Skip to content

Covariant Electromagnetic Fluids

Scrollable formula. Use Left and Right arrow keys to reveal hidden content.

Scrollable table. Use Left and Right arrow keys to reveal additional columns.

Scrollable figure. Use Left and Right arrow keys to inspect hidden content.

A relativistic plasma exchanges energy and momentum internally between charged matter and the electromagnetic field. Maxwell’s equations make that exchange exact: the divergence of the electromagnetic stress tensor is minus the Lorentz four-force on matter, so the sum of field and matter stress tensors is conserved. This covariant bookkeeping is the foundation on which ideal MHD, resistive MHD, and force-free electrodynamics impose different constitutive closures.

Required background. Hilbert Stress–Energy from Metric Variation supplies symmetric stress tensors and covariant conservation; Exterior Algebra and Differential Forms supplies the electromagnetic two-form and its homogeneous equation.

Helpful background. The Einstein Equation gives gravitational coupling; Relativistic Matter and Compact Objects places magnetofluids among the volume’s matter models; and Computational Relativity provides evolution formulations.

Use Heaviside–Lorentz electromagnetic units, so there is no 4π4\pi in Maxwell’s source equation. With F=dAF=dA and the electric-field definition below, the antisymmetric field tensor FabF_{ab} and current JaJ^a obey

∇aFba=Jb,∇[aFbc]=0.\nabla_aF^{ba}=J^b, \qquad \nabla_{[a}F_{bc]}=0.

The order of indices in the source equation matters: equivalently, ∇aFab=−Jb\nabla_aF^{ab}=-J^b. The homogeneous equation is dF=0dF=0 and locally permits F=dAF=dA. Taking a divergence of the source equation gives

∇aJa=0,\nabla_aJ^a=0,

because the double divergence of an antisymmetric tensor vanishes after commuting covariant derivatives. Charge conservation is therefore a compatibility condition of Maxwell’s equations, not an optional material equation.

For a unit timelike observer nan^a, define measured electric and magnetic fields by

Ea=Fabnb,Ba=−∗Fabnb,naEa=naBa=0.E^a=F^{ab}n_b, \qquad B^a=-{}^*F^{ab}n_b, \qquad n_aE^a=n_aB^a=0.

Here the Hodge dual is ∗Fab=12ϵabcdFcd{}^*F^{ab}=\tfrac12\epsilon^{abcd}F_{cd}. The minus sign in BaB^a follows from the site’s spacetime orientation, ϵ0123=+−g\epsilon_{0123}=+\sqrt{-g}, when the observer’s right-handed spatial basis has ϵ123=+1\epsilon^{123}=+1. In that orthonormal frame,

F0i=Ei,Fij=ϵijkBk.F^{0i}=E^i,\qquad F^{ij}=\epsilon^{ijk}B_k.

These components give ∇⋅E=J0\boldsymbol\nabla\cdot\boldsymbol E=J^0 and ∂tE=∇×B−J\partial_t\boldsymbol E=\boldsymbol\nabla\times\boldsymbol B-\boldsymbol J, fixing the convention without ambiguity. A reference that uses the opposite volume-form orientation may define BaB^a with the opposite dual sign.

The invariant FabFab=2(B2−E2)F_{ab}F^{ab}=2(B^2-E^2) and the vanishing of F∧FF\wedge F distinguish magnetic dominance and field degeneracy, respectively. In particular, F∧F=0F\wedge F=0 is equivalent to EaBa=0E_aB^a=0. Those classifications do not depend on the observer; see Gralla and Jacobson 2014, § 3.2.1, Eqs. (11)–(14).

The electromagnetic stress–energy tensor is

TEMab=FacFbc−14gabFcdFcd.T_{\rm EM}^{ab} =F^a{}_cF^{bc}-\frac14g^{ab}F_{cd}F^{cd}.

Using both Maxwell equations and antisymmetry gives

∇aTEMab=−FbcJc.\nabla_aT_{\rm EM}^{ab}=-F^b{}_cJ^c.

Thus fb=FbcJcf^b=F^b{}_cJ^c is the four-force density delivered to matter. If TmabT_{\rm m}^{ab} denotes all material stress, its equation is

∇aTmab=FbcJc,\nabla_aT_{\rm m}^{ab}=F^b{}_cJ^c,

and hence

∇a ⁣(Tmab+TEMab)=0.\nabla_a\!\left(T_{\rm m}^{ab}+T_{\rm EM}^{ab}\right)=0.

The sign can be checked in a local inertial frame: the temporal component transfers power J⋅E\boldsymbol J\cdot\boldsymbol E to matter, while the spatial component gives charge force plus current cross magnetic field. The field loses exactly what matter gains.

The observer projections provide a second check. Write γab=gab+nanb\gamma^{ab}=g^{ab}+n^an^b for the spatial projector and ϵabc=ndϵdabc\epsilon^{abc}=n_d\epsilon^{dabc} for the observer’s spatial volume tensor. In the nan^a frame, field energy density, momentum density, and spatial stress are

UEM=12(E2+B2),Sa=ϵabcEbBc,U_{\rm EM}=\frac12(E^2+B^2), \qquad S^a=\epsilon^{abc}E_bB_c, SEMab=12(E2+B2)γab−EaEb−BaBb.S_{\rm EM}^{ab}=\frac12(E^2+B^2)\gamma^{ab}-E^aE^b-B^aB^b.

The stress is tensile along magnetic field lines and pressurized transverse to them. Its trace satisfies TEMaa=0T_{{\rm EM} a}{}^a=0, consistent with the absence of a classical electromagnetic mass scale in four spacetime dimensions.

Relative to a material four-velocity uau^a, split the current into

Ja=qua+ja,uaja=0,J^a=q u^a+j^a, \qquad u_a j^a=0,

where q=−uaJaq=-u_aJ^a is comoving charge density and jaj^a is conduction current. Maxwell’s equations relate the field derivatives to the total current, but do not close the joint field–matter evolution. Matter conservation, an equation of state, and a constitutive relation for jaj^a are needed; Gauss’s law remains a constraint on any proposed charge density.

Ideal MHD imposes vanishing comoving electric field Fabub=0F_{ab}u^b=0. Resistive MHD permits a finite conductivity; its simplest Ohm law is algebraic, while more detailed models evolve the conduction current with a relaxation time. Force-free electrodynamics removes material inertia from the leading momentum balance and demands FbcJc=0F^b{}_cJ^c=0. These are different limits of a plasma, not interchangeable algebraic rewritings. The ideal stress and conservation laws appear in Martí and Müller 2015, § 3.1, Eqs. (1)–(7).

Charge neutrality also needs care. A plasma may be quasineutral on macroscopic scales while carrying a nonzero conduction current because positive and negative charges drift oppositely. Setting q=0q=0 does not imply Ja=0J^a=0. Conversely, a one-fluid current closure is trustworthy only when unresolved charge-separation and kinetic scales remain small compared with the scales being modeled.

With Ja=0J^a=0, electromagnetic stress is conserved separately. With Fab=0F_{ab}=0, the matter equation reduces to ordinary fluid conservation. In a local frame with Ea=0E^a=0, the Poynting vector vanishes for the observer comoving with that frame and the magnetic energy is B2/2B^2/2; a different observer generally sees an electric field and Poynting flux. The invariant statement is not “electric field zero for everyone” but the existence of a timelike frame in which it vanishes.

The decomposition assumes classical electromagnetism and minimal coupling. Polarization, magnetization, quantum electrodynamic vacuum effects, and non-Abelian fields alter the constitutive stress or field equations. An algebraic scalar Ohm law coupled to the full Maxwell equations can already have causal field propagation: finite conductivity by itself does not make Maxwell’s equations parabolic. Magnetic diffusion arises only after eliminating the fast electric-field dynamics, as the resistive-MHD page demonstrates.

In a flat-space region with a spatially uniform electric field Ex(t)E_x(t), zero magnetic field, and conductivity σ>0\sigma\gt0, derive the decay of field energy and the matter heating rate. Explain how this checks both the Maxwell-source sign and total stress conservation.

Solution

Ampère’s equation and Jx=σExJ_x=\sigma E_x give E˙x=−σEx\dot E_x=-\sigma E_x, hence Ex(t)=Ex(0)e−σtE_x(t)=E_x(0)e^{-\sigma t}. The field energy satisfies

ddtEx22=−σEx2,f0=J⋅E=σEx2.\frac{d}{dt}\frac{E_x^2}{2}=-\sigma E_x^2, \qquad f^0=\boldsymbol J\cdot\boldsymbol E=\sigma E_x^2.

Thus the field loses exactly the heat gained by matter. More generally,

∇aTmab+∇aTEMab=FbcJc−FbcJc=0.\nabla_aT_{\rm m}^{ab}+\nabla_aT_{\rm EM}^{ab} =F^b{}_cJ^c-F^b{}_cJ^c=0.

The opposite signs describe internal exchange. Positive work on matter is an equal loss from the electromagnetic field, so no external four-momentum is created.

  • Gralla, Samuel E., and Ted Jacobson. “Spacetime Approach to Force-Free Magnetospheres.” Monthly Notices of the Royal Astronomical Society 445, 2500–2534 (2014). Open preprint.
  • Martí, José María, and Ewald Müller. “Grid-Based Methods in Relativistic Hydrodynamics and Magnetohydrodynamics.” Living Reviews in Computational Astrophysics 1, 3 (2015). DOI.