pubs.acs.org/Langmuir © 2009 American Chemical Society
Highly Quasi-Monodisperse Ag Nanoparticles on Titania Nanotubes by Impregnative Aqueous Ion Exchange J. A. Toledo-Antonio,*,† M. A. Cortes-Jacome,† C. Angeles-Chavez,† E. Lopez-Salinas,† and P. Quintana‡ †
Molecular Engineering Program, Instituto Mexicano del Petr oleo, Eje Central L azaro C ardenas 152, 07730 M exico, D.F., M exico, and ‡CINVESTAV-M erida, Depto. de Fı´sica Aplicada Km 6 Ant Carr. a Progreso, C.P. 97310, Cordemex, M erida, Yuc, M exico Received March 19, 2009. Revised Manuscript Received May 8, 2009
Silver nanoparticles were homogenously dispersed on titania nanotubes (NT), which were prepared by alkali hydrothermal methodology and dried at 373 K. Ag+ incorporation was done by impregnative ion exchange of aqueous silver nitrate onto NT. First, Ag+ ions incorporate into the layers of nanotube walls, and then, upon heat treatment under N2 at 573 and 673 K, they migrate and change into Ag2O and Ag0 nanoparticles, respectively. In both cases, Ag nanoparticles are highly dispersed, decorating the nanotubes in a polka-dot pattern. The Ag particle size distribution is very narrow, being ca. 4 ( 2 nm without any observable agglomeration. The reduction of Ag2O into Ag0 octahedral nanoparticles occurs spontaneously and topotactically when annealing, without the aid of any reducing agent. The population of Ag0 nanoparticles can be controlled by adjusting the annealing temperature. An electron charge transfer from NT support to Ag0 nanoparticles, because of a strong interaction, is responsible for considerable visible light absorption in Ag0 nanoparticles supported on NT.
Introduction Nanotechnology is devoted to designing novel engineered material structures in the range of 1-100 nm in order to modify their mechanical, optical, magnetic, and electronic properties as well as their chemical reactivity, which in turn leads to surprising and unpredictable effects.1,2 Nobel metal nanoparticles interacting with a semiconductor nanostructured material undergo charge equilibration under photoexcitation, and the composite Fermi level shifts closer to the conduction band of the semiconductor,3-6 improving the adsorptive, catalytic, and electronic efficiency of the composite system.7 Among such nanocomposite structures, the Ag/TiO2 system has attracted the interest of researchers for biological applications due to the antibacterial activity of both components.8-10 On one hand, TiO2 is the most studied semiconductor for environmental cleanup applications because of its unique ability to photooxidize organic compounds and microorganisms in wastewater and air and because of its chemical inertness, nontoxicity, and biological compatibility. However, silver is probably the most powerful antimicrobial that exhibits strong cytotoxicity toward a broad *
[email protected]. (1) Renn and, O.; Roco, M. C. J. Nanopart. Res. 2006, 8, 153. (2) Siegel, R. W., Hu, E., Roco, M. C., Eds.; Nanostructure Science and Technology; Springer: Dordrecht, Netherlands, 1999. Available at http://www. wtec.org/loyola/nano/. (3) Subramanian, V.; Wolf and, E. E; Kamat, P. V. J. Phys. Chem. B 2003, 107, 7479. (4) Wood, A.; Giersig, M.; Mulvaney, P. J. Phys. Chem. B 2001, 105, 8810. (5) Jakob, M.; Levanon, H.; Kamat, P. V. Nano Lett. 2003, 3, 353. (6) Subramanian, V.; Wolf, E. E; Kamat, P. V. J. Am. Chem. Soc. 2004, 126, 4943. (7) Hirakawa, T.; Kamat, P. V. J. Am. Chem. Soc. 2005, 127, 3928. (8) Ni~no-Martinez, N.; Martinez-Cata~non, G. A.; Aragon-Pi~na, A.; MartinezGutierrez, F.; Martinez-Mendoza, J. R.; Ruiz, F. Nanotechnology 2008, 19, 065711 (9) Liu, Y.; Wang, X.; Yang and, F.; Yang, X. Microporous Mesoporous Mater. 2008, 114, 431. (10) Shah, M. S. A. S.; Nag, M.; Kalagara, T.; Singh, S.; Manorama, S. V. Chem. Mater. 2008, 20, 2455.
Langmuir 2009, 25(17), 10195–10201
range of microorganisms and remarkable low human toxicity compared to other heavy metal ions.11-13 Moreover, silver ions are capable of causing a bacteriostatic (grown inhibitions) or even bactericide (antibacterial) impact;14-16 therefore, it is not surprising to find silver nanoparticles in multiple medical applications. With a view toward controlling particle size and dispersion, various methods have been applied to deposit Ag nanoparticles onto TiO2. The most used bulk-solution method for obtaining metal nanoparticles is the chemical reduction of metallic salts. The synthesis of metal particles usually involves a soluble metal salt, a reducing agent, and a stabilizing agent. Reducing agents such as sodium boron hydride and alcohols are commonly used for the preparation of metal nanoparticles. A stabilizing additive, such as polymers and organic molecules, cap the particle and hinder further growth or aggregation. For instance, in a colloidal approach, organic-capped anatase TiO2 was synthesized by the hydrolysis of titanium tetra-isopropoxide using oleic acid as a surfactant, and after yielding TiO2 solids, they were admixed with a very dilute solution of AgNO3 in CHCl3/EtOH.17 In another approach, the addition of citrate ions, as a reducing and as a complexing agent, to obtain silver colloids (Ag-capped SiO2) produced particles of 50-100 nm of varying shape and size.18 Ag nanoparticles (10-80 nm) were synthesized by aqueous adsorption at pH 7 on the surface of commercial TiO2 using aqueous NaBH4 reduction at pH 108. Recently, in a sol-gel (11) Lok, C. N.; Ho, C. M.; Chen, R.; He, Q. Y.; Yu, W. Y.; Sun, H.; TAM, P. K. H.; Chiu, J. F.; Che, C. M. J. Proteome Res. 2006, 5, 916. (12) Gogoi, S. K.; Gopinath, P.; Paul, A.; Ramesh, A.; Ghosh, S. S.; Chattopadhyay, A. Langmuir 2006, 22, 9322. (13) Kong, H.; Jang, J. Langmuir 2008, 24, 2051. (14) Feng, Q. L.; Wu, J.; Chen, G. Q.; Cui, F. Z.; Kim, T. N.; Kim, J. O. J. Biomed. Mater. Res. 2000, 52, 662. (15) Batarseh, K. I. J. Antimicrob. Chemother. 2004, 54, 546. (16) Morones, J. R.; Elechiguerra, J. L.; Camacho, A.; Holt, K.; Kouri, J. B.; Ramirez, J. T.; Yacaman, M. J. Nanotechnology 2005, 16, 2346. (17) Cozzoli, P. D.; Comparelli, R.; Fanizza, E.; Curri, M. L.; Agostiano, A.; Baub, D J. Am. Chem. Soc. 2004, 126, 3868. (18) Pillai, Z. S.; Kamat, P. V. J. Phys. Chem. B 2004, 108, 945.
Published on Web 06/01/2009
DOI: 10.1021/la9009702
10195
Article
Toledo-Antonio et al.
method, AgNO3 was admixed with tetra-n-butyl titanate to yield, after sol formation, Ag/TiO2 powders that were then hydrothermally treated with 5-10 M NaOH to obtain 2-5 nm Ag particles on TiO2 nanotubes or nanosheets.19 In spite of these advanced methods, the implementation of a simple process to produce Ag nanoparticles without the use of organic stabilizers and/or chemical reducing agents would be a significant issue in larger-scale preparations. Also, it is known that both the photoactivity of TiO2 and the plasmon resonance effect of Ag nanoparticles are size- and shapedependent.18,20 Then, by tailoring the crystallite size and morphology of TiO2 and Ag particles, the specific surface area (SSA) available for chemical reactions and surface interactions is increased, and the antimicrobial performance is expected to be more efficient. Accordingly, the conversion of the 3D anatase structure into a 1D structure, such as nanotubes,21 significantly changes the photocatalytic properties and considerably increases the SSA. TiO2 has been converted into titania nanotubes or nanofibers through a relatively simple alkaline hydrothermal method,22,23 representing a suitable methodology for rapid manufacturing scale up. Nanotubes exhibit large internal and external surfaces, along with a surface in their vertex and a surface in the interlayer regions that make up the nanotube walls.24,25 The transformation of TiO2 into nanotubes yields materials with an SSA as large as ca. 400 m2/g.26 The dispersion of novel metallic nanoparticles on this support may represent advantages on the multiple and potential applications of these nanocomposites. In fact, Pt, Ru, Au, and Pd have been homogeneously distributed on the surface of titania nanotubes showing considerable improvement in their catalytic performance.27-30 Given the cation-exchange capabilities of TiO2 nanotubes, in this work we aim to obtain nanoparticles by first ion exchanging Ag+ cations into the layered walls of the nanotubes and then by annealing at the indicated temperatures, eventually generating Ag nanoparticles. The size, shape, location, mobility, and reduction state of Ag nanoparticles are examined in detail.
Experimental Section Synthesis of Nanotubular Hydrous Titania and Ag-Impregnative Ion Exchange. Nanotubular hydrous titania (NT) was synthesized by a hydrothermal method as described elsewhere.26 A portion of NT powder was placed in contact with 1 N AgNO3 solution for 1 h. Then, the sample was filtered and dried at 373 K without washing. In typical ion exchange methods, a final aqueous washing step is applied before drying in order to eliminate excess unexchanged ions. In this case, an impregnative ion-exchange method was used, and under these conditions, NT
(19) Lai, Y.; Chen, Y.; Zhuang, H.; Lin, C. Mater. Lett. 2008, 62, 3688. (20) Murakami, Y.; Matsumoto, J.; Takasu, Y. J. Phys. Chem. B 1999, 103, 1836. (21) Rao, C. N. H.; Nath, M. Dalton Trans. 2003, 1. (22) Kasuga, T.; Hiramatsu, M.; Hoson, A.; Sekino, T.; Niihara, K. Langmuir 1998, 14, 3160. (23) Du, G. H.; Chen, Q.; Che, R. C.; Yuan, Z. Y.; Peng, L. M. Appl. Phys. Lett. 2001, 79, 3702. (24) Tenne, R. Nature 2004, 431, 640. (25) Tenne, R. Angew. Chem., Int. Ed. 2003, 42, 5124. (26) Toledo-Antonio, J. A.; Capula, S.; Cortes-Jacome, M. A.; Angeles-Chavez, C.; Lopez-Salinas, E.; Ferrat, G.; Navarrete, J.; Escobar, J. J. Phys. Chem. C 2007, 111, 10799. (27) Bavykin, D. V.; Lapkin, A. A.; Plucinski, P. K.; Friedrich, J. M.; Walsh, F. C. J. Catal. 2005, 235, 10. (28) Bavykin, D. V.; Lapkin, A. A.; Plucinski, P. K.; Torrente-Murciano, L.; Friedrich, J. M.; Walsh, F. C. Topics Catal. 2006, 39, 151. (29) Torrente-Murciano, L.; Lapkin, A. A.; Bavykin, D. V.; Walsh, F. C.; Wilson, K J. Catal. 2007, 245, 272. (30) Yu, K. P.; Yu, W. Y.; Kuo, M. C.; Liou, Y. C.; Chien, S. H.; Appl. Catal., B 2008, 84, 112.
10196 DOI: 10.1021/la9009702
was always exposed to very acidic pH values, similar to that at which it was formed. Two fractions of the sample were annealed at 573 and 673 K in flowing N2 for 4 h. From now on, the samples will be referred as Ag-NT-x, where Ag indicates that it contains silver, NT stands for nanotubes, and x is the annealing temperature. Characterization. Transmission electron microscopy (TEM) and scanning transmission electron microscopy (STEM) were both performed on a JEM-2200FS microscope with an accelerating voltage of 200 kV. The microscope is equipped with a Schottky-type field emission gun and an ultrahigh-resolution configuration (Cs=0.5 mm; Cc=1.1 mm; point-to-point resolution= 0.19 nm) and an on-column omega-type energy filter. Dark field images highlighting silver nanoparticles were obtained by using a high-angle annular dark field (HAADF) detector. The particle size was determined directly on images by counting and measuring 152 and 160 particles for Ag-NT-573 and Ag-NT-673, respectively. X-ray photoelectron spectroscopy (XPS) spectra were recorded on a THERMO-VG scalab 250 spectrometer equipped with an Al KR X-ray source (1486.6 eV) and a hemispherical analyzer. The base pressure during the analysis was 10-9 Torr. The XPS analyses were performed in a static system on the samples annealed in situ under an inert atmosphere in the heating chamber. The experimental peaks were decomposed into components using a mixed Gaussian-Lorentzian function, a nonlinear squares fitting algorithm, and Shirley-type background subtraction by using XPS peak-fit software. The binding energies (BE) were referenced to the adventitious C (1s) carbon peak at 284.6 eV. The Ag/Ti atomic ratio was determined from the corresponding peak areas, after subtracting the inelastic background and correction of the peak areas by the corresponding sensitive factors. X-ray diffraction (XRD) patterns of the samples packed in a glass holder were recorded at room temperature in a Siemens D-5000 diffractometer with Cu KR radiation (λ = 1.5418 A˚) at 35 kV and 30 mA. The samples were measured in the 2θ range between 5 and 90°, with a 2θ step size of 0.02° and a step time of 12 s. A PerkinElmer lambda 900 UV-vis spectrometer was used to record the diffuse reflectance spectra (DRS). Samples were annealed in situ in a homemade cell equipped with a quartz window by flowing nitrogen during the thermal treatment from 373 to 773 K.
Results and Discussion The XRD pattern of the anatase phase (crystal size ∼8.0 nm) was used as a starting material as shown in Figure 1a. This anatase phase was transformed into TiO2 nanotubes (Figure 2a) by an alkaline hydrothermal treatment at 373 K, followed by an annealing treatment under nitrogen flow at 573 °C; the XRD pattern showed four broad peaks at 2θ = 11.9, 24.4, 28-30, and 48.5° (Figure 1b). These peaks have been assigned to the diffraction lines of titanates with monoclinic,31,32 orthorhombic,33,34 and lepidocrocite-type structure.35,36 A more detailed study of nanotubular structure transformation was done by Raman and HRTEM techniques and was reported elsewhere.26,37 The broad (31) Chen, Q.; Du, G. H.; Zhang, S.; Peng, L. M. Acta Crystallogr., Sect. B 2002, 58, 587. (32) Chen, Q.; Zhou, W. Z.; Du, G. H.; Peng, L. M. Adv. Mater. 2002, 14, 1208. (33) Yang, J.; Jin, Z.; Wang, X.; Li, W.; Zhang, J.; Zhang, S.; Guo, X.; Zhang, Z. J. Chem. Soc., Dalton Trans. 2003, 3898. (34) Zhang, M.; Jin, Z.; Zhang, J.; Guo, X.; Yang, J.; Li, W.; Wang, X.; Zhang, Z. J. Mol. Catal. A: Chem. 2004, 217, 20. (35) Ma, R.; Bando, Y.; Sasaki, T. Chem. Phys. Lett. 2003, 380, 577. (36) Ma, R.; Fukuda, K.; Sasaki, T.; Osada, M.; Bando, Y. J. Phys. Chem. B 2005, 109, 62. (37) Cortes-Jacome, M. A.; Ferrat, G.; Flores Ortiz, L. F.; Angeles-Chavez, C.; Lopez-Salinas, E.; Escobar, J.; Mosqueira, M. L.; Toledo-Antonio, J. A. Catal. Today. 2007, 126, 248.
Langmuir 2009, 25(17), 10195–10201
Toledo-Antonio et al.
Figure 1. XRD of anatase precursor, nanotubular TiO2 (NT) and Ag on NT at the indicated annealing temperature.
peak centered at 2θ ≈ 11.9° corresponds to the interlayer distance on the layered titanates,38 with dimensions around 0.74 nm in agreement with the interlayer space measured by TEM micrographs (Figure 2b). The intensity of the broad peak around 2830° has been related to Na+ ions remaining in the interlayer space of the layered structure of titanates,33 and it depends on the degree of replacement of Na+ by H+ during the acid treatment.39,40 When silver is incorporated (e.g., as in the Ag-NT-573 sample; Figure 1c), the diffraction peak at 11.9° almost vanishes whereas the intensity of the one at 28-30° increases. This suggests that Ag+ ion exchange occurs in the interlayer space of the nanotubular structure during the ion-exchange process and Ag incorporation causes a staking disorder on the layered structure of titanates, resulting in the peak disappearance at 2θ=11.9°. Additionally, a diffraction reflection around 38.1° suggests the presence of Ag nanoparticles (JCPDS-004-0783). In the case of the Ag-NT-673 sample, the XRD pattern in Figure 1d indicates that the nanotubular structure is transformed into anatase. Moreover, the (111), (200), and (220) diffraction peaks of face-centered Ag nanoparticles were more evident. However, a less intense broad line at 2θ=28-30° suggests that some layered structure remains, probably with the Ag+ ion in their interlayer position. Ag Particle Size and Distribution. The nanotubular features of the NT support, after annealing at 573 K, are shown in Figure 2a,b. The nanotubes have outer diameters of ca. 7 to 10 nm and inner diameters of 4.5 to 5.5 nm, with the walls being composed of two or three stacked layers with an interlayer space of 0.74 nm, as indicated in Figure 2b, exactly matching the interplanar distance described by the XRD peak at 2θ=11.9°. Ag samples were observed by applying the HAADF-STEM technique in order to highlight the difference produced by Ag, a heavy element in comparison with titanium and oxygen atoms, which are both lighter elements. The obtained result is displayed in HAADF-STEM images in Figure 2c,d, showing bundles of nanotubes dotted by a large number of very small white particles. Because the intensity in HAADF images depends on the atomic number, the small white particles are likely to be Ag nanoparticles with an average size of 4.5 nm. Additionally, the surface of the bundles appears to be decorated by very small white dots, which are likely associated with Ag+ ions remaining in strong (38) Clearfield, A.; Letho, J. J. Solid State Chem. 1988, 73, 98. (39) Tsai, C. C.; Teng, H. Chem. Mater. 2006, 18, 367. (40) Morgado, E. Jr.; De Abreu, M. A. S.; Moure, G. T.; Marickovic, B. A.; Jardim, P. M.; Araujo, A. S. Mater. Res. Bull. 2007, 42, 1748.
Langmuir 2009, 25(17), 10195–10201
Article
interaction with the support. The size of silver nanoparticles does not change after the annealing treatment at 673 K, as can be compared by an amplified HRTEM images in Figure 2e,f, where several well-faceted Ag nanoparticles appear to decorate the walls of the nanotubes. Noticeably, the spacing between the layers characteristic of the nanotubular structure in the support was not evident in Ag-containing samples, which is in agreement with the XRD results. Particle size distributions of silver in Ag-NT-573 (Figure 3a) have an average size of 4.3 nm, where 89% of the measurements were between 2 and 6 nm, whereas in Ag-NT-673 (Figure 3b), the distribution has an average size of 4.2 nm and 95% of measurements are between 3 and 6 nm. The silver bulk crystallite size determined by XRD by applying Scherrer’s formula gives somewhat higher values of 7.5 and 11.8 nm for Ag-NT-573 and Ag-NT-673, respectively. The difference in particle and/or crystallite size between the two techniques arises from the sample size (i.e., a local and statistical analysis by HAADF-STEM and a bulk larger analysis by XRD, where the later may be averaging larger particles not observed in HAADF-STEM). Ag Particle Image Simulation. A detailed crystallographic analysis in HRTEM images to determine the structure of the Ag nanoparticles deposited on NT was performed. Figure 4a shows an HRTEM image of a single 10 nm Ag particle of the sample, annealed at 573 K. The lattice d spacings were 0.267, 0.269, and 0.234 nm, corresponding to (1 1 1), (-1 -1 1), and (0 0 2) planes of the cubic Ag2O structure, respectively, according to JCPDS card number 41-1104. This nanocrystallite was found to be oriented along the [1 -1 0] direction, in relation to that of the electron beam. The nanoparticle showed a hexagonal shape exposing four {1 1 1} and two {0 0 2} planes. Therefore, Ag2O nanoparticles can be represented as a truncated octahedron with {1 1 1} and {0 0 2} truncated faces. For HRTEM image interpretation, image simulation of a real structural model is required.41 A model was obtained from inorganic crystal structure database number 35540, which correspond to a cubic Ag2O structure. The structure projected along the [1 -1 0] crystal direction, which is the same as determined in the HRTEM image, is displayed in Figure 4b. Using SimulaTem software,42 theoretically calculated images for several defocus settings with a crystal thickness of 6.7 nm were obtained. The calculated image is similar to the experimental image contrast features, shown in Figure 4c, with a defocus of -154 A˚. As can be observed under this condition, only white dots appeared in the calculated image, where Ag atoms show a strong contrast. Then, each white dot observed in the experimental image corresponds to a single atomic column of silver. From these results, a 3D shape of the nanoparticle can be built (Figure 4d). An HRTEM image of a single 5 nm nanoparticle of Ag-NT673 is displayed in Figure 5a. The crystallite shows interplanar distances of 0.233, 0.232, and 0.201 nm that correspond to (1 1 1), (-1 -1 1), and (0 0 2) planes of the cubic silver metallic phase (JCPDS 4-0783), oriented along the [1 -1 0] direction. The hexagonal shape and the {1 1 1} and {0 0 2} surface planes exposed in the Ag nanoparticle strongly suggest that the nanoparticle is an octahedral nanoparticle truncated in {0 0 2} planes. In this case, the structural model used to simulate the image was an fcc Ag lattice. The atomic model projected in the same direction determined in the experimental HRTEM image is illustrated in Figure 5b. The closest calculated image to the (41) Perez-Ramirez, J. G.; Perez, R.; Yacaman, M. J. Scripta Met. 1986, 20, 1523. (42) Gomez, A.; Beltran del Rio, L. M. Metal. Mater. 2001, 21, 46.
DOI: 10.1021/la9009702
10197
Article
Toledo-Antonio et al.
Figure 2. (a, b) TEM of NT annealed at 573 K. (c, d) HAADF-STEM images of Ag-NT-573 and Ag-NT-673. (e, f) HRTEM images of nanoparticles on NT of Ag-NT-573 and Ag-NT-673 samples.
experimental one was obtained at a -204 A˚ defocus with a crystal thickness of 4.8 nm, as shown in Figure 5c. Then, according to this model and the similarity between the image contrast features in the theoretical image and the experimental one, the white dots are atomic columns of Ag. The hexagonal shape of this nanoparticle is basically that of an octahedra with {1 1 1} surfaces and truncated in {0 0 2} planes, such as that shown in the 3D image in Figure 5d. Valence State of Surface Ag Nanoparticles. XPS measurements were carried out in order to examine any possible oxidation state variation in Ag nanoparticles as a function of an eventual thermal autoreduction. XPS of the Ag 3d region of samples AgNT-573 and Ag-NT-673 is shown in Figure 6. One or two doublets were necessary to fit each spectrum, yielding the XPS parameters, BE, and fwhm, as reported in Table 1. The BE values (43) Waterhouse, G. I. N.; Bowmaker, G. A.; Metson, J. B. Appl. Surf. Sci. 2001, 183, 191.
10198 DOI: 10.1021/la9009702
of the doublets were assigned to different oxidation states of silver nanoparticles.43 Accordingly, the BE values of the 3d5/2 signal reported for Ag0 nanoparticles, Ag+ in Ag2O, and Ag2+ in AgO are around 368.4, 367.8, and 367.4 eV, respectively.43,44 The Ag 3d spectrum of the sample annealed at 573 K, Figure 6a, was made up of a doublet with 3d5/2 and 3d3/2 signals at 367.8 and 373.9 eV, respectively, corresponding to Ag+. The second doublets used to fit the spectrum were 3d5/2 and 3d3/2 signals at 368.4 and 374.4 eV, which correspond to Ag0, indicating that after annealing at 573 K, silver ions were partially selfreduced, giving rise to a Ag0/(Ag0 + Ag+) ratio of 0.22, which means that only around 22% of total number of Ag ions were reduced to Ag0. The surface coverage by silver ions, being 0.17, was determined by the Ag/(Ag+Ti) ratio taking into account the whole Ag 3d and Ti 2p signals. On the Ag-NT-673 sample in Figure 6b, only one doublet with 3d5/2 and 3d3/2 signals at 368.1 (44) Zemlyanov, D. Y.; Nagy, A.; Schlogl, R. Appl. Surf. Sci. 1998, 133, 171.
Langmuir 2009, 25(17), 10195–10201
Toledo-Antonio et al.
Figure 3. Particle size distributions. (a) Ag-NT-573 and (b) AgNT-673.
Article
Figure 5. (a) Experimental HRTEM of a Ag-NT-673 nanoparticle with an inset FFT pattern. (b) Structural atomic model view along the [1 -1 0] direction. (c) Simulated HRTEM image focused at Δf = -204 A° and thickness of 4.8 nm; (d) Structural 3D atomic model.
Figure 6. XPS of the Ag 3d signal in (a) Ag-NT-573 and (b) AgNt-673 samples. Figure 4. (a) Experimental HRTEM of a Ag-NT-573 nanoparticle with an inset FFT pattern. (b) Structural atomic model view along the [1 -1 0] direction. (c) Simulated HRTEM image focused at Δf = -154 A˚ and a thickness of 6.7 nm. (d) Structural 3D atomic model.
and 374.1 eV was necessary to fit the spectrum corresponding to Ag0 nanoparticles, in agreement with the obtained XRD and HRTEM results. After annealing at 673 K, silver nanoparticles remain in a completely reduced state (e.g., Ag0). Therefore, silver nanoparticles remain mainly as Ag+ after autothermal reduction in a nitrogen atmosphere at 573 K, whereas after autoreduction at 673 K the silver ions were completely reduced to metallic Ag0. The surface coverage by silver atoms, Ag/(Ag + Ti) = 0.27, which is considerably higher than for the Ag-NT-573 sample, suggests a migration of Ag ions from the interlayer space to the outer Langmuir 2009, 25(17), 10195–10201
nanotube surface. In fact, these results can help to explain the difference in intensity observed in HAADF images of Figure 2c,d; that is, less-intense white nanoparticles observed in Ag-NT-573 than in Ag-NT-673 sample. In the first case, the nanoparticles consisted of Ag and O atoms (Figure 2c), and in the second case, the nanoparticles consisted mainly of Ag atoms (Figure 2d). The contrast in HAADF-STEM is obtained by differences in the electron current that is deflected to angles of 4 and 10° at different locations in the samples. Accordingly, the deflected current is proportional to the number of atoms in the beam, multiplied by an atom-specific cross section for electron scattering.45,46 Then, the intensity in the HAADF-STEM images (45) Treacy, M. M. J.; Rice, S. B. J. Microsci. (Oxford, U.K.) 1989, 156, 211. (46) Carlsson, A.; Puig-Molina, A.; Janssens, T. V. W. J. Phys. Chem. B 2006, 110, 5286.
DOI: 10.1021/la9009702
10199
Article
Toledo-Antonio et al. Table 1. XPS Parameters of Ag 3d and Ti 2p Signals of Samples
sample NT-573 Ag-NT-573 Ag-NT-673
binding energy (eV) 458.4 367.8 368.4 458.6 368.1 459.0
fwhm (eV)
peak assignment
1.1 1.0 1.0 1.4 1.2 1.4
2p3/2 - Ti 3d5/2 - Ag+ 3d5/2 - Ag0 2p3/2 - Ti4+ 3d5/2 - Ag0 2p3/2 - Ti4+
Ag0/(Ag0 + Ag+)
Ag/(Ag + Ti)
0.22
0.175
1.0
0.274
4+
Figure 8. Representation of Ag distribution on titania nanotubes.
Figure 7. XPS of the Ti 2p signal in (a) NT-573, (b) Ag-NT-573, and (c) Ag-NT-673 samples.
becomes proportional to the number of atoms at a specific location, assuming that all particles have the same chemical composition. Because the cross section of scattering increases approximately with the square of the atomic number (Z2), the intensity must be higher in the particles with high concentration of Ag atoms (e.g., heavy atoms) than in those where Ag and O atoms coexist. From these results, it can be established that silver nanoparticles on titania nanotubes can be formed by the simple impregnation of silver nitrate solution followed by thermal treatment at 673 K in an inert atmosphere. Highly and homogeneously dispersed silver nanoparticles can be obtained without employing any stabilizer organic ligand or additives for silver,47,48 any reductive agents such as sodium boronhydride or ammonia complex,17,47,48 and any UV illumination process for silver photoreduction.49,50 It is interesting that Ag nanocluster aggregation did not occur with increasing annealing temperature (below 673 K) and the size and morphology of Ag2O nanoparticles nearly correspond to those of Ag0 clusters, obtained after annealing at 673 K. Titania nanotubes prevent the aggregation of silver clusters because of the strong interaction of silver nanoparticles with the titania support, as observed in TEM images in Figure 2d,f. These highly crystalline silver nanoparticles in close contact with the titania support are expected to favor electron transfer between the TiO2 support (47) Zhang, L.; Yu, J. C.; Yip, H. Y.; Li, Q.; Kwong, K. W.; Xu, A. W.; Wong, P. K. Langmuir 2003, 19, 10372. (48) Rodriguez-Gatorno, G.; Dı´ az, D.; Rendon, L.; Hernandez-Segura, G. O. J. Phys. Chem. B 2002, 106, 2482. (49) Jin, M.; Zhang, X.; Nishimoto, S.; Liu, Z.; Tryk, D. A.; Emeline, A. V.; Murakami, T.; Fujishima, A. J. Phys. Chem. C 2007, 111, 658. (50) Sudeep and, P. K.; Kamat, P. V. Chem. Mater. 2005, 17, 5404. (51) Amazu, K.; Fujimaki, M.; Rockstuhl, C.; Tominaga, J.; Murakami, H.; Ohki, Y; Yoshida and, N.; Watanabe, T J. Am. Chem. Soc. 2008, 130, 1676. (52) Hirakawa and, T.; Kamat, P. V. Langmuir 2004, 20, 5645.
10200 DOI: 10.1021/la9009702
and silver nanoparticles,3,7,51,52 which is highly desirable in the development of semiconductor-metal nanocomposites. XPS of the Ti 2p signal in all samples was well fitted with only one doublet with a 2p3/2 peak at 458.4-459.0 eV and a 2p1/2 peak at 364.1-364.6 eV, corresponding to Ti4+. The fwhm of the peaks around 1.1-1.4 eV indicates no evidence of any reduced Ti species. The presence of either Ag2O or Ag0 nanoparticles on titania nanotubes gradually shifts the binding energy of the XPS 2p3/2 signal of Ti4+ atoms from 458.4 to 459.0 eV, as indicated in Figure 7 and Table 1, suggesting that electron transfer occurs between silver nanoparticles and the TiO2 support. That means that a transfer of electrons occurs from the semiconducting TiO2 to the metallic centers. The electrons are accumulated on silver particles and generate holes (h+) on the semiconductor, giving rise to more electropositive Ti4+ atoms and shifting the Ti4+ BE toward higher values. Then, by exciting Ag-NT samples with the X-ray beam during XPS experiments, TiO2 provides photogenerated electrons that can reduce Ag+ ions or can be accumulated on Ag0 nanoparticles, as indicated elsewhere.17 From XRD, HRTEM (HAADF), and XPS results, a graphical model of the silver oxidation state and location on titania nanotubes as a function of the heat-treatment temperature is proposed in Figure 8. UV-Vis Spectra. The UV-vis spectra of transformed Kubelka-Munk functions versus the band gap energy (BGE) of NT573, Ag-NT-573, and Ag-NT-673 samples are shown in Figure 9. The extrapolation of the linear portion of the modified spectra to zero absorption determines the BGE. Two band transitions were observed in the NT-573 and Ag-containing samples. The first one in the UV region (hν ≈ 3.0 eV) correspond to the fundamental absorption edge of titania, and the other in the visible region (hν ≈ 2.0 eV) correspond to the surface plasmonic resonance of silver nanoparticles. The UV-vis spectrum of anatase was also included as a reference. The BGE estimated for the nanotubular support, NT-573, was 2.84 eV, which is considerably red-shifted from that of anatase (e.g., 3.13 eV). Additionally, the NT-573 sample showed considerable absorption in the visible region, which is not observed in anatase. The differences between the optical properties on NT and anatase can be explained by assuming that the vacancies generated during the dehydroxylation of the tubular structure, where highly distorted TiO6 octahedra are bonded onto the curved structure, are not created in anatase titania nanoparticles. After annealing at higher Langmuir 2009, 25(17), 10195–10201
Toledo-Antonio et al.
Article
tion in the visible region in NT-573 can be attributed to the residual OH groups remaining in the interlayer space of the nanotubular structure, such as that occurring for the photoluminescence properties.53 When the sample (Ag-NT-673) was annealed at 673 K, although the support structure collapsed into anatase, for the transformation of Ag2O into Ag0 nanoparticles, the BGE of anatase shifts toward the red from 3.13 to 2.65 eV, indicating that electron transfer took place between the TiO2 support and Ag0 nanoparticles with a strong interaction, in agreement with XPS and HRTEM results. The absorption observed in the visible region increased again in Ag-NT-673; in this case, it corresponds to the surface plasmonic resonance of Ag0 nanoparticles.7
Conclusions Figure 9. UV-vis spectra of anatase, NT, and Ag-NT at the indicated annealing temperature.
temperature (i.e., 773 K), the nanotubular structure collapses and anatase is obtained, therefore blue shifting the band gap energy up to 3.2 eV. Accordingly, another optical property of titania nanotubes, such as photoluminescence, has also been found to depend on the annealing temperature of nanotubes, showing a maximum after annealing at 573 K, and at a higher temperature the nanotubes collapse and their photoluminescence properties decrease.53 The photoluminescence of titania nanotubes has been attributed to the luminescent centers of Ti-OH complexes in the nanotubular structure. The structure of the electronic band changes by means of a highly efficient charge recombination, which is expected to be stronger in 1D tubular structure than in 0D anatase nanoparticles. However, Ag containing samples (e.g., Ag-NT-573) showed quite similar band gap energy to that of the nanotubular support (e.g., NT-573), as shown in Figure 9. The absorption in the visible region of nanotubes considerably decreased with Ag incorporation into the Ag-NT-573 sample. As Ag incorporation is carried out, generating a staking disorder in the interlayer space of the nanotube walls as observed in the XRD results, the OH groups are expected to be eliminated more efficiently. Then, the absorp(53) Qian, L.; Jin, Z.-S.; Yang, S.-Y.; Du, S.-L.; Xu.-Rong, X. Chem. Mater. 2005, 17, 5334.
Langmuir 2009, 25(17), 10195–10201
The formation of quasi-monodisperse Ag nanoparticles on TiO2 nanotubes can be easily carried out by means of an impregnative ion exchange method using aqueous AgNO3 and then annealing at temperatures between 573 and 673 K. The resulting Ag nanoparticles have an average size of 5 nm, and as indicated by HRTEM, they do not agglomerate when annealing below 673 K, showing a clear polka-dot pattern. The Ag particles show truncated octahedron shapes. Initially in the fresh materials, the Ag+ cations partially intercalcate into the layers of the nanotube walls, and then upon annealing, they form Ag2O and/ or Ag0 nanoparticles, which migrate to the surface of the nanotubes. The oxidation state of surface Ag nanoparticles, in an inert annealing atmosphere, depends on the heat-treatment temperature, thus from 22 to 100% Ag0 can be generated by annealing at 573 and 673 K, respectively. Therefore, a reducing agent is not necessary when forming Ag0 nanoparticles on titania nanotubes. The electronic spectra of Ag-containing samples indicate that only in Ag-NT-673 visible light can be considerably absorbed, but not when annealing at 573 K. Given these physicochemical properties, the application of these Ag nanoparticles on titania nanotubes as antimicrobial agents and/or photocatalysts can be anticipated. Acknowledgment. This work was supported by IMP Project D.00483. The authors are grateful to Mr. Juan Fernando BonillaAguilar for technical assistance in UV-vis measurements.
DOI: 10.1021/la9009702
10201
