Abstract
We define a new family of random spin models with one-dimensional structure, finite-range multi-spin interactions, and bounded average degree (number of interactions in which each spin participates). Unfrustrated ground states can be described as solutions of a sparse, band diagonal linear system, thus allowing for efficient numerical analysis. In the limit of infinite interaction range, we recover the so-called XORSAT (diluted p-spin) model, that is known to undergo a random first order phase transition as the average degree is increased. Here we investigate the most important consequences of a large but finite interaction range: (i) Fluctuation-induced corrections to thermodynamic quantities; (ii) The need of an inhomogeneous (position dependent) order parameter; (iii) The emergence of a finite mosaic length scale. In particular, we study the correlation length divergence at the (mean-field) glass transition.
Introduction
Understanding the physics of glasses—materials that are structurally liquid but mechanically solid—is one of the "holy grails" of statistical mechanics. Standard mean-field models, such as the $p$-spin or XORSAT models, are mathematically elegant because every particle interacts with every other particle (infinite range). These models predict a "dynamical phase transition" where the system gets stuck in an exponential number of "pure states."
However, real-world materials are finite-dimensional; an infinite number of states cannot coexist in a small volume. The "mosaic state" scenario suggests that a finite-dimensional glass is actually a patchwork of these different states, each occupying a region defined by a "mosaic length scale" ($\ell_s$). The challenge has been finding a model that is structurally realistic enough to show this mosaic behavior but simple enough to solve. The Kac-XORSAT model bridges this gap by providing a numerically tractable look at how these length scales emerge and diverge.
The Kac-XORSAT Model
The authors propose a one-dimensional chain of spins where each spin participates in a few random constraints (linear equations mod 2) involving other spins within a finite range $R$.
Key technical innovations include:
- Linear Algebra Mapping: Because the constraints are XOR-based, finding the ground state is equivalent to solving a system of linear equations ($Hx = b \pmod 2$). This makes the model "unfrustrated" and much easier to handle than typical spin glasses.
- Optimized Transfer Matrix Algorithm: A specialized algorithm exploits the linear structure to achieve polynomial complexity ($O(LR^3)$), allowing simulations of ranges up to $R=100$ and system sizes $L$ up to 15,000 spins.
- Replica Method: Analytically, the replica method is used to calculate thermodynamic quantities and $1/R$ corrections, providing a theoretical benchmark for numerical results.
Key Insights: The Point-to-Set Experiment
To measure the mosaic length, the authors perform a "point-to-set" experiment. They fix the spins outside a box of size $\ell$ to a specific reference configuration and check if the spin at the center is forced to match.
- Free vs. Frozen: In small boxes, the center is "frozen" by the boundaries; in large boxes, the center becomes "free." The transition point defines the mosaic length $\ell_s$.
- Inhomogeneity: Unlike uniform mean-field models, the Kac model requires a position-dependent order parameter, as spins near boundaries are more likely to be frozen than those in the bulk.
Key Results
The study demonstrates that as the interaction range $R \to \infty$, the Kac-XORSAT model perfectly recovers the properties of the standard XORSAT model.
- Entropy Corrections: The authors verified $1/R$ corrections to the ground state entropy, showing how finite-range interactions "smear" the sharp transitions seen in mean-field theory.
- Correlation Divergence: Most significantly, they confirmed that the mosaic length scale $\ell_s$ diverges at the glass transition point ($\gamma_s$). They found $\ell_s \sim (\gamma_s - \gamma)^{-1}$, aligning with theoretical "mosaic" predictions.
Future Directions
While this work provides a powerful framework, it focuses on the "unfrustrated" case at zero temperature. Real glasses are often "frustrated," meaning no single configuration satisfies all constraints. Future work could extend this to finite temperatures or study Glauber dynamics to see if relaxation times scale with the mosaic length as predicted.
Summary updated