Quantum simulations of spin chemistry on IBM Q quantum comuter
Abstract
Spin-1/2 particles like electrons are "nature's qubits", and spin systems are convenient to map onto quantum devices. Because of that, chemistry problems involving electron spin dynamics, such as spin chemistry simulations, are an interesting application for quantum computing. Spin chemistry is an interdisciplinary subfield of physics and chemistry dealing with magnetic and spin effects in chemical reactions, with important applications in biochemistry, quantum biology, and solar energy conversion. Spin-chemical effects connect quantum phenomena like superposition and entanglement directly to macroscopic chemical parameters such as reaction yields. In this work, we used IBM Q quantum devices, as well as Qiskit AER quantum simulator, to simulate the spin dynamics underlying the Radical Pair Mechanism (RPM), the most important mechanism of spin chemistry. As a simple model system for the RPM simulation, we chose quantum beats in recombination fluorescence of radical ion pairs. To reduce the number of hyperfine couplings (hfc's) to be accounted for, we focused on deuterated radical ions. The systems we addressed include the following radical pairs: diphenyl sulfide-d10+./para-terphenyl-d14 (PTP)-. (no significant hfc's); 2,2,6,6-tetramethylpiperidine (TMP)+./PTP-. (2 different hfc's); (TMP)2+./PTP-. (1 nitrogen hfc); 9,10-octaline+./PTP-. (8 equivalent proton hfc's); and 2,4-dimethylbutane+./PTP-. (2 groups of equivalent proton hfc's). Results of quantum beats simulation are in good agreement with both theory and experimental data for all radical pairs. More complex systems can be addressed in the future with NISQ devices. In order to simulate T1 and T2-relaxation in radical pairs, we used a novel approach leveraging the inherent qubit noise. We demonstrated that the similarity of T1 and T2-decoherence in qubits and radical spins makes it possible to use qubit noise (amplified with delay cycles) to simulate paramagnetic relaxation in spin systems with no additional computational overhead, which has interesting implications for simulation of open quantum systems on NISQ devices.