{"id":"99d27e3b-da97-41ef-a253-edb9c6e19316","arxiv_id":"2505.08569","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Silicon interlayers in Ti/Ni multilayers suppress Ni3Ti intermetallic formation by reacting with titanium, while surface silicon forms a passivating oxide.","lead":"This paper measures how thin silicon layers change the chemical and magnetic structure of titanium/nickel multilayer coatings used in X-ray and neutron optics. It finds that silicon buffers suppress the formation of nickel-titanium intermetallic compounds and that a surface silicon layer protects the top layers from oxidation.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The quantitative 1.2 nm/0.6 nm transition-layer widths rest on a slab-model XRR fit with fixed densities and no error bars or alternative-model comparison; an independent or model-free check is needed before these numbers can be treated as measured.","rationale":"The reader identified the XRR model in Section 3.4 as the weakest assumption, and I agree this is the single most load-bearing issue. The paper's qualitative conclusions are well supported by convergent XPS, XRD, and magnetometry evidence, and the authors are appropriately cautious about the silicide identification. What is not supported is the precision implied by reporting transition-layer thicknesses of 1.2 nm and 0.6 nm to two decimal places in Table 4 without error bars, without a graded-interface or density-profile alternative, and without an independent structural measurement such as TEM or atom-probe tomography. The 8% deviation between fitted and nominal period in the 4 nm Si-containing sample is an additional warning that the slab model is not uniquely constrained. These are addressable conditions rather than fatal flaws, so the reader's CONDITIONAL verdict stands without adjustment.","tokens_in":15728,"tokens_out":6811,"duration_ms":72746,"concrete_test":"Refit the [Ti/Ni]42 reflectivity curve with a graded-interface or continuous-SLD model, keeping layer thicknesses free but describing the interfaces with error functions or a multi-slice spline, and propagate fit parameters with MCMC to obtain credible intervals. If the integrated Ni-on-Ti transition width and its composition integral remain within ~0.2 nm of the Table 4 values, the concern is resolved; if the best-fit width shifts by more than ~0.3 nm or the credible interval is wider than ~0.3 nm, the quantitative headline should be downgraded to 'of order 1 nm' and presented as model-dependent.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative assertion—asymmetric transition layers of ~1.2 nm (Ni-on-Ti) and ~0.6 nm (Ti-on-Ni), and the claim that Si buffers reduce them—rests entirely on the XRR model in Section 3.4 and Table 4. In that model, the layer sequence is fixed a priori to abrupt slabs (NixTi/Ti/NixTi/Ni, plus TiSiy layers in Si-containing samples), the densities of Ti, Ni, NiO, TiO2, and C are fixed to bulk/tabulated values, and only the NixTi and TiSiy densities are allowed to vary inside narrow intervals. XRR has a strong thickness-density correlation, and the reported values carry no uncertainty estimates, no fit-quality metrics, and no comparison with alternative structural models. The same caveat extends to the Si-containing samples, where the fitted period of the 4 nm system (4.32 nm) deviates from nominal (4.0 nm) by about 8%, suggesting that the slab model may be absorbing unmodeled roughness or grading into layer thicknesses. The chemical evidence (XPS, XRD, magnetometry) supports the qualitative story and is consistent with the model, but it does not independently calibrate the absolute 1.2/0.6 nm lengths. The silicide mechanism is already acknowledged to be uncertain in Section 3.1 because no elemental Si reference was measured, so I do not treat that as the decisive issue.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports a multi-technique study of Ti/Ni multilayer structures with nominal period thicknesses of 4 nm and 10 nm, with and without ultra-thin Si interlayers. The authors use XPS, XRD, XRR, and SQUID magnetometry to characterize the interfacial transition layers. They identify asymmetric transition layers of approximately 1.2 nm at the Ni-on-Ti interface and 0.6 nm at the Ti-on-Ni interface, attribute them primarily to the intermetallic phase Ni3Ti (with a smaller NiTi2 contribution in the 10 nm-period sample), and argue that Si interlayers suppress intermetallic formation, likely by reacting with Ti to form titanium silicides. The paper also reports that a surface Si layer oxidizes in air and acts as a passivation layer. The abstract, results, and conclusions consistently present the XRR thicknesses as the quantitative basis for the asymmetry and suppression claims.","tokens_in":16014,"tokens_out":5076,"duration_ms":47819,"significance":"If the quantitative results hold, the manuscript provides useful interface-engineering guidance for Ti/Ni multilayer optics in X-ray and neutron applications, where interfacial intermixing degrades optical contrast. The qualitative conclusion that Si interlayers suppress Ni-Ti intermetallic formation is supported by converging evidence from XRD peak shifts, XPS line shapes, and magnetometry, and the authors are appropriately cautious about the limitations of XPS decomposition and about the missing elemental Si reference. The main weakness is that the central quantitative assertion—the 1.2 nm/0.6 nm transition-layer asymmetry and its reduction by Si—rests entirely on a slab-model XRR fit without uncertainty estimates or alternative-model validation. The paper would be significantly strengthened by adding such analysis, but the underlying qualitative mechanism is credible and worth publishing after revision.","major_comments":[{"comment":"The reported 1.2 nm and 0.6 nm transition-layer thicknesses are outputs of a slab-model fit in which the layer sequence is fixed a priori and the base-layer densities are fixed to bulk values, with only the NixTi and TiSiy densities varied inside narrow intervals. No uncertainty estimates, fit-quality metrics, or alternative structural models are given. Because XRR has a strong thickness-density correlation, the absolute values and the asymmetry claim are not yet established. Please add parameter confidence intervals or a correlation analysis, report fit residuals or chi-squared values, and test at least one alternative interface model (e.g., continuous grading or a different layer sequence) to show that the 1.2/0.6 nm result is not an artifact of the chosen model.","section":"Section 3.4, Table 4"},{"comment":"For the 4 nm-period Si-containing sample, the fitted period is 4.32 nm, about 8% above the nominal 4.0 nm, while the Si-free 4 nm sample fits to 4.09 nm. This discrepancy suggests that the slab model may be absorbing unmodeled roughness or interlayer grading into the fitted layer thicknesses. The authors should discuss how this affects the reliability of the individual layer thicknesses in that sample and, ideally, report the fitted period with an explicit uncertainty.","section":"Section 3.4, Table 4"},{"comment":"The XPS-derived stoichiometry of the NixTi reference film ranges from x about 1.7-2.4 (as-deposited) to x about 3.0-4.2 (after sputtering) depending on the assumed analyzer transmission exponent, and the authors state that peak decomposition of the multilayer spectra would be 'unstable and potentially misleading.' While this caution is appropriate, it means the XPS data alone cannot confirm Ni3Ti as the dominant intermetallic. The paper should state more explicitly that the Ni3Ti identification rests primarily on the XRD data and literature precedents, not on the XPS quantification.","section":"Section 3.1"}],"minor_comments":[{"comment":"The statement that the XPS setup provides 'an energy resolution of up to 2 meV' is implausible for a laboratory Al K-alpha source; please correct the unit (likely 2 eV or 0.2 eV).","section":"Section 2"},{"comment":"The binding energy '154.65 эВ' contains a Cyrillic character; replace with 'eV'.","section":"Figure 4c caption"},{"comment":"The phrase 'n ranges from 0.4 до 1.1' contains a Russian word; replace 'до' with 'to'.","section":"Section 3.1"},{"comment":"Please add explicit labels distinguishing the NixTi layer at the Ni-on-Ti interface from the NixTi layer at the Ti-on-Ni interface, since the asymmetry claim depends on this distinction.","section":"Table 4"},{"comment":"The caption states that the fitted curves are red lines, but the curves are difficult to distinguish in the figure; use distinct line styles or colors that remain visible in grayscale.","section":"Figure 8"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is within the scope of the journal and the qualitative conclusions are plausible and supported by multiple techniques. The main issue is the quantitative XRR modeling, which needs uncertainty quantification and alternative-model testing before the 1.2 nm/0.6 nm claim can be accepted. I see no circularity or novelty concern; the authors cite their prior work appropriately."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper is worth engaging with. It gives a multi-technique picture of what happens at Ti/Ni interfaces and shows—convincingly on the qualitative level—that thin Si interlayers suppress Ni3Ti/NiTi2 formation, most plausibly by reacting with Ti to form a silicide. The XPS reference measurements on a deliberately intermixed NixTi film are a nice touch, and the consistency between XPS, XRD, and magnetometry is genuine evidence, not just fitting. The XRD decomposition showing reduced Ni3Ti and NiTi2 peaks with Si is solid, and the magnetometry point (more ferromagnetic Ni with Si) is a good independent check. This is a legitimate extension of the earlier structural study by Smertin et al., adding chemical speciation and two period thicknesses.\n\nThe soft spot is exactly where the reader put it: the quantitative 1.2 nm / 0.6 nm transition-layer thicknesses come from a single XRR slab model with fixed layer sequence, fixed base-layer densities, and only NixTi/TiSiy densities varied. There are no error bars, no fit-quality metrics, no comparison with alternative models. The period mismatch for the 4 nm system (4.32 vs 4.0 nm) suggests the model may be absorbing roughness or grading into layer thicknesses. The authors are honest about the Si 2s ambiguity—no elemental Si reference was measured, so the titanium silicide assignment is plausible but not confirmed. They also admit the XPS decomposition would be unstable. None of this sinks the qualitative conclusion, because the XRD and magnetometry data point the same way. But the 1.2 nm number should be presented as a model-dependent estimate, not a measured value.\n\nWho is this for? Anyone working on Ti/Ni multilayers for X-ray or neutron optics, or on interface engineering in metal multilayers generally. The paper deserves a serious referee. It should be accepted only after the authors quantify the XRR fitting uncertainty, show some robustness checks (vary the model, give error bars), and ideally make the raw reflectivity curves and fit parameters available. If they do that, I would cite it.\n\nI would not desk-reject this. The science is honest, the methods are appropriate, and the limitations are mostly acknowledged. It needs revision, not rejection.","headline":"A competent, useful experimental study of Si interlayers in Ti/Ni multilayers, but the headline transition-layer thicknesses (1.2/0.6 nm) are slab-model fit outputs without error bars, not direct measurements.","tokens_in":16599,"tokens_out":1278,"would_cite":true,"duration_ms":15643,"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":"Silicon interlayers switch Ti/Ni interface chemistry from Ni3Ti intermetallic formation to titanium silicide, preserving the contrast these multilayer mirrors need.","keywords":["Ti/Ni multilayer","transition layers","Ni3Ti intermetallic","silicon interlayer","titanium silicide","X-ray photoelectron spectroscopy","X-ray reflectometry","multilayer mirrors"],"falsifier":"Cross-sectional scanning transmission electron microscopy or atom-probe tomography across a Ni-on-Ti interface would directly test whether a roughly 1.2 nm layer of stoichiometric Ni3Ti exists as a plateau in the elemental profile; if the measured profile shows only a graded concentration change without a 3:1 Ni:Ti plateau, the central claim fails. A fitting competition that lets the XRR model choose between a discrete intermetallic layer and a continuous diffusion profile would settle the same question more cheaply.","tokens_in":15514,"feed_emoji":"🪞","tokens_out":4246,"duration_ms":40986,"temperature":0.7,"pith_summary":"The paper sets out to identify what actually forms where nickel meets titanium in Ti/Ni multilayer mirrors and whether ultra-thin silicon layers can prevent it. Using XPS, XRD, XRR, and magnetometry on 4 nm and 10 nm period structures, it argues that a transition layer up to 1.2 nm thick grows at the Ni-on-Ti interface and consists mainly of the intermetallic Ni3Ti, with a thinner ~0.6 nm layer at the Ti-on-Ni interface. Inserting ~0.3–0.4 nm silicon buffers replaces intermetallic formation with titanium silicide, preserving more pure Ni. The result matters for X-ray and neutron optics because transition layers degrade the optical contrast that makes Ti/Ni mirrors useful.","feed_headline":"Silicon interlayers suppress Ni3Ti growth in Ti/Ni multilayers","feed_subtitle":"Thin Si buffer layers shrink the 1.2 nm intermetallic transition layer and passivate the surface, preserving Ni for X-ray and neutron…","key_machinery":"The load-bearing mechanism is a competition between two interfacial reactions: Ni + Ti forming Ni3Ti, which is thermodynamically favored and kinetically fastest among NixTi phases, versus Ti + Si forming titanium silicide, which is energetically even more favorable. The diagnostic machinery combines XPS core-level shifts (Ni 2p and Ti 2p positions and the Ni satellite splitting indicate Ni-rich intermetallic), XRD peak decomposition (Ni(111) vs Ni3Ti(004) and NiTi2(511)), XRR model fits that assign layers of NixTi and TiSiy with thicknesses, and SQUID magnetometry showing higher remanent magnetization when Ni is preserved.","core_discovery":"The central claim is that the intermixed region at Ni–Ti interfaces in as-deposited Ti/Ni multilayers is not a diffuse alloy but an extended, asymmetric intermetallic layer: about 1.2 nm of Ni-rich NixTi, primarily Ni3Ti, at the Ni-on-Ti interface and about 0.6 nm at the Ti-on-Ni interface, in both 4 nm and 10 nm period samples. Silicon interlayers of 0.3–0.4 nm react preferentially with titanium to form titanium silicide, which suppresses Ni3Ti (and NiTi2 in the 10 nm structure) and leaves more elemental Ni; surface silicon also oxidizes to SiOx, passivating the top period. The paper thus claims a chemical mechanism, not just an empirical roughness reduction, for the barrier effect.","pith_inferences":["The same thermodynamic competition should apply to other strongly intermetallic-forming pairs such as Co/Ti or Fe/Ti: a silicide-forming buffer would likely suppress intermetallic nucleation wherever the buffer's reaction enthalpy beats the intermetallic's, which is testable with the same XPS/XRD recipe.","If Ni3Ti is the kinetically favored first phase, a barrier that forms an even more stable compound with titanium than silicides do may outperform silicon; comparing Si with B4C or carbide buffers on identical period stacks would separate thermodynamic from kinetic suppression.","The claim that surface SiOx passivates the top period implies a practical recipe: a sacrificial silicon cap could protect Ni/Ti mirrors during air transport, which one could verify by tracking O 1s, Ni 2p, and Ti 2p intensities over weeks of ambient exposure.","Because the transition-layer thicknesses come from one XRR model, an independent structural check by cross-sectional electron microscopy or atom-probe tomography would also reveal whether the 1.2 nm region is a stoichiometric Ni3Ti compound or a graded concentration profile, a distinction that matters for predicting neutron-optical contrast."],"forward_implications":["At both 4 nm and 10 nm periods, the Ni-on-Ti interface carries about 1.2 nm of Ni3Ti, while the Ti-on-Ni interface carries about 0.6 nm, making the transition-layer asymmetry a robust feature of deposition order.","A 0.3–0.4 nm silicon buffer at each interface converts the intermetallic reaction into titanium silicide formation, shrinking the intermetallic layers to roughly 0.25–0.56 nm and increasing the effective pure-Ni thickness.","In the 4 nm-period system, silicon interlayers suppress titanium crystallization entirely; in the 10 nm system, the crystalline Ni3Ti and NiTi2 components are reduced.","Surface silicon oxidizes in air to SiOx and prevents oxidation of the top Ni/Ti period, whereas unprotected samples show NiO and TiO2 formation beneath the surface.","Because the transition region is dominated by intermetallic compounds rather than simple atomic mixing, strategies that block compound nucleation, not just diffusion smoothing, are the ones that preserve optical contrast."],"supporting_citations":[{"why":"Supplies first-principles formation energies showing Ni3Ti is the most stable NixTi phase and that titanium silicides are even more favorable, grounding the reaction-competition argument.","marker":"[25]"},{"why":"Provides diffusion kinetics showing Ni3Ti has the lowest activation energy for interfacial nucleation, explaining why this phase forms first.","marker":"[26,27]"},{"why":"Gives the XPS core-level shift systematics used to identify the intermetallic as Ni-rich rather than Ti-rich.","marker":"[28]"},{"why":"Establishes the prior baseline of ~0.7 nm and ~1.2 nm asymmetric transition layers in Ti/Ni multilayers that this paper extends with chemical identification.","marker":"[7–10]"},{"why":"Prior study showing silicon interlayers affect roughness and diffusion in Ni/Ti mirrors, which this paper extends to chemical composition and two period thicknesses.","marker":"[16]"},{"why":"Provides the reflectometry fitting software used to extract the layer thicknesses and densities in the XRR models.","marker":"[23,24]"},{"why":"Supports assignment of the 2θ ≈ 41.3°–41.5° peak to the fcc NiTi2(511) reflection in annealed Ti/Ni multilayers.","marker":"[63]"}],"fun_headline_variants":["Si barriers block Ni3Ti growth in Ti/Ni multilayers","Thin Si layers shrink intermetallic zones in Ti/Ni stacks","Silicon interlayers tame Ni-Ti intermetallic formation","Surface Si passivates Ti/Ni and suppresses Ni3Ti"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The claimed transition-layer thicknesses rest on an X-ray reflectivity model whose layer sequence is fixed in advance and whose base-layer densities are held at tabulated values, so if a different stacking or density profile fits the data equally well, the 1.2 nm and 0.6 nm numbers would not stand.","fun_headline_variants_meta":{"raw":{"variants":["Si barriers block Ni3Ti growth in Ti/Ni multilayers","Thin Si layers shrink intermetallic zones in Ti/Ni stacks","Silicon interlayers tame Ni-Ti intermetallic formation","Surface Si passivates Ti/Ni and suppresses Ni3Ti"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000191,"raw_usage":{"total_tokens":1334,"prompt_tokens":926,"completion_tokens":408,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":542,"completion_tokens_details":{"reasoning_tokens":332}},"tokens_in":542,"tokens_out":408,"duration_ms":4227,"temperature":1.0,"reasoning_tokens":332,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T21:50:52.559579+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Cross-sectional scanning transmission electron microscopy or atom-probe tomography across a Ni-on-Ti interface would directly test whether a roughly 1.2 nm layer of stoichiometric Ni3Ti exists as a plateau in the elemental profile; if the measured profile shows only a graded concentration change without a 3:1 Ni:Ti plateau, the central claim fails. A fitting competition that lets the XRR model choose between a discrete intermetallic layer and a continuous diffusion profile would settle the same question more cheaply.","supporting_citations":[{"cited_title":"Jain, S.P","cited_arxiv_id":null,"evidence_quote":"Supplies first-principles formation energies showing Ni3Ti is the most stable NixTi phase and that titanium silicides are even more favorable, grounding the reaction-competition argument."},{"cited_title":"Senkovskiy, D.Yu","cited_arxiv_id":null,"evidence_quote":"Gives the XPS core-level shift systematics used to identify the intermetallic as Ni-rich rather than Ti-rich."},{"cited_title":"Smertin, E","cited_arxiv_id":null,"evidence_quote":"Prior study showing silicon interlayers affect roughness and diffusion in Ni/Ti mirrors, which this paper extends to chemical composition and two period thicknesses."},{"cited_title":"Shi, Z.H","cited_arxiv_id":null,"evidence_quote":"Supports assignment of the 2θ ≈ 41.3°–41.5° peak to the fcc NiTi2(511) reflection in annealed Ti/Ni multilayers."}],"review_version":1}