Quantum Diagonalization of Iron-Sulfur Clusters
sample-based quantum diagonalization for strongly correlated iron-sulfur clusters on IBM quantum processors
collaborators: IBM Research, TCS Research, Prof. Phani Motamarri
Ongoing work.
Iron-sulfur clusters such as [2Fe-2S], [4Fe-4S] and the iron-molybdenum cofactor (FeMoco) of nitrogenase are textbook examples of strongly correlated molecules: many nearly degenerate electronic configurations matter at once, and even the best classical methods struggle to describe them accurately. Sample-based quantum diagonalization (SQD) uses a quantum computer to sample configurations from a correlated quantum state, and a classical computer to diagonalize the Hamiltonian in the subspace those samples span.
This project improves each stage of that pipeline with one question in mind: what does it take for quantum samples to match, and eventually go beyond, the best classical selected configuration interaction methods? We work on
- circuits built from physically motivated starting points and mapped efficiently onto IBM’s heavy-hex qubit layout;
- recovering useful configurations from noisy hardware samples;
- getting more accuracy out of the sampled configurations on the classical side;
- classical baselines at the same subspace size, so that the contribution of the quantum computer is measured fairly.
We test these ideas on [4Fe-4S] (54 electrons in 36 orbitals, 72 qubits) on IBM Heron processors. Two companion projects feed into this pipeline: randomized Krylov circuits (SqDRIFT) for generating configurations, and a faster selected-CI eigensolver for the classical diagonalization.
