pubs.acs.org/Langmuir © 2009 American Chemical Society
Unravelling the Chemical Morphology of a Mesoporous Titanium Dioxide Interface by Confocal Raman Microscopy: New Clues for Improving the Efficiency of Dye Solar Cells and Photocatalysts Andre L.A. Parussulo, Juliano A. Bonacin, Sergio H. Toma, Koiti Araki,* and Henrique E. Toma Instituto de Quı´mica, Universidade de S~ ao Paulo, Avenida Prof. Lineu Prestes 748, C. Postal 26077, CEP: 05508-000, Sao Paulo, SP, Brazil Received July 13, 2009. Revised Manuscript Received August 20, 2009 The presence of anatase and rutile domains on nanocrystalline films of P25 TiO2, as well as the distinct coordination modes of carboxylates on those phases, were revealed by confocal Raman microscopy, a technique that showed to be suitable for imaging the chemical morphology down to submicrometric size.
The advancement of nanoscience and nanotechnology is leading to the development of increasingly intricate systems, very often, exhibiting unpredictable combinations of two or more materials or polymorphic phases.1,2 This is particularly critical in the cases where inorganic/organic hybrid interfaces are involved, turning into a real challenge the evaluation and understanding of their detailed properties, which ultimately depend on the chemical composition, chemical morphology, and interfaces. However, techniques such as scanning electron microscopy (SEM), energy-dispersive spectroscopy (EDS), and transmission electron microscopy (TEM), which are nicely suited for inorganic solids, become rather limited in the case of molecular or hybrid materials.3,4 Titanium dioxide is an abundant, nontoxic and biocompatible wide band gap semiconductor, generally found as rutile and anatase polymorphs. One of the most important commercial sources is P25 titanium dioxide, supplied by Degussa in the nanocrystalline form. It is a mixture of anatase (∼80%) and rutile (∼20%) and has been used in many scientific and technological applications, particularly in sensors, photocatalysis, and photoelectrochemical cells.5-8 In the last two areas, the results obtained with Degussa P25 have been considered as a reference. However, there is not a consensus about how those phases are distributed in the colloidal dispersion and mesoporous films. In 1991, Bicley et al.9 proposed that the anatase nanocrystals are covered by a layer of rutile. Soon after, Datye et al. (1995),10 found out that much bigger individual rutile nanoparticles are in
fact dispersed in smaller anatase nanocrystals. More recently (2001), Ohno et al.11 showed that rutile and anatase nanocrystals are present as agglomerates in P25 powder, which was proposed to disassemble during typical processing conditions. On the other hand, in recent years, many attempts have been made to push the overall efficiency of TiO2-based photoelectrochemical cells to levels higher than the 12% reported by O’Regan and Gr€atzel12 in 1993. In this regard, among the several factors influencing the efficiency of those devices, some are intrinsically related to the mesoporous TiO2 interface. For most applications, not only the crystalline phase but the average grain size and phase distribution, the pore size, and the connections between the grains are quite relevant. The binding properties and electronic coupling of the molecular species adsorbed on the mesoporous surface are also very important.13-16 Thus, in this communication we report, for the first time, the direct imaging of the chemical morphology of a mesoporous titanium dioxide interface prepared with Degussa P25 TiO2. The presence of relatively large rutile agglomerates dispersed in anatase, as well as the distinct coordination modes of adsorbed carboxylated species on those polymorphs, were clearly evidenced. Confocal Raman microscopy,17,18 a technique combining the chemical analysis capability of Raman spectroscopy and the enhanced optical resolution of confocal microscopy (xy = 400 nm and z = 1000 nm), was used for that purpose. Mesoporous films of TiO2 (4-8 μm thick)19 were prepared by dispersing a colloidal paste on glass coverslips (for confocal
*Corresponding author. E-mail:
[email protected].
(1) Toma, H. E.; Araki, K. Prog. Inorg. Chem. 2009, 56, 379–485. (2) Descalzo, A.; - Man~ez, B. R. M.; Sancenon, F.; Hoffmann, K.; Rurack, K. Angew. Chem., Int. Ed. 2006, 45, 5924–5948. (3) Burnside, S. D.; Shklover, V.; Barde, C.; Comte, P.; Arendse, F.; Brooks, K.; Gratzel, M. Chem. Mater. 1998, 10, 2419–2425. (4) Park, J.; Joo, J.; Kwon, S. G.; Jang, Y.; Hyeon, T. Angew. Chem., Int. Ed. 2007, 46, 4630–4660. (5) Furtado, L. F. O.; Alexiou, A. D. P.; Gonc-alves, L.; Toma, H. E.; Araki, K. Angew. Chem., Int. Ed. 2006, 45, 3143–3146. (6) Kolmakov, A.; Moskovits, M. Annu. Rev. Mater. Res. 2004, 34, 151–180. (7) Linsebigler, A. L.; Lu, G.; Yates, J. T. Chem. Rev. 1995, 95, 735–758. (8) Kroon, J. M.; Bakker, N. J.; Smit, H. J. P.; Liska, P.; Thampi, K. R.; Wang, P.; Zakeeruddin, S. M.; Gratzel, M.; Hinsch, A.; Hore, S.; Wurfel, U.; Sastrawan, R.; Durrant, J. R.; Palomares, E.; Pettersson, H.; Gruszecki, T.; Walter, J.; Skupien, K.; Tulloch, G. E. Prog. Photovoltaics 2007, 15, 1–18. (9) Bickley, R. I.; Gonzalezcarreno, T.; Lees, J. S.; Palmisano, L.; Tilley, R. J. D. J. Solid State Chem. 1991, 92, 178–190. (10) Datye, A. K.; Riegel, G.; Bolton, J. R.; Huang, M.; Prairie, M. R. J. Solid State Chem. 1995, 92, 236–239.
Langmuir 2009, 25(19), 11269–11271
(11) Ohno, T.; Sarukawa, K; Tokieda, K.; Matsumura, M. J. Catal. 2001, 203, 82–86. (12) Nazeeruddin, M. K.; Kay, A.; Rodicio, I.; -Baker, R. H.; M€uller, E.; Liska, P.; Vlachopoulos, N.; Gr€atzel, M. J. Am. Chem. Soc. 1993, 115, 6382–6390. (13) Tang, H.; Prasad, K.; Sanjines, R.; Schmid, P. E.; Levy, F. J. Appl. Phys. 1994, 75, 2042–2047. (14) Frank, A. J.; Kopidakis, N.; van de Lagemaat, J. Coord. Chem. Rev. 2004, 248, 1165–1179. (15) Wang, G.; Wu, F.; Zhang, X.; Luo, M.; Deng, N. J. Photochem. Photobiol. A 2006, 179, 49–56. (16) Schlichthorl, G.; Huang, S. Y.; Sprague, J.; Frank, A. J. Phys. Chem. B 1997, 101, 8141–8155. (17) Donev, E. U.; Lopez, R.; Feldman, L. C.; Haglund, R. F., Jr. Nano Lett. 2009, 9, 702–706. (18) Robinson, J. A.; Puls, C. P.; Staley, N. E.; Stitt, J. P.; Fanton, M. A.; Emtsev, K. V.; Seyller, T.; Liu, Y. Nano Lett. 2009, 9, 964–968. (19) Nogueira, A. F.; Formiga, A. L. B.; Winnischofer, H.; Nakamura, M.; Engelmann, F. M.; Araki, K.; Toma, H. E. Photochem. Photobiol. Sci. 2004, 3, 56–62.
Published on Web 08/31/2009
DOI: 10.1021/la902551k
11269
Letter
Parussulo et al.
Figure 1. SEM images showing the morphology of mesoporous P25 film fired at 100 (left) and 450 °C (right), for 30 min.
Figure 3. Confocal Raman images showing the CM-βCD distribution on a rutile/anatase nanocrystalline film obtained simultaneously with the images shown in the Figure 2, but monitoring the νas(COO) bands at 1652 (A) and 1605 cm-1 (B). An image (generated using the 398 cm-1 band) showing a larger area (40 40 μm, 50 50 points) of the Figure 2A sample is shown in C. The Raman spectra of the molecular probe in the 1400 to 1900 cm-1 region in (a) rutile and (b) anatase rich domains are shown in D.
Figure 2. Confocal Raman Images showing the rutile/anatase distribution on a nanocrystalline P25 film fired at 450 °C, 10 10 μm wide, 75 75 points, monitoring the intensity of the (A) 398 cm-1 and (B) 448 cm-1 bands. Raman spectra of the (a) rutile-rich and (b) anatase-rich domains are shown in C. The plot of Raman scattering intensity at (a) 398 and (b) 448 cm-1 bands, along the cross-section lines in A and B, are shown in D.
Raman microscopy) or fluorine-doped tin oxide (FTO) glass (for SEM) using the doctor blade method, then allowing them to dry at room temperature and firing at 100-450 °C, for 30 min. The samples modified with CM-βCD were prepared by soaking those TiO2 films overnight into a 1.0 10-4 mol 3 dm-3 methanol solution, washing with plenty of ethanol, and drying at room temperature in a desiccator, under vacuum. The paste was prepared by dispersing 6 g of nanocrystalline TiO2 (P25 from Degussa) in 2 mL of deionized (DI) water and 0.2 mL of acetylacetone (Sigma-Aldrich) in a mortar, for 40 min. Then, 0.1 mL of Triton-X100 (Aldrich) and 8.0 mL of DI water were added. Confocal Raman microscopy images were recorded on a WITec Alpha 300R microscope, equipped with a 532 nm solid state laser (100 mW) and a 100 objective (N.A. = 0.8). SEM images were recorded on a Jeol JSM-7401F field emission scanning electron microscope. Films of P25 titanium dioxide appear as a network of interconnected 20-70 nm grains, in the SEM image, as shown in Figure 1. No significant changes in the grain size and interconnections between the grains were observed when the samples were fired at higher temperatures, for example, 450 instead of 100 °C, for 30 min. The two polymorphs constituting the film can be easily differentiated by the characteristic Raman bands of anatase (Figure 2C) at 144 (Eg), 197 (Eg), 398 (B1g), 519 (A1g, B1g) and 11270 DOI: 10.1021/la902551k
639 (Eg) cm-1, and of rutile at 448 (Eg) and 612 (A1g) cm-1. Those bands are slightly shifted20,21 from those in the pure polymorph samples. Contrasting Raman images of mesoporous nanocrystalline TiO2 films were obtained by plotting the intensity of typical rutile Raman peaks, e.g., at 448 cm-1 (a), and anatase peaks, e.g., at 398 cm-1 (b), as a function of the xy position, thus revealing their actual distribution in the film, as shown in Figure 2A,B. One can see that a small amount of rutile (Figure 2C, spectrum b) is dispersed in the anatase phase (light areas). However, a significant fraction of rutile nanoparticles is present as agglomerates in relatively large rutile domains (up to 2 μm wide and ∼55% rutile), spanning a significant fraction of the analyzed areas (black spots in Figure 2A, light spots in 2B). The complementarity of those images is remarkable, allowing the unequivocal discrimination of rutile and anatase domains. This contrasting chemical morphology is also evidenced by the cross section analysis plot, shown in Figure 2D. Similar results were obtained for larger areas as shown in Figure 3C, using the 398 cm-1 band. This is a rather unexpected result since, according to Ohno et al.,11 rutile and anatase nanocrystals were supposed to be separated and homogeneously distributed in colloidal dispersions used for the preparation of nanocrystalline films. On the other hand, the thermal treatment up to 450 °C did not promote significant changes in the film morphology or in the relative phase distribution. Since none or insignificant mass transport process and phase transformation seem to be taking place during the firing process, one can infer that the aggregates observed in the nanocrystalline films should be present in the colloidal paste. This hypothesis was confirmed by examining samples fired at 100 and 200 °C (Supporting Information, Figure S2). Thus, the aggregates of rutile nanocrystals (20) Ohsaka, T.; Izumi, F.; Fujiki, Y. J. Raman Spectrosc. 1978, 7, 321–324. (21) Mazza, T.; Barborini, E.; Piseri, P.; Milani, P.; Cattaneo, D.; Bassi, A. L.; Bottani, C. E.; Ducati, C. Phys. Rev. B 2007, 75, 45416.
Langmuir 2009, 25(19), 11269–11271
Parussulo et al.
present in P25 powder can not be disassembled as easily as previously reported. Another relevant issue is the distribution and binding mode of molecular species on rutile and anatase nanocrystals present in mesoporous TiO2 films. For this purpose, carboxymethylβ-cyclodextrin (CM-βCD) was used as a Raman probe. That molecular host species was previously anchored onto TiO2 through its carboxylate groups, generating a supramolecular UV dosimeter,22 which mimics the skin response for erytheme formation. Comparison of Raman spectra of CM-βCD adsorbed on rutile and anatase domains evidenced the shift of the νas(COO) stretching mode respectively from 1605 to 1652 cm-1. In this way, images distinguishing the two polymorphs domains can be generated using those bands. The image obtained using the νas(COO) peak at 1652 cm-1 (Figure 3A) is analogous to that shown in Figure 2A. This may suggest that the probe molecule is preferentially bonded to anatase nanoparticles. However, a complementary image, matching with that shown in Figure 2B, came out when the 1605 cm-1 peak was used for imaging (Figure 3B). In fact, the intensity of the νas(COO) peak in the rutile was shown to be similar to that in anatase domains (Figure 3D). Therefore, from the Raman confocal imaging, the amount of CM-βCD bound to rutile and anatase nanocrystals should be similar, but involving different carboxylate coordination modes responsible for the shift in the vibrational frequency. As a matter of fact, it is known that the carboxylate asymmetric stretching frequency decreases when the coordination changes from unidentate to bidentate mode. Therefore, we propose that CM-βCD is bound unidentate on anatase and bidentate on rutile.23 This finding, in addition to the similarity of the ν(COO) band intensities (Figure 3D), seems to deny the lower adsorption capability previously reported in the literature for rutile, in spite of its theoretically larger number of binding sites.24 (22) Toma, S. H.; Bonacin, J. A.; Araki, K.; Toma, H. E. Surf. Sci. 2006, 600, 4591–4597. (23) Deacon, G. B.; Phillips, R. J. Coord. Chem. Rev. 1980, 33, 227–250. (24) Fahmi, A.; Minot, C.; Fourre, P.; Nortier, P. Surf. Sci. 1995, 343, 261–272. (25) Park, N.-G.; de Lagemaat, J.; van Frank, A. J. J. Phys. Chem. B 2000, 104, 8989–8994. (26) Hurum, D. C.; Agrios, A. G.; Gray, K. A.; Rajh, T.; Thurnauer, M. C. J. Phys. Chem. B 2003, 107, 4545–4549.
Langmuir 2009, 25(19), 11269–11271
Letter
Our finding corroborates the proposal of Park et al.25 that the 30% lower photoconversion efficiency observed for rutile-based solar cells, as compared to anatase-based ones, is a consequence of a proportionally smaller surface area of the mesoporous interface, rather than a lower chemical affinity. Interestingly, the presence of rutile in the anatase phase has also been reported to improve the activity of mesoporous TiO2 films as photocatalysts.9 In this case, rapid electron and hole transfer processes seem to be taking place at the anatase/rutile interface, decreasing the recombination and generation of photocatalytic “hot spots” in the mixed phase materials.26 However, the photocatalytic activity of a 20:80 mixture of rutile and anatase nanocrystals exhibited lower activity compared with P25, in spite of the analogous composition.11 Accordingly, the presence of larger rutile agglomerates may be contributing in someway to increase the photoinduced charge-separation process and the amount of hot spots. In conclusion, mesoporous films of P25 TiO2 exhibit distinct anatase-rich and rutile-rich domains, consistent with the presence of relatively large agglomerates of small nanocrystals of rutile polymorph dispersed in the complementary anatase phase. This knowledge on the chemical morphology of nanocrystalline P25 TiO2 interfaces should provide new insights to be explored in the area of photocatalysis and dye-sensitized solar cells. Acknowledgment. This research was supported by Fundac-~ao de Amparo a Pesquisa do Estado de S~ao Paulo (FAPESP), Conselho Nacional de Desenvolvimento Cientı´ fico e Tecnologico (CNPq), Instituto do Mil^enio de Materiais Complexos (IMMC), and Petrobras. We also thank Degussa for the kind supply of P25 TiO2. Supporting Information Available: Optical images of TiO2 fired at 100 and 450 °C, 100 objective, are shown Figure S1. Confocal Raman microscopy images (sample area = 50 50 μm2, resolution = 65 65 points) showing the rutile/anatase distribution in nanocrystalline P25 films fired at 100 and 200 °C, obtained by monitoring the 398 cm-1 and 448 cm-1 bands, are presented in Figure S2. This material is available free of charge via the Internet at http://pubs.acs.org.
DOI: 10.1021/la902551k
11271
