Weak first-order phase transition out of the classical kagome spin liquid

Using a spin-component expansion, this paper resolves a long-standing debate by demonstrating that the classical kagome Heisenberg antiferromagnet undergoes a weak first-order phase transition into a $\sqrt{3}\times\sqrt{3}$ ordered state at low temperatures, rather than remaining a spin liquid as previously suggested by Monte Carlo simulations.

The Gist

Imagine a crowded dance floor where everyone is trying to hold hands with their neighbors, but the room is shaped in a way that makes it impossible for everyone to be happy at the same time. This is the world of frustrated magnets, specifically the kagome lattice (a pattern of triangles inside triangles, like a woven basket).

For over 30 years, physicists have been arguing about what happens to the dancers (the magnetic spins) when the music stops and the room gets freezing cold.

The Great Debate: A Smooth Slide or a Hard Crash?

The Old Story (Monte Carlo Simulations):
Previous computer simulations suggested that as the room cooled down, the dancers didn't suddenly snap into a rigid formation. Instead, they slowly drifted from a chaotic, swirling mess (a "spin liquid") into a more organized, flat pattern (the $\sqrt{3} \times \sqrt{3}$ phase). It was thought to be a gentle, smooth transition, like water slowly turning into slush.

The New Story (This Paper):
Cecilie Glittum and Olav F. Sylju˚asen used a new mathematical tool called Nematic Bond Theory (NBT) to look at the problem again. They found that the old story was missing a crucial detail.

They discovered that the transition isn't a smooth slide. It's a weak first-order phase transition.

• The Analogy: Imagine a ball rolling down a hill. In the old view, the ball rolled smoothly into a valley. In this new view, the ball rolls down, hits a small, sharp cliff, and drops into the valley.
• The "Weak" Part: The cliff isn't a giant mountain; it's a tiny step. The energy difference (latent heat) is so small it's almost invisible, which is why previous computer simulations missed it. They were looking for a big crash, but the transition was a subtle "clunk."

The Mystery of the "Frozen" Dance

Once the dancers finally settle into their organized $\sqrt{3} \times \sqrt{3}$ pattern, do they stop moving completely?

• The Old View: Simulations suggested the dancers kept wiggling and stumbling around, never fully locking into place. The "order" was weak and suppressed by invisible walls (domain walls) and swirling vortices.
• The New View: The authors show that as the temperature hits absolute zero, the dancers do lock in perfectly. The "ordered moment" (how perfectly they align) reaches its maximum possible value. The chaos is gone; the dance is complete.

Why Did the Old Computers Miss This?

The authors explain that the old computer methods (Monte Carlo simulations) are like trying to watch a movie through a foggy window at low temperatures.

1. The Fog: At very low temperatures, the computer algorithms get "stuck" in local loops, unable to explore the whole room efficiently.
2. The Mix-Up: Because the computers got stuck, they saw a messy mixture of the chaotic state and the ordered state, making it look like a smooth crossover rather than a sharp drop.
3. The New Tool: NBT doesn't try to simulate every single dancer's move one by one. Instead, it calculates the "energy score" of the whole room directly. It's like looking at the blueprint of the building rather than trying to count every person walking through the door. This allowed them to see the tiny "cliff" (the phase transition) that the others missed.

A Tale of Two Lattices

To prove their method wasn't just making things up, the authors tested it on a different shape called the pyrochlore lattice (a 3D version of the problem).

• The Result: On this 3D shape, the dancers never lock into a rigid pattern, no matter how cold it gets. They stay in a chaotic spin liquid forever.
• The Lesson: This proves that the "locking in" behavior on the kagome lattice is a real, unique feature of that specific shape, not a glitch in their new mathematical tool.

Summary

This paper settles a 30-year-old argument by showing that the classical kagome spin liquid doesn't just slowly fade into order. Instead, it undergoes a tiny, sharp, first-order jump into a perfectly ordered state as it reaches absolute zero. The "weakness" of this jump is why it was hidden for so long, but with a better mathematical lens, the authors have finally seen the cliff edge.

Technical Summary

Technical Summary: Weak First-Order Phase Transition out of the Classical Kagome Spin Liquid

Problem Statement
The low-temperature fate of the classical Heisenberg antiferromagnet on the kagome lattice has been a subject of debate for over three decades. While the system exhibits a highly correlated, disordered "spin liquid" regime at intermediate temperatures, the question remains whether this state persists down to zero temperature or if thermal fluctuations eventually select an ordered state via an "order-by-disorder" mechanism. Previous Monte Carlo (MC) simulations suggested a sharp crossover into a coplanar $\sqrt{3} \times \sqrt{3}$ phase with a substantially reduced ordered moment, potentially suppressed by proliferating domain walls or vortices. However, these simulations face severe algorithmic slowing down at low temperatures, making it difficult to distinguish between a true thermodynamic phase transition and a crossover, and to determine if the ordered moment truly saturates at $T=0$.

Methodology
To address these limitations, the authors employ Nematic Bond Theory (NBT), a self-consistent approach based on a large-$N_s$ expansion (where $N_s=3$ is the number of spin components). NBT extends the Self-Consistent Gaussian Approximation (SCGA) by:

1. Including momentum-dependent self-energies.
2. Enforcing the local unit-length spin constraint more accurately by introducing a fluctuating constraint field. The uniform component is treated self-consistently, while non-uniform components are incorporated diagrammatically.

This method allows for the direct comparison of free energies for competing phases in large system sizes without relying on direct configuration sampling, thereby bypassing the critical slowing down issues inherent in MC simulations at low temperatures. The authors solve the resulting Dyson equations via iteration, tracking converged branches from different initial self-energies to identify the thermodynamically stable state (the branch with the lowest free energy).

Key Results

1. Nature of the Transition: The study establishes that the classical kagome spin liquid and the $\sqrt{3} \times \sqrt{3}$ ordered phase are distinct thermodynamic phases separated by a weak first-order phase transition. The transition occurs at a critical temperature $T_c \approx 1.59 \times 10^{-3}$ (in units of the exchange coupling $J_1=1$).
2. Latent Heat: The transition is characterized by a very small latent heat per spin, $\ell_h \approx 1.04 \times 10^{-4}$, confirming its weak first-order nature.
3. Ordered Moment Saturation: Contrary to MC findings where the ordered moment remains suppressed, the NBT results show that the ordered moment of the $\sqrt{3} \times \sqrt{3}$ phase saturates as $T \to 0$. The system orders completely at zero temperature.
4. Specific Heat: The specific heat $c_v$ exhibits a discontinuity at $T_c$ and follows a $\sqrt{T}$ behavior at very low temperatures in the ordered phase, consistent with theoretical predictions for coplanar states.
5. Phase Diagram with $J_2$: When second-neighbor interactions ($J_2$) are introduced, the transition belongs to a line of first-order transitions that terminates at critical points. A triple point exists where the spin liquid, $\sqrt{3} \times \sqrt{3}$, and $\vec{q}=0$ phases coexist.
6. Comparison with Pyrochlore: As a control, the authors applied NBT to the pyrochlore antiferromagnet. In that case, no ordered branch overtakes the spin liquid at low temperatures, and the system remains disordered down to the lowest studied temperatures. This contrast suggests the kagome result is a genuine physical distinction rather than an artifact of the NBT method.

Discrepancy with Monte Carlo Simulations
The authors attribute discrepancies with earlier MC results to two main factors:

• Sampling Issues: The weak first-order transition involves a latent heat that creates a peak in the specific heat that is extremely narrow and sits on top of a discontinuity. In finite-size MC simulations, this peak is difficult to resolve and may appear as a broad plateau or crossover.
• Equilibration: MC algorithms may struggle to equilibrate across different domain-wall sectors at low temperatures, potentially resulting in configurations that are mixtures of phases, leading to a permanently reduced apparent ordered moment.

Significance and Claims
The paper claims to resolve long-standing questions regarding the kagome Heisenberg antiferromagnet by demonstrating that:

• The system does not merely drift into a coplanar regime via crossover physics.
• Instead, it exits the spin-liquid regime through a genuine, albeit weak, first-order phase transition.
• The low-temperature $\sqrt{3} \times \sqrt{3}$ phase is fully ordered with a saturated moment at $T=0$.

The authors note that NBT is an approximate method (neglecting vertex corrections) and typically overestimates critical temperatures by 10-20%; thus, the specific temperature values presented should be viewed as estimates. However, the qualitative distinction between the kagome and pyrochlore behaviors, and the identification of a first-order transition, are presented as robust findings supported by the method's ability to compare free energies directly.