Observation of centrality-dependent dijet transverse momentum imbalance in O+O and Ne+Ne collisions at $\sqrt{s_{NN}}$ = 5.36 TeV with the ATLAS detector

The ATLAS experiment observed a centrality-dependent dijet transverse momentum imbalance in O+O and Ne+Ne collisions at $\sqrt{s_{NN}}$ = 5.36 TeV, providing evidence for medium-induced partonic energy loss and establishing a new regime for studying quark-gluon plasma effects in small collision systems.

The Gist

The Big Picture: Smashing Tiny Balls to Find a "Super-Fluid"

Imagine you are a scientist trying to understand what happens when you mix two things together at incredibly high speeds. In the world of particle physics, the "things" are atomic nuclei, and the "speed" is nearly the speed of light.

For decades, scientists have been smashing huge, heavy balls (like Lead or Xenon atoms) together to create a state of matter called Quark-Gluon Plasma (QGP). Think of QGP as a super-hot, super-dense "soup" where the tiny particles that make up atoms (quarks and gluons) melt together and flow freely, like water molecules in a boiling pot.

The big question has always been: How small can this "soup" be? Does it only form when you smash giant atoms together, or can it happen even when you smash tiny atoms?

This paper from the ATLAS experiment at CERN says: It can happen with tiny atoms too.

The Experiment: The "Light" Collision

Usually, to make this soup, scientists smash heavy nuclei (like Lead, which has 82 protons and neutrons). In this study, they decided to try something smaller. They smashed together:

• Oxygen atoms (16 particles inside)
• Neon atoms (20 particles inside)

These are much lighter than Lead. It's like comparing smashing two bowling balls together (Lead) versus smashing two tennis balls together (Oxygen/Neon).

They did this at the Large Hadron Collider (LHC) in 2025, using a massive detector called ATLAS. They collected data from thousands of these collisions.

The Test: The "Jet" and the "Mud"

To see if a "soup" (QGP) was formed, the scientists didn't just look at the mess after the crash. They looked at something specific: Jets.

The Analogy:
Imagine two powerful water hoses (jets) spraying water in opposite directions. In a normal, empty room (a proton-proton collision), the water streams shoot out perfectly balanced. They have the same pressure and go in opposite directions.

Now, imagine spraying those same hoses through a thick wall of mud (the QGP soup).

• One hose might shoot through a thin patch of mud and keep its pressure.
• The other hose might shoot through a thick patch of mud, get slowed down, and lose energy.

When the hoses stop, the one that hit the mud will be weaker than the other. This imbalance tells you that something "muddy" was in the way.

In the paper, they call this imbalance $x_J$.

• If the jets are balanced, $x_J$ is close to 1.
• If one jet loses energy to the "mud," $x_J$ gets smaller.

The Discovery: The "Mud" Exists in Tiny Collisions

The scientists looked at the Oxygen and Neon collisions and compared them to collisions of single protons (which have no "mud").

1. The "Central" Crash: When the Oxygen or Neon atoms hit each other dead-center (like two billiard balls hitting perfectly), the jets became unbalanced. One jet lost significant energy compared to the other.
2. The "Peripheral" Crash: When the atoms just grazed each other (like a glancing blow), the jets stayed balanced, just like in the proton collisions.

What this means:
The fact that the jets lost energy only in the "dead-center" crashes proves that a dense, fluid-like medium was created in the middle of the collision. The jets had to travel through this medium and got slowed down.

This is a huge deal because it shows that you don't need giant atoms to create this "soup." Even with tiny atoms like Oxygen and Neon, if you smash them hard enough and head-on, you create a tiny droplet of Quark-Gluon Plasma.

Why This Matters (According to the Paper)

The paper claims this discovery helps scientists understand how far particles have to travel through this "soup" to lose energy.

• In big collisions (Lead), the "soup" is huge, and particles travel a long way through it.
• In these tiny collisions (Oxygen/Neon), the "soup" is a tiny droplet.

By seeing that energy loss still happens in these tiny droplets, the scientists have found a new, smaller playground to study how this "super-fluid" behaves. They can now measure exactly how the size of the droplet changes the amount of energy lost, which helps refine our understanding of the laws of physics that govern the early universe.

Summary in One Sentence

The ATLAS experiment proved that even when smashing tiny atoms like Oxygen and Neon together, they create a microscopic "super-fluid" soup that slows down high-speed particles, showing that this exotic state of matter exists in much smaller systems than previously confirmed.

Technical Summary

Technical Summary: Observation of Centrality-Dependent Dijet Transverse Momentum Imbalance in O+O and Ne+Ne Collisions

Problem and Motivation
A primary objective of the heavy-ion program at the Large Hadron Collider (LHC) is to characterize the quark–gluon plasma (QGP), a deconfined state of strongly interacting matter. High-transverse-momentum ($p_T$) jets serve as powerful probes of this medium, undergoing "jet quenching"—a reduction in jet energy and redistribution of energy to large angles—as they traverse the QGP. While evidence for path-length-dependent energy loss has been established in non-central lead-lead (Pb+Pb) collisions, the interpretation of such effects in smaller collision systems remains challenging. In proton-nucleus collisions, unambiguous evidence for jet quenching has been elusive, and in peripheral Pb+Pb collisions, fluctuations in initial-state nucleon configurations complicate the determination of effective path lengths.

Light-ion collisions, specifically Oxygen-Oxygen (O+O) and Neon-Neon (Ne+Ne), offer a new regime to investigate these phenomena. These systems possess a substantially smaller transverse extent than Pb+Pb, resulting in shorter in-medium path lengths (approximately a factor of two smaller). Furthermore, central light-ion collisions provide a more symmetric overlap geometry compared to peripheral Pb+Pb interactions, potentially reducing geometric uncertainties associated with interpreting path-length-dependent effects. Recent measurements have indicated collective behavior in O+O and Ne+Ne collisions, suggesting QGP formation, yet hard probes (such as jets) have not yet definitively demonstrated medium-induced effects in these small systems. A key challenge in small systems is "centrality bias," where correlations between hard-process production and the soft underlying event used to define centrality can distort measurements based on absolute yields.

Methodology
The ATLAS Collaboration analyzed data collected in 2025 at a nucleon–nucleon center-of-mass energy of $\sqrt{s_{NN}} = 5.36$ TeV. The dataset comprises:

• $8.0 \text{ nb}^{-1}$ of O+O collisions.
• $1.0 \text{ nb}^{-1}$ of Ne+Ne collisions.
• $386 \text{ pb}^{-1}$ of proton-proton ($pp$) collisions at the same energy, serving as the reference.

Event Selection and Centrality:
Events were selected using minimum-bias and jet triggers. Centrality intervals (0–10%, 10–20%, 20–40%, 40–60%, and 60–80%) were defined based on the total transverse energy ($\Sigma E_T^{FCal}$) measured in the ATLAS Forward Calorimeters (FCal), mapped to the inelastic cross-section using a Monte Carlo Glauber model. Pileup events were rigorously suppressed by requiring a single high-quality reconstructed vertex and utilizing correlations between Zero Degree Calorimeter (ZDC) energy and $\Sigma E_T^{FCal}$.

Jet Reconstruction and Observable:
Jets were reconstructed using the anti-$k_t$ algorithm with a radius parameter $R=0.4$ within $|y| < 2.1$. The analysis focused on dijet pairs where the leading ($p_{T1}$) and sub-leading ($p_{T2}$) jets are nearly back-to-back in azimuth ($\Delta\phi > 7\pi/8$). The primary observable is the dijet transverse momentum imbalance, $x_J$:
$$x_J = \frac{p_{T2}}{p_{T1}}$$
where $0.32 < x_J < 1$ and $63 < p_{T1} < 251$ GeV.

Analysis Techniques:
To mitigate centrality bias and dependence on absolute rates, the analysis utilized "self-normalized" distributions, defined as $\frac{1}{N_{pair}} \frac{dN_{pair}}{dx_J}$. This approach normalizes the distribution to unit area, making it less sensitive to selection biases and uncertainties in nuclear parton distribution functions (nPDFs).

• Background Subtraction: An event-by-event background subtraction procedure estimated the underlying-event transverse energy density ($\rho$). Uncorrelated jet pairs were estimated using a sideband method in $\Delta\phi$ and subtracted.
• Unfolding: Detector effects, including jet energy scale (JES), jet energy resolution (JER), and reconstruction inefficiencies, were corrected using an iterative Bayesian unfolding procedure. Two-dimensional unfolding was performed to account for bin migration and the interchange of leading/sub-leading jet ordering.
• Systematics: Uncertainties were evaluated for JES, JER, combinatorial background subtraction, and the unfolding prior. The total systematic uncertainty generally remained below 10–15% for most $x_J$ bins.

Key Results
The measured self-normalized $x_J$ distributions in O+O and Ne+Ne collisions exhibit a clear centrality dependence when compared to the $pp$ reference:

1. Centrality Trend: As collisions become more central (increasing event activity), the $x_J$ distributions show increasingly large deviations from the $pp$ baseline. Specifically, there is an increase in the relative contribution of dijets with lower $x_J$ values (indicating greater energy imbalance).
2. Statistical Significance: The deviation from the null hypothesis (no modification, i.e., ratio = 1) was quantified using a $\chi^2$ test.
   • In O+O collisions, significant modifications ($>5\sigma$) were observed in the 0–10% and 10–20% centrality intervals, persisting up to the 20–40% interval. The significance reached $10.2\sigma$ in the 0–10% centrality bin for $79 < p_{T1} \le 89$ GeV.
   • In Ne+Ne collisions, significant modifications were observed in the 0–10% and 10–20% centrality intervals, with a significance of $8.9\sigma$ in the 0–10% bin for the same $p_{T1}$ range.
   • Modifications were not statistically significant in more peripheral collisions (40–60% and 60–80%) for either system.
3. Energy Dependence: The modification remains significant in O+O (Ne+Ne) collisions up to leading-jet $p_{T1}$ of approximately 141 GeV (126 GeV) in the most central collisions.
4. Theoretical Comparison:
   • The observed modifications are incompatible with predictions based solely on nuclear PDF (nPDF) and isospin effects (using EPPS21 and CT18 sets), which predict deviations of less than 4% from unity.
   • Comparisons with the Hybrid and Jewel+Trajectum models show that while both predict a modification in central O+O collisions, the Hybrid model generally underestimates the magnitude of the effect, while Jewel+Trajectum overestimates it. The inclusion of Molière scatterings in the Hybrid model produced only minor changes.
5. System Size Comparison: When compared to peripheral Pb+Pb collisions (60–80% centrality) matched by event activity, the $x_J$ modifications in central O+O and Ne+Ne collisions are found to be comparable.

Significance and Claims
The paper claims that these results constitute an observation of centrality-dependent dijet transverse momentum imbalance in O+O and Ne+Ne collisions. The findings demonstrate that medium-induced partonic energy loss persists in collision systems considerably smaller than Pb+Pb and Xe+Xe.

The significance of this work lies in:

• Establishing a New Regime: It extends the study of jet quenching to the smallest quark–gluon plasma droplets accessible at the LHC.
• Path-Length Dependence: The results provide new constraints on the onset and path-length dependence of jet quenching, utilizing a system with a more symmetric geometry to reduce geometric uncertainties.
• Robustness: By using self-normalized distributions, the measurement explicitly constrains centrality-bias effects and event-activity selection biases that complicate interpretations in small systems.

The authors conclude that the observed modifications are qualitatively similar to those in nucleus-nucleus collisions but occur in a regime where initial-state configuration effects are better controlled, thereby strengthening the case for the formation of a QGP-like medium in light-ion collisions.