The initial photocatalytic reaction rate follows a Langmuir-Hinshelwood type of relationship with respect to initial organic concentration [9,26-29,31,33,39,42,48,49,59,60,63-65]:


where CQ is the initial concentration of the target organic compound. The constants k and K represent collections of oxidation, recombination, and possibly adsorption terms. This expression describes a first-order reaction at low substrate concentrations, with a transition to a rate independent of organic concentration at high substrate loadings (Figure 2). The rate expression can be linearized by reciprocation, resulting in

thus describing a linear plot for l/rQ versus 1/C„ (Figure 2, curve B). Secondary organic substances, formed via the oxidation process, complete with the original substrate for the oxidant (hydroxy 1 radicals), so later reaction rates may be slower than the initial rate.

The effects of several operational and solution parameters on the PCO reaction rate have been evaluated. In numerous cases, effects of these parameters are compared using rate constants (kobs) derived from first-order plots that are linear over a short range. Others have made comparisons among various reaction conditions using initial rates (rj, where a linear decrease in concentration is measured over a brief time period.

A Langmuir-Hinshelwood dependence on oxygen concentration has been determined [33,63]:

Oxygen is necessary to complete the oxidation reaction by reacting with the photoproduced electrons [surficial Ti(III)] to maintain electroneutrality via reaction (6).

The incident light intensity controls electron/hole production and thus the hydroxyl radical formation rate. For example, a direct linear relationship between light intensity and PCO rate has been found for m-cresol [43]. However, a transition from first-order to 1/2-order is found at high light intensities [63,66]. This is apparently due to an increase in hydroxyl radical recombination,

Figure 2 (A) Initial 3-chlorophenol concentration versus initial PCO rate. (B) 1/CQ versus l/rD. (Reprinted with permission from D'Oliveira et al. [39]. Copyright 1990 American Chemical Society.)

Figure 2 (A) Initial 3-chlorophenol concentration versus initial PCO rate. (B) 1/CQ versus l/rD. (Reprinted with permission from D'Oliveira et al. [39]. Copyright 1990 American Chemical Society.)

yielding a recombination rate proportional to [OH-]"2. This consequently produces an overall organic oxidation rate proportional to the square root of light intensity [67].

Quantum yield is a measure of the efficiency of light utilization and is defined as the ratio of the number of photons entering into a photochemical reaction to the number of photons applied. The former parameter is usually evaluated by the conversion of the substrate, the latter by measurement of the incident light characteristics.

The degradation of 3,4-dichlorobiphenyl over illuminated TiOz was examined by Tunesi and Anderson [47], Two light intensities were evaluated; increases in light intensity result in a faster removal rate but a smaller quantum yield, due to inefficient light utilization and increased radical recombination, as discussed above. Intermediate products isolated include linear and branched hydrocarbons and some phenols. Quantum yields ranged from 7.4 X 10-4 to 1.6 X 1(T3. PCO quantum yields of 0.06 for acetic acid [51] and 0.022 for salicylic acid [48] have also been reported.

Since key steps to Ti02 photocatalytic oxidation occur at the Ti02 surface [i.e., reactions (2) and (6)], the reaction rate would be expected to increase linearly with available catalyst. At dilute TiOz concentrations, such a relationship is observed. However, above certain concentrations, the rate of oxidation does not increase [34,36,63], as shown in Figure 3, and may decrease with further increasing Ti02 concentration. For example, Matthews [20] found that the rate of salicylate formation in the oxidation of sodium benzoate increased with the quantity of titanium dioxide up to 2 g/L but decreased slightly at higher loading. Augugliaro et al. [33] observed the same phenomenon with a maximum phenol PCO rate at 1 g/L TiOz. Matthews


Figure 3 Dependence of kobs on photocatalyst concentration in the photocatalytic oxidation of 4-fluorophenol. The inset shows a linear reciprocation of a Langmuir-Hinshelwood expression. (Reprinted with permission from Minero et al. [36]. Copyright 1991 American Chemical Society.)

Figure 3 Dependence of kobs on photocatalyst concentration in the photocatalytic oxidation of 4-fluorophenol. The inset shows a linear reciprocation of a Langmuir-Hinshelwood expression. (Reprinted with permission from Minero et al. [36]. Copyright 1991 American Chemical Society.)

[61], analyzing 4-chlorophenol degradation using Equation (11a), found that k decreased while K increased as Ti02 loading varied from 0.2 to 2 g/L.

These results are explained by the fact that above certain Ti02 concentrations, there is a stoichiometric Ti02 loading that is sufficient to use all available photons emitted at a given intensity. Increasing the Ti02 concentration above this level becomes inconsequential because all available light is being utilized. In addition, other factors that contribute to the rate independence on Ti02 concentration include the reactor configuration, reflection, and solution opacity, which may prohibit some of the light from activating available photocatalyst.

The effect of pH on PCO reaction kinetics is still very much unresolved. Some type of pH dependence, although it is usually slight, is noted with almost every organic substrate. However, it has not been possible to draw any general conclusions with respect to pH about photocatalytic oxidation kinetics. In some cases, there are discrepancies about the pH of the maximum rate for the same organic compound. For example, the maximum rate for phenol PCO has been reported at pH 3 by Augugliaro et al. [33], at pH 4-5 by Okamoto et al. [24], and a pH 5-9 by Tseng and Huang [34].

The PCO rates of some nonionic compounds, such as solvents, are affected by solution pH. Matthews [65] found that rates of C02 production from benzene, nitrobenzene, and chloroform [as quantified by the product kK using Equation (11)] were faster at pH 4.5 than at pH 3.0. However, dioxin PCO was found to be more rapid under basic conditions [10].

The effect of pH on the PCO of dissociating organic acids is complicated by the speciation changes of these compounds. In most cases, anionic organics are more reactive than the pro-tonated molecular species, which is common for electrophilic oxidation. Oxidation of the pen-tachlorophenolate ion at high pH proceeds faster than that of molecular pentachlorophenol at pH 3 [35]. Similar results have been found for m-cresol [43] and 4-chlorophenol [37]. Palm-isano et al. [68] studied the effect of pH on the initial reaction rates of phenol and 2-, and

3-, and 4-nitrophenols. The highest photoactivity was found in the alkaline region. 3-Nitrophenol photocatalytic oxidation was less pH-dependent than that of the 2- and 4-isomers. The PCO rate of oxalic acid was maximum at pH 2.3 and decreased at higher pH, apparently due to the change in speciation from HC204~ to C2042- [9]; however, acetic acid photocatalytic oxidation is maximum at pH 3 [51]. The photocatalytic oxidation of three commercial azo dyes and a model compound, 4-hydroxyazobenzene, was examined by Hustert and Zepp [53], Orange G was oxidized much faster at pH 12 than at pH 7, apparently due to the deprotonation of this weak organic acid dye.

Mechanisms other than substrate speciation that may be responsible for pH dependencies include surface hydrolysis of the Ti02 (with subsequent changes in the electrical double layer), pH dependencies of any adsorption of products or reactants (02, H20, OH-, or organic), and pH dependencies of the reaction rates of specific organic substrates. For example, Kormann et al. [66] correlated photocatalytic oxidation rates as a function of pH with calculated surface speciation for both trichloroacetate and chloroethylammonium ion adsorption onto Ti02.

Titanium dioxide has two common crystal structures, anatase and rutile. Several studies have noted that rutile is not an active photocatalyst [24,33], However, two of the three rutile catalysts examined by Auguliaro et al. [33] have specific surface areas significantly lower than that of the anatase sample; thus a direct comparison may not be completely valid. Nevertheless, Sclafani et al. [69] noted an inactive rutile that has a specific surface area larger (20 m2/g) than that of an active anatase (14 m2/g). An active rutile photocatalyst was found by Davis et al. [70], Apparently the reactivity of rutile Ti02 depends on the method of Ti02 synthesis [69], The bandgap energy for rutile is 3.0 eV, as compared to 3.2 for anatase. Thus the oxidation-reduction potentials are slightly less for the rutile phase and, thermodynamically, some reactions may not be favored with rutile. Sclafani et al. [71] attributed PCO activity differences between anatase and rutile Ti02 to the thermodynamics and kinetics of the reduction reactions. Although crystal structure plays a role in the photocatalytic reactivity of Ti02, there are other controlling parameters that must be considered when evaluating Ti02 from different sources.

In theory, a higher Ti02 specific surface area will increase its photocatalytic activity due to increased area for adsorption of HzO and OH- and the corresponding subsequent generation of OH- radicals [reaction (2)]. Supporting this assumption, Matthews [59] observed that the degradation rate of 4-chlorophenol was much lower with an equal concentration of La Porte TiOz (specific surface area = 9 m2/g) than with Degussa TiOz (50 m2/g), presumably because of its much lower surface area.

On the other hand, this theory has not been found to be consistent in several other studies. Investigations by Cuendet and Grätzel [72] on pyruvate photocatalytic reactions found similar rates for two Ti02 samples with different specific surface areas (145 m2/g versus 50 m2/g). Tanaka et al. [73] studied the degradation of trichloroethylene (TCE), methylchloride acid, and phenol on 12 commercially available Ti02 samples. The degradation rates varied among Ti02 types, but there was no correlation with the specific surface area. The rates were found to be dependent on the crystallite size of the anatase form present in the Ti02 sample; the larger the crystallite size, the faster the reaction rate. The two Ti02 samples (with 100% anatase) in which the fastest rates were observed had specific surface areas of 17.3 m2/g (Fujititan TP-2) and 9.5 m2/g (Aldrich). Sclafani et al. [69] examined several commercial and synthetic types of Ti02 possessing a wide range of physical and chemical properties, such as differing crystal structures and specific surface areas; correspondingly, a wide range of PCO rates was found. There was no photocatalyst property that produced a correlation with respect to reaction rates.

Particle size may affect PCO rates through an effect on the degree of electron/hole trapping. As particle size increases, the distance that the electron-hole pair must diffuse through the solid before reacting at the Ti02 surface increases. This concurrently increases the proba-

Sampling Port


a '



TiO, Suspension

Magnetic Stirrer




TiO, Suspension

Magnetic Stirrer

Longwave UV Lamp

Figure 4 Typical recirculating batch reactor used in photocatalytic oxidation kinetic studies.

bility that recombination will occur. Consequently, any decreased rate of PCO for larger Ti02 particles may be due to the greater degree of electron-hole recombination. Davis et al. [70] found that the photocatalytic activity of a specific TiOz catalyst depends on the synthesis (manufacturer) of the TiOz, the crystal structure, and any pretreatment of the solid. No correlation was found between particle size of the Ti02 and initial PCO rate of toluene.

Several investigators have reported on the detrimental effect of chloride ions on PCO rates [26,31,33,38,66], The presence of chloride significantly decreases organic oxidation rates, possibly by scavenging an active radical species [74]:

Inhibition of the PCO process by chloride is a serious concern because chloride is formed during the mineralization of chlorocarbon compounds, many of which are found in contaminated waters. Abdullah et al. [74] and Tseng and Huang [38] examined the effect of several other anions on the PCO rate of organic substrates. Phosphates and sulfates decreased reaction rates, apparently by adsorbing onto the Ti02, resulting in the deactivation of some active sites; N03~ and C104~ had no effect on PCO reaction rates.

Photocatalytic oxidation rates are decreased in the presence of bicarbonate ions. The presence of 500 mg/L bicarbonate decreased TCE rate constants by a factor of 2-5 in bench-scale studies [75]. Bicarbonate ions are well known as radical scavengers through work investigating ozonation kinetics. Apparently the HC03~ acts in the same manner during PCO, scavenging hydroxyl radicals, thus preventing reaction with the target substrate.

In many instances PCO kinetics have been investigated using recirculating systems in which a mixed reactor feeds an isolated plug-flow photoreactor, similar to that in Figure 4. In this manner, sampling and suspension chemistry monitoring and adjustments are performed in the mixed reactor, away from the photoreactor. Studies have shown that PCO first-order rate constants (kobs) are a function of flow rate (FR) through these reactors and are described by the equation [56]

At small flow rates, an increase in flow results in a higher reaction rate constant; the rate becomes independent of flow at higher flow rates. Similar results were found by Al-Ekabi and Serpone [21] in an examination of flow through a coiled glass tube with an internal coating of Ti02.

Minero et al. [36] noted that the PCO rates of five different fluorophenols and difluorophe-nols were very similar. Analogously, the rates of phenol, the three chlorophenol isomers, five different dichlorophenols, and 2,4,6-trichlorophenol were all within about 30% [38]. Both of these investigations illustrate the nonselectivity of PCO systems. However, in contrast, Terzian et al. [43] found that the PCO rate of m-cresol was about 2 1/2 times slower than that of either o- or p-cresol under identical conditions.

Competitive interactions among 4-chlorophenol, 2,4-dichlorophenol, and 2,4,5-trichloro-phenol were examined by Al-Ekabi et al. [42]. The PCO rates of all three compounds were slowed in a ternary mixture. However, the sum of the three individual rates was equal to that for PCO of a single compound, indicating that a fixed amount of oxidant was being produced by the illuminated Ti02 system.

Absolute comparison of organic PCO rates among various investigators is not possible due to variations in substrate and catalyst loading, light intensity, and reactor configurations. Some have based rates on target organic disappearance, while others have monitored the production of C02 or CI-. Batch and flow reactors using suspended as well as immobilized photocatalyst have been examined. Discussions on the effect of reactor dynamics on PCO rates are presented by Davis and Hao [76] and Turchi and Wolfrum [77],

0 0

Post a comment