{"id":"95cc6bed-5328-443f-9382-647826b56adb","arxiv_id":"1908.06620","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"First-principles calculations predict hot electron thermalization in titanium nitride is about 50 times faster than in gold, while nitrogen vacancies slow it to about 215 femtoseconds.","lead":"This paper computes from first principles how fast hot electrons cool in titanium nitride, a cheap material used in plasmonics, including samples with nitrogen vacancies and oxygen impurities. It finds cooling happens far faster than in gold, with the slowest cases around 200 femtoseconds, which matters for designing hot-carrier devices.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"VCA defect predictions are unbenchmarked; the 215 fs vacancy value is an upper bound that could change under explicit supercell treatment.","rationale":"The reader's weakest assumption and my stress-test converge on VCA. The paper is strong on the pristine system: the Au validation, consistent linewidth analysis, and the large ratio between TiN and Au coupling all support sub-picosecond thermalization in defect-free TiN. The unvalidated VCA is the only route to the defective-system numbers, and the paper's own caveat turns the vacancy result into an upper bound. Since actual defective TiN could cool faster, the specific claim that vacancies increase thermalization time from 0.15 to 0.215 ps may fail once explicit local defects are included. A single explicit supercell calculation for one near-stoichiometric composition would settle this. I also note that Eq. (7) as printed appears to contain d f/d T_e rather than -d f/d epsilon, which would not reduce to Eq. (8); given the excellent Au benchmark, this is likely a transcription/OCR artifact rather than a real error, and I do not treat it as the primary concern. The appropriate verdict remains the reader's CONDITIONAL.","tokens_in":11586,"tokens_out":12790,"duration_ms":125669,"concrete_test":"Construct a 2x2x2 rock-salt supercell of TiN with one N vacancy (Ti8N7, composition TiN0.875, comparable to the studied TiN0.88), relax it with the same BLYP ultrasoft pseudopotentials, and compute the Eliashberg function, lambda<omega^2>, G(Te), and tau_ep(Te=5000 K) using the same Quantum Espresso/EPW workflow used for the pristine cells. Compare this explicit supercell result with the VCA curve for TiN0.88. If the supercell tau_ep differs from the VCA value by more than ~25%, or if it no longer exceeds the pristine TiN value of 150 fs, the vacancy trend is an artifact of VCA. Also inspect the supercell DOS near the Fermi level for vacancy-induced states, whose absence in VCA would explain any discrepancy.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The defect-engineering conclusion rests entirely on the virtual crystal approximation (Computational details, second paragraph). For TiN0.84, a 16% nitrogen vacancy concentration, VCA replaces real vacancies by an average atom; it cannot produce localized defect states, vacancy-induced lattice relaxation, or the extra electron-phonon scattering channels of a real defective lattice. The paper concedes this: \"the VCA does not capture all defect-induced relaxation channels and therefore the calculated relaxation times should be interpreted as upper bounds.\" Consequently the headline value 215 fs for TiN0.84 is only an upper bound. The actual value could be close to or below the 150 fs of pristine TiN, so the claimed vacancy-induced increase — the paper's defect-engineering message — is not established. The companion claim that defect-free TiN thermalizes much faster than Au is independent of VCA and well supported by the Au benchmark; the quantitative defect trends are the soft spot.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper computes electron-phonon coupling and hot-electron thermalization times in pristine and defective titanium nitride (TiN) using a two-temperature model whose parameters (electron and phonon heat capacities and the electron-phonon coupling factor G) are obtained from density-functional theory. Defects (nitrogen vacancies and oxygen substitutionals) are modeled with the virtual crystal approximation. The authors report that hot-electron thermalization in TiN is much faster than in gold, with a maximum thermalization time of about 0.15 ps for pristine TiN and about 0.215 ps for nitrogen-deficient TiN0.84, and they conclude that defect engineering can tune hot-carrier dynamics in TiN.","tokens_in":11659,"tokens_out":8067,"duration_ms":81562,"significance":"If the results hold, the paper provides a useful first-principles parameter set for two-temperature-model simulations of TiN, a material of growing interest for hot-carrier plasmonic devices. The work is genuinely parameter-free: all model inputs are computed from DFT, with no fitting to the TiN target data. The gold validation (G = 2.2×10^16 W/m^3/K, λ = 0.19, thermalization times of order 1–10 ps) gives confidence in the workflow for the pristine case. The paper also makes a clear, falsifiable prediction: sub-picosecond hot-electron cooling in TiN, much faster than in Au. The main weakness is the use of an unbenchmarked virtual-crystal approximation for the defect calculations, which is load-bearing for the claimed defect-engineering trend.","major_comments":[{"comment":"The defect-engineering conclusion is not established because the VCA treatment of vacancies is unbenchmarked and the paper itself concedes that 'the VCA does not capture all defect-induced relaxation channels and therefore the calculated relaxation times should be interpreted as upper bounds.' The reported 215 fs maximum for TiN0.84 is thus an upper bound, and the true value could be close to or below the pristine 150 fs value. The statement in the Conclusions that 'introducing nitrogen vacancies increases the maximum thermalization time to 0.215 ps' overinterprets an upper bound as an established increase. To support the central defect trend, the authors should provide at least one explicit supercell calculation for a defective TiN system (e.g., TiN0.84 with a real vacancy) or otherwise benchmark the VCA for electron-phonon coupling against a method that captures localized defect states.","section":"Computational details, second paragraph; Results, Fig. 5; Conclusions"},{"comment":"Equation (7) as written contains an error: the electron-phonon coupling factor should involve the derivative of the Fermi-Dirac distribution with respect to energy, -∂f/∂ε, not with respect to temperature, ∂f/∂Te. With ∂f/∂Te, the integrand acquires an additional factor (ε - ε_F)/Te, which for a smooth density of states makes G vanish linearly as Te→0, contradicting the low-temperature constancy of G stated after Eq. (8) and shown in Fig. 5. The standard Allen-type expression used in the literature (e.g., Brown et al., PRB 94, 075120) has -∂f/∂ε. Please correct Eq. (7) and confirm that the actual numerical implementation uses the corrected expression.","section":"Methods, Eq. (7)"}],"minor_comments":[{"comment":"The abstract states that the largest thermalization times in TiN with nitrogen vacancies occur 'for electron temperatures around 4000 K', but the Results and Fig. 5 report the maximum for TiN0.84 at about 5000 K; please make these consistent.","section":"Abstract"},{"comment":"There is a missing space in 'Ithasbeenused' in the first paragraph; please correct this and perform a general pass for similar typographical errors.","section":"Introduction"},{"comment":"The caption of Fig. 4 appears to contain duplicated axis tick labels ('0 1 2 3 4 5 6 7 8' repeated four times), which makes it unclear which panel corresponds to Cel(T) and which to Cph(T); please clarify the panel layout.","section":"Figure 4"},{"comment":"The sentence following Eq. (7) says 'As a result, G is independent of the phonon temperature,' but the expression shown contains only Te; this is consistent, but the notation G(Te) should be used consistently throughout, including in Eqs. (1) and (2), where G(Te,Tp) appears.","section":"Methods, Eq. (7)"},{"comment":"The VCA implementation for nitrogen vacancies is not described in sufficient detail (e.g., how the missing atom is represented in the pseudopotential and how the valence electron count is adjusted). A short description would improve reproducibility.","section":"Computational details"}],"recommendation":"major_revision","confidential_remarks":"The paper's core comparison between pristine TiN and Au is well supported and likely correct. However, the defect-trend claim, which is a headline conclusion, rests on an unbenchmarked VCA that the authors themselves state gives only upper bounds. In addition, Eq. (7) has a likely typo in the derivative of the Fermi function that should be corrected and verified. These are fixable with additional calculations and a careful revision, so I recommend major revision rather than rejection. I would also note that the abstract and conclusions should be tempered to reflect the upper-bound nature of the vacancy result unless the benchmark is provided."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a well-executed, honest computational paper. The thing to know: the central result — hot-electron thermalization in TiN is sub-picosecond and an order of magnitude faster than in gold — is solid and worth citing. The secondary claim, that nitrogen vacancies raise the maximum thermalization time to 215 fs, is not established; it rests entirely on the virtual crystal approximation, and the paper admits that VCA misses defect-induced relaxation channels, so the 215 fs is an upper bound that could easily drop below the pristine 150 fs if real vacancies add scattering.\n\nWhat's new: although Habib et al. looked at defect-free nitrides, this is the first systematic look at how nitrogen vacancies and oxygen substitutionals change the electron-phonon coupling and thermalization times in TiN. The workflow is standard — Allen's formula, two-temperature model, all parameters from DFT — and the gold benchmark is reassuring. The comparison of phonon linewidths across systems is a nice way to identify the relevant relaxation channels. They also state the VCA limitation explicitly in the Computational details, which I appreciate.\n\nSoft spots, in proportion: the main one is the VCA issue. A 16% vacancy concentration is a lot; VCA replaces vacancies with an average atom and cannot create localized states or the extra scattering channels of a real defective lattice. They don't benchmark VCA against any explicit supercell calculation, and the claimed increase from 150 to 215 fs is small compared to the likely error of the method. So the defect-engineering message should be read as a tentative prediction, not a result. The other issues are minor: the claim that Eq. (2) was verified numerically isn't shown, and there are no error bars on any of the defect trends. The gold validation doesn't cover the VCA part, so it doesn't rescue the defect numbers.\n\nWho this is for: anyone working on hot-carrier plasmonics, ultrafast dynamics in transition metal nitrides, or e-ph coupling in defective metals. The pristine TiN result and the methodological setup deserve a serious referee; the defect part needs a follow-up with supercell calculations before it's quoted quantitatively.\n\nRecommendation: send it to peer review. It's an honest, carefully done study with a clear new contribution and a clearly disclosed limitation. A referee should ask for a VCA benchmark on at least one stoichiometry, or a softer framing of the defect predictions, but the paper is not a desk reject.","headline":"Solid, honest computational study; the pristine TiN sub-picosecond thermalization result is robust, but the VCA-based defect trend (215 fs) is an unbenchmarked upper bound and shouldn't be quoted as established.","tokens_in":12250,"tokens_out":2628,"would_cite":true,"duration_ms":26191,"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":"Titanium nitride thermalizes hot electrons in under a picosecond, about ten times faster than gold.","keywords":["hot electrons","two-temperature model","titanium nitride","electron-phonon interaction","virtual crystal approximation","density functional theory","thermalization time","defect engineering"],"falsifier":"Pump-probe measurements on TiN films with independently characterized nitrogen-vacancy concentrations would settle the claim: if observed electron thermalization times exceed the predicted sub-picosecond range, or if increasing vacancy concentration shortens rather than lengthens the times, the central result fails. A second check is to compute electron-phonon coupling in an explicit supercell containing a real nitrogen vacancy and compare it with the virtual-crystal prediction.","tokens_in":11310,"feed_emoji":"⚡","tokens_out":5505,"duration_ms":52282,"temperature":0.7,"pith_summary":"This paper sets out to determine how quickly photoexcited electrons in titanium nitride give up their heat to the lattice, and whether the nitrogen vacancies and oxygen substitutionals that real TiN samples contain change that speed. Working from density-functional-theory inputs fed into a two-temperature model, it concludes that electron-phonon thermalization in TiN is sub-picosecond: the maximum is about 0.15 ps for defect-free TiN, and 0.215 ps when nitrogen vacancies are introduced, compared with 10.2 ps for gold. A sympathetic reader would take the central claim to be that TiN's strong electron-phonon coupling makes it a fast hot-carrier material whose thermalization can be tuned by defect engineering.","feed_headline":"Hot electrons in titanium nitride cool in under a picosecond","feed_subtitle":"Defect-free TiN relaxes in ~0.15 ps; nitrogen vacancies stretch that to 0.215 ps.","key_machinery":"The central object is the two-temperature model with all parameters computed from density-functional theory, specifically the electron-phonon coupling factor $G(T_e)$, which enters the thermalization time $\tau_{\\mathrm{ep}}$ through $1/\\tau_{\\mathrm{ep}} = G(1/C_e + 1/C_p)$. $G$ is evaluated from the second moment of the Eliashberg function, $\\lambda\\langle\\omega^2\\rangle$, and at low $T_e$ reduces to Allen's formula $\\tau_{\\mathrm{ep}} = \\pi k_{\\mathrm{B}}T_e / (3\\hbar\\lambda\\langle\\omega^2\\rangle)$. Defective TiN is treated by the virtual crystal approximation, which replaces the defect with a fictive atom of averaged properties. The machinery does three jobs: it produces the materials-specific $C_e$, $C_p$, and $G$, it connects phonon linewidths to the dominant scattering channels, and it makes the defect trend in thermalization times computable.","core_discovery":"The paper's central claim is that electron-phonon coupling in titanium nitride is strong enough to thermalize hot carriers within a few hundred femtoseconds, an order of magnitude faster than in gold, and that nitrogen vacancies weaken the coupling just enough to push the maximum thermalization time from 0.15 ps to 0.215 ps. The result is obtained by computing the electron and phonon densities of states, the Eliashberg function and its second moment, and the electron-phonon coupling parameter $G$ from first principles, then solving the two-temperature model, using the virtual crystal approximation for the defective systems. Oxygen substitutional defects are found to leave the thermalization times nearly unchanged, whereas nitrogen vacancies reduce the Fermi-level density of states and the coupling strength, lengthening the time. The paper also connects phonon linewidths to relaxation channels: acoustic modes near $W$ and $L$ and optical modes throughout the Brillouin zone dominate the coupling in TiN.","pith_inferences":["Because the paper itself notes that the virtual crystal approximation omits some defect-induced relaxation channels, the quoted times are upper bounds; real nitrogen-vacancy TiN may thermalize even faster than 0.215 ps.","The correlation between $g(E_F)$, $\\lambda\\langle\\omega^2\\rangle$, and $\\tau_{\\mathrm{ep}}$ suggests that any disorder shifting the Fermi level in TiN, not just vacancies, will move thermalization times.","The same first-principles two-temperature-model pipeline could be applied to other non-stoichiometric refractory nitrides and carbides to screen for fast hot-carrier coolers."],"forward_implications":["Hot-carrier devices made from TiN will have to harvest or lose electron energy on a sub-picosecond timescale set by electron-phonon coupling, not by diffusion.","Varying the nitrogen-vacancy concentration tunes the maximum thermalization time between 0.15 ps and 0.215 ps, while oxygen substitution is a weak lever.","Pump-probe experiments on TiN should see thermalization signatures about ten times faster than on gold at comparable electron temperatures.","The computed phonon linewidths identify which vibrational modes dominate electron cooling, giving a target for defect engineering."],"supporting_citations":[{"why":"Supplies the Allen formula for $\\tau_{\\mathrm{ep}}$ and the two-temperature-model reduction used to convert coupling strength into a thermalization time.","marker":"[45]"},{"why":"Provides the Green's-function formalism behind the Eliashberg function, its moments, and the phonon linewidth expressions.","marker":"[56]"},{"why":"Earlier study of electron-phonon interactions in defect-free transition metal nitrides that this paper extends to defective TiN.","marker":"[32]"},{"why":"Source of the ab initio methodology for electron-phonon coupling and hot-electron response in plasmonic metals, supplying the gold comparison baseline.","marker":"[51]"},{"why":"Introduces the virtual crystal approximation used to model nitrogen vacancies and oxygen substitutionals.","marker":"[33]"},{"why":"Pump-probe measurement of electron thermalization in gold used to validate the model's timescale against experiment.","marker":"[66]"},{"why":"Further gold pump-probe data giving the 1-10 ps thermalization range used as the comparison benchmark.","marker":"[67]"}],"fun_headline_variants":["TiN hot electrons thermalize 10x faster than gold","Defect-free TiN cools hot electrons in 150 femtoseconds","Nitrogen vacancies slow TiN hot-electron cooling to 215 fs","TiN hot carriers cool in 150 fs, 215 fs with vacancies","First-principles model predicts fast hot-carrier thermalization in TiN"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the virtual crystal approximation faithfully represents how nitrogen vacancies and oxygen substitutionals alter electron-phonon coupling, even though real vacancies may create localized states and extra scattering channels that an averaged virtual atom cannot capture.","fun_headline_variants_meta":{"raw":{"variants":["TiN hot electrons thermalize 10x faster than gold","Defect-free TiN cools hot electrons in 150 femtoseconds","Nitrogen vacancies slow TiN hot-electron cooling to 215 fs","TiN hot carriers cool in 150 fs, 215 fs with vacancies","First-principles model predicts fast hot-carrier thermalization in TiN"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000798,"raw_usage":{"total_tokens":3458,"prompt_tokens":842,"completion_tokens":2616,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":458,"completion_tokens_details":{"reasoning_tokens":2523}},"tokens_in":458,"tokens_out":2616,"duration_ms":17949,"temperature":1.0,"reasoning_tokens":2523,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:38:45.819706+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Pump-probe measurements on TiN films with independently characterized nitrogen-vacancy concentrations would settle the claim: if observed electron thermalization times exceed the predicted sub-picosecond range, or if increasing vacancy concentration shortens rather than lengthens the times, the central result fails. A second check is to compute electron-phonon coupling in an explicit supercell containing a real nitrogen vacancy and compare it with the virtual-crystal prediction.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the Allen formula for $\\tau_{\\mathrm{ep}}$ and the two-temperature-model reduction used to convert coupling strength into a thermalization time."},{"cited_title":"Electron-phonon interactions from first principles","cited_arxiv_id":null,"evidence_quote":"Provides the Green's-function formalism behind the Eliashberg function, its moments, and the phonon linewidth expressions."},{"cited_title":"Hot carrier dynamics in plasmonic transition metal nitrides","cited_arxiv_id":null,"evidence_quote":"Earlier study of electron-phonon interactions in defect-free transition metal nitrides that this paper extends to defective TiN."},{"cited_title":"M.; Sundararaman, R.; Narang, P.; Goddard, W","cited_arxiv_id":null,"evidence_quote":"Source of the ab initio methodology for electron-phonon coupling and hot-electron response in plasmonic metals, supplying the gold comparison baseline."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces the virtual crystal approximation used to model nitrogen vacancies and oxygen substitutionals."},{"cited_title":"Vallée,; Acioli, L","cited_arxiv_id":null,"evidence_quote":"Pump-probe measurement of electron thermalization in gold used to validate the model's timescale against experiment."},{"cited_title":"E.; Juhasz, T","cited_arxiv_id":null,"evidence_quote":"Further gold pump-probe data giving the 1-10 ps thermalization range used as the comparison benchmark."}],"review_version":1}