{"id":"f700d3fd-9f9c-4394-8866-81a3404607d1","arxiv_id":"1908.07443","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"A nearby metallic screen that weakens long-range Coulomb interactions does not speed up the slow relaxation of amorphous indium-oxide electron glasses, pointing to disorder as the controlling factor.","lead":"This experiment places a metal screen near a disordered indium-oxide film to weaken the long-range part of the Coulomb interaction, then watches how the film relaxes after being pushed out of equilibrium. The relaxation is just as slow with the screen as without it, suggesting that disorder, not long-range Coulomb forces, controls the slow thermalization of these Anderson insulators.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The claimed irrelevance of long-range Coulomb interaction for relaxation rests on a single un-replicated screened-reference pair; a hidden disorder difference could mask a real effect.","rationale":"My stress-test converges with the reader's weakest assumption: the decisive dynamics comparison is a single screened-reference pair, and the paper explicitly warns that gold deposition can alter disorder. The memory-dip reduction across six pairs is a useful positive control, but it does not establish that the two samples in Fig. 9 have identical disorder. The 13-sample scaling of tau with kF l is independent evidence for disorder's role, but it cannot separate long-range interaction from disorder because carrier concentration changes both. The paper is careful and the experiments are relevant, but the central claim is under-supported by one unreplicated pair. This matches the CONDITIONAL verdict, so no change is recommended.","tokens_in":13425,"tokens_out":4868,"duration_ms":48547,"concrete_test":"Measure tau (gate-protocol) on all six screened-reference pairs reported in Figs. 3, 5, and 6, not only the 'best-matched' pair. Plot tau_screened/tau_reference against MD reduction and against the low-temperature R_sheet ratio. If the ratio is consistent with unity across pairs with MD reductions ranging 12-23%, the null result is robust; if it correlates with R_sheet mismatch or scatters beyond a factor of 2, the single-pair comparison in Fig. 9 cannot support the central claim. Ideally, add a pair with d ~ 20 nm (negligible screening) to confirm that the gold layer itself does not alter tau.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that relaxation times do not hinge on long-range Coulomb interaction rests on the null result in Fig. 9, which compares the exponential relaxation time tau of one screened and one reference sample. The paper's own caveat in 'Modifying the memory-dip by a screening-plane' acknowledges that depositing the gold layer can change disorder (annealing, strain) and that a few percent difference in room-temperature resistance can become orders of magnitude at helium temperature. The authors mitigate this by matching room-temperature resistivities, but for the dynamics the comparison is a single pair with no error bars or statistical test. The low-temperature sheet resistances of this 'best-matched' pair differ by a factor of 2 (6.3 MOhm vs 3.1 MOhm in Fig. 5). Since tau is exponentially sensitive to disorder, an accidental disorder difference in this pair could mask a genuine effect of screening on tau. The 12-23% reduction in memory-dip demonstrates that screening affects the Coulomb gap, but it does not verify identical disorder on the length scales controlling relaxation. Therefore the conclusion that long-range interaction is irrelevant is plausible but not established.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports experiments on amorphous indium-oxide Anderson insulators configured with a nearby gold screening plane, intended to shorten the spatial range of the Coulomb interaction. It makes three main claims: (i) the memory dip, a conductance feature associated with the Coulomb gap, is reduced by 12–23% in screened samples, consistent with the theoretical estimate of Hadley et al.; (ii) the long-time exponential relaxation time tau, extracted from a gate-protocol, is essentially the same for a screened sample and a reference sample, so relaxation over thousands of seconds does not require the long-range Coulomb interaction or a true gap in the density of states; and (iii) tau decreases with the disorder parameter k_F*l and tends to zero near the metal-insulator transition, supporting disorder as the dominant control parameter for the slow thermalization dynamics.","tokens_in":13539,"tokens_out":3662,"duration_ms":40259,"significance":"If the central null result is correct, it would be a useful constraint on theories of electron glasses and on the role of long-range interactions in slow thermalization of Anderson insulators. The experimental methodology is well conceived: the screening-plane geometry is a direct way to modify the interaction range, the gate-protocol is validated against thermal quench cooling, and the memory-dip reduction is shown for six sample pairs. The paper also gives explicit credit to prior theoretical work and to the author's earlier structural studies that support the interpretation of the disorder dependence. However, the central claim about dynamics rests on a single un-replicated pair comparison, and the paper itself acknowledges the difficulty of ensuring that screened and reference films differ only by image charges. Because tau is exponentially sensitive to disorder, this missing control is load-bearing rather than cosmetic.","major_comments":[{"comment":"The statement that 'limiting the range of the Coulomb interaction to approximately 8 nm does not have a significant effect on the system relaxation-time' is not established by the data as presented. The comparison is a single pair labeled the 'best-matched' screened-reference pair, with no error bars, no repeated measurements, and no statistical test. Given the exponential sensitivity of tau to disorder shown in Fig. 10, and the factor-of-two difference in low-temperature sheet resistance between the screened (R_sq=3.1 MOhm) and reference (R_sq=6.3 MOhm) samples in Fig. 5, an accidental disorder difference could mask a real effect of screening on tau. The paper's own caveat in 'Modifying the memory-dip by a screening-plane' acknowledges that depositing the gold layer can change disorder and that a few percent difference in room-temperature resistance can become orders of magnitude at helium temperatures; this caveat directly applies to the dynamics comparison. I ask the authors to provide multiple screened-reference pairs, quantitative uncertainty estimates, or a systematic variation of spacer thickness to rule out this alternative.","section":"Dynamics of screened-reference samples, Fig. 9"},{"comment":"The quantitative evidence that the screening layer is actually limiting the Coulomb interaction is weakened by the absence of a systematic dependence on spacer thickness d. The paper reports that in the six pairs, 'it was not possible to see a systematic dependence on the spacer d' and attributes this to thickness fluctuations of roughly ±8%. This means the claimed 12–23% reduction in memory-dip magnitude, and the inferred d*N^(1/3) values of 2.8 to 1.6, are only a consistency check with the theory of Hadley et al., not a controlled test. Because the gold deposition itself could alter disorder (annealing, strain), the possibility remains that part of the observed memory-dip reduction is not due to image-charge screening. A series with intentionally varied d, or a demonstration that the MD reduction is independent of R_sq across a broader set of samples, would make the central screening claim load-bearing.","section":"Modifying the memory-dip by a screening-plane"},{"comment":"The proposed scaling relation tau = tau(N) * [k_F*l - (k_F*l)_C] is introduced as 'seeming to suggest' but is not fitted or tested; tau(N) is not determined and the scatter in Fig. 10 is not quantified. Since the paper argues that disorder dominates over interaction, this relation is part of the evidence for that conclusion. I ask for either a quantitative fit to the data, a collapse test for the two carrier densities, or an explicit statement that the relation is only a qualitative guide. This is not a fatal issue, but it needs to be addressed if the 'dominant role of disorder' claim is to be supported by the data set.","section":"Disorder vs. Interaction, Fig. 10 and proposed scaling"}],"minor_comments":[{"comment":"The title contains a formatting artifact: 'An derson insulators' should be 'Anderson insulators'.","section":"Title"},{"comment":"In the paragraph introducing the gate-protocol, 'An effective and way to take the system far from equilibrium' appears to be missing a noun; it should probably read 'An effective way...'.","section":"Thermalization dynamics"},{"comment":"The paper calls this pair 'best-matched' in terms of room-temperature resistance, but the low-temperature sheet resistances differ by a factor of two. Please state explicitly what matching criterion was used and what tolerance was accepted.","section":"Fig. 5"},{"comment":"The abbreviation 'MD' is used repeatedly after the phrase 'memory-dip' but is never explicitly defined; please define it at first use.","section":"General notation"},{"comment":"The scaling relation for tau should specify units, the range of validity, and the uncertainty in the critical value (k_F*l)_C; the latter is quoted elsewhere as 0.31±0.03 but is not carried through to the scaling plot.","section":"Unnumbered equation"}],"recommendation":"major_revision","confidential_remarks":"The paper is likely to be of interest to the readership of cond-mat.dis-nn, but the central claim about the irrelevance of long-range Coulomb interactions for thermalization rests on a single pair comparison. The missing controls are experimentally feasible, so I am recommending major revision rather than rejection. I would also ask the editor to ensure the author addresses the inconsistency between the strong abstract wording and the paper's own caveats about disorder matching."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear X,\n\nQuick take: Ovadyahu has a genuinely new experimental trick here—a nearby gold plane that screens the long-range Coulomb interaction in an Anderson-insulating InO film—and he uses it to make a strong claim: long-range Coulomb forces don't control the slow relaxation dynamics. The memory-dip part is solid. The dynamics part is plausible but under-supported.\n\nWhat's good: The screening-plane technique is clever, and the six-pair demonstration that the memory dip shrinks by 12–23%, with the size roughly consistent with Hadley et al.'s prediction, is a real, citable experimental constraint. The gate-protocol validation against a thermal quench (Fig. 8) is careful. The 13-sample tau vs kFl dataset is a useful addition, especially since the relaxation time vanishes near the metal-insulator transition, which strengthens the electronic origin of the glassy dynamics.\n\nWhere I'd push back: The central null result—the irrelevance of long-range interactions for tau—rests on exactly one screened-reference pair (Fig. 9). No error bars, no repeat, no spacer-thickness systematic. And that one pair has a factor-of-two difference in low-temperature sheet resistance (6.3 vs 3.1 M). Since tau is exponentially sensitive to disorder, the near-equality of the two tau values could actually reflect a cancellation: the screened film is less disordered (faster dynamics) but maybe screening slows it, bringing it back into agreement. You can't rule that out with one pair. The paper acknowledges the disorder-matching caveat but doesn't resolve it.\n\nAlso, the memory-dip reduction across six pairs doesn't show a systematic dependence on spacer thickness, which weakens the quantitative link to Hadley's theory—though the author attributes this to film-thickness fluctuations, which is fair enough. The proposed scaling tau = tau(N)[kFl − (kFl)C] is a one-line suggestion with an unspecified prefactor; it's not a firm result.\n\nBottom line: plausible but not definitive. The memory-dip result deserves to be published; the dynamics conclusion needs more replication and a matched pair where the 4K resistances are also comparable. As a referee I'd ask for that. As a reader, I'd take the conclusion as a hypothesis, not a resolution.\n\nRecommendation: send to peer review. A serious referee can help the author either strengthen the case with another pair or soften the claim. This is exactly the kind of paper where the community benefits from scrutiny.\n\nBest,\n[You]","headline":"Clever screening method and solid memory-dip data, but the key claim about relaxation times rests on one unreplicated pair and should be read as a hypothesis, not a resolution.","tokens_in":14129,"tokens_out":3361,"would_cite":true,"duration_ms":32133,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["72.15.Rn","61.43.-j"],"model":"deepseek-v4-flash","headline":"A nearby gold plane that weakens the long-range Coulomb interaction in an Anderson-insulating indium-oxide film makes the memory-dip shallower but leaves the film's thousand-second relaxation time essentially unchanged, so slow…","keywords":["Anderson localization","electron glass","Coulomb gap","memory dip","slow thermalization","Coulomb screening","amorphous indium oxide","disorder"],"falsifier":"Measure relaxation times on a series of screened indium-oxide films with spacer thickness varied from about 5 nm to 30 nm while holding k_F l fixed; the paper's claim predicts τ does not change systematically with spacer thickness, whereas any Coulomb-gap-controlled relaxation would predict τ increasing as the screening plane is moved away.","tokens_in":13123,"feed_emoji":"⚡","tokens_out":6469,"duration_ms":59062,"temperature":0.7,"pith_summary":"This paper asks whether the very slow electronic relaxation seen in Anderson-insulating films (electron glasses) is caused by the long-range Coulomb interaction or by disorder alone. The author places a thin gold layer a few nanometres from amorphous indium-oxide films to screen the long-range part of the Coulomb interaction, and compares 'screened' with 'reference' films from the same deposition. The gold layer shrinks the memory-dip—the conductance dip that records where the system relaxed—by 12–23%, showing the screening works. Yet relaxation times of thousands of seconds are essentially unchanged, leading to the claim that slow thermalization in these systems does not require long-range Coulomb interaction or a real gap in the density of states; disorder dominates. If true, this narrows the theoretical problem: models of electron-glass dynamics can focus on disorder-generated distributions of transition rates, with short- and medium-range interactions as a secondary ingredient.","feed_headline":"Long-range Coulomb force doesn't set electron-glass relaxation times","feed_subtitle":"Screened films show a shallower memory dip yet identical thousand-second relaxation, pointing to disorder as the key.","key_machinery":"The central object is the screened-reference pair: two simultaneously deposited amorphous indium-oxide films, one with a 200 Å gold plane separated by a 7–11 nm SiO₂ spacer. The gold plane's image charges screen the long-range Coulomb interaction, while the pair isolates the interaction's effect from other sample properties. The other machinery is the gate protocol—jumping the gate voltage and fitting the late-time conductance relaxation to exp(−t/τ)—which yields the thermalization time, together with the disorder parameter k_F l used to place each sample relative to the metal-insulator transition.","core_discovery":"On the paper's own terms, the discovery is that the long-range Coulomb interaction can be switched off experimentally—by a nearby metallic plane that introduces image charges and cuts the interaction range to roughly 8 nm—and the glassy dynamics of the Anderson insulator does not care. Six screened-reference pairs of amorphous In_xO films were compared: the screened film's memory-dip is consistently shallower by 12–23%, the expected signature of a modified Coulomb gap, so the screening plane is doing its job. But the exponential relaxation time τ measured by the gate protocol is the same in screened and reference films, meaning thousands-of-seconds relaxation persists without the long-range part of the interaction. Combined with the observation that τ decreases linearly with k_F l and vanishes at the metal-insulator transition, the paper concludes that quenched disorder, not the long-range Coulomb interaction, controls slow thermalization, and that a finite density of states at the chemical potential is compatible with these long times.","pith_inferences":["A testable extension is to vary only the spacer thickness d on otherwise matched samples: the paper's claim predicts τ independent of d, while Coulomb-gap-controlled relaxation would predict τ growing with d.","The result suggests that other glassy electronic systems, such as doped semiconductors or phase-change materials, should also be re-examined with a screening plane to separate interaction from disorder contributions to memory effects.","Because both disorder and interaction grow with carrier concentration in these films, the observed N-dependence of τ may be mostly a disorder effect (higher N requires stronger disorder for localization); an experiment that tunes N by gate doping at fixed quench disorder could settle this."],"forward_implications":["Theory of electron-glass relaxation can drop the long-range Coulomb tail and still account for thousand-second relaxation times, because the experiments show a finite density of states at the Fermi level suffices.","The memory-dip magnitude and the relaxation time are decoupled: screening changes the dip but not τ, so measurements of one cannot be used to infer the other.","The vanishing of τ as k_F l approaches the critical value 0.31 supports treating the slow dynamics as an electronic, Anderson-localization-driven effect rather than a structural-defect effect.","The gate protocol is a valid substitute for thermal quench in measuring τ, provided the sample is equilibrated longer than τ; this makes repeated, history-controlled dynamics measurements practical."],"supporting_citations":[{"why":"Supplies the theoretical estimate (Hadley et al.) for how a nearby metallic plane should reduce the Coulomb-gap contribution to the memory-dip in a 2D system, used to interpret the 12–23% reduction.","marker":"[18]"},{"why":"Establishes that the memory-dip shape in these electron glasses is independent of disorder and depends on carrier concentration, which is the basis for using the dip shape as an interaction diagnostic.","marker":"[7]"},{"why":"Provides the prior characterization of how annealing changes disorder in In_xO and shows dynamics are sensitive to disorder, supporting the use of k_F l as the disorder measure.","marker":"[17]"},{"why":"Altshuler et al.'s image-charge account of modified density of states near a metallic electrode is the physical mechanism expected to make the screening plane effective.","marker":"[21]"},{"why":"Documents the slow relaxation and memory effects in electron glasses, including log(t) relaxation over many decades, which is the long-standing phenomenon the paper addresses.","marker":"[24]"},{"why":"One of the two independent determinations of the critical disorder (k_F l)_C = 0.31 for the metal-insulator transition in In_xO, used to show relaxation times vanish at the transition.","marker":"[26]"},{"why":"The second independent determination of (k_F l)_C for the lower-carrier version of In_xO, anchoring the disorder axis in Fig. 10.","marker":"[27]"}],"fun_headline_variants":["Screened Coulomb gap leaves electron-glass relaxation unchanged","Disorder, not long-range Coulomb, sets Anderson insulator relaxation","Coulomb screening fails to alter thousand-second electron-glass relaxation","Long-range Coulomb off, slow relaxation persists in Anderson insulators","Memory dip shallow, but relaxation time same: Disorder rules electron-glasses"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The screened and reference films are identical in disorder and differ only by the image charges induced by the nearby gold layer; if depositing the gold layer changes the film's disorder, the comparison between the pair cannot isolate the Coulomb interaction.","fun_headline_variants_meta":{"raw":{"variants":["Screened Coulomb gap leaves electron-glass relaxation unchanged","Disorder, not long-range Coulomb, sets Anderson insulator relaxation","Coulomb screening fails to alter thousand-second electron-glass relaxation","Long-range Coulomb off, slow relaxation persists in Anderson insulators","Memory dip shallow, but relaxation time same: Disorder rules electron-glasses"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000623,"raw_usage":{"total_tokens":2858,"prompt_tokens":888,"completion_tokens":1970,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":504,"completion_tokens_details":{"reasoning_tokens":1885}},"tokens_in":504,"tokens_out":1970,"duration_ms":11984,"temperature":1.0,"reasoning_tokens":1885,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:47:37.140106+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure relaxation times on a series of screened indium-oxide films with spacer thickness varied from about 5 nm to 30 nm while holding k_F l fixed; the paper's claim predicts τ does not change systematically with spacer thickness, whereas any Coulomb-gap-controlled relaxation would predict τ increasing as the screening plane is moved away.","supporting_citations":[{"cited_title":"Hadley, M","cited_arxiv_id":null,"evidence_quote":"Supplies the theoretical estimate (Hadley et al.) for how a nearby metallic plane should reduce the Coulomb-gap contribution to the memory-dip in a 2D system, used to interpret the 12–23% reduction."},{"cited_title":"Vaknin, Z","cited_arxiv_id":null,"evidence_quote":"Establishes that the memory-dip shape in these electron glasses is independent of disorder and depends on carrier concentration, which is the basis for using the dip shape as an interaction diagnostic."},{"cited_title":"Ovadyahu, Phys","cited_arxiv_id":null,"evidence_quote":"Provides the prior characterization of how annealing changes disorder in In_xO and shows dynamics are sensitive to disorder, supporting the use of k_F l as the disorder measure."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Altshuler et al.'s image-charge account of modified density of states near a metallic electrode is the physical mechanism expected to make the screening plane effective."},{"cited_title":"Ovadyahu, and M","cited_arxiv_id":null,"evidence_quote":"Documents the slow relaxation and memory effects in electron glasses, including log(t) relaxation over many decades, which is the long-standing phenomenon the paper addresses."},{"cited_title":"Shahar and Z","cited_arxiv_id":null,"evidence_quote":"One of the two independent determinations of the critical disorder (k_F l)_C = 0.31 for the metal-insulator transition in In_xO, used to show relaxation times vanish at the transition."},{"cited_title":"Givan and Z","cited_arxiv_id":null,"evidence_quote":"The second independent determination of (k_F l)_C for the lower-carrier version of In_xO, anchoring the disorder axis in Fig. 10."}],"review_version":1}