4-Vector Electrodynamics
Adrian Down January 30, 2006
1
1.1
Review
4-scalars
dt Proper time d Rest mass
1.2
4-vectors
x = (ct, x)
Position
4-momentum p = m where
dx d
dx = mc (, ) d
E ,p c
is the 4-velocity.
Wave vector k = ,k c
2
2.1
Relativistic equations
Motivation
We will see that almost all equations we have written down previously can be written in a 4-vector form. Also, these equations must be manifestly invariant.
2.2
2.2.1
Momentum
Scalar product
The scalar product of two vectors is independent of the frame of reference. In particular, consider the scalar product of the momentum with itself. In the rest frame of the body, all components of the velocity are 0. 1 mc 1 0 = m 2 c2 p p = mc 0 0 0 0 1 1 0 Evaluating the scalar product in an arbitrary frame, E 1 c 1 p1 E 2 E = |p|2 p p = c p1 p2 p3 p2 1 c2 1 p3 Equating these expressions for the scalar product yields the most useful equation in relativistic kinematics, E2 |p|2 = m2 c2 c2 2.2.2 Wave vector
Two familiar equations are combined in the relativistic form of the de Broglie equation, p = k E = =h p
This denition also shows that the phase of a wave is invariant. The phase is given by phase = k x This scalar product is invariant between frames. Physically, this is because the number of wavefronts observed by a moving emitter and a stationary observer must be the same. This formulation the wave 4-vector also naturally results in the relativistic doppler shift.
2.3
Derivatives
We would like to form more 4-vectors by dierentiation. The logically consistent denition of the derivative is = 1 , c t
Note. The negative sign results because of the form of the transformation matrix, x = x x = x Since the transformation matrix has upper and lower subscripts, it involves a metric tensor, which switches the sign of the space-like components. Using this denition, we can nd the 4D analog to the 3D equation, x= x +y + z x y z (x + y + z) = 1 + 1 + 1 = 3 x y z
In four dimensions, ct x =4 y z
x =
1 c t
Note. The above equation used the non-standard form derivative, so the sign of the spatial components is switched. 3
3
3.1
Potential and sources
4-potential
From the Lorentz gauge condition, 1 V =0 c2 t We can write this as a 4-vector equation if we dene a new 4-vector A, A+ A = 0 A = V ,A c
Since we can write the Lorentz gauge condition as a scalar product, it must hold in any frame if it holds in one particular frame.
3.2
DAlembertian
1 2 c2 t2 V 0 = A 0 J
The sourceful Maxwells equations satisfy the wave equations,
2
We can write this as a 4-vector equation if we dene the quantity in parenthesis as an operator. We can also relate it to the 4-derivative, =
2
1 c2
This operator is called the DAlembertian. Note. The negative sign in the denition cancels the single negative sign in the scalar product.
3.3
3.3.1
4-current density
Denition
The sourceful Maxwells equations can be written using the DAlembertian, ( ) A = 0 J if we dene the 4-current density J, J = (cp, J) 4
3.3.2
Charge conservation
Charge conservation can be expressed as t Alternatively, dv =
J da
= t
+ J=0 t This can be written as a 4D equation using the 4-current density, J = 0 This equation provides yet another justication for the denition of the 4current density, and shows that charge conservation is a Lorentz invariant.
3.4
Field-strength tensor
F = A A
For completion, we include the eld strength 4-tensor,
We can write this as a matrix. On the diagonal, = , so the matrix elements are 0. Switching the indices switches the sign of the matrix element, so the matrix must be anti-symmetric. Therefore, there are 6 independent components. Since the B and E elds are obtained by dierentiation of the potentials, it is reasonable to expect that the elements of the matrix should be components of the E and B elds. The Lorentz transformations for the E and B elds come about automatically from our understanding of 4-tensors, F
= F
Working out the transformation would yield, E|| = E|| E = (E + cB) B|| = B|| cB = (cB E)
Note. There is another possible 4-tensor which we have yet to dene called G. The best form of the Maxwells equations are then F = 0 J for the sourceful equations, and G = 0 for the sourceless equations.