Vortex molecules in thin films of layered superconductors

The seminal work of A.A. Abrikosov on the vortex lattices in type-II superconductors has triggered the creation and rapid development of the whole new field of condensed matter physics, namely, the physics of vortex matter in a large variety of superfluid and superconducting systems. Among these sys...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Физика низких температур
Datum:2018
Hauptverfasser: Samokhvalov, A.V., Mel’nikov, A.S., Buzdin, A.I.
Format: Artikel
Sprache:English
Veröffentlicht: Фізико-технічний інститут низьких температур ім. Б.І. Вєркіна НАН України 2018
Schlagworte:
Online Zugang:https://nasplib.isofts.kiev.ua/handle/123456789/176151
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Назва журналу:Digital Library of Periodicals of National Academy of Sciences of Ukraine
Zitieren:Vortex molecules in thin films of layered superconductors / A.V. Samokhvalov, A.S. Mel’nikov, A.I. Buzdin // Физика низких температур. — 2018. — Т. 44, № 6. — С. 691-700. — Бібліогр.: 34 назв. — англ.

Institution

Digital Library of Periodicals of National Academy of Sciences of Ukraine
id nasplib_isofts_kiev_ua-123456789-176151
record_format dspace
spelling Samokhvalov, A.V.
Mel’nikov, A.S.
Buzdin, A.I.
2021-02-03T19:05:23Z
2021-02-03T19:05:23Z
2018
Vortex molecules in thin films of layered superconductors / A.V. Samokhvalov, A.S. Mel’nikov, A.I. Buzdin // Физика низких температур. — 2018. — Т. 44, № 6. — С. 691-700. — Бібліогр.: 34 назв. — англ.
0132-6414
PACS: 74.25.Hа, 74.25.Uv, 74.78.–w
https://nasplib.isofts.kiev.ua/handle/123456789/176151
The seminal work of A.A. Abrikosov on the vortex lattices in type-II superconductors has triggered the creation and rapid development of the whole new field of condensed matter physics, namely, the physics of vortex matter in a large variety of superfluid and superconducting systems. Among these systems and compounds the layered superconductors are known to form a very important class which is associated with a number of fascinating new discoveries in vortex physics. The phenomenon of the vortex attraction in tilted magnetic fields provides one of the examples of such findings which affected theoretical and experimental research in the field for almost two decades. In our paper we review some recent advances in this direction focusing on the intervortex interaction and equilibrium vortex structures in thin films of layered superconductors in magnetic field tilted with respect to the layers. In such a case the magnetic field penetrates superconductor in the form of tilted vortices or a crossing array of Josephson vortices and pancake stacks. We study the interplay between two different long-range potentials: (i) attraction of tilted vortices or deformed stacks; (ii) the Pearl's repulsion. This interplay is responsible for the formation of the minimum in the total interaction energy and resulting decay of vortex chains in clusters. The number of Abrikosov vortices in these clusters (or vortex molecules) depends on field tilting angle and film thickness.
This work was supported, in part, by the Russian Foundation for Basic Research, the EU COST Action CA16218, the French ANR project “Optofluxonics” (AIB) and Foundation for the Advancement of Theoretical Physics “Basis” (ASM). The numerical simulations were carried out with the support of the Russian Science Foundation (project No. 17-12-01383). AIB thanks the Leverhulme Trust for supporting his stay in Cambridge 768 University
en
Фізико-технічний інститут низьких температур ім. Б.І. Вєркіна НАН України
Физика низких температур
Специальный выпуск К 90-летию со дня рождения A.A. Абрикосова
Vortex molecules in thin films of layered superconductors
Article
published earlier
institution Digital Library of Periodicals of National Academy of Sciences of Ukraine
collection DSpace DC
title Vortex molecules in thin films of layered superconductors
spellingShingle Vortex molecules in thin films of layered superconductors
Samokhvalov, A.V.
Mel’nikov, A.S.
Buzdin, A.I.
Специальный выпуск К 90-летию со дня рождения A.A. Абрикосова
title_short Vortex molecules in thin films of layered superconductors
title_full Vortex molecules in thin films of layered superconductors
title_fullStr Vortex molecules in thin films of layered superconductors
title_full_unstemmed Vortex molecules in thin films of layered superconductors
title_sort vortex molecules in thin films of layered superconductors
author Samokhvalov, A.V.
Mel’nikov, A.S.
Buzdin, A.I.
author_facet Samokhvalov, A.V.
Mel’nikov, A.S.
Buzdin, A.I.
topic Специальный выпуск К 90-летию со дня рождения A.A. Абрикосова
topic_facet Специальный выпуск К 90-летию со дня рождения A.A. Абрикосова
publishDate 2018
language English
container_title Физика низких температур
publisher Фізико-технічний інститут низьких температур ім. Б.І. Вєркіна НАН України
format Article
description The seminal work of A.A. Abrikosov on the vortex lattices in type-II superconductors has triggered the creation and rapid development of the whole new field of condensed matter physics, namely, the physics of vortex matter in a large variety of superfluid and superconducting systems. Among these systems and compounds the layered superconductors are known to form a very important class which is associated with a number of fascinating new discoveries in vortex physics. The phenomenon of the vortex attraction in tilted magnetic fields provides one of the examples of such findings which affected theoretical and experimental research in the field for almost two decades. In our paper we review some recent advances in this direction focusing on the intervortex interaction and equilibrium vortex structures in thin films of layered superconductors in magnetic field tilted with respect to the layers. In such a case the magnetic field penetrates superconductor in the form of tilted vortices or a crossing array of Josephson vortices and pancake stacks. We study the interplay between two different long-range potentials: (i) attraction of tilted vortices or deformed stacks; (ii) the Pearl's repulsion. This interplay is responsible for the formation of the minimum in the total interaction energy and resulting decay of vortex chains in clusters. The number of Abrikosov vortices in these clusters (or vortex molecules) depends on field tilting angle and film thickness.
issn 0132-6414
url https://nasplib.isofts.kiev.ua/handle/123456789/176151
citation_txt Vortex molecules in thin films of layered superconductors / A.V. Samokhvalov, A.S. Mel’nikov, A.I. Buzdin // Физика низких температур. — 2018. — Т. 44, № 6. — С. 691-700. — Бібліогр.: 34 назв. — англ.
work_keys_str_mv AT samokhvalovav vortexmoleculesinthinfilmsoflayeredsuperconductors
AT melnikovas vortexmoleculesinthinfilmsoflayeredsuperconductors
AT buzdinai vortexmoleculesinthinfilmsoflayeredsuperconductors
first_indexed 2025-11-26T10:30:19Z
last_indexed 2025-11-26T10:30:19Z
_version_ 1850618726897942528
fulltext Low Temperature Physics/Fizika Nizkikh Temperatur, 2018, v. 44, No. 6, pp. 691–700 Vortex molecules in thin films of layered superconductors A.V. Samokhvalov1,2, A.S. Mel’nikov1,2, and A.I. Buzdin3,4 1Institute for Physics of Microstructures, Russian Academy of Sciences, Nizhny Novgorod, GSP-105, 603950, Russia 2N.I. Lobachevsky State University of Nizhny Novgorod, Nizhny Novgorod 603950, Russia 3University of Bordeaux, LOMA UMR-CNRS 5798, F-33405 Talence Cedex, France 4Department of Materials Science and Metallurgy, University of Cambridge, CB3 0FS, Cambridge, United Kingdom E-mail: samokh@ipmras.ru Received December 27, 2017, published online April 25, 2018 The seminal work of A.A. Abrikosov on the vortex lattices in type-II superconductors has triggered the crea- tion and rapid development of the whole new field of condensed matter physics, namely, the physics of vortex matter in a large variety of superfluid and superconducting systems. Among these systems and compounds the layered superconductors are known to form a very important class which is associated with a number of fascinat- ing new discoveries in vortex physics. The phenomenon of the vortex attraction in tilted magnetic fields provides one of the examples of such findings which affected theoretical and experimental research in the field for almost two decades. In our paper we review some recent advances in this direction focusing on the intervortex interac- tion and equilibrium vortex structures in thin films of layered superconductors in magnetic field tilted with re- spect to the layers. In such a case the magnetic field penetrates superconductor in the form of tilted vortices or a crossing array of Josephson vortices and pancake stacks. We study the interplay between two different long- range potentials: (i) attraction of tilted vortices or deformed stacks; (ii) the Pearl's repulsion. This interplay is re- sponsible for the formation of the minimum in the total interaction energy and resulting decay of vortex chains in clusters. The number of Abrikosov vortices in these clusters (or vortex molecules) depends on field tilting angle and film thickness. PACS: 74.25.Hа Magnetic properties including vortex structures and related phenomena; 74.25.Uv Vortex phases; 74.78.–w Superconducting films and low-dimensional structures. Keywords: Abrikosov vortices, layered superconductors, vortex chains. 1. Introduction A.A. Abrikosov contributed to many domains of theo- retical physics: condensed matter, plasma physics, physics of high pressure, and quantum electrodynamics. However, the Nobel Prize in Physics 2003 was awarded to A.A. Abrikosov for his theoretical discovery of vortices and the theory of type-II superconductivity. These vortices are now often called Abrikosov vortices and the understanding of the properties of the mixed state of type-II superconductors is ultimately related to the paradigm of “vortex matter”. Thermal excitations, vortex pinning, crystal anisotropy, spatial and time dependent magnetic field, all these reveal panoply of different transitions in the vortex matter, which makes its physics very rich [1]. Both the equilibrium and transport properties of the vor- tex matter are essentially affected by the behavior of the potential of the interaction between vortices. In isotropic bulk superconductors the intervortex interaction potential is well known to be always repulsive [2] and screened at intervortex distances R greater than the London penetration depth λ. As a result, in perfect crystals quantized Abrikosov vortices (vortex lines) form a triangular flux lattice [3]. The magnetic flux through the unit cell of such flux line lattice equals to the flux quantum 0 = /c eϕ π , and each unit cell is occupied by the single vortex. This textbook picture is known to change drastically in layered superconductors in tilted magnetic field due to the phenomenon of the long-range attraction between tilted [4–8] or deformed [9,10] vortex lines composed of two dimension- al 2D pancake vortices (PVs). For bulk layered superconduc- tors with a moderate anisotropy (like NbSe2, YBaCuO), the low magnetic field penetrates mainly in form of the tilted vortex lines which are arranged in chains (Fig. 1(a)). These vortex chains have been observed experimentally by the dec- oration technique in YBa2Cu3O7 [11], scanning-tunneling © A.V. Samokhvalov, A.S. Mel’nikov, and A.I. Buzdin, 2018 A.V. Samokhvalov, A.S. Mel’nikov, and A.I. Buzdin microscopy in NbSe2 [12], and Lorentz microscopy meas- urements in YBa2Cu3O7 [13] (see [14] for a review). For strongly anisotropic superconductors like Bi2Sr2CaCu2O8+δ (BSCCO) the chain consists of a lattice of in-plane Josephson vortices (JVs) crossing with pancake stacks perpendicular to the planes (Figs. 1(b),(c)). The attraction between PVs and JVs [15] leads to the deformation of the pancake stacks de- tected by the modern vortex imaging methods such as Lo- rentz microscopy [13], high-resolution scanning Hall probe [17], Bitter decoration techniques [18], and magneto-optical measurements [19] (see also Ref. 14 as a review). A lot of other ground-state vortex chain configurations and phase transitions between them were studied in detail in Ref. 20. The potential of the interaction between vortices in layered superconductors depends on the internal structure of vortex lines and is determined by two different kinds of interac- tion between pancake vortices in different layers: magnetic and Josephson interactions. The relative strength of these two interactions depends on the anisotropy parameter = /c abΓ λ λ and interlayer spacing s, where cλ is the pen- etration depth for currents along the c axis and abλ is the penetration depth for currents in the ab plane (parallel to the layers). The distinctive feature of anisotropic super- conductors is known to be an unusual attractive part of the vortex–vortex interaction potential in tilted magnetic field [4–10]. Tilted or deformed PVs stacks attract each other in the plane defined by the anisotropy axes and vortex line, and the value of the attractive force is controlled by the tilting angle of the vortex line with respect to the anisotropy axis. Certainly, the interaction between vortex stacks at small dis- tances R λ remains repulsive. Interestingly, the attractive long-range intervortex potential has the same asymptotic behavior 21/aU R− ( )R λ for both cases: for tilted vortex lines [4–8] and for deformed stacks [9,10]. Note, that the attractive interaction between any two vortices in the chain (except the nearest neighbors) provides the stability of the whole vortex chain. In thin films of layered superconductors, however, such van der Waals type potential appears to be strongly modified by the long-ranged Pearl’s repulsion [21] which could mask the subtle attraction of vortices. The Pearl’s potential ( )PU r of intervortex interaction is always repulsive, decays ex- tremely slowly ( 1/PU R for > )R λ and always domi- nates at large distances > .mR R λ It means that the re- pulsion between two vortices is restored at intervortex dis- tance > mR R and stability of the vortex chain of a length mL R is destroyed. As a result, the formation of an infi- nite vortex chain becomes unfavorable. Adding vortices one by one we can find an optimal number of vortices which can be arranged in a chain of a finite length. As a result, there appears an intriguing possibility to form a vortex structure consisting of finite size chains, i.e., of vortex molecules. By varying either the film thickness or the tilting angle we can modify the balance between the attractive and repulsive in- teractions, which should determine the size of the energeti- cally favorable vortex configurations. Modern vortex imaging techniques provide a possibility to observe the crossover between different intervortex inter- action regimes in thin film samples. In particular, the mag- netic field distributions induced by vortices in thin films can be probed by the penetrating electron beam used in Lorentz microscopy measurements [22,23]. This technique, owing to the low penetration power of the existing 300-kV field- emission beam, permits us to work with films of thicknesses smaller than 0.5–1 µm. It is, therefore, “par excellence” an ideal tool to study the peculiarities of the vortex structures in thin films. In Ref. 24 it was demonstrated how the Lorentz microscopy technique permits to discover the special charac- ters of the intervortex interaction in YBaCuO films of thick- ness 0.5 m,d µ placed in a tilted magnetic field. The vor- tex structure changes qualitatively for a fixed external magnetic field 0B direction by increasing absolute value of the field 0B . The increase in the field value 0B causes an increase in the tilting angles of vortex lines. Therefore, the vortex attraction prevails at higher 0B values, while at low fields the attraction force is overcome by the repulsive force due to Pearl's effect, and the vortex chains are expected to disappear. Indeed, such a qualitative change in the vortex structure is confirmed [24] by the experimental data: at low fields, which correspond to small tilting values, the vortex chains are completely absent, while at rather high fields *0 > ,B B where *B is the critical field, the formation of vortex chains is clearly observed. Fig. 1. (Color online) Schematic representation of a layered superconductor with tilted (a) and crossing (b), (c) vortex structures, s is a distance between the layers, N is the number of layers, D = (N – 1)s is the total thickness of the structure. 692 Low Temperature Physics/Fizika Nizkikh Temperatur, 2018, v. 44, No. 6 Vortex molecules in thin films of layered superconductors In this paper we focus on the calculations of the pair po- tential of the intervortex interaction in thin films of layered superconductors. Our calculations are based on the Lowrence–Doniach theory [25] which is simplified neglect- ing the effects of weak interlayer Josephson coupling. This approximation of Josephson-decoupled superconducting (SC) layers [26] is known to be useful in studies of the vortex lattice structure for very weak coupling of the layers = ,J absλ Γ λ where the Josephson length =J sλ Γ is the in-plane size of the Josephson core radius [27]. The equi- librium shape of a single vortex line depends on the in-plane magnetic field || || 0= BB x distribution and can be very dif- ferent. Without the in-plane field, the pancakes form a verti- cal stack. If the in-plane field ||B is rather small 2 || 0 0( = / )B H sϕ Γ and does not penetrate inside super- conductor in form of Josephson vortices, then screening Meissner currents tilt and bend the vortex line. Bending of the vortex line is essential only near the surface of the SC film and often can be neglected [28]. In Sec. 2.2 we study the interaction between two straight stacks of PVs tilted at a cer- tain angle γ with respect to the c direction, as shown in Fig. 1(a). In fact in layered superconductors with very weak interlayer coupling the Josephson vortices will appear at much lower field. As a result, at tilted magnetic field, cross- ing lattice of pancakes, formed by Abrikosov vortices, and Josephson vortices is more energetically preferable than a lattice of tilted vortex stacks [15–17]. The interaction be- tween pancakes and in-plane field in the form of Josephson vortices produces zigzag deformation of the stack of the pan- cakes [30]. This deformation is responsible for a long-range attraction between such stacks [9] which is quite similar to the case of tilted vortex lines. In Sec. 2.3 we study the ex- treme case when the in-plane magnetic field || 0B H cre- ates the dense triangular lattice of JVs which are strongly compressed along the c axis due to a very high anisotropy 1Γ  . This dense triangular lattice of JVs produces zigzag displacement of PVs [30] along the x axis, so that the pan- cake centers for each vortex line are positioned along the broken line (see Fig. 1(c)). We calculate the interaction po- tential between two zigzag lines in a thin SC film. In Sec. 2.4 we compare these two extreme cases and discuss qualitative- ly the condition of the vortex attraction if the in-plane mag- netic field is relatively small || 0( )B H and the crossing points between JVs and PVs are rare. In Sec. 3 we discuss the condition of vortex cluster formation and phase transitions between the vortex lattices with different number of vortices per unit cell. 2. Interaction potential for two stacks of pancakes As a model of a film of strongly anisotropic supercon- ductor we consider a finite set of N thin SC layers with the thickness d much smaller λ located at a distance s from each other, so that the nth SC layer coincides with the plane = =nz z ns (1 ).n N≤ ≤ The field component per- pendicular to the layers creates the vortex line which pierc- es the film as a stack of PVs centered at the points 0 0=n n nx y+r x y in the nth layer. We derive general ex- pressions for interaction potential of two identical stacks of PVs taking into account both long-range attraction and repulsion phenomena. We study both the limits of weak || 0( )B H and strong || 0( )B H in-plane magnetic field. The shape of the interacting vortex lines is assumed to be fixed and not affected by the vortex–vortex interac- tion potential. Certainly, such assumption is valid only in the limit of rather larger distances between the vortex lines when the effect of interaction on the vortex shape can be viewed as a small perturbation. 2.1. Basic equations Within the model of the set of Josephson-decoupled SC layers, pancakes interact with each other only via magnetic fields. General equation for the vector potential A distribu- tion in such system reads , =1 4rot rot = ( ) ( ) N m n n n m z z c π δ −∑A J r . (1) The sheet current at the nth layer created by the pancake at mth layer takes the form ( ) = ( ) ( , ) , 4 m m n m nm n c z − δ − πΛ J r Φ r r A r (2) where 2 2= / = /abd sΛ λ λ is the effective penetration depth in a superconducting film, ( , )m zA r is the vector potential induced by the only pancake vortex located in the mth layer. The vector ( )Φ r in the Eq. (2) is given by the expression [ ]00 2( ) = 2 ×ϕ π z r Φ r r . (3) For the layered system without Josephson coupling a general expression for the free energy can be written in the form ( ) 2 2 2 =1 1 4= rot ( ) ( ) . 8 N n n n F dV z z c  π + Λ δ −   π    ∑∫ A J r (4) The total vector potential ( , )zA r and the sheet current in the nth layer ( )nJ r , produced by an arbitrary vortex line are the sum of the contributions induced by N pancakes: =1 =1 ( , ) = ( , ), ( ) = ( ), N N m m n n m m z z∑ ∑A r A r J r J r and can by found from Eqs. (1)–(3) using an approach simi- lar to that proposed in Refs. 32, 33 (see Ref. 28 for details). Using the gauge div = 0A and the Fourier transforms 2( , ) = e ( , ),i ikzk d dz z+∫ qrA q r A r (5) 2 2( ) = e ( , ), ( ) = e ( ),i i n n n nd z d∫ ∫qr qrA q r A r J q r J r (6) Low Temperature Physics/Fizika Nizkikh Temperatur, 2018, v. 44, No. 6 693 A.V. Samokhvalov, A.S. Mel’nikov, and A.I. Buzdin one can rewrite the basic Eq. (1) for each vortex line in the momentum representation as follows: [ ]2 2 1( ) ( , ) = ( ) ( ) e ,ikns n n n q k k+ − Λ ∑A q Φ q A q (7) where [ ]0 0 2( ) = ( )e , ( ) = .i nn i q × − ϕqr z q Φ q Φ q Φ q These equations can be reduced to the scalar form | |1 e = e , 2 iqn m qs nn m m f f q − −+ Λ ∑ r (8) where we introduce the new functions ( )nf q : ( )( ) = ( ) ( ) = ( ) ( ) 4 4n n n n c c f− πΛ πΛ J q Φ q A q Φ q q . (9) The solution of the linear system (8) for a fixed configura- tion of pancakes nr determines the distribution of the vec- tor potential ( , )zA r which is created by an arbitrary vor- tex line in a finite stack of superconducting layers. For two vortex lines we can write the total vector poten- tial and the total sheet current as superpositions of contri- butions coming from the first (1)( nA , (1) )nJ and second (2)( nA , (2) )nJ vortices. Calculating the interaction energy intε of vortex lines we should keep in the free energy only the terms which contain the products of fields correspond- ing to different vortex lines. As a result, in the momentum representation the general expression (4) for the free ener- gy of the layered system without Josephson coupling reads ( )2 (1) (1) (2) int 3 1= ( ) ( ) ( ) 32 n n n n d ε − − +π Λ ∑∫ q Φ q A q Φ q ( )(2) (2) (1)( ) ( ) ( ) .n n n + − − Φ q A q Φ q (10) For the particular case of two identical (parallel) vortex lines which are shifted at the vector (2) (1)= ( = 1 )n n n N− −R r r in the xy plane we get following general expression for the interaction energy via the scalar functions ( )nf q [28]: 2 2 0 int 3 2( ) = cos( ) ( )e . 16 i nn n d f q −ϕ ε π Λ ∑∫ qrqR qR q (11) The expression (11) and equations (8) determine the inter- action energy of two identical vortex lines. 2.2. Interaction potential of tilted stacks of pancakes In the limit of weak in-plane magnetic field || 0( )B H we neglect the vortex line bending and restrict ourselves to the case of straight vortex lines parallel to the plane xy and tilted at a certain angle γ with respect to the anisotropy axis c (z axis) as shown in Fig. 1(a). Let us evaluate the interaction energy (11) of two tilted parallel vortex lines taking (1) (2) (1) 0= tan , =n n nns γ +r x r r R . For the case of a large intervortex distance ,R s the solution of the problem (8) may be simplified, because for the relevant wave vectors 1/q R the conditions of the continuous limit are satisfied: 1qs  and tan 1xq s γ  . Introducing a continuous coordinate =z ns and continu- ous function ( )f zq , the linear system of Eq. (8) reduces to the following integral equation /2 tan| | 2 /2 1( ) e ( ) = e . 2 D iq zq z z x ab D f z dz f z q γ− − ′ − + ′ ′ λ ∫q q (12) The Eq. (12) can be rewritten as a differential one ( ) ( ) 2 tan2 2 2 22 2 ( ) = etan iq zxab x d f q f z q q dz γ−− λ + − γ +q q (13) at the interval /2 < < /2D z D− with the boundary conditions tan /2 /2 = ( tan )e ,iq Dxx D df q f iq q dz ± γ ±   ± γ ±   q q (14) where = ( 1)D N s− is the thickness of superconducting film. Skipping the details of the solution of Eqs. (13), (14), we get the following expression for the interaction energy (11) in the continuous limit [28]: 2 2 0 int 3 2 2= cos( ) ( ), 16 ab d S q ϕ ε π λ ∫ q qR q (15) where the expression for the kernel ( )S q takes form ____________________________________________________ 2 2 2 22 2 2 2 2 2 2 2 2(1 ) (1 )sinh ( )(cosh cos( )) 2 sin( ) ( ) = , 1 (1 ) 2 cosh (1 )sinhab k k p L k p L pL kp pLp kS D p q p k L k L−  − − + − − ++  +  + + λ + + +  q (16) 2 2 2 2 2 2= , = / , = tan / .ab ab x abL D q k q q p q q− − −+ λ + λ γ + λ _______________________________________________ The first term in the kernel (16) describes the intervortex interaction in the bulk system, while the second term is responsible for the Pearl effect due to film boundaries. The minimum energy configuration corresponds to the case = 0yR . In Fig. 2 we present some typical plots of the interaction energy int ( , = 0)x yR Rε vs the distance =xR R 694 Low Temperature Physics/Fizika Nizkikh Temperatur, 2018, v. 44, No. 6 Vortex molecules in thin films of layered superconductors for = 3 abD λ which corresponds to the Lorentz microscopy experiments in YBCO [24] and Bi-2212 [13] samples. Ana- lyzing the dependence int ( ),Rε one can separate three con- tributions to the energy of vortex–vortex interaction: (i) a short-range repulsion which decays exponentially with in- creasing intervortex distance R (for > );abR λ (ii) an intervortex attraction which is known to be specific for tilted vortices in anisotropic systems; this attraction energy term decays as 2R− and strongly depends on the angle γ be- tween the vortex axis and the c direction; (iii) long-range (Pearl) repulsion which decays as 1R− and results from the surface contribution to the energy. Note that the third term does exist even for a large sample thickness D (see Ref. 29) although in the limit abD λ it is certainly masked by the dominant bulk contribution. At abR λ the short-range interaction term vanishes and the interaction energy vs R takes the simple form 2 2 0 eff int 2 2 2tan , 8 D RR  ϕ γ ε − +  π  (17) where eff = 2 tanh( /2 )ab abD D D− λ λ is the effective film thickness. One can observe here an interplay be- tween the long-range attractive (first term in Eq. (17)) and the long-range repulsive (second term in Eq. (17)) forces. Note that the abλ value increases with an increase in temperature, thus, the effective thickness decreases and the long-range attraction force appears to be suppressed with increasing temperature. For large R the energy is always positive and corresponds to the vortex repulsion similar to the one between the pancakes in a single layer system. With a decrease in the distance R the attraction force comes into play resulting in the change of the sign of the energy. Such behavior points to the appearance of the minimum in the interaction potential. 2.3. Interaction potential of zigzag stacks of pancakes In the limit of strong in-plane magnetic field || 0( )B H the tilted magnetic field penetrates in the form of the crossing lattice comprising the in-plane Josephson vortices and per- pendicular to the layers stacks of pancakes (see Fig. 1(c)). In high-field limit, || 0 0> = / ,JB H sϕ λ all interlayer junctions are homogeneously filled with Josephson vortices, which form the dense triangular lattice. The period of the lattice in z direction is equal 2s which is much less than the period along the y axis 0 ||= / .Ja B sϕ λ The interaction between pancakes and in-plane field in the form of JVs is known to produce zigzag deformation of PV stacks (see Fig. 1(c)) [15,20,30]. The zigzag deformation is somewhat larger near the surface (due to the decrease of the stifness of the PVs stack) but for simplicity we neglect this relatively small ef- fect. The amplitude of the zigzag deformation || 0 | | = , = 1, ln Jab n ab B u u H λλ β β β λ   (18) is assumed to be the same for all N layers and corre- sponds to the equilibrium form of the zigzag vortex in a bulk ( )N → ∞ layered superconductor [9,30]. In this case the centers of pancakes for both vortex lines are positioned at a zigzag line (1) 1 (2) (1) 0= ( 1) =n n n nu−− +r x r r R , (19) and the Eq. (8) with respect to functions ( )nf q looks dif- ferently for odd and even n . Introducing two new func- tions 1 ( )kf q and 2 ( )kf q 1 2 , if — odd, = ( 1)/2, = , if — even, = /2, k n k f n k n f f n k n +   the Eq. (8) can be rewrited as follows: 2| | 1,2 1,2 1 e 2 k m qs k m m f f q − −+ + Λ ∑ (2| | 1) 2,1 1 e = e . 2 iq uk m qs xm m f q ±− − ++ Λ ∑ (20) The interaction energy (11) expressed in terms of the scalar functions 1 ( )kf q and 2 ( )kf q reads 2 2 0 int 3 2( ) = cos( ) 16 d q ϕ ε × π Λ ∫ qR qR 1 2( )e ( )e . N iq u iq ux xk k k f f− × +  ∑ q q (21) The expression (21) and the system of algebraic equation (20) determine the interaction energy of two identical zigzag vortex lines which are shifted at the vector R in the xy plane. Fig. 2. (Color online) Typical plots of the interaction energy per vortex [Eqs. (15) and (16)] vs the distance R between two vorti- ces for a film of thickness D = 3λab and different tilting angles γ = 70°, 75°, 78°, 80° 2 3 0 0( = /16 )abε ϕ π λ . Low Temperature Physics/Fizika Nizkikh Temperatur, 2018, v. 44, No. 6 695 A.V. Samokhvalov, A.S. Mel’nikov, and A.I. Buzdin For the case of a large intervortex distance , ,R s u we may use the continuous limit for the solution of the problem (20) because for the relevant wave vectors 1qs  and 1xq u  . We introduce a continuous coordi- nate = 2z ks and continuous functions 1,2 ( , )f zq defined at the interval | | /2z D≤ , where D is the total thickness of superconducting film. Thus, the linear system of Eqs. (20) reduces to the following integral equations: /2 | | 1,2 1,2 /2 1( , ) e ( , ) 4 D q z z D f q z dz f q z q s − − ′ − + +′ ′ Λ ∫ /2 | | 2,1 /2 e e ( , ) = e . 4 Dqs iq uq z z x D dz f q z q s − ±− − ′ − + ′ ′ Λ ∫ (22) Using the solutions of Eqs. (22), we may rewrite the ex- pression for interaction energy (21) as follows (see Ref. 34 for details): ____________________________________________________ 2 2 2 2 20 int 3 2 2 2 4 ( ) ( ) 2 ( )cos sin cos( ) = cos( ) , 16 (1 / tanh ) x x x ab q u q u q q u d D s p q p q p L+ + + +   ϕ  ε + +   π Λ λ +     ∫R q qR (23) where 2 2 2 2 21/ , , = /2.abp q p q L p D+ − ± ±+ λ  (24) Performing in (23) the necessary integration, we finally obtain the following expression for the interaction energy of two identical zigzag pancake stacks which are shifted at the vector 0= RR x in the xy plane: 2 2 2 0 int 0 0 02 2 2 2 4( ) = 2 ln 32 ab ab ab R R u R u R uR D K K K s R         ϕ − + −ε + + + +        λ λ λ     π Λ       0 0 0 2 2 2 2 2 2 2 21 0 2 ( ) ( ( 2 )) ( ( 2 ))2 , ( ) 1 / ( /2)tanhab ab ab ab J qR J q R u J q R udq q q Dq q ∞ − − −− + − + + +  λ + λ + + λ + λ  ∫ (25) _______________________________________________ where 0J and 0K are the Bessel and modified Bessel functions of zero order, respectively. The first term propor- tional to D in Eq. (25) describes the interaction between the zigzag stacks in the bulk layered system [9], while the last term is responsible for the influence of film bounda- ries. In Fig. 3 we present some typical plots of the interac- tion energy per one layer int /Nε vs the intervortex dis- tance R for different thickness of the film D and the in- plane magnetic field 0=xB H . At long distances abR λ , the modified Bessel func- tions decay exponentially, and a bulk short-range repulsion between two PVs stacks is negligible small. The leading bulk contribution comes from the logarithmic term in (25), which describes an attraction between the zigzag PV lines. The last term in (25) describes long-range Pearl's repulsion which decays as 1/R and results from the surface contribu- tion to the energy. Assuming that the effect of the zigzag deformation abu λ is small, the long-range part of inter- action energy (25) for abR λ can be written as 2 2 2 0 int 2 2 2 3 2 4( ) . 8 ab Du uR RR R  ϕ ε − + +  π λ    (26) One can observe an interplay of the long-range attractive and the repulsive forces between two zigzag deformed PV stacks, similar to the one between two tilted vortices in ani- sotropic SC films [24,28]. The last term in (26) describes a Fig. 3. (Color online) Typical plots of the interaction energy per layer εint(R)/2Nε0 [Eq. (25)] vs the distance R between two zigzag deformed stacks of pancakes for in-plane magnetic field B||/H0 = 1. The numbers near the curves denote the values of film thickness D/λab. The dashed line shows the interaction energy between two zigzag vortices in bulk (D → ∞) layered SC. Dotted lines show the long-range part of interaction energy (26). Here 2 2 0 0= /32 ,abε ϕ π λ s = 0.01λab, γ = 300. 696 Low Temperature Physics/Fizika Nizkikh Temperatur, 2018, v. 44, No. 6 Vortex molecules in thin films of layered superconductors modification of long-range Pearl’s repulsion due to zigzag deformation of the PVs stack. Certainly, in a bulk sample ( )D → ∞ the Pearl’s term in (26) vanishes, and at long dis- tances the dominant interaction between the zigzag PV lines is an attraction [9]. As a result, in a bulk sample there always exists the minimum of the interaction potential, which reali- zes with logarithmic accuracy at 2 ln( / )m ab abR u≈ λ λ . However the second term in (26) is very important even for a rather thick films: for large R the energy intε is always positive and corresponds to the vortex repulsion. With a decrease of the distance R the attraction force comes into play and can result in the change of the sign of the energy at 2 2 0 /2 ,ab abR Du≈ λ λ while short-range repulsion is still weak. Such behavior should be accompanied by the appear- ance of the minimum in the interaction potential. The condi- tion 0 > mR R gives us a rough estimate of the critical film thickness crD for which the minimum in the interaction potential can exist and the formation of vortex chains can be energetically favorable: 22 ||2 3 cr 0 4 4 (ln ) ,J ab ab B D H  σλ ≈ σλ β β  λ    (27) where the constant σ is of the order unity. The more accu- rate criterion of attraction of zigzag PVs stacks should be based on the expression for the interaction energy (25) and may be found from the conditions int int( ) = ( ) = 0.R Rε ε′ The typical dependence of the critical film thickness crD as a function of in-plane magnetic field ||B is shown in Fig. 4. So, in the presence of a dense lattice of Josephson vortices, the AVs penetrate in the form of chains only for a large film thickness cr>D D . Otherwise, if cr<D D , the formation of the usual Abrikosov lattice of zigzag de- formed stacks of PVs occurs. It is interesting to note that following (26) the intervortex attraction increases near the critical temperature of the superconducting transition ,cT when abλ becomes large (except for the region very close to cT where the condition ab Jλ λ is not satisfied). This behavior is in contrast with the moderately anisotropic case [28] when the critical thickness for intervortex attraction to be observed decreases near cT and the repulsion between vortices prevails. 2.4. Interaction potential of two stacks in the crossing lattices In the high in-plane field limit || 0B H considered in Sec. 2.3 the triangular lattice of JVs is so dense that the cur- rents of adjacent JVs ovrlap setrongly and the amplitude of the zigzag deformation decreases [30]. As a result, the criti- cal thickness crD grows rapidly with an increase of the in- plane field ||B (see Fig. 4). Also the potential dip in the intervortex interaction energy (25) strongly decreases with an increase of ||B [9]. Optimal regime for the long-ranged attraction corresponds to || 0B H . If the magnetic field ||B is small, the crossings between JVs and PVs are rare and the distance between Josephson vortices along the z axis be- comes much larger than 2s, as well as the distance between the deformed parts of the PVs line (see Fig. 1(b)). Let us estimate the attraction between two deformed stacks in this limit. The Josephson vortices are well separated and posi- tioned at the distance 0 || 0 ||3 /2 /s sB H Bγϕ   along the z axis. As has been demonstrated in Ref. 9, the contribution to the attraction from one crossing is 22 0 att 2( ) = , ln( / ) ab J J ab s R AR  ϕ λ ε −  λ λ λ   (28) where 3.5A ≈ (see Ref. 15). Neglecting the effect of a zigzag deformation on the Pearl's repulsion, the long- ranged part of interaction energy int ( )Rε in low-field limit || 0B H at distances abR λ may be estimated as 2 2 0 int 2 2 2 2( ) 8 ab DuR RR  ϕ ε − +  π λ      , (29) where the effective displacement [9,15] 1/42 || 0 2 2 ln( / ) ab J J ab B u A H  λ  λ λ λ     (30) grows slowly with increasing in-plane magnetic field ||.B The expressions (29), (30) result in the following estimate of the critical film thickness cr :D 1/22 2 || cr 0 ( / )ln . 2 J J ab ab BA D H − λ λ λ  λ     (31) Fig. 4. Critical film thickness Dcr as a function of in-plane mag- netic field B|| ( ).− −• The dashed line shows the estimate of the critical film thickness (27). Here β = B||λJ/H0λab, s = 0.01λab, γ = = 300, λJ = 3λab, σ = 2. Low Temperature Physics/Fizika Nizkikh Temperatur, 2018, v. 44, No. 6 697 A.V. Samokhvalov, A.S. Mel’nikov, and A.I. Buzdin As before, the case || 0B H is optimal for observation of intervortex attraction and this case corresponds to the crossover between the regimes described by formulas (27) and (31). For typical experimental value of the in-plane magnetic field [17] B|| = 20–30 G we obtain cr 70 μmD  . Thus, the relations (27), (31) give us the estimate of the threshold value of the thickness cr || 0 cr || 0 , if < , > , if > , D B H D D B H     for which the minimum in the interaction potential can exist and the interplay between the long-range attraction and repulsion between zigzag vortex stacks lines in the films of layered superconductors takes place. We should note that the Pearl's interaction plays an important role even for rather thick film. Indeed, for the most favorable case || 0B H this interaction completely masks the attrac- tion for the film thicknesses smaller than ∼50–200 λab. 3. Vortex molecules The vortex attraction in bulk layered superconductors is known to result in the formation of infinite vortex chains. But even in the regime when the intervortex attraction ex- ists, the formation of infinite chains can be questioned for rather thin films. The point is that, despite the fact that two vortices attract each other, further increase in the number of vortices arranged in a chain can be energetically unfavorable because of the slower decay of the long-ranged Pearl's re- pulsive force compared to the attractive one. Therefore, for rather thin samples, there appears an intriguing possibility to observe vortex chains of finite length, i.e., vortex molecules or clusters. The calculations [24,28] confirm this and indeed the number of vortices energetically favorable in a molecule grows as we increase the film thickness and/or the tilting angle because of the increasing attraction term in the pair potential intε (15), (16). In Ref. 28 the intervortex interac- tion has been also calculated and analyzed taking account of the equilibrium shape of tilted vortex lines. Figure 5 shows schematic pictures of vortex matter con- sisting of dimeric (Fig. 5(a)) and trimeric (Fig. 5(b)) mole- cules, indicating that the vortex lattices contain more than one vortex per unit cell. The transitions between different multiquanta vortex lattices should occur with the change of the tilting angle and field strength [28]. Finally, for rather thick samples with abd λ we get a standard infinite chain structure typical of bulk systems (Fig. 5(c)). Note that the formation of an infinite vortex chain may be considered in some sense as a polymerization of the vortex molecules. Certainly, the crossover from the vortex molecule state to the infinite chain structure is strongly influenced by the in- crease in the vortex concentration governed by the compo- nent of the external magnetic field perpendicular to the film. Indeed, one can expect such a crossover to occur when the mean intervortex spacing approaches the molecular size. The experiments [24] were performed at constant orien- tation of the applied magnetic field. Therefore, by varying the magnetic field, changes of both the tilting angle and vor- tex concentration were realized. As the vortex concentration was relatively high, we could not expect to observe the mol- ecules (in this regime the average distance between vortices must be much larger than the size of the molecule). To ob- serve the vortex molecules (or multiquanta vortex lattices) it would be preferable to change only the parallel component of the magnetic field, by varying the vortex tilting angle while not affecting the vortex concentration, which must be very low to avoid the inter-molecule interaction. Qualitatively all conclusions of references [24,28] are ap- plicable for the case of crossing lattices. Similarly, the zig- zags of PVs stacks can be arranged in vortex molecules. To estimate the size of such molecule let's find the cohesion en- ergy of equidistant chain of deformed PVs stacks. The per- pendicular component of the magnetic field B⊥ fixes the number M of AVs in the film, so that in the square film W W× it should be 2 0= /M B W⊥ ϕ vortices. We consider the case of small concentration of vortices when at large dis- tances the intervortex interaction is given by Pearl’s term. For roughly uniform vortex distribution (usual vortex lattice) the energy per one vortex can be estimated as 2 0 0 /4 .B W⊥ε ϕ π If vortices form clusters with m vortices each, it may be demonstrated that interaction between clusters gives the same contribution 0ε per one vortex. The remaining energy origi- Fig. 5. Schematic pictures of vortex matter consisting of dimeric molecules (a), trimeric molecules (b) and infinite vortex chain (c). Vor- tex positions are denoted by filled ellipses. 698 Low Temperature Physics/Fizika Nizkikh Temperatur, 2018, v. 44, No. 6 Vortex molecules in thin films of layered superconductors nates from the interaction between vortices inside cluster, and for the case of a low in-plane magnetic field || 0 ,B H it may be expressed via the long-ranged part of interaction en- ergy intε (29): int int > 1= ( ), m m ij i j R m ε ε∑ (32) where ijR are the distances between ith and jth vortices in the chain molecule. Taking the characteristic distance be- tween vortices in a cluster 2 ln( / )m ab J abR ≈ λ λ λ we find 2 2 0 int 2 2 2 2 ln . 8 m mab m Du m RR  ϕ ε − +  π λ       (33) Certainly the cluster formation occurs if int < 0mε and the number of vortices in the cluster is given by the expression 2 2exp 2 ab m Dum R   ≈   λ    . (34) The condition > 1m 2 2( /2 > 1)ab mDu Rλ   gives us the low boundary of the in-plane magnetic field restricting the in- terval of vortex molecules existence 22 3|| 0 ( / )ln> . 2 J J ab ab B H D  λ λ λ   λ   (35) For = 100 abD λ , = 3J abλ λ and 0 = 3 TH we obtain || > 60 GB . Note that the number of vortices m in the molecules rapidly increases as the in-plane magnetic field ||B grows. Naturally this scenario realizes only for low concentra- tion of AVs when the average distance between vortices 0 /B⊥ϕ is much larger than the intervortex distance in the chain mR . In the opposite case we may expect the existence of multiquanta flux lattice with several vortices per cell similar to the case of tilted vortices considered in Ref. 28. 4. Conclusions To sum up, we have analyzed some distinctive features of the vortex matter in thin films of layered superconduc- tors focusing mainly on the effects caused by the tilted magnetic field. The interplay of electrodynamic interac- tions inside and outside the film results in a very peculiar behavior of the intervortex interaction potential which in- cludes both attractive and repulsive terms. This interplay can change essentially the structural properties of the vor- tex matter which become quite different from the ones suggested in the original work by A.A. Abrikosov. In par- ticular, we predict the possible formation of vortex clusters or molecules which can be also viewed as the multiplica- tion of the unit cell for a regular vortex array. This work was supported, in part, by the Russian Foundation for Basic Research, the EU COST Action CA16218, the French ANR project “Optofluxonics” (AIB) and Foundation for the Advancement of Theoreti- cal Physics “Basis” (ASM). The numerical simulations were carried out with the support of the Russian Science Foundation (project No. 17-12-01383). AIB thanks the Leverhulme Trust for supporting his stay in Cambridge 768 University. ________ 1. G. Blatter, M.V. Feigel'man, V.B. Geshkenbein, A.I. Larkin, and V.M. Vinokuret, Rev. Mod. Phys. 66, 1125 (1994). 2. P.G. De Gennes, Superconductivity of Metals and Alloys, Benjamin, New York (1966). 3. A.A. Abrikosov, Fundamentals of the Theory of Metals, North-Holland, Amsterdam (1988). 4. A.I. Buzdin and A.Yu. Simonov, Pis'ma Zh. Eksp. Teor. Fiz. 51, 168 (1990) [JETP Lett. 51, 191 (1990)]. 5. A.I. Buzdin and A.Yu. Simonov, Physica C 168, 421 (1990). 6. A.M. Grishin, A.Yu. Martynovich, and S.V. Yampolskii, Sov. Phys. JETP 70, 1089 (1990) [Zh. Eksp. Teor. Fiz. 97, 1930 (1990)]. 7. V.G. Kogan, N. Nakagawa, and S.L. Thiemann, Phys. Rev. B 42, 2631 (1990). 8. A.I. Buzdin and A.Yu. Simonov, Physica C 175, 143 (1991). 9. A. Buzdin and I. Baladie, Phys. Rev. Lett. 88, 147002 (2002). 10. M.J.W. Dodgson, Phys. Rev. B 66, 014509 (2002). 11. P.L. Gammel, D.J. Bishop, J.P. Rice, and D.M. Ginsberg, Phys. Rev. Lett. 68, 3343 (1992). 12. H.F. Hess, C.A. Murray, and J.V. Waszczak, Phys. Rev. Lett. 69, 2138 (1992). 13. A. Tonomura, H. Kasai, O. Kamimura, T. Matsuda, K. Harada, T. Yoshida, T. Akashi, J. Shimoyama, K. Kishio, T. Hanaguri, K. Kitazawa, T. Masui, S. Tajima, N. Koshizuka, P.L. Gammel, D. Bishop, M. Sasase, and S. Okayasu, Phys. Rev. Lett. 88, 237001 (2002). 14. S.J. Bending and M.J.W. Dodgson, J. Phys.: Condens. Matter 17, R955 (2005). 15. A.E. Koshelev, Phys. Rev. Lett. 83, 187 (1999). 16. L.N. Bulaevskii, M. Ledvij, and V.G. Kogan, Phys. Rev. B 46, 366 (1992). 17. A. Grigorenko, S. Bending, T. Tamegai, S. Ooi, and M. Henini, Nature (London) 414, 728 (2001). 18. C.A. Bolle, P.L. Gammel, D.G. Grier, C.A. Murray, D.J. Bishop, D.B. Mitzi, and A. Kapitulnik, Phys. Rev. Lett. 66, 112 (1991); I.V. Grigorieva, J.W. Steeds, G.Balakrishnan, and D.M. Paul, Phys. Rev. B 51, 3765 (1995); M. Tokunaga, T. Tamegai, Y. Fasano, and F. de la Cruz, Phys. Rev. B 67, 134501 (2003). 19. V.K. Vlasko-Vlasov, A. Koshelev, U. Welp, G.W. Crabtree, and K. Kadowaki, Phys. Rev. B 66, 014523 (2002). 20. A. E. Koshelev, Phys. Rev. B 71, 174507 (2005) 21. J. Pearl, Appl. Phys. Lett. 5, 65 (1964). 22. A. Tonomura, Electronic Holography, 2nd ed., Springer Series in Optical Sciences 70, Springer, Berlin, Heidelberg (1999). Low Temperature Physics/Fizika Nizkikh Temperatur, 2018, v. 44, No. 6 699 https://doi.org/10.1103/RevModPhys.66.1125 https://doi.org/10.1016/0921-4534(90)90537-O https://doi.org/10.1103/PhysRevB.42.2631 https://doi.org/10.1016/0921-4534(91)90245-T https://doi.org/10.1103/PhysRevLett.88.147002 https://doi.org/10.1103/PhysRevB.66.014509 https://doi.org/10.1103/PhysRevLett.68.3343 https://doi.org/10.1103/PhysRevLett.69.2138 https://doi.org/10.1103/PhysRevLett.88.237001 https://doi.org/10.1088/0953-8984/17/35/R01 https://doi.org/10.1088/0953-8984/17/35/R01 https://doi.org/10.1103/PhysRevLett.83.187 https://doi.org/10.1103/PhysRevB.46.366 https://doi.org/10.1038/414728a https://doi.org/10.1103/PhysRevLett.66.112 https://doi.org/10.1103/PhysRevB.51.3765 https://doi.org/10.1103/PhysRevB.67.134501 https://doi.org/10.1103/PhysRevB.66.014523 https://doi.org/10.1103/PhysRevB.71.174507 https://doi.org/10.1063/1.1754056 https://doi.org/10.1007/978-3-540-37204-2 A.V. Samokhvalov, A.S. Mel’nikov, and A.I. Buzdin 23. K. Harada, T. Matsuda, J. Bonevich, M. Igarashi, S. Kondo, G. Pozzi, U. Kawabe, and A. Tonomura, Nature 360, 51 (1992). 24. A.I. Buzdin, A.S. Mel'nikov, A.V. Samokhvalov, T. Akashi, T. Masui, T. Matsuda, S. Tajima, H. Tadatomo, and A. Tonomura, Phys. Rev. B 79, 094510 (2009). 25. W.E. Lowrence and S. Doniach, in Proceedings of 12th Intern. Conf. on Low Temperature Physics, Kyoto 1970, E. Kanda (ed.), Keigaku, Tokyo (1970), p. 361. 26. A.I. Buzdin and D. Feinberg, J. Phys. (Paris) 51, 1971 (1990); J.R. Clem, Phys. Rev. B 43, 7837 (1991); K. Efetov, Sov. Phys. JETP 49, 905 (1979); S.N. Artemenko and A.N. Kruglov, Phys. Lett. A 143, 485 (1990). 27. L. Bulaevskii and J.R. Clem, Phys. Rev. B 44, 10234 (1991). 28. A.V. Samokhvalov, D.A. Savinov, A.S. Mel'nikov, and A.I. Buzdin, Phys. Rev. B 82, 104511 (2010). 29. J. Pearl, J. Appl. Phys. 37, 4139 (1966). 30. L.N. Bulaevskii, M. Maley, H. Safar, and D. Dominguez, Phys. Rev. B 53, 6634 (1996). 31. J. Mirkovic, S.E. Savel'ev, E. Sugahara, and K. Kadowaki, Phys. Rev. Lett. 86, 886 (2001); J. Mirkovic, S.E. Savel’ev, E. Sugahara, and K. Kadowaki, Phys. Rev. B 66, 132505 (2002); M. Konczykowski, C.J. van der Beek, A.E. Koshelev, V. Mosser, M. Dodgson, and P.H. Kes, Phys. Rev. Lett. 97, 237005 (2006). 32. V. Pudikov, Physica C 212, 155 (1993). 33. T. Pe, M. Benkraouda, and J.R. Clem, Phys. Rev. B 55, 6636 (1997). 34. A.V. Samokhvalov and A.S. Mel'nikov and A.I. Buzdin, Phys. Rev. B 85, 184509 (2012). ___________________________ 700 Low Temperature Physics/Fizika Nizkikh Temperatur, 2018, v. 44, No. 6 https://doi.org/10.1038/360051a0 https://doi.org/10.1103/PhysRevB.79.094510 https://doi.org/10.1051/jphys:0199000510170197100 https://doi.org/10.1103/PhysRevB.43.7837 https://doi.org/10.1016/0375-9601(90)90418-N https://doi.org/10.1103/PhysRevB.44.10234 https://doi.org/10.1103/PhysRevB.82.104511 https://doi.org/10.1063/1.1707989 https://doi.org/10.1103/PhysRevB.53.6634 https://doi.org/10.1103/PhysRevLett.86.886 https://doi.org/10.1103/PhysRevB.66.132505 https://doi.org/10.1103/PhysRevLett.97.237005 https://doi.org/10.1103/PhysRevLett.97.237005 https://doi.org/10.1016/0921-4534(93)90497-E https://doi.org/10.1103/PhysRevB.55.6636 1. Introduction 2. Interaction potential for two stacks of pancakes 2.1. Basic equations 2.2. Interaction potential of tilted stacks of pancakes 2.3. Interaction potential of zigzag stacks of pancakes 2.4. Interaction potential of two stacks in the crossing lattices 3. Vortex molecules 4. Conclusions