Rare earth-doped TiO2 nanocrystalline thin films: Preparation and thermal stability

Available online at www.sciencedirect.com

ScienceDirect Journal of the European Ceramic Society 34 (2014) 4457–4462

Short Communication

Rare earth-doped TiO2 nanocrystalline thin films: Preparation and thermal stability Mario Borlaf a , María T. Colomer a , Rodrigo Moreno a , Angel L. Ortiz b,∗ a

b

Instituto de Cerámica y Vidrio, Consejo Superior de Investigaciones Científicas (CSIC), C/Kelsen 5, 28049 Madrid, Spain Departamento de Ingeniería Mecánica, Energética y de los Materiales, Universidad de Extremadura, Avda. de Elvas, S/N, 06006 Badajoz, Spain Received 7 April 2014; received in revised form 30 June 2014; accepted 5 July 2014 Available online 26 July 2014

Abstract A colloidal sol–gel route was used for the synthesis of nanoparticulate TiO2 and Ln3+ -doped TiO2 sols (Ln = Eu or Er; contents of 1, 2, or 3 mol.%), from which the corresponding functional nanocrystalline thin films were subsequently obtained by the dip-coating method. It was found that the as-synthesized sols are not entirely suitable for the preparation of homogeneous thin films due to the water’s high surface tension, a problem that is however solved by diluting the sols in ethanol. Appropriate dilution conditions were then determined, and the effect of this dilution on the sol viscosity identified. Finally, the phase composition in the as-deposited condition and the thermal stability of the dip-coated thin films were investigated by X-ray thermodiffractometry up to 1000 ◦ C. It was found that the as-deposited thin films are homogenous and formed by the desired anatase nanoparticles, which eventually start to transform into rutile particles at high temperature. However, no precipitation of titanates occurs in the temperature range investigated. Also, it was observed that increasing the Ln3+ content improves the thermal stability of these anatase nanocrystalline thin films, an effect that is, if any, slightly more marked for Eu3+ than for Er3+ . © 2014 Elsevier Ltd. All rights reserved. Keywords: Ceramic thin films; TiO2 ; Phase transformation

1. Introduction There is great interest in the preparation of ceramic thin films for their utilization in many functional applications, including photo-induced applications. TiO2 , a well-known semiconductor oxide, is one of those ceramics with an attractive set of properties for the fabrication of thin films used in photocatalytic,1 photoluminescent,2 electrochemical3 and photoelectric4 devices, to name a few potential applications. Anatase, especially if doped with trivalent lanthanides and having a particle size in the nanoscale, is the preferred polymorph of the TiO2 variants (i.e., anatase, rutile, brookite, and ␤-TiO2 ) for photo-induced applications, which justifies the research efforts put on its synthesis5–8 and on the study of its thermal stability.8,9

∗

Corresponding author. Tel.: +34 924289600x86726; fax: +34 924289601. E-mail addresses: [email protected], [email protected] (A.L. Ortiz). http://dx.doi.org/10.1016/j.jeurceramsoc.2014.07.008 0955-2219/© 2014 Elsevier Ltd. All rights reserved.

Polycrystalline ceramic thin films, including TiO2 -based thin films, can in principle be deposited using various methods. Dipcoating is one of the most widely-used methods for depositing oxide ceramics onto large surfaces, where the substrate is first immersed into and subsequently extracted from a precursor liquid solution or suspension thus creating the thin film. This method also offers certain advantages in terms of simplicity, versatility, cost, processing temperatures, quantities produced, etc. Normally, the ceramic precursor solutions/suspensions have been prepared by the polymeric sol–gel route,10–14 which uses organic solvents (i.e., alcohols). Today, however, there is a growing demand for preparing these solutions/suspensions using an aqueous media, and therefore the colloidal sol–gel route, which uses water instead of alcohol and is thus clearly more environmentally friendly and cost-effective, is being the subject of increasing attention.6,15 With these premises, the present study was undertaken with two objectives in mind. The first is to prepare TiO2 nanocrystalline thin films of anatase phase, both undoped and doped with

4458

M. Borlaf et al. / Journal of the European Ceramic Society 34 (2014) 4457–4462

different concentrations of Ln3+ (Ln = Eu or Er; contents of 1, 2, or 3 mol.%) ions, by the dip-coating method from nanoparticulate sols synthesized by a colloidal sol–gel route. The second is to assess the thermal stability of these anatase thin films, as a function of temperature up to 1000 ◦ C. The details of the experiments and the major findings are described below. 2. Experimental The TiO2 sols were prepared by adding titanium(IV) isopropoxide (97%, Sigma-Aldrich, Germany) to a stirring mixture of deionized water (18.2 M cm−1 , ultrapure Milli-Q, France) and nitric acid (65%, PANREAC, Spain) in water:alkoxide and H+ :Ti4+ molar ratios of 50:1 and of 1:5, respectively; HNO3 was used as a catalyst and dispersing agent. For the preparation of the Ln3+ -doped TiO2 sols (Ln = Er or Eu), the corresponding lanthanide(III) acetate hydrate (Ln(OOCCH3 )3 ·xH2 O, SigmaAldrich, Germany) was dissolved in the H2 O–HNO3 mixture to a molar ratio Ln3+ :TiO2 of 1:99, 2:98, and 3:97 (hereafter referred to simply as 1, 2, or 3 mol.% Ln3+ ) prior to the addition of the titanium(IV) isopropoxide. The synthesis temperature was maintained constant during the entire process at 35 ◦ C. The aqueous sols were used in their as-prepared condition, and were also diluted in absolute ethanol (99.5%, Panreac, Spain) using different dilution conditions (i.e., ethanol contents in the range 8–60 vol.%). The viscosity of the as-prepared and diluted sols was measured using a rotational rheometer (RS50, Haake, Thermo, Germany) with a double-cone/plate sensor configuration (DC60/2◦ , Haake, Thermo) operated in controlled shear rate mode with the following measurement cycle: linear increase of the shear rate from 0 to 1000 s−1 in 300 s, plateau at 1000 s−1 for 60 s, and decrease to zero shear rate in 300 s. The average particle size in selected sols was also measured, using dynamic light scattering (DLS; Zetasizer Nano ZS, Malvern, UK).

Thin films were deposited from the sols at room-temperature onto well-cleaned, single-crystal (plane (1 0 0)) Si substrates (∼2 × 1 cm2 ) by the dip-coating method, using a withdrawal rate of 0.5 mm s−1 . Atomic force microscopy (AFM; Nanotec Electronica, Spain) was used to observe the surface topography of selected thin films, with the sole objective of confirming their homogeneity and nanocrystalline nature. Finally, the phase composition in the as-deposited condition and the thermal stability of the thin films were investigated by X-ray thermodiffractometry (XRTD), using to that end a high-resolution diffractometer (D8 Advance, Bruker AXS, Karlsruhe, Germany) of pure Cu-K␣1 ˚ that is equipped with a linear incident radiation (λ = 1.5406 A) ultra-fast detector and with a micro-furnace high-temperature chamber.9 The XRTD patterns were acquired in situ in the temperature range 30–1000 ◦ C with a temperature step of 3 ◦ C, heating ramp of 3 ◦ C s−1 , and delay time of 2 s, and the acquisition was done with the detector operating in snapshot mode (fixed 2θ mode) for 60 s with its window being opened at 11◦ 2θ to register simultaneously the angular range 21–32◦ 2θ. The crystalline phases were identified with the aid of the PDF2 database. 3. Results and discussion The first step was the preparation of TiO2 -based thin films using directly the sols in their as-synthesized condition. However, as can be seen in Fig. 1 for the undoped TiO2 sols, the resulting thin films were, unfortunately, quite inhomogeneous. This is because the water’s high surface tension (∼72 mN m−2 at 25 ◦ C)16 induces the formation of many drops during the slow drying stage at room temperature (water’s boiling temperature is ∼100 ◦ C). Based on this and earlier results7 we decided to dilute the sols in ethanol, which has a lower surface tension (∼22 mN m−2 at 25 ◦ C) and boiling temperature (∼78 ◦ C)16 than water, and investigate the effect of the ethanol concentration on the quality of the resulting thin films. Photographs

Fig. 1. Photographs of the undoped TiO2 thin films obtained by the dip-coating method from the as-synthesized and diluted sols. The number in each photograph indicates the dilution condition (i.e., ethanol concentration in vol.%). The substrate dimensions are 2 cm length × 1 cm width.

M. Borlaf et al. / Journal of the European Ceramic Society 34 (2014) 4457–4462

4459

Table 1 Viscosity values (mPa s) taken at a shear rate of 1000 s−1 for the as-synthesized and diluted sols (60 vol.% ethanol). Sol

Undiluted Diluted

Viscosity (mPa s; ±0.8) TiO2

1 mol.% Eu3+

2 mol.% Eu3+

3 mol.% Eu3+

1 mol.% Er3+

2 mol.%Er3+

3 mol.%Er3+

3.8 10.3

3.8 11.3

3.5 10.0

11.9 9.9

4.1 9.2

4.5 10.3

20.6 8.4

Table 2 Onset temperature for the anatase-to-rutile phase transformation determined for the undoped and Ln3+ -doped thin films and xerogels (taken from Refs. [7] for the xerogels doped with Eu3+ and [8] for the xerogels doped with Er3+ ). Sample

Thin film Xerogel

Temperature (◦ C; ±3) TiO2

1 mol.% Eu3+

2 mol.% Eu3+

3 mol.% Eu3+

1 mol.% Er3+

2 mol.%Er3+

3 mol.%Er3+

660 480

910 635

935 690

955 765

910 570

930 640

950 730

Fig. 3. XRTD patterns of the Si substrate measured at three different conditions: as-received at 25 ◦ C, in situ at 1000 ◦ C, and after cooling down at 25 ◦ C. The patterns have been shifted along the Y-axis to facilitate their comparison.

Fig. 2. AFM micrographs of the TiO2 thin films doped with 1 mol.% of Ln3+ prepared from diluted sols (60 vol.% ethanol): (A) Ln = Eu, and (B) Ln = Er. The insets show the corresponding three-dimensional micrographs.

of the undoped TiO2 thin films obtained for ethanol concentrations in the range 8–60 vol.% are also shown in Fig. 1. Clearly, the thin-film homogeneity increases with increasing the degree of dilution, with the thin films already covering the substrate for ethanol concentrations of 40 vol.% or superior. However, the gradual shifting of the position of the drying line (a typical defect of the thin films prepared by dipping) towards the

edge of the substrate indicates that the ethanol concentration of 60 vol.% is preferable over those of 40 and 48 vol.%, and therefore it was the one selected as dilution condition (surface tension ∼27 mN m−2 at 25 ◦ C)16 for the subsequent preparation of homogeneous undoped and Ln3+ -doped TiO2 thin films. Since the deposition of the thin films will be performed by the dipping technique, it is convenient to investigate at this moment the effect of this particular dilution condition (i.e., 60 vol.% ethanol) on the viscosity of the sols. Table 1 lists the viscosity values determined at a shear rate of 1000 s−1 for all undoped and Ln3+ -doped TiO2 sols prepared in this study. The comparison of values reveals that the sols diluted in ethanol are indeed more viscous than the undiluted sols when the Ln3+ doping content is 0, 1, and 2 mol.%, but are however less viscous when the Ln3+ doping content is 3 mol.%. This trend seems to be the net result of the complex balance between the dilution that reduces the solid content and therefore the sol viscosity, and the polymerization of the organic molecules catalyzed by the acid

4460

M. Borlaf et al. / Journal of the European Ceramic Society 34 (2014) 4457–4462

Fig. 4. Two-dimensional intensity contours extracted from the XRTD patterns as a function of the temperature in the range 30–1000 ◦ C for the TiO2 thin films prepared from diluted sols (60 vol.% ethanol). Ln3+ -doping contents of: (A) 0 mol.%, (B) 1 mol.% Eu3+ , (C) 2 mol.% Eu3+ , (D) 3 mol.% Eu3+ , (E) 1 mol.% Er3+ , (F) 2 mol.% Er3+ , and (G) 3 mol.% Er3+ . The phase identification is included. Anatase is identified by its 1 0 1 peak, and rutile by its 1 1 0 peak.

M. Borlaf et al. / Journal of the European Ceramic Society 34 (2014) 4457–4462

conditions that increases the sol viscosity. It appears that dilution predominates when the Ln3+ doping content is 3 mol.%, and polymerization in the rest of cases, the latter implying necessarily that those sols must have greater particle size after the dilution in ethanol due to the agglomeration of nanoaggregates. To confirm this hypothesis we have measured the particle sizes in the undiluted and diluted sols with 0, 1, and 2 mol.% Ln3+ doping by DLS, obtaining, respectively: 22 ± 2 vs. 46 ± 8 for TiO2 , 11 ± 4 vs. 35 ± 9 for 1 mol.% Eu3+ -doped TiO2 , 32 ± 2 vs. 69 ± 23 for 2 mol.% Eu3+ -doped TiO2 , 32 ± 2 vs. 35 ± 20 for 1 mol.% Er3+ -doped TiO2 , and finally 21 ± 2 vs. 41 ± 8 for 2 mol.% Er3+ -doped TiO2 . Whichever the case, an important implication of the ethanol dilution in practice is that, unlike the as-synthesized sols, the viscosity of the diluted TiO2 sols varies little, if any, with the Ln3+ doping, thus eliminating the dependence of the deposition conditions on the chemical composition of the TiO2 sols. Next, undoped and Ln3+ -doped TiO2 thin films were prepared from the diluted sols (60 vol.% ethanol), which have a uniform microstructure as shown by way of example in the AFM images of Fig. 2 for the two cases with 1 mol.% Ln3+ doping. In addition, these AFM images also reveal the nanocrystalline nature of the thin films prepared here. The thermal stability of the entire set of nanocrystalline thin films was subsequently investigated as a function of temperature up to 1000 ◦ C by XRTD, which is the central objective of the present work. Given that this type of thermal stability studies rests on the correct indexing of the peaks in the XRTD patterns, it is then necessary to first monitor the evolution of the nude Si substrate within the temperature range 25–1000 ◦ C. Fig. 3 shows the XRTD patterns collected at 25 ◦ C prior to heating, in situ at 1000 ◦ C, and again at 25 ◦ C after the corresponding cooling down, from where it is evident that the Si substrate undergoes no changes and that the only effect of the heating cycle is the expected peak shifting due to the thermal expansion (during heating)/contraction (during cooling down). Fig. 4A–G shows now the two-dimensional intensity contours generated from the indexed XRTD patterns collected as a function of temperature for the TiO2 thin films without and with Ln3+ doping. Clearly, the thin films are formed all by anatase phase, which is the most desired phase for photo-induced applications. Initially, the only visible effect in the XRTD patterns as the temperature increases is that the 1 0 1 peak of anatase increasingly shifts towards lower diffraction angles, which reflects the corresponding thermal expansion of the unit cell. Eventually, a certain temperature is reached at which the anatase phase starts to transform into the rutile phase. The temperature for the onset of the anatase-to-rutile phase transformation depends however on the type and relative concentration of Ln3+ cations used to dope the TiO2 , and is listed in Table 2. It is evident that the Ln3+ doping increasingly shifts the phase transformation towards higher temperatures, and that this shifting is, if any, only a little more pronounced for Eu3+ than for Er3+ . These effects have been observed in earlier studies on xerogels obtained by drying the undiluted sols at ambient conditions,8,9 and are due to the formation of solid solutions with Ln3+ interstitial solutes in the TiO2 host. Note that these solid solutions are more stable with temperature than pure TiO2 due to their internal lattice stresses,

4461

obstructed Ti4+ and O2− diffusion, and hindered migration of the crystallite boundary.8,9 It can nevertheless be seen in Table 2 that a notable difference is that the thin films have a much greater thermal stability than the xerogels, as the temperatures for the onset of phase transformation are much higher and there is no precipitation of titanates up to 1000 ◦ C. The effect of the type of Ln3+ doping (Eu3+ or Er3+ ) is however less marked. These observations are reasonable considering that the lower the grain growth during the heat-treatment the greater the thermal stability of the anatase phase, and that the driving force for diffusion in the nanocrystalline thin films is lower than in the xerogels (i.e., constrained sintering with two-dimensional grains).17 4. Concluding remarks Both undoped and Ln3+ -doped TiO2 nanocrystalline thin films (Ln = Er or Eu; contents of 1, 2, or 3 mol.%) of anatase phase have been prepared by the dip-coating method from nanoparticulate sols in turn synthesized by an aqueous colloidal sol–gel route. It has been observed that dilution of the as-synthesized sols in ethanol is a key step for the preparation of homogenous thin films. It has also been observed that, regardless of the type and amount of Ln3+ doping, the resulting thin films are formed by nanocrystalline anatase, which is the desired TiO2 phase for photo-induced applications. In addition, these anatase nanocrystalline thin films are highly stable thermally, by exhibiting greater temperatures for the onset of anatase-to-rutile phase transformation than their xerogel counterparts and inhibiting the precipitation of titanates (at least up to 1000 ◦ C). Finally, it has also been observed that the thermal stability of the thin films increases with increasing the content of Ln3+ doping, with the Eu3+ solute being, if any, slightly more effective than the Er3+ solute. Acknowledgements This work was supported by the Ministerio de Economía y Competitividad (Government of Spain) under Grant no. MAT2012-31090 and MAT 2010-16848. Dr. Mario Borlaf thanks Consejo Superior de Investigaciones Científicas for his PhD Grant no. JAE-Pre 083. References 1. Arconada N, Castro Y, Durán A, Héquet V. Photocatalytic oxidation of methyl ethyl ketones over sol–gel mesoporous and meso-structured TiO2 films obtained by EISA method. App Catal B: Environ 2011;107:52–8. 2. Leroy CM, Cardinal T, Jubera V, Treguer-Delapierre M, Majimel J, Manaud JP, Backov R, Boissière C, Grosso D, Sanchez C, Viana B, Pellé F. Europium-doped mesoporous titania thin films: rare-earth locations and emission fluctuations under illumination. ChemPhysChem 2008;9:2077–84. 3. Colomer MT. Nanoporous anatase thin films as fast-proton-conducting materials. Adv Mater 2006;18:371–4. 4. Shen Y, Tao J, Gu F, Huang L, Bao J, Zhang J, Dai N. Preparation and photoelectric properties of ordered mesoporous titania thin films. J Alloys Compd 2009;474:326–9. 5. Borlaf M, Poveda JM, Moreno R, Colomer MT. Synthesis and characterization of TiO2 /Rh3+ nanoparticulate sols, xerogels and cryogels for photocatalytic applications. J Sol-Gel Sci Technol 2012;63:408–15.

4462

M. Borlaf et al. / Journal of the European Ceramic Society 34 (2014) 4457–4462

6. Colomer MT. Nanoporous anatase ceramic membranes as fast-protonconducting materials. J Eur Ceram Soc 2006;26:1231–6. 7. Borlaf M, Colomer MT, Cabello F, Serna R, Moreno R. Electrophoretic deposition of TiO2 /Er3+ nanoparticulate sols. J Phys Chem B 2013;117:1556–62. 8. Borlaf M, Moreno R, Ortiz AL, Colomer MT. Synthesis and photocatalytic activity of Eu3+ -doped nanoparticulate TiO2 sols and thermal stability of the resulting xerogels. Mater Chem Phys 2014;144:8–16. 9. Borlaf M, Colomer MT, Moreno R, Ortiz AL. Effect of Er3+ doping on the thermal stability of TiO2 nanoparticulate xerogels. J Nanopart Res 2013;15:1752 (10 pp). 10. Henrist C, Dewalque J, Mathis F, Cloots R. Control of the porosity of anatase thin films prepared by EISA: influence of thickness and heat treatment. Microporous Mesoporous Mater 2009;117:292–6. 11. Crepaldi EL, Soler-Illia GJdAA, Grosso D, Cagnol F, Ribot F, Sanchez C. Controlled formation of highly organized mesoporous titania thin films: from mesostructured hybrids to mesoporous nanoanatase TiO2 . J Am Chem Soc 2003;125:9770–86.

12. Chen W, Sun X, Cai Q, Weng D, Li H. Facile synthesis of thick ordered mesoporous TiO2 film for dye-sensitized solar cell use. Electrochem Commun 2007;9:382–5. 13. Yang P, Zhao D, Margolese DI, Chmelka BF, Stucky GD. Generalized syntheses of large-pore mesoporous metal oxides with semicrystalline frameworks. Nature 1998;396:152–5. 14. Nicole L, Boissière C, Grosso D, Quach A, Sanchez C. Mesostructured hybrid organic–inorganic thin films. J Mater Chem 2005;15: 3598–627. 15. Colomer MT, Guzmán J, Moreno R. Peptization of nanoparticulate titania sols prepared inder different water–alkoxide molar ratios. J Am Ceram Soc 2010;93:59–64. 16. Wohlfarth Ch, Wohlfarth B. Surface tension of pure liquids and binary liquid mixtures. Heidelberg (Germany): Springer-Verlag; 1997. 17. Kang S-JL. Sintering: densification, grain growth and microstructure. Oxford: Elsevier; 2005.