---
title: Rosenfeld–Tarazona Scaling in Liquids & Plasmas
url: https://www.emergentmind.com/topics/rosenfeld-tarazons-scaling
type: topic
---

# Rosenfeld–Tarazona Scaling in Liquids & Plasmas

Rosenfeld–Tarazona Scaling (RT Scaling) describes a set of power-law and exponential scaling relations between thermodynamic, structural, and transport properties of condensed-matter systems—primarily simple liquids and dense plasmas—based on excess thermodynamic quantities. It originated from the observation that, in strongly interacting fluids, the excess isochoric heat capacity and certain transport coefficients exhibit nearly universal, model-independent dependence on temperature and excess entropy, provided these are expressed in appropriately reduced units. The theory has been repeatedly confirmed and extended both by analytic theory and by high-precision numerical simulations, but also exhibits limitations dependent on the specific fluid, path in the phase diagram, and strength of structural correlations. The so-called "generalized RT scaling" augments the original framework to accommodate material- and density-dependent exponents.

## 1. Phenomenology and Historical Context

Rosenfeld–Tarazona scaling was introduced to rationalize simulation data for dense simple liquids, such as Lennard–Jones (LJ) and one-component plasma (OCP) systems, where the temperature dependence of excess thermodynamic quantities and transport coefficients appeared to follow regular scaling forms over broad thermodynamic ranges. Rosenfeld's original hypothesis postulated that reduced transport coefficients (e.g., diffusion, viscosity) collapse onto master curves when plotted against excess entropy per particle, $S_\text{ex}$, as $D^* = A\, \exp(B S_\text{ex})$, with weakly system-dependent prefactors. Tarazona and Rosenfeld subsequently provided perturbative arguments demonstrating that the thermal component of the excess energy—and its temperature derivative, the excess isochoric heat capacity—obey nontrivial but universal power laws in temperature for a large class of "strongly correlating" (Roskilde-simple) model fluids [1305.1776, 2601.01659].

## 2. RT Scaling for Excess Isochoric Heat Capacity

The canonical form of RT scaling for the excess isochoric heat capacity of simple liquids is
\[
C_V^{\rm ex}(T) \propto T^{-\alpha}
\]
where $\alpha$ is an exponent whose value characterizes the decay rate of thermal correlations as temperature increases above the melting line [2512.11520, 2505.15986]. The original theory predicted $\alpha = 2/5$, that is, $C_V^{\rm ex} \propto T^{-2/5}$ [1305.1776, 2601.01659]. For a system with $N$ particles, the total isochoric heat capacity is partitioned as
\[
C_V = \frac{3}{2} N k_B + C_V^{\rm ex}
\]
with the ideal-gas part ($\frac{3}{2} N k_B$) subtracted to define the excess. In reduced units (per particle), this reads $c_V^{\rm ex} = A (T_m/T)^{\alpha}$, where $T_m$ is the melting temperature and $A$ is a constant of order unity.

Extensive numerical studies confirm that the $T^{-2/5}$ scaling holds for the OCP and for many van der Waals and metallic atomic liquids in their strongly coupled regimes [2601.01659, 1305.1776]. However, deviations are observed for certain fluids, densities, and ranges of temperature, necessitating the generalized RT scaling.

## 3. Two-Phase Model and Microscopic Origin

A rigorous microscopic rationale for RT scaling is provided by the two-phase model, in which the liquid is conceived as a binary mixture of "solid-like" and "gas-like" microstates [2512.11520]. In this framework, the fraction $x(T)$ of "solid-like" vibrational modes governs the energy and heat capacity:
\[
E(T) = N k_B T [3x(T) + \frac{3}{2}(1-x(T))]
\]
\[
C_V = \frac{dE}{dT} = \frac{3}{2} N k_B \left[ 1 + x(T) + T \frac{dx}{dT} \right]
\]
Above the melting point, scale invariance implies a power-law form $x(T) = (T_m/T)^{\alpha}$ with temperature, reproducing the generalized RT scaling $C_V^{\rm ex} \propto T^{-\alpha}$. Physically, $x(T)$ is the liquid rigidity parameter, providing a quantitative link between thermodynamic and instantaneous normal mode analyses.

The exponent $\alpha$ controls the thermal decay of correlations: large $\alpha$ leads to rapid decay (quick crossover to gas-like behavior), while small $\alpha$ sustains solid-like character at higher temperatures. Empirical fits place $\alpha$ in the range $0.3$–$0.4$ for many monatomic liquids; e.g., $\alpha \approx 1/3$ is a particularly robust value near melting for model systems [2512.11520]. The special case $\alpha = 1/3$ corresponds to occurrence of the "Frenkel line" criterion (rigidity loss) at $T \approx 10\, T_m$.

| Exponent $\alpha$ | Representative System | Regime of Validity          |
|-------------------|----------------------|-----------------------------|
|    $2/5$          | OCP, high-density LJ | Strong coupling, near-freezing|
|   $\approx 1/3$   | Argon, low-$\rho$ LJ | Near triple point, low density|
|$0.27 \!-\! 0.43$  | Noble Metals         | Varies with $\rho$, species  |

## 4. Generalized RT Scaling and System Dependence

Simulations and NIST data analyses on noble gas liquids and metals have established that no single universal exponent $\alpha$ describes all systems across all densities and temperatures [2505.15986]. Instead, the generalized RT scaling
\[
C_V^{\rm ex} \propto (T_m/T)^{\alpha}
\]
is treated as a phenomenological law in which $\alpha$ is a material- and density-dependent parameter, compiled empirically (see Table I in [2505.15986]). At low densities or for "softer" interparticle potentials, the best-fit $\alpha$ can be closer to $1/3$; at higher densities and for stiffer potentials, it recovers the classic RT exponent $2/5$. For very steep (hard-sphere–like) potentials, temperature dependence is suppressed ($\alpha \rightarrow 0$).

No general microscopic theory exists that predicts $\alpha(\rho, \phi(r))$ solely from pair potential characteristics; heuristic arguments (e.g., $\alpha \simeq 1-3/n$ for $\phi(r) \sim r^{-n}$) overestimate observed values because attractive forces and anharmonicity reduce the effective exponent. Ongoing efforts aim to link RT scaling exponents to fundamental physical parameters.

## 5. RT Scaling, Isomorph Theory, and Thermodynamic Path Dependence

Rosenfeld–Tarazona scaling is most accurate in liquids that exhibit strong correlations between equilibrium fluctuations of virial and potential energy—i.e., Roskilde-simple liquids, characterized by a high correlation coefficient $R \gtrsim 0.9$ [1305.1776]. For these systems, isomorph theory provides a formal underpinning: excess entropy and structure remain invariant along isomorphs in the $(\rho,T)$ phase diagram, and RT scaling gives $C_V^{\rm ex}(\rho, T) = [h(\rho)/T]^{2/5}$ with $h(\rho)$ specified by the potential and correlation slope $\gamma$ at a reference state.

A key limitation is the path-dependence of RT scaling in anomalous fluids or those with competing length scales (e.g., core-softened, water-like anomalies) [1004.1894, 1008.0939]. In such cases, the monotonic mapping between excess entropy and dynamical or thermodynamic quantities can break down along specific thermodynamic paths. RT scaling is valid along isochores (fixed density) where monotonicity is preserved, but may fail along isotherms or across coexistence/critical anomalies.

## 6. Broader Generalizations: Diffusion, Transport, and Complex Fluids

The RT paradigm has been extended to scaling relations for transport and dynamic properties—such as reduced diffusivity and viscosity—via exponential functions of excess entropy $S_\text{ex}$. These relations remain robust across a wide range of monatomic, molecular, and even colloidal and Brownian systems, provided the mapping is adjusted (via so-called generalized Rosenfeld scaling) for system-specific prefactors and reductions [1101.1982, 1508.06337]. For example, Brownian dynamics in ultrasoft-interaction fluids is successfully described using a generalized RT master curve, predicting long-time diffusivity to within $\sim 20\%$ of simulation data [1101.1982]. Analytic derivations for random energy landscapes confirm the exponential diffusion-entropy scaling when site energies are uncorrelated, but this scaling fails in the presence of static spatial correlations [1508.06337].

## 7. Practical Guidelines, Applications, and Open Problems

RT scaling serves as a practical predictive tool for thermodynamic modeling and extrapolation in engineering and computational chemistry. Upon empirical determination of the appropriate scaling exponent from either experiment or simulation—typically via fitting $C_V^{\rm ex}$ near the melting line—one can extrapolate heat capacities or predict behavior in unexplored thermodynamic regions, provided the underlying system satisfies the criteria for Roskilde-simplicity and remains in the relevant phase. For Roskilde-simple systems, the isomorph theory in conjunction with RT scaling enables a fully predictive two-parameter model for $C_V^{\rm ex}(\rho,T)$ using only the virial–potential-energy correlation coefficient at a single reference point [1305.1776].

Open questions remain regarding the first-principles prediction of $\alpha$ for arbitrary interaction potentials, quantification of deviations in systems with mixed or network bonding, and the adaptation of RT-like scaling to critical and highly anomalous regimes. The interplay of RT scaling with glassy dynamics, fragility, and higher-order entropy contributions continues to be an active area of research [1706.01715].

---

**References:**  
[2512.11520], [2505.15986], [1305.1776], [2601.01659], [1004.1894], [1008.0939], [1101.1982], [1508.06337], [1706.01715]

Source: https://www.emergentmind.com/topics/rosenfeld-tarazons-scaling