Electrochemical interfacial catalysis in Co-based battery electrodes involving spin-polarized electron transfer

Significance

Electrode materials undergo dynamic evolution under different operating voltages, and such electrochemically driven transformation further dictates the catalytic reactivity and electrochemical performance. However, the comprehension of the structural and compositional evolution of electrodes with direct visualization and quantification is a grand challenge. Here, we develop a protocol for studying the dynamic evolution of electrodes during the electrochemically driven evolution process by integrating microscopic and spectroscopic analysis, operando magnetometry techniques, and density functional theory calculations. Such a methodology can help optimize interfacial catalysts by recognizing how the evolving properties affect the catalytic reactivity, which leads to an avenue to build next-generation energy systems.

Abstract

Interfacial catalysis occurs ubiquitously in electrochemical systems, such as batteries, fuel cells, and photocatalytic devices. Frequently, in such a system, the electrode material evolves dynamically at different operating voltages, and this electrochemically driven transformation usually dictates the catalytic reactivity of the material and ultimately the electrochemical performance of the device. Despite the importance of the process, comprehension of the underlying structural and compositional evolutions of the electrode material with direct visualization and quantification is still a significant challenge. In this work, we demonstrate a protocol for studying the dynamic evolution of the electrode material under electrochemical processes by integrating microscopic and spectroscopic analyses, operando magnetometry techniques, and density functional theory calculations. The presented methodology provides a real-time picture of the chemical, physical, and electronic structures of the material and its link to the electrochemical performance. Using Co(OH)₂ as a prototype battery electrode and by monitoring the Co metal center under different applied voltages, we show that before a well-known catalytic reaction proceeds, an interfacial storage process occurs at the metallic Co nanoparticles/LiOH interface due to injection of spin-polarized electrons. Subsequently, the metallic Co nanoparticles act as catalytic activation centers and promote LiOH decomposition by transferring these interfacially residing electrons. Most intriguingly, at the LiOH decomposition potential, electronic structure of the metallic Co nanoparticles involving spin-polarized electrons transfer has been shown to exhibit a dynamic variation. This work illustrates a viable approach to access key information inside interfacial catalytic processes and provides useful insights in controlling complex interfaces for wide-ranging electrochemical systems.

Increasing global energy demand has intensified the pursuit of higher performance and more cost-effective energy technologies. The electrochemical interfacial catalysis is generally present in various energy-related devices, which can reduce the kinetic energy barriers and thus improve the electrochemical performance (1–4). Interfacial catalytic reactions are typically accompanied by dynamic evolutions of electrode materials, such as phase/morphological transformations, surface/bulk rearrangements, and conversion reactions, which in turn influence the electrochemical performance at a device level (5, 6). Now that the electrochemically driven structural and compositional evolution complicates such processes, it calls for the development of advanced methodologies to probe the local chemical/physical environment and reveal the detailed catalytic mechanisms. Transition metal and metal oxides have been widely used in a variety of electrochemical catalytic reaction processes because of their high catalytic activity and excellent selectivity (7–10). Nevertheless, in the past decades, electrochemical interfacial catalysis in these materials has not been thoroughly investigated, mainly due to the lack of ideal analytical probes that should possess a high spatial and temporal resolution for their evolution with operating voltage and predictive accuracy for their complex chemistry, as well as operando and noninvasive characteristics to minimize external interferences. As a very powerful analytical technique applied thus far, even high-resolution multinuclear/multidimensional solid-state NMR spectroscopy (11, 12), although it reveals important interfacial structural and chemical information, still fails to provide the dynamic evolution of the electronic structures in interfacial catalysis under operando conditions.

The most challenging part of investigating electrochemical interfacial catalysis is the interfacial electronic structure dynamic evolution under operating conditions (13–15). Taking lithium-ion battery as an example, theoretically, the interfacial catalysis inside an electrode is expected to have significant influences on its resulting capacity, as this process could potentially enable storage reactions that otherwise thermodynamically constrained (16). Up to date, however, experimental proof of electrochemical interfacial catalysis is rather sparse, and its connection to interfacial electronic structure remains a conjecture (6, 17). Establishing such a significant link experimentally requires an operando characterization tool that can provide access to the dynamic electronic structure information during the complicated catalytic processes. Although in situ transmission electron microscopy (TEM) (18, 19), in situ Raman spectroscopy (20, 21), and synchrotron X-ray absorption spectroscopy (22, 23) are able to provide accurate and relevant chemical and physical information, the possibility of directly linking interfacial electronic structure exclusively with the electrochemical interfacial catalysis remains low.

In this work, using advanced operando magnetometry (24–27), we successfully achieved monitoring the real-time evolution of the electronic structure and chemical/physical environment of a transition metal compound involving interfacial catalytic processes. Specifically, a Co(OH)₂/Li battery was selected because its continuous and reversible conversion process takes place under an applied voltage. Moreover, the conversion generates nano-scale metallic Co nanoparticles, which usually exhibit high catalytic activity during electrochemical processes (28–30). By tracing the dynamic magnetic response of the electrode and its implied electronic structure change, we were able to not only reveal the chemical picture of two consecutive conversion reactions in the whole operating voltage range but also ravel the physical picture of electron accumulation and release centered around the transition metal atoms. Furthermore, these findings are corroborated with theoretical analyses covering both thermodynamic models and atomic-level simulations.

Results

The Co(OH)₂ active material used in this study was synthesized by a modified coprecipitation reaction method (31, 32), and its microscopic morphology is shown in SI Appendix, Fig. S1. The coordination structure and the corresponding electronic configuration of the transition metal cation in Co(OH)₂ are illustrated in Fig. 1A. X-ray powder diffraction (XRD) pattern (Fig. 1B), high-resolution TEM image (HR-TEM, SI Appendix, Fig. S2), and Raman spectrum (SI Appendix, Fig. S3) of the as-synthesized material confirm it to be hexagonal Co(OH)₂. Its crystal structure indicates that it is a typical layered compound with large interlayer spacing, constructed by cobalt ions aligned in rows in an octahedral environment of oxygen atoms of hydroxide ion (33) (SI Appendix, Fig. S4). The magnetic hysteresis (MH) curve shows that the present Co(OH)₂ is a paramagnetic material (Fig. 1C).

Fig. 1.

Characterization of the Co(OH)₂ active material and its electrochemical behavior. (A) Coordination structure and corresponding electronic configuration of Co ion. The Co, O, and H atoms are represented by blue, red, and pink balls, respectively. (B) Powder XRD pattern of the as-synthesized Co(OH)₂. (C) Magnetic hysteresis (MH) curve of the Co(OH)₂. (D) Schematic diagram comparing the experimental and the theoretical capacity of the Co(OH)₂/Li battery.

The electrochemical lithium storage activity of the Co(OH)₂ was assessed in a CR2032 coin-type cell with metallic lithium foil as the counter electrode. Cyclic voltammetry (CV) measurements were conducted in the potential range of 0.01 to 3.0 V vs. Li/Li⁺ at a scan rate of 0.5 mV s⁻¹ (SI Appendix, Fig. S5). During the first negative scan, a huge reduction peak centered around ~0.82 V was observed, which can be ascribed to the reduction of divalent cobalt to cobalt nanoparticle (34). Fig. 1D shows the comparison between the measured capacities of the first three cycles and the theoretical capacity (577 mAh g⁻¹) computed for a two-electron conversion in the Co(OH)₂/Li system. In detail, the galvanostatic discharge–charge profiles at a current density of 200 mA g⁻¹ are shown in SI Appendix, Fig. S6. Obviously, the initial discharge capacity of the Co(OH)₂ electrode is ~1,800 mAh g⁻¹ and stabilizes to ~1,200 mAh g⁻¹ for the subsequent cycles, suggesting an additional electrochemical reaction capacity beyond the two-electron conversion process. Even at higher current densities, stable and considerable specific capacities can still be attained (SI Appendix, Fig. S7). Galvanostatic intermittent titration technique (GITT) (35) was employed to gain insight into the kinetic aspects of Li⁺ diffusion in the Co(OH)₂ electrode at various stages (SI Appendix, Fig. S8. A detailed derivation of Li⁺ diffusion coefficient sees SI Appendix, Note S1). Due to the slow kinetics of the conversion reaction, the Li⁺ diffusion coefficient decreases dramatically upon reaching the plateau region. After that, it gradually increases at the end of the conversation. When the voltage reaches the last sloping region, it achieves a steady-state maximum which is possibly associated with interfacial reactions (36, 37) with fast reaction kinetics.

To correlate the electrochemistry process with the evolution of crystal structure and chemical composition, ex situ XRD, X-ray photoelectron spectroscopy (XPS), in situ Raman, and HR-TEM/ Selected Area Electron Diffraction (SAED) studies were conducted. Fig. 2A shows a typical galvanostatic discharge–charge profile of the Co(OH)₂/Li battery cycled at a current density of 200 mA g⁻¹, and ex situ samples from various states of discharge and charge were prepared by setting appropriate cutoff values for voltage during the galvanostatic battery cycling. The resulting XRD patterns (Fig. 2B) show, right after the start of discharge, the intensities of (001) and (101) diffraction peaks of the electrode material decrease rapidly owing to the transformation of crystalline Co(OH)₂ into an amorphous phase. Thus, due to the loss of long-range geometrical order after the early stage of the first discharge, XRD cannot provide much useful insight into the nature of the ongoing electrochemical reactions. So, we further carried out ex situ XPS characterization of electrode material at pristine, fully discharged, and fully charged states, and the results clearly indicate that the original transition metal hydroxide has been converted to a metallic state after discharge (Fig. 2C).

Fig. 2.

Structure and chemical composition evolution of the Co(OH)₂ electrode. (A) Representative charge–discharge profile of the Co(OH)₂/Li battery cycled at 200 mA g⁻¹. (B) A series of ex situ XRD patterns of the Co(OH)₂ electrode during the first cycle. (C) XPS spectra of Co 2p in the Co(OH)₂ electrode during the first cycle (fresh electrode, fully discharged, and fully charged). (D) A series of in situ Raman spectra for one full cycle recorded at the Co(OH)₂ interface. (E) Magnified views of key regions (I and II) marked in (D). (F) HR-TEM images and corresponding SAED patterns taken on the Co(OH)₂ electrodes (1) when discharged to 0.9 V, (2) when discharged to 0.01 V, and (3) when charged to 3.0 V.

The further intermediate species and the reaction process of the deeply discharged Co(OH)₂ were investigated by in situ measurements of Raman spectroscopy. Raman is well suited to probe interfacial reactions over bulk ones in this case due to a signal enhancement coming from the electromagnetic fields located between surface of the formed metal nanoparticles and electrolyte solution. The acquired Raman spectra of the Co(OH)₂ electrode and the Co(OH)₂/Li cell show one prominent characteristic Co(OH)₂ peak at 667.5 cm⁻¹ (SI Appendix, Figs. S9 and S10). A typical collection of in situ Raman spectra is shown in Fig. 2D. The increase of a Raman peak at 617.4 cm⁻¹, which can be assigned to LiOH (38) (also SI Appendix, Fig. S11), distinctly clarifies the complete conversion reaction from Co(OH)₂ to Co and LiOH since further variation in the spectrum cannot be observed. With the proceeding of discharge, another two Raman peaks assigned to lithium-containing species gradually appear and grow (the stretching modes for Li₂O at 576 cm⁻¹ and the vibration mode for LiH around 1,047 cm⁻¹) (39), while the peak intensity for LiOH correspondingly decreases (Fig. 2E). The evolution of these characteristic peaks with potential indicates that the discharge product LiOH can be further lithiated and decompose to form LiH and Li₂O. To confirm the above conclusion, ex situ HR-TEM and SAED measurements were carried out to characterize the Co(OH)₂ electrodes taken at selected states of the charge/discharge process (Fig. 2F). Upon discharged to 0.9 V, the observed characteristic interplanar lattice spacings are measured as 0.218, 0.24 and 0.274 nm, which can be assigned to the (100) plane of Co, the (101) plane of Co(OH)₂ and the (101) plane of LiOH, respectively (40). The SAED pattern recorded from this sample also exhibits apparent diffraction rings of Co(OH)₂, Co, and LiOH, confirming that the initial conversion reaction occurred between Co(OH)₂ and Li. When further discharged to 0.01 V, lattice fringes of Co (0.21 nm), LiOH (0.27 nm) and Li₂O (0.26 nm) are recognized in the HR-TEM image. The SAED pattern obtained from this stage indicates the coexistence of Co metal, Li₂O, LiOH, and LiH (34, 39, 41). Upon charged back to 3.0 V, the resolved lattice fringes with a spacing of 0.23 nm, 0.24 nm, and 0.27 nm are consistent with the characteristic interplanar distance of (002), (101), and (100) planes of Co(OH)₂. The decomposition of Li-containing species (in this case LiOH) has also been observed in other transition metal compound/Li systems and is able to provide appreciable capacity, and the reaction has been suggested to be related to the newly reduced transition metal nanoparticles (11, 39, 42–44).

The electrochemical processes in the Co(OH)₂ electrode were further examined by operando magnetometry. Here, we first assembled a flexible strip-shaped pouch battery with Co(OH)₂ as the working electrode (Fig. 3A), which was then subjected to magnetic tests. Fig. 3 C and D shows its variation in magnetic properties in the potential range of 0.01 to 3.0 V together with the voltage as a function of the cycle time for the initial three consecutive CV cycles. The magnetic response of the Co(OH)₂ electrode changes periodically with the voltage, indicating that the electrochemical reactions are highly reversible and stable. Taking the second cycle as an example, the temporal change operando magnetization data have multiple peaks and valleys, which can be subdivided into six distinct stages based on the voltage changes (marked as V1 to V7, Fig. 3E). When the Co(OH)₂/Li battery is discharged from V1 (3.0 V) to V2 (0.9 V), the reduction of the Co(OH)₂ to metallic Co is manifested by a sharp increase in magnetization (Ms). HR-TEM and SEAD images taken at 0.9 V (Fig. 2F) also indicate that metallic Co nanoparticles have been generated and dispersed inside a LiOH matrix. Subsequently, the increasing of Ms becomes sluggish until the potential reaches V3 (0.6 V), indicating that some additional reactions may present and weaken the magnetic change resulting from the conversion reaction. This unique magnetic response could be explained by a spin-polarized surface capacitance proposed in our previous work (24–27). Once the metallic Co nanoparticles are generated during the discharge process, the spin-polarized surface capacitance appears, where numerous spin-polarized electrons can be injected into the Co nanoparticles within a Thomas–Fermi estimate and form a space charge region, while the Li⁺ ions as a charge compensator are stored outside the grain boundaries and surfaces. Due to the dominant population of spin-down states near Fermi level on the surface of the metallic Co nanoparticles, the electrons preferentially accumulate in these spin-down d-bands (Fig. 3B), suggesting that the overall magnetization should display a reduction trend during the discharge process.

Fig. 3.

Operando magnetometry characterization of the Co(OH)₂ electrode. (A) Schematics of the Co(OH)₂/Li flexible pouch battery. (B) Diagram of the spin polarization state density on the surface of Co. (C) CV curves of the Co(OH)₂ electrode for the first three cycles. (D) Operando magnetic response on the Co(OH)₂/Li battery at an applied magnetic field of 3 T over the potential range of 0.01 to 3 V. (E) Magnified view of the dotted area in (D).

More intriguingly, the Ms undergoes a second increase until the end of the discharge process (V4, 0.01 V). This extraordinary magnetic response is quite unexpected, and it seems that part of the spin-polarized electrons stored inside the metallic Co nanoparticles are released. When the Co(OH)₂/Li battery is charged from V4 (0.01 V) to V5 (0.9 V), the Ms goes down, which is opposite to the observation in the discharge process, demonstrating that a reverse reaction has happened during the charge process. The significant increase of Ms from V5 (0.9 V) to V6 (1.8 V) can be ascribed to the release of spin-polarized electrons from metallic Co⁰, and the reoxidation of metallic Co to Co(OH)₂ upon the charge [V6 (1.8 V) → V7 (3.0 V)] leads to a drastic decrease of Ms once again. Whereas the origin of the magnetic response evolution in the voltage range of 0.01 to 0.9 V remains unclear and requires further exploration. In a previous report, Hu et al. proposed that LiOH could react with Li to form Li₂O and LiH at a range of 0 to 0.85 V in the RuO₂/Li system by high-resolution magic-angle spinning multinuclear/multidimensional solid-state NMR techniques (11). Similarly, our in situ Raman and ex situ TEM analyses of the Co(OH)₂ electrode in the voltage range of 0.01 to 0.9 V have demonstrated that indeed the decomposition reaction of LiOH to LiH and Li₂O has occurred. However, the self-decomposition of LiOH is most likely thermodynamically limited (45), and it is easily verified through control experiments (SI Appendix, Fig. S12). In fact, combined with the unique variation of the magnetization under the LiOH decomposition potential, we speculate that the metallic Co nanoparticles may be able to catalyze the decomposition, which involves the transfer of spin-polarized electrons as we observed from the operando magnetometry.

In order to obtain stronger experimental evidences to verify our hypothesis above, we proceeded to perform operando magnetic measurements between partitions within a given operating potential window and analyzed the corresponding responses. It is worth noting that the Co(OH)₂/Li battery used for the tests was first discharged to 0.01 V in advance. Then, it was charged up to 1.8 V, and cycled between 1.8 V and 1.0 V. Meanwhile, the corresponding CV profiles retain a quasi-rectangle shape with no obvious redox peaks, indicating a nearly ideal capacitive charge storage behavior (SI Appendix, Fig. S13) with fast kinetics. The high-resolution Co 2p XPS spectrum (SI Appendix, Fig. S14) collected at 1.8 V shows two characteristic peaks at binding energies of 778.6 and 794.1 eV, which indicate that the Co does not undergo any oxidation reactions. In this voltage sweep interval, the magnetization decreases from 1.8 V to 1 V during the discharge process and increases again during the subsequent charge process (Fig. 4A), which is in good agreement with the spin-polarized surface capacitance effect (24, 46, 47). We further applied the previously established thermodynamic model (48) to explore the possible electrochemical reactions of the Co(OH)₂ electrode over this specific voltage interval (Fig. 4 D and E and SI Appendix, Fig. S15). All fitting parameters of the extracted parameters in this voltage range are satisfactorily within the framework of the applied thermodynamic model (details are provided in SI Appendix, Note S2), describing Li⁺ ions and electrons separated through phase contact, electrons injected into the metallic Co nanoparticles while Li⁺ ions accumulated on surfaces of the grain boundaries (Fig. 4C). Next, the Co(OH)₂/Li battery was discharged to 0.01 V, and reoperated in the voltage range of 0.01 to 0.9 V, and the corresponding CV curves in this voltage range are shown in SI Appendix, Fig. S16. During this process, the Ms goes up to the maximum value as the voltage reduces to 0.01 V and goes down to the minimum when the voltage rises to 0.9 V (Fig. 4B). However, the XPS data taken at 0.9 V again indicates no change in the valence state of Co (SI Appendix, Fig. S17). The magnetic response obtained from this voltage range is just the opposite to that between 1.0 and 1.8 V voltage range, and it may be closely related to the electrochemical interfacial catalysis of the metallic Co nanoparticles on the decomposition and formation of LiOH. Some reasonable explanations can be put forward as follows: during the discharge from 0.9 V to 0.01 V, the spin-polarized electrons are released from Co 3d orbitals and transferred to LiOH during the decomposition reaction of LiOH catalyzed by the metallic Co nanoparticles to form LiH and Li₂O, leading to an increase in magnetization; while in the reversible process (LiH and Li₂O react to regenerate LiOH), a large number of electrons are released, and these electrons refill into the Co 3d orbitals, resulting in a decrease in magnetization. Since the catalytic activity of transition metals is closely associated with their electronic structures, which can be accurately reflected by their magnetization, therefore, operando magnetometry is able to build the relationship between catalysis and magnetism. Highly sensitive magnetic measurements of the electronic structure could provide direct access to the information of electron transfer in dynamic catalytic processes.

Fig. 4.

Experimental and theoretical evidences for spin-polarized electrons transfer. Corresponding operando magnetic response of the Co(OH)₂ electrode over the potential range of (A) 1.0 to 1.8 V and (B) 0.01 to 0.9 V. (C) Schematical illustration of the formation of a space charge zone. (D) Dependence of lithium interfacial storage on lithium activity, differentiating high (large Q) and low (small Q) storage regions. (E) Dependence of E + (nkBT/e)lnQ on interfacial storage Q, taking account of both chemical and electrostatic storage. Energy profiles for the decomposition reaction of LiOH (F) with Co (111) and (G) without Co (Insets: the optimized intermediate species structures for LiOH + 2Li → Li₂O + LiH, color code: Co in yellow, Li in green, O in red, and H in white). Differential charge density of (H) Co-LiOH interface and (I) Co-Li₂O and Co-LiH interface.

Finally, to provide a more comprehensive picture of the chemical, physical, and electronic structure evolutions during the electrochemical interfacial catalysis of the metallic Co nanoparticles at LiOH decomposition voltage, the density functional theory (DFT) calculations were executed. The Co (111) surface has been a model platform to study chemisorption, we performed similar calculations of LiOH adsorption on a Co (111) surface (49), and its optimized structure is shown in SI Appendix, Fig. S18. The chemisorption of LiOH on Co (111) demonstrates an average adsorption energy of −3.07 eV, which is favorable for its reaction due to their strong adsorption (50) (SI Appendix, Fig. S18). To go a step further, two models of matrix without and with Co atoms were applied in our theoretical simulation. It is considered that the formation of Li₂O and LiH by LiOH and Li is a reversible overall reaction. The energy profiles of LiOH decomposition reaction were calculated for the presence and absence of Co (Fig. 4 F and G and SI Appendix, Table S1). The structure of the optimized intermediates and their free energy spectrum are shown in SI Appendix, Fig. S19. It is observed that the transformations from LiOH to LiO+H are endothermic and the following steps involving the formation of Li₂O and LiH are exothermic. The largest positive energy change occurs during the conversion of LiOH to LiO+H, indicating its role as a rate-determining step throughout the lithiation decomposition process. LiOH decomposition energy of support with Co is much lower than that without Co, indicating that the presence of Co is more advantageous to the lithiation decomposition reaction of LiOH. The calculated LiOH conversion energy barriers of support with Co (0.82 eV) are smaller than that without Co (6.90 eV), which reveals that atomic Co nanoparticles could act as active sites to promote the decomposition of LiOH and enhance the battery performance. In addition, on the basis of a basic understanding of the two phases involved in the reaction, the in-process charge transfer is further studied through charge density difference analyses. The plots of charge density differences provide further insights into the local charge distribution around the metallic Co/LiOH heterointerface, where the yellow and cyan isosurfaces indicate the charge accumulation and depletion regions, respectively (51). Fig. 4H illustrates the electronic density redistribution that occurs during the discharge process of Co/LiOH nanocomposites at low potentials, showing a charge density decrease around the metallic Co nanoparticles and increases around the LiOH. In the following low potential charge process, it can be spotted that Co gains and accumulates electrons, accompanied by the reaction of Li₂O and LiH to form LiOH (Fig. 4I). The theoretical calculation results are well consistent with the magnetic signal variation observed in our operando magnetometry experiments. From a theoretical point of view, the metallic Co nanoparticles filled with spin-polarized electrons indeed exhibit a very good catalytic activity for the decomposition conversion of LiOH during the low-voltage cycling process in the Co(OH)₂/Li battery. The schematic diagram in SI Appendix, Fig. S20 illustrates the dynamic reaction process of the conversion of LiOH on the Co nanoparticles, highlighting the catalytic role of the latter.

Up to this point, a picture of the electrochemical interfacial catalysis in a representative Co(OH)₂/Li battery has been revealed through both microscopic and spectroscopic investigations, and the results are shown schematically in Fig. 5. The overall electrochemical reaction process starts with a pure hexagonal Co(OH)₂ material that is converted into a metallic Co⁰ and LiOH nanocomposite under reducing potential, while a remarkable increase in magnetic response is observed. As the potential further reduces, a spin-polarized surface capacitance is triggered to form, due to the spin-polarized electrons gathering in the Thomas–Fermi shielding length below the surface of the metallic Co nanoparticles. When reaching a lower LiOH decomposition potential, an electrochemical interfacial catalytic reaction occurs at the metallic Co nanoparticles accompanied by the spin-polarized electrons transfer, and LiOH catalytically decomposing to LiH and Li₂O marks the end of this process.

Fig. 5.

Schematics of the electrochemical interfacial catalysis in the Co(OH)₂/Li battery. (A) The summary of reaction pathway. (B) Schematic illustration representing the evolution of phase distribution at each stage of discharge. Color code: Co in yellow, Li in gray, O in red, and H in white.

Discussion

In summary, this work provides an effective means of combining microscopic, spectroscopic analyses, operando magnetometry techniques and DFT calculations, to reveal the local and operational chemical and electronic structure of Co-based battery electrode material during its electrochemically driven evolution process. We demonstrated on a model Co(OH)₂/Li battery to establish the link between the dynamic structural and compositional evolution the unusual electrochemical performance. The dynamic picture established by operando magnetometry also reveals a structured interphase, in which the metal can dynamically inject and release spin-polarized electrons, while the Li-containing species matrix stores Li ions and catalytically decomposes at their respective potentials. Not only does such an interface exists in many transition metal compounds after conversion and can be likewise investigated, in principle, the same methodology could also be used to study other systems involving electrochemical catalysis, such as catalytic ammonia synthesis, CO₂ reduction, and hydrogen evolution reaction.

Materials and Methods

Chemicals and Materials.

All the chemicals and solvents are of analytical grade and use without further purification. Cobalt chloride hexahydrate (CoCl₂·6H₂O, Sigma-Aldrich ≥99.9%), polyvinyl pyrrolidone [PVP-K30, (C₆H₉NO)n average M.W. 58000, Aladdin], and potassium hydroxide (KOH, Aladdin ≥95.0%) were obtained from commercial supplies.

Preparation of the Samples.

The synthesis was carried out in a beaker in ambient air. All chemicals used in this work were used as purchased without further purification. In a typical chemical precipitation process, 1.43 g CoCl₂·6H₂O and 0.8 g PVP-K30 were dissolved in 30 mL deionized water to form a homogenous solution A. In the meanwhile, 0.68 g KOH was dissolved in 15 mL deionized water to form solution B, which was then slowly added to A, accompanied by the formation of pink precipitates. Next, this suspension was placed in an oil bath kept at 90 °C with magnetic stirring for 1 h. After that, the final product was collected by centrifugation and washed alternately with deionized water and ethanol. Finally, the obtained material, which verified to be pure Co(OH)₂, was dried overnight at 60 °C under a vacuum. Commercial LiOH was purchased from Aladdin Reagent Co., Ltd.

Materials Characterizations.

The crystal phase of the as-synthesized product was characterized by X-ray diffraction (XRD, Bruker D8 21 Advance, Germany) with a Cu Kα radiation at 40 kV and 30 mA. To confirm the surface chemical composition and valent states of the product, X-ray photoelectron spectroscopy (XPS) was performed on an ESCALAB250Xi X-ray photoelectron spectrometer. TEM (JEOL, JEM-2100F) and scanning electron microscopy (ZEISS, Sigma 300) were carried out to analyze the structure and morphology of the product. A Thermo Fisher Scientific-DXR2 Raman microscope was employed to acquire Raman spectra under an excitation laser with wavelength of λ = 532 nm.

Electrochemical Measurements.

The Co(OH)₂ electrodes were fabricated using a conventional tape casting method in the following steps. A slurry of 70 wt% active material, 20 wt% conductive additives (Super-P carbon black), and 10 wt% sodium carboxymethyl-cellulose (CMC) dissolved in deionized water was casted by a doctor blade onto a copper foil. Then, the foil was placed in a vacuum oven and dried at 60 °C overnight to allow the residual solvent to evaporate before being punched into circular electrodes (diameter = 1 cm). LiOH electrodes using commercial LiOH power were prepared similarly. The 2032-coin cells were assembled in a glovebox full of high-purity Ar (O₂ and H₂O < 0.1 ppm) with the electrodes, a lithium metal counter electrode, a separator (Celgard 2250 film, Whatman), and a 1 M solution of LiPF₆ dissolved in a 1:1 w/w mixture of ethyl carbonate and dimethyl carbonate. LiOH/Li batteries were prepared and measured under the similar conditions as for the Co(OH)₂/Li batteries.

The galvanostatic charge–discharge measurements were conducted at room temperature between 0.01 and 3.0 V (versus Li⁺/Li) using a NEWARE battery testing system (CT-4008 T-5V20 mA-164, Shenzhen, China). The CV measurements were performed on an electrochemical workstation (IVIUM technologies, Vertex. One. EIS) at a scanning rate of 0.5 mV s⁻¹ within the selected potential. The GITT measurements were carried out in the voltage range of 0.01 to 3.0 V by applying repeated current pulses at a current rate of 0.1 C for 20 min followed by a 1.5-h rest.

Ex Situ XRD and XPS Characterization.

The postcycled electrodes for XRD and XPS measurements were prepared by disassembling the coin cell at appropriate cutoff voltage values in an Ar-filled glovebox (H₂O < 0.1 ppm, O₂ < 0.1 ppm). The working electrodes were removed and rinsed using propylene carbonate (99%, Sigma-Aldrich) repeatedly and then dried in the glovebox to evaporate the solvents. When preparing test samples, the dried electrodes were cut into appropriate sizes and transferred through sealed blue cap bottles. The blue caps were sealed in the glove box and can not be opened before being transferred to the test instrument.

For XRD measurements, the electrodes were measured in the same aforementioned diffractometer, and the XRD patterns were collected from 10° to 80° at a scan rate of 1 min⁻¹. Contributions from the Cu current collector were fitted using experimental data from pristine Cu foil and subtracted. For XPS measurements, the dried electrodes were first sealed into a vessel under argon in the glovebox and exposed to air for less than 1 min during transfer into the XPS ultrahigh vacuum chamber. The XPS spectra were collected using the same aforementioned spectrometer with an Al Kα X-ray source. All resulting spectra were calibrated with the disordered C1s component at a binding energy of 284.8 eV.

Ex Situ TEM Characterization.

The postcycled working electrodes were loaded into 12 mm diameter plastic tubes with anhydrous N-Methyl-2-pyrrolidone and then transferred out from the glovebox. Then the electrodes in NMP were placed in an ultrasound machine for 10 min and the resulting dispersions were quickly drop-casted onto TEM grids. All TEM samples were placed and dried under vacuum at 80 °C overnight and then transported into the TEM column in an airtight vial under argon, in which the air exposure was limited to less than 30 s.

Operando Magnetometry Characterization.

The flexible pouch-type battery for operando magnetization measurements was assembled using the same electrolyte and counter electrode as the above coin cells in a glove box filled with argon at room temperature. For the working electrode, the already-prepared Co (OH)₂ active material /conductive carbon black/CMC slurry with a weight ratio of 7:2:1 was coated onto a rectangular copper current collector and then dried at 60 °C in a vacuum oven for 12 h before use. Polyethylene terephthalate (PET) sheets were used for sealing the battery and making it flexible for the required measurements. All the measurements were performed on a Quantum Design physical property measurement system (PPMS) at 300 K. Initially, these test batteries were cycled and connected in galvanostatic mode at 200 mA g⁻¹ in a selected voltage interval at 300 K, and simultaneously, the magnetization was sampled under a constant magnetic field. The magnetic hysteresis (MH) curve measurements for the electrode material were carried out ex situ. The magnetic background signals from other battery components (such as lithium metal, copper foil, and PET sheets) were subtracted from the total magnetization. The electrochemical data could be collected simultaneously with the corresponding magnetic signal, and the magnetization values given in emu g⁻¹ were defined based on the unit weight of active materials Co(OH)₂.

Theoretical Calculation.

All the calculations were performed within the framework of the DFT as implemented in the Vienna Ab initio Software Package (VASP 5.4.4) code within the Perdew–Burke–Ernzerhof (PBE) generalized gradient approximation and the projected augmented wave (PAW) method (52–55). The cutoff energy for the plane-wave basis set was set to 400 eV. The Brillouin zone of the surface unit cell was sampled by Monkhorst–Pack (MP) grids, and Γ points were used for structure optimizations (56). The convergence criterion for the electronic self-consistent iteration and force was set to 10⁻⁵ eV and 0.01 eV/Å, respectively. We constructed Co materials to explore the reaction of LiOH. The barrier is determined by the CINEB (Climbing Image Nudged Elastic Band) method (57), where three images are considered for our CINEB calculations (58). All of the 3D visualization models are constructed by using VESTA (59).

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

All study data are included in the article and/or SI Appendix.