Phase Structure and Luminescence Properties of ... - ACS Publications

Dec 24, 2005 - Phase Structure and Luminescence Properties of Eu3+-Doped TiO2 Nanocrystals Synthesized by Ar/O2 Radio Frequency Thermal Plasma ...
5 downloads 0 Views 359KB Size
Download PDF
J. Phys. Chem. B 2006, 110, 1121-1127

1121

Phase Structure and Luminescence Properties of Eu3+-Doped TiO2 Nanocrystals Synthesized by Ar/O2 Radio Frequency Thermal Plasma Oxidation of Liquid Precursor Mists Ji-Guang Li,* Xiaohui Wang, Kenji Watanabe, and Takamasa Ishigaki National Institute for Materials Science, AdVanced Materials Laboratory, Namiki 1-1, Tsukuba, Ibaraki 305-0044, Japan ReceiVed: June 21, 2005; In Final Form: NoVember 16, 2005

Eu3+-doped TiO2 luminescent nanocrystals have been synthesized in this work via Ar/O2 thermal plasma oxidizing mists of liquid precursors containing titanium tetra-n-butoxide and europium(III) nitrate, with varied O2 input in the plasma sheath (10-90 L/min) and Eu3+ addition in the precursor solution (Eu/(Ti + Eu) ) 0-5 atom %). The resultant nanopowders are mixtures of the anatase (30-36 nm) and rutile (64-83 nm) polymorphs in the studied range, but the rutile fraction increases steadily at a higher Eu3+ addition, as revealed by X-ray diffraction (XRD) and Raman spectroscopy, because of the creation of oxygen vacancies in the TiO2 gas clusters by substitutional Eu3+ doping. The amount of Eu3+ that can be doped into a TiO2 lattice was limited up to 0.5 atom %, above which Eu2Ti2O7 pyrochlore was formed in the final products. High resolution transmission electron microscopy (HRTEM) observation indicates that the particles are dense and have sizes ranging from several nanometers up to 180 nm. Efficient nonradiative energy transfer from the TiO2 host to Eu3+ ions, which was seldom reported in the wet-chemically derived nanoparticles or thin films of the current system, was confirmed by combined studies of excitation, UV-vis (ultraviolet-visible), and PL (photoluminescence) spectroscopy. As a consequence of this, bright red emissions were observed from the plasma-generated nanopowders either by exciting the TiO2 host with UV light shorter than 405 nm or by directly exciting Eu3+ at a wavelength beyond the absorption edge (405 nm) of TiO2.

1. Introduction Rare-earth-doped materials find usage in a wide variety of applications, including phosphors, X-ray imaging, scintillators, lasers, display monitors, and amplifiers for fiber-optic communications.1 Up to date, a variety of materials have been investigated as the host lattice for rare-earth (RE) cations, including oxides (single cation or multication),2,3 sulfides,4 selenides,5,6 and metallorganic complexes.7 Rare-earth complexes with organic ligands, especially β-diketones, show enhanced emissions due to the nonradiative energy transfer from the triplet state of the ligands to the crystal field states of the central RE ions7 but usually possess poor thermal stability and weak mechanical properties, which prevent them from some practical applications, such as tunable solid-state lasers and phosphor devices. Current research tends toward doping semiconducting compounds (such as ZnO,8,9 ZnS,4,10 CdS,4 CdSe,5,6 TiO2,11-14 and BaTiO314) with RE ions, with the expectation that the optical properties might be controlled or tuned by utilizing the unique band-gap structure of the semiconductor host. Up to date, a handful of solution-based procedures have been used for the synthesis of luminescent nanoparticles, mainly including sol-gel, precipitation, hydrothermal treatment, and inverse micelle techniques.4,11-14 Effective doping, however, is still rather difficult to achieve in many cases, as the dopant and host ions have different ionic sizes and especially different chemical properties, which frequently lead to sequential precipitation. As a result, the RE ions might be simply absorbed on the surfaces of the semiconductor particles, and hence the * Corresponding author. Telephone: +81-29-860-4394. Fax: +81-29860-4701. E-mail: [email protected].

observed emissions are not due to the energy transferred from the host lattice but due to the direct absorptions by the RE activators themselves, as was recently reported by Bol et al.4 In addition, wet-chemical processing frequently generates amorphous or low-crystallinity particles with surfaces covered with hydroxyls, which are known to increase the probability of nonradiative relaxations of RE ions.15 Postannealing, however, may lead to hard aggregation of the nanoparticles. Eu3+ is a widely used activator for red phosphors, due to its 5D0 f 7F2 electronic transition. TiO2, on the other hand, is a promising host material for RE ions, due to its high transparency in the visible wavelength region and its good thermal, chemical, and mechanical properties. Very recently, Frindell et al.16 have demonstrated that mesoporous TiO2 with nanocrystalline walls is an excellent host for Eu3+, while Xin et al.17 successfully utilized lamellar aggregates of titania nanosheets to accommodate Eu3+ ions. In both the cases, characteristic emissions from Eu3+ were observed arising from energy transfer from the host to the activator Eu3+ ions. Such a phenomenon, however, was not reported and/or not observed in Eu3+-doped TiO2 nanoparticles and thin films prepared via precipitation and solgel techniques.11,13 Eu3+-doped TiO2 luminescent nanocrystals, with efficient nonradiative energy transfer from the TiO2 host to the activator Eu3+ ions, have been synthesized in this work via radio frequency (rf) Ar/O2 thermal plasma oxidizing mists of liquid precursors containing titanium tetra-n-butoxide and europium nitrate. Radio frequency thermal plasma proves to be a powerful tool for synthesizing well-dispersed nanoparticles of good crystallinity within a very short period of time. Its high

10.1021/jp053329l CCC: $33.50 © 2006 American Chemical Society Published on Web 12/24/2005

1122 J. Phys. Chem. B, Vol. 110, No. 3, 2006

Li et al. TABLE 1: Experimental Conditions for TiO2:Eu3+ Nanopowder Synthesis via Ar/O2 Thermal Plasma Oxidation of Liquid Precursor Mists

Figure 1. A schematic illustration showing the experimental setup of this work.

processing temperature (up to 1.5 × 104 K), fast quenching rate (∼105-106 K/s) at the plasma tail, and its thermal nonequilibrium effects provide a unique reaction field for solid solution formation.18 Meanwhile, the use of mist droplets ensures rapid chemical reactions between the starting material and the plasma field and hence high-degree supersaturation of evaporated species, favoring nanoparticle formation upon gas-phase condensation.19 The plasma generated TiO2:Eu3+ nanoparticles were found to exhibit good red emissions either by exciting the TiO2 host or by directly exciting the Eu3+ ions at a wavelength beyond the absorption edge of TiO2, indicating the occurrence of efficient nonradiative energy transfer. In the following sections, we report the synthesis, phase structure, and photoluminescence properties of the plasma generated Eu3+-doped TiO2 nanocrystals. 2. Experimental Section 2.1. Nanoparticle Synthesis. The experimental apparatus for powder synthesis mainly consists of a water-cooled induction plasma torch (model PL-50, TEKNA Plasma System Inc., Sherbrooke, QC, Canada), a 2-MHz radio frequency power supply system (Nihon Koshuha Co. Ltd., Yokohama, Japan), a water-cooled stainless steel reactor, and a stainless steel filter connecting the reactor and a vacuum pump. Details of the experimental setup are shown in Figure 1. All the chemicals used in this work are reagent grade supplied by Wako Pure Chemical Industries Ltd., Tokyo, Japan, and were used as received. The liquid precursor for powder synthesis was made according to the following procedure. A total of 0.1 mol (34 g) of titanium tetra-n-butoxide (Ti(OC4H9)4, TTBO) was added to 0.4 mol (42 g) of diethanolamine (HN(OC2H5)2, DEA, a chelate for Ti4+) under magnetic stirring to obtain a clear solution (solution 1). A weak exothermic reaction was observed upon mixing, which is attributable to the chelating reaction between the two reagents. Such a chelating reaction stabilizes the vivid reactive TTBO against hydrolysis, even in the presence of water. Separately, europium(III) nitrate (Eu/(Eu + Ti) ) 0, 0.05, 0.1, 0.2, 0.3, 0.5, 0.75, 1.0, 2.0, 3.0, and 5.0 atom %) and citric acid (C6H8O7, CA, a chelate for Eu3+, CA-Eu3+) 1:1 in molar ratio) were dissolved in 20 mL of distilled water (solution 2). Acids catalyze the hydrolysis of TTBO even in the presence of DEA, and to avoid this, the pH of solution 2 was adjusted to 9.0 with 3 mL of 25% ammonia solution. Mixing solutions 1 and 2 yields a stable clear solution to be used as the liquid precursor for powder synthesis.

parameter

value

central gas and flow rate sheath gas and flow rate atomization gas and flow rate precursor feeding rate induction power for plasma generation chamber pressure

Ar, 30 L/min Ar + O2, 90 L/min in total Ar, 5 L/min 4.5 g/min 25 kW 66.7 kPa

For powder synthesis, the liquid precursor was delivered by a peristaltic pump into the center of the plasma plume through an atomization probe. Flow rate of the liquid precursor was controlled at 4.5 g/min, which corresponds to 4.5 × 10-3 mol/ min in terms of TTBO. The atomization probe (model SA792260-100, TEKNA Plasma System Inc., Sherbrooke, QC, Canada) was water-cooled to resist the extreme temperature of the plasma, and the precursor was atomized into mists at the tip of it by Ar carrier gas flowing through the probe at 5 L/min. The Ar/O2 thermal plasma was generated by mixing O2 gas to the Ar sheath. Total flow rate of the sheath gas (Ar + O2) was set at 90 L/min but in which the O2 input was varied in the range 10-90 L/min to investigate its effects on powder properties. Other details of the processing parameters were summarized in Table 1. 2.2. Characterization Techniques. Phase identification was performed via X-ray diffraction (XRD) on a Philips PW1800 X-ray diffractometer (Philips Research Laboratories, Eindhoven, The Netherlands) operating at 40 kV/50 mA using nickel-filtered Cu KR radiation with a scanning speed of 0.15°/2θ per minute. Rutile and anatase contents of the powders were determined according to the method of Zhang and Banfield.20 Crystallite sizes of the anatase and rutile phases were calculated from the Scherrer equation via broadening analysis performed on the anatase (101) and the rutile (110) diffractions, respectively. Raman spectroscopy was made using Ar+ laser excitation (514.5 nm) with a source power of 50 mW and a resolution of 1 cm-1 (model NR-1800, Jasco Co., Tokyo, Japan). Particle morphology was observed using a JEM-2100F transmission electron microscope (TEM, JEOL, Tokyo, Japan) with an acceleration voltage of 200 kV. Ultraviolet-visible (UV-vis) diffuse reflectance spectroscopy was conducted on a Jasco V-570 spectrophotometer (Jasco Co., Tokyo, Japan). Photoluminescence (PL) spectroscopy under 325 nm He-Cd laser excitation was performed on a Renishaw spectrophotometer (Renishaw plc, Gloucestershire, U.K.). Excitation and PL spectra were also collected on a Hitachi F-4500 fluorescence spectrophotometer (Hitachi, Tokyo, Japan). 3. Results and Discussion 3.1. Phase Structure of the Plasma Generated Nanoparticles. The TiO2 nanoparticles formed via rapid oxidation of the atomized liquid precursor by Ar/O2 thermal plasma mainly deposit on the filter and the inner walls of the reactor. Elemental analysis on europium by standard ICP (inductively coupled plasma) procedures confirmed that the prepared Eu/(Ti + Eu) molar ratios in the precursor solutions have been kept to the final products. Figure 2 shows XRD patterns of the powders synthesized with 40 L/min of O2 flow in the plasma sheath, for some typical dopant concentrations. In all the cases the powders are mixtures of the anatase and rutile polymorphs. Eu2Ti2O7 pyrochlore (JCPDS: no. 23-1072) appeared in the powder doped with 0.5 atom % of Eu3+, and its diffraction intensity successively increased at a higher Eu3+ addition. This indicates that

Phase Structures and Luminescence Properties

Figure 2. XRD patterns of the nanopowders synthesized under 40 L/min of O2 input in the plasma sheath, with varied Eu3+ additions in the precursor solution. A, R, and P denote anatase, rutile, and Eu2Ti2O7 pyrochlore, respectively.

Figure 3. Phase composition and average crystallite size of the nanopowders synthesized with 40 L/min of O2 input in the plasma sheath as a function of the Eu3+ content in the precursor solution.

the solubility of Eu3+ in the TiO2 lattice is rather limited, which is conceivable from the quite big size discrepancy between Ti4+ (0.0605 nm for 6-fold coordination) and Eu3+ (0.0947 nm for 6-fold coordination) ions.21 The addition of Eu3+ to the precursor solution, however, has appreciable effects on phase constituents of the resultant powders, which can be perceived from the relative intensities of the anatase (101) and rutile (110) peaks. Phase composition and crystallite size of the TiO2 powders synthesized under 40 L/min of O2 input are shown in Figure 3 as a function of the Eu3+ content in the liquid precursor. Without Eu3+ doping, the powder contains 22 wt % of rutile, which increases steadily to 52 wt % at 5 atom % of Eu3+ addition. The average size of anatase crystallites stays almost constant in the range 30-36 nm for all the powders, while the average rutile size, varying between 64 and 83 nm in the studied range, is always much bigger than the anatase size. Changing the O2 input in the plasma sheath does not significantly affect the phase constituent and crystallite size, as is shown in Figure 4 for the powders doped with 0.5 atom % of Eu3+. Figure 5a exhibits overall morphology of the 0.5 atom % Eu3+-doped TiO2 nanopowder synthesized with 40 L/min of O2 flow. The particles are dense and largely dispersed, but a wide size distribution was observed ranging from several

J. Phys. Chem. B, Vol. 110, No. 3, 2006 1123

Figure 4. Phase composition and average crystallite size of the nanopowders as a function of the O2 flow rate in the plasma sheath. The Eu3+ content in the precursor was kept at Eu/(Eu + Ti) ) 0.5 atom %.

nanometers up to 180 nm. The nanosized particles tend to be faceted while the submicrometer-sized particles are almost perfect spheres. The wide size distribution and the different particle morphologies might arise from the different trajectories, or in other words the different thermal histories, of the mist droplets in the hot zone of the thermal plasma. Individual anatase (Figure 5b) and rutile (Figure 5c) nanocrystallites were identified by HRTEM observations through d-spacing calculations, which were further confirmed by selected area diffraction (SAD) analysis (the insets in Figure 5c and Figure 5d). The main process of nanocrystal formation in the present work might be explained as follows. The liquid precursor mists are rapidly oxidized by the Ar/O2 thermal plasma, and the oxidation products undergo simultaneous evaporation due to the high plasma temperature (∼104 K) and the additional heat generated by the combustive oxidation reaction. TiO2 gas clusters22,23 are then formed just before liquid phase (melt) formation (melting point: 1870 °C)24 upon the evaporated species being cooled at the plasma tail. Further cooling of the melts induces nucleation and growth of TiO2 nanocrystals. The size distribution of particles shown in Figure 5 might mainly be related to that of the melt droplets. It is well-known that metastable anatase transforms to thermodynamically stable rutile by routine annealing at relatively low temperatures of ∼4001000 °C, though the exact transition temperature is affected by crystallite size, impurity (dopant) type and concentration, and atmosphere.25-27 The formation of metastable anatase, despite the high plasma temperature, is due to the fast quenching effect at the tail of the plasma plume. Theoretical analysis through calculation of the critical free energy for nucleation implies that metastable anatase tends to nucleate preferentially from deeply undercooled TiO2 melts,28 which explains the formation of anatase even as a major phase. On the other hand, a dopant affects the phase structure of TiO2. Vemury and Pratsinis29 studied the effects of a dopant on the phase structure of TiO2 powders obtained via flame pyrolysis and found that rutile formation was enhanced either by introducing dopant oxides with the same crystal structure as rutile or by creating oxygen vacancies in the gas clusters by doping with subvalent cations such as Al3+. Their observations conformed to the findings of Shannon and Pask27 obtained through phase transformation studies performed under thermal equilibrium conditions. A similar explanation holds for the effects of Eu3+ on the phase composition observed in Figure 3. Subvalent Eu3+ ions create oxygen vacancies (for charge compensation) in the TiO2 gas

1124 J. Phys. Chem. B, Vol. 110, No. 3, 2006

Li et al.

Figure 5. TEM micrographs showing (a) overall morphology of the 0.5 atom % Eu3+-doped TiO2 particles synthesized with 40 L/min of O2 input in the plasma sheath, (b) an individual anatase nanocrystallite, (c) an individual rutile nanocrystallite, and (d) the enlarged lattice fringe shown in part b. The insets in parts c and d are the SAD patterns of the nanocrystallites with zone axis of and , respectively.

clusters at the high temperature of thermal plasma and therefore enhance rutile formation. Unlike Al3+, which forms substitutional solid solutions with TiO2 up to a high level of about Al/ Ti ) 0.2 molar ratio,29 most of the Eu3+ was expelled out during the cooling processes because of its very limited solubility in TiO2, causing the formation of Eu2Ti2O7 pyrochlore. The different crystallite sizes between anatase and rutile polymorphs might be related to the different cooling rates of the TiO2 melt droplets arising from the temperature profile of the thermal plasma and the complex trajectories of the precursor mists. As mentioned earlier, anatase nucleates with priority from deeply undercooled TiO2 melts while rutile nucleates from less undercooled ones.28 More crystallite growth might occur in the latter case, yielding larger rutile crystallites. Previous work on the synthesis of pure TiO2 nanocrystallites via in-flight oxidation of TiN or TiC particles with Ar/O2 thermal plasma demonstrated unambiguously that the O2 flow rate in the plasma sheath significantly affects phase constituents of the resultant powders through altering the cooling rate at the plasma tail and especially the stoichiometry of the TiO2 gas clusters.23,30 Varying the O2 input in the present study, however, gives no substantial effects on the final products (Figure 4). This is due to the fact that the O2 input in the present study has been overstoichiometric for oxidation of the liquid precursor. Assuming CO (CO tends to be more stable than CO2 at the high plasma temperature),31 NO2, H2O, and TiO2 as the oxidation products, a complete oxidation requires 0.16 mol/min of O2, as calculated from the chemical composition and feeding

rate (4.5 g/min or 4.5 × 10-3 mol/min for TTBO) of the liquid precursor, while the lowest O2 input in this work (10 L/min) is 0.294 mol/min at a reactor pressure of 66.7 kPa. In such a case, the rutile content is overwhelmingly influenced by the oxygen deficiency in the TiO2 gas clusters created by substitutional Eu3+ doping for charge compensation. A similar phenomenon was recently observed in Fe3+-doped TiO2 nanocrystals synthesized via Ar/O2 thermal plasma oxidation of liquid precursor containing TTBO and ferrocene.32 Figure 6 exhibits Raman spectra of the powders containing various amounts of Eu3+, with that of pure Eu2Ti2O7 included for comparison. The pyrochlore phase shows diffuse and weak Raman scatterings in the selected wavenumber region, which is in contrast to the TiO2:Eu3+ nanoparticles. The sharp Raman bands of TiO2:Eu3+ powders confirm the good crystallinity revealed by XRD. The scatterings at 146 (Eg), 200 (Eg), 401 (B1g), 519 (B1g), and 641 cm-1 (Eg) were identified for the anatase phase, while those at 449 (Eg) and 614 cm-1 (A1g) were found for the rutile phase.33,34 The B1g mode of rutile, frequently observed at 143 cm-1, apparently overlaps the strongest scattering from anatase (146 cm-1) and was not clearly seen. Both the Eg and A1g modes of rutile increase in intensity at a higher Eu3+ addition, indicating increased fractions of the rutile phase in the powders. Such an observation is consistent with the results of XRD analysis (Figure 3). Raman bands of the pyrochlore phase were not unambiguously identified in the powders doped up to 5 atom % of Eu3+, attributable to the small content of the pyrochlore phase and the weakness of the Raman

Phase Structures and Luminescence Properties

Figure 6. Raman spectra of the powders synthesized with various Eu3+ additions in the precursors. The O2 input in the plasma sheath is 40 L/min. A and R represent anatase and rutile, respectively.

Figure 7. UV-vis diffuse reflectance spectra of the TiO2:Eu3+ nanopowders compared with those of TiO2 and Eu2Ti2O7 pyrochlore.

bands. Increased oxygen vacancies in the crystal structure of TiO2 cause increased wavenumber of the anatase Eg mode (146 cm-1) and decreased wavenumber of the rutile Eg mode (449 cm-1),30,35,36 which were not observed in this work up to 5 atom % of Eu3+ addition. This may imply that the discrepancies in the oxygen vacancy level have not been significant enough to cause Raman band shifting. This is understandable from the rather limited substitutional Eu3+ doping in the TiO2 lattice. 3.2. Luminescence Properties of the Plasma Generated Nanopowders. Figure 7 shows UV-vis diffuse reflectance spectra of the TiO2:Eu3+ nanopowders compared with pure TiO2 and Eu2Ti2O7 synthesized with thermal plasma under 40 L/min of O2 input in the plasma sheath. Notice that data offsetting and profile enlargement (for Eu2Ti2O7) have been made here for overall clearness of the figure. The pure TiO2 sample exhibits an onset of absorption at 405 nm, corresponding to a band gap of 3.06 eV. This band gap value is almost identical to that (3.04 eV) of Degussa P25,37 which has a phase composition (∼20 wt % of rutile and ∼80 wt % of anatase) close to the pure TiO2

J. Phys. Chem. B, Vol. 110, No. 3, 2006 1125

Figure 8. Excitation and emission spectra of the nanopowders: (a) excitation spectrum obtained by tracking the 617 nm emission of 0.5 atom % Eu3+ doped TiO2; (b-d) emission spectra with sample composition and excitation wavelength indicated. Direct comparison of the emission intensities can be made among parts b-d.

powder obtained in this work (Figure 3). In addition to the absorption in the UV region, the pyrochlore powder shows additional absorptions at 395, 416, 467, and 538 nm, which are assignable to the intraconfigurational 4f f 4f transitions of Eu3+ ions.3,17 These absorption peaks should also appear in the UVvis spectra of TiO2:Eu3+ nanopowders but are not clearly observed up to 5.0 atom % of Eu3+ addition due to the weakness of the peaks. Besides, the 395 nm peak occurs at a wavelength where absorption by the TiO2 host lattice is still relatively strong, and hence the absorption by Eu3+ ions at 395 nm might be quenched. Figure 8a demonstrates the excitation spectrum of the 0.5 atom % Eu3+-doped TiO2 nanopowder, measured by monitoring the 617 nm emission from the 5D0 f 7F2 transition of Eu3+ ions. The three peaks at 416, 467, and 538 nm, which show features similar to those observed in the UV-vis spectrum of Eu2Ti2O7, can be assigned to the 7F0,1 f 5D3, 7F0,1 f 5D2, and 7F0,1 f 5D1 transitions of the Eu3+ ions, respectively.3 The broad peak at 360 nm, whose position coincides well with the absorption band of pure TiO2 (Figure 7), indicates that Eu3+ ions can be effectively excited through the TiO2 host lattice (notice that the right-hand tail of the 360 nm peak overlaps the left-hand tail of the 416 nm peak). Parts b-d of Figure 8 show emission spectra of pure TiO2 (Figure 8b) and 0.5 atom % Eu3+ doped TiO2 (Figure 8c and Figure 8d). When excited at 467 nm, a wavelength longer than the absorption edge of TiO2 (405 nm), the Eu3+-doped nanopowder exhibits characteristic emissions in the range 550-750 nm (Figure 8c), which are associated with the electronic transitions from the excited 5D0 level to 7Fj (j ) 1, 2, 3, 4) levels of Eu3+ activators as commonly observed.3,11,17 Clearly, these emissions come from the absorption by the Eu3+ ions themselves, as confirmed by the excitation (Figure 8a) and the UV-vis spectra (Figure 7). Excited at a wavelength (360 nm) shorter than the absorption edge of TiO2, the host lattice exhibits a luminescence behavior (Figure 8b) almost identical to that of the Ti0.91O2 nanosheets reported by Xin et al.17 On the other hand, the 0.5 atom % Eu3+-doped

1126 J. Phys. Chem. B, Vol. 110, No. 3, 2006

Li et al.

Figure 10. Relative intensity of the 617 nm emission as a function of the Eu3+ addition in the precursor solution. All the powders were synthesized under 40 L/min of O2 input in the plasma sheath.

Figure 9. Emission spectra of the TiO2:Eu3+ nanopowders under 325 nm He-Cd laser excitation compared with those of Eu2O3 and Eu2Ti2O7. For the four samples, the original PL signals have been divided by a common factor for clearer demonstration and easier intensity comparison. Satellite emissions (Stark splits) of each 5D0 f 7 Fj (j ) 1, 2, 3, 4) transition is indicated by an arrowed bracket in the figure.

sample shows luminescence from both the TiO2 host and the Eu3+ ions under 360 nm excitation (Figure 8d). It was noticed that the Eu3+ emission is clearly stronger than that from the host lattice and also stronger than an emission under 467 nm excitation, suggesting an efficient nonradiative energy transfer from the TiO2 host to Eu3+ ions. Luminescence properties of the TiO2:Eu3+ nanopowders have been studied thoroughly via 325 nm He-Cd laser excitation. Figure 9 displays typical emission spectra of TiO2: Eu3+ nanoparticles as well as those of pure Eu2O3 and Eu2Ti2O7. It is the 5D0 f 7F2 transition that gives a sharp red color when a 325 nm He-Cd laser beam impinges upon sample surfaces. The Eu3+-doped samples exhibit emissions clearly different from those of Eu2O3 and Eu2Ti2O7, in terms of peak positions and peak shapes, implying different local environments of Eu3+. It is known that both anatase (space group D19 4h) and 38 while Eu O (T7)39 rutile (D14 ) assume tetragonal symmetry 2 3 4h h and Eu2Ti2O7 (Fd3hm)40 adopt cubic symmetry at room temperature. In addition, the substitutional Eu3+ dopants in TiO2 are less well-located on the lattice sites as compared to the Eu3+ ions in the latter two compounds, due to lattice distortion and the existence of point defects such as oxygen vacancies. Figure 9, therefore, provides direct evidence that the emissions from TiO2:Eu3+ arise from the Eu3+ ions doped into the TiO2 lattice and not from separated Eu2O3 or Eu2Ti2O7 particles. Besides, photoluminescence arising from Eu2Ti2O7 pyrochlore was clearly observed in the nanopowder containing 5 atom % of Eu3+. The 5D0 f 7F1 lines (three Stark splits) are known to originate from magnetic dipole transition while the 5D0 f 7F2 lines (five Stark splits) originate from electric dipole transition. According to the Judd-Ofelt theory,41,42 the magnetic dipole transition is permitted while the electric dipole transition is forbidden. The latter is allowed only on condition that the Eu3+ ions occupy a site without an inversion center and are sensitive to local symmetry. For the TiO2:Eu3+ nanopowders synthesized

in this work, the electric dipole transition is much stronger than the magnetic dipole transition, suggesting that Eu3+ takes a low symmetry site without an inversion center in the TiO2 host lattice. Anatase and rutile adopt tetragonal symmetry with space 14 groups of I41/amd (D19 4h) and P42/mnm (D4h), respectively, and 4+ the site symmetries for Ti are D2d in anatase and D2h in rutile.38 The substitution of Eu3+ for Ti4+ creates oxygen vacancies and lattice distortions in the TiO2 host, making the site symmetry of Eu3+ deviate from the exact D2d and D2h symmetries. So the 5D f 7F transition of Eu3+ dominates the emission spectra. It 0 2 is well-known that the relative intensities of 5D0 f 7F1 and 5D0 f 7F2 transitions are hypersensitive to the local symmetry of Eu3+ ions. The intensity ratio of 5D0 f 7F2 transition (monitored at 617 nm) to 5D0 f 7F1 transition (monitored at 599 nm) keeps almost constant at 9.7 in this work for the samples doped up to 0.5 atom % of Eu3+, implying that the overall Eu3+ local environments do not differ from sample to sample. Figure 10 displays relative luminescence intensity of the nanopowders as a function of the Eu3+ concentration in the precursor solution. The PL intensity (monitored as the 617 nm emission from the 5D f 7F transition) increases sharply and almost linearly with 0 2 increasing the Eu3+ content up to 0.5 atom % and then levels off. Such a phenomenon is primary due to the formation of Eu2Ti2O7 pyrochlore arising from the limited solubility of Eu3+ in the TiO2 lattice (Figure 2). Luminescence properties of the nanopowders processed with different O2 inputs were also investigated, but no significant difference was observed. The energy transfer from the TiO2 host to Eu3+ ions is a defect-mediated process, a model of which has been proposed by Frindell et al.16 for the Eu3+-doped mesoporous TiO2 films and later adopted by Xin et al.17 for the Eu3+-accommodating Ti0.91O2 nanosheets. According to this model, UV light is absorbed in the band gap of TiO2 and the energy is relaxed to the defect states. As the defect energy levels of the TiO2 host are higher than that of the emitting state (5D0) of Eu3+ ions, energy transfer to the crystal field states of Eu3+ ions then takes place, resulting in efficient photoluminescence from the plasma generated nanoparticles. The defects commonly encountered in TiO2 might be oxygen vacancies, interstitial or substitutional Ti3+ ions, interstitial Ti4+ ions, and so forth.27 In this work, the energy mediating defects may dominantly be the oxygen vacancies generated by the substitutional doping of subvalent Eu3+ ions into TiO2 lattice.

Phase Structures and Luminescence Properties 4. Conclusions Eu3+-doped

TiO2 nanopowders have been synthesized via radio frequency Ar/O2 thermal plasma oxidizing mists of liquid precursors containing titanium tetra-n-butoxide and europium(III) nitrate, with varied O2 inputs (10-90 L/min) in the plasma sheath and Eu3+ concentrations (0-5 atom %) in the precursor solution. The resultant powders are mixtures of the anatase (30-36 nm) and rutile (64-83 nm) polymorphs in the studied range, but the rutile fraction increases steadily at a higher Eu3+ addition, owing to the creation of oxygen vacancies by replacing the Ti4+ sites with subvalent Eu3+ ions in the TiO2 gas clusters. Eu3+ was hardly dissolved in the TiO2 lattice (>0.5 atom %) but mainly formed a Eu2Ti2O7 pyrochlore. Varying the O2 input yields little influence on phase composition and crystallite size. Bright red emissions were observed from the nanopowders either by exciting the TiO2 host with UV light (λ < 405 nm) or by directly exciting the Eu3+ ions at a longer wavelength where there is no absorption by TiO2, confirming the occurrence of efficient nonradiative energy transfer from the host lattice to Eu3+ ions. The nanoparticles synthesized in this work might find applications in optoelectronic devices. References and Notes (1) Blasse, G.; Grabmaier, B. C. Luminescent Materials; SpringerVerlag: Berlin, Germany, 1994. (2) Williams, D. K.; Bihari, B.; Tissue, B. M.; McHale, J. M. J. Phys. Chem. B 1998, 102, 916. (3) Riwotzki, K.; Haase, M. J. Phys. Chem. B 1998, 102, 10129. (4) Bol, A. A.; van Beek, R.; Meijerink, A. Chem. Mater. 2002, 14, 1121. (5) Firth, A. V.; Cole-Hamilton, D. J.; Allen, J. W. Appl. Phys. Lett. 1999, 75, 3120. (6) Zhang, X.; Liang, L.; Zhang, J.; Su, Q. Mater. Lett. 2005, 59, 749. (7) Kido, J.; Okamoto, Y. Chem. ReV. 2002, 102, 2357. (8) Mais, N.; Reithmaier, J. P.; Forchel, A.; Kohls, M.; Spanhel, L.; Muller, G. Appl. Phys. Lett. 1999, 75, 2005. (9) Zhou, Z.; Komori, T.; Yoshino, M.; Morinaga, M. Appl. Phys. Lett. 2005, 86, 041107. (10) Xu, S. J.; Chua, S. J.; Liu, B.; Gan, L. M.; Chew, C. H.; Xu, G. Q. Appl. Phys. Lett. 1998, 73, 478. (11) Conde-Gallardo, A.; Garcia-Rocha, M.; Hernandez-Calderon, I.; Palomino-Merino, R. Appl. Phys. Lett. 2001, 78, 3436.

J. Phys. Chem. B, Vol. 110, No. 3, 2006 1127 (12) Jeon, S.; Braun, P. V. Chem. Mater. 2003, 15, 1256. (13) Ovenstone, J.; Titler, P. J.; Withnall, R.; Silver, J. J. Phys. Chem. B 2001, 105, 7170. (14) Patra, A.; Friend, C. S.; Kapoor, R.; Prasad, P. N. Chem. Mater. 2003, 15, 3650. (15) Bahtat, A.; DeLucas, M. C. M.; Jacquier, B.; Varrel, B.; Bouazaoui, M.; Mugnier, J. Opt. Mater. 1997, 7, 173. (16) Frindell, K. L.; Bartl, M. H.; Robinson, M. R.; Bazan, G. C.; Popitsch, A.; Stucky, G. D. J. Solid State Chem. 2003, 172, 81. (17) Xin, H.; Ma, R.; Wang, L.; Ebina, Y.; Takada, K.; Sasaki, T. Appl. Phys. Lett. 2004, 85, 4187. (18) Boulos, M. I.; Fauchais, P.; Pfender, E. Thermal Plasmas: Fundermentals and Applications; Plenum Press: New York, 1994; Vol. 1. (19) Ishigaki, T.; Oh, S.-M.; Li, J.-G.; Park, D.-W. Sci. Technol. AdV. Mater. 2005, 6, 111. (20) Zhang, H. Z.; Banfield, J. F. J. Phys. Chem. B 2000, 104, 3481. (21) Shannon, R. D. Acta Crystallogr. 1976, A32, 751. (22) Suyama, Y.; Ito, K.; Kato. A. J. Inorg. Nucl. Chem. 1975, 37, 1883. (23) Oh, S.-M.; Li, J.-G.; Ishigaki, T. J. Mater. Res. 2005, 20, 529. (24) Cvetkovic, K.; Petric, A. Am. Ceram. Soc. Bull. 2000, 79, 65. (25) Kumar, K. N. P.; Keizer, K.; Burggraaf, A. J. J. Mater. Chem. 1993, 11, 1141. (26) Zhang, H. Z.; Banfield, J. F. J. Mater. Res. 2000, 15, 437. (27) Shannon, R. D.; Pask, J. A. J. Am. Ceram. Soc. 1965, 48, 391. (28) Li, Y. L.; Ishigaki, T. J. Cryst. Growth 2002, 242, 511. (29) Vemury S.; Pratsinis, S. E. J. Am. Ceram. Soc. 1995, 78, 2984. (30) Li, Y. L.; Ishigaki, T. J. Phys. Chem. B 2004, 108, 15536. (31) Yargeau, V.; Soucy, G.; Boulos, M. I. Plasma Chem. Plasma Process. 1999, 19, 327. (32) Wang, X. H,; Li, J.-G.; Kamiyama, H.; Katada, M.; Ohashi, N.; Moriyoshi, Y.; Ishigaki, T. J. Am. Chem. Soc. 2005, 127, 10982. (33) Tompsett, G. A.; Bowmaker, G. A.; Coony, R. P.; Metson, J. B.; Rodgers, K. A.; Seakins, J. M. J. Raman Spectrosc. 1995, 26, 57. (34) Zhang, Y. H.; Chan, C. K.; Porter, J.; Guo, W. J. Mater. Res. 1998, 13, 2602. (35) Parker, J. C.; Siegel, R. W. Appl. Phys. Lett. 1990, 57, 943. (36) Parker, J. C.; Siegel, R. W. J. Mater. Res. 1990, 5, 1246. (37) Yin, S.; Yamaki, H.; Komatsu, M.; Zhang, Q.; Wang, J.; Tang, Q.; Saito, F.; Sato, T. J. Mater. Chem. 2003, 13, 2996. (38) Evarestov, R. A.; Smirnov, V. P. Phys. Status Solidi B 1999, 215, 949. (39) Mochizuki, S.; Suzuki, Y.; Nakanishi, T.; Ishi, K. Physica B 2001, 308-310, 1046. (40) Nag Chattopadhyay, A.; Dasgupta, P.; Jana, Y. M.; Ghosh, D. J. Alloys Compd. 2004, 384, 6. (41) Judd, B. R. Phys. ReV. 1962, 127, 750. (42) Ofelt, G. S. J. Chem. Phys. 1962, 37, 511.