{
  "id": "cpt-theorem",
  "title": "CPT Theorem",
  "lang": "en",
  "source_lang": "en",
  "summary": "The CPT theorem establishes a combined symmetry of charge conjugation, spatial inversion, and time reversal in local relativistic quantum field theory.",
  "category": "science",
  "subcategory": "physics",
  "type": "concept",
  "aliases": [
    "CPT invariance theorem",
    "CPT symmetry theorem",
    "charge–parity–time theorem",
    "TCP theorem"
  ],
  "generated_at": "2026-11-05",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/cpt-theorem/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/cpt-theorem/",
      "json": "https://aiwiki.page/zh/cpt-theorem.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "The CPT theorem states that a broad class of quantum field theories is invariant under the combined operation of charge conjugation, parity inversion, and time reversal. Charge conjugation exchanges particles with their antiparticles, parity reverses spatial coordinates, and time reversal reverses the temporal direction of a physical process. Although any one of these transformations—or certain pairs of them—may fail to be a symmetry of nature, their combined action is guaranteed under the theorem’s assumptions.\n\nThe theorem links the mathematical foundations of special relativity and quantum mechanics to experimentally testable statements about matter and antimatter. Among its consequences are equality of particle and antiparticle masses and total lifetimes, together with precise relations between their electromagnetic properties and reaction probabilities. It also connects violations of combined charge conjugation and parity to violations of time-reversal symmetry.\n\nCPT invariance is a structural result rather than a separate empirical rule added to the Standard Model. Its status nevertheless depends on assumptions, notably relativistic covariance and locality. Precision comparisons of particles and antiparticles therefore test not only particular particle models but also the framework in which ordinary relativistic quantum field theory is formulated."
    },
    {
      "heading": "Charge conjugation, parity, and time reversal",
      "text": "The three operations entering CPT act differently on physical systems. Charge conjugation, denoted \\(C\\), exchanges a particle species with its antiparticle and reverses additive internal quantum numbers such as electric charge. An electron is mapped to a positron, while a proton is mapped to an antiproton. Charge conjugation is not simply reversal of electric charge: electrically neutral particles can carry other quantum numbers that distinguish them from their antiparticles. Some neutral particles, such as the photon, are their own antiparticles.\n\nParity, denoted \\(P\\), reverses all three spatial coordinates:\n\n\\[(t,\\mathbf{x})\\longmapsto(t,-\\mathbf{x}).\\]\n\nA spatial inversion changes the sign of polar vectors such as position and momentum, but not axial vectors such as angular momentum. It also exchanges left- and right-handed spatial arrangements. In three spatial dimensions, inversion of all coordinates can be expressed as a single-plane reflection followed by a rotation, which explains the frequent description of parity as a mirror transformation. Its action on fields also includes intrinsic parity factors and, for spinor fields, a matrix acting on their components.\n\nTime reversal, denoted \\(T\\), changes \\(t\\) to \\(-t\\). In a time-reversed motion, momentum and angular momentum reverse sign. Position does not. For electromagnetic fields, conventional time reversal leaves the electric field unchanged and reverses the magnetic field, when field values are compared at the corresponding reversed times. Reversing a process requires transforming all relevant degrees of freedom, including external fields and preparation conditions; merely playing a recording backward is not a complete physical definition.\n\nTheir combined spacetime action is\n\n\\[(t,\\mathbf{x})\\longmapsto(-t,-\\mathbf{x}),\\]\n\ntogether with particle–antiparticle exchange. For a one-particle state, the momentum reversals produced by \\(P\\) and \\(T\\) cancel. CPT therefore maps it to an antiparticle state with the same energy and momentum but reversed spin projection. Since momentum is unchanged and spin is reversed, helicity—the projection of spin along momentum—changes sign.\n\nThe assertion of CPT invariance concerns this complete transformation, not three independent invariances. A theory can distinguish particles from antiparticles, distinguish left from right, and distinguish a reaction from its time reverse while still satisfying CPT exactly. The weak interaction provides the central physical example: its violations of several simpler discrete symmetries are compatible with the combined theorem."
    },
    {
      "heading": "Mathematical statement and assumptions",
      "text": "In a standard formulation, the theorem concerns quantum fields on four-dimensional flat spacetime. The fields act on a positive-definite Hilbert space, transform covariantly under the proper orthochronous Lorentz group, and obey an appropriate locality condition. The energy–momentum spectrum lies in the closed forward light cone, ensuring nonnegative energy and excluding states with spacelike total four-momentum. A Poincaré-invariant vacuum is also assumed. These conditions belong to the framework commonly called the Wightman axioms, although equivalent CPT results can be established in other formulations.\n\nLorentz covariance specifies how fields behave under continuous rotations and boosts. The proper orthochronous Lorentz group preserves spatial orientation and the direction of time; neither parity nor time reversal is included in its connected component. CPT is therefore not initially assumed as one of these continuous spacetime symmetries. The theorem establishes that the additional combined discrete transformation follows from the other structural properties.\n\nLocality, or microcausality, requires observables in spacelike-separated regions to commute. For elementary fields, the corresponding condition is expressed through commutators for bosonic fields and anticommutators for fermionic fields. With metric signature \\(+---\\), a representative condition is\n\n\\[[\\mathcal{O}(x),\\mathcal{O}(y)]=0 \\qquad\\text{when}\\qquad (x-y)^2<0.\\]\n\nThis is a statement about the compatibility of independently performed local measurements. It does not exclude quantum entanglement, whose correlations do not permit controllable faster-than-light signalling.\n\nThe conclusion is the existence of an antiunitary operator \\(\\Theta\\) implementing CPT. It leaves the vacuum invariant up to an unobservable phase and maps each field to an appropriately transformed conjugate field at \\(-x\\). The exact field formula depends on spin, tensor structure, and phase conventions. For a complex scalar field, one may write, with a suitable convention,\n\n\\[\\Theta\\phi(x)\\Theta^{-1} =\\eta\\,\\phi^\\dagger(-x), \\qquad |\\eta|=1.\\]\n\nSpinor and tensor fields require additional transformation factors.\n\nA frequently used Lagrangian version states that a local, Hermitian, Lorentz-invariant field theory with the usual spin–statistics assignments possesses CPT symmetry. This is a practical criterion for conventional particle theories, not a substitute for specifying mathematical assumptions in every setting. In particular, a gauge-fixed formulation may contain unphysical states that do not have a positive-definite inner product. Its relationship to the theorem must be understood through the physical states and gauge-invariant observables rather than by applying every Hilbert-space assumption directly to auxiliary fields."
    },
    {
      "heading": "Antiunitarity and the structure of the proof",
      "text": "The antiunitary character of CPT is essential. A unitary operator is complex-linear and preserves inner products. An antiunitary operator is conjugate-linear: for complex numbers \\(a\\) and \\(b\\),\n\n\\[\\Theta(a|\\psi\\rangle+b|\\chi\\rangle) =a^*\\Theta|\\psi\\rangle+b^*\\Theta|\\chi\\rangle.\\]\n\nConsequently,\n\n\\[\\Theta i\\Theta^{-1}=-i.\\]\n\nThis complex conjugation is necessary for a time-reversing transformation to be compatible with a Hamiltonian whose energy spectrum is bounded below. If time reversal were represented by a unitary operator in the ordinary way, reversing the sign of time in quantum evolution would instead require reversing the sign of the Hamiltonian.\n\nThe operator formulation makes this explicit. If the Hamiltonian \\(H\\) is invariant under CPT, then\n\n\\[\\Theta H\\Theta^{-1}=H.\\]\n\nNevertheless, because \\(\\Theta\\) is antiunitary, the time-evolution operator satisfies\n\n\\[\\Theta e^{-iHt/\\hbar}\\Theta^{-1} =e^{+iHt/\\hbar}.\\]\n\nThus the dynamics is invariant while the temporal direction of evolution is reversed. In the conventional representation, \\(C\\) and \\(P\\) are unitary and \\(T\\) is antiunitary, making their combined product antiunitary as well.\n\nAxiomatic proofs use vacuum correlation functions, also called Wightman functions. These have the form\n\n\\[W(x_1,\\ldots,x_n) =\\langle0|\\phi_1(x_1)\\cdots\\phi_n(x_n)|0\\rangle.\\]\n\nThe positive-energy spectral condition allows these functions to be continued analytically into domains of complex spacetime coordinates. Lorentz covariance enlarges the available analytic domains. Within this complexified setting, total spacetime inversion is connected to the continuous Lorentz transformations used in the analytic argument, even though it is disconnected from ordinary proper orthochronous transformations at real coordinates.\n\nLocality supplies the other crucial step. At suitable configurations of spacelike-separated points, it permits reversal of field ordering, with the signs required for fermionic fields. Analytic continuation then extends the resulting relation beyond those special configurations. Combining covariance, analyticity, and reordered fields produces the correlation-function identity from which a CPT operator can be reconstructed.\n\nIn this approach, a condition called weak local commutativity at particular spacelike configurations is closely tied to CPT invariance. This reveals why the theorem is not merely an algebraic coincidence in familiar interaction terms. Its deeper content concerns analytic properties of quantum fields and the relationship between relativistic causality and operator ordering. The connection to the spin–statistics theorem is correspondingly close: both results derive strong restrictions on particle theories from overlapping assumptions, although they are distinct theorems and should not be identified."
    },
    {
      "heading": "Historical development",
      "text": "The CPT theorem emerged from mid-twentieth-century attempts to understand the general structure of relativistic quantum fields. Early work by Wolfgang Pauli on the relation between spin and statistics helped establish the importance of Lorentz covariance, energy positivity, and locality. Julian Schwinger’s 1951 analysis of the theory of quantized fields contained a closely related combined reflection result, although the modern theorem and its assumptions were clarified through several subsequent contributions.\n\nGerhart Lüders gave an important formulation in 1954, and Pauli presented a further treatment in 1955. Their work showed that a combined discrete symmetry could hold even when its separate factors did not. The terminology “TCP theorem” was common in early literature, reflecting a different conventional ordering of the same three transformations. Raymond Jost’s 1957 work supplied a particularly influential axiomatic treatment, expressing the result through analyticity and local commutativity of vacuum expectation values.\n\nThese developments initially preceded the experimental collapse of parity conservation as a universal principle. In 1956, Tsung-Dao Lee and Chen-Ning Yang examined the evidence for parity conservation in weak interactions and proposed decisive tests. The experiment led by Chien-Shiung Wu, reported in 1957, studied polarized cobalt-60 nuclei undergoing beta decay. The asymmetric emission of electrons relative to nuclear spin demonstrated that the weak interaction does not respect parity.\n\nCombined charge conjugation and parity, or CP symmetry, then became a possible replacement for the abandoned parity principle. That possibility was overturned in 1964 when James Cronin, Val Fitch, and collaborators observed a small decay component of the long-lived neutral kaon into two pions. The result established CP violation in neutral-kaon physics. Since CPT could remain valid, the discovery also implied time-reversal noninvariance within a CPT-invariant description.\n\nLater experiments separated the logical and experimental questions more sharply. Neutral-kaon measurements at CPLEAR reported a time-reversal asymmetry in 1998, and the BaBar experiment reported a direct time-reversal-violation measurement using neutral \\(B\\) mesons in 2012. These developments did not turn CPT into a consequence of observed CP or T violation. Rather, they showed how the theorem organizes distinct experimental comparisons and determines which combinations of asymmetries can coexist.\n\nThe theorem’s historical significance lies partly in this sequence. A mathematical result about the combined transformation survived the discovery that several apparently natural component symmetries fail. Discrete symmetries consequently became independent objects of precision measurement rather than interchangeable expressions of an assumed mirror symmetry of nature."
    },
    {
      "heading": "Particle–antiparticle masses, lifetimes, and moments",
      "text": "One of the theorem’s most direct consequences is equality of the mass of a particle and its antiparticle. CPT transforms a one-particle state into the corresponding antiparticle state without changing its energy or momentum. Since invariant mass is determined by the energy–momentum relation,\n\n\\[E^2=c^2\\mathbf{p}^{\\,2}+m^2c^4,\\]\n\nthe two states have the same mass. This applies to elementary particles and to corresponding composite particles, such as a proton and an antiproton, provided the comparison involves genuinely CPT-related states.\n\nFor unstable particles, CPT also requires equality of total decay widths and hence total lifetimes. In the usual exponential-decay approximation,\n\n\\[\\tau=\\frac{\\hbar}{\\Gamma},\\]\n\nwhere \\(\\Gamma\\) is the total energy width. The qualification “total” is important. CPT does not generally require every exclusive decay channel to have the same rate as its charge-conjugate channel. Direct CP violation can produce differences between individual particle and antiparticle decay rates, while the full sums over final states remain equal.\n\nThe theorem also constrains electromagnetic properties. Corresponding particles and antiparticles have equal-magnitude and opposite-sign electric charges. Their magnetic moments likewise have equal magnitudes and opposite signs when defined relative to the same spin direction. In the common expression\n\n\\[\\boldsymbol{\\mu} =g\\,\\frac{q}{2m}\\,\\mathbf{S},\\]\n\nequal masses and equal \\(g\\)-factors, together with opposite charges, produce opposite magnetic moments for equal spin orientations. Care is required because experimental reports sometimes quote magnitudes or define signed \\(g\\)-factors differently.\n\nA permanent electric dipole moment illustrates an important distinction between CPT and its components. For a nondegenerate spin state, an intrinsic electric dipole moment directed along spin violates parity and time reversal. It does not necessarily violate CPT. Electric-dipole searches can therefore probe new sources of CP violation while remaining completely consistent with the CPT theorem.\n\nNone of these statements means that matter and antimatter behave identically in an unchanged environment made of ordinary matter. An antiproton entering a material target can annihilate, whereas a proton usually does not. The two experiments are not related by a complete CPT transformation, which would also transform the target and reverse the relevant spacetime conditions. Similarly, differences induced by external electric or magnetic fields must be distinguished from intrinsic particle–antiparticle differences.\n\nThe theorem concerns the complete physical system and its dynamics. Translating its consequences into a laboratory comparison requires matching spin states, accounting for fields, and specifying which properties are intrinsic rather than environmental."
    },
    {
      "heading": "Scattering, decay, and neutral-meson mixing",
      "text": "CPT invariance constrains transition amplitudes by relating a reaction to the inverse reaction involving transformed antiparticle states. Let \\(i\\) and \\(f\\) denote initial and final asymptotic states. Up to convention-dependent phases, the relevant relation has the schematic form\n\n\\[\\mathcal{M}(i\\rightarrow f) = \\mathcal{M}(\\Theta f\\rightarrow\\Theta i).\\]\n\nThe interchange of initial and final states is indispensable. CPT does not simply identify a reaction with the process obtained by replacing every particle with its antiparticle while leaving all other conditions unchanged.\n\nIn scattering theory, the relation can be expressed through the scattering operator \\(S\\):\n\n\\[\\Theta S\\Theta^{-1}=S^\\dagger.\\]\n\nTogether with antiunitarity, this gives equality of the corresponding transition probabilities. Translating probabilities into cross sections or decay rates also requires the appropriate flux and phase-space factors. This formulation explains why CPT can coexist with different partial decay rates for a particle and its antiparticle: the direct CPT partner of a decay is a suitably transformed inverse process.\n\nNeutral mesons provide especially sensitive tests because a neutral meson and its antimeson can mix. The neutral kaon system consists of \\(K^0\\) and \\(\\overline{K}{}^0\\), distinguished by their quark content and strangeness. Weak interactions connect the two, producing propagating states with different masses and lifetimes. Analogous mixing occurs in neutral \\(B\\) and \\(D\\) mesons.\n\nWithin the standard effective description, time evolution in the meson–antimeson subspace is governed by\n\n\\[H_{\\mathrm{eff}}=M-\\frac{i}{2}\\Gamma,\\]\n\nwhere \\(M\\) and \\(\\Gamma\\) are Hermitian \\(2\\times2\\) matrices representing dispersive and absorptive contributions. The effective Hamiltonian is not Hermitian because probability can leave the subspace through decay. CPT invariance requires equality of its diagonal elements,\n\n\\[(H_{\\mathrm{eff}})_{11} =(H_{\\mathrm{eff}})_{22},\\]\n\nwithin this approximation. CP- and T-violating effects can still appear through mixing and decay amplitudes.\n\nInterference between mixing and decay allows experiments to extract tiny differences that would be inaccessible through a direct weighing of short-lived particles. Neutral-kaon analyses constrain CPT-sensitive mass differences at approximately the \\(10^{-18}\\,\\mathrm{GeV}\\) scale or below, depending on the fitted parameters and assumptions. The interpretation must specify whether decay-amplitude relations or other symmetry conditions have been imposed.\n\nEntangled meson pairs add further possibilities. By identifying one meson’s decay products, an experiment can infer a property of the partner state and compare ordered transitions. These measurements can distinguish CP, T, and CPT tests more directly than an undifferentiated comparison of matter and antimatter decay distributions."
    },
    {
      "heading": "Precision experimental tests",
      "text": "Experimental tests of CPT seek violations of consequences that should hold exactly within its domain. A null result does not prove the theorem’s assumptions for every physical scale, but it constrains departures in particular observables. The strongest tests are complementary: a bound on a neutral-meson mass parameter, a magnetic moment, and an atomic transition frequency probes different combinations of possible symmetry-breaking effects.\n\nOne major approach compares particle and antiparticle motion in electromagnetic traps. A Penning trap uses a magnetic field and an electrostatic trapping potential to confine charged particles. In the idealized nonrelativistic limit, the cyclotron angular frequency is\n\n\\[\\omega_c=\\frac{|q|B}{m}.\\]\n\nComparing cyclotron frequencies therefore compares charge-to-mass ratios, after accounting for trap imperfections and field variations. The BASE collaboration reported in 2022 a proton–antiproton charge-to-mass-ratio comparison at a fractional precision of about \\(16\\) parts per trillion.\n\nMagnetic-moment measurements use spin-precession frequencies as well as motional frequencies. Electron–positron comparisons have long supplied stringent tests, while trapped antiprotons allow analogous comparisons with protons. BASE reported a measurement of the antiproton magnetic moment at approximately \\(1.5\\) parts per billion precision in 2017. Such results constrain CPT-sensitive spin interactions in addition to testing the equality of ordinary magnetic-moment magnitudes.\n\nAntihydrogen provides a different class of comparison. It consists of an antiproton bound to a positron, corresponding to ordinary hydrogen. Since electrically neutral atoms cannot be confined by an ordinary charged-particle trap, experiments use magnetic trapping of suitable internal states. Laser spectroscopy then compares transition frequencies with those known from hydrogen. In 2018, the ALPHA collaboration reported agreement of the hydrogen and antihydrogen \\(1S\\)–\\(2S\\) transition frequencies at about two parts in a trillion.\n\nThese measurements require more than comparing two numerical frequencies. Magnetic fields shift atomic levels, trapped atoms occupy distributions of velocities, and interrogation methods introduce line-shape and systematic effects. Experiments must either remove such contributions or calculate their influence on the specific transition. A CPT test is consequently tied to a defined physical model of the apparatus and of the states being compared.\n\nOther searches examine neutral mesons, neutrinos, charged leptons, and the orientation dependence of precision clocks. Neutrino–antineutrino comparisons require particular care because propagation through ordinary matter can produce different effective behavior without fundamental CPT violation. Likewise, a very small fractional discrepancy is not automatically the strongest test in every theory: sensitivity depends on which symmetry-breaking parameter enters the measured observable."
    },
    {
      "heading": "CPT, CP violation, and the arrow of time",
      "text": "CPT invariance connects CP and T symmetry without making them the same transformation. If the complete dynamics obeys CPT and also obeys CP, it must obey T. Conversely, observed CP violation requires T violation when CPT holds. This inference concerns the exact dynamics; extracting it from a particular experimental asymmetry may additionally require control of initial-state preparation, final-state interactions, and detector effects.\n\nThe Standard Model accommodates CP violation through a complex phase in the Cabibbo–Kobayashi–Maskawa matrix, which describes charged-current mixing among quark flavors. This phase produces CP-violating effects in meson mixing and decay while preserving CPT. Consequently, the discovery of CP violation did not invalidate local relativistic quantum field theory. It constrained the permitted interactions within that framework.\n\nMicroscopic T violation should also be distinguished from the thermodynamic arrow of time. The second law of thermodynamics concerns the typical evolution of macroscopic systems from special low-entropy conditions. Irreversible diffusion, heat flow, and the production of records can occur in models whose microscopic dynamics is time-reversal invariant. Conversely, microscopic T violation does not by itself explain the large-scale growth of entropy or the observed difference between past and future boundary conditions.\n\nCPT likewise does not imply that the observable universe contains equal amounts of matter and antimatter. A symmetry of dynamical laws is not necessarily a symmetry of a particular state. The standard conditions identified by Andrei Sakharov in 1967 for baryogenesis are baryon-number violation, C and CP violation, and departure from thermal equilibrium. These conditions describe how an asymmetry can develop in conventional CPT-invariant settings; they do not require different proton and antiproton masses.\n\nThe equilibrium qualification is significant. Under the usual CPT-invariant assumptions and without an imposed conserved-charge asymmetry, thermal equilibrium does not generate a net baryon number. Departures from equilibrium permit processes and their inverses to occur with unequal effectiveness as the universe evolves. Alternative baryogenesis scenarios sometimes investigate CPT violation, but such mechanisms are distinct from the standard Sakharov framework.\n\nThe theorem therefore constrains relationships among laws and transition probabilities, not the existence of a particular cosmic history. A universe with a low-entropy past and a matter excess can be described by CPT-invariant equations even when its actual state is not invariant under CPT."
    },
    {
      "heading": "Scope, extensions, and possible violations",
      "text": "The theorem is powerful precisely because its assumptions are identifiable. A genuine CPT violation would indicate that at least one condition needed for the applicable theorem does not describe the system. Possible departures include nonlocal dynamics, failure of ordinary Lorentz covariance, or a modification of the standard Hilbert-space and time-evolution framework. A proposed violation is not meaningful without identifying which physical quantities differ and how the underlying theory avoids the theorem.\n\nThere is an important logical distinction between Lorentz violation and CPT violation. In standard local, unitary relativistic field-theoretic settings, CPT violation entails a failure of Lorentz invariance under the conditions of the relevant results. The converse does not hold: Lorentz symmetry can be violated while CPT remains intact. A preferred background tensor, for example, can select directions in spacetime through interactions that are CPT-even.\n\nA widely used framework for organizing searches is the Standard-Model Extension. Developed by Don Colladay and V. Alan Kostelecký in the 1990s, it supplements familiar field theories with coordinate-covariant interactions whose coefficients can encode violations of particle Lorentz symmetry. Some operators are CPT-odd and others CPT-even. For a Dirac field, an illustrative CPT-odd term has the form\n\n\\[\\mathcal{L}_b =-b_\\mu\\overline{\\psi}\\gamma_5\\gamma^\\mu\\psi,\\]\n\nwhere \\(b_\\mu\\) is a background coefficient. Which coefficient combinations are observable depends on the species and experimental arrangement.\n\nSuch terms can produce spin-dependent shifts or variations tied to laboratory orientation and velocity. Since a terrestrial apparatus rotates with Earth, precision observables may exhibit sidereal modulation relative to a fixed celestial frame. Comparing data across orientations and times can therefore reveal effects that a single static particle–antiparticle comparison would miss.\n\nThe familiar theorem is formulated in flat spacetime. In general relativity, a generic curved geometry need not possess a global map corresponding to simultaneous inversion of space and time. More general results in locally covariant quantum field theory relate theories on spacetimes with reversed time orientation under additional assumptions. These are more careful statements than simply declaring the flat-spacetime theorem unchanged in every gravitational setting.\n\nQuestions involving quantum gravity, nonlocality, or nonstandard quantum evolution remain dependent on the particular framework. The absence of a familiar local field description does not automatically imply CPT violation. Conversely, an observed particle–antiparticle discrepancy would require independent checks against environmental shifts, incomplete state matching, and conventional interaction effects before it could be interpreted as evidence that the theorem’s underlying framework had failed."
    }
  ],
  "links": [
    "quantum-field-theory",
    "special-relativity",
    "quantum-mechanics",
    "standard-model",
    "electric-charge",
    "electron",
    "positron",
    "proton",
    "photon",
    "parity-physics",
    "momentum",
    "angular-momentum",
    "electric-field",
    "magnetic-field",
    "weak-interaction",
    "spacetime",
    "quantum-entanglement",
    "hamiltonian-operator",
    "fermion",
    "spin-statistics-theorem",
    "beta-decay",
    "cp-violation",
    "mass",
    "electric-dipole",
    "scattering",
    "hydrogen",
    "spectroscopy",
    "neutrino",
    "second-law-of-thermodynamics",
    "entropy",
    "earth",
    "general-relativity",
    "quantum-gravity"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}