Tsallis–Pareto Distribution

• Tsallis–Pareto distribution is a generalized statistical model based on Tsallis entropy, incorporating a temperature parameter and an entropic index to describe non-Gaussian, power-law behavior in diverse systems.
• It is derived by maximizing Tsallis entropy under constraints, offering thermodynamically consistent fits to particle spectra in high-energy collisions and capturing long-range correlations and intrinsic fluctuations.
• The framework enables the extraction of a 'Tsallis thermometer', linking non-extensive temperatures to conventional measurements through calibration curves that reconcile traditional and non-extensive thermodynamics.

The Tsallis–Pareto distribution constitutes a central component of non-extensive thermostatistics and provides a robust phenomenological and theoretical framework for describing a wide range of physical systems exhibiting power-law tails, long-range correlations, or intrinsic fluctuations. It is especially prominent in the statistical mechanics of high-energy hadronic and nuclear collisions, non-equilibrium plasma physics, and complex systems with emergent non-Gaussian stationary distributions. In contexts where the conventional Boltzmann–Gibbs formalism fails to account for observed spectra or correlations, the Tsallis–Pareto form enables an internally consistent extension with direct thermodynamic interpretation, including a temperature parameter that can act as a “Tsallis thermometer.”

1. Definition and Mathematical Structure

The Tsallis–Pareto distribution is a generalization of the Boltzmann (exponential) or Maxwell–Boltzmann–Gibbs distributions, derived from maximizing the Tsallis entropy under appropriate constraints. For a single-particle observable (e.g., energy $E$ or transverse mass $m_T$), the canonical Tsallis–Pareto distribution takes the form

$$f(E) = \left[1 + (q-1)\frac{E-\mu}{T}\right]^{-1/(q-1)}, \quad (q>1)$$

where:

• $T$ is the Tsallis temperature (Lagrange multiplier for mean energy),
• $q$ is the entropic index (non-extensivity parameter),
• $\mu$ is the chemical potential.

In particle spectra, and particularly in high-energy collisions, the invariant yield at mid-rapidity ($y\approx 0$) is typically parameterized as

$$\frac{1}{2\pi p_T} \frac{d^{2}N}{dp_T\,dy} = A\,m_T\,\left[1 + \frac{q-1}{T}(m_T - m)\right]^{-\frac{q}{q-1}},$$

with $m_T = \sqrt{p_T^2 + m^2}$. This reduces to the standard exponential in the $q \to 1$ limit. The exponent $m_T$0 is required for thermodynamic consistency, ensuring that all thermodynamic relations (such as Maxwell relations) are satisfied (Cleymans et al., 2015, Azmi et al., 2015).

2. Thermodynamic Consistency and Physical Interpretation

A defining feature of the Tsallis–Pareto framework is thermodynamic consistency, mainly realized when the correct exponent is used in the distribution. For the canonical ensemble with Tsallis entropy,

$m_T$1

the maximization procedure leads to the q-exponential forms above, accompanied by the thermodynamic definitions:

• $m_T$2
• $m_T$3
• $m_T$4

This construction preserves the equations of state, entropy, and heat capacity as generalized functions of $m_T$5, $m_T$6, and $m_T$7, with the physical temperature identified via the generalized Euler or Maxwell relations (Bhattacharyya et al., 2016, Cleymans, 2012).

Thermodynamically consistent Tsallis–Pareto distributions allow $m_T$8 to act as a “Tsallis thermometer,” enabling direct extraction of a non-extensive temperature from observables and yielding consistent physical predictions for energy, entropy, and fluctuations (Ishihara, 2021, Cleymans et al., 2015).

3. Operational Construction and Calibration of the Tsallis Thermometer

The practical extraction of Tsallis–Pareto parameters proceeds via:

• Measuring single-particle spectra (e.g., $m_T$9-distributions in collider data).
• Fitting the spectra with the Tsallis–Pareto form, with free parameters $$f(E) = \left[1 + (q-1)\frac{E-\mu}{T}\right]^{-1/(q-1)}, \quad (q>1)$$0, $$f(E) = \left[1 + (q-1)\frac{E-\mu}{T}\right]^{-1/(q-1)}, \quad (q>1)$$1, $$f(E) = \left[1 + (q-1)\frac{E-\mu}{T}\right]^{-1/(q-1)}, \quad (q>1)$$2 (and possibly volume and chemical potential).
• Using maximum likelihood or least-squares fitting methods to determine the best-fit values.

Once $$f(E) = \left[1 + (q-1)\frac{E-\mu}{T}\right]^{-1/(q-1)}, \quad (q>1)$$3 is extracted, it plays the canonical role of temperature, defining the mean transverse or kinetic energy per degree of freedom, with $$f(E) = \left[1 + (q-1)\frac{E-\mu}{T}\right]^{-1/(q-1)}, \quad (q>1)$$4 quantifying the degree of deviation from extensivity, encoding information about event-by-event fluctuations, correlations, or system memory effects (Marques et al., 2015, Cleymans et al., 2015, Ishihara, 21 Jul 2025).

In oscillator or free-particle models, calibration is achieved by coupling a Tsallis-thermometer (e.g., a set of $$f(E) = \left[1 + (q-1)\frac{E-\mu}{T}\right]^{-1/(q-1)}, \quad (q>1)$$5 oscillators) to the system, measuring their energy distributions, and fitting to extract the physical $$f(E) = \left[1 + (q-1)\frac{E-\mu}{T}\right]^{-1/(q-1)}, \quad (q>1)$$6, using closed-form expressions that relate $$f(E) = \left[1 + (q-1)\frac{E-\mu}{T}\right]^{-1/(q-1)}, \quad (q>1)$$7, $$f(E) = \left[1 + (q-1)\frac{E-\mu}{T}\right]^{-1/(q-1)}, \quad (q>1)$$8, and $$f(E) = \left[1 + (q-1)\frac{E-\mu}{T}\right]^{-1/(q-1)}, \quad (q>1)$$9 to $T$0 and $T$1 (Ishihara, 2021, Ishihara, 3 Aug 2025).

4. Phenomenological Applications and Empirical Trends

The Tsallis–Pareto distribution is now the standard tool for describing both soft (thermal) and hard (power-law) components of particle spectra in relativistic nuclear and hadronic collisions. Key empirical findings include:

• A universal Tsallis temperature, e.g., $T$2–$T$3 MeV in high-energy $T$4 and $T$5–$T$6 collisions at the LHC, independent of collision energy, system size, or hadron species, suggesting the existence of an intrinsic non-extensive temperature scale (Marques et al., 2015, Cleymans, 2012, Azmi et al., 2015).
• Entropic index values of $T$7–$T$8, again largely system- and species-independent in central rapidity intervals (Azmi et al., 2013).
• Quantitative relations and scaling laws between $T$9, $q$0, and global observables such as multiplicity: $q$1 increases and $q$2 decreases toward 1 with increasing event multiplicity or system size, converging to an apparent QCD scale (the Tsallis-thermometer equilibrium point) for large systems or high multiplicity classes (Bíró et al., 2020).
• Direct linear proportionality between the Tsallis temperature and the mean transverse momentum, $q$3, offering a simplified operational definition for experiments (Gyulai et al., 7 Jan 2026).

In small systems or when using event-shape classifiers (multiplicity, spherocity, flattenicity), the Tsallis–Pareto parameters are sensitive to both soft and hard processes, with $q$4 reflecting the soft sector and $q$5 encoding deviations due to hard or jetty events (Gyulai et al., 7 Jan 2026).

5. Comparison with Standard Distributions and Calibration

When $q$6, the Tsallis–Pareto distribution reduces smoothly to the standard Maxwell–Boltzmann, Bose–Einstein, or Fermi–Dirac result, and the extracted $q$7 matches the kinetic or chemical freeze-out temperature. For $q$8, the effective Tsallis temperature is linearly related to the “standard” effective temperature (e.g., from BE/FD or Boltzmann fits), allowing construction of calibration curves or universal slope-intercept relations bridging the two frameworks (Zhang et al., 31 May 2025). This mapping enables direct comparison and translation of Tsallis-based measurements into the conventional temperature scale.

Systematic trends indicate that Tsallis $q$9 is always numerically less than BE/FD $\mu$0 for fixed data, with robust linear correlations observed across particle species, collision centralities, and system sizes, facilitating cross-experiment unification of temperature measurements.

6. Theoretical and Experimental Considerations

Although the Tsallis–Pareto distribution is phenomenologically successful and internally consistent, key theoretical questions remain regarding the microscopic/statistical origin of $\mu$1 in QCD or other many-body quantum systems. Competing interpretations include superstatistics (event-by-event $\mu$2 fluctuations), fractal phase spaces, and non-ergodic dynamics.

Experimental limitations include:

• The necessary fitting range and minimal acceptance biases.
• The covariance between $\mu$3 and $\mu$4 in restricted data intervals.
• The role of resonance decays, radial/collective flow, and nonthermal sources in high-$\mu$5 tails.

The use of alternative definitions of entropy or expectation values (standard vs. escort averages), as well as the domain of validity for $\mu$6 (e.g., $\mu$7 for oscillators, $\mu$8 for certain free-particle ensembles), must be carefully accounted for in precise applications (Ishihara, 2021, Ishihara, 3 Aug 2025, Ishihara, 21 Jul 2025).

7. Summary Table: Core Tsallis–Pareto Distribution Relations and Calibrations

Quantity/Concept	Mathematical Formulation	Key Reference
Canonical Probability	$\mu$9	(Ishihara, 21 Jul 2025)
Invariant Yield (fit)	$y\approx 0$0	(Cleymans et al., 2015, Marques et al., 2015)
Physical Temperature	$y\approx 0$1	(Bhattacharyya et al., 2016)
Thermodynamic Consistency	$y\approx 0$2, $y\approx 0$3	(Cleymans, 2012, Cleymans et al., 2015)
Tsallis–thermometer relation	$y\approx 0$4 (linear $y\approx 0$5–$y\approx 0$6)	(Bíró et al., 2020)
Calibration with standard $y\approx 0$7	$y\approx 0$8	(Zhang et al., 31 May 2025)

The Tsallis–Pareto distribution provides a comprehensive and practical extension of statistical mechanics for systems with nontrivial correlations and fluctuations, offering both operational and theoretical means to characterize temperatures and non-extensivity in diverse physical contexts. The thermodynamically consistent implementation ensures that parameter extraction yields physically interpretable observables, with $y\approx 0$9 acting as a universal thermometer across systems from high-energy collisions to classical oscillator baths, subject to appropriate constraints and calibrations (Marques et al., 2015, Cleymans et al., 2015, Azmi et al., 2013, Bhattacharyya et al., 2016, Bíró et al., 2020, Zhang et al., 31 May 2025, Ishihara, 2021, Ishihara, 21 Jul 2025, Ishihara, 2021).