A self-actuated electrocaloric polymer heat pump design exploiting the synergy of electrocaloric effect and electrostriction

Caloric cooling is an attractive family of technologies owing to their environmental friendliness and potential for higher efficiency than present refrigeration systems. Cooling devices based on the electrocaloric (EC) effect specifically have the added benefit of being easily miniaturized, enabling applications in electronic thermal management, wearables and localized cooling. A challenge in prior compact EC cooling devices has been the need for a separate actuation mechanism to cyclically contact the EC material with hot and cold interfaces. Here, we propose a self-actuated EC polymer heat pump, exploiting recent discoveries of giant EC and electromechanical responses under low electric fields in P(VDF-TrFE-CFE-FA) (VDF: vinylidene fluoride, TrFE: trifluoroethylene, CFE: chlorofluoroethylene, FA: fluorinated alkynes) relaxor tetrapolymers. We show that the transverse electroactuation of P(VDF-TrFE-CFE-FA) relaxor tetrapolymer films can be tailored over a broad range, from strong actuation to weak actuation, without affecting the high EC response. Using this principle, a unimorph actuator was constructed from two EC tetrapolymer layers with large differences in electroactuation. This device autonomously achieves a large displacement between the heating and cooling cycles of the EC films, which could be used to switch thermal contact between hot and cold interfaces. This concept could thus enable highly efficient and compact EC heat pumps.

In general, a cooling device transports heat (entropy) from a cold end at a low temperature (T c ) to a heat sink at a high temperature T h . In EC coolers, this process is realized by switching the EC material connection between the hot and cold interfaces. Prior methods of thermal contact switching include cycling the flow direction of heat transfer fluid [13], physically moving EC material between cold and hold interfaces by electrostatic actuation [18,19]and external mechanical actuation [14,15,17,21]. All of these approaches require additional systems to independently control the thermal switching, which may increase cost and complexity.
This paper proposes a self-switched EC heat pump, which exploits the synergy of the large EC and electrostrictive properties of relaxor P(VDF-TrFE-CFE-FA) (VDF: vinylidene fluoride, TrFE: trifluoroethylene, CFE: chlorofluoroethylene, FA: fluorinated alkynes) tetrapolymers. Applying an electric field to a relaxor polymer generates an EC response [5,22]. If the same polymer under the same field also generates large electroactuation, cyclic thermal contact between the EC material and hot and cold interfaces can be achieved automatically.
An ideal material for this concept should possess large electrostriction and ECEs and must have long-term reliability while undergoing large deformations. Polyvinylidene fluoride (PVDF) based ferroelectric relaxors exhibit both large ECE (∆T > 15 K) [23] and electrostriction (S > 5%) [24,25], meeting these requirement. One challenge in using the terpolymer ferroelectric relaxors, i.e. P(VDF-TrFE-CFE) and P(VDF-TrFE-CTFE) (CTFE: Chlorotrifluoroethylene) in practical EC heat pumps is the high electric field strength needed for inducing the large ECE. Recently, a new class of polymer ferroelectric relaxor P(VDF-TrFE-CFE-FA) was reported that achieves both giant ECE and electromechanical (EM) performances at low electric fields [8,26]. Both the giant ECE and EM responses in tetrapolymer relaxors are understood to be induced by the conformational change of polymer chains under electric fields. This suggests that both of these effects can be maximized at similar compositions [8,26]. This motivates the use of P(VDF-TrFE-CFE-FA) in the present investigation of the self-switched EC heat pump concept.
A unimorph actuator device architecture was adopted. With the large electroactuation of the tetrapolymers, a small strip coupon (2 cm × 0.7 cm) displaces by >1 mm when switched between the heating and cooling phases at 50 MV m −1 electric field [27]. In general, electrostrictively actuated unimorphs are constructed by bonding active and inactive layers. Large electroactuation in the active layer perpendicular to the applied field direction (transverse strain S 1 ), which is constrained by the inactive layer, causes bending and large displacement perpendicular to the film thickness. For EC cooling devices, the inactive layer would be an additional thermal load. If it is not EC active, the inactive layer would reduce heat pump performance. Therefore materials should be adopted that offer high EC effect, but selective electroactuation.
Transverse strain S 1 in the stretched P(VDF-TrFE-CFE) relaxor polymer films (along the uniaxial stretching direction) can be much greater than that in non-stretched films. As will be shown in this paper, this is also the case for the P(VDF-TrFE-CFE-FA) tetrapolymers. Moreover, tetrapolymers derived from the initial P(VDF-TrFE-CFE) 63/29.7/7.3 terpolymer composition exhibit similar large EC response in both non-stretched and uniaxially stretched films. Thus, stretched and unstretched P(VDF-TrFE-CFE-FA) films are promising candidates for the EC unimorph heat pump (EUHP) layers. This paper will present the EC performances and transverse strain (S 1 ) of stretched and unstretched tetrapolymers, as well as the cooling and actuation performance of a derived EUHP.

Materials preparation
P(VDF-TrFE-CFE) 63/29.7/7.3 terpolymer was provided by Piezo Technologies, France. FA is induced by an elimination reaction as the same method as indicated in [26]. The FA content was determined by proton nuclear magnetic resonance (1 H NMR) with acetonitrile-d3 as the solvent. In preparation of the unstretched polymer films, 0.4 g polymers were dissolved in 10 ml dimethylformamide (DMF) and stirred overnight at room temperature. The homogeneous solution was then deposited onto a 50 mm × 75 mm clean glass substrate in an oven for 12 h at 60 • C. After drying, the films were placed in deionized (DI) water to be peeled off from the substrate. Afterward, the polymer films were placed in a vacuum oven and annealed at 120 • C for 24 h, followed by a natural cooling in the oven to room temperature. Then the unstretched polymer films were ready for characterization. The thickness of unstretched (US) films were in the range from of 27 µm to 35 µm.
To prepare stretched samples, a higher concentration of polymer solution was prepared (2 g polymer dissolved in 10 ml DMF). The other processing steps were the same for the unstretched films before annealing. After peeling from the substrate, the polymers films were stretched to 7× initial length through uniaxial zone-stretching. After stretching, the polymer films were annealed at 120 • C for 24 h. Afterward, the films were ready to use. The thickness of stretched films were similar to that of unstretched films. Gold electrodes were sputtered on the polymer films with an EMITECH K550X system, with nominal thickness of 30 nm.

Property and performance characterizations
The total EC heat flow Q to/from the polymer films when applying/removing electric fields was measured with a heat flux sensor. This yields the isothermal entropy change ∆S of the EC polymers (Q = T∆S). The details of this method were reported in [9].
A thermal imaging infrared (IR) camera (Infratec ® ImageIR 8300) was used to directly measure surface temperature changes of the EUHP device during electric field cycles. A thin layer of carbon-powder-loaded black paint was applied to cover the gold surface to reduce reflections and increase emissivity for thermal imaging measurements.
Transverse strain S 1 was measured using a photonic sensor (MTI 2000 FOTONIC SENSOR) connected to a Polarization Loop & Dielectric Breakdown Test System (PolyK Technologies). A 1 Hz triangle wave was applied to the polymer films, and the strain (along the stretching direction for uniaxially stretched films) was recorded. The displacement of the EUHP was measured with a photonic sensor (Polytec OFV 3001S). The electric field was applied via a function generator at 1 Hz sine wave amplified through a Trek (601D) voltage supply.
Polarization   figure 1(c). It can also be seen that the stretched tetrapolymer films exhibit slightly higher ∆S compared with unstretched films, especially at high electric field such as 70 MV m −1 . It has been suggested that the mechanical stretching can facilitate the inclusion of FA in the crystalline phase which makes them more effective in yielding the ferroelectric related responses [26].

ECE of the P(VDF-TrFE-CFE-FA) tetrapolymers
The polarization-temperature curves under unipolar electric field of 50 MV m −1 for the tetrapolymers with different FA concentrations are presented in figure 1(d). This reveals that the polarization of the tetrapolymer peaks at the terpolymers with 1.9 mol% FA-the same composition of the ECE peak. The peak position of polarization also suggests that 1.9 mol% FA tetrapolymer locates at a composition that close to the end critical point, which is considered as an efficient approach to enhance EM and EC performances of ferroelectric relaxors [28][29][30][31]. ∆S and P data were used to determine the phenomenological β coefficient, ∆S = 1 /2 β P 2 (figure 1(e)). The β coefficient of all the tetrapolymers in the figure is higher than that of the terpolymer. At the EC peak composition (1.9 mol% FA), the β value is more than 2.5 time of that of the terpolymer. This indicates that the polarization processes in the tetrapolymers are much more effective in generating ECE compared with the terpolymer.

Transverse strain S 1 of the tetrapolymers
As discussed in the section 1, the two polymer layers in an ideal EUHP should possess similar and large ECEs and significantly different EM responses. S 1 is the key parameter governing unimorph actuation, and should be substantially different in the two layers to achieve a large displacement. The transverse strains S 1 were characterized for the tetrapolymers and results are presented in figures 2(a) and (b). Analogous to the ECE and polarization data presented in figure 1, the transverse strains also peak at 1.9% FA concentration. This suggests that the enhancements of both EM and ECE due to FA addition originate from similar mechanisms. The stretched tetrapolymer with 1.9% FA generated a 2% S 1 under 50 MV m −1 . In contrast, the unstretched films had much lower S 1 . The tetrapolymer with 1.9% FA generated 1% S 1 at the same field-half of that of the stretched film. This large difference in S 1 between the stretched and unstretched films suggests that the tetrapolymers with 1.9% FA are ideal for the proposed EUHPs.

Electro-actuation of EUHP
The stretched and unstretched tetrapolymers with 1.9% FA were selected as the base materials for the EUHP. The assembly is illustrated in figure 2(c) and a photograph of the physical specimen is presented in figure 2(d). The thicknesses of epoxy glue layer (<1 µm) and electrodes (∼30 nm) are much smaller than the those of the tetrapolymer films (∼30 µm for each of the two tetrapolymer films). The glue and electrode layers can thus be assumed to have negligible effects on actuation. The measurement of actuation displacement of EUHP is illustrated in figure 2(e), the two ends of the EUHP were fixed at x = ±L/2, where L is 2.0 cm. When an electric field is applied on the device, the actuator moves upward through the mechanism described in [27]. The maximum displacement Z max occurs at the center of the strip (x = 0), which can be estimated from [27] Z max = a L 4 16 .
Here, a is a parameter determined by the effective strain S e of the plate under electric field, and L is the distance between the two fixed ends. For a given electric field, the strain of active layer (S-tetrapolymer) is S a , and the strain of less-active layer (US-tetrapolymer) is S L . The effective strain S e can be calculated from a modified form of the classic equation of general unimorph [27], Here, k = E L δ L /E a δ a , where E L and E a are the Young's moduli of less-active layer and active layer, respectively. δ L and δ a are the thicknesses of the less-active layer and active layer, respectively. In the fabricated device, δ L ≈ δ a , yielding k ≈ E L /E a = 220 MPa/530 MPa = 0.415. S a and S L of the tetrapolymer films were shown in figures 2(a) and (b), and can be used to determine S e as a function of electric field. Using S e , the coefficient a in equation (1) can be determined as described in [27]. The theoretically predicted Z max is compared with experimental measurements in figure 2(f). The good agreement between predictions and experimental data supports the proposed mechanism of the EUHP actuators. At 50 MV m −1 , the EUHP actuator can switch the device more than 1.5 mm.

EC performance of EUHP
To assess the EC performance of the EUHP, a high-speed thermal imaging IR camera was used to measure the temperature change as the device was switched [18]. As schematically illustrated in figure 3(a), in the IR imaging of the EUHP heat pumping process, the EUHP was hung on two fixed ends and was slightly tilted toward the IR camera to increase the EUHP area facing the IR camera. Eventually, the thickness of EUHP seen in the IR camera is about 3.2 mm, while the actual thickness of the EUHP is ca. 60 µm. Figure 3(b) presents thermal images of the EUHP as electric fields were applied and removed. Upon applying electric field, the EUHP moves upward and the temperature of the EC film increases. When the field is removed, the EUHP returns to its initial position and cools ( figure 3(b)). This displacement of the EUHP can be observed in the IR images. For instance, at 53 MV m −1 , the displacement between the positions of the EUHP with the field ON and OFF is greater than 1 mm. The temperature changes are summarized in figure 3(c). The asymmetry of EC heating and cooling ∆T is due to the P 2 dependence of ∆T of the EC polymer. As shown in supporting materials (figure S1), the ∆T measured under a step E field in the heating cycle will be smaller than that in the cooling cycle in short time after the change of E, which is the case in this study. The predicted temperature drops ∆T for S-tetrapolymer and US-tetrapolymer can be calculated from the measured ∆S data (figure 1) as ∆T = T∆S/C E . Here, C E is the material specific heat capacity: 1404 J kg −1 K −1 at room temperature. The material-level ∆Ts for the S-tetrapolymer and US-tetrapolymers are ∼8.8 K at 50 MV m −1 . The measured ∆T for the full EUHP is ∼6.5 K at 53.3 MV m −1 . The difference may be due to heat dissipation to the surroundings and the added thermal mass from the glue and carbon-loaded paint layers. The EUHP still exhibits great EC performance considering the low electric field.

Conclusion
This manuscript proposed and demonstrated a self-switched EUHP module, which exploits the multi-functional performance of relaxor tetrapolymer and can be used in EC cooling devices. The proposed EC module could be self-actuated to toggle contact between hot and cold thermal interfaces, as needed for heat pumping. The physical prototype achieved a high ∆T ∼ 6.5 K at moderate field strength (53.3 MV m −1 ). As this device does not require additional controls for actuation, a stacked arrangement of EUHP modules could easily be assembled to amplify overall temperature lift. The tetrapolymer P(VDF-TrFE-CFE-FA) has been demonstrated as a suitable EC material for such EUHP applications, owning to its outstanding EC and EM characteristics.

Data availability statement
All data that support the findings of this study are included within the article (and any supplementary files).