Skip to content

Thermodynamic Temperature

The thermal state quantity T of a specified equilibrium system, independent of the thermometer or temperature scale used to assign its value.

Version
v1 · 2026-10-07 · History
Domain-specific #
14034
Domain group
Natural Sciences
Origin domain
Physics
Subdomains
Thermodynamics, Temperature Metrology → Physics
Aliases
Absolute Thermodynamic Temperature

Core Idea

Thermodynamic temperature, denoted T, is the thermal state quantity of a specified physical system in equilibrium under a stated thermodynamic convention. It characterizes that state independently of the device or temperature scale used to assign a number. The kelvin is its SI unit; the present definition fixes the numerical value of the Boltzmann constant k exactly. A temperature scale such as ITS-90 supplies a practical approximation T₉₀, not a second definition of the underlying T.[1][2]

Thermodynamic relations connect T to the state. In a regular formulation with internal energy U(S,V,{nᵢ}), the fundamental relation gives T = (∂U/∂S) with volume and composition fixed. This is one characterization with declared variables, not a claim that every physical realization uses that derivative or that it remains differentiable across every phase boundary. Gas acoustics and electrical thermal noise are unlike ways of realizing the same quantity.[3][2]

Structural Signature

Sig role-phrases:

  • Declared equilibrium carrier — specifies the system whose thermal state is at issue and the conditions under which a definite T is assigned. A free-standing sound speed or noise voltage has no identified temperature subject by itself.[3][2]
  • Thermal state quantity T — remains the target when the measuring apparatus, substance or practical scale changes. It is distinct from thermal energy kT, total internal energy and a unit label.[1][2]
  • Applicable thermal relation — connects T to the declared state under stated assumptions, whether through an equilibrium thermodynamic derivative or a specified primary thermometry equation. No single Carnot, gas or electrical formula is universal.[3][2]
  • Quantity–realization distinction — separates a state property from the observable and the convention assigning numerical values. Acoustic resonance and resistor noise can infer T through different physical laws; neither signal is itself T.[2][4]

What It Is Not

It is not the kelvin. The kelvin is a unit defined through fixed k; the quantity being measured can exist without anyone reporting its value in that unit. It is not identical by definition to T₉₀: the International Temperature Scale of 1990 defines a close practical approximation, with specified fixed points and interpolation. Nor is it simply the mean particle energy or the total internal energy; the SI Brochure cautions that T need not scale with the latter.[1][2]

It is not a thermometer reading stripped of its model. A measured resonance frequency or voltage fluctuation is evidence for T only when the relevant gas or conductor relation, parameters, corrections and uncertainty are supplied. A Carnot-cycle ratio can express a thermodynamic temperature relation under its own assumptions, but it is not a membership condition for every T realization.[2]

Scope of Application

The admitted scope is equilibrium thermal states for which a thermodynamic temperature can be assigned under a specified convention. Primary thermometry infers that quantity from a well-understood physical system and independent measurements. Absolute primary routes use the fixed value of k without requiring a temperature fixed point. Relative primary methods can use a reference temperature; defined scales such as ITS-90 serve practical assignment and dissemination.[2][1]

The old kelvin definition fixed the water triple point at exactly 273.16 K. Under the present definition, its thermodynamic temperature must be determined experimentally. The triple point remains a valuable realization setting, but it is not the constitutive anchor of T or its present unit definition.[1][2]

A bounded-spectrum negative-temperature question depends on the entropy convention and physical preparation. The two inspected author abstracts take opposing interpretive positions. This entry does not use a nuclear-spin claim as a positive case, nor does it settle that controversy from abstract-level evidence.[5][6]

Clarity

Ask first, “Which system and equilibrium state?” Then state the relation used to connect that state to T, the observable that was measured, and the unit or practical scale used in the report. These are different roles. Changing a resonator to a resistor changes the realization route, not the identity of thermodynamic temperature; replacing T with T₉₀ or an uncorrected readout changes the quantity being claimed.[1][2]

For a derivative statement, name what is fixed. The MIT lecture writes U(S,V,{nᵢ}) and dU = T dS − p dV + Σ μᵢ dnᵢ. It warrants T=(∂U/∂S)_{V,n} where that equilibrium relation is differentiable. The reciprocal entropy derivative requires local invertibility and finite nonzero T; these qualifications matter near singular or degenerate regimes.[3]

Manages Complexity

Temperature measurements involve gases, cavities, electrical conductors, quantum voltage sources, corrections and calibration histories. The abstraction keeps the target quantity T stable while allowing those measurement chains to differ. The Metre Convention's practical document organizes methods by how directly they realize thermodynamic T, preventing a convenient defined scale from silently replacing the target.[2][4]

The compression is useful only if its limits travel with it. An acoustic formula uses an ideal-gas and zero-frequency limit; a noise formula uses a low-frequency condition and an electrical response model. Labeling both results “temperature” without those conditions would erase the evidence connecting each signal to the state quantity.[2]

Abstract Reasoning

On a regular equilibrium branch, the fundamental relation makes T the entropy-conjugate coefficient in dU at fixed volume and composition. The equation tells us which variable is held fixed and permits local state reasoning. It does not turn T into an arbitrary proxy for heat content, nor prove a global smooth derivative across every equilibrium boundary.[3][1]

Primary thermometry supports a different inference pattern. For an ideal monatomic gas, the zero-frequency sound-speed relation has u² = γkT/m, with γ=5/3; measured resonances and corrections can infer T. For a conductor in the hf ≪ kT resistor limit, mean-square noise voltage over bandwidth Δf is 4kTRΔf. These relations do not make sound speed equal to noise voltage or make either quantity a universal equation of state for all materials. They converge on the same target T under their separate physical assumptions.[2]

Knowledge Transfer

The quantity–realization distinction transfers across thermal metrology. The same T can be approached through an acoustic gas resonator or a resistor with a quantum voltage noise source, with different dominant corrections and uncertainty budgets. A result obtained by one route can be compared with another only after their assumptions and traceability are stated.[2][4]

Outside equilibrium thermal physics, “temperature” often labels an effective or metaphorical scale. That vocabulary alone cannot transfer this entry's thermodynamic identity. A field-specific effective temperature requires its own state relation and validity conditions. This entry remains domain-specific; the present live catalog does not supply a proved strict full-signature parent across its admitted scope.[3][2]

Examples

Canonical realization: acoustic gas thermometry

The BIPM's kelvin mise en pratique describes primary acoustic gas thermometry using resonances of a monatomic gas in an isothermal cavity. In the ideal, zero-frequency limit, u² = γkT/m; real-gas behavior is handled through corrections and pressure extrapolation. Mapped back: the gas in the cavity is the declared equilibrium carrier; its T is the thermal state quantity; the sound-speed equation is the applicable relation; and cavity resonances are the realization readout, not the quantity. BIPM reports triple-point-water implementations but notes that the cited lowest acoustic uncertainties had not yet been independently confirmed in that document.[2]

Applied realization: resistor Johnson-noise thermometry

The BIPM also describes absolute primary Johnson-noise thermometry. Thermal fluctuations of charge carriers in a conductor produce a voltage-noise spectrum; in the applicable resistor and low-frequency limit, the measured mean-square voltage is related to T by 4kTRΔf. Above 1 K, an electronic comparison can use a quantum voltage noise source. Mapped back: the resistor/conductor is the equilibrium carrier; its T is the target quantity; the Nyquist limit is the thermal relation; and measured noise, resistance, bandwidth and comparison waveform belong to the realization. Benz and colleagues review this electronic method; the method is not a separate kind of temperature.[2][4]

Structural Tensions

No all-instance opposed pressure is established by the source packet. Precision, convenience and model complexity are practical metrology choices, while T itself is a state quantity rather than an optimization process. Acoustic and electrical routes have different corrections; that variation is evidence for method independence, not a structural conflict inside every instance. Diagnostic: which model and correction chain justifies identifying a readout with this system's T?[2][4]

Structural–Framed Character

T is a structural physical quantity: a system's thermal state is not brought into existence by its SI name or an instrument choice. The institutional SI definition, practical scales and traceability rules frame how laboratories assign and compare numerical reports. Thus the entry lies on the structural side of the spectrum with a metrological frame around its measurement. It has little inherent evaluative weight; hotter or colder is a state comparison, not a judgment of quality. The term travels widely, but literal recognition requires an equilibrium carrier and a valid thermal relation; importing it to an unrelated “temperature” metaphor changes the identity. Its character: a physical equilibrium quantity whose numerical realization is convention-governed and method-dependent, while the target remains one T.[1][2]

Structural Core vs. Domain Accent

The broad skeleton is a state quantity inferred through multiple observables while retaining one target identity. Here the residual is thermal equilibrium physics: T, entropy-energy relations where applicable, physical primary thermometry laws, and a kelvin assigned by fixed k. Prime bar: remove thermal carrier and governing relations and this named quantity no longer exists; only a generic state-property or measurement analogy remains. Accordingly, the entry is domain-specific. The generic state-property/measurement skeleton is a possible future Prime question, not a newly accepted Prime or a strict edge from this name. State Function is a close live neighbor, but its full exact-differential and equation-of-state obligations are not proved for every admitted boundary case, so no strict edge is asserted.[3][1][2]

Measurement is relevant when T is actually measured, but an unmeasured equilibrium system still has a thermal state; the measurement operation is not its genus. Invariance describes why a quantity does not depend on the thermometer route, yet the live strict edge proof is not established through that broad analogy. The provisional DAG placement is an unparented root with no strict upward edge; it can be revised when an all-instance parent with a fitting signature is curated.[2]

Neighborhood in Abstraction Space

Thermodynamic Temperature sits in a sparse region of the domain-specific corpus (66th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Condensed Matter & Physical Chemistry Models (26 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-08

Not to Be Confused With

  • Kelvin: the SI unit of thermodynamic temperature, presently defined by fixed k.[1]
  • ITS-90 T₉₀: a defined-scale quantity close to, but not definitionally identical with, thermodynamic T.[2]
  • State Function: a close thermodynamic neighbor; its current live entry requires an exact differential/equation-of-state proof not supplied across this entry's full admitted scope.
  • State Variable: a chosen coordinate in a model; T can be a thermal state quantity without being selected as a sufficient coordinate in every model.
  • Thermometer readout: an acoustic frequency or noise voltage used to infer T through a model, not T itself.[2]

References

[1] Bureau International des Poids et Mesures, The International System of Units (SI Brochure), 9th ed., version 3.01, 2024, §2.2 and §2.3.1, printed pp.130, 134. Full official PDF inspected. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i ↩j

[2] BIPM Consultative Committee for Thermometry, Mise en pratique for the definition of the kelvin in the SI, MeP-K-19D, 2019 core with 2024 digital reference update, §§3.1–3.2, 4.1, 4.4. Full official PDF inspected. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i ↩j ↩k ↩l ↩m ↩n ↩o ↩p ↩q ↩r ↩s ↩t ↩u ↩v ↩w ↩x

[3] Massachusetts Institute of Technology, Physical Chemistry II Lecture 1, Spring 2008, printed p.3. Instructor lecture supports the equilibrium fundamental relation, not a new experiment. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g

[4] Samuel P. Benz et al., “Practical realisation of the kelvin by Johnson noise thermometry”, Metrologia 61 (2024), 022001, DOI 10.1088/1681-7575/ad2273. Full author review PDF inspected; not cited as a separate new laboratory experiment. registry ↩a ↩b ↩c ↩d ↩e

[5] Daan Frenkel and Patrick B. Warren, “Gibbs, Boltzmann, and negative temperatures”, arXiv:1403.4299 (2014), author abstract only; included to mark a disputed convention boundary, not as a positive case. registry ↩

[6] Jörn Dunkel and Stefan Hilbert, “Inconsistent thermostatistics and negative absolute temperatures”, arXiv:1304.2066 (2013 preprint), author abstract only. The cited preprint does not resolve the negative-temperature debate here. registry ↩