{"id":"c58a55d3-575b-4c99-bc24-299a8eb5e9a1","arxiv_id":"1908.02421","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"In 55-atom sodium-potassium alloy clusters, surface atoms bind their electrons more tightly than interior atoms, and the size of the shift tracks composition.","lead":"Researchers used density functional theory to calculate how the binding energy of core electrons shifts as sodium atoms replace potassium atoms in 55-atom alloy clusters. The results connect each atom's position and coordination to that shift, which may help guide the experimental preparation of alloy nanoclusters.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Overinterpretation of DFT eigenvalue shifts as BOLS coordination numbers: Kohn-Sham level shifts are not core-level binding-energy shifts, and bulk XPS constants are applied to 55-atom clusters without validation.","rationale":"The reader identified the transferability of bulk XPS constants to finite clusters and the identification of DFT eigenvalues with measured binding-energy shifts as the weakest assumption. My stress-test agrees: this is the single most load-bearing link, because all quantitative quantities in Table 2 (z, γ, δEd, εz) are derived from it. I also note a further internal inconsistency in Table 2 (identical ΔE'v for Na15K40 and Na7K48 Na S1 giving identical z) that reinforces the concern. However, the paper's qualitative claim about surface vs bulk binding-energy ordering appears to be supported by its own DFT data, and the paper correctly does not claim experimental validation for the screening tool beyond consistency. Therefore, the correct verdict is unchanged: conditional acceptance pending direct geometric validation of z and/or core-hole-aware shifts.","tokens_in":8353,"tokens_out":1933,"duration_ms":17918,"concrete_test":"Compute the radial positions and coordination numbers directly from the optimized DFT geometries for Na55, K55, Na7K48, and Na15K40 (e.g., using a cutoff of the first minimum of the pair correlation function) and compare these geometric z values with the BOLS-derived z values in Table 2 for each labeled shell. Additionally, recompute the core-level shifts with an explicit core-hole method (e.g., full core-hole or GW-like approach) for the same clusters and compare with the plain Kohn-Sham eigenvalue shifts used in the paper. If the reconstructed geometric z values disagree by more than 1–2 units with Table 2, or if core-hole-corrected shifts differ from the reported eigenvalues by more than 0.1 eV, the mapping from DFT eigenvalues to coordination numbers is not quantitatively reliable.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central bridge in this paper is the conversion of DFT-derived 'binding energy shifts' (Kohn-Sham eigenvalue differences for Na 2p and K 3p levels) into coordination numbers z via Eq. (2), which is calibrated using bulk XPS surface shifts for Na(110) and K(110) surfaces (ΔEv(12)=2.401 eV for Na, 2.754 eV for K). This transfer assumes that (a) changes in Kohn-Sham core-level eigenvalues in a 55-atom cluster equal measured binding-energy shifts in the XPS sense, including initial- and final-state effects, and (b) the BOLS z↔ΔE relation parameterized on semi-infinite bulk surfaces applies quantitatively to finite, mixed Na/K clusters. The paper itself notes only that its ordering is 'consistent with the trends' observed by Aguado et al. (Ref. 21) and Tchaplyguine et al. (Ref. 18), not that its derived z, γ, δEd, and εz values are validated. The internal inconsistency: Table 2 reports Na15K40(Na) S1 as ΔE'v=0.122 eV and z=6.76, identical to Na7K48(Na) S1, but a plot of Eq. (2) with m=1 would give a monotonic one-to-one mapping from ΔE to z; identical ΔE therefore implies identical z, yet the text claims positions with different compositions have different binding energies. More importantly, the paper does not demonstrate that its Kohn-Sham 'binding energy' values (e.g., Na 2p at -25.034 eV for Na55 vs measured XPS values in Ref. 18) can be directly identified with the physical core-level binding energies of XPS. The BOLS interpretation of Eq. (2) uses the bulk shift ΔEv(12) calibrated to XPS, so the entire chain ΔE_eigenvalue → z → γ, δEd, εz in Table 2 inherits this unvalidated identification. If this fails, the quantitative claim—that atoms with lower coordination have higher binding energies and bond energy densities—still holds qualitatively for the cluster itself, but the specific z values and derived bond strains are not reliable, and the screening rule is not validated as a predictor of stable experimental NaK alloy structures.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports DFT calculations for NaxK55-x alloy clusters (x=0–55) using VASP with PBE functionals, with starting structures taken from Aguado's Gupta-potential global minima. The authors extract Na 2p and K 3p 'binding energies' (Kohn-Sham eigenvalues) for each composition and, for four selected clusters (Na55, K55, Na7K48, Na15K40), decompose the DOS into bulk and shell components. Using BOLS theory, they convert the core-level shifts into coordination numbers, bond energy ratios, bond energy densities, and local strains (Table 2), concluding that compositional effects are determined by atomic arrangement, with lower-coordinated surface atoms having higher binding energies and bond energy densities.","tokens_in":8725,"tokens_out":13043,"duration_ms":120809,"significance":"The paper's strength is the systematic DFT screening of all 56 compositions and the clear shell-resolved LDOS, which qualitatively show that surface atoms have deeper core levels than bulk atoms. If the BOLS-based conversion were validated, the approach would offer a low-cost screening rule for chemical ordering in alkali alloy clusters and a direct connection to XPS. However, the central quantitative bridge—identifying Kohn-Sham eigenvalue shifts with XPS binding-energy shifts and applying bulk surface constants to finite clusters—is not justified, and Table 2 contains an internal inconsistency that undercuts the headline monotonic claim. The result is therefore a plausible but not yet established screening tool.","major_comments":[{"comment":"The row for Na15K40(K) S1 lists z=3.23, γ=1.202, and δEd=108.92%. Using Eq. (3) with the bond contraction relation of Eq. (1) (m=1) gives δEd = γ C^{-3} ≈ 1.202 × (1/0.832)^3 ≈ 209%, not 108.92%. The printed value also violates the paper's stated monotonic trend, since the S2 row with higher coordination (z=4.63) has δEd=151.82%, larger than the S1 value. Because the central conclusion rests on this table, the error must be corrected and the monotonic claim re-examined.","section":"Sec. 3.3, Table 2"},{"comment":"The manuscript does not state how the Na 2p and K 3p binding energies are extracted from VASP: which eigenvalues are used, how the DOS peak positions are defined, and what energy reference (e.g., vacuum level in the supercell) is applied. No convergence tests are reported for the 400 eV cutoff, Gamma-only sampling, or the 15 Å vacuum spacer. These omissions make the numerical shifts in Table 1 and Fig. 2 irreproducible; at minimum, a brief description of the extraction procedure and a convergence check for one cluster are needed.","section":"Secs. 2.1 and 3.1"},{"comment":"The conversion of DFT eigenvalue shifts into coordination numbers via Eq. (2) relies on two unverified assumptions: (i) Kohn-Sham core-level eigenvalue differences are quantitatively equal to XPS binding-energy shifts, including final-state effects; and (ii) the bulk shift constants ΔEv(12)=2.401 eV (Na) and 2.754 eV (K) from Na(110) and K(110) XPS measurements transfer unchanged to finite mixed clusters. The derived z, γ, δEd, and εz in Table 2 are therefore entirely model-dependent. The authors should validate at least one case by comparing derived coordination numbers with the known shell structure of the DFT-optimized clusters.","section":"Sec. 3.3, Eq. (2)"},{"comment":"The statement that the DFT results 'are consistent with the XPS data reported by Tchaplyguine et al.' is not backed by a quantitative comparison or a discussion of reference alignment (Fermi vs vacuum, initial vs final state). The absolute eigenvalue positions (e.g., Na 2p at −25.034 eV for Na55) are not directly comparable to measured binding energies without such corrections. The concluding claim that compositional effects are 'principally determined by the arrangement of the different atoms within its structure' is also stronger than the evidence, because the detailed shell-resolved analysis is limited to four clusters.","section":"Secs. 3.2 and 4"}],"minor_comments":[{"comment":"The Na 2p binding energy for Na36K19 is listed as '24.877' without the minus sign used for all other entries.","section":"Table 1"},{"comment":"The cluster label 'Na15K4' should be 'Na15K40'.","section":"Table 2"},{"comment":"Panel (c) uses the label 'Core' for the bulk-like component, whereas other panels use 'B'; the notation should be made uniform.","section":"Fig. 4"},{"comment":"Equation (2) is very hard to read as typeset; please ensure all subscripts, exponents, and the 'ln' argument are clearly rendered.","section":"Eq. (2)"},{"comment":"The axis label 'Relative change(%)' is ambiguous; please specify which quantity is plotted on the x-axis and what the relative change refers to.","section":"Fig. 5"},{"comment":"The heading 'Energy (eV)' should state that this is the DFT total energy without entropy, not a binding energy.","section":"Table 1"}],"recommendation":"major_revision","confidential_remarks":"The paper applies the group's BOLS formalism to Na-K alloy clusters. The DFT screening part is simple and the qualitative surface/bulk ordering in the LDOS is plausible. The main risk is that the quantitative BOLS-derived quantities are presented as established facts even though the transfer of bulk XPS constants to finite clusters is untested and the extraction of the core-level eigenvalues is undocumented. The Table 2 inconsistency is a concrete error that must be fixed. I would advise the editor to require the authors to either validate the BOLS bridge or substantially temper the quantitative claims; with those changes the paper could be publishable as a screening study."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper gives you the first systematic DFT scan of Na 2p and K 3p level positions across Na_xK_55-x (x=0 to 55), using the global-minimum structures from Aguado. That scanning dataset is new and, as far as I can tell, the qualitative finding—shell atoms have larger core-level binding energies than bulk atoms—is solidly supported by the LDOS decompositions. The DOS figures show the ordering clearly. For people modeling alkali nanoalloys, this is a useful reference.\n\nWhat is not solid is the BOLS layer. The authors convert DFT level shifts into coordination numbers z via Eq. (2), using bulk XPS shifts from Na(110) and K(110). Two problems. First, they never state how the 'binding energy' values are extracted from VASP: are these Kohn-Sham eigenvalues? PAW core levels? Projected DOS peaks? Without that, the numbers are not reproducible, and the identification of Kohn-Sham shifts with XPS binding-energy shifts is not innocent. Second, the z, gamma, deltaEd, and eps_z in Table 2 are outputs of the same BOLS relation that is used to invert them; they are model-interpretations, not independently validated structural predictions. The bulk surface constants may not transfer quantitatively to 55-atom mixed clusters. The authors themselves only claim consistency with earlier experiments or calculations. So the table should be seen as an illustration of BOLS, not a result.\n\nThe stress-test note flags an 'internal inconsistency' about identical DeltaE' giving identical z for two different clusters. I don't think that lands: if two atoms have the same local undercoordination, they can plausibly have the same shift even in different compositions. The real issue is transferability, not the math.\n\nThere are also some sloppy details: a row reads 'Na15K4' instead of Na15K40, a missing minus sign in Table 1, no convergence tests, no data or coordinates, gamma-only sampling, and a vague sentence about predicting structures that overstates what was done. These are fixable.\n\nWho is this for? Computational folks studying alkali nanoalloys, and BOLS practitioners. The DFT scan itself deserves a serious referee; the BOLS interpretation needs to be either validated or hedged. I'd send it to review but expect major revision on the extraction methodology and a rewrite of the predictive claims.","headline":"Useful new DFT screening data for alkali nanoalloys, wrapped in a BOLS analysis that reaches beyond what the method supports.","tokens_in":9322,"tokens_out":3620,"would_cite":false,"duration_ms":41046,"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 shows that in 55-atom sodium–potassium alloy clusters, atoms with fewer neighbors bind more tightly, and that this local bonding picture can be read off from computed core-level binding-energy shifts.","keywords":["sodium-potassium alloy clusters","binding energy shift","density functional theory","BOLS notation","coordination number","bond energy density","surface segregation","nanoalloy stability"],"falsifier":"Run explicit core-hole ($Z+1$) DFT on Na55, K55, Na7K48, and Na15K40 and compare the core-level shifts with the plain DFT eigenvalue shifts used here; if the ordering or magnitudes differ substantially, the BOLS conversion needs reparameterization. Alternatively, measure XPS of size-selected NaxK55-x clusters and check whether the predicted coordination numbers match the known structural coordination numbers from the global-minimum geometries.","tokens_in":8159,"feed_emoji":"⚛️","tokens_out":11121,"duration_ms":112176,"temperature":0.7,"pith_summary":"This paper asks what controls the binding energy in 55-atom sodium–potassium alloy clusters, Na$_x$K$_{55-x}$, and answers that it is the arrangement of the atoms, not just overall composition: atoms sitting in the surface shell bind more tightly than atoms inside, because surface atoms have fewer neighbours and therefore shorter, stronger bonds. Using DFT total-energy calculations on global-minimum geometries and connecting the resulting core-level energy shifts to coordination numbers through BOLS (bond-order–length–strength) notation, the authors identify which compositions place Na in bulk or shell sites and extract per-atom bond energy ratios, bond energy densities, and local bond strains. They report that lower-coordinated atoms have higher binding energies and bond energy densities, and that their calculated Na $2p$ and K $3p$ shifts track available XPS measurements. If this screening picture is right, the same DFT-plus-BOLS recipe can rank alloy compositions by stability and provide a practical guide for preparing NaK nanoalloys with wanted surface/bulk ordering.","feed_headline":"Lower-coordinated atoms bind tighter in NaK alloy clusters","feed_subtitle":"DFT screening links core-level shifts to coordination, offering a route to predict stable NaK alloy clusters.","key_machinery":"The central object is the BOLS correlation of Eq. (2), which ties a core-level binding-energy shift $\\Delta E'_v(i)$ of level $v$ at site $i$ to the site coordination number $z$ through $\\Delta E'_v(i)=\\Delta E_v(z)-\\Delta E_v(12)$ and an inversion formula that uses bulk shift constants $\\Delta E_v(12)$ taken from Na(110) and K(110) XPS measurements (2.401 eV for Na, 2.754 eV for K). This object converts the DFT-computed Kohn–Sham eigenvalue shifts into per-atom coordination numbers, and Eq. (3) then converts those numbers into the bond energy ratio $\\gamma$, the bond energy density $\\delta E_d$, and the local bond strain $-\\varepsilon_z$. The load-bearing step is treating the DFT eigenvalue shifts as measurable core-level binding-energy shifts, and using bulk surface constants inside 55-atom alloy clusters.","core_discovery":"The paper's central claim is that compositional effects in NaxK55-x nanoclusters are principally determined by the atomic arrangement within the structure, and that atoms with lower coordination numbers have higher binding energies and bond energy densities. Concretely, the DFT+BOLS analysis resolves each cluster into bulk, S2, and S1 components, and finds that the more negative (higher-magnitude) core-level binding energies always belong to the lower-coordinated surface components: in Na55 the Na $2p$ level peaks at –24.839 eV (bulk), –25.034 eV (S2), and –25.217 eV (S1), with the same ordering for K $3p$ in K55 and in the alloys. Extracted coordination numbers fall from $z=12$ in the bulk to roughly $z=3$–$7$ at the surface, while the bond energy ratio rises to 1.05–1.20 and the local bond strain to 5–17 percent. The composition scan also reveals a clear structural crossover: in K-rich clusters (x = 1–26) Na atoms occupy bulk layers; for x = 27–42 Na moves into the shell; and in Na-rich clusters K atoms sit on the shell with no inward segregation. The largest Na $2p$ binding energy occurs at Na4K51 (–25.228 eV) and the smallest at Na15K40 (–24.753 eV); K $3p$ extrema occur at Na39K16 (–15.803 eV) and Na13K42 (–15.523 eV).","pith_inferences":["A natural extension is to invert the procedure: given an experimental XPS spectrum of an unknown NaK nanoalloy, the BOLS formula could be used to infer the distribution of coordination numbers and hence the segregation pattern, turning the screening into a spectroscopy-based structure probe.","The same DFT+BOLS pipeline could be applied to other simple-metal alloys (for example Li–Na, Na–Rb, or K–Cs) to test whether the 'low coordination binds tighter' ordering and the composition-dependent surface/bulk crossover are universal.","The identification of DFT eigenvalue shifts with measured XPS binding-energy shifts is the main point a skeptical reader would probe; a direct test would be a core-hole ($Z+1$) calculation, which the paper does not include.","If the trend holds, the extreme binding energies at specific stoichiometries (Na4K51, Na15K40, Na39K16, Na13K42) could serve as composition markers in co-deposition experiments, where XPS peak positions would signal which stoichiometry has formed."],"forward_implications":["For K-rich compositions (x = 1–26), Na atoms sit in the bulk atomic layers; for x = 27–42 they move into the shell; and for Na-rich compositions the minority K atoms sit on the shell with no inward segregation, so composition alone predicts which species will be surface-enriched.","Surface and shell atoms always bind more tightly than bulk atoms, meaning the alloy's core-level spectrum carries direct information about its chemical ordering pattern (bulk–shell, onion, or random).","The calculated Na $2p$ and K $3p$ shifts track the XPS measurements and the excess-energy trends of the reference global-minimum structures, so the screening can be used to rank candidate stoichiometries before synthesis.","The lower the coordinated atom, the higher its binding energy and bond energy density, giving a simple local rule that can be checked atom-by-atom in any proposed alloy geometry."],"supporting_citations":[{"why":"Supplies the Gupta-potential optimized geometries used as starting configurations, and the excess-energy trends the binding-energy shifts are compared with.","marker":"[21]"},{"why":"Reports the global-minimum structures of 55-atom NaK nanoalloys that supply the candidate atomic arrangements screened in this work.","marker":"[19]"},{"why":"Provides the XPS core-level binding-energy measurements against which the calculated Na $2p$ and K $3p$ shifts are benchmarked.","marker":"[18]"},{"why":"Is the source of BOLS notation, the bond contraction and strengthening relations underlying the shift-to-coordination conversion.","marker":"[26]"},{"why":"Provides the bulk core-level shift constants for Na(110) and K(110) and the formula linking binding-energy shift to coordination number z.","marker":"[30]"},{"why":"Defines the bond-energy-ratio, bond-energy-density, and local-bond-strain relations used to quantify the bonding trends.","marker":"[31]"},{"why":"Establishes the projector augmented wave method used for the electronic-structure calculations.","marker":"[27]"},{"why":"Documents the plane-wave DFT implementation that produced all optimized geometries and binding energies.","marker":"[28]"},{"why":"Supplies the PBE exchange-correlation functional used throughout the DFT calculations.","marker":"[29]"}],"fun_headline_variants":["Coordination controls binding in NaK alloy clusters","NaK cluster binding energy tied to atomic arrangement","Low-coordinate atoms bind stronger in NaK alloys","NaK alloy shifts follow coordination, not just size","DFT shows NaK cluster binding driven by site coordination"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing assumption is that bulk-surface XPS shift constants for Na(110) and K(110) remain valid inside a 55-atom mixed alloy cluster, and that DFT Kohn–Sham eigenvalue shifts are identical to the measured core-level binding-energy shifts; if either part fails, the derived coordination numbers and bond parameters in Table 2 are not reliable.","fun_headline_variants_meta":{"raw":{"variants":["Coordination controls binding in NaK alloy clusters","NaK cluster binding energy tied to atomic arrangement","Low-coordinate atoms bind stronger in NaK alloys","NaK alloy shifts follow coordination, not just size","DFT shows NaK cluster binding driven by site coordination"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000174,"raw_usage":{"total_tokens":1280,"prompt_tokens":944,"completion_tokens":336,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":560,"completion_tokens_details":{"reasoning_tokens":261}},"tokens_in":560,"tokens_out":336,"duration_ms":4917,"temperature":1.0,"reasoning_tokens":261,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:44:24.950668+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run explicit core-hole ($Z+1$) DFT on Na55, K55, Na7K48, and Na15K40 and compare the core-level shifts with the plain DFT eigenvalue shifts used here; if the ordering or magnitudes differ substantially, the BOLS conversion needs reparameterization. Alternatively, measure XPS of size-selected NaxK55-x clusters and check whether the predicted coordination numbers match the known structural coordination numbers from the global-minimum geometries.","supporting_citations":[{"cited_title":"Bruma, R","cited_arxiv_id":null,"evidence_quote":"Supplies the Gupta-potential optimized geometries used as starting configurations, and the excess-energy trends the binding-energy shifts are compared with."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Reports the global-minimum structures of 55-atom NaK nanoalloys that supply the candidate atomic arrangements screened in this work."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the XPS core-level binding-energy measurements against which the calculated Na $2p$ and K $3p$ shifts are benchmarked."},{"cited_title":"Aguado, J","cited_arxiv_id":null,"evidence_quote":"Is the source of BOLS notation, the bond contraction and strengthening relations underlying the shift-to-coordination conversion."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the bulk core-level shift constants for Na(110) and K(110) and the formula linking binding-energy shift to coordination number z."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines the bond-energy-ratio, bond-energy-density, and local-bond-strain relations used to quantify the bonding trends."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes the projector augmented wave method used for the electronic-structure calculations."},{"cited_title":"Rubinovich, M","cited_arxiv_id":null,"evidence_quote":"Documents the plane-wave DFT implementation that produced all optimized geometries and binding energies."},{"cited_title":"Arslan, A","cited_arxiv_id":null,"evidence_quote":"Supplies the PBE exchange-correlation functional used throughout the DFT calculations."}],"review_version":1}