- Email: [email protected]
Photoluminescence of Eu3+:Y2O3 as an indication of crystal structure and particle size in nanoparticles synthesized by flame spray pyrolysis
Aerosol Science 37 (2006) 402 – 412 www.elsevier.com/locate/jaerosci
Photoluminescence of Eu3+: Y2 O3 as an indication of crystal structure and particle size in nanoparticles synthesized by flame spray pyrolysis Dosi Dosev, Bing Guo, Ian M. Kennedy∗ Department of Mechanical and Aeronautical Engineering, University of California, Davis, CA 95616, USA Received 25 September 2004; received in revised form 3 May 2005; accepted 18 August 2005 Dedicated to the 70th birthday of Professor Daniel E. Rosner
Abstract Nanoparticles of europium-doped yttrium oxide (Eu: Y2 O3 ) were synthesized by flame spray pyrolysis. The nanoparticles were separated by centrifugation into two size groups (5–60 nm and 50–200 nm), each characterized by laser induced fluorescence spectroscopy, Transmission Electron Microscopy (TEM) and X-ray Diffraction (XRD). The fluorescence spectra, the electron diffraction pattern, and the XRD pattern of the large particles were typical of the stable cubic (Mn2 O3 type) phase of bulk Y2 O3 while those of small particles were quite different and indicated the possible presence of higher density metastable mixed phases—including monoclinic with some indication of a face-centered cubic phase. The size dependence of the particle properties could be attributable to the effect of surface free energy that elevated the internal particle pressure as size decreased. Doping with the lanthanide ion provided a new and useful diagnostic method for determining the crystal structure of flame-synthesized materials. 䉷 2005 Elsevier Ltd. All rights reserved. Keywords: Nanoparticles; Yttrium; Europium; Fluorescence; Spray pyrolysis
1. Introduction Gas processing of nanostructured materials offers significant advantages over liquid phase chemistry. The process is scalable to high production rates; it can yield material of high purity; a wide range of materials can be formed; and the process can be designed to be both environmentally benign, with no toxic by-products, and energetically efficient. The important characteristics of the product include the particle size distribution, composition and morphology. Rosner, McGraw, and Tandon (2003) and Rosner and Pyykonen (2002) have recognized the importance of multiple variables in the design and operation of gas phase synthesis processes and have developed an appropriate formalism for treating this problem numerically. Crystal structure may ultimately be predictable with such methods, and in some materials and applications, such as yttrium oxide (yttria), the crystal phase may be an important process variable. Yttrium oxide (Y2 O3 ) has often been used as a host material for phosphors and other optical applications and is conventionally processed from micron-sized powders that almost always contain small amounts of impurities. While ∗ Corresponding author. Tel.: +1 530 752 2796; fax: +1 120 620 22448.
E-mail address: [email protected] (I.M. Kennedy). 0021-8502/$ - see front matter 䉷 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.jaerosci.2005.08.009
D. Dosev et al. / Aerosol Science 37 (2006) 402 – 412
403
the size and quality of micron-sized powders may be adequate for conventional technologies, durability, mechanical strength, and infrared transparency (in the window from 3 to 5 m and beyond) is sought in refractory ceramics like yttria for missile and aerodynamic applications. Realization of these critical properties is highly dependent on the ability to reproducibly synthesize nanometer sized ceramic powders of single phases. The doping of lanthanides into yttria provides additional functionalities for this material. Lanthanide-doped nanoparticles have attracted a great deal of interest because of their high fluorescent intensity, large Stokes shift and long fluorescence lifetime (Bhargava, 1996; Tissue, 1998). They are used in the display industry (Wakefield, Holland, Dobson, & Hutchison, 2001) and show promise in sensor applications (Feng et al., 2003). This type of application requires a method for the production of nanopowders (ultra-fine particles with diameters below 100 nm) with high production rates (grams per hour range), at low cost, and with the ability to obtain materials with different photoluminescent spectra. Yttrium oxide (Y2 O3 ) is one of the best hosts for lanthanide ions (Hao, Studenikin, & Cocivera, 2001; Yang et al., 1999) because its ionic radius and crystal structure are very similar to many lanthanide oxides. Doping with a variety of lanthanide ions (Eu for red, Tb for green, Dy for yellow, Tm for blue) (Hao et al., 2001; Vetrone, Boyer, Capobianco, Speghini, & Bettinelli, 2004) can yield materials with different fluorescent spectra. The doping concentration of lanthanide ions into Y2 O3 is of key importance in determining the efficiency of fluorescence emission of these materials (Bazzi et al., 2003; Kang, Park, Lenggoro, & Okuyama, 1999). A wide variety of synthesis techniques have been developed for the production of pure and doped nanopowders, including wet chemical methods (Bazzi et al., 2003), laser ablation (Eilers & Tissue, 1996; Jones, Kumar, Singh, & Holloway, 1997) and combustion techniques (Hao et al., 2001; Kang, Roh, & Park, 2000; Kang, Roh, Park, & Park, 2002). Different sets of parameters for each synthesis method determine the structural and optical properties of the final products. The ability to measure and control these properties with good reproducibility is an important characteristic for any method of synthesis. In fact, it is very desirable to have an analytical method that may provide an online process control so that flow rates, temperatures, and feedstock can be adjusted to yield the desired product. In general, the physical characterization of nanoparticles for luminescent applications is performed by means of X-ray diffraction (XRD), transmission electron microscopy (TEM) or scanning electron microscopy (SEM) (Tissue, 1998). These techniques provide crystallographic characterization and enable evaluation of the particle size distribution, degree of aggregation and morphology. However, they are slow and require expensive equipment. Optical methods may provide a useful alternative in some cases. A number of optical methods have been used for the in situ characterization of combustion-generated nanoparticles. Elastic light scattering (Xing, Rosner, Koylu, & Tandon, 1997; Xing, Koylu, & Rosner, 1996, 1999) has been used to infer particle size and the fractal dimension of aggregates. Laser-induced incandescence has been used to obtain size characteristics of carbonaceous materials such as carbon nanotubes (Vander Wal, Berger, Ticich, & Patel, 2002). The presence of trace metals in aerosols can be measured with laser breakdown spectroscopy (Vander Wal, Ticich, West, & Householder, 1999). Arabi-Katbi, Pratsinis, Morrison, and Megaridis (2001) used FTIR spectroscopy to measure in situ flame and particle temperatures in the synthesis of anastase and rutile TiO2 nanoparticles—the crystallinity and phase were determined by ex situ thermophoretic sampling and XRD analysis. Spectroscopy has not been explored as a possible method for the rapid, and ideally in situ, determination of crystallinity and phase. Lanthanide-doped nanophosphors may offer the potential for a diagnostic of this type. Lanthanide atoms can occupy different crystallographic sites in the host crystal lattice. Different sites give rise to unique fluorescent spectra. As a characterization technique, optical studies of the lanthanide emission spectra are very sensitive probes of the crystal structure (Chen, Stump, Haire, & Peterson, 1992). This makes it possible to use laser-induced fluorescence (LIF) to study the crystal structure of the nanoparticles—so called site-selective optical spectroscopy (Eilers & Tissue, 1996; Williams, Yuan, & Tissue, 1999). The structural properties of Eu-doped Y2 O3 (Eu: Y2 O3 ) nanoparticles and their fluorescent spectra have been found to depend on the particle size in the case of material obtained by laser ablation followed by condensation (Tissue & Yuan, 2003). These changes arise from alteration to the crystal structure. Several crystal structures of Y2 O3 are possible. A cubic (Mn2 O3 type) crystal lattice is the stable equilibrium form for lanthanide oxides under standard state conditions. However, a monoclinic, high-density phase can be obtained during high pressure synthesis (Hoekstra & Gingerich, 1964). We have employed a conventional flame spray pyrolysis technique to produce europium-doped yttrium oxide (Eu: Y2 O3 ) nanoparticles with the ultimate purpose to use them as luminescent labels in bioassays. Our
404
D. Dosev et al. / Aerosol Science 37 (2006) 402 – 412
immediate goal is to examine the impact of nanoparticle size on crystal phase, and to demonstrate the feasibility of using spectroscopy as a diagnostic for the synthesis of materials such as yttria and lanthanide-doped yttria.
2. Experimental 2.1. Nanoparticle synthesis A schematic diagram of the burner used in this study is shown in Fig. 1. The burner consists of a nebulizer and a co-flow jacket. The nebulizer has an inner nozzle made of 20 gage SS304 capillary tube (0.81 mm OD) and an outer jacket. The inner nozzle extends through a hole in the outer jacket approximately 1 mm in diameter and ends flush with the top of the outer jacket. A narrow annular gap is formed between the inner nozzle and the outer jacket. An ethanol solution containing 2.5 mM Eu(NO3 )3 and 50 mM Y(NO3 )3 was pumped by a syringe pump (Cole–Parmer) at 40 mL/h into the inner nozzle of the nebulizer. Air, at 2 L/min, flowed through the annular gap surrounding the inner nozzle at high speed to atomize the ethanol solution containing the precursor materials. The solution was atomized to form a spray above the tip of the nebulizer. The co-flow jacket supplied H2 at 2 L/min and co-flow air at 10 L/min, to form a hydrogen-air diffusion flame surrounding the outlet of the nebulizer. The flame temperature in the H2 flame was measured with a coated thermocouple to be about 2100 ◦ C. The H2 diffusion flame served as a pilot for the spray flame that was formed by the combustible ethanol droplets containing the europium and yttrium precursors. Reactions took place within the flame to form Eu: Y2 O3 nanoparticles. A cold finger was used for collecting the Eu: Y2 O3 particles thermophoretically above the flame. The production rate of this synthesis procedure was about 400–500 mg/h. Fig. 2 shows the spray droplet size distribution, which was measured using a standard laser diffraction system (Model 2600, Malvern Instruments, UK). The droplet sizes were measured without ignition of the flame. Droplet diameters between 1 and 100 m were found. This wide droplet size distribution leads to a correspondingly wide particle size distribution in the nanoparticle powder that is formed. 2.2. Nanoparticle separation The as-synthesized particles were suspended in methanol in an ultrasonic bath for 30 min in order to break any weak agglomerates formed during the collection process. The primary Eu: Y2 O3 particles had a relatively wide size distribution from 5 to 300 nm as determined by TEM. Two portions with narrower size distributions were extracted using size selective centrifugation with a Sorvall䉸 RC5B Plus Centrifuge (Kendro Laboratory Products). First, the
Fig. 1. Schematic description of the burner used for synthesis of the nanoparticles.
D. Dosev et al. / Aerosol Science 37 (2006) 402 – 412
405
Fig. 2. Size distribution of the spray droplets injected in the flame.
Fig. 3. Size distributions of the two nanoparticle fractions, separated by centrifugation: (A) small particles; (B) big particles.
nanoparticles-methanol suspension was centrifuged at 1000 g for 16 min. In this step, all the nanoparticles with diameters bigger than about 50 nm were settled. The supernatant containing nanoparticles with diameters smaller than 50 nm was collected and the methanol was evaporated. 2.3. Nanoparticle characterization The size distributions of the two separated nanoparticle suspensions are shown in Fig. 3. They were determined using a Philips CM-12 TEM. More than 100 individual particle sizes were measured for each sample. The smaller particle portion contained particles with diameters from 5 to 60 nm. The larger particle portion contained particles with diameters from about 50 to 200 nm. There was a small size overlap between the two samples at about 50 nm as one can see from Fig. 3. This is expected with the centrifugal separation method. Representative TEM images for the two fractions are shown in Fig. 4. The collected particles were non-aggregated, a requirement for sensor applications. Both small and big particles exhibited a fully dense morphology. Many of the smaller particles (Fig. 4A) appeared to be faceted. On the other hand, the larger particles (Fig. 4B) had an overall
406
D. Dosev et al. / Aerosol Science 37 (2006) 402 – 412
Fig. 4. Representative bright field TEM images of nanoparticles separated by centrifugation: (A) small particle fraction; (B) large particle fraction.
spherical shape but in some cases, as seen in this example, smaller inclusions of the type seen in Fig. 4A were apparent in the center of the larger sphere. Close examination of images such as Fig. 4B suggested that the larger particles were polycrystalline with multiple domains apparent in the material surrounding the core. Optical characterization of the Eu: Y2 O3 nanoparticles was carried out by using LIF. Colloidal suspensions of 1 mg ml−1 in methanol were prepared from each of the two sized fractions. The nanoparticles were excited nonselectively with a pulsed laser beam of approximately 100 J at 260 nm wavelength, using an OpoletteTM tunable pulsed optical parametric oscillator (OPO) laser (Opotek, CA); we found that it was not possible to obtain useful spectra with site-selective excitation of the particles in the visible region of the spectrum due to the small concentration of particles and the relatively small absorption cross-sections at the visible wavelengths that correspond to crystal sites. The fluorescence spectra were recorded using a Spectra Pro 300i gated intensified spectrometer (Princeton Instruments Inc). The flame emitted a visible light when it was fed with yttrium and europium precursor. Thermal emission along an approximately 5 mm diameter beam across the flame, defined by collimating apertures, was recorded with the same spectrometer set-up.
3. Results The fluorescence spectrum of the small particle fraction is shown in Fig. 5. The spectrum exhibits a broad red emission with weakly discernible peaks at 616.2, 614.8 and 617.7 nm. According to Bihari, Eilers, and Tissue (1997), these peaks correspond to the 5 D0 → 7 D2 transition of Eu3+ ions, occupying the sites A (617.7), B (616.2) and C (614.8) of the monoclinic phase of Y2 O3 . Measurements at room temperature generally exhibit some inhomogeneous broadening of the spectral lines. In addition, small particles are more susceptible to surface defects and disorder than larger particles. Disordered states contribute to the broadening of the spectrum and could be reduced by annealing; our material was interrogated “as is” without annealing in order to avoid phase transformations. In addition, there is some evidence for the possible presence of mixed phases that have not been clearly identified. The joint effect of these phenomena leads to a spectrum in which the monoclinic A, B and C site Eu3+ transitions are small features in a broader emission. The fluorescent spectrum of the larger particle fraction is presented in Fig. 6. It is readily distinguished from the small particle spectrum of Fig. 5. The predominant peak is at 611.4 nm, which corresponds to the 5 D0 → 7 D2 transition of the C2 site of the cubic phase of Y2 O3 (Williams et al., 1999). In view of the hypersensitivity of the Eu3+ ion to its environment (Bihari et al., 1997; Chen et al., 1992), the results from Figs. 5 and 6 are a strong indication that nanoparticles in the two fractions (small and big) have different crystal structures. For the small particles, the spectroscopic data indicate the presence of some amount, possibly small, of monoclinic Y2 O3 , in agreement with the TEM results (discussed below). It is likely that surface defects and mixed phases are important contributors to the properties of the small particles. The larger particles are clearly cubic in light of the unambiguous spectroscopic information, although they are obtained by the same synthesis process. We attempted
D. Dosev et al. / Aerosol Science 37 (2006) 402 – 412
407
Fig. 5. Fluorescent spectrum of the small particles fraction after laser excitation at 260 nm. Top-right inset shows the predominant emission peak at a magnified scale.
Fig. 6. Fluorescent spectrum of the large particles fraction after laser excitation at 260 nm. Top-right inset shows the predominant emission peak at a magnified scale.
to measure LIF of the nanoparticles in situ, within the flame. Unfortunately, the OPO laser did not provide enough pulse energy (≈ 100 J) to enable us to detect emission from Eu3+ against the thermal emission from the flame. A high energy, pulsed Nd: YAG laser operating at 266 nm should be able to excite nanoparticle LIF at detectable levels in these materials as long as conditions permit crystallization. In order to confirm our spectroscopic observation of two distinct Y2 O3 particle morphologies, we preformed an analysis of the two samples by means of selected area electron diffraction. This allows sensitive studies to be performed on a very small amount of material under TEM. The electron diffraction pattern obtained from small particles (5–50 nm) is shown in Fig. 7A. Three rings are clearly distinguishable. From the ring radii, we calculated the d-spacings to be ˚ respectively. These numbers match very well with the d-spacings of monoclinic Y2 O3 according 2.95, 1.82 and 1.51 A, to the JCPDS database (JCPDS—International Centre for Diffraction Data). Fig. 7B shows the electron diffraction pattern of the large particle fraction. Two rings can be clearly observed with radii that indicate d-spacings of 2.60 and ˚ According to the JCPDS database, these d-spacings correspond to the cubic Y2 O3 phase. 1.22 A.
408
D. Dosev et al. / Aerosol Science 37 (2006) 402 – 412
Fig. 7. Electron diffraction patterns of Eu: Y2 O3 nanoparticles separated by centrifugation: (A) small particles fraction; (B) large particles fraction.
Fig. 8. X-ray diffraction spectra of the small particle fraction (top) and big particle fraction (bottom).
The crystal structure of the nanoparticles was also studied by means of X-ray diffraction using an Inel XRG3000 X-ray Diffractometer. The measured XRD spectra for the Eu: Y2 O3 nanoparticles of the two size groups are compared in Fig. 8. The XRD pattern of the large particles has low background and strong peaks that unambiguously match the JCPDS data for cubic (Mn2 O3 type) Y2 O3 . This result is in agreement with the fluorescence measurements and electron diffraction data. On the other hand, the measured XRD spectrum of the small particles has higher noise as a result of the very limited amount of sample that we could obtain; only two peaks can be clearly distinguished. A group of weak peaks centered at 2 = 32.1◦ is consistent with the presence of the monoclinic phase of Y2 O3 although the evidence is ambiguous. A relatively strong and sharp peak is observed at 2 = 29.4◦ . It does not match any of the expected peaks for cubic Y2 O3 (at 29.16◦ ) or monoclinic Y2 O3 (at 29.78). A peak at 2 = 29.41◦ has been reported for face-centered cubic (fcc)-structured Y2 O3 that was obtained by fast temperature quenching (Katagiri, Ishizawa, & Marumo, 1993).
D. Dosev et al. / Aerosol Science 37 (2006) 402 – 412
409
Fig. 9. Thermal emission spectra taken from the spray flame at different heights above the burner.
Spectroscopic examination of the flame and its products can provide information about the mechanism of aerosol formation. Particle formation from sprays can take place in the gas phase or entirely in the liquid phase. The optical thermal emission from gas phase species may indicate the presence of volatile species that can lead to aerosol product. Thermal emission from the spray flame was measured at a series of heights above the burner; the results are shown in Fig. 9. The spectra are similar throughout the flame with the double headed, red-degraded bands typical of a thermal emission sequence with peaks in the spectra at 598 and 614 nm. These features are consistent with the A2 1/2,3/2 → X2 + transitions in YO that have been reported by Chalek and Gole (1976) and Wijchers, Dijkerman, Zeegers, and Alkemade (1980). We do not observe any additional spectral lines in the region between 612 and 630 nm that could be associated with Eu and we conclude that EuO or associated species are not prevalent or detectable in the gas phase, although the low concentration of Eu precursor makes it relatively difficult to detect the presence of gas phase EuO.
4. Discussion The measured spray size distribution (Fig. 2) can be used to estimate the expected nanoparticle size distribution following pyrolysis. About 80% of the droplets have diameters between 2 and 70 m. With a 50 mMol solution of Y(NO3 )3 , there are 2 × 10−16 mol of dissolved nitrate in a 2 m droplet and 9 × 10−12 mol in a 70 m droplet. The oxidation reaction that leads to the formation of M2 O3 from M(NO3 )3 (where M = Y or Eu) is (Shikao & Jiye, 2001) 4M(NO3 )3 → 2M2 O3 + 12NO2 + 3O2 . Therefore, 2 mol of nitrate generate 1 mol of oxide, leading to between 1 × 10−16 and 4.5 × 10−12 mol of Y2 O3 product from droplets between 2 and 70 m. Using an overall molecular weight for 5% doped Eu: Y2 O3 of 232 g gmol−1 , we calculate that the mass of the oxide particles from this range of droplet diameters falls between 4.5 × 10−19 g and 2 × 10−14 g per particle. This corresponds to particle diameters between 5 and 200 nm, assuming a sphere with the density of Y2 O3 . The range of predicted particle sizes is consistent with the sizes that we observe in the final product. Although the thermal emission spectra showed that YO is present in the gas phase, it does not appear to be the dominant route to the formation of doped Eu: Y2 O3 . In fact, a gas phase route would most likely lead to separate yttrium and europium phases in the product that would manifest their presence as shortened lifetimes of the pure Eu2 O3 emission as a result of concentration quenching. We measured emission lifetimes in methanol: the smaller particles exhibited lifetimes between 800 s and 1 ms while the larger particles emitted with lifetimes between 1 and 1.5 ms. Pure cubic Eu2 O3 particles emit with a lifetime of about 100 s (Feng et al., 2003), significantly shorter than the doped Y2 O3
410
D. Dosev et al. / Aerosol Science 37 (2006) 402 – 412
particles and due to concentration quenching in the pure europium oxide. We conclude that separate Eu2 O3 phases are not significant in the product and that the gas phase route to nanoparticle formation is not dominant. The role of particle size in determining the crystal phase of nanoparticles is subject to continuing uncertainty for one of the most studied materials, TiO2 . Gribb and Banfield (1997) found that the anatase to rutile transformation in the temperature region around 500 ◦ C was favored with finer crystallite sizes. Zhang and Banfield (1998) noted that anatase TiO2 was the stable phase for crystallites less than 14 nm, a value that was explained by accounting for the roles of surface free energies and the related surface stress, the latter manifesting itself via an excess pressure within the particles. On the other hand, Kim, Nakaso, Xia, and Okuyama (2005) found that grain size in nanoparticles did not correlate with the anatase to rutile transformation. The primary effect seemed to be due to precursor concentration in their furnace synthesis; the authors speculated that carbon impurities may explain the observations. However, for pure particles it seems likely that size can have an impact on properties via surface stresses and internal pressure for several materials. For example, Ehrman and co-workers (Ehrman, 1998; Ehrman, Friedlander, & Zachariah, 1998) showed that pressure can affect the solid-state diffusivity in nanoparticles and may have an impact on the sintering and coalescence of SiO2 nanoparticles. In order to explain the observed differences in crystal structure between the large and the small particle fractions, one must recognize that the cubic phase of Y2 O3 is an equilibrium phase that is favored under normal ambient conditions of temperature and pressure, while the monoclinic phase is typical of synthesis processes that involve high pressures (Hoekstra & Gingerich, 1964; Skandan et al., 1992). For example, Husson, Proust, Gillet, and Itie (1999) used Raman spectroscopy to find a high pressure phase transition in Y2 O3 from cubic to monoclinic at room temperature. The formation of monoclinic Eu: Y2 O3 nanoparticles with diameters about 20 nm has been reported for a laser ablation gas-phase condensation synthesis, and has been attributed to the increased surface stress that leads to the formation of particles in the higher density phase (Tissue & Yuan, 2003). In spray pyrolysis, we can estimate the pressure within a particle by assuming that each particle is formed from a single spherical molten droplet of Y2 O3 . The pressure p in excess of ambient within a particle of diameter D can be obtained from a hydrostatic force balance as p = 4/D, where the surface energy is taken to be 1.5 J m−2 , a value that is typical of many ceramics. The equation shows that smaller particles will experience a higher internal pressure. For a particle diameter of 20 nm (see Fig. 3A), the estimated pressure is 0.3 × 109 Pa (3 kbar). The actual pressure in our case may be higher than this value because the surface energy, , can be higher in nanoscale particles than for a flat surface (Skandan et al., 1992) although the effect is still open to some debate. For bulk material, the pressure at the cubic-to-monoclinic transition is about 1.5 × 109 Pa (Skandan et al., 1992). Our results suggest that in the case of small particles, values of the pressures may be sufficiently large to permit a monoclinic crystalline structure to form. It should also be noted that the diffusion flame that was used in this work is a very inhomogeneous environment for particle synthesis, with different temperature histories that can be experienced by particles, although the peak temperatures (about 2100 ◦ C) that are seen by all particles will be similar. These variations may also play some role in influencing the ultimate morphology of the product and may lead to the mixed phases that we see in the smallest particles. 5. Conclusions The fluorescent and crystalline properties of Eu-doped Y2 O3 nanoparticles obtained by flame spray pyrolysis showed a dependence on the particle size, a result that has not been reported before with this synthesis method. Fluorescent spectra, electron diffraction and X-ray diffraction demonstrated that particles larger than about 50 nm had a cubic Mn2 O3 -type structure that was typical of bulk material. Particles smaller than 50 nm exhibited a more complex structure with an indication, based on electron diffraction data, of the presence of the monoclinic phase with minor spectroscopic features that are consistent with this phase. There was also an indication that mixed phases may be present in the smaller size fraction. This size dependence of particle morphology may be attributable to an elevated internal pressure as particle diameter decreases. LIF of the nanopowder samples obtained from the spray flame provided a relatively straightforward diagnostic for crystal phase to the extent that a well-ordered cubic phase is clearly distinguished from a less well-ordered mixed phase in the smaller particles. Fluorescence lifetime measurements indicated the absence
D. Dosev et al. / Aerosol Science 37 (2006) 402 – 412
411
of phase separation in Eu–Y mixtures. Spectroscopy can be used as a simpler alternative to complex TEM and XRD as a method for determining crystal phases in materials such as Y2 O3 , and may have application to process control in flame synthesis processes. Higher energy pulsed laser excitation and more efficient collection optics should permit nanoparticle LIF to be used as an in situ diagnostic tool for crystal formation and phase in aerosol processing of nanophosphor materials. Modeling of the dynamics of particle morphology in multi-dimensional calculations would be aided by diagnostic tools of this sort. Acknowledgements The authors acknowledge the assistance and cooperation of Dr. K.D. Giles and Dr. D. Downey from the Department of Biological and Agricultural Engineering, UC Davis, for droplet size measurements. The assistance of Mr. J. Neil and Professor A. Navrotsky with X-ray diffraction is also appreciated. The authors wish to acknowledge the support of the National Science Foundation, NIRT Grant DBI-0102662 and the Superfund Basic Research Program with Grant 5P42ES04699 from the National Institute of Environmental Health Sciences, NIH. References Arabi-Katbi, O. I., Pratsinis, S. E., Morrison, P. W., & Megaridis, C. M. (2001). Monitoring the flame synthesis of TiO2 particles by in-situ FTIR spectroscopy and thermophoretic sampling. Combustion and Flame, 124, 560. Bazzi, R., Flores-Gonzalez, M. A., Louis, C., Lebbou, K., Dujardin, C., Brenier, A. et al. (2003). Synthesis and luminescent properties of sub-5-nm lanthanide oxides nanoparticles. Journal of Luminescence, 102–103, 445. Bhargava, R. N. (1996). Doped nanocrystalline materials—Physics and applications. Journal of Luminescence, 70, 85. Bihari, B., Eilers, H., & Tissue, B. M. (1997). Spectra and dynamics of monoclinic Eu2 O3 and Eu3+ : Y2 O3 nanocrystals. Journal of Luminescence, 75, 1. Chalek, C. L., & Gole, J. L. (1976). Chemiluminescence spectra of Sc and YO: Observation and analysis of the A 2 → X2 + band system. Journal of Chemical Physics, 65, 2845. Chen, G., Stump, N. A., Haire, R. G., & Peterson, J. R. (1992). Study of the phase behavior of Eu2 O3 under pressure via luminescence of Eu3+ . Journal of Alloys and Compounds, 181, 503. Ehrman, S. H. (1998). Effect of internal pressure on nanoparticle coalescence. Journal of Aerosol Science, 29, S31. Ehrman, S. H., Friedlander, S. K., & Zachariah, M. R. (1998). Characteristics of SiO2 /TiO2 nanocomposite particles formed in a premixed flat flame. Journal of Aerosol Science, 29, 687. Eilers, H., & Tissue, B. M. (1996). Laser spectroscopy of nanocrystalline Eu2 O3 and Eu3+ : Y2 O3 . Chemical Physics Letters, 251, 74. Feng, J., Shan, G. M., Maquieira, A., Koivunen, M. E., Guo, B., Hammock, B. D. et al. (2003). Functionalized europium oxide nanoparticles used as a fluorescent label in an immunoassay for atrazine. Analytical Chemistry, 75, 5282. Gribb, A. A., & Banfield, J. F. (1997). Particle size effects on transformation kinetics and phase stability in nanocrystalline TiO2 . American Mineralogist, 82, 717. Hao, J., Studenikin, S. A., & Cocivera, M. (2001). Blue, green and red cathodoluminescence of Y2 O3 phosphor films prepared by spray pyrolysis. Journal of Luminescence, 93, 313. Hoekstra, H. R., & Gingerich, K. A. (1964). High-pressure B-type polymorphs of some rare-Earth sesquioxides. Science, 146, 1163. Husson, E., Proust, C., Gillet, P., & Itie, J. P. (1999). Phase transitions in yttrium oxide at high pressure studied by Raman Spectroscopy. Materials Research Bulletin, 34, 2085–2092. Jones, S. L., Kumar, D., Singh, R. K., & Holloway, P. H. (1997). Luminescence of pulsed laser deposited Eu doped yttrium oxide films. Applied Physics Letters, 71, 404. Kang, Y. C., Park, S. B., Lenggoro, I. W., & Okuyama, K. (1999). Gd2 O3 :Eu phosphor particles with sphericity, submicron size and non-aggregation characteristics. Journal of Physics and Chemistry of Solids, 60, 379. Kang, Y. C., Roh, H. S., & Park, S. B. (2000). Preparation of Y2 O3 :Eu phosphor particles of filled morphology at high precursor concentrations by spray pyrolysis. Advances in Materials, 12, 451. Kang, Y. C., Roh, H. S., Park, S. B., & Park, H. D. (2002). High luminescence Y2 O3 :Eu phosphor particles prepared by modified spray pyrolysis. Journal of Materials Science Letters, 21, 1027. Katagiri, S., Ishizawa, N., & Marumo, F. (1993). A new high temperature modification of face-centered cubic Y2 O3 . Powder Diffraction, 8, 60. Kim, C. S., Nakaso, K., Xia, B., & Okuyama, K. (2005). A new observation on the phase transformation of TiO2 nanoparticles produced by a CVD method. Aerosol Science and Technology, 39, 104. Rosner, D. E., McGraw, R., & Tandon, P. (2003). Multivariate population balances via moment and Monte Carlo simulation methods: An important sol reaction engineering bivariate example and “mixed” moments for the estimation of deposition, scavenging, and optical properties for populations of nonspherical suspended particles. Industrial & Engineering Chemistry Research, 42, 2699. Rosner, D. E., & Pyykonen, J. J. (2002). Bivariate moment simulation of coagulating and sintering nanoparticles in flames. AIChE Journal, 48, 476. Shikao, S., & Jiye, W. (2001). Combustion synthesis of Eu3+ activated Y3 Al5 O12 phosphor nanoparticles. Journal of Alloys and Compounds, 327, 82.
412
D. Dosev et al. / Aerosol Science 37 (2006) 402 – 412
Skandan, G., Foster, C. M., Frase, H., Ali, M. N., Parker, J. C., & Hahn, H. (1992). Phase characterization and stabilization due to grain size effects of nanostructured Y2 O3 . Nanostructured Materials, 1, 313. Tissue, B. M. (1998). Synthesis and luminescence of lanthanide ions in nanoscale insulating hosts. Chemical Materials, 10, 2837. Tissue, B. M., & Yuan, H. B. (2003). Structure, particle size, and annealing of gas phase-condensed Eu3+ : Y2 O3 nanophosphors. Journal of Solid State Chemistry, 171, 12. Vander Wal, R. L., Berger, G. M., Ticich, T. M., & Patel, P. D. (2002). Application of laser-induced incandescence to the detection of carbon nanotubes and carbon nanofibers. Applied Optics, 41, 5678. Vander Wal, R. L., Ticich, T. M., West, J. R., & Householder, P. A. (1999). Trace metal detection by laser-induced breakdown spectroscopy. Applied Spectroscopy, 53, 1226. Vetrone, F., Boyer, J. C., Capobianco, J. A., Speghini, A., & Bettinelli, M. (2004). A spectroscopic investigation of trivalent lanthanide doped Y2 O3 nanocrystals. Nanotechonology, 15, 75. Wakefield, G., Holland, E., Dobson, P. J., & Hutchison, J. L. (2001). Luminescence properties of nanocrystalline Y2 O3 :Eu. Advances in Materials, 13, 1557. Wijchers, T., Dijkerman, H. A., Zeegers, P. J. T., & Alkemade, C. T. J. (1980). Fluorescence spectra of laser-excited YO molecules in a H2 -O2 -Ar flame. Spectrochimica Acta, 35B, 271. Williams, D. K., Yuan, H., & Tissue, B. M. (1999). Size dependence of the luminescence spectra and dynamics of Eu3+ : Y2 O3 nanocrystals. Journal of Luminescence, 83–84, 297. Xing,Y. C., Koylu, U. O., & Rosner, D. E. (1996). Synthesis and restructuring of inorganic nano-particles in counterflow diffusion flames. Combustion and Flame, 107, 85. Xing, Y. C., Koylu, U. O., & Rosner, D. E. (1999). In situ light-scattering measurements of morphologically evolving flame-synthesized oxide nanoaggregates. Applied Optics, 38, 2686. Xing, Y., Rosner, D. E., Koylu, U. O., & Tandon, P. (1997). Morphological evolution of nanoparticles in diffusion flames: Measurements and modeling. AIChE Journal, 43, 2641. Yang, H.-S., Hong, K. S., Feofilov, S. P., Tissue, B. M., Meltzer, R. S., & Dennis, W. M. (1999). Electron-phonon interaction in rare earth doped nanocrystals. Journal of Luminescence, 83–84, 139. Zhang, H. Z., & Banfield, J. F. (1998). Thermodynamic analysis of phase stability of nanocrystalline titania. Journal of Materials Chemistry, 8, 2073.
Photoluminescence of Eu3+: Y2 O3 as an indication of crystal structure and particle size in nanoparticles synthesized by flame spray pyrolysis Dosi Dosev, Bing Guo, Ian M. Kennedy∗ Department of Mechanical and Aeronautical Engineering, University of California, Davis, CA 95616, USA Received 25 September 2004; received in revised form 3 May 2005; accepted 18 August 2005 Dedicated to the 70th birthday of Professor Daniel E. Rosner
Abstract Nanoparticles of europium-doped yttrium oxide (Eu: Y2 O3 ) were synthesized by flame spray pyrolysis. The nanoparticles were separated by centrifugation into two size groups (5–60 nm and 50–200 nm), each characterized by laser induced fluorescence spectroscopy, Transmission Electron Microscopy (TEM) and X-ray Diffraction (XRD). The fluorescence spectra, the electron diffraction pattern, and the XRD pattern of the large particles were typical of the stable cubic (Mn2 O3 type) phase of bulk Y2 O3 while those of small particles were quite different and indicated the possible presence of higher density metastable mixed phases—including monoclinic with some indication of a face-centered cubic phase. The size dependence of the particle properties could be attributable to the effect of surface free energy that elevated the internal particle pressure as size decreased. Doping with the lanthanide ion provided a new and useful diagnostic method for determining the crystal structure of flame-synthesized materials. 䉷 2005 Elsevier Ltd. All rights reserved. Keywords: Nanoparticles; Yttrium; Europium; Fluorescence; Spray pyrolysis
1. Introduction Gas processing of nanostructured materials offers significant advantages over liquid phase chemistry. The process is scalable to high production rates; it can yield material of high purity; a wide range of materials can be formed; and the process can be designed to be both environmentally benign, with no toxic by-products, and energetically efficient. The important characteristics of the product include the particle size distribution, composition and morphology. Rosner, McGraw, and Tandon (2003) and Rosner and Pyykonen (2002) have recognized the importance of multiple variables in the design and operation of gas phase synthesis processes and have developed an appropriate formalism for treating this problem numerically. Crystal structure may ultimately be predictable with such methods, and in some materials and applications, such as yttrium oxide (yttria), the crystal phase may be an important process variable. Yttrium oxide (Y2 O3 ) has often been used as a host material for phosphors and other optical applications and is conventionally processed from micron-sized powders that almost always contain small amounts of impurities. While ∗ Corresponding author. Tel.: +1 530 752 2796; fax: +1 120 620 22448.
E-mail address: [email protected] (I.M. Kennedy). 0021-8502/$ - see front matter 䉷 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.jaerosci.2005.08.009
D. Dosev et al. / Aerosol Science 37 (2006) 402 – 412
403
the size and quality of micron-sized powders may be adequate for conventional technologies, durability, mechanical strength, and infrared transparency (in the window from 3 to 5 m and beyond) is sought in refractory ceramics like yttria for missile and aerodynamic applications. Realization of these critical properties is highly dependent on the ability to reproducibly synthesize nanometer sized ceramic powders of single phases. The doping of lanthanides into yttria provides additional functionalities for this material. Lanthanide-doped nanoparticles have attracted a great deal of interest because of their high fluorescent intensity, large Stokes shift and long fluorescence lifetime (Bhargava, 1996; Tissue, 1998). They are used in the display industry (Wakefield, Holland, Dobson, & Hutchison, 2001) and show promise in sensor applications (Feng et al., 2003). This type of application requires a method for the production of nanopowders (ultra-fine particles with diameters below 100 nm) with high production rates (grams per hour range), at low cost, and with the ability to obtain materials with different photoluminescent spectra. Yttrium oxide (Y2 O3 ) is one of the best hosts for lanthanide ions (Hao, Studenikin, & Cocivera, 2001; Yang et al., 1999) because its ionic radius and crystal structure are very similar to many lanthanide oxides. Doping with a variety of lanthanide ions (Eu for red, Tb for green, Dy for yellow, Tm for blue) (Hao et al., 2001; Vetrone, Boyer, Capobianco, Speghini, & Bettinelli, 2004) can yield materials with different fluorescent spectra. The doping concentration of lanthanide ions into Y2 O3 is of key importance in determining the efficiency of fluorescence emission of these materials (Bazzi et al., 2003; Kang, Park, Lenggoro, & Okuyama, 1999). A wide variety of synthesis techniques have been developed for the production of pure and doped nanopowders, including wet chemical methods (Bazzi et al., 2003), laser ablation (Eilers & Tissue, 1996; Jones, Kumar, Singh, & Holloway, 1997) and combustion techniques (Hao et al., 2001; Kang, Roh, & Park, 2000; Kang, Roh, Park, & Park, 2002). Different sets of parameters for each synthesis method determine the structural and optical properties of the final products. The ability to measure and control these properties with good reproducibility is an important characteristic for any method of synthesis. In fact, it is very desirable to have an analytical method that may provide an online process control so that flow rates, temperatures, and feedstock can be adjusted to yield the desired product. In general, the physical characterization of nanoparticles for luminescent applications is performed by means of X-ray diffraction (XRD), transmission electron microscopy (TEM) or scanning electron microscopy (SEM) (Tissue, 1998). These techniques provide crystallographic characterization and enable evaluation of the particle size distribution, degree of aggregation and morphology. However, they are slow and require expensive equipment. Optical methods may provide a useful alternative in some cases. A number of optical methods have been used for the in situ characterization of combustion-generated nanoparticles. Elastic light scattering (Xing, Rosner, Koylu, & Tandon, 1997; Xing, Koylu, & Rosner, 1996, 1999) has been used to infer particle size and the fractal dimension of aggregates. Laser-induced incandescence has been used to obtain size characteristics of carbonaceous materials such as carbon nanotubes (Vander Wal, Berger, Ticich, & Patel, 2002). The presence of trace metals in aerosols can be measured with laser breakdown spectroscopy (Vander Wal, Ticich, West, & Householder, 1999). Arabi-Katbi, Pratsinis, Morrison, and Megaridis (2001) used FTIR spectroscopy to measure in situ flame and particle temperatures in the synthesis of anastase and rutile TiO2 nanoparticles—the crystallinity and phase were determined by ex situ thermophoretic sampling and XRD analysis. Spectroscopy has not been explored as a possible method for the rapid, and ideally in situ, determination of crystallinity and phase. Lanthanide-doped nanophosphors may offer the potential for a diagnostic of this type. Lanthanide atoms can occupy different crystallographic sites in the host crystal lattice. Different sites give rise to unique fluorescent spectra. As a characterization technique, optical studies of the lanthanide emission spectra are very sensitive probes of the crystal structure (Chen, Stump, Haire, & Peterson, 1992). This makes it possible to use laser-induced fluorescence (LIF) to study the crystal structure of the nanoparticles—so called site-selective optical spectroscopy (Eilers & Tissue, 1996; Williams, Yuan, & Tissue, 1999). The structural properties of Eu-doped Y2 O3 (Eu: Y2 O3 ) nanoparticles and their fluorescent spectra have been found to depend on the particle size in the case of material obtained by laser ablation followed by condensation (Tissue & Yuan, 2003). These changes arise from alteration to the crystal structure. Several crystal structures of Y2 O3 are possible. A cubic (Mn2 O3 type) crystal lattice is the stable equilibrium form for lanthanide oxides under standard state conditions. However, a monoclinic, high-density phase can be obtained during high pressure synthesis (Hoekstra & Gingerich, 1964). We have employed a conventional flame spray pyrolysis technique to produce europium-doped yttrium oxide (Eu: Y2 O3 ) nanoparticles with the ultimate purpose to use them as luminescent labels in bioassays. Our
404
D. Dosev et al. / Aerosol Science 37 (2006) 402 – 412
immediate goal is to examine the impact of nanoparticle size on crystal phase, and to demonstrate the feasibility of using spectroscopy as a diagnostic for the synthesis of materials such as yttria and lanthanide-doped yttria.
2. Experimental 2.1. Nanoparticle synthesis A schematic diagram of the burner used in this study is shown in Fig. 1. The burner consists of a nebulizer and a co-flow jacket. The nebulizer has an inner nozzle made of 20 gage SS304 capillary tube (0.81 mm OD) and an outer jacket. The inner nozzle extends through a hole in the outer jacket approximately 1 mm in diameter and ends flush with the top of the outer jacket. A narrow annular gap is formed between the inner nozzle and the outer jacket. An ethanol solution containing 2.5 mM Eu(NO3 )3 and 50 mM Y(NO3 )3 was pumped by a syringe pump (Cole–Parmer) at 40 mL/h into the inner nozzle of the nebulizer. Air, at 2 L/min, flowed through the annular gap surrounding the inner nozzle at high speed to atomize the ethanol solution containing the precursor materials. The solution was atomized to form a spray above the tip of the nebulizer. The co-flow jacket supplied H2 at 2 L/min and co-flow air at 10 L/min, to form a hydrogen-air diffusion flame surrounding the outlet of the nebulizer. The flame temperature in the H2 flame was measured with a coated thermocouple to be about 2100 ◦ C. The H2 diffusion flame served as a pilot for the spray flame that was formed by the combustible ethanol droplets containing the europium and yttrium precursors. Reactions took place within the flame to form Eu: Y2 O3 nanoparticles. A cold finger was used for collecting the Eu: Y2 O3 particles thermophoretically above the flame. The production rate of this synthesis procedure was about 400–500 mg/h. Fig. 2 shows the spray droplet size distribution, which was measured using a standard laser diffraction system (Model 2600, Malvern Instruments, UK). The droplet sizes were measured without ignition of the flame. Droplet diameters between 1 and 100 m were found. This wide droplet size distribution leads to a correspondingly wide particle size distribution in the nanoparticle powder that is formed. 2.2. Nanoparticle separation The as-synthesized particles were suspended in methanol in an ultrasonic bath for 30 min in order to break any weak agglomerates formed during the collection process. The primary Eu: Y2 O3 particles had a relatively wide size distribution from 5 to 300 nm as determined by TEM. Two portions with narrower size distributions were extracted using size selective centrifugation with a Sorvall䉸 RC5B Plus Centrifuge (Kendro Laboratory Products). First, the
Fig. 1. Schematic description of the burner used for synthesis of the nanoparticles.
D. Dosev et al. / Aerosol Science 37 (2006) 402 – 412
405
Fig. 2. Size distribution of the spray droplets injected in the flame.
Fig. 3. Size distributions of the two nanoparticle fractions, separated by centrifugation: (A) small particles; (B) big particles.
nanoparticles-methanol suspension was centrifuged at 1000 g for 16 min. In this step, all the nanoparticles with diameters bigger than about 50 nm were settled. The supernatant containing nanoparticles with diameters smaller than 50 nm was collected and the methanol was evaporated. 2.3. Nanoparticle characterization The size distributions of the two separated nanoparticle suspensions are shown in Fig. 3. They were determined using a Philips CM-12 TEM. More than 100 individual particle sizes were measured for each sample. The smaller particle portion contained particles with diameters from 5 to 60 nm. The larger particle portion contained particles with diameters from about 50 to 200 nm. There was a small size overlap between the two samples at about 50 nm as one can see from Fig. 3. This is expected with the centrifugal separation method. Representative TEM images for the two fractions are shown in Fig. 4. The collected particles were non-aggregated, a requirement for sensor applications. Both small and big particles exhibited a fully dense morphology. Many of the smaller particles (Fig. 4A) appeared to be faceted. On the other hand, the larger particles (Fig. 4B) had an overall
406
D. Dosev et al. / Aerosol Science 37 (2006) 402 – 412
Fig. 4. Representative bright field TEM images of nanoparticles separated by centrifugation: (A) small particle fraction; (B) large particle fraction.
spherical shape but in some cases, as seen in this example, smaller inclusions of the type seen in Fig. 4A were apparent in the center of the larger sphere. Close examination of images such as Fig. 4B suggested that the larger particles were polycrystalline with multiple domains apparent in the material surrounding the core. Optical characterization of the Eu: Y2 O3 nanoparticles was carried out by using LIF. Colloidal suspensions of 1 mg ml−1 in methanol were prepared from each of the two sized fractions. The nanoparticles were excited nonselectively with a pulsed laser beam of approximately 100 J at 260 nm wavelength, using an OpoletteTM tunable pulsed optical parametric oscillator (OPO) laser (Opotek, CA); we found that it was not possible to obtain useful spectra with site-selective excitation of the particles in the visible region of the spectrum due to the small concentration of particles and the relatively small absorption cross-sections at the visible wavelengths that correspond to crystal sites. The fluorescence spectra were recorded using a Spectra Pro 300i gated intensified spectrometer (Princeton Instruments Inc). The flame emitted a visible light when it was fed with yttrium and europium precursor. Thermal emission along an approximately 5 mm diameter beam across the flame, defined by collimating apertures, was recorded with the same spectrometer set-up.
3. Results The fluorescence spectrum of the small particle fraction is shown in Fig. 5. The spectrum exhibits a broad red emission with weakly discernible peaks at 616.2, 614.8 and 617.7 nm. According to Bihari, Eilers, and Tissue (1997), these peaks correspond to the 5 D0 → 7 D2 transition of Eu3+ ions, occupying the sites A (617.7), B (616.2) and C (614.8) of the monoclinic phase of Y2 O3 . Measurements at room temperature generally exhibit some inhomogeneous broadening of the spectral lines. In addition, small particles are more susceptible to surface defects and disorder than larger particles. Disordered states contribute to the broadening of the spectrum and could be reduced by annealing; our material was interrogated “as is” without annealing in order to avoid phase transformations. In addition, there is some evidence for the possible presence of mixed phases that have not been clearly identified. The joint effect of these phenomena leads to a spectrum in which the monoclinic A, B and C site Eu3+ transitions are small features in a broader emission. The fluorescent spectrum of the larger particle fraction is presented in Fig. 6. It is readily distinguished from the small particle spectrum of Fig. 5. The predominant peak is at 611.4 nm, which corresponds to the 5 D0 → 7 D2 transition of the C2 site of the cubic phase of Y2 O3 (Williams et al., 1999). In view of the hypersensitivity of the Eu3+ ion to its environment (Bihari et al., 1997; Chen et al., 1992), the results from Figs. 5 and 6 are a strong indication that nanoparticles in the two fractions (small and big) have different crystal structures. For the small particles, the spectroscopic data indicate the presence of some amount, possibly small, of monoclinic Y2 O3 , in agreement with the TEM results (discussed below). It is likely that surface defects and mixed phases are important contributors to the properties of the small particles. The larger particles are clearly cubic in light of the unambiguous spectroscopic information, although they are obtained by the same synthesis process. We attempted
D. Dosev et al. / Aerosol Science 37 (2006) 402 – 412
407
Fig. 5. Fluorescent spectrum of the small particles fraction after laser excitation at 260 nm. Top-right inset shows the predominant emission peak at a magnified scale.
Fig. 6. Fluorescent spectrum of the large particles fraction after laser excitation at 260 nm. Top-right inset shows the predominant emission peak at a magnified scale.
to measure LIF of the nanoparticles in situ, within the flame. Unfortunately, the OPO laser did not provide enough pulse energy (≈ 100 J) to enable us to detect emission from Eu3+ against the thermal emission from the flame. A high energy, pulsed Nd: YAG laser operating at 266 nm should be able to excite nanoparticle LIF at detectable levels in these materials as long as conditions permit crystallization. In order to confirm our spectroscopic observation of two distinct Y2 O3 particle morphologies, we preformed an analysis of the two samples by means of selected area electron diffraction. This allows sensitive studies to be performed on a very small amount of material under TEM. The electron diffraction pattern obtained from small particles (5–50 nm) is shown in Fig. 7A. Three rings are clearly distinguishable. From the ring radii, we calculated the d-spacings to be ˚ respectively. These numbers match very well with the d-spacings of monoclinic Y2 O3 according 2.95, 1.82 and 1.51 A, to the JCPDS database (JCPDS—International Centre for Diffraction Data). Fig. 7B shows the electron diffraction pattern of the large particle fraction. Two rings can be clearly observed with radii that indicate d-spacings of 2.60 and ˚ According to the JCPDS database, these d-spacings correspond to the cubic Y2 O3 phase. 1.22 A.
408
D. Dosev et al. / Aerosol Science 37 (2006) 402 – 412
Fig. 7. Electron diffraction patterns of Eu: Y2 O3 nanoparticles separated by centrifugation: (A) small particles fraction; (B) large particles fraction.
Fig. 8. X-ray diffraction spectra of the small particle fraction (top) and big particle fraction (bottom).
The crystal structure of the nanoparticles was also studied by means of X-ray diffraction using an Inel XRG3000 X-ray Diffractometer. The measured XRD spectra for the Eu: Y2 O3 nanoparticles of the two size groups are compared in Fig. 8. The XRD pattern of the large particles has low background and strong peaks that unambiguously match the JCPDS data for cubic (Mn2 O3 type) Y2 O3 . This result is in agreement with the fluorescence measurements and electron diffraction data. On the other hand, the measured XRD spectrum of the small particles has higher noise as a result of the very limited amount of sample that we could obtain; only two peaks can be clearly distinguished. A group of weak peaks centered at 2 = 32.1◦ is consistent with the presence of the monoclinic phase of Y2 O3 although the evidence is ambiguous. A relatively strong and sharp peak is observed at 2 = 29.4◦ . It does not match any of the expected peaks for cubic Y2 O3 (at 29.16◦ ) or monoclinic Y2 O3 (at 29.78). A peak at 2 = 29.41◦ has been reported for face-centered cubic (fcc)-structured Y2 O3 that was obtained by fast temperature quenching (Katagiri, Ishizawa, & Marumo, 1993).
D. Dosev et al. / Aerosol Science 37 (2006) 402 – 412
409
Fig. 9. Thermal emission spectra taken from the spray flame at different heights above the burner.
Spectroscopic examination of the flame and its products can provide information about the mechanism of aerosol formation. Particle formation from sprays can take place in the gas phase or entirely in the liquid phase. The optical thermal emission from gas phase species may indicate the presence of volatile species that can lead to aerosol product. Thermal emission from the spray flame was measured at a series of heights above the burner; the results are shown in Fig. 9. The spectra are similar throughout the flame with the double headed, red-degraded bands typical of a thermal emission sequence with peaks in the spectra at 598 and 614 nm. These features are consistent with the A2 1/2,3/2 → X2 + transitions in YO that have been reported by Chalek and Gole (1976) and Wijchers, Dijkerman, Zeegers, and Alkemade (1980). We do not observe any additional spectral lines in the region between 612 and 630 nm that could be associated with Eu and we conclude that EuO or associated species are not prevalent or detectable in the gas phase, although the low concentration of Eu precursor makes it relatively difficult to detect the presence of gas phase EuO.
4. Discussion The measured spray size distribution (Fig. 2) can be used to estimate the expected nanoparticle size distribution following pyrolysis. About 80% of the droplets have diameters between 2 and 70 m. With a 50 mMol solution of Y(NO3 )3 , there are 2 × 10−16 mol of dissolved nitrate in a 2 m droplet and 9 × 10−12 mol in a 70 m droplet. The oxidation reaction that leads to the formation of M2 O3 from M(NO3 )3 (where M = Y or Eu) is (Shikao & Jiye, 2001) 4M(NO3 )3 → 2M2 O3 + 12NO2 + 3O2 . Therefore, 2 mol of nitrate generate 1 mol of oxide, leading to between 1 × 10−16 and 4.5 × 10−12 mol of Y2 O3 product from droplets between 2 and 70 m. Using an overall molecular weight for 5% doped Eu: Y2 O3 of 232 g gmol−1 , we calculate that the mass of the oxide particles from this range of droplet diameters falls between 4.5 × 10−19 g and 2 × 10−14 g per particle. This corresponds to particle diameters between 5 and 200 nm, assuming a sphere with the density of Y2 O3 . The range of predicted particle sizes is consistent with the sizes that we observe in the final product. Although the thermal emission spectra showed that YO is present in the gas phase, it does not appear to be the dominant route to the formation of doped Eu: Y2 O3 . In fact, a gas phase route would most likely lead to separate yttrium and europium phases in the product that would manifest their presence as shortened lifetimes of the pure Eu2 O3 emission as a result of concentration quenching. We measured emission lifetimes in methanol: the smaller particles exhibited lifetimes between 800 s and 1 ms while the larger particles emitted with lifetimes between 1 and 1.5 ms. Pure cubic Eu2 O3 particles emit with a lifetime of about 100 s (Feng et al., 2003), significantly shorter than the doped Y2 O3
410
D. Dosev et al. / Aerosol Science 37 (2006) 402 – 412
particles and due to concentration quenching in the pure europium oxide. We conclude that separate Eu2 O3 phases are not significant in the product and that the gas phase route to nanoparticle formation is not dominant. The role of particle size in determining the crystal phase of nanoparticles is subject to continuing uncertainty for one of the most studied materials, TiO2 . Gribb and Banfield (1997) found that the anatase to rutile transformation in the temperature region around 500 ◦ C was favored with finer crystallite sizes. Zhang and Banfield (1998) noted that anatase TiO2 was the stable phase for crystallites less than 14 nm, a value that was explained by accounting for the roles of surface free energies and the related surface stress, the latter manifesting itself via an excess pressure within the particles. On the other hand, Kim, Nakaso, Xia, and Okuyama (2005) found that grain size in nanoparticles did not correlate with the anatase to rutile transformation. The primary effect seemed to be due to precursor concentration in their furnace synthesis; the authors speculated that carbon impurities may explain the observations. However, for pure particles it seems likely that size can have an impact on properties via surface stresses and internal pressure for several materials. For example, Ehrman and co-workers (Ehrman, 1998; Ehrman, Friedlander, & Zachariah, 1998) showed that pressure can affect the solid-state diffusivity in nanoparticles and may have an impact on the sintering and coalescence of SiO2 nanoparticles. In order to explain the observed differences in crystal structure between the large and the small particle fractions, one must recognize that the cubic phase of Y2 O3 is an equilibrium phase that is favored under normal ambient conditions of temperature and pressure, while the monoclinic phase is typical of synthesis processes that involve high pressures (Hoekstra & Gingerich, 1964; Skandan et al., 1992). For example, Husson, Proust, Gillet, and Itie (1999) used Raman spectroscopy to find a high pressure phase transition in Y2 O3 from cubic to monoclinic at room temperature. The formation of monoclinic Eu: Y2 O3 nanoparticles with diameters about 20 nm has been reported for a laser ablation gas-phase condensation synthesis, and has been attributed to the increased surface stress that leads to the formation of particles in the higher density phase (Tissue & Yuan, 2003). In spray pyrolysis, we can estimate the pressure within a particle by assuming that each particle is formed from a single spherical molten droplet of Y2 O3 . The pressure p in excess of ambient within a particle of diameter D can be obtained from a hydrostatic force balance as p = 4/D, where the surface energy is taken to be 1.5 J m−2 , a value that is typical of many ceramics. The equation shows that smaller particles will experience a higher internal pressure. For a particle diameter of 20 nm (see Fig. 3A), the estimated pressure is 0.3 × 109 Pa (3 kbar). The actual pressure in our case may be higher than this value because the surface energy, , can be higher in nanoscale particles than for a flat surface (Skandan et al., 1992) although the effect is still open to some debate. For bulk material, the pressure at the cubic-to-monoclinic transition is about 1.5 × 109 Pa (Skandan et al., 1992). Our results suggest that in the case of small particles, values of the pressures may be sufficiently large to permit a monoclinic crystalline structure to form. It should also be noted that the diffusion flame that was used in this work is a very inhomogeneous environment for particle synthesis, with different temperature histories that can be experienced by particles, although the peak temperatures (about 2100 ◦ C) that are seen by all particles will be similar. These variations may also play some role in influencing the ultimate morphology of the product and may lead to the mixed phases that we see in the smallest particles. 5. Conclusions The fluorescent and crystalline properties of Eu-doped Y2 O3 nanoparticles obtained by flame spray pyrolysis showed a dependence on the particle size, a result that has not been reported before with this synthesis method. Fluorescent spectra, electron diffraction and X-ray diffraction demonstrated that particles larger than about 50 nm had a cubic Mn2 O3 -type structure that was typical of bulk material. Particles smaller than 50 nm exhibited a more complex structure with an indication, based on electron diffraction data, of the presence of the monoclinic phase with minor spectroscopic features that are consistent with this phase. There was also an indication that mixed phases may be present in the smaller size fraction. This size dependence of particle morphology may be attributable to an elevated internal pressure as particle diameter decreases. LIF of the nanopowder samples obtained from the spray flame provided a relatively straightforward diagnostic for crystal phase to the extent that a well-ordered cubic phase is clearly distinguished from a less well-ordered mixed phase in the smaller particles. Fluorescence lifetime measurements indicated the absence
D. Dosev et al. / Aerosol Science 37 (2006) 402 – 412
411
of phase separation in Eu–Y mixtures. Spectroscopy can be used as a simpler alternative to complex TEM and XRD as a method for determining crystal phases in materials such as Y2 O3 , and may have application to process control in flame synthesis processes. Higher energy pulsed laser excitation and more efficient collection optics should permit nanoparticle LIF to be used as an in situ diagnostic tool for crystal formation and phase in aerosol processing of nanophosphor materials. Modeling of the dynamics of particle morphology in multi-dimensional calculations would be aided by diagnostic tools of this sort. Acknowledgements The authors acknowledge the assistance and cooperation of Dr. K.D. Giles and Dr. D. Downey from the Department of Biological and Agricultural Engineering, UC Davis, for droplet size measurements. The assistance of Mr. J. Neil and Professor A. Navrotsky with X-ray diffraction is also appreciated. The authors wish to acknowledge the support of the National Science Foundation, NIRT Grant DBI-0102662 and the Superfund Basic Research Program with Grant 5P42ES04699 from the National Institute of Environmental Health Sciences, NIH. References Arabi-Katbi, O. I., Pratsinis, S. E., Morrison, P. W., & Megaridis, C. M. (2001). Monitoring the flame synthesis of TiO2 particles by in-situ FTIR spectroscopy and thermophoretic sampling. Combustion and Flame, 124, 560. Bazzi, R., Flores-Gonzalez, M. A., Louis, C., Lebbou, K., Dujardin, C., Brenier, A. et al. (2003). Synthesis and luminescent properties of sub-5-nm lanthanide oxides nanoparticles. Journal of Luminescence, 102–103, 445. Bhargava, R. N. (1996). Doped nanocrystalline materials—Physics and applications. Journal of Luminescence, 70, 85. Bihari, B., Eilers, H., & Tissue, B. M. (1997). Spectra and dynamics of monoclinic Eu2 O3 and Eu3+ : Y2 O3 nanocrystals. Journal of Luminescence, 75, 1. Chalek, C. L., & Gole, J. L. (1976). Chemiluminescence spectra of Sc and YO: Observation and analysis of the A 2 → X2 + band system. Journal of Chemical Physics, 65, 2845. Chen, G., Stump, N. A., Haire, R. G., & Peterson, J. R. (1992). Study of the phase behavior of Eu2 O3 under pressure via luminescence of Eu3+ . Journal of Alloys and Compounds, 181, 503. Ehrman, S. H. (1998). Effect of internal pressure on nanoparticle coalescence. Journal of Aerosol Science, 29, S31. Ehrman, S. H., Friedlander, S. K., & Zachariah, M. R. (1998). Characteristics of SiO2 /TiO2 nanocomposite particles formed in a premixed flat flame. Journal of Aerosol Science, 29, 687. Eilers, H., & Tissue, B. M. (1996). Laser spectroscopy of nanocrystalline Eu2 O3 and Eu3+ : Y2 O3 . Chemical Physics Letters, 251, 74. Feng, J., Shan, G. M., Maquieira, A., Koivunen, M. E., Guo, B., Hammock, B. D. et al. (2003). Functionalized europium oxide nanoparticles used as a fluorescent label in an immunoassay for atrazine. Analytical Chemistry, 75, 5282. Gribb, A. A., & Banfield, J. F. (1997). Particle size effects on transformation kinetics and phase stability in nanocrystalline TiO2 . American Mineralogist, 82, 717. Hao, J., Studenikin, S. A., & Cocivera, M. (2001). Blue, green and red cathodoluminescence of Y2 O3 phosphor films prepared by spray pyrolysis. Journal of Luminescence, 93, 313. Hoekstra, H. R., & Gingerich, K. A. (1964). High-pressure B-type polymorphs of some rare-Earth sesquioxides. Science, 146, 1163. Husson, E., Proust, C., Gillet, P., & Itie, J. P. (1999). Phase transitions in yttrium oxide at high pressure studied by Raman Spectroscopy. Materials Research Bulletin, 34, 2085–2092. Jones, S. L., Kumar, D., Singh, R. K., & Holloway, P. H. (1997). Luminescence of pulsed laser deposited Eu doped yttrium oxide films. Applied Physics Letters, 71, 404. Kang, Y. C., Park, S. B., Lenggoro, I. W., & Okuyama, K. (1999). Gd2 O3 :Eu phosphor particles with sphericity, submicron size and non-aggregation characteristics. Journal of Physics and Chemistry of Solids, 60, 379. Kang, Y. C., Roh, H. S., & Park, S. B. (2000). Preparation of Y2 O3 :Eu phosphor particles of filled morphology at high precursor concentrations by spray pyrolysis. Advances in Materials, 12, 451. Kang, Y. C., Roh, H. S., Park, S. B., & Park, H. D. (2002). High luminescence Y2 O3 :Eu phosphor particles prepared by modified spray pyrolysis. Journal of Materials Science Letters, 21, 1027. Katagiri, S., Ishizawa, N., & Marumo, F. (1993). A new high temperature modification of face-centered cubic Y2 O3 . Powder Diffraction, 8, 60. Kim, C. S., Nakaso, K., Xia, B., & Okuyama, K. (2005). A new observation on the phase transformation of TiO2 nanoparticles produced by a CVD method. Aerosol Science and Technology, 39, 104. Rosner, D. E., McGraw, R., & Tandon, P. (2003). Multivariate population balances via moment and Monte Carlo simulation methods: An important sol reaction engineering bivariate example and “mixed” moments for the estimation of deposition, scavenging, and optical properties for populations of nonspherical suspended particles. Industrial & Engineering Chemistry Research, 42, 2699. Rosner, D. E., & Pyykonen, J. J. (2002). Bivariate moment simulation of coagulating and sintering nanoparticles in flames. AIChE Journal, 48, 476. Shikao, S., & Jiye, W. (2001). Combustion synthesis of Eu3+ activated Y3 Al5 O12 phosphor nanoparticles. Journal of Alloys and Compounds, 327, 82.
412
D. Dosev et al. / Aerosol Science 37 (2006) 402 – 412
Skandan, G., Foster, C. M., Frase, H., Ali, M. N., Parker, J. C., & Hahn, H. (1992). Phase characterization and stabilization due to grain size effects of nanostructured Y2 O3 . Nanostructured Materials, 1, 313. Tissue, B. M. (1998). Synthesis and luminescence of lanthanide ions in nanoscale insulating hosts. Chemical Materials, 10, 2837. Tissue, B. M., & Yuan, H. B. (2003). Structure, particle size, and annealing of gas phase-condensed Eu3+ : Y2 O3 nanophosphors. Journal of Solid State Chemistry, 171, 12. Vander Wal, R. L., Berger, G. M., Ticich, T. M., & Patel, P. D. (2002). Application of laser-induced incandescence to the detection of carbon nanotubes and carbon nanofibers. Applied Optics, 41, 5678. Vander Wal, R. L., Ticich, T. M., West, J. R., & Householder, P. A. (1999). Trace metal detection by laser-induced breakdown spectroscopy. Applied Spectroscopy, 53, 1226. Vetrone, F., Boyer, J. C., Capobianco, J. A., Speghini, A., & Bettinelli, M. (2004). A spectroscopic investigation of trivalent lanthanide doped Y2 O3 nanocrystals. Nanotechonology, 15, 75. Wakefield, G., Holland, E., Dobson, P. J., & Hutchison, J. L. (2001). Luminescence properties of nanocrystalline Y2 O3 :Eu. Advances in Materials, 13, 1557. Wijchers, T., Dijkerman, H. A., Zeegers, P. J. T., & Alkemade, C. T. J. (1980). Fluorescence spectra of laser-excited YO molecules in a H2 -O2 -Ar flame. Spectrochimica Acta, 35B, 271. Williams, D. K., Yuan, H., & Tissue, B. M. (1999). Size dependence of the luminescence spectra and dynamics of Eu3+ : Y2 O3 nanocrystals. Journal of Luminescence, 83–84, 297. Xing,Y. C., Koylu, U. O., & Rosner, D. E. (1996). Synthesis and restructuring of inorganic nano-particles in counterflow diffusion flames. Combustion and Flame, 107, 85. Xing, Y. C., Koylu, U. O., & Rosner, D. E. (1999). In situ light-scattering measurements of morphologically evolving flame-synthesized oxide nanoaggregates. Applied Optics, 38, 2686. Xing, Y., Rosner, D. E., Koylu, U. O., & Tandon, P. (1997). Morphological evolution of nanoparticles in diffusion flames: Measurements and modeling. AIChE Journal, 43, 2641. Yang, H.-S., Hong, K. S., Feofilov, S. P., Tissue, B. M., Meltzer, R. S., & Dennis, W. M. (1999). Electron-phonon interaction in rare earth doped nanocrystals. Journal of Luminescence, 83–84, 139. Zhang, H. Z., & Banfield, J. F. (1998). Thermodynamic analysis of phase stability of nanocrystalline titania. Journal of Materials Chemistry, 8, 2073.