{"id":"524f1f5d-6a0d-49f4-87be-d04232fdccbf","arxiv_id":"2506.08207","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Ag adatoms migrate along the tops of Si(100) dimer rows at temperatures as low as 100 K, enabling addimer formation that a naive barrier picture would forbid.","lead":"Ag atoms deposited on silicon can still pair up into dimers at very low temperatures, and this paper shows why: they land on top of the silicon's dimer rows and migrate along them even when thermal motion should be frozen out. The work replaces an earlier 'hot atom impact' explanation with a thermal migration mechanism and quantifies it with density functional theory and kinetic Monte Carlo simulations.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 0.09 eV barrier ordering at the T2 site that drives along-row hopping is within DFT error; if reversed, no low-T dimerization occurs.","rationale":"The reader's weakest-assumption analysis and mine converge on the same point: the mechanism hinges on the 0.09 eV preference for T2→T3 over T2→B. This is not an external-consensus disagreement but a genuine robustness gap. At 140 K, kBT ≈ 0.012 eV, so a 0.09 eV difference gives a rate ratio of roughly 1800; even a 0.05 eV functional or convergence error would materially change the number of along-row hops before an adatom falls into a B valley. Since the B site is essentially a trap at 140 K (B→T barrier 0.68 eV, hop time ~10^12 s), dimerization depends critically on the ordering and magnitude of those two barriers. The paper's own numbers make this clear: the effective activation energy of 0.55 eV is presented as about 20% below the nominal 0.68 eV barrier, but the mechanism that produces that reduction is the same fragile 0.09 eV difference. I therefore agree with the reader's conditional verdict: the proposed explanation is plausible, self-consistent, and supported by standard DFT/NEB and KMC practice, but the decisive energy difference is within typical DFT error and no error estimate is given. The requested additional calculations are concrete and feasible, and they would settle whether the concern is fatal or merely cosmetic. I see no reason to move the verdict away from conditional on the basis of this stress test.","tokens_in":10292,"tokens_out":3644,"duration_ms":44872,"concrete_test":"Recompute the competing barriers T2→T3 and T2→B with an independent electronic-structure method—preferably HSE06 or PBE+vdW—using the same 4×4 slab and 7×7×1 k-points, plus a 6×6 slab check; also re-converge the NEB paths to <0.005 eV/Å and verify that T3 is a real local minimum by finite differences. The mechanism survives only if ΔE ≡ E_a(T2→T3) − E_a(T2→B) remains negative by ≳0.05–0.1 eV across these checks; if it changes sign or collapses, the low-T dimerization claim should be revised.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim reduces to a kinetic competition at the T2 site (Fig. 2 inset): after leaving a B site via B→T (0.68 eV) and T→T2 (0.15 eV), an Ag adatom either continues along the row (T2→T3, 0.36 eV) or drops to a B valley (T2→B, 0.45 eV). The entire 'surprisingly low temperature' dimerization mechanism—and the claimed anisotropy—comes from this 0.09 eV difference. At 140 K the Boltzmann ratio is exp(0.09/0.012) ≈ 1800, so the prediction is exponentially sensitive to the barrier difference. PBEsol barrier errors for metal adatoms on semiconductors are routinely ~0.1 eV or more, and no error estimate or convergence tests (functional, k-points, cell size, vdW) are reported. If the true difference is smaller or reversed, adatoms would slide into B sites and become immobile (B→T is 0.68 eV), killing dimerization at 140 K. The shallow T3 'minimum' (0.005 eV barriers either side, with NEB force tolerance 0.01 eV/Å) adds a second fragility: if T3 is not a true stationary point, the along-row T2→T3→T2 hops are an artifact of the NEB interpolation. The reader's weakest assumption is therefore the right one; the claim is internally consistent but not robust at the claimed temperature.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a thermal-migration mechanism to explain the experimentally observed formation of Ag addimers on Si(100) at temperatures as low as 140 K, in contrast to the previously suggested hot-atom mechanism. The authors combine Tersoff-potential molecular dynamics of Ag deposition, adaptive kinetic Monte Carlo (AKMC) with the same empirical potential, DFT/PBEsol calculations of adsorption sites and CI-NEB minimum-energy paths, and KMC simulations based on the DFT barriers. Their central finding is that, although the static barriers for diffusion parallel and perpendicular to the Si dimer rows are nearly equal (about 0.68 eV), an Ag adatom that has escaped a stable inter-row B site preferentially migrates along a dimer row because the T2→T3 hop (0.36 eV) is 0.09 eV lower than the T2→B drop (0.45 eV). This kinetic bias produces an effective activation energy for parallel diffusion of about 0.55 eV and, in KMC simulations that assume barrierless and irreversible Ag dimer formation, leads to dimerization at 140 K. The authors conclude that the puzzling low-temperature dimerization is due to structural guidance by the Si dimer rows rather than to transient hot-atom motion.","tokens_in":10624,"tokens_out":7652,"duration_ms":82466,"significance":"The paper addresses a real experimental puzzle and offers a physically plausible mechanism that is, in principle, falsifiable: it predicts strongly anisotropic low-temperature diffusion with an effective activation energy about 20% lower than the static barrier, and a preference for Ag to land on top of the dimer rows. The methodological combination of deposition MD, systematic saddle-point searches, and KMC is standard and appropriate, and the paper contains useful comparison with previous DFT studies. If the mechanism is correct, it provides a general lesson that small static-barrier differences can create large kinetic anisotropy at low temperature. However, the central quantitative claims rest on several load-bearing assumptions that are not secured by the presented evidence, in particular the sign and magnitude of the 0.09 eV barrier difference, the barrierless dimer-formation rule, and the stability of the barely bound T3 site.","major_comments":[{"comment":"The entire low-temperature mechanism is governed by the 0.09 eV difference between the T2→T3 barrier (0.36 eV) and the T2→B barrier (0.45 eV). At 140 K the Boltzmann ratio of these rates is exp(0.09 eV/kBT) ≈ 1800, so the predicted anisotropy and dimerization yield are exponentially sensitive to this difference. The paper reports no uncertainty estimates for the barriers and no convergence tests (functional, k-points, supercell size, van der Waals treatment) for the 0.09 eV difference. Since typical DFT errors for such surface barriers are of order 0.1 eV, the sign of the difference is not secured. Please provide a systematic convergence study and a KMC sensitivity analysis in which this barrier difference is varied within a plausible range (e.g., ±0.05 eV) to show that the dimerization conclusion is robust.","section":"Fig. 2 inset and KMC results (Figs. 3-4)"},{"comment":"The dimerization simulations assume that Ag dimer formation is barrierless and irreversible when two adatoms occupy the merged B–(T+T2) or (T+T2)–T3 configurations. No DFT or NEB calculation for the two-adatom system, the dimer binding energy, or the dimer-formation barrier is presented. This rule is load-bearing because the central claim is dimerization. In addition, the T and T2 sites have different binding energies and different rate constants, and the paper does not explain how the merging of these two sites into a single species in the KMC lattice is implemented or how the elementary rates are adjusted. Please compute the dimer-formation minimum energy paths or otherwise justify the barrierless and irreversible assumption, and describe the merging procedure in detail.","section":"KMC dimerization simulations (paragraph beginning 'To investigate this, we performed KMC simulations at 140 K')"},{"comment":"The T3 site is described as 'barely a minimum' with a barrier of only 0.005 eV on either side, while the CI-NEB optimization is carried out until forces perpendicular to the path are below 0.01 eV/Å. These two numbers are of the same order, so it is not established that T3 is a true stationary point on the DFT potential energy surface. If T3 is an artifact of the NEB interpolation, the along-row T2→T3→T2 hops and the perpendicular path through Tu and T3 would not be real. Please report tighter NEB convergence (e.g., force tolerance below 0.001 eV/Å) and confirm the existence of T3 with a local minimization or a more refined search.","section":"T3 site and NEB convergence (Fig. 2 and Methods)"},{"comment":"The KMC dimerization simulations use a top-row landing probability 'four times greater than that of landing in between rows' based on counting four top-row sites (T, T2, T3, Tu) versus one B site, i.e., 80% top-row. However, the MD deposition simulations gave 91% top-row landing, and the Tu site is not included in the KMC lattice (the elementary transitions list only B, T, T2, T3). The resulting inconsistency (80% versus 91%, and an undefined treatment of Tu) affects the initial conditions of the dimerization simulations. Please use a landing probability consistent with the KMC lattice and with the MD result, and provide a sensitivity analysis of the dimerization yield to this probability.","section":"Landing-site statistics and KMC input (paragraph 'Here, the T and T2 sites were merged...')"}],"minor_comments":[{"comment":"The abstract states that the experimental observation is dimerization 'even as low as 140 K', but the body text says 'even after deposition at temperature as low as 100 K' when citing Ref. 10. Please clarify which temperature corresponds to the experimental data.","section":"Abstract and text (paragraph starting 'The reduction of the effective activation energy...')"},{"comment":"The text states that the average displacement during energy dissipation is 1.2 Å and that 'roughly half the time an atom lands, it performs one hop due to transient mobility', yet concludes that hot-atom motion is not a significant contributor to dimerization. Please quantify how the transient hop contributes to the dimerization yield, or qualify the conclusion to acknowledge a small but non-negligible transient contribution.","section":"MD deposition results (paragraph beginning 'To assess the initial impingement...')"},{"comment":"The effective activation energy of 0.55 eV is obtained from a linear fit of ln D versus 1/T over 100–300 K, but no error bars or goodness-of-fit information are reported, and the analysis uses only 10 random seeds per temperature. Please report standard errors or confidence intervals for the fitted activation energy.","section":"KMC diffusion model (paragraph containing D(T) = <L^2>/2τ)"},{"comment":"A single pre-exponential factor of 10^12 s^-1 is assumed for all elementary processes without justification or sensitivity analysis. While the kinetic competition at a fixed temperature depends only on barrier differences, the absolute timescale for dimerization (e.g., the 12800 s in the AKMC simulation) is sensitive to the prefactor, so a brief discussion or sensitivity check would be appropriate.","section":"Methods and dimerization simulations"},{"comment":"There are several minor typographical and stylistic issues: 'optiml B site' in the paragraph describing perpendicular diffusion, 'skidding on top of the dimer rows' in the dimerization paragraph, and the reference 'jeong Kong, K.' should be formatted consistently with the author's name. These do not affect the scientific content.","section":"General presentation"},{"comment":"The paper states that 'most of the deposited Ag atoms end up forming dimers' but does not report a quantitative fraction or its dependence on simulation time or coverage. Reporting the yield and its statistical uncertainty would strengthen the comparison with the experimental observation of addimer formation.","section":"Quantitative dimerization yield"}],"recommendation":"major_revision","confidential_remarks":"The paper presents an interesting and plausible mechanism, and the methodological framework is sound, but the central claim currently rests on a set of interrelated assumptions that are not fully verified: the 0.09 eV barrier ordering, the barrierless dimer-formation rule, the existence of the T3 minimum, and the landing statistics. These are all within the scope of the manuscript to address with additional calculations and sensitivity analyses, so I do not recommend rejection. However, without these additions, the quantitative prediction of dimerization at 140 K is not robust enough for publication in a journal of this standard. The inconsistency between the 80% landing ratio used in KMC and the 91% from MD, and the merging of T and T2 sites without description, should also be resolved."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe paper explains the long-standing puzzle of Ag addimer formation on Si(100) at 140 K by thermal migration on top of the Si dimer rows, explicitly rejecting the hot-atom picture of Huang et al. The mechanism itself is not entirely new — the same group demonstrated an indirect top-row path for Au on Si(100) — but this is a serious application to Ag with fresh DFT/PBEsol barriers, KMC simulations over 100–300 K, and a consistent story for why no isolated Ag adatoms are seen in STM. Credit where due: the effective 0.55 eV activation energy is an emergent output of KMC, not fitted to the experimental dimerization; the control run that raises the T2→T3 barrier to 1 eV and kills dimerization is a good sanity check; and the paper compares its binding sites against earlier PBE and LDA results.\n\nThe soft spot is exactly where the stress-test lands. The whole low-T mechanism rests on a 0.09 eV difference between T2→T3 (0.36 eV) and T2→B (0.45 eV). At 140 K that difference gives a Boltzmann ratio around 1800, so the prediction is exponentially sensitive. PBEsol barrier errors for metal adatoms on semiconductors are routinely on the order of 0.1 eV or larger, and the paper reports no error estimate, no functional dependence check, and no convergence tests for k-points, cell size, or vdW. If the true ordering is reversed or even closer, the adatom slides into B and becomes immobile at 140 K (B→T is 0.68 eV), and dimerization would not occur. A second fragility is the T3 site itself: with 0.005 eV barriers on either side and a NEB force tolerance of 0.01 eV/Å, calling T3 a minimum is borderline. If T3 is an artifact of the interpolation, the along-row T2→T3→T2 hops are not real.\n\nMinor points: the KMC landing probability for top-row sites was set to 4:1 (80% top), but the MD deposition gives 91%; the text claims agreement, which is arithmetically off. And the dimer-formation rule — two adatoms in adjacent sites make a dimer — is assumed, not computed; a DFT barrier for Ag2 formation would strengthen the claim considerably.\n\nOverall the central argument is plausible and internally consistent, but not yet robust at the claimed temperature. This deserves a serious referee who asks for convergence tests and a sensitivity analysis on the 0.09 eV difference. I would not desk-reject it.","headline":"Plausible mechanism for low-T Ag dimerization on Si(100), but the 0.09 eV barrier ordering that drives it is within DFT error and needs convergence checks before the claim is secure.","tokens_in":11132,"tokens_out":2850,"would_cite":false,"duration_ms":32188,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["68.43.Jk","68.43.Bc"],"model":"deepseek-v4-flash","headline":"Ag adatoms dimerize on Si(100) at 140 K because they migrate along the tops of the Si dimer rows by ordinary thermal motion, not by impact-driven hot-atom kicks.","keywords":["Ag adatoms","Si(100) surface","dimer-row migration","addimer formation","low-temperature deposition","kinetic Monte Carlo","density functional theory","hot-atom mechanism"],"falsifier":"A cryogenic STM experiment on a dilute Ag deposit on Si(100) at 100–140 K settles it: the explanation predicts that isolated adatoms hop along the tops of dimer rows and pair into dimers, while atoms that reach the valleys stay put; seeing adatoms move across rows, or seeing dimers form anyway when a higher-level calculation reverses the sign of the 0.09 eV barrier difference, would falsify the mechanism.","tokens_in":10102,"feed_emoji":"⚛️","tokens_out":14145,"duration_ms":147974,"temperature":0.7,"pith_summary":"The paper sets out to explain why silver adatoms deposited on the Si(100) surface at temperatures as low as 140 K are observed to form dimers, even though isolated adatoms should be frozen in place. It argues that the accepted explanation, transient hot-atom mobility from the energy of impact, is wrong, and that the real mechanism is ordinary thermal migration on top of the silicon dimer rows. The key quantitative claim is that the effective activation energy for diffusion parallel to the dimer rows is about 0.55 eV, roughly 20% lower than the 0.68 eV barrier the static energy landscape appears to impose, because an adatom makes many fast hops along a row before dropping into the stable valley between rows. If correct, this resolves an apparent contradiction between nearly isotropic energy barriers and strongly anisotropic, low-temperature diffusion, and explains the experimental absence of isolated Ag adatoms in cryogenic STM images.","feed_headline":"Silver atoms hop along silicon rows, forming dimers at 140 K","feed_subtitle":"The impact hotspot fades in 2 ps; a 0.09 eV barrier gap keeps atoms gliding along the dimer rows.","key_machinery":"The central mechanism is a multi-hop migration path along the top of a Si dimer row, through sites the paper labels T (on the row edge), T2 (between two dimers of the row) and T3 (on top of a dimer, a barely stable intermediate with a 0.005 eV barrier back to T2), with B the optimal binding site in the valley between rows. The load-bearing numbers are the barriers from the T2 site: 0.36 eV to hop along the row versus 0.45 eV to drop to B; that 0.09 eV difference is what keeps an adatom on top of the row at low temperature, enabling long runs of hops before it falls down. The supporting machinery is a three-layer simulation chain: classical molecular dynamics of deposition with a Tersoff-type potential to get landing sites and impact energy dissipation; climbing-image nudged elastic band with DFT/PBEsol forces to map the minimum energy paths; and kinetic Monte Carlo with Arrhenius rates (a pre-exponential factor of $10^{12}$ s$^{-1}$) to reach experimental timescales. The construction that carries the argument is the comparison between the static energy barrier, which looks isotropic at 0.68 eV, and the effective activation energy of about 0.55 eV extracted from the simulated temperature dependence of the parallel diffusion coefficient.","core_discovery":"The central claim is that the low-temperature aggregation of Ag on Si(100) is caused by thermally activated migration on top of the Si dimer rows, and not by transient hot-atom motion following impact. The argument rests on three computed results. First, molecular dynamics of deposition shows that 91% of impinging Ag atoms land on top of a dimer row, and that the kinetic energy of impact is dissipated within 1–2 ps, with an average surface displacement of about 1.2 Å, at most one hop and only half the time, so the hot spot cannot carry an atom to a meeting with another adatom. Second, density functional theory (PBEsol) with nudged elastic band calculations gives a nearly isotropic energy landscape, with a barrier of 0.68 eV for leaving the optimal binding site between rows, but the mechanism differs by direction: along a row the atom passes through a sequence of shallow sites and hops repeatedly on top of the row, because hopping along the row (0.36 eV from the T2 site) costs 0.09 eV less than dropping down to the stable valley (0.45 eV). Third, kinetic Monte Carlo simulations over 100–300 K show that the number of hops on top of a row before falling into the valley grows exponentially as temperature drops, so the effective activation energy for parallel diffusion becomes about 0.55 eV, 20% below the static barrier, and at 140 K the majority of deposited Ag atoms end up as dimers; artificially raising the row-hopping barrier to 1 eV kills dimer formation. The paper concludes that dimerization stems from structural guidance by the Si dimer rows rather than impact-driven mobility.","pith_inferences":["A direct experimental discriminator: cool the substrate below about 70 K so that the 0.36 eV top-row hop is thermally frozen; the paper's mechanism predicts dimer formation should essentially stop, while a residual hot-atom contribution would not feel the temperature at all.","The argument implies the dimer yield depends on the landing-site lottery: if deposition were tuned so that most atoms land directly in the inter-row valleys, dimer formation at 140 K should vanish, because the paper's own rates put the valley-to-row escape time at about $5\\times10^{11}$ s.","Because the sign of a single 0.09 eV barrier difference carries the mechanism, recomputing the T2-to-T3 and T2-to-B barriers with an electronic-structure method that can resolve the sign with confidence would settle the explanation without any new experiment.","Read as a general principle, the result shows how a nominally isotropic adsorption energy landscape can still produce strongly anisotropic, low-temperature-active diffusion whenever a low-barrier highway of shallow sites exists on the surface, a pattern worth testing for other metal/semiconductor pairs before invoking non-thermal mechanisms."],"forward_implications":["At 100–300 K an Ag adatom that finds itself on top of a dimer row makes many consecutive hops along the row before dropping into the stable valley, so at 140 K most deposited atoms dimerize within the simulation timescale.","The effective activation energy for diffusion parallel to the rows is about 0.55 eV, roughly 20% below the 0.68 eV static barrier, because the repeated top-row hops renormalize the rate.","The anisotropy is a low-temperature effect: above about 300 K the 0.09 eV difference between the competing hops becomes negligible, and diffusion approaches isotropic.","Cryogenic STM should find Ag dimers but no isolated adatoms on the surface, since any adatom landing on a row either finds a partner or falls into a valley and freezes there.","The same row-top migration mechanism previously documented for Si and Au adatoms on Si(100) is now shown for Ag, so it is a general pattern for this surface rather than a silver-specific effect."],"supporting_citations":[{"why":"The experimental report of Ag aggregate formation at low temperature whose hot-atom explanation this paper disputes; the observation the whole argument is built to explain.","marker":"[10]"},{"why":"The earlier DFT study of Ag adsorption and migration on Si(001) reporting a nearly isotropic 0.51 eV barrier, which the present PBEsol landscape closely matches.","marker":"[15]"},{"why":"The earlier ab initio calculation giving an isotropic 0.5 eV migration barrier, corroborating the near-isotropy of the static landscape.","marker":"[16]"},{"why":"The parametrized Tersoff-type Ag/Si potential used for the deposition molecular dynamics and the adaptive kinetic Monte Carlo simulations.","marker":"[17]"},{"why":"The precedent showing Si adatoms migrate on top of dimer rows and form addimers there, the same mechanism invoked here for Ag.","marker":"[21]"},{"why":"The companion study finding an indirect, top-row migration mechanism for Au adatoms on Si(100), extending the pattern the paper builds on.","marker":"[22]"}],"fun_headline_variants":["Ag dimers form at 140 K via thermal hops, not hot spots","Thermal migration on Si rows explains Ag dimerization at 140 K","Ag adatoms pair up at 140 K by hopping along Si dimer rows","Why Ag on Si(100) dimerizes at 140 K: it's the rows, not the impact","Low-temp Ag dimerization on Si(100) comes from row-guided hopping"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The mechanism collapses if the computed 0.09 eV advantage of hopping along a dimer row over dropping into the stable valley has the wrong sign, because density-functional errors for such barriers are typically of that same size and the paper gives no error estimate; it also assumes from an empirical potential that 91% of depositing atoms land on top of a row.","fun_headline_variants_meta":{"raw":{"variants":["Ag dimers form at 140 K via thermal hops, not hot spots","Thermal migration on Si rows explains Ag dimerization at 140 K","Ag adatoms pair up at 140 K by hopping along Si dimer rows","Why Ag on Si(100) dimerizes at 140 K: it's the rows, not the impact","Low-temp Ag dimerization on Si(100) comes from row-guided hopping"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000904,"raw_usage":{"total_tokens":4032,"prompt_tokens":1231,"completion_tokens":2801,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":847,"completion_tokens_details":{"reasoning_tokens":2706}},"tokens_in":847,"tokens_out":2801,"duration_ms":23624,"temperature":1.0,"reasoning_tokens":2706,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T05:17:20.462182+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A cryogenic STM experiment on a dilute Ag deposit on Si(100) at 100–140 K settles it: the explanation predicts that isolated adatoms hop along the tops of dimer rows and pair into dimers, while atoms that reach the valleys stay put; seeing adatoms move across rows, or seeing dimers form anyway when a higher-level calculation reverses the sign of the 0.09 eV barrier difference, would falsify the mechanism.","supporting_citations":[],"review_version":1}