Crystal Structure Manipulation of the Exchange Bias in an Antiferromagnetic Film

Exchange bias is one of the most extensively studied phenomena in magnetism, since it exerts a unidirectional anisotropy to a ferromagnet (FM) when coupled to an antiferromagnet (AFM) and the control of the exchange bias is therefore very important for technological applications, such as magnetic random access memory and giant magnetoresistance sensors. In this letter, we report the crystal structure manipulation of the exchange bias in epitaxial hcp Cr2O3 films. By epitaxially growing twined oriented Cr2O3 thin films, of which the c axis and spins of the Cr atoms lie in the film plane, we demonstrate that the exchange bias between Cr2O3 and an adjacent permalloy layer is tuned to in-plane from out-of-plane that has been observed in oriented Cr2O3 films. This is owing to the collinear exchange coupling between the spins of the Cr atoms and the adjacent FM layer. Such a highly anisotropic exchange bias phenomenon is not possible in polycrystalline films.


Results and Discussion
The (1010) oriented Cr 2 O 3 films are grown on the (001) oriented rutile TiO 2 substrates via laser molecular beam epitaxy (LMBE) with a base pressure of 2 × 10 −8 mbar (see methods for details). Fig. 1a,b show the in-situ reflection high-energy electron diffraction (RHEED) characterization of the (001) oriented TiO 2 substrate's surface viewed from the [100] and [110] directions. After the initial growth of 3 nm Cr 2 O 3 , the RHEED pattern of the TiO 2 substrate disappears and that of Cr 2 O 3 starts to appear, as shown in Fig. 1c increases, its RHEED pattern becomes brighter. Fig. 1e-h show the RHEED patterns of 10 nm and 27 nm Cr 2 O 3 films, respectively. To be noted, four satellite RHEED spots are observed around each main diffraction spot viewed from the [100] direction of the TiO 2 substrates (Fig. 1e,g).
The crystalline structural properties of these Cr 2 O 3 films are further characterized by x-ray diffraction (XRD). The θ -2θ scans of the rutile TiO 2 substrate, 10 nm, 20 nm, and 27 nm Cr 2 O 3 films are shown in Fig. 2a. The peak at 2θ of ~63 degrees corresponds to the (002) peak of the TiO 2 substrates. For the 10 nm Cr 2 O 3 film, a peak at 2θ of ~65 degrees is observed, which corresponds to the (3030) peak of the Cr 2 O 3 crystal. As the thickness of Cr 2 O 3 increases, the intensity of the peak at ~65 degrees becomes stronger. We note that for the 27 nm Cr 2 O 3 film, only (3030) peak is detectable for the whole scan range (see supplementary information; Fig. S1), indicating good crystalline properties of the (1010) oriented Cr 2 O 3 thin films. The surface morphology is characterized by atomic force microscopy (AFM). The root-mean-square (RMS) roughness is 0.09 nm for the rutile TiO 2 substrate after annealing in the chamber (Fig. 2b). After the growth of 13 nm Cr 2 O 3 film, the RMS roughness increases to 0.28 nm (Fig. 2c), indicating that the surface of the epitaxial Cr 2 O 3 films is quite smooth.
The epitaxial growth of the (1010) oriented Cr 2 O 3 films on TiO 2 is quite interesting, given the fact that Cr 2 O 3 and TiO 2 belong to totally different space groups. Cr 2 O 3 has a hexagonal crystal structure, which belongs to the    R3c group 17 , while rutile TiO 2 has a cubic structure, which belongs to the P4 2 group 18 . However, the c lattice constant of Cr 2 O 3 is 13.599 Å, and the a lattice constant of TiO 2 is 4.584 Å, which results in a coincidental anion alignment with a lattice mismatch of only ~1.1% . Hence, the Cr 2 O 3 films could be grown with the c axis lying in-plane and parallel to a or b axis of the TiO 2 substrates (Fig. 3a). As the TiO 2 crystal's ab plane has four-fold symmetry, which could result in four-fold in-plane rotational symmetry of the crystalline structure of the Cr 2 O 3 thin films. To investigate this, HRTEM is used to characterize the interfacial structure properties between Cr 2 O 3 and TiO 2 viewed from the [010] direction of the TiO 2 substrate. As shown in Fig. 3b 15,19,20 . The growth mode of (1010) Cr 2 O 3 film in our study is similar to that observed in an earlier report for the (1010) oriented Fe 2 O 3 films grown on rutile (001).TiO 2 substrates 21 . Interestingly, with the c axis lying in the film plane, the spin orientations of the Cr atoms also lie in-plane in these (1010) oriented Cr 2 O 3 film, as schematically shown in Fig. 3a. This is very different from previously  Fig. 4a,b). The in-plane magnetic hysteresis loop are first measured. Prior to the measurement, the sample is cooled from 400 to 10 K in an in-plane magnetic field of 1000 Oe along the [100] direction of the TiO 2 substrate, which is much smaller than the spin-flop field of several Tesla for Cr 2 O 3 reported previously 22 . By cooling through the blocking temperature, the magnetization direction of the Py sets the surface spin configurations of the Cr 2 O 3 films. Then, we measure the magnetization of the Py as a function of the in-plane magnetic field along the [100] direction of the TiO 2 substrate (Fig. 4a) from 10 to 300 K. After subtracting a linear background which is mainly due to the diamagnetic response of the rutile TiO 2 substrate, the shifted magnetic hysteresis loops of Py are displayed in Fig. 4c. At 10 K, as the magnetic field ramps from negative to positive, a sharp jump in magnetic moment occurs at ~110 Oe, but the jump occurs at ~ − 400 Oe on the return sweep. These two switching fields are labeled as H 1 and H 2 , respectively, as indicated in the top panel of the Fig. 4c . As the temperature increases, the exchange bias field steadily decreases from the low temperature value.
To characterize the anisotropy of the exchange bias effect, magnetic hysteresis loops are measured with a magnetic field perpendicular to the films (Fig. 4b). The same field cooling procedure as with the in-plane magnetic fields is adopted. The out-of-plane magnetization curves measured at 10, 20, 40 and 60 K are shown in Fig. 4d. The symmetric magnetization hysteresis loops indicate a negligible perpendicular exchange, which is in stark contrast to the out-of-plane exchange bias observed in the (0001) oriented Cr 2 O 3 films. The highly anisotropic exchange bias phenomenon can be attributed to the crystalline orientation difference of the Cr 2 O 3 films. In hcp structures, the c-axis direction dictates the spin orientations of the magnetic atoms. In the (0001) oriented Cr 2 O 3 films, the spin orientations of the Cr atoms are perpendicular to the films, whereas in the (1010) oriented Cr 2 O 3 films, the spins of the Cr atoms lie in the film plane. These results further indicate the direct collinear exchange coupling between the spins of the Cr atoms and the adjacent Py layer. Fig. 5a,b summarize the in-plane exchange bias field and in-plane conceive field (H C ), where , for the sample consisting of the 13 nm Cr 2 O 3 films and 10 nm Py as a function of the temperature. As the temperature increases from 10 to 60 K, the in-plane exchange bias field (Fig. 5a, Blue dots) decreases quickly from ~ − 150 Oe to almost 0 Oe, where the sign depends on the magnetic field direction during magnetic cooling. No exchange bias is observable at and above 60 K, which implies a blocking temperature (T B ) of ~60 K. An abrupt increase in Hc below 60 K is another property of exchange biased Py, which is due to the formation of the AFM order in this 13 nm Cr 2 O 3 thin film 2 . As there are two crystalline zones of (1010) oriented Cr 2 O 3 , as indicated in zones [0001] and [1210], we also measure the exchange bias in the direction along the TiO 2 [010] direction. Almost identical exchange biases are observed at each temperature (Fig. 5a, Green dots).
The measured T B of 13 nm Cr 2 O 3 is ~60 K, which is much lower compared to the value reported on (0001) oriented bulk Cr 2 O 3 single crystals 15 . In antiferromagnetic films, it has been known that T B is highly related to Neel temperature (T N ), and is usually slightly lower than the T N . Both T B and T N increase as the AFM thickness increases due to finite-size effects 2,12,23,24 . To obtain the T B as a function of the thicknesses of the Cr 2 O 3 thin films, the in-plane exchange bias for the samples consisting of 7, 10, 20 and 27 nm Cr 2 O 3 films and 10 nm Py bilayer films are also measured. Fig. 6a,b show the exchange bias as a function of temperature for the bilayer structures consisting of Cr 2 O 3 (7 nm)/Py (10 nm) and Cr 2 O 3 (27 nm)/Py (10 nm), respectively. The blocking temperatures of these two structures are determined to be 40 K and 100 K. The blocking temperature increases as the thickness of the Cr 2 O 3 films increases, as shown in Fig. 6c. For the 27 nm Cr 2 O 3 film, the blocking temperature is only ~100 K, which is far below the T B of bulk Cr 2 O 3 . One possible reason is the non-trivial finite size effects arising from the grain boundaries or oxygen defects in the Cr 2 O 3 25 .

Conclusion
In summary, we have demonstrated the manipulation of the exchange bias of the Cr 2 O 3 thin films by controlling the surface spin orientations of the Cr atoms via crystal orientation design. For the epitaxial growth of (1010) oriented Cr 2 O 3 films, the spin configurations of Cr atoms give rise to only in-plane exchange bias at the interface between Py and the Cr 2 O 3 thin films, while no perpendicular exchange bias is observed. Our results along with previous studies on (0001) oriented Cr 2 O 3 films indicate the collinear exchange coupling between the spins of the Cr atoms and the adjacent FM layer.

Methods
Cr 2 O 3 films growth. The (1010) oriented Cr 2 O 3 films are grown on the (001) oriented rutile TiO 2 substrates via laser molecular beam epitaxy (LMBE) with a base pressure of 2 × 10 −8 mbar. Prior to the Cr 2 O 3 growth, the substrate temperature is increased to 350 °C with a rate of 20 °C/min in the chamber with an oxygen partial pressure of 0.08 mbar. Then the Cr 2 O 3 film is deposited from a Cr 2 O 3 target with a laser power of (8.0 ± 0.2) mJ and a frequency of 2.0 Hz. The thickness of the Cr 2 O 3 thin film (t) is determined from the cross section high resolution transmission electron microscopy.
Exchange bias measurement. A 10 nm Py is grown on top of the (1010) oriented Cr 2 O 3 films by RF magnetron sputtering with a growth rate of 0.02 Å/s. A capping layer of 20 nm aluminum is deposited prior to taking the samples out of this sputtering chamber to prevent oxidization of Py. Magnetic Properties Measurement System (MPMS; Quantum Design) is used to measure the magnetic hysteresis loops to determine the exchange bias.