Embedded magnetic phases in (Ga,Fe)N: Key role of growth temperature

The local chemistry, structure, and magnetism of (Ga,Fe)N nanocomposites grown by metal-organic vapor-phase epitaxy are studied by synchrotron x-ray diffraction and absorption, high-resolution transmission electron microscopy, and superconducting quantum interference device magnetometry as a function of the growth temperature $T_{\text{g}}$. Three contributions to the magnetization are identified: (i) paramagnetic—originating from dilute and noninteracting ${\text{Fe}}^{3+}$ ions substitutional of Ga and dominating in layers obtained at the lowest considered $T_{\text{g}}$ $\left(800°\text{C}\right)$; (ii) superparamagneticlike—brought about mainly by ferromagnetic nanocrystals of $\epsilon {\text{-Fe}}_3\text{N}$ but also by ${\gamma }^{\prime }{\text{-Fe}}_4\text{N}$ and by inclusions of elemental $\alpha \text{-Fe}$, and prevalent in films obtained in the intermediate $T_{\text{g}}$ range; (iii) component linear in the magnetic field and associated with antiferromagnetic interactions—found to originate from highly nitridated ${\text{Fe}}_x\text{N}$ $\left(x\le 2\right)$ phases, like $\zeta {\text{-Fe}}_2\text{N}$, and detected in samples deposited at the highest employed temperature, $T_{\text{g}}=950°\text{C}$. Furthermore, depending on $T_{\text{g}}$, the Fe-rich nanocrystals segregate toward the sample surface or occupy two-dimensional planes perpendicular to the growth direction.

Figure 1

(Color online) SXRD spectra for (Ga,Fe)N layers deposited at different growth temperatures. Inset: peak at $35.3°$ deconvoluted into two components assigned to diffraction maxima (111) of $\epsilon {\text{-Fe}}_3\text{N}$ and (110) of $\alpha \text{-Fe}$ [experiment (dotted line) and fit (smooth line)].

Figure 2

(Color online) SXRD spectra for a dilute (Ga,Fe)N sample (S690) as grown and upon in situ annealing at $T_{\text{a}}=900°\text{C}$ for 1 h, indicating that postgrowth annealing does not induce, in the SXRD sensitivity range, crystallographic decomposition. Inset: shift of the ${\text{Al}}_2{\text{O}}_3\left(006\right)$ vs $T_{\text{a}}$ in this work (open squares) and in Ref. 33 (full squares).

Figure 3

SXRD for a sample (S874) grown at $950°\text{C}$ evidencing the aggregation of (200) $\gamma \text{-Fe}$ in the (Ga,Fe)N layer.

Figure 4

(Color online) Average size vs $T_{\text{g}}$ of nanocrystals in the different ${\text{Fe}}_x\text{N}$ phases as determined from SXRD.

Figure 5

(Color online) Normalized XANES spectra (main plot) and the amplitude of the Fourier transforms (inset) of the $k^2$-weighted EXAFS in the range from 2.5 to $10.0{\text{Å}}^{-1}$ for three samples (points) grown at different temperatures with their relative fits (lines) summarized in Table .

Figure 6

Transmission electron micrographs of different ${\text{Fe}}_x\text{N}$ phases: (a) HRTEM image of a $\epsilon {\text{-Fe}}_3\text{N}$ nanocrystal and (b) the corresponding FFT image revealing that the $d$ spacing along the growth direction is about 0.216 nm. (c) HRTEM image on $\alpha \text{-Fe}$ nanocrystal in sample S687 and (d) SAED pattern on the enclosing area. (e) HRTEM image of a $\zeta {\text{-Fe}}_2\text{N}$ nanocrystal and (f) the corresponding FFT image revealing that the $d$ spacing along the growth direction is about 0.211 nm.

Figure 7

TEM images: distribution of the Fe-rich nanocrystals with increasing growth temperature. (a) $T_{\text{g}}=800°\text{C}$ (S690)—dilute (Ga,Fe)N; (b) $T_{\text{g}}=850°\text{C}$ (S687), Fe-rich nanocrystals concentrated in proximity of the samples interface solely; $T_{\text{g}}=950°\text{C}$ (S691)—Fe-rich nanocrystals segregating in proximity of the sample surface (c) and of the interface between the GaN buffer and the Fe-doped layer (d); [(e) and (f)] $T_{\text{g}}=950°\text{C}$ (S876)—to be compared with the TEM images of S691 in (c) and (d): reproducibility in the distribution of the Fe-rich nanocrystals for different samples grown at the same $T_{\text{g}}$.

Figure 8

A phase diagram of (Ga,Fe)N as a function of the growth temperature.

Figure 9

(Color online) (a) Magnetic field dependence of the nanocrystals magnetization $M_{\text{N}}$ at selected temperatures for sample S687 $\left(T_{\text{g}}=850°\text{C}\right)$. Each $M_{\text{N}}\left(H\right)$ curve has been measured from the maximum positive to the maximum negative field, only the dotted lines obtained by numerical invertion are guides for the eye. The dashed lines represent the saturation level of magnetization at each temperature. (b) Bullets, temperature dependence of the saturation magnetization $M_{\text{N}}^{\text{Sat}}$ obtained from panel (a). Dashed line, the Brillouin function for the magnetic moment of $2{\mu }_{\text{B}}$ per Fe atom. (c) Three major contributions to the total magnetic signal (bold solid) for sample S691 $\left(T_{\text{g}}=950°\text{C}\right)$: (i) paramagnetic from ${\text{Fe}}^{3+}$ (bold dashed), (ii) high-$T_{\text{C}}$ superparamagneticlike (thin solid) from the nanocrystals, and (iii) slowly saturating component (short dashed). Also here only a half of the full hysteresis loop was measured and the dotted lines obtained by numerical reflection are guides for the eye. (d) Bullets, temperature dependence of $M_{\text{N}}^{\text{Sat}}$ for sample S691. Dashed line, the Brillouin function for the magnetic moment of $2{\mu }_{\text{B}}$ per atom Fe. The dotted line follows the excess of $M_{\text{N}}^{\text{Sat}}$ over the contribution from high-$T_{\text{C}}$ ferromagnetic nanocrystals.

Figure 10

(Color online) Estimated lower limit for the Fe concentration that precipitates in the form of various Fe-rich nanocrystals $x_{{\text{Fe}}_{\text{N}}}$ as a function of the growth temperature.

Figure 11

(Color online) ZFC and FC curves measured in applied magnetic field of 200 Oe for samples grown at (a) $800°\text{C}$ and (b) $950°\text{C}$.