Effect of Ag NPs on the radiant absorption of photocatalyst film

1.   Introduction

Noble metal nanoparticles (NPs), such as Ag, Au, and Pt, can harvest visible light effectively due to their Localized Surface Plasmon Resonance (LSPR). Incident irradiation is highly localized at NPs. Generated electrons and holes are efficiently separated at the metal-semiconductor interface due to Fermi level equilibration ^[1,2]. These properties can be applied for photocatalytic activity. Thus, noble metal NPs are combined with photocatalysts to form plasmonic photocatalysts ^[3,4]. As the radiant absorption of photocatalyst is the precondition for the photocatalysis reaction, it is necessary to investigate the radiative properties, including the spectral absorptivity and internal radiant absorption, for plasmonic photocatalysts. The simulated research in Ref. ^[5] have provided a comprehensive analysis for the radiative properties of hedgehog-like ZnO-Au composite particles, their particle models were established based on discrete dipole approximation (DDA) and the optimized morphology was summarized for the application of photocatalysis. Considering some assumptions and simplifications of the model, they have also developed experimental measurement for the hedgehog-like ZnO-Au composite particles ^[6].

It is reported that the Ag NPs in the Ag/TiO₂ system would undergo corrosion or dissolution during photocatalytic reaction ^[7,8]. Thus, a thin insulating shell is employed to protect the Ag NPs. Different types of insulating oxides are suggested, but electron transfer from metal NPs to the insulating shell does not occur ^[9,10]. When the charge carriers generated by LSPR-induced electromagnetic field are localized to the vicinity of the metal NPs, they cannot migrate to the active sites near photocatalyst surface, so the photocatalytic efficiency is just enhanced by the localized irradiation. Alumina (Al₂O₃) is an earth-abundant and stable metal oxide, which has been widely applied as a supporter for photocatalyst due to its low cost, chemical inertness and efficient photoinduced electron-hole separation ^[11,12,13]. In addition, dual layers composed of Al₂O₃ and TiO₂ have been experimentally proven to suppress the recombination rate of electron and hole pairs ^[14]. From the aspect of incident radiant absorption, the imaginary component of the Al₂O₃ dielectric constant is nearly zero at the wavelengths between 300 nm and 900 nm. Thus, Al₂O₃ can be considered as transparent for the incident irradiation in this wavelength region. So Al₂O₃ was chosen to protect the metal NPs from corrosion for designing and preparing stable and efficient plasmonic photocatalysts. Titanium dioxide (TiO₂) is a sustainable candidate material for photocatalysts due to its good stability, nontoxicity, high efficiency and low cost ^[15]. The band gap of anatase TiO₂ is 3.2 eV. This characteristic indicates that electrons and holes can be generated in anatase TiO₂ when the wavelength of incident radiation is less than 387.5 nm. The generated carriers can be used for catalytic reaction before the recombination of electron and hole pairs.

In this work, a composite film composed of Ag, Al₂O₃ and TiO₂ was proposed. The Ag NPs were immersed in the host medium of Al₂O₃, and the TiO₂ film was sputtered on the Ag/Al₂O₃ film. A theoretical model for the radiant absorption of Ag/Al₂O₃-TiO₂ film was established based on Finite Difference Time Domain (FDTD) method ^[16]. The simulations were implemented by commercial FDTD package to analyze the effect of Ag NP radiant absorption on the composite film. The local radiant absorption of TiO₂ film in the Ag/Al₂O₃-TiO₂ composite film was also analyzed.

2.   Radiant absorption model for Ag/Al₂O₃-TiO₂ composite film

The effect of Ag NPs on the TiO₂ film radiant absorption in the Ag/Al₂O₃-TiO₂ composite film is mainly related to the following factors: the Ag-TiO₂ distance and Ag NP size, distribution and morphology. The thicknesses of Al₂O₃ film and TiO₂ film could be controlled during the preparation and measured through ellipsometer method (Semilab R & D). The absorptance of the composite film was measured using an UV-VIS spectrophotometer (Evolution 220).

The radiant absorption model for Ag/Al₂O₃-TiO₂ composite film was established based on electromagnetic theory. The electromagnetic field in irradiated composite film could be described by Maxwell equations. The Finite Difference Time Domain (FDTD) method was applied to solve the equations. (see Supplementary Information Section S2). As the morphology, size and distribution of Ag NPs in the prepared composite films cannot be regulated precisely, the simulated results for the Al₂O₃-TiO₂ composite film without Ag NPs were compared with the measurable absorptance of the prepared Al₂O₃-TiO₂ films with the same layer thickness. The optical constants of Al₂O₃ film and TiO₂ films, which were needed for the simulation, were also investigated through ellipsometry method (Semilab R & D). The measured and simulated data shown in Figure 1 were both transformed into the same unit. The FDTD model for Ag NPs was verified based on Mie theory (Figure 2). The results in Figure 2 were for the spherical Ag NPs with different radius R at the wavelength of 413.3 nm. The Relative Errors (RE) between them were calculated as follows:

where x is the simulated result, $\mu$ is the measured result or the result calculated from Mie theory.

The data in Figure 1 indicated that simulated results were close to the measured results. The data in Figure 1 indicated that the REs between simulated and measured results increased with the wavelength. The differences were probably due to measurement errors, the FDTD model's numerical error and the simplified Al₂O₃-TiO₂ interface structure in the simulated model. The data in Figure 2 indicated that the REs between the FDTD model and Mie theory increased with particle's Size Parameter. The differences could arise from numerical error.

The model for analyzing the radiant absorption of the Ag/Al₂O₃-TiO₂ composite film was shown in Figure 3. The main objective of this work was to investigate the influence of Ag NPs on the radiant absorption of the Ag/Al₂O₃-TiO₂ composite film, especially on the radiant absorption of the TiO₂ film. A single layer of dispersed Ag NPs was assumed to be immersed in the host medium of Al₂O₃. The TiO₂ film with thickness of H₁ was located on the surface. Figure 3(a) presented a sectional view of Ag/Al₂O₃-TiO₂ composite film. H₂ was the distance between Ag and TiO₂. D was the diameter of Ag NPs, and L was the distance between two adjacent Ag NPs. The parameters and conditions of the model were set based on confirmation with experimental or analytical results, which were introduced at the beginning of this section. The unit cell for the FDTD modeling was sketched in Figure 3(b). The unite cell was periodic along the X and Y coordinate axes. The boundary conditions at the Z coordinate axis were set as the perfectly matched layer, which was an absorbing boundary. The incident source was set as a monochromatic plane wave propagating along the negative Z coordinate axis shown as an arrow in Figure 3(b). The incident polarization shown as a double-headed arrow in Figure 3(b) was along the X coordinate axis.

The radiant absorption model for the Ag/Al₂O₃-TiO₂ composite film was constructed based on electromagnetic theory (see Supplementary Information Section S2). The radiant absorption of Ag/Al₂O₃-TiO₂ composite film and TiO₂ film could be obtained using Eqs (2) and (3), respectively.

where ${A_\lambda }$ was the spectral absorptivity of the Ag/Al₂O₃-TiO₂ composite film at the wavelength of $\lambda$, ${r_\lambda }$ was the spectral reflectivity gained in the plane R, ${t_1}_\lambda$ was the spectral transmissivity gained in the plane T₁, ${a_\lambda }$ was the spectral absorptivity of the TiO₂ layer in the composite film, and ${t_2}_\lambda$ was the spectral transmissivity gained in the plane T₂.

3.   Effect of Ag NPs on the radiant absorption of TiO₂ in the Ag/Al₂O₃-TiO₂ film

3.1.   Influence of distance between Ag NPs and TiO₂

The influence of Ag-TiO₂ distance on the spectral absorptivity of Ag/Al₂O₃-TiO₂ composite film was analyzed using the radiant absorption model. The thickness (H₁) of the TiO₂ film was set as 60 nm. The distance (L) between two adjacent Ag NPs was set as 30nm. The diameter (D) of Ag NP was set as 20 nm. The distance (H₂) between Ag and TiO₂ was set at the value listed as "H₂-n" in Table 1. The Cases of TiO₂ film with the thickness of 60 nm and Ag/Al₂O₃ composite film without TiO₂ were also simulated. The results for the cases listed in Table 1 are compared in Figure 4, where Figure 4(a) shows the spectral absorptivity (${A_\lambda }$) of Ag/Al₂O₃-TiO₂ composite film and Figure 4(b) shows the spectral absorptivity (${a_\lambda }$) of the TiO₂ layer in the Ag/Al₂O₃-TiO₂ composite film. The radiant absorption of TiO₂ was intense in the short wavelength region. The value of the spectral absorptivity for the Ag/Al₂O₃-TiO₂ composite film decreased with the increasing wavelength. The main reason for this decline was the reduced absorption of the TiO₂ film. When the incident wavelength exceeded 300 nm, the effect of LSPR from Ag NPs became apparent. The spectral absorptivity for the Ag/Al₂O₃-TiO₂ composite film reached a small peak near 350 nm and the main peak near 530 nm. The value of spectral absorptivity increased with the increasing Ag-TiO₂ distance around the 350 nm peak and decreased with the increasing Ag-TiO₂ distance around the 530 nm peak. The influence of Ag-TiO₂ distance on the spectral absorptivity of the TiO₂ layer in Ag/Al₂O₃-TiO₂ composite film had the similar tendency as that shown in Figure 4(b).

However, the increased spectral absorptivity of the TiO₂ layer was considerably less than that of the Ag/Al₂O₃-TiO₂ composite film at the 350 nm peak and 530 nm peak mainly because the incident irradiation was highly localized at Ag NPs. Given that spectral absorptivity is proportional to the square of the local electric field in the composite film, the distribution of spectral absorptivity could be analyzed through the electric field maps. The maps in Figure 5 shows the internal relative electric-field intensity for Ag/Al₂O₃-TiO₂ composite film with the Ag-TiO₂ distance of 2 nm. The values shown in the maps are the internal electric-field intensity compared with incident electric-field intensity, which could be described by E/E₀. The internal electric field at the wavelengths below 387.5 nm can facilitate charge separation. The radiant absorption in the spectral range above 387.5 nm can generate heat in the composite film. Then temperature is increased in some localized regions, which may influence the photocatalytic performance. Based on these mechanisms, the internal electric field distributions at the two wavelengths of 550 nm and 370 nm were analyzed in Figure 5(a) and (b) respectively. The maps in Figure 5, which show the relative electric-field distribution in T₂ plane, indicate that the localized electric field in the TiO₂ cross section near the interface of TiO₂-Al₂O₃ could reach to 1.6 times the incident electric-field intensity at the wavelength of 550 nm and 0.78 times that at 370 nm. The simulated results for the case without Ag NPs showed that the maximum electric-field intensity in the T₂ plane was 0.616 times the incident electric-field intensity at wavelength of 550 nm and 0.651 times that at 370 nm. Incident radiation declined because of the absorption of TiO₂.

The photocatalytic activities of the Ag/Al₂O₃-TiO₂ film, which were characterized based on their decomposition of methylene blue under the irradiation of HPMV, showed that the composite film with low Ag-TiO₂ distance possessed high catalytic activity (Supplementary Information). It was indicated that the generated electrons and holes in this composite film could be efficiently used for catalytic reaction. The simulated results presented in this section showed that the TiO₂ layer in the composite film with wide Ag-TiO₂ distance exhibited high spectral absorptivity in the incident wavelength region of 280 nm to 400 nm. Thus, the increased radiant absorption in the TiO₂ layer could not directly trigger the enhancement of the photocatalytic activity. Tuning the spots of the localized incident irradiation to the catalytic sites is important. Then the photocatalytic activity of the composite film is effectively enhanced. In the prepared composite film with low Ag-TiO₂ distance, the catalytic sites were close to the neighborhood of Ag NPs where the enhanced electric field was triggered. Thus, these composite films exhibited enhanced catalytic activities. The distribution of internal radiant absorption was needed to analyze for the optimal design of composite catalysts.

3.2.   Influence of Ag NP morphology

In nanotechnology, many methods have been used to produce metal NPs, including chemical synthesis, light-induced rapid annealing et al, the NPs morphology can be modified by carefully controlling the synthetic condition. Then the metal NPs morphology can be optimized for their potential applications in plasmonic photocatalyst. In this part, the influence of Ag NP morphology was also analyzed through the constructed theoretical model sketched in Figure 4(b). Five different nanostructures considered for Ag NPs are shown in Figure 6. These nanostructures were spherical morphology, pyramid structure, cylinder, hollow cylinder (also called ring) and cube. The heights of these nanostructures were set at the same value of 40 nm. The underside of the pyramid structure was a square with the side-length of 40 nm. The underside of the cylinder was a circle with the radius of 40 nm. The underside of the hollow cylinder was a circular ring with the outside diameter of 40 nm and the inside diameter of 20 nm. The diameter of the sphere and the side-length of cube were both set at the value of 40 nm. The thickness H₁ of the TiO₂ film was set at 60 nm, and the distance H₂ between Ag and TiO₂ was 2 nm. Given that the distance L between two adjacent Ag NPs was 120 nm, so the interaction between the Ag NPs was sufficiently weak. Other parameters in this simulation were the same as those in section 3.2.2. The results for the five morphologies are compared in Figure 7, where Figure 7(a) shows the spectral absorptivity (${A_\lambda }$) of the Ag/Al₂O₃-TiO₂ composite film, and Figure 7(b) shows the spectral absorptivity (${a_\lambda }$) of the TiO₂ layer in the Ag/Al₂O₃-TiO₂ composite film. The radiant absorption of the TiO₂ layer played the main role in the wavelength reigon of 280~320 nm. Thus, the influence of Ag NPs morphology was not obvious in the wavelength region of 320~800 nm, the peaks of the composite films were located at different wavelengths because of the Ag NP morphology. The composite film with spherical Ag NPs had a main peak at 510 nm. The composite film with circular ring Ag NPs had a main peak at 650 nm and secondary peak at 360 nm. The spectral absorptivity of the composite film with cylindrical or cubic Ag NPs was also enhanced, and the peak wavelengths of the composite films with cylindrical and cubic Ag NPs were 575 nm and 630 nm respectively. The spectral absorptivity peak of the composite films with pyramidal Ag NPs was considerably less than those of other composite films in the wavelength region of 280~320 nm. The main reason for the result was the low absorptivity of pyramidal Ag NP under the irradiance from summit to underside of pyramidal Ag NP.

The spectral absorptivity of TiO₂ layers in the composite film with and without Ag NPs are compared in Figure 7(b). In the wavelength region of 320~380 nm, the enhanced radiant absorption intensified because of the LSPR effect from Ag NPs. The plasmonic effect of Ag nanocube was stronger than that of the other samples in this wavelength region because of the secondary plasmonic peak near 360 nm. When the wavelength was in the 400~700 nm region, the spectral absorptivity of the TiO₂ layer had the highest peak under the influence of cylindrical Ag NPs, and it had a wide enhanced region under the influence of cubic Ag NPs. The total effects of the five morphologies in the 400~700 nm region could be compared by using the following expression

where ${a_i}(n)$ was for spectral absorptivity of TiO₂ layers in the composite film with Ag NPs, subscript i represented each morphology, and ${a_T}(n)$ represented the spectral absorptivity of TiO₂ layers in the composite film without Ag NPs. The calculated results indicated that the plasmonic effect of Ag nanocylinders was stronger than that of the other samples in the wavelength region of 400–700 nm. The main reason for the result was the Ag nanocylinder enhanced radiant absorption in TiO₂ layer exist in a wider wavelength region.

Although the radiant absorption of the TiO₂ layer was considerably stronger below the wavelength of 330 nm, the incident radiation (e.g., solar irradiance) usually focused in the wavelength range from visible to near-infrared. The LSPR effect of Ag NPs was intense in the wavelength region of 400–700 nm. Thus, the electric field maps of the composite films were analyzed at the wavelength of main peak. The maps in the X-Z plane through the center of Ag NPs in Figure 8 showed the internal relative electric-field intensity for the composite film. The values in the maps also showed the internal electric-field intensity E compared with the incident electric-field intensity E₀. The electric field was highly enhanced around Ag NPs, especially near its angles. The relative electric-field intensity (E/E₀) could reach 70 in cylindrical Ag NPs. Given the highly localized internal electric field near the Ag NP surface, their distribution could not be identified in the TiO₂ layer as shown in Figure 8. The internal relative electric-field intensity maps in the TiO₂ layer are shown in Figure 9 at the peak wavelength, Figure 9(a) shows the X-Z section crossing the centers of Ag NPs and Figure 9(b) shows the X-Y cross section near the Al₂O₃-TiO₂ interface. The relative electric-field intensity in the TiO₂ layer was enhanced in different degrees and domains because of the difference in the Ag NP morphology. Among the five morphologies, Ag NPs with cylindrical morphology or hollow cylindrical morphology triggered wider and stronger internal radiant absorption in the TiO₂ layer. This is because most absorbed radiant energy in Ag NPs was localized near the surface with bigger curvature. In the composite films with cylindrical or hollow cylindrical Ag NPs, these localized regions were wider near TiO₂ layer.

4.   Conclusions

The photocatalytic activity of TiO₂ film was promoted through combination with plasmonic Ag NPs. Al₂O₃ was proposed as the protective medium for Ag NPs. Thus, a thin layer of Al₂O₃ was placed between Ag NPs and the TiO₂ film. The degradation of aqueous methylene blue under the incident irradiation of HPMV indicated that the Ag/Al₂O₃-TiO₂ composite film with the Ag-TiO₂ distance of 2 nm or 5 nm exhibited high degradation rate. One possible reason for this phenomenon was that, in the composite film with low Ag-TiO₂ distance, the metal-semiconductor interface where photoinduced charge carriers were efficiently separated and the location where radiant absorption was localized were near the catalytic sites. The simulated results verified that the incident radiation was highly localized in the surrounding of Ag NPs.

The photocatalytic activity of Ag/Al₂O₃-TiO₂ composite film was related to several factors. This work was focused on the effect of Ag NPs on the radiant absorption of the photocatalyst film. The influences of Ag-TiO₂ distance and Ag NP morphology on the spectral absorptivity and radiant absorption distribution were systematically studied through FDTD simulation. The simulated results indicated that the nanostructure of the composite film could be regulated to an optimal case wherein spectral absorptivity and radiant absorption distribution were both beneficial to improve the photocatalytic activity.

Acknowledgment

This work was supported by National Natural Science Foundation of China (Grant No.51476078).

Conflict of interest

The authors declare that there are no conflict of interest in this paper.

Author contributions

Li Jiayu: Conceptualization, Methodology, Writing—review & Editing the manuscript, Funding acquisition. Guo Jiewen: Investigation, Modeling, Data analysis.