TY - UNPB
T1 - Self-consistent Quantum Linear Response with a Polarizable Embedding environment
AU - Reinholdt, Peter
AU - Kjellgren, Erik Rosendahl
AU - Ziems, Karl Michael
AU - Coriani, Sonia
AU - Sauer, Stephan P. A.
AU - Kongsted, Jacob
PY - 2024
Y1 - 2024
N2 - Quantum computing presents a promising avenue for solving complex problems, particularly in quantum chemistry, where it could accelerate the computation of molecular properties and excited states. This work focuses on hybrid quantum-classical algorithms for near-term quantum devices, combining the quantum linear response (qLR) method with a polarizable embedding (PE) environment. We employ the self-consistent operator manifold of quantum linear response (q-sc-LR) on top of a unitary coupled cluster (UCC) wave function in combination with a Davidson solver. The latter removes the need to construct the entire electronic Hessian, improving computational efficiency when going towards larger molecules. We introduce a new superposition-state-based technique to compute Hessian-vector products and show that this approach is more resilient towards noise than our earlier gradient-based approach. We demonstrate the performance of the PE-UCCSD model on systems such as butadiene and para-nitroaniline in water and find that PE-UCCSD delivers comparable accuracy to classical PE-CCSD methods on such simple closed-shell systems. We also explore the challenges posed by hardware noise and propose simple error correction techniques to maintain accurate results on noisy quantum computers.
AB - Quantum computing presents a promising avenue for solving complex problems, particularly in quantum chemistry, where it could accelerate the computation of molecular properties and excited states. This work focuses on hybrid quantum-classical algorithms for near-term quantum devices, combining the quantum linear response (qLR) method with a polarizable embedding (PE) environment. We employ the self-consistent operator manifold of quantum linear response (q-sc-LR) on top of a unitary coupled cluster (UCC) wave function in combination with a Davidson solver. The latter removes the need to construct the entire electronic Hessian, improving computational efficiency when going towards larger molecules. We introduce a new superposition-state-based technique to compute Hessian-vector products and show that this approach is more resilient towards noise than our earlier gradient-based approach. We demonstrate the performance of the PE-UCCSD model on systems such as butadiene and para-nitroaniline in water and find that PE-UCCSD delivers comparable accuracy to classical PE-CCSD methods on such simple closed-shell systems. We also explore the challenges posed by hardware noise and propose simple error correction techniques to maintain accurate results on noisy quantum computers.
U2 - 10.48550/arXiv.2411.03852
DO - 10.48550/arXiv.2411.03852
M3 - Preprint
VL - 2411.03852
BT - Self-consistent Quantum Linear Response with a Polarizable Embedding environment
PB - arxiv.org
ER -