Unveiling the bifunctional role of morphological differences of self-supported Cu(OH)₂ in electrocatalysis†

Brahmari Honnappa ^a, T. R. Naveen Kumar ^a, Prince J. J. Sagayaraj‡ ^b, Sulakshana Shenoy ^c, Chitiphon Chuaicham ^c, Manova Santhosh Yesupatham‡ ^b, Anantharaj Sengeni ^d, Bernaurdshaw Neppolian ^b, Keiko Sasaki *^c and Karthikeyan Sekar *^b
aDepartment of Physics and Nanotechnology, SRM Institute of Science and Technology, Kattankulathur 603203, Tamil Nadu, India
^bSustainable Energy and Environmental Research Laboratory, Department of Chemistry, Faculty of Engineering and Technology, SRM Institute of Science and Technology, Kattankulathur 603203, Tamil Nadu, India. E-mail: Karthiks13@srmist.edu.in
^cDepartment of Earth Resources Engineering, Kyushu University, Fukuoka 819-0395, Japan
^dLaboratory for Electrocatalysis and Energy, Department of Chemistry, Indian Institute of Technology, Kanpur, Uttar Pradesh 2018 016, India

First published on 15th November 2023

---

---

To alleviate the dependence on fossil fuels, electrochemical energy conversion and energy storage devices, like polymer electrolyte membrane fuel cells (PEMFCs), water electrolysis, metal–air batteries (MAB), etc. have garnered unprecedented interest in recent years.¹ The overall efficiency of these devices is hindered due to the oxygen evolution reactions (OER). In general, the OER are four-electron/proton-coupled reactions that require higher energy to form the OO bond and other intermediates, respectively.² Currently, IrO₂/RuO₂ catalysts are considered as the state-of-the-art catalysts;^3,4 however, their high cost and low stability at high current density have encouraged research groups to develop new catalysts. In the meantime, few groups have desired to perform the OER with other anodic reactions. With this insight, researchers have focused on direct methanol fuel cells (DMFCs) due to their higher efficiency and less emission of greenhouse gases. In DMFCs, the methanol oxidation reaction (MOR) occurs at the anode and reduction at the cathode. To date, noble metals have been widely employed in both OER and MOR applications, due to their high catalytic efficiency and redox properties. However, low stability and high cost made various research groups develop an ideal catalyst for both the OER and MOR, which should increase the overall efficiency and reduce the cost of the device. Among different types of non-noble metals, polymers, MOFs, etc., transition metal^5,6 (particularly copper)-based materials have notable prospects owing to their high conductivity, redox properties, optimal adsorption energies and their abundance in nature. Besides, copper-based materials have also been shown to catalyse the electrooxidation of organic compounds, such as alcohols, hydrocarbons and so on.^7–9 It is also worth mentioning that synthesizing Cu-based materials¹⁰ with different morphology/structure can augment their electrocatalytic activity towards OER and MOR applications. To date different synthesis techniques like hydrothermal, wet-chemical, and combustion methods, as well as electrochemical anodization, and electrodeposition have been adopted to synthesize different morphologies of Cu-based materials.¹¹ Owing to the simplicity and efficiency (time and energy) we chose the electrochemical anodization technique.¹²

Interestingly, we employed a copper substrate as the monolith, which serves as both substrate and the source for a catalyst to give that added advantage towards commercialization. To date, anodization is performed by galvanostatic and potentiostatic techniques using both two-electrode and three-electrode systems.¹² However, to avoid unnecessarily lengthy reaction time and to save energy, we chose to anodize by the potentiostatic method at −0.06 V vs. Ag/AgCl in 1.0 M KOH (Scheme 1), and the respective i–t curve is shown in Fig. S1.†¹³ The potential of −0.06 V was chosen because of the oxidization capability of Cu to Cu²⁺ that results in Cu(OH)₂ formation instead of CuO. The formation of Cu(OH)₂ prevents the phase transformation in the electrolyte that plays a crucial role in OER and MOR applications.¹³

Further, in order to obtain different morphology, we introduced structure-directing agents, like urea and acetic acid in the KOH solution. With this insight, we have successfully grown Cu(OH)₂/CuO with different morphologies, like nanoneedles on copper foil. Unlike other conventional techniques, we obtained these morphologies within 200 s that privileged our work. Briefly, while anodizing the copper foil, metallic copper was oxidized and the re-deposition of copper took place.¹⁴ Interestingly, adding different structure-directing agents like urea and acetic acid to KOH leads to a change in pH and the growth rate of the Cu(OH)₂ that resulted in different morphology.¹³

Displayed in Fig. 1(a and b) are the Field Emission Scanning Electron Microscopy (FE-SEM) images of the anodized Cu foils in the presence of KOH (CuK), while CuKU and CuKAA are observed when urea and acetic acid were added to the KOH (Fig. 1(c–f) and S2†), respectively. High-Resolution Transmission Electron Microscopy with Energy Dispersive X-ray (HRTEM-EDX) spectroscopy mapping images of CuK indicated the formation of nanoneedles (Fig. 1(g and h), and S3†) with the presence of elements namely Cu and O as presented in Fig. S4† intrigued by the morphology differences, other physicochemical characteristics of prepared catalysts were also studied along with the bare Cu foil. Firstly, the pristine copper substrate displayed uneven sheet-like morphology and the change in morphology was observed in FE-SEM. To understand the crystal structure of bare foil, X-ray diffraction (XRD) analysis was performed and the diffraction pattern of the bare foil displayed peaks at 43°, 50° and 73° corresponding to the (111), (200) and (220) planes of metallic Cu, respectively. After anodization the peak intensity at the (111) facet drastically reduced,¹⁵ which indicated the growth of Cu(OH)₂ (Fig. S5†). It is also worth mentioning that the peak that appeared around 24° corresponds to the metallic copper and after anodization, the peak shifted towards the higher angle, which might be due to the deficiency of metallic copper on the surface.¹⁶ Further, the interplanar distance of the CuK sample was found to be 0.21 nm from HR-TEM which is in accordance with the XRD pattern. Moreover, the calculated hkl value of Selected Area Electron Diffraction (SAED) pattern for the CuK sample is shown in Fig. S6.† In addition, the Raman spectra of the bare and anodized samples, recorded to confirm the formation of Cu(OH)₂ and CuO, are presented in Fig. S7.†

The pristine electrocatalyst displayed A_1g and E_g stretching vibrations along with highlighting the Cu(OH)₂ and CuO. The peaks that appeared around 220, 245, 260, 270 and 550 cm⁻¹ corresponded to the CuO structures.¹⁷ After anodization, all the catalysts displayed new peaks at 490 and 1050 cm⁻¹ that are attributed to the formation of Cu(OH)₂.¹⁸ The decrease in Cu metallic state along with CuO and the formation of Cu(OH)₂ in anodized samples strongly suggest that the formation of Cu(OH)₂ was successful. Further, to have a clear understanding of the chemical compositions and oxidation state of the synthesized sample, X-ray photoelectron spectroscopy (XPS) measurement was carried out and the results are shown in Fig. 2, S8 and S9.†

The core-level spectra of Cu 2p_3/2 and Cu 2p_1/2, along with the satellite peak in CuK catalyst confirm the formation of Cu(OH)₂.¹⁹ Similarly, the characteristic peaks of Cu 2p_3/2 and Cu 2p_1/2 in CuKU were observed at 931.15 eV and 951.05 eV and in case of CuKAA catalyst the spin–orbit binding energy at 932.25 eV and 952.15 eV reveals the presence of Cu(OH)₂ and copper oxide.²⁰ Here, the ratio of Cu (OH)₂ was found to be (42.95%), (46.17%), (41.95%) in Cu 2p_3/2 and (21.96%), (20.5%) and (20.17%) in Cu 2p_1/2 for CuK, CuKU and CuKAA samples showed in Fig. 2(a–c). Further, O 1s spectra were deconvoluted and results are given in Fig. 2(d–f). In the CuK catalyst, the binding energy at 529.4 eV, 530.6 eV and 531.5 eV correspond to Cu(OH)₂, and CuO,^13,21 respectively. Interestingly, the Cu(OH)₂ ratio showed a drastic change in CuKU and CuKAA catalysts. Moreover, CuKU and CuKAA shifted towards higher binding energy compared with CuK catalysts, due to the anodization.

Moreover, we integrated X-ray Absorption Spectroscopy (XAS) with Auger Emission Spectroscopy (AES) to acquire the exact electrical structure and chemical bonding of a catalyst. X-ray Absorption Near Edge Structure (XANES) spectra of CuO, Cu(OH)₂, Cu standard and Cu foil are investigated and Cu(I), Cu(II) and metallic Cu are distinguished based on their binding energies.^18,22 As shown in Fig. 3(a) the pre-edge XANES spectra Cu species exhibited a transition between 8975 and 8980 eV. The initial transition observed between 8981 and 8984 eV is due to the dipole-allowed 1s-to-4p transition of Cu(I) and metallic copper (8999–9001 eV). To distinguish the Cu(I) and Cu(II) peak, the first-order derivates of XANES spectra was demonstrated. Briefly, the Cu(I) peak was observed at 8981.7 eV and Cu(II) peak was found at 8987.8 eV. Here Cu(II) spectra displayed an octahedral symmetry. Notably, a shoulder at 8984–8988 eV corresponds to 1s-to-4p_z and an intense feature at 8995–9002 eV was attributed to 1s-to-4p_xy transitions, potentially depicting the existence of Cu(II).²³ In the case of CuO, the transition of 1s to 3d is responsible for the XANES pre-edge feature, whereas the 1s to vacant 4p and 3d states are responsible for the major edge and post-edge, respectively.

The Fourier-transformed Extended X-ray Absorption Fine Structure (EXAFS) spectra of copper in radial distance (R) are exhibited in Fig. 3(b). R space value of the coordination number for the copper atoms is measured between 0 and 2.5 Å. Here, the Cu–O coordination distance is the cause of the EXAFS spectrum in CuO showing a noticeable peak at approximately 1.95 Å. In octahedral geometry, for Cu(OH)₂, the coordination number for copper(II) is typically six, with the closest neighbours at a distance of around 2 Å being the six oxygen atoms. In the specific EXAFS investigation, the R-space value can change, although it commonly falls between 1.5 Å.²⁴ In particular, the excitation of 2p electrons to the vacant 3d orbitals of the copper atoms in CuO leads to a significant rise in X-ray absorption in the R edge area. The copper ions in Cu(OH)₂ have a high charge density, which allows them to attract and bind to the oxygen molecules more effectively than other catalysts. The bulk composition of Cu(OH)₂ and Cu ratio has been shown in Fig. S10.†

To elucidate the role of hydroxide and effect of different morphology, the as-prepared electrocatalyst was evaluated for the OER in 0.1 M KOH and the MOR in 1 M KOH along with the 0.5 M methanol solution and the results are discussed below.

The CV curves of the OER were asserted in 0.1 M KOH at the potential window of 0.65 to 2.0 V vs. RHE (Fig. S11†). For pristine and anodized samples, the oxidation peak indicates the transition of Cu₂O → Cu(OH)₂/CuO and CuO → Cu alterations, respectively. Linear sweep voltammetry (LSV) curves of bare and all anodized samples are exhibited in Fig. 4(a). The overpotential was calculated at 10 mA cm⁻². For the pristine electrocatalyst, the overpotential was observed at 570 mV vs. RHE.

Comparison plot of overpotential and onset potential for bare Cu, CuK, CuKU and CuKAA in 0.1 M KOH presented in Fig. 4(b). After anodization, the overpotential was reduced by 200 mV for CuK (400 mV vs. RHE) sample and for CuKU and CuKAA samples, the overpotential was found to be at 500 and 510 mV vs. RHE, respectively. Tafel value for bare and all anodized electrocatalysts was found to be 144.3, 59.33, 82.82 and 90.2 mV dec⁻¹ as picturized in Fig. 4(c).

To study the reaction kinetics and the charge transfer characteristics of the prepared electrocatalyst, the Tafel plot and Nyquist plot were inspected. The Tafel value indicates that the CuK sample required a lower overpotential than all the anodized samples. The difference in activity among different nanostructures is directly correlated to their morphologies and their CuO:Cu(OH)₂ ratio.²⁵ In our case, the synthesized catalyst with a high CuO ratio exhibited relatively lower activity than the one rich in Cu(OH)_2, indicating that they have the highest efficiency for oxidising water. In line with the Tafel plot, Nyquist plots also demonstrated that the CuK sample has lower R_ct resistance, which reveals that the CuK sample has the highest conductivity than other catalysts. The R_ct value was calculated from the Nyquist plot of CuK, CuKU and CuKAA materials presented in Fig. S12.† The relative order of R_ct value for bare, CuK, CuKU, and CuKAA, as follows 79.15, 18.41, 37.34 and 56.91 Ω. Electrical double-layer capacitance was used to measure the electrochemically active surface area (ECSA) of the prepared electrocatalyst, which confirmed the increase in electrochemical activity. Fig. S13 and Table S1† show the calculated turn over frequency (TOF), ECSA and roughness factor (R_f) from eqn S1–S4,† respectively. Compared to the bare, both ECSA and R_f values were increased with anodized catalysts, particularly, the values were high for CuK. The R_f was calculated and it was found that the CuK had the maximum value of 9.66, which justifies the better activity of CuK. Comparatively, the other two materials CuKU and CuKAA showed the R_f values of 9.5 and 5.51, respectively. Alongside, ECSA was calculated from the eqn S3† and attained values of 20.74, 19.12 and 11.01 cm² corresponding to CuK, CuKU and CuKAA, respectively, indicating CuK sample has the highest active sites for the reaction. Moreover, the prepared electrocatalyst was compared with the existing literature. In comparison to CuO with a 470 mV²⁶ overpotential, Cu(OH)₂NWAs/Cu with a 560 mV²⁷ overpotential, Cu₂Se–Cu₂O/TF with a 465 mV²⁸ overpotential, CuO film with an 810 mV²⁹ overpotential, and Cu₂Se with a 470 mV³⁰ overpotential for the OER, our catalyst displayed a 400 mV overpotential as shown in Fig. 4(d) and Table S2.† These results show prepared electrocatalyst displayed better activity than other existing materials. Moreover, recent studies show that the higher performance of Cu-based catalysts is caused by the conversion of copper oxides into metastable Cu(III) ions during water oxidation. The overall kinetics of the cyclic reaction are controlled by the correct wetting of the active sites with the ionic electrolyte and the quick conversion produced on the active sites during electrocatalysis.

The stability of the CuK was tested at 0.4 V vs. RHE using CA studies. The prepared CuK electrocatalyst showed remarkable stability for 72 h as exhibited in Fig. S14.† The current density was retained at almost 94% at the end of 72 h shown in Fig. S15.† Further, the collected sample after CA studies was analysed using FE-SEM and a significant change in the morphology of CuK was observed in Fig. S16.† Moreover, the impedance analysis after the reaction reveals a similar R_ct value of nanoneedles (Fig. S17†). Similarly, post-stability of XPS studies in the CuK sample reveals a decrease in the Cu(OH)₂ and an increase in CuO percentage (Fig. S18†). However, the current density was retained due to the conversion of nanoneedle to nanorod.

We have also demonstrated the use of Cu(OH)₂–CuO heterostructure for the MOR in alkali electrolyte. Initially, the MOR of the bare electrocatalyst was performed in 1 M KOH and different (0.25, 0.5, 1.0 and 1.5) M MeOH solutions (Fig. S19 and S20†). After the evaluation of bare catalyst, all anodized sample was performed in 1 M KOH with 0.5 M MeOH solution as displayed in Fig. 5(a and b). A coupled proton–electron transfer (CPET) mechanism has been proposed where MeOH was oxidized into CO₂, liberating six electrons.

The bare catalyst showed a current density of 8.65 mA cm⁻², whereas, the anodized sample exhibited a remarkable current density of 133 mA cm⁻², 126 mA cm⁻² and 119 mA cm⁻² for CuK, CuKU and CUKAA catalysts. The increase in current density was mainly attributed to the formation of CuOOH, which prevents the phase transformation and quick conversion of CuO. It is also worth mentioning that the onset potential of the anodized sample was reduced by almost 300 mV compared to pristine copper foil.

Moreover, the Tafel value was calculated in the presence of MeOH and values indicate that the electrocatalyst follows the Langmuir isotherm process indicating the OH bond scissoring as depicted in Fig. 5(c). Compared to the other catalyst (Fig. 5(d) and Table S3†), the CuK sample showed a lower Tafel value of 34.67 mV dec⁻¹, whereas CuKU and CuKAA displayed the Tafel values of 52.76 and 171 mV dec⁻¹, respectively. Further different scan rate of bare and CuK catalyst was performed in KOH and in the presence of MeOH (Fig. S21†). Furthermore, the self-supported catalyst compared with the commercial Pt/C in 1 M KOH with 0.50 M of methanol with a scan rate of 10 mV s⁻¹ solution as presented in Fig. S22,† the Pt/C shows 57.4 mA cm⁻² as the initial current densities. To evaluate the long-term stability of the catalyst, CA studies were employed and the results are picturized in Fig. S23,† The pristine and CuK electrocatalysts were performed at 0.6 V for 7200 s. Notably, the current density of the pristine catalyst declined by 50% of its initial value, however, the CUK electrocatalyst retained almost 90% (180 to 140 mA cm⁻²).

In conclusion, straightforward anodization was utilized in synthesizing different morphologies of Cu(OH)₂ like nanoneedle, nano-cuboid and nano-cuboctahedron. The role of Cu(OH)₂ structure has been elaborated for the OER and MOR applications. In the OER, the CUK catalyst showed an overpotential of 400 mV at 10 mA cm⁻² and an onset potential of 835 mV. Moreover, the stability of Cu(OH)₂ in the OER was retained by ∼94% after 72 h. Similarly, in the MOR the current density of 180 mA cm⁻² was obtained with an overpotential of 315 mV vs. RHE for CuK. The increase in activity was mainly due to the presence of Cu(OH)₂ than CuO. The simple anodization technique pays the path for depositing Cu(OH)₂ and increases the active site.

Author contributions

Brahmari Honnappa: experimentation, data curation, investigation, writing original draft & editing. T. R. Naveen Kumar: formal analysis, writing original draft & editing. Prince J. J. Sagayaraj: formal analysis. Sulakshana Shenoy: formal analysis. Chitiphon Chuaicham: formal analysis. Manova Santhosh Yesupatham: formal analysis. Anantharaj Sengeni: formal analysis. Bernaurdshaw Neppolian: formal analysis. Keiko Sasaki: resources, review & editing. Karthikeyan Sekar: supervision and formal analysis.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

K. S. would like to thank the Royal Society-Newton International Fellowship Alumni follow-on funding support AL\211016 and AL\221024. K. S. would like to thank the Department of Chemistry at SRM- Institute of Science and Technology. K. S. also thanks the SERB Start-up Research Grant (SRG/2023/000658).

Notes and references

1. D. Zhang, R. Wang, X. Wang and Y. Gogotsi, Nat. Energy, 2023, 8(6), 567–576 CrossRef CAS .
2. X. Li, H. Lei, L. Xie, N. Wang, W. Zhang and R. Cao, Acc. Chem. Res., 2022, 55, 878–892 CrossRef CAS PubMed .
3. Z.-Y. Wu, F.-Y. Chen, B. Li, S.-W. Yu, Y. Z. Finfrock, D. M. Meira, Q.-Q. Yan, P. Zhu, M.-X. Chen, T.-W. Song, Z. Yin, H.-W. Liang, S. Zhang, G. Wang and H. Wang, Nat. Mater., 2023, 22, 100–108 CrossRef CAS PubMed .
4. H. Tian, Y. Yu, Q. Wang, J. Li, P. Rao, R. Li, Y. Du, C. Jia, J. Luo, P. Deng, Y. Shen and X. Tian, Int. J. Hydrogen Energy, 2021, 46, 31202–31215 CrossRef CAS .
5. B. H. Honnappa, S. Mohan, M. Shanmugam, A. Augustin, P. J. Sagayaraj, C. Chuaicham, S. Rajendran, T. K. A. Hoang and K. Sekar, Energy Adv., 2022, 738–760 RSC .
6. D. A. Agyeman, Y. Zheng, T.-H. Lee, M. Park, W. Tamakloe, G.-H. Lee, H. W. Jang, K. Cho and Y.-M. Kang, ACS Catal., 2021, 11, 424–434 CrossRef CAS .
7. R. Rajak, M. Saraf, P. Kumar, K. Natarajan and S. M. Mobin, Inorg. Chem., 2021, 60, 16986–16995 CrossRef CAS PubMed .
8. T. R. Naveen Kumar, L. R. Harishitha and B. Neppolian, Ceram. Int., 2022, 48, 29025–29030 CrossRef .
9. M. Nallal, K. H. Park, S. Park, J. Kim, S. Shenoy, C. Chuaicham, K. Sasaki and K. Sekar, ACS Appl. Nano Mater., 2022, 5, 17271–17280 CrossRef CAS .
10. C.-Z. Yuan, S. Wang, K. San Hui, K. Wang, J. Li, H. Gao, C. Zha, X. Zhang, D. A. Dinh, X.-L. Wu, Z. Tang, J. Wan, Z. Shao and K. N. Hui, ACS Catal., 2023, 13, 2462–2471 CrossRef CAS .
11. M. B. Gawande, A. Goswami, F.-X. Felpin, T. Asefa, X. Huang, R. Silva, X. Zou, R. Zboril and R. S. Varma, Chem. Rev., 2016, 116, 3722–3811 CrossRef CAS PubMed .
12. S. Anantharaj, H. Sugime and S. Noda, ACS Appl. Mater. Interfaces, 2020, 12, 27327–27338 CrossRef CAS PubMed .
13. S. Anantharaj, H. Sugime, S. Yamaoka and S. Noda, ACS Appl. Energy Mater., 2021, 4, 899–912 CrossRef CAS .
14. S. Y. Lee, H. Jung, N.-K. Kim, H.-S. Oh, B. K. Min and Y. J. Hwang, J. Am. Chem. Soc., 2018, 140, 8681–8689 CrossRef CAS PubMed .
15. N. Ulrich, M. Schäfer, M. Römer, S. D. Straub, S. Zhang, J. Brötz, C. Trautmann, C. Scheu, B. J. M. Etzold and M. E. Toimil-Molares, ACS Appl. Nano Mater., 2023, 6, 4190–4200 CrossRef CAS .
16. D. P. Singh, A. K. Ojha and O. N. Srivastava, J. Phys. Chem. C, 2009, 113, 3409–3418 CrossRef CAS .
17. Q. Lei, L. Huang, J. Yin, B. Davaasuren, Y. Yuan, X. Dong, Z.-P. Wu, X. Wang, K. X. Yao, X. Lu and Y. Han, Nat. Commun., 2022, 13, 4857 CrossRef CAS PubMed .
18. Y. Deng, A. D. Handoko, Y. Du, S. Xi and B. S. Yeo, ACS Catal., 2016, 6, 2473–2481 CrossRef CAS .
19. S. D. Giri, S. M. Mahajani, A. K. Suresh and A. Sarkar, Mater. Res. Bull., 2020, 123, 110702 CrossRef CAS .
20. Z. L. Wang, X. Y. Kong, X. Wen and S. Yang, J. Phys. Chem. B, 2003, 107, 8275–8280 CrossRef CAS .
21. S. Gong, X. Wu, J. Zhang, N. Han and Y. Chen, CrystEngComm, 2018, 20, 3096–3104 RSC .
22. S. Kühl, A. Tarasov, S. Zander, I. Kasatkin and M. Behrens, Chem.–Euro. J., 2014, 20, 3782–3792 CrossRef PubMed .
23. M. C. Hsiao, H. P. Wang, Y.-J. Huang and Y. W. Yang, J. Synchrotron Radiat., 2001, 8, 931–933 CrossRef CAS PubMed .
24. T. Wu, S. N. MacMillan, K. Rajabimoghadam, M. A. Siegler, K. M. Lancaster and I. Garcia-Bosch, J. Am. Chem. Soc., 2020, 142, 12265–12276 CrossRef CAS PubMed .
25. M. S. Burke, L. J. Enman, A. S. Batchellor, S. Zou and S. W. Boettcher, Chem. Mater., 2015, 27, 7549–7558 CrossRef CAS .
26. X. Liu, S. Cui, M. Qian, Z. Sun and P. Du, Chem. Commun., 2016, 52, 5546–5549 RSC .
27. C.-C. Hou, W.-F. Fu and Y. Chen, ChemSusChem, 2016, 9, 2069–2073 CrossRef CAS PubMed .
28. H. Chen, Y. Gao, L. Ye, Y. Yao, X. Chen, Y. Wei and L. Sun, Chem. Commun., 2018, 54, 4979–4982 RSC .
29. T.-T. Li, S. Cao, C. Yang, Y. Chen, X.-J. Lv and W.-F. Fu, Inorg. Chem., 2015, 54, 3061–3067 CrossRef CAS PubMed .
30. J. Masud, W. P. R. Liyanage, X. Cao, A. Saxena and M. Nath, ACS Appl. Energy Mater., 2018, 1, 4075–4083 CrossRef CAS .

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d3ta05471c

‡ These authors have contributed equally.

This journal is © The Royal Society of Chemistry 2023