Probing One Antiferromagnetic Antiphase Boundary and Single

Mar 29, 2010 - College of Science King Saud University, Riyadh 11451, Saudi Arabia, ... a high saturation field are observed as compared with the case...
8 downloads 0 Views 2MB Size
pubs.acs.org/NanoLett

Probing One Antiferromagnetic Antiphase Boundary and Single Magnetite Domain Using Nanogap Contacts Han-Chun Wu,*,† Mohamed Abid,*,†,‡ Byong S. Chun,† Rafael Ramos,† Oleg N. Mryasov,§ and Igor V. Shvets† †

CRANN, School of Physics, Trinity College Dublin, Dublin 2, Ireland, ‡ King Abdullah Institute for Nanotechnology, College of Science King Saud University, Riyadh 11451, Saudi Arabia, and § Department of Physics and MINT, University of Alabama, Tuscaloosa, Alabama 35487 ABSTRACT We have probed one antiferromagnetic (AF) antiphase boundary (APB) and a single Fe3O4 domain using nanogap contacts. Our experiments directly demonstrate that, in the case of probing one AF-APB, a large magnetoresistance (MR), high resistivity, and a high saturation field are observed as compared with the case of probing a single Fe3O4 domain. The shape of the temperaturedependent MR curves is also found to differ between the single domain and one of the AF-APB measurements, with a characteristic strong temperature dependence for the single domain and temperature independence for the one AF-APB case. We argue that these observations are indicative of profound changes in the electronic transport across APBs. The investigated APB defects increase the activation energy and disturb the long-range charge ordering of monodomain Fe3O4. KEYWORDS Magnetite, Antiphase Boundary, Monodomain, Half-metal

O

field.10,11 Voogt et al.9 and Eerenstein et al.12 observed superparamagnetic behavior in Fe3O4 films. It has been shown recently from Lorentz microscopy that the exchange interaction at APBs can be overcome by a magnetic field.13 Also Arora et al. showed that the ultrathin Fe3O4 films are ferromagnetic and their magnetization is appreciably larger than that of bulk magnetite.14 It was proposed that the miscompensation of spin moments at the surface and APBs are main factors contributing to the observed enhanced magnetic moment. The presence of APBs leads to a change of magnetic properties; they may well be beneficial in attaining a greater MR response,8,15,16 provided these nanoscale defects are properly manipulated. For the emerging spin electronic devices such as spin torque magnetic random access memory (ST-MRAM) or nonvolatile logic,17 a critical issue is scalability, or the ability to make ever smaller diameter devices while maintaining excellent properties which have been demonstrated at large length scales. This prompted us to investigate the role of specific defects which are expected to be typical for oxide-based nanoelectronic and spintronic devices. Apart from that, measuring the MR of one AF-APB or one magnetite domain is also much more interesting because it occurs on an even shorter length scale and a large MR ratio is expected across one AF-APB.18 Despite all of the intensive research over the past decade, experimental investigation of a single or few AF-APBs’ effect on magnetotransport has not been reported or compared with a single domain case. Two main challenges are to be overcome to fill this gap. First, the magnetic coupling of APBs is not always AF in nature and is difficult to control. Second, the Fe3O4 domain size is so small that it prevented re-

ver the last few decades, significant attention has been devoted to Fe3O4 due to its fascinating electrical and magnetic properties. Fe3O4 is a ferrimagnet with a high critical temperature of about 850 K and is expected to have a nearly fully spin polarized electron band at the Fermi level, making it a strong candidate for spintronic devices.1,2 Initial efforts in exploiting its half metallic nature inmagnetictunneljunctionshavebeenfarfrompromising.3-6 The subexpectation tunnel magnetoresistance (TMR) observed in these investigations was attributed to imperfections of the interface between Fe3O4 and MgO layers, which would alter the scattering and the spin polarization.7 However, a sharp interface was observed by transmission electron microscopy (TEM)5,8 and the magnetic properties of the interface and within the magnetic layers were reported to be not significantly different.9 It is well established that epitaxial Fe3O4 films8,15 and chemical synthesis nanowires16 grown on MgO substrates are reported to contain antiphase boundaries (APBs). Due to a modified cationic configuration at the APBs, the nature of magnetic exchange interactions is expected to be altered. The magnetic coupling over a large fraction of these boundaries is suggested to be antiferromagnetic (AF) in nature.10 The presence of these APBs defects contributes to the unusual magnetic properties of Fe3O4, such as the nonsaturation of magnetization, even at a very high magnetic

* To whom correspondence should be addressed, [email protected] and mabid@ chimie.u-strasbg.fr. Received for review: 09/8/2009 Published on Web: 03/29/2010 © 2010 American Chemical Society

1132

DOI: 10.1021/nl902973p | Nano Lett. 2010, 10, 1132–1136

FIGURE 2. The resistivity (F) as a function of temperature (R-T) during the warming cycle with zero applied field for probing one AF-APB (red line) and single domain (green line). The black line is the R-T for 60 nm thick film. (Inset) F plotted as a function of 1000/ T.

epitaxial growth and establish the growth mode. The presence of the RHEED intensity oscillations confirmed that the films grew in a layer-by-layer mode with a growth rate of 0.3 Å/s. A multicrystal high-resolution X-ray diffractometer was further used to confirm the single phase structural and epitaxial nature of the Fe3O4 films. Device fabrication was carried out by EBL using single-layer positive tone resist PMMA supplied by MicroChem Corp. After development, thick metal contacts consisting of Ti (5 nm)/Au (50 nm) were deposited through e-beam evaporation. All these nanogap contacts are along [100] or [010] directions of the films. Subsequently, after liftoff by using acetone, UV lithography was carried out in order to have macroscopic metal contacts. Figure 2 shows the resistivity (F) as a function of temperature (R-T) during the warming cycle with zero applied field for probing one AF-APB and a single magnetite domain. In order to compare with the ordinary thin film case and also to estimate the number of AF-APBs between the contacts, we show the R-T curve for a 60 nm thick film with a 1 mm contact gap in Figure 2. The resistances were measured by a dc-four-probe method where two probes are used for both current and voltage bias measurements due to the devices’ design. For probing one AF-APB, the gap between the contacts is around 80 nm and the resistivity is around 0.029 Ω cm at room temperature, which is twice the value of a 60 nm thick film and is also 1 order of magnitude larger than the bulk value of 0.005 Ω cm. It was shown by Knittel et al. that the AF-APBs can be directly imaged by conventional scanning probe microscopy.20,21 In this work we focus on AF-APBscharacterizationviaresistivitymeasurement.Roughly, the double resistivity implies the local density of AF-APBs in the current path has been doubled in comparison with macroscopic contact gap measurements. If two AF-APBs are present along the current path, the linear AF-APBs density for a 60 nm film would be one AF-APB per 80 nm, which seems unreasonably large. As it follows from the magnetic force microscopy, for a 100 nm thick film, the linear AFAPBs density is around 1 AF-APB per 200 nm.20 A similar linear density of AF-APBs is expected for 60 nm film.19 On the basis of these arguments, we expect that in our experiments we are probing only one AF-APB and that would correspond to a linear AF-APBs density of about 1 AF-APB per 160 nm. For probing single domain magnetite, the gap between the contacts is around 30 nm. The resistivity in this

FIGURE 1. (a) Typical TEM image of epitaxial Fe3O4 film taken with (220) two beam imaging conditions showing the network of APBs. Schematic of the setup to probe (b) a single APB and (c) a single magnetite domain. (d) Scanning electron microscope image of contacts with a 30 nm gap.

searches up to now from attempting to fabricate even smaller gap electrical contacts. In this work, based on electron beam lithography (EBL) technology, we have demonstrated the feasibility of probing one AF-APB and single magnetic domain of Fe3O4 using contacts with a nanoscale gap. These experiments show that, in the case of probing one AF-APB, a large MR, high resistivity, and a high saturation field are observed as compared to the case of a single Fe3O4 domain. Our temperature-dependent measurements further show that the shape of MR curves for a single Fe3O4 domain strongly depends on the temperature region. Furthermore, we directly demonstrate that the APB increases the activation energy and disturbs the long-range charge ordering of Fe3O4. Figure 1a shows a typical TEM image of an epitaxial Fe3O4 film showing the network of APBs. It is known that the domain size in thin films is dependent on the film thickness.19 For a 60 nm thick Fe3O4 film, the average domain size is around 50 nm and the thickness of APB is in the order of 2-3 nm. To make sure we indeed measure at least one but fewer than three APBs, the gap of the contacts should be between 50 and 100 nm (Figure 1b). To probe a single magnetite domain, a pair of contacts with a gap less than 50 nm is necessary (Figure 1c). The lateral geometry of the devices is shown in Figure 1d with a gap of 30 nm. All microstructures discussed in this paper were fabricated on Fe3O4 epitaxial films of 60 nm. Thin films of Fe3O4 were grown on (001) oriented MgO single crystal substrates using an oxygen plasma assisted molecular beam epitaxy (MBE) with a base pressure of 5 × 10-10 Torr. Details of the growth procedure are given elsewhere.6,8,15 Reflection high-energy electron diffraction (RHEED) was employed to confirm the © 2010 American Chemical Society

1133

DOI: 10.1021/nl902973p | Nano Lett. 2010, 10, 1132-–1136

case is drastically smaller (around 0.0026 Ω cm at room temperature) than that for probing one AF-APB. We find that the single domain resistivity is even less than the bulk value. This observation is consistent with the assumption of the narrow gap measurement corresponding to a defect-free single domain case. This unusually small resistivity allows us to easily investigate the transport properties of magnetite ultrathin films down to a temperature of 40 K. This is a notable result because we built devices for probing one AFAPB and also also single domain magnetite free from APBs and thus made a step forward in understanding how the AFAPBs influence spin-dependent transport and also other physical properties, such as magnetic exchange interactions across AF-APBs. These nanoscale devices also allow us to understand how the AF-APBs influence the Verwey transition22,23 and the activation energy. One can clearly see that the Verwey transition disappears for probing one AF-APB due to a increase in local APBs density, which has been observed in very thin epitaxial Fe3O4 films. However when probing single domain magnetite, the Verwey transition is well-defined. This implies that APBs disturb the long-range charge ordering and, thus, oppose to the phase transition at the Verwey temperature (Tv). We also estimated the activation energies of the three sets of devices by fitting the curves using

FIGURE 3. (a) Magnetoresistance vs field curve at different temperatures and (b) MR ratio at H ) 2T as a function of temperature for probing one AF-APB. (c) Schematic drawing of spin structure disturbance due to the AF-APB with and without an in-plane external field.

the external magnetic field, where φAF ) 0 and φAF ) π/2 correspond to the cases for probing a single domain and a single AF-APB, respectively. Using the values from the resistivity measurements, we get σ0 ) 34.48(1/Ω cm) and t02 ) 350.1(1/Ω cm). In a 2 T external field, the conductivity increases to 35.1(1/Ω cm) and φAF rotates to 87.6°, which is still close to π/2. Using eq 4 in ref 18, we get

[ ]

F ) F0 exp

Ea KBT

where Ea relates to the temperature region 120 K < T < 200 K. Fitting the results shows that the activation energy for probing one AF-APB, a 60 nm thick film and probing single domain magnetite are 55.7, 47.98 (film), and 33.6 meV, respectively. Therefore, the APB also increases the activation energy. Figure 3a shows the MR measurements (up to 2 T) for probing one AF-APB. The external in-plane field is applied along the current direction. From Figure 3a, one can clearly see that the resistance shows a linear response to the external field at all temperatures, which suggests we are measuring one AF-APB.18 In Figure 3c we present sketch a typical spin structure disturbance due to the AF-APB with and without an in-plane external field. We assume that magnetotransport properties of the nanoscale device considered here are dominated by transport across the AF-APB which is consistent with observed linear response to the external field.18 Thus, we can write the conductivity for probing a single AF-APB in the following form

σ ) σ0 + t02 cos2 φAF

AAF ) - √AFd2HMs /(cos2 φAF + cos3 φAF)

where H is the external field and Ms is the saturation magnetization. AAF is the (negative) exchange stiffness for the AF exchange interaction at the boundary. AF is exchange stiffness for the ferromagnetic coupling of the spins, and d is the average distance between two neighboring spin chains along the boundary. Taking the literature value for AF (3 K)10 and d is set to 5 Å, we get AAF around -27.6 K, which is slightly larger than the literature value (-25 K) estimated by Motida et al.29 Thus, to the best of our knowledge this is the first evaluation of the exchange interaction across the AFAPBs on the basis of the transport measurements. For this strong AF coupling, a magnetic field can align the spins far from the boundary, whereas the spins close to the boundary still require a stronger external field to align. To help in the understanding of the nonsaturation, the effect of magnetic anisotropy can be neglected. Figure 3b shows the MR ratio (up to 2 T applied field) as a function of temperature. A MR ratio of -1.8% was achieved at room temperature. One can also notice from Figure 3b that the MR ratio increases with decreasing temperature and peaks at the Verwey transition

(1)

where σ0 is due to the spin-dependent scattering at the interface without an external field and the second term is responsible for the MR effect. φAF (marked in Figure 3) is the angle between the spins at the left side of the APB and © 2010 American Chemical Society

(2)

1134

DOI: 10.1021/nl902973p | Nano Lett. 2010, 10, 1132-–1136

[ ]

F ) F0 exp

Ea KBT

Since the MR , 1, we can write the MR as

MR ≈

Ea(H) - Ea KBT

where Ea(H) is the field-dependent activation energy. It is well-known that a long-range charge ordering is present at T < TV.22,23 Therefore, in region III, we can write the temperature-dependent resistivity as

[ B(H) T ]

F ) F0 exp

1/4

where B (H) is the field-dependent activation energy (in units of K).26 For a small field

B(H) ) B0 - α(T)H + β(T)H2

FIGURE 4. (a-c) Magnetoresistance vs field curve at different temperatures and (d) MR ratio at H ) 2T as a function of temperature for probing a single magnetite domain.

where parameters R(T) and β(T) are found to be only slightly dependent on temperature26 which is in good agreement with our experimental results (Figure 4c). The difference between the spin-dependent transport mechanisms in region I and region III contributes to the difference in MR curve shapes. In region II, the discontinuous change of the entropy27 and the phonon-magnon interaction28 play important roles which make the shape of the MR curve different from those above or below the Verwey temperature and also increase the MR ratio. In summary, we performed the magnetotransport measurements through one AF-APB and a single magnetite domain. Our experiments clearly demonstrate that, in the case of probing one AF-APB, a large MR, high resistivity, and a high saturation field are observed as compared with the case of a single Fe3O4 domain. The former is indicative of the frustrated/perturbed magnetic state in and around AFAPBs, while the MR measurements show how these nanoscale magnetic regions influence the spin-dependent scattering. Further we find that the shape of MR curves for a single Fe3O4 domain is distinctly temperature dependent while essentially independent for the one AF-APB case. Our findings shed light on the origin of spin-dependent transport in Fe3O4 and the role of specific defects which are expected to be typical for oxide-based nanoelectronic and spintronic devices.

temperature. Further decreasing the temperature leads to a lowering of the MR ratio. We would like to mention here the shape of the MR curves for this case is independent of temperature. MR measurements for probing a single magnetite domain at different temperatures are shown in Figure 4a-c. The external in-plane field is applied along the current direction and Figure 4d shows the MR ratio as a function of temperature at 2 T of applied field. A small MR ratio of -0.3% was observed at room temperature which is comparable to the value of the bulk magnetite.24 As can be seen in parts a-c of Figure 4, nonlinear MR curves were observed for a single domain at all temperatures which is different from the MR curve shapes observed in the case of probing one AF-APB. We divide the studied temperature range into three regions: region I, region II, and region III (see Figure 4d). It can be noted from Figure 4 that the shape of the MR curve as well as the MR ratio varies from one identified temperature region to another. This variation can be attributed to different mechanisms of spin-dependent transport dominating at different temperatures. As there are no APBs and the gap between the contacts is close to the domain size, the MR effect is mainly due to the rearrangement of spin moments of atoms. When the temperature is above or below TV (such as region I and region III, respectively), for weak external fields, the configuration of the spins is only affected by the magnetic anisotropy field. A relatively low field is needed to align the spins far from the boundary, which corresponds to the jump in the MR curves. At strong fields, the spins close to the boundary will start to rotate; a linear behavior was observed. In region I, we can write the temperature-dependent resistivity as25 © 2010 American Chemical Society

Acknowledgment. This work was supported by the Science Foundation of Ireland (SFI) under Contract No. 06/IN.1/I91. ONM acknowledges CNMS User support by Oak Ridge National Laboratory Division of Scientific User facilities, U.S. Department of Energy. REFERENCES AND NOTES (1) (2) (3)

1135

Walz, F. J. Phys.: Condens. Matter 2002, 14, R285. Ziese, M. Rep. Prog. Phys. 2002, 65, 143. Brabers, J. H.; V., J.; Walz, F.; Kronmuller, H. J. Phys.: Condens. Matter 2002, 12, 5437. Ghosh, K.; Ogale, S. B.; Pai, S. P.; Robson, M.; Li, E.; Jin, I.; Dong, Z. W.; Greene, R. L.; Ramesh, R.; Venkatesan, T.; Johnson, M. Appl. Phys. Lett. 1998, 73, 689. DOI: 10.1021/nl902973p | Nano Lett. 2010, 10, 1132-–1136

(4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16)

Sensor, P.; Fert, A.; Maurice, J. L.; Montaigne, F.; Petroff, F.; Vaurs, A. Appl. Phys. Lett. 1999, 74, 4017. Li, X. W.; Gupta, A.; Xiao, G.; Qian, W.; Dravid, V. P. Appl. Phys. Lett. 1998, 73, 3282. Wu, H. C.; Arora, S. K.; Mryasov, O. N.; Shvets, I. V. Appl. Phys. Lett. 2008, 92, 182502. Wu, H. C.; Mryasov, O. N.; Radican, K.; Shvets, I. V. Appl. Phys. Lett. 2009, 94, 262506. Arora, S. K.; Sofin, R. G. S.; Shvets, I. V.; Luysberg, M. J. Appl. Phys. 2006, 100, No. 073908. Voogt, F. C.; Palstra, T. T. M.; Niesen, L.; Rogojanu, O. C.; James, M. A.; Hibma, T. Phys. Rev. B 1998, 57, R8107. Margulies, D. T.; Parker, F. T.; Rudee, M. L.; Spada, F. E.; Chapman, J. N.; Aitchison, P. R.; Berkowitz, A. E. Phys. Rev. Lett. 1997, 79, 5162. Margulies, D. T.; Parker, F. T.; Spada, F. E.; Goldman, R. S.; Li, J.; Sinclair, R.; Berkowitz, A. E. Phys. Rev. B 1996, 53, 9175. Eerenstein, W.; Hibma, T.; Celotto, S. Phys. Rev. B 2004, 70, 184404. Kasama, T.; Dunin-Borkowski, R. E.; Eerenstein, W. Phys. Rev. B 2006, 73, 104432. Arora, S. K.; Wu, H. C.; Choudhary, R. J.; Shvets, I. V.; Mryasov, O. N.; Yao, H. Z.; Ching, W. Y. Phys. Rev. B 2008, 77, 134443. Arora, S. K.; Sofin, R. G. S.; Shvets, I. V. Phys. Rev. B 2005, 72, 134404. Zhang, D.; Liu, Z.; Han, S.; Li, C.; Lei, B.; Stewart, M. P.; Tour, J. M.; Zhou, C. Nano Lett. 2004, 4, 2151.

© 2010 American Chemical Society

(17) Parkin, S.; Jiang, X.; Kaiser, C.; Panchula, A.; Roche, K.; Samant, M. Proc. IEEE 2003, 91, 661. (18) Eerenstein, W.; Palstra, T. T. M.; Saxena, S. S.; Hibma, T. Phys. Rev. Lett. 2002, 88, 247204. (19) Eerenstein, W.; Palstra, T. T. M.; Hibma, T.; Celotto, S. Phys. Rev. B 2002, 68, No. 014428. (20) Knittel, I.; Wei, J.; Zhou, Y.; Arora, S. K.; Shvets, I. V.; Luysberg, M.; Hartmann, U. Phys. Rev. B 2006, 74, 132406. (21) Wei, J. D.; Knittel, I.; Hartmann, U.; Zhou, Y.; Murphy, S.; Shvets, I. V.; Parker, F. T. Appl. Phys. Lett. 2006, 89, 122517. (22) Anderson, P. W. Phys. Rev. 1956, 102, 1008. Verwey, E. J. W. Nature 1939, 144, 327. (23) Wright, J. P.; Attfield, J. P.; Radaelli, P. G. Phys. Rev. Lett. 2001, 87, 266401. (24) Ziese, M.; Blythe, H. J. J. Phys.: Condens. Matter 2000, 12, 13. (25) Ziese, M.; Srinitiwarawong, C. Phys. Rev. B 1998, 58, 11519. (26) Gong, G. Q.; Gupta, A.; Xiao, G.; Qian, W.; Dravid, V. P. Phys. Rev. B 1997, 56, 5096. (27) Gridin, V. V.; Hearne, G. R.; Honig, J. M. Phys. Rev. B 1996, 53, 15518. (28) Ogale, S. B.; Ghosh, K.; Sharma, R. P.; Greene, R. L.; Ramesh, R.; Venkatesan, T. Phys. Rev. B 1998, 57, 7823. (29) Motida, K.; Miyahara, S. J. Phys. Soc. Jpn. 1970, 28, 1188. Boekema, C.; van der Woude, F.; Sawatzky, G. A. Int. J. Magn. 1971, 3, 341.

1136

DOI: 10.1021/nl902973p | Nano Lett. 2010, 10, 1132-–1136