{"id":"3a6a3162-52d9-437c-b68f-c64591e3669a","arxiv_id":"2501.05578","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A combined optical and microwave pumping scheme concentrates TlF molecules into the J=0, F=0 hyperfine ground state, achieving a factor 20.1(4) population gain in a cryogenic beam.","lead":"The CeNTREX collaboration demonstrates a laser and microwave pumping scheme that concentrates thallium fluoride molecules into a single internal state, giving a 20-fold increase in the useful beam population. This is a key step for a future precision search for time-reversal symmetry violation in the thallium nucleus.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The quoted gain is inferred via Eq. (4) from a branching-fraction ratio bf_F=0/bf_F=1 that Appendix D admits differs significantly from Ref. [9]; with only statistical errors propagated, a 10% error in that ratio shifts G0 by ~1.9, well outside the 0.4 uncertainty.","rationale":"The reader's weakest-assumption analysis identifies the correctness of the branching fractions used in Eq. (4) as the load-bearing point, and I agree. The paper is an honest experimental report with a plausible and conservatively chosen primary method, but the central number is an inference, not a direct measurement. The sensitivity of G0 to r is large: a 10% error in r shifts G0 by ~1.9, which is five times the quoted statistical uncertainty of 0.4. The paper itself flags an unresolved disagreement with a prior published calculation, and the note in Table I provides no quantitative systematic analysis. The depletion-based check is not usable for this purpose because the authors state that microwave leakage corrupted those measurements, and the two alternative Schemes 2 and 3 are explicitly biased by dark states. Therefore no independent constraint on r exists within the paper. This does not warrant rejection — the measurement is careful and the gain is likely real — but the quoted uncertainty is incomplete, and the claim as stated is conditional on a disputed theoretical input. The reader's CONDITIONAL verdict is appropriate, and I see no reason to change it.","tokens_in":15124,"tokens_out":21745,"duration_ms":202128,"concrete_test":"Have an independent group (or independent code) recompute the P(2) F1'=3/2, F'=1 eigenstate using the Hamiltonian of Ref. [10] and again using the model/parameters of Ref. [9], and extract the ratio r=bf_F=0/bf_F=1. Then plug both r values into Eq. (4) with the measured G1=3.78. If the two r values differ by more than ~5%, the G0^(1)=20.1(4) error bar must be enlarged to include the spread; the paper should publish the systematic budget. If available, a direct microwave-optical double-resonance measurement of the J=0 F=0/F=1 population ratio after cooling would bypass the branching-fraction calculation entirely.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section IV quotes G0^(1)=20.1(4) as the central result, but no probe measures the J=0,F=0 population directly. Scheme 1 instead measures the F=1 gain G1 and multiplies by the theoretical ratio r=bf_F=0/bf_F=1 via Eq. (4): G0=3r(G1-1)+1. From the quoted numbers, G1≈3.78 and r≈2.29, so dG0/dr=3(G1-1)≈8.3. An unquantified 10% error in r changes G0 by ±1.9, about five times the reported statistical error; a factor-2 error would place G0 anywhere from ~12 to ~26. Appendix D's Table I note states that the branching fractions 'differ significantly from those in [9]' and gives only 'after correspondence ... we believe the values presented here to be correct' as justification — no systematic uncertainty is assigned. The excited state is strongly mixed in J' and F1', so the ratio is not fixed by simple angular-momentum algebra; it depends on the same Hamiltonian diagonalization that is in dispute. The depletion-based estimate (G0≈24, Appendix C) is dismissed because of microwave leakage in the detection chamber, and Schemes 2 and 3 are explicitly biased by dark states, so no in-paper cross-check constrains r. The headline population gain is therefore conditional on an unresolved, unquantified theoretical input.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports a demonstration of rotational and hyperfine cooling in a cryogenic beam of 205TlF, aimed at increasing the population in the J=0, F=0 ground hyperfine level that is used for the CeNTREX Schiff-moment search. The method uses one UV laser on the P(2) F1'=3/2, F'=1 transition of the B3Pi1-X1Sigma+ system, together with two microwave fields coupling J=1-2 and J=2-3, to funnel population into J=0. The gain is inferred indirectly: Scheme 1 (claimed as primary) measures the gain in the J=0, F=1 population via the R(0) F1'=3/2, F'=2 transition, then converts to the F=0 gain using a computed branching-fraction ratio bf_F=0/bf_F=1. The result quoted in Section IV and the abstract is G0^(1)=20.1(4), averaged over |v_perp|<2 m/s. Two other schemes (R(0) F1'=1/2, F'=1 branching method and a differential method) give 22.1(4) and 22.9(6) but are explicitly regarded as biased by dark states. Appendix C gives a depletion-based estimate of 24.1(11), dismissed as unreliable because of microwave leakage. Appendix D contains a new calculation of the B3Pi1 branching fractions; Table I notes that the values 'differ significantly' from those of Ref. [9], with the justification being correspondence with the authors of Ref. [10] rather than an independent measurement.","tokens_in":15375,"tokens_out":2682,"duration_ms":28309,"significance":"If the central claim is right, this is an important experimental milestone for CeNTREX: a population gain of ~20 in the single usable hyperfine state translates directly into improved statistical sensitivity for the Schiff-moment measurement, and the paper is the first demonstration of hyperfine-level rotational cooling in a closed-shell molecule. The paper is also valuable for its detailed description of the multi-pass UV scheme, polarization/dark-state management, and the three-way cross-check of gain diagnostics. The authors are appropriately conservative in choosing the lowest of the three gain estimates as the headline result and in flagging the dark-state bias of Schemes 2 and 3. However, the headline number is not a direct measurement of the J=0,F=0 population: it is the product of a directly measured F=1 gain and a computed branching-fraction ratio. The manuscript nowhere assigns a systematic uncertainty to that ratio, even though Appendix D states that its values disagree significantly with a prior published calculation. This is the load-bearing point for the paper's quantitative claim.","major_comments":[{"comment":"The quoted central result G0^(1)=20.1(4) relies on Eq. (4), G0 = 3 r (G1-1)+1 with r = bf_F=0/bf_F=1, and the stated 0.4 is only the statistical uncertainty from the fluorescence ratio. From the numbers given, G1 ~ 3.78 and r ~ 2.29, so dG0/dr = 3(G1-1) ~ 8.3. A 10% uncertainty in r alone would shift G0 by ~1.9, about five times the reported statistical error. The manuscript does not provide any estimate of the systematic uncertainty in r, and Appendix D explicitly states that the values in Table I 'differ significantly' from those in Ref. [9]. Because the excited state is strongly mixed in J' and F1', r is not fixed by simple angular-momentum algebra but depends on the same Hamiltonian diagonalization that is in dispute. The headline gain is therefore conditional on an unresolved theoretical input; the paper needs either an independent calibration of r, a quantified systematic error, or a differently framed claim that avoids making r load-bearing.","section":"Section IV and Eq. (4)"},{"comment":"Table I's note that the branching fractions 'differ significantly' from Ref. [9], with 'after correspondence ... we believe the values presented here to be correct' as the only justification, is not a sufficient basis for a quantitative measurement claim. The manuscript should report the actual differences between the two calculations, identify which Hamiltonian parameters cause them, and propagate the resulting uncertainty into Eq. (4). In particular, the paper should state whether the discrepancy arises from the hyperfine-mixing Hamiltonian of Ref. [10] versus Ref. [9], and whether the predicted decay branching fractions to J=0 (0.4841 and 0.5159 from the relevant excited state) are testable by any existing or proposed measurement.","section":"Appendix D and Table I"},{"comment":"The depletion-based estimate G0 ~ 24 is dismissed because of microwave leakage in the detection chamber, but this estimate is also the only in-paper cross-check that does not use the disputed branching fractions for the conversion (it uses only the sum bf_F=0+bf_F=1 through the factor in Eq. C1). If the microwave-leakage concern applies equally to the depletion measurements and to the Scheme-1 fluorescence measurements, the paper should say so explicitly; if the leakage affects only the depletion measurement, the 24.1(11) value provides some independent support for the central claim, and the reason for not using it as a systematic bound on r should be stated. As written, the manuscript leaves the reader unable to tell whether the leakage is a generic systematic that could also affect G1.","section":"Section IV and Appendix C"},{"comment":"The dark-state bias of Schemes 2 and 3 is quantified only through a phenomenological efficiency ratio epsilon1/epsilon0, and Figure 5 shows that the apparent gain depends strongly on this ratio. The paper reports G0^(2)=22.1(4) and G0^(3)=22.9(6) but gives no measured value or uncertainty for epsilon1/epsilon0. It would strengthen the paper to report the inferred epsilon1/epsilon0 from the difference between the schemes, and to check that the inferred value is consistent with the dark-state model. Without this, the claim that Scheme 1 is the 'more reliable' method rests on a qualitative argument.","section":"Section III E and Appendix B"},{"comment":"The sentence 'Given the drawbacks of employing the R(0) F1'=1/2 F'=1 transition ... we only use the result Scheme 1 for our quantitative conclusions' is in tension with the abstract and conclusion, which present 20.1(4) as an unconditional achieved gain. The conditional nature of the number (conditioned on the computed branching-fraction ratio, and on the correctness of the dark-state model for Schemes 2 and 3) should be stated in the abstract or at least in the conclusion, so that the quantitative claim is not overstated.","section":"Section IV, first paragraph"}],"minor_comments":[{"comment":"The derivation of Eq. (4) assumes that the thermal population ratio rho_1/rho_0 = 3 exactly and that the branching fractions are the only mechanism changing the F=0 population. The paper should state explicitly that hyperfine-changing collisions, off-resonant excitation of other P(2) hyperfine components, and decays from the F'=1 excited state to J=0,F=1 with subsequent microwave transfer to J=1,2 are all neglected or included in the quoted uncertainty; if they are neglected, a bound should be given.","section":"Eq. (4) and Appendix A"},{"comment":"The caption of Fig. 1(b) states 'Decays back to J=2+ are not shown', but the text says 'roughly half the decays from J'=1 end up in J=0, and nearly all of the remainder returns to J=2'. A reader would benefit from a single figure showing both the J=0 and J=2 decay channels with the branching fractions 0.484, 0.516 (and the 0.337/0.147 sub-branching), since the numerical values are central to the scheme.","section":"Section II, Fig. 1 caption"},{"comment":"The statement that the R(0) F1'=1/2, F'=1 transition 'excites from the unresolved F=0 and F=1 hyperfine levels' should be clarified: the ground-state hyperfine splitting of 13 kHz is indeed unresolved by the laser, but the branching fractions for the detection transition into the F'=1 excited state are not presented here. Since the detection transition's own branching fractions affect the interpretation of Schemes 2 and 3, a reference or table entry would be helpful.","section":"Section III D"},{"comment":"The 9-j symbol in Eq. (D4) is written with a 1 in the bottom row, which likely denotes a rank-1 tensor component; this notation is nonstandard and could be confused with a scalar. Please define the symbol (e.g., as a 9-j symbol involving q) and ensure the phase convention is stated.","section":"Appendix D, Eq. (D4)"},{"comment":"The prefactor 3*omega^3/(3 hbar c^3) is dimensionally correct but written in a redundant way; the factor 3 in the numerator and denominator cancels. This is cosmetic but could be simplified to avoid confusion with the standard expression.","section":"Appendix D, Eq. (D6)"},{"comment":"The abstract says 'a factor of 20.1(4) gain in the population of the J=0, F=0 hyperfine sublevel'. As noted in the major comments, this is an inferred quantity. Even if the systematic issue is resolved, the uncertainty should include the branching-fraction contribution; if it is not resolved, the abstract should say 'inferred'.","section":"Abstract and Conclusion"},{"comment":"The statement that 'the only loss is branching to other vibrational states which amounts to ≲1%' cites Refs. [9-11]; however, the Franck-Condon factor for B-X vibrational decay is typically a few percent for TlF, so the authors should clarify whether the 1% figure refers to the total loss per optical cycle or to some other quantity.","section":"Section II"},{"comment":"The time-dependent phase offset Delta_Omega t with Delta_Omega = 65 kHz and the claim that molecules see 'approximately two complete cycles' of the sideband variation is not derived. A short estimate of the interaction time (beam diameter / forward velocity) would make this quantitative.","section":"Appendix E"},{"comment":"The two figures show gain and depletion as functions of detuning with a grey-shaded region corresponding to +/-2 m/s. The text should explicitly state that the average over the shaded region is unweighted over the transverse-velocity distribution, or if it is weighted, how the weighting was determined.","section":"Fig. 6 and Fig. 7"},{"comment":"Ref. [9] (Norrgard et al., PRA 95, 062506 (2017)) is the prior hyperfine-structure calculation that allegedly disagrees with Table I. The manuscript should give the specific numbers from Ref. [9] for the branching fractions of the P(2) F1'=3/2,F'=1 excited state, so the reader can see the magnitude of the discrepancy without chasing the reference.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is a solid experimental demonstration with a real and important application, and the authors are careful in many respects (choosing the lowest of three gain estimates, flagging dark-state effects, providing a depletion cross-check). The central issue is purely quantitative: the headline 20.1(4) has no systematic error term for the branching-fraction ratio, and the paper itself admits that the ratio differs from a prior calculation. This is a fixable issue if the authors can either bound the discrepancy from the Hamiltonian parameters or reframe the claim, but as it stands the number is over-quoted. I would not recommend reject: the measurement logic is sound and the issue is the propagation of an acknowledged uncertainty. The appropriate outcome is major revision, with the specific requirement that the branching-fraction systematic be quantified and either included in the uncertainty or removed from the headline. If the authors can provide that quantification, the paper is close to publishable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: the cooling works and the experiment is real, but the headline gain carries a systematic uncertainty from the branching-fraction ratio that is not quantified, so the 20.1(4) number is not as precise as it looks.\n\nThe paper adapts optical pumping and microwave repumping to a closed-shell diatomic and demonstrates, for the first time, hyperfine-resolved rotational cooling. That is a genuine technical step. The multi-pass UV geometry, phase-modulation to cover Doppler shifts, polarization switching to destabilize dark states, and the three detection schemes show careful engineering. Choosing the most conservative scheme (Scheme 1) as primary and explaining why the other two are biased by dark states is the right call. The depletion-based estimate (G0 ~ 24) is also mentioned but not used, and the reason (microwave leakage) is plausible.\n\nThe soft spot is the conversion from the measured F=1 gain to the F=0 gain. Equation (4) uses the branching-fraction ratio r = bf_F0/bf_F1, with r ~ 2.29. The derivative dG0/dr = 3(G1 - 1) ~ 8.3, so a 10% error in r shifts G0 by about 1.9, roughly five times the reported statistical error of 0.4. Table I says the branching fractions differ significantly from Ref. [9], and the justification is \"after correspondence we believe the values presented here to be correct\"—no error bar. Since the excited state is strongly mixed in J' and F1', the ratio is not fixed by simple angular momentum algebra; it depends on the same Hamiltonian diagonalization that is in dispute. So the precision claim rests on an unquantified theoretical input. This is not fatal to the main point—the gain is certainly large, and likely >10 even if r is off by a factor of two—but it does mean the quoted uncertainty should not be taken at face value.\n\nAlso, no probe measures the J=0, F=0 population directly; every gain number is inferred. The authors are transparent about that. But the absence of a direct cross-check on r leaves the headline precision conditional.\n\nWho this is for: anyone working on molecular beams, optical cycling, or precision measurements with diatomics. It is a meaningful step for CeNTREX. I would bring it to reading group and cite it for the cooling scheme. For peer review: yes, it deserves a serious referee. The review should focus on requiring a systematic uncertainty for the branching fractions, or a direct measurement of the F=0 population, before the 20.1(4) can be taken as a calibrated number. With that fixed, this is a solid paper.","headline":"The cooling scheme works and is worth citing, but the headline gain's precision is hostage to an unquantified branching-fraction ratio.","tokens_in":15975,"tokens_out":2923,"would_cite":true,"duration_ms":28144,"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":"Rotational-hyperfine cooling of a cryogenic thallium-fluoride beam, driven by one ultraviolet laser and two microwave fields, concentrates the population into the $J=0, F=0$ ground state with a measured gain of 20.1(4).","keywords":["rotational cooling","hyperfine cooling","thallium fluoride","cryogenic molecular beam","optical pumping","dark states","Schiff moment","time-reversal violation"],"falsifier":"Measure the branching-fraction ratio directly by driving the $P(2)$ transition and counting decays into each resolved $J=0$ hyperfine level, or compare the inferred gain with a probe that resolves the 13 kHz splitting; if $\\mathrm{bf}_{F=0}/\\mathrm{bf}_{F=1}$ differs from $0.337/0.147$, the reported 20.1(4) gain changes proportionally.","tokens_in":14900,"feed_emoji":"⚛️","tokens_out":10935,"duration_ms":85792,"temperature":0.7,"pith_summary":"The paper reports a method to dramatically increase the population of the single rotational-hyperfine state of thallium fluoride that the experiment uses to search for time-reversal violation. At the beam's 6.3 K rotational temperature, only about one percent of molecules sit in the usable $J=0, F=0$ level; the rest are spread over dozens of states. By optically pumping the $P(2)$ ultraviolet transition while microwaves mix the $J=1,2,3$ rotational levels, most of the population is driven into $J=0$, with decay branching preferentially filling $F=0$ over $F=1$. The measured gain is 20.1(4) over the transverse velocity range accepted by the experiment's electrostatic lens, sufficient for its projected statistical sensitivity.","feed_headline":"TlF beam gains 20-fold in its one usable state","feed_subtitle":"One ultraviolet laser and two microwave beams concentrate a cryogenic molecule beam into the state needed for a time-reversal-violation…","key_machinery":"The cooling cycle is driven by the $P(2)$ $\\tilde{F}_1'=3/2$, $F'=1$ transition at 271.75 nm, which excites $J=2$ molecules to a $B\\,^3\\Pi_1$ level whose decays land about 48% in $J=0$ and 52% back in $J=2$; microwaves at 26.6 and 40.0 GHz continuously repopulate $J=2$ from $J=1$ and $J=3$. The hyperfine selectivity comes from the excited-state decay branching fractions, which put about 70% of the $J=0$ decays into $F=0$ rather than $F=1$; the paper computes these fractions by diagonalizing the strongly mixed $B\\,^3\\Pi_1$ Hamiltonian. Rapidly switched laser and microwave polarizations break the coherent dark states that would otherwise trap population and slow the scattering rate.","core_discovery":"The central claim is that rotational-hyperfine cooling of a cryogenic-beam molecule is practical and effective: a single ultraviolet laser plus two microwave fields transfer the majority of the Boltzmann distribution of TlF into the $|J=0, F=0\\rangle$ ground state, yielding a gain of $G_0 = 20.1(4)$ in that level's population. This is the first demonstration of hyperfine cooling in a closed-shell molecule, and it works despite long-lived coherent dark states that would otherwise stall the optical pumping. The paper argues that this gain is sufficient for the experiment's projected sensitivity and that straightforward upgrades, more laser power and a second pump transition, could raise it to roughly 40.","pith_inferences":["The same optical-pumping-plus-microwave-mixing cycle should transfer to other heavy closed-shell diatomics whose excited-state hyperfine structure is resolved in the optical transition, so a gain near 20 is not specific to TlF.","Because the reported $F=0$ gain is obtained by multiplying the measured $F=1$ gain by the computed branching-fraction ratio, an independent measurement of that ratio—for instance by resolving the 13 kHz ground-state splitting in a microwave-optical double-resonance experiment—would directly test the headline number.","If the microwave-leakage effect flagged in the depletion measurements were eliminated, the depletion-based gain of 24.1(1.1) would provide a sharp cross-check of the branching-fraction calculation."],"forward_implications":["The experiment can now operate with the beam population concentrated in the useful $J=0, F=0$ state, reaching its projected statistical sensitivity for the $^{205}$Tl Schiff-moment search.","Increasing the cooling-laser power to about 500 mW should fully deplete the $J=1,2,3$ states and raise the gain above 25.","Adding a second laser on the $R(0)$ transition to empty the $J=0, F=1$ manifold could push the total gain to nearly 40.","The method provides a template for rotational-hyperfine cooling of other heavy closed-shell molecules used in symmetry-violation searches.","The demonstration that rapid polarization switching destabilizes dark states at 1 MHz rates is directly useful for optical cycling in other molecular species."],"supporting_citations":[{"why":"Defines the cryogenic beam source, the 6.3 K rotational temperature, the single usable sublevel, and the projected sensitivity the gain is benchmarked against.","marker":"[1]"},{"why":"Provides the analytically compiled branching fractions for decays of the B^3Pi_1 hyperfine levels used in the cooling scheme.","marker":"[8]"},{"why":"Gives the hyperfine structure of the B^3Pi_1 state and prior optical-cycling predictions whose branching fractions the present calculation revises.","marker":"[9]"},{"why":"Supplies the B^3Pi_1 Hamiltonian eigenstates used to compute the new branching fractions in Table I.","marker":"[10]"},{"why":"Underpins the assumption that vibrational branching losses in the X to B cycle are at the one-percent level.","marker":"[11]"},{"why":"Establishes the dark-state destabilization mechanism that the paper exploits by rapid polarization switching.","marker":"[12]"},{"why":"Supports the use of alternating polarizations to maintain a radiative force during optical cycling.","marker":"[13]"},{"why":"Demonstrates microwave mixing of rotational states, the basis of the J=1<->2 and J=2<->3 transfers.","marker":"[14]"},{"why":"Provides the bound on photon scattering rate from the number of excited and ground sublevels that sets the required laser interaction time.","marker":"[15]"}],"fun_headline_variants":["20.1-fold gain from TlF hyperfine cooling","Single UV laser and two microwaves funnel TlF into one state","TlF beam cooling: 20x population in the needed state","20-fold state boost from laser-microwave cooling of TlF","Hyperfine cooling compresses TlF beam into single quantum state"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the computed decay branching fractions from the $P(2)$ excited state into the $J=0, F=0$ and $J=0, F=1$ ground levels are correct, since the reported gain scales with the ratio of those two numbers.","fun_headline_variants_meta":{"raw":{"variants":["20.1-fold gain from TlF hyperfine cooling","Single UV laser and two microwaves funnel TlF into one state","TlF beam cooling: 20x population in the needed state","20-fold state boost from laser-microwave cooling of TlF","Hyperfine cooling compresses TlF beam into single quantum state"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001103,"raw_usage":{"total_tokens":4564,"prompt_tokens":872,"completion_tokens":3692,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":488,"completion_tokens_details":{"reasoning_tokens":3617}},"tokens_in":488,"tokens_out":3692,"duration_ms":23721,"temperature":1.0,"reasoning_tokens":3617,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T21:14:29.146452+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the branching-fraction ratio directly by driving the $P(2)$ transition and counting decays into each resolved $J=0$ hyperfine level, or compare the inferred gain with a probe that resolves the 13 kHz splitting; if $\\mathrm{bf}_{F=0}/\\mathrm{bf}_{F=1}$ differs from $0.337/0.147$, the reported 20.1(4) gain changes proportionally.","supporting_citations":[{"cited_title":"The 205Tl and 19F nuclear spins give rise to hyperfine substructure in the rotational states","cited_arxiv_id":null,"evidence_quote":"Defines the cryogenic beam source, the 6.3 K rotational temperature, the single usable sublevel, and the projected sensitivity the gain is benchmarked against."},{"cited_title":"Timgren,Progress towards a measurement of time- reversal symmetry violation in thallium fluoride, Ph.D","cited_arxiv_id":null,"evidence_quote":"Provides the analytically compiled branching fractions for decays of the B^3Pi_1 hyperfine levels used in the cooling scheme."},{"cited_title":"Norrgard, E","cited_arxiv_id":null,"evidence_quote":"Gives the hyperfine structure of the B^3Pi_1 state and prior optical-cycling predictions whose branching fractions the present calculation revises."},{"cited_title":"Courageux, A","cited_arxiv_id":null,"evidence_quote":"Supplies the B^3Pi_1 Hamiltonian eigenstates used to compute the new branching fractions in Table I."},{"cited_title":"Hunter, S","cited_arxiv_id":null,"evidence_quote":"Underpins the assumption that vibrational branching losses in the X to B cycle are at the one-percent level."},{"cited_title":"After this interaction region, the molecules travel downstream∼40 cm to a detection region","cited_arxiv_id":null,"evidence_quote":"Establishes the dark-state destabilization mechanism that the paper exploits by rapid polarization switching."},{"cited_title":"Meijer and B","cited_arxiv_id":null,"evidence_quote":"Supports the use of alternating polarizations to maintain a radiative force during optical cycling."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Demonstrates microwave mixing of rotational states, the basis of the J=1<->2 and J=2<->3 transfers."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the bound on photon scattering rate from the number of excited and ground sublevels that sets the required laser interaction time."}],"review_version":1}