Characterisation of YAG:Ce powders thermal treated at different temperatures

Characterisation of YAG:Ce powders thermal treated at different temperatures

Applied Surface Science 238 (2004) 469–474 Characterisation of YAG:Ce powders thermal treated at different temperatures G. Del Rosarioa,*, S. Oharab,...

372KB Sizes 0 Downloads 51 Views

Applied Surface Science 238 (2004) 469–474

Characterisation of YAG:Ce powders thermal treated at different temperatures G. Del Rosarioa,*, S. Oharab, L. Mancicc, O. Milosevicc a

Technological Support Center, University Rey Juan Carlos I. Tulipan s/n. Mo´stoles, Madrid 28933, Spain b Institute of Multidisciplinary Research for Advanced Materials, Tohoku University, 2-1-1 Katahira, Aoba-ku, Sendai, 980-8577, Japan c Institute of Technical Sciences of Serbian Academy of Science and Arts, K. Mihajlova 35/IV, 11000 Belgrade, Yugoslavia Available online 11 September 2004

Abstract Submicronic, spherical, polycrystalline YAG:Ce powders with luminescence properties were synthesised through aerosol processing route from the corresponding nitrates solution. Additional heat treatment was performed in the temperature range from 1000 to 1200 8C in order to increase the crystallinity of the obtained cubic garnet phase. SEM examination and subsequent morphological analysis allowed studying the surface properties and particle size distributions. An Image Processor was used to measure particle surface roughness. Quantitative SEM/EDS analysis indicated the synthesised materials present high purity and compositional homogeneity. TEM and selected area electron diffraction (SAED) showed a high crystallinity of the samples. XRD patterns of the powders were recorded in the region of 2y ¼ 108–808. The evolution of crystallite size was performed measuring of the broadening of a particular peak using the Scherrer equation. It was found that the crystallite size and bulk particles size vary with the applied temperature. The observed changes in function of the different thermal treatments were correlated with the photoluminescence (PL) properties of these materials. # 2004 Elsevier B.V. All rights reserved. Keywords: Aerosol; SEM/EDS; TEM; XRD; SAED; Luminescence

1. Introduction Y3Al5O12:Ceþ3 (YAG:Ce) presents an excellent thermal stability, good mechanical properties and luminescence properties. Since, the radiation emitted reaching wavelength surrounds to 380–500 nm, these compounds are considered to be of interest in several applications (e.g. field emission display (FED), cathode-ray tubes (CTRs), LED and pigments and thermal *

Corresponding author. Tel.: þ34 91 488 7191; fax: þ34 91 488 7184. E-mail address: [email protected] (G. Del Rosario).

barrier coatings (TBC)). The most of the studies [1,2] about these materials are approached the analysis of influence of chemical composition and, in particular the cerium concentration. In this paper, a new approach is studied. We have analysed the influence of morphology and crystallinity of samples, with the same initial composition, in the emitted light intensity. The homogeneity of materials, synthesised through aerosol, induces energy transfer process well defined which contributes to the higher quantum efficiency of luminescent materials [3,4]. In polycrystalline YAG:Ce, the transitions from the 4f ground state to the lowest 5d energy level occur in the visible part of

0169-4332/$ – see front matter # 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.apsusc.2004.05.253

470

G. Del Rosario et al. / Applied Surface Science 238 (2004) 469–474

the spectrum. The 5d electrons have very interactions with the ligands. In general, the electronic d-state splits into several energy levels whose degeneracy in the crystalline field depends on the site symmetry. For this reason, we pretend to correlate the modifications over structure and morphology of YAG:Ce powders, caused by a different applied thermal processing, with its luminescent properties. A comparative study of the characteristics of the powders treated at different temperatures 1000–1200 8C helps to understand the evolution of the system with the temperature.

2. Experimental Starting precursor solution was prepared by mixing the corresponding aqueous solutions of Y(NO3)36H2O (Aldrich) with Al(NO3)39H2O (Merck) in the desired 3:5 Y:Al molar ratio and adding Ce(NO3)36H2O (Aldrich) while stirring magnetically at room temperature for approximately 20 min before being added to the ultrasonic atomisation chamber. Such obtained common solution was ultrasonically atomised (resonant frequency of 800 kHz) into fine aerosol droplets, introduced into a high temperature tubular flow reactor with air as a carrier gas and decomposed at 900 8C. The characteristics of the used experimental installation were exposed in a previous paper [5,6]. The precursors solution was control by measuring the concentration (0.06 mol dm3), pH (2.7), viscosity (1.043  103 Pa s, MLW Viskosimeter B3), density (1.064 g cm3, AP PAAR Density meter DMA55) and surface tension (112.29 mN m1, Tensiometer K10T KRUSS). The droplet/particle residence time was calculated from the carrier gas flow rates and the reactor geometry, obtaining a value of 3 s for the gas flow rate set at 1.67 dm3/min. Based on these values, the average droplet size, the mean particle size and the aerosol droplet number density were estimated as follows, respectively: 2.316 mm, 745 nm and <106 cm3. As-prepared powders were additionally thermal treated at 1000, 1100 and 1200 8C for 2 h. 2.1. Characterisation Scanning electron micrographs and morphological analysis was carried out on a XL30 ESEM Philips with an energy dispersive spectrometry system (EDS),

operating at 30 kV. Digital images with lateral resolution of 143 pixels/mm was analysed by Scanning Probe Image Processor (SPIP) software [7], containing automatic tools for calibration and measurement of surface topography. The roughness parameters give quantitative information about the particle average surface. Calculating the ratio between the surface area (taking the Z height into account) and the area of the surface projected in a flat XY, is possible to know to increase of area. X-ray diffraction patterns were recorded in the region of 2y ¼ 108–808 with a step size of 0.048 on a Diffratometer Philips X’Pert PRO, using Cu Ka radiation. Crystallite size is performed by measuring of the broadening of a particular peak associated to a planar reflection from crystal unit cell. Conventional transmission electron microscopy was carried out on a TECNAI 20 Philips, operating at 200 kV. The samples were prepared by powder deposition over copper grids. The emission spectra and luminescent intensity were measured in the wavelength region from 400 to 700 nm, using a light source (excitation spectra): 351 nm of Ar laser (7 mW).

3. Results and discussion The effect of applied temperature in the morphology and size distribution of particles was studied by SEM. Fig. 1 shows obtained micrographs at high magnification of the YAG:Ce powders, which exhibit irregularly shape in its surface. The formation of this surface texture is attributed to the growing of nanocrystallites during the crystallisation progress from amorphous precursors. Particle size distribution, together with chemical composition which was measured by EDS (Fig. 1), show changes with the post heat treatment. This means that, it is possible both to eliminate volatiles compounds by simple decomposition of nitrates and to modify the crystallinity with the temperature increase [3]. The equivalent surface and volume were estimated based on the average particle diameter. Statistical analysis of the particle size distribution is presented in Table 1, where are observed differences between samples. The sample treated to 1100 8C has the average range of size distribution of 152–1830 nm, with the major frequency from 400 to 1000 nm. Also, this sample presents higher mean diameter and its fre-

G. Del Rosario et al. / Applied Surface Science 238 (2004) 469–474

471

Fig. 1. SEM photomicrographs of powder after thermal treatment at: 1000 8C (a), 1100 8C (b) and 1200 8C (c). EDS spectra of YAG-Ce powder (d).

quency histogram is more uniform than the rest of studied samples. It is known that aerosol synthesis allows obtaining powders with a high surface roughness, which enhances its surface reactivity and avoids the compositional segregation at the droplet level [1–4]. The calculus of surface roughness parameters was performed, using the deep of focus calculated from

Raleigh principle [8] as a step-height reference. These results are summarised in Table 1, where is also collected the calculated surface area of particle. The growth of the surface in the sample treated to 1100 8C implies higher reactivity and must provide better uniformity in coating applications. Fig. 2 shows the variation of X-ray diffraction patterns with the temperature. The three principal

Table 1 Statistical analysis, roughness data and surface area calculated Treatment (8C)

Mean diameter (nm)

S.E. (yEr)

Average volume (mm3)

Roughness average surface Sa (nm)

Surfaces area ratio Sdr (%)

Area surface (m2)

Mass by particle (g)

Relation (m2/g)

1000 1100 1200

664 788 599

32 44 36

0.154 0.257 0.112

68.0 78.9 71.6

37.9 73.9 32.1

52.6 1012 144.3 1012 36.2 1012

7.0 1013 1.2 1012 5.1 1013

74.5 122.3 71

Note: S.E. is standard error of the mean. Used density: (Dx-YAG) 4.553 g cm3 (JCPDS—International Centre for Diffraction data, 33–0040). (Dx-Ce) 6.814 g cm3 (JCPDS—International Centre for Diffraction data, 78–0638).

472

G. Del Rosario et al. / Applied Surface Science 238 (2004) 469–474

Fig. 2. XRD pattern of YAG:Ce powder after thermal treatment at 1000 8C (a), 1100 8C (b), 1200 8C (c). Details of crystallographic directions ˚ , (b) 407 A ˚ and (c) 474 A ˚. 4 2 0 overlap pattern and the results of the crystallite size calculated for: (a) 318 A

peaks belong to the directions 2 1 1, 4 0 0, 4 2 0 of the YAG phase (JCPDS file card 33–0040). The intensity of peaks significantly increases in function of the applied temperature. It was probed that this variation is not due to preferential orientations. Also, the diffracted intensity is function of the volume irradiated, but this was equal in the three studied specimens. Therefore, the explication of this observation is that the crystallite size increases with thermal treatment. Average crystallite size was determined from the broadening of the peak corresponding to the (4 2 0) reflection, using the Scherrer formula [9,10]: Dhkl ¼

Kl ðb cos yÞ

˚ ), l is the where Dhkl is the average crystallite size (A ˚ wavelength of X-ray used (1.54056 A) and y is the angle of diffraction. b is a full width at half maximum observed in radian and K is a constant which depends of the crystallite form (0.9  57.3). The obtained values for crystallite size are 318, 407 ˚ for samples treated at 1000, 1100 and and 474 A 1200 8C, respectively. The temperature increase causes dimensional changes (Table 1) and variations of chemical composition (Fig. 1). These variations can be associated to

the growth of crystallite size and the loss of volatile compounds. It has been found that the relationship between the crystallite size and average particle volume present a maximum for the sample treated at 1100 8C, indicating that in the studied thermal range, this is the optima temperature because it is reached a maximum growth of crystallite size without significance loss of material. Fig. 3a shows the PL spectra. The curves exhibit broad emission in the range from 450 to 700 nm with a luminescent maximum at 533 nm ( 18.76  103 cm1), attributed to the Ce3þ inter-shell transition (5d ! 4f ) in YAG lattice [8]. The higher intensity was detected for the powders treated at 1100 8C. In an attempt to explain these results, the average distance (dCe–Ce) between luminescence centres has been calculated, basing on the proposition that luminescence intensity is influenced by the content of Ce3þ ions, substituted for Y lattice sites (CeY) in YAG matrix. The calculus was performed from the average volume occupied by Ce atoms, assuming ideal dispersion. Because of the close ˚ for Ce3þ and matching of ionic radius (1.01 A 3þ ˚ for Y ), 0.9 A   VCe 1=3 dCeCe ¼ 2 ð4=3pÞ

G. Del Rosario et al. / Applied Surface Science 238 (2004) 469–474

473

Fig. 3. (a) PL spectra for powders of YAG:Ce treated to: 1000 8C (a), 1100 8C (b), 1200 8C (c). (b) The calculated distance of luminescent centres (dCe–Ce) as a function of energy emission (Eem) and intensity normalised emission for a wavelength fix of 533 nm.

where: VCe ¼ (1/Na)(Vm/XCe), Vm is the molar volume calculated from the weighed-out composition (YAG ¼ 130.38 cm3/mol); Na the Avogardo number; XCe the Ce molar fraction. Fig. 3b shows the obtained results, which are similar to reported [11–15]. High content of cerium reduces the distance between luminescent centres and increases the intensity of emission. In this case, two samples treated at different temperatures, 1000 and 1200 8C, present similar energy and different emission intensities because of the differences of Ce3þ content substituted in the garnet. As expected, the sample treated at 1100 8C presents the higher PL intensity due to the luminescent centres distance is lower. Therefore, it is confirmed that the observed changes in the morphology and crystanillity of the YAG:Ce powders, caused by the

applied thermal treatment, affect significantly at PL properties. TEM analysis was applied in order to evaluate crystallinity changes with temperature. Inner particle structure of all studied samples is porous, constituted by primary particles (<60 nm) randomly oriented (Fig. 4a). Selected area electron diffraction (SAED) patterns showed a cubic symmetry for all samples. Nevertheless, the specimen treated at lower temperature presents several zones with different orientations and interplannar spacing, which varies from 0.405 to 0.841 nm. The sample treated at 1100 8C (Fig. 4b) is the most homogeneous, with a constant interplannar spacing of 0.486 nm. These results are in agreement with detailed XRD evaluation and peak fitting in YAG:Ce system [11]. Twin domain has both cubic structure, as prevailing phase and also an intercrystal-

474

G. Del Rosario et al. / Applied Surface Science 238 (2004) 469–474

revealed by TEM, it is deduced that the studied samples have a high content of interfaces.

4. Conclusions The heat treatment of YAG:Ce powders modifies its morphology and crystallinity, varying the volume and texture of particle, the crystallite size and structure crystalline. The tendency of Y3Al5O12 phase to crystalline order with the temperature has been confirmed by the high growth rate of crystallite size ˚ 8C1). On the other hand, it has been probed (0.78 A that the treatment at 1200 8C causes an important loss of volatile compounds. Therefore, it is concluded that 1100 8C is the optima temperature, with that it is reached the highest homogeneity and the lowest distance between luminescent centres, increasing its PL properties. References

Fig. 4. TEM images of powder samples treated to 1100 8C. (a) Low magnification show particles transparent to electron beam. (b) High-resolution image at the structural level.

line phase, associated with high defect content. Taking into consideration high heating rates during synthesis (up to 300 8C/s) and the nanophased inner structure

[1] D. Jia, R.S. Meltzer, W.M. Yen, J. Lumin. 99 (2002) 1. [2] E.J. Bosze, G.A. Hirata, L.E. Shea-Rohwer, J. McKittrick, J. Lumin. 104 (2003) 47–54. [3] M.B. Kakabe, S. Ramanathan, P.V. Ravindran, J. Alloys Compd. 350 (2003) 123–139. [4] S.N. Patankar, D. Zhang, G. Adam, F.H. (Sam) Froes, J. Alloys Compd. 353 (1–2) (2003) 307–309. [5] N. Niolic, L. Mancic, Z. Marinkovic, O. Milosevic, M.M. Ristic, Ann. Chim. Sci. Mater. 26 (5) (2001) 35–41. [6] L. Mancic, O. Milosevic, B. Marinkovic, M. de F da Silva Lopez, F. Rizzo, Physica C 341–348 (2000) 503–504. [7] http://www.imagemet.com/. [8] C.E. Hall, Introduction to Electron Microscopy, Mc Graw Hill, New York, 1953 [9] R.C. Pullar, M.D. Taylor, A.K. Bhattacharya, J. Eur. Ceram. Soc. 18 (1988) 1759–1764. [10] Joaquı´n Bermu´ dez Polonio, Difraccio´ n de Rayos X, principios y aplicaciones, Madrid, 1981, pp. 354–362. [11] P. Dorenbos, J. Lumin. 99 (2002) 283–299. [12] J.L. Kennedy, N. Djeu, J. Lumin. 101 (2003) 147–153. [13] D. de Graaf, H.T. Hintzen, G. de With, J. Lumin. 104 (2003) 131–136. [14] O. Milosevic, L. Mancic, S. Ohara, G. del Rosario, P. Vulic, in: Paper Presented at ICCCI Conference, 24–27 September 2003, Kurashiki, Japan. [15] R.M. Macfarlane, J. Lumin. 100 (2002) 1–20.