Heavy Quark Transport is Non-Gaussian Beyond Leading Log

arXiv:2604.21895 · hep-ph, hep-th, nucl-th · Submitted 2026-04-23 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: High Energy Physics: Phenomenology. The day in particle phenomenology: models, signatures and constraints.

Kai: We're looking at the paper: "Heavy Quark Transport is Non-Gaussian Beyond Leading Log".

Mira: Here's the gist of the paper. Heavy quark transport beyond leading logarithm at weak coupling is intrinsically non-Gaussian: the longitudinal momentum transfer distribution has asymmetric exponential tails that are crucial for equilibration dynamics.

Kai: This finding is significant because it suggests a robust feature of realistic Quark-Gluon Plasma rather than being peculiar to weak or strong coupling, conformality, or supersymmetry.

Mira: Heavy quarks provide a particularly valuable probe of the quark-gluon plasma (QGP) produced in ultra-relativistic heavy-ion collisions (HICs).

Kai: The Core Finding on Non-Gaussian Structure The paper establishes that heavy quark transport at weak coupling beyond leading logarithm exhibits manifest nonGaussian structure, characterized by a Gaussian core and asymmetric exponential tails in the longitudinal momentum transfer distribution.

Paper summary: Kai: This paper, "Heavy Quark Transport is Non-Gaussian Beyond Leading Log," investigates heavy quark transport in weakly coupled non-Abelian plasmas. The core thesis here is that heavy quark transport at weak coupling exhibits intrinsic non-Gaussian structure.

Mira: Specifically, they found that the longitudinal momentum transfer distribution has asymmetric exponential tails, and these tails are crucial for understanding the equilibration dynamics of the system.

Lev: The authors do this by computing the leading order momentum transfer kernel for relativistic heavy quarks in weakly coupled non-Abelian plasmas and matching it to hard thermal loop resummed soft physics.

Kai: They also pointed out that this specific structure is not unique to weak or strong coupling, nor is it peculiar to conformality or supersymmetry.

Mira: This finding is significant because it suggests that this non-Gaussian behavior might be a feature of the physical quark-gluon plasma we see in ultra-relativistic heavy-ion collisions.

Lev: The paper aims to determine two momentum diffusion coefficients and also recovers the known leading order diffusion coefficient at zero velocity, which helps recover appropriate Gaussian limits.

Kai: They satisfy the generalized equilibration condition from reference thirty which provides a nontrivial consistency check on their resulting non-Gaussian dynamics <ref:2604.21895#pg2,a nontrivial consistency check on>.

Mira: The methodology involves using Heaviside step functions to separate contributions in each regime of momentum transfer, which is reminiscent of the original methods by Braaten and Thoma.

Lev: They find that when they write the kernel S L v in terms of these components, they can omit the hard thermal loop pieces for momenta greater than T because those are just higher order contributions.

Kai: This leaves them with a result that is manifestly fixed-order except for the HTL resummed contribution in the soft region where momentum is less than T.

Mira: In that soft region, since the g to zero limit and the momentum integral don't commute, they have to perform a radial integral exactly before expanding in m D over T proportional to g.

Lev: This expansion is what allows them to correctly determine those logarithmically enhanced g to the fourth power log of one over g terms.

Conclusion: Kai: So, to wrap up the paper "Heavy Quark Transport is Non-Gaussian Beyond Leading Log," the authors showed that heavy quark transport has a non-Gaussian structure beyond leading order at weak coupling.

Mira: The main implication for us is that we can't rely on purely Gaussian descriptions when modeling heavy-flavor observables in these plasmas.

Lev: It means that any phenomenological models based only on Gaussian or Langevin truncations are not microscopically justified by default; they need to incorporate those non-Gaussian features.

Kai: They suggest a more realistic approach might involve describing the dynamics using eta D of v, kappa L of v, and that volatility exponent R of v.

Mira: And they found that this scaling behavior for R of v is robust, scaling as one over four gamma squared T across both weak and strong coupling expansions.

Lev: The fact that the leading contribution to the volatility exponent is independent of the coupling, with coupling dependence only appearing at subleading orders, tells us a lot about its fundamental nature.

Kai: Heavy quarks give us a particularly valuable probe because they are sensitive to the QGP produced in ultra-relativistic heavy-ion collisions.

Mira: So, this work provides a concrete structural insight into the transport properties that we might find when studying the QGP from those experiments.

Massachusetts Institute of Technology · University of California Santa Barbara

hep-ph, hep-th, nucl-th

Submitted: 2026-04-23

Updated: 2026-11-10

Comments: 17 pages, 2 figures, v2 matches published version

Journal ref: Phys. Rev. Lett. 137, 192302 (2026)

DOI: 10.1103/7fd5-1kq5

License: http://creativecommons.org/licenses/by/4.0/

Importance score: 97/100

The gist: Heavy quark transport beyond leading logarithm at weak coupling is intrinsically non-Gaussian: the longitudinal momentum transfer distribution has asymmetric exponential tails that are crucial for…

Key concepts

Non-Gaussian Structure
The distribution of longitudinal momentum transfer for heavy quarks is not purely Gaussian. It has a central Gaussian part and distinct asymmetric exponential tails. This asymmetry is crucial for understanding how heavy quarks equilibrate within the Quark-Gluon Plasma.
Longitudinal Momentum Transfer Distribution
This describes how the momentum of a heavy quark changes during its transport. The paper shows this distribution has a Gaussian center (from the central limit theorem) and exponential tails, which are key to describing equilibration dynamics in HICs.
Volatility Exponent R(v)
This exponent describes the scaling behavior of transport coefficients, like the diffusion coefficient at zero velocity. In ultra-relativistic limits, it scales as R(v) ∼ 1/(4γ2T), showing a coupling-independent robust feature across both weak and strong coupling expansions.

Terminology

Summary

Heavy quark transport beyond leading logarithm at weak coupling is intrinsically non-Gaussian: the longitudinal momentum transfer distribution has asymmetric exponential tails that are crucial for equilibration dynamics. p. 0

Where the paper says it

MIT-CTP/6020 Heavy Quark Transport is Non-Gaussian Beyond Leading Log Jean F. Du Plessis1, 2, ∗ and Bruno Scheihing-Hitschfeld2, † 1Center for Theoretical Physics — A Leinweber Institute, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, USA 2Kavli Institute for Theoretical Physics, University of California, Santa Barbara, California 93106, USA (Dated: April 24, 2026) We find that heavy quark transport beyond leading logarithm at weak coupling is intrinsically non-Gaussian: the longitudinal momentum transfer distribution has asymmetric exponential tails that are crucial for equilibration dynamics.

This finding is significant because it suggests a robust feature of realistic Quark-Gluon Plasma rather than being peculiar to weak or strong coupling, conformality, or supersymmetry. p. 4
Where the paper says it

1, together with the same qualitative weak-coupling structure in QCD, suggests that Gaussian core and exponential tail momentum transfer is not peculiar to weak or strong coupling, conformality, or supersymmetry, but rather a robust feature that realistic QGP should also exhibit, with its independent Gaussian characteristics ηD(v), κL(v), and the volatility exponent R(v) as the primary properties of the dynamics. We have shown that the longitudinal momentum transfer distribution of a relativistic heavy quark in generic weakly coupled gauge theory plasmas has a Gaussian core and exponential tails, and that, in particular, weakly coupled QCD exhibits this behavior.

Heavy quarks provide a particularly valuable probe of the quark-gluon plasma (QGP) produced in ultra-relativistic heavy-ion collisions (HICs). p. 0
Where the paper says it

We therefore expect that this is a robust feature that physical quark-gluon plasma should also exhibit. Heavy quarks (HQs) provide a particularly valuable probe of the quark-gluon plasma (QGP) produced in ultra-relativistic heavy-ion collisions (HICs).

The Core Finding on Non-Gaussian Structure

The paper establishes that heavy quark transport at weak coupling beyond leading logarithm exhibits manifest nonGaussian structure, characterized by a Gaussian core and asymmetric exponential tails in the longitudinal momentum transfer distribution. The Gaussian core follows as a natural consequence of the central limit theorem. p. 0

Where the paper says it

The tails are essential for equilibration in light of the violation of the Einstein relation [30], meaning that HQ transport is intrinsically non-Gaussian. The Gaussian core follows as a natural consequence of the central limit theorem.

The exponential tails are a new and surprising result, which is also present in the strongly coupled calculation of [31]. p. 0
Where the paper says it

On the other hand, the exponential tails are a new and surprising result. In fact, this structure is also present in the strongly coupled calculation of [31] — a fact that has remained underappreciated so far.

This indicates that Gaussian core and exponential tail momentum transfer is not peculiar to weak or strong coupling, conformality, or supersymmetry. p. 0
Where the paper says it

In fact, this structure is also present in the strongly coupled calculation of [31] — a fact that has remained underappreciated so far. With weakly coupled QCD and both weakly and strongly coupled N = 4 SYM exhibiting this same structure, this indicates that Gaussian core and exponential tail momentum transfer is not peculiar to weak or strong coupling, conformality, or supersymmetry, but is instead likely to be a robust feature of realistic QGP.

Methodology for Kernel Determination

The authors determine the longitudinal momentum transfer distribution by matching the ultraviolet perturbative contribution to the infrared HTL sector. They isolate the strict fixed-order kernel and determine the distribution at all momenta via a Legendre transform. p. 0

Where the paper says it

With weakly coupled QCD and both weakly and strongly coupled N = 4 SYM exhibiting this same structure, this indicates that Gaussian core and exponential tail momentum transfer is not peculiar to weak or strong coupling, conformality, or supersymmetry, but is instead likely to be a robust feature of realistic QGP. To establish this result, we consistently match the ultraviolet perturbative contribution to the infrared HTL sector, isolate the strict fixed-order kernel, and determine the longitudinal momentum transfer distribution at all momenta via a Legendre transform.

The full matched kernel reproduces known leading-log results for drag and momentum diffusion coefficients, as well as the known leading-order diffusion coefficient at v = 0. p. 0
Where the paper says it

To establish this result, we consistently match the ultraviolet perturbative contribution to the infrared HTL sector, isolate the strict fixed-order kernel, and determine the longitudinal momentum transfer distribution at all momenta via a Legendre transform. The full matched kernel reproduces the known leading-log results for the drag and arXiv:2604.

This process also satisfies the generalized equilibration condition of Ref. p. 1
Where the paper says it

2 momentum diffusion coefficients, as well as the known leading-order diffusion coefficient at v = 0, thereby recovering the appropriate Gaussian limits. It also satisfies the generalized equilibration condition of Ref.

[30].

Verification of Consistency and Robustness

The calculation involves isolating the strict fixed-order kernel by using Heaviside step functions to separate contributions in each regime. The structure of the integrand in Eq. p. 1

Where the paper says it

(6) then provides a nontrivial consistency check on that construction. Because this symmetry is a structural property of the integrand in Eq.

(5) provides a nontrivial consistency check on the construction, as it satisfies the generalized equilibration condition order by order in perturbation theory. p. 1
Where the paper says it

It also satisfies the generalized equilibration condition of Ref. [30], providing a nontrivial consistency check on the resulting non-Gaussian dynamics.

Furthermore, the longitudinal part of the HTL-resummed contributions in Eq. p. 2
Where the paper says it

This can be seen by examining the exponential suppression of IHard(p, v) at large p by statistical factors (see End Matter). We have explicitly verified that replacing the separa3 It is also reassuring to note that the longitudinal part of the HTLresummed contributions (in the first two lines) in the kernel (10) satisfies the Einstein relation.

(10) satisfies the Einstein relation, which is mandated by the fact that in the extreme weakly coupled limit equilibration should still be attained. p. 2
Where the paper says it

We have explicitly verified that replacing the separa3 It is also reassuring to note that the longitudinal part of the HTLresummed contributions (in the first two lines) in the kernel (10) satisfies the Einstein relation. This is mandated by the fact that in the extreme weakly coupled limit equilibration should still be attained — and the fact that this part of the kernel is quadratic means that the associated stochastic process is Gaussian.

The second moment of the kernel at v = 0 reproduces the known diffusion coefficient exactly at order g4. p. 3
Where the paper says it

4 tion scale T by any Λ ∼ T leaves the fixed-order kernel SFO invariant. As an additional cross-check, we have verified that the second moment of the kernel at v = 0 reproduces the known diffusion coefficient [1, 20] exactly at order g 4 .

Implications for Phenomenological Descriptions

The non-Gaussian momentum transfer dynamics can modify heavy-flavor observables and the transport coefficients extracted from them. A purely Gaussian description is not consistent with first-principles calculations for generic relativistic heavyquark transport. p. 5

Where the paper says it

At finite velocity, however, the remaining fixed-order contributions control the shape of the distribution and invalidate a purely Gaussian truncation, as the universal equilibration condition [30] is only satisfied through a detailed balance of all of the cumulants. A Gaussian description is therefore not consistent with first-principles calculations for generic relativistic heavyquark transport.

This result suggests that phenomenological descriptions based purely on Gaussian or Langevin truncations are not microscopically justified by default. p. 5
Where the paper says it

These results have direct implications for heavy-flavor phenomenology. Since non-Gaussian momentum transfer can modify heavy-flavor observables and the transport coefficients extracted from them [27–29], phenomenological descriptions based purely on Gaussian or Langevin truncations are not microscopically justified by default.

A more realistic, non-Gaussian description may be organized in terms of ηD(v), κL(v), and a single additional parameter R(v), the volatility exponent. p. 5
Where the paper says it

Our calculation therefore motivates revisiting heavyquark transport in the QGP with non-Gaussian kernels if one aims to extract transport coefficients that can be meaningfully compared to lattice determinations. A more realistic, non-Gaussian description may be organized in terms of ηD(v), κL(v), and a single additional parameter R(v), the volatility exponent, characterizing the distribution of (rare) large momentum kicks, which sets the slopes of the asymmetric exponential tails.

Scaling Behavior of Transport Coefficients

The analysis reveals scaling behaviors for the transport coefficients, such as the diffusion coefficient at v = 0. The diffusion coefficient is found to be proportional to g2CF m2DT 6π ln(TmD) + TF Nf ln 2Nc + TF Nf + ξ. p. 8

Where the paper says it

9 The poles cancel in the sum, as they must. We obtain κ = g2CF m2DT 6π " ln 2 TmD + TF Nf ln 2 Nc + TF Nf + ξ # (34) with ξ = 12 − γ E + ζ ′ (2) ζ(2) , in agreement with the known leading-order heavy quark diffusion coefficient [ 1 , 20].

In the ultra-relativistic limit at weak coupling, the volatility exponent is found to scale as R(v) ∼ 1/(4γ2T). p. 8
Where the paper says it

The conclusion of this analysis is that as γ → ∞ one has R(v) ∼ 1/(4γ2T). (36) R(v) in the ultra-relativistic limit at strong coupling.

This scaling behavior is robust across both weak and strong coupling expansions. p. 4
Where the paper says it

1, together with the same qualitative weak-coupling structure in QCD, suggests that Gaussian core and exponential tail momentum transfer is not peculiar to weak or strong coupling, conformality, or supersymmetry, but rather a robust feature that realistic QGP should also exhibit, with its independent Gaussian characteristics ηD(v), κL(v), and the volatility exponent R(v) as the primary properties of the dynamics. We have shown that the longitudinal momentum transfer distribution of a relativistic heavy quark in generic weakly coupled gauge theory plasmas has a Gaussian core and exponential tails, and that, in particular, weakly coupled QCD exhibits this behavior.

The leading contribution to the volatility exponent R(v) is independent of the coupling — coupling dependence only appears at subleading order in either expansion. p. 3
Where the paper says it

This defines a volatility exponent R(v), which encodes the likelihood of experiencing (rare) large momentum kicks. Note that, both in the weak and strong coupling expansions, the leading contribution to the volatility exponent R(v) is independent of the coupling — coupling dependence only appears at subleading order in either expansion.

Higher Moment Analysis at Zero Velocity

The higher projected longitudinal moments at zero velocity are derived by focusing on the hard contribution for n ≥ 4. The even moments Mn(0) are given by a closed-form result involving integrals of sinh and cosh functions. p. 7

Where the paper says it

Because the cumulants are given by derivatives of K as Mn = −∂ nK/∂xn|x=0, and the only scale in the problem is T, all that is needed is to evaluate the integrals J˜(sinh) n,m ≡ Z ∞ 0 dpZ ∞ 1 dk p nk m sinh2 pk 4 = 2 n+3n! ζ(n) (n − m) , (30) J˜(cosh) n,m ≡ Z ∞ 0 dpZ ∞ 1 dk p nk m cosh2 pk 4 = (1 − 2 1−n )J˜(sinh) n,m .

The formula for Mn(0) is valid for all even n ≥ 4, while Mn(0) = 0 for odd n. p. 7
Where the paper says it

(31) Using these to evaluate the cumulants yields Mn(0) = 3 g 2CF m2 D 2 n−1T n−2 (n − 1)! ζ(n) π 3(n + 1)Nc + NcNX 2 + TF Nf " Nc NXn(n + 2) + 2(n 2 − 2n + 8) 8(n − 2)(n + 2) + (1 − 2 1−n) TF Nf n − 2 # (32) for all even n ≥ 4, while Mn(0) = 0 for odd n.

This confirms the vanishing of odd moments due to parity. p. 7
Where the paper says it

(31) Using these to evaluate the cumulants yields Mn(0) = 3 g 2CF m2 D 2 n−1T n−2 (n − 1)! ζ(n) π 3(n + 1)Nc + NcNX 2 + TF Nf " Nc NXn(n + 2) + 2(n 2 − 2n + 8) 8(n − 2)(n + 2) + (1 − 2 1−n) TF Nf n − 2 # (32) for all even n ≥ 4, while Mn(0) = 0 for odd n.

The pole in the formula signals the special infrared sensitivity of the diffusion coefficient. p. 7
Where the paper says it

ζ(n) π 3(n + 1)Nc + NcNX 2 + TF Nf " Nc NXn(n + 2) + 2(n 2 − 2n + 8) 8(n − 2)(n + 2) + (1 − 2 1−n) TF Nf n − 2 # (32) for all even n ≥ 4, while Mn(0) = 0 for odd n. The explicit (n − 2)−1 pole signals the special infrared sensitivity of the diffusion coefficient, and is precisely why the n = 2 case must be treated separately.

Matching-Scale Invariance Proof

The matching-scale invariance is verified by differentiating Eq. (10) with respect to the matching scale Λ and showing that the derivative is zero. p. 0

Where the paper says it

We show this by computing the leading-order momentum transfer kernel for relativistic heavy quarks in weakly coupled non-Abelian plasmas, matching perturbative momentum transfer on the thermal scale to hard-thermal-loop-resummed soft physics. This is the same structure previously found in strongly coupled holographic plasmas, showing that it is not peculiar to weak or strong coupling, conformality, or supersymmetry.

This proof confirms that the structure of the fixed-order kernel remains invariant under a change in separation scale. p. 3
Where the paper says it

4 tion scale T by any Λ ∼ T leaves the fixed-order kernel SFO invariant. As an additional cross-check, we have verified that the second moment of the kernel at v = 0 reproduces the known diffusion coefficient [1, 20] exactly at order g 4 .

This invariance is a structural property of the integrand in Eq. p. 1
Where the paper says it

(6) then provides a nontrivial consistency check on that construction. Because this symmetry is a structural property of the integrand in Eq.

(5), independent of the specific forms of IHard, ISoft, and IMatch. p. 1
Where the paper says it

Because this symmetry is a structural property of the integrand in Eq. (5), independent of the specific forms of IHard, ISoft, and IMatch, the equilibration condition holds order by order in perturbation theory.

The result is shown to be zero, proving matching-scale independence. p. 0
Where the paper says it

Du Plessis1, 2, ∗ and Bruno Scheihing-Hitschfeld2, † 1Center for Theoretical Physics — A Leinweber Institute, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, USA 2Kavli Institute for Theoretical Physics, University of California, Santa Barbara, California 93106, USA (Dated: April 24, 2026) We find that heavy quark transport beyond leading logarithm at weak coupling is intrinsically non-Gaussian: the longitudinal momentum transfer distribution has asymmetric exponential tails that are crucial for equilibration dynamics. We show this by computing the leading-order momentum transfer kernel for relativistic heavy quarks in weakly coupled non-Abelian plasmas, matching perturbative momentum transfer on the thermal scale to hard-thermal-loop-resummed soft physics.

The final result shows the diffusion coefficient is independent of the arbitrary matching scale choice. p. 7
Where the paper says it

(28) Invariance under change in separation scale. Choosing an arbitrary matching scale Λ ∼ g 0T leads to Eq.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Non-Gaussian Transport Modeling for Heavy Quarks: The improved AI system can perform first-principles calculations of heavy quark transport in weakly coupled gauge theory plasmas by calculating a fixed-order momentum transfer kernel that exhibits a Gaussian core and asymmetric exponential tails, which is crucial for equilibration dynamics.

  2. Improved Phenomenological Description: The system can move beyond purely Gaussian or Langevin truncations to utilize a non-Gaussian description organized in terms of ηD(v), κL(v), and the volatility exponent R(v) to characterize rare large momentum kicks, providing a more microscopically justified framework for extracting transport coefficients.

  3. Velocity and Coupling Robustness Testing: The system can test the robustness of this non-Gaussian structure by comparing results across different coupling strengths (weak vs strong) and different theories (QCD vs N=4 SYM), confirming that Gaussian core and exponential tail momentum transfer is not peculiar to weak or strong coupling, conformality, or supersymmetry.

  4. Quantitative Moment Prediction: The AI can accurately predict the higher projected longitudinal moments at zero velocity for even n greater than two using closed-form results derived from the kernel analysis, providing a precise formula for these moments.

  5. Diffusion Coefficient Extraction: The system can determine the heavy quark momentum diffusion coefficient κ by combining the log-enhanced contribution and the remaining fixed-order contributions, yielding an expression that recovers known leading-order coefficients at zero velocity.

Abstract

We find that heavy quark transport beyond leading logarithm at weak coupling is intrinsically non-Gaussian: the longitudinal momentum transfer distribution has asymmetric exponential tails that are crucial for equilibration dynamics. We show this by computing the leading-order momentum transfer kernel for relativistic heavy quarks in weakly coupled non-Abelian plasmas, matching perturbative momentum transfer on the thermal scale to hard-thermal-loop-resummed soft physics. This is the same structure previously found in strongly coupled holographic plasmas, showing that it is not peculiar to weak or strong coupling, conformality, or supersymmetry. We therefore expect that this is a robust feature that physical quark-gluon plasma should also exhibit.

Sources

Related papers