{"id":"3d73b79b-5296-4a5b-a1bd-6a769131baf4","arxiv_id":"2607.17832","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Multisine EIS matches single-sine EIS in steady state and stays valid during cell operation, exposing charge/discharge kinetics asymmetry that steady-state impedance cannot see.","lead":"Multisine impedance spectroscopy (many frequencies at once) measures the same Li-ion battery impedance as the standard single-frequency method in steady state, in a fraction of the time. The paper shows it also keeps working during charging, discharging, relaxation, and temperature changes—where single-frequency checks produce distorted low-frequency data.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Operando impedance rests on an unvalidated trend-removal step [60]; KK compliance, the only reported check, is insensitive to the nonstationarity the operando spectra themselves exhibit.","rationale":"The paper's strongest and best-supported claim is the steady-state equivalence: Fig. 6 shows overlap at multiple SOCs, Fig. 7 gives measurement-model residuals with uncertainty bounds, and matched rms excitation addresses the natural confound. My concern is confined to §4.3. All operando impedance values are outputs of the [60] trend-removal method, which is not re-derived or independently validated here, and the KK pass the paper reports cannot certify stationarity by the paper's own admission. This is exactly the reader's weakest assumption. I see no reason to move the CONDITIONAL verdict; rather, the proposed synthetic recovery test and a replicated cell would upgrade or reject the operando claims cleanly. I therefore agree with the reader and recommend no change to the verdict.","tokens_in":73907,"tokens_out":5224,"duration_ms":55441,"concrete_test":"Using the raw operando current/voltage bursts (or a re-measurement with the same protocol), run a synthetic recovery test: inject a known time-varying impedance (e.g., a Randles circuit whose Rct ramps linearly and then quadratically by several mΩ over a 200 s burst) into the measured current-voltage data, and process with the [60] trend-removal routine. If the recovered impedance deviates from the known windowed impedance by more than the Appendix B 95.4% uncertainty band, or if the result changes when the number of basis functions is varied over a plausible range, the §4.3 operando spectra cannot be considered unbiased. If recovery is unbiased and basis-order independent, the drift-removal concern is settled. A repeat of Fig. 11 on a second cell would additionally confirm the asymmetry is not cell-specific.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The steady-state equivalence is well supported, but the operando leg of the central claim passes through the trend-removal algorithm of Hallemans et al. [60]. Section 4.3 states that the reported charge/discharge, relaxation, and temperature spectra are obtained by 'modelling a linear time-variation and removing the drifts and transients' with that method, and that the results satisfy Kramers-Kronig through the measurement model. The basis-function model is not specified in this paper, no independent validation or sensitivity analysis is shown, and KK compliance is admitted in §4.2 (citing You et al. [59]) to be insensitive to nonstationarity—the very condition the operando spectra exhibit as 'skirts' in Figs. 10 and 13. If the drift basis or the assumed linear time-variation partially absorbs the evolving low-frequency or charge-transfer response, then the reported operando impedance is a model-dependent artifact rather than the true linearized cell impedance. In addition, the charge/discharge asymmetry (Fig. 11), relaxation arc growth (Fig. 12), and temperature trend all rest on a single cell without replicates, so they cannot yet be separated from cell-to-cell or run-order variability.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents a practical demonstration of multisine electrochemical impedance spectroscopy (EIS) on a modified commercial potentiostat, applied to a commercial Li-ion cell over 20 mHz–1 kHz. It compares single-sine and multisine impedance under steady-state conditions at several SOCs, showing overlap, and verifies linearity/stationarity using non-excited spectral bins, Lissajous plots, and an external measurement-model/Kramers-Kronig tool. It then presents operando multisine measurements during C/3 charge/discharge, relaxation, and a temperature ramp, using a drift/transient-removal procedure from an earlier paper, and interprets the resulting spectra in terms of Butler-Volmer linearization around nonzero current, relaxation dynamics, and thermal activation. The paper argues that multisine excitation resolves the erroneous low-frequency impedance produced by sequential single-sine excitation during operation.","tokens_in":74026,"tokens_out":4308,"duration_ms":54553,"significance":"If the operando results are valid, the paper would be a valuable practical demonstration that broadband multisine EIS on commercial hardware can characterize battery dynamics away from steady state, with direct time-series validation of linearity and stationarity. The steady-state leg is the strongest part: the comparison uses rms-matched excitations, the spectra show non-excited bins at the noise floor, the Kramers-Kronig check is done with an external measurement-model tool, the 1 kHz point is transparently excluded, and Appendix B provides a period-to-period variance estimate. These are concrete strengths. The operando leg is weaker: it depends on a trend-removal method that is only cited, not specified or revalidated here, and the reported physical conclusions derive from a single cell and single run.","major_comments":[{"comment":"The operando impedance data in Figs 10–13 are obtained by ‘modelling a linear time-variation and removing the drifts and transients using the method of Hallemans et al. [38,60]’. This basis-function procedure is not specified in this paper, no sensitivity analysis is shown, and no independent validation is provided. The only reported check is Kramers-Kronig compliance through the measurement model, yet Section 4.2 itself states (citing You et al. [59]) that Kramers-Kronig analysis is insensitive to the type of nonstationarity seen as ‘skirts’ in Figs 10 and 13. The central operando claim therefore currently rests on an unvalidated, model-dependent step. Please specify the trend-removal model and demonstrate, e.g. with synthetic time-varying impedance or an independent benchmark, that the drift/transient basis does not partially absorb the evolving low-frequency or charge-transfer respons","section":"Section 4.3, Figs 10–13"},{"comment":"The physical conclusions — charge/discharge asymmetry at equal SOC, growth of the charge-transfer arc during relaxation, and the temperature trend — are all derived from a single cell with a single repetition. With one cell and one run order, the observed asymmetries and trends cannot be separated from cell-to-cell variability or run-order effects. This is load-bearing for the claim that the charge/discharge asymmetry is a genuine kinetic effect (Butler-Volmer linearization around nonzero current). Please add replicate measurements, at minimum repeated runs on the same cell for the charge/discharge comparison, and report scatter or confidence intervals; ideally use multiple cells.","section":"Section 4.3, Figs 11–13"},{"comment":"The comparison between single-sine and multisine operando data is not fully on equal footing: the single-sine impedance is computed directly from the spectra without drift removal, while the multisine impedance is computed after drift/transient removal. The text says that even after drift removal the single-sine data still fail Kramers-Kronig, but the corresponding processed single-sine spectra are not shown. To support the claim that ‘multisine excitation resolves this issue’, please display single-sine operando impedance after the same or an equivalent drift-removal treatment (or explain quantitatively why the same treatment cannot be applied), so the comparison isolates the benefit of simultaneous excitation from the benefit of postprocessing.","section":"Section 4.3, Fig. 10"}],"minor_comments":[{"comment":"Please define T_transients(f_m) explicitly; currently it appears in both expressions without a formal definition.","section":"Section 2, Eqs (3)–(4)"},{"comment":"The anti-aliasing filter is stated to have cut-off 65 kHz while the sampling frequency is 15.625 kHz (Nyquist ≈7.8 kHz). The wording is imprecise; if the signal bandwidth is limited by the 1.6 kHz reconstruction filter, say so explicitly, otherwise readers may question the anti-aliasing design.","section":"Section 3.2"},{"comment":"The legend labels ‘C/3 charge’ and ‘C/3 discharge’ plus ‘Steady-state’ are clear, but the corresponding symbols for charge vs discharge are difficult to distinguish in the printed Nyquist plot; please increase marker differentiation or use separate panels.","section":"Fig. 11"},{"comment":"The variance formula is stated without derivation. A brief explanation of the noise assumptions (stationary, zero-mean, independent of signal) and why the covariance term enters with this sign would help readers apply the result.","section":"Appendix B, Eq. (15)"},{"comment":"The low-frequency increase in the voltage spectrum is attributed to ‘a small drift signal’. This is acknowledged, but it would be useful to state whether that drift is quantitatively negligible for the impedance estimate, given that non-excited bins are otherwise at the noise floor.","section":"Section 4.1"}],"recommendation":"major_revision","confidential_remarks":"The steady-state comparison is convincing and could support acceptance after revision, but the operando leg of the paper — which is central to the title and conclusions — is not yet established without explicit validation of the trend-removal procedure and at least repeat measurements. The issues are fixable within the manuscript's scope, so I recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know two things about this one. First, the steady-state claim is solid: on the same cell, with rms-matched excitation and the raw current/voltage captured from a modified commercial potentiostat, multisine and single-sine produce overlapping impedance spectra across multiple SOCs. The checks are the right ones — non-excited bins at the noise floor, measurement-model Kramers-Kronig via Orazem's external software, period-to-period variance, and a transparent exclusion of the 1 kHz point. Second, the operando section is the weaker leg. The charge/discharge, relaxation, and temperature spectra are all produced by the authors' own trend-removal method, cited as [60] but not specified or re-validated here, and the only reported sanity check for those results is Kramers-Kronig compliance, which the paper itself admits (§4.2, citing You et al.) is insensitive to nonstationarity — precisely what the 'skirts' in Figures 10 and 13 show. That is a real circularity in the operando claim, not a manufactured one.\n\nWhat the paper does well: it gives practical design guidance for odd-random-phase multisines, demonstrates the hardware/anti-aliasing improvements that fix aliasing errors in the authors' earlier work, shows the SOC-swing advantage of multisine over low-frequency single-sines, and is unusually candid about what Kramers-Kronig can and cannot detect. The acknowledgement about the previous aliasing error is a mark of honesty.\n\nSoft spots, in proportion: the load-bearing drift-removal step is a citation, not a derivation; the single cell with no replicates means the charge/discharge asymmetry and relaxation arc growth are not yet separated from cell or run-order variability; the abstract promises a dataset that is not linked anywhere in the text; and no processing code is released. These are structural, not fatal. The steady-state equivalence and the single-sine-versus-multisine guidance stand on their own, and the operando results are plausible but need the drift-removal sensitivity analysis and replicate data before they can be taken as physical findings.\n\nVerdict: send to peer review. The methods contribution and the validation style deserve referee time. A good referee will ask for the drift-removal sensitivity study, the dataset/code, and at least one more cell. This is for measurement-methodology practitioners and battery characterization labs, and it would be a useful reading-group paper — the discussion of how to verify linearity/stationarity in raw spectra is worth having on the table.","headline":"A careful, honest methods paper: steady-state multisine vs. single-sine equivalence is nailed, but the operando results lean on an unvalidated drift-removal step and a single cell, so accept conditionally.","tokens_in":74703,"tokens_out":992,"would_cite":true,"duration_ms":14940,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Multisine EIS matches single-sine in steady state and reads impedance during battery operation.","keywords":["multisine excitation","electrochemical impedance spectroscopy","operando EIS","lithium-ion battery","Butler-Volmer kinetics","linearity stationarity verification","Kramers-Kronig","trend removal"],"falsifier":"Repeat the C/3 charge and discharge multisine bursts on the same cell at several C-rates and on at least one fresh cell, and compare the charge/discharge asymmetry at equal SOC and temperature with an independently parametrised Butler-Volmer prediction (from DC pulses or nonlinear EIS). If the asymmetry disappears when the trend-removal basis-function order is changed, or fails to follow the predicted C-rate scaling, the paper's separation of drift from dynamics and its kinetic-asymmetry claim would both collapse.","tokens_in":73630,"feed_emoji":"🔋","tokens_out":4895,"duration_ms":54914,"temperature":0.7,"pith_summary":"This paper argues that a carefully designed multisine excitation—a short sum of many sine waves applied at once—measures exactly the same steady-state impedance of a lithium-ion cell as the standard sequential single-sine method, while cutting measurement time from about 976 s to 200 s. Its more consequential claim is that multisine bursts can do something single-sines cannot: extract credible impedance while the battery is charging, discharging, relaxing, or changing temperature, because all frequencies see the same operating condition simultaneously. Using such operando measurements, the paper finds that charge-transfer resistance is lower during cycling than at rest, and that charge and discharge do not give the same impedance even at equal state of charge and temperature. The paper interprets this asymmetry as the signature of Butler-Volmer reaction kinetics linearised around a non-zero current—information that steady-state EIS cannot provide. If correct, this turns multisine EIS into a practical tool for parameterising battery models and quantifying reaction heat during real operation.","feed_headline":"Multisine impedance measures batteries while charging","feed_subtitle":"Broadband 20 mHz–1 kHz bursts catch charge/discharge kinetic asymmetry that steady-state EIS cannot see.","key_machinery":"The engine of the method is the odd random-phase multisine: a sum of sine waves at odd-integer multiples of a 20 mHz fundamental (40 excited frequencies spanning 20 mHz–1 kHz), with randomised phases that keep the crest factor near the single-sine value. Exciting only odd harmonics and leaving every seventh odd harmonic unexcited creates spectral 'traps' where even and odd nonlinear distortions land, so linearity and stationarity can be read directly from the current and voltage spectra. Impedance is the ratio of DFT voltage to DFT current at excited harmonics; several periods yield a variance estimate and confidence circles. For operando bursts, the prior trend-removal method fits basis fun","core_discovery":"The central discovery, on the paper's own terms, is that an odd random-phase multisine signal delivered through a modified commercial potentiostat yields the same impedance spectra as single-sine excitation when the cell is linear and stationary, and that this equivalence is what makes operando impedance measurement legitimate: because all excited frequencies overlap in time, each frequency is evaluated at the same drift, the same state of charge, and the same temperature. After removing slow trends and transients with the authors' basis-function method, the resulting impedance during C/3 charge and discharge satisfies Kramers-Kronig compliance where single-sine operando data fails, and reve","pith_inferences":["A testable corollary: repeating the same multisine bursts on multiple cells, C-rates, and repeated runs would establish whether the charge/discharge asymmetry is a general kinetic property or partly a single-cell artefact—the paper does not claim to settle cell-to-cell variation.","If the drift/trend-separation assumption holds, operando impedance at non-zero current could feed physics-based battery models with parameters valid under load, which would change how reaction-heat and overpotential estimates are derived from standard impedance data.","The trap-harmonic design generalises: choosing different harmonic-excitation patterns could separate nonlinearity order from drift, and might allow time-varying nonlinear models—not just linearised averages—to be identified from the same bursts."],"forward_implications":["Multisine bursts cut impedance characterisation from about 976 s to 200 s while giving the same steady-state result as single-sines at equal rms current.","Operando impedance during C/3 charging and discharging is measurably lower than steady-state impedance in the charge-transfer frequency range, and charge and discharge differ even at matched SOC and temperature.","The charge/discharge asymmetry is a linearised Butler-Volmer effect around non-zero current, so operando EIS can probe kinetic asymmetry that classical steady-state EIS cannot see.","Impedance measured in short bursts during relaxation and temperature ramps tracks the growth of the charge-transfer arc and the Arrhenius-like temperature dependence of exchange current, supporting use as a thermal and state diagnostic."],"fun_headline_variants":["Multisine EIS: same spectra, but measured while charging","Charge and measure: multisine impedance for live cells","Operando EIS done right: multisine matches single-sine","K-K compliant impedance during C/3 charge via multisine"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The operando conclusions rest on the assumption that the trend-removal method—which fits basis functions to the slow evolution of the spectra and is specified in a cited prior paper, not re-derived here—cleanly separates drift and transients from the true impedance without leaking into the reported spectra; Kramers-Kronig compliance alone cannot certify this because the paper itself notes that it is insensitive to nonstationarity.","fun_headline_variants_meta":{"raw":{"variants":["Multisine EIS: same spectra, but measured while charging","Charge and measure: multisine impedance for live cells","Operando EIS done right: multisine matches single-sine","K-K compliant impedance during C/3 charge via multisine"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000656,"raw_usage":{"total_tokens":2814,"prompt_tokens":690,"completion_tokens":2124,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":434,"completion_tokens_details":{"reasoning_tokens":2054}},"tokens_in":434,"tokens_out":2124,"duration_ms":17773,"temperature":1.0,"reasoning_tokens":2054,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T16:52:45.389613+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Repeat the C/3 charge and discharge multisine bursts on the same cell at several C-rates and on at least one fresh cell, and compare the charge/discharge asymmetry at equal SOC and temperature with an independently parametrised Butler-Volmer prediction (from DC pulses or nonlinear EIS). If the asymmetry disappears when the trend-removal basis-function order is changed, or fails to follow the predicted C-rate scaling, the paper's separation of drift from dynamics and its kinetic-asymmetry claim would both collapse.","supporting_citations":[],"review_version":1}