Hawking radiation in optics
and beyond
rsta.royalsocietypublishing.org
Research
Article submitted to journal
Subject Areas:
vacuum physics, nonlinear optics
Keywords:
analogue gravity, Hawking radiation,
optical fibres
Author for correspondence:
David Bermudez
e-mail: dbermudez@fis.cinvestav.mx
Raul Aguero-Santacruz and
David Bermudez
Departamento de Física. Cinvestav, A.P. 14-740,
07000 Ciudad de México, Mexico
Hawking radiation was originally proposed in
astrophysics, but it has been generalized and extended
to other physical systems receiving the name of
analogue Hawking radiation. In the last two decades
several attempts have been made to measure it in a
laboratory, one of the most successful systems is in
optics. Light interacting in a dielectric material suffers
an analogue Hawking effect, in fact, its stimulated
version has already been detected and the search for
the spontaneous signal is currently ongoing. We do a
brief review of the derivation of Hawking radiation in
general, then we mainly focus on the optical analogue
and present some novel numerical results.
1. Introduction
The fascinating phenomenon of Hawking radiation was
proposed over 45 years ago by Stephen W. Hawking [1].
Using a semi-classical approximation, i.e., considering
quantum fields around a classical geometry, Hawking
realized that the event horizon of a black hole in the
Schwarzschild metric [2] should emit massless particles.
Black holes are not entirely black.
This is a significant phenomenon because it gives
black holes, known as the ultimate devourers, something
unthinkable, i.e., a mechanism to lose mass. This is
known as black hole evaporation and could dramatically
change the cosmological evolution and the final state of
the universe.
However, astrophysical Hawking radiation turns out
to be too small to be directly detected by any conceivable
experiment, satellite, or telescope up to now. This is
partly due to the fact that gravity is the weakest
known force [3]. For example, in the Millikan oil drop
experiment, a single electron charge can counteract the
gravitational attraction of the whole planet earth. For
this reason, a quantum effect due to gravity would be
extremely hard to measure.
c The Authors. Published by the Royal Society under the terms of the
Creative Commons Attribution License http://creativecommons.org/licenses/
by/4.0/, which permits unrestricted use, provided the original author and
source are credited.
 However, this situation did not stop the theoretical studies of the effect. By 1981, while William 2
Unruh [4] was teaching a course on fluid dynamics [5], he realized that the equations for the
propagation of sonic waves in a moving fluid could be expressed in a similar fashion as those for

rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 0000000

..................................................................
light waves around the black hole geometry, as studied by Hawking. This work slowly inspired
new ideas on how to study phenomena usually related to gravity in other physical systems, giving
birth to a new research area known now as analogue gravity.
Inversely, general relativity inspired a connection between spacetime geometry and optical
media through what became known as transformation optics [6,7], using the tools developed by
Walter Gordon [8] and Jerzy Plebański [9] to study electromagnetic fields in curved spacetimes.
Furthermore, transformation optics has been used to obtain another system for the search of
analogue Hawking radiation, light pulses interacting inside optical fibres [10–13].
It is important to note that analogue gravity does not reproduce the dynamics of general
relativity given by Einstein equations, but rather, the kinematic of its solutions given by spacetime
metrics that describe the curvature of spacetime. In this way, it can reproduce only the kinematic
phenomena due to gravity and not the dynamical ones. Luckily, there are several kinematic effects
to analyse, such as Hawking radiation [14].
In order to study the regime of “semi-classical” gravity, the field propagating on top of
the curved background should be quantized [15]. Several systems have been proposed for this
purpose. Unruh suggested liquid Helium for the first quantum analogue in 1981 [4]—see also
work by Grigory Volovik [16]. Since then, several systems were proposed, including phonons
in Bose-Einstein condensates [17,18], polaritons [19], and photons in optical fibres [10,13,63].
Classical systems such as water have also been explored [20,21].
In this work we will present briefly the general theory of analogue gravity, for more complete
reviews see Refs. [11,22–26]. Then, we will focus on the optical analogue and in particular on the
theoretical and numerical aspects. For a review on the experimental side see Ref. [27]. We will
explain how the Hawking effect works in the optical analogue and what are the particularities
of this system, such as the strong dispersive regime and the completely non-thermal Hawking
spectrum [28–30,62].
Research on gravity-like effects in analogue systems can give new clues into how to quantise
gravity in some emergent gravity models by analysing the corresponding underlying theory
in the analogue system. For example, by investigating the role of high-frequency modes [31],
quantum vacuum, or dispersion in the derivation of Hawking radiation. This can help us
understand if astrophysical black holes radiate [32]. Furthermore, there are several interesting
phenomena to study in analogue gravity besides Hawking radiation, such as cosmological
creation of particles [33], dark energy [34], and superradiance [35].
In Section 2, we review important aspects of astrophysical Hawking radiation and describe
the kinematics of horizons. Section 3 is a summary of the theory of analogue systems describing
modes around the horizon. In Section 4, we translate the developed tools to the optical
analogue case. We describe briefly the dynamics of the stimulated optical analogue with classical
electromagnetism in Section 5. At last, we present our conclusions in Section 6.
Finally, given some resistance from a part of the scientific community, it is important to argue in
favour of using the name Hawking radiation to describe those phenomena in analogue systems.
We will show here why the name is appropriate and call for a generalization of the term Hawking
radiation to include the same effect in systems different from astrophysics.

2. Astrophysical Hawking radiation

In the first half of the 1970s, Jacob Bekenstein’s ideas about the black holes thermodynamics [36]
were spreading and began to gain even more relevance once Hawking deduced his famous
relation1 between the mass and the temperature of a radiating black hole [1]: 3
~c3
T= . (2.1)

rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 0000000

..................................................................
8πGM kB
This deduction works for black holes formed by gravitational collapse and whose event horizon
is therefore an apparent horizon in view of its finite creation time. If the mass of the star is larger
than both the Chandrasekar and the Tolman-Oppenheimer-Volkoff (TOV) limits, the dying star
cannot counterbalance its own weight and it collapses into a black hole.
Some of Bekenstein’s equations relate the entropy with the surface area of the black hole
and have a resemblance to the usual laws of thermodynamics. His deductions came from
the dynamics of general relativity, given by Einstein equations, whereas the Hawking formula
comes from the Schwarzschild metric and it only needs the kinematics of general relativity [14].
Although the connection seems to be strong between both works, a true link is yet to be proven.
On its own, Hawking deduction has some issues that have not been solved yet. Being a semi-
classical effect, quantum gravity effects are neglected and the nature of pair production is not
entirely clear. For an observer far away from the horizon, the measured radiation is going to
be red-shifted, meaning that wave packets are blue-shifted without limit the closer they are to
the horizon, quickly getting even below the Planck length. This is known as the trans-Planckian
problem [32].
For these reasons, it is valuable to learn about Hawking radiation in analogue systems, where
we can understand different aspects of this phenomenon. For example, it has been proven that a
cut-off can be introduced in high-frequency modes, and the Hawking effect still persists [37].
Although these modifications usually lead to grey factors that change the thermality of the
spectrum.
Another important factor is the possibility offered by analogue experiments to measure the
effect. If astrophysical black holes were to radiate, their temperature would be of the order of
∼ 10−8 K. Considering that the cosmic microwave background (CMB) temperature is ∼ 2.7 K and
its fluctuations of ∼ 10−4 K, there is a lack of current technological capacity to perform a direct
measurement. Thus making the analogues the most viable alternative in search of experimental
evidence for the Hawking effect.
The basic assumptions to derive Hawking radiation in any system were worked out by
Matt Visser [14]. The essential feature is the existence of a metric in the form of Eq. (3.3) with
the following characteristics: (1) apparent horizon, (2) non-zero surface gravity, and (3) slow
evolution.
We will see that these points are fulfilled for the optical analogue. It will be clear that
no dynamical considerations are needed in this derivation. In this sense, there is no need for
Bekenstein theory of black hole thermodynamics.
To elaborate on how the analogues work, we can picture Schwarzschild spacetime as a moving
fluid for both of its solutions, the black and white holes. For a black hole, spacetime can be seen
as being dragged into the centre of the hole, meanwhile for the white hole, which is nothing more
than a black hole evolving backwards in time, spacetime is being expelled out from the centre.
For each case, waves travelling against the curvature are counter-propagating to the spacetime
fluid. This is depicted in the embedding diagram in Fig. 1.
One of the main ingredients of Hawking radiation is an effective geometry that is divided
by a horizon. We can picture this idea in a more familiar scenario using an analogy inspired by
Unruh [4] . Imagine a fish swimming in a river. The river has a flux speed V , considered as the
background medium with an effective geometry. The fish represents a perturbation propagating
in this background, assuming it can only swim up to a maximum speed c, then its destiny will
depend on its relation with V . If the river flow speed changes with position V (x), e.g., increasing
towards a waterfall (see Fig. 2a), or in a similar case, if the fish speed depends on position c(x),
e.g., if the water is becoming muddier and fish find it more difficult to swim there (see Fig. 2b).
Then, in a region where V < c, the fish will be able to swim in counter current in an ever increasing
1
In fact, Hawking asked for this equation to be written on his grave stone on Westminster Abbey, UK.
 Black hole White hole 4

rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 0000000

..................................................................
rS
v
rS v

Figure 1. Embedding diagram of a black hole (left) and a white hole (right). Waves travelling against the curvature (red)
can be seen as counter-propagating waves in a fluid flow (black). The event horizon is marked with rS (yellow).

Figure 2. River flow V and fish with speed c. In (a) the river flow depends on the position V = V (x) as its slope is
changing towards a waterfall, this is the situation in condensed-matter analogues. In (b) the fish speed changes with
position c = c(x) as water becomes muddier, this is the situation in optical analogues.

speed up to a maximum of c, this corresponds to the right-hand side in both cases of Fig. 2. In
the opposite case, where V > c, the fish is dragged by the current, corresponding to the left-hand
side of Fig. 2. The horizon then becomes the point where V = c, the only place where the fish will
stay put, marked with the sign “point of no return”. We can summarize the important aspects as
follows: (1) fluid flow V , (2) fish speed c, and (3) change of fluid flow V = V (x) or of fish speed
c = c(x). Experiments in condensed-matter analogues are based in a change of fluid flow V (x),
while optical analogues are based on varying the wave speed c(x).

3. General theory of analogue gravity

We tend to associate the idea of a metric with coordinates in astrophysics, however, it can also be
used to describe effective geometries in a wide variety of physical scenarios, such as in fluid flux
or dielectric media configurations. Hawking radiation was first discovered in astrophysics using
the Schwarzschild metric that in Painlevé-Lemaître-Gullstrand (PLG) coordinates is
 r 2
rS
ds2 = c2 dt2 − dr + c dt + r2 dΩ 2 , (3.1)
r
where rS is the Schwarzschild radius. We can generalize this metric to other velocity profiles c(r, t) 5
and V (r, t) as
ds2 = −c(r, t)2 dt2 + [dr − V (r, t)dt]2 + r2 dΩ 2 , (3.2)

rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 0000000

..................................................................
or equivalently as

ds2 = −[c(r, t)2 − V (r, t)2 ]dt2 − 2V (r, t)drdt + dr2 + r2 dΩ 2 . (3.3)

where in general V (r, t) represents the velocity of a propagating medium and c(r, t) the velocity of
a perturbation propagating through such medium. In most systems the analysis is performed by
considering only one spatial dimension and therefore ignoring the In the matrix representation,
the PLG metric takes the form
!
−(c2 − V 2 ) −V r̂j
gµν (r, t) = . (3.4)
−V r̂i δij

Wep can return to the Schwarzschild solution setting c(r, t) = c the speed of light and V (r, t) =
−c r/rS the velocity profile of an observer in a radial trajectory. Due to radial symmetry, the
horizon is found at the spatial point where the condition c(r, t) = |V (r, t)| is fulfilled. This could
also be an extended region or several points and it is not a truly singular point since det(gµν ) =
−c(r, t)2 . It is convenient to define the quantity gH (t) evaluated at the horizon as

1 d[c(r, t)2 − V (r, t)2 ] 

 
d[c(r, t) − |V (r, t)|] 
gH (t) = = cH  . (3.5)
2 dr 
H dr H

For the static geometry case, the definition of surface gravity α can be recovered [14,38]:
gH
α= . (3.6)
cH
The procedure to derive the Hawking temperature is the same as in astrophysical case, i.e.,
by considering a free massless scalar quantum field φ(r, t) over a classical spacetime background
with an associated curvature. This is why Hawking radiation is considered as a semi-classical
effect. Taking this field over the PLG spacetime in 1 + 1 dimensions, i.e., ignoring the angular
part dΩ = 0 and r → x, the field Lagrangian can be expressed as [39]:
1√
L= −gg µν ∂µ φ† ∂ν φ, (3.7)
2
When the analysis is done in 3+1 dimensions we arrive at a similar result: Hawking radiation
is still emitted but its spectrum is not thermal, since it is now modified by a grey body factor
[39]. If we substitute the metric (3.3) and use the variational method or plug in the result in the
Euler-Lagrange equation, we arrive to the wave equation

(∂t + ∂x V )(∂t + V ∂x )φ − c2 ∂x2 φ = 0. (3.8)

At this point we could call upon the eikonal approximation for the s wave
  Zx 
φ(x, t) = A(x, t) exp ∓i ωt − k(y)dy , (3.9)

where A is a slowly-varying envelope and the argument of the exponential is a rapidly-varying

phase. It can be proven that this approximation is enough to derive Hawking radiation [14].
Nonetheless, we can obtain a general solution to the wave equation (3.8) if we define two new
coordinates u and v as
Zx Zx
dy dy
u=t− , v=t+ , (3.10)
c + V (y) c − V (y)
that solve the wave equation ∂u ∂v φ = 0. Outside the region where c = |V |, the solution of Eq. (3.7)
can be expressed as the sum of two arbitrary functions, one as an only u function and the other as
an only v function: 6
 Zx   Zx 
u dy v dy
φ=φ t− +φ t+ . (3.11)

rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 0000000

c + V (y) c − V (y)

..................................................................
Concerning a medium moving to the left, mode φu moves to the right and is known as the
counter-propagating mode, since it propagates against the fluid, while mode φv is the co-
propagating mode, as it propagates along with the fluid. Considering that Lagrangian (3.7) is
invariant to temporal translations, there exists solutions of the form

φ(x, t) = e−iωt φω (x), (3.12)

which are known as stationary modes. Modes are then separated into co-propagating
and counter-propagating parts in the form φ = Cu φu v
ω (x) + Cv φω (x). Since the case under
consideration does not include dispersive effects, it is expected that such functional forms
preserve its structure, where all wave components have speed c measured from the moving
Rx u,v
medium. A local wavenumber k(x) is defined such that φu,v


ω (x) = exp i kω (y)dy , thus if
we use Eqs. (3.11) and (3.12) we get
u ω v ω
kω (x) = , kω (x) = − . (3.13)
c + V (x) c − V (x)
The same result can be obtained through the eikonal approximation (3.9). In general terms, the
previous equations constitute the dispersion relation

[ω − kV (x)]2 = c2 k2 . (3.14)
0
We identify the Doppler-shifted frequency to the co-moving frame ω = ω − kV (x) which is of
particular importance for the theory of optical analogues.
Assuming the existence of a horizon, i.e., a region where V = −c, counter-propagating modes
in the moving medium will have a zero velocity in there. For simplicity, suppose that the horizon
is located at x = 0 and that the first derivative of V (x) is not zero there. Since the u coordinate
becomes singular at this point, a linear approximation of the velocity profile is usually performed
as 
dV (x) 
V (x) ≈ −c + αx, α= . (3.15)
dx  x=0
If α > 0, the speed increases in the same direction as the moving medium, which is equivalent to
the behaviour of a black hole. On the other hand, when α < 0, the speed is reduced in the direction
of the flux and the behaviour is that of a white hole.
Returning to the behaviour of the u coordinate near the horizon, under the approximation
(3.15) we can obtain
1 α 
u ≈ t − ln |x| . (3.16)
α c
This functional form of u establishes the presence of a horizon at x = 0, and a separation of space
into two regions delimited by x = 0. Since there must exist a full set of modes supporting the
existence of waves in all space, there must be two different counter-propagating modes for all
frequencies in each region of space separated by the horizon, this is
1 1
φu,L
ω (x, t) = θ(−x) √ eiωu , φu,R
ω (x, t) = θ(x) √ e−iωu , (3.17)
4πcω 4πcω
where the change of sign in the exponent for the left L and right R regions is included due to the
change of sign in the quantity c + V , assuring a positive norm for the modes.
Since we are in the dispersionless case, the speed of light is effectively a constant c(r, t) = c,
so the value of α given in Eq. (3.15) reduces to the relation (3.6). Equations (3.17) correspond to
a single outgoing wave, now identified as out-modes. We can expect as well a set of modes that
correspond to a single ingoing wave, taking into account that a proper definition of these in-
modes requires a finite time for the formation of the horizon (sometimes referred to as apparent
horizon), such that we can construct these modes when the horizon has not yet been formed. The
in- and out-modes form two different sets of orthonormal u-modes that can be related through a 7
Bogoliubov transformation:

φout,R αj (ω)φin,L βj (ω)φin,R

rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 0000000

X X

..................................................................
ω = ω,j + ω,j , (3.18)
j j

where the coefficients α and β fulfil the norm-conservation identity

[|αj (ω)|2 − |βj (ω)|2 ] = 1.

X
(3.19)
j

The βj (ω) term in the Bogoliubov transformation (3.18) provides a mixture of positive and
negative frequency modes as a result of a dispersion process generated by the existence of a
horizon. As Hawking originally did, we can consider the vacuum state, defined as the absence of
particles, which mathematically reads as

â|0i = 0, ∀ â, (3.20)

where â is the annihilation operator and |0i the vacuum state. This means that the operator â
annihilates all modes in the basis. Different vacuum states arise when at least one coefficient β in
Eq. (3.18) is different from zero. Then, in a scattering process the in-modes would be different to
the out-modes. Existence of negative modes in Eq. (3.18) is essential for the creation of particles
from vacuum, making norm conservation more akin to electric charge conservation (with positive
and negative charged particles) than to mass conservation in classical mechanics (with only
positive massive particles).
It is known from quantum field theory in curved spacetimes that the definition of vacuum
can change depending on the set of orthonormal modes taken as basis [15]. We can calculate the
expectation value of the number of outgoing particles in the in-vacuum and find that it has a
bosonic thermal distribution, which is seen far from the horizon as a Planckian flux of outgoing
particles [39]:
δ(ω − ω 0 )
hnout,R
ω i=   . (3.21)
2πω
exp −1
α
In a similar way to the astrophysical Hawking case, certainly as in the Unruh effect, in this
generalized derivation the radiation is thermal and its effective temperature depends on the
acceleration of the moving medium [40], this is
~α
kB T = . (3.22)
2π
It should be appreciated that this equation is the central piece of the general theory of Hawking
radiation that can be applied to any system, astrophysical or analogue. However, let us remark
an important difference with the astrophysical case. Despite the fact that the dispersionless case
was considered in this derivation, in analogue systems the presence of some kind of dispersion is
unavoidable, which usually leads to a change in the spectrum, i.e., the Hawking spectrum may no
longer be thermal. In the astrophysical case it is not clear if a dispersion mechanism exists, but if
spacetime discretisation were true, dispersion effects could be introduced into Hawking radiation
theory as in the analogue systems [41,42].
The most efficient way of solving the Bogoliubov transformations for several in- and out-
modes relies on the scattering matrix formalism. This method has been used to derive the
Hawking spectrum for several systems and in particular for the optical case that is a highly
dispersive system [12,29,30,43,44].

4. Analogue Hawking radiation in optics

Optical analogues are of particular interest because its practical realization is already on its way
[10,13,45,63]. The first obstacle for the creation of an optical black hole lies in the necessity to have
 8

rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 0000000

..................................................................
Figure 3. Transformation optics. On the left, flat space with non-isotropic media (blue). On the right, curved space (curved
lines) in isotropic media. Also, an equivalent geodesic (red dashed) is shown in both spaces.

a medium moving at a speed close to the speed of light in the material. Initially, Ulf Leonhardt
proposed a slow-light model [46], where the group velocity of light is reduced with the aid of
induced electromagnetic transparencies. However, we will see that the modification of both phase
and group velocities is necessary to obtain the analogue, for this reason slow light could not be
implemented, as it only affects the group velocity [47]. However, it inspired a closer look into the
relation between optics and general relativity. Similarly to Unruh’s story, Leonhardt’s interest in
analogue gravity also arose from teaching a course, in his case of general relativity [48].
The behaviour of light depends on the propagating medium. For instance, non-uniform media
can bend light due to the change in the index of refraction caused, for example, by a hot air
gradient as in a mirage. Furthermore, according to Einstein equations, a highly massive body
can also bend light due to the spacetime curvature, as light propagates through geodesics. In the
first case there is a flat geometry and non-isotropic media (g ij = δ ij , ij 6= δ ij ), whereas in the
second case there is curved geometry and isotropic media (g ij 6= δ ij , ij = δ ij ), see Fig. 3 and Eq.
(4.4). The equivalence between both systems was initially derived [8,9] and eventually led to the
establishment of the field of transformation optics [6,7]. However, to pass from studying light
bending in space to light bending in spacetime we need a moving horizon.
With a change of perspective, moving an optical medium to a speed close to the speed of light
is actually possible and a phenomenon that occurs all the time in optical telecommunications,
where light pulses propagate through optical fibres. Acknowledging that the effective medium
properties are the ones that determine the appearance of an event horizon, Philbin et al. proposed
[10] to take advantage of the localized perturbation generated by the optical Kerr effect. This
effect introduces a local change in the index of refraction of the medium proportional to the pulse
intensity [49]. Since this contribution travels along with the pulse, the pulse itself constitutes a
moving medium. We call this the pump pulse. Note that no fibre material is being displaced, only
light itself. In the reference frame of the pulse, the fibre material appears to be moving in the
opposite direction. The index of refraction n can be expressed as the sum of the intrinsic index of
the medium n0 (ω) plus a small contribution due to Kerr effect δn(t):

n(ω, t) = n0 (ω) + δn(t). (4.1)

For the first part of the calculations we ignore the ω-dependence of n and include it back in Eq.
(4.16). Once we consider the dispersion relation we take it back into account. The speed of the
effective medium is equal and opposite to the group velocity V of the pump in the fibre. Another
pulse, called probe, would see a non-uniform index of refraction. Considering that in dispersive
materials the speed of light depends on the frequency, we can work with probe light of a different
 PHR 9
BH PHR WH

rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 0000000

..................................................................
NHR

V NHR V

Figure 4. Pump pulse (yellow) is fixed at the centre and the fibre (green) moves to the left with speed −V in the co-moving
frame. On the left, the probe pulse (orange) moves to the left and interacts with a black hole (BH) horizon, on the right the
probe pulse (red) moves to the right and interacts with a white hole (WH) horizon. In both cases the probe produces two
signals that correspond to the analogues of positive- (PHR) and negative- (NHR) frequency Hawking radiation.

frequency such that it moves slower [faster] than the pump, i.e., moving to the left [right] in the
co-moving frame. It is then reversed by the interaction with the pump due to the change in the
refractive index in a soft crash or blockage. This establishes a horizon at the point where the probe
is slowed down to a stop or a point of no return. Depending on the relative motion of the probe
with respect to the pump we have an artificial black [white] hole horizon, as shown in Fig. 4 left
[right].
The effective geometry is established by the dielectric properties of the medium. These are
obtained from the quantum field of light described by the vectorial potential  that generates
the electromagnetic fields. Using the constitutive equations under a Lorentz transformation, in an
inertial frame called laboratory frame, we get the relations [41]:
  
V V
D̂ = 0 γ 2 (Ê − V B̂) + B̂ − 2 Ê , (4.2)
µ c
c2
  
V
Ĥ = 0 γ 2 V (Ê − V B̂) + B̂ + Ê , (4.3)
µ µ
p
where γ = 1/ 1 − V 2 /c2 is the relativistic factor and , µ are the electric permittivity and
magnetic permeability, respectively. We use the matrix g µν to introduce the analogy of the
geometric background
!
µν 1 c2 − µ−1 V 2 ( − µ−1 )cV
g = 2 . (4.4)
c − V 2 ( − µ−1 )cV c2 − µ−1 V 2

Thus we can establish the relations

D̂ = −0 cg 0ν ∂ν Â, Ĥ = −0 c2 g 1ν ∂ν Â. (4.5)

where the Einstein sum convention has been used. Likewise, the wave equation for light in a
moving medium can be written as
∂µ g µν ∂ν Â = 0. (4.6)

In a similar way as the change of variables was previously used in Eq. (3.10), we define a change
of coordinates of the form
Z
dx
t± = t − , (4.7)
v±

where the relativistic sum of velocities in the medium is done according to the Einstein’s velocity
addition theorem
c
V ±
v± = n . (4.8)
V
1±
cn
Therefore, the wave equation can be rewritten as 10
Λ ∂ 2 Â
4 2 = 0, (4.9)

rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 0000000

..................................................................
c ∂t+ ∂t−

where Λ = n(c2 − V 2 )/(c2 − n2 V 2 ) is a conformal factor. It can be appreciated that solution (4.9)
leads to (3.11). A moving medium establishes an effective geometry given by Eq. (4.4), such that
the horizon is formed when the velocity of the medium equals the speed of light in that medium,
this is
c
V =− , (4.10)
n
where, once again the new coordinate t+ contains a pole in the horizon, because v+ is cancelled
at this point and thus we realize a first-order approximation of the form

dv+ 
v+ = αx, α= . (4.11)
dx x=0
Using Eqs. (4.8) and (4.10), it can be showed that
 
1 dV c dn 
α= − . (4.12)
1 − n−2 dx n2 dx x=0
Therefore, the corresponding Hawking temperature is given by
~α
kB T = , (4.13)
2π
i.e., the same as in Eq. (3.22). However, the previous calculations were made for a reference
frame moving with the propagating pulse. Since Hawking radiation is ultimately observed in the
laboratory frame, it is necessary to perform a Lorentz transformation on the result (4.12). Before
this step, new coordinates for the reference frame co-moving with the pump are defined as
x x
τ =t− , ζ= , (4.14)
V V
where V appears as we choose a frame moving with the pump. As a consequence, the retarded
time or delay τ has now the role of distance, while the propagation time ζ plays the role of time.
Then, α can be expressed as
1 dδn
α= . (4.15)
δn dτ
The Hawking temperature depends on the relative change of δn, thus if we have a very short
pulse duration, a spectrum with considerable temperature could be produced, making Hawking
radiation an observable phenomenon in optical fibres [10]. To include the dispersive properties
of the medium modelled by its dispersion relation, we consider a first-order approximation
that is valid around the horizon and allows us to keep the same Eq. (4.13) with a varying
α = α(τ ) by evaluating Eq. (4.15). The dispersion relation is usually expressed as a variation of
wavenumber with frequency k(ω) = n(ω)ω/c. Even more, non-linear effects in optical fibres have
to be considered. From Eq. (4.14), the phase transformation of the modes in both reference frames
can be found. In the laboratory frame we have
Z
ϕ = (kdx − ωdt), (4.16)

and under the change of coordinates, in the co-moving frame we obtain

Z
ϕ = − (ωdτ + ω 0 dζ), (4.17)

where the co-moving frequency ω 0 is defined as

 
V
ω 0 (ω) = γ (ω − V k(ω)) = γ 1 − n(ω) ω. (4.18)
c
 GH 11
!0 probe PHR GH NHR
!2

rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 0000000

!1

..................................................................
<latexit sha1_base64="csR7G8aqOn5tobrszlFoWTgeyr8=">AAAB73icbVDLSgNBEOyNrxhfUY9eBoPgKexGQY9BLx4jmAckS5idzCZD5rHOzAphyU948aCIV3/Hm3/jJNmDJhY0FFXddHdFCWfG+v63V1hb39jcKm6Xdnb39g/Kh0cto1JNaJMornQnwoZyJmnTMstpJ9EUi4jTdjS+nfntJ6oNU/LBThIaCjyULGYEWyd1ekrQIe7X+uWKX/XnQKskyEkFcjT65a/eQJFUUGkJx8Z0Az+xYYa1ZYTTaamXGppgMsZD2nVUYkFNmM3vnaIzpwxQrLQradFc/T2RYWHMRESuU2A7MsveTPzP66Y2vg4zJpPUUkkWi+KUI6vQ7Hk0YJoSyyeOYKKZuxWREdaYWBdRyYUQLL+8Slq1anBRrd1fVuo3eRxFOIFTOIcArqAOd9CAJhDg8Ayv8OY9ei/eu/exaC14+cwx/IH3+QO8WY/E</latexit>

<latexit sha1_base64="4jwydYJy+j1aJe4fyf4qCtcM7Ds=">AAAB73icbVDLSgNBEOyNrxhfUY9eBoPgKexGQY9BLx4jmAckS5idzCZD5rHOzAphyU948aCIV3/Hm3/jJNmDJhY0FFXddHdFCWfG+v63V1hb39jcKm6Xdnb39g/Kh0cto1JNaJMornQnwoZyJmnTMstpJ9EUi4jTdjS+nfntJ6oNU/LBThIaCjyULGYEWyd1ekrQIe4H/XLFr/pzoFUS5KQCORr98ldvoEgqqLSEY2O6gZ/YMMPaMsLptNRLDU0wGeMh7ToqsaAmzOb3TtGZUwYoVtqVtGiu/p7IsDBmIiLXKbAdmWVvJv7ndVMbX4cZk0lqqSSLRXHKkVVo9jwaME2J5RNHMNHM3YrICGtMrIuo5EIIll9eJa1aNbio1u4vK/WbPI4inMApnEMAV1CHO2hAEwhweIZXePMevRfv3ftYtBa8fOYY/sD7/AG61Y/D</latexit>

RRIR pump RR NRR

PH
!

Figure 5. Dispersion relation in the co-moving frame, i.e., a relation ω 0 (ω). Horizontal lines mark the conservation of ω 0
for the pump ω1 (yellow) and for the probe ω2 (red). The group- (GH) and phase- (PH) velocity horizons are also marked.

By comparing (4.16) and (4.17), we show that −ω and ω 0 play the roles of k and ω, respectively
[29,50]. The functional form of n(ω) is modelled by considering the dispersive properties of the
medium, e.g., a real dispersion for a photonic crystal fibre (PCF) has a finite region of anomalous
group-velocity dispersion (GVD). The general shape of the co-moving frequency dispersion ω 0 (ω)
for PCFs is depicted in Fig. 5. The pump with laboratory frequency ω1 (yellow) sits in the
anomalous GVD with positive curvature, and the probe with ω2 in the normal GVD region with
negative curvature.
The pump creates non-linear effects by itself, in particular, the conservation of its co-moving
frequency creates new signals as resonant radiation (RR and RRIR ), also known as Cherenkov
radiation or dispersive waves. Furthermore, in the high-frequency regime the dispersion relation
goes to zero at the phase-velocity horizon (PH) and becomes negative. We plot the negative part
of the dispersion as a dotted line to show it in the first quadrant.
We already said that negative frequencies are an essential part of the derivation of Hawking
radiation. In fact, the study of optical analogues led to the prediction of a new phase-matching
condition for resonant radiation. This was previously overlooked because it needs a mixing of
positive and negative frequencies. The new signal is called negative-frequency resonant radiation
(NRR), and has been already measured in several experiments [13,51,52]. Furthermore, it has been
explained using non-linear optics by Conforti et al. [53].
Looking at the dispersion relation in Fig. 5, the need for both the group- (GH) and phase-
(PH) velocity horizons becomes clear. The change of refractive index blocks the in-mode at the
GH and undergoes a Hawking effect—redshifting and blueshifting of the in-modes occurs at the
GH [54]. Meanwhile, the PH is needed to have negative modes that lead to a mixing of opposite-
norm modes that can create particles while still conserving the norm. The phase horizon makes
nonzero β coefficients in the Bogoliubov transformation in Eq. (3.18).
The Bogoliubov transformation that describes the mixing quantum states with positive and
negative frequencies in Eq. (3.18) can be solved efficiently using the scattering matrix. In this way
we get the Hawking spectrum, i.e., the intensity of Hawking radiation for different frequencies. As
we mentioned earlier, the Hawking radiation still exists even under dispersion, but its thermality
is lost, or more precisely, the effective temperature depends now on the frequency. With this
method [29,43], we obtained a Hawking spectrum for the PCF used in experiment [13]. See Fig. 6.

We can now discuss the essential points for the Hawking effect to be realized in optics. In this
case, the existence of a Lorentzian metric is given by Eq. (4.4). We recall the apparent horizon is
established by a pulse travelling in the medium, causing a change in the local index of refraction
by means of the Kerr effect. In this case the concept of apparent horizon depends on the frequency
of the modes that get close to it. Each frequency mode will experience its own horizon because
 200
12

rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 0000000

..................................................................
150

100

50

0
0.5 1.0 1.5 2.0

Figure 6. Spectrum from the spontaneous Hawking effect in a PCF from solving the mixing of Bogoliubov coefficients
with the scattering matrix. The GH and PH are marked, as well as the frequencies where the peaks of Hawking emission
are expected (PHR, NHR).

its group velocity will vary with frequency and each mode will be blocked at a different position
(in τ ) of the pump. This is called a “fuzzy horizon”. Besides, the value of the surface gravity
α is given by (4.15), it also depends on the frequency, but it is non-zero for all the frequencies
that experience the horizon. As for the slow evolution, we recall that a soliton in an anomalous
dispersion regime (β2 < 0) will retain its shape since dispersive effects are counterbalanced by the
nonlinear ones. The optical production of light through the Kerr effect has a response time of the
electronic polarization of the material 10−15 s [55], while the evolution of the pump in the fibre is
on the order of 10−13 s.
Note that in Fig. 6 there is no probe λ2 as this is a quantum Hawking effect that is caused
directly from quantum vacuum. In a sense, it means we have a probe in all wavelengths, but in the
vacuum state. Also, all the signals from PH and redder are positive-frequency Hawking radiation
(PHR)—this includes the peaks obtained from the first-order approximation marked as λPHR —
and from PH and bluer are negative-frequency Hawking radiation (NHR). The complicated
structure is partly due to the two GHs in the realistic PCF dispersion.

5. Stimulated Hawking radiation in optics

Hawking radiation was originally obtained with quantum vacuum as the input state. However,
the Hawking effect still occurs if replace the vacuum with a different state, this is called stimulated
radiation. The stronger stimulated effect is obtained if we replace vacuum with a classical state.
This is classical stimulated Hawking radiation and is the usual effect studied in optical analogues.
In this case we use a probe pulse or a continuous wave (CW) along with the pump pulse to
generate stimulated Hawking radiation which is easier to detect. All optical experiments so far
have measured the classical stimulated effect, in 2008 it was done with a CW [10] and the latest in
2019 with a pulse [13]. Let us study now this version of the effect in optics.
In order to explain the production of new signals from the point of view of optics, it is
important to study the pulse dynamics inside the fibre. The most used model for this is the
non-linear Schrödinger equation (NLSE):

∂E |β | ∂ 2 E
i − 2 + γ|E|2 E = 0, (5.1)
∂z 2 ∂t2
that describes the electric part of the optical field E(z, t) based on the second-order dispersion
β2 and the non-linear coefficient γ. To increase the intensity of the Hawking signal we need to
increase the surface gravity α and from Eq. (4.15) this means a fast changing δn. Therefore, we 13
should use the shortest available optical pulses, with duration of around 7 fs. However, these
pulses cannot be described by the NLSE, but with a different version with less approximations

rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 0000000

..................................................................
[56] called unidirectional pulse propagation equation (UPPE):
∂Eω ω
i + β(ω)Eω + P = 0, (5.2)
∂z 2cn(ω) NL,ω
where Eω is the Fourier transform of the electric field, β(ω) = k(ω) is the full dispersion relation,
and PNL = χ(3) E 3 /8 is the non-linear polarization.
Dissipation in optical systems can be included by an extra term in the evolution equation,
sometimes called generalized nonlinear Schrödinger equation (GNLSE). However, the dynamics
of the optical Hawking effect are much faster than such dissipation and neglecting it is a very
good approximation. On the other hand, dispersion produces soliton fission and the emission
of dispersive waves at a similar length scale as the Hawking effect. For example, for the PCF
used dissipation should be considered at 100m, dispersion at 10mm, and the Hawking effect
is produced in 10mm [57,58]. It was shown in the numerical solutions and the experiment in
Ref. [13] that the Hawking effect is robust enough to survive these strong dispersive effects.
In this way, a description of the evolution of the optical fields inside the optical fibre can be
obtained. This evolution is only classical, as our starting equations are Maxwell equations, and
it can only describe the emission of the stimulated Hawking radiation, not the spontaneous one.
We set the pump and probe signals as part of the electric field E and we obtain the signal coming
from the stimulated Hawking effect.
However, as the same frequency conservation is fulfilled in both the stimulated and the
spontaneous cases, we obtain Hawking signals in the correct frequencies described by the
conservation of co-moving frequency of the probe ω 0 (ω2 ), but these signals come from a classical
frequency shifting of the probe energy and not created directly from vacuum energy [59].
The existence of the probe pulse at frequency ω2 causes a different co-moving frequency
conservation, as seen in Fig. 5. The probe (red) causes the creation of signals at two frequencies, the
positive- (PHR, orange) and negative- (NHR, blue) frequency Hawking radiation. This situation
corresponds to the black-hole horizon shown in Fig. 4 (left). For the white-hole horizon, probe
and PHR should be interchanged.
In this way, working out the pulse dynamics we can obtain the output spectrum that includes
new stimulated signals at both the PHR and NHR frequencies. An example of these spectra is
shown in Fig. 7 by solving Eq. (5.2) for a 7fs, 800nm, 50mW pump and a 50fs, 1500nm, 5mW
probe propagating through 7mm of PCF [13].

6. Conclusion
First, we studied the emission of Hawking radiation in its original derivation in the astrophysical
case. We argued for the generalization of such concept to other systems. These are called analogue
systems and the emission is called analogue Hawking radiation. Then we mentioned some
examples, in particular two that have experimental evidence, light pulses in optical fibres and
then we mentioned examples that have experimental evidence, and focus only in light pulses in
optical fibres.
We also presented a derivation of Hawking radiation in a general setting that works for any
non-dispersive system, in particular, this includes the original derivation in astrophysics but also
in any analogue system under a non-dispersive approximation. In this way it is clear that the
Hawking effect is more general than previously thought.
Then, we focused on the optical analogues. We explained the main ideas and established that
the conditions for the emission of analogue Hawking radiation are fulfilled in that case, apparent
horizon, surface gravity, and slow evolution.
Furthermore, we studied the expected signals in the quantum or spontaneous case (Fig. 6) and
in the classical or stimulated one (Fig. 7). Each with its own way of calculating the resulting
 3 2 1.5 1 0.8 0.6 0.5 0.4 0.35 0.3 0.27 0.24 0.22
20 14

rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 0000000

..................................................................
-20

-40

-60

-80

-100

-120
2 4 6 8

Figure 7. Spectrum of pump ω1 - probe ω2 interaction after 7mm of propagation in a PCF fibre obtained through
numerical solution of the UPPE (5.2). The stimulated Hawking effect is shown as NHR and PHR and the resonant
radiation as RRIR, RR, and NRR.

Astro Aquatic Sonic Optical

Figure 8. Instead of collecting “stamps” or several instances of phenomena similar to Hawking radiation, we should
generalize the concept and call them all Hawking radiation. Images: [Astro] Ute Kraus, Universität Hildesheim [61];
[Aquatic] Germain Rousseaux, Institut Pprime [21], [Sonic] Jeff Steinhauer, Technion [18]; [Optical] Yuval Rosenberg,
Weizmann Institute of Science [13].

spectrum for the Hawking emission. For the spontaneous radiation we used the scattering
matrix to solve the mixing of Bogoliubov coefficients [12,29,43]. For the stimulated one, the pulse
dynamics gives the classical emission from non-linear optics [13,60].
Finally, we would like to finish with a comment on the current state of the field of analogue
gravity in front of the scientific community. “All science is either physics or stamp collecting”,
this phrase is usually attributed to Ernest Rutherford. It is a controversial phrase, specially if
it is used by physicists. However, we would like to use it to counter an argument usually
made by physicists against the use of the name Hawking radiation to describe the so-called
analogue Hawking radiation, i.e., a Hawking-type effect in analogue systems, see Fig. 8. The
naming of physical effects is important to emphasize if they are related at all and how closely.
In this case, instead of collecting several related facts or “stamps” and giving a different name to
each one: astrophysical Hawking radiation, aquatic Hawking radiation, sonic Hawking radiation,
optical Hawking radiation; we should generalize the concept of Hawking radiation. Yes, it was
discovered originally in astrophysics, but it is realized in the same way in all analogue systems,
as it was shown in the general derivation in Section 3. This phenomenon is nothing more than
Hawking radiation.
Authors’ Contributions. RAG drafted the manuscript and perform the bibliographical review. DB 15
proposed the topic and expanded the manuscript. All authors corrected and approved the manuscript.

rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 0000000

Competing Interests. The authors declare that they have no competing interests.

..................................................................
Funding. Funding provided by Fondo SEP-Cinvestav 2018-60 (Mexico).
Acknowledgements. The authors would like to acknowledge the valuable discussions with Maxime
Jacquet, Yuval Rosenberg, and Ulf Leonhardt. Also, we are grateful to Friderich König, Silke Weinfurtner,
and Maxime Jacquet as the organisers of the scientific meeting on “The next generation of analogue systems”
at the Royal Society on December 2019 that sparked this discussion.

References
1. Hawking SW. 1974 Black hole explosions? Nature 248 30. (doi:10.1038/248030a0).
2. Misner CW, Thorne KS, Wheeler JA. 2017 Gravitation. Princeton University Press, Princeton,
USA. (isbn:9780691177793).
3. Feynman R. 2006 The Feynman Lectures on Physics. Addison-Wesley, New York, USA.
(isbn:9780465040834).
4. Unruh WG. 1981 Experimental black-hole evaporation? Phys. Rev. Lett. 46 1351.
(doi:10.1103/PhysRevLett.46.1351)
5. Confirmed by William Unruh, private communication.
6. Leonhardt U. 2006 Optical conformal mapping. Science 312 1777–1780.
(doi:10.1126/science.1126493).
7. Pendry JB, Schurig D, Smith DR. 2006 Controlling electromagnetic fields. Science 312
1780–1782. (doi:10.1126/science.1125907)
8. Gordon W. 1923 Zur Lichtfortpflanzung nach der Relativitätstheorie. Ann. Phys. 72 421
(doi:10.1002/andp.19233772202)
9. Plebanski J. 1960 Electromagnetic waves in gravitation fields. Phys. Rev. 118 1396.
(doi:10.1103/PhysRev.118.1396)
10. Philbin TG, Kuklewicz C, Robertson S, Hill S, König F, Leonhardt U. 2008 Fiber-optical analog
of the event horizon.Science 319 1367–1370. (doi:10.1126/science.1153625)
11. Bermudez D, Drori J, Leonhardt U. 2016 Dialogues about geometry and light. Frontiers in
Modern Optics 190 1 (doi:10.3254/978-1-61499-647-7-1)
12. Jacquet M. 2018 Negative frequency at the horizon. Springer Theses, Cham, Switzerland.
(isbn:9783319910710).
13. Drori J, Rosenberg Y, Bermudez D, Silberberg Y, Leonhardt U. 2019 Observation of
stimulated Hawking radiation in an optical analogue. Phys. Rev. Lett. 122 010404
(doi:10.1103/PhysRevLett.122.010404).
14. Vissser M. 2003 Essential and inessential features of Hawking radiation. Int. J. Mod. Phys. D
12 649 (doi:10.1142/S0218271803003190).
15. Birrell ND, Davies PCW. 1982 Quantum fields in curved space. Cambridge University Press,
Cambridge, UK. (isbn:0521278589).
16. Volovik G. 2003 The Universe in a helium droplet. Clarendon Press, Oxford, UK.
(isbn:9780199564842).
17. Garay LJ, Anglin JR, Cirac JI, Zoller P. 2000 Sonic analog of gravitational black holes in Bose-
Einstein condensates Phys. Rev. Lett 85 4643 (doi:10.1103/PhysRevLett.85.4643)
18. Munoz de Nova JR, Golubkov K, Kolobov VI, Steinhauer J. 2019 Observation of thermal
Hawking radiation and its temperature in an analogue black hole. Nature 569 688
(doi:10.1038/s41586-019-1241-0)
19. Nguyen HS, Gerace D, Carusotto I, Sanvitto D, Galopin E, Lemaitre A, Sagnes I, Bloch J, Amo
A. 2015 Acoustic black hole in a stationary hydrodynamic flow of microcavity polaritons Phys.
Rev. Lett 114 036402 (doi:10.1103/PhysRevLett.114.036402)
20. Weinfurtner S, Tedford EW, Penrice M, Unruh WG, Lawrence GA. 2011 Measurement
of stimulated Hawking emission in an analogue system. Phys. Rev. Lett. 106 2
(doi:10.1103/PhysRevLett.106.021302).
21. Euvé LP, Michel F, Parentani R, Philbin T, Rousseaux G. 2016 Observation of noise
correlated by the Hawking effect in a water tank. Phys. Rev. Lett. 117 121301
(doi:10.1103/PhysRevLett.117.121301)
22. Unruh W, Schutzhold R. 2007 Quantum analogues: From phase transitions to black holes and 16
cosmology. Springer, Berlin. (isbn:9783540708582)
23. Barcelo C, Liberati S, Visser M. 2011 Analogue gravity. Liv. Rel. Rev. 14 3 (doi:10.12942/lrr-

rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 0000000

..................................................................
2005-12).
24. Faccio D, Belgiorno F, Cacciatori S, Gorini V, Liberati S, Moschella U. 2013 Analogue gravity
phenomenology. Springer, Switzerland. (isbn:9783319002668)
25. Belgiorno FD, Cacciatori SL, Faccio D. 2019 Hawking radiation. From astrophysical black holes to
analogue systems in the lab. World Scientific, Singapore. (isbn:9789814508537)
26. Barcelo C. 2019 Analogue black-hole horizons. Nature Physics 15 210 (doi:10.1038/s41567-018-
0367-6)
27. Rosenberg Y. 2020 Optical analogues of black-hole horizons. Phil. Trans. A ?, ?. (doi:?)
28. Leonhardt U, Robertson S. 2012 Analytical theory of Hawking radiation in dispersive media
New J. Phys. 14 5 (doi:10.1088)
29. Bermudez D, Leonhardt U. 2016 Hawking spectrum for a fiber-optical analog of the event
horizon. Phys. Rev. A 93 053820. (doi:10.1103/PhysRevA.93.053820)
30. Moreno-Ruiz A, Bermudez D. 2019 Hawking temperature in dispersive media: Analytical and
numerical study. arXiv Preprint 1908.02368. (arXiv:1908.02368 [gr-qc])
31. Unruh W. 1995 Sonic analogue of black holes and the effects of high frequencies on black hole
evaporation. Phys. Rev. D 51 2827. (doi:10.1103/physrevd.51.2827)
32. Helfer AD. 2003 Do black holes radiate? Rep. Prog. Phys. 66 (2003) 943. (doi: 10.1088/0034-
4885/66/6/202)
33. Fedichev PO, Fischer UR. 2017 “Cosmological” quasiparticle production in harmonically
trapped superfluid gases. Nature Phys. 13 833. (doi:10.1038/nphys4151)
34. Leonhardt U. 2019 Lifshitz theory of the cosmological constant. Ann. Phys. 411 167973.
(doi:10.1016/j.aop.2019.167973)
35. Torres T, Patrick S, Coutant A, Richartz M, Tedford EW, Weinfurtner S. 2017 Rotational
superradiant scattering in a vortex flow. Nature Phys. 13 833. (doi:10.1038/nphys4151)
36. Bekenstein J. 1973 Black holes and entropy. Phys. Rev. D 7 2333. (doi:10.1103/physrevd.7.2333)
37. Jacobson T. 1991 Black hole evaporation and ultrashort distances. Phys. Rev. D 44 1731.
(doi:10.1103/physrevd.44.1731)
38. Visser M. 1993 Acoustic propagation in fluids: an unexpected example of Lorentzian geometry
arXiv 9311028 [gr-qc]
39. Mukhanov V, Winitzki S 2007 Introduction to quantum effects in gravity. Cambridge University
Press, Cambridge. (isbn:9780521868341)
40. Robertson S. 2012 The theory of Hawking radiation in laboratory analogues. J. Phys. B: At.
Mol. Opt. Phys. 45 163001 (doi:10.1088/0953-4075/45/16/163001)
41. Leonhardt U. 2010 Essential quantum optics. Cambridge University Press, Cambridge.
(isbn:9780521869782).
42. Bermudez D, Leonhardt U. 2019 Resonant Hawking radiation as an instability. Class. Quantum
Grav. 36 024001. (doi:10.1088/1361-6382/aaf435)
43. Robertson S, Leonhardt U. 2014 Integral method for the calculation of Hawking
radiation in dispersive media. I. Symmetric asymptotics. Phys. Rev. E 90 053302
(doi:10.1103/PhysRevE.90.053302)
44. Finazzi S, Carusotto I. 2014 Spontaneous quantum emission from analog white holes in a
nonlinear optical medium. Phys. Rev. A 89 5 (doi:10.1103/PhysRevA.89.053807).
45. Petev M, Westerberg N, Moss D, Rubino E, Rimoldi C, Cacciatori SL, Belgiorno F, Faccio D.
2013 Blackbody emission from light interacting with an effective moving dispersive medium
Phys. Rev. Lett.. 111 043902 (doi:10.1103/PhysRevLett.111.043902)
46. Leonhardt U. 2002 A laboratory analogue of the event horizon using slow light in ann atomic
medium. Nature 415 406 (doi:10.1038/417102a)
47. Unruh W, Schützhold R. 2003 On slow light as a black hole analogue. Phys. Rev. D 68 024008
(doi:10.1103/PhysRevD.68.024008)
48. Confirmed by Ulf Leonhardt, private communication.
49. Agrawal GP. 2013 Nonlinear fiber optics. Academic Press, Oxford, UK. (isbn:9780123970237).
50. Gaona-Reyes JL, Bermudez D. 2017 The theory of optical black hole lasers. Ann. Phys. 380 41.
(doi:10.1016/j.aop.2017.03.005)
51. Rubino E, McLenaghan J, Kehr SC, Belgiorno F, Townsend D, Rohr S, Kuklewicz C, Leonhardt
U, König F, Faccio D. 2012 Negative-frequency resonant radiation. Phys. Rev. Lett. 108 253901
(doi:10.1103/PhysRevLett.108.253901).
52. McLenaghan J and König F. 2014 Few-cycle fiber pulse compression and evolution of negative 17
resonant radiation. New J. Phys. 16 6 (doi:10.1088/1367-2630/16/6/063017).
53. Conforti M, Marini A, Tran TX, Faccio D, Biancalana F. 2013 Interaction between optical fields

rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 0000000

..................................................................
and their conjugates in nonlinear media. Opt. Exp. 21 21239. (doi:10.1364/oe.21.031239)
54. Choudhary A, König F. 2012 Efficient frequency shifting of dispersive waves at solitons. Opt.
Exp. 20 5 (doi:/10.1364/OE.20.005538).
55. Boyd RW. 2008 Nonlinear optics. Academic Press, Oxford, UK. (isbn:9780123694706).
56. Couairon A, Brambilla E, Corti T, Majus D, Ramírez-Góngora OJ, Kolesik M. 2011
Practitioner’s guide to laser pulse propagation models and simulation. Eur. Phys. J. Special
Topics 199 5 (doi:10.1140/epjst/e2011-01503-3)
57. Amiranashvili S, Radziunas M, Bandelow U, Čiegis R. 2019 Numerical methods for accurate
description of ultrashort pulses in optical fibers. Communications in Nonlinear Science and
Numerical Simulation 67 391 (doi:10.1016/j.cnsns.2018.07.031).
58. Dudley J M, Genty G, Coen S. 2006 Supercontinuum generation in photonic crystal fiber. Rev.
of Mod. Phys. 78 4 (doi:10.1103/RevModPhys.78.1135).
59. Aguero-Santacruz R, Rosenberg, Y, Leonhardt U, Bermudez D. 2020 On the origin of the
stimulated Hawking radiation in nonlinear optics. In preparation
60. Rubino E, Lotti A, Belgiorno F, Cacciatori SL, Couairon A, Leonhardt U, Faccio D. 2012 Soliton-
induced relativistic-scattering and amplification. Sci. Rep. 2 932 (doi:10.1038/srep00932).
61. Ute Kraus, Institute of Physics, Universität Hildesheim, Space Time Travel
(https://www.spacetimetravel.org/)
62. Linder MF, Schützhold R, Unruh WG. 2016 Derivation of Hawking radiation in dispersive
dielectric media. Phys. Rev. D 93 10 (doi:10.1103/PhysRevD.93.104010).
63. Belgiorno F, Cacciatori SL, Clerici M, Gorini V, Ortenzi G, Rizzi L, Rubino E, Sala VG, Faccio
D. 2010 Hawking radiation from ultrashort laser pulse filaments. Phys. Rev. Lett. 105 20
(doi:10.1103/PhysRevLett.105.203901).