{"id":"b4804be8-8a74-4804-a1b2-64888867f4ed","arxiv_id":"2502.06495","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":7,"one_line_summary":"Resonant spin amplification in MAPbI3 single crystals yields a record electron spin dephasing time of 21 ns, with longitudinal spin relaxation up to 30 ns.","lead":"This paper measures how long electron spins remember their direction in thin MAPbI3 perovskite crystals at cryogenic temperatures, finding record-long spin memory of 21 nanoseconds. The result suggests perovskite crystals could be useful for spintronic devices that store or process information in electron spin.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"RSA T2* extraction ignores dynamic nuclear polarization, which can narrow the zero-field peak and inflate the 21 ns record.","rationale":"The reader's weakest assumption concerns passive nuclear Overhauser field fluctuations broadening the RSA peak, which would make the extracted T2* an underestimate and hence would not threaten the record claim's direction. My concern is different and more damaging: DNP can actively narrow the zero-field RSA peak through nuclear polarization or frequency focusing, which would make the extracted T2* an overestimate of the intrinsic spin dephasing time. The paper provides direct evidence of DNP in the same crystal (Section II F), and the RSA setup does not appear to suppress it. This is a concrete, mechanism-based risk to the central comparative claim, not merely a lineshape nit. The correct response is to keep the verdict CONDITIONAL, adding the requirement that RSA be repeated under DNP-suppressing conditions or that the rôle of nuclear feedback be quantified. I do not think the paper should be rejected: the experimental work is substantial, the RSA analysis is internally consistent with Eq. (6), and the DNP effect may turn out to be negligible at the low pump power used. But the record claim should not be treated as settled until this check is performed.","tokens_in":21019,"tokens_out":7826,"duration_ms":81429,"concrete_test":"Repeat the RSA measurement of Fig. 5 with alternating pump helicity (σ+/σ- modulation at a frequency high enough to suppress DNP buildup, e.g., ≫1/T_N where T_N is the nuclear spin relaxation time) and compare the zero-field RSA peak width with the fixed-helicity result. Also sweep the magnetic field upward and downward to check for hysteresis. If the extracted T2,0,e* changes by more than the fitting uncertainty, the 21 ns record is not an intrinsic material parameter and must be re-qualified.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central record claim T2,0,e* = 21 ns is obtained from the zero-field RSA peak (Fig. 5c) fitted with Eq. (5), a Lorentzian that assumes a single exponential decay τs and no nuclear-spin feedback. The paper itself demonstrates strong hyperfine coupling and dynamic nuclear polarization (DNP) in this material (Section II F, Eqs. (12)-(13), Fig. 8). In the RSA experiments the pump helicity is fixed (amplitude modulated, per the Experimental Section), so DNP can accumulate over the 76 MHz pulse train. The resulting Overhauser field acts on the electron spins; if it partially polarizes or 'focuses' the nuclear spin bath, it can reduce the effective dephasing or lock precession near the pulse repetition rate, narrowing the central RSA peak. Eq. (4) contains no nuclear-field distribution, no DNP feedback, and no check for scan-direction hysteresis. Because the abstract's claim is explicitly comparative ('longest reported so far'), an uncontrolled DNP contribution is a load-bearing threat: the measured 21 ns may reflect the pumped nuclear-spin state rather than the intrinsic electron spin dephasing time. At minimum, the value should be reported as conditional on the experimental helicity and nuclear-spin history.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports time-resolved Kerr ellipticity studies of carrier spin dynamics in a 20-µm-thick MAPbI3 single crystal at 1.6 K. It claims a zero-field electron spin dephasing time T2,0,e* = 21 ns obtained from resonant spin amplification (RSA), a longitudinal spin relaxation time T1 reaching 30 ns at BF = 20 mT from spin inertia, a small electron g-factor dispersion Δge = 0.006, and dynamic nuclear polarization with Overhauser fields up to +3.3 mT for electrons and -17.2 mT for holes. The paper introduces RSA and spin-inertia techniques to lead halide perovskites and demonstrates spin accumulation with a 76 MHz pulsed laser.","tokens_in":21343,"tokens_out":18511,"duration_ms":157222,"significance":"If the central claim holds, 21 ns is the longest electron spin dephasing time reported for lead halide perovskites and would strengthen the case for these materials in spin-based technologies. The paper has notable strengths: the zero-field RSA peak width is independently fit by a Lorentzian (19.4 ns) and by the full RSA model (21 ns), and these values are quantitatively consistent with the 3.3 ns dephasing measured at 0.5 T through Eq. (6). The first demonstration of RSA in perovskites and the explicit measurements of Overhauser fields are valuable. The main uncertainties concern the nuclear-spin-state dependence of the RSA peak and the uncontrolled comparison underlying the 'longest reported' claim.","major_comments":[{"comment":"The record value T2,0,e* = 21 ns is extracted from the zero-field RSA peak using Eq. (4) and the Lorentzian approximation Eq. (5), both of which assume a single spin dephasing time and no nuclear-spin feedback. The RSA experiment is performed with an amplitude-modulated pump of fixed helicity (Sec. V), so a helicity-dependent Overhauser field can accumulate over the 76 MHz pulse train. The authors themselves demonstrate strong dynamic nuclear polarization in this material (Sec. II F, Fig. 8) and note that nuclear frequency focusing can modify coherent carrier spin dynamics. If the pumped nuclear bath reduces transverse Overhauser-field fluctuations, the zero-field RSA peak would be narrowed and T2,0,e* would reflect the nuclear spin state rather than the intrinsic electron dephasing. Conversely, if the nuclear-field distribution simply broadens the peak, the extracted value is an upper bound on the homogeneous dephasing time. Because the abstract's claim is explicitly comparative ('longest reported so far'), this possibility must be excluded or stated as a condition of the measurement. A control comparing RSA with opposite helicities, with helicity modulation, or with scan-direction/hysteresis checks would address this.","section":"II C, Eqs. (4)-(5), and II F"},{"comment":"The statement that T2,0,e* = 21 ns is 'the longest reported so far for lead halide perovskite semiconductors' compares a zero-field value with literature values (11.5 ns for FAPbBr3 [16], 11 ns for MAPbI3 [15]) without specifying the magnetic fields at which those values were obtained. Since T2* in this work decreases strongly with field (Eq. (6), Fig. 5d), a zero-field value is not directly comparable with a finite-field value unless the field is stated. The record claim should be restricted to the same field condition or reformulated with the field dependence explicitly stated.","section":"II C"}],"minor_comments":[{"comment":"The abstract reports 'T1 = 30 ns' without stating that this is the spin inertia time Ts measured at BF = 20 mT and without the Ts ≈ T1 caveat from Eq. (10); the zero-field value from the same method is about 20 ns (Fig. 7d). Please state the field and the Ts/T1 distinction.","section":"Abstract and II E"},{"comment":"The hole spin dephasing time is reported as 0.35 ns from the time-domain fit but about 1.0 ns from the FFT linewidth (Δωh = 6.03 rad/ns); this factor-of-three discrepancy is not discussed.","section":"II B and Fig. 2"},{"comment":"The sentence 'The spin dephasing time is decreasing with magnetic field due to the decreasing impact of the g-factor dispersion' should read 'increasing impact', since the g-factor dispersion has a larger effect at higher fields.","section":"II C, text near Eq. (6)"},{"comment":"The word 'Lorenzian' should be 'Lorentzian'.","section":"Fig. 5 caption"},{"comment":"The phrase 'Vice verse' should be 'Vice versa'.","section":"II B"},{"comment":"In the sample synthesis description, 'PI2' presumably should be 'PbI2'.","section":"V"},{"comment":"The zero-field KE decay gives T1 = 26 ns (Fig. 2a), while spin inertia gives Ts ≈ 20 ns at zero field; the origin of this 30% difference is not commented on.","section":"II D and Fig. 7d"},{"comment":"The symbol T2,0* is used both for the zero-field dephasing time in Eq. (6) and for the temperature-independent dephasing time in Eq. (14); these are different quantities and using separate symbols would avoid confusion.","section":"Eqs. (6) and (14)"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is from a group with a strong track record in perovskite spin physics, and the RSA implementation is technically sound. The main risk is overinterpretation of the 'longest reported' claim without a nuclear-spin control; given the group's own prior work on nuclear spin squeezing in these materials, they are well positioned to add the requested control. The paper fits the journal scope, and the requested changes are achievable within a revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: you should send this to review. The central claim—electron T2* = 21 ns, the longest reported for a lead halide perovskite—is supported by a clean RSA measurement and by internal consistency with the time-domain Kerr ellipticity at 0.5 T. The group knows these techniques well and the data look honest.\n\nWhat's new: first application of resonant spin amplification to lead halide perovskites. The authors transfer the standard RSA/spin-inertia/PR toolkit from III-Vs to MAPbI3 thin crystals, and they get a nice result: a record 21 ns electron dephasing time, exceeding the 11.5 ns FAPbBr3 and 11 ns MAPbI3 records. The g-factor dispersion of 0.006 (0.2%) is a good measure of sample quality and matches an estimate from the pump spectral width. The multi-technique approach (RSA, PR, spin inertia, DNP, temperature dependence) gives a fairly complete picture.\n\nSoft spots, in descending order of importance.\n\nFirst, the abstract reports T1 = 30 ns as if it were a zero-field value. It's actually the spin-inertia value at BF = 20 mT; zero-field spin inertia gives 20 ns and the time-domain KE decay gives 26 ns. The abstract and Table I should say 'up to 30 ns at 20 mT.' That's a presentation flaw, not a physics problem.\n\nSecond, the RSA zero-field peak is fitted with a Lorentzian that has no nuclear-spin feedback. Since the pump helicity is fixed (amplitude modulation), DNP could in principle accumulate and reduce nuclear-fluctuation dephasing, narrowing the peak and inflating the 21 ns. The authors do demonstrate strong DNP at high pump power (32 W/cm^2), but the RSA runs at 0.5 W/cm^2 where DNP should be much weaker. Still, they should show a scan with alternating helicity or a power dependence of the extracted T2* to rule out nuclear effects. This is a referee question, not a fundamental flaw.\n\nThird, the quantum technology framing in the abstract is overreach. Ensemble spin dephasing in a perovskite crystal is nice spintronics material physics, but it doesn't establish a platform for quantum technologies on its own.\n\nMinor: no error bars on fitted parameters, raw data only 'upon request,' and the PR-component assignment leans on analogy with a different perovskite from the same group. The citation pattern looks fine; the self-citations are to their own prior perovskite spin work, which is the relevant literature.\n\nWho's it for: the perovskite spin-dynamics community, and anyone benchmarking spin coherence in new semiconductors. The paper deserves a serious referee; I'd send it out with a request to address the T1 labeling and the DNP caveat.","headline":"First RSA in lead halide perovskites reports a record 21 ns electron dephasing; the measurement is plausible but the abstract overstates T1 and the zero-field peak is not checked against DNP effects.","tokens_in":21932,"tokens_out":10172,"would_cite":true,"duration_ms":83722,"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":"A 20-micrometre MAPbI3 perovskite single crystal keeps electron spins coherent for 21 nanoseconds at 1.6 K, the longest such time reported in any lead halide perovskite, and long enough that spins from successive laser pulses add up and…","keywords":["lead halide perovskites","MAPbI3","carrier spin dynamics","time-resolved Kerr ellipticity","resonant spin amplification","spin accumulation","spin inertia","spintronics"],"falsifier":"Measure the same zero-field RSA peak with a two-pulse spin echo or Hahn echo sequence: if the echo-derived homogeneous dephasing time is clearly longer than $T_2^* = 21$ ns, then the RSA Lorentzian width is broadened partly by inhomogeneous nuclear Overhauser fields and the quoted value is a lower bound. A complementary check is to compare the zero-field RSA peak width with the width predicted by Equation (5) and to repeat the measurement at fields of tens of millitesla, where the polarization recovery data show nuclear fluctuations are suppressed; a narrowing beyond the g-factor dispersion model would confirm the nuclear contribution.","tokens_in":20763,"feed_emoji":"🧲","tokens_out":15808,"duration_ms":115145,"temperature":0.7,"pith_summary":"This paper reports that electron spins in a 20-micrometre MAPbI3 perovskite single crystal remain coherent for about $T_{2,e}^{*}=21$ ns at 1.6 K, the longest transverse spin dephasing time reported so far for any lead halide perovskite, and that the longitudinal spin relaxation time $T_1$ reaches about 30 ns in a 20 mT field. Because these lifetimes exceed the 13.2 ns period of the 76 MHz laser pulses, the spin polarization created by one pulse survives until the next pulse arrives, producing spin accumulation. The authors use that accumulation to perform resonant spin amplification, polarization recovery, and spin inertia measurements, extracting electron and hole g-factors, spin dephasing times, a small g-factor spread for the electron ensemble, and nuclear Overhauser fields from dynamic nuclear polarization. The paper's point is to show that lead halide perovskites support the coherent spin phenomena previously seen in conventional III-V and II-VI semiconductors, establishing them as a platform for spintronic and spin-based quantum technologies.","feed_headline":"Perovskite crystal holds electron spin for 21 ns","feed_subtitle":"A record for lead halide perovskites: spins outlive the laser pulse period, enabling spin accumulation.","key_machinery":"The experiments detect spin precession through time-resolved Kerr ellipticity, the change in elliptical polarization of a linearly polarized probe beam after reflection from the spin-polarized sample. The load-bearing mechanism is resonant spin amplification (RSA): when the spin dephasing time $T_2^*$ exceeds the laser repetition period $T_R = 13.2$ ns, spin polarization from successive pulses adds coherently, and the detected Kerr ellipticity is amplified whenever the Larmor precession frequency $\\omega_L$ satisfies the phase synchronization condition $\\omega_L = n\\,\\omega_R$ with $\\omega_R = 2\\pi/T_R$. The paper sums the pulse train analytically in Equation (4), which gives the RSA peak shape and its magnetic-field-dependent amplitude; for sharp peaks and long spin memory this reduces to Equation (5), a Lorentzian in the detuning $\\omega_L T_R - 2\\pi n$ whose width is set by $T_R/T_2^*$. To reproduce the decay of peak amplitude at higher fields, the paper averages Equation (4) over a Gaussian distribution of electron g-factors, extracting the median g-factor $g_{e,0}=2.676$ and the spread $\\Delta g_e = 0.006$. Supporting techniques are the polarization recovery curve (Equation (8)), which separates electron and hole hyperfine couplings; the spin inertia frequency response (Equation (11)), which yields $T_1$; and the pump-helicity dependence of Larmor frequencies (Equation (13)), which gives the nuclear Overhauser fields. The RSA peaks are spaced by about 2 mT, reflecting the electron g-factor.","core_discovery":"The central claim is that a thin, structurally high-quality MAPbI3 single crystal in its low-temperature orthorhombic phase supports electron spin coherence lasting $T_{2,0,e}^{*}=21$ ns, with a Lorentzian fit of the zero-field resonant spin amplification peak giving 19.4 ns, which the paper identifies as the longest electron spin dephasing time reported for lead halide perovskites. The long-lived spin component is assigned to localized electrons based on the electron g-factor $g_{e,0}=2.676$ obtained by modeling the RSA pattern with Equation (4) averaged over a Gaussian distribution of g-factors of width $\\Delta g_e = 0.006$; the hole component dephases much faster, with $T_{2,h}^{*}\\approx 0.8$ ns. Longitudinal spin relaxation, measured by the spin inertia technique, increases from about 20 ns at zero field to about 30 ns at $B_F = 20$ mT, so the paper concludes $T_1 \\ge 30$ ns. The experiments also expose the nuclear spin bath: the polarization recovery curve contains components at 3 mT and 21 mT assigned to electrons and holes, and circularly polarized pumping in a tilted magnetic field produces nuclear Overhauser fields of about $+3.3$ mT on electrons and $-17.2$ mT on holes. In the authors' interpretation, these results make spin accumulation and resonant spin amplification workable tools for halide perovskites and position MAPbI3 as a candidate platform for spin-based quantum technologies.","pith_inferences":["The quoted 21 ns value is extracted from the zero-field RSA peak width under an exponential-dephasing assumption; if random nuclear Overhauser fields inhomogeneously broaden that peak, the true homogeneous dephasing time could be longer, making 21 ns a lower bound rather than the limiting coherence time.","A two-pulse spin echo or Hahn echo measurement on the same crystal would separate homogeneous from inhomogeneous contributions to $T_2^*$ and directly test whether nuclear spin fluctuations set the zero-field linewidth.","The same RSA protocol could be transferred to perovskite nanocrystals and two-dimensional halide perovskites, where shorter spin times and different g-factor tunability would change the resonance pattern but the same phase synchronization condition should still govern the spin accumulation.","The weak hole spin signal (about one fifth of the electron amplitude) suggests a sample-specific resident hole population; surface treatments or doping that alter the hole density should change the hole RSA and polarization recovery amplitudes in a testable way."],"forward_implications":["If the 21 ns electron dephasing time is correct, MAPbI3 thin single crystals hold electron spin coherence about twice as long as the previous best single-crystal value of 11 ns, and far longer than the sub-nanosecond times typical of perovskite films.","Because $T_1 \\approx 30$ ns and $T_2^* \\approx 21$ ns both exceed the 13.2 ns laser period, spin accumulation is accessible with a standard 76 MHz pulsed laser, so resonant spin amplification, polarization recovery, and spin inertia measurements do not require a special low-repetition-rate source.","The measured g-factor spread $\\Delta g_e/g_e \\approx 0.2\\%$ shows that the electron spin ensemble is highly homogeneous, which the paper attributes to the structural quality of the thin single crystal and which underlies the sharp RSA peaks.","The demonstration transfers established resonant spin amplification and spin inertia protocols from III-V and II-VI semiconductors to lead halide perovskites, and the paper expects similarly long spin times in other high-quality perovskite compositions.","Dynamic nuclear polarization produces measurable Overhauser fields on both electrons and holes, so the nuclear spin bath in these crystals can be addressed and read out through carrier spin precession."],"supporting_citations":[{"why":"provides the original demonstration of resonant spin amplification in n-type GaAs that this paper transfers to MAPbI3.","marker":"[25]"},{"why":"supplies the theoretical treatment of RSA versus spin mode locking, including the pulse-train summation in Equation (4) and the g-factor averaging procedure.","marker":"[31]"},{"why":"is the reference for time-resolved Kerr rotation and Faraday/Kerr techniques, Larmor precession, and the field dependence of $T_2^*$ used throughout.","marker":"[10]"},{"why":"gives the bulk MAPbI3 spin dynamics parameters, including an electron spin dephasing time of 0.4 ns, that the paper compares with its 21 ns value.","marker":"[14]"},{"why":"reports the previous 11 ns electron spin dephasing time in MAPbI3 single crystals, the baseline that the new 21 ns value exceeds.","marker":"[15]"},{"why":"provides the universal g-factor versus band-gap relation used to identify the electron component and to estimate the g-factor spread from the pump spectral width.","marker":"[7]"},{"why":"supplies the polarization recovery components for electrons and holes in a similar lead halide perovskite that the paper uses to assign its PR Lorentzians.","marker":"[36]"},{"why":"introduces the spin inertia technique and the frequency-response formula used to extract $T_1$ from the modulation-frequency dependence.","marker":"[43]"},{"why":"establishes the lead-dominated hyperfine interaction and dynamic nuclear polarization in lead halide perovskites, supporting the Overhauser field interpretation.","marker":"[13]"}],"fun_headline_variants":["Longest perovskite spin dephasing: 21 ns","Perovskite spins outlast laser pulses, enabling spin memory","MAPbI3 crystal sets spin coherence record at 21 ns","Electron spin in perovskite stays coherent for 21 ns","Spin accumulation in MAPbI3: a new quantum resource"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The 21 ns coherence claim rests on assuming the zero-field resonant spin amplification peak has a purely Lorentzian, exponential-dephasing shape and that random magnetic fields from nuclear spins do not materially broaden it; if those nuclear field fluctuations do broaden the peak, the extracted time would be an underestimate and the lineshape model would be wrong.","fun_headline_variants_meta":{"raw":{"variants":["Longest perovskite spin dephasing: 21 ns","Perovskite spins outlast laser pulses, enabling spin memory","MAPbI3 crystal sets spin coherence record at 21 ns","Electron spin in perovskite stays coherent for 21 ns","Spin accumulation in MAPbI3: a new quantum resource"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000907,"raw_usage":{"total_tokens":3960,"prompt_tokens":1062,"completion_tokens":2898,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":678,"completion_tokens_details":{"reasoning_tokens":2813}},"tokens_in":678,"tokens_out":2898,"duration_ms":18865,"temperature":1.0,"reasoning_tokens":2813,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-08T15:16:25.141550+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the same zero-field RSA peak with a two-pulse spin echo or Hahn echo sequence: if the echo-derived homogeneous dephasing time is clearly longer than $T_2^* = 21$ ns, then the RSA Lorentzian width is broadened partly by inhomogeneous nuclear Overhauser fields and the quoted value is a lower bound. A complementary check is to compare the zero-field RSA peak width with the width predicted by Equation (5) and to repeat the measurement at fields of tens of millitesla, where the polarization recovery data show nuclear fluctuations are suppressed; a narrowing beyond the g-factor dispersion model would confirm the nuclear contribution.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"provides the original demonstration of resonant spin amplification in n-type GaAs that this paper transfers to MAPbI3."},{"cited_title":"Kirstein, N","cited_arxiv_id":null,"evidence_quote":"supplies the theoretical treatment of RSA versus spin mode locking, including the pulse-train summation in Equation (4) and the g-factor averaging procedure."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"is the reference for time-resolved Kerr rotation and Faraday/Kerr techniques, Larmor precession, and the field dependence of $T_2^*$ used throughout."},{"cited_title":"Kirstein, D","cited_arxiv_id":null,"evidence_quote":"gives the bulk MAPbI3 spin dynamics parameters, including an electron spin dephasing time of 0.4 ns, that the paper compares with its 21 ns value."},{"cited_title":"Kirstein, D","cited_arxiv_id":null,"evidence_quote":"reports the previous 11 ns electron spin dephasing time in MAPbI3 single crystals, the baseline that the new 21 ns value exceeds."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"provides the universal g-factor versus band-gap relation used to identify the electron component and to estimate the g-factor spread from the pump spectral width."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"supplies the polarization recovery components for electrons and holes in a similar lead halide perovskite that the paper uses to assign its PR Lorentzians."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"introduces the spin inertia technique and the frequency-response formula used to extract $T_1$ from the modulation-frequency dependence."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"establishes the lead-dominated hyperfine interaction and dynamic nuclear polarization in lead halide perovskites, supporting the Overhauser field interpretation."}],"review_version":1}