1: \documentclass[aps,prb,twocolumn]{revtex4}
2: \usepackage{graphicx}% Include figure files
3:
4: \textwidth 16.5cm
5: \hoffset 0.65cm
6: \voffset 0.7cm
7:
8: \begin{document}
9: \draft
10: \title{Transverse and secondary voltages in
11: $\rm{Bi_{2}Sr_{2}CaCu_{2}O_{8}}$ single crystals}
12: \author{K. Vad$^1$, S. M\'esz\'aros$^1$, and B. Sas$^2$}
13: \address{Institute of Nuclear Research, P.O.Box 51, H--4001 Debrecen,
14: Hungary\\
15: $^{2}$ Research Institute of Solid State Physics, P.O.Box 49,
16: H--1525 Budapest, Hungary}
17:
18: \date{\today}
19:
20: \begin{abstract}
21: Multicontact configuration is one of the most powerful arrangements
22: for electrical transport measurements applied to study vortex phase
23: transition and vortex phase dimensionality in strongly anisotropic
24: high-$T_c$ superconducting materials. In this paper we present electrical
25: transport measurements using a multiterminal configuration, which prove
26: both the existence of guided vortex motion in
27: ${\rm{Bi_{2}Sr_{2}CaCu_{2}O_{8}}}$ single crystals near the
28: transition temperature and that secondary voltage in zero external magnetic
29: field is induced by thermally activated vortex loop unbinding. The phase
30: transition between the bound and unbound states of the vortex loops was
31: found to be below the temperature where the phase coherence of the
32: superconducting order parameter extends over the whole volume of the
33: sample. We show experimentally that 3D/2D phase transition in vortex
34: dimensionality is a length-scale-dependent layer decoupling process and
35: takes place simultaneously with the 3D/2D phase transition in
36: superconductivity at the same temperature.
37: \end{abstract}
38: \pacs{74.25.Dw; 74.25.Fy; 74.72.Hs}
39: \maketitle
40:
41: \section{Introduction}
42: In the mixed state of isotropic type II superconductors the vortices can
43: move in the direction of the Lorentz force due to transport currents. The
44: angle between the local transport current density and the direction of the
45: electric field induced by vortex motion is the Hall angle, which can
46: be determined experimentally by measuring the Hall voltage. If the
47: superconducting system has structural inhomogeneities, the direction
48: of vortex velocity can differ from that of the Lorentz force. This
49: misalignment between the directions of vortex velocity and the Lorentz
50: force can be identified by observing an even transverse voltage, which, as
51: opposed to the odd Hall voltage, does not change sign upon magnetic field
52: reversal. This transverse voltage can be attributed to guided vortex motion,
53: where vortices move along in a particular direction determined by the
54: structure of the sample.
55:
56: Guided vortex motion has been observed in classical type II
57: superconductors \cite{danckwert} and is expected to be more pronounced
58: in high-$T_c$ materials. In ${\rm {YBa_{2}Cu_{3} O_{7-\delta}}}$
59: single crystals it can be induced by twin boundaries, as it has been proved
60: by magneto-optical technique \cite{vlasko}, magnetization \cite{oussena}
61: and transverse voltage \cite{chabanenko} measurements. These results
62: confirm that vortices move in a channel, formed at one edge of a twin
63: boundary along the twin planes. The other possibility of guided vortex
64: motion is the intrinsic channeling between CuO$_2$ planes due to intrinsic
65: pinning. This effect in La-Sr-Cu-O superconductors was observed by
66: C.A. Dur\'{a}n {\em et al.} \cite{duran} and in
67: ${\rm {YBa_{2} Cu_{3}O_{7-\delta }}}$ films by P. Berghuis
68: {\em et al.} \cite{berghuis}. Channel formation in 2H-NbSe$_2$ at
69: relatively low temperatures was experimentally studied by Z. Xiao {\em et
70: al.} \cite{xiao}. It was shown that depinning of a strongly pinned vortex
71: lattice starts through the formation of weakly pinned regions in which the
72: vortices start moving first. In ${\rm {Bi_{2}Sr_{2}CaCu_{2} O_{x}}}$
73: guided vortex motion was studied by the Hall resistivity measurements in
74: textured samples \cite{vasek}. To our knowledge, in
75: ${\rm {Bi_{2} Sr_{2}CaCu_{2}O_{8}}}$ single crystals the existence of
76: guided vortex motion has not been proven yet.
77:
78: Of high-$T_c$ superconducting compounds
79: $\rm{Bi_{2}Sr_{2}CaCu_{2}O_{8}}$ is preferred because, due to its
80: extremely high anisotropy, its discrete superconducting layers play an
81: important role in current conduction properties even in the transition
82: temperature range, and, due to the highly anisotropic conductivity, the
83: vortex matter has a very rich phase diagram with numerous phase
84: transitions. The multicontact configuration is one of the most promising
85: arrangements of electrical transport measurements in the study of
86: superconducting phase transition and vortex dimensionality. In this
87: configuration four electrical contacts are attached to one side of a single
88: crystal and two or four contacts to the opposite side. If the current injected
89: into one side of a crystal is high enough, a voltage drop on both sides can
90: be measured. The side where the current is injected is called primary, the
91: opposite side is secondary. Due to this contact arrangement and the
92: anisotropic conductivity, a non-uniform current distribution develops in the
93: sample, and the current injected into the primary side is confined to a very
94: thin surface layer, which is thinner than the thickness of the crystal. If the
95: pancakes in adjacent layers belonging to the same vortex line are strongly
96: coupled to each other, the primary current induces vortex motion not only
97: in the primary, but also in the secondary surface, causing non-local
98: dissipation in the sample. Moreover, if the coupling strength between the
99: pancakes is high enough to produce three-dimensional (3D) type vortex
100: lines across the sample, primary and secondary voltages can be equal. This
101: configuration reminds us of the geometry used by I. Giaever in his famous
102: experiment \cite{giaever}, which is called dc flux transformer
103: configuration.
104:
105: The first measurements on $\rm{Bi_{2}Sr_{2}CaCu_{2}O_{8}}$ single
106: crystals using the dc flux transformer configuration were performed by H.
107: Safar {\em et al}. \cite{safar} and R. Busch {\em et al}. \cite{busch}.
108: They observed the dimensionality of the vortex system over a wide range
109: of the phase diagram. Y.M. Wan {\em et al}. \cite{wan1,wan2} and C.D.
110: Keener {\em et al}. \cite{keener} also used the multicontact dc flux
111: transformer configuration to study the secondary voltage and found that in
112: magnetic fields near the transition temperature the interlayer vortex coupling
113: was responsible for the secondary voltage. However, the origin of the
114: secondary voltage in zero external magnetic field still remained problematic.
115: S.W. Pierson \cite{pierson1} suggested that it originated from thermally
116: activated vortex loop unbinding. A vortex loop is a correlated
117: vortex-antivortex line pair. Using a real-space renormalization group
118: analysis, the author identified three characteristic critical currents and
119: calculated their temperature dependence. He found that the temperature
120: dependence of the secondary voltage is a horizontal slice of the
121: current-temperature phase diagram.
122:
123: In this paper we present electrical transport measurements using a
124: multiterminal configuration that prove the existence of guided vortex
125: motion in chemically and mechanically homogeneous
126: ${\rm{Bi_{2}Sr_{2}CaCu_{2}O_{8}}}$ single crystals. We also
127: present the implications of experimental results for the 3D/2D phase
128: transition in vortex dimensionality near the Ginzburg-Landau transition
129: temperature and show that this phase transition is a
130: length-scale-dependent layer decoupling process. We show that the
131: temperature dependence of the secondary voltage in zero magnetic field
132: has a double peak structure.
133:
134: \section{Experimental arrangement}
135: Single-crystalline $\rm{Bi_{2}Sr_{2}CaCu_{2}O_{8}}$ compounds
136: were prepared by the melt cooling technique, described in
137: Reference \cite{keszei}. Optically smooth rectangular crystals were
138: carefully cleaved from these compounds, and heated in flowing oxygen for
139: 15 minutes at 900 K in order to stabilize the oxygen content. Chemical
140: homogeneity of the samples was checked by microbeam PIXE and
141: (O$^{16}, \alpha $) resonant elastic scattering \cite{simon} with a spatial
142: resolution of 5 $\mu $m. No chemical inhomogeneities were identified. The
143: surface smoothness was measured by atomic force microscopy. The
144: surfaces were found to be flat with a typical roughness of 10 nm. Before
145: preparing the electrical contacts, the sample quality was checked by
146: magnetization measurements using a SQUID or a vibrating sample
147: magnetometer, and by AC susceptibility measurements. Electrical contacts
148: were made by bonding 25 $\mu $m gold wires with Dupont 6838 silver
149: epoxy fired for five minutes at 900 K. The contact resistance was a few
150: ohms. The geometrical position of electrical contacts was precisely
151: measured by optical microscope.
152:
153: Two current and four potential electrodes were attached to both faces of
154: the crystals. The scheme of the electrode configuration is shown in
155: Fig.\ \ref{fig1}. The current was injected into one face of the single crystal
156: through the current contacts (this is the primary current $I_P$), while
157: primary and secondary longitudinal voltages, primary and secondary
158: transverse voltages, and the voltage between the two faces were recorded
159: simultaneously, using a six-channel data acquisition system. Applying this
160: symmetrical contact arrangement, this configuration made it possible to
161: check the sample homogeneity from the point of view of electrical
162: conductivity.
163:
164: We have performed measurements on a number of crystals, but in this
165: paper we present the results we measured on two of them fabricated from
166: the same batch (samples $A$ and $B$). The mean-field Ginzburg-Landau
167: (GL) transition temperatures $T_{c0}$ of the samples were 88.5 K and
168: 88 K, respectively. The sample dimensions were about $1\times 1.5$ mm
169: $^2$, the thickness was 8 and 3 $\mu $m, respectively. As we pointed out
170: in our previous papers \cite{sas}, most of the Joule heat due to the
171: transport current was generated in the current contacts. In order to reduce
172: the heat dissipation, and also to eliminate the thermoelectric force, we used
173: current pulses with a duration time of 1 ms and repetition time of 100 ms.
174: This arrangement and the fact that the sample temperature was regulated by
175: a temperature controlled He gas stream instead of exchange gas, made it
176: possible to avoid heating during the pulse up to the amplitude of 10 mA.
177: \begin{figure}
178: \includegraphics[width=6cm, height=3.5cm]{fig1.eps}
179: \caption{Electrode configuration used for transverse and secondary
180: voltage measurements. For definition of voltages see text. $I_P$
181: denotes the primary current.}
182: \label{fig1}
183: \end{figure}
184:
185: In our experiments the following voltages, shown in Fig.\ \ref{fig1}, were
186: simultaneously recorded: (i) longitudinal voltages measured parallel to the
187: current direction on the surface of the crystal where the current was
188: injected and on the opposite surface ($V_{PL}$ and $V_{SL}$, the
189: primary and secondary longitudinal voltages, respectively); (ii) the primary
190: and secondary transverse voltages, measured perpendicularly to the current
191: direction ($V_{PT}$ and $V_{ST}$); (iii) the $c$-axis direction voltage
192: measured between the two surfaces of the crystal ($V_{C}$).
193:
194: Using the analysis of R. Busch {\em et al}. \cite{busch}, we could define
195: the temperature dependence of the $ab$ plane and $c$-axis resistivities
196: from the voltages $V_{PL}$ and $V_{SL}$. For the anisotropy ratio
197: $\gamma$ we received
198: $\gamma =\sqrt(\rho _{c}/\rho _{ab}) \approx 500$ with
199: $\rho _{ab} \approx 100$ $\mu \Omega$cm at 90 K.
200:
201: \section{Experimental results}
202: The temperature dependence of the primary longitudinal and transverse
203: voltages, $V_{PL}$ and $V_{PT}$, of a
204: ${\rm{Bi_{2}Sr_{2} CaCu_{2}O_{8}}}$ single crystal (sample $A$)
205: measured in zero and 1 T magnetic fields is shown in Fig.\ \ref{fig2}(a and
206: b). In zero magnetic field $V_{PT}$ shows a sharp change in the transition
207: temperature range with a sign reversal. This sign reversal also exists in
208: weak magnetic fields and seems to be a universal characteristic of vortex
209: motion. Fig.\ \ref{fig2}(c) shows the calculated ratio of the measured
210: primary transverse and longitudinal voltages in zero and 1 T magnetic
211: fields. The maximum value of this ratio in zero magnetic field is 6, and it is
212: larger than 1 in a temperature range of $\sim 3$ K. In 1 T magnetic field
213: the maximum value is only 1.5 and the temperature range where the ratio is
214: larger than 1 is 14 K wide. The Hall voltage can be calculated from the
215: difference of the two transverse voltages measured in a magnetic field
216: (here 1 T) with opposite field directions (Fig.\ \ref{fig2}(b)).
217: \begin{figure}
218: \includegraphics[width=6.5cm, height=8cm]{fig2.eps}
219: \caption{Temperature dependence of the primary longitudinal voltage
220: $V_{PL}$ (a), primary transverse voltage $V_{PT}$ (b) and the absolute
221: value of the $V_{PT}/V_{PL}$ ratio (c) of a
222: $Bi_{2}Sr_{2}CaCu_{2}O_{8}$ single crystal measured in zero and 1 T
223: magnetic fields. In (b) $V_{PT}$ is measured in two opposite magnetic
224: field directions. The magnetic field is perpendicular to the $ab$ planes,
225: $I_P$= 1 mA.}
226: \label{fig2}
227: \end{figure}
228:
229: On the secondary side of the crystal we found that in zero magnetic field
230: the ratio of transverse and longitudinal voltages, $V_{ST}$ and
231: $V_{SL}$, is higher than 1 in a temperature range of $\sim 0.7$ K
232: (Fig.\ \ref{fig3}). This range is about 4 times narrower than it is on the
233: primary side. In higher magnetic fields, with the field perpendicular to the
234: $ab$ plane, we could not determine this ratio because both $V_{ST}$ and
235: $V_{SL}$ decreased to lower than 100 nV at 5 mA measuring current
236: already in 100 mT magnetic field. Sample $B$, prepared to study the
237: parallel case, was placed in the cryostat so that the $ab$ planes were
238: parallel to the magnetic field direction. In this arrangement $V_{SL}$ is
239: detectable in high magnetic fields. The result is presented in Fig.\
240: \ref{fig4}.
241:
242: \begin{figure}
243: \includegraphics[width=6.5cm, height=7cm]{fig3.eps}
244: \caption{Temperature dependence of the secondary longitudinal
245: $V_{SL}$ and transverse $V_{ST}$ voltages (a) and the absolute value
246: of the $V_{ST}/V_{SL}$ ratio (b), $I_P$=1 mA.}
247: \label{fig3}
248: \end{figure}
249:
250: \begin{figure}
251: \includegraphics[width=6cm, height=4.5cm]{fig4.eps}
252: \caption{Secondary and primary voltages vs. temperature in different
253: magnetic fields (sample $B$). The magnetic field is parallel to the $ab$
254: planes, $I_{P}$ = 1 mA. $T_{UB}$, $T_c$ and $T_{c0}$ denote the
255: unbinding, transition and Ginzburg-Landau transition temperature,
256: respectively. The different vortex dimensionalities are also shown.}
257: \label{fig4}
258: \end{figure}
259:
260: In zero applied magnetic field the secondary voltage is determined by
261: interlayer vortex coupling. In order to gain information about this coupling
262: strength, we measured the voltage $V_C$ between the two surfaces of the
263: crystal while applying a primary current $I_P$. Depending on the current
264: distribution in the crystal, some part of this current flows in the $c$-axis
265: direction and $V_C$ depends on this current. The temperature
266: dependence of $V_C$ is of metallic ($d\rho/dT>0$) type, except in a small
267: temperature range near the mean-field Ginzburg-Landau transition
268: temperature. In Fig.\ \ref{fig5}(a) this range is between 85 and 86 K. This
269: temperature dependence is also reflected in the current-voltage
270: characteristics (Fig.\ \ref {fig5}(b)). Far above the GL transition
271: temperature, at 256 K the current-voltage characteristic is linear. Near the
272: GL transition temperature, at 87.3 K there is a slight curvature in it. On
273: further cooling the sample from 87.3 K to 86 K the curvature increases. In
274: the temperature range between 86 K and 85 K the $I_{P}-V_C$ curves are
275: concave in shape (for clarity in Fig.\ \ref{fig5} we present only the curve
276: measured at 85.8 K). At temperatures lower than 85 K both the
277: $V_{C}-I_{P}$ and $V_{C}-T$ curves show the same metallic
278: behaviour, as above 86 K.
279: \begin{figure}
280: \includegraphics[width=6cm, height=9cm]{fig5.eps}
281: \caption{Temperature dependence of $V_{C}$ at 1 mA primary current
282: (a), and current-voltage characteristics measured at different temperatures
283: (b) in zero magnetic field. The arrows in (a) denote the temperatures where
284: the current-voltage characteristics were measured.}
285: \label{fig5}
286: \end{figure}
287:
288: \section{Discussion}
289: \subsection{Guided vortex motion}
290: The experimental results discussed above can help us to understand the
291: origin of transverse voltages. In general, transverse voltage consists of
292: three components: a component coming from the geometrical misalignment
293: of the contacts, a Hall voltage component, and a component originating
294: from the guided motion of vortices. The geometrical misalignment can be
295: detected by normal-state measurements. The other two components can be
296: distinguished by comparing the voltages measured in two opposite
297: magnetic field directions: while in the Hall voltage the sign of the measured
298: voltage depends on the external magnetic field direction because the vortex
299: motion is only governed by the Lorentz force, in the case of the guided
300: vortex motion it is independent of the field direction. This difference gives
301: us the chance to distinguish between these two voltage generation
302: mechanisms. On the basis of our results we can conclude that the guided
303: motion of vortices is reflected in both $V_{PT}$ and $V_{ST}$
304: transverse voltages (Fig.\ \ref{fig2} and \ref{fig3}), and that in our samples
305: the guided vortex motion is a dominating effect in the Hall voltage
306: measurements, since the Hall voltage is superimposed on a voltage, which
307: is found to be much higher than the Hall voltage itself (Fig.\ \ref{fig2}(b)).
308:
309: In our electrode configuration the main part of the injected current $I_P$
310: flows in the $ab$ planes, but has a component in the $c$-axis direction,
311: too. If the external magnetic field is parallel to the $c$-axis, the latter
312: component of the current is parallel to the vortex lines, so, it does not
313: contribute to the Lorentz force. The dominating component of the
314: transverse voltage is brought about by the intralayer current.
315:
316: There is a well-defined temperature range, where the ratio of transverse and
317: longitudinal voltages is larger than 1, while the geometrical misalignment
318: between the transverse voltage contacts is less than 10 \% of the distance
319: between the longitudinal voltage contacts, as it was measured by an optical
320: microscope. This supports the view that the transverse voltage cannot
321: originate from a geometrical misalignment of the potential contacts.
322: Chemical inhomogeneity does not explain this phenomenon, either. We
323: believe that these results can be understood within the concept of guided
324: vortex motion. If the magnetic field is perpendicular to the $ab$ planes,
325: slight inhomogeneities in pinning potentials can bring about channels for
326: moving vortices, which results in guided motion. Some vortices are
327: localized by stronger pinning centers and form trapped vortex regions,
328: other mobile vortices flow in channels between these trapped regions. The
329: existence of such vortex channels decreases the critical current because the
330: effective pinning potential is lower, consequently, vortex mobility is higher
331: in a channel than outside. Guided vortex motion is sensitive to applied
332: magnetic fields. Higher magnetic fields restricts the guided motion of
333: vortices through vortex-vortex interaction.
334:
335: \subsection{Secondary longitudinal voltage}
336: In 2D superconducting layers the phase fluctuation of the order parameter
337: generates vortex-antivortex pairs as topological excitations\cite{tinkham}.
338: The phase transition in an isolated 2D superconducting layer, where the
339: vortex-antivortex pairs are bound below the phase transition temperature
340: and are unbound above it, is described by the Kosterlitz-Thouless theory
341: \cite {kosterlitz,halperin}. In thin films, if the film thickness is less than the
342: superconducting coherence length, this Kosterlitz-Thouless phase
343: transition ($T_{KT}$) can be observed.
344:
345: In high-temperature superconductors, the high transition temperature, short
346: coherence length and layered structure make the phase fluctuation of the
347: order parameter dominant over the other fluctuations in the transition
348: temperature range, but the coupling between superconducting layers
349: modifies the 2D Kosterlitz-Thouless picture. The interaction between
350: superconducting layers leads to vortex-antivortex interaction different from
351: the 2D case. In $\rm{Bi_{2}Sr_{2}CaCu_{2}O_{8}}$ single crystals the
352: CuO$_2$ planes serve as 2D superconductor layers, the vortices are the
353: pancake vortices. Due to the layered structure a high anisotropy exists in
354: conductivity and the vortex matter has a very rich phase diagram with
355: numerous phase transitions. Among others the 3D phase appears, where
356: the coupling between neighbouring layers arranges the thermally excited
357: pancake vortices into 3D flux lines\cite{blatter}. These 3D flux lines can
358: form vortex loops as correlated vortex-antivortex line pairs. They are called
359: thermally activated vortex loops, because they are the result of a combined
360: effect of thermally activated vortex excitation and interlayer vortex
361: coupling.
362:
363: The 3D character modifies the structure of the phase transition
364: between the bound and unbound states. Although the 2D signatures are
365: prevalent, a narrow 3D window appears around the phase transition
366: temperature ($T_c$) and a nonzero critical current appears. The 3D
367: temperature region is theoretically predicted and it has been shown that the
368: size of this 3D region is much smaller above $T_c$ than
369: below\cite{pierson2}. Above the phase transition temperature the
370: behaviour of vortices becomes 2D due to decoupling of the
371: superconducting layers (3D/2D phase transition). It has been also shown in
372: another paper\cite{pierson3} that this layer decoupling is a
373: length-scale-dependent process: the layers become decoupled at length
374: scales larger than an interlayer screening length, while for lengths below this
375: scale they remain coupled. The manifestation of the 3D/2D transition in
376: electrical transport properties has not been investigated before.
377:
378: In an isolated 2D superconducting layer the vortex pair energy is the
379: sum of the creation energy of the two vortices and the intralayer logarithmic
380: coupling energy between vortices. In layered systems like
381: $\rm{Bi_{2}Sr_{2}CaCu_{2}O_{8}}$ this vortex pair energy is modified
382: by the interlayer Josephson coupling, which not only strengthens the
383: intralayer interaction, but causes an interaction between vortices located in
384: neighbouring layers. The intralayer vortex pair energy can be written in the
385: following simple form:
386: $E(r)=2E_{c}+K_{\parallel} ln(r/\xi_{0})+K_{\perp} r/\xi_{0}$,
387: where $r$ is the distance between the vortex and antivortex, $E_{c}$ is the
388: creation energy of a vortex, $\xi_{0}$ is the zero-temperature correlation
389: length, $K_{\parallel}$ is the intralayer vortex-vortex coupling constant and
390: $K_{\perp}$ is the interlayer Josephson coupling constant. While the 2D
391: logarithmic interaction (second term in this equation) dominates at short
392: distances, the Josephson-coupling mediated 3D linear interaction (third
393: term) dominates at large distances. The intralayer vortex length scale
394: $R_{\lambda}(T)$ is the characteristic length which divides this
395: logarithmic and linear regimes\cite{pierson3}.
396: $R_{\lambda}(T)=\xi_{0}/\sqrt \lambda$, where $\lambda$ is the ratio of
397: the interlayer Josephson coupling to the intralayer coupling,
398: $\lambda=K_{\perp}/K_{\parallel}$. $\lambda$ depends on the size $r$ of
399: the vortex pairs, and above $T_c$ it has a maximum. The size $r$ which
400: belongs to this maximum $\lambda$ is the other characteristic length, the
401: interlayer screening length ${\it l}\,_{3D/2D}(T)$. If the separation
402: between two vortices located in neighbouring layers is larger than
403: ${\it l}\,_{3D/2D}$, the Josephson-coupling mediated linear interaction is
404: screened out and the layers are decoupled. Around $T_c$ the dimensional
405: behaviour of the system is determined by these two characteristic lengths.
406: The layers are coupled and the behaviour of vortices is 3D if the separation
407: between vortices is greater than the intralayer vortex length scale
408: $r>R_{\lambda}(T)$ and less than the interlayer screening length
409: $r<{\it l}\,_{3D/2D}(T)$. The temperature dependence of
410: $R_{\lambda}(T)$ and ${\it l}\,_{3D/2D}(T)$ shows\cite{pierson4} that
411: the intralayer vortex length scale $R_{\lambda}(T)$ is constant for
412: $T \ll T_c$, but it increases as the temperature approaches $T_c$ from
413: below. The interlayer screening length ${\it l}\,_{3D/2D}(T)$ decreases
414: continuously as the temperature increases.
415:
416: While in the 3D regime below $T_{c}$ the electrical transport behaviour is
417: dominated by vortex loops, above $T_{c}$ it is dominated by vortex lines
418: and pancake vortices. The multiterminal configuration is a good
419: arrangement to distinguish between these two regimes. In this paper
420: we show that studying the temperature dependence of the secondary
421: voltage can help us to understand the length-scale-dependent layer
422: decoupling, i.e., the 3D/2D phase transition.
423:
424: The secondary voltage has already been studied both theoretically
425: \cite{pierson1} and experimentally\cite{wan1}. It can now be accepted that
426: in zero applied magnetic field it originates from thermally activated vortex
427: loop unbinding. At low temperatures where $V_{SL}$ and $V_{PL}$ are
428: zero, the thermally excited 3D flux lines form vortex loops which are
429: 'pinned' to the crystal. With increasing temperature, the transport current
430: splits these vortex loops into free vortex-antivortex line pairs. The
431: temperature where this splitting starts is the unbinding temperature
432: ($T_{UB}$). This is the lowest temperature where both $V_{SL}$ and
433: $V_{PL}$ are observable. Above the unbinding temperature the free
434: vortices move in the sample like 3D vortex lines due to the Lorentz force,
435: producing the same voltage drop on the primary and secondary side of the
436: crystal, $V_{PL}=V_{SL}$. This 3D character of the vortex lines remains
437: up to a temperature where the secondary voltage has a local maximum. At
438: sample $B$ the 3D temperature range is around 85 K where the zero field
439: primary and secondary longitudinal voltages $V_{PL}(0T)$ and
440: $V_{SL}(0T)$ coincide (see Fig.\ \ref{fig4}). This is the same 3D
441: temperature range which was predicted theoretically by renormalization
442: group analysis \cite{pierson2}. With increasing temperature the 3D
443: character of flux motion disappears and consequently $V_{SL}$ becomes
444: lower than $V_{PL}$, but another local maximum of $V_{SL}$ can be
445: found as the temperature approaches $T_{c0}$. The temperature
446: dependence of the secondary voltage has two peaks with a higher and a
447: lower amplitude.
448:
449: This double peak structure of $V_{SL}(0T)$ can be explained by the
450: motion of different types of vortex lines. In zero applied magnetic field free
451: vortex lines can be produced in two ways. First, they can be the result
452: of vortex-antivortex depairing of thermally activated vortex loops due
453: to the Lorentz force of the transport current. In this case the number of
454: free vortex lines depends on the transport current density and a non-Ohmic
455: behaviour characterizes the system. Secondly, free vortex lines can be
456: spontaneously created by thermal activation, mainly above $T_{c}$. The
457: number of free vortex lines increases with increasing temperature and the
458: system is characterized by an Ohmic behaviour.
459:
460: The effect of current on vortex-antivortex depairing is twofold. On the one
461: hand the current reduces the creation energy whereby increases the density
462: of vortex pairs. On the other hand the current exerts a force (the Lorentz
463: force) on vortex loops and can blow them out. The number of blowouts
464: depends on the size $r$ of the vortex pairs. If $r$ is higher than the
465: intralayer vortex length scale $R_{\lambda}(T)$, the Josephson-coupling
466: mediated 3D linear interaction is energetically favoured over 2D intralayer
467: logarithmic interaction and the energy of a vortex loop is smaller than the
468: energy of a pair of independent vortex lines. Below $T_c$ in
469: $\rm{Bi_{2}Sr_{2}CaCu_{2}O_{8}}$
470: $R_{\lambda}(T)=\xi_{0}/\sqrt \lambda \approx 1 \mu$m, where
471: $\xi_{0} \sim 3$nm and $\lambda \sim 10^{-5}$. Consequently, if
472: $r>1 \mu$m, creation of vortex loops is energetically favoured over free
473: vortex-antivortex line pairs. This happens in the 3D temperature range
474: where the dominant topological excitation is the vortex loop. With
475: increasing temperature the number of blowouts and, so, the longitudinal
476: secondary voltage increases. However, approaching the transition
477: temperature the intralayer vortex length scale $R_{\lambda}(T)$ starts to
478: increase, therefore the number of vortex pairs which can be blown out by
479: the transport current decreases which results in the decrease of the
480: secondary voltage. The temperature which belongs to the peak value of the
481: secondary voltage is the transition temperature $T_c$. Although at this
482: temperature the behaviour of vortices is still 3D type as it was shown
483: theoretically\cite{pierson4}, the secondary voltage is somewhat lower than
484: the primary voltage (see Fig.\ \ref{fig4}). This means that the secondary
485: voltage starts to decrease before layer decoupling. At higher temperatures
486: the 3D character of flux motion disappears and $V_{SL}$ becomes
487: significantly lower than $V_{PL}$. Above the 3D temperature range
488: another local maximum of $V_{SL}$ can be found as the temperature
489: approaches $T_{c0}$, because the temperature is high enough to
490: produce free vortex lines by thermal activation. The number of free vortex
491: lines and, so, the secondary voltage increases with increasing temperature.
492: The system is characterized by a continuous 3D/2D transition due to
493: continuous decrease of the interlayer screening length
494: ${\it l}\,_{3D/2D}(T)$ as the temperature approaches the mean-field
495: Ginzburg-Landau transition temperature $T_{c0}$. Near $T_{c0}$ the
496: amplitude of the order parameter decreases, just as the number of the free
497: vortex lines. This effect evokes the decrease of the secondary voltage. The
498: temperature which belongs to the minimum secondary voltage above the
499: double peak structure is $T_{c0}$.
500:
501: The local maxima of $V_{SL}$ can be decreased by magnetic field. If the
502: magnetic field is perpendicular to the $ab$ planes, the external field
503: prevents the free motion of vortex-antivortex line pairs. While one part of a
504: pair, which is parallel to the external magnetic field, cannot move in the
505: sample because of the vortex-vortex interaction, the other antiparallel part
506: is annihilated in the external magnetic field. If the magnetic field is parallel
507: to the $ab$ planes, there is no interaction between the vortex lattices
508: formed by the external magnetic field and vortex loops, and the secondary
509: voltage is not so sensitive to the applied magnetic field. In this arrangement
510: the magnetic field reduces the coupling strength between the CuO$_2$
511: bilayers by decreasing both the magnetic and Josephson coupling. This has
512: a twofold consequence. First, the thermally excited 2D vortex-antivortex
513: pair creation in CuO$_2$ bilayers is suppressed by magnetic field because
514: the creation of a vortex-antivortex pair is equivalent to the creation of a
515: dislocation in the interlayer flux line lattice which increases the elastic
516: energy of the lattice. Secondly, the magnetic field reduces the probability
517: of arrangement of thermally excited pancake vortices into 3D flux lines.
518: That is why only one peak exists in low magnetic fields (50 mT) instead of
519: a double peak (Fig.\ \ref{fig4}).
520:
521: Information about the strength of Josephson coupling and the 2D/3D phase
522: transition in superconductivity can be obtained by study the temperature
523: dependence of $I_{P}-V_C$. In the temperature range where the
524: $I_{P}-V_C$ curves are concave (between 85 and 86 K) the
525: superconducting coherent state exists in the CuO$_2$ bilayers, but the
526: interlayer coupling is not strong enough to establish the $c$-axis
527: coherence. In consequence, the current in $c$-axis direction is carried by
528: single particle tunnelling instead of Cooper pair tunnelling (2D type
529: superconductivity). With the decrease in temperature, at 85 K, the
530: Josephson coupling, and consequently, a phase coherence between
531: superconducting bilayers develops and extends over the whole volume of
532: the sample, producing maximum values in both $V_{SL}$ and $V_C$
533: (2D/3D phase transition). At temperatures lower than 85 K the Josephson
534: coupling energy increases, the phase coherence of the superconducting
535: order parameter extends over the whole volume of the sample and
536: develops the 3D type superconductivity. The higher peak value of
537: $V_{SL}$ is also at 85 K which is the upper end of the temperature range
538: where vortex lines have 3D character. Consequently 2D/3D phase transition
539: in superconductivity and in vortex dimensionality takes place at the same
540: temperature. This can be valid inversely, too. If the vortex dimensionality
541: decreases from 3D to 2D, the dimensionality of superconductivity can also
542: decrease. This result was experimentally supported by transport current
543: measurements\cite{pethes}, where the authors proved that
544: the Bardeen-Stephen model for the flux flow resistance
545: $\rho _{f}=\rho _{n}\cdot B/B_{c2}$ ($\rho _{n}$, $B$ and $B_{c2}$
546: are the normal state resistivity, magnetic field and critical magnetic field) is
547: not valid at high current density in $\rm{Bi_{2}Sr_{2}CaCu_{2}O_{8}}$
548: single crystals. Due to intensive flux motion both the phase coherence
549: between superconducting bilayers and the interlayer screening length
550: ${\it l}\,_{3D/2D}(T)$ decrease resulting in a 3D/2D phase transition in
551: dimensionality of superconductivity.
552:
553: Thermally unbound vortex-antivortex line pairs take part in guided motion.
554: The maximum values of $\mid V_{PT}/V_{PL}\mid $ and
555: $\mid V_{ST}/V_{SL}\mid$ are near the unbinding temperature in the 3D
556: type superconducting regime. So, the guided vortex motion has a 3D
557: character in $\rm{Bi_{2}Sr_{2}CaCu_{2}O_{8}}$ single crystals and
558: develops due to melting of vortex loop lattice. The external magnetic field
559: which is perpendicular to the $ab$ planes restricts the guided motion of
560: vortices through vortex-vortex interaction. That is why weak magnetic
561: fields can prevent the development of secondary transverse voltage.
562:
563: \section{Conclusions}
564: In conclusions, we found that temperature dependence of secondary
565: longitudinal voltage has a double-peak structure and its higher maximum
566: value is at the temperature where the phase coherence of the order
567: parameter extends over the whole sample. Secondary
568: voltage originates from correlated vortex-antivortex line pair unbinding, i.e.,
569: from vortex loop unbinding due to the Lorentz force of the transport
570: current. Near $T_{c0}$ free vortex-antivortex line pairs are also generated
571: by thermally activated vortex excitation. We think that the two types of
572: vortex-antivortex line pairs are responsible for the double peak structure of
573: the secondary longitudinal voltage. Lacking of theories describing the
574: double peak structure in $\rm{Bi_{2}Sr_{2}CaCu_{2}O_{8}}$ single
575: crystals indicates a need for better description of length-scale dependence
576: in layered superconductors. In order to improve the theoretical description,
577: one can perform the renormalization group analysis of Reference
578: \cite{pierson5} for non-constant current density or one can use different
579: renormalization group methods, e.g., field theoretical renormalization group
580: approaches\cite{nandori}.
581:
582: \section{Acknowledgement}
583: We take great pleasure in acknowledging discussion with P.F. de
584: Ch\^{a}tel and I. N\'{a}ndori. This work was supported by the Hungarian
585: Science Foundation (OTKA) under contract no. T037976.
586:
587: \begin{references}
588: \bibitem{danckwert} M. Danckwert, A.R. Go\~{n} and C. Thomsen,
589: Phys. Rev. B {\bf 59}, R6624 (1999 ).
590: \bibitem{vlasko} V.K. Vlasko-Vlasov, L.A. Dorosinskii, A.A. Polyaskii,
591: V.I. Nikitenko, U. Welp, B.W. Veal and G.W. Crabtree, Phys. Rev. Lett.
592: {\bf 72}, 3246 (1994).
593: \bibitem{oussena} M. Oussena, P.A.J. de Groot, S.J. Porter, R. Gagnon
594: and L. Taillefer, Phys. Rev. B {\bf 51}, 1389 (1995).
595: \bibitem{chabanenko} V.V. Chabanenko, A.A. Prodan, V.A. Shklovskij,
596: A.V. Bondarenko, M.A. Obolemskii, H. Szymczak and S. Piechota,
597: Physica C {\bf 314}, 133 (1999).
598: \bibitem{duran} C.A. Dur\'an, P.L. Gammel and D.J. Bishop, J. Appl.
599: Phys. {\bf 79}, 6592 (1996).
600: \bibitem{berghuis} P. Berghuis, E. Di Bartolomeo, G.A. Wagner and J.E.
601: Evetts, Phys. Rev. Lett. {\bf 79}, 2332 (1997).
602: \bibitem{xiao} Z. Xiao, L.E.Y. Andrei, P. Shuk and M. Greenblatt, Phys.
603: Rev. Lett. {\bf 86}, 2431 (2001).
604: \bibitem{vasek} P. Va\u{s}ek, I. Jane\u{c}ek and V. Plech\'a\u{c}ek,
605: Physica C {\bf 247}, 381 (1995).
606: \bibitem{giaever} I. Giaever, Phys. Rev. Lett. {\bf 15}, 825 (1965).
607: \bibitem{safar} H. Safar, E. Rodriguez, F. de la Cruz, P.L. Gammel, L.F.
608: Schneemeyer and D.J. Bishop, Phys. Rev. B {\bf 46}, 14238 (1992).
609: \bibitem{busch} R. Busch, G. Ries, H. Werthner, G. Kreiselmeyer and G.
610: Saemann-Ischenko, Phys. Rev. Lett. {\bf 69}, 522 (1992).
611: \bibitem{wan1} Y.M. Wan, S.E. Hebboul, D.C. Harris and J.C. Garland,
612: Phys. Rev. Lett. {\bf 71}, 157 (1993).
613: \bibitem{wan2} Y.M. Wan, S.E. Hebboul and J.C. Garland, Phys. Rev.
614: Lett.{\bf 72}, 3867 (1994).
615: \bibitem{keener} C.D. Keener, M.L. Trawick, S.M. Ammirata, S.E.
616: Hebboul and J.C. Garland, Phys. Rev. B {\bf 55}, R708 (1997); Phys.
617: Rev. Lett. {\bf 78}, 1118 (1997).
618: \bibitem{pierson1} S.W. Pierson, Phys. Rev. Lett. {\bf 74}, 2359 (1995);
619: Phys. Rev. B {\bf 55}, 14536 (1997).
620: \bibitem{keszei} B. Keszei, Gy. Szab\'o, J. Vandlik, L. Pog\'any, and G.
621: Oszl\'anyi, J. Less Common Metals {\bf 155}, 229 (1989);. J.R. Cooper,
622: L. Forr\'o and B. Keszei, Nature {\bf 343}, 444 (1990).
623: \bibitem{simon} I. Rajta, I. Borb\'{e}ly-Kiss, Gy. M\'{o}rik, L. Bartha, E.
624: Koltay and \'{A}.Z. Kiss, Nucl. Inst. and Methods B {\bf 109}, 148
625: (1996); T. Sz\"{o}r\'{e}nyi, Zs. Geretovszky, L. Kelemen, J. T\'{o}th and
626: A. Simon, Vacuum {\bf 59}, 327 (1998).
627: \bibitem{sas} B. Sas, F. Portier, K. Vad, B. Keszei, L.F. Kiss, N.
628: Hegmann, I. Puha, S. M\'{e}sz\'{a}ros and F.I.B. Williams, Phys. Rev. B
629: {\bf 61}, 9118 (2000); F. Portier, G. Kriza, B. Sas, L.F. Kiss, I. Pethes, K.
630: Vad, B. Keszei, N. and F.I.B. Williams, Phys. Rev. B {\bf 66}, 140511(R)
631: (2002).
632: \bibitem{tinkham} M.Tinkham, Introduction to Superconductivity
633: \linebreak[2]
634: (McGraw-Hill, New York, 1996).
635: \bibitem{kosterlitz} J.M. Kosterlitz and D.J. Thouless, J. Phys. C {\bf 6},
636: 1181 (1973).
637: \bibitem{halperin} B.I. Halperin and D.R. Nelson, J. Low Temp. Phys.
638: {\bf 36}, 599 (1979).
639: \bibitem{blatter} G. Blatter, M.V. Feigel'man, V.B. Geshkenbein, A.I.
640: Larkin and V.M. Vinokur, Reviews of Modern Physics {\bf 66}, 1125
641: (1994).
642: \bibitem{pierson2} S.W. Pierson, Phys. Rev. B {\bf 51}, 6663 (1995).
643: \bibitem{pierson3} S.W. Pierson, Phys. Rev. Lett. {\bf 75}, 4674 (1995).
644: \bibitem{pierson4} S.W. Pierson, Phys. Rev. B {\bf 54}, 688 (1995).
645: \bibitem{pethes} I. Pethes, B. Sas, G.Kriza, F. Portier, F.I.B. Williams,
646: K. Vad and S. M\'esz\'aros, Synthetic Metals {\bf 120}, 1013 (2001).
647: \bibitem{pierson5} S.W. Pierson, O.T. Valls, and H. Bahlouli, Phys. Rev.
648: B {\bf 45}, 13035 (1992).
649: \bibitem{nandori} I. N\'andori, U.D. Jentschura, K.Sailer and G. Soff,
650: Phys. Rev. D {\bf 69}, 025004 (2004) and J. Phys. G .{\bf 28}, 607
651: (2002); I. N\'andori, J. Polonyi and K. Sailer, Phil. Mag. B {\bf 81}, 1655 (
652: 2001) and Phys. Rev. D {\bf 63}, 045022 (2001).
653: \end{references}
654:
655: \end{document}
656:
657:
658:
659:
660:
661:
662:
663:
664:
665:
666:
667:
668:
669:
670:
671:
672:
673: