{"id":"c5e9fe7c-1043-4aa9-bb7c-94b606b35d13","arxiv_id":"2510.23303","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"New ab initio potential energy surfaces for CaF+Ca show a barrierless ground-state Ca-isotope exchange and an excited (2)2A' surface lying >1000 cm^-1 below the ground-state asymptote.","lead":"This paper maps the energy surfaces for interactions between laser-coolable CaF molecules and Ca atoms, covering nine electronic states. It predicts a barrierless calcium-isotope exchange reaction and finds an excited state that dips more than 1000 cm^-1 below the ground-state energy asymptote.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"MRCI asymptotic gap not validated; the >1000 cm^-1 below-asymptote dip may be a size-consistency artifact.","rationale":"The reader's weakest assumption focused on the active-space sensitivity broadly. I agree that the hand-tuned active space is a concern, but the more precise and directly testable issue is whether the MRCI calculation reproduces the asymptotic excitation energies that set the zero of the energy scale. The entire 'below ground asymptote' claim is a relative energy statement, and size-consistency errors in MRCI are a documented risk. Without a reported check of the asymptote, the headline claim is not robust. The proposed test is inexpensive and would either validate the feature or show it is an artifact. This does not reject the paper—the ground-state PES and isotope-exchange analysis may still be valid—but it makes the central excited-state pathway claim conditional on a numerical check, which is exactly the reader's conditional verdict.","tokens_in":14773,"tokens_out":6674,"duration_ms":59017,"concrete_test":"From the same MRCI calculation, extract the energy difference between the (X)2A' and (2)2A' states at the largest computed R (e.g., 50 bohr) and compare with the experimental Ca(3P)-Ca(1S) splitting of 15315 cm^-1; also compare the (X)2A' energy with the 2Π states against the CaF A2Π-X2Σ+ gap of 16490 cm^-1. If either deviation exceeds about 200 cm^-1, then the >1000 cm^-1 dip in Fig. 5 is within the method error and the central claim is unsubstantiated.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The paper's central claim is that the (2)2A' surface lies more than 1000 cm^-1 below the CaF(2Σ+)+Ca(1S) asymptote (Sec. III.D.2, Fig. 5), and this is used to argue for spin-orbit-driven rovibronic mixing and a non-adiabatic reaction pathway. The placement of this surface relative to the ground asymptote is determined entirely by the MRCI calculation with a hand-selected active space (Sec. II). MRCI is not size-consistent, and the chosen active space (CaF HOMO + four LUMOs + Ca 4s/4p/3d) may describe Ca(3P) and Ca(1S) with unequal accuracy. The paper does not report the MRCI-computed asymptotic Ca(3P)-Ca(1S) splitting, nor does it state whether the plotted surfaces were shifted to the experimental value of 15315 cm^-1. If the MRCI gap is overestimated by even a few percent, the >1000 cm^-1 dip below the ground asymptote could be an artifact. This is the weakest link in the chain from ab initio data to the conclusion of an excited-state chemical pathway.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports ab initio potential energy surfaces for the CaF+Ca system, computed with CCSD(T) for the ground X2A' state and CASSCF/MRCI for nine excited states arising from the lowest three asymptotes, within the rigid-rotor approximation for CaF. The authors build 2D surfaces for the ground state and the (2)2A' state, extract Legendre moments and long-range C6 coefficients, and analyze isotope-exchange reactions. They find a barrierless ground-state atom-exchange path and claim that the (2)2A' surface, correlated with CaF(2Σ+)+Ca(3P), lies more than 1000 cm^-1 below the ground-state asymptote, which they propose may enable spin-orbit-driven non-adiabatic reactivity with metastable Ca(3P).","tokens_in":15116,"tokens_out":3968,"duration_ms":40130,"significance":"If the excited-state crossing is robust, this paper provides a valuable starting point for ultracold collision and photoassociation studies of CaF+Ca. The ground-state PES and Legendre components are directly usable in scattering calculations, and the C6 coefficients are derived from independent polarizability integrals, not fitted to the target result. The monomer benchmark calculations agree well with experiment. However, the central excited-state claim currently rests on an unvalidated MRCI asymptotic gap and a hand-selected active space, so the significance of the paper for the ultracold chemistry community depends on that validation.","major_comments":[{"comment":"The claim that the (2)2A' surface lies more than 1000 cm^-1 below the CaF(2Σ+)+Ca(1S) asymptote is central to the paper. This placement is set entirely by the MRCI calculation with the active space described in Sec. II. The manuscript does not report the MRCI-computed Ca(3P)-Ca(1S) asymptotic splitting, nor does it state whether the plotted surfaces were shifted to the experimental value of 15315 cm^-1. MRCI is not size-consistent, and the chosen active space may describe Ca(3P) and Ca(1S) with unequal accuracy; a few-percent error in the asymptotic gap would erase the >1000 cm^-1 dip. Please provide the computed atomic excitation energy at the same active-space/basis level, compare it with experiment, and state whether the surfaces were shifted. A sensitivity test with a slightly larger active space would also bound the uncertainty.","section":"Sec. II / Sec. III.D.2 (Fig. 5)"},{"comment":"The barrierless ground-state isotope-exchange conclusion is obtained from a 2D PES in which the Ca-F-Ca angle is fixed at the equilibrium value (~138°) while r1 and r2 are varied. A barrier may appear for other angles. The Conclusions state that 'no barriers have been found along the reaction coordinates,' which overstates the evidence from this fixed-angle cut. Please either scan the bending angle, or explicitly restrict the claim to the fixed-angle geometry and adjust the abstract/conclusions accordingly.","section":"Sec. III.D.1 (Fig. 4)"},{"comment":"The proposal that spin-orbit-driven rovibronic mixing opens a non-adiabatic pathway from CaF(2Σ+)+Ca(3P) to the ground electronic state is not substantiated by any calculation of spin-orbit coupling matrix elements or non-adiabatic coupling terms. As written, it is a plausible hypothesis, not a conclusion from the PES data. The authors themselves note that no direct surface crossings between entrance and exit channels were found. Recommend either softening the language to clearly label this as speculation, or adding an estimate of the spin-orbit coupling between (X)2A' and (2)2A' to support the proposed mechanism.","section":"Sec. III.D.2"}],"minor_comments":[{"comment":"Typos: 'chmeical' in Introduction; 'timer' for trimer in Sec. II; 'Winger' for Wigner in Sec. III.D.1; 'crosss' and 'indicateing' in Sec. III.D.2. Please proofread.","section":"Throughout"},{"comment":"The notation 'α Cad2 CaF' is ambiguous. Use α_Ca d_CaF^2 to indicate the product of the calcium polarizability and the square of the CaF dipole moment.","section":"Eq. (3)"},{"comment":"The description of the optimal active space gives the orbitals included but not the number of active electrons. Specify the active-electron count to make the calculation reproducible.","section":"Sec. II"},{"comment":"The caption does not identify the line styles or colors used for the various 2Σ, 2Π, 4Σ, and 4Π states. Add a legend or a table of line styles.","section":"Fig. 2"},{"comment":"The ΔE values are presumably experimental asymptotic splittings. Clarify whether the calculations are referenced to these values or use the ab initio computed splittings; this is related to the major comment on asymptotic-gap validation.","section":"Table I"}],"recommendation":"major_revision","confidential_remarks":"This is a solid ab initio study with a potentially interesting excited-state crossing. The main risk is that the >1000 cm^-1 below-asymptote dip is an artifact of the MRCI asymptotic gap. The requested benchmark and active-space sensitivity test should be mandatory before publication. The ground-state barrierless claim also needs to be qualified to the fixed-angle cut. With these revisions, the paper could be suitable for publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: this is a useful, competently executed ab initio study that fills a real gap. The CaF+Ca surfaces are new, the isotope-exchange exothermicities (1-8 cm^-1) are a clean, experimentally meaningful prediction, and the observation that the (2)2A' state dips below the ground asymptote is the kind of qualitatively interesting feature that justifies a careful look. The monomer benchmarks against experiment are good, and the C6 coefficients come from polarizability integrals rather than fitted to the target result, so the long-range part is on solid ground.\n\nThe soft spots are the ones the reader flagged. There are no error bars or basis-set extrapolations, and the MRCI active space is described as 'optimal' without a sensitivity test. The ground-state barrierless conclusion rests on a fixed-angle 2D cut, which is a standard approximation but should be labeled as such. And the excited-state pathway is explicitly not checked for barriers, so the non-adiabatic reaction argument is speculative, though the authors are honest about that.\n\nThe stress-test concern about the asymptotic gap is worth taking seriously. The unshifted MRCI surfaces are put on an absolute scale, and the >1000 cm^-1 dip below the ground asymptote depends on the calculated Ca(3P)-Ca(1S) separation. The paper doesn't report the MRCI value against the experimental 15315 cm^-1, so we can't tell whether the dip is an artifact of size-consistency error. This is a legitimate worry, not a manufactured one. That said, a few percent error in the gap would still likely leave the dip negative, just not as deep. The fix is easy: report the asymptotic gaps, shift to experimental values if needed, and test a couple of active spaces.\n\nWho's this for? People working on ultracold molecule-atom collisions or photoassociation with CaF. It's a serious contribution, not a desk reject. I'd send it to referees with a request to check the asymptotic placement and the active-space sensitivity. I'd cite it myself if I worked on this system.","headline":"Solid first PESs for CaF+Ca; the excited-state dip below the ground asymptote is the one claim I'd want stress-tested before trusting.","tokens_in":15545,"tokens_out":2384,"would_cite":true,"duration_ms":25074,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["31.15.Ar","34.20.Cf","34.50.Lf"],"model":"deepseek-v4-flash","headline":"For the ultracold CaF+Ca pair, an excited electronic state of the Ca2F trimer dips more than 1000 cm-1 below the ground-state collision threshold, opening a spin-orbit-driven route to atom-exchange chemistry.","keywords":["potential energy surfaces","ultracold collisions","CaF molecule","calcium atom","isotope exchange reactions","multireference configuration interaction","spin-orbit coupling","atom-exchange chemistry"],"falsifier":"A converged multireference calculation with an enlarged active space (for example, including additional Rydberg or all valence orbitals) that moves the (2)2A' minimum above the ground asymptote, or a spectroscopic search that fails to find a Ca2F bound state below the CaF+Ca(1S) threshold in the linear geometry, would falsify the submerged-surface claim; an ultracold collision measurement that finds no enhanced reaction products when Ca is prepared in the metastable 3P state would also test it.","tokens_in":1623,"feed_emoji":"🧪","tokens_out":3969,"duration_ms":155661,"temperature":0.7,"pith_summary":"This paper computes the potential energy surfaces that govern ultracold collisions between laser-cooled CaF molecules and calcium atoms. The ground-state surface is deep, anisotropic, and permits a barrierless isotope-exchange reaction: exchanging a heavier calcium isotope into the molecule releases 1-8 cm-1, while the reverse is endothermic. The central finding is that an excited state, (2)2A', correlated with Ca in its metastable 3P state, dips more than 1000 cm-1 below the ground-state asymptote in the linear Ca-F-Ca geometry. Because this excited state carries spin-orbit coupling from the 3P atom, the authors argue it can mix with ground-state bound and scattering states, opening a non-adiabatic reaction pathway. If correct, these surfaces give experimenters the input needed for scattering calculations and photoassociation searches.","feed_headline":"Excited Ca2F state dips 1000 cm-1 below ground threshold","feed_subtitle":"New surfaces predict a barrierless isotope swap and an excited-state dip that could drive ultracold reactions.","key_machinery":"The central objects are two-dimensional surfaces V(R, theta) of the Ca2F trimer in Jacobi coordinates, with CaF treated as a rigid rotor for the ground state and relaxed for the excited state. The ground X 2A' surface uses CCSD(T); the nine excited surfaces use MRCI with a hand-selected active space (CaF HOMO, four lowest unoccupied orbitals, plus Ca 4s/4p/3d). Surfaces are expanded in Legendre polynomials to quantify anisotropy, and long-range C6,0 and C6,2 coefficients derive from dynamic polarizabilities. The load-bearing feature is the submerged (2)2A' surface: its dip below the ground asymptote, combined with spin-orbit coupling from metastable Ca(3P), is the proposed mechanism coupling","core_discovery":"The central discovery is that one of nine ab initio electronic states of Ca2F, the (2)2A' state arising from CaF(2Σ+)+Ca(3P), crosses below the ground-state CaF(2Σ+)+Ca(1S) asymptote by more than 1000 cm-1 at a linear geometry. The authors interpret this as enabling spin-orbit-mediated mixing between excited and ground rovibronic levels and the ground scattering continuum, potentially funneling metastable-channel population into ground-state products. They note that no direct crossings between entrance and exit surfaces were found and the excited-channel reaction was not tested for barrierlessness, so the pathway is a motivated possibility rather than a calculated rate. For the ground state,","pith_inferences":["If the submerged surface is physical, photoassociation spectroscopy near the CaF+Ca(1S) threshold should reveal bound or quasibound levels of Ca2F in the linear geometry; their positions would directly test the predicted dip.","A dedicated calculation of the spin-orbit coupling matrix element between (2)2A' and X 2A' would turn the qualitative pathway into a reactivity estimate; the paper does not compute this coupling.","The same active-space approach could be applied to AlF+Al, SrF+Sr, and BaF+Ba to see whether even deeper submerged surfaces are expected for heavier laser-coolable fluorides."],"forward_implications":["Ground-state CaF+Ca collisions are predicted to allow a barrierless isotope-exchange reaction, exothermic only when the product CaF contains the heavier calcium isotope, releasing 1-8 cm-1.","Because the exothermicity is far below the molecular vibrational spacing (581 cm-1) but above rotational spacings, product molecules should remain vibrationally cold but rotationally hot, allowing final-state-resolved detection.","The submerged (2)2A' surface suggests that spin-orbit coupling can mix the excited state with ground-state internal levels and unbound scattering states, offering a non-adiabatic pathway for reactions initiated in the CaF + Ca(3P) channel.","The published Legendre components and C6 coefficients provide quantitative input for coupled-channel scattering calculations of ultracold CaF+Ca collisions.","The isotope-exchange exothermicity (a few kelvin) is far above the characteristic van der Waals energy E* = 292 microkelvin, so the reaction is not in the quantum-threshold regime."],"fun_headline_variants":["Excited Ca2F state dips 1000 cm-1 below ground asymptote","New surfaces may enable ultracold isotope swap in CaF+Ca","Ca2F excited well found 1000 cm-1 below ground limit","Metastable Ca2F state may mix with ground scattering states"],"cache_read_input_tokens":16896,"weakest_assumption_plain":"The hand-selected active space used for the excited-state MRCI calculations—CaF's HOMO plus four unoccupied orbitals plus Ca's 4s, 4p, and 3d—is what places the (2)2A' surface below the ground asymptote; if a larger or differently constructed active space shifts that surface upward, the submerged-state claim becomes a calculation artifact.","fun_headline_variants_meta":{"raw":{"variants":["Excited Ca2F state dips 1000 cm-1 below ground asymptote","New surfaces may enable ultracold isotope swap in CaF+Ca","Ca2F excited well found 1000 cm-1 below ground limit","Metastable Ca2F state may mix with ground scattering states"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000794,"raw_usage":{"total_tokens":3338,"prompt_tokens":757,"completion_tokens":2581,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":501,"completion_tokens_details":{"reasoning_tokens":2499}},"tokens_in":501,"tokens_out":2581,"duration_ms":19637,"temperature":1.0,"reasoning_tokens":2499,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-04T07:57:13.383920+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A converged multireference calculation with an enlarged active space (for example, including additional Rydberg or all valence orbitals) that moves the (2)2A' minimum above the ground asymptote, or a spectroscopic search that fails to find a Ca2F bound state below the CaF+Ca(1S) threshold in the linear geometry, would falsify the submerged-surface claim; an ultracold collision measurement that finds no enhanced reaction products when Ca is prepared in the metastable 3P state would also test it.","supporting_citations":[],"review_version":1}