{"id":"b0e004e4-e4d7-484d-ae50-a7594851ada0","arxiv_id":"2607.08667","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":4,"one_line_summary":"Amorphous Zn-Sn-O is a viable IGZO alternative for transistor channels but shares IGZO's susceptibility to hydrogen-induced doping, as shown by statistically sampled DFT modeling of 1250+ structural models.","lead":"This paper uses large-scale density functional theory to evaluate amorphous Zn-Sn-O as a potential replacement for IGZO in transistors, finding it has promising electronic properties but similar vulnerability to hydrogen doping. A smart generalist would read it for its methodological framework showing how to properly model amorphous materials statistically rather than relying on single small supercells.","discovery_kind":"unclear","skeptic_critique":{"model":"glm-5.2","headline":"Hydrogen 'doping' claim rests on binding energies alone, with no charge-state or transition-level analysis to show H actually acts as a donor.","rationale":"The reader correctly identifies ISWO as an unvalidated metric, but this is a known limitation of the methodology that affects the quantitative mobility comparison, not the qualitative conclusion that Zn-Sn-O is a reasonable alternative. The more load-bearing concern is on the hydrogen side: the paper claims 'similar sensitivity to hydrogen induced doping' but only computes binding energies, never examining whether hydrogen actually introduces donor states near the conduction band. This is a gap between what was computed (thermodynamic stability of H binding) and what was claimed (doping sensitivity). The ISWO concern is real but secondary; even if ISWO is imperfect, the paper uses it comparatively between similar material classes and acknowledges its limitations. The hydrogen doping claim, by contrast, is missing a necessary physical ingredient (charge-state analysis) to support the stated conclusion. The verdict remains CONDITIONAL: the performance claim is adequately supported for a computational study, but the hydrogen doping claim needs charge transition level analysis to be credible. The paper's honest reporting of non-significant oxygen results and acknowledgment of ISWO limitations are points in its favor, but they do not address the missing charge-state analysis for hydrogen.","tokens_in":22478,"tokens_out":2059,"duration_ms":131118,"concrete_test":"Compute the charge transition level (e.g., (+/0)) for interstitial hydrogen in the amorphous ZnSnO3 models already generated, using the same PBEsol level of theory. Specifically, calculate the formation energy of H in different charge states as a function of Fermi-level position within the bandgap. If the (+/0) transition level lies above or near the CBM, hydrogen acts as a shallow donor and the doping claim is supported. If it lies deep in the gap, the claim about 'hydrogen induced doping' is unsupported despite stable binding.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The paper's central claim has two prongs: (1) Zn-Sn-O is a viable performance alternative to IGZO, and (2) it shares IGZO's sensitivity to hydrogen-induced doping. The second prong is the weakest link. The hydrogen analysis (Figs. 7-9) reports only binding energies and bond types (H-to-O vs H-to-metal). It shows that hydrogen binds stably to oxygen in amorphous ZnSnO3, with negative binding energies relative to H2 gas. From this, the paper concludes that 'a similar sensitivity to hydrogen induced doping may be present as in IGZO.' However, stable hydrogen binding to oxygen does not automatically imply n-type doping. Whether interstitial hydrogen acts as a shallow donor depends on its charge transition level relative to the conduction band minimum. In IGZO, hydrogen is known to act as a shallow donor because its (+/0) transition level lies near or above the CBM. The paper does not compute charge transition levels, defect formation energies as a function of Fermi level position, or the electronic structure of the H-doped system (e.g., whether a donor state appears near the CBM). Without this, the paper demonstrates similar hydrogen *binding* but not similar hydrogen *doping*. These are distinct physical phenomena: hydrogen could bind stably to oxygen forming a passive O-H bond that introduces a deep level in the gap, which would not cause the threshold voltage shifts that motivate the concern. The qualitative similarity in binding-site distributions between ZnSnO3 and IGZO (both show stable O-binding and metastable metal-binding) is insufficient to establish similar doping behavior. Additionally, the hydrogen study is restricted to only the 1:1 Zn:Sn ratio and extrapolated to the full system, which the authors acknowledge but do not validate.","agreement_with_reader":"partial"},"referee_report":{"model":"glm-5.2","summary":"This manuscript presents a first-principles DFT study of amorphous Zn-Sn-O oxide semiconductors, framed as both a methodological best-practices guide and a materials assessment. The authors generate ensembles of ~20 amorphous supercell models (~200 atoms each) for four Zn:Sn stoichiometric ratios, compute electronic gaps and an inverse state weighted overlap (ISWO) mobility proxy, and compare oxygen off-stoichiometry and interstitial hydrogen binding to the IGZO benchmark. The central conclusions are: (1) Zn-Sn-O is a promising performance alternative to IGZO due to favorable conduction-band character and compositional stability of the gap; (2) Zn-Sn-O exhibits similar sensitivity to hydrogen-induced doping as IGZO. The statistical framework—Shapiro-Wilk normality tests, t-tests with reported p-values, and honest reporting of non-significant results—is a genuine strength. However, the second conclusion overreaches the evidence presented, as detailed below.","tokens_in":23272,"tokens_out":1472,"duration_ms":252489,"significance":"The paper's methodological advocacy for statistical sampling of amorphous structural models is well-motivated and timely. The generation of over 1250 DFT calculations with systematic convergence checks, the transparent reporting of non-significant p-values for oxygen off-stoichiometry (Table 4), and the exhaustive hydrogen site screening (~1000 sites) are commendable. The ISWO metric provides a falsifiable, parameter-free proxy for mobility that avoids the well-identified pitfall of applying Bloch theorem concepts to amorphous phases. The identification of gap stability across compositions as a device-relevant advantage is a concrete, testable prediction.","major_comments":[{"comment":"The claim of 'similar sensitivity to hydrogen induced doping' (Abstract, Conclusions section) is not supported by the analysis presented. Figs. 7-9 report only binding energies and nearest-neighbor bond types for interstitial hydrogen. Stable hydrogen binding to oxygen does not automatically imply n-type doping; whether interstitial H acts as a shallow donor depends on its charge transition level relative to the conduction band minimum. The paper does not compute charge transition levels, defect formation energies as a function of Fermi level position, or the electronic structure of the H-doped system (e.g., whether a donor state appears near the CBM). Without at least one of these analyses, the paper demonstrates similar hydrogen *binding* but not similar hydrogen *doping*. The authors should either (a) compute the H (+/0) transition level or the defect formation energy vs. Fermi level,","section":null},{"comment":"The oxygen off-stoichiometry analysis (Fig. 5, Table 4) reports p-values of 0.36-0.86, indicating no statistically significant difference between stoichiometric and oxygen-reduced models for any composition. The text acknowledges this ('not a statistically significant trend'), yet the paper still draws physical conclusions from the median reaction energies (e.g., 'oxygen seems to be bound more weakly' with increasing Sn content, and 'all compositions are vulnerable to oxygen scavenging by H2 gas exposure'). The latter claim relies on comparing the computed binding energies to the H2O formation energy (2.58 eV), but since the distributions overlap substantially and the t-tests are non-significant, the vulnerability claim should be stated more cautiously, or additional sampling should be performed to achieve statistical significance. The tension between reporting non-significant p-values","section":null},{"comment":"The ISWO metric (Refs. 52, 90) is the sole basis for all mobility conclusions, including the central claim that Zn-Sn-O is a 'promising material class' from a performance standpoint. While the conceptual motivation for replacing effective mass with a state-overlap metric in amorphous systems is sound, the paper provides no validation that ISWO correlates with experimentally measured field-effect mobilities in amorphous oxides. The IGZO reference value (green line in Figs. 3-4) is itself computed with the same metric and code. Without at least a qualitative comparison to experimental mobility data for Zn-Sn-O (which exists, e.g., Refs. 56-57), the performance assessment remains internally circular. The authors should either provide such a comparison or explicitly qualify that the performance claim rests on the ISWO proxy being a valid relative mobility indicator.","section":null}],"minor_comments":[{"comment":"Fig. 8 caption and legend contain typos: 'Singel O bond' should be 'Single O bond', 'Doubel O bond' should be 'Double O bond', and the y-axis label reads 'Binding energ' (missing 'y'). Fig. 9 has the same 'Binding energ' typo.","section":null},{"comment":"The ISWO metric (Refs. 52, 90) is introduced without a compact mathematical definition in the main text. A brief equation or formal definition would improve accessibility for readers unfamiliar with the authors' prior work.","section":null},{"comment":"The extension of hydrogen binding results from the 1:1 Zn:Sn ratio to other ratios is acknowledged as an assumption ('we estimate that the defect state results can be extended to other Zn:Sn ratios'). Given that the authors themselves note Zn-rich ratios optimize field-effect mobility (Ref. 57), this limitation is material. The text should state this more prominently, e.g., in the Conclusions rather than only in the Results section.","section":null},{"comment":"Table 1 lists Zn2SnO2 with formula implying Zn:Sn = 2:1, but the text and figures refer to Zn2SnO4 for this ratio. The table entry 'Zn2SnO2' appears inconsistent with the stoichiometry discussed elsewhere (Zn2+ Sn4+ with 2:1:4 ratio). This should be corrected or clarified.","section":null},{"comment":"The paper states that 20 models per composition are used, but the supplementary information (Figs. S1-S4) shows radial distribution functions averaged over 'all models' without specifying the count. The figure captions should state the number of models averaged.","section":null},{"comment":"Section 'Substoichiometric oxygen content': the statement 'When calculating the distributions of only ten amorphous models per composition, no significantly different distributions... are obtained' is important for justifying the sample size but is mentioned only in passing. A brief table or figure comparing the 10-model vs. 20-model p-values would strengthen the best-practices argument.","section":null}],"recommendation":"major_revision","confidential_remarks":"The methodological framework is the paper's strongest asset and could be publishable. However, the hydrogen 'doping' claim is the weakest link and is load-bearing for the abstract's conclusion. The authors may be able to address this by softening the language to 'hydrogen binding' rather than 'doping,' or by computing at least one charge transition level. The ISWO circularity concern is less acute—it is a known limitation of using proprietary metrics—but should at least be acknowledged. I would encourage the editor to allow revision rather than rejection, as the statistical sampling methodology is a genuine contribution to best practices in this subfield."},"author_rebuttal":null,"desk_editor":{"model":"glm-5.2","letter":"Two things matter here. First, the statistical framework for amorphous oxide DFT is genuinely useful and well-executed. Second, the hydrogen doping claim is overstated — they show binding, not doping, and those are different physics. The stress-test concern on this point lands squarely. The paper's real contribution is methodological. They generate 20 amorphous models per composition (~200 atoms each) across four Zn-Sn ratios, apply Shapiro-Wilk normality tests, report t-test p-values honestly (including non-significant ones at 0.36–0.86 for oxygen off-stoichiometry), and argue convincingly that single-model studies of amorphous phases are unreliable. The critique of applying k-point band structures and effective mass extraction to amorphous materials — where Bloch's theorem doesn't hold — is correct and worth stating. The ISWO metric for mobility is reasonable as a delocalization proxy, though it lacks experimental validation and was introduced by these authors. The oxygen scavenging analysis is properly done: they generate fresh off-stoichiometric models rather than yanking atoms from relaxed structures, which avoids a real pitfall. The soft spot is the hydrogen section. They compute binding energies at ~1000 sites in amorphous ZnSnO3 and show stable H-to-O binding, qualitatively similar to IGZO. From this they conclude 'similar sensitivity to hydrogen induced doping may be present.' But stable binding to oxygen doesn't establish donor behavior. They don't compute charge transition levels, defect formation energies as a function of Fermi level, or the electronic structure of H-doped systems to show a donor state near the CBM. Hydrogen could bind stably and form a passive O-H bond with a deep level — no threshold voltage shift. The word 'doping' does real work in the conclusion and isn't backed by the computations shown. Also, the hydrogen study covers only the 1:1 Zn:Sn ratio and is extrapolated, which the authors acknowledge. The PBEsol bandgaps (1.8–2.0 eV vs. experimental 3.0–3.3 eV) are a known DFT limitation, not a fatal flaw, but the paper could state more clearly that quantitative gap values aren't the point. This is for computational materials scientists working on amorphous oxides and for experimentalists evaluating Zn-Sn-O as an IGZO replacement. The methodological framework deserves publication; the hydrogen doping claim needs either charge-state analysis or softer language. It deserves a serious referee who should push back specifically on the doping language.","headline":"Solid statistical DFT framework for amorphous Zn-Sn-O, but the hydrogen 'doping' claim overreaches the binding-energy evidence","tokens_in":23603,"tokens_out":608,"would_cite":false,"duration_ms":180648,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["61.43.Dq","71.23.Cq","73.61.Le"],"model":"glm-5.2","headline":"Statistical DFT shows Zn-Sn-O matches IGZO performance but shares its hydrogen flaw","keywords":[],"falsifier":"If experimental field-effect mobility measurements on amorphous Zn-Sn-O films with varying zinc-to-tin ratios show strong composition-dependent variation, the claim that the electronic properties are composition-insensitive would be undermined. If hydrogen-exposure experiments on Zn-Sn-O transistors show no threshold-voltage shift, the claim of shared hydrogen sensitivity with IGZO would be falsified.","tokens_in":22783,"feed_emoji":"🔬","tokens_out":891,"duration_ms":308985,"temperature":0.7,"pith_summary":"This paper argues that first-principles computer modeling of amorphous oxide semiconductors is only trustworthy when researchers generate enough structural models to build statistical distributions of the properties they care about, rather than reporting results from a single small simulated cell. Using the Zn-Sn-O material system as a test case, the authors generate over 1,250 atomic-scale models and show that its electronic gap and charge-carrier mobility are largely insensitive to the zinc-to-tin ratio, a stability advantage over the industry-standard amorphous oxide IGZO. However, the same statistical approach reveals that Zn-Sn-O binds oxygen weakly and binds interstitial hydrogen strongly to oxygen sites, meaning it is likely just as vulnerable as IGZO to the hydrogen-induced doping shifts that plague transistor reliability. The central methodological claim is that three common shortcuts in the field—using too few structural models, applying crystal-only concepts like band-structure effective mass to amorphous materials, and creating defects by yanking atoms out of unrelaxed structures—produce results that can be misleading or simply wrong.","feed_headline":"Statistical DFT shows Zn-Sn-O matches IGZO performance but shares its hydrogen flaw","feed_subtitle":"Over 1,250 atomic models reveal the indium-free oxide is stable across compositions but just as vulnerable to hydrogen doping as the thing它要","key_machinery":"The Inverse State Weighted Overlap (ISWO), a dimensionless metric that measures how spatially connected and energetically close the electronic states near the conduction-band edge are, serving as a proxy for electron mobility in materials where the band-structure concept of effective mass is invalid because translational symmetry is absent. The decorate-and-relax method for generating amorphous structural models, combined with ensemble sizes of ~20 models per composition and ~200 atoms per model, provides the statistical power needed to distinguish real trends from noise.","core_discovery":"When amorphous Zn-Sn-O is modeled with proper statistical sampling across many independently generated structural models, its electronic gap and a mobility-proxy metric called the Inverse State Weighted Overlap (ISWO) are remarkably stable across different zinc-to-tin ratios, suggesting compositional variation in real devices would not degrade performance. But the same computations show that oxygen is bound more weakly than in IGZO and that interstitial hydrogen binds stably to oxygen atoms throughout the amorphous network, indicating that the hydrogen-doping instability observed in IGZO transistors would likely persist in Zn-Sn-O.","pith_inferences":[],"forward_implications":["Researchers publishing DFT results on amorphous oxides should be expected to report distributions over multiple structural models, not single-cell values, and to test whether those distributions are statistically significant.","The ISWO metric could be applied to other candidate amorphous oxide semiconductors to rank their likely electron mobility without requiring experimentally validated effective masses.","The finding that oxygen binding in Zn-Sn-O is weaker than in IGZO suggests that process integration steps involving forming-gas anneals or oxygen-poor deposition conditions may require the same engineering mitigations developed for IGZO.","The compositional stability of the bandgap in Zn-Sn-O could relax manufacturing tolerances for the metal ratio in deposited films, potentially lowering production cost."],"fun_headline_variants":["Statistical DFT: Zn-Sn-O gaps stay stable across ratios, but hydrogen flaw persists","1,250 models show Zn-Sn-O matches IGZO mobility but shares hydrogen doping risk","Proper sampling finds Zn-Sn-O compositionally robust but hydrogen-vulnerable","Zn-Sn-O mobility holds across compositions but hydrogen binds like in IGZO","Statistical framework ties Zn-Sn-O hydrogen instability to weak oxygen bonds"],"cache_read_input_tokens":0,"weakest_assumption_plain":"The ISWO metric, introduced by the same authors in prior work, is assumed to be a valid proxy for real electron mobility in amorphous oxides, but no experimental comparison between ISWO values and measured field-effect mobilities is provided to confirm this connection.","fun_headline_variants_meta":{"raw":{"variants":["Statistical DFT: Zn-Sn-O gaps stay stable across ratios, but hydrogen flaw persists","1,250 models show Zn-Sn-O matches IGZO mobility but shares hydrogen doping risk","Proper sampling finds Zn-Sn-O compositionally robust but hydrogen-vulnerable","Zn-Sn-O mobility holds across compositions but hydrogen binds like in IGZO","Statistical framework ties Zn-Sn-O hydrogen instability to weak oxygen bonds","Amorphous Zn-Sn-O: stable gaps, shared hydrogen flaw with IGZO"]},"model":"glm-5.2","effort":"low","cost_usd":0.0,"raw_usage":{"total_tokens":1885,"prompt_tokens":521,"completion_tokens":1364,"prompt_tokens_details":null},"tokens_in":521,"tokens_out":1364,"duration_ms":88314,"temperature":1.0,"reasoning_tokens":1201,"cache_read_input_tokens":0,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-10T03:19:34.886685+00:00","model_set":{"reader":"glm-5.2"},"falsifier":"If experimental field-effect mobility measurements on amorphous Zn-Sn-O films with varying zinc-to-tin ratios show strong composition-dependent variation, the claim that the electronic properties are composition-insensitive would be undermined. If hydrogen-exposure experiments on Zn-Sn-O transistors show no threshold-voltage shift, the claim of shared hydrogen sensitivity with IGZO would be falsified.","supporting_citations":[],"review_version":1}