Background

The transition from bulk or micron to nanosize domain greatly affects a material, altering, for example, its mechanical, optical, and electrical properties [1,2,3,4,5,6]. These changes are mostly attributed to the size and associating non-equilibrium structure. An example is the unique phosphorescence and emission properties achievable by nanoparticles [2, 7]. The emission properties of nanophosphors can be modulated by doping, in addition to the quantum confinement effects described [8, 9]. The charge valence of a dopant and the site in the structure that it occupies generally affect the emission properties of a phosphor. Dopants can often be located in a nanophosphor at sites (e.g., on a particle’s surface) other than the usual sites in a micron-sized phosphor. Thus, the surfaces of nanoparticles become important sites for dopants that do not normally occupy such sites in bulk or micron systems. The changes in the emission behavior have been reported due to the occupation site, which is associated with other defects [8, 9].

Pure MgAl2O4 has an intrinsic defect of an Mg2+ vacancy, VMg′′, which is the center of a red emission at 720 nm. Strong blue emission is observed from single crystals of Ti-doped MgAl2O4; the disappearance of the red emission is attributed to charge compensation through the addition of Ti4+ [10, 11]. However, in our previous work, we found that Ti-doped micron-sized MgAl2O4 powder heat-treated in air produced a white emission [12]. The difference was explained via the occurrence of red and green emissions in addition to the blue observed from Ti-doped MgAl2O4 single crystals. Our previous work [12] also simulated the mechanism for the red emission via intrinsic Schottky defect associate, (VO ··–VMg′′)x. The present work reports the visual observation of extrinsic Schottky defects on the surface of Ti-doped MgAl2O4 nanopowder and relates it to the difference in the emission spectra between micron and nano systems.

Methods

Mg(NO3)2·6H2O (Mg nitrate; 2.46 g, Aldrich), Al(NO3)3·9H2O (Al nitrate; 7.246 g, Aldrich), CO(NH2)2 (urea; 5.231 g, Aldrich), and C10H14O5Ti (Ti oxy-acetyl-acetonate; 0.1 g, Aldrich,) were used as starting materials for the synthesis of Ti-doped nano MgAl2O4. Mg nitrate and Al nitrate were used in a 1:2 molar ratio in the synthesis, and 2 mol% Ti doping was provided by Ti oxy-acetyl-acetonate. The starting materials were dissolved in deionized water, and the mixture was homogenized by stirring, before the water was evaporated on a hot plate. The remaining mixture was placed in an alumina crucible and fired at 500 for 1 h in air.

The phases of the synthesized nanopowder were analyzed by X-ray diffractometry (XRD; Rigaku), and photoluminescence properties were measured by fluorescence spectrophotometry (PSI, PL Darsa pro-5000 system) using monochromated 260 and 360 nm light from a Xe lamp. Powder morphology and size were observed by high-resolution transmission electron microscopy (TEM; JEOL, JEM-2100F). Images of the Ti dopant and the Al vacancies were also obtained by high-resolution scanning TEM (HR-STEM; JEOL, JEM-2100F).

First principles density functional theory (DFT) calculations were performed based on the generalized gradient approximation of Perdew–Burke–Ernzerhof and projector-augmented plane wave pseudopotentials implemented in the Vienna ab initio simulation package (VASP) [13,14,15] with an energy cutoff of 500 eV and a self-consistency field convergence of 10−5 eV. The stabilities of various defect associates were examined by calculation to investigate their dependence upon the positions of the dopants and vacancies and the relative distances between them.

The surface energy of the (100) surface plane in the MgAl2O4 crystal was calculated; its variation with the Ti site was also examined. The unit cell—whose crystal structure was previously optimized allowing full relaxation of the lattice parameter, crystal shape, and atomic positions—was expanded to a 4 × 1 supercell. {100} surfaces were created by inserting a vacuum slab inside the supercell. Insertion position of the vacuum slab which is of the size of 2 × 1 supercell was varied to examine change of the surface formation energy with the distance between the surface and the Ti dopant. Surface termination with 50% of Mg layer was mainly considered because this was found to be the most stable (1 0 0) surface of MgAl2O4.

Results and Discussion

Figure 1 shows the XRD pattern for Ti-doped MgAl2O4 nano phosphors with TEM graphs. Figure 1a clearly confirms that MgAl2O4 was synthesized at 500 °C by the combustion method, given its similarity to the JCPDS XRD pattern for pure MgAl2O4. The broad peaks indicate the presence of nanocrystallites and are related to the particles of < 20 nm shown in Fig. 1b. In contrast, the Ti-doped micron-sized MgAl2O4 of our previous work [12] shows high crystallinity, attributed to the high-temperature treatment of the MgAl2O4 powder (1300 °C for 2 h).

Fig. 1
figure 1

a XRD pattern for Ti-doped nano MgAl2O4 powder synthesized at 500 °C 1 h with reference peaks of MgAl2O4 from JCPDS and b, c TEM images of the powder

Full size image

The photoluminescence emission spectra of Ti-doped MgAl2O4 show white emission at 260 nm excitation (Fig. 2a) for nano- and micron-sized samples synthesized at 500 °C for 1 h and 1300 °C for 2 h in Fig. 2b, c, respectively. However, the two emission bands result in slightly different colors: that of the nanopowder synthesized at 500 °C is blue-shifted relative to that of the micron-scale powder prepared at 1300 °C. The blue emission of Ti-doped MgAl2O4 single crystals is attributed to Ti4+ in Al (octahedral) sites, which was the only form of Ti ions in the single crystals [10, 11]. However, Ti-doped MgAl2O4 micron-sized powder was shown to have both Ti3+ and Ti4+ equally occupying both Al (octahedral) and Mg (tetrahedral) sites [12].

Fig. 2
figure 2

Photoluminescence of Ti-doped MgAl2O4: a 260 nm excitation, b nano, and c micron [12] powders synthesized at 500 °C, 1 h, and 1300 °C, 2 h, respectively

Full size image

Figure 3a shows an HR-STEM image taken near the surface of Ti-doped nano MgAl2O4. The magnified image in Fig. 3b shows the distance between the arrays to be 0.2057 nm, which matches well with the (400) planar distance of MgAl2O4 (0.202 nm). It shows that the atomic arrangement left a relatively dark vacancy among the spots (see arrows in Fig. 3a, b). The minor brightness at the vacancy might have originated from atoms in the lower layers. The defect point is also identified in the contrast intensity plot in the inset, which shows the contrast peaks for the atoms inside the red box of Fig. 3b. The vacancy is clearly shown by the low contrast-intensity of the fifth site from left. To identify the vacancy site, we performed the Fourier transformation of the image in Fig. 3a and found that the beam axis is close to [001] (inset, Fig. 3a). It is noted from the [001] projected view of a MgAl2O4 crystal that Mg atoms are located independently in the (004) plane, whereas Al and O atoms appear overlapped together in the same plane. Thus, if the fluctuation shown in the contrast intensity is only due to the constituent atoms on the plane, it is more probable for the vacancy to originate from a vacant Al site rather than a vacant Mg site.

Fig. 3
figure 3

HR-STEM images of Ti-doped nano MgAl2O4; a. The image shows Al3+ vacancy and Ti4+ dopant in the vicinity with Fourier transformation of the image. b The magnified image in red box of (a) and Al3+vacancy is revealed from STEM image with a corresponding contrast intensity, inset of (b). c The contrast intensity, inset of (c), confirms that Ti4+ dopant occupies an Al site. The arrows indicate the locations of Al vacancy and Ti in Al site, respectively

Full size image

In Fig. 3c, the lattice point indicated by an arrow in the red box is much brighter than the others. Considering that Mg and Al atoms cannot be distinguished by z-contrast due to their similar atomic numbers and that it is difficult to detect oxygen atoms owing to its low atomic number, this brighter point is concluded to be due to the Ti dopant. The corresponding contrast intensity plot (inset, Fig. 3c) emphasizes the brighter spot, indicating the presence of an element of higher atomic number, definitely Ti in this system. Ti in an Al site causes displacement error, because its charge valence and ionic radius are different from those of Al3+. The brighter atom in the figure appears larger than the others, in agreement with the larger effective ionic radii of Ti3+ (0.081 nm) and Ti4+ (0.0745 nm) in comparison to that of Al3+ (0.0675 nm) [16]. The effective ionic radius of Mg2+ is reported as 0.086 nm, which is larger than those of the Ti ions. Thus, we concluded that the defects shown in Fig. 3 (i.e., Fig. 3b, c) are VAl′′′ and TiAl ·, respectively, expecting that Ti4+ ions of a smaller size (0.0745 nm) have more chance to take vacant Al sites than Ti3+ (0.081 nm).

Figure 4a shows the change in the surface energy of a Ti-doped MgAl2O4 perfect crystal calculated with respect to the position of the dopant. The surface energy, which may be a major factor influencing the formation energy of a nanosystem, decreases as Ti approaches the surface, indicating that the crystal is more stable when Ti is closer to the surface. The result indicates a common trend for Ti at an Al site and Ti at an Mg site; however, the dopant is more stable at an Mg site than at an Al site. This is attributed to the larger effective ionic radius of Mg2+(0.086 nm) than either Ti4+(0.0745 nm) or Al3+(0.0675 nm) [16].Thus, the trend is more probable when Ti-doped MgAl2O4 has a high crystallinity. However, it might not be always true for a nanosystem of a low crystallinity, at least near the surface region.

Fig. 4
figure 4

a Stabilities of Ti in Mg or Al vacancy site on the surface of MgAl2O4: the red circle shows surface energy of Ti in Al site and the black dot is that of Ti in Mg site and b the binding energy of defect associates, TiAl ·–VAl′′′ as a function of distance

Full size image

DFT calculations were also run to investigate the positioning of the Ti dopant and the Al vacancy. The calculated energy of a Ti1Mg15Al31O64 crystal, spinel containing a Ti dopant (TiAl ·) and an Al vacancy (VAl′′′), increases as the dopant and vacancy are moved apart as shown in Fig. 4b. Therefore, greater stability is achieved when the two defects are close to each other and form defect associates such as (TiAl ·–VAl′′′)′′ that are responsible for the blue emission. This result is attributed to the structural stability and the Coulombic force between the two point defects. However, a compromise emerges between these factors and the configurational entropy to stabilize the system at elevated temperature, which results in the two defects being spaced 2–3 atoms apart, as shown in Fig. 3a.

In general, the formation energy of an Al or Mg vacancy is much lower (~4.5 eV) than that of an oxygen interstitial (~ 7.0 eV) in MgAl2O4 [17, 18]. Also, the formation energy of intrinsic Schottky defects for MgAl2O4 (4.15 eV/defect) is much lower than those for individual oxides, MgO (7.7 eV) and α-Al2O3 (4.2–5.1 eV). According to Coulomb estimates, the defect association energies of extrinsic Schottky pairs are smaller than those of intrinsic Schottky pairs in various ionic systems [19]. When Ti-doped MgAl2O4 is synthesized chemically by combustion method via nucleation and precipitation process, as for the nano system of this study, instead of by solid-state diffusion, the formation of defects and defect associates, including O2− vacancies which are commonly observed in oxide ceramics, would be significantly facilitated on the particle surfaces. Overall results indicate that the defect associates, i.e., (TiAl ·–VAl′′′)′′, prevail on the surface of Ti-doped MgAl2O4 nanopowders, causing the blue shift in the white emission of nano-powders compared to that of micron powders.

Conclusions

The substitution of Ti in the Al sites of MgAl2O4 was observed by HR-STEM. An Al vacancy and Ti dopant were detected near the surface of Ti-doped nano MgAl2O4. These observations demonstrate the presence of Ti4+ in Al sites. The blue shift relative to the spectrum of the micron-scale system is attributed to the presence of more Ti4+ ions in Al sites at the surface. It would be energetically more favorable for Ti4+ ions to take Mg sites in the spinel structure. However, Ti4+ ions tend to take Al sites in the Ti-doped nano MgAl2O4. This difference in the luminescence of the nanosystem arose from its low crystallinity that is resulted from the low processing temperature.