{"id":"72dc365d-9c96-47b1-9b7c-21b6aff5ab1a","arxiv_id":"2506.19205","paper_version":1,"verdict":"UNVERDICTED","confidence":"UNKNOWN","novelty_score":5.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":0,"one_line_summary":"Etching ITO, dip-coating for 12 hours, and adding the bis-phosphonic acid 6dPA each increase phosphonic acid coverage, and together they reduce nonradiative recombination and raise efficiency in this perovskite solar cell study.","lead":"Researchers systematically varied how phosphonic acid coatings are applied to ITO contacts in perovskite solar cells, comparing spin coating versus 12-hour dip coating, acid etching versus no etching, and adding a second spacer molecule. All three changes increased coating coverage, and the combined treatment gave the longest carrier lifetimes and the highest device efficiencies in their wide-bandgap perovskite cells.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Coverage-to-lifetime claim is confounded by uncharacterized perovskite bulk quality; the authors themselves note devices are 'limited more by film quality than by surface recombination dynamics,' yet no XRD/SEM/AFM of the perovskite is reported.","rationale":"The reader's weakest assumption—that perovskite bulk quality is comparable across modified contacts—is the same as the most load-bearing concern I find, so I mark agreement as agree and leave the reader's UNVERDICTED verdict unchanged. The central causal claim is 'coverage controls recombination,' and the only evidence tying coverage to recombination is the correlation between pre-perovskite XPS/UPS metrics on bare ITO and post-deposition trPL/QFLS on full half-stacks. Because the surface treatments also alter the template on which perovskite nucleates, bulk-quality differences are a natural alternative explanation. The authors themselves acknowledge this twice: once in the trPL section ('possibly because the perovskite grown from those surfaces itself has fewer defects') and once in the device section ('our devices are limited more by film quality than by surface recombination dynamics'). My proposed SEM/AFM/XRD check directly tests whether the alternative explanation is real. I do not treat this as grounds for rejection: the manuscript is internally consistent, reports multiple controls (6dPA-only half-stacks, Me-4PACz/MeO-2PACz comparisons, PbI2 adsorption experiments), and candidly states that the best devices are below state of the art. The issue is a missing control, not a demonstrated error. If the proposed structural characterization were added and showed flat morphology and crystallinity, the coverage-to-interfacial-recombination claim would be substantially strengthened.","tokens_in":18312,"tokens_out":6510,"duration_ms":77936,"concrete_test":"Deposit Cs17Br25 perovskite in a single run on the four key contacts—bare etched ITO, spin-coated I-2PACz, dip-coated I-2PACz, and dip-coated I-2PACz/6dPA—and characterize the perovskite films by top-view SEM grain size, AFM roughness, and XRD peak width/intensity (e.g., the 100/220 reflections). If median grain size or XRD crystallinity increases monotonically with phosphonic-acid coverage, the coverage-lifetime correlation is confounded by bulk quality; if morphology and crystallinity are statistically indistinguishable across the four contacts, the interfacial-recombination interpretation is supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central claim is causal: deposition protocol raises phosphonic-acid coverage on ITO, and this higher coverage lowers nonradiative recombination at the ITO/perovskite interface, improving trPL lifetimes, QFLS, and PCE. The load-bearing step is that the coverage differences measured by XPS/UPS on bare modified ITO are the cause of the recombination differences in the perovskite films grown on top. That step is not protected against a bulk-quality confound. The authors flag it explicitly: 'We postulate that this type of interface modification minimizes direct contact between the perovskite and regions of the ITO responsible for faster surface recombination... but also possibly because the perovskite grown from those surfaces itself has fewer defects.' Later they state that their devices 'are limited more by film quality than by surface recombination dynamics.' Etching, 12-hour dip coating, and the 6dPA second layer all change surface chemistry, wettability, and roughness; any of these can alter nucleation, grain size, crystallinity, and bulk defect density of the perovskite. The manuscript reports no XRD, SEM, or AFM of the perovskite films, so the 55x lifetime increase (3.8 ns to 210 ns) could reflect systematically better bulk perovskite rather than interfacial passivation. The same confound weakens the device argument: PCE gains are dominated by JSC and FF, while VOC barely improves despite large QFLS increases, consistent with the authors' own film-quality interpretation.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript studies how the deposition protocol for phosphonic acid interface modifiers on ITO affects the coverage of the modifier and the performance of wide-bandgap (1.7 eV) perovskite optoelectronic devices. The authors vary three parameters: ITO surface etching (HCl/FeCl3 vs. oxygen plasma), deposition method (spin coating vs. 12-hour dip coating), and use of a secondary modifier (1,6-hexylenediphosphonic acid, 6dPA). They characterize the modified ITO surfaces with UPS and XPS, and half-stack ITO/modifier/perovskite films with PLQY-derived quasi-Fermi level splitting (QFLS) and time-resolved photoluminescence (trPL), then correlate these with full device J-V performance. They find that etching, prolonged dip coating, and the dual I-2PACz/6dPA layer each increase the XPS-based coverage, work-function shift, trPL lifetime, QFLS, and device efficiency, with the best combination achieving a top PCE of 14.35% and a 57% improvement over the control. The central claim is that higher phosphonic acid coverage reduces nonradiative recombination at the buried ITO/perovskite interface, leading to improved lifetimes, QFLS, and device performance.","tokens_in":18604,"tokens_out":3094,"duration_ms":33606,"significance":"If the causal interpretation holds, the paper makes a valuable practical contribution: it identifies concrete, easily reproduced processing choices that systematically improve SAM coverage on ITO and, correspondingly, perovskite film quality metrics and device performance. The study is unusually thorough in correlating multiple independent measurements (XPS, UPS, trPL, QFLS, J-V) across a matrix of deposition conditions, and it includes important controls such as 6dPA-only layers and a demonstration with the more common Me-4PACz modifier. The authors also explicitly acknowledge limitations, including uncertainty about perovskite bulk quality and the observation that their devices are limited by film quality rather than surface recombination dynamics. That honesty is a strength, but it also exposes a gap between the strength of the empirical correlations and the strength of the causal claim. The paper does not provide machine-checked proofs, but as an experimental correlation study it is reproducible in design.","major_comments":[{"comment":"The central claim that increased phosphonic acid coverage causes reduced nonradiative recombination at the ITO/perovskite interface is confounded by uncharacterized variations in perovskite bulk quality. The authors themselves flag this: 'We postulate that this type of interface modification minimizes direct contact between the perovskite and regions of the ITO responsible for faster surface recombination... but also possibly because the perovskite grown from those surfaces itself has fewer defects.' Later, in the device section, they state that 'our devices are limited more by film quality than by surface recombination dynamics.' The manuscript reports no XRD, SEM, or AFM of the perovskite films, so a 55x increase in trPL lifetime (3.8 ns to 210 ns) and the large QFLS gains could in principle arise from systematically better bulk perovskite crystallization on etched, dip-coated, dual-modifier surfaces rather than from interfacial passivation. This is a load-bearing ambiguity for the paper's central message. I recommend adding structural characterization of the perovskite films (e.g., XRD crystallinity, SEM/AFM morphology, grain size) and, if possible, a thickness-dependent trPL or a controlled experiment that isolates interfacial recombination.","section":"Results and Discussion, 'Studies of surface modification protocols' (Figure 2 and ensuing text)"},{"comment":"The device data are internally inconsistent with the interfacial-passivation narrative. Although QFLS improves by ~7.8% and trPL lifetime by over an order of magnitude, the device VOC improves only slightly, which the authors attribute to film-quality limitations. Yet the PCE gains are dominated by JSC (+15-18%) and FF (+31% with etching). These parameters can be strongly affected by perovskite nucleation, coverage, and bulk transport properties that change with the wetting and chemistry of the underlying substrate. The conclusion that 'improved phosphonic acid coverage in turn leads to reduced nonradiative decay... and consequently to improved photovoltaic device performance' is therefore not uniquely supported. The manuscript should either present additional evidence separating interface from bulk contributions, or soften the causal language to describe correlations and explicitly discuss the alternative that the modified substrates improve perovskite film formation.","section":"Results and Discussion, 'Characterization of photovoltaic responses' (Figure 4c-4f)"},{"comment":"The XPS-based coverage metric conflates surface coverage with multilayer thickness. The authors report that prolonged dip coating and 6dPA addition increase P/In, O/In, and effective thickness, but they also present ARXPS and LKE-edge evidence for disordered layers, patchy coatings, and 6dPA multilayer formation ('clear evidence for multilayered regions', 'it is likely that some areas of multilayer coverage exist'). If the thicker layers are partially multilayered, then the correlation between 'coverage' and improved lifetimes may depend on factors other than the fraction of ITO surface passivated, such as the density of exposed phosphonic-acid head groups or the surface energy presented to the perovskite precursor. The manuscript should separate the concept of monolayer surface coverage from total adsorbed mass/thickness, and ideally use a direct coverage-sensitive probe (e.g., a calibrated monolayer or a surface-selective technique) to support the statement that the best devices have the highest actual passivation of the ITO surface.","section":"Results and Discussion, 'Characterization of clean, etched and phosphonic acid coverage of ITO using UPS and XPS'…"}],"minor_comments":[{"comment":"The text reads 'higher shunt and lower series resistance' but contains a typographical error: 'sunt' for 'shunt.'","section":"Characterization of photovoltaic responses, paragraph on FF"},{"comment":"Atomic force microscopy is mentioned as a funding source ('Atomic force microscopy and wide-bandgap perovskite semiconductor growth was supported by...'), but no AFM data are presented in the paper or Supporting Information. Either provide the AFM characterization or remove this reference to avoid implying measurements that are not reported.","section":"Acknowledgments"},{"comment":"The UPS work-function table reports two values for the pure 6dPA samples because of LKE-edge variability, and the text discusses patchy coatings. A representative set of LKE-edge spectra displayed consistently (rather than in the Supporting Information) would help the reader evaluate the magnitude of the spatial heterogeneity that the authors emphasize throughout.","section":"Table 1 and Figure S2-S8"},{"comment":"Reference 8 is missing an author list and full title, appearing simply as 'Electron-Hole Diffusion Lengths Exceeding 1 Micrometer in an Organometal Trihalide Perovskite Absorber.' Please reformat for completeness.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The paper is a solid empirical study with careful cross-correlation of surface and device measurements, but the central causal claim is currently underdetermined because the perovskite bulk quality is not characterized. The authors' own statements ('possibly because the perovskite grown from those surfaces itself has fewer defects'; 'limited more by film quality than by surface recombination dynamics') indicate that they are aware of the confound, yet the abstract and conclusion present the coverage-lifetime-performance link as established. Adding perovskite film structure/morphology data or a more direct interface-sensitive experiment would resolve this. The manuscript is probably within scope for the journal, though the practical device efficiencies are below state of the art; the contribution is in the mechanistic processing-structure-property insight."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a useful empirical study that systematically varies three deposition parameters for phosphonic acid SAMs on ITO - spin vs prolonged dip coating, etching vs not, and single I-2PACz vs dual I-2PACz/6dPA - and correlates the results with XPS/UPS coverage, trPL lifetime, QFLS, and device performance. The individual ingredients have all been reported before, but the factorial comparison on one platform with directly connected coverage and performance data is a legitimate new contribution. The work also does a good job of showing that work function alone does not explain the trends, and the angle-resolved XPS evidence for disordered multilayers rather than clean SAMs is a useful corrective to loose 'self-assembled monolayer' language in the field. The PbI2 uptake experiments are preliminary but a nice probe of residual adsorption sites, and the transfer to MeO-2PACz/Me-4PACz shows the practical reproducibility point is general. The main soft spot is the bulk-quality confound, and the authors flag it themselves. They write that the perovskite grown from those surfaces 'itself has fewer defects,' and later that devices are 'limited more by film quality than by surface recombination dynamics.' That is close to conceding that the coverage-lifetime correlation may be mediated by changes in crystallization and bulk defect density, not purely by interfacial passivation. Without XRD, SEM, or AFM of the perovskite films, the 55x lifetime increase (3.8 ns to 210 ns) cannot be cleanly assigned to reduced interfacial recombination. The device data support the authors' own film-quality reading: VOC barely moves while JSC and FF dominate the PCE gains. This does not kill the paper - the correlation stands and the practical guidance is useful - but it does mean the mechanistic headline should be read as a hypothesis, not an established fact. The error bars are large in places, but the trends are consistent across multiple independent measurements, and the limitations are stated honestly. Who gets value from this: anyone working on SAM-based perovskite devices, especially labs chasing reproducibility or trying to understand why their I-2PACz results vary. It deserves serious refereeing, not desk rejection, and I would cite it for the deposition-coverage relationship even while treating the recombination mechanism as open. My advice: send it to review, and ask the authors to add structural characterization of the perovskite or soften the causal language in the abstract and conclusion.","headline":"A solid, honest methods paper whose coverage-lifetime correlation is real but whose mechanistic causal claim is undermined by the authors' own film-quality caveat; worth refereeing and worth citing for the systematic comparison.","tokens_in":758,"tokens_out":922,"would_cite":true,"duration_ms":28778,"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":"The deposition protocol for phosphonic acid modifiers on ITO determines how much of the molecule covers the contact, and that coverage—not molecular chemistry alone—controls nonradiative recombination and device efficiency in wide-bandgap…","keywords":["perovskite solar cells","self-assembled monolayers","phosphonic acid modifiers","ITO surface modification","surface coverage","nonradiative recombination","time-resolved photoluminescence","wide-bandgap perovskite"],"falsifier":"Measure the structural quality of the perovskite films—grain size, texture, crystallinity via XRD, SEM, or AFM—on each modified ITO; if the film on the etched/dip/6dPA stack has larger grains or fewer bulk defects, the greater-than-50x lifetime improvement cannot be unambiguously assigned to phosphonic acid coverage.","tokens_in":18144,"feed_emoji":"☀️","tokens_out":9366,"duration_ms":91623,"temperature":0.7,"pith_summary":"This paper argues that how a phosphonic acid interface modifier is put onto an indium tin oxide (ITO) contact matters as much as which molecule is used. By short chemical etching of the ITO, soaking it for 12 hours in the phosphonic acid solution instead of spin coating, and adding a second bis-phosphonic acid (1,6-hexylenediphosphonic acid), the authors systematically increase the surface coverage of the carbazole phosphonic acid I-2PACz. Higher coverage, read from XPS peak area ratios and work-function shifts, is accompanied by longer photoluminescence lifetimes (from about 3.8 ns to 210 ns), higher quasi-Fermi level splitting (1.16 to 1.25 eV), and higher device efficiencies for 1.7 eV bandgap perovskite solar cells. The same recipe applied to a more common modifier, Me-4PACz, raises average efficiency from about 9.75% to 18.40%. If correct, this gives the field a simple, transferable route to make buried ITO/perovskite interfaces more reproducible and less recombination-active.","feed_headline":"Etch, dip, and a second molecule lift perovskite lifetimes 55-fold","feed_subtitle":"Buried-interface coverage, not chemistry alone, explains the 55-fold lifetime and efficiency gains.","key_machinery":"The load-bearing object is the coverage of phosphonic acid molecules on ITO, treated as a measurable quantity through XPS peak-area ratios (P/In, N/In, I/In, O/In) and effective work function from UPS. I-2PACz, a diiodinated carbazole phosphonic acid, and 6dPA, a six-carbon bis-phosphonic acid, are the two modifiers studied. The three interventions—HCl/FeCl3 etching of ITO, 12-hour dip coating in place of spin coating, and sequential addition of 6dPA to I-2PACz—each raise coverage. Coverage is the variable then correlated with photoluminescence lifetime, quasi-Fermi level splitting, and device parameters. The authors also invoke the chemical and electrical heterogeneity of ITO and the presence of accessible surface hydroxyl sites as the reason uncovered regions act as recombination hot spots.","core_discovery":"The central claim is that phosphonic acid layers on ITO are not self-assembled monolayers in the usual sense: their coverage is limited by the reactivity and heterogeneity of the oxide surface, and coverage can be raised by activating the ITO, giving the molecules more time to bind, and co-depositing a small diphosphonic acid. On perovskite films grown over these modified contacts, the paper observes a direct correlation between phosphonic acid coverage, quantified by XPS In 3d attenuation and P/In, N/In, I/In, and O/In peak area ratios, and reduced nonradiative recombination: average trPL lifetimes improve 55-fold, quasi-Fermi level splitting rises step by step from 1.16 eV to 1.25 eV, and device JSC, fill factor, and PCE follow the same ordering. The authors conclude that etching, prolonged dip coating, and dual-modifier layers act on coverage, which in turn controls recombination at the buried interface; work function changes alone do not explain the improvements, since the mixed I-2PACz/6dPA layer has a lower work function yet the best lifetimes.","pith_inferences":["Beyond the paper's data, the coverage-lifetime link predicts that the passivation benefit should track coverage rather than molecular dipole, so other phosphonic acids should show the same lifetime gains if deposited by etch plus prolonged dip coating; device efficiency would still depend on work function and transport.","A testable extension: applying the same etch, 12-hour dip, and 6dPA recipe to other metal-oxide contacts such as FTO, NiOx, or AZO should reproduce the passivation gains if oxide heterogeneity is the root cause.","The paper's Pb-adsorption results suggest the interface layer changes perovskite precursor uptake; if that is general, deposition protocols could be tuned to control nucleation and film growth, not just recombination."],"forward_implications":["Protocols that raise phosphonic acid coverage—ITO etching, 12-hour dip coating, and 6dPA co-modification—each independently improve trPL lifetime, QFLS, JSC, fill factor, and PCE, and combining all three gives the best devices, with a top PCE of 14.35% for I-2PACz.","Reported performance of SAM-based perovskite devices depends sensitively on surface preparation, so literature comparisons that omit deposition protocol are likely to contain large variability.","The carbazole group is not required for low nonradiative recombination: pure 6dPA layers give among the longest lifetimes, although they make poor devices because their work function and transport properties are unsuitable.","Because VOC gains are small despite large QFLS gains, these devices are limited by film quality as well as interface recombination, so further efficiency gains need to address both."],"supporting_citations":[{"why":"Supplies the phosphonic-acid binding chemistry, dipole, and work-function framework for transparent conductive oxide modification.","marker":"[27]"},{"why":"Provides the prolonged dip-coating adsorption methodology and coverage analysis used for phosphonic acids on ITO.","marker":"[28]"},{"why":"Documents the surface composition and electrical properties of freshly deposited versus acid-etched ITO, motivating the etching step.","marker":"[33]"},{"why":"Supplies the XPS O 1s and P 2p peak assignments and binding-mode interpretation used to compare modifier layers.","marker":"[35]"},{"why":"Documents spin coating as a common phosphonic acid deposition method and the field's tendency to underreport deposition details.","marker":"[41]"},{"why":"Shows that a six-carbon diphosphonic acid can outperform carbazole phosphonic acids for passivation, serving as the baseline for 6dPA comparisons.","marker":"[50]"},{"why":"Establishes wettability improvement of carbazole-based monolayers by a second molecule, supporting the dual-modifier rationale.","marker":"[19]"},{"why":"Demonstrates electrical heterogeneity at ITO/organic interfaces with conducting-tip AFM, supporting the hot-spot recombination model.","marker":"[29]"},{"why":"Provides the photoluminescence kinetics interpretation used to extract carrier lifetimes from trPL decays.","marker":"[55]"}],"fun_headline_variants":["Coverage, not just chemistry, sets perovskite buried-interface lifetimes","Etch, dip, and co-deposit: recipe for 55-fold perovskite lifetime gains","Substrate prep dictates phosphonic acid coverage and perovskite lifetimes","55-fold lifetime gain from etching, dip coating, and a second phosphonic acid"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument assumes that the perovskite films grown on the different ITO treatments have comparable bulk quality, so the measured lifetime and QFLS gains come from the interface rather than from changes in film crystallization; the authors note this possibility themselves, and no XRD, SEM, or AFM data rule it out.","fun_headline_variants_meta":{"raw":{"variants":["Coverage, not just chemistry, sets perovskite buried-interface lifetimes","Etch, dip, and co-deposit: recipe for 55-fold perovskite lifetime gains","Substrate prep dictates phosphonic acid coverage and perovskite lifetimes","55-fold lifetime gain from etching, dip coating, and a second phosphonic acid"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001507,"raw_usage":{"total_tokens":6128,"prompt_tokens":1112,"completion_tokens":5016,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":728,"completion_tokens_details":{"reasoning_tokens":4935}},"tokens_in":728,"tokens_out":5016,"duration_ms":36064,"temperature":1.0,"reasoning_tokens":4935,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T23:06:51.889896+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the structural quality of the perovskite films—grain size, texture, crystallinity via XRD, SEM, or AFM—on each modified ITO; if the film on the etched/dip/6dPA stack has larger grains or fewer bulk defects, the greater-than-50x lifetime improvement cannot be unambiguously assigned to phosphonic acid coverage.","supporting_citations":[],"review_version":1}