Solid‐State Electrochemical Thermal Switches with Large Thermal Conductivity Switching Widths

Abstract

Thermal switches that switch the thermal conductivity (κ) of the active layers are attracting increasing attention as thermal management devices. For electrochemical thermal switches, several transition metal oxides (TMOs) are proposed as active layers. After electrochemical redox treatment, the crystal structure of the TMO is modulated, which results in the κ switching. However, the κ switching width is still small (<4 W m⁻¹ K⁻¹). In this study, it demonstrates that LaNiO _x ‐based solid‐state electrochemical thermal switches have a κ switching width of 4.3 W m⁻¹ K⁻¹. Fully oxidized LaNiO₃ (on state) has a κ of 6.0 W m⁻¹ K⁻¹ due to the large contribution of electron thermal conductivity (κ _ele, 3.1 W m⁻¹ K⁻¹). In contrast, reduced LaNiO_2.72 (off state) has a κ of 1.7 W m⁻¹ K⁻¹ because the phonons are scattered by the oxygen vacancies. The LaNiO _x ‐based electrochemical thermal switch is cyclable of κ and the crystalline lattice of LaNiO _x . This electrochemical thermal switch may be a promising platform for next‐generation devices such as thermal displays.

Solid‐state electrochemical thermal switches that exhibit a wide thermal conductivity switching width of 4.3 W m⁻¹K⁻¹ are demonstrated. The oxidized on‐state exhibited 6.0 W m⁻¹K⁻¹, due to a large contribution of electrons, whereas the reduced off‐state exhibited 1.7 W m⁻¹K⁻¹. The present results will boost the development of solid‐state electrochemical thermal switches for thermal management devices.

1. Introduction

Reuse of waste heat resulting from the low conversion rate of primary energy is crucial for sustainable development. Low‐ to medium‐temperature (100–300 °C) waste heat is the most difficult to reuse; the temperature is too low to generate jet steam for power generation. Although thermoelectric energy conversion technology is a solution, its efficiency is too low in this temperature range in the air.^{[ 1} , ² , ³ , ^{4 ]} Thermal management technologies^{[ 5 ]} such as thermal diodes^{[ 6} , ⁷ , ^{8 ]} and thermal switches (or sometimes they are called as thermal transistors)^{[ 9} , ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ , ^{16 ]} have recently attracted attention. Thermal diodes rectify the heat flow; thermal switches electrically switch the heat flow on and off. We expect that thermal displays that visualize heat contrast using infrared cameras can be realized using thermal switches. Thus, thermal switches may be useful for the reuse of waste heat.

For this purpose, electrical control of thermal conductivity (κ) in the active materials of thermal switches is paramount. It necessitates a switch between the on‐state (high κ) and off‐state (low κ). Electrochemical^{[ 10} , ¹¹ , ¹² , ¹³ , ^{14 ]} and electrostatic^{[ 15} , ^{16 ]} approaches offer pathways to govern the κ of active materials. Although electrostatic methods provide rapid κ control, their suitability for thermal display applications is limited by the requirement of an extremely thin active material around the heterointerface between the gate dielectric and the active material. We focus on electrochemical methods because they control the κ of entire materials. Many studies have used ionic liquids such as organic electrolytes and water for electrochemical modulation of materials.^{[ 10} , ¹¹ , ¹² , ^{14 ]} However, this method is incompatible with integrated circuits, limiting its application. In our pursuit of thermal displays with substantial thermal conductance differences between the on‐ and off‐states, we used all‐solid‐state electrochemical thermal switches.^{[ 13} , ¹⁷ , ^{18 ]}

All‐solid‐state electrochemical thermal switches, a cornerstone of advanced thermal management, harness the redox modulation of transition metal oxides (TMOs). Among many candidates,^{[ 9} , ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ , ^{16 ]} TMOs have emerged as promising materials. When subjected to electrochemical redox treatment by inserting and extracting metal ions^{[ 10 ]} and oxide ions,^{[ 12} , ¹³ , ^{14 ]} TMOs undergo structural transformations leading to a switch in their κ width. Despite these advancements, the challenge lies in achieving a substantial κ switching width, a critical factor for practical application that is often limited (<4 W m⁻¹ K⁻¹).^{[ 10} , ¹² , ¹³ , ¹⁴ , ^{19 ]}

To overcome this limitation, we chose LaNiO₃ as the active material for solid‐state electrochemical thermal switches. Bulk LaNiO₃ has comparably high electrical and thermal conductivity,^{[ 20 ]} indicating its potential for κ modulation (Figure S1, Supporting Information). The schematic depicted in Figure  1 introduces LaNiO _x as the active layer, offering a wide κ switching width. In the on state, LaNiO₃ has heightened electrical conductivity when fully oxidized, resulting in a significant contribution from the electron thermal conductivity (κ _ele). Conversely, the off state achieved through electrochemical reduction leads to oxygen vacancies, reduced electrical conductivity, and negligible κ _ele. The scattering of phonons by these vacancies manifests as a low κ.

Figure 1.

Strategy of electrochemical thermal switches with transition metal oxide active layer with large κ switching width. (Left) Diagram of on state. Fully oxidized TMOs show high electrical conductivity. Both electrons and phonons carry heat. The thermal switch shows high κ. (Right) Diagram of off state. Electrochemical reduction treatment produces oxygen vacancies. The reduced TMOs show low electrical conductivity. The κ _ele is negligible. Phonons are scattered by oxygen vacancies. The thermal switch shows low κ.

This study focuses on the use of LaNiO _x ‐based electrochemical thermal switches to address the limitations in κ switching width. Our investigations indicated a κ switching width of 4.3 W m⁻¹ K⁻¹. Fully oxidized LaNiO₃ (on‐state) exhibited a high κ of 6.0 W m⁻¹ K⁻¹, primarily attributed to the contribution of κ _ele (3.1 W m⁻¹ K⁻¹). Conversely, reduced LaNiO_2.72 (off state) had a low κ of 1.7 W m⁻¹ K⁻¹ due to the scattering of phonons by oxygen vacancies. Both the reduction and oxidation processes exhibited a nearly linear change in κ.

Furthermore, our study investigates the cyclability of κ and the crystalline lattice of LaNiO _x ‐based thermal switches. The performance of these electrochemical thermal switches positions them as viable candidates for integration into next‐generation devices, particularly thermal displays.

2. Results and Discussion

2.1. Electrochemical Thermal Switch Fabrication and Operation

Figure  2a shows a schematic of the thermal switch device structure, which is similar to those of our previous thermal switches.^{[ 13} , ¹⁷ , ^{18 ]} In this study, we inserted an extremely thin SrCoO _x layer between the LaNiO₃ and solid electrolyte (Gd‐doped CeO₂/YSZ) (Figure S2, Supporting Information). As shown in Figure S2 (Supporting Information), when LaNiO₃ was grown on the SrCoO _x /GDC‐buffered (001) YSZ, the crystallographic orientation was stronger than that without a buffer layer. X‐ray reciprocal space mapping (RSM) around 113 YSZ diffraction spots (data not shown) confirmed that LaNiO₃ grown on SrCoO _x /GDC‐buffered (001) YSZ demonstrated the highest quality. Moreover, there was a positive correlation between the quality of LaNiO₃ and its thermal conductivity (Table S3, Supporting Information). Thus, the use of the thermal switch structure shown in Figure 2a is crucial.

Figure 2.

LaNiO _x ‐based thermal switch operation. a) Structure of the thermal switch that is composed of six layers: 50‐nm‐thick Pt film, 80‐nm‐thick LaNiO _x film, 2‐nm‐thick SrCoO _x film, 10‐nm‐thick Gd‐doped CeO₂ (GDC) film, 0.5‐mm‐thick (001) YSZ single‐crystal substrate, and 40‐nm‐thick Pt film. b) Schematic of LaNiO _x thermal switch. The switch was placed on a Pt‐coated glass substrate and heated at 280 °C in air. A K‐type thermocouple was used to monitor the switch surface temperature. The switch measured 5 mm × 5 mm. A constant negative current (−10 µA) was applied for reduction; a constant positive current (+10 µA) was applied for oxidation. c,d) Changes in observed DC voltage of thermal switch during reduction from LaNiO₃ to LaNiO_2.72 and oxidation from LaNiO_2.72 to LaNiO₃ with a step of Q = 2 × 10²¹ cm⁻³.

The setup for the operation of the LaNiO _x thermal switch is shown in Figure 2b. Electrochemical redox treatment was performed at 280 °C in the air by applying a constant current of ±10 µA. During the redox reaction, we controlled the flown electron density Q = (I·t)/(e·V) through the current application time (Figure 2c,d), with a step of Q = 2 × 10²¹ cm⁻³ marked as A–K, where I is the flown current, t is the application time, e is the electron charge, and V is the volume of the LaNiO _x layer in the thermal switch. In this study, electrochemical redox treatments were performed according to Faraday's law of electrolysis.

Reduction:

$$\mathrm{LaNi}{\mathrm{O}}_3+0.56e^{-}\to \mathrm{LaNi}{\mathrm{O}}_{2.72}+0.28{\mathrm{O}}^{2-}$$ (1)

Oxidation:

$$\mathrm{LaNi}{\mathrm{O}}_{2.72}+0.28{\mathrm{O}}^{2-}\to \mathrm{LaNi}{\mathrm{O}}_3+0.56e^{-}$$ (2)

The econductivity of the reduced state of LaNiO_2.72 (9 S cm⁻¹) also confirmed its identity as LaNiO_2.72.^{[ 21 ]}

Electrochemical redox treatment was initiated by applying a negative current to reduce LaNiO₃ to LaNiO_2.72 (Figure 2c). As Q increased, the absolute value of the voltage increased. The slope decreased after 2 × 10²¹ cm⁻³ and slightly increased after 2 × 10²¹ cm⁻³. When the current became positive, LaNiO_2.72 was oxidized to LaNiO₃ (Figure 2d). As Q increased, fluctuations occurred at 2 × 10²¹ cm⁻³ and 8 × 10²¹ cm⁻³. Overall, there was still an increasing trend and an absolute value between 3 and 5 V, with minimal variations. The absence of steps in the process curve indicates that there were no new thermodynamically stable phases.

2.2. Crystalline Lattice and κ Changes During Redox Treatment

Redox treatment induced reversible changes in the crystalline lattice of LaNiO _x step by step (reduction A → F, oxidation: F → K), as evidenced by the out‐of‐plane X‐ray diffraction patterns (Figure  3a). Slight shifts in the 002 diffraction peak indicated modulations in the crystal structure, with lattice expansion observed after reduction and shrinkage observed after oxidation. The change in the lattice parameter c is shown in Figure 3b.

Figure 3.

Change in the crystalline lattice of the LaNiO _x layer after redox treatment. a) Change in out‐of‐plane XRD patterns after redox treatment with a step of Q = 2 × 10²¹ cm⁻³. The 002 diffraction peak shifted to a smaller q _z side after reduction and a larger q _z side after oxidation. These shifts were entirely reversible. b) Changes in lattice parameter c of LaNiO _x as a function of Q. A lattice expansion of ≈0.6% occurred after reduction.

Figures  4 and S3 (Supporting Information) illustrate the significant changes in the κ of the LaNiO _x during redox treatment. Time‐domain thermoreflectance (TDTR) decay curves show a decrease in κ from 5.9 to 1.8 W m⁻¹ K⁻¹ after reduction (A → F) and an increase from 1.8 to 5.9 W m⁻¹ K⁻¹ after oxidation (F → K). Both the reduction and oxidation processes exhibited a nearly linear change in κ. The slight variations in lattice constants between the oxidized and reduced states are attributed to the disappearance and generation of oxygen vacancies. Through phonon scattering, oxygen vacancies contribute to a decrease in the lattice thermal conductivity (κ _lat).

Figure 4.

Change in κ of LaNiO _x layer after redox treatment. a) TDTR decay curves of the thermal switch after (upper) reduction and (lower) oxidation treatment. b) Changes in κ of LaNiO _x layer after (left) reduction and (right) oxidation treatment with a step of Q = 2 × 10²¹ cm⁻³. The κ of LaNiO₃ was 5.9 W m⁻¹ K⁻¹; it decreased almost linearly with Q after reduction. After oxidation, the κ increased almost linearly with Q and returned to 5.9 W m⁻¹ K⁻¹.

However, there is a significant contrast in electrical conductivity (σ) between the oxidized (4250 S cm⁻¹) and reduced (9 S cm⁻¹) states (Table S4, Supporting Information). We estimated the κ κ _ele by assuming the Wiedemann–Franz law; κ _ele = L·σ·T, where L is the Lorentz number (2.44 × 10⁻⁸ W Ω K⁻²), and T is the absolute temperature. The κ _ele of the oxidized state reached 3.1 W m⁻¹ K⁻¹. According to the principle that observable κ is the sum of κ _lat and κ _ele,^{[ 22 ]} we estimated the κ _lat of LaNiO₃ (on‐state) to be 2.9 W m⁻¹ K⁻¹. This value aligns with previously reported values for bulk LaNiO₃,^{[ 20 ]} confirming the consistency of our findings. This substantial difference enabled the LaNiO _x thermal switch to modulate its κ over a wide range. Furthermore, the linear and reversible changes in κ indicate the potential of the LaNiO _x ‐based thermal switch for precise thermal modulation.

2.3. Cycle Properties of Thermal Switch

As shown in Figures  5 and S4 (Supporting Information), the LaNiO _x ‐based thermal switch exhibits cycling properties. The on‐state (LaNiO₃) has a higher average κ (6.0 W m⁻¹ K⁻¹) than the off‐state (LaNiO_2.72, 1.7 W m⁻¹ K⁻¹). Seven cycles are shown in the figure. The TDTR decay of LaNiO₃ was faster than that of LaNiO_2.72 (Figure 5a). The TDTR curves for each cycle overlapped significantly, indicating good repeatability.

Figure 5.

Cycle properties of LaNiO _x ‐based thermal switch. a) Changes in TDTR decay curves (seven cycles overlapped). The TDTR decay of the LaNiO₃ layer was faster than that of the LaNiO_2.72 layer. b) Change in κ of LaNiO₃ and LaNiO_2.72 layers after redox cycling. The average κ of the LaNiO₃ (on‐state) layer and LaNiO_2.72 (off‐state) layer were 6.0 and 1.7 W m⁻¹ K⁻¹, respectively. The κ _ele of the LaNiO₃ (on‐state) layer was high (3.1 W m⁻¹ K⁻¹). The κ _ele of the LaNiO_2.72 (off‐state) layer was negligible.

2.4. Comparison with Other TMO Thermal Switches

Figure  6 compares the κ switching widths of different TMOs, indicating the unparalleled performance of LaNiO _x ‐based thermal switches, with a κ switching width of ≈4.3 W m⁻¹ K⁻¹, greater than that of other TMOs. The wide switching range indicates the exceptional versatility of LaNiO _x in modulating its κ. One more important property of thermal switches is their on‐to‐off κ ratio. Currently, La_0.5Sr_0.5CoO _x (2.5 ≦ x ≦ 3) ^{[ 14 ]} based thermal switches show the largest on‐to‐off κ ratio of ≈5.4. It is still required to reduce the off‐state κ of LaNiO_{3− δ} to improve the on‐to‐off κ ratio (≈3.5). We will tackle this issue in future. The inherent ability of LaNiO _x ‐based thermal switches to linearly transition between high κ in the fully oxidized state and low κ in the reduced state positions them at good candidate TMOs for thermal management applications. The nuanced control over κ and the stability demonstrated in the cycling properties (Figure 5) underscore the potential of LaNiO _x ‐based thermal switches for application in next‐generation thermal displays and other advanced thermal management systems.

Figure 6.

Comparison of κ switching widths of several transition metal oxides. LaNiO _x ‐based thermal switches exhibited large κ switching widths (≈4.3 W m⁻¹ K⁻¹). Data for LiCoO₂ ↔ Li_{1− δ} CoO₂ are from Ref. [10] SrCoO₃ ↔ SrCoO_2.5 and SrCoO_2.5 ↔ HSrCoO_2.5 are from Ref. [12] SrCoO₃ ↔ SrCoO₂ are from Ref. [13] WO₃ ↔ H _δ WO₃ are from Ref. [19] and La_0.5Sr_0.5CoO₃ (LSCO₃) ↔ La_0.5Sr_0.5CoO_2.5 (LSCO_2.5) are from Ref. [14].

3. Conclusion

This study presents a breakthrough with LaNiO _x ‐based electrochemical thermal switches, with an exceptional κ switching width of 4.3 W m⁻¹ K⁻¹. The fully oxidized LaNiO₃ (on‐state) produced a κ of 6.0 W m⁻¹ K⁻¹, primarily attributed to a substantial contribution from κ _ele (3.1 W m⁻¹ K⁻¹). In contrast, the reduced LaNiO_2.72 (off‐state) had a low κ of 1.7 W m⁻¹ K⁻¹ due to negligible κ _ele (0.007 W m⁻¹ K⁻¹) and phonon scattering caused by oxygen vacancies. The LaNiO _x ‐based electrochemical thermal switch demonstrated κ cyclability while maintaining the structural integrity of the LaNiO _x crystalline lattice, making it a promising candidate for integration into next‐generation devices, particularly thermal displays.

4. Experimental Section

Fabrication of Thermal Switches

LaNiO₃ films were heteroepitaxially grown on SrCoO _x /10%‐Gd‐doped CeO₂ (GDC)‐buffered (001)‐oriented YSZ substrates using pulsed laser deposition (PLD). First, ≈10 nm‐thick GDC was heteroepitaxially grown on a YSZ (10 mm × 10 mm × 0.5 mm, double‐sided polished, crystal base) substrate at 770 °C in an oxygen atmosphere (10 Pa). Approximately 2‐nm‐thick SrCoO _x was grown on GDC in the same conditions. Focused KrF excimer laser pulses (λ = 248 nm, fluence ≈2 J cm⁻² pulse⁻¹, repetition rate = 10 Hz) were irradiated onto the ceramic target of GDC. Subsequently, an ≈80‐nm‐thick LaNiO₃ film was heteroepitaxially grown on the GDC film at 625 °C in an oxygen atmosphere (25 Pa). The laser fluence was ≈1.6 J cm⁻² pulse⁻¹. After film growth, the sample was cooled to room temperature in a PLD chamber in an oxygen atmosphere (25 Pa). An ≈50‐nm‐thick Pt film was sputtered on the top surface of the LaNiO₃ epitaxial film, followed by an ≈50‐nm‐thick Pt film sputtered on the backside of the YSZ substrate. DC sputtering of Pt was performed at room temperature. The samples were cut into four squares (≈5 mm × 5 mm).

Electrochemical Redox Treatment

Thermal switch (≈5 mm × 5 mm) was placed on a Pt‐coated glass substrate and heated at 280 °C in air. Electrochemical redox treatment was performed by applying a constant current of ±10 µA, after which the sample was immediately cooled to room temperature. During the redox treatment, it controlled the flown electron density Q = (I·t)/(e·V) through the current application time.

Crystallographic Analyses

The crystalline phase, orientation, and lattice parameters of the resultant films were analyzed using high‐resolution X‐ray diffraction (Cu Kα₁, λ = 1.54059 Å, ATX‐G, Rigaku). Out‐of‐plane Bragg diffraction patterns and reciprocal space mappings (RSMs) were measured at room temperature to clarify changes in the crystalline phase of LaNiO _x . The lattice parameters were calculated from the diffraction peaks. The atomic arrangements of the LaNiO₃ films were visualized using a STEM (JEM‐ARM200CF, JEOL) operated at 200 keV.

Measurement of Electrical Properties of LaNiO_x Layers

To measure the electrical conductivity (σ) of the LaNiO _x layers after redox treatment, it mechanically attached Au foil on the film surface while Pt films were deposited only on the backside of the YSZ substrate.^{[ 13 ]} The LaNiO _x films were oxidized and reduced electrochemically at 280 °C in air using the Au foil as the electrode. The σ of the LaNiO₃ (on‐state) and LaNiO_2.72 (off‐state) films was measured using the DC four‐probe method with a van der Pauw electrode configuration at room temperature in air.

Thermal Conductivity Measurements

The κ of the LaNiO _x layers perpendicular to the substrate surface was measured through time‐domain thermoreflectance (TDTR, PicoTR, Netzsch Japan). The top Pt film was used as the transducer. The decay curves of the TDTR signals were simulated to obtain κ. The specific heat capacity of the LaNiO₃ ceramic target was measured by differential scanning calorimetry (DSC, DSC200, Hitachi High‐Tech Co.). The specific heat capacities of the layers used for the TDTR simulation were Pt: 132 J kg⁻¹ K^{−1[ 23 ]}; LaNiO₃: 420 J kg⁻¹ K⁻¹ (measured) (cf. 448 J kg⁻¹ K^{−1[ 24 ]}), and YSZ: 460 J kg⁻¹ K^{−1[ 25 ]}. Details of the TDTR method are described in the previous studies.^{[ 13} , ¹⁷ , ²⁶ , ²⁷ , ^{28 ]} Note that it simulated the TDTR decay curves by assuming the interfacial resistance due to 2‐nm‐thick SrCoO _x and 10‐nm‐thick GDC layers is constant (1 × 10⁻⁹ K W⁻¹). It confirmed this assumption works by measuring the κ of the thicker (151‐nm‐thick) LaNiO₃ active layer. In the treatment of the κ values, as there were uncertainties such as the position of the baseline, position of time zero, and noise in the signal, it used error bars for ±15% of the obtained values.

Conflict of Interest

The authors declare no conflict of interest.

Supporting information

Supporting Information

Bian Z., Yoshimura M., Jeong A., Li H., Endo T., Matsuo Y., Magari Y., Tanaka H., Ohta H., Solid‐State Electrochemical Thermal Switches with Large Thermal Conductivity Switching Widths. Adv. Sci. 2024, 11, 2401331. 10.1002/advs.202401331

Data Availability Statement

The data that support the findings of this study are available in the supplementary material of this article.