Perfect Fluid Stress-Energy

The perfect fluid is the workhorse matter model of general relativity — the source term you reach for whenever spacetime's matter is too complicated to track particle by particle but simple enough to summarize by what a comoving observer measures. Strip a continuum of every anisotropic transport phenomenon — no viscosity, no heat conduction, no shear stress — and what remains, in the local rest frame, is described by exactly two thermodynamic scalars: the energy density $\rho$ (rest-mass plus internal energy) and an isotropic pressure $p$ equal in every direction. That radical economy is the entire physical content of the word perfect, and it is why one tensor with two scalars models galaxies, starlight, neutron-star interiors, and the dark energy accelerating the cosmos. The covariant packaging is $T^{\mu\nu}=(\rho+p)u^\mu u^\nu + p\,g^{\mu\nu}$ , the most general symmetric rank-2 tensor buildable from the four-velocity $u^\mu$ and the metric $g^{\mu\nu}$ that is isotropic about the rest frame. Fed into the Einstein field equations $G_{\mu\nu}=8\pi G\,T_{\mu\nu}$ , it couples the fluid to curvature; and because the contracted Bianchi identity forces $\nabla_\mu G^{\mu\nu}=0$ , the field equations automatically impose $\nabla_\mu T^{\mu\nu}=0$ — the fluid's equations of motion are consequences of geometry, not independent postulates. Projecting that conservation law along and across the flow delivers the relativistic continuity and Euler equations. Supply an equation of state $p=p(\rho)$ and the system closes; the perfect fluid then underlies the Friedmann equations, the Tolman–Oppenheimer–Volkoff equation, and the energy conditions of the singularity theorems.

Key takeaways

A perfect fluid is matter that, in its local rest frame, is fully characterized by just two scalars — energy density $\rho$ and isotropic pressure $p$ — with no shear stress, no viscosity, and no heat conduction.
Its stress-energy tensor is $T^{\mu\nu}=(\rho+p)u^\mu u^\nu + p\,g^{\mu\nu}$ in the $(-,+,+,+)$ signature, the unique symmetric tensor built from $u^\mu$ ( $u^\mu u_\mu=-1$ ) and $g^{\mu\nu}$ that is isotropic about the rest frame.
In the comoving frame $T^\mu{}_\nu=\mathrm{diag}(-\rho,p,p,p)$ with trace $T=-\rho+3p$ ; the pure-diagonal spatial block is the mathematical face of "perfect".
Local conservation $\nabla_\mu T^{\mu\nu}=0$ splits, along and across $u^\mu$ , into the relativistic continuity and Euler equations — in which the inertia is $\rho+p$ , not $\rho$ , because pressure gravitates.
An equation of state $p=w\rho$ closes the system and specializes it: dust ( $w=0$ ), radiation ( $w=1/3$ , traceless), stiff fluid ( $w=1$ ), and the cosmological constant ( $w=-1$ ); cosmic densities dilute as $\rho\propto a^{-3(1+w)}$ .
Key idea: A perfect fluid is fully described by the scalars $\rho$ (rest-frame energy density) and $p$ (isotropic pressure) plus the four-velocity $u^\mu$ ; its stress-energy $T^{\mu\nu}=(\rho+p)u^\mu u^\nu + p\,g^{\mu\nu}$ is the unique symmetric tensor built from these that is isotropic in the rest frame, and $\nabla_\mu T^{\mu\nu}=0$ is its complete dynamics.

Where this lives today

The perfect fluid is no mere idealization — it is the source term in the equations that fit modern data to many decimals. In cosmology, the ΛCDM concordance model treats the universe as a mixture of non-interacting perfect fluids — cold dark matter, baryons, photons, neutrinos, and a cosmological constant — each with its own $w$ , and the dilution law $\rho\propto a^{-3(1+w)}$ lets Planck CMB and DESI surveys back out the cosmic energy budget and test whether dark energy is really $w=-1$ . In compact-object astrophysics, the Tolman–Oppenheimer–Volkoff equation with a perfect-fluid interior underlies every neutron-star structure calculation, and the 2017 merger GW170817 constrained the dense-matter equation of state $p(\rho)$ feeding that model. And at the frontier, the sound speed $c_s^2=\mathrm{d}p/\mathrm{d}\rho$ and the energy conditions the perfect fluid makes transparent are the tools for asking whether dark energy could be phantom ( $w<-1$ ), whether inflation violates the strong energy condition, and whether exotic matter could thread a wormhole.

The physical picture

Sit inside a tiny parcel of the fluid and ride along with it — the comoving, or rest, frame. A perfect fluid is one in which, from this vantage, the world looks identical in every direction: the only force per unit area on any surface is a pressure $p$ pushing outward, equal on every face, with no sideways shear and no heat crossing. That complete rest-frame isotropy is the whole meaning of perfect, and it forces the mixed stress-energy into a strikingly simple diagonal form — energy density along the time axis, one equal pressure on each spatial axis, nothing off the diagonal, as the first figure shows.

Figure

In the comoving rest frame a perfect fluid is isotropic: energy density ρ acts along the time axis while an equal pressure p pushes outward on every spatial axis, with no shear and no heat flux. The mixed tensor is therefore T^μ_ν = diag(−ρ, p, p, p) and its trace is T = −ρ + 3p

The tensor is not a guess — it is forced. The only objects on hand to build a symmetric two-index tensor are $u^\mu$ , which singles out the time direction, and $g^{\mu\nu}$ : the product $u^\mu u^\nu$ carries the energy density along the flow, while $h^{\mu\nu}\equiv g^{\mu\nu}+u^\mu u^\nu$ projects onto the three spatial directions and carries the isotropic pressure. Any anisotropy would need another tensor and break the rest-frame isotropy that defines the model. Once an equation of state is chosen, the fluid evolves in a way that depends only on $w=p/\rho$ , as the second figure traces for the real universe: three cosmic fluids diluting at their own rates and crossing at the epochs that organize cosmic history.

Figure

Cosmic fluids dilute at rates fixed by their equation of state w = p/ρ, since ρ ∝ a^(−3(1+w)): radiation (w = 1/3) falls as a^−4, matter (w = 0) as a^−3, and dark energy (w = −1) stays constant. With Ω_r ≈ 9×10⁻⁵ and Ω_m ≈ 0.315 the radiation and matter curves cross at matter–radiation equality, 1 + z ≈ 3400, and matter meets dark energy near z ≈ 0.3

The core equations

The defining object is the stress-energy tensor itself, in the mostly-plus signature $(-,+,+,+)$ used throughout, assembled from the four-velocity $u^\mu$ ( $u^\mu u_\mu=-1$ ), the metric, and the two scalars $\rho$ and $p$ . Every equation below is a projection or specialization of this one line.

T^{\mu\nu} = (\rho+p)\,u^\mu u^\nu + p\,g^{\mu\nu} \qquad (1)

where: $\rho$ — rest-frame total energy density (rest mass + internal energy); $p$ — isotropic rest-frame pressure; $u^\mu$ — fluid four-velocity, $u^\mu u_\mu=-1$ ; $g^{\mu\nu}$ — inverse metric, signature $(-,+,+,+)$ . The first term carries energy along the flow, the second spreads pressure isotropically. The same tensor is clearer with the pressure written through the spatial projector $h^{\mu\nu}$ , which makes the energy/pressure split manifest and is the language for the projections that follow.

T^{\mu\nu} = \rho\,u^\mu u^\nu + p\,h^{\mu\nu}, \qquad h^{\mu\nu}\equiv g^{\mu\nu}+u^\mu u^\nu \qquad (2)

where: $h^{\mu\nu}$ — projector onto the 3-space orthogonal to $u^\mu$ , with $h^{\mu\nu}u_\nu=0$ and $h^\mu{}_\mu=3$ ; it is the rest-frame spatial metric, so $p\,h^{\mu\nu}$ is equal pressure on every spatial axis. Lowering one index and evaluating in the comoving frame ( $u^\mu=(1,0,0,0)$ , $u_\mu=(-1,0,0,0)$ ) exposes the diagonal structure that names the model.

T^\mu{}_\nu = \mathrm{diag}(-\rho,\,p,\,p,\,p) \qquad (3)

where: the timelike entry $-\rho$ carries the lowered-index sign; the three equal spatial entries $p$ are the isotropic pressure. Reading it: no off-diagonal terms means no shear and no heat flux — the signature of a perfect fluid. Contracting the mixed tensor gives a Lorentz-invariant scalar, the trace — the cleanest diagnostic of an equation of state, vanishing precisely for radiation.

T \equiv T^\mu{}_\mu = -\rho + 3p \qquad (4)

where: $T$ — scalar trace $=g_{\mu\nu}T^{\mu\nu}$ . Reading it: $T=0$ demands $p=\rho/3$ (radiation); $T=-\rho$ for dust; $T=-4\rho$ for a cosmological constant. Coupling to gravity is automatic: because the Einstein tensor is divergence-free by the contracted Bianchi identity, the field equations $G_{\mu\nu}=8\pi G\,T_{\mu\nu}$ force the fluid's stress-energy to be conserved — one covariant statement that contains all of the fluid's dynamics.

\nabla_\mu T^{\mu\nu} = 0 \qquad (5)

where: $\nabla_\mu$ — metric-compatible (Levi-Civita) covariant derivative. The fluid has no independent equation of motion — its dynamics are dictated by the geometry it sources. Projecting the conservation law along the flow (contracting with $u_\nu$ ) yields the relativistic continuity, or energy, equation — mass conservation generalized to include the work pressure does during expansion.

u^\mu\nabla_\mu\rho + (\rho+p)\,\nabla_\mu u^\mu = 0 \qquad (6)

where: $\nabla_\mu u^\mu\equiv\theta$ — the expansion scalar, the fractional rate of change of a comoving volume. Energy density falls both from dilution ( $\rho\,\theta$ ) and from the $p\,\theta$ work pressure does on it. Projecting orthogonal to the flow (with $h^\nu{}_\sigma$ ) yields the relativistic Euler equation, whose defining feature is the inertial factor $\rho+p$ : pressure contributes to inertia.

(\rho+p)\,a^\nu = -\,h^{\nu\mu}\nabla_\mu p \qquad (7)

where: $a^\nu=u^\mu\nabla_\mu u^\nu$ — the four-acceleration; $h^{\nu\mu}\nabla_\mu p$ — the pressure gradient projected into the rest-frame 3-space. Reading it: the gradient accelerates the enthalpy density $\rho+p$ , not just $\rho$ — a genuinely relativistic effect. One relation is still missing to close the system — an equation of state. The adiabatic sound speed is the derivative that closes the equations and enforces causality.

c_s^2 = \left(\frac{\partial p}{\partial\rho}\right)_s \qquad (8)

where: $c_s$ — speed of sound; $s$ — specific entropy held fixed; causality requires $c_s\le c$ . Reading it: a barotropic $p=p(\rho)$ fixes $c_s$ ; radiation gives $c_s=c/\sqrt3$ , and $p=\rho$ saturates the causal bound $c_s=c$ .

Derivation AWhy the perfect-fluid tensor is the unique isotropic form

The goal here is to show that $T^{\mu\nu}=(\rho+p)u^\mu u^\nu+p\,g^{\mu\nu}$ is not an ansatz to memorize but the only symmetric two-index tensor buildable from the four-velocity $u^\mu$ and the metric $g^{\mu\nu}$ that respects complete isotropy about the rest frame — so the entire physical assumption of "perfect" is encoded in a counting argument about which tensors are available. We build it from the ground up and check it reproduces the diagonal rest-frame components.

Step 1. List the building blocks. A symmetric rank-2 tensor made from the data of a fluid at a point can only use $u^\mu$ , the metric $g^{\mu\nu}$ , and Lorentz scalars, so the two available symmetric tensors are $u^\mu u^\nu$ and $g^{\mu\nu}$ : the most general form is $T^{\mu\nu}=A\,u^\mu u^\nu + B\,g^{\mu\nu}$ . Why: any additional term would need a new object — a heat-flux $q^\mu$ or an anisotropic-stress $\pi^{\mu\nu}$ — exactly what a perfect fluid forbids.

Step 2. Decompose $g^{\mu\nu}=h^{\mu\nu}-u^\mu u^\nu$ with the projector $h^{\mu\nu}=g^{\mu\nu}+u^\mu u^\nu$ . Substituting, $T^{\mu\nu}=(A-B)u^\mu u^\nu + B\,h^{\mu\nu}$ : the first piece lives along the flow, the second in the orthogonal 3-space. Why: this separates the time-time energy content from the space-space stress content, the split isotropy demands.

Step 3. Read off the scalars in the rest frame. Contracting with $u_\mu u_\nu$ (using $u_\mu u^\mu=-1$ , $h^{\mu\nu}u_\nu=0$ ) gives $T^{\mu\nu}u_\mu u_\nu=A-B\equiv\rho$ ; the spatial trace gives $B\,h^\mu{}_\mu=3B\equiv3p$ , so $B\equiv p$ . Why: $\rho$ is the comoving energy density and $p$ is the single pressure isotropy allows on the three spatial axes.

Step 4. Reassemble. With $A-B=\rho$ and $B=p$ we get $A=\rho+p$ , hence $T^{\mu\nu}=(\rho+p)u^\mu u^\nu + p\,g^{\mu\nu}$ , which in the rest frame gives $T^\mu{}_\nu=\mathrm{diag}(-\rho,p,p,p)$ , matching equation (3). Why: construction and component check agree, so isotropy alone fixes the tensor up to the two scalars $\rho$ and $p$ .

Go deeper · Derivation B · 1 minThe conservation law splits into continuity and relativistic Euler

Derivation BThe conservation law splits into continuity and relativistic Euler

The aim is to show that the single statement $\nabla_\mu T^{\mu\nu}=0$ contains the entire dynamics of the fluid: projecting it along and orthogonal to the four-velocity separates it into the relativistic continuity (energy) equation and the relativistic Euler equation, whose Newtonian limits are ordinary hydrodynamics. Throughout we use $u^\mu u_\mu=-1$ , which differentiates to $u_\nu\nabla_\mu u^\nu=0$ .

Step 1. Take the divergence of equation (1). Since $\nabla_\mu g^{\mu\nu}=0$ (metric compatibility), $\nabla_\mu(p\,g^{\mu\nu})=\nabla^\nu p$ , so $\nabla_\mu T^{\mu\nu}=\nabla_\mu[(\rho+p)u^\mu]\,u^\nu+(\rho+p)u^\mu\nabla_\mu u^\nu+\nabla^\nu p=0$ . Why: separating a scalar-times- $u^\nu$ piece from an acceleration piece lets the two projections pick out one equation each.

Step 2. Project along the flow by contracting with $u_\nu$ . Using $u_\nu u^\nu=-1$ and $u_\nu\nabla_\mu u^\nu=0$ , the acceleration term dies, leaving $-\nabla_\mu[(\rho+p)u^\mu]+u^\mu\nabla_\mu p=0$ ; expanding, the $u^\mu\nabla_\mu p$ pieces cancel to give $u^\mu\nabla_\mu\rho+(\rho+p)\nabla_\mu u^\mu=0$ — equation (6). Why: this is energy conservation for comoving observers, with pressure entering only through the $p\,\theta$ work term, the first law $\mathrm{d}(\rho V)=-p\,\mathrm{d}V$ .

Step 3. Project orthogonal to the flow with $h^\nu{}_\sigma=\delta^\nu_\sigma+u^\nu u_\sigma$ . The along-flow terms are annihilated ( $h^\nu{}_\sigma u^\sigma=0$ ), leaving $(\rho+p)u^\mu\nabla_\mu u^\nu + h^{\nu\mu}\nabla_\mu p=0$ — the relativistic Euler equation (7). Why: this is the spatial force balance in the rest frame, with the enthalpy density $\rho+p$ as inertial mass.

Step 4. Take the Newtonian limit: $p\ll\rho$ , $u^i\to v^i$ , $u^0\to1$ , $\nabla_\mu\to\partial_\mu$ plus a weak potential $\Phi$ . Equation (6) becomes $\partial_t\rho+\nabla\!\cdot(\rho\mathbf v)=0$ and equation (7) becomes $\rho(\partial_t\mathbf v+\mathbf v\!\cdot\!\nabla\mathbf v)=-\nabla p-\rho\nabla\Phi$ . Why: recovering classical hydrodynamics confirms the relativistic forms, and shows the $(\rho+p)$ inertia stays invisible until $p\sim\rho$ .

Go deeper · Derivation C · 1 minThe Friedmann fluid equation and cosmic dilution

Derivation CThe Friedmann fluid equation and cosmic dilution

The objective is to feed the perfect fluid into a homogeneous, isotropic (FLRW) universe and show the parallel projection of $\nabla_\mu T^{\mu\nu}=0$ becomes the energy-conservation equation $\dot\rho+3H(\rho+p)=0$ , then integrate it with $p=w\rho$ to obtain the dilution law $\rho\propto a^{-3(1+w)}$ governing every component of the cosmic energy budget. We work in comoving coordinates with $u^\mu=(1,0,0,0)$ and spatial metric scaling as $a(t)$ .

Step 1. Compute the expansion scalar. For spatially flat FLRW, $\sqrt{-g}\propto a^3$ , so $\nabla_\mu u^\mu=\frac{1}{\sqrt{-g}}\partial_\mu(\sqrt{-g}\,u^\mu)=\frac{1}{a^3}\partial_t(a^3)=3\dot a/a\equiv3H$ . Why: the expansion scalar is just the fractional growth rate of a comoving volume, and $H=\dot a/a$ is the Hubble rate — the only kinematic quantity a homogeneous flow carries.

Step 2. Insert into the continuity equation (6). With $u^\mu\nabla_\mu\rho=\dot\rho$ and $\nabla_\mu u^\mu=3H$ , equation (6) becomes $\dot\rho+3H(\rho+p)=0$ — exactly the first law $\mathrm{d}(\rho a^3)=-p\,\mathrm{d}(a^3)$ for adiabatic expansion of a comoving volume $\propto a^3$ . Why: cosmic expansion does $p\,\mathrm{d}V$ work, so pressure — not only dilution — controls how energy density falls.

Step 3. Close with $p=w\rho$ (constant $w$ ). Then $\dot\rho/\rho=-3(1+w)\dot a/a$ , which integrates to $\rho\propto a^{-3(1+w)}$ . Why: a constant equation-of-state parameter turns the conservation law into a power law, so each species dilutes at a rate set entirely by its $w$ .

Step 4. Read off the components: dust ( $w=0$ ) gives $\rho\propto a^{-3}$ ; radiation ( $w=1/3$ ) gives $\rho\propto a^{-4}$ (an extra factor of $a$ from redshifting each quantum); vacuum ( $w=-1$ ) gives $\rho=\text{const}$ ; a stiff fluid ( $w=1$ ) gives $\rho\propto a^{-6}$ . Why: these four laws are the backbone of the dilution figure and the entire thermal history — radiation first, then matter, then dark energy.

Limiting cases

Choosing an equation of state $p=p(\rho)$ specializes the one perfect-fluid tensor into physically distinct media — each row is the same $T^{\mu\nu}$ with a different relation between $p$ and $\rho$ .

Case	Result	Meaning
Pressureless dust	$p = 0$	Galaxies, cold dark matter, any non-interacting cold matter. $T^{\mu\nu}=\rho\,u^\mu u^\nu$ ; fluid elements follow timelike geodesics, and $\rho\propto a^{-3}$ .
Radiation / ultrarelativistic gas	$p=\tfrac13\rho$	Photons, neutrinos, any relativistic species. Traceless ( $T=-\rho+3p=0$ , conformal invariance); $\rho\propto a^{-4}$ ; sound speed $c_s=c/\sqrt3$ .
Vacuum / cosmological constant	$p=-\rho$	$T^{\mu\nu}=-\rho\,g^{\mu\nu}$ is boost-invariant (same to every observer); $\rho_\Lambda=\Lambda/(8\pi G)$ , $\rho=\text{const}$ , drives acceleration and violates the strong energy condition.
Stiff fluid	$p=\rho$	Sound speed equals light speed, $c_s^2=\mathrm{d}p/\mathrm{d}\rho=1$ — the causal ceiling for a barotropic fluid (Zel'dovich); $\rho\propto a^{-6}$ .
Newtonian limit	$p\ll\rho$ , $v\ll c$	Continuity and Euler reduce to classical hydrodynamics; $\rho\to$ rest-mass density and pressure drops out of the inertia.

Common trapPressure is just a force per area, so it cannot affect gravity or inertia

It does both, and forgetting this is the classic perfect-fluid trap. In Newtonian gravity only mass density sources the field, so pressure looks dynamically passive. In general relativity the source is the full stress-energy tensor, with pressure on its diagonal, and two consequences follow. First, gravity: the relativistic analogue of the Poisson source is $\rho+3p$ , so large positive pressure adds to attraction — which is why a sufficiently pressurized star still collapses, giving a maximum neutron-star mass with no Newtonian counterpart. Second, inertia: the relativistic Euler equation accelerates the enthalpy density $\rho+p$ , not $\rho$ alone. Both effects vanish when $p\ll\rho$ and dominate when $p\sim\rho$ . And the flip side: negative pressure with $\rho+3p<0$ produces gravitational repulsion — exactly how a cosmological constant accelerates the universe.

Worked examples

1 · Matter–radiation equality from cosmic dilution (fully worked)

The present cosmic energy budget has a radiation fraction $\Omega_{r,0}\approx9.0\times10^{-5}$ and a matter fraction $\Omega_{m,0}\approx0.315$ (in units of the critical density). Radiation dilutes as $a^{-4}$ and matter as $a^{-3}$ . At what scale factor, and redshift, were the two densities equal?

Step 1. Write each density using the dilution law from Derivation 3, normalized to today ( $a_0=1$ ): $\rho_r(a)=\rho_{r,0}\,a^{-4}$ and $\rho_m(a)=\rho_{m,0}\,a^{-3}$ , with $\rho_{r,0}\propto\Omega_{r,0}$ , $\rho_{m,0}\propto\Omega_{m,0}$ in critical-density units.

Step 2. Set them equal: $\Omega_{r,0}\,a_\text{eq}^{-4}=\Omega_{m,0}\,a_\text{eq}^{-3}$ . One power of $a$ cancels, giving $a_\text{eq}=\Omega_{r,0}/\Omega_{m,0}=9.0\times10^{-5}/0.315=\mathbf{2.86\times10^{-4}}$ .

Step 3. Convert to redshift via $1+z=1/a$ : $1+z_\text{eq}=\Omega_{m,0}/\Omega_{r,0}=0.315/(9.0\times10^{-5})=\mathbf{3500}$ . Radiation falls one power of $a$ faster than matter, so the universe was radiation-dominated before this epoch and matter-dominated after. Equality at a ≈ 2.9×10⁻⁴, redshift 1 + z ≈ 3500, matching the observed z_eq ≈ 3400.

2 · The radiation fluid: tracelessness and sound speed

Show that a gas of radiation ( $p=\rho/3$ ) has a traceless stress-energy tensor, and compute its adiabatic sound speed in SI units.

Step 1. The trace from equation (4) is $T=-\rho+3p$ . Substituting $p=\rho/3$ gives $T=-\rho+3(\rho/3)=-\rho+\rho=\mathbf{0}$ . The vanishing trace is the imprint of the conformal invariance of the electromagnetic field.

Step 2. The adiabatic sound speed from equation (8) is $c_s^2=(\partial p/\partial\rho)_s$ . For $p=\rho/3$ this is $c_s^2=1/3$ (units $c=1$ ), so $c_s=1/\sqrt3\approx\mathbf{0.577\,c}$ .

Step 3. In SI units, $c_s=c/\sqrt3=(2.998\times10^8)/1.732=\mathbf{1.73\times10^8\ m/s}$ . This is the speed of acoustic oscillations in the primordial photon–baryon plasma, whose travel before recombination sets the sound horizon — the standard ruler behind baryon acoustic oscillations. Radiation is traceless (T = 0) with sound speed c/√3 ≈ 1.73×10⁸ m/s.

3 · The cosmological constant as a boost-invariant fluid

Take $p=-\rho$ and show the resulting stress-energy looks identical to every observer, then determine which energy conditions it violates.

Step 1. Substitute $p=-\rho$ into equation (1): $T^{\mu\nu}=(\rho+(-\rho))u^\mu u^\nu+(-\rho)g^{\mu\nu}=\mathbf{-\rho\,g^{\mu\nu}}$ . The four-velocity has dropped out entirely.

Step 2. Because $T^{\mu\nu}\propto g^{\mu\nu}$ , any Lorentz boost leaves it invariant: the fluid looks identical to every observer, with no rest frame and no flow. Identifying $\rho_\Lambda=\Lambda/(8\pi G)$ shows the cosmological constant is precisely a perfect fluid with $w=-1$ , the simplest dark energy.

Step 3. Test the energy conditions with $p=-\rho$ , $\rho>0$ . Weak/null need $\rho+p\ge0$ : here $\rho+p=0$ , marginally satisfied. Strong needs $\rho+3p\ge0$ : here $\rho+3p=\rho-3\rho=\mathbf{-2\rho<0}$ , violated. Dominant needs $\rho\ge|p|$ : here $|p|=\rho$ , marginally satisfied. The cosmological constant violates only the strong energy condition (ρ + 3p = −2ρ < 0), which is exactly why it produces gravitational repulsion and accelerates the universe.

Physics of the portrait

The perfect fluid is where relativistic gravity, thermodynamics, and astrophysics meet — and no one did more to fuse those threads than the chemist who turned Einstein's equations into a thermodynamics of matter in curving spacetime.

Origins · 1 minRichard Tolman and how this was discovered

Richard Tolman

1881–1948, American physical chemist and mathematical physicist · relativistic thermodynamics and cosmology of perfect fluids

Richard Chace Tolman began as a physical chemist — his early fame came from statistical mechanics — before turning, in the 1920s and 1930s, to the young theory of general relativity. His 1934 treatise Relativity, Thermodynamics and Cosmology taught a generation how to treat matter and radiation as relativistic fluids: how energy density and pressure enter the stress-energy tensor, how the first law reads for a comoving volume in an expanding universe, and how an equation of state closes the Einstein equations. In 1939, with J. Robert Oppenheimer and George Volkoff, he combined the perfect-fluid source with static spherical symmetry to derive the hydrostatic-equilibrium equation now bearing their initials — the Tolman–Oppenheimer–Volkoff equation — and discovered that relativity imposes a maximum mass on a cold, degenerate star. That result, resting entirely on the perfect-fluid stress-energy and the way pressure gravitates, is the theoretical seed of neutron stars and black holes.

History

1915

— Einstein completes the field equations Gμν = 8πG Tμν, making matter the source of curvature

1922

— Friedmann models an expanding universe with a homogeneous perfect-fluid source

1934

— Tolman publishes Relativity, Thermodynamics and Cosmology, the relativistic thermodynamics of fluids

1939

— Tolman, Oppenheimer and Volkoff derive the equation of relativistic stellar structure

1965

— Penrose's singularity theorem turns energy conditions on Tμν into statements about collapse

Going further

Texts. Misner, Thorne & Wheeler, Gravitation (1973), §5.5, §22.2 (note MTW uses mostly-minus, flipping the sign of the $p\,g^{\mu\nu}$ term) · Wald, General Relativity (1984), §4.3, §6.2 · Carroll, Spacetime and Geometry (2004), §1.9, §4.5, §8.3 · Schutz, A First Course in General Relativity (2009), §4.6, §7.2 · Hartle, Gravity (2003), §22.2 · Tolman, Relativity, Thermodynamics and Cosmology (1934). The key results of this article are: the rest-frame components $\mathrm{diag}(-\rho,p,p,p)$ from equation (1) with $u^\mu=(1,0,0,0)$ ; the trace $-\rho+3p$ ; the parallel and orthogonal projections of $\nabla_\mu T^{\mu\nu}=0$ ; the FLRW dilution $\rho\propto a^{-3(1+w)}$ ; matter–radiation equality $1+z_\text{eq}=\Omega_m/\Omega_r\approx3500$ ; the radiation sound speed $c_s=c/\sqrt3\approx1.73\times10^8\ \mathrm{m/s}$ ; and the energy-condition signs for $p=-\rho$ . The equation-of-state values and energy-condition inequalities are standard textbook results; $\Omega_{m,0}\approx0.315$ and $\Omega_{r,0}\approx9\times10^{-5}$ are Planck-consistent.

Prerequisites. The Stress-Energy Tensor · Local Conservation of Energy-Momentum · The Einstein Field Equations. Leads to. The Cosmological Constant · The Energy Conditions · The Newtonian Limit.

Ask the tutor: Ask me to derive the Tolman–Oppenheimer–Volkoff equation from this stress-energy, or to show step by step how $\rho\propto a^{-3(1+w)}$ builds the thermal history of the universe.

What is the difference between the energy density ρ and the rest-mass density?

They coincide only for cold dust. The scalar $\rho$ is the total rest-frame energy density: rest-mass energy plus all internal energy — thermal motion, radiation, binding energy. The rest-mass density $\rho_0$ counts only the constituent particles' mass. For a cold gas $\rho\approx\rho_0$ ; a hot or photon gas differs — a photon gas has $\rho_0=0$ yet $\rho>0$ . This is why $\rho$ , not $\rho_0$ , gravitates: in relativity all forms of energy weigh.

Why does pressure gravitate in general relativity but not in Newtonian gravity?

In Newtonian gravity the field is sourced by mass density alone, so pressure — a momentum flux — has no gravitational role. In general relativity the source is the entire stress-energy tensor, and the relativistic analogue of the Poisson source for a static fluid is $\rho+3p$ , not $\rho$ , so positive pressure adds to gravity. Inside a neutron star, where $p$ approaches $\rho$ , the extra self-gravity from pressure makes collapse unstoppable above a critical mass — the Tolman–Oppenheimer–Volkoff maximum mass that Newtonian theory cannot produce.

Does the sign of the pressure term depend on the metric signature?

Yes — the most common source of sign errors. In the mostly-plus signature $(-,+,+,+)$ used here, $u^\mu u_\mu=-1$ and $T^{\mu\nu}=(\rho+p)u^\mu u^\nu+p\,g^{\mu\nu}$ , giving $T^{00}=\rho$ in the rest frame. In MTW's mostly-minus convention $(+,-,-,-)$ , $u^\mu u_\mu=+1$ and the same physics reads $T^{\mu\nu}=(\rho+p)u^\mu u^\nu-p\,g^{\mu\nu}$ : the components $\rho$ and $p$ are identical, only the sign in front of $p\,g^{\mu\nu}$ flips. Always confirm the signature before trusting a sign.

In what sense is dark energy "really" a perfect fluid?

A cosmological constant is exactly the $p=-\rho$ member of the family. Substituting $p=-\rho$ collapses the tensor to $T^{\mu\nu}=-\rho\,g^{\mu\nu}$ , invariant under Lorentz boosts — no rest frame, the same to every observer, precisely the property vacuum energy must have. So the simplest dark energy is a perfect fluid with $w=-1$ and constant density. More general "quintessence" has a slowly varying $w(a)$ near $-1$ , and cosmology's central question is whether the data prefer exactly $w=-1$ or a dynamical equation of state.

What does an equation of state do, and why is one needed?

The Einstein equations plus $\nabla_\mu T^{\mu\nu}=0$ relate $\rho$ , $p$ , and $u^\mu$ , but fall one relation short of closure: two unknown scalars with the conservation law constraining only their combination. An equation of state $p=p(\rho)$ (or $p=p(\rho,s)$ ) supplies the missing thermodynamic information about what the fluid is made of — cold matter, hot radiation, degenerate neutrons, vacuum — and only with it can you solve for the actual evolution: cosmic dilution, a star's pressure profile, the sound speed of perturbations. It is where microphysics enters an otherwise purely geometric framework.

Can a perfect fluid ever carry heat flow or viscosity?

No — by definition. Allowing heat conduction forces a heat-flux vector $q^\mu$ into the stress-energy, and viscosity introduces an anisotropic-stress tensor $\pi^{\mu\nu}$ ; both break the rest-frame isotropy that makes the fluid "perfect", and both are exactly what Derivation 1's uniqueness argument excludes. Real fluids have them, so the perfect fluid is an idealization — excellent when mean free paths are short (cosmic plasma, stellar interiors), poor when transport matters (dissipative accretion, supernova heating), where imperfect-fluid or Israel–Stewart extensions take over and reduce to it when the dissipative coefficients vanish.