Journal of Magnetism and Magnetic Materials 191 (1999) 339-344 Induced anisotropy, magnetic domain structure and magnetoimpedance effect in CoFeB amorphous thin films D. Garci´a *, J.L. Munoz , G. Kurlyandskaya , M. Va´zquez , M. Ali , M.R.J. Gibbs Instituto de Magnetismo Aplicado. UCM-RENFE and Instituto de Ciencia de Materiales, CSIC. P.O. Box 155, 28230 Las Rozas, Madrid, Spain Department of Physics, University of Sheffield, Sheffield, S3 7RH, UK Received 5 May 1998; received in revised form 24 June 1998 Abstract CoFeB amorphous thin films have been prepared by radio frequency magnetron sputtering. The objective of the work has been to obtain very soft thin films, and to correlate their domain structure with the induced anisotropy and giant magneto-impedance (GMI) features. Tailored anisotropies were induced by either growth under stress, or by post- deposition annealing in the presence of a magnetic field. In certain cases these induced anisotropies give rise to significant GMI effect at a frequency, 5 MHz, relatively low for thin films. 1999 Elsevier Science B.V. All rights reserved. PACS: 75.30.Gw; 72.15.Gd; 75.60.Ch; 75.70.!i Keywords: Magnetic anisotropy; Giant magnetoimpedance effect; Magnetic domains; RF sputtering; Thin films 1. Introduction [3], but is weak. In this work we have aimed to produce controlled MA in CoFeB amorphous thin Magnetic anisotropy (MA) in thin films is well films, where the magnetostriction is low, but in- known to be important for both the control of duced anisotropy may be high. This has involved magnetic properties and technological applications controlled stress on the substrate during growth, or [1]. Magnetic anisotropy can be formed in thin post-deposition heat treatment in the presence of films during preparation [2], and also using post- a magnetic field. The GMI effect is a promising deposition heat treatments. In-plane MA has been magneto-transport property for technological ap- found in FeBSi amorphous RF-sputtered thin films plications. This effect has been explained in terms of the skin depth changes suffered by the magnetic material when the permeability induced by a polarising AC electric current is varied through a * Corresponding author: Tel.: ##34 1 630 1724; fax: ##34 1 630 1625; e-mail: imass24@fresno.csic.es. bias magnetic field [4-6]. The magneto-impedance Permanent address: Institute of Metal Physics UD RAS response of the magnetic material is highly depen- Kovalevskaya str. 18, GSP 70, 620219 Ekaterinburg, Russia. dent on its MA since magnetic anisotropy controls 0304-8853/99/$ - see front matter 1999 Elsevier Science B.V. All rights reserved. PII: S 0 3 0 4 - 8 8 5 3 ( 9 8 ) 0 0 2 6 2 - 5 340 D. Garci&a et al. / Journal of Magnetism and Magnetic Materials 191 (1999) 339-344 the films permeability. We have aimed to induce the pression produced an easy magnetisation direction most favourable MA for the GMI response of the parallel to the long dimension owing to the nega- thin films. tive magnetostriction. The samples of this type There are a number of publications connected chosen for measurements were transversely cut to with MA in Co-based amorphous ribbons. The 2 mm;20 mm, and finally exhibit a transverse underlying physics is understood in terms of inelas- magnetic anisotropy. tic response or pair ordering [7-9]. Currently there Amorphicity was checked before and after treat- are no reported systematic investigations of the ments using X-ray diffraction on a D-5000 Siemens induced MA for amorphous CoFeB thin films. Diffractometer using Cu K Our CoFeB samples have been sputtered on radiation. Two hysteresis loops taken with a vibrating pre-stressed substrates to induce magnetoelastic sample magnetometer (VSM) at room temperature anisotropy via the non-zero magnetostriction, or were measured for each sample: one along the long sputtered onto non-stress substrate and afterwards in-plane sample direction, which is described as annealed in a DC or AC magnetic field post-pro- a longitudinal loop and another one along the duction. The aim of this investigation has been to short in-plane sample direction, described as a correlate their domain structure with features of the transverse loop. magnetisation process and of the GMI ratio behav- The domain structure of the samples for different iour, which can easily illustrate the anisotropy applied magnetic fields was observed using a mag- peculiarities. netooptical Kerr effect (MOKE) imaging system. The magneto-impedance has been measured us- ing a set-up with a complete computer control, as 2. Experimental techniques described in Ref. [10]. The impedance changes of the samples were obtained by measuring the AC The films were deposited using a planar magne- voltage drop across the film at a constant AC tron radio frequency sputtering system. The base current amplitude. The contacts were carefully pre- pressure was in the low 10\ mbar range and argon pared using silver electric paint. The current was pressure during deposition was 5;10\ mbar. The flowing along the longitudinal direction of the deposition rate was 2.3 A>/s under these growing films. The GMI ratio was defined as conditions. The composition of the target was Co Z Z(H)!Z(H FeB which, as an amorphous ribbon, has (H)"100 ), (1) a magnetostriction constant !1;10\. The Z Z(H ) samples were grown on 40 mm;20 mm glass sub- strates, and then cut transversely to 2 mm;20 mm. where H  is the maximum bias longitudinal mag- The thickness of the films ranged from 1.5 to netic field, H "10.5 kA/m. This DC field was 4.3 m. Uniaxial transverse magnetic anisotropy produced by a Helmholtz coils system. The AC along the short dimension was induced using two current flowing through the film was kept at a con- different methods. The first method consisted of stant amplitude of I "5 mA. The frequency of thermal treatments in the presence of either a DC the current was 5 MHz. or AC (50 Hz) 1.5 kA/m magnetic field oriented along the short in-plane sample direction. The treatments were carried out either at 300°C or 3. Results and discussion 250°C in an Ar atmosphere for 1 h. The second method consisted in growing the films onto Fig. 1 shows the domain structure in the re- a 40 mm;20 mm glass substrate bent along an manent state for an as-grown film, a compressed axis parallel to the long dimension. The radius of film, and films thermally treated in the presence of curvature was 400 mm. A compressive stress was a magnetic field. All of the images show 180° do- thus developed in the film upon removal of the main wall patterns, typical of magnetic materials substrate from the sputtering chamber. This com- with uniaxial anisotropy, where the easy axis is D. Garci&a et al. / Journal of Magnetism and Magnetic Materials 191 (1999) 339-344 341 Fig. 1. Domain patterns of remanent magnetisation of different samples: (a) as-grown, (b) compressed, (c) 300°C AC field annealed and (d) 300°C DC field annealed samples. determined by the particular treatment each been too low for sufficient thermal activation of the sample has received. The as-grown sample (Fig. 1a) diffusion. exhibits a longitudinal easy magnetisation direc- Fig. 2 shows the longitudinal hysteresis loops of tion, which may be determined by magnetron effect the as-grown, 300°C DC and 300°C AC field an- during preparation. The direction of the field while nealed and compressed samples. The low values of annealing and applied stress account for the trans- the coercive fields, from 14 up to 30 A/m, and of the verse orientation of the magnetisation in the ther- remanent magnetisation, below 5% of saturation mally treated and compressed samples. magnetisation, are typical for soft magnetic mate- The compressed (Fig. 1b) and the DC thermal rials with no structural and/or shape-related treated (Fig. 1d) samples show stripe domains alig- features which could give rise to high magnetic ned with the transverse axis of the films together anisotropies. As-grown and 300°C AC field an- with small spike domains near the edges. The nealed samples displayed coercive fields of 14 and 300°C AC field annealed sample (Fig. 1c) presents 16 A/m, respectively. These values are smaller than a different domain structure consisting of wider the coercive fields of the 300°C DC field annealed curved stripes. The domain walls adopt a reversed and compressed samples, which are 30 and 20 A/m, `S' shape. The running direction of the walls forms respectively. Remanent magnetisation values were an angle of around 17° with the transverse direction in the central zone of the film and progressively increases up to an angle of about 65° near the edges. The 250°C AC field annealed sample has a similar domain structure but with an angle of 52° in the central region increasing up to 70° near the edges. It has to be inferred that the AC anneal induces a weaker anisotropy, and that shape effects compete in determining the net anisotropy direc- tion. The mean effective field under AC excitation is certainly less than under DC excitation. But at the same time AC annealing could lead to destabilisa- tion of the domain structure, which is very impor- tant for AC field demagnetisation. The sample was given a 250°C AC field annealing and it retained a large longitudinal component to its easy Fig. 2. Longitudinal hysteresis loops of samples. Note that magnetization axis. The temperature may have transverse induced anisotropy can be deduced. 342 D. Garci&a et al. / Journal of Magnetism and Magnetic Materials 191 (1999) 339-344 similar for all the samples and they correlate well Table 1 Transverse hysteresis loops data with the remanent domain patterns shown in Fig. 1, where a large number of antiparallel Sample Remanence Coercive Magnetisation domains are present, with approximately equal (in units of M) field (A/m) work areas of the two magnetisation orientations. (kJ/m) The possible magnetic anisotropy sources in As-grown 0.010 19 1.65 these films are reduced to directional ordering of AC field 0.006 12 1.39 atoms and stress induced effects which can be annealed modified externally by appropriate treatments and DC field 0.010 22 1.53 the contribution of the geometry of the samples annealed through the demagnetising fields created at their Compressed 0.015 17 0.85 surfaces. The influence of the last source of an- isotropy in the magnetisation process is considered to be similar in all the films since their demagnetis- ing factors are the same. So, in the following we domain walls produces such low figures for re- focus the discussion on the anisotropies induced by manent magnetisation and coercive field. The high- the various treatments. est coercive field corresponds to the DC field As long as the longitudinal loops shown in Fig. 2 annealed film, where the domain structure stabilis- correspond to magnetisation processes in the direc- ation hardens the film due to non-uniform domain tion near the hard magnetisation direction of the wall displacements. treated samples, a straight comparison of the effec- The magnetisation work (energy required to tive anisotropies induced by each method can be saturate the sample) considered in Table 1 is cal- made by taking into account the different values of culated taking a saturation magnetic polarisation the longitudinal fields needed to reach the satura- of 1.18 T [12]. The as-grown sample needs more tion magnetisation (H ) deduced from the loops. energy than the rest of the films due to its longitudi- The as-grown sample, whose easy magnetisation nal component of easy magnetisation axis. The DC axis roughly lies in the longitudinal direction, has field annealed sample requires more energy to be a H of 370 A/m and for 300°C AC field annealed saturated than the AC field annealed sample, al- sample it is 820 A/m. The AC annealing has turned though transverse anisotropy generated by DC the easy axis towards the transverse sample direc- field annealing is higher than by AC field annealing. tion, although a longitudinal component of this The reason for that is attributed again to the stron- axis still remains, as shown in Fig. 1c. ger pinning of the domain walls at the edges of the A further increase of H field is obtained for DC field annealed sample. The lowest magnetisa- the 300°C DC field annealed sample, with tion work corresponds to the strained sample, H "1100 A/m. This field is larger than the one for which confirms that the stress method is the most the AC annealed sample. This is explained in terms effective to induce the transverse anisotropy in of the lower net field seen under AC field annealing these films. which, for a given temperature, induce a lower The scientific society undertakes special attempts anisotropy energy. The maximum H field is ob- to find thin films with low frequency (in the interval tained for the strained sample, with a value of 5-15 MHz) magnetoimpedance response. The anal- 1770 A/m. This is a typical value in comparison ysis of the GMI features together with anisotropy with the magnitudes of field and stress induced parameters and domain structure seems promising anisotropy parameters seen in ribbon samples [11]. to predict the states with this response. Fig. 3 Table 1 summarises the main parameters of the shows the variations of the GMI ratios with the transverse hysteresis loops of the samples. The AC bias longitudinal magnetic field for the 300°C AC field annealed sample presents the lowest remanent field annealed (Fig. 3a) and compressed (Fig. 3b) magnetisation and the lowest coercive field. The samples. No appreciable magnetoimpedance effects absence of effective pinning centres due to wide were found on the other samples. Although the D. Garci&a et al. / Journal of Magnetism and Magnetic Materials 191 (1999) 339-344 343 4. Conclusions Magnetic anisotropy induced in soft amorphous CoFeB thin films by different techniques is compar- atively analysed. Low temperature heat treatments in the presence of an external magnetic field pro- duce a complete change of the domain structure of the films and leads to changes of the position of easy magnetisation axis. The domain wall statics and dynamics can be directly correlated with the strength of induced anisotropy. Permanent stresses applied to the magnetic films through their substrates profit by the negative nature of mag- netostriction of CoFeB films to easily induce magnetoelastic anisotropy. Under certain treat- ments to produce transversal anisotropy together with destabilised domain structures the CoFeB thin films show a magnetoimpedance response at a low 5 MHz frequency of AC polarising current. Acknowledgements Fig. 3. GMI curves of the 300°C AC field annealed (a), com- pressed (b) and 300°C DC field annealed (c) samples. The insets G. Kurlyandskaya acknowledges the Spanish show the magnetoimpedance response at low bias fields for (a) and (b) samples. Solid symbols refer to low bias fields GMI Ministerio de Educacio´n y Ciencia for supporting response and open symbols refer to high bias field's GMI re- the project SAB 95-0421. M. Ali acknowledges the sponse. provision of a studentship by the UK Engineering and Physical Sciences Research Council. The authors acknowledge Dr. A.V. Korolyov for helpful overall changes of the Z/Z ratios were below 1% discussion. in both samples, the peaks associated with maxima in the permeability induced by the AC polarising current are clearly distinguished. These maxima in References the GMI ratios are obtained for bias fields which correspond to the longitudinal H [1] A. Hernando, M. Va´zquez, in: H.H. Liebermann (Ed.), fields of the two films. The presence of appreciable GMI effect sug- Rapidly Solidified Alloys: Processes, Structures, Properties and Applications, ch. 17. gests that the treatments carried out on these two [2] A. Miktus, J. Wenda, K. Kulakowski, J. Magn. Magn. samples favour the rotation of their magnetisations. Mater. 160 (1996) 341. In the same time 300°C DC annealed sample [3] J. Wenda, L.J. Maksymowicz, J. Magn. Magn. Mater. 87 (Fig. 3c), which shows a higher transversal anisot- (1990) 286. ropy than the AC field annealed film, presents no [4] R.S. Beach, A.E. Berkowitz, J. Appl. Phys. 76 (1994) 6209. appreciable magnetoimpedance probably due to [5] L.V. Panina, K. Mohri, Appl. Phys. Lett. 65 (1994) the processes of domain structure stabilisation. 1189. Also, no hysteretic GMI behaviour was observed [6] R.L. Sommer, C.L. Chien, Appl. Phys. Lett. 67 (22) (1995) for the films with nonzero Z/Z ratios, but authors 3346. consider that no conclusion can be done on this [7] O.V. Nielsen, A. Hernando, V. Madurga, J.M. Gonzalez, J. Magn. Magn. Mater. 46 (3) (1985) 341. subject, because the absolute values of the mag- [8] K. Zaveta, O.V. Nielsen, K. Jurek, J. Magn. Magn. Mater. netoimpedance ratios were rather small. 117 (1-2) (1992) 61. 344 D. Garci&a et al. / Journal of Magnetism and Magnetic Materials 191 (1999) 339-344 [9] G.V. Kurlyandskaya, N.V. Dmitrieva, A.P. Potapov, [11] M.R.J. Gibbs, J. Magn. Magn. Mater. 83 (1990) 329. V.A. Lukshina, L.M. Voronova, I.V. Gervase´va, N.G. [12] E.P. Wohlfrarth, Ferromagnetic Materials, vol. 1, North- Bebenin, Phys. Met. Metall. 83 (5) (1997) 487. Holland, Amsterdam, 1980, p. 489. [10] J.P. Sinnecker, P. Tiberto, G.V. Kurlyandskaya, E.H.C.P. Sinnecker, M. Va´zquez, A. Hernando, J. Appl. Phys., sub- mitted.