The herbicide MCPA sodium salt monohydrate (sodium (4-chloro-2 methylphenoxy) acetate pure standard (99% purity) was purchased from Sigma-Aldrich Corporation (USA) and used as received. Formic acid and HPLC grade Acetonitrile were purchased from Aldrich and used as received. Water was Milli-Q quality. TiO₂ P-25 from Degussa (anatase/rutile = 3.6/1, surface area 50 m²/g, non-porous) was used for photo-catalytic experiments. Grafted TiO₂ thin blue glasses were obtained from Pilkington (UK) (Figure 2). All the solutions were daily prepared in ultra-pure water from a Millipore purification system. The pH of the solution was monitored using a Basic pH Meter from Denver Instrument Company. In the experiments with heterogeneous was separated by filtration (through 0.2 μm membrane filters, Schleicher and Schuell, Germany. Cat. No. 10462200) before analysis of MCPA sodium salt monohydrate.

Photolysis, photocatalysis with TiO₂ powder and photocatalysis with immobilized TiO₂ on thin blue glass experiments were accomplished using batch reactor system, this system consists of:

1) Suntest CPS + Solar Simulator (Heraeus Instruments, Germany) equipped with a xenon lamp, temperature sensor and water-cooling circuit. The xenon lamp was filtered by an optically stable borosilicate UV filter (Atlas Material Testing, France) delivering a light emission spectrum similar to that of the sun with a UV cut-off at 290 nm. The pyrex reactor that contains the solution placed inside the chamber of the suntest device.

2) Pyrex batch reactor. The capacity of the reactor is 300 ml. Its outer perimeter reactor is covered by aluminum foils and only the upper surface is exposed to radiation.

3) Magnetic stirrer device (Falc F20 Mini Magnetic stirrer. Progen Scientific, Merton. London) is employed to maintain a continuous stirring for the solution inside the pyrex reactor during the experiments.

A schematic drawing of the batch reactor system is shown in (Figure 3).

1) High Performance Liquid Chromatography system (HPLC) (Agilent technologies 1200 series) (Avondale, PA, USA) is equipped with a C₁₈ (YMC), 4.6 × 150 mm diameter with 3 μm column (Phenomenex-USA) using a diode array detector.

2) Scanning electron microscope (SEM): scanning electron microscope (SEM) (LEO model EVO50XVP, Carl Zeiss AG-EVO® 50 Series, Germany).

In batch reactor system, samples of 2 mL were taken at determined time intervals. Samples of photocatalysis/TiO₂ powder experiments were filtered through a 0.2 µm filters to remove TiO₂ particles. Changes in the concentration of each drug were observed from its characteristic absorption at selected nm, using HPLC system through the following method:

- Column: C₁₈ (YMC), 4.6 × 150 mm diameter with 3 μm.

- Mobile phase: 1% CH₂O₂ (V/V): ACN (60:40).

- Wavelength: 350 nm.

- Flow rate: 1.0 mL/min.

- Injected volum: 20 µL.

The elemental analysis for the grafted TiO₂ thin Pilkington blue glasses was performed by using scanning electron microscope. Some cross sections obtained from the Pilkington Active^TM Blue glass were analyzed. The thin sections were

coated with a 30 nm-thick carbon films. Semi quantitative analyses of the elemental composition of the different layers were obtained using a Ge ED Oxford-Link detector equipped with a Super Atmosphere Thin Window. Operating conditions of the SEM were: 15 kV accelerating potential, 500 pA probe current and about 10 mm of working distance (WD).

Thin sections of glass were prepared by the Department of Health and Environmental Science, Bari University. Samples were embedded in resin epoxy plugs and then polished.

The knowledge of reaction order is essential for finding the accurate integrated rate equation. By trying to fit data of various integrated rate equations, it is possible to verify the reaction order. Kinetic parameters were calculated using integrated equations describing zero-, first- and second- (Langmuir-Hinshelwood) order equations. The determination coefficient (R²) was used to check the best fit.

Kinetic parameters (reaction order (n), determination coefficient (R²), half-life (t_1/2), kinetic constant (k) were obtained by linear regression of logarithmic concentration values determined as a function of time according to the following equations [15] [16] .

Zero-order: C t = C 0 – k t (1)

First-order: L n C t = L n C 0 – k t (2)

Second-order: C 0 / C t = 1 + ( 1 / t 1 / 2 ) t (3)

(Figure 4) shows the SEM image of the cross section sample of the Blue glass, where the thickness of TiO₂ film was found to be 397.2 nm. (Figure 5) displays the fine-tooth comb geometry of TiO₂ coating on the Blues glass surface.

As shown in (Table 1), TiO₂ is a component present only on the glass surface along with other metal oxides such as iron oxide, while in the core glass it is absent and other metal oxides are countable. As reported in the Pilkington patent [17] , cobalt oxide is present in low amounts (less than 75 µg/g) but it was not detectable by surface analysis used. Cobalt oxide may confer the blue color to the glass.

Preliminary experiments in the dark showed that the drug dissolved in aqueous solution with and without catalyst after 48 hours was stable and no reactivity was observed at all.

Figure 6 represents the behavior of photolysis and photocatalysis reactions. Removal of MCPA sodium monohydrate during photolysis experiment was very poor, only 12 % from its initial concentration has been removed after 27 hours with half life equal 115.5 hours. Using immobilized TiO₂ glass slaps in the photocatalytic degradation experiments accelerate to some extent the degradation rate, about 40 % of initial concentration was removed after 27 hours with half life equal to 37.3 hours as shown in Table 2. But in general, the degradation rate is still substandard.

While, using TiO₂ dispersed powder to the photocatalytic process accelerated the degradation process dramatically (half-life = 36.5 min and complete mineralization in about 4 hours).

The knowledge of reaction order is essential for finding the accurate integrated rate equation. By trying to fit data of various integrated rate equations, it is possible to verify the reaction order. Kinetic parameters were calculated using integrated equations describing zero-, first- and second- (Langmuir-Hinshelwood) order equations. The determination coefficient (R²) was used to check the best fit.

The photo-degradation for MCPA sodium monohydrate plotted as the natural logarithm of the ratios of the observed over the initial concentrations, exhibited approximately liner behavior. From the overall results, a comparative kinetic study was carried out for the chemical process with photolysis and with photocatalysis. Initial mineralization rate constants (k) were determined, for all processes, from the slope of (ln CA_(t)) vs t (time) plots where CA_(t) are MCPA sodium monohydrate concentration at time t. It is also presented the half-life times (t), i.e. the time necessary to reduce 50% of initial MCPA sodium monohydrate concentration for each process. The half-life times are determined by interpolation from experimental data (Table 2).

Determining the transient photoproducts that resulted from degradation processes is very important, since some of these photoproducts could be more toxic than the parent compound accordingly. Previous studies and literatures related to possible pathways of MCPA photo degradation have focused on the expected photoproducts.

According to Topalov et al. (2001 and 2003), the photo assisted oxidative degradation of MCPA in the presence of TiO₂ involves formation of a variety of intermediate species. In these studies the NMR analysis of suspensions of TiO₂ containing MCPA was performed and the proton NMR spectral profiles were monitored on a Bruker AC-250 instrument, and ¹³C NMR spectra were recorded by the same instrument. The results suppose that degradation pathway followed the hydroxylation of the aromatic ring and chloride. Accordingly, the degradation of the parent MCPA molecule yields the formation of 4-chloro-2-methylphenol, by elimination of CO₂ and H₂O as shown in (Table 3). Further transformations of the aromatic ring are possible by the chlorine substitution with hydroxyl group, or by electrophilic addition of the OH radical [14] [18] .

As a result of degradation of the aromatic part of the MCPA molecule, acetic acid is formed. By further action of OH radicals, acetic acid is decomposed into CO₂ and H₂O [17] [19] [20] . Hydroxyl radicals continued the attacking of the substrate caused hydroxylation products, and finally the opening of the ring yielding CO₂ and H₂O as presented in Figure 7 [14] [18] .