Electrodeposition of atmosphere-sensitive ternary sodium transition metal oxide films for sodium-based electrochemical energy storage

Significance

Layered sodium transition metal oxides constitute an important class of materials with applications including electrochemical energy storage, high-temperature superconductivity, and electrocatalysis. However, electrodeposition of these compounds, an approach commonly used to grow oxides, has been elusive due to their atmosphere instability and intrinsic incompatibility with aqueous electrolytes. Using a molten sodium hydroxide electrolyte, we demonstrate electrodeposition of O3 (O′3)- and P2-type layered sodium transition metal oxides and apply these electrodeposits as high areal capacity cathodes in sodium-ion batteries. The electrodeposits are micrometers thick, polycrystalline, and structurally similar to materials synthesized classically at high temperature. This work enables fabrication of previously inaccessible alkali and alkaline earth ion intercalated, higher valent transition group oxides in important thick film form factors.

Abstract

We introduce an intermediate-temperature (350 °C) dry molten sodium hydroxide-mediated binder-free electrodeposition process to grow the previously electrochemically inaccessible air- and moisture-sensitive layered sodium transition metal oxides, Na_xMO₂ (M = Co, Mn, Ni, Fe), in both thin and thick film form, compounds which are conventionally synthesized in powder form by solid-state reactions at temperatures ≥700 °C. As a key motivation for this work, several of these oxides are of interest as cathode materials for emerging sodium-ion–based electrochemical energy storage systems. Despite the low synthesis temperature and short reaction times, our electrodeposited oxides retain the key structural and electrochemical performance observed in high-temperature bulk synthesized materials. We demonstrate that tens of micrometers thick >75% dense Na_xCoO₂ and Na_xMnO₂ can be deposited in under 1 h. When used as cathodes for sodium-ion batteries, these materials exhibit near theoretical gravimetric capacities, chemical diffusion coefficients of Na⁺ ions (∼10⁻¹² cm²⋅s⁻¹), and high reversible areal capacities in the range ∼0.25 to 0.76 mA⋅h⋅cm⁻², values significantly higher than those reported for binder-free sodium cathodes deposited by other techniques. The method described here resolves longstanding intrinsic challenges associated with traditional aqueous solution-based electrodeposition of ceramic oxides and opens a general solution chemistry approach for electrochemical processing of hitherto unexplored air- and moisture-sensitive high valent multinary structures with extended frameworks.

Electrochemical synthesis of materials has contributed to significant breakthroughs in materials processing by replacing high-temperature, cost- and energy-intensive pyrometallurgical processes (1). Noteworthy examples include aluminum extraction by Bayer’s process (2), electrowinning of copper (3), titanium extraction through the Kroll process (4), electrolytic production of steel (5), and electrochemical synthesis of cement (6). Additional advantages of electrochemical synthesis include controllable access to metastable polymorphs (7), prevention of crystallite coarsening, and generation of asymmetric, highly oriented structures (8). Specific to the work discussed here on transition metal oxides for electrochemical energy storage, electrodeposition allows direct thermodynamic and kinetic control of the phase formation along with the ability to conformally deposit on topologically complex structures, resulting in improved electron and ion transport kinetics (9). Even though electrodeposition as a technology is well developed for reduction of elements and alloys, studies on multicomponent systems grown by electrolytic oxidation are sparse. Finally, synthesis of materials directly using electrons as the energy packet provides the potential of significant reductions in greenhouse gas emissions and enhanced thermodynamic efficiencies (10).

We focus on materials for sodium-ion (Na-ion) batteries due to their potential cost advantages over lithium-ion systems and the equitable distribution of raw materials required for the synthesis of the cathodes relative to lithium-based systems (11). Of particular interest for Na-ion battery cathodes is O3-type layered NaMO₂ (M = Co, Mn, Ni, Fe) with an alpha NaFeO₂-type crystal structure and octahedrally coordinated transition metal ion. Along with electrochemical energy storage applications (12), this class of oxides has also been studied for its catalytic (13) and superconducting (14) properties. Fabrication of NaMO₂-based battery cathodes involves first high-temperature (>700 °C) synthesis of NaMO₂ for over 24 h, followed by grinding, mixing with binders and additives, and slurry casting. Throughout the process, water must be rigorously excluded. While vapor-phase deposition of Na-ion cathodes by methods including pulsed laser deposition and radio frequency magnetron sputtering has been attempted, deposition of polytypes of Na_xCoO₂ (15–18), alpha Na_xMnO₂ (19), and polyanion compounds (20) has faced processing challenges, only the first of which is maintaining Na stoichiometry under vacuum during growth on a substrate heated to around 700 °C. Additionally, the vapor-phase methods have been limited to cathodes with thicknesses of ≤750 nm and growth rates of only tens of nanometers per hour. As a result, the areal capacities are only a few microamps per hour per square centimeter even at modest discharge currents of order 10 µA⋅cm⁻². The direct electrochemical synthesis of thick layered sodium transition metal oxides (STMOs) at reasonable growth rates and temperatures would overcome the abovementioned challenges and serve as a potential starting point for the scalable manufacturing for Na-ion cathodes. As we will show, such electrochemically grown cathodes also have higher volumetric and gravimetric capacities than slurry-based electrodes due to the absence of binder and conductive additives.

In early work, most ceramic oxides, including binary (21–23) and ternary oxides (24–27), were formed via aqueous electrodeposition. These systems were generally limited to thinner layers (≤5 µm), exhibited slow growth rates (micrometers per day), and exhibited poor crystallinity. An annealing step is often needed to oxygenate and crystallize the deposit. The general mechanism for anodic electrolytic deposition of ceramic oxides (28) involves the complexation of a transition metal ion in an alkaline solution as stable hydroxo-aquo complexes of the type M^(m+)(OH)_x(H₂O)_y (m = 2–3) (details depend on the transition metal ion concentration, charge on the transition metal ion center, pH, and temperature) (29–32). During the anodic oxidation step, M^m+ is electrochemically oxidized and precipitates as a higher valent oxide and/or proton intercalated oxide. The poor kinetics of the reaction and quality of electrodeposits can be traced back to the choice of water as the solvent, the low temperature of synthesis (room temperature to 80 °C), and the low hydrolysis ratio (i.e., OH⁻/M^m+ = 3 to 5 at pH 14 for low transition metal concentrations). Synthesis temperatures of prior electrodeposited oxide/oxyhydroxides by anodic electrolytic deposition and equivalent solid-state synthesis are summarized in SI Appendix, Fig. S1A. A high hydrolysis ratio (>75) (SI Appendix, Fig. S1B) and high temperature of deposition (>250 °C) (SI Appendix, Fig. S1C) synergistically kinetically benefit the growth of thicker (>10 µm) electrodeposits. Moreover, since the O3 and O′3 polytypes of the STMOs most interesting for energy storage applications are air and moisture sensitive (33), aqueous electrodeposition cannot be employed to grow this class of materials.

Here we report a dry molten sodium hydroxide-based solvent-mediated electrodeposition of high-quality thick STMOs with controlled crystallography and phase. By using dry hydroxide salts, we overcome the intrinsic limitations imposed by the inclusion of water in the solvent and broaden the temperature range of the growth conditions. Oxides of transition metals (Co₃O₄, Mn₃O₄, Fe₃O₄, Ni(OH)₂) are first solvated as stable hydroxo complexes in molten NaOH at 350 °C. In comparison to water-based anodic electrodeposition, here, electrodeposition is conducted at an extremely high hydrolysis ratio ([OH⁻]/[M²⁺] = 300:1) and high pH (pH₂O = 10 in Lux Flood bascisity scale) and at an elevated temperature of 350 °C. The combination of these factors improves the kinetics of the deposition rate limiting base catalyzed oxolation (inorganic S_N2) process, due to the extremely high nucleophile concentration [OH⁻], catalyzed by the “base” [OH⁻], and removal of the leaving group H₂O from the bath (as vapor at 350 °C) (31). We specifically demonstrate the direct electrochemical deposition of various thick film (>10 µm) thermodynamically and kinetically controlled highly crystalline STMO polymorphs. Fig. 1 outlines the general electrodeposition scheme. We explore the effects of the applied electrode potential, basicity of the hydroxide solvent medium, transition metal precursor concentration, and temperature of synthesis on the resulting phase assemblage, phase distribution, and electrochemical performance of the electrodeposited Na-M-O phases across various transition metal chemistries.

Fig. 1.

Reaction conditions for molten hydroxide-mediated anodic electrodeposition of layered STMOs of general form Na-M-O (M = Co, Mn, Ni, Fe) at 350 °C. SEM image below the reaction shows ∼60-µm-thick Na_xCoO₂ on a nickel foil (nickel foil is the brighter band on left side of image).

Results

Pourbaix Diagrams and Cyclic Voltammetry of the Na-M-O-H Systems.

Fig. 2 compares the Pourbaix diagrams for the investigated Na-M-O chemistries, corresponding cyclic voltammetry (CV) curves, and the suggested electrodeposition pathways. Pourbaix diagrams are constructed as detailed in SI Appendix, section 1 and ref. 34. Cyclic voltammetry was conducted with the respective transition metal precursor at a molar ratio of 300:1 ([Na⁺]:[M²⁺]) in molten NaOH solvent at 350 °C. The solvent medium follows a Lux–Flood acid base concept where the pH₂O dictates the solvent basicity (35). The self-dissociation constant for the reaction 2OH⁻ ⇌ H₂O +O²⁻ is 10⁻¹⁰ at 350 °C (36–38) and the solvent pH₂O is assumed to be ≃ 10 for all experiments at 350 °C. Electrolytic deposition of the respective STMOs is a function of 1) the chemical stability of the transition metal hydroxo complex in +2 state, 2) the oxidation potential of the transition metal hydroxo complex in comparison to the electrolyte decomposition potential, and 3) the chemical stability of the STMO in the solvent at the synthesis temperature. For a successful electrolytic deposition, the complex must be chemically stable, with the oxidation potential of the M²⁺|M³⁺ being less than that of the electrolyte decomposition, and the electrodeposited STMO must be chemically stable in the electrolyte. The Pourbaix diagrams provide a qualitative representation of the abovementioned factors and the experimental CV curves corroborate the main inferences from the Pourbaix diagrams. We choose the mixed valent oxides M₃O₄ (M = Co, Mn, Fe) as the precursor for electrodeposition, at a concentration less than the solubility limit of M₃O₄ in molten NaOH at 350 °C. M₃O₄ is composed of equimolar quantities of MO and M₂O₃ with M^x+ formal oxidation states of +2 and +3, respectively. A stable M²⁺ hydroxo complex from MO can be electrochemically oxidized to M₃O₄ or NaMO₂ based on the basicity of the solvent and temperature. We aim to electrodeposit M₃O₄-free NaMO₂. By using M₃O₄ as the precursor, at a concentration below its solubility limit NaOH, enables deposition of M₃O₄-free NaMO₂. Basically, should any M₃O₄ be electrodeposited, it remains soluble in the bath and will dissolve in the solvent.

Fig. 2.

A, D, G, and J show the Pourbaix diagrams of (A) Na-Co-O, (D) Na-Mn-O, (G) Na-Ni-O, and (J) Na-Fe-O systems in molten NaOH at 350 °C. Lines 1, 2, 3, 4, and 5 indicate interphase stability between M|MO, MO|M₃O₄, M₃O₄|NaMO₂, OH⁻→O₂, and MO|NaMO₂, respectively. The color of the (MO)_complex domain is representative of the color of the complex in the electrolyte. There exist two domains labeled I and II where oxidation of the +2 complex results in the higher-valent mixed oxide (M₃O₄) and the Na intercalated MO₂⁻ (light yellow), respectively. B, E, H, and K show cyclic voltammograms of the (B) Na-Co-O, (E) Na-Mn-O, (H) Na-Ni-O, and (K) Na-Fe-O systems in molten NaOH at 350 °C for the first (red) and second (blue) cycles at a scan rate of 10 mV⋅s⁻¹. C, F, I, and L present schematics of the possible mechanistic pathway for electrodeposition of the (C) Na-Co-O, (F) Na-Mn-O, (I) Na-Ni-O, and (L) Na-Fe-O systems.

Among the different Na-M-O systems, only Pourbaix diagrams of Na-Co-O (Fig. 2A) and Na-Mn-O (Fig. 2D) show the presence of a thermodynamic zone (light yellow) where complexes of the type Co(OH)₄²⁻ and Mn(OH)₄²⁻ with a formal oxidation state of +2 can be electrochemically oxidized to +3 in zone II before electrochemical decomposition of the electrolyte. CV during the first cycle for the Na-Co-O system (Fig. 2B) shows the electrochemical oxidation of Co²⁺|Co³⁺ at 0.45 V (vs. Mo) (denoted by * in Fig. 2B) and for Na-Mn-O shows the oxidation of Mn²⁺|Mn³⁺ (denoted by * in Fig. 2E) at 0.86 V (vs. W). For the subsequent cycle, the electrodeposited material in the +3 state undergoes further partial oxidation for Co³⁺|Co⁴⁺ at 0.91 V (vs. Mo) and Mn³⁺|Mn⁴⁺ at 1.01 V (vs. W) (denoted by ♣ in Fig. 2E). For all the systems, the first cycle shows the oxidation and partial/complete passivation of the substrate by NiO (marked by ♦ in Fig. 2 B, E, H, and K). Complete passivation is defined by a very low amplitude of the Ni|Ni²⁺ redox current (corresponding to the oxidation of nickel-to-nickel oxide) during the second CV cycle at the same potential as where the nickel is oxidized in the first cycle. The mechanism of deposition for Na_xCoO₂ (Fig. 2C) and Na_xMnO₂ (Fig. 2F) is hypothesized as an electrochemically initiated, flux-assisted electrocrystallization reaction where the electrochemical oxidation (EO) of the transition metal hydroxo complex (from +2 to +3) triggers a base catalyzed oxolation reaction (31) producing precipitation (P) of product insoluble in the solvent electrolyte.

The Pourbaix diagram for the Na-Ni-O system (Fig. 2G) indicates that Ni(OH)₄²⁻ cannot be electrochemically oxidized to +3 before electrochemical decomposition of the electrolyte. CV (Fig. 2H) further corroborates that the Ni²⁺ complex cannot be electrochemically oxidized to Ni³⁺ (Na_xNiO₂) in molten NaOH and there is no change in cycle 1 and cycle 2. Thus, it is not possible to electrolytically deposit Na_xNiO₂ under the adopted solvent conditions. Similar observations have been found in electrolysis of Ni(OH)₂ in a NaOH-KOH solvent at 227 °C (39). As a result, a different approach was adopted for deposition of Na-Ni-O as outlined in Fig. 2I. A pulsed anodic potential of 1.20 V (vs. W) was applied to electrochemically decompose (ED) (in Fig. 2I) the electrolyte, generating O₂ (OH⁻ → H₂O + O₂) near the working electrode. The O₂ subsequently chemically oxidizes and precipitates the Ni²⁺ complex as Na_xNiO₂ on the working electrode.

The Fe²⁺ hydroxo complex is chemically unstable in dry molten NaOH at 350 °C and thus is chemically oxidized to the precipitate Na_xFe_yO_z in this solvent. This is corroborated by the Pourbaix diagram (Fig. 2J), which shows a very narrow stability domain for FeO in a highly basic solvent and a wide domain for NaFeO₂. Chemical oxidation of the Fe²⁺ hydroxo complex in molten NaOH at 350 °C is further supported by CV (Fig. 2K), which shows no stable Fe²⁺|Fe³⁺ redox in the solution after addition of the Fe₃O₄ precursor. As we will discuss at greater length, electrodepositing from the Na-Fe-O system required a different strategy based on controlled addition of CsOH.

Electrodeposition from the Na-Co-O System.

Fig. 3A shows the X-ray diffractogram (XRD) of sodium cobalt oxide (SCO) electrodeposited at 0.6 V (vs. Mo) from 2.0 wt/wt% Co₃O₄ in NaOH at 350 °C. The as-electrodeposited material contains both O3 (major, 97%) and O′3 (minor, 3%) phases. O3 is indexed in a rhombohedral system with space group R$\overline{3}$m and O′3 is indexed in a monoclinic system with space group C2/m (SI Appendix, Table S3A). No Co₃O₄ is found. Inductively coupled plasma (ICP) mass spectrometry measurements (SI Appendix, Table S4A) show the stoichiometry to be Na_0.98CoO₂. The crystal structure of this O3 SCO is visualized using Vesta (40) and presented in Fig. 3B. The refined unit cell parameters (a = b = 2.8855(5) Å, c = 15.6087(9) Å) closely match those reported for solid-state synthesis (41–43). Fig. 3C shows the surface morphology of electrodeposited SCO. Electrodeposited SCO grows as interpenetrating hexagonal platelets (Fig. 3 E and F) with a (003) plane (green), bounded by Na-ion conducting surface terminating (104) (pink), (110) planes (light blue) (Fig. 3D). The platelet type surface morphology has been predicted earlier for O3 LiCoO₂ (44) prepared under synthesis conditions with high oxygen chemical potential (similar to that of OH⁻ in a molten hydroxide flux). Cross-sectional analysis of the sample (Fig. 3E) shows electrodeposits can be grown to ≃ 300 µm thick in 2 h. To understand the physical distribution of the O3 and O′3 phases, a platelet from this sample is analyzed with a 5-nm nanobeam via scanning electron nanobeam diffraction (SEND) over a 500 × 500-nm area across the edge of the platelet (shown in Fig. 3G). The phase map of this area (Fig. 3H) shows the central major part of the platelet is O3 and a small fraction of O′3 is formed at the edge of the platelet (∼60 to 120 nm). Diffraction patterns acquired from the center (light brown square, Fig. 3I), middle (dark brown square, Fig. 3J), and edge (black square, Fig. 3K) of the platelet show a gradual transition of O3 (blue dotted) to O′3 (dark green dotted) near the edge, with the phase fraction of O′3 increasing markedly near the edge. Diffraction patterns are indexed with a [001] zone axis and the unit cell parameters used in the analysis of the diffraction patterns are listed in SI Appendix, Fig. S3A. This observation can be explained from the Na-Co-O phase diagram (41, 45) where O3 NaCoO₂ occurs as a line compound and a small off-stoichiometry leads to the formation of O′3 (0.83 < x < 0.99) in Na_xCoO₂. The appearance of trace amounts O′3 at the edge of the platelet is hypothesized as a result of concomitant charging (desodiation) of the electrodeposited Na_0.98CoO₂ during deposition. This appears predominantly at the edges of the platelet as an intraparticle phase separation occurs from O3 to O′3. We note a 2-h hold in the electrolyte postdeposition chemically converts the trace O′3 to O3, resulting in platelets that are homogeneous in phase (SI Appendix, Fig. S3 B and C) and chemical composition (SI Appendix, Fig. S3D).

Fig. 3.

Electrodeposition from the Na-Co-O system. (A) XRD of electrodeposited O3 layered NaCoO₂. (B) Visualization of the O3 crystal structure electrodeposited at 0.6 V (vs. Mo). Na, Co, and O atoms are color coded in yellow, blue, and red, respectively. (C) Top-view SEM image of electrodeposited dense NaCoO₂. (D) Visualization of the surface terminating planes in the electrodeposited sodium cobalt oxide with platelet morphology. (E and F) Cross-sectional (E) and top-view (F) SEM images of electrodeposited NaCoO₂. (G) TEM image of a section of an electrodeposited platelet. (H) A spatial phase map by SEND from the area (marked by red dotted border in G) shows that the central major part of the platelet is O3 whereas a thin layer of O′3 is formed at the edge of the platelets (∼60 to 120 nm). Diffraction patterns are acquired from (I) center (light brown square), (J) middle (dark brown square), and (K) edge of the platelet (black square). (L) Cyclic voltammetry curves in no precursor, 0.2 wt/wt% Co₃O₄, and 2.0 wt/wt% Co₃O₄ in molten NaOH. (M) Phase map for the electrodeposited samples of the Na-Co-O system as a function of the peak applied voltage and Co₃O₄ concentration. Dashed lines are a guide to the eye.

Via XRD and Na half-cell electrochemical studies, the effect of precursor (Co₃O₄) concentration (Na⁺:Co²⁺ molar ratio) and applied pulse peak potential during plating (Fig. 3M) was investigated. A unique feature of the electrodeposition of layered STMOs is the electrochemical oxidation-induced intraparticle phase heterogeneity and concurrent relaxation of the particle due to back sodiation by the molten sodium hydroxide solvent. Here we demonstrate two ways to capture and identify the state of desodiation/phase heterogeneity of the as-electrodeposited phase assemblage by ex situ XRD (indicated by ▲ in Fig. 3M) and corroborate the measurements with a complementary measurement of the equilibrated open circuit voltage in a Na half-cell (indicated by ▼ in Fig. 3M). Agreement of the two methods by coincidence of two triangles takes the appearance of a star in the graph (Fig. 3M). Open circuit voltage in an equilibrated Na half-cell is particularly useful to indicate the phase assemblage (single-phase solid solution or biphasic mixture) as illustrated in detail in SI Appendix, Fig. S4. CV measurements at the respective precursor concentrations chosen were collected (Fig. 3L). The peaks that we ascribe to the transition metal redox are present only when the transition metal precursors are present. Cyclic voltammograms conducted in pure molten NaOH (black line) do not show any of the redox peaks (Fig. 3L). The Co²⁺|Co³⁺ redox is a function of precursor activity/concentration, ranging from 0.60 V (vs. Mo) at 0.2 wt/wt% Co₃O₄ to 0.45 V (vs. Mo) at 2.0 wt/wt% Co₃O₄, beyond which the redox potential becomes lower due to the higher cobalt activity/concentration. At a high precursor concentration (2.0 to 7.0 wt/wt%) and low overpotential (50 to 250 mV), we observe the growth of O3. Increasing the applied voltage results in formation of an O3 and O′3 phase mixture. At high overpotentials (450 to 550 mV) and low precursor concentrations (0.2 to 2.0 wt/wt%), there is presence of P′3 phase. The observed phase behavior can be rationalized as follows: layered O3 Na_xCoO₂ undergoes electrochemical transformation in the order O3 → O′3 → P′3 (41, 46, 47) as the voltage is increased and sodium is extracted. Thus, an increase in the applied voltage first results in an O3 and O′3 phase mixture. Finally, at high overpotentials both the O3 and O′3 phases form, and some O′3 converts to P′3. Thermal annealing of the as-electrodeposited O3 Na_0.98CoO₂ to 900 °C and quenching the resultant phase arrest the high-temperature P2 polymorph (SI Appendix, Fig. S5), thus enabling access to both electrochemically important polytypes (O3 and P2).

Electrodeposition from the Na-Mn-O System.

Na-Mn-O systems are electrodeposited at a pulsed voltage with a peak potential of 0.97 V vs. W for 2 h at 350 °C from a bath consisting of 3.04 g of Mn₃O₄ in 160 g NaOH ([Na⁺]:[Mn²⁺] 300:1). The resulting electrodeposit is composed of O′3 (α′) Na_1.3MnO_2+x (94.56%), β NaMnO₂ (1.99%), and Cmcm Na_0.67MnO₂ (3.45%) (Fig. 4A). ICP measurements (SI Appendix, Table S4B) show the Na to Mn molar ratio to be 1.3:1.0. Assuming the remainder is oxygen, the stoichiometry of this sodium excess structure is Na_1.3MnO_2+x. O′3 (α′) Na_1.3MnO_2+x is indexed in a monoclinic system with space group C2/m with unit cell parameters (a = 5.6777(6) Å, b = 2.8571(6) Å, c = 5.7597(9) Å, β = 113.06°) (SI Appendix, Table S3B) close to those reported for high-temperature synthesis of α′ NaMnO₂ (48–50). Synthesis diagrams constructed from solid-state reactions do not show presence of α′ Na₁MnO₂ or β Na₁MnO₂ below 600 °C; instead, phases such as Na₃MnO₄ and Na_0.70MnO_2+y form at lower temperatures at Na:Mn molar ratios of 1:1 (50). It has been shown that α and β polymorphs coexist in the room temperature as the difference in formation energy between the two polymorphs lies between 5 and 18 meV (51–53). At a synthesis temperature of 350 °C (K_BT of 53.72 meV), it is possible for the phases to coexist if low-energy stacking faults are present (53). Stacking faults are indeed present as shown in the diffraction patterns of the structures acquired by transmission electron microscopy (TEM) (SI Appendix, Fig. S12). Similar to electrodeposited Na_xCoO₂, concurrent charging of the electrodeposit during deposition results in formation of a partially charged α′ phase (indexed to Cmcm space group). Cmcm Na_0.67MnO₂ is also known to be produced as a partially charged product during charging α NaMnO₂ (48). The crystal structure of the major α′ (O′3) phase is visualized in Fig. 4B and consists of a layered O′3 structure in a monoclinic lattice. Surface morphology (Fig. 4C) of the electrodeposits shows elongated rod-like structures (Fig. 4D). Cross-sectional analysis (Fig. 4F) of ∼75% dense 20-µm-thick electrodeposited Na_1.3MnO_2+x shows a uniform distribution of Na and Mn across the thickness (Fig. 4G). The electrodeposit undergoes a charge-dependent phase transition during deposition (Fig. 4E). The charge reported is integrated from the current vs. time curve on a surface area of 2.25 cm². For 14 mAh of charge passed, there is a dominant presence of a Mn₂O₇ type phase along with α′ (major, O′3) and β (minor, P2) phases. It is difficult to find an exact match for the two peaks occurring at two theta values of 9.80° and 10.98° but Mn₂O₇ provides the most accurate match for the two primary peaks. With the passage of additional charge (23 mAh), the two peaks gradually diminish in intensity. Finally, after 37 mAh of charge the structure finally stabilizes to only α′ (major), β (minor), and partially charged α′ (minor) phases. Thermal annealing of the as-electrodeposited O′3 Na_0.98MnO₂ to 900 °C and quenching the resultant phase arrest the high-temperature orthorhombic P2 polymorph (with space group Cmcm) (SI Appendix, Fig. S11).

Fig. 4.

Electrodeposition from the Na-Mn-O system. (A) XRD of as-synthesized electrodeposit at 0.97 V (vs. W) from 3.04 g of Mn₃O₄ in 160 g NaOH ([Na⁺] or [OH⁻]:[Mn²⁺] = 300:1) at 350 °C showing O′3 (α′) Na_1.3MnO_2+x (94.56%), β NaMnO₂ (1.99%), and Cmcm Na_0.67MnO₂ (3.45%). (B) Visualization of the unit cell of the electrodeposited O′3 (α') phase. Na (yellow), Mn (magenta), and O (red) atoms are color coded. (C) Top-view SEM image of the electrodeposit showing elongated rodlike structures. (D) Visualization of the surface terminating planes in the rodlike structures. (E) XRD illustrating charge-dependent phase transition in electrodeposits from the Na-Mn-O system. (F and G) Cross-sectional SEM image (F) and line EDS (G) of ∼75% dense electrodeposited sodium manganese oxide along dashed line in F.

Electrodeposition from the Na-Fe-Cs-O System.

NaFeO₂ is soluble in alkaline solutions at room temperature and Fe²⁺ disproportionates in alkaline solutions above 60 °C (54). Hence, molten NaOH chemically oxidizes Fe₃O₄ at high temperature (350 °C). This has been illustrated in Fig. 2K where Fe²⁺|Fe³⁺ or higher valent redox systems are found to be absent. To improve the chemical stability of the Fe²⁺ hydroxo complex, the solvent is changed from NaOH to a NaOH-CsOH mixture. Cs⁺, being a more polarizable cation than Na⁺, has a lower dissociation constant in a molten state, which in turn leads to a lower effective basicity in molten NaOH-CsOH relative to pure NaOH. As the basicity of the solvent (Fig. 2J) is reduced, the pH₂O is reduced, and oxidation power of the solvent is no longer sufficient to completely oxidize the Fe²⁺ complex. The stabilized Fe hydroxo complex with a Fe²⁺|Fe³⁺ redox can now be successfully electrochemically oxidized to Na_xFeO₂ and/or CsFeO₂ based on the phase fraction of the hydroxides present. Phase formation at different mole fractions of NaOH, i.e., X_NaOH = m_NaOH/(m_NaOH + m_CsOH) (m indicates the number of moles), is shown with the XRD of the corresponding electrodeposited phases in Fig. 5A and a synthesis plot is constructed in Fig. 5B with a possible mechanistic pathway in Fig. 5C. There is no deposit formation with pure NaOH as solvent. With the addition of CsOH, at X_CsOH = 0.03 (X_NaOH = 0.97), Na_xFeO₂ phases are electrodeposited as two polymorphs, α (major) and β (minor). With the increase of X_CsOH up to 0.08, CsFeO₂ starts to appear in conjunction with the Na_xFeO₂ polymorphs. As the CsOH fraction increases further, even though it stabilizes the Fe²⁺ complex, once X_CsOH exceeds 0.21 (X_NaOH = 0.79), CsFeO₂ becomes the major product. At X_CsOH = 0.65, CsFeO₂ forms as the sole product. There exists only a narrow zone of solvent basicity (X_NaOH = 0.92 < x < 0.97), where Na_xFeO₂ is the only product. The loading achieved in this narrow zone is <0.1 mg⋅cm⁻², limiting the practical utility. Replacing all of CsOH with KOH leads solely to the formation of KFeO₂ (SI Appendix, Table S6A) and changing the NaOH with molar equivalents of NaCl, Na₂CO₃, or combinations thereof (SI Appendix, Table S6B) does not improve the Na_xFeO₂ loading.

Fig. 5.

Electrodeposition from the Na-Fe-Cs-O system. (A and B) XRD (A) and synthesis map (B) created as a function of solvent basicity, i.e., mole fraction of NaOH, X_NaOH = m_NaOH/(m_NaOH + m_CsOH) and applied peak potential. Dashed lines are a guide to the eye. (C) Schematic of the possible deposition mechanism of Na-Fe-O and Cs-Fe-O oxides as a function of X_NaOH.

Thermodynamic and Kinetic Factors of Electrodepositing Sodium Transition Metal Oxides.

Multiple factors, including the redox potential (E_o′ for M²⁺|M³⁺) and chemical stability of the transition metal complex in the +2-oxidation state, the chemical stability of the electrodeposited oxide, and the oxidative electrochemical stability of the electrolyte, strongly influence the viability of phase formation, growth rate, and loading, which varies across different transition metal chemistries. The optimized conditions for the depositions are summarized in Table 1. The criteria discussed in Fig. 2 are favorably present only for the Na-Co-O and Na-Mn-O systems for the solvent chemistries discussed in this work. Layered oxides have multiple temperature-dependent metastable and thermodynamically stable polytypes that are critically linked to synthesis conditions. The electrodeposits of the sodium layered oxide family crystallize in thermodynamically stable low-temperature rhombohedral O3 (or monoclinic O′3) polytype (as in Na-Co-O, Na-Mn-O, and Na-Fe-O), as the major phase along with slight presence of the high-temperature orthorhombic beta polymorph (as in Jahn Teller active Na-Mn-O and Na-Fe-O). The connecting thread across the chemistries grown by electrodeposition presented here is the concurrent oxidation of the deposited material and formation of a partially desodiated phase in conjunction with a primary fully sodiated phase. The challenge of electrodepositing phase pure NaMO₂ lies in the fact that fully sodiated NaMO₂ are line compounds and any compositional variations during the electrosynthesis generate partially sodiated polytypes as minority phases. Attractively, the partially desodiated minority phases are all electrochemically active. If desired, they can be fully sodiated in a Na half-cell with an organic electrolyte. Growth rate and thickness during aqueous electrodeposition of oxide materials at room temperature suffer from intrinsic electronic conductivity challenges of the electrodeposited oxide, resulting in a poor growth rate (approximately nanometers per hour) and a self-limiting thickness (<5 µm). However, layered sodium transition metal oxides are good electronic conductors in the partially and fully sodiated states [∼350 S⋅cm⁻¹ for Na_0.85CoO₂ at 300 K (55), ∼10⁻¹ S⋅cm⁻¹ for Na_0.87NiO₂ at ∼600 K (40), and ∼10⁻⁴ S⋅cm⁻¹ for NaFeO₂ at 350 °C (56)]. The good electronic conductivity of the electrodeposited oxides is reflected in the current vs. time plots (SI Appendix, Fig. S2), which show high currents (60 mA/cm² for Na-Co-O, 95 mA/cm² for Na-Mn-O, ∼620 mA/cm² for Na-Ni-O, and ∼30 mA/cm² for Na-Fe-O) that remain roughly invariant as the oxide layer grows, indicating minimal resistances. The current profiles for Na-Co-O, Na-Mn-O, and Na-Fe-O reflect that the depositions are conducted in a kinetically controlled zone without diffusion limitations. O3 Na_0.98CoO₂ and O′3 Na_1.3MnO₂ can be grown to ∼20 to 300 µm thick at deposition rates of ∼14 to 22 mg⋅cm⁻²⋅h⁻¹ and ∼5.5 mg⋅cm⁻²⋅h⁻¹ for O3 Na_0.98CoO₂ and O′3 Na_1.3MnO_2+x, respectively, while Na_xNiO₂ (∼0.5 mg⋅cm⁻²⋅h⁻¹) and Na_xFeO₂ (∼ <0.1 mg⋅cm⁻²⋅h⁻¹) are formed only to much lower loadings. Coulombic efficiency (CE) is an important parameter to quantify the amount of charge useful for the growth of the electrodeposit. CE for SCO and SMO growth is ∼45 and ∼72%, respectively, while the CE is extremely low (∼1%) for the Na-Ni-O system because almost the entire charge is decomposing the electrolyte and only a small fraction is oxidizing the complex near the substrate. Minimally successful attempts to improve the loading by adopting strategies such as expanding the electrochemical stability window, use of more soluble precursors, and controlled evolution of oxygen are detailed in SI Appendix, section 2.

Table 1.

Optimized electrochemical processing parameters for Na-M-O (M = Co, Mn, Ni, Fe) systems

System	Precursor	Na+:M2+ molar ratio	Electrolyte	Synthesis temperature (°C)	Applied voltage	Phase	Loading (mg/cm2) (2 h deposition)	Faradaic efficiency
Na-Co-O	Co3O4 (3.2 to 11.2 g)	85–300	NaOH (160 g)	350	0.60–0.70 V (vs. Mo)	O3 Na0.98CoO2 (97%)	28.0 ± 3.3	45.0 ± 4.0
							44.4 ± 6.6	46.1 ± 3.2
Na-Mn-O	Mn3O4 (3.04 g)	300	NaOH (160 g)	350	0.97 V (vs. W)	O′3 Na1.3MnO2+x (95%)	11.0 ± 1.4	72.0±-2.1
Na-Ni-O	Ni(OH)2 (6.4 g)	58	NaOH (160 g)	350	1.20 V (vs. W)	O′3 NaxNiO2	1.4	1%
Na-Fe-O	Fe3O4 (3.08 g)	225	NaOH (120 g): CsOH (40 g)	350	0.85 V (vs. W)	α (O3) + β (P2) NaxFeO2	<0.1	—

Electrochemical Performance of As-Synthesized Electrodeposited Materials.

Electrochemical performance of the as-deposited electrodeposits is analyzed in a half-cell configuration with sodium metal reference and counter electrodes. CV at a sweep rate of 0.1 mV⋅s⁻¹ serves to elucidate the phase transformations in the first and second cycles in the Na-M-O systems. All the electrodeposited STMOs (O3 Na_0.98CoO₂, Fig. 6 A, i; O′3 Na_1.3MnO_2+x, Fig. 6 B, i; O′3 Na_xNiO₂, Fig. 6C; and O3 Na_xFeO₂, Fig. 6D) are electrochemically active in the as-synthesized state and show distinct phase transitions corroborating the crystallinity of the as-deposited material. Peaks (and corresponding phase transformations) are observed at similar voltages (vs. Na) for samples prepared by electrodeposition and solid-state synthesis (summarized in SI Appendix, Table S2). The magnitude of currents for the cathodic and anodic reactions reflects the deposit loading. All the O3 (and O′3) systems undergo an irreversible phase transition to O′3 in the low-voltage region (2.68 V in SCO, 2.81 V in SMO) for the first cycle. Electrodeposited O3 Na_0.98CoO₂ shows a first-cycle charge capacity of 132.7 mAh/g (charged to 3.8 V) and a first-cycle discharge capacity of 92.7 mAh/g (Fig. 6 A, ii). This is in close agreement with solid-state synthesized O3 Na₁CoO₂, where charging O3 Na_1.0CoO₂ in the 2.0- to 3.8-V range corresponds to a desodiation of 0.59 mol of Na out of which 0.42 mol can be reintercalated back into the structure due to the irreversible formation of O′3 from O3 (41). The structure becomes electrochemically reversible after conversion of the O′3 to P′3 (SI Appendix, Fig. S10C), resulting in a highly reversible capacity of 61 mAh/g from the 25th to the 100th cycle at C/10. The reversible areal capacity achieved is 0.47 mAh/cm² (Fig. 6 A, iii) with a capacity loss of 6% from the 50th to the 100th cycle. The rate capability of electrodeposited SCO is good at 20 µA (C/20), 40 µA (C/10), 66 µA (C/6), 0.15 mA (C/2), and 2.0 mA (15C) with areal discharge capacities of 0.37, 0.33, 0.29, 0.25, and 0.11 mAh/cm² (sample area of 1.13 cm² and loading of 3.94 mg/cm² [SI Appendix, Fig. S6A]). The diffusion coefficient of Na⁺ ions (1.03 × 10⁻¹² cm²/s in electrodeposited O3 Na_0.98CoO₂) is not adversely affected by the lower synthesis temperature (SI Appendix, Fig. S6 B, i and ii). Electrodeposited O′3 Na_1.3MnO_2+x shows a first-cycle charge capacity of 252.2 mAh/g and a first-cycle discharge capacity of 188.8 mAh/g (Fig. 6 B, ii). A total of 0.85 mol of Na can be deintercalated out of the structure of which 0.63 mol can be reintercalated into the structure in the first discharge cycle. Electrodeposited sodium manganese oxide exhibits discharge capacities of 152.9, 127.4, and 115.5 mAh/g for the 10th, 50th, and 100th cycles, respectively, and the 50th to 100th cycle fade is 9%. The electrodeposited materials deliver reversible areal capacities of the order of ∼0.2 to 0.5 mAh/cm² (Fig. 6 A and B, iii), higher than capacities reported for binder-free sodium cathodes such as grown by vacuum deposition techniques at similar discharge current densities (SI Appendix, Table S5 and Fig. S8A).

Fig. 6.

Electrochemical performance of as-synthesized electrodeposited materials. CV is shown at a sweep rate of 0.1 mV/s in the first and second cycles for O3 Na_0.98CoO₂ (SCO) (A, i), O′3 Na_1.3MnO_2+x (SMO) (B, i), O′3 NaNiO₂ (SNO) (C), and O3 alpha NaFeO₂ (SIO) (D). Sodium intercalation and deintercalation profiles are shown of O3 Na_0.98CoO₂ (A, ii) and O′3 Na_1.3MnO_2+x (B, ii) between 2 and 3.8 V, at a charge and discharge rate of C/10. O3 Na_0.98CoO₂ has a loading of 8.0 mg/cm² and a discharge current of 132 µA/cm² is applied. O′3 Na_1.3MnO_2+x has a loading of 1.4 mg/cm² and a discharge current of 35 µA/cm² is applied. A, iii and B, iii show cycling studies of Na_0.98CoO₂ and O′3 Na_1.3MnO_2+x at a charge and discharge rate of C/10 for 100 cycles.

Discussion

Hydroxo complexes form the building blocks of the electrodeposited layered oxide structure, which on application of an oxidizing potential undergo polycondensation and precipitation. The coordination environment (octahedral/tetrahedral) of the complex is translated from solution to the solid phase and retained in the final product as sodium-ion intercalated edge-sharing polyhedral structures. During the electrocrystallization process, the kinetic pathways are presumed to be of low-energy barrier and the temperature is sufficiently high to aid surface and bulk mobility of adatoms (or adions) to create crystalline deposits. The skewed [Na⁺]:[M²⁺] ratio of 300:1 expands the thermodynamic domain for NaMO₂ stabilization, yielding layered oxides across all transition metal chemistries with none of the higher-valent oxides (as in Co₃O₄, Mn₃O₄, Fe₃O₄) being present in the electrodeposit. This molar ratio, however, does not lead to sodium excess structures in the layered oxides except in the Na-Mn-O system, where layered Na_1.3MnO_2+x is electrodeposited. Moreover, since we use molten NaOH as the solvent, NaOH provides a sodium-ion reservoir and actively prevents loss of sodium during synthesis; thus the Na stoichiometry is maintained at or near the fully sodiated stage. Across the materials, Na-Co-O (Fig. 2C and SI Appendix, Fig. S6), Na-Ni-O (SI Appendix, Fig. S14), and Na-Fe-O (SI Appendix, Fig. S17), the basic building blocks are hexagonal platelets, whereas rodlike structures are observed for the Na-Mn-O system (Fig. 4C).

Na-ion cathode materials synthesized by conventional routes have areal loadings of 1 to 5 mg/cm² and areal capacities in the range of 0.1 to 0.6 mAh/cm². We compare our dense electrodeposited cathodes with state-of-the-art solid-state synthesized Na cathodes (layered oxides and polyanion compounds), directly grown sodium cathodes by sputtering and pulsed laser deposition, and commercially used cathodes for Na-ion batteries in SI Appendix, Table S5 and Fig. S8A. Electrodeposited oxides of the Na-Co-O and Na-Mn-O system can match the electrochemical performance of bulk solid-state synthesized cathodes at even greater loadings of 2 to 10 mg/cm² and reversible capacities in the range 0.2 to 0.5 mAh/cm² and can be grown as high as ∼2.5 mAh/cm² (with 21.0 mg/cm² loading ) (SI Appendix, Fig. S7C). Electrodeposited O3 Na_0.98CoO₂ (SI Appendix, Fig. S7A), O′3 Na_1.3MnO_2+x (SI Appendix, Fig. S13A), and O′3 Na_xNiO₂ (SI Appendix, Fig. S15A) exhibit similar gravimetric capacities and intercalation/deintercalation profiles in the as-deposited state, at the same C rates as compared to the solid-state thermochemically synthesized equivalents. The binder and additive-free nature of the >75% dense electrodes also improve the equivalent gravimetric and volumetric energy density by ∼20 and ∼40%, respectively, due to the absence of low-density polymer binder and carbon-based additives. Even though our cathodes are free of any additives, these exhibit minimal ohmic resistance during charging, which is corroborated by minimal overpotential during the ordered phase transitions at a constant composition, as shown in SI Appendix, Fig. S7A. The directly grown binder and additive-free electrodeposits serve as model cathode systems for all solid-state Na-ion batteries (57). Design principles for electrodeposition of A_xM_yO₂ (A = alkali/alkaline earth, M = one or multiple late transition metals) can be formulated based on the observations reported here. The concept is compatible with multiple cathode chemistries such as NaMO₂ (e.g., M = Cr, V, Ru), which crystallize in an alpha NaFeO₂-type layered O3-type structure. Changing the alkali ion from Na to K or Li would lead to layered oxide cathodes of electrochemical importance such as K_xCoO₂ (58) from a KOH-based molten eutectic. The as-deposited O3-type cathodes can be thermally converted to the corresponding high-temperature metastable P2 polytypes by thermal activation as previously observed in ref. 59. In this study, we focus our attention on the electrochemical properties of the low-temperature thermodynamically stable O3 (or O′3) type of NaMO₂.

As we show here, it is possible to electrodeposit multiple important sodium transition metal oxides for electrochemical energy storage applications. To make further strides in this endeavor, expansion of the oxidative voltage stability window of the electrolyte would be crucial. This would accommodate redox couples with an E^o′ beyond the oxidative decomposition of the electrolyte (such as Ni²⁺|Ni³⁺ demonstrated here in this work). This would enable dense, adherent electrodeposits at a good faradaic efficiency for the electrodeposition process. Moreover, tuning of the dielectric constant and basicity ([O²⁻]) of the solvent would dictate controllable selective precipitation/dissolution of desired/unwanted phases. Some proof-of-concept approaches to address these challenges are demonstrated for Na_xNiO₂ (SI Appendix, section 2) and Na_xFeO₂ (SI Appendix, Table S6 A and B) electrosynthesis.

Conclusion

We have demonstrated an intermediate-temperature (350 °C) molten hydroxide-mediated electrodeposition-based fabrication technique that enables electrosynthesis of layered STMOs at the lowest reported synthesis temperature and reaction time, in the form of thick polycrystalline films. These films successfully cycle, in a binder and additive-free configuration, as intercalation cathode materials for sodium ion batteries with near theoretical gravimetric capacity. Perhaps surprisingly, even when thick, the films exhibit adequate Na-ion and electrical conductivity at reasonable charge and discharge rates. Atmosphere-sensitive oxides including cathodes for Na ion batteries, previously considered incompatible with electrochemical growth, can now be accessed electrochemically by the general paradigm demonstrated here.

Based on the solubility of alkaline/alkaline earth group and late transition group/group 13/low group 14–15 (Sn, Sb, Pb, Bi) oxides in molten salt hydroxides, our method, in principle, can be extended to fabricate oxide films with extended, such as layered and spinel, structures. The processing method can be applied to functional cobaltates, cuprates, and nickelates of interest for high-temperature superconductivity applications. We hypothesize that there exists an unexplored experimental space for electrochemically crystallizing high-valent inorganic structures with extended frameworks from their respective hydro solvents (−OH, −SH) in the form of crystalline solid films via electrochemically initiated, thermally assisted polycondensation reactions.

Materials and Methods

A total of 160 g (4 mol) of NaOH (reagent grade, ≥98%, anhydrous pellets) was dried in an Inconel crucible (250 mL capacity; Sigma) at 150 °C in a vacuum oven for 1 d and subsequently heated to 350 °C on a hotplate in a nitrogen or Ar-filled glovebox atmosphere (oxygen and water level <0.1 ppm), until the NaOH (T_m = 318 °C) solution was clear and dry. Molten NaOH is thermally stable at 350 °C and there is minimal loss of solvent due to evaporation. There is no detectable reaction of the crucible with the molten hydroxide and nickel metal is known to be chemically stable up to 600 °C under Ar or N₂ (60, 61). The crucibles are enclosed in ceramic thermal insulation pads to prevent loss of heat. The transition metal precursor (Co₃O₄, Mn₃O₄, Ni(OH)₂, Fe₃O₄) was added slowly to the molten electrolyte to the desired Na⁺:M²⁺ molar ratio while stirring with a nickel spatula. The bath turns deep blue, light brown, dark green, and reddish brown on addition of Co, Mn, Ni, and Fe precursors, respectively, due to the formation of the corresponding hydroxo complex Co(OH)₄²⁻, Ni(OH)₃¹⁻, Mn(OH)₄²⁻, Fe(OH)₄²⁻. Nickel foil is chosen as the working (2.25 cm²) and counter electrode (3.0 cm²) for the CV experiments (Fig. 2 B, E, H, and K) conducted at a scan rate of 10 mV/s. A potentiostatic pulse waveform with a peak potential was used for deposition: 0.60 V (vs. Mo), 1.20 V (vs. W), 0.97 V (vs. W), and 0.85 V (vs. W) for electrodeposition of O3 Na_xCoO₂, O′3 Na_xMnO₂, O′3 Na_xNiO₂, and O3 Na_xFeO₂, respectively. Duty cycle for all the depositions was 0.25 s t-_on and 8.0 s t-_off to reduce local pH gradients and diffusion limitations. Mo and W wire were used as pseudoreference electrodes for Na-Co-O-based depositions and Na-Mn-O-, Na-Ni-O-, and Na-Fe-O-based depositions, respectively. We note that metals with a melting point >400 °C and low chemical reactivity with NaOH are suitable substrates (working electrodes). Please note that our electrodeposited cathodes can be grown on battery-grade aluminum, stainless steel, and battery-grade nickel foil and are compatible with other Na-ion battery electrolytes such as 0.6 M NaPF₆ in ethylene carbonate:propylene carbonate (1:1 by vol). Most typically used was a nickel foil that was at least 90% submerged into the electrodeposition solution and connected to the potentiostat using a copper alligator clip. For the thermal conversion to the high-temperature beta (P2) phase, the as-electrodeposited O3 NaCoO₂ and O′3 NaMnO₂ are annealed to 900 °C at a ramp rate of 5 °C/min, held at 900 °C for 6 h and quenched.

Electrodeposited samples were washed with dry absolute ethanol (<0.005 wt% water; Sigma) to remove excess solid hydroxide on the surface of the deposit. Note that prior to use as a washing solvent, the as-received alcohol is dried with 3-Å molecular sieves (20% mass/vol) for 1 wk. After washing, samples were dried at 80 °C for 1 d. Electrolyte used in cells was 1.0 M NaClO₄ in propylene carbonate:fluoroethylene carbonate (98:2 by vol%). Two glass fiber filters (Whatman glass microfiber filters, grade GF/F) were used as the separator. Swagelok cells were tested in a half-cell configuration with a sodium metal anode. Sodium cubes were purchased from Sigma, dipped in hexane to remove residual oil, rolled into 20-µm thin Na foils, and punched as 12-mm diameter disks.

X-ray diffraction measurements were conducted in a Bruker D8 Advance with a Cu source in reflection mode for the electrodeposited film samples sealed with a mylar/Kapton film. For structural analysis, the deposits were scraped off, finely ground and sealed inside a capillary tube, and measured in the transmission mode using a Mo source. Rietveld refinements were carried out in Topas5.

Cross-sectional samples were prepared using a Gatan PECS II ion miller in single-modulation configuration with a sample rotation speed of 0.5 rpm at a gun voltage of 8 kV for 1 h with a 0° gun tilt. Scanning electron micrographs were taken using a Hitachi S4700 scanning electron microscope (SEM). Density of the electrodeposited materials was calculated based on the theoretical density of the respective deposited material, the cross-sectional area, the deposit thickness, and the deposit mass.

Cross-sectional SEMs for analysis of cycled samples were performed with a Thermo Scientific Scios 2 Dual Beam focused ion beam (FIB). Before sectioning, the top surface was protected with a 2-µm layer of platinum via ion beam-induced deposition (IBID). Then ion milling was performed with a series of beam currents from high to low with 100 pA as the last step ensuring minimal beam damage-related artifacts.

The TEM images, selected area electron diffraction (SAED) patterns, and SEND data were acquired with a JEOL 2100 Cryo TEM equipped with a LaB₆ emitter operating at 200 kV. SAED patterns were acquired using a camera length of 25 cm. In SEND (62, 63) a nanometer-sized electron beam is scanned across a sample and diffraction patterns corresponding to each spot of the scan are collected. The diffraction patterns and relevant information (e.g., diffraction intensity, intensity ratio between different phases) are collected using a slow scan CCD camera and mapped according to their probe positions. The electron probe used for SEND acquisition is a semiconvergent electron beam of 5 nm (full-width half maximum). The scanning diffraction patterns are acquired over an area of 500 × 500 nm in 25 × 25 pixels, corresponding to a step size of 20 nm. In our SEND experiment, 4 times binning (512 × 512 pixels) is used. The typical exposure time for each diffraction pattern is 0.1 s. A total of 625 diffraction patterns are recorded within 15 min without a beam stop. The obtained diffraction images are further processed to extract structural information.

Data Availability

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