{"id":"15b902a9-287e-4b51-8fb9-26b730c483f5","arxiv_id":"2505.14194","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A new open-source Python toolbox for HAXPES analysis offers grouped curve fitting with cross-spectrum parameter constraints and a signal-ratio model for overlayer thickness.","lead":"POLEVAL is a Python package for analyzing hard X-ray photoelectron spectroscopy (HAXPES) data, combining data handling, constrained group fitting, and adsorbate layer thickness estimation. It aims to make synchrotron and lab XPS analysis faster, more reproducible, and self-documenting by working inside Jupyter notebooks.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central claim that grouped fitting with equality constraints yields 'more confident fit results' is not validated: if constrained parameters vary across spectra, the fit becomes biased while uncertainties shrink, and no benchmark is provided.","rationale":"The reader's primary weakest assumption is exactly the grouped-fit commonality premise, and I agree that it is the most load-bearing concern. The paper is a software description, so the central claim is an empirical capability claim: grouped fitting improves confidence for noisy spectra. That claim rests on a bias-variance trade-off, and without a synthetic-data benchmark the reader cannot know whether the constraint reduces total error or merely hides bias. The thickness model's limitations are stated in §3, so they are less concerning for the central claim, though they still limit quantitative accuracy. I found no algebraic error in Equations 3-7; they are standard attenuation formulas and appear internally consistent. Because the missing validation is straightforward to add in a revision, the appropriate disposition remains conditional rather than acceptance or rejection.","tokens_in":3791,"tokens_out":6470,"duration_ms":65882,"concrete_test":"Use the package itself to run a Monte Carlo validation: generate, for example, 100 synthetic groups of 10 spectra with known Voigt line shapes and a realistic spread (say 5% relative standard deviation) in one constrained parameter, such as peak width or position. Fit each group twice, once with the grouped equality constraint and once spectrum-by-spectrum, and compare the recovered parameters to ground truth in terms of bias, standard deviation, and root-mean-square error. If the constrained RMSE exceeds the unconstrained RMSE for that realistic spread, the 'more confident' claim must be qualified; if it is lower, the claim is supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The core advantage claimed for POLEVAL over per-spectrum fitting is the grouped fitter with dependency constraints (Abstract; Section 2). The load-bearing premise is that shared parameters such as peak width, position, or intensity ratios are physically common across all spectra in the group. No simulation, synthetic-data benchmark, or statistical analysis is provided to test this premise. If a constrained parameter varies with temperature, pressure, time, or sample condition, forcing equality introduces systematic bias even as the reported parameter uncertainties decrease; the fit then looks more confident but is less accurate. The thickness model's flat-homogeneous assumption is explicitly acknowledged in §3, but the grouped-fit commonality assumption is presented without qualification and is the paper's main novelty. The claim that no published software provides this functionality may also warrant a broader literature check, though the absence of validation is the primary risk.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"POLEVAL is presented as a Python toolbox for HAXPES/XPS analysis, organized around Jupyter notebooks and custom data classes (Region, Sequence, SequenceGroup) to support persistent, collaborative, and reproducible workflows. The paper describes three quantitative analysis functions: cross-section normalization using haxquantpy, a grouped fitting routine built on lmfit that enforces shared peak parameters across spectra, and an adsorbate-overlayer thickness model based on signal ratios (Eqs. 1-7). The central claimed advantage is that grouped fitting with equality constraints yields \"more confident fit results\" for noisy data, and the authors further assert that no published software provides this grouping functionality. The paper contains no experimental data, synthetic benchmarks, or statistical validation of these claims; the thickness model equations appear standard and correct, and the flat-homogeneous geometry assumption is explicitly acknowledged in §3.","tokens_in":3992,"tokens_out":3715,"duration_ms":38348,"significance":"If the central claims are validated, POLEVAL would be a useful contribution to the HAXPES community: it is open source under an MIT license, integrates literature photoionization parameters via haxquantpy, and offers a notebook-based workflow that could improve reproducibility in synchrotron-based experiments. The overlayer-thickness formalism in Eqs. (1)-(7) is a standard textbook result and is presented clearly. The paper's main novelty, the grouped fitting routine with dependency constraints, is conceptually appealing, and the free availability of the code is a concrete strength. However, the paper currently does not demonstrate the key advantage with any example: there is no fit output, no comparison against unconstrained fitting, no uncertainty calibration, and no discussion of when the equality constraints are physically valid. The literature claim of novelty is also under-supported. These gaps make the present significance claims speculative rather than established.","major_comments":[{"comment":"The central claim that grouped fitting with equality constraints yields \"more confident fit results\" is not validated anywhere in the manuscript. No fit results, synthetic-data experiments, or statistical analyses are provided. Please add a benchmark with known ground-truth parameters, compare constrained versus unconstrained fits in terms of both bias and reported uncertainty, and show that the uncertainties are calibrated. Without such evidence, the main novelty of the package is only asserted, not demonstrated.","section":"Abstract and §2 (Implementation)"},{"comment":"The premise that constrained parameters such as peak widths, positions, or ratios are common across all spectra in a group is stated without qualification. If any constrained parameter varies with temperature, pressure, time, or sample condition, forcing equality will introduce systematic bias even while reported uncertainties shrink. Please state the physical conditions under which the constraint is valid, and provide a diagnostic or a sensitivity analysis showing the effect of violating the commonality assumption on fitted parameters and their uncertainties.","section":"§2 (Implementation)"},{"comment":"The claim that \"no published software so far provides this functionality\" is too strong given the limited survey in the paper (only two open-source tools, LG4X and PyARPES, and one commercial package are cited). Many XPS analysis packages and general peak-fitting libraries support parameter linking or shared constraints; a broader literature and software search should be reported, or the claim should be narrowed to \"no published HAXPES-specific software that we are aware of.\" Otherwise, the novelty statement is not substantiated.","section":"§1 (Statement of Need)"},{"comment":"The reproducibility argument would be much stronger if the manuscript referenced a runnable example notebook, a test suite, or a documented API. The current text only points to a GitLab repository. Please add a link to documentation and include a minimal worked example (e.g., a synthetic or small experimental dataset fitted with the grouped routine) either in the paper or as supplementary material, so that readers and reviewers can verify the claimed behavior.","section":"§4 (Availability)"}],"minor_comments":[{"comment":"The phrase \"common open-source source algorithms\" contains a typo; it should likely read \"open-source algorithms\" or \"open-source analysis algorithms.\"","section":"§1"},{"comment":"The caption reads \"showing the key classes and and their most important properties\"; the duplicate \"and\" should be removed.","section":"Figure 1 caption"},{"comment":"The sentence preceding Table 1 uses \"absorbate\" instead of \"adsorbate\" in the description of the subscripts.","section":"§3, Table 1"},{"comment":"The name \"POLEV AL\" appears with an extra space in several places (e.g., the title, abstract, and running text); this appears to be a formatting artifact and should be corrected to \"POLEVAL.\"","section":"Throughout"},{"comment":"The sentence beginning \"The rationale for this is that the differences in X-ray intensity...\" is unclear: please rephrase to specify what is negligible and why, since the scale argument as written is ambiguous.","section":"§3"},{"comment":"Reference [1] contains an extra space in the author name \"V . Yarzhemsky\" and in the volume/page formatting; please ensure consistent citation style.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The manuscript appears to be written in a software-journal format, and the software itself may be a valuable contribution. My recommendation assumes that the authors can add a validation section with synthetic and/or real examples, a sensitivity analysis for the grouped-fit constraint, and a more careful literature survey within a revision cycle. I did not run the software; the assessment is based solely on the manuscript text."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"POLEVAL is a serious software contribution: an MIT-licensed Python package that wraps lmfit and haxquantpy into a notebook-oriented workflow for HAXPES/XPS, with object layers (Region, Sequence, SequenceGroup) that genuinely help handle interleaved beamtime data. The grouped fitting approach — tying peak widths, positions, or ratios across spectra via dependency constraints — is the real novelty, and it is a good idea for noisy data. The thickness formulas (Eqs. 1–7) are standard slab-model results; the authors are upfront that roughness and standing-wave effects are ignored.\n\nWhat's missing is any validation. There is no test suite, no synthetic benchmark, no example dataset shipped with the paper, and no quantification of the claim that grouped fitting gives 'more confident fit results.' That matters because the stress-test point is exactly right: if a constrained parameter varies with temperature, pressure, or time, forcing equality introduces systematic bias while the reported uncertainty shrinks. The paper does not acknowledge this asymmetry, and because the grouping is the headline feature, an empty-handed claim leaves the central advantage unproven. A simple Monte Carlo study or one real case with ground truth would have closed most of the gap.\n\nThe novelty statement ('no published software so far provides this functionality') is also a bit broad without a literature survey — LG4X and PyARPES are cited, but a quick look suggests some commercial packages allow parameter linking, though perhaps not across spectra in this way. This is a minor overclaim, fixable by softening the language.\n\nOn the other hand, the code appears clean enough, the integration with haxquantpy is sensible, and the notebook-first design is genuinely aligned with synchrotron practice. The paper is not pretending to derive new physics; it's a software announcement, and as such it is clearly written and honest about the model's limits.\n\nWho should read it: anyone doing HAXPES analysis or building open tools for spectroscopy. It deserves a serious referee, and a JOSS-style review would be appropriate. My recommendation: engage with it, but require the authors to add a benchmark or example analysis demonstrating that grouped fitting recovers known parameters at least as accurately as per-spectrum fitting, and to discuss the bias-risk of equality constraints.","headline":"A useful, openly licensed HAXPES analysis package with a clever grouped-fitting feature; the paper underdelivers on validation of that feature, but it deserves referee time.","tokens_in":4453,"tokens_out":2690,"would_cite":false,"duration_ms":27631,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Grouped fitting makes noisy HAXPES fits more confident","keywords":["HAXPES","X-ray photoelectron spectroscopy","peak fitting","shared constraints","adsorbate thickness","reproducible research","Python","data analysis"],"falsifier":"Simulate a series of HAXPES spectra from a known ground truth with controlled noise and a peak width that drifts linearly across the series, then fit with the grouped constraint that widths are equal and compare with unconstrained fits. If the constrained fits yield confidence intervals that exclude the true drifting width, or if the unconstrained fits recover the drift with no larger uncertainty, the central claim that grouping gives more confident and more consistent fits is falsified.","tokens_in":3627,"feed_emoji":"📈","tokens_out":8390,"duration_ms":75849,"temperature":0.7,"pith_summary":"POLEVAL is a freely available Python package for analysing hard X-ray photoelectron spectroscopy (HAXPES) data from multi-sample or time-resolved campaigns. The paper's central claim is that fitting all spectra of a group together, with shared peak parameters enforced as constraints, gives more consistent and more confident fits than fitting each noisy spectrum independently. The package also provides a signal-ratio formalism that estimates adsorbate overlayer thickness from the relative intensities of substrate and adsorbate core levels, and it integrates literature photoionization cross-sections so intensities from different core levels can be compared quantitatively. The authors argue that no other published XPS software offers the grouped fitting capability, making POLEVAL a new option for reproducible, notebook-based beamtime analysis.","feed_headline":"Grouped fitting makes noisy HAXPES fits more confident","feed_subtitle":"POLEVAL, a free notebook-based Python toolbox, constrains peak widths and positions across entire measurement series.","key_machinery":"The data model is built from Region, Sequence, SequenceGroup, and Config objects, which organise spectra by core level, by loaded file, and by experimental group so that trends in time, temperature, or pressure can be explored in a few lines of code. The load-bearing fitting mechanism is a grouped peak fit that operates on all Regions of the same core level in a SequenceGroup at once, using dependency constraints between parameters to encode prior knowledge; the paper states this is the feature that no published XPS software provides. For overlayer thickness, the key identity is the effective probing depth $1/\\lambda' = 1/\\lambda_p + 1/\\lambda_e$, which enters the signal ratio $S_a/S_b$ (Eq. 5); the package solves this ratio for d with a numerical root finder and supplies a closed-form approximation when the two effective probe depths in the adsorbate are equal.","core_discovery":"On the paper's own terms, the contribution is a working software implementation of grouped HAXPES analysis. The novel step is the fitting routine: a single fit operates on every Region of the same core level within a SequenceGroup, and the user can impose dependencies such as equal peak widths, fixed positions, or fixed ratios between peaks across all spectra in the set. This transfers prior physical knowledge into the fit, restricting the information that must be extracted from noisy data and, the authors contend, reducing the uncertainty of the results. The second quantitative contribution is the overlayer thickness model (Eqs. 5-7), which estimates the adsorbate thickness d by equating the measured signal ratio $S_a/S_b$ to an expression involving effective probing depths and material densities, solving numerically for d and using the simplified closed form (Eq. 6) as an initial guess.","pith_inferences":["If grouped fitting reduces fit uncertainty as claimed, the same constraint-based strategy should apply to other spectroscopy series where parameters are known to drift smoothly, such as temperature-dependent X-ray absorption or resonant photoemission; the paper does not discuss these settings.","The benefit of grouping should grow with the number of spectra and the amount of shared structure; a quantitative benchmark against synthetic data would let users know how many spectra justify the constraints, which the paper does not provide.","The thickness model's neglect of roughness and standing-wave effects, which the authors acknowledge, means its estimates are only as good as the flat, semi-infinite geometry assumption; extending the ratio formalism to account for a rough interface would be a natural next step that the paper leaves undone.","The claim that no published software provides grouped fitting is a strong comparative statement; it could be tested only by a wider literature survey, and if a similar feature exists elsewhere, the novelty claim would narrow."],"forward_implications":["Beamtime users can treat a full measurement campaign as one fit, so a noisy single spectrum does not have to be fitted alone; shared parameters across the series stabilise the result.","Analysis reports become reproducible because the notebook records every data treatment step, intermediate spectrum, and fit constraint, and can be handed off between researchers mid-campaign.","Quantitative comparisons between different core levels are possible with cross-section normalization from literature values, without leaving the analysis environment.","Adsorbate thicknesses on the nanometre scale can be estimated directly from signal ratios, including for a composite substrate measured at two core levels (Eq. 7)."],"supporting_citations":[{"why":"Supplies the Dirac-Fock photoionization cross-section parameters used for quantitative comparisons between core levels.","marker":"[1]"},{"why":"Existing open-source XPS software that the paper positions POLEVAL against.","marker":"[2]"},{"why":"An analysis framework for photoemission spectroscopies that the paper distinguishes from the HAXPES-focused tool.","marker":"[3]"},{"why":"A practical guide for curve fitting in XPS, cited to justify why consistent peak shape parameters across a series are essential.","marker":"[4]"},{"why":"The earlier codebase that POLEVAL uses as its starting point.","marker":"[5]"}],"fun_headline_variants":[],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The grouped fitting routine assumes that the parameters it constrains to be equal are physically common across all spectra in the group; if a peak width, position, or ratio actually drifts with temperature, pressure, or time, forcing equality biases every fit, and the paper provides no simulation or statistical test of this assumption.","fun_headline_variants_meta":{"error":"Client error '402 Payment Required' for url 'https://api.deepseek.com/chat/completions'\nFor more information check: https://developer.mozilla.org/en-US/docs/Web/HTTP/Status/402"},"cache_creation_input_tokens":0},"created_at":"2026-08-07T15:37:55.106588+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Simulate a series of HAXPES spectra from a known ground truth with controlled noise and a peak width that drifts linearly across the series, then fit with the grouped constraint that widths are equal and compare with unconstrained fits. If the constrained fits yield confidence intervals that exclude the true drifting width, or if the unconstrained fits recover the drift with no larger uncertainty, the central claim that grouping gives more confident and more consistent fits is falsified.","supporting_citations":[{"cited_title":"Dirac–Fock photoionization parameters for HAXPES applications,","cited_arxiv_id":null,"evidence_quote":"Supplies the Dirac-Fock photoionization cross-section parameters used for quantitative comparisons between core levels."},{"cited_title":"Nakajima","cited_arxiv_id":null,"evidence_quote":"Existing open-source XPS software that the paper positions POLEVAL against."},{"cited_title":"PyARPES: An analysis framework for multimodal angle-resolved photoemission spectroscopies,","cited_arxiv_id":null,"evidence_quote":"An analysis framework for photoemission spectroscopies that the paper distinguishes from the HAXPES-focused tool."},{"cited_title":"Practical guide for curve fitting in x-ray photoelectron spectroscopy,","cited_arxiv_id":null,"evidence_quote":"A practical guide for curve fitting in XPS, cited to justify why consistent peak shape parameters across a series are essential."},{"cited_title":"Shipilin","cited_arxiv_id":null,"evidence_quote":"The earlier codebase that POLEVAL uses as its starting point."}],"review_version":1}