--- name: voltpeer-metrics description: Audit comparability of battery performance numbers and mechanism claims across papers, tables, or figures. Use when cell conditions and denominators matter. --- # Battery evidence and comparison rules For model/tool adaptation or resuming a task, read [the execution guide](references/EXECUTION.md). DeepSeek Flash uses short stages and checkpoints; DeepSeek Pro can plan larger text/evidence batches, with the same scientific checks. Reply in the user's language with the result, usable result/preview links and material unresolved questions. Keep mappings, configuration, logs and recovery records inside the project's `.voltpeer/` folder; do not link them in a normal final reply. Provide the corresponding source record only when the user explicitly requests provenance. Use only capabilities actually available in the current model and host. Use this when extracting performance numbers, building a benchmark, interpreting a mechanism, or assessing translation claims. These are prompts for what to extract, not a universal requirement that every paper report every field. Use [BATTERY_COMPARISON.csv](assets/templates/BATTERY_COMPARISON.csv) when a project has no existing extraction schema. `NR` means *not reported after checking the relevant source*; `NV` means *not yet verified*; `NA` means not applicable. Never infer a reporting gap from an abstract-only paper or fill any gap by guesswork. ## Identify the observation before quoting the number For a numerical claim, record the material and treatment, chemistry and electrode pair, cell configuration (`half`, `symmetric`, `full`) and package (`coin`, `pouch`, etc.) as separate axes, charge/discharge or insertion/extraction branch, cycle index and formation protocol, measurement method, value and uncertainty, and source page/figure/SI. A charge branch is not automatically evidence of fast charging. A plateau fraction at one current is not capacity retention across rates. Record the denominator and boundary: active-material mass, composite electrode mass, electrode area/volume, cell mass/volume, or pack/system boundary. Distinguish specific capacity, areal capacity, energy density, power density, Coulombic efficiency, retention, calendar/cycle life, and throughput. If recalculating, save formula, inputs, assumptions, and propagation of uncertainty. Mark each entry `reported`, `recalculated`, or `modelled`. A theoretical value or projected production design must not be presented as a measured cell result. Typical conditions that can determine interpretation include loading and areal capacity, electrode thickness/density, counter-electrode excess or N/P, electrolyte amount and formulation, separator, voltage window, C-rate definition or current density, temperature, pressure, state-of-charge/depth-of-discharge window, rest periods, cycle/retention baseline, sample count, and error bars. Use chemistry-specific additions below. The [ACS battery reporting checklist](https://pubs.acs.org/doi/10.1021/acsenergylett.1c00870) and the [academic-to-industrial metrics discussion](https://www.nature.com/articles/s41565-019-0371-8) explain why these conditions matter. ## Build a comparison, not a leaderboard Group studies by the question and comparable boundary. Assign an explicit comparability label for *that comparison*: | Label | Meaning | Suitable use | | --- | --- | --- | | A | Same chemistry, configuration, denominator, and key operating conditions are reported and aligned | Direct comparison with residual caveats | | B | Partly aligned; a named variable still differs or is missing | Conditional comparison, sensitivity analysis | | C | Useful to show context or a trend, but not to rank | Qualitative discussion | | D | Different boundary or essential conditions make the requested comparison invalid | Exclude from that quantitative contrast | State the rule used to assign labels and show `NR` and `NV` fields in the table. Do not invent universal pass/fail thresholds for “practical” cells; application, chemistry, and test format change the meaning. Missing methods can themselves be a finding, but calculate reporting rates only for the checked source set and name that denominator. Hard interpretation boundaries: - Half-cell capacity does not establish full-cell energy density; symmetric-cell stability does not establish full-cell lifetime. - Coin-cell results and pouch-cell results are different package claims, independent of the half/full distinction. - “Fast charging” requires the charging branch, a defined time/current and capacity or energy endpoint, and operating conditions. - Cycle life needs the retention numerator and baseline cycle, current, temperature, depth of discharge, and cell type. - Energy density needs measured versus modelled status and the included mass/volume boundary. Material-level, cell-level, and pack-level projections are separate. - Scale-up, manufacturing energy, LCA, safety, and cost require their own system boundaries, assumptions, and source; laboratory measurements and models must be labeled separately. ## Chemistry profiles: load the relevant one - **Lithium-sulfur:** sulfur fraction in composite, sulfur loading/areal capacity, electrolyte-to-sulfur ratio (E/S), lithium excess or N/P, shuttle and retention protocol. Do not rank only by specific capacity per gram of sulfur. See [Li-S benchmarking discussion](https://www.nature.com/articles/s41467-025-60528-4). - **Solid-state batteries:** solid-electrolyte identity and thickness, fabrication pressure and *operating stack pressure* separately, interlayers, anode/reference (including In/Li), temperature, cycling protocol and reproducibility. “Solid-state” does not by itself imply a practical pressure or energy boundary. See [interlaboratory study](https://www.nature.com/articles/s41560-024-01634-3). - **Aqueous zinc:** zinc thickness or inventory, Zn utilization/depth of discharge, cathode/anode balancing, electrolyte amount, loading, side-reaction/gas protocol, and replicates. An oversized zinc foil may hide poor utilization. See [experimental-practice analysis](https://www.nature.com/articles/s41467-022-28381-x). - **Sodium-ion:** hard-carbon initial Coulombic efficiency, sodium inventory and presodiation, N/P, full-cell voltage, and energy normalization. Separate sodiation/desodiation branches and first/subsequent cycles. See [full-cell example](https://www.nature.com/articles/s41467-025-66492-3). - **Flow batteries:** coulombic, voltage, and energy efficiency separately; state-of-charge range, concentration, membrane, current density, capacity utilization, cycle versus calendar time, symmetric versus full-cell setup, and whether pump power is included. See [performance assessment framework](https://www.nature.com/articles/s41560-020-00772-8). For lithium-ion, lithium-metal, multivalent, metal-air, lead-acid, and emerging chemistries, derive an analogous condition set from the actual cell and claim. Add a profile only when repeated work justifies it; avoid assuming one chemistry's metric transfers unchanged to another. ## Mechanism and causal language Separate what was directly measured from a correlation, computational support, or a hypothesis. A DFT adsorption-energy difference alone does not establish a kinetic catalytic pathway. Powder characterization may not represent an operating electrode. Ex situ, in situ, and operando experiments have different perturbations and limits; report those rather than using the labels as a hierarchy of prestige. [Operando methods discussion](https://www.nature.com/articles/s41467-022-32245-9) gives useful cautions. EIS fits are model-dependent; a smaller semicircle or assigned charge-transfer resistance is not stand-alone mechanism proof. Check measurement conditions and fit validity; see [EIS methods commentary](https://www.nature.com/articles/s43246-022-00284-w). In prose and artwork, use calibrated language: `observed`, `consistent with`, `supports`, `suggests`, or `hypothesized`. Record contradictory results and likely causes, including configuration and protocol differences, before declaring consensus.