Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.
APPLIED PHYSICS LETTERS 86, 152505 ͑2005͒Nanoengineered Curie temperature in laterally patterned ferromagneticsemiconducto...
152505-2       Eid et al.                                                                                        Appl. Phy...
152505-3     Eid et al.                                                                                      Appl. Phys. L...
Upcoming SlideShare
Loading in …5

Nanoengineered Curie temperature in laterally patterned ferromagnetic semiconductor heterostructures


Published on

  • Be the first to comment

  • Be the first to like this

Nanoengineered Curie temperature in laterally patterned ferromagnetic semiconductor heterostructures

  1. 1. APPLIED PHYSICS LETTERS 86, 152505 ͑2005͒Nanoengineered Curie temperature in laterally patterned ferromagneticsemiconductor heterostructures K. F. Eid, B. L. Sheu, O. Maksimov, M. B. Stone,a͒ P. Schiffer, and N. Samarthb͒ Department of Physics and Materials Research Institute, Pennsylvania State University, University Park, Pennsylvania 16802 ͑Received 20 December 2004; accepted 1 March 2005; published online 5 April 2005͒ We demonstrate the manipulation of the Curie temperature of buried layers of the ferromagnetic semiconductor ͑Ga,Mn͒As using nanolithography to enhance the effect of annealing. Patterning the GaAs-capped ferromagnetic layers into nanowires exposes free surfaces at the sidewalls of the patterned ͑Ga,Mn͒As layers and thus allows the removal of Mn interstitials using annealing. This leads to an enhanced Curie temperature and reduced resistivity compared to unpatterned samples. For a fixed annealing time, the enhancement of the Curie temperature is larger for narrower nanowires. © 2005 American Institute of Physics. ͓DOI: 10.1063/1.1900938͔ Heterostructures derived from the ferromagnetic semi- dry etching. For the patterning process, the samples are firstconductor ͑Ga,Mn͒As are important for studying spin- spin coated with a bilayer of about 400-nm-thick electron-dependent device concepts1–5 relevant to semiconductor beam resist P͑MMA-MAA͒ copolymer/PMMA with molecu-spintronics.6–8 In contrast to metallic ferromagnets where lar weight of 950. The desired patterns are defined on thecrystalline defects rarely affect the magnetic ordering tem- sample using direct-write electron-beam lithography at elec-perature, it is now well established that Mn interstitial tron energy of 100 keV. After development of the resist indefects—double donors that compensate holes—play a key MIBK: IPA 1 : 1 solution, a metallic layer of either 45 nm Alrole in limiting the hole-mediated Curie temperature ͑TC͒ of or Al/ Au is deposited on the sample using thermal evapora-͑Ga,Mn͒As to below 110 K in as-grown samples.9 In thin tion. Using standard liftoff techniques, we then obtain metal͑Ga,Mn͒As epitaxial layers, post-growth annealing10,11 wires that serve as a hard mask for the subsequent chlorine-drives Mn interstitials to benign regions at the free surface based dry etching process. After dry etching, for sampleswhere they are passivated, resulting in significant increases with an Al mask, the metal layer is dissolved in CD-26 pho-in both the hole density ͑p͒ and TC ͑which can be as high as toresist developer, leaving just the semiconductor nanowire;160 K͒.12–14 Such values of TC should in principle allow the for samples with an Al/ Au mask, the metal is retained onfabrication of proof-of-concept spintronic devices operating portions of the sample. We note that in the latter case theabove liquid nitrogen temperatures. Unfortunately, the undoped GaAs capping layer serves as an insulator that pre- vents the shunting of current through the metal during theannealing-induced enhancement of TC is almost completely measurement.17 Figures 1͑a͒ and 1͑b͒ show plan view SEMsuppressed in heterostructure devices containing buried images of patterned nanowires. Three identical rows of wires͑Ga,Mn͒As layers.15,16 In this letter, we describe a nanoengi- are patterned on each wafer: one set is measured as-grownneered solution to this problem: lithographic patterning of͑Ga,Mn͒As heterostructures into nanowires. Such patterningopens up new pathways for defect diffusion to free surfacesat the sidewalls, and thus restores the annealing-enhanced TCto buried ͑Ga,Mn͒As layers. These results suggest routes to-wards the flexible fabrication of ͑Ga,Mn͒As-based spintronicdevices with relatively high TC. Laterally patterned wires are fabricated fromGaAs/ ͑Ga, Mn͒As/ GaAs heterostructures grown by molecu-lar beam epitaxy on epiready semi-insulating GaAs ͑100͒substrates. Crystal growth conditions are identical to thosedescribed in an earlier publication.15 The samples consist ofa 5 nm thick low temperature GaAs buffer layer, followed bya Ga1−xMnxAs layer ͑Mn content ϳ6%, thickness of 15 or 50nm͒ and finally a low temperature grown 10 nm GaAs cap-ping layer. The data shown are for 50 nm thick magneticlayers, but the results were qualitatively the same for the 15nm thick samples. We pattern these samples into wires thatare ϳ4 ␮m long and with two different widths ͑1 ␮m and 70nm͒ using electron beam lithography followed by liftoff anda͒ Currently at the Condensed Matter Sciences Division of the Oak Ridge National Laboratory. FIG. 1. SEM images of nanowires patterned along two different crystallo-b͒ Electronic mail: graphic directions: ͑a͒ ͓110͔ and ͑b͒ ͓010͔.0003-6951/2005/86͑15͒/152505/3/$22.50 86, 152505-1 © 2005 American Institute of PhysicsDownloaded 06 Apr 2005 to Redistribution subject to AIP license or copyright, see
  2. 2. 152505-2 Eid et al. Appl. Phys. Lett. 86, 152505 ͑2005͒FIG. 2. ͑a͒ Magnetization vs temperature for macroscopic pieces ͑area ofϳ5 mm2͒ of as-grown and annealed GaAs/ ͑Ga, Mn͒As͑50 nm͒ /͑10 nm͒GaAs heterostructure. The data are measured in a magnetic field of50 Oe in-plane upon warming after field cooling at 10 kOe. The sample isannealed at 190 °C in ultrahigh purity nitrogen gas ͑5N͒ for 5 h. The datashow that TC is not affected by annealing. ͑b͒ Resistance vs temperature͑measured in van der Pauw geometry͒ for a 1 mmϫ 2 mm mesa patternedfrom the same wafer as in ͑a͒. FIG. 3. Temperature dependence of the ͑four-probe͒ resistivity for as-grown and annealed single wires of ͑a͒ 1 ␮m width and ͑b͒ 70 nm width. Both sets of wires are patterned along the ͓110͔ direction. Note that the data for theand patterned, while the other two sets are annealed after annealed 70 nm wire in ͑b͒ are plotted using a different scale ͑axis on right hand side͒. Plot ͑c͒ shows the same measurements for four individual 70 nmpatterning at 190 °C for 5 h. wires patterned along different crystalline orientations. All wires are pat- We probe the ferromagnetic phase transition in single terned from the same wafer used in Fig. 2, and the annealing conditions arewires using four probe measurements of the temperature- identical to those used in Fig. 2.dependent resistance R͑T͒; it is well-established7,8 that R͑T͒shows a well-defined peak close to TC in ͑Ga,Mn͒As, espe-cially for samples with TC Ͻ 100 K. For higher TC, the peak 190 °C for 5 h. The data indicate that annealing results in ais less well-defined but still yields a reasonable estimate of slight increase in TC, accompanied by a slight increase in thethe ordering temperature ͑typically an overestimate of resistivity of the sample. These observations suggest that an-ϳ10 K͒.7,8 In addition, we measure the temperature depen- nealing induces modest diffusion of Mn interstitials to thedence of both the magnetization ͓using a commercial super- free surfaces at the sidewalls of the wire. The diffusion co-conducting quantum interference device ͑SQUID͒ magneto- efficient ͑estimated to be D ϳ 100 nm2 / h at ϳ190 ° C14͒ ismeter͔ and the resistivity in macroscopic pieces of the parent simply not large enough to allow significant removal of Mnwafers ͑using the van der Pauw method͒. interstitials from the bulk of the 1 ␮m wide wire within the Figure 2͑a͒ shows the magnetization as a function of 5 h annealing time. There is however some alteration of thetemperature for both as-grown and annealed ͑ϳ5 mm2͒ defect states, as indicated by the small increase in the highpieces of a GaAs/ 50 nm͑Ga, Mn͒As/ 10 nm GaAs hetero- temperature resistivity; we do not have an explanation forstructure. As found in earlier studies,15 the cap layer com- this observation but note that a similar effect has been seenpletely suppresses any annealing-induced enhancement of in previous studies of long anneals, albeit at higher annealingTC. This suppression of the annealing effect has been attrib- temperatures.11uted to the formation of a p-n junction at each of the two Figure 3͑b͒ shows the effect of annealing for a 70 nmGaAs/ ͑Ga, Mn͒As interfaces as some Mn interstitials wide nanowire ͑also patterned along ͓110͔͒. Here, annealing͑double donors͒ initially diffuse across the boundaries into produces a striking increase in TC of almost 50 K, accompa-GaAs.14,15 The resulting Coulomb barrier is expected to nied by a correspondingly significant decrease in the resis-strongly inhibit the further diffusion of Mn interstitials from tivity of the wire. Both these observations strongly suggestthe bulk of the ͑Ga,Mn͒As layer towards the interfaces. The the successful removal of the Mn interstitials from the bulktemperature dependence of the sample resistance is in com- of the ͑Ga,Mn͒As nanowire. Lateral diffusion of the Mn in-plete agreement with the magnetization measurements. Fig- terstitials provides the most likely explanation for these ob-ure 2͑b͒ shows van der Pauw measurements of R͑T͒ for a servations. We note that the time and length scales for diffu-macroscopic mesa ͑1 mmϫ 2 mm͒; the peak of R͑T͒ does sion observed in our experiments ͑ϳ35 nm in 5 h͒ arenot change upon annealing. In addition, there is no signifi- consistent with earlier low-temperature annealing studies ofcant reduction in the resistivity of the sample with annealing, ͑Ga,Mn͒As epilayers.14indicating that the hole density has not changed. It is important to examine whether defect-diffusion may Lateral patterning produces distinctly different behavior. depend on the crystalline direction and/or the nature of theFigure 3͑a͒ shows the temperature-dependent resistivity for a free surface where the interstitials eventually reside. We1 ␮m wide wire patterned along the ͓110͔ direction. The data hence pattern four nanowires ͑each with 70 nm width͒, alongare shown for both an as-grown wire and one annealed at ¯ the principal cubic directions ͓͑110͔, ͓110͔, ͓100͔, andDownloaded 06 Apr 2005 to Redistribution subject to AIP license or copyright, see
  3. 3. 152505-3 Eid et al. Appl. Phys. Lett. 86, 152505 ͑2005͓͒010͔͒. All four nanowires are annealed under identical con- 2 D. Chiba, M. Yamounichi, F. Matsukura, and H. Ohno, Science 301, 943ditions ͑190 °C for 5 h͒. The results are shown in Fig. 3͑c͒ ͑2003͒. 3 M. Tanaka and Y. Higo, Phys. Rev. Lett. 87, 026602 ͑2001͒.and indicate that, at least for the processing and annealing 4 S. H. Chun, S. J. Potashnik, K. C. Ku, P. Schiffer, and N. Samarth, Phys.protocol followed in this study, defect diffusion does not Rev. B 66, 100408͑R͒ ͑2002͒.vary significantly with crystalline direction. 5 D. Chiba, Y. Sato, T. Kita, F. Matsukura, and H. Ohno, Phys. Rev. Lett. In summary, we have shown that nanolithography allows 93, 216602 ͑2004͒. 6the engineering of defect diffusion pathways, hence provid- S. A. Wolf, D. D. Awschalom, R. A. Buhrman, J. M. Daughton, S. voning a means of tailoring impurity-controlled magnetism in Molnar, M. L. Roukes, A. Y. Chtchelkanova, and D. M. Treger, Science 294, 1488 ͑2001͒.ferromagnetic semiconductor heterostructures. Areas with 7 H. Ohno in Semiconductor Spintronics and Quantum Computation, editeddifferent TC and resistivity can be made on the same wafer by D. D. Awschalom, D. Loss, and N. Samarth ͑Springer, Berlin, 2002͒, p.simply by having different feature sizes. Smaller features 1. 8have a higher Curie temperature and lower resistivity. Al- N. Samarth, in Solid State Physics, edited by H. Ehrenreich and F.though we do not observe any obvious crystalline anisotropy Spaepen ͑Elsevier/Academic, San Diego 2004͒, Vol. 58. 9in the diffusion constant of the Mn interstitials, such direc- K. M. Yu, W. Walukiewicz, T. Wojtowicz, I. Kuryliszyn, X. Liu, Y. Sasaki, and J. Furdyna, Phys. Rev. B 65, 201303͑R͒ ͑2002͒.tional dependence may still exist and may be revealed by in 10 T. Hayashi, Y. Hashimoto, S. Katsumoto, and Y. Iye, Appl. Phys. Lett. 78,situ resistance monitoring while annealing the nanowires. 1691 ͑2001͒.Our findings auger well for prospects of designing 11 S. J. Potashnik, K. C. Ku, S. H. Chun, J. J. Berry, N. Samarth, and P.͑Ga,Mn͒As-based spintronic device heterostructures such as Schiffer, Appl. Phys. Lett. 79, 1495 ͑2001͒. 12magnetic tunnel junctions with TC well above 77 K. In addi- K. W. Edmonds, K. Y. Wang, R. P. Campion, A. C. Neumann, C. T.tion, systematic studies of annealing of nanowires of differ- Foxon, B. L. Gallagher, and P. C. Main, Appl. Phys. Lett. 81, 3010 ͑2002͒.ing width may provide new insights into ongoing controver- 13 K. C. Ku, S. J. Potashnik, R. F. Wang, S. H. Chun, P. Schiffer, N. Samarth,sies about the microscopic details of the diffusion and M. J. Seong, A. Mascarenhas, E. Johnston-Halperin, R. C. Myers, A. C.passivation of Mn interstitials in ͑Ga,Mn͒As.18 Gossard, and D. D. Awschalom, Appl. Phys. Lett. 82, 2302 ͑2003͒. 14 K. W. Edmonds, P. Boguslawski, K. Y. Wang, R. P. Campion, S. N. No- This research has been supported by Grant Nos. ONR vikov, N. R. Farley, B. L. Gallagher, C. T. Foxon, M. Sawicki, T. Dietl, M.N0014-05-1-0107, DARPA/ONR N00014-99-1093, -00-1- Buongiorno Nardelli, and J. Bernholc, Phys. Rev. Lett. 92, 0372010951, University of California-Santa Barbara Subcontract ͑2004͒. 15KK4131, and NSF DMR-0305238 and -0401486. This work M. B. Stone, K. C. Ku, S. J. Potashnik, B. L. Sheu, N. Samarth, and P. Schiffer, Appl. Phys. Lett. 83, 4568 ͑2003͒.was performed in part at the Penn State Nanofabrication Fa- 16 D. Chiba, K. Takamura, F. Matsukura, and H. Ohno, Appl. Phys. Lett. 82,cility, a member of the NSF National Nanofabrication Users 3020 ͑2003͒.Network. 17 This is confirmed by comparing the resistivity of the wires with van der Pauw resistivity measured on samples that had no metal.1 18 Y. Ohno, D. K. Young, B. Beschoten, F. Matsukura, H. Ohno, and D. D. M. Adell, J. Kanski, L. Ilver, V. Stanciu, and P. Svedlindh, cond-mat/ Awschalom, Nature ͑London͒ 402, 790 ͑1999͒. 0412006.Downloaded 06 Apr 2005 to Redistribution subject to AIP license or copyright, see