Spatially Resolved Imaging of Electrocaloric Effect ... - ACS Publications

Aug 10, 2016 - ... strongly suggests that all future electrocaloric cooling devices should rely on a design favoring heat exchange through the MLC ter...
0 downloads 0 Views 4MB Size
Subscriber access provided by Northern Illinois University

Letter

Spatially Resolved Imaging of Electrocaloric Effect and the Resultant Heat Flux in Multilayer Capacitors Yang Liu, Brahim Dkhil, and Emmanuel Defay ACS Energy Lett., Just Accepted Manuscript • DOI: 10.1021/acsenergylett.6b00232 • Publication Date (Web): 10 Aug 2016 Downloaded from http://pubs.acs.org on August 10, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Energy Letters is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Energy Letters

Spatially Resolved Imaging of Electrocaloric Effect and the Resultant Heat Flux in Multilayer Capacitors Yang Liu1,2,*, Brahim Dkhil1, and Emmanuel Defay2 1

Laboratoire Structures, Propriétés et Modélisation des Solides (SPMS), CentraleSupélec, CNRS UMR8580, Université Paris-Saclay, 92290 Châtenay-Malabry, France 2

Materials Research & Technology Department, Luxembourg Institute of Science and Technology (LIST), 41 Rue du Brill, L-4422 Belvaux, Luxembourg

AUTHOR INFORMATION Corresponding Author *Email:[email protected] (Y.L.)

1 ACS Paragon Plus Environment

ACS Energy Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ABSTRACT Multilayer capacitors (MLCs) are one of the most studied electrocaloric (EC) systems since they are regarded as the most promising configuration for developing electrocaloric prototype devices. Indeed, they combine 1) thin films prone to exhibit the socalled giant EC effect, 2) they are made of several hundreds of active layers and 3) their internal interdigitated structure alternating highly thermally conductive electrodes with EC layers should dramatically increase its propensity to extract heat. Although the latter has been foreseen by modelling, it has never been observed experimentally. Here we use infra-red camera to observe for the first time the heat flux crossing standard EC MLCs while operating. Interestingly, we observed that heat flux perpendicular to the terminal-terminal direction accounts only for 20 % of the total heat flux. This observation strongly suggests that all future electrocaloric cooling devices should rely on a design favouring heat exchange through the MLCs terminals.

TOC GRAPHICS

2 ACS Paragon Plus Environment

Page 2 of 28

Page 3 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Energy Letters

Electrocaloric effect (ECE) is defined as the adiabatic temperature change ∆T or isothermal entropy change ∆S of a material upon application/withdrawal of an external electric field ∆E .1-2 Electrocaloric refrigeration based on ECE is regarded as a solid-state alternative

to

conventional

vapor-compression

cooling

technology.3-7

Interestingly,

electrocaloric refrigerator prototypes based on multilayer capacitors (MLCs) (see the typical structure as shown in Figure 1a) were recently designed attesting the potential of MLCs for cooling applications.8-11 The promise of using MLCs is that heat flow can be conducted by the inner interdigitated electrodes.12,13 Previous works assumed a homogeneous distribution of temperature change in the electrocaloric active ceramic regions and the behaviour of electrocaloric heat flux between the centre of ceramic regions and the metallic terminals was only foreseen by modelling.12,14 Direct experiments are still lacking whereas understanding the heat flow behaviour in specific caloric prototype devices is crucial to optimize refrigeration efficiency.6 For instance, in active magnetic regeneration the control of heat flux direction has been shown to be very useful to enhance the operating frequency and thus the power density of the device.15 Interestingly, infra-red (IR) imaging can nowadays enable a direct determination of caloric effects both temporally and spatially.16-24 For instance, previous studies have demonstrated the ability of this tool to measure real time caloric responses with good precision, which was confirmed by simultaneously monitoring using a thermocouple17,19,21 or by comparison with differential scanning calorimetry (DSC).20 Non-contact IR camera is compact and fast compared to other techniques i.e., DSC, and its use is rather simple. More importantly, IR camera can directly image spatial caloric response, which is the main advantage over other techniques such as DSC, thermocouple, and other specifically designed calorimeters. As a result, spatially resolved measurements have been conducted to study the local inhomogeneity of elastocaloric and magnetocaloric responses relevant to the microstructure nucleation dynamics of phase transition,17,23 and correlated to local magnetic 3 ACS Paragon Plus Environment

ACS Energy Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

field strength.16 However, spatial imaging of ECE is rarely reported. ECE is usually characterized in literature by isotropic ∆T with homogeneous distribution along samples except in very few studies.21,25,26 For instance, anisotropy of electrocaloric cooling (~0.200.25 K) due to strong anisotropic polar direction in uniaxial relaxor Sr0.75Ba0.25Nb2O6 single crystals,25 and depending on crystal orientations in Pb(Mg1/3Nb2/3)O3-PbTiO3 single crystals26 were recently reported. In addition, extrinsic variation of ECE (~0.2 K temperature difference) due to specific geometry in MLCs was observed between the terminals and ceramic regions in MLCs21 (see MLC structure in Figure 1a). This underlines that spatially distributed ECE should exist over MLCs (see Figure 1b) in contrast to theoretical assumption,12,14 and its quantitative understanding is highly demanded for a better use of MLCs in future devices. In this work, we use an IR camera to directly image the variation of electrocaloric response in Ni-electroded commercial MLCs made of a proprietary BaTiO3-based dielectric (Multicomp). We find the existence of a significant variation of ECE up to 0.285 K between the central region and regions near the terminals. This variation corresponds to about 50% of the average value in the sample, which is larger than the previous ones on anisotropy/inhomogeneity of electrocaloric cooling.21,25,26 Direct experimental evidence allows us to qualitatively and quantitatively analyse the dynamic electrocaloric flux which is found to be mainly transferring unidirectionally and inhomogeneously from the central ceramic regions to the terminals. We reveal the existence of measurable changes in ECE from region to region including the central zones. Our findings offer a promising route to detect ECE locally and the resultant heat flux in a specific electrocaloric material or device by using IR imaging in order to design and optimize the prototype cooling device efficiency. The MLCs were purchased from Multicomp (MC1210F476Z6R3CT). The structure schematic is shown in Figure 1a. The MLCs consist of around 150 doped BaTiO3 ceramic layers of thickness 11 µm separated by interdigitated Ni inner electrodes of thickness 2.0 µm. 4 ACS Paragon Plus Environment

Page 4 of 28

Page 5 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Energy Letters

Tin is deposited onto the terminals in order to realize electrical connection by thin copper wires between the MLCs and generator (KEITHLEY). For the IR measurements, we used a FLIR X6580SC camera and an arrangement of lenses. The X6580sc camera features digital InSb detector corresponding to broadband spectral sensitivity of 1.5-5.1 µm. Thermal images captured by the camera have an effective spatial resolution of 6.5 µm with a thermal sensitivity of less than 25 mK (20 mK typical). The frame frequency of camera we used is 100 Hz, which is sensitive enough to capture thermal change in electrocaloric active MLCs (see Figure S14, Supporting Information). Moreover, the smallest frame frequency likely exhibits a remarkable variation of electrocaloric response, supporting that at higher frequencies (i.e. 100 Hz) the adiabatic regime can be always ensured. In order to determine the temperature of MLCs, a thermocouple was attached to MLCs using scotch tape. 800-grit and 1200-grit abrasive papers are sequentially used to polish the top surface of MLCs. After that, ethanol is employed to wash and clean the MLCs. Indeed, temperature inhomogeneity due to the inevitable variations in emissivity and photon collections in the terminals has already been evidenced in previous work.21 Due to the rough surface of the metallic terminals, it is difficult to extract any small spatial variations of ECE. In this study, we focus on the flat electrocaloric active ceramic region in MLCs rather than the terminals, avoiding rough surface and poor emissivity with lots of reflections on terminals. The MLC/IR camera set-up is similar to the one used in ref. 21. Moreover, we develop the mode of IR measurements by setting the temperature of the MLC to a specific temperature determined by a Pt-100 thermocouple (TE Connectivity Ref NB-PTCO-155 class F 0.1 with accuracy of +/-100mK regarding the absolute temperature) connected to the centre of the MLC. This can be achieved by manually adjusting the emissivity of different selected regions without necessarily compromising the resolution of the camera. It can even happen that emissivity varies up to 1. Note that this happens exclusively in regions Bj (Figure 1d and Table S1, Supporting Information), meaning extremely close to terminals regions but not on 5 ACS Paragon Plus Environment

ACS Energy Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

terminals themselves. As terminals are made of metal, their emissivity is low (less than 10%) and therefore their reflection is very intense. This high reflection infers an extrinsic increase of the effective emissivity of the ceramic regions close to the terminals, even when polishing has been performed (Table S1, Supporting Information). For instance, for unpolished MLC#1, the average emissivity of A-areas is 0.876 whereas it is as high as 0.944 for B-areas.. Reaching saturation with an effective emissivity of 1 may induce a slight error on the temperature extraction, estimated to be less than 2%. Therefore, we may quantitatively image spatial distributions of ECE and gain qualitative information about how the electrocaloric heat conducted between electrocaloric ceramic layers and terminals, which is obviously of great importance for the design of electrocaloric prototype based on MLCs.8-11 A sketch of the IR camera and MLC structure is presented in Figure 1a. We choose three different fresh MLCs: namely MLC #1 with its both sides (parallel to the electrode plates) unpolished, MLC #2 and #3 with one side polished and the other side unpolished. In particular, we only measure the polished side of MLC #3. The IR camera with black-body calibration by the provider is deliberately set to focus on the side (parallel to the electrode plates as shown in Figure 1a). In this regard, there is one top ceramic layer which is electrocaloric inactive. Figure 1c depicts the typical thermal images in MLC #3 (polished) with no applied voltage. In order to locally display electrocaloric properties in MLC #3, 15 different regions are selected with regions Ai (i=1-9) away from the terminals (Figure 1c) and Bj (j=1-6) close to the terminals (Figure 1d). The area is about 50-90 times as large as that selected on the terminals in the previous work,21 which is far above the spatial resolution (~6.5 µm × 6.5 µm) of the IR camera. In order to ensure that the sample as a whole should always show the same temperature without any applied voltages, we manually adjust the emissivity of different regions to overcome the spatial distribution of temperature profile (see Table S1, Supporting 6 ACS Paragon Plus Environment

Page 6 of 28

Page 7 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Energy Letters

Information). As a result, the temperature of different regions can now reach a fixed temperature which is the same as that measured by the thermocouple (also see an example in MLC #1 as shown in Figure S1 and S2, Supporting Information). The electrocaloric temperature change is obtained by counting the photons emitted by the sample. More precisely, the number of emitted photons is interpreted as a variation of temperature according to the surface emissivity of the sample, which has been calibrated thanks to the procedure with the thermocouple. Therefore it is a variation of photons number that we actually use to extract the variation of temperature when we run the experiments.21 The change in photons correlated with local emissivity then represents the real temperature change, which is managed by the software thanks to the provider’s calibration. Typical electrocaloric responses in these 15 regions are shown in Figure 1e and 1f. It can be seen that the magnitude of |∆T | near the terminals (Bj, j=1-6) in Figure 1f is smaller than those which are not close to the terminals (Ai, i=1-9) in Figure 1e. This is in line with previous results reporting a decrease (~0.2 K) in the temperature change on moving from the MLCs as a whole to the terminals.21 In addition, the contribution of Joule heating due to leakage current can be negligible since the IR signal returns to the baseline temperature after applying/removing the voltage (see Figure 1e and 1f), in agreement with previous work.13 As will be shown later, it does not even compromise under numerous cycles of high fields of about 180 kV/cm. Figure 2a-2b show typical results on spatial determination of electrocaloric peak values under different voltages. In this work, we present the ECE in terms of the maximum temperature drop when applied voltage is removed in one cycle (data correspond to application of voltage can be found in Figure S23 and S23, Supporting Information). Interestingly, the non-uniformity of ECE is directly evidenced with a significant distribution in Figure 2a: the closer to the central regions (A4, A5 and A6), and the larger the magnitude of |∆T | will be. The largest variation is 0.285 K between regions A4 (only very slightly larger

7 ACS Paragon Plus Environment

ACS Energy Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

than A5) and B5 when a voltage of 200 V is removed, which exceeds all previous results on electrocaloric temperature difference (~0.20-0.25 K).21,25,26 This temperature variation corresponds to 51% of the ECE (0.57 K) averaged over 15 different regions (A1-9 and B1-6). Similar variation (0.20 K) in MLC #1 (unpolished) is found with a change of 34% of the total ECE (0.58 K) as shown in Figure S5, Supporting Information. The significant spatial distribution of ECE is always found in all MLCs regardless of whether the surface is polished or not (see Figure S5 in MLC #1 and Figure S11 and S15 in MLC #2 in Supporting Information). In other words, the temperature change after activating the ECE is not the same throughout one given MLC. Moreover, Figure 2a shows that remarkably inhomogeneous distribution along x direction (perpendicular to the terminal) is observed to be more significant than that in y direction (along the terminal) as shown in Figure 2b. Moreover, it is shown in Figure 2c that the largest difference between |∆Tx3 | and |∆Tx5 | is about 0.216 K, which is nearly three times as large as that (0.08 K) between |∆Ty1 | and |∆Ty3 | as shown in Figure 2d. The existence of |∆T | gradient provides direct evidence of heat transfer between the central region of MLCs which is insulated and the region with the terminals which is not insulated (see sketch in Figure 1b). Therefore, our results explicitly indicate that the electrocaloric heat (negative here) is mainly drawn from the terminals towards the ceramic regions when the voltage is removed (Figure 1b, bottom panel). In the case of application of voltage, the heat (positive) transfer is inversed (see Figure 1b, middle panel, Figure S22 and S23, Supporting Information). Besides, we carried out measurements under cyclic variation of applied voltage and also under different voltage frequency f (see Figure S19, Supporting Information). An illustrative example of recorded ECE upon 10 successive voltage cycles between 0 and 200 V is summarized in Figure 3a-3c. Reproducibility is confirmed in every region notably for high frequency pulses (see Figure S20, Supporting Information). This is due to the fact that the

8 ACS Paragon Plus Environment

Page 8 of 28

Page 9 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Energy Letters

impact arising from the thermal surroundings can be minimized by high frequency pulses. For instance, the fluctuation in ∆T defined as δ Tmax is found to decrease (the magnitude) from ± 0.030 K at f=0.02 Hz to ± 0.010 K at f=0.1 Hz for the region A3 with the average ∆T slightly increasing from -0.568 K to -0.563 K (see Table S2, Supporting Information) and this change (5 mK) is far below the camera resolution of 25 mK. Moreover, almost all δ Tmax at f=0.1 Hz are smaller than the resolution of IR camera, and the spatial distribution of ECE aforementioned at f=0.02 Hz can be well reproduced in the case of f=0.1 Hz. This strongly supports the reliability of our data, and the results shown in Figure 1-3 can therefore represent the electrocaloric properties of MLC #3. In addition, previous work carried out direct measurement on MLCs using a thermocouple and reported three distinct types of electrocaloric response depending on frequency.27 At a high frequency (>>1 Hz), it was shown that MLCs become hotter due to limited thermal conductivity of MLCs, which cannot afford such high frequency of heat transfer. Typical electrocaloric response like here can only be observed under relatively low frequency (i.e. 0.017 Hz).27 However, the frequency dependence of ECE in the intermediate range (i.e. from 0.017 Hz to 1 Hz) was not reported in this previous study. This frequency range is actually located within the operational window as suggested in Ref. 26 and thus its frequency dependence is of importance for design of electrocaloric devices. Also recalling the controversial frequency dependence of ECE in P(VDF-TrFE-CFE) terpolymer by IR imaging,20,24 the temperature change in MLCs is found to be independent of voltage frequency in the range of 0.02-0.1Hz (see Figure 3a-3c and Figure S19-S21 and S23, Supporting Information). Moreover, our findings indicate that IR camera is a powerful tool to spatially map the dynamic performances of electrocaloric prototypes such as field waveform,27 pulse duration,27,28 frequency11,20,23,24,27,28 and electrocaloric fatigue under cyclic operating. These unexplored data can provide extremely valuable information for design of electrocaloric prototypes in the further studies. 9 ACS Paragon Plus Environment

ACS Energy Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

In order to gain further insight into the electrocaloric heat conducted between the MLCs and the terminals, the average temperature change ∆Tav , the largest average temperature difference in different regions along x direction ∆ Txmax and along y direction ∆Tymax are plotted in Figures 3d-3f, respectively. Figure 3d shows that the average temperature change ∆Tav is nearly identical for these three different MLCs. The largest ∆Tav is around -0.55 K

corresponding to the removal of 200 V. It is also in the same range of temperature changes for other previous results.10,13,21 It is shown that all MLC samples display a significantly larger variation of ECE in x direction than in y direction ( ∆ Txmax >> ∆Tymax as shown in Figure 3e and 3f). In other words, the closer to the terminals, and the smaller magnitude of temperature change is observed. This is due to the large thermal mass of the terminals21 which are electrocaloric inactive and lead to massively diluted electrocaloric response as we observed in this work. Moreover, heat flux mainly follows the path towards the terminals and only dissipates weakly in direction vertical to this path (see Figure 1b). Since the terminals are directly responsible for heat transfer between the MLCs and heat/cold sink in the prototype,811

geometrical optimization is highly desired. For instance, according to a finite element

analysis, optimizing the length of active ceramic layers can shorten the path of electrocaloric heat transfer, which is accompanied by a significant increase in the cooling power and operating frequency.14 In this regard, our findings imply that IR camera is emerging as an ideal technique to precisely image the dynamics of electrocaloric heat flow in MLCs. For instance, the existence of |∆T | gradient not only indicates the main direction of heat flux (Figure 1b) but also implies the heat flux is not homogeneous i.e. weak heat flux exists in the direction of along the terminals (y direction) due to the temperature gradient ∆Tymax . This also demands further studies to directly quantify this inhomogeneous heat flux by solving the heat equations as will be shown later.

10 ACS Paragon Plus Environment

Page 10 of 28

Page 11 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Energy Letters

The aim of polishing the top surface (using thin sandpapers) is to ensure that surfaceroughness-induced variations in emissivity and the resultant thermal inhomogeneity are as small as possible. Although polishing of top layer can make the surface emissivity more uniform (see comparison of thermal images in two sides of MLC #2 in Figure S9 and S13, Supporting Information), no systematic dependence of ECE on the treatment of polishing is found. For instance, electrocaloric properties are found to be exactly the same between the polished side and unpolished side in MLC #2. Interestingly, given that ∆Tav is nearly the same for MLC #1 (unpolished) and MLC #3 (polished), both ∆ Txmax and ∆Tymax shows remarkable different values when the same voltage is withdrawn (see Figure 3e and 3f). Moreover, the largest temperature change along the y direction is changing subtly either located at the side (i.e. |∆Ty1 | is the largest in MLC #1 and MLC #3) or central position (i.e.

|∆Ty2 | in MLC #2). It provides direct experimental evidence demonstrating that heat flow behavior is actually more complex than either simple assumption in theoretical simulations14 or general expectation from common knowledge about thermal conduction. One possible reason may be related to intrinsic microstructural inhomogeneity of MLCs either or extrinsic contributions introduced by the polishing. Nevertheless, provided that this layer is over 200 times thinner than the active electrocaloric MLCs, its contribution to the electrocaloric heat transfer is unlikely to dominate since the same results are observed in MLC #2 regardless of polishing of the inactive layer. Figure 3e and 3f again provide direct experimental evidence to confirm that the electrocaloric heat is mainly transferred in a path along interdigitated Ni inner electrodes away from (towards) the terminals (x direction) and it is weakly dissipated in the direction along the terminals (y direction) in a complex manner. Also note that we cannot simply use the temporal electrocaloric thermal images (Figure S17, Supporting Information) to directly reproduce the heat flux as schematically shown in Figure 1b due to the foregoing problem of emissivity under no voltage or/and inhomogeneous polishing of top layer. In this 11 ACS Paragon Plus Environment

ACS Energy Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

case, the temporal profile also may provide a false thermal image when the electrocaloric effect is triggered. Our preliminary result implies that more careful polishing of top surface is highly desired in further works. Finally, in order to gain deeper insights into dynamic behaviour of the electrocaloric heat flux, we concentrate on the temporal evaluation of ∆T distribution as shown in Figure 4. A significant reduction (over 50%) in the magnitude of ∆T is observed within 2 s in y2-regions after the voltage is removed at 250 s (Figure 4a and its inset). The difference in ∆T close to the central regions (A2,A5,A8) becomes trivial after 255 s (5 s after the voltage is removed). However, a noticeable difference (>25 mK) in ∆T between the central ceramic regions and those close to the metallic terminals still exists (see comparison between regions B5 and A5 or regions B2 and A5 as shown in Figure 4a). Moreover, the difference in |∆T | between the central regions and terminals is the largest corresponding to the time when the voltage is removed (see Figure 4b) or applied (see Figure S27, Supporting Information) and it then decays exponentially-like with time towards zero (see Figure 4a). The counterpart in ydirection firstly shows a small gradient of about 0.045 K and becomes even smaller with a nearly constant value of about 0.017 K (Figure 4c). This also indicates that heat flux parallel with the terminal-terminal direction should be at least 3 times larger than that perpendicular to the terminal-terminal direction. The typical temporal data again confirm our foregoing analysis on the main direction of electrocaloric heat flux at peak position (see Figure 2 and 3) and our proposed sketch in Figure 1b (and see schematic of the heat flux in three MLCs in Figure S28 for detailed information in Supporting Information). More importantly, our study provides a practical route to extract thermal time constant τ (at which |∆T | reaches 63.2% of the steady-state value relative to its initial value) for heat flow between central ceramic layers and metallic terminal, which is one of the key parameters to obtain the cooling power ( ∝ 1/ τ ) according to recent finite element modelling.14 We estimate this τ constant of about 1 s based 12 ACS Paragon Plus Environment

Page 12 of 28

Page 13 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Energy Letters

on our results in Figure 4b, which is significantly larger than those (0.01-0.2 s) used in theoretical simulations14 and the one (0.17 s) calculated by lumped thermal model12 (see Supporting Information). τ is also affected by the external voltage because of the existence of surroundings as shown in Figure 1e and 1f. The larger measured τ is likely attributed to the enhancement in the total thermal resistance and the large thermal mass of BaTiO3 might play an important role in affecting the thermal resistance and thus the relaxation time (Supporting Information). In addition, a summary on the locally average ∆T along MLCs (see Figure 4d and Figure 4f) well reproduces the foregoing results in Figure 4a-4c, which supports that the typical results in y2-regions and x3-regions can represent the behaviour of the whole MLC #3. Indeed, IR camera with spatially resolved temperature gradient is probably the only technique which can be used to directly determine the electrocaloric heat flux. The dynamic heat flux behaviour in MLC #3 is summarized in Figure 4f, which is obtained through a lumped thermal model12 by taking into account the experimental complications (Supporting Information). It is found in Figure 4f that large heat flux 1, 2, and 3 exist only within ~5 s and it decays dramatically from 738.9 W/m2, 529.3 W/m2, and 168.0 W/m2 after removing the external voltage. Heat flux 1 is 30% larger than heat flux 2, which is well consistent with the temperature distribution profile as shown in Figure. 4b. The difference between heat flux 1 and 2 arises from the different thermal mass on terminals due to the different amount of SnPb we deposited to connect leads. Although the magnitude of heat flux 3 (average value) in xdirection is over three times smaller than those in y-direction (average value), this noticeable heat dissipation should be avoided or reduced as much as possible for practical cooling applications (the inset of Figure 4f). Electrocaloric performances of MLCs such as the magnitude of ECE, thermal conductivity need to be further optimized to design an efficient prototype for practical applications.29 For instance, it was reported that MLCs with thinner electrocaloric layer 13 ACS Paragon Plus Environment

ACS Energy Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

thickness (around 1 µm) can permit high temperature change up to 7.2 K under an ultrahigh electric field of 800 kV/cm at 60 ℃ .30,31 Substitution of the BaTiO3-based ceramics by inorganic relaxors,32 polymers,9 antiferroelectric,33,34 composite materials35-38 or multiferroic39 would be desired. Mechanical stress40-43 and pressure44 may also be useful to tune the electrocaloric properties. Geometrical optimization14 may be useful to improve the cooling powers of devices taking into account Joule heating associated with leakage and the thermal coupling of the device to the environment.45,46 Further evidence of the cooling efficiency47,48 and coefficient of performance (COP)6 is also desired in future studies. IR camera can be a very powerful tool to testify all these possible optimizations. In addition, we considered an MLC barely connected to the holder, except air. Further consideration of heat convection between the air and the MLC is beyond the scope of our article, which requires implementing other experiments. In summary, we have reported direct IR measurements on the uniformity and quality of ECE in MLCs as a function of the amplitude and frequency of voltage pulses at room temperature. Our experimental findings not only evidence the large variation in electrocaloric response across the MLCs for the first time, but also demonstrate the advance of IR imaging in precisely probing the electrocaloric heat flux and local electrocaloric performances. Our results suggest that it is attractive to minimize the terminal mass in the optimized devices. Our work highlights the promise of IR imaging in correlating the ECE and heat flux with structural degrees of freedom and invites future direct measurements by this sensitive technique.

Supporting Information Details of all experimental results on electrocaloric responses for all MLCs together with the calculations details about heat flux. ACKNOWLEDGMENT

14 ACS Paragon Plus Environment

Page 14 of 28

Page 15 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Energy Letters

Y.L. and B.D. acknowledge the China Scholarship Council (CSC) for funding Y.L.'s stay in France and a public grant overseen by the French National Research Agency (ANR) as part of the

“Investissements

d’Avenir”

program

(reference:

ANR-10-LABX-0035,

Labex

NanoSaclay). Y.L. and B.D. are grateful to J. F. Scott for critical reading of this manuscript and fruitful discussions. E.D. acknowledges Fonds National de la Recherche from Luxembourg to support COFERMAT PEARL project. E.D. and B.D. acknowledge Fonds National de la Recherche (FNR) du Luxembourg through the InterMobility project 16/1159210 "MULTICALOR".

Notes The authors declare no competing financial interest. REFERENCES (1)

Mischenko, A. S.; Zhang, Q.; Scott, J. F.; Whatmore, R. W.; Mathur, N. D. Giant

Electrocaloric Effect in Thin-Film PbZr0.95Ti0.05O3. Science 2006, 311, 1270-1271. (2)

Neese, B.; Chu, B. J.; Lu, S. G.; Wang, Y.; Furman, E.; Zhang, Q. M. Large

Electrocaloric Effect in Ferroelectric Polymers Near Room Temperature. Science 2008, 321, 821-823. (3)

Lu, S. G.: Zhang, Q. M. Electrocaloric Materials for Solid-State Refrigeration. Adv.

Mater. 2009, 21, 1983-1987. (4)

Scott, J. F. Electrocaloric Materials. Annu. Rev. Mater. Res. 2011, 41, 229-240.

(5)

Moya, X. ; K.-Narayan, S. ; Mathur, N. D.; Caloric Materials Near Ferroic Phase

Transitions. Nat. Mater. 2014, 13, 439-450. (6)

Kitanovski, A.; Plaznik, U.; Tomc, U.; Poredoš, A. Present and Future Caloric

Refrigeration and Heat-Pump Technologies. Int. J. Refrig. 2015, 57, 288-298.

15 ACS Paragon Plus Environment

ACS Energy Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(7)

Liu, Y.; Scott, J. F.; Dkhil, B. Direct and indirect measurements on electrocaloric

effect: recent developments and perspectives. Applied Physics Review 2016, 3, 031102. DOI: http://dx.doi.org/10.1063/1.4958327 (8)

Epstein, R. I.; Malloy, K. J. Electrocaloric Devices Based on Thin-Film Heat Switches.

J. Appl. Phys. 2009, 106, 064501. (9)

Gu, H. M.; Qian, X. S.; Li, X. Y.; Craven, B.; Zhu,W. Y.; Cheng, A. L.; Yao, S. C.;

Zhang, Q. M. A Chip Scale Electrocaloric Effect Based Cooling Device. Appl. Phys. Lett.

2013, 102, 122904. (10)

Jia,Y. B.; Ju, Y. Sungtaek. A Solid-State Refrigerator Based on the Electrocaloric

Effect. Appl. Phys. Lett. 2012, 100, 242901. (11)

Wang, Y. D.; Smullin,S. J.; Sheridan, M. J.; Wang, Q.; Eldershaw, C.; Schwartz, D. E.

A Heat-Switch-Based Electrocaloric Cooler. Appl. Phys. Lett. 2015, 107, 134103. (12)

Kar-Narayan, S.; Mathur, N. D. Predicted Cooling Powers for Multilayer Capacitors

Based on Various Electrocaloric and Electrode Materials. Appl. Phys. Lett. 2009, 95, 242903. (13)

Kar-Narayan, S.; Mathur, N. D. Direct and Indirect Electrocaloric Measurements

Using Multilayer Capacitors. J. Phys. D: Appl. Phys. 2010, 43, 032002. (14)

Crossley, S.; McGinnigle, J. R.; Kar-Narayan, S.; Mathur, N. D. Finite-element

optimisation of electrocaloric multilayer capacitors. Appl. Phys. Lett. 2014, 104, 082909. (15)

Tomc,U.; Tušek, J.; Kitanovski, A.; Poredoš, A. A New Magnetocaloric Refrigeration

Principle With Solid-State Thermoelectric Thermal Diodes. Applied Thermal Engineering

2013, 58, 1-10. (16)

Christensen, D. V.; Bjørk, R.; Nielsen,K. K.; Bahl,C. R. H.; Smith, A.; Clausen, S.

Spatially Resolved Measurements of the Magnetocaloric Effect and the Local Magnetic Field Using Thermography. J. Appl. Phys. 2010, 108, 063913.

16 ACS Paragon Plus Environment

Page 16 of 28

Page 17 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Energy Letters

(17)

Vives, E.; Burrows, S.; Edwards, R. S.; Dixon, S.; Manosa, L.; Planes, A.; Romero, R.

Temperature Contour Maps At the Strain-Induced Martensitic Transition of A Cu–Zn–Al Shape-Memory Single Crystal. Appl. Phys. Lett. 2011, 98, 011902. (18)

Bechtold, C. ; Chluba,C. ; Lima de Miranda, R.; Quandt, E. High Cyclic Stability of

the Elastocaloric Effect in Sputtered TiNiCu Shape Memory Films. Appl. Phys. Lett. 2012,

101, 091903. (19)

Cui, J.; Wu, Y. M.; Muehlbauer, J.; Hwang,Y. H.; Radermacher, R.; Fackler, S.;

Wuttig, M.; Takeuchi, I. Demonstration of High Efficiency Elastocaloric Cooling with Large

∆T using NiTi wires.Appl. Phys. Lett. 2012, 101, 073904. (20)

Sebald, G.; Seveyrat, L.; Capsal, J.-F.; Cottinet, P.-J.; Guyomar, D. Differential

Scanning Calorimeter and Infrared Imaging For Electrocaloric Characterization of Poly(vinylidene fluoride-trifluoroethylene-chlorofluoroethylene) Terpolymer. Appl. Phys. Lett.

2012, 101, 022907. (21)

Kar-Narayan, S.; Crossley, S.; Moya, X.; Kovacova, V.; Abergel, J.; Bontempi, A.;

Baier, N.; Defay, E.; Mathur, N. D. Direct Electrocaloric Measurements of A Multilayer Capacitor Using Scanning Thermal Microscopy and Infra-Red Imaging. Appl. Phys. Lett.

2013, 102, 032903. (22)

Ossmer, H.; Lambrecht, F.; Gültig, M.; Chluba, C.; Quandt, E.; Kohl, M. Evolution of

Temperature Profiles in TiNi Flms For Elastocaloric Cooling. Acta Mater. 2014, 81, 9-20. (23)

Tušek, J.; Engelbrecht, K.; Mikkelsen, L. P.; Pryds, N. Elastocaloric Effect of Ni-Ti

Wire For Application In A Cooling Device. J. Appl. Phys. 2015, 117, 124901. (24)

Guo, D. Z.; Gao, J. S.; Yu, Y.-J.; Santhanam, S.; Fedder, G. K.; McGaughey, A. J. H.;

Yao, S. C. Electrocaloric Characterization of A Poly(vinylidene fluoride-trifluoroethylenechlorofluoroethylene) Terpolymer By Infrared Imaging. Appl. Phys. Lett. 2015, 105, 031906.

17 ACS Paragon Plus Environment

ACS Energy Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(25)

Goupil, F. L.; Axelsson, A. K.; Dunne, L. J.; Valant, M.; Manos, G.; Lukasiewicz, T.;

Dec, J.; Berenov, A.; Alford, N. M. Anisotropy of the Electrocaloric Effect In Lead-free Relaxor Ferroelectrics. Adv. Energy Mater. 2014, 4, 1301688. (26)

Bai,Y.; Wei, D.; Qiao, L.-J. Control Multiple Electrocaloric Effect Peak In

Pb(Mg1/3Nb2/3)O3-PbTiO3 By Phase Composition and Crystal Orientation. Appl. Phys. Lett.

2015, 107, 192904. (27)

Kwon, B.; Roh, I.-J.; Baek, S.-H.; Kim, S. K.; Kim, J.-S.; Kang, C.-Y. Dynamic

Temperature Response Of Electrocaloric Multilayer Capacitors. Appl. Phys. Lett. 2014, 104, 213902. (28)

Chukka, R.; Vandrangi, S.; Shannigrahi, S.; Chen, L. An Electrocaloric Device

Demonstrator For Solid-State Cooling. EPL (Europhysics Letters) 2013, 103, 47011. (29)

Liu, Y.; Scott, J. F.; Dkhil, B. Some Strategies For Improving Caloric Responses

With Ferroelectrics. APL Materials 2016, 4, 064109. (30)

Y. Bai, G. P. Zheng, S. Q. Shi, Direct Measurement of Giant Electrocaloric Effect In

BaTiO3 Multilayer Thick Film Structure Beyond Theoretical Prediction. Appl. Phys. Lett.

2009, 96, 192902. (31)

Bai,Y.; Zheng, G. P.; Ding, K.; Qiao, L.; Shi, S. Q.; Guo, D. The Giant Electrocaloric

Effect and High Effective Cooling Power Near Room Temperature for BaTiO3 Thick Film. J.

Appl. Phys. 2011, 110, 094103. (32)

Ye, H.-J.; Qian, X.-S.; Jeong, D.-Y.; Zhang, S. J.; Zhou, Y.; Shao, W.-Z.; Zhen, L.;

Zhang, Q. M. Giant Electrocaloric Effect in BaZr0.2Ti0.8O3 Thick Film. Appl. Phys. Lett. 2014,

105, 152908. (33)

Geng, W. P.; Liu, Y.; Meng, X. J.; Bellaiche, L.; Scott, J. F.; Dkhil, B.; Jiang, A. Q.

Giant Negative Electrocaloric Effect in Antiferroelectric La-Doped Pb(ZrTi)O3 Thin Films Near Room Temperature. Adv. Mater. 2015, 27, 3165-3169.

18 ACS Paragon Plus Environment

Page 18 of 28

Page 19 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Energy Letters

(34)

Vats, G.; Kumar, A.; Ortega, N.; Bowen, C. R.; Katiyar, R. S. Giant Pyroelectric

Energy Harvesting and A Negative Electrocaloric Effect in Multilayered Nanostructures

Energy Environ. Sci. 2016, 9, 1335-1345. (35)

Zhang, G. Z.; Li, Q.; Gu, H. M.; Jiang, S. L.; Han, K.; Gadinski, M. R.; Haque, Md A.;

Zhang, Q. M.; Wang, Q. Ferroelectric Polymer Nanocomposites for Room-Temperature Electrocaloric Refrigeration. Adv. Mater. 2015, 25, 1450-1454. (36)

Li, Q.; Zhang, G. Z.; Zhang, X. S.; Jiang, S. L.; Ye, Y. K.; Wang, Q. Relaxor

Ferroelectric Based Electrocaloric Polymer Nanocomposites with A Broad Operating Temperature Range and High Cooling Energy. Adv. Mater. 2015, 27, 2236-2241. (37) Li, Q.; Zhang, G. Z.; Liu, F. H.: Han, K.; Gadinski, M. R.; Xiong, C.; Wang, Q. Solution-Processed Ferroelectric Terpolymer Nanocomposites With High Breakdown Strength and Energy Density Utilizing Boron Nitride Nanosheets. Energy Environ. Sci. 2015,

8, 922-931. (38)

Yang, L.; Qian, X. S.; Koo, C. M.; Hou, Y.; Zhang, T.; Zhou, Y.; Lin, M. R.; Qiu, J.-

H.; Zhang, Q. M. Graphene Enabled Percolative Nanocomposites With Large Electrocaloric Efficient Under Low Electric Fields Over A Broad Temperature Range. Nano Energy 2016,

22, 461-467. (39)

Liu, Y.; Phillips, L. C.; Mattana, R.; Bibes, M.; Barthélémy, A.; Dkhil, B. Large

Reversible Caloric Effect in FeRh Thin Films Via A Dual-Stimulus multicaloric cycle. Nat.

Commun. 2016, 7, 11614. (40)

Lisenkov, S.; Mani, B. K.; Chang, C.-M.; Almand, J.; Ponomareva, I. Multicaloric

effect in ferroelectric PbTiO3 from first principles. Phys. Rev. B 2013, 87, 224101. (41)

Liu, Y.; Wei, J.; Janolin, P.-E.; Infante, I. C.; Kreisel, J.; Lou, X. J.; Dkhil, B.

Prediction of giant elastocaloric strength and stress-mediated electrocaloric effect in BaTiO3 single crystals. Phys. Rev. B 2014, 90, 104107.

19 ACS Paragon Plus Environment

ACS Energy Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(42)

Liu, Y.; Infante, I. C.; Lou, X. J.; Bellaiche, L.; Scott, J. F.; Dkhil, B. Giant Room-

Temperature Elastocaloric Effect in Ferroelectric Ultrathin Films. Adv. Mater. 2014, 26, 6132-6137. (43)

Liu, Y.; Wei, J.; Lou, X. J.; Bellaiche, L.; Scott, J. F.; Dkhil, B. Influence of Epitaxial

Strain on Elastocaloric Effect in Ferroelectric Thin Films. Appl. Phys. Lett. 2015, 106, 032901. (44)

Liu, Y.; Wei, J.; Janolin, P.-E.; Infante, I. C.; Kreisel, J.; Lou, X. J.; Dkhil, B. Giant

room-temperature barocaloric effect and pressure-mediated electrocaloric effect in BaTiO3 single crystal. Appl. Phys. Lett. 2015, 104, 162904. (45)

Quintero, M.; Ghivelder, L.; Gomez-Marlasca, F.; Parisi, F. Decoupling Electrocaloric

Effect From Joule Heating in A Solid State Cooling Device. Appl. Phys. Lett. 2011, 99, 232908. (46)

Quintero, M.; Gaztañaga, P.; Irurzun, I. Intrinsic Leakage and AdsorptionCurrents

Associated With the Electrocaloric Effect in Multilayer Capacitors. Appl. Phys. Lett. 2015,

107, 151901. (47)

Defay, E.; Crossley, S.; Kar-Narayan, S.; Moya, X.; Mathur, N. D. The Electrocaloric

Efficiency of Ceramic and Polymer Films. Adv. Mater. 2013, 25, 3337-3342. (48)

Moya, X.; Defay, E.; Heine,V.; Mathur, N. D. To Cool to Work. Nature Physics 2015,

11, 202-205

20 ACS Paragon Plus Environment

Page 20 of 28

Page 21 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Energy Letters

Figure 1. (a) Schematic representation of measurement setup. (b) Schematic of heat transfer for the MLCs: top panel (no field), middle panel (field applied) and bottom panel (field removed). Solid black lines indicate the boundaries of the terminals. The black arrow indicates increasing temperature while the colored arrows refer to the direction of heat flow. Thermal image of top surface of MLC #3 (polished) at room temperature with no applied voltage, where the ~470 µm-diameter regions numbered A1-A9 and regions numbered B1-B6 (~520 µm × ~200 µm) are specifically marked, respectively in (c) and (d). The variation of shape of the region of interest has no influence on the extracted electrocaloric response (Fig. S1-S8, Supporting Information). Recorded temperatures (e) and (f) corresponding to different regions in (c) and (d) as a function of real time while the voltage pulse is applied/removed as schematically shown in the inset in (e).

21 ACS Paragon Plus Environment

ACS Energy Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 2. Typical electrocaloric temperature change ∆T as a function of voltage (0-200 V): (a) y2-regions (B2, A2, A5, A8, B5) and (b) x3-regions (A4, A5, A6). (c) Average temperature change ∆Tx for the regions (i.e., A1, A2 and A3) with the same x (i.e., x1) values and (d) ∆Ty for the regions with similar y values. The inset cartoons depict the corresponding locations of selected regions and the arrows indicate the directions of heat flow.

22 ACS Paragon Plus Environment

Page 22 of 28

Page 23 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Energy Letters

Figure 3. The local electrocaloric properties averaged by 10 successive voltage cycles between 0 and 200 V under various frequency in MLC #3 (polished): (a) A1-A9; (b) y1-3regions [(A1, A4, A7), (A2, A5, A8) and (A3, A6, A9)] and (c) x2-4-regions [(A1, A2, A3), (A4, A5, A6) and (A7, A8, A9)]. The average temperature change ∆ Tav (d) obtained by averaging all the measured regions; (e) ∆ Txmax indicates the largest difference in ∆Tx along x direction while (f) ∆Tymax indicates the largest difference in ∆Ty along y direction. The cartoons depict the corresponding locations of selected regions. The error bars for (a)-(c) decrease from ± 0.060 K at f=0.02 Hz to less ± 0.020 K at f=0.1 Hz.

23 ACS Paragon Plus Environment

ACS Energy Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 4. (a) Temporal ∆T in y2-regions (B2, A2, A5, A8, B5) after the voltage is removed at 250 s. The dashed line and arrows act as a guide for axes of coordinates; (b) Temporal T difference between region A5 and B2 (B5) after the voltage is removed at 250 s and the inset is an enlarged view of (a) specifically in time range of [250-256 s] in MLC #3 (polished). (c) the temporal ∆T in x3-regions (A4, A5, A6) after the voltage is removed at 250 s; Average temperature change (d) ∆Tx and (e) ∆Ty as a function of real measurement time; (f) Heat flux 1, 2, and 3 along x and y directions corresponding to removal of 200 V as a function of real time. Accordingly, the inset corresponds to a schematic of heat flux 1, 2, and 3 in MLC #3 (polished) after 200 V is removed. The inset cartoons in (a)-(f) depict the corresponding locations of selected regions and the arrows indicate the directions of heat flow.

24 ACS Paragon Plus Environment

Page 24 of 28

Page 25 of 28

(b)

(c)

(e)

T (°C)

(a)

(d)

25.6 25.4 25.2 25.0 24.8 24.6 24.4 24.2 24.0 23.8 23.6

A1 A2 A3

A4 A5 A6

A7 A8 A9 100 V on

200 V on

150 V on

50 V on

25 V on

25 V off

50 V off 100 V off 150 V off

Voltage Pulse Sequence

0

50

100

150

Time (s)

200

200 V off

250

(f) 25.6

T (°C)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42

ACS Energy Letters

ACS Paragon Plus Environment

25.4 25.2 25.0 24.8 24.6 24.4 24.2 24.0 23.8 23.6

100 V on 25 V on

0

50

200 V on

50 V on

25 V off

B1 B2 B3

150 V on

50 V off

100 V off

B4 B5 B6

150 V off 200 V off

100

150

Time (s)

200

250

0.0 ACS Energy Letters

(a) 0.0

-0.1

-0.3

B2

-0.4 -0.5

A5

-0.6

B5

-0.7

0

50

(c)

100

Voltage (V)

-0.1

∆Tx (K)

-0.2

B2 A5

-0.5 -0.6 -0.7

50

-0.4

150

-0.7

200

A4 A5 A6

0

50

(d)

100

Voltage (V)

150

200

0.0 B1+A1+A4+A7+B4 B2+A2+A5+A8+B5 B3+A3+A6+A9+B6

-0.1 -0.2 -0.3 -0.4 -0.5

A4 A5 A6

-0.6

B5 0

-0.3

-0.6

B1+B2+B3 A1+A2+A3 A4+A5+A6 A7+A8+A9 B4+B5+B6

-0.4

A4 A5 A6

-0.1

-0.5

0.0

-0.3

Page 26 of 28

-0.2

∆T (K)

-0.2

∆Ty (K)

∆T (K)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42

(b)

B2 A2 A5 A8 B5

100

Voltage (V)

-0.7

150 ACS Paragon 200Plus Environment 0

50

100

Voltage (V)

150

200

Page 27 of 28

ACS Energy Letters

(c) 0.8

f=0.02 Hz f=0.025 Hz

0.7

0.8

f=0.05 Hz f=0.1 Hz

0.7 0.6

0.5

0.5

|∆T | (K)

0.6 0.4 0.3

0.4 0.3

0.2

0.2

0.1

0.1

0.0

0.0

A1+A4+A7 A2+A5+A8 A3+A6+A9

f=0.05 Hz f=0.1 Hz

f=0.02 Hz f=0.025 Hz

A1+A2+A3 A4+A5+A6 A7+A8+A9

(f)

(e) 0.24

0.09

MLC#1 unpolished MLC#2 unpolished MLC#2 polished MLC#3 polished

0.20 0.16

MLC#1 unpolished MLC#2 unpolished MLC#2 polished MLC#3 polished

0.06

∆Tymax (K)

∆Tav (K)

(d)

(b)

|∆T | (K)

|∆T |(K)

(a)

∆Txmax (K)

1 2 3 4 0.8 f=0.05 Hz f=0.02 Hz f=0.1 Hz f=0.025 Hz 5 0.7 6 7 0.6 8 0.5 9 100.4 11 120.3 130.2 14 150.1 160.0 1 2 3 4 5 6 7 8 9 17 Region Number 18 19 20 21 22 23 0.0 MLC#1 unpolished 24 MLC#2 unpolished 25 -0.1 MLC#2 polished 26 MLC#3 polished -0.2 27 28 -0.3 29 30 -0.4 31 32 -0.5 33 -0.6 34 35 -0.7 36 0 50 100 150 200 37 Voltage (V) 38 39 40 41 42

0.12 0.08

0.03

0.04 0.00

0.00 0

50

100

Voltage (V)

150

ACS Paragon Plus Environment

200

0

50

100

Voltage (V)

150

200

ACS Energy Letters

260

255

250 0.0 -0.1 -0.2 -0.3

B2 A2 A5 A8 B5

255

260

265

Time (s)

-0.4 -0.5 -0.6 270

-0.7 275

(d)

(b)

(c) 0.00

-0.05

0.0

A5-B2 A5-B5

-0.10 -0.15

-0.30 250

251

255

252

253

254

Time (s)

260

265

Time (s)

-0.2

-0.2

-0.3

-0.3

-0.5

270

∆T (K)

∆T (K)

-0.4

Time (s)

-0.7 250

1 2 3

y

400

-0.5

275

600

-0.3

265

Time (s)

270

275

255

260

265

Time (s)

270

ACS Paragon Plus Environment

275

0 250

3 2

1 x

200

B1+A1+A4+A7+B4 B2+A2+A5+A8+B5 B3+A3+A6+A9+B6

-0.6 270

Heat flux (W/m2)

A4 A5 A6

-0.2

265

260

800

-0.1

260

255

(f) 0.0

255

-0.7 250

275

(e)

B1+B2+B3 A1+A2+A3 A4+A5+A6 A7+A8+A9 B4+B5+B6

A4 A5 A6

-0.6

-0.7 256

255

-0.4 -0.5

-0.6

-0.25 250

-0.1

-0.4

B2 A2 A5 A8 B5

-0.20

-0.1

∆T (K)

265

T difference (K)

270

∆T (K)

(a)

∆T (K)

1 2 3 275 4 0.0 5 6 -0.1 7 -0.2 8 9 -0.3 10 11-0.4 12 -0.5 13 14-0.6 15 16-0.7 17 250 18 19 20 21 22 23 0.0 24 25-0.1 26 -0.2 27 28-0.3 29 30-0.4 31 32-0.5 33-0.6 34 35-0.7 36 250 37 38 39 40 41 42

Page 28 of 28

255

260

265

Time (s)

270

275