Challenges to magnetic doping of thin films of the Dirac semimetal ${\mathrm{Cd}}_3{\mathrm{As}}_2$

Magnetic doping of topological quantum materials provides an attractive route for studying the effects of time-reversal symmetry breaking. Thus motivated, we explore the introduction of the transition metal Mn into thin films of the Dirac semimetal ${\mathrm{Cd}}_3{\mathrm{As}}_2$ during growth by molecular beam epitaxy. Scanning transmission electron microscopy measurements show the formation of a Mn-rich phase at the top surface of Mn-doped ${\mathrm{Cd}}_3{\mathrm{As}}_2$ thin films grown using both uniform doping and delta doping. This suggests that Mn acts as a surfactant during epitaxial growth of ${\mathrm{Cd}}_3{\mathrm{As}}_2$, resulting in phase separation. Magnetometry measurements of such samples indicate a ferromagnetic phase with out-of-plane magnetic anisotropy. Electrical magneto-transport measurements of these films as a function of temperature, magnetic field, and chemical potential reveal a lower carrier density and higher electron mobility compared with pristine ${\mathrm{Cd}}_3{\mathrm{As}}_2$ films grown under similar conditions. This suggests that the surfactant effect might also serve to remove impurities from the bulk of the film. We observe robust quantum transport (Shubnikov-de Haas oscillations and an incipient integer quantum Hall effect) in very thin (7 nm) ${\mathrm{Cd}}_3{\mathrm{As}}_2$ films despite being in direct contact with a structurally disordered surface ferromagnetic overlayer.

Figure 1

(a) In situ RHEED patterns during MBE of a nominally 15-nm-thick Mn-doped ${\mathrm{Cd}}_3{\mathrm{As}}_2$ heterostructure grown using the uniform doping approach. The electron beam is directed along $\left[0\overline{1}1\right]$ (left) and $\left[\overline{2}11\right]$ (right) direction. (b) Ex situ AFM images of the same sample as in panel (a). (c) Out-of-plane XRD of a nominally 25-nm-thick Mn-doped ${\mathrm{Cd}}_3{\mathrm{As}}_2$ heterostructure grown using the uniform Mn-doping approach. The “extra peaks” in the XRD indicate the presence of a Mn-rich phase of unknown composition.

Figure 2

(a) Low-magnification cross-sectional HAADF-STEM image of a Mn-rich phase/${\mathrm{Cd}}_3{\mathrm{As}}_2$ heterostructure that results from uniform Mn-doping during MBE growth of ${\mathrm{Cd}}_3{\mathrm{As}}_2$. Composite and individual EDX maps show the spatial distribution of Mn, Cd, and Sb in the heterostructure as well as an atomic percent concentration profile across the heterostructure. Instead of being incorporated into the ${\mathrm{Cd}}_3{\mathrm{As}}_2$ lattice, a segregated Mn-rich phase formed on the top of the ${\mathrm{Cd}}_3{\mathrm{As}}_2$ layer. (b) Schematic of the intended sample using the uniform doping method. (c) Schematic of the actual Mn-rich phase/${\mathrm{Cd}}_3{\mathrm{As}}_2$ heterostructure produced by the uniform doping growth.

Figure 3

Investigation of ferromagnetism in a Mn-rich layer/${\mathrm{Cd}}_3{\mathrm{As}}_2$ heterostructure. (a)–(c) SQUID magnetometry measurements of a 1 nm Mn-rich phase/7 nm ${\mathrm{Cd}}_3{\mathrm{As}}_2$ heterostructure at 10 K, respectively. The data suggest ferromagnetic ordering with an out-of-plane magnetic anisotropy. The inconsistency between the saturated magnetization values in the easy- and hard-axis measurements is an artifact created by the difficulty in properly handling the large filling factor when the thin film is mounted for field in plane measurements in the SQUID. (d) Observation of Shubnikov de Haas oscillations and an incipient integer quantum Hall effect in a 1 nm Mn-rich phase/7 nm ${\mathrm{Cd}}_3{\mathrm{As}}_2$ heterostructure at 2 K. (e) An expanded view of the low-field MR and Hall effect in panel (d).

Figure 4

Shubnikov-de Haas quantum oscillations in a 1 nm Mn-rich phase/7 nm ${\mathrm{Cd}}_3{\mathrm{As}}_2$ heterostructure. (a), (b) The temperature dependence of the quantum oscillations and Hall effect, respectively. (c) Temperature dependence of the carrier density and Drude mobility. The values of the carrier density and Drude mobility for a pristine ${\mathrm{Cd}}_3{\mathrm{As}}_2$ film of the same thickness are also shown at 2 K for comparison. (d) Temperature dependence of the quantum oscillation amplitude at different Landau fan index ($m=3,3.5,4,4.5$), with integer and half-integer values of $m$ corresponding to MR peaks and valleys, respectively. Figure S4 in the Supplemental Material [35] shows the oscillations in more detail. The dashed lines are fits using the well-known Lifshitz-Kosevich formula. (e), (f) Gate-voltage dependence of the quantum oscillations and Hall effect, respectively.