Similarities in the evaporation of saturated solitons and black holes Giacomo Contri ID, Gia Dvali, Otari Sakhelashvili Abstract It has been suggested some time ago that many black hole properties are not specific to gravity, but rather are shared by a large class of objects, the so-called saturons, that saturate the quantum field theoretic upper bound on microstate degeneracy. By now, various aspects of this universality have been understood and demonstrated in a number of explicit examples. In the present paper, we add one more brick to the building by showing that the decay of a simple two-dimensional saturated soliton copies some key aspects of the black hole decay as well as of the information retrieval. In particular, we study the evaporation process of a classically-stable vacuum bubble of a spontaneously broken S ​ U ​ ( N ) SU(N) -symmetry, coupled to massless fermions. We show that the decay rate as well as the characteristic energy of the emitted quanta are given by the inverse size of the object, in striking similarity with the Hawking evaporation of a black hole. The time-scale of information retrieval also matches the one previously suggested for a black hole by Page. We give the semiclassical derivation of the phenomenon as well as its fully quantum resolution as a decaying coherent state of Goldstone bosons. The universal nature of the effect and its microscopic understanding support the analogous quantum portrait of a black hole as a saturated coherent state of gravitons. I Introduction Black holes have long been considered as objects in their own category, characterized by certain unique properties. Perhaps, the feature that has been considered to be the most mysterious is the Bekenstein-Hawking entropy, which follows an area law in Planck units Bekenstein 1973 , S\sim A\,M_{P}^{d-2}\,, (1) where d is the dimensionality of the space-time and M_{P} the corresponding Planck mass. This entropy also saturates the Bekenstein Bound Bekenstein 1981 S\leq 2\pi ER , where E is the energy and R is the radius of the object. The black hole exhibits other striking features. In the semiclassical description it is characterized by an information horizon as well as by the thermal-like Hawking evaporation rate Hawking 1975 , \Gamma\sim\frac{1}{R}\,. (2) Both the rate as well as the characteristic energy of emitted quanta are given by the inverse size of the system. Another mysterious property of a black hole is a macroscopically-long time-scale of information retrieval. Namely, despite having a maximal information storage capacity, a semiclassical black hole radiates energy but no (or negligible) information. The duration of the validity of the semiclassical regime, which ignores the quantum backreaction, is a separate important question entangled with the question of information retrieval. This shall be discussed separately. However, there is little doubt that, within the validity of the semiclassical treatment, the information carried by radiation is unreadable. This is puzzling. One may be tempted to attribute the above features to peculiarities of quantum gravity, however this turned out not to be the case. First, in Dvali 2018a it was shown that the feature of the delayed information-retrieval is a universal property of all systems with enhanced capacity of information storage, regardless of their origin. All such systems at the initial stages radiate energy but maintain the load of information internally for a macroscopically long time-scale. However, the universality of black hole features does not stop here. The unifying framework was proposed in Dvali 2021a ; Dvali 2021b ; Dvali 2021c , where it was argued that none of the above-discussed features are unique to black holes. Rather, they are common for a universality class of objects, the so-called “saturons”, that saturate the quantum field theoretic (QFT) bound on the microstate degeneracy. In Dvali 2021a ; Dvali 2021b ; Dvali 2021c the bound has been formulated in three distinct forms, which emphasize different physical meanings of the saturation effect. First, the bound says that in any QFT the maximal microstate entropy attainable by a localized object of radius R is given by S_{\rm max}\,=\,\frac{1}{\tilde{\alpha}}\,, (3) where \tilde{\alpha} is a dimensionless quantum coupling of the interaction responsible for the existence of the object. In the above expression, the running coupling \tilde{\alpha} has to be evaluated at the scale R . Moreover, \tilde{\alpha} also sets the upper bound on the occupation number of quanta N in the bound state by the self-sustainability condition \tilde{\alpha}N=1 . Thus, for a saturated system the microstate entropy is bounded by the number of constituents Dvali 2021b , S_{\rm max}\,=\,N\,. (4) The alternative (and perhaps a more striking) presentation of the same bound is that the maximal entropy is given by the area S_{\rm max}\,=\,A\,f_{P}^{d-2}\,, (5) where f_{P} is the scale of spontaneous breaking of the Poincaré symmetry by the object. The connection between ( 5 ) and ( 3 ) becomes very clear by noticing that the scale f_{P} is the scale (“decay constant”) of the Poincaré Goldstone. Correspondingly, the dimensionless coupling of this Goldstone, evaluated at the momentum scale 1/R is ( 3 ). In other words, at the point when a system reaches the maximal microstate degeneracy, the coupling \tilde{\alpha} of the interaction responsible for the existence of the object matches the coupling of the Goldstone mode of the spontaneously-broken Poincaré symmetry. It has been shown that going beyond the bounds ( 3 ), ( 4 ) and ( 5 ) invalidates the QFT description, in particular, by the breakdown of the loop-expansion in \tilde{\alpha} Dvali 2021b ; Dvali 2021c , and by the violation of unitarity in the scattering processes Dvali 2021a . Therefore, the phenomenon of saturation gives a transparent meaning to the area-form of the entropy: it manifests the universal feature of the (inverse) Goldstone coupling which at wavelength R in an arbitrary number of space-time dimensions scales as the corresponding area A\sim R^{d-2} . Once we understand the above, it becomes immediately clear that the black hole entropy ( 1 ) is a particular manifestation of the more general expressions ( 5 ) and ( 3 ) Dvali 2021a ; Dvali 2021b ; Dvali 2021c . Indeed, for equating ( 1 ) and ( 5 ), it suffices to notice that for a black hole the scale of the Poincaré-breaking is the Planck scale, f_{P}=M_{P} . Likewise, it is clear that the relation between the Bekenstein-Hawking entropy ( 1 ) and the quantum coupling of graviton evaluated at the scale R , \tilde{\alpha}=(RM_{P})^{d-2} , satisfies the relation, as it was originally pointed out in Dvali and Gomez 2013a . The bound described by ( 3 ) and ( 5 ) is therefore a more fundamental bound, which encompasses in itself the area law, for which the Bekenstein-Hawking entropy represents a particular case. We note that in many instances, the saturation of the bounds ( 3 ), ( 4 ), ( 5 ) also implies the saturation of the Bekenstein bound. However, the three former bounds appear more general, as they must be respected regardless of the latter. Now, whenever the bound ( 3 ) (and thereby ( 4 ) and( 5 )) is saturated, it turns out that all the properties that previously were thought to be exclusive to black holes also appear. In particular, all saturons display the relations ( 2 ). This has been demonstrated by explorations of saturons in renormalizable theories in which calculations can be performed reliably Dvali 2021a ; Dvali 2021d ; Dvali and Venugopalan 2022 ; Dvali and Sakhelashvili 2022 ; Dvali et al. 2022a . In addition, it has been shown Dvali et al. 2022b ; Dvali et al. 2024a that saturons exhibit the relation between the maximal spin and the entropy identical to black holes. The understanding of all aspects of the phenomenon of saturation is of fundamental importance because of the following reasons. First, it allows us to understand the mysterious black hole properties in terms of a universal underlying phenomenon that goes far beyond gravity. Due to this universality, saturons in renormalizable theories can be used as test laboratories not only for explaining the known black hole properties but also for predicting new effects. One example of a new phenomenon is provided by the so-called “memory burden” effect Dvali 2018a ; Dvali et al. 2020 ; Dvali et al. 2019 , which suggests that the information load carried by any system of enhanced capacity of information storage (such as a black hole), exerts a stabilizing back-reaction on it. The stabilization of saturated solitons by the memory burden effect has been discussed in Dvali and Sakhelashvili 2022 ; Dvali et al. 2022a ; Dvali et al. 2024b . Various potentially-observational implications of this effect have been discussed in the literature, but shall not be considered here 1 1 1 For early work on implications for primordial black holes see, Dvali et al. 2020 ; Alexandre et al. 2024 ; Thoss et al. 2024 ; Dvali et al. 2024b . . Next, the understanding of saturons in renormalizable theories allows us to shed light on the microscopic structure of a black hole. In particular, it provides a framework for testing the microscopic description of a black hole as a saturated coherent state of gravitons of occupation number N=\tilde{\alpha}^{-1} Dvali and Gomez 2013a ; Dvali and Gomez 2014a ; Dvali and Gomez 2014b . This picture, called the black hole’s quantum N -portrait, explains the black hole properties microscopically. In particular, one can explicitly recover features of the Hawking radiation ( 2 ) from the quantum rescattering of the graviton constituents. The study of saturons in renormalizable theories allows us to verify this correspondence, since saturons admit an explicit microscopic resolution in terms of coherent states of quanta. They can therefore demonstrate how the black hole-like features emerge from such states. For example, the saturation of the entropy bound ( 4 ) by a black hole is automatic within the N -portrait description Dvali and Gomez 2013a ; Dvali and Gomez 2014a ; Dvali and Gomez 2014b . This feature has been explicitly verified for other saturons and shall also be the case in the present work. Finally, the study of saturons is important for the general approach of understanding the strong field non-perturbative semiclassical phenomena in terms of fully quantum multi-particle effects. This program was initiated in Dvali and Gomez 2013a ; Dvali and Gomez 2014a ; Dvali and Gomez 2013b ; Dvali and Gomez 2012 ; Dvali and Gomez 2014b ; Dvali et al. 2013 ; Dvali et al. 2017 ; Dvali and Gomez 2016 ; Dvali 2025 for strong field gravitational systems such as black holes and de Sitter, and it was generalized to non-gravitational soliton and instanton processes in a series of papers Dvali 2021b ; Dvali 2021c ; Dvali 2021a ; Dvali and Venugopalan 2022 ; Dvali and Sakhelashvili 2022 ; Dvali et al. 2022a ; Dvali and Eisemann 2022 ; Berezhiani and Zantedeschi 2021 ; Berezhiani et al. 2022a ; Berezhiani et al. 2022b ; Berezhiani et al. 2024 ; Berezhiani et al. 2025 ; Dvali and Zell 2018 . The main focus of the present paper will be on retrieving the Hawking-like evaporation rate ( 2 ) for a specific example of a saturon in a 1+1 -dimensional theory. This decay rate has been reproduced previously in Dvali and Venugopalan 2022 ; Dvali and Sakhelashvili 2022 ; Dvali et al. 2022a in 1+1 and 3+1 dimensional examples of saturons, using the picture of the S -matrix scattering of constituents. In Dvali and Venugopalan 2022 , the role of the saturon was played by a bound-state of the Gross-Neveu model Gross and Neveu 1974 , whereas in Dvali and Sakhelashvili 2022 ; Dvali et al. 2022a the saturon is represented by a stabilized version of the vacuum bubble of a spontaneously broken SU(N) -symmetry Dvali 2021a . The example considered in the present paper is a 1+1 -dimensional version of the above-type vacuum bubble. The novelty is that we shall perform computations both in a semiclassical as well as in a fully quantum regime, and demonstrate the matching. The semiclassical computation can be regarded as a non-gravitational analog of the Hawking analysis. In this treatment, the saturon is taken as a background which sources quantum particles in the zero-back-reaction limit. In this regime, the particle creation is a vacuum process very similar to Hawking’s semiclassical effect in a black hole. Thanks to the simplicity of the model, some results can be obtained analytically. On the other hand, in the fully quantum description, the decay rate is obtained as a result of the re-scattering and decay of the constituent quanta. For the vacuum bubble, these constituents represent the Nambu-Goldstone modes of the spontaneously broken SU(N) -symmetry localized within the bubble interior. On the gravitational side, this analysis represents a counterpart of the black hole decay described within the quantum N -portrait Dvali and Gomez 2013a . In the proper limit, the two approaches match, as expected. The presented saturon model thus can be viewed as a calculable laboratory that allows to explicitly interpolate between the Hawking-type semiclassical and quantum-corpuscular regimes. The advantage of the corpuscular picture is that, being more fundamental, it allows us to go beyond the leading-order approximation and estimate the corrections to the semiclassical behaviour. In particular, it allows us to explicitly track the process of the information retrieval and to establish the required time-scales, which are in accordance with the previous estimates Dvali 2021a ; Dvali and Venugopalan 2022 ; Dvali and Sakhelashvili 2022 ; Dvali et al. 2022a . Within the quantum picture, these time-scales are finite but go to infinity in the semiclassical limit. We thus explicitly observe, at the level of a calculable unitary theory, that seeming non-recovery of information is an artifact of the semiclassical approximation. The paper is structured as follows: in the second section, we introduce the model and we discuss the bubble-like solutions which appear in its spectrum, with a particular focus on saturation. In the third section, we describe the setting of the semiclassical computation and we obtain the radiation spectrum of the bubble in the saturation limit. In the fourth section, we give the quantum description of the evaporation process. In the fifth section, we interpret the results and estimate the corrections to the semiclassical behaviour. In the sixth section, we present the quantum derivation of the time-scale of the information retrieval. In the seventh section, we conclude. II A model of a saturon II.1 The classical soliton We shall study the two-dimensional version of the model of a saturated vacuum bubble originally introduced in Dvali 2021a . We shall focus on the stabilized version of the bubble derived in Dvali et al. 2022a . The theory is defined by the following Lagrangian: \displaystyle\mathcal{L} \displaystyle=\frac{1}{2}\operatorname{tr}\left[\left(\partial_{\mu}\Phi\right)\left(\partial^{\mu}\Phi\right)\right]-V[\Phi] (6) \displaystyle V[\Phi] \displaystyle=\frac{\alpha}{2}\operatorname{tr}\left[\left(f\Phi-\Phi^{2}+\frac{I}{N}\operatorname{tr}\left[\Phi^{2}\right]\right)^{2}\right]\,, (7) where \Phi is a scalar field in the adjoint representation of the SU(N) -group. Since we are in 1+1 space-time dimensions, \Phi is dimensionless. So is the parameter f , whereas the parameter \alpha has dimensionality of [mass] 2 . The dimensionless coupling, \tilde{\alpha} , controlling the strength of the interaction, is determined as \tilde{\alpha}\equiv\frac{\alpha}{m^{2}}=\frac{1}{f^{2}}\,. (8) In order to obtain saturated states, we shall work in the limit of large N and large f (weak coupling). Defining the collective (’t Hooft) coupling as N/f^{2} , unitarity imposes the following upper bound on it, N/f^{2}\lesssim 1\,. (9) The potential ( 7 ) admits the SU(N) -symmetric vacuum as well as a number of degenerate vacua in which the SU(N) symmetry is spontaneously broken down to various maximal little groups, SU(N-K)\times SU(K)\times U(1) with 0R . The interior of a bubble ( |x|0 . The generalization of our analysis to a field defined on the entire real axis is trivial. For \dfrac{\omega^{2}}{\alpha f^{2}}\,<\,1 (or, equivalently, for \omega^{2}\,<\,m^{2} ), the above potential has two maxima, \phi=0 and \phi_{\text{max}}=\frac{f}{4}(3+\sqrt{1+\frac{8\omega^{2}}{\alpha f^{2}}}) , separated by a minimum at \phi_{\text{min}}=\frac{f}{4}(3-\sqrt{1+\frac{8\omega^{2}}{\alpha f^{2}}}) . In addition to \phi=0 , the potential also becomes zero at the points \phi_{0}=f\pm\frac{\omega}{\sqrt{\alpha}} . In this regime, the existence of a stationary bubble solution is rather transparent. The “particle” starts at the “time” x=-\infty with the initial position \phi=0 and with zero initial velocity, \phi^{\prime}=0 . At some point it rolls down the hill, climbs on the other side of the slope and reaches the point \phi_{0}=f-\frac{\omega}{\sqrt{\alpha}} with the zero velocity. It then bounces and repeats the excursion backward: rolls down, then up and reaches the initial maximum \phi=0 at x=+\infty . Thus, the bouncing point \phi_{0}=f\left(1-\frac{\omega}{f\sqrt{\alpha}}\right) (18) determines the value of the field at the center of the bubble. The bubble size, R , can be defined as the characteristic “time” interval required for the entire round trip. Each one-way journey represents a kink and an anti-kink, separated by the time interval that the field spends near the bouncing point \phi_{0} . This time interval is controlled by the frequency \omega . Notice that the time interval spent at the bouncing point \phi_{0}=f-\frac{\omega}{\sqrt{\alpha}} goes to infinity for \omega\rightarrow 0 , since in this limit the bounce-point coincides with the maximum: \phi_{0}=\phi_{\text{max}}=f . Correspondingly, the particle spends an infinite time there before starting the return journey. That is, the time interval separating the onward and backwards transitions becomes infinite. In other words, in this limit, the bubble becomes a kink and an anti-kink, given by ( 11 ), at infinite separation, R\rightarrow\infty . Thus, as in the 3+1 case of Dvali et al. 2022a , choosing a certain frequency \omega for the Goldstone oscillations in the bubble interior fixes the size R and the height \phi_{0} of the bubble profile as in ( 18 ). However, in contrast to the 3+1 -dimensional case Dvali et al. 2022a , in the 1+1 case for \omega\rightarrow 0 the size of the stationary bubble R increases only logarithmically. Namely, in the thin-wall approximation m\gg\omega we have R\sim\frac{1}{m}\ln\left(\frac{m}{\omega}\right)\,. (19) The reason for the slow growth is that for R\gg m^{-1} , the attractive potential between the kink and anti-kink is suppressed exponentially. Thus, in the thin-wall limit \phi_{0}\simeq f . Moreover, we have R\omega\sim\frac{\omega}{m}\ln\left(\frac{m}{\omega}\right)\ll 1 (20) and \frac{\omega}{\phi_{0}}\ll\sqrt{\alpha}\,. (21) For thick-wall bubbles, on the other hand, we have \omega\sim m , R\omega\sim 1 and \phi_{0}1/R . Figure 7: A would be N\rightarrow 2 process depicting the annihilation of N Goldstones into a pair of highly energetic fermions. Processes of this kind are forbidden by charge conservation. There is a natural explanation for this. In the case of a black hole, we can understand the suppression microscopically in the following way. Thinking of a black hole as a coherent state of gravitons Dvali and Gomez 2013a , we have gravitons of all possible frequencies. The distribution is peaked around the frequency 1/R . However, much higher frequencies E\gg 1/R are also present albeit in exponentially suppressed numbers \sim e^{-ER} . The production of energetic particles in the case of a black hole has at least two sources. The first is a direct decay of high-frequency constituent gravitons. Obviously, because of their suppressed number, the corresponding rate is exponentially suppressed. The second source for production of asymptotic particles of energy E\gg 1/R is the rescattering of a large number ( \Delta N\sim ER ) of constituent gravitons into a quantum of energy E . Such rates are exponentially suppressed as e^{-ER} Dvali and Gomez 2013a ; Dvali et al. 2015b ; Dvali and Eisemann 2022 . The suppression of the transitions with big changes in the number of particles is a generic effect. For gravitons, it has been observed in the multi-graviton scattering amplitudes Dvali et al. 2015b ; Addazi et al. 2017 . The same feature is responsible for the suppressed production of solitons in particle scattering Brown 1992 ; Voloshin 1992 ; Argyres et al. 1993 ; Gorsky and Voloshin 1993 ; Libanov et al. 1994 ; Libanov et al. 1995 ; Son 1996 ; Dvali 2018b ; Monin 2018 ; Cornwall 1990 ; Goldberg 1990 . In the case of a saturated stationary bubble, none of these channels is available in the semiclassical limit. First, there are no constituent Goldstones of frequencies different from 1/R . Thus, the production of a high energy quantum from the decay of a single constituent is impossible. Secondly, each Goldstone carries a charge. Correspondingly, the annihilation of \Delta N=ER\gg 1 Goldstones must produce a quantum of a corresponding charge \Delta N . However, such quanta do not exist in the theory. Notice that, since the bubble is not oscillating, the radial modes carry no frequencies but only momenta. Due to this, the rescattering of radial modes cannot increase the energy of the produced particle. In this sense, the bubble shares some features of static topological solitons, which cannot decay due to exact same reason: no constituents of non-zero frequencies are available Dvali et al. 2015c ; Berezhiani et al. 2025 . The above reasoning, presented in the zero back-reaction limit ( 50 ), explains in simple terms the peculiarities of the bubble spectrum as well as some of its differences from the spectrum of a semiclassical black hole. Once we go away from this limit, the spectrum can get corrected due to various 1/N -effects. For example, since for finite- N the bubble has a finite mass, the spectrum will be corrected due to a non-zero recoil. Also, if we make the bubble oscillate radially around the classical equilibrium point, the radiation of modes with higher frequencies can appear, and will be exponentially suppressed Dvali and Eisemann 2022 . VI time-scale of information retrieval Let us now discuss the question of the time-scale of the information retrieval. Our reasoning and conclusions will be similar to the ones of the earlier papers Dvali 2021a ; Dvali 2021d ; Dvali and Sakhelashvili 2022 ; Dvali et al. 2022a ; Dvali 2016 addressing this issue. First, within the validity of the semiclassical regime, it is impossible to retrieve any information from saturons, in full analogy with the case of black holes. In contrast, in the quantum treatment the time-scale of information retrieval is finite, but macroscopic. This can be understood in very general terms of the necessary departure from a Hawking-like self-similar evaporation regime due to 1/S -corrections to thermality Dvali 2016 . Such corrections are in full correspondence with the prediction of the black hole’s N -portrait Dvali and Gomez 2013a . However, the lesson from saturons allows us to understand the universality of this phenomenon, and to visualize it within a simple calculable framework. For a saturated bubble, the information is stored in the SU(N) quantum numbers of the Goldstones. As classified in Dvali 2021a , for an outside observer the information readout is possible in two ways, passive and active. The passive regime implies the detection of the outgoing radiation and the analysis of its SU(N) -flavor content. The active regime implies the readout of the bubble’s flavor state by means of the scattering of external quanta at the bubble. Let us first discuss the passive method. There are two factors that delay the information read-out. The first factor is the weak coupling. For a saturated bubble, the emitted radiation has a frequency \omega\sim 1/R and it interacts via the coupling constant \tilde{\alpha}\sim 1/N\sim 1/S . The rate for resolving such a quantum in a detector is, \Gamma_{\rm det}\sim\frac{1}{R}\frac{1}{N^{2}}{N_{\rm det}}\,. (82) where, the factor 1/N^{2} comes from the coupling and the N_{\rm det} accounts for the capacity of the detector. At its maximum, this quantity is N_{\rm det}\sim N . The rate ( 82 ) fully resonates with the general black hole argument about the resolution of information encoded in 1/S -corrections to thermality Dvali 2016 . It also matches the picture of black hole N -portrait according to which the information is encoded in radiation at the level of 1/N\sim 1/S effects Dvali and Gomez 2013a ; Dvali and Gomez 2013b ; Dvali and Gomez 2014a ; Dvali and Gomez 2014b . However, the detection of a particle is just a beginning of the information read-out, since the identification of a single quantum gives only an exponentially small part of the bubble’s information content. Thus, even with a most sensitive detector, the start of the information retrieval is bounded from below by the time, t\gtrsim\frac{1}{\Gamma_{\rm det}}\sim\,S\,R\,. (83) As already pointed out in Dvali 2021a ; Dvali and Venugopalan 2022 ; Dvali and Sakhelashvili 2022 ; Dvali et al. 2022a , this time-scale is strikingly similar to the one proposed for a black hole by Page Page 1993 . As discussed in Dvali 2021d , a universal way for understanding the information retrieval-time ( 83 ) is that it is equal to the time-scale ( 38 ) given by the inverse gaps of the memory modes ( 37 ). It is clear that ( 38 ), and correspondingly ( 83 ), is the minimal time required for processing the information encoded in any mode of frequency ( 37 ). The story with the active retrieval of information is different Dvali 2021a . This way of information retrieval for black holes has not been discussed in the standard semiclassical context. In general, it is rather sensitive to the memory burden effect Dvali 2018a ; Dvali et al. 2020 . Following Dvali 2021a , let us first discuss the active retrieval in a pre-memory-burden phase. This can be achieved by scattering the external radiation at a bubble. In our model, the role of the external probes can be played by the massive quanta of the adjoint field or by the massless fermion species that are coupled to it. The scattering rate of a fermion of energy \sim 1/R at a saturated bubble is of order one. This is because the suppression by the coupling 1/N is compensated by the occupation number of Goldstones N_{G}\sim N . However, this leading order effect does not reveal the bubble’s SU(N) -flavor content. The resolution of the flavor quantum numbers is again suppressed by ( 82 ), amounting to the time-scale ( 83 ). However, the above reasoning does not take into account the memory burden effect Dvali 2018a ; Dvali et al. 2020 . Here, we shall limit ourselves by a qualitative discussion. The detailed analysis of memory burden in saturated vacuum bubbles can be found in Dvali et al. 2024b . The memory burden of a black hole can be imitated by the bubble if we endow the fermions with bulk masses above the bubble’s Goldstone frequency \omega . This creates an energy difference between the bubble constituents and the asymptotic quanta of the same SU(N) -charge. Notice that within the composite framework of black holes Dvali and Gomez 2013a ; Dvali and Gomez 2014a ; Dvali 2018c ; Dvali 2018a this gap is automatic, and goes up to \sim M_{P} . This is because in a black hole the information is encoded in high angular momentum modes, which outside of the black hole have very high frequencies but are gapless within the black hole. The prototype Hamiltonians capturing this effect have been constructed in Dvali 2018c ; Dvali 2018a . Once the bubble is in the memory-burden phase, the emission is suppressed, and the active retrieval of information is the most effective way of reading it out. From here on, the time-scale depends on the particularities of the system and can be different for black holes and other saturons (see Dvali et al. 2024b ). VII conclusions One of the key features of black hole physics is that the Hawking decay rate, \Gamma , as well as the characteristic energy of the emitted particles, E , are both set by the size, R , of the object, \Gamma\sim\frac{1}{R}\,,~~~\,E\sim\frac{1}{R}\,. (84) At the same time, despite radiating the energy, the information stored in a black hole is not coming out until a macroscopically long time. On the other hand, in recent years it has been shown Dvali 2021a ; Dvali 2021d ; Dvali and Venugopalan 2022 ; Dvali and Sakhelashvili 2022 ; Dvali et al. 2022a ; Dvali et al. 2022b ; Dvali et al. 2024a that many features of a black hole are universally shared by the objects, the so-called saturons, that exhibit the maximal microstate degeneracy permitted by the QFT consistency. The corresponding bound on the microstate degeneracy is given by three equivalent expressions ( 3 ), ( 4 ) and ( 5 ). Within a consistent QFT, all bounds are saturated simultaneously. Thereby, they express different physical meanings of the same fundamental law. In particular, all three bounds are saturated by a black hole. It has therefore been suggested Dvali 2021a that black hole features are not specific to gravity, but rather represent a universal outcome of the phenomenon of saturation. In the present paper, we have demonstrated that the particular features ( 84 ) are indeed exhibited by a class of saturated solitons in 1+1 dimensions. The soliton of our interest represents a bubble of a vacuum with spontaneously broken SU(N) -symmetry embedded in the symmetric vacuum. Due to the spontaneous breaking of the symmetry, the bubble houses a large number ( \sim N ) of gapless Goldstone modes. These modes represent the memory modes responsible for the maximal microstate degeneracy of the bubble. Classically, the bubble is stabilized by the Goldstone charge. However, in the quantum theory, it can decay via the emission of bulk particles. The black hole-like features of saturon decay were previously demonstrated via the S -matrix analysis in Dvali and Sakhelashvili 2022 , where the saturon states have been shown to exist in Gross-Neveu theory Gross and Neveu 1974 . The case closer to the present paper is the S -matrix analysis of the decay of the saturon bubbles in 3+1 -dimensions presented in Dvali et al. 2022a . The lower-dimensional case studied in the present work allows us to derive the bubble decay more explicitly and to see a clear connection between the semiclassical and quantum-composite pictures. We analyzed the evaporation spectrum of a saturated bubble coupled to a massless fermionic field, using the semiclassical methods as well as the fully quantum picture of the bubble. As predicted from the general phenomenon of saturation, when the bubble enters the saturation regime, its decay exhibits the black hole features given by ( 84 ). The evaporation rate is obtained from the explicit computation of the Bogoliubov coefficients of the fermionic field at asymptotic times. We show that the semiclassical result matches the estimation from the fully quantum corpuscular description of the bubble. In this picture, the bubble evaporation is the result of the decay of the excited Goldstone modes, localized in the bubble interior, into the free bulk particles. The presented model allows to explicitly monitor the mechanisms of information storage and retrieval at the level of a calculable microscopic theory. As in the previous work on saturons Dvali 2021a ; Dvali and Venugopalan 2022 ; Dvali and Sakhelashvili 2022 ; Dvali et al. 2022a , we observe the emergence of the time-scale of information retrieval ( 83 ). We note that this time-scale reproduces both the general estimate of 1/S -departures from thermality of black hole evaporation Dvali 2016 as well as the explicit predictions from the black hole N -portrait Dvali and Gomez 2013a about the strength of such departures. It is striking (but perhaps hardly surprising) that the time-scale ( 83 ) fully matches the one previously introduced by Page Page 1993 in the context of a black hole. The saturon model gives a transparent microscopic meaning to this time-scale and shows why it extends far beyond gravity. At the same time, the saturon model also shows that the information retrieval process can be strongly influenced by the memory burden effect Dvali 2018a ; Dvali et al. 2020 ; Dvali et al. 2019 , which however we only discussed very briefly (for a recent detailed study of the effect in saturated bubbles, see Dvali et al. 2024b ). The success of this quantum picture, which is rather transparent in non-gravitational theories, also provides an indirect support for the analogous quantum picture of a black hole as of a saturated coherent state of gravitons Dvali and Gomez 2013a . The observed behaviour shows on an explicit example that saturons, even in a renormalizable 1+1 -dimensional theory that is drastically different from gravity, share the black hole properties such as the decay rate ( 84 ). This further reinforces the idea that these objects all belong to the same universality class of saturons. Appendix A derivation of the evaporation spectrum In this appendix, we provide some details of derivation of the radiation spectrum ( 62 ). We begin by considering the Eq ( 60 ). As explained above, it is sufficient to consider the two flavors that transform non-trivially under the generator of SU(2) -subgroup responsible for the bubble charge. Then, the evaporation dynamics is reduced to four relevant degrees of freedom. In the thin-wall approximation, the Eq ( 60 ) splits into three free Dirac-like equations, two in the two asymptotic regions outside of the bubble and one in its interior. Notice that there is a time-dependent symmetry of the fermionic system in the background of the bubble. It is generated by a simultaneous time-translation and an SU(2) -rotation in the flavor space, Q=\partial_{t}-\omega T_{1} . Thus, we can label the modes according to the associated charge. This charge is exactly the energy \epsilon of the modes, after the time-dependent fermion redefinition that we performed using ( 57 ). On the other hand, translational invariance is broken by the bubble itself. Hence, from now on, it will be convenient to express the momentum k of the modes, both asymptotically and inside the bubble, as a function of \epsilon , and not vice-versa, as it is customary. This will make the matching of the mode behaviour between the different regions more transparent. In the two asymptotic regions, the modes describe a standard two-flavored massless fermion, where each flavor dispersion relation is modified by the field redefinition: k_{1/2}(\epsilon)=\pm\epsilon\pm\frac{\omega}{2}\,. (85) Inside the bubble, both fermions acquire an effective mass m_{\psi}\sim g\phi_{0} , and interact with the excited Goldstones. This leads to a flavor-dependent modified dispersion relation: k_{1/2}^{2}=\epsilon^{2}-\left(\frac{m_{\psi}}{2}\right)^{2}+\left(\frac{\omega}{2}\right)^{2}\pm\frac{1}{2}\sqrt{4\omega^{2}\epsilon^{2}+m_{\psi}^{4}-\omega^{2}m_{\psi}^{2}}\,. (86) The polarization vectors are modified accordingly. Note that for some frequencies the momenta of the modes become imaginary. This signals that the fermionic particles cannot propagate on-shell inside the bubble, as their energy would be smaller than their effective mass inside the bubble. Nevertheless, these off-shell modes must not be dismissed as they can connect the two asymptotic regions through tunneling. The modes of Eq ( 60 ) hence are of the following form: \psi_{\epsilon}(x,t)=\begin{cases}\sum_{j}\tilde{c}^{j}\tilde{\mathbf{u}}^{\epsilon}_{j}e^{\text{i}\left(\tilde{k}_{j}(\epsilon)x-\epsilon t\right)}&|x|R\,,\end{cases} (87) where quantities with a tilde refer to the solution inside the bubble, and the summation goes over both spin and flavor degrees of freedom. This expression holds for all frequencies \epsilon , both positive and negative. Note that the coefficients c^{i} have to be interpreted either as creation or annihilation operators, depending on their frequencies. This is crucial for understanding how particle creation from the vacuum can occur. In order to obtain the full solution for the modes, we need to match the coefficients of the plane-wave expansion in the different regions for a given frequency. This allows us to obtain c_{L}^{i} as a function of c_{R}^{i} c^{Li}_{\epsilon}=S_{\epsilon}{}^{i}{}_{j}c^{R\ j}_{\epsilon}\,. (88) The complete expression for S_{ij} is quite lengthy and we do not report it here. However, it contains all the necessary ingredients for extracting the coefficients of the Bogoliubov transformation ( 58 ). Below, for illustrative purposes, we write explicitly the Bogoliubov coefficients for one of the outgoing ladder operators in the expansion for small g\phi_{0}/\omega . As explained in the main body of the paper, the asymptotic vacuum can be defined at t=\pm\infty by considering modes that propagate incoming or outgoing waves respectively. Once the full spectrum of modes is obtained by ( 88 ), one can write the coefficients c_{i} corresponding to outgoing waves as functions of the incoming ones as in ( 58 ). A particle creation can occur when the Eq ( 58 ) mixes modes with positive and negative frequencies. If we undo our time-dependent fermionic field redefinition, it is clear from the Eq ( 85 ) that this happens for \epsilon\leq\frac{\omega}{2} . Indeed, given that \epsilon is conserved, the only mixing that can take place is among the modes that in the time-dependent basis have frequencies \epsilon^{\prime} and \epsilon^{\prime}-\frac{\omega}{2} . Consequently, positive and negative frequencies overlap only in the range discussed above. Outside of this range, particle creation from the vacuum is excluded in the semiclassical limit. However, we still have a non-trivial scattering against the bubble. We report one of the Bogoliubov equations for the outgoing ladder operator associated with the flavor of the bubble (the others are similar and can be obtained as described above): \begin{split}c_{1,\text{out}~\epsilon}^{L\dagger}=\;&c_{2,\text{in}~\epsilon}^{L\dagger}\left(\frac{ig\phi_{0}\sin(R (2 \epsilon- \omega))}{2\epsilon-\omega}\right)+c_{3,\text{in}~\epsilon}^{R}\left(\frac{ig\phi_{0}\sin(2 R \epsilon)}{2\epsilon}\right)\\ &+c_{1,\text{in}~\epsilon}^{R\dagger}\left(-\frac{2i\epsilon^{2}(g\phi_{0})^{2}e^{-iR\omega}\sin(4 R \epsilon)}{(\omega^{2}-4\epsilon^{2})^{2}}+\frac{8i\epsilon^{3}(g\phi_{0})^{2}\sin(R \omega)}{\omega(\omega^{2}-4\epsilon^{2})^{2}}\right.\\ &\qquad-\frac{\omega\epsilon(g\phi_{0})^{2}e^{-iR\omega}\cos(4 R \epsilon)}{(\omega^{2}-4\epsilon^{2})^{2}}+\frac{\omega\epsilon(g\phi_{0})^{2}(\cos(R \omega)-3i\sin(R \omega))}{(\omega^{2}-4\epsilon^{2})^{2}}\\ &\qquad+\frac{\omega^{2}(g\phi_{0})^{2}\sin(4 R \epsilon)(\sin(R \omega)+i\cos(R \omega))}{2(\omega^{2}-4\epsilon^{2})^{2}}-\left.\frac{\omega^{3}(g\phi_{0})^{2}e^{-iR\omega}\sin^{2}(2R\epsilon)}{2\epsilon(\omega^{2}-4\epsilon^{2})^{2}}\right)\\ &+c_{4,\text{in}~\epsilon}^{L}\left(\frac{2\epsilon^{2}(g\phi_{0})^{2}(\sin(R \omega)+i\cos(R \omega))\sin(R (4 \epsilon- \omega))}{(\omega^{2}-4\epsilon^{2})^{2}}-\frac{8iR\epsilon^{3}(g\phi_{0})^{2}}{(\omega^{2}-4\epsilon^{2})^{2}}\right.\\ &\qquad+\frac{\omega\epsilon(g\phi_{0})^{2}e^{-2iR\omega}\left(e^{4iR\epsilon}-e^{2iR\omega}\right)}{(\omega^{2}-4\epsilon^{2})^{2}}-\frac{2iR\omega\epsilon^{4}(g\phi_{0})^{2}}{(\omega^{2}-4\epsilon^{2})^{2}}\\ &\qquad+\frac{\omega^{2}(g\phi_{0})^{2}e^{-2iR(\omega+2\epsilon)}\left(2e^{2iR\omega}-3e^{2iR(\omega+2\epsilon)}+e^{8iR\epsilon}\right)}{4(\omega^{2}-4\epsilon^{2})^{2}}\\ &\qquad+\frac{3iR\omega^{2}\epsilon^{3}(g\phi_{0})^{2}}{(\omega^{2}-4\epsilon^{2})^{2}}+\frac{iR\omega^{3}\epsilon^{2}(g\phi_{0})^{2}}{2(\omega^{2}-4\epsilon^{2})^{2}}\\ &\qquad+\frac{\omega^{4}(g\phi_{0})^{2}(1-e^{-4iR\epsilon})}{16(\omega^{2}-4\epsilon^{2})^{2}}-\left.\frac{iR\omega^{4}\epsilon(g\phi_{0})^{2}}{4(\omega^{2}-4\epsilon^{2})^{2}}+1\right)\,.\end{split} (89) The precise computation of the particle spectrum follows directly from the expression for the modes contained in Eq. ( 88 ). This has been obtained exactly in the thin-wall limit, and then expanded in series both for small g\phi_{0}/\omega and in the opposite regime, in order to give the radiation spectra ( 62 ) and ( 64 ). Appendix B calculation of the wavefunction overlap suppression In order to justify the suppression factors given in IV , we estimate the decay rate of a Goldstone particle \theta of energy \omega into a pair of asymptotic fermions of energy \omega_{1} and \omega_{2} and corresponding momenta |p_{1}|=\omega_{1} and |p_{2}|=\omega_{2} . At tree level, the total decay rate reads: \Gamma=\frac{1}{T}\int\frac{dp_{1}}{2\omega_{1}}\frac{dp_{2}}{2\omega_{2}}\frac{|\bra{\theta}g\int d^{2}x\theta\bar{\psi}\psi\ket{\bar{\psi}\psi}|^{2}}{\bra{\theta}\ket{\theta}}\,. (90) The Goldstone field is localized in the bubble of size R . Thus, the modes of the Goldstone particles populating the bubble are normalized by a factor 1/\sqrt{\omega R} . Let us denote by \theta_{\omega} the mode function that corresponds to a Goldstone particle with frequency \omega . On the other hand, the fermionic field in the background of the bubble is expanded in a continuum of modes. Again, let us denote by \psi_{i} the mode that is associated with an asymptotic fermion of energy \omega_{i} . Note that if we were in an empty space, the process would not be kinematically allowed by momentum conservation. We will show below that this constraint gets relaxed in the presence of the bubble, and a larger kinematical window is allowed. Since the fermion is massive inside the bubble but massless outside, for energies smaller than the fermionic mass the modes will resemble a plane wave outside the bubble, and will be exponentially decaying inside with the penetration length l_{p}\sim 1/m_{\psi} . Indeed, inside the bubble the modes behave like \psi_{i}(x)\sim\exp\left(ix\sqrt{\omega^{2}-m_{\psi}^{2}}\right) . Notice that, since the modes are oscillatory outside of the bubble, they are normalized with the usual continuum normalization. Namely, for each of them we have <\psi_{i}|\psi_{i}>=(2\pi)\omega_{i}\delta(0)=(2\pi)\omega_{i}V , where V is the volume of the full space, and it should be understood as \frac{1}{V}\sim dp . When the energy \omega exceeds the fermionic mass induced by the bubble, the fermion starts propagating on-shell also inside the bubble, with an effective momentum of order k\sim\sqrt{\omega^{2}-m_{\psi}^{2}} . Given the premises, we can estimate the decay rate of a Goldstone into a fermion pair. Summing over polarizations, we obtain: \Gamma\sim g^{2}\frac{1}{R\omega}\int\frac{dp}{\omega_{1}}\frac{dq}{\omega_{2}}\left|\int dx\theta^{*}_{\omega}\bar{\psi}_{1}\psi_{2}\right|^{2}\omega_{1}\omega_{2}\delta(\omega-\sum_{i}\omega_{i})\,. (91) The wavefunction overlap factor can be estimated depending on the regimes. In the thin-wall limit, R\omega\ll 1 , when m_{\psi}\gg\omega , the integral is approximately independent of the value of the asymptotic momenta \displaystyle\left|\int dx\theta^{*}_{\omega}\bar{\psi}_{1}\psi_{2}\right|^{2}\sim \displaystyle\left|\int_{0}^{R}dxe^{i\omega x}e^{-2m_{\psi}x}\right|^{2} (92) \displaystyle\sim \displaystyle\begin{cases}\frac{1}{m^{2}_{\psi}}&m_{\psi}\gg 1/R\\ R^{2}&m_{\psi}\ll 1/R\,.\end{cases} (93) The \omega_{i} -dependence of the wavefunction-overlap factor arises in the form of the \omega/m_{\psi} -corrections, which are contributing to the phase space integral. Energy conservation ensures that they remain suppressed compared to the leading-order contribution. When \omega\geq m_{\psi} , the fermions can be produced on-shell. The wavefunction overlap again is: \left|\int dx\theta^{*}_{\omega}\bar{\psi}_{1}\psi_{2}\right|^{2}\sim\begin{cases}R^{2}&\omega\ll 1/R\\ \frac{1}{\omega^{2}}\sim R^{2}&\omega\sim 1/R\,.\end{cases} (94) Thus, the rate for a single particle scattering is of the order: \Gamma\sim\begin{cases}\frac{g^{2}}{\omega}\frac{1}{m_{\psi}^{2}}\frac{\omega}{R}&m_{\psi}\gg 1/R\\ \frac{g^{2}}{\omega}(\omega R)&m_{\psi}\ll 1/R\,.\end{cases} (95) Taking into account the occupation number of Goldstone particles, this leads to the results of IV . Appendix C Modes of a bubble with Gaussian profile It can be instructive to consider modes of a bubble with the Gaussian profile \phi_{b}^{2}(x)=e^{-(x\omega)^{2}/2} . Although such a bubble is not a solution of the presented model, it can be viewed as an auxiliary construct for gaining an intuition about the Goldstone mode structure, which is exactly solvable. The equation for the mode functions ( 34 ) takes the form Dvali and Vilenkin 2003 , \Theta_{k}^{\prime\prime}+x\omega^{2}\,\Theta_{k}^{\prime}+k^{2}\Theta_{k}=0\,. (96) which is solved by Hermite polynomials H_{n}(\omega x/\sqrt{2}) with k=\omega\sqrt{n} . Also, notice that in this case all the modes would be localized. Of course, in the present case the bubble profile is not a Gaussian, but this gives a qualitative idea about the structure of lowest lying modes. Notice that there exists an upper bound k^{2}