Optical Materials 46 2015 345-349

Optical Materials 46 (2015) 345–349

Contents lists available at ScienceDirect

Optical Materials journal homepage: www.elsevier.com/locate/optmat

Red shifts of the Eg(1) Raman mode of nanocrystalline TiO2:Er monoliths grown by sol–gel process R. Palomino-Merino a, P. Trejo-Garcia a, O. Portillo-Moreno b, S. Jiménez-Sandoval c, S.A. Tomás d, O. Zelaya-Angel d, R. Lozada-Morales a, V.M. Castaño e,⇑ a Posgrado en Física Aplicada, Facultad de Ciencias Físico-Matemáticas, Benemérita Universidad Autónoma de Puebla, 14 Av. Sur y Av. San Claudio, Col. San Manuel, 72570 Puebla, Mexico b Facultad de Ciencias Químicas, Benemérita Universidad Autónoma de Puebla, 14 Av. Sur y Av. San Claudio, Col. San Manuel, 72570 Puebla, Mexico c Laboratorio de Investigación en Materiales, Centro de Investigación y de Estudios Avanzados del IPN, Unidad Querétaro, Apdo. Postal 1-798, Querétaro 76001, Mexico d Departamento de Física, Centro de Investigación y de Estudios Avanzados del IPN, P.O. Box 14-740, México 07360 D.F., Mexico e Centro de Física Aplicada y Tecnología Avanzada, U.N.A.M, Boulevard Juriquilla 3001, Santiago de Querétaro, Querétaro 76230, Mexico

a r t i c l e

i n f o

Article history: Received 18 February 2014 Received in revised form 26 March 2015 Accepted 17 April 2015 Available online 5 May 2015 Keywords: Titania nanocrystalline Sol–gel Erbium ions Raman red shift

a b s t r a c t Nanocrystalline monoliths of Er doped TiO2 were prepared by the sol–gel technique, by controlling the Er-doping levels into the TiO2 precursor solution. As-prepared and annealed in air samples showed the anatase TiO2 phase. The average diameter of the nanoparticles ranged from 19 to 2.6 nm as the nominal concentration of Er varies from 0% to 7%, as revealed by EDS analysis in an electron microscope. Photo Acoustic Spectroscopy (PAS) allowed calculate the forbidden band gap, evidencing an absorption edge at around 300 nm, attributed to TiO2 and evidence of electronic transitions or Er3+. The Raman spectra, corresponding to the anatase phase, show the main phonon mode Eg(1) band position at 144 cm1 with a red shift for the annealing samples. Ó 2015 Elsevier B.V. All rights reserved.

1. Introduction Since the work of Fujishima and Honda in 1972 [1] intensive research on titanium dioxide has been carried out in order to understand the basic phenomena and to extend the industrial applications of the material. Nowadays, TiO2 is an oxide with a large number of applications due to its important optical, chemical and mechanical properties. In particular, TiO2 is employed as gas and humidity sensor, catalyst, optical coating, photovoltaic cells, rechargeable batteries [2–6]. Many theoretical and experimental works have been reported with the purpose to understand the properties, especially in nanocrystalline form, of TiO2. The Raman spectrometry technique has been a useful and important tool and there have been a variety of publications that interpret of different manners the behavior of the vibrational modes of a Titania matrix as the radius of nanoparticles changes due to the incorporation of different types of ions. Among these interpretations are phonon confinement, non-stoichiometry or internal/stress surface tension effects [7–12]. In this work a Raman analysis on nanocrystalline monoliths of Er doped TiO2 prepared by employing the sol–gel method is ⇑ Corresponding author. http://dx.doi.org/10.1016/j.optmat.2015.04.042 0925-3467/Ó 2015 Elsevier B.V. All rights reserved.

reported. Er doped TiO2 monoliths have potential applications in photocatalytic processes for polluted air and water cleaning and in energy storage. The radius of quantum dots lies in the 2.5 ± 0.4–19.0 ± 0.5 nm. The Raman spectra belong to the anatase phase of TiO2. The main phonon mode Eg(1) band position at 144 cm1 shows a red shift for annealed in air samples. These experimental facts are discussed in base of current theoretical and experimental results. 2. Experimental Nanocrystalline monoliths of titania were prepared by mixing, at room temperature (RT), 1.0 mol of titanium isopropoxide (TIPO) (Aldrich) with 2.0 mol of acetic acid (Aldrich) and 2.0 mol of isopropanol (Baker) under strong stirring. Separately, 0.5 mol of Er(NO3)3:5H2O (Aldrich) was dissolved in a mixture of 2.0 mol of water, 2.0 mol of isopropanol and 2.0 mol of acetic acid. Subsequently, both solutions were mixed under vigorous agitation at RT. With this procedure it was possible to obtain titanium dioxide crystals doped with 0.01, 0.03, 0.04, 0.05, 0.06 and 0.07 M concentrations of Er. The gel obtained from each sample was slowly dried during the gelation for 2 months and immediately underwent thermal annealing in air at 700 °C during 1 h.

346

R. Palomino-Merino et al. / Optical Materials 46 (2015) 345–349

Diffractograms of the TiO2:Er crystalline structure were obtained by using a standard X-ray Siemens D-5000 powder diffractometer with the Cu Ka radiation. Optical absorption was measured by means of a home-made closed photoacoustic cell. Semi-quantitative measurements of atomic concentration of Ti, O, and Er elements were achieved by means of the electron dispersion spectroscopy (EDS) technique, utilizing a Voyager II X-ray quantitative microanalysis in an 1100/1110 EDX system from Noran Instruments. Raman measurements were achieved by means of a Jobin Yvon model Labram micro Raman spectrometer.

3. Results and discussion X-ray diffraction (XRD) patterns, in the 2h range 10–70° of the samples after thermal annealing (TA) in air are displayed in Fig. 1. The material exhibits the anatase crystalline structure of TiO2, some peaks reveal the presence of small amounts of rutile and brookite phases in undoped TiO2 [13], nevertheless, in Er doped material these phases are no present. Samples with no TA show similar XRD patterns. Reflections due to Er and Er-oxides were not observed, like in TiO2:Er nanoparticles reported by other authors [14]. The XRD peaks become broader with the Er presence, which can be consequence of the size reduction of particles as more Er is incorporated into TiO2. The Er content in atomic percentage (at.%) in the nanocrystalline TiO2, measured by using EDS, versus the nominal concentration (in percentage units) of Er [NEr] in the growing solution is displayed in Fig. 2. An approximated linear dependence is observed for the whole range of [NEr] studied. The solid line represents the best linear least square calculation to the data. The slope of the straight line is 0.67 ± 0.04, which indicates that, in general, only 2/3 of nominal Er percentage is observed in each monolith, the 1/3 remaining is probably either evaporated during gelation (2 months) or sublimated during the thermal annealing at 700 °C. The optical absorption (OA) spectra, measured by using photoacoustic (PA) spectroscopy, in the range 1.5–4.0 eV, are presented in Fig. 3 for all the TA–TiO2:Er samples. The advantage of using the PA spectroscopy for OA measurements is that it is less sensitive to light scattering effects than conventional optical techniques [15]. The spectra evidence an absorption edge at photon energies of around 3.0 eV and six electronic transitions which indicate some well defined Er atomic energy absorption levels. The six electronic transitions, identified with triply ionized erbium ions (Er3+), are 4I9/ 4 4 2 4 2, F9/2, S3/2, H11/2, and F3/2, all of them are transitions until the ground level 4I15/2 [16]. The inset in Fig. 3 displays how the 4S3/2 level is resolved from the 2H11/2 by a deconvolution procedure using Gaussian lineshapes.

Fig. 2. EDS measured concentration of Er versus the nominal [NEr] concentration of Er in the growing precursor solution.

Fig. 3. Photoacoustical optical absorption spectra (PAS) for all the [NEr] values analyzed and shown the electron transitions of Er3+ ions. The inset describes the deconvolution process applied to resolve the 4S3/2 and 2H11/2 emissions.

Fig. 1. X-ray diffraction patterns for all the [NEr] values studied. A, B and R letters joined to the indexes on XRD-peaks denotes anatase, brookite and rutile, respectively.

The nanoparticle average grain size (GS) as a function of [NEr] for TA–TiO2:Er is exhibited in the inset of Fig. 4. A similar trend is observed for unannealed samples. GS was calculated from the full width at half maximum (FWHM) of reflection-peaks of XRD patterns and the Scherrer’s formula by considering, as an approximation, a spherical shape of the grains. It can be observed that GS decreases when [NEr] increases. Bahtat and co-workers have suggested that Er3+ strongly affects the crystallization of sol–gel derivated titania by stabilizing amorphous material [17]. Then, in the

R. Palomino-Merino et al. / Optical Materials 46 (2015) 345–349

347

Fig. 5. Raman spectra of the all the doped samples. The left inset displays how the Eg(1) mode band red shift move as [NEr] varies. The right inset illustrates that undoped TiO2 does not show the bands with asterisks associated with impurities.

Fig. 4. The direct Egd and the indirect Egi band gaps as a function of R. The inset displays the radius of the nanoparticle versus [NEr].

present case, thermal annealing of TiO2:Er samples produced nanocrystalline grains with smaller GS as [NEr] rises, this fact is valid for TA–TiO2:Er and unannealed samples. A minimum value of GS is observed in Fig. 4 for [NEr] = 5%, for [NEr] = 6% and 7% GS has higher values. This behavior indicates that GS does not follow a monotonically decreasing process as [NEr] increases. This fact could be related to a limited solubility of Er in TiO2 since due to this effect Li and co-workers [18] have reported that for Er around 3 at.% in TiO2 crystalline phases, O2 and Er start exchanging. In our samples no other phases different of TiO2 were observed, and probably that limit of solubility occurs around [NEr]  5% (3.3 at.% measured by EDS according to Fig. 2) and for [NEr] > 5% no more Er ions enter into the nanocrystalline lattice of TiO2. For [NEr] = 6% and 7% amorphous material composed by Ti, O, and/or Er elements, aside from crystalline TiO2:Er, can be present covering the nanoparticles as suggested by the broader bands observed in Fig. 1 at low 2h values for these two values of [NEr]. The direct (Egd) and indirect (Egi) forbidden energy band gaps of TiO2:Er nanoparticles were calculated by employing the relation (ahm)n / (hm  Eg), n = 1/2 for the indirect band gap and n = 2 for the direct one. Here, a is the optical absorption coefficient, and hm is the photon energy. Taking into account that under the conditions studied in this work the PA amplitude is directly proportional to a, the relation (PA amplitude ⁄ hm)n / (hm  Eg) can be employed to calculate Egd and Egi. In Fig. 4, Egd and Egi versus the average radius of particles (R) is plotted. It can be observed that Egd and Egi follows an inverse of R function, actually is R2 as fitted by us (not shown). The quantum confinement occurs in the intermediate-weak confinement region since the Bohr radius of anatase is 0.8 nm [19]. The Raman spectra of TA–TiO2:Er nanoparticles is displayed in Fig. 5, where the phonon modes associated to anatase Eg(1), Eg(2), B1g(1), A1g(1) + B1g(2) and Eg(3) at 144, 192, 399, 520, and 640 cm1, respectively, are present, all of them correspond to the six modes of TiO2 in the 110–700 cm1 range: 1A1g + 2B1g + 3Eg [8,20]. The left inset shows the position of the Eg(1) phonon mode at

144 cm1 for the undoped sample and shown a red shift movement (observe the left direction of the arrow) of the dopped samples as [NEr] changes in the samples, these bands also shows an asymmetric lineshape. This mode is the most intense Eg Raman mode of anatase and is commonly chosen for analysis [8]. Raman modes pointed out with an asterisk do not belong to anatase. The right inset shows the Raman spectrum of undoped TiO2, where these two modes are not observed. In the inset of Fig. 6, other new Raman modes at the right side of the mode at 144 cm1, the B2g and the B3g, are present in TiO2:Er samples which means that these new modes are due to the presence of Er in the TiO2 lattice. Hardwick and coworkers [21] have reported that Li (Lithium ions) insertion into anatase-type nanocrystalline TiO2 deformates the pure-TiO2 tetragonal lattice of anatase (symmetry group I41/amd-D19 4h) to transform it to an orthorhombic structure (Imma-D28 2h), where the new Raman active modes are nine: Ag(1), Ag(2), Ag(3), B2g(1), B2g(2), B2g(3), B3g(1), B3g(2), B3g(3), i.e., 3Ag + 3B2g + 3B3g. Both A1g and B1g become the Ag modes, and Eg modes split into B2g and B3g. In the present case, the insertion of Er in the TiO2 can produce a similar behavior. A transformation of

Fig. 6. Eg(1) mode position versus the radius R of the nanoparticle. The inset illustrates the deconvolution method to fit the B2g and the B3g modes.

348

R. Palomino-Merino et al. / Optical Materials 46 (2015) 345–349

this type could not be detected in the diffractograms of TiO2:Er (see Fig. 2), however, Raman spectra registered small contribution of the B2g and B3g modes, as Eg(1) at 144 cm1 splits into the B2g, B3g doublet at 163 and 177 cm1, respectively. This unfolding can be observed in the inset of Fig. 6 in a detail of Raman spectrum in the 115–185 cm1 of a representative nanocrystalline TiO2:Er sample, with a deconvolution process using a Lorentzian lineshapes. The lack of symmetry at left side of Eg(1) band (the experimental points fall above the Lorentzian curve) is originated by the nanocrystalline character of the TiO2:Er samples [22]. The two Raman bands indicated with asterisks in Fig. 5 are located at around 280 and 340 cm1, were observed in TiO2:Li (Lithium ions) samples at 276 and 339 cm1, respectively. Nevertheless, they were not identified in Ref. [21], together with other four modes within the interval 276–464 cm1, with any of the 3Ag + 3B2g + 3B3g bands coming from the orthorhombic symmetry of TiO2:Li samples. They have been momentary assigned to different positions of Li atoms in the TiO2 lattice [21]. Therefore the ions into the matrix induced polymorph of anatase and the Raman modes have been associated with impurities present in natural TiO2 material [29]. We can assume that in our TiO2:Er material, erbium induces a slight structural deformation which gives origin to new Raman bands in a similar way as in Refs. [21] and [23]. The central-position of the Eg(1) Raman band, determined from the center of the fitted Lorentzian-like curves (see inset in Fig. 6), versus the radius of nanoparticle (R or Grain Size) has been plotted in Fig. 6. In spite of the quantum confinement of nanoparticles is not in the strong regime, since the Bohr radius of anatase is 0.8 nm [19], the experimental results follow a well-defined tendency in accordance, qualitatively at least, with the theoretical expression xn2 = x2LO b2L (ln/r)2, where the square frequency of the observed first LO phonon x1, is proportional to the square inverse of the radius [24,25]. Fig. 7 displays the FWHM of the Eg(1) Raman phonon versus R. The FWHM decreases as R also decreases (observe the direction of arrow), like cubic CeO2

Fig. 8. Unit cell volume as a function the radius R of the nanoparticle.

nanoparticles where FWHM diminishes as R decreases [26]. But, actually, this result is contrary to the tendency expected in most peak line-width versus R data [8,27–29], where FWHM increases when R decreases. Besides, the inset of Fig. 7, show the Eg(1) peak position as a function of the FWHM resembles a linear dependence with the Eg(1) position, similar to several TiO2 nanocrystalline samples [8,30]. On the other hand, the unit cell volume (UCV) decreases as R decreases as can be observed in Fig. 8. This fact indicates that under the influence of Er doping the TiO2 nanoparticles experience a compressive stress. In CeO2 and BaTiO3 QD’s UCV follows a contrary tendency [8,31]. The apparent controversial results can be explained by considering the following hypothesis: in the present case, the compressive stress in TiO2 nanoparticles due to Er is not the same for the different R values. In general, smaller R larger Er concentration and more compressive stress. Besides, Er induced compressive stress is not similar to an external induced one, this last stress is exerted in a uniform manner into all the particle volume, and the case of internal induced stress, like the produced by Er into TiO2 particle volume, the compression can be relaxed at the surface border. If this internal stress at the surface is larger than the surface tension stress, the effect of surface relaxation can be more evident on the physical properties of nanoparticles. 4. Conclusions

Fig. 7. Full Width at Half Maximum (FWHM) of the Eg(1) versus the radius R of the nanoparticle for all samples. The inset shows the Eg(1) mode position versus the FWHM.

Nanocrystalline Er doped TiO2 monoliths were prepared by mean of the sol–gel technique. The average diameter of the nanoparticles considered as spheres depends on the Er concentration incorporated into the TiO2 lattice. The Er concentration in the crystalline monoliths is around 2/3 of the Er nominal concentration in the growing precursor solution. By employing XRD patterns and TEM images, the nanocrystalline character of the oxide was determined. The direct and indirect band gap energies depend on the size of the particle approximately following the well known theoretical Eg versus R2 linear dependence. Photoacoustic spectra evidence the 4I9/2, 4F9/2, 4S3/2, 2H11/2, and 4F3/2, electronic transitions of Er3+ ions. Analysis of area under the 4S3/2, 2H11/2, and 4F9/2 bands is discussed in terms of the Er concentration in TiO2 nanoparticles.

R. Palomino-Merino et al. / Optical Materials 46 (2015) 345–349

We report a red shift of a Raman mode Eg(1) and the split of this mode into B2g and B3g due the insertion of the Er ions.

References [1] A. Fujishima, K. Honda, Nature 238 (1972) 37. [2] Y. Yan, S.R. Chaudhuri, D.G. Chen, A. Sarkar, Chem. Mater. 7 (1995) 2007. [3] G. Gusamano, G. Montesperelli, P. Nuziante, E. Traversa, A. Montenero, M. Braghini, G. Mattogno, A. Bearzotti, J. Ceram. Soc. Jpn. 101 (1993) 1095. [4] Y. Hamasaki, S. Ohkubo, K. Murakami, H. Sei, G. Nogami, J. Electrochem. Soc. 141 (1994) 660. [5] A. Bahtat, M. Bouazaoui, M.M. Bahtat, C. Garapon, B. Jacquier, J. Mugnier, J. Non-Cryst. Solids 202 (1996) 16. [6] S.Y. Huang, L. Kavan, A. Kay, M. Grätzel, J. Electrochem. Soc. 142 (1995) 142. [7] L.-H. Liang, C.-M. Shen, X.-P. Chen, W.-M. Liu, H.-J. Gao, J. Phys.: Condens. Matter 16 (2004) 267. [8] V. Swamy, A. Kuznetsov, L.S. Dubrovinsky, R.A. Caruso, D.G. Shchukin, B.C. Muddle, Phys. Rev. B 71 (2005) 184302. [9] K.-R. Zhu, M.-S. Zhang, Q. Chen, Z. Yin, Phys. Lett. A 340 (2005) 220. [10] V. Swamy, D. Menzies, B. Muddle, A. Kuznetsov, L. Dubrovinsky, Q. Dai, V. Dmitrev, Appl. Phys. Lett. 88 (2006) 243103. [11] L. Djerdj, A.M. Tonejc, J. Alloys Compd. 413 (2006) 159. [12] A. Orendorz, A. Brodyanski, J. Lösch, L.H. Bai, Z.H. Chen, Y.K. Le, C. Ziegler, G. Gnaser, Phys. Status Solidi (c) 6 (2007) 1822. [13] Anatase: PDF 73-1764. Rutile: PDF 78-1509. Brookite: PDF 76-1936. [14] S. Jeon, P. Braun, Chem. Mater. 15 (2003) 1256. [15] T. Toyoda, I. Tsuboya, Rev. Sci. Instrum. 74 (2003) 782.

349

[16] A.A. Kaminskii, Laser Crystals: Their Physics and Properties, Springer-Verlag, Berlin, Heidelberg, 1990, p. 10. [17] A. Bahtat, B. Bouderbala, M. Bahtat, M. Boauzaoui, J. Mugnier, M. Druetta, Thin Solid Films 59 (1998) 323. [18] J.-G. Li, X.-H. Wanga, H. Kamiyamaa, T. Ishigakia, T. Sekiguchib, Thin Solid Films 506/507 (2006) 292. [19] H.I. Elim, W. Ji, A.H. Yuwono, J.M. Sue, J. Wang, Appl. Phys. Lett. 82 (2003) 2691. [20] M. Mikami, S. Nakamura, O. Kitao, H. Arakawa, Phys. Rev. B 66 (2002) 155213. [21] L.J. Hardwick, M. Holzapfel, P. Novák, L. Dupont, E. Baudrin, Electrochim. Acta 52 (2007) 5357. [22] M. Ivanda, S. Music´, A. Turkovic´, A.M. Tonejc, O. Gamulin, J. Mol. Struct. 480/ 481 (1999) 641. [23] J.C. Jackson, J.W. Horton Jr., I.-Ming Chou, H.E. Belkin, Am. Mineral. 91 (2006) 604. [24] C. Trallero-Giner, A. Debernardi, M. Cardona, E. Menéndez-Proupin, A.I. Ekimov, Phys. Rev. B 57 (1998) 4664. [25] W. Wei, G. Xu, J. Wang, T. Wang, Vacuum 81 (2007) 656. [26] F. Zhang, S.-W. Chan, J.E. Spanier, E. Apak, Q. Jing, R.D. Robinson, I.P. Herman, Appl. Phys. Lett. 80 (2002) 127. [27] J.E. Spanier, R.D. Robinson, F. Zhang, S.-W. Chan, I.P. Herman, Phys. Rev. B 64 (2007) 245407. [28] A. Orendorz, A. Brodyanski, J. Lösch, L.H. Bai, Z.H. Chen, Y.K. Le, C. Ziegler, H. Gnaser, Phys. Status Solidi (c) 4 (2007) 1822. [29] X.L. Wu, T. Gao, X.M. Bao, F. Yan, S.S. Jiang, D. Feng, J. Appl. Phys. 82 (1997). [30] Y. Djaoued, S. Badilescu, P.V. Ashrit, D. Bersani, P.P. Lottici, J. Robichaud, J. Sol– Gel Sci. Technol. 24 (2002) 255. [31] T.-C. Huang, M.-T. Wang, H.-S. Sheu, W.-F. Hsieh, J. Phys.: Condens. Matter 19 (2007) 476212.