10—Partial Differential Equations 302
The separated solutions are then
f(x)g(z) =enπix/L
(
Aenπz/L+Be−nπz/L
)
(35)
The solution forz > 0 is therefore the sum
V(x,z) =
∑∞
n=−∞
enπix/L
(
Anenπz/L+Bne−nπz/L
)
(36)
The coefficientsAnandBnare to be determined by Fourier techniques. First however, look at thez-behavior.
As you move away from the plane toward positivez, the potential should not increase without bound. Terms
such aseπz/Lhowever increase withz. This means that the coefficients of the terms that increase exponentially
inzcannot be there.
An= 0forn > 0 , and Bn= 0forn < 0
V(x,z) =A 0 +B 0 +
∑∞
n=1
enπix/LBne−nπz/L+
∑−^1
n=−∞
enπix/LAnenπz/L (37)
The combined constantA 0 +B 0 is really one constant; you can call itC 0 if you like. Now use the usual Fourier
techniques given that you know the potential atz= 0.
V(x,0) =C 0 +
∑∞
n=1
Bnenπix/L+
∑−^1
n=−∞
Anenπix/L
The scalar product ofemπix/Lwith this equation is
〈
emπix/L,V(x,0)
〉
=
2 LC 0 (m= 0)
2 LBm (m > 0 )
2 LAm (m < 0 )