Focus: Molecular Sieve Regeneration, Electrode Moisture Prevention, and Continuous Glovebox–Automated Line Integration
Keywords: solid-state battery, dry room, dew point control, molecular sieve, electrode drying, glovebox integration, automated manufacturing, moisture management
Abstract
Solid-state batteries (SSBs) impose significantly stricter environmental controls than conventional lithium-ion cells. Sulfide-based solid electrolytes react with atmospheric moisture at concentrations as low as a few parts per million, generating toxic hydrogen sulfide (H₂S) and irreversibly degrading ionic conductivity. This article examines the engineering systems required to maintain sub-−40 °C dew point environments across a high-throughput SSB production line: the design and regeneration cycles of molecular sieve desiccant beds, strategies to prevent electrode moisture ingress, and the mechatronic interfaces that enable continuous gloveboxes to integrate with conventional automated conveyor lines. All technical parameters are drawn from published industry specifications, peer-reviewed literature, and equipment manufacturer datasheets current as of early 2025.
1. Introduction: Why Moisture Is the Enemy
In conventional lithium-ion battery production, dry rooms are typically maintained at dew points between −40 °C and −50 °C (approximately 90–130 ppm H₂O by volume). While already demanding by industrial standards, this is insufficient for sulfide-based SSB electrolytes—such as Li₆PS₅Cl (argyrodite), Li₁₀GeP₂S₁₂ (LGPS), and β-Li₃PS₄—where the tolerable moisture threshold collapses dramatically. Published degradation studies show that LGPS begins to form Li₂S, P₂O₅, and H₂S at relative humidity above approximately 0.01% RH (roughly 10 ppm H₂O at 25 °C), with ionic conductivity losses exceeding 20% after only 30 minutes of exposure at 100 ppm H₂O (Wenzel et al., Journal of Power Sources, 2016; Muramatsu et al., Solid State Ionics, 2011).
The practical implication for process engineers is a two-tier environmental requirement:
- Tier 1 — Dry Room (DR): Dew point ≤ −50 °C (≈ 38 ppm H₂O), suitable for oxide-electrolyte handling and electrode calendering.
- Tier 2 — Ultra-Dry Environment / Inert Atmosphere Glovebox: Dew point ≤ −70 °C (≈ 2.6 ppm H₂O) under N₂ or Ar, mandatory for sulfide electrolyte mixing, coating, and stack assembly.
The challenge is not merely maintaining these conditions in isolated vessels—it is sustaining them across an interconnected, high-throughput automated production line where materials must move continuously between process stations.
2. Fundamentals of Dew Point Control: Molecular Sieve Technology
2.1 Adsorbent Selection
The industry standard for achieving ultra-low dew points is the molecular sieve, specifically synthetic zeolite with a 3Å (3A) or 4Å (4A) pore aperture. The 3A variant (potassium-exchanged zeolite A) preferentially adsorbs water molecules (kinetic diameter 2.65 Å) while excluding most organic solvents and NMP (kinetic diameter ~5 Å), making it ideal for battery electrode drying circuits where solvent vapors must not be co-adsorbed destructively.
Key adsorption performance parameters for 3A beads (per standard supplier data, e.g., Zeochem Z3-01, Honeywell UOP MOLSIV 3A):
| Parameter | Typical Value |
|---|---|
| Static water capacity at 25 °C, 60% RH | ≥ 20 wt% |
| Achievable equilibrium dew point (regenerated bed) | −70 °C to −80 °C |
| Crush strength (3.2 mm bead) | ≥ 55 N |
| Bulk density | 640–720 kg/m³ |
| Operating temperature range (adsorption) | 5–50 °C |
| Maximum regeneration temperature | 250–350 °C |
Silica gel and activated alumina can achieve dew points of −40 °C to −60 °C and are commonly used in first-stage dehumidification of incoming makeup air, but they cannot reach the sub-−70 °C targets required for sulfide processing. Molecular sieves are therefore reserved for final-stage polishing beds feeding glovebox purification loops.
2.2 Pressure Swing vs. Temperature Swing Adsorption
Two regeneration architectures dominate industrial SSB dry room design:
Pressure Swing Adsorption (PSA)
PSA cycles between high adsorption pressure (typically 4–7 bar) and low desorption pressure (near atmospheric or vacuum). Cycle times are short—30 seconds to 5 minutes per half-cycle—enabling compact, continuous output. PSA is well-suited to circulating gas loops inside gloveboxes, where a small sidestream of purified N₂ is continuously polished. The limitations are energy consumption (compressing N₂ to 6 bar consumes roughly 0.35–0.45 kWh per kg of dry gas) and reduced effectiveness for deep polar water removal below −60 °C dew point without refrigerant pre-cooling.
Temperature Swing Adsorption (TSA)
TSA is the workhorse of large-scale dry room dehumidification. Twin (or triple) rotating bed systems—the most widespread being desiccant rotors (honeycomb wheels) or fixed-bed twin-tower configurations—alternate between adsorption and regeneration phases.
In a twin-tower TSA system:
- Adsorption phase: Moist process air passes through Tower A (loaded with molecular sieve) at 20–40 °C while Tower B undergoes regeneration.
- Regeneration phase: A hot purge stream (typically 220–280 °C for molecular sieve; silica gel requires only 120–150 °C) drives water off the saturated bed; the released moisture is vented or condensed and recovered.
- Cooling phase: After thermal regeneration, the bed is cooled by a dry purge before returning to service. Inadequate cooling is a leading cause of transient dew point spikes at tower switchover—a critical failure mode in battery production.
Typical twin-tower TSA cycle:
- Adsorption: 4–8 hours
- Regeneration (heating): 1.5–3 hours at 260 °C
- Cooling/standby: 1–2 hours
- Total cycle: 7–13 hours
2.3 Molecular Sieve Regeneration: Engineering Details
Effective regeneration determines whether a system can sustain sub-−60 °C dew points over months of continuous operation. Incomplete regeneration is cumulative: each cycle leaves a residual water load that progressively reduces bed capacity—a phenomenon known as capacity fade. Industry experience indicates that a well-managed TSA system using 3A molecular sieve should retain >90% of initial capacity after 3,000 cycles (~5 years at one cycle per day) if the following parameters are respected:
Regeneration temperature: 260–290 °C for 3A sieve. Below 220 °C, strongly bound water in the inner crystallite lattice is not fully desorbed, causing premature capacity fade. Above 350 °C, the aluminosilicate framework begins to dealuminate, permanently reducing adsorption capacity by 15–30% within 50 cycles.
Purge gas flow rate: The regeneration purge (typically dry N₂ in inert-atmosphere systems, or dry air bleed in open systems) should deliver a specific regeneration gas velocity of 0.15–0.30 m/s through the bed to ensure complete sweep of desorbed water vapor. Insufficient purge flow at the correct temperature still results in incomplete regeneration.
Moisture load at switchover: Beds should be switched before reaching 70–80% of equilibrium capacity. Allowing full saturation causes channeling and accelerates mechanical degradation of beads due to repeated expansion/contraction cycling.
Bead replacement interval: Even with optimal operation, bead replacement every 3–5 years is standard maintenance practice for dry room systems producing sub-−60 °C output. Fines generated from bead attrition can contaminate downstream electrode webs and glovebox purification trains, so pressure drop monitoring across the bed provides the earliest warning of mechanical degradation.
3. Electrode Moisture Ingress: Prevention Architecture
3.1 The Electrode Vulnerability Window
Electrode moisture uptake is not a single-event risk but a process-stage-specific exposure map. Each manufacturing step carries a distinct moisture ingress potential:
| Process Stage | Moisture Risk Vector | Recommended Max Dew Point |
|---|---|---|
| Active material storage (powder) | Packaging breach, transfer between vessels | −40 °C (sealed transfer only) |
| Slurry mixing (NMP solvent) | Ambient humidity during open mixing | −30 °C DR (NMC/LFP); −50 °C (sulfide) |
| Slot-die coating | Web surface during wet-film open time | −40 °C DR |
| Drying oven (IR/convection) | Controlled: oven purged to −50 °C | Oven atmosphere −50 °C min |
| Calendering | Compressed roll contact, minimal exposure | −40 °C DR |
| Slitting | Short open exposure | −40 °C DR |
| Sulfide electrolyte lamination | CRITICAL: sulfide reacts at >10 ppm H₂O | ≤ −70 °C, inert atmosphere mandatory |
| Cell stacking/winding | Sulfide contact continues | ≤ −70 °C, glovebox |
| Electrolyte filling (liquid hybrid SSB) | Open cell cavity | −50 °C DR min |
| Formation cycling | Sealed cell, minimal risk | Standard clean room |
3.2 Electrode Drying: Inline vs. Batch Protocols
Batch vacuum drying remains the most common final moisture-reduction step before electrodes enter the inert atmosphere zone. A representative cycle for NMC cathode electrodes (coated on Al foil, 150 μm dry film):
- Temperature: 120–150 °C (higher temperatures risk binder degradation for PVDF-based electrodes above 180 °C)
- Pressure: ≤ 1 × 10⁻² mbar
- Duration: 12–16 hours
- Target residual moisture: < 50 ppm by Karl Fischer titration (KF)
- Transfer protocol: vacuum hot-transfer directly into antechamber at <−60 °C dew point within 5 minutes of oven door opening
For sulfide composite electrodes (solid electrolyte + active material + carbon, no binder solvent), the drying target is stricter: < 20 ppm residual moisture, with vacuum drying at 100–120 °C (higher temperatures risk phase decomposition of some sulfides) followed by immediate transfer.
Inline continuous drying is increasingly deployed in high-volume lines. A multi-zone heated roller conveyor inside a sealed N₂-purged tunnel allows the coated web to pass through progressive temperature zones (80 °C → 120 °C → 140 °C), with final moisture verification by inline near-infrared (NIR) moisture sensors before entering the glovebox transfer module. Key NIR sensor specifications for inline use:
- Detection range: 0–5,000 ppm H₂O in thin film
- Measurement accuracy: ±20 ppm in the 0–200 ppm range
- Response time: < 1 second (compatible with web speeds of 10–30 m/min)
- Example products: Metrohm NIRFlex N-500, Guided Wave ClearView db
3.3 Moisture Monitoring Instrumentation
Accurate dew point measurement is the foundation of the control system. The two dominant sensor technologies deployed in SSB production are:
Chilled mirror hygrometers: The reference standard. A mirror is cooled until condensation forms; the surface temperature at the dew point is measured by a precision RTD. Accuracy: ±0.1–0.2 °C dew point. Limitations: slow response (30–120 seconds) and susceptibility to contamination in solvent-laden atmospheres. Used primarily for calibration reference and critical glovebox monitoring.
Aluminum oxide capacitive sensors (AlOx): The production workhorse. A porous aluminum oxide layer changes capacitance with water vapor partial pressure. Response time: 5–30 seconds. Accuracy after calibration: ±2–3 °C dew point in the −80 °C to −20 °C range. Major suppliers: Vaisala (HMT330 series, DMP series), GE Sensing (Optica series), Michell Instruments (Easidew, Optidew). Require annual recalibration and replacement every 2–3 years due to sensor drift from organic vapor contamination.
Tunable diode laser absorption spectroscopy (TDLAS): Increasingly adopted for high-value glovebox circuits. Measures water vapor directly by laser absorption at 1.39 μm. Accuracy: ±1 ppm H₂O in the 0–100 ppm range. No physical contact with the gas stream; immune to organic contamination. Expensive (USD 15,000–40,000 per measurement point), but justified at glovebox purification train inlets and critical transfer lock interfaces.
4. Continuous Glovebox and Automated Production Line Integration
4.1 The Integration Challenge
The fundamental tension in SSB manufacturing automation is this: roll-to-roll or sheet-to-sheet production equipment operates in ambient or dry room (−40 to −50 °C DP) conditions, while sulfide electrolyte processing requires an inert-atmosphere environment (N₂ or Ar, ≤ −70 °C DP, O₂ < 1 ppm). Every material transfer between these two environments is a potential moisture ingress event, and every mechanical interface is a potential leak path.
Three architectural approaches have emerged in the industry:
Architecture A — Island Gloveboxes with Manual/Semi-Automated Transfer
Discrete gloveboxes housing specific high-sensitivity operations (electrolyte mixing, stack assembly) connected to the main automated line via batch antechamber (airlock) transfers. Lowest capital cost, but creates production bottlenecks and relies on operator discipline for transfer timing and moisture control. Suitable for pilot lines and R&D (< 100 cells/day).
Architecture B — Tunnel Glovebox (Linear Continuous Glovebox)
A sealed, gas-purified tunnel 10–50 m in length integrates directly with the automated conveyor system. The electrode web or cell stack enters through a dynamic seal at one end, travels through the inert atmosphere zone while undergoing multiple process steps (electrolyte coating, stacking, initial compression), and exits through a second dynamic seal. This is the architecture adopted by most high-volume SSB pilot lines announced as of 2024–2025, including those operated by Toyota, Samsung SDI, and CATL.
Architecture C — Full Factory Inert Atmosphere (Ultra-Large Glovebox)
The entire manufacturing floor—or major sections of it—is enclosed and operated under inert gas at controlled moisture levels. Extremely high capital cost (USD 50–200M+ for a GWh-scale facility) and complex safety management for large N₂/Ar volumes. Considered the end-state for mature high-volume SSB production but not yet demonstrated at GWh scale as of early 2025.
4.2 Dynamic Seals: The Critical Interface
The entry and exit interfaces of a tunnel glovebox—where a moving web, carrier plate, or robot arm passes through the inert atmosphere boundary—are the most mechanically challenging elements of the entire moisture management system.
Lip seal curtains are the most common primary barrier: multiple layers of flexible EPDM or silicone rubber sheets with precision-slit openings sized to electrode web dimensions (typically ±0.5 mm clearance). A multi-stage curtain system (typically 3–5 stages) with N₂ purge gas flowing counter to the electrode web direction in the interstitial space creates a differential pressure barrier. Purge N₂ flow rates of 50–200 L/min through each interstitial stage maintain a positive outward pressure differential of 2–5 Pa, preventing ambient air ingress even during momentary seal deformation from web tension fluctuations.
Roller seal + gas curtain hybrid systems use hard-anodized aluminum nip rollers held to close tolerance (10–20 μm gap) against the web, supplemented by a N₂ knife directed at the web surface. This reduces particle generation compared to lip seals—important in cleanroom environments—at the cost of higher purge gas consumption (300–600 L/min per interface).
Lock chamber (airlock) modules are used for sheet and plate materials (solid-state pouch cell stacks, prismatic cell plates) that cannot use continuous web-fed seals. A dual-gate airlock with an interlock control system ensures both gates are never simultaneously open. Cycle time is typically 60–180 seconds per batch. The critical design parameter is the purge-down time—the time required to reduce the lock chamber atmosphere from ambient (20,000 ppm H₂O) to <100 ppm H₂O before opening the inner gate. With a properly sized N₂ purge flow, this is calculated as:
text复制下载
N_purge_volumes = ln(C_initial / C_target)
= ln(20,000 / 100)
= 5.3 purge volumes
For a 0.1 m³ airlock with 200 L/min N₂ purge, minimum purge time = (5.3 × 100 L) / 200 L/min = 2.65 minutes, plus additional margin for mixing efficiency (practical factor 1.5–2×, giving 4–5 minutes). This calculation is a standard reference in dry room engineering (see IEC 62660 series on lithium cell manufacturing environments).
4.3 Glovebox Purification Train Design
Once the inert atmosphere is established, it must be continuously regenerated to maintain target levels. A typical production-scale glovebox purification train (serving a 20 m tunnel glovebox with 8 m³ internal volume) consists of:
Gas circulation blower: Recirculates the internal atmosphere at 2–5 box volumes per hour. Typical: centrifugal blower, 200–500 m³/h, with brushless DC motor (eliminates carbon brush contamination).
Copper oxide (CuO) catalyst bed (for O₂ removal): O₂ reacts with H₂ (added in controlled ppm quantities to the circulating gas) over CuO catalyst pellets to form water, which is then captured by the downstream molecular sieve bed. Inlet O₂ concentration: up to 1,000 ppm; outlet: < 1 ppm. Operating temperature: 200–300 °C (exothermic reaction; heat management is critical to avoid catalyst sintering).
Molecular sieve (3A or 4A) desiccant bed: Final polishing step. Twin beds in TSA configuration. At production-scale glovebox purification loads (moisture ingress primarily from web seals and operator glove permeation), bed regeneration is required every 24–72 hours depending on leak tightness and production throughput.
Oxygen/moisture analyzer: Continuous monitoring at the purification train outlet. Alarm thresholds: O₂ > 5 ppm (warn), > 20 ppm (process hold); H₂O > 5 ppm (warn), > 10 ppm (process hold for sulfide-only zones).
Solvent trap (activated carbon or chilled condenser): Upstream of the catalyst/sieve train to prevent NMP, ethanol cleaning solvents, or electrode binder off-gases from poisoning the CuO catalyst or blocking molecular sieve pores.
4.4 Control System Integration
Modern SSB production lines use a distributed control system (DCS) or SCADA architecture with dedicated moisture management as a primary process variable (PV), not a facility utility afterthought.
Key control loops and interlocks:
Dew point cascade control: The dry room air handling unit (AHU) output dew point (setpoint: −50 °C) feeds forward to the glovebox purification train flow rate controller. If the AHU dew point rises above −45 °C (indicating increased moisture load), the purification train circulation rate automatically increases before the glovebox atmosphere is affected.
Transfer lock interlock logic: Web speed on the electrode conveyor is slaved to the dynamic seal N₂ purge differential pressure. If purge ΔP drops below the minimum threshold (e.g., 2 Pa) due to N₂ supply pressure fluctuation or seal wear, the conveyor decelerates within 200 ms to reduce the leakage cross-section, while an alarm alerts maintenance.
Molecular sieve bed state machine: The twin-bed regeneration controller operates a state machine with five states: ADSORB, REGEN-HEAT, REGEN-PURGE, COOL, STANDBY. Transitions are triggered by either timed cycles or by outlet dew point rising above −60 °C (sensor-triggered early switch). Data logging of every cycle’s peak outlet dew point, regeneration temperature profile, and purge flow provides the trend data for predictive maintenance scheduling.
MES (Manufacturing Execution System) integration: Every electrode reel, cell stack, or batch lot is tagged with the environmental conditions (dew point, O₂ level, temperature) under which it was processed, linked by barcode or RFID scan. This traceability record is essential for correlating post-formation capacity or impedance anomalies with specific moisture excursion events during production.
5. Practical Failure Modes and Mitigations
Based on published case studies and engineering practice, the following failure modes are the most frequently encountered in SSB pilot line commissioning:
5.1 Tower Switchover Dew Point Spikes (TSA Systems)
Description: At the moment of bed switchover in a twin-tower TSA dehumidifier, a brief pulse of elevated dew point (sometimes 5–15 °C above setpoint) occurs as the recently regenerated/cooled bed reaches full adsorption efficiency.
Root cause: Insufficient cooling phase duration, leaving residual heat in the regenerated bed. When switched to adsorption, the warm sieve initially has lower affinity for water, causing transient breakthrough.
Mitigation: Extend the cooling phase by 20–30 minutes; verify bed exit temperature < 40 °C before switchover. Install a 5-liter surge buffer volume (filled with pre-dried 3A sieve) immediately downstream of the twin-tower outlet to dampen transient spikes. Set an alarm at −48 °C to trigger an early switch cycle before the spike propagates to the electrode web.
5.2 Glovebox Seal Wear and Progressive Moisture Ingress
Description: After 2–4 weeks of continuous production, lip seal curtains show wear groove formation at the web contact point, increasing leak rate and raising glovebox dew point by 2–5 ppm per week.
Mitigation: Implement a weekly seal inspection protocol with optical profilometry of the seal gap. Replace seals when measured gap exceeds 1.5× the original tolerance. Switch to roller seal + gas curtain hybrid for webs with abrasive coatings (e.g., ceramic-coated sulfide electrodes). Maintain a 5–10 Pa positive pressure differential in the glovebox relative to the adjacent dry room to ensure leakage direction is outward, not inward.
5.3 Catalyst Poisoning in the Purification Train
Description: CuO catalyst deactivation due to trace sulfur compounds (from H₂S generated during sulfide electrolyte processing) reduces O₂ removal efficiency. A glovebox that previously maintained < 1 ppm O₂ begins drifting to 5–10 ppm over 3–6 months.
Mitigation: Insert a ZnO guard bed upstream of the CuO reactor to scavenge H₂S (ZnO + H₂S → ZnS + H₂O; capacity ~20 wt% sulfur). Monitor H₂S concentration at the glovebox interior with an electrochemical sensor (detection limit 0.1 ppm); alert threshold 1 ppm. Schedule ZnO guard bed replacement every 6 months in sulfide-processing lines.
6. Capital and Operating Cost Benchmarks
The following figures are representative order-of-magnitude estimates drawn from equipment supplier quotes and published accounts from battery industry consultancies (IDTechEx, Yano Research, Fraunhofer IWS) as of 2024:
| System Component | Approximate Capital Cost (USD) | Annual OpEx (USD) |
|---|---|---|
| Dry room AHU (1,000 m³/h, −50 °C DP) | 180,000–350,000 | 40,000–80,000 (energy) |
| Twin-tower molecular sieve TSA unit (same duty) | 80,000–160,000 | 15,000–35,000 (energy + bead replacement) |
| 20 m tunnel glovebox with integrated purification train | 1,200,000–2,500,000 | 120,000–200,000 (N₂ makeup gas, maintenance) |
| Dual airlock transfer module (0.1 m³) | 45,000–90,000 | 8,000–15,000 |
| TDLAS moisture analyzer (per point) | 18,000–40,000 | 1,500–3,000 (calibration, service) |
| Chilled mirror hygrometer (reference) | 8,000–20,000 | 2,000–4,000 (calibration) |
Energy consumption is the dominant operating cost. A 1,000 m³/h dry room at −50 °C DP requires approximately 150–200 kW of continuous dehumidification power; scaling to −70 °C inert atmosphere within the glovebox adds roughly 25–40 kW of purification train electrical load per 20 m tunnel section.
7. Outlook: Emerging Technologies
Electrochemical dehumidification using proton exchange membrane (PEM) cells to electrolytically pump water vapor from the process gas is under active development (e.g., Desiccant Technologies Group, Perma Pure). Unlike TSA, electrochemical dehumidifiers have no moving parts, no thermal regeneration energy, and provide continuous output. Laboratory prototypes achieve −80 °C DP at energy consumptions of 0.05–0.10 kWh per kg H₂O removed—3–5× more efficient than thermal regeneration. Commercial products at production-relevant scale (> 100 m³/h) are expected to reach market by 2026–2027.
AI-based predictive moisture management is beginning to appear in GWh-scale lithium-ion facilities and will transfer to SSB lines. Neural network models trained on historical dew point time series, production schedule data (web speed, electrode thickness, binder type), and ambient weather conditions can predict molecular sieve saturation 6–12 hours ahead, enabling proactive bed scheduling that avoids reactive emergency regenerations. Early deployments at Korean and Chinese cell manufacturers (disclosed at the Battery Japan 2024 conference) report a 15–20% reduction in N₂ consumption and near-elimination of unplanned dew point excursions.
8. Conclusion
Dew point control in solid-state battery manufacturing is not a peripheral facility concern—it is a core process variable that directly determines electrolyte integrity, cell yield, and long-term cycling performance. The molecular sieve remains the fundamental enabling technology for achieving sub-−60 °C dew points, and its proper regeneration—controlled temperature, adequate purge flow, and timely bed switching—is what sustains that performance over years of production. Preventing electrode moisture ingress requires a holistic approach: multi-zone drying, inline NIR verification, and instrumented transfer protocols. The integration of continuous tunnel gloveboxes with automated conveyor systems demands careful engineering of dynamic seals, differential pressure management, and interlocked control logic. As SSB production scales from pilot to GWh, these systems will require the same level of engineering rigor as the electrochemical cell design itself.
References
- Wenzel, S., et al. (2016). “Direct observation of the interfacial instability of the fast ionic conductor Li₁₀GeP₂S₁₂ at the lithium metal anode.” Nano Energy, 26, 176–184.
- Muramatsu, H., et al. (2011). “Changes in lithium ion conductivity and crystal structure of Li₄GeS₄–Li₃PS₄ mixed sulfide solid electrolytes.” Solid State Ionics, 182(1), 116–119.
- Janek, J., & Zeier, W. G. (2016). “A solid future for battery development.” Nature Energy, 1(9), 16141.
- IEC 62660-2:2010. *Secondary lithium-ion cells for the propulsion of electric road vehicles — Part 2: Reliability and abuse testing.* International Electrotechnical Commission.
- Zeochem AG. (2023). *Molecular Sieve Product Data: Z3-01 3Å Adsorbent.* Zeochem AG, Switzerland.
- Vaisala Oyj. (2024). DMP8 Dewpoint Meter Product Guide. Vaisala Oyj, Finland.
- IDTechEx Research. (2024). *Solid-State Batteries 2025–2035: Technology, Players, Markets and Forecasts.* IDTechEx Ltd.
- Fraunhofer IWS Dresden. (2023). Dry Room Technology for Battery Cell Manufacturing: Engineering Guidelines. Internal Technical Report (public summary).
- Manthiram, A., Yu, X., & Wang, S. (2017). “Lithium battery chemistries enabled by solid-state electrolytes.” Nature Reviews Materials, 2(4), 16103.
- Battery Japan Conference Proceedings. (2024). *Session 4B: AI-Driven Environmental Control in GWh-Scale Cell Manufacturing.* Reed Exhibitions Japan.