From Protonation and Li‑rich Contamination to Grain‑Boundary Segregation: Assessing the Impact of Solvent‑Free vs. Wet Methods in LLZO Solid Electrolyte Preparation

1. Research Background

Metallic lithium, with its exceptionally high specific capacity (3,860 mAh g⁻¹), has garnered intense interest for next‑generation batteries. Solid‑state lithium batteries (SSLBs), by virtue of their safety and potential energy‑density advantages, are strong contenders to replace commercial liquid‑electrolyte Li‑ion batteries (LiBs). Among oxide solid electrolytes, the garnet‑structured Li₇La₃Zr₂O₁₂ (LLZO) is a leading candidate for lithium‑metal SSLBs due to its electrochemical stability against Li metal and high Li‑ion conductivity.

Conventionally, LLZO is synthesized via a wet route using isopropanol (IPA) as the dispersion solvent. However, IPA‑borne moisture can trigger proton–lithium exchange and Li‑enriched surface contamination during mixing and forming, degrading LLZO performance. To mitigate these issues, Dr. Xiao Huang and colleagues developed a solvent‑free process employing hydrophobic polytetrafluoroethylene (PTFE) and polyurethane (PU) binders. They then conducted a systematic comparison—orthogonal evaluation—of wet vs. dry mixing and forming methods, examining microstructure, phase purity, ionic/electronic conductivity, and electrochemical performance.

2. Experimental Scheme

As illustrated in Figure 1, starting reagents LiOH·H₂O, La₂O₃, Ta₂O₅, and ZrO₂ were combined under four process variants:

• 1WW – Wet mixing × Wet forming
• 2WD – Wet mixing × Dry forming
• 3DW – Dry mixing × Wet forming
• 4DD – Dry mixing × Dry forming

Figure 1. Orthogonal experimental design for LLZO preparation.

Table 1. LLZO preparation parameters

Detailed process parameters are listed in Table 1. The resulting powders were characterized by laser particle sizing, SEM/EDS morphology, and phase analysis. Ionic conductivity was measured with a Toyo LN‑Z2‑HF impedance spectrometer across –55 °C to 25 °C. Li-Li symmetric cells were used for long‑term cycling and critical current density (CCD) tests.

3. Results & Characterization

3.1 Powder Morphology & Surface Chemistry

Figure 2. LLZO preparation process.

The slurry prepared by the wet method using IPA as the solvent exhibited a viscous, cheese-like consistency, resulting in increased processing difficulty. However, the sintered particles showed smaller sizes, better uniformity, and lower tap density. In contrast, dry-mixed powders displayed a flaky structure with larger particles and poorer mixing homogeneity. After sintering, the dry-method powders exhibited larger particle sizes and a tap density twice that of wet-method samples. Particle size distribution tests (Figure 3a) revealed that the as-calcined Ta₅Li₆ powders from wet and dry methods aligned with the morphological observations in Figure 2(b). The trimodal distributions of wet-method samples (1WW and 3DW) matched the SEM-derived morphologies, while the unimodal distributions of dry-method samples (2WD and 4DD) deviated significantly from their actual morphologies. SEM images further demonstrated that wet-method LLZO particles were surrounded by flaky impurities, whereas dry-method powders had cleaner surfaces. This indicates that dry-method samples exhibited significantly reduced protonation compared to the severe surface contamination observed in wet-method counterparts. In summary, the wet method offers advantages in mixing and forming uniformity, effectively reducing particle size due to IPA’s solvent support. The dry method, however, requires less time, simplifies processing, and achieves double the tap density of wet-method powders.

3.2 Phase Analysis (XRD)

From XRD patterns, the calcined Ta₅Li₆ powders were primarily composed of the cubic LLZO phase. However, dry-method samples contained detectable tetragonal phases and La₂O₃ impurities, with XRD peak intensities decreasing more significantly than those of wet-method samples. This may correlate with the broadening of full width at half maximum (FWHM) and the degree of particle amorphization. The solvent’s wetting and grinding media in the wet method induced soft lithium-rich contamination at LLZO particle interfaces. In contrast, the solvent-free dry method effectively disrupted the LLZO lattice. Additionally, the c-LLZO lattice expanded significantly after wet-method grinding, primarily attributed to protonation. Dry-method routes resulted in minimal lattice expansion, likely due to the amorphization of LLZO particles.

Figure 3. Experimental particle size results and XRD test results

Table 2. XRD quantitative phase analysis

3.3 Protonation and Contamination

FTIR and TG-DSC (Figure 4) demonstrated higher weight loss in wet-formed samples (1WW, 3DW) compared to dry methods (2WD, 4DD), confirming severe protonation in solvent-based routes. SEM cross-sections of wet-method pellets revealed visible contaminants, absent in dry-method samples.

Figure 4. FTIR, TG and SEM results

Figure 5. Ceramic sheet and SEM-EDS test results

3.4 Ionic Conductivity and Activation Energy

High-frequency impedance testing from 100 MHz to 1 kHz was employed to measure the ionic resistance of samples at various temperatures. The grain boundary resistance, calculated from the distribution of relaxation times (DRT) results, was fitted and used to determine the activation energy in two temperature ranges: -55°C to 25°C (low-temperature region) and -15°C to 25°C (mid-temperature region). The activation energy data revealed that the interfaces of 1WW, 2WD, 3DW, and 4DD samples exhibited activation energies of approximately 0.3 eV. Such low activation energy values indicate extremely fast lithium-ion transport between the Li anode and LLZO solid electrolyte. Notably, the activation energies obtained in this study are higher than those previously reported in literature, which is primarily attributed to the limited testing frequencies (<1 MHz) of earlier equipment, resulting in underestimated values. This demonstrates that the impedance testing frequency significantly impacts the accuracy of ionic conductivity and activation energy measurements.

Figure 7. Activation energy test results

3.5 Li-Li Symmetric‑Cell Performance

As shown in Figure 8, under long‑term cycling at 0.1 mA cm⁻² current density and 0.1 mAh cm⁻² areal capacity, the polarization voltage of all four ceramics (1WW, 2WD, 3DW, and 4DD) increased only marginally over 2,000 hours, demonstrating stable interfaces with the Li metal anode and no side reactions. In critical current density (CCD) tests at 60 °C, the 4DD sample sustained a high CCD of 1.6 mA cm⁻² at an areal capacity of 1.6 mAh cm⁻², whereas the other three formulations short‑circuited at 1.1–1.2 mA cm⁻².

Figure 8. Li-Li symmetric long cycle and limiting current density test

4. Summary