Fast ion transport for synthesis and stabilization of β-Zn₄Sb₃

The study investigates whether ionic mass transport can be harnessed for rapid materials synthesis and whether the interplay of mobile ions with thermal and electric fields can yield microstructures that stabilize phases beyond typical electrochemical stability thresholds. β-Zn₄Sb₃, a mixed ionic–electronic conductor with complex structure and strong thermoelectric performance in the 400–700 K range, serves as the model system. Its performance stems from mobile Zn interstitials that enable high Zn²⁺ mobility and ultralow thermal conductivity, aligning with the Electron-Crystal Phonon-Liquid paradigm. However, broader application is hindered by two issues: the lack of time- and cost-efficient synthesis methods and poor phase stability, including re-entry instability and Zn precipitation under electric fields. The authors propose using fast Zn²⁺ migration under an applied electric field to accelerate synthesis and engineering microstructures—especially amorphous grain-boundary layers—that limit long-range ion migration to stabilize β-Zn₄Sb₃, while exploring Cd/Ge doping to further suppress Zn²⁺ mobility.

Prior work has established β-Zn₄Sb₃ as a promising thermoelectric material due to disordered, weakly bonded Zn interstitials that lower thermal conductivity and enable mixed ion–electron conduction, though plagued by phase instability and electromigration issues. Conventional synthesis involves high-temperature melt growth (~1023 K) or SPS (~700 K). Phase instability (425–565 K), Zn migration during processing (e.g., SPS), and Zn whisker precipitation under electric fields are well documented. Strategies such as Cd and Ge substitution have been explored for thermal stability and band structure modifications. The present work builds on this by leveraging field-assisted ion transport to both accelerate synthesis at low temperatures and generate stabilizing microstructures, while quantifying improvements in ionic conductivity and electromigration resistance.

Electric field-assisted synthesis (EFAS): Stoichiometric mixtures of Zn and Sb powders, as well as Cd- or Ge-doped variants, were loaded into a BN-coated graphite die (16 mm diameter). A thermocouple monitored the powder bed temperature. The system was evacuated to ≤20 Pa. A pulsed sawtooth DC waveform (0.4 Hz) with current amplitude 0–9 A and voltage 1.8–3.6 V was applied for 60 s under load. The maximum measured temperature during current flow was 598 K. The process yielded pellets (~16 mm diameter, ~3 mm thickness) with >98% relative density. Doped compositions included (Zn₁₋ₓCdₓ)₄Sb₃ with x = 0.005, 0.01, 0.015 and (Zn₁₋ₓGeₓ)₄Sb₃ with x = 0.0025, 0.005, 0.0075, synthesized under identical conditions. A reference β-Zn₄Sb₃ was made via melt-quench (1023 K, 12 h, quench) followed by SPS (713 K, 4 min, 30 MPa). Characterization: Phase purity was checked by XRD (Cu Kα). Microstructures were examined by FESEM with EDS, EPMA, TEM (Talos F200s) and aberration-corrected TEM (Titan Themis G2), including in situ TEM biasing using a TEM-STM holder for observing grain boundary evolution under applied voltage. Samples for TEM were prepared by ion milling and FIB. Thermal analysis measured heat flow; thermoelectric properties were assessed using ZEM-3 for σ and Seebeck coefficient, Hall measurements (van der Pauw, ±1.5 T) for R_H and derived carrier properties, thermal diffusivity via laser flash (LFA457), specific heat via Dulong–Petit, and density by Archimedes. Ionic conductivity: Zn²⁺ ionic transport was isolated using DC polarization with electron-blocking zinc-loaded montmorillonite electrodes and corroborated by AC EIS. A solid-state Au | zinc-loaded montmorillonite | Zn₄Sb₃-based sample | zinc-loaded montmorillonite | Au pseudo-galvanic cell was constructed; all samples had similar dimensions (~8 × 8 × 1.3 mm³). Room-temperature ionic conductivities were calculated from polarization data. Electromigration tests: Conducted at 473 K under DC current density of 20 A cm⁻² for 24 h in 100 Pa high-purity Ar to prevent shorting. Post-test fracture surfaces were examined for cracking and Zn precipitation. Process studies: Time-resolved EFAS runs tracked pressure, temperature, punch displacement, current, voltage, and sintering function (displacement derivative) to identify reaction onset and consolidation. Microstructure evolution at discrete times (6, 12, 36, 48, 56 s) was imaged to map phase formation dynamics and development of a 3D ionic network and amorphous grain boundaries.

- Ultra-fast synthesis: Single-phase β-Zn₄Sb₃ and doped variants were obtained within 60 s using EFAS at a maximum measured temperature of 598 K, far below conventional melt growth (~1023 K) and SPS (~700 K). Relative densities exceeded 98%. - Temporal window: β-Zn₄Sb₃ forms rapidly; extending charging time beyond ~83 s induces ZnSb at the upstream side and Zn whisker precipitation downstream. Inverting the sample polarity for 18 s restores homogeneous β-Zn₄Sb₃, indicating β-Zn₄Sb₃ serves as a fast Zn²⁺ transport channel and highlighting a synthesis time window before decomposition. - Microstructure: EFAS produces a core–shell composite with crystalline β-Zn₄Sb₃ grains and amorphous, off-stoichiometric grain boundaries. Upstream boundaries are Zn-poor; downstream boundaries are Zn-rich. A 3D β-Zn₄Sb₃ network forms by ~36 s, facilitating rapid Zn²⁺ transport and densification via nanocrystal formation at ion channel bulges that later quench amorphous. - Transport properties and stability: EFAS samples show highly repeatable TE properties across measurement cycles, unlike MQ+SPS references. Room-temperature electrical conductivity changes after a test cycle: MQ+SPS decreases from 6.7×10⁴ to 4.4×10⁴ S m⁻¹; EFAS pristine from 5.9×10⁴ to 5.8×10⁴ S m⁻¹; EFAS Zn₃.₉₆Cd₀.₀₄Sb₃ and Zn₃.₉₇Ge₀.₀₃Sb₃ remain ~5.3–5.4×10⁴ S m⁻¹. - Ionic conductivity (room temperature): MQ+SPS Zn₄Sb₃: 5.9×10⁻⁵ S m⁻¹; EFAS Zn₄Sb₃: 7.5×10⁻⁶ S m⁻¹; EFAS Zn₃.₉₆Cd₀.₀₄Sb₃: 4.5×10⁻⁶ S m⁻¹; EFAS Zn₃.₉₇Ge₀.₀₃Sb₃: 4.2×10⁻⁶ S m⁻¹. Suppression by >10× relative to MQ+SPS correlates with improved stability. - Thermoelectric performance: EFAS Zn₃.₉₆Cd₀.₀₄Sb₃ achieves ZT_max = 1.2 at 693 K; EFAS Zn₃.₉₇Ge₀.₀₃Sb₃ achieves ZT_max = 1.12 at 700 K. Doping increases ZT across the measured temperature range. - Electromigration resistance (473 K, 20 A cm⁻², 24 h): MQ+SPS shows cracks and Zn whiskers; EFAS pristine shows roughened grains without Zn precipitation; EFAS Cd-doped shows clean grains with few microcracks; EFAS Ge-doped shows clean, crack-free surfaces, indicating strongest robustness. - In situ TEM under bias: In EFAS samples, grain boundaries widen with applied voltage up to a critical level (~0.69 V), after which local crystallization of the amorphous boundary to β-Zn₄Sb₃ occurs and Zn metal precipitates at grain junctions; Zn migrates downstream, and upstream grains can crack from Zn loss. MQ+SPS samples crack and disintegrate continuously upon bias without boundary widening, indicating pervasive ion channels. - Mechanism: Fast Zn²⁺ migration under field forms β-Zn₄Sb₃ nuclei and crystals that connect into a 3D ionic network, accelerating reaction and densification. Rapid cooling upon current cutoff quenches nanocrystals at grain boundaries to amorphous layers that block intergranular ion transport. Cd/Ge substitution on Zn sites further suppresses intra-grain Zn²⁺ migration via steric hindrance, enhancing thermodynamic and kinetic stability.

The findings demonstrate that directed ionic mass transport can be exploited for rapid materials synthesis: under a pulsed electric field, highly mobile Zn²⁺ in nascent β-Zn₄Sb₃ channels enable fast interdiffusion and reaction between separated Zn and Sb, reducing synthesis time to seconds and temperature to below 600 K. Simultaneously, the interplay of electric and thermal fields with ion migration produces self-assembled, core–shell microstructures where amorphous, off-stoichiometric grain-boundary layers impede long-range Zn²⁺ migration across grains. This disrupts the formation of percolating ion channels, suppressing Zn precipitation and re-entry phase instability that have historically limited β-Zn₄Sb₃. Doping with Cd or Ge acts as intra-grain steric hindrance, further lowering Zn²⁺ mobility and improving stability, while maintaining or enhancing thermoelectric performance (ZT ≥ 1.1–1.2 near 700 K). In situ TEM clarifies a critical-voltage behavior: below threshold, amorphous boundaries act as ion blockers; above ~0.69 V, they crystallize locally into β-Zn₄Sb₃, enabling ion transport and eventual degradation. Compared to conventionally processed materials, EFAS confines ion channels within grains, blocking intergranular networks and delivering reproducible TE transport and strong electromigration resistance. The strategy has broader relevance to other mixed ionic–electronic conductors and could be extended to systems like ZnSb, Cu₂Se, and Cu₂S.

A rapid electric field-assisted synthesis (EFAS) was developed to produce dense, single-phase β-Zn₄Sb₃ within 60 s at sub-600 K temperatures. The process leverages fast Zn²⁺ migration to accelerate reaction and fosters a core–shell crystalline–amorphous microstructure with Zn-rich and Zn-poor amorphous grain boundaries that block intergranular ion transport. Combined with Cd or Ge substitution at Zn sites, Zn²⁺ ionic conductivity is suppressed by over an order of magnitude relative to conventional samples, enabling robust, reproducible thermoelectric performance (ZT_max up to 1.2 at ~693 K) and strong resistance to electromigration (20 A cm⁻² at 473 K for 24 h). The results answer key questions about using ionic mass transport for fast synthesis and about microstructural stabilization via field–ion–temperature interactions. Future work could optimize pulse parameters and dopant chemistry to widen the operational window below the critical voltage for boundary crystallization, explore scalability and module-level integration, and extend the EFAS strategy to other mixed-conducting or superionic thermoelectrics and functional materials.

- Temporal and voltage windows: Prolonged charging (>~83 s) or bias beyond a critical voltage (~0.69 V in TEM) leads to decomposition, grain-boundary crystallization, Zn metal precipitation, and eventual instability, indicating a limited processing and operating window that must be carefully controlled. - Residual ionic conduction: Although reduced by >10× in EFAS and doped samples, finite Zn²⁺ mobility remains and may impose constraints under extreme fields, temperatures, or long durations. - Microstructural sensitivity: Stabilization relies on maintaining amorphous grain-boundary layers; conditions that anneal or crystallize these layers can re-enable intergranular ion channels and degrade stability. - Scope and scaling: While preliminary demonstrations suggest applicability to other systems (ZnSb, Cu₂Se, Cu₂S), process transferability, uniformity in larger geometries, and long-term device-level reliability require further validation.