Critical Schwinger pair production

Holger Gies1, 2, ∗ and Greger Torgrimsson3, †

1
Theoretisch-Physikalisches Institut, Abbe Center of Photonics,
Friedrich-Schiller-Universität Jena, Max-Wien-Platz 1, D-07743 Jena, Germany
2
Helmholtz-Institut Jena, Fröbelstieg 3, D-07743 Jena, Germany
3
Department of Applied Physics, Chalmers University of Technology, SE-41296 Gothenburg, Sweden
We investigate Schwinger pair production in spatially inhomogeneous electric backgrounds. A
critical point for the onset of pair production can be approached by fields that marginally provide
sufficient electrostatic energy for an off-shell long-range electron-positron fluctuation to become a
real pair. Close to this critical point, we observe features of universality which are analogous to
continuous phase transitions in critical phenomena with the pair-production rate serving as an order
parameter: electric backgrounds can be subdivided into universality classes and the onset of pair
arXiv:1507.07802v1 [hep-ph] 28 Jul 2015

production exhibits characteristic scaling laws. An appropriate design of the electric background
field can interpolate between power-law scaling, essential BKT-type scaling and a power-law scaling
with log corrections. The corresponding critical exponents only depend on the large-scale features of
the electric background, whereas the microscopic details of the background play the role of irrelevant
perturbations not affecting criticality.

INTRODUCTION cal interpretation – can be viewed as a classical deter-

ministic field equation. In fact, the first observation of
Universality is an overarching concept in physics, sig- this phenomenon relied on this formulation [5], and so
nifying the independence of general gross properties of do many more modern approaches at least indirectly [6–
a physical system of the details of its microscopic re- 11]. On the other hand Schwinger pair production is
alizations. Most prominently, critical phenomena near also encoded in the photon correlation functions derived
a continuous phase transition reveal a remarkably high from the full functional integral of QED, as seen from
degree of universality, such that different systems con- the derivations of Euler, Heisenberg [12], and Schwinger
sisting microscopically of rather different building blocks [13]. Again many variants of this fluctuation-based de-
exhibit quantitatively identical long-range behavior near scriptions exist [14–16]. The fact that both descriptions
the phase transition [1]. The quantification of universal- are equivalent is a manifestation of the optical theorem
ity by means of fixed points is one of the great successes of (for a recent discussion in the context of pair production,
the renormalization group that provides a map from the see [17, 18]).
microscopic details to the effective long-range properties In this work, we use the worldline formalism with back-
[2]. ground fields [19], as this method makes universality in
the language of fluctuations of electrons in spacetime
As a consequence, critical systems can be associated
most transparent. We use the imaginary part of the
with universality classes which are characterized by only
QED effective action Γ as the order parameter for the
a few properties such as the symmetries of the order pa-
onset of criticality. It is related to the probability of vac-
rameter, the dimensionality, and the number and type of
uum decay via P = 1 − exp(−2 Im Γ[E]) in the presence
long-range degrees of freedom. It is therefore not surpris-
of an electric field E; to lowest order, it is also related to
ing that universality and a notion of criticality can also
the pair production rate [20, 21]. The seminal Schwinger
be found beyond the realm of statistical physics. For
formula
instance, the onset of black-hole formation shows a sur-
prising insensitivity to the initial data. Generically, the ∞
(eE)2 ! 1
"
πm2
#
black hole mass as a function of a single control param- Im Γ = V4 exp −n (1)
8π 3 n=1 n2 eE
eter parametrizing the initial data scales according to a
power-law with the universal Choptuik exponent [3, 4].
exhibits no signature of criticality, as it assumes the pres-
Whereas universality in statistical physics is typically as-
ence of an electric field being constant all throughout
sociated with the presence of fluctuations on all scales,
space and time. By contrast, a critical point can arise for
the example of gravitational collapse is observed in a
spatially inhomogeneous fields, as can be read off from
purely classical deterministic setting.
Nikishov’s exact solution for the electric field with the lo-
In the present work, we identify for the first time as- calized Sauter-profile E(x) = Esech2 kx of inverse width
pects of universality in the phenomenon of Schwinger pair k [21]. The order parameter for pair production Im Γ for
production in quantum electrodynamics (QED). This such a spatial profile drops to zero at
sets a dual example as the phenomenon of pair produc-
tion in strong external fields can be derived from the 2eE ! km
$
Dirac equation which – despite its quantum mechani- e dxE(x) = = 2m ⇒ γcr := = 1. (2)
k eE
 2

This equation has a simple meaning: the work done by spinor QED consists only in an irrelevant prefactor [25])
the electric field on a particle of charge e propagating $ ∞
ds −im2 s
$ " 2 #
along the whole real axis has to be equal to the rest mass i s dσ ẋ4 −eA·ẋ
!
Γ[A] = − e Dxe 0 ,
of the particle–anti-particle pair to be created. Trans- 0 s x(s)=x(0)
lated into the language of fluctuations, a virtual pair cre- (3)
ated at some spacetime point in such a field with adia- where the path integral can be interpreted as an average
baticity (Keldish) parameter γ = (km/eE) < γcr ≡ 1 can over all trajectories of electron fluctuations within the
become real if particle and anti-particle separate from one background field A. Though the electron mass m explic-
another sufficiently far to acquire enough electrostatic en- itly sets a scale, effectively constraining the (proper-)time
ergy. In all cases discussed below, the electrostatic energy s available for the fluctuations, the free path integral has
becoming equal to 2m, or γ → 1, always characterizes the a Gaußian velocity distribution such that the ensemble
onset of critical Schwinger pair production, and a scaling contains paths of arbitrary length scale [23]. This is the
? origin of universality for localized fields, as the near crit-
behavior Im Γ ∼ (1 − γ 2 )β with some critical exponent
β seems already suggestive at this stage. ical regime is dominated by the trajectories of largest
relevant extent which become less and less sensitive to
the microscopic details of the background field.
An important difference to standard critical phenom-
In the following, we study universality in the weak-field
ena of the type mentioned in the beginning is the oc-
regime,
currence of an explicit finite mass scale: the electron
mass. While universality arising near continuous phase " #2
eE
transitions is related to a diverging correlation length, ≪ 1 − γ 2 ≪ 1. (4)
i.e., long-range interactions mediated by an excitation m2
becoming exactly massless at the critical point, the elec- Although this prevents us from going all the way to γ =
tron mass remains as a finite scale in QED. This prevents 1, it is experimentally relevant given the large value of
us from associating the critical point with the notion of 2
the critical field strength Ec = me . This is precisely
scale invariance and self-similarity in a straightforward
the regime, where the semiclassical approximation of the
way. We find that this leads to a reduced degree of uni-
path integral as well as the propertime integral in Eq. (3)
versality, implying that critical pair production will not
become exact. In this semiclassical critical limit, the
be characterized by a universal scaling law or exponent,
path integral is dominated by the stationary points of the
but rather by a set of scaling laws for different large-
worldline action: the worldline instantons [24–30] which
scale properties of the spatial electric field profile. Still a
in general can be complex stationary paths [22]. Up to
rather large degree of universality, i.e., independence of
finite prefactors, the order parameter for pair production
the microscopic profile details, remains, such that electric
near semiclassical criticality scales as [26]
fields fall into universality classes of field profiles.
% 2
&
exp − πm eE g(γ 2
) d
In the present work, we confine ourselves to simple Im Γ ∼ ' , (. . . )′ ≡ 2 (. . . ), (5)
unidirectional electric fields that vary only in one spa- 2 ′
(γ g) (γ g) 2 ′′ dγ
tial coordinate, which also specifies the direction of the
field. More precisely, we assume that the x component of where the field dependence is contained in a single func-
the electric field can be written as E(x) = Ef ′ (u), where tion related to the worldline instanton action
the potential function f is antisymmetric, monotonic and $uγ
normalized such that max f = 1, and u = kx is a dimen- 2 1 4 '
g(γ ) = 2 du γ 2 − f 2 . (6)
sionless coordinate with 1/k being a suitable length scale γ π
0
of the spatial profile. With this restricted class of fields
we avoid pathological cases where large microscopic de- Here, ±uγ correspond to the semiclassical turning points
tails could dominate the pair production process. The defined by f (uγ ) = γ (because of the anticipated anti-
latter type of fields would require a case by case study symmetry of f (u), it suffices to consider u > 0 here and in
along the lines of fields with compact support included the following). Heuristically, these turning points corre-
below, possibly accompanied by interference effects [22]. spond to those points, where a separated virtual pair has
Still, the present class of fields is sufficiently general to acquired sufficient electrostatic energy to become real.
illustrate aspects of universality and gives access to a va- Eq. (5) has the standard semiclassical form of an expo-
riety of interesting universality classes. nential tunneling amplitude arising from the action along
a classical path, and a prefactor from the fluctuations
We begin with the worldline representation of the ef- about the classical path. The order parameter Im Γ van-
fective action of scalar QED in an external field [19] (for ishes if the prefactor vanishes, i.e., g ′ , g ′′ diverge, or if the
the following discussion of universality, the difference to exponent ∼ g diverges.
 3

Universality becomes already apparent from the de-

pendence of g on the potential function f in Eq. (6):
any field strength profile that leads to the same divergence
structure of g or its derivatives for γ → 1 belongs to the
same universality class. In the limit γ → 1, the turning
point uγ approaches the point u0 where the electric field
vanishes and f attains its maximum

f (uγ ) = γ → 1 = max f =: f (u0 ). (7)

Substituting sin θ = f (u)/γ, we find

$π/2
2 2 2 ′ 1
(γ g(γ )) = dθ ′ , (8) FIG. 1. Various examples for critical field profiles with ex-
π f ponent β = 54 . The onset of criticality is determined by the
0
asymptotic behavior (exponential in these cases). The critical
demonstrating that the divergence of g ′ comes from the scaling law Eq. (9) is independent of the local details of the
field profiles.
region close to zero field strength and the maximum of the
potential, f ′ → 0. Actually, while g ′′ always diverges, g ′
can be finite for certain compact fields (see below). Fur- insertion into Eq. (5) – produces a field-dependent power
ther, for fields vanishing asymptotically, u0 → ∞, also g in addition to β = 2,
can diverge; otherwise, e.g., for fields with compact sup- " √
m2
#
port, g is finite for regular fields and the tunneling am- 2 2 1+ c eE
Im Γ ∼ (1 − γ ) . (11)
plitude (i.e. the exponent in Eq. (5)) cannot contribute
eE
to criticality. Since m 2 ≪ 1, cf. Eq. (4), this field-dependent part
The resulting scaling laws can analytically be extracted dominates the exponent, indicating the approach to ex-
by expanding f near the leading-order divergence of 1/f ′ . ponential scaling.
Let us consider several paradigmatic examples, starting (III) The latter becomes manifest for a field decaying
with localized fields that decay asymptotically with a with 1 < p < 3, since the instanton action ∼ g diverges
c as a power near criticality, resulting in the scaling law
power, E(x) ∼ E (kx) p as x → ∞, with some constant
c 1
c. Then, f ≈ 1 − p−1 up−1 and u0 → ∞. Depending on "
πm2 C
#
the power p, three different cases occur: (I) for p > 3, the Im Γ ∼ (1 − γ 2 )β exp − , (12)
eE (1 − γ 2 )λ
function g stays finite and the scaling law arises purely
from the fluctuation prefactor, yielding a standard power- where the essential exponent λ and the constant C,
law " p
# p−1
3−p 2 2c %3 3 − p &
5p + 1 λ= , C= B , ,
Im Γ ∼ (1 − γ 2 )β , β= . (9) 2(p − 1) πc p − 1 2 2(p − 1)
4(p − 1) (13)
are both universal. (For p < 5/3, the exponent in
We emphasize that all field profiles with the same power- Eq. (12) can acquire universal subleading singularities
law decay exhibit the same universal critical scaling in- e.g. 1/(1 − γ 2 )λ−1 or ln(1 − γ 2 ).) In critical phenom-
dependently of the details of the profile at finite x (at ena, a scaling of this type is known as essential scaling
least within the class of fields specified above). Equa- or BKT (or Miransky) scaling [31]. It is known to occur
tion (9) also includes exponentially decaying fields: in in a wide range of systems, in particular those exhibiting
the limit p → ∞, we discover a unique exponent β = 54 . a transition from a conformal to a non-conformal phase
This agrees, for instance, with the exact result [21] for [32]. While our scaling law includes the BKT-scaling
the sech2 kx profile which in the regime (4) reads law with exponent λ = 12 for an electric field decaying
#3/2 with power p = 2, any essential exponent λ > 0 can be
L 2 T m3
"
eE 2πm2 realized for appropriate decay powers p. Equation (12)
Im Γ = (1 − γ 2 )5/4 e− eE . (10)
2(2π)3 m2 also has a universal powerlaw prefactor which is reminis-
cent to the many-flavor phase transition in gauge theories
We emphasize that the exponent β = 45 also holds for [33]. We also observe that λ diverges for p → 1 where the
other exponentially localized fields, different examples electrostatic energy receives dominant contributions from
are shown in Fig. 1. long-range fluctuations. Essential scaling of critical pair
(II) The powerlaw decay p = 3 is special, since the production hence is obviously related to a dominance of
function g itself diverges logarithmically which – upon electron-positron fluctuations at the largest length scales.
 4

Let us now turn to electric fields of compact support agreement with the Euclidean worldline picture, since the
in x direction. Within the class of fields considered contributions from large-scale fluctuations (large proper-
here, this implies that the potential function f (u) at- times) are suppressed by the electron mass scale.
tains its maximum at a finite value u0 . Correspondingly, The similarity of critical Schwinger pair production to
E(x) = 0 for |x| > u0 /k. The worldline action (6) can- critical phenomena discovered in this work appears to call
not become singular in this case, so the scaling law is for a renormalization group description. It is conceivable
solely determined by the fluctuation prefactor and thus that the critical point corresponds to a fixed point of a
by the way in which f ′ (u) approaches zero for u → u0 , suitable coarse-graining procedure involving the world-
cf. (8). Let us assume that the electric field drops to zero lines, the background field or both. Such a description
as E(x) ∼ (u0 /k − x)n . For n > 1, the order parameter would be rewarding as it could give access to potential
satisfies power-law scaling (9) with exponent further aspects of criticality such as (hyper-)scaling rela-
5n − 1 tions.
β= . (14) Our results can straightforwardly be generalized to
4(n + 1)
field profiles with only asymptotic symmetry as well as
It is interesting to see that the universal exponent for to an arbitrary number of translation invariant transver-
exponential decay β = 54 is rediscovered in the limit n → sal directions (y, z, t in the present work). As the latter
∞. Note also that (14) can be obtained by replacing only influences the propertime integrand, 1/s → 1/s 2 ,
d−2

p → −n in (9). the order parameter receives additional scaling factors ac-

Another special case is n = 1, where we find a logarith- 4−d
cording to Im Γd ∼ (γ 2 g)′ 2 Im Γ4 with corresponding
mic divergence in the prefactor, g ′ ∼ − ln(1 − γ 2 ), such consequences for the scaling exponents.
that the scaling law receives log corrections
From the underlying picture in terms of virtual fluc-
(1 − γ 2 )1/2 tuations needing to acquire sufficient energy to become
Im Γ ∼ . (15) real, we expect that our results analogously persist also
− ln(1 − γ 2 )
for static fields that are localized in more than one space
Again, this has an analog in statistical physics, as log- dimension, even though the analysis can become rather
corrections are known to arise in cases where marginal involved for non-unidirectional fields [27]. For time-
operators contribute to criticality [34], such as in the 2d dependencies slower than the Compton scale, the exis-
4-state Potts model [35]. For n < 1 only g ′′ diverges and tence of a critical point γcr ≃ 1 has been observed in
we again find power-law scaling (9) with [37, 38]. We expect though a qualitative change for fields
3n + 1 varying rapidly in time. Indeed, there are no critical
β= . (16) points for fields depending on lightfront time t+x [39, 40].
4(n + 1)
As soon as the fields vary in time, pair production does
In the limit n → 0, the electric field becomes step-like not have to rely only on (instantaneous) electrostatic en-
with exponent β = 41 . ergy, but can also be supported by finite (multi-)photon
The diversity of universality classes given above can be energies of the varying field. Hence, we expect the crit-
covered in a unified description using an implicit defini- ical point in the spatial adiabaticity parameter to shift
tion of the electric field in terms of differential equations. to values larger than γcr = 1 for increasing time vari-
One such possible definition is ations. This claim is supported by a recent paper [41]
which shows that for electric fields depending on a coor-
f ′ = (1 − f 2 )b , (17) dinate q that interpolates continuously between x, t + x
and t, the critical point increases from γcr = 1 at q = x to
with the implicit solution u = f 2 F1 (1/2, b, 3/2, f 2) in
infinity at q = t+x (i.e. effectively vanishes); for timelike
terms of a hypergeometric function. The case, b > 1 cov-
q there is no critical point.
ers all examples of asymptotically vanishing fields given
above with power p = b/(b − 1), and b < 1 corresponds to For rapidly varying fields, pair production via multi-
fields with compact support with n = b/(1−b). The cases photon effects may remove any singularity associated
b = 1/2, 1 and 3/2 correspond to to the fields E ∼ cos kx, with criticality. The would-be critical point due to the
sech2 kx and (1 + kx2 )−3/2 , respectively, studied explic- spatial profile may then still be visible as a cross-over in
itly in [26]. We have g = 2 F1 (1/2, b, 2, γ 2), and thus the production rate.
g ′′ ∼ (1 − γ 2 )−(b+1/2) , cf. [36], which agrees with all the We emphasize that our results for universality hold
different scalings above. in the regime defined by Eq. (4). We expect analo-
This unified analysis also verifies the general trend of gous features also in the deeply critical regime where
( eE )2
the relation between field profile and scaling: field profiles 1 − γ2 ≪ m 2 ≪ 1, even though the values of the
which are more spread out exhibit a steeper scaling. So, exponents might differ. For instance, the exact solu-
the order parameter Im Γ vanishes faster for an asymp- tion for the sech2 kx case scales differently in this regime,
totically decaying field than for a compact field. This is in Im Γ ∼ (1 − γ 2 )3 [21]. Determining the degree of uni-
 5

versality in this regime remains an interesting problem. A. Huet, S. P. Kim and C. Schubert, Phys. Rev. D 90,
As the electron mass scale is less dominant, universality no. 12, 125033 (2014) [arXiv:1411.3074 [hep-th]]; A. Otto,
could even be substantially enhanced. D. Seipt, D. Blaschke, B. Kämpfer and S. A. Smolyansky,
Phys. Lett. B 740, 335 (2015) [arXiv:1412.0890 [hep-ph]].
Radiative corrections will also take a subleading quan- [8] S. P. Kim and D. N. Page, Phys. Rev. D 65, 105002
titative influence on our results. E.g., the two-loop cor- (2002) [hep-th/0005078].
rection includes mass-shift effects [42, 43]; a resummation [9] I. Bialynicki-Birula, P. Gornicki and J. Rafelski, Phys.
could also account for a production into a positronium Rev. D 44, 1825 (1991); F. Hebenstreit, R. Alkofer
bound state. As these effects modify the invariant mass and H. Gies, Phys. Rev. D 82, 105026 (2010)
of the final state, they may primarily lead to a devia- [arXiv:1007.1099 [hep-ph]].
tion of the critical point γcr ≃ 1 but could preserve the [10] M. Ruf, G. R. Mocken, C. Muller, K. Z. Hatsagortsyan
and C. H. Keitel, Phys. Rev. Lett. 102, 080402 (2009)
universal critical exponents. [arXiv:0810.4047 [physics.atom-ph]]; T. Heinzl, A. Ilder-
In summary, we have discovered an analogy between ton and M. Marklund, Phys. Lett. B 692, 250 (2010)
Schwinger pair production and continuous phase tran- [arXiv:1002.4018 [hep-ph]].
sitions. This analogy is quantitatively manifest in uni- [11] F. Hebenstreit, J. Berges and D. Gelfand, Phys. Rev. D
versal scaling laws for the onset of pair production in 87, no. 10, 105006 (2013) [arXiv:1302.5537 [hep-ph]].
spatially inhomogeneous electric backgrounds. The scal- [12] W. Heisenberg and H. Euler, Z. Phys. 98 (1936) 714
[physics/0605038].
ing laws show a high degree of universality as the cor-
[13] J. S. Schwinger, Phys. Rev. 82 (1951) 664.
responding critical exponents only depend on the large- [14] W. Dittrich and M. Reuter, Lect. Notes Phys. 220, 1
scale properties of the background (for monotonic poten- (1985); V. I. Ritus, IN *GINZBURG, V.L. (ED.): IS-
tials) but become insensitive to the microscopic details. SUES IN INTENSE-FIELD QUANTUM ELECTRODY-
Hence, localized electric backgrounds fall into universal- NAMICS* 1-62 (1989); W. Dittrich and H. Gies, Springer
ity classes each giving rise to a characteristic scaling law. Tracts Mod. Phys. 166, 1 (2000); G. V. Dunne, In *Shif-
As a particularly fascinating aspect, we discovered uni- man, M. (ed.) et al.: From fields to strings, vol. 1* 445-
522 [hep-th/0406216].
versality classes covering essentially all types of scaling
[15] G. V. Dunne, H. Gies and R. Schutzhold, Phys. Rev. D
laws familiar from continuous phase transitions. 80, 111301 (2009) [arXiv:0908.0948 [hep-ph]]; V. N. Baier
We thank Gerald Dunne, Anton Ilderton, Felix Karb- and V. M. Katkov, Phys. Lett. A 374, 2201 (2010)
stein, and Ralf Schützhold for interesting discussions. [arXiv:0912.5250 [hep-ph]].
G.T. thanks TPI, FSU Jena, and HI Jena for hospi- [16] H. Gies and K. Klingmuller, Phys. Rev. D 72, 065001
tality during a research visit. We acknowledge support (2005) [hep-ph/0505099].
[17] V. Dinu, T. Heinzl, A. Ilderton, M. Marklund and
by the DFG under grants No. GRK1523/2, and SFB-
G. Torgrimsson, Phys. Rev. D 89 (2014) 12, 125003
TR18 (H.G.), and the Swedish Research Council, con- [arXiv:1312.6419 [hep-ph]].
tract 2011-4221 (P.I.: A. Ilderton), (G.T). [18] S. Meuren, K. Z. Hatsagortsyan, C. H. Keitel and
A. Di Piazza, Phys. Rev. D 91, no. 1, 013009 (2015)
[arXiv:1406.7235 [hep-ph]].
[19] M. G. Schmidt and C. Schubert, Phys. Lett. B 318,
438 (1993) [hep-th/9309055]; M. Reuter, M. G. Schmidt
∗
[email protected] and C. Schubert, Annals Phys. 259, 313 (1997) [hep-
†
[email protected] th/9610191]; C. Schubert, Phys. Rept. 355, 73 (2001)
[1] L. P. Kadanoff, In *Varenna 1970, Proceedings, Critical [hep-th/0101036].
Phenomena*, New York 1971, 100-117; J. Zinn-Justin, [20] T. D. Cohen and D. A. McGady, Phys. Rev. D 78, 036008
Int. Ser. Monogr. Phys. 113, 1 (2002). (2008) [arXiv:0807.1117 [hep-ph]].
[2] F. J. Wegner and A. Houghton, Phys. Rev. A 8, 401 [21] A. I. Nikishov, Nucl. Phys. B 21, 346 (1970).
(1973); K. G. Wilson and J. B. Kogut, Phys. Rept. 12, [22] C. K. Dumlu and G. V. Dunne, Phys. Rev. D 84 (2011)
75 (1974); J. Polchinski, Nucl. Phys. B 231, 269 (1984); 125023 [arXiv:1110.1657 [hep-th]].
C. Wetterich, Phys. Lett. B 301, 90 (1993); [23] H. Gies and K. Langfeld, Nucl. Phys. B 613, 353 (2001)
[3] M. W. Choptuik, Phys. Rev. Lett. 70, 9 (1993). [hep-ph/0102185]; Int. J. Mod. Phys. A 17, 966 (2002)
[4] C. Gundlach and J. M. Martin-Garcia, Living Rev. Rel. [hep-ph/0112198].
10, 5 (2007) [arXiv:0711.4620 [gr-qc]]. [24] I. K. Affleck, O. Alvarez and N. S. Manton, Nucl. Phys.
[5] F. Sauter, Z. Phys. 69 (1931) 742. B 197 (1982) 509.
[6] E. Brezin and C. Itzykson, Phys. Rev. D 2, 1191 (1970); [25] G. V. Dunne and C. Schubert, Phys. Rev. D 72 (2005)
V. S. Popov and M. S. Marinov, Yad. Fiz. 16, 809 (1972). 105004 [hep-th/0507174].
[7] S. P. Gavrilov and D. M. Gitman, Phys. Rev. D 53, 7162 [26] G. V. Dunne, Q. h. Wang, H. Gies and C. Schubert, Phys.
(1996) [hep-th/9603152]; S. A. Smolyansky, G. Ropke, Rev. D 73 (2006) 065028 [hep-th/0602176].
S. M. Schmidt, D. Blaschke, V. D. Toneev and A. V. Pro- [27] G. V. Dunne and Q. h. Wang, Phys. Rev. D 74 (2006)
zorkevich, hep-ph/9712377; Y. Kluger, E. Mottola and 065015 [hep-th/0608020].
J. M. Eisenberg, Phys. Rev. D 58, 125015 (1998) [hep- [28] D. D. Dietrich and G. V. Dunne, J. Phys. A 40, F825
ph/9803372]; S. M. Schmidt, D. Blaschke, G. Ropke, (2007) [arXiv:0706.4006 [hep-th]].
S. A. Smolyansky, A. V. Prozorkevich and V. D. Toneev, [29] J. Gordon and G. W. Semenoff, arXiv:1407.0987 [hep-th].
Int. J. Mod. Phys. E 7, 709 (1998) [hep-ph/9809227]; [30] E. Strobel and S. S. Xue, Nucl. Phys. B 886 (2014) 1153
 6

[arXiv:1312.3261 [hep-th]]. ume 3, ed. by H. Yurij, World Scientific Publishing, pp.

[31] V. L. Berezinsky, Sov. Phys. JETP 32, 493 (1971); 1-46 (2013) [arXiv:1205.4252].
J. M. Kosterlitz and D. J. Thouless, J. Phys. C 6, 1181 [35] L. Cardy, M. Nauenberg, D.J. Scalapino, Phys. Rev. B
(1973); V. A. Miransky and K. Yamawaki, Mod. Phys. 22, 2560 (1980).
Lett. A 4, 129 (1989). [36] NIST Digital Library of Mathematical Functions,
[32] V. A. Miransky and K. Yamawaki, Phys. Rev. D 55, 5051 http://dlmf.nist.gov/.
(1997) [Phys. Rev. D 56, 3768 (1997)] [hep-th/9611142]; [37] F. Hebenstreit, R. Alkofer and H. Gies, Phys. Rev. Lett.
H. Gies and J. Jaeckel, Eur. Phys. J. C 46, 433 (2006) 107, 180403 (2011) [arXiv:1106.6175 [hep-ph]].
[hep-ph/0507171]; D. B. Kaplan, J. W. Lee, D. T. Son [38] C. Schneider and R. Schutzhold, arXiv:1407.3584 [hep-
and M. A. Stephanov, Phys. Rev. D 80 (2009) 125005 th]; M. F. Linder, C. Schneider, J. Sicking, N. Szpak and
[arXiv:0905.4752 [hep-th]]. R. Schutzhold, arXiv:1505.05685 [hep-th].
[33] J. Braun and H. Gies, JHEP 1005 (2010) 060 [39] T. N. Tomaras, N. C. Tsamis and R. P. Woodard, JHEP
[arXiv:0912.4168 [hep-ph]]; J. Braun, C. S. Fischer and 0111 (2001) 008 [hep-th/0108090].
H. Gies, Phys. Rev. D 84 (2011) 034045 [arXiv:1012.4279 [40] A. Ilderton, JHEP 1409 (2014) 166 [arXiv:1406.1513
[hep-ph]]; M. P. Lombardo, K. Miura, T. J. Nunes da [hep-th]].
Silva and E. Pallante, Int. J. Mod. Phys. A 29, no. 25, [41] A. Ilderton, G. Torgrimsson and J. Wårdh,
1445007 (2014) [arXiv:1410.2036 [hep-lat]]; J. A. Bon- arXiv:1506.09186 [hep-th].
net and C. S. Fischer, Phys. Lett. B 718, 532 (2012) [42] V. I. Ritus, Sov. Phys. JETP 42, 774 (1975) [Pisma Zh.
[arXiv:1201.6139 [hep-ph]]. Eksp. Teor. Fiz. 69, 1517 (1975)]; Zh. Eksp. Teor. Fiz.
[34] F. J. Wegner, Phys. Rev. B 5, 4529 (1972); in 73, 807 (1977).
Phase Transitions and Critical Phenom- ena VI, ed. by [43] G. V. Dunne and C. Schubert, JHEP 0206, 042 (2002)
C. Domb and M. S. Green, Academic Press, London [hep-th/0205005].
(1976); R. Kenna, in Order, Disorder and Criticality: [44] G. Von Gersdorff and C. Wetterich, Phys. Rev. B 64
Advanced Problems of Phase Transition Theory - Vol- (2001) 054513 [hep-th/0008114].