Plasmonic gold-embedded TiO₂ thin films as photocatalytic self-cleaning coatings

Highlights

• Colloidal plasmonic Au nanoparticles are made compatible with TiO₂ sol-gel process.

• One-pot coating suspension contains both TiO₂ precursor and plasmonic Au particles.

• Homogeneous Au-embedded TiO₂ coatings are obtained that are highly transparent.

• Up to 40 % photocatalytic efficiency improvement is measured under solar light.

• 1 wt% Au is an economically favorable metal loading with good enhancement.

Abstract

Transparent photocatalytic TiO₂ thin films hold great potential in the development of self-cleaning glass surfaces, but suffer from a poor visible light response that hinders the application under actual sunlight. To alleviate this problem, the photocatalytic film can be modified with plasmonic nanoparticles that interact very effectively with visible light. Since the plasmonic effect is strongly concentrated in the near surroundings of the nanoparticle surface, an approach is presented to embed the plasmonic nanostructures in the TiO₂ matrix itself, rather than deposit them loosely on the surface. This way the interaction interface is maximised and the plasmonic effect can be fully exploited. In this study, pre-fabricated gold nanoparticles are made compatible with the organic medium of a TiO₂ sol-gel coating suspension, resulting in a one-pot coating suspension. After spin coating, homogeneous, smooth, highly transparent and photoactive gold-embedded anatase thin films are obtained.

Introduction

When TiO₂ is used in self-cleaning applications, it is usually deposited as a thin film that is capable of removing pollutants from the surface using (sun)light. The film thickness typically ranges from a few nanometres to several micrometres [[1], [2], [3]]. TiO₂ thin films can be deposited by a wide variety of methods, ranging from target sputtering over sol-gel deposition, thermal methods, vapour deposition techniques (PVD/CVD/ALD), anodic oxidation and plasma-assisted deposition processes [[4], [5], [6], [7], [8], [9], [10], [11]]. In this work, the sol-gel method is selected because it is straightforward, affordable and it can be applied directly onto a wide variety of substrates using easy and low-cost techniques such as spin or dip coating [3]. Furthermore, this simple yet versatile sol-gel process is an ideal method to obtain thin film coatings of high transparency, which is a crucial parameter when applying self-cleaning films onto glass surfaces. One of the main constraints of photocatalytic self-cleaning materials on the market, is their limited solar-light response. Photocatalytic films based on unmodified TiO₂ as the active ingredient are only activated by the UV-component of solar light, and thus only ca. 4–5 % of the incident solar light spectrum is actually used effectively [12,13].

To further improve the photocatalytic self-cleaning activity under solar light, we propose to modify the films by plasmonic nanoparticles embedded in the active layer. Surface plasmon resonance (SPR) is induced by incidence of photons of a resonant wavelength, which results in the collective oscillation of free electrons in the conduction band. This unique optical property of (noble) metal nanoparticles allows the concentration and manipulation of light at the nanoscale, which has resulted in different applications such as sensors, energy production, surface enhanced Raman spectroscopy and photocatalytic environmental remediation [12,14]. In contrast to traditional methods, where modification with plasmonic nanoparticles occurs by simply mixing, depositing or (photo-) impregnating them on the surface of the photocatalyst in a multi-step process [[15], [16], [17], [18], [19], [20], [21], [22]], here, the nanoparticles are first stabilised and mixed with the titanium precursor sol. Hence, this procedure provides a ‘one-pot’ coating suspension that can be readily applied. This is expected to facilitate a homogeneous dispersion and stable embedment of the plasmonic nanoparticles throughout the resulting TiO₂ matrix. The additional advantage of fully embedding the nanoparticles into the photocatalytic TiO₂ film can be rationalised as follows: (i) under UV light irradiation electron-hole pairs are generated in the TiO₂ semiconductor, while the metal nanoparticles solely act as passive electron sinks, resulting in less recombination events [23]. (ii) On the other hand, previous research has also shown that for Au/TiO₂ composites under visible light illumination, direct “hot electron” injection from the excited plasmonic nanoparticles into the conduction band of TiO₂ plays an important role [24,25], as well as an increase of the electromagnetic near-field within a 3 nm radius surrounding the plasmonic nanoparticle, as investigated by Asapu et al. [26,27]. Regardless of the excitation wavelength, maximising the direct contact interface between the nanoparticle and the semiconductor is of paramount importance for both of these electron and energy transfer processes. As a consequence, by partially or fully embedding the nanoparticles in the TiO₂ matrix, the contact interface is increased substantially and may lead to considerably higher photocatalytic activities (Fig. 1) [28,29]. In addition, partially or fully embedding the nanoparticles in a rigid matrix is hypothesised to protect them from chemical corrosion, reshaping, agglomeration and detachment during post-treatment or photocatalytic testing, thereby adding to the stability of the entire composite system [30,31].

A vast amount of literature can be found on several synthesis procedures enabling the preparation of nanoparticles of different compositions, shapes and sizes, whether or not functionalised with specific ligands. In the proposed approach, the plasmonic metal nanoparticles are first prepared separately after which they are introduced in the sol-gel precursor solution (“ex-situ” method). This entails several advantages over an “in-situ” approach, where a metal salt precursor and TiO₂ precursor are mixed, simultaneously forming metal nanoparticles and TiO₂ [[32], [33], [34], [35]]. In other words, the ex-situ method enables much more accurate and precise control over the nanoparticle properties, tailored to the envisaged application [36]. In-situ synthesis procedures on the other hand only enable limited control over the nucleation process and final nanoparticle size and shape [33]. The only difficulty that lies in the proposed strategy is that the plasmonic metal nanoparticles (typically prepared in an aqueous medium) have to be made compatible with the titania precursor coating sol (typically organic solvent based). One possible solution to this problem is proposed by Sonawane et al. [37]. It involves complete condensation of a Ti-alkoxide precursor followed by re-dissolution in aqueous hydrogen peroxide to render it compatible with the aqueous nanoparticle colloids. In our work, the use of such harsh conditions is avoided by achieving the opposite: adapting the colloids to the coating sol. A similar strategy has been investigated by the group of Mulvaney, with the eye on gas-sensing applications [33]. The metal loadings used in that particular study are, however, quite high (ranging from 4 to 8 wt%) and no photocatalytic activity data have been collected. In the present work, thin TiO₂ films are modified with gold nanoparticles up to 3 wt% loading. The resulting effect on transparency is evaluated and the photocatalytic activity is measured by means of stearic acid degradation under simulated solar light, with the eye on real outdoor applications. This is a widely recognised model reaction as stearic acid is a representative of the group of organic fouling compounds that typically contaminate glass surfaces [1,5,38]. This study thus goes beyond the state-of-the-art of (photocatalytic) self-cleaning surfaces in that sense that currently commercially available photocatalytic self-cleaning materials can only utilise the UV-component of solar light. The goal is to expand the activity window towards the entire solar spectrum. This would not only increase the efficiency for outdoor use, but could also enable the use of such surfaces under artificial (indoor) light.