Kappa distributions in the language of superstatistics

Sergio Davis, Biswajit Bora, Cristian Pavez, and Leopoldo Soto · 2025

Read Kappa distributions in the language of superstatistics by Sergio Davis, Biswajit Bora, Cristian Pavez, and Leopoldo Soto (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-03757.txt.

Introduction

In the solar wind and Earth's magnetosphere, particles rarely settle into thermal balance, and their speeds follow a fat-tailed pattern called the kappa distribution. Sergio Davis and co-authors review in 2025 how superstatistics, a framework that layers one distribution over another, explains that pattern without exotic statistics. They rebuild both the single-particle and many-particle versions inside this frame.

The approach also simplifies computing averages under kappa distributions, a daily chore in plasma work. The authors discuss what the framework implies for correlations between particles, and for temperature and entropy in such plasmas, giving space physicists a more standard toolbox for a stubborn distribution.

Catalog record

CREATORSergio Davis, Biswajit Bora, Cristian Pavez, and Leopoldo Soto
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE31 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2507-03757

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Kappa distributions in the language of superstatistics Sergio Davis, Biswajit Bora, Cristian Pavez, and Leopoldo Soto Abstract The kappa distribution of velocities appears routinely in the study of collisionless plasmas present in Earth’s magnetosphere, the solar wind among other contexts where particles are unable to reach thermal equilibrium. Originally justified through the use of Tsallis’ non-extensive statistics, nowadays there are alternative frameworks that provide insight into these distributions, such as superstatistics. In this work we review the derivation of the multi-particle and single-particle kappa distributions for collisionless plasmas within the framework of superstatistics, as an alternative to the use of non-extensive statistics. We also show the utility of the superstatistical framework in the computation of expectation values under kappa distributions. Some consequences of the superstatistical formalism regarding correlations, temperature and entropy of kappa-distributed plasmas are also discussed. 1 Introduction The Maxwell-Boltzmann distribution is the probability density of particle velocities observed in systems that reach thermal equilibrium. For a particle with mass m at temperature T , the Maxwell-Boltzmann distribution is the three-dimensional Gaussian P(\bm{v}|\beta,m)=\left(\frac{m\beta}{2\pi}\right)^{\frac{3}{2}}\exp\left(-\frac{\beta m\bm{v}^{2}}{2}\right), (1) with \beta\mathrel{\mathop{:}}=1/(k_{B}T) the inverse temperature. On the other hand, collisionless plasmas such as the ones present in the Earth’s magnetosphere [ 1 , 2 , 3 , 4 ] , the solar wind [ 5 , 6 , 7 ] and the interstellar medium [ 8 , 9 ] are unable to reach thermal equilibrium, and in that case what is commonly observed is that particle velocities are best described by the kappa distribution [ 10 , 11 ] , a generalization of ( 1 ) given by P(\bm{v}|\kappa,v_{\text{th}})=\left[\pi\left(\kappa-\frac{3}{2}\right)v_{\text{th}}^{2}\right]^{-\frac{3}{2}}\frac{\Gamma(\kappa+1)}{\Gamma(\kappa-\frac{1}{2})}\left[1+\frac{1}{\kappa-\frac{3}{2}}\frac{\bm{v}^{2}}{v_{\text{th}}^{2}}\right]^{-(\kappa+1)}, (2) where \kappa is referred to as the spectral index and v_{\text{th}} is the thermal velocity . Outside space plasmas, kappa distributions have only recently been proposed [ 12 , 13 ] as plausible models for velocity distributions in laboratory plasmas such as Z-pinch discharges, even when power law distributions in such configurations have been reported previously [ 14 , 15 , 16 , 17 , 18 ] . It is particularly interesting to explore the existence of kappa distributions in plasma focus discharges [ 19 , 20 ] , where phenomena such as jets and filaments similar to the ones observed in astrophysical environments, have been observed [ 21 , 22 , 23 ] . In the limit \kappa\rightarrow\infty , the kappa distribution in ( 2 ) reduces to a Maxwellian as in ( 1 ) with temperature T given by k_{B}T=\frac{mv_{\text{th}}^{2}}{2}. (3) The thermodynamic foundations leading to the Maxwell-Boltzmann distribution in ( 1 ) are well understood. Let us consider a system of N classical particles with microstate \bm{\Gamma}\mathrel{\mathop{:}}=(\bm{P},\bm{R}) , where \bm{P}\mathrel{\mathop{:}}=(\bm{p}_{1},\ldots,\bm{p}_{N}) are the kinetic degrees of freedom and \bm{R}\mathrel{\mathop{:}}=(\bm{r}_{1},\ldots,\bm{r}_{N}) the configurational degrees of freedom, and such that the Hamiltonian of the system is \mathcal{H}(\bm{P},\bm{R})=\sum_{i=1}^{N}\frac{\bm{p}_{i}^{2}}{2m_{i}}+\Phi(\bm{R}), (4) with \Phi the interaction (potential) energy. The distribution of microstates p(\bm{\Gamma}) that maximizes the Boltzmann-Gibbs entropy \mathcal{S}[p]\mathrel{\mathop{:}}=-k_{B}\int d\bm{\Gamma}p(\bm{\Gamma})\ln p(\bm{\Gamma}) (5) subject to the constraint of fixed mean energy \big<\mathcal{H}\big>_{p}=\int d\bm{\Gamma}p(\bm{\Gamma})\mathcal{H}(\bm{\Gamma})=\overline{E} (6) is known as the canonical ensemble, the equilibrium distribution given by P(\bm{P},\bm{R}|\beta)=\frac{\exp\big(-\beta\mathcal{H}(\bm{P},\bm{R})\big)}{Z(\beta)}. (7) The equilibrium distribution of momenta P(\bm{P}|\beta) is then obtained by integration of the configurational degrees of freedom \bm{R} in ( 7 ), that is, P(\bm{P}|\beta)=\int\frac{d\bm{R}}{Z(\beta)}\,\exp\left(-\beta\left[\sum_{i=1}^{N}\frac{\bm{p}_{i}^{2}}{2m_{i}}+\Phi(\bm{R})\right]\right) (8) which reduces to P(\bm{P}|\beta)=\left(\frac{\beta}{2\pi\tilde{m}}\right)^{\frac{3N}{2}}\;\exp\left(-\beta\sum_{i=1}^{N}\frac{\bm{p}_{i}^{2}}{2m_{i}}\right), (9) with \tilde{m} the geometric mean mass, \tilde{m}\mathrel{\mathop{:}}=\left(\prod_{i=1}^{N}m_{i}\right)^{\frac{1}{N}}=m_{A}^{N_{A}/N}\cdot m_{B}^{N_{B}/N}, (10) where in the last equality we have assumed, for simplicity, a two-species plasma with species A and B , having masses m_{A} and m_{B} , and number of particles N_{A} and N_{B} respectively, such that N=N_{A}+N_{B} . The distribution in ( 9 ) is universal, in the sense that it does not depend on the interaction potential \Phi . Moreover, it clearly factors as the product of N independent, single-particle momentum distributions, P(\bm{P}|\beta)=\prod_{i=1}^{N}P(\bm{p}_{i}|\beta), (11) where P(\bm{p}_{i}|\beta)=\left(\frac{\beta}{2\pi m_{i}}\right)^{\frac{3}{2}}\exp\left(-\frac{\beta\bm{p}_{i}^{2}}{2m_{i}}\right). (12) The single-particle momentum distribution P(\bm{p}|\beta) is equivalent to ( 1 ) if we recall that \bm{p}=m\bm{v} and use the scaling property P(\bm{v}|\beta)=m^{3}\,P(\bm{p}=m\bm{v}|\beta). (13) Taking this procedure as a basis, an early theoretical attempt [ 24 ] to explain the origin of the kappa distribution lead to Tsallis’ non-extensive statistics [ 25 ] , which postulates that the functional that must be maximized under appropriate constraints is not ( 5 ) but the more general q -entropy, defined by S_{q}[p]\mathrel{\mathop{:}}=\frac{k_{B}}{q-1}\left(1-\int d\bm{\Gamma}\,p(\bm{\Gamma})^{q}\right), (14) where q is known as the entropic index , and such that S_{q} reduces to the Boltzmann-Gibbs entropy in ( 5 ) in the limit q\rightarrow 1 . However, Tsallis statistics has not been free of criticisms, due to the fact that its application can lead to inconsistencies [ 26 , 27 , 28 , 29 ] . Moreover, nowadays there exist alternative formalisms that can explain the existence of kappa distributions without the need to postulate a different definition of entropy. Among them, one of the most promising is the theory of superstatistics [ 30 , 31 ] , where the inverse temperature \beta is promoted from a constant to a random variable having its own probability density. This mechanism cleanly produces non-equilibrium ensembles, including the power laws such as ( 2 ) commonly associated with non-extensive statistics, using instead mixtures of canonical ensembles. In this work we will review the derivation of the kappa distribution from the point of view of superstatistics, using an invariant parameterization of the distribution of inverse temperatures, as well as derive new results regarding correlations, entropy and temperature of kappa-distributed plasmas. 2 Superstatistics The formalism of superstatistics [ 30 , 31 ] is an alternative to non-extensive statistics that effectively recovers non-canonical ensembles in an elegant manner. It has been successfully applied to plasma physics [ 32 , 33 , 34 , 35 , 36 ] , condensed matter systems [ 37 , 38 , 39 ] , high-energy physics and gravitation [ 40 , 41 , 42 ] , among other fields [ 43 , 44 ] . In superstatistics the inverse temperature \beta is not a constant but a random variable, having a probability density P(\beta|\bm{\lambda}) where \bm{\lambda} represents the parameters defining a particular ensemble. The canonical ensemble in ( 7 ) is then replaced by a joint distribution P(\bm{\Gamma},\beta|\bm{\lambda}) of the microstates \bm{\Gamma} and the inverse temperature \beta , and this joint distribution is factorized as P(\bm{\Gamma},\beta|\bm{\lambda})=P(\beta|\bm{\lambda})P(\bm{\Gamma}|\beta,\bm{\lambda}) (15) using the product rule of probability theory. Because \bm{\lambda} represents knowledge about \beta , we see that the state of knowledge (\beta,\bm{\lambda}) can be replaced by \beta itself and thus we can replace P(\bm{\Gamma}|\beta,\bm{\lambda}) by P(\bm{\Gamma}|\beta) as given by ( 7 ). Therefore, we can write ( 15 ) as P(\bm{\Gamma},\beta|\bm{\lambda})=P(\beta|\bm{\lambda})\frac{\exp\big(-\beta\mathcal{H}(\bm{\Gamma})\big)}{Z(\beta)}, (16) and the superstatistical distribution of microstates P(\bm{\Gamma}|\bm{\lambda}) as the integral over \beta of ( 16 ), P(\bm{\Gamma}|\bm{\lambda})=\int_{0}^{\infty}d\beta\,P(\beta|\bm{\lambda})\frac{\exp\big(-\beta\mathcal{H}(\bm{\Gamma})\big)}{Z(\beta)}. (17) Because \beta and \Gamma are correlated in ( 16 ), the marginal distribution P(\bm{\Gamma}|\bm{\lambda}) is, in general, different from the canonical ensemble. This provides us with a mechanism to obtain non-canonical ensembles by postulating different inverse temperature distributions without the need to replace the Boltzmann-Gibbs entropy. The canonical ensemble is a particular case of superstatistics where \beta becomes again a constant, that is, when the uncertainty over \beta vanishes. This is of course achieved for P(\beta|\beta_{0})=\delta(\beta-\beta_{0}), (18) as in this case P(\bm{\Gamma}|\bm{\lambda}) in ( 17 ) reduces to ( 7 ) with \beta=\beta_{0} . We will

Read the complete text