Magnetic properties of the layered III-VI diluted magnetic semiconductor Ga_1−xFe_xTe

Gallium selenide (GaSe) has emerged as one of the best non-linear optical materials.^1–11 GaSe exhibits a rich collection of important properties for THz generation and detection applications including a large non-linear coefficient, high damage threshold, high-temperature operation, and a wide transparency range.^1–4 GaSe is highly efficient as a broadband tunable source up to 40 THz and as a sensor up to 100 THz.^4,5 In short, GaSe (and related doped and un-doped III-VI semiconductors such as GaTe) have emerged as valuable systems in the very active THz research field.^{1–3,5,6,8–11} Work on GaTe and doped GaTe crystals includes the growth and characterization of these systems for broadband tunable THz sources and sensors.⁵ 

Doping with Te,^3,5 S, Cr,^2,6 Ag,¹ and Er⁷ was found to strengthen GaSe. Doping with In was found to significantly enhance the physical properties and strengthen the crystals enough to allow optical surfaces to be cut and polished along additional directions. Unexpectedly, the In doping did not degrade the useful properties of GaSe but actually resulted in enhanced non-linear optical properties.^{2,3,5,6,8–11} First-principles theoretical studies investigating the mechanical and electrical properties Te-based systems were also conducted.³ 

Unlike doping with In, the incorporation of a transition metal element raises intriguing possibilities for coupling the magnetic properties of the transition metal ion with the host III-VI semiconductor leading to optical or electrical transport effects. Specifically, the sp-d exchange coupling in related materials can sometimes have dramatic physical consequences such as giant Faraday rotation, bound magnetic polarons, or induced metal to insulator transitions.^12,13 Doping with the transition metal ion Cr in Ga_1−xCr_xSe^2,6 has already been studied and shows improved mechanical properties while preserving the advantages of the host GaSe as a THz material.^2,6

In this work we present magnetic and x-ray measurements on single-crystalline Ga_1−xFe_xTe. This work complements work done on the other III-VI DMS systems investigated to date (Ga_1−xMn_xS,^14–18 In_1−xMn_xS,^19,20 In_1−xMn_xSe,^21–23 Ga_1−xMn_xSe,²⁴ and Ga_1−xFe_xSe).²⁵ 

A 0.0288 g single-crystalline Ga_1−xFe_xTe sample was taken from a boule grown using the vertical Bridgman method with a nominal concentration of x = 0.05. Magnetization measurements were performed using a Quantum Design MPMS XL7 superconducting quantum interference device (SQUID) magnetometer at temperatures between 1.8 and 400 K in fields up to 7 T. A pure GaTe crystal was measured to determine the value of the diamagnetic signal (−3.5 × 10⁻⁷ emu/g G) due to the semiconductor host GaTe, which was subtracted from the data.

Structural analysis was performed on a small crystalline specimen taken from the same boule as the magnetization crystal. The crystalline fragment was mounted on a loop holder, suitable for single crystal x-ray analysis, using Paratone-N oil. It was then transferred to the goniostat where it was cooled to 100 K for characterization and data collection. The data were collected from a combination of phi and omega scans on a Bruker D8 Venture diffractometer equipped with an Iμ S Mo Kα source and a CMOS detector.

Magnetization versus temperature measurements taken on a single crystalline Ga_1−xFe_xTe sample with a magnetic field oriented perpendicular to the two-dimensional GaTe layers are shown in Figs. 1(a) and 1(b). The magnetization remains roughly constant and featureless for the entire temperature range from 1.8 to 400 K for measurements taken in both 0.1000 and 0.0500 T. In 7 T, the magnetization increases by only 20% as the temperature is decreased from 400 to 1.85 K. In contrast, the sharp cusp shown in Fig. 1(c) is due to the long-range collective interaction of spins resulting in the spin-glass^17,18 transition at 11.2 K in Ga_1−xMn_xS. At higher temperatures, the broad thermal hysteresis^21–23 shown in Fig. 1(d) is likely due in part to a long-range collaborative spin-density-wave in In_1−xMn_xSe. No evidence for comparable long-range interactions of spins is observed in Ga_1−xFe_xTe (Fig. 1(a)) for the entire temperature range. The long-range interaction of spins in Ga_1−xFe_xTe apparently does not play the same prominent role in Ga_1−xFe_xTe that the transition metal ions play in related III-VI DMS systems. Nonetheless, it is important to note that the magnetization in Ga_1−xFe_xTe remains high even as the temperature increases well above room temperature. This is reminiscent of Van-Vleck paramagnetism.

To explore the possibility of a ferromagnetic minority phase, magnetization hysteresis loops were taken on Ga_1−xFe_xTe at 5 and 300 K (Figs. 2(a) and 2(b)). The magnetization increases rapidly at low field, reaching over 50% of the 7 T value in a field of only 0.2 T and over 90% by 1 T at both 5 and 300 K. Since, there is no apparent coercive field on this scale, the region near zero field was examined further and is shown in Figs. 3(a) and 3(b). In this region the coercive field was measured to be less than 0.0005 T at 5 and 300 K as well as 1.8 and 10 K (not shown), which is within the experimental uncertainty of the instrument.

Having ruled out multi-domain ferromagnetic regions, single domain super-paramagnetic regions might account for the rapid rise observed in the magnetization at low fields that is clearly seen in Figs. 2(a) and 2(b). If this data is plotted together, the magnetization at 300 K should increase significantly slower than at 5 K. However, the magnetization at both 300 and 5 K rise nearly identically as can be seen in Fig. 4(a). This is true when the two-dimensional Ga_1−xFe_xTe layers are aligned perpendicular to the field (up and down triangles) or parallel to the field (circles and squares). This suggests that we have a good quality crystal with neither multi-domain ferromagnetic regions nor single domain super-paramagnetic clusters present in our sample.

To further investigate the crystal quality of our sample, we collected single crystal x-ray diffraction data on a small fragment taken from the same boule as the magnetization crystal. An initial survey of the reciprocal space revealed a set of reflections with a monoclinic lattice, with unit cell dimensions a = 23.20 Å, b = 5.00 Å, and c = 6.36 Å, with α = 90°, β = 92.3°, and γ = 90°. These result in a unit cell volume of 737 ų. Our findings are in agreement with earlier work.^26,27 Upon analysis of a full data set, additional symmetry elements became apparent, resulting in a rhombohedral space group bearing hexagonal axes, namely R3c. The unit cell dimensions were a = 5.01 Å, b = 5.01 Å, and c = 17.02 Å, with α = 90°, β = 90°, and γ = 120°, and a unit cell volume of 369 ų. This is in agreement with a recent paper by Borisenko and coworkers,²⁸ where the authors found a hexagonal space group from single-crystal studies of pure GaTe.

A significant magnetocrystalline anisotropy is evident in the magnetization of our Ga_1−xFe_xTe (x = 0.05) crystal between 0 and 2 T at both 10 and 300 K. The easy axis is observed along a random direction parallel to the two-dimensional Ga_1−xFe_xTe layers. These layers have one-third of the Ga-Ga direct bonds aligned in the plane of the layers and two-thirds aligned perpendicular to the layers. The hard axis is along the direction perpendicular to the layers or along the direction with the majority of the direct Ga-Ga bonds.

Similar magnetocrystalline anisotropy was observed in Ga_1−xFe_xSe and is shown in Fig. 4(b).²⁵ Additionally, preliminary calculations were conducted for Ga_1−xFe_xSe using a model Hamiltonian including crystal-field, spin-orbit, spin-spin, and Zeeman terms. The model Hamiltonian used for Ga_1−xFe_xSe would be similar to a corresponding model Hamiltonian for Ga_1−xFe_xTe. The Ga_1−xFe_xTe and Ga_1−xFe_xSe data as well as the theoretical calculations have the same easy and hard axis, where the hard axis is along the direction perpendicular to the layers and along the prominent direction of the direct Ga-Ga bonds. The easy axis in all cases is parallel to the two-dimensional layers. Although good qualitative agreement is made between theory and experiment, further work is needed to refine the parameter values (e.g. spin-orbit coupling constant, bond lengths, etc.) used in the theory before quantitative agreement is achieved.

Magnetic properties of single crystalline Ga_1−xFe_xTe (x = 0.05) have been measured. The magnetization versus field for an x = 0.05 sample reaches a maximum of 0.68 emu/g at 2 K in 7 T. The magnetization of Ga_1−xFe_xTe saturates rapidly even at room temperature where the magnetization reaches 50% of saturation in a field of only 0.2 T. In 0.1 T at temperatures between 50 and 400 K, the magnetization drops to a roughly constant 0.22 emu/g. In 0 T, the magnetization drops to zero with no hysteresis present. The data is consistent with Van-Vleck paramagnetism combined with a pronounced crystalline anisotropy. Single crystal x-ray diffraction data yield a rhombohedral space group bearing hexagonal axes, namely R3c. The unit cell dimensions were a = 5.01 Å, b = 5.01 Å, and c = 17.02 Å, with α = 90°, β = 90°, and γ = 120° giving a unit cell volume of 369 ų. The x-ray data corroborates our observations from the magnetism studies and affirms that the crystal lattice of Ga_1−xFe_xSe is nearly the same as GaTe. As such we believe that Fe substitutes Ga in the lattice, as predicted by theory, and consistent with the magnetization studies presented.

This research was supported by the UNF Terry Presidential Professorship, by the Florida Space Grant Consortium/NASA, a Cottrell College Science Award by the Research Corporation for Science Advancement, a Purdue Univ. Academic Reinvestment Program, and by the National Science Foundation (NSF) Grants Nos. DMR-07-06593 and DMR-14-29428.