# Maxwell’s Equations in Differential Form

v r r v ∂B r ∇× E = − − M i = −M d − M i ∂t v v v v ∂D v v v ∇ × H = J i + σE + = Ji + Jc + J d ∂t v ∇ ⋅ D = ρ ev v ∇ ⋅ B = ρ mv v v v v v v ∂B ∂D Jd = Md = , J c = σE , ∂t ∂t

v E≡ v B≡

Electric field intensity [V/m] Magnetic flux density [Weber/m2 = V s/m2 = Tesla]

v M i ≡ Impressed (source) magnetic current density [V/m2]

v M d ≡ Magnetic displacement current density [V/m2] v H≡ v Ji ≡ v D≡ v Jc ≡
Magnetic field intensity [A/m] Impressed (source) electric current density [A/m2] Electric flux density or electric displacement [C/m2] Electric conduction current density [A/m2]

v J d ≡ Electric displacement current density [A/m2]

ρ ev ≡ Electric charge volume density [C/m3] ρ mv ≡ Magnetic charge volume density [Weber/m3] σ ≡ Electric conductivity [1/Ω m]

1

Remarks:

v 1. Impressed magnetic current density ( M i ) and magnetic charge density ( ρ mv ) are unphysical quantities introduced through “generalized” current to balance Maxwell’s equations. v v 2. Although unphysical, M i , ρ mv similar to J i and ρ ev can be considered as energy sources that generate the fields. v Jc

v Ji
Current source

v v ∂D Jd = dt v J i generates the electric displacement current

v 3. Through “equivalent principle” M i and ρ mv can be used to simplify the solutions to some boundary value problems. v v ∂B 4. M d = (magnetic displacement current density [V/m2]) is introduced ∂t v v ∂D analogous to J d = (electric displacement current density [A/m2]) ∂t

2

v v v v v v v ∂B v ∂ v − M i ⇒ ∫∫ ∇ × E ⋅ ds = − ∫∫ B ⋅ ds − ∫∫ M i ⋅ ds • ∇× E = − ∂t ∂t S S S 3 .Integral form of Maxwell’s Equations Elementary vector calculus: S C Stokes’ Theorem: v ∫∫ (∇ × A) ⋅ ds = ∫ A ⋅ dl S C v v v • It says that if you want to know what is happening in the interior of a surface bounded by a curve just go around the curve and add up the field contributions. divergence theorem states that if you want to know what is happening within a volume of V just go around the surface S (bounding volume V ) and add up the field contributions. volume V surface S Divergence Theorem: v ∫∫∫ (∇ ⋅ A) dV = ∫∫ A ⋅ ds V S v v • In simple words. Null Identities: v v v v ∇ ⋅ (∇ × A) = 0 ⇔ ∇ ⋅ B = 0 ⇔ B = ∇ × A v v ∇ × (∇φ ) = 0 ⇔ ∇ × E = 0 ⇔ E = −∇φ (electrostatic) • The Divergence and Stokes’ theorems can be used to obtain the integral forms of the Maxwell’s Equations from their differential form.

S C v v v v v ∂ v ⇒ ∫ E ⋅ dl = − ∫∫ B ⋅ ds − ∫∫ M i ⋅ ds ∂t S C S v v v v v v v v v v v ∂D ∂D v • ∇× H = + σE + J i ⇔ ∫∫ ∇ × H ⋅ ds = ∫∫ ⋅ ds + ∫∫ σE ⋅ ds + ∫∫ J i ⋅ ds ∂t ∂t S S S S v v ∂ v v v v v v ⇒ ∫ H ⋅ dl = ∫∫ D ⋅ ds + ∫∫ J c ⋅ ds + ∫∫ J i ⋅ ds ∂t S C S S v v Where J c = σ E v v v v • ∇ ⋅ D = ρ ev ⇔ ∫∫∫ ∇ ⋅ D dv = ∫∫∫ ρ ev dv ⇒ ∫∫ D ⋅ ds = ∫∫∫ ρ ev dv = Qe v • ∇ ⋅ B = ρ mv ⇔ v v ∫∫∫ ∇ ⋅ Bdv = ∫∫∫ ρ mv dv ⇒ ∫∫ B ⋅ ds = Qm = ∫∫∫ ρ mv dv V V V V V S V S v v Since Qm = 0 ⇒ ∫∫ B ⋅ ds = 0 S V 4 .

v ∇×M = C v v v and M n (the normal component of M on the boundary) is also known. dt dt However such an approach to classical electromagnetism will be unnecessarily cumbersome. Newtonian mechanic is formulated in terms of force ( F ) and torque ( τ ). 5 . then M is uniquely defined. The question is then whether or not such a description (in terms of curls and divergences) is sufficient and unique? The answer to this question is provided by Helmholtz Theorem • A vector field is determined to within an additive constant if both its divergence and its curl are specified everywhere. τ = where L is the angular momentum. Instead. that is if: S and C are known and given by v ∇ ⋅ M = Sv.Helmholtz Theorem r r • Traditionally. v v ρ ∂ v Ex.: ∇ × E = − B and ∇ ⋅ E = ev ∂t ε Irrotational & Solenoidal Fields (Use of Helmholtz Theorem) Definition: • A field is irrotational if its curl is zero v ∇ × Fi = 0 ≡ Fi is irrotational • A field is solenoidal (divergenceless) if its divergence is zero v v ∇ ⋅ Fs = 0 ≡ Fs is solenoidal Theorem: • A vector field which its divergence and curl vanishes at infinity can be written as the sum of an irrotational & a solenoidal fields. r r r r dP r dL F= . the description of electromagnetics starts with Maxwell’s equations which are written in terms of curls and divergences. • Equivalent statement: A vector field is uniquely specified by giving its divergence and its curl v within a region and its normal component over the boundary. Remark: Helmholtz’s theorem allows us to appreciate the importance of the Maxwell’s v v equations in which E and H are defined by their divergence and curl.

the vector field M can be written as v v v (1) M = Fi + Fs v v v • Since Fi is irrotational then ∇ × Fi = 0 ⇒ Fi = −∇φ where φ is a scalar function. χ e are dimensionless. respectively. χ m . • Index of refraction is defined as (3) n = ε r μ r n ≡ index of refraction or phase index [#] • If we are mostly concerned with non-magnetic materials then μ r ≈ 1 ⇒ μ = μ0 ⇒ n = ε r Polarization and Magnetization v v v v •vPolarization vector P and magnetization vector M are related to D and E and v B and H according to: 6 .85 ×10-12 [F/m] ε r ≡ Relative permittivity or dielectric constant [#] μ ≡ permeability [H/m] μ 0 ≡ free space permeability = 4π × 10-7 [H/m] μ r ≡ relative permeability [#] • We also write (1) μ r = 1 + χ m (2) ε r = 1 + χ e Where χ m and χ e are the magnetic and electric susceptibility.v • According to the theorem stated above. v v v v v v • Since Fs is solenoidal then ∇ ⋅ Fs = 0 ⇒ Fs = ∇ × A then (1) ⇒ M = −∇φ + ∇ × A Constitutive Relations v D=ε v B=μ ε = εr v E v H ε0 μ = μ r μ0 ε ≡ permittivity [F/m] ε o ≡ vacuum permittivity = 8.

e. it can still be a scalar function. then: v v v v v v v D = ε 0 E + P = ε 0 E + ε 0 χ e E = ε 0 (1 + χ e )E = ε 0 ε r E ⇒ v v D = εE v v • Assuming M = χ m H then v v v v v B = μ 0 H + μ 0 M = μ 0 (1 + χ m )H = μ o μ r H ⇒ v v B = μH • ε and μ describe the macroscopic response of the media. media is non-homogeneous. ε E . Note that in this case even though permittivity is a function of the filed strength. Linear vs. Homogeneous vs.. We then write: v (1) D = ε E where ⎡ε xx ε xy ε xz ⎤ ⎡ E x ⎤ ⎡ Dx ⎤ ⎥ ⎢ D ⎥ = ε ⎢ε (2) ⎢ y ⎥ ε yy ε yz ⎥ ⎢ E y ⎥ 0 ⎢ yx ⎢ ⎥ ⎢ε zx ε zy ε zz ⎥ ⎢ E z ⎥ ⎢ Dz ⎥ ⎣ ⎦ ⎣ ⎦⎣ ⎦ v • If ε depends on the magnitude of the applied field.v v v (4) D = P + ε 0 E v v v v v (5) B = μ 0 H + μ 0 M = μ 0 (H + M ) v v • Assuming P = ε 0 χ e E . v v • If ε depends on the direction of the applied field. Anisotropic. i. ε (r ) . ε characterizes the electric response while μ describes the magnetic response. we say medium is ( ) nonlinear. • In the case of anisotropic medium ε is a tensor (for our purposes a matrix of rank 2). Non-Linear • If ε depends on position. D and E are not co-linear. 7 . • An example of non-linear medium is when −1 / 2 ε ⎡ 1 2 2 ⎤ = ⎢1 + 2 (c B − E 2 )⎥ . Inhomogeneous. i. Isotropic vs.e. then the medium is said to be anisotropic. (3) ε0 ⎣ b ⎦ Where b is the maximum field strength. i. In the following we assume our medium is nonmagnetic.e.

(3) So far. we may note that in the orbits of electrons in atoms. there has been no experimental evidence of the existence of this kind of classical nonlinearities. yet ordinary quantum theory with linear superposition is sufficient to describe the observed phenomena with a high degree of accuracy. Spatial dispersion means that the parameters describing the medium response (e. For heavier atoms. i. ε and μ ) are functions of time derivatives. ε and μ ) are functions of space derivatives. ε (ω ) . (1) The theory suffers from arbitrariness in the manner in which the nonlinearities occur. This ε ⎡ 1 ⎤ = ⎢1 + 2 (c 2 B 2 − E 2 )⎥ classical non-linear effect is given by ε0 ⎣ b ⎦ • However.g.. its electromagnetic energy grows more and more and becomes unbounded.e. field strengths of 1011-1017 V/m are present. HW: Consider a hydrogen atom unexcited and in thermal equilibrium. field strength has an upper bound. there are few problems with Born & Infeld classical non-linear vacuum response. To avoid this infinite self-energy we can think that some saturation of field strength takes place.• Interesting thing about (3) is the fact that it describes the response of the vacuum. we say the medium is dispersive (frequency dispersion) ω p2 ε = 1+ 2 (1) ε r = ε0 ω0 + j γ ω − ω 2 • Note that from (1) we can write (2) ε = ε ′ − jε ′′ • Remarks: Temporal dispersion means that the parameters describing the medium response (e.g.e. these fields can be even as large as 1021 V/m at the edge of the nucleus. Calculate the magnitude of the electric field due to its nucleus at the site of its electron. we say the medium is simple. homogeneous. Infinite Self Energy • A charge particle can be thought as the localization of the charge density. 1/ 2 Temporal dispersion • If ε depends on frequency. (2) There are problems with transitions to the quantum domain. and isotropic. • As to the last point. As a charge distribution localizes to a point charge. Electric Field • Electric field due to a point charge in origin 8 . • If a medium is linear. i. (proposed by Born & Infeld) in order to address the problem of vacuum infinite selfenergy.

v v v ˆ q1 a r q1 r q1 r E= where v = v = 4πε 0 r 2 4πε 0 r 3 4πε 0 r 3 v v r r ˆ a r = v = and we use the shorthand r r v notation r = r . ρ v (r ′) = ρ v ( x ′. y ′. x z ˆ ar q1 v r A Observation Point y • Electric filed due to a point charge not at the origin v v v v ˆ q1 aR q1 R q1 r − r ′ E= v = v = v v 4πε 0 R 2 4πε 0 R 3 4πε 0 r − r ′ 3 z v r Observation Point A v v v R = r − r′ v q1 v r′ y x • Superposition principle ⎡ ˆ ⎤ v ˆ ˆ 1 ⎢ aR1 a a E= q1 v 2 + q2 vR 22 + q3 vR 32 + L⎥ 4πε 0 ⎢ R1 ⎥ R2 R3 ⎣ ⎦ v v v ⎡ ⎤ 1 ⎢ R1 R2 R3 = q1 v + q2 v 3 + q3 v 3 + L⎥ 4πε 0 ⎢ R1 3 ⎥ R2 R3 ⎣ ⎦ v v v v v v (r − r ′) (r − r ′) ⎤ 1 ⎡ (r − r1′) = ⎢q1 v v 3 + q2 v v 2 3 + q3 v v33 + L⎥ 4πε 0 ⎢ r − r1′ r − r2′ r − r3′ ⎥ ⎣ ⎦ v v N r − r′ 1 E= ∑ qk r − r ′k 3 v v 4πε 0 k =1 k A Observation Point z v r v R1 v v R2 R3 v r1′ v q1 y v r3′ x v r2′ v q2 v q3 Electric Field & Potential due to Continuous Charge Distribution v ′ ′ • Volume charge density. z ′) 9 .

ρ s (r ′) = ρ s ( x′.z v v 1 R ′ dE = v ρ v dv ′ 4πε 0 R 3 v v v 1 r − r′ ′ dE = v v ρ v dv ′ 4πε 0 r − r ′ 3 x v v v 1 r − r′ ′ E = ∫∫∫ v v ρ v dv ′ 4πε 0 r − r ′ 3 v′ ′ ′ ′ 1 ρ v dv′ 1 ρ v dv′ 1 ρ v dv′ = ∫∫∫ V = ∫∫∫ v = ∫∫∫ v v 4πε 0 R 4πε 0 R 4πε 0 r − r ′ v′ v′ v′ A (Observation point) v r v r′ v v v R = r − r′ y Differential volume charge ′ density ρ v HW: From the above can you guess what forms the potentials due to a single or a collection of charges must have. y′. v ′ ′ • Surface charge density. ρ l′(r ′) = ρ l′( x′. y′. z′) z v r v r′ v v v R = r − r′ Differential line charge ρ l′ y x 10 . z ′) z v r v r′ x A (Observation point) v v v R = r − r′ y v v ˆ 1 aR 1 R 1 ′ ′ E = ∫∫ v 2 ρ s ds′ = ∫∫ v 3 ρ s ds′ = 4πε 0 R 4πε 0 R 4πε 0 s′ s′ ∫∫ s′ v v r − r′ ′ v v 3 ρ s ds′ r − r′ Differential surface charge ′ density ρ s V = 1 4πε 0 ∫∫ s′ ′ ρ s ds′ v R = 1 4πε 0 ∫∫ s′ ′ ρ s ds′ R = 1 4πε 0 v v ∫∫ r − r ′ s′ ′ ρ s ds′ v v • E and V due to a line charge.

v E= V = 1 4πε 0 1 4πε 0 ∫ l′ ˆ aR 1 v 2 ρ l′ dl ′ = 4πε 0 R ∫ l′ v R 1 v 3 ρ l′ dl ′ = 4πε 0 R = 1 4πε 0 l′ ∫ l′ v v r − r′ v v 3 ρ l′ dl ′ r − r′ ∫ l′ ρl′ dl ′ v R = 1 4πε 0 ∫ l′ ρ l′ dl ′ R v v ∫ r − r′ ρl′ dl ′ Electric Field of a Dipole z +q v d v r1 A v r2 v r y −q x v v r2 ⎤ q ⎡ r1 ⎢v − v ⎥ 4πε 0 ⎢ r1 3 r2 3 ⎥ ⎦ ⎣ v v v d r1 = r − 2 v v v d r2 = r + 2 Then ⎡ v v v q ⎢ r −d 2 E= − v v 4πε 0 ⎢ r − d 2 3 ⎣ v E= v v ⎤ r +d 2 ⎥ v v 3 r +d 2 ⎥ ⎦ v v⎤ v v −3 v −3 ⎡ 3 r ⋅ d r − d 2 ≈ r ⎢1 + v 2 ⎥ ⎢ 2 r ⎥ ⎦ ⎣ v v v v − 3 v −3 ⎡ 3 r ⋅ d ⎤ r + d 2 ≈ r ⎢1 − v 2 ⎥ ⎢ 2 r ⎥ ⎦ ⎣ v v v v v q v −3 ⎡ v v ⎛ 3 r ⋅ d ⎞ v v ⎛ 3 r ⋅ d ⎞ ⎤ E≈ r ⎢ r − d 2 ⎜1 + v 2 ⎟ − r + d 2 ⎜ 1 − v 2 ⎟ ⎥ ⎜ 2 r ⎟ ⎜ 2 r ⎟⎥ 4πε 0 ⎢ ⎝ ⎠ ⎝ ⎠⎦ ⎣ v −3 ⎡ v v v qr r ⋅ d v v⎤ ⎢3 v 2 r − d ⎥ E= 4πε 0 ⎢ r ⎥ ⎣ ⎦ v r Recall qd = p then ( ) ( ) 11 .

φ )aφ v ˆ ˆ ˆ A = Ax ( x. θ . y. then let us express E in the spherical coordinate v E= 1 v ˆ ˆ ˆ A = Ar (r . z )a x + Ay ( x. y. φ )a r + Aθ (r . θ . z )a z ⎡ Ar ⎤ ⎡ sin θ cos φ ⎢ ⎥ ⎢ ⎢ Aθ ⎥ = ⎢cos θ cos φ ⎢ Aφ ⎥ ⎢ − sin φ ⎣ ⎦ ⎣ x = r sin θ cos φ y = r sin θ sin φ z = r cos θ ⎡ pr ⎤ ⎡ cos θ p z ⎤ r ⎢ ⎥ ⎢ ⎥ ˆ ˆ ˆ ⎢ pθ ⎥ = ⎢− sin θ p z ⎥ ⇒ p = p az = (cos θ ar − sin θ aθ ) p ⎢ pφ ⎥ ⎢ ⎥ 0 ⎦ ⎣ ⎦ ⎣ • Or finally from v r ⎡ 3(r ⋅ p ) v r ⎤ v 1 E= v 3 ⎢ v 2 r − p⎥ 4πε 0 r ⎢ r ⎥ ⎦ ⎣ we get v ˆ ˆ ˆ 1 1 ⎡ 3r ar ⋅ (cosθ ar − sin θ aθ ) p ˆ ˆ ˆ E= r ar − (cos θ ar − sin θ aθ ) v3 ⎢ v2 4πε 0 r ⎢ r ⎣ v r p ˆ ˆ E= v 3 [2 cos θ ar + sin θ aθ ] . φ )aθ + Aφ (r . where r = r 4πε 0 r sin θ sin φ cos θ sin φ cos φ cos θ ⎤ ⎡ Az ⎤ − sin θ ⎥ ⎢ Ay ⎥ with ⎥⎢ ⎥ ⎥⎢ ⎥ 0 ⎦ ⎣ Az ⎦ ⎤ p⎥ ⇒ ⎥ ⎦ HW: Show that potential at point A for an electric dipole is given by +q A v r d r r ˆ ˆ p ⋅ ar p ⋅ ar V = v 2 = 4πε r 2 4πε 0 r 0 −q 12 . y. θ .v r ⎡ 3(r ⋅ p ) v r ⎤ v 3 ⎢ v 2 r − p⎥ 4πε 0 r ⎢ r ⎥ ⎣ ⎦ r ˆ • For our coordinate system p = p a z v v • r is the position vector in spherical coordinate. z )a y + Az (x.

13 .

We want to calculate the potential V [Volt] set up by this polarized dielectric. z v r A v R y v r′ x v′ Differential volume element dv′ • Potential due to a single dipole r ˆ p ⋅ aR V = 4πε 0 R 2 14 . r r dp • In differential form: P = dv′ Potential due to Bound (Polarized) Surface & Volume Charge Densities A dielectric of volume v ′ is polarized. (b) alignment of the permanent dipole moments.v Electric Polarization P NΔv′ k v P = limt k =1 Δv′→0 Δv′ r p [C·m] Electric dipole moment r P [C/m2] Electric polarization vector N [#/m3] is the number of dipoles per unit volume ∑p r r r • P [C/m2] is the volume density of electric dipole moment p [C·m] r v v v v Note P and D have the same units [C/m2]: D = ε 0 E + P r • Polarization vector P may come to exist due to (a) induced dipole moment. or (c) migration of ionic charges.

y ′. z ′) ˆ ∇f ( R ) = a R ∂f (R ) ∂R v 1 1 v v − ∇ 2 v = 4πδ 3 (R ) . or − ∇ 2 v v = 4πδ 3 (r − r ′) r − r′ R and furthermore r v v v v v ∇′ ⋅ ( f A) = f ∇′ ⋅ A + A ⋅ ∇′f ⇒ A ⋅ ∇′f = ∇′ ⋅ ( f A) − f ∇′ ⋅ A v r 1 * Let A = P and f = then R r r v⎞ ⎛P⎞ ⎛1⎞ ′ ⎛1 P ⋅ ∇′ ⎜ ⎟ dv = ∫∫∫ ∇′ ⋅ ⎜ ⎟ dv′ − ∫∫∫ ⎜ ∇′ ⋅ P ⎟ dv′ ∫∫∫ ⎜R⎟ R ⎝R⎠ ⎠ ⎝ ⎠ v′ v′ v′ ⎝ Use divergence theorem ⇒ v r v⎞ P r ⎛1⎞ ′ ⎛1 ∫∫∫ P ⋅ ∇′ ⎜ R ⎟ dv = ∫∫ R ⋅ ds′ − ∫∫∫ ⎜ R ∇′ ⋅ P ⎟ dv′ ⎝ ⎠ ⎠ v′ s′ v′ ⎝ The potential then can be written as 15 . z ) R y x source (x ′.r • An elemental electric dipole. r ˆ dp ⋅ aR will set up a differential potential dV = 4πε 0 R 2 r r But from our definition of polarization we dp = P dv′ ⇒ r r ˆ ˆ dp ⋅ aR P ⋅ aR dV = = dv′ 2 4πε 0 R 4πε 0 R 2 Total potential V is found by integrating the above: r ˆ P ⋅ aR 1 V = ∫∫∫ R 2 dv′ 4πε 0 v′ v2 2 2 2 Where R 2 = R = ( x − x ′) + ( y − y ′) + ( z − z ′) ˆ ⎛1⎞ a Note that ∇ ′⎜ ⎟ = R (see Remarks below) then 2 ⎝R⎠ R r 1 ⎛1⎞ V = ∫∫∫ P ⋅ ∇′ ⎜ R ⎟ dv′ 4πε 0 v′ ⎝ ⎠ Remarks: Few useful identities ˆ ⎛1⎞ a ∇ ′⎜ ⎟ = R 2 ⎝R⎠ R ˆ ⎛ 1⎞ −a ∇⎜ ⎟ = 2R R ⎝R⎠ v v R R ˆ ∇R = a R = v = R R z A ( x. y . having a differential electric dipole moment of dp [C·m].

i. we define r ˆ • Bound or polarized surface charge density: ρ sP = P ⋅ an [C/m2] r • Bound or polarized volume charge ρ vP = −∇ ⋅ P [C/m3] A polarized dielectric can be replaced by a bound (polarized) surface and volume charge densities ( ρ sP & ρ vP ). dropping the prim notation since we know that integration is carried with respect to the prim coordinate. ε0 • When a medium is polarized we must take into account the effects of the bound charge. The potential setup by these bound charges then can be calculated.r r ˆ′ ∇′ ⋅ P ⎤ 1 ⎡ P ⋅ an V = ds′ − ∫∫∫ dv′⎥ .: ′ 1 ρ ′ ds ′ 1 ρ v dv ′ and V = ∫∫∫ V = ∫∫ ⇒ 4πε 0 R 4πε 0 R s′ v′ r ′ ˆ′ P ⋅ an = ρ s r ′ − ∇′ ⋅ P = ρ v Or in general. ⎢ R 4πε 0 ⎣ ∫∫ R ′ S ⎦ ′ ˆ n is perpendicular to surface S ′ bounding volume v ′ .e. Where a Compare above to the previously obtained expressions for V due to surface and volume charge densities. hence r v ρ v ρ vp ρ v ∇ ⋅ P ∇⋅E = + = − ⇒ r v v r ∇ ⋅ ε 0 E + ∇ ⋅ P = ρv ⇒ ∇ ⋅ ε 0 E + P = ρv v v v Let’s define D = ε 0 E + P then v ∇ ⋅ D = ρ v ← Generalized Gauss’ Law r v v v v v r v • Also note that for D = ε 0 E + P if P = ε 0 χ e E then D = ε 0 (1 + χ e )E = ε 0 ε r E ε0 ε0 ε0 ε0 ( ) Where ε r = 1 + χ e then v v D = ε 0ε r E = ε E where ε = ε 0 ε r Magnetization & Permeability 16 . v v Generalized Gauss’ Law & Constitutive Relation D = ε E v ρ • In free space ∇ ⋅ E = v .

square.e. its shape can be circular. magnetization) when subjected to an applied magnetic field • This magnetization is the result of alignment of the magnetic dipoles of material with the applied magnetic field.. • For a material of volume Δv which contains N m magnetic dipoles (orbiting electrons) per unit volume. accurate qualitative and quantitative description can be found using simple atomic model (semi-classical) • The electron orbiting the nuclei can be though of as a small current loop of area dsi with current I i • As long as loop is small.e− ≡ ≡ dsi Ii dsi • The magnetic dipole moment is given by v ˆ dmi = ni I i dsi [A·m2]. Magnetic Dipole & Magnetic Dipole Moment • To accurately describe magnetic behaviour of materials quantum theory of matter is needed. i. a magnetic dipole. or any other closed curve e− - ˆ ni ˆ ni + e− . i. This is similar to electric polarization which is the result of alignment of electric dipoles of the material with the applied electric field. • The magnetic field of the current carrying loop at large distance is similar to the field of a linear bar magnet. magnetization ( M ) is given by 17 . However.r • Magnetic materials exhibit magnetic polarization ( M . ˆ where ni is perpendicular to the loop surface.e. the total magnetic moment is given by N m Δv N m Δv v v ˆ mt = ∑ dmi = ∑ ni I i dsi i =1 i =1 v • The magnetic polarization.

v ⎡ 1 N m Δv v ⎤ ⎡ 1 N m Δv ⎤ ⎡1 v⎤ ˆ M = limt ⎢ mt ⎥ = limt ⎢ dmi ⎥ = limt ⎢ ∑ ∑ ni I i dsi ⎥ [A/m] Δv → 0 ⎣ Δv ⎦ Δv → 0 ⎣ Δv i =1 ⎦ Δv → 0 ⎣ Δv i =1 ⎦ v • Note that magnetization ( M ) is the volume densityvof the total magnetic dipole v moment ( mt ). we writev v v v (2) B = μ 0 H a + μ 0 M = μ 0 (H a + M ) v v • However. Where μ r = 1 + χ m is the relative permeability and μ is the permeability. 18 . and also the fact that magnetization ( M ) has the same units as the v magnetic field intensity. Ba ) = I i dsi Ba sin Ψi • Subjected to the above torque. v However. v • In absence of an applied field ( Ba = 0 ) the magnetic dipoles point in random directions. when a magnetic material is present. Where χ m is a scalar (or tensor) function then we have v v v (4) B = μ0 [1 + χ m ] H a = μ0 μ r H a = μ H a . Ba ) = I i dsi Ba sin(∠ni . when Ba ≠ 0 . M is ultimately related to the applied field H a . a magnetic polarization ( M ) is also present and an additional term must be added to (1). v magnetic dipoles realign themselves such that their the v moment ( dmi ) is collinear with Ba (see figure in the next page) v Ψi → 0 ⇒ Δτ → 0 Remark: Comparing the similarities between the torque & potential energy for electric & magnetic dipoles r v v v v Δτ B = dm × Ba ΔU B = − dm ⋅ Ba r r v r v Δτ E = dp × Ea ΔU E = − dp ⋅ Ea • From next page figure we see that in absence of an applied magnetic field. If we assume v v (3) M = χ m H a . we can write v v (1) Ba = μ0 H a . In order to take into account the influence of the material. the dipoles will experience a torque given by v v v v v v v v v ˆ Δτ = dmi × Ba = dmi Ba sin(∠dmi . v But. H [A/m].

19 .

v • This bound magnetization surface current density ( J sm ) is responsible for producing the v magnetization ( M ). ρvP and ρ sP are the volume and ˆ surface bound charges. Conductivity (DC & AC) • Recall the time harmonic form of the Ampere’s Law: 20 . the magnetic moments line up with Ba to minimize the potential energy as shown in the figure. the currents of the loops within the interior part of the medium will cancel each v other and only a surface current ( J sm ) on the exterior of the slab remains. where P is the electric polarization. and a n is the normal to the surface. • In general the magnitude of the spin magnetic moment is ≈ ±9 × 10 −24 [A·m2]. and a n is the normal to the surface. • Only electrons in the atomic shell that are not completely filled will contribute to the spin magnetic moment. • There is also a magnetic moment associated with the nucleus spin. • So far we have only considered the magnetic moment of the orbiting electron. in the limit. which is approximately 1000 times less than the moment associated with the electron spin. J sm is the bound magnetization surface current density [A/m]. however. • Since the number of dipoles is very large and therefore they are closely packed. • Similarly. J vm is the bound ˆ magnetization volume current density [A/m2]. Remark: v The origin of magnetization ( M ) can also be visualized by the following: v v • When Ba ≠ 0 . a magnetic moment can also be assigned to the spin of electron. for magnetic field applied to a medium we have v v ˆ J sm = M × an v v J vm = ∇ × M v v • Here.Bound Magnetization Current Density • Recall that for an electric field applied to a medium we had r ˆ ρ sP = P ⋅ an r r ρ vP = −∇ ⋅ P .

selective heating of human tissue ..microwave cooking . where ωε ′ ⎠ ⎝ 21 . • σ a is due to “resistance” of the dipoles as they attempt to align (rotate) themselves with the applied field. i.e. which contributes to σ a is sometimes called dielectric hysteresis. • The phenomenon of dipole rotation.v v v v v v v ∇ × H = J i + J C + J D = J i + σ s E + jωε E • Where ε in general is complex: ε = ε ′ − jε ′′ and σ s is the static (DC) conductivity (this is due to free carriers. • Thev Ampere law then can be written as: v v v v v v v v v ∇ × H = J i + σ s E + jω (ε ′ − jε ′′)E = J i + (σ s + ω ε ′′)E + jωε ′E = J i + σ e E + jωε ′E where we have defined the followings: σ e = σ s + ω ε ′′ ≡ Equivalent (effective) conductivity [1/Ω·m] σ a = ωε ′′ ≡ Alternating (AC) conductivity [1/Ω·m] σs = ⎨ μe N e q (for conductors) = static (DC) conductivity [1/Ω·m] ⎩μe N e q + μ h N h q (for semiconductors) σ e = σ s + σ a = σ s + ωε ′′ ⎧ • Note σ s is due to free charges at ω = 0 (a signature of true conductors). • For good dielectrics such as glass or plastic σ s ≈ 0 . but these materials when exposed to alternating fields ( σ a ≠ 0 ) can dissipate large amount of energy. Example of large σ a and its application are: .removing sulfur from mineral coal to produce clean coal (selective heating) v v v v v v v • Note ∇ × H = J i + σ e E + jωε ′E = J i + J C + J d v J i ≡ Impressed current density v v v v J C ≡ σ e E = (σ s + σ a )E = (σ s + ωε ′′) E : Effective conduction current density v v J d ≡ jωε ′E : Effective displacement current density Loss Tangent • we can rewrite Ampere Law as: v v v σ ⎞v v ⎛ ∇ × H = J i + jωε ′⎜1 − j e ⎟ E = J i + jωε ′(1 − j tan δ e ) E . electrons at ω = 0 ).

where ωε ′ ⎠ ⎝ σ ⎞ σ +σa σ ⎛ ε c = ε ′⎜1 − j e ⎟ = ε ′ − j ( s ) = ε ′ − j ( s + ε ′′) ωε ′ ⎠ ω ω ⎝ • In the expression for ε c the free carrier losses and dielectric losses are clearly evident. where electrons are moving with v velocity v . DC Conductivity • Consider a small cylinder containing N electrons per unit volume. v v v σ ⎞v v ⎛ • To see this we write: ∇ × H = J i + jωε ′⎜1 − j e ⎟ E = J i + jωε c E . We could have also formulated the problem in terms of complex permittivity. • Remark: The presence of static conductivity as a separate mechanism of loss in addition to the dielectric loss ( ε ′′ ) can also be observed in the Kramers-Kronig relations which connects the real and imaginary parts of the dielectric constant. When a medium has static conductivity σ s then Kramers-Kronig relations are given by σ s 2ω +∞{Re[ε (ω ′)] ε 0 } − 1 − Im[ε (ω )] = P∫ dω ′ ω π ω ′2 − ω 2 0 ω ′ {Im[ε (ω ′)] ε 0 } Re[ε (ω )] = 1 + P ∫ dω ′ π 0 ω ′2 − ω 2 2 +∞ .σs : Static (DC) loss tangent ωε ′ ε ′′ tan δ a = : Alternating (AC) loss tangent ε′ • Manufacturer usually provides loss tangent or the conductivity. tan δ s = tan δ e ≡ Effective electric loss tangent σ σ σ σ ωε ′′ σ s ε ′′ = e = s + a = s + = + ωε ′ ω ε ′ ωε ′ ω ε ′ ω ε ′ ω ε ′ ε ′ = tan δ s + tan δ a with • Note that in the above discussion we have expressed the conduction (DC) and dielectric losses (AC) in terms of effective conductivity ( σ e ) or effective loss tangent ( tan δ e ). N ≡ Number of electrons per unit volume [1/m3] v v ≡ Velocity vector of electrons e ≡ Electron charge v v ˆ n ΔS ) v n ⋅ v Δt 22 . where P stands for the principle value integral.

which implies v ˆ ⇒ I = ∫∫ J ⋅ n ds ← This is our standard equation for calculating current from current density. v ˆ then we have ΔI = J ⋅ n ΔS Remark: v v ˆ ˆ • ΔI = J ⋅ n ΔS can be written as dI = J ⋅ n ds in differential form. where J is the current density vector [A/m2] and ΔQ Δt = ΔI .602 × 10−19 [C] v v ⇒J =qN μ E v v • Compare the above to J = σ s E ⇒ σ s = q N μ . however when positive charges (ions of holes) are present we must consider the contributions of both carriers to the conductivity. electron density. for electron e = −q = −1. v = − μ E . This implies ΔQ v ˆ = ΔI = N e ΔS n ⋅ v Δt v v v • We define N e v = J . where μ is called mobility v v [m2/V·s] (note E and v are anti-parallel) v E v v - v v v • Then J = N e v = − N e μ E . • In our analysis so far we have only considered the electrons. i. v • Let us assume a linear relationship between velocity ( v ) and v v v electric filed ( E ). they are valid when fields are: single valued.ˆ n ≡ Normal to the surface ΔV ≡ Volume of the cylinder • The total chare ( ΔQ ) contained within the volume ( ΔV ) is given by v ˆ ΔQ = N e ΔV . bounded. continuous.e. The static conductivity is then modified according to: σ s = q N e μe + q N h μh μ e ≡ Electron mobility μ h ≡ Hole mobility N e and N h are electron and holes densities [1/m3] Boundary Conditions • Maxwell’s equations in differential forms are point equations.. This says that static conductivity is the product of electron charge. 23 .e. and have continuous derivatives. where ΔV = ΔS n ⋅ v Δt hence v ˆ ΔQ = N e ΔS n ⋅ v Δt . and electron mobility. i.

v This all can be summarized as v ˆ n × (H 2 − H 1 ) = 0 y ˆ n C0 2 ε 2 . • Also the fact that we can carry the same analysis invthe y-z plane which results in v ˆ ˆ ˆ a z ⋅ (H 2 − H1 ) = 0 . • Note that: v ˆ a x ⋅ H 2 ≡ Tangential component of H 2 WRT the interface. σ 1 z S0 24 . μ1 . J i = 0) v v v v ∂ v v ∫ H ⋅ dl = ∫∫σ E ⋅ ds + ∂t ∫∫ D ⋅ ds (1) C0 S0 S0 Δy →0 Δy →0 ˆ n C0 2 ε 2 . hence to find the fields we must y rely on their integral form. (1) can be written as: v v v ˆ lim ∫∫ σE ⋅ ds = lim ∫∫ σE ⋅ dxdy a z Δy →0 C0 v ˆ = lim σEΔxΔy ⋅ az = 0 Δy →0 S0 Δy →0 [ ] S0 • The second term on the RHS of Eq. σ 2 ≠ ∞ ) and no sources on boundary v v (M i = 0.• When boundaries are present. μ1 . μ 2 .σ 2 1 x Δx Δy ε 1 . μ 2 . v v ˆ a x ⋅ H 1 ≡ Tangential component of H1 WRT the interface. r • Boundary conditions for tangential H : Assume finite conductivity (σ 1 . The conclusion is then the r following: tangential components of H are continuous across the boundary between two dielectrics. (1) can be written as: v v v ∂ v ∂ ∂ ˆ ˆ lim D ⋅ ds = lim ∫∫ D ⋅ dxdy az = lim ( D ⋅ ΔxΔy a z ) = 0 ∫∫ Δy →0 ∂t Δy →0 ∂t Δy →0 ∂t S S0 0 • Putting it all together: v v v v ˆ ˆ a x ⋅ (H1 − H 2 ) Δx = 0 ⇒ a x ⋅ (H 2 − H1 ) = 0 . the Left hand side (LHS) can be written as: v v v v v v lim ∫ H ⋅ dl = lim ∫ H1 ⋅ dl1 + ∫ H 2 ⋅ dl2 [ ] v v v v ˆ ˆ ˆ = H1 ⋅ Δx a x − H 2 ⋅ Δx a x = (H1 − H 2 )⋅ Δx a x • The first term on the right hand side (RHS) of Eq. fields are discontinuous. σ 1 z S0 • Taking the limit of the both sides of Eq. (1). with a z ⋅ H 2 and a z ⋅ H1 designating the tangential components of the r H fields.σ 2 1 x Δx Δy ε 1 .

μ1 .σ 2 1 x ε 1 .Boundary Condition on Normal Components (Not corrected) Medium (1) and (2) are non conductors (dielectrics) ( σ 1 . μ 2 . σ 1 z A0 ∫∫ D ⋅ ds = ∫∫∫ ρ dv v v v v LHS: v v v v v v v v ˆ ˆ lim ∫∫ D ⋅ ds = lim ∫∫ D2 ⋅ ds + ∫∫ D1 ⋅ ds = lim ∫∫ D2 ⋅ dxdza y − ∫∫ D1 ⋅ dxdza y Δy →0 Δy →0 Δy →0 v v ˆ ˆ = lim (D2 A0 a y − D1 A0 a y ) [ ] [ ] Δy →0 RHS: lim ∫∫∫ ρ v dv = lim [ρ v ΔyA0 ] = A0 lim ρ v Δy = A0 ρ s = 0 Δy →0 v Δy →0 Δy →0 Then v v v v v v ˆ ˆ ˆ (D2 − D1 )⋅ a y = 0 ⇔ n ⋅ (D2 − D1 ) = 0 ⇔ n ⋅ (ε 2 E 2 − ε 1 E1 ) = 0 25 . σ 2 ≠ ∞ ) and there are no sources at the boundary ρ es = ρ ms = 0 y ˆ n A0 2 ε 2 .

v v ˆ n ⋅ (D2 − D1 ) = 0 . μ1 . the v v electrodynamics assumption.e. since perfect conductor ⇒ E1 = D1 = 0 then r v ∂ v ∂ v ∇ × E1 = − B1 ⇒ 0 = B1 . • J s is the surface current due to the free charges on the metal (not the bound charges) 26 . v v ˆ n ⋅ (B2 − B1 ) = 0 Boundary Conditions for Two Media v which One Medium Is a Perfect Conductor in ( σ 1 = ∞ ). B1 = H1 = 0 v v v v ˆ ˆ • n × (E 2 − E1 ) = − M s ⇒ n × E 2 = 0 Electric filed has no tangential component on the boundary between perfect conductor and dielectric. Therefore. σ 1 z v v ˆ • n ⋅ (B2 − B1 ) = ρ ms ρ ms : Fictitious magnetic surface charge density [Weber/m2] Boundary Conditions Between to Perfect Dielectrics: v v ˆ n × (E 2 − E1 ) = 0 . ρ ms = 0 ): v v • In medium-1. But this means that B1 must be a constant function of time ∂t ∂t which contradicts the assumption of time varying electric and magnetic fields.σ 2 1 x ε 1 . v v ˆ n × (H 2 − H 1 ) = 0 . r v v v v ˆ ˆ • n × (H 2 − H1 ) = J s ⇒ n × H 2 = J s v v Tangential component of H is discontinuous by amount of surface current J s at the boundary between perfect conductor and dielectric. With no Sources Present ( M s = 0 .Summary of boundary conditions General Case: v v v ˆ • n × (E 2 − E1 ) = − M s M s : Fictitious magnetic current density [V/m] y ˆ n v v v ˆ • n × (H 2 − H 1 ) = J s v J s : Electric surface current density [A/m] v v ˆ • n ⋅ (D2 − D1 ) = ρ es ρ es : Electric surface charge density [C/m2] 2 ε 2 . i. μ 2 .

v v v ˆ ˆ • n ⋅ (D2 − D1 ) = ρ es ⇒ n ⋅ D2 = ρ es Electric field has only normal component on the boundary between perfect conductor and dielectric. v H 1t = 0 ) and no sources are present v ( ρ es = 0 . v v v ˆ ˆ • n ⋅ (B2 − B1 ) = ρ ms ⇒ n ⋅ B2 = 0 Magnetic field has no normal component on the boundary between perfect conductor and dielectric. i. J s = 0 ) v v v v • Here H 1 = 0 ⇒ B1 = 0 . v E v E metal Boundary Conditions Between Two Medium one of which Is a Perfect Magnetic Material (the medium has infinite magnetic conductivity.e. E1 = D1 = 0 v v v v v ˆ ˆ • n × (E2 − E1 ) = − M s ⇒ n × E2 = − M s Electric filed is tangential to the boundary v v v v ˆ ˆ • n × (H 2 − H1 ) = J s ⇒ n × H 2 = 0 Magnetic filed has no tangential component on the boundary v v v ˆ ˆ • n ⋅ (D2 − D1 ) = ρ es ⇒ n ⋅ D2 = 0 Electric filed has no normal component at the boundary v v v ˆ ˆ • n ⋅ (B2 − B1 ) = ρ ms ⇒ n ⋅ B2 = ρ ms Magnetic field is normal to the boundary 27 .