{"id":"25ad08b1-f7cc-4676-a06f-7f91da4e8d1d","arxiv_id":"2507.12572","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"The SHG intensity at the 1S exciton-polariton resonance in ZnSe is nonreciprocal in magnetic field for light along [111], due to interference of crystallographic and magnetic-field-induced SHG.","lead":"Researchers found that the intensity of light emitted at twice the input frequency from a zinc selenide crystal depends on the direction of an applied magnetic field, an effect called nonreciprocal second harmonic generation. This is the first time the effect has been seen in a semiconductor exciton-polariton system, and it could become a tool for probing exciton state symmetries.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Quantitative inconsistency: Eq. (4) with C/A = +0.02/T predicts I(+10T) = 1.64 and I(-10T) = 0.77, opposite to the reported doubling at -10 T and no change at +10 T; the microscopic model (A11) predicts a sub-percent effect.","rationale":"The reader's verdict of CONDITIONAL is appropriate, but the most load-bearing weakness is sharper than the line-assignment/z'-contribution issue highlighted as the weakest assumption. Even if the y' assignment is correct, the quantitative validation fails: the phenomenological fit parameters reported in Eq. (4) produce the opposite sign of nonreciprocity from the data description, and the microscopic model with the stated Δ and the measured Γ predicts a nonreciprocity two orders of magnitude weaker than observed. This is an internal inconsistency, not merely a disagreement with external expectations. The raw observation of a B-direction-dependent SHG in the [111] geometry is still plausibly real, and the [001] control supports a B-induced SHG contribution, so the paper should not be rejected outright. However, the central claim that the effect is quantitatively explained by interference of crystallographic and field-induced SHG with phases set by the exciton damping cannot be assessed until the numbers are reconciled. Re-fitting the data and clarifying the field sign convention would settle the issue. This preserves the CONDITIONAL verdict: the authors must demonstrate that a single consistent set of parameters reproduces both the sign and the magnitude of the reported nonreciprocity.","tokens_in":11420,"tokens_out":17558,"duration_ms":197711,"concrete_test":"Re-analyze the raw SHG peak intensities of the y' line from Fig. 3(b): fit I(B) to |A + CB + C'B^2|^2 allowing A, C, C' complex, and also to the microscopic expression (A9) with g, Δ, Γ as free parameters. Check whether the best-fit Re(C/A) has the sign needed to produce an increase at -10 T and no change at +10 T, and whether the best-fit g/Δ ratio can account for a ~100% effect at 10 T. In addition, verify the solenoid polarity convention used to assign +B and -B in Figs. 2 and 3, as a swapped convention would reverse the sign of the reported nonreciprocity.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central mechanistic claim rests on the fit of the y'-component SHG intensity vs magnetic field to I = |A + CB + C'B^2|^2. With the reported C/A = 0.02/T and C'/A = 7.9e-4/T^2, and taking A real, the normalized intensities at ±10 T are |1 ± 0.2 + 0.079|^2 = 1.64 and 0.77: the signal is stronger at +10 T and weaker at -10 T. The text states the opposite: 'more than doubled' at -10 T and 'no change in intensity' at +10 T. The sign of C/A (or the stated field convention) is therefore irreconcilable with the data description. Independently, the microscopic model, Eq. (A11), gives a linear-in-B intensity correction 2gμB B Γ/(Δ^2+Γ^2). Using the quoted Δ ≈ 5 meV, the measured FWHM 0.18 meV (Γ ≈ 0.09 meV), and g ≈ 1, this correction is ≈ 0.005 at B = 10 T, i.e., a 0.5% nonreciprocity, two orders of magnitude below the observed ~100% effect. Either Δ is much smaller than stated or g is much larger than typical for ZnSe; neither is discussed. The reported fit parameters and the model's quantitative predictions are thus inconsistent with the central observation, so the claimed interference mechanism, and the extracted phase, are not validated by the presented numbers.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports magnetic-field-dependent second harmonic generation (SHG) at the 1S exciton-polariton resonance in bulk ZnSe. For kω ∥ [111] with B perpendicular to k, the authors observe that the SHG intensity of the y′ exciton-polariton component is strongly nonreciprocal in B: it is roughly doubled at B = −10 T and nearly unchanged at B = +10 T, while the x′ component changes only weakly and symmetrically. The effect is absent for kω ∥ [001], where crystallographic SHG is symmetry-forbidden. The authors develop a phenomenological model in which the SHG intensity is I = |A + CB + C′B²|² and attribute the nonreciprocity to interference between crystallographic and magnetic-field-induced SHG contributions, with phases set by exciton damping. A microscopic model based on B-induced mixing of the y′ and z′ exciton states is presented in Appendix A. The paper claims the first observation of nonreciprocal SHG in semiconductor crystals and the first for exciton-polaritons.","tokens_in":11773,"tokens_out":7814,"duration_ms":96508,"significance":"If quantitatively validated, this would be a notable experimental result: a clean, spectrally resolved demonstration of magnetic-field-direction-dependent SHG at an exciton-polariton resonance, with a control geometry that supports the interference interpretation. The raw spectral observation—the nonreciprocal intensity at ±10 T in the [111] geometry—appears robust and is largely independent of the model. The paper also gives credit for including explicit symmetry arguments, rotational anisotropy data, and a microscopic model. However, the quantitative support for the claimed mechanism is currently not convincing: the reported fit parameters are internally inconsistent with the stated field dependence, and the microscopic model as written predicts an effect roughly two orders of magnitude smaller than observed. These issues are load-bearing for the central mechanistic claim, so the paper needs substantial revision before publication.","major_comments":[{"comment":"The reported fit parameters are inconsistent with the stated nonreciprocity. The text says the y′ component is “more than doubled” at −10 T and shows “no change in intensity” at +10 T, and it earlier quotes χ1/χ0 ≈ −0.024/T for the linear-in-B susceptibility ratio. However, the Fit-2 parameters reported in §IV are C/A = 0.02/T and C′/A = 7.9×10⁻⁴/T². Taking A real, Eq. (4) gives I(+10 T) = |1 + 0.2 + 0.079|² ≈ 1.64 and I(−10 T) = |1 − 0.2 + 0.079|² ≈ 0.77, i.e., the opposite asymmetry from what is described. Since the sign (and in general the phase) of C/A is exactly the quantity used to extract the relative phase of the crystallographic and magnetic-field-induced SHG contributions, this inconsistency must be resolved. Please report the fitted complex ratios C/A and C′/A unambiguously, state the sign convention for B, and show the resulting fit curve together with the data.","section":"§IV, Eq. (4), Fig. 3(b)"},{"comment":"The microscopic model as presented cannot quantitatively account for the observed effect. Using the paper's own values—Δ ≈ 5 meV, Γ ≈ 0.09 meV (half of the 0.18 meV FWHM), g ≈ 1, and B = 10 T—the linear-in-B intensity correction in Eq. (A11), 2gμB B Γ/(Δ² + Γ²), is of order 4×10⁻³, i.e., about 0.4%. The observed effect is a near doubling of the intensity, and the C′B² term in Eq. (4) is even in B and cannot produce the asymmetry. Thus the Zeeman-mixing mechanism as formulated does not explain the magnitude of the nonreciprocity. The paper must either provide values of g and Δ (with justification) that make Eq. (A11) compatible with the data, or quantitatively include the magneto-spatial dispersion terms γ1 and γ2 from Eq. (A12), or identify another mechanism. Without this, the central claim that the interference of crystallographic and magnetic-field-induced SHG explains the observation is not quantitatively validated.","section":"Appendix A, Eq. (A11)"},{"comment":"The assignment of the two magnetic-field-split lines to the x′ and y′ polariton components, and specifically the claim that the y′ state is mixed with the longitudinal z′ state while z′ does not contribute to the detected SHG, is central to the inferred phase relationship. This assignment should be checked against the measured splitting. With the quoted Δ ≈ 5 meV and g ≈ 1, the Zeeman-induced second-order shift of the y′ level is (gμB B)²/Δ ≈ 0.07 meV at B = 10 T, whereas the observed splitting between the two lines is 0.23 meV. The paper should explain whether the additional splitting comes from polariton dispersion, magneto-spatial dispersion, or other terms; otherwise the inferred phase and the extracted C/A ratio rest on an unvalidated assignment.","section":"§III, §IV, Appendix A"}],"minor_comments":[{"comment":"The definition of “nonreciprocal” should be stated explicitly. The manuscript opens with the source–detector exchange definition of reciprocity, but the experiment compares I(+B) and I(−B) at fixed k. These are related but not identical notions; please clarify the definition used and how the B-reversal measurement establishes nonreciprocity.","section":"§I"},{"comment":"The caption refers to “Fit 1” and “Fit 2” in panel (b), but the printed figure does not clearly distinguish the two curves. Please use distinct line styles and clearly label them, and consider showing the residuals of both fits.","section":"Fig. 3"},{"comment":"The sentence “We are firm, that the nonreciprocal SHG can be observed in a manifold of semiconductor crystals” should be rephrased as an evidence-based statement about the generality of the mechanism, rather than an assertion of certainty.","section":"§V"},{"comment":"There are several typographical errors, including “Among then are GaAs,” “metalic multilayers,” “surfces and interfaces,” and “due toits cubic noncentrosymmetric crystal lattice.” These should be corrected before publication.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The sign inconsistency in C/A is likely a typographical error that can be fixed, but the order-of-magnitude disagreement between the Appendix A model and the observed 100% nonreciprocity is a deeper problem. If the authors can provide a corrected fit and a quantitative microscopic mechanism (or a clear justification for much smaller Δ or larger g), the paper would be a valuable contribution. In its current form, however, the central mechanistic claim is not supported by the numbers as printed."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a genuine experimental result, but the quantitative backing as written doesn't hold together. The observation of nonreciprocal SHG in ZnSe at the 1S exciton-polariton resonance looks real, and the control geometry along [001] showing no nonreciprocity is a good check. The rotational anisotropy analysis and the symmetry argument (interference of crystallographic and field-induced SHG) are cleanly presented and the assignment of the two split lines to x' and y' is plausible.\n\nThe problems are in the numbers. First, the fit to Eq. (4) with C/A = +0.02/T and C'/A = 7.9e-4/T^2 gives I(+10T) = 1.64 and I(-10T) = 0.77 in normalized units, the opposite of the reported 'more than doubled at -10 T' and 'no change at +10 T'. Either the sign is misprinted or the field convention in the text disagrees with the fit; either way, the paper as written contradicts itself. That's a referee-level fix, but it is load-bearing for the claimed quantitative agreement.\n\nSecond, the microscopic model (Eq. A11) predicts a linear-in-B intensity correction of order 2g muB B Gamma/(Delta^2+Gamma^2). With Delta ~ 5 meV, Gamma ~ 0.09 meV, and g ~ 1, that is about 0.5% at 10 T, two orders of magnitude below the observed ~100% effect. The paper mentions magneto-spatial dispersion terms (gamma1, gamma2) but gives no estimate for them, so the Zeeman mixing mechanism as quantified does not explain the data.\n\nThe novelty claim also needs qualification: nonreciprocal SHG has been seen in CrI3 bilayers and other magnets; the new part is the semiconductor/exciton-polariton setting, not the phenomenon itself. That remains a legitimate first.\n\nBottom line: the experimental observation is probably right and interesting, and the symmetry framework is sound. But the fit parameters and the microscopic estimate need to be reconciled with the data. This deserves a serious referee, not a desk reject. If the inconsistency is a typo, it is minor; if not, the interpretation fails. I'd send it out.","headline":"A real first observation with a quantitative story that doesn't add up; send to referees but expect major revision.","tokens_in":12403,"tokens_out":3126,"would_cite":false,"duration_ms":32707,"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":"Reversing a magnetic field more than doubles the second-harmonic intensity of the y' exciton-polariton component in ZnSe; the cause is interference of crystallographic and field-induced SHG, with the phase set by exciton damping.","keywords":["second harmonic generation","nonreciprocal optics","exciton polaritons","ZnSe","Zeeman effect","external magnetic field","Voigt geometry","nonlinear magneto-optics"],"falsifier":"Measure $I(+B)-I(-B)$ for the $y'$ component as a function of temperature between 5 K and 200 K, where the 1S damping $\\Gamma$ grows. The model predicts that this difference scales as $\\Gamma/(\\Delta^2+\\Gamma^2)$, rising to a maximum when $\\Gamma\\approx\\Delta$ and then declining; observing instead a monotonic increase, or a nonreciprocity in the $k\\parallel[001]$ geometry beyond the pure $B^2$ term, would contradict the interference-plus-damping explanation and point to the magneto-spatial dispersion terms.","tokens_in":11178,"feed_emoji":"🧲","tokens_out":12380,"duration_ms":110792,"temperature":0.7,"pith_summary":"This paper reports that second harmonic generation (SHG) at the 1S exciton-polariton resonance in bulk ZnSe depends on the direction of an external magnetic field even though the light path is unchanged. In the $k \\parallel [111]$ geometry, the $y'$-polarized resonance component is more than doubled in intensity at $-10\\,\\mathrm{T}$ and shows no intensity change at $+10\\,\\mathrm{T}$, while the $x'$ component changes by only about 25%. The authors explain the nonreciprocity as interference between the crystallographic SHG, allowed by the zinc-blende point group, and a magnetic-field-induced SHG channel produced by the Zeeman mixing of the $y'$ and $z'$ exciton states. The relative phase of the two channels is controlled by the exciton damping; if damping were zero, the linear-in-field interference term would vanish and the response would be reciprocal. The paper claims this is the first observation of nonreciprocal SHG in semiconductor crystals and the first for exciton-polaritons.","feed_headline":"Reversing the magnet more than doubles light emission in ZnSe","feed_subtitle":"Interference between crystallographic and magnetic channels breaks the mirror symmetry between +B and −B.","key_machinery":"The central mechanism is a linear-in-$B$ interference between two SHG channels: the crystallographic channel $\\chi_0$ (two-photon excitation of the $|y\\rangle$ state) and the Zeeman-induced channel $\\chi_1$ (excitation of the $|z\\rangle$ state admixed into $|\\tilde y\\rangle$). The microscopic model uses the effective Hamiltonian $H=\\Delta L_z^2 + g\\mu_B \\mathbf{B}\\cdot\\mathbf{L}$ in the Cartesian basis $x\\parallel[1\\bar10]$, $y\\parallel[11\\bar2]$, $z\\parallel[111]$, which yields $\\chi_1=-(i g\\mu_B B)/(2\\hbar\\omega-E_z+i\\Gamma)\\,\\chi_0$ and $\\chi_2=0$. The relative phase between $\\chi_0$ and $\\chi_1$ is governed by the damping $\\Gamma$: the cross term in the intensity, $1+2g\\mu_B B\\Gamma/(\\Delta^2+\\Gamma^2)$, is odd in $B$ only when $\\Gamma\\neq0$. This is what converts an otherwise reciprocal $\\propto B^2$ magneto-SHG response into a nonreciprocal one. A secondary set of $k\\cdot B$ magneto-spatial dispersion terms (coefficients $\\gamma_1,\\gamma_2$ in Eq. A12 of the paper) can add a phase-locked contribution that would also break reciprocity, but the measured anisotropy diagrams indicate the Zeeman-mixing path dominates.","core_discovery":"The central claim is that nonreciprocal magnetic-field-induced SHG in ZnSe arises from the interference of the crystallographic second-order susceptibility $\\chi_0$ with a field-induced contribution $\\chi_1$, with the exciton damping $\\Gamma$ setting the phase that allows a linear-in-$B$ term to survive in the intensity. In the $k \\parallel [111]$ geometry, the magnetic field mixes the $y'$ and $z'$ exciton-polariton states as $|\\tilde y\\rangle \\approx |y\\rangle + i\\,g\\mu_B B_x/\\Delta\\, |z\\rangle$; the $z'$ state is two-photon active through $E_x^2+E_y^2$ but emits no SHG at $B=0$ because it is polarized along the light propagation direction. Phenomenologically, the second-harmonic polarization reads $P_y^{2\\omega}=(\\chi_0+\\chi_1 B_x)(E_x^2-E_y^2)+\\chi_1 B_x(E_x^2+E_y^2)$ with $\\chi_1 = -i\\,g\\mu_B B/(2\\hbar\\omega - E_z + i\\Gamma)\\,\\chi_0$ and $\\chi_2=0$. For $\\Gamma=0$ the phases of $\\chi_0$ and $\\chi_1$ differ by $\\pi/2$ and the $B$-linear term cancels in the intensity; only nonzero damping produces the cross term visible as different SHG intensities for opposite field directions. The experimental signature is that the $y'$ component's SHG intensity is more than doubled at $-10\\,\\mathrm{T}$ and unchanged at $+10\\,\\mathrm{T}$, the $x'$ component stays nearly field independent, and the $k\\parallel[001]$ geometry, where crystallographic SHG is forbidden, shows no nonreciprocity.","pith_inferences":["Beyond the paper, the predicted temperature dependence $\\Gamma/(\\Delta^2+\\Gamma^2)$ for $I(+B)-I(-B)$ could be checked by raising the lattice temperature from 5 K toward 200 K; the difference should peak near $\\Gamma\\approx\\Delta$ and then decline.","Beyond the paper, the same Hamiltonian predicts that increasing the longitudinal-transverse splitting $\\Delta$ suppresses the effect, while raising damping up to $\\Gamma\\sim\\Delta$ enhances it, giving a design rule for materials that show this nonreciprocity.","A testable extension: reverse only the light propagation direction at fixed $\\mathbf{B}$; the Zeeman mechanism predicts the same intensity asymmetry as reversing $\\mathbf{B}$ at fixed $\\mathbf{k}$, whereas magneto-spatial dispersion terms would produce a different dependence on crystal axes.","The $\\gamma_1$ magneto-spatial dispersion channel enters with the same phase as $\\chi_0$ rather than with the $i$-phase of the Zeeman term, so tilting the light direction away from [111] should shift the phase of the nonreciprocity; detecting such a shift would separate that contribution from the pure Zeeman mechanism."],"forward_implications":["Reversing the magnetic field direction with a fixed light path is equivalent to reversing the light path through the crystal, so the measured intensity difference is a true optical nonreciprocity at the exciton-polariton resonance.","Because crystallographic SHG is forbidden along $k\\parallel[001]$, that geometry shows no nonreciprocity even though field-induced SHG is present, confirming that interference of two channels, not a single field-induced channel, is required.","Tied to the 1S exciton binding energy of 20 meV, the effect is expected to survive up to roughly 200 K in ZnSe, far above the N\\'eel temperature of the antiferromagnets where nonreciprocal SHG was previously observed.","The same interference mechanism should operate in other noncentrosymmetric semiconductors with pronounced exciton-polariton effects; the paper specifically points to ZnO and Cu2O, where higher exciton binding energies could push the effect toward room temperature.","Measuring the nonreciprocity of the $y'$ component gives a direct probe of the Zeeman mixing of the $y'$ and $z'$ exciton-polariton states, i.e., of the longitudinal-transverse splitting $\\Delta$ and the exciton $g$-factor."],"supporting_citations":[{"why":"Supplies the standard electric-dipole $\\chi^{(2)}$ description, Eq. (1), and the general nonlinear optics framework used throughout.","marker":"[6]"},{"why":"Gives the symmetry analysis: in zinc-blende the $\\chi^{(2)}$ tensor has only the independent component $\\chi_{xyz}$, fixing the zero-field selection rules used to explain the $k\\parallel[001]$ silence and the $k\\parallel[111]$ six-fold anisotropy.","marker":"[26]"},{"why":"Provides the ZnSe samples and the prior characterization of their linear and nonlinear optical properties, including the exciton-polariton dispersion and SHG line position, which the present experiment relies on for resonance assignment.","marker":"[20]"},{"why":"Supplies the high-resolution SHG spectroscopy technique (femtosecond pulses, 30 $\\mu$eV spectral resolution) used for the resonant measurements.","marker":"[18]"},{"why":"Establishes the exciton-polariton upper-branch harmonic generation picture in bulk semiconductors with magnetic field, the framework the microscopic model extends to the 1S resonance in ZnSe.","marker":"[16]"},{"why":"Earlier demonstration of nonreciprocal SHG in the antiferromagnet CuB2O4; its temperature range provides the contrast for the claim that the semiconductor effect survives to much higher temperatures.","marker":"[7]"},{"why":"Independent observation of nonreciprocal SHG in CuB2O4, used as a comparison for the first-observation claim and for the mechanism being interference of crystallographic and induced signals.","marker":"[8]"},{"why":"Provides the $I=|A+CB|^2$ nonlinear magneto-optical Kerr-effect fit that the paper adapts as Fit 1 to the $y'$ field dependence.","marker":"[11]"},{"why":"Recent nonlinear magneto-optical Kerr-effect analysis of field-induced SHG at interfaces, supporting the treatment of the linear-in-$B$ susceptibility term.","marker":"[13]"}],"fun_headline_variants":["Magnet flips SHG intensity in ZnSe exciton polaritons","Nonreciprocal SHG first seen in semiconductor crystal","Interference of SHG signals causes +B/-B asymmetry in ZnSe","Magnetic field breaks mirror symmetry in ZnSe SHG"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The interpretation depends on assigning the lower-energy split line to the $x'$ exciton-polariton component and the higher-energy line to the $y'$ component, and on the assumption that the $z'$ state is not detected because it is polarized along the light direction; if that assignment, the magneto-spatial dispersion terms, or field-induced changes in the fundamental-beam polarization contributed to the detected signal, the inferred phase between the crystallographic and field-induced signals, and hence the sign and size of the nonreciprocity, would change.","fun_headline_variants_meta":{"raw":{"variants":["Magnet flips SHG intensity in ZnSe exciton polaritons","Nonreciprocal SHG first seen in semiconductor crystal","Interference of SHG signals causes +B/-B asymmetry in ZnSe","Magnetic field breaks mirror symmetry in ZnSe SHG"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000703,"raw_usage":{"total_tokens":3233,"prompt_tokens":1066,"completion_tokens":2167,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":682,"completion_tokens_details":{"reasoning_tokens":2093}},"tokens_in":682,"tokens_out":2167,"duration_ms":18759,"temperature":1.0,"reasoning_tokens":2093,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T16:45:36.578898+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure $I(+B)-I(-B)$ for the $y'$ component as a function of temperature between 5 K and 200 K, where the 1S damping $\\Gamma$ grows. The model predicts that this difference scales as $\\Gamma/(\\Delta^2+\\Gamma^2)$, rising to a maximum when $\\Gamma\\approx\\Delta$ and then declining; observing instead a monotonic increase, or a nonreciprocity in the $k\\parallel[001]$ geometry beyond the pure $B^2$ term, would contradict the interference-plus-damping explanation and point to the magneto-spatial dispersion terms.","supporting_citations":[{"cited_title":"Faraday, Experimental researches in electricity","cited_arxiv_id":null,"evidence_quote":"Supplies the standard electric-dipole $\\chi^{(2)}$ description, Eq. (1), and the general nonlinear optics framework used throughout."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the symmetry analysis: in zinc-blende the $\\chi^{(2)}$ tensor has only the independent component $\\chi_{xyz}$, fixing the zero-field selection rules used to explain the $k\\parallel[001]$ silence and the $k\\parallel[111]$ six-fold anisotropy."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the ZnSe samples and the prior characterization of their linear and nonlinear optical properties, including the exciton-polariton dispersion and SHG line position, which the present experiment relies on for resonance assignment."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the high-resolution SHG spectroscopy technique (femtosecond pulses, 30 $\\mu$eV spectral resolution) used for the resonant measurements."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes the exciton-polariton upper-branch harmonic generation picture in bulk semiconductors with magnetic field, the framework the microscopic model extends to the 1S resonance in ZnSe."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Earlier demonstration of nonreciprocal SHG in the antiferromagnet CuB2O4; its temperature range provides the contrast for the claim that the semiconductor effect survives to much higher temperatures."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Independent observation of nonreciprocal SHG in CuB2O4, used as a comparison for the first-observation claim and for the mechanism being interference of crystallographic and induced signals."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the $I=|A+CB|^2$ nonlinear magneto-optical Kerr-effect fit that the paper adapts as Fit 1 to the $y'$ field dependence."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Recent nonlinear magneto-optical Kerr-effect analysis of field-induced SHG at interfaces, supporting the treatment of the linear-in-$B$ susceptibility term."}],"review_version":1}