Structural and Spectroscopic Properties of Luminescent Er3+-Doped SiO2-Ta2O5 Nanocomposites

Journal

J. Am. Ceram. Soc., 94 [4] 1230–1237 (2011) DOI: 10.1111/j.1551-2916.2010.04191.x r 2010 The American Ceramic Society

Structural and Spectroscopic Properties of Luminescent Er31-Doped SiO2–Ta2O5 Nanocomposites

Jefferson L. Ferrari,z Karmel O. Lima,z Lauro J. Q. Maia,y Sidney J. L. Ribeiro,z and Roge´ria R. Gon@alvesw,z z

Departamento de Quı´ mica, Faculdade de Filosofia, Cieˆncias e Letras de Ribeira˜o Preto, USP, 14040-901 Ribeira˜o Preto/SP, Brazil

y

Grupo Fı´ sica de Materiais, Instituto de Fı´ sica, UFG, Campus Samambaia, 74001-970 Goiaˆnia/GO, Brazil z

Laborato´rio de Materiais Fotoˆnicos, Instituto de Quı´ mica, UNESP, 14801-970 Araraquara/SP, Brazil

range. In the case of Er31-doped materials, they can be used as amplifiers for operation in the S, C, and L bands (1460–1610 nm), depending fundamentally on the chemical composition of the host. Most works on broadband emitters in the near-infrared region reported in the literature are related to tellurite glasses, and most of them are doped with Er31 or Tm31 ions, or codoped with Er31/Tm31.3–7 Concerning silica-based glass– ceramic systems, Tikhomirov et al.,7 have published some results about broadband emission in the near-infrared region of an oxyfluoride glass–ceramics. An Er31-doped oxyfluoride glass– ceramic containing CaF2 nanocrystals with sizes ranging between 5 and 38 nm was reported to have full-width at halfmaximum (FWHM) between 70 and 60 nm for the emission band located around 1550 nm. The FWHM of the spectra was depicted to be directly dependent on the annealing temperature.8 On the bases of the facts mentioned above, different methods for the preparation of optical materials for photonic application have been reported in the literature. These methodologies have yielded materials with good optical properties, and many of those processes are currently used in the manufacture of devices. In this context, the sol–gel route has emerged as one of the cheapest and most versatile routes for the preparation of SiO2based materials with valuable prospects for photonic application.9 Homogeneous multicomponent systems can be easily obtained by mixing the molecular precursor solutions, thereby furnishing high homogeneous multicomponent materials. Moreover, the sol–gel method is potentially applicable in the production potential of optical materials based on silica, and this route allows one to achieve materials at low cost, with excellent composition homogeneity. As stated in many literature papers, the matrix influences the optical properties of Er31-doped materials, and hence the radiative transition probabilities, luminescence lifetime, emission and absorption cross sections, luminescence quantum efficiency, and emission bandwidth are directly dependent on host composition and rare earth concentration. In an attempt to study the spectroscopic properties of the Er31 ions in oxides, some authors have reported on the preparation of sol–gel derived silicabased materials, including Sc2SiO5, Y2SiO5,10 TiO2–SiO2,11 ZrO2–SiO2, HfO2–SiO2,12–16 among others. Such Er31-doped binary oxide systems display luminescence around 1550 nm, with a bandwidth of 45 up to 51 nm, depending on the titanium, zirconium, or hafnium oxide content as well as annealing time and temperature. The emission bandwidth can be a limiting factor in the preparation of an optical amplifier operating in the near-infrared region. Therefore, it is useful to obtain compounds presenting broadband emission of a hundred nanometers. A previous work reported by us revealed that Er31-doped SiO2– Ta2O5 nanocomposites represent excellent candidates for photonic applications, especially those related to broadband emission in the near-infrared region.17

This paper reports on the preparation and structural, morphological, and luminescence properties of Er31-doped nanocomposites based on SiO2–Ta2O5 prepared by the sol–gel method. The influence of the tantalum oxide content on the structural and spectroscopic properties was analyzed for Si/Ta molar ratios of 90:10, 80:20, 70:30, 60:40, and 50:50. The sols were kept at 601C for formation of the xerogel, followed by annealing at 9001, 10001, and 11001C for 2 h for production of the nanocomposites. The densification, phase separation, and crystallization processes were monitored through vibrational spectroscopy (FTIR), X-ray diffraction, and high-resolution transmission electron microscopy. Er31 emission in the infrared region, assigned to the 4I13/2-4I15/2 transition, was observed for all the nanocomposites. Evolution from a vitreous-like environment to a crystalline one was identified upon increasing the annealing temperature and tantalum content. According to the results obtained, the Er31 ions are preferentially localized close to the region rich in Ta2O5 nanoparticles. The localization of Er31 ions was shown to be dependent on the amount of tantalum. Moreover, the fact that the Er31 ions are located close to Ta2O5 nanoparticles promotes a broadband emission with fullwidth at half-maximum of 90 nm around 1550 nm. I. Introduction

O

XIDE materials have been demonstrated to be useful hosts for rare earth ions for use in applications such as solid-state lasers, photo and electroluminescent materials, optical amplifiers, and active planar waveguides. Owing to the high demand in the telecommunication segment, the development of new materials with excellent optical properties and the improvement of the properties of the existing materials are of utmost importance. Among the oxide materials designed for photonic applications, the Er31-doped silica-based materials have been the object of most of the literature reports in this field. It is well known that their utilization as optical amplifiers operating in the C-telecom band, localized in 1530–1565 nm range,1 coincides with the lowloss window of the standard silica host.2 In this sense, many efforts have been made toward the development of materials with broad bandwidth emission in the near-infrared region, which is required for waveguide divisor multiplexing (WDM) devices and amplifier systems operating in a large wavelength

H. H. Du—contributing editor

Manuscript No. 28279. Received July 4, 2010; approved September 16, 2010. This work was financially supported by FAPESP and CNPq. w Author to whom correspondence should be addressed. e-mail: rrgoncalves@ ffclrp.usp.br

1230

April 2011

Structural and Spectroscopic Properties of SiO2–Ta2O5 Nanocomposites

In this work, we report on the structural and spectroscopic properties of luminescent Er31-doped SiO2–Ta2O5 nanocomposites for photonic application. The influence of the tantalum oxide content on the structural and spectroscopic properties was analyzed. Tantalum oxide has some interesting properties regarding optical applications. First of all, high concentrations of rare earth ions may be incorporated into this oxide during the synthetic procedure. Besides, tantalum oxide is transparent over a wide spectral window including the visible and near-infrared regions. Moreover, this oxide has high refractive index and relatively low cutoff phonon energy (o700 cmÿ1); hence, nanocomposites containing Ta2O5 are suitable candidates for rareearth hosts. The combinations of these properties are desirable for the enhancement of the optical properties of rare earth ions, because efficient radiative emission output requires a relatively high refractive index and a low nonradiative decay for excited states. Here, we focus on the structural and spectroscopic properties of the Er31-doped SiO2–Ta2O5 system prepared by the sol–gel method as a function of tantalum content, as well as annealing time and temperature. On the basis of the results reported in a previous work,8–10 the Er31 concentration was chosen so that the best compromise between signal intensity and lifetime would be achieved.

II. Experimental Procedure 31

The Er -doped SiO2–Ta2O5 nanocomposites containing 0.3 mol% Er31 ions were prepared by the sol–gel method. The solutions with a total concentration of Si1Ta 5 0.445 mol/L were obtained at Si/Ta molar ratios of 90:10, 80:20, 70:30, 60:40, or 50:50. Tantalum ethoxide (Aldrich—99.98%, Steinheim, Germany) and tetraethylorthosilicate (TEOS) (Merck—98%, Hohenbrunn, Germany) were used as precursors. The Er31 ion (0.3 mol% in relation to the Si1Ta ions) was added from the corresponding chloride solution, which was prepared from the respective oxide by dissolution in 0.1 mol/L hydrochloric acid aqueous solution, followed by careful drying and dilution in anhydrous ethanol for achievement of the stock ethanolic solution. This solution was then standardized with EDTA 0.01 mol/ L. A mixture of TEOS, anhydrous ethanol, and hydrochloric acid (solution 1) was firstly prepared at a TEOS/HCl volume ratio of 50:1. Tantalum ethoxide, 2-ethoxyethanol, and Er31 ions (solution 2) were mixed in a separate container at a 2-ethoxyethanol/tantalum ethoxide volume ratio of 10:1. Next, the solutions containing TEOS and tantalum ethoxide, namely solutions 1 and 2, were mixed. The final solution was kept under stirring at room temperature for 30 min. Later, a 0.27 mol/L hydrochloric acid aqueous solution was added to the final solution at room temperature using a TEOS/HCl molar ratio of 1:0.007. The volume of the final solution was 20 mL. Subsequently, all the solutions were filtered through a 0.2 mm Millipore filter (Millipore, Billerica, MA) and left to stand for 16 h. Afterwards they were maintained at 601C in an oven, so that xerogels with different Si/Ta molar ratios. Xerogels containing only tantalum oxide were also prepared following the same procedure (solution 2). Next, the xerogels were annealed in an electrical furnace at 9001, 10001, or 11001C for 2 h, for formation of the nanocomposites. To facilitate the discussion, the nanocomposites with Si/Ta molar ratios of 90:10, 80:20, 70:30, 60:40, and 50:50 were labeled S1, S2, S3, S4, and S5, respectively. The powder samples were mixed with KBr and pressed into pellets, and the corresponding Fourier transform infrared (FTIR) spectra were collected in the 4000–400 cmÿ1 range using a Bomem MB102 spectrometer (ABB Bomem Inc., Quebe´c, Canada) with 2 cmÿ1 resolution. The structures of the resulting nanocomposites were checked by X-ray diffraction (XRD) carried out using a diffractometer (Siemens-Bruker D5005, AXS GmbH, Karlsruhe, Germany), with CuKa radiation, l 5 1.5418 A˚, graphite monochromator at 0.031/2 s in the 151–701 2y range. High-resolution transmission electron microscopy (HRTEM) images were acquired using a microscope

1231

(Philips CM 200, Philips Electron Optics, Endhoven, the Netherlands). Statistical histograms were obtained analyzing the HRTEM images. The photoluminescence (PL) spectra at room temperature were registered on a spectrofluorometer (SpectraPro, model 300i, Roper Scientific Germany, Ottobrun, Germany), equipped with an InGaAs detector (model ID-441C). The excitation source was the 488 nm line of an Ar1 laser (250 mW) (model Innova 301C, Coherent Inc., Santa Clara, CA). The lifetime at room temperature was acquired with a Ge detector using an Nd:YAG laser upon excitation at 532 nm and an optical filter RG1000 to eliminate the excitation signal.

III. Results and Discussion Figure 1 presents the FTIR spectra of the samples S1, S3, and S5, where the vibrational bands assigned to the SiO2–Ta2O5 system can be detected. Table I describes the position of the bands and also the bands assignments reported in the literature. The bands around 3433 and 1643 cmÿ1 are assigned to the stretching and bending vibration of the OH group.18 These bands indicate that desorbed water or OH groups still remain in the nanocomposite. The intensity of both bands decreases as the annealing temperature increases, suggesting that desorbed water and OH groups are removed from the pore when the nanocomposites are submitted to heat treatment. Fixing the annealing time, only the annealing temperature of 11001C was enough to eliminate most of the OH groups that remained from the synthetic precursors. The OH groups were eliminated as the temperature was elevated from 9001 to 11001C. OH elimination from the nanocomposites obtained is very important because when these species are present in materials with emission around 1550 nm, they act as fluorescence-quenching centers due to their vibrational frequency. For instance, the two vibrational quanta of OH groups are required to bridge the energy gap between the 4I13/2-4I15/2 transition of the Er31 ion, and consequently an emission quenching occurs due to nonradiative decay by multiphonon relaxation. The bands located around 2988 and 2872 cmÿ1 in the spectra of the materials annealed at 601C for 2 h are assigned to organic residues19 from ethanol, TEOS, or 2-ethoxyethanol, used as precursors. No signals due to organic groups were detected after annealing at temperatures above 9001C, indicating that the organic groups were removed from the nanocomposite, at least at this detection level. The main peaks in the spectra, localized around 1000 and 1200 cmÿ1, are typical of silica-based glassy materials. In accordance to Almeida et al.,18 the band localized at 1075 cmÿ1 is assigned to the asymmetric stretching of the transversal optical component, while the shoulder around 1185 cmÿ1 is attributed to the stretching of the longitudinal optical component of the Si– O–Si group, evidencing formation of the SiO2 network. The authors also report that the shoulder around 1185 cmÿ1 is more evident in systems containing larger quantities of pores. In this sense, this work shows that the relative intensity of the shoulder around 1185 cmÿ1 decreases in relation to the intensity of the band at 1075 cmÿ1 as a function of the annealing temperature. This behavior is an indication that the heat treatment promotes densification of the materials, as observed from the band assigned to the OH groups, whose intensity reduced as a function of the annealing temperature. The bands around 800 cmÿ1 are ascribed to the Si–O–Si symmetric stretching, which is also evidence that the silica matrix is well connected. The bands around 950 cmÿ1 can be attributed to the presence of Si–O–Ta and Si–OH stretching,20 because both the vibrational groups appear at the same wavenumber. Regularly, for all the spectra, the intensity of the band around 950 cmÿ1 decreases with the rising temperature, as observed for the bands localized around 3433 and 1693 cmÿ1. This strongly suggests that Si–OH elimination occurs. It is noteworthy that the band centered at 950 cmÿ1 continues to appear with significant intensity even in

1232

Vol. 94, No. 4

Journal of the American Ceramic Society—Ferrari et al.

Table I. Infrared Bands and their Assignments Position of bands (cmÿ1) in this work

3433 2988–2872 1643 1075 950 963–498

800

Position of band (cmÿ1) in other works

B3400 (3090–2800) B1650 1075 940 1000–200 1215, 600, and 585 1000 and 500 910 and 730 825 B800

Band assignments

OH stretching19 C–H stretching20 H2O angular deformation19 Si–O–Si stretching19 Si–Oÿ, Si–O–Ta, and Si–OH stretching21 Ta2O5 phonons band22 O3Ta22,23 Ta–Oÿ22,23 Ta–O–Ta22,23 Ta 5 O stretching22,23 Si–O–Si symmetric stretching19

The introduction of tantalum oxide into the silica promotes a structural change, and a chemical bond between the silica and the tantalum oxide could be observed for all the compositions. Furthermore, the bands attributed to the Ta2O5 phonons21,22 localized below 900 cmÿ1 could be observed, as depicted in Fig. 2. The wide band displays peaks at 896, 840, 756, 720, 614, and 543 cmÿ1, assigned to the vibrational modes relative to the many bondings between Ta51 and O2ÿ present in the Ta2O5. In fact, the relative intensity of this band augments as a function of the amount of tantalum. The positions of the bands assigned to the tantalum oxide are displayed in Table I and comparison with the values reported in the literature for this same oxide are also included. The XRD reflection patterns of the nanocomposites are shown in Fig. 3, where the peaks are indentified and compared with those reported in the literature. A structural change clearly occurs as a function of the variation of composition and heat treatment, which denotes a direct dependence of the crystalline phase, as well as size value and distribution on these factors. A large background between 151 and 371 in 2y was detected in all the diffractograms, which corresponds to the amorphous silicabased host, in accord to JCPDS Card number 029-0085,23 because it is typical of silicate glasses. Basically, the observed background is due to the SiO2 network of the nanocomposites, as also discussed in the case of the FTIR results. The background intensity decreases as the temperature and the content of tantalum oxide increase, which can be directly associated with partial crystallization. The crystalline fraction in each nanocom-

Fig. 1. FTIR spectra of 0.3 mol% Er31-doped (A) S1, (B) S3, and (C) S5 nanocomposites, annealed at 601, 9001, 10001, or 11001C for 2 h.

the nanocomposite, which does not contain OH groups. Thus, it might be hypothesized that a greater contribution to the appearance of this band is related to the vibrational mode of the Si–O–Ta group. One can, therefore, see the covalent bond between the two components of the nanocomposite.

Fig. 2. FTIR spectra of 0.3 mol% Er31-doped S1, S3, and S5 nanocomposites, annealed at 11001C for 2 h, region between 1700 and 459 cmÿ1.

April 2011

Structural and Spectroscopic Properties of SiO2–Ta2O5 Nanocomposites

1233

Fig. 3. X-ray diffraction pattern of 0.3 mol% Er31-doped (A) S1, (B) S2, (C) S3, (D) S4, (E) S5 nanocomposites, and (F) Ta2O5, annealed at 601, 9001, 10001, and 11001C for 2 h.

posite becomes larger with the increasing tantalum concentration, thereby resulting in a segregated crystalline Ta2O5 phase dispersed in the amorphous SiO2. The diffraction patterns were identified and also compared with literature data, and the planes are listed in Fig. 3. Figure 3(A) reveals that no fraction corresponding to the crystalline structure can be observed in the XRD pattern of all the nanocomposite containing an Si/Ta molar ratio of 90:10 identified to be independent of the annealing temperature. However, the diffractograms of all the other samples, i.e., S2, S3, S4, and S5, recorded after heat treatment at 10001 and 11001C for 2 h display diffraction patterns relative to the L-Ta2O5 nanocrystalline phase. The diffraction patterns observed in the diffractograms corroborate the orthorhombic crystalline structure of Ta2O5 with

cell parameters a 5 6.1980 A˚, b 5 40.2900 A˚, c 5 3.8880 A˚, and a 5 b 5 d 5 901, which show that this oxide belongs to the space group P21212, in agreement with the JCPDS Card number 250922.24 According to Stephenson and Roth,25 the L-Ta2O5 orthorhombic structure can be represented by a chain of eight edge-sharing pentagons where the unit cell contains 22 Ta atoms and 55 O atoms, and different 12 symmetry sites are occupied by tantalum atoms. No phase composed by the Er31 ion was detected in the XRD diffractograms, indicating that no crystalline erbium oxide was formed, at least within the detection limits of this technique. Using the XRD diffractograms and the Scherrer’s formula, it was possible to estimate the average nanocrystallite size of the Ta2O5 dispersed in the SiO2 amorphous matrix for the S2, S3,

1234

Vol. 94, No. 4

Journal of the American Ceramic Society—Ferrari et al.

Table II. Average Nanocrystallite size Estimated by the Scherrer’s Formula for the S2, S3, S4, and S5 Samples Annealed at 11001C for 2 h Based on the (001) Reflections and the Average Nanoparticles Size Calculated from High-Resolution Transmission Electron Microscopic (HRTEM) Images Samples

Average crystallite size by Scherrer’s formula (nm) (70.5 nm) Average nanoparticles size by HRTEM (nm) (70.5 nm)

S4, and S5 nanocomposites annealed at 11001C for 2 h. The equation is described by t 5 0.89l/Bcosy, where t is an average nanocrystallite size, lA˚ is the X-ray wavelength, B is the FWHM, and yrad is the position of the peak used in the calculation.26 These calculations were accomplished on the basis of the reflection peaks localized at 2y 5 22.761 assigned to the (001) plane. The S1 nanocomposite was not considered in the analysis because it did not display any diffraction patterns assigned to the Ta2O5 phase. Table II summarizes the average crystallite size, which changes from 3.8 up to 11.4 nm on going from the S2 to the S5 sample. The higher the content of tantalum, the larger the nanocrystallite size, which accounts for this observation. Glass–ceramics with orthorhombic Ta2O5 nanocrystals dispersed into the amorphous SiO2-based host were observed from analysis of the XRD data relative to the samples containing more than 20 mol% tantalum (Ta/Si1Ta) and annealed at temperatures of 10001 or 11001C for 2 h. Figure 4 displays the HRTEM images of the samples S1, S3, and S5, annealed at (A) 900, (B) 1000, and (C) 11001C for 2 h. The presence of Ta2O5 spherical shape nanoparticles dispersed into the SiO2 amorphous matrix can be noted even for the com-

Temperature of annealing (2 h)

S1

S2

S3

S2

S5

11001C 9001C 10001C 11001C

o3.8 2.1 3.2 5.2

3.8 — — —

7.5 3.1 5.3 7.9

10.8 — — —

11.4 4.2 6.7 10.0

pounds containing the lowest content of tantalum oxide. The densification process in our system occurs together with the onset of crystallization, where Ta2O5 nanocrystals with sizes around 2 nm are dispersed into the SiO2-based host. Figure 5 corresponds to the statistical histogram of the size distribution of the Ta2O5 nanoparticles, which is narrow. The average size of the Ta2O5nanoparticles dispersed into the SiO2 amorphous matrix can be taken from the histograms presented in Fig. 5, and the values of the sizes are depicted in Table II. The achieved results agree with the average nanocrystallite size calculated by the Scherrer’s formula. Glass–ceramics was also detected for the S1 samples for all the annealing temperatures, as shown in Fig. 4(1), where the Ta2O5 nanoparticles appear to be well dispersed in the amorphous SiO2 matrix, as it can be confirmed by electron diffraction (inset). Even though no diffraction peaks were detected for sample S1, the HRTEM image clearly shows the occurrence of an initial crystallization process, with average sizes around 2.1, 3.2, and 5.2 nm for the nanocrystals in the samples annealed at 9001, 10001, and 11001C, respectively. The volume of nanocrystals detected for the sample S1 is smaller those of samples S3 and S5, which would justify the absence of diffraction peaks in the first case.

Fig. 4. HRTEM images of (1) S1, (2) S3, and (3) S5 Er31-doped nanocomposites annealed at (A) 9001C, (B) 10001C, and (C) 11001C for 2 h.

April 2011

1235

Structural and Spectroscopic Properties of SiO2–Ta2O5 Nanocomposites

Fig. 6. Photoluminescence spectra in the 1550 nm range assigned to the 4 I13/2-4I15/2 transitions of 0.3 mol% Er31-doped (A) S1, (B) S2, (C) S3, (D) S4, (E) S5 nanocomposites, and (F) Ta2O5, annealed at 11001C for 2 h upon excitation at 488 nm.

Fig. 5. Histograms of 0.3 mol% Er31-doped (A) S1, (B) S3, and (C) S5 nanocomposite annealed at 9001, 10001, and 11001C for 2 h. Statistical histograms were obtained from the HRTEM images.

The emission spectra of the nanocomposites acquired in the 1400–1700 nm range are shown in Fig. 6. The emission spectra of the corresponding Er31-doped Ta2O5 powder are also included for comparison. Luminescence from the excited 4I13/2 state to the ground 4I15/2 state is observed from a broadened band with a maximum around 1535 nm, whose spectral bandwidth changes with the content of tantalum oxide. Table III summarizes the FWHM values for the nanocomposites, which increased from 74 up to 93 nm as the tantalum oxide content became larger. The broadest band, around 92 nm (75 nm), was registered for the S4 (Fig. 6(D)) and S5 (Fig. 6(E)) samples annealed at 11001C for 2 h. Such inhomogeneous broadening

probably indicates the presence of Er31 ions in many different sites in the structure of the nanocomposite. The Er31 ions can substitute the Ta51 ions and/or be also placed in the interstitial space present inside of the structure of Ta2O5. As discussed in the XRD results, the Ta51 ions are localized in different symmetry sites in orthorhombic Ta2O5, more precisely 12 sites, so if the Er31 ions substitute the Ta51 ions, the observed inhomogeneous broadening in the emission spectra can be explained. In the case of SiO2–Ta2O5-based materials, Ta51 may be promoting the rupture of the SiO2 network. Consequently, the number of SiOÿ species rises and the matrix can accommodate Er31 ions in different environments without appreciable strains on the matrix. The spectral width of the emission band is due to inhomogeneous broadening plus additional Stark splitting27,28 of the first excited and ground states of the Er31 ions. This FWHM value is considered high compared with that of silicate systems7–10 and of other systems reported in the literature, such as in the case of Er31-doped niobic tellurite glass,29 lead tellurite glass B67 nm,30 phosphate B50 nm,31 and silicate glass B40 nm.32 Duverger et al.33 have reported that the bandwidth of the Er31 emission around 1530 nm also changes between 34 (72 nm) and 60 nm (72 nm), depending on the Er31 concentration. For comparison purposes, we have prepared orthorhombic tantalum oxide doped with Er31, whose emission spectrum can be seen in Fig. 6(F). A bandwidth of 92 nm (75 nm) was registered, as observed for the samples S4 (Fig. 6(D)) and S5 (Fig. 6(E)) annealed at 11001C for 2 h. It is noteworthy that the bandwidth increases as the amount of tantalum oxide in the nanocomposite increases, and it reaches similar values to those obtained for the samples containing tantalum oxide only. This strongly indicates that most of the Er31 ions are localized in an environment rich in tantalum oxide. For the samples containing low tantalum oxide content, the Er31 ions could also be accommodated in different environments in the silica-based matrix. Compared with other literature works, where the authors describe the advantages of materials for application in telecom-

Table III. Values of Bandwidth and Central Position of the Band Emission Around 1550 nm for the Nanocomposites Annealed at 11001C for 2 h Samples

FWHM (nm)75 nm

Central position (nm)

S1 S2 S3 S4 S5 Ta2O5

74 77 81 92 91 93

1538 1538 1536 1539 1539 1530

FWHM, full-width at half-maximum.

1236

Vol. 94, No. 4

Journal of the American Ceramic Society—Ferrari et al.

munications due to broadband emission around 1550 nm, in this work a broadband emission in this region was acquired with relatively low Er31 concentration (0.3 mol%). Here, it is very important to mention that it is necessary to know about the suitable concentration of rare earth inside the matrix for good optical quality to be achieved.12,14 In this sense, a more detailed study about the optimal Er31 concentration in the matrix reported in this work is in progress in our laboratory. An important meaning of large bandwidth values is related to obtaining a more flat-gain region of Er31-doped materials, in order to increase the number of WDM channels. The large optical bandwidth of the Er31-doped SiO2–Ta2O5 prepared by the sol– gel method makes it a suitable candidate for WDM applications. Figure 7(A) depicts the PL decay curve for the 4I13/2 level of Er31 ions for the S3 nanocomposite annealed at 10001C for 2 h, which is well representative of all other samples. The best fit adjusted for the exponential curves was the first-order type. The measured and calculated linear measures are listed in Fig. 7(A), and the lifetime values as a function of the content of tantalum oxide are shown in Fig. 7(B). The lifetime values diminish as a function of the content of which is strong indication that the Er31 ions prefer to diffuse to Ta2O5 nanoparticles instead of being localized in the SiO2 matrix. In accordance to Sloof et al.,34 the Er31 lifetime in the SiO2 environments is around 12 ms, while the Er31 lifetime in the Ta2O5 matrix observed in this work is around 1.40 ms. An important goal for future studies on rare earth-doped SiO2–Ta2O5 films is the possibility of making erbium-doped waveguide amplifiers for several channels in the near-infrared region in a cost-efficient manner.

IV. Conclusions In conclusion, new Er31-doped SiO2–Ta2O5 nanocomposites with different Si/Ta molar ratios were prepared by the sol–gel method. Samples containing the lowest tantalum content (10 mol%), annealed up to 11001C, exhibit phase separation with nanocrystal sizes of up to 3.8 nm. Upon elevation of the tantalum content, more pronounced phase separation and crystallization process are observed, depending on the annealing temperature and nanocomposite composition. The densification process as a function of the annealing time was also accompanied by FTIR, which enabled detailed observation of the elimination of OH groups, formation of the SiO2 network, as well as Ta2O5. Consequently, glass–ceramics with orthorhombic Ta2O5 nanocrystals well dispersed in amorphous SiO2-based matrix were observed for all the nanocomposites. An evolution from a vitreous-like environment to a crystalline one was identified upon increasing the annealing temperature and tantalum content. The technologically important Er31 emission in the infrared region, assigned to the 4I13/2-4I15/2 transition, was detected for all the nanocomposites. All the nanocomposites annealed at 11001C for 2 h displayed a broadband emission, with a maximum at 1535 nm. Bandwidth values increased from 64 up to 92 nm as the tantalum content augmented. The broader band, around 92 nm (75 nm), was registered for the Si/Ta molar ratios of 60:40 and 50:50 annealed at 11001C for 2 h. Such inhomogeneous broadening probably indicates the presence of Er31 ions in many different sites in the glass–ceramic structure. The lifetime value of the 4I13/2 state indicates that the Er31 ions are preferentially localized in a region rich in Ta2O5 nanoparticles. The large optical bandwidth of the Er31-doped SiO2– Ta2O5 prepared by the sol–gel method makes this material a suitable candidate for optical amplifier and wavelength division multiplexing applications, where an increased number of channels can be considered.

Acknowledgments The authors would like to acknowledge Capes for the scholarship. The authors also acknowledge Mrs. Cynthia Maria de Campos Prado Manso and Mr. Rafael I. Estrada Mejı´ a for reviewing the text.

References 1

Fig. 7. (A) Photoluminescence decay curve from the 4I13/2 metastable state of Er31 ions in the S3 nanocomposite annealed at 10001C/2 h, upon excitation at 532 nm. (B) Lifetime values as a function of the tantalum oxide concentration for the S1, S2, S3, S4, and S5 samples annealed at 9001, 10001, and 11001C/2 h.

A. Polman and F. C. J. M. van Veggel, ‘‘Broadband Sensitizers for ErbiumDoped Planar Optical Amplifiers: Review,’’ J. Opt. Soc. Am. B, 21 [5] 871–92 (2004). 2 L. J. Q. Maia, V. R. Mastelaro, S. Pairis, A. C. Hernandes, and A. Ibanez, ‘‘A Sol–Gel Route for the Development of Rare-Earth Aluminum Borate Nanopowders and Transparent Thin Films,’’ J. Solid State Chem., 180 [2] 611–8 (2007). 3 G. E. Keiser, ‘‘A Review of WDM Technology and Applications,’’ Opt. Fiber Technol., 5 [1] 3–39 (1999). 4 M. Djeneka and B. Samson, ‘‘Rare-Earth-Doped Fibers for Telecommunications Applications,’’ MRS Bull., 24 [9] 39–45 (1999). 5 E. Desurvire, Erbium-Doped Fiber Amplifiers. John Wiley, New York, 1994. 6 R. Balda, J. Fernandez, and J. M. Fernandez-Navarro, ‘‘Study of Broadband Near-Infrared Emission in Tm31–Er31 Codoped TeO2–WO3–PbO Glasses,’’ Opt. Express, 17 [11] 8781–8 (2009). 7 V. K. Tikhomirov, D. Furniss, A. B. Seddon, I. M. Reaney, M. Beggiora, M. Ferrari, M. Montagna, and R. Rolli, ‘‘Fabrication and Characterization of Nanoscale, Er31-Doped, Ultratransparent Oxy-Fluoride Glass Ceramics,’’ Appl. Phys. Lett., 81 [11] 1937–9 (2002). 8 D. Chen, Y. Wang, Y. Yu, and E. Ma, ‘‘Improvement of Er31 Emissions in Oxyfluoride Glass Ceramic Nano-Composite by Thermal Treatment,’’ J. Sol. State Chem., 179 [5] 1445–52 (2006). 9 L. L. Hench and J. K. West, ‘‘The Sol–Gel Process,’’ Chem. Rev., 90 [1] 33–72 (1990). 10 L. Fornasiero, K. Petermann, E. Heumann, and G. Huber, ‘‘Spectroscopic Properties and Laser Emission of Er31 in Scandium Silicates Near 1.5 mm,’’ Opt. Mater., 10 [1] 9–17 (1998). 11 R. M. Almeida, A. C. Marques, S. Pelli, G. C. Righini, A. Chiasera, M. Mattarelli, M. Montagna, C. Tosello, R. R. Gon@alves, H. Portales, S. Chaussedent, M. Ferrari, and L. Zampedri, ‘‘Spectroscopic Assessment of Silica– Titania and Silica–Hafnia Planar Waveguides,’’ Philos. Mag., 84 [13] 1659–66 (2004). 12 R. R. Gon@alves, G. Carturan, L. Zampedri, M. Ferrari, M. Montagna, A. Chiasera, G. C. Righini, S. Pelli, S. J. L. Ribeiro, and Y. Messaddeq, ‘‘Sol–Gel Er-

April 2011

Structural and Spectroscopic Properties of SiO2–Ta2O5 Nanocomposites

Doped SiO2–HfO2 Planar Waveguides: A Viable System for 1.5 mm Application,’’ Appl. Phys. Lett., 81 [1] 28–30 (2002). 13 R. R. Gon@alves, G. Carturan, M. Montagna, M. Ferrari, L. Zampedri, S. Pelli, G. C. Righini, S. J. L. Ribeiro, and Y. Messaddeq, ‘‘Erbium-Activated HfO2Based Waveguides for Photonics,’’ Opt. Mater., 25 [2] 131–9 (2004). 14 R. R. Gon@alves, G. Carturan, L. Zampedri, M. Ferrari, A. Chiasera, M. Montagna, G. C. Righini, S. Pelli, S. J. L. Ribeiro, and Y. Messaddeq, ‘‘InfraredTo-Visible CW Frequency Upconversion in Erbium Activated Silica–Hafnia Waveguides Prepared by Sol–Gel Route,’’ J. Non-Cryst. Solids, 322 [1–3] 306–10 (2003). 15 R. R. Gon@alves, J. J. Guimara˜es, J. L. Ferrari, L. J. Q. Maia, and S. J. L. Ribeiro, ‘‘Active Planar Waveguides Based on Sol–Gel Er31-Doped SiO2–ZrO2 for Photonic Applications: Morphological, Structural and Optical Properties,’’ J. Non-Cryst. Solids, 354 [42–44] 4846–51 (2008). 16 L. Zampedri, C. G. Righini, H. Portales, S. Pelli, L. Zampedri, G. C. Nunzi, M. Montagna, M. Mattarelli, R. R. Gon@alves, M. Ferrari, A. Chiasera, M. Bouazaoui, and C. Armellini, ‘‘Sol–Gel-Derived Er-Activated SiO2–HfO2 Planar Waveguides for 1.5 ı´ m Application,’’ J. Non-Cryst. Solids, 345 & 346 [15] 580–4 (2004). 17 J. L. Ferrari, K. O. Lima, L. J. Q. Maia, S. J. L. Ribeiro, A. S. L. Gomes, and R. R. Gon@alves, ‘‘Broadband NIR Emission in Sol–Gel Er31-Activated SiO2– Ta2O5 Glass Ceramic Planar and Channel Waveguides for Optical Application,’’ J. Nanosc. Nanotech., 10, 1–5 (2010). 18 R. M. Almeida, T. A. Guiton, and C. G. Paˆntano, ‘‘Characterization of Silica Gels by Infrared Reflection Spectroscopy,’’ J. Non-Cryst. Solids, 121 [1–3] 193–7 (1990). 19 D. A. Neumeyer and E. Cartier, ‘‘Materials Characterization of ZrO2–SiO2 and HfO2–SiO2 Binary Oxides Deposited by Chemical Solution Deposition,’’ J. Appl. Phys., 90 [4] 1801–8 (2004). 20 D. Blanc, W. Zhang, C. Massard, and J. Mugnier, ‘‘Synthesis and Characterisation of Tantalum-Incorporating Silica Hybrid Sol–Gel Thin Films for Optical Applications,’’ Opt. Mater., 28 [4] 331–5 (2006). 21 A. P. Huang and P. K. Chu, ‘‘Crystallization Improvement of Ta2O5 Thin Films by the Addition of Water Vapor,’’ J. Cryst. Growth, 274 [1–2] 73–7 (2005). 22 N. Wada, M. Kubo, N. Maeda, M. Akira, and K. Kojima, ‘‘Fluorescence Property and Dissolution Site of Er31 in Ta2O5 Film Prepared by Sol–Gel Method and Dip-Coating Technique,’’ J. Mater. Res., 19 [2] 667–75 (2004).

1237

23 Joint Committee on Powder Diffraction Standards, Diffraction Data File, No. 029-0085—JCPDS International Center for Diffraction Data, Pennsylvania, 1991. 24 Joint Committee on Powder Diffraction Standards, Diffraction Data File, No. 025-0922—JCPDS International Center for Diffraction Data, Pennsylvania, 1991. 25 N. C. Stephenson and R. S. Roth, ‘‘Structural Systematics in the Binary System Ta2O5–WO3. V. The Structure of the Low-Temperature Form of Tantalum Oxide L-Ta2O5,’’ Acta Cryst., B27, 1037–44 (1971). 26 B. D. Cullity, Elements of X-Ray Diffraction, 2nd edition, Addison-Wesley, Massachusetts, 1978, 102pp. 27 Y. D. Huang, M. Mortier, and F. Auzel, ‘‘Stark Levels Analysis for Er31-Doped Oxide Glasses: Germanate and Silicate,’’ Opt. Mater., 15 [4] 243–60 (2001). 28 Y. D. Huang, M. Mortier, and F. Auzel, ‘‘Stark Level Analysis for Er31Doped ZBLAN Glass,’’ Opt. Mater., 17 [4] 501–11 (2001). 29 X. Wang, ‘‘Investigation of Thermal Stability and Spectroscopic Properties in Er31/Yb31 Co-Doped Niobic Tungsten Tellurite Glasses,’’ Spectrochim. Acta Part A: Mol. Biomol. Spectrosc., 70 [1] 99–103 (2008). 30 T. F. Xu, X. Shen, Q. H. Nie, and Y. Gao, ‘‘Spectral Properties and Thermal Stability of Er31/Yb31 Codoped Tungsten–Tellurite Glasses,’’ Opt. Mater., 28 [3] 241–5 (2006). 31 S. Jiang, T. Luo, B. C. Hwang, F. Smekatala, K. Seneschal, J. Lucas, and N. Peyghambarian, ‘‘Er31-Doped Phosphate Glasses for Fiber Amplifiers with High Gain Per Unit Length,’’ J. Non-Cryst. Solids, 263–264 [1] 364–8 (2000). 32 D. Ning, Y. Lv-Yun, P. Ming-Ying, Z. Qing-Ling, C. Dan-Ping, T. Akai, and K. Kadono, ‘‘Preparation and Spectroscopic Properties of Er31-Doped High Silica Glass Fabricated by Sintering Nanoporous Glass,’’ Mater. Lett., 60 [16] 1987–9 (2006). 33 C. Duverger, M. Montagna, R. Rolli, S. Ronchin, L. Zampedri, M. Fossi, S. Pelli, G. C. Righini, A. Monteil, C. Armellini, and M. Ferrari, ‘‘Erbium-Activated Silica Xerogels: Spectroscopic and Optical Properties,’’ J. Non-Cryst. Solids, 280 [1–3] 261–8 (2001). 34 L. H. Sloof, M. J. A. de Dood, A.van Blaaderen, and A. Polman, ‘‘Effects of Heat Treatment and Concentration on the Luminescence Properties of ErbiumDoped Silica Sol–Gel Films,’’ J. Non-Cryst. Solids, 296 [3] 158–64 (2001). &