Unveiling the charge storage mechanisms of Co-based perovskite fluoride in a mild aqueous electrolyte†

Yuzhen Zhang‡ , Miao Liu‡ , Rui Ding *, Yi Li , Jian Guo , Qi Fang , Miao Yan and Jinmei Xie
Key Laboratory of Environmentally Friendly Chemistry and Applications of Ministry of Education, College of Chemistry, Xiangtan University, Xiangtan, Hunan 411105, P.R. China. E-mail: drm8122@xtu.edu.cn; drm8122@163.com

First published on 30th August 2024

---

---

Introduction

Given the looming threat of resource shortages and environmental pollution, there is a critical need to explore reliable technologies for storing green energy.^1–3 Electrochemical energy storage systems (EESs) are receiving global attention as an intermediary platform for the versatile and efficient deployment of renewable energy sources. The supercapacitor has attracted increasing attention as an important energy storage device known for its high specific power and excellent cycling stability.^4–11 The move towards using water-based electrolytes instead of flammable and expensive organic electrolytes in aqueous supercapacitors has drawn the attention of many researchers, which is attributed to its low pollution, low cost, high safety, high ionic conductivity and other advantages.^12–15 Based on the energy storage mechanisms, supercapacitors are usually classified into two categories: electrical double-layer capacitors (EDLCs) and pseudo-capacitors. EDLCs perform charge storage by forming an interfacial double layer between the electrodes and the electrolyte, which is a non-faradaic process. Carbon-based materials with extensive surface areas are mainly utilized as active electrodes.^16,17 Conversely, the energy storage mechanism of the pseudo-capacitor is the transfer of the electronic charge between the electrodes and the electrolyte, and their electrode materials are mainly transition metal oxides and conductive redox polymers.^18–20

Specific energy (E = 1/2 CV²) and specific power (P = 1/2 VI) stand as essential benchmarks for evaluating the capacitive capabilities of supercapacitors, guided by the respective formulas that define them. For a supercapacitor to be deemed exceptional, it needs to maintain a high cell voltage. One of the major limitations to the widespread use of aqueous supercapacitors is the relatively low cell voltage, which is due to the decomposition voltage of 1.23 V for water. The overpotentials for hydrogen-extraction reactions and oxygen-extraction reactions are usually higher in mild neutral electrolytes than in rigorous acidic and basic electrolytes.^14,21 This means that, under neutral conditions, water splitting (generating hydrogen and oxygen) requires a higher voltage, thereby expanding the operational voltage window of the electrolyte. Additionally, acidic or alkaline environments may accelerate the corrosion or degradation of electrode materials; hence, supercapacitors utilized in mild neutral electrolytes could have a longer lifespan and better long-term stability. In a study conducted by Fedorov et al., the electrochemical properties of electrodeposited cobalt hydroxide and cobalt oxide in 1 M Na₂SO₄ solution were investigated with the anodic potential in the range up to 0.8–1.3 V vs. Ag/AgCl.²² Zhao et al. studied the electrochemical performance of NiCo₂O₄ nanosheets and showed that the potential window is 1.2 V in the Na₂SO₄ neutral electrolyte, which is almost twice as much as that of the alkaline electrolyte (0–0.6 V).²³ Overall, the development of EESs based on neutral electrolytes is of increasing interest, and it is therefore crucial to elucidate the charge storage mechanism of electrode materials in neutral media.

In recent years, transition metal perovskite fluoride (ABF₃) electrode materials with structures similar to ABO₃ perovskites have shown good application potential due to their strong electronegativity of F that can lower the Fermi energy level and increase the operating voltage window. ABF₃ features a robust open frame, intersecting tetragonal lumen chains, and three-dimensional diffusion channels that facilitate ion migration for excellent rate performance. Perovskite fluoride is a hot research material in the field of energy storage due to its superior structural properties.^24–26 Recently, the results of research on the use of different metal elements to replace perovskite-type fluoride positive materials with A or B sites in different EESs have been reported by Ding's research group; these energy storage devices exhibit typical bulk conversion mechanisms in alkaline media.^27–31 However, studies on the mechanism of ABF₃ materials in mild aqueous electrolytes are still limited. Our group has recently reported the charge storage mechanism of KMnF₃ in a neutral electrolyte,³² mainly the transformation reaction between the Mn (+2/+3/+4) redox species, the adsorption of H₂O molecules, and the intercalation/extraction of Na⁺ and H⁺ in the electrolyte. This provides us insights for deeply understanding the energy storage mechanisms of ABF₃ electrode materials in mild aqueous electrolytes, laying a solid foundation.

Over the years, cobalt (Co) has been extensively used as a key electrode material component in various aqueous and non-aqueous EESs. In this paper, the charge storage mechanism of KCoF₃ in neutral aqueous electrolytes is elucidated by using electrochemical quartz crystal microbalance (EQCM) combined with a variety of ex situ characterization techniques and electrochemical testing methods. Specifically, KCoF₃ nanocrystals are converted into amorphous Co(OH)₂, Co₃O₄, CoOOH, and (hydrated) CoO₂ using adsorption, conversion and insertion/extraction mechanisms. During the first charging stage, KCoF₃ generates Co(OH)₂ in the presence of water, and subsequently Co(OH)₂ is converted into Co₃O₄ and CoOOH. CoOOH is partially extracted by H⁺ to form CoO₂, while Co₃O₄ is further converted into CoOOH and (hydrated) CoO₂. During the first discharging period, part of the H⁺ is inserted into (hydrated) CoO₂ to form CoOOH and Co₃O₄, and CoOOH is converted into Co₃O₄, and both CoOOH and Co₃O₄ are further converted into Co(OH)₂. Furthermore, we also added the KCoF₃ material into aqueous SCs, leading to the attainment of an extensive voltage range from 0 to 1.9 V, coupled with exceptional stability. This study notably advances our understanding of the charge storage mechanism of ABF₃ materials in mild aqueous electrolytes and is important for the development of advanced EESs.

Results and discussion

XRD was used to confirm the phase compositions and crystallinity of the synthesized KCoF₃ material. Fig. 1a and Fig. S1† show the XRD results of KCoF₃. The sharp diffraction peaks indicate that the material has good crystallization, and the characteristic peaks located at 31.06°, 44.50°, 55.26° and 64.75° correspond to the crystal facets of KCoF₃ (PDF# 18-1006) (110), (200), (211) and (220), respectively. Furthermore, it is noteworthy that the XRD pattern for the KCoF₃ sample shows a trace of impurity spots at around 2θ = 27°. This is attributed to the comparatively weak metal oxyhydroxide (CoOOH) phase, which results from the water (H₂O) utilized as a co-solvent in trace amounts throughout the synthesis process. The elemental content of the KCoF₃ material was ascertained using ICP-OES, and the outcomes are displayed in Fig. 1b, with the chemical formula of K_0.92CoF_3.02, close to the theoretical composition of ABF₃ (Fig. S2†). Additionally, using Rietveld refinement of the XRD data examined at a scan rate of 2° min⁻¹, the exact lattice parameters of the KCoF₃ sample were determined (Fig. 1c), which slightly deviated from the standard value for KCoF₃ (4.0708 Å).

In KCoF₃, Co is primarily bivalent with a trace quantity of trivalent, in accordance with the principle of electrical neutrality. The surface morphology, particle size and distribution, fine crystalline structure, and element distribution of the KCoF₃ material were further investigated in detail using SEM and TEM (Fig. 1d and e; Fig. S3a–d†). The KCoF₃ sample is represented by rectangular nanocrystals in the size range of 75–95 nm, with 85 nm being the predominant size, as seen in Fig. 1d and e; however, bigger size particles (micro-level) can also be detected. Fig. 1f shows the magnified region of the (110) crystal plane of the KCoF₃ material in a high-resolution TEM (HRTEM) image with a d-spacing value of 2.951 Å, and SAED patterns in Fig. 1g, which demonstrate strong crystallinity based on regular lattice fringes and clear diffraction spots and are well matched with the XRD results, which further validate the typical crystalline planes of (100), (110), (111), (200), (210), and (211) identified in the XRD patterns formerly. EDS can be used to verify elements K, Co, and F in the KCoF₃ sample. The existence and even distribution of K, Co, and F elements are shown in Fig. 1h; Fig. S4,† further demonstrating the material's effective synthesis. The specific surface and pore structure of the KCoF₃ material have been further examined using nitrogen sorption isotherms, pore volume and pore size distribution tests (Fig. 1i and j), showing that the material has a specific surface area, cumulative pore volume, and average pore diameter of 5.99 m² g, 0.044 cm³ g⁻¹ and 24.90 nm, respectively, with the majority of the pore size located at 2.47 nm and 49.19 nm.

The electrochemical properties of as-prepared KCoF₃ electrodes were evaluated by using a three-electrode system in 1 mol L⁻¹ Na₂SO₄ electrolyte, which was also compared with the nickel foam (NF) current collector, the kinetic behaviour of the KCoF₃ electrode and the change in the internal resistance to charge transfer R_ct before and after electrode cycling were analyzed, and the findings are shown in Fig. 2. The first-loop CV map has been the focus of charge storage mechanism studies. Fig. 2a displays the initial CV cycle of the KCoF₃ electrode at a scan rate of 10 mV s⁻¹ with a voltage window of −1.1 to 1.25 V. It can be seen that the CV cycle has multiple pairs of oxidation peaks during the charge process, and its appearance may be attributed to the conversion of the B-site Co element into a higher valency phase. A₁ may be attributed to pristine KCoF₃ adsorbing water to form an electric double layer and the conversion of the pristine KCoF₃ material into Co²⁺ (Co(OH)₂) in the presence of water, and the subsequent oxidation of Co²⁺ (Co(OH)₂) to form Co^2+/3+ (Co₃O₄) and Co³⁺ (CoOOH), In A₂, the continuous reduction of Co²⁺ (Co(OH)₂) to Co^2+/3+ (Co₃O₄) and Co³⁺ (CoOOH) and the partial extraction of H⁺ from Co³⁺ (CoOOH) might lead to the formation of Co⁴⁺ (CoO₂) at the same time. More prominently, an oxidation peak appears in A₃ and A₄ possibly due to the oxidation of Co²⁺ (Co(OH)₂), Co³⁺ (CoOOH) to Co^2+/3+ (Co₃O₄), Co³⁺ (CoOOH), and Co⁴⁺ (CoO₂). Additionally, this might involve the oxidation of Co^2+/3+ (Co₃O₄) to Co³⁺ (CoOOH) and Co⁴⁺ (CoO₂ or hydrated CoO₂). During the discharge process, high-valence Co compounds are reduced, resulting in reduction peaks. A distinct reduction peak in the C₁ area could be attributed to the embedding of H⁺ into Co⁴⁺ (CoO₂ or hydrated CoO₂) to form Co³⁺ (CoOOH) and Co^2+/3+ (Co₃O₄), which is possibly accompanied by the reduction of Co³⁺ (CoOOH) to Co^2+/3+ (Co₃O₄) and both Co³⁺ (CoOOH) and Co^2+/3+ (Co₃O₄) to Co²⁺ (Co(OH)₂). A weak reduction peak in C₂ and C₃ area might be the continuous reduction of Co⁴⁺ (CoO₂) to Co^2+/3+ (Co₃O₄) and the reduction of Co³⁺ (CoOOH) and Co^2+/3+ (Co₃O₄) to Co²⁺ (Co(OH)₂). Additionally, a more distinct reduction peak in the C₄ area could result from the reduction of Co³⁺ (CoOOH) and Co^2+/3+ (Co₃O₄) back to Co²⁺ (Co(OH)₂).

The CV curve for the third cycle of conventional electrochemical testing is shown in the inset of Fig. 2a, where the intensity of the oxidation peaks is significantly higher in the third lap compared to the first cycle, but the peak positions are almost the same, probably due to further activation of the reactants. However, in the A₅ region, a small and less pronounced oxidation peak can be seen, which could be attributed to the further oxidation of the substances generated by the previous reaction. Fig. 2b shows the first cycle CV curve obtained from the EQCM device test, and the inset shows the third cycle CV curve. In Fig. 2a and b, it is clear that the CV curves obtained from the conventional device test with the potential positions of the redox peaks are similar to the curves obtained from the EQCM equipment test. The primary distinction lies in the fact that the redox peak intensities in the initial three turns of the CV curve obtained from EQCM are nearly identical. Additionally, the A₃ region during the charging process encompasses the A₄ region generated by the conventional electrochemical test. This is because the active material on the quartz wafer surface reaches the microgram level, resulting in a more comprehensive reaction of the active material.

We tested the CV curves at different scanning speeds from 2 to 160 mV s⁻¹ to evaluate the kinetic behaviour of the KCoF₃ electrode (Fig. 2d), from which it can be seen that there are four pairs of redox peaks (A₁/C₁, A₂/C₂, A₃/C₃, and A₄/C₄), and the peak current and the area of the CV curve are increasing with the increase of the sweep rate, implying that the electrode has superior rate capability. It is worth noting that these results were obtained after two turns of CV at 10 mV s⁻¹, in which the CV curve at 10 mV s⁻¹ is consistent with the results of the first four turns (Fig. S5†). In addition, a very small fraction of the current to the KCoF₃ electrode in the neutral electrolyte can be supplied by the nickel foam current collector (Fig. S6†). Fig. 2e shows the GCD curves (the 2nd cycle) of the KCoF₃ electrode at different current densities from 1 to 16 A g⁻¹ obtained after achieving the CV of Fig. 2d, with a voltage window of −0.9 to 1.2 V, and the variation curves of the specific capacity of the electrode with current density are shown in Fig. 2c, which shows that the electrode specific capacity reaches to 40 C g⁻¹ (1 A g⁻¹). The corresponding coulombic efficiency increases gradually with the increase of the current density and finally approaches 100%. Fig. 2f shows the cycling performance of the KCoF₃ electrode for 10000 cycles at 2 A g⁻¹, and the GCD curves at different cycling turns are shown in the inset. From the figure we can see that the capacity undergoes a slight decay in the first 529 cycles, because diffusion governs the reaction (as shown later in the kinetics results), and the rate is faster at the beginning. As more turns are added, the electrolyte gets deeper into the electrode and the rate slows. However, after 529 cycles, the internal channels of the electrode material gradually open and the rate rises, at which point the specific capacity shows an increasing trend and exhibits exceptional cycling stability. To understand the kinetic behaviour of the KCoF₃ electrode, a series of CV curves ranging from −1.1 to 1.25 V were collected over a scan rate range of 2–10 mV s⁻¹. According to i = k₁v + k₂v^1/2, the percentage of capacitance and diffusion contributions to the total capacity can be quantified by dividing the response current at the same potential into surface-dominated processes (k₁v) and diffusion-dependent processes (k₂v^1/2). As demonstrated in Fig. 2g, the slope of the straight line (b) can be obtained by fitting the peak intensity values of each current peak at different scanning rates. It is evident that some of the b values are close to 0.5, and some of the b values far more than 0.8 for both cathodic and anodic reactions, indicating that the KCoF₃ electrode exhibits a kinetic behaviour controlled by a mixture of pseudocapacitance and diffusion in the Na₂SO₄ neutral electrolyte. The shaded area of the curves in Fig. 2h represents the capacitance contributed by diffusion control, and its corresponding percentage contribution is shown in Fig. 2i. It can be seen that the contribution of the surface capacitance to the total charge storage increases from about 3.31 to 7.10% at 2–10 mV s⁻¹, and it is clear that the diffusion-controlled contribution is much larger than that of the surface capacitance, which further indicates diffusion-dominated kinetics. Fig. 2j and k provides a clear illustration of the GITT curves for the KCoF₃ electrode as well as a zoomed-in view of the selected area. It is evident from the figure that the diffusion coefficients of the KCoF₃ electrode are 1.4 × 10⁻⁹–9.6 × 10⁻⁸ and 4.28 × 10⁻¹¹–2.6 × 10⁻⁷ cm² s⁻¹ during the charging and discharging processes, respectively. The Nyquist plots before and after the KCoF₃ cycle are shown in Fig. 2i, and the equivalent circuit models are LR(Q(RW))(QR) and LR(QR)(QR)Q, respectively. The charge transfer resistance (R_ct) of the redox species during the kinetics is represented by the second R in the model. It is worth noting that the R_ct value of the KCoF₃ electrode is lower after the cycling than it was before the cycling (Table S2†). It shows that the kinetics are faster after cycling and hence there is an increase in the electrode specific capacity with the increase in the number of cycles.

The surface phases of the KCoF₃ powder/electrode in pristine, charged and discharged states after the first CV cycle at a sweep speed of 10 mV s⁻¹ were analysed using ex situ XPS to elucidate the charge storage mechanism of the KCoF₃ electrode. Fig. 3a shows the characteristic peaks of Co 2p_3/2 and their corresponding satellite peaks (Sat.) in these three states. In the initial state, the presence of Co²⁺ (KCoF₃) is mainly observed, but a small amount of CoOOH is present, which is consistent with the XRD results. In contrast, the peak binding energy of Co 2p_3/2 after charging and discharging has a significant shift, which reveals that the first charge and discharge induce an irreversible electrochemical phase transition process, and the original state of Co²⁺ (KCoF₃) is converted into Co²⁺ (Co(OH)₂), Co³⁺ (CoOOH), Co⁴⁺ (CoO₂), and Co⁴⁺ (CoO₂)/Co³⁺ (CoOOH), Co^2+/3+ (Co₃O₄) and Co²⁺ (Co(OH)₂). It is worth noting that the K 2p and F 1s spectra in the charged/discharged state, as shown in Fig. 3c and d, differ significantly from the initial state and the typical peaks present in the original state are almost all lost in the charged/discharged state (the C–F bond of F 1s at around 688 eV is derived from polyvinylidene fluoride (PVDF), and the CoF_x component of F 1s in its original state is formed when materials are synthesized), indicating that the phase change has occurred in the original material. To further verify the species produced by the transformation, in Fig. 3e, the O 1s peaks in the charged and discharged states can be broken down into four peaks, corresponding to Co–O, Co–OOH, Co–OH and H₂O, respectively, while the O 1s peaks in the initial state are mainly composed of Co–OOH and adsorbed water (H₂O), where the peaks of Co–OOH are caused by CoOOH impurities formed during the synthesis of materials. These changes in the K 2p, F 1s and O 1s orbital peaks further confirm that KCoF₃ undergoes an irreversible electrochemical phase transition during the first charge–discharge period.

We further used ex situ XRD, Raman, and FTIR testing to confirm the phases generated before and after the corresponding redox peaks at different potentials to better understand the phase transition process. In the ex situ XRD patterns of the KCoF₃ electrodes in the charged/discharged states (Fig. 4a), as can be seen, the near disappearance of the characteristic peaks of the KCoF₃ perovskite phase indicates the irreversible bulk phase conversion of the crystalline KCoF₃ perovskite structure into amorphous cobalt oxide, cobalt hydroxyl oxide and cobalt hydroxide, while these products supported on the nickel foam current collector are difficult to accurately detect using the conventional XRD. Therefore, we used ex situ Raman and FT-IR to better characterize the specific phases generated. Fig. 4b showcases the FT-IR spectrum of the KCoF₃ electrode powder. A weak Co–F stretching in KCoF₃ was observed at 417 cm⁻¹ in the pristine state.³³ Notably, this signal was also observed in both the charged and discharged states, which may be due to incomplete conversions of KCoF₃. Furthermore, an absorption band at 556 cm⁻¹ was detected in the charged and discharged states, which is attributed to the Co–OH bending modes of CoOOH.³⁴ In addition, stretching modes of Co–O weak broad absorption bands at 572 and 672 cm⁻¹, which confirm the presence of the CoO₂ structure,³⁵ and absorption bands were observed at 542 and 655 cm⁻¹, which may be related to the coupling mode between the Co–O stretching modes of Co₃O₄.³⁶ In addition, two weak absorption bands are noted at 515 and 620 cm⁻¹, which may be related to the Co–O and Co–OH stretching of Co(OH)₂,³⁷ respectively. In further studies, ex situ RS at different voltage states (charged to −0.8, −0.5, 0.15, 0.45, 0.7, 1.0, and 1.25 V; discharged to 0.5, 0.1, −0.35, −0.75, −0.88, −0.97, and −1.1 V) is shown in Fig. 4c and d. In the pristine state, the vibrational peaks at 197 and 267 cm⁻¹ can be attributed to KCoF₃,³⁸ and due to incomplete conversions, the signal can be detected in other states as well. In the presence of potentials, we assume that the reaction starts with the binding of H₂O to the Co-active centre to generate Co–OH₂, along with the successive shedding of H⁺ to form the intermediate (CoO), and H₂O takes advantage of the nucleophilic attack on CoO to produce Co(OH)₂, which can be accounted for by the Raman signal observed near 572 cm⁻¹ (the Co–OH stretching of Co(OH)₂).^34,39 When charged to −0.5 V, the Co–O characteristic peak of Co₃O₄ can be identified around 525 and 694 cm⁻¹,^40,41 while the Co–OH stretching signals of CoOOH were also observed at about 499 and 612 cm⁻¹,³⁴ the characteristic peak is detected at the subsequent voltage points, which may be obtained by Co(OH)₂ oxidation. In addition, the Raman signal of the characteristic peak of CoO₂ can be observed at about 417 and 664 cm⁻¹,⁴² where CoO₂ may be oxidized by CoOOH and Co₃O₄, and the characteristic peak can also be detected in Raman measurements from 0.45 charged to 1.25 V and the fully discharged. During the discharge process, the embedding of H⁺ may cause the reduction of CoO₂ to CoOOH and a few Co₃O₄. At the same time, it may be accompanied by the partial reduction of CoOOH to Co₃O₄ and both CoOOH and Co₃O₄ to form Co(OH)₂. The relevant characteristic peaks and displacement data are listed in Fig. 4e. Combined with the above ex situ Raman and FTIR test results, it was further confirmed that cobalt oxide and cobalt hydroxyl oxide were formed during the first CV cycle, which was consistent with the XPS results.

To further confirm the generated phases, TEM/SAED/HRTEM tests were carried out on the fully charged/discharged KCoF₃ electrodes. As shown in Fig. 5, the TEM images illustrate that KCoF₃ transformed rectangular nanocrystals to ultrathin nanosheets (Fig. S9†). Some typical crystal planes of KCoF₃, CoOOH and CoO₂, Co₃O₄ and Co(OH)₂ are detected in both the charged-discharged state in Fig. 5a and b. In addition, faint diffraction spots and faint lattice streaks show that these cobalt hydroxides and oxides are amorphous structures (the crystalline information is listed in Table S3†). This is consistent with the results of various previous characterisation methods and was confirmed later by EQCM tests.

The EQCM can record any small mass changes in nanoscale electrodes in real time and is becoming a powerful tool for electrochemical research. In order to gain a more comprehensive and systematic understanding of the charge storage mechanism of the KCoF₃ electrode, an in situ EQCM test was conducted using a voltage range of −1.1–1.25 V vs. Ag/AgCl and a sweep speed of 10 mV s⁻¹. The first five CV cycles of the KCoF₃ material tested using the EQCM are shown in Fig. S10a,† and the CV curves are generally consistent. In addition, as can be seen in Fig. S10b,† there is a very small change in resonant resistance (ΔR) compared to the resonant frequency change (ΔF) during the electrochemical measurement. Therefore, we can use the Sauerbrey equation to work with the EQCM data. The EQCM results show that the KCoF₃ electrode undergoes a complex process involving changes in both electrode mass and charge (Fig. 6a–d). Specifically, Fig. 6a–d show the CV curves, the current versus time curves, the cumulative mass (Δm) versus potential curves and the accumulated charge (ΔQ) versus potential curves, respectively, for the EQCM with the frequency monitoring function. During the anodic scanning, the first stage (from −1.1 V to −0.8 V) is shown in Fig. 6f, where there is a significant increase in the total mass-to-charge ratio (MCR) of 13.8 and 2.8 g mol⁻¹ e⁻¹, which is a direct result of the adsorption of H₂O molecules leading to the formation of electric double-layer (EDL) effect. The second stage (from −0.8 V to −0.5 V) increases first with an MCR of 12.2 g mol⁻¹ e⁻¹ and then loses weight with an MCR of 15.3 g mol⁻¹ e⁻¹ (Fig. 6g), the process of mass increase is the same as in the first stage, with adsorbed water forming a double electric layer, and it is speculated that Co(OH)₂ might have been generated from the KCoF₃ (the process does not involve electron transfer) in conjunction with the Raman results. Subsequently, Co(OH)₂ oxidises under the action of H₂O to form Co₃O₄ and CoOOH. This reduces the mass of the process. During the anodic scanning from −0.5–0.45 V, Fig. 6h shows a decrease in mass of 14.1 and 4.7 g mol⁻¹ e⁻¹, which may be attributed to the continuous formation of Co₃O₄ and CoOOH by Co(OH)₂, the detachment of H⁺ from CoOOH to form CoO₂ with higher chemical valence. Fig. 6i shows that during the forward scanning from 0.45 to 1.25 V, the mass on the electrode decreases first and then increases at a rate of 19.5 and 11.4 g mol⁻¹ e⁻¹, respectively, which may be due to the continuous oxidation of Co(OH)₂ to form Co₃O₄ and CoOOH, the continuous oxidation of CoOOH to CoO₂ at the same time, and the partial oxidation of Co₃O₄ to the higher value compound CoOOH and (hydrated) CoO₂ at higher potentials. In the process of reverse cathodic scanning, as shown in Fig. 6j, from the voltage range of 1.25–0.1 V, the mass on the electrode first increased and then decreased, and the rates were 10.8 and 17.2 g mol⁻¹ e⁻¹, respectively, which may be related to the formation of CoOOH by H⁺ intercalation of CoO₂, and the continuous reduction of (hydrated) CoO₂ to form Co₃O₄, accompanied by the reduction of CoOOH and Co₃O₄ to form Co(OH)₂. The reason for the decrease may be the continuous reduction of CoOOH and (hydrated) CoO₂ to produce Co₃O₄. The electrode mass increased by MCR of 7.8 g mol⁻¹ e⁻¹ during the discharging within the voltage range of 0.1 V to −0.75 V (Fig. 6k), which is likely attributed to the intercalation of H⁺ into CoO₂ to produce CoOOH and the continuous reduction of CoOOH to Co(OH)₂ at the same time accompanied by the reduction of Co₃O₄ to form Co(OH)₂. A 26.9 g mol⁻¹ e⁻¹ increase in MCR from −0.75 V to −1.1 V (Fig. 6l) could be due to the reduction of CoOOH and Co₃O₄ to Co(OH)₂ at low potential.

Based on the above results of our tests and experiments, the electrochemical reaction process that may occur in the first lap CV test can be described as follows:

The forward anodic scan:I: from −1.1 V to −0.8 V

KCoF3 + H2O → KCoF3·(OH2)x (EDL effect)

II: from −0.8 V to −0.5 V

KCoF3 + H2O → KCoF3·(OH2)x (EDL effect)

KCoF3 + 2H2O → Co(OH)2 + K+ + 3F− + 2H+

3Co(OH)2 → Co3O4 + 2H2O + 2H+ + 2e−

Co(OH)2 → CoOOH + H+ + e−

III: from −0.5 V to 0.45 V

3Co(OH)2 → Co3O4 + 2H2O + 2H+ + 2e−

Co(OH)2 → CoOOH + H+ + e−

CoOOH → CoO2 + H+ + e−

IV: from 0.45 V to 1.25 V

3Co(OH)2 → Co3O4 + 2H2O + 2H+ + 2e−

Co(OH)2 → CoOOH + H+ + e−

CoOOH → CoO2 + H+ + e−

Co3O4 + 2H2O → 3CoOOH + H+ + e−

Co3O4 + 2H2O → 3CoO2 + 4H+ + 4e−

Co3O4 + (2 + 3x)H2O → 3CoO2·(H2O)x + 4H+ + 4e−

The reverse cathodic scan:V: from 1.25 V to 0.1 V

CoO2 + H+ + e− → CoOOH

3CoO2 + 4H+ + 4e− → Co3O4 + 2H2O

CoOOH + H+ + e− → Co(OH)2

3CoOOH + H+ + e− → Co3O4 + 2H2O

Co3O4 + 2H2O + 2H+ + 2e− → 3Co(OH)2

3CoO2·(H2O)x + 4H+ + 4e− → Co3O4 + (2 + 3x)H2O

VI: from 0.1 V to −0.75 V

3CoO2 + 4H+ + 4e− → Co3O4 + 2H2O

CoOOH + H+ + e− → Co(OH)2

Co3O4 + 2H+ + 2H2O + 2e− → 3Co(OH)2

VII: from −0.75 V to −1.1 V

CoOOH + H+ + e− → Co(OH)2

Co3O4 + 2H+ + 2H2O + 2e− → 3Co(OH)2

Aqueous asymmetric electrochemical capacitors consisting of KCoF₃ as the positive electrode and activated carbon (AC) as the negative electrode were further investigated to evaluate the practical application of the KCoF₃ material in electrochemical energy storage (Fig. S11–S13†). According to the charge balance equation, the mass ratio of KCoF₃ and AC electrodes was designed as 4:1 based on the specific capacity at 1 A g⁻¹. Also, it tested the performance of electrochemical capacitors with positive-to-negative mass ratios of 1:1, 1:2, 2:1, 1:3 and 3:1. We chose an operating voltage range of 0–2.0/2.1 V for the full cell, which is higher than that of the perovskite fluoride-based alkaline aqueous SCs. The large triangle and box shapes for GCD and CV plots suggest a pseudocapacitive response. Fig. S13a† shows the energy/power densities of different ratios of AC//KCoF₃ aqueous SCs in 1 M Na₂SO₄ electrolytes, in which the positive-to-negative mass ratios of 1:2 have better overall performance, and the max value can reach over 6.5 W h kg⁻¹ and 2.5 kW kg⁻¹, respectively. In addition, the change of specific capacity in the cycling period for AC//KCoF₃ aqueous SCs is similar to that in the single electrode cycling, after the early activation cycle, the specific capacity shows a large increase (Fig. S13b†), and its coulombic efficiency is stable after thousands of cycles. In all, perovskite fluoride KCoF₃ has great potential for energy storage in the mild aqueous media, as suggested by the aforementioned findings.

Conclusions

In summary, we have demonstrated the charge storage mechanism of the KCoF₃ electrode material in 1 M Na₂SO₄ by using in situ EQCM, integrating a variety of ex situ physicochemical characterization methods and electrochemical tests, showing a bulk phase conversion reaction involving multiple electron co-transfer during the charge–discharge process, as well as adsorption of H₂O molecules reactions and H⁺ intercalation/extraction reactions. This achievement holds substantial significance for advancing the understanding of charge storage processes in ABF₃ electrode materials within mild aqueous environments. It serves as a critical reference point for the creation of innovative aqueous EESs and enhances comprehension of their charge storage mechanisms.

Data availability

The data supporting this article have been included as part of the ESI.†

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The authors are thankful for the financial support from the National Natural Science Foundation of China (22278346 and 22078279) and the Distinguished Young Scholar Fund Project of Hunan Province Natural Science Foundation (2022JJ10044).

References

1. L. Yu and G. Z. Chen, Electrochem. Energy Rev., 2020, 3, 271–285 CrossRef .
2. D. Larcher and J. M. Tarascon, Nat. Chem., 2014, 7, 19–29 CrossRef PubMed .
3. R. Iranpour, M. Stenstrom, G. Tchobanoglous, D. Miller, J. Wright and M. Vossoughi, Science, 1999, 285, 706–711 CrossRef CAS PubMed .
4. Z. Jia, W. Shi, R. Ding, W. Yu, Y. Li, C. Tan, X. Sun and E. Liu, Chem. Commun., 2021, 57, 7962–7965 RSC .
5. Y. Li, R. Ding, Z. Jia, W. Yu, A. Wang, M. Liu, F. Yang, Y. Zhang, Q. Fang, M. Yan, J. Xie, X. Sun and E. Liu, Energy Storage Mater., 2023, 57, 334–345 CrossRef .
6. W. Shi, W. Yu, R. Ding, Z. Jia, Y. Li, Y. Huang, C. Tan, X. Sun and E. Liu, J. Mater. Chem. A, 2021, 9, 9624–9633 RSC .
7. B. Dunn, H. Kamath and J.-M. Tarascon, Science, 2011, 334, 928–935 CrossRef CAS PubMed .
8. Z. Yang, J. Zhang, M. C. W. Kintner-Meyer, X. Lu, D. Choi, J. P. Lemmon and J. Liu, Chem. Rev., 2011, 111, 3577–3613 CrossRef CAS PubMed .
9. D. Chao, W. Zhou, F. Xie and C. Ye, Sci. Adv., 2020, 6, eaba4098 CrossRef CAS PubMed .
10. D. Tie, S. Huang, J. Wang, J. Ma, J. Zhang and Y. Zhao, Energy Storage Mater., 2019, 21, 22–40 CrossRef .
11. A. Muzaffar, M. B. Ahamed and C. M. Hussain, Renewable Sustainable Energy Rev., 2024, 195, 114324 CrossRef CAS .
12. J. Huang, Z. Guo, Y. Ma, D. Bin, Y. Wang and Y. Xia, Small Methods, 2019, 3, 1800272 CrossRef .
13. A. Noori, M. F. El-Kady, M. S. Rahmanifar, R. B. Kaner and M. F. Mousavi, Chem. Soc. Rev., 2019, 48, 1272–1341 RSC .
14. N. Choudhary, C. Li, J. Moore, N. Nagaiah, L. Zhai, Y. Jung and J. Thomas, Adv. Mater., 2017, 29, 1605336 CrossRef PubMed .
15. M. Winter and R. J. Brodd, Chem. Rev., 2004, 104, 4245–4270 CrossRef CAS PubMed .
16. P. Xie, W. Yuan, X. Liu, Y. Peng, Y. Yin, Y. Li and Z. Wu, Energy Storage Mater., 2021, 36, 56–76 CrossRef .
17. L. L. Zhang and X. S. Zhao, Chem. Soc. Rev., 2009, 38, 2520–2531 RSC .
18. Y. Liang, Y. Li, H. Wang and H. Dai, J. Am. Chem. Soc., 2013, 135, 2013–2036 CrossRef CAS PubMed .
19. K. Krishnamoorthy, P. Pazhamalai and S. J. Kim, Electrochim. Acta, 2017, 227, 85–94 CrossRef CAS .
20. Q. Meng, K. Cai, Y. Chen and L. Chen, Nano Energy, 2017, 36, 268–285 CrossRef CAS .
21. T. Guo, D. Zhou, L. Pang, S. Sun, T. Zhou and J. Su, Small, 2022, 18, 2106360 CrossRef CAS PubMed .
22. F. S. Fedorov, J. Linnemann, K. Tschulik, L. Giebeler, M. Uhlemann and A. Gebert, Electrochim. Acta, 2013, 90, 166–170 CrossRef CAS .
23. N. Zhao, H. Fan, M. Zhang, J. Ma, W. Zhang, C. Wang, H. Li, X. Jiang and X. Cao, Electrochim. Acta, 2019, 321, 134681 CrossRef CAS .
24. T. Yan, Y. Huang, R. Ding, W. Shi, D. Ying, Z. Jia, C. Tan, Y. Huang, X. Sun and E. Liu, Nanoscale Adv., 2021, 3, 5703–5710 RSC .
25. T. Yan, R. Ding, Y. Huang, D. Ying, C. Tan, Y. Huang, F. Yang, X. Sun, P. Gao and E. Liu, J. Mater. Chem. A, 2021, 9, 14276–14284 RSC .
26. C. Tan, R. Ding, Y. Huang, T. Yan, Y. Huang, F. Yang, X. Sun, P. Gao and E. Liu, Electrochim. Acta, 2021, 389, 138713 CrossRef CAS .
27. D. Ying, Y. Li, R. Ding, W. Shi, Q. Xu, Y. Huang, Z. Jia, W. Yu, X. Sun, P. Gao, E. Liu and X. Wang, Adv. Funct. Mater., 2021, 31, 2101353 CrossRef CAS .
28. F. Yang, R. Ding, Z. Jia, W. Yu, Y. Li, A. Wang, M. Liu, J. Xie, M. Yan, Q. Fang, Y. Zhang, X. Sun and E. Liu, Energy Storage Mater., 2022, 53, 222–237 CrossRef .
29. A. Wang, R. Ding, Y. Li, M. Liu, F. Yang, Y. Zhang, Q. Fang, M. Yan, J. Xie, Z. Chen, Z. Yan, Y. He, J. Guo, X. Sun and E. Liu, Small, 2023, 19, 2302333 CrossRef CAS PubMed .
30. W. Yu, R. Ding, Z. Jia, Y. Li, A. Wang, M. Liu, F. Yang, X. Sun and E. Liu, Adv. Funct. Mater., 2022, 32, 2112469 CrossRef CAS .
31. Z. Jia, R. Ding, W. Yu, Y. Li, A. Wang, M. Liu, F. Yang, X. Sun and E. Liu, Adv. Funct. Mater., 2022, 32, 2107674 CrossRef CAS .
32. M. Liu, R. Ding, Y. Li, A. Wang, F. Yang, Y. Zhang, Q. Fang, M. Yan, J. Xie, Z. Chen, Z. Yan, Y. He, J. Guo, X. Sun and E. Liu, Energy Storage Mater., 2023, 63, 103004 CrossRef .
33. N. Tyagi, E. Ghanti, N. Gupta, N. Lalla and R. Nagarajan, Bull. Mater. Sci., 2009, 32, 583–587 CrossRef CAS .
34. K. K. Lee, P. Y. Loh, C. H. Sow and W. S. Chin, Biosens. Bioelectron., 2013, 39, 255–260 CrossRef CAS PubMed .
35. M. T. Al-Helaly and N. A. Al-Tememee, J. Kufa-Phys., 2021, 13, 33–42 CrossRef .
36. A. I. Rabee, C. B. Gaid, G. A. Mekhemer and M. I. Zaki, Mater. Chem. Phys., 2022, 289, 126367 CrossRef CAS .
37. T. Wu and C. Yuan, Mater. Lett., 2012, 85, 161–163 CrossRef CAS .
38. J. Nouet, D. Toms and J. Scott, Phys. Rev. B: Solid State, 1973, 7, 4874 CrossRef CAS .
39. P. Gao, Y. Zeng, P. Tang, Z. Wang, J. Yang, A. Hu and J. Liu, Adv. Funct. Mater., 2022, 32, 2108644 CrossRef CAS .
40. I. Lorite, J. Romero and J. Fernández, J. Raman Spectrosc., 2012, 43, 1443–1448 CrossRef CAS .
41. V. Hadjiev, M. Iliev and I. Vergilov, J. Phys. C: Solid State Phys., 1988, 21, L199 CrossRef .
42. H. Yang, C. Nie, Y. Shi, H. Yu, S. Ding, Y. Liu, D. Wu, N. Wang and J. Li, Solid State Commun., 2005, 134, 403–408 CrossRef CAS .

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4nr02522a

‡ These authors equally contributed to the work.

This journal is © The Royal Society of Chemistry 2024