Quantum Computers Overcome Noise to Reveal Molecular Spectra with Greater Clarity

Scientists have developed a new method for extracting excitation spectra from complex many-electron systems, improving spectral reconstruction accuracy despite limitations in current quantum hardware resolution and environmental noise. Taichi Kosugi and colleagues at Quemix Inc, in collaboration with MITSUI KINZOKU COMPANY, National Institutes for Quantum Science and Technology (QST), The University of Tokyo, and Quantum Materials and Applications Research Centre, present QPE averaged over variable grids, or QAVG, which combines low-resolution quantum phase estimation with multiple origin shifts and continuous parametrization. They accurately determined the spectra of a CO molecule adsorbed onto a $χ$-Fe$_$5C$_2$ surface using Quantinuum H2-2, employing both physical and logical quantum phase estimation circuits with Steane code and offline bit-flip correction. QAVG effectively suppresses local minima during optimisation and offers a strong pathway towards quantum simulations of correlated spectra, enabling advancements in the field as fault-tolerant quantum computers develop.

Variable grid averaging enhances quantum spectral reconstruction for molecular modelling

Deviations in spectral reconstruction were reduced to less than the nominal QPE resolution, representing a two-fold improvement over previous methods and enabling accurate analysis previously impossible with limited quantum hardware. This breakthrough stemmed from employing a new technique, QPE averaged over variable grids, or QAVG, which combines multiple low-resolution measurements to overcome limitations imposed by noise and grid resolution in quantum computers. The fundamental principle behind QAVG lies in its ability to mitigate the effects of spectral leakage, a common artefact in Fourier-transform based spectroscopy where energy from a given spectral feature spreads into adjacent frequencies. Traditional quantum phase estimation (QPE) relies on accurately determining the eigenvalues of the system’s Hamiltonian, which requires a finely discretised grid of energy levels. However, current noisy intermediate-scale quantum (NISQ) devices struggle to maintain the coherence necessary for high-resolution QPE. QAVG circumvents this by performing multiple QPE measurements with slightly shifted energy grids, effectively sampling the underlying continuous spectrum and reconstructing it with enhanced precision. Quantinuum’s H2-2 quantum computer facilitated the successful application of QAVG to model a carbon monoxide molecule adsorbed onto a χ-Fe 5 C 2 surface, stabilising optimisation processes by suppressing local minima arising from spectral leakage.

Both physical QPE circuits and logical QPE circuits, encoded with Steane code and offline bit-flip correction, were utilised in experiments confirming the accuracy of the reconstructed spectra even with noisy data. Steane code is a quantum error-correcting code capable of protecting against bit-flip errors, a prevalent source of noise in quantum computations. The implementation of Steane code, coupled with offline bit-flip correction, demonstrates the potential for enhancing the robustness of QAVG against environmental disturbances. The use of logical qubits, which encode quantum information across multiple physical qubits to provide redundancy, further improves the reliability of the calculations. Shifting grids during averaging substantially reduced the number of local minima, aiding parameter optimisation, a process reminiscent of hyperacuity where the human eye achieves resolution beyond the spacing of photoreceptors. In hyperacuity, the brain integrates information from multiple slightly different viewpoints to perceive finer details than the physical resolution of the retina allows. Similarly, QAVG integrates information from multiple QPE measurements with shifted grids to reconstruct a spectrum with higher resolution than the individual measurements would permit. The optimisation process, crucial for determining the system’s Hamiltonian parameters, benefits significantly from this reduction in local minima, leading to more accurate and reliable results. Despite this promising advance in spectral reconstruction, extending the methodology to more complex and realistic materials remains a considerable challenge.

Predicting the behaviour of electrons in larger, disordered systems is notoriously difficult, and applying this technique to such systems represents a significant undertaking. The complexity arises from the exponential scaling of the Hilbert space with the number of electrons, requiring increasingly powerful quantum computers to simulate even moderately sized systems. Furthermore, real materials often exhibit structural imperfections and electronic correlations that further complicate the calculations. Demonstrating accurate spectral reconstruction on a real quantum computer, even with a simplified system like carbon monoxide on iron carbide, validates the approach as a viable pathway forward and opens possibilities for exploring the limitations of current quantum hardware in more demanding scenarios. The adsorption of CO onto the χ-Fe 5 C 2 surface serves as a well-defined model system, allowing researchers to benchmark the performance of QAVG and identify areas for improvement. QAVG, a novel technique for accurately determining the energy levels of complex materials, combines multiple low-resolution estimations with strategically shifted starting points, mitigating errors caused by limitations in current quantum hardware and environmental noise, similar to enhancing a blurry image through multiple exposures. The significance of accurate excitation spectra lies in their ability to reveal crucial information about the electronic structure and properties of materials. These spectra can be used to identify energy gaps, understand the nature of electronic transitions, and predict the material’s response to external stimuli. Future work will focus on extending this method to more complex materials, potentially ushering in a new era in quantum material simulations and providing insights into the scalability of this approach. This includes exploring more sophisticated error mitigation techniques and developing algorithms that can efficiently handle larger systems, ultimately paving the way for the design and discovery of novel materials with tailored properties.

Researchers successfully reconstructed excitation spectra with greater accuracy than standard methods using a technique called QAVG, which combines multiple low-resolution estimations and shifts in the starting parameters. This matters because accurately determining the energy levels of materials is crucial for understanding their electronic properties and behaviour. The study, performed on a CO molecule adsorbed onto a $χ$-Fe$_5$C$_2$ surface using the Quantinuum H2-2 computer, demonstrates a robust approach to quantum simulation, even with noisy data. The authors intend to extend this method to more complex materials and explore further error mitigation strategies.