{"id":"b7c54e73-abe2-42a4-b687-cf3d99b6c730","arxiv_id":"2507.06559","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":7,"one_line_summary":"Persistent photoconductivity in hydrogenated diamond arises from random local potential fluctuations and percolative transport, and it weakens as oxygen termination reduces the surface carrier density.","lead":"Persistent photoconductivity in hydrogenated diamond, where conductivity lingers after light is turned off, is traced to random potential fluctuations at the diamond surface. The authors show that partial oxygen termination reduces the decay time from 232 seconds to 5 seconds, tying the effect to carrier density.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Bulk-defect exclusion rests on an unmeasured similarity assumption; since 400 nm excitation can access nitrogen-related bulk states, the surface-RLPF attribution is not yet separated from a bulk-trap explanation.","rationale":"The paper's central claim is that PPC in surface-conducting H-diamond is governed by random local potential fluctuations and percolation rather than by bulk traps. The data are internally consistent: with increasing ozonation, carrier density drops by roughly two orders of magnitude, decay time falls from 232 to 5 s, and the recombination barrier falls from about 150 to 54 meV; stretched-exponential decay and the temperature dependence are qualitatively consistent with the RLPF picture. The paper also gives credit to earlier surface-state calculations and to its own reproducible measurements, so the overall structure is plausible. However, the attribution to surface-only disorder depends on excluding bulk traps. The Discussion asserts that 'grain boundaries and other bulk defects are similar across all three films,' but no bulk defect density measurement is reported. This matters especially because 400 nm (3.1 eV) excitation lies in the range of nitrogen-related defect levels (1.7–3.2 eV) that the paper itself cites for CVD diamond; if bulk trap density varies between growth runs or as a side effect of ozonation, the observed monotonic trends in decay time and barrier would be equally consistent with bulk trap-assisted recombination. The statement that traps related to E_N states have 'already been ruled out' is not supported by any measurement; it rests on the same similarity assumption. This is a concrete missing-evidence issue, not a disagreement with consensus. The reader's weakest-assumption analysis identified exactly this point, so my stress-test agrees. The fix is straightforward: measure bulk defect densities or use pieces of a single film and show the defect spectra are invariant under ozonation. Until that is done, a conditional verdict is appropriate; I do not see a basis for moving to reject or accept.","tokens_in":61,"tokens_out":5257,"duration_ms":133297,"concrete_test":"Take a single as-grown hydrogen-terminated diamond film and cut it into three pieces; measure bulk defect density on each piece before ozonation using deep-level transient spectroscopy or photothermal deflection spectroscopy, plus secondary-ion mass spectrometry for nitrogen content. Then ozone-treat two pieces for 60 s and 90 s, remeasure the bulk defect spectra on the same pieces, and repeat the PPC decay-time and Arrhenius-barrier measurements. If the bulk-defect spectra are unchanged while τd and ΔE follow the reported trend, the surface-RLPF attribution is supported; if bulk defect density changes with ozonation or differs across pieces, the central claim is undercut and a bulk-trap model must be reconsidered.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central mechanistic claim is that PPC in surface-conducting hydrogenated diamond is caused by random local potential fluctuations and percolative transport at the surface, not by bulk traps. The decisive step is in Section 4 (Discussion): 'since the grain boundaries and other bulk defects are similar across all three films, their contribution to the observed PPC is likely minimal.' No measurement of bulk defect density is presented, and the text does not state whether the three samples are pieces of one film or separate growths. This is load-bearing because 400 nm (3.1 eV) excitation falls inside the 1.7–3.2 eV range of unintentional nitrogen-related defect levels that the paper itself cites for CVD diamond [20]; sub-bandgap light can directly populate those bulk levels. If ozonation altered bulk traps, or if the films differed in nitrogen content or grain-boundary defect density, the monotonic trends in decay time (232→69→5 s) and recombination barrier (150→80→54 meV) could be explained by bulk trap-assisted recombination without invoking surface potential fluctuations. The manuscript also states that traps related to E_N states have 'already been ruled out,' but the only basis offered is the same similarity assumption, not a measurement. This is a missing-evidence gap at the center of the attribution, not merely a stylistic caveat.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper investigates the origin of persistent photoconductivity (PPC) in surface-conducting hydrogen-terminated diamond (HD) films. The authors prepare three samples with different degrees of oxygen termination (HD, OHD-60s, OHD-90s) and measure photocurrent rise and decay under sub-bandgap 400 nm illumination. They fit the decay with a stretched exponential, extract recombination barriers from Arrhenius plots, and analyze the temperature-dependent photocurrent with a percolation model. They report that both the PPC decay time (232 to 5 s) and the recombination barrier (~150 to 54 meV) decrease with increasing oxygen termination. The central claim is that PPC in HD arises from random local potential fluctuations at the surface—caused by inhomogeneous hydrogen termination and adsorbate distribution—and from percolative transport, rather than from bulk traps.","tokens_in":13256,"tokens_out":3745,"duration_ms":40112,"significance":"If the mechanistic attribution is correct, the work provides a useful framework for understanding and controlling PPC in diamond-based optoelectronic devices. The experimental design is systematic: a single growth process, controlled ozonation, and consistent measurement protocols yield a monotonic trend in a key observable. The authors also make reasonable qualitative arguments against the large-lattice-relaxation and macroscopic-barrier models. However, the central claim relies on an unverified assumption that bulk defect densities are identical across the three samples, and the percolation model is only demonstrated on two of the three samples. These gaps do not necessarily invalidate the qualitative conclusion, but they require either additional experimental evidence or a more cautious interpretation before the surface-confined mechanism can be considered established.","major_comments":[{"comment":"The exclusion of bulk traps as the origin of PPC rests entirely on the assumption that \"grain boundaries and other bulk defects are similar across all three films\" (Section 4, third paragraph). No measurement of bulk defect density in the three samples is presented, and the manuscript does not state whether the samples are pieces of a single film or separate growths. This matters because 400 nm (3.1 eV) excitation lies within the 1.7–3.2 eV range of unintentional nitrogen-related defect levels that the paper itself cites for CVD diamond [20]; sub-bandgap light can directly populate such bulk states. If ozonation altered the bulk trap distribution, or if the films differ in nitrogen content or grain-boundary defects, the observed monotonic trends in decay time and recombination barrier could be explained by bulk trap-assisted recombination without invoking surface potential fluctuations. The later statement that \"we have already ruled out the traps related to EN states\" (Section 4, near Fig. 7b) is not supported by any measurement; the only basis is the same similarity assumption. This is a load-bearing gap in the attribution and should be addressed, either by measuring defect densities (e.g., photoluminescence, sub-bandgap absorption, or comparing samples from the same growth), or by explicitly reframing the conclusion as a surface mechanism that is plausible but not uniquely determined.","section":"Section 4 (Discussion)"},{"comment":"The percolation model fit is presented only for the HD and OHD-60s samples; the OHD-90s sample is excluded with the statement that its critical temperature TC falls below 80 K. This exclusion is not substantiated by a fit attempt or a quantitative criterion. Since the paper's broader conclusion is that \"the observed PPC behavior is closely associated with percolative transport processes within the HD film,\" support from only two of the three measured samples weakens the generality of the claim. The authors should either show the OHD-90s data and its fit deviation, or provide evidence (for example, from the temperature dependence of tau_d) that the percolation transition indeed occurs below the accessible temperature range for that sample.","section":"Section 3.4 and Fig. 6"},{"comment":"Several quantitative results are reported without fit uncertainties, including the stretched-exponential decay time tau_d (232, 69, and 5 s), the stretching exponent beta (0.54, 0.41, 0.38), and the growth time constants tau_1 and tau_2 in Eq. (1). Without error bars or goodness-of-fit metrics, it is difficult to assess whether the differences between samples, which are central to the trend claims, are statistically meaningful. The Arrhenius barriers are given with uncertainties (150 +/- 51, 80 +/- 11, 54 +/- 13 meV), but the decay times and exponents are not. I request that the authors provide uncertainties for all fitted parameters, or at least for tau_d, and report a measure of fit quality (e.g., R^2 or residuals) for the stretched-exponential and percolation fits.","section":"Sections 3.3 and 3.4, Eqs. (1)-(3)"}],"minor_comments":[{"comment":"There are typographical errors: \"diamand Raman band\" should be \"diamond Raman band,\" and \"qulaity\" should be \"quality.\"","section":"Section 3.1"},{"comment":"The phrase \"When H atoms on the diamond surface are partially placed by O atoms\" should read \"partially replaced by O atoms.\"","section":"Section 3.2"},{"comment":"The text refers to \"Fig. 5a, 5b and 5c respectively for HD, OHD-60s and OHD-90s,\" but Fig. 5 panels a, c, and e show the decay curves, while panels b, d, and f show the temperature dependence. The panel references in the text should be corrected to match the figure panels.","section":"Section 3.4"},{"comment":"The photocurrent is denoted In in Eq. (1) but I(t) elsewhere; using a consistent notation would improve clarity.","section":"Eq. (1)"},{"comment":"The statement that the double exponential growth fitting \"represents that two distinct and dominant processes are involved\" is not elaborated; if the authors cannot identify the processes, it would be more precise to say that two exponential components are empirically needed to describe the growth.","section":"Section 3.3"}],"recommendation":"major_revision","confidential_remarks":"The manuscript has already been accepted and published in Diamond and Related Materials (per the header note), so this report applies to the arXiv version. My main concern is the unmeasured bulk-defect similarity assumption, which is the sole basis for excluding a bulk-trap explanation; this is a load-bearing point. I also note that the percolation model is only fit to two of three samples. These issues could be resolved by adding a short experimental section on sample characterization (e.g., PL or sub-bandgap spectroscopy) or by carefully softening the exclusivity of the claim. The paper is otherwise well structured and the experimental trends are internally consistent."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a careful, systematic study with a genuine new dataset—PPC in H-diamond tuned by partial ozonation—and the RLPF/percolation interpretation is plausible. The central weakness is that the bulk-defect story is dismissed on an unmeasured similarity claim, and that is load-bearing.\n\nWhat's new: three carrier densities (1.1e13 to 2.5e11 cm^-2) from controlled ozonation, with clear trends: decay time 232→69→5 s, barrier 150→80→54 meV, onset temperature dropping. The stretched-exponential description and temperature dependence are consistent with RLPF; the percolation fit gives sensible TC for two samples. This is the first systematic handle on PPC in H-diamond via surface termination, and the data should be useful to people building diamond photodetectors or optical memory. Citation pattern is fine; the model references are standard.\n\nSoft spots, in order of importance. First, the exclusion of bulk traps: 'grain boundaries and other bulk defects are similar across all three films' is asserted, not measured. The manuscript says EN traps are 'ruled out' but the basis is that same assertion. Since 400 nm light sits inside the 1.7–3.2 eV window of N-related states in CVD diamond you cite, bulk excitation is not ruled out. You don't state whether the three samples are pieces of one film or separate growths, which matters a lot for that assumption. This is the biggest gap. Second, the percolation model is fitted, not predicted, and one of three samples (OHD-90s) is excluded because TC falls below the measurement range; that weakens the percolation claim but is honestly disclosed. Third, several fit parameters (stretched-exponential beta, growth time constants) are quoted without uncertainties; the barrier and percolation fits do have errors. These are fixable.\n\nOverall, the central claim—surface potential fluctuations, not bulk traps, control PPC in this system—is defensible but not airtight. The data don't falsify it; they support it, but the bulk-trap alternative remains a live possibility unless defect densities are measured or samples are demonstrably from the same film.\n\nWho this is for: the diamond device community and anyone working on PPC in wide-bandgap surfaces. It deserves a serious referee; if it crossed my desk I'd send it to review with a request for bulk-defect characterization (PL or EPR), explicit statement of film origin, and uncertainties on all reported fit parameters. Note the header says it has already been accepted in Diamond and Related Materials, so this may be moot, but as a contribution it's solid enough to engage with.","headline":"A useful systematic dataset on tunable PPC in H-diamond, but the bulk-trap exclusion is asserted rather than shown and needs the same scrutiny as the surface model.","tokens_in":13789,"tokens_out":3221,"would_cite":true,"duration_ms":36702,"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":"Persistent photoconductivity in hydrogenated diamond is controlled by surface disorder, not bulk traps.","keywords":["hydrogenated diamond","surface conductivity","persistent photoconductivity","random local potential fluctuations","percolation transport","ozonation","two-dimensional hole gas","surface states"],"falsifier":"Measure the bulk defect density in each film (for example with deep-level transient spectroscopy or sub-bandgap absorption). If the defect density varies across HD, OHD-60s, and OHD-90s in step with the decay times, then bulk traps could explain the trend and the surface-fluctuation claim would be undercut. Alternatively, show via Kelvin probe microscopy that surface potential fluctuations do not decrease with oxygen termination; the fitted recombination barriers then lose their assigned origin.","tokens_in":12766,"feed_emoji":"💎","tokens_out":4671,"duration_ms":46801,"temperature":0.7,"pith_summary":"This paper tries to pin down why hydrogen-terminated diamond keeps conducting after the light goes off. It argues that the persistent photoconductivity comes from random local potential fluctuations at the surface, which trap photoexcited carriers and slow their recombination, rather than from bulk defects. To test this, the authors progressively replaced hydrogen with oxygen by ozonation, which reduced the sheet carrier density and shortened the photocurrent decay time from 232 to 5 seconds. A sympathetic reader should care because identifying the true mechanism determines how diamond photodetectors and optoelectronic devices should be engineered to be fast and stable.","feed_headline":"Diamond's light memory traced to surface disorder","feed_subtitle":"Oxygen termination shrinks the recombination barrier from 150 to 54 meV and decay from 232 to 5 s.","key_machinery":"The argument is carried by the random-local-potential-fluctuation (RLPF) model, in which surface disorder creates a landscape of energy maxima and minima that prevents immediate electron-hole recombination. This model is tested against the standard alternatives: the large-lattice-relaxation model predicts stronger PPC at low temperature and the macroscopic-barrier model predicts single-exponential decay, both of which the data reject. The quantitative tools are stretched-exponential fits of the photocurrent decay, which give a decay time $\\tau_d$ and stretching exponent $\\beta$; Arrhenius fits of $\\tau_d$ above a critical temperature $T_C$, which give the recombination barrier $\\Delta E$; and fits of the photocurrent buildup to $I_{\\text{build-up}} \\propto (T - T_C)^\\mu$, which support percolative transport. Surface states introduced by hydrogen termination provide the midgap levels through which sub-bandgap photons are absorbed.","core_discovery":"The paper's central claim is that persistent photoconductivity in surface-conducting hydrogenated diamond is governed by random local potential fluctuations arising from inhomogeneous hydrogen termination and non-uniform surface adsorbates. These fluctuations create spatially separated minima in the valence band and midgap states, so photoexcited electron-hole pairs are held apart and recombine slowly, producing stretched-exponential decay. As oxygen termination increases, the surface becomes more homogeneous, Coulomb interactions between the two-dimensional hole gas and the adsorbate layer weaken, the recombination barrier falls from about 150 to 54 meV, and the decay time falls from 232 to 5 seconds. Above a critical temperature, transport proceeds by percolation between localized states, and this percolative picture fits the measured photocurrent buildup. The authors therefore conclude that bulk traps and grain boundaries, being similar across all three films, are not the source of the observed PPC.","pith_inferences":["If RLPF is correct, then depositing a uniform, strongly bonded monolayer that eliminates adsorbate inhomogeneity should suppress PPC even more than ozonation does; this is a testable prediction the paper does not make.","Kelvin probe force microscopy across the three films should show a monotonic decrease in the amplitude of surface potential fluctuations with increasing oxygen termination; the paper cites such methods but does not report those maps.","The model implies a direct link between PPC decay time and the spatial correlation length of the potential fluctuations; engineered patterns of partial termination could act as a lithographic test of percolation-limited recombination.","A cleaner test would compare single-crystal and polycrystalline hydrogenated diamond with identical termination: if PPC is truly surface-controlled, the decay times should be similar despite very different grain-boundary densities."],"forward_implications":["Controlled ozonation can reduce diamond's persistent photocurrent from minutes to seconds, making photodetectors based on hydrogen-terminated diamond respond faster.","The mechanism implies that reducing surface inhomogeneity, not passivating bulk defects, should be the design goal for low-PPC diamond devices.","The critical temperature for percolative conduction depends on carrier density, so device operating temperature and surface termination must be chosen together.","The stretched-exponential relaxation and the barrier trend give a direct lifetime metric to optimize: minimizing $\\Delta E$ below roughly 50 meV corresponds to nearly negligible PPC.","Surface chemistry, not just band structure, sets the recombination barrier, linking adsorbate control to device memory time."],"supporting_citations":[{"why":"Supplies the bulk-trap model of PPC in CVD diamond that this paper argues against.","marker":"[9]"},{"why":"Reports PPC in H-terminated single-crystal diamond, showing that surface conductivity can produce PPC without bulk defects.","marker":"[12]"},{"why":"Provides DFT evidence for unoccupied surface states below the conduction band that enable sub-bandgap absorption.","marker":"[17]"},{"why":"Shows a broad energy distribution of localized surface states within the bandgap, supporting the surface-state pathway.","marker":"[19]"},{"why":"Supplies the percolation-transition expression and the concept of a critical temperature used to fit the photocurrent buildup.","marker":"[23]"},{"why":"Defines the large-lattice-relaxation model that the paper rules out on temperature-behavior grounds.","marker":"[24]"},{"why":"Gives the Arrhenius form $\\tau_d = \\tau_0 \\exp(\\Delta E/k_BT)$ used to extract recombination barriers.","marker":"[22]"},{"why":"Provides the stretched-exponential model used to fit the non-exponential photocurrent decay.","marker":"[21]"},{"why":"Reports Kelvin probe force microscopy evidence of surface potential inhomogeneity, used to support the RLPF origin.","marker":"[37]"}],"fun_headline_variants":["Persistent photoconductivity in diamond traced to surface fluctuations","Oxygen termination shrinks diamond's photoconductivity decay","Diamond's light memory from random surface potential fluctuations","Surface disorder explains persistent photoconductivity in hydrogenated diamond","How oxygen termination controls diamond's persistent photoconductivity"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole attribution to surface disorder rests on the assumption that the three films have the same bulk defect population, because the paper does not measure defect densities and simply infers similarity from identical growth.","fun_headline_variants_meta":{"raw":{"variants":["Persistent photoconductivity in diamond traced to surface fluctuations","Oxygen termination shrinks diamond's photoconductivity decay","Diamond's light memory from random surface potential fluctuations","Surface disorder explains persistent photoconductivity in hydrogenated diamond","How oxygen termination controls diamond's persistent photoconductivity"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000573,"raw_usage":{"total_tokens":2719,"prompt_tokens":967,"completion_tokens":1752,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":583,"completion_tokens_details":{"reasoning_tokens":1669}},"tokens_in":583,"tokens_out":1752,"duration_ms":12712,"temperature":1.0,"reasoning_tokens":1669,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T18:59:55.494158+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the bulk defect density in each film (for example with deep-level transient spectroscopy or sub-bandgap absorption). If the defect density varies across HD, OHD-60s, and OHD-90s in step with the decay times, then bulk traps could explain the trend and the surface-fluctuation claim would be undercut. Alternatively, show via Kelvin probe microscopy that surface potential fluctuations do not decrease with oxygen termination; the fitted recombination barriers then lose their assigned origin.","supporting_citations":[{"cited_title":"Nebel, A","cited_arxiv_id":null,"evidence_quote":"Supplies the bulk-trap model of PPC in CVD diamond that this paper argues against."},{"cited_title":"Zakaria, Persistent photoconductivity and transport properties of the air-induced surface conducting diamond, ARPN Journal of Engineering and Applied Sciences 13 (2018) 3570–3578","cited_arxiv_id":null,"evidence_quote":"Reports PPC in H-terminated single-crystal diamond, showing that surface conductivity can produce PPC without bulk defects."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides DFT evidence for unoccupied surface states below the conduction band that enable sub-bandgap absorption."},{"cited_title":"Sobaszek, M","cited_arxiv_id":null,"evidence_quote":"Shows a broad energy distribution of localized surface states within the bandgap, supporting the surface-state pathway."},{"cited_title":"Jiang, J.Y","cited_arxiv_id":null,"evidence_quote":"Supplies the percolation-transition expression and the concept of a critical temperature used to fit the photocurrent buildup."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines the large-lattice-relaxation model that the paper rules out on temperature-behavior grounds."},{"cited_title":"Chen, R.S","cited_arxiv_id":null,"evidence_quote":"Gives the Arrhenius form $\\tau_d = \\tau_0 \\exp(\\Delta E/k_BT)$ used to extract recombination barriers."},{"cited_title":"Moore, C","cited_arxiv_id":null,"evidence_quote":"Provides the stretched-exponential model used to fit the non-exponential photocurrent decay."},{"cited_title":"Rezek, C.E","cited_arxiv_id":null,"evidence_quote":"Reports Kelvin probe force microscopy evidence of surface potential inhomogeneity, used to support the RLPF origin."}],"review_version":1}