{"id":"381118be-2266-4356-be80-74a5532a7ce2","arxiv_id":"2608.04219","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Including GW corrections to both band structure and electron-phonon couplings in Boltzmann transport yields electron mobilities within about 11% of experiment for Si, GaAs, GaP, diamond, and SiC.","lead":"A new calculation method that adds many-body GW corrections to both the electronic bands and the electron-phonon couplings used in Boltzmann transport predicts electron mobilities of five benchmark semiconductors within 11% of experiment on average. The result suggests that a long-standing DFT-level shortfall in mobility prediction can be overcome, which could speed up computational screening of materials for electronics.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Polar Fröhlich coupling and constant-screening approximation are left at DFPT level in two of five benchmark materials, so the 11% MARE for GaAs/GaP may rest on untested many-body corrections.","rationale":"The paper is a serious computational study with open data, convergence tests, jackknife stability, and a full-frequency check in diamond. I read the central claim as: GW corrections to both bands and electron-phonon couplings give 11% MARE for electron mobility in five benchmark semiconductors, roughly four times better than DFT+DFPT. The most load-bearing condition is that the electron-phonon couplings used in the mobility calculation are actually the GW-corrected ones for the scattering processes that dominate. This condition is not fully met for GaAs and GaP: the long-range Fröhlich part, which dominates transport in these polar materials, remains at the DFPT level, and the variation of W is neglected. The paper explicitly acknowledges both limitations in the Supplemental Methods. The 11% figure therefore rests on an assumption about the long-range vertex that is plausible but untested in this work. I do not see a fatal flaw: the GWPT framework is established, the short-range corrections are demonstrated, and the authors are transparent about the long-range treatment. However, a conditional verdict remains appropriate, requiring either a long-range GWPT test or an explicit statement that the 11% claim is benchmark-specific and does not yet cover the dominant polar coupling. This is the same concern the reader identified, hence 'agree.'","tokens_in":20721,"tokens_out":5245,"duration_ms":49393,"concrete_test":"Using the GWPT long-range Fröhlich method of Zhu et al. (arXiv:2512.12479), recompute the electron-phonon matrix elements for GaAs and GaP with the same GWPT settings, then rerun the GWBTE mobility at 300 K with all other inputs unchanged. If either mobility shifts by more than about 10% relative to the DFPT-long-range result, the central 11% MARE claim is not robust for polar materials. A cheaper preliminary check is to compare GWPT matrix elements with and without the constant-screening approximation at the q→Γ Fröhlich point in GaAs.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim requires that GW corrections be applied to both the band structure and the electron-phonon couplings. The band-structure part is performed, but the electron-phonon part is incomplete for polar materials: the Supplemental Methods state that long-range dipole and quadrupole corrections are applied 'at the DFPT level,' and that a recently developed GWPT Fröhlich approach would 'further improve the accuracy of interpolation for polar materials.' In addition, the GWPT workflow neglects the variation of the screened Coulomb interaction (constant-screening approximation, ΔW=0), citing prior validation. For GaAs and GaP, Fröhlich scattering is the dominant electron-phonon mechanism at 300 K, so their mobilities are controlled by precisely the matrix elements that are not GW-corrected. The 11% MARE could therefore reflect a cancellation between GWPT-corrected short-range couplings and unconverged long-range couplings. The paper reports no per-material test of the constant-screening approximation or of the DFPT long-range vertex against a GWPT long-range reference. This is not an internal inconsistency, but it is a gap between the advertised 'fully many-body' method and the calculation actually performed for two of the five benchmark semiconductors.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This Letter reports GWBTE, a first-principles workflow that combines G0W0 quasiparticle band structures with GW perturbation theory (GWPT) electron-phonon matrix elements and solves the ab initio Boltzmann transport equation (aiBTE). The authors benchmark room-temperature electron mobilities for five cubic semiconductors (Si, GaAs, GaP, diamond, 3C-SiC), reporting a mean absolute relative error (MARE) of 11% with respect to high-quality experimental references, compared with 106% for DFT+DFPT and 42% for GW+DFPT. They also present a temperature-dependent comparison against 66 experimental measurements (24% MARE), a Drude-like decomposition that separates band-structure and electron-phonon coupling contributions, and extensive convergence tests. The central claim is that many-body corrections to both band structure and electron-phonon couplings are needed for predictive mobility calculations, with GW+DFPT being potentially worse than DFT+DFPT for some materials.","tokens_in":20774,"tokens_out":5308,"duration_ms":48444,"significance":"If the claims hold, this is a notable advance: a parameter-free, fully ab initio method reaching 11% MARE on the tested semiconductors, roughly four times more accurate than the common GW-band-only scheme. The paper's strengths include extensive numerical convergence checks (Supplemental Figs. S3, Tabs. S8-S9), a full-frequency GW comparison in diamond (Tab. S6), jackknife cross-validation (Tab. S3), and open data and code availability. The main caveat is that the benchmark set is small (five materials) and that the many-body treatment of the electron-phonon coupling is incomplete for polar materials: for GaAs and GaP, the long-range Fröhlich part is still interpolated at the DFPT level, and the constant-screening approximation is used without a per-material test. These limitations do not invalidate the work, but they should be addressed before the 'record accuracy' claim can be accepted as stated.","major_comments":[{"comment":"The claim that the 11% MARE arises from GW corrections to both the band structure and the electron-phonon couplings is not fully supported for GaAs and GaP. The Supplemental Methods state that long-range dipole and quadrupole corrections are applied at the DFPT level, and that the GWPT workflow neglects the variation of the screened Coulomb interaction (constant-screening approximation). Since Fröhlich scattering dominates transport in these polar materials at 300 K, the mobilities of two of the five benchmark materials are controlled by matrix elements that are not GW-corrected. The paper provides no quantitative test of the constant-screening approximation or of the DFPT long-range vertex against a GWPT reference for these materials. I request either adding such tests (e.g., using the GWPT Fröhlich approach cited as Ref. 15) or substantially qualifying the abstract and main-text claims to state that the long-range Fröhlich coupling is treated at the DFPT level.","section":"Supplemental Methods (Wannier-Fourier interpolation; GW perturbation theory calculations)"},{"comment":"The headline MARE of 11% is computed on only five materials, all cubic semiconductors with similar bonding and transport characteristics. The jackknife analysis shows robustness within this set, but it does not establish that 'record accuracy' holds more generally. The authors already caution about statistical accuracy, but the abstract and title still assert a general record. I recommend either expanding the benchmark set or explicitly restricting the claim to 'the five semiconductors studied here.'","section":"Fig. 2(a) and Table S3"},{"comment":"The Drude-like decomposition used to attribute mobility corrections to band-structure versus electron-phonon effects is performed in the SERTA approximation, but Table S5 shows that SERTA deviates strongly from the full iterative BTE for GaAs (5924 cm2/Vs versus 9680 cm2/Vs, a 39% discrepancy). The conclusion that GaAs's mobility improvement is driven mainly by band-structure corrections may therefore not carry over to the full BTE results. Please either provide the corresponding decomposition using the iterative BTE mobilities or explicitly discuss the SERTA/BTE discrepancy for GaAs when interpreting Fig. 3.","section":"Table S5 and Fig. 3(g)-(i)"}],"minor_comments":[{"comment":"The caption states that all samples have doping concentration ≤10^16 cm^-3, but Table S1 lists diamond at 7×10^16 cm^-3; please reconcile these values.","section":"Fig. 2(a) caption"},{"comment":"The quantity ρ in Eq. (1) is called the 'average density of states,' but the Supplemental Methods define it as a scattering density of states (Eq. S9) that depends on phonon occupations and delta functions; this should be clarified in the main text to avoid confusion.","section":"Eq. (1) and surrounding text"},{"comment":"There are several typos, including 'Monkhort-Pack' instead of 'Monkhorst-Pack' and 'bwteen' instead of 'between' in the Fig. S3 caption; please proofread.","section":"Supplemental Methods and Figure S3 caption"},{"comment":"The statement that the constant-screening approximation 'has been widely tested to be valid' cites Ref. 23, but no test specific to the present materials is provided; a short benchmark or a clearer justification of transferability to GaAs and GaP would strengthen the claim.","section":"Supplemental Methods (GW perturbation theory calculations)"}],"recommendation":"major_revision","confidential_remarks":"The paper is technically strong and likely to be influential if the polar-material limitation is resolved or clearly delineated. The main obstacle is that two of the five benchmark materials do not receive GW corrections to their dominant electron-phonon coupling channel, so the 'fully many-body' phrasing in the abstract overstates what was computed. The small benchmark set also makes the 'record accuracy' claim premature. I would be willing to reconsider after a revision that adds a quantitative sensitivity test or softens the claims accordingly."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe headline result is real: combining GWPT vertex corrections with the ab initio BTE cuts the mean absolute relative error in electron mobilities from ~106% (DFT) to ~11% across five cubic semiconductors, and the finding survives jackknife. What's new is not GWPT itself (Li et al. 2019) but its use inside BTE mobility with Wannier interpolation, plus the useful counterintuitive observation that GW bands with DFPT couplings can worsen agreement for diamond, SiC, and GaP. That claim is solid, and the paper is careful: convergence extrapolation, full-frequency checks for diamond, band-cutoff tests, open data and code patches, SERTA-only decomposition for analysis. The self-citation is appropriate; GWPT is the correct method reference.\n\nThe soft spots, in order of importance. First, the 'fully many-body' label overstates the polar case. The supplemental methods state plainly that long-range dipole/quadrupole corrections remain at DFPT level and that a recently developed GWPT Fröhlich approach would improve interpolation for polar materials; the constant-screening (ΔW=0) approximation is inherited from prior work without per-material verification. Since Fröhlich scattering dominates GaAs and GaP at 300 K, the 11% for those two materials rests on precisely the matrix elements that aren't GW-corrected. This is a gap between the advertising and the calculation, not a fatal flaw—the paper is open about it—but the reader should not take 11% as a verified statement about fully many-body polar transport.\n\nSecond, the 11% number is fragile: five materials, one experimental 'most accurate' reference per material, diamond being the sensitive case (the 7e16 cm^-3 sample in Table S1 violates the caption's <=1e16 filter; a caption fix is needed). The broader 66-point dataset gives 24% MARE, which is more honest but still good. Third, the band-structure-only GW results are worse for some materials—presented as a caution about incomplete corrections, which is fine.\n\nThe stress-test note I was given lands: the polar gap is real and should be addressed with either a GWPT Fröhlich calculation or explicit disclaimers when quoting 11%. The reader's verdict of conditional is about right; I don't think the main argument is wrong, but the headline needs qualification.\n\nWho is it for? People doing ab initio transport or materials screening. It deserves a serious referee—the computation is formidable, the data/code are open, and the result will be a reference point regardless of whether 11% survives expansion to more materials. Send it out.\n\nBest,","headline":"Genuinely new and mostly delivers: GWPT vertex corrections in the BTE cut mobility errors to ~11% on five benchmarks, though the 'fully many-body' label overstates the polar case and the headline number is a bit friendlier than the evidence warrants.","tokens_in":21473,"tokens_out":3749,"would_cite":true,"duration_ms":26385,"reading_group":"yes","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 demonstrates that correcting electron-phonon couplings with GW perturbation theory, alongside GW band structures, yields electron mobilities within 11% of experiment for Si, GaAs, GaP, diamond, and 3C-SiC.","keywords":["carrier mobility","GW perturbation theory","electron-phonon coupling","Boltzmann transport equation","ab initio transport","semiconductors","many-body perturbation theory","first-principles prediction"],"falsifier":"Recomputing the electron mobility of GaAs and GaP with the missing W-variation and a GW-level long-range Fröhlich coupling would settle the matter: if the mean error across the five benchmark materials rises above 20% instead of staying near 11%, the central claim fails.","tokens_in":20369,"feed_emoji":"⚡","tokens_out":10764,"duration_ms":79801,"temperature":0.7,"pith_summary":"The paper claims that the dominant error in first-principles carrier mobility calculations is not just the band structure but the electron-phonon coupling, which density functional theory overscreens. Applying many-body GW corrections to both quantities and solving the Boltzmann transport equation gives electron mobilities within 11% of experiment on average for Si, GaAs, GaP, diamond, and 3C-SiC at room temperature. Correcting only the bands with GW leaves the average error at 42%, and can even worsen agreement for diamond, SiC, and GaP. If correct, this lifts a long-standing limitation of DFT-based transport and makes predictive mobility screening practical.","feed_headline":"GW fixes electron-phonon coupling, cutting mobility error to 11%","feed_subtitle":"Correcting band structure and electron-phonon coupling with GW theory beats DFT by four times in five semiconductors.","key_machinery":"The load-bearing object is the GWPT electron-phonon matrix element, $g^{\\mathrm{GW}}_{mn\\nu}(\\mathbf{k},\\mathbf{q}) = g^{\\mathrm{DFT}}_{mn\\nu}(\\mathbf{k},\\mathbf{q}) + \\langle \\psi_{m\\mathbf{k}+\\mathbf{q}} | \\Delta_{\\mathbf{q}\\nu}(\\Sigma^{\\mathrm{GW}} - V_{\\mathrm{xc}}) | \\psi_{n\\mathbf{k}} \\rangle$, which replaces the DFT exchange-correlation potential variation by the variation of the GW self-energy and thereby corrects the overscreening of the bare vertex. The argument is carried by the exact Drude-like rewriting of the mobility, $\\mu = e\\tau/m^*$ with $1/\\tau = (2\\pi/\\hbar) g^2 \\rho$, which separates band-structure renormalization ($m^*$ and $\\rho$) from coupling renormalization ($g$) and shows that the $g^2$ enhancement dominates the error reduction in most materials. Wannier interpolation of the GW-corrected quantities onto ultra-dense Brillouin-zone grids is what makes the fully many-body Boltzmann solution computationally feasible.","core_discovery":"The central discovery is that GW self-energy effects in the electron-phonon vertex, rather than band-structure corrections alone, are required to bring computed mobilities into agreement with high-purity measurements. The GWPT-corrected matrix element $g^{\\mathrm{GW}} = g^{\\mathrm{DFT}} + \\langle \\psi | \\Delta(\\Sigma^{\\mathrm{GW}} - V_{\\mathrm{xc}}) | \\psi \\rangle$ increases the coupling strength by up to 80% for band-edge states (43% on average in GaP), correcting the overscreening that makes DFT couplings too weak. With these couplings and GW bands in the fully iterative Boltzmann equation, the mean absolute relative error at 300 K drops from 106% (DFT), or 42% (GW bands only), to 11% across the five benchmark crystals. A hybrid scheme that keeps DFPT couplings with GW bands is not sufficient and, for diamond, SiC, and GaP, it moves the answer further from experiment.","pith_inferences":["If the 11% accuracy survives a broader dataset, first-principles mobility could replace empirical mobility models in semiconductor device simulation and materials screening.","The wavevector-dependent spread of the GWPT corrections, spanning a factor of two around the average, indicates that simple scalar rescaling of DFPT couplings cannot capture many-body effects; this indirectly motivates machine-learned vertex corrections for larger unit cells.","A natural test is to apply the protocol to a strongly polar or low-dimensional semiconductor (e.g., GaN or a transition-metal dichalcogenide monolayer), where the constant-screening and DFPT long-range approximations are most stressed; degradation of the 11% error would identify the next approximation to fix."],"forward_implications":["The common practice of combining GW bands with DFPT electron-phonon couplings should be abandoned; it can worsen agreement and is never as accurate as the fully corrected scheme.","The 11% mean error over Si, GaAs, GaP, diamond, and 3C-SiC provides a benchmark that future first-principles transport methods should be measured against.","The same GWBTE workflow can be applied to other weakly-to-moderately correlated semiconductors, where phonon-limited scattering governs room-temperature mobility.","For polar materials, upgrading the long-range Fröhlich coupling from DFPT to GWPT is the most likely next step to preserve the accuracy outside the current benchmark set.","The material-specific decomposition means that no single correction applies everywhere: GaAs needs band renormalization, while SiC, GaP, and diamond need vertex renormalization, and silicon needs neither beyond DFT."],"supporting_citations":[{"why":"Introduces the GWPT method for correcting electron-phonon matrix elements with the variation of the GW self-energy.","marker":"[17]"},{"why":"Provides the computational implementation of GWPT used for the vertex corrections in this work.","marker":"[30]"},{"why":"Supplies the ab initio Boltzmann transport framework with Wannier interpolation used to evaluate scattering rates and solve the BTE.","marker":"[7]"},{"why":"Gives the G0W0 approximation with the generalized plasmon-pole model used for quasiparticle band energies.","marker":"[12]"},{"why":"Provides the iterative Jacobi solver for the Boltzmann equation and the prior GW-band mobility scheme that this work extends.","marker":"[15]"},{"why":"Sets out density-functional perturbation theory, the starting point for the phonons and electron-phonon couplings that GWPT corrects.","marker":"[25]"},{"why":"Documents the large errors of DFT-based mobility predictions, the motivation for adding many-body corrections.","marker":"[11]"},{"why":"Defines the Fan-Migdal self-energy and electron-phonon matrix elements formalism on which the transport calculations rest.","marker":"[1]"},{"why":"Supplies the reference experimental electron mobility for silicon used in the 11% benchmark.","marker":"[31]"},{"why":"Supplies the reference experimental electron mobility for GaAs used in the benchmark.","marker":"[32]"}],"fun_headline_variants":["GW vertex correction cuts mobility error to 11% across five semiconductors","Full GW corrections bring mobility error to 11% for five crystals","Electron-phonon GW correction is key to accurate mobility predictions","GW-corrected electron-phonon coupling cuts mobility error to 11%","Vertex GW correction, not just bands, yields 11% mobility error"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The calculation assumes that the screening of the electron-phonon interaction stays frozen when atoms move (the constant-screening approximation) and, for the polar materials GaAs and GaP, the long-range Fröhlich interaction is still taken from the DFT-level theory rather than from the many-body correction.","fun_headline_variants_meta":{"raw":{"variants":["GW vertex correction cuts mobility error to 11% across five semiconductors","Full GW corrections bring mobility error to 11% for five crystals","Electron-phonon GW correction is key to accurate mobility predictions","GW-corrected electron-phonon coupling cuts mobility error to 11%","Vertex GW correction, not just bands, yields 11% mobility error"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001548,"raw_usage":{"total_tokens":6168,"prompt_tokens":899,"completion_tokens":5269,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":515,"completion_tokens_details":{"reasoning_tokens":5174}},"tokens_in":515,"tokens_out":5269,"duration_ms":30682,"temperature":1.0,"reasoning_tokens":5174,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-08T00:13:25.670974+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Recomputing the electron mobility of GaAs and GaP with the missing W-variation and a GW-level long-range Fröhlich coupling would settle the matter: if the mean error across the five benchmark materials rises above 20% instead of staying near 11%, the central claim fails.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the computational implementation of GWPT used for the vertex corrections in this work."},{"cited_title":"Baroni, S","cited_arxiv_id":null,"evidence_quote":"Sets out density-functional perturbation theory, the starting point for the phonons and electron-phonon couplings that GWPT corrects."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the reference experimental electron mobility for silicon used in the 11% benchmark."}],"review_version":1}