PPT

advertisement
BH Astrophys Ch6.6
The Maxwell equations
– how charges produce fields
𝛻·π΅ =0
𝛻 · 𝐷 = 4πœ‹πœŒ
𝛻×𝐸 =−
1πœ•π΅
𝑐 πœ•π‘‘
4πœ‹
1 πœ•π·
𝛻×𝐻 =
𝐽 +
𝑐
𝑐 πœ•π‘‘
Total of 8 equations, but only 6 independent variables (3 components each for E,B)
Where are the two extra equations hidden?
Conservation of (electric) charge
See Jackson 3ed Ch 6.11 for more about this
“Constrained transport” of magnetic field
Conservation of magnetic monopoles ( equal zero in our universe)
Therefore the Maxwell equations actually have in-built conservation of electric and
magnetic charges! (therefore total independent equations are 6 only)
Energy momentum equations
– How fields affect particles
Work done by EM field on charges
Lorentz force equation
For fluids, they translate to
Recall energy equation of fluid
The scalar and vector potentials
Given the Maxwell equations,
and
One can define a scalar potential πœ™ and a vector potential 𝐴 which have the relations
with 𝐸 and 𝐡 as
𝐡 =π›»× π΄
𝐸 = −π›»πœ™ −
1πœ•π΄
𝑐 πœ•π‘‘
However, we can see that any change of πœ™ and 𝐴 that follow
πœ™′ = πœ™ +
1 πœ•π‘
𝑐 πœ•π‘‘
′
𝐴 = 𝐴 + 𝛻𝑍
Will still give the same 𝐸 and 𝐡 fields
This is called “gauge invariance”, we have the freedom to choose some Z that would
make problems easier.
The Lorenz Gauge
Applying the definitions
𝐡 =𝛻×𝐴
𝐸 = −π›»πœ™ −
One can transform the two source equations
1πœ•π΄
𝑐 πœ•π‘‘
𝛻 · 𝐷 = 4πœ‹πœŒ
𝛻×𝐻 =
4πœ‹
1 πœ•π·
𝐽 +
𝑐
𝑐 πœ•π‘‘
into
Often when the problem involves waves, a convenient gauge is the “Lotenz Gauge”
1 πœ•πœ™
𝛻· 𝐴 +
= 0 Which, through
𝑐 πœ•π‘‘
πœ™′ = πœ™ +
′
1 πœ•π‘
𝑐 πœ•π‘‘
also tells us that
𝐴 = 𝐴 + 𝛻𝑍
1 πœ•2𝑍
− 2 2 + 𝛻 2𝑍 = 0
𝑐 πœ•π‘‘
Applying the Lorenz Gauge condition, the two potential equations become
1 πœ•2πœ™
− 𝛻 2 πœ™ = 4πœ‹πœŒ
2
2
𝑐 πœ•π‘‘
1 πœ•2 𝐴
4πœ‹
2 𝐴 =
−
𝛻
𝐽
𝑐 2 πœ•π‘‘ 2
𝑐
This is not the Lorentz as in Lorentz transformations!
Ludvig Lorenz (1829~1891) Hendrik Lorentz (1853~1928)
Linearly polarized plane EM waves
in vacuum
In vacuum, the wave equations become
1 πœ•2πœ™
− 𝛻 2πœ™ = 0
2
2
𝑐 πœ•π‘‘
1 πœ•2 𝐴
4πœ‹
2
−
𝛻
𝐴
=
𝐽
𝑐 2 πœ•π‘‘ 2
𝑐
Which gives
The Lotenz gauge condition
1 πœ•πœ™
𝛻· 𝐴 +
=0
𝑐 πœ•π‘‘
Since Z used in the gauge transforms also satisfies
It also has the wave solution
πœ™′ = πœ™ +
′
1 πœ•π‘
𝑐 πœ•π‘‘
𝐴 = 𝐴 + 𝛻𝑍
Then tells us that
gives
1 πœ•2𝑍
− 2 2 + 𝛻 2𝑍 = 0
𝑐 πœ•π‘‘
Building the spacetime compatible
form of electromagnetism
To work in spacetime, we need equations to be in tensor form such that they will be
valid in any frame, in harmony with the principle of relativity.
It is therefore crucial to rewrite the whole set of equations in 4-form, either with 4vectors or tensors.
Previously, we started off with the electric and magnetic fields from the Maxwell’s
equations and finally came to another fully equivalent form in describing things – the
potentials.
Here, we will do the reverse, we work out the potential in 4-form then find some
suitable tensor to fit in the electric and magnetic fields.
The current 4-vector
Recall the charge continuity equation
πœ•πœŒ
+𝛻· 𝐽 =0
πœ•π‘‘
Observing closely, we see that if we define π½πœ‡ = 𝜌, 𝐽 , then using
π‘₯ πœ‡ = 𝑑, π‘₯, 𝑦, 𝑧 , the continuity equation can very simply be written as
πœ•π½πœ‡
≡ π½πœ‡ ,πœ‡ = πœ•πœ‡ π½πœ‡ = 0
πœ‡
πœ•π‘₯
(contracting a vector and a one-form gives a scalar)
What this says is that π½πœ‡ is the correct 4-vector that compatible with space-time
transforms!
Note:
πœ•
πœ• πœ• πœ• πœ•
πœ•πœ‡ ≡ πœ‡ =
, , ,
πœ•π‘₯
πœ•π‘‘ πœ•π‘₯ πœ•π‘¦ πœ•π‘§
πœ•πœ‡ ≡
πœ•
πœ• πœ• πœ• πœ•
= − , , ,
πœ•π‘₯πœ‡
πœ•π‘‘ πœ•π‘₯ πœ•π‘¦ πœ•π‘§
The potential 4-vector
Recall the wave equations
1 πœ•2πœ™
2 πœ™ = 4πœ‹πœŒ
−
𝛻
𝑐 2 πœ•π‘‘ 2
We can see that in both cases we have the operator
πœ•2
πœ•π‘‘ 2
1 πœ•2 𝐴
4πœ‹
2
−𝛻 𝐴 =
𝐽
𝑐 2 πœ•π‘‘ 2
𝑐
− 𝛻 2 acting on either πœ™ or 𝐴 .
Is that operator also related to some 4-vector?
Apparently, Yes! It is the square of the gradient operator!
2
πœ•
πœ•πœ‡ πœ•πœ‡ = −( 2 − 𝛻 2 )
πœ•π‘‘
Then, since π½πœ‡ = 𝜌, 𝐽 is a 4-vector as we have demonstrated, we can also define the
4-potential
π΄πœ‡ ≡ πœ™, 𝐴
Then, the wave equations are simply, in 4-form,
πœ• 𝛼 πœ•π›Ό 𝐴𝛽 = −4πœ‹π½π›½
And the Lorenz gauge condition
𝐴𝛼 ,𝛼 = 0
Reminder: 𝛻 · 𝐴 +
1 πœ•πœ™
𝑐 πœ•π‘‘
=0
Gauge-free form
Recall that before we took the Lorenz gauge, our equations for the potentials looked like:
By applying the definitions
π½πœ‡
= 𝜌, 𝐽 ;
π΄πœ‡
We find the 4-form for the above 2 equations as
≡ πœ™, 𝐴 ;
πœ•πœ‡ πœ•πœ‡
=
(𝛻 2
−
πœ•2
)
πœ•π‘‘ 2
πœ• 𝛼 πœ•π›Ό 𝐴𝛽 − πœ•π›½ πœ•π›Ό 𝐴𝛼 = −4πœ‹π½π›½
Since the equations for the potentials originally came from
𝛻 · 𝐷 = 4πœ‹πœŒ
𝛻×𝐻−
1 πœ• 𝐷 4πœ‹
=
𝐽
𝑐 πœ•π‘‘
𝑐
We should be able to manipulate πœ• 𝛼 πœ•π›Ό 𝐴𝛽 − πœ•π›½ πœ•π›Ό 𝐴𝛼 = −4πœ‹π½π›½ into something that
gives us the Maxwell equations in terms of E and B fields
Towards the Maxwell equations
Lets rewrite πœ• 𝛼 πœ•π›Ό 𝐴𝛽 − πœ•π›½ πœ•π›Ό 𝐴𝛼 = −4πœ‹π½π›½ as πœ•π›Ό (πœ• 𝛼 𝐴𝛽 − πœ•π›½ 𝐴𝛼 ) = −4πœ‹π½π›½
Comparing with
𝛻 · 𝐷 = 4πœ‹πœŒ
𝛻×𝐻−
1 πœ• 𝐷 4πœ‹
=
𝐽
𝑐 πœ•π‘‘
𝑐
It should be clear that as the Maxwell equations contain only 1st derivatives,
πœ• 𝛼 𝐴𝛽 − πœ•π›½ 𝐴𝛼 should be some 2nd rank tensor that contains that E and B fields.
𝐹 αβ ≡ πœ• 𝛼 𝐴𝛽 − πœ•π›½ 𝐴𝛼
𝐹 αβ =
0
=
0
πœ• 0 𝐴1 − πœ• 1 𝐴0
πœ• 0 𝐴2 − πœ• 2 𝐴0
πœ• 0 𝐴3 − πœ• 3 𝐴0
1 0
1 2
2 1
− πœ• 𝐴 −πœ• 𝐴
0
πœ• 𝐴 −πœ• 𝐴
πœ•1 𝐴3 − πœ• 3 𝐴1
− πœ• 0 𝐴2 − πœ• 2 𝐴0 − πœ•1 𝐴2 − πœ• 2 𝐴1
0
πœ• 2 𝐴3 − πœ• 3 𝐴2
− πœ• 0 𝐴3 − πœ• 3 𝐴0 − πœ•1 𝐴3 − πœ• 3 𝐴1 − πœ• 2 𝐴3 − πœ• 3 𝐴2
0
πœ•π΄π‘¦ πœ•πœ™
πœ•π΄π‘₯ πœ•πœ™
πœ•π΄π‘§ πœ•πœ™
−
−
−
−
−
−
πœ•π‘‘
πœ•π‘₯
πœ•π‘‘
πœ•π‘¦
πœ•π‘‘
πœ•π‘§
0
𝐸π‘₯
𝐸𝑦
𝐸𝑧
πœ•π΄π‘¦ πœ•π΄π‘₯
πœ•π΄π‘§ πœ•π΄π‘₯
−𝐸π‘₯
0
𝐡𝑧 −𝐡𝑦
0
−
−
=
πœ•π‘₯
πœ•π‘¦
πœ•π‘₯
πœ•π‘§
−𝐸𝑦 −𝐡𝑧
0
𝐡π‘₯
πœ•π΄π‘§ πœ•π΄π‘¦
−𝐸𝑧 𝐡𝑦 −𝐡π‘₯
0
0
−
πœ•π‘¦
πœ•π‘§
1πœ•π΄
𝐡 =𝛻×𝐴 0
𝐸 = −π›»πœ™ −
𝑐 πœ•π‘‘
0 1
The Maxwell’s equations in 4-form
As we defined the Minkowski metric as πœ‚αβ =
−1
0
0
0
0
1
0
0
0
0
1
0
0
0
0
1
Same in Griffiths,
Introduction to
Electrodynamics 3ed
the Faraday tensor has the form we derived
𝐹 αβ ≡ πœ• 𝛼 𝐴𝛽 − πœ•π›½ 𝐴𝛼 =
0
−𝐸π‘₯
−𝐸𝑦
−𝐸𝑧
𝐸π‘₯
0
−𝐡𝑧
𝐡𝑦
𝐸𝑦
𝐡𝑧
0
−𝐡π‘₯
𝐸𝑧
−𝐡𝑦
𝐡π‘₯
0
and 𝐹αβ ≡
0
𝐸π‘₯
𝐸𝑦
𝐸𝑧
−𝐸π‘₯
0
−𝐡𝑧
𝐡𝑦
−𝐸𝑦
𝐡𝑧
0
−𝐡π‘₯
−𝐸𝑧
−𝐡𝑦
𝐡π‘₯
0
The source equation πœ•π›Ό (πœ• 𝛼 𝐴𝛽 − πœ•π›½ 𝐴𝛼 ) = −4πœ‹π½π›½ then becomes πœ•π›Ό 𝐹 αβ = −4πœ‹π½π›½
Or, since 𝐹 αβ = −𝐹 βα (2nd rank anti-symmetric tensor), πœ•π›Ό 𝐹 βα = 4πœ‹π½π›½ → πœ•β 𝐹 αβ = 4πœ‹π½α
The Maxwell’s equations in 4-form
(cont’)
Also, we can define another 2nd rank tensor
(sometimes called the Maxwell tensor)
𝐡 =𝛻×𝐴
𝐸 = −π›»πœ™ −
1πœ•π΄
𝑐 πœ•π‘‘
𝐺 αβ ≡
𝐡π‘₯
0
𝐸𝑧
−𝐸𝑦
𝐡𝑦
−𝐸𝑧
0
𝐸π‘₯
𝐡𝑧
𝐸𝑦
−𝐸π‘₯
0
Can also be expressed as πœ•π›½ 𝐺 αβ = 0
𝐹 αβ ≡
Faraday tensor
Note that in Jackson, one finds
0
−𝐡π‘₯
−𝐡𝑦
−𝐡𝑧
𝐹 αβ ≡
0
𝐸π‘₯
𝐸𝑦
𝐸𝑧
−𝐸π‘₯
0
𝐡𝑧
−𝐡𝑦
This is because Jackson uses the metric
−𝐸𝑦
−𝐡𝑧
0
𝐡π‘₯
πœ‚αβ =
−𝐸𝑧
𝐡𝑦
−𝐡π‘₯
0
0
−𝐸π‘₯
−𝐸𝑦
−𝐸𝑧
𝐸π‘₯
0
−𝐡𝑧
𝐡𝑦
𝐸𝑦
𝐡𝑧
0
−𝐡π‘₯
𝐸𝑧
−𝐡𝑦
𝐡π‘₯
0
Different by a minus sign
1 0
0
0 −1 0
0 0 −1
0 0
0
0
0
0
−1
Looking at page 11 where we derived the Faraday tensor will show why.
The Maxwell’s equations in 4-form
(cont’)
We can also show that πœ•π›½ 𝐺 αβ = 0 can be written as πœ•π›Ό 𝐹βγ + πœ•π›½ 𝐹γα + πœ•π›Ύ 𝐹αβ = 0
which basically is just an identity.
This means that by defining 𝐹 αβ ≡ πœ• 𝛼 𝐴𝛽 − πœ•π›½ 𝐴𝛼 we have already included the
homogeneous Maxwell equations.
4-force
Previously, we have defined the equations π‘ˆ 𝛼 ≡
Now, compare
𝐹 αβ =
0
−𝐸π‘₯
−𝐸𝑦
−𝐸𝑧
𝐸π‘₯
0
−𝐡𝑧
𝐡𝑦
𝐸𝑦
𝐡𝑧
0
−𝐡π‘₯
dx𝛼
;
dτ
𝑃 𝛼 ≡ π‘š0 π‘ˆ 𝛼 ; 𝐹 α =
dPα
dτ
𝐸𝑧
−𝐡𝑦
𝐡π‘₯
0
Then one can see that what is left to write out the electromagnetic 4-force is π‘ˆ 𝛼
Thus,
dP α π‘ž αβ
= 𝐹 π‘ˆπ›½
dτ
𝑐
Simultaneously contains the work done by the E field and the Lorenz force.
Equation 6.116 is wrong!! It should be the one-form 4-velocity
Plane EM waves in 4-form
Previously, we found that in the Lorenz gauge πœ• 𝛼 πœ•π›Ό 𝐴𝛽 = −4πœ‹π½π›½ ,
Therefore the source free equation is πœ• 𝛼 πœ•π›Ό 𝐴𝛽 = 0, which has the solution
We see that the phase term π‘˜ · π‘₯ = −ωt + π‘˜π‘₯ π‘₯ + π‘˜π‘¦ 𝑦 + π‘˜π‘§ 𝑧
Then, since phase is also a scalar, (sorry I don’t have a very good explanation for this),
we can rewrite π‘˜ · π‘₯ = −ωt + π‘˜π‘₯ π‘₯ + π‘˜π‘¦ 𝑦 + π‘˜π‘§ 𝑧 = πœ‚αβ π‘˜ 𝛼 π‘₯ 𝛽
π‘˜ 𝛼 = πœ”, π‘˜
πœ• 𝛼 πœ•π›Ό 𝐴𝛽 = 0 then gives πœ‚αβ π‘˜ 𝛼 π‘˜π›½ = −πœ”2 + π‘˜ 2 = 0 → photons travel along null vectors
Fluid aspects next week
Download