{"id":"d27aca64-ad63-4252-906a-27209fb7fb78","arxiv_id":"1909.01449","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"First-principles calculations predict that tensile strain raises in-plane electron and hole mobilities in boron arsenide by 60 to 80 percent, and that boron arsenide forms nearly lattice-matched type-II junctions with InGaN and ZnSnN2.","lead":"This paper uses density functional theory to predict how squeezing or stretching the semiconductor boron arsenide changes its band gap, band offsets, and carrier mobility. It proposes that boron arsenide can serve as a lattice-matched, heat-spreading substrate for InGaN and ZnSnN2 optoelectronic devices.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The claimed 68% hole-mobility increase compares a strained no-SOC calculation to an unstrained SOC-included literature baseline, so the headline number is not a same-framework comparison.","rationale":"The reader's formal weakest_assumption concerned band-offset alignment across different calculation frameworks, which is a legitimate concern about the heterojunction claims. I agree that concern exists, but I see the hole-mobility comparison as more directly load-bearing because it affects a headline quantitative claim in the abstract and conclusion and involves an internal inconsistency in baseline choice. The reader did flag this issue as condition (1) in the rationale, so there is partial overlap. The underlying physics of strain-enhanced mobility is supported by the electron channel, where the 80% increase is computed consistently within the same no-SOC framework and agrees with Liu et al. for the unstrained electron mobility. The hole channel has a plausible mechanism, but the claimed 68% is obtained by mixing SOC and no-SOC results. Because this can be settled by one additional calculation, the appropriate verdict remains CONDITIONAL, and no change to the reader's verdict is needed beyond making this check an explicit condition.","tokens_in":11393,"tokens_out":4950,"duration_ms":49763,"concrete_test":"Recompute the 1% tensile-strained in-plane hole mobility with spin-orbit coupling included in the EPW/Wannier calculation, using the same 6x6x6 phonon grid and the same dense BZ extrapolation as the no-SOC strained run. Compare the resulting value to (a) a recomputed unstrained SOC hole mobility in the same code and (b) the no-SOC strained value of 3550 cm2/Vs. If the SOC-corrected strained mobility is close to 3550, the issue is mainly which baseline is quoted; if it differs by more than ~10%, the headline strain-enhancement number should be revised.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The least secure part of the central claim is the hole-mobility enhancement, because the strained and unstrained numbers are not computed on the same footing. Section 3.2 reports the authors' own unstrained hole mobility as 1387 cm2/Vs from their no-SOC EPW calculation, and notes that the literature SOC value from Liu et al. is 2110 cm2/Vs, i.e., exactly 1.5 times larger. For 1% tensile strain, the same no-SOC framework gives an in-plane hole mobility of 3550 cm2/Vs. The abstract and conclusion quote a '68% increase' by comparing 3550 to 2110, the SOC-included unstrained literature value. The only same-framework comparison, 3550 versus 1387, gives a ~156% increase. The authors argue in Section 3.2 that SOC should not strongly affect the strained hole mobility because the strained VBM is non-degenerate with or without SOC (Figure 3(e)-(f)), but this is an inference, not a calculation; spin-orbit coupling can renormalize effective masses and electron-phonon matrix elements even when the degeneracy is already lifted. The qualitative direction of the effect is plausible, but the quantitative 68% figure depends on an unverified cross-framework normalization and could misstate the magnitude of the strain enhancement.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript uses hybrid-functional DFT and EPW electron-phonon calculations to study biaxially strained boron arsenide (BAs), reporting that tensile strain reduces the band gap, increases in-plane electron and hole mobilities (to 2417 and 3550 cm2/Vs at 1% strain), and that BAs forms nearly lattice-matched, type-II heterojunctions with InGaN and ZnSnN2. The authors propose BAs as both a strain-engineered thin-film channel material and a high-thermal-conductivity substrate for optoelectronic junctions.","tokens_in":11607,"tokens_out":3100,"duration_ms":32142,"significance":"If the quantitative claims hold, this work would provide a useful predictive framework for BAs heterostructures and identify specific device-relevant interfaces. The methodological strengths are the converged EPW mobility calculations with Brillouin-zone extrapolation, the direct comparison to prior electron-mobility results from Liu et al., and the physically clear mechanism attributed to valley and valence-band degeneracy lifting. The band-alignment section also addresses an important gap in the literature. However, the central hole-mobility enhancement claim rests on a cross-framework comparison (no-SOC strained vs. SOC-included unstrained), and the band offsets are assembled from three different computational frameworks, so the quantitative conclusions need revision or careful qualification.","major_comments":[{"comment":"The claimed 68% hole-mobility increase is not a same-framework comparison. The paper's own unstrained hole mobility, computed without spin-orbit coupling, is 1387 cm2/Vs, and the strained in-plane value is 3550 cm2/Vs, which is a 156% increase. The 68% figure instead compares 3550 cm2/Vs to the SOC-included unstrained literature value 2110 cm2/Vs from Liu et al. The argument in Section 3.2 that SOC should not strongly affect the strained hole mobility because the strained VBM is non-degenerate with or without SOC (Figure 3(e)-(f)) is an inference, not a quantitative calculation; SOC can renormalize effective masses and electron-phonon matrix elements even when degeneracy is already lifted. Because the abstract and conclusion advertise the 'more than 60%' improvement, this load-bearing number must either be computed with SOC in the strained case or be restated with explicit caveats about the different baselines.","section":"Section 3.2, Figure S5, and Conclusion"},{"comment":"The reported band offsets (0.42/0.30 eV for InGaN, 0.503/0.516 eV for ZnSnN2) are assembled from three different reference schemes: BAs is aligned to vacuum using an unrelaxed HSE slab, InGaN data are taken from ref. [47] without recomputation, and ZnSnN2 is aligned using an HSE mixing parameter of 31% fitted to the experimental gap. If the vacuum-level references are not mutually consistent, the offsets and the type-II classification could shift. The manuscript should either recompute InGaN in the same slab framework, test the sensitivity of the ZnSnN2 offsets to the mixing parameter, or explicitly quantify the expected uncertainty from these procedural differences. This point is central because the proposed heterojunction applications depend on the type-II alignment and the absolute offset magnitudes.","section":"Section 2 and Section 3.3"},{"comment":"The temperature-dependent mobility analysis compares strained no-SOC results with the SOC-included literature data of Liu et al. in Figure 4(a)-(b), and the text states that the power laws in the 200-500 K range are 'similar to the values found by Liu et al.' This is an additional instance of the same cross-framework issue: the unstrained no-SOC hole mobility differs from the SOC-included value by a factor of about 1.5, so any quantitative comparison of scattering mechanisms or power-law exponents should be made on a consistent footing. The qualitative conclusions about acoustic versus optical deformation potential may survive, but the current presentation conflates two different computational settings.","section":"Section 3.2, temperature dependence"}],"minor_comments":[{"comment":"The opening sentence 'BAs is III-V semiconductor' is missing an article; it should read 'BAs is a III-V semiconductor.'","section":"Abstract and Section 1"},{"comment":"The Supporting Information text contains apparent figure-reference mismatches: the ordered ZnSnN2 slab potential is said to be shown in 'Figure S4,' but Figure S4 is a band-structure plot, and the disorder-alignment discussion refers to 'Figure S5' for electrostatic potentials while Figure S5 is the mobility convergence plot. The numbering should be corrected.","section":"Supporting Information, Figure S4/S5 references"},{"comment":"The abscissa of Figure 5(b) mixes two different variables (In content x for InGaN and long-range order parameter S for ZnSnN2) on a single axis, which makes the plot difficult to read; a two-panel layout or separate axes would improve clarity.","section":"Figure 5(b)"},{"comment":"The mobility calculations use LDA while the band-structure and band-alignment calculations use HSE; the manuscript should state explicitly whether the strained geometries used in the two parts are consistent, since the strain relaxation procedure is described separately for each code.","section":"Section 2"}],"recommendation":"major_revision","confidential_remarks":"The core issue is the headline hole-mobility enhancement, whose 68% figure is obtained by comparing a no-SOC strained calculation with an SOC-included unstrained literature value. This is fixable by re-presenting the comparison or adding a strained SOC calculation, but as written it misstates the magnitude of the strain effect. The band-alignment section similarly needs a consistency check across the three frameworks. The manuscript is otherwise informative and likely publishable after these revisions."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper does something useful: it gives the first strain-dependent band structure, absolute band positions, and phonon-limited mobilities for BAs, then uses those numbers to propose two nearly lattice-matched, type-II heterojunctions (BAs/InGaN and BAs/ZnSnN2). The BAs part is mostly parameter-free DFT/EPW work, benchmarked against Liu et al., and the calculated unstrained electron mobility (1341 vs 1400 cm2/Vs) is a good sanity check. The qualitative mobility story is also sensible: tensile strain lowers the out-of-plane CBM valleys, the in-plane electron mass drops, and strain lifts the VBM degeneracy and removes hole scattering channels. I find the core physics credible.\n\nThe soft spot is the one the stress-test note flags. The abstract and conclusion present the 68% hole-mobility increase by comparing the strained no-SOC value (3550) to Liu et al.'s unstrained SOC-included value (2110). That is not a same-framework comparison. The authors' own unstrained no-SOC hole mobility is 1387, and compared to that the strain increase is about 156%, not 68%. They do argue that SOC should not strongly affect the strained hole mobility because the strained VBM is non-degenerate either way, which is a reasonable inference, but they have not calculated it. The headline number should either be recomputed with SOC for the strained structure or re-baselined against their own unstrained no-SOC value. This is a fixable presentation issue, not a sign the physics is wrong.\n\nThe band-alignment section has smaller soft spots. The InGaN offsets are taken from literature (Moses and Van de Walle) rather than recomputed in the same slab framework, and the ZnSnN2 offsets use an HSE mixing parameter fitted to the experimental gap. Any absolute offset carries a few tenths of an eV of uncertainty, and the paper should say so. The type-II classification is probably robust, but the precise 0.42/0.30 and 0.503/0.516 eV numbers are firmer than the methodology supports.\n\nWho benefits: computational materials scientists working on BAs, III-nitride alloys, or Zn-IV-N2 compounds, and experimentalists thinking about substrate choices for InGaN or ZnSnN2. It deserves serious refereeing. I would send it to review with a major-revision request focused on the hole-mobility comparison and on quantifying alignment uncertainty.","headline":"Useful predictive study of strained BAs and two heterojunctions, but the headline hole-mobility gain mixes a no-SOC strained calculation with an SOC-included unstrained baseline and should be reframed before publication.","tokens_in":12232,"tokens_out":2805,"would_cite":true,"duration_ms":27116,"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":"Boron arsenide, already prized for its thermal conductivity, is predicted to get faster and more useful under tensile strain: in-plane electron and hole mobilities rise to 2417 and 3550 cm²/Vs, and the material becomes a…","keywords":["boron arsenide","biaxial strain","carrier mobility","electron-phonon coupling","band alignment","type-II heterojunction","InGaN","ZnSnN2"],"falsifier":"Grow a BAs film on a substrate imposing about 1% biaxial tensile strain, measure the 300 K in-plane and out-of-plane electron and hole mobilities (predicted 2417/482 and 3550/2956 cm²/Vs), and check the predicted valley splitting (roughly 200 meV) and valence-band splitting (roughly 115 meV) by optical or angle-resolved photoemission measurements. If the mobilities do not increase and the band splittings are absent, the central claim fails. For the alignments, measure the valence-band offset at a BAs/InGaN interface by X-ray photoelectron spectroscopy; a value far from 0.30 eV would overturn the type-II prediction.","tokens_in":11124,"feed_emoji":"⚡","tokens_out":9049,"duration_ms":80414,"temperature":0.7,"pith_summary":"Boron arsenide (BAs) is usually studied for its record thermal conductivity; this paper asks what it can do electronically and as a heterostructure partner. Using hybrid-functional density functional theory and first-principles electron-phonon calculations, the authors argue that 1% biaxial tensile strain raises the in-plane electron mobility from 1341 to 2417 cm²/Vs and the hole mobility from 1387 to 3550 cm²/Vs at 300 K. The strain works by splitting the conduction-band valleys, so electrons move with a lighter transverse mass, and by splitting the valence-band maximum, so holes lose scattering channels. The same calculations place BAs as a nearly lattice-matched substrate for In-rich InGaN and ZnSnN2, with type-II band alignments that would separate photogenerated carriers. If these predictions hold, strained BAs films and BAs-based heterojunctions become concrete options for fast transistors, full-visible-spectrum LEDs, and solar cells.","feed_headline":"Stretching boron arsenide boosts carrier mobility by 60 percent","feed_subtitle":"1 percent tensile strain lifts band degeneracies, raising in-plane electron and hole mobilities to 2417 and 3550 cm2/Vs.","key_machinery":"The central object is the strain-split band edge of cubic BAs. In the unstrained crystal the six conduction minima near X are equivalent, so transport averages a light transverse mass and a heavy longitudinal mass; 1% biaxial strain lowers the out-of-plane pair of valleys by about 200 meV, so in-plane electrons carry only the lighter transverse mass ($m^* = 0.24$), while out-of-plane electrons see the heavier mass ($m^* = 1.09$). At the valence edge, strain removes the degeneracy of the top two bands by about 115 meV, cutting the hole interband scattering channels from two to one. For heterojunctions, the key instrument is the vacuum-level alignment of slab electrostatic potentials, applied directly to BAs and ZnSnN2 and taken from published hybrid-functional results for InGaN; that alignment produces the type-II offsets.","core_discovery":"The paper's central discovery is that tensile strain is a powerful lever on boron arsenide's electronic performance, not just its phonons. At 1% biaxial tensile strain, the in-plane electron mobility rises from 1341 to 2417 cm²/Vs and the in-plane hole mobility from 1387 to 3550 cm²/Vs at 300 K, an 80% and 68% increase, because strain splits the conduction-band valleys, so electrons move with a lighter transverse effective mass, and splits the valence-band maximum, so holes lose an interband scattering channel. The paper also finds that BAs is nearly lattice-matched to In-rich InGaN and to ZnSnN2, with type-II (staggered) band alignments: conduction- and valence-band offsets of 0.42/0.30 eV against In0.5Ga0.5N and 0.503/0.516 eV against ordered ZnSnN2. These numbers make p-BAs/n-InGaN and p-BAs/n-ZnSnN2 junctions plausible for carrier-separating optoelectronic devices on a substrate that also removes heat efficiently.","pith_inferences":["A direct experimental test is to measure the mobility anisotropy and the roughly 200 meV valley splitting of a strained BAs film; if the anisotropy and splitting do not appear together, the proposed mechanism, rather than the mobility numbers, would be the point of failure.","The same valley-splitting logic that works for BAs and strained silicon should apply to other cubic multi-valley semiconductors, so the calculation could be repeated for boron phosphide or boron antimonide to see whether strain is a general mobility lever in the boron pnictides.","The paper stops at intrinsic junctions; a natural extension is to model the actual space-charge region and doping levels at BAs/ZnSnN2 interfaces, where charged defects in ZnSnN2 may bend bands and shift the effective barrier from the ideal intrinsic offsets.","The InGaN offsets come from an external calculation rather than a fresh alignment in this paper's framework, so a direct photoemission measurement of the BAs/InGaN valence-band offset is the cleanest way to confirm or revise the type-II assignment."],"forward_implications":["At 1% tensile strain, in-plane electron and hole mobilities reach 2417 and 3550 cm²/Vs, an 80% and 68% improvement over unstrained BAs, making strained BAs a candidate channel for fast, low-power transistors.","The out-of-plane electron mobility falls to 482 cm²/Vs, so strained BAs is strongly anisotropic; this matters for any vertical device geometry and is a direct fingerprint of the valley-repopulation mechanism.","BAs substrates introduce less than 1% misfit strain for InGaN near 47% In, which would let InGaN devices reach red wavelengths without the dislocation problem that limits GaN substrates.","Type-II alignments with In0.5Ga0.5N and ZnSnN2 mean photogenerated electrons and holes separate across the junction, supporting photovoltaic and photodetector designs on a high-thermal-conductivity substrate.","Strain of 4% can lower the BAs band gap from 1.78 eV to roughly 0.89 to 1.15 eV, a range that could supply the 1.1 eV subcell in a multijunction solar cell."],"supporting_citations":[{"why":"Establishes the native p-type dopability of BAs that makes it the p-side of the proposed heterojunctions.","marker":"[4]"},{"why":"Supplies the transverse and longitudinal electron effective masses used to explain the strain-induced mobility increase.","marker":"[11]"},{"why":"Provides the converged bulk BAs electron and hole mobility values that serve as the unstrained baseline and comparison.","marker":"[13]"},{"why":"Documents the strained-silicon valley-splitting and degeneracy-lifting mechanisms that this paper transfers to BAs.","marker":"[21]"},{"why":"Supplies the lattice constant and structure data that show ZnSnN2 is nearly lattice-matched to BAs.","marker":"[27]"},{"why":"Defines the hybrid functional used for all band-structure, gap, and band-alignment calculations.","marker":"[31]"},{"why":"Supplies the electron-phonon interpolation capability used to compute carrier mobilities from the calculated coupling.","marker":"[38]"},{"why":"Provides the InGaN band-alignment data that, combined with the BAs alignment, yields the InGaN/BAs offsets.","marker":"[47]"},{"why":"Supports the 31% mixing parameter and the n-type defect context assumed for ZnSnN2.","marker":"[35]"}],"fun_headline_variants":["Tensile strain boosts BAs carrier mobility by 60%+","Strained BAs electrons and holes get 60% faster","BAs nearly matches InGaN and ZnSnN2 lattices","Type-II band alignment predicted for BAs with InGaN/ZnSnN2","Strained BAs: type-II bands with InGaN and ZnSnN2"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The reported offsets assume that the way the paper lines up each material's energy bands to the vacuum level is consistent across BAs, ZnSnN2 (computed with a fitted mixing parameter), and InGaN (taken from an earlier calculation); if those references disagree, the offsets and even the type-II classification could shift.","fun_headline_variants_meta":{"raw":{"variants":["Tensile strain boosts BAs carrier mobility by 60%+","Strained BAs electrons and holes get 60% faster","BAs nearly matches InGaN and ZnSnN2 lattices","Type-II band alignment predicted for BAs with InGaN/ZnSnN2","Strained BAs: type-II bands with InGaN and ZnSnN2"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001787,"raw_usage":{"total_tokens":7090,"prompt_tokens":1039,"completion_tokens":6051,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":655,"completion_tokens_details":{"reasoning_tokens":5951}},"tokens_in":655,"tokens_out":6051,"duration_ms":41067,"temperature":1.0,"reasoning_tokens":5951,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T05:17:50.634018+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Grow a BAs film on a substrate imposing about 1% biaxial tensile strain, measure the 300 K in-plane and out-of-plane electron and hole mobilities (predicted 2417/482 and 3550/2956 cm²/Vs), and check the predicted valley splitting (roughly 200 meV) and valence-band splitting (roughly 115 meV) by optical or angle-resolved photoemission measurements. If the mobilities do not increase and the band splittings are absent, the central claim fails. For the alignments, measure the valence-band offset at a BAs/InGaN interface by X-ray photoelectron spectroscopy; a value far from 0.30 eV would overturn the type-II prediction.","supporting_citations":[],"review_version":1}