Luminescence and high temperature ferromagnetism in YAlO nanophosphors: materials for efficient next generation LEDs and spintronic applications

Abstract

Current research on environment friendly and efficient luminescent materials for white LEDs, envisaged as next generation energy efficient lighting devices, has spurred research into the different forms of yttrium-aluminum-oxide (YAlO) (YAG: garnet structured YAlO and YAM: monoclinic structured YAlO), including the discovery of a new phase of tetragonal structured YAlO named YAT. Appropriate luminescence in the visible-yellow region has been realized by doping rare earth (Ce) in the YAlO host. Interestingly, this work indicates that all the studied YAlO nanophosphors (NPs) exhibit dilute ferromagnetism up to 600 K. The observed magnetic character of YAlO NPs is attributed to the oxygen vacancies and remains unperturbed even after Ce doping, thereby establishing the independent nature of the two phenomena. This non-interfering luminescence and magnetic character of the YAlO NPs is anticipated to have a wide application potential in high temperature spintronics, LEDs and other biological applications.

---

1. Introduction

Materials that exhibit optical and magnetic properties simultaneously are touted as one class of advanced multifunctional materials expedient in magneto-optic (MO) and spintronic devices.^1–4 Conventional MO materials are normally composites constituted by optical and magnetic materials. However, it is not trivial to comprehend the interaction between optical and magnetic properties (i.e., tuning the optical properties by a magnetic field or vice versa) in these materials owing to the extrication between the constituent optical and magnetic phases. Nevertheless, designing and producing materials which exhibit both optical and magnetic properties independently would be highly desirable in developing advanced MO devices for high-accuracy communications, aircraft guidance and magnetic field detection.^4–7 MO interaction in a single-phase material implies interaction amongst the atoms in the same crystal lattice leading to the simultaneous exhibition of optical and magnetic properties. Here it should be noted that the MO interaction in composites occurs at the interfaces (between the constituent phases) which would be associated with a high density of defects. Recent reports show that the MO interactions can be observed in single-crystal or single-phase oxides doped with lanthanides; for example, in NaGd-F₄:Yb³⁺,Er³⁺ and Gd₂O₃:Yb³⁺,Er³⁺, the host materials containing the Gd³⁺ ions exhibited paramagnetic properties while the dopant Er³⁺ ions formed luminescent centers.^6,8–10 Further, the luminescence of the Er³⁺ ions was found to depend on the coordinated magnetic ion, Gd³⁺. In other words, if the impact of the coordinating magnetic ions on the luminescent centers is strong enough, they could alter or influence the luminescence of the lanthanide-doped oxides depending on the applied magnetic field. However, in the case of nanophosphors (NPs) it would be convenient to independently control both optical and magnetic properties. Here subsists the motivation behind this work, which is about the luminescence and magnetic properties of unique yttrium-aluminum-oxide (YAlO) NPs.

Among several stable YAlO crystallographic phases garnet cubic Y₃Al₅O₁₂ (named as YAG) and Y₄Al₂O₉ (named as YAM) exhibit luminescence in the visible region depending on the dopant used.^11–16 For example, YAG doped with Ce is a yellow-light emitting luminescent phosphor material and by coating this on high-brightness blue InGaN diode, a white light emitting device (LED) can be fabricated. This white LED has distinct advantages such as ease of fabrication and low cost compared with homo/hetero junction based devices. However, the drawback is the light itself which is colder and bluer than that from traditional lamps due to the poor color rendering index. The most viable proposition to address this issue is to improve the efficiency of YAG:Ce NPs by enhancing the luminescence of phosphor-converted white LEDs.^17,18

Through this work, Y₃Al₅O₁₂ (named as YAT), a tetragonal YAlO phase, is being showcased as an excellent alternative material which addresses the above mentioned issues. While YAT was identified in PM2000 steel¹⁹ and its thermal expansion²⁰ studies were separately reported, its application as a luminescent material has not yet been reported. It will be shown in this work that the room temperature (RT) luminescence efficiency of YAT:Ce is much better in comparison to YAG:Ce and YAM:Ce while undoped YAT exhibits only defect-related luminescence in the visible region. Through this work, the ferromagnetism (FM) in such standalone YAlO NPs insulators with a closed shell band and, more significantly, its highly stable nature both with and without Ce doping at temperatures up to 600 K and beyond, without the addition of any magnetic impurity will also be highlighted. Stable FM from non-magnetic insulating materials (namely undoped HfO₂ (ref. 21) and CaB₆, SrB₆, BaB₆,^22–24 Ca_1−xLa_xB₆,²⁵ and CaO²⁶) has been previously observed and it was attributed to the presence of intrinsic defects. In this work, the observation of a stable magnetic behavior (at temperatures from 5 K to above 600 K) of all the NPs is one of the significant findings. The present study demonstrates a promising way to understand the role of intrinsic defects and dopants in YAlO insulators to obtain the desired luminescence along with the magnetism. This work elucidates the origin of the FM which is attributed to the strong exchange interaction between localized electrons' spin moments resulting from oxygen vacancies (V_o). This work also throws light on the extrication between luminescence and FM in these YAlO NPs.

2. Experimental section

2.1. Synthesis of the materials

YAlO NPs are synthesized by controlled post-heated solidified combustion technique.²⁷ Al(NO₃)₃·9H₂O (A.R., Aldrich, 99.997% trace metals basis), Ce(NO₃)₃·6H₂O (Aldrich, 99.9%) and Y₂O₃ (A.R., Aldrich, 99.999% trace metals basis) were used as raw materials. The yttrium nitrate solution was prepared by dissolving Y₂O₃ in diluted HNO₃ and 5 ml of deionized water followed by heating at 80 °C for 30 min to get a clear solution and evaporate excess acid. Al(NO₃)₃·9H₂O and Ce(NO₃)₃·6H₂O (varied from 0.025 to 0.5 M) were added in to the above solution. 0.1 M% of urea was then added into the solution and kept in preheated furnace (550 °C) for 20 min. The resulting fluffy mass samples were grounded and subsequently fired at 1300 °C for 4 h in an air atmosphere. Stoichiometry of Y/Al was maintained as per the reported procedure.²⁰

2.2. Characterization

The synthesized materials were characterized for their crystal structure by a powder X-ray diffraction (XRD) technique. XRD patterns were recorded in the 2θ range of 2–60° using Cu Kα radiation (1.5406 Å) with a standard monochromator using Ni filter on a Siemens D5000 (Cheshire, UK) instrument operated at 40 kV and 30 mA. The average crystallite size of the samples was estimated by considering major XRD peaks and by applying the Scherrer equation. The synthesized materials were characterized for morphology by Transmission Electron Microscopy (TEM) on a Philips Technai G² FEI F12 operated at a maximum accelerating voltage of 120 kV. For TEM, the samples were dispersed in ethanol by ultra-sonication and then loaded on to formvar coated copper grids which were then air-dried.

²⁷Aluminum (²⁷Al) solid state Magic Angle Spinning-Nuclear Magnetic Resonance (MAS-NMR) was used to confirm the oxidation states pertaining to different YAlO NPs. These measurements were performed on a Bruker ultrashield 500 MHz WB NMR spectrometer in an 11.75 T field producing a ²⁷Al Larmor frequency at 130.31 MHz. The samples were placed in a 3.2 mm rotor and the spinning rate was set at 10 kHz. The applied pulse width corresponding to an angle of π/2 for ²⁷Al was 4 μs with a pulse delay of 10 s while the pulse sequence applied was the normal single pulse. The chemical shift reference was set at 0 ppm corresponding to the aluminum nitrate single peak. The X-ray photo electron spectroscopy (XPS) spectra were obtained using the M/s. SPECS XPS system (M/s. SPECS Surface Nano Analysis GmbH, Germany) equipped with a twin anode X-ray source and a hemispherical analyzer with a single-channel detector. The spectra were obtained with 150 W non-monochromatic Al Kα radiation (1486.6 eV). The RT photoluminescence (PL) spectrum was recorded on a double monochromator based Cary Eclipse luminescence spectrometer (AGILENT Technologies) with a xenon flash lamp as the source of excitation.

Magnetization versus magnetic field (M–H) of the YAlO NPs powder samples was measured at RT using a micro-sense vibrating sample magnetometer (VSM) using a glass rod wrapped with aluminum foil with YAlO samples as the sample holder. Variation in magnetization and coercive field with temperature was recorded using a VSM for the Curie temperature determination. A continuous flow of nitrogen gas (under a liquid nitrogen atmosphere) was maintained to decrease the temperature and argon gas was used for high temperature measurements. The results of the magnetic measurements were reproducible for all the samples. Magnetic measurements carried out on different batches of YAlO samples showed consistent results. Low temperature (<130 K) magnetic measurements were performed using a quantum design superconducting quantum interference device (SQUID) magnetometer MPMS-5P, operating between 5 and 300 K for dc-applied fields ranging from −5 to 5 T. Zero-field cool (ZFC)/field cool (FC) and hysteresis measurements were performed on centrifuged and dried NP samples wrapped in a polyethylene membrane. ZFC/FC measurements were recorded at temperatures in the range 5–300 K and at an applied field of 100 Oe. The hysteresis measurements were performed at RT.

Magnetic force microscopy (MFM) measurements were carried out to confirm the RT FM of the samples as well as to study their magnetic domain structure. High resolution magnetic phase shift mapping of the NPs was carried out by using an MFM module in the scanning probe microscopy from NT-MDT (Russia, Model-Aura), in semi-contact mode. Silicon cantilevers (NSG-01/Co, from NT-MDT) coated with cobalt/chromium film having coating thickness 30–40 nm, tip radius ∼40 nm, and resonance frequency 222 kHz, were used. All the results reported in the present work were carried out under an external magnetic field of 1000 gauss. The samples for the measurements were prepared by compacting the nanophosphor particulate powder. Throughout this work, a nonmagnetic SMENA head (NT-MDT make) and sample holder were used so as to omit any magnetic interference during the measurements.

3. Results and discussion

Powder XRD patterns of the undoped and Ce doped YAT NPs are shown in Fig. 1(a) while those of YAG and YAM NPs are shown in the ESI (Fig. ESI1(a) and (b),† respectively). XRD patterns shown in Fig. 1(a) are in good agreement with JCPDS card no. 09-0310 (Fig. ESI1(c)†), confirming the tetragonal crystal structure of YAT samples without any other impurity phase. XRD patterns shown in Fig. ESI1(a) and (b)† are in good agreement with JCPDS card no. 33-0040 and 34-0368, confirming the cubic and monoclinic crystal structure of the YAG and YAM samples, respectively. The mean crystalline size of these NPs was estimated as in the range 25–37 nm. Calculated cell parameters of undoped YAT (a = 7.5156 Å and c = 4.2406 Å) are in agreement with standard cell parameter values (a = 7.51 Å and c = 4.24 Å as per JCPDS card no. 09-0310) establishing that the obtained sample is a tetragonal phased Y₃Al₅O₁₂. The peak broadening compared with standard bulk samples are due to a finite size effect indicating that the sample consists of nanocrystalline domains²⁸ and the peak shift to lower 2θ (Table ESI1†) values in the case of doped NPs compared with undoped ones can be ascribed to the incorporated activator atom effect. The observed increase in the value of cell parameters for Ce doped YAT compared with undoped YAlO (as an example, the YAT:Ce case is shown in Table ESI1†) is due to the incorporation of the larger Ce ion (0.115 nm) at the Y ion (0.104 nm) position with a higher probability of replacement due to the identical valence of +3 (ref. 29) of both the ions and the presence of Ce. ²⁷Al solid state NMR (MAS-NMR) is used for further confirmation of the phase formation and the assignment of different coordination states of Al centers in the NPs. In the garnet Y₂O₃–Al₂O₃ structure, the lattice consists of a 3D oxygen network where Al³⁺ resides both in elongated octahedral and compressed tetrahedral interstitial sites, in the ratio of 3 : 2, whereas yttrium atoms occupy dodecahedral sites with distorted polyhedra.^30,31 Considering the structure of pure YAlO, the Al ions tend to occupy two different sites in the crystal lattice and consequently have two coordination numbers. One site is positioned at the center of a tetrahedron (tetra-coordinated Al³⁺, 90–60 ppm), and the other one is at the center of the octahedron (hexa-coordinated Al³⁺, 15–10 ppm).^31–33 The NMR peaks observed for the YAT, YAG and YAM NPs at 9.88, 9.91 and 1.53 ppm, respectively, as shown in Fig. 1(b) are in good agreement with the values observed for octahedral AlO₆ sites, while the minor peaks in the range 88–55 ppm are due to tetrahedral AlO₄ sites. Small peaks observed at 9.88 and 1.47 ppm in the YAT and YAM cases are similar to YAT and YAM octahedral AlO₆ peaks, suggesting the possibility of a by-product corresponding to these NPs.³⁴ Similar peak positions at 9.88 and 9.91 ppm of YAT and YAG prove that both belong to the same garnet group. Further, peaks observed at −65 to −75 ppm could be due to anti-site defects;³⁵ a separate study is essential to ascertain this.

Fig. 1 (a) XRD patterns of undoped and Ce doped YAT NPs, (b) Al coordination states in YAT and YAG as per MAS-NMR, and (c) XPS data of undoped and Ce doped YAT NPs.

For a confirmation of the stoichiometry pertaining to the samples and oxidation state of the dopant Ce ion, XPS measurements were carried out and the results for undoped and Ce doped YAT NPs are presented in the Fig. 1(c). The results establish the presence of yttrium and aluminum in the expected stoichiometry of 3 : 5 (Y/Al) for YAG and YAT and 4 : 2 for YAM NPs (Table ESI2†). The Ce valence band spectra show two distinct characteristic peaks corresponding to the Ce ion at the binding energy positions of 886 and 903 eV, which correspond to the Ce 3d_5/2 and Ce 3d_3/2, respectively, due to the 3d orbital split of Ce ions by the spin–orbit coupling effect.^18,36 Another observation is the decrease in the concentration of yttrium in doped NPs, indicating the substitution of Y by Ce ions.

The surface morphology of the YAlO NPs is revealed by the FESEM and TEM images, Fig. 2(a–c) shows the formation of a foam-like macroporous structure with a roughly uniform network with average pore size of about 50 to 100 nm for undoped and 0.1 M% Ce doped YAT NPs. These morphology studies suggested that the hierarchically macroporous YAlO NPs can be prepared by a simple post heated solidified combustion technique without the need of any external templates and a detailed study will be undertaken. The SAED pattern of the NPs (inset of Fig. 2(c)), a diffuse ring pattern, is an indication of the polycrystalline nature of the sample.^37,38 The morphology and SAED patterns of undoped and Ce doped YAT and YAG, and YAM NPs (Fig. ESI2†) reveals a similar observation as that in the undoped and doped YAT material.

Fig. 2 FESEM images of (a) undoped YAT and (b and c) FESEM and TEM images of 0.1 M% Ce doped YAT NPs. Inset in (c) is the selected area electron diffraction pattern.

PL excitation and emission spectra of Ce doped (with different Ce concentrations) YAT, YAG and YAM NPs observed at identical conditions are shown in Fig. 3. Undoped YAT, YAG and YAM NPs (Fig. 3(a)) show a low intensity broad emission spectrum between 490 to 535 nm. However, higher peak intensity was observed in the case of YAG NPs compared to YAT and YAM NPs which is attributed to the difference in intrinsic defects.^39,40 It is important to note that the doubly ionized oxygen vacancies (+2e) are the easiest to form and are more stable than the other charged species.⁴¹ The XPS analysis of YAG NPs also shows the presence of more oxygen vacancies compared to YAT and YAM, directly corroborating with the observed PL efficiency of corresponding NPs confirming that the observed defect related PL emission is due to V_o. Furthermore, the absence of longer wavelength emission (∼800 nm) which is associated with iron impurities^39,42 confirms the purity of YAlO NPs. The single broad strong emission spectrum observed at 574, 555 and 540 nm refers to 460 nm excitation of YAT:Ce, YAG:Ce and YAM:Ce NPs, respectively. Maximum PL outputs attained at 0.1 M% Ce doped NPs confirms that the NPs have attained the optimum concentration, as a decrease in emission intensity has been observed with a further increase of Ce above 0.1 M% concentration (up to 0.5 M%). PL intensity increased with increasing Ce concentration up to 0.1 M% and was stable with an increase in Ce concentration (0.2 M%) and decreased with a further increase in Ce to 0.5 M%. The observed PL saturation could be attributed to several factors, such as the presence of non-radiative de-excitation pathways brought about by Ce–Ce interactions or lattice defects that become more important at higher Ce concentrations.^18,43,44 Two broad excitation bands: a strong band centered around 460 nm and a very low intensity band at 340 nm are observed upon Ce emissions, originating from the 4f–5d transition of Ce.¹⁸ The two excitation bands are considered to be related to the electron transition from the 4f ground state (²F_5/2, ²F_7/2) to the different crystal-field splitting (CFS) levels of the 5d state (²D_5/2, ²D_3/2) of Ce. The observed low intensity peak at 340 nm is much lower than the reported intensity values,^17,18,29,45 implying that the population levels of ²F_5/2, (or) ²F_7/2 to ²D_3/2 are totally quenched, leading to an efficient absorption at the ²F_5/2, (or) ²F_7/2 to ²D_5/2 levels at room temperature and resulting in efficient luminescence at 540–574 nm. This emission is ascribed to the electron transitions from the lowest CFS component of the 5d level to the ground state of Ce.

Fig. 3 Variation in the intensity of the PL from (a) undoped and Ce doped (b) YAT (c) YAG and (d) YAM NPs. Insets of the corresponding images show body color of the powders under a fluorescent room light. Inset in (d) is the excitation spectrum of YAM:Ce NPs. Slit width was 10 nm, 10 nm in the case of YAG:Ce and YAM:Ce NPs.

The peak shift (red shift) in the case of YAT:Ce compared to YAM:Ce and YAG:Ce is due to a change in the crystal field surroundings because the radial wave function of the excited 5d electron extends spatially, well beyond the closed 5s² 5p⁶, implying that its states are strongly perturbed by the crystal field. Thus, both the strongest excitation band and the strongest emission band are associated with the lowest lying 5d state, which is affected by the crystal field surroundings. Interestingly, YAT:Ce falls in the perfect yellow range (570–590 nm) compared to the greenish yellow (540–555 nm) of YAM:Ce and YAG:Ce NPs, and this difference is reflected in the body color of the powder under fluorescent room light, which is greenish yellow for YAM:Ce and YAG:Ce and perfect yellow for YAT:Ce NPs (respective insets in Fig. 3(b–d)). There is a four-fold increase in the luminescence emission intensity of YAT:Ce NPs in comparison to YAG:Ce and YAM:Ce NPs (excitation and emission slit width was 2.5 nm, 2.5 nm in the case of YAT:Ce compared with 10 nm, 10 nm in the case of YAG:Ce and YAM:Ce NPs) confirming that YAT:Ce NPs are more suitable for WLEDs when compared with YAG:Ce.

Absorption spectra (Fig. ESI3†) show negligible absorption at 340 nm with a strong major broad peak between 400 and 600 nm centered at 489 nm in the case of YAT:Ce compared with YAG:Ce and YAM:Ce NPs, which a show significant peak at 340 nm, which further establishes the high population level of ²F_7/2 to ²D_5/2 level compared to ²F_7/2 to ²D_3/2. Such a low population level of the 340 nm peak is reported for low temperature (4.5 K) absorption measurements with a high level population of 489 nm and vice versa for high temperature (293 K) measurements.^46,47 The position of the absorption band exactly matches with the reported value (489.16 nm) of the zero-phonon line for the lowest energy d-state of the YAG:Ce (489.2 nm).⁴⁷ The low population level of 340 nm at RT for YAT:Ce NPs could be attributed to the larger CFS, since the intensity, position of the absorption band and broadness of the absorption band are very sensitive even to small variations in the local coordination, which affects the covalency and the CFS.⁴⁶ Generally, V_o leads to the low CFS due to active absorption of the incident photon by these defects, which leads to a low absorption and emission intensity. The observed red shift in YAT:Ce NPs (compared with YAM:Ce and YAG:Ce) resulting from a high CFS⁴⁶ leads to a high efficiency. The absence of shorter wavelength absorption at 256 nm due to charge transfer transitions from the oxygen ligands to the empty orbits of Fe impurities confirmed the absence of magnetic impurities.^42,48

The excess oxygen deficiency in YAG:Ce and YAM:Ce, as mentioned earlier, could be one of the factors responsible for the low efficiency of the PL, as it is well known that trap levels related to V_o act as luminescence quenchers⁴⁹ because the excited electrons could recombine or get trapped at the nearest defective sites and cause a non-radiative relaxation. As all the NPs are prepared under similar experimental conditions, the presence of an elevated defect (V_o) level in YAG:Ce and YAM:Ce compared with YAT:Ce could be attributed to the different crystal structures, as there is a direct correlation between the crystal structure and the luminescence properties.⁵⁰

Temperature dependent magnetic properties of the NPs were determined from VSM and SQUID magnetometry analysis. Fig. 4 shows the hysteresis loops of undoped and Ce doped YAT, YAG and YAM NPs at RT. All the NPs show similar hysteresis behavior with a maximum magnetization (M) of ∼32.04 memu g⁻¹, coercivity (H_c) of ∼48.94 Oe and remnant magnetization (M_r) ranging from 6.75 memu g⁻¹ to 3.17 memu g⁻¹. H_c increased with a decrease in temperature (from 200 K to 130 K) while it decreased with an increase in temperature above RT (400, 500 and 600 K) while maintaining M nearly constant, as shown in Fig. 5 and ESI4.† Fig. 5(b) shows the hysteresis loops of undoped YAT at 130 and 600 K. H_c and M values of different samples at RT and at other temperatures are tabulated in Tables 1 and 2, respectively.

Fig. 4 RT VSM plots of undoped and Ce doped (a) YAT (b) YAG (c) YAM samples.

Fig. 5 Temperature dependent VSM plots of (a) undoped and (c) 0.1 M% Ce doped YAT NPs. (b) Hysteresis loops of undoped YAT NPs at 130 and 600 K. The external magnetic field is in parallel with the sample surface.

Table 1 Magnetic properties, namely H_c and M values, of different YAlO NPs measured at RT. The units of H_c and M are Oe and memu g⁻¹, respectively

Sample		YAT		YAG		YAM
		Hc	M	Hc	M	Hc	M
Undoped		38.67	23.07	40.25	24.20	40.83	28.24
Ce doped	0.025 M%	45.88	20.28	48.94	27.69	40.70	26.23
	0.05 M%	41.01	20.63	46.77	28.96	47.78	22.75
	0.1 M%	46.41	28.89	41.14	32.04	38.55	22.34
	0.5 M%	42.77	27.13	42.28	24.35	45.92	26.76

---

Table 2 Magnetic properties, namely H_c and M values, of different YAlO NPs measured at various temperatures. The units of H_c and M are Oe and memu g⁻¹, respectively. ‘Ce doped’ implies 0.1 M% Ce doped samples

Sample	Temperature (K)
	130		200		400		500		600
	Hc	M	Hc	M	Hc	M	Hc	M	Hc	M
Undoped YAT	98.43	27.99	60.26	28.63	25.58	25.37	17.35	26.19	11.88	16.58
Ce doped YAT	144.09	12.17	135.33	11.08	39.68	6.62	41.63	80.63	32.81	7.84
Undoped YAG	120.50	9.21	158.59	7.59	66.60	11.15	51.98	10.22	42.12	8.93
Ce doped YAG	154.98	8.29	95.06	5.80	53.94	10.09	36.02	11.23	60.12	9.72
Undoped YAM	120.76	10.63	88.61	10.71	24.08	8.24	23.30	8.17	—	6.60
Ce doped YAM	98.08	14.87	57.47	17.56	49.76	6.90	50.89	7.73	39.39	6.67

---

M_r at 130 and 600 K is 7.81 memu g⁻¹ and 0.56 memu g⁻¹ respectively. Similar behavior was observed in the case of undoped and Ce (0.1 M%) doped YAM NPs. But H_c of the YAG NPs showed a dissimilar behavior when compared with the normal increasing H_c behavior observed from RT to 200 K (158.59 Oe); with a further decrease in temperature to 130 K, instead of following the increasing trend, coercivity was found to decrease to 120 Oe. Similarly, H_c values further increased with an increase in temperature up to 400 K and decreased with a further increase in temperature (Table 2). Some distortion in the garnet YAlO could be the cause for this unexpected magnetic behavior. A saturation magnetization of ∼320 × 10⁻⁴ emu g⁻¹ and a coercive field of ∼46 Oe were measured.

For a further insight into the magnetic properties of YAlO NPs at low temperatures (<130 K), the magnetization versus temperature data were obtained using a SQUID magnetometer in the temperature range 5–300 K. Fig. 6 shows the magnetization of undoped and 0.1 M% Ce doped YAT NPs as a function of temperature obtained at ZFC and FC conditions with an applied magnetic field of 100 Oe. In the temperature regime 150 K to 300 K both measurements show a similar behavior, the magnetization of the NPs slowly increases with decreasing temperature below ∼150 K and drastically increases at temperatures less than 50 K. Thermally stable ZFC–FC magnetization curves without blocking temperature (or super paramagnetic behavior)^51–53 confirm the ferromagnetic nature of the NPs and a clear separation between the FC and ZFC processes further supports this claim.⁵² Measuring ZFC/FC curves is also an effective method for determining the presence of any magnetic impurities.⁵⁴ Magnetic impurities, if present, show up as a blocking temperature that can be observed in the ZFC/FC curves. No signal for a blocking temperature was observed in the ZFC curves, which implies that no magnetic impurities exist in any of the samples within the measured temperature range. This finding further supports the XRD and EDS results, which also do not show the presence of any impurities. In addition, the presence of magnetic hysteresis loops with saturation magnetization and coercive field values of approximately 1.27, 1.02 memu g⁻¹ and 106, 168 Oe observed in the RT SQUID measurements taken for undoped and 0.1 M% Ce doped YAT, respectively (Fig. 6(b)), further supports the inferences above. To further confirm the existence of oxygen vacancies, EPR spectra of the undoped and 0.1 M% Ce doped YAlO NPs have been obtained (Fig. 7). It is seen that all the samples have a broad strong symmetric EPR peak in the range g = 2.02–2.15 while a weak but sharp peak appeared at g = 3.01 in the case of undoped and doped YAG and doped YAT NPs.

Fig. 6 (a) ZFC and FC magnetization curves (at an applied field of 100 Oe) and (b) hysteresis loops at RT for the undoped YAT and 0.1 M% Ce doped YAT NPs.

Fig. 7 ESR spectra of (a) undoped and (b) 0.1 M% Ce doped YAlO NPs at RT.

A broad signal appearing at the lower field is attributed to ferromagnetic-resonance which arises from transition within the ground state of the ferromagnetic domain,^55,56 and a narrow ESR signal arising from the paramagnetic states of surface defects^57,58 is observed. However, it is believed that this narrow signal is not involved in the ferromagnetic ordering.^59,60 The EPR signal with g = 2.00 to 2.15 was reported and attributed to oxygen vacancies.^58,59,61 Similar values have been reported for V_o in oxides like TiO₂ and ZnO.^58,62

Fig. 8 shows the magnetic phase mapping of undoped and Ce doped YAT, YAG and YAM NPs. From the contrast/brightness profile in the magnetic phase map, it can be clearly inferred that the magnetic signals are definitely not due to any surface effect but due to the primary nature of the sample. The bright and dark contrast domains correspond to the regions of high concentrations of repulsive and attractive forces, respectively. Regularly structured magnetic domains are clearly observed that are extended over many agglomerated nanophosphor particulates, with edges corresponding to agglomeration domains. The magnetic phase shift is almost constant (Δφ = 0.11°) in all the samples except the Ce doped YAM (Δφ = 0.25°).

Fig. 8 2D magnetic phase mapping of the undoped and 0.1 M% Ce doped YAlO NPs. MFM images (8 μm × 8 μm) of undoped YAT, YAG and YAM (a–c) and corresponding Ce doped YAT, YAG and YAM (d–f) recorded at RT.

This study establishes a direct correspondence between V_o from XPS studies and the PL efficiency. This establishes that the higher concentrations of oxygen vacancies for undoped YAG are responsible for the observed highly efficient defect emission when compared with YAT prepared under identical conditions. The FM behavior of these materials has been confirmed, but its origin needs to be explained. Neither Y nor Al are magnetic, but the YAlO NPs (for example, undoped YAT) exhibit a well-defined hysteresis at 600 K with a saturation magnetization value at 16.58 memu g⁻¹ and a coercive field of 11.88 Oe, which are far above the values at RT. Having established the purity of the phase by XRD, EDAX, ZFC and EPR, which rules out the presence of any magnetic impurities, the large magnetic moment needs explanation. Another justification to support our claim of the absence of any magnetic impurity is that, in the case of magnetic impurities being present, their effect would be consistent throughout the temperature range but, as observed, the coercive field has a characteristic temperature response.

This leads to the inference that defects like oxygen vacancies (V_o) are responsible for the observed magnetic moment²¹ of YAlO NPs. The XPS and EPR results of undoped and Ce doped YAlO NPs and PL results of undoped YAlO NPs indicate the presence of V_o. The observed high coercive field for undoped and doped YAG NPs in VSM analysis compared with the analogous YAT NPs (Table 2) and the existence of more numbers of V_o from XPS and the highly efficient defect emission in PL analysis for undoped YAG compared with YAT prepared under identical conditions all combine to reason that the observed FM can be directly correlated to the oxygen vacancies. Based on the above luminescence and FM results, the following hypothesis is put forward:

(i) The basic luminescence character in these phosphor materials can be associated with the oxygen vacancies but the enhancement in the specific luminescence is possible only after incorporation of the rare earth ions.

(ii) The inferences drawn from the various characterization techniques attribute the presence of oxygen vacancies as the primary responsible species for the observed FM.

(iii) The FM is exhibited by both the pure YAlO systems of all morphologies and also the Ce doped samples. Even though the PL shows a significant improvement on Ce incorporation, this has little or no influence on the magnetic behavior.

(iv) These observations support our deduction that the FM behavior that originates when the rare earth atoms get coupled with the negatively charged oxygen vacancies is unfounded in the present case. It can thus be proposed that defects like oxygen vacancies are solely responsible for the observed magnetic moment of YAlO NPs. The key issue here is to understand the role of oxygen vacancies in the FM. The exchange mechanisms that could establish the FM are Ruderman–Kittel–Kasuya–Yosida (RKKY) interactions and/or bound magnetic polarons (BMP). The former require a metallic system whereas the latter can exist in semiconducting/insulating materials like our YAlO NPs. Taking into account the existence of oxygen vacancies in the present system, we consider the F-center exchange (FCE) FM.^12,63 The FCE mechanism corroborates with the FM arising due to the presence of V_o. There are three possible charge states of a V_o: (a) F²⁺ center with no trapped electrons; (b) F⁺ center with one trapped electron and (c) F⁰ center with two trapped electrons. The F⁰ center charge states form shallow donor levels or lie above the conduction band edge as shown in Fig. 8. This can form an impurity band, and in the condition when this overlaps with the conduction band of the host oxide, the material exhibits ferromagnetic behavior.⁶⁴

(v) A schematic diagram (Fig. 9) shows that mechanisms supporting luminescence and magnetic properties of YAlO NPs, even though they are non-interfering and independent, can be present in the same material.

Fig. 9 Schematic diagram of the possible mechanism of the luminescence and magnetic properties from YAlO nanophosphors. O_v, VB, CB are the abbreviations for oxygen vacancy, valence band, conduction band respectively.

The above two sections reveal a direct correlation between the luminescence and magnetic behavior of YAlO NPs. Undoped YAG, with a high concentration of oxygen vacancies (Fig. 3(a)) compared with the remaining two compounds (YAT and YAM) exhibits a higher magnetic behavior (Table 2), bringing a one to one correspondence between the oxygen vacancies and the luminescence. Further, to improve a specific luminescence, Ce is introduced as a dopant. This results in an improved luminescence without in any way deteriorating the magnetic behavior. This, in fact, results in obtaining the desired multifunctionality of improving the low efficiency of undoped YAlO NPs, by doping Ce while maintaining the magnetic character.

4. Conclusions

In summary, we have successfully achieved luminescence and magnetic behavior from a single phosphor material without any magnetic impurity, making these independent phenomena in the same host lattice. Improved luminescence while maintaining the FM behavior in undoped and all the Ce doped NPs proves this claim. The magnetic characterization studies supported by XPS, PL and EPR categorically establish that the YAlO NPs exhibit a ferromagnetic behavior and these materials can be classified as ferromagnetic insulators. The findings of high temperature FM from insulating undoped and YAT, YAG and YAM have evoked a new phenomenon termed d⁰ magnetism, or magnetism due to defects like oxygen vacancies. Such a strong FM and efficient luminescence behavior makes these NPs multifunctional materials apposite for spintronics and magneto-optic devices and the stability of the FM above 600 K makes such devices suitable for room temperature and high temperature applications.

Acknowledgements

Dr K. Jayanthi Rajan acknowledges CSIR-Nehru Science Postdoctoral Research Fellowship and a Senior Research Associate (Scientists' Pool Scheme) for financial support. The corresponding authors are thankful to Dr V. Jayatirtha Rao, Head CPC Division, for extending the fluorescence spectrometer facility and Dr L. Giribabu and Dr P. R. Bangal, Senior Scientists, I & PC Division, for discussions on the luminescence studies.

References

1. J.-L. Bridot, A.-C. Faure, S. Laurent, C. Rivièr, C. Billotey, B. Hiba, M. Janier, V. Josser, J.-L. Coll, L. V. Elst, R. Muller, S. Roux, P. Perriat and O. Tillement, J. Am. Chem. Soc., 2007, 129, 5076.
2. H. Kim, M. Achermann, L. P. Balet, J. A. Hollingsworth and V. I. Klimov, J. Am. Chem. Soc., 2005, 127, 544.
3. H.-W. Gu, R.-K. Zheng, X.-X. Zhang and B. Xu, J. Am. Chem. Soc., 2004, 126, 5664.
4. G.-F. Wang, Q. Peng and Y.-D. Li, Acc. Chem. Res., 2011, 44, 322.
5. V. K. Tikhomirov, L. F. Chibotaru, D. Saurel, P. Gredin, M. Mortier and V. V. Moshchalkov, Nano Lett., 2009, 9, 721.
6. S. K. Singh, K. Kumar, M. K. Srivastava, D. K. Rai and S. B. Rai, Opt. Lett., 2010, 35, 1575.
7. R. Valiente, M. Millot, F. Rodríguez, J. González, J. M. Broto, S. George, S. Garcia-Revilla, Y. Romanyuk and M. Pollnau, High Pressure Res., 2009, 29, 748.
8. Z.-L. Wang, J. H. Hao and H. L. W. Chan, J. Mater. Chem., 2010, 20, 3178.
9. D. Wang, L. Ren, X. Zhou, X.-Z. Wang, J. Zhou, Y. Han and N. Kang, Nanotechnology, 2012, 23, 225705.
10. C.-B. Tan, B. Ma, J. Zhang, Y. Zuo, W. Zhu, Y.-X. Liu, W.-B. Li and Y.-T. Zhang, Mater. Lett., 2012, 73, 147–149.
11. D. M. Kaz, C. G. Bischak, C. L. Hetherington, H. H. Howard, X. Marti, J. D. Clarkson, C. Adamo, D. G. Schlom, R. Ramesh, S. Shaul Aloni, D. F. Ogletree and N. S. Ginsberg, ACS Nano, 2013, 7, 10397.
12. L. Yang, Y.-J. Dong, D. Chen, C. Wang, X. Hu, N. Da, G. Zhao, J. Xu, X. Jiang, C. Zhu and J. Qiu, Opt. Express, 2006, 14, 243.
13. B. Dong, J. Wang, J. Sun, S. Xu, X. Bai, Z. Jiang, L. Xia, L. Sun and H. Song, RSC Adv., 2012, 2, 3897.
14. M. L. Saladino, A. Zanotto, D. C. Martino, A. Spinella, G. Nasillo and E. Caponetti, Langmuir, 2010, 26, 13442.
15. S. Lee and S. Y. Seo, J. Electrochem. Soc., 2002, 149, J85.
16. C. C. Lin and R.-S. Liu, J. Phys. Chem. Lett., 2011, 2, 1268.
17. A. A. Setlur, W. J. Heward, M. E. Hannah and U. Happe, Chem. Mater., 2008, 20, 6277.
18. N. Wei, T. Lu, F. Li, W. Zhang, B. Ma, Z. Lu and J. Qi, Appl. Phys. Lett., 2012, 101, 061902.
19. Y. Z. Shen, T. T. Zou, S. Zhang and L. Z. Sheng, ISIJ Int., 2013, 53, 304.
20. J. W. Nowok and J. P. Kay, J. Mater. Sci. Lett., 2001, 20, 1331.
21. M. Venkatesan, C. B. Fitzgerald and J. M. D. Coey, Nature, 2004, 430, 630.
22. R. Monnier and B. Delley, Phys. Rev. Lett., 2001, 87, 157204.
23. H. J. Tromp, P. van Gelderen, P. J. Kelly, G. Brocks and P. A. Bobbert, Phys. Rev. Lett., 2001, 87, 016401.
24. D. P. Young, D. Hall, M. E. Torelli, Z. Fisk, J. D. Thomson, H. R. Ott, S. B. Oseroff, R. G. Goodrich and R. Zysler, Nature, 1999, 397, 412.
25. I. S. Elfimov, S. Yunoki and G. A. Sawatzky, Phys. Rev. Lett., 2002, 89, 216403.
26. S. Murakami, R. Shindou, N. Nagaosa and A. S. Mishchenko, J. Phys. Soc. Jpn., 2002, 71, 309.
27. K. Jayanthi and S. V. Manorama, J. Mater. Chem. C, 2014, 2, 10322.
28. S. F. Wuister, C. D. M. Donega and A. Meijerink, Phys. Chem. Chem. Phys., 2004, 6, 1633.
29. A. Birkel, K. A. Denault, N. C. George, C. E. Doll, B. Héry, A. A. Mikhailovsky, C. S. Birkel, B.-C. Hong and R. Seshadri, Chem. Mater., 2012, 24, 1198.
30. M. Veith, S. Mathur, A. Kareiva, M. Jilavi, M. Zimmer and V. Huch, J. Mater. Chem., 1999, 9, 3069.
31. N. K. Nasikas, S. Sen and G. N. Papatheodorou, Chem. Mater., 2011, 23, 2860.
32. A. Potdevin, G. Chadeyron, V. Briois, F. Leroux, C. V. Santilli, M. Dubois, D. Boyer and R. Mahiou, Dalton Trans., 2010, 39, 8706.
33. K. J. D. Mackenzie and R. H. Meinhold, J. Mater. Chem., 1994, 4, 1595.
34. R. Ji, W. Yin, C. Fang and Y. Zeng, J. Mater. Chem. C, 2013, 1, 1763.
35. Y. S. Lin, Y. H. Tseng, R. S. Liu and J. C. C. Chan, J. Electrochem. Soc., 2007, 154, 16.
36. V. Babin, V. Gorbenko, I. Kondakova, T. Kärner, V. V. Laguta, M. Nikl, S. Zazubovich and Y. Zorenko, J. Phys. D: Appl. Phys., 2011, 44, 315402.
37. J. Li, D. Li, X. Hong, L. Wang, K. Zhao, J. Li, Y. Bai and T. Lia, Chem. Commun., 2004, 982.
38. H. M. Lim, S. K. Batabyal, S. S. Pramana, L. H. Wong, S. Magdassi and S. G. Mhaisalkar, Chem. Commun., 2013, 49, 5351.
39. V. Singh, R. P. S. Chakradhar, J. L. Rao and H. Y. Kwak, Appl. Phys. B: Lasers Opt., 2010, 98, 407.
40. D. T. Mackay, C. R. Varney, J. Buscher and F. A. Selim, J. Appl. Phys., 2012, 112, 023522.
41. F. A. Selim, C. R. Varney, M. C. Tarun, M. C. Rowe, G. S. Collins and M. D. McCluskey, Phys. Rev. B: Condens. Matter Mater. Phys., 2013, 88, 174102.
42. C. R. Varney, S. M. Reda, D. T. Mackay, M. C. Rowe and F. A. Selim, AIP Adv., 2011, 1, 042170.
43. W. L. Warren, K. Vanheusden, M. A. Rodriguez, C. H. Seager, D. R. Tallant, P. D. Rack, P. H. Holloway, B. K. Wagner, C. J. Summers and P. N. Yocom, Appl. Phys. Lett., 1997, 70, 478.
44. D. L. Dexter and J. H. Schulman, J. Chem. Phys., 1954, 22, 1063.
45. L. T. Su, A. I. Y. Tok, F. Y. C. Boey, X. H. Zhang, J. L. Woodhead and C. J. Summers, J. Appl. Phys., 2007, 102, 083541.
46. J. M. Ogiegło, A. Zych, K. V. Ivanovskikh, T. Jüstel, C. R. Ronda and A. Meijerink, J. Phys. Chem. A, 2012, 116, 8464.
47. D. J. Robbins, J. Electrochem. Soc., 1979, 126, 1550.
48. C. R. Varney, D. T. Mackay, S. M. Reda and F. A. Selim, J. Phys. D: Appl. Phys., 2012, 45, 015103.
49. R. Kasuya, T. Isobe, H. Kuma and J. Katano, J. Phys. Chem. B, 2005, 109, 22126.
50. G.-H. Lee and S. Kang, J. Lumin., 2011, 131, 2606.
51. Y. Köseoğlu, H. Yildiz and R. Yilgin, J. Nanosci. Nanotechnol., 2012, 12, 2261.
52. I. Rhee, J. Korean Phys. Soc., 2009, 54, 1721.
53. J.-P. Fan, X. L. Li, Z. H. Quan and X. H. Xu, Appl. Phys. Lett., 2013, 102, 102407.
54. W. Q. Zou, C. N. Ge, G. Venkataiah, H. L. Su, H. S. Hsu, J. C. A. Huang, X. C. Liu, F. M. Zhang and Y. W. Du, J. Appl. Phys., 2012, 111, 113704.
55. R. M. Kadam, M. K. Bhide, M. D. Sastry, J. V. Yakhmi and O. Kahn, Chem. Phys. Lett., 2002, 357, 457.
56. D. Karmakar, S. K. Mandal, R. M. Kadam, P. L. Paulose, A. K. Rajarajan, T. K. Nath, A. K. Das, I. Dasgupta and G. P. Das, Phys. Rev. B: Condens. Matter Mater. Phys., 2007, 75, 144404.
57. S. V. Chong, K. Kadowaki, J. Xia and H. Idriss, Appl. Phys. Lett., 2008, 92, 232502.
58. X. W. Wang, X. P. Gao, G. R. Li, L. Gao and T. Y. Yan, Appl. Phys. Lett., 2007, 91, 143102.
59. N. F. Djaja and R. Saleh, Mater. Sci. Appl., 2012, 3, 245.
60. R. V. Sagar and S. Buddhudu, Spectrochim. Acta, Part A, 2010, 75, 1218.
61. S. Colis, A. Bouaine, G. Schmerber, C. Ulhaq-Bouillet, A. Dinia, S. Choua and P. Turek, Phys. Chem. Chem. Phys., 2012, 14, 7256.
62. X. Y. Xu, C. X. Xu, J. Dai, J. G. Hu, F. J. Li and S. Zhang, J. Phys. Chem. C, 2012, 116, 8813.
63. G. Blasse and A. Bril, Appl. Phys. Lett., 1967, 11, 53.
64. L. R. Shah, B. Ali, H. Zhu, W. G. Wang, Y. Q. Song, H. W. Zhang, S. I. Shah and J. Q. Xiao, J. Phys.: Condens. Matter, 2009, 21, 486004.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra01425a

---