materials, i) minimizing the absorption of sunlight and atmospheric thermal radia­ tion and (ii) enhancing their emittance in the mid-IR, as the energy radiated is directly proportional to the surface emissivity. This broad­ band mid-IR emittance is particularly relevant for above-ambient cool­ ing applications, e.g. when the substrate emits a large amount of heat, such as in a thermoelectric generator, because it maximizes the power evacuated as thermal radiation. Furthermore, for outdoor applications, a net heat exchange between the surface and the outer space takes place due to the surface high emittance from 8 to 13 µm, the spectral range at which the atmosphere is transparent. As the blackbody thermal radia­ tion peaks at ~ 10 µm at 300 K, the highly emitting surface is able to dissipate heat through the infrared transparency window (8–13 μm). Many radiative cooling materials have by now been demonstrated, using nanotechnology and surface engineering. Remarkable approaches Fig. 1. Thermoelectric Generator scheme and temperatures at different loca­ include thin film multi-layer structures [24–28], microsphere-based tions of the TEG with integrated heat pipes. photonic media [29–31] metal-dielectric nanophotonic structures [32], double-layer nanoparticle-based coatings [33,34], aerogels law of heat conduction. With the same effects, but making the energy [35,36], hybrid dielectric-polymer materials [37–40], porous synthetic balance on the cold side, the rate of heat flow that the thermoelectric polymer based coatings [35,38,41,42] and sustainable cellulose-based module releases to the environment through the heat exchanger is given films [43–49]. by Eq. (2). Likewise, and due to the Seebeck effect, an electromotive However, the application of radiative cooling materials for thermo­ force is produced in the thermoelectric module whose expression is electric generators has been barely studied, despite its potential impact given by Eq. (3). The values for the thermal conductance, electric to enhance electric generation. For example, Zhan et al. [50] applied this resistance, Seebeck coefficient and temperature operation range vary radiative enhancement technique to thermoelectricity. In their work, with the TEM used. In this work Marlow TG12-8L [52] thermoelectric this surface treatment was applied to a finned heatsink in natural con­ modules have been employed. vection, increasing the electric generation from 0.018 W to 0.028 W, using a heat source at a temperature of 250 ◦ C. Their work showed that Q̇H = K(TH − TC ) − I 2 R0 + αITH (1) the improvement achieved in finned dissipaters thanks to a radiative 2 treatment, results in an increased thermoelectric generation of more I 2 R0 than 50 % under these conditions. Since finned heat sinks were used as Q̇C = K(TH − TC ) + + αITC (2) 2 heat exchangers, the generation obtained is rather low, around 28 mW. However, it showed that advances in this technology could lead to more Et = α(TH − TC ) (3) efficient heat exchangers. In summary, there are many published papers that study phase If the thermoelectric generator is connected to a load resistance, RL , change heat exchanger in thermoelectric generators and a few very there will be an output voltage in the load resistance given by Eq. (4). recent works about radiative cooling and thermoelectric generators. V = Et − R 0 I = R L I (4) However, radiative cooling and phase change heat exchangers com­ bined in a thermoelectric generator have not been reported so far. In this So, the electric power produced by the generator will be determined work a passive thermoelectric generator, which includes phase change by Eq. (5), where m = RL /R0 ; and its efficiency is given by Eq. (6). heat exchangers and radiative cooling has been developed and tested for m the first time. Thanks to this novel heat exchanger, the passive ther­ Ẇ e = α(TH − TC )2 (5) R0 (m + 1)2 moelectric generator produces 58 % more electrical energy under nat­ ural convection. This approach could be used to improve applications of Ẇ e TH − TC m thermoelectric generation that do not incorporate moving parts, thus ∊= = (6) Q̇H TH 1+m− (TH − TC ) + (1 + m)2 αR20TkH preserving the intrinsic advantages of thermoelectricity such as 2TH robustness, reliability and lack of noise and vibrations. The temperatures of the hot, TH , and cold, TC , sides of the thermo­ electric modules will depend on the temperatures of the hot source and 2. Methodology cold sink (environment), as well as on the heat transfer rate and the thermal resistances of the heat exchangers used, whose expressions are 2.1. Thermoelectric generators and heat exchangers given by Eq. (7) and Eq. (8). Tsource − TH Thermoelectric generators are devices capable of converting heat RH = (7) Q̇H into electric energy. Their principle of operation is based on the heat transfer from the hot source to the hot side of the thermoelectric mod­ TC − T∞ ules (TEM), where part of that heat flow will be transformed into electric RC = (8) Q̇C energy, thanks to the Seebeck effect. The rest of the heat will be released to the environment by another heat exchanger placed at the cold side of As can be deduced from these equations, a decrease in the thermal the modules, as shown in Fig. 1. resistance of the heat exchangers will mean a greater temperature dif­ According to the basic calculation method [51], the rate of heat flow ference between the faces of the thermoelectric modules, increasing the that reaches the hot face of the thermoelectric modules is given by Eq. generation of electric power and the efficiency of the system, see Eq. (5) (1) and contemplates the Peltier effect, the Joule effect and the Fourier’s and (6). This effect was already demonstrated by Astrain et al. [10], 3 D. Astrain et al. Applied Thermal Engineering 225 (2023) 120162 showing that the efficiency of the TEG was improved by 8 % when the temperature probe, as shown in Fig. 1. Likewise, the heat flux provided thermal resistance of the heat exchangers was decreased by 10 %. by the electric resistances acting as the heat source has been measured This experimental study is only focused on the cold side of the by an ammeter and voltmeter. Each thermoelectric generator is con­ thermoelectric modules and using heat pipes as heat exchangers on the nected to a variable load resistance and has an ammeter and voltmeter to cold side since, as shown in [15,16], these heat exchangers improve the be able to register the power output coming from them. All these vari­ electric generation of the TEGs, with respect to those that use finned ables have been recorded during the experimentation with an Ahlborn heat sinks. Almemo 5690-1M09 data acquisition system. Table 1 includes the res­ Two identical TEG prototypes have been built, so that each of them olution and accuracy for every measurement probe that has been consists of a heating plate responsible for supplying a heat flux to the employed in the tests. thermoelectric generator. Over this plate a thermoelectric module, From these experimental results, the two variables under study can model Marlow TG12-8L [52], with 127 BiTe pairs, has been placed. On be estimated: the thermal resistance of the HP heat exchangers on the the cold side of the module, a 40 mm × 40 mm × 12 mm aluminum plate cold side of the thermoelectric modules can be found using Eq. (8), has been located, where 3 heat pipes, whose length is 400 mm and have where Q̇C , which is the rate of heat flow released to the environment, can 8 mm diameter, have been inserted to transfer the heat to the environ­ be estimated following the energy balance in Eq. (9). ment, as shown in Fig. 2. All these components have been assembled with two steel metal plates and screws, making a tightening of 1 Nm. Q̇H = Ẇ e + Q̇C (9) Finally, with the aim that all the heat generated by the heating plate is Here, Q̇H , can be obtained from Eq. (10), assuming that the device is transferred to the thermoelectric modules, all this assembly has been properly insulated, and the thermal losses can be neglected; and Ẇe is properly insulated, except for the HP, which are responsible for the electric power output from the thermoelectric module, see Eq. (11). releasing the heat to the environment. For this, a first layer of 60 mm A temperature sensor was placed outside the insulation box covering the rockwool material Pro Rox WM 970 ES [53] has been placed, with a heating plate. Since the temperature measurements were close to the thermal conductivity of 0.04 W/mK within the range of the working environmental temperature sensor, the thermal losses throughout this temperatures. Over this material a second 20-mm layer of ARMAFLEX material can be neglected. has been placed [54]. Q̇H = Vre Ire (10) 2.2. Experimental methodology Ẇ e = Vp Ip (11) One of the objectives of this work is to determine the influence of the The performance of the thermoelectric generators was studied under application of the radiative treatment, described in Section 2.3, on the several conditions. They were connected to different load resistances to performance of heat-pipe based heat exchangers for the cold side of obtain the characteristic curve of the generator. The goal was to prove thermoelectric generators under different working conditions (forced the influence of the radiative surface treatment (described in Section and natural convection, indoor and outdoor conditions). To do this, it is necessary to determine the thermal resistance of the HP, as well as the electric power output obtained with the TEGs. Consequently, two pro­ Table 1 totypes have been built: TEG_RC, a thermoelectric generator using a HP Resolution and accuracy of the measurement sensors. that includes a radiative cooling treatment, TEG_w/o_RC, another Sensor Resolution Accuracy thermoelectric generator using a heat-pipe too, but without any radia­ Temperature (◦ C) 0.1 ± 0.5 tive treatment. Both prototypes include NiCr-Ni thermocouples tem­ Ammeter (A) 0.01 ± 0.02 perature probes on the hot and cold sides of the thermoelectric modules. Voltmeter (V) 0.1 ± 0.4 The environment temperature is also measured using an additional Fig. 2. Thermoelectric Generator prototype with a heat-pipe on the cold side. 4 D. Astrain et al. Applied Thermal Engineering 225 (2023) 120162 2.3) not only on the performance of the heat exchangers, but also on the electric generation of the prototypes under different ambient conditions. The results obtained from a thermoelectric generator whose HPs have been coated with the radiative treatment (TEG_RC) were compared with the generator without the radiative cooling treatment (TEG_w/o_RC). In the first test, both prototypes with no treatment were tested to know the initial conditions of each of them (TEST_0), with the aim of determining the uncertainty that could exist in the operation of both identical generators, due to possible differences in the manufacture and assembly processes. Once it was verified that they offer equivalent re­ sults, as shown in Fig. 4, one of them (TEG_RC) was treated with the radiative coating, as described in Section 2.3. Again, both generators were tested simultaneously inside a climatic chamber to maintain con­ stant the environment temperature under conditions of forced convec­ Fig. 4. TEST-0 for both prototypes TEG_RC and TEG_w/o_RC without any tion with air speed of 1 m/s (TEST_1) and natural convection (TEST_2). surface radiative treatment, at an air velocity of 1 m/s. Finally, tests (TEST_3) were carried out outdoors under real environ­ ment conditions. Table 2 summarises the tests performed in this work. Table 2 Conditions of the tests carried out in this work. 2.3. Micro-patterning of the HPs integrated in the TEG system TEST-0 TEST-1 TEST-2 TEST-3 The passive radiative cooling coating was applied on the external Indoor Indoor Indoor Outdoor Initial conditions Forced Natural Free - forced surface of the copper HPs integrated in the TEG system by means of UV convection convection assisted Nanoimprint Lithography (UV-NIL). The radiative cooling TEG_RC: No treat. TEG_RC: With TEG_RC: With TEG_RC: With coating consists of a commercial resist with inherent high emissivity in TEG_w/o_RC: No treat. treat. treat. the mid-IR. Its surface is texturized by using this nano- and micro- treat. TEG_w/o_RC: No TEG_w/o_RC: No TEG_w/o_RC: No treat. treat. treat. replication technology to further enhance its directional spectral emis­ sivity, even at high emission angles. In UV-NIL, the pattern is replicated from a mold into a low molecular working mold was fabricated using soft-lithography [57] by casting and weight photo-curable resist layer which is previously applied on the curing the polymer precursors in direct contact with a 10 cm diameter desired surface. Then, the polymer is crosslinked under UV light irra­ silicon master mold, containing the topography to be replicated (fabri­ diation to produce a hard structured coating. Due to the HṔ s curved cated with photolithography and electroplating techniques) [58]. surface, the mold used in this case for the UV-NIL process was a thin soft Fig. 3a shows the soft-lithography procedure. A PDMS (Sylgard 184, flexible mold fabricated in poly dimethyl siloxane (PDMS) [55,56] Dow Corning, Sigma Aldrich) precursor in dilution with the curing agent which allows conformal contact with the whole curved surface. This Fig. 3. Micro-patterning of the HPs integrated in a TEG system: a) PDMS working mold fabrication by soft-lithography technique. The PDMS precursor mixture is poured over the master silicon mold and after polymer crosslinking, the flexible PDMS is peeled off. b) After cleaning, the HP surface is coated by using a syringe, c) the PDMS working mold is wrapped around the coated HP surface, and irradiated by UV light to perform the UV-NIL process. d) Once the resist is crosslinked and the mold peeled off, the coated surface exhibits the required hardened topography. 5 D. Astrain et al. Applied Thermal Engineering 225 (2023) 120162 (10:1) was mixed and degassed under vacuum. First, the PDMS solution prototypes have been connected to an electrical load resistance and a was casted on top of the silicon master mold and then cured for 4 h at heat flow is applied to them, recording the temperatures, the voltage and 80⁰C. After polymer crosslinking, the PDMS mold was peeled-off current in the thermoelectric modules as a function of time, until steady manually and a flexible working mold, with about 2 mm thickness, state is reached. Fig. 5 difference between the cold face of the thermo­ was obtained. electric modules and the environment, corresponding to two tests, one The surface treatment began by cleaning the copper HPs with IPA under forced convection with a wind velocity of 1 m/s (TEST-1) and the and subsequently drying them properly. The commercial resist (3 ml) other under natural convection conditions (TEST-2). In both tests, a heat was applied on the copper HP surface (placed in a planar position) using flux of 40 W has been applied to the two prototypes, with the thermo­ a syringe (Fig. 3b). The PDMS flexible working mold was put into con­ electric modules connected to a load resistance of 2.2 Ω. tact with the coated copper surface, wrapped around the pipe and finally It can be observed that the increase in the temperatures produced tightened with metallic clips. In this way, a constant pressure was between the cold face of the module and the environment is greater in guaranteed along the whole contact area and the uncured resist was able the case of the TEG_w/o_RC, which does not incorporate the radiative to flow through the mold cavities. Afterwards, the wrapped HP was surface treatment, as a result of a worse heat transfer by radiation. This placed inside a homemade UV-NIL tool, whose chamber was previously effect, which becomes more significant in the case of natural convection, fully covered with aluminium paper (to promote homogeneous UV determines that the thermal resistance of the HP in this TEG_w/o_RC is illumination and facilitate the curing of the resist all around the curved higher (see Fig. 6) and the electric power output of the generator is surface of the HP during the exposure step). Then, the surface was lower, due to the decrease in the temperature difference between the irradiated by the UV lamp (OmniCure S2000-XLA), as shown in Fig. 3c module’s faces (see Fig. 7). These tests have been repeated varying the (exposure time ~ 15 s). Finally, the mold was manually peeled off load resistance as well as the heat flux from the hot source, to find the leaving the cured resist with the negative shape of the mold pattern on thermal resistance values of the heat exchangers and the electric gen­ the surface of the copper bars, required to improve the material radiative eration, as described in the methodology. properties (Fig. 3d). Fig. 6 shows the results of the thermal resistance of the HPs, comparing the tests in forced and free convection for each of the two 3. Results and discussion prototypes: TEG_RC and TEG_w/o_RC. If the results of the tests in forced convection (TEST-1) are compared, in the same conditions as in the 3.1. Uncertainty analysis and TEST-0 initial TEST-0, it is observed that the TEG_w/o_RC maintains the value of its resistance, 1.72 K/W. However, the TEG_RC, that includes the radi­ In the tests, the performance of each heat-pipe heat exchanger and ative coating, offers a lower thermal resistance of 1.6 K/W, showing that thermoelectric generator has been analysed varying the heat flux the surface treatment improves a 7.5 % the thermal resistance value of absorbed from the hot side. When studying the thermal resistance of the the heat-pipes. heat exchangers, the tests were replicated three times to estimate the The greatest change occurs when testing the devices under natural overall uncertainty according to Coleman [59], which is composed of convection conditions (TEST-2), where the radiative treatment was ex­ the random standard uncertainty (using the replicas done) and the pected to have a higher effect, due to the decrease in the convection systematic uncertainty, using for every parameter the energy balances in coefficient and, therefore, to the greater influence of radiation. In the Eq. (9), (10), (11) and the following Eq. (12). case of the TEG_w/o_RC, the thermal resistance increases an 86.4 % ⎛ ⎞2 ⎛ ⎞2 compared with the case of wind velocity of 1 m/s, reaching an average value of 3.2 K/W. However, in the case of the TEG_RC, the thermal ⎜1⎟ ⎜− 1⎟ (TC − T∞ )2 2 b2RC = ⎝ ⎠ b2TC + ⎝ ⎠ b2T∞ + 4 bQ̇C (12) resistance of the heat exchanger corresponds to 54.7 %, reaching on Q̇C Q̇C Q̇C average 2.48 K/W. This represents a 29 % improvement in the thermal resistance of the HPs that include the radiative treatment. In this case, Fig. 4 shows the results of the preliminary tests (TEST-0), that pro­ the improvement in the radiation heat transfer makes up for the vides a comparison of the starting point of the heat exchangers and the decrease in the free convection coefficient when there is no air velocity. initial operation of the two thermoelectric generators (TEG_RC and The next analysis carried out is about the electric generation of the TEG_w/o_RC). In this way, it is possible to evaluate if the original pro­ thermoelectric prototypes. As described in Eq. (5), the electric power totypes exhibit differences in the thermal behavior, before applying any output depends on the load resistance to which the TEG is connected, so treatment. The objective is to eliminate the uncertainty in the results of the performance of each prototype has been tested varying the electric the subsequent tests due to possible variations in the manufacture and load resistance to obtain the characteristic curve of these devices. Fig. 7 assembly process of the prototypes. These tests have been carried out shows the results obtained from the experimentation under forced and inside a climatic chamber with an air velocity of 1 m/s and an ambient temperature of 15 ◦ C, for different heat fluxes supplied. As it can be seen from the results, both prototypes offer very similar thermal resistance, with maximum deviations of 0.05 K/W, which is below the calculated overall uncertainty, which for these tests is 0.07 K/W. In this way, it is ensured that, once the radiative surface treatment has been applied to one of the prototypes, the differences in their thermal and electrical behavior will be exclusively due to the applied coating and the resulting thermal improvement. 3.2. Results of indoor experiments: TEST-1 and TEST-2 Once it has been verified that there are no appreciable differences between the two prototypes, the HPs of one of them (TEG_RC) have been coated with the radiative surface treatment described in Section 2.3 and the thermal characterization tests have been repeated in both natural Fig. 5. Evolution of the temperature difference between the cold face of the convection and forced convection with an air velocity of 1 m/s. modules and the environment of both prototypes (TEG_RC and TEG_w/o_RC), As mentioned in the methodology section, in each test, both TEG for 40 W of heat provided and 2.2 Ω of load resistance. 6 D. Astrain et al. Applied Thermal Engineering 225 (2023) 120162 radiative treatment or not. Thus, for example, for 30 W of heat input under forced convection, TEG_RC is generating 11 % more electric power than the TEG_w/o_RC thanks to the radiative coating. In the case of natural convection, this improvement rises to 38.8 %. When 40 W are applied to the generators this effect becomes more visible, reaching an increase in the electric generation of 54.8 % thanks to the treatment. It is noteworthy that even under natural convection, the electric power output of the TEG_RC is very similar to the electric generation of the TEG_w/o_RC under forced convection. This effect means that the electric production of the thermoelectric generators can be maintained at its best values even in more unfavorable conditions, such as the absence of external air velocity. This improvement can have a large impact on thermoelectric applications such as those developed in [20–22], avoid­ Fig. 6. Results of the thermal resistance of the heat-pipes for the TEG_RC and ing energy losses in conditions of no wind. TEG_w/o_RC, both under forced convection with an air velocity of 1 m/s (TEST- 1) and under natural convection (TEST-2), for different heat power inputs. 3.3. Results of outdoor experiments: TEST-3 natural convection, and it is observed that the trend is the expected one In this section, the results obtained in the outdoor tests are shown. In for a thermoelectric generator: an initial increase in the electric gener­ these experiments both the ambient temperature and the wind velocity ation as the load resistance increases up to the optimum point, followed have been recorded since these are conditions that cannot be controlled by a decrease in the power output. Comparing the behavior of both as they depend on the weather. prototypes under forced and free conditions, it can be seen that there is Fig. 9 shows the temporal evolution of the temperature difference hardly any difference in the electric production for low heat fluxes, between the cold face of the thermoelectric modules and the environ­ lower than 20 W. However, with higher heat fluxes, the differences ment temperature, as well as the wind velocity during the test. As it can become significant. Under forced convection, TEG_w/o_RC generates be seen, there is a clear relationship between that temperature differ­ 0.48 W, while the TEG_RC reaches a production of 0.52 W, which rep­ ence achieved and the presence of gusts of wind. With higher wind ve­ resents an increase of 8.3 % of the electric power output. Once the forced locities, where heat transfer mainly occurs by means of forced component of the convection is eliminated and a free convection is convection, a temperature difference of around 50 ◦ C is achieved, while, established, there is a decrease in the power generation from both pro­ with very low wind velocities, the temperature difference increase, totypes, since, as seen in Fig. 6, the resistance of the heat exchangers on meaning that the heat dissipation is poorer and consequently the ther­ the cold side worsens and, therefore, the generation decreases, as mal resistance of the HPs increases. However, as it has been seen in the described in Eq. (5), (6) and (8). However, it is remarkable that, while tests inside, this rise in the thermal resistance is smaller in the TEG_RC, the TEG_w/o_RC reduces its electric generation by 35 %, down to 0.31 reaching a temperature difference of 80 ◦ C. Whereas in the TEG_w/o_RC W, in the absence of air velocity; the TEG_RC only reduces the electric the temperature difference rises up to more than 100 ◦ C. production by 7.7 %, generating 0.48 W. These results show that the For both prototypes, these results show that the thermal resistance of application of the radiative surface treatment to the HPs avoids the the HPs of has a direct influence on the electric generation. As shown in drastic decrease in the generated electric power that usually occurs in Fig. 10, electric power output rises with an increase in the wind velocity, thermoelectric generators when they work under free convection, reaching an average generation of 0.57 W with TEG_RC and 0.56 W with keeping the production of electrical energy much more constant. TEG_w/o_RC. This electric generation decreases when the wind slows Additional information of TEST-1 and TEST-2 as the hot and cold temperatures of the thermoelectric modules, environment temperature, Table 3 electric power output and cold-side thermal resistance of the heat ex­ Experimental values for TEST-1 and TEST-2 under optimal electric load resis­ changers; is included in Table 3. tance and heat flux of 40 W. Fig. 8 shows the electric power generation for different heat fluxes Test Device TH (◦ C) TC (◦ C) T∞ (◦ C) Ẇe (W) RC (K/W) supplied to the prototypes that are connected to the optimal load resistance. It can be seen that, at low heat fluxes, the differences in 1 (forced) TEG_RC 130.6 76.2 15.3 0.52 1.60 generation are really small, but from 20 W of thermal power on, there TEG_w/o_RC 133.1 82.3 15.3 0.48 1.72 2 (natural) TEG_RC 151.9 103.2 14.8 0.48 2.29 are noticeable differences between the electric power output under TEG_w/o_RC 171.8 126.0 14.8 0.31 3.04 natural and forced convection depending on whether the device has the Fig. 7. Electric power output of both TEG prototypes under forced convection, with a wind velocity of 1 m/s (TEST-1) and natural convection (TEST-2), for different load resistances and two heat fluxes provided (20 and 40 W). 7 D. Astrain et al. Applied Thermal Engineering 225 (2023) 120162 Fig. 8. Electric power output from both TEG prototypes with the optimal load resistance, under natural and free convection (with an air velocity of 1 m/s), varying the heat flux provided. Fig. 9. Evolution of the temperature difference between the cold face of the thermoelectric modules in the generators and the environment temperature; and wind velocity over time during TEST-3. Fig. 10. Electric power output from both prototypes and wind velocity at the test site. down showing a similar behavior detected in the tests inside the labo­ counteracts the reduction in the electric production caused by a poorer ratory. During periods with low wind velocities, in which the forced convective heat transfer. convection component disappears, the electric generation from TEG_RC Finally, Table 4 compares the electric generation achieved with both drops to 0.49 W, which represents a 14 % decrease. While in the case of prototypes in all the conditions studied, highlighting the increase in the the TEG_w/o_RC, the generation falls to 0.33 W, a 41.6 % reduction. production of electric power as a result of the radiative treatment placed Again, it is observed that the radiative treatment considerably on the TEG_RC prototype. It is observed that, in all situations, the 8 D. Astrain et al. Applied Thermal Engineering 225 (2023) 120162 Table 4 Table 5 Comparison of electric power output under all studied conditions, for a heat flux Comparison of the conversion efficiency of the generators under all studied of 40 W. conditions, for a heat flux of 40 W. Indoors Outdoors Indoors Outdoors Forced Natural Forced No Forced Natural Forced No wind convection (1 m/ convection convection (1.5 wind convection (1 convection convection (1.5 s) m/s) m/s) m/s) TEG_RC 0.52 W 0.48 W 0.57 W 0.49 TEG_RC 1.28 % 1.19 % 1.40 % 1.38 % W TEG_w/o_RC 1.19 % 0.77 % 1.21 % 0.82 % TEG_w/o_RC 0.48 W 0.31 W 0.56 W 0.33 Improvement 8.2 % 54.2 % 16.1 % 68.73 % W with RC Improvement 8.3 % 54.8 % 1.8 % 48.5 with RC % difference between the cold face of the thermoelectric modules and the environment increases, although to a lesser extent in the TEG that in­ TEG_RC generates more electricity than the TEG_w/o_RC. This increase cludes de radiative coating. This effect has a positive influence on the in generation is much more noticeable when natural convection condi­ electric generation, which, although it decreases when there is no wind, tion is dominant (as shown during the indoor tests under free convection the reduction is only of 14 % thanks to the radiative surface treatment. or outdoors without wind). Whereas in those cases where forced con­ Meanwhile, in the TEG without the radiative coating, the generation vection is important, the improvements achieved are reduced. falls by 41.6 %. In the cases of indoor tests, the air velocity recorded in the tests In short, it has been proven that the radiative coating over the heat- carried out inside the climatic chamber was 1 m/s, while in the case of pipes of the cold side heat exchangers has a positive influence on the outdoor wind tests, the speed was 1.5 m/s, which explains why the heat dissipation from the thermoelectric modules to the environment. percentage of improvement due to the radiative treatment is lower, This translated into an increase in the electric production of the ther­ noting the fact that for high wind velocities, forced convection has a moelectric generator of 8.3 % with outdoor wind velocities of 1 m/s and dominant effect over radiation. up to 54.8 % when the device is working under free convection condi­ Table 5 gathers the values of the efficiency achieved by the ther­ tions. This increase in the electric generation of thermoelectric devices is moelectric generators under all the studied conditions. To estimate the very interesting, since it is an improvement in the heat exchangers that conversion efficiency of the generators the electric power output is being does not imply the incorporation of moving parts, nor loss of robustness divided by the rate of heat flow absorbed from the hot source, according or compactness, typical of thermoelectric systems. to Equation (13). Ẇ e CRediT authorship contribution statement η= (13) Q̇H David Astrain: Conceptualization, Methodology, Validation, Inves­ As it can be seen, the radiative coating improves again the perfor­ tigation, Writing – original draft, Supervision. Juliana Jaramillo-Fer­ mance of the device specially under natural or free convection condi­ nandez: Methodology, Validation, Investigation, Writing – original tions. The maximum efficiency obtained from the experiments gets a draft, Writing – review & editing, Supervision. Miguel Araiz: Method­ value of 1.4 %. ology, Investigation, Data curation, Writing – original draft, Writing – review & editing. Achille Francone: Methodology, Investigation. Leyre 4. Conclusions Catalán: Methodology, Investigation, Data curation, Writing – review & editing. Alejandra Jacobo-Martín: Methodology, Investigation, For the first time, a radiative surface treatment has been applied to Writing – review & editing. Patricia Alegría: Methodology, Investiga­ the heat-pipe heat exchanger of a thermoelectric generator to improve tion, Writing – review & editing. Clivia M. Sotomayor-Torres: Writing the electric power output of the device. In the results of the tests carried – review & editing, Supervision. out under forced convection conditions, with an air velocity of 1 m/s, the heat-pipes with the coating improved by 7 % the thermal resistance Declaration of Competing Interest with respect to the heat exchanger without the radiative film. In the case of the tests under natural convection, the thermal resistances of both The authors declare that they have no known competing financial devices increased. However, while in the case of the TEG without the interests or personal relationships that could have appeared to influence radiative treatment, the thermal resistance increased by 86.4 %, the the work reported in this paper. thermal resistance of the TEG with the treatment only increased by 54.7 %, which means that the heat-pipes with the radiative coating offer a 29 Data availability % less thermal resistance. This improvement in the heat exchangers is reflected in the electric No data was used for the research described in the article. power output that both prototypes are capable of generating. Under forced convection a maximum electric production of 0.52 W is achieved Acknowledgements with the TEG with the radiative cooling treatment, and 0.48 W with the TEG without the coating. As it could be expected, under natural con­ The authors acknowledge the support of the Spanish Ministry of vection conditions, the electric generation decreases. However, this Science, Innovation and Universities, and the European Regional generation drop is only of 7.7 % with the TEG_RC, while with the Development Fund, under grants PID2021-124014OB-I00 (VIVOTEG), TEG_w/o_RC the decrease is 35.4 %, proving that the radiative film TED2021-129359B-I00 (GEOTEG), PGC2018-101743-B-I00 (SIP) and greatly minimizes the loss of electric generation that occurs in the RTI2018-093921-A-C44 (SMOOTH). The ICN2 authors thank the absence of forced convection with the heat-pipes. CERCA Program/Generalitat de Catalunya. Open access funding pro­ The experimental tests performed outdoors, with changing weather vided by Universidad Pública de Navarra. conditions, are in agreement with those carried out in the laboratory, showing that when the wind velocity is very low or null, the temperature 9 D. Astrain et al. Applied Thermal Engineering 225 (2023) 120162 Appendix A. Supplementary material [24] 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. [25] Y. Zhu, D. Wang, C. Fang, P. He, Y.H. Ye, A multilayer emitter close to ideal solar Supplementary data to this article can be found online at https://doi. reflectance for efficient daytime radiative cooling, Polymers (Basel). 11 (2019) org/10.1016/j.applthermaleng.2023.120162. 1–10, https://doi.org/10.3390/polym11071203. [26] A. Hervé, J. Drévillon, Y. Ezzahri, K. Joulain, Radiative cooling by tailoring surfaces with microstructures: Association of a grating and a multi-layer structure, References J. Quant. Spectrosc. Radiat. Transf. 221 (2018) 155–163, https://doi.org/10.1016/ j.jqsrt.2018.09.015. [1] S.M. Pourkiaei, M.H. Ahmadi, M. Sadeghzadeh, S. Moosavi, F. Pourfayaz, L. Chen, [27] S. Meng, L. Long, Z. Wu, N. Denisuk, Y. Yang, L. Wang, F. Cao, Y. Zhu, Scalable M.A. Pour Yazdi, R. Kumar, Thermoelectric cooler and thermoelectric generator dual-layer film with broadband infrared emission for sub-ambient daytime devices: a review of present and potential applications, modeling and materials, radiative cooling, Sol. Energy Mater. Sol. Cells 208 (2020), 110393, https://doi. Energy 186 (2019), 115849, https://doi.org/10.1016/j.energy.2019.07.179. org/10.1016/j.solmat.2020.110393. [2] D. Champier, Thermoelectric generators: A review of applications, Energy Convers [28] E. Blandre, R.A. Yalçin, K. Joulain, J. Drévillon, Microstructured surfaces for Manag. 140 (2017) 167–181, https://doi.org/10.1016/j.enconman.2017.02.070. colored and non-colored sky radiative cooling, Opt. Express 28 (2020) [3] W. He, G. Zhang, X. Zhang, J. Ji, G. Li, X. Zhao, Recent development and 29703–29713, https://doi.org/10.1364/OE.401368. application of thermoelectric generator and cooler, Appl. Energy 143 (2015), [29] J. Jaramillo-Fernandez, G.L. Whitworth, J.A. Pariente, A. Blanco, P.D. Garcia, https://doi.org/10.1016/j.apenergy.2014.12.075. C. Lopez, C.M. Sotomayor-Torres, A Self-Assembled 2D Thermofunctional Material [4] A. Montecucco, J. Siviter, A.R. Knox, A Combined Heat and Power System for for Radiative Cooling, Small 15 (2019) 1905290, https://doi.org/10.1002/ Solid-fuel Stoves Using Thermoelectric Generators, Energy Procedia 75 (2015) smll.201905290. 597–602, https://doi.org/10.1016/j.egypro.2015.07.462. [30] G.L. Whitworth, J. Jaramillo-Fernandez, J.A. Pariente, P.D. Garcia, A. Blanco, [5] X.F. Zheng, C.X. Liu, R. Boukhanouf, Y.Y. Yan, W.Z. Li, Experimental study of a C. Lopez, C.M. Sotomayor-Torres, Simulations of micro-sphere/shell 2D silica domestic thermoelectric cogeneration system, Appl. Therm. Eng. 62 (2014) 69–79, photonic crystals for radiative cooling, Opt. Express 29 (2021) 16857–16866, https://doi.org/10.1016/j.applthermaleng.2013.09.008. https://doi.org/10.1364/OE.420989. [6] A. Mohammadnia, A. Rezania, B.M. Ziapour, F. Sedaghati, L. Rosendahl, Hybrid [31] S. Atiganyanun, J.B. Plumley, S.J. Han, K. Hsu, J. Cytrynbaum, T.L. Peng, S. energy harvesting system to maximize power generation from solar energy, Energy M. Han, S.E. Han, Effective Radiative Cooling by Paint-Format Microsphere-Based Convers Manag. 205 (2020), 112352, https://doi.org/10.1016/j. Photonic Random Media, ACS Photonics 5 (2018) 1181–1187, https://doi.org/ enconman.2019.112352. 10.1021/acsphotonics.7b01492. [7] M.A. Karri, E.F. Thacher, B.T. Helenbrook, Exhaust energy conversion by [32] E. Rephaeli, A. Raman, S. Fan, Ultrabroadband Photonic Structures To Achieve thermoelectric generator: Two case studies, Energy Convers Manag. 52 (2011) High-Performance Daytime Radiative Cooling, Nano Lett. 13 (2013) 1457–1461, 1596–1611, https://doi.org/10.1016/j.enconman.2010.10.013. https://doi.org/10.1021/nl4004283. [8] F. Huang, J. Zheng, J.M. Baleynaud, J. Lu, Heat recovery potentials and [33] Z. Huang, X. Ruan, Nanoparticle embedded double-layer coating for daytime technologies in industrial zones, J. Energy Inst. 90 (2017) 951–961, https://doi. radiative cooling, Int. J. Heat Mass Transf. 104 (2017) 890–896, https://doi.org/ org/10.1016/j.joei.2016.07.012. 10.1016/j.ijheatmasstransfer.2016.08.009. [9] D. Zhao, G. Tan, A review of thermoelectric cooling: Materials, modeling and [34] C. Yijun, M. Jyotirmoy, L. Wenxi, S.-W. Ajani, T. Cheng-Chia, H. Wenlong, S. Sajan, applications, Appl. Therm. Eng. 66 (2014) 15–24, https://doi.org/10.1016/j. Y. Nanfang, H.R.P. S., C. Anyuan, Y. Yuan, Colored and paintable bilayer coatings applthermaleng.2014.01.074. with high solar-infrared reflectance for efficient cooling, Sci Adv. 6 (2021) [10] D. Astrain, J.G. Vián, A. Martínez, A. Rodríguez, Study of the influence of heat eaaz5413. https://doi.org/10.1126/sciadv.aaz5413. exchangers’ thermal resistances on a thermoelectric generation system, Energy 35 [35] A. Leroy, B. Bhatia, C.C. Kelsall, A. Castillejo-Cuberos, M. di Capua H., L. Zhao, L. (2010) 602–610, https://doi.org/10.1016/j.energy.2009.10.031. Zhang, A.M. Guzman, E.N. Wang, High-performance subambient radiative cooling [11] X. Niu, J. Yu, S. Wang, Experimental study on low-temperature waste heat enabled by optically selective and thermally insulating polyethylene aerogel, Sci thermoelectric generator, J. Power Sources 188 (2009) 621–626, https://doi.org/ Adv. 5 (2019) eaat9480. https://doi.org/10.1126/sciadv.aat9480. 10.1016/j.jpowsour.2008.12.067. [36] M. Yang, W. Zou, J. Guo, Z. Qian, H. Luo, S. Yang, N. Zhao, L. Pattelli, J. Xu, D. [12] T.Y. Kim, A.A. Negash, G. Cho, Waste heat recovery of a diesel engine using a S. Wiersma, Bioinspired “Skin” with Cooperative Thermo-Optical Effect for thermoelectric generator equipped with customized thermoelectric modules, Daytime Radiative Cooling, ACS Appl. Mater. Interfaces 12 (2020) 25286–25293, Energy Convers Manag. 124 (2016) 280–286, https://doi.org/10.1016/j. https://doi.org/10.1021/acsami.0c03897. enconman.2016.07.013. [37] H. Zhang, K.C.S. Ly, X. Liu, Z. Chen, M. Yan, Z. Wu, X. Wang, Y. Zheng, H. Zhou, [13] P. Aranguren, D. Astrain, M.G. Pérez, Computational and experimental study of a T. Fan, Biologically inspired flexible photonic films for efficient passive radiative complete heat dissipation system using water as heat carrier placed on a cooling, PNAS 117 (2020) 14657–14666, https://doi.org/10.1073/ thermoelectric generator, Energy 74 (2014) 346–358, https://doi.org/10.1016/j. pnas.2001802117. energy.2014.06.094. [38] D. Li, X. Liu, W. Li, Z. Lin, B. Zhu, Z. Li, J. Li, B. Li, S. Fan, J. Xie, J. Zhu, Scalable [14] Y. Liu, Z. Shi, G. Wang, Y. Yan, Y. Zhang, Experimental Investigation for a Novel and hierarchically designed polymer film as a selective thermal emitter for high- Prototype of a Thermoelectric Power Generator With Heat Pipes, Front. Energy performance all-day radiative cooling, Nat. Nanotechnol. 16 (2021) 153–158, Res. 9 (2021) 1–13, https://doi.org/10.3389/fenrg.2021.744366. https://doi.org/10.1038/s41565-020-00800-4. [15] P. Aranguren, D. Astrain, A. Rodríguez, A. Martínez, Experimental investigation of [39] X. Wang, X. Liu, Z. Li, H. Zhang, Z. Yang, H. Zhou, T. Fan, Scalable Flexible Hybrid the applicability of a thermoelectric generator to recover waste heat from a Membranes with Photonic Structures for Daytime Radiative Cooling, Adv. Funct. combustion chamber, Appl. Energy 152 (2015) 121–130, https://doi.org/10.1016/ Mater. 30 (2020) 1907562, https://doi.org/10.1002/adfm.201907562. j.apenergy.2015.04.077. [40] W. Huang, Y. Chen, Y. Luo, J. Mandal, W. Li, M. Chen, C.-C. Tsai, Z. Shan, N. Yu, [16] A. Date, A. Date, C. Dixon, R. Singh, A. Akbarzadeh, Theoretical and experimental Y. Yang, Scalable Aqueous Processing-Based Passive Daytime Radiative Cooling estimation of limiting input heat flux for thermoelectric power generators with Coatings, Adv. Funct. Mater. 31 (2021) 2010334, https://doi.org/10.1002/ passive cooling, Sol. Energy 111 (2015) 201–217, https://doi.org/10.1016/j. adfm.202010334. solener.2014.10.043. [41] J. Mandal, Y. Fu, A.C. Overvig, M. Jia, K. Sun, N.N. Shi, H. Zhou, X. Xiao, N. Yu, [17] B. Orr, B. Singh, L. Tan, A. Akbarzadeh, Electricity generation from an exhaust heat Y. Yang, Hierarchically porous polymer coatings for highly efficient passive recovery system utilising thermoelectric cells and heat pipes, Appl. Therm. Eng. 73 daytime radiative cooling, Science 362 (2018) (1979) 315–319. (2014) 586–595, https://doi.org/10.1016/j.applthermaleng.2014.07.056. [42] M. Yang, M. Yang, W. Zou, W. Zou, J. Guo, J. Guo, Z. Qian, Z. Qian, H. Luo, H. Luo, [18] M. Araiz, A. Martínez, D. Astrain, P. Aranguren, Experimental and computational S. Yang, S. Yang, N. Zhao, N. Zhao, L. Pattelli, L. Pattelli, J. Xu, D.S. Wiersma, D. study on thermoelectric generators using thermosyphons with phase change as S. Wiersma, Bioinspired “skin” with Cooperative Thermo-Optical Effect for heat exchangers, Energy Convers Manag. 137 (2017) 155–164, https://doi.org/ Daytime Radiative Cooling, ACS Appl. Mater. Interfaces 12 (2020) 25286–25293, 10.1016/j.enconman.2017.01.046. https://doi.org/10.1021/acsami.0c03897. [19] Y. Huang, D. Xu, J. Kan, W. Li, Study on field experiments of forest soil [43] S. Gamage, D. Banerjee, M.M. Alam, T. Hallberg, C. Åkerlind, A. Sultana, thermoelectric power generation devices, PLoS One 14 (2019) 1–13, https://doi. R. Shanker, M. Berggren, X. Crispin, H. Kariis, D. Zhao, M.P. Jonsson, Reflective org/10.1371/journal.pone.0221019. and transparent cellulose-based passive radiative coolers, Cellul. (2021), https:// [20] L. Catalan, M. Araiz, P. Aranguren, G.D. Padilla, P.A. Hernandez, N.M. Perez, C. doi.org/10.1007/s10570-021-04112-1. G. de la Noceda, J.F. Albert, D. Astrain, Prospects of autonomous volcanic [44] B. Xiang, R. Zhang, Y. Luo, S. Zhang, L. Xu, H. Min, S. Tang, X. Meng, 3D porous monitoring stations: Experimental investigation on thermoelectric generation from polymer film with designed pore architecture and auto-deposited SiO2 for highly fumaroles, Sensors (Switzerland). 20 (2020) 1–21, https://doi.org/10.3390/ efficient passive radiative cooling, Nano Energy 81 (2021), 105600, https://doi. s20123547. org/10.1016/j.nanoen.2020.105600. [21] L. Catalan, A. Garacochea, A. Casi, M. Araiz, P. Aranguren, D. Astrain, [45] X. Chen, M. He, S. Feng, Z. Xu, H. Peng, S. Shi, C. Liu, Y. Zhou, Cellulose-based Experimental evidence of the viability of thermoelectric generators to power porous polymer film with auto-deposited TiO2 as spectrally selective materials for volcanic monitoring stations, Sensors (Switzerland). 20 (2020) 1–25, https://doi. passive daytime radiative cooling, Opt Mater (Amst). 120 (2021), 111431, https:// org/10.3390/s20174839. doi.org/10.1016/j.optmat.2021.111431. [22] P. Alegria, L. Catalan, M. Araiz, A. Rodriguez, D. Astrain, Experimental [46] Y. Chen, B. Dang, J. Fu, C. Wang, C. Li, Q. Sun, H. Li, Cellulose-Based Hybrid development of a novel thermoelectric generator without moving parts to harness Structural Material for Radiative Cooling, Nano Lett. 21 (2021) 397–404, https:// shallow hot dry rock fields, Appl. Therm. Eng. 200 (2022), 117619, https://doi. doi.org/10.1021/acs.nanolett.0c03738. org/10.1016/j.applthermaleng.2021.117619. [47] S. Gamage, E.S.H. Kang, C. Åkerlind, S. Sardar, J. Edberg, H. Kariis, T. Ederth, [23] W.M. Rohsenow, Handbook of Heat Transfer, 1998. M. Berggren, M.P. Jonsson, Transparent nanocellulose metamaterial enables 10 D. Astrain et al. Applied Thermal Engineering 225 (2023) 120162 controlled optical diffusion and radiative cooling, J Mater Chem C Mater. 8 (2020) [53] R.P. S.A.U., ProRox WM 970 Datasheet, 2012. 11687–11694, https://doi.org/10.1039/D0TC01226B. [54] Armacell, ArmaFlex insulation Datasheet, 2021. [48] W. Wei, Y. Zhu, Q. Li, Z. Cheng, Y. Yao, Q. Zhao, P. Zhang, X. Liu, Z. Chen, F. Xu, [55] M.A. Verschuuren, M. Megens, Y. Ni, H. van Sprang, A. Polman, Large area Y. Gao, An Al2O3-cellulose acetate-coated textile for human body cooling, Sol. nanoimprint by substrate conformal imprint lithography (SCIL), Adv. Opt. Technol. Energy Mater. Sol. Cells 211 (2020), 110525, https://doi.org/10.1016/j. 6 (2017) 243–264, https://doi.org/10.1515/aot-2017-0022. solmat.2020.110525. [56] B. Kwon, J.H. Kim, Importance of Molds for Nanoimprint Lithography: Hard, Soft, [49] J. Jaramillo-Fernandez, H. Yang, L. Schertel, G.L. Whitworth, P.D. Garcia, and Hybrid Molds, J. Nanosci. 2016 (2016) 6571297, https://doi.org/10.1155/ S. Vignolini, C.M. Sotomayor-Torres, Highly-scattering cellulose-based lms for 2016/6571297. radiative cooling, Adv. Sci. 2104758 (2021), https://doi.org/10.1002/ [57] I. Rodríguez, J. Hernández, Soft thermal nanoimprint and hybrid processes to advs.202104758. produce complex structures, Nanofabrication: Nanolithography Techniques and [50] Z. Zhan, M. ElKabbash, Z. Li, X. Li, J. Zhang, J. Rutledge, S. Singh, C. Guo, Their Applications. 7 (2020) 1–33. Enhancing thermoelectric output power via radiative cooling with nanoporous [58] N. Kehagias, A. Francone, M. Guttmann, F. Winkler, A. Fernández, C.M. Sotomayor alumina, Nano Energy 65 (2019), 104060, https://doi.org/10.1016/j. Torres, Fabrication and replication of re-entrant structures by nanoimprint nanoen.2019.104060. lithography methods, J. Vac. Sci. Technol. B 36 (2018) 06JF01, https://doi.org/ [51] D.M. Rowe, Thermoelectrics Handbook: Macro to Nano, 1st ed., Boca Raton, 2006. 10.1116/1.5048241. [52] Marlow Industries Inc., Marlow TG12-8 Datasheet, (2016) 1–2. http://www.marl [59] S.WG. Coleman HW, Experimentation, validation and uncertainty analysis for ow.com/media/marlow/product/downloads/tg12-8-01ls/TG12-8_Data_Sheet_ engineers., n.d. https://doi.org/https://doi.org/10.1017/ RevM.pdf (accessed January 1, 2016). CBO9781107415324.004. 11