Sustainable one step process for making carbon-free TiO₂ anodes and sodium-ion battery electrochemistry

Abstract

Electrodes with a carbon-free architecture impart a sustainable solution by eliminating side reactions. Such designs are required for anode systems, especially for sodium-ion batteries where the capacity contribution from carbon is noticed. Herein, carbon-free TiO₂ is prepared in the monolithic dendritic form via a facile single-step aerosol vapor deposition technique, grown onto a stainless steel current collector. The single crystalline nature of the anatase dendrites explicitly provides detailed crystallographic analysis. Herein, we have realized an escalation of the rutile phase with the existing anatase phase upon cycling, which is depicted as a moiré pattern, while evaluating the structure through TEM studies. As the anode is carbon and binder-free, the overall material is electrochemically active, and hence appreciably improves the volumetric energy-density. Here, we observe a long cycle life of 1000 cycles from the additive-free anatase dendrites with an adequate capacity retention of 85.1%. Therefore, we believe that our single-step fabrication process to procure additive-free monolithic electrodes is a forthcoming paramount technique to produce sodium-ion batteries on mass.

---

Introduction

Devising sodium-ion batteries (SIBs) is one of the most topical interests in current materials research.¹ Although sodium-ion batteries produce comparatively low energy-density systems,² making the electrode carbon and binder-free is noteworthy to reach the specifications of state-of-the-art lithium technology, as the electrodes contain only active materials.³ Currently, sodium electrochemistry requires a durable anode, while on the other hand, the cathode materials, such as Na₃V₂(PO₄)₃,⁴ and Prussian blue,⁵ have been adequately ameliorated to meet the requirements of energy density as well as durability. Lithium electrochemistry mainly uses four commercialized anodes: graphite,⁶ Li₄Ti₅O₁₂,⁷ Sn,⁸ and TiO₂.⁹ However, carbonaceous anodes are still unable to deliver adequate durability over 1000 cycles at high rate, resulting in rapid capacity fading.¹ Hence, TiO₂ has become the spotlight of current research. TiO₂ is well-known as a high-rate and durable anode for lithium-ion batteries, but producing similar performance with sodium systems^10–12 has become a tangible challenge due to the sluggish kinetics of Na-ions.¹³

Reviewing the recent literature, it can be observed that TiO₂ has been studied as an anode for sodium-ion batteries, effectuating the concept of making binder and carbon-free electrodes.⁹ In relation to this, we have found that carbon additives have an electrochemical property to adsorb sodium onto their surface.¹⁰ However, this storage property has low durability, which might affect the overall battery performance. Hence, the absence of carbon also unveils the actual electrochemical outcome, facilitating a high durability. Still, the rate performance of these carbon and binder-free systems requires an improved capacity. Studies conducted by Tarascon's group³ and by Yi Cui's group¹⁴ have shown that columnar-type morphologies can give more surface area of the active materials. Modifying such ideas, an electrode can be sculpted into a monolithic form, which reduces the volume of the extended current collector while maintaining well-defined particle–particle connections. In addition, ample space is required for the electrolyte to avail the active material interface for storage reactions. This can be achieved by making a nano-pillar grown perpendicular to the current collector.¹⁵ In a nutshell, a fossilized columnar dendritic morphology^16,17 is ideal to generate a high energy-density along with an improved rate performance.

Presently, we report a monolithic dendritic TiO₂ material as an anode for sodium-ion batteries, for the first time. Using a one-step aerosol chemical vapor deposition (ACVD) process, we have grown the dendritic morphology on a stainless-steel current collector, and further used it as an electrode. Upon rate cycling followed by a prolonged run until 1000 cycles, the retention of the morphology can been observed, depicting the structural benefit of the procured structure. Here, long-term cycling until 1000 cycles is reported for the first time for monolithic TiO₂, in the case of SIBs. Interestingly, the one-dimensionally grown anatase TiO₂ explicitly reveals a phase change to rutile upon cycling. In the current study, the phase change has been looked into in detail via ex situ XRD, TEM, XPS and Raman studies. For the first time in the case of a sodium-ion anode, a rotational moiré fringe has been observed upon sodium-ion insertion. We believe that this concept involving the deployment of one-step electrode fabrication towards a tolerable dendritic morphology will enrich the next generation of research on metal-ion batteries.

Experimental section

Synthesis of the TiO₂ electrode

TiO₂ nanostructures were synthesized using a single-step ACVD process.^17,18 Vapors of titanium tetraisopropoxide at a concentration of 1.29 × 10⁻³ mol m⁻³ were introduced into the reactor, where they thermally decomposed to form TiO₂ molecules, which nucleated and grew to form particles. These particles were deposited onto a 1 cm diameter stainless steel substrate (25 μm thickness) maintained at 500 °C, on which the particles sintered to form 1D columnar nanostructures. The deposition was carried out for 30 min.

Material characterization

The morphology of the nanostructures was analyzed using field emission gun scanning electron microscopy (FEGSEM, NOVA NanoSEM 230, FEI Co.), and their crystal structure was characterized using X-ray diffraction (XRD) (Bruker D8 Advance) with CuKα radiation (wavelength = 1.5406 Å) at 35 kV and 35 mA. The scattering angle (2θ) was from 20° to 60°, with a step size of 0.02° and a dwell time of 2 s. Rietveld refinement to obtain the lattice parameters from the XRD spectra was carried out using Topas 5 software (Bruker). Before and after cycling, detailed crystal orientation, selected area electron diffraction (SAED) and morphology studies were done with high-resolution transmission electron microscopy (HR-TEM, JEOL J2100F) operated at 200 kV.

Electrochemical cell fabrication and measurements

Anatase TiO₂, deposited on a stainless-steel substrate with a diameter of 1 cm, was used as a working electrode (WE) to study the electrochemical behavior of TiO₂ as an electrode for Na-ion batteries. Galvanostatic charge–discharge tests were carried out using a lab scale Swagelok-type stainless steel setup, with a cell configuration of Na|electrolyte|TiO₂. All electrochemical cells were assembled in an argon-filled glovebox (Lab Star, MBraun, Germany) under controlled moisture and oxygen concentrations below 1 ppm. Na foil (Alfa Aesar, 99.8%) was used as the counter electrode as well as the reference electrode (CE/RE). Electrolyte preparation for this study was done inside a glove box. A solution (EC : PC) of ethylene carbonate (EC) (Sigma, 98%) and propylene carbonate (PC) (Sigma, 99.7% anhydrous), in a 1 : 1 volume ratio, was prepared. NaClO₄ salt (Sigma, 98%), which was vacuum dried for two days and inserted into the glove box, was dissolved into the EC : PC solution to achieve a 1 M concentration. This solution was used as the electrolyte. The separator was a porous borosilicate glass microfiber filter (Whatman), soaked with a few drops of the electrolyte. Electrochemical charge–discharge experiments were performed to evaluate the durability of the prepared cells, using an Arbin battery testing instrument (BT2000, USA) at various constant current rates. Potentiostatic electrochemical impedance spectroscopy (PEIS) was carried out for the first discharge cycle using a Bio-Logic VMP-3 electrochemistry workstation. The technique was performed within a frequency range of 1 MHz to 1 Hz and with a voltage amplitude of ΔV = 10 mV. Furthermore, the galvanostatic intermittent titration technique (GITT) was used to study the kinetics of Na-ion diffusion into the TiO₂ matrix. These electrochemical experiments were carried out at a constant current density of 50 mA g⁻¹ keeping a constant temperature of 20 °C.

Ex situ analysis

Ex situ analysis of the cycled electrodes was carried out using XRD, Raman spectroscopy, X-ray photoelectron spectroscopy (XPS), HR-TEM, SAED, and scanning transmission electron microscopy-energy dispersive X-ray spectroscopy (STEM-EDX) to examine the structural changes in the nanocolumns upon Na-ion insertion and de-insertion. Raman spectroscopy (using a HOBIBA LabRAM HR 800 spectrometer with a 532 nm solid-state laser source of 50 mW) confirmed the composition of the electrode using a 100× lens with a 10% edge filter. STEM-EDX analysis was carried out in the HR-TEM instrument at the operating parameters mentioned previously. Line scan EDX analysis was conducted along the length and the width of the column. XPS measurements were carried out using a PHI 5000 VersaProbe II equipped with a monochromatic Al Kα (1486.6 eV) X-ray source. Fine scans were done for Na, Ti, C, and O. XPS peak analysis was done using Multipack (v9.6, ULVAC-PHI), and the peaks were fitted to a Gaussian–Lorentzian mixed function via the iterated Shirley method. The peaks were calibrated using the peak for adventitious surface carbon at 284.80 eV, corresponding to C 1s photoemission. These ex situ experiments were carried out by cycling the cells at a current rate of 50 mA g⁻¹. In addition, a few more ex situ characterizations were performed after 1000 cycles, where the cell was in a de-sodiated state.

Results and discussion

Electrode fabrication and rate cycling

Carbon and binder-free monolithic anatase TiO₂ dendrites were directly grown on the stainless steel current collector in a single-step ACVD process as shown in Fig. 1a. We have previously elucidated the formation mechanism of different morphologies that can be synthesized by the ACVD process.^19,20 The procured morphology was columnar in structure, which favors adequate interfacial area for the electrolyte.¹⁷ The fabricated electrodes were subjected to discharge–recharge at different rates for the first 30 cycles (Fig. 1b). The as-prepared electro-active material exhibits 357 mA h g⁻¹ (2^nd cycle), 223 mA h g⁻¹ (6^th cycle), 180 mA h g⁻¹ (11^th cycle), 151 mA h g⁻¹ (16^th cycle), 115 mA h g⁻¹ (21^st cycle), and 89 mA h g⁻¹ (26^th cycle) discharge capacities and 180 mA h g⁻¹ (2^nd cycle), 177 mA h g⁻¹ (6^th cycle), 156 mA h g⁻¹ (11^th cycle), 136 mA h g⁻¹ (16^th cycle), 105 mA h g⁻¹ (21^st cycle), and 83 mA h g⁻¹ (26^th cycle) charge capacities at constant current rates of 20 mA g⁻¹ (0.06 C), 50 mA g⁻¹, 100 mA g⁻¹, 200 mA g⁻¹, 500 mA g⁻¹, and 1000 mA g⁻¹ (3 C), respectively. After the initial rate cycling, the performance was prolonged to 1000 cycles at a current rate of 100 mA g⁻¹. At the 1000^th cycle, the TiO₂ columns exhibited a discharge capacity of 120 mA h g⁻¹, thus retaining 85.1% capacity with respect to the 31^st cycle (141 mA h g⁻¹). At the end of 1000 cycles, the cell was disassembled and subjected to different physical characterizations. Here, we have compared the morphology of the TiO₂ columns after 1000 cycles with that of the as-prepared material. Subsequently, the retention of the morphology has been explicitly depicted through an SEM study (Fig. 1c). In addition, a TEM study also shows retention of the columnar morphology (Fig. 1d) after 1000 cycles when analyzing a single column. However, the single-crystalline TiO₂ becomes polycrystalline after cycling, which has been discussed in detail in the ESI (Fig. S2†).

Fig. 1 (a) Synthesis pathway, (b) cycling performance, and morphological comparison of TiO₂ through (c) SEM and (d) TEM for the as-prepared material and after 1000 cycles.

Electrochemistry

To evaluate the reactions at different voltages with respect to Na/Na⁺, the dQ/dV [d(capacity)/d(voltage)] for different cycles at varying current rates was plotted (Fig. 2a). Initial cycling and the GITT at the 2^nd cycle have been elaborated in the ESI (Fig. S3†). In this part, only extended cycling at 100 mA g⁻¹ has been described. Two prominent reduction peaks were observed during the reduction process in all cycles. The peak near 0.74 V vs. Na/Na⁺ can be attributed to Na-ion insertion into the TiO₂ host matrix. Vice versa, de-insertion of Na-ions from the TiO₂ was observed in the oxidation cycles near 0.78 V vs. Na/Na⁺.²¹ The as-prepared material exhibited an excellent reversibility of 99.55% at the 1000^th cycle (Table T1, ESI†). The potential difference obtained from the dQ/dV plot was observed to be 41 mV (at the 1000^th cycle), taking the Na-ion insertion peak at 0.738 V and the Na-ion de-insertion peak at 0.779 V. On average, the overall battery performance is nearly 0.76 V.

Fig. 2 (a) Charge–discharge performance and (b) dQ/dV plots on extended cycling.

Structural and phase change upon cycling

In this part, we have aimed to map the mechanism of reversible Na-ion storage into the single-crystalline anatase TiO₂. Here, ex situ XRD, XPS, Raman and TEM studies were performed on the cycled electrodes. Initially, ex situ XRD was performed in the discharged (0.05 V) and charged state (3 V) for the 1^st and 2^nd cycle. With the obtained X-ray diffraction patterns (Fig. 3a), the lattice parameters (a and c) were measured for the cycled electrodes and compared with those of pristine TiO₂. A slight increase in the lattice constants was observed for all of the cycled TiO₂ samples, as compared to the pristine TiO₂ samples (Fig. 3b). Going from pristine to the first sodiation, both unit cell parameters, a and c, increase from 3.7845 Å to 3.7868 Å and from 9.4957 Å to 9.5027 Å, respectively. This expansion is consistent with previously reported detailed in situ and ex situ XRD analyses for Li and Na-ion intercalation into anatase TiO₂.^22,23 After the first insertion, the lattice parameters changed with a well-defined trend. After charging, a increases and c decreases, while the opposite phenomenon happened after discharge. Here, the lattice constants of the tetragonal crystal structure (I4₁/amd) were obtained via Rietveld refinement of the XRD spectra of the pristine TiO₂ and cycled TiO₂ samples. For this typical calculation, the anatase standard diffraction pattern was taken from ICSD PDF no. 01-071-1166 and the rutile standard diffraction pattern was taken from ICSD PDF no. 04-001-7096. The ex situ XPS study on the cycled electrodes (Fig. 3c) shows the presence of Ti³⁺ after discharge, which disappeared after charging. The peak at 457.64 eV corresponds to the Ti³⁺ binding energy, while the peak at 458.53 eV corresponds to the Ti⁴⁺ binding energy.²⁴ This reflects the Na-ion insertion into the lattice, which is quite similar to Li electrochemistry.²⁵ Interestingly, the Raman study after 1000 cycles (Fig. 3d) shows an additional phase of rutile subsumed with anatase. The ex situ Raman study is discussed in the ESI (Fig. S5†). The presence of rutile was also depicted in the X-ray diffraction pattern (Fig. 3d inset), in which the (1 1 0) plane of the rutile phase was explicitly observed. It was assessed that the change in lattice parameters upon cycling and the occurrence of the rutile phase can culminate in a mechanistic pathway of Na-ion insertion and de-insertion into the TiO₂ matrix.

Fig. 3 Ex situ (a) XRD and corresponding (b) lattice parameter analyses, and (c) XPS of the nanostructured TiO₂ anode during the first two charge and discharge cycles. (d) Ex situ Raman after 1000 cycles; inset: XRD after 1000 cycles.

To scrutinize them in detail, the first sodiated and de-sodiated electrodes were subjected to TEM analysis. Initially, a mapping study was performed on these two electrodes. For this typical study, the surface adsorbed sodium compounds were removed with water, which has been detailed in the ESI (Fig. S6†). SETM-EDS mapping of Na, Ti and O was carried out. This shows that Na was well distributed throughout the column during discharging and purged out in a uniform manner after charging. To support this, line scans were performed with the same single columns across the width and length of the column. This also depicts a similar distribution. In the next step, the SAED pattern of these electrodes was compared, fixing the zone axis to the (h 0 l) direction. A fascinating outcome was noticed explicitly through this study. For the first time, a rotational moiré fringe pattern was observed on the column surface for the sodiated samples (Fig. 4c), suggesting the presence of two overlaying lattices. The observed rotation for this typical case was 4.5°. This phenomenon occurred after Na-ion insertion into the TiO₂ matrix and disappeared after charging. This part has been further described in the ESI (Fig. S7†).

Fig. 4 STEM-EDX mapping of TiO₂ columns after (a) the 1^st discharge and (b) the 1^st charge along with the line scan data for the same columns and (c) the SAED pattern and HR-TEM image to evaluate the Na-ion storage.

Brief discussion

Initially, this study focused on the morphological benefit of the monolithic dendrites when used as a carbon and binder-free electrode for Na-ion batteries. We have recently reported that a significant capacity is contributed by conductive carbon in the potential range of 0.01–1 V. Therefore, in the current study, we have used a carbon-binder free material to judge the actual capacity contribution from the electrode. The advantage of this morphology is the well-defined connection between the grown columns and the current collector, and the separation between the columns helps the electrolyte to access the electrolyte–electrode interface. Hence, a homogeneous surface reaction occurs (observed through STEM-EDS mapping, Fig. 4a and b). However, these additive-free electrodes need an initial activation (reaction) to rearrange the crystallinity, facilitating easy Na-ion insertion and de-insertion. This assessment is reflected in an improvement in capacity when cycling was initially performed at a very slow rate (20 mA g⁻¹) in comparison with a higher rate (discussed in the ESI, Fig. S1†). At the time of slow rate cycling, for the first two cycles, the obtained capacity output was higher than the theoretical capacity (335 mA h g⁻¹),¹¹ which is most likely due to the surface adsorption of Na-compounds onto the surface. This appraisal has been reflected in the STEM-EDS mapping, XPS and PEIS study (discussed in the ESI, Fig. S4†). In addition, a stable electrolytic resistance (∼13.35 Ω) was obtained from the PEIS study (discussed extensively in the ESI, Fig. S4†). After the first discharge, the ex situ SEM does not show any additional morphology, as observed by Passerini's group.²⁵ This observation indicates that the additional morphology comes from the reaction with the conductive carbon, which is in agreement with our previous study on conductive carbon.²⁶ Hence, this carbon-free system does offer an easy-fueling of Na-ions at the interface, which resulted in long-term cycling up to 1000 cycles. Interestingly, the GITT study shows that the diffusion co-efficient (8.54 × 10⁻¹⁸ cm² s⁻¹) is ∼20 times higher than that for the previously reported additive-free TiO₂ (discussed extensively in the ESI, Fig. S3d†).¹⁵ Although the overall electrochemical outcome is similar to that of lithium, it differs significantly in the operational voltage. Generally, TiO₂ performs near 1.5–1.6 V vs. Li/Li⁺,¹⁷ while here the same material performs at much lower potential (∼0.76 V). It is well known that a low-voltage anode results in enhanced energy-density and power-density.² Therefore, the removal of unnecessary components from the electrode volume and the lower potential operation significantly improve the volumetric energy-density, which is a rudimentary requisite to ameliorate sodium electrochemistry. In comparison with similar reports on TiO₂, our anatase dendrites perform much better in terms of capacity. Nearly 30% improvement in specific capacity has been observed, which has been detailed in Fig. S8 (ESI†).

As the material is 1D-grown anatase TiO₂, the structural and phase changes were noticed explicitly through ex situ studies. The insertion of Na-ions into the TiO₂ matrix affects the (h 0 l) direction and produces a distorted lattice, which is reflected as a lattice rotation (depicted as a moiré fringe in SAED, Fig. 4c). Probably, an amount of strain gets induced into the TiO₂ lattice. During oxidation (de-insertion of Na-ions), the material seeks to retain its original anatase structure. Interestingly, the crystalline structure does not convert back to the exact pristine TiO₂. This phenomenon has been observed as an initial increment in the lattice parameters (a and c) that does not decrease to its original value even in the de-sodiated state. Initially, the material induces strain due to sodiation, which is reflected as a moiré pattern. After de-sodiation, the strain releases and the material incorporates the rutile phase along with anatase. Upon cycling, the single-crystalline TiO₂ becomes polycrystalline (discussed extensively in the ESI, Fig. S2†) and keeps the additional rutile phase, retaining the initial morphology. As a result, after de-sodiations (after the 1^st, 2^nd, 1000^thetc.) the X-ray diffraction patterns and Raman spectra show the presence of the rutile phase along with the anatase phase.

Conclusion

In summary, for the first time, a monolithic anatase TiO₂ electrode has been proposed as a binder and conductive carbon-free anode for sodium-ion batteries. Here, TiO₂ exhibits a long cycle life of 1000 cycles including initial rate cycling, without changing its preparatory morphology. The retention in capacity is 81.5% with a high coulombic efficiency of 99.55% (at the 1000^th cycle). Na-ion insertion and de-insertion happen near 0.76 V, which is a significant advantage of the Na-ion battery anode over currently used lithium systems. Ex situ SEM, TEM, XRD and XPS confirms the additional phase incorporation of rutile after cycling. From SAED, a rotational moiré pattern has been observed, for the first time, to describe Na-ion insertion into the anode system. Our collaborative study on single-step monolithic additive-free electrode fabrication at both Washington University in St. Louis and the Indian Institute of Technology Bombay, addressing long-term durability, is on its way to monetization in attenuated devices like pacemakers, smart watches, etc. Currently, we are working on further improving the specific capacity by introducing significant dopants in a cost-effective way and developing a scaled-up roll-to-roll ACVD process for large-scale manufacturing and maintenance of environmental benignity.

Author contributions

TSC, PKD, SM and PB designed the study. TSC and PKD carried out the synthesis of the nanostructured anode and the electrochemical characterization. RR aided with the XPS and TEM measurements. TSC and PKD wrote the manuscript, SM and PB revised it and all authors approved the final version before submission. SM and PB provided the overall supervision on this project.

Conflicts of interest

There are no conflicts of interest to declare.

Acknowledgements

This paper is based upon work supported by the Solar Energy Research Institute for India and the U.S. (SERIIUS) funded jointly by the U.S. Department of Energy subcontract DE AC36-08G028308 (Office of Science, Office of Basic Energy Sciences, and Energy Efficiency and Renewable Energy, Solar Energy Technology Program, with support from the Office of International Affairs) and the Government of India subcontract IUSSTF/JCERDC-SERIIUS/2012 dated 22nd Nov. 2012. The authors acknowledge Dr Huafang Li at the Institute of Material Science and Engineering (IMSE), Washington University in St. Louis for help with STEM measurements and Professor James C. Ballard for reviewing the manuscript.

Notes and references

1. M. D. Slater, D. Kim, E. Lee and C. S. Johnson, Adv. Funct. Mater., 2013, 23, 947–958.
2. V. Palomares, P. Serras, I. Villaluenga, K. B. Hueso, J. Carretero-Gonzalez and T. Rojo, Energy Environ. Sci., 2012, 5, 5884–5901.
3. P. L. Taberna, S. Mitra, P. Poizot, P. Simon and J. M. Tarascon, Nat. Mater., 2006, 5, 567–573.
4. Y. Xu, Q. Wei, C. Xu, Q. Li, Q. An, P. Zhang, J. Sheng, L. Zhou and L. Mai, Adv. Energy Mater., 2016, 6, 1600389.
5. Y. Jiang, S. Yu, B. Wang, Y. Li, W. Sun, Y. Lu, M. Yan, B. Song and S. Dou, Adv. Funct. Mater., 2016, 26, 5315–5321.
6. J. O. Besenhard and G. Eichinger, J. Electroanal. Chem. Interfacial Electrochem., 1976, 68, 1–18.
7. X. Han, M. Ouyang, L. Lu and J. Li, Energies, 2014, 7, 4895–4909.
8. W. M. Zhang, J. S. Hu, Y. G. Guo, S. F. Zheng, L. S. Zhong, W. G. Song and L. J. Wan, Adv. Mater., 2008, 20, 1160–1165.
9. D. Wang, D. Choi, J. Li, Z. Yang, Z. Nie, R. Kou, D. Hu, C. Wang, L. V. Saraf and J. Zhang, ACS Nano, 2009, 3, 907–914.
10. J.-Y. Hwang, S.-T. Myung and Y.-K. Sun, Chem. Soc. Rev., 2017, 46, 3529–3614.
11. Y. Zhang, C. W. Foster, C. E. Banks, L. Shao, H. Hou, G. Zou, J. Chen, Z. Huang and X. Ji, Adv. Mater., 2016, 28, 9391–9399.
12. G. Zou, H. Hou, P. Ge, Z. Huang, G. Zhao, D. Yin and X. Ji, Small, 2018, 14, 1702648.
13. D. Prutsch, M. Wilkening and I. Hanzu, ACS Appl. Mater. Interfaces, 2015, 7, 25757–25769.
14. C. K. Chan, H. Peng, G. Liu, K. McIlwrath, X. F. Zhang, R. A. Huggins and Y. Cui, Nat. Nanotechnol., 2008, 3, 31.
15. H. Xiong, M. D. Slater, M. Balasubramanian, C. S. Johnson and T. Rajh, J. Phys. Chem. Lett., 2011, 2, 2560–2565.
16. S. Wenzel, T. Hara, J. Janek and P. Adelhelm, Energy Environ. Sci., 2011, 4, 3342–3345.
17. T. S. Chadha, A. M. Tripathi, S. Mitra and P. Biswas, Energy Technol., 2014, 2, 906–911.
18. T. S. Chadha, P. Biswas and W.-J. An, US Pat., Application No. 14/830,361, 2016.
19. T. S. Chadha, M. Yang, K. Haddad, V. B. Shah, S. Li and P. Biswas, Chem. Eng. J., 2017, 310, 102–113.
20. K. Haddad, A. Abokifa, S. Kavadiya, T. Chadha, P. Shetty, Y. Wang, J. Fortner and P. Biswas, CrystEngComm, 2016, 18, 7544–7553.
21. Z. Le, F. Liu, P. Nie, X. Li, X. Liu, Z. Bian, G. Chen, H. B. Wu and Y. Lu, ACS Nano, 2017, 11, 2952–2960.
22. K.-T. Kim, G. Ali, K. Y. Chung, C. S. Yoon, H. Yashiro, Y.-K. Sun, J. Lu, K. Amine and S.-T. Myung, Nano Lett., 2014, 14, 416–422.
23. R. van de Krol, A. Goossens and E. A. Meulenkamp, J. Electrochem. Soc., 1999, 146, 3150–3154.
24. J. Huang, D. Yuan, H. Zhang, Y. Cao, G. Li, H. Yang and X. Gao, RSC Adv., 2013, 3, 12593–12597.
25. L. Wu, D. Bresser, D. Buchholz, G. A. Giffin, C. R. Castro, A. Ochel and S. Passerini, Adv. Energy Mater., 2015, 5, 1401142.
26. P. K. Dutta and S. Mitra, J. Power Sources, 2017, 349, 152–162.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c8se00082d

‡ These authors contributed equally.

---