Supergiant Barocaloric Effects in Acetoxy Silicone Rubber over a Wide Temperature Range: Great Potential for Solid-state Cooling

Solid-state cooling based on caloric effects is considered a viable alternative to replace the conventional vapor-compression refrigeration systems. Regarding barocaloric materials, recent results show that elastomers are promising candidates for cooling applications around room-temperature. In the present paper, we report supergiant barocaloric effects observed in acetoxy silicone rubber—a very popular, low-cost and environmentally friendly elastomer. Huge values of adiabatic temperature change and reversible isothermal entropy change were obtained upon moderate applied pressures and relatively low strains. These huge barocaloric changes are associated both to the polymer chain rearrangements induced by confined compression and to the first-order structural transition. The results are comparable to the best barocaloric materials reported so far, opening encouraging prospects for the application of elastomers in near future solid-state cooling devices.


INTRODUCTION
The current priorities in sustainability and energy efficiency lead to the study and development of new refrigeration technologies. In this context, solid-state cooling based on caloric effects (also called i-caloric effects [1] ) has shown to be a promising alternative to the conventional vapor-compression systems. The caloric effects can be described by an isothermal entropy change (ΔS T ) and an adiabatic temperature change (ΔT S ), both manifested when an external field is applied on a material. Depending on the nature of this external field (magnetic field, electric field, or stress field), the caloric effects can be categorized as magnetocaloric effect, electrocaloric effect, and mechanocaloric effect (σ-CE). The latter can still be divided in elastocaloric effect, driven by uniaxial stress, and barocaloric effect, driven by isotropic stress variations.
The first caloric effect reported was the elastocaloric effect in natural rubber (NR), observed by John Gough in the begin-ning of the nineteenth century. [2] Nevertheless, the research into caloric effects has strengthened only in the last two decades due to the discoveries of giant magnetocaloric effect in Gd 5 Si 2 Ge 2 compound [3] and giant electrocaloric effect in PbZr 0.95 Ti 0.05 O 3 thin film. [4] Even though σ-CE is the least researched among caloric effects, interesting results were reported for vulcanized natural rubber (VNR) and other synthetic elastomers already in the 1940 decade. [5,6] Shapememory alloys also exhibit promising σ-CE around room temperature. [7][8][9][10][11] More recently, giant barocaloric effects were reported in (NH 4 ) 2 SO 4 and molecular crystal below room temperature. [12,13] Another promising class of material is plastic crystals, which demonstrated colossal barocaloric effects near room temperature. [14,15] Regarding elastomers, the number of studies reporting large σ-CE values is gradually growing in the last years. [16][17][18][19][20][21][22][23] On the contrary of what is observed in shape memory alloys and ionic salts, elastomers show significant elastocaloric and barocaloric effects even in the absence of phase transitions. This behavior is assigned to the rearrangement of polymer chains induced by the application of mechanical stress. [16] In this context, the silicone rubber can be considered a po-tential mechanocaloric material, since it is generally elastomeric and may present favorable structural transitions. The popular term "silicone" includes any organosilicon compound containing at least one pair of silicon atoms linked by an oxygen atom (Si-O-Si). Properly speaking, the correct terminology for these compounds is "polysiloxane", whose formula is (RR′SiO) n , where R and R′ are alkyl groups. [24] Among the polysiloxanes, there is a group named room-temperature vulcanizing silicone rubbers (RTV-SR), which consist of polydimethylsiloxane, curing agent, fillers, and additives. Acetoxy silicone rubber (ASR) is a type of RTV-SR, which releases acetic acid during vulcanization. Recently, our preprint results [25] motivated studies on ASR as a viable alternative for using as a refrigerant in solid-state devices based in barocaloric effects. [26][27][28] In the present paper, we systematically investigate the barocaloric effect in ASR around room temperature. ΔS T was evaluated following an indirect method using a Maxwell relation, and ΔT S was directly measured. The results allowed us to establish a link between the crystallineamorphous transition and the very large barocaloric changes observed in ASR. The experiments with pressure were performed in the experimental setup detailed elsewhere; the isostatic condition for barocaloric measurements was also discussed. [20,21,23,29] Fig. S1 (in the electronic supplementary information, ESI), presents a schematic representation of the apparatus used in the barocaloric experiments. Two versions of the pressure chamber are available, consisting of carbon-steel cylinders with an 8 or 12 mm bore. The maximum pressure attainable with each chamber is 173 MPa (12 mm) and 390 MPa (8 mm). Samples are prepared in the proper shape to perfectly fit the chamber. A piston is the only movable part, responsible for compressing the sample. The bottom closure of the chamber has a narrow hole allowing a thermocouple to be placed inside the sample, monitoring its temperature. Another thermocouple, placed inside the chamber wall, monitors the chamber temperature, and its reading is used as a feedback to the temperature control system. A copper coil guides fluids (water or liquid nitrogen) around the chamber, and together with the heating elements, is responsible for controlling the chamber temperature. The chamber is placed on a load gauge and a hydraulic press actuates the piston. The pressure applied to the sample is directly given by the force measured by the load gauge divided by the piston cross-section area. A length gauge monitoring the displacement of the piston retrieves information about the deformation of the sample.

EXPERIMENTAL
ΔT S as a function of temperature (223-333 K) was directly measured using a quasi-adiabatic compression/decompression process. In such process, pressure varies quickly enough to preclude significant heat exchange between the sample and its surroundings. Pressure changes between the minimum and the maximum pressures (p 1 and p 2 , respectively) were in the range of 26.0-390 MPa; compression corresponds to the process p 1 → p 2 , while decompression corresponds to the process p 2 → p 1 . Strain versus temperature curves were measured at different constant pressures (0.9-332 MPa), varying the temperature at a rate of ~3 K·min −1 for cooling and heating processes, within the 213-333 K temperature range. Strain (ε) is defined as ε(p,T) ≡ (l p,T − l 0 )/l 0, where l p,T is the final length of the sample at pressure p, for each temperature T, and l 0 is its initial length measured at ambient pressure (p 0 ) and room temperature (T 0 = 293 K). Since the cross-sectional area (A) of the samples does not vary in confined pressure, the volume change (ΔV) is directly proportional to the strain because ε(p,T) ≡ (Al p,T − Al 0 )/Al 0 = ΔV/V 0 . These strain versus temperature curves up to 173(3) MPa were used in the calculation of ΔS T versus T.
X-ray diffraction (XRD) patterns were measured at the XRD1 beamline, [30,31] at the Brazilian Synchrotron Light Laboratory (LNLS). The beam energy used at XRD1 was 12 keV, and the samples were cooled down by a Cryojet5 (Oxford Instruments), within the 300-100 K temperature range at ambient pressure.
Fourier transform infrared spectroscopy (FTIR) was performed by a spectrometer from PerkinElmer ® (model Spectrum Two), in the spectral range from 4000 cm −1 to 450 cm −1 , with spectral resolution of 1 cm −1 (Fig. S2, in ESI).

RESULTS AND DISCUSSION
In each ε versus T curves on heating process (Fig. 1a), one can see a narrow region where the derivative abruptly increases, and the transition temperature (T TR ) shifts toward higher temperatures for larger applied pressures. This behavior, in addition to the hysteresis shown in Fig. 1(b), strongly indicates a first-order transition. In fact, as observed in the XRD patterns ( Fig. 1c), there is an amount of amorphous phase that becomes crystalline phase when the temperature decreases. To verify the influence of the pressure on T TR , we calculated each T TR from a local maximum in dε/dT versus T curves. Considering a linear fit, we found dT TR /dp = 0.27(1) K·MPa −1 on heating and 0.23(1) K·MPa −1 on cooling (Fig. 1d).
Temperature as a function of time for ASR (see an example in Fig. 2a) was measured at the pressure range of 26.0(5)-390(12) MPa and different initial temperatures (223-333 K). The observed behavior is reversible in a large temperature range (i.e., the adiabatic temperature change during compression and decompression is similar). The asymmetry found at the highest pressure in Fig. 2(a) (the absolute value of temperature change in decompression is higher than in compression for 390 MPa) is related to the partial amorphous-crystalline phase transition. In Fig. 2(b), ΔT S corresponding to the decompression process is displayed. We observe a maximum barocaloric effect of 41.1 K, at ~298 K, for |Δp| = 390 MPa. This ΔT S value, which we classify as supergiant (|ΔT S | ≥ 30 K), is significantly higher than those reported for any barocaloric materials around room temperature (e.g., VNR [21] at ~315 K presents |ΔT S | = 24.9 K for |Δp| = 390 MPa; PDMS [20] at ~283 K presents |ΔT S | = 28.5 K for |Δp| = 390 MPa). It is easy to see that a |ΔT S | maximum appears for |Δp| = 173 MPa, and this maximum shifts to higher temperatures when the pressure increases. Considering a linear rate for the temperatures at the maximum ΔT (T m,ΔT ) with pressure, we have dT m,ΔT /dp = 0.22(8) K·MPa −1 . The outstanding ΔT S values registered in ASR can be understood as a combination of the structural changes associated to the crystalline-amorphous transition and polymer chain rearrangements unrelated to phase transitions. Above T m,ΔT , the contribution to the barocaloric effect comes entirely from the amorphous phase. If we compare the |ΔT S | values at T m,ΔT with the |ΔT S | values above T m,ΔT , an increase of 33%, 34%, and 47% is obtained for 173, 273, and 390 MPa, respectively. In addition, |ε| is less than 25% up to |Δp| = 390 MPa.

≅
T curves (Fig. 2b, dT m,ΔT /dp = 0.22(8) K·MPa −1 ). Thus, ΔS T and ΔT S maxima shift similarly with pressure (taking the errors into account), although obtained from different processes. As expected, dT m,ΔS /dp dT TR /dp, since ΔS T curves from Fig. 3(a) are obtained from ε versus T data on heating process. Similar behaviors can be observed taking ΔS T on cooling process (Fig. 3b). Moreover, ΔS T on heating and on cooling show irreversibility due to the hysteresis (i.e., the difference between ΔS T from heating and cooling processes cannot be negligible). In order to eliminate the irreversible contribution, the reversible ΔS T as a function of temperature was estimated from the overlapping ΔS T on heating and on cooling processes (Fig. 3c). We can see that the reversible |ΔS T | is still supergiant (maximum of 182(46) J·kg −1 ·K −1 for |Δp| = 173 MPa). This reversibility is essentially due to the amorphous phase. Regarding the peaks observed on Fig. 3, we must note that local fluctuations occur across the entire temperature and pressure ranges; such fluctuations are artifacts generated by the experimental and numerical precision of the measurements and data analysis procedures. Finally, we compared the barocaloric properties of ASR around room temperature with promising barocaloric materials in literature. [9,11,[19][20][21][32][33][34][35] The normalized adiabatic temperature change (|ΔT S /Δp|) as a function of |ΔT S | is plotted in Fig. 4(a), where simultaneously larger |ΔT S | and |ΔT S /Δp| values indicate higher potential for barocaloric cooling applications. |ΔT S /Δp| reaches a huge maximum value of 120 K·GPa −1 for |ΔT S | = 20.4 K. It is noteworthy the results for ASR exceed those of any other barocaloric material in a broad temperature range. We also calculated the normalized refrigerant capacity (NRC) as a function of the temperature difference between hot reservoir and cold reservoir (ΔT h-c ≡ T hot − T cold ), following the equation: [20,21] as shown in Fig. 4(b). For ASR, we fixed the hot reservoir at 300 K, |Δp| = 173 MPa and we used ΔS T data from Fig. 3(a) and reversible ΔS T data from Fig. 3(c). Again, the NRC values of ASR surpasses all barocaloric materials in the full ΔT h-c range, and this difference increases as a function of ΔT h-c , reaching ~15 kJ·kg −1 ·GPa −1 for ΔT h-c = 25 K. Moreover, the curve keeps a clear trend to increase, following a distinct behavior to that observed for the other non-elastomeric barocaloric materials in the comparison. Also, the relative cooling power (RCP ≡ |∆S max × δT FWHM |, where ∆S max is the maximum entropy change and δT FWHM is the full width of the entropy change peak at half maximum) for ASR taking reversible ΔS T is 13 (3)  are due to combined effects of the first-order crystallineamorphous transition and the polymer chain rearrangements unrelated to the crystallization process. These are the largest barocaloric effects known among elastomers so far. The major concern with elastomers is related to the low thermal conductivity, which can be improved by developing polymers with high conductivity fillers. But considering all the favorable characteristics exhibited by ASR concerning solid-state cooling, we can foresee a practical interest in the development of energy-efficient and environmental-friendly refrigeration devices based on barocaloric effect in polysiloxanes and other elastomers. As a further matter, pressure-induced crystallization on an amorphous polymer can dramatically modify its barocaloric properties; controlling or tuning the degree of crystallization in the barocaloric processes may lead to even greater values of ΔT S , ΔS T , refrigerant capacity, and relative cooling power.   [20] VNR (|Δp| = 173 MPa), [19,21] Mn 3 GaN (maximum |ΔT S | reported for |Δp| = 93 MPa), [32] Gd 5 Si 2 Ge 2 (maximum |ΔT S | reported for |Δp| = 200 MPa), [33] La-Fe-Si-Co (maximum |ΔT S | reported for |Δp| = 200 MPa), [34] Mn-Co-Ge-In (maximum |ΔT S | reported for |Δp| = 300 MPa), [11] and MnNiSi-FeCoGe

Electronic Supplementary Information
Electronic supplementary information (ESI) is available free of charge in the online version of this article at https://doi.org/10.1007/s10118-020-2423-9.