{"id":"177af5f4-0288-4539-82a0-75c2e2b25f38","arxiv_id":"2411.14180","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Spin-valley polarization is preserved during ultrafast interlayer charge transfer in MoSe2/WSe2 at low temperature and lost at high temperature due to phonon scattering.","lead":"This paper uses ultrafast laser pulses to watch electrons and holes move between two atom-thick semiconductor layers, MoSe2 and WSe2. It finds that at low temperature the spin-valley memory of the particles survives the transfer, and at high temperature phonon vibrations erase it.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Rate-equation model excludes interlayer transfer from the depolarized MoSe2 K′ hole population; this structural choice inflates ΓCT,NP, so the headline rate comparison is not yet established.","rationale":"The paper is a well-executed helicity-resolved TA study, and the qualitative observation—that the two layers show the same circular dichroism at 8 K and that this polarization is lost at higher temperature—is directly supported by the displayed traces and consistent with earlier work (e.g., Refs. 28 and 23). I do not see grounds for a rejection-level objection. The concern is about the quantitative model used to extract rates. The reader flagged the neglect of separate electron dynamics and many-body contributions to the A-exciton bleach; that is a legitimate but diffuse worry. Sharper is the dead-end treatment of NMo- in the four rate equations. The fitted ΓVDP,Mo is comparable to ΓCT,NP, so the K′ MoSe2 population is not negligible during the transfer window; setting its CT rate to zero is a structural choice that directly determines what ΓCT,NP means. This is not an external-consensus issue: it is an internal identifiability problem in the model as written. The proposed extension (adding γ NMo- transfer) is a simple re-fit and would settle whether ΓCT,P > ΓCT,NP survives. Until that re-fit is shown, or until an independent measurement (e.g., time-resolved Kerr rotation or momentum-resolved photoemission) constrains the K′ MoSe2 transfer channel, the quantitative headline should be treated as conditional. I therefore keep the reader's CONDITIONAL verdict. The empty data-availability link and missing SI make this test currently impossible from the preprint alone; the authors should be asked to provide the data and the extended-model fit.","tokens_in":13366,"tokens_out":8216,"duration_ms":84898,"concrete_test":"Extend the rate equations to allow NMo- to transfer: add -γ NMo- to dNMo-/dt and +γ NMo- to dNW-/dt, with γ either free or constrained equal to ΓCT,P; refit the 8, 100, and 300 K datasets using the original fitting procedure. Report the fitted ΓCT,P, ΓCT,NP, and the confidence interval for ΔΓ = ΓCT,P - ΓCT,NP. If the interval includes zero or changes sign under this extension, the central claim is not robust; if ΔΓ remains positive with comparable fit quality, the model-based objection is resolved. The check can be done on the published raw traces once the data link is provided.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The four rate equations in Sec. 2 (after Fig. 3a) contain the term dNMo-/dt = ΓVDP,Mo(NMo+ - NMo-) with no ΓCT term. The paper justifies this by assuming \"only the sub-ps dynamics of the directly photoexcited AMo exciton at K serves as the source of both inter- and intra-layer spin-valley charge dynamics.\" This is the load-bearing point: NMo- is the depolarized hole population in MoSe2, and it is created at rate ΓVDP,Mo ≈ 0.83 ps^-1, only ~1.4 times slower than the fitted ΓCT,NP ≈ 1.19 ps^-1. A hole in MoSe2 K′ should be able to undergo the same type-II interlayer hopping to WSe2 K′ (spin/valley conserving) or to WSe2 K (spin-flip) with rates of the same order as ΓCT,P. By setting those rates to zero, all NW- population is forced through ΓCT,NP acting on the directly excited K population. The extracted ΓCT,NP therefore conflates genuine spin-flip transfer with transfer of holes that already depolarized inside MoSe2, and the reported inequality ΓCT,P > ΓCT,NP is not identified by the data under this model. The qualitative CD dynamics may still indicate valley conservation, but the quantitative headline rate comparison—and the temperature-dependent ratio in the inset of Fig. 3b—is contingent on this exclusion. The SI/data link is also empty, so the fit cannot currently be checked.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports helicity-resolved transient absorption measurements on a MoSe2/WSe2 heterostructure with a twist angle of about 2 degrees, resonantly exciting the MoSe2 A-exciton with circularly polarized light. The authors observe a delayed rise of the WSe2 A-exciton bleaching relative to MoSe2, a temperature-dependent rise delay, and a circular dichroism that decays on nanosecond timescales at 8 K but much faster at higher temperatures. A four-population rate-equation model is fit to the sub-picosecond dynamics to extract valley-preserving and non-valley-preserving interlayer hole transfer rates, along with valley depolarization rates. The authors conclude that spin-valley polarization is conserved during interlayer charge transfer at low temperature, that phonon scattering mediates the transfer and erases valley polarization at elevated temperature, and that the long-lived circular dichroism of intralayer excitons reports on interlayer exciton formation.","tokens_in":13777,"tokens_out":9938,"duration_ms":94076,"significance":"If the conclusions hold, this is a valuable experimental contribution to the question of whether spin-valley information survives ultrafast interlayer charge transfer in TMD heterostructures. The qualitative data are likely robust: the delayed WSe2 response, the difference between K and K' rise times at 8 K, and the nanosecond CD in the heterostructure compared with roughly 20 ps in the monolayer are all clean and interesting observations. The quantitative rate comparison, however, is contingent on a rate-equation model with a structural omission, and the missing SI and data link currently prevent verification of the fits. The paper is potentially significant for valleytronics applications, but the quantitative claims need to be placed on firmer footing.","major_comments":[{"comment":"The model omits interlayer transfer from the depolarized MoSe2 K' population NMo-. The equation dNMo-/dt = ΓVDP,Mo(NMo+ - NMo-) contains no ΓCT terms, and no corresponding source terms appear in the WSe2 equations. Because ΓVDP,Mo ≈ 0.83 ps^-1 is comparable to ΓCT,NP ≈ 1.19 ps^-1, NMo- is substantially populated during the transfer window. A hole in K' MoSe2 should be able to transfer to WSe2 with valley-preserving or valley-flipping rates of the same order as those from K. Setting those rates to zero forces the observed WSe2 K' population to be generated exclusively by ΓCT,NP applied to NMo+, so the fitted ΓCT,NP conflates spin-flip transfer with transfer of holes that had already depolarized within MoSe2. The headline inequality ΓCT,P > ΓCT,NP and the temperature-dependent ratio in the inset of Fig. 3(b) are therefore not established by the model as written. Please extend the equations to include transfer out of NMo- (for example, -ΓCT,P NMo- - ΓCT,NP NMo- in the NMo- equation, with corresponding sources in the NW- and NW+ equations), re-fit, and discuss the identifiability of the resulting rates, or restrict the quantitative conclusions to what the reduced model can actually support.","section":"Section 2, rate equations following Fig. 3(a)"},{"comment":"The data availability statement reads \"Data pertaining to this work is available by following link:\" and then gives no URL. The SI, which is repeatedly cited for the fitting procedure, the temperature-dependent dynamics, and the twist-angle determination, is not included with the arXiv version. Since the central quantitative results are defined by a fit to these data, the fitting code/data and the SI must be supplied before the claims can be checked.","section":"Data and materials availability statement"},{"comment":"The model assumes that the transient bleaching of the intralayer A-excitons is a direct readout of hole population and that electrons in MoSe2 influence the AMo signal only through phase space filling. The AMo circular dichroism, however, depends on the valley occupation of both electrons and holes; without separate electron dynamics the fitted ΓVDP,Mo and the CT rates may absorb electron valley relaxation or many-body contributions to the bleaching. The qualitative CD evolution is likely unaffected, but the quantitative rates should be presented with this caveat explicitly tested, for example by comparing fits with and without an electron population or by using a doping-dependent measurement.","section":"Section 2, assumptions after Fig. 3(a)"}],"minor_comments":[{"comment":"The fit curves are not shown overlaid on the data, and no residuals or goodness-of-fit statistics are given; please provide these for all temperatures shown in the inset.","section":"Fig. 3(b)"},{"comment":"The ΓCT,P/ΓCT,NP ratio is plotted versus temperature without error bars and without displaying the fits at 100 K and 300 K; please present those results or give a clear cross-reference to an SI section that is actually included.","section":"Inset of Fig. 3(b)"},{"comment":"The abstract and conclusions state that spin-valley polarization is \"conserved\" during charge transfer, but the data show a nanosecond decay of the CD; a phrase like \"preserved on the timescale of the experiment\" or \"long-lived\" would be more precise.","section":"Abstract and Conclusions"},{"comment":"The claim that interlayer excitons and their spin-valley polarization can be probed in the intralayer exciton response is an inference from the common nanosecond decay of AMo and AW CD, not a direct measurement of ILEs; please soften the wording or provide an independent ILE signature.","section":"Section 4, long-time dynamics"},{"comment":"The extracted AW rise delays of about 50 fs at 300 K and 200 fs at 8 K are comparable to the 100 fs instrument response; please report uncertainties and a statistical test for the temperature dependence of the rise delay.","section":"Fig. 2(d)-(f)"},{"comment":"There are minor editorial issues: the text after Fig. 2(b) says \"the dynamics of AMo and AMo\" where the second should presumably be AW, the figure axes use comma decimal separators inconsistently, and the title contains \"T ransfer\" with an extra space.","section":"General presentation"}],"recommendation":"major_revision","confidential_remarks":"The experimental dataset appears to be of high quality and the qualitative spin-valley conservation story is plausible, but the quantitative rate comparison is not identifiable from the reduced model as presented, and the missing SI/data link prevents verification. I recommend major revision rather than rejection: the authors should either extend the rate-equation model to include transfer from the depolarized MoSe2 population and re-fit, or remove the quantitative rate claims. The final manuscript must also include the SI and a working data link. The paper is within scope for this journal and, once the model issue is addressed, could be a solid contribution."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper is a careful helicity-resolved transient absorption study of MoSe2/WSe2 heterobilayers, and the experimental core is solid. The 8/100/300 K series is the real addition: it cleanly shows that the WSe2 A-exciton rise delays as temperature drops, that the circular dichroism of both layers tracks together at 8 K for nanoseconds, and that the intralayer bleaching carries a long-lived fingerprint of the interlayer exciton. The authors are also transparent about their model assumptions, which I appreciate. The delayed rise and the CD dynamics genuinely support the qualitative picture: low-temperature CT preserves spin-valley polarization, high temperature washes it out via phonon scattering.\n\nThe soft spot is the rate-equation model used to extract the quantitative rates. The four equations decouple NMo-, the depolarized hole population in MoSe2, from any interlayer transfer. That choice is justified by saying only the directly photoexcited K population serves as the source of CT dynamics. But the VDP rate in MoSe2 is 0.83 ps^-1, only about 1.4 times slower than the fitted non-valley-preserving CT rate of 1.19 ps^-1. A hole depolarized in MoSe2 should have the same type-II hopping channels available to it, with comparable rates. By forcing all transfer through NMo+, the model bundles genuine spin-flip transfer together with holes that already depolarized in the source layer. The headline inequality ΓCT,P > ΓCT,NP is therefore not actually identified by the data under this model — it's an artifact of a structural choice. That does not kill the qualitative valley-conservation claim, but it does mean the quantitative rates and the temperature-dependent ratio in the inset should be read as model-dependent, not measured facts.\n\nThere are two smaller issues. The paper assumes electrons only affect the AMo bleach through phase-space filling; bandgap renormalization or exciton-exciton interactions could bias the fitted rates, though the direction is unclear. And the data availability link is empty, with no SI or code included, so the fits can't currently be checked. Given the model issue, that's a real problem for verification.\n\nThe paper is for the ultrafast spectroscopy and 2D valleytronics crowd. It deserves a serious referee because the experiment is well executed and the qualitative result is likely right, but the authors need to either extend the model to include CT from the depolarized MoSe2 population, or reframe the quantitative claims as order-of-magnitude estimates. I'd send it to peer review with that demand.","headline":"Solid temperature-series helicity-resolved TA data on MoSe2/WSe2, but the headline rate comparison rests on a rate-equation model that silently excludes transfer from depolarized MoSe2 holes, so the quantitative claim isn't yet established.","tokens_in":14326,"tokens_out":2049,"would_cite":false,"duration_ms":20989,"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":"Valley-polarized holes survive interlayer transfer at 8 K","keywords":["spin-valley polarization","interlayer charge transfer","transient absorption spectroscopy","transition metal dichalcogenides","heterostructures","phonon scattering","valleytronics","interlayer excitons"],"falsifier":"Prepare the same heterostructure in a gate-tunable device and measure the A-exciton bleaching while injecting electrons without holes (or vice versa). If the bleaching amplitude does not scale linearly with the injected carrier density, the phase-space-filling-only assumption fails and the extracted $\\Gamma_{\\mathrm{CT,P}}/\\Gamma_{\\mathrm{CT,NP}}$ ratio would need to be re-evaluated. Alternatively, measuring the ratio as a function of twist angle should show whether the valley-preserving advantage is intrinsic to the interface or a property of the near-zero-twist sample.","tokens_in":13184,"feed_emoji":"⚛️","tokens_out":8841,"duration_ms":69520,"temperature":0.7,"pith_summary":"This paper uses helicity-resolved ultrafast transient absorption spectroscopy to track what happens to spin-valley polarization when an exciton is created in the MoSe2 layer of a MoSe2/WSe2 heterostructure and its hole transfers to WSe2. It claims that at 8 K the transfer is valley-conserving: the rate for a hole to hop while keeping its valley index is about $\\Gamma_{\\mathrm{CT,P}} \\approx 2.08\\ \\mathrm{ps}^{-1}$, versus about $\\Gamma_{\\mathrm{CT,NP}} \\approx 1.19\\ \\mathrm{ps}^{-1}$ for hopping into the opposite valley, so the spin-valley polarization of the photoexcited state survives charge transfer and the formation of interlayer excitons. At higher temperatures, phonon scattering makes both channels equally likely and the circular dichroism of the two layers decays together, showing that spin-valley memory is lost. The paper also argues that interlayer excitons, which are hard to see directly because of their weak oscillator strength, can be monitored through the much stronger transient absorption of the intralayer excitons. If correct, this opens a practical way to read out interlayer spin-valley polarization and sets a temperature constraint for valleytronic devices.","feed_headline":"Valley-polarized holes survive interlayer transfer at 8 K","feed_subtitle":"Phonons wash out the valley memory at higher temperatures, so cryogenic operation is key for valleytronic devices.","key_machinery":"The argument rests on helicity-resolved broadband transient absorption spectroscopy combined with a four-state rate-equation model. The observables are the transient bleaching amplitudes of the MoSe2 and WSe2 A-excitons measured with co- and cross-circularly polarized probe pulses; their difference defines the circular dichroism that tracks valley polarization. The rate equations couple populations in the four valley/layer states (K and K' in each layer) via valley-preserving interlayer transfer $\\Gamma_{\\mathrm{CT,P}}$, non-valley-preserving transfer $\\Gamma_{\\mathrm{CT,NP}}$, and intralayer valley depolarization rates $\\Gamma_{\\mathrm{VDP,Mo}}$ and $\\Gamma_{\\mathrm{VDP,W}}$. Fitting this model to the sub-picosecond dynamics yields the transfer-rate ratio whose temperature dependence is the central evidence for phonon-mediated spin-valley-selective charge transfer.","core_discovery":"The central discovery is that, following resonant circular-polarized excitation of the MoSe2 A-exciton in a nearly aligned MoSe2/WSe2 heterostructure, the hole that transfers to WSe2 retains its valley index at low temperature. From a four-population rate-equation model of the K and K' valley populations in both layers, the paper extracts a valley-preserving interlayer transfer rate of $\\Gamma_{\\mathrm{CT,P}} = 2.08 \\pm 0.02\\ \\mathrm{ps}^{-1}$ that is nearly twice the valley-flipping rate $\\Gamma_{\\mathrm{CT,NP}} = 1.19 \\pm 0.05\\ \\mathrm{ps}^{-1}$ at 8 K, and finds that the circular dichroism of the two layers remains identical for nanoseconds. The polarization is lost at elevated temperatures, where thermally populated phonons make the two transfer channels equally probable, consistent with a phonon-mediated charge-transfer mechanism. In addition, the long-lived dynamics of the intralayer exciton bleaching in both layers are attributed to the presence of interlayer excitons, indicating that interlayer exciton population and spin-valley polarization can be probed through the stronger intralayer response.","pith_inferences":["The model's assumption that electrons influence the MoSe2 A-exciton bleaching only through phase space filling could be tested by measuring the response under electron-only injection in a gate-tunable device; if many-body effects such as bandgap renormalization contribute, the extracted transfer rates would be biased.","Because the sample has a near-zero twist angle of about 2 degrees, the valley-preserving transfer rate may be specific to nearly commensurate interfaces; larger moiré angles introduce momentum mismatch that could suppress $\\Gamma_{\\mathrm{CT,P}}$, an extension the paper does not address.","The long-lived coincident circular dichroism of both layers implies an interlayer-exciton spin-valley lifetime of at least a nanosecond at 8 K, suggesting that the heterostructure acts as a cryogenic spin-valley memory element.","The temperature dependence of $\\Gamma_{\\mathrm{CT,P}}/\\Gamma_{\\mathrm{CT,NP}}$ suggests a crossover temperature below which the valley-preserving channel dominates; mapping that crossover precisely as a function of twist angle and dielectric environment would turn this observation into a design rule for valleytronic devices."],"forward_implications":["Low-temperature operation preserves spin-valley information across ultrafast charge separation, a prerequisite for using such heterostructures as valley filters or spin memories.","The intralayer exciton bleaching can serve as a non-invasive optical probe of interlayer exciton population and spin-valley polarization, bypassing the difficulty of detecting weak interlayer exciton absorption.","The extracted rates quantify the competition between spin-orbit coupling, which blocks valley flips, and phonon scattering, which enables them, giving a concrete target for phonon-engineering strategies like strain or encapsulation.","At elevated temperatures the valley-preserving advantage disappears, so any valleytronic application that relies on spin-valley-selective transfer would need thermal management.","The similar nanosecond decay of circular dichroism in both layers points to a common spin-valley relaxation channel set by the interlayer exciton, rather than independent intralayer relaxation."],"supporting_citations":[{"why":"Provides the phonon-mediated interlayer charge separation mechanism in MoSe2/WSe2 that this work confirms and quantifies.","marker":"[7]"},{"why":"Supports the role of collective atomic motion in ultrafast interlayer charge transfer, grounding the phonon-mediated interpretation.","marker":"[8]"},{"why":"Shows spin-polarized charge separation in a TMD heterobilayer at room temperature, establishing the spin-valley transfer phenomenon this paper extends to cryogenic temperatures.","marker":"[20]"},{"why":"Reports ultralong valley lifetime in WSe2/MoS2 heterostructures, the expectation of extended spin-valley coherence that motivates the measurement.","marker":"[24]"},{"why":"Observed negative circular polarization in WSe2/MoSe2 heterobilayers, the signature the paper uses to infer same-sign circular dichroism in both layers after transfer.","marker":"[28]"},{"why":"Provides the phonon-activated valley relaxation temperature scale around 100 K used to interpret the temperature dependence.","marker":"[34]"},{"why":"Measures spin/valley coupled dynamics at a MoS2/MoSe2 interface with delayed rise times, the experimental signature the authors reproduce.","marker":"[36]"},{"why":"Models phonon-assisted ultrafast charge transfer at van der Waals interfaces, the microscopic mechanism invoked for the temperature-dependent rates.","marker":"[39]"},{"why":"Establishes that phonons limit valley polarization in TMDs, supporting the claim that phonon scattering erases valley memory at higher temperatures.","marker":"[41]"}],"fun_headline_variants":["Spin memory survives interlayer hop in MoSe2/WSe2","Phonons decide fate of valley-polarized holes","Valley polarization preserved in ultrafast charge transfer","Cryogenic temps key to valleytronic memory in 2D stacks","Interlayer excitons probed via intralayer response"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The central assumption is that the transient bleaching of the MoSe2 A-exciton is a direct readout of the hole population in that layer, with electrons influencing the signal only through phase space filling; if many-body effects such as bandgap renormalization or exciton-exciton interactions also contribute to the bleaching, the fitted charge-transfer rates—and with them the spin-valley conservation conclusion—would be biased.","fun_headline_variants_meta":{"raw":{"variants":["Spin memory survives interlayer hop in MoSe2/WSe2","Phonons decide fate of valley-polarized holes","Valley polarization preserved in ultrafast charge transfer","Cryogenic temps key to valleytronic memory in 2D stacks","Interlayer excitons probed via intralayer response"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000649,"raw_usage":{"total_tokens":2996,"prompt_tokens":978,"completion_tokens":2018,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":594,"completion_tokens_details":{"reasoning_tokens":1934}},"tokens_in":594,"tokens_out":2018,"duration_ms":13693,"temperature":1.0,"reasoning_tokens":1934,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T15:26:16.604456+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Prepare the same heterostructure in a gate-tunable device and measure the A-exciton bleaching while injecting electrons without holes (or vice versa). If the bleaching amplitude does not scale linearly with the injected carrier density, the phase-space-filling-only assumption fails and the extracted $\\Gamma_{\\mathrm{CT,P}}/\\Gamma_{\\mathrm{CT,NP}}$ ratio would need to be re-evaluated. Alternatively, measuring the ratio as a function of twist angle should show whether the valley-preserving advantage is intrinsic to the interface or a property of the near-zero-twist sample.","supporting_citations":[{"cited_title":"Phonon-mediated interlayer charge separation and recombination in a MoSe _2 /WSe _2 heterostructure","cited_arxiv_id":null,"evidence_quote":"Provides the phonon-mediated interlayer charge separation mechanism in MoSe2/WSe2 that this work confirms and quantifies."},{"cited_title":"The role of collective motion in the ultrafast charge transfer in van der W aals heterostructures","cited_arxiv_id":null,"evidence_quote":"Supports the role of collective atomic motion in ultrafast interlayer charge transfer, grounding the phonon-mediated interpretation."},{"cited_title":"Spin-Polarized Charge Separation in a Photoexcited Transition Metal Dichalcogenide Heterobilayer at Room Temperature","cited_arxiv_id":null,"evidence_quote":"Shows spin-polarized charge separation in a TMD heterobilayer at room temperature, establishing the spin-valley transfer phenomenon this paper extends to cryogenic temperatures."},{"cited_title":"F.; Wang, F","cited_arxiv_id":null,"evidence_quote":"Reports ultralong valley lifetime in WSe2/MoS2 heterostructures, the expectation of extended spin-valley coherence that motivates the measurement."},{"cited_title":"Negative circular polarization emissions from WSe _2 /MoSe _2 commensurate heterobilayers","cited_arxiv_id":null,"evidence_quote":"Observed negative circular polarization in WSe2/MoSe2 heterobilayers, the signature the paper uses to infer same-sign circular dichroism in both layers after transfer."},{"cited_title":"Evidence for line width and carrier screening effects on excitonic valley relaxation in 2D semiconductors","cited_arxiv_id":null,"evidence_quote":"Provides the phonon-activated valley relaxation temperature scale around 100 K used to interpret the temperature dependence."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Measures spin/valley coupled dynamics at a MoS2/MoSe2 interface with delayed rise times, the experimental signature the authors reproduce."},{"cited_title":"A.; Xie, Y.; Lan, Z.; Prezhdo, O","cited_arxiv_id":null,"evidence_quote":"Models phonon-assisted ultrafast charge transfer at van der Waals interfaces, the microscopic mechanism invoked for the temperature-dependent rates."},{"cited_title":"Phonon-limited valley polarization in transition-metal dichalcogenides","cited_arxiv_id":null,"evidence_quote":"Establishes that phonons limit valley polarization in TMDs, supporting the claim that phonon scattering erases valley memory at higher temperatures."}],"review_version":1}