{"id":"b636ab46-6f0f-46ee-b601-b829492e9607","arxiv_id":"1908.03956","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For Pd nanoparticles, potential-energy histograms and the U/Ec surface-energy ratio mark melting and cub-to-ico transitions, but both reduce to previously published methods.","lead":"Simulations of heated palladium nanoparticles compare standard melting criteria with two alternatives: a histogram of per-atom potential energies and a surface-energy ratio computed directly from total potential energy. The proposed surface-energy ratio signals the cuboctahedron-to-icosahedron shape change before melting and tracks expected size and temperature trends without needing a surface-area estimate.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eq. (9) is an unvalidated substitution that reduces gamma_p to a rescaled copy of U(T); as written it gives a negative ratio, so the claimed surface-energy advantages are unsupported.","rationale":"The reader's weakest assumption identifies exactly the unproven identity gamma_p/gamma_b = U/E_c, and my read agrees. I sharpen it: because gamma_p is proportional to U, the perceived advantages of Eq. (9) reduce to plotting the potential-energy caloric curve on a rescaled axis, and the sign convention is not specified. This removes the main support for the abstract's claim that gamma_p offers advantages over previous criteria. The rest of the paper—PED, CNA, caloric curves, size-dependent melting—is a competent qualitative MD study, and the PED histograms in Fig. 5 do support the distinction of interior, plane, edge, and corner atoms. The data-availability statement ('data cannot be shared at this time') further prevents independent verification of Fig. 6. The paper should remain CONDITIONAL, with the condition being a derivation or direct validation of Eq. (9) and a stated sign convention; it should not be accepted as a validated surface-energy criterion as written.","tokens_in":11736,"tokens_out":9616,"duration_ms":103805,"concrete_test":"Recompute gamma_p(T) for the 8-cub and 8-ico clusters using the same EAM trajectories but an independent surface-energy definition that does not invoke Eq. (9)—for example, the finite-temperature slab/cleaving surface free energy or the virial surface-stress route—and normalize by gamma_b from the same potential. Overlay this on the Eq. (9) curve of Fig. 6 over 300–1500 K, including the 1070 K cub-to-ico transition. If the independent gamma_p does not share the slope, magnitude, and local-minimum signature of U/E_c, then Eq. (9) is not a valid surface-energy proxy. As a minimal analytic check, re-derive Eq. (9) from Eqs. (8) and (4) with one consistent sign convention; Eq. (9) should follow from the surface-excess formula gamma_p = (U_p - N_t E_c)/(4 pi R_p^2).","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central new quantity is Eq. (9), gamma_p/gamma_b = U/E_c, obtained by replacing the cluster cohesive energy E_p in Jiang-Lu's Eq. (8) with the instantaneous total potential energy U. The step is not derived, and it conflates two different energies: E_p is the 0 K cohesive energy per atom, whereas U is the finite-temperature total potential energy per atom containing vibrational and configurational contributions. Because Eq. (9) makes gamma_p proportional to U, every feature in Fig. 6—the negative temperature slope, the size trend, and the local minimum at the cub-to-ico transition—is inherited from the caloric curve U(T). A scaled copy of U cannot make an allotropic transition appear more clearly than U itself; the claimed advantage is a plotting artifact. There is also a sign problem: U is negative in Fig. 2, so Eq. (9) as written yields a negative gamma_p/gamma_b at every temperature unless an unstated absolute value is used. With either sign, the relation has no independent physical content without a derivation, a comparison to a direct finite-temperature surface-energy calculation, or an independent surface-area estimate. The agreement with Guggenheim-Katayama and Tolman is therefore a restatement of the temperature and size dependence of the total potential energy, not a test of a surface-energy proxy.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports molecular dynamics simulations of Pd nanoparticles in cuboctahedral and icosahedral shapes (55-28741 atoms) using an EAM potential, with simulated heating at 1.4e12 K/s in the NVT ensemble. It compares conventional melting diagnostics (potential energy caloric curves, C_p, G(r), CNA) with two proposed diagnostics: per-atom potential energy distributions (PED) and a surface-energy ratio gamma_p/gamma_b defined by Eq. (9) as U/E_c. The PED analysis identifies interior, plane, edge, and corner atoms and shows the cub-to-ico transition as the disappearance of the third surface peak. The authors claim Eq. (9) avoids surface-area estimation and predicts correct temperature and size trends consistent with Guggenheim-Katayama and Tolman, in contrast to the slab model Eq. (4). Size-dependent melting temperatures and enthalpies are also compared with prior simulations and models.","tokens_in":12013,"tokens_out":6089,"duration_ms":62624,"significance":"The PED diagnostic is a genuine contribution: it is a simple, surface-sensitive descriptor that can distinguish surface sites and shows a clear signature of the allotropic transition. If Eq. (9) were valid, the surface-area-free gamma_p proxy would be practically useful. However, the paper's central surface-energy claim rests on an unproven substitution and, as written, a sign inconsistency; the agreement with Guggenheim-Katayama and Tolman is largely a restatement of the temperature and size dependence of U. The manuscript also lacks uncertainty estimates for transition temperatures and C_p. The work is therefore of moderate significance in its present form, with the PED part salvageable and the gamma_p part requiring substantial additional validation.","major_comments":[{"comment":"The identity gamma_p/gamma_b = U/E_c is asserted by substituting the instantaneous total potential energy U for the cluster cohesive energy E_p in Jiang-Lu's Eq. (8). This substitution is not derived, and it conflates two distinct quantities: E_p is a 0 K cohesive energy per atom, while U is a finite-temperature total potential energy per atom that includes vibrational and configurational contributions. Moreover, U is negative in Fig. 2, so Eq. (9) as written gives a negative gamma_p/gamma_b, while Fig. 6 reports positive gamma_p values in mJ/m^2; an unstated absolute value or sign convention is needed. With either sign, gamma_p is proportional to U, so the negative temperature slope, the size trend, and the local minimum at the cub-to-ico transition in Fig. 6 are inherited from U(T), not independent surface-energy predictions. The authors should either derive Eq. (9) from a physical model or validate it against a direct finite-temperature surface-energy calculation (e.g., from the surface-area-dependent Eq. (4) or an explicit surface construction). Without this, the claimed agreement with Guggenheim-Katayama and Tolman is not established.","section":"3.2.5, Eq. (9)"},{"comment":"The size-dependent comparison in Fig. 9 does not validate Eq. (9). At fixed T, Eq. (9) is simply U/E_c; its increase toward unity with diameter is the well-known size dependence of the average potential energy of the cluster, not a new surface-energy result. The figure compares Eq. (9) with Eq. (8), which has the same functional dependence on E_p, and with the slab-model Eq. (4), which gives a different trend; no independent gamma_p measurement at finite temperature is provided. Thus the statement that Eq. (9) 'predicts the correct temperature and size-dependent trend' is a restatement of U's behavior. The authors should provide a quantitative comparison with an independent surface-energy estimator and report the actual numerical values of gamma_p/gamma_b, not just normalized curves.","section":"3.3.3, Fig. 9"},{"comment":"Transition temperatures and the claim that the allotropic transition appears more clearly in C_p and gamma_p are based on visual inspection without uncertainty quantification. C_p is computed as a numerical derivative of averaged U, but no convergence tests with respect to sampling length, block size, or heating rate are reported, and no error bars are given for T_mp or for the local minimum in gamma_p. Because the central advantage claimed for gamma_p over U is that the transition 'appears more clearly,' the authors need a quantitative criterion (e.g., peak height relative to noise, or a statistical test) to support this comparison. The same applies to the step at ~1070 K in the caloric curve and the CNA percentages.","section":"3.2.1, 3.2.5, Figs. 2 and 6"}],"minor_comments":[{"comment":"The y-axis is in mJ/m^2, while Eq. (9) defines a dimensionless ratio gamma_p/gamma_b; state explicitly how the ratio is converted to absolute values (presumably multiplying by gamma_b = 2050 mJ/m^2).","section":"Fig. 6"},{"comment":"There are several typos: 'slop' should be 'slope' in Sec. 3.2.5 and Sec. 4; 'clusterss' in the Fig. 1 caption should be 'clusters'; and in ref. [16], 'M. tukesh' appears to be a typo that should be verified.","section":"Throughout"},{"comment":"The text says '2-8-cub clusters are showing a different trend,' but earlier it is stated that 2-cub and 4-cub transform to ico during relaxation; clarify which structures are actually used in the T_mp comparison.","section":"3.3.1"},{"comment":"The legend entries '12', '13', and '14' are unexplained; these appear to be cluster sizes from ref. [44] but should be labeled for the reader.","section":"Fig. 7"},{"comment":"The statement that raw/processed data cannot be shared limits reproducibility; providing at least representative trajectories or derived data for the key figures would strengthen the manuscript.","section":"Data Availability"}],"recommendation":"major_revision","confidential_remarks":"The manuscript overlaps significantly with the authors' previous work (ref. [2]); the new surface-energy relation is the main claimed advance, and it is currently unsupported. If the authors cannot provide a derivation or independent validation, the editor may consider whether the surface-energy claims should be removed and the paper refocused on the PED diagnostic."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Thanks for the report. I read the paper. The useful core is the systematic MD comparison on Pd cub/ico clusters: caloric curves, C_p, G(r), CNA, and per-atom potential energy distributions, with snapshots. They document the cub-to-ico transition below melting, distinguish it from surface melting, and show that PED resolves corner/edge/plane surface atoms while G(r) and CNA do not. That part is careful and worth keeping.\n\nThe problem is Eq. (9), the centerpiece claim. They take Jiang–Lu's broken-bond relation gamma_p/gamma_b = E_p/E_c (Eq. 8) and substitute the finite-temperature total potential energy U for the 0 K cohesive energy E_p. That substitution is asserted, not derived. U includes vibrational and configurational energy; it is not the cohesive energy, and the normalization changes meaning. There is also a sign inconsistency: U is negative in their Fig. 2 and E_c is positive, so Eq. (9) as written gives a negative ratio, yet Fig. 6 shows positive gamma_p in mJ/m^2. They never state that they take an absolute value. Either way, gamma_p becomes a linear rescaling of U(T), so the temperature slope, size trend, and local minimum at the cub-to-ico transition are inherited from the caloric curve. Calling that agreement with Guggenheim-Katayama and Tolman is not an independent test; it is restatement.\n\nThe size-dependent section, using Eqs. (4) and (9) against Tolman and Jiang models, suffers from the same issue. The critique of the slab model (Eq. 4 predicts the wrong temperature trend) is fine, but replacing it with Eq. (9) does not validate a new surface energy. No error bars, no convergence tests on C_p, and the raw data are withheld, so reproducibility is limited.\n\nWhat would fix it: derive or justify the substitution, compare gamma_p to a direct finite-temperature surface-area calculation (e.g., from a Gibbs dividing surface or stress tensor), and address the sign. As it stands, the central methodological claim is unsupported. The PED analysis, however, stands on its own and may be the more durable contribution.\n\nMy recommendation: if this comes to a journal, send it to peer review so referees can force the fix, but the paper should not be accepted with Eq. (9) as a new surface-energy criterion. Worth a reading-group discussion as a case study in why substituting energies into a broken-bond model is dangerous.","headline":"A careful MD comparison of Pd nanoparticle melting criteria is undermined by an unvalidated, sign-inconsistent surface-energy proxy that simply rescales the total potential energy.","tokens_in":12528,"tokens_out":3059,"would_cite":false,"duration_ms":31770,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proposes that the ratio of total potential energy to bulk cohesive energy gives the surface energy of a Pd nanoparticle, marking melting and cub-to-ico transitions without computing surface area.","keywords":["molecular dynamics","palladium nanoparticles","surface energy","melting transition","allotropic transition","potential energy distribution","embedded atom method","common neighbor analysis"],"falsifier":"Take one of the simulated clusters, for example the 8-cub or 8-ico cluster, compute the surface energy at several temperatures by an independent method that actually measures the surface area or uses a finite-temperature slab, and compare the result with $U/E_c$ from the same trajectory; a systematic disagreement in magnitude or in the sign of $\\partial\\gamma_p/\\partial T$ would falsify Eq. (9).","tokens_in":11542,"feed_emoji":"⚛️","tokens_out":6219,"duration_ms":72804,"temperature":0.7,"pith_summary":"Molecular-dynamics simulations of palladium cuboctahedron and icosahedron clusters with 55 to 28,741 atoms are used to argue that two simple diagnostics beat standard criteria for detecting phase transitions in nanoparticles. First, the distribution of per-atom potential energy separates interior atoms from surface atoms and even distinguishes plane, edge, and corner sites, making it possible to tell surface melting from a solid-state cub-to-ico transition. Second, the paper proposes that the normalized total potential energy $U/E_c$ directly gives the cluster surface-energy ratio $\\gamma_p/\\gamma_b$, so surface energy can be tracked without estimating surface area. If correct, these tools give a cheap, surface-area-free way to locate both melting temperatures and allotropic transitions in simulated nanoparticles, and they reproduce the known temperature and size trends of surface energy.","feed_headline":"One energy ratio flags melting and shape shifts in nanoparticles","feed_subtitle":"No surface-area estimate needed: potential energy alone gives the right temperature and size trends for cluster surface energy.","key_machinery":"The identity $\\gamma_p/\\gamma_b = U/E_c$ is the load-bearing object: $U$ is the total potential energy of the cluster, $E_c$ is the per-atom cohesive energy of the bulk, and $\\gamma_p/\\gamma_b$ is the cluster-to-bulk surface-energy ratio. The paper obtains it by substituting $U$ for the cluster cohesive energy in an existing broken-bond formula for size-dependent surface energy, and the substitution removes the need to estimate cluster surface area. A second tool is the potential-energy-distribution histogram, which resolves surface atoms by site type, namely plane, edge, and corner atoms, and therefore carries the argument that surface-sensitive phase changes can be read from energy spectra.","core_discovery":"The paper's central claim is that $\\gamma_p/\\gamma_b = U/E_c$, obtained by replacing the cluster cohesive energy in a broken-bond size-dependent surface-energy model with the cluster's total potential energy $U$, predicts the surface energy of a hot, finite nanoparticle. On this basis, $\\gamma_p$ falls with temperature, rises sharply at the cub-to-ico shape transition, and drops at melting, with the predicted size dependence following the established nonlinear size-dependent trend rather than the linear slab-model trend. The same simulation data show that a histogram of per-atom potential energies has separate peaks for interior, plane, edge, and corner atoms at low temperature; the disappearance of the corner or edge features near 1100 K flags the allotropic transition, while the main peak shifts near 1300 K at melting. Common-neighbor analysis and the radial distribution function, by contrast, cannot by themselves separate surface melting from an allotropic change.","pith_inferences":["If the identity $\\gamma_p/\\gamma_b = U/E_c$ holds generally, similar normalized-potential-energy diagnostics could be applied to other metals and alloys, and possibly to estimate surface stress or evaluate shape stability at finite temperature.","Because per-atom potential energy is inexpensive to histogram, the method could scale to larger nanocrystals or long annealing runs where structural classification becomes costly.","A direct test on a single cluster would be to compute the surface energy independently via a finite-temperature slab or thermodynamic integration and compare it with $U/E_c$; the paper does not provide that check."],"forward_implications":["Surface energy of Pd nanoparticles can be monitored throughout a heating run from the same potential-energy trajectory used for caloric curves, with no separate surface-area calculation.","Cub-to-ico transitions appear as a local minimum in $\\gamma_p$ and a minor peak in $C_p$, so allotropic changes no longer need to be inferred solely from structure analysis.","The simple slab-model expression, which gives increasing $\\gamma_p$ with temperature and a linear size dependence, is shown to contradict both empirical temperature behavior and established size dependence; the normalized-ratio form corrects this.","For small clusters of about 2 to 8 shells, the cub structure transforms to ico during relaxation or before melting, so size-dependent melting models that assume a fixed static shape systematically misestimate $T_{mp}$."],"supporting_citations":[{"why":"establishes the prior combination of caloric and structural criteria and the cub-to-ico transition that this paper extends with potential-energy-distribution and surface-energy diagnostics.","marker":"[2]"},{"why":"is the earlier MD study that interpreted the low-temperature step as shell-by-shell or surface melting, the interpretation this paper corrects.","marker":"[3]"},{"why":"supplies the specific-heat caloric-curve definition and the conventional melting-temperature reading used for comparison.","marker":"[12]"},{"why":"provides the previous MD melting-temperature and spherical-cluster surface-energy data used as benchmarks in the size-dependence comparisons.","marker":"[40]"},{"why":"is the source of the empirical temperature dependence of surface energy that Eq. (9) is required to reproduce.","marker":"[41]"},{"why":"supplies the broken-bond size-dependent surface-energy model whose cluster form Eq. (9) modifies by substituting total potential energy for cohesive energy.","marker":"[42]"},{"why":"gives the size-dependent melting-temperature model against which the simulated $T_{mp}$ values are compared.","marker":"[43]"},{"why":"provides the established size-dependent surface-tension relation used to judge the size trend of $\\gamma_p$.","marker":"[50]"}],"fun_headline_variants":["Energy distribution spots melting and shape shifts in Pd clusters","Potential energy peaks reveal atom roles in nanoparticle melting","Surface energy from potential: no area estimate needed","Nanoparticle phase shifts seen in per-atom energy histograms","Melting and shape change detected via potential energy ratio"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the cluster's total potential energy divided by the bulk cohesive energy ($U/E_c$) really equals its surface-energy ratio, a substitution asserted from a broken-bond model that the paper does not independently derive or verify.","fun_headline_variants_meta":{"raw":{"variants":["Energy distribution spots melting and shape shifts in Pd clusters","Potential energy peaks reveal atom roles in nanoparticle melting","Surface energy from potential: no area estimate needed","Nanoparticle phase shifts seen in per-atom energy histograms","Melting and shape change detected via potential energy ratio"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000176,"raw_usage":{"total_tokens":1306,"prompt_tokens":976,"completion_tokens":330,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":592,"completion_tokens_details":{"reasoning_tokens":253}},"tokens_in":592,"tokens_out":330,"duration_ms":4168,"temperature":1.0,"reasoning_tokens":253,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:56:31.814078+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take one of the simulated clusters, for example the 8-cub or 8-ico cluster, compute the surface energy at several temperatures by an independent method that actually measures the surface area or uses a finite-temperature slab, and compare the result with $U/E_c$ from the same trajectory; a systematic disagreement in magnitude or in the sign of $\\partial\\gamma_p/\\partial T$ would falsify Eq. (9).","supporting_citations":[{"cited_title":"Kateb, M","cited_arxiv_id":null,"evidence_quote":"establishes the prior combination of caloric and structural criteria and the cub-to-ico transition that this paper extends with potential-energy-distribution and surface-energy diagnostics."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"is the earlier MD study that interpreted the low-temperature step as shell-by-shell or surface melting, the interpretation this paper corrects."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"supplies the specific-heat caloric-curve definition and the conventional melting-temperature reading used for comparison."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"provides the previous MD melting-temperature and spherical-cluster surface-energy data used as benchmarks in the size-dependence comparisons."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"is the source of the empirical temperature dependence of surface energy that Eq. (9) is required to reproduce."},{"cited_title":"Jiang, H","cited_arxiv_id":null,"evidence_quote":"supplies the broken-bond size-dependent surface-energy model whose cluster form Eq. (9) modifies by substituting total potential energy for cohesive energy."},{"cited_title":"Safaei, The e ffect of the averaged structural and energetic features on the cohesive energy of nanocrystals, Journal of Nanoparticle Research 12 (2010) 759–776","cited_arxiv_id":null,"evidence_quote":"gives the size-dependent melting-temperature model against which the simulated $T_{mp}$ values are compared."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"provides the established size-dependent surface-tension relation used to judge the size trend of $\\gamma_p$."}],"review_version":1}