A Quantum-HPC Hybrid Workflow for Reaction-Center Electronic Dynamics: Application to a Cytochrome P450-Inspired Iron-Complex Model

arXiv:2607.05786 · quant-ph, physics.chem-ph · Submitted 2026-07-07 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Quantum Radio. Generated commentary on the latest quantum physics and condensed matter papers.

Kai: We're looking at the paper: "A Quantum-HPC Hybrid Workflow for Reaction-Center Electronic Dynamics".

Mira: Here's the gist of the paper. A Quantum-HPC Hybrid Workflow for Reaction-Center Electronic Dynamics: Application to a Cytochrome P450-Inspired Iron-Complex Model introduces a practical workflow that uses classical high-performance computing to construct reduced Hamiltonians and then maps them onto quantum circuits for dynamics on current hardware, establishing a diagnostics route for assessing the adequacy of active-space models Workflow Overview The paper introduces a four-stage workflow that moves from classical construction to quantum execution and diagnostic analysis.

Kai: This workflow involves constructing an effective multistate Hamiltonian along the reaction coordinate using classical electronic-structure calculations.

Mira: Next, its spectral properties are verified, and then the Hamiltonian is used to perform quantum time evolution to obtain state populations from a reactant initial state.

Kai: Finally, the product-manifold population is analyzed as a diagnostic observable to identify near-degeneracy regions and possible deficiencies in the reduced model.

Title and authors: Kai: Let's talk about the authors for a second. We've got Shintaro Maekawa, Takao Otsuka, Riku Masui, Juan W. Pedersen, David Muñoz Ramo, Yasushi Okuno and Kentaro Yamamoto on this paper. They seem to have a good mix of expertise here.

Mira: I think their background supports the whole approach because you need strong chemistry to build that initial model and then quantum mechanics to test it. The title itself points directly at what they are doing: a hybrid workflow for reaction-center dynamics, using this iron complex example.

Lev: From my side, I'm interested in how they bridge that gap between the classical HPC part and the quantum circuit part because that’s usually where the overhead gets really messy on real hardware.

Kai: Exactly. They’re not just running a quantum simulation; they are using classical calculations to create a reduced Hamiltonian, which then gets mapped onto quantum circuits for dynamics on current hardware. This is a hybrid approach, which is key here.

Mira: So what this means for the broader field of molecular modeling is that we don't have to rely solely on the high precision required for full active space calculations across the entire pathway if we can build a reliable reduced model first.

Lev: I worry about that first step, building the effective Hamiltonian from state-averaged complete active space self-consistent field calculations. That’s computationally heavy, and if that classical construction is flawed, the quantum simulation will just be running on garbage data.

The paper's summary: Kai: To summarize what they did in "A Quantum-HPC Hybrid Workflow for Reaction-Center Electronic Dynamics: Application to a Cytochrome P450-Inspired Iron-Complex Model," they introduce population transfer dynamics as a practical validation observable for active space derived reduced Hamiltonians in multistate reaction center chemistry.

Mira: What that means simply is that they are using how the populations move between different electronic states to check if their reduced model is accurate, especially in transition metal systems where missing couplings can really mess things up.

Lev: That makes sense because as I said before, agreement in relative energies alone doesn't guarantee you have a consistent description of how states mix along the reaction coordinate.

Kai: They use a specific iron complex model, treating

Fe(CN)four(O2): two− as the core complex <ref:2607.05786#pg1>. They fix the active space to CAS eight e six o, which includes those iron three d orbitals and the oxygen two pi star orbitals to focus on what’s essential for the reaction center.

Mira: And they built a reduced Hamiltonian using fourteen energetically relevant states, excluding state fifteen because it was way too far away from the lower energy states. They then checked their spectral level comparison and found an RMS deviation of zero point zero three zero electron volts and a maximum absolute deviation of zero point one four three electron volts across all twenty one points times fourteen states.

Lev: Those error numbers sound manageable, but I’m always looking for where the assumptions are weakest in that initial state averaging process they used to get that reduced model.

Kai: The paper then uses this Hamiltonian to perform quantum time evolution starting from a reactant ground state to obtain those state populations over time. Then, they analyze the product manifold population as a diagnostic observable to spot near degeneracy regions and potential deficiencies in the reduced model.

The paper's improvements: Kai: Now let’s talk about what they suggest are the real improvements this workflow offers. First, they use population transfer dynamics as a diagnostic tool to assess the adequacy of those reduced Hamiltonians. This lets AI systems find near-degeneracy regions around x equals zero point three where state mixing is strongest, which is a much better check than just looking at energy metrics alone.

Mira: That’s interesting because it moves us away from just checking if the energy matches, to checking if the states are actually interacting correctly along the path. That feels chemically intuitive for understanding reactivity.

Lev: I agree that population transfer is better, but what about when you try to make this run on actual NISQ hardware? The paper mentions they used coupling pruning with a cutoff of zero point zero two electron volts to reduce the nonzero coupling set from thirty-two down to seven while keeping the main transfer pathways.

Kai: They also chose a Trotter step number of thirty as the best practical operating point among values they tested. They quantified the trade-off between dynamical fidelity and circuit resources by choosing that cutoff and that step number because increasing it further gave only a small gain relative to the circuit cost.

Mira: So, they’re not just showing it works in theory; they’re showing how you can optimize the workflow for real devices using those hardware constraints. It moves this from a theoretical exercise to something executable under realistic device limitations.

Conclusion: Kai: So to wrap up, the main implication of this paper is that we can now use population transfer dynamics as a validation observable for active space derived reduced Hamiltonians in multistate reaction center chemistry. They successfully demonstrated that chemically interpretable multistate electronic dynamics can already be executed on current trapped-ion hardware under realistic device constraints.

Mira: It’s a big step because it shows that reproducing the adiabatic spectrum alone isn't enough to know if your reduced Hamiltonian is valid; you need to look at the population transfer dynamics instead. That's a more sensitive diagnostic tool for finding where state mixing is happening most strongly, like around x equals zero point three.

Lev: From an error correction standpoint, I’m glad they tested the trade-off between that coupling cutoff and the Trotter step number because it gives us a concrete starting point for how to set up simulations on noisy hardware. It shows what you need to sacrifice in circuit depth for practical results.

Kai: Exactly. The whole point of this paper, "A Quantum-HPC Hybrid Workflow for Reaction-Center Electronic Dynamics: Application to a Cytochrome P450-Inspired Iron-Complex Model," is that it proves we can actually execute these kinds of chemically interpretable multistate electronic dynamics on current trapped-ion hardware.

Mira: It’s a demonstration that the classical construction, when coupled with this quantum execution and diagnostic check, gives us a robust way to assess the validity of those reduced models in transition metal systems.

Lev: Yeah, it sets up a framework where we can systematically test those reduced models against real hardware constraints using dynamical observables.

Kai: That’s what we have here. We’ll take this workflow and see how we can extend it to include nuclear motion and environmental fluctuations next.

Quantinuum K.K. · Graduate School of Medicine, Kyoto University · Quantinuum Terrington House · RIKEN Center for Computational Science

quant-ph, physics.chem-ph

Submitted: 2026-07-07

Updated: 2026-07-14

Comments: 15 pages, 12 figures; Updated acknowledgments and funding information to match the submitted manuscript. No changes to the scientific results or conclusions

DOI: 10.1021/acs.jctc.6c01410

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Importance score: 96/100

The gist: A Quantum-HPC Hybrid Workflow for Reaction-Center Electronic Dynamics: Application to a Cytochrome P450-Inspired Iron-Complex Model introduces a practical workflow that uses classical…

Key concepts

Reduced Hamiltonian Construction
Classical calculations are used to create a simplified mathematical description of the complex's energy states. By selecting only the most important electronic configurations (the active space), researchers create a smaller Hamiltonian that captures the essential dynamics along a reaction pathway, making quantum simulation possible.
Product-Manifold Population (pP(t))
This is a key diagnostic observable that measures how much of the system's initial state has transferred to the product states over time. A rapid increase in pP(t) signals a region where electronic states are strongly mixing, which helps identify critical areas in the reaction mechanism.
Coupling Pruning and Trotterization
These are techniques used to make the quantum simulation practical on limited hardware. Coupling pruning reduces the number of interactions (couplings) considered, while Trotterization breaks down time evolution into small steps. These trade-offs allow researchers to find a balance between simulating accurate dynamics and staying within the computational limits of current quantum devices.

Terminology

Summary

A Quantum-HPC Hybrid Workflow for Reaction-Center Electronic Dynamics: Application to a Cytochrome P450-Inspired Iron-Complex Model introduces a practical workflow that uses classical high-performance computing to construct reduced Hamiltonians and then maps them onto quantum circuits for dynamics on current hardware, establishing a diagnostics route for assessing the adequacy of active-space models

Workflow Overview

The paper introduces a four-stage workflow that moves from classical construction to quantum execution and diagnostic analysis. This workflow involves constructing an effective multistate Hamiltonian along the reaction coordinate using classical electronic-structure calculations. p. 1 Next, its spectral properties are verified, and then the Hamiltonian is used to perform quantum time evolution to obtain state populations from a reactant initial state. p. 1 Finally, the product-manifold population is analyzed as a diagnostic observable to identify near-degeneracy regions and possible deficiencies in the reduced model. p. 1

Model Construction and Validation

The study focuses on a cytochrome P450-inspired Fe-complex model, specifically treating [Fe(CN)4(O2)]2− as the core complex. The active space was fixed to CAS(8e,6o), including the Fe 3d orbitals and the O2(π∗) orbitals to focus on essential reaction-center degrees of freedom. p. 3 The reduced Hamiltonian is constructed using 14 energetically relevant states (State 1–State 14), excluding State 15 which remains well separated from the lower-energy states. p. 4 The spectral level comparison showed a consistency of RMS = 0. p. 6030 eV and a maximum absolute deviation of 0. p. 0 p. 0143 eV across all 21 points × 14 states. p. 6

Dynamics Diagnostics and Control Parameters

The product-manifold population pP (t) is used as a diagnostic observable to identify near-degeneracy windows where state mixing is strongest. The classical exact time evolution at x = 0. p. 03 shows a marked and rapid increase in pP(t), indicating the region of strongest state mixing. p. 7 To render the workflow executable on current trapped-ion hardware, coupling pruning was used with a cutoff of ϵ = 0. p. 002 eV to reduce the nonzero coupling set from 32 to 7 while preserving dominant transfer pathways. p. 0 A Trotter step number of M = 30 was chosen as the best practical operating point among tested values. p. 0

Hardware Performance and Trade-offs

The end-to-end workflow was demonstrated on Quantinuum’s trapped-ion quantum computer Reimei, which reproduced the key reaction-coordinate trend identified by the classical reduced model, including a pronounced maximum at x = 0.3. The hardware measured product population was 0. p. 042 on hardware and 0. p. 0 p. 043 on the matched emulator. p. 0 Analysis of the accuracy-resource trade-off showed that increasing M beyond 30 yields only a marginal improvement relative to the circuit cost. p. 9 The coupling cutoff of ε = 0. p. 002 eV was adopted as a practical threshold that preserves dynamics while reducing circuit cost. p. 9

Conclusion and Outlook

This work establishes a dynamics-based diagnostic route for assessing the adequacy of active-space-derived reduced Hamiltonians. It shows that chemically interpretable multistate electronic dynamics can already be executed on current trapped-ion hardware under realistic device constraints. p. 0 The workflow successfully demonstrates that reproducing the adiabatic spectrum alone is not sufficient to assess the validity of the reduced Hamiltonian, as population-transfer dynamics provide a more sensitive and chemically relevant diagnostic tool. p. 8 Future work will extend this framework to include nuclear motion and environmental fluctuations. p. 8

--- Page 1 ---

A Quantum-HPC Hybrid Workflow for Reaction-Center Electronic Dynamics: Application to a Cytochrome P450-Inspired Iron-Complex Model We introduce population-transfer dynamics as a practical validation observable for active-spacederived reduced Hamiltonians in multistate reaction-center chemistry In transition-metal systems, missing couplings or configurations can compromise the reliability of reduced active-space models along a reaction coordinate, even when the overall spectrum is reproduced reasonably well Using a cytochrome P450-inspired Fe-complex model, we construct a reaction-coordinate-dependent effective Hamiltonian from state-averaged complete active-space self-consistent field (SA-CASSCF) calculations, map it to a quantum-circuit representation suitable for current hardware, and propagate the dynamics from the reactant-side ground state At the spectrum level, the reduced Hamiltonian reproduces the SA-CASSCF reference with an RMS deviation of 0.030 eV and a maximum absolute deviation of 0.143 eV As a dynamics-based diagnostic, the product-manifold population pP (t), identifies a pronounced near-degeneracy window around x ≈ 0. p. 03, where state mixing is strongest; specifically, pP (t) computed from classical exact time evolution reaches 0. p. 0488 at x = 0. p. 03 after 10 fs, compared with 7. p. 026 × 10−2 at x = 0.2 and5.90 × 10−3at x = 0.0 To render the workflow executable on current trapped-ion hardware, we quantify the NISQ-relevant trade-off between dynamical fidelity and circuit resources through coupling pruning and first-order Trotterization A coupling cutoff of ϵ = 0. p. 002 eV reduces the nonzero coupling set from 32 to 7 while preserving the dominant transfer pathways, and a Trotter step number of M = 30 provides the best practical operating point among the tested values Finally, we demonstrate the end-to-end workflow on Quantinuum’s trapped-ion quantum computer Reimei The hardware reproduces the key reaction-coordinate trend identified by the classical reduced model, including the pronounced maximum at x = 0. p. 03, where the measured product population is 0. p. 042 on hardware and 0.43 on the matched emulator Overall, this work establishes a dynamics-based diagnostic route for assessing the adequacy of active-space-derived reduced Hamiltonians and shows that chemically interpretable multistate electronic dynamics can already be executed on current trapped-ion hardware under realistic device constraints

--- Page 2 ---

Many chemical reactions in biomolecular systems proceed by exploiting a diverse manifold of electronic states localized at the reaction center Representative examples include transition-metal enzyme reaction centers such as cytochrome P450 and the photosynthetic oxygen-evolving complex (Mn4CaO5) [1], a stronglycorrelated transition-metal complex surrounded by thousands of atoms of the photosystem II (PSII) protein complex [2], which serves as its environment If such biomolecular systems could be simulated accurately, they would enable clearer mechanistic interpretation of spectroscopy/kinetics and provide design principles for bioinspired catalysts and selective oxidation chemistry In practice, the interpretation of experimental results often relies on reduced models that do not explicitly account for environmental effects [2, 3] For a rigorous first-∗ shintaro.maekawa@quantinuum. p. 0com † kentaro. p. 0yamamoto@quantinuum. p. 0com principles description [4, 5], however, a hierarchical modeling strategy is essential Specifically, while the reaction center must be treated with high-precision quantum chemistry, the surrounding environment must be handled using approximations suited for systems with many degrees of freedom Despite this separation of roles, the combination of strong electronic correlations in the reaction center and the vast configuration space of the surrounding environment remains a major barrier to predictive simulations The most serious bottleneck is that, even with state-of-the-art high-performance computing (HPC), the computational resources required to evaluate the electronic states of the reaction center can be prohibitively large This difficulty is particularly acute in transitionmetal enzymes and photochemical reaction systems In transition-metal systems, nontrivial electronic-structure effects and inherent multiconfigurational character present significant challenges to standard computational approaches In photochemical and nonadiabatic dynamics, multiple electronic states often interact and become strongly mixed, and their character and couplings may evolve along the reaction coordinate BearXiv:2607. p. 005786v2 [quant-ph] 14 Jul 2026

--- Page 3 ---

cause nonadiabatic couplings depend on the quality and character of the underlying electronic wave functions, agreement in relative energies alone may not guarantee a consistent description of interstate mixing [11] Consequently, elucidating the reaction mechanism requires a consistent multistate model of the reaction center that includes interstate couplings and nonadiabatic effects A practical framework for achieving such a description is multireference electronic-structure theory within a suitably chosen active space [12, 13] However, the construction of the active space is inherently path-dependent Even if the active space is optimal near the reactant region, it may become incomplete in regions where the dominant electronic configurations change, such as near conical intersections In practice, as the reaction coordinate evolves, “important configurations” and “dominant orbitals” can interchange; therefore, an active space defined by a single geometric structure may not remain appropriate throughout the entire pathway Therefore, a framework to verify whether the effective (reduced) model derived from the selected active space is valid across the entire reaction coordinate is essential for ensuring the reliability of mechanistic interpretations In particular, the reduced model should reproduce not only energy but also the interstate mixing that governs electron transfer and reactivity, which can be examined via population-transfer dynamics as an experimentally interpretable surrogate observable

--- Page 4 ---

Maintaining consistency across the entire reaction pathway often requires a large active space that integrates essential orbitals from different regions, and the computational cost of methods such as complete activespace self-consistent field (CASSCF) increases exponentially with the size of the active space Consequently, even if the environment is hierarchically separated in a conceptually sound manner, the electronic structure problem of the reaction center itself remains a key practical limitation Furthermore, the Born–Oppenheimer approximation [15], which allows the dynamic coupling between electronic and nuclear motions to be ignored, generally breaks down when electronic properties change dramatically along the reaction coordinate, further complicating the problem for classical computers Such nonadiabatic and multistate effects are now widely recognized as central elements of reactivity and photochemistry in transition-metal systems and strongly correlated systems [8, 9] Quantum computing [18] has long been considered a promising approach for overcoming classical limits caused by the exponential growth of computational resources [19, 20] Aspuru-Guzik et al. demonstrated that quantum algorithms can suppress the scaling of key resource requirements to polynomial levels In recent years, efforts to apply quantum computing to computational chemistry have made progress in various areas, including ground-state energy estimation, improved treatment of electronic correlations, and quantum simulation of reduced chemical Hamiltonians Research at the interface of quantum computing and chemistry has developed along two main directions, although the distinction between them is not always strict Much of the work has focused on electronic structure problems It typically relies on variational approaches or phase estimation to obtain ground- and excited-state energies In practice, many studies have concentrated on relatively small molecular systems and on reducing measurement cost and circuit overhead Another direction explores the simulation of model Hamiltonians and dynamical processes In this context, time evolution is a natural target for quantum devices To make such problems tractable on current hardware, simplified descriptions are often introduced, for example active-space models or effective diabatic Hamiltonians With the advent of the NISQ (noisy intermediate-scale quantum) era, the focus has shifted from asymptotic advantage toward what can be achieved under realistic constraints A key question is how reliably chemically meaningful observables can be reproduced in the presence of noise and limited circuit depth Despite ongoing progress, it remains difficult to describe biomolecular reaction centers with high multireference accuracy across extended reaction coordinates This has motivated the development of practical workflows for validating reduced Hamiltonians and their couplings In this work, we introduce a quantum–HPC hybrid approach and demonstrate its feasibility on current quantum hardware We focus on the implementation of multistate dynamics and on the diagnostic observables that can be obtained from it Closer integration between classical workflows and quantum implementations remains an important direction The workflow proceeds in four stages First, an effective multistate Hamiltonian is constructed along the reaction coordinate using classical electronic-structure calculations Next, its spectral properties are verified The Hamiltonian is then used to perform quantum time evolution and to obtain the dynamics of state populations from a reactant initial state Finally, the population of the product manifold is analyzed as a diagnostic observable to identify near-degeneracy regions and possible deficiencies in the reduced model Unlike energy-based metrics, which are often insensitive to localized near-degeneracy, population-transfer observables depend explicitly on interstate couplings through the time evolution of electronic amplitudes In standard nonadiabatic dynamics, the time evolution of state populations is described in terms of transition or hopping probabilities governed by nonadiabatic couplings [24], and in perturbative electron-transfer regimes by rates that scale with the square of the electronic coupling [25] Consequently, missing states or couplings can lead to significant deviations in population dynamics, providing a more stringent and chemically intuitive validation of reduced Hamiltonians We validate the dynamics using Quantinuum’s System Model H1 trapped-ion quantum computer “Reimei,”

--- Page 5 ---

examining whether chemically interpretable multistate population transfer can be reproduced under realistic NISQ conditions Here, “chemically interpretable” refers to linking population transfer with electronic-state interactions and state mixing, and with measurable observables such as spectroscopic signals The platform provides all-to-all connectivity and supports mid-circuit measurement and conditional branching, which makes it suitable for such benchmarks This work makes two contributions.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Dynamic diagnostic capability for reduced Hamiltonians: The improved system can assess the adequacy of active-space-derived reduced Hamiltonians by using population-transfer dynamics as a practical validation observable. This allows AI systems to identify near-degeneracy regions around x ≈ 0.3, where state mixing is strongest, which provides a more stringent and chemically intuitive validation than energy metrics alone.

  2. Hardware constraint optimization: The system can employ coupling pruning (epsilon = 0.02 eV) and first-order Trotterization (M = 30) to quantify the NISQ-relevant trade-off between dynamical fidelity and circuit resources. This enables the AI to determine a practical operating point for current trapped-ion hardware, ensuring that simulations are both chemically meaningful and implementable on real devices.

  3. Fidelity assessment under noise: The system can distinguish between errors in simulated dynamics and hardware noise by comparing deviation metrics like pP(t) with leakage rates. This allows the AI to interpret results by separating approximation errors (e.g., Trotterization and truncation of coupling terms) from errors caused by hardware noise.

  4. Coordinate-resolved error mapping: The system can perform a coordinate-resolved analysis of emulator accuracy and error structure to see how residual errors depend on the reaction coordinate. This enables the AI to identify regions where coherent amplification is amplified, such as the near-degeneracy region around x ≈ 0.25–0.40, which is crucial for understanding physical mechanisms like electron transfer enhancement.

  5. End-to-end workflow execution: The improved system can execute an end-to-end workflow on quantum hardware to demonstrate that chemically interpretable multistate electronic dynamics can already be executed on current trapped-ion hardware under realistic device constraints. This validates the entire process from classical Hamiltonian construction to quantum simulation.

Related papers