⚛️ quantum physics

Fast simulations of X-ray absorption spectroscopy for battery materials on a quantum computer

This paper presents a highly optimized quantum algorithm for simulating X-ray absorption spectroscopy in battery materials, which leverages specific Hamiltonian representations, eigenvalue-focused error bounds, and sample distribution strategies to enable efficient fault-tolerant simulations of complex systems like Li $_4$ Mn $_2$ O with significantly reduced resource requirements.

Original author: Stepan Fomichev, Pablo A. M. Casares, Jay Soni, Utkarsh Azad, Alexander Kunitsa, Arne-Christian Voigt, Jonathan E. Mueller, Juan Miguel Arrazola

Published 2026-11-06

📖 5 min read🧠 Deep dive

CC BY 4.0

Original author: Stepan Fomichev, Pablo A. M. Casares, Jay Soni, Utkarsh Azad, Alexander Kunitsa, Arne-Christian Voigt, Jonathan E. Mueller, Juan Miguel Arrazola

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the author. For technical accuracy, refer to the original paper. Read full disclaimer

Batteries that power electric vehicles and store renewable energy are constantly being pushed to hold more charge. One promising path involves using special cathode materials rich in lithium, but these materials often degrade quickly, losing their ability to hold power after just a few charge cycles. To understand why this happens, scientists need to see the invisible changes occurring deep inside the material's atomic structure, specifically how the atoms bond and what their electrical states are during use. The primary tool for peering into this hidden world is a technique called X-ray absorption spectroscopy. By firing X-rays at a material, researchers can measure how much energy the atoms absorb, creating a unique fingerprint that reveals the local arrangement of electrons and atoms. However, reading this fingerprint is incredibly difficult. The raw data is complex, and to translate it into a clear picture of the material's condition, scientists must run massive computer simulations. For the large, complicated molecules found in advanced batteries, these simulations are so computationally expensive that they often take too long or require resources that simply do not exist on today's classical computers.

A team of researchers has now developed a new way to perform these simulations using quantum computers, a type of machine designed to handle the specific complexity of quantum systems. In their work, they focused on a model cluster of atoms representing a lithium-excess battery material, a system that is currently beyond the reach of standard supercomputers. Instead of trying to simulate the entire history of the material's behavior all at once, which is the traditional and costly approach, they devised a method that watches how the system evolves over time. They broke the simulation down into a series of small, manageable steps, using a mathematical strategy that compresses the information needed to describe the atoms' interactions. This allowed them to simulate the X-ray absorption spectrum with far fewer computational resources than previously thought possible.

The researchers tested their method on a specific cluster containing four lithium atoms, two manganese atoms, and one oxygen atom, a setup that represents the Li4Mn2O system. While the full model for this cluster involves eighteen spatial orbitals and twenty-two electrons, the specific simulation described in the validation phase used a smaller active space labeled CAS(14e,11o), which focuses on the ten Mn 3d orbitals and the core O 1s orbital. In a classical simulation, solving for the behavior of this many electrons would require calculating the energy of countless possible states, a task that grows exponentially harder as the system gets larger. The new quantum algorithm, however, avoids this explosion of complexity. By using a technique that groups the atomic interactions into efficient blocks and by carefully choosing how many times to repeat the simulation steps, they reduced the computational load dramatically. They found that they could simulate this challenging system using only one hundred logical qubits, the fundamental units of information in a quantum computer, and fewer than four hundred million specific logic gates per circuit. This is a reduction in cost of five orders of magnitude compared to unoptimized methods, bringing the task within the reach of future fault-tolerant quantum devices.

To ensure their results were accurate, the team first ran their algorithm on a classical simulator to verify that it could correctly reproduce the known absorption spectrum of a simpler nitrogen molecule. Once the method was validated, they applied it to the lithium-manganese-oxygen cluster using the reduced active space. The simulation successfully generated a spectrum that matched the expected physical behavior of the system, confirming that the algorithm could access the core-excited states that are crucial for X-ray analysis. The researchers also compared their approach to the best available classical methods, such as the restricted active space technique. While a classical supercomputer might take weeks to simulate this specific system by calculating state after state, their quantum algorithm suggests a runtime of less than two days on a hypothetical quantum computer operating at standard speeds. This difference arises because the quantum method scales much more gently with system size, whereas classical methods become exponentially more difficult as more atoms are added.

The significance of this work lies in its potential to accelerate the development of better batteries. By making it feasible to simulate the X-ray absorption spectra of complex battery materials, scientists can finally deduce the exact oxidation states of atoms during degradation. This information is the key to understanding why lithium-excess cathodes fail and how to fix them. The paper does not claim to have solved the battery problem or to have run these simulations on a physical quantum computer yet; rather, it demonstrates that a highly optimized algorithm exists which can do the job with realistic resource estimates. The team showed that by refining how the computer handles the math of atomic interactions and how it samples the data, the barrier to entry for these critical simulations can be lowered. This paves the way for using quantum computers not just as theoretical tools, but as practical instruments for designing the high-capacity energy storage systems needed for a sustainable future.

Technical Summary: Fast Simulations of X-ray Absorption Spectroscopy for Battery Materials on a Quantum Computer

Problem Statement
X-ray absorption spectroscopy (XAS), particularly X-ray absorption near-edge spectroscopy (XANES), is a critical technique for understanding structural changes in advanced battery materials, such as lithium-excess cathodes. These materials are promising for increasing battery capacity but suffer from rapid performance decline linked to irreversible structural degradation, often involving oxygen dimerization or changes in transition metal oxidation states. Determining these oxidation states from experimental spectra requires high-accuracy simulations. However, existing classical methods face a dichotomy: density functional theory (DFT) lacks the necessary accuracy for strongly correlated systems, while wavefunction-based methods (e.g., configuration interaction) are computationally prohibitive for the system sizes required to model these degradation mechanisms. There is a specific need for simulation methods capable of resolving the local electronic structure of bulk materials with strongly correlated transition metals, a challenge where quantum computers are well-positioned to contribute.

Methodology
The paper proposes a highly optimized quantum algorithm for simulating XAS spectra based on a time-domain approach. The core objective is to compute the absorption cross-section, which is related to the imaginary part of the frequency-domain Green's function. Instead of direct matrix inversion, the algorithm utilizes time evolution under the electronic Hamiltonian, which is more natural for quantum hardware.

The implementation relies on several key technical components:

Time Evolution via Compressed Double Factorization (CDF): The electronic Hamiltonian is decomposed using the CDF technique (Cartan subalgebra approach). This reduces the number of terms required for product formula time evolution from $O(N^4)$ (typical of Jordan-Wigner mapping) to $\tilde{O}(N^3)$ . The Hamiltonian is split into one-body and two-body fragments, each diagonalized in a specific single-particle basis, allowing for efficient implementation via Givens rotations and Pauli Z rotations.
Perturbative Error Analysis for Trotter Steps: A critical optimization involves re-evaluating the error requirements for spectroscopy. Rather than controlling the generic error of the full time evolution operator, the algorithm focuses on controlling the error in the eigenvalues of the approximate Hamiltonian. Using perturbation theory, the authors demonstrate that significantly fewer Trotter steps are required to achieve the necessary spectral accuracy (specifically for peak positions) compared to standard time-evolution error bounds.
Optimized Sampling Strategy: The algorithm employs a Monte Carlo sampling scheme that leverages the exponentially decaying Lorentzian kernel inherent in the Fourier transform of the time-domain signal. Instead of uniform sampling across evolution times, the shot budget is allocated according to the kernel's decay, significantly reducing the number of measurements needed for longer evolution times.
Circuit Optimizations: Several circuit-level tricks are implemented to reduce gate counts:
- Double Phase Trick: Replacing controlled rotations with uncontrolled rotations sandwiched by CNOTs, effectively doubling the simulated time steps per circuit execution and reducing Toffoli gate costs by a factor of 4.
- Combining Consecutive Rotations: Merging unitary rotations between adjacent fragments to reduce the number of basis rotation gates.
- Double Measurement Trick: Extracting additional information from the system register after the Hadamard test to gain a multiplicative saving in shots.
- Rotation Precision: Setting rotation precision based on the final observable's accuracy requirements rather than arbitrary high-precision standards.

Key Contributions and Results
The authors apply this algorithm to a prototypical model system: the $\text{Li}_4\text{Mn}_2\text{O}$ cluster with a complete active space of 22 electrons in 18 orbitals (CAS(22e, 18o)). This system was previously identified as being beyond the reach of existing classical simulation methods for this specific application.

Resource Estimates: Through constant-factor resource estimation, the authors show that the optimized algorithm requires approximately 100 logical qubits and $3.54 \times 10^8$ Toffoli gates for the longest circuit. When compiled for an active volume architecture with 350 logical qubits and a 1 MHz logical clock rate, the estimated runtime is less than two days.
Performance Gains: The combination of optimizations results in a reduction of the Toffoli gate count by five orders of magnitude compared to an unoptimized implementation. Figure 1 in the paper details the specific contributions of each optimization (e.g., CDF provides a $\times 20$ reduction, perturbative analysis provides a $\times 10.8$ reduction).
Validation: The algorithm was implemented on a simulator using the PennyLane library. It was first benchmarked against a classical Full Configuration Interaction (FCI) reference for the $\text{N}_2$ molecule, accurately reproducing the absorption spectrum. It was then applied to the $\text{Li}_4\text{Mn}_2\text{O}$ cluster using the core-valence separation (CVS) approximation to isolate core-excited states, demonstrating convergence of the spectrum with respect to simulation parameters.
Classical Comparison: A rough comparison with the state-of-the-art classical Restricted Active Space (RAS) method suggests the quantum algorithm could be approximately 30 times faster for the target system, with the gap widening significantly for larger system sizes due to the exponential scaling of classical methods versus the polynomial scaling ( $O(N^3)$ ) of the quantum approach.

Significance
The paper claims that these optimizations make XAS simulation a realistic and attractive application for early fault-tolerant quantum devices. By drastically reducing the resource requirements, the work demonstrates that quantum computers can potentially solve specific, industrially relevant problems in battery material science that are currently intractable for classical high-performance computing. The ability to simulate these spectra with high accuracy could enable the determination of oxidation states in lithium-excess cathodes, thereby aiding in the understanding of degradation mechanisms and accelerating the development of high-capacity battery technologies. The authors frame this not as a definitive solution to all battery challenges, but as a proof-of-concept that a highly optimized quantum algorithm can bridge the gap between theoretical simulation and industrial application in the near term.

More like this