Perfect Fluid Stress-Energy

Every source of gravity in general relativity enters through one object: the stress-energy tensor $T^{\mu\nu}$ , the relativistic ledger that bundles energy density, momentum density, energy flux, and mechanical stress into a single symmetric array. The Einstein field equations set the geometry of spacetime equal to this ledger, $G^{\mu\nu}=8\pi G\,T^{\mu\nu}$ , so whatever matter you pour into a spacetime, its gravity is fixed by how it fills $T^{\mu\nu}$ . The trouble is that real matter is complicated — it flows, shears, conducts heat, and radiates — and writing its full stress-energy is hopeless. The perfect fluid is the idealization that cuts through this: strip away every dissipative effect and a continuous medium is described by just two numbers measured by an observer riding along with it, the rest-frame energy density $\rho$ and an isotropic pressure $p$ . That radical economy is not an assumption bolted on for convenience; it is forced. Building on the four-velocity, the Minkowski metric, and proper time, one asks for the most general symmetric rank-2 tensor that looks isotropic in the local rest frame — no preferred spatial direction, no shear, no heat flow. The answer is unique and has exactly two free scalars, giving $T^{\mu\nu}=(\rho+p)\,u^\mu u^\nu + p\,g^{\mu\nu}$ . Feed it to Einstein and demand consistency, and the contracted Bianchi identity hands you $\nabla_\mu T^{\mu\nu}=0$ for free: local conservation of energy and the relativistic Euler equation, the dynamics of the fluid. Supply one more ingredient — an equation of state $p=p(\rho)$ — and the system closes. This single tensor is the engine of Friedmann cosmology, stellar structure, and relativistic astrophysics.

Key takeaways

A perfect fluid is a continuous medium with no viscosity, no shear stress, and no heat conduction, described entirely by two rest-frame scalars: energy density $\rho$ and isotropic pressure $p$ .
Local isotropy makes its stress-energy tensor unique: $T^{\mu\nu}=(\rho+p)\,u^\mu u^\nu + p\,g^{\mu\nu}$ , which reads $\mathrm{diag}(\rho,p,p,p)$ in the rest frame (signature $-+++$ , $c=1$ ).
The combination $(\rho+p)$ is the enthalpy density — the true relativistic inertia of a fluid element, so pressure resists acceleration and gravitates; the trace is $T^\mu{}_\mu=-\rho+3p$ .
Conservation $\nabla_\mu T^{\mu\nu}=0$ splits into the relativistic continuity equation and the relativistic Euler equation, recovering Newtonian fluid dynamics in the low-speed, low-pressure limit.
An equation of state $p=w\rho$ selects the regime — dust ( $w=0$ ), radiation ( $w=1/3$ ), stiff matter ( $w=1$ ), vacuum energy ( $w=-1$ ) — and fixes the cosmological dilution $\rho\propto a^{-3(1+w)}$ .
Key idea: isotropy in the rest frame plus the two scalars $\rho$ and $p$ uniquely fix $T^{\mu\nu}=(\rho+p)u^\mu u^\nu + p\,g^{\mu\nu}$ , and its conservation is the relativistic continuity and Euler equations; the enthalpy $(\rho+p)$ and the source $(\rho+3p)$ are where relativity departs from Newton.

The core insight

Isotropy in the rest frame plus the two scalars $\rho$ and $p$ uniquely fix the stress-energy tensor to $T^{\mu\nu}=(\rho+p)u^\mu u^\nu + p\,g^{\mu\nu}$ ; its conservation gives the relativistic continuity and Euler equations, and its enthalpy $(\rho+p)$ and source combination $(\rho+3p)$ are where pressure acquires inertia and weight.

Where this lives today

The perfect fluid is the matter model behind the precision cosmology of the 2020s. When the Planck satellite, and now surveys like DESI, fit the cosmic microwave background and the clustering of galaxies, they are fitting the density and pressure of a small set of perfect fluids — baryons, cold dark matter, photons, neutrinos, and dark energy — each with its own equation of state $w$ , evolving by the very scaling law $\rho\propto a^{-3(1+w)}$ derived above. The headline question is whether dark energy is a pure cosmological constant ( $w=-1$ exactly) or a fluid with $w$ drifting in time; DESI's 2024–2025 baryon-acoustic-oscillation results hint at an evolving $w$ , which if confirmed would mean the vacuum is not constant after all. At the other extreme of scale, the same tensor with a nuclear equation of state feeds the Tolman-Oppenheimer-Volkoff equation to predict neutron-star radii, now measured directly: NASA's NICER telescope times X-ray hot spots to pin down the mass-radius relation, and the LIGO-Virgo detection of the merging neutron stars GW170817 constrained the fluid's stiffness through the stars' tidal deformability. And in the laboratory, the quark-gluon plasma created at RHIC and the LHC behaves as one of the most nearly perfect fluids known — a droplet of near-ideal relativistic fluid a trillion times hotter than the Sun's core.

The physical picture

Ride along with a small parcel of gas — sit in the frame where it is momentarily at rest. Look around: there is no preferred direction. The parcel holds an energy density $\rho$ , and it pushes on its neighbours equally hard in every direction with a pressure $p$ . "Equally hard in every direction" is the whole physical content of the word perfect: the fluid is isotropic and dissipationless in its rest frame. Isotropy forbids any shear stress (a force parallel to a face, which would single out a direction), and dissipationlessness forbids heat flux (energy leaking across faces). What survives is only energy density on the time slot and equal pressures on the three spatial diagonals — the tensor collapses to $\mathrm{diag}(\rho,p,p,p)$ , two numbers instead of ten.

Figure

A comoving fluid element pushes equally hard in every direction: isotropy in the rest frame kills all shear and heat-flux terms, leaving only the energy density ρ on the time-time slot and equal pressures p on the three spatial diagonals, so the stress-energy tensor is diag ρ p p p

Two features of this simple object have no Newtonian counterpart, and both come from the fact that in relativity pressure is energy per volume and energy has inertia and weight. First, if you boost the fluid, its momentum density is $T^{0i}=(\rho+p)\,\gamma^2 v^i$ — the pressure adds to the flow of momentum even though in the rest frame it was pure stress. The effective inertia of a fluid element is therefore $(\rho+p)$ , the enthalpy density, not $\rho$ alone. Second, the fluid gravitates through the combination $\rho+3p$ , not $\rho$ : pressure has weight. A gas held at high pressure is harder to support against its own gravity than a naive count of its energy would suggest, and a medium with sufficiently negative pressure gravitates repulsively. That sign flip, $\rho+3p<0$ , is exactly the condition for cosmic acceleration, and it is why the equation of state — the relation between $p$ and $\rho$ — is the single most consequential input in cosmology.

Figure

Cosmic energy budget from the perfect-fluid conservation law: radiation with w = 1/3 dilutes as a to the minus four, matter with w = 0 as a to the minus three, and vacuum energy with w = minus one stays constant, so the universe passes from radiation to matter domination near redshift 3400 and into dark-energy domination near redshift 0.3

The core equations

The whole subject is the unpacking of one tensor. It is assembled from the metric $g^{\mu\nu}$ , the fluid four-velocity $u^\mu$ , and the two rest-frame scalars $\rho$ and $p$ ; nothing else is available, and isotropy fixes how they combine. Work in geometrized units $c=1$ , signature $(-,+,+,+)$ .

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

where: $T^{\mu\nu}$ — stress-energy tensor, the flux of $\mu$ -momentum through a surface of constant $x^\nu$ ; $\rho$ — energy density in the rest frame (rest mass plus internal energy); $p$ — isotropic rest-frame pressure; $u^\mu=\mathrm dx^\mu/\mathrm d\tau$ — fluid four-velocity; $g^{\mu\nu}$ — inverse metric. Reading: the tensor is two rank-2 building blocks, $u^\mu u^\nu$ and $g^{\mu\nu}$ , weighted by the enthalpy $(\rho+p)$ and the pressure $p$ . In the local rest frame the four-velocity is $u^\mu=(1,0,0,0)$ and equation (1) collapses to a diagonal matrix — the operational meaning of " $\rho$ and $p$ are the only observables".

T^{\mu\nu}_{\text{rest}} = \begin{pmatrix} \rho & 0 & 0 & 0\\ 0 & p & 0 & 0\\ 0 & 0 & p & 0\\ 0 & 0 & 0 & p \end{pmatrix} \qquad (2)

where: the $00$ entry is energy density $\rho$ ; the equal spatial diagonal entries are the isotropic pressure $p$ ; every off-diagonal entry vanishes because there is no shear and no heat flux. Reading: ten independent components of a symmetric tensor are reduced to two by isotropy. The four-velocity is a unit timelike vector — this normalization is what fixes the coefficient $(\rho+p)$ and threads through every projection below.

u_\mu u^\mu = g_{\mu\nu}\,u^\mu u^\nu = -1 \qquad (3)

where: $u_\mu=g_{\mu\nu}u^\nu$ — covariant four-velocity. Reading: the flow is always timelike and future-directed; differentiating this identity gives $u_\mu a^\mu=0$ , so four-acceleration is spatial. Once the fluid is coupled to gravity, its dynamics are not a separate postulate. The contracted Bianchi identity $\nabla_\mu G^{\mu\nu}=0$ combined with the field equations forces local conservation.

\nabla_\mu T^{\mu\nu} = 0 \qquad (4)

where: $\nabla_\mu$ — covariant derivative compatible with $g_{\mu\nu}$ . Reading: this one equation contains both energy conservation (its projection along $u^\nu$ ) and the momentum balance (its projection orthogonal to $u^\nu$ ). Contracting the tensor with the metric gives its trace, the invariant that distinguishes matter from radiation and controls how the fluid sources gravity.

T^\mu{}_\mu = g_{\mu\nu}T^{\mu\nu} = -\rho + 3p \qquad (5)

where: $T^\mu{}_\mu$ — Lorentz-invariant trace, using $u_\mu u^\mu=-1$ and $g^\mu{}_\mu=4$ . Reading: the trace vanishes exactly when $p=\rho/3$ (radiation); the closely related combination $\rho+3p$ (the trace of the active gravitational source) sets the sign of cosmic acceleration. Projecting (4) along the flow gives the relativistic generalization of mass-energy conservation — how compression changes the energy density.

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

where: $u^\mu\nabla_\mu\rho$ — rate of change of $\rho$ along the flow; $\nabla_\mu u^\mu=\theta$ — expansion (fractional volume rate) of the fluid congruence. Reading: energy density drops when the fluid expands, and pressure does work in the process, so it is $(\rho+p)$ that multiplies the expansion. Projecting (4) orthogonal to the flow gives the relativistic Euler equation — Newton's second law for a fluid element, with a strikingly different inertia.

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

where: $a^\nu=u^\mu\nabla_\mu u^\nu$ — four-acceleration; $h^{\mu\nu}$ — projector orthogonal to $u^\mu$ , extracting the spatial pressure gradient; $(\rho+p)$ — effective inertial mass density. Reading: a pressure gradient across a fluid element accelerates it, but the element's resistance to acceleration is the enthalpy $(\rho+p)$ , not $\rho$ . Closing the system requires one more relation, the equation of state, whose derivative is the sound speed — bounded by $c$ for causality.

p = p(\rho,s),\qquad c_s^2 = \left(\frac{\partial p}{\partial \rho}\right)_{\!s}\le 1 \qquad (8)

where: $s$ — specific entropy, held fixed for adiabatic perturbations; $c_s$ — speed of small sound waves. Reading: the conservation laws (6)–(7) are four equations for five unknowns ( $\rho$ , $p$ , and the three independent components of $u^\mu$ ); the equation of state supplies the missing relation and selects the physical regime.

Derivation AWhy the perfect-fluid tensor is unique, and how a boost gives pressure its inertia

The goal is to show that the two-parameter form $T^{\mu\nu}=(\rho+p)u^\mu u^\nu+p\,g^{\mu\nu}$ is not a guess but the only symmetric rank-2 tensor that is isotropic in the rest frame and built from the available ingredients, and then to boost it and watch pressure enter the momentum density — the relativistic fact with no Newtonian analogue. We fix the coefficients by matching rest-frame components, then apply a Lorentz boost to expose $(\rho+p)$ as the true inertia.

Step 1. List the building blocks. From a single four-velocity field $u^\mu$ and the metric, the only symmetric rank-2 tensors available are $u^\mu u^\nu$ and $g^{\mu\nu}$ , so the most general isotropic ansatz is $T^{\mu\nu}=A\,u^\mu u^\nu + B\,g^{\mu\nu}$ with scalar coefficients $A,B$ . Why: any spatial vector we could add would pick out a direction and violate rest-frame isotropy, so it is excluded from the start.

Step 2. Match the rest frame. With $u^\mu=(1,0,0,0)$ and $g^{\mu\nu}=\mathrm{diag}(-1,1,1,1)$ , the time-time entry is $T^{00}=A-B$ and each spatial diagonal is $T^{ii}=B$ . Demanding $T^{00}=\rho$ and $T^{ii}=p$ gives $B=p$ and $A=\rho+p$ . Why: the definitions of energy density and pressure are exactly the rest-frame components, so matching them fixes both scalars with no freedom left — the form is unique.

Step 3. Boost along $x$ with rapidity giving $u^\mu=\gamma(1,v,0,0)$ , $\gamma=(1-v^2)^{-1/2}$ . The momentum density is $T^{0x}=(\rho+p)u^0u^x+p\,g^{0x}=(\rho+p)\gamma^2 v$ , since $g^{0x}=0$ . Why: the pressure term rides along in $(\rho+p)$ , so a moving pressurized fluid carries more momentum than its energy density alone would predict — pressure contributes to inertia.

Step 4. The coefficient multiplying the velocity in the momentum density is $(\rho+p)$ , the enthalpy density, playing the role Newtonian mechanics assigns to mass density; the boosted energy density likewise carries it as $T^{00}=(\rho+p)\gamma^2-p$ . Why: this is the precise sense in which "pressure has inertia" — it is the enthalpy, not the energy density, that resists being pushed, and the same combination reappears in the Euler equation. *Why this works end-to-end: the rest frame gives $\mathrm{diag}(\rho,p,p,p)$ ; a boost gives $T^{0x}=(\rho+p)\gamma^2 v$ and $T^{00}=(\rho+p)\gamma^2-p$ ; for radiation at $v=0.6c$ , $\gamma^2=1.5625$ , so $T^{0x}=1.25\rho$ and $T^{00}=1.75\rho$ , both larger than the dust values — pressure genuinely adds to inertia.*

Go deeper · Derivation B · 1 minContinuity and the relativistic Euler equation from conservation

Derivation BContinuity and the relativistic Euler equation from conservation

The goal is to unpack the compact law $\nabla_\mu T^{\mu\nu}=0$ into its physical content by projecting it along and orthogonal to the flow, yielding a scalar energy equation and a vector momentum equation. This is the derivation that turns a geometric identity into fluid dynamics, and it shows cleanly why the enthalpy $(\rho+p)$ governs both how energy dilutes under expansion and how a fluid responds to a pressure gradient. We differentiate the tensor, then use the four-velocity normalization to separate the two projections.

Step 1. Differentiate (1) with the product rule, using metric compatibility $\nabla_\mu g^{\mu\nu}=0$ : $\nabla_\mu T^{\mu\nu}=u^\mu u^\nu\nabla_\mu(\rho+p)+(\rho+p)[u^\nu\nabla_\mu u^\mu+u^\mu\nabla_\mu u^\nu]+g^{\mu\nu}\nabla_\mu p=0$ . Why: every term is now explicit in $\rho$ , $p$ , $u^\mu$ and its derivatives, ready to be sorted by projecting parallel or perpendicular to the flow.

Step 2. Project along the flow by contracting with $u_\nu$ . Using $u_\nu u^\nu=-1$ and its derivative $u_\nu\nabla_\mu u^\nu=0$ (so $u_\nu a^\nu=0$ ), the acceleration term drops and one is left with $-u^\mu\nabla_\mu(\rho+p)-(\rho+p)\nabla_\mu u^\mu+u^\mu\nabla_\mu p=0$ . Why: the normalization $u_\mu u^\mu=-1$ is precisely what annihilates the acceleration piece, isolating the scalar energy balance.

Step 3. Simplify: the $u^\mu\nabla_\mu p$ terms cancel, leaving the relativistic continuity equation $u^\mu\nabla_\mu\rho+(\rho+p)\nabla_\mu u^\mu=0$ . Why: this says the energy density carried along the flow decreases in proportion to the expansion $\nabla_\mu u^\mu$ , and the pressure does the compression work — the origin of $\dot\rho=-3H(\rho+p)$ in cosmology.

Step 4. Project orthogonal to the flow with $h_{\sigma\nu}=g_{\sigma\nu}+u_\sigma u_\nu$ , which satisfies $h_{\sigma\nu}u^\nu=0$ . The projector kills every term carrying a bare $u^\nu$ , leaving $(\rho+p)a_\sigma=-h_\sigma{}^\mu\nabla_\mu p$ — the relativistic Euler equation. Why: only the spatial pressure gradient survives to push the element, and the enthalpy $(\rho+p)$ sits where Newtonian mass would, closing the loop with the boost argument. *Why this recovers Newtonian fluid dynamics: in the limit $v\ll1$ , $p\ll\rho$ , $\rho\to\rho_0$ , the pair reduces to $\partial_t\rho_0+\nabla\cdot(\rho_0\mathbf{v})=0$ and $\rho_0(\partial_t\mathbf{v}+\mathbf{v}\cdot\nabla\mathbf{v})=-\nabla p-\rho_0\nabla\Phi$ , the classical continuity and Euler equations.*

Go deeper · Derivation C · 1 minCosmic dilution and the scaling ρ ∝ a to the minus 3(1+w)

Derivation CCosmic dilution and the scaling ρ ∝ a to the minus 3(1+w)

The goal is to apply the continuity equation to a homogeneous, isotropic universe and derive how each kind of matter dilutes as space expands, obtaining the master scaling $\rho\propto a^{-3(1+w)}$ that underlies the whole cosmic energy budget. This single result explains why radiation once dominated, why matter took over, and why a cosmological constant eventually wins — the three regimes plotted in the second figure. We specialize the covariant continuity equation to a comoving fluid in an expanding metric with scale factor $a(t)$ .

Step 1. Take the fluid comoving with the cosmic expansion, $u^\mu=(1,0,0,0)$ , in a spatially flat Friedmann metric where $\sqrt{-g}=a^3$ . The expansion is $\nabla_\mu u^\mu=\tfrac{1}{a^3}\partial_t(a^3)=3\dot a/a\equiv 3H$ . Why: the divergence of the comoving four-velocity is just the fractional rate of change of a comoving volume, which is three times the Hubble rate $H$ .

Step 2. Insert into the continuity equation (6): $\dot\rho+3H(\rho+p)=0$ . Why: this is the first law of thermodynamics for the universe — the energy in a comoving volume $a^3$ changes only by the pressure work $p\,\mathrm d(a^3)$ done as it expands.

Step 3. Introduce a barotropic equation of state $p=w\rho$ with constant $w$ , so $\dot\rho/\rho=-3(1+w)\dot a/a$ . Integrating gives $\rho\propto a^{-3(1+w)}$ . Why: separating variables turns the balance law into a pure power law; the exponent counts both the volume dilution ( $a^{-3}$ ) and the extra redshift work set by $w$ .

Step 4. Read off the regimes: dust $w=0\Rightarrow\rho\propto a^{-3}$ ; radiation $w=1/3\Rightarrow\rho\propto a^{-4}$ (the extra factor is the redshift of each photon's momentum); vacuum $w=-1\Rightarrow\rho=\text{const}$ . Why: because radiation falls faster than matter and matter faster than vacuum, the dominant component necessarily changes with time — radiation, then matter, then dark energy. *Why the numbers work out: the continuity result $\dot\rho=-3H(\rho+p)$ integrates to $\rho\propto a^{-3(1+w)}$ ; with $\Omega_m=0.315$ and $\Omega_r=9.2\times10^{-5}$ the $a^{-3}$ and $a^{-4}$ curves cross at $a=\Omega_r/\Omega_m=2.9\times10^{-4}$ , i.e. $z\approx3400$ .*

Limiting cases

The dimensionless equation of state $w\equiv p/\rho$ selects the physical regime; each row is one corner of the perfect-fluid story.

Case	Result	Meaning
Pressureless dust ( $w=0$ )	$p=0$ , $\;T^{\mu\nu}=\rho\,u^\mu u^\nu$	Non-interacting massive particles — galaxies, cold dark matter. Dilutes as $a^{-3}$ ; the four-velocity alone carries the tensor.
Radiation ( $w=1/3$ )	$p=\rho/3$ , $\;T^\mu{}_\mu=0$	Photons and ultrarelativistic gas; the trace vanishes, signalling conformal invariance. Dilutes as $a^{-4}$ ; sound speed $c_s=1/\sqrt3$ .
Vacuum energy ( $w=-1$ )	$p=-\rho$ , $\;T^{\mu\nu}=-\rho\,g^{\mu\nu}$	Constant density, isotropic in every frame. Source $\rho+3p=-2\rho<0$ drives accelerated expansion — dark energy.
Stiff fluid ( $w=1$ )	$p=\rho$ , $\;c_s=c$	Sound speed equals light speed, the causal ceiling for a barotropic fluid; relevant to a kinetic-energy-dominated scalar field in the early universe.
Newtonian limit	$p\ll\rho$ , $\;v\ll c$	Energy density is dominated by rest mass $\rho\to\rho_0$ ; the relativistic Euler and continuity equations reduce to classical fluid dynamics.

Common trapPressure is just force per area — it cannot affect gravity or a body's inertia

It does both, and that is the heart of the relativistic story. Inertially, the resistance of a fluid element to acceleration is the enthalpy density $(\rho+p)$ , not the energy density $\rho$ : this is the coefficient that multiplies the four-acceleration in the Euler equation and the momentum in a boosted frame, so a high-pressure gas is genuinely harder to push. Gravitationally, the active source in the Einstein equations is not $\rho$ but the combination $\rho+3p$ (the trace-reversed source). A hot, high-pressure star therefore weighs more than its energy content alone would suggest — pressure gravitates — which is exactly why the Tolman-Oppenheimer-Volkoff equation makes neutron stars so much harder to support than Newtonian intuition allows, and why raising the pressure to fight collapse eventually backfires. Turn the sign around: a medium with $p<-\rho/3$ makes $\rho+3p<0$ and gravitates repulsively, the mechanism behind cosmic acceleration. Pressure is never "just" a mechanical stress in general relativity.

Worked examples

1 · When did matter overtake radiation? (fully numeric)

The present universe holds matter with density parameter $\Omega_m=0.315$ and radiation (photons plus neutrinos) with $\Omega_r=9.2\times10^{-4}\times0.1\approx9.2\times10^{-5}$ . Using the perfect-fluid scaling law, find the redshift of matter-radiation equality.

Step 1. From Derivation C, matter dilutes as $\rho_m=\rho_{m,0}\,a^{-3}$ and radiation as $\rho_r=\rho_{r,0}\,a^{-4}$ , because $w=0$ and $w=1/3$ respectively. Their ratio evolves as $\rho_r/\rho_m\propto a^{-1}$ , so radiation, though tiny today, must have dominated at early times.

Step 2. Set the densities equal: $\rho_{m,0}\,a_{\rm eq}^{-3}=\rho_{r,0}\,a_{\rm eq}^{-4}$ , which gives $a_{\rm eq}=\rho_{r,0}/\rho_{m,0}=\Omega_r/\Omega_m$ . Numerically $a_{\rm eq}=9.2\times10^{-5}/0.315=2.9\times10^{-4}$ .

Step 3. Convert to redshift with $1+z=1/a$ : $1+z_{\rm eq}=1/(2.9\times10^{-4})=3.4\times10^{3}$ . Matter-radiation equality is at $z_{\rm eq}\approx3400$ , in excellent agreement with the value inferred from the cosmic microwave background. Before this epoch the universe expanded as $a\propto t^{1/2}$ (radiation era); afterward as $a\propto t^{2/3}$ (matter era) — the entire structure-formation history hinges on this crossover, and it followed from one line of the conservation law.

2 · Does the universe accelerate today? (fully numeric)

The Friedmann acceleration equation reads $\ddot a/a=-\tfrac{4\pi G}{3}(\rho+3p)$ . Using $\Omega_m=0.315$ , $\Omega_\Lambda=0.685$ (and negligible radiation), evaluate the active gravitational density $\rho+3p$ today, in units of the critical density, and decide the sign of $\ddot a$ .

Step 1. The relevant source is $\rho+3p$ , the trace combination from equation (5). Matter is pressureless ( $w=0$ ), so it contributes $\rho_m(1+3\cdot0)=\rho_m$ . Vacuum energy has $w=-1$ , so it contributes $\rho_\Lambda(1+3(-1))=-2\rho_\Lambda$ .

Step 2. In units of the critical density $\rho_c$ , add the pieces: $(\rho+3p)/\rho_c=\Omega_m-2\Omega_\Lambda=0.315-2(0.685)=0.315-1.370=\mathbf{-1.06}$ .

Step 3. Because $\rho+3p<0$ , the acceleration equation gives $\ddot a>0$ : the expansion is accelerating today. Matter alone ( $+0.315$ ) would decelerate it, but the negative pressure of dark energy wins by more than a factor of four. Tracing the balance backward, matter and vacuum densities were equal at $a=(\Omega_m/\Omega_\Lambda)^{1/3}=0.77$ , i.e. $z\approx0.3$ ; acceleration set in around then. This sign flip — impossible in Newtonian gravity, where only mass gravitates — is the single most important consequence of pressure appearing in $T^{\mu\nu}$ .

3 · The momentum a moving radiation fluid carries (fully numeric)

A photon gas (radiation, $p=\rho/3$ ) drifts with bulk velocity $v=0.6c$ . Compute its energy density and momentum density in the lab frame, and compare the momentum with that of pressureless dust of the same rest-frame $\rho$ moving at the same speed.

Step 1. The Lorentz factor gives $\gamma^2=1/(1-0.6^2)=1/0.64=1.5625$ . From the boosted tensor (Derivation A), the momentum density is $T^{0x}=(\rho+p)\gamma^2 v$ . For radiation $\rho+p=\tfrac43\rho$ , so $T^{0x}=\tfrac43(1.5625)(0.6)\,\rho=\mathbf{1.25\,\rho}$ .

Step 2. The lab energy density is $T^{00}=(\rho+p)\gamma^2-p=\tfrac43(1.5625)\rho-\tfrac13\rho=2.083\rho-0.333\rho=\mathbf{1.75\,\rho}$ . Both exceed the naive rest-frame $\rho$ , because kinetic energy and the pressure's inertia both add to what the lab observer measures.

Step 3. For dust of the same $\rho$ , $p=0$ and $\rho+p=\rho$ , so its momentum density is $T^{0x}_{\rm dust}=\rho\gamma^2 v=(1.5625)(0.6)\rho=0.9375\rho$ . The radiation carries $1.25/0.9375=1.33$ times as much momentum as dust of identical energy density — precisely the factor $(\rho+p)/\rho=1+w=4/3$ . The pressure of the photon gas is not a spectator; it is a full one-third contribution to the fluid's inertia.

Physics of the portrait

The perfect-fluid tensor is where the abstract idea of energy-momentum as geometry meets the concrete matter of stars and universes — and the person who made it the working language of cosmology and stellar structure was a chemist who became a relativist.

Origins · 1 minRichard C. Tolman and how this was discovered

Richard C. Tolman

1881–1948 · the physical chemist who made the relativistic perfect fluid the working language of cosmology and stellar structure

Richard Chace Tolman trained as a physical chemist at MIT before turning, in his Caltech years, into one of the deepest relativists of his generation. His 1934 treatise Relativity, Thermodynamics and Cosmology did for the matter side of Einstein's equations what Einstein had done for the geometry: it systematized the perfect fluid as the source of gravity, worked out its thermodynamics in an expanding universe, and made explicit the two facts that most sharply separate relativistic from Newtonian gravity — that the inertia of a fluid element is its enthalpy $\rho+p$ , and that its gravitational weight is set by $\rho+3p$ , so pressure gravitates. The "Tolman mass" and the Tolman surface-brightness test both descend from this insight. In 1939, in the same volume of Physical Review as Oppenheimer and Volkoff's paper on neutron cores, Tolman published the general static-fluid solutions of Einstein's equations, and the equation of hydrostatic equilibrium for a relativistic star — the Tolman-Oppenheimer-Volkoff equation — carries his name first. He showed that pressure, which in a Newtonian star always helps resist collapse, in a relativistic star also adds to the gravitating mass, so that raising the pressure to resist collapse eventually works against the star. That result, born from the humble two-parameter tensor, is the theoretical seed of the neutron-star maximum mass and, ultimately, of the black hole.

History

1908

— Minkowski casts energy, momentum and stress as a single spacetime tensor

1911

— von Laue shows internal stresses contribute to a closed system's energy-momentum

1915

— Einstein sets Gᵘᵛ = 8πG Tᵘᵛ, making the stress-energy tensor the source of gravity

1934

— Tolman systematizes the relativistic perfect fluid and shows that pressure gravitates

1939

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

Going further

Texts. Misner, Thorne & Wheeler, Gravitation (1973), §5.5–5.6, §22.2 · Wald, General Relativity (1984), §4.3, Appendix E · Carroll, Spacetime and Geometry (2004), §1.9, §4.4, §8.1 · Schutz, A First Course in General Relativity (2nd ed., 2009), Ch. 4 · Hartle, Gravity (2003), Ch. 22 · Tolman, Relativity, Thermodynamics and Cosmology (1934), §§152–165. The key results to keep in mind: equation (1) reduces to $\mathrm{diag}(\rho,p,p,p)$ in the rest frame and the coefficients are uniquely fixed by that match; the trace is $T^\mu{}_\mu=-\rho+3p$ in four dimensions; the projected conservation laws give continuity $\dot\rho=-3H(\rho+p)$ and the relativistic Euler equation; the scaling $\rho\propto a^{-3(1+w)}$ yields the equality redshift $z_{\rm eq}=\Omega_m/\Omega_r-1\approx3400$ with $\Omega_m=0.315$ , $\Omega_r\approx9.2\times10^{-5}$ ; the active-density sign gives $\Omega_m-2\Omega_\Lambda=-1.06<0$ for acceleration; and the boosted radiation values at $v=0.6c$ are $T^{0x}=1.25\rho$ , $T^{00}=1.75\rho$ .

Prerequisites. Four-Vectors · Minkowski Metric and Signature · Proper Time · Lorentz Boosts and Rapidity. Leads to. Einstein Field Equations · Friedmann Equations · Tolman-Oppenheimer-Volkoff Equation · Relativistic Euler Equation.

Ask the tutor: Ask me to project $\nabla_\mu T^{\mu\nu}=0$ for any equation of state, or to trace how pressure enters the inertia $\rho+p$ and the gravitating source $\rho+3p$ .

What exactly makes a fluid "perfect"?

"Perfect" is a statement about the rest frame, not about the flow. In the frame momentarily comoving with a fluid element, a perfect fluid is isotropic — no direction is special — and dissipationless — it generates no entropy. Those two conditions remove every stress except an isotropic pressure: there is no viscosity (which would produce shear or bulk stresses opposing the flow) and no heat conduction (which would produce an energy flux across fluid elements). What remains is captured by just the energy density $\rho$ and the pressure $p$ . Real fluids are imperfect: adding shear-viscous, bulk-viscous, and heat-flux terms gives the relativistic Navier-Stokes, Eckart, and Israel-Stewart theories, which are corrections layered on top of the perfect-fluid tensor.

Does pressure really gravitate, and how is that different from energy?

Yes. In Newtonian gravity only mass density sources the field. In general relativity the source is the whole stress-energy tensor, and when you write the field equations in trace-reversed form the effective gravitating density for a static fluid is $\rho+3p$ , not $\rho$ . Each of the three pressures on the spatial diagonal adds its own contribution, hence the factor of three. For ordinary matter $p\ll\rho$ and the correction is tiny, but inside a neutron star $p$ is a sizeable fraction of $\rho$ and the extra weight is decisive: it is one reason such stars have a maximum mass. The same term with the opposite sign is what lets vacuum energy, with $p=-\rho$ , produce repulsive gravity and accelerate the universe.

Is ρ the mass density or the energy density?

It is the total energy density measured in the rest frame, which includes rest mass and internal energy (thermal motion, binding, radiation), not the rest-mass density $\rho_0=nm$ alone. The distinction rarely matters for cold matter, where almost all the energy is rest mass, but it is essential for hot or relativistic matter. In the conservation laws it is always the total energy density that appears, and it enters the inertial and gravitational combinations $(\rho+p)$ and $(\rho+3p)$ . A common slip is to insert the rest-mass density into the relativistic Euler equation; the correct inertia is the full enthalpy density $\rho+p$ , which reduces to $\rho_0$ only in the Newtonian limit where pressure and internal energy are negligible.

How can vacuum energy have negative pressure — isn't that strange?

It is unusual but not paradoxical. Vacuum energy must look the same to every observer, so its stress-energy has to be proportional to the only frame-independent rank-2 tensor available, the metric: $T^{\mu\nu}=-\rho\,g^{\mu\nu}$ . Comparing with the perfect-fluid form forces $p=-\rho$ . Physically, if you stretch a box of vacuum its energy grows with volume rather than dropping, which is exactly what a tension (negative pressure) does — the vacuum pulls inward on its walls. There is nothing acausal here: for a genuine cosmological constant $\rho$ is fixed, so no sound wave propagates. The negative pressure's only job is to make $\rho+3p<0$ and drive acceleration.

Why is the sound speed limited by c, and what is the stiffest fluid?

Small adiabatic disturbances in the fluid travel at $c_s=\sqrt{(\partial p/\partial\rho)_s}$ . Causality — no signal faster than light — requires $c_s\le c$ , which constrains the equation of state: $p$ cannot rise faster than $\rho$ . The limiting case $c_s=c$ is the "stiff" fluid $p=\rho$ ( $w=1$ ), the hardest barotropic equation of state consistent with relativity, realized for instance by a scalar field whose kinetic energy dominates its potential. Radiation, by contrast, has $p=\rho/3$ and hence $c_s=c/\sqrt3\approx1.7\times10^8$ m/s. A would-be equation of state with $p>\rho$ would imply superluminal sound and is excluded. This bound is one of the main theoretical constraints on the ultra-dense matter inside neutron stars.

Where does the conservation law ∇ₘTᵘᵛ = 0 come from — is it a separate assumption?

Not once gravity is switched on. In flat spacetime you would impose $\partial_\mu T^{\mu\nu}=0$ by hand as energy-momentum conservation. In general relativity it is automatic: the Einstein tensor obeys the contracted Bianchi identity $\nabla_\mu G^{\mu\nu}=0$ as a pure geometric fact, so the field equations $G^{\mu\nu}=8\pi G\,T^{\mu\nu}$ immediately force $\nabla_\mu T^{\mu\nu}=0$ . Conservation of energy-momentum is therefore built into the structure of the theory rather than added to it — and it is a strong consistency check, because it means the matter dynamics (the continuity and Euler equations) are not independent of the gravitational field equations but a consequence of them. This is why you cannot specify an arbitrary $T^{\mu\nu}$ and an arbitrary motion; the geometry enforces the fluid's equations of motion.

How is a perfect fluid different from the water in a real pipe?

Real water has viscosity and conducts heat, so it is not a perfect fluid — dragging it past a wall sets up shear stresses and dissipates energy, none of which the perfect-fluid tensor contains. Yet the idealization is often superb, because for many flows the dissipative terms are small compared with pressure and inertia. In astrophysics and cosmology it is excellent: interstellar gas, stellar interiors, and the cosmic matter distribution have negligible viscosity on the relevant scales. The most spectacular real example is the quark-gluon plasma, whose viscosity-to-entropy ratio sits near the smallest value quantum mechanics seems to allow. When dissipation does matter, one adds Navier-Stokes-type corrections rather than abandoning the framework.