Fault-tolerant quantum computations of vibrational wave functions

Marco Majland, Rasmus Berg Jensen, Patrick Ettenhuber, Irfansha Shaik, Nikolaj Thomas Zinner, Ove Christiansen · 2025

Read Fault-tolerant quantum computations of vibrational wave functions by Marco Majland, Rasmus Berg Jensen, Patrick Ettenhuber, Irfansha Shaik, Nikolaj Thomas Zinner, Ove Christiansen (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-16253.txt.

Introduction

Marco Majland and five co-authors set out, in a 2025 paper, ways to work out how molecules vibrate on fault-tolerant quantum computers. How a molecule vibrates shapes how it stores energy and reacts, so exact values matter for chemistry. Quantum machines may one day beat classical ones on such tasks. Yet fault-tolerant forms of these methods are still little explored.

The team gives several ways to encode vibrational Hamiltonians, the math that describes such motion, using qubitization. They try low-rank approximations of that math, two kinds of coordinate systems, and ways to split the work or group terms. Tests span small molecules and large ones with more than one hundred modes of vibration. The tests show which choices keep costs low. They give a base for later work in quantum chemistry.

Catalog record

CREATORMarco Majland, Rasmus Berg Jensen, Patrick Ettenhuber, Irfansha Shaik, Nikolaj Thomas Zinner, Ove Christiansen
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE118 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2508-16253

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Fault-tolerant quantum computations of vibrational wave functions Marco Majland, Rasmus Berg Jensen, Patrick Ettenhuber, Irfansha Shaik, Nikolaj Thomas Zinner, Ove Christiansen Abstract Quantum computation of vibrational properties of molecules is a promising platform to obtain computational advantages for computational chemistry. However, fault-tolerant quantum computations of vibrational properties remain a relatively unexplored field in quantum computing. In this work, we present different algorithms for efficient encodings of vibrational Hamiltonians using qubitization. Specifically, we investigate different encoding representations, high order tensor decomposition to obtain low rank approximations for the vibrational Hamiltonian, rectilinear and polyspherical coordinate systems, parallelization and grouping algorithms. To investigate the performance of the different methods, we perform benchmark computations for both small and large molecules with more than one hundred vibrational modes. I Introduction The accurate computation of the quantum states of molecules, which comprise many interacting electrons and nuclei, remains a formidable challenge, even for the most powerful high-performance computers of today. Quantum simulation of such many-body systems offers a promising path forward, with quantum chemistry widely considered one of the fields where quantum algorithms may provide substantial computational advantages daley_practical_2022 ; lee_is_2023 ; elfving_how_2020 ; aspuru-guzik_simulated_2005 . In this context, the quantum phase estimation algorithm formally presents a powerful framework for achieving quantum speedups in the calculation of molecular properties, such as ground and excited state energies and other observables aspuru-guzik_simulated_2005 ; knill_optimal_2007 ; rall_quantum_2020 ; steudtner_fault-tolerant_2023 ; obrien_efficient_2022 ; abrams_quantum_1999 ; reiher_elucidating_2017 . To date, most quantum computational efforts have concentrated on the electronic structure problem, namely, computing electronic energies and wave functions. Although these are central and logical starting points, as the electronic energies define the Potential Energy Surface when viewed as functions of nuclear positions, a complete molecular description must also account for the quantum states due to nuclear motion. These nuclear motions contribute critically to molecular free energies, key drivers of chemical reactivity and stability, and are often observed directly in spectroscopy. In recent years, studies have begun to explore the calculation of molecular vibrational properties using quantum computers, employing both Noisy Intermediate Scale Quantum (NISQ) devices and fault-tolerant quantum computing approaches mcardle_digital_2019 ; ollitrault_hardware_2020 ; sawaya_near-_2021 ; sawaya_resource-efficient_2020 ; sparrow_simulating_2018 ; magann_digital_2021 ; richerme_quantum_2023 ; sawaya_quantum_2019 ; jahangiri_quantum_2020 . Variational algorithms based on variants of unitary vibrational coupled cluster theory have been proposed to variationally encode vibrational wavefunctions for NISQ devices mcardle_digital_2019 ; ollitrault_hardware_2020 ; majland_optimizing_2023 . In the context of fault-tolerant algorithms primarily Trotter methods being investigated. sawaya_resource-efficient_2020 ; trenev_refining_2023 . Using rough estimates for several vibrational Hamiltonians, is has been argued that the cost of simulating vibrational properties may be tractable earlier compared to its electronic counterpart sawaya_near-_2021 . As part of such further speculations lower bounds and gate estimates of Trotter steps for acetylene-like polyynes were considered in Ref. trenev_refining_2023 using the methods presented in Ref. childs_theory_2021 . However, in the field of electronic structure, qubitization has been shown to require fewer Toffoli gates for practical calculations compared to other methods babbush_encoding_2018 ; lee_even_2021 ; berry_qubitization_2019 ; kivlichan_quantum_2018 ; babbush_low-depth_2018 . Qubitization encodes the Hamiltonian as a Szegedy walk operator which may be implemented with zero error excluding the error arising from rotation synthesis low_hamiltonian_2019 ; babbush_encoding_2018 . In contrast, Trotter methods implement the time evolution operator of the Hamiltonian which yield finite errors in the implementation kivlichan_quantum_2018 ; babbush_low-depth_2018 ; childs_theory_2021 . Since current encodings for vibrational wave functions rely on Trotter methods, any potential advantages of utilizing qubitization for encoding the vibrational Hamiltonian remains to be investigated. Furthermore, although substantial progress has been made in resource reduction techniques for electronic structure computations, only a limited number of optimization strategies have been suggested to minimize computational overhead in fault-tolerant vibrational computations sawaya_resource-efficient_2020 ; trenev_refining_2023 . We here note the recent interesting work on Trotter simulation with cost estimates that employs resource reduction techniques both similar and different to ours. malpathak2025trottersimulationvibrationalhamiltonians Cost estimates for vibrational computations necessitate consideration of vibrational Hamiltonians, including coordinate selection, Kinetic Energy Operator representation, and the Potential Energy Surface. Past research often focused on simple vibrational potentials describing small systems, using only rectilinear coordinates like normal coordinates. These are inadequate for large amplitude nuclear motions, such as torsional motions crucial in many biomolecular systems. Choosing non-optimal coordinates can lead to strong inter-coordinate couplings. Therefore, employing curvilinear coordinates for describing molecular internal motion, which naturally include angles and bonds, can be beneficial for reducing Potential Energy Surface and wave function couplings. In this work, we present an algorithm for encoding the vibrational Hamiltonian using qubitization. Detailed quantum circuits for block encodings of vibrational Hamiltonians are discussed. In order to optimize relevant parameters for qubitization, we discuss different representations for encoding vibrational second quantization operators. To reduce gate costs of block encoding when using chemically relevant many-term Hamiltonians, we present tensor decomposition methods utilizing a combination of the Tucker and canonical decompositions of multidimensional tensors which allow for efficient low-rank approximations of the vibrational Hamiltonian. In addition, we present efficient grouping algorithms to perform parallel block encodings which further reduce the circuit depths. To illustrate the applicability of our method to general curvilinear coordinate systems, we apply it to hydrogen peroxide using polyspherical coordinates. Leveraging recent techniques bader_efficient_2024 , we obtain an efficient representation of the vibrational Kinetic Energy Operator with low coupling complexity and manageable computational cost. Similar to how the Potential Energy Surface is often approximated in practice, the Kinetic Energy Operator can also be simplified. By following general frameworks for constructing exact Kinetic Energy Operators lauvergnat_exact_2002 , one can systematically derive approximate forms that retain accuracy while drastically reducing the number of coupled terms bader_efficient_2024 . To investigate the scaling of larger systems, we provide gate costs for encoding the vibrational Hamiltonian for large molecules ranging from 10 to 100 vibrational modes to perform realistic estimates of the required quantum resources for performing vibrational wave function calculations on fault-tolerant quantum computers. II Theoretical background In this section, we discuss the theoretical background relating to working conditions for the algorithms and computations discussed later. Specifically, we discuss many-mode second quantization formalism for vibrational wave function theory, the choice of coordinates and Kinetic Energy Operators, the representation of the full Hamiltonian, and finally qubitization. II.1 Many-mode second quantization We shall from the outset use many-mode second quantization as introduced in Ref. christiansen_second_2004 . Let \{m_{1},m_{2},...\} denote distinguishable vibrational modes and let \{p^{m},q^{m},r^{m},s^{m}\} index general one-mode functions (modals) for a mode m . We use N_{m} to denote the number of modals for mode m . We define creation and annihilation operators a^{m_{1}\dagger}_{p^{m_{1}}} and a^{m_{2}}_{q^{m_{2}}} which create and annihilate occupation of modals with indices p^{m_{1}} and q^{m_{2}} for two given modes m_{1} and m_{2} respectively. The space of occupation vectors defining occupation for all mode and modals is larger than the relevant physical space defined by \{\forall m:\sum_{p^{m}}k_{r^{m}}^{m}=1\} where k_{r^{m}}^{m} is the occupation of the modal p^{m} for the mode m . The commutation relations for the creation and annihilation operators may be written as \displaystyle[a_{p^{m_{1}}}^{m_{1}},a^{m_{2}\dagger}_{q^{m_{2}}}]=\delta_{p^{m_{1}}q^{m_{2}}}\delta_{m_{1}m_{2}} (1) and \displaystyle[a^{m_{1}\dagger}_{p^{1_{2}}},a^{m_{2}\dagger}_{q^{m_{2}}}]=[a_{p^{m_{1}}}^{m_{1}},a_{q^{m_{2}}}^{m_{2}}]=0 (2) Importantly, the creation and annihilation operators are not to be confused with harmonic oscillator ladder operators. The modal functions are not assumed to be harmonic oscillator functions but are general one-mode basis functions. In practical calculations, these

Read the complete text