Order Parameter Discovery for Quantum Many-Body Systems

Nicola Mariella, Tara Murphy, Francesco Di Marcantonio, Khadijeh Najafi, Sofia Vallecorsa, Sergiy Zhuk, E. RicoarXiv 2024 · DOI · arXiv

Abstract

Quantum phase transitions reveal deep insights into the behavior of many-body quantum systems, but identifying these transitions without well-defined order parameters remains a significant challenge. In this work, we introduce a novel approach to constructing phase diagrams using the vector field of the reduced fidelity susceptibility (RFS). This method maps quantum phases and formulates an optimization problem to discover observables corresponding to order parameters. We demonstrate the effectiveness of our approach by applying it to well-established models, including the Axial Next Nearest Neighbour Interaction (ANNNI) model, a cluster state model, and a chain of Rydberg atoms. By analyzing observable decompositions into eigen-projectors and finite-size scaling, our method successfully identifies order parameters and characterizes quantum phase transitions with high precision. Our results provide a powerful tool for exploring quantum phases in systems where conventional order parameters are not readily available.

Introduction

In condensed matter physics, characterizing phase transitions typically necessitates identifying an order parameter—a physical quantity that distinguishes between different phases of matter. However, many modern quantum systems, particularly those exhibiting topological phases, lack obvious local order parameters. Traditional methods for identifying these transitions often require extensive prior knowledge of system symmetries or computationally expensive full state tomography. This work introduces a novel framework to identify phase transitions and discover their corresponding order parameters using only local measurements.

Methodology: The RFS Framework

The core of the proposed method is the Reduced Fidelity Susceptibility (RFS). Fidelity measures the sensitivity of a quantum state to infinitesimal changes in system parameters. By analyzing the reduced density matrix of a small subsystem, the authors calculate the RFS and treat its gradient as a vector field.

Overview of the RFS-based framework for phase diagram construction and order parameter discovery.

In this vector field representation, phase transitions emerge as "sources" (divergence points), while stable ground states act as "sinks" (convergence points). Once phases are mapped, the discovery of order parameters is formulated as a Quadratically Constrained Quadratic Program (QCQP). This optimization seeks a Hermitian operator $M$ that maximizes the contrast in expectation values between distinct identified phases.

Experimental Evaluation

The framework was validated using Density Matrix Renormalization Group (DMRG) simulations across three benchmark models:

  1. ANNNI Model: A 1D spin chain with competing interactions that create a complex phase diagram, including a "floating" phase.
  2. Cluster Hamiltonian: A model featuring Symmetry Protected Topological (SPT) phases, which traditionally require non-local "string" order parameters.
  3. Rydberg Atom Chain: A representation of experimental quantum simulators with long-range interactions and various crystalline phases ($Z_2, Z_3$).

Results and Discussion

The RFS method successfully mapped complex phase diagrams with high precision. For the ANNNI model, the vector field analysis clearly delineated phase boundaries.

RFS map (a) and vector field (b) for the ANNNI model, illustrating how transitions act as sources and phases as sinks.

Key findings include:

  • Universality: Finite-size scaling in the ANNNI model yielded a critical exponent $\beta \approx 0.123$, aligning closely with the Ising universality class theoretical value of $1/8$.
  • Topological Discovery: In the cluster model, the optimization routine successfully "rediscovered" the non-local string order parameter required to identify the SPT phase, using only 5-site local information.
  • High Resolution: The RFS maps for Rydberg atoms revealed intricate "ripple-like" features, suggesting a more nuanced phase structure than previously documented by other methods.
  • Efficiency: Because it relies on local reduced density matrices, the method is significantly more scalable for experimental quantum hardware than methods requiring full state reconstruction.

Conclusion

This research demonstrates that global phase information and non-local order parameters can be extracted from local reduced density matrices. By treating susceptibility as a topological landscape, the method provides a symmetry-agnostic tool for exploring quantum phases. While current work focuses on linear order parameters, future research will extend this to more exotic topological orders and real-time characterization on quantum hardware simulators.

Summary updated

Abstract

Quantum phase transitions reveal deep insights into the behavior of many-body quantum systems, but identifying these transitions without well-defined order parameters remains a significant challenge. In this work, we introduce a novel approach to constructing phase diagrams using the vector field of the reduced fidelity susceptibility (RFS). This method maps quantum phases and formulates an optimization problem to discover observables corresponding to order parameters. We demonstrate the effectiveness of our approach by applying it to well-established models, including the Axial Next Nearest Neighbour Interaction (ANNNI) model, a cluster state model, and a chain of Rydberg atoms. By analyzing observable decompositions into eigen-projectors and finite-size scaling, our method successfully identifies order parameters and characterizes quantum phase transitions with high precision. Our results provide a powerful tool for exploring quantum phases in systems where conventional order parameters are not readily available.
BibTeX
@misc{mariella2024order,
  author = {Nicola Mariella and Tara Murphy and Francesco Di Marcantonio and Khadijeh Najafi and Sofia Vallecorsa and Sergiy Zhuk and E. Rico},
  title = {{Order Parameter Discovery for Quantum Many-Body Systems}},
  year = {2024},
  eprint = {2408.01400},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2408.01400},
}
Figure