Effects of Varying Alx Moles on the Structure and Luminescence Properties of ZnAlxO1.5x + 1:0.1% Tb3+ Nano Phosphor Prepared using Citrate sol-gel Method

. Un-doped and ZnAl x O (1.5x + 1) :0.1% Tb 3+ (ZAOT) nano-powders were synthesized via citrate sol-gel method. The Al x moles were varied in the range of 0.2 5 ≤ x ≤ 5.0. The X-ray powder diffraction (XRD) data revealed that for the x < 1.5, the prepared samples crystal structure consists of mixed phases of the cubic phase of cubic ZnAl 2 O 4 . This was confirmed by the Raman and Fourier-Transform Infrared (FTIR) vibrational spectroscopy. Scanning electron microscopy (SEM) showed that varying Al x moles influences the morphology while Transmission electron microscopy (TEM) shows the dual morphology at x < 1.5. The photoluminescence (PL) revealed intense and distinct emissions attributed to both the host and Tb 3+ transitions. The emission intensity highly depends on the Al x moles. The International Commission on Illumination (CIE) colour chromaticity showed that the emission colour could be tuned by varying the Al x moles. atoms available to completely form ZnAl 2 O 4 . This results suggest that the transition from dual ZnAl 2 O 4 /ZnO mixed phases to ZnAl 2 O 4 single phase is somewhere around x = 1.5. Thus, the mixed phase/s in ZAOT highly depends on the Al x moles. The XRD patterns of the doped (x = 2.0) and undoped (x = 2.0*) are similar, signifying a successful incorporation of Tb 3+ within the ZnAl 2 O 4 crystal lattice.


Introduction
In the past few decades, metal oxides (MO) have enticed many researches in different field due to their unique physical and chemical properties [1]. MO are on the cutting edge of technology in various field of science such as water splitting [2][3][4], energy storage and conversions [5,6], photoluminescence (PL) [7][8][9] and transparent conducting oxide [1,10,11]. This is because of their band gap located directly at the Г point which was determined theoretically and experimentally to be around few electron volts (eV) [12][13][14]. The MO denoted by the AB2O4 formula are usually referred to as spinel family group, which contains around 120 compound members with ZnAl2O4 (known as zinc aluminate or gahnite) being one of the prominent species in the PL field [1,15]. The ZnAl2O4 normally crystallizes in the cubic crystal structures of the 3 ̅ space group, where Zn and Al respectively occupies the tetrahedral (Td) and octahedral (Oh) sites in the crystal matrix while oxygen atoms occupies the Wyckoff 32e position located at (u, u, u), where u ≈ 0.25 [1,16,17]. The positioning of the atoms or the structural configuration of the ZnAl2O4 depends on synthesis parameter such as the temperature [18,19] or concentration of the constituent elements [9,15,20]. These can lead in the crystallisation of metal ion in the normal (ZnAl2O4) configuration as described above (where Zn occupies Td and Al occupies Oh sites) or in an inversion (Al[ZnAl]2O4) configuration where the Td is now fully occupied by Al ion and the Oh sites is half-sheared by Al and Zn ions. The spinel can also crystallise in between the normal and inversion configurations where ions occupy those sites spatial. Raman spectroscopy has been an wonderful technique of detecting such behaviours [17,20,21].
Various studies have been conducted on ZnAl2O4 in order to optimize and tune the emission colour.
The intra-ET within Tb 3+ ions and between Tb 3+ to Mn 2+ , which were described as follows Tb 3+ ( 5 D3) + Tb 3+ ( 7 F6) → Tb 3+ ( 5 D4) + Tb 3+ ( 7 F2,1,0) and Tb 3+ → Mn 2+ [23], respectively. The intra-ET happened via overlapping of the Tb 3+ band and ET between Tb and Mn happened via the absorption of excitation energy at 350 nm from Tb 3+ by the 654 nm emission peak of Mn 2+ transitions. Most of the reported work on ZnAl2O4 focuses on the dopants and synthesis conditions. The effects of the metal ions (i.e. Zn and Al) moles in ZnAl2O4:0.1% Tb 3+ has not been investigated in detail as far as the structural and optical properties are concern. Few studies were reported in the field such as that of catalysis [30,31]. ground into fine powder samples using the pestle and mortar.

Characterization
The powder samples were analyzed with the Bruker D8-Advance powder X-ray powder diffraction (XRD) with a CuKα (1.5405 Ǻ) radiation, Witec alpha 300 RAS+ (Raman spectroscopy) using the 532 nm at 2 mW and Bruker Alpha platinum-ATR Fourier-transform infrared spectroscopy (FTIR) in the range of 400 to 1000 cm -1 was used to analyze the crystal structure, phase, functional group and the vibration bands. The structural phase from the XRD were identified using X'pert Highscore plus software and the relative phase quantification (%) were estimated using the Rietveld method. The particle shape, surface morphology and elementary constituents were characterized by the Zeiss Supra 55 scanning electron microscope (SEM) at 20 kV coupled with an energy dispersive X-ray spectroscope (EDS). Crystallite shape and sizes was analysed via the JEOL JEM 1010 transmission electron microscopy (TEM). The room temperature (RT) photoluminescence (PL) spectra and lifetime measurements were performed by the Hitachi F-7000 fluorescence spectrophotometer using a 150 W monochromatized Xenon lamp as an excitation source.

X-ray powder diffraction
The XRD patterns of the prepared samples are shown in Fig. 1  to the availability of enough Al atoms to form ZnAl2O4 single phase. As x increases beyond 1.5 the ZnAl2O4 phase diffraction intensity seems to be decreasing since the excess Al might possibly be incorporated within the ZnAl2O4 matrix, which is leads to the absence of alumina (Al2O3) related diffraction patterns. The excess Al within ZnAl2O4 influenced the diffraction angle (see Fig. 3 which subsequently affect the lattice constant. The stacked zoomed analysis of the ZnO (101) and ZnAl2O4 (311) displayed in Fig. 3 (b) shows that both peaks are shifting towards a lower diffraction angle as x increases up to 1.5 resulting in the expansion of the lattice constant [9,15]. On the ZnAl2O4 phase, this can be caused by the excess Zn 2+ into ZnAl2O4 crystal lattice, as for the ZnO phase, the lattice constant increases due to multiple incorporation of Al 3+ ion and/or Zn intrastitial within the ZnO matrix which intend pave the formation of single phase ZnAl2O4. When x ≥ 1.5, the (311) diffraction peak shifts towards higher diffraction angle signifying a decrease in lattice constant. Meaning that some of the Al 3+ (ionic radius = 0.53 Å [7,34]) are occupying the Zn 2+ (ionic radius = 0.74 Å [9,34]) site. This behaviour can be explained using the Vegard's law [35,36] which suggests that the substitution or replacement of a certain ion with the smaller or bigger ionic radius is expected to shrink or expand the lattice constant, respectively. The lattice constant of the ZnAl2O4 and ZnO were respectively estimated using the equation (1) The lattice constant as a function of Alx moles is depicted in Fig. 4 (a).
The lattice constant increase from x = 0.25 to 1 due to the excess of Zn atoms compared to Al atoms, the dramatic decrease to x = 1.5 is due to the completely removal of ZnO phase due to reasonable Zn:Al ratio to start forming ZnAl2O4 single phase. An increase in lattice parameters from x = 1.5 to 2.5 might be attributed to the full formation of stable ZnAl2O4 single phase. The observed decrease in lattice parameter at x > 2.5 might be due to Al 3+ replacing the Zn 2+ ions in ZnAl2O4 crystal lattice. The crystallite size for the ZnO and ZnAl2O4 were estimated from the dominant diffraction peaks ZnO (101) and ZnAl2O4 (311) using Scherrer's formula [40]. The estimated values are presented in Table   1. The crystallite size of the ZnAl2O4 as a function of Alx mole is presented on Fig. 4 (a). The graph of crystallite resembles a similar behaviour as that of the lattice constant. The strain on the ZnAl2O4 and ZnO phases was calculated from Williams and Hall method [41], and the values are presented in Table 1. Strain as a function of the Alx is shown in Fig. 4 (b). When 0.25 ≥ x > 1.0 the strain decreases due to multiple phases present in the ZAOT system and that it is easy to form both the ZnO and ZnAl2O4. At x = 1.5, the strain increases due to the formation of ZnAl2O4 single phase. At the decrease in strain above x > 1.5 is probably due to easy formation of ZnAl2O4 at these Alx moles. Above x ≥ 2.5, the strain increases linearly as more Al ions are being incorporated into the ZnAl2O4 matrix.

Raman Spectroscopy
Raman spectroscopy was employed to identify the molecular structure of the prepared samples and the obtained spectra are shown in Fig. 5. The results show that there are five Raman peaks located at 62, 101, 420, 440 and 660 cm -1 . The peaks at 101 and 440 cm -1 are attributed to the ZnO phase specifically from the 2 and 2 ℎ ℎ resulting from the Zn and O atomic oscillations (see Fig. 2 (a)) [42,43]. These peaks disappear at x > 1.5, which is a clear confirmation that they are indeed from ZnO phase as it was suggested by the XRD results in Fig. 1. The Raman peaks at 420 and 660 cm -1 are due to ZnAl2O4. It is generally accepted that Raman modes within the medium frequency range (300-600 cm -1 ) modes are due to the MO6 octahedral while the high range frequency (> 600 cm -1 ) modes are due to the MO4 tetrahedral. Thus, the 420 and 660 cm -1 bands can be attributed to the Eg and F2g mode ascribed to the Oh and Td site [20,21], respectively. These observed two peaks from ZnAl2O4 is in good agreement with the previous reports [16,17,20,21]. Previous studies [16,44] have shown that Eg and F2g modes of Raman spectrum of the inverse spinel occurred in the 350 -400 and 490 -640 cm -1 region. Zn0.9Cu0.1Al2O4 and Zn0.9Ni0.1Al2O4 have also shown a lower frequency Raman shift located in the 395-410 and 630-650 cm -1 which was attributed to appearance of minor inverse spinel [16,20]. Therefore, these results suggest that the prepared ZnAl2O4 spinel has crystallised in the normal configuration. The Raman peak located at 62 cm -1 can be attributed to the Zn interstitial [43] and/or the presents of Tb 3+ within the ZnO or ZnAl2O4 hosts matrix. The Raman confirms or revealed the structural transition from multiphase ZnO/ZnAl2O4 to ZnAl2O4 single phase. Note that the intensity of the Raman peak at 62 cm -1 for the x = 2.0 (un-doped) sample is lower than the doped sample, which suggest that the Zn interstitial sites are more pronounced for the doped compared to the undoped sample [43]. Generally, the Raman and XRD result supports each other.

Fourier-transform infrared spectroscopy
The FTIR spectra of the sample are depicted in Fig. 6. It is well known that infrared spectra of spinels are characterized by absorption band in the range between 400 -700 cm -1 [16,21].  [47][48][49]. These peaks are through-out the entire ZAOT samples because they all contain ZnAl2O4 phase as observed on the XRD and Raman results. The additional absorption band at 803 cm -1 , which is clearly noticeable from x ≥ 3 might be also from the ZnAl2O4 due to the Al-O bond arising from Al in the Td site of ZnAl2O4 as suggest on the lattice constant and strain analysis.

Energy-dispersive X-ray spectroscopy
The EDS technique was deployed to analyse the chemical composition of the un-doped and ZAOT nano-powders and the spectra for the selected is shown in Fig. 7. The expected elements namely Zn, Al, and O are observed in all spectra. Except for the element Tb, due to its low doping concentration.
The additional peak of carbon (C) observed at the lower energy is due to the sample coating during the sample preparation for the EDS analysis.

Scanning Electron Microscopy
The SEM micrographs of the selected samples are shown in Fig. 8. The micrograph displayed on Fig.   8 (a) shows the x = 0.25, which is a dual morphology consisting of rods and randomly disturbed irregular crystallites. Arguing from the XRD and Raman results, the morphology of the spherical particles can be attributed to the ZnAl2O4 since it is observed in all sample as shown in Fig. 8 (a) -(d). Therefore, the rods structures are possibly due to the ZnO as they are only present at the lower Alx. It is also clear that Tb 3+ doping and varying the Alx moles does not influence the crystallites agglomeration.

Transmission Electron Microscopy
The prepared samples were further analysed by the TEM to investigate the phase, crystal size and shapes. The TEM image of the selected samples is depicted in Fig. 9. In general, the TEM images confirm what has been observed in XRD, Raman and SEM in terms of the transition from dual morphology (or phase) to a single phase of ZnAl2O4. The average particle size of the investigated nanopowders are in nanometric regime (i.e. below 30 nm), which agrees very well with the XRD results.
In addition, the TEM images shows the presence of the hexagonal like-particles (indicated by the red arrows), which are certainly attributed to the ZnO. This results suggest that there might still be the presence of hexagonal ZnO phase at x = 2.0. However, the ZnO phase is totally vanished at higher Alx moles (e.g. x = 4.0). Thus, TEM results also suggest the Alx moles influenced the phase present on the prepared samples.

Photoluminescence spectroscopy
The ZnAlxO1.5x + 1:0.1% Tb 3+ (x = 2.0) sample was excited at various exitation wavelength in a range of 200 -280 nm and the obtained emission spectra is illustrated in Fig. 10  showed that the luminescence active defects within ZnAl2O4 decreases dramatically after the annealling temperature of 600 o C. The emission intensity of the most intense peak at 545 nm as a function of excitation wavelength is shown in Fig. 10 (b), which shows a Gaussian behaviour. The Gaussian fit results shows that the optimum excitation wavelength is at 225 nm. This might probably serve as a good reason why a similar excitation wavelength of 228 nm was used to excite the ZnAl2O4:4 mol. % Tb 3+ system [51]. The RT PL excitation and emission spectra of the x = 2.0* and ZAOT series are presented in Fig. 11.
The excitation spectra of the un-doped (x = 2.0*) and doped (x = 2.0) displayed on Fig. 11 (a) were measured when monitoring the 545 nm emission which shows the presences of a broad excitation band located at 225 nm. This excitation band at 225 nm can be attributed to the spin allowed transition (∆S = 0) from the 4F 8 (ground state) → 4F 7 5d 1 of the Tb 3+ ions [27]. The emission spectra show the presences of eight peaks which were similar to those observed in Fig. 10 (a). To further investigate the emissions originating from the un-doped, the emission spectra of the doped x = 2.0 and un-doped x = 2.0 samples where normalized as shown in Fig. 11 (b). Fig. 11 (c) shows the deconvolution emission spectra of the un-doped (x = 2.0*) sample. From Fig. 11 (b), the excitation band is most likely to be due to the band-to-band transition of AlO6 anion grouping in ZnAl2O4 [52]. From Fig. 11 (c), the observed emission bands from the host at 407 nm can ascribed to the intrinsic intraband gap defects, such as oxygen vacancies (Vo*) [52]. The 458 nm peak can be ascribed to the second order of 225 nm excitation wavelength. Based on the TEM results for the x = 2.0 and x = 2.0* (and all samples with ZnO phase), it is therefore reasonable to also attribute the 545 nm emission peak to the deep level defect emission within ZnO nanorods [53]. Since the ZnO might still be present although at the small proportions. The excitation and emission spectrums of the ZAOT series is shown in Fig. 11 (d). The results show that there were no new emission peaks which were observed except the ones discussed in Figs. 10 and 11 (a). It is important to emphasize that the 458 nm emission peak is attributed to both the contribution from Tb 3+ ( 5 D3 → 7 F2 transition) and second order emission. The most intense emission peak at 545 nm is attributed to a magnetic dipole transition which hardly varies in the presents of crystal field strength, while the 489 nm ascribed to the electrical dipole which is sensitive to the local environment and depends on the symmetry of the crystal field [27,54]. Fig. 11 (e) shows the emission intensity of the 545 nm emission peak as a function of Alx moles. The result shows a dramatic increase in the emission intensity when x > 1.5, which can be attributed to the multiphase change to single phase on the prepared phosphors. Furthermore, the results imply that the presence of the ZnO at high quantities quenches the luminescence. In order to properly deduce the transitions within the band gap, the most intense sample x = 2.5 (ZnAl2.5O4.75:0.1% Tb 3+ ) was further examined by monitoring excitation of each emission bands. Fig.   12 (a) shows the obtained excitation for each individual emission peak. Fig. 12 (b) shows an additional excitation band located at 308 nm when monitoring an emission peak at 621 nm (insert shows the zoomed version for the wavelength range of 365 -650 nm. The new absorption peak might be from the transition of Tb 3+ arising from the ground state to the 5 D4 level. All the observed emission peaks are similar to those discussed in Fig. 10 and 11. The proposed pathway channels for the excitations and emissions discussed in this paper are shown in Fig. 13. The energy level location of Tb 3+ within ZnAl2O4 was predicted based on the Dorenbo's diagram [26,55]. Note that the deep level defect emission within ZnO at 545 nm [58] is not shown on the excitation and emission pathway mechanism because it is only present on the samples with ZnO phase.

Fig. 12
Excitation and emission spectra of the x = 2.5 sample Fig. 13 The proposed excitation and emission pathways mechanism for ZnAl2O4:0.1%Tb 3+ series

Lifetime
Fig. 14 shows the RT phosphorescence lifetime of the prepared ZAOT series. The lifetime was taken when monitoring the 545 nm emission wavelength and 225 nm excitation wavelengths. Fig. 14 (b) shows the natural logarithm of the lifetime intensity as a function of time. All the samples where fitted using the first order exponential decay presented on equation 3 [26,56].
where represents the phosphorescent intensity, is the fitting parameters, is the time of measurement and 1 time values. The fitting parameter and decay times are presented in Table 2.

Colour chromaticity
The International Commission on Illumination (CIE) chromaticity diagram with vertex region of different colour and the co-ordinates of the luminescent material are displayed in Fig. 15. The colour co-ordinates of the prepared sample were calculated using the CIE co-ordinate calculator software [59].
The colour co-ordinates for the x = 2.0 sample excited at different excitation wavelength is shown in Fig. 15 (a). The results show that the excitation wavelength influence the emission colour. Fig. 15 (b) compare the emission colour co-ordinates of the un-dope and Tb 3+ dope for x = 2.0 sample. The undoped sample shows a blue colour emission while the Tb 3+ doped shows the green emission, which clearly shows that doping influences the emission of the host material. The colour co-ordinates for the ZAOT series is shown in Fig. 15 (c) and the (x;y) values are also presented in Table 2. The results clearly show that varying the Alx moles influence the emission colour. The emission colour could be tuned from greenish to bluish. Fig. 15 (d) shows the CIE colour co-ordinates for the x = 2.5 sample when excited at 225 and 308 nm. The results show that the emission colour highly depends on the excitation wavelength.   The Spectra for the x = 2.0* (undoped) and ZnAlxO1.5x + 1:0.1% Tb3+ (0.25 ≤ x ≤ 5.0) Figure 6 The FTIR spectra for the x = 2.0* and ZnAlxO1.5x + 1:0.1% Tb3+ (0.25 ≤ x ≤ 5.0)  Excitation and emission spectra of the (a) un-doped (x = 2.0*) and doped (x = 2.0), (b) normalized of undoped and doped, (c) deconvolution of the un-doped emission, (d) ZnAlxO1.5x + 1:0.1% Tb3+ (0.25 ≤ x ≤ 5.0), and (e) emission intensity of the peak located at 545 nm as a function of Alx moles.

Figure 12
Excitation and emission spectra of the x = 2.5 sample Figure 13 The proposed excitation and emission pathways mechanism for ZnAl2O4:0.1%Tb3+ series