
a timelike conformal boundary at r= 0. On top of
Mwe consider a real scalar field Ψ : AdS2×S2→R
whose dynamics is ruled by the Klein-Gordon equa-
tion
PΨ := (−m2
0)Ψ = 0,(2)
where is the D’Alembert wave operator built out
of Equation (1), while m0is the mass parameter.
In Section 2, we recollect the main ingredients nec-
essary to construct the two-point functions for ground
and thermal states for Ψ. The procedure follows the
prescription outlined in [1] and it generalizes the re-
sults obtained in [15] by considering two different fam-
ilies of generalized (γ, v)-boundary conditions at the
conformal boundary, dubbed (γ, vκ) with κ= 1,2.
The analysis for κ= 2 coincides with that of [15],
while for κ= 1 it yields novel two-point functions.
Nonetheless, since also in this case the overall analy-
sis is structurally identical to that for κ= 2, we do
not dwell into many details and we limit ourselves to
sketching the main steps of the construction.
Subsequently, in Section 3 we obtain an explicit
expression for the transition rate of a static Unruh-
DeWitt detector interacting either with the ground
or with a thermal state of a Klein-Gordon field. We
scrutinize the response of the detector to grasp if and
how it is affected by the choice of boundary condi-
tion. In particular, we focus on the consequences of
the freedom in selecting a secondary solution in the
construction of Section 2. In Section 4, we summarize
the main results of this work.
2. Dynamics and Boundary Conditions
Working with the coordinates of Equation (1), a
real Klein-Gordon field Ψ that solves Equation (2)
can be decomposed as
Ψ(t, r, θ, ϕ) = e−iωtψ(r)Ym
`(θ, ϕ),
where Y`(θ, ϕ) are the spherical harmonics on the 2-
dimensional unit sphere. Setting λ=−`(`+ 1), the
function ψ(r) solves the radial equation
Lψ(r) :=d2ψ(r)
dr2+ω2+λ−m2
0
r2ψ(r) = 0.(3)
2.1. The radial equation
In the following we assume that m0≥0 and we in-
troduce the auxiliary quantities
ν:= 1
2q1−4λ+ 4m2
0and p:= √ω2,(4)
subject to the constraint ν > 0,∀`≥0. Equation
(3) is a Sturm-Liouville problem with eigenvalue ω2,
which admits a basis of solutions written in terms of
Bessel functions of first and second kind:
y1(r) = √rJν(pr),(5a)
y2(r) = √rYν(pr).(5b)
Following [16], it is convenient to work on the
Hilbert space L2((0,∞), dr) and allow pto take a
priori any complex value. Hence, according to Weyl’s
endpoint classification r→ ∞ is a limit point, see also
[15] for a succinct summary of the nomenclature. As a
matter of fact, assuming that Im p6= 0, the most gen-
eral solution of Equation (3) lying in L2((c, ∞), dr),
∀c∈(0,∞) can be written in terms of Hankel func-
tions of first and second kind:
ψ∞(r) = √r[H(1)
ν(pr)Θ(Im p) + H(2)
ν(pr)Θ(−Im p)] .
On the other hand, it turns out that r= 0 is a
limit circle point if ν∈(0,1), since both y1, y2∈
L2((0, c0), dr), ∀c0∈(0,∞), as one can infer from the
following asymptotic expansion close to the origin:
|y1(r)|2r→0
∼r1+2ν,
|y2(r)|2r→0
∼r1−2ν.
Observe that
ν∈[0,1) ⇐⇒ `= 0 and m2
0∈−1
4,3
4.(6)
Still following the theory of Sturm-Liouville, this en-
tails that only the `= 0 mode calls for a boundary
condition at r= 0, provided that the mass lies in the
range individuated in Equation (6). When ν≥1, no
boundary condition needs to be imposed at both ends
and since this case is not of interest in this work, we
shall not consider it further.
2.2. Generalized (γ, v)-boundary conditions
At the limit circle endpoint, r= 0, we impose
generalized (γ, v)-boundary conditions as introduced
in [1]. These identify self-adjoint extensions of the
radial operator Lin L2((0,∞), dr) and they are fully
characterized in terms of a parameter γ∈R, of the
principal solution, u, and of a secondary solution, v,
which we introduce in the following. The principal
solution at r= 0 reads
u:= √rJν(pr) = y1(r).
This is unambiguously identified by the condition
that, for any solution vof Equation (3) such that
v6=λu,λ∈C, then lim
r→0
u
v= 0. On the contrary
choosing a secondary solution at r= 0, linearly in-
dependent from u, is fully arbitrary. In the following
we consider two possible choices vκ,κ= 1,2:
v1:= p√rYν(pr) (7a)
v2:= −p2ν√rJ−ν(pr).(7b)
2