
Uncorrected Proof
BookID 160928 ChapID 14 Proof# 1 - 29/07/09
426 14 Electrodynamics of Metals
E(z,t)=E
0
e
iωt−z/δ
. (14.22)
61
What we find is that electromagnetic waves do not propagate in the metal 62
(for frequencies lower than ω
p
), and that the electric field in the solid drops 63
off exponentially with distance from the surface. The distance δ =
c
ω
p
is called 64
the normal skin depth. 65
If ωτ 1, (this is usually true at rf frequencies, even at low temperatures 66
with pure materials) we have 67
ε(q,ω) 1 − i
ω
2
p
τ
ω
∼−i
ω
2
p
ωτ
ω
2
when ω
p
ω. (14.23)
The solution of the wave equation is given by 68
q = ±
ω
p
c
ωτ
2
1/2
(1 − i), (14.24)
so that the field E is of the form 69
E(z,t)=E
0
e
iωt
e
−(i+1)
ω
p
c
(
ωτ
2
)
1/2
z
. (14.25)
Thus, the skin depth is given by
70
δ =
c
ω
p
2
ωτ
1/2
. (14.26)
Ifthemeanfreepathl is much greater than the skin depth, l δ, then the
71
local theory is not valid. In good metals at low temperatures, it turns out that 72
l v
F
τ 10
7
nm and δ 10nm, so that l δ, and we must use the nonlocal 73
theory. 74
Anomalous Skin Effect 75
The normal skin effect was derived under the assumption that the q depen- 76
dence of σ was unimportant. Remember that this assumption is valid if 77
ql = qv
F
τ 1. We have found that the electric field varies like e
−z/δ
.If 78
δ turns out to be smaller than l = v
F
τ, our initial assumption was certainly 79
incorrect. The skin depth δ is of the order of 80
δ
c
ω
p
. (14.27)
Therefore, if
81
ωc
ω
p
v
F
<ωτ, (14.28)
the theory is inconsistent because the field E(z) changes appreciably over a
82
mean free path l contradicting the assumption that the q dependence of σ can 83
be neglected. The theory for this case in which the q dependence of σ must 84