Gap-free proof confirms magnetic order in the 2D random-bond quantum Ising model

Peierls's argument starts from a simple picture of a magnet. In the Ising model, each spin sits on a lattice, points up or down, and interacts with its nearest neighbours ^[14]. "An individual spin does not care whether it is up or down; it only wants to align with its neighbor. The lowest-energy configurations have all spins pointing up, or all pointing down," Lucas said ^[9].

Where an up region meets a down region, the mismatched bonds form a domain wall, and the wall's energy cost grows with its length ^[7]. A long wall is expensive. It can also take an enormous number of shapes, and each shape is another chance for one of them to appear ^[7]. At low temperature the energy cost wins, and the chance of a wall spanning the whole system becomes vanishingly small ^[7]. The requirement that large walls cost more energy than their number of shapes can offset is called the Peierls condition ^[8].

Quantum mechanics changes the ground state itself. "In quantum mechanics, the lowest-energy (ground) state is (roughly) a macroscopic superposition of the up state and the down state," Lucas said ^[15]. Proving that such a phase is stable means showing the system can be changed slightly without leaving it. Earlier proofs did that through the energy gap, the minimum energy needed to excite the system above its ground state ^[13].

Lucas explained why the gap helps. "Intuitively, this is because you can imagine 'tuning a knob' to convert the system between two states in the same phase, and due to the gap, the knob tuning does not excite the gapless modes (e.g., sound), preserving the character of the state," he said ^[12]. Gapped phases are the exception. "Most phases we are familiar with are not gapped," Lucas said ^[11]. Fluids and solids can be excited by sound waves of arbitrarily low energy, he said, but an Ising ferromagnet is gapped ^[11]. The disordered random-bond version is gapless, so the gap-based proofs did not reach it ^[3].

Lucas said the gap-free route began with a technique that he and Chao, his former Ph.D. student, developed in 2024 to place strong constraints on where many-body quantum states are supported ^[5][10]. "This was a highly unusual idea, and it allowed us to look at the problem in the current paper from a fresh perspective," Lucas said ^[10].

The result holds only under the model's assumptions. It concerns a model magnet, and the phys.org account describes no test on a physical sample ^[2]. Within the model, I think the conjecture can be treated as settled for weak fluctuations in two dimensions ^[1][2].