Observation of universal non-Gaussian statistics of the order parameter across a continuous phase transition

Using single-atom-resolved detection in an interacting lattice Bose gas, researchers experimentally mapped the full non-Gaussian probability distribution of the order parameter across a continuous phase transition, revealing critical scaling in high-order cumulants that aligns with quantum models rather than classical ones.

The Gist

Imagine a crowded dance floor where everyone is moving to music. Sometimes, the crowd moves in perfect, synchronized unison (like a superfluid). Other times, everyone is just jiggling randomly in their own spots (like a normal gas). The moment the music changes and the crowd shifts from dancing together to dancing alone is called a phase transition.

For decades, scientists have studied these transitions by looking at the "average" behavior of the crowd. They knew that right at the moment of change, things get chaotic. But this new paper takes a much closer look at that chaos, not just by counting heads, but by watching the exact pattern of movement for every single person.

Here is a simple breakdown of what the researchers did and found:

1. The Experiment: A Giant Quantum Dance Floor

The scientists used a cloud of about 4,000 super-cold helium atoms trapped in a grid of light (an optical lattice). Think of this grid as a giant, invisible dance floor where the atoms can hop from one spot to another.

• The Control Knob: They could adjust how much the atoms liked to interact with each other (like turning up the bass on the music).
• The Camera: They had a special camera that could see every single atom individually. When they turned off the trap, the atoms fell like rain, and the camera caught their positions. This allowed them to reconstruct exactly how the atoms were moving inside the trap.

2. The Discovery: It's Not Just "Random Noise"

Usually, when things fluctuate (like the height of waves in the ocean), we expect them to follow a "bell curve" (a Gaussian distribution). This means most of the time, things are average, and extreme highs or lows are rare but predictable.

The researchers found that right at the critical point (the exact moment the phase transition happens), the atoms stopped behaving like a normal bell curve.

• The Analogy: Imagine a crowd that is usually calm. As they get ready to switch from "dancing together" to "dancing alone," they don't just get a little jittery. They start doing wild, unpredictable, and highly specific patterns of movement that a simple average cannot describe.
• The Result: The "shape" of the crowd's movement became non-Gaussian. It had a weird, unique shape that is the same for all systems in this specific "universe" of physics, regardless of the tiny details of the atoms.

3. The "Energy Landscape" Map

To understand this weird shape, the scientists built a map called an "effective thermodynamic potential."

• The Analogy: Think of this as a landscape of hills and valleys.
  • When the atoms are dancing together (Ordered): The landscape has a deep valley where the atoms like to sit.
  • When they are dancing alone (Disordered): The landscape is a flat hill with the lowest point right in the middle (zero movement).
  • At the Transition: The valley disappears, and the landscape becomes a strange, flat plateau. The paper shows that the shape of this landscape changes in a very specific, universal way right at the moment of the switch.

4. The "Sign Flip" Surprise

The most exciting part of the paper involves something called cumulants. In simple terms, these are mathematical tools used to measure how "weird" or "non-average" the fluctuations are.

• The Observation: As the scientists turned the control knob through the transition, these mathematical numbers didn't just get bigger or smaller. They abruptly flipped signs (from positive to negative or vice versa).
• The Analogy: Imagine a thermometer that doesn't just go up and down, but suddenly jumps from "hot" to "cold" and back again in a specific pattern right when the weather changes.
• Why it matters: The researchers found that this "sign flip" is a universal rule. It happens in their quantum experiment, and it matches what computer simulations predict for similar systems. However, it doesn't happen in old-fashioned classical models. This proves that the weird behavior is driven by quantum mechanics, not just simple heat.

5. Why This is a Big Deal

Usually, to see these weird quantum effects, you need a perfect, infinite system. But the researchers showed that even in a small, finite system (like their trap with 4,000 atoms), you can still see these universal rules.

In summary:
The paper is like taking a high-definition video of a crowd changing its dance style. Instead of just saying "they changed," the researchers showed that the way they changed followed a strict, universal script that only quantum physics can write. They proved that by looking at the "weirdness" of the fluctuations (the non-Gaussian statistics and the sign flips), we can understand the fundamental rules of how matter changes states, even in small, imperfect systems.

Technical Summary

Technical Summary: Observation of Universal Non-Gaussian Statistics of the Order Parameter Across a Continuous Phase Transition

Problem Statement
Second-order phase transitions are fundamentally characterized by critical scaling and universality, typically described by critical exponents derived from the singular behavior of thermodynamic quantities. While Landau mean-field theory successfully distinguishes ordered and disordered phases via the average order parameter, it neglects the large fluctuations that drive the transition. These fluctuations are captured by the probability distribution function (PDF) of the order parameter. Theoretical frameworks, including Wilson's renormalization group, predict that near the critical point, the order-parameter statistics become non-Gaussian, particularly in finite-size systems where the correlation length $\xi$ exceeds the system size $L$. However, experimentally characterizing these non-Gaussian statistics has remained largely unexplored due to the narrowness of the critical regime in macroscopic systems and the inherent difficulty in measuring full counting statistics. Previous experimental reports of non-Gaussian statistics have been limited to classical, mesoscopic systems like liquid crystals.

Methodology
The authors investigate the superfluid-to-normal fluid phase transition in an interacting lattice Bose gas, realized using approximately 4,000 ultracold metastable helium ($^4$He*) atoms in a cubic optical lattice. This system implements the Bose-Hubbard model, where the control parameter $u = U/J$ (the ratio of on-site interaction energy to tunneling energy) is tuned by adjusting the lattice depth. The experiment is conducted in a harmonic trap, creating an inhomogeneous system where the superfluid and normal phases can coexist.

The core experimental technique involves single-atom-resolved detection in momentum space. After releasing the atoms from the lattice, they are detected one-by-one on a micro-channel plate. This allows for the reconstruction of the full 3D momentum distribution for each experimental shot. The order parameter amplitude $|\psi|$ is defined as the square root of the number of atoms $N_0$ detected in the central momentum mode ($k \approx 0$), corresponding to the condensate fraction. By performing approximately 1,200 experimental repetitions for various values of $u$, the authors reconstruct the full probability distribution $P(|\psi|)$. To mitigate shot-to-shot fluctuations in the total atom number $N$, the data is post-selected on $N$.

Key Contributions and Results

1. Measurement of Full Order-Parameter PDF:
   The study provides the first experimental measurement of the full probability distribution of the order-parameter amplitude across a continuous phase transition in a quantum many-body system.

   • Ordered Phase ($u < u_c$): The distribution is Gaussian with a non-zero mean, reflecting the macroscopic population of the condensate.
   • Disordered Phase ($u > u_c$): The distribution is Gaussian with a maximum at zero, consistent with thermal statistics.
   • Critical Regime ($u \approx u_c$): The distribution exhibits a distinct, non-Gaussian shape with a large width, capturing the enhanced fluctuations of the order parameter.

2. Effective Thermodynamic Potential:
   The authors reconstruct an effective thermodynamic potential $F(|\psi|) = -\ln P(|\psi|)$, analogous to the Landau free energy. This potential displays a non-trivial minimum in the superfluid phase which vanishes at the transition point. While the bulk of the distribution is well-described by a low-order polynomial (Landau-like form), the authors note that beyond-mean-field effects (specific to the 3D XY universality class) manifest only in the tails of the distribution as rare events, which are difficult to resolve with the current number of experimental shots.

3. Critical Scaling and Exponents:
   The average order parameter $\langle|\psi|\rangle$ exhibits power-law scaling $\langle|\psi|\rangle \propto (u_c - u)^\beta$ with a critical exponent $\beta = 0.35(1)$. This value is consistent with the 3D XY universality class ($\beta_{3DXY} \approx 0.3485$) and distinct from the mean-field prediction ($\beta_{MF} = 0.5$). The critical point $u_c$ is determined to be $38.1(2)$, which aligns with Quantum Monte Carlo (QMC) predictions for a trapped system.

4. High-Order Cumulants and Sign Changes:
   A primary finding is the characterization of the non-Gaussian nature of the statistics through high-order cumulants ($\kappa_n$ for $n \geq 3$).

   • Non-Gaussian Crossover: Higher-order cumulants vanish deep in the ordered and disordered phases but take large values near the transition, signaling the crossover from Gaussian to non-Gaussian statistics.
   • Abrupt Sign Changes: The authors observe abrupt sign changes in the high-order cumulants near the transition point.
   • Universality and Scaling: Numerical simulations of homogeneous systems belonging to the 3D XY universality class reproduce these sign changes, demonstrating that they obey critical scaling.
   • Quantum vs. Classical: The experimental behavior of the cumulants (specifically for $n \geq 4$) is not reproduced by classical models, even when accounting for system geometry and inhomogeneity. Instead, the data is captured by a low-temperature quantum model (quantum rotor model), indicating that the observed deviations arise from universal corrections to scaling induced by quantum fluctuations.

Significance
The paper underscores the crucial role of order-parameter statistics in probing critical phenomena and universality. By leveraging single-atom-resolved detection in mesoscopic quantum simulators, the authors demonstrate a complementary approach to standard critical exponent measurements. This method is highly sensitive to the nature of critical fluctuations, including subtle quantum effects at thermal transitions. The observation of universal sign changes in cumulants provides a new diagnostic tool for identifying phase transitions, with potential relevance to other fields such as the study of the QCD phase diagram. The work challenges theoretical and numerical methods to accurately describe order-parameter fluctuations in finite-size, inhomogeneous quantum systems.