View PDF Version

 

Sulfur-based thermally regenerative cascade batteries enabling dual electrochemical reactions for enhanced power generation

Yu Shi† ^ab, Dong Li†^ab, Yichao An^ab, Shuai Tang^ab, Liang Zhang*^ab, Jun Li^ab, Xun Zhu^ab and Qiang Liao^ab
aKey Laboratory of Low-grade Energy Utilization Technologies and Systems, Chongqing University, Ministry of Education, Chongqing, 400030, China. E-mail: liangzhang@cqu.edu.cn
^bInstitute of Engineering Thermophysics, School of Energy and Power Engineering, Chongqing University, Chongqing, 400030, China

First published on 5th August 2025

Traditional TRBs face challenges in achieving long-term cycling stability. This limitation primarily stems from the dual role of the Cu electrode, which not only participates in electrochemical reactions as an active material but also functions as a current collector. The Cu anode suffers from poor anodic Coulombic efficiency due to side reactions (Cu(NH₃)₄²⁺ + e⁻ ↔ Cu(NH₃)₂⁺) and corrosion, ultimately resulting in significant capacity degradation.⁷ To address these issues, a sulfur-based TRB (SRB) system utilizing S/Cu₂S electrodes was recently proposed.⁸ Owing to its more favorable anode potential, the S/Cu₂S system avoids the potential window of the side reaction, thereby enhancing electrode reversibility and improving cycling stability. While SRBs exhibit promising performance, their capacity remains closely dependent on sulfur loading. Although increasing sulfur content is a straightforward strategy to enhance capacity, excessive loading, especially under high current densities, leads to performance degradation due to increased electrode thickness, prolonged ion diffusion paths, reduced mass transfer efficiency and elevated internal resistance.^9,10 These factors collectively impair rate capability and long-term cycling stability.¹¹

In this work, we propose a thermally regenerative cascade battery (TRCB) that incorporates dual electrochemical reactions within a single electrode to overcome the mentioned limitations. As illustrated in Fig. 1a, the TRCB sequentially involved the Cu₂S/S redox reaction and the Cu²⁺/Cu redox reaction. The product Cu₂S can act catalytically to facilitate the Cu²⁺/Cu redox process, thereby introducing additional capacity at relatively low cost. Moreover, ethylenediamine (en) is employed as the ligand in the anolyte, significantly improving the cycling stability of the second electrochemical stage. Compared to conventional TRBs and SRBs, this cascade reaction strategy maximizes the utilization of active species and expands the reversible capacity of the electrode, offering a new pathway for efficient and durable thermoelectric energy conversion.

	Fig. 1 (a) Diagrammatic sketch of a TRCB coupling dual electrochemical reactions. (b) Potential during anodic reaction and cathodic reaction, (c) and the corresponding XRD results at five points during anodic reaction and cathodic reaction. (d) Dual electrochemical conversion stages of the anode and cathode.

To investigate the electrochemical reactions occurring on the electrode, a S-loaded carbon cloth (S@CC) electrode was subjected to chronopotentiometry (CP), as illustrated in Fig. 1b. During the electrochemical conversion of the cathode, two distinct potential plateaus (from points 1 to 3) were observed, indicating two successive electrochemical processes. Similarly, two potential plateaus (points 3, 4 and 5) were identified during the anodic conversion. To elucidate the phase evolution at each stage, samples collected at points 1 to 5 were analyzed via X-ray diffraction (XRD) and scanning electron microscopy (SEM) (Fig. 1c and Fig. S1). The XRD pattern at point 1 confirmed that the initial state of the electrode was predominantly elemental S. From stage 1 to 2, the disappearance of S diffraction peaks and the emergence of diffraction peaks indicated the formation of Cu₂S via a reducing reaction between S and Cu²⁺.^12,13 At the end of the cathodic process (point 3), the diffraction peaks corresponding to Cu₂S remained, while metallic Cu diffraction peaks appeared. The SEM and EDS mapping results confirmed the uniform deposition of Cu on the electrode surface (Fig. S2). The electrode obtained after point 3 was subsequently transferred to the anolyte for anodic conversion. At point 4, the XRD and SEM results closely resembled those of point 2, suggesting a similar composition and morphology, primarily consisting of Cu₂S. Moreover, the analysis of the electrode at point 5 revealed diffraction peaks and morphologies similar to those at point 1, confirming the reduction to elemental S. Therefore, S was first reduced to Cu₂S, followed by the electrodeposition of Cu²⁺ on the surface during cathodic reaction, while during anodic discharge, the Cu was first stripped from the electrode, and then the exposed Cu₂S was oxidized back to S (Fig. 1(d)).

Importantly, digital images (Fig. S3) further revealed that copper predominantly deposits in regions containing S/Cu₂S. This suggests that the intermediate Cu₂S formed during the first step of the cathodic reaction may act as the catalyst in the second step to accelerate Cu electrodeposition. To verify this hypothesis, linear sweep voltammetry (LSV) was conducted using Cu₂S-loaded carbon cloth (Cu₂S@CC) and bare carbon cloth (CC) electrodes in the catholyte from 0 to −0.3 V vs. Cu²⁺/Cu (0.34 V vs. SHE). The Cu₂S@CC electrode exhibited a significantly higher current density, as shown in Fig. S4a. The Tafel slope derived from the LSV results further demonstrates the improved electrochemical kinetics. Specifically, the Cu₂S@CC electrode showed a Tafel slope of 147 mV dec⁻¹, which is markedly lower than that of the bare CC electrode (230 mV dec⁻¹), indicating enhanced reaction kinetics for Cu deposition (Fig. S4b). Electrochemical impedance spectroscopy (EIS) further confirmed that Cu₂S@CC showed significantly lower charge-transfer resistance compared to bare CC (Fig. S4c). Moreover, potentiostatic deposition at −200 mV demonstrated that the current density for Cu₂S@CC was considerably greater, confirming the boosted reaction kinetics (Fig. 2a). Since catalytic materials generally accelerate both forward and reverse reactions by lowering the activation energy, Cu₂S is also expected to facilitate Cu²⁺ stripping during the anodic process. To confirm this, LSV was performed on Cu₂S@CC and CC electrodes in the anolyte. The Cu₂S@CC electrode consistently showed higher current density (Fig. S5a). The Tafel slopes and EIS results reveal significantly lower Tafel slopes and reduced charge-transfer impedance, respectively, highlighting the superior catalytic activity of the Cu₂S@CC electrode for Cu stripping (Fig. S5b and c). Potentiostatic stripping at 200 mV also showed a larger I–t curve area for Cu₂S@CC, suggesting more efficient Cu oxidation (Fig. 2b).

	Fig. 2 Current density of Cu (a) deposition and (b) stripping on bare CC and Cu2S@CC. Differential charge density map and charge density distribution plot of (c) and (d) CC and (e) and (f) Cu2S. The (g) P–I curve, (h) discharge curve and (i) potentials of a TRCB.

Density functional theory (DFT) calculations were performed to investigate the interaction between Cu²⁺ and Cu₂S.¹⁴ The adsorption energy of Cu²⁺ on Cu₂S was calculated to be −11.3 eV, which was significantly lower than that on graphite (−1.18 eV), indicating a much stronger binding affinity. Electron density difference maps reveal charge transfer from Cu₂S to Cu²⁺, confirming strong electronic interaction (Fig. 2c–f). Moreover, charge density distributions show substantial orbital overlap between Cu₂S and Cu²⁺, suggesting the formation of chemical bonds, in contrast to the weak physical adsorption observed with graphite. These results demonstrate that Cu₂S exhibits superior adsorption capability toward Cu²⁺, which underpins its catalytic behavior.

The bifunctional catalytic activity of Cu₂S toward the Cu²⁺/Cu redox couple markedly enhances the performance of TRCBs. As shown in Fig. 2g, the device delivered an open-circuit voltage of ∼900 mV and a peak power density of 14.2 mW cm⁻², surpassing previously reported values for TRBs using en ligands (11.9 mW cm⁻²). During the discharge of a TRCB, three distinct voltage drops are observed, corresponding to transitions in the electrode reactions (Fig. 2h and i). Initially, the anodic potential shifted rapidly, while the cathodic potential remained stable. Once the anodic Cu was depleted, a sharp increase in anodic potential resulted in the first voltage drop. The second drop was triggered by the completion of the conversion of S to Cu₂S, followed by the process of Cu deposition. Moreover, the third voltage drop marked the exhaustion of Cu₂S on the anode, terminating in the discharge of the TRCB. The TRCB achieves a high areal capacity of 32.8 mAh cm⁻² and an energy density of 788.8 Wh m⁻³, demonstrating the significant role of Cu₂S in enhancing both the kinetics and energy output of the system.

The above investigations revealed that the discharge curve of TRCBs exhibits three distinct voltage plateaus, underscoring the importance of analysing the individual power generation characteristics of each stage. To investigate the performance of each reaction stage independently, four TRCB configurations were fabricated with pre-defined electrode compositions: Cu/S-TRCB, Cu₂S/S-TRCB, Cu/Cu₂S-TRCB, and Cu₂S/Cu₂S-TRCB. The discharge performances of these configurations are summarized in Fig. 3a. Cu/S-TRCB exhibited the highest power density of 14.2 mW cm⁻², which was attributed to the favourable redox pair of metallic copper and S. As discharge progresses, the TRCB could transition to either the Cu/Cu₂S-TRCB or Cu₂S/S-TRCB state depending on the initial stoichiometry of Cu and S. When the anodic copper content exceeded the cathodic S capacity, the system would enter the Cu/Cu₂S-TRCB stage, wherein Cu oxidation at the anode couples with copper deposition at the cathode. This configuration maintains a relatively high power output with a peak power density of 9.1 mW cm⁻². Alternatively, with excess S, the TRCB would transition to the Cu₂S/S-TRCB stage, in which Cu₂S was oxidized at the anode while S reduction continued at the cathode. This stage exhibited a moderate power density of 7.5 mW cm⁻². In the final stage, Cu₂S/Cu₂S-TRCB, Cu deposition occurred at the cathodic reaction, while Cu₂S was fully oxidized back to elemental S at the anode. This process was accompanied by a significantly lower open-circuit voltage (360 mV), resulting in the lowest power density of 3.1 mW cm⁻². These findings indicated that the output performance of the TRCBs was not solely governed by reactant depletion, but also by intrinsic changes in the electrochemical reaction pathways as the discharge process evolves.

	Fig. 3 (a) P–I and U–I curves of a TRCB at different discharge stages. (b) Discharge curves and (c) energy density and capacity at different Cu loadings. (d) Relationship between energy and capacity.

The power density of the TRCBs varied significantly across the different electrochemical stages, and this variation was primarily governed by the relative contributions of Cu and S to the overall charge storage. However, the Cu²⁺ deposition process resulted in a sharp decrease in potential at the cathode, which became particularly pronounced in the later stages of discharge. Consequently, the Cu-to-S ratio played a crucial role in determining the voltage profile and energy output of the TRCBs across different stages. As shown in Fig. 3b, when the charge contribution from copper was lower than that from S or Cu₂S, the discharge process comprised three distinct stages. The TRCB initially operated in the Cu/S-TRCB stage, in which Cu was gradually consumed at the anode, leading to a continuous increase in its potential. Meanwhile, S underwent reduction at the cathode, resulting in a relatively stable cathodic potential and a slow decrease in the total output voltage. As the anodic Cu was depleted, Cu₂S became the anodic active material, and the system entered the Cu₂S/S-TRCB stage. During this period, both electrode potentials stabilized and formed a well-defined voltage plateau. Eventually, when the cathodic S had also been consumed, the TRCB entered the Cu₂S/Cu₂S-TRCB stage. In contrast, when the loading of Cu exceeded that of S, the initial and final discharge stages remained the same, but the intermediate stage differed. Upon S depletion, Cu deposition began earlier at the cathode, transitioning the system into the Cu/Cu₂S-TRB stage. Interestingly, when the charge capacities of Cu and S were well balanced, the discharge process was simplified to two dominant stages. However, when the charge imbalance was increased, the intermediate stage tended to vanish.

The performance metrics further highlight the importance of the Cu loading. For instance, in the TRB with Cu₂S/S electrodes, the battery delivered a capacity of just 22.7 mAh cm⁻² and an energy density of 529 Wh m⁻³. In contrast, increasing the Cu-to-S ratio to 1.2 boosted the capacity and energy density to 50.7 mAh cm⁻² and 1066 Wh m⁻³, respectively. Moreover, as shown in Fig. 3c, both capacity and energy density scale linearly with increasing copper loading under a constant S loading of 10 mg cm⁻². Despite this linear trend, Fig. S6 revealed that the anodic potential continuously shifted forward during the Cu oxidation process, resulting in the absence of a stable voltage plateau. This explained the reason for the energy density not doubling with the capacity, unlike in the other three reactions. Nevertheless, a clear linear relationship persists between energy density and capacity (Fig. 3d), with a slope of approximately 20.5 Wh m⁻³ per mAh.

The TRCBs had demonstrated high power and energy output; however, their ability to deliver stable and continuous power output is also crucial for realizing sustained thermoelectric conversion. To evaluate the cycling stability of the TRCBs, galvanostatic discharge tests were carried out at a constant discharge current of 20 mA cm⁻² using a TRCB with a Cu-to-S ratio of 1.0. To eliminate the influence of electrolyte degradation and the depletion of active species, fresh electrolyte was replenished and the anolyte and catholyte were alternated after each discharge cycle. As shown in Fig. 4a, the output voltage profiles remained largely consistent over 20 consecutive discharge cycles, and the voltage–capacity relationship suggested that the electrochemical reactions at each stage were reproducible across batches. To provide a clearer comparison, discharge curves from the 1st, 5th, 10th, 15th, and 20th cycles are plotted in Fig. 4b. These curves confirmed that the voltage trends remained similar throughout the cycling process. However, a gradual decline in maximum discharge capacity was observed with increasing cycles.

	Fig. 4 (a) U–I curves over 20 cycles under a constant current density of 20 mA cm−2, and (b) V–I curves for discharge cycles 1, 5, 10, 15, and 20. (c) Energy density and capacity of each cycle.

Fig. 4c shows the capacity and energy density for each discharge batch. Although slight fluctuations in energy density were noted, both metrics exhibited an overall downward trend. The initial cycle delivered a maximum capacity of 46.2 mAh cm⁻² and an energy density of 950 Wh m⁻³. By the 20th cycle, the capacity had decreased by 9.7%, demonstrating excellent cycling durability compared to the traditional TRB with Cu metal electrodes (as shown in Table S1).^4,15–18 Furthermore, the developed TRCBs achieved an average Coulombic efficiency exceeding 99% over the 20 cycles, underscoring their excellent electrochemical reversibility and long-term operational stability.

Herein, a TRCB based on the electrochemical conversion between Cu₂S and S was designed to improve both capacity and cycling stability. By coupling this conversion process with the Cu²⁺/Cu redox reaction, the TRCB achieved significant enhancements in overall performance. Notably, Cu₂S was found to effectively accelerate both anodic Cu stripping and cathodic Cu²⁺ deposition, resulting in improved kinetics and stable cycling behaviour. Moreover, the cascade electrochemical reaction strategy offers a viable route toward capacity scalability and overcomes the limitations of conventional battery configurations, presenting a new paradigm for the design of integrated two-in-one thermally regenerative batteries. These advancements will be essential for realizing the full potential of TRCBs in continuous and efficient low-grade thermoelectric energy conversion applications.

Foundation for Innovative Research Group Project of the National Natural Science Foundation of China (No. 52021004), the National Natural Science Foundation of China (No. 52376045), Natural Science Foundation of Chongqing, China (CSTB2022NSCQ-MSX1596).

Conflicts of interest

Data availability

SEM and digital images; electrochemical results of LSV and EIS. See DOI: https://doi.org/10.1039/d5cc03329b.

1. K. Li, L. Wang, Y. Wang, X. Feng, F. Jiang and M. Ouyang, escience, 2024, 100226 CrossRef .
2. J. D. Kocher, J. Woods, A. Odukomaiya, A. Mahvi and S. K. Yee, Energy Environ. Sci., 2024, 17, 226–2218 RSC .
3. R. Haldar, S. Naskar, B. Jana, D. Mandal and M. Shanmugam, Chem. Commun., 2024, 61, 318–321 RSC .
4. F. Zhang, J. Liu, W. Yang and B. E. Logan, Energy Environ. Sci., 2015, 8, 343–349 RSC .
5. J. Yang, Q. Yu, Y. Lei, S. Chen, Y. Yu and F. Yan, Appl. Energy, 2025, 391, 125917 CrossRef .
6. M. Rahimi, A. D'Angelo, C. A. Gorski, O. Scialdone and B. E. Logan, J. Power Sources, 2017, 351, 45–50 CrossRef CAS .
7. Y. Shi, L. Zhang, Y. Zhang, J. Li, Q. Fu, X. Zhu and Q. Liao, J. Phys. Chem. Lett., 2023, 14, 1663–1673 CrossRef CAS PubMed .
8. Y. Shi, T. Yang, Q. Ma, Y. An, L. Zhang, J. Wei, R. Gao, J. Li, Q. Fu, X. Zhu, Q. Liao and Z. Chen, Nano Energy, 2025, 141, 111118 CrossRef CAS .
9. G. Tang, Y. Li, J. Wu, Q. Kang and L. Ma, Nano Res. Energy, 2025 DOI:10.26599/NRE.2025.9120179 .
10. D. Luo, C. Li, Y. Zhang, Q. Ma, C. Ma, Y. Nie, M. Li, X. Weng, R. Huang, Y. Zhao, L. Shui, X. Wang and Z. Chen, Adv. Mater., 2022, 34, 2105541 CrossRef CAS PubMed .
11. C. Dai, L. Hu, X. Jin, H. Chen, X. Zhang, S. Zhang, L. Song, H. Ma, M. Xu, Y. Zhao, Z. Zhang, H. Cheng and L. Qu, Adv. Mater., 2021, 33, 2105480 CrossRef CAS PubMed .
12. Y. Zhang, M. Yao, P. Jing, Y. Fu, W. Wang, Y. Zhang, Q. Lai and Q. Wang, Angew. Chem., Int. Ed., 2025, 64, e202501205 CrossRef CAS PubMed .
13. X. Wu, A. Markir, L. Ma, Y. Xu, H. Jiang, D. P. Leonard, W. Shin, T. Wu, J. Lu and X. Ji, Angew. Chem., Int. Ed., 2019, 58, 12640–12645 CrossRef CAS PubMed .
14. Y. Wang, Z. Yang, N. Zhang, D. Wang, R. Wang, W. Peng, J. Zhang, J. Liu and J. Zhang, Appl. Surf. Sci., 2023, 610, 155340 CrossRef CAS .
15. X. Zhu, M. Rahimi, C. A. Gorski and B. Logan, ChemSusChem, 2016, 9, 873–879 CrossRef CAS PubMed .
16. Y. Zhang, Y. Shi, L. Zhang, J. Li, Q. Fu, X. Zhu and Q. Liao, J. Power Sources, 2021, 495, 229815 CrossRef CAS .
17. Y. Shi, L. Zhang, Y. Zhang, J. Li, Q. Fu, X. Zhu and Q. Liao, Chem. Eng. J., 2021, 423, 130339 CrossRef CAS .
18. Y. M. Ko, S. Lee, S. Kim and Y. T. Kang, Appl. Energy, 2025, 396, 126337 CrossRef CAS .