Scholarly article on topic 'Large enhanced perpendicular magnetic anisotropy in CoFeB/MgO system with the typical Ta buffer replaced by an Hf layer'

Large enhanced perpendicular magnetic anisotropy in CoFeB/MgO system with the typical Ta buffer replaced by an Hf layer Academic research paper on "Materials engineering"

Share paper
Academic journal
AIP Advances
OECD Field of science

Academic research paper on topic "Large enhanced perpendicular magnetic anisotropy in CoFeB/MgO system with the typical Ta buffer replaced by an Hf layer"

Large enhanced perpendicular magnetic anisotropy in CoFeB/MgO system with the typical Ta buffer replaced by an Hf layer

T. Liu, J. W. Cai' , and Li Sun

Citation: AIP Advances 2, 032151 (2012); doi: 10.1063/1.4748337 View online: http://dx.doi.Org/10.1063/1.4748337 View Table of Contents: Published by the American Institute of Physics

Articles you may be interested in

Interfacial perpendicular magnetic anisotropy in CoFeB/MgO structure with various underlayers AIP Advances 115, 17C72417C724 (2014); 10.1063/1.4864047

Large enhanced perpendicular magnetic anisotropy in CoFeB/MgO system with the typical Ta buffer replaced by an Hf layer

T. Liu,1 J. W. Cai,1a and Li Sun2

1 Beijing National Laboratory for Condensed Matter Physics, Institute of Physics, Chinese Academy of Sciences, Beijing 100190, P. R. China

2Department ofMechanical Engineering and Texas Center for Superconductivity (TcSUH), University of Houston, Houston, Texas 77204, USA

(Received 3 July 2012; accepted 14 August 2012; published online 22 August 2012)

By systematically comparing the magnetic properties of the Ta/CoFeB/Ta and MgO/CoFeB/MgO structures with and without a submonolayer of MgO, Ta, V, Nb, Hf and W inserted in the middle of the CoFeB layer, we have proved that the observed perpendicular magnetic anisotropy (PMA) in Ta/CoFeB/MgO sandwiches is solely originated from the CoFeB/MgO interface with the Ta buffer acting to enhance the CoFeB/MgO interface anisotropy significantly. Moreover, replacing Ta with Hf causes the CoFeB/MgO interfacial PMA further enhanced by 35%, and the CoFeB layer with perpendicular magnetization has a much larger critical thickness accordingly, leaving a wider thickness margin for the CoFeB/MgO-based perpendicular magnetic tunnel junction optimization. Also the sputter deposited thin Hf films are amorphous with low surface roughness. These results will ensure the Hf/CoFeB/MgO more promising material system for PMA device development. Copyright 2012 Au-thor(s). This article is distributed under a Creative Commons Attribution 3.0 Un-ported License. []


Magnetic tunnel junctions (MTJs) with ferromagnetic electrodes possessing perpendicular magnetic anisotropy (PMA) have attracted a great deal of attention in recent years.1-10 Such MTJs combine several important advantages over the MTJs with in-plane anisotropy, i.e., higher energy barrier against thermal agitation at reduced dimensions due to larger anisotropy energy; and smaller critical current density and faster reversal speed for current induced magnetization switching (CIMS) due to the absence of the demagnetization term.11-13 So far most of the available PMA materials have been explored as MTJ electrodes, including the rare-earth/ transition-metal alloys,3-5 L10-ordered (Co, Fe)-(Pt, Pd) alloys,6,7 and Co/(Pt, Pd, Ni) multilayers.8-10 Attempts have also been made to search for strong PMA in Co or CoFeB layers sandwiched by Pd (Pt) and MgO (or other oxides) layers.14,15 However, all these material systems suffer from either difficulties in integrating them into MTJs with high magnetoresistance ratio, or insufficient chemical/thermal stability, or large critical current needed for CIMS due to their large Gilbert damping constant. A real breakthrough is the discovery of appreciable interfacial PMA in the widely adopted Ta/CoFeB/MgO films for conventional in-plane MTJs,2 and by incorporating this interfacial PMA high performance perpendicular MTJs have been demonstrated. Further study showed that, electric field can be used to assist magnetization switching by reducing the current density by two orders of magnitude in the perpendicular Ta/CoFeB/MgO based MTJs.16 These exciting findings represent the significant steps towards the next-generation spintronic devices. However, for practical device development, the interfacial PMA in the Ta/CoFeB/MgO system is insufficient to some extent. It will not only

aAuthor to whom correspondence should be addressed; electronic mail:

2158-3226/2012/2(3)/032151/7 2,032151-1 ~ ' I ill I l( ) "Tf MIB

limit the device density scalability, but also challenge the layer deposition and MTJ fabrication technique considering the ultrathin CoFeB layer with a very narrow tunable thickness range. Physically, whether the Ta/CoFeB interface contributes to PMA in the Ta/CoFeB/MgO system is still an open question although the presence of PMA at CoFeB/MgO interface is widely accepted.17 Ikeda et at.2 first ascribed the observed PMA in Ta/CoFeB/MgO to the contribution of the CoFeB/MgO interface. Wang et at.16 further claimed that the PMA in Ta/CoFeB/MgO structure was totally originating from the CoFeB/MgO interface based on the observation that PMA only developed when MgO thickness exceeded 1 nm. However, Worledge et at.18 asserted that the Ta/CoFeB interface also introduces substantial PMA judging from the fact that the Ta/CoFeB/MgO films have an appreciably larger PMA than the Ru/CoFeB/MgO films. In this paper, we demonstrated that the PMA in Ta/CoFeB/MgO sandwich structure solely comes from the CoFeB/MgO interface, yet the Ta buffer essentially enhances CoFeB/MgO interface anisotropy significantly. Furthermore, replacing the Ta buffer by Hf can cause an increase in the CoFeB/MgO interfacial PMA by as much as 35%, leading to a wider CoFeB thickness margin for perpendicular MTJ optimization. In addition, similar to the Ta buffer the sputter deposited thin Hf films also possess amorphous structure with small surface roughness, which can be very beneficial for MgO barrier MTJs by eliminating the harmful template effects existed in the Pt or Pd/CoFeB/MgO based structures.14,15


Films were deposited on the thermally oxidized Si wafers at room temperature by magnetron sputtering. The base pressure of the sputtering system was better than 4x 10-5 Pa and a working argon pressure was 0.5 Pa. Three sets of samples have been studied, these include:

(I) Ta(5)/Co4oFe4oB2o(1.1-4)/Ta(5) with and without a 0.2 nm MgO layer inserted in the middle of CoFeB layer;

(II) MgO(5)/Co40Fe40B20(1.1-4)/MgO(5)/Ta(5) with and without a dusting layer of Ta, V, Nb, Hf and W (o.1 and o.2 nm in nominal thickness) inserted in the middle of CoFeB layer;

(III) Ta(5) or Hf(5)/Co40Fe40B20(0.8-2)/MgO(2)/Ta(5);

where the numbers in parentheses are the nominal thickness of the individual layers in nanometers. The metal and CoFeB films were deposited by dc sputtering while the MgO layer was deposited by rf sputtering. Up to 18 uniform films were prepared in a single deposition run. Samples were subsequently annealed at 300°C for 1 hour in a vacuum furnace (3x 10-5 Pa). Magnetic properties were studied using vibrating sample magnetometer (VSM). The crystalline structure and surface morphologies were checked by x-ray diffraction (XRD) and atomic force microscope (AFM), respectively. All presented results are obtained from room temperature measurements on the annealed samples unless otherwise specified.


To clarify the effect of Ta on PMA in the Ta/CoFeB/MgO structure, magnetic properties of the Ta/Co40Fe40B20/Ta and the MgO/Co40Fe40B20/MgO multilayers were studied first. The annealed Ta/Co40Fe40B20/Ta sandwiches are magnetically soft in the film plane with the areal saturation moment varying linearly with the CoFeB thickness, from which the magnetic dead layer thickness and saturation magnetization (MS) of the CoFeB layers were determined to be 0.38 nm and 1280 emu/cm3, respectively. Meanwhile, the out-of-plane magnetization saturation field for all the Ta/Co40Fe40B20/Ta samples is about 16 kOe, equals to 4nMS regardless the CoFeB layer thickness. Apparently neither volume nor interfacial PMA presents in the annealed Ta/CoFeB/Ta sandwiches. As for the MgO/Co40Fe40B20/MgO films, the saturation magnetization is almost same as that in Ta/CoFeB/Ta films without the magnetic dead layer formation, besides, all MgO/CoFeB/MgO samples also exhibit in-plane anisotropy, but the perpendicular saturation field is smaller than 4nMS, especially for the films with a thin CoFeB layer. As an example, Fig. 1(a) shows the in plane and perpendicular M-H curves of the MgO/CoFeB(1.8)/MgO film. Note that the film normal direction is the magnetic hard axis with a perpendicular saturation field of about 3.6 kOe, much smaller than the

FIG. 1. Representative M-H curves of the (a) MgO(5)/Co40Fe40B20(1.8) /MgO(5)/Ta(5) and (b) MgO(5)/ Co40Fe40B20(1.1)/Ta(0.2)/Co40Fe40B20(1.1)/MgO(5)/ Ta(5) (in nm) structures in in-plane and perpendicular fields. (c) Dependences of Keff iCoFeB on iCoFeB for the samples MgO(5)/Co40Fe40B20(tCoFeByMgO(5)/Ta(5) and with and without the 0.2 nm Ta insertion layer in the middle of the CoFeB layers. The straight lines are the linear fitting results.

demagnetizing field (16 kOe). The effective perpendicular anisotropy (Keff) of theMgO/CoFeB/MgO samples is determined and summarized in Fig. 1(c). By fitting the tCoFeB dependence of the product of Keff and tCoFeB through the equation,

Keff = (Kv - 2nMS) + 2KS/tCoFeB,

the CoFeB volume anisotropy (KV) is found to be negligible and each CoFeB/MgO interface has an interfacial anisotropy (KS) of about 0.6 erg/cm2. Indeed, interfacial PMA is formed at the two MgO/CoFeB interfaces, but apparently this PMA is insufficient to overcome the shape anisotropy. From the results obtained above, the Ta/Co40Fe40B20/MgO films could be naively inferred to possess in-plane anisotropy. This is in conflict with the well known experimental fact. A plausible cause is that in the Ta/Co40Fe40B20/MgO films, Ta may appreciably enhance the CoFeB/MgO interfacial PMA or vice versa, instead of uncorrelated with each other. To prove this argument, a submonolayer of Ta (0.2 nm) is inserted into the middle of the CoFeB layer for the MgO/Co40Fe40B20/MgO structure. Remarkably, the Ta doped films with tCoFeB < 2.2 nm exhibit perpendicular magnetic anisotropy. Figure 1(b) displays the M-H curves obtained from a MgO(5)/CoFeB(1.1)/Ta(0.2)/ CoFeB(1.1)/MgO(5) film. The film normal direction is the magnetic easy axis with a large in-plane saturation field of about 1.5 kOe, in contrast to the pure MgO/CoFeB/MgO films. The measured Keff for the Ta doped films is also presented in Fig. 1(c), showing significantly enhanced Keff with a positive value at tCoFeB < 2.2 nm. The slight deviation of the Keff ■ tCoFeB product from the straight line at small tCoFeB might be due to the deterioration of CoFeB/MgO interface caused by Ta diffusion. Similar data fitting as performed on the pure MgO/CoFeB/MgO films gives a negligible KV and a doubled KS (1.25 erg/cm2). Since the sputter deposited submonolayer should be discontinuous by nature, it is unlikely that appreciable interfacial PMA would occur around such dusting Ta atoms. This speculation is further supported by the same KS enhancement observed in the samples with the MgO(5)/CoFeB(0.8)/Ta(0.1)/CoFeB(0-1.6)/Ta(0.1)/ CoFeB(0.8)/MgO(5)/Ta(5) structure. Therefore, Ta layers even as thin as 0.1 nm can significantly increase the CoFeB/MgO interface anisotropy. On the other hand, when a dusting layer of MgO(0.2 nm) is inserted into the middle of the CoFeB layer in the Ta/Co40Fe40B20/Ta structure, both KV and KS are still negligible. This means that the MgO layer has much less or little effect on the CoFeB/Ta interfacial anisotropy. It thus can be concluded that the widely observed PMA in Ta/CoFeB/MgO sandwiches is mostly originated from the CoFeB/MgO interface with the Ta buffer acting to enhance the CoFeB/MgO interface anisotropy significantly.

In order to further elucidate the origin of the CoFeB/MgO interfacial PMA enhancement through a Ta layer across a thin CoFeB film, a submonolayer (0.2 nm in thickness) of different elements

Hf buffer

Ta buffer

I-1 i 1

" S7~ 13 nm 1.5 nm -O- 1.6 nm

(a) out

-V 13 nm 1.5 nm -O- 1.6 nm

(b) in

—1.1 nm —□— 1.2 nm

(c) out

y 1.1 nm —□— 1.2 nm M

(d) in

H (kOe)

FIG. 2. Representative M-H curves of samples Hf(5), Ta(5)/ Co4oFe4oB2o(icoFeB)/ MgO(2)/Ta(5) (in nm) with external field applied at perpendicular direction in (a), (c) and in-plane in (b), (d).

adjacent to Ta in the periodic table, including V, Nb, Hf and W, has also been inserted into the middle of the CoFeB layers in the MgO/Co40Fe40B20/MgO structure. The corresponding KS of each CoFeB/MgO interface is determined to be 0.8, 0.95, 1.4 and 1.2 erg/cm2 and KV negligible. Theory proposed by Yang et al. suggested that the CoFeB/MgO interfacial PMA comes from the overlap between the O-pz and transition metal dz2 orbits with larger spin-orbit coupling (SOC) induced splitting around the Fermi level for the perpendicular magnetization orientation.17 The SOC of the CoFe(B) layer can be strengthened by the addition of the heavy elements, while the hybridization between CoFe(B) and O, which is related with their electronegativity difference, may also be reinforced by the insertion atoms with smaller electronegativity. Therefore, a heavier element dusting layer has more appreciable effect on KS enhancement, i.e., Hf, Ta, W > Nb > V by sequence, and the dusting layer of elements on the left side in the periodic table with smaller electronegativity19 does likewise, i.e., Hf > Ta > W in effect. Parenthetically, the present method of inserting a submonolayer of Hf, Ta and W into the CoFeB films encapsulated by MgO layers with the magnetization switched to the film normal direction has a potential application in MgO-based double-barrier perpendicular MTJs.

As noted above, the Hf insertion layer enhanced the CoFeB/MgO interfacial anisotropy more effectively than any other dusting layer did. It would be expected that the CoFeB/MgO films grown on an Hf buffer layer have a larger PMA and accordingly a thicker CoFeB layer with perpendicular magnetization as compared with the currently used Ta/CoFeB/MgO films. Figures 2(a) and 2(b) show the representative M-H curves of the Hf(5)/Co4oFe4oB2o(tCoFeB)/ MgO(2)/Ta(5) films with the magnetic field applied in-plane and perpendicular direction, respectively. The samples with tCoFeB < 1.5 nm have perpendicular magnetization easy axis, and the effective perpendicular anisotropy field increases as tCoFeB decreases. Figures 2(c) and 2(d) plot the representative M-H curves for the Ta(5)/Co40Fe40B20(tCoFeB)/MgO(2)/ Ta(5) films prepared under the same condition. Apparently, the critical thickness of the CoFeB layer to show perpendicular easy axis is only 1.1 nm. It should be stressed that, the significant PMA in Hf/CoFeB/MgO films has different origin in comparison with that observed in the Pd or Pt/CoFeB/MgO films.14,15 For the present case, the Hf/CoFeB interface itself does not introduce interfacial PMA, in contrast to the appreciable interfacial anisotropy at Pd or Pt/CoFeB interface. In fact, the Hf/CoFeB/Hf films with and without a 0.2 nm MgO dusting layer at the CoFeB layer center have also been fabricated and magnetically they do not exhibit any PMA component.

The inset of Fig. 3 shows the CoFeB thickness dependences of the saturation magnetic moment per unit area for the Hf and Ta buffered samples, which coincides with each other. The saturation magnetization is determined to be 1650 emu/cm3 and magnetically dead-layer thickness is zero for both of them. Such a large MS value suggests the well crystallization of the CoFeB layers with

&FeB (nm)

FIG. 3. Dependence of ^"effiCoFeB on icoFeB for Hf(5) or Ta(5)/Co4oFe4oB2o/MgO(2)/ Ta(5) (in nm). Inset: the CoFeB thickness dependence of the areal saturation moment. The straight lines are the linear fitting results.

part of boron absorbed by the Hf or Ta underlayer as demonstrated by Hindmarch et al.20 Figure 3 compares the dependence of A"efftCoFeB on tCoFeB for the Hf and Ta buffered samples. Since there is no interfacial anisotropy between the CoFeB layer and the Ta or Hf buffer, the Keff for the Ta or Hf/CoFeB/MgO films should be written as,

Keff = (Kv - 2nMS) + Ks/tCoFeB-

From the data fitting by using this equation, the KS for the Hf/CoFeB/MgO and Ta/CoFeB/MgO structures is determined to be 2.3 and 1.7 erg/cm2, respectively, while KV is essentially zero for both of them. With the Ta buffer replaced by an Hf underlayer, the PMA at the CoFeB/MgO interface is enhanced by 35%. It should be pointed out that, the KS for the Hf or Ta buffered samples is larger than that obtained in the MgO/CoFeB/MgO films with a 0.2 nm Hf or Ta insertion layer. One possibility is the insufficient amount of Ta or Hf atoms involved in the doped MgO/CoFeB/MgO samples. In addition, the less crystallization of the CoFeB layers, reflected by the smaller MS value (125o emu/cm3) obtained in MgO/CoFeB/MgO films with or without the inserting layer, may also contribute to the smaller KS as suggested by Lee et al.21 In fact, Gan et al. have also found that the CoFeB encapsulated by MgO is hard to crystallize during annealing.22

Finally the crystal structure and surface morphology of the sputter deposited Hf films were characterized by x-ray diffraction (XRD) and atomic force microscope (AFM), respectively. Using in-plane grazing incidence (IP-GID) geometry, XRD spectra of Hf films of various thicknesses with a 5 nm Ta cover were collected and some representative results are displayed in Fig. 4(a). While the thick Hf films exhibit appreciable hcp (100) and (110) diffraction peaks, the thin Hf films (tHf < 10 nm) does not have any diffraction peak, indicating the amorphous structure of the Hf film under 10 nm in thickness. The amorphous nature of the buffer layer makes the Hf/CoFeB/MgO system superior to the Pd or Pt buffered CoFeB/MgO films, where the template effect of the (111) oriented Pd and Pt films deteriorates the (001) orientation of CoFeB/MgO and thus leads to a small tunneling magnetoresistanc ratio.14,15 Figures 4(b) and 4(c) show the AFM surface topography of the Hf(5)/Co40Fe40B20(1.2)/MgO(2) and Ta(5)/Co40Fe40B20(1.2)/ MgO(2), respectively. Both samples have a root mean square roughness of about 0.2 nm in a scan area of 1 x 1 fim2, almost equal to that value of the substrate. Therefore, structurally the Hf buffer is as good as the widely adopted Ta films for the CoFeB/MgO based MTJs.

30 40 50 60

26 <°)

FIG. 4. (a) x-ray diffraction (XRD) spectra for the Hf(5, 10, and 20 nm) films covered by a 5 nm Ta, data collected using in-plane grazing incidence (IP-GID) model; The atomic force microscope surface topography of (b) Hf(5)/ Co40Fe40B20(1.2)/MgO(2), and (c) Ta(5)/Co40Fe40B20(1.2)/MgO(2) (in nm).


In summary, the observed PMA in the Ta/CoFeB/MgO sandwiches is solely originated from the CoFeB/MgO interface with the Ta buffer significantly enhancing this CoFeB/MgO interface anisotropy. Most important, by using an Hf underlayer, much higher (a 35% increase) perpendicular magnetic anisotropy can be achieved for CoFeB/MgO interface. In addition, the sputter deposited Hf films are amorphous at small thickness with atomically smooth surface. These characteristics of the Hf/CoFeB/MgO structure can lead to the development of more promising high performance PMA devices.


This work was supported by the National Basic Research Program of China under Grant No. 2009CB929201, and the National Natural Science Foundation ofChina under Grant Nos. 51171205, 51021061 and 50831002.

1 S. Mangin, D. Ravelosona, J. A. Katine, M. J. Carey, B. D. Terris, and E. E. Fullerton, Nat. Mater. 5, 210 (2006).

2 S. Ikeda, K. Miura, H. Yamamoto, K. Mizunuma, H. D. Gan, M. Endo, S. Kanai, J. Hayakawa, F. Matsukura, and H. Ohno, Nat. Mater. 9, 721 (2010).

3 K. Yakushiji, K. Noma, T. Saruya, H. Kubota, A. Fukushima, T. Nagahama, S. Yuasa, and K. Ando, Appl. Phys. Express 3, 053003 (2010).

4 N. Nishimura, T. Hirai, A. Koganei, T. Ikeda, K. Okano, Y. Sekiguchi, and Y. Osada, J. Appl. Phys. 91, 5246 (2002).

5H. Ohmori, T. Hatori, and S. Nakagawa, J. Appl. Phys. 103, 07A911 (2008).

6M. Yoshikawa, E. Kitagawa, T. Nagase, T. Daibou, M. Nagamine, K. Nishiyama, T. Kishi, and H. Yoda, IEEE Trans. Magn. 44, 2573 (2008).

7 G. Kim, Y. Sakuraba, M. Oogane, Y. Ando, and T. Miyazaki, Appl. Phys. Lett. 92, 172502 (2008).

8 K. Yakushiji, T. Saruya, H. Kubota, A. Fukushima, T. Nagahama, S. Yuasa, and K. Ando, Appl. Phys. Lett. 97, 232508 (2010).

9B. Carvello, C. Ducruet, B. Rodmacq, S. Auffret, E. Gautier, G. Gaudin, and B. Dieny, Appl. Phys. Lett. 92, 102508 (2008).

10 J. H. Park, C. Park, T. Jeong, M. T. Moneck, N. T. Nufer, and J. G. Zhu, J. Appl. Phys. 103, 07A917 (2008).

11 J. Z. Sun, Phys. Rev. B 62, 570 (2000).

12 J. C. Slonczewski, J. Magn. Magn. Mater. 159, L1 (1996).

13 D. C. Worledge, G. Hu, D. W. Abraham, J. Z. Sun, P. L. Trouilloud, J. Nowak, S. Brown, M. C. Gaidis, E. J. O'Sullivan, and R. P. Robertazzi, Appl. Phys. Lett. 98, 022501 (2011).

14 J. H. Jung, S. H. Lim, and S. R. Lee, Appl. Phys. Lett. 96, 042503 (2010).

15 L. E. Nistor, B. Rodmacq, S. Auffret, and B. Dieny, Appl. Phys. Lett. 94, 012512 (2009).

16 W. G. Wang, M. Li, S. Hageman, and C. L. Chien, Nat. Mater. 11, 64 (2012).

17 H. X. Yang, M. Chshiev, B. Dieny, J. H. Lee, A. Manchon, and K. H. Shin, Phys. Rev. B 84, 054401 (2011).

18 D. C. Worledge, G. Hu, David W. Abraham, J. Z. Sun, P. L. Trouilloud, J. Nowak, S. Brown, M. C. Gaidis, E. J. O'Sullivan, and R. P. Robertazzi, Appl. Phys. Lett. 98, 022501 (2011).

19 D. R. Lide, CRC Handbook of Chemistry and Physics, 2000-2001 (CRC, Boca Raton, FL, 2000), Sec. 9, p. 75.

20A. T. Hindmarch, V. Harnchana, A. S. Walton, A. P. Brown, R. M. D. Brydson, and C. H. Marrows, Appl. Phys. Express 4,013002(2011).

21 K. Lee, J. J. Sapan, S. H. Kang, and E. E. Fullerton, J. Appl. Phys. 109, 123910 (2011).

22 H. D. Gan, S. Ikeda, W. Shiga, J. Hayakawa, K. Miura, H. Yamamoto, H. Hasegawa, F. Matsukura, T. Ohkubo, K. Hono, and H. Ohno, Appl. Phys. Lett. 96, 0192507 (2010).