Poynting’s Theorem and Conservation of Energy and Momentum for a System of Charged Particles and Electromagnetic Fields The forms of the laws of conservation of energy and momentum are important results to establish for the electromagnetic field. We begin by considering con- servation of energy, often called Poynting’s theorem (1884). For a single charge q the rate of doing work by external electromagnetic fields E and B is qv-E, where v is the velocity of the charge. The magnetic field does no work, since the magnetic force is perpendicular to the velocity. If there exists a continuous dis- tribution of charge and current, the total rate of doing work by the fields in a finite volume V isI J Ed’x (6.103) This power represents a conversion of electromagnetic energy into mechanical or thermal energy. It must be balanced by a corresponding rate of decrease of energy in the electromagnetic field within the volume V. To exhibit this conser- vation law explicitly, we use the Maxwell equations to express (6.103) in other terms. Thus we use the Ampére-Maxwell law to eliminate J: D [reas [fee xm £2] ax (6.104)If we now employ the vector identity, V-(Ex H)=H-(V x E)-E-(V x H) and use Faraday’s law, the right-hand side of (6.104) becomes [aear=-[[eexm+e- Ban aB ax (6. 1 | x (6.105)To proceed further we make two assumptions: (1) the macroscopic medium is linear in its electric and magnetic properties, with negligible dispersion or losses, and (2) the sum of (4.89) and (5.148) represents the total electromagnetic energy density, even for time-varying fields. With these two assumptions and the total energy density denoted by un 5 (E-D + B-H) (6.106)(6.105) can be written a -[ JR ax = i E +V-(Ex »| x (6.107) v v [at Since the volume V is arbitrary, this can be cast into the form of a differential continuity equation or conservation law, ou —+4V-S=-J-E (6.108) otThe vector S, representing energy flow, is called the Poynting vector. It is given by S=EXxH (6.109) and has the dimensions of (energy/area X time). Since only its divergence appears in the conservation law, the Poynting vector seems arbitrary to the extent that the curl of any vector field can be added to it. Such an added term can, however, have no physical consequences. Relativistic considerations (Section 12.10) show that (6.109) is unique.The physical meaning of the integral or differential form (6.107) or (6.108) is that the time rate of change of electromagnetic energy within a certain volume, plus the energy flowing out through the boundary surfaces of the volume per unit time, is equal to the negative of the total work done by the fields on the sources within the volume. This is the statement of conservation of energy. The assumptions that follow (6.105) really restrict the applicability of the simple ver- sion of Poynting’s theorem to vacuum macroscopic or microscopic fields. Even for linear media, there is always dispersion (with accompanying losses). Then the right-hand side of (6.105) does not have the simple interpretation exhibited in (6.107). The more realistic situation of linear dispersive media is discussed in the next section.The emphasis so far has been on the energy of the electromagnetic fields. The work done per unit time per unit volume by the fields (J - E) is a conversion of electromagnetic energy into mechanical or heat energy. Since matter is ulti- mately composed of charged particles (electrons and atomic nuclei), we can think of this rate of conversion as a rate of increase of energy of the charged particles per unit volume. Then we can interpret Poynting’s theorem for the microscopic fields (E, B) as a statement of conservation of energy of the combined system ofparticles and fields. If we denote the total energy of the particles within the volume V as Ejnech and assume that no particles move out of the volume, we have dE, ech | mech _ fy. eg gs 6.110 dt v * (6-10) Then Poynting’s theorem expresses the conservation of energy for the combined system as dE d 7 ap (Emecn + Eneta) = -$ n+ S da (6.111)where the total field energy within V is Esea = [ u dx = 2 (E? + 2B?) dx (6.112) v The conservation of linear momentum can be similarly considered. The total electromagnetic force on a charged particle is F = q(E+vxB) (6.113) If the sum of all the momenta of all the particles in the volume V is denoted by Prnechs WE can write, from Newton’s second law, AP meen a= [ (pE + J x B) d°x (6.114)where we have converted the sum over particles to an integral over charge and current densities for convenience in manipulation. In the same manner as for Poynting’s theorem, we use the Maxwell equations to eliminate p and J from (6.114): p=«V-E, J=tvxB— 6 Ho ot With (6.115) substituted into (6.114) the integrand becomes (6.115) ms rxpe aleve sex on x (x m|Then writing dE a OB x — = -—(E x B)+Ex — ot an ( ) ot and adding c’B(V - B) = 0 to the square bracket, we obtain pE + J x B= &[E(V +E) + 2B(V -B) — Ex (Vx E) —- Bx (¥ x B)] «> (Ex B)The rate of change of mechanical momentum (6.114) can now be written AP nec, A a 3 Get Gy Jy SE % B) ax (6.116) = 6 | [eV -B) — Bx (Vx B) + CBIV-B) ~ CB x (Vx BY] aeWe may tentatively identify the volume integral on the left as the total electro- magnetic momentum Px.1g in the volume V: Prea = €0 I, E x Bd’x = poe I, Ex Hd’x (6.117) The integrand can be interpreted as a density of electromagnetic momentum. We note that this momentum density is proportional to the energy-flux density S, with proportionality constant c~*.We may tentatively identify the volume integral on the left as the total electro- magnetic momentum Pyeig in the volume V: Preia = €0 I, E x B d?x = poe I Ex Hd*x (6.117) The integrand can be interpreted as a density of electromagnetic momentum. We note that this momentum density is proportional to the energy-flux density S, with proportionality constant c~*.[E(V-E) -Ex (Vx Eh, =p, (1 4 ME, Es) (Ma _ EL) (Ex _ as Ox, oe 0x3 Ox, oe 0X3 Ox, a = 2) + 2 Esk) + (EE) ~ 35, (EE + B+ BD) : ts This means that we can write the ath component as [E(V-E)-Ex(V xB), => = (EqEg ~ $E+E8q,) (6.119) B Band have the form of a divergence of a second rank tensor on the right-hand side. With the definition of the Maxwell stress tensor Tag as Tap = €[EaEp + CB.By — (E+E + CB + B)d,9] (6.120) we can therefore write (6.116) in component form as d a (Panecn + Pretada = z vin, Tag 2x (6.121)Application of the divergence theorem to the volume integral gives d a Pech + Preisda = f D Topnp da (6.122) it SB where n is the outward normal to the closed surface S. Evidently, if (6.122) rep- resents a statement of conservation of momentum, = ,T g/g is the ath component of the flow per unit area of momentum across the surface S into the volume V. In other words, it is the force per unit area transmitted across the surface S and acting on the combined system of particles and fields inside V. Equation (6.122) can therefore be used to calculate the forces acting on material objects in elec- tromagnetic fields by enclosing the objects with a boundary surface S and adding up the total electromagnetic force according to the right-hand side of (6.122).The conservation of angular momentum of the combined system of particles and fields can be treated in the same way as we have handled energy and linear momentum. This is left as a problem for the student (see Problem 6.10). The discussion of electromagnetic momentum and the stress tensor in fluids and solids entails analysis of interplay of mechanical, thermodynamic, and elec- tromagnetic properties (e.g., d/0T and de/ap). We refer the reader to Landay and Lifshitz, Electrodynamics of Continuous Media (Sections 10, 15, 16, 31, 35), Stratton (Chapter 2), and, for a statistical mechanical approach, to de Groot (Sec- tion 13). We note only that, although a treatment using the macroscopic Maxwell equations leads to an apparent electromagnetic momentum, g = D x B (Min- kowski, 1908), the generally accepted expression for a medium at rest is1 2 SE XH = wok xH= 48 (6.123) 2 We note that g is the electromagnetic momentum associated with the fields There is an additional co-traveling momentum within the medium from the me- chanical momentum of the electrons in the molecular dipoles in response to the incident traveling wave.* The Minkowski momentum of a plane wave is the “pseudomomentum” of the wave vector (k = nw/c or hk = n(hw)/c for a photon).Poynting’s Theorem in Linear Dispersive Media with Losses In the preceding section Poynting’s theorem (6.108) was derived with the restric- tion to linear media with no dispersion or losses (i.c., D = €E and B = wH), with and y real and frequency independent. Actual materials exhibit dispersion and losses. To discuss dispersion it is necessary to make a Fourier decomposition in time of both E and D (and B and H). Thus, with E(x, 1) = E dw E(x, we" D(x, 1) — fp do D(x, wen"the assumption of linearity (and, for simplicity, isotropy) implies that D(x, w) = €(w)E(x, w), where €(w) is the complex and frequency-dependent susceptibility. Similarly, B(x, @) = 4(@)H(x, @). The reality of the fields implies that E(x, —«) = E*(x, w), D(x, —@) = D*(x, w), and e(—@) = e*(w). The presence of dispersion carries with it a temporally nonlocal connection between D(x, t) and E(x, £), discussed in detail in Section 7.10. As a consequence, the term E + (aD/dt) in (6.105) is not simply the time derivative of (E + D/2). We write out E - (D/dt) in terms of the Fourier integrals, with the spatial dependence implicit, aD B-T= [de | dune%(w')[-iwe(w)] + E(w)e oeSplit the integrand into two equal parts and in one make the substitutions, @— —', o' > — , and use the reality constraints to obtain @ - 1 E- iT dw J do'E*(«')[-iwe(w) + iv'e*(o")] - E(w)e (6.124) We now suppose that the electric field is dominated by frequency components in a relatively narrow range compared to the characteristic frequency interval over which e() changes appreciably. We may then expand the factor iw'e*(w') in the square brackets around w' = @ to get [ = 2w Im €(w) — (w - 0’) 4 (we¥()) + +++Insertion of this approximation into (6.124) leads to aD a7 dw { do! EX(o') + E(w)o Im e(w)e"°™ é (6.125) + 22 do [ ae E*(o') + Blo) foet(o)e te There is a corresponding expression for H - dB/dt with E — H and e > » on the right-hand side.First of all note that if «and pare real and frequency independent we recover the simple connection between the time derivative terms in (6.105) and du/dt, with w given by (6.106). Second, the first term in (6.125) evidently represents the conversion of electrical energy into heat (or more generally into different forms of radiation*), while the second term must be an effective energy density. A more transparent expression, consistent with our assumption of the dominance of E and H by a relatively narrow range of frequencies can be obtained by supposing that E = E(t) cos(wpt + a), H = A(t) cos(wot + 8), where E(’) and A(t)are slowly varying relative to both 1/wp and the inverse of the frequency range over which e(w) changes appreciably. If we substitute for the Fourier transforms E(@) and H(w) and average both sides of the sum of (6.125) and its magnetic counterpart over a period of the “carrier” frequency wo, we find (after some straightforward manipulation), aD B\ _ ; (e FP +H- 4) = wy Im e(w)(E(x, ft) + E(x, t)) (6.126a) + oy Im (Hx, 1) + HU, 0) +where the effective electromagnetic energy density is Moe esr = 5 Re [see don) dw (on fee 1) - E(x, ))) (6.126b) +3 Re| (os) foe 1) H(x. ))The presence of the factors d(we)/dw and d(wp)/dw was first noted by Brillouin (see Brillouin, pp. 88-93). Our treatment is similar to Landau and Lifshitz, Elec- trodynamics of Continuous Media (Section 80). Poynting’s theorem in these circumstances reads Mel sy. S = —J-E — ay Im e(wo)(E(X, 1) + E(x, 1) a , , (6.127) —@» Im u(w)(H(x, t) + H(x, 1) The first term on the right describes the explicit ohmic losses, if any, while the next terms represent the absorptive dissipation in the medium, not counting con- duction loss. If the conduction current contribution is viewed as part of the di- electric response (see Section 7.5), the —J+E term is absent. Equation (6.127) exhibits the local conservation of electromagnetic energy in realistic situations where, as well as energy flow out of the locality (V-S # 0), there may be losses from heating of the medium (Im € # 0, Im pu # 0), leading to a (presumed) slow decay of the energy in the fields.Poynting’s Theorem for Harmonic Fields; Field Definitions of Impedance and Admittance* Lumped circuit concepts such as the resistance and reactance of a two-terminal linear network occur in many applications, even in circumstances where the size of the system is comparable to the free-space wavelength, for example, for a resonant antenna. It is useful therefore to have a general definition based on field concepts. This follows from consideration of Poynting’s theorem for harmonic time variation of the fields. We assume that all fields and sources have a time dependence e~™, so that we write E(x, t) = Re[E(x)e~'] = S[E(ne" + E*(x)e"“"] (6.128)The field E(x) is in general complex, with a magnitude and phase that change with position. For product forms, such as J(x, t) « E(x, f), we have S(x, 1) + Ew, 2) = H[I(xe "+ F*(x)e'"] « [E(xe" + E*(x)e"] (6.129) = } Re[J*(x) + E(x) + I(x) « E(x)e77'] For time averages of products, the convention is therefore to take one-half of the real part of the product of one complex quantity with the complex conjugate of the other. For harmonic fields the Maxwell equations become V-B=0, Vx E — ioB =0 6.13 V-D=p, VxH+ioD=J (6.130)where all the quantities are complex functions of x, according to the right-hand side of (6.128). Instead of (6.103) we consider the volume integral 1 i 6B x ah) FE dx whose real part gives the time-averaged rate of work done by the fields in the volume V. In a development strictly paralleling the steps from (6.103) to (6.107). we have tf tf > *. By = = . * — igD*) a. J) FE ade = 5 | B[V x He — ioD*] dx wn) a3 [ey ex - ioe De - BHO) axWe now define the complex Poynting vector E x H*) (6.132) and the harmonic electric magnetic energy densities, w. = }(E-D*), — w,, = 3(B+ H*) (6.133) Then (6.131) can be written as Sf agate + aie { (We = Wy) Bx +f s-mda =0 (6.134) This is the analog of (6.107) for harmonic fields. It is a complex equation whose real part gives the conservation of energy for the time-averaged quantities and whose imaginary part relates to the reactive or stored energy and its alternating flow. If the energy densities w, and w,, have real volume integrals, as occurs for systems with lossless dielectrics and perfect conductors, the real part of (6.134) is I 4 Re(J* + E) d°x + ¢ Re(S + n) da = 0 showing that the steady-state, time-averaged rate of doing work on the sources in V by the fields is equal to the average flow of power into the volume V through the boundary surface S, as calculated from the normal component of Re S. This is just what would be calculated from the earlier form of Poynting’s theorem (6.107) if we assume that the energy density w has a steady part and a harmoni- cally fluctuating part. With losses in the components of the system, the second term in (6.134) has a real part that accounts for this dissipation. The complex Poynting theorem (6.134) can be used to define the input im- pedance of a general, two-terminal, linear, passive electromagnetic system. We imagine the system in the volume V surrounded by the boundary surface S, with only its input terminals protruding, as shown in Fig. 6.4. If the complex harmonic input current and voltage are J, and V;, the complex power input is $/7V,. ThisFigure 6.4 Schematic diagrams of arbitrary, two-terminal linear, passive electromagnetic systems. The surface S completely surrounds the system; only the input terminals protrude. At these terminals, the harmonic input current and voltage are J, tnd Vj, with the input impedance Z defined by V; ~ 21. The upper diagratn upplies at low frequencies where radiation losses are negligible, while the lower one with its coaxial-line input permits discussion of radcan be written in terms of the Poynting vector by using (6.134) applied to all of space on the outside of S as iV, = -$ S-nda (6.135) where the unit normal n is outwardly directed, as shown in Fig. 6.4, and we have assumed that the input power flow is confined to the area S; (the cross section of the coaxial line in the lower diagram of Fig. 6.4). By now considering (6.134) for the volume V surrounded by the closed surface S, the right-hand side of (6.135) can be written in terms of integrals over the fields inside the volume V:1 siV; = a ROE dx + rie { (We = Wm) Bx + 2Jv v f. _ Srnda (6.136) The surface integral here represents power flow out of the volume V through the surface S, except for the input surface S;. If the surface (S — S;) is taken to infinity, this integral is real and represents escaping radiation (see Chapter 9). At low frequencies it is generally negligible. Then no distinction need be made between S, and S; the upper diagram in Fig. 6.4 applies. The input impedance Z = R — iX (electrical engineers please read as Z = R + jX!) follows from (6.136) with its definition, V; = Z/,, Its real and imaginary parts are at i ‘ 3 $ i _ 3 R= rie {Re pV Ed'x +29 Senda + dolm | (wy, — w.) dx (6.137)ie {+ Re [ (Wm — we) dx = Im I BOE ex} (6.138) In writing (6.137) and (6.138) we have assumed that the power flow out through Sis real. The second term in (6.137) is thus the “radiation resistance,” important at high frequencies. At low frequencies, in systems where ohmic losses are the only appreciable source of dissipation, these expressions simplify to R= i f o |EP dx (6.139) x iw f (Wm — We) Bx (6.140)Here cis the real conductivity, and the energy densities w,, and w, (6.133) are also real over essentially the whole volume. The resistance is clearly the value expected from consideration of ohmic heat loss in the circuit. Similarly, the re- actance has a plausible form: if magnetic stored energy dominates, as for a lumped inductance, the reactance is positive, etc. The different frequency de- pendences of the low-frequency reactance for inductances (X = wL) and capac- itances (X = —1/wC) can be traced to the definition of L in terms of current and voltage (V = L di/dt) on the one hand, and of C in terms of charge and voltage (V = QIC) on the other. The treatment of some simple examples is left to the problems at the end of the chapter, as is the derivation of results equivalent to (6.139) and (6.140) for the conductance and susceptance of the complex admittance Y.Transformation Properties of Electromagnetic Fields and Sources Under Rotations, Spatial Reflections, and Time Reversal The fact that related physical quantities have compatible transformation prop- erties under certain types of coordinate transformation is so taken for granted that the significance of such requirements and the limitations that can be thereby placed on the form of the relations is sometimes overlooked. It is useful therefore to discuss explicitly the relatively obvious properties of electromagnetic quanti- ties under rotations, spatial inversions, and time reversal. The notions have direct application for limiting phenomenological constitutive relations, and are applied in the next section where the question of magnetic monopoles is discussed. This assumed that the idea of space and time coordinate transformations and their relation to the general conservation laws is familiar to the reader from classical mechanics (see, e.g,, Goldstein). Only a summary of the main results is given here.A. Rotations A rotation in three dimensions is a linear transformation of the coordinates of a point such that the sum of the squares of the coordinates remains invariant, Such a transformation is called an orthogonal transformation. The transformed coordinates x, are given in terms of the original coordinates xg by x6 =D aapxXp (6.141) P The requirement to have (x’) = (x)’ restricts the real transformation coefficients Gqg to be orthogonal, D ange, = day (6.142) The inverse transformation has (a~')4, = dg, and the square of the determinant of the matrix (a) is equal to unity. The value det(a) = +1 corresponds to a proper rotation, obtainable from the original configuration by a sequence of infinitesimal steps, whereas det(a) = —1 represents an improper rotation, a reflection plus a rotation. Physical quantities are classed as rotational tensors of various ranks depend- ing on how they transform under rotations. Coordinates x;, velocities v;, momenta P, have components that transform according to the basic transformation law (6.141) and are tensors of rank one, or vectors. Scalar products of vectors, such SX; + Xp OF Vj + Pp, are invariant under rotations and so are tensors of rank zero, or scalars. Groups of quantities that transform according to 0 =D dante Bs 6.143)are called second-rank tensors or, commonly, tensors. The Maxwell stress tensor is one such group of quantities. Higher rank tensor transformations follow obviously. In considering electromagnetic fields and other physical quantities, we deal with one or more functions of coordinates and perhaps other kinematic variables. There then arises the choice of an “active” or a “passive’”’ view of the rotation. We adopt the active view—the coordinate axes are considered fixed and the physical system is imagined to undergo a rotation. Thus, for example, two charged particles with initial coordinates x; and x form a system that under a rotation is transformed so that the coordinates of the particles are now x} and x}, as shown in Fig. 6.5. The components of each coordinate vector transform according to (6.141), but electrostatic potential is unchanged because it is a func- tion only of the distance between the two points, R = |x, — x,|, and R? is a sum of scalar products of vectors and so is invariant under the rotation. The electro- static potential is one example of a scalar under rotations. In general, if a physical quantity , which is a function of various coordinates denoted collectively by x; (possibly including coordinates such as velocities and momenta), is such that when the physical system is rotated with x, —> x/, the quantity remains unchanged, $'(xi) = o(x) (6.144) then ¢ is a scalar function under rotations. Similarly, if a set of three physical quantities V,(x;) (a = 1, 2, 3) transform under rotation of the system according to Vas DY aaVa(%,) (6.145) B then the V, form the components of a vector, and so on for higher rank tensors.Figure 6S Active rotation of a system of two charges. Differential vector operations have definite transformation properties under rotations. For example, the gradient of a scalar, V¢, transforms as a vector, the divergence of a vector, V+ V. isa scalar, and the Laplacian operator V* is a scalar ‘operator in the sense that its application to a function or set of functions does not alter their rotational transformation properties. ‘Special mention must be made of the cross product of two vectors: A=BxC (6.146) In component form this vector shorthand reads Aa = & éapy Ba Cy where éyay = +1 for a = 1, 6 = 2, y= 3 and eyelic permutations, é,y, = —I for ‘ther permutations, and vanishes for two or more indices equal. Because of the presence of two vectors on the right-hand side, the cross product has some at- tributes ofa traceless antisymmetric second-rank tensor. Since such a tensor has only three independent components, we treat it as @ vector. This practice 18 jus tified, of course, only insofar as it transforms under rotations according to (6.141). In actual fact, the transformation law for the cross product (6.146) is Az = det(a) Daas (6187) For proper rotations, the only kind we have considered so far, det(a) = +1; thus (6.147) is in agreement with the basic coordinate transformation (6.141). Under proper rotations, the cross product transforms as a vector.D. Electromagnetic Quantities Just as with the laws of mechanics, it is true (i.c., consistent with all known experimental facts) that the forms of the equations governing electromagnetic phenomena are invariant under rotations, space inversion, and time reversal. This, implies that the different electromagnetic quantities have well-defined transfor- mation properties under these operations. It is an experimental fact that electric charge is invariant under Galilean and Lorentz transformations and is a scalar under rotations. It is natural, convenient, and permissible to assume that charge