{"id":"7cd6b92c-782b-4454-b9ad-1bb241798d37","arxiv_id":"2411.17657","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A quantum computing formulation of electron nuclear dynamics is derived with the Fukutome unitary representation and demonstrated on H2+ with simulator circuits.","lead":"This paper adapts a method for simulating chemical reactions, called electron nuclear dynamics, so that part of its calculation can be done on a quantum computer, using a mathematical trick to make the electron states unitary. The authors design small quantum circuits for a one-electron molecule and show that the resulting electron motion matches the known analytical answer.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The H2+ 'full END/QC/VQS simulation' never uses the circuit-evaluated M and V: the reported trajectory is obtained from analytical expressions, so the central claim that the method enables the simulation is not yet demonstrated end-to-end.","rationale":"In my reading, the formalism and the individual circuit evaluations are largely sound. The factorization in Eq. (21) is appropriate for the declared model systems, and Approximations 3 and 4 are explicit scope restrictions rather than hidden errors; the reader is right that they limit generality, but for a proof of concept that is acceptable. The weakest load-bearing point is internal to the H2+ demonstration: the abstract and Section 6 claim a full END/QC/VQS simulation, but the simulation branch is executed with analytical formulas. If the objective is only to validate circuit elements for M and V, the paper already provides a reasonable check; if the objective is the claimed simulation, the missing closed-loop experiment is essential. A finite-shot integrator can accumulate errors, especially because the circuit tests are performed at parameter values different from those of the trajectory. The recommended verdict stays conditional: the manuscript should either add the closed-loop test or temper the language from 'illustrate an END/QC/VQS simulation' to 'validate the circuit evaluation of M and V.'","tokens_in":28871,"tokens_out":8184,"duration_ms":78970,"concrete_test":"Rerun the H2+ example of Section 6 replacing the analytical M,V in Eq. (49) with M,V assembled from the circuit outputs of Figs. 2-6 at n_s = 10^3, 10^4, 10^5 shots (parameters from Table 1, initial ρ0=50°, ω0=0), integrating with the same Shampine-Gordon solver. Compare the Mulliken population trajectory with Eq. (58); require the error to decrease roughly as n_s^{-1/2} and remain below a stated tolerance over several periods T=8.0844 a.u. Without such a closed-loop result, the claim to have illustrated an END/QC/VQS simulation is not supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"To support the central claim, it must be shown that shot-limited circuit evaluations of M and V can be composed across time steps into an accurate END/QC/VQS trajectory. Section 6 does not do this. After validating circuits at the fixed parameters of Table 1 (ρ=240°, ω=45°/180°), the H2+ dynamics is computed by 'numerical integration of the END/QC equations, Eq. (49)' and by the analytical solution Eq. (58), not by feeding circuit outputs into the integrator. The reported RMSE plots (Figs. 12-13) assess only static, pointwise accuracy against analytical M and V; the trajectory in Figs. 14-15 is therefore a classical simulation of the analytically known equations. The observed -1/2 slopes are the expected shot-noise scaling and say nothing about closed-loop stability or bias accumulation. Also, the actual H2+ trajectory has ρ=50° and time-dependent ω, while the circuit tests use ρ=240°, so even the pointwise validation is not at the parameters of the claimed simulation. The paper should either add a real closed-loop run or restrict the claim to validation of M/V circuit elements.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript develops a quantum-computing formulation of the electron-nuclear dynamics (END) method, called END/QC/VQS, based on Fukutome unitary representation of single-determinantal states. For a restricted family of model systems defined by four approximations, the authors show that the Fukutome unitary matrix factorizes into one-qubit SU(2) rotations, enabling a natural one-electron-per-qubit encoding. They derive the TDVP equations of motion, express the metric matrix M and gradient vector V in terms of circuit-evaluable quantities, design five ancillary-qubit circuits to evaluate these quantities, and validate the circuits on Qiskit for selected elements of M and V and for the full M and V constructs. The reported error metrics decay as the inverse square root of the number of shots, with regression slopes close to -1/2 and R^2 near 1. The paper also presents an H2+ example, including an analytic solution for its pure electronic dynamics, electron-density snapshots, and Mulliken-population oscillations.","tokens_in":29105,"tokens_out":2079,"duration_ms":21558,"significance":"If the claims are supported, the paper would be a useful proof of concept for porting END into the variational-quantum-simulator framework. The identification of Fukutome unitary matrices as tensor products of one-qubit rotations for the model systems is a genuine structural insight, and the circuit constructions with explicit analytical benchmarks are valuable as a first step. The Qiskit evaluations are reproducible in principle, and the documented shot-noise scaling is consistent with what one expects from sampling-based estimation. However, the demonstrated result is narrower than the title and abstract suggest: the paper validates the circuit-evaluated building blocks of M and V at fixed parameter values, while the H2+ dynamics itself is computed classically from the analytically known equations. The central claim of a full END/QC/VQS simulation therefore needs either an actual closed-loop demonstration or a carefully restricted claim.","major_comments":[{"comment":"The paper's central claim that END/QC/VQS 'simulates' H2+ pure electronic dynamics is not supported by the reported workflow. After validating the circuits at the fixed parameters of Table 1 (rho = 240 degrees, omega = 45/180 degrees), the trajectory in Figs. 14-15 is obtained by numerical integration of Eq. (49) with PACE and by the analytic solution Eq. (58), not by feeding shot-limited circuit evaluations of M and V into the integrator. The RMSE plots in Figs. 12-13 assess only static, pointwise accuracy against analytical M and V at a single parameter point; they say nothing about closed-loop stability, bias accumulation, or whether the hybrid loop would remain accurate over many time steps. To support the stated claim, the authors should either present a genuine closed-loop run in which circuit-evaluated M and V are used at each time step, or restrict the paper's claim to validation of the circuit elements of M and V.","section":"Section 6, Figs. 14-15"},{"comment":"Even the pointwise validation is not performed at the parameters of the claimed H2+ simulation. The circuit tests use rho = 240 degrees and fixed omega = 45 or 180 degrees, whereas the H2+ dynamics starts at rho = 50 degrees and has a time-dependent omega following Eq. (58). Since the circuit outputs depend on rho and omega through the matrix elements in Eqs. (47)-(48), the observed -1/2 slopes do not by themselves establish accuracy at the actual operating point of the dynamics. The authors should justify why the chosen test point is representative, or add validation at the dynamics parameters.","section":"Section 6 and Table 1"},{"comment":"The load-bearing factorization in Eq. (21) depends on Approximations 3 and 4, which restrict particle-hole excitations to same-spin, same-unit pairs. As the authors note, Approximation 4 is exact only with well-separated units carrying localized spin-orbitals. The manuscript's framing as a general QC formulation of END is therefore too broad: the derived circuits and the one-electron-per-qubit encoding apply to the restricted model family, not to general molecules with inter-unit excitations or general basis sets. This limitation should be stated prominently in the abstract and conclusions, and the scope of the 'first installment' claims should be tightened accordingly.","section":"Section 4, Approximations 3 and 4"},{"comment":"The RMSE definitions in Eq. (59) compare circuit outputs with 'Exact' analytical values of M and V. For a variational hybrid algorithm, the quantity that matters for the dynamics is the accuracy of the propagated trajectory under repeated inexact M and V evaluations. The current static RMSE analysis is necessary but not sufficient for the claimed simulation, and the text should not describe Fig. 15 as an END/QC/VQS result without stating explicitly that the trajectory uses analytical M and V rather than circuit outputs.","section":"Section 6, Eq. (59) and Figs. 12-13"}],"minor_comments":[{"comment":"The phrase 'on-the-flight' should read 'on-the-fly', and the abstract contains 'END/QC/QVS' in the last sentence, which should be 'END/QC/VQS'.","section":"Abstract"},{"comment":"The notation in Approximation 3 is confusing: the condition '0 if z_z s_mu neq z_z s_alpha' is written with a hat on the operator but the subscripts are not defined in that equation. Clarifying the spin labels would help the reader.","section":"Section 4, Eq. (19)"},{"comment":"The word 'unocupied' is a typo for 'unoccupied'.","section":"Section 6, text below Eq. (39)"},{"comment":"The definitions of the h-tilde coefficients could be written more explicitly; in particular, the text should state which combination of h_alpha_alpha and h_mu_mu gives h_Z tilde, since the caption of Fig. 9 uses a slightly different expression with h_mu_mu - h_mu_mu.","section":"Section 6, Eq. (46)"},{"comment":"The caption says the figure shows SDs for 'the individual circuits and of the M matrix and V vector', but the main text also refers to Fig. 11 for SDs of RMSEs in Figs. 12-13. Please clarify which quantities are plotted in Fig. 11.","section":"Section 6, Fig. 11 caption"}],"recommendation":"major_revision","confidential_remarks":"The paper is a reasonable proof of concept, but the gap between the validated circuit elements and the claimed H2+ simulation is substantial and should be resolved before publication. The authors may either add a genuine closed-loop demonstration, even for a few time steps, or explicitly reframe the manuscript as a validation of circuit-evaluated M and V components. I would also encourage the authors to move the scope limitations from Section 4 into the abstract, since the current title and abstract promise more than the model-system analysis delivers. The manuscript is not circular and does not fit a reject path, but the central claim currently overstates what is demonstrated."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First, the new thing: this is the first END/QC/VQS formulation, and using Fukutome's unitary representation instead of the non-unitary Thouless form is a sensible move for quantum circuits. Under the stated approximations, the factorization into one-qubit rotations and the natural one-electron-per-qubit encoding are clearly derived, and the five circuits for M and V are worked out in detail. The Qiskit validation against analytical expressions behaves as expected, and the −1/2 slopes are consistent with shot noise; it's good that the paper reports the raw error distributions, not just point estimates. The H2+ example is a nice illustration of symmetry-breaking-driven electron transfer, and the analytical solution matching the numerical integration is reassuring.\n\nWhere it overclaims: the H2+ \"full END/QC/VQS simulation\" never uses the circuit-evaluated M and V. The trajectory in Figs. 14–15 comes from numerical integration of the analytical equations or the closed-form solution, not from feeding shot-limited circuit outputs into the integrator. The circuit tests are at ρ=240°, ω=45°/180°, while the actual dynamics starts at ρ=50° with time-dependent ω. So the paper demonstrates the circuit elements are correct at fixed parameters, but not that the closed-loop hybrid scheme works. The −1/2 scaling is the expected statistical behavior, not a dynamical result. The paper itself is transparent about the approximations (same-spin, same-unit excitations, exact only for well-separated units), which is good, but the word \"full\" is doing too much work.\n\nAlso, no code or data are shipped, which makes it hard to check the circuits independently, though the derivation is detailed enough to reimplement. The OCR corruption in the provided text may not be the authors' fault, but the published version should have clean equations.\n\nI'd send it to peer review because the formalism is new, the derivations are coherent, and the circuit-element validation is solid; but the authors should either add a genuine closed-loop run or restrict the claim to validation of the M/V elements. It's a proof of concept, and as such it's useful for people working on VQS or END. I wouldn't cite it myself in the next year unless I worked on this exact problem.","headline":"First END/QC/VQS formulation with a clean Fukutome-based derivation, but the claimed H2+ simulation is actually classical integration of analytically known equations.","tokens_in":29611,"tokens_out":1644,"would_cite":false,"duration_ms":14476,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["03.67.Lx","31.15.-p"],"model":"deepseek-v4-flash","headline":"END/QC/VQS is established as a quantum computing formulation of electron nuclear dynamics: its five circuits evaluate the metric matrix and gradient vector, with errors that scale as the inverse square root of the number of shots, and the…","keywords":["electron nuclear dynamics","time-dependent variational principle","Fukutome unitary representation","variational quantum simulator","quantum computing","single-determinantal state","quantum circuits","H2+ dynamics"],"falsifier":"For a diatomic molecule described with a basis set larger than the minimal two orbitals per unit, compute the exact metric matrix from the full Fukutome unitary matrix and compare it with the tensor-product-factored form used by the circuits; any appreciable difference would falsify the factorization claim, and the circuits would need to be replaced or supplemented.","tokens_in":28669,"feed_emoji":"⚛️","tokens_out":11678,"duration_ms":92788,"temperature":0.7,"pith_summary":"Electron nuclear dynamics (END) is a variational, non-adiabatic method that simulates chemical reactions by evolving frozen Gaussian wave packets for nuclei and a single-determinantal wavefunction for electrons, through symplectic equations $\\mathbf{M}\\dot{\\xi} = \\mathbf{V}$ involving a metric matrix $\\mathbf{M}$ and an energy gradient vector $\\mathbf{V}$. This paper tries to establish that those two quantities can be evaluated on a quantum computer, by recasting the electron state in a unitary representation that factorizes into one-qubit rotations for a class of model systems. If true, it opens a route to run END dynamics on quantum hardware with one electron per qubit and with circuit errors that shrink like $O(n_s^{-1/2})$ as the number of shots grows. The authors demonstrate the construction with five quantum circuits and a full pure-electronic simulation of $\\mathrm{H}_2^+$, where the electron density and atomic populations oscillate with the HOMO-LUMO gap.","feed_headline":"Quantum circuits run electron nuclear dynamics on H2+","feed_subtitle":"One electron per qubit; errors drop as shots to the -1/2 power.","key_machinery":"The central machinery is the factorization of the Fukutome unitary matrix $\\mathbf{u}(\\lambda)\\in U(K)$, a Lie-group element generated by particle-hole pair operators, into a tensor product of one-qubit matrices $\\mathbf{u}_{\\mu\\alpha}(\\lambda_{\\mu\\alpha})\\in SU(2)$ under the model-system approximations. Each one-qubit matrix is written as a rotation $\\mathbf{R}_{\\mu\\alpha}(\\rho_{\\mu\\alpha},\\omega_{\\mu\\alpha}) = \\mathbf{R}_z(\\omega_{\\mu\\alpha})\\mathbf{R}_y(2\\rho_{\\mu\\alpha})\\mathbf{R}_z(-\\omega_{\\mu\\alpha})$, expressed through two real variational parameters $\\rho_{\\mu\\alpha}$ and $\\omega_{\\mu\\alpha}$. This factorization is what allows the assignment of one electron to each qubit and underlies the five quantum circuits (Figures 2-6) that evaluate the basic elements $\\tilde{M}_{jk}$ and $\\tilde{V}_{j,k}$ of $\\mathbf{M}$ and $\\mathbf{V}$ by measuring an ancilla qubit.","core_discovery":"The central claim is that the END/QC/VQS formalism evaluates the metric matrix $\\mathbf{M}$ and the energy gradient vector $\\mathbf{V}$ of the END equations of motion on a quantum computer, and that this enables a simulation of the pure electronic dynamics of $\\mathrm{H}_2^+$. The route is to replace the non-unitary single-determinantal representation of the original END with the Fukutome unitary representation, whose unitary operator fits directly into quantum gates. For the model systems defined by four approximations (two-electron chemical units, minimal basis sets, and spin- and unit-conserving particle-hole excitations), the Fukutome unitary matrix factorizes into a tensor product of one-qubit $SU(2)$ rotations, each further decomposable into Euler rotations $\\mathbf{R}_z(\\omega)\\mathbf{R}_y(2\\rho)\\mathbf{R}_z(-\\omega)$. This yields a natural encoding of one electron per qubit and five quantum circuits, each with one ancillary qubit, that evaluate the non-zero elements of $\\mathbf{M}$ and $\\mathbf{V}$ through the expectation value of a measured Pauli operator. In log-log plots against the number of shots $n_s$, the mean absolute errors, root mean square errors, and standard deviations of these evaluations all decrease linearly with slope $-1/2$, i.e. they scale as $O(n_s^{-1/2})$. The paper closes with an END/QC/VQS simulation of $\\mathrm{H}_2^+$ whose density snapshots and atomic populations show periodic intramolecular electron transfer with period $2\\pi/\\Delta$, where $\\Delta$ is the HOMO-LUMO energy gap.","pith_inferences":["A natural stress test would be to gradually delocalize the model-system orbital basis while monitoring the factorization error; the breakdown of the tensor-product form at some inter-unit distance would mark the practical boundary of the one-electron-per-qubit encoding.","The slope $-1/2$ is exactly what classical shot-noise statistics predicts for an unbiased estimator, so the circuit results can be read as evidence that the circuits are essentially sampling the correct expectation values rather than being limited by algorithmic error.","An alternative END/QC route hinted at in the paper, keeping the original non-unitary representation by expressing it as a linear combination of unitaries, would, if developed, provide a complementary formulation that may behave differently under noisy hardware.","Because the $\\mathrm{H}_2^+$ dynamics is triggered by a spatial symmetry breaking in the initial wavefunction, the END/QC/VQS formalism could be used to study time-dependent symmetry breaking in small molecules, connecting to existing classical END studies."],"forward_implications":["For model systems with two-electron units, the END/QC/VQS equations require only as many qubits as electrons, plus one ancilla, so quantum resources grow linearly with system size in this regime.","If the $O(n_s^{-1/2})$ error scaling persists on real hardware, users can trade shots for accuracy in computing $\\mathbf{M}$ and $\\mathbf{V}$, with the classical integration of the symplectic equations done off-device.","The $\\mathrm{H}_2^+$ example shows a chemically meaningful observable, the electron-transfer period $2\\pi/\\Delta$ equal to the inverse HOMO-LUMO gap, emerging directly from the quantum-circuit evaluation of the equations of motion.","The same circuit framework, generalized to full electron-nuclear dynamics and general basis sets, is the stated path to applying END to larger molecules on quantum computers."],"supporting_citations":[{"why":"Supplies the real-parameter time-dependent variational principle that defines the metric matrix $\\mathbf{M}$ and gradient vector $\\mathbf{V}$.","marker":"[2]"},{"why":"Defines the END method and its symplectic TDVP equations of motion that are being ported to quantum computing.","marker":"[3]"},{"why":"Provides the original non-unitary single-determinantal representation that is replaced by the Fukutome unitary form.","marker":"[5]"},{"why":"The VQS hybrid quantum/classical scheme and base circuit that the five END/QC circuits adapt.","marker":"[17]"},{"why":"Supplies the unitary representation of single-determinantal states, the key replacement that makes END fit into quantum gates.","marker":"[19]"},{"why":"Computes the basis-function integrals in the classical task of the hybrid algorithm.","marker":"[10]"},{"why":"The quantum-circuit simulator used to execute the five circuits and produce the error-scaling data.","marker":"[43]"}],"fun_headline_variants":["Quantum circuits simulate H2+ electron dynamics via END","Electron nuclear dynamics on quantum hardware: H2+ simulation","Quantum END: one electron per qubit, error scales as -1/2","First QC formulation of END: H2+ dynamics on a quantum simulator","Quantum circuits run END on H2+ with error dropping as shot^-1/2"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The factorization of the Fukutome unitary matrix into one-qubit rotations assumes that particle-hole excitations occur only within the same spin and within the same chemical unit; if real molecules need inter-unit or spin-flipping excitations, the factorization, the one-electron-per-qubit encoding, and the five circuits no longer apply.","fun_headline_variants_meta":{"raw":{"variants":["Quantum circuits simulate H2+ electron dynamics via END","Electron nuclear dynamics on quantum hardware: H2+ simulation","Quantum END: one electron per qubit, error scales as -1/2","First QC formulation of END: H2+ dynamics on a quantum simulator","Quantum circuits run END on H2+ with error dropping as shot^-1/2"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000757,"raw_usage":{"total_tokens":3466,"prompt_tokens":1147,"completion_tokens":2319,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":763,"completion_tokens_details":{"reasoning_tokens":2224}},"tokens_in":763,"tokens_out":2319,"duration_ms":15414,"temperature":1.0,"reasoning_tokens":2224,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T11:54:16.301459+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For a diatomic molecule described with a basis set larger than the minimal two orbitals per unit, compute the exact metric matrix from the full Fukutome unitary matrix and compare it with the tensor-product-factored form used by the circuits; any appreciable difference would falsify the factorization claim, and the circuits would need to be replaced or supplemented.","supporting_citations":[{"cited_title":"Geometry of The Time-Dependent Variational Principle in Quantum Mechanics; Springer-Verlag: New York, 1981","cited_arxiv_id":null,"evidence_quote":"Supplies the real-parameter time-dependent variational principle that defines the metric matrix $\\mathbf{M}$ and gradient vector $\\mathbf{V}$."},{"cited_title":"Time-dependent theoretical treatments of the dynamics of electrons and nuclei in molecular systems","cited_arxiv_id":null,"evidence_quote":"Defines the END method and its symplectic TDVP equations of motion that are being ported to quantum computing."},{"cited_title":"Unrestricted Hartree-Fock Theory and Its Applications to Molecules and Chemical Reactions","cited_arxiv_id":null,"evidence_quote":"Supplies the unitary representation of single-determinantal states, the key replacement that makes END fit into quantum gates."},{"cited_title":"Efficient electronic integrals and their generalized derivatives for object oriented implementations of electronic structure calculations","cited_arxiv_id":null,"evidence_quote":"Computes the basis-function integrals in the classical task of the hybrid algorithm."}],"review_version":1}