Enhanced ferromagnetism in single crystalline Co-doped ZnO thin films by Al codoping

Journal of Alloys and Compounds 494 (2010) 392–395

Contents lists available at ScienceDirect

Journal of Alloys and Compounds journal homepage: www.elsevier.com/locate/jallcom

Enhanced ferromagnetism in single crystalline Co-doped ZnO thin films by Al codoping Z.L. Lu a,b,c,∗ , W. Miao a , W.Q. Zou a , M.X. Xu c , F.M. Zhang a a b c

National Laboratory of Solid State Microstructure, and Department of Physics, Nanjing University, Nanjing 210093, China Physics Department and Institute of Innovations and Advanced Studies (IIAS), National Cheng Kung University, Tainan 701, Taiwan, ROC Department of Physics, Southeast University, Nanjing 210096, China

a r t i c l e

i n f o

Article history: Received 10 September 2009 Received in revised form 9 January 2010 Accepted 11 January 2010 Available online 18 January 2010 PACS: 75.50.Pp 61.05.cj 73.61.Ga 75.25.+z Keywords: Co-doped ZnO Diluted magnetic semiconductors X-ray absorption fine structure Single crystalline thin films

a b s t r a c t High-quality (Co, Al)-codoped ZnO single crystalline films have been grown on a-plane sapphire substrates using molecular-beam epitaxy. Al codoping yields the Co-doped ZnO thin films to exhibit metallic conducting behavior with high fee carrier concentration. X-ray absorption studies confirm that nearly all Co ions are in divalent state and actually substituted into the ZnO lattice without formation of any detectable secondary phase. Compared with weak ferromagnetism (0.32 ␮B /Co2+ ) in the Co-doped ZnO single crystalline film, the film with additional Al codoping was found to have much stronger ferromagnetism (0.83 ␮B /Co2+ ) at room temperature. The magnetic anisotropy was also observed in the (Co, Al)-codoped ZnO thin film, indicating the ferromagnetism is intrinsic. Our experimental observations suggest that additional electron doping is effective to strengthen the ferromagnetism in Co-doped ZnO films, which is well consistent with the recent theoretical description of carrier-mediated magnetism in Co-doped ZnO. © 2010 Elsevier B.V. All rights reserved.

1. Introduction Diluted magnetic semiconductors (DMSs) have been expected as sources of spin-polarized carriers for future semiconductor devices which can simultaneously utilize the charge and spin degrees of freedom of the carrier, and are currently attracting a considerable amount of attention [1,2]. Transition metal (TM) doped ZnO is especially exciting in this context due to possible room temperature (RT) ferromagnetism compared with the GaMnAs and InMnAs. Theoretically, Dietl et al. first predicted that p-type Mn doped ZnO would show ferromagnetism with TC well above RT [3]. In contrast, Co-doped ZnO was theoretically predicated to become ferromagnetic with n-type doping [4] and a flood of experimental studies of this subject have been reported upon [5–24]. However, the presented data were plagued by instability and a lack of reproducibility. In some cases, the ferromagnetism has been attributed to uncontrolled ferromagnetic clusters or secondary phases [10,11], which are absent from single phase specimens [12].

∗ Corresponding author at: National Laboratory of Solid State Microstructure, Nanjing University, Nanjing 210093, China. Tel.: +86 25 8320 5360; fax: +86 25 8359 5535. E-mail address: [email protected] (Z.L. Lu). 0925-8388/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.jallcom.2010.01.056

In some cases, high-temperature ferromagnetism could have been caused by defects, such as Zn interstitials, O vacancies and H contamination, which are present in the TM-doped ZnO, as explained by the bound magnetic polaron (BMP) model [13–15]. Additionally, it is surprising that several groups observed distinct ferromagnetic signals in ZnO thin films and nano-particles without TM doping [16,17], and it seems that the charge-transfer ferromagnetism model can give a satisfactory explanation for them [18]. However, in all the above cases, they are not true DMSs. In a true DMS, the spins of the magnetic dopant retain remanent alignment under the influence of spin-polarized free carriers [19]. Recently, careful theoretical studies indicate substitutional Co in ZnO has a weak preference for antiferromagnetic ordering and stabilization of ferromagnetism can be achieved through n-type doping [20]. Behan et al. observed robust carrier-mediated ferromagnetism and strong magneto-optic signals in the metallic Al-doped ZnCoO film with high carrier density, which they claimed to be a genuine magnetic semiconductor [21]. However, Kaspar et al. argued that lack of ferromagnetism was observed in the ZnO:CoAl films even the carrier density is very high (1020 cm−3 ) [22]. The reasons for the distinct discrepancy between their works are not clear until now. In fact, most of the above films were typical polycrystalline, and the reproducibility was usually poor. Therefore, high-quality single crystalline films with good reproducibility, as well as more care-

Z.L. Lu et al. / Journal of Alloys and Compounds 494 (2010) 392–395

393

ful studies of the local structural and magnetic characteristics, are required to clarify the mechanism of ferromagnetism in Co-doped ZnO DMSs. In the present work, we report our experimental studies of metallic and magnetic (Co, Al)-codoped ZnO single crystalline films that are grown by molecular-beam epitaxy (MBE). 2. Experimental details Zn0.96 Co0.04 O (ZnCoO) and Zn0.95 Co0.04 Al0.01 O (ZnCoAlO) thin films with a thickness of 500 Å were prepared in an MBE system with a base pressure of 5 × 10−10 Torr on Al2 O3 (1 1 2¯ 0) substrate at 400 ◦ C using the ␦-doping technique [23,24]. A 50 Å pure ZnO film was grown on the substrate as a buffer layer before the film was deposited. Pure (99.995%) ZnO and Co in a specific ratio were evaporated from two independent e-beam sources, while Al ions from a source were co-deposited on the ZnO layer. The atomic concentrations of the doping element were checked by energy dispersive X-ray analysis (EDAX). The Al concentration in the ZnCoAlO film was controlled to be about 1%. The typical growth pressure of the doped films was below 5 × 10−8 Torr and the deposition rate was about 0.02 Å/s. The film structure and crystalline quality were determined by in situ high-energy electron diffraction (RHEED) and high-resolution X-ray diffraction (XRD) at the BL17B beamline of the Taiwan Light Source (TLS) in Hsinchu, Taiwan. The X-ray-absorption fine structure (XAFS) measurements of the Zn and Co K-edges were performed in the wiggler-C beamline of TLS to determine the electronic state and local environment of the Zn and Co in the doped films. Magnetization studies were carried out using a superconducting quantum interference device (SQUID) magnetometer (MPMS, XL-7T) in the temperature range 5–350 K. Electrical transport including Hall measurements was carried out using a four terminal van der Pauw configuration in magnetic fields of up to 1 T.

3. Results and discussion Because of neglectable lattice mismatch (<0.1%) between ZnO (0 0 0 1) and Al2 O3 (1 1 2¯ 0) [25], epitaxial growth of high-quality ZnCoO and ZnCoAlO films were realized on Al2 O3 (1 1 2¯ 0) substrates. For both the as-grown films, clear c-oriented ZnO [1 0 1¯ 0] and [1 1 2¯ 0] streaky in situ RHEED patterns were observed (Fig. 1(a) inset for ZnCoAlO), indicating the doped ZnO epilayers were grown in the Frank–Van der Merwe (FM) model with two-dimensional layer-by-layer growth [26]. The out-of-plane and in-plane orientational relationships are ZnCoAlO [0 0 0 1]//sapphire [1 1 2¯ 0] and ZnCoAlO [1 1 2¯ 0]//sapphire [0 0 0 1]. It seems the doped films grown on a-plane sapphire are high-quality single crystalline thin films with atomic flat surface morphology, which are suitable for multilayer devices. High-resolution X-ray measurements also confirm that both films exhibit a strong c-axis texture without any second phase; typical result was shown in Fig. 1(a). The full width at half maximum (FWHM) of (0 0 0 2) peak is about 0.20◦ , suggesting good crystallinity of the doped films. The typical  scan from a (1 0 1¯ 1) plane of the ZnCoAlO thin film was also performed to verify its single crystalline characteristics, as shown in Fig. 1(b). The sixfold symmetry of the ZnO hexagonal structure crystal basal plane is clearly revealed by the six peaks separated by 60◦ . The FWHM of the peaks is about 1.4◦ , revealing very good epitaxy. It is important but difficult to detect whether there are a trace of Co clusters in Co-doped ZnO film. XAFS based on synchronization radiation is one of the most powerful techniques in this area and is very sensitive to formation of metal bonding to the level of a few percent of the dopant atoms [11]. Here the electronic state and local environment of Co and Ga in ZnCoO and ZnCoAlO films were studied by XAFS. The X-ray absorption near-edge structure (XANES) is highly sensitive to the presence of metal clusters in host oxides. Fig. 2(a) shows the Co K-edge XANES spectra of the samples as well as those of the standard Co metal and the oxide for comparison. In contrast to the marked shoulder around 7712 eV for the Co metal, the Co K-edge spectra of both the samples show clear 1s to 3d preedge features around 7709 eV, which are characteristic of Co2+ substitution for Zn2+ in ZnO [27]. Extended X-ray absorption fine structure (EXAFS) is also adopted to clarify the local structures around the Co atoms. Fig. 2(b) plots the radial distribution function (RDF), the Fourier transform amplitude of EXAFS, at the Co and

Fig. 1. (a) X-ray –2 scan and (b)  scan from a (1 0 1¯ 1) plane of the ZnCoAlO thin film grown on a-plane sapphire substrate, respectively. The inset shows the typical RHEED patterns.

Zn K-edge for both the samples. For the Zn RDF, the first and second peaks observed can be identified as Zn–O and Zn–Zn bonding, respectively. The RDFs for both the films at the Co K-edge are very similar to those of the Zn K-edge spectra, implying that Co ions have similar local structures, as do the Zn ions in ZnO. The results reveal that nearly all Co atoms actually substitute into the ZnO lattice without the formation of any detectable metallic Co or other secondary phase. Measurements of the resistivity and Hall effect of the ZnCoO and ZnCoAlO samples were made, and the data shown in Table 1. Clearly, the resistivity  of the ZnCoO film greatly decreases from 7.1 × 10−3 to 4.5 × 10−4  cm as the Al doping, because of a rapid increase in both the carrier concentration nc and the Hall mobility . The low resistivity and high mobility suggest the carriers are weakly localized in the films. In order to further study the electric properties of the films, the Fermi temperature TF and the mean free path  at RT were estimated with equation TF = 2

2/3

1/3

h ¯ (32 nc ) /(2me kB ) and  = h ¯ (32 ) /(nc 2/3 e2 ), respectively, where  is the Planck’s constant, me the electron mass, kB Boltzmann constant, and e the electron charge [28]. Clearly, for both the Table 1 Electrical and magnetic parameters of the ZnCoO samples with and without Al codoping at 300 K. Sample ZnCoO ZnCoAlO

 ( cm) −3

nc (cm−3 )

7.1 × 10 4.7 × 10 4.5 × 10−4 4.8 × 1020

19

 (cm2 V−1 s−1 )

TF (K)

 (Å)

MS (␮B /Co2+ )

18.7 29.3

496 2596

15.4 46.2

0.32 0.83

394

Z.L. Lu et al. / Journal of Alloys and Compounds 494 (2010) 392–395

Fig. 2. (a) Normalized absorption spectra of the ZnCoO and ZnCoAlO thin films on Co K-edges and (b) the radial distribution functions of Co from Fourier transform magnitude of EXAFS of Co K-edges, respectively.

films, TF is higher than RT and  is much greater than the lattice spacing (3 Å), which satisfy the criteria for identifying the metallic regime of the ZnO semiconductor [21]. So it is reasonable to conclude that the charge carriers are dislocalized and itinerant in both the films, which would be effective to mediate the doped magnetic ions. The field dependence of in-plane magnetization of the ZnCoO and ZnCoAlO films was measured at 300 K, as shown in Fig. 3(a). To eliminate the influence of substrates, the M–H curves of the films were obtained after the contribution of the sapphire substrate had been subtracted. The ZnCoO thin film shows ferromagnetic ordering with a low magnetic moment (Ms ) of only 0.32 ␮B /Co2+ and coercive field (Hc ) of 120 Oe at 300 K. Because the possibility of existing Co clusters or any other second phase can be excluded by XAFS studies discussed above, the observed ferromagnetism should be intrinsic due to Co doping into the ZnO lattice. Notably, the ferromagnetism became much stronger when additional Al was codoped in the film, and the values of Ms and Hc increase greatly to 0.83 ␮B /Co2+ and 150 Oe, respectively. The enhanced ferromagnetism can be attributed to the significant increase of the free carrier concentration, which was predicted to be effective to strengthen the ferromagnetic interaction in Co-doped ZnO films [20]. Moreover, when the applied magnetic field is perpendicular to the film plane, the ZnCoAlO thin film gets saturated harder with a less coercivity of about 100 Oe (Fig. 3(b)), which indicates that the film has magnetic anisotropy with the easy axis of magnetization parallel to the film plane. In fact, the magnetic anisotropy is a further proof of the intrinsic nature of the ferromagnetism in

Fig. 3. (a) Magnetic hysteresis loops measured at 300 K with the applied magnetic field parallel to the film plane for both the ZnCoO and ZnCoAlO thin films. (b) Typical M–H curves of the ZnCoAlO sample at 300 K with the applied magnetic field parallel and perpendicular to the film plane, respectively.

ZnO DMS materials, since cluster-related ferromagnetism is generally isotropic [29]. Finally, in order to shed light on the magnetic origin in the films, both the zero-field-cooled (ZFC) and fieldcooled (FC) magnetizations as a function of temperature at applied field of 50 Oe were measured, and the results for the films with and without Al codoping are shown in Fig. 4. Obviously, the ZFC curves appear rather smooth between 300 and 20 K and no blocking temperature due to ferromagnetic nano-clusters could be seen.

Fig. 4. Magnetization of the ZnCoO and ZnCoAlO thin films as a function of temperature in a field of 50 Oe with zero field cooling (ZFC) and field cooling (FC). The fitting results of the FC data are also shown.

Z.L. Lu et al. / Journal of Alloys and Compounds 494 (2010) 392–395

Moreover, it is found that the FC M(T) data can be well fitted by a standard Block spin-wave model (M(T) = M0 (1 − AT3/2 )) [30], where M0 is the zero temperature magnetization, added a Curie–Weiss model ( = 0 + C/(T − )), revealing a mixed magnetic phase, i.e. a ferromagnetic phase plus a paramagnetic one. Indeed, there is just a fraction of doped Co2+ ions in ferromagnetic ordering and the others are still in paramagnetic or antiferromagnetic ordering, with considering the saturated magnetic moment is only several tenths ␮B /Co2+ , much below the expected value of 3 ␮B /Co2+ for tetrahedrally coordinated high-spin Co2+ ions. Moreover, for the paramagnetic region, the fitted Curie temperature  is about −10 K, indicating the existence of antiferromagnetic interactions of doped Co2+ ions [8]. It is noted that Dietal et al. has interpreted (Zn,Co)O in terms of spinodal decomposition into Co-rich antiferromagnetic nanocrystals embedded in the Co-poor (Zn,Co)O host and attribute ferromagnetic signatures to uncompensated spins at the surface of the antiferromagnetic nanocrystals [31]. However, in that case, the ferromagnetism is very weak, typically lower than 0.1 ␮B /Co2+ at RT, which cannot explain the much larger magnetic moments here. Certainly, there is still the possibility of such non-uniform Co distributions in our samples, an additional effort aiming at establishing the Co nano-scale distribution is necessary and it will be studied carefully at the next step in the future. The magnetic mechanisms—either carrier-mediated or defectinduced ferromagnetism in oxide DMSs, such as Co:ZnO, remain debatable because of the intimate correlation between carrier concentration and structural defects (e.g. O vacancies, Zn interstitials). Structural defects are considered to be essential to ferromagnetism in the insulating and semiconducting regimes of Co:ZnO with a low free carrier concentration or mobility [21]. In such cases, most carriers tend to be localized, favoring the forming of BMPs, which lead to the ferromagnetism. However, the BMP mechanism is not applicable here because the free carrier concentration is high (>1020 cm−3 ) and the mean free path  (>10 Å) is large. Notably, both the doped samples here are single crystalline films with high quality, and there should be much fewer structure defects in the films as compared with that in other reports. Therefore, defect-induced ferromagnetism can reasonably be excluded, or its contribution to the ferromagnetism is not dominant in the ZnCoAlO single crystalline film. Very recently, Walsh et al. reported a theoretical description of carrier-mediated magnetism in Co-doped ZnO and demonstrated that this system is intrinsically antiferromagnetic, while the addition of charge carrier density above a threshold (nc /ni ∼ 0.03, where ni is the density of Co) would result in a transition to ferromagnetic ordering through enhanced coupling between the Co t2d states [20]. Clearly, our experimental observation is consistent with the theoretical study. With the Al codoping, the trend of strengthened ferromagnetic interaction by additional free carrier is clear, which strongly support the theoretical investigation. Finally, we note that the moment of the ZnCoAlO film (0.83 ␮B /Co2+ ) is much below the expected value of 3 ␮B /Co2+ for tetrahedrally coordinated high-spin Co2+ ions. Further experimental and theoretical studies are thus necessary. Magnetic circular dichroism (MCD), neutron diffraction, and spin detection through tunnel junctions are underway to investigate. 4. Conclusions In conclusion, high-quality (Co, Al)-codoped ZnO single crystalline films have been grown on a-plane sapphire substrates by MBE. Al doping yields the ZnCoO thin film with metallic conducting behavior at high fee carrier concentration. Compared with weak ferromagnetism in the Co-doped ZnO single crystalline film, the

395

film with additional Al codoping was found to have much stronger ferromagnetic ordering at room temperature. Our experimental observations indicate that additional electron doping is effective to strengthen the ferromagnetism in Co-doped ZnO films, which is well consistent with the recent theoretical description of carriermediated magnetism in Co-doped ZnO. Acknowledgements This work was partially supported by National Science Foundation of China (NSFC Grant No. 10804017), National Science Foundation of Jiangsu Province of China (Grant No. BK2007118), Research Fund for the Doctoral Program of Higher Education of China (Grant No. 20070286037), Cyanine-Project Foundation of Jiangsu Province of China (Grant No. 1107020060), Foundation for Climax Talents Plan in Six-Big Fields of Jiangsu Province of China (Grant No. 1107020070) and New Century Excellent Talents in University (NCET-05-0452). References [1] H. Ohno, Science 281 (1998) 951. [2] S.A. Wolf, D.D. Awschalom, R.A. Buhrman, J.M. Daughton, S. von Molnar, M.L. Roukes, A.Y. Chtchelkanova, D. Treger, Science 294 (2001) 1488. [3] T. Dietl, H. Ohno, F. Matsukura, J. Cubert, D. Ferrand, Science 287 (2000) 1019. [4] K. Sato, H. Katyama-Yoshida, Semicond. Sci. Technol. 17 (2002) 367. [5] Ü. Özgür, Ya.I. Alivov, C. Liu, A. Teke, M.A. Reshchikov, S. Do˘gan, V. Avrutin, S.J. Cho, H. Morkoc, J. Appl. Phys. 98 (2005) 041301. [6] S.A. Chambers, Surf. Sci. Rep. 61 (2006) 345. [7] J.M.D. Coey, S.A. Chambers, MRS Bull. 33 (2008) 1053. [8] F. Pan, C. Song, X.J. Liu, Y.C. Yang, F. Zeng, Mater. Sci. Eng. R 62 (2008) 1. [9] Z.L. Lu, H.S. Hsu, Y.H. Tzeng, F.M. Zhang, Y.W. Du, J.C.A. Huang, Appl. Phys. Lett. 95 (2009) 102501. [10] Y. Wang, L. Sun, L.G. Kong, J.F. Kang, X. Zhang, R.Q. Han, J. Alloys Compd. 423 (2006) 256. [11] S.M. Heald, T. Kaspar, T. Droubay, V. Shutthanandan, S. Chambers, A. Mokhtari, A.J. Behan, H.J. Blythe, J.R. Neal, A.M. Fox, G.A. Gehring, Phys. Rev. B 79 (2009) 075202. [12] G.Q. Pei, C.T. Xia, F. Wu, J. Xu, J. Alloys Compd. 467 (2009) 539. [13] X.J. Liu, X.Y. Zhu, J.T. Luo, F. Zeng, F. Pan, J. Alloys Compd. 482 (2009) 224. [14] S.H. Lee, Y.C. Cho, S.J. Kim, C.R. Cho, S.Y. Jeong, S.J. Kim, J.P. Kim, Y.N. Choi, J.M. Sur, Appl. Phys. Lett. 94 (2009) 212507. [15] J.L. MacManus-Driscoll, N. Khare, Y. Liu, M.E. Vickers, Adv. Mater. 19 (2007) 2925. [16] M. Kapilashrami, J. Xu, V. Ström, K.V. Rao, L. Belova, Appl. Phys. Lett. 95 (2009) 033104. [17] Q.Y. Xu, S.Q. Zhou, H. Schmidt, J. Alloy Compd. 487 (2009) 665. [18] J.M.D. Coey, W. Kwanruthai, J. Alaria, M. Venkatesan, J. Phys. D: Appl. Phys. 41 (2008) 134012. [19] S.J. Pearton, C.R. Abemathy, M.E. Overberg, G.T. Thaler, D.P. Norton, N. Theodoropoulou, A.F. Hebard, Y.D. Park, F. Ren, J. Kim, L.A. Boatner, J. Appl. Phys. 93 (2003) 1. [20] A. Walsh, J.L.F. Da Silva, S.H. Wei, Phys. Rev. Lett. 100 (2008) 256401. [21] A.J. Behan, A. Mokhtari, H.J. Blythe, D. Score, X.-H. Xu, J.R. Neal, A.M. Fox, G.A. Gehring, Phys. Rev. Lett. 100 (2008) 047206. [22] T.C. Kaspar, T. Droubay, S.M. Heald, P. Nachimuthu, C.M. Wang, V. Shutthanandan, C.A. Johnson, D.R. Gamelin, S.A. Chambers, New J. Phys. 10 (2008) 055010. [23] J.C.A. Huang, H.S. Hsu, Y.M. Hu, C.H. Lee, Y.H. Huang, M.Z. Lin, Appl. Phys. Lett. 85 (2004) 3815. [24] Z.L. Lu, H.S. Hsu, Y.H. Tzeng, J.C.A. Huang, Appl. Phys. Lett. 94 (2009) 152507. [25] P. Fons, K. Iwata, A. Yamada, K. Matsubara, S. Niki, K. Nakahara, T. Tanabe, H. Takasu, Appl. Phys. Lett. 77 (2000) 1801. [26] G.L. Liu, Q. Cao, J.X. Deng, P.F. Xing, Y.F. Tian, Y.X. Chen, S.S. Yan, L.M. Mei, Appl. Phys. Lett. 90 (2007) 052504. [27] X.C. Liu, E.W. Shi, Z.Z. Chen, B.Y. Chen, W. Huang, L.X. Song, K.J. Zhou, M.Q. Cui, Z. Xie, B. He, S.Q. Wei, J. Alloys Compd. 463 (2008) 435. [28] Q.Y. Xu, L. Hartmann, H. Schmidt, Phys. Rev. B 76 (2007) 134417. [29] M. Venkatesan, C.B. Fitzgerald, J.G. Lunney, J.M.D. Coey, Phys. Rev. Lett. 93 (2004) 177206. [30] S.J. Potashnik, K.C. Ku, R. Mahendiran, S.H. Chun, R.F. Wang, N. Samarth, P. Schiffer, Phys. Rev. B 66 (2002) 12408. ´ [31] T. Dietl, T. Andrearczyk, A. Lipinska, M. Kiecana, M. Tay, Y. Wu, Phys. Rev. B 76 (2007) 155312.