A thorough spectroscopic study of luminescent precursor solution of Y₃Al₅O₁₂:Tb³⁺: influence of acetylacetone

Abstract

Undoped and Tb³⁺-doped precursor solutions of Y₃Al₅O₁₂ (YAG) have been prepared by the sol–gel process using alkoxide precursors stabilized by a chelating agent: acetylacetone (acacH), with various complexation ratios R_C = 0, 1, 2 and 3 defined as the ratio between [Al(OPrⁱ)₃] and [acacH]. Combining in situ UV-visible investigation of the hydrolysis process of these sols with X-ray diffraction study on the powders extracted after water addition to the sols, drying and subsequent heating, an optimal R_C value has been determined. It leads to sols stable over a long period of time and resulting in a pure YAG phase. Structural and optical properties of Tb³⁺-doped precursor sols characterized by R_C = 0 (as a reference for the influence of acacH) and R_C = 1 (determined optimal value) were studied by means of X-ray Absorption Spectroscopy (XAS) and photoluminescence, respectively. Y K and Tb L₃ edges XAS results show the presence of double heterometallic Y or Tb/Al alkoxides for the sol with R_C = 1 involving pre-nuclei cluster with YAG structure. Emission and excitation spectra as well as decay curves were recorded and compared. Results revealed that the stabilized sol is much more efficient upon UV excitation than its unmodified counterpart thanks to an efficient energy transfer from acetylacetonate groups to active ions. Upon a 485 nm excitation, acac-modified samples remain the most efficient ones. It could be related to the dissimilar Tb local environments of the sols suggested by both UV-visible and photoluminescence excitation results.

---

Introduction

Metal–organic precursors such as metal alkoxides are very powerful precursors allowing, via soft chemistry routes, a wide range of functional ceramics to be obtained.^1–3 In particular, the alkoxide-based sol–gel process has shown potential advantages in terms of element homogeneity and versatile control of size and morphology of nanoparticles.^2,4 Indeed, metal alkoxides (M–OR) produce via hydrolysis–condensation reactions oligomeric structures, including homo or heterometallic oxo-bridges (M–O–M or M′–O–M). Depending on the nature of the ligands surrounding the metal M, hydrolysis and condensation steps can be controlled in order to favour a specific polymerization process. This tailoring permits to orientate structural and morphological properties of the target material.^1–3

Hence, chemical additives are frequently employed to better control the hydrolysis–condensation process^5–8 in order to obtain materials with tuned properties. A literature survey shows that most commonly used chemical modifiers are organic acids, e.g. acetic and malic acids,^6,7,9,10 and β-diketones or allied derivatives.^3,5 These complexing ligands prevent or at least slow down the condensation step and act as surface protecting entities. As a result, complexation strongly affects not only synthesis parameters such as gelation time, but also final morphology of the as-prepared solid such as porosity, particle shape and size, allowing the fine tuning of nanostructural feature of powder.^11,12 For example, their use allows for obtaining monodisperse non-aggregated nanoparticles.¹³

Yttrium aluminium garnet (YAG or Y₃Al₅O₁₂) is one of the most important matrices in optical devices. In particular, when doped with Ce³⁺ or Tb³⁺, it represents a very efficient compound for solid-state lighting.^14–17 Thanks to a chloride-modified alkoxide-based sol–gel route, we synthesized both powders and films at temperatures lying from 800 °C to 900 °C (ref. 18 and 19) which is much lower than that required by traditional solid-state preparation route.²⁰ Furthermore, the very good rare-earth distribution obtained in the matrix through these synthesis conditions leads to better photoluminescence properties compared to materials prepared by other processes.⁴ In recently published papers,^19,21 we investigated the influence of acetylacetone on the structural, morphological and optical properties of YAG:Tb³⁺ powders. In particular, Small Angle X-ray Scattering (SAXS) and photoluminescence techniques have shown that acac-modified powders are characterized by an open morphology resulting from monomer–cluster aggregation process and exhibit, for powders heated between 400 and 900 °C, a significant improvement of the luminescence intensity upon UV radiation compared to samples prepared free of acac. These results strongly suggest that acacH plays a role since the early stage of the synthesis that is the precursor sol formation.

In this framework, this paper deals with the structural organization and optical properties of Tb³⁺ doped unmodified and acac-modified sols, molecular precursors of the YAG:Tb powders studied elsewhere.^19,21 The main goal of this work is to better understand the mechanisms involved at the early stages of the YAG synthesis and, especially, to unravel the role played by acetylacetone during the sol elaboration. For this purpose, the influence of the quantity of acacH on the properties of sols and derived powders has been studied using UV-visible spectroscopy and X-ray diffraction. Then, X-ray absorption spectroscopy (XAS) has been used to investigate oxidation state and local environment of terbium ions for unmodified as well as for acac-modified sols. Their photoluminescence properties were recorded upon UV and blue excitation in Tb^{3+ 5}D₄ level. These results are discussed in light of the structural properties of the samples.

It is noteworthy that numerous papers have been devoted to the hydrolysis of “simple” alkoxides such as Ti or Zr derived ones, and to their chemical modifications.^11,22–28 Except results reported for Ln[Al(OiPr)₃]₄,^29,30 only very few studies concern heterometallic alkoxides involving both isopropoxide and acetylacetonate groups. Those works mainly concern a transition metal^31,32 as metallic centre. In this general context, this paper presents an original work on heterometallic alkoxide systems, which could allow to better understand and to master the early stages of the sol–gel synthesis of functional ceramics. A preliminary study¹⁸ carried out on sols had highlighted that acac-groups are truly linked to metals and that acac-modified sols exhibit larger and more stable colloids than unmodified ones. The present study will particularly focus on the structural and optical modifications of the sol, induced by the presence of acetylacetone with a first insight in the features of functional heterometallic alkoxides.

Experimental section

Materials

YCl₃ (99.99% pure, Aldrich), TbCl₃ (99.99% pure, Aldrich), metallic potassium (98% pure, Aldrich), aluminium isopropoxide (99.99+% pure, Aldrich), acetylacetone (2,4-pentanedione, ≥99%, Aldrich) and anhydrous isopropanol (iPrOH, 99.8+% pure, Aldrich) have been used as received as starting materials.

Synthesis

The sol–gel synthesis procedure used has been detailed in previous works.¹⁸ It consists in preparing separately a solution of anhydrous yttrium chloride and terbium chloride dissolved in anhydrous isopropanol, labelled solution A, and a solution of potassium isopropoxide, labelled solution B. Solution B is slowly added to solution A under vigorous stirring: a precipitate of KCl immediately appears. After 1 hour reflux at 85 °C of the mixed solution, a stoichiometric quantity of aluminium isopropoxide powder is poured directly into the solution. After further reflux for 4 h at 85 °C, a clear and homogeneous solution (solution C) is obtained together with the KCl precipitate. This latter is eliminated by centrifugation carried out subsequently to cooling. All the synthesis steps have to take place under dry argon atmosphere since alkoxide reactants are very sensitive to moisture.

Solution C is generally hydrolysed by introducing an excess of water, then dried at 80 °C and heat-treated to form crystallized YAG powders. But it can also be stabilized by adding acetylacetone after 2 hours of the above mentioned reflux at 85 °C. After the introduction of the desired quantity of acacH into the reactional solution, the reflux is maintained during 2 supplementary hours. Then, after cooling, centrifugation is used for removing KCl precipitate. This gives rise to a stabilized acac-modified sol which is usually used to elaborate thin films by dip-coating for example or which is hydrolysed with the same procedure aforementioned to yield acac-modified powders.

In this work, in order to determine the optimal quantity of acacH required to obtain a stable sol leading to pure YAG phase, several syntheses were carried out with various complexation ratios: R_C = 0, 1, 2 and 3, R_C being defined as:

Undoped and Tb³⁺-activated (20 mol%) matrices were prepared via the synthesis procedure described above. The doping rate of 20 mol% Tb³⁺ was chosen in relation to the observed optimal luminescence performances under blue excitation (485 nm).³³ The different solutions with R_C = 0, 1, 2 or 3 were stored at 10 °C in order to slow down hydrolysis and condensation processes. After several weeks at 10 °C, some crystals appeared in these solutions. To better understand the nature of these crystals, they have been analysed by Raman spectroscopy and X-ray diffraction.

In order to help us for the interpretation of our in-depth spectroscopic study of the sols, simple (homometallic) alkoxide solutions have been also elaborated, following the synthesis process afore described. Three solutions have been prepared from aluminium, terbium or yttrium chlorides dissolved in isopropanol, using the same quantities of acetylacetone.

Characterization

UV-Vis absorption spectra of sols during hydrolysis were measured in quartz cells of 10 mm optical path within the 250–850 nm wavelength range using optical fibres connected to a Varian Cary 50 UV-visible spectrophotometer. Hydrolysis reactions initiated by water addition (10% vol) were carried out at 60 °C in order to accelerate the gelation process. Besides, UV-Vis spectra of acac-modified and unmodified solutions C, deposited on KBr substrates have been recorded using a Shimadzu UV-2101 PC spectrometer equipped with an integrating sphere. They have been compared to the spectra obtained for the non-hydrolyzed homometallic modified alkoxides.

XRD measurements were performed on a Siemens D501 diffractometer operating with a Cu-Kα radiation monochromatized whit a curved graphite monocristal.

Yttrium K-edge (17 080 eV) and terbium L₃-edge (7514 eV) X-ray absorption spectra of references and samples were collected on the beamline BL-11-1 of the Elettra storage ring (Trieste, Italy) operating at 2/2.4 GeV with an optimal current around 300 mA. Measurements were performed using a Si(111) double crystal monochromator. All the data were recorded at liquid nitrogen temperature (77 K) in transmission mode using two ionization chambers filled with a mixture of nitrogen and argon optimized in pressure at each edge in order to achieve absorptions of 20% and 80% for incident I₀ and transmitted I₁ photon flux, respectively. At the Y K-edge and Tb L₃ edge, EXAFS spectra were collected with a step of 2 eV with integration time of 2 s whereas, at the L₃ Tb-edge, XANES spectra were recorded with a step of 0.2 eV and 1 s of integration time. Typically, three or four scans were merged for each sample to get a good signal to noise ratio. A liquid cell with Kapton® windows placed in a liquid nitrogen-chilled cryostat was filled by the sols for their absorption spectra recording.

EXAFS data analysis was carried out using the Athena and Artemis programs.³⁴ The EXAFS extraction and fitting at the Y K edge (respectively Tb L₃ edge) were performed using the same functions and S₀² and e_not parameters than those described in our previous papers¹⁹ (respectively²¹). The reliability of the fit is assessed by the value of the residual factor R_F, which was minimized during the least square fitting procedure. For the XANES study, Tb₄O₇ was used as a standard and data were normalized at 7609 eV.

Raman spectra were recorded with a T64000 Jobin-Yvon confocal micro-Raman spectrograph. The confocal configuration of the micro-Raman instrument allows depth profiling of the samples, permitting the detection of the Raman spectrum from a volume as small as 1 μm³ focusing at different depths into the sample. The excitation source used was the 514.5 nm wavelength line from a Coherent model 70C5 Ar⁺ laser operating at a power of 400 mW. A 100-fold objective lens was used for the focusing so that only a volume of 1 μm³ was sampled. The data was collected for 2 × 30 s. The resolution on the wavenumbers is approximately 1 cm⁻¹.

Excitation and UV-excited emission spectra were recorded with a Jobin-Yvon set-up consisting of a Xe lamp operating at 400 W and two monochromators (Triax 550 and Triax 180) combined with a cryogenically cooled charge coupled device (CCD) camera (Jobin-Yvon Symphony LN₂ series) for emission spectra and with a Hamamatsu R980 photomultiplier for excitation ones. Excitation spectra have been corrected for instrument response and Xe lamp intensity thanks to sodium salicylate. The resolution of the system is better than 0.1 nm in both emission and excitation configurations. Emission spectra upon blue excitation as well as decay curves were recorded with a monochromator Jobin-Yvon HR 1000 spectrometer, using a dye laser (continuum ND62) pumped by a frequency doubled pulsed YAG:Nd³⁺ laser (continuum surelite I). The dye solution was prepared by mixing rhodamines 610 and 640. To achieve a resonant pumping in the blue wavelength range, the output of the dye laser was up-shifted to 4155 cm⁻¹ by stimulated Raman scattering in a high pressure gaseous H₂ cell. Fluorescence decays were measured with a LeCroy 400 MHz digital oscilloscope. All luminescence properties have been studied at room temperature.

Results and discussion

Determination of optimal complexation ratio R_C

The preparation of phosphors with controlled morphological, structural and optical properties by sol–gel method can be obtained through the optimization of the chemical composition of the precursor sol. In our study, solution C must lead to pure YAG phase after water addition, drying and heating treatment. The determination of the amount of acetylacetone to use during the synthesis process to prepare a solution C both suitable for long-term kinetics stability and easy post-synthesis processability is of paramount importance and studied in-depth in the following.

The hydrolysis–condensation step of the precursor solution C with various R_C ratios was studied by UV-Vis spectroscopy. Fig. 1 reports the absorbance change measured at 800 nm of the system heated at 60 °C after addition of distilled water (10% vol).

Fig. 1 Time evolution of the absorbance of YAG:Tb (20%) sols synthesized with different R_C at 800 nm during hydrolysis at 60 °C.

At 800 nm, precursors do not absorb (see Fig. S1†) whatever the R_C ratio, then the increase in absorbance level can be mainly related to scattering phenomenon by the size of particles embedded in solution. Herein, a significant increase of absorbance is expected during the gel formation for which the media becomes turbid due to increase of connectivity between particles. For R_C = 1 and 2, the absorbance slowly increases in time confirming the stability of these sols. For R_C = 3, the absorbance first slowly increases at a rate similar to the variations observed for R_C = 1 and 2 before to rise abruptly due to the formation of large aggregates in the medium owing to the progression of the sol–gel transition (hydrolysis–condensation processes). Finally, the unmodified sol (R_C = 0) is unstable since gelation instantaneously occurs after water addition, as indicated by the significant increase in the absorbance after few minutes. These results show that there is an optimal quantity of acacH achieved with R_C = 1 and 2 which preserves stable the sols for a long time. This kind of behaviour has also been observed by Spijksma and co-workers³⁵ who studied stabilization and destabilization of zirconium isopropoxide sols by acetylacetone. Depending on the added acacH quantity, mixed zirconium isopropoxide/acetylacetonate [Zr(OiPr)₂(acac)₃] turned into more stable Zr(acac)₄.

In order to appreciate the influence of acacH on the YAG crystallinity, the different as-prepared sols (R_C = 0, 1, 2 and 3) were hydrolysed, dried for 12 hours at 80 °C and fired 4 hours at 1100 °C. Fig. 2 displays the XRD patterns of these so-obtained powders. Whereas for R_C ≤ 1, pure YAG phase is obtained, a secondary phase i.e. YAM (Y₄Al₂O₉) is identified for R_C = 2 and 3, together with the YAG phase. The intensity of the diffraction peaks assignable to this phase remains weak for R_C = 2 but strongly increases for R_C = 3. Thanks to this XRD analysis, the value R_C = 2 has been further banished in our synthesis protocol since it leads to impure material.

Fig. 2 XRD patterns recorded from acac-modified YAG:Tb (20%) samples sintered at 1100 °C for 4 h with different R_C.

This behaviour can be related to the nature of crystals formed after storing the solutions at 10 °C for several weeks. For R_C = 1, luminescent small needles are obtained. Unfortunately as these needles are changed to an amorphous powder analogous to xerogels when they are dried, their further investigation by XRD was unsuccessful.

For R_C = 3, non-luminescent cubic-like crystals have been isolated. They are insensitive to drying contrary to needles obtained for R_C = 1. For R_C = 2, both kinds of crystals have been observed. The non-luminescent crystals have been analysed by XRD and Raman spectroscopy. As an example, Fig. 3 exhibits the Raman spectrum of the solid phase crystallizing in the sol characterized by R_C = 3. All the lines observed correspond to vibration modes usually obtained for aluminium tris-acetylacetone.³⁶ This result is in good agreement with the non-luminescence of these crystals since luminescence properties are only due to the presence of Tb³⁺ ions. The XRD pattern obtained is close to the one previously reported by Le Bihan et al.³⁷ and ascribed to Al(acac)₃. Then we propose that, in the presence of acacH with R_C > 1, several acac groups link preferably to the aluminium atoms to form the stable Al(acac)₃.

Fig. 3 Raman spectrum of crystals obtained from the sol synthesized with R_C = 3.

Since reagents are used in stoichiometric proportions to form Y₃Al₅O₁₂, the formation of Al(acac)₃ for R_C > 1 explains that, upon hydrolysis–condensation processes and subsequently drying, a lower Al-rich phase than YAG identified by XRD as YAM (Y₄Al₂O₉), is obtained.

Based on the communications of Spijksma et al.,³⁵ Ribot et al.,³⁸ and Harlan et al.,³⁹ we assume that Al(acac)₃ is more stable than Y(acac)₃ or Tb(acac)₃ and the appearance of Al(acac)₃ entails the destabilization of sols after addition of water by leaving unstable [M(OⁱPr)₃]_n or [M(OⁱPr)₂(acac)]_n (M = Y or Tb and n to be presumably equal to 2, 3 or 4) species alone. This can explain the specific behaviour observed for R_C = 3 in Fig. 1: a slight increase in absorbance during the first 30 minutes (time needed for destabilization mechanism) followed by a rapid sol–gel transition. This assumption is consistent with the ⁸⁹Y NMR study of Harlan et al.³⁹ showing that when mixed with alumoxane, Y(acac)₃ tends to create an yttrium-doped alumoxane together with the formation of Al(acac)₃. In this case, Al(acac)₃ naturally appears as more stable than Y(acac)₃.

To conclude this part, an optimal quantity of acacH defined by a value of R_C = 1 has been determined for the preparation of solution C: after water-addition, this ratio leads to a more stable sol (Fig. 1) resulting in a pure single YAG phase with good crystallinity as evidenced by XRD (Fig. 2). As a consequence, the next sections of this paper will be devoted only to the comparison of the structural and optical properties between unstabilized (R_C = 0) and stabilized (R_C = 1) sols.

UV-Vis spectroscopy

Fig. 4 gathers the UV-Vis spectra recorded for unmodified (R_C = 0) and modified (R_C = 1) precursor solutions C of YAG:Tb as well as for free acetylacetone and homometallic alkoxides solutions made of Tb, Y or Al isopropoxides mixed with acetylacetone. Contrary to the previous section where water was added to solutions C for the study of long term stability, the solutions studied in this section are the as-prepared homo and hetero solutions C only presenting slight hydrolysis–condensation characteristics induced by exposure to atmosphere moisture. They are named “sols”.

Fig. 4 UV-Vis spectra recorded for (a) acac-modified (R_C = 1) and (b) unmodified (R_C = 0) precursor sols of YAG:Tb, simple alkoxides made of (c) Tb, (d) Al or (e) Y isopropoxides mixed with acetylacetone and (f) free acetylacetone.

The spectrum obtained for the acac-modified precursor sols of YAG:Tb (Fig. 4a) only presents a broad absorption band lying from approximately 240 to 340 nm whereas the unmodified sample does not display any absorption band within this wavelength range (Fig. 4b). The absorption signal observed for the modified sample seems to result from different contributions, with two maxima located at 274 and 302 nm, respectively. This broad band can be ascribed to the π→π* transition of acetylacetonate groups.⁴⁰

This confirms the chelating effect of acetylacetone, already reported by IR spectroscopy for this sol.¹⁸ Furthermore, considering the absorption spectra obtained for the simple acac-modified isopropoxide sols of Tb (Fig. 4c), Al (Fig. 4d) and Y (Fig. 4e) and for free acetylacetone (Fig. 4f), it can be concluded that the modified sol possesses a complex architecture where acac-groups have chelated at least two different metallic atoms, among them Tb ones. As explained by Holm and Cotton,⁴⁰ the wavelength of the main absorption band is dependent on the coordinated metal and its position is nearly random; consequently, it is difficult, from UV-Vis spectra, to conclude if the sol is only composed of a mixture of homometallic alkoxides, of heterometallic alkoxides, chelated by acac groups or contains both of them.

XAS

It is of paramount importance to gain knowledge on the local surrounding environment of lanthanides even in sols because luminescence properties are strongly dependant on the local structure around the active ion.⁴¹ XAS has been proved to be a very efficient tool to study alkoxides in solution, particularly when they are modified by chelating agent such as acetylacetone.^38,42 In order to obtain more insights in the structure of metallic species forming both sols, a multi-edge EXAFS investigation was done at the Y K and Tb L₃ edges. The PRDFs (uncorrected from phase shift) determined from the Y K-edge and Tb L₃-edge EXAFS spectra of both Tb-doped sols are shown on Fig. 5a and b respectively. Whatever the absorbing atoms, both sols are characterized by a main contribution located between 1 and 2 Å related to the first oxygen coordination shell around Y or Tb with the same respective position. Namely, irrespective of the absorbing atom, it appears than the main contribution characterizing the PRDFs is located at longer distance for the unmodified sol compared to the acac-modified sample. Regarding the contributions at longer distances, if the local order around Tb displays a resolved peak at around 2.8 Å for both sols (Fig. 5b), only the acac-modified sample displays an additional resolved contribution centred at around 2.8 Å from the yttrium absorbing atom. These features evidence that the modified sol appears as structured Tb and Y entities with organization around yttrium clearly promoted by acac groups. This phenomenon has also been reported for acac-modified and acetic acid modified zirconium-based sols,^43,44 for which the increase in second neighbours has been ascribed to a higher degree of polycondensation with more compact structural units.

Fig. 5 (a) Y K and (b) Tb L₃-edges Fourier transforms |F(r)| uncorrected from phase shift and calculated from kχ(k) EXAFS oscillations for both sols (R_C = 0 and R_C = 1); comparison with the Fourier transform obtained for stabilized xerogel.

Structural parameters characterizing the local order around Y and Tb of both sols were refined by least square fitting procedure and gathered in Tables 1 and 2 respectively. Fits for stabilized acac-modified sol (R_C = 1) are compared with the filtered experimental EXAFS spectra at the Tb L₃ edge in the 1–4 Å range in Fig. 6. Results obtained for acac-modified sample at the Y K-edge are presented in Fig. S2.†

Table 1 Structural parameters obtained from the EXAFS data recorded at the Y K-edge for both sols

Sample	Y neighbourhood	N	R (Å)	Sigma (Å)
Unstabilized sol RC = 0	O	1.4	2.01	0.094
	O	7.5	2.33	0.102
Stabilized sol RC = 1	O	3.3	2.29	0.049
	O	4.0	2.42	0.071
	Al	1.9	3.22	0.065

---

Table 2 Structural parameters obtained from the EXAFS data recorded at the Tb L₃-edge for both sols

Sample	Tb neighbourhood	N	R (Å)	Sigma (Å)
Unstabilized sol RC = 0	O	1.0	2.20	0.116
	O	7.0	2.39	0.116
	Al	5.3	3.24	0.130
Stabilized sol RC = 1	O	4.7	2.30	0.086
	O	3.3	2.44	0.086
	Al	4.4	3.22	0.110

---

Fig. 6 Experimental (solid line) and fitted (dot line) two first-shell filtered EXAFS spectra at Tb L₃-edge for stabilized acac-modified sol (R_C = 1).

Data and theoretical signals are in good agreement (R_F comprised between 0.57 and 1.03%). It is noteworthy that for the acac-modified sol, the oxygen contributions around both absorbing atoms are located at distances in agreement with those found for the first coordination shell around yttrium embedded in the YAG structure, i.e. R₁ = 2.30 Å and R₂ = 2.43 Å.⁴⁵ Irrespective of the absorbing atom, the acac-modified sol displays a coordination shell with almost equivalent number of oxygen atoms around 4 in both sub-shells whereas the first coordination shell of the unmodified sol is mainly composed by 7 oxygen atoms at the longer distance, giving rise to the apparent shift at longer distance of the corresponding PRDFs peak for this sample compared to the modified one.

Besides the first coordination shell, least square fitting of EXAFS data for the modified sol reveals as second neighbours aluminium atoms located at 3.23 ± 0.01 Å for both absorbing atom. This contribution is at longer distance than the one characterizing the 2 aluminium atoms in tetrahedral site of the YAG structure (3.00 Å) but shorter than the one reported for the 4 second aluminium nearest neighbours in octahedral site (3.35 Å). We assume that pre-nuclei of YAG are already formed in the sol state but the lack of contributions at longer distances of aluminium/yttrium/terbium shells at 3.66 Å in the YAG or TAG structures does not constrain these first aluminium neighbours to be located in the tetrahedral and octahedral sites of the garnet structures but rather at the mean distance of both sites (3.24 Å = (2 × 3.00 Å + 4 × 3.35 Å)/6, 2 and 4 being the occupancy of tetrahedral and octahedral sites respectively). It is noteworthy that any attempt to fit the medium range order around yttrium for the unmodified sol beyond the first coordination shell with an aluminium contribution was unsuccessful. This contribution is probably due to light element contributions belonging to isopropoxide ligands.

Although the EXAFS data does not unambiguously revealed the formation of the triple heterometallic alkoxide (Y–Tb–Al) by the presence of Y or Tb contribution at longer distances, we have clearly evidenced the formation of heterometallic alkoxide with Al atoms linked to Tb atoms for both sols and to Y atoms for the modified sol. Actually, the latter does not only consist of a mix of simple homometallic alkoxide [M(OiPr)₂(acac)]_n M = Al, Y or Tb and n to be presumably equal to 2, 3 or 4) as it has been observed by Ion et al.⁴⁴ for zirconium-based sols. This is in good agreement with UV-visible results reported in the Fig. 4 which emphasize the presence of a mixture of homo and/or heterometallic alkoxides, associated with acetylacetonate groups.

Finally, the similarity of the local order around Tb and Y when acac ligands are added in the synthesis route, suggests that pre-nuclei of YAG:Tb structure are formed. We assume that the aluminium second neighbours around both cations are themselves bonded to acac ligands to form a capping shell leading to the apparent stabilization of the sol in presence of water (Fig. 1). Similar order around yttrium, with aluminium atoms as second nearest neighbours, is not evidenced for the unmodified sol. This feature is probably responsible for the unstable behaviour of this sol when water is added (Fig. 1).

A Tb L₃ edge XANES study has been undertaken on both Tb-doped sols in order to assess a possible influence of acacH on the Tb oxidation degree. Indeed, the presence of Tb⁴⁺ ions would be prejudicial to the optical properties: actually, Tb³⁺ always leads to radiative emission whereas Tb⁴⁺ does not and can even promote luminescence quenching.^46,47

Fig. 7 displays XANES spectra recorded for both sols as well as the one corresponding to the standard mixed valent Tb₄O₇. Sols are characterized by similar XANES profiles for which Tb⁴⁺ signal has not been observed.⁴⁶ The synthesis procedure does not entail the formation of Tb⁴⁺, even if acacH is used.

Fig. 7 Normalized Tb L₃-edge XANES spectra of both YAG:Tb (20% mol) precursor sols and Tb standard Tb₄O₇.

Optical studies

In order to assess the application potentialities of synthesized phosphors in devices using UV or blue excitations, excitation spectra of both Tb-doped sols have been recorded at room temperature in these spectral ranges, monitoring the ⁵D₄→⁷F₅ transition located at 544 nm. As shown in Fig. 8, samples exhibit very different spectral profiles. For the stabilized sol (Fig. 8a), a broad and intense absorption band located at 330 nm is mainly observed as well as a shoulder around 376 nm and a very weak signal at 485 nm. Besides, the unmodified sol displays very weak absorption bands centred at 260, 300, 352, 376 and 485 nm (Fig. 8a and b). The absorption bands located at 352, 376 and 485 nm are ascribed, for both sols, to the Tb^{3+ 7}F₆→⁵D₂ ⁷F₆→⁵D₃ and ⁷F₆→⁵D₄ (4f–4f) transitions, respectively, according to YAG:Tb³⁺ powders luminescence studies.^47,48

Fig. 8 (a) Excitation spectra of YAG:Tb (20%) sols in the UV and blue ranges recorded at room temperature, monitoring the ⁵D₄→⁷F₅ transition located at 544 nm and (b) zoom for the excitation spectrum of the unmodified sol (×5).

The main absorption band obtained for R_C = 1 is characteristic of the π→π* transition of acetylacetonate groups. It can be assumed that it is broad and strong enough to hide the excitation peaks relative to 4f→5d Tb³⁺ transitions. The significant differences recorded between both sols below 350 nm are directly related to the modifications of Tb³⁺ local environment induced by acetylacetone, as evidenced by the XAS study (see Table 2). Indeed, the 5d orbit, being the outer orbit of the ion, is strongly influenced by the strength and symmetry of crystal-field undergone by the active ions. In the framework of crystal-field theory, amorphous phases exhibit lower symmetry and weaker crystal-field strength induced by the neighbouring ions if compared with crystallized ones. Consequently, the next neighbour could significantly alter the crystal-field parameters and so the spectral distribution, as observed here.

Emission spectra have been recorded, exciting the sols at the wavelengths corresponding to their maximum of absorption, which are 330 nm for acac-modified sols and 300 nm for unmodified ones. They are presented in Fig. 9a and b respectively. The spectral profiles are similar and exhibit the characteristic ⁵D₄→⁷F_J (J = 6 to 0) Tb³⁺ transitions.^47,48 They are characteristic of Tb³⁺ ions embedded in an amorphous matrix with relatively broad signals and not clearly defined Stark components.⁴⁹

Fig. 9 Room temperature emission spectra of YAG:Tb (20 mol%) (a) acac-modified sol and (b) unmodified sol upon UV excitations. The inset presents a photograph of acac-modified sol upon UV excitation.

For both samples, the ⁵D₄→⁷F₅ transition which usually confers to the Tb³⁺-doped matrices their green fluorescence (see the picture in the inset of Fig. 9) is the most intense. If compared the emission intensities between the acac-modified (Fig. 9a) and unmodified (Fig. 9b) sols, the first one is much more efficient upon UV-excitation. It can be related to an efficient energy transfer from the acac-groups that absorb the 330 nm excitation to the Tb³⁺ ions:^50,51 it is the well-known “antenna effect” of rare-earth doped β-diketonates⁵² described through the schema presented in Fig. S3.†

Emission spectra recorded for the same sols under blue excitation (485 nm) are displayed in Fig. 10. They are similar and the green emission observed arises from the spectral profiles. Thus, the relative outputs of luminescence related to both sols were determined by integrating the area under this emission band. Results are shown in Fig. S4.†

Fig. 10 Room temperature emission spectra of YAG:Tb (20 mol%) sols upon blue excitation.

The higher luminescence yield is obtained for the stabilized sol (R_C = 1) which is five times more efficient than the unmodified one upon blue excitation (485 nm). This can be explained by changes in Tb³⁺ local environment since Tb^{3+ 5}D₄ level is directly excited in this case.

It is in good agreement with both dissimilar Tb³⁺ surroundings evidenced by XAS study and the assumption of the presence of pre-nuclei in acac-modified sol made based on Y³⁺ local environment (see discussion in XAS part and results gathered in Tables 1 and 2).

It is noteworthy to precise that the addition of acetylacetone during the synthesis has not only an influence on sols properties but also on structural, morphological and optical features of the derived crystallized powders, as evidenced in our previous works.^19,21 Indeed, acac-modified powders were characterized by a different network aggregation process resulting in an earlier crystallization and a better removal of organic residue compared to non-modified powders. Thanks to this behaviour, acac-modified YAG:Tb powders presented a significantly improved luminescence under both blue and UV excitation for heating temperatures comprised between 400 and 900 °C.

Finally, the decay curves of these sols recorded at room temperature under blue excitation, monitoring ⁵D₄→⁷F₅ transition, are presented in Fig. 11a and b for sol with R_C = 0 and R_C = 1, respectively. For both investigated sols (Fig. 11), a single exponential decay is observed with time constants of 1.8 ms and 1.1 ms for R_C = 0 and R_C = 1 respectively.

Fig. 11 Luminescence decay traces at 300 K of the ⁵D₄→⁷F₅ emission of Tb³⁺ in (a) unstabilized YAG sol (R_C = 0) and (b) stabilized YAG sol (R_C = 1).

These lifetimes are much lower than those recorded for YAG:Tb (20% mol) crystallized powders for which values around 3 ms were obtained.^53,54

This shortening can be ascribed to the presence of organic residues which act as non-radiative relaxation promoters, via vibrations associated to C–H and O–H groups. For sol prepared with R_C = 1, the ⁵D₄ luminescence decay is faster than for R_C = 0, which can be ascribed to energy transfer between Tb³⁺ followed by trapping of part of the luminescence by additional organic impurities due to acacH as previously evidenced by TGA.¹⁸ Furthermore, the time constant corresponding to the acac-modified sol is close to those observed for Tb³⁺ β-diketonates.

Conclusion

The precursor solutions of Y₃Al₅O₁₂:Tb³⁺ phosphors obtained through sol–gel process have been studied. Thanks to XRD and UV-visible spectroscopy, an optimal rate of acetylacetone defined by a value of R_C = 1 has been determined: it leads to a more stable sol than unmodified one while resulting in a pure single YAG phase with good crystallinity.

The second part of this paper has been focused on the relationship between optical properties of the solutions defined by R_C = 0 and R_C = 1 and the local environment of Y³⁺ and Tb³⁺ ions. XAS has highlighted the presence of heterometallic alkoxides Y/Al in acac-modified solutions whereas XANES has established that the trivalent oxidation state of Tb cations is preserved during the sol–gel process for both kinds of solutions. Thanks to the chelating effect of acac-groups, acac-modified solutions have turned out to be much more efficient upon UV-excitation than unmodified ones. Upon blue excitation acac-modified solutions remain more efficient but, in this case, it has to be related to local environment around Tb³⁺. Indeed, modified sols are organized differently compared to unmodified ones, with the assumed presence of pre-nuclei of YAG:Tb structure entailing a stronger crystal-field undergone by Tb³⁺ ions. This high luminescence of sols is very interesting for using them for applications and substrates where heating treatment is not possible. By the way, we have already used this acac-modified sol to elaborate luminescent silica monoliths⁵⁵ usable for mood lighting or interior design.

Acknowledgements

The authors are grateful to ELETTRA (Sincrotrone) and the European Community for having financially supported the XAS measurements carried out at Elettra (Contract RII3-CT-2004-506008 (IA-SFS)).

Notes and references

1. S. Mathur, M. Driess, H. C. Robert and D. M. P. Mingos, in Comprehensive Organometallic Chemistry III, Elsevier, Oxford, 2007, pp. 35–70.
2. C. J. Brinker and G. W. Scherer, Sol–Gel Science: The Physics and Chemistry of Sol–Gel Processing, Academic Press, New York, 1990.
3. C. Sanchez and J. Livage, New J. Chem., 1990, 14, 513–521.
4. M. Veith, J. Sol-Gel Sci. Technol., 2008, 46, 291–298.
5. M. Henry, J. P. Jolivet and J. Livage, in Ultrastructure Processing of Advanced Materials, ed. D. R. Uhlmann and D. R. Ulrich, John Wiley & Sons, New York, 1992, pp. 23–45.
6. J. Livage, M. Henry and C. Sanchez, Prog. Solid State Chem., 1988, 18, 259–341.
7. C. Sanchez, J. Livage, M. Henry and F. Babonneau, J. Non-Cryst. Solids, 1988, 100, 65–76.
8. D. Jia, Chem. Eng. Commun., 2007, 194, 1666–1687.
9. P. Vaqueiro and M. A. Lopez-Quintela, Chem. Mater., 1997, 9, 2836–2841.
10. H. Izutsu, F. Mizukami, T. Sashida, K. Maeda, Y. Kiyozumi and Y. Akiyama, J. Non-Cryst. Solids, 1997, 212, 40–48.
11. M. M. Haridas, S. Datta and J. R. Bellare, Ceram. Int., 1999, 25, 601–606.
12. H. Schmidt, J. Sol-Gel Sci. Technol., 2006, 40, 115–130.
13. P. Padmaja, G. M. Anilkumar, P. Mukundan, G. Aruldhas and K. G. K. Warrier, Int. J. Inorg. Chem., 2001, 3, 693–698.
14. P. Schlotter, J. Baur, C. Hielscher, M. Kunzer, H. Obloh, R. Schmidt and J. Schneider, Mater. Sci. Eng., B, 1999, 59, 390–394.
15. S. Nakamura, S. Pearton and G. Fasol, The Blue Laser Diode: The Complete Story, Springer, Berlin, 2nd updated and extended edn, 2000.
16. P. F. S. Pereira, M. G. Matos, L. R. Avila, E. C. O. Nassor, A. Cestari, K. J. Ciuffi, P. S. Calefi and E. J. Nassar, J. Lumin., 2010, 130, 488–493.
17. A. Potdevin, G. Chadeyron and R. Mahiou, Chem. Phys. Lett., 2010, 490, 50–53.
18. A. Potdevin, G. Chadeyron, D. Boyer and R. Mahiou, J. Sol-Gel Sci. Technol., 2006, 39, 275–284.
19. A. Potdevin, G. Chadeyron, V. Briois, F. Leroux, C. V. Santilli, M. Dubois, D. Boyer and R. Mahiou, Dalton Trans., 2010, 39, 8706–8717.
20. C. W. Won, H. H. Nersisyan, H. I. Won, J. H. Lee and K. H. Lee, J. Alloys Compd., 2011, 509, 2621–2626.
21. A. Potdevin, G. Chadeyron, V. Briois, F. Leroux and R. Mahiou, Dalton Trans., 2010, 39, 8718–8724.
22. W. C. LaCourse and S. Kim, Ceram. Eng. Sci. Proc., 1987, 8, 1128–1134.
23. A. Ponton, S. Barboux-Doeuff and C. Sanchez, J. Non-Cryst. Solids, 2005, 351, 45–53.
24. D. Peter, T. Ertel and H. Bertagnolli, J. Sol-Gel Sci. Technol., 1995, 5, 5–14.
25. M. Chatry, M. Henry, M. In, C. Sanchez and J. Livage, J. Sol-Gel Sci. Technol., 1994, 1, 233–240.
26. J. C. Debsikdar, J. Non-Cryst. Solids, 1986, 86, 231–240.
27. M. J. Percy, J. R. Bartlett, J. L. Woolfrey, L. Spiccia and B. O. West, J. Mater. Chem., 1999, 9, 499–505.
28. M. Sedlar and M. Sayer, J. Sol-Gel Sci. Technol., 1995, 5, 27–40.
29. A. Mehrotra and R. C. Mehrotra, Inorg. Chem., 1972, 11, 2170–2174.
30. C. K. Narula, W. H. Weber, J. Y. Ying and L. F. Allard, J. Mater. Chem., 1997, 7, 1821–1829.
31. E. Ghanti and R. Nagarajan, Dalton Trans., 2010, 39, 6056–6061.
32. V. G. Kessler, S. Gohil and S. Parola, Dalton Trans., 2003, 544–550.
33. A. Potdevin, G. Chadeyron, D. Boyer and R. Mahiou, J. Non-Cryst. Solids, 2006, 352, 2510–2514.
34. B. Ravel and M. Newville, J. Synchrotron Radiat., 2005, 12, 537–541.
35. G. I. Spijksma, H. J. M. Bouwmeester, D. H. A. Blank and V. G. Kessler, Chem. Commun., 2004, 1874–1875.
36. S. F. Tayyari, H. Raissi and Z. Ahmadabadi, Spectrochim. Acta, Part A, 2002, 58, 2669–2682.
37. L. Le Bihan, F. Dumeignil, E. Payen and J. Grimblot, J. Sol-Gel Sci. Technol., 2002, 24, 113–120.
38. F. Ribot, P. Toledano and C. Sanchez, Chem. Mater., 1991, 3, 759–764.
39. C. J. Harlan, A. Kareiva, B. MacQueen, R. Cook and A. R. Barror, Adv. Mater., 1997, 9, 68–71.
40. R. H. Holm and F. A. Cotton, J. Am. Chem. Soc., 1958, 80, 5658–5663.
41. J. McKittrick, M. E. Hannah, A. Piquette, J. K. Han, J. I. Choi, M. Anc, M. Galvez, H. Lugauer, J. B. Talbot and K. C. Mishra, ECS J. Solid State Sci. Technol., 2013, 2, R3119–R3131.
42. M. Bauer, S. Muller, G. Kickelbick and H. Bertagnolli, New J. Chem., 2007, 31, 1950–1959.
43. I. Arcon, B. Malic, M. Kosec and A. Kodre, Mater. Res. Bull., 2003, 38, 1901–1906.
44. E. D. Ion, B. Malic, I. Arcon, J. P. Gomilsek, A. Kodre and M. Kosec, J. Eur. Ceram. Soc., 2010, 30, 569–575.
45. D. Rodic, M. Mitric, R. Tellgren and H. Rundlof, J. Magn. Magn. Mater., 2001, 232, 1–8.
46. A. Martinez-Arias, A. B. Hungria, M. Fernandez-Garcia, A. Iglesias-Juez, J. C. Conesa, G. C. Mather and G. Munuera, J. Power Sources, 2005, 151, 43–51.
47. S. Zhou, Z. Fu, J. Zhang and S. Zhang, J. Lumin., 2006, 118, 179–185.
48. G. Blasse and B. C. Grabmaier, Luminescent Materials, Springer-Verlag, Berlin, 1994.
49. K. Tonooka and O. Nishimura, J. Mater. Sci., 1999, 34, 5039–5044.
50. D. F. Parra, A. Mucciolo and H. F. Brito, J. Appl. Polym. Sci., 2004, 94, 865–870.
51. R. Wang, D. Song and S. Wang, Chem. Commun., 2002, 368–369.
52. K. Y. Jung, D. Y. Lee and Y. C. Kang, J. Lumin., 2005, 115, 91–96.
53. T. Welker, J. Lumin., 1991, 48–49, 49–56.
54. J. P. van der Ziel, L. Kopf and L. G. Van Uitert, Phys. Rev. B: Condens. Matter Mater. Phys., 1972, 6, 615.
55. R. Pązik, K. Zawisza, A. Watras, K. Maleszka-Bagińska, P. Boutinaud, R. Mahiou and P. J. Dereń, Mater. Res. Bull., 2013, 48, 337–342.

---

Footnote

† Electronic supplementary information (ESI) available: Fig. S1: absorbance spectra of the precursor solutions; Fig. S2: experimental (solid line) and fitted (dot line) two first-shell filtered EXAFS spectra at Y K-edge for stabilized acac-modified sol (R_C = 1); Fig. S3: general scheme representing the energy transfer from a ligand to Tb³⁺ ions, evidencing the antenna effect; Fig. S4: comparison of relative luminescence yields determined under blue excitation for Tb-doped sols between an acac-modified sol and an unmodified one. See DOI: 10.1039/c6ra06444b

---