The peaks of the correlation function in acoustic black holes

Abstract

Renaud Parentani was one of the leading figures in Quantum Field Theory in curved spacetime, in particular concerning its applications to Hawking-like radiation in analogue models. In this paper dedicated to him, we discus the features of the characteristic peaks appearing in the correlation functions in an acoustic black hole formed by a Bose-Einstein condensate, considered as signature of the presence of Hawking radiation in this system.

I Introduction

In 1981 Unruh unruh81 suggested that Hawking’s black hole (hereafter BH) radiation can have an analogue in a fluid whose flow undergoes a transition from a subsonic regime to a supersonic one. The locus where this happens is the so called ”sonic horizon””, since sound waves are trapped inside the supersonic region and cannot propagate upstream; they are trapped by the flow and dragged downstream. Unruh showed that in this situation one should expect an emission in the subsonic region of thermal phonons at a temperature proportional to the surface gravity of the sonic horizon exactly as predicted by Hawking for gravitational BHs hawking74 ; hawking75 .

In the following years many systems were proposed to experimentally detect this analogue Hawking radiation Barcelo:2005fc . The most promising appeared to be the ones constructed by Bose-Einstein condensates (BECs), since in this case one can arrange the experimental setup so that the associated Hawking temperature is expected to be just one order of magnitude smaller than the BEC background temperature ($100\ nK$). Nevertheless even this difference has so far prevented any direct detection of these thermal phonons.

In 2008 it was shown that a characteristic peak in the in-out density-density correlation function should appear paper1 ; cfrbf . The reason is that the Hawking effect involves a genuine pair creation process in which for each thermal phonon in the subsonic region there is a corresponding negative energy partner inside the horizon. This is the smoking gun of the Hawking effect. This peak was observed in a series of experiments performed by Steinhauer and his group jeff2016 ; jeff2019 ; jeff2021 . This represents up to today the best evidence of the presence of Hawking radiation in sonic BHs.

Renaud Parentani suggested that besides this main peak, two other minor peaks should appear in the density-density correlation function because of backscattering effects on the modes mp . While there is as yet no experimental evidence for these peaks, numerical calculations have confirmed their presence paper2 ; rpc ; paper2011 .

In this paper, dedicated to Renaud Parentani for his invaluable contributions to this field, we shall investigate, using the framework of Quantum Field Theory in curved space (see for example bd ; fu ; pt ), the features of the main and secondary peaks of the density-density correlation function for two flow profiles having the same asymptotic sound speed limits and horizon surface gravity, outlining similarities and differences.

II The setting

In a BEC the phase fluctuation $\hat{\theta}$ on top of the condensate under the hydrodynamical approximation obeys an equation which is formally identical to a wave equation for a massless scalar field propagating in a fictitious curved space-time described by the line element

$$ds^{2}=\frac{n}{mc}[-c^{2}dT^{2}+(d\vec{x}-\vec{v}dT)(d\vec{x}-\vec{v}dT)]\ ,$$ (1)

where $n$ is the condensate density, $c$ the local speed of sound, $\vec{v}$ the velocity field and $m$ the mass of a single atom. The wave equation reads

$$\Box\hat{\theta}=0\ ,$$ (2)

where $\Box\equiv\nabla_{\mu}\nabla^{\mu}$ is the covariant d’Alembertian calculated with the metric (1).

This system can be treated by using the methods of Quantum Field Theory in curved space-time. This has been done in a paper written in collaboration with R. Parentani paper2013 . Here we just outline the main points.

We shall consider for simplicity a stationary unidimensional flow directed along the $\hat{x}$ axis with a constant velocity $\vec{v}$, the density $n$ is also constant. By an appropriate rescaling of the phase operator $\hat{\theta}=\sqrt{\frac{mc}{n\hbar L_{\perp}^{2}}}\ {\hat{\theta}^{(2)}}$, where $L_{\perp}$ is the size of the transverse direction with $L_{\perp}\ll\frac{\hbar}{mc}$ so that excitations with transverse momenta are frozen, eq. (2) can be rewritten as

$$\left(\Box^{(2)}-V\right)\hat{\theta}^{(2)}=0\ ,$$ (3)

where $V$ is given by

$$V=-\frac{1}{2}\left(1-\frac{v^{2}}{c^{2}}\right)\frac{d^{2}c}{dx^{2}}+\left(\frac{1}{4c}-\frac{5v^{2}}{4c^{3}}\right)\left(\frac{dc}{dx}\right)^{2}\;$$ (4)

and $\Box^{(2)}$ is the two dimensional (2D) d’Alembertian associated with the 2D section of the line element (1), namely

$$ds^{2}=\frac{n}{m}\left[-\frac{c(x)^{2}-v^{2}}{c(x)}dt^{2}+\frac{c(x)}{c(x)^{2}-v^{2}}dx^{2}\right]\;.$$ (5)

Here we have introduced a “Schwarzschild” time $t$ such that

$$t=T-\int^{x}dy\frac{v}{c(y)^{2}-v^{2}}\;.$$ (6)

By considering the coordinate $x^{*}$ , given by

$$x^{*}=\int^{x}\frac{c(y)dy}{c(y)^{2}-v^{2}}\ ,$$ (7)

we can rewrite (3) in the form

$$\left(\frac{\partial^{2}}{\partial t^{2}}+\frac{\partial}{\partial x^{*2}}+V_{\rm eff}\right)\theta^{(2)}=0\ ,$$ (8)

with effective potential

$$V_{\rm eff}=\frac{c^{2}-v^{2}}{c}V.$$ (9)

The BEC flows along the $\hat{x}$ direction from right to left (i.e. $\vec{v}=-v_{0}\hat{x}$, with $v_{0}>0$). By varying $c(x)$ one can engineer the flow so that it is subsonic for $x>0$ ($R$ region) and supersonic for $x<0$ ($L$ region). $x=0$ is the sonic horizon. $c(x)$ can be further assumed to approach constant values asymptotically; namely $\lim_{x\to+\infty}c(x)=c_{R}>v_{0}$ and $\lim_{x\to-\infty}c(x)=c_{L}<v_{0}$. Note that the effective potential $V_{\rm eff}$ (9) vanishes asymptotically and on the horizon. The Penrose diagram of the spacetime is depicted in Fig. (1).

Figure 1: Penrose diagram of the spacetime of eq. (5). $H^{+}$ is the future horizon, $H^{-}$ the past one, $I^{\pm}_{R,L}$ are null infinities. $R$ is the region outside the horizon, $L$ is the one inside.

For further use we introduce the retarded ($u$) and advanced ($v$) null Eddington-Finkelstein like coordinates as

$$u = t-x^{*}\ ,$$ (10)

$$v = t+x^{*}\ ,$$ (11)

and the Kruskal one

$$U=\pm\frac{1}{\kappa}e^{-\kappa u}\ ,$$ (12)

where $\kappa$ is the surface gravity of the horizon

$$\kappa=\left.\frac{dc}{dx}\right|_{x=0}$$ (13)

and in eq. (12) the plus sign is for the $L$ region and the minus sign is for the $R$ region.

The quantum state of our field $\hat{\theta}^{(2)}$, as is well known, can be approximated at late times after the formation of the BH by the Unruh state $|U\rangle$ unruh76 . This corresponds to an expansion of the quantum operator as

$$\hat{\theta}^{(2)}=\int_{0}^{\infty}d\omega_{K}\left[\hat{a}_{K}(\omega_{K})u_{H}^{K}+h.c.\right]+\int_{0}^{\infty}d\omega\left[\hat{a}_{I}(\omega)u_{I}^{R}+h.c.\right]\ ,$$ (14)

where the form of the mode $u_{I}^{R}$ on $I_{R}^{-}$ is

$$u_{I}^{R}(\omega,x)=\frac{e^{-i\omega v}}{\sqrt{4\pi\omega}}\ ,$$ (15)

while that for the mode $u_{H}^{K}$ on $H^{-}$ is

$$u_{H}^{K}(\omega_{K},x)=\frac{e^{-i\omega_{K}U}}{\sqrt{4\pi\omega_{K}}}\ .$$ (16)

The Unruh state $|U\rangle$ is defined as

$$\hat{a}_{K}(\omega_{K})|U\rangle=0=\hat{a}_{I}(w)|U\rangle\$$ (17)

for every $\omega,\omega_{K}$. In Fig. (2) we illustrate these modes on the Cauchy surface $H^{-}\cup I_{R}^{-}$.

Figure 2: Modes for the Unruh state.

Note that while $u_{I}^{R}$ is a positive frequency mode with respect to Schwarzschild time, $u_{H}^{K}$ is positive frequency with respect to Kruskal time. The Unruh state describes a situation in which one has no incoming radiation on $H^{-}\cup I_{R}^{-}$ , while at $I_{R}^{+}$ one has a thermal flux at the Hawking temperature $\displaystyle{T_{H}=\frac{\hbar\kappa}{2\pi k_{B}}}$, where $k_{B}$ is Boltzmann’s constant.

It is convenient for the calculations of the 2-point function to express the modes $u_{H}^{K}$ in terms of modes (Boulware modes bo ) that on $H^{-}$ behave as

$$u_{H}^{R} = \frac{e^{-i\omega u}}{\sqrt{4\pi\omega}}\ ,x>0,$$ (18)

$$u_{H}^{L} = \frac{e^{i\omega u}}{\sqrt{4\pi\omega}}\ ,x<0\ .$$ (19)

Note the $+$ sign in front of the exponential in eq. (19). $u_{H}^{L}$ has negative frequency (while having positive norm), it is associated with the negative (Killing) energy partners. We have

$$u_{H}^{K}(\omega_{K},x)=\int_{0}^{\infty}d\omega\left[\alpha_{\omega_{K}\omega}^{R}u_{H}^{R}+\beta^{R}_{\omega_{K}\omega}u_{H}^{R*}\right]+R\leftrightarrow L\ .$$ (20)

The Bogoliubov coefficients are given in Ref. paper2013 and are summarised in Appendix A. Because of the presence of the effective potential $V_{\rm eff}$ (9), the incoming modes will be modified from their asymptotic form Eqs. (15), (18), (19) and backscattered. In Ref. paper1 , $V$ was neglected and the modes maintained their expression (15), (18), (19) throughout the space-time. In Figs. (3) - (5) we schematically describe the backscattering of each mode.

Figure 3: Backscattering of the $u_{I}^{R}$ modes.

Figure 4: Backscattering of the $u_{H}^{R}$ modes.

Figure 5: Backscattering of the $u_{H}^{L}$ modes.

In any case, since $V_{\rm eff}$ vanishes asymptotically, at future null infinity each incoming mode will be a linear combination of $e^{-i\omega v}$ and $e^{\pm i\omega u}$.

We will consider two profiles of the sound velocity. The first one has been proposed in a numerical simulation fully based on the Bogoliubov theory of a BEC cfrbf which confirmed the presence of a peak in the in-out correlation function as predicted by paper1 using only QFT in curved space methods. It is

$$c(x) = \sqrt{c_{L}^{2}+\frac{1}{2}(c_{R}^{2}-c_{L}^{2})\left[1+\frac{2}{\pi}\tan^{-1}\left(\frac{x+b}{\sigma_{v}}\right)\right]}\;,$$ (21)

$$b = \sigma_{v}\tan\left[\frac{\pi}{c_{R}^{2}-c_{L}^{2}}\left(v_{0}^{2}-\frac{1}{2}(c_{R}^{2}+c_{L}^{2})\right)\right]\;,$$ (22)

where $\sigma_{v}$ is an arbitrary positive constant. The horizon is at $x=0$. The surface gravity for this profile is

$$\kappa=\left.\frac{dc}{dx}\right|_{x=0}=\frac{1}{2\pi v_{0}\sigma_{v}}(c_{R}^{2}-c_{L}^{2})\sin^{2}\left[\frac{\pi(c_{R}^{2}-v_{0}^{2})}{c_{R}^{2}-c_{L}^{2}}\right]\ .$$ (23)

We call this the “original” profile. In this case the equation for the modes eq. (3) has to be solved numerically. The explicit construction of the modes has been given in Ref. paper2013 , written in collaboration with Renaud Parentani, where all the details can be found. The second profile, which we call “analytical”, is

$$c(x)=\frac{c_{R}}{\sqrt{1+\left(\frac{c_{R}^{2}}{v_{0}^{2}}-1\right)\exp\left[-\frac{2\frac{c_{R}^{2}}{v_{0}^{2}}kx}{(\frac{c_{R}^{2}}{v_{0}^{2}}-1)}\right]}}\theta(x)+\frac{c_{L}}{\sqrt{1-\left(1-\frac{c_{L}^{2}}{v_{0}^{2}}\right)\exp\left[-\frac{2k\frac{c_{L}^{2}}{v_{0}^{2}}x}{\left(\frac{c_{L}^{2}}{v_{0}^{2}}-1\right)}\right]}}\theta(-x)\;,$$ (24)

where $x=0$ is the horizon and $k$ is a positive constant of dimension $L^{-1}$ and the corresponding surface gravity is

$$\kappa=\left.\frac{dc}{dx}\right|_{x=0}=k\,v_{0}\ .$$ (25)

The profile has been introduced in Ref. fba2016 . The advantage of this profile is that the modes can be computed analytically in terms of hypergeometric functions.

III Correlation functions

As mentioned in the Introduction, the only experimental support for the existence of Hawking-like radiation in an acoustic BH formed by a BEC comes from the observation of a peak appearing in the in-out (one point inside the horizon and the other outside) equal time density-density correlation function, in agreement with the theoretical prediction paper1 .

Defining the operator $\hat{n}$ as the quantum density fluctuation on top of the condensate, the density-density correlation function is

$$G_{2}\left(T,x;T^{\prime},x^{\prime}\right)=\langle U|\hat{n}(T,x)\hat{n}(T^{\prime},x^{\prime})|U\rangle\ .$$ (26)

In the hydrodynamical approximation we have Barcelo:2005fc

$$\hat{n}=\frac{\hbar n}{mc^{2}}\left[v_{0}\partial_{x}\hat{\theta}-\partial_{T}\hat{\theta}\right]$$ (27)

and one gets, in our setting,

$$G_{2}\left(T,x;T,x^{\prime}\right)=\frac{\hbar n}{2mL_{\perp}^{2}c^{2}(x)c^{2}(x^{\prime})}D\sqrt{c(x)c(x^{\prime})}\langle U|\left\{\{\hat{\theta^{(2)}}(t,x),\hat{\theta^{(2)}}(t^{\prime},x^{\prime})\right\}|U\rangle\ ,$$ (28)

where

$$D\equiv\partial_{T}\partial_{T^{\prime}}-v_{0}\partial_{x}\partial_{T^{\prime}}-v_{0}\partial_{T}\partial_{x^{\prime}}+v_{0}^{2}\partial_{x}\partial_{x^{\prime}}\ .$$ (29)

Using eqs. (15) and (20) and integrating over $\omega_{K}$ (see paper2013 ) the two-point function entering (28) can be written as paper1

$$\langle U|\left\{\hat{\theta}^{(2)}(t,x),\hat{\theta}^{(2)}(t^{\prime},x^{\prime})\right\}|U\rangle=I+J\ ,$$ (30)

where

$$I = \int_{0}^{\infty}d\omega\frac{1}{\sinh\left(\frac{\pi\omega}{\kappa}\right)}\left\{u^{L}_{H}(\omega,t,x)\,u^{R}_{H}(\omega,t^{\prime},x^{\prime})+u^{L*}_{H}(\omega,t,x)\,u^{R*}_{H}(\omega,t^{\prime},x^{\prime})\right.$$ (31)

$$\;\;\;\left.+u^{R}_{H}(\omega,t,x)\,u^{L}_{H}(\omega,t^{\prime},x^{\prime})+u^{R*}_{H}(\omega,t,x)\,u^{L*}_{H}(\omega,t^{\prime},x^{\prime})\right.$$

$$\;\;\left.+\cosh\left(\frac{\pi\omega}{\kappa}\right)\left[u^{L}_{H}(\omega,t,x)\,u^{L*}_{H}(\omega,t^{\prime},x^{\prime})+u^{L*}_{H}(\omega,t,x)\,u^{L}_{H}(\omega,t^{\prime},x^{\prime})\right.\right.$$

$$\;\;\;\left.\left.+u^{R}_{H}(\omega,t,x)\,u^{R*}_{H}(\omega,t^{\prime},x^{\prime})+u^{R*}_{H}(\omega,t,x)\,u^{R}_{H}(\omega,t^{\prime},x^{\prime})\right]\right\}\;,$$

$$J = \int_{0}^{\infty}d\omega\,\left[u^{R}_{I}(\omega,t,x)\,u^{{R}\,*}_{I}(\omega,t^{\prime},x^{\prime})+u^{{R}\,*}_{I}(\omega,t,x)\,u^{R}_{I}(\omega,t^{\prime},x^{\prime})\right]\;$$ (32)

and the relation between Schwarzschild time $t$ and $T$ is given by eq. (6).

If one neglects the effective potential in eq. (8), the modes maintain the form given by eqs. (15), (18), (19) throughout the entire space-time. In this case one can obtain an analytical expression for $G_{2}(T,x;T^{\prime},x^{\prime})$ which, taking the point $x$ in the asymptotic $R$ region where $c(x)\sim c_{R}$ and the point $x^{\prime}$ in the asymptotic $L$ region where $c(x)\sim c_{L}$, can be written as

$$G_{2}(T,x;T,x^{\prime}) = \frac{\hbar n}{2mL_{\perp}^{2}c_{R}^{3/2}c_{L}^{3/2}}\Big{\{}-\frac{1}{(c_{R}-v_{0})(v_{0}-c_{L})}\frac{\kappa^{2}}{\cosh^{2}\frac{\kappa}{2}(u-u^{\prime})}$$ (33)

$$+ \frac{1}{(c_{R}+v_{0})(c_{L}+v_{0})}\frac{1}{(v-v^{\prime})^{2}}\Big{\}}\ .$$

We see that this function has a negative peak at $u=u^{\prime}$ which corresponds (in the geometrical optics approximation) to the trajectory of the Hawking quanta ($u=const$) and its partner ($u^{\prime}=const$). Beside this no other structure is present. The peak is the one observed by Steinhauer’s group jeff2016 ; jeff2019 ; jeff2021 .

We see that the no-backscattering asymptotic correlation function (33) has the same form for all profiles having the same $c_{R},c_{L},v_{0}$ and surface gravity $\kappa$; in particular the height, width, and location of the peak are identical. We shall impose this to be the case for our two profiles introduced in the previous section. In Fig. (6) we have plotted the two profiles for $v_{0}=\frac{3}{4},c_{L}=\frac{1}{2},c_{R}=1,\sigma_{v}=8$ and numerically matched the two surface gravities.

Figure 6: Plot for the profiles of Eqs. (21) (blue curve), (24) (orange curve) arranging them to have the same surface gravity and asymptotic behaviour.

Even with all these parameters matched, there are noticeable differences in the comparison of the sound profiles. This leads to differences in the scattering of the modes as can be seen in Fig. (7), where the effective potential for the two profiles is plotted. The peaks of the potential appear to be higher and narrower for the original profile. All this has a significant signature in the correlation functions as we shall see.

Figure 7: Plot of the effective potential (9), (4) for the profiles of eqs. (21) (blue curve), (24) (orange curve) with the same asymptotic behaviour and surface gravity..

In Fig. (8) we have represented the correlation function Eq. (28) at equal time $T=T^{\prime}$ for the original and the analytical profile respectively.

Figure 8: Density-density correlation function for the two sound speed profiles. Left: Original profile (21). Right: Result using the analytical profile (24). Each figure is oriented so that the bottom left quadrant where both points are in the interior and the upper right quadrant where both points are in the exterior of the analog black hole.

This function is diverging when the points come together and thus the region $x=x^{\prime}$ is cutoff for this reason. In each figure one can clearly see the large negative peak when one point is in the interior and the other point is in the exterior region. This is the one predicted by the no-backscattering asymptotic expression (33). One can also see two much smaller peaks predicted by R. Parentani: a negative one when both points are inside the horizon and a positive one, when one point is inside the horizon and the other outside. These two secondary peaks exist because of the backscattering of the modes.

To see the differences between the two profiles, we have taken in Fig. (9) a slice at $x^{\prime}=-250$ of the in-out region ($x^{\prime}<0,x>0$) of Fig. (8).

Figure 9: The density-density correlation function at $x^{\prime}=-250$ when one point is in the interior and one point in in the exterior of the analog black hole. Blue dot: Original profile (21). Orange diamond: Analytic profile (24).

The large main peak is clearly visible but the peak height and location are offset for the two profiles. The main negative peak for the analytic profile appears slightly smaller and shifted to the left as compared to that of the original one. The opposite occurs for the smaller secondary peak as seen in Fig. (10), where we have magnified the scale to better appreciate this point.

Figure 10: A comparison of the secondary peak at $x^{\prime}=-250$ when one point is in the interior and one point in in the exterior of the analog black hole. Blue dot: Original profile (21). Orange diamond: Analytic profile (24).

Figure 11: Density-density correlation function for both points in the interior at a fixed ${x^{\prime}=400}$. Blue dot: Original profile (21). Orange diamond: Analytic profile (24).

More striking is the relative difference appearing in the negative peak in the in-in region, see Fig. (11).

Figure 12: A comparison of the main peak at $x^{\prime}=-250$ when one point is in the interior and one point in in the exterior of the analog black hole. Left: Blue(Solid): Curve for the analytic expression (33). Blue dot: Numeric data for original profile (21). Right: Orange (solid): Curve for the analytic expression (33). Orange diamond: Numeric result for analytic profile (24).

The backscattering also affects the main peak. In Fig. (12) a comparison is made for the two profiles with the no-backscattering approximation eq. (33). The differences are more significant for the original profile.

Finally, it is interesting to compare the ratios of the heights of each of the two minor peaks to that of the main peak. The ratio of the height of the positive minor peak with one point inside and one point outside the horizon to that of the main peak for the original profile is $0.0293$. For the analytical profile it is $0.0302$. The ratio of the height of the negative peak when both points are inside the horizon to the main peak for the original profile is $0.261$. For the analytical profile it is $0.191$. In both cases there is agreement in the first digit only, so differences in the profiles lead to relatively significant differences between the two profiles.

IV Scaling

There is a scaling related to the surface gravity $\kappa$ that works for both sound speed profiles used in this paper, (21) and (24). It is

$$\bar{\omega} = \frac{\omega}{\kappa}\;,$$ (34a)

$$\tau = \kappa t\;,$$ (34b)

$$\xi = \kappa x\;.$$ (34c)

It is easy to see that for this scaling both sound speed profiles, written in terms of $\xi$, are independent of $\kappa$.

In general, for any sound speed profile that, when written in terms of $\xi$, is independent of $\kappa$ one can substitute  (34) into the Boulware and Kruskal modes of (14) and show that they both scale as $\kappa^{-1/2}$. Using these results, one can show that the two-point function (30) is independent of $\kappa$. Then one finally has from (28) that

$$G_{2}(T,x;T^{\prime},x^{\prime})=\kappa^{2}\;G_{2}(\bar{T},\xi;\bar{T}^{\prime},\xi^{\prime})\;.$$ (35)

This means that the heights and depths of the correlation peaks are larger for larger values of $\kappa$. Since $x=\frac{\xi}{\kappa}$, the widths of the correlation peaks in terms of the space coordinate $x$ are narrower for larger values of $\kappa$. See Figs. (13, 14).

Figure 13: Scaling preset in the density-density correlation function: Left: Result using the original profile (21) with $\sigma_{v}=8$ with axis limits at $-400\leq x\leq 400$. Right: Result using the Original profile 21 with $\sigma_{v}=4$ with axis limits at $-200\leq x\leq 200$.

Figure 14: Comparison of the main peak in analytic and numerical results for with $\sigma_{v}=8$ and $\sigma_{v}=4$ along a slice at $x=-100$. Blue (solid ): Theoretical curve (33) for the original profile (21) with $\sigma_{v}=8$. Blue diamond: Numerical result for the original profile (21) with $\sigma_{v}=8$. Red (solid): Theoretical curve for original profile (21 with $\sigma_{v}=4$. Red square: Numerical result for the original profile (21) with $\sigma_{v}=4$.

V Conclusions

While awaiting the direct detection of the thermal phonons, the peaks in the density-density correlation function are the major tool to experimentally investigate the analogue of Hawking radiation in a sonic BH formed by a BEC. Of the three characteristic peaks only one, the main one predicted in paper1 , has being observed so far. The detection of the other much smaller peaks, whose existence was predicted by R. Parentani, represents the next challenge for the experimentalists in this field. In this paper we have shown how sensitive the peaks are to the different sound velocity profiles used, exhibiting significant differences even when the profiles have the same asymptotic and horizon limits.

Appendix A Bogoliubov coefficients relating the Kruskal and Boulware modes

We report here the Bogoliubov coefficients appaearing in eq. (20),

$$u_{H}^{K}(\omega_{K},x)=\int_{0}^{\infty}d\omega\left[\alpha_{\omega_{K}\omega}^{R}u_{H}^{R}+\beta^{R}_{\omega_{K}\omega}u_{H}^{R*}+\alpha_{\omega_{k}\omega}^{L}u_{H}^{L}+\beta^{L}_{\omega_{k}\omega}u_{H}^{L*}\right]\ ,$$ (36)

$$\alpha^{R}_{\omega_{k}\omega} = \frac{1}{2\pi\kappa}\sqrt{\frac{\omega}{\omega_{K}}}\Gamma(-i\omega/\kappa)\;\left(-i\frac{\omega_{K}}{\kappa}\right)^{i\omega/\kappa}\ ,$$

$$\beta^{R}_{\omega_{k}\omega} = \frac{1}{2\pi\kappa}\sqrt{\frac{\omega}{\omega_{K}}}\Gamma(i\omega/\kappa)\;\left(-i\frac{\omega_{K}}{\kappa}\right)^{-i\omega/\kappa}\ ,$$

$$\alpha^{L}_{\omega_{k}\omega} = \frac{1}{2\pi\kappa}\sqrt{\frac{\omega}{\omega_{K}}}\Gamma(i\omega/\kappa)\;\left(i\frac{\omega_{K}}{\kappa}\right)^{-i\omega/\kappa}\ ,$$

$$\beta^{L}_{\omega_{k}\omega} = \frac{1}{2\pi\kappa}\sqrt{\frac{\omega}{\omega_{K}}}\Gamma(-i\omega/\kappa)\;\left(i\frac{\omega_{K}}{\kappa}\right)^{i\omega/\kappa}\;.$$ (37)