UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 19 Prof. Steven Errede
LECTURE NOTES 19
LORENTZ TRANSFORMATION OF ELECTROMAGNETIC FIELDS (SLIGHT RETURN)
Before continuing on with our onslaught of the development of relativistic electrodynamics via tensor analysis, I want to briefly discuss an equivalent, simpler method of Lorentz transforming the EM fields EB and from one IRF(S) to another IRF(S'), which also sheds some light (by contrast) on how the EM fields Lorentz transform vs. “normal” 4-vectors.
In P436 Lecture Notes 18.5 {p. 18-22} we discussed the tensor algebra method for Lorentz transformation of the electromagnetic field e.g. in the lab frame IRF(S), represented by the EM field strength tensor F v to another frame IRF(S'), represented by the EM field strength tensor Fv via the relation:
vv T FF n.b. in matrix form: FF F since is symmetric.
Analytically carrying out this tensor calculation by hand can be tedious and time-consuming. If such calculations are to be carried out repeatedly/frequently, we encourage people to code this up and simply let the computer do the repetitive work, which it excels at.
For 1-dimensional Lorentz transformations (only) there is a simpler, less complicated, perhaps somewhat more intuitive method. Starting with the algebraic rules for Lorentz- transforming { EB and } in one IRF(S) to { EB and } in another IRF(S') e.g. moving with relative velocity vvx ˆ with respect to IRF(S):
2 component(s): EEx x Bx Bx 11
components: EEcByyz ByyzBEc vc
EEcBzzy Bzzy BEc
We can write these relations more compactly and elegantly by resolving them into their and components relative to the boost direction: here, is along vvx ˆ and is perpendicular to v , defined as follows {n.b in general, v could be e.g. to xˆˆ,,yzˆ or r ˆ}:
EE vc EEvBEcB 11 2
B B 11 B BvEB2 E cc
Now since vvx ˆ {here} then: EE x , B Bx and: B Byyzˆ Bzˆ , E Eyyzˆ Ezˆ {and similarly for corresponding quantities in IRF(S')}.
© Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 1 2005-2015. All Rights Reserved. UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 19 Prof. Steven Errede
Since vvx ˆ and EE x then: vE vExx ˆ 0 And likewise, since B Bx then: vB vBxx ˆ 0 .
Thus, we can {safely} write: vE vE and vB vB, as long as v is always to one of the components of E and B - e.g. xˆˆ or yz or ˆ .
Then we can write the Lorentz transformation of EM fields as:
EE vc EEvBEcB vc B B 11 2 B BvEB2 E 11 cc
This can be written more compactly in 2-D matrix form as:
EM Fields: “Normal” 4-Vector: EE → EE and BB compare to x x cB cB EE10 EE x x ↔ cB 01 cB cB cB ct ct
Unit Matrix Operator Matrix Scalar Matrix
Thus, we see that for the EM fields vs. the 3-D space-part of a “normal” 4-vector, the vs. components are switched, B transforms “sort of” like time t, but 2 2 Lorentz boost matrices for ( EB and ) vs. 4-vectors are not the same (they are similar, but they are not identical).
We can also write compact inverse Lorentz transformations (e.g. from IRF(S') rest frame → IRF(S) lab frame):
EM Fields: “Normal” 4-Vector: EE → E EBB and compare to x x cB cB EE10 EE x x ↔ cB 01 cB cB cB ct ct
Unit Matrix Operator Matrix Scalar Matrix
2 © Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 2005-2015. All Rights Reserved. UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 19 Prof. Steven Errede For a general Lorentz transformation (i.e. no restriction on the orientation of v {arbitrary}):
A.) Lorentz transformation from IRF(S) → IRF(S'):
2 EEcB E vc vc 1 2 1 2 B BE B 11 c 1 Or: 1 EE 1 cB cB 1 1 operator matrix
B.) Inverse Lorentz transformation from IRF(S') → IRF(S):
2 EEcB E 1 Switch , E E and 2 1 B B in above relations B BE B c 1 Or: 1 EE 1 cB cB x 1 1 operator matrix
Electrodynamics in Tensor Notation
So now that we know how to represent the EM field in relativistic tensor notation (as FGvv or ), we can also reformulate all laws of electrodynamics (e.g. Maxwell’s equations, the Lorentz force law, the continuity equation {expressing electric charge conservation}, etc. . . ) in the mathematical language of tensors.
In order to begin this task, we must first determine how the sources of the EM fields – the electric charge density (a scalar quantity) and the electric current density J (a 3-D vector quantity) Lorentz transform.
The electric charge density QV = charge per unit volume (Coulombs/m3)
© Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 3 2005-2015. All Rights Reserved. UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 19 Prof. Steven Errede
Imagine a cloud of electric charge drifting by. Concentrate on an infinitesimal volume V containing charge Q moving at (ordinary) velocity u :
Then: QV = charge density (Coulombs/m3)
And: Ju = current density (Amps/m2).
A subtle, but important detail:
If there is only one species (i.e. kind / type) of charge carrier contained within the infinitesimal volume V, then all charge carriers travel at the same (average / mean) speed u .
However, if there are multiple species (kinds / types) of charge carriers (e.g. with different masses) and/or different signs of charge carriers contained within the the infinitesimal volume V 2 2 (e.g. electrons e with rest masses mec and protons p with rest masses mpc ) then the different constituents / species must be treated separately in the following:
If N species: th Current density Juiii for the i species (i = 1, . . ., N), the electric charge density ii QV NN And: JJiii u ii11 We also need to express and J in terms of the proper charge density 0 = volume charge density defined in the rest frame of the charge Q, IRF(S0).
The infinitesimal rest volume / proper volume = V0 {defined in the rest/proper frame IRF(S0)}
The proper charge density: QV ← Recall that electric charge Q (like c) is 00 a Lorentz invariant scalar quantity
Because the longitudinal direction of motion undergoes Lorentz contraction from the rest frame IRF(S0) in the Lorentz transformation → another reference frame, e.g. lab frame IRF(S)
1 1 u Then: VV 0 where: Vwd and: Vwd , where: and: 0000 u 2 u c u 1 u
If the Lorentz transformation is along (i.e. || to) the length, 0 of the infinitesimal volumes 1 Then: 0 and the components of the volumes are unchanged: ww0 , dd0 . u
1 QQ Then if: VV 0 → uu 0 Ju uu00 u u u VV0 d Recall that the 3-D vector associated with the proper velocity is: uu J 0 d
4 © Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 2005-2015. All Rights Reserved. UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 19 Prof. Steven Errede
dx0 dt The zeroth (i.e. temporal/scalar) component of the proper 4-velocity is: 0 cc ddu
The corresponding zeroth (i.e. temporal/scalar) component of the current density 4-vector J is:
00 Jcccc00uu 0
0 2 The current density 4-vector is: JJJcJcJJJ,,,,,x yz (SI units: Amps/m )
0 0 Then: J 000, where: ,,uuc u u cu , u cu ,,, xyz u u constant scalar quantity
J is a proper four vector, i.e. J = proper current density 4-vector.
Thus: JJ JJ is a Lorentz invariant quantity. Is it ???
2 22 1 u cu 222 JJJJ 2 222222 cuuu 2 c c c 00 uxyz 0uu22 0 0 1122 u2 cc
2 22 22 2 JJ JJ00 c 0 0 c Since: c
Yes, JJ JJ is a Lorentz invariant quantity!
The 3-D continuity equation mathematically expresses local conservation of electric charge (using differential vector calculus):
rt, Jrt, rt, = scalar point function, JJrt , = 3-D vector point function t
xˆˆ yz ˆ (in Cartesian coordinates) x yz
We can also express the continuity equation in 4-vector tensor notation: n.b. 3 i 00 repeated J J y J J 11 c JJ J x z and: ( Jc0 ) indices x yz xi tc t ct x0 implies i1 summation!
33JJii00 J J 3 JJ 0 0 Then: J J 0 ii00 = t t ii11xx x x 0 xx J Thus: J or J 0 0 Continuity equation (local charge conservation) t t x
© Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 5 2005-2015. All Rights Reserved. UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 19 Prof. Steven Errede
J Physically, note that is the 4-dimensional space-time divergence of the current density x J 4-vector JcJ , . The 4-current density JcJ , is divergenceless because 0 . x The 4-vector operator is called the 4-D gradient operator, (a.k.a the quad operator x or “quad” for short). However, because the 4-D gradient operator functions like a covariant x 4-vector, e.g. when it operates on contravariant J (or any other contravariant 4-vectors), it is often alternatively given the shorthand notation . Similarly, because the 4-D x gradient operator functions like a contravariant 4-vector, e.g. when it operates on x covariant J (or any other covariant 4-vectors), it is given the shorthand notation . x {See/work thru Griffiths Problem 12.55 (p. 543) for more details.}
1 The contravariant quad/gradient 4-vector operator: , xct
1 The covariant quad/gradient 4-vector operator: , xct
22 1 22 Then: 22 xx xx xx ct = D’Alembertian 4-vector operator = 4-D Laplacian operator = Lorentz-invariant quantity!
The 4-vector product of any two (bona-fide) relativistic 4-vectors is a Lorentz invariant quantity (i.e. the same value in any/all IRF’s)!
We can equivalently write the relativistic 4-D continuity equation as:
J JJ0 n.b. A Lorentz-invariant quantity!!! x
Electric charge is (locally) conserved in any/all IRF’s (as it must be!!!)
6 © Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 2005-2015. All Rights Reserved. UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 19 Prof. Steven Errede
Maxwell’s Equations in Tensor Notation
Maxwell’s Equations: 1 1) Gauss’ law: E o 2) No Magnetic Monopoles: B 0 B 3) Faraday’s law: E t 1 E 4) Ampere’s law: BJ with Maxwell’s Displacement Current o ct2
Can be written as 4-derivatives of the relativistic EM field strength tensors F v and Gv :
F v Gv FJv and: G v 0 voxv v xv
n.b. summation over v = 0:3 is implied
1 F v F 0v E FJv 00v 1) Gauss’ law . If μ = 0 in: vov , i.e. voFJv , v = 0:3 o x x
Physically, μ = 0 is the temporal/scalar component of any space-time 4-vector.
FFFFF0v 00010203 Then: μ = 0 (first row of F v ) xv xxxx0123
Row #
0 Ecxyz Ec Ec Ecxzy0 B B F v Ecyz B0 B x Eczyx B B 0 Column #
0v E F 11Ex y Ez 0 0 0 E and: ooJc xcxyzc
1 2 2 1 1 E oc or: Eco but: oc E c o o
Gauss’ law arises from the 0 (scalar / temporal) component of the 4-vector relation:
F v FJv voxv
© Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 7 2005-2015. All Rights Reserved. UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 19 Prof. Steven Errede 1 E F v 4) Ampere’s law: BJ If 1 in FJv o ct2 voxv
FFFFF1v 10111213 Then: 1 (second row of F v ) xv xxxx0123 1v B FE11Ex Bz y 1 0 B and: 0 JJ ox xctv 22 yzct x FE1v 1 B oxJ xctv 2 x
Then for 2 and 3 (third and fourth rows of F v ), likewise we find that: FE2v 1 FE3v 1 B oyJ and: B ozJ xctv 2 xctv 2 y z F v 1 E J B J xv 1:3o 1:3 ct2 o
3-D spatial components of 4-vector J 1 E Or: BJ Ampere’s law with Maxwell’s displacement current term !!! o ct2
Ampere’s law arises from the 1:3 (3-D spatial / vector component) of 4-vector relation:
1 E (0 temporal / scalar component) F v o FJv v xv 0 1 E B J (1:3 3-D spatial / vector component) ct2 o
Thus, Gauss’ law and Ampere’s law form a 4-vector:
0 o J v 11 1EFv oovJE , B2 J v F cc ct x o
Gauss’ law temporal / scalar Ampere’s law 3-D spatial / vector 0 component 0 J of 4-vector o J component 0 J of 4-vector o J
22 And: JJ JJ0 c = Lorentz invariant quantity {from above, page 5}.
8 © Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 2005-2015. All Rights Reserved. UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 19 Prof. Steven Errede
G v 2) B 0 no magnetic monopoles / no magnetic charges. If 0 in G v 0 v xv Gv 0 is the temporal (scalar) component of space-time “null” 4-vector G v 00,00 v xv GGGGG0v 00010203 Then: 0 0 (First row of Gv ) xxxxxv 0123
Row #
0 BBBxyz Bxzy0 Ec Ec Gv ByzEc0 Ec x BEcEczy x 0 Column #
G0v B B B G0v 000x y z B B 00 xxyzv xv B Gv 3) Faraday’s law: E If 1 in G v 0 t v xv
GGGGG110111213v Then: 0 1 (Second row of Gv ) xxxxxv 0123 1v GB1111B E Ey x 00 z E xctcyczctv x G1v B 0 gives E 0 xv t x Likewise, for 2 and 3 (third and fourth rows of Gv ) G2v B G3v B 0 gives E 0 and: 0 gives E 0 xv t xv t y z Gv B B 0 gives E 0 or: E xv 1:3 t t B 0 ( 0 temporal / scalar component) Gv Thus: G v 0 v v x B E ( 1:3 3-D spatial / vector component) t Gv Arise from temporal ( 0 ) and spatial ( 1:3) component of the “null” 4-vector G v 0 v xv
© Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 9 2005-2015. All Rights Reserved. UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 19 Prof. Steven Errede
Thus, in relativistic 4-vector / tensor notation, Maxwell’s 4 equations {written in language of 3-D differential vector calculus}:
Maxwell’s Equations: 1 1) Gauss’ law: E o 2) No Magnetic Monopoles: B 0 B 3) Faraday’s law: E t 1 E 4) Ampere’s law: BJ with Maxwell’s displacement current o ct2 are elegantly represented by two simple 4-vector equations:
1 0 temporal/scalar component: E 1) Gauss’ law o F v FJv voxv 1 E 1:3 spatial/vector component: B J 4) Ampere’s law ct2 o With Maxwell’s displacement current
0 temporal/scalar component: B 0 2) No Magnetic Charges G v G v 0 v xv B 1:3 spatial/vector component: E 0 3) Faraday’s law t
Griffiths Problem 12.53: B We can show that Maxwell’s two equations B 0 and E 0 that are contained in t G v G v 0 can also be obtained from (the more cumbersome/inelegant relation): v xv
G v FFF n.b. note the cyclic G v 0 FFF v v 0 permutations of v xv vvv xxxv indices , and v.
Since there are 3 indices in the latter equation 0 :3, v 0 :3, 0 :3 , there are actually 64 (= 43) separate equations!!! However, many of these 64 equations are either trivial or redundant.
10 © Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 2005-2015. All Rights Reserved. UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 19 Prof. Steven Errede
Suppose two indices are the same (e.g. v )
FFF Then: 0 . But the EM field strength tensor F (like F v ) is anti-symmetric. xxx v
F 0 and: FF . Thus, 2 indices the same gives the trivial relation: 0 = 0.
Thus, in order to obtain a / any non-trivial result, μ, v, and λ must all be different from each other.
1) The indices μ, v, and λ could all be spatial indices, such as: μ = 1 (x), v = 2 (y), λ = 3 (z) (or permutations thereof).
Or:
2) One index could be temporal, and two indices could be spatial, such as: μ = 0, v = 1, λ = 2 (or permutations thereof), or: μ = 0, v = 1, λ = 3
1) For the case(s) of all spatial indices, e.g. μ = 1, v = 2, λ = 3:
B F12 FF23 31 Bz Bx y 0 = 0 = B 0 xxx312 zxy All other permutations involving the all-spatial indices {1, 2, 3} yield the same relation B 0 or minus it: i.e. B 0 .
2) For the case of one temporal and two spatial indices, e.g. μ = 0, v = 1, λ = 2:
E E FF01F12 20 111Ex Bz y 11Bz y Ex 2010 = 0 = 0 xxx cy ct cx ct c x y B = E 0 t z Other Permutations:
For v = 0, μ & λ = 1:3 and λ = 0, μ & v = 1:3 get redundant results (same as above).
If μ = 0, v = 1, λ = 3 get y – component of above relation! If μ = 0, v = 2, λ = 3 get x – component of above relation! FFF B v v 0 contains: B 0 and: E xxxv t
© Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 11 2005-2015. All Rights Reserved. UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 19 Prof. Steven Errede
Duality Transformation of the Relativistic EM Field Strength Tensors F v and Gv : The duality transformation for the specific case of space-time “rotating” EcB and o vv cB E ( duality = 90 ) takes FG , and can be mathematically represented in tensor notation as: vv1 GF 2
where: F is the {doubly} covariant form of the contravariant tensor F . and: v is the totally anti-symmetric rank-four tensor.
+1 for all even permutations of μ = 0, v = 1, λ = 2, σ = 3 v = 0 if any two indices are equal/identical/the same. 1 for all odd permutations of μ = 0, v = 1, λ = 2, σ = 3
Since v is a rank-four tensor (= 4-dimensional “matrix”) we can’t write it down on 2-D paper all at once! v has (μ, v, λ, σ = 0:3) → 44 elements = 256 elements!!!
We could write out 16 {44} matrices – e.g. one μ-v matrix for each unique combination of λ and σ:
000 vvv 0 v 012(4 4) (4 4) (4 4) 3 (4 4) 111 vvv 1 v v 012(4 4) (4 4) (4 4) 3 (4 4) 222 vvv 2 v (4 4) (4 4) (4 4) (4 4) 01 2 3 333 vv vv 3 012(4 4) (4 4) (4 4) 3 (4 4)
Define = 4×4 totally anti-symmetric rank-two tensor (aka the 4×4 Levi-Cività symbol):
0111 1 0 1 1 11 01 111 0
Then, we can define v in terms of the product of two 's: vv
12 © Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 2005-2015. All Rights Reserved. UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 19 Prof. Steven Errede
The Minkowski / Proper Force on a Point Electric Charge
The Minkowski force (a.k.a. proper force) K acting on a point electric charge q can be written in 4-vector / tensor notation in terms of the EM field strength tensor F v and the proper 4-velocity . Recall that:
dp dp dp KF where the ordinary force: F and: 11 2 ddt uu dt uu
v However, we can equivalently write the Minkowski/proper force as: KqF v v where v is the covariant form of the contravariant proper 4-velocity . v i.e. we contract the EM field strength tensor F with the covariant proper 4-velocity v .
2 Since: KF u and: vuv u , where: uu11 and: u uc
v v v KqF v uuvFquF or: FquF v ordinary force on point charge
1 1v 010111212 313 If μ = 1 (i.e. row #1): KqFqv F F F F
2 ucuv , ↔ vuuuvcu, u where: uu11 and: uc
Row # 0 0 Ecxyz Ec Ec K 0 Ecxyz Ec Ec0 Ec0 B B K1 Ec0 B B F v xzyqxzy1 2 Ecyz B0 B x K Ecyz B0 B x 2 Ec B B 0 3 Ec B B 0 zyx K zyx 3 Column # 1 Kquxxyzzyuxyzzyu cEcu0 uBuB qEuBuBqEuB x
1 KqEuB u Kq u EuB ← Minkowski 3-D Force law x
Similarly, for μ = 2, μ = 3: 2 KqEuB u But: KF u y 3 KqEuB u FqEuB ← Lorentz 3-D Force law z
© Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 13 2005-2015. All Rights Reserved. UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 19 Prof. Steven Errede
For μ = 0 (the temporal / scalar component) {see/work Griffiths Problem 12.54, page 541}: 0 Kquxxyxzxu c0 uEcuEcuEcquEc
000 101 202 303 0v FF F FF v
n.b. this relation explicitly shows that Ex , Ey , Ez are temporal-spatial (or spatial-temporal) v v components of F , whereas Bx , By , Bz are pure spatial-spatial components of F !!!
dp0 11 dE dt dE 1 dE 1 dE We also know that: K 0 KquEc0 dcdcddtcdt u uucdt
dE or: PquE = {ordinary} relativistic power delivered to point charged particle (> 0) q dt
dE dW Time rate of change of work done PquEqEuFu = q dt dt on changed particle by EM field
Note: The {ordinary} Lorentz force: FqEquB
Fu quE q u u B quE
But: uB u uuB0 magnetic forces do no work !!!
v v Thus we have the relations: KF u = KqF v and also: FquF v with vuv u .
The Relativistic 4-Vector Potential A We know that the electric and magnetic fields EB and can be expressed in terms of a scalar potential V and a vector potential A as:
Art, Ert,, Vrt and: Brt,, Art t Thus, it should not be surprising to realize that the scalar potential V and the vector potential A form the temporal and spatial components (respectively) of the relativistic 4-vector potential A :
SI Units: Newtons/Ampere = “p/q” The 4-Vector Potential: AVcAVcAAA,,,,x yz {= momentum per Coulomb!}
Newton- meters V N-- m m N sec Newtons n.b. SI units of V: Volts then: == Coulomb c Coul sec Coul Amp
14 © Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 2005-2015. All Rights Reserved. UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 19 Prof. Steven Errede
The EM field strength tensor F v can be written in terms of covariant space-time derivatives of the 4-vector potential field A as:
AAv FAAvvv n.b. covariant differentiation here!! x xv n.b. We must change the sign of the temporal/scalar component of the covariant derivatives
v and relative to that of the contravariant derivatives and v v . x xv x x Explicitly evaluate a few terms: AVcA ,
For μ = 0 and v = 1: 10 01 AAAE Vc 11 A A FVVx x x xctxctc tc 01 xx
Likewise, for (μ = 0, v = 2) and (μ = 0, v = 3) we obtain: 02 1 A Ey 03 1 A E FV and: FV z ctc ctc y z
For μ = 1 and v = 2:
21 12 AAAy Ax FAB z z xx12 xy
Likewise, for (μ = 1, v = 3) and (μ = 2, v = 3) we obtain: 13 23 FAB y and: FAB x y x
Note that the relativistic 4-potential formulation automatically takes care of the G v B homogeneous Maxwell equation G v 0 {it gives B 0 and: E } v xv t G v FFF because G v 0 is equivalent to v v 0 . v xv xxxv
{See/read pages 10-11 of these lecture notes – also see/work Griffiths Problem 12.53, page 541}.
AAv A A FAAvvv FAA v And since: vvv v x xv x x
© Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 15 2005-2015. All Rights Reserved. UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 19 Prof. Steven Errede
FFF Thus: v v 0 xxxv
AAAAvvAA = vv v 0 xx x xx x xx x
AAAAAA = vv0 xx xx vv xx xx xx vv xx
AAvv AA AA = vvvv 0 xx xx xx xx xx xx
= AAv vv vv A0 xx xx xx xx xx xx 0 0 0
But: i.e. can change the order of differentiation – has no effect! x xxx 22 = 00 x xxx AAv F v The relativistic 4-potential formulation does indeed automatically satisfy x xv G v FFF Gv G v 0 because v v 0 {shown to be equivalent to 0 } v xv xxxv xv AAv FAAvvv is satisfied / obeyed for . x xv AAv FAAvvv Does the relativistic 4-potential formulation x xv F v satisfy the inhomogeneous Maxwell relation FJv ??? voxv
FAAAAvv22 v FJv v vv v v o xxxxvv xxxx
Switching the order of derivatives: vv v FAA voFJvvv xxxxx v
This is an intractable equation, as it stands now…
16 © Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 2005-2015. All Rights Reserved. UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 19 Prof. Steven Errede
However, from our formulation of F v in terms of (differences) in space-time derivatives of AAv FAAvvv the 4-vector potential A : it is clear that we can add to the x xv 4-vector potential A the space-time gradient of any scalar function λ: AAA* !!! x The scalar and vector potentials V and A are not uniquely determined by the EM fields E and B .
Thus:
AA**vv A A AA v 2 2 F *v xxvvvvvv xxxxxxxx xxxv x 0 AAv F v !!! xx v
FF*vv by AAA* is gauge invariance associated with the EM field F v !!! x
We can exploit the gauge invariant properties of F v to simplify the seemingly intractable relation:
vv v FAA voFJvvv xxxxx v
1 V 1 V A Using the Lorenz gauge condition: A A 0 A 0 ct2 ct2 x
0 !! FAAvv 2 A 2 A v v AJ Thus: voFJvvv = v o J or: vov xxxxx v xxv xxv
22 2 vv 2 vvvv2 1 But: vv v , and: vv v v vv 22 x xv xvvxxx ct D’Alembertian operator (4-dimensional Laplacian operator)
2 A Single 4-vector equation! 2 v = the most elegant and simple formulation of Maxwell’s AAJ vo = v o J xxv equations – it contains all four of Maxwell’s equations!!
Taken together with the continuity equation (charge conservation): JJ 0 , these two relations compactly describe virtually all of {non-matter/free space} EM phenomena!!!
© Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 17 2005-2015. All Rights Reserved. UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 19 Prof. Steven Errede Note that the choice of the (instantaneous) Coulomb gauge A 0 is a bad one for use in relativistic electrodynamics, because A {alone} is not a Lorentz invariant quantity!
1 Vrt,, A rt However: Art,,0 A rt is a Lorentz invariant quantity ct2 x because it is the product of two relativistic 4-vectors: and A . x
n.b. A 0 is “destroyed” by any Lorentz transformation from one IRF(S) to another IRF(S') !!!
In order to restore A 0 , one must perform an appropriate gauge transformation for each new inertial system entered, in addition to carrying out the Lorentz transformation itself !!!
In the Coulomb gauge A is not a “true” relativistic 4-vector, because AA AA is actually not a Lorentz invariant quantity in the Coulomb gauge !!!
n.b. The Coulomb gauge A 0 is useful when vc , i.e. for non-relativistic problems.
18 © Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 2005-2015. All Rights Reserved.