III. RESULTS AND DISCUSSION

Similar to our previous report,^{Reference Li, Pan, Li, Liu, Wu and Peng17} the Ag NPs on the fresh films show the irregular spherical shapes of isotropic growth, irrespective of the AgNO₃ concentration, which ranged from 200 to 3200 mg/L (1.2–19.2 mM) in this study. Figure 1(a) depicts a typical SEM image and the corresponding size distribution of Ag NPs. The sizes of the Ag NPs were defined as the square root of the cross-sectional area of the Ag NP SEM images. For comparison, the SEM image of the bare TiO₂ film is also presented in the inset at the bottom-left corner of Fig. 1(a). We found that the asymmetric distribution of the Ag NPs' size can be well fitted with the extreme function, y = Aexp(−exp(−z) + z + 1), where y is the percentage of the Ag NPs with the size of x, z = (x − x _c)/σ, x _c Ag NP feature size at the most probable value, and σ is the standard deviation. The feature size (x _c) is ∼80 nm at 5 min growth time, as shown in the inset at the top-right corner of Fig. 1(a). With the increase in growth time (t), the feature size is on the increase, as shown in Fig. 1(b). The average size (d) of Ag NPs fit the relation ${d^3} - d_0^3 = kt$, indicating that OR mechanism dominates the crystal growth of Ag NPs.^{Reference Viswanatha, Santra, Dasgupta and Sarma18} Such Ag–TiO₂ hybrids have been reported in the photocatalytic synthesis of Ag NPs and adopted in the studies related to LSPR-assisted TiO₂ photocatalysis and SERS.^{Reference Li, Pan, Li, Liu, Wu and Peng17,Reference Mills, Hill, Stewart, Graham, Smith, Hodgen, Halfpenny, Faulds and Robertson19–Reference Sudnik, Norrod and Rowlen23}

FIG. 1.

Typical SEM image (a) and the size distribution (b) of Ag NPs grown on fresh TiO₂ films in the AgNO₃ solution of 3200 mg/L. The insets at the bottom-left and top-right corners of Fig. 1(a) are the SEM image of the bare TiO₂ film and the size distribution of Ag NPs at the growth time of 5 min, respectively.

Under the same growth condition as the case in Fig. 1, the growth behavior of Ag NPs on the oil-decorated TiO₂ films was varied, as shown in Fig. 2. In the initial growth stage (at 1 min), Ag NPs show sphere-like shapes that are similar to those grown on the fresh films, but their sizes are smaller. By increasing the growth time to 2 and 4 min, the anisotropic Ag NPs, or the small Ag nanoplates, appear in the products. At 8 min growth time, the well-defined Ag nanoplates, upright standing or tiled, were popularly observed in the product. The formation of Ag nanoplates is likely due to the aggregation of the primary particles (nuclei) in the nucleation stage. As indicated by the circles in the figure, the spherical Ag NPs in some regions aggregate together or form Ag nanochains, as shown in Figs. 2(a) and 2(b), and then the small Ag nanoplates are gradually developed to be the perfect regular triangles or hexagons, as shown in Figs. 2(c) and 2(d). Such nanoplates have three major LSPR modes, namely, out-of-plane mode and in-plane longitudinal and transverse modes, and respond selectively to light polarized in a specific angle. Thereby, the Ag nanoplates can be applied to the fields such as stereoscopic visions, storing multiple information, certification, and anticounterfeit technologies and polarizers.^{Reference Sakai, Tanabe and Tatsuma24,Reference Tanabe, Matsubara, Sakai and Tatsuma25} Using the upright standing and the tiled Ag nanoplates, the thickness of the Ag nanoplates was estimated to be 30–40 nm, whereas the lengths of their edges ranged from 0.1 to 1 μm. TEM analysis reveals that the Ag nanoplates are faced with (111) planes. The above results indicate that the oil-decorated TiO₂ films successfully induce the anisotropic growth of Ag NPs.

FIG. 2.

(a–d) SEM images of Ag NPs grown on the oil-decorated TiO₂ films in the AgNO₃ solution of 3200 mg/L at the given growth times. The insets are the enlarged SEM images of the encircled regions in the figures.

Figure 3 shows the size distributions of the Ag NPs on the oil-decorated films at the growth times of 1, 2, 4, and 8 min. Initially, the size distribution of the Ag NPs on the oil-decorated films is similar to that of the Ag NPs on the fresh films, as shown in Fig. 3(a). With an increase in the growth time, the feature size of the distribution does not obviously increase, but the percentage of large Ag NPs increases considerably, forming a distribution tail, as shown in Figs. 3(b)–3(e). This type of evolution is much different from that of Ag NPs on the fresh films. When compared with the SEM images in Fig. 2, the feature sizes were found to be determined by the distributions of irregularly spherical Ag NPs in the products, and the appearance of the Ag nanoplates results in the formation of distribution tails. Because the feature size was not increased by increasing the growth time, we conclude that the isotropic growth of Ag NPs on the oil-decorated films is restrained. The change in growth behavior was also observed in the in situ extinction spectra, as shown in Fig. 3(f). The extinction intensities in the visible region slowly increase by increasing the growth time, whereas the extinction in the NIR region is quickly enhanced. The small peak at ∼340 nm, which was first observed at 8 min, is due to the out-of-plane quadrupole resonance of the Ag nanoplates in AgNO₃ solution, and the peak position is associated with the thickness of the Ag nanoplates.^{Reference Jin, Cao, Hao, Me, Schatz and Mirkin13} According to the effective medium theory,^{Reference Moores and Goettmann26,Reference Ung, Liz-Marza and Mulvaney27} isotropic Ag NPs contribute to the extinction in the visible region and Ag nanoplates give rise to an enhancement of the extinction in the NIR region. Thereby, we also conclude that the oil-decorated films play a role in restraining the isotropic growth of the Ag NPs.

FIG. 3.

(a–e) Size distribution of Ag NPs grown on the oil-decorated TiO₂ films in the AgNO₃ solution of 3200 mg/L and (f) the corresponding extinction spectra.

The Ag nanoplates in the product were found to be dependent on the oil-decoration time of the TiO₂ film. Figure 4 shows the SEM images of the Ag NPs grown on the TiO₂ films with their corresponding time of oil decoration. The 3200 mg/L AgNO₃ solution was used, and the growth time was controlled at 5 min. The number of Ag nanoplates increases with an increase in the oil-decoration time until 4 min, and then no obvious difference is observed. In addition, we found that a shorter time of oil decoration leads to the production of smaller Ag nanoplates. Besides the oil-decoration time, the AgNO₃ concentration is found to be important in determining the formation of Ag nanoplates on the oil-decorated films. Figure 5 shows the SEM images of Ag NPs grown on the oil-decorated films in the 1600, 800, 400, and 200 mg/L AgNO₃ solutions, in which Ag NPs were grown in solution for 6 min and the oil-decoration time of TiO₂ film was 4 min. Ag nanoplates cannot be synthesized in the AgNO₃ solution below 200 mg/L. In AgNO₃ solutions with concentration >400 mg/L, the density of the Ag nanoplates increases with an increase in the AgNO₃ concentration, indicating that a critical AgNO₃ concentration is required for the formation of Ag nanoplates on the oil-decorated films. In this study, the critical AgNO₃ concentration is between 200 and 400 mg/L. We found that the critical concentration is necessary to induce the formation of Ag nanoplates on the oil-decorated films, irrespective of the decoration time.

FIG. 4.

SEM images of Ag NPs grown on the films with oil decoration of 1, 2, 4, and 8 min, with the use of 3200 mg/L AgNO₃ solution at 5 min growth time.

FIG. 5.

SEM images of Ag NPs grown on the oil-decorated films in the AgNO₃ solutions of 1600, 800, 400, and 200 mg/L; the growth time is 6 min and the time of oil decoration is 4 min.

To characterize the surface change, the water contact angles of the films were measured. The contact angle of water on the fresh film is ∼6°, and it increases to 12°, 23°, 27°, and 33° for the films with oil decoration of 1, 2, 4, and 8 min, respectively, evidencing that the surface properties of the films were modified by oil decoration. Because oil molecules are easily desorbed in ultrahigh vacuum at room temperature, they cannot be characterized with the conventional surface analysis methods, such as x-ray photoelectron spectroscopy and Auger electron spectroscopy. Therefore, SERS was used to estimate the number of oil molecules on the films. To be reliable, three kinds of SERS samples, namely, oil-decorated Ag NPs on fresh film (S1), Ag NPs on oil-decorated TiO₂ film (S2), and Ag NPs on fresh film (S3), were prepared for the Raman spectrum comparison with the pump oil in capillary. Sample S1 was made by evaporating the pump oil onto Ag NPs grown on fresh films. Samples S2 and S3 were Ag NPs grown on TiO₂ films with and without oil decoration, respectively. As shown in Fig. 6, the Raman peaks of C–H bonds in the range of 2800–3000 cm⁻¹ were detected for the samples with oil molecules (S1 and S2), but no Raman signal was detected for the oil-free sample (S3). Thus, the detected Raman signals are attributed to the oil decoration.

FIG. 6.

Raman spectra of the pump oil in capillary and those taken from the samples of the TiO₂ films with and without oil decoration as indicated in the figure.

As all organic solvents exhibit the same Raman peaks of C–H bonds as the oil, the number of oil molecules on the films can only be estimated using SERS enhanced factors (EF) of the Ag–TiO₂ film, which was determined using a certain concentration of rhodamine 6G as the probe. In this study, 1 μM rhodamine 6G ethanol solution was used. Assuming that the determined SERS EF was valid for the oil-decorated samples, we obtained 10¹⁵–10¹⁶ cm⁻² of C–H bond density using the Raman intensity of the pump oil in capillary as the reference. Because such an estimation depends on the probe concentration, photodegradation of methylene blue, which is discussed in the following text, was also used for the estimation by supposing the photodegradation rate of oil molecules on TiO₂ film to be the same as that of methylene blue in solution, and ∼10¹⁶ C–H/cm² per minute of oil deposition rate was obtained. The two methods resulted in approximately the same conclusion that a sublayer to a few layers of oil molecules was deposited on the TiO₂ film after oil decoration.

The above results prove that the oil-decorated films are capable of promoting the anisotropic growth of Ag NPs in the photocatalysis, and the oil layer on the films is responsible for the formation of Ag nanoplates. However, obtaining direct evidence to reveal the oil-layer function in the formation of Ag nanoplates is difficult because the growth of Ag NPs is a dynamic process, and no available technique can be used for an in situ observation of the Ag NP growth. Based on our observation, nucleation theory, and photocatalysis principle, a simple discussion on the growth of Ag NPs on the oil-decorated films is presented.

First, two kinds of Ag NPs, isotropic NPs (∼60 nm size) and anisotropic nanoplates (30–40 nm thickness), were obtained using the oil-decorated films. However, the initial nucleation of Ag NPs on the oil-decorated films is similar to that on the fresh film except that the size of the primary particle is smaller. According to the nucleation theory of crystal growth,^{Reference Ohring28} high supersaturation of the neutral atom leads to the small size and high density of the primary particles (nuclei). Thereby, the formation of Ag nanoplates is related to the increase in the density of Ag atoms. In other words, the reduction rate of Ag atoms on the oil-decorated films is likely higher than that on the fresh film. The critical AgNO₃ concentration for the formation of Ag nanoplates also suggests that the anisotropic growth needs a high Ag reduction rate, which is proportional to the concentration of Ag ions in solution. The increase in the Ag reduction rate can be attributed to the efficient charge separation of photocatalysis due to the introduction of the oil layer on TiO₂ films. As is well-known, two kinds of photoinduced carriers, electron and hole, are generated in the TiO₂ film and are transferred to its surface. The electron reduces the ions in solution and the hole oxidates the neutral particles into ions. Thus, the reoxidation of the reduced Ag atoms decreases the Ag reduction rate. We believe that the oil layer on the TiO₂ film is the substitute for the reduced Ag atoms. Because the Ag atoms avoid reoxidation with the oil layer, the net reduction rate of the Ag atoms is increased.

To demonstrate our assumption, the photodegradation of 10 mg/L methylene blue aqueous solution was compared using three samples, namely, the fresh film, the oil-decorated TiO₂ film, and a bare glass, as shown in Fig. 7. The fresh film exhibits the highest photodegradation rate. After oil decoration, the photodegradation is considerably reduced, indicating that the oil layer prevents the oxidation of methylene blue. The glass sample shows the lowest photodegradation rate of methylene blue, which can be attributed to the UV-induced photodegradation. The experiments clearly demonstrated that the oil layer is capable of preventing oxidation of the Ag atoms in the growth of Ag NPs, leading to the increase in the density of Ag atoms. An interesting phenomenon in using the oil-decorated films is that the photodegradation rate sharply increased at time > ∼15 min, which is likely the time when the oil layer was consumed due to the photodegradation.

FIG. 7.

Photodegradation curves of methylene blue aqueous solution obtained by using the samples of fresh TiO₂ film, oil-decorated TiO₂ film, and bare glass. The initial solution concentration (C ₀) was 10 mg/L, and the concentration (C) of methylene blue in photodegradation was determined by using the absorption spectra collected. A low-pressure mercury lamp was used as the irradiation source for the photodegradation.

Second, we found that most of the Ag nanoplates pile up over each other or lay over the irregularly shaped Ag NPs, as shown in Figs. 4(c) and 4(d). In addition, these Ag nanoplates are mobile and easily stripped from the substrates. These observations suggest that the formation of Ag nanoplates likely took place in the area adjacent to the film surface, and they adhered to the surface when they were taken out from the solution. In other words, the Ag nanoplates probably originated from the self-nucleation of Ag NPs in the solution, but not from the heterogeneous nucleation on the TiO₂ film. However, studies involving the self-nucleation or heterogeneous nucleation of Ag NPs via photocatalytic reduction are unavailable. On the basis of the given observation, we suggest a scenario for the growth of Ag NPs via photocatalytic reduction with TiO₂ film, as shown in Fig. 8. Considering Ag atom mobility and reoxidation, a specific Ag atom distribution is suggested, as shown in Fig. 8(a). Depending on the nucleation free energy and the supersaturation, self-nucleation in the solution and heterogeneous nucleation on the film surface are both possible, as shown in Fig. 8(b). After nucleation, the Ag NPs grow through two mechanisms, OR and OA. The OR mechanism often leads to an isotropic growth, whereas the OA mechanism results in the formation of nanoplates, nanorods, and nanosheets, etc. We believe that the formation of the Ag nanoplates in this study is likely related to the OA growth.

FIG. 8.

Photocatalytic growth model of Ag NPs in AgNO₃ solution with TiO₂ films (a) reduction of Ag atoms, (b) nucleation of Ag primary particles, and (c) formation of Ag NPs or nanoplates.

The OA mechanism is a particle-mediated mesoscopic transformation process,^{Reference Niederberger and Cölfen29,Reference Zhang, Huang and Lin30} and has been demonstrated to be one of the important mechanisms leading to the anisotropic growth of Ag NPs in solution-phase synthesis. Yin et al.^{Reference Yin, Huang, Zhang, Zheng and Lin31} found that both OA and OR mechanisms coexist in the growth of NPs. At the initial stage, increasing the concentration of the primary particles facilitates the OA-based growth, and the OR growth is restrained. Our observation with regard to the growth of Ag NPs is consistent with that by Yin et al. The HRTEM observation also implies that the OA growth is likely responsible for the formation of Ag nanoplates on the oil-decorated films. Figure 9 shows two typical images of Ag NPs in contact, which were stripped from the sample of Ag–TiO₂ hybrids at 8 min growth time. The contacted Ag NPs exhibit the lattice spacings of 0.24–0.25 nm, which are closed to the spacing of Ag (111) plane (0.2361 nm). Similar Ag NPs in contact were never observed on the fresh films. In the in situ HRTEM study of crystal growth controlled by the OA mechanism in solution-phase synthesis,^{Reference Li, Nielsen, Lee, Frandsen, Banfield and De Yoreo32} Li et al. took a series of TEM images of the NPs in contact, which are interpreted as the particles undergoing continuous rotation and interaction before they find a perfect match.

FIG. 9.

Typical HRTEM images of contacted Ag NPs grown on the oil-decorated TiO₂ films in 3200 mg/L AgNO₃ solution.

The proposed photocatalytic growth model helps us to understand the growth behavior of Ag NPs on TiO₂ films, and provide information on developing a more effective method to control the growth of Ag NPs. For example, we have tried to use TiO₂ films that were kept in air for weeks to grow Ag NPs, and a large amount of triangular and polygonal Ag NPs were obtained. This result suggests that the surface properties of TiO₂ films changed in a high-moisture environment. However, the surface-capping agents, such as polyvinylpyrrolidone (PVP), polyethylene glycol (PEG), sodium citrate, etc., were found to be useless in controlling the anisotropic growth of Ag NPs on TiO₂ films, possibly due to their high water solubility.