Facile synthesis of cobalt fluoride (CoF₂)/multi-walled carbon nanotube (MWCNT) nanocomposites and improvement of their electrochemical performance as cathode materials for Li-ion batteries†

Hayoung Park ^ab, Il-Seop Jang ^ab, Bo-Ye Song ^ab, Yun Chan Kang ^b, Seongseop Kim *^c and Jinyoung Chun *^a
aEmerging Materials R&D Division, Korea Institute of Ceramic Engineering and Technology (KICET), Jinju 52851, Republic of Korea. E-mail: jchun@kicet.re.kr
^bDepartment of Materials Science and Engineering, Korea University, Seoul 02841, Republic of Korea
^cSchool of Chemical Engineering, Clean Energy Research Center, Jeonbuk National University, Jeonju 54896, Republic of Korea. E-mail: seongseopkim@jbnu.ac.kr

First published on 14th June 2023

---

---

Seongseop Kim

Dr Seongseop Kim is an assistant professor in the School of Chemical Engineering at Jeonbuk National University (JBNU). He received his B.S. degree in Chemical Engineering from Hanyang University (2013), and PhD degree in Chemical Engineering from POSTECH (2019). During the PhD course, he joined Rensselaer Polytechnic Institute as a visiting student. After his PhD, he worked as a postdoctoral research at KAIST and Max Planck Institute für Kohlenforschung. His research interests focus on the design of nano- and meso-porous inorganic materials for electrocatalysis and energy applications.

Introduction

Cathode materials for Li-ion batteries (LIBs) with high energy-densities are essential to meet the increased energy consumption of portable devices, rapid growth of the electric vehicle industry, and increasing demand for advanced energy storage systems. Intercalation-based high-Ni cathode materials, such as LiNi_0.8Co_0.15Al_0.05O₂ (NCA) and LiNi_0.8Co_0.1Mn_0.1O₂ (NCM811), have attracted significant attention because of their improved energy densities due to the increased availability of lithium ions.^1–3 However, the number of lithium ions available per unit molecule of active material is still limited, having a theoretical capacity of approximately 250 mA h g⁻¹, thus restricting the achievement of the energy density required by modern industry. Therefore, to solve the intrinsic capacity limitation issues of lithium-ion intercalation materials, researchers have focused on new electrode materials that react with lithium ions through conversion or alloying reactions, which can take up multiple lithium ions per unit formula.^4–9

Among the metal compounds that undergo conversion reactions, such as metal oxides, nitrides, sulfides, and fluorides, metal fluorides (MF_x) are a promising cathode candidate for LIBs.^10–17 MF_x exhibit high operating voltages because of the strong ionic characteristics of the M–F bond,¹⁰ as well as large gravimetric and volumetric specific capacities, as 2–3 lithium ions can be stored per transition metal atom (redox center, e.g., Fe, Ni, Cu, and Co) in MF_x cathodes.¹¹ However, utilizing the potentially high energy density of MF_x has been a challenge over the last several decades owing to their synthesis difficulties and inherent chemical and electrochemical properties.^12–14,18

Cobalt fluoride (CoF₂), a member of the MF_x family, provides high theoretical specific capacities (553 mA h g⁻¹ and 2038 mA h cm⁻³) that relate to a two-electron transfer per formula unit of CoF₂ and promising energy densities (828 W h kg⁻¹ and 1565 W h L⁻¹) in CoF₂–Li batteries (Fig. S1†).^19,20 However, CoF₂ also has the common drawbacks of MF_x cathode materials: (i) limited synthesis and nanostructuring methods, (ii) low electronic conductivities originating from high-ionic bonding characteristics and a large bandgap (4.44 eV), (iii) large voltage hysteresis caused by conversion reactions, and (iv) dissolution of metal ions at high voltages.

Among these challenges, the low electronic conductivity and poor kinetics of bulk-sized materials are critical drawbacks of using CoF₂ as a cathode for LIBs. Hence, the introduction of nanostructures and the formation of composites with conducting agents are essential preconditions for the application of CoF₂ as a cathode material. However, the synthesis methods developed in previous studies have several disadvantages, such as complicated manufacturing processes,²¹ expensive precursors,^21,22 and the need for heat treatment facilities to handle high-risk fluorine-based gases.²³ In several studies, the relatively large particle size of CoF₂ resulted in poor electrochemical performance.^24–26 Although the preparation of designed CoF₂ nanocomposites incorporated in conducting carbons and respective improved electrochemical properties have been reported recently,^23,27,28 a large portion of inactive carbon components (approximately 40 wt% or more) and the requirement of complex synthetic processes are challenges remaining to be addressed. In addition, the rapid dissolution of transition metal cations at high voltages is observed during the conversion reactions of CoF₂. This phenomenon significantly reduces the reversibility of reactions and cause fatal side reactions within the cell, thus hindering the long cycle life of batteries.^23,29 Therefore, to improve the cycle performance of MF_x, electrolyte optimization and surface modification of the electrodes to prevent the dissolution of active materials should be conducted simultaneously.

In this study, a facile synthesis method for CoF₂ nanocomposites was developed and additional in situ and ex situ surface protection strategies were introduced to improve their electrochemical performance. CoF₂/multi-walled carbon nanotube (MWCNT) nanocomposites were prepared by simple precipitation of cobalt precursors onto MWCNTs, followed by heat treatment with ammonium fluoride (NH₄F). During this synthesis process, special facilities for high-pressure solvothermal reactions and handling of toxic gases are not necessary. Consequently, the synthesis procedure is not only simple and versatile but also relatively safe. Compared with the CoF₂ particles obtained in the absence of MWCNTs, CoF₂/MWCNTs exhibited improved specific capacities through conversion reactions and superior rate capabilities at high current densities. In addition, the in situ formation of a cathode electrolyte interphase (CEI) layer using a fluoroethylene carbonate (FEC)-containing electrolyte and ex situ coating of an aluminium oxide (Al₂O₃) thin layer on the electrode by atomic layer deposition (ALD) effectively mitigated rapid capacity decay originating from metal ion dissolution. This study contributes to the development of advanced MF_x cathodes based on conversion reactions for high-energy-density batteries with long cycle lives.

Experimental section

Synthesis of CoF₂/MWCNT nanocomposites

Co(NO₃)₂·6H₂O (97.0%, Junsei Chemical), NH₄HCO₃ (95.0%, Junsei Chemical), and NH₄F (97.0%, Junsei Chemical) were used as received, without further purification. MWCNTs (MR99, Carbon Nano-material Technology Co., Ltd) were used after acid treatment by following a previously reported procedure.³⁰ Cobalt precursor–MWCNT nanocomposites were obtained using a precipitation method. Co(NO₃)₂·6H₂O ethanolic solution (5 mmol) was prepared, and MWCNTs were added, at a mass ratio of CoF₂:MWCNTs of 7:3. MWCNTs were dispersed by ultrasonication. Subsequently, NH₄HCO₃ was dissolved in the solution and stirred for 12 h to form the cobalt precursor. The products were washed with deionized water to remove impurities and dried at 80 °C in a vacuum oven. The resultant powder was mechanically ground with NH₄F at a mass ratio of 1:1 using a mortar and pestle. The mixture was heat treated at 250 °C for 2 h under a nitrogen atmosphere to prevent oxidation. The final product was denoted as CoF₂/MWCNTs. For comparison, CoF₂ particles were synthesized using only cobalt precursors and NH₄F in the absence of MWCNTs and denoted as bare-CoF₂.

Al₂O₃ coatings onto the electrodes through atomic layer deposition (ALD)

Al₂O₃ coating on the CoF₂/MWCNT electrode was carried out at 120 °C with a growth rate of 0.1 nm per cycle using an ALD system (Lucida D100, NCD). Trimethylaluminum (TMA, Al(CH₃)₃) and H₂O were used as the precursors and a complete ALD cycle included: (1) pulsing TMA for 0.2 s, (2) purging the chamber for 10 s, (3) pulsing H₂O for 0.2 s, and (4) purging the chamber for 10 s. The CoF₂/MWCNT electrodes were coated with Al₂O₃ through 20 or 50 ALD cycles, and named 20A- and 50A-CoF₂/MWCNTs, respectively.

Materials characterization

The crystal structure of the prepared samples was obtained by powder X-ray diffraction (XRD; D8 Advance, Bruker) using Cu Kα radiation (λ = 1.5418 Å) in the diffraction angle (2θ) range of 10–80°. The morphologies were observed using field-emission scanning electron microscopy (FE-SEM; JSM-7610F, JEOL) at a working voltage of 10.0 kV, and field-emission transmission electron microscopy (FE-TEM; JEM-4010, JEOL) at a working voltage of 300 kV. The elemental mapping results were investigated by FE-TEM (Tecnai G2 F30, FEI). The chemical bonding states were analyzed using X-ray photoelectron spectroscopy (XPS; Nexsa, Thermo Fisher Scientific) with Al Kα radiation (1486.6 eV). The XPS depth profile analysis was performed by etching with an Ar ion beam. The MWCNT content was measured by thermogravimetric analysis (TGA; STA7300, Hitachi) under an air atmosphere at a heating rate of 10 °C min⁻¹.

Electrochemical characterization

To prepare the electrode, the active material, conducting carbon (Super P), and polyvinylidene difluoride (PVDF) were mixed in a weight ratio of 8:1:1 with N-methyl-2-pyrrolidone (NMP) using a Thinky mixer (ARE-310, THINKY Co.). The resulting slurries were coated onto aluminum foils using the doctor-blade method. The electrode was dried in a vacuum oven at 80 °C overnight and punched into 12-pi circular discs for use as the working electrode. Coin-type cells (CR2032) were assembled in a dry room, in which Li foil was used as the reference and counter electrodes. A polypropylene (PP) film (Celgard 2500) was used as the separator and 1.0 M LiPF₆ in ethylene carbonate/dimethyl carbonate (EC/DMC, 1:1 volume ratio, Enchem Co.) and 1.0 M LiPF₆ in fluoroethylene carbonate/ethyl methyl carbonate (FEC/EMC, 30:70 volume ratio, Enchem Co.) were used as the electrolyte. Using a WBCS-3000 battery cycler (WonATech Co.), cyclic voltammetry (CV) and galvanostatic charge/discharge analyses were conducted in the potential range of 1.0 to 4.0 V (vs. Li/Li⁺). The specific capacities and C-rates were calculated based on the masses of the active materials. The rate of 1C corresponded to a current density of 553 mA g⁻¹. Electrochemical impedance spectroscopy (EIS) measurements were performed using a potentiostat (SP-300, BioLogic) over the frequency range between 0.1 Hz and 1 MHz. For galvanostatic intermittent titration technique (GITT) analysis, a current pulse of 100 mA g⁻¹ was applied for 1 h, followed by a relaxation period of 10 h.

Results and discussion

Synthesis of CoF₂/MWCNT nanocomposites and their electrochemical properties

The procedure developed for the synthesis of CoF₂/MWCNTs is shown in Fig. 1. Through the precipitation method using NH₄HCO₃, nano-sized Co precursors were well distributed on the MWCNTs (Fig. S2a and b†). After mechanical grinding with NH₄F and heat treatment, the Co precursor/MWCNTs were easily converted to CoF₂/MWCNT nanocomposites. Through these processes, CoF₂/MWCNTs can be effectively prepared without complicated synthetic procedures or the direct use of hazardous reagents (e.g., fluorine gas and hydrogen fluoride solutions).

The obtained XRD patterns of CoF₂/MWCNTs and bare-CoF₂ are shown in Fig. 2a. All diffraction peaks matched those of tetragonal CoF₂ (JCPDS No. 00-033-0417) and no other impurity peaks were observed. A small shoulder peak was observed at approximately 26° because of the carbon substrates, which demonstrates the presence of MWCNTs in CoF₂/MWCNTs (Fig. S2c†). The XRD patterns of bare-CoF₂ also presented the same CoF₂ crystal structure without the MWCNT peak. Considering that the XRD peaks of CoF₂/MWCNTs and bare-CoF₂ appeared broad rather than narrow, the crystallite size was assumed to be small. SEM and TEM analyses were performed to confirm the particle size and morphology. In the case of CoF₂/MWCNTs, nano-sized (10–15 nm) CoF₂ particles were spread on the MWCNTs (Fig. 2b, d and e). Compared with Co precursors/MWCNTs (Fig. S2a and b†), the morphology of CoF₂ nanoparticles in CoF₂/MWCNTs was well maintained after heat treatment. By contrast, bare-CoF₂ consisted of randomly aggregated particles composed of 20 nm-sized primary particles (Fig. 2c and f). These results show that the synthetic method used in this study was effective in preparing nanocomposites in which small-sized CoF₂ nanoparticles were well-dispersed on the conductive nanocarbons.

The presence of CoF₂ nanoparticles in the synthesized nanocomposite was further confirmed by elemental mapping of CoF₂/MWCNTs (Fig. 3). The mapping signals confirmed that the nanoparticles on the MWCNTs consisted of cobalt and fluorine. XPS analysis was performed to confirm the composition of the nanocomposites (Fig. 4). The high-resolution Co 2p XPS spectrum exhibited two distinct peaks at 783.9 and 800.0 eV for Co²⁺ 2p_3/2 and Co²⁺ 2p_1/2, respectively (Fig. 4a) and an F 1s peak located at 686.0 eV which are characteristic to CoF₂.³¹ Regarding the C 1s spectrum, the peak located at 284.6 eV corresponds to C–C bonding from the MWCNTs. The small peak observed at 286.0 eV is associated with the contribution of C–O functionalities. These functional groups containing oxygen atoms provide nucleation sites for cobalt precursors and prevent excessive particle aggregation.^32,33

The loading amount of CoF₂ in the CoF₂/MWCNTs was confirmed by TGA (Fig. S3†). CoF₂/MWCNTs were fully converted into Co₃O₄ through heat treatment at 700 °C under an air atmosphere, thus leading to a mass loss of approximately 43%. Considering the mass changes originating from the oxidation of CoF₂ to Co₃O₄ and the decomposition of MWCNTs simultaneously, the loading amount of CoF₂ calculated from the TGA results was approximately 68 wt%, which was consistent with the target value (70 wt%). This result indicates that heat treatment with NH₄F in a solventless system is effective for the fluorination of cobalt precursors without product loss.

The electrochemical reactions of CoF₂/MWCNTs were examined using CV tests at a scan rate of 0.1 mV s⁻¹ between 1.0 and 4.0 V (vs. Li/Li⁺) (Fig. 5a). A large intensity reduction peak was observed in the voltage range of approximately 1.12 V, which corresponds to the lithiation of CoF₂ through conversion reactions.^25,34 Next, small anodic peaks appeared at 3.1–3.2 V, and the small cathodic peaks were displayed at 2.0–2.1 V. In contrast to the first reduction peak, a pair of broad anionic and cathodic peaks were observed. The voltage difference between the first reduction and oxidation peaks of CoF₂/MWCNTs was approximately 2.0 V and decreased to 1.1–1.2 V as the battery cycled. These are unique characteristics of the conversion reactions: as CoF₂ converted to Co and LiF, the crystallite sizes of each component significantly decreased.^24,35 Next, reconversion/conversion reactions occurred in these small crystallite sizes; hence, broad anodic/cathodic peaks were observed after the first reduction step, and the redox potential difference decreased. On the other hand, it was observed that the potential difference slightly increased in the 3rd cycle compared to the 2nd cycle. The dissolution of metallic Co during cycling led to the accumulation of insulating LiF, which is detrimental to the electronic and ionic transport in the electrode.³⁶ Therefore, it is assumed that partial dissolution of metallic Co led to the slight increase of the potential difference in the 3rd cycle. The dissolution of metallic Co during cycling will be further discussed in the next section.

The voltage profiles of CoF₂/MWCNTs are shown in Fig. 5b. The first wide voltage plateau at 0.2C appeared above 1.3 V with a discharge capacity of 799 mA h g⁻¹, which is higher than the theoretical capacity of CoF₂ (553 mA h g⁻¹). The extra capacity can be attributed primarily to the interfacial Li storage within the Co/LiF matrix.²⁴ The charging (delithiation) capacity in the first cycle, that is, the reversible capacity, was 554 mA h g⁻¹ at a cutoff voltage of 4.0 V. The sloping voltage curves after the initial lithiation are comparable to the broad peaks of CoF₂/MWCNTs in the CV curves. Conversely, bare-CoF₂ exhibited much lower discharge and charge capacities of 612 and 338 mA h g⁻¹, respectively, in the first cycle (Fig. 5b). The polarization between the discharge and charge voltages of bare-CoF₂ also increased (Fig. 5b and S4†). In addition, the discharge capacity of bare-CoF₂ rapidly decreased to 343 mA h g⁻¹ in the second cycle, whereas CoF₂/MWCNTs retained a discharge capacity of 613 mA h g⁻¹. Fig. 5c and d present the rate capabilities of CoF₂/MWCNTs and bare-CoF₂, respectively. CoF₂/MWCNTs exhibited excellent rate performance (Fig. 5c), with a discharge capacity at 2C of 650 mA h g⁻¹. The first charge capacity of CoF₂/MWCNTs at 2C was 469 mA h g⁻¹, which corresponds to a reversible capacity of 84.7% at 0.2C. Table S1† shows the reversible capacity of CoF₂/MWCNTs at 2C in comparison to that of other CoF₂-based cathode materials reported previously,^{19,21–23,28,31,35} which demonstrates the superior rate properties of CoF₂/MWCNTs. As CoF₂ nanoparticles were introduced in MWCNTs, ion/electron conduction was facilitated and the reversibility and kinetics of the conversion reaction of CoF₂ improved. However, bare-CoF₂ exhibited an immediate capacity decrease as the C-rate increased, and only 0.3% of the capacity was maintained at 2C compared with the discharge/charge capacity at 0.2C (Fig. 5d).

The overall results indicate that the reversible capacities of CoF₂/MWCNTs were improved, and the rate performance of CoF₂/MWCNTs was considerably enhanced by the addition of MWCNTs and the nanosizing of CoF₂ particles. Nanosized CoF₂ particles and conductive MWCNTs efficiently addressed the intrinsic insulating issue of CoF₂ and the poor kinetics of the conversion reactions. This was further demonstrated by EIS measurements and GITT analysis. Fig. 6a shows the EIS results for CoF₂/MWCNTs and bare-CoF_2. The diameter of the semicircular arc corresponds primarily to the electrochemical charge-transfer resistance (R_ct) of the lithium-ion redox reaction at the electrode interface. The small semicircle of CoF₂/MWCNTs indicates improved kinetics of the electrochemical reaction. Electrons are efficiently transferred through the MWCNTs, and the nanoparticles facilitate the conversion reactions of CoF₂.^23,37 GITT was applied to analyze the overall potential response of CoF₂/MWCNTs and bare-CoF₂ during the first cycle (Fig. 6b). When the current supply was interrupted, the closed-circuit voltage (CCV) during the current pulse changed to quasi-open-circuit-voltage (QOCV) at the end of the relaxation period owing to internal resistance. The internal resistance was calculated by substituting the voltage difference between OCV and QOCV into the equation of Ohm's law (Fig. 6c).³⁸ The overpotential of CoF₂/MWCNTs was less than that of bare-CoF₂, demonstrating that the conversion reaction kinetics were improved when the small-sized CoF₂ nanoparticles were well-dispersed on the conductive MWCNTs. Note that both CoF₂/MWCNTs and bare-CoF₂ exhibited a decreasing resistance as the CCV was lowered. This trend was likely owing to the formation of conductive Co metal nanoparticles in the composite electrode by the conversion reaction (CoF₂ + 2Li⁺ + 2e⁻ → Co⁰ + 2LiF).³⁹

Cycle performance tests were conducted at 0.2C (111 mA g⁻¹). To closely observe the capacities after the first cycle, the discharge capacities were plotted except for the first cycle (Fig. S5a†). Compared with bare-CoF₂, CoF₂/MWCNTs exhibited better cycle performance, maintaining a capacity of 98 mA h g⁻¹ over 30 cycles, which is nearly three times higher than that of bare-CoF₂. The discharge capacity increased throughout the cycle owing to the addition of the electroconductive media. Furthermore, the discharge capacity of pure MWCNTs in the 30th cycle was only 19 mA h g⁻¹, considering that the mass ratio of CoF₂ to MWCNTs was 70:30 (Fig. S5b†). Thus, MWCNTs serve as conductive materials that facilitate the efficient conversion reaction of CoF₂ in the CoF₂/MWCNT nanocomposites. In particular, the capacity of CoF₂/MWCNTs was attributed primarily to CoF₂. Because the theoretical volume expansion of CoF₂ is not very large (∼21 vol%),²⁰ severe cracks or volumetric changes were not observed on the CoF₂/MWCNT and bare-CoF₂ electrodes (Fig. S6†). However, because rapid capacity fading occurred owing to Co²⁺ dissolution into the electrolyte during cycling,¹⁶ further research to prevent capacity degradation was conducted and will be presented in the next section.

Electrolyte modification for in situ formation of the CEI layer

The cycle stability of CoF₂/MWCNTs in different liquid electrolytes was investigated. When an FEC-containing electrolyte was used, the cyclability of CoF₂/MWCNTs increased significantly. As shown in Fig. 7a, CoF₂/MWCNTs with the FEC/EMC electrolyte exhibited a discharge capacity of 241 mA h g⁻¹ in the 30th cycle and retained a high capacity of 180 mA h g⁻¹ in the 100th cycle. This is approximately 5.5 times higher than the 100th discharge capacity in the EC/DMC electrolytes (33 mA h g⁻¹). These results indicate that the FEC/DMC electrolyte mitigated the capacity degradation of CoF₂/MWCNTs more effectively.

Fig. 7b shows the EIS spectra of the CoF₂/MWCNT cells operated in the FEC/EMC and EC/DMC electrolytes after the 1st cycle. The larger semicircle of the CoF₂/MWCNTs cell operated in the FEC/EMC electrolytes indicates a larger resistance of the CoF₂/MWCNT electrode after the 1st cycle. This increase in resistance was primarily due to the in situ formation of CEI layers. Compared with Fig. 5c, the voltage profiles in Fig. S7† show that the irreversible discharge capacities of CoF₂/MWCNTs with the FEC/EMC electrolyte increased significantly during the 1st cycle. In particular, the in situ formation of numerous CEI layers led to an increase in the irreversible capacity and electrode resistance. Even though the operating voltage was higher than 1.0 V (vs. Li/Li⁺), solid-electrolyte interphase layers could be formed because the reduction potentials of FEC-containing electrolytes are higher than those of other electrolytes.^23,40–42 Although this reduced the initial discharge/charge efficiency, the sufficiently formed CEI layers mitigated the dissolution of cobalt ions during cycling.

XPS depth profiling of the electrodes after the 1st cycle provided further information on the CEI layers on the electrodes. When using the EC/DMC electrolyte, Co 2p peaks were observed. The peaks at approximately 781.5 and 797.2 eV correspond to the binding energies of Co²⁺ 2p_3/2 and Co²⁺ 2p_1/2, respectively. Considering this result and the voltage profiles of CoF₂/MWCNTs, Co/LiF formed through the discharging process was re-converted to CoF₂ during the charging process and not to other cobalt compounds. Note that the binding energies of Co 2p in Fig. 7c are slightly lower than those of the pristine CoF₂/MWCNTs (Fig. 4a). This phenomenon is presumed to be related to the conditions under which CoF₂ was formed. Previous studies have reported that the binding energy of Co 2p decreases as CoF₂ is synthesized at low temperatures.^43,44 Therefore, compared with the pristine CoF₂/MWCNTs synthesized at 250 °C, the binding energy of CoF₂ formed through the electrochemical conversion reaction at room temperature was lowered. Peaks corresponding to metallic Co were observed as the etching depth increased. The Co 2p peaks at 778.2 and 793.2 eV correspond to Co⁰ 2p_3/2 and Co⁰ 2p_1/2, respectively. These peaks originate from Co/LiF that did not re-convert to CoF₂ near the core of the particle. By contrast, when the FEC/EMC electrolyte was used, almost no Co 2p peak was observed (Fig. 7c). Even after etching to approximately 10 nm, the intensities of the Co peaks did not increase significantly, indicating the presence of a large number of CEI layers. The appearance of the F 1s peak originated from the components in the CEI layers and not from the Co–F bonding (Fig. 7d). To confirm the stability of the CEI layers, TEM analysis was further conducted using the CoF₂/MWCNT electrodes cycled in the FEC/EMC electrolyte. As shown in Fig. S8,† CEI layers were generated on the surface of CoF₂ particles after the 1st cycle (Fig. S8a†) and stably maintained similar thickness even after the 100th cycle (Fig. S8b–d†).

As mentioned above, such CEI layers improved the cyclability of CoF₂/MWCNTs by mitigating active material dissolution, although it was accompanied by a decrease in the coulombic efficiency and an increase in the interfacial resistance. The post-mortem SEM and EDS analyses of a part of the cycled separators also support this assumption. In the case of the EC/DMC electrolyte cell, severely clogged parts were observed on the surface of the cycled separator (Fig. S9†). The EDS mapping images revealed a high accumulation of cobalt, indicating that the active materials in CoF₂/MWCNTs were significantly dissolved. As the voltage increased during charging, a portion of the metallic cobalt formed during the discharging process could be dissolved as cobalt ions into the electrolyte. The repeated dissolution of cobalt ions during cycling led to the growth of numerous side products and severe degradation of the cells. By sharp contrast, when using the FEC/EMC electrolyte cell, the number of side products on the separator was relatively reduced, and the corresponding EDS mapping showed the partial presence of cobalt (Fig. S10†). Owing to the significant influence of the CEI layers formed by the FEC-containing electrolyte, excessive dissolution of active materials was efficiently mitigated, and the cyclability of CoF₂/MWCNTs could be improved.

Ex situ Al₂O₃ coating by LAD to reduce capacity fading

Ex situ Al₂O₃ layer coating using ALD was further conducted to reduce undesirable reactions at the interface of the electrode and electrolyte. To explore the effect of the coating thickness, the CoF₂/MWCNT electrodes were coated with Al₂O₃ for 20 or 50 ALD cycles. As shown in Fig. S11a,† the high-resolution TEM image suggests the presence of amorphous Al₂O₃ layers on the CoF₂/MWCNT electrode. Fig. S11b† shows the Al 2p XPS spectra of 20A- and 50A-CoF₂/MWCNTs. The binding energy of 74.5 eV in the XPS spectrum is attributed to the Al–O bond,⁴⁵ confirming the existence of Al₂O₃ on the electrode. The peak intensity increased with an increasing number of ALD cycles.

Fig. 8a shows the cycling performance of CoF₂/MWCNTs with different Al₂O₃ coating thicknesses at 0.2C, within a voltage range of 1.0–4.0 V, using the FEC/EMC electrolyte. To closely observe the capacity after the first cycle, the second to 100th cycles were plotted. The discharge capacity of uncoated-CoF₂/MWCNTs continued to decay and exhibited the lowest capacity of 180 mA h g⁻¹ in the 100th cycle. In comparison, the discharge capacities of 20A- and 50A-CoF₂/MWCNTs in the 100th cycle were 11% (199 mA h g⁻¹) and 37% (247 mA h g⁻¹) higher than that of the uncoated CoF₂/MWCNTs, respectively. The better cyclability of 50A-CoF₂/MWCNTs indicates that a sufficient Al₂O₃ coating effectively reduced the active material dissolution and improved the electrochemical performance.¹⁹ Interestingly, even though the inactive and non-electrically conductive Al₂O₃ layers are present on 50A-CoF₂/MWCNTs, the results of EIS analysis after the 100th cycle show that 50A-CoF₂/MWCNTs had a lower charge transfer resistance compared to CoF₂/MWCNTs. Therefore, a sufficient Al₂O₃ coating on CoF₂/MWCNTs through ALD suppressed the formation of undesirable resistive species originating from the dissolution of active materials.⁴⁶ The long-term cycling stabilities of 20A- and 50A-CoF₂/MWCNTs were further examined. Fig. 8c shows the stable cycling performance of both electrodes over 500 cycles at 0.2C. In particular, 50A-CoF₂/MWCNTs retained a higher capacity of 206 mA h g⁻¹ after 500 cycles. The changes in the separator after 500 cycles in the 50A-CoF₂/MWCNT cell were investigated by SEM imaging and EDS mapping (Fig. S12†). Only small amounts of impurities were present on the surface of the separator in the 50A-CoF₂/MWCNT cell with the FEC/EMC electrolyte; a further decrease in side reactions was observed compared with the uncoated-CoF₂/MWCNT cell with the FEC/EMC electrolyte (Fig. S10†). Overall, the results demonstrate that the ex situ coating of Al₂O₃ layers is an effective strategy to reduce active material dissolution and improve the long-term cyclability of CoF₂-based cathodes. One thing to note is that the coulombic efficiencies (CEs) of CoF₂/MWCNTs at the early stage of the cycle tests were lower than those of commercial cathode materials for the following reasons. (i) CEI layers formed by the irreversible reduction reactions of the FEC-containing electrolyte. (ii) Some CoF₂ particles that were not closely complexed with MWCNTs exhibited irreversible conversion reactions. (iii) Even though Al₂O₃ and CEI layers existed on the surface of active materials, small amounts of metallic Co became exposed to the electrolyte during the repeated conversion/re-conversion reaction; they could be dissolved in the high voltage region. However, such phenomena mainly occurred at the early stage of the cycle test. Therefore, as the cycle progressed the CEs of CoF₂/MWCNTs increased and they approached approximately 100% at a later stage (Fig. 8a and c).

Conclusions

CoF₂/MWCNTs were synthesized using a facile method developed in this study via the precipitation of Co precursors onto MWCNTs, followed by an easy fluorination process with NH₄F. Well-dispersed CoF₂ nanoparticles on electrically conductive MWCNTs facilitated the conversion reaction of CoF₂. As a result, CoF₂/MWCNTs exhibited an improved reversible capacity (554 mA g⁻¹ at 0.2C) and excellent rate capability (a capacity retention of 84.7% at 2C versus 0.2C) in comparison with bare-CoF₂. In addition, the electrochemical performance of CoF₂/MWCNTs was further improved by in situ and ex situ surface protection approaches. In situ formation of CEI layers using an FEC-containing electrolyte and ex situ Al₂O₃ layer coating on the electrode by ALD effectively reduced the dissolution of active materials during cycling. By protecting the active materials from undesirable side reactions with the electrolyte, CoF₂/MWCNTs retained a capacity of more than 200 mA h g⁻¹ after 500 cycles. This study proposes a practical material design with advanced synthetic and surface modification methods to improve the electrochemical performance of CoF₂-based cathodes. This study is expected to provide new insights for achieving high-performance LIBs based on new MF_x cathode materials.

Author contributions

H. P. designed and conducted the whole experiments and wrote the manuscript. I.-S. J. and B.-Y. S. carried out material characterization and electrochemical analyses. Y. C. K. supervised this work. S. K. supervised this work and revised the manuscript. J. C. designed this work and contributed to writing the manuscript.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the Korea Institute of Energy Technology Evaluation and Planning (KETEP) grant funded by the Korea government (MOTIE) (20219410100120, commercialization technology development of lithium-ion capacitors with sacrificial cathodes for solar ESS). This research was also supported by a National Research Foundation of Korea (NRF) grant funded by the Korea government (RS-2023-00214588).

Notes and references

1. C. Liu, Z. G. Neale and G. Cao, Mater. Today, 2016, 19, 109–123 CrossRef CAS .
2. W. Li, E. M. Erickson and A. Manthiram, Nat. Energy, 2020, 5, 26–34 CrossRef CAS .
3. C.-H. Jung, H. Shim, D. Eum and S.-H. Hong, J. Korean Ceram. Soc., 2021, 58, 1–27 CrossRef CAS .
4. F. Wu and G. Yushin, Energy Environ. Sci., 2017, 10, 435–459 RSC .
5. K. Turcheniuk, D. Bondarev, V. Singhal and G. Yushin, Nature, 2018, 559, 467–470 CrossRef CAS PubMed .
6. L. Wang, Z. Wu, J. Zou, P. Gao, X. Niu, H. Li and L. Chen, Joule, 2019, 3, 2086–2102 CrossRef CAS .
7. K. Turcheniuk, D. Bondarev, G. G. Amatucci and G. Yushin, Mater. Today, 2021, 42, 57–72 CrossRef CAS .
8. R. Saroha, J.-H. Ahn and J. S. Cho, Korean J. Chem. Eng., 2021, 38, 461–474 CrossRef CAS .
9. Y. Lu, L. Yu and X. W. D. Lou, Chem, 2018, 4, 972–996 CAS .
10. R. Prakash, C. Wall, A. K. Mishra, C. Kübel, M. Ghafari, H. Hahn and M. Fichtner, J. Power Sources, 2011, 196, 5936–5944 CrossRef CAS .
11. F. Wu, J. Maier and Y. Yu, Chem. Soc. Rev., 2020, 49, 1569–1614 RSC .
12. C. P. Guntlin, T. Zünd, K. V. Kravchyk, M. Wörle, M. I. Bodnarchuk and M. V. Kovalenko, J. Mater. Chem. A, 2017, 5, 7383–7393 RSC .
13. G. G. Amatucci and N. Pereira, J. Fluorine Chem., 2007, 128, 243–262 CrossRef CAS .
14. F. Klein, B. Jache, A. Bhide and P. Adelhelm, Phys. Chem. Chem. Phys., 2013, 15, 15876–15887 RSC .
15. C. Li, C. Yin, L. Gu, R. E. Dinnebier, X. Mu, P. A. van Aken and J. Maier, J. Am. Chem. Soc., 2013, 135, 11425–11428 CrossRef CAS PubMed .
16. C. Li, C. Yin, X. Mu and J. Maier, Chem. Mater., 2013, 25, 962–969 CrossRef CAS .
17. Y. Lu, Y. Zhou, Q. Yan and E. Fong, J. Mater. Chem. A, 2016, 4, 2691–2698 RSC .
18. Y. Lu, Z.-y. Wen, J. Jin, X.-w. Wu and K. Rui, Chem. Commun., 2014, 50, 6487–6490 RSC .
19. Y. Wang, M. Zhang, Y. Zhang, Y. Wang, W. Liu, C. Yang, V. Kondratiev and F. Wu, Energy Fuels, 2022, 36, 8439–8448 CrossRef CAS .
20. N. Nitta, F. Wu, J. T. Lee and G. Yushin, Mater. Today, 2015, 18, 252–264 CrossRef CAS .
21. X. Wang, W. Gu, J. T. Lee, N. Nitta, J. Benson, A. Magasinski, M. W. Schauer and G. Yushin, Small, 2015, 11, 5164–5173 CrossRef CAS PubMed .
22. C. Wall, R. Prakash, C. Kübel, H. Hahn and M. Fichtner, J. Alloys Compd., 2012, 530, 121–126 CrossRef CAS .
23. F. Wu, V. Srot, S. Chen, M. Zhang, P. A. van Aken, Y. Wang, J. Maier and Y. Yu, ACS Nano, 2021, 15, 1509–1518 CrossRef CAS PubMed .
24. J. Tan, L. Liu, S. Guo, H. Hu, Z. Yan, Q. Zhou, Z. Huang, H. Shu, X. Yang and X. Wang, Electrochim. Acta, 2015, 168, 225–233 CrossRef CAS .
25. Y. T. Teng, S. S. Pramana, J. Ding, T. Wu and R. Yazami, Electrochim. Acta, 2013, 107, 301–312 CrossRef CAS .
26. M. J. Armstrong, A. Panneerselvam, C. O'Regan, M. A. Morris and J. D. Holmes, J. Mater. Chem. A, 2013, 1, 10667–10676 RSC .
27. Z. Wang, D. Luan, S. Madhavi, Y. Hu and X. W. D. Lou, Energy Environ. Sci., 2012, 5, 5252–5256 RSC .
28. J. Chun, C. Jo, S. Sahgong, M. G. Kim, E. Lim, D. H. Kim, J. Hwang, E. Kang, K. A. Ryu and Y. S. Jung, ACS Appl. Mater. Interfaces, 2016, 8, 35180–35190 CrossRef CAS PubMed .
29. W. Choi and A. Manthiram, J. Electrochem. Soc., 2006, 153, A1760–A1764 CrossRef CAS .
30. D. F. Katowah, S. M. Saleh, S. A. Alqarni, R. Ali, G. I. Mohammed and M. A. Hussein, Sci. Rep., 2021, 11, 5056 CrossRef CAS PubMed .
31. Q. Cheng, Y. Chen, X. Lin, J. Liu, Z. Yuan and Y. Cai, J. Phys. Chem. C, 2020, 124, 8624–8632 CrossRef CAS .
32. J. Rojas, M. Toro-Gonzalez, M. Molina-Higgins and C. Castano, Mater. Sci. Eng., B, 2016, 205, 28–35 CrossRef CAS .
33. R. Sitko, E. Turek, B. Zawisza, E. Malicka, E. Talik, J. Heimann, A. Gagor, B. Feist and R. Wrzalik, Dalton Trans., 2013, 42, 5682–5689 RSC .
34. L. Li, F. Meng and S. Jin, Nano Lett., 2012, 12, 6030–6037 CrossRef CAS PubMed .
35. Q. Zhang, Y. T. Huang, X. Chen, A. Pan, Z. Cai, S. Liu and Y. Zhang, J. Alloys Compd., 2019, 805, 539–544 CrossRef CAS .
36. C. Li, K. Chen, X. Zhou and J. Maier, npj Comput. Mater., 2018, 4, 22 CrossRef .
37. A. Nulu, V. Nulu, J. S. Moon and K. Y. Sohn, Korean J. Chem. Eng., 2021, 38, 1923–1933 CrossRef CAS .
38. J. Oh, E. Lim, J. Chun and C. Jo, J. Power Sources, 2022, 521, 230935 CrossRef CAS .
39. X. Hua, R. Robert, L.-S. Du, K. M. Wiaderek, M. Leskes, K. W. Chapman, P. J. Chupas and C. P. Grey, J. Phys. Chem. C, 2014, 118, 15169–15184 CrossRef CAS .
40. T. Hou, G. Yang, N. N. Rajput, J. Self, S.-W. Park, J. Nanda and K. A. Persson, Nano Energy, 2019, 64, 103881 CrossRef CAS .
41. Q. Huang, T. P. Pollard, X. Ren, D. Kim, A. Magasinski, O. Borodin and G. Yushin, Small, 2019, 15, 1804670 CrossRef PubMed .
42. E. Markevich, G. Salitra and D. Aurbach, ACS Energy Lett., 2017, 2, 1337–1345 CrossRef CAS .
43. Z.-W. Fu, C.-L. Li, W.-Y. Liu, J. Ma, Y. Wang and Q.-Z. Qin, J. Electrochem. Soc., 2004, 152, E50 CrossRef .
44. M. Li, Y. Gu, Y. Chang, X. Gu, J. Tian, X. Wu and L. Feng, Chem. Eng. J., 2021, 425, 130686 CrossRef CAS .
45. E. Zhao, O. Borodin, X. Gao, D. Lei, Y. Xiao, X. Ren, W. Fu, A. Magasinski, K. Turcheniuk and G. Yushin, Adv. Energy Mater., 2018, 8, 1800721 CrossRef .
46. G. Xu, C. Pang, B. Chen, J. Ma, X. Wang, J. Chai, Q. Wang, W. An, X. Zhou and G. Cui, Adv. Energy Mater., 2018, 8, 1701398 CrossRef .

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d3ta01918g

This journal is © The Royal Society of Chemistry 2023