1 Introduction
Dielectric mediums have been extensively studied in literature, both theoretically [1
–
5] and experimentally
[6
–
9]. In the past years, analogue models of gravity [10
–
20] considering a nonlinear dielectric medium
received attention in the scientific community [21
–
39]. These models predicted that nonlinear materials
can produce, in laboratory, an effective (optical) horizon with similar properties to the event horizon of
black holes [40]. At these horizons, light is trapped, such that a ring could be observed due to birefringence
phenomenon [41
–
44]. Mathematically, this is described by an effective geometry [21,24,33,34,36,38]
generated by the medium, such that light follows geodesics in this effective geometry, and it determines
the position of the optical horizons. In this paper, we are interested in applying this method to curved
spacetimes. For that, instead of considering the flat spacetime of a laboratory as the background geometry,
we are going to consider a curved background generated by a spherically symmetric object. This choice is
motivated by studies concerning a system formed by black holes surrounded by a plasma [45,46]. If this
medium has nonlinear properties, we can describe its effects on the propagation of light by means of an
effective geometry, in a similar manner to analogue models.
In order to apply this method to a system formed by a compact object surrounded by a nonlinear
electromagnetic material, we are going to consider a dielectric medium
µ
=
const.
,
=
(
E
) with negligible
mass, compared to the compact object, such that we can idealize a fixed background. In this paper, we
show that this simple model is enough to predict some important consequences in the propagation of light,
that includes corrections on the geometrical redshift and on the light deflection by a compact body. If the
medium surround a charged body, we can have two polarization modes for light, such that this model
gives a mechanism to evaluate the consequences of birefringence effects on curved spacetimes. Each mode
is associated to a different correction on these quantities, such that we can have two possible paths for the
light ray. These paths only coincide if the electrical field is turned off or if the permittivity is constant.
The vacuum appears as a particular case in this configuration.
In this paper, what we call event horizon depends only on the background metric, and what we call
killing horizon depends only on the effective metric, such that we are going to refer to it as an optical
horizon. Formally, only in the vacuum we could use the background metric to evaluate the optical horizon,
because in that case the effective metric and the background metric are equivalent. This is a consequence
of the definition of a killing horizon [47,48], which is described as the radius where the redshift diverges
and the time killing vector changes its sign, implying that we need to consider the effective metric to
determine it. We proved that the position of the optical horizons coincide with the position of the killing
horizons predicted by the background metric. This is a direct consequence of the form of the effective
metric.
The paper is organized as follows: In section 2we review Hadamard’s method of discontinuities and
construct the effective geometry for a nonlinear dielectric medium, in order to discuss birefringence effects
on flat spacetimes. In section 3, we apply this method to a spherically symmetric curved background
and show its consequences on the propagation of light. In section 4, we consider three particular cases
for our analyzes. The first, is a special case that doesn’t produce alterations on either the geometrical
redshift or the gravitational deflection of light. However, it changes the effective potential of light, which
coincides with the vacuum predictions only if
=
0
. The second, is a less restrict case that produce
alterations on the effective potential and on the deflection angle of light, but maintain the predictions for
the geometrical redshift of the positive polarization. The last, is a general case that account all possible
consequences on curved spacetimes. Both can be reduced to the first case in the particular situation where
the permittivity is constant. Finally, we end our analyzes showing that the effective metric associated to
the negative polarization can be written in a conformally flat form, for a given dependence of the dielectric
3