---
title: Contact Temperature in Non-equilibrium Systems
url: https://www.emergentmind.com/topics/contact-temperature
type: topic
---

# Contact Temperature in Non-equilibrium Systems

Searching arXiv for recent and foundational papers on contact temperature and closely related formulations.
arxiv_search(query="all:\"contact temperature\" OR ti:\"contact temperature\"", max_results=10, sort_by="submittedDate")
arxiv_search(query="ti:\"contact temperature\" OR abs:\"contact temperature\"", max_results=10, sort_by="relevance")
Contact temperature is an operationally defined temperature for a non-equilibrium system: it is the thermostatic temperature of an equilibrium environment for which the net heat exchange with the system vanishes. In the formulation introduced and developed by Muschik for closed discrete systems, the contact temperature \(\Theta\) is treated as an internal variable, so that it enters the non-equilibrium state space, entropy rate, and entropy production without appearing in the mechanical work term of the first law [1709.03156]. Recent work extends the same operational logic to arbitrary quantum states through a universal thermometer that assigns a unique inverse contact temperature \(\beta_{\rm op}\) by the condition of zero heat flow [2606.31969], while related nonequilibrium steady-state constructions identify a unique weighted average temperature \(T_e\) that restores equilibrium-like thermodynamic relations at the level of averages and response [2406.05801].

## 1. Definition by thermal contact and zero heat flow

For a closed discrete system, the defining construction is to place the system boundary in thermal contact, through an inert partition, with an equilibrium heat reservoir at thermostatic temperature \(T^*\). If \(Q\) is the net heat flow from the reservoir into the system, then the contact temperature \(\Theta\) is defined by  
\[
Q\left(\frac{1}{\Theta}-\frac{1}{T^*}\right)\ge 0
\qquad \text{for all } T^*,
\]
with the immediate consequence
\[
Q=0 \iff T^*=\Theta.
\]
Accordingly, \(\Theta\) is that thermostatic temperature of an equilibrium environment for which the net heat exchange with the non-equilibrium system vanishes [1709.03156].

Several properties follow directly from this definition. The contact temperature depends on the system’s non-equilibrium state and on the properties of the partition, but not on the reservoir’s state except through \(Q\). Since the boundary is inert, the construction is purely thermal: the partition is energy-nonabsorbing and nonmaterial-transporting. In a closed system, and with no work exchange through the same boundary, \(\Theta(t)\) and the internal energy \(U(t)\) are independent variables. These are precisely the properties used to classify \(\Theta\) as an internal variable rather than a control parameter [1709.03156].

The operational character of \(\Theta\) is essential. It is measurable by adjusting \(T^*\) until \(Q=0\), yet it is not a work variable and does not enter the work term of the first law. In true equilibrium, it reduces to the ordinary thermostatic temperature \(T(U,a)\). This gives a non-equilibrium notion of temperature that remains meaningful even when different “thermometers” would otherwise give different readings.

## 2. Contact temperature as a state variable and entropy-rate contribution

To include contact temperature in the thermodynamic description, the state space of the closed system is enlarged from the usual equilibrium variables to  
\[
Z=(U,a,\Theta,\xi),
\]
where \(a\) denotes the mechanical work variables and \(\xi\) denotes other internal variables. The entropy is then postulated as a state function
\[
S=S(U,a,\Theta,\xi).
\]
Along a process trajectory \(Z(t)=(U(t),a(t),\Theta(t),\xi(t))\), the first law is written as
\[
\dot U = Q+W,\qquad W=A\cdot \dot a,
\]
and the entropy rate becomes
\[
\dot S
= \frac{1}{\Theta}\dot U - \frac{A}{\Theta}\cdot \dot a + \alpha \dot\Theta + B\cdot \dot\xi
= \frac{1}{\Theta}Q + \alpha \dot\Theta + B\cdot \dot\xi,
\]
with
\[
\alpha:=\frac{\partial S}{\partial \Theta},
\qquad
B:=\frac{\partial S}{\partial \xi}.
\]
In this formulation, the term \(\alpha \dot\Theta\) is the explicit entropy-rate contribution associated with the contact temperature [1709.03156].

For an isolated system, where \(Q=0\) and \(\dot a=0\), the entropy production is
\[
E:=\dot S=\alpha \dot\Theta + B\cdot \dot\xi \ge 0.
\]
This shows that both \(\Theta\) and \(\xi\) must be endowed with non-negative dissipation forces \(\alpha\) and \(B\), respectively. The consistency condition further imposes
\[
\frac{\partial S}{\partial U}=\frac{1}{\Theta},
\]
which replaces the usual Clausius definition \(1/T=\partial S/\partial U\) by its non-equilibrium analogue, valid along any process [1709.03156].

A central consequence is that the entropy balance of a non-equilibrium discrete system can be written entirely within an enlarged state description, without requiring a global equilibrium temperature field. This makes contact temperature structurally comparable to other internal variables used in non-equilibrium thermodynamics.

## 3. Evolution law and constitutive closure

Once \(\Theta\) is promoted to the rank of a state variable, a kinetic law is required. Starting from the defining inequality, the simplest constitutive relation for the heat flux through the inert partition is
\[
Q=\kappa\,[T^*-\Theta],\qquad \kappa>0,
\]
with the possibility that \(\kappa=\kappa(T^*-\Theta)\). Since \(T^*\) and \(Q\) are measurable quantities, this constitutive equation allows \(\Theta\) to be determined experimentally and treated as an internal variable [1709.03156].

Differentiation with respect to time gives
\[
\dot Q=
\bigl[\kappa'(T^*-\Theta)(T^*-\Theta)+\kappa\bigr](\dot T^*-\dot\Theta),
\]
and hence an explicit evolution law
\[
\dot\Theta
=
\dot T^*
-
\bigl[\kappa'(T^*-\Theta)(T^*-\Theta)+\kappa\bigr]^{-1}\dot Q.
\]
This relation shows that \(\Theta(t)\) responds to changes in the heat flow \(Q(t)\) and thus behaves as a genuine internal variable [1709.03156].

In the special case of a fixed environment, \(\dot T^*=0\), and with \(\kappa\) taken constant, the evolution reduces to a first-order relaxation law,
\[
\dot\Theta = -(\kappa/\alpha)(\Theta-T^*),
\]
where \(\alpha\) may be absorbed into a relaxation time. This form makes explicit the tendency of the contact temperature to relax toward the reservoir temperature under continued thermal contact.

The constitutive role of \(\kappa\) is limited to heat exchange; \(\Theta\) still does not appear in the mechanical work term. This separation between thermal kinetics and work variables is one of the reasons \(\Theta\) is classified as an internal variable rather than a degree of freedom in the mechanical sense.

## 4. Clausius-type inequalities and finite-time heat engines

By enlarging the state space to \(Z=(U,a,\Theta,\xi)\), the entropy-rate balance acquires the additional term \(\alpha \dot\Theta\). In isolation one must have \(\alpha\dot\Theta\ge 0\), so that changes in the contact temperature are driven by the generalized force \(\alpha\). Through Clausius’ integral theorem one recovers the extended Clausius inequality
\[
\oint \frac{Q}{\Theta}\,dt \ge 0,
\]
which is valid without assuming equilibrium or a single global temperature field [1709.03156].

The same formalism yields a finite-time heat-engine result. Consider a cyclic, power-producing system operating between two heat reservoirs at fixed thermostatic temperatures \(T_H>T_L\). In non-equilibrium the two contacts acquire contact temperatures \(\Theta_H(t)\) and \(\Theta_L(t)\), with corresponding heat flows \(Q_H(t)<0\) into the system and \(Q_L(t)>0\) out of the system. From the defining inequality,
\[
T_H \ge \Theta_H(t), \qquad \Theta_L(t)\ge T_L \qquad \forall\, t.
\]
If overlines denote time-averages over one cycle, then
\[
|Q_H|/\bar\Theta_H \le |Q_H|/T_H,
\qquad
Q_L/\bar\Theta_L \ge Q_L/T_L.
\]
Combining these relations with the first-law constraint on the cycle and defining the work output \(W\), one obtains the non-equilibrium efficiency
\[
\eta_{\rm neq}=1-\frac{\bar\Theta_L}{\bar\Theta_H},
\]
which is strictly less than the Carnot efficiency
\[
\eta_C=1-\frac{T_L}{T_H}.
\]
In this sense, the contact temperatures \(\Theta_H\) and \(\Theta_L\) contain the finite-time and irreversible corrections that reduce efficiency below the ideal Carnot bound [1709.03156].

A plausible implication is that contact temperature provides a compact way to encode irreversible boundary effects without introducing ad hoc efficiency corrections. In the cited formulation, these corrections emerge directly from the zero-heat-flow definition and the entropy balance.

## 5. Related formulations in steady states and quantum systems

The operational idea behind contact temperature has been generalized beyond closed discrete classical systems. For a harmonic chain whose two ends are coupled to two baths at temperatures \(T_L\) and \(T_R\), the effective temperature in the nonequilibrium steady state is taken to be a weighted average
\[
T_e=C_L T_L + C_R T_R,
\qquad
C_L+C_R=1,
\]
with weights determined by the bath couplings and by internal asymmetries encoded in the steady-state covariance structure. In this framework, the average kinetic energy per mode is precisely
\[
\langle p^T p\rangle/(2N)=T_e,
\]
the internal energy is
\[
E=\langle H\rangle = N T_e,
\]
and equilibrium-like formulas for entropy and free energy are restored after the replacement \(T\to T_e\) [2406.05801].

For arbitrary finite-dimensional quantum states, a different construction uses a universal thermometer. The sample \(A\) is a finite \(d\)-level system with nondegenerate Hamiltonian \(H_A\), and the thermometer \(B\) is prepared in a Gibbs state at inverse temperature \(\beta\). A thermal process \(U\) is any unitary satisfying
\[
[U,H_{AB}]=0,
\]
so that total energy is conserved. If the initial state is \(\rho\otimes \tau_\beta\), the heat flow into \(A\) is
\[
Q_A(\beta)
=
\mathrm{Tr}_{AB}\!\big[U(\rho\otimes\tau_\beta)U^*(H_A\otimes \mathbf 1)\big]
-
\mathrm{Tr}_A(\rho H_A).
\]
The contact temperature is then defined by the unique value \(\beta_{\rm op}\in\mathbb R\) for which the heat flow vanishes. The existence and uniqueness of \(\beta_{\rm op}\) follow from the fact that the heat-flow function is strictly decreasing and continuous, together with its boundary values [2606.31969].

| Context | Operational condition | Resulting temperature |
|---|---|---|
| Closed discrete non-equilibrium system | \(Q(1/\Theta-1/T^*)\ge 0\), with \(Q=0 \iff T^*=\Theta\) | Contact temperature \(\Theta\) |
| Harmonic chain in a NESS | \(T_e=C_L T_L + C_R T_R\) and \(\langle p^T p\rangle/(2N)=T_e\) | Weighted average temperature \(T_e\) |
| Arbitrary quantum state | Unique \(\beta_{\rm op}\) such that \(\bar Q_A(\beta_{\rm op})=0\) | Contact temperature \(\beta_{\rm op}^{-1}\) |

The quantum construction reproduces the usual equilibrium temperature for Gibbs states, and negative \(\beta_{\rm op}\) occurs for active, population-inverted states. For a qubit with populations \(p_1,p_2\) and energy gap \(\varepsilon_2-\varepsilon_1\), the unique solution is
\[
\beta_{\rm op}
=
\frac{\ln(p_1)-\ln(p_2)}{\varepsilon_2-\varepsilon_1},
\]
which exactly reproduces the usual effective temperature that equates populations [2606.31969].

## 6. Terminological distinctions and adjacent usages

The phrase “contact temperature” is not used uniformly across the literature. In the discrete non-equilibrium framework discussed above, it denotes the thermostatic temperature of an equilibrium environment for which the net heat exchange with the non-equilibrium system vanishes [1709.03156]. In a classical two-body heat-exchange setting, however, the same phrase can denote the final uniform temperature of the composite after direct contact,
\[
T_c=\frac{T_1 C_1 + T_2 C_2}{C_1+C_2},
\]
where \(C_1\) and \(C_2\) are heat capacities. Mishchenko and Pshenichka showed that if each body is split into \(N\) equal sub-bodies and these are brought into thermal contact sequentially, the two bodies can completely exchange their temperatures reversibly as \(N\to\infty\), with total entropy production tending to zero [1702.08845]. This is a distinct use of the phrase.

A common source of confusion is the ultracold-gas literature, where “contact” usually refers not to a temperature but to Tan’s contact or related short-range correlation parameters. In a unitary Fermi gas, Tan’s contact \({\cal I}\) is the amplitude of the large-momentum tail,
\[
{\cal I}=\lim_{k\to\infty} k^4 n_\sigma(k),
\]
and it governs short-range pair correlations and the large-\(k\) structure factor [1012.2626]. In a Tonks–Girardeau gas, the contact \(C\) is the coefficient of the \(1/k^4\) tail of the momentum distribution [1209.1545]. In a weakly interacting single-component Fermi gas, the \(p\)-wave contact \(C_v(T)\) is defined thermodynamically by
\[
C_v(T)=-(2m/3\hbar^2)\,\partial F/\partial(1/v_p),
\]
and controls the \(k^{-2}\) tail and the two-body loss rate [2306.15904]. These objects are universal correlation measures, not temperatures.

There is also a purely geometrical adjacency in the phrase “contact line.” In liquid nitrogen, two-color temperature-sensitive paint has been used to measure the temperature field around a cryogenic three-phase contact line under heat fluxes of \(110\), \(430\), and \(900\ \mathrm{W/m^2}\), revealing a localized temperature minimum of about \(78\ \mathrm K\) at the observed contact line while the liquid region remains near \(81\ \mathrm K\) and the gas region heats with increasing flux [2606.24533]. This concerns a temperature distribution near a contact line, not the thermodynamic contact temperature of a system defined by zero net heat flow to a reservoir.

Taken together, these distinctions show that “contact temperature” is a specific thermodynamic notion tied to operational thermal contact and vanishing heat exchange. Where the literature uses “contact” for short-range correlations, effective temperatures, or spatial contact-line measurements, the underlying concepts are different even when the terminology partially overlaps.

Source: https://www.emergentmind.com/topics/contact-temperature