Interface-centred co-design of solid-state lithium batteries: An evidence-graded assessment of electrolytes, electrodes and ai-guided discovery
Year: 2026
Type: article
Abstract: Solid electrolytes are widely expected to replace flammable liquid electrolytes in lithium batteries, yet cell-level progress has lagged behind the large gains in bulk ionic conductivity achieved over the past decade. This review argues that the electrode–electrolyte interface, rather than the bulk electrolyte, should be the unit of design, and examines the consequences of this view for engineering practice and for artificial intelligence (AI)-guided discovery. Using a documented, non-systematic literature selection (mainly 2018–2026, with earlier foundational work), we first re-examine the safety premise: removing the liquid removes one fuel, but thermal-runaway pathways remain through cathode oxygen release, lithium-metal reactivity, and stored electrical energy. We then compare oxide, sulfide, halide/oxyhalide, solid polymer, gel/quasi-solid, and composite electrolytes, together with the cathodes and anodes they must contact, explain how solid–solid interfaces form and degrade chemically, electrochemically, and mechanically, and critically assess the experimental techniques used to diagnose them. An area-specific resistance analysis indicates that once separators are about 30 µm thick and conductivities exceed ∼1 mS cm⁻¹, interfacial and composite-electrode resistances, rather than bulk conductivity, usually dominate ohmic loss. Three ‘translation gaps’ are identified in which bulk metrics or computational descriptors (ionic conductivity, thermodynamic stability windows and elastic-modulus criteria) fail to predict cell behaviour. AI studies are graded on two complementary scales: the type of validation (Levels L1–L4) and the cell-level validation scale (grades A–D). Experimentally validated successes are concentrated in bulk properties and low-dimensional synthesis optimisation, whereas interface prediction remains largely computational, trained on small datasets and bounded by an experimental ‘noise floor’ (inter-laboratory relative standard deviations of 35–50% in the conductivity of identical samples). We propose an Interface Co-Design Framework with four operationally defined validity gates, chemical pairing, mechanical contact, transport translation, and manufacturing and safety validation, together with measurable reporting criteria and a Minimum Interface Data Record. AI is likely to accelerate practical solid-state batteries only if its training labels describe the failure modes that limit cells and if electrolyte, electrode, interface and processing are optimised jointly rather than sequentially.
Source: Results in Engineering
Language: English
Subfield: Electrical and Electronic Engineering
Field: Engineering
Domain: Physical Sciences
Keywords: solid-state lithium batteries, electrode-electrolyte interface, solid electrolytes, thermal runaway, oxide electrolytes, sulfide solid electrolytes, solid polymer electrolytes, gel electrolyte, composite electrolytes, area specific resistance, interfacial resistance, battery safety, ionic conductivity
Open Access status: gold
Estimated APC paid: $1,940
OpenAlex ID: W7225271207
