English

• Nowadays, the energy utilization efficiency in industrial production is not high, which is due to a large amount of energy waste, especially the low-grade waste heat resources (temperature < 130 ℃) [1]. In recent years, harvesting low-grade waste heat to generate electrical power has attracted considerable attention due to its vast potential for power generation [2]. At present, the direct thermal-electric energy conversion technologies mainly include solid-state semiconductor-based thermoelectric conversion systems [3] and liquid-based thermoelectrochemical cells (TECs) [4], while their power densities and thermal-electricity energy conversion efficiencies need to be improved to make them commercially viable [5].

  A thermally regenerative battery (TRB) [6] was recently developed for converting low-grade waste heat into electricity with a relatively high power density and low cost. The TRB system includes an electrical generation process and a thermally regenerative process, as shown in Fig. 1. During the electricity generation process, ammonia corrodes the copper anode, which produces the electrons that flow to the cathode against external resistance. The copper ions in the catholyte are deposited on the cathode after receiving electrons from the anode. After the discharging, ammonia is separated from the anodic product by a distillation process utilizing low-grade waste heat. The distilled ammonia will be transported to the anode chamber, and the next cycle is initiated. Both copper electrodes are alternately operated as anodes or cathodes in successive cycles. Since TRBs were first proposed, a lot of research has been conducted on improving the electricity generation performance, which mainly includes the investigation of flow battery structure design [7-9], simulation models construction [10, 11], novel electrolyte and electrode development [12-17], and applications in wastewater treatment [18, 19]. With regards to the thermal regeneration process, a preliminary study on the influence of operating parameters was conducted [20].

  Figure 1

  

  Figure 1.  Schematic of the two processes and ammonia crossover in a TRB system.

  DownLoad: Full-Size Img PowerPoint

  Through the efforts of previous research, the performance of TRBs has been significantly improved, but it is still limited by several key factors. Among these, ammonia crossover through the anion exchange membrane (AEM) from the anode to the cathode would cause self-discharge, which was commonly observed in the previous TRB studies [14, 18, 21]. As shown in Fig. 1, the ammonia crossover from the anode to the cathode chamber results in an unfavorable chemical consumption instead of electrochemical consumption of copper ions. This will cause the corrosion of the cathode copper and induces a mixed potential, then resulting in a discharge voltage curve without a relatively stable platform. It was reported that improving the operational temperature enhanced the reaction kinetics and then improved the battery performance, while the ammonia crossover was also boosted [21]. An electrolyte decoupling strategy achieved by an intermediate chamber with a buffer solution was proposed, which significantly delayed the emergency of ammonia crossover, but also reduced the power density [22]. Besides, a poly(phenylene oxide) anion exchange membrane was developed with high energy density in TRBs, but the result also demonstrated that there was a tradeoff between the power density and ammonia crossover [5]. Thus, investigating new solutions to alleviate the ammonia crossover in a TRB without reducing the power output is of vital importance.

  In this study, a graphene oxide (GO) modified AEM was proposed for alleviating the ammonia crossover and enhancing the energy density of TRBs. Due to the high specific surface area and abundant functional groups [23, 24], GO is expected to effectively absorb ammonia [25] and then improve the performance of TRBs. The surface modification was conducted by a simple electrophoretic deposition method. The surface characterization of membranes indicates that GO was effectively deposited on the surface by electrophoretic deposition. Besides, the performance of the TRB and the visualization of ammonia crossover during the discharging demonstrate that the ammonia crossover was effectively alleviated.

  Fig. 2 shows the modification principle of AEM by the electrophoretic deposition method. Graphene oxide has many carboxyl and hydroxyl functional groups and is negatively charged [26, 27]. Thus, graphene oxide could move from the negative stainless steel electrode side to the positive carbon cloth electrode side in a lab-made device with a DC power supply (Fig. S1 in Supporting information). In this experiment, a quaternary amine-type anion-exchange membrane with a positive charge on the membrane surface was used. It can be seen from Fig. 2 that the AEM has changed from the original faint yellow surface to the modified dark brown surface, and the surface deposition layer was evenly distributed, which proves that GO was successfully modified on the surface.

  Figure 2

  

  Figure 2.  Schematic diagram of the membrane surface modification principle.

  DownLoad: Full-Size Img PowerPoint

  To prove that the graphene oxide was successfully deposited on the surface of AEM, the modified membrane was photographed by SEM. As shown in Fig. 3a, the surface of the original membrane was uneven, possessing some bulges and sunken. This was mainly because the membranes applied in this experiment were manufactured by mixing small particles of a homogeneous ion exchange resin and an inert binder, thus resulting in the heterogeneous surface of AEMs [23]. After the deposition of GO nanosheets, the GO-AEM presented a wrinkled lamellar structure, as shown in Fig. 3b. With the high surface specific area of GO, the modified membrane could be able to block the ammonia by physical anchoring. AFM 3D images of the original membrane and GO modified membranes were shown in Figs. 3c and d. The root-mean-squared (RMS) roughness of the GO modified membrane was 28.5 nm, which was much lower than that of the original membrane (107 nm). With the increasing amount of deposited GO, GO nanosheets could cover the peaks and valleys on the surface of the original membrane, resulting in the obvious decrease of surface roughness of the GO modified AEM. XPS was also performed on the original membrane and GO modified membrane to further analyze chemical composition. Fig. 3e shows the wide XPS spectrum with three major peaks corresponding to O 1s (532.3 eV), N 1s (402.4 eV) and C 1s (284.8 eV). The atomic percentages of carbon, nitrogen and oxygen elements were also obtained. The oxygen content in the GO modified membrane (32.33%) was significantly higher than that of the original membrane (12.87%), which was mainly due to the functional groups of modified GO [24]. This also shows that graphene oxide was effectively deposited on the membrane.

  Figure 3

  

  Figure 3.  Surface characterization of membranes. (a) SEM images of the original AEM. (b) SEM images of the GO modified membrane. (c) AFM 3D image of the original and (d) GO modified membrane. (e) Wide XPS spectrum of membranes.

  DownLoad: Full-Size Img PowerPoint

  The above results indicate that the surface of AEM could be effectively modified by the electrophoretic deposition of graphene oxide. With a high specific surface and abundant functional groups, the modified GO would be favourable for blocking the ammonia crossover by physical anchoring and chemical adsorption [26]. To demonstrate the block effect on ammonia, the ammonia permeability (Fig. 4) of AEMs was tested in a two-chamber diffusion cell (Fig. S2 in Supporting information). The ammonia permeability of the GO-AEM (7.68×10^–10 cm²/s) was 39.5% lower than that of the AEM (1.27×10^–9 cm²/s), which indicates that the graphene oxide modified membrane could help to alleviate ammonia crossover in the TRBs. To investigate the effect of modified GO on conductivity, electrochemical impedance spectroscopy (EIS) was conducted in a cubic TRB, and the Nyquist plot was fitted according to the equivalent circuit described in Fig. S3 (Supporting information). The electrical resistance of the AEM and GO-AEM was 1.89 and 1.76 Ω respectively (Fig. S4 in Supporting information). The conductivity of the AEM and GO-AEM were 2.77 and 2.97 S/m, which indicated that the deposition of GO caused no obvious change in the conductivity of the membrane. The above results show that surface modification with GO can effectively reduce the ammonia permeability, without any significant influence on the conductivity of the membrane.

  Figure 4

  

  Figure 4.  Ammonia permeability and electrical resistance of the original and GO modified membranes.

  DownLoad: Full-Size Img PowerPoint

  To prove the feasibility of applying GO modified membrane in TRBs, the performance of TRB-GO-AEM was contrastively studied with that of TRB-AEM, as shown in Fig. 5a. Notably, the GO-modified surface was on the cathode side of the TRB. There was no obvious difference between the open-circuit voltage of TRBs with AEM and GO-AEM. The maximum power density of the TRB with the original membrane was 9.9 W/m², while the TRB with the GO modified membrane delivered a 24.3% higher maximum power (12.3 W/m²). This was possibly due to the decreased ammonia crossover and membrane resistance induced by the GO modification. Electricity generation of the TRBs was performed over a complete discharge batch, at the external resistance that produced the maximum power in the polarization test, as shown in Fig. 5b. During the discharge process, the voltage output of the TRBs decreased slowly in the main discharge period because of the constant ammonia consumption [14]. Although there was a similar total batch time (approximately 2.5 h), the TRB-GO-AEM had a relatively higher voltage output and a longer main discharging period compared with that of the TRB-AEM. The main discharging time of the TRB-AEM was about 80 min, and the voltage of the TRB-AEM dropped rapidly after that moment. This was mainly related to the self-discharge induced by ammonia crossover from anode to cathode, which induced the decay of ammonia in the anode and corrosion of copper in the cathode, then subsequently affected the electrical generation. With regards to the TRB-GO-AEM, the main discharge period lasted for about 110 min, which was relatively longer than that of the TRB-AEM.

  Figure 5

  

  Figure 5.  Battery performance and ammonia crossover visualization of TRBs with the original and GO modified membranes. (a) Power curves and voltages; (b) discharging curve; (c) visualization test; (d) total charge and energy density; (e) coulombic efficiency; (f) thermal energy efficiency and that relative to Carnot efficiency.

  DownLoad: Full-Size Img PowerPoint

  To demonstrate the effective alleviation of ammonia crossover, visualization was also conducted through the color evolution of the catholyte during the discharge process [22]. As shown in Fig. S5 (Supporting information), the initial catholyte of the TRBs was light blue with a pH of 4.1. With regards to the TRB-AEM, the dark blue color in the cathode chamber indicated that ammonia crossover was severe after 80 min, as shown in Fig. 5c. Besides, the cathode potential started to drop sharply, which indicates that there was a mixed potential induced by the reaction between the shuttled ammonia and cathode copper. However, the catholyte of the TRB-GO-AEM remained the initial light blue color, indicating that the ammonia crossover was weak at 80 min. As for this, the cathode potential was much higher than that of the TRB-AEM. After discharging process, the final pH value of the catholyte in the TRB-AEM was 6.90, while that of the catholyte in the TRB-GO-AEM was 6.16, as shown in Table S1 (Supporting information). The result of ammonia crossover visualization was consistent with the cathode potential and voltage plots, indicating that the ammonia crossover in the TRB-GO-AEM was effectively alleviated during the discharging process.

  Concerning the total charge, there was an 18.7% increase in the TRB-AEM (297.9 C) compared to that of the TRB-GO-AEM (250.9 C), as shown in Fig. 5d. The energy density of the TRB-GO-AEM (689.3 Wh/m³) was 20.2% higher than that of the TRB-AEM (573.3 Wh/m³). Coulombic efficiencies (CCE and ACE) were calculated to study electrode reversibility, as shown in Fig. 5e. The CCE and ACE of TRBs with original membrane were 39.72% and 90.3% respectively, which were comparable to CEs from previous studies that used copper mesh and foam electrodes [9, 13, 15]. The ammonia crossover was significantly reduced, which led to less cathode copper being corroded and more anode copper reacting with anodic ammonia, resulting in higher CCE (48.7%) and ACE (93.6%) for TRBs with modified membranes than for TRBs with original membranes. The TRAB thermal energy efficiency was estimated by comparing the electrical energy recovered with the energy required for the distillation of the ammonia from the anolyte solution [21]. The distillation process was simulated according to the diagram shown in Fig. S6 (Supporting information). As the TRB-GO-AEM produced a higher energy density, its thermal energy efficiency (0.56%) was higher than that of the TRB-AEM (0.46%). The efficiency relative to the Carnot efficiency increased from 3.6% to 5.2% when a GO modified membrane was applied (Fig. 5f). The above results indicated that the ammonia crossover was effectively alleviated and the battery energy generation was significantly improved.

  With regards to the stability of the GO modifying layer, it was evaluated by a long-term discharging test. As shown in Fig. S7a (Supporting information), the cell was discharged for successive 6 cycles in 30 h. Though the decreased discharging time for each batch and the darker blue color of the catholyte indicate that the ammonia crossover became more severe in long-term tests, this was mainly attributed to the decreased sites for ammonia adsorption. The optical and SEM images of the membranes before and after the stability test were obtained, as shown in Figs. S7b and c (Supporting information). The optical images show that the modifying GO layer was stable on the surface of AEM, only in a small part of the modified area did the GO fall off because of contact with the electrode. The SEM images also show no obvious difference between the membrane surfaces before and after the test, which also proves the good stability of the modified membrane.

  In summary, the graphene oxide modified AEM was proposed for the thermally regenerative battery to restrain the self-discharge and enhance the energy density. The properties of the GO-modified membranes were characterized and the results indicated that GO was effectively modified on the surface of the membrane. The chemical adsorption and physical anchoring of the graphene oxide were beneficial to alleviate the ammonia crossover in the TRBs. Compared with the original AEM, the GO modified AEM presented a higher conductivity of 2.97 S/m and a 39.5% lower ammonia permeability (7.68×10^–10 cm²/s). With regards to the battery performance, a 24.3% higher power density (12.3 W/m²), 18.9% higher total charge (297.9 C), and a 20.2% higher energy density (689.3 Wh/m³) were obtained in the TRB with the GO modified membrane (TRB-GO-AEM). Along with the visualization test, it is demonstrated that the ammonia crossover in the TRB was effectively alleviated by the GO modified membrane during the discharging process. The long-term discharging test and surface characterization also indicated that the GO modified membrane has good stability. The present study provides a simple and effective method to significantly alleviate the ammonia crossover without reducing the power output, showing a promising application in future TRBs.

  Declaration of competing interest

  The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

  Acknowledgments

  This work was supported by Innovative Research Group Project of National Natural Science Foundation of China (No. 52021004), National Natural Science Foundation of China (No. 51976018), Scientific Research Foundation for Returned Overseas Chinese Scholars of Chongqing, China (No. cx2021088), and Research Funds of Key Laboratory of Low-grade Energy Utilization Technologies and Systems (No. LLEUTS-2018005).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2022.07.047.

  
  
• 1. [1]

     Z.X. Su, M.L. Zhang, P.H. Xu, et al., Energy Convers. Manage. 229 (2021) 113769. doi: 10.1016/j.enconman.2020.113769

  2. [2]

     D. Zhao, A. Würger, X. Crispin, J. Energy Chem. 61 (2021) 88–103. doi: 10.1016/j.jechem.2021.02.022

  3. [3]

     M. Rahimi, A.P. Straub, F. Zhang, et al., Energy Environ. Sci. 11 (2018) 276–285. doi: 10.1039/c7ee03026f

  4. [4]

     M.F. Dupont, D.R. MacFarlane, J.M. Pringle, Chem. Commun. 53 (2017) 6288–6302. doi: 10.1039/C7CC02160G

  5. [5]

     M. Rahimi, L. Zhu, K.L. Kowalski, et al., J. Power Source. 342 (2017) 956–963. doi: 10.1016/j.jpowsour.2017.01.003

  6. [6]

     F. Zhang, J. Liu, W. Yang, B.E. Logan, Energy Environ. Sci. 8 (2015) 343–349. doi: 10.1039/C4EE02824D

  7. [7]

     V.M. Palakkal, T. Nguyen, P. Nguyen, et al., ACS Appl. Energy Mater. 3 (2020) 4787–4798. doi: 10.1021/acsaem.0c00400

  8. [8]

     Y.S. Zhang, L. Zhang, J. Li, et al., Electrochim. Acta 331 (2020) 135442. doi: 10.1016/j.electacta.2019.135442

  9. [9]

     Y.S. Zhang, Y. Shi, L. Zhang, et al., J. Power Source 495 (2021) 229815. doi: 10.1016/j.jpowsour.2021.229815

  10. [10]

      W.G. Wang, G.Q. Shu, H. Tian, X.P. Zhu, J. Power Source 388 (2018) 32–44. doi: 10.5539/enrr.v8n4p32

  11. [11]

      H. Tian, W. Jiang, G.Q. Shu, et al., J. Electrochem. Soc. 166 (2019) A2814–A2825. doi: 10.1149/2.0711912jes

  12. [12]

      M. Rahimi, A. D'angelo, C.A. Gorski, O. Scialdone, B.E. Logan, J. Power Source 351 (2017) 45–50. doi: 10.1016/j.jpowsour.2017.03.074

  13. [13]

      M. Rahimi, T. Kim, C.A. Gorski, B. Logan, J. Power Source. 373 (2018) 95–102. doi: 10.1016/j.jpowsour.2017.10.089

  14. [14]

      L. Zhang, Y.X. Li, X. Zhu, et al., Ind. Eng. Chem. Res. 58 (2019) 7408–7415. doi: 10.1021/acs.iecr.9b00616

  15. [15]

      Y. Shi, L. Zhang, J. Li, et al., Renew. Energ. 159 (2020) 162–171. doi: 10.1016/j.renene.2020.05.147

  16. [16]

      Y. Shi, L. Zhang, Y.S. Zhang, et al., Chem. Eng. J. 423 (2021) 130339. doi: 10.1016/j.cej.2021.130339

  17. [17]

      P.Y. Chen, L. Zhang, Y. Shi, et al., J. Power Source. 517 (2022) 230719. doi: 10.1016/j.jpowsour.2021.230719

  18. [18]

      M. Rahimi, Z. Schoener, X.P. Zhu, et al., J. Hazard. Mater. 322 (2017) 551–556. doi: 10.1016/j.jhazmat.2016.10.022

  19. [19]

      W.G. Wang, G.Q. Shu, H. Tian, X.P. Zhu, Sep. Purif. Technol. 235 (2020) 116230. doi: 10.1016/j.seppur.2019.116230

  20. [20]

      F. Vicari, A. D'angelo, Y. Kouko, et al., J. Appl. Electrochem. 48 (2018) 1381–1388. doi: 10.1007/s10800-018-1240-0

  21. [21]

      F. Zhang, N. Labarge, W. Yang, J. Liu, B.E. Logan, ChemSusChem 8 (2015) 1043–1048. doi: 10.1002/cssc.201403290

  22. [22]

      W.G. Wang, G.Q. Shu, X.P. Zhu, H. Tian, J. Mater. Chem. A 8 (2020) 12351–12360. doi: 10.1039/d0ta03236k

  23. [23]

      Y.J. Li, S.Y. Shi, H.B. Cao, Z.J. Zhao, H. Wen, Appl. Surf. Sci. 442 (2018) 700–710. doi: 10.1016/j.apsusc.2018.02.166

  24. [24]

      Y. Wang, S.S. Li, H.T. Yang, J. Luo, RSC Adv. 10 (2020) 15328–15345. doi: 10.18632/aging.103518

  25. [25]

      Y. Peng, J.H. Li, Front. Environ. Sci. Eng. 7 (2013) 403–411. doi: 10.1007/s11783-013-0491-6

  26. [26]

      Y.J. Li, S.Y. Shi, H.B. Cao, et al., ACS Appl. Nano Mater. 3 (2020) 588–596. doi: 10.1021/acsanm.9b02142

  27. [27]

      M. Yang, C.W. Zhao, S.F. Zhang, P. Li, D.Y. Hou, Appl. Surf. Sci. 394 (2017) 149–159. doi: 10.1016/j.apsusc.2016.10.069

• 

  Figure 1  Schematic of the two processes and ammonia crossover in a TRB system.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  Schematic diagram of the membrane surface modification principle.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  Surface characterization of membranes. (a) SEM images of the original AEM. (b) SEM images of the GO modified membrane. (c) AFM 3D image of the original and (d) GO modified membrane. (e) Wide XPS spectrum of membranes.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  Ammonia permeability and electrical resistance of the original and GO modified membranes.

  下载: 全尺寸图片 幻灯片
  

  Figure 5  Battery performance and ammonia crossover visualization of TRBs with the original and GO modified membranes. (a) Power curves and voltages; (b) discharging curve; (c) visualization test; (d) total charge and energy density; (e) coulombic efficiency; (f) thermal energy efficiency and that relative to Carnot efficiency.

  下载: 全尺寸图片 幻灯片
•