- Email: [email protected]
Formation of metallic Ni nanoparticles on titania surfaces by chemical vapor reductive deposition method
Journal of Colloid and Interface Science 309 (2007) 149–154 www.elsevier.com/locate/jcis
Formation of metallic Ni nanoparticles on titania surfaces by chemical vapor reductive deposition method Masaki Yoshinaga a , Hideyuki Takahashi b , Katsutoshi Yamamoto a , Atsushi Muramatsu a,∗ , Takeshi Morikawa c a Institute of Multidisciplinary Research for Advanced Materials, Tohoku University, Sendai 980-8577, Japan b Graduate School of Environmental Studies, Tohoku University, Sendai 980-8579, Japan c Toyota Central Research and Development Laboratories, Inc., Nagakute, Aichi 480-1192, Japan
Received 27 October 2006; accepted 23 January 2007
Abstract Metallic Ni nanoparticles were successfully prepared on the surface of titania thin film substrate by a novel method, named as chemical vapor reductive deposition (CVRD) method. The growth of the nanoparticles was based on the specific adsorption and heterogeneous nucleation on the surface of substrate, not via vapor-phase formation and subsequent sedimentation. The nanoparticle size was found to be well controllable between 10 and 30 nm by the preparation time and vapor pressure of metal complex precursor. ESCA and electron diffraction results clearly demonstrated Ni nanoparticles as metallic. Titania thin film with metallic Ni nanoparticles on its surface showed high efficiency in their photocatalysis of hydrogen evolution from decomposition of ethanol. © 2007 Elsevier Inc. All rights reserved. Keywords: Ni nanoparticles; Heteronucleation; Growth control; High dispersibility; Titania thin films; Photocatalysis; Hydrogen; Ethanol decomposition
1. Introduction Metallic nanoparticles as catalysts have been vigorously investigated because of their specific properties, such as large surface area, and their superior properties, different from those of bulk materials. However, catalytic activities of the nanoparticles was often reduced due to their tremendous aggregation, because of their high surface energy. To solve this problem, many researchers have reported various synthetic methods for metallic nanoparticles supported on the surfaces of other materials [1]. Among many synthesis techniques, the liquid phase reduction method is one of the main candidate methods, since metallic Ni nanoparticles with a diameter of 1–10 nm can be supported on the surfaces of titania particles [2,3]. These studies have been based on nanoparticle growth on specific fine particle supports in the liquid phase [3–6]. On the other * Corresponding author. Fax: +81 22 217 5163.
E-mail address: [email protected] (A. Muramatsu). 0021-9797/$ – see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2007.01.051
hand, there are few reports dealing with nanoparticles grown on films. Although the synthesis of nanoparticles on films has been simply and traditionally carried out by the chemical vapor deposition (CVD) method, the growth mechanism of the nanoparticles was predominated by spontaneous homogeneous nucleation in the gas phase and aggregative deposition of primary particles. Thus, the CVD method was known to have less uniformity and dispersibility in nanoparticle preparation [1]. Although many studies of the preparation of metallic nanoparticles with a substrate, for examples, SiO2 , single-crystal Si or Au, or some highly ordered substrate, have been carried out [7,8] by CVD methods, little attention is focused on heterogeneous nucleation and nanoparticle growth on substrate surfaces. Therefore, in this report, a novel synthesis technique based on heterogeneous nucleation and growth on a substrate surface, named the chemical vapor reductive deposition (CVRD) method, is developed.
150
M. Yoshinaga et al. / Journal of Colloid and Interface Science 309 (2007) 149–154
2. Experimental 2.1. Preparation Nickelocene (Ni(C5 H5 )2 , bis(η5 -cyclopentadienyl)nickel(II), 99% purity, Wako Pure Chemicals Ind. Ltd.) and hydrazine monohydrate ((NH2 )2 ·H2 O, 99.99% purity, Wako Pure Chemicals Ind. Ltd.) were used as Ni precursor and reducing agent, respectively. A quartz glass (12 mm square and 1 mm thick) was used as substrate. Before any treatments, the quartz substrate was ultrasonically cleaned in acetone, ethanol, and distilled water for 10 min each and then dried in air at 80 ◦ C. Titania thin film was then prepared on the quartz-glass substrate by the pulsed laser deposition (PLD) method [9]. PLD was carried out by Nd: YAG laser (λ = 1064 nm) irradiation of a rutiletype titania target at room temperature under vacuum less than 6.7 × 10−9 kPa. Obtained titania thin films, which have ca. 50 nm thickness, showed amorphous structure by XRD measurement but were found to have fine crystal structures that consisted of anatase and rutile, from electron diffraction patterns. Surface morphology observation revealed that although there were some island-like particles, the large part was relatively smooth. From UV–vis measurement, the absorption edge was around 310 nm, shorter than the bulk titania of ca. 400 nm. This seems to be caused by the quantum size effect as shown in the literature [9]. Fig. 1 illustrates schematic drawing of experimental equipment. The substrate was put at the center of a horizontal Pyrex glass tube (35 mm external diameter, 350 mm length). The glass tube was heated to 100 ◦ C by an electric furnace under vacuum less than 1.00 × 10−2 kPa. After the substrate temperature was stabilized, the precursor vapor was introduced into the glass tube, and vaporized hydrazine gas was immediately introduced. After reductive reaction for 1–10 min, the reaction tube was evacuated to less than 1.00 × 10−2 kPa again and cooled down to room temperature. Prepared materials were characterized by AFM (SPM9500J2, Shimadzu), XRF (System3270EL, Rigaku), UV–vis (UV-2550, Shimadzu), ESCA (ESCA5600, ULVAC-PHI), and TEM (HF-2000EX, Hitachi, at 200 kV). 2.2. Photocatalytic activity evaluation Photocatalytic activities of the synthesized Ni/titania thin film catalysts were evaluated through the decomposition of
Fig. 1. Schematic drawing of CVRD equipment used in this study.
ethanol. Photocatalytic decomposition of ethanol was brought about as follows [10]: + CH3 CH2 OH + h+ (VB) → CH3 CHOH + H , + CH3 CHOH + h+ (VB) → CH3 CHO + H ,
1 − → H2 . H+ + e(CB) 2 Thus, photocatalytic activities can be evaluated by employing the amount of hydrogen generated. The Ni/titania catalyst was placed in a quartz reaction vessel. After evacuation of the vessel to less than 1.00 × 10−2 kPa, 2.67 kPa of ethanol and 46.7 kPa of Ar gas were introduced. A super-high-pressure mercury lamp (BMO-500DJ and Hx-500, WACOM R&D) was used for UV light irradiation. The amount of hydrogen evolved was estimated by TCD (GC-8A, Shimadzu). 3. Results and discussions 3.1. Heterogeneous nucleation of Ni nanoparticles and their growth First of all, we prepared Ni nanoparticles directly on a quartz-glass substrate without making titania, in order to confirm the usefulness of this technique. Fig. 2 shows the AFM top-view images; one is the original quartz-glass substrate and another is CVRD treated under the condition of 0.03 kPa precursor and preparation time 1 min. The latter image (Fig. 2b) clearly demonstrated that many nanoparticles were successfully grown on the substrate as compared with the untreated substrate (Fig. 2a). From the cross-sectional scanning images shown in the figure, the nanoparticle size (Fig. 2b) was found to be 20–40 nm. Thus, it was indicated that the nanoparticles could be successfully prepared on the surface of the substrate. In addition, Ni nanoparticles were not formed at all without the reducing agent. Apparently, Ni nanoparticles were nucleated on the substrate surface through the reduction of nickelocene. Fig. 3 shows AFM images of materials obtained at 0.30 kPa with preparation time 1 (a), 5 (b), and 10 min (c)–(d), respectively. The difference between (c) and (d) was the setting direction of the substrate in the reaction tube, (c) horizontally and (d) vertically. The size of the nanoparticles was increased with elongation of the preparation time. The thickness of those materials was measured as (a) 24, (b) 34, and (c) 62 nm, respectively. Moreover, even if the setting direction of the substrate was different, the morphology in both (c) and (d) was almost the same. This fact indicted that the nanoparticles were obtained on the substrate surface, not via homogeneous nucleation and deposition as chemical vapor deposition, but via adsorption of nickelocene and its reduction by reducing agent, followed by heteronucleation in the chemical vapor reductive deposition. Fig. 4a shows the ESCA Ni2p spectra of the sample of Fig. 3c and the Ar sputtered one (Fig. 3b). The peak position was normalized by the C1s peak at 284.6 eV. Only one peak of the Ni2− species (855.6 eV) was detected from the as-prepared sample. This result agreed well with the result of a Ni/silica specimen reported by J.P. Espinos et al. [11]. On the other hand,
M. Yoshinaga et al. / Journal of Colloid and Interface Science 309 (2007) 149–154
151
Fig. 2. AFM top-view surface images of quartz glass (a) and Ni nanoparticles prepared at 0.03 kPa with the preparation time of 1 min (b).
Fig. 3. AFM surface images of Ni nanoparticles prepared at 0.30 kPa with preparation times of 1 (a), 5 (b), and 10 min (c, d), respectively. Specimen (d) was prepared by setting the substrate vertically. Each value at the lower right indicates the thickness of the respective film.
after Ar sputtering for 1 min, a new intense peak assigned to metallic Ni0 species (852.7 eV) was clearly observed, together with a small amount of the Ni2− species [12]. According to a report by J.C.D. Jesus et al., the peak position around 855.6 eV was considered to be NiO and Ni(OH)2 species [13]. These oxidized Ni compounds could be formed by the exposure of the as-prepared sample to air. Thus, the Ni nanoparticles prepared on the substrate surface must have been originally metallic; nevertheless the surface of the nanoparticles was hereafter oxidized. Fig. 5 shows the TEM micrographs of the Ni nanoparticles prepared at 0.30 kPa for 1 min, which is the same condition as
Fig. 4a. The Ni nanoparticles with size less than 20 nm were clearly observed in the bright-field image (Fig. 5a). This result corresponds well with the AFM observation (Fig. 4a). Judging from the electron diffraction patterns in Fig. 5b, three kinds of Ni phases, metallic Ni, NiO, and Ni(OH)2 , were formed, consistent with the results of ESCA analysis, as mentioned above. In addition, the dark field image (Fig. 5c) demonstrated that the Ni nanoparticles had polycrystalline structure. Consequently, the formation route of metallic Ni nanoparticles in this study is considered to be as follows: (1) precursor molecules adsorbed on the surface of the substrate [1]; (2) heterogeneous nucleation took place through the reduction
152
M. Yoshinaga et al. / Journal of Colloid and Interface Science 309 (2007) 149–154
of the adsorbed precursor [2,3]; and (3) metallic Ni nanoparticles grew. Clearly, the CVRD method can successfully provide metallic nanoparticles with size controllability on a substrate surface. 3.2. Chemical vapor reductive deposition of Ni nanoparticles on titania Fig. 6a shows an AFM image of the surface of titania deposited on the substrate by PLD [9]. From the AFM images after CVRD treatment, Ni nanoparticles with uniform size and
Fig. 4. ESCA Ni2p spectra of Ni nanoparticles before (a) and after Ar sputtering (b). The preparation conditions are 0.30 kPa of precursor and 10 min of time.
high density were formed on the titania surface. The particle size was found to be 10–17 nm (0.03 kPa in Fig. 6b), 12–20 nm (0.10 kPa in Fig. 6c), and 15–30 nm (0.30 kPa in Fig. 6d), respectively. Thus, it was suggested that the size of the Ni nanoparticles on titania was successfully controlled by the preparation conditions. In addition, the loading amount of Ni was gradually increased with increasing precursor amount, as shown in Fig. 7. Brissonneau and Vahlas [14] attempted to prepare Ni thin films by the CVD method with nickelocene and hydrogen and Carpenter and Wronski [15] reported the synthesis of Ni nanoparticles by CVD with nickel carbonyl. Unfortunately, their Ni size was surprisingly large, several hundred nanometers in thickness for the former, and several dozen nanometers in diameter with uniform morphology for the latter. Compared to their reports, Ni nanoparticles obtained on titania in the present study have very fine size with good uniformity in structure and composition. The size of the Ni nanoparticles formed on the titania surface was found to be much less than that in the absence of titania, as mentioned above. Since metal complexes can selectively adsorb onto the surface hydroxyl groups of the surface of the metal oxide support [1], the number of precursor molecules adsorbing onto the substrate surface is increased with increasing number of hydroxyl groups, so that the number of the nucleation sites could be increased. Generally, metal oxide thin films prepared by the PLD method have some defects, e.g., oxygen vacancies [16]. The oxygen vacancies located are not only at the inside but also at the surface, where unsaturated Ti atoms would form Ti–OH bonds. Hence, the fact that the Ni particle
Fig. 5. TEM bright field (a), electron diffraction patterns (b), and dark field photographs (c). The preparation conditions are 0.30 kPa of precursor and 1 min of time.
M. Yoshinaga et al. / Journal of Colloid and Interface Science 309 (2007) 149–154
153
Fig. 6. AFM images of titania alone and Ni/titania thin films. The conditions are 10 min of preparation time and precursor at 0.03 (b), 0.10 (c), and 0.30 kPa (d), respectively.
Fig. 7. XRF NiKα intensity change with vapor pressure of precursor. Inset figure shows a typical profile of NiKα in Ni/titania thin films.
size on titania was rather small by PLD is presumably the result of many metal–hydroxyl bonds on the titania surface. Fig. 8 shows the ESCA Ni2p spectra of titania alone (a) and Ni/titania prepared under the condition of 0.03 kPa for 10 min (b) and 0.30 kPa for 10 min (c), respectively. The spectra indicate two peaks, in which higher and lower binding energy peaks were attributed to Ni2+ and Ni0 species (Figs. 8b and 8c), whereas no Ni2p peak appeared on the titania thin film (Fig. 8a). As mentioned above, a higher binding energy peak at 856.1 eV suggests that original metallic Ni species were oxidized by exposure to air. On the other hand, Ni0 species attributed to the metallic phase were also detected in the spectra. The peaks around 852.5 (a) and 852.2 eV (b) were slightly lower than the bulk Ni0 of 852.7 eV [12]. It may be due to the same effect as reported by Bradford and Vannice [17], in which
Fig. 8. ESCA Ni2p spectra of titania alone (a) and Ni/titania in 0.03 kPa (b) and 0.30 kPa with a preparation time of 10 min (c).
they suggested a strong metal–support interaction between Ni and titania [17,18]. 3.3. Photocatalytic activities of Ni/titania thin film catalysts As shown in Fig. 9, the UV–vis spectra of Ni/titania and titania were almost the same, suggesting that the Ni nanoparticles did not disturb the photoabsorption of titania. Photocatalytic activities in hydrogen evolution rate from ethanol decomposition are summarized in Table 1. Every Ni/titania catalyst showed higher activity than titania alone (33 × 10−8 mol cm−2 h−1 W−1 ). On the other hand, with increasing Ni loading, the catalytic activity was drastically enhanced. Among them, the Ni/titania catalyst prepared under the conditions of 0.03 kPa nickelocene and 10 min of preparation time showed the highest production rate, 117 ×
154
M. Yoshinaga et al. / Journal of Colloid and Interface Science 309 (2007) 149–154
decreased with the reaction time; this is possibly due to the oxygen defects that cause the high catalytic activity but gradually disappeared during the photoirradiation [21]. On the other hand, Ni/titania showed stable catalytic activity. This may be because the Ni nanoparticles formed on titania surface were quite stable due to the strong interaction between Ni and titania [22], and because the sintering of Ni to grow was consequently suppressed. 4. Conclusion
Fig. 9. UV–vis transmittance spectra of titania alone (dotted line) and Ni/titania (solid line) thin films. Table 1 Photocatalytic activities of titania alone and Ni/titania catalysts Preparation conditions Ni(C5 H5 )2 (kPa)
Time (min)
0.03 0.03 0.03 0.30 0.10 0.30 Titania alone
1 5 8 10 10 10
H2 production (×10−8 mol cm−2 h−1 W−1 ) 40.12 55.68 62.83 117.42 91.75 68.64 33.27
Fig. 10. Photocatalytic stability of titania alone (open circles) and Ni/titania (closed squares) of the optimal catalyst in Table 1.
10−8 mol cm−2 h−1 W−1 . However, the activity was lowered when too large an amount of Ni was loaded. According to the AFM images shown in Fig. 6, the Ni nanoparticle size was larger under the condition of 0.30 kPa of precursor vapor pressure than under that of 0.03 kPa. Because the catalytic activity is often strongly related to the specific surface area and particle size [19,20], the Ni/titania thin film catalysts which have small particle size and high dispersibility are expected to show high photocatalytic performance. Fig. 10 shows the photocatalytic stability against the irradiation time of Ni/titania catalyst, in comparison with of titania alone. In the case of titania, the photocatalytic activity
In this study, we propose a new nanoparticle preparation technique, the CVRD (chemical vapor reductive deposition) method, which could accomplish the formation of metallic Ni nanoparticles on the surface of a titania thin film substrate. The heterogeneous nucleation and the particle growth were well monitored and controlled through various investigations. The Ni nanoparticles successfully enhanced the photocatalytic activity of titania thin film. It is concluded that the CVRD method can be one of the best candidate techniques for preparing metallic nanoparticles on the supporting material, in particular as the first development in a nanoparticle–thin film preparation system. References [1] K.P. de Jong, Curr. Opin. Solid State Mater. Sci. 4 (1999) 55–62. [2] H. Takahashi, Y. Sunagawa, S. Myagmarjav, K. Yamamoto, N. Sato, A. Muramatsu, Mater. Trans. 44 (2003) 2414–2416. [3] S. Myagmarjav, H. Takahashi, Y. Sunagawa, K. Yamamoto, N. Sato, E. Matsubara, A. Muramatsu, Mater. Trans. 45 (2004) 2035–2038. [4] P. Burattin, M. Che, C. Louis, J. Phys. Chem. B 104 (2000) 10482–10489. [5] A. Jasik, R. Wojcieszak, S. Monteverdi, M. Ziolek, M.M. Bettahar, J. Mol. Catal. A 242 (2005) 81–90. [6] P. Li, J. Liu, N. Nag, P.A. Crozier, Surf. Sci. 600 (2006) 693–702. [7] Z. Bastl, J. Franc, P. Janda, H. Pelouchová, Z. Samec, Nanotechnology 17 (2006) 1492–1500. [8] J. Wang, H. Zhang, J. Ge, Y. Li, J. Phys. Chem. B 110 (2006) 807–811. [9] T. Nakamura, T. Ichitsubo, E. Matsubara, A. Muramatsu, N. Sato, H. Takahashi, Acta Mater. 53 (2005) 323–329. [10] B.R. Müller, S. Majoni, D. Meissner, R. Memming, J. Photochem. Photobiol. A 151 (2002) 253–265. [11] J.P. Espinòs, A.R.G. Elipe, A. Caballero, J. Garcìa, G. Munuera, J. Catal. 136 (1992) 415–422. [12] C.D. Wagner, W.M. Riggs, L.E. Davis, J.F. Moulder, Handbook of X-ray Photoelectron Spectroscopy, Physical Electronics Industries, Eden Prairie, MN, 1983. [13] J.C.D. Jesus, I. González, A. Quevedo, T. Puerta, J. Mol. Catal. A 228 (2005) 283–291. [14] L. Brissonneau, C. Vahlas, Chem. Vap. Deposition 5 (1999) 135–142. [15] G.J.C. Carpenter, Z.S. Wronski, J. Nanopart. Res. 6 (2004) 215–221. [16] V. Craciun, R.K. Singh, Appl. Phys. Lett. 76 (2000) 1932–1934. [17] M.C.J. Bradford, M.A. Vannice, Appl. Catal. A 142 (1996) 73–96. [18] J.S. Smith, P.A. Thrower, M.A. Vannice, J. Catal. 68 (1981) 270–285. [19] J.L. Carter, J.A. Cusumano, J.H. Sinfelt, J. Phys. Chem. 70 (1966) 2257– 2263. [20] T. Hayashi, K. Tanaka, M. Haruta, J. Catal. 178 (1998) 566–575. [21] T. Ihara, M. Miyoshi, Y. Iriyama, O. Matsumoto, S. Sugihara, Appl. Catal. B 42 (2003) 403–409. [22] K. Otsuka, S. Takenaka, Catal. Surv. Asia 8 (2004) 77–90.
Formation of metallic Ni nanoparticles on titania surfaces by chemical vapor reductive deposition method Masaki Yoshinaga a , Hideyuki Takahashi b , Katsutoshi Yamamoto a , Atsushi Muramatsu a,∗ , Takeshi Morikawa c a Institute of Multidisciplinary Research for Advanced Materials, Tohoku University, Sendai 980-8577, Japan b Graduate School of Environmental Studies, Tohoku University, Sendai 980-8579, Japan c Toyota Central Research and Development Laboratories, Inc., Nagakute, Aichi 480-1192, Japan
Received 27 October 2006; accepted 23 January 2007
Abstract Metallic Ni nanoparticles were successfully prepared on the surface of titania thin film substrate by a novel method, named as chemical vapor reductive deposition (CVRD) method. The growth of the nanoparticles was based on the specific adsorption and heterogeneous nucleation on the surface of substrate, not via vapor-phase formation and subsequent sedimentation. The nanoparticle size was found to be well controllable between 10 and 30 nm by the preparation time and vapor pressure of metal complex precursor. ESCA and electron diffraction results clearly demonstrated Ni nanoparticles as metallic. Titania thin film with metallic Ni nanoparticles on its surface showed high efficiency in their photocatalysis of hydrogen evolution from decomposition of ethanol. © 2007 Elsevier Inc. All rights reserved. Keywords: Ni nanoparticles; Heteronucleation; Growth control; High dispersibility; Titania thin films; Photocatalysis; Hydrogen; Ethanol decomposition
1. Introduction Metallic nanoparticles as catalysts have been vigorously investigated because of their specific properties, such as large surface area, and their superior properties, different from those of bulk materials. However, catalytic activities of the nanoparticles was often reduced due to their tremendous aggregation, because of their high surface energy. To solve this problem, many researchers have reported various synthetic methods for metallic nanoparticles supported on the surfaces of other materials [1]. Among many synthesis techniques, the liquid phase reduction method is one of the main candidate methods, since metallic Ni nanoparticles with a diameter of 1–10 nm can be supported on the surfaces of titania particles [2,3]. These studies have been based on nanoparticle growth on specific fine particle supports in the liquid phase [3–6]. On the other * Corresponding author. Fax: +81 22 217 5163.
E-mail address: [email protected] (A. Muramatsu). 0021-9797/$ – see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2007.01.051
hand, there are few reports dealing with nanoparticles grown on films. Although the synthesis of nanoparticles on films has been simply and traditionally carried out by the chemical vapor deposition (CVD) method, the growth mechanism of the nanoparticles was predominated by spontaneous homogeneous nucleation in the gas phase and aggregative deposition of primary particles. Thus, the CVD method was known to have less uniformity and dispersibility in nanoparticle preparation [1]. Although many studies of the preparation of metallic nanoparticles with a substrate, for examples, SiO2 , single-crystal Si or Au, or some highly ordered substrate, have been carried out [7,8] by CVD methods, little attention is focused on heterogeneous nucleation and nanoparticle growth on substrate surfaces. Therefore, in this report, a novel synthesis technique based on heterogeneous nucleation and growth on a substrate surface, named the chemical vapor reductive deposition (CVRD) method, is developed.
150
M. Yoshinaga et al. / Journal of Colloid and Interface Science 309 (2007) 149–154
2. Experimental 2.1. Preparation Nickelocene (Ni(C5 H5 )2 , bis(η5 -cyclopentadienyl)nickel(II), 99% purity, Wako Pure Chemicals Ind. Ltd.) and hydrazine monohydrate ((NH2 )2 ·H2 O, 99.99% purity, Wako Pure Chemicals Ind. Ltd.) were used as Ni precursor and reducing agent, respectively. A quartz glass (12 mm square and 1 mm thick) was used as substrate. Before any treatments, the quartz substrate was ultrasonically cleaned in acetone, ethanol, and distilled water for 10 min each and then dried in air at 80 ◦ C. Titania thin film was then prepared on the quartz-glass substrate by the pulsed laser deposition (PLD) method [9]. PLD was carried out by Nd: YAG laser (λ = 1064 nm) irradiation of a rutiletype titania target at room temperature under vacuum less than 6.7 × 10−9 kPa. Obtained titania thin films, which have ca. 50 nm thickness, showed amorphous structure by XRD measurement but were found to have fine crystal structures that consisted of anatase and rutile, from electron diffraction patterns. Surface morphology observation revealed that although there were some island-like particles, the large part was relatively smooth. From UV–vis measurement, the absorption edge was around 310 nm, shorter than the bulk titania of ca. 400 nm. This seems to be caused by the quantum size effect as shown in the literature [9]. Fig. 1 illustrates schematic drawing of experimental equipment. The substrate was put at the center of a horizontal Pyrex glass tube (35 mm external diameter, 350 mm length). The glass tube was heated to 100 ◦ C by an electric furnace under vacuum less than 1.00 × 10−2 kPa. After the substrate temperature was stabilized, the precursor vapor was introduced into the glass tube, and vaporized hydrazine gas was immediately introduced. After reductive reaction for 1–10 min, the reaction tube was evacuated to less than 1.00 × 10−2 kPa again and cooled down to room temperature. Prepared materials were characterized by AFM (SPM9500J2, Shimadzu), XRF (System3270EL, Rigaku), UV–vis (UV-2550, Shimadzu), ESCA (ESCA5600, ULVAC-PHI), and TEM (HF-2000EX, Hitachi, at 200 kV). 2.2. Photocatalytic activity evaluation Photocatalytic activities of the synthesized Ni/titania thin film catalysts were evaluated through the decomposition of
Fig. 1. Schematic drawing of CVRD equipment used in this study.
ethanol. Photocatalytic decomposition of ethanol was brought about as follows [10]: + CH3 CH2 OH + h+ (VB) → CH3 CHOH + H , + CH3 CHOH + h+ (VB) → CH3 CHO + H ,
1 − → H2 . H+ + e(CB) 2 Thus, photocatalytic activities can be evaluated by employing the amount of hydrogen generated. The Ni/titania catalyst was placed in a quartz reaction vessel. After evacuation of the vessel to less than 1.00 × 10−2 kPa, 2.67 kPa of ethanol and 46.7 kPa of Ar gas were introduced. A super-high-pressure mercury lamp (BMO-500DJ and Hx-500, WACOM R&D) was used for UV light irradiation. The amount of hydrogen evolved was estimated by TCD (GC-8A, Shimadzu). 3. Results and discussions 3.1. Heterogeneous nucleation of Ni nanoparticles and their growth First of all, we prepared Ni nanoparticles directly on a quartz-glass substrate without making titania, in order to confirm the usefulness of this technique. Fig. 2 shows the AFM top-view images; one is the original quartz-glass substrate and another is CVRD treated under the condition of 0.03 kPa precursor and preparation time 1 min. The latter image (Fig. 2b) clearly demonstrated that many nanoparticles were successfully grown on the substrate as compared with the untreated substrate (Fig. 2a). From the cross-sectional scanning images shown in the figure, the nanoparticle size (Fig. 2b) was found to be 20–40 nm. Thus, it was indicated that the nanoparticles could be successfully prepared on the surface of the substrate. In addition, Ni nanoparticles were not formed at all without the reducing agent. Apparently, Ni nanoparticles were nucleated on the substrate surface through the reduction of nickelocene. Fig. 3 shows AFM images of materials obtained at 0.30 kPa with preparation time 1 (a), 5 (b), and 10 min (c)–(d), respectively. The difference between (c) and (d) was the setting direction of the substrate in the reaction tube, (c) horizontally and (d) vertically. The size of the nanoparticles was increased with elongation of the preparation time. The thickness of those materials was measured as (a) 24, (b) 34, and (c) 62 nm, respectively. Moreover, even if the setting direction of the substrate was different, the morphology in both (c) and (d) was almost the same. This fact indicted that the nanoparticles were obtained on the substrate surface, not via homogeneous nucleation and deposition as chemical vapor deposition, but via adsorption of nickelocene and its reduction by reducing agent, followed by heteronucleation in the chemical vapor reductive deposition. Fig. 4a shows the ESCA Ni2p spectra of the sample of Fig. 3c and the Ar sputtered one (Fig. 3b). The peak position was normalized by the C1s peak at 284.6 eV. Only one peak of the Ni2− species (855.6 eV) was detected from the as-prepared sample. This result agreed well with the result of a Ni/silica specimen reported by J.P. Espinos et al. [11]. On the other hand,
M. Yoshinaga et al. / Journal of Colloid and Interface Science 309 (2007) 149–154
151
Fig. 2. AFM top-view surface images of quartz glass (a) and Ni nanoparticles prepared at 0.03 kPa with the preparation time of 1 min (b).
Fig. 3. AFM surface images of Ni nanoparticles prepared at 0.30 kPa with preparation times of 1 (a), 5 (b), and 10 min (c, d), respectively. Specimen (d) was prepared by setting the substrate vertically. Each value at the lower right indicates the thickness of the respective film.
after Ar sputtering for 1 min, a new intense peak assigned to metallic Ni0 species (852.7 eV) was clearly observed, together with a small amount of the Ni2− species [12]. According to a report by J.C.D. Jesus et al., the peak position around 855.6 eV was considered to be NiO and Ni(OH)2 species [13]. These oxidized Ni compounds could be formed by the exposure of the as-prepared sample to air. Thus, the Ni nanoparticles prepared on the substrate surface must have been originally metallic; nevertheless the surface of the nanoparticles was hereafter oxidized. Fig. 5 shows the TEM micrographs of the Ni nanoparticles prepared at 0.30 kPa for 1 min, which is the same condition as
Fig. 4a. The Ni nanoparticles with size less than 20 nm were clearly observed in the bright-field image (Fig. 5a). This result corresponds well with the AFM observation (Fig. 4a). Judging from the electron diffraction patterns in Fig. 5b, three kinds of Ni phases, metallic Ni, NiO, and Ni(OH)2 , were formed, consistent with the results of ESCA analysis, as mentioned above. In addition, the dark field image (Fig. 5c) demonstrated that the Ni nanoparticles had polycrystalline structure. Consequently, the formation route of metallic Ni nanoparticles in this study is considered to be as follows: (1) precursor molecules adsorbed on the surface of the substrate [1]; (2) heterogeneous nucleation took place through the reduction
152
M. Yoshinaga et al. / Journal of Colloid and Interface Science 309 (2007) 149–154
of the adsorbed precursor [2,3]; and (3) metallic Ni nanoparticles grew. Clearly, the CVRD method can successfully provide metallic nanoparticles with size controllability on a substrate surface. 3.2. Chemical vapor reductive deposition of Ni nanoparticles on titania Fig. 6a shows an AFM image of the surface of titania deposited on the substrate by PLD [9]. From the AFM images after CVRD treatment, Ni nanoparticles with uniform size and
Fig. 4. ESCA Ni2p spectra of Ni nanoparticles before (a) and after Ar sputtering (b). The preparation conditions are 0.30 kPa of precursor and 10 min of time.
high density were formed on the titania surface. The particle size was found to be 10–17 nm (0.03 kPa in Fig. 6b), 12–20 nm (0.10 kPa in Fig. 6c), and 15–30 nm (0.30 kPa in Fig. 6d), respectively. Thus, it was suggested that the size of the Ni nanoparticles on titania was successfully controlled by the preparation conditions. In addition, the loading amount of Ni was gradually increased with increasing precursor amount, as shown in Fig. 7. Brissonneau and Vahlas [14] attempted to prepare Ni thin films by the CVD method with nickelocene and hydrogen and Carpenter and Wronski [15] reported the synthesis of Ni nanoparticles by CVD with nickel carbonyl. Unfortunately, their Ni size was surprisingly large, several hundred nanometers in thickness for the former, and several dozen nanometers in diameter with uniform morphology for the latter. Compared to their reports, Ni nanoparticles obtained on titania in the present study have very fine size with good uniformity in structure and composition. The size of the Ni nanoparticles formed on the titania surface was found to be much less than that in the absence of titania, as mentioned above. Since metal complexes can selectively adsorb onto the surface hydroxyl groups of the surface of the metal oxide support [1], the number of precursor molecules adsorbing onto the substrate surface is increased with increasing number of hydroxyl groups, so that the number of the nucleation sites could be increased. Generally, metal oxide thin films prepared by the PLD method have some defects, e.g., oxygen vacancies [16]. The oxygen vacancies located are not only at the inside but also at the surface, where unsaturated Ti atoms would form Ti–OH bonds. Hence, the fact that the Ni particle
Fig. 5. TEM bright field (a), electron diffraction patterns (b), and dark field photographs (c). The preparation conditions are 0.30 kPa of precursor and 1 min of time.
M. Yoshinaga et al. / Journal of Colloid and Interface Science 309 (2007) 149–154
153
Fig. 6. AFM images of titania alone and Ni/titania thin films. The conditions are 10 min of preparation time and precursor at 0.03 (b), 0.10 (c), and 0.30 kPa (d), respectively.
Fig. 7. XRF NiKα intensity change with vapor pressure of precursor. Inset figure shows a typical profile of NiKα in Ni/titania thin films.
size on titania was rather small by PLD is presumably the result of many metal–hydroxyl bonds on the titania surface. Fig. 8 shows the ESCA Ni2p spectra of titania alone (a) and Ni/titania prepared under the condition of 0.03 kPa for 10 min (b) and 0.30 kPa for 10 min (c), respectively. The spectra indicate two peaks, in which higher and lower binding energy peaks were attributed to Ni2+ and Ni0 species (Figs. 8b and 8c), whereas no Ni2p peak appeared on the titania thin film (Fig. 8a). As mentioned above, a higher binding energy peak at 856.1 eV suggests that original metallic Ni species were oxidized by exposure to air. On the other hand, Ni0 species attributed to the metallic phase were also detected in the spectra. The peaks around 852.5 (a) and 852.2 eV (b) were slightly lower than the bulk Ni0 of 852.7 eV [12]. It may be due to the same effect as reported by Bradford and Vannice [17], in which
Fig. 8. ESCA Ni2p spectra of titania alone (a) and Ni/titania in 0.03 kPa (b) and 0.30 kPa with a preparation time of 10 min (c).
they suggested a strong metal–support interaction between Ni and titania [17,18]. 3.3. Photocatalytic activities of Ni/titania thin film catalysts As shown in Fig. 9, the UV–vis spectra of Ni/titania and titania were almost the same, suggesting that the Ni nanoparticles did not disturb the photoabsorption of titania. Photocatalytic activities in hydrogen evolution rate from ethanol decomposition are summarized in Table 1. Every Ni/titania catalyst showed higher activity than titania alone (33 × 10−8 mol cm−2 h−1 W−1 ). On the other hand, with increasing Ni loading, the catalytic activity was drastically enhanced. Among them, the Ni/titania catalyst prepared under the conditions of 0.03 kPa nickelocene and 10 min of preparation time showed the highest production rate, 117 ×
154
M. Yoshinaga et al. / Journal of Colloid and Interface Science 309 (2007) 149–154
decreased with the reaction time; this is possibly due to the oxygen defects that cause the high catalytic activity but gradually disappeared during the photoirradiation [21]. On the other hand, Ni/titania showed stable catalytic activity. This may be because the Ni nanoparticles formed on titania surface were quite stable due to the strong interaction between Ni and titania [22], and because the sintering of Ni to grow was consequently suppressed. 4. Conclusion
Fig. 9. UV–vis transmittance spectra of titania alone (dotted line) and Ni/titania (solid line) thin films. Table 1 Photocatalytic activities of titania alone and Ni/titania catalysts Preparation conditions Ni(C5 H5 )2 (kPa)
Time (min)
0.03 0.03 0.03 0.30 0.10 0.30 Titania alone
1 5 8 10 10 10
H2 production (×10−8 mol cm−2 h−1 W−1 ) 40.12 55.68 62.83 117.42 91.75 68.64 33.27
Fig. 10. Photocatalytic stability of titania alone (open circles) and Ni/titania (closed squares) of the optimal catalyst in Table 1.
10−8 mol cm−2 h−1 W−1 . However, the activity was lowered when too large an amount of Ni was loaded. According to the AFM images shown in Fig. 6, the Ni nanoparticle size was larger under the condition of 0.30 kPa of precursor vapor pressure than under that of 0.03 kPa. Because the catalytic activity is often strongly related to the specific surface area and particle size [19,20], the Ni/titania thin film catalysts which have small particle size and high dispersibility are expected to show high photocatalytic performance. Fig. 10 shows the photocatalytic stability against the irradiation time of Ni/titania catalyst, in comparison with of titania alone. In the case of titania, the photocatalytic activity
In this study, we propose a new nanoparticle preparation technique, the CVRD (chemical vapor reductive deposition) method, which could accomplish the formation of metallic Ni nanoparticles on the surface of a titania thin film substrate. The heterogeneous nucleation and the particle growth were well monitored and controlled through various investigations. The Ni nanoparticles successfully enhanced the photocatalytic activity of titania thin film. It is concluded that the CVRD method can be one of the best candidate techniques for preparing metallic nanoparticles on the supporting material, in particular as the first development in a nanoparticle–thin film preparation system. References [1] K.P. de Jong, Curr. Opin. Solid State Mater. Sci. 4 (1999) 55–62. [2] H. Takahashi, Y. Sunagawa, S. Myagmarjav, K. Yamamoto, N. Sato, A. Muramatsu, Mater. Trans. 44 (2003) 2414–2416. [3] S. Myagmarjav, H. Takahashi, Y. Sunagawa, K. Yamamoto, N. Sato, E. Matsubara, A. Muramatsu, Mater. Trans. 45 (2004) 2035–2038. [4] P. Burattin, M. Che, C. Louis, J. Phys. Chem. B 104 (2000) 10482–10489. [5] A. Jasik, R. Wojcieszak, S. Monteverdi, M. Ziolek, M.M. Bettahar, J. Mol. Catal. A 242 (2005) 81–90. [6] P. Li, J. Liu, N. Nag, P.A. Crozier, Surf. Sci. 600 (2006) 693–702. [7] Z. Bastl, J. Franc, P. Janda, H. Pelouchová, Z. Samec, Nanotechnology 17 (2006) 1492–1500. [8] J. Wang, H. Zhang, J. Ge, Y. Li, J. Phys. Chem. B 110 (2006) 807–811. [9] T. Nakamura, T. Ichitsubo, E. Matsubara, A. Muramatsu, N. Sato, H. Takahashi, Acta Mater. 53 (2005) 323–329. [10] B.R. Müller, S. Majoni, D. Meissner, R. Memming, J. Photochem. Photobiol. A 151 (2002) 253–265. [11] J.P. Espinòs, A.R.G. Elipe, A. Caballero, J. Garcìa, G. Munuera, J. Catal. 136 (1992) 415–422. [12] C.D. Wagner, W.M. Riggs, L.E. Davis, J.F. Moulder, Handbook of X-ray Photoelectron Spectroscopy, Physical Electronics Industries, Eden Prairie, MN, 1983. [13] J.C.D. Jesus, I. González, A. Quevedo, T. Puerta, J. Mol. Catal. A 228 (2005) 283–291. [14] L. Brissonneau, C. Vahlas, Chem. Vap. Deposition 5 (1999) 135–142. [15] G.J.C. Carpenter, Z.S. Wronski, J. Nanopart. Res. 6 (2004) 215–221. [16] V. Craciun, R.K. Singh, Appl. Phys. Lett. 76 (2000) 1932–1934. [17] M.C.J. Bradford, M.A. Vannice, Appl. Catal. A 142 (1996) 73–96. [18] J.S. Smith, P.A. Thrower, M.A. Vannice, J. Catal. 68 (1981) 270–285. [19] J.L. Carter, J.A. Cusumano, J.H. Sinfelt, J. Phys. Chem. 70 (1966) 2257– 2263. [20] T. Hayashi, K. Tanaka, M. Haruta, J. Catal. 178 (1998) 566–575. [21] T. Ihara, M. Miyoshi, Y. Iriyama, O. Matsumoto, S. Sugihara, Appl. Catal. B 42 (2003) 403–409. [22] K. Otsuka, S. Takenaka, Catal. Surv. Asia 8 (2004) 77–90.