{"id":"b8dea8c2-a48b-4586-9ca7-9f56a6c167a3","arxiv_id":"2501.12966","paper_version":2,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"First measurement of the 2s-1s exciton interaction strength and first observation of a 2s-1s biexciton in WSe2 and MoSe2 monolayers.","lead":"Researchers used ultrafast laser pulses to measure how excited (2s) excitons interact with ground (1s) excitons in atomically thin WSe2 and MoSe2, finding the interaction is nearly as strong as between two 1s excitons. They also detected a new bound state, the 2s-1s biexciton, and measured how tightly it binds.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 2s-1s biexciton claim rests on an unverified two-level assumption: the zero-crossing of the 2s cross-circular Stark shift is taken to equal the biexciton binding energy, but multilevel background terms could shift this crossing and bias the extracted values.","rationale":"The reader's weakest_assumption identifies exactly the load-bearing point: the 2s-1s binding energy is extracted from a zero-crossing that is assumed to be a pure two-level biexciton effect. I agree with that assessment. The paper's own co-circular analysis (Eq. (1), Eq. (2)) demonstrates that the 2s shift contains multiple contributions, including a 1/δ1s^2 term from the virtual 1s density. In the cross-circular configuration, that term does not change sign and will displace the zero-crossing unless it is negligible. No argument is given for why it is negligible for the 2s state, and the intensity independence of the crossing cannot distinguish a background that is linear in intensity. The 1s result is validated against literature and therefore limits the bias for the 1s-1s biexciton, but the 2s case is new and has no independent cross-check. The proposed reanalysis of the existing Zenodo data is concrete and would settle whether the reported binding energies are correct. If the reanalysis shows the background is negligible, the original ACCEPT stands. Until then, the key quantitative claim is conditional on that check. This is not an ad hominem or a disagreement with consensus; it is a request for an internally consistent multilevel analysis that the paper currently lacks. The rest of the paper, including the repulsive 2s-1s interaction ratio and the 1s-1s benchmarking, is solid and reproducible in principle, so I do not recommend REJECT or UNVERDICTED.","tokens_in":17454,"tokens_out":16680,"duration_ms":180399,"concrete_test":"Using the published Zenodo data, re-fit the cross-circular 2s detuning curves with Δ2s/I = A/(δ1s − E_bind) + B/δ1s + C/δ1s^2, fixing C from the co-circular b2s1s and U2s1s values (Tab. I) and B from a2s1s, with signs appropriate for the attractive cross-valley interaction. If the best-fit E_bind differs from the zero-crossing by more than the quoted ±1.0/±0.6 meV, the zero-crossing interpretation is biased and the reported binding energies need revision.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central quantitative claim, binding energies of (24.2 ± 1.0) meV in WSe2 and (29.7 ± 0.6) meV in MoSe2 for the 2s-1s biexciton, is read off from the zero-crossing of the 2s optical Stark shift in the cross-circular configuration (Fig. 4(c)). This interpretation assumes the 2s shift is dominated by the two-level coupling |2s> ↔ |2s,1s'>, so that the sign change occurs exactly at δ1s = E_bind. The paper does not derive this from a multilevel calculation. In the co-circular data (Eq. (2), Fig. 3), the same pump creates a virtual 1s density whose interaction with the 2s state contributes a sizeable b2s1s/δ1s^2 term; in cross-circular this interaction is attractive and gives a 1/δ1s^2 background redshift that does not change sign. Additional 1/δ terms arise from the pump coupling to the 2s' polarization and to higher biexciton manifolds such as |2s,2s'>. The observed intensity independence of the zero-crossing does not rule out such backgrounds, since both the BX coupling and the background scale linearly with pump intensity. The 1s-1s validation against photoluminescence values is reassuring, but the 2s state has a different XP enhancement (a2s1s/a1s1s up to 1.77) and its BX transition dipole is not independently calibrated, so the bias could be comparable to or larger than the quoted uncertainties.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript reports femtosecond pump-probe measurements of the excitonic optical Stark effect on the 1s and 2s exciton resonances in charge-neutral WSe2 and MoSe2 monolayers. In the co-circular configuration, the detuning-dependent shift of each resonance is fitted with a two-parameter form Δλ/I = a/δ + b/δ², yielding the 1s-1s and 2s-1s interaction strengths and their ratio. In the cross-circular configuration, the sign change of the 1s and 2s shifts as a function of pump detuning is interpreted as the signature of biexciton resonances, giving 1s-1s binding energies consistent with literature and 2s-1s binding energies of 24.2 ± 1.0 meV (WSe2) and 29.7 ± 0.6 meV (MoSe2). The authors claim the first observation of a bound 2s-1s biexciton state and report that the repulsive 2s-1s interaction is only slightly larger than the 1s-1s one.","tokens_in":17818,"tokens_out":13428,"duration_ms":135600,"significance":"The co-circular analysis is careful and the ratio U2s1s/U1s1s is a robust quantity: it is formed from two fits that share the same virtual 1s density, so systematic errors largely cancel. The 1s-1s BX binding energies reproduce independent photoluminescence measurements, which is a valuable external check. If the 2s-1s BX interpretation survives closer scrutiny, the observation would be a significant advance for Rydberg-exciton physics in TMDCs and would provide a benchmark for biexciton theory. The authors also make the data openly available on Zenodo and are transparent about the model dependence of absolute U values. The principal weakness is that the central zero-crossing interpretation for the 2s state is not backed by a multilevel calculation.","major_comments":[{"comment":"The central claim of a 2s-1s biexciton rests on the assumption that the cross-circular 2s optical Stark shift is dominated by the two-level coupling |2s⟩ ↔ |2s,1s′⟩, so that the zero-crossing in Fig. 4(c) occurs exactly at δ1s = E_bind(2s-1s). This assumption is not derived: the SM contains the SBE calculation only for the co-circular configuration (Eqs. (15)-(17)), and no multilevel expression is given for the cross-circular case. In the same SBE framework, the pump-induced virtual 1s′ density (∝ 1/δ1s²) interacts attractively with the 2s state and would produce a background redshift that does not change sign, and additional 1/δ terms from coupling to other manifolds (for example |2s,2s′⟩) can contribute as well. Any such background displaces the zero-crossing, and the observed intensity independence does not rule it out because both the biexciton coupling and the background scale linearly with pump intensity. The agreement of the 1s-1s BX values with previous work is reassuring, but it does not transfer quantitatively to the 2s state, which has a different exciton-photon enhancement (a2s1s/a1s1s up to 1.77, Table I) and whose biexciton dipole is not independently calibrated. Please quantify the background shift at δ ≈ E_bind (for example with the measured U and n1s) or include such terms in the fit before claiming the binding energies.","section":"Biexcitonic state / Fig. 4(c)"},{"comment":"The displayed formula for the oscillator strength reads fλ = |dcv|² (2/π)(2n−1)³ (rBλ)^−2, which makes f increase with n and directly contradicts the statement in the main text that the oscillator strength decreases rapidly for higher Rydberg states. The standard 2D-hydrogen result has (2n−1)³ in the denominator. Please correct this equation and confirm that the numerical calculations (Table IV and the virtual-density expressions) use the denominator form; as printed, the formula would invalidate the neglect of higher Rydberg states in Eq. (2).","section":"Supplemental Material, Eq. (8)"}],"minor_comments":[{"comment":"The heading contains a typo: 'Adressing' should be 'Addressing'.","section":"2s-1s interaction"},{"comment":"The factor-of-three difference between the SBE and OBE virtual-density calibrations is discussed only in the SM; since Table I presents absolute U1s1s values as benchmarks and the text claims the first measurement of U1s1s in WSe2, a sentence in the main text noting this model dependence would help readers.","section":"Table I and SM 'OBE: Virtual exciton density'"},{"comment":"The experimental ratio U2s1s/U1s1s for MoSe2 (1.08 ± 0.12) and the SBE value (1.40 ± 0.18) differ by more than one combined standard error; the word 'agreement' should be qualified in the discussion.","section":"Table I"},{"comment":"Error bars on the plotted shifts and a description of how the zero-crossing and its uncertainty were obtained would strengthen the central figure.","section":"Fig. 4(c)"}],"recommendation":"major_revision","confidential_remarks":"The co-circular data analysis is solid and the ratio claim is likely to survive. My main concern is the cross-circular zero-crossing interpretation: it is a genuine gap, not a presentation issue. If the authors can provide a numerical bound on the background terms or fit them explicitly, I would be willing to accept a revised version. The Eq. (8) formula error in the SM should be corrected regardless."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First, the new results: the paper measures the 2s–1s repulsive interaction strength (co-circular) and reports a 2s–1s biexciton in cross-circular for WSe2 and MoSe2. The 1s–1s benchmark values reproduce previous work, the data are publicly available, and the fitting to Eq. (2) with fixed slopes is robust. The central ratio U2s1s/U1s1s is clean—it comes from the same virtual density, so calibration errors cancel—and the values (1.19, 1.08) are close to the SBE predictions (1.33, 1.40). This part I find convincing.\n\nThe soft spot is the biexciton binding energy extraction. The zero-crossing of the 2s shift in cross-circular is read as the two-level resonance between |2s> and the 2s–1s biexciton. But the paper never shows that other terms are negligible. The same pump that creates the virtual 1s density in the co-circular case also acts in cross-circular, where the 2s–1s interaction is attractive and gives a 1/δ^2 term that does not change sign, plus there are 1/δ terms from coupling to other polarizations and higher biexciton manifolds. The claim that the crossing is intensity-independent doesn't remove this worry, since both the two-level term and the background scale linearly with pump intensity. The 1s–1s BX values match PL, which is reassuring, but the 2s state lacks an independent calibration: its XP enhancement already disagrees with theory in MoSe2 (1.77 vs 1.51), so the BX dipole could be off too.\n\nI don't think this sinks the paper. The co-circular results stand on their own, and the 2s–1s BX is a plausible first observation with qualitative support (Autler–Townes splitting in WSe2). But the quoted 0.6–1.0 meV uncertainties on the binding energies are probably too tight if a few-meV background is present. A referee should ask for a multilevel estimate of the cross-circular shift, or at least a quantitative bound on the background terms.\n\nBottom line: this is a careful, honest experimental paper with a solid ratio result and a plausible but under-defended new biexciton. Worth sending to review; the authors should be asked to strengthen the BX analysis. I'd bring it to a reading group for the TMDC exciton crowd.","headline":"Careful Stark-shift study of 2s–1s exciton interactions; the interaction-strength ratio is solid, but the 2s–1s biexciton binding energies rely on an undefended two-level assumption.","tokens_in":18456,"tokens_out":4175,"would_cite":true,"duration_ms":40999,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["71.35.-y","78.47.-p"],"model":"deepseek-v4-flash","headline":"Coherent pump-probe measurements in WSe2 and MoSe2 monolayers show a repulsive 2s-1s exciton interaction nearly as strong as the 1s-1s interaction and reveal a bound 2s-1s biexciton with binding energies of 24.2 meV and 29.7 meV.","keywords":["optical Stark effect","biexciton","exciton-exciton interaction","transition-metal dichalcogenide monolayers","Rydberg excitons","semiconductor Bloch equations","WSe2","MoSe2"],"falsifier":"Measure the 2s-1s biexciton binding energy by two-photon absorption; if it disagrees with the detuning at which the 2s Stark shift crosses zero (24.2 meV in WSe2, 29.7 meV in MoSe2), the central claim is wrong.","tokens_in":17203,"feed_emoji":"⚛️","tokens_out":18450,"duration_ms":133720,"temperature":0.7,"pith_summary":"This paper uses a pump pulse tuned below the 1s exciton resonance to create a fleeting, virtual population of 1s excitons in WSe2 and MoSe2 monolayers and probes how it shifts the energy of the 2s exciton. In co-circular polarization the shift is a blueshift, showing that same-valley 2s-1s interactions are repulsive and only slightly stronger (about 10–20%) than 1s-1s interactions, as the semiconductor Bloch equations predict. In cross-circular polarization the 2s shift changes sign at a specific pump detuning, which the authors identify as the binding energy of a 2s-1s biexciton: 24.2 meV in WSe2 and 29.7 meV in MoSe2. This gives the first measurement of a 2s-1s biexciton and a benchmark for theories of how Rydberg excitons interact.","feed_headline":"2s-1s biexciton appears in WSe2 and MoSe2 monolayers","feed_subtitle":"Pump-probe spectroscopy measures its binding energy at 24.2 meV in WSe2 and 29.7 meV in MoSe2.","key_machinery":"The load-bearing mechanism is the optical Stark effect in the low-intensity limit, decomposed via the semiconductor Bloch equations into an exciton-photon term scaling as $1/\\delta$ and an exciton-exciton term scaling as $1/\\delta^2$, where $\\delta$ is the pump detuning from the 1s state. Fitting the measured 1s and 2s shifts as a function of detuning yields the prefactors $a$ and $b$, whose ratio $b_{2s1s}/b_{1s1s}$ gives the interaction-strength ratio with systematic errors cancelling. In the cross-circular geometry, the pump couples the $\\lvert ns\\rangle$ exciton to the $\\lvert ns,1s'\\rangle$ biexciton, and the two-level dressed-state shift changes sign when the detuning equals the biexciton binding energy; the zero-crossing of the 2s shift therefore serves as a direct measure of the 2s-1s biexciton binding energy.","core_discovery":"We use the optical Stark effect in the low-intensity limit to show that the shift of the 2s exciton in WSe2 and MoSe2 monolayers separates into an exciton-photon term and an exciton-exciton term, distinguished by their $1/\\delta$ and $1/\\delta^2$ detuning dependence. Fitting this detuning dependence yields a 2s-1s interaction strength only 19% (WSe2) and 8% (MoSe2) larger than the 1s-1s strength, in agreement with semiconductor Bloch equation calculations based on the Rytova-Keldysh potential. In the cross-circular configuration, the 2s light shift changes sign at a pump detuning that is independent of intensity; we identify this zero-crossing with the transition to a bound 2s-1s biexciton and extract binding energies of 24.2±1.0 meV (WSe2) and 29.7±0.6 meV (MoSe2). To our knowledge this is the first observation of a 2s-1s biexciton, and the non-resonant method isolates the bright-bright intervalley biexciton without populating dark states.","pith_inferences":["The same zero-crossing technique could measure 3s-1s and 4s-1s biexciton binding energies, revealing whether the near-equality in MoSe2 (29.7 vs 29.1 meV) is accidental or reflects a weak dependence on principal quantum number.","Because the ratio $U_{2s1s}/U_{1s1s}$ is insensitive to the absolute calibration of virtual density (which differs by a factor of three between the two methods discussed in the supplement), future studies of interaction strengths in doped or moiré systems could rely on such ratios rather than absolute values.","The discrepancy between theory and experiment for the absolute 1s-1s interaction strength in MoSe2 (theory overestimates by about a factor of three) suggests the approximate 2D hydrogen wave functions used in the semiconductor Bloch equation calculation miss part of the screening physics.","A cavity or waveguide geometry that enhances the coherent signal might recover the predicted attractive 2s-2s interaction, which the paper reports was obscured by incoherent absorption in a blue-detuned pump configuration."],"forward_implications":["The measured 2s-1s biexciton binding energies (24.2 meV in WSe2, 29.7 meV in MoSe2) become the first benchmarks against which theoretical models of biexcitons in two-dimensional semiconductors must be tested.","The near-unity ratio of 2s-1s to 1s-1s interaction strength (1.19 in WSe2, 1.08 in MoSe2) shows that Rydberg-exciton interaction enhancement is modest in these materials, informing proposals for strong optical nonlinearities.","The non-resonant cross-circular method isolates the bright-bright intervalley biexciton without populating dark states, and the WSe2 value is close to previously reported bright-dark biexciton binding energies, indicating similar binding for the two configurations.","The work opens perspectives for the coherent manipulation of Rydberg polaritons in optical cavities, as the authors state."],"supporting_citations":[{"why":"Supplies the method of extracting exciton-exciton interaction strengths from the optical Stark shift and the virtual-density calibration used here.","marker":"[20]"},{"why":"First demonstration of the intervalley biexcitonic optical Stark effect in a monolayer, which the cross-circular measurement extends to the 2s state.","marker":"[22]"},{"why":"Previous observation of biexcitonic optical Stark effects in MoSe2 including the sign change and Autler-Townes splitting that the zero-crossing measurement builds on.","marker":"[23]"},{"why":"Gives the exciton-photon/exciton-exciton decomposition of the light shift used to separate interaction contributions.","marker":"[34]"},{"why":"Provides the semiconductor Bloch equations that underpin the shift formula and the theoretical interaction strengths.","marker":"[36]"},{"why":"Supplies theoretical 1s-1s and 2s-1s interaction strengths and the Rytova-Keldysh screened potential for comparison with the measured ratios.","marker":"[46]"},{"why":"Used for the semiconductor Bloch equation analysis of the optical Stark shifts, yielding the theoretical shift ratios reported in Table I.","marker":"[47]"},{"why":"Provides the reference 1s-1s biexciton binding energy in WSe2 against which the cross-circular zero-crossing method is validated.","marker":"[9]"}],"fun_headline_variants":["First 2s-1s biexciton observed in TMD monolayers","2s-1s biexciton binding energies measured in WSe2 and MoSe2","Optical Stark effect uncovers 2s-1s biexciton in TMDs","New biexciton state: 2s-1s in WSe2 and MoSe2 monolayers"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The binding energy of the 2s-1s biexciton is read directly from the detuning at which the 2s optical Stark shift changes sign, which assumes that this shift is dominated by the two-level coupling to the biexciton and that no comparable background shift (for example from the direct ac Stark effect on the 2s state or from other virtual populations) displaces the zero-crossing.","fun_headline_variants_meta":{"raw":{"variants":["First 2s-1s biexciton observed in TMD monolayers","2s-1s biexciton binding energies measured in WSe2 and MoSe2","Optical Stark effect uncovers 2s-1s biexciton in TMDs","New biexciton state: 2s-1s in WSe2 and MoSe2 monolayers"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000168,"raw_usage":{"total_tokens":1255,"prompt_tokens":935,"completion_tokens":320,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":551,"completion_tokens_details":{"reasoning_tokens":221}},"tokens_in":551,"tokens_out":320,"duration_ms":3271,"temperature":1.0,"reasoning_tokens":221,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T16:35:00.985552+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the 2s-1s biexciton binding energy by two-photon absorption; if it disagrees with the detuning at which the 2s Stark shift crosses zero (24.2 meV in WSe2, 29.7 meV in MoSe2), the central claim is wrong.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the method of extracting exciton-exciton interaction strengths from the optical Stark shift and the virtual-density calibration used here."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"First demonstration of the intervalley biexcitonic optical Stark effect in a monolayer, which the cross-circular measurement extends to the 2s state."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Previous observation of biexcitonic optical Stark effects in MoSe2 including the sign change and Autler-Townes splitting that the zero-crossing measurement builds on."},{"cited_title":"Schmitt-Rink, D","cited_arxiv_id":null,"evidence_quote":"Gives the exciton-photon/exciton-exciton decomposition of the light shift used to separate interaction contributions."},{"cited_title":"Haug and S","cited_arxiv_id":null,"evidence_quote":"Provides the semiconductor Bloch equations that underpin the shift formula and the theoretical interaction strengths."},{"cited_title":"Shahnazaryan, I","cited_arxiv_id":null,"evidence_quote":"Supplies theoretical 1s-1s and 2s-1s interaction strengths and the Rytova-Keldysh screened potential for comparison with the measured ratios."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Used for the semiconductor Bloch equation analysis of the optical Stark shifts, yielding the theoretical shift ratios reported in Table I."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the reference 1s-1s biexciton binding energy in WSe2 against which the cross-circular zero-crossing method is validated."}],"review_version":1}