Sample-based quantum simulation of vibrational structure of polyyne chains

Molecules are not static sculptures; they are constantly vibrating, with their atoms jiggling and stretching in complex patterns. To understand how a molecule behaves, how it absorbs light, or how it might react with another, scientists must map out these vibrations with extreme precision. For decades, the standard approach has been to treat these vibrations as simple, independent springs, a simplification known as the harmonic approximation. However, real molecules are more complicated. Their bonds stretch and squeeze in ways that interact with one another, creating a tangled web of motion called anharmonicity. Capturing this complexity is essential for predicting the exact colors of light a molecule absorbs, which is the basis of infrared spectroscopy. Yet, as molecules grow larger or as scientists try to model these interactions with greater detail, the mathematical problem becomes so vast that even the world's most powerful supercomputers struggle to solve it. The number of possible vibrational states grows so quickly that it overwhelms traditional calculation methods, leaving a gap in our ability to predict the behavior of complex chemical systems.

To bridge this gap, researchers have turned to quantum computers, which operate on the principles of quantum mechanics and are naturally suited to simulating other quantum systems like molecules. A new study by a team from IBM Research and ExxonMobil demonstrates a practical way to use these machines to tackle the vibrational structure of polyyne chains, which are molecules made of alternating single and triple carbon bonds. The team focused on two specific molecules: acetylene, a simple two-carbon chain, and diacetylene, a slightly larger four-carbon chain. Instead of trying to calculate every single possible vibration at once—a task that would require more computing power than exists—the researchers developed a method to sample the most important vibrations and ignore the rest. They used a technique called sample-based quantum diagonalization. In this approach, the quantum computer does not try to find the perfect answer immediately. Instead, it generates a large number of samples, or snapshots, of how the molecule might vibrate. These samples are then fed into a classical computer, which filters out the unlikely possibilities and focuses only on a small, manageable group of the most physically relevant states. By solving the problem within this tiny, selected group, the researchers could reconstruct the molecule's energy levels and its infrared spectrum with high accuracy.

The researchers tested two different ways to generate these samples. The first method, called vib-LUCJ, is based on a specific mathematical guess about how the molecule's vibrations are connected. The second, vib-SqDRIFT, works by simulating the molecule's evolution over a short period of time and seeing where it ends up. Both methods were applied to the two polyyne molecules, developing quantum representations that scale up to 104 qubits for the larger molecule. The results were striking. Even though the full mathematical space of possible vibrations for the larger molecule was astronomically huge—containing roughly 550 billion possible states—the quantum sampling methods managed to identify a tiny subset of just 16,000 to 26,000 states that contained almost all the necessary information. When the researchers calculated the energy levels from these small subsets, the results matched the best available classical calculations with an error margin of less than 10 units of energy, a level of precision known as chemical accuracy. This proved that the quantum computer could effectively find the "needles" in the "haystack" of vibrational possibilities without needing to search the entire haystack.

However, the study revealed that getting the energy right is only half the battle. The researchers also calculated the infrared spectra, which show how strongly the molecule absorbs light at different frequencies. This is a much harder test because it depends not just on the energy levels, but on the detailed shape of the vibrational waves themselves. They found that even when the energy levels appeared to have settled and stopped changing, the infrared intensities continued to shift as they added more detail to their model. This means that a calculation can look perfect for energy but still miss important details about how the molecule interacts with light. The study showed that to get a true picture of the molecule's behavior, one must look beyond just the lowest energy states and ensure that the complex mixing of different vibrations is captured correctly. The team discovered that the infrared spectrum is a far more sensitive indicator of whether a simulation is truly accurate than the energy values alone.

This work represents a significant step forward in using quantum computers for chemistry, moving beyond simple energy calculations to full spectroscopic predictions. The researchers successfully demonstrated that sample-based methods can handle vibrational problems that are far too large for traditional computers to solve directly. They showed that by combining quantum sampling with classical processing, it is possible to recover accurate vibrational states and transition properties for molecules with representations requiring up to 104 qubits. While the current simulations were performed on specific hardware and used specific sampling strategies, the results suggest a clear path forward. The study highlights that the future of molecular simulation lies not in brute-force calculation, but in smart sampling—finding the right few states that tell the whole story. As quantum hardware improves, these methods could allow scientists to predict the properties of much larger and more complex molecules, opening new doors for understanding materials and chemical reactions that are currently out of reach.