Strain effect of multilayer FeN structure on GaAs substrate Xiaowei Zhang, Nian Ji, Valeria Lauter, Hailemariam Ambaye, and Jian-Ping Wang Citation: J. Appl. Phys. 113, 17E149 (2013); doi: 10.1063/1.4800086 View online: http://dx.doi.org/10.1063/1.4800086 View Table of Contents: http://jap.aip.org/resource/1/JAPIAU/v113/i17 Published by the American Institute of Physics.
Additional information on J. Appl. Phys. Journal Homepage: http://jap.aip.org/ Journal Information: http://jap.aip.org/about/about_the_journal Top downloads: http://jap.aip.org/features/most_downloaded Information for Authors: http://jap.aip.org/authors
JOURNAL OF APPLIED PHYSICS 113, 17E149 (2013)
Strain effect of multilayer FeN structure on GaAs substrate Xiaowei Zhang,1,2,a) Nian Ji,1,2 Valeria Lauter,3 Hailemariam Ambaye,3 and Jian-Ping Wang1,2
1 The Center for Micromagnetics and Information Technologies (MINT) & Electrical and Computer Engineering Department, University of Minnesota, Minneapolis, Minnesota 55455, USA 2 School of Physics and Astronomy, University of Minnesota, Minneapolis, Minnesota 55455, USA 3 Neutron Science Scattering Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831, USA
(Presented 15 January 2013; received 6 November 2012; accepted 24 January 2013; published online 10 April 2013) Overly doped FeN multilayer structure on GaAs substrate was fabricated. After the post-annealing process, FeN martensite in each Fe/FeN layer formed partially chemically ordered Fe16N2, which was observed by X-ray diffraction. To detect the saturation magnetization (Ms) depth profile, polarized neutron reflectivity was conducted. Fe/FeN layer showed a significant improvement of Ms for each layer compared to Ms of Fe. More importantly, different FeN layers showed different Ms according to the physical distance to the substrate GaAs. The most enhanced Ms (exceeding the limit of Fe65Co35 Ms) observed at the bottom part of the film, consistent with previous reports, should be attributed to the lattice strain by GaAs substrate. In order to detect the lattice constant, C 2013 In-plane X-ray Diffraction was done and a large in-plane lattice constant was determined. V AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4800086]
I. INTRODUCTION
II. EXPERIMENT
Giant saturation magnetization (Ms) material, e.g., Fe16N2, is highly demanded by magnetic recording.1 Fe16N2 has been first claimed to possess giant Ms since 1970s.2 Since then, a great deal of research work3–6 has been put in this material. Despite the controversial results of different groups, Sugita et al.7 claimed the successful fabrication of pure phase Fe16N2 thin film and its Ms reached 2.9 T at room temperature, which was substantially higher than FeCo alloy. Conventionally, Fe65Co35 was known as the material with the highest Ms. Through years of research and study on Fe16N2, we have established a reliable and repeatable thin film fabrication process using facing target sputtering process and a direct measurement of giant Ms with spatial resolution in the film normal direction by polarized neutron reflectivity (PNR). High Ms8,9 and large perpendicular anisotropy10 of Fe16N2 were recently reported and discussed. However, it was also observed that giant Ms of Fe16N2 was dependent on the growth condition and different growth procedures would result in different magnetic behaviors. In this work, overly doped FeN samples were fabricated on GaAs single crystal substrate. Both single layered (Fe(5 nm)/FeN(15 nm)) and multi-layered ((Fe(5 nm)/ FeN(15 nm)) 3) structures showed high Ms that was more than the value of FeCo alloy after post-annealing process, which promoted the diffusion of N atoms in N rich FeN layer into Fe underlayer. Strain effect was proposed to explain the giant enhancement of Ms at the bottom of the film.
A. Sample fabrication
Two FeN thin films were fabricated using facing target sputtering process. For sample L1, Fe seed layer was deposited on GaAs single crystal substrate to induce (001) texture, which was carried out at 250 C for 30 s. FeN layer was deposited subsequently at room temperature. In these two samples, N atoms were overly doped in the FeN layer. This constituted one double layer Fe/FeN. Three of this Fe/FeN were repeated for sample L3. After in situ post-annealing treatment, partially ordered Fe16N2 was formed in the film. B. Structure characterization
Crystal structure of FeN films was characterized by a Siemens D5005 x-ray diffractometer with Cu Ka radiation. Sample thickness was measured by X-ray reflectivity (XRR) using Philips X’Pert Pro X-ray diffractometer. To better determine the Ms of our samples and reveal the spatial resolution, both samples were measured by polarized neutron reflectometry using the Magnetism Reflectometer on Beamline 4A at Spallation Neutron Source, Oak Ridge National Lab. During the experiment, a magnetic field of 1.15 T was applied in the plane of the film to saturate the magnetization. Neutron beam was reflected from the film and momentum transfer was perpendicular to the film plane. Spin up and spin down reflectivities were recorded instantaneously. Inplane X-ray diffraction (XRD) was also done by Philips X’Pert Pro X-ray diffractometer. III. RESULTS AND DISCUSSION
a)
Author to whom correspondence should be addressed. Electronic mail:
[email protected].
0021-8979/2013/113(17)/17E149/3/$30.00
XRD was measured on L1 and L3 to investigate this structural order. Fe16N2 is a metastable phase11 with body center tetragonal (BCT) structure, in which N atoms occupy
113, 17E149-1
C 2013 AIP Publishing LLC V
17E149-2
Zhang et al.
J. Appl. Phys. 113, 17E149 (2013)
FIG. 1. XRD patterns of Fe16N2 thin films L1 and L3. Both samples showed finger print peak (002) at 2h ¼ 28.5 .
the interstitials between Fe atoms in an orderly fashion. Since samples L1 and L3 were both overly doped with N atoms, it was not expected to see Fe16N2. However, in Fig. 1, finger print peak for Fe16N2 which is (002), still occurred at 2h ¼ 28.5 . The structure factor for this particular peak is 1.08fFe þ 2fN. In Fe8N martensite, which is the case before annealing treatment, this peak does not exist due to the disordered occupation of N atoms in c direction interstitials. Both L1 and L3 showed this peak, which indicated the formation of ordered Fe16N2 phase. The peak at 2h ¼ 58.9 corresponded to both Fe16N2 (004) and Fe8N (002).The appearance of Fe16N2 (002) demonstrated that N rich FeN layer with Fe underlayer could form Fe16N2 after post annealing process. To further analyze the structure, XRR was conducted. In Fig. 2(a), the decay of both curves followed 1/q4. Oscillatory fringes came from the interference of the reflections beams.
FIG. 2. (a) X-ray reflectivity of Fe16N2 samples L1 and L3, (b) electron SLD of L3 was homogeneous through the whole sample, and (c) SLD of L1 stayed the same after the first 2 nm bump.
Fitting experimental data will yield the thickness of the film and more importantly the electron scattering length density (SLD), which will reveal the film’s chemical structure (Figs. 2(b) and 2(c)) and information about the interface. For the Ms measurement of Fe16N2 thin films, VSM can show the integrated Ms value of the whole sample. However, considerable errors including the alignment of the sample to the electrode magnet and volume determinations have a tendency to smear the Ms value. PNR, on the other hand, takes advantage of the direct interaction of neutrons and magnetization of the magnetic material. Spin up and spin down neutrons will experience different potentials and, therefore, behave differently in reflectivity. For L1 and L3 (Figs. 3(a) and 3(c)),
FIG. 3. (a) Reflectivity curve of L1, (b) NSLD and Ms depth profile of L1, (c) reflectivity curve of L3, and (d) NSLD and Ms depth profile of L3. Both (b) and (d) curves showed high Ms region near the substrate with a thickness of 5 nm.
17E149-3
Zhang et al.
FIG. 4. In plane X-ray reflectivity of Fe/FeN on GaAs substrate. FeN was ˚ due to the peak FeN shown to possess a larger lattice constant of 5.75 A (220).
reflectivity curves showed different oscillatory periods due to film thickness differences. For neutron scattering length density (NSLD), both samples exhibited a nearly smooth depth profile as shown in brown lines. For both samples, N atoms were overly doped in FeN layer, therefore, Fe16N2 should not be expected to be formed in L1 and L3, which means low Ms should dominate the whole samples. These samples were different from the samples with right amount of N atoms in FeN layer (Fe/N ¼ 8/1), whose high Ms values were determined by PNR and were homogeneous through the whole sample. However, green lines of both samples in Figs. 3(b) and 3(d) indicated giant Ms at the bottom of the film, with a thickness of 5 nm, respectively. The only option for this enhancement should come from Fe16N2 since all the other FeN phases possess low Ms values. Considering the fact that Fe seed layer was deposited before FeN layer, N atoms diffusion should contribute to the formation of Fe16N2 at the bottom as N atoms were initially rich in top layer. With the right stoichiometric amount of diffused N atoms, Fe16N2 was formed at this specific location near the substrate. On top of this giant Ms layer, we would expect a N rich phase since N atoms were initially rich and a lower Ms region was detected by PNR. This transition of Ms from high to low also occurred on L3. Multilayer structure of Fe/FeN provided the same environment of single Fe/FeN layer but the difference was that higher Fe/FeN layers were not next to substrate; they did not exhibit giant Ms. From the perspective of NSLD, although lattice is expanded in c direction for Fe16N2, it also contains two more N atoms, which makes its SLD close to original bcc Fe. In both samples, N atom diffusion created Fe/FeN layer and their ˚ 2 in Figs. 3(b) and NSLD varied between 7 and 8 106 A 3(d). This showed the little NSLD difference between Fe and FeN layer. To better demonstrate the fitting of PNR, chemical structure extracted from NSLD was checked and the correspondent model also applied XRR data. It is noticeable that the bottom parts of samples L1 and L3 have different electron and NSLD from XRR and PNR result. This region is most likely Fe and GaAs compound due to the inter-diffusion of Fe-Ga or Fe-As atoms at their interface. The addition of this
J. Appl. Phys. 113, 17E149 (2013)
layer actually improved our fitting to the experiment data and was also confirmed by SLD depth profile. After post-annealing, low Ms FeN phase, through a N atom diffusion process into Fe underlayer, was partially converted into high Ms Fe16N2 phase. In addition, high Ms values occurring at the bottom of both samples L1 and L3 rather than the bottom of each Fe/FeN layer suggested the substrate’s importance for developing giant Ms phase. The preference of forming Fe16N2 phase at this location was associated with the minimum energy of this ordered FeN phase. Intuitively, lattice distortion due to substrate was checked first. In plane XRD was performed to reveal the in plane lattice constant that was influenced by substrate. In Fig. 4, sample L3 was placed in the configuration the same as inset figure. Momentum transfer was along [110] direction. The peak at 2h ¼ 45.2 belonged to GaAs (220) d spacing. The higher peak next to GaAs (220) corresponded to a ˚ . The lattice bigger in plane lattice spacing (220) d ¼ 2.033 A ˚, constant of Fe16N2 calculated from this value was 5.75 A ˚. which was slightly larger than the nominal value 5.72 A This lattice expansion could create a minimum energy state that favored the formation of Fe16N2 with giant Ms. The volume increase was responsible for the charge transfer between Fe and N atoms, which resulted in a magnetization difference from the smaller lattice constant case.12
IV. CONCLUSIONS
Overly doped FeN thin films were fabricated using facing target sputtering process. Giant Ms was promoted by the annealing treatment, which promoted N atoms diffusion with the formation of Fe16N2. PNR revealed the Ms depth profile and showed giant Ms at the bottom of the film. This could be attributed to the lattice strain from the GaAs substrate. ACKNOWLEDGMENTS
This work was partially supported by Seagate Technology and Western Digital. Parts of this work were carried out in the Characterization Facility through NSF MRSEC program at University of Minnesota. 1
J. P. Wang, N. Ji, X. Liu, Y. Xu, C. S anchez-Hanke, F. M. F. de Groot, Y. Wu, L. F. Allard, and E. Lara-Curzio, IEEE Trans. Magn. 48, 1710 (2012), and references therein. 2 T. K. Kim and M. Takahashi, Appl. Phys. Lett. 20, 492 (1972). 3 M. A. Russac, C. V. Jahnes, E. Klokholm, J. Lee, M. E. Re, and B. C. Webb, J. Appl. Phys. 70, 6427 (1991). 4 C. Gao and W. D. Doyle, J. Appl. Phys. 73, 6579 (1993). 5 K. Nakajima and S. Okamoto, Appl. Phys. Lett. 56, 92 (1990). 6 M. Takahashi, H. Shoji, H. Takahashi, T. Wakiyama, M. Kinoshita, and W. Ohta, IEEE Trans. Magn. 29, 3040 (1993). 7 Y. Sugita, H. Takahashi, M. Komuro, K. Mitsuoka, and A. Sakuma, J. Appl. Phys. 76, 6637 (1994). 8 N. Ji, L. Allard, E. Curzio, and J. P. Wang, Appl. Phys. Lett. 98, 092506 (2011). 9 N. Ji, Y. Wu, and J. P. Wang, J. Appl. Phys. 109, 07B767 (2011). 10 N. Ji, M. S. Osofsky, V. Lauter, L. F. Allard, X. Li, K. L. Jensen, H. Ambaye, E. Lara-Curzio, and J. P. Wang, Phys. Rev. B 84, 245310 (2011). 11 K. H. Jack, Proc. R. Soc. London, Ser. A 208, 216 (1951). 12 S. Okamoto, O. Kitakami, and Y. Shimada, J. Appl. Phys. 85, 4952 (1999).
