pith. sign in

arxiv: 2605.30967 · v1 · pith:TBTCU2LVnew · submitted 2026-05-29 · 🪐 quant-ph

Evaluating higher-order product formulae for molecular ground-state energy estimation

Hiromu Abe , Keita Kanno , Ryosuke Kimura , Masahiko Kamoshita , Kosuke Mitarai This is my paper

Pith reviewed 2026-06-28 22:00 UTC · model grok-4.3

classification 🪐 quant-ph
keywords product formulaequantum phase estimationmolecular ground-state energyhydrogen chainsgate countTrotterizationfault-tolerant quantum simulation
0
0 comments X

The pith

A new fourth-order product formula achieves the lowest total gate count for molecular ground-state energy estimation near chemical accuracy on hydrogen chains.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper compares deterministic higher-order product formulae inside quantum phase estimation for computing molecular ground-state energies. It uses one-dimensional hydrogen chains from H2 to H15 as benchmarks and tracks both total gate count and the depth of RZ-rotation layers needed to reach chosen target errors. Among earlier constructions the eighth-order formula minimizes cost at chemically relevant accuracies, yet the authors also build a new fourth-order formula that undercuts all others in total gate count near chemical accuracy while cutting RZ-layer depth. This matters because gate count directly limits the feasibility of fault-tolerant quantum chemistry simulations when non-Clifford operations carry their own cost.

Core claim

The paper claims that a newly constructed fourth-order product formula achieves the lowest total gate count among the formulae considered for all H-chain instances near chemical accuracy and over much of the 0.1-10 mHa target-error window for most instances, while also reducing the RZ-layer depth; among previously published formulae the eighth-order construction minimizes both cost metrics at a chemically relevant target error, and increasing formal order does not automatically reduce total cost.

What carries the argument

The perturbative method that estimates the eigenvalue error induced by each product formula, allowing the cost of the corresponding phase-estimation procedure to be evaluated at larger system sizes without full simulation; the new fourth-order product formula carries the reported efficiency gain.

If this is right

  • The tenth-order formula can require more total gates than the eighth-order formula near chemical accuracy.
  • The new fourth-order formula also reduces RZ-layer depth relative to the higher-order alternatives.
  • Product-formula order for molecular simulations must be chosen by explicit benchmarking rather than by assuming higher order is always cheaper.
  • The relative advantage of any given order changes with the target error window.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The observed advantage of the new fourth-order formula may persist for molecular systems whose Hamiltonians have different locality or interaction structure than linear hydrogen chains.
  • In hardware models where non-Clifford gates are generated locally at low cost, the optimal choice of product-formula order may shift even further toward lower orders with favorable constants.
  • Systematic optimization of the free parameters inside a fixed-order product formula could produce additional reductions in gate count for specific error targets.

Load-bearing premise

The perturbative method used to estimate the eigenvalue error induced by each product formula is accurate enough to reliably predict the gate cost of the corresponding phase-estimation procedure across the tested system sizes and error targets.

What would settle it

A direct numerical computation of the actual eigenvalue error produced by applying the new fourth-order formula inside a phase-estimation circuit on one of the larger hydrogen chains (H10 or H15) and comparing that error to the perturbative prediction.

Figures

Figures reproduced from arXiv: 2605.30967 by Hiromu Abe, Keita Kanno, Kosuke Mitarai, Masahiko Kamoshita, Ryosuke Kimura.

Figure 1
Figure 1. Figure 1: FIG. 1. For H-chains at the target energy error [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2. For H-chains at the target energy error [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3. As in Fig [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4. As in Fig [PITH_FULL_IMAGE:figures/full_fig_p003_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5. Values of [PITH_FULL_IMAGE:figures/full_fig_p005_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: FIG. 6. Relationship between the total gate count [PITH_FULL_IMAGE:figures/full_fig_p005_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: FIG. 7. System-size scaling of the per-step quantities for the H-chain instances considered in this work: (a) [PITH_FULL_IMAGE:figures/full_fig_p007_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: FIG. 8. Best (lowest gate count) product formula for each H [PITH_FULL_IMAGE:figures/full_fig_p007_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: FIG. 9. Distribution of the phase-estimation error [PITH_FULL_IMAGE:figures/full_fig_p008_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: FIG. 10. Relative difference between two evaluations of the [PITH_FULL_IMAGE:figures/full_fig_p010_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: FIG. 11. Reference values of the error-model parameters [PITH_FULL_IMAGE:figures/full_fig_p011_11.png] view at source ↗
read the original abstract

We evaluate deterministic higher-order product formulae for molecular ground-state energy estimation. Motivated by recent fault-tolerant architectures in which non-Clifford operations may be generated more locally and cheaply than in conventional assumptions, we re-examine such formulae as practical candidates for quantum chemistry. Using one-dimensional hydrogen chains from $\mathrm{H}_2$ to $\mathrm{H}_{15}$ as benchmarks, we estimate both the total gate count and the depth of $R_Z$-rotation layers required to reach a target energy error. To make this comparison feasible at larger system sizes, we use a perturbative method to estimate the eigenvalue error induced by each product formula and thereby evaluate the cost of the corresponding phase-estimation procedure. Among the previously considered formulae, the eighth-order construction introduced by Morales et al. [M. E. S. Morales et al., "Greatly improved higher-order product formulae for quantum simulation," arXiv:2210.15817v2 (2024)] minimizes both cost metrics in the benchmark at a chemically relevant target error. We also find that increasing the formal order does not automatically reduce the total cost: near chemical accuracy, the tenth-order formula introduced in the same work can be less efficient than the eighth-order one. Motivated by this observation, we construct a new fourth-order formula; it achieves the lowest total gate count among the formulae considered for all H-chain instances near chemical accuracy and over much of the 0.1-10 mHa target-error window for most instances, while also reducing the $R_Z$-layer depth. These results clarify how deterministic higher-order product formulae should be selected for molecular ground-state energy estimation.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript evaluates deterministic higher-order product formulae for Trotterized simulation in quantum phase estimation of molecular ground-state energies. Using one-dimensional hydrogen chains H2–H15 as test cases, it compares total gate counts and RZ-layer depths required to reach target energy errors in the 0.1–10 mHa range. A perturbative estimator is used to predict the eigenvalue error induced by each formula, which is then fed into the phase-estimation resource formula. Among existing formulae the eighth-order construction of Morales et al. performs best; the authors introduce a new fourth-order formula that yields the lowest total gate count near chemical accuracy for all tested instances and reduces RZ depth over much of the error window.

Significance. If the perturbative error estimates prove reliable, the work supplies concrete, system-size-dependent guidance on formula selection for fault-tolerant quantum chemistry. It demonstrates that formal order does not monotonically improve cost and supplies an explicit new fourth-order construction whose performance is competitive on the chosen benchmarks. The emphasis on both total gate count and RZ-layer depth is a useful addition to the resource-estimation literature.

major comments (2)
  1. The central cost claims rest on feeding perturbative eigenvalue-error estimates into the phase-estimation resource formula. No section reports a direct numerical check of these estimates against exact diagonalization or exact Trotter simulation of the effective Hamiltonian, even for the smallest chains (H2–H4) where such validation is computationally trivial. Without this check it is impossible to know whether higher-order perturbative corrections or non-perturbative effects alter the relative ordering of the formulae near chemical accuracy.
  2. The perturbative estimator is described only at leading order. The manuscript should state the explicit form of the leading correction (including any dependence on the number of Trotter steps and the norm of the commutators) and quantify the expected size of the neglected terms for the H-chain Hamiltonians at the step counts used.
minor comments (2)
  1. The abstract states that the new fourth-order formula “achieves the lowest total gate count … for all H-chain instances near chemical accuracy,” yet the text should clarify whether this holds after rounding the number of steps to the nearest integer required by the phase-estimation formula.
  2. Figure captions and axis labels should explicitly indicate whether the plotted gate counts include the cost of the phase-estimation circuit or only the simulation cost per Trotter step.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their careful reading and constructive feedback on our manuscript. We address each major comment below and will incorporate the suggested revisions.

read point-by-point responses

  1. Referee: The central cost claims rest on feeding perturbative eigenvalue-error estimates into the phase-estimation resource formula. No section reports a direct numerical check of these estimates against exact diagonalization or exact Trotter simulation of the effective Hamiltonian, even for the smallest chains (H2–H4) where such validation is computationally trivial. Without this check it is impossible to know whether higher-order perturbative corrections or non-perturbative effects alter the relative ordering of the formulae near chemical accuracy.

    Authors: We agree that direct numerical validation would strengthen the central claims. Although the perturbative estimator follows from standard Trotter-error analysis, we will add comparisons of the perturbative predictions against exact diagonalization and full Trotter simulations for H2–H4 in a new appendix of the revised manuscript. These checks will confirm whether the relative ordering of the formulae is preserved near chemical accuracy. revision: yes

  2. Referee: The perturbative estimator is described only at leading order. The manuscript should state the explicit form of the leading correction (including any dependence on the number of Trotter steps and the norm of the commutators) and quantify the expected size of the neglected terms for the H-chain Hamiltonians at the step counts used.

    Authors: We will revise the manuscript to state the explicit leading-order correction formula, including its dependence on the number of Trotter steps and the norms of the relevant commutators. We will also add quantitative bounds on the size of the neglected higher-order terms evaluated at the Trotter step counts used for the H-chain benchmarks. revision: yes

Circularity Check

0 steps flagged

No significant circularity; evaluation uses independent perturbative estimates and explicit benchmarks

full rationale

The paper's central results rest on applying a perturbative estimator for eigenvalue error induced by each product formula, then feeding those estimates into a phase-estimation resource formula to obtain gate counts and RZ depths on H-chain instances. This estimator is an approximation technique whose validity is external to the paper's own fitted values or self-referential definitions. The new fourth-order formula is constructed from prior observations and then ranked via the same benchmark procedure; no equation reduces the reported ordering or cost metric to a tautology by construction. The cited Morales et al. work is by different authors and supplies an independent higher-order construction. No self-citation chain, ansatz smuggling, or renaming of known results is load-bearing for the headline claim. The derivation chain is therefore self-contained against the stated numerical benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claims rest on the accuracy of the perturbative eigenvalue-error estimator and on the assumption that 1D hydrogen chains are representative enough for the cost comparisons to be useful; no free parameters or invented entities are mentioned.

axioms (1)
  • domain assumption The perturbative method accurately estimates the eigenvalue error induced by each product formula for the purpose of cost comparison.
    Explicitly invoked to make comparisons feasible at larger system sizes without full simulation.

pith-pipeline@v0.9.1-grok · 5844 in / 1277 out tokens · 30929 ms · 2026-06-28T22:00:58.889493+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

31 extracted references · 1 canonical work pages

  1. [1]

    8th (Morales et al.)

    Quantum phase estimation Quantum phase estimation (QPE) [1] is an algorithm to estimate eigenphasesφof a unitaryUusing controlled- version ofU. Throughout this work, we assume the ide- alized case in which the input state to QPE is the exact eigenstate of interest, namely the true molecular ground state. Thus, the resource estimates below do not include t...

  2. [2]

    Product formulae approximate the time-evolution oper- atore iHt via the product ofe iHjt

    Product formulae Suppose a HamiltonianHis expressed as a linear com- bination ofJtermsH j withj∈{1,...,J}: H= J∑ j=1 Hj. Product formulae approximate the time-evolution oper- atore iHt via the product ofe iHjt. The second-order for- mula approximates the time evolution as S2(t) := J−1∏ p=1 e 1 2iHpteiHJt J−1∏ q=1 e 1 2iHJ−qt.(2) It satisfies eiHt =S 2(t) ...

  3. [3]

    We then translate it into Pauli operators using Jordan- Wigner transformation [22] withOpenFermion[23, 24]

    Perform restricted Hartree-Fock (HF) calculations usingPySCF[20, 21], and construct the second- quantized HamiltonianHusing the HF basis. We then translate it into Pauli operators using Jordan- Wigner transformation [22] withOpenFermion[23, 24]. We also calculate the true ground state en- ergyEby performing full configuration interaction calculation usingPySCF

  4. [4]

    In the QPE resource estimates below, we fixpto the formal order of each product formula and fit onlyα

    Evaluate the errorδEbetweenEandE pf for mul- tiple values oft, and fit it to the formαtp using the perturbative method described in Appendix D. In the QPE resource estimates below, we fixpto the formal order of each product formula and fit onlyα. For reference only, Appendix F reports the values obtained from an unconstrained fit in which both αandpare tr...

  5. [5]

    (9) to calculateF, and also evaluate DRZ

    Substitute the fittedαtogether with the fixed value ofpinto Eq. (9) to calculateF, and also evaluate DRZ. We then compare these costs for each prod- uct formula as functions of the target errorεE

  6. [6]

    (T) median

    Analyze how each cost metric scales with system size. Specifically, for systems that achieve the tar- get errorεE, we fitFandD RZ as functions of the number of qubitsNusing the models F(N) =a FNbF, D RZ(N) =a DNbD, wherea F ,b F ,a D, andb D are fitting parameters. We then use the fitted models to extrapolate these costs to larger qubit counts. Note that,...

  7. [7]

    A. Y. Kitaev, arXiv:quant-ph/9511026 [quant-ph] (1995)

    Pith/arXiv arXiv 1995

  8. [8]

    D. S. Abrams and S. Lloyd, Physical Review Letters83, 5162 (1999)

    1999

  9. [9]

    H. F. Trotter, Proceedings of the American Mathematical Society10, 545 (1959)

    1959

  10. [10]

    Suzuki, Communications in Mathematical Physics51, 183 (1976)

    M. Suzuki, Communications in Mathematical Physics51, 183 (1976)

    1976

  11. [11]

    G. H. Low and I. L. Chuang, Quantum3, 163 (2019)

    2019

  12. [12]

    Gily´ en, Y

    A. Gily´ en, Y. R. Sanders, G. H. Low, and N. Wiebe, in Proceedings of the 51st Annual ACM SIGACT Sympo- sium on Theory of Computing (STOC ’19)(ACM, 2019) pp. 193–204

    2019

  13. [13]

    Bravyi and A

    S. Bravyi and A. Kitaev, Physical Review A71, 022316 (2005)

    2005

  14. [14]

    Gidney, N

    C. Gidney, N. Shutty, and C. Jones, arXiv:2409.17595 [quant-ph] (2024). 12

    Pith/arXiv arXiv 2024

  15. [15]

    Hirano, T

    Y. Hirano, T. Itogawa, and K. Fujii, in2024 IEEE In- ternational Conference on Quantum Computing and En- gineering (QCE)(IEEE, 2024) pp. 843–853

    2024

  16. [16]

    Akahoshi, K

    Y. Akahoshi, K. Maruyama, H. Oshima, S. Sato, and K. Fujii, PRX Quantum5, 010337 (2024)

    2024

  17. [17]

    M. E. S. Morales, P. Costa, G. Pantaleoni, D. K. Bur- garth, Y. R. Sanders, and D. W. Berry, arXiv:2210.15817 [quant-ph] (2024), arXiv:2210.15817v2

    arXiv 2024

  18. [18]

    Yoshida, Physics letters A150, 262 (1990)

    H. Yoshida, Physics letters A150, 262 (1990)

    1990

  19. [19]

    M. E. S. Morales, P. C. S. Costa, G. Pantaleoni, D. K. Burgarth, Y. R. Sanders, and D. W. Berry, Quan- tum Information and Computation 10.2478/qic-2025- 0001 (2025)

    work page doi:10.2478/qic-2025- 2025

  20. [20]

    G¨ unther, F

    J. G¨ unther, F. Witteveen, A. Schmidhuber, M. Miller, M. Christandl, and A. W. Harrow, arXiv:2503.05647 [quant-ph] (2025)

    Pith/arXiv arXiv 2025

  21. [21]

    Kanasugi, R

    S. Kanasugi, R. Toshio, K. Maruyama, and H. Oshima, arXiv:2603.22778 [quant-ph] (2026)

    arXiv 2026

  22. [22]

    Reiher, N

    M. Reiher, N. Wiebe, K. M. Svore, D. Wecker, and M. Troyer, Proceedings of the National Academy of Sci- ences114, 7555 (2017)

    2017

  23. [23]

    Suzuki, Physics Letters A146, 319 (1990)

    M. Suzuki, Physics Letters A146, 319 (1990)

    1990

  24. [24]

    Blanes and F

    S. Blanes and F. Casas, Linear algebra and its applica- tions378, 135 (2004)

    2004

  25. [25]

    Kahan and R.-C

    W. Kahan and R.-C. Li, Mathematics of Computation 66, 1089 (1997)

    1997

  26. [26]

    Q. Sun, T. C. Berkelbach, N. S. Blunt, G. H. Booth, S. Guo, Z. Li, J. Liu, J. D. McClain, E. R. Sayfutyarova, S. Sharma, S. Wouters, and G. K.-L. Chan, Wiley Inter- disciplinary Reviews: Computational Molecular Science 8, e1340 (2018)

    2018

  27. [27]

    Q. Sun, X. Zhang, S. Banerjee, P. Bao, M. Barbry, N. S. Blunt, N. A. Bogdanov, G. H. Booth, J. Chen, Z.-H. Cui, J. J. Eriksen, Y. Gao, S. Guo, J. Hermann, M. R. Hermes, K. Koh, P. Koval, S. Lehtola, Z. Li, J. Liu, N. Mardirossian, J. D. McClain, M. Motta, B. Mussard, H. Q. Pham, A. Pulkin, W. Purwanto, P. J. Robin- son, E. Ronca, E. R. Sayfutyarova, M. Sc...

    2020

  28. [28]

    Ortiz, J

    G. Ortiz, J. E. Gubernatis, E. Knill, and R. Laflamme, Physical Review A64, 022319 (2001)

    2001

  29. [29]

    J. R. McClean, I. D. Kivlichan, K. J. Sung, D. Steudt- ner, Y. Cao, C. Dai, E. S. Fried, C. Gidney, B. Gimby, T. H¨ aner, J. Izaac, Z. Jiang, M. Kusumoto, Y. Li, S. McArdle, M. Neeley, T. E. O’Brien, B. O’Gorman, M. Ozols, A. D. Patterson, J. Romero, N. C. Rubin, N. P. D. Sawaya, K. Setia, S. Sim, M. Steudtner, Q. Sun, W. Sun, D. Wang, F. Zhang, P. D. Nat...

    Pith/arXiv arXiv 2017

  30. [30]

    J. R. McClean, N. C. Rubin, K. J. Sung, I. D. Kivlichan, X. Bonet-Monroig, Y. Cao, C. Dai, E. S. Fried, C. Gidney, B. Gimby, P. Gokhale, T. H¨ aner, T. Hardikar, V. Havl´ ıˇ cek, O. Higgott, C. Huang, J. Izaac, Z. Jiang, X. Liu, S. McArdle, M. Neeley, T. E. O’Brien, B. O’Gorman, M. Ozols, D. Packwood, R. Santagati, M. S. Scheurer, K. Setia, S. Sim, S. Sin...

    2020

  31. [31]

    Inoue, K

    W. Inoue, K. Aoyama, Y. Teranishi, K. Kanno, Y. O. Nakagawa, and K. Mitarai, Phys. Rev. Res.6, 013096 (2024)

    2024