The Effect of Interfacial Chemical Bonding in TiO2-SiO2 Composites on Their Photocatalytic NOx Abatement Performance

The chemical bonding of particulate photocatalysts to supporting material surfaces is of great importance in engineering more efficient and practical photocatalytic structures. However, the influence of such chemical bonding on the optical and surface properties of the photocatalyst and thus its photocatalytic activity/reaction selectivity behavior has not been systematically studied. In this investigation, TiO2 has been supported on the surface of SiO2 by means of two different methods: (i) by the in situ formation of TiO2 in the presence of sand quartz via a sol-gel method employing tetrabutyl orthotitanium (TBOT); and (ii) by binding the commercial TiO2 powder to quartz on a surface silica gel layer formed from the reaction of quartz with tetraethylorthosilicate (TEOS). For comparison, TiO2 nanoparticles were also deposited on the surfaces of a more reactive SiO2 prepared by a hydrolysis-controlled sol-gel technique as well as through a sol-gel route from TiO2 and SiO2 precursors. The combination of TiO2 and SiO2, through interfacial Ti-O-Si bonds, was confirmed by FTIR spectroscopy and the photocatalytic activities of the obtained composites were tested for photocatalytic degradation of NO according to the ISO standard method (ISO 22197−1). The electron microscope images of the obtained materials showed that variable photocatalyst coverage of the support surface can successfully be achieved but the photocatalytic activity towards NO removal was found to be affected by the preparation method and the nitrate selectivity is adversely affected by Ti-O-Si bonding.


Introduction
Concrete structures are ubiquitous in our society. Typically associated with our urban centers, their significant surface area represents an important interface with the urban atmosphere 1,2 . With increasing concerns over the economic and health impacts of deteriorating urban air quality this interface presents an important opportunity for atmospheric remediation. TiO 2 -based photocatalysts have been utilized for some time in the remediation of NOx-contaminated air, and their support on these high surface area concrete structures offers concrete the additional functionality previously associated with photocatalytic materials: (i) easy-cleaning, whereby materials which bind dirt to surfaces are photocatalytically degraded enabling dirt to readily wash off with rain water 3 ; (ii) photo-induced hydrophilicity, which also enhances the selfcleaning effect 3 ; and (iii) purification of the urban atmosphere which today, is typically polluted by vehicle emissions at levels that significantly exceed maximum permissible levels, particularly with respect to NOx 4 . TiO 2 is the most commonly employed photocatalyst in environmental applications due to its chemical stability, relatively low price, high photocatalytic activity, and more importantly its eco-safety as indicated by currently available TiO 2 toxicology data 5 .
Photocatalytic concretes have already demonstrated their potential for atmospheric remediation on trial sites throughout Europe and elsewhere. Numerous studies on photocatalytic cementitious materials over the last two decades have predominantly dealt with catalyst activity, often expressed in terms of NOx concentration reduction 1,6,7,8,9 . However, activity alone is an insufficient indicator of photocatalytic effectiveness. A reduction in NOx concentration, defined as the sum of the concentrations of the atmospheric nitrogen oxides, does not by itself represent a useful impact on air quality because the relative toxicities of the constituent gases are not equivalent 10 .
Photocatalytic oxidation of NOx gases follow the sequence NO → HONO → NO 2 → HONO 2 (NO 3 -) The higher toxicity of NO 2 relative to NO (by, conservatively a factor of 3 .
Significantly improved performance can therefore be expected when accessible catalyst surface area is better preserved in more efficient photocatalytic structures. These have included catalysts supported on concrete surface exposed aggregates and in zeolite structures 2,12 . The durability of these structures depends very much on how well bound the catalyst is to the various supports. The benefits of chemically bonding TiO 2 to substrates have often been referred to in the literature 8,13 but the means of characterizing the degree of binding has been ambiguous.
Nevertheless, the integrity of a chemical bond relative to a physical attraction presents an opportunity to develop robust structures on the surface of the concrete. However, the influence of a chemical bond between TiO 2 and a substrate, e.g. quartz, to provide a Ti-O-Si linkage, on the optical and photocatalytic properties of the supported TiO 2 has not previously been studied. Therefore, the focus of the present work has been in establishing means to generate and quantify levels of Ti-O-Si linkages and to correlate these with the photocatalytic properties of the supported TiO 2 . For this purpose, commercial as well as synthesized TiO 2 have been bonded, by different methods, onto quartz SiO 2 sand (Q; as a simple example of an aggregate). NOTE: T2: Pure TiO 2 was also prepared by the same sol-gel method but in the absence of TEOS.

Characterization
1. Record IR spectra using a spectrophotometer equipped with UATR (Single Reflection Diamond) 15 . 2. Obtain X-ray diffraction (XRD) patterns using a PAN analytical diffractometer equipped with a CuKa1 1.54 Å X-ray source 16 .
3. Analyze the morphology of the samples via scanning electron microscopy (SEM), equipped with ED X-ray analyzer and BSE detector with operating voltage between 10 -20 kV. Use energy dispersive X-ray analysis and capture images with a Digital Image Acquisition System. 4. Perform transmission electron microscopy (TEM) on a microscope operated with an accelerating voltage of 200 kV. Capture images with a camera. 5. Record UV-Vis diffuse reflectance spectra of the samples using a UV-Vis spectrophotometer equipped with fiber optic coupler.
Use barium sulphate as reference in the range of 250 to 600 nm. Transform the resulting reflectance spectra into apparent absorption spectra by using the Kubelka−Munk function

Photocatalytic Performance Test
1. Test the photocatalytic activities of the prepared materials using the removal of NOx from polluted air test 18 . 1. For this purpose, establish an air-purification test set-up (see Figure 1) consisting of gas supplies, humidifier (2), gas flow controllers (1), photocatalytic reactor (3), UV(A) light source (4) and NOx analyzer (5). The gas supplies were NO (100 ppm) in N 2 , and synthetic air( BOC). 2. Use mass flow controllers (1) to provide NO at 1 ppmv (0.5 ppmv, for ST1 and T1 samples) and the relative humidity to ca. 40%, confirmed by Rotronic hygropalm, to the laminar flow reactor (3)    Diffuse reflectance spectroscopy Figure 4 shows  Fourier Transform Infrared Spectroscopy (FTIR) Figure 5 shows the FTIR spectra of the SiO 2 /TiO 2 mixed oxides samples and of the TiO 2 -Q composites. Evidence for the chemical binding of TiO 2 to SiO 2 may be observed in the range between 900 -960 cm -1 assignable to the Si-O-Ti stretching vibrational mode 15 ; as expected, no absorption peak due to this mode was observed for SiO 2 or TiO 2 .  Samples (QT1, QT2). For clarity, the spectrum for T2 is not shown but it is identical to T1. Please click here to view a larger version of this figure.

Scanning electron microsocopy (SEM)
The effectiveness of a silicate-based film on quartz (QT2) for the efficient support of TiO 2 has been examined by SEM. Much depends on how well the film itself coats the quartz substrate. Figure 6 compares the SEM-EDS of commercial TiO 2 (PC105) dispersed within this film derived from TEOS with TiO 2 in a 1:1 molar ratio (QT2). The silicate film was found to have been immobilized inhomogenously on the grains as some areas remain clear of the silicate coating. Consequently, in this case, TiO 2 , associated with the silicate-based gel phase, is also inhomogeneously distributed and is not bonded directly to the quartz surface. This is consistent with the TEM image in Figure 3 b(2). The silicate coating (top right of image) gives an EDS analyses comparable with that reported in Figure 6(d) indicating the association of TiO 2 with the silicate layer. It can also be noticed from Figure 7 that the concentration of NO increased slightly and continuously during the entire irradiation time. This illustrates an approach to the steady state condition and can be attributed to the accumulation on the available active sites of photocatalytically generated NO oxidation products, i.e., HNO 2 /NO 2 -; NO 2 ; and HNO 3 /NO 3 -, which may influence NO adsorption rates. Bloh et al. reported that achieving a steady-state in this system requires several hours of illumination. To determine and compare the activities of the obtained TiO 2 -SiO 2 composites for NOx abatement, the photonic efficiencies (ξ) for the removal of NO, NOx and the formation of NO 2 was calculated and illustrated in Figure 8.  Figure 8 shows quite significant differences between the NO photonic efficiencies for each of the photocatalytic materials. The advantages of supporting the photocatalyst to increase accessibility to reactive surface is now well established and it is worth noting the difference between the photonic efficiencies for NO oxidation measured for PC105 and for PC105 supported on treated quartz (QT2). ξ NO (QT2) was measured at 73% of that for PC105 but QT2 had only 6.5% of the TiO 2 loading. Clearly, activity improvements are significant on supported systems but care should be applied when interpreting measurements with significant morphological differences.

Discussion
A key characteristic of the photocatalytic test system which can be expected to influence the measurement is the surface texture of the sample supported in the photocatalyst reactor. This influences the effective surface area. The calculation of ξ includes an area term but this is a twodimensional area of illumination defined by the reactor sample holder. The particle size distribution of TiO 2 powders, i.e. PC105, T1 and T2, are quite different from the composites, where TiO 2 'powder' is supported on SiO 2 of diameter in the range 0.4-50 μm. This means that the photocatalyst surface textures are quite variable and are expected to influence the reported photonic efficiencies. It also influences reactor flow characteristics. The rougher the texture, due to packing characteristics, the more likely that the laminar flow regime required is disrupted. This is expected to influence rates of gas molecule diffusion to surface and consequently the photonic efficiency measurement.
As a consequence of these effects, the most useful comparison of photocatalyst types must be based on properties derived from measurements on individual catalysts. In this study, nitrate selectivity, which is based on ξ NO and ξ NO 2 (Equation 10), both measured on the same sample are used in subsequent discussion.
(10) The factors which control nitrate selectivity appear to be complex and relevant variables include TiO 2 polymorphism, defect state, availability of water, etc. 7 , but the role of substrate binding, often considered to be advantageous to photocatalytic performance, can now also be considered. It is beneficial therefore to discuss the nitrate selectivity differences between non-bonded and bonded systems, i.e. stand-alone photocatalyst vs photocatalyst-support composites, e.g. PC105 vs QT2; where QT2 represents PC105 supported in a silicate coating on quartz. These nitrate selectivity differences are summarized in Table 1 ) were obtained from Figure 5 using Origin Peak Analyses software. The dimensionless area ratio indicated in Table 1 is taken as a measure of the degree of Ti-O-Si bonding in composite systems. Also shown are the peak center positions associated with the Ti-O-Si linkage. These data are summarized in Figure 10. The greatest selectivity reduction on composite formation, i.e. that which would show the largest negative impact on ambient air quality, is indicated for photocatalyst T2 when it is combined with a silicate precursor. A highly dispersed gel is produced in which Ti-O-Si linkages are maximized. The peak area analyses indicates that around 65 mole % of the TiO 2 is associated with SiO 2 through Ti-O-Si connections, which is approaching the stoichiometric TiO 2 :SiO 2 ratio of the preparation (80%) and providing confidence in the peak area ratio analysis. It is also noteworthy that the Ti-O-Si peak center is located at the lowest wavenumber observed for the composites and suggests that compositional information may be embedded in the Ti-O-Si peak characteristics. All other composites display considerably lower (Ti-O-Si)/SiO 2 peak area ratios, indicating lower levels of Ti-O-Si bonding. Figure 10 shows that this level of bonding is correlated with Selectivity, expressed as a percentage reduction from the free standing catalyst selectivity, indicating that Ti-O-Si binding has a negative impact on photocatalytic NOx abatement.
The consequences of these findings are that a compromise must be met to ensure the physical durability of a bonded system without a significant loss of photocatalytic performance. Possible approaches could include: (i) increasing the supported TiO 2 particle size such that the beneficial Ti-O-Ti bonding, which defines the intrinsic photocatalytic properties of 'stand-alone' photocatalysts, are not diluted by the Ti-O-Si linkages, and/or (ii) engineering a thin, porous and durable surface coatings for the substrate such that the photocatalyst is trapped in pores accessible to reactant gas molecules and illumination.
Silica in the form of quartz sand or reactive silica spheres has been successfully modified with TiO 2 either via binding commercial TiO 2 photocatalyst (PC105), utilizing a silicate-based binder or via the hydrolysis-condensation reactions of different Ti precursors. The photocatalytic performance of the resulting composites has been compared with that of a sol-gel derived mixed oxide system promoting high levels of Ti-O-Si binding linkages. The key findings show that: (i) the degree of TiO 2 -SiO 2 binding in the mixed oxide preparation is high (65%) as expected and approaches the stoichiometric TiO 2 :SiO 2 ratio in the preparation. This composite gel system displayed no nitrate selectivity compared with the comparable sol-gel derived TiO 2 (T2) which showed a selectivity of 33%, (ii) as the reactivity of the silicate surface reduces, the degree of Ti-O-Si