P H Y S I C A L R E V I E W L E T T E R S week ending VOLUME 90, NUMBER 19 16 MAY 2003 Nanoengineered Magnetic-Field-Induced Superconductivity Martin Lange,* Margriet J. Van Bael, Yvan Bruynseraede, and Victor V. Moshchalkov Laboratorium voor Vaste-Stoffysica en Magnetisme, Katholieke Universiteit Leuven, Celestijnenlaan 200D, 3001 Leuven, Belgium (Received 4 September 2002; published 15 May 2003) The perpendicular critical fields of a superconducting film have been strongly enhanced by using a nanoengineered lattice of magnetic dots (dipoles) on top of the film. Magnetic-field-induced super- conductivity is observed in these hybrid superconductor/ferromagnet systems due to the compensation of the applied field between the dots by the stray field of the dipole array. By switching between different magnetic states of the nanoengineered field compensator, the critical parameters of the superconductor can be effectively controlled. DOI: 10.1103/PhysRevLett.90.197006 PACS numbers: 74.25.Dw, 74.78.Db, 75.75.+a When the applied magnetic field exceeds a certain an amorphous Si=SiO2 substrate, which is held at liquid critical value, superconductivity is suppressed due to nitrogen temperature during deposition. This thin Pb film orbital and spin pair breaking effects. This very general behaves as a type-II superconductor. For protection property of superconductors sets strong limits for their against oxidation, the Pb is covered by a 10 nm Ge layer practical applications, since, in addition to applied mag- that is insulating at low temperatures and thus prevents netic fields, the current sent through a superconductor the influence of the proximity effect between Pb and also generates magnetic fields, which can lead to a loss Co=Pd. The Ge=Pb=Ge trilayer is patterned into a of zero resistance. Materials that are not only able to transport bridge (width 200 m, distance between volt- withstand magnetic fields, but in which superconductivity age contacts 630 m) using optical lithography and can even be induced by applying a magnetic field, are very chemical wet etching. The ferromagnetic dots are made rare, and up to now only EuSn Mo6S8 [1,2], organic by defining a resist mask on the transport bridge by - BETS 2FeCl4 materials [3,4], and HoMo6S8 [5] show electron-beam lithography and subsequent evaporation this unusual behavior. The appearance of magnetic-field- of a Pd 3:5 nm = Co 0:4 nm =Pd 1:4 nm 10 multilayer induced superconductivity (FIS) in the former two com- pounds was interpreted in terms of the Jaccarino-Peter effect [6], in which the exchange fields from the para- (a) H = 0 z magnetic ions compensate an applied magnetic field, so m B that the destructive action of the field is neutralized. Co/Pd Here we report that FIS can also be realized in hybrid Ge superconductor/ferromagnet nanostructured bilayers. The Pb basic idea is quite straightforward (see Fig. 1): a lattice of magnetic dots with magnetic moments aligned along the Ge positive z direction is placed on top of a superconducting Si/SiO2 film. The magnetic stray field of each dot has a positive z component of the magnetic induction B (b) z under the dots and a negative one in the area between the dots. Added to H a homogeneous magnetic field H-see Fig. 1(b)-these dipole fields enhance the z component of the effective magnetic field 0Heff 0H Bz in the small area just under the dots and, at the expense of that, reduce Heff everywhere else in the Pb film, thus providing the con- dition necessary for the FIS observation. This new field compensation effect is not restricted to specific super- conductors, so that FIS could be achieved in any super- conducting film with a lattice of magnetic dots. To implement the idea of the nanoengineered FIS, we FIG. 1 (color). Schematic drawing of the investigated hybrid have prepared a sample, which reminds us of other sys- superconductor/ferromagnet sample. (a) The magnetic stray field B of the dots is comparable with the field of a magnetic tems used during the last decade for studying flux pinning dipole. (b) A magnetic field H applied in the z direction can be by periodic arrays of magnetic dots [7­11] and by mag- compensated by the dipole stray field between the dots, result- netic domains [12]. The sample consists of a 85 nm super- ing in the conditions necessary for the observation of magnetic conducting Pb film evaporated on a 1 nm Ge base layer on field-induced superconductivity. 197006-1 0031-9007=03=90(19)=197006(4)$20.00 2003 The American Physical Society 197006-1 P H Y S I C A L R E V I E W L E T T E R S week ending VOLUME 90, NUMBER 19 16 MAY 2003 into the resist mask. The resist mask is finally removed in a lift-off procedure. The dots are arranged in a regular square array with period 1:5 m. They have a square shape (side length about 0:8 m) with slightly irregular edges. The dots on the superconducting Pb film consist of Co=Pd multilayers having an easy axis of magnetization perpendicular to the sample surface [13]. The hysteresis loop of the dots is measured with H perpendicular to the surface by the magneto-optical Kerr effect, revealing a high magnetic remanence of Mr 0:8Ms, where Mr and Ms are the remanent and saturation magnetization, re- spectively, and a large coercive field 0Hcoe 150 mT. This makes it possible to produce quite stable remanent magnetic domain states in the dots by using different magnetization procedures. These domain states were in- vestigated by magnetic force microscopy (MFM) in a Digital Instruments nanoscope III. After demagnetiza- tion, the signal from each of the dots consists of dark and bright spots, as shown in Fig. 2(a), indicating the pres- ence of several magnetic domains in the dots; compare Ref. [14]. The net magnetic moment m of each dot in this state is approximately zero (m jmj 0). The demag- netization is carried out by oscillating H (perpendicular to the sample surface) around zero with decreasing am- plitude. Saturating the dots in a large positive perpen- dicular field aligns all m along the positive z direction (mz > 0), so that the dots appear brighter compared with the signal between the dots; see Fig. 2(b). In contrast to that, when the dots have been saturated in a large negative field, resulting in mz < 0, they give a darker contrast in the MFM image, as shown in Fig. 2(c). Simultaneous recording of magnetic and topographic images shows that the spots visible on dots in 2(b) and 2(c) are of topographic origin, and are not due to a magnetic signal. The magnetic field (H)-temperature (T)-phase dia- grams of the Pb film were constructed for the three magnetic states of the dots from T measurements carried out in a Quantum Design physical properties measurement system applying a four-probe ac technique with an ac current of 10 A at a frequency of 19 Hz. H is applied perpendicular to the sample surface. We defined FIG. 2. MFM images of the hybrid superconductor/ferromag- the critical temperature as Tc T 50% n , with net structure in H 0 and at room temperature. The images the resistivity and n 1:4 cm the normal state re- show a 10 10 m2 region of the sample in the remanent state sistivity at 7.3 K. We did not observe any indication that after (a) demagnetization, (b) magnetization in H 1 T, and the small magnetic fields jHj Hcoe applied during (c) magnetization in H ˙1 T. these measurements altered the domain state of the dots, although minor microscopic changes cannot be excluded. applied flux per unit cell of the dot array is exactly one The H-T-phase boundary separating the normal (N) superconducting flux quantum 0 2:07 mT m2. In from the superconducting (S) state is clearly altered by contrast to that, the H-T-phase boundary is strongly changing the magnetic state of the dot array. A conven- asymmetric with respect to H when the dots are magne- tional symmetric (with respect to H) phase boundary is tized in positive or negative directions; see Figs. 3(b) and obtained when m 0; see Fig. 3(a). Two kinks in the 3(c). Moreover, the maximum Tc is shifted to 2H1 when curve can be seen at H H1, with H1 the first matching mz > 0 and to ˙2H1 when mz < 0. This shift gives rise to field 0H1 0= 1:5 m 2 0:92 mT, at which the FIS when mz > 0 and mz < 0, as is demonstrated in 197006-2 197006-2 P H Y S I C A L R E V I E W L E T T E R S week ending VOLUME 90, NUMBER 19 16 MAY 2003 4 m = 0 4 m N z > 0 4 mz < 0 N 2 2 S 2 H 1 N / 0 S 0 0 H N -2 -2 -2 S -4 N (a) -4 (b) -4 (c) N 7.18 7.18 7.20 7.22 7.24 7.18 7.20 7.22 7.24 7.18 7.20 7.22 7.24 T (K) T (K) T (K) 1.0 1.0 1.0 0.8 0.8 0.8 n 0.6 0.6 0.6 / N S N N S N N S N 0.4 0.4 0.4 0.2 m = 0 0.2 mz > 0 0.2 mz < 0 T = 7.22 K (d) T = 7.20 K (e) T = 7.20 K (f) 0 0 0 -4 -2 0 2 4 -4 -2 0 2 4 -4 -2 0 2 4 µ0H (mT) µ0H (mT) µ0H (mT) FIG. 3 (color). Field-induced superconductivity (FIS) in a Pb film with an array of magnetic dots. Blue and yellow areas correspond to the superconducting (S) and the normal (N) states, respectively. The H-T-phase diagrams are obtained after (a) demagnetization (m 0), (b) saturation of the dots in a large positive H (mz > 0), and (c) saturation in a large negative H (mz < 0). H is shown for the magnetic states (d) m 0, (e) mz > 0, and (f) mz < 0, measured at lines of constant T as indicated by red arrows in the corresponding phase diagrams. Figs. 3(e) and 3(f). For instance, for mz > 0 and T coinciding with the first matching field H1. These kinks 7:20 K, the sample is in the normal state in zero field, are due to fluxoid quantization effects [16], confirming but when a positive field between 0:6 and 3:3 mT is that superconductivity indeed nucleates in multiply con- applied, the Pb film becomes superconducting, as shown nected regions of the film, like in superconducting wire in Fig. 3(e). Similarly, when the magnetic state of the dots networks [17] or thin films with periodic arrays of anti- is switched to mz < 0, superconductivity is induced by dots [18,19]. The maximum Tc at exactly H 2H1 can applying a negative field between ˙3:3 and ˙0:6 mT; see therefore be understood in terms of fluxoid quantization: Fig. 3(f). Contrary to that, the H curve shows the the flux created by the stray field between the dots can be typical NSN transition of a conventional superconductor estimated from the magnetostatical calculations to be for m 0. about ˙2:1 0 per unit cell of the dot array. This makes In the present system, the FIS can be explained by H 2H1 a favorable field for fulfilling the fluxoid taking into account the local magnetic induction of the quantization constraint. Similar arguments can also be dots B; see the discussion of Fig. 1. To support further applied for the dots in the mz < 0 state to explain the shift these arguments, we give in Fig. 4(a) the distribution of of the maximum Tc to H ˙2H1. For m 0, B is Bz, the z component of B for mz > 0 in the x-y plane, strongly reduced due to the domain structure in the calculated by using a magnetostatic model (see, e.g., dots. This means that the stray field only weakly influen- Ref. [15]). In zero field [Fig. 4(a)] the magnetic dipoles ces the Pb film, leading to a phase boundary without generate stray fields exceeding the upper critical field of peculiarities except the weak kinks at H H1. To the Pb film when T > 7:185 K , and, as a result, the Pb understand all features in the phase boundaries in more film is in the normal state. In an applied field of H detail, one should solve the linearized Ginzburg-Landau H2, the compensation of Bz takes place in the interdot (GL) equations, taking explicitly into account the addi- area where the Pb film is now in the superconducting state tional vector potential created by the magnetic dots. This [see the blue color in Fig. 4(b)], thus providing the per- work has not been done yet, but the GL analysis of the flux colation through dominantly superconducting areas, and distribution around magnetic dots has already been re- making possible the continuous flow of Cooper pairs and ported in Ref. [20]. zero film resistance. Moreover, the simple picture of field compensation is An important feature to note here is the appearance of not applicable anymore deeper in the superconducting periodic kinks in the H-T-phase boundary with a period state. At zero applied field, the dipole stray field of a 197006-3 197006-3 P H Y S I C A L R E V I E W L E T T E R S week ending VOLUME 90, NUMBER 19 16 MAY 2003 (a) H=0 which superconductivity is controlled by switching be- tween the two polarities of the magnetized dot array. 4 -6 mT In conclusion, we have shown that a nanoengineered lattice of magnetic dipoles can be used to selectively 3 -3 mT enhance the critical fields of superconducting films. )m -1 mT Magnetic-field-induced superconductivity is observed (µ 2 y 1 mT due to the compensation of the applied field by the stray field of the dipoles. 9 mT 1 The authors are thankful to E. Claessens for help with 15 mT the measurements and to S. Raedts, M. Morelle, and K. Temst for their contribution to the sample preparation. 00 1 2 3 4 This work was supported by the Belgian IUAP and the x (µm) Flemish GOA programs, by the ESF ``VORTEX'' pro- gram, and by the Fund for Scientific Research (F.W.O.)- (b) H=2H1 Flanders. M. J.V. B. acknowledges support from the F.W.O. 4 -6 mT 3 -3 mT ) -1 mT m *Electronic address: martin.lange@fys.kuleuven.ac.be (µ 2 y 1 mT [1] S. A. Wolf et al., Phys. Rev. B 25, 1990 (1982). 9 mT [2] H.W. Meul et al., Phys. Rev. Lett. 53, 497 (1984). 1 [3] S. Uji et al., Nature (London) 410, 908 (2001). 15 mT [4] L. Balicas et al., Phys. Rev. Lett. 87, 067002 (2001). [5] M. Giroud et al., J. Low Temp. Phys. 69, 419 (1987). 00 1 2 3 4 [6] V. Jaccarino and M. Peter, Phys. Rev. Lett. 9, 290 (1962). x (µm) [7] O. Geoffroy et al., J. Magn. Magn. Mater. 121, 223 (1993). FIG. 4 (color). Contour plots of the z component of the [8] J. I. Marti´n et al., Phys. Rev. Lett. 79, 1929 (1997). effective magnetic field 0Heff 0H Bz in the supercon- [9] D. J. Morgan and J. B. Ketterson, Phys. Rev. Lett. 80, ductor, calculated using a magnetostatic model, (a) for H 0, 3614 (1998). (b) for H 2H1. [10] M. J. Van Bael et al., Phys. Rev. B 59, 14 674 (1999). [11] M. J. Van Bael et al., Physica (Amsterdam) 332C, 12 magnetic dot generates two vortex-antivortex pairs, with (2000). the vortices located on the dot sites, and the antivortices [12] M. Lange et al., Appl. Phys. Lett. 81, 322 (2002). between the dots. When an external field is applied, [13] P. F. Carcia, A. D. Meinhaldt, and A. Suna, Appl. Phys. vortices enter the sample and interact with the vortex- Lett. 47, 178 (1985). antivortex pairs associated with the dots. This process [14] M. Hehn et al., Science 272, 1782 (1996). gives rise to interesting new effects dealing with vortex- [15] J. D. Jackson, Classical Electrodynamics (Wiley, New York, 1999), 3rd ed. antivortex patterns (see, e.g., Ref. [21]). These novel ef- [16] W. A. Little and R. D. Parks, Phys. Rev. Lett. 9, 9 (1962). fects cannot be described in more detail here due to a lack [17] B. Pannetier et al., in Quantum Coherence in Mesoscopic of space, but will be reported elsewhere. Systems, edited by B. Kramer (Plenum Press, New York, The field region in which FIS is observed can be tuned 1991), p. 457. by changing the period of the dot array or the magnitude [18] A. Bezryadin and B. Pannetier, J. Low Temp. Phys. 98, of the stray field. For instance, an increase of the fields 251 (1995). emanating from the dots could be achieved by using [19] V.V. Moshchalkov et al., in Handbook of Nanostructured magnetic dipoles with larger magnetic moments, shifting Materials and Nanotechnology, edited by H. S. Nalwa the maximum of T (Academic Press, San Diego, 2000), Vol. 3, Chap. 9, c to an even higher applied field. Good candidates for that are arrays of nanodots [22] and nano- p. 451. pillars [23]. For instance, dot arrays with a period of [20] M.V. Milosevic, S.V. Yampolskii, and F. M. Peeters, Phys. Rev. B 66, 024515 (2002). 70 nm have been fabricated [24], corresponding to H1 [21] S. Erdin, cond-mat/0211117. 0:4 T, which is already a remarkably high field. Besides [22] C. A. Ross et al., J. Appl. Phys. 91, 6848 (2002). improving the critical fields, the dipole array field com- [23] S.Y. Chou et al., J. Appl. Phys. 76, 6673 (1994). pensator can also be used to design logical devices in [24] K. Koike et al., Appl. Phys. Lett. 78, 784 (2001). 197006-4 197006-4