Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

A matching theory to characterize sound emission during vortex reconnection in quantum fluids

A matching theory to characterize sound emission during vortex reconnection in quantum fluids A matching theory to characterize sound emission during vortex reconnection in quantum uids Davide Proment School of Mathematics, University of East Anglia, Norwich Research Park, Norwich NR4 7TJ, United Kingdom Giorgio Krstulovic Laboratoire J. L. Lagrange, Observatoire de la C^ ote d'Azur, Universit e de la C^ ote d'Azur, CNRS, Nice 06304, France In a concurrent work, Villois et al. 2020 [1] reported the evidence that vortex reconnections in quantum uids follow an irreversible dynamics, namely vortices separate faster than they ap- proach; such time-asymmetry is explained by using simple conservation arguments. In this work we develop further these theoretical considerations and provide a detailed study of the vortex recon- nection process for all the possible geometrical con gurations of the order parameter (super uid) wave function. By matching the theoretical description of incompressible vortex laments and the linear theory describing locally vortex reconnections, we determine quantitatively the linear mo- mentum and energy exchanges between the incompressible (vortices) and the compressible (density waves) degrees of freedom of the super uid. We show theoretically and corroborate numerically, why a unidirectional density pulse must be generated after the reconnection process and why only certain reconnecting angles, related to the rates of approach and separations, are allowed. Finally, some aspects concerning the conservation of centre-line helicity during the reconnection process are discussed. arXiv:2005.02047v2 [physics.flu-dyn] 24 Sep 2020 2 I. SCIENTIFIC MOTIVATIONS A. Introduction Solving the dynamics of the vorticity eld is a fundamental problem in uid mechanics [2]. A uid possesses in general a continuous three-dimensional vorticity eld, but it occurs often in nature that vorticity takes very high values in narrow uid structures supported in vortex patches like tubes, sheets or laments. As such is the case, studying the dynamics of the vorticity eld then results in characterizing the interaction between these structures. A vortex reconnection is a seminal example of such problem: how do two vortex tubes approach, interact, interchange, and eventually separate? Vortex reconnections are an important phenomenon in uid dynamics. They are observed in the context of plasma physics [3], and both classical [4] and quantum uids [5, 6]. Reconnections play an important role transferring energy in super uid turbulence, ne scale mixing of classical uid turbulence [7] and solar physics [8]. Addressing vortex reconnections in (classical) incompressible viscous uids is an extremely dicult task. Viscosity is responsible for the di usion of the vorticity eld: an initially nite-supported, or very intense, vorticity tube will gradually loose its distinctiveness in time. Moreover, even for reconnecting processes where the vorticity eld of the tubes is so intense that the reconnection time-scale is much faster than viscous di usion, a variety of new vortex patches like bridges [4, 9], ribbons [10], pancakes [11], or extra vortex tubes [12, 13] emerge. Hence, the presence of viscosity is two-folded when studying classical vortex reconnections. It is a blessing as otherwise, following Kelvin's theorem, the circulation around a vortex lament will simply be transported by the ow and vortex patches will not reconnect; but it is also a curse as it makes the problem mathematically prohibitive. Furthermore, any description of vortex lament dynamics needs to be accompanied with some modeling of the vortex core. At the moment of reconnection, where the scales at play are of the order of the core size, viscous dissipation becomes extremely strong and it is very dicult to disentangle universal aspects of vortex reconnection from the model-depending physics. What universal aspects of vortex reconnection should prevail in the limit of in nite Reynolds number is an unsolved question. In that sense and from a theoretical point of view, it would be natural to consider the limit where the vortex core is very small. Such limit, naturally arises in quantum uids. Quantum uids, also known as super uids, are exotic uids characterized by the complete absence of viscosity. This (non-classical) property is consequence of their inherent quantum mechanical nature: super uidity is closely related to Bose{Einstein condensation, a phase transition occurring when a three-dimensional system of bosons is cooled down below a critical temperature or reaches a critical density [14]. In a quantum uid the vorticity eld is lamentary (-supported) [15] and because its circulation takes only discrete values, these vortex laments are called quantized (or quantum) vortices. Quantum vortices are actually topological defects of a macroscopic wave function and they are thus topologically stable non-linear structures. At large distances, they interact each other in the same manner as classical hydrodynamic laments do. However, unlike classical inviscid vortex laments, despite the total absence of viscosity, quantized vortex reconnections naturally occur in this system thanks to the presence of dispersive e ects [16]. The previous reasons make a quantum uid the ideal system to study the uid mechanical problem of lamentary vortex reconnections. In this paper we study super uid vortex reconnections theoretically and numerically. We provide a theory to explain the origin of the time irreversibly reported in [1] within the framework of the Gross-Pitaevskii model. Our theory explains momentum and energy exchanges during the reconnection process and it is well supported by numerical data. B. The Gross{Pitaevskii model The simplest model that mimics the dynamics of a quantum uid is probably the Gross{Pitaevskii (GP) equation. Formally derived for Bose{Einstein condensates made of a dilute gas of bosons [14], it is a mean eld nonlinear Schroedinger-type equation that reads 2 2 @ ~ 4~ a 2 2 i~ = r + j j + V ; (1) ext @t 2m m where (r; t) 2 C is the wave-function of the condensate order parameter, ~ is the reduced Planck constant, and m and a are the mass and s-wave scattering length of the bosons, respectively. In this work, for the sake of simplicity, we assume no external con nement, that is V  0, and consider the system in a periodic cubic box of volume V . ext The following three integrals of motion, corresponding to the total number of bosons, the Hamiltonian (or energy), and linear momentum, respectively, exist N = j j dV ; (2) V 3 2 2 ~ 2~ a 2 4 H = jr j + j j dV ; (3) 2m m and i~ P = ( r r ) dV ; (4) where () stands for complex conjugation. Introducing the bulk mass density  = mN=V , the healing length  = m=(8a  ) and the speed of sound 0 s 0 2 3 c = (4~ a  )=m , Eq.(1) with V  0 results in s 0 ext @ c m 2 2 2 i = p  r + j j : (5) @t 2 0 Written in this manner, the GP equation has the advantage of putting into evidence the time-scale =c and the length-scale  at which dispersive e ects are important. The super uid nature of the system appears evident when one recasts the model in a uid mechanical framework by using the Madelung transformation p p (r; t) = (r; t)=m exp[i(r; t)=( 2c)] : (6) The imaginary and real parts of eq. (5) results, respectively, in +r ( v) = 0 @t ; where v(r; t) = r(r; t) : (7) @v c 2 2 + (vr) v = r + c  r @t These are nothing but the mass and linear momentum continuity equations for a barotropic, compressible, irrotational, inviscid uid [17], that is a super uid of density  and velocity eld v. Despite the irrotational property, vortices arise as lamentary topological defects where vanishes and its argument, see eq. (6), changes by n-multiples of 2. If so happens, the uid mechanical circulation evaluated over a closed curve around one of these defects is thus quantized and reads I I p p C = v dl = 2c r arg( ) dl = n ; where n 2 Z and = 2 2c = (8) is the quantum of circulation. This constraint implies that the velocity eld v diverges as r , where r is the distance to the vortex. Note that even if arg( ) is ill-de ned at any vortex point, the wave-function is a regular eld. The pretended singularity is just a consequence of the Madelung transformation that is not de ned on the topological defects. By de nition, the density vanishes on the vortex lines making, for instance, the energy (3) and the momentum (4) well de ned quantities. The typical size of a vortex core is controlled by the dispersion of the system and it is thus of the order of one healing length . Numerically, vortex lines are easily detectable by plotting the iso-surfaces of a low density value compare the bulk, or, if great precision is needed, by tracking accurately the nodal lines themselves [18, 19]. In a nutshell, if one forgets about density waves, a super uid described by the GP equation can be considered pictorially as a collection of vortex laments with a vortex core of size of order , whose interaction at large distances is roughly given by the Biot-Savart based vortex lament model [20]. C. Irreversible dynamics in quantized vortex reconnections Although the possibility of quantum vortex reconnections were suggested by Feynman in the 50's [21], the rst numerical evidence of quantized vortex reconnections, within the GP model, was only given almost 40 years later by Koplik and Levine [16]. Following this seminal work, many numerical and theoretical studies were undertaken in the following decades. Many of those works focused on characterizing their rate of approach and separation, on the macroscopic angle between two reconnecting vortex laments [6, 22{28], on the sound emission following a reconnection event [24, 29], and on the evolution of the length of the lament and super uid helicity throughout the 4 2.5 1.5 0.5 0 0.5 1 1.5 2 2.5 FIG. 1. (Color online) Values of approaching and separation pre-factors A and A . Red circles correspond to data of [1]. Gray left and right triangles correspond to reconnections of free and trapped vortices respectively, from Galantucci et al.[35]. Other symbols are taken from Villois et al. [25]. 0.976 0.992 1.01 1.02 1.04 a) b) c) d) FIG. 2. (Colour online) Snapshots of density iso-surfaces of a decaying Hopf link: times correspond to t = 0 (a), t = 53c= (b), t = 71c= (c), and t = 101==c (d). After the reconnection takes place (b), a clear sound pulse is created and propagates unidirectionally. reconnection process [30{34]. Despite some initial contradicting results, there is now a general consensus that about a reconnection event the laments approach and separate following the law 1=2 (t) = A (jt t j) ; (9) where  is the distance between the laments, t is the time at which the reconnection event takes place, and the 1=2 signs  stand for what happens before (-) and after (+) the reconnection event. Even if the scaling  / jt t j is universal [25], the (dimensionless) pre-factors A are not: importantly, one usually nds that A & A , that is laments approach slower than they separate. Figure 1 reports a collection of (A ; A ) values obtained in the literature for reconnections of very di erent nature following: the decay of Hopf links (red circles) [1]; interactions between vortex lines and rings in homogeneous and trapped super uids (triangles) [27] and regular and random con gurations of vortex laments (all other symbols) [25]. These results are a clear evidence of the irreversible dynamics of the reconnection process in quantum uids. Despite the fact that the GP dynamics is time-reversible and conservative, the data plotted in Fig.1 cluster in the region A > A , thus exhibiting a time-asymmetry. In other words, following the experience provided by these data, an educated observer could in principle guess the time direction of a reconnection event. This fact is clear a manifestation of the irreversibility of the process. Moreover, as the reconnecting laments accelerate during a reconnection process, a directional sound pulse is generated in the super uid. Figure 2 shows a set of snapshots of the evolution of some iso-surfaces of the super uid density eld to highlight the sound emission during a reconnection event following the decay of an Hopf link [1]. After the reconnection takes place, here t ' 53=c, a distinctive variation of the bulk density (yellow-to-violet colors) rec emerges at the reconnection point and propagates non-isotropically. Explaining the origin of the asymmetry between the pre-factors A and characterizing the directionality and intensity of the sound pulse are the main scopes of this work. In our reasoning and calculations we use two di erent limits of the GP model: the linear model, essentially the Schroedinger equation where the nonlinear term of GP is 5 (a) (b) FIG. 3. (Colour online) Here p = 1; A = 2; B = 1; C = 2; D = 1;  = =3 and lengths are rescaled by  . (a) An example of the two intersecting iso-surfaces Re( ) = 0 and Im( ) = 0 of eq. (11). (b) The hyperbola resulting from the projection of r r the nodal lines onto the z = 0 plane for the times before, t = 1, and after, t = 1, the reconnection time t = 0. neglected, and the Biot-Savart model, where the compressible degrees of freedom of the super uid are neglected. In a nutshell, our main results follow a simple matching between these two limits and make use of the conservation of total super uid linear momentum and energy, Eqs (4) and (3), respectively. The work is organized as follows. Section II describes exhaustively all the possible geometrical con gurations taken by two reconnecting laments: subsections II A and II B explore the range of parameters of the wave-function about a reconnection event, subsection II C relates these wave-function's parameters to a set of geometrical parameters for the vortex laments, nally subsection II D introduces a useful parameterization of the laments in terms of the geometrical parameters. Section III is devoted to the variations in time of the linear momentum and energy of the laments during a reconnection, and relate these to the emission of the sound pulse: subsection III A set the framework for such study, while subsections III B and III C contain the detailed calculations of the linear momentum di erence and energy variations, respectively. Section IV is left for the conclusions and future perspectives. II. THE GEOMETRY OF THE RECONNECTING FILAMENTS A. The reconnecting wave-function of the laments As previously obtained in [22, 25], about a reconnection event, the nonlinear term of the Gross-Pitaevskii (GP) model can be neglected, as the super uid density vanishes at the vortex core. The dynamics is thus driven by the linear Schroedinger equation i@ = r : (10) Without any loss of generality we set the reconnection time t = 0 and let the reconnection point be the origin of our reference frame. The most general second-order-polynomial wave-function initial condition having two nodal lines intersecting at the origin results in 2 2 2 2 1 A(x cos  + y sin ) + B(x cos  + y sin ) Cx + Dy (x; y; z) = p z + i z ; (11) 5=2 2 2 where p = 1, (A; B; C; D) 2 R and  2 [0; ] are dimensionless parameters, we will call them wave-function parameters in what follows.  > 0 is a characteristic length. The nodal lines of Eq. (11) can be easily identi ed by intersecting the two iso-surfaces Re( ) = 0 and Im( ) = 0: Figure 3(a) shows an example of this intersection when r r setting the parameters to p = 1; A = 2; B = 1; C = 2; D = 1;  = =3 and where lengths are rescaled by  . 6 Under the linear Schroedinger operator in Eq. (10) the solution in time reads it r 2 (x; y; z; t) = e (x; y; z) = 1 + it r (x; y; z) : (12) r r After some tedious algebra (see the Mathematica notebook available as Supplemental Material [36]), we obtain that the evolution of the wave-function nodal lines Re( ) = 0 and Im( ) = 0 results in the equations 2 2 (A + B)(x + y ) + (A B) [(x y)(x + y) cos(2) + 2xy sin(2)] C + D z = t (13) 4 4p and 2 2 Cx + Dy A + B z = + t ; (14) 2 4p respectively. By simplifying the z-dependence in Eqs. (13) and (14), one nds that the projection of the nodal lines onto the z = 0 plane satis es the equation A + B + C + D 2 2 (A + B 2C )x + (A + B 2D)y + (A B) [(x y)(x + y) cos(2) + 2xy sin(2)] = t : (15) For a suitable choice of the dimensionless parameters, this relation identi es a hyperbola; an example of it is shown in Fig. 3(b) where, again, p = 1; A = 2; B = 1; C = 2; D = 1;  = =3 and lengths are rescaled by  . It is important to notice that the (x; y) axes can always be rotated in order to ensure that the hyperbola asymptotes are mirrored with respect to the two axes. This property simply re ects the fact that Eq. (15) can be re-expressed in its normal form by a suitable rotation that depends on the chosen value of the wave-function parameter . For the sake of simplicity, we set the speci c value 0: (16) Note however that all the following considerations are actually valid for any value of  (see Appendix). The projection of the nodal lines onto the z = 0 plane now results in C A A + B + C + D 2 2 x + y = t ; (17) B D 2(B D)p by assuming, from now onwards, that B 6= D. Finally, by simplifying the y-dependence in Eqs. (13) and (14), we nd that the projection of the nodal lines onto the y = 0 plane satis es the equation BC AD D(C + D) + B(A + B) z = x + t ; (18) 2(B D) 4(B D)p that is a parabola that shifts along the z-axis at constant speed. B. Region of validity of the other wave-function parameters We denote by R and R the sets of points of the two vortex laments before reconnection and by R and 1 2 1 R the ones after reconnection. Without loss of generality we may assume: that (i) about the reconnection point, + + R  fy > 0g and R  fy < 0g whereas R  fx < 0g and R  fx > 0g; and that (ii) the orientation of the 1 2 1 2 vorticity follows the arrows as in the sketch displayed in Fig. 4(a). In order to nd the range of the admissible values of the wave-function parameters p 1 and (A; B; C; D) 2 R of eq. (11), we thus need to impose the following validity conditions. Existence of the hyperbola. At the reconnection time t = 0 we want the hyperbola asymptotes y = (C A)=(B D) x set by Eq. (17) to be real (in other words we want the equation to describe a hy- perbola and not an ellipse). This reduces to the condition C A 0 : (19) B D 7 y/ζ z/ζ (a) (b) concavity − − t < 0, R t < 0, R 1,2 1,2 parameter Λ + + t > 0, R t > 0, R 1,2 1,2 + + φ φ x/ζ x/ζ FIG. 4. (Colour online) Sketch of the reconnecting laments projected (a) onto the z = 0 plane and (b) onto the y = 0 plane. Convention on the location of the laments. Our convention adopts that, about the reconnection point, the + + positions of the laments satisfy R  fy > 0g and R  fy < 0g whereas R  fx < 0g and R  fx > 0g. 1 2 1 2 Hence, by evaluating eq. (17) at times t < 0 and t > 0 we obtain that the following conditions, respectively, must hold (B D) p (C + D + A + B) < 0 and (C A) p (C + D + A + B) < 0 : (20) Convention on the vorticity orientation of the laments. The orientation of the laments can be evaluated by computing the pseudo-vorticity ! = rRe( )rIm( ) of the wave-function in Eq. (12) at its nodal lines [18]. In order to impose the vorticity orientations as the arrows sketched in Fig. 4(a) the following conditions 8 8 > p = 1 > p = 1 < < B < D or B > D (21) > > : : A > C A < C must be realised. C. The geometrical parameters of the laments As shown in the previous subsection, provided that a suitable choice of the wave-function parameters is taken about the reconnection event, the projection of the laments onto the z = 0 plane corresponds to the hyperbola represented in Eq. (17). We can de ne the macroscopic reconnecting angle  as the angle formed by the hyperbola asymptotes when considering the laments after the reconnection, as shown in Fig. 4(a). By simple geometrical considerations, and using Eq. (17), we nd that this angle is related to the wave-function parameters as C A  C A + 2 = 2 arctan () tan = : (22) B D 2 B D Also, about the reconnection event, the projection of the laments onto the y = 0 plane is the parabola found in Eq. (18). Its concavity results simply in BC AD  BC AD = ; given the dimensionless concavity parameter  = . (23) (B D)  (B D) A sketch of the projection displaying the role of the concavity parameter is shown in Fig. 4(b). 8 FIG. 5. (Colour online) Region of the admissible parameters (B; D) for p = 1 (blue areas) and p = 1 (red areas), given a speci c choice of the geometrical parameters ( ; ). We can express these two geometrical parameters versus the wave-function parameters characterizing the wave- function initial condition. A possible choice is the following > A = B tan + : (24) > 2 C = D tan + In terms of the geometrical parameters, the region of validity given by Eqs. (16), (19), (20), and (21) results in 2 < tan 1 (B + D) \ (p = 1 \ D > B) [ (p = 1 \ D < B) : (25) This set of validity conditions are better summarized using the diagrams in Fig. 5, where we have set T = tan 1 (26) to simplify the notation. For a given choice of the geometrical parameters  2 (0; =2) [ (=2; ) and  2 R, two important remarks should be made: rstly, it is always possible to nd a set of values (p; B; D) that satis es the validity conditions in Eq. (25); secondly, the area spanned by the parameters (B; D) is unbounded. The limit ! =2, causing T ! 0, needs a special consideration: in this case, using Eq. (25), we nd that  2 R , that is the concavity can only be negative, otherwise the convention on the positions of the laments is not satis ed. Actually, a completely symmetrical reconnection ( = =2) can not be realized in a fully planar con guration ( = 0) with a quadratic wave function. 9 D. Parametrization of the laments The vortex laments can be parametrized in terms of the reconnecting angle  and the concavity parameter  as follows: (t)   (t) R (`; t) = cot sinh (`); cosh (`); z (`; t) (27) 2 2 2 (t)   (t) R (`; t) = cot sinh (`); cosh (`); z (`; t) (28) 2 2 2 + + + (t)  (t) + + R (`; t) = cosh (`); tan sinh (`); z (`; t) (29) 2 2 2 + + + (t)  (t) + + R (`; t) = cosh (`); tan sinh (`); z (`; t) ; (30) 2 2 2 where ` 2 (1; +1) is the parameter spanning the entire length of the laments (note however that ` does not correspond to the arc-length parametrization of the laments). Here 2 tan ( =2) 1 (B + D) 4 A (t) = A jtj ; A = ; A = ; (31) (B D)p tan( =2) and z (`; t) can be expressed by using eq. (18) and the square of the x-component of the laments before and after the reconnection, respectively, resulting in z (`; t) = z (t) +  (t) cot sinh(`) 8 2 ; (32) + + z (`; t) = z (t) +  (t) cosh(`) where 2 2 2 + (B + D ) tan ( =2) 1 + (B + D) z (t) = t : (33) 4(B D)p This choice of parametrization is particularly useful as one can see easily that the distance between the laments before and after the reconnection is given by jR (` = 0; t) R (` = 0; t)j =  (t) ; (34) 1 2 1=2 which immediately demonstrates the scaling  / jtt j [22, 25]. Furthermore, we can notice that the ratio between the rate of approach and separation is independent on the concavity parameter  and reads + + A = = cot : (35) A 2 Finally, we can understand by looking at Eqs. (31) and (33) that the choice of the parameters (B; D) only in uences the time-scale of the reconnection process. III. MOMENTUM AND ENERGY TRANSFERS DURING A RECONNECTION As it has been observed in previous works [24, 29, 37] and clearly displayed in Fig.2, when a vortex reconnection takes place in a quantum uid, a sound pulse is excited. Energy and momentum are thus transferred from the incompressible to the compressible degrees of freedom of the super uid in an irreversible manner. The aim of this section, and the main result of this work, is to develop an asymptotic matching theory that allows for quantifying such energy and momentum exchanges. In the GP model, the total energy (3) and linear momentum (4) are conserved during the reconnection process, as they are integrals of motion. Well before a reconnection event, practically only the presence of vortex laments 10 contribute to the invariants, whereas after reconnection, both laments and compressible waves add up their con- tributions to them. If one is able to estimate the contribution of the laments, then the contribution of density waves can be deduced using the conservation of the invariants. In the case of energy, such decomposition can be easily done numerically by splitting the kinetic energy term into the incompressible and compressible parts [17]. Such measurements were performed in [1] and will be reproduced below in our discussions. Our analytical treatment of the problem is as follows. When the laments are far from each other, i.e. (t)  , their dynamics of mainly driven by the Biot-Savart model. In that region we might use the vortex lament description to evaluate their energy and momentum of the super uid. On the other hand, when  (t)  , the dynamics is governed by the linear regime given by the Schroedinger equation. Vortices then reconnect following the laws described in the preceding section. We thus describe the reconnection matching, sketched in Fig.6, as follows. Before reconnection, some Biot-Savart dynamics leads to the pre-reconnection input con guration R for the laments 1;2 about the vortex reconnection point. The Biot-Savart description is assumed to be valid down to a distance  =  , lin where  is of the order of a few healing lengths. From there, the laments are driven by the Schroedinger equation lin allowing them to reconnect. After the reconnection, this linear regime is valid until the vortices separate up to + + a distance  =  . The linear evolution thus provides the output post-reconnection con guration R for the lin 1;2 laments. From there onwards, the dynamics is again governed by the Biot-Savart model. Note that the linear regime time (t) ⇠ (t) ⌧ ⇠ (t) ⇠ <latexit sha1_base64="Ya0DY5y/uo686nVffTbTb0EXnpo=">AAACEnicbVDLTsJAFJ36RHxVXLqZSExwIWnRRJdENy4xkUdCkUynA0yYdpqZWwNp+AvXbvUb3Bm3/oCf4F84BRYKnuQmJ+fcm3vv8WPBNTjOl7Wyura+sZnbym/v7O7t2weFhpaJoqxOpZCq5RPNBI9YHTgI1ooVI6EvWNMf3mR+85EpzWV0D+OYdULSj3iPUwJG6toFL2ACyMNZCU69fh97I961i07ZmQIvE3dOimiOWtf+9gJJk5BFQAXRuu06MXRSooBTwSZ5L9EsJnRI+qxtaERCpjvp9PYJPjFKgHtSmYoAT9XfEykJtR6HvukMCQz0opeJ/3qZAlIKvXAA9K46KY/iBFhEZ/t7icAgcZYPDrhiFMTYEEIVNy9gOiCKUDAp5k027mISy6RRKbvn5crdRbF6PU8ph47QMSohF12iKrpFNVRHFI3QM3pBr9aT9Wa9Wx+z1hVrPnOI/sD6/AHSH518</latexit> <latexit sha1_base64="x3GdmI1tmJLmLHqbEY/335CXFgI=">AAACFHicbVDLTgJBEJzFF+Jrfdy8TCQmeCG7aKJHohePmMgjYZHMDgNMmNnZzPQakfAbnr3qN3gzXr37Cf6Fs8BBwUo6qVR1p7srjAU34HlfTmZpeWV1Lbue29jc2t5xd/dqRiWasipVQulGSAwTPGJV4CBYI9aMyFCweji4Sv36PdOGq+gWhjFrSdKLeJdTAlZquwdBhwkgd0EsC3ASCIGDB952817RmwAvEn9G8miGStv9DjqKJpJFQAUxpul7MbRGRAOngo1zQWJYTOiA9FjT0ohIZlqjyfVjfGyVDu4qbSsCPFF/T4yINGYoQ9spCfTNvJeK/3qpAkoJM3cAdC9aIx7FCbCITvd3E4FB4TQh3OGaURBDSwjV3L6AaZ9oQsHmmLPZ+PNJLJJaqeifFks3Z/ny5SylLDpER6iAfHSOyugaVVAVUfSIntELenWenDfn3fmYtmac2cw++gPn8wfsXp6m</latexit><latexit sha1_base64="vqgSSZr8nhlK16ec2D1Xx6GjNX4=">AAACEnicbVDLSsNAFJ34rPUV69LNYBEqQkmqoMuiG5cV7AOaWiaTaTt0kgkzN9IS+heu3eo3uBO3/oCf4F84abvQ1gMXDufcy733+LHgGhzny1pZXVvf2Mxt5bd3dvf27YNCQ8tEUVanUkjV8olmgkesDhwEa8WKkdAXrOkPbzK/+ciU5jK6h3HMOiHpR7zHKQEjde2CFzAB5OGsBKdev4+9Ee/aRafsTIGXiTsnRTRHrWt/e4GkScgioIJo3XadGDopUcCpYJO8l2gWEzokfdY2NCIh0510evsEnxglwD2pTEWAp+rviZSEWo9D33SGBAZ60cvEf71MASmFXjgAeledlEdxAiyis/29RGCQOMsHB1wxCmJsCKGKmxcwHRBFKJgU8yYbdzGJZdKolN3zcuXuoli9nqeUQ0foGJWQiy5RFd2iGqojikboGb2gV+vJerPerY9Z64o1nzlEf2B9/gDO1Z16</latexit> 1,2 1,2 <latexit sha1_base64="8fNreUZBNgXcmMFVCQLOMbLg4vI=">AAACD3icbVDLSsNAFJ34rPXRqEs3g0VwoSWpgi6LblxWsQ9oY5hMJ+3QyUyYmQgl5CNcu9VvcCdu/QQ/wb9w0mahrQcuHM65l3vvCWJGlXacL2tpeWV1bb20Ud7c2t6p2Lt7bSUSiUkLCyZkN0CKMMpJS1PNSDeWBEUBI51gfJ37nUciFRX8Xk9i4kVoyGlIMdJG8u1K2g9CeJc9nPqpe1LPfLvq1Jwp4CJxC1IFBZq+/d0fCJxEhGvMkFI914m1lyKpKWYkK/cTRWKEx2hIeoZyFBHlpdPDM3hklAEMhTTFNZyqvydSFCk1iQLTGSE9UvNeLv7r5YoWgqm5A3R46aWUx4kmHM/2hwmDWsA8HDigkmDNJoYgLKl5AeIRkghrE2HZZOPOJ7FI2vWae1ar355XG1dFSiVwAA7BMXDBBWiAG9AELYBBAp7BC3i1nqw36936mLUuWcXMPvgD6/MHUtScIA==</latexit> <latexit sha1_base64="HI63mBs/BbivdJEnIYA94mzeKGw=">AAACD3icbVDLSsNAFJ34rPXRqEs3g0UQlJJUQZdFNy6r2Ae0MUymk3boZCbMTIQS8hGu3eo3uBO3foKf4F84abPQ1gMXDufcy733BDGjSjvOl7W0vLK6tl7aKG9ube9U7N29thKJxKSFBROyGyBFGOWkpalmpBtLgqKAkU4wvs79ziORigp+rycx8SI05DSkGGkj+XYl7QchvMseTvzUPa1nvl11as4UcJG4BamCAk3f/u4PBE4iwjVmSKme68TaS5HUFDOSlfuJIjHCYzQkPUM5iojy0unhGTwyygCGQpriGk7V3xMpipSaRIHpjJAeqXkvF//1ckULwdTcATq89FLK40QTjmf7w4RBLWAeDhxQSbBmE0MQltS8APEISYS1ibBssnHnk1gk7XrNPavVb8+rjasipRI4AIfgGLjgAjTADWiCFsAgAc/gBbxaT9ab9W59zFqXrGJmH/yB9fkDT5KcHg==</latexit> input output Biot-Savart model Schrödinger equation Biot-Savart model FIG. 6. (Colour online) A sketch of a reconnection process and matching asymptotics. When vortices are far apart their dynamics if governed by the Biot-Savart equation, whereas when they are about to reconnect the process is driven by the Schroedinger equation. corresponds only to the dynamics inside the orange-greenish box in Fig. 6. We can thus consider the linear regime as the regularization mechanism allowing vortex reconnections in the Biot-Savart model. Summarizing, in order to compute the di erences before and after the reconnection in the incompressible energy and momentum of the super uid, we use the theoretical description R for the laments given in Eqs. (27-30). 1;2 Namely, we use such parametrization when the distance is =  =  &  ; (36) lin as illustrated in Figs. 7(a) and 7(b), respectively. Note that the assumption that the linear regime description may be still valid at distances beyond the healing length  is justi ed by numerical evidence [25]. A. The cylindrical region of integration As detailed in the following, the calculations of the linear momentum and energy of the vortices involve the integration over the full length of the laments. As we are interested in their di erences, to simplify the problem we will consider only the segments of the laments which lie inside the cylinder of circle of radius R centered at the origin and having the cylindrical axis parallel to the z axis; the projection of the cylinder onto the z = 0 plane, corresponding to the circle of radius R centered at the origin, is also sketched in Fig. 7. 11 (a) (b) t < 0 t > 0 lin lin R R FIG. 7. (Colour online) Projections of the reconnecting laments and the cylinder onto the plane z = 0 before the reconnection (a) and after the reconnection (b). Here the reconnection angle is  = =3. The vortex laments lie inside the cylinder when their parametrization satis es j`j  L before and after the reconnection, respectively, given 8 9 h ih i > 2 2 > 2 2 > > 8(R= ) + (A 1) + 2 4 (R= ) 1 4 (R= ) + A < lin lin lin = r r L (R= ) = ln (37) lin > > 2 A + 1 > r > : ; 8 9 h ih i > 2 2 > 2 2 2 > > 8A (R= ) + (1 A ) + 2A 4 (R= ) 1 4A (R= ) + 1 < lin r lin lin = r r r L (R= ) = ln : (38) lin 2 > A + 1 > > r > : ; Physically, it is natural to assume that R is larger, or much larger, than  . Note that, keeping A nite, we have lin r lim L (R= ) = ln(R= ) + ln + : : : lin lin R= !1 lin 1 + A (39) lim L (R= ) = ln(R= ) + ln + ln(A ) + : : : lin lin r R= !1 lin 1 + A B. Linear momentum di erence Following L. Pismen [38], the linear momentum of an incompressible and inviscid uid with lamentary vorticity eld of intensity , that is within the framework of the Biot-Savart model, reads I Z @R P = R dR = p (`) d` ; where p (`) = R (40) l l l 2 2 @` laments is the momentum density (per unit of lament length) and the integration interval follows the parametrization of the laments. In our case ` 2 (1; +1) and there are only two laments before and two laments after the reconnection. The incompressible super uid momenta density before and after reconnection can be therefore computed by using Eqs. (27-30). Due to the particular symmetries of the reconnecting con guration, their expressions are rather simple. 12 Interestingly they are independent of  and `, and read ! ! " # 2 2 2 L (R= ) lin A  A  A r r r p = 0; 0; =) P ( ; A ; R= ) = 0; 0; d` = 0; 0; L (R= ) r lin lin l l 4 4 2 L (R= ) lin ! ! " # 2 2 2 L (R= ) + + lin + + + + + p = 0; 0; =) P ( ; A ; R= ) = 0; 0; d` = 0; 0; L (R= ) : r + lin l l 4A 4A + 2A r r r L (R= ) lin (41) These results immediately tell us that the di erence of the linear momentum of vortex laments during a reconnection is non zero only along the z-direction and in the limit of large R= is given by lin 1 + A + r + P ( ; A ; R= ) = P ( ; A ; R= ) P ( ; A ; R= ) / = 2 csc  : (42) l;z lin r lin lin r lin lin r lin l;z l;z This is a remarkable result: following our convention, refer again to Fig. 4 and for all ( ; A ; R= ), the incom- lin r lin pressible super uid linear momentum changes during a reconnection only along the z-axis and, most importantly, its variation is always negative with a global maximum for  = =2. As a consequence, assuming that the total linear momentum within the cylinder of radius R remains conserved during the reconnection process, a density/phase compressible excitation, a sound pulse, must carry the missing momentum P = P ; (43) pulse l that is, must certainly move along the positive direction of the z-axis. C. Energy di erence In rst approximation, the (kinetic) energy of an incompressible inviscid uid consisting on a collection of vortex laments is proportional to the total length L of the vortex con guration. This result arises from the local induction approximation (LIA) of the Biot-Savart model. The (kinetic) incompressible energy of the super uid, in the LIA assumption, is thus given by @R E (L) = T L = T d` ; (44) LIA @` laments 2 2 where T =  log(L =a )=(4 ) is the so-called vortex line tension [38, 39], with L is a characteristic length order 0 0 0 0 of the mean radius and a is the vortex core size. The precise de nition of the vortex tension is not relevant for the next considerations. The incompressible super uid energy, before and after the reconnection, thus results in L (R= ) lin @R @R 1 2 E (A ; =;  ; R= ) = T + d` ; (45) r lin lin LIA @` @` L (R= ) lin and the energy di erence during the reconnection is simply E(A ; =;  ; R= ) = E (A ; =;  ; R= ) E (A ; =;  ; R= ) / L(A ; =;  ; R= ) ; (46) r lin lin r lin lin r lin lin r lin lin where Z + Z L (R= ) + + L (R= ) lin lin @R @R @R @R 1 2 1 2 L(A ; =;  ; R= ) = + d` + d` (47) r lin lin + @` @` @` @` L (R= ) L (R= ) lin lin is the di erence of the total length of the laments during the reconnection process. The integrals in (47) can be evaluated analytically for  = 0 and the nal result reads 2i 1 lin 2 + 2 L(A ; = = 0;  ; R= ) = A Ell iL (R= ) 1 + Ell iL (R= )jA + 1 ; (48) r lin lin lin lin r r A A r 13 (a) (b) 1 2 3 4 1 2 3 4 -1 -1 -2 -2 FIG. 8. (Colour online) Plots of the di erence in the total length of the laments during the reconnection process when  = 0 and its limiting case when R=  1, that is eq.s (48) and (49), versus A for di erent values of R= ; here  = 1. lin r lin lin where Ell(j) is the incomplete elliptic integral of the second kind. The behavior of eq. (48) for di erent values of R= is shown in Fig. 8(a). Note that L appears to be monotonically decreasing versus A and to saturate for large lin r R= . This latter property can be corroborated by taking the limit R= ! 1 keeping A nite: using eq. (39) one lin lin r gets ( " # ) ( " # )! 2i 4 (R= ) 1 4A (R= ) lin lin r lin 2 2 L(A ; = = 0; R= ! 1) = A Ell i ln p 1 + Ell i ln p jA + 1 : r lin r r 2 2 A A r 1 + A 1 + A r r (49) Figure 8(b) shows its behavior versus A for larger and larger values of R= ; one has however to be cautious in using r lin this limiting expression, as Eq. (49) presents an unphysical behavior when A ! 0 which is absent in the original Eq. (48). Also, we can note that due to symmetry reasons, the two following properties hold L(A = 1; =;  ; R= ) = 0 (50) r lin lin and L(A ; =;  ; R= ) = L(A ;=;  ; R= ) : (51) r lin lin r lin lin In the general case where  6= 0 and nite, the integrals in (47) have to be computed numerically. As for  = 0, L appears to converge in the limit R= ! 1; also, as shown in Figs. 9(a) and 9(b), we observe numerically that lin L(A 6= 0;jj= ! 1;  ; R= ) = 0 : (52) r lin lin Hence, let us note that the value of L(A ; ;  ; R= ) is always bounded by Eqs. (48) and (52). r lin lin Using the same rationale as in the linear momentum di erence subsection, we now assume that the total (kinetic) super uid energy within the cylinder is conserved during the reconnection. Hence, to enforce this conservation, a density/phase compressible excitation, a sound pulse, must be created with energy fraction E E (A ;j=j;  ; R= ) L(A ;j=j;  ; R= ) pulse LIA r lin lin r lin lin = = ; (53) E E L tot LIA;0 0 given E the total super uid energy, E is the LIA energy of the initial con guration of the laments, and L tot LIA;0 0 the total initial length of the laments. As L appears to be monotonically decreasing versus A and equal to zero at A = 1, then E is positive only for A > 1. This is a very important result, as it shows that, under our r pulse r assumptions, reconnections cannot happen for A  1, that is for   =2, because a sound pulse with negative energy should be created, which is clearly unphysical. 14 (a) (b) 1.5 1.5 1.0 1.0 0.5 0.5 1 2 3 4 -0.5 1 2 3 4 -1.0 -0.5 FIG. 9. (Colour online) Plots of the di erence in the total length of the laments during the reconnection process, that is a numerical estimation of eq (47), versus A for di erent values of jj= ; here  = 1 and R= = 2. r lin lin IV. DISCUSSIONS AND FUTURE PERSPECTIVES Summarizing, following the calculations presented in Section III, we are able to estimate the linear momentum and energy of the sound pulse emitted during a reconnection event. These reads P = [0; 0;P ( ; A ; R)] pulse l;z lin r ; (54) L(A ;jj=;  ; R= ) r lin lin E = E pulse tot where the details of the functions can be found in Eqs (42) and (47), respectively. In particular, we underline that for all (jj=;  ; R= ) lin lin P > 0 ; for all A pulse;z r ; (55) E > 0 ; for A > 1 pulse r explain the origin of the asymmetry observed in the distribution of the the pre-factors A , and the directionality observed in the sound pulse emission. Let us now to discuss further the nature of the sound pulse. As veri ed in reference [1] and apparent in Fig.2, the pulse indeed propagates in a well de ned direction. Figure 10(a) shows such a sound pulse, rescaled by the bulk density, propagating along the positive z-direction after a reconnection characterized by the geometrical parameter A = 1:67. The pulse is plotted versus time in Fig.10.a and versus the retarded time in Fig.10.b for di erent values of the z-coordinate. We can observe that the pulse appears to move slightly slower than the speed of sound c. The signal shows some dispersive e ects while time advances, an evidence that the pulse probably contains more than one Bogoliubov perturbation, perhaps including high wave numbers. Also, the depth of its trough  decays as min the pulse propagates following the scaling / z typical of a three-dimensional wave signal originating from a point source, as depicted in Fig. 10(c). It is however still unclear if the the pulse consists of a simple linear superposition of Bogoliubov perturbations, or if a fully nonlinear subsonic coherent structure, like a Robert{Jones solitary wave [40], is also superimposed. A complete analysis on the spectrum of the sound pulse and the possible presence of coherent structures is left for future works. Concerning the energy transferred from the vortices to the pulse, we present in Fig.11a-b the comparison between our theoretical prediction, Eq. (53), and the GP reconnection data obtained in [1] for di erent choices of  . The region lin of validity of the theory, the colored regions depicted in Fig. 11, represents all the accessible values of the concavity parameter  2 R, with boundaries obtained from Eqs (48) and (52). We note that the speci c value of the energy di erence in Eq. (53) depends on the choice of  , but the conclusion that reconnection with A < 1 are unlucky to lin r occur, remains valid. Moreover, the fact that no energy exchange takes place for a symmetric reconnection A = 1, 0:71 suggests to plot E versus A 1. Such plot is displayed in Fig.11.c, where the scaling E  (A 1) , wav r wav r 15 1.005 1.01 1.005 1 1 1 0.995 0.995 0.995 0.99 0.99 0.99 0.985 0.99 0.985 0.985 0.98 0.98 0.98 0.98 0.975 0.975 0.975 0.97 0.97 0 20 40 -20 0 20 20 25 30 35 40 0 50 FIG. 10. (Color online) Traveling pulse emitted during a reconnection with A = 1:67. (a) Density around the pulse as function of time for di erent values of the distance z to the reconnection plane. (b) Same as (b), but as a function of the retarded time. (c) Depth of the traveling pulse trough  at di erent values of z compared with the 1=z decay law predicted form min an acoustic pulse emitted from a point source. The theoretical formula is obtained by imposing the 1=z decay and matching the measured value of  at z = 20, the smallest value of z at which the pulse was measured. The speed of sound is c and min 0 = =c. -1 -2 -3 -4 0.3 0.5 1 2 3 4 5 FIG. 11. (Colour online) Energy radiated during reconnections as a function of A = A =A for di erent values of  . r lin Numerical data from GP simulations obtained in [1] is confronted to our theoretical prediction in Eq. (53). (a)  = 4. (b) lin = 8. (c) LogLog plot of energy as a function of A 1 for di erent values of  . lin r lin obtained by t, is clearly observed for at least one decade and all values of  . A more accurate theory, providing lin such scaling exponent is out of the scope of this work. It is interesting to note that in a recent experimental work [41], P. Svancara and M. La Mantia observed that in super uid helium, the skewness of Langrangian velocity increments exhibits the opposite trend with respect to the one of classical ows, namely, in super uids particles accelerate on average faster than they decelerate. The authors interpreted their ndings as a consequence of quantum vortex reconnections because particles could probe individual vortex reconnections. This interpretation is consistent with the recent numerical calculations reported in [42], where quantum vortex reconnections mediated by particles were studied. In that work, the authors observed that in addition to the momentum exchange between the vortices and sound waves, there is also a net transfer towards particles that incurs on a fast acceleration of particles at the reconnection event. We should also remark on the possible role played by the center-line helicity H during a reconnection event. As rst observed in [30], the evolution of H shows a sudden drop (but still being continuous) during a reconnection if the initial linking number between the vortex laments is non-zero. This is certainly the case analyzed in [1], where the initial con guration of the laments, an Hopf link, has linking number equal to 2. The temporal evolution of a Hopf link, for a case with A = 3:5, is displayed Fig. 12.a, where the helicity drop is clearly visible. One might be tempted to think that the sudden drop in the center-line helicity H is related to the properties of the reconnecting laments, for example their parameter A . Figure 12.b shows that H as a function of the parameter A . Numerical data r c r do not correlate as well as the energy and no clear trend is observed. We can however speculate that the center-line helicity drop may in uence the amplitude and polarization of the Kelvin waves forming on the vortex laments after the reconnection process. This research proposal constitutes another interesting direction for future works. In conclusion, our theoretical results help to understand not only how vortex reconnections take place in quantum uids, but also why they do occur. Albeit the GP model is time reversible, the vortex reconnection process presents 16 -0.2 -0.4 -0.6 0 1 2 3 4 FIG. 12. (Color online) a) Temporal evolution of central line helicity H for a reconnection having A = 3:5. The green zone c r correspond to the time interval de ned by  (t ) =  = 8. The red dot is the reconnection time. a) Drop of central line lin + + helicity H = H (t ) H (t ) as a function of A = A =A . is the circulation of the laments. c c c r a time asymmetry so that the system can naturally transfer part of the (kinetic) incompressible energy into its compressible counterpart. In some sense, this observed temporal asymmetry can be interpreted as consequence of the system being in an out-of-equilibrium stage and reconnections being a fast route towards reaching thermal equilibrium. It will be very interesting to study vortex reconnections at nite temperatures, where a thermal bath of Bogoliubov modes is present, to see if whether this asymmetry is reduced or destroyed completely. Furthermore, the situation can be di erent when open conditions, like external forcing and damping terms acting at di erent length scales, are introduced. In a fully developed turbulent state, uctuations could provide enough energy to allow reconnections with A < 1, but one might still expect skew distributions towards A > 1, as dissipation of turbulent ows do not vanishes r r in the limit of in nite Reynolds numbers because of the dissipative anomaly of turbulence [43]. This is undoubtedly, another interesting direction of research for future studies. Finally, it is important to remark that the theoretical predictions obtained in this work may complement future experiments in several physical systems. In Bose{Einstein condensates made of dilute gases, experimentalists are now able to observe the real-time dynamics of few reconnecting vortex laments [44] and are now working to create reproducible reconnections [45]. Even if measuring the incompressible and compressible energy components of the super uid is still a huge experimental challenge, assessing the directionality and shape of the sound pulse should be nowadays achievable using, for example, density imaging or Bragg spectroscopy. In super uid liquid helium-4 experiments, the challenge is to track single vortex reconnections, especially in the limit of very low temperatures where the normal component is negligible. However, measuring the production of compressible energy excitations (and/or rotons) during many reconnections occurring in a turbulent tangle might be achievable [30, 46, 47]; in order to be able to theoretically estimate this production rate, a statistical extension to our detailed analysis, including possibly nite temperature e ects, might be needed. Ultimately, let us point out there are impressive similarities between vortex reconnections occurring in quantum uids and classical uids at high Reynolds numbers, especially when the classical vortex tubes possess an almost hollow core [48, 49]. It would be therefore very interesting to measure the pre-factors A , reconnection angle  and concavity parameter  in the classical uid experiments and quantitatively compare them with our theoretical predictions. Appendix A: Choice of system of coordinates and value of parameter To simplify the analysis of the wave-function parameters and the following calculations we have set the parameter = 0. Actually, one can always perform a rotation in the x y plane such that Eq.(15) becomes Eq.(17) after a rede nition of the coecient A; B; C; D. Once this change of variable is performed, Eq.(14) will become in the new variables 2 2 C x + C xy + C y A + B xx xy yy z = + t ; (A1) 2 4p where C ; C and C are new coecient depending on the old ones. The analisys concerning the region of xx xy yy parameteres involves only the coecients of the hyperbola, therefore it remains unchanged. The concavity parameter 17 can not be de ned so straighforwardly as now the projection on the x = 0 or y = 0 are not simple parabolas. However, the concavity increases with the magnitud of the coeciens C ; C and C . xx xy yy The momentum di erence (42) remains unchanged if (A1) is taken into consideration for the vortex parametrization. Finally, the expression for the di erence of vortex length are slightly more complicated, but the property that L  0 for A  1 and L  0 for A  1 remains valid. r r ACKNOWLEDGMENTS The authors acknowledge fruitful discussions with J. Hannay that led to the considerations of the most general normal form discussed in the Appendix of this manuscript. The authors acknowledge L. Galantucci for providing data displayed in Fig.1 and A. Villois for supplying the raw data used in Fig.10. G.K., D.P. were supported by the cost- share Royal Society International Exchanges Scheme (IE150527) in conjunction with CNRS. D.P. was supported by the EPSRC First Grant scheme (EP/P023770/1). G.K. and D.P. acknowledge the Federation Doeblin for supporting DP during his sojourn in Nice. G.K. was also supported by the ANR JCJC GIANTE ANR-18-CE30-0020-0.1 and by the EU Horizon 2020 research and innovation programme under the grant agreement No 823937 in the framework of Marie Skodowska-Curie HALT project. Computations were carried out at M esocentre SIGAMM hosted at the Observatoire de la C^ ote d'Azur and on the High Performance Computing Cluster supported by the Research and Specialist Computing Support service at the University of East Anglia. Part of this work has been presented at the workshop \Irreversibility and Turbulence" hosted by Fondation Les Treilles in September 2017. The authors acknowledge Fondation Les Treilles and all participants of the workshop for the frightful scienti c discussions and support. D.P. acknowledges J. Hannay for his frequent visits to Norfolk that always lead to fruitful conversations. [1] A. Villois, D. Proment, and G. Krstulovic, Irreversible dynamics of vortex reconnections in quantum uids, arXiv:2005.02048 (2020). [2] P. G. Sa man, Cambridge Monographs on Mechanics (Cambridge University Press, Cambridge, 1993). [3] E. R. Priest, Heating the solar corona by magnetic reconnection, in Plasma Astrophysics And Space Physics: Proceedings of the VIIth International Conference held in Lindau, Germany, May 4{8, 1998 , edited by J. Buc  hner, I. Axford, E. Marsch, and V. Vasyliun  as (Springer Netherlands, Dordrecht, 1999) pp. 77{100. [4] S. Kida and M. Takaoka, Vortex reconnection, Annual Review of Fluid Mechanics 26, 169 (1994), http://dx.doi.org/10.1146/annurev. .26.010194.001125. [5] E. Fonda, D. P. Meichle, N. T. Ouellette, S. Hormoz, and D. P. Lathrop, Direct observation of kelvin waves excited by quantized vortex reconnection, Proceedings of the National Academy of Sciences 111, 4707 (2014). [6] S. Sera ni, L. Galantucci, E. Iseni, T. Bienaim e, R. N. Bisset, C. F. Barenghi, F. Dalfovo, G. Lamporesi, and G. Ferrari, Vortex reconnections and rebounds in trapped atomic bose-einstein condensates, Physical Review X 7, 021031 (2017). [7] F. Hussain and K. Duraisamy, Mechanics of viscous vortex reconnection, Physics of Fluids 23, 021701 (2011). [8] Xue Zhike, Yan Xiaoli, Cheng Xin, Yang Liheng, Su Yingna, Kliem Bernhard, Zhang Jun, Liu Zhong, Bi Yi, Xiang Yongyuan, Yang Kai, and Zhao Li, Observing the release of twist by magnetic reconnection in a solar lament eruption, Nat Commun 7, http://dx.doi.org/10.1038/ncomms11837 10.1038/ncomms11837 (2016). [9] J. Yao and F. Hussain, A physical model of turbulence cascade via vortex reconnection sequence and avalanche, Journal of Fluid Mechanics 883, A51 (2020). [10] A. Pumir and R. M. Kerr, Numerical simulation of interacting vortex tubes, Physical Review Letters 58, 1636 (1987). [11] D. S. Agafontsev, E. A. Kuznetsov, and A. A. Mailybaev, Development of high vorticity structures and geometrical properties of the vortex line representation, Physics of Fluids, Physics of Fluids 30, 095104 (2018). [12] R. M. Kerr, Swirling, turbulent vortex rings formed from a chain reaction of reconnection events, Physics of Fluids, Physics of Fluids 25, 065101 (2013). [13] R. McKeown, R. Ostilla-M onico, A. Pumir, M. P. Brenner, and S. M. Rubinstein, Turbulence generation through an iterative cascade of the elliptical instability, Science Advances 6, eaaz2717 (2020). [14] L. Pitaevskii and S. Stringari, Bose-Einstein condensation and super uidity, Vol. 164 0191076686 (Oxford University Press, 2016). [15] This is true for mean- eld models of quantum uids, when quantum uctuations are considered this picture is more complicated. [16] J. Koplik and H. Levine, Vortex reconnection in super uid helium, Phys. Rev. Lett. 71, 1375 (1993). [17] C. Nore, M. Abid, and M. Brachet, Decaying kolmogorov turbulence in a model of super ow, Physics of Fluids (1994- present) 9, 2644 (1997). [18] A. Villois, G. Krstulovic, D. Proment, and H. Salman, A vortex lament tracking method for the gross{pitaevskii model of a super uid, Journal of Physics A: Mathematical and Theoretical 49, 415502 (2016). 18 [19] A. Villois, D. Proment, and G. Krstulovic, Evolution of a super uid vortex lament tangle driven by the gross-pitaevskii equation, Phys. Rev. E 93, 061103(R) (2016). [20] M. D. Bustamante and S. Nazarenko, Derivation of the biot-savart equation from the nonlinear schr odinger equation, Phys. Rev. E 92, 053019 (2015). [21] R. Feynman, Chapter fIIg application of quantum mechanics to liquid helium (Elsevier, 1955) pp. 17 { 53. [22] S. Nazarenko and R. West, Analytical solution for nonlinear schr odinger vortex reconnection, Journal of low temperature physics 132, 1 (2003). [23] M. Kursa, K. Bajer, and T. Lipniacki, Cascade of vortex loops initiated by a single reconnection of quantum vortices, Phys. Rev. B 83, 014515 (2011). [24] S. Zuccher, M. Caliari, A. W. Baggaley, and C. F. Barenghi, Quantum vortex reconnections, Physics of Fluids 24, 125108 (2012). [25] A. Villois, D. Proment, and G. Krstulovic, Universal and nonuniversal aspects of vortex reconnections in super uids, Phys. Rev. Fluids 2, 044701 (2017). [26] C. Rorai, J. Skipper, R. M. Kerr, and K. R. Sreenivasan, Approach and separation of quantised vortices with balanced cores, Journal of Fluid Mechanics 808, 641 (2016). [27] L. Galantucci, A. W. Baggaley, N. G. Parker, and C. F. Barenghi, Crossover from interaction to driven regimes in quantum vortex reconnections, Proceedings of the National Academy of Sciences 116, 12204 (2019). [28] S. Rica, Self-similar vortex reconnection, Patterns and dynamics: homage to Pierre Coullet / Formes et dynamique: hommage  a Pierre Coullet, Comptes Rendus M ecanique 347, 365 (2019). [29] M. Leadbeater, T. Winiecki, D. C. Samuels, C. F. Barenghi, and C. S. Adams, Sound emission due to super uid vortex reconnections, Phys. Rev. Lett. 86, 1410 (2001). [30] M. W. Scheeler, D. Kleckner, D. Proment, G. L. Kindlmann, and W. T. M. Irvine, Helicity conservation by ow across scales in reconnecting vortex links and knots, Proceedings of the National Academy of Sciences 111, 15350 (2014), http://www.pnas.org/content/111/43/15350.full.pdf. [31] C. E. Laing, R. L. Ricca, and L. S. De Witt, Conservation of writhe helicity under anti-parallel reconnection, Scienti c reports 5 (2015). [32] P. Clark di Leoni, P. D. Mininni, and M. E. Brachet, Helicity, topology, and kelvin waves in reconnecting quantum knots, Phys. Rev. A 94, 043605 (2016). [33] H. Salman, Helicity conservation and twisted seifert surfaces for super uid vortices, Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences 473, 20160853 (2017). [34] S. Zuccher and R. L. Ricca, Helicity conservation under quantum reconnection of vortex rings, Phys. Rev. E 92, 061001 (2015). [35] L. Galantucci, A. W. Baggaley, N. G. Parker, and C. F. Barenghi, Crossover from interaction to driven regimes in quantum vortex reconnections, Proceedings of the National Academy of Sciences 116, 12204 (2019). [36] D. Proment and G. Krstulovic, Supplemental material for: A matching theory to characterize sound emission during vortex reconnection in quantum uids, Mathematica Notebook available as Supplemental Material (2020). [37] R. M. Kerr, Vortex stretching as a mechanism for quantum kinetic energy decay, Phys. Rev. Lett. 106, 224501 (2011). [38] L. M. Pismen and L. M. Pismen, Vortices in nonlinear elds: from liquid crystals to super uids, from non-equilibrium patterns to cosmic strings, Vol. 100 (Oxford University Press, 1999). [39] E. B. Sonin, Vortex oscillations and hydrodynamics of rotating super uids, Reviews of Modern Physics 59, 87 (1987). [40] C. A. Jones and P. H. Roberts, Motions in a bose condensate. iv. axisymmetric solitary waves, Journal of Physics A: Mathematical and General 15, 2599 (1982). [41] P. Svancara and M. La Mantia, Flight-crash events in super uid turbulence, Journal of Fluid Mechanics 876, R2 (2019). [42] U. Giuriato and G. Krstulovic, Quantum vortex reconnections mediated by trapped particles, Phys. Rev. B 102, 094508 (2020). [43] U. Frisch, Turbulence: The Legacy of A. N. Kolmogorov (Cambridge University Press, 1995). [44] S. Sera ni, M. Barbiero, M. Debortoli, S. Donadello, F. Larcher, F. Dalfovo, G. Lamporesi, and G. Ferrari, Dynamics and interaction of vortex lines in an elongated bose-einstein condensate, Phys. Rev. Lett. 115, 170402 (2015). [45] K. Xhani, E. Neri, L. Galantucci, F. Scazza, A. Burchianti, K. L. Lee, C. F. Barenghi, A. Trombettoni, M. Inguscio, M. Zaccanti, G. Roati, and N. P. Proukakis, Critical transport and vortex dynamics in a thin atomic josephson junction, Physical Review Letters 124, 045301 (2020). [46] I. Amelio, D. E. Galli, and L. Reatto, Probing quantum turbulence in He by quantum evaporation measurements, Phys. Rev. Lett. 121, 015302 (2018). [47] D. Proment, M. Onorato, and C. F. Barenghi, Vortex knots in a bose-einstein condensate, Phys. Rev. E 85, 036306 (2012). [48] D. Kleckner, L. H. Kau man, and W. T. M. Irvine, How super uid vortex knots untie, Nature Physics 12, 650 EP (2016). [49] S. V. Alekseenko, P. A. Kuibin, S. I. Shtork, S. G. Skripkin, and M. A. Tsoy, Vortex reconnection in a swirling ow, JETP Lett. 103, 455 (2016). http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Nonlinear Sciences arXiv (Cornell University)

A matching theory to characterize sound emission during vortex reconnection in quantum fluids

Proment, Davide; Krstulovic, Giorgio
Nonlinear Sciences , Volume 2020 (2005) – May 5, 2020
Download PDF
18 pages

Loading next page...
 
/lp/arxiv-cornell-university/a-matching-theory-to-characterize-sound-emission-during-vortex-Fce6u7QYGR

References

References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.

ISSN
2469-990X
eISSN
ARCH-3338
DOI
10.1103/PhysRevFluids.5.104701
Publisher site
See Article on Publisher Site

Abstract

A matching theory to characterize sound emission during vortex reconnection in quantum uids Davide Proment School of Mathematics, University of East Anglia, Norwich Research Park, Norwich NR4 7TJ, United Kingdom Giorgio Krstulovic Laboratoire J. L. Lagrange, Observatoire de la C^ ote d'Azur, Universit e de la C^ ote d'Azur, CNRS, Nice 06304, France In a concurrent work, Villois et al. 2020 [1] reported the evidence that vortex reconnections in quantum uids follow an irreversible dynamics, namely vortices separate faster than they ap- proach; such time-asymmetry is explained by using simple conservation arguments. In this work we develop further these theoretical considerations and provide a detailed study of the vortex recon- nection process for all the possible geometrical con gurations of the order parameter (super uid) wave function. By matching the theoretical description of incompressible vortex laments and the linear theory describing locally vortex reconnections, we determine quantitatively the linear mo- mentum and energy exchanges between the incompressible (vortices) and the compressible (density waves) degrees of freedom of the super uid. We show theoretically and corroborate numerically, why a unidirectional density pulse must be generated after the reconnection process and why only certain reconnecting angles, related to the rates of approach and separations, are allowed. Finally, some aspects concerning the conservation of centre-line helicity during the reconnection process are discussed. arXiv:2005.02047v2 [physics.flu-dyn] 24 Sep 2020 2 I. SCIENTIFIC MOTIVATIONS A. Introduction Solving the dynamics of the vorticity eld is a fundamental problem in uid mechanics [2]. A uid possesses in general a continuous three-dimensional vorticity eld, but it occurs often in nature that vorticity takes very high values in narrow uid structures supported in vortex patches like tubes, sheets or laments. As such is the case, studying the dynamics of the vorticity eld then results in characterizing the interaction between these structures. A vortex reconnection is a seminal example of such problem: how do two vortex tubes approach, interact, interchange, and eventually separate? Vortex reconnections are an important phenomenon in uid dynamics. They are observed in the context of plasma physics [3], and both classical [4] and quantum uids [5, 6]. Reconnections play an important role transferring energy in super uid turbulence, ne scale mixing of classical uid turbulence [7] and solar physics [8]. Addressing vortex reconnections in (classical) incompressible viscous uids is an extremely dicult task. Viscosity is responsible for the di usion of the vorticity eld: an initially nite-supported, or very intense, vorticity tube will gradually loose its distinctiveness in time. Moreover, even for reconnecting processes where the vorticity eld of the tubes is so intense that the reconnection time-scale is much faster than viscous di usion, a variety of new vortex patches like bridges [4, 9], ribbons [10], pancakes [11], or extra vortex tubes [12, 13] emerge. Hence, the presence of viscosity is two-folded when studying classical vortex reconnections. It is a blessing as otherwise, following Kelvin's theorem, the circulation around a vortex lament will simply be transported by the ow and vortex patches will not reconnect; but it is also a curse as it makes the problem mathematically prohibitive. Furthermore, any description of vortex lament dynamics needs to be accompanied with some modeling of the vortex core. At the moment of reconnection, where the scales at play are of the order of the core size, viscous dissipation becomes extremely strong and it is very dicult to disentangle universal aspects of vortex reconnection from the model-depending physics. What universal aspects of vortex reconnection should prevail in the limit of in nite Reynolds number is an unsolved question. In that sense and from a theoretical point of view, it would be natural to consider the limit where the vortex core is very small. Such limit, naturally arises in quantum uids. Quantum uids, also known as super uids, are exotic uids characterized by the complete absence of viscosity. This (non-classical) property is consequence of their inherent quantum mechanical nature: super uidity is closely related to Bose{Einstein condensation, a phase transition occurring when a three-dimensional system of bosons is cooled down below a critical temperature or reaches a critical density [14]. In a quantum uid the vorticity eld is lamentary (-supported) [15] and because its circulation takes only discrete values, these vortex laments are called quantized (or quantum) vortices. Quantum vortices are actually topological defects of a macroscopic wave function and they are thus topologically stable non-linear structures. At large distances, they interact each other in the same manner as classical hydrodynamic laments do. However, unlike classical inviscid vortex laments, despite the total absence of viscosity, quantized vortex reconnections naturally occur in this system thanks to the presence of dispersive e ects [16]. The previous reasons make a quantum uid the ideal system to study the uid mechanical problem of lamentary vortex reconnections. In this paper we study super uid vortex reconnections theoretically and numerically. We provide a theory to explain the origin of the time irreversibly reported in [1] within the framework of the Gross-Pitaevskii model. Our theory explains momentum and energy exchanges during the reconnection process and it is well supported by numerical data. B. The Gross{Pitaevskii model The simplest model that mimics the dynamics of a quantum uid is probably the Gross{Pitaevskii (GP) equation. Formally derived for Bose{Einstein condensates made of a dilute gas of bosons [14], it is a mean eld nonlinear Schroedinger-type equation that reads 2 2 @ ~ 4~ a 2 2 i~ = r + j j + V ; (1) ext @t 2m m where (r; t) 2 C is the wave-function of the condensate order parameter, ~ is the reduced Planck constant, and m and a are the mass and s-wave scattering length of the bosons, respectively. In this work, for the sake of simplicity, we assume no external con nement, that is V  0, and consider the system in a periodic cubic box of volume V . ext The following three integrals of motion, corresponding to the total number of bosons, the Hamiltonian (or energy), and linear momentum, respectively, exist N = j j dV ; (2) V 3 2 2 ~ 2~ a 2 4 H = jr j + j j dV ; (3) 2m m and i~ P = ( r r ) dV ; (4) where () stands for complex conjugation. Introducing the bulk mass density  = mN=V , the healing length  = m=(8a  ) and the speed of sound 0 s 0 2 3 c = (4~ a  )=m , Eq.(1) with V  0 results in s 0 ext @ c m 2 2 2 i = p  r + j j : (5) @t 2 0 Written in this manner, the GP equation has the advantage of putting into evidence the time-scale =c and the length-scale  at which dispersive e ects are important. The super uid nature of the system appears evident when one recasts the model in a uid mechanical framework by using the Madelung transformation p p (r; t) = (r; t)=m exp[i(r; t)=( 2c)] : (6) The imaginary and real parts of eq. (5) results, respectively, in +r ( v) = 0 @t ; where v(r; t) = r(r; t) : (7) @v c 2 2 + (vr) v = r + c  r @t These are nothing but the mass and linear momentum continuity equations for a barotropic, compressible, irrotational, inviscid uid [17], that is a super uid of density  and velocity eld v. Despite the irrotational property, vortices arise as lamentary topological defects where vanishes and its argument, see eq. (6), changes by n-multiples of 2. If so happens, the uid mechanical circulation evaluated over a closed curve around one of these defects is thus quantized and reads I I p p C = v dl = 2c r arg( ) dl = n ; where n 2 Z and = 2 2c = (8) is the quantum of circulation. This constraint implies that the velocity eld v diverges as r , where r is the distance to the vortex. Note that even if arg( ) is ill-de ned at any vortex point, the wave-function is a regular eld. The pretended singularity is just a consequence of the Madelung transformation that is not de ned on the topological defects. By de nition, the density vanishes on the vortex lines making, for instance, the energy (3) and the momentum (4) well de ned quantities. The typical size of a vortex core is controlled by the dispersion of the system and it is thus of the order of one healing length . Numerically, vortex lines are easily detectable by plotting the iso-surfaces of a low density value compare the bulk, or, if great precision is needed, by tracking accurately the nodal lines themselves [18, 19]. In a nutshell, if one forgets about density waves, a super uid described by the GP equation can be considered pictorially as a collection of vortex laments with a vortex core of size of order , whose interaction at large distances is roughly given by the Biot-Savart based vortex lament model [20]. C. Irreversible dynamics in quantized vortex reconnections Although the possibility of quantum vortex reconnections were suggested by Feynman in the 50's [21], the rst numerical evidence of quantized vortex reconnections, within the GP model, was only given almost 40 years later by Koplik and Levine [16]. Following this seminal work, many numerical and theoretical studies were undertaken in the following decades. Many of those works focused on characterizing their rate of approach and separation, on the macroscopic angle between two reconnecting vortex laments [6, 22{28], on the sound emission following a reconnection event [24, 29], and on the evolution of the length of the lament and super uid helicity throughout the 4 2.5 1.5 0.5 0 0.5 1 1.5 2 2.5 FIG. 1. (Color online) Values of approaching and separation pre-factors A and A . Red circles correspond to data of [1]. Gray left and right triangles correspond to reconnections of free and trapped vortices respectively, from Galantucci et al.[35]. Other symbols are taken from Villois et al. [25]. 0.976 0.992 1.01 1.02 1.04 a) b) c) d) FIG. 2. (Colour online) Snapshots of density iso-surfaces of a decaying Hopf link: times correspond to t = 0 (a), t = 53c= (b), t = 71c= (c), and t = 101==c (d). After the reconnection takes place (b), a clear sound pulse is created and propagates unidirectionally. reconnection process [30{34]. Despite some initial contradicting results, there is now a general consensus that about a reconnection event the laments approach and separate following the law 1=2 (t) = A (jt t j) ; (9) where  is the distance between the laments, t is the time at which the reconnection event takes place, and the 1=2 signs  stand for what happens before (-) and after (+) the reconnection event. Even if the scaling  / jt t j is universal [25], the (dimensionless) pre-factors A are not: importantly, one usually nds that A & A , that is laments approach slower than they separate. Figure 1 reports a collection of (A ; A ) values obtained in the literature for reconnections of very di erent nature following: the decay of Hopf links (red circles) [1]; interactions between vortex lines and rings in homogeneous and trapped super uids (triangles) [27] and regular and random con gurations of vortex laments (all other symbols) [25]. These results are a clear evidence of the irreversible dynamics of the reconnection process in quantum uids. Despite the fact that the GP dynamics is time-reversible and conservative, the data plotted in Fig.1 cluster in the region A > A , thus exhibiting a time-asymmetry. In other words, following the experience provided by these data, an educated observer could in principle guess the time direction of a reconnection event. This fact is clear a manifestation of the irreversibility of the process. Moreover, as the reconnecting laments accelerate during a reconnection process, a directional sound pulse is generated in the super uid. Figure 2 shows a set of snapshots of the evolution of some iso-surfaces of the super uid density eld to highlight the sound emission during a reconnection event following the decay of an Hopf link [1]. After the reconnection takes place, here t ' 53=c, a distinctive variation of the bulk density (yellow-to-violet colors) rec emerges at the reconnection point and propagates non-isotropically. Explaining the origin of the asymmetry between the pre-factors A and characterizing the directionality and intensity of the sound pulse are the main scopes of this work. In our reasoning and calculations we use two di erent limits of the GP model: the linear model, essentially the Schroedinger equation where the nonlinear term of GP is 5 (a) (b) FIG. 3. (Colour online) Here p = 1; A = 2; B = 1; C = 2; D = 1;  = =3 and lengths are rescaled by  . (a) An example of the two intersecting iso-surfaces Re( ) = 0 and Im( ) = 0 of eq. (11). (b) The hyperbola resulting from the projection of r r the nodal lines onto the z = 0 plane for the times before, t = 1, and after, t = 1, the reconnection time t = 0. neglected, and the Biot-Savart model, where the compressible degrees of freedom of the super uid are neglected. In a nutshell, our main results follow a simple matching between these two limits and make use of the conservation of total super uid linear momentum and energy, Eqs (4) and (3), respectively. The work is organized as follows. Section II describes exhaustively all the possible geometrical con gurations taken by two reconnecting laments: subsections II A and II B explore the range of parameters of the wave-function about a reconnection event, subsection II C relates these wave-function's parameters to a set of geometrical parameters for the vortex laments, nally subsection II D introduces a useful parameterization of the laments in terms of the geometrical parameters. Section III is devoted to the variations in time of the linear momentum and energy of the laments during a reconnection, and relate these to the emission of the sound pulse: subsection III A set the framework for such study, while subsections III B and III C contain the detailed calculations of the linear momentum di erence and energy variations, respectively. Section IV is left for the conclusions and future perspectives. II. THE GEOMETRY OF THE RECONNECTING FILAMENTS A. The reconnecting wave-function of the laments As previously obtained in [22, 25], about a reconnection event, the nonlinear term of the Gross-Pitaevskii (GP) model can be neglected, as the super uid density vanishes at the vortex core. The dynamics is thus driven by the linear Schroedinger equation i@ = r : (10) Without any loss of generality we set the reconnection time t = 0 and let the reconnection point be the origin of our reference frame. The most general second-order-polynomial wave-function initial condition having two nodal lines intersecting at the origin results in 2 2 2 2 1 A(x cos  + y sin ) + B(x cos  + y sin ) Cx + Dy (x; y; z) = p z + i z ; (11) 5=2 2 2 where p = 1, (A; B; C; D) 2 R and  2 [0; ] are dimensionless parameters, we will call them wave-function parameters in what follows.  > 0 is a characteristic length. The nodal lines of Eq. (11) can be easily identi ed by intersecting the two iso-surfaces Re( ) = 0 and Im( ) = 0: Figure 3(a) shows an example of this intersection when r r setting the parameters to p = 1; A = 2; B = 1; C = 2; D = 1;  = =3 and where lengths are rescaled by  . 6 Under the linear Schroedinger operator in Eq. (10) the solution in time reads it r 2 (x; y; z; t) = e (x; y; z) = 1 + it r (x; y; z) : (12) r r After some tedious algebra (see the Mathematica notebook available as Supplemental Material [36]), we obtain that the evolution of the wave-function nodal lines Re( ) = 0 and Im( ) = 0 results in the equations 2 2 (A + B)(x + y ) + (A B) [(x y)(x + y) cos(2) + 2xy sin(2)] C + D z = t (13) 4 4p and 2 2 Cx + Dy A + B z = + t ; (14) 2 4p respectively. By simplifying the z-dependence in Eqs. (13) and (14), one nds that the projection of the nodal lines onto the z = 0 plane satis es the equation A + B + C + D 2 2 (A + B 2C )x + (A + B 2D)y + (A B) [(x y)(x + y) cos(2) + 2xy sin(2)] = t : (15) For a suitable choice of the dimensionless parameters, this relation identi es a hyperbola; an example of it is shown in Fig. 3(b) where, again, p = 1; A = 2; B = 1; C = 2; D = 1;  = =3 and lengths are rescaled by  . It is important to notice that the (x; y) axes can always be rotated in order to ensure that the hyperbola asymptotes are mirrored with respect to the two axes. This property simply re ects the fact that Eq. (15) can be re-expressed in its normal form by a suitable rotation that depends on the chosen value of the wave-function parameter . For the sake of simplicity, we set the speci c value 0: (16) Note however that all the following considerations are actually valid for any value of  (see Appendix). The projection of the nodal lines onto the z = 0 plane now results in C A A + B + C + D 2 2 x + y = t ; (17) B D 2(B D)p by assuming, from now onwards, that B 6= D. Finally, by simplifying the y-dependence in Eqs. (13) and (14), we nd that the projection of the nodal lines onto the y = 0 plane satis es the equation BC AD D(C + D) + B(A + B) z = x + t ; (18) 2(B D) 4(B D)p that is a parabola that shifts along the z-axis at constant speed. B. Region of validity of the other wave-function parameters We denote by R and R the sets of points of the two vortex laments before reconnection and by R and 1 2 1 R the ones after reconnection. Without loss of generality we may assume: that (i) about the reconnection point, + + R  fy > 0g and R  fy < 0g whereas R  fx < 0g and R  fx > 0g; and that (ii) the orientation of the 1 2 1 2 vorticity follows the arrows as in the sketch displayed in Fig. 4(a). In order to nd the range of the admissible values of the wave-function parameters p 1 and (A; B; C; D) 2 R of eq. (11), we thus need to impose the following validity conditions. Existence of the hyperbola. At the reconnection time t = 0 we want the hyperbola asymptotes y = (C A)=(B D) x set by Eq. (17) to be real (in other words we want the equation to describe a hy- perbola and not an ellipse). This reduces to the condition C A 0 : (19) B D 7 y/ζ z/ζ (a) (b) concavity − − t < 0, R t < 0, R 1,2 1,2 parameter Λ + + t > 0, R t > 0, R 1,2 1,2 + + φ φ x/ζ x/ζ FIG. 4. (Colour online) Sketch of the reconnecting laments projected (a) onto the z = 0 plane and (b) onto the y = 0 plane. Convention on the location of the laments. Our convention adopts that, about the reconnection point, the + + positions of the laments satisfy R  fy > 0g and R  fy < 0g whereas R  fx < 0g and R  fx > 0g. 1 2 1 2 Hence, by evaluating eq. (17) at times t < 0 and t > 0 we obtain that the following conditions, respectively, must hold (B D) p (C + D + A + B) < 0 and (C A) p (C + D + A + B) < 0 : (20) Convention on the vorticity orientation of the laments. The orientation of the laments can be evaluated by computing the pseudo-vorticity ! = rRe( )rIm( ) of the wave-function in Eq. (12) at its nodal lines [18]. In order to impose the vorticity orientations as the arrows sketched in Fig. 4(a) the following conditions 8 8 > p = 1 > p = 1 < < B < D or B > D (21) > > : : A > C A < C must be realised. C. The geometrical parameters of the laments As shown in the previous subsection, provided that a suitable choice of the wave-function parameters is taken about the reconnection event, the projection of the laments onto the z = 0 plane corresponds to the hyperbola represented in Eq. (17). We can de ne the macroscopic reconnecting angle  as the angle formed by the hyperbola asymptotes when considering the laments after the reconnection, as shown in Fig. 4(a). By simple geometrical considerations, and using Eq. (17), we nd that this angle is related to the wave-function parameters as C A  C A + 2 = 2 arctan () tan = : (22) B D 2 B D Also, about the reconnection event, the projection of the laments onto the y = 0 plane is the parabola found in Eq. (18). Its concavity results simply in BC AD  BC AD = ; given the dimensionless concavity parameter  = . (23) (B D)  (B D) A sketch of the projection displaying the role of the concavity parameter is shown in Fig. 4(b). 8 FIG. 5. (Colour online) Region of the admissible parameters (B; D) for p = 1 (blue areas) and p = 1 (red areas), given a speci c choice of the geometrical parameters ( ; ). We can express these two geometrical parameters versus the wave-function parameters characterizing the wave- function initial condition. A possible choice is the following > A = B tan + : (24) > 2 C = D tan + In terms of the geometrical parameters, the region of validity given by Eqs. (16), (19), (20), and (21) results in 2 < tan 1 (B + D) \ (p = 1 \ D > B) [ (p = 1 \ D < B) : (25) This set of validity conditions are better summarized using the diagrams in Fig. 5, where we have set T = tan 1 (26) to simplify the notation. For a given choice of the geometrical parameters  2 (0; =2) [ (=2; ) and  2 R, two important remarks should be made: rstly, it is always possible to nd a set of values (p; B; D) that satis es the validity conditions in Eq. (25); secondly, the area spanned by the parameters (B; D) is unbounded. The limit ! =2, causing T ! 0, needs a special consideration: in this case, using Eq. (25), we nd that  2 R , that is the concavity can only be negative, otherwise the convention on the positions of the laments is not satis ed. Actually, a completely symmetrical reconnection ( = =2) can not be realized in a fully planar con guration ( = 0) with a quadratic wave function. 9 D. Parametrization of the laments The vortex laments can be parametrized in terms of the reconnecting angle  and the concavity parameter  as follows: (t)   (t) R (`; t) = cot sinh (`); cosh (`); z (`; t) (27) 2 2 2 (t)   (t) R (`; t) = cot sinh (`); cosh (`); z (`; t) (28) 2 2 2 + + + (t)  (t) + + R (`; t) = cosh (`); tan sinh (`); z (`; t) (29) 2 2 2 + + + (t)  (t) + + R (`; t) = cosh (`); tan sinh (`); z (`; t) ; (30) 2 2 2 where ` 2 (1; +1) is the parameter spanning the entire length of the laments (note however that ` does not correspond to the arc-length parametrization of the laments). Here 2 tan ( =2) 1 (B + D) 4 A (t) = A jtj ; A = ; A = ; (31) (B D)p tan( =2) and z (`; t) can be expressed by using eq. (18) and the square of the x-component of the laments before and after the reconnection, respectively, resulting in z (`; t) = z (t) +  (t) cot sinh(`) 8 2 ; (32) + + z (`; t) = z (t) +  (t) cosh(`) where 2 2 2 + (B + D ) tan ( =2) 1 + (B + D) z (t) = t : (33) 4(B D)p This choice of parametrization is particularly useful as one can see easily that the distance between the laments before and after the reconnection is given by jR (` = 0; t) R (` = 0; t)j =  (t) ; (34) 1 2 1=2 which immediately demonstrates the scaling  / jtt j [22, 25]. Furthermore, we can notice that the ratio between the rate of approach and separation is independent on the concavity parameter  and reads + + A = = cot : (35) A 2 Finally, we can understand by looking at Eqs. (31) and (33) that the choice of the parameters (B; D) only in uences the time-scale of the reconnection process. III. MOMENTUM AND ENERGY TRANSFERS DURING A RECONNECTION As it has been observed in previous works [24, 29, 37] and clearly displayed in Fig.2, when a vortex reconnection takes place in a quantum uid, a sound pulse is excited. Energy and momentum are thus transferred from the incompressible to the compressible degrees of freedom of the super uid in an irreversible manner. The aim of this section, and the main result of this work, is to develop an asymptotic matching theory that allows for quantifying such energy and momentum exchanges. In the GP model, the total energy (3) and linear momentum (4) are conserved during the reconnection process, as they are integrals of motion. Well before a reconnection event, practically only the presence of vortex laments 10 contribute to the invariants, whereas after reconnection, both laments and compressible waves add up their con- tributions to them. If one is able to estimate the contribution of the laments, then the contribution of density waves can be deduced using the conservation of the invariants. In the case of energy, such decomposition can be easily done numerically by splitting the kinetic energy term into the incompressible and compressible parts [17]. Such measurements were performed in [1] and will be reproduced below in our discussions. Our analytical treatment of the problem is as follows. When the laments are far from each other, i.e. (t)  , their dynamics of mainly driven by the Biot-Savart model. In that region we might use the vortex lament description to evaluate their energy and momentum of the super uid. On the other hand, when  (t)  , the dynamics is governed by the linear regime given by the Schroedinger equation. Vortices then reconnect following the laws described in the preceding section. We thus describe the reconnection matching, sketched in Fig.6, as follows. Before reconnection, some Biot-Savart dynamics leads to the pre-reconnection input con guration R for the laments 1;2 about the vortex reconnection point. The Biot-Savart description is assumed to be valid down to a distance  =  , lin where  is of the order of a few healing lengths. From there, the laments are driven by the Schroedinger equation lin allowing them to reconnect. After the reconnection, this linear regime is valid until the vortices separate up to + + a distance  =  . The linear evolution thus provides the output post-reconnection con guration R for the lin 1;2 laments. From there onwards, the dynamics is again governed by the Biot-Savart model. Note that the linear regime time (t) ⇠ (t) ⌧ ⇠ (t) ⇠ <latexit sha1_base64="Ya0DY5y/uo686nVffTbTb0EXnpo=">AAACEnicbVDLTsJAFJ36RHxVXLqZSExwIWnRRJdENy4xkUdCkUynA0yYdpqZWwNp+AvXbvUb3Bm3/oCf4F84BRYKnuQmJ+fcm3vv8WPBNTjOl7Wyura+sZnbym/v7O7t2weFhpaJoqxOpZCq5RPNBI9YHTgI1ooVI6EvWNMf3mR+85EpzWV0D+OYdULSj3iPUwJG6toFL2ACyMNZCU69fh97I961i07ZmQIvE3dOimiOWtf+9gJJk5BFQAXRuu06MXRSooBTwSZ5L9EsJnRI+qxtaERCpjvp9PYJPjFKgHtSmYoAT9XfEykJtR6HvukMCQz0opeJ/3qZAlIKvXAA9K46KY/iBFhEZ/t7icAgcZYPDrhiFMTYEEIVNy9gOiCKUDAp5k027mISy6RRKbvn5crdRbF6PU8ph47QMSohF12iKrpFNVRHFI3QM3pBr9aT9Wa9Wx+z1hVrPnOI/sD6/AHSH518</latexit> <latexit sha1_base64="x3GdmI1tmJLmLHqbEY/335CXFgI=">AAACFHicbVDLTgJBEJzFF+Jrfdy8TCQmeCG7aKJHohePmMgjYZHMDgNMmNnZzPQakfAbnr3qN3gzXr37Cf6Fs8BBwUo6qVR1p7srjAU34HlfTmZpeWV1Lbue29jc2t5xd/dqRiWasipVQulGSAwTPGJV4CBYI9aMyFCweji4Sv36PdOGq+gWhjFrSdKLeJdTAlZquwdBhwkgd0EsC3ASCIGDB952817RmwAvEn9G8miGStv9DjqKJpJFQAUxpul7MbRGRAOngo1zQWJYTOiA9FjT0ohIZlqjyfVjfGyVDu4qbSsCPFF/T4yINGYoQ9spCfTNvJeK/3qpAkoJM3cAdC9aIx7FCbCITvd3E4FB4TQh3OGaURBDSwjV3L6AaZ9oQsHmmLPZ+PNJLJJaqeifFks3Z/ny5SylLDpER6iAfHSOyugaVVAVUfSIntELenWenDfn3fmYtmac2cw++gPn8wfsXp6m</latexit><latexit sha1_base64="vqgSSZr8nhlK16ec2D1Xx6GjNX4=">AAACEnicbVDLSsNAFJ34rPUV69LNYBEqQkmqoMuiG5cV7AOaWiaTaTt0kgkzN9IS+heu3eo3uBO3/oCf4F84abvQ1gMXDufcy733+LHgGhzny1pZXVvf2Mxt5bd3dvf27YNCQ8tEUVanUkjV8olmgkesDhwEa8WKkdAXrOkPbzK/+ciU5jK6h3HMOiHpR7zHKQEjde2CFzAB5OGsBKdev4+9Ee/aRafsTIGXiTsnRTRHrWt/e4GkScgioIJo3XadGDopUcCpYJO8l2gWEzokfdY2NCIh0510evsEnxglwD2pTEWAp+rviZSEWo9D33SGBAZ60cvEf71MASmFXjgAeledlEdxAiyis/29RGCQOMsHB1wxCmJsCKGKmxcwHRBFKJgU8yYbdzGJZdKolN3zcuXuoli9nqeUQ0foGJWQiy5RFd2iGqojikboGb2gV+vJerPerY9Z64o1nzlEf2B9/gDO1Z16</latexit> 1,2 1,2 <latexit sha1_base64="8fNreUZBNgXcmMFVCQLOMbLg4vI=">AAACD3icbVDLSsNAFJ34rPXRqEs3g0VwoSWpgi6LblxWsQ9oY5hMJ+3QyUyYmQgl5CNcu9VvcCdu/QQ/wb9w0mahrQcuHM65l3vvCWJGlXacL2tpeWV1bb20Ud7c2t6p2Lt7bSUSiUkLCyZkN0CKMMpJS1PNSDeWBEUBI51gfJ37nUciFRX8Xk9i4kVoyGlIMdJG8u1K2g9CeJc9nPqpe1LPfLvq1Jwp4CJxC1IFBZq+/d0fCJxEhGvMkFI914m1lyKpKWYkK/cTRWKEx2hIeoZyFBHlpdPDM3hklAEMhTTFNZyqvydSFCk1iQLTGSE9UvNeLv7r5YoWgqm5A3R46aWUx4kmHM/2hwmDWsA8HDigkmDNJoYgLKl5AeIRkghrE2HZZOPOJ7FI2vWae1ar355XG1dFSiVwAA7BMXDBBWiAG9AELYBBAp7BC3i1nqw36936mLUuWcXMPvgD6/MHUtScIA==</latexit> <latexit sha1_base64="HI63mBs/BbivdJEnIYA94mzeKGw=">AAACD3icbVDLSsNAFJ34rPXRqEs3g0UQlJJUQZdFNy6r2Ae0MUymk3boZCbMTIQS8hGu3eo3uBO3foKf4F84abPQ1gMXDufcy733BDGjSjvOl7W0vLK6tl7aKG9ube9U7N29thKJxKSFBROyGyBFGOWkpalmpBtLgqKAkU4wvs79ziORigp+rycx8SI05DSkGGkj+XYl7QchvMseTvzUPa1nvl11as4UcJG4BamCAk3f/u4PBE4iwjVmSKme68TaS5HUFDOSlfuJIjHCYzQkPUM5iojy0unhGTwyygCGQpriGk7V3xMpipSaRIHpjJAeqXkvF//1ckULwdTcATq89FLK40QTjmf7w4RBLWAeDhxQSbBmE0MQltS8APEISYS1ibBssnHnk1gk7XrNPavVb8+rjasipRI4AIfgGLjgAjTADWiCFsAgAc/gBbxaT9ab9W59zFqXrGJmH/yB9fkDT5KcHg==</latexit> input output Biot-Savart model Schrödinger equation Biot-Savart model FIG. 6. (Colour online) A sketch of a reconnection process and matching asymptotics. When vortices are far apart their dynamics if governed by the Biot-Savart equation, whereas when they are about to reconnect the process is driven by the Schroedinger equation. corresponds only to the dynamics inside the orange-greenish box in Fig. 6. We can thus consider the linear regime as the regularization mechanism allowing vortex reconnections in the Biot-Savart model. Summarizing, in order to compute the di erences before and after the reconnection in the incompressible energy and momentum of the super uid, we use the theoretical description R for the laments given in Eqs. (27-30). 1;2 Namely, we use such parametrization when the distance is =  =  &  ; (36) lin as illustrated in Figs. 7(a) and 7(b), respectively. Note that the assumption that the linear regime description may be still valid at distances beyond the healing length  is justi ed by numerical evidence [25]. A. The cylindrical region of integration As detailed in the following, the calculations of the linear momentum and energy of the vortices involve the integration over the full length of the laments. As we are interested in their di erences, to simplify the problem we will consider only the segments of the laments which lie inside the cylinder of circle of radius R centered at the origin and having the cylindrical axis parallel to the z axis; the projection of the cylinder onto the z = 0 plane, corresponding to the circle of radius R centered at the origin, is also sketched in Fig. 7. 11 (a) (b) t < 0 t > 0 lin lin R R FIG. 7. (Colour online) Projections of the reconnecting laments and the cylinder onto the plane z = 0 before the reconnection (a) and after the reconnection (b). Here the reconnection angle is  = =3. The vortex laments lie inside the cylinder when their parametrization satis es j`j  L before and after the reconnection, respectively, given 8 9 h ih i > 2 2 > 2 2 > > 8(R= ) + (A 1) + 2 4 (R= ) 1 4 (R= ) + A < lin lin lin = r r L (R= ) = ln (37) lin > > 2 A + 1 > r > : ; 8 9 h ih i > 2 2 > 2 2 2 > > 8A (R= ) + (1 A ) + 2A 4 (R= ) 1 4A (R= ) + 1 < lin r lin lin = r r r L (R= ) = ln : (38) lin 2 > A + 1 > > r > : ; Physically, it is natural to assume that R is larger, or much larger, than  . Note that, keeping A nite, we have lin r lim L (R= ) = ln(R= ) + ln + : : : lin lin R= !1 lin 1 + A (39) lim L (R= ) = ln(R= ) + ln + ln(A ) + : : : lin lin r R= !1 lin 1 + A B. Linear momentum di erence Following L. Pismen [38], the linear momentum of an incompressible and inviscid uid with lamentary vorticity eld of intensity , that is within the framework of the Biot-Savart model, reads I Z @R P = R dR = p (`) d` ; where p (`) = R (40) l l l 2 2 @` laments is the momentum density (per unit of lament length) and the integration interval follows the parametrization of the laments. In our case ` 2 (1; +1) and there are only two laments before and two laments after the reconnection. The incompressible super uid momenta density before and after reconnection can be therefore computed by using Eqs. (27-30). Due to the particular symmetries of the reconnecting con guration, their expressions are rather simple. 12 Interestingly they are independent of  and `, and read ! ! " # 2 2 2 L (R= ) lin A  A  A r r r p = 0; 0; =) P ( ; A ; R= ) = 0; 0; d` = 0; 0; L (R= ) r lin lin l l 4 4 2 L (R= ) lin ! ! " # 2 2 2 L (R= ) + + lin + + + + + p = 0; 0; =) P ( ; A ; R= ) = 0; 0; d` = 0; 0; L (R= ) : r + lin l l 4A 4A + 2A r r r L (R= ) lin (41) These results immediately tell us that the di erence of the linear momentum of vortex laments during a reconnection is non zero only along the z-direction and in the limit of large R= is given by lin 1 + A + r + P ( ; A ; R= ) = P ( ; A ; R= ) P ( ; A ; R= ) / = 2 csc  : (42) l;z lin r lin lin r lin lin r lin l;z l;z This is a remarkable result: following our convention, refer again to Fig. 4 and for all ( ; A ; R= ), the incom- lin r lin pressible super uid linear momentum changes during a reconnection only along the z-axis and, most importantly, its variation is always negative with a global maximum for  = =2. As a consequence, assuming that the total linear momentum within the cylinder of radius R remains conserved during the reconnection process, a density/phase compressible excitation, a sound pulse, must carry the missing momentum P = P ; (43) pulse l that is, must certainly move along the positive direction of the z-axis. C. Energy di erence In rst approximation, the (kinetic) energy of an incompressible inviscid uid consisting on a collection of vortex laments is proportional to the total length L of the vortex con guration. This result arises from the local induction approximation (LIA) of the Biot-Savart model. The (kinetic) incompressible energy of the super uid, in the LIA assumption, is thus given by @R E (L) = T L = T d` ; (44) LIA @` laments 2 2 where T =  log(L =a )=(4 ) is the so-called vortex line tension [38, 39], with L is a characteristic length order 0 0 0 0 of the mean radius and a is the vortex core size. The precise de nition of the vortex tension is not relevant for the next considerations. The incompressible super uid energy, before and after the reconnection, thus results in L (R= ) lin @R @R 1 2 E (A ; =;  ; R= ) = T + d` ; (45) r lin lin LIA @` @` L (R= ) lin and the energy di erence during the reconnection is simply E(A ; =;  ; R= ) = E (A ; =;  ; R= ) E (A ; =;  ; R= ) / L(A ; =;  ; R= ) ; (46) r lin lin r lin lin r lin lin r lin lin where Z + Z L (R= ) + + L (R= ) lin lin @R @R @R @R 1 2 1 2 L(A ; =;  ; R= ) = + d` + d` (47) r lin lin + @` @` @` @` L (R= ) L (R= ) lin lin is the di erence of the total length of the laments during the reconnection process. The integrals in (47) can be evaluated analytically for  = 0 and the nal result reads 2i 1 lin 2 + 2 L(A ; = = 0;  ; R= ) = A Ell iL (R= ) 1 + Ell iL (R= )jA + 1 ; (48) r lin lin lin lin r r A A r 13 (a) (b) 1 2 3 4 1 2 3 4 -1 -1 -2 -2 FIG. 8. (Colour online) Plots of the di erence in the total length of the laments during the reconnection process when  = 0 and its limiting case when R=  1, that is eq.s (48) and (49), versus A for di erent values of R= ; here  = 1. lin r lin lin where Ell(j) is the incomplete elliptic integral of the second kind. The behavior of eq. (48) for di erent values of R= is shown in Fig. 8(a). Note that L appears to be monotonically decreasing versus A and to saturate for large lin r R= . This latter property can be corroborated by taking the limit R= ! 1 keeping A nite: using eq. (39) one lin lin r gets ( " # ) ( " # )! 2i 4 (R= ) 1 4A (R= ) lin lin r lin 2 2 L(A ; = = 0; R= ! 1) = A Ell i ln p 1 + Ell i ln p jA + 1 : r lin r r 2 2 A A r 1 + A 1 + A r r (49) Figure 8(b) shows its behavior versus A for larger and larger values of R= ; one has however to be cautious in using r lin this limiting expression, as Eq. (49) presents an unphysical behavior when A ! 0 which is absent in the original Eq. (48). Also, we can note that due to symmetry reasons, the two following properties hold L(A = 1; =;  ; R= ) = 0 (50) r lin lin and L(A ; =;  ; R= ) = L(A ;=;  ; R= ) : (51) r lin lin r lin lin In the general case where  6= 0 and nite, the integrals in (47) have to be computed numerically. As for  = 0, L appears to converge in the limit R= ! 1; also, as shown in Figs. 9(a) and 9(b), we observe numerically that lin L(A 6= 0;jj= ! 1;  ; R= ) = 0 : (52) r lin lin Hence, let us note that the value of L(A ; ;  ; R= ) is always bounded by Eqs. (48) and (52). r lin lin Using the same rationale as in the linear momentum di erence subsection, we now assume that the total (kinetic) super uid energy within the cylinder is conserved during the reconnection. Hence, to enforce this conservation, a density/phase compressible excitation, a sound pulse, must be created with energy fraction E E (A ;j=j;  ; R= ) L(A ;j=j;  ; R= ) pulse LIA r lin lin r lin lin = = ; (53) E E L tot LIA;0 0 given E the total super uid energy, E is the LIA energy of the initial con guration of the laments, and L tot LIA;0 0 the total initial length of the laments. As L appears to be monotonically decreasing versus A and equal to zero at A = 1, then E is positive only for A > 1. This is a very important result, as it shows that, under our r pulse r assumptions, reconnections cannot happen for A  1, that is for   =2, because a sound pulse with negative energy should be created, which is clearly unphysical. 14 (a) (b) 1.5 1.5 1.0 1.0 0.5 0.5 1 2 3 4 -0.5 1 2 3 4 -1.0 -0.5 FIG. 9. (Colour online) Plots of the di erence in the total length of the laments during the reconnection process, that is a numerical estimation of eq (47), versus A for di erent values of jj= ; here  = 1 and R= = 2. r lin lin IV. DISCUSSIONS AND FUTURE PERSPECTIVES Summarizing, following the calculations presented in Section III, we are able to estimate the linear momentum and energy of the sound pulse emitted during a reconnection event. These reads P = [0; 0;P ( ; A ; R)] pulse l;z lin r ; (54) L(A ;jj=;  ; R= ) r lin lin E = E pulse tot where the details of the functions can be found in Eqs (42) and (47), respectively. In particular, we underline that for all (jj=;  ; R= ) lin lin P > 0 ; for all A pulse;z r ; (55) E > 0 ; for A > 1 pulse r explain the origin of the asymmetry observed in the distribution of the the pre-factors A , and the directionality observed in the sound pulse emission. Let us now to discuss further the nature of the sound pulse. As veri ed in reference [1] and apparent in Fig.2, the pulse indeed propagates in a well de ned direction. Figure 10(a) shows such a sound pulse, rescaled by the bulk density, propagating along the positive z-direction after a reconnection characterized by the geometrical parameter A = 1:67. The pulse is plotted versus time in Fig.10.a and versus the retarded time in Fig.10.b for di erent values of the z-coordinate. We can observe that the pulse appears to move slightly slower than the speed of sound c. The signal shows some dispersive e ects while time advances, an evidence that the pulse probably contains more than one Bogoliubov perturbation, perhaps including high wave numbers. Also, the depth of its trough  decays as min the pulse propagates following the scaling / z typical of a three-dimensional wave signal originating from a point source, as depicted in Fig. 10(c). It is however still unclear if the the pulse consists of a simple linear superposition of Bogoliubov perturbations, or if a fully nonlinear subsonic coherent structure, like a Robert{Jones solitary wave [40], is also superimposed. A complete analysis on the spectrum of the sound pulse and the possible presence of coherent structures is left for future works. Concerning the energy transferred from the vortices to the pulse, we present in Fig.11a-b the comparison between our theoretical prediction, Eq. (53), and the GP reconnection data obtained in [1] for di erent choices of  . The region lin of validity of the theory, the colored regions depicted in Fig. 11, represents all the accessible values of the concavity parameter  2 R, with boundaries obtained from Eqs (48) and (52). We note that the speci c value of the energy di erence in Eq. (53) depends on the choice of  , but the conclusion that reconnection with A < 1 are unlucky to lin r occur, remains valid. Moreover, the fact that no energy exchange takes place for a symmetric reconnection A = 1, 0:71 suggests to plot E versus A 1. Such plot is displayed in Fig.11.c, where the scaling E  (A 1) , wav r wav r 15 1.005 1.01 1.005 1 1 1 0.995 0.995 0.995 0.99 0.99 0.99 0.985 0.99 0.985 0.985 0.98 0.98 0.98 0.98 0.975 0.975 0.975 0.97 0.97 0 20 40 -20 0 20 20 25 30 35 40 0 50 FIG. 10. (Color online) Traveling pulse emitted during a reconnection with A = 1:67. (a) Density around the pulse as function of time for di erent values of the distance z to the reconnection plane. (b) Same as (b), but as a function of the retarded time. (c) Depth of the traveling pulse trough  at di erent values of z compared with the 1=z decay law predicted form min an acoustic pulse emitted from a point source. The theoretical formula is obtained by imposing the 1=z decay and matching the measured value of  at z = 20, the smallest value of z at which the pulse was measured. The speed of sound is c and min 0 = =c. -1 -2 -3 -4 0.3 0.5 1 2 3 4 5 FIG. 11. (Colour online) Energy radiated during reconnections as a function of A = A =A for di erent values of  . r lin Numerical data from GP simulations obtained in [1] is confronted to our theoretical prediction in Eq. (53). (a)  = 4. (b) lin = 8. (c) LogLog plot of energy as a function of A 1 for di erent values of  . lin r lin obtained by t, is clearly observed for at least one decade and all values of  . A more accurate theory, providing lin such scaling exponent is out of the scope of this work. It is interesting to note that in a recent experimental work [41], P. Svancara and M. La Mantia observed that in super uid helium, the skewness of Langrangian velocity increments exhibits the opposite trend with respect to the one of classical ows, namely, in super uids particles accelerate on average faster than they decelerate. The authors interpreted their ndings as a consequence of quantum vortex reconnections because particles could probe individual vortex reconnections. This interpretation is consistent with the recent numerical calculations reported in [42], where quantum vortex reconnections mediated by particles were studied. In that work, the authors observed that in addition to the momentum exchange between the vortices and sound waves, there is also a net transfer towards particles that incurs on a fast acceleration of particles at the reconnection event. We should also remark on the possible role played by the center-line helicity H during a reconnection event. As rst observed in [30], the evolution of H shows a sudden drop (but still being continuous) during a reconnection if the initial linking number between the vortex laments is non-zero. This is certainly the case analyzed in [1], where the initial con guration of the laments, an Hopf link, has linking number equal to 2. The temporal evolution of a Hopf link, for a case with A = 3:5, is displayed Fig. 12.a, where the helicity drop is clearly visible. One might be tempted to think that the sudden drop in the center-line helicity H is related to the properties of the reconnecting laments, for example their parameter A . Figure 12.b shows that H as a function of the parameter A . Numerical data r c r do not correlate as well as the energy and no clear trend is observed. We can however speculate that the center-line helicity drop may in uence the amplitude and polarization of the Kelvin waves forming on the vortex laments after the reconnection process. This research proposal constitutes another interesting direction for future works. In conclusion, our theoretical results help to understand not only how vortex reconnections take place in quantum uids, but also why they do occur. Albeit the GP model is time reversible, the vortex reconnection process presents 16 -0.2 -0.4 -0.6 0 1 2 3 4 FIG. 12. (Color online) a) Temporal evolution of central line helicity H for a reconnection having A = 3:5. The green zone c r correspond to the time interval de ned by  (t ) =  = 8. The red dot is the reconnection time. a) Drop of central line lin + + helicity H = H (t ) H (t ) as a function of A = A =A . is the circulation of the laments. c c c r a time asymmetry so that the system can naturally transfer part of the (kinetic) incompressible energy into its compressible counterpart. In some sense, this observed temporal asymmetry can be interpreted as consequence of the system being in an out-of-equilibrium stage and reconnections being a fast route towards reaching thermal equilibrium. It will be very interesting to study vortex reconnections at nite temperatures, where a thermal bath of Bogoliubov modes is present, to see if whether this asymmetry is reduced or destroyed completely. Furthermore, the situation can be di erent when open conditions, like external forcing and damping terms acting at di erent length scales, are introduced. In a fully developed turbulent state, uctuations could provide enough energy to allow reconnections with A < 1, but one might still expect skew distributions towards A > 1, as dissipation of turbulent ows do not vanishes r r in the limit of in nite Reynolds numbers because of the dissipative anomaly of turbulence [43]. This is undoubtedly, another interesting direction of research for future studies. Finally, it is important to remark that the theoretical predictions obtained in this work may complement future experiments in several physical systems. In Bose{Einstein condensates made of dilute gases, experimentalists are now able to observe the real-time dynamics of few reconnecting vortex laments [44] and are now working to create reproducible reconnections [45]. Even if measuring the incompressible and compressible energy components of the super uid is still a huge experimental challenge, assessing the directionality and shape of the sound pulse should be nowadays achievable using, for example, density imaging or Bragg spectroscopy. In super uid liquid helium-4 experiments, the challenge is to track single vortex reconnections, especially in the limit of very low temperatures where the normal component is negligible. However, measuring the production of compressible energy excitations (and/or rotons) during many reconnections occurring in a turbulent tangle might be achievable [30, 46, 47]; in order to be able to theoretically estimate this production rate, a statistical extension to our detailed analysis, including possibly nite temperature e ects, might be needed. Ultimately, let us point out there are impressive similarities between vortex reconnections occurring in quantum uids and classical uids at high Reynolds numbers, especially when the classical vortex tubes possess an almost hollow core [48, 49]. It would be therefore very interesting to measure the pre-factors A , reconnection angle  and concavity parameter  in the classical uid experiments and quantitatively compare them with our theoretical predictions. Appendix A: Choice of system of coordinates and value of parameter To simplify the analysis of the wave-function parameters and the following calculations we have set the parameter = 0. Actually, one can always perform a rotation in the x y plane such that Eq.(15) becomes Eq.(17) after a rede nition of the coecient A; B; C; D. Once this change of variable is performed, Eq.(14) will become in the new variables 2 2 C x + C xy + C y A + B xx xy yy z = + t ; (A1) 2 4p where C ; C and C are new coecient depending on the old ones. The analisys concerning the region of xx xy yy parameteres involves only the coecients of the hyperbola, therefore it remains unchanged. The concavity parameter 17 can not be de ned so straighforwardly as now the projection on the x = 0 or y = 0 are not simple parabolas. However, the concavity increases with the magnitud of the coeciens C ; C and C . xx xy yy The momentum di erence (42) remains unchanged if (A1) is taken into consideration for the vortex parametrization. Finally, the expression for the di erence of vortex length are slightly more complicated, but the property that L  0 for A  1 and L  0 for A  1 remains valid. r r ACKNOWLEDGMENTS The authors acknowledge fruitful discussions with J. Hannay that led to the considerations of the most general normal form discussed in the Appendix of this manuscript. The authors acknowledge L. Galantucci for providing data displayed in Fig.1 and A. Villois for supplying the raw data used in Fig.10. G.K., D.P. were supported by the cost- share Royal Society International Exchanges Scheme (IE150527) in conjunction with CNRS. D.P. was supported by the EPSRC First Grant scheme (EP/P023770/1). G.K. and D.P. acknowledge the Federation Doeblin for supporting DP during his sojourn in Nice. G.K. was also supported by the ANR JCJC GIANTE ANR-18-CE30-0020-0.1 and by the EU Horizon 2020 research and innovation programme under the grant agreement No 823937 in the framework of Marie Skodowska-Curie HALT project. Computations were carried out at M esocentre SIGAMM hosted at the Observatoire de la C^ ote d'Azur and on the High Performance Computing Cluster supported by the Research and Specialist Computing Support service at the University of East Anglia. Part of this work has been presented at the workshop \Irreversibility and Turbulence" hosted by Fondation Les Treilles in September 2017. The authors acknowledge Fondation Les Treilles and all participants of the workshop for the frightful scienti c discussions and support. D.P. acknowledges J. Hannay for his frequent visits to Norfolk that always lead to fruitful conversations. [1] A. Villois, D. Proment, and G. Krstulovic, Irreversible dynamics of vortex reconnections in quantum uids, arXiv:2005.02048 (2020). [2] P. G. Sa man, Cambridge Monographs on Mechanics (Cambridge University Press, Cambridge, 1993). [3] E. R. Priest, Heating the solar corona by magnetic reconnection, in Plasma Astrophysics And Space Physics: Proceedings of the VIIth International Conference held in Lindau, Germany, May 4{8, 1998 , edited by J. Buc  hner, I. Axford, E. Marsch, and V. Vasyliun  as (Springer Netherlands, Dordrecht, 1999) pp. 77{100. [4] S. Kida and M. Takaoka, Vortex reconnection, Annual Review of Fluid Mechanics 26, 169 (1994), http://dx.doi.org/10.1146/annurev. .26.010194.001125. [5] E. Fonda, D. P. Meichle, N. T. Ouellette, S. Hormoz, and D. P. Lathrop, Direct observation of kelvin waves excited by quantized vortex reconnection, Proceedings of the National Academy of Sciences 111, 4707 (2014). [6] S. Sera ni, L. Galantucci, E. Iseni, T. Bienaim e, R. N. Bisset, C. F. Barenghi, F. Dalfovo, G. Lamporesi, and G. Ferrari, Vortex reconnections and rebounds in trapped atomic bose-einstein condensates, Physical Review X 7, 021031 (2017). [7] F. Hussain and K. Duraisamy, Mechanics of viscous vortex reconnection, Physics of Fluids 23, 021701 (2011). [8] Xue Zhike, Yan Xiaoli, Cheng Xin, Yang Liheng, Su Yingna, Kliem Bernhard, Zhang Jun, Liu Zhong, Bi Yi, Xiang Yongyuan, Yang Kai, and Zhao Li, Observing the release of twist by magnetic reconnection in a solar lament eruption, Nat Commun 7, http://dx.doi.org/10.1038/ncomms11837 10.1038/ncomms11837 (2016). [9] J. Yao and F. Hussain, A physical model of turbulence cascade via vortex reconnection sequence and avalanche, Journal of Fluid Mechanics 883, A51 (2020). [10] A. Pumir and R. M. Kerr, Numerical simulation of interacting vortex tubes, Physical Review Letters 58, 1636 (1987). [11] D. S. Agafontsev, E. A. Kuznetsov, and A. A. Mailybaev, Development of high vorticity structures and geometrical properties of the vortex line representation, Physics of Fluids, Physics of Fluids 30, 095104 (2018). [12] R. M. Kerr, Swirling, turbulent vortex rings formed from a chain reaction of reconnection events, Physics of Fluids, Physics of Fluids 25, 065101 (2013). [13] R. McKeown, R. Ostilla-M onico, A. Pumir, M. P. Brenner, and S. M. Rubinstein, Turbulence generation through an iterative cascade of the elliptical instability, Science Advances 6, eaaz2717 (2020). [14] L. Pitaevskii and S. Stringari, Bose-Einstein condensation and super uidity, Vol. 164 0191076686 (Oxford University Press, 2016). [15] This is true for mean- eld models of quantum uids, when quantum uctuations are considered this picture is more complicated. [16] J. Koplik and H. Levine, Vortex reconnection in super uid helium, Phys. Rev. Lett. 71, 1375 (1993). [17] C. Nore, M. Abid, and M. Brachet, Decaying kolmogorov turbulence in a model of super ow, Physics of Fluids (1994- present) 9, 2644 (1997). [18] A. Villois, G. Krstulovic, D. Proment, and H. Salman, A vortex lament tracking method for the gross{pitaevskii model of a super uid, Journal of Physics A: Mathematical and Theoretical 49, 415502 (2016). 18 [19] A. Villois, D. Proment, and G. Krstulovic, Evolution of a super uid vortex lament tangle driven by the gross-pitaevskii equation, Phys. Rev. E 93, 061103(R) (2016). [20] M. D. Bustamante and S. Nazarenko, Derivation of the biot-savart equation from the nonlinear schr odinger equation, Phys. Rev. E 92, 053019 (2015). [21] R. Feynman, Chapter fIIg application of quantum mechanics to liquid helium (Elsevier, 1955) pp. 17 { 53. [22] S. Nazarenko and R. West, Analytical solution for nonlinear schr odinger vortex reconnection, Journal of low temperature physics 132, 1 (2003). [23] M. Kursa, K. Bajer, and T. Lipniacki, Cascade of vortex loops initiated by a single reconnection of quantum vortices, Phys. Rev. B 83, 014515 (2011). [24] S. Zuccher, M. Caliari, A. W. Baggaley, and C. F. Barenghi, Quantum vortex reconnections, Physics of Fluids 24, 125108 (2012). [25] A. Villois, D. Proment, and G. Krstulovic, Universal and nonuniversal aspects of vortex reconnections in super uids, Phys. Rev. Fluids 2, 044701 (2017). [26] C. Rorai, J. Skipper, R. M. Kerr, and K. R. Sreenivasan, Approach and separation of quantised vortices with balanced cores, Journal of Fluid Mechanics 808, 641 (2016). [27] L. Galantucci, A. W. Baggaley, N. G. Parker, and C. F. Barenghi, Crossover from interaction to driven regimes in quantum vortex reconnections, Proceedings of the National Academy of Sciences 116, 12204 (2019). [28] S. Rica, Self-similar vortex reconnection, Patterns and dynamics: homage to Pierre Coullet / Formes et dynamique: hommage  a Pierre Coullet, Comptes Rendus M ecanique 347, 365 (2019). [29] M. Leadbeater, T. Winiecki, D. C. Samuels, C. F. Barenghi, and C. S. Adams, Sound emission due to super uid vortex reconnections, Phys. Rev. Lett. 86, 1410 (2001). [30] M. W. Scheeler, D. Kleckner, D. Proment, G. L. Kindlmann, and W. T. M. Irvine, Helicity conservation by ow across scales in reconnecting vortex links and knots, Proceedings of the National Academy of Sciences 111, 15350 (2014), http://www.pnas.org/content/111/43/15350.full.pdf. [31] C. E. Laing, R. L. Ricca, and L. S. De Witt, Conservation of writhe helicity under anti-parallel reconnection, Scienti c reports 5 (2015). [32] P. Clark di Leoni, P. D. Mininni, and M. E. Brachet, Helicity, topology, and kelvin waves in reconnecting quantum knots, Phys. Rev. A 94, 043605 (2016). [33] H. Salman, Helicity conservation and twisted seifert surfaces for super uid vortices, Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences 473, 20160853 (2017). [34] S. Zuccher and R. L. Ricca, Helicity conservation under quantum reconnection of vortex rings, Phys. Rev. E 92, 061001 (2015). [35] L. Galantucci, A. W. Baggaley, N. G. Parker, and C. F. Barenghi, Crossover from interaction to driven regimes in quantum vortex reconnections, Proceedings of the National Academy of Sciences 116, 12204 (2019). [36] D. Proment and G. Krstulovic, Supplemental material for: A matching theory to characterize sound emission during vortex reconnection in quantum uids, Mathematica Notebook available as Supplemental Material (2020). [37] R. M. Kerr, Vortex stretching as a mechanism for quantum kinetic energy decay, Phys. Rev. Lett. 106, 224501 (2011). [38] L. M. Pismen and L. M. Pismen, Vortices in nonlinear elds: from liquid crystals to super uids, from non-equilibrium patterns to cosmic strings, Vol. 100 (Oxford University Press, 1999). [39] E. B. Sonin, Vortex oscillations and hydrodynamics of rotating super uids, Reviews of Modern Physics 59, 87 (1987). [40] C. A. Jones and P. H. Roberts, Motions in a bose condensate. iv. axisymmetric solitary waves, Journal of Physics A: Mathematical and General 15, 2599 (1982). [41] P. Svancara and M. La Mantia, Flight-crash events in super uid turbulence, Journal of Fluid Mechanics 876, R2 (2019). [42] U. Giuriato and G. Krstulovic, Quantum vortex reconnections mediated by trapped particles, Phys. Rev. B 102, 094508 (2020). [43] U. Frisch, Turbulence: The Legacy of A. N. Kolmogorov (Cambridge University Press, 1995). [44] S. Sera ni, M. Barbiero, M. Debortoli, S. Donadello, F. Larcher, F. Dalfovo, G. Lamporesi, and G. Ferrari, Dynamics and interaction of vortex lines in an elongated bose-einstein condensate, Phys. Rev. Lett. 115, 170402 (2015). [45] K. Xhani, E. Neri, L. Galantucci, F. Scazza, A. Burchianti, K. L. Lee, C. F. Barenghi, A. Trombettoni, M. Inguscio, M. Zaccanti, G. Roati, and N. P. Proukakis, Critical transport and vortex dynamics in a thin atomic josephson junction, Physical Review Letters 124, 045301 (2020). [46] I. Amelio, D. E. Galli, and L. Reatto, Probing quantum turbulence in He by quantum evaporation measurements, Phys. Rev. Lett. 121, 015302 (2018). [47] D. Proment, M. Onorato, and C. F. Barenghi, Vortex knots in a bose-einstein condensate, Phys. Rev. E 85, 036306 (2012). [48] D. Kleckner, L. H. Kau man, and W. T. M. Irvine, How super uid vortex knots untie, Nature Physics 12, 650 EP (2016). [49] S. V. Alekseenko, P. A. Kuibin, S. I. Shtork, S. G. Skripkin, and M. A. Tsoy, Vortex reconnection in a swirling ow, JETP Lett. 103, 455 (2016).

Journal

Nonlinear Sciences – arXiv (Cornell University)

Published: May 5, 2020

There are no references for this article.