Structure and magnetism of MnGe thin films grown with a non-magnetic CrSi template

TL;DR

This work demonstrates a method to grow MnGe(111) thin films on Si(111) using a non-magnetic B20 CrSi template to probe intrinsic MnGe magnetism in the ultrathin limit. MnGe films 2–40 nm thick crystallize in the B20 structure with a rhombohedral distortion and exhibit a predominantly conical spin state below the ordering temperature, with a low-temperature remanent moment developing below about 35 K. Transport measurements reveal a complex Hall response, including a small topological-like contribution that is not unambiguously linked to a triple-$Q$ hedgehog phase; the observed low-temperature phase could be a multi-domain single-$Q$ conical state or a true triple-$Q$ texture. The results show that CrSi templates enable templated epitaxy of MnGe without magnetic interference and highlight finite-size effects that motivate direct scattering studies to resolve the exact magnetic textures in ultrathin MnGe films.

Abstract

We report on a novel method to grow B20 MnGe thin films which employs an ultrathin CrSi template layer on Si(111). This layer is expected to be non-magnetic, in contrast to MnSi and FeGe buffer layers that have been used previously, allowing an investigation of the intrinsic properties the MnGe in the ultrathin film limit without the influence of a neighboring magnetic layer. Single-phase MnGe(111) films were grown with thicknesses between 2 and 40 nm, which exhibited low interfacial roughnesses on the order of 0.6 nm. The films crystallized in a B20 structure with a small rhombohedral distortion. Magnetometry measurements in out-of-plane fields are consistent with a conical state. However, an unexpected remanent moment develops below 35 K, concomitant with features in the field dependence of the transport data. This provides indirect evidence for the presence of a low-temperature phase which has been identified by others as either a triple-Q topological spin-hedgehog lattice, or a multi-domain single-Q conical state.

Figures (11)

• Figure 1: (a) Unit cell of B20 MnGe. (b) Unit cell of rhombohedrally-distorted $R3$ MnGe shown for both in-plane tension ($\alpha>\ang{90}$) and compression ($\alpha<\ang{90}$). The arrows indicate the direction of the in-plane strain. (c)-(f) Illustration of the stacking of the MnGe structure, for the four possible morphologies. The structures (c) and (e) comprise a left-handed crystal whereas (d) and (f) are right-handed. Crystal structures were visualized using the vesta software Vesta.
• Figure 2: In-situ RHEED and ex-situ AFM images of (a)-(c) Si(111) substrate, (d)-(f) 2CrSi template (20nm Si overlayer), (g)-(i) 6.0nm CrSi film (20nm Si overlayer), and (j)-(l) an 7.2nm MnGe film (no Si overlayer) atop a 2CrSi template. The arrows in (e) indicate the fractional-order diffraction streaks visible in the $\surd 3 \times \surd 3$ pseudo-reconstruction of 2 CrSi with the beam along Si$[11\bar{2}]$. These are not present for the thicker CrSi film in (h). The RHEED patterns in each column share the same orientation relative to the substrate direction annotated in (a) and (b). All micrographs share the same scalebar.
• Figure 3: $\theta-2\theta$ XRD measurements of (a) the MnGe (111) peak for films grown atop 2 CrSi using the two-stage annealing procedure, and (b) the MnGe and CrSi (111) peak for a film grown atop a 6nm CrSi template using the same procedure. The inset in (b) shows phase purity of the B20 structure. Ticks in (c) denote the bulk peak positions for several crystal structures. Fits are performed using the model in Eq. \ref{['eq:xrd_peak']}.
• Figure 4: (a) $\theta-2\theta$ XRD measurements comparing the effect of temperature on the deposition of Mn and Ge atop 2 CrSi, leading to the formation of MnSi by diffusion above 250. (b) $\theta-2\theta$ XRD scans for Mn-rich films deposited on 2 CrSi. The as-grown film exhibits Mn5Ge3 impurities, which are removed when annealed to 250 after annealing in excess Ge. This impurity phase remains upon heating above 250, where diffusion-formed MnSi is still observed in coexistence with MnGe and (smoother) Mn5Ge3. The fit in (b) was performed using the model in Eq. \ref{['eq:xrd_peak']}. Ticks in (c) denote the bulk peak positions for several crystal structures.
• Figure 5: (a) HAADF-STEM image of a MnGe film depicting two chiral domains. (b) SADP calculated from the FFT of a lower magnification image of the domains in (a) encompassing the entire film thickness. (c) EDS elemental mapping of Mn, Ge, Cr, and Si. The substrate and capping layer are blue, and the MnGe film is shown in red. The orange layer at the Si-MnGe interface is the 2 CrSi template.