16.5 Diocotron modes 475the wall boundary conditionψ(a) =0. Since the solutions of Eq.(16.82) areψ∼r±l, the
inner solution must be
ψ=α(r
a)l
for 0 ≤r<s (16.83)and the outer solution must be
ψ=β((
r
a)l
−(r
a)−l)
fors<r≤a (16.84)where the coefficientsαandβare to be determined.
Integration of Eq.(16.80) across the delta function fromr=s−tor=s+gives the
jump condition [
d
dr
ψ(r,s)]s+s−=− 1 (16.85)
and integrating a second time shows thatψmust be continuous atr=s.These jump and
continuity conditions give two coupled equations inαandβ,
βl
a((
s
a)l− 1
+(s
a)−l− 1 )
−
αl
a(s
a)l− 1
= − 1 (16.86)β((
s
a)l
−(s
a)−l)
−α(s
a)l
= 0. (16.87)Solving forαandβgives the Green’s function,
ψ(r,s)=
−
a
2 l(s
a)l+1((r
a)l
−(r
a)−l)
forr>s−a
2 l(r
a)l((s
a)l+1
−(s
a)−l+1)
forr<s;(16.88)
this satisfies Eq.(16.80) and also the boundary conditions atr= 0andr=a.Using the
Green’s function in Eq.(16.81) gives
̃φl(r) = q
ε 0[∫r0dsψ(r,s) ̃nl(s)+∫ardsψ(r,s) ̃nl(s)]
= −
a
2 lq
ε 0
∫r
0 ds((
s
a)l+1((r
a)l
−(r
a)−l))
̃nl(s)+∫a
rds((
r
a)l((s
a)l+1
−(s
a)−l+1))
n ̃l(s)
(16.89).
Finally, using Eq.(16.77) to substitute for ̃nl(s)gives
φ ̃l(r)= q
2 ε 0 B
∫r0dssl
al(
rl
al−
al
rl) ̃
φl(s)
ω−lω 0 (s)dn 0
ds
+∫ards
rl
al(
sl
al−
al
sl) ̃
φl(s)
ω−lω 0 (s)dn 0
ds
. (16.90)
This integral equation forφ ̃l(r)not only prescribes the mode dynamics, but also explic-
itly incorporates the spatial boundary conditions. The resonant denominatorsω−lω 0 (s)in