⚛️ quantum physics

Quasiparticle quantum simulation of materials with the Bethe-Salpeter equation

This paper presents a quasiparticle-based quantum simulation algorithm for the Bethe-Salpeter equation that leverages first-quantized block encoding and symmetry optimizations to achieve polynomial reductions in Toffoli gate counts and exponential reductions in qubit requirements for simulating multi-excitonic states in realistic materials, demonstrating that classical heuristic frameworks are essential for designing efficient future fault-tolerant quantum algorithms.

Original authors: Jielun Chen, Jiace Sun, Garnet Kin-Lic Chan

Published 2026-10-05

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Jielun Chen, Jiace Sun, Garnet Kin-Lic Chan

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 authors. For technical accuracy, refer to the original paper. Read full disclaimer

Materials science often feels like trying to understand a massive, bustling city by watching every single person move at once. In a solid piece of matter, such as a semiconductor chip, there are trillions of electrons zipping around, interacting with one another and with the atomic nuclei. To predict how this material will behave—how it conducts electricity, how it absorbs light, or how it might fail under stress—scientists must solve a complex set of rules that govern all these particles simultaneously. For decades, the most powerful computers have struggled with this task because the number of possibilities grows so fast that even the best supercomputers run out of time and memory when the material gets large. The standard approach has been to simulate every single electron, a method that works for tiny systems but becomes impossible for the realistic materials engineers actually want to build.

However, nature often offers a shortcut. In many materials, the interesting behavior at low energies comes not from the entire crowd of electrons, but from just a few special ones. When a material absorbs light, it doesn't excite every electron; instead, it creates a few specific disturbances, often described as pairs of a missing electron (a hole) and an excited electron. These pairs, and groups of them, behave like distinct particles called quasiparticles. By focusing only on these few active players rather than the entire sea of electrons, scientists can use simpler, faster models to describe the material's properties. This is the foundation of the Bethe-Salpeter equation, a well-established tool in classical physics that has successfully predicted the behavior of semiconductors for years. The challenge has always been that while these simplified models are easier for classical computers, they still become incredibly difficult to solve when we try to look at complex scenarios, such as when multiple pairs of these quasiparticles interact at the same time.

A team of researchers at the California Institute of Technology has now shown how to bring these simplified models into the realm of quantum computing, creating a path to simulate complex material behaviors that are currently out of reach. They developed a new way to translate the mathematics of these quasiparticle interactions into instructions that a future, error-corrected quantum computer could follow. Their work demonstrates that by sticking to the quasiparticle framework rather than trying to simulate every single electron, a quantum computer could solve these problems with a dramatic reduction in the resources required. In their simulations, they found that for a material containing a thousand atoms, their method could reduce the number of fundamental computational steps needed by a factor of a million to ten million compared to the standard method of simulating the full system of electrons.

The researchers focused specifically on a difficult problem: calculating the energy states of three interacting quasiparticle pairs, known as a triexciton. This is a scenario that is extremely expensive for classical computers to handle and has rarely been studied in detail for realistic material sizes. They built a detailed blueprint for a quantum algorithm that encodes the physics of these interactions directly into the quantum computer's operations. Instead of trying to store the state of every electron in the material, their method stores only the information about the few quasiparticles that matter. This approach allows the quantum computer to represent the system using far fewer memory units, known as qubits, than would be required to simulate the full system. In fact, for the materials they tested, their optimized design required only a few hundred logical qubits, a number that is manageable for future machines, whereas simulating the full system would require millions.

The study also revealed a massive difference in the time it would take to run these calculations. By using a technique that minimizes the number of complex logic gates the computer must perform, the researchers estimated that a quantum computer could solve the problem for a diamond crystal in roughly two weeks. In contrast, using current methods to simulate the full system of electrons for the same material would take anywhere from ten thousand to one hundred thousand years. This gap highlights a crucial insight: the most effective way to use quantum computers for materials science may not be to abandon the successful shortcuts developed by classical physicists, but to refine them. The researchers argue that these classical heuristic frameworks, which have been validated by decades of experimental success, contain essential information that should be preserved and utilized in quantum algorithms.

The team tested their approach on several common semiconductor materials, including diamond, magnesium oxide, and cadmium selenide. For each material, they adjusted the simulation to account for the specific symmetries of the crystal structure, which further reduced the computational load. They found that even with these optimizations, the savings remained enormous. The method works by breaking down the complex interactions between the quasiparticles into smaller, manageable pieces that can be processed efficiently. It avoids the need to load vast amounts of data into the computer's memory at once, instead using a strategy that loads information on demand. This allows the algorithm to scale much more gently as the size of the material increases, making it possible to study larger, more realistic systems than ever before.

Ultimately, this work provides a concrete roadmap for how quantum computers can tackle real-world problems in materials science. It suggests that the path forward is not to wait for quantum computers to be powerful enough to simulate every electron in a material, but to use them to solve the specific, simplified equations that physicists already know work. By combining the efficiency of classical approximations with the power of quantum algorithms, scientists can potentially unlock the ability to design new materials with tailored properties, such as more efficient solar cells or faster electronic devices. The results show that with the right approach, the quantum advantage is not just a theoretical possibility but a practical reality that could be realized in the coming era of fault-tolerant quantum computing.

Technical Summary: Quasiparticle Quantum Simulation of Materials with the Bethe-Salpeter Equation

Problem Statement
The first-principles quantum simulation of the full electronic Hamiltonian for materials of realistic size (e.g., $\sim 10^3$ atoms, $\sim 10^4$ electrons) remains a formidable challenge even for future fault-tolerant quantum computers. Standard full Hamiltonian simulation techniques require resource counts (Toffoli gates and logical qubits) that exceed available capabilities by many orders of magnitude to achieve relevant precision. Conversely, classical simulations of material excited states typically rely on heuristic quasiparticle frameworks, such as the $GW$ approximation and the Bethe-Salpeter equation (BSE), which model only a few quasiparticles (electrons and holes) rather than the full many-body system. While these classical methods are efficient with polynomial cost in system size $N$ (specifically $O(N^{2m})$ for $m$ excitons), they become intractable for high-order multi-excitonic states (e.g., triexcitons) due to the high degree of the polynomial scaling. The paper addresses the gap in simulating these complex excited states on quantum computers by leveraging the quasiparticle picture to achieve significant resource reductions compared to full Hamiltonian simulation.

Methodology
The authors develop a detailed framework for quasiparticle quantum simulation of multi-excitonic states within the BSE formalism. The methodology involves:

First-Quantized, Orbital-Based Formulation: The authors formulate the multi-exciton BSE Hamiltonian in a first-quantized, orbital basis. The computational basis consists of antisymmetric product states of $m$ electrons and $m$ holes (where $m$ is the number of excitons). This representation requires $O(m \log N)$ qubits, offering an exponential compression over classical dense vector storage ( $O(N^{2m})$ ) and standard first/second-quantized quantum simulation ( $O(N \log N)$ or $O(N)$ ).
Block Encoding and Qubitization: To enable quantum phase estimation (QPE), the authors construct an explicit block encoding of the BSE Hamiltonian ( $H_{BSE}$ $H_{B S E}$ ). They utilize a "walk operator" $W$ $W$ derived from a Linear Combination of Unitaries (LCU) approach.
- Hamiltonian Decomposition: $H_{BSE}$ is decomposed into a Fock term ( $H_0$ ) and interaction terms: direct electron-electron ( $H_{vv}^{dir}$ ), direct hole-hole ( $H_{oo}^{dir}$ ), direct electron-hole ( $H_{ov}^{dir}$ ), and exchange electron-hole ( $H_{ov}^{ex}$ ).
- Integral Factorization: To manage the $O(N^4)$ data size of interaction integrals (Coulomb $V$ and screened $W$ ), the authors employ Tensor Hypercontraction (THC). This reduces the data to $O(N^2)$ in general, and further to $O(N)$ when combined with translational symmetry (Bloch states).
- Symmetry Adaptation: The authors incorporate crystal space-group symmetries to reduce the number of independent data blocks that must be loaded, further lowering resource costs.
- Optimized Synthesis: A novel block encoding strategy is used, particularly for the exchange term, which avoids the unfavorable subnormalization scaling associated with straightforward LCU summations over crystal momentum. The synthesis utilizes interferometer-based unitary synthesis and QROAM (Quantum Random Access Memory) for data loading.
Resource Optimization: Two optimization strategies are explored:
- Toffoli-Optimized: Minimizes gate count by trading space for time (using $O(\sqrt{N})$ ancillas).
- Qubit-Optimized: Minimizes logical qubit count (using $O(\log N)$ qubits) at the cost of higher gate counts.

Key Contributions

Explicit Compilation: The paper provides a detailed description and an explicit first-quantized, orbital-based block encoding of the multi-exciton BSE Hamiltonian for qubitization, detailing the block encoding circuits for direct and exchange terms.
Asymptotic Speedups: The authors derive asymptotic scaling advantages. For a system of size $N$ $N$ and a fixed number of quasiparticles $m$ $m$ :
- Toffoli Count: The algorithm achieves a $\tilde{O}(N)$ scaling (or $\tilde{O}(\sqrt{N})$ with space-time tradeoffs), compared to $\tilde{O}(N^3)$ for full Hamiltonian simulation and $\tilde{O}(N^{2m+2})$ for classical matrix-vector methods. For triexcitons ( $m=3$ ), this represents a $\sim 6$ th-order polynomial speedup over full Hamiltonian simulation and a $\sim 12$ th-order speedup over classical methods.
- Qubit Count: The algorithm requires $\tilde{O}(\log N)$ logical qubits (qubit-optimized) or $\tilde{O}(\sqrt{N})$ (Toffoli-optimized), offering an exponential reduction in qubit count compared to full Hamiltonian simulation ( $O(N)$ or $O(N \log N)$ ).
Subnormalization Analysis: The authors demonstrate that the subnormalization factor $\lambda$ of the BSE Hamiltonian scales favorably (theoretically $\tilde{O}(1)$ , numerically $\sim N^{0.2}$ ), which is crucial for minimizing the QPE overhead.

Results
The authors performed resource estimates for triexcitonic ( $m=3$ ) eigenvalue estimation across several semiconductor models (Diamond, MgO, BN, AlN, CdSe) with realistic physical sizes ( $N \sim 10^3$ atoms).

Toffoli Reduction: For a Diamond system with a $10 \times 10 \times 10$ k-point mesh ( $N_k = 1000$ ), the Toffoli-optimized BSE simulation requires approximately $5.5 \times 10^{11}$ Toffoli gates. In contrast, standard full Hamiltonian simulation (second and first quantization) requires $3.7 \times 10^{17}$ and $4.2 \times 10^{18}$ gates, respectively. This constitutes a reduction of $10^6$ to $10^7$ .
Qubit Requirements: The qubit-optimized BSE implementation requires only a few hundred logical qubits (e.g., 266 for Diamond), whereas full Hamiltonian simulation requires $\sim 10^5$ qubits.
Runtime Estimates: Assuming a gate time of 25 $\mu$ s, the optimized BSE simulation for Diamond is estimated to take roughly 14 days. This is a reduction from the estimated $10^4$ – $10^5$ years required for full Hamiltonian simulation.
Classical Comparison: The classical cost for solving the same triexcitonic problem via dense matrix-vector multiplication is estimated at $3.3 \times 10^{27}$ FLOPs and $6.8 \times 10^{20}$ words of memory, rendering it intractable on current hardware.

Significance
The paper argues that classical heuristic frameworks, specifically the quasiparticle picture used in $GW$ -BSE theory, are essential for the design of future fault-tolerant quantum algorithms for materials. By restricting the simulation to the relevant low-energy quasiparticle subspace, quantum algorithms can achieve massive resource reductions (up to 6th power in time and exponential in space) compared to simulating the full electronic Hamiltonian. The work demonstrates that quantum advantage in materials science may not require simulating the full many-body problem exactly, but rather leveraging established classical approximations to target specific, classically intractable regimes (such as high-order multi-excitonic dynamics) with feasible quantum resources. This supports the thesis that designing quantum algorithms within the regimes where classical heuristics are valid is a critical direction for the field.

Technical Summary: Quasiparticle Quantum Simulation of Materials with the Bethe-Salpeter Equation

More like this