Abstract:We generalize the proof of stability of topological order, due to Bravyi, Hastings and Michalakis, to stabilizer Hamiltonians corresponding to low-density parity check (LDPC) codes without the restriction of geometric locality in Euclidean space. We consider Hamiltonians $H_0$ defined by $[[N,K,d]]$ LDPC codes which obey certain topological quantum order conditions: (i) code distance $d \geq c \log(N)$, implying local indistinguishability of ground states, and (ii) a mild condition on local and global compatibility of ground states; these include good quantum LDPC codes, and the toric code on a hyperbolic lattice, among others. We consider stability under weak perturbations that are quasi-local on the interaction graph defined by $H_0$, and which can be represented as sums of bounded-norm terms. As long as the local perturbation strength is smaller than a finite constant, we show that the perturbed Hamiltonian has well-defined spectral bands originating from the $O(1)$ smallest eigenvalues of $H_0$. The band originating from the smallest eigenvalue has $2^K$ states, is separated from the rest of the spectrum by a finite energy gap, and has exponentially narrow bandwidth $\delta = C N e^{-\Theta(d)}$, which is tighter than the best known bounds even in the Euclidean case. We also obtain that the new ground state subspace is related to the initial code subspace by a quasi-local unitary, allowing one to relate their physical properties. Our proof uses an iterative procedure that performs successive rotations to eliminate non-frustration-free terms in the Hamiltonian. Our results extend to quantum Hamiltonians built from classical LDPC codes, which give rise to stable symmetry-breaking phases. These results show that LDPC codes very generally define stable gapped quantum phases, even in the non-Euclidean setting, initiating a systematic study of such phases of matter.
What problem does this paper attempt to address?
### Problems the Paper Attempts to Solve
This paper attempts to address the problem of stabilizing quantum phases defined by quantum error-correcting codes, particularly low-density parity-check (LDPC) codes. Specifically, the authors aim to demonstrate that LDPC stabilizer codes defined on non-Euclidean geometric structures can maintain their topological order and energy gap stability when subjected to graph-local perturbations.
### Background and Motivation
1. **Quantum Information Theory and Quantum Phases**:
- Over the past 20 years, quantum information theory has made significant progress in understanding quantum phases of matter, especially in the study of zero-temperature topological order.
- Stabilizer codes are among the most extensively studied error-correcting codes. They define gapped and exactly solvable Hamiltonians, which are sums of commuting projection operators, with their ground state space coinciding with the code space.
2. **Limitations of Existing Results**:
- The work of Bravyi, Hastings, and Michalakis (BHM) demonstrated that gapped, commuting projection operator Hamiltonians on Euclidean geometric structures remain stable under sufficiently weak local perturbations.
- However, these results are limited to geometrically local Hamiltonians and do not apply to non-Euclidean geometric structures or graph-local Hamiltonians.
3. **New Developments in Non-Euclidean LDPC Codes**:
- Recently, the field of quantum error correction has seen breakthroughs with low-density parity-check codes (LDPC codes) defined on non-Euclidean expander graphs.
- These codes exhibit optimal error-correcting capabilities under certain metrics, which cannot be achieved in local Euclidean geometric structures.
### Specific Problems
1. **Can Non-Local and/or Non-Euclidean LDPC Stabilizer Codes be Associated with Stable Quantum Phases?**
- The authors aim to answer this question by proving the stability of these codes under graph-local perturbations.
- This is not only a theoretically important question but also has significant implications for practical applications in quantum error correction and quantum computing.
2. **Challenges in Stability Analysis**:
- Despite the physical intuition that all LDPC codes should be stable, there are some technical challenges.
- For instance, good codes have an exponential ground state degeneracy and extensive ground state entropy, which may lead to large energy differences in degenerate perturbation theory.
- Additionally, the stability of ground state degeneracy might imply a stable violation of the third law, which is a very unusual phenomenon in statistical physics.
### Main Contributions
1. **Proving the Stability of Non-Euclidean LDPC Codes under Graph-Local Perturbations**:
- The authors use an iterative rotation method to gradually eliminate non-frustration terms in the Hamiltonian, thereby proving the stability of these codes under graph-local perturbations.
- They show that in the perturbed Hamiltonian, the energy gap remains finite, and the ground state splitting decays exponentially with the code distance.
2. **Extension to Classical LDPC Codes**:
- The authors also extend their results to classical LDPC codes with symmetry-breaking order, showing that these codes exhibit stability under symmetry-preserving perturbations.
3. **New Technical Tools**:
- The authors introduce an iterative process called the "KAM" (Kolmogorov-Arnold-Moser) process to track the growth of the support of rotation operators and the number of excited stabilizers.
### Conclusion
Through rigorous mathematical proofs, this paper establishes the stability of non-Euclidean LDPC stabilizer codes under graph-local perturbations, providing an important theoretical foundation for understanding and designing more complex quantum error-correcting codes. This result not only extends existing stability theories but also provides theoretical support for the experimental realization and verification of these codes.