{"id":"f6276ccd-036a-4fbd-b311-b78ddd20712e","arxiv_id":"2411.14408","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Correlated PL, TEM, and AFM show that electron-beam imaging, laser exposure, oxygen plasma, and annealing each change hBN's surface residue and optical emission, often opposing common cleaning assumptions.","lead":"This paper measures how microscopy and cleaning treatments change the structure and light emission of thin flakes of hexagonal boron nitride. It shows that the measurement tools themselves, and common treatments like oxygen plasma and annealing, alter the very properties researchers want to study.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The conclusion that oxygen plasma 'brightens' PL emission is contradicted by the paper's own Samples B and F; the evidence only supports thinning, residue reduction, and a shift to diffuse emission, with brightness sample-dependent.","rationale":"The reader's weakest assumption centered on the image-processing threshold parameters (SI Sec. IV, Fig. S6). That concern is real for the quantitative residue values, but the SI already shows the non-monotonic residue trend persists across three parameter sets, so it is less load-bearing for the core claim than the internal inconsistency on oxygen-plasma brightening. I therefore partially agree with the reader: the same small-sample caution applies, but the more decisive defect is that a headline treatment effect is asserted categorically while the paper's own data show it is not reproducible. A conditional acceptance with a mandatory revision of the brightening claim is the appropriate outcome; no new experiments are needed to settle this because the existing images contain the information.","tokens_in":17317,"tokens_out":6789,"duration_ms":72046,"concrete_test":"Re-analyze the raw confocal images for Samples B, D, and F before and after oxygen plasma: integrate the total and background-subtracted PL intensity over the same suspended-region mask with identical normalization, and report each sample separately with uncertainty. If Sample B (identical recipe to D) and Sample F show no statistically significant brightening, revise the Conclusions and Abstract to state that oxygen plasma diffuses PL emission and thins hBN while the brightness change is sample-dependent; if they do brighten, the current wording is justified.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's strongest claim includes 'oxygen plasma treatment results in thinning of hBN and reduces surface residue while brightening and diffusing PL emission' (Conclusions, Sec. IV). The brightening portion is not supported by the paper's own data. Sample D (Fig. 5c-d) shows a large increase in diffuse background emission, but Sample B, treated with the identical 5-min, 50-W, 50-sccm recipe, shows 'background brightness returns to the same level as the initial image' after plasma (SI Sec. VII, Fig. S10c), and Sample F, treated with the lower-flow recipe, shows 'little change in overall optical activity' (Fig. S11c). The main text itself concedes that 'changes in brightness were less consistent' (Sec. III.B). The categorical 'brightening' in the abstract and conclusions is therefore an overgeneralization from a single flake; if taken literally it misstates the treatment's effect and would mislead attempts to reproduce oxygen-plasma brightening. The robust, reproducible effects are morphological (thinning, reduced residue) and a conversion from localized to diffuse emission, with brightness varying across samples.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports a correlative structural and optical study of exfoliated hexagonal boron nitride (hBN) flakes suspended over holes in silicon nitride membranes, combining confocal photoluminescence (PL) microscopy, transmission electron microscopy (TEM), atomic force microscopy (AFM), and energy-dispersive X-ray spectroscopy (EDS). The authors present evidence that the measurement techniques themselves alter the samples: prolonged laser exposure causes photobleaching of localized emitters, while electron-beam irradiation in the TEM restructures surface contamination (attributed to PDMS transfer residue) and changes the optical emission, including a shift from localized to diffuse emission. They also compare two common treatments: oxygen plasma (studied on Samples B, D, and F) is reported to thin the flakes, reduce surface residue, and convert localized to diffuse emission, with a large brightness increase in Sample D; annealing (Samples C and E) is reported to increase roughness, warp/crack the flake (Sample C), increase or redistribute residue, and reduce localized emission. The central message is that correlated characterization is necessary because microscopy and treatments themselves modify hBN's optical activity and surface chemistry.","tokens_in":17526,"tokens_out":4658,"duration_ms":47797,"significance":"If the qualitative findings hold, this is a useful and timely cautionary study for the hBN quantum-emitter community. Its strengths include the genuinely correlative use of PL and TEM on the same flakes, the EDS confirmation of PDMS-related residue, the direct visual evidence for electron-beam-induced residue restructuring, and the explicit reporting of image-processing parameters with a partial robustness check for Sample A. The paper is also honest in several places about sample-to-sample variability. However, the significance is moderated by the small number of samples per treatment (often one or two), the reliance on user-tuned image segmentation for quantitative residue claims, and the overgeneralized wording in the abstract and conclusions regarding oxygen-plasma 'brightening' and annealing-induced 'warping and cracking.' These issues are fixable with rephrasing and additional uncertainty reporting, but they are load-bearing for the paper's headline conclusions.","major_comments":[{"comment":"The statement that oxygen plasma treatment results in 'brightening and diffusing PL emission' is not supported by the paper's own data across all samples. In SI Sec. VII, Sample B, treated with the identical 5-min, 50-W, 50-sccm recipe as Sample D, shows background brightness returning to the initial level after plasma (Fig. S10c), and Sample F, treated with the lower-flow recipe, shows little change in overall optical activity (Fig. S11c). The main text itself concedes that 'changes in brightness were less consistent' (Sec. III.B). The robust, reproducible effects are morphological (thinning, residue reduction) and the localized-to-diffuse transition. The abstract and conclusions should be reworded to describe brightness changes as sample-dependent rather than a categorical brightening effect.","section":"Abstract and Sec. IV (Conclusions)"},{"comment":"The conclusion that 'annealing induces warping and cracking in hBN flakes' is based primarily on TEM observation of Sample C (Sec. III.A, Fig. 4i,j). Sample E, annealed with the same recipe, is reported to show substantially increased roughness but no mention of cracking or warping (Sec. III.B, Fig. 5f-i). The categorical wording overgeneralizes from a single flake. The conclusion should be qualified to state that annealing can induce warping and cracking (observed in Sample C) and that increased roughness is the more consistently observed morphological change.","section":"Abstract and Sec. IV (Conclusions)"},{"comment":"The quantitative contamination coverage values (Sample A: 19% after 40 min; Sample B: 25% to 7%; Sample C: 7% to 23%) rest on image-processing parameters that were selected by visual inspection on the same TEM images being quantified. For Sample A, the authors provide a welcome robustness check across three parameter sets (Fig. S6k-l) showing that the non-monotonic trend is stable. However, no equivalent sensitivity analysis is provided for Samples B and C, where the parameters were re-optimized (SI Sec. V). Since the plasma-vs-annealing comparison uses these numbers, the authors should either provide similar robustness data for Samples B and C or explicitly label the coverage values as semi-quantitative. This is a load-bearing issue for the quantitative residue claims, though the qualitative visual trends are clear.","section":"SI Secs. IV-V and Figs. S6-S8"}],"minor_comments":[{"comment":"The caption reads 'Effects of annealing on Sample B and O2 plasma on Sample C', which is the reverse of the text and figure content: Sample B received oxygen plasma and Sample C received annealing. This should be corrected.","section":"Fig. 4 caption"},{"comment":"The sentence 'Figure 2e illustrates the dramatic qualitative change in the fluorescence properties following TEM imaging' should refer to Fig. 3e, not Fig. 2e, since the PL images and emitter family analysis are in Fig. 3.","section":"Sec. II.B, para. 5"},{"comment":"The sentence 'oxygen plasma at 50W with flow rate of 25 or 50 sccm for 5 or 10 minutes, respectively' is ambiguous about which flow rate pairs with which duration. The main text and SI indicate 50 sccm for 5 min and 25 sccm for 10 min; the sentence should be rephrased to avoid confusion.","section":"Methods G"},{"comment":"The emitter family analysis selects the number of groups based on reduced chi-squared, but no uncertainties are reported for the fitted density and brightness parameters, and no threshold for a 'significant' change (e.g., the reported factor-of-3 decrease in density/brightness) is defined. Reporting confidence intervals or fit residuals would strengthen the quantitative interpretation.","section":"Methods D and Fig. 3c,f"},{"comment":"In the paragraph describing Fig. S10, the text refers to 'Fig. 4.2' in one place; this should be 'Fig. 4' or a more specific panel reference.","section":"SI Sec. VII"}],"recommendation":"major_revision","confidential_remarks":"The paper is within scope for a physical-chemistry or materials-characterization journal. The empirical correlative approach is valuable, and the qualitative conclusions are likely correct, but the abstract and conclusions contain categorical claims about treatment effects that are contradicted by the authors' own control samples. The revision should focus on qualifying these claims and on providing sensitivity analyses for the residue quantification on Samples B and C. I would also encourage the authors to deposit raw TEM images and the segmentation code to support reproducibility, but this is not a requirement for acceptance."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe paper is a useful correlative study: same hBN flakes measured with PL and TEM before and after the standard treatments. The genuinely new stuff is quantitative: electron beam changes residue non-monotonically (coverage rises to 25 min then falls by 40 min) with correlated PL brightening then dimming; oxygen plasma thins flakes by ~6 nm and etches triangular pits; annealing warps and cracks flakes and redistributes residue, not cleans it. Those are real observations with direct images behind them. The central message—that the measurement tools and cleaning treatments themselves perturb the material you are trying to characterize—is well supported.\n\nThe main soft spot is that the conclusions overstate the oxygen-plasma brightening. Section III.B admits brightness changes were \"less consistent,\" and the SI shows Sample B returns to its initial background and Sample F barely changes, while only Sample D clearly brightens. The robust effects are thinning, residue reduction, and the shift from localized to diffuse emission. The abstract and conclusions should be fixed to say that. It is a real overgeneralization, not a nitpick.\n\nThe residue quantification rests on a user-tuned threshold and morphology parameters, but the authors do a sensitivity check (Fig. S6): three parameter sets all show the same non-monotonic trend, with coverage peaking at 25 minutes and falling by 40. That largely answers the worry about the trend being an artifact. The absolute coverage numbers have uncertainty, but the relative trend is robust. No raw images or code are included, which is unfortunate for reproducibility but not disqualifying.\n\nSample size is one flake per main treatment condition. That makes the generality claims weaker than the tone suggests, but the qualitative features are consistent across samples (all TEM-imaged samples brighten; all annealed samples dim). The paper is honest about the inconsistencies.\n\nWho is it for? Anyone doing hBN emitter work with TEM or PL, and anyone using oxygen plasma or annealing as \"cleaning\" steps. It should be sent to referees—two or three competent people will sort out the abstract overreach. The data and images justify revision rather than rejection. I'd cite it as a cautionary methodological reference.\n\nRecommendation: send it to peer review, with an eye on the brightness claim.","headline":"A useful correlative study that overstates oxygen-plasma brightening in the conclusions, but the core observation that measurements and treatments perturb hBN is solid and deserves referee time.","tokens_in":18099,"tokens_out":1716,"would_cite":true,"duration_ms":16175,"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":"Characterization tools and common cleaning treatments themselves alter the optical activity, surface residue, and morphology of hexagonal boron nitride, so any hBN quantum-emitter experiment must account for these induced changes.","keywords":["hexagonal boron nitride","hBN quantum emitters","photoluminescence","transmission electron microscopy","poly(dimethylsiloxane) residue","oxygen plasma treatment","annealing","correlative microscopy"],"falsifier":"A direct test would be to acquire TEM images of the same hBN region before and after electron irradiation using a parameter-free residue quantification method (for example, manual segmentation by multiple independent observers or machine learning trained on labeled images) and compare the resulting coverage curves; if the non-monotonic peak at 25 minutes disappears under ground-truth segmentation, the claimed dose-dependent residue evolution would not be supported.","tokens_in":17126,"feed_emoji":"🔬","tokens_out":2154,"duration_ms":23704,"temperature":0.7,"pith_summary":"This paper establishes that the standard tools and treatments used to study hexagonal boron nitride (hBN) are not passive: photoluminescence (PL) microscopy photobleaches the material's emission, transmission electron microscopy (TEM) changes both surface residue and emission characteristics, oxygen plasma thins the flake and diffuses its emission, and annealing warps and cracks the flake while redistributing residue and reducing optical activity. By correlating PL, TEM, and atomic-force microscopy on the same suspended flakes, the authors show that surface contamination from polymer transfer is substantial and that its evolution under electron irradiation tracks non-monotonic changes in luminescence. The practical payoff is a cautionary but constructive message: experiments aiming to identify or engineer quantum emitters in hBN must treat the measurement itself as a variable, and the correlated multi-modal platform they demonstrate can be adapted to other two-dimensional materials.","feed_headline":"Microscopes change the hBN they image","feed_subtitle":"Correlated PL, TEM, and AFM show that measurement and cleaning steps themselves reshape hBN's surface and emission.","key_machinery":"The central object is a correlated measurement platform: mechanically exfoliated multilayer hBN flakes (about 10–30 nm thick) suspended over focused-ion-beam-milled holes in silicon nitride membranes, which are compatible with both confocal photoluminescence and transmission electron microscopy. The argument is carried by three quantitative tools: an image-processing routine that thresholds TEM micrographs to compute contamination coverage and residue-size distributions; an emitter-family analysis that fits PL image histograms to extract background, density, and brightness of emitter groups; and energy-dispersive x-ray spectroscopy mapping that ties the residue to silicon, carbon, and oxygen, consistent with polydimethylsiloxane transfer contamination.","core_discovery":"The authors claim that the microscopy techniques themselves induce changes in hBN's optical activity and residue morphology: PL measurements induce photobleaching, whereas TEM measurements alter surface residue and emission characteristics. They further claim that oxygen plasma treatment results in thinning of hBN and reduces surface residue while brightening and diffusing PL emission, whereas annealing induces warping and cracking in hBN flakes and results in an increase or re-distribution of surface residue while reducing optical activity overall. These conclusions rest on a custom sample platform—hBN flakes suspended over holes in silicon nitride membranes—that allows the same flake to be examined sequentially with confocal PL, TEM, and AFM, together with quantitative image analysis of contamination coverage and emitter-family statistics.","pith_inferences":["A likely editorial extension is that some previously reported 'activation' of hBN emitters by electron irradiation or annealing may actually reflect electron-beam-induced residue redistribution or Fermi-level shifts rather than the creation of new intrinsic defects.","The non-monotonic residue trend (coverage rising to 25 minutes, then falling by 40 minutes) suggests a dose-dependent competition between hydrocarbon polymerization and beam-induced removal; this could be tested directly with in-situ mass spectrometry or controlled hydrocarbon partial pressures.","A testable extension would be to perform the same correlated measurements on flakes transferred by a residue-free method (for example, dry transfer with clean stamps) to isolate whether the observed optical changes are intrinsic to hBN or mediated by PDMS residue.","The reported recipe dependence of oxygen plasma effects implies that an optimized cleaning protocol might remove residue without etching the hBN; systematic parameter sweeps with the correlated platform could identify such a window."],"forward_implications":["If the central claim is correct, any experiment that first images hBN with TEM and then measures PL cannot assume the observed emission reflects the pristine material; the electron beam itself modifies residue and emission.","Oxygen plasma, often used as a cleaning step, measurably thins hBN and can brighten and diffuse its emission, so its use changes the very defect landscape under study.","Annealing, commonly expected to stabilize or brighten emitters, can instead warp the flake, increase surface residue, and reduce optical activity, at least for flakes in this thickness range.","Quantitative contamination coverage provides a concrete metric for comparing sample-transfer methods and cleaning recipes, enabling future optimization that avoids etching the hBN itself.","The same correlated platform can be applied to other two-dimensional materials to separate intrinsic properties from measurement-induced changes."],"supporting_citations":[{"why":"Provides the emitter-family analysis method used to quantify density and brightness of emitter groups from PL images.","marker":"[21]"},{"why":"Documents PDMS and poly(bisphenol A carbonate) transfer residue on hBN, supporting the identification of the observed surface contamination.","marker":"[36]"},{"why":"Earlier correlated TEM and PL work on hBN that this study extends, including the effect of electron irradiation on emission.","marker":"[26]"},{"why":"Explains electron-beam-induced hydrocarbon diffusion and polymerization, invoked to interpret the non-monotonic residue changes.","marker":"[33]"},{"why":"Supplies the competing mechanism of electron-beam-induced material removal, used to interpret the residue decrease at longer exposure.","marker":"[34]"},{"why":"Reports photobleaching in hBN quantum emitters, providing prior evidence for the PL-induced changes observed here.","marker":"[19]"},{"why":"Proposes organic compounds as a source of visible emission in hBN, cited as a possible link between residue changes and optical changes.","marker":"[40]"},{"why":"Shows that oxygen ion implantation can alter hBN optical properties and shift emission to a diffuse morphology, supporting the plasma-treatment interpretation.","marker":"[55]"}],"fun_headline_variants":["Microscopes alter hBN's surface and glow","PL and TEM reshape hBN while probing it","Imaging hBN changes the sample itself","Probing hBN leaves its own mark","Correlative microscopy reveals hBN's self-alteration"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The quantitative contamination coverage and residue size distributions rest on an image-processing routine whose parameters (threshold intensity, minimum area, morphological kernel) were chosen by visual inspection on the same TEM images from which the reported values are drawn, with no ground-truth masks or raw images provided, so the non-monotonic residue trend could shift with parameter choice.","fun_headline_variants_meta":{"raw":{"variants":["Microscopes alter hBN's surface and glow","PL and TEM reshape hBN while probing it","Imaging hBN changes the sample itself","Probing hBN leaves its own mark","Correlative microscopy reveals hBN's self-alteration"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000301,"raw_usage":{"total_tokens":1718,"prompt_tokens":911,"completion_tokens":807,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":527,"completion_tokens_details":{"reasoning_tokens":735}},"tokens_in":527,"tokens_out":807,"duration_ms":9075,"temperature":1.0,"reasoning_tokens":735,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T15:12:23.119627+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A direct test would be to acquire TEM images of the same hBN region before and after electron irradiation using a parameter-free residue quantification method (for example, manual segmentation by multiple independent observers or machine learning trained on labeled images) and compare the resulting coverage curves; if the non-monotonic peak at 25 minutes disappears under ground-truth segmentation, the claimed dose-dependent residue evolution would not be supported.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the emitter-family analysis method used to quantify density and brightness of emitter groups from PL images."},{"cited_title":"Bhatia, T","cited_arxiv_id":null,"evidence_quote":"Documents PDMS and poly(bisphenol A carbonate) transfer residue on hBN, supporting the identification of the observed surface contamination."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Earlier correlated TEM and PL work on hBN that this study extends, including the effect of electron irradiation on emission."},{"cited_title":"Rykaczewski, W","cited_arxiv_id":null,"evidence_quote":"Supplies the competing mechanism of electron-beam-induced material removal, used to interpret the residue decrease at longer exposure."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Reports photobleaching in hBN quantum emitters, providing prior evidence for the PL-induced changes observed here."},{"cited_title":"Neumann, X","cited_arxiv_id":null,"evidence_quote":"Proposes organic compounds as a source of visible emission in hBN, cited as a possible link between residue changes and optical changes."},{"cited_title":"Mendelson, D","cited_arxiv_id":null,"evidence_quote":"Shows that oxygen ion implantation can alter hBN optical properties and shift emission to a diffuse morphology, supporting the plasma-treatment interpretation."}],"review_version":1}