Use VQE to calculate the ground energy of hydrogen molecules on IBM Quantum

Maomin Qing, Wei XiearXiv 2023 · DOI · arXiv

TL;DR

This paper demonstrates the practical implementation of the Variational Quantum Eigensolver (VQE) algorithm on IBM’s quantum hardware to calculate the ground state energy of a hydrogen molecule ($H_2$). By combining quantum state preparation with classical optimization, the researchers achieved results within approximately 1.47% of the actual molecular energy, highlighting both the potential and the current noise-related limitations of Noisy Intermediate-Scale Quantum (NISQ) devices.

What Problem Does This Paper Address?

Calculating the ground state energy of molecules is a cornerstone of computational chemistry, essential for designing new drugs, catalysts, and materials. However, classical computers struggle with this because the complexity of the Schrödinger equation grows exponentially with the number of electrons.

Quantum computers offer a natural way to simulate these systems, but we are currently in the "NISQ" era, where quantum bits (qubits) are prone to errors and decoherence. The VQE algorithm was designed specifically for this era. It is a hybrid algorithm that offloads the most "quantum" part of the problem—representing the complex wave function—to a quantum processor, while using a standard classical computer to handle the optimization. This paper explores whether current IBM Quantum hardware is up to the task for a simple system like $H_2$.

Methods

The authors followed a structured pipeline to bridge the gap between chemistry and quantum circuits:

  1. Hamiltonian Construction: They started with the electronic Hamiltonian for $H_2$, simplifying it using the Born-Oppenheimer approximation (assuming nuclei are stationary).
  2. Mapping: To run this on a quantum computer, they converted fermionic operators (representing electrons) into qubit operators (Pauli matrices) using the Jordan-Wigner transformation.
  3. The Ansatz: They used the Unitary Coupled Cluster Singles and Doubles (UCCSD) method to create a parameterized trial wavefunction (the "Ansatz"). This is the "guess" that the quantum computer starts with.
  4. Hybrid Loop:
    • Quantum: The IBM Quantum device (ibmq_jakarta) prepares the state and measures the energy.
  5. Classical: The L-BFGS optimizer adjusts the parameters of the Ansatz to find a lower energy state.
  6. Software Stack: The implementation relied on Qiskit (IBM’s quantum SDK) and PySCF for initial molecular data.

Datasets

The study did not use a traditional large-scale dataset. Instead, it utilized molecular geometry data for the hydrogen molecule ($H_2$) provided by the PySCF library. The researchers evaluated the energy across a range of atomic distances, starting from 0.2 Å to 4.0 Å, with a step size of 0.05 Å, to plot the dissociation curve.

Key Insights and Ideas

  • The Power of Hybridization: The paper reinforces that VQE is robust because the classical optimizer can "soak up" some of the systematic errors present in the quantum hardware.
  • Mapping is Crucial: The transition from the Schrödinger equation to a string of Pauli matrices (like $Z \otimes I \otimes Z \otimes I$) is the critical "translation" layer that allows chemistry to be spoken in the language of quantum gates.
  • Initial States Matter: The authors initialized the circuit to the Hartree-Fock state $|0101\rangle$. A good starting "guess" significantly accelerates the convergence of the classical optimizer.

Key Results

  • Accuracy: At the equilibrium geometric distance of 0.725 Å, the VQE algorithm calculated an energy of -1.134167 Hartree. Compared to the actual value of -1.117506 Hartree, this represents an error of approximately 1.47%.
  • Energy Curve: The VQE results closely followed the "exact" numerical solution curve across various atomic distances, demonstrating that the algorithm can successfully capture the physics of molecular bonding and dissociation.
  • Hardware Validation: The experiment successfully ran on a real 7-qubit quantum processor (ibmq_jakarta), proving the accessibility of these tools for researchers.

Limitations and Future Work

The authors identify three primary hurdles for moving beyond simple molecules like $H_2$:

  1. Hardware Noise: Background noise and decoherence on the quantum device remain the biggest contributors to error. As circuits get deeper (for larger molecules), this noise will eventually overwhelm the signal.
  2. The "Stationary Nucleus" Assumption: The model assumes nuclei don't move, which is a simplification of real-world molecular dynamics.
  3. Scaling: While $H_2$ requires few qubits, larger molecules require significantly more qubits and more complex "Ansatz" circuits, which are currently difficult to optimize and execute reliably on NISQ hardware.

Future work will likely focus on error mitigation techniques to reduce the 1.47% gap and exploring more efficient mapping methods to handle larger chemical systems.

Summary updated

Abstract

Quantum computing has emerged as a promising technology for solving problems that are intractable for classical computers. In this study, we introduce quantum computing and implement the Variational Quantum Eigensolver (VQE) algorithm using Qiskit on the IBM Quantum platform to calculate the ground state energy of a hydrogen molecule. We provide a theoretical framework of quantum mechanics, qubits, quantum gates, and the VQE algorithm. Our implementation process is described, and we simulate the results. Additionally, experiments are conducted on the IBM Quantum platform, and the results are analyzed. Our fi ndings demonstrate that VQE can effi ciently calculate molecular properties with high accuracy. However, limitations and challenges in scaling the algorithm for larger molecules are also identifi ed. This work contributes to the growing body of research on quantum computing and highlights the potential applications of VQE for real-world problem-solving.
BibTeX
@misc{qing2023use,
  author = {Maomin Qing and Wei Xie},
  title = {{Use VQE to calculate the ground energy of hydrogen molecules on IBM Quantum}},
  year = {2023},
  eprint = {2305.06538},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2305.06538},
}
Figure