XMCD study of magnetism and valence state in iron-substituted strontium titanate

Room-temperature ferromagnetism was characterized for thin films of $\mathrm{SrT}{\mathrm{i}}_{0.6}\mathrm{F}{\mathrm{e}}_{0.4}{\mathrm{O}}_{3\text{−}\delta }$ grown by pulsed laser deposition on $\mathrm{SrTi}{\mathrm{O}}_3$ and Si substrates under different oxygen pressures and after annealing under oxygen and vacuum conditions. X-ray magnetic circular dichroism demonstrated that the magnetization originated from $\mathrm{F}{\mathrm{e}}^{2+}$ cations, whereas $\mathrm{F}{\mathrm{e}}^{3+}$ and $\mathrm{T}{\mathrm{i}}^{4+}$ did not contribute. Films with the highest magnetic moment ($0.8{\mu }_{\mathrm{B}}$ per Fe) had the highest measured $\mathrm{F}{\mathrm{e}}^{2+}$:$\mathrm{F}{\mathrm{e}}^{3+}$ ratio of 0.1 corresponding to the largest concentration of oxygen vacancies $(\delta =0.19)$. Postgrowth annealing treatments under oxidizing and reducing conditions demonstrated quenching and partial recovery of magnetism respectively, and a change in Fe valence states. The study elucidates the microscopic origin of magnetism in highly Fe-substituted $\mathrm{SrT}{\mathrm{i}}_{1-x}\mathrm{F}{\mathrm{e}}_x{\mathrm{O}}_{3\text{−}\delta }$ perovskite oxides and demonstrates that the magnetic moment, which correlates with the relative content of $\mathrm{F}{\mathrm{e}}^{2+}$ and $\mathrm{F}{\mathrm{e}}^{3+}$, can be controlled via the oxygen content, either during growth or by postgrowth annealing.

Figure 1

(a) Model of Fe-substituted $\mathrm{SrTi}{\mathrm{O}}_3$ demonstrating cubic perovskite structure, oxygen octahedra surrounding the B sites, and cation substitution of Ti with Fe. (b)–(e) XRD $\omega \text{−}2\theta$ scans for $\mathrm{SrT}{\mathrm{i}}_{60}\mathrm{F}{\mathrm{e}}_{40}{\mathrm{O}}_{3\text{−}\delta }$ samples on (b) Si and (c)–(e) $\mathrm{SrTi}{\mathrm{O}}_3$ substrates, subjected to different base and growth pressures (b)–(d), or annealing treatments (e). XRD $\omega \text{−}2\theta$ scans at ${40}^{\circ }<2\theta <{50}^{\circ }$ (d) are left of their corresponding wide-range scans (c). (f) TEM image of $1-1.2\mu \mathrm{Torr}$ (STO) with Fe elemental map showing homogeneous distribution of Fe.

Figure 2

(a), (b) Out-of-plane (OP) and in-plane (IP) VSM hysteresis curves of $\mathrm{SrT}{\mathrm{i}}_{60}\mathrm{F}{\mathrm{e}}_{40}{\mathrm{O}}_{3\text{−}\delta }$ thin films grown at 1 μTorr base pressure, $1.2\mu \mathrm{Torr}$ growth pressure on (a) Si substrate and (b) $\mathrm{SrTi}{\mathrm{O}}_3$ substrate. (c) XMCD hysteresis curves for the same two films collected at the energy of the maximum Fe signal, shown with an arrow in (d). (d) Fe-$L$ edge and (e) Ti-$L$ edge XAS and XMCD spectra of $\mathrm{SrT}{\mathrm{i}}_{0.60}\mathrm{F}{\mathrm{e}}_{0.40}{\mathrm{O}}_{3\text{−}\delta }$ thin film grown at 1 μTorr base pressure, 1.2-μTorr growth pressure on $\mathrm{SrTi}{\mathrm{O}}_3$. All XAS and XMCD data shown here are TFY scans.

Figure 3

(a), (b) Out-of-plane VSM hysteresis curves of $\mathrm{SrT}{\mathrm{i}}_{60}\mathrm{F}{\mathrm{e}}_{40}{\mathrm{O}}_{3\text{−}\delta }$ thin films grown at different base and growth pressures on Si (a) and $\mathrm{SrTi}{\mathrm{O}}_3$ (b). (c) TFY Fe-$L$ edge XAS and XMCD of $\mathrm{SrT}{\mathrm{i}}_{0.60}\mathrm{F}{\mathrm{e}}_{0.40}{\mathrm{O}}_{3\text{−}\delta }$ thin films grown on Si at different base and growth pressures. (d) The relation between the Fe valence states determined from XAS and the net magnetic moment measured by VSM. A representative error bar is shown.

Figure 4

(a) XRD $\omega \text{−}2\theta$ scans at ${40}^{\circ }<2\theta <{50}^{\circ }$, (b) out-of-plane and (c) in-plane VSM hysteresis curves for the $\mathrm{SrT}{\mathrm{i}}_{65}\mathrm{F}{\mathrm{e}}_{35}{\mathrm{O}}_{3\text{−}\delta }$ thin film subjected to various annealing treatments. (d) XMCD hysteresis curves collected on Fe signal and (e) XAS and XMCD for the same film subjected to various annealing treatments.