Optimizing the number of measurements for vibrational structure on quantum computers: coordinates and measurement schemes

Marco Majland, Rasmus Berg Jensen, Mads Greisen Højlund, N. T. Zinner, Ove Christiansen

Calculating the properties of molecules is one of the most promising "killer apps" for quantum computers. However, we face a massive bottleneck: the measurement problem. To get a precise energy estimate using algorithms like the Variational Quantum Eigensolver (VQE), we often need to run the quantum circuit millions or even billions of times.

While researchers have spent years optimizing this for electronic structures (how electrons move), the vibrational structure (how the nuclei themselves dance and shake) has remained relatively unexplored. A new paper by Majland et al. changes that, demonstrating that by simply choosing the right "language" to describe these vibrations, we can significantly reduce the quantum workload.

TL;DR

The researchers show that the number of measurements required to simulate molecular vibrations on a quantum computer can be reduced by an average of 3-fold (and up to 7-fold) by optimizing the vibrational coordinate system. By moving beyond standard "Normal Coordinates" to localized hybrid systems, they demonstrate a practical path toward making vibrational quantum chemistry more feasible on near-term hardware.

What Problem Does This Paper Address?

In the quantum world, estimating a physical quantity like energy isn't a one-and-done deal. Because quantum mechanics is probabilistic, you have to sample the system many times to get a stable average. The number of measurements required is directly proportional to the variance of the Hamiltonian (the mathematical description of the molecule's energy).

For vibrational systems, the Hamiltonian can be incredibly complex because atoms don't just vibrate independently; they are coupled together in "anharmonic" ways. If these couplings are strong, the variance—and thus the measurement count—skyrockets. The challenge is finding a way to represent the molecule such that these couplings are minimized, making the energy easier to "read" with fewer samples.

Methods

The authors attacked the problem from two angles:

  1. Coordinate Transformations: Instead of using standard Normal Coordinates (NCs)—which treat vibrations as delocalized movements across the whole molecule—they employed Hybrid Optimized and Localized Coordinates (HOLCs). These allow for a "tuning" of the coordinate system to find a sweet spot between energy decoupling and spatial localization.
  2. Measurement Grouping Schemes: They tested different ways to group the "Pauli strings" (the basic building blocks of the quantum Hamiltonian) so they can be measured simultaneously. This included:
  3. QWC (Qubit-Wise Commutativity): A simpler grouping that requires less complex quantum circuits.
  4. FC (Full Commutativity): A more aggressive grouping that can reduce measurements further but requires more complex "change-of-basis" circuits.
  5. MCR (Mode Combination Range): A novel approach that exploits the fact that vibrational modes are distinguishable (unlike electrons), allowing terms to be batched based on which specific atoms they involve.

Datasets

The study benchmarked a diverse set of 18 molecules:

  • 9 Triatomic (3-mode) systems: Including $SO_2, O_3, S_3, H_2O,$ and $ClNO$.
  • 9 Tetratomic (6-mode) systems: Including $H_2CS, HFCO, FNO_2,$ and $H_2CO$.

The researchers used the MidasCpp software package to generate high-quality Potential Energy Surfaces (PES) for these molecules, ensuring the quantum simulations were grounded in realistic chemical data.

Key Insights and Ideas

  • Coordinates are a "Free" Optimization: Unlike changing the quantum algorithm itself, changing the coordinate system is a classical pre-processing step. It doesn't make the quantum circuit harder to run, but it can drastically lower the statistical noise.
  • Distinguishability is a Superpower: In electronic structure, all electrons are identical. In vibrations, the "modes" (e.g., a stretch vs. a bend) are distinguishable. The authors showed how to use this to create "Mode Combination" groupings, which could be a major advantage for scaling up to larger molecules.
  • The "FC vs. QWC" Surprise: In electronic chemistry, Full Commutativity (FC) is almost always better. Here, the authors found that for vibrations, the simpler QWC often performs just as well, meaning we might not need the extra circuit complexity that FC requires.
  • Symmetry isn't Everything: Molecules with identical symmetry (like $O_3$ and $S_3$) showed wildly different variance behaviors, suggesting that the specific "shape" of the energy surface matters more than just the molecule's point group.

Key Results

  • Measurement Reduction: For 3-mode molecules, they achieved an average 3-fold reduction in measurements, with a peak of 7-fold for some systems. For 6-mode molecules, the average reduction was 1.5-fold (up to 2.5-fold).
  • Optimal Localization: They found that a moderate amount of localization (using a parameter $w_l$ around $10^{-5}$ to $10^{-7}$) usually yielded the best results. Extreme localization, however, was counterproductive and increased the variance.
  • Total Counts: For the molecules studied, reaching "chemical accuracy" ($1 \mu E_h$) still requires roughly $10^7$ to $10^8$ measurements. While high, the reductions provided by this work bring these numbers much closer to the realm of what near-term quantum hardware can handle.

Limitations and Future Work

The authors acknowledge that their study focused on relatively small molecules (3 and 6 modes). While the results are promising, the real test will be scaling to larger systems. In larger molecules, localized coordinates are expected to provide even more dramatic benefits because they can effectively "isolate" vibrations in different parts of the molecule.

Additionally, the study used a "direct mapping" to qubits. Future work could explore more compact encodings, which might interact with the coordinate choice in interesting ways to further shrink the measurement overhead. Finally, the MCR grouping—while not the top performer for these small molecules—shows potential for larger systems where the classical overhead of sorting thousands of terms becomes a bottleneck.

Summary updated

Abstract

One of the primary challenges prohibiting demonstrations of practical quantum advantages for near-term devices amounts to excessive measurement overheads for estimating relevant physical quantities such as ground state energies. However, with major differences between the electronic and vibrational structures of molecules, the question of how the resource requirements of computing anharmonic, vibrational states can be reduced remains relatively unexplored compared to its electronic counterpart. Importantly, bosonic commutation relations, distinguishable Hilbert spaces and vibrational coordinates allow manipulations of the vibrational system that can be exploited to minimize resource requirements. In this work, we investigate the impact of different coordinate systems and measurement schemes on the number of measurements needed to estimate anharmonic, vibrational states for a variety of three-mode (six-mode) molecules. We demonstrate an average of 3-fold (1.5-fold), with up to 7-fold (2.5-fold), reduction in the number of measurements required by employing appropriate coordinate transformations, based on an automized construction of qubit Hamiltonians from a conventional vibrational structure program.
BibTeX
@inproceedings{majland2023optimizing,
  author = {Marco Majland and Rasmus Berg Jensen and Mads Greisen Højlund and N. T. Zinner and Ove Christiansen},
  title = {{Optimizing the number of measurements for vibrational structure on quantum computers: coordinates and measurement schemes}},
  booktitle = {Chemical},
  year = {2023},
  doi = {10.1039/d3sc01984e},
}

Loading related content...

Figure