Molecular beam epitaxy growth and magnetic properties of Cr-Co-Ga Heusler alloy films

We have re-investigated growth and magnetic properties of Cr2CoGa films using molecular beam epitaxy technique. Phase separation and precipitate formation were observed experimentally again in agreement with observation of multiple phases separation in sputtered Cr2CoGa films by M. Meinert et al. However, significant phase separation could be suppressed by proper control of growth conditions. We showed that Cr2CoGa Heusler phase, rather than Co2CrGa phase, constitutes the majority of the sample grown on GaAs(001) at 450 oC. The measured small spin moment of Cr2CoGa is in agreement with predicted HM-FCF nature; however, its Curie temperature is not as high as expected from the theoretical prediction probably due to the off-stoichiometry of Cr2CoGa and the existence of the disorders and phase separation.


I. INTRODUCTION
The search for high-performance materials is a priority for the development of spintronic devices such as spin valves and spin memories. 13][4][5] In addition, an interesting proposed addition to the family of HM-FMs is the zero-moment half-metallic ferrimagnet, named half-metallic fully-compensated ferrimagnets (HM-FCFs). 6HM-FCF has both 100% spin polarization and zero net magnetic moment, which causes the generated stray fields and energy losses in spintronics devices are lower compared with HM-FM with large net magnetic moment.Thereby, HM-FCFs are more suitable in the real application of spintronics devices.Some Heusler alloys and perovskite compounds have been found to be HM-FCFs. 4,7,8Especially the half-metallic Heusler compounds having the formula X 2 YZ attracted intense interest because most of those alloys possess high Curie temperatures, compatible crystalline structure with the zincblende of semiconductors, and generation of rich electronic structures and novel properties by varying the valence of X, Y, and Z. 9 CrMnSb have antiparallel spin magnetic moments of about the same magnitude in Cr and Mn sites. 10Mn 3 Ga, 6,11 Cr 2 MnSb, 12 (Mn 0.5 Co 0.5 ) 2 VAl, 1,13 Co 2 CrAl 14 and Cr2CoAl 4,14 are half-metallic ferrimagnetic alloys with zero or nearly zero spin moment.I. Galanakis and E. Sasıoglu have searched for High Curie temperature (T C ) HM-FCF Heusler compounds by ab-initio electronic structure calculations and found that Cr 2 CoGa (CCG) is a nice HM-FCF with high degree of spin-polarization for a wide range of lattice constants and very high T C . 7However, Cr 2 CoGa was found to be unstable with respect to their elemental constituents according to the density functional theory calculation of the formation enthalpy. 15Manuel P. Geisler et al. have prepared Cr 2 CoGa films by magnetron co-sputtering and observed experimentally the phase separation and precipitate formation in dependence on the heat treatment. 16So it seemed that the instability of the Cr 2 CoGa alloy makes this material unsuitable for the spintronics applications.
Here we re-investigated the growth and properties of Cr 2 CoGa film by molecular beam epitaxy (MBE) and attempted to stabilize the metastable Heusler phase of Cr 2 CoGa.The generation of phase separation and precipitate was again confirmed experimentally.However, our results showed that the majority of Cr 2 CoGa phase could be stabilized and the phase separation could be suppressed to the utmost extent by proper control of growth conditions.The measured spin moment of the sample was small in agreement with the calculated zero-moment; however, the T C was not as high as expected from the theoretical prediction probably due to the off-stoichiometry of Cr 2 CoGa and the existence of the disorders and phase separation.

II. EXPERIMENTAL
Cr 2 CoGa films with thickness of 60 nm were grown on Si(001) and GaAs(001) substrates, respectively, in a standard solid-source MBE chamber (VG Semicon model V80).Substrates were acid dipped to remove surface oxide and heated at 550 o C for around 30 min before sample growth.The ratio of elemental constituents was controlled by adjusting the evaporation rate of respective effusion cells.The accurate composition of samples determined by the electron probe microanalysis (EPMA) measurement.The Cr 2 CoGa films were deposited under an ultrahigh vacuum of 10 −9 Torr at different temperatures.Some of samples were capped by 2 nm GaAs to avoid the contamination of inner layers.Reflection high-energy electron diffraction (RHEED) was applied to monitor the growth process of the Cr 2 CoGa films.The crystal structure and lattice constant were determined by X-ray diffraction (XRD) using Cu K α radiation.The surface morphology of the sample was measured by scanning electron microscope (SEM).The magnetic properties were characterized by superconducting quantum interference device magnetometer (SQUID, Quantum Design).Temperature dependence of resistance (measured with a four point configuration) and magneto-resistance was carried out by a home-made transport property measurement system (TPMS).

A. Growth of Cr 2 CoGa on Si(001) substrate
We first attempted to grow Cr 2 CoGa on Si(001) substrate.By monitoring RHEED pattern during film growth, it was found that crystalline phase can only be formed at the high growth temperature.Fig. 1(a) shows RHEED pattern for the 450 o C-grown Cr 2 CoGa/Si(001) film.The pattern of ordered spots indicated presence of crystalline phase in the sample.As shown in the XRD spectra in Fig. 2, lots of diffraction peaks appeared when sample was grown on Si(001) at 450 o C, which could not be assigned to the Heusler phase of Cr 2 CoGa.Instead, phase separation including Cr 3 Ga, Cr 3 Ga 4 , CoGa 3 , and CrCo could be assigned, in agreement with observations by M. P. Geisler and M. Meinert. 16Conspicuous phase separation could also be seen in the SEM image in Fig. 3(b).Probably due to the large lattice mismatch (6.7%) between Cr 2 CoGa and Si(001), Cr 2 CoGa phase could not be stabilized.The obtained sample composed of multi-phases separation is non-ferromagnetic according to the magnetization vs magnetic field curves shown in Fig. 4. Therefore, it was concluded that Cr 2 CoGa phase could not be stabilized on Si(001) substrate and the choice of substrate with appropriate lattice parameter and symmetry was probably important for the stabilization of Cr 2 CoGa by MBE growth.

B. Growth of Cr 2 CoGa on GaAs(001) substrate
We then grew Cr 2 CoGa on GaAs(001) substrate.As shown in Fig. 1(b), it could be seen that when grown on GaAs(001) at 70 o C, Cr 2 CoGa sample takes on ring-shaped RHEED pattern indicating the formation of non-crystalline or amorphous state of the sample, in agreement with previous observation that crystalline phase can only be formed at the high growth temperature.The absence of crystalline phase in the 70 o C-grown Cr 2 CoGa/GaAs(001) was confirmed by the XRD spectra in Fig. 2, where no distinct diffraction peaks from the film were observed.It indicated that Cr 2 CoGa Heusler alloy does not crystallize at low temperature.Meanwhile, the surface morphology of the sample was homogeneous as seen from the SEM image in Fig. 3(a).It means that phase separation and precipitates also do not happen when Cr 2 CoGa was grown on GaAs(001) at low temperature.As shown in Fig. 4, the non-crystalline Cr 2 CoGa grown at low temperature is also of non-ferromagnetic ordering.001) is 2.4%, around 3 times smaller than that between Cr 2 CoGa on Si(001).However, there were still some three dimensional features every other streak, which was in agreement with existence of more than one pure phase in the sample as seen from XRD in Fig. 2. The possible release of As from GaAs surface during heating substrate at high temperature might lead to formation of defects on the GaAs surface, which would affect the epitaxial growth of film.It was also noted that the as-used GaAs(001) wafer was of poor quality as seen from the GaAs (004) and (002) diffraction intensity in the XRD spectra;  222) facets to relieve the lattice mismatch in between.The corresponding lattice parameter of Cr 2 CoGa is a = 5.79 Å, in agreement with previous reports. 15,17The appearance of Cr 3 Ga phase separation was also in agreement with Ref. 16.The surface morphology of the corresponding sample shown in Fig. 3(c) was relatively rough compared with the sample grown at low temperature; however, almost no distinct phase separation could be observed even in the magnified image in Fig. 3(d).Therefore, it is clear that Cr 2 CoGa phase could be stabilized on GaAs(001) substrate at 450 o C probably due to the small lattice mismatch.The phase separation still exist, however, it was significantly suppressed.We speculated that pure phase Cr 2 CoGa sample could be achieved by proper preparation of a perfect substrate surface and precise control of slow growth rate for the epitaxy in the further research.
In order to determine the relative content of Cr 2 CoGa, Rietveld refinement was performed for the XRD data of 450 o C-grown Cr 2 CoGa/GaAs(001) sample, as shown in Fig. 5. Rietveld refinement was carried out via the computer software General Structure Analysis System (GSAS) program 18 with the Hg 2 CuTi(102972-ICSD), Cr 3 Ga(626025-ICSD) and GaAs(41674-ICSD) crystal structure as starting models.A quantitative analysis of this sample was performed with Rwp = 6.51%,Rp = 4.36% and χ2 = 2.01.A relative content of Co 2 CrGa:Cr 3 Ga = 10:1.In view of the poor quality of XRD data, even though error accounts for 20%, Cr 2 CoGa still account for 72%.Therefore, the Cr 3 Ga phase separation was a minority.Actually according to previous research, 16 if there is Cr 3 Ga, there should also be some Co-based compounds as well.Therefore, some Co-based compounds, such as CoCr, which does not show distinct peaks in the XRD spectra, might also exist in the sample except Cr 3 Ga.The Co-based compound might partially contribute to the observed non-zero magnetic moment.
Figure 6 shows the magnetization vs applied field (M-H) loops for the 450 o C-grown Cr 2 CoGa/ GaAs(001) film.Distinct M-H hysteresis curves were observed at both 300 K and 10 K indicating ferro(ferri)magnetic ordering of Cr 2 CoGa.The saturation magnetization is small, around 70 emu/cm 3 at 10 K.The saturation moment at 10 K was determined to be 0.31 µ B per formula unit, which is larger than the predicted zero moment probably due to the off-stoichiometry of Cr 2 CoGa and the existence of phase separation and disorders.In addition, the distinct difference between the in-plane and out-of-plane saturated magnetization in Fig. 6 indicated existence of the shape anisotropy which could be induced by the tetragonal distortion of Cr 2 CoGa.The tetragonal distortion of Cr 2 CoGa was likely to happen during epitaxy of film due to lattice-mismatch and could also be the reason explaining the observed magnetic moment different from the predicted value.M. P. Geisler et al. have also calculated the formation energies of the phase separation of Cr 2 CoGa. 16The most favorable reaction is 2(2Cr + Co + Ga) → Co 2 CrGa + Cr 3 Ga, with formation energy of -0.61 eV per formula unit.Therefore, the Co 2 CrGa phase might also exist in the 450 o C-grown Cr 2 CoGa/GaAs(001) sample because the Cr 3 Ga phase was detected in the XRD spectra.The crystal structure and lattice parameter of Co 2 CrGa is similar to that of Cr 2 CoGa, therefore, these two phases cannot be distinguished by XRD.However, due to the small content of Cr 3 Ga phase as judged from Rietveld refinement in Fig. 5, the content of Co 2 CrGa should also be few.Moreover, the Co 2 CrGa phase has the magnetic moment of 3 µ B per formula unit according to previous experiments and calculations. 19,20Therefore, the measured 0.31 µ B /f.u. also indicated that it is Cr 2 CoGa phase, rather than Co 2 CrGa, that constitutes the majority of the sample.is probably a ferrimagnet, however, the ZFC M-T curve in Fig. 6 has non-standard ferrimagnetic behavior which probably due to the coexistence of antiferromagnetic phase in the sample.We found that the FC M-T curve could not fit to M(T) = M(0) * [1 − bT 3/2 ], which indicated that our Cr 2 CoGa sample does not conform to the Bloch T 3/2 law.Instead, the M-T curve could be fitted empirically to M(T) = M(0) * [1 − (T/Tc) 2 ] 1/2 , as seen in the inset in Fig. 7.The T C for the Cr 2 CoGa film can be estimated to be 317 K which is quite lower than the predicted 1520 K 7 probably due to the off-stoichiometry of Cr 2 CoGa and the existence of phase separation and disorders.Moreover, the T 2 dependence of magnetization has been previously observed in half-metallic Heusler alloys and can be attributed to itinerant-like ferro(ferri)magnetism. 21,22he resistance as a function of temperature (R-T) is shown in Fig. 8. R-T curves indicated metallic behavior of the samples.A significant degree of lattice defects and phase separation were inferred from the large residual resistance (R 20K ) of the Cr 2 CoGa/Si(001) sample, with residual resistance ratio R 300K /R 20K = 1.56.For the Cr 2 CoGa/GaAs(001) sample, the residual resistance ratio R 300K /R 20K increased to 2.04 indicating a relatively small contribution to the resistance from lattice defects and other forms of atomic disorder.By assuming the functional R(T) = R 0 + cT n for our R-T data, a plot of ln(R-R 0 ) as a function of ln(T) was shown in the inset of Fig. 8.It demonstrated almost purely phononic linear dependence above 26K typifying the Bloch-Gruneisen formula and thus signified a dominance of phonon scattering in the sample. 21he results of the magnetoresistance (MR) measurements at different temperatures are displayed in Fig. 9.Note that the applied magnetic field was perpendicular to the film plane.The MR exhibits a negative dependence on the applied field at room temperature.The negative MR was due to the decreased scattering centers caused by an increase of magnetic domain with increasing magnetic field, which is also evidence for ferrimagnetic ordering in the Cr 2 CoGa film. 23Usually the negative MR originating from spin dependent scattering should enhance with lowering temperature, while in contrary, upon decreasing temperature below 150 K, MR turns to be positive dependence.It was probably due to the complicated state of the sample consisting of multi-phases at low temperature.The impurity scatterings became significant at low temperatures and thus reduce the negative MR effect.The positive MR is usually due to the dominant ordinary magnetoresistance (OMR) effect. 24The OMR is the increase in resistance with increased magnetic field due to bending of the electron trajectories by the Lorentz force, which is anisotropic and only exist when the magnetic field is not parallel to the current.In view of the complicated state of the sample, the positive MR might also be induced by the spin fluctuations or flip.Similar positive MR behavior was also reported in the spin gapless semiconductor of Mn 2 CoAl sample.Note that the field is applied perpendicular to film plane.

IV. CONCLUSIONS
In summary, we have prepared Cr 2 CoGa film on GaAs(001) using MBE technique.Significant phase separation was suppressed by proper control of growth conditions.It was demonstrated that Cr 2 CoGa Heusler phase, rather than Co 2 CrGa phase, constitutes the majority of the sample when grown on GaAs(001) at 450 o C. The measured spin moment of Cr 2 CoGa is small, in agreement with predicted HM-FCF nature; however, its T C is not as high as expected from the theoretical prediction probably due to the off-stoichiometry of Cr 2 CoGa and the existence of the disorders and phase separation.

FIG. 4 . 5 Feng
FIG. 4. Magnetic field dependence of the magnetization curves for the 70 o C-grown Cr 2 CoGa/GaAs(001) and 450 o C-grown Cr 2 CoGa/Si(001) films, respectively.Note that the diamagnetic signal of the substrate was not subtracted.

117223- 8 Feng
FIG. 9. Magnetoresistance vs applied field curves at different temperatures for 450 o C-grown Cr 2 CoGa/GaAs(001) film.Note that the field is applied perpendicular to film plane.