High-power lithium-selenium batteries enabled by atomic cobalt electrocatalyst in hollow carbon cathode

The study addresses critical limitations of lithium–selenium (Li–Se) batteries, namely sluggish Se reaction kinetics with Li, volumetric expansion, and polyselenide shuttle that cause rapid capacity fading. While Se offers higher electronic conductivity than S and high volumetric capacity, practical Li–Se cathodes under high power remain underdeveloped. Prior strategies confine Se in conductive porous carbons to reduce resistance and mitigate shuttle, yet long-life, high-rate Li–Se performance has been elusive. Single-atom catalysts (SACs) offer maximal atom utilization, uniform active sites, and unique electronic structures and have improved catalysis in other battery chemistries, but had not been realized for Li–Se due to synthesis challenges. This work hypothesizes that atomically dispersed Co on N-doped hollow porous carbon can catalyze Se redox, immobilize polyselenides, and enable high-power, long-life Li–Se batteries.

Se/carbon composites (e.g., carbon nanospheres, nanofibers, hierarchical and hollow carbons) improve Se cathodes by enhancing conductivity and confining polyselenides, but still fall short at high currents and long cycling. SACs, including MOF-derived systems, have shown superior activity and durability in various electrochemical processes (e.g., metal–air, metal–sulfur batteries). However, controllable incorporation of atomic metals into Se hosts had not been reported for Li–Se. The gap motivates deploying Co single atoms in a hollow N-doped carbon host to combine catalytic activity with effective polyselenide confinement and mass transport advantages.

Synthesis: Polystyrene (PS) spheres were synthesized via emulsion polymerization (PVP-stabilized; KPS initiated). PS@ZIF core–shell particles were formed by growing bimetallic Zn/Co zeolitic imidazolate frameworks (2-methylimidazole ligand) onto PVP-modified PS in methanol. Three Zn/Co precursor ratios yielded PS@ZIF-1 (20:1), PS@ZIF-2 (21:0), and PS@ZIF-3 (17:4). One-step pyrolysis under N2 at 700 °C (5 °C min−1, 5 h) removed PS to form hollow carbons: CoSA-HC (single-atom Co), HC (no Co), and CoNP-HC (Co nanoparticles). Selenium was infused by mixing Se with the carbons (1:1 wt) and heating at 300 °C in Ar to obtain Se@CoSA-HC, Se@HC, Se@CoNP-HC; a high-Se variant used Se:CoSA-HC = 3:1. Characterization: SEM, TEM, HAADF-STEM, aberration-corrected HAADF-STEM, and EDS mapped morphology and elemental distribution; XRD identified phases; N2 sorption (BET, pore size) quantified porosity; Raman assessed graphitization (ID/IG); XPS probed C, N, Co environments; XANES/EXAFS at Co K-edge determined oxidation state and coordination; ICP-OES measured Co content; TGA measured Se loading. Electrochemistry: CR2032 half-cells used Se–carbon composite:carbon black:PVDF = 8:1:1 on Al foil; electrolyte 1 wt% LiNO3 in DOL/DME (1:1 v/v), Celgard 2300 separator, ~20 µL electrolyte, Se areal loading ~0.8 mg cm−2 (and tests at ~5 mg cm−2). Galvanostatic cycling 1.0–3.0 V at 0.1–50 C; CV 0.1–0.5 mV s−1; EIS 100 kHz–10 mHz (5 mV). GITT measured diffusion coefficients; exchange current from EIS via Butler–Volmer. Visual H-cell observation compared polyselenide dissolution using cycled electrodes. Computation: DFT (VASP, PAW, GGA-PBE, DFT-D2) on N-doped graphene (NC) and Co–N–C (Co-NC). Energy cutoff 500 eV; k-point 2×3×1. Gibbs free energies for Se8 → Li2Se reaction steps on NC vs Co-NC; NEB computed Li2Se transformation barriers during charge. Convergence: 1e−5 eV/atom energy, 0.01 eV Å−1 forces.

Materials and structure: Co atoms are atomically dispersed and positively charged within N-doped hollow carbon (Co–N3/N4 coordination; EXAFS Co–N peak ~1.43 Å; no Co–Co peak). Co–N coordination number ~3.3. BET surface areas: CoSA-HC 221 m2 g−1, HC 265 m2 g−1, CoNP-HC 136 m2 g−1. Raman ID/IG ~0.99–1.01. Se loadings by TGA: ~57 wt% (typical) and ~73 wt% (high loading). ICP-OES Co content ~1.3 wt%. Electrochemical performance (Se@CoSA-HC): - Rate capability (from 2nd cycle): 613, 579, 569, 548, 516, 467, 427, 385, and 311 mAh g−1 at 0.1, 0.2, 0.5, 1, 2, 5, 10, 20, and 50 C, respectively; recovery to 537 mAh g−1 when returning to 1 C. Voltage plateaus ~2.0 V; low hysteresis vs controls. - Cycling: 563 mAh g−1 after 100 cycles at 0.1 C (94% retention); 457 mAh g−1 after 1700 cycles at 0.5 C; 237 mAh g−1 after 2500 cycles at 20 C with 0.015% capacity decay per cycle; 267 mAh g−1 after 5000 cycles at 50 C with 0.0067% decay per cycle; Coulombic efficiency ~100% throughout. High-Se loading cells retained 242 mAh g−1 at 0.2 C (100 cycles) and 220 mAh g−1 at 0.5 C (100 cycles). - Kinetics: CV b-values ~0.85 (R1), 0.85 (R2), 0.84 (O), indicating capacitive-dominated behavior; capacitive contribution ~85% at 0.5 mV s−1, increasing with scan rate. GITT-derived DLi+: ~1.02×10−13–1.7×10−13 for Se@CoSA-HC vs ~1.03×10−14–8.35×10−14 for Se@HC. Exchange current density io: 1.14 mA cm−2 (Se@CoSA-HC) vs 0.76 mA cm−2 (Se@HC). EIS shows stable charge-transfer resistance over cycling and formation of a stable interfacial layer. - Mechanistic evidence: XPS after cycling shows C–Se bonding; TEM after 1700 cycles shows preserved hollow structure and a ~20 nm uniform surface layer (assigned to Li2Se2/Li2Se). Visual H-cell test: electrolyte remains clear with Se@CoSA-HC (suppressed polyselenide dissolution) but turns yellow with Se@HC. DFT insights: The rate-determining reduction step is Li2Se2 → Li2Se; Co-NC lowers Gibbs free energy barrier (0.85 eV vs 0.96 eV on NC). During charge, Li2Se transformation barrier is smaller on Co-NC (1.82 eV) than on NC (2.04 eV). Li–Se bond elongation on Co-NC indicates weakened Li interaction facilitating delithiation. Overall, atomic Co sites catalyze both discharge and charge transformations and enhance Se utilization.

The integration of single Co atoms into a nitrogen-doped hollow porous carbon host directly addresses Se redox sluggishness and polyselenide shuttle. Atomically dispersed Co–N sites catalyze the conversion of polyselenides, decreasing thermodynamic and kinetic barriers (lower Gibbs free energy and Li2Se transformation barrier), which accelerates both lithiation and delithiation. The hollow architecture provides large internal voids to buffer Se volume changes, short ion/electron transport pathways, and extensive electrode–electrolyte contact, collectively enabling high rate performance. Experimental kinetics (higher DLi+, larger io, capacitive-dominated CV, stable/low overpotentials, and stable EIS) corroborate the catalytic effect and fast reaction dynamics. The formation of a stable surface layer and strong interactions (C–Se) with suppressed polyselenide dissolution contribute to remarkable long-term cyclability, achieving near-100% Coulombic efficiency and record stability at ultra-high current densities (e.g., 50 C over 5000 cycles). These findings validate the hypothesis that SAC-enabled hosts can unlock high-power, durable Li–Se batteries.

This work presents a facile PS@ZIF templating and pyrolysis route to synthesize nitrogen-doped hollow porous carbon hosting atomically dispersed, positively charged Co single atoms (Co–N3/N4). Embedding Se in this host yields Se@CoSA-HC cathodes with outstanding rate capability (311 mAh g−1 at 50 C) and unprecedented long-term stability (e.g., 267 mAh g−1 after 5000 cycles at 50 C; ~100% Coulombic efficiency). The performance arises from: (i) catalytic Co single atoms that immobilize polyselenides and lower reaction barriers, (ii) hollow architecture accommodating volume changes and enhancing transport, and (iii) conductive carbon confining soluble species. The authors state this provides an efficient route to single-atom materials and a new strategy for high-power electrochemical energy storage devices.