Aus der Elektrodynamik folgt die Elektrostatik über die Forderung
∂
t
E
→
=
∂
t
B
→
=
∂
t
D
→
=
0
{\displaystyle \partial _{t}{\vec {E}}=\partial _{t}{\vec {B}}=\partial _{t}{\vec {D}}=0}
,
wodurch sich die (makroskopischen) Maxwellgleichungen wie folgt vereinfachen:
∇
→
⋅
B
→
=
0
{\displaystyle {\vec {\nabla }}\cdot {\vec {B}}=0}
(keine magnet. Monopole),
∇
→
⋅
D
→
=
4
π
ρ
{\displaystyle {\vec {\nabla }}\cdot {\vec {D}}=4\pi \rho }
(woraus u.a.
auch das Coulomb-Gesetz folgt),
∇
→
×
H
→
=
4
π
c
J
→
{\displaystyle {\vec {\nabla }}\times {\vec {H}}={\frac {4\pi }{c}}{\vec {J}}}
(Ampère'sches Durchflutungsgesetz) und
∇
→
×
E
→
=
0
{\displaystyle {\vec {\nabla }}\times {\vec {E}}=0}
(rotationsfreies elektrisches
Feld).
Aus diesen Grundgleichungen werden wir in den nächsten Kapiteln die
Gesetze der Elektrostatik herleiten.
Die Kontinuitätsgleichung nimmt wegen des Ampère'schen Durchflutungsgesetzes
die folgende Form an
∇
→
⋅
J
→
=
c
4
π
∇
→
⋅
(
∇
→
×
H
→
)
=
0
{\displaystyle {\vec {\nabla }}\cdot {\vec {J}}={\frac {c}{4\pi }}{\vec {\nabla }}\cdot \left({\vec {\nabla }}\times {\vec {H}}\right)=0}
wegen
d
i
v
r
o
t
≡
0
{\displaystyle div\,rot\equiv 0}
.
(1) Ein skalares Potential können wir folgendermaßen einführen:
∇
→
×
E
→
=
0
,
r
o
t
g
r
a
d
≡
0
⇒
E
→
=
−
∇
→
A
0
{\displaystyle {\vec {\nabla }}\times {\vec {E}}=0,rot\,grad\equiv 0\;\Rightarrow {\vec {E}}=-{\vec {\nabla }}A_{0}}
.
Die Konsequenzen hieraus sind die Wegunabhängigkeit für Kurvenintegrale
über das elektrische Feld, der Begriff der Spannung bzw. der Potentialdifferenz
sowie die Möglichkeit, das Skalarpotential angeben zu können. Dies
alles zeigen wir unter den folgenden beiden Punkten (a) und (b)
(a) Die Wegunabhängigkeit:
Fig:
C
1
{\displaystyle C_{1}}
: Weg von
x
→
A
{\displaystyle {\vec {x}}_{A}}
nach
x
→
B
{\displaystyle {\vec {x}}_{B}}
;
C
2
{\displaystyle C_{2}}
: Weg von
x
→
B
{\displaystyle {\vec {x}}_{B}}
nach
x
→
A
{\displaystyle {\vec {x}}_{A}}
Aus
∇
→
×
E
→
=
0
{\displaystyle {\vec {\nabla }}\times {\vec {E}}=0}
folgt mit dem Satz
von Stokes:
0
=
∫
A
d
a
→
⋅
(
∇
→
×
E
→
)
=
∫
C
=
∂
A
=
C
1
+
C
2
d
x
→
⋅
E
→
=
∫
C
1
d
x
→
⋅
E
→
+
∫
C
2
d
x
→
⋅
E
→
⇒
{\displaystyle 0={\underset {A}{\int }}d{\vec {a}}\cdot \left({\vec {\nabla }}\times {\vec {E}}\right)={\underset {C=\partial A=C_{1}+C_{2}}{\int }}d{\vec {x}}\cdot {\vec {E}}={\underset {C_{1}}{\int }}d{\vec {x}}\cdot {\vec {E}}+{\underset {C_{2}}{\int }}d{\vec {x}}\cdot {\vec {E}}\;\Rightarrow }
∫
C
1
d
x
→
⋅
E
→
=
−
∫
C
2
d
x
→
⋅
E
→
=
∫
−
C
2
d
x
→
⋅
E
→
{\displaystyle {\underset {C_{1}}{\int }}d{\vec {x}}\cdot {\vec {E}}=-{\underset {C_{2}}{\int }}d{\vec {x}}\cdot {\vec {E}}={\underset {-C_{2}}{\int }}d{\vec {x}}\cdot {\vec {E}}}
,
obwohl
C
1
≠
−
C
2
{\displaystyle C_{1}\neq -C_{2}}
.
Es gibt daher eine Spannung zwischen
x
→
A
{\displaystyle {\vec {x}}_{A}}
und
x
→
B
{\displaystyle {\vec {x}}_{B}}
:
U
=
−
∫
C
1
d
x
→
⋅
E
→
=
∫
C
1
d
x
→
⋅
∇
→
A
0
=
A
0
(
x
→
B
)
−
A
0
(
x
→
A
)
{\displaystyle U=-{\underset {C_{1}}{\int }}d{\vec {x}}\cdot {\vec {E}}={\underset {C_{1}}{\int }}d{\vec {x}}\cdot {\vec {\nabla }}A_{0}=A_{0}\left({\vec {x}}_{B}\right)-A_{0}\left({\vec {x}}_{A}\right)}
,
eine sog. Potenzialdifferenz, die mit der potentielle Energie für
eine Punktladung q wie folgt zusammenhängt:
E
p
o
t
(
x
→
B
)
−
E
p
o
t
(
x
→
A
)
=
−
∫
C
1
d
x
→
⋅
F
→
=
{\displaystyle E_{pot}\left({\vec {x}}_{B}\right)-E_{pot}\left({\vec {x}}_{A}\right)=-{\underset {C_{1}}{\int }}d{\vec {x}}\cdot {\vec {F}}=}
−
∫
C
1
d
x
→
⋅
q
E
→
=
q
U
=
q
(
A
0
(
x
→
B
)
−
A
0
(
x
→
A
)
)
⇒
{\displaystyle -{\underset {C_{1}}{\int }}d{\vec {x}}\cdot q{\vec {E}}=qU=q\left(A_{0}\left({\vec {x}}_{B}\right)-A_{0}\left({\vec {x}}_{A}\right)\right)\;\Rightarrow }
E
p
o
t
=
q
A
0
{\displaystyle E_{pot}=qA_{0}}
.
(b) Bestimmen von
A
0
{\displaystyle A_{0}}
:
∇
→
⋅
D
→
=
4
π
ρ
{\displaystyle {\vec {\nabla }}\cdot {\vec {D}}=4\pi \rho }
, isotrope Dielektrika:
D
→
=
ε
E
→
,
∇
→
2
1
r
=
−
4
π
δ
3
(
r
→
)
{\displaystyle {\vec {D}}=\varepsilon {\vec {E}},{\vec {\nabla }}^{2}{\frac {1}{r}}=-4\pi \,\delta ^{3}\left({\vec {r}}\right)}
.
−
∇
→
2
A
0
(
x
→
)
=
∇
→
⋅
E
→
(
x
→
)
=
4
π
ε
ρ
(
x
→
)
=
4
π
ε
∫
d
3
x
′
ρ
(
x
→
′
)
δ
3
(
x
→
−
x
→
′
)
{\displaystyle -{\vec {\nabla }}^{2}A_{0}\left({\vec {x}}\right)={\vec {\nabla }}\cdot {\vec {E}}\left({\vec {x}}\right)={\frac {4\pi }{\varepsilon }}\rho \left({\vec {x}}\right)={\frac {4\pi }{\varepsilon }}\int d^{3}x^{\prime }\,\rho \left({\vec {x}}^{\prime }\right)\delta ^{3}\left({\vec {x}}-{\vec {x}}^{\prime }\right)}
=
1
ε
∫
d
3
x
′
ρ
(
x
→
′
)
∇
→
2
1
|
x
→
−
x
→
′
|
=
∇
→
2
1
ε
∫
d
3
x
′
ρ
(
x
→
′
)
|
x
→
−
x
→
′
|
{\displaystyle ={\frac {1}{\varepsilon }}\int d^{3}x^{\prime }\,\rho \left({\vec {x}}^{\prime }\right){\vec {\nabla }}^{2}{\frac {1}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}={\vec {\nabla }}^{2}{\frac {1}{\varepsilon }}\int d^{3}x^{\prime }\,{\frac {\rho \left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}}
,
A
0
(
x
→
)
=
1
ε
∫
d
3
x
′
ρ
(
x
→
′
)
|
x
→
−
x
→
′
|
{\displaystyle A_{0}\left({\vec {x}}\right)={\frac {1}{\varepsilon }}\int d^{3}x^{\prime }\,{\frac {\rho \left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}}
.
Für das Beispiel einer Punktladung, d.h.
ρ
(
x
→
′
)
=
q
δ
3
(
x
→
−
x
→
′
)
{\displaystyle \rho \left({\vec {x}}^{\prime }\right)=q\delta ^{3}\left({\vec {x}}-{\vec {x}}^{\prime }\right)}
,
erhalten wir
A
0
(
x
→
)
=
q
ε
r
,
E
→
=
−
∇
→
A
0
(
x
→
)
=
q
ε
x
→
r
3
{\displaystyle A_{0}\left({\vec {x}}\right)={\frac {q}{\varepsilon r}},\;{\vec {E}}=-{\vec {\nabla }}A_{0}\left({\vec {x}}\right)={\frac {q}{\varepsilon }}{\frac {\vec {x}}{r^{3}}}}
,
also das berühmte Coulomb-Feld.
(2) Ein Vektorpotential resultiert aus der Gleichung
∇
→
⋅
B
→
=
0
,
d
i
v
r
o
t
≡
0
⇒
B
→
=
∇
→
×
A
→
{\displaystyle {\vec {\nabla }}\cdot {\vec {B}}=0,div\,rot\equiv 0\;\Rightarrow {\vec {B}}={\vec {\nabla }}\times {\vec {A}}}
.
Mit Hilfe des Ampère'schen Durchflutungsgesetztes
∇
→
×
H
→
=
4
π
c
J
→
{\displaystyle {\vec {\nabla }}\times {\vec {H}}={\frac {4\pi }{c}}{\vec {J}}}
,
und unter der Annahme isotroper Magnetika, d.h.
B
→
=
μ
H
→
{\displaystyle {\vec {B}}=\mu {\vec {H}}}
,
erhalten wir über die Greensfunktion
1
r
{\displaystyle {\frac {1}{r}}}
der
Elektrostatik, für die
∇
→
2
1
r
=
−
4
π
δ
3
(
r
→
)
{\displaystyle {\vec {\nabla }}^{2}{\frac {1}{r}}=-4\pi \,\delta ^{3}\left({\vec {r}}\right)}
gilt, folgende Bestimmungsgleichung für das Vektorpotential:
∇
→
(
∇
→
⋅
A
→
)
−
∇
→
2
A
→
=
∇
→
×
(
∇
→
×
A
→
)
=
∇
→
×
B
→
=
4
π
μ
c
J
→
{\displaystyle {\vec {\nabla }}\left({\vec {\nabla }}\cdot {\vec {A}}\right)-{\vec {\nabla }}^{2}{\vec {A}}={\vec {\nabla }}\times \left({\vec {\nabla }}\times {\vec {A}}\right)={\vec {\nabla }}\times {\vec {B}}={\frac {4\pi \mu }{c}}{\vec {J}}}
.
In Coulomb-Eichung, d.h.
∇
→
⋅
A
→
=
0
{\displaystyle {\vec {\nabla }}\cdot {\vec {A}}=0}
,
wird hieraus
∇
→
2
A
→
(
x
→
)
=
−
4
π
μ
c
J
→
(
x
→
)
=
−
4
π
μ
c
∫
d
3
x
′
J
→
(
x
→
′
)
δ
3
(
x
→
−
x
→
′
)
{\displaystyle {\vec {\nabla }}^{2}{\vec {A}}\left({\vec {x}}\right)=-{\frac {4\pi \mu }{c}}{\vec {J}}\left({\vec {x}}\right)=-{\frac {4\pi \mu }{c}}\int d^{3}x^{\prime }\,{\vec {J}}\left({\vec {x}}^{\prime }\right)\delta ^{3}\left({\vec {x}}-{\vec {x}}^{\prime }\right)}
=
μ
c
∫
d
3
x
′
J
→
(
x
→
′
)
∇
→
2
1
|
x
→
−
x
→
′
|
=
∇
→
2
μ
c
∫
d
3
x
′
J
→
(
x
→
′
)
|
x
→
−
x
→
′
|
{\displaystyle ={\frac {\mu }{c}}\int d^{3}x^{\prime }\,{\vec {J}}\left({\vec {x}}^{\prime }\right){\vec {\nabla }}^{2}{\frac {1}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}={\vec {\nabla }}^{2}{\frac {\mu }{c}}\int d^{3}x^{\prime }\,{\frac {{\vec {J}}\left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}}
,
woraus für das Vektorpotential
A
→
(
x
→
)
=
μ
c
∫
d
3
x
′
J
→
(
x
→
′
)
|
x
→
−
x
→
′
|
{\displaystyle {\vec {A}}\left({\vec {x}}\right)={\frac {\mu }{c}}\int d^{3}x^{\prime }\,{\frac {{\vec {J}}\left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}}
folgt.
Die Lösung der Poisson-Gleichung,
∇
→
2
A
0
(
x
→
)
=
−
4
π
ρ
(
x
→
)
{\displaystyle {\vec {\nabla }}^{2}A_{0}\left({\vec {x}}\right)=-4\pi \rho \left({\vec {x}}\right)}
,
wird erst durch die an sie gestellten Randbedingungen eindeutig.
Zunächst aber ein kleiner mathematischer Exkurs über die Sätze von
Green.
Aus dem Satz von Gauß ergibt sich der Satz I von Green:
∫
V
d
3
x
(
ϕ
∇
→
2
ψ
+
(
∇
→
ϕ
)
⋅
(
∇
→
ψ
)
)
=
∫
V
d
3
x
∇
→
⋅
(
ϕ
∇
→
ψ
)
=
∫
S
=
∂
V
d
2
a
ϕ
n
→
⋅
∇
→
ψ
{\displaystyle {\underset {V}{\int }}d^{3}x\,\left(\phi {\vec {\nabla }}^{2}\psi +\left({\vec {\nabla }}\phi \right)\cdot \left({\vec {\nabla }}\psi \right)\right)={\underset {V}{\int }}d^{3}x\,{\vec {\nabla }}\cdot \left(\phi {\vec {\nabla }}\psi \right)={\underset {S=\partial V}{\int }}d^{2}a\,\phi {\vec {n}}\cdot {\vec {\nabla }}\psi }
und hieraus resultiert wiederum der Satz II von Green:
∫
V
d
3
x
(
ϕ
∇
→
2
ψ
−
ψ
∇
→
2
ϕ
)
=
∫
S
=
∂
V
d
2
a
(
ϕ
n
→
⋅
∇
→
ψ
−
ψ
n
→
⋅
∇
→
ϕ
)
{\displaystyle {\underset {V}{\int }}d^{3}x\left(\phi {\vec {\nabla }}^{2}\psi -\psi {\vec {\nabla }}^{2}\phi \right)={\underset {S=\partial V}{\int }}d^{2}a\left(\phi {\vec {n}}\cdot {\vec {\nabla }}\psi -\psi {\vec {n}}\cdot {\vec {\nabla }}\phi \right)}
.
Setze jetzt in Green II
ϕ
=
A
0
{\displaystyle \phi =A_{0}}
,
ψ
=
1
|
x
→
−
x
→
′
|
{\displaystyle \psi ={\frac {1}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}}
ein und verwende
∇
→
2
1
|
x
→
−
x
→
′
|
=
−
4
π
δ
3
(
x
→
−
x
→
′
)
{\displaystyle {\vec {\nabla }}^{2}{\frac {1}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}=-4\pi \delta ^{3}\left({\vec {x}}-{\vec {x}}^{\prime }\right)}
:
A
0
(
x
→
)
=
∫
V
d
3
x
′
ρ
(
x
→
′
)
|
x
→
−
x
→
′
|
+
1
4
π
∫
S
=
∂
V
d
2
a
′
n
→
′
⋅
∇
→
′
A
0
(
x
→
′
)
|
x
→
−
x
→
′
|
−
1
4
π
∫
S
=
∂
V
d
2
a
′
A
0
(
x
→
′
)
n
→
′
⋅
∇
→
′
1
|
x
→
−
x
→
′
|
{\displaystyle A_{0}\left({\vec {x}}\right)={\underset {V}{\int }}d^{3}x^{\prime }\,{\frac {\rho \left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}+{\frac {1}{4\pi }}{\underset {S=\partial V}{\int }}d^{2}a^{\prime }\,{\frac {{\vec {n}}^{\prime }\cdot {\vec {\nabla }}^{\prime }A_{0}\left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}-{\frac {1}{4\pi }}{\underset {S=\partial V}{\int }}d^{2}a^{\prime }\,A_{0}\left({\vec {x}}^{\prime }\right){\vec {n}}^{\prime }\cdot {\vec {\nabla }}^{\prime }{\frac {1}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}}
.
Als Green'sche Funktion erhalten wir hieraus:
G
(
x
→
,
x
→
′
)
=
1
|
x
→
−
x
→
′
|
+
F
(
x
→
,
x
→
′
)
{\displaystyle G\left({\vec {x}},{\vec {x}}^{\prime }\right)={\frac {1}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}+F\left({\vec {x}},{\vec {x}}^{\prime }\right)}
,
∇
→
2
F
(
x
→
,
x
→
′
)
=
0
,
{\displaystyle {\vec {\nabla }}^{2}F\left({\vec {x}},{\vec {x}}^{\prime }\right)=0,}
für die
∇
→
2
G
(
x
→
,
x
→
′
)
=
∇
→
2
G
(
x
→
−
x
→
′
)
=
−
4
π
δ
3
(
x
→
−
x
→
′
)
{\displaystyle {\vec {\nabla }}^{2}G\left({\vec {x}},{\vec {x}}^{\prime }\right)={\vec {\nabla }}^{2}G\left({\vec {x}}-{\vec {x}}^{\prime }\right)=-4\pi \delta ^{3}\left({\vec {x}}-{\vec {x}}^{\prime }\right)}
gilt, d.h.
A
0
(
x
→
)
=
∫
V
d
3
x
′
ρ
(
x
→
′
)
|
x
→
−
x
→
′
|
+
1
4
π
∫
S
=
∂
V
d
2
a
′
G
(
x
→
−
x
→
′
)
n
→
′
⋅
∇
→
′
A
0
(
x
→
′
)
{\displaystyle A_{0}\left({\vec {x}}\right)={\underset {V}{\int }}d^{3}x^{\prime }\,{\frac {\rho \left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}+{\frac {1}{4\pi }}{\underset {S=\partial V}{\int }}d^{2}a^{\prime }\,G\left({\vec {x}}-{\vec {x}}^{\prime }\right){\vec {n}}^{\prime }\cdot {\vec {\nabla }}^{\prime }A_{0}\left({\vec {x}}^{\prime }\right)}
−
1
4
π
∫
S
=
∂
V
d
2
a
′
A
0
(
x
→
′
)
n
→
′
⋅
∇
→
′
G
(
x
→
−
x
→
′
)
{\displaystyle -{\frac {1}{4\pi }}{\underset {S=\partial V}{\int }}d^{2}a^{\prime }\,A_{0}\left({\vec {x}}^{\prime }\right){\vec {n}}^{\prime }\cdot {\vec {\nabla }}^{\prime }G\left({\vec {x}}-{\vec {x}}^{\prime }\right)}
.
Es können folgende Arten der Randbedingungen unterschieden werden:
(a) Dirichlet'sche Randbedingungen:
G
(
x
→
−
x
→
′
)
=
0
{\displaystyle G\left({\vec {x}}-{\vec {x}}^{\prime }\right)=0}
für
x
→
′
∈
S
{\displaystyle {\vec {x}}^{\prime }\in S}
sowie
(b) Neumann'sche Randbedingungen: für ein
x
→
∈
V
{\displaystyle {\vec {x}}\in V}
gilt mit dem Gauß'schen Satz
−
4
π
=
−
4
π
∫
V
d
3
x
′
δ
3
(
x
→
−
x
→
′
)
=
∫
V
d
3
x
′
∇
→
′
2
G
(
x
→
−
x
→
′
)
=
∫
S
=
∂
V
d
2
a
′
⏟
=
A
n
→
′
⋅
∇
→
′
G
(
x
→
−
x
→
′
)
⇒
{\displaystyle -4\pi =-4\pi {\underset {V}{\int }}d^{3}x^{\prime }\,\delta ^{3}\left({\vec {x}}-{\vec {x}}^{\prime }\right)={\underset {V}{\int }}d^{3}x^{\prime }\,{{\vec {\nabla }}^{\prime }}^{2}G\left({\vec {x}}-{\vec {x}}^{\prime }\right)=\underbrace {{\underset {S=\partial V}{\int }}d^{2}a^{\prime }} _{=A}\,{\vec {n}}^{\prime }\cdot {\vec {\nabla }}^{\prime }G\left({\vec {x}}-{\vec {x}}^{\prime }\right)\;\Rightarrow }
n
→
′
⋅
∇
→
′
G
(
x
→
−
x
→
′
)
=
−
4
π
A
⟶
S
=
∂
R
3
:
A
→
∞
0
{\displaystyle {\vec {n}}^{\prime }\cdot {\vec {\nabla }}^{\prime }G\left({\vec {x}}-{\vec {x}}^{\prime }\right)=-{\frac {4\pi }{A}}{\underset {S=\partial \mathbb {R} ^{3}:\,A\rightarrow \infty }{\longrightarrow }}0}
für ein
x
→
′
∈
S
{\displaystyle {\vec {x}}^{\prime }\in S}
.
Über die Green'sche Funktion können u.a. die folgenden beiden Aussagen
getroffen werden.
Die Green'sche Funktion ist in ihren Argumenten symmetrisch, d.h.
G
(
x
→
,
x
→
′
)
=
G
(
x
→
′
,
x
→
)
{\displaystyle G\left({\vec {x}},{\vec {x}}^{\prime }\right)=G\left({\vec {x}}^{\prime },{\vec {x}}\right)}
:
−
4
π
(
G
(
x
→
,
x
→
′
)
−
G
(
x
→
′
,
x
→
)
)
=
∫
V
d
3
y
[
G
(
x
→
,
y
→
)
(
−
4
π
δ
3
(
x
→
′
−
y
→
)
)
⏟
=
∇
→
y
→
2
G
(
x
→
′
,
y
→
)
−
G
(
x
→
′
,
y
→
)
(
−
4
π
δ
3
(
x
→
−
y
→
)
)
⏟
=
∇
→
y
→
2
G
(
x
→
,
y
→
)
]
{\displaystyle -4\pi \left(G\left({\vec {x}},{\vec {x}}^{\prime }\right)-G\left({\vec {x}}^{\prime },{\vec {x}}\right)\right)={\underset {V}{\int }}d^{3}y\,\left[G\left({\vec {x}},{\vec {y}}\right)\underbrace {\left(-4\pi \delta ^{3}\left({\vec {x}}^{\prime }-{\vec {y}}\right)\right)} _{={\vec {\nabla }}_{\vec {y}}^{2}G\left({\vec {x}}^{\prime },{\vec {y}}\right)}-G\left({\vec {x}}^{\prime },{\vec {y}}\right)\underbrace {\left(-4\pi \delta ^{3}\left({\vec {x}}-{\vec {y}}\right)\right)} _{={\vec {\nabla }}_{\vec {y}}^{2}G\left({\vec {x}},{\vec {y}}\right)}\right]}
=
∫
V
d
3
y
[
G
(
x
→
,
y
→
)
∇
→
y
→
2
G
(
x
→
′
,
y
→
)
−
G
(
x
→
′
,
y
→
)
∇
→
y
→
2
G
(
x
→
,
y
→
)
]
=
G
r
e
e
n
I
I
{\displaystyle ={\underset {V}{\int }}d^{3}y\,\left[G\left({\vec {x}},{\vec {y}}\right){\vec {\nabla }}_{\vec {y}}^{2}G\left({\vec {x}}^{\prime },{\vec {y}}\right)-G\left({\vec {x}}^{\prime },{\vec {y}}\right){\vec {\nabla }}_{\vec {y}}^{2}G\left({\vec {x}},{\vec {y}}\right)\right]{\underset {Green\,II}{=}}}
=
∫
S
=
∂
V
d
2
a
[
G
(
x
→
,
y
→
)
⏟
=
0
,
y
→
∈
S
n
→
y
⋅
∇
→
y
→
G
(
x
→
′
,
y
→
)
−
G
(
x
→
′
,
y
→
)
⏟
=
0
,
y
→
∈
S
n
→
y
⋅
∇
→
y
→
G
(
x
→
,
y
→
)
]
=
0
{\displaystyle ={\underset {S=\partial V}{\int }}d^{2}a\,\left[\underbrace {G\left({\vec {x}},{\vec {y}}\right)} _{=0,\,{\vec {y}}\in S}{\vec {n}}_{y}\cdot {\vec {\nabla }}_{\vec {y}}G\left({\vec {x}}^{\prime },{\vec {y}}\right)-\underbrace {G\left({\vec {x}}^{\prime },{\vec {y}}\right)} _{=0,\,{\vec {y}}\in S}{\vec {n}}_{y}\cdot {\vec {\nabla }}_{\vec {y}}G\left({\vec {x}},{\vec {y}}\right)\right]=0}
wegen der Dirichlet'schen Randbedingung
G
(
x
→
,
y
→
)
=
0
{\displaystyle G\left({\vec {x}},{\vec {y}}\right)=0}
für
y
→
∈
S
{\displaystyle {\vec {y}}\in S}
.
Die Lösung der Poisson-Gleichung wir mit z.B. der Dirichlet'schen
Randbedingung bis auf eine Konstante eindeutig.
Um dies zu zeigen, nehmen wir zunächst das Gegenteil an, d.h. es existieren
zwei Lösungen der Poisson-Gleichung zur selben Ladungsverteilung
ρ
{\displaystyle \rho }
:
∇
→
2
ϕ
1
=
−
4
π
ρ
{\displaystyle {\vec {\nabla }}^{2}\phi _{1}=-4\pi \rho }
,
∇
→
2
ϕ
2
=
−
4
π
ρ
{\displaystyle {\vec {\nabla }}^{2}\phi _{2}=-4\pi \rho }
,
U
:=
ϕ
1
−
ϕ
2
⇒
∇
→
2
U
=
0
{\displaystyle U:=\phi _{1}-\phi _{2}\;\Rightarrow {\vec {\nabla }}^{2}U=0}
Mit Green I folgt hieraus:
∫
V
d
3
x
(
U
∇
→
2
U
⏟
=
0
+
(
∇
→
U
)
2
)
=
∫
S
=
∂
V
d
2
a
U
⏟
=
0
,
x
→
∈
S
n
→
⋅
∇
→
U
=
0
⇒
(
∇
→
U
)
2
=
0
{\displaystyle {\underset {V}{\int }}d^{3}x\,\left(U\,\underbrace {{\vec {\nabla }}^{2}U} _{=0}+\left({\vec {\nabla }}U\right)^{2}\right)={\underset {S=\partial V}{\int }}d^{2}a\,\underbrace {U} _{=0,\,{\vec {x}}\in S}{\vec {n}}\cdot {\vec {\nabla }}U=0\;\Rightarrow \;\left({\vec {\nabla }}U\right)^{2}=0}
,
wenn z.B. Dirichlet'sche Randbedingungen, d.h.
U
(
x
→
)
=
0
,
x
→
∈
S
{\displaystyle U\left({\vec {x}}\right)=0,\;{\vec {x}}\in S}
,
angewandt werden. Wegen
(
∇
→
U
)
2
=
0
{\displaystyle \left({\vec {\nabla }}U\right)^{2}=0}
gilt somit auch
c
o
n
s
t
.
=
U
=
ϕ
1
−
ϕ
2
{\displaystyle const.=U=\phi _{1}-\phi _{2}}
, d.h.
ϕ
1
=
ϕ
2
+
c
o
n
s
t
.
{\displaystyle \phi _{1}=\phi _{2}+const.}
Da wir bereits das Vektorpotential in der Elektrostatik bestimmt haben,
A
→
(
x
→
)
=
μ
c
∫
d
3
x
′
J
→
(
x
→
′
)
|
x
→
−
x
→
′
|
{\displaystyle {\vec {A}}\left({\vec {x}}\right)={\frac {\mu }{c}}\int d^{3}x^{\prime }\,{\frac {{\vec {J}}\left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}}
,
erhalten wir aus
B
→
=
∇
→
×
A
→
{\displaystyle {\vec {B}}={\vec {\nabla }}\times {\vec {A}}}
und
∇
→
1
r
=
−
x
→
r
3
{\displaystyle {\vec {\nabla }}{\frac {1}{r}}=-{\frac {\vec {x}}{r^{3}}}}
unmittelbar:
B
→
(
x
→
)
=
∇
→
×
A
→
(
x
→
)
=
∇
→
×
μ
c
∫
d
3
x
′
J
→
(
x
→
′
)
|
x
→
−
x
→
′
|
=
{\displaystyle {\vec {B}}\left({\vec {x}}\right)={\vec {\nabla }}\times {\vec {A}}\left({\vec {x}}\right)={\vec {\nabla }}\times {\frac {\mu }{c}}\int d^{3}x^{\prime }\,{\frac {{\vec {J}}\left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}=}
−
μ
c
∫
d
3
x
′
J
→
(
x
→
′
)
×
∇
→
1
|
x
→
−
x
→
′
|
=
μ
c
∫
d
3
x
′
J
→
(
x
→
′
)
×
x
→
−
x
→
′
|
x
→
−
x
→
′
|
3
{\displaystyle -{\frac {\mu }{c}}\int d^{3}x^{\prime }\,{\vec {J}}\left({\vec {x}}^{\prime }\right)\times {\vec {\nabla }}{\frac {1}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}={\frac {\mu }{c}}\int d^{3}x^{\prime }\,{\vec {J}}\left({\vec {x}}^{\prime }\right)\times {\frac {{\vec {x}}-{\vec {x}}^{\prime }}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|^{3}}}}
.
Mit
H
→
=
1
μ
B
→
{\displaystyle {\vec {H}}={\frac {1}{\mu }}{\vec {B}}}
resultiert hieraus
der berühmte Biot-Savart'sche Satz,
H
→
(
x
→
)
=
1
c
∫
d
3
x
′
J
→
(
x
→
′
)
×
x
→
−
x
→
′
|
x
→
−
x
→
′
|
3
=
1
c
I
∫
d
x
→
′
×
(
x
→
−
x
→
′
)
|
x
→
−
x
→
′
|
3
{\displaystyle {\vec {H}}\left({\vec {x}}\right)={\frac {1}{c}}\int d^{3}x^{\prime }\,{\vec {J}}\left({\vec {x}}^{\prime }\right)\times {\frac {{\vec {x}}-{\vec {x}}^{\prime }}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|^{3}}}={\frac {1}{c}}I\int {\frac {d{\vec {x}}^{\prime }\times \left({\vec {x}}-{\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|^{3}}}}
,
mit dem in der Elektrostatik z.B. das Magnetfeld eines stromdurchflossenen
Leiters bestimmt werden kann.
Dipolentwicklung des elektrischen Feldes
Bearbeiten
Der Beobachter sei weit von der Quelle des elektrischen Feldes, einer
Ladungsverteilung um
x
→
′
{\displaystyle {\vec {x}}^{\prime }}
, entfernt, d.h.
r
≫
r
′
{\displaystyle r\gg r^{\prime }}
.
Daher kann eine Taylor-Entwicklung bis zur ersten Ordnung in
x
→
′
{\displaystyle {\vec {x}}^{\prime }}
um Null durchgeführt werden:
1
|
x
→
−
x
→
′
|
=
1
r
+
x
→
′
⋅
x
→
r
3
{\displaystyle {\frac {1}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}={\frac {1}{r}}+{\frac {{\vec {x}}^{\prime }\cdot {\vec {x}}}{r^{3}}}}
,
die man auch wie folgt einsehen kann:
|
x
→
−
x
→
′
|
=
(
x
→
−
x
→
′
)
2
=
r
2
+
r
′
2
−
2
r
r
′
cos
ϑ
=
{\displaystyle \left|{\vec {x}}-{\vec {x}}^{\prime }\right|={\sqrt {\left({\vec {x}}-{\vec {x}}^{\prime }\right)^{2}}}={\sqrt {r^{2}+r^{\prime 2}-2rr^{\prime }\cos \vartheta }}=}
r
1
+
(
r
′
r
)
2
−
2
r
′
r
cos
ϑ
≈
r
≫
r
′
r
(
1
−
r
′
r
cos
ϑ
)
{\displaystyle r\,{\sqrt {1+\left({\frac {r^{\prime }}{r}}\right)^{2}-2{\frac {r^{\prime }}{r}}\cos \vartheta }}{\underset {r\gg r^{\prime }}{\approx }}r\left(1-{\frac {r^{\prime }}{r}}\cos \vartheta \right)}
mit dem Winkel
ϑ
{\displaystyle \vartheta }
zwischen
x
→
{\displaystyle {\vec {x}}}
und
x
→
′
{\displaystyle {\vec {x}}^{\prime }}
, d.h.
x
→
′
⋅
x
→
=
r
r
′
cos
ϑ
{\displaystyle {\vec {x}}^{\prime }\cdot {\vec {x}}=rr^{\prime }\cos \vartheta }
.
Somit erhalten wir
1
|
x
→
−
x
→
′
|
≈
r
≫
r
′
1
r
+
r
′
r
cos
ϑ
r
3
=
1
r
+
x
→
′
⋅
x
→
r
3
{\displaystyle {\frac {1}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}{\underset {r\gg r^{\prime }}{\approx }}{\frac {1}{r}}+{\frac {r^{\prime }r\,\cos \vartheta }{r^{3}}}={\frac {1}{r}}+{\frac {{\vec {x}}^{\prime }\cdot {\vec {x}}}{r^{3}}}}
.
Bei Abwesenheit von Dielektrika (d.h.
ε
=
1
{\displaystyle \varepsilon =1}
)
gilt:
A
0
(
x
→
)
=
∫
d
3
x
′
ρ
(
x
→
′
)
|
x
→
−
x
→
′
|
≈
1
r
∫
d
3
x
′
ρ
(
x
→
′
)
⏟
=
q
+
x
→
r
3
⋅
∫
d
3
x
′
ρ
(
x
→
′
)
x
→
′
⏟
=
p
→
=
q
r
+
p
→
⋅
x
→
r
3
{\displaystyle A_{0}\left({\vec {x}}\right)=\int d^{3}x^{\prime }\,{\frac {\rho \left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}\approx {\frac {1}{r}}\underbrace {\int d^{3}x^{\prime }\,\rho \left({\vec {x}}^{\prime }\right)} _{=q}+{\frac {\vec {x}}{r^{3}}}\cdot \underbrace {\int d^{3}x^{\prime }\,\rho \left({\vec {x}}^{\prime }\right){\vec {x}}^{\prime }} _{={\vec {p}}}={\frac {q}{r}}+{\vec {p}}\cdot {\frac {\vec {x}}{r^{3}}}}
,
mit der Ladung
q
=
∫
d
3
x
′
ρ
(
x
→
′
)
{\displaystyle q=\int d^{3}x^{\prime }\,\rho \left({\vec {x}}^{\prime }\right)}
und dem Dipolmoment
p
→
=
∫
d
3
x
′
ρ
(
x
→
′
)
x
→
′
{\displaystyle {\vec {p}}=\int d^{3}x^{\prime }\,\rho \left({\vec {x}}^{\prime }\right){\vec {x}}^{\prime }}
.
A
0
{\displaystyle A_{0}}
ist also eine Summe aus einem Monopolterm (Term
mit q ) und einem Dipolterm (Term mit
p
→
{\displaystyle {\vec {p}}}
).
Aus dem Skalarpotential können wir das elektrische Feld
E
→
=
−
∇
→
A
0
{\displaystyle {\vec {E}}=-{\vec {\nabla }}A_{0}}
,
mittels
∇
→
1
r
=
−
r
→
r
3
{\displaystyle {\vec {\nabla }}{\frac {1}{r}}=-{\frac {\vec {r}}{r^{3}}}}
,
und
∂
∂
x
i
x
j
r
3
=
1
r
3
(
δ
i
j
−
3
x
i
r
x
j
r
)
=
1
r
3
(
δ
i
j
−
3
n
i
n
j
)
{\displaystyle {\frac {\partial }{\partial x_{i}}}{\frac {x_{j}}{r^{3}}}={\frac {1}{r^{3}}}\left(\delta _{ij}-3{\frac {x_{i}}{r}}{\frac {x_{j}}{r}}\right)={\frac {1}{r^{3}}}\left(\delta _{ij}-3n_{i}n_{j}\right)}
errechnen:
E
→
=
−
∇
→
A
0
≈
q
r
→
r
3
+
3
n
→
(
p
→
⋅
n
→
)
−
p
→
r
3
{\displaystyle {\vec {E}}=-{\vec {\nabla }}A_{0}\approx q{\frac {\vec {r}}{r^{3}}}+{\frac {3{\vec {n}}\left({\vec {p}}\cdot {\vec {n}}\right)-{\vec {p}}}{r^{3}}}}
.
Dipolentwicklung des magnetischen Feldes
Bearbeiten
Der Beobachter sie weit von der Quelle des magnetischen Feldes, einer
Stromverteilung um
x
→
′
{\displaystyle {\vec {x}}^{\prime }}
, entfernt, d.h.
r
≫
r
′
{\displaystyle r\gg r^{\prime }}
:
1
|
x
→
−
x
→
′
|
=
1
r
+
x
→
′
⋅
x
→
r
3
{\displaystyle {\frac {1}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}={\frac {1}{r}}+{\frac {{\vec {x}}^{\prime }\cdot {\vec {x}}}{r^{3}}}}
.
Bei Abwesenheit von Magnetika (d.h.
μ
=
1
{\displaystyle \mu =1}
) gilt daher:
A
k
(
x
→
)
=
1
c
∫
d
3
x
′
J
k
(
x
→
′
)
|
x
→
−
x
→
′
|
≈
1
c
1
r
∫
d
3
x
′
J
k
(
x
→
′
)
+
x
→
r
3
⋅
∫
d
3
x
′
J
k
(
x
→
′
)
x
→
′
{\displaystyle A_{k}\left({\vec {x}}\right)={\frac {1}{c}}\int d^{3}x^{\prime }\,{\frac {J_{k}\left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}\approx {\frac {1}{c}}{\frac {1}{r}}\int d^{3}x^{\prime }\,J_{k}\left({\vec {x}}^{\prime }\right)+{\frac {\vec {x}}{r^{3}}}\cdot \int d^{3}x^{\prime }\,J_{k}\left({\vec {x}}^{\prime }\right){\vec {x}}^{\prime }}
Wegen der Quellfreiheit des Stromes in der Elektrostatik, d.h.
∇
→
⋅
J
→
(
x
→
)
=
0
{\displaystyle {\vec {\nabla }}\cdot {\vec {J}}\left({\vec {x}}\right)=0}
,
vereinfacht sich folgendes Integral
∫
V
d
3
x
f
(
x
→
)
J
→
(
x
→
)
⋅
∇
→
g
(
x
→
)
=
∫
V
d
3
x
∇
→
⋅
(
f
(
x
→
)
g
(
x
→
)
J
→
(
x
→
)
)
{\displaystyle {\underset {V}{\int }}d^{3}x\,f\left({\vec {x}}\right){\vec {J}}\left({\vec {x}}\right)\cdot {\vec {\nabla }}g\left({\vec {x}}\right)={\underset {V}{\int }}d^{3}x\,{\vec {\nabla }}\cdot \left(f\left({\vec {x}}\right)g\left({\vec {x}}\right){\vec {J}}\left({\vec {x}}\right)\right)}
−
∫
V
d
3
x
g
(
x
→
)
J
→
(
x
→
)
⋅
∇
→
f
(
x
→
)
−
∫
V
d
3
x
f
(
x
→
)
g
(
x
→
)
∇
→
⋅
J
→
(
x
→
)
⏟
=
0
{\displaystyle -{\underset {V}{\int }}d^{3}x\,g\left({\vec {x}}\right){\vec {J}}\left({\vec {x}}\right)\cdot {\vec {\nabla }}f\left({\vec {x}}\right)-{\underset {V}{\int }}d^{3}x\,f\left({\vec {x}}\right)g\left({\vec {x}}\right)\underbrace {{\vec {\nabla }}\cdot {\vec {J}}\left({\vec {x}}\right)} _{=0}}
,
in dem nach dem Satz von Gauß der Divergenzterm
∫
V
d
3
x
∇
→
⋅
(
f
(
x
→
)
g
(
x
→
)
J
→
(
x
→
)
)
=
∫
A
=
∂
V
d
2
a
n
→
⋅
(
f
(
x
→
)
g
(
x
→
)
J
→
(
x
→
)
)
⟶
r
→
∞
,
∂
V
=
∂
R
3
0
{\displaystyle {\underset {V}{\int }}d^{3}x\,{\vec {\nabla }}\cdot \left(f\left({\vec {x}}\right)g\left({\vec {x}}\right){\vec {J}}\left({\vec {x}}\right)\right)={\underset {A=\partial V}{\int }}d^{2}a\,{\vec {n}}\cdot \left(f\left({\vec {x}}\right)g\left({\vec {x}}\right){\vec {J}}\left({\vec {x}}\right)\right){\underset {r\rightarrow \infty ,\,\partial V=\partial \mathbb {R} ^{3}}{\longrightarrow }}0}
,
auf dem Rande des Integrationsvolumens verschwindet, weil im Unendlichen
kein Strom fließe. Hieraus resultiert der im Folgenden noch oft verwendete
Hilfssatz:
Für
∇
→
⋅
J
→
(
x
→
)
=
0
{\displaystyle {\vec {\nabla }}\cdot {\vec {J}}\left({\vec {x}}\right)=0}
gilt
0
=
∫
d
3
x
(
f
(
x
→
)
J
→
(
x
→
)
⋅
∇
→
g
(
x
→
)
+
g
(
x
→
)
J
→
(
x
→
)
⋅
∇
→
f
(
x
→
)
)
{\displaystyle 0=\int d^{3}x\,\left(f\left({\vec {x}}\right){\vec {J}}\left({\vec {x}}\right)\cdot {\vec {\nabla }}g\left({\vec {x}}\right)+g\left({\vec {x}}\right){\vec {J}}\left({\vec {x}}\right)\cdot {\vec {\nabla }}f\left({\vec {x}}\right)\right)}
.
Hierin setzten wir nacheinander
(a)
f
=
1
,
g
=
x
k
⇒
∇
→
f
=
0
,
∇
→
g
=
e
^
k
{\displaystyle f=1,\,g=x_{k}\;\Rightarrow \;{\vec {\nabla }}f=0,\,{\vec {\nabla }}g={\hat {e}}_{k}}
:
∫
d
3
x
J
k
(
x
→
)
=
∫
d
3
x
J
→
(
x
→
)
⋅
e
^
k
=
∫
d
3
x
J
→
(
x
→
)
⋅
∇
→
x
k
=
0
{\displaystyle \int d^{3}x\,J_{k}\left({\vec {x}}\right)=\int d^{3}x\,{\vec {J}}\left({\vec {x}}\right)\cdot {\hat {e}}_{k}=\int d^{3}x\,{\vec {J}}\left({\vec {x}}\right)\cdot {\vec {\nabla }}x_{k}=0}
.
D.h. der Monopol-Anteil verschwindet!
(b)
f
=
x
i
,
g
=
x
k
⇒
∇
→
f
=
e
^
i
,
∇
→
g
=
e
^
k
{\displaystyle f=x_{i},\,g=x_{k}\;\Rightarrow \;{\vec {\nabla }}f={\hat {e}}_{i},\,{\vec {\nabla }}g={\hat {e}}_{k}}
:
∫
d
3
x
(
x
i
J
k
(
x
→
)
+
x
k
J
i
(
x
→
)
)
=
0
{\displaystyle \int d^{3}x\,\left(x_{i}J_{k}\left({\vec {x}}\right)+x_{k}J_{i}\left({\vec {x}}\right)\right)=0}
,
(c)
x
i
J
k
−
x
k
J
i
=
∑
p
,
q
x
p
J
q
(
δ
i
p
δ
k
q
−
δ
i
q
δ
k
p
)
=
∑
p
,
q
,
l
ε
i
k
l
ε
l
p
q
x
p
J
q
=
∑
l
ε
i
k
l
[
x
→
×
J
→
]
l
{\displaystyle x_{i}J_{k}-x_{k}J_{i}={\underset {p,q}{\sum }}x_{p}J_{q}\left(\delta _{ip}\delta _{kq}-\delta _{iq}\delta _{kp}\right)={\underset {p,q,l}{\sum }}\varepsilon _{ikl}\varepsilon _{lpq}x_{p}J_{q}={\underset {l}{\sum }}\varepsilon _{ikl}\left[{\vec {x}}\times {\vec {J}}\right]_{l}}
.
Aus (a) bis (c) erhalten wir somit für das Vektorpotential
A
k
(
x
→
)
≈
(
a
)
1
c
∑
i
x
i
r
3
∫
d
3
x
′
x
i
′
J
k
(
x
→
′
)
=
1
2
c
∑
i
x
i
r
3
∫
d
3
x
′
(
x
i
′
J
k
(
x
→
′
)
+
x
i
′
J
k
(
x
→
′
)
)
=
(
b
)
{\displaystyle A_{k}\left({\vec {x}}\right){\underset {(a)}{\approx }}{\frac {1}{c}}{\underset {i}{\sum }}{\frac {x_{i}}{r^{3}}}\int d^{3}x^{\prime }\,x_{i}^{\prime }J_{k}\left({\vec {x}}^{\prime }\right)={\frac {1}{2c}}{\underset {i}{\sum }}{\frac {x_{i}}{r^{3}}}\int d^{3}x^{\prime }\,\left(x_{i}^{\prime }J_{k}\left({\vec {x}}^{\prime }\right)+x_{i}^{\prime }J_{k}\left({\vec {x}}^{\prime }\right)\right){\underset {(b)}{=}}}
1
2
c
∑
i
x
i
r
3
∫
d
3
x
′
(
x
i
′
J
k
(
x
→
′
)
−
x
k
′
J
i
(
x
→
′
)
)
=
(
c
)
1
2
c
∑
i
,
l
ε
i
k
l
x
i
r
3
∫
d
3
x
′
[
x
→
′
×
J
→
(
x
→
′
)
]
l
=
{\displaystyle {\frac {1}{2c}}{\underset {i}{\sum }}{\frac {x_{i}}{r^{3}}}\int d^{3}x^{\prime }\,\left(x_{i}^{\prime }J_{k}\left({\vec {x}}^{\prime }\right)-x_{k}^{\prime }J_{i}\left({\vec {x}}^{\prime }\right)\right){\underset {(c)}{=}}{\frac {1}{2c}}{\underset {i,l}{\sum }}\varepsilon _{ikl}{\frac {x_{i}}{r^{3}}}\int d^{3}x^{\prime }\,\left[{\vec {x}}^{\prime }\times {\vec {J}}\left({\vec {x}}^{\prime }\right)\right]_{l}=}
[
1
2
c
∫
d
3
x
′
(
x
→
′
×
J
→
(
x
→
′
)
)
⏟
=:
m
→
×
x
→
r
3
]
k
{\displaystyle \left[\underbrace {{\frac {1}{2c}}\int d^{3}x^{\prime }\,\left({\vec {x}}^{\prime }\times {\vec {J}}\left({\vec {x}}^{\prime }\right)\right)} _{=:{\overrightarrow {m}}}\times {\frac {\vec {x}}{r^{3}}}\right]_{k}}
,
d.h.
A
→
(
x
→
)
≈
m
→
×
x
→
r
3
{\displaystyle {\vec {A}}\left({\vec {x}}\right)\approx {\vec {m}}\times {\frac {\vec {x}}{r^{3}}}}
mit dem magnetischen Moment
m
→
=
1
2
c
∫
d
3
x
′
(
x
→
′
×
J
→
(
x
→
′
)
)
{\displaystyle {\vec {m}}={\frac {1}{2c}}\int d^{3}x^{\prime }\,\left({\vec {x}}^{\prime }\times {\vec {J}}\left({\vec {x}}^{\prime }\right)\right)}
.
Aus dem jetzt bekannten Vektorpotential lässt sich selbstverständlich
wieder ein magnetisches Feld berechnen:
B
→
=
∇
→
×
A
→
≈
−
∇
→
×
(
m
→
×
∇
→
1
r
)
=
m
→
=
c
o
n
s
t
.
∇
→
×
(
∇
→
×
m
→
r
)
=
{\displaystyle {\vec {B}}={\vec {\nabla }}\times {\vec {A}}\approx -{\vec {\nabla }}\times \left({\vec {m}}\times {\vec {\nabla }}{\frac {1}{r}}\right){\underset {{\vec {m}}=const.}{=}}{\vec {\nabla }}\times \left({\vec {\nabla }}\times {\frac {\vec {m}}{r}}\right)=}
=
−
∇
→
2
m
→
r
+
∇
→
(
∇
→
⋅
m
→
r
)
=
−
m
→
∇
→
2
1
r
⏟
=
0
,
r
≠
0
+
∇
→
(
m
→
⋅
∇
→
1
r
)
{\displaystyle =-{\vec {\nabla }}^{2}{\frac {\vec {m}}{r}}+{\vec {\nabla }}\left({\vec {\nabla }}\cdot {\frac {\vec {m}}{r}}\right)=-{\vec {m}}\underbrace {{\vec {\nabla }}^{2}{\frac {1}{r}}} _{=0,\,r\neq 0}+{\vec {\nabla }}\left({\vec {m}}\cdot {\vec {\nabla }}{\frac {1}{r}}\right)}
,
∂
∂
x
i
x
j
r
3
=
1
r
3
(
δ
i
j
−
3
n
i
n
j
)
{\displaystyle {\frac {\partial }{\partial x_{i}}}{\frac {x_{j}}{r^{3}}}={\frac {1}{r^{3}}}\left(\delta _{ij}-3n_{i}n_{j}\right)}
,
n
→
=
x
→
r
{\displaystyle {\vec {n}}={\frac {\vec {x}}{r}}}
,
woraus ein magnetisches Dipolfeld resultiert:
B
→
=
∇
→
×
A
→
≈
3
n
→
(
m
→
⋅
n
→
)
−
m
→
r
3
{\displaystyle {\vec {B}}={\vec {\nabla }}\times {\vec {A}}\approx {\frac {3{\vec {n}}\left({\vec {m}}\cdot {\vec {n}}\right)-{\vec {m}}}{r^{3}}}}
.
Außerdem ist es noch möglich, mittels (a) bis (c) die magnetische
Dipolenergie und das Drehmoment auf eine lokalisierte Stromverteilung
zu bestimmen.
Die magnetische Dipolenergie können wir durch das Betrachten einer
Kraft
F
→
=
q
c
∫
d
3
x
′
J
→
(
x
→
′
)
×
B
→
(
x
→
′
)
{\displaystyle {\vec {F}}={\frac {q}{c}}\int d^{3}x^{\prime }\,{\vec {J}}\left({\vec {x}}^{\prime }\right)\times {\vec {B}}\left({\vec {x}}^{\prime }\right)}
auf eine wenig ausgedehnte Stromverteilung im äußeren Magnetfeld gewinnen,
d.h.
B
i
(
x
→
)
≈
B
i
(
0
)
+
∑
k
x
k
∂
B
i
∂
x
k
(
0
)
{\displaystyle B_{i}\left({\vec {x}}\right)\approx B_{i}\left(0\right)+{\underset {k}{\sum }}x_{k}{\frac {\partial B_{i}}{\partial x_{k}}}\left(0\right)}
:
F
l
≈
1
c
∑
p
,
q
ε
l
p
q
(
B
q
(
0
)
∫
d
3
x
′
J
p
(
x
→
′
)
⏟
=
(
a
)
0
+
∑
k
∂
B
i
∂
x
k
(
0
)
∫
d
3
x
′
x
k
′
J
p
(
x
→
′
)
)
{\displaystyle F_{l}\approx {\frac {1}{c}}{\underset {p,q}{\sum }}\varepsilon _{lpq}\left(B_{q}\left(0\right)\underbrace {\int d^{3}x^{\prime }\,J_{p}\left({\vec {x}}^{\prime }\right)} _{{\underset {(a)}{=}}0}+{\underset {k}{\sum }}{\frac {\partial B_{i}}{\partial x_{k}}}\left(0\right)\int d^{3}x^{\prime }\,x_{k}^{\prime }J_{p}\left({\vec {x}}^{\prime }\right)\right)}
=
1
2
c
∑
p
,
q
ε
l
p
q
∑
k
∂
B
q
∂
x
k
(
0
)
∫
d
3
x
′
(
x
k
′
J
p
(
x
→
′
)
+
x
k
′
J
p
(
x
→
′
)
)
{\displaystyle ={\frac {1}{2c}}{\underset {p,q}{\sum }}\varepsilon _{lpq}{\underset {k}{\sum }}{\frac {\partial B_{q}}{\partial x_{k}}}\left(0\right)\int d^{3}x^{\prime }\,\left(x_{k}^{\prime }J_{p}\left({\vec {x}}^{\prime }\right)+x_{k}^{\prime }J_{p}\left({\vec {x}}^{\prime }\right)\right)}
=
(
b
)
1
2
c
∑
p
,
q
ε
l
p
q
∑
k
∂
B
q
∂
x
k
(
0
)
∫
d
3
x
′
(
x
k
′
J
p
(
x
→
′
)
−
x
p
′
J
k
(
x
→
′
)
)
{\displaystyle {\underset {(b)}{=}}{\frac {1}{2c}}{\underset {p,q}{\sum }}\varepsilon _{lpq}{\underset {k}{\sum }}{\frac {\partial B_{q}}{\partial x_{k}}}\left(0\right)\int d^{3}x^{\prime }\,\left(x_{k}^{\prime }J_{p}\left({\vec {x}}^{\prime }\right)-x_{p}^{\prime }J_{k}\left({\vec {x}}^{\prime }\right)\right)}
=
(
c
)
1
2
c
∑
p
,
q
ε
l
p
q
∑
k
∂
B
q
∂
x
k
(
0
)
∑
r
ε
k
p
s
∫
d
3
x
′
[
x
→
′
×
J
→
(
x
→
′
)
]
s
{\displaystyle {\underset {(c)}{=}}{\frac {1}{2c}}{\underset {p,q}{\sum }}\varepsilon _{lpq}{\underset {k}{\sum }}{\frac {\partial B_{q}}{\partial x_{k}}}\left(0\right){\underset {r}{\sum }}\varepsilon _{kps}\int d^{3}x^{\prime }\,\left[{\vec {x}}^{\prime }\times {\vec {J}}\left({\vec {x}}^{\prime }\right)\right]_{s}}
=
1
2
c
∑
k
,
s
,
q
∂
B
q
∂
x
k
(
0
)
∫
d
3
x
′
[
x
→
′
×
J
→
(
x
→
′
)
]
s
⏟
=
2
c
m
→
∑
p
ε
l
q
p
ε
p
k
s
⏟
=
δ
l
k
δ
q
s
−
δ
l
s
δ
q
k
{\displaystyle ={\frac {1}{2c}}{\underset {k,s,q}{\sum }}{\frac {\partial B_{q}}{\partial x_{k}}}\left(0\right)\underbrace {\int d^{3}x^{\prime }\,\left[{\vec {x}}^{\prime }\times {\vec {J}}\left({\vec {x}}^{\prime }\right)\right]_{s}} _{=2c{\vec {m}}}\underbrace {{\underset {p}{\sum }}\varepsilon _{lqp}\varepsilon _{pks}} _{=\delta _{lk}\delta _{qs}-\delta _{ls}\delta _{qk}}}
=
∂
B
→
∂
x
l
(
0
)
⋅
m
→
−
m
l
(
∇
→
⋅
B
→
)
⏟
=
0
(
0
)
=
−
∂
∂
x
l
(
−
m
→
⋅
B
→
)
{\displaystyle ={\frac {\partial {\vec {B}}}{\partial x_{l}}}\left(0\right)\cdot {\vec {m}}-m_{l}\underbrace {\left({\vec {\nabla }}\cdot {\vec {B}}\right)} _{=0}\left(0\right)=-{\frac {\partial }{\partial x_{l}}}\left(-{\vec {m}}\cdot {\vec {B}}\right)}
,
weil ja
m
→
=
c
o
n
s
t
.
{\displaystyle {\vec {m}}=const.}
Die magnetische Dipolenergie beträgt somit:
U
=
−
m
→
⋅
B
→
{\displaystyle U=-{\vec {m}}\cdot {\vec {B}}}
.
Drehmoment auf eine lokalisierte Stromverteilung:
Aus
d
F
→
≈
1
c
J
→
(
x
→
)
×
B
→
(
0
)
d
3
x
{\displaystyle d{\vec {F}}\approx {\frac {1}{c}}{\vec {J}}\left({\vec {x}}\right)\times {\vec {B}}\left(0\right)d^{3}x}
folgt unmittelbar
d
N
→
=
x
→
×
d
F
→
≈
1
c
x
→
×
(
J
→
(
x
→
)
×
B
→
(
0
)
)
d
3
x
=
{\displaystyle d{\vec {N}}={\vec {x}}\times d{\vec {F}}\approx {\frac {1}{c}}{\vec {x}}\times \left({\vec {J}}\left({\vec {x}}\right)\times {\vec {B}}\left(0\right)\right)d^{3}x=}
1
c
[
(
B
→
(
0
)
⋅
x
→
)
J
→
(
x
→
)
−
(
J
→
(
x
→
)
⋅
x
→
)
B
→
(
0
)
]
d
3
x
{\displaystyle {\frac {1}{c}}\left[\left({\vec {B}}\left(0\right)\cdot {\vec {x}}\right){\vec {J}}\left({\vec {x}}\right)-\left({\vec {J}}\left({\vec {x}}\right)\cdot {\vec {x}}\right){\vec {B}}\left(0\right)\right]d^{3}x}
.
Für
∇
→
⋅
J
→
(
x
→
)
=
0
{\displaystyle {\vec {\nabla }}\cdot {\vec {J}}\left({\vec {x}}\right)=0}
gilt ja:
0
=
∫
d
3
x
(
f
(
x
→
)
J
→
(
x
→
)
⋅
∇
→
g
(
x
→
)
+
g
(
x
→
)
J
→
(
x
→
)
⋅
∇
→
f
(
x
→
)
)
{\displaystyle 0=\int d^{3}x\,\left(f\left({\vec {x}}\right){\vec {J}}\left({\vec {x}}\right)\cdot {\vec {\nabla }}g\left({\vec {x}}\right)+g\left({\vec {x}}\right){\vec {J}}\left({\vec {x}}\right)\cdot {\vec {\nabla }}f\left({\vec {x}}\right)\right)}
.
Hierin setzen wir ein:
(
g
=
f
=
r
⇒
∇
→
g
=
∇
→
f
=
x
→
r
=
e
^
r
)
⇒
0
=
∫
d
3
x
(
J
→
(
x
→
)
⋅
x
→
)
{\displaystyle \left(g=f=r\,\Rightarrow \,{\vec {\nabla }}g={\vec {\nabla }}f={\frac {\vec {x}}{r}}={\hat {e}}_{r}\right)\,\Rightarrow 0=\int d^{3}x\,\left({\vec {J}}\left({\vec {x}}\right)\cdot {\vec {x}}\right)}
,
woraus sich für das Drehmoment
N
i
=
1
c
B
→
(
0
)
⋅
∫
d
3
x
x
→
J
i
(
x
→
)
=
1
2
c
∑
k
B
k
(
0
)
∫
d
3
x
(
x
k
J
i
+
x
k
J
i
)
{\displaystyle N_{i}={\frac {1}{c}}{\vec {B}}\left(0\right)\cdot \int d^{3}x\,{\vec {x}}J_{i}\left({\vec {x}}\right)={\frac {1}{2c}}{\underset {k}{\sum }}B_{k}\left(0\right)\int d^{3}x\,\left(x_{k}J_{i}+x_{k}J_{i}\right)}
=
(
b
)
1
2
c
∑
k
B
k
(
0
)
∫
d
3
x
(
x
k
J
i
−
x
i
J
k
)
{\displaystyle {\underset {(b)}{=}}{\frac {1}{2c}}{\underset {k}{\sum }}B_{k}\left(0\right)\int d^{3}x\,\left(x_{k}J_{i}-x_{i}J_{k}\right)}
=
(
c
)
1
2
c
∑
k
,
l
ε
k
i
l
B
k
(
0
)
∫
d
3
x
[
x
→
×
J
→
]
l
⏟
=
2
c
m
l
=
m
→
×
B
→
(
0
)
{\displaystyle {\underset {(c)}{=}}{\frac {1}{2c}}{\underset {k,l}{\sum }}\varepsilon _{kil}\,B_{k}\left(0\right)\underbrace {\int d^{3}x\,\left[{\vec {x}}\times {\vec {J}}\right]_{l}} _{=2cm_{l}}={\vec {m}}\times {\vec {B}}\left(0\right)}
ergibt.
Skalarpotential bei Anwesenheit von Dielektrika
Bearbeiten
Bei Anwesenheit von Dielektrika, die nicht unbedingt isotrop sein
müssen, gilt:
ρ
~
=
ρ
−
∇
→
⋅
P
→
{\displaystyle {\tilde {\rho }}=\rho -{\vec {\nabla }}\cdot {\vec {P}}}
.
Für das Skalarpotential ergibt sich somit
A
0
(
x
→
)
=
∫
d
3
x
′
ρ
~
(
x
→
′
)
|
x
→
−
x
→
′
|
=
∫
d
3
x
′
(
ρ
−
∇
→
′
⋅
P
→
)
(
x
→
′
)
|
x
→
−
x
→
′
|
=
{\displaystyle A_{0}\left({\vec {x}}\right)=\int d^{3}x^{\prime }\,{\frac {{\tilde {\rho }}\left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}=\int d^{3}x^{\prime }\,{\frac {\left(\rho -{\vec {\nabla }}^{\prime }\cdot {\vec {P}}\right)\left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}=}
∫
d
3
x
′
ρ
(
x
→
′
)
|
x
→
−
x
→
′
|
−
∫
d
3
x
′
∇
→
′
⋅
(
P
→
(
x
→
′
)
|
x
→
−
x
→
′
|
)
⏟
⟶
0
,
G
a
u
s
s
+
∫
d
3
x
′
P
→
(
x
→
′
)
⋅
∇
→
′
1
|
x
→
−
x
→
′
|
=
{\displaystyle \int d^{3}x^{\prime }\,{\frac {\rho \left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}-\underbrace {\int d^{3}x^{\prime }\,{\vec {\nabla }}^{\prime }\cdot \left({\frac {{\vec {P}}\left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}\right)} _{\longrightarrow 0,\,Gauss}+\int d^{3}x^{\prime }\,{\vec {P}}\left({\vec {x}}^{\prime }\right)\cdot {\vec {\nabla }}^{\prime }{\frac {1}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}=}
∫
d
3
x
′
ρ
(
x
→
′
)
|
x
→
−
x
→
′
|
−
∫
d
3
x
′
P
→
(
x
→
′
)
⋅
∇
→
1
|
x
→
−
x
→
′
|
=
{\displaystyle \int d^{3}x^{\prime }\,{\frac {\rho \left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}-\int d^{3}x^{\prime }\,{\vec {P}}\left({\vec {x}}^{\prime }\right)\cdot {\vec {\nabla }}{\frac {1}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}=}
∫
d
3
x
′
ρ
(
x
→
′
)
|
x
→
−
x
→
′
|
+
∫
d
3
x
′
P
→
(
x
→
′
)
⋅
x
→
−
x
→
′
|
x
→
−
x
→
′
|
3
{\displaystyle \int d^{3}x^{\prime }\,{\frac {\rho \left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}+\int d^{3}x^{\prime }\,{\vec {P}}\left({\vec {x}}^{\prime }\right)\cdot {\frac {{\vec {x}}-{\vec {x}}^{\prime }}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|^{3}}}}
,
wobei wir hier den Gauß'schen Integralsatz verbunden mit der Tatsache
verwendet haben, dass im Unendlichen keine polarisierbare Materie vorhanden
ist. Außerdem haben wir
∇
→
′
1
|
x
→
−
x
→
′
|
=
−
∇
→
1
|
x
→
−
x
→
′
|
=
x
→
−
x
→
′
|
x
→
−
x
→
′
|
3
{\displaystyle {\vec {\nabla }}^{\prime }{\frac {1}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}=-{\vec {\nabla }}{\frac {1}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}={\frac {{\vec {x}}-{\vec {x}}^{\prime }}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|^{3}}}}
genutzt.
Deutung des Ergebnisses: Das Skalarpotential
A
0
(
x
→
)
=
∫
d
3
x
′
ρ
(
x
→
′
)
|
x
→
−
x
→
′
|
+
∫
d
3
x
′
P
→
(
x
→
′
)
⋅
x
→
−
x
→
′
|
x
→
−
x
→
′
|
3
{\displaystyle A_{0}\left({\vec {x}}\right)=\int d^{3}x^{\prime }\,{\frac {\rho \left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}+\int d^{3}x^{\prime }\,{\vec {P}}\left({\vec {x}}^{\prime }\right)\cdot {\frac {{\vec {x}}-{\vec {x}}^{\prime }}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|^{3}}}}
entsteht aus der Dipolentwicklung von
A
0
{\displaystyle A_{0}}
,
A
0
(
x
→
)
≈
q
r
+
p
→
⋅
x
→
r
3
{\displaystyle A_{0}\left({\vec {x}}\right)\approx {\frac {q}{r}}+{\vec {p}}\cdot {\frac {\vec {x}}{r^{3}}}}
,
durch eine Volumenintegration über:
d
A
0
(
x
→
)
=
A
0
(
x
→
′
)
δ
3
(
x
→
−
x
→
′
)
d
3
x
′
≈
d
q
|
x
→
−
x
→
′
|
+
d
p
→
⋅
x
→
−
x
→
′
|
x
→
−
x
→
′
|
3
{\displaystyle dA_{0}\left({\vec {x}}\right)=A_{0}\left({\vec {x}}^{\prime }\right)\delta ^{3}\left({\vec {x}}-{\vec {x}}^{\prime }\right)d^{3}x^{\prime }\approx {\frac {dq}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}+d{\vec {p}}\cdot {\frac {{\vec {x}}-{\vec {x}}^{\prime }}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|^{3}}}}
mit
d
q
=
ρ
(
x
→
′
)
d
3
x
′
,
d
p
→
=
P
→
(
x
→
′
)
d
3
x
′
{\displaystyle dq=\rho \left({\vec {x}}^{\prime }\right)d^{3}x^{\prime },\;d{\vec {p}}={\vec {P}}\left({\vec {x}}^{\prime }\right)d^{3}x^{\prime }}
.
Vektorpotential bei Anwesenheit von Magnetika
Bearbeiten
Bei Anwesenheit von Magnetika, die nicht unbedingt isotrop sein müssen,
gilt:
J
→
~
=
J
→
+
c
∇
→
×
M
→
+
∂
t
P
→
{\displaystyle {\tilde {\vec {J}}}={\vec {J}}+c{\vec {\nabla }}\times {\vec {M}}+\partial _{t}{\vec {P}}}
.
Zeitableitungen werden aber in der Elektrostatik vernachlässigt -
daher:
J
→
~
=
J
→
+
c
∇
→
×
M
→
{\displaystyle {\tilde {\vec {J}}}={\vec {J}}+c{\vec {\nabla }}\times {\vec {M}}}
.
Diese Stromdichte erfüllt dann auch wieder die Divergenzfreiheit (Kontinuitätsgleichung):
∇
→
⋅
J
→
~
=
0
{\displaystyle {\vec {\nabla }}\cdot {\tilde {\vec {J}}}=0}
, weil
d
i
v
r
o
t
≡
0
{\displaystyle div\,rot\equiv 0}
.
D.h. wir erhalten
A
→
(
x
→
)
=
1
c
∫
d
3
x
′
J
→
~
(
x
→
′
)
|
x
→
−
x
→
′
|
=
1
c
∫
d
3
x
′
J
→
(
x
→
′
)
|
x
→
−
x
→
′
|
+
∫
d
3
x
′
∇
→
′
×
M
→
(
x
→
′
)
|
x
→
−
x
→
′
|
{\displaystyle {\vec {A}}\left({\vec {x}}\right)={\frac {1}{c}}\int d^{3}x^{\prime }\,{\frac {{\tilde {\vec {J}}}\left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}={\frac {1}{c}}\int d^{3}x^{\prime }\,{\frac {{\vec {J}}\left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}+\int d^{3}x^{\prime }\,{\frac {{\vec {\nabla }}^{\prime }\times {\vec {M}}\left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}}
.
Im zweiten Integral können wir den Integranden mittels
∇
→
′
×
(
M
→
(
x
→
′
)
|
x
→
−
x
→
′
|
)
=
∇
→
′
×
M
→
(
x
→
′
)
|
x
→
−
x
→
′
|
−
M
→
(
x
→
′
)
×
∇
→
′
1
|
x
→
−
x
→
′
|
{\displaystyle {\vec {\nabla }}^{\prime }\times \left({\frac {{\vec {M}}\left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}\right)={\frac {{\vec {\nabla }}^{\prime }\times {\vec {M}}\left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}-{\vec {M}}\left({\vec {x}}^{\prime }\right)\times {\vec {\nabla }}^{\prime }{\frac {1}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}}
umformen:
∫
d
3
x
′
∇
→
′
×
M
→
(
x
→
′
)
|
x
→
−
x
→
′
|
=
∫
d
3
x
′
∇
→
′
×
(
M
→
(
x
→
′
)
|
x
→
−
x
→
′
|
)
+
∫
d
3
x
′
M
→
(
x
→
′
)
×
∇
→
′
1
|
x
→
−
x
→
′
|
{\displaystyle \int d^{3}x^{\prime }\,{\frac {{\vec {\nabla }}^{\prime }\times {\vec {M}}\left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}=\int d^{3}x^{\prime }\,{\vec {\nabla }}^{\prime }\times \left({\frac {{\vec {M}}\left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}\right)+\int d^{3}x^{\prime }\,{\vec {M}}\left({\vec {x}}^{\prime }\right)\times {\vec {\nabla }}^{\prime }{\frac {1}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}}
.
Es gibt zudem eine spezielle Variante des Gauß'schen bzw. Stokes'schen
Satzes, die wir hier anwenden können:
∫
V
d
3
x
∇
→
×
F
→
=
∫
A
=
∂
V
d
a
→
×
F
→
{\displaystyle {\underset {V}{\int }}d^{3}x\,{\vec {\nabla }}\times {\vec {F}}={\underset {A=\partial V}{\int }}d{\vec {a}}\times {\vec {F}}}
,
wobei
d
a
→
=
d
2
a
n
→
{\displaystyle d{\vec {a}}=d^{2}a\,{\vec {n}}}
. Der folgende Oberflächenterm
verschwinde:
∫
d
3
x
′
∇
→
′
×
(
M
→
(
x
→
′
)
|
x
→
−
x
→
′
|
)
=
∫
A
=
∂
R
3
d
a
→
×
(
M
→
(
x
→
′
)
|
x
→
−
x
→
′
|
)
⟶
0
{\displaystyle \int d^{3}x^{\prime }\,{\vec {\nabla }}^{\prime }\times \left({\frac {{\vec {M}}\left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}\right)={\underset {A=\partial \mathbb {R} ^{3}}{\int }}d{\vec {a}}\times \left({\frac {{\vec {M}}\left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}\right)\longrightarrow 0}
.
Somit erhalten wir
∫
d
3
x
′
M
→
(
x
→
′
)
×
∇
→
′
1
|
x
→
−
x
→
′
|
=
∫
d
3
x
′
M
→
(
x
→
′
)
×
(
x
→
−
x
→
′
)
|
x
→
−
x
→
′
|
3
{\displaystyle \int d^{3}x^{\prime }\,{\vec {M}}\left({\vec {x}}^{\prime }\right)\times {\vec {\nabla }}^{\prime }{\frac {1}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}=\int d^{3}x^{\prime }{\frac {{\vec {M}}\left({\vec {x}}^{\prime }\right)\times \left({\vec {x}}-{\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|^{3}}}}
,
woraus schließlich
A
→
(
x
→
)
=
1
c
∫
d
3
x
′
J
→
~
(
x
→
′
)
|
x
→
−
x
→
′
|
=
1
c
∫
d
3
x
′
J
→
(
x
→
′
)
|
x
→
−
x
→
′
|
+
∫
d
3
x
′
M
→
(
x
→
′
)
×
(
x
→
−
x
→
′
)
|
x
→
−
x
→
′
|
3
{\displaystyle {\vec {A}}\left({\vec {x}}\right)={\frac {1}{c}}\int d^{3}x^{\prime }\,{\frac {{\tilde {\vec {J}}}\left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}={\frac {1}{c}}\int d^{3}x^{\prime }\,{\frac {{\vec {J}}\left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}+\int d^{3}x^{\prime }{\frac {{\vec {M}}\left({\vec {x}}^{\prime }\right)\times \left({\vec {x}}-{\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|^{3}}}}
resultiert.
Hierauf kommen wir auch über (s. das Kap. über das skalare Potential
bei Anwesenheit von Dielektrika)
A
→
(
x
→
)
≈
m
→
×
x
→
r
3
{\displaystyle {\vec {A}}\left({\vec {x}}\right)\approx {\vec {m}}\times {\frac {\vec {x}}{r^{3}}}}
:
A
→
(
x
→
)
=
∫
d
A
→
(
x
→
)
{\displaystyle {\vec {A}}\left({\vec {x}}\right)=\int d{\vec {A}}\left({\vec {x}}\right)}
entsteht durch Volumenintegration über
d
A
→
(
x
→
)
=
A
→
(
x
→
′
)
δ
3
(
x
→
−
x
→
′
)
d
3
x
′
≈
1
c
d
J
→
|
x
→
−
x
→
′
|
+
d
m
→
×
x
→
−
x
→
′
|
x
→
−
x
→
′
|
3
,
d
J
→
=
J
→
(
x
→
′
)
d
3
x
′
,
d
m
→
=
M
→
(
x
→
′
)
d
3
x
′
{\displaystyle d{\vec {A}}\left({\vec {x}}\right)={\vec {A}}\left({\vec {x}}^{\prime }\right)\delta ^{3}\left({\vec {x}}-{\vec {x}}^{\prime }\right)d^{3}x^{\prime }\approx {\frac {1}{c}}{\frac {d{\vec {J}}}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}+d{\vec {m}}\times {\frac {{\vec {x}}-{\vec {x}}^{\prime }}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|^{3}}},\;d{\vec {J}}={\vec {J}}\left({\vec {x}}^{\prime }\right)d^{3}x^{\prime },\;d{\vec {m}}={\vec {M}}\left({\vec {x}}^{\prime }\right)d^{3}x^{\prime }}
.
Daraus können wir die magnetische Flussdichte bestimmen:
B
→
(
x
→
)
=
∇
→
×
A
→
(
x
→
)
=
∇
→
×
1
c
∫
d
3
x
′
J
→
(
x
→
′
)
|
x
→
−
x
→
′
|
⏟
=
H
→
(
x
→
)
+
∇
→
×
∫
d
3
x
′
M
→
(
x
→
′
)
×
(
x
→
−
x
→
′
)
|
x
→
−
x
→
′
|
3
{\displaystyle {\vec {B}}\left({\vec {x}}\right)={\vec {\nabla }}\times {\vec {A}}\left({\vec {x}}\right)=\underbrace {{\vec {\nabla }}\times {\frac {1}{c}}\int d^{3}x^{\prime }{\frac {{\vec {J}}\left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}} _{={\vec {H}}\left({\vec {x}}\right)}+{\vec {\nabla }}\times \int d^{3}x^{\prime }{\frac {{\vec {M}}\left({\vec {x}}^{\prime }\right)\times \left({\vec {x}}-{\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|^{3}}}}
=
H
→
(
x
→
)
−
∇
→
×
∫
d
3
x
′
M
→
(
x
→
′
)
×
∇
→
1
|
x
→
−
x
→
′
|
{\displaystyle ={\vec {H}}\left({\vec {x}}\right)-{\vec {\nabla }}\times \int d^{3}x^{\prime }\,{\vec {M}}\left({\vec {x}}^{\prime }\right)\times {\vec {\nabla }}{\frac {1}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}}
=
H
→
(
x
→
)
+
∇
→
×
∫
d
3
x
′
∇
→
×
M
→
(
x
→
′
)
|
x
→
−
x
→
′
|
=
H
→
(
x
→
)
+
∇
→
×
∇
→
×
∫
d
3
x
′
M
→
(
x
→
′
)
|
x
→
−
x
→
′
|
{\displaystyle ={\vec {H}}\left({\vec {x}}\right)+{\vec {\nabla }}\times \int d^{3}x^{\prime }\,{\vec {\nabla }}\times {\frac {{\vec {M}}\left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}={\vec {H}}\left({\vec {x}}\right)+{\vec {\nabla }}\times {\vec {\nabla }}\times \int d^{3}x^{\prime }\,{\frac {{\vec {M}}\left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}}
.
Der letzte Summand lässt sich noch weiter umformen:
∇
→
×
(
∇
→
×
∫
d
3
x
′
M
→
(
x
→
′
)
|
x
→
−
x
→
′
|
)
=
∇
→
(
∇
→
⋅
∫
d
3
x
′
M
→
(
x
→
′
)
|
x
→
−
x
→
′
|
)
−
∇
→
2
∫
d
3
x
′
M
→
(
x
→
′
)
|
x
→
−
x
→
′
|
{\displaystyle {\vec {\nabla }}\times \left({\vec {\nabla }}\times \int d^{3}x^{\prime }\,{\frac {{\vec {M}}\left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}\right)={\vec {\nabla }}\left({\vec {\nabla }}\cdot \int d^{3}x^{\prime }\,{\frac {{\vec {M}}\left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}\right)-{\vec {\nabla }}^{2}\int d^{3}x^{\prime }\,{\frac {{\vec {M}}\left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}}
,
worin wiederum der letzte Term berechnet werden kann:
−
∇
→
2
∫
d
3
x
′
M
→
(
x
→
′
)
|
x
→
−
x
→
′
|
=
−
∫
d
3
x
′
M
→
(
x
→
′
)
∇
→
2
1
|
x
→
−
x
→
′
|
=
4
π
∫
d
3
x
′
M
→
(
x
→
′
)
δ
3
(
x
→
−
x
→
′
)
=
4
π
M
→
(
x
→
)
{\displaystyle -{\vec {\nabla }}^{2}\int d^{3}x^{\prime }\,{\frac {{\vec {M}}\left({\vec {x}}^{\prime }\right)}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}=-\int d^{3}x^{\prime }\,{\vec {M}}\left({\vec {x}}^{\prime }\right){\vec {\nabla }}^{2}{\frac {1}{\left|{\vec {x}}-{\vec {x}}^{\prime }\right|}}=4\pi \int d^{3}x^{\prime }\,{\vec {M}}\left({\vec {x}}^{\prime }\right)\delta ^{3}\left({\vec {x}}-{\vec {x}}^{\prime }\right)=4\pi {\vec {M}}\left({\vec {x}}\right)}
.
Aus der Rotation dieser Gleichung für
B
→
{\displaystyle {\vec {B}}}
und
der Maxwell-Gleichung
∇
→
×
H
→
=
4
π
c
J
→
{\displaystyle {\vec {\nabla }}\times {\vec {H}}={\frac {4\pi }{c}}{\vec {J}}}
erhalten wir
∇
→
×
B
→
=
∇
→
×
H
→
+
4
π
∇
→
×
M
→
=
∇
→
×
(
H
→
+
4
π
M
→
)
=
4
π
c
(
J
→
+
c
∇
→
×
M
→
)
{\displaystyle {\vec {\nabla }}\times {\vec {B}}={\vec {\nabla }}\times {\vec {H}}+4\pi {\vec {\nabla }}\times {\vec {M}}={\vec {\nabla }}\times \left({\vec {H}}+4\pi {\vec {M}}\right)={\frac {4\pi }{c}}\left({\vec {J}}+c{\vec {\nabla }}\times {\vec {M}}\right)}
,
woraus wir schließlich
B
→
=
H
→
+
4
π
M
→
{\displaystyle {\vec {B}}={\vec {H}}+4\pi {\vec {M}}}
ablesen können.
Elektrostatische Energie in Dielektrika
Bearbeiten
Die potentielle Energie einer Probeladung q im elektrischen Feld
beträgt
E
p
o
t
=
q
A
0
{\displaystyle E_{pot}=qA_{0}}
,
woraus sich
δ
E
p
o
t
=
∫
d
3
x
δ
ρ
(
x
→
)
⏟
1
4
π
∇
→
⋅
δ
D
→
(
x
→
)
A
0
(
x
→
)
=
1
4
π
∫
d
3
x
A
0
(
x
→
)
∇
→
⋅
δ
D
→
(
x
→
)
{\displaystyle \delta E_{pot}=\int d^{3}x\underbrace {\delta \rho \left({\vec {x}}\right)} _{{\frac {1}{4\pi }}{\vec {\nabla }}\cdot \delta {\vec {D}}\left({\vec {x}}\right)}A_{0}\left({\vec {x}}\right)={\frac {1}{4\pi }}\int d^{3}x\,A_{0}\left({\vec {x}}\right){\vec {\nabla }}\cdot \delta {\vec {D}}\left({\vec {x}}\right)}
ergibt, wenn man z.B. die Platten eines Kondensators auflädt. Hieraus
erhalten wir durch partielle Integration mittels Green I, d.h.
∫
V
d
3
x
(
ϕ
∇
→
2
ψ
+
(
∇
→
ϕ
)
⋅
(
∇
→
ψ
)
)
=
∫
S
=
∂
V
d
2
a
ϕ
n
→
⋅
∇
→
ψ
{\displaystyle {\underset {V}{\int }}d^{3}x\,\left(\phi {\vec {\nabla }}^{2}\psi +\left({\vec {\nabla }}\phi \right)\cdot \left({\vec {\nabla }}\psi \right)\right)={\underset {S=\partial V}{\int }}d^{2}a\,\phi {\vec {n}}\cdot {\vec {\nabla }}\psi }
mit
ϕ
=
A
0
{\displaystyle \phi =A_{0}}
und
∇
→
ψ
=
δ
D
→
{\displaystyle {\vec {\nabla }}\psi =\delta {\vec {D}}}
,
δ
E
p
o
t
=
−
1
4
π
∫
R
3
d
3
x
δ
D
→
⋅
∇
→
A
0
+
∫
∂
R
3
d
2
a
A
0
n
→
⋅
δ
D
→
{\displaystyle \delta E_{pot}=-{\frac {1}{4\pi }}{\underset {\mathbb {R} ^{3}}{\int }}d^{3}x\,\delta {\vec {D}}\cdot {\vec {\nabla }}A_{0}+{\underset {\mathbb {\partial R} ^{3}}{\int }}d^{2}a\,A_{0}{\vec {n}}\cdot \delta {\vec {D}}}
.
Das Oberflächenintegral
∫
∂
R
3
d
2
a
A
0
n
→
⋅
δ
D
→
⟶
0
{\displaystyle {\underset {\mathbb {\partial R} ^{3}}{\int }}d^{2}a\,A_{0}{\vec {n}}\cdot \delta {\vec {D}}\longrightarrow 0}
,
da entsprechend dem Coulomb-Feld
D
→
∝
x
→
r
3
{\displaystyle {\vec {D}}\propto {\frac {\vec {x}}{r^{3}}}}
,
A
0
∝
1
r
{\displaystyle A_{0}\propto {\frac {1}{r}}}
und
n
→
=
x
→
r
{\displaystyle {\vec {n}}={\frac {\vec {x}}{r}}}
,
d
2
a
∝
r
2
{\displaystyle d^{2}a\propto r^{2}}
(Kugelkoordinaten), sodass
d
2
a
A
0
n
→
⋅
δ
D
→
∝
1
r
⟶
r
→
∞
0
{\displaystyle d^{2}a\,A_{0}{\vec {n}}\cdot \delta {\vec {D}}\propto {\frac {1}{r}}{\underset {r\rightarrow \infty }{\longrightarrow }}0}
.
Wegen
E
→
=
−
∇
→
A
0
{\displaystyle {\vec {E}}=-{\vec {\nabla }}A_{0}}
erhalten wir schließlich
δ
E
p
o
t
=
−
1
4
π
∫
R
3
d
3
x
δ
D
→
⋅
∇
→
A
0
=
1
4
π
∫
R
3
d
3
x
δ
D
→
⋅
E
→
{\displaystyle \delta E_{pot}=-{\frac {1}{4\pi }}{\underset {\mathbb {R} ^{3}}{\int }}d^{3}x\,\delta {\vec {D}}\cdot {\vec {\nabla }}A_{0}={\frac {1}{4\pi }}{\underset {\mathbb {R} ^{3}}{\int }}d^{3}x\,\delta {\vec {D}}\cdot {\vec {E}}}
.
Bei einem linearen Zusammenhang zwischen
E
→
{\displaystyle {\vec {E}}}
und
D
→
{\displaystyle {\vec {D}}}
, d.h.
δ
D
→
⋅
E
→
=
1
2
δ
(
D
→
⋅
E
→
)
{\displaystyle \delta {\vec {D}}\cdot {\vec {E}}={\frac {1}{2}}\delta \left({\vec {D}}\cdot {\vec {E}}\right)}
,
folgt:
E
p
o
t
=
1
8
π
∫
d
3
x
D
→
⋅
E
→
{\displaystyle E_{pot}={\frac {1}{8\pi }}\int d^{3}x\,{\vec {D}}\cdot {\vec {E}}}
.
Eine sog. Multipolentwicklung der Energie einer Ladungsverteilung
im äußeren Feld
E
→
{\displaystyle {\vec {E}}}
(d.h.
∇
→
⋅
E
→
=
0
{\displaystyle {\vec {\nabla }}\cdot {\vec {E}}=0}
)
können wir folgendermaßen vornehmen: In
E
p
o
t
=
∫
d
3
x
ρ
(
x
→
)
A
0
(
x
→
)
{\displaystyle E_{pot}=\int d^{3}x\,\rho \left({\vec {x}}\right)A_{0}\left({\vec {x}}\right)}
entwickeln wir das Skalarpotential nach Taylor um den Ursprung:
A
0
(
x
→
)
≈
x
→
≈
0
A
0
(
0
)
+
x
→
⋅
(
∇
→
A
0
)
(
0
)
⏟
=
−
E
→
(
0
)
+
1
2
∑
i
,
j
x
i
x
j
(
∂
2
∂
x
i
∂
x
j
A
0
)
(
0
)
⏟
=
−
∂
E
j
∂
x
i
(
0
)
+
.
.
.
{\displaystyle A_{0}\left({\vec {x}}\right){\underset {{\vec {x}}\approx 0}{\approx }}A_{0}\left(0\right)+{\vec {x}}\cdot \underbrace {\left({\vec {\nabla }}A_{0}\right)\left(0\right)} _{=-{\vec {E}}\left(0\right)}+{\frac {1}{2}}{\underset {i,j}{\sum }}x_{i}x_{j}\underbrace {\left({\frac {\partial ^{2}}{\partial x_{i}\partial x_{j}}}A_{0}\right)\left(0\right)} _{=-{\frac {\partial E_{j}}{\partial x_{i}}}\left(0\right)}+...}
,
woraus
E
p
o
t
=
q
A
0
(
0
)
−
p
→
⋅
E
→
(
0
)
−
1
2
⋅
1
3
∑
i
,
j
∫
d
3
x
ρ
(
x
→
)
(
3
x
i
x
j
∂
E
j
∂
x
i
(
0
)
−
r
2
(
∇
→
⋅
E
→
)
⏟
=
δ
i
j
∂
E
j
∂
x
i
⏞
=
0
(
0
)
)
{\displaystyle E_{pot}=qA_{0}\left(0\right)-{\vec {p}}\cdot {\vec {E}}\left(0\right)-{\frac {1}{2}}\cdot {\frac {1}{3}}{\underset {i,j}{\sum }}\int d^{3}x\,\rho \left({\vec {x}}\right)\left(3x_{i}x_{j}{\frac {\partial E_{j}}{\partial x_{i}}}\left(0\right)-r^{2}\overbrace {\underbrace {\left({\vec {\nabla }}\cdot {\vec {E}}\right)} _{=\delta _{ij}{\frac {\partial E_{j}}{\partial x_{i}}}}} ^{=0}\left(0\right)\right)}
mit
q
=
∫
d
3
x
ρ
(
x
→
)
{\displaystyle q=\int d^{3}x\,\rho \left({\vec {x}}\right)}
und
p
→
=
∫
d
3
x
ρ
(
x
→
)
x
→
{\displaystyle {\vec {p}}=\int d^{3}x\,\rho \left({\vec {x}}\right){\vec {x}}}
resultiert. Diese Entwicklung besteht also aus einem Monopol-Term
mit der Ladung q , einem Dipolterm mit dem Dipolmoment
p
→
{\displaystyle {\vec {p}}}
sowie einem sog. Quadrupolterm mit dem Quadrupolmoment
Q
i
j
=
∫
d
3
x
ρ
(
x
→
)
(
3
x
i
x
j
−
r
2
δ
i
j
)
{\displaystyle Q_{ij}=\int d^{3}x\,\rho \left({\vec {x}}\right)\left(3x_{i}x_{j}-r^{2}\delta _{ij}\right)}
:
E
p
o
t
=
q
A
0
(
0
)
−
p
→
⋅
E
→
(
0
)
−
1
6
∑
i
,
j
Q
i
j
∂
E
j
∂
x
i
{\displaystyle E_{pot}=qA_{0}\left(0\right)-{\vec {p}}\cdot {\vec {E}}\left(0\right)-{\frac {1}{6}}{\underset {i,j}{\sum }}Q_{ij}{\frac {\partial E_{j}}{\partial x_{i}}}}
.
Über die durch Joule'sche Wärme abgegebene Leistung, d.h.
−
d
d
t
E
p
o
t
=
∫
d
3
x
J
→
⋅
E
→
=
I
∫
d
x
→
⋅
E
→
=
I
U
{\displaystyle -{\frac {d}{dt}}E_{pot}=\int d^{3}x\,{\vec {J}}\cdot {\vec {E}}=I\int d{\vec {x}}\cdot {\vec {E}}=IU}
,
und die Induktionsspannung
U
=
∫
C
=
∂
A
d
x
→
⋅
E
→
=
−
1
c
d
d
t
∫
A
d
2
a
n
→
⋅
B
→
{\displaystyle U={\underset {C=\partial A}{\int }}d{\vec {x}}\cdot {\vec {E}}=-{\frac {1}{c}}{\frac {d}{dt}}{\underset {A}{\int }}d^{2}a\,{\vec {n}}\cdot {\vec {B}}}
können wir auf eine Änderung der potentiellen Energie schließen:
δ
E
p
o
t
=
1
c
I
∫
A
d
2
a
n
→
⋅
δ
B
→
⏟
=
∇
→
×
δ
A
→
=
1
c
I
∫
A
d
2
a
n
→
⋅
(
∇
→
×
δ
A
→
)
=
S
t
o
k
e
s
1
c
I
∫
C
=
∂
A
d
x
→
⋅
δ
A
→
{\displaystyle \delta E_{pot}={\frac {1}{c}}I{\underset {A}{\int }}d^{2}a\,{\vec {n}}\cdot \underbrace {\delta {\vec {B}}} _{={\vec {\nabla }}\times \delta {\vec {A}}}={\frac {1}{c}}I{\underset {A}{\int }}d^{2}a\,{\vec {n}}\cdot \left({\vec {\nabla }}\times \delta {\vec {A}}\right){\underset {Stokes}{=}}{\frac {1}{c}}I{\underset {C=\partial A}{\int }}d{\vec {x}}\cdot \delta {\vec {A}}}
=
1
c
I
∫
V
d
3
x
J
→
⏟
=
c
4
π
∇
→
×
H
→
⋅
δ
A
→
=
1
4
π
∫
V
d
3
x
(
∇
→
×
H
→
)
⋅
δ
A
→
{\displaystyle ={\frac {1}{c}}I{\underset {V}{\int }}d^{3}x\,\underbrace {\vec {J}} _{={\frac {c}{4\pi }}{\vec {\nabla }}\times {\vec {H}}}\cdot \delta {\vec {A}}={\frac {1}{4\pi }}{\underset {V}{\int }}d^{3}x\,\left({\vec {\nabla }}\times {\vec {H}}\right)\cdot \delta {\vec {A}}}
,
wobei wir hierin von
∇
→
⋅
(
H
→
×
δ
A
→
)
=
δ
A
→
⋅
(
∇
→
×
H
→
)
−
H
→
⋅
(
∇
→
×
δ
A
→
)
⏟
=
δ
B
→
{\displaystyle {\vec {\nabla }}\cdot \left({\vec {H}}\times \delta {\vec {A}}\right)=\delta {\vec {A}}\cdot \left({\vec {\nabla }}\times {\vec {H}}\right)-{\vec {H}}\cdot \underbrace {\left({\vec {\nabla }}\times \delta {\vec {A}}\right)} _{=\delta {\vec {B}}}}
Gebrauch machen, um
δ
E
p
o
t
=
1
4
π
∫
V
d
3
x
∇
→
⋅
(
H
→
×
δ
A
→
)
+
1
4
π
∫
V
d
3
x
H
→
⋅
δ
B
→
{\displaystyle \delta E_{pot}={\frac {1}{4\pi }}{\underset {V}{\int }}d^{3}x\,{\vec {\nabla }}\cdot \left({\vec {H}}\times \delta {\vec {A}}\right)+{\frac {1}{4\pi }}{\underset {V}{\int }}d^{3}x\,{\vec {H}}\cdot \delta {\vec {B}}}
zu erhalten. Mit dem Satz von Gauß gilt:
∫
V
=
R
3
d
3
x
∇
→
⋅
(
H
→
×
δ
A
→
)
=
∫
∂
R
3
d
2
a
n
→
⋅
(
H
→
×
δ
A
→
)
⟶
0
{\displaystyle {\underset {V=\mathbb {R} ^{3}}{\int }}d^{3}x\,{\vec {\nabla }}\cdot \left({\vec {H}}\times \delta {\vec {A}}\right)={\underset {\partial \mathbb {R} ^{3}}{\int }}d^{2}a\,{\vec {n}}\cdot \left({\vec {H}}\times \delta {\vec {A}}\right)\longrightarrow 0}
,
sodass sich
δ
E
p
o
t
=
1
4
π
∫
V
d
3
x
H
→
⋅
δ
B
→
{\displaystyle \delta E_{pot}={\frac {1}{4\pi }}{\underset {V}{\int }}d^{3}x\,{\vec {H}}\cdot \delta {\vec {B}}}
ergibt.
Bei einem linearen Zusammenhang zwischen
B
→
{\displaystyle {\vec {B}}}
und
H
→
{\displaystyle {\vec {H}}}
, d.h.
H
→
⋅
δ
B
→
=
1
2
δ
(
H
→
⋅
B
→
)
{\displaystyle {\vec {H}}\cdot \delta {\vec {B}}={\frac {1}{2}}\delta \left({\vec {H}}\cdot {\vec {B}}\right)}
,
gilt:
E
p
o
t
=
1
8
π
∫
V
d
3
x
H
→
⋅
B
→
{\displaystyle E_{pot}={\frac {1}{8\pi }}{\underset {V}{\int }}d^{3}x\,{\vec {H}}\cdot {\vec {B}}}
.