1: \documentclass[aps,groupedaddress,epsfig,showpacs,floats]{revtex4}
2: \usepackage{graphicx}
3:
4:
5: \begin{document}
6:
7: \title{Influence of the stray field of magnetic dot on the nucleation of superconductivity in a disk}
8: \author{ D. S. Golubovi\'{c}\footnote{Dusan.Golubovic@fys.kuleuven.ac.be}, W. V. Pogosov, M. Morelle and V. V. Moshchalkov}
9: \affiliation{Nanoscale Superconductivity and Magnetism Group,
10: Laboratory for Solid State Physics and Magnetism , K. U. Leuven,
11: Celestijnenlaan 200 D, B-3001 Leuven, Belgium}
12:
13:
14: \begin{abstract}
15: We have investigated the nucleation of superconductivity in an Al
16: mesoscopic disk, with a magnetic dot on the top. The dot is
17: magnetized perpendicularly, and its magnetic field is
18: inhomogeneous. The Al disk and magnetic dot are separated by an
19: insulating layer, which ensures that there is only magnetic
20: interaction between them. This hybrid superconductor/ferromagnet
21: structure exhibits the maximum critical temperature for a finite
22: value of the perpendicular applied magnetic field, which is
23: parallel to the magnetization of magnetic dot. We have found a
24: good agreement between the experimental data for the
25: superconductor/normal metal phase boundary and the theoretical
26: predictions based on the Ginzburg-Landau theory.
27: \end{abstract}
28:
29: \pacs{74.78.Na,74.25.Dw}
30: \maketitle
31:
32:
33:
34: \section{Introduction}
35: {\it Hybrid} superconductor/ferromagnet structures have attracted
36: a lot of attention
37: \cite{victor,ja,pokro,rusi,misko,miskoa,mjvb,martin,mjvbb}. So
38: far, the experimental efforts have mainly been focused on
39: superconducting thin films with arrays of magnetic dots. Recently,
40: a superconducting disk with a perpendicularly magnetized dot was
41: fabricated and its superconducting $T_{c}(B)$ phase boundary
42: determined \cite{ja}. However, the nucleation of superconductivity
43: in this case was strongly affected by the proximity effect between
44: the disk and dot.
45:
46: In this paper we have investigated the onset of superconductivity
47: in an Al disk with a perpendicularly magnetized Co/Pd magnetic dot
48: on the top. The dot is separated from the disk by an insulating
49: spacer layer, which ensures that there is no suppression of the
50: order parameter in the disk due to the proximity effect and that
51: the interaction between the disk and dot has only {\it magnetic}
52: character \cite{ja}.
53:
54: The superconducting $T_{c}(B)$ phase boundary, obtained by
55: transport measurements, exhibits an asymmetry with respect to the
56: polarity of the applied magnetic field. The maximum critical
57: temperature, higher than the zero-field critical temperature, is
58: attained for a finite applied magnetic field which is parallel to
59: the magnetization of magnetic dot.
60:
61: \section{Sample preparation and experimental technique }
62: The sample was prepared on a SiO$_{2}$ substrate by electron beam
63: lithography on PMMA950K and the co-polymer electron beam resists
64: in three phases. Each phase involves the patterning of a desired
65: structure, thermal evaporation of the material and ultrasonic
66: assisted lift-off procedure.
67:
68: The contact pads and leads, as well as the alignment markers are
69: made up of $5\,$nm Cr and $30\,nm$ Au. The superconducting disk is
70: a $60\,$nm thick Al, whereas the magnetic dot consists of
71: $2.5\,nm$ Pd buffer layer and $10$ bilayers of $0.4\,$nm Co and
72: $1\,$nm Pd. A $10\,$nm thick Si spacer layer was evaporated before
73: the magnetic dot. A careful alignment procedure was needed to
74: position the dot at the centre of the disk.
75:
76: \begin{figure}[htb]
77: \centering
78: \includegraphics*[width=8cm]{disk.eps}
79: \caption{An AFM topography image of the sample. \label{afm}}
80: \end{figure}
81:
82: Fig. \ref{afm} presents an atomic force microscopy image of the
83: sample. The radius of the Al disk is $1.08\,{\rm \mu m}$, whereas
84: the dot has the radius of $270\,$nm. The electric contacts are
85: wedge-shaped, with the opening angle of $15\,^{\circ}$, as these
86: have proved to be the least invasive for transport measurements of
87: superconducting nanostructures \cite{proka}.
88:
89: The magnetic properties of patterned Co/Pd structures were
90: thoroughly investigated in Ref. \cite{martin}. We have determined
91: the magnetic properties of the Co/Pd multilayer at room
92: temperature from the magneto-optical Kerr measurements of the
93: co-evaporated plane film. These data have confirmed that the Co/Pd
94: multilayer has a perpendicular anisotropy, with a complete
95: remanence and coercive field of approximately $70\,$mT. Prior to
96: the measurements the sample was magnetized perpendicularly in the
97: magnetic field of $300\,$mT. As the applied magnetic fields in the
98: experiment never exceeded $30\,$mT, we have assumed that the
99: magnetization of the dot remains unaltered during the
100: measurements.
101:
102: The onset of superconductivity in the structure was studied by
103: measuring the superconducting $T_{c}(B)$ phase boundary. The phase
104: boundary was found resistively, from four-point transport
105: measurements, in a cryogenic setup at temperatures down to
106: $1.11\,$K, with the temperature and field resolution of $0.5\,$mK
107: and $5\,{\rm \mu T}$, respectively. The transport current with the
108: effective value of $100\,$nA and frequency $27.7\,$Hz was used,
109: whilst the signal-to-noise ratio was being improved by a lock-in
110: amplifier.
111:
112: The resistance of the structure at room temperature is $4.85\,{\rm
113: \Omega}$, the low temperature resistance is $R_{n}=2.8\,{\rm
114: \Omega}$, whereas the maximum critical temperature is
115: $T_{cm}=1.421\,$K.
116:
117:
118: \section{Experimental results}
119:
120:
121: \begin{figure}[htb]
122: \centering
123: \includegraphics*[width=8.5cm]{rt.EPS}
124: \caption{$R(T)$ transition in different applied magnetic fields.
125: Filled symbols indicate the transitions in the applied fields
126: parallel to the magnetization of the dot, whereas open symbols
127: present the transition in the antiparallel magnetic fields.
128: \label{rt}}
129: \end{figure}
130:
131: Fig. \ref{rt} presents resistive transitions of the structure in a
132: constant applied magnetic field. We have taken parallel magnetic
133: fields as positive and antiparallel magnetic fields as negative,
134: and will be using this convention throughout the paper.
135:
136: Typically, superconducting structures have resistive transitions
137: that are symmetric with respect to the polarity of an applied
138: magnetic field. Likewise, the maximum transition temperature is
139: achieved in zero applied field. The transitions in Fig. \ref{rt}
140: are strongly asymmetric with respect to the polarity of an applied
141: field. The critical temperature, defined conventionally as the
142: temperature at which the resistance is $R_{n}/2$, in the applied
143: magnetic field of $+2\,$mT is equal to the transition temperature
144: in $-1\,$mT. This dependence of the resistive transitions on the
145: field polarity is reproduced for higher fields, as well. More
146: importantly, the critical temperature in zero applied field is not
147: the maximum critical temperature of the structure. The structure
148: attains the maximum critical temperature when exposed to the
149: magnetic field of $+0.3\,mT$. The difference between the maximum
150: and zero-field critical temperature
151: is approximately $2.5\,$mK.
152:
153: The transitions for the applied fields $-1 < B_{a}<1\,{\rm [mT]}$
154: have considerable overshoots in resistance, with respect to the
155: normal state. This phenomenon is related to the formation of
156: normal/superconducting (NS) interfaces and nonequilibrium charge
157: imbalance effects \cite{vital,vit,proka}. The critical temperature
158: of the mesoscopic contacts is slightly higher than the critical
159: temperature of the disk \cite{proka}. For this reason,
160: superconductivity nucleates nonuniformly, thus giving rise to the
161: formation of the NS interfaces. Due to a finite local magnetic
162: induction in the disk generated by the dot, the difference in the
163: critical temperatures of the disk and the mesoscopic contacts is
164: greater than for a disk without a magnetic dot. As a positive
165: applied magnetic field increases, the total magnetic induction in
166: the disk around the magnetic dot decreases, thereby effectively
167: reducing the difference in the critical temperatures of the disk
168: and mesoscopic contacts. For this reason, as the positive applied
169: field increases the amplitude of the resistance anomaly decreases.
170: On the other hand, a negative applied field gives rise to a
171: greater difference in the local critical temperatures, which is
172: responsible for more pronounced peaks. When increasing the
173: negative applied field, the contribution of the stray field to the
174: total field becomes less significant and the typical amplitudes of
175: the overshoots are recovered \cite{victor,proka}.
176:
177: \begin{figure}[htb]
178: \centering
179: \includegraphics*[width=8.5cm]{tcb.EPS}
180: \caption{ The experimentally obtained superconducting phase
181: boundary plotted as $1-T_{c}(B)/T_{cm}$ versus the applied field
182: $B_{a}$. $T_{cm}$ is the maximum critical temperature. The inset
183: shows the phase boundary around $T_{c0}$. \label{fig4}}
184: \end{figure}
185:
186:
187: Fig. \ref{fig4} shows the experimental phase boundary in the plane
188: of the normalized critical temperature $1-T_{c}(B)/T_{cm}$ and the
189: applied magnetic field, with $T_{cm}$ being the maximum measured
190: critical temperature. The inset displays the phase boundary around
191: the zero-field critical temperature $T_{c0}$. The phase boundary,
192: except for the low fields, has a {\it linear} background, which is
193: the hallmark of the nucleation process in the disk \cite{victor}.
194: In addition, the quasiperiodicity of the phase boundary is
195: entirely consistent with what has been obtained for a mesoscopic
196: superconducting disk without the dot \cite{vb}. When there is a
197: pronounced proximity effect between the superconducting disk and
198: magnetic dot, the order parameter is suppressed below the dot and
199: the disk can be approximated with a superconducting loop with a
200: finite width \cite {ja}. In this case, the superconducting
201: $T_{c}(B)$ phase boundary has a crossover in the background - from
202: a parabolic for low fields to a linear for high fields \cite{vb}.
203: Therefore, the presence of a pronounced proximity effect between
204: the dot and the disk can confidently be ruled out, and specific
205: features of the phase boundary can be attributed to the magnetic
206: interaction between the disk and magnetic dot.
207:
208: The most striking feature of the phase boundary is its asymmetry.
209: The minimum in the phase boundary, which corresponds to the
210: maximum critical temperature of the structure, has been obtained
211: for a finite applied magnetic field which is parallel to the
212: magnetization of magnetic dot.
213:
214: \section{Theoretical results}
215:
216: The experimental data have been analysed by using the
217: Ginzburg-Landau (GL) theory. Near the transition from the normal
218: to the superconducting state the total magnetic field in the disk
219: is equal to the sum of the stray field of the magnetic dot and a
220: uniform externally applied field. The stray field we have used in
221: the analysis, was obtained by magnetostatic calculations. For
222: details, we refer to \cite{ja,misko}.
223:
224: As the sample in our experiment is thinner than the coherence
225: length $\xi (T)$, the order parameter is constant along $z$ axis
226: and the magnetic field in the GL equations can be averaged out
227: over the thickness of the disk. Thus, the problem is reduced to
228: the 2D case. The axial symmetry of the sample allows us to further
229: reduce the dimensionality of the problem to the 1D case, since
230: near the phase boundary the modulus of the order parameter is
231: axially symmetric. Using the cylindrical coordinate system with
232: coordinates $r$, $\varphi $, $z$, the dimensionless order
233: parameter $\psi(r,\phi)$ can be expressed as
234: \begin{equation}
235: \psi(r,\phi)=f(r)\, {\rm exp}(-iL\phi) \\\ , \label{ord}
236: \label{eq1}
237: \end{equation}
238: where $f(r)$ is the modulus of the order parameter and $L$ stands
239: for the winding number (vorticity). Instead of solving the GL
240: equations, which are nonlinear with respect to $f(r)$, we have
241: applied a variational procedure, similar to what has been used in
242: Ref. \cite{pogosov}. The trial function used for the modulus of
243: the order parameter is
244: \begin{eqnarray}
245: f(r) = p_{1}\cdot {\rm exp} \left(-q {r^{2}\over{R^{2}}} \right)
246: \cdot \left(\left({r\over{R}}\right)^{L}+p_{2}\left({r\over{R}}
247: \right)^{L+1}+p_{3}\left({r\over {R}}\right)^{L+2}+p_{4}\left(
248: {r\over{R}} \right)^{L+3} \right) . \label{eq2}
249: \end{eqnarray}
250: where $p_{1}$, $p_{2}$, $p_{3}$ and $p_{4}$ are the variational
251: parameters, $R$ is the radius of the disk, whereas $q$ is found
252: from the vacuum boundary condition for the order parameter
253: ($f'(R)=0$)
254: \begin{equation}
255: q={L+p_{2}(L+1)+p_{3}(L+2)+p_{4}(L+3)
256: \over{2(1+p_{2}+p_{3}+p_{4})}} .
257: \end{equation}
258:
259: Using Eqs. (\ref{eq1}) and (\ref{eq2}) the GL energy is found as a
260: function of the variational parameters (see Ref. \cite{pogosov}).
261: The values of the variational parameters are calculated by
262: minimizing the GL energy. Comparing the energies of states with
263: different $L$ the superconducting $T_{c}(B)$ phase boundary of the
264: disk is found.
265:
266: Fig. \ref{ter} displays the theoretical fit of the experimental
267: phase boundary in the low field regime, along with the theoretical
268: phase boundary of an identical superconducting disk without
269: magnetic dot (dashed line). The critical temperatures are
270: normalized to the zero-field critical temperature of the
271: superconducting disk without magnetic dot, as obtained by the
272: calculations. The best agreement between the theory and the
273: experiment has been found for $\xi (T=0)=90\,$nm, which is
274: consistent with the value of $\xi(T=0)$ obtained in Ref.
275: \cite{victor} for mesoscopic Al superconductors.
276: %The periodicity,
277: %as well as the asymmetry of the $T_{c}(B)$ curve have been
278: %reproduced well.
279: Each cusp in the phase boundary corresponds to
280: the transition between the states with different vorticities.
281: According to our results, there is one vortex in the disk in the
282: absence of the external magnetic field.
283:
284: \begin{figure}[htb]
285: \centering
286: \includegraphics*[width=8.5cm]{tcbt.EPS}
287: \caption{The experimental data and the corresponding theoretical
288: phase boundary of the superconducting disk with magnetic dot
289: (solid line), as well as the theoretical phase boundary of an
290: identical superconducting disk without magnetic dot (dashed line).
291: The temperatures are normalized to the zero-field critical
292: temperature of the disk without magnetic dot. The numbers indicate
293: the vorticity of the hybrid structure. \label{ter}}
294: \end{figure}
295:
296: The superconducting phase boundary strongly depends upon the
297: polarity of an external magnetic field. The magnetization of the
298: dot $m$ has been chosen to provide the best qualitative and
299: quantitative agreement between the theory and the experiment in
300: the vicinity of $T_{c0}$. The direction of the shift of $T_{c}(B)$
301: phase boundary near $T_{c0}$, for a fixed orientation of the
302: magnetization $m$, depends upon the intensity of the stray field
303: of magnetic dot. The shift can come about as a result of the
304: cancellation of the total flux generated by the magnetic dot, or
305: due to a change in the kinetic energy of the superconducting
306: condensate in the disk, accompanied by a switch in the vorticity
307: by one. The former shifts the phase boundary in the direction
308: opposite to the magnetization of the dot and the maximum critical
309: temperature is observed for a finite {\it negative} applied field,
310: whereas the latter provides that the maximum critical temperature
311: is achieved for a finite applied field parallel to the
312: magnetization of the dot, that is for a finite {\it positive}
313: field. Which of these competing effects prevails strongly depends
314: upon the intensity of magnetization of magnetic dot, as well as
315: upon the parameters of the superconducting structure. In our case
316: the shift is positive.
317:
318: \begin{figure}[htb]
319: \centering
320: \includegraphics*[width=5.8cm]{psi0.EPS}
321: \includegraphics*[width=5.8cm]{psi1.EPS}
322: \caption{ The moduli of the order parameter in (a) the Meissner
323: state and (b) vortex state with the vorticity $L=1$. The solid
324: line describes the superconducting disk with magnetic dot, whereas
325: the dashed line presents the disk without magnetic dot. In the
326: Meissner state the temperatures are $T/T_{c0}=0.987$ and
327: $T/T_{c0}=0.995$ for the disk with and without magnetic dot,
328: respectively. For the vortex state $L=1$ the respective
329: temperatures are $T/T_{c0}=0.990$ and $T/T_{c0}=0.995$. Here
330: $T_{c0}$ is the zero-field temperature of the plane
331: superconducting disk. \label{psi}}
332: \end{figure}
333:
334:
335: In addition to the shift along the $B$-axis, the phase boundary of
336: the hybrid structure is also shifted along the $T$-axis. The shift
337: along the $T$-axis is caused by a difference in the maximum
338: critical temperature of the disk with and without the dot. As the
339: stray field of the dot is spatially inhomogeneous, it cannot be
340: cancelled out by an externally applied homogeneous magnetic field.
341: For this reason, there is always a finite flux through a disk with
342: a magnetic dot and its maximum critical temperature is reduced
343: compared to the disk without a magnetic dot. Even though the
344: maximum critical temperature of the hybrid structure is less than
345: 1$\, \%$ lower than the maximum critical temperature of the plane
346: disk, this difference is sufficient to modify the phase boundary.
347: If the additional homogeneous magnetic field were applied, there
348: would be no differences in maximum critical temperature, and,
349: consequently, the $T_{c}(B)$ phase boundary would not be
350: substantially modified along the $T$-axis. We can, therefore,
351: conclude that the effect of the {\it inhomogeneous} magnetic field
352: on the nucleation of superconductivity is {\it twofold}: it shifts
353: the phase boundary along the $B$-axis, as well as distorts the
354: phase boundary along the $T$-axis, altering the values at which
355: the structure switches between different vorticities.
356:
357:
358: The stray field of the magnetic may change the typical spatial
359: profile of the superconducting condensate density within the disk.
360: Fig. \ref{psi}(a) and \ref{psi}(b) present the moduli of the order
361: parameter for the disk with magnetic dot (solid line) and disk
362: without magnetic dot (dashed line). Fig. \ref{psi}(a) shows the
363: Meissner state, whereas Fig. \ref{psi}(b) shows the state with the
364: vorticity $L=1$. In the Meissner state, for the plane
365: superconducting disk, the modulus of the order parameter has a
366: maximum at the centre of the disk. On the other hand, the
367: amplitude of the order parameter is reduced below the magnetic
368: dot, where the intensity of the stay field is the highest, and
369: exhibits a minimum at the centre of the disk. The spatial profile
370: of the order parameter is not strongly affected by the stray field
371: in the vortex state, because the vortex core, where the order
372: parameter is equal to zero, is in the region below the dot.
373:
374:
375:
376: In conclusion, we have fabricated a mesoscopic superconducting
377: disk made up of Al with a perpendicularly magnetized magnetic dot
378: on the top. The superconducting properties of the system have been
379: investigated by measuring the superconducting/normal state phase
380: boundary $T_{c}(B)$. It has been demonstrated that the phase
381: boundary is asymmetric with respect to the direction of the
382: applied field. The maximum critical temperature has been attained
383: for a finite value of the applied magnetic field, which is
384: oriented parallel to the magnetization of magnetic dot. The
385: experimental data are in a good agreement with the theoretical
386: results. It has also been shown that the inhomogeneity of the
387: stray field gives rise to a modification of the $T_{c}(B)$ phase
388: boundary along the $T$-axis, which would not be present if an
389: additional magnetic field were homogeneous.
390:
391: \section{Acknowledgements}
392: The authors would like to thank G. Rens for AFM measurements. This
393: work has been supported by the Belgian IUAP, the Flemish FWO and
394: the Research Fund K. U. Leuven GOA/2004/02 programmes, as well as
395: by the ESF programme "VORTEX". W. V. P. acknowledges the support
396: from the Research Council of the K.U. Leuven and DWTC.
397:
398: \begin{thebibliography}{20}
399: \bibitem{victor} MOSHCHALKOV V. V. {\it et al.}, in Handbook of
400: Nanostructured Materials and Nanotechnology 3, edited by H. S.
401: Nalwa (Academic Press, San Diego) 2000, pp. 451-525.
402: \bibitem{vb} BRUYNDONCX V. {\it et al.}, Phys. Rev. B, {\bf 60},
403: 10468 (1999).
404: \bibitem{ja}GOLUBOVI\'{C} D. S. {\it et al.}, Appl. Phys. Lett., {\bf 83}, 1593 (2003).
405: \bibitem{pokro} POKROVSKY V. L. {\it et al.}, cond-mat/0305153 (2003).
406: \bibitem{rusi} ALADYSHKIN A. Yu. {\it et al.}, cond-mat/0305551 (2003).
407: \bibitem{misko} MILO\v{S}EVI\'{c} M. V. {\it et al.}, Phys. Rev. B, {\bf 66}, 024515 (2002).
408: \bibitem{miskoa}MILO\v{S}EVI\'{c} M. V. {\it et al.}, Phys. Rev. B, {\bf 66}, 174519 (2002).
409: \bibitem{mjvb} VAN BAEL M. J. {\it et al.}, Phys. Rev. Lett., {\bf 86}, 1 (2001).
410: \bibitem{martin} LANGE M. {\it et al.}, Phys. Rev. Lett., {\bf 90}, 197006 (2003).
411: \bibitem{mjvbb} VAN BAEL M. J. {\it et al.}, Physica C, {\ 332}, 12
412: (2000).
413: \bibitem{vital} STRUNK C. {\it et al.}, Phys. Rev. B, {\bf 54}, R12701
414: (1996).
415: \bibitem{pogosov} POGOSOV W.V., Phys. Rev. B, {\bf 65}, 224511 (2002).
416: \bibitem{vit} BRUYNDONCX V. {\it et al.}, Europhys. Lett. {\bf 60}, 6
417: (1996).
418: \bibitem{proka} MORELLE M. {\it et al.}, Phys. Stat. Sol., {\bf 237}, 1 (2003).
419: \end{thebibliography}
420:
421: \end{document}
422: