Search

In this dissertation, the growth, surface and interface dynamics of kagome Mn3Sn film isinvestigated in detail. The studies explore how the structural, orientation and interfacedynamics vary with temperature for bulk films. Molecular beam epitaxy is used to preparethe clean sample surfaces that are investigated in-situ using reflection high energy electrondiffraction to monitor the growth. A variety of ex-situ techniques are used to obtaininformation of the morphology, composition and crystallinity of the film. Additionally,theoretical calculations are carried out to support the experimental findings.The first study refers to the deposition of c-plane Mn3Sn on Al2O3 (0001) at 524 ± 5 C. The experimental findings indicates that the resulting film is predominantly c-planeoriented. The samples prepared in this way were found to be discontiguous, showing a3-dimensional morphology. According to first-principles calculations, Mn3Sn exhibits adisplaced Kagome structure in the very first stages of growth, for 2ML and 4ML growthon Al2O3 (0001). This result is then corroborated by calculating the surface formationenergies which shows that the Mn3Sn kagome structure is thermodynamically unstableduring the initial stages of growth with a separation of the Mn and Sn sub-lattices. Thisseparation helps to explain the observable RHEED pattern.With the aim of understanding of how the orientation would be affected bytemperature, the second study discusses the deposition of a-plane Mn3Sn on Al2O3 (0001)at 453 ± 5 C. The experimental analysis shows the in-plane lattice constants of cM =4.117 ± 0.027 Å and bM = 4.943 ± 0.033 Å, which is a very unexpected result indicatingthat the film is strained. Additionally, the deposition under these conditions appear to have4an overall 3D island morphology. Furthermore, in an effort to explain these results, twopossible orientation relationships between the film and the substrates are proposed, fromwhich we found that the Mn3Sn film is strained to the unconventional rectangle of theAl2O3 substrate. This finding is then corroborated by first principle calculations whichprovided atomic models of the Mn3Sn (11-20) lattice on Al2O3 (0001) high symmetry sitesin which the L3 and L3 R-90 degree models agree with the experimental findings. Additionally,an interface model of Mn3Sn on Al2O3 is proposed where it is shown that the strainobserved in the film is due to O-O spacing in the sapphire substrate.With the aim of improving the a-plane Mn3Sn film, a different growth protocol iscarried out and investigated at different temperatures. In this study, the Mn3Sn film isdeposited in a two step process where the first step involves depositing a thin seed layerwhich is allowed to crystallize and the second step involves growing the main layer of thedesired thickness. Our experimental findings show that the two step process with seed andmain layers deposited at room temperature leads to a flat and contiguous film consists ofdifferent rotated mis-oriented grains. Hence, to improve the crystallinity of the film, thetwo step process is further modified by adding an annealing step.The new growth method now involves annealing the seed layer at 350 ± 5 C for 35mins and later the main layer is grown at that elevated temperature. This annealed filmshows presence of islands & terraces indicating a crystalline layer. The roughness of thefilm is investigated in which it is seen the annealed case is 10× as rough as thenon-annealed sample, which is attributed to the crystalline layering. In both cases it wasobserved that the film experiences large in-plane compressive strain in the cM-axis. Thisexperimental strain is explained by understanding the mechanical properties using densityfunctional theory. From the theoretical calculations we observe that the cM axisexperiences a plastic deformation while the bM and aM axis experience compressive andtensile strain respectively. Furthermore, this strain is seen to modify the direction of theMn magnetic moment by decreasing its value to 2.51 μB.

Exchange bias, a shift in the center of the magnetic hysteresis loop of a magnetic material, originally discovered in 1956, has gained a lot of attention recently due to its applications in spintronics. This property has been observed in layered magnetic structures like antiferromagnetic /ferromagnetic bilayers. It is dependent on the interface interaction of the bilayer. Lots of work has been carried out on different systems such as Mn/CoFe and Fe/FeMn, and exchange bias has been reported to be present. The technological motivation for study of this property is because of its widespread usage in biosensors, hard disk read heads, and other applications. Scientifically, greater fundamental understanding of exchange bias is still needed. In this work, the L10 MnGa/ θ-MnN bilayer on MgO substrate was studied for the investigation of exchange bias motivated by recent reports of giant exchange bias using MnN as the antiferromagnet [10]. The bilayer was grown by using ultra-high vacuum molecular beam epitaxy. The growth was monitored using reflection high energy electron beam diffraction (RHEED). The sample was then taken for ex-situ measurements using various techniques such as x-ray diffraction (XRD), magnetic force microscopy, Rutherford backscattering (RBS), and either vibrating sample magnetometry (VSM) or SQUID magnetometry. Three different bilayers were grown with 15, 3 and 1 nanometer thick L10 MnGa layers but with constant MnN layer thickness. RHEED images showed spot patterns and a partial disorder within the crystalline films resulting in a bright background from diffuse scattering. The XRD analysis showed that the growth orientation of the MnGa was (001). The (001) peak was not so sharp indicating that the film is very thin and may not yet have a uniform layer spacing, which was consistent with the results from RHEED images. This may be due to large lattice mismatch resulting in a non-ideal or non-simple heteroepitaxy, and due to the films being very thin. Furthermore, the SQUID measurements showed the presence of a small but finite exchange bias in the case of the 3 nm L10 MnGa/θ-MnN sample in the in-plane direction only, and the amount of loop shift from the origin was estimated to be 300 Oersted. The results were compared with a recent publication for the case of CoFeB coupled with θ-MnN which showed giant exchange bias (3600 Oersted), and the similarities and differences are discussed. In the end, both the magnitude and anisotropy direction of the observed exchange bias are highly sensitive to the choice of ferromagnetic material; otherwise, it is dependent on the growth method details and/or the field cooling procedure.