Structural, Optical, and Magnetic Properties of Erbium-Substituted Yttrium Iron Garnets.

We synthesized a series of slightly erbium-substituted yttrium iron garnets (Er:YIG), Y3-x Er x Fe5O12 at different Er concentrations (x = 0, 0.01, 0.05, 0.10, and 0.20) using a solid-state reaction and investigated their structural, magnetic, and optical properties as a function of Er concentration. The volume of the unit cell slightly increased with Er concentration and Er atoms predominately replaced Y atoms in the dodecahedrons of YIG. The optical properties exhibited certain decreases in reflectance in the 1500-1600 nm wavelength range due to the presence of Er3+. Despite the many unpaired 4f electrons in Er3+, the total magnetic moments of Er:YIG showed similar trends with temperatures and magnetic fields above 30 K. An X-ray magnetic circular dichroism study confirmed the robust Fe 3d magnetic moments. However, the magnetic moments suddenly decreased to below 30 K with Er substitution, and the residual magnetism (M R) and coercive field (H C) in the magnetic hysteresis loops decreased to below 30 K with Er substitution. This implies that Er substitution in YIG has a negligible effect on magnetic properties over a wide temperature range except below 30 K where the Er 4f spins are coupled antiparallel to the majority Fe 3d spins. Our studies demonstrated that above 30 K the magnetic properties of YIG are retained even with Er substitution, which is evidence that the Er doping scheme is applicable for YIG-based magneto-optical devices in the mid-infrared regime.

Introduction

Magnetic garnets have been regarded as key materials for quantum computing and communication systems because of their intriguing quantum phenomena, such as hybridized magnon–phonon excitation,[1] the conversion between magnons and phonons,[2] spin pumping with elastic waves,[3] and the spin Seebeck effect,[4] which are based on a strong magnon and phonon interactions. Compared with the magnon–phonon interaction in dilute spin ensemble systems such as rare-earth ion-doped crystals[5] or nitrogen-vacancy centers in diamonds,[6] insulating ferrimagnet yttrium iron garnets (Y3Fe5O12, YIG) have been widely used for various magneto-optical applications. Many researchers have made efforts to achieve a high spin density and a low damping rate, which are important factors for magnetic spintronic devices, as well as high-frequency microwave devices.[7−10]

It has been reported that transition-metal-substituted YIG exhibits a giant Faraday rotation or a magneto-optical Kerr effect, indicating that the interaction between magnetic and optical properties can be modulated by substitution or doping in the YIG system.[11−13] Rare-earth atoms substituted garnets have attracted much attention for optical applications because of their strong fluorescence emission in the mid-infrared regime.[14] Er-substituted crystals exhibit a high radiative decay rate and a long optical coherent time, as well as mid-infrared emissions at a wavelength of 1.5 μm, which is the C-band for telecommunications.[15] In recent years, the magnetic, optical, and structural properties of rare earth or transition-metal-substituted garnets have been widely investigated. Ce-substituted YIG has enhanced magnetic moments because of the preferred-(100) orientation.[16] The coercive fields of magnetic garnets have been studied by changing the grain sizes.[17,18] However, many garnets with substitutional atoms have shown degraded magnetic and optical properties due to changes of sample properties such as crystallinity, shape, size, and impurities.[19,20] Therefore, the search for appropriate substitutional atoms to enhance the magnetic properties as well as optical properties is highly demanded.

Here, we synthesized a series of Er-substituted YIG, which had the chemical compositions of Y3–ErFe5O12 at different Er concentrations (x = 0, 0.01, 0.05, 0.10, and 0.20) using a solid-state reaction method, and studied their changed magnetic and optical properties with small concentrations of Er atoms. Fe L2,3-XAS and XMCD spectra showed that the Er atoms seemed to replace Y atoms in the dodecahedrons of YIG. We found that the optical properties were almost invariant under Er substitution except for small changes in reflectance in the range of 1500–1600 nm wavelengths because of the distinct optical transition lines of the Er3+ ions. The magnetic properties of YIG were retained in a moderate temperature range, 30–300 K, however, below 30 K the magnetic moment and coercive force decreased as functions of the Er concentration, indicating an antiparallel spin alignment of the Er 4f spins at Y sites below that temperature. Our study showed that the magnetic properties of YIG could be retained in a broad temperature range of 30–300 K with characteristic optical transition lines in a range relevant to optical telecommunication systems when Er atoms were substituted in YIG at small concentrations.

Experimental Procedures

II.1. Sample Synthesis

Powders of Y3–ErFe5O12 were synthesized by a solid-state reaction with various Er contents (x). The stoichiometric mixtures of high purity powders, Y2O3 (99.99%, Sigma-Aldrich), Fe2O3 (99.995%, Sigma-Aldrich), and Er2O3 (99.9%, Sigma-Aldrich), were well ground using an agate mortar. These mixtures were calcined at 850 °C for 12 h, then cooled down to room temperature in air. This mixture was reheated three times to a high temperature of 1300 °C with intermediate grinding.

II.2. Sample Characteristics

We studied the crystal structures of the Er-substituted YIG using X-ray powder diffraction (Expert pro-MPD, PANalytical) at the Western Seoul Center of Korea Basic Science Institute (KBSI) using Cu Kα radiation (λ = 0.15406 nm) in the 2θ range of 10–120° at room temperature. The actual Er concentration in Er-substituted YIG was also studied using an inductively coupled plasma optical emission spectrometer (ICP-OES) (OPTIMA 5110, Agilent) at the Yonsei Center for Research Facilities (YCRF). Differential scanning calorimetry (DSC250, TA Instruments) at Ewha Womans University was used to estimate the Curie temperatures of the Er-substituted YIG in a temperature range of 300–700 K. A SQUID Magnetometer (Quantum Design MPMS3, Quantum Design North America) at the KBSI was used to observe the magnetic properties of Er-substituted YIG in a temperature range of 2–300 K from −0.1 to 0.1 T. The optical properties of the Er-substituted YIG were determined by Fourier-transform infrared spectroscopy (Nicolet iS50, Thermo Fisher Scientific Instrument) and a UV–VIS/NIR spectrophotometer (Lambda 1050, Perkin Elmer) at the KAIST Analysis Center for Research Advancement (KARA) within a wavelength range of 400–7000 cm–1.

X-ray absorption spectroscopy (XAS) at the Fe L2,3-edge of the powder samples was conducted at 2A beamline in the Pohang Light Source (PLS) at room temperature using circularly polarized X-rays. The absorption coefficient was collected as the total electron yield as a function of increasing photon energy. An external magnetic field of 0.7 T was applied during the measurement parallel and antiparallel to the normal direction of the specimen, and the difference in the absorption intensities (so-called X-ray magnetic circular dichroism; XMCD) was monitored to examine the long-range magnetic moments of Fe 3d spins. Hard XAS at the Er L3-edge was conducted at the 8C beamline at the PLS. The extended X-ray absorption fine structure (EXAFS) analysis was processed using the UWXAFS package.[21]

Results and Discussion

Figure shows a schematic of Er-substituted Y3Fe5O12 (Er:YIG) with a cubic structure (space group -3). One unit cell contains Y24Fe40O96 as shown in the inset of Figure . Since Y atoms with a trivalent state (Y3+) have no magnetic moment, the net magnetic moment of YIG is determined by two different Fe3+ atoms, which align antiferromagnetically along the (110) crystal direction of the YIG. For convenience, the Fe atoms in the octahedral and tetrahedral coordination are designated as FeO and FeT, respectively. FeO and FeT have opposite spin directions, resulting in an expected magnetic moment of 5 μB/YIG at 0 K, which is consistent with the experimental value of 4.73 μB/YIG.[22] Since Er atoms have large magnetic moments and strong emission at around 1550 nm, the magnetic and optical properties of YIG might be modified by Er substitution. We synthesized polycrystalline Er-substituted YIG powders with chemical compositions of Y3–ErFe5O12 (x = 0, 0.05, 0.10, 0.15, and 0.20) and characterized their magnetic and optical properties as functions of the Er concentration. To prevent the unexpected formation of secondary phases, we investigated Er-substituted YIG with only a low Er concentration (x ≤ 0.20). The actual Er concentrations were determined by ICP-OES measurements and they well matched the nominal Er concentrations of Er:YIG (see Table ).

Figure 1

Crystal structure of YIG. One unit cell of YIG contains 24 dodecahedrons of Y–O, 24 octahedrons of FeO–O, and 16 tetrahedrons of FeT–O connected to each other by shared edges and corners.

Table 1

Nominal and Measured Er Concentrations (x) of Er:YIG

nominal x	measured x
x = 0	x = 0.0001
x = 0.01	x = 0.0088
x = 0.05	x = 0.0498
x = 0.10	x = 0.1003

To obtain a solid solution of Er-substituted YIG (referred to as Er:YIG, hereafter) with a single phase, Er:YIG was sintered three times at a high temperature above 1300 °C with subsequent inter-grinding, and their crystallographic properties were examined using X-ray diffraction, as shown in Figure . The as-synthesized Er:YIG powder had a cubic structure, and the lattice constants slightly increased with increasing Er content from x = 0 to 0.20. Considering the differences in ionic radii, 0.63 Å (Fe3+), 1.019 Å (Y3+), and 1.144 Å (Er3+), the small increase in the lattice constants of Er:YIG can be explained by Er3+ substitution at the Y3+ sites in the YIG system.[23]

Figure 2

Rietveld refinements of the X-Ray diffraction data of Y3–ErFe5O12 with the Er content (x = 0, 0.01, 0.05, 0.10, 0.15, and 0.20). The experimental and simulated X-ray diffraction are depicted as red dots and black lines, respectively. The difference between the experimental and simulated X-ray diffraction is depicted as a blue line. X-ray diffraction patterns well match the reference peaks of YIG (green bars, space group -3) with a low Er content. In x = 0.20, X-ray diffraction patterns are refined with YIG (upper green bars, space group -3) and YFeO3 (lower green bars, space group ).

To study the optical properties of Er:YIG, we conducted FT-IR, UV–VIS, and near IR spectroscopy analyses with Y3–ErFe5O12 powder (x = 0, 0.01, 0.05, and 0.10). The transmittance measurement of Y3–ErFe5O12 in the region of wavenumbers from 400 to 1800 cm–1 showed that YIG has strong adsorption at around 450–750 cm–1. Figure a shows the FT-IR spectrum of Y3–ErFe5O12 with the Er content (x = 0, 0.01, 0.05, and 0.10) in the region of wavenumbers from 450 to 750 cm.–1 There were three intense absorption bands, located at around 543, 581, and 646 cm–1, which can be assigned to the asymmetrical stretching vibration of a FeT–O tetrahedral in YIG.[24,25] McDevitt et al. reported that the absorption bands in garnets could be shifted linearly with the ionic radius of rare-earth atoms.[26] In the case of Er:YIG, the absorption bands k1 and k3 of the Fe–O tetrahedra were red-shifted while k2 was slightly blue-shifted by increasing Er concentrations (see the trend curves of k1, k2, and k3 in the inset of Figure a). Considering the similarity in the ionic radius of Er3+ and Y3+, the red and blue shifts in the adsorption bands might be caused by the local environment changes of the Fe–O tetrahedra, as Er atoms are substituted into Y atomic sites in the YIG system. The absorption bands at around 400 cm–1 could be matched to the vibrations of the Fe–O octahedra and Y–O bonds.[24,25]

Figure 3

(a) FT-IR spectrum of Y3–ErFe5O12 with the Er content (x = 0, 0.01, 0.05, and 0.10) with three absorption bands, k1, k2, and k3. The inset shows the shifts of k1, k2, and k3 as a function of Er contents (x). (b) UV–VIS–NIR spectrum of Y3–ErFe5O12 with the Er content (x = 0, 0.01, 0.05, and 0.10). The inset shows the optical transition of Er3+ in a wavelength range of 1450–1550 nm. (c) Tauc plot of Y3–ErFe5O12 with the Er content (x = 0, 0.01, 0.05, and 0.10) where α is the absorbance and hv is an incident photon energy.

We also studied the optical properties of Er:YIG in the UV–VIS–NIR range. Figure b shows that the Er:YIG has very low reflectance in the VIS range (300–700 nm), and a steep increase in reflectance was observed at around 700 nm. Note that high reflectance exceeding 80% was observed above a wavelength of 1200 nm, which is consistent with the reported high transmittance of YIG in this wavelength range. In Er:YIG with x = 0.05 and 0.10, there are small dips in the wavelengths range of 1450–1550 nm, which correspond to the optical transition between I15/2 and I13/2 in Er3+ ions (see the inset figure in Figure b).

We checked the optical bandgaps of Er:YIG using the Tauc plot method, which has been used to estimate the optical bandgap in many semiconducting systems.[27,28] According to the calculated band structures of YIG, YIG with a cubic structure (-3) exhibits a direct bandgap.[29,30] Therefore, we plotted (αhv)2 as a function of the photon energy as shown in Figure c. The optical bandgaps deduced from the Tauc plot were around 1.69–1.70 eV over all of the Er concentrations, indicating that the optical bandgaps are almost invariant with Er substitution in a slightly Er-substituted YIG system.

The magnetization of YIG originates from the antiferromagnetically coupled 3d spins of FeT and FeO. Since Er atoms have a large magnetic moment, exceeding 9.6 μB/Er atom, Er substitution at the Y sites can change the magnetic moments of the YIG system. The temperature (T)-dependence of magnetization (M) in Er:YIG under a magnetic field of 0.05 T shows that Y3–ErFe5O12 has ferromagnetic behavior as shown in Figure a. The room temperature magnetic moments of Y3–ErFe5O12 with x = 0, 0.01, 0.05, and 0.10 were ∼3.0 to 3.3 μB per one formula unit. The temperature-dependent magnetic moments of Er:YIG showed that the magnetic moments were saturated or slightly decreased with a decrease in temperature below 100 K. Interestingly, antiferromagnet-like behaviors (dips in magnetization) appeared at low temperatures (<30 K) for all of the samples, except for the undoped YIG, which is likely due to the ordering of Er 4f spins in the low-temperature range.[31] The contribution of the Er 4f spins should be dominated by the interactions between Er and the YIG lattice rather than a direct Er–Er interaction because Er 4f orbitals should be much more localized than Fe 3d orbitals and furthermore much farther apart from each other (∼3.8 Å) compared to Er–FeT (∼3.10 Å) or Er–FeO (∼3.46 Å).[32] The onset temperatures of the dips (referred to as TEr and are denoted by triangles in Figure a) increased and the amount of decrease in the magnetic moment below TEr increased with increasing Er concentration, which confirms the conclusion that the dips originated from the weak Er 4f spin ordering. Nevertheless, the presence of Er atoms did not significantly alter the magnetic moments at temperatures above TEr (<30 K), clearly showing that Er 4f spins had a negligible impact on room temperature magnetism.

Figure 5

Magnetic field (H) dependency of magnetization in Y3–ErFe5O12 (x = 0, 0.01, 0.05, and 0.10) at various temperatures.

We estimated the Curie temperatures (TC) where the ferromagnetic ordering disappeared in Er:YIG using DSC measurements. Upon cooling and heating Er:YIG, an exothermic and endothermic transition occurs at certain temperatures, which are referred to as Curie temperatures (TC) of Er:YIG.[33]Figure shows the changes of TC during heating and cooling and are marked with black dashed lines; TC decreases by roughly 10 K with increasing content of Er from x = 0 to 0.20. Considering a reported TC of YIG at around 560 K,[34] the TC of the Er:YIG slightly decreases by substituting Er into YIG.

Figure 4

(a) Temperature-dependent magnetic moment of Er:YIG (x = 0, 0.01, 0.05, 0.10, and 0.20) from 2 to 300 K. (b) DSC measurements of Y3–ErFe5O12 with Er content (x = 0, 0.05, 0.10, and 0.20) with a changing temperature between 300 and 700 K upon heating and cooling.

The magnetic field (H) dependence of the magnetization of Er:YIG was measured at various temperatures. Figure b shows that the magnetic moments of Y3–ErFe5O12 were saturated above 0.1 T at a value of around 4.0 μB/YIG, which is similar to the theoretically predicted net magnetic moment of YIG (5.0 μB/YIG).[35] In a temperature range between 2 and 300 K, the saturated magnetic moments of Y3–ErFe5O12 under a high magnetic field changed slightly with temperature and Er concentrations; however, there were no significant changes in the magnetic properties. The magnetic hysteresis curves of Y3–ErFe5O12 (x = 0.01, 0.05, and 0.10) taken at 2 and 300 K were compared as shown in Figure a. The magnetic hysteresis curves of Y3–ErFe5O12 were retained up to room temperature with the coercive field (HC) and residual magnetization (MR) of Y3–ErFe5O12. The changes in HC and MR are explained by impeditive spin ordering at the surface of the Er:YIG powder or the formation of secondary phases.[16,36,37]Figure b shows that the HC and MR are almost invariant to the Er content at 300 K, while the HC and MR at 2 K decrease with the Er content. Therefore, the decrease of HC and MR at low temperature are hardly related to particle sizes or impurity phases in Er:YIG. This is likely because antiferromagnetic ordering of Er ions would weaken the ferromagnetic ordering of Fe ions, leading to the decrease of MS and HC below TEr (<30 K).

Figure 6

(a) M (H) loops for Er:YIG measured at 2 and 300 K. (b) The coercive force and residual magnetization of Er:YIG as a function of the Er content.

For the XAS and XMCD measurements, the intensities of the XAS signals were collected using circularly polarized X-rays at a magnetic field of B = 0.7 T with alternating directions parallel (I+) and antiparallel (I–) to the specimens’ normal direction. The isotropic spectra (=I+ + I–) taken at Fe L3- (photon energy ∼709 eV) and L2-edges (∼723 eV) reflect the excited electronic structure of Fe 3d utilizing the strict 2p3/2(2p1/2) → 3d dipole transition, thereby showing the overall valence and coordination symmetry of the Fe ions irrespective of the magnetism. This implies that the signals from the Fe-containing secondary phase, i.e., YFeO3, were also unavoidably captured. Meanwhile, the difference spectra (=I+ – I–), called XMCD, capture the signals from the ferromagnetically ordered Fe ions so as to reveal the electronic structure in ferromagnetic material only. This can help to distinguish the signals from the (Er- substituted) ferromagnetic YIG from that of the antiferromagnetic YFeO3.

Figure a shows the isotropic XAS spectra of the Er:YIG powders with targeted Er concentrations of x = 0, 0.05, 0.1, 0.15, and 0.2 for Y3–ErFe5O12. Overall, the XAS peak energies can be utilized to measure the average valence of the Fe ions. The energies of the most intense peaks were ∼709 eV for all of the samples, which is indicative of a predominant valence of +3 in the Fe ions.[38] Moreover, the XAS lineshapes reflect details of the Fe–O coordination, such as its symmetry and the strength of orbital hybridizations. The spectra are very similar to each other. This suggests identical coordination symmetry of the Fe ions regardless of the Er concentration. The presence of a secondary phase, o-YFeO3 (see Figure ), might also contribute to the XAS data. However, no significant changes in the lineshapes were observed, even for the Er-rich sample (x = 0.2).

Figure 7

(a) Isotropic term of XAS and (b) XMCD of Er:YIG. The reference spectra of (c) XAS and (d) XMCD from a YIG crystal together with the simulated spectra, adapted with permission from ref (40). Copyrighted by the American Physical Society.

The XMCD spectra are displayed in Figure b. The intensities of the dips and peaks in the five spectra are almost the same, again indicating negligible evolution of the Fe magnetism with increasing Er concentration. Since o-YFeO3 is almost antiferromagnetic[39] and Er2O3 has no Fe ions, the XMCD spectra may be solely due to the local structure and magnetism of the Er:YIG. This indicates that the structure and magnetism in Er:YIG is robust with the inclusion of Er up to x = 0.2 despite a small increase in the lattice constant (see Figure ), and these results are consistent with the robust magnetism of Er:YIG measured using SQUID.

For comparison, the XAS and XMCD spectra of a bulk YIG crystal taken from Chin et al. are appended in Figure c,d, respectively.[40] Both spectra can be decomposed into the contributions from two different coordination symmetries of the Fe3+ ions in YIG, octahedral (Oh) and tetrahedral (Td). The results of the Fe 2p–3d atomic multiplet calculations, which have taken into account the configuration interactions under either the Oh or Td O ligand field, successfully reproduce the experimental data of YIG.[40]

Thus, each of the peaks and dips in Figure b can be due to the signatures of Fe3+–Oh (FeO) and Fe3+–Td (FeT): Fe3+–Oh for the two peaks at ∼707 and ∼710 eV, and Fe3+–Td for the dips at ∼709 eV in the L3-edge region. The values of FeO at the L3-edge are overall positive (peaks) while those of FeT are overall negative (dips). This is because the spins of the FeO and FeT sites tend to align antiferromagnetically to each other due to super-exchange interactions. The sum of the two XMCDs reproduces the experimental data of Er:YIG, as well as the YIG crystal. Therefore, this confirms that the ratios of the Fe3+ ions, FeT/FeO = 3/2, are indeed maintained in Er:YIG.

To elucidate the Y replacement by Er, we conducted Er L3-edge XAS for Er:YIG with x = 0.2 as a representative. Figure shows the Fourier-transform (FT) magnitudes of the extended X-ray absorption fine structures (EXAFS) of the x = 0.2 sample and a reference Er2O3 powder. The FT was processed on k2-weighted (k: electron momentum) EXAFS in a range of 2–10 Å–1 using a Hanning window.[41] The FT EXAFS magnitudes show the bonding distributions of Er ions in both samples. The overall FT spectra are very different from each other, suggesting that the local environment of Er in the x = 0.2 sample is fundamentally different from that of Er2O3. This confirms that the presence of Er2O3 in Er:YIG samples is negligible.

Figure 8

FT magnitudes of k2-weighted Er L3-edge EXAFS oscillations (χ) showing the bonding distributions of Er ions in the x = 0.2 sample, and the reference Er2O3 powder. The XANES results are displayed in the inset.

The major peaks were assigned according to the phase uncorrected atomic distance, R, as Er–O, Er–FeT, and Er–FeO for the x = 0.2 sample, and as Er–O and Er–Er for Er2O3. The roughly estimated bond lengths (R + Δ) after the scattering phase corrections (Δ) were ∼2.2 to 2.4 Å (Er–O), 2.84 Å (Er–FeT), and 3.21 Å (Er–FeO), a bit reduced from the bond lengths of 2.37–2.43 Å (Y–O), 3.09 Å (Y–Fetet), and 3.46 Å (Y–Feoct) in pristine YIG.[32] The reduction in bond lengths (by 0.1–0.2 Å) appears contradictory to the slightly different size (radius) of Er3+ ions (1.144 Å) compared to Y3+ ions (1.019 Å). Perhaps, certain local lattice shrinkage is accompanied near Er3+ in substituted YIG. We also considered how Er substitution influences three absorption bands in the FT-IR spectra, which correspond to the vibration of the FeT–O tetrahedra. However, the composition dependence of the absorption wavenumbers is rather weak, implying the structural shrinkage occurs only locally.

Meanwhile, the average valence of the Er ions was the same as in Er2O3 (that is, 3+). This was confirmed because the first peaks had the same energy (highlighted by the dashed vertical line) in the X-ray absorption near-edge structures (XANES) displayed in the inset. The overall similarity in the Er local structure (EXAFS) as well as in the average valence of Er (XANES) strongly suggests that the Er3+ ions indeed occupy the Y sites in the YIG structure. This is consistent with the finding of robustness of the local environment (and thereby the magnetism) of the Fe ions in Figure .

The Er3+ ions substituting the Y sites can have limited influence on the magnetism in YIG because the magnetism is dominated by the spin interactions of the relatively dispersed Fe 3d orbital states, not the localized Er 4f orbital state. If Er3+ occupied FeT or FeO sites instead, the Fe3+ ions leaching from Er:YIG should have constituted a different Fe compound with a distinct local structure (as well as magnetism) from YIG. This is contradictory to the findings of XAS and XMCD in Figure , where there are no noticeable changes in the lineshapes observed upon Er doping. Therefore, we concluded that Er doping indeed occurred in YIG (apart from the formation of YFeO3), replacing the Y ions, while the electronic or spin configurations of the Fe ions hardly changed.

Therefore, the decrease in magnetic moments with an increasing Er concentration cannot account for the degradation in Fe 3d ferrimagnetism in YIG. The temperature-dependent magnetization data in Figure a show that the magnetic moment decreases gradually as the temperature decreases, even at temperatures above TEr in the low-temperature region (<150 K). Although the Er 4f spins may contribute to the magnetic moments at low temperatures, as in the case of Er2O3,[42] the Er 4f electron states should be highly localized in contrast to the delocalized Fe 3d states, so as to be less relevant to the spin transfer, which is the key ingredient in spin devices. This indicates that the loss in spin-transfer functionality induced by the Er substitution would not be substantial.

On the other hand, there is a small but clear evolution in the optical properties (reflectance in the wavelength range of 1450–1550 nm in Figure b) due to optical transitions within the Er3+ 4f states. Absorption will lead to radiation in the C-band, which is useful for telecommunications. Thus, adding Er offers a promising way of improving optical properties while preserving the magnetic properties in YIG. Further studies are needed to optimize the efficiency of the dopant-related radiation to use YIG-based materials in magneto-optical devices.

Conclusions

Y3–ErFe5O12 was successfully synthesized with various Er concentrations up to x = 0.20 using a solid-state reaction method, and the effects of the Er substitution on the magnetic and optical properties of YIG were scrutinized using various techniques. The Er atoms replacing the Y sites in YIG seemed to modify the optical properties, whereas the magnetic properties were barely modified by the Er substitution except for the contribution of Er 4f spins in the low-temperature regime (TEr < 30 K). The magnetism near room temperature was retained even for the Er-substituted YIG, suggesting that rare-earth ion doping is an effective way of controlling optoelectronic properties in the mid-infrared range, which is useful for magneto-optical device applications.