Figures 1(a)~(e)-4(a)~(e) show the room-temperature transmission Mӧssbauer spectra of the xGd₂O₃-(1-x)α-Fe₂O₃ nanoparticles system for molar concentration x

= 0.1, 0.3, 0.5, 0.7 and ball milling times (BMT) of 0, 2, 4, 8 and 12 hours, respectively. The refined values of the hyperfine parameters extracted from these spectra by nonlinear least-squares fitting are listed in Table 1. The spectra in Figures 1(a)-4(a) were analyzed considering one six-line pattern, with the magnetic hyperfine field (BHF) characteristic to hematite. A small quadrupole-split doublet was added to these fits to account for the presence of ultrafine hematite nanoparticles, with the dimension less than 7 nm.

A second sextet was introduced in the analysis of the Mӧssbauer spectra in Figures 1(b)~(e)-4(b)~(e) with the magnetic hyperfine field smaller than that of hematite and assigned to gadolinium-doped hematite nanoparticles. This sextet originates from substitutions of Fe atoms by Gd atoms and is consistent with the model of local atomic environment. Moreover, a third sextet was necessary in the Mӧssbauer spectrum of Figure 1(c) demonstrating an increased level of the substitutions in the sample milled for 4 h. The dependence of the BHF values of these sextets on the ball milling time is depicted in Figure 5 and indicates the formation of a limited solid solution in the studied samples. A higher level of

substitution is not possible due to the difference in the ionic radii of the Gd³⁺ ions (0.938 Å) and Fe³⁺ ions (0.645 Å). It may be noted that the formation of a solid solution was not observed in the xNd₂O₃-(1-x)α-Fe₂O₃ system reported previously, due to the specific chemistry of the Nd ions [26].

The Mӧssbauer spectra in Figure 1(d), Figure 1(e) and Figures 2(b)~(e)-4(b)~(e) were analyzed by considering an additional quadrupole-split doublet, with the hyperfine parameters (quadrupole splitting QS and isomer shift, IS, relative to Fe) characteristic to the gadolinium perovskite, GdFeO₃. Figure 6 displays the dependence of the relative abundance of this doublet as function of the ball milling time for all values of the molarities used. It can be seen that the contribution of this doublet increases with increasing milling time and reaches 41.64% for the sample with x = 0.7 ball-milled for 12 hours. The occurrence of the gadolinium perovskite can be understood by considering the reaction Gd₂O₃ + Fe₂O₃ à 2GdFeO₃, which is believed to occur during the mechanochemical activation process.

Figures 7(a)-(e) shows the XRPD patterns of the xGd₂O₃-(1-x)α-Fe₂O₃ nanoparticles system for x = 0.5 and ball milling times of 0, 2, 4, 8 and 12 hours, respectively. The pattern in Figure 7(a) is consistent with the reflections from α-Fe₂O₃ (PDF card 01-089-0597) and Gd₂O₃ (PDF card 03-065-3181), which represent the starting material. It can be observed that after milling, the reflections indicative of the gadolinium perovskite appear (PDF card 01-074-1900). The lines of GdFeO₃ dominate the XRPD pattern after 12 h of exposure to mechanochemical activation. These results are in good, qualitative agreement with the Mӧssbauer findings presented previously.

Figure 8(a) shows the hysteresis curve for the gadolinium oxide-hematite equimolar composition (x = 0.5) at 0 h of milling, recorded at 300 K with a 5 T applied magnetic field. The plot is consistent with the presence of a paramagnetic component specific to Gd₂O₃ (known as a perfect paramagnet) and a weak

ferromagnetic component, due to hematite. The magnetic state of hematite is antiferromagnetic with the Neel temperature T_N = 960 K [1] - [5]. Specific to hematite is the fact that its two magnetic sublattices have equal moment and antiparallel orientation below the Morin temperature, T_M = 262 K, for pure hematite [7]. At temperatures higher than the Morin temperature, the two sublattices are slightly canted, leading to weak ferromagnetism [27]. This phenomenon explains the presence of the coercive field of H_c = 158 Oe.

Figure 8(b) displays the hysteresis loop of the 0-h equimolar sample recorded at 5 K and 5 T. It can be seen that the magnetization varies linearly with the applied magnetic field, dependence due to the paramagnetic behavior of gadolinium oxide and the antiferromagnetic behavior of hematite. The order-of-magnitude increase in the magnetization at 5 K compared to the magnetization at 300 K can be explained by the temperature dependence of magnetization given in Figure 8(c). The magnetization evolution with temperature was measured via the

zero-field-cooling-field cooling (ZFC-FC) procedure: first, the sample was cooled in zero field, then a 200 Oe magnetic field was applied, the ZFC dependence on heating was recorded and next, the temperature dependence of the magnetization on the FC was measured. The thermal hysteresis between 145 and 215 K given by the ZFC-FC curves reveals a first-order thermodynamic transition, which is specific to the Morin transition of hematite. In our case, this transition occurs at lower temperatures, due to the smaller grain sizes and the equimolar mixture of hematite and gadolinium oxide.

It can be seen in Figure 8(d) for the 4-h sample that the coercive field evidenced at 300 K significantly decreased, while the magnetization increased, without reaching saturation in a field of 5 T. Further increase in the magnetization

can be observed in the hysteresis loop recorded at 5 K and depicted in Figure 8(e). As can be inferred from Figure 8(f), the thermal hysteresis specific to the Morin transition is drastically affected after the 4-h milling time. Thus, the Morin transition occurs over a broad temperature range and the antiferromagnetic coupling is replaced by the weak ferromagnetism down to 20 K. This is due to the decrease in the dimension of the nanoparticles formed.

The hysteresis curves recorded for the sample milled for 12 h reveal an increase in magnetization and coercivity both at 300 K (Figure 8(g)) and 5 K (Figure 8(h)). As a result of the ball milling performed, Gd substitutions for Fe in the hematite lattice add increasingly higher contributions to magnetization. The ZFC-FC dependence specific to the 12-h milled sample (Figure 8(i)) shows a small Morin transition, between 60 and 100 K, result which is consistent with the consumption of hematite in the mechanochemical activation reaction and the formation of the gadolinium perovskite phase. In GdFeO₃, Gd-Gd, Gd-Fe and Fe-Fe interactions result in its antiferromagnetic behavior along with a weak ferromagnetic behavior [22] [23]. Consequently, in the equimolar mixture of Gd₂O₃-Fe₂O₃ we deal with a superposition of magnetic behaviors, with multiple magnetic phase transitions for both intrinsic causes and changes induced by mechanical grinding.

Figure 9(a) shows the optical absorption spectra of the gadolinium oxide-hematite equimolar mixture as function of energy over the UV-VIS-NIR spectral range for all milling times employed. For the starting material, hematite has a band gap of 1.9 - 2.2 eV in the visible region, while gadolinium oxide has a gap value of 5.2 eV in the ultraviolet range. For the ball-milled material, it can be seen that the absorbance is considerably enhanced and broadened, an effect we believe to result from the substitution of Gd ions for Fe.

Since the 2.2 eV transition in hematite is indirect, a plot of (αE)² as function of energy (eV) is able to yield the band gap of the compound, according to the Tauc plots [28] [29] [30]. Indeed, it can be observed in Figure 9(b)) that the intercepts give a value of ~2.1 eV for the band gap; moreover, this value is independent of

the milling time employed. These results show that the gadolinium oxide-hematite mixed-oxide nanostructures have semiconductor properties.