Epitaxial structure and magnetic anisotropies of metastable single crystal Co0.70Mn0.30 film H. W. Zhao National Laboratory of Solid State Microstructures, Center of Materials Analysis, Nanjing University, Nanjing 210 093, People's Republic of China Y. Chen, W. R. Zhu, G. S. Dong, and X. F. Jin National Key Laboratory of Applied Surface Physics, Department of Physics, Fudan University, Shanghai 200 433, People's Republic of China M. Lu and H. R. Zhaia) National Laboratory of Solid State Microstructures, Center of Materials Analysis, Nanjing University, Nanjing 210 093, People's Republic of China Received 19 June 1996; accepted for publication 10 November 1996 Alloy films of Co0.70Mn0.30 were grown on GaAs 001 substrates by molecular beam epitaxy at room temperature. The metastable tetragonally distorted single crystal structure was confirmed by reflection high energy electron diffraction and x-ray diffraction measurement, which exhibited a 2.87 Å in-plane lattice parameter. Ferromagnetic resonance measurements and theoretical fitting were performed and showed that, with a Mn capping layer, a uni-directional anisotropy existed. In addition, a fourfold tetragonal magnetocrystalline anisotropy as well as a uniaxial term in the film plane was also confirmed. The hysteresis loops recorded by longitudinal magneto-optical Kerr-effect also demonstrated the existence of different kinds of in-plane magnetic anisotropy. The origins of the anisotropy are explained tentatively. © 1997 American Institute of Physics. S0021-8979 97 03204-0 Molecular beam epitaxy MBE has been shown to be was analyzed once by XPS and confirmed the approximate suitable for growing single crystal films of transition metal coincidence with the nominal composition but the data of on semiconductor substrates.1,2 Various metastable phases the resonance fields of two samples in Fig. 2 implies that have been grown and stabilized at room temperature on composition deviations may exist . The substrate tempera- lattice-matched substrates. For example, metastable body- ture was maintained at T 300 K during deposition. centered-cubic bcc Co and body-centered-tetragonal bct Upon deposition, the RHEED pattern of GaAs 001 Mn films were grown on the GaAs 001 surface3 and on the faded. Continuing the film growth up to about 5 Å yielded a Pd 001 surface,4 respectively. However, the epitaxy of alloy new pattern shown in Fig. 1 a with the incidence electron films has seldom been reported. As a continuation of the beam along 110 direction the crystal azimuth in this article previous work of epitaxial growth of face-centered-cubic refers to the GaAs crystalline axis unless special description fcc Mn films on GaAs 001 surfaces,2 we prepared single is made . This RHEED pattern did not change during growth crystal Co0.70Mn0.30 alloy films on GaAs 001 substrates at up to 150 Å. A Cu layer was utilized to prevent oxidation. room temperature and studied their structure and magnetic The rectangular distribution of the streaks corresponds to the characteristics. To our knowledge, this study is the first one projection of the three-dimensional reciprocal lattice along to involve epitaxial growth on this particular alloy. the 110 direction. When the sample was rotated 90° around The sample fabrication was carried out in a MBE system the normal direction, so that the electron beam was along the connected with a VG-ESCALAB electron spectrometer.2 Re- 11¯0 direction, the resulting RHEED pattern was exactly flection high energy electron diffraction RHEED is the same as that obtained from the 110 direction, which equipped to monitor the growth in situ. After sputtering with indicates that the in-plane lattice net is square. Fig. 1 b dis- Ar ions at 800 eV for 55 min, the GaAs 001 substrate was plays the RHEED pattern for the sample rotating 45° with then annealed at 500 °C in ultrahigh vacuum for 25 minutes the e-beam along 010 direction. By comparing with the until a streaklike RHEED pattern was observed, implying a RHEED data of the clean GaAs 001 , the lattice constant a ``good'' flat surface. The pressure during evaporation was of the surface net was estimated to be 2.87 Å. It is difficult to below 1 10 7 Pa. The absence of contamination of oxygen determine the out-of-plane lattice spacing from the observed and carbon was indicated by Auger electron spectroscopy RHEED pattern alone because RHEED is generally not sen- AES and x-ray photoelectron spectroscopy XPS analyses. sitive to the vertical lattice spacing; therefore, we performed Finally, the Co0.70Mn0.30 alloy film was deposited by direct x-ray diffraction on the same sample ex situ. It is found co-evaporation of Co and Mn on the GaAs 001 surface at that in the obtained XRD spectrum not shown here the appropriate deposition rates measured by a quartz-crystal the huge peak around 66° and a peak at 58.9° are diffractions thickness monitor. The composition of the Co­Mn alloy film of the GaAs 004 , respectively. At about 74.2°, correspond- ing to the lattice spacing of 2.56 Å, a peak appears which is a Electronic mail: hrzhai@nju.edu.cn the only diffraction peak from the Co0.70Mn0.30 thin film, 2036 J. Appl. Phys. 81 (4), 15 February 1997 0021-8979/97/81(4)/2036/3/$10.00 © 1997 American Institute of Physics Downloaded¬14¬Mar¬2001¬to¬148.6.169.65.¬Redistribution¬subject¬to¬AIP¬copyright,¬see¬http://ojps.aip.org/japo/japcpyrts.html FIG. 2. In-plane angular dependence of the FMR field Hres for Co­Mn alloy films with different overlayers. The open circles are the experimental data and the solid line represents a theoretical fit using the parameters listed in Table I for a Cu 20 Å /Co­Mn 65 Å and b Mn 30 Å /Co­Mn 50 Å . FIG. 1. RHEED patterns for a 100 Å Co­Mn film epitaxially grown on the GaAs 001 surface with a e-beam 110 and b 010 , respectively. KD. , , and H stand, respectively, for the angles of the magnetization vector in spherical coordinates and the azi- however, the corresponding value does not accord with a fcc muthal angle of the magnetic field in the film plane. The or a bcc structure. Then it is realized that the Co-Mn films three edges of tetragonal crystal structure are taken as the were neither rotated fcc nor unrotated bcc structures,2 but reference axes of the coordinates. Note that the easy axes of rather a bct or an equivalent face-centered-tetragonal fct the fourfold, uniaxial, and uni-directional anisotropy are all structure. Considering its fcc structure for bulk phase,5 in this along the 100 article we assume the Co­Mn film to be a rotated fct struc- Co­Mn direction which is also the 110 axis of GaAs. The general condition for ferromagnetic resonance6 ture which shows 001 plane that rotating 45° with respect yields the following resonance equation: to the plane of the substrate, so 110 GaAs 100 Co­Mn , 110 / 2 H cos GaAs 010 Co­Mn . H 4 M 2H1 4 H2 The magnetic properties of Co­Mn alloy films were H studied at room temperature by means of ferromagnetic reso- 2 cos 4 2Hu cos 2 HD cos nance FMR and longitudinal magnetic optical Kerr effect H cos H 2Hu cos 2 LMOKE ex situ. The FMR experiments were performed at the X band of 9.78 GHz with the external magnetic field 16H2 cos 4 HD cos ], 2 rotated in the film plane, starting from the 100 Co­Mn direc- in which we define Hi Ki /M. By assuming the gyromag- tion. The angular dependences of the ferromagnetic reso- netic ratio to be the value of bulk Co, the best fitting nance field Hres , for Co­Mn films with different overlayers shown in Figs. 2 a and 2 b was achieved by taking the of Mn and Cu are displayed in Fig. 2. In the case of a 20 parameters listed in Table I. It is found that Ku is comparable Å Cu capping layer, it can be speculated from Fig. 2 a that to K1 and KD if not zero , which could also be deduced a fourfold anisotropy exists which corresponds to the tetrag- from the relative values of Hres in Fig. 2. It is noted that onal magnetocrystalline anisotropy in the 001 plane, be- KD( 5 104 erg/cm3) for Co­Mn film with Mn overlayer sides a contribution of a uniaxial term with the easy axis is smaller than the corresponding value of Jt HexM along 100 Co­Mn direction. For the film with 30 Å Mn over- 16 104erg/cm3 in Ref. 7 where stronger exchange cou- layer, it is seen from Fig. 2 b that Hres 0° is not equal to pling existed. Hres 180° . This means that an additional uni-directional an- The easy axis of the uniaxial anisotropy Ku term was isotropy field HD is present, causing 100 Co Mn to be the found along 110 GaAs in all the samples we made, which easiest direction. Following the above idea, a theoretical fit- gives us an insight into its origin. Considering the fact that ting of FMR field was made by using the following expres- there is a slight miscut of about 2° from 001 to 110 plane, sion of total free energy density: which leads to steps along 110 GaAs direction, the uniaxial E HM sin cos term may result from step-induced anisotropy as proposed in H K1 sin2 K2 sin4 Ref. 8. In a Co/Mn/Co sandwich system it was found that a K2 sin4 cos 4 Ku sin2 cos2 stronger induced uniaxial anisotropy also resulted from the 2 M2 cos2 K substrate steps.9 As for the uni-directional anisotropy KD , it D sin cos 1 could be considered to originate from an exchange interac- which includes the contributions from the Zeeman energy, tion between the ferromagnetic Co­Mn alloy and the Mn the tetragonal magnetocrystalline anisotropy terms overlayer. In the early 1960's, Kouvel discovered the exist- (K1 ,K2 ,K2 ), the in-plane uniaxial anisotropy (Ku term , the ence of exchange anisotropy in Co­Mn disordered alloys of demagnetizing energy, and a uni-directional anisotropy term about 25, 30, and 35 at. % Mn, and explained it with a sta- J. Appl. Phys., Vol. 81, No. 4, 15 February 1997 Zhao et al. 2037 Downloaded¬14¬Mar¬2001¬to¬148.6.169.65.¬Redistribution¬subject¬to¬AIP¬copyright,¬see¬http://ojps.aip.org/japo/japcpyrts.html TABLE I. Anisotropy constants deduced from the FMR data of Co­Mn film with different coverages ( 104erg/cm3). Sample K1 K2 K2 Ku KD Mn 30 Å /Co­Mn 50 Å 8.0 4.0 1.1 7.5 5.0 Cu 20 Å /Co­Mn 65 Å 3.4 2.3 0.4 2.45 0 tistical composition fluctuation model.10 However, in this ex- periment the absence of KD when the Mn overlayer was replaced by Cu indicates that it could not be due to the in- homogeneity of the Co­Mn film. Recently, it was discovered by Henry and Ounadjela11 that in Co/Mn multilayers the ex- change interaction existed at the Co/Mn interfaces, which supports our experimental results. The experimental LMOKE setup we used to record the FIG. 3. Hysteresis loops for samples of Mn 30 Å /Co­Mn 100 Å measured magnetic hysteresis loops of the Mn/Co­Mn system is simi- by LMOKE with the applied field in-plane along different from 110 lar to that described by Qui et al.12 By rotating the sample direction: a 0° b 90°, c 135°, d 180°, respectively. around the normal direction of the film plane, the hysteresis loops with the magnetic field along different directions in the film plane were recorded. prepared on GaAs 001 surface via MBE. FMR studies show Figures 3 a ­3 d display typical hysteresis loops of the that with the Mn capping layer, a uni-directional anisotropy sample Mn 30 Å /Co­Mn 100 Å with varying angle be- appears besides the fourfold magnetocrystalline anisotropy tween the 100 Co Mn direction and the incidence plane of and a uniaxial term in the film plane. It is proposed that this light. Fig. 3 a , 3 b , 3 c , and 3 d are the Kerr loops with results from the exchange interaction between the Co­Mn 0°, 90°, 135°, and 180°, which correspond to the four alloy and the Mn overlayer. The anisotropy of the hysteresis directions marked by A, B, C, and D in Fig. 2 b . It is found loops recorded by the LMOKE measurement can be under- that the rectangular loop in Figs. 3 a and 3 d is character- stood qualitatively as a result of the interplay of three istic of the two easy magnetization directions of anisotropies. 100 Co­Mn and 110 Co­Mn , in which the former is the easi- This work is supported by NNSFC, National Key Labo- est due to the positive uni-directional anisotropy field, caus- ratory of Applied Surface Physics of Fudan University and ing the shift of the loop to negative side while 3 d is shifted National Laboratory of Magnetism, Academy of Sciences of to positive side. A complicated hysteresis loop composed of China. two steps in Fig. 3 b appears in the 010 Co­Mn direction Note added in proof. Recently, we found that the along which the total anisotropy energy is at a minimum Co­Mn film is conclusively bct, which changes the values of with higher energy at position B in Fig. 2 b . To change the anisotropy constants and the easy directions with no influ- direction of magnetization by 180°, corresponding to the one ence on the whole physics. branch in the hysteresis loop, requires the magnetization to go through two potential maxima, e.g., from B to D across C and from D to F across E, which may be the origin of the 1 G. A. Prinz, Science 250, 1092 1990 . stepped loop. The tilted loop of Fig. 3 c shows the feature of 2 X. Jin, M. Zhang, G. S. Dong, M. Xu, Y. Chen, Xun Wang, X. G. Zhu, the hard 110 and X. L. Shen, Appl. Phys. Lett. 70, 3078 1994 . Co­Mn direction. It is obvious that under the 3 G. A. Prinz, Phys. Rev. Lett. 54, 1051 1985 . maximum field of about 150 G the sample cannot be satu- 4 D. Tian, A. M. Begley, and F. Jona, Surf. Sci. 273, L393 1992 . rated, which causes the unique unsaturated loop. This is in 5 K. Adachi, K. Sato, M. Matsui, and S. Mitani, J. Phys. Soc. Jpn. 35, 426 contrast to the other three loops, which show a nearly satu- 1973 . 6 rated state under the same maximum field. However, the in- H. Suhl, Phys. Rev. 97, 555 1955 . 7 D. Mauri, E. Kay, D. Scholl, and J. K. Howard, J. Appl. Phys. 62, 2929 consistency of the experimental value of coercive force of 1987 . about 40 G and the calculated value is explained by assum- 8 W. Weber, C. H. Back, A. Blschof, D. Pescia, and R. Allenspach, Nature ing a coherent rotation process,13 which means that the mag- London 374, 788 1995 . 9 netization process in the loop is not likely through uniform H. W. Zhao, Y. Chen, G. S. Dong, X. F. Jin, and H. R. Zhai unpub- lished . rotation. It should be a domain nucleation and domain wall 10 J. S. Kouvel, Phys. Chem. Solids 16, 107 1960 . displacement process. Therefore a quantitative explanation 11 Y. Henry and K. Ounadjela, Phys. Rev. Lett. 76, 1944 1996 . for the different loop shapes needs more study. 12 Z. Q. Qiu, J. Pearson, A. Berger, and S. D. Bader, Phys. Rev. Lett. 68, In conclusion, the metastable single crystal 1389 1992 . 13 B. Dieny, J. G. Gavigan, and J. P. Rebouillat, J. Phys. Condens. Matter 2, Co0.70Mn0.30 alloy film with tetragonal symmetry has been 159 1990 . 2038 J. Appl. Phys., Vol. 81, No. 4, 15 February 1997 Zhao et al. Downloaded¬14¬Mar¬2001¬to¬148.6.169.65.¬Redistribution¬subject¬to¬AIP¬copyright,¬see¬http://ojps.aip.org/japo/japcpyrts.html