TiO2 Reaction with

Gary O. Rankin , Adam Sweeney , Christopher Racine , Travis Ferguson , Deborah Preston , Dianne K. Anestis. Chemico-Biological Interactions 2014 222, ...
1 downloads 0 Views 124KB Size
8184

Ind. Eng. Chem. Res. 2005, 44, 8184-8189

KINETICS, CATALYSIS, AND REACTION ENGINEERING Destruction of o-Chloroaniline in UV/TiO2 Reaction with Photosensitizing Additives W. K. Choy and W. Chu* Department of Civil and Structural Engineering, Research Centre for Environmental Technology and Management, The Hong Kong Polytechnic University, Hung Hom, Hong Kong

The photodecay of o-chloroaniline (o-ClA) in TiO2 suspension with and without the use of additives was examined under an ultraviolet light at 300 nm. An analysis showed that the disappearance of o-ClA followed a pseudo first-order reaction and can be described by the LangmuirHinshewood (LH) model. The application of various dosages of TiO2 indicated a discontinuous linear increment at a break point of 0.1 g L-1 TiO2. The change in the system pH has a significant effect on the performance of the process. This is a mixed result of the surface charge of TiO2 and the chemical characteristic of o-ClA, which is discussed in detail in this paper. The further enhancement of the UV/TiO2 process by adding H2O2 is feasible if the dosage is carefully controlled. The addition of excess amounts of H2O2 should be avoided to minimize the scavenging of hydroxyl radicals in the solution and valence-band holes at the TiO2 surface. The use of the sensitizer Rose Bengal (RB) in the UV/TiO2 process, however, is not suggested. This is because RB is not sensitive at the optimal wavelength for the photocatalytic process involving TiO2; in addition, the RB may become an additional competitor (with the probe) for the free radicals and active sites on the TiO2 surface. Introduction The removal of organic matter through a photocatalytic process has proven to be effective in the fields of air and wastewater treatment. A wide variety of organic pollutants including alkanes, alcohols, carboxylic acids, alkenes, phenols, dyes, PCBs, and pesticides were treated by UV illumination,1,2 and mineralization was reported in many cases.3,4 The generation of powerful oxidants, hydroxyl (OH•), or other radicals (O2•-) by UV illumination is the major mechanism of photodegradation. However, direct photolysis only causes little or slow decay with insignificant number of radical generation, while a higher reaction rate is always observed with photon-assisted catalysts.5 Therefore, scientists have tried to screen or generate different catalysts to improve the process. The application of heterogeneous photocatalysts has received great consent because of its effectiveness and because of its wide adaptability in organic destruction.6,7 Potential metal oxide semiconductors have been investigated by researchers over the years. The decolorization of azo dye with TiO2, ZnO, CdS, Fe2O3, and SnO2 has been carried out under solar light.8 Among them, titanium dioxide (TiO2) is considered one of the best catalysts for the photooxidation process.9-10 It is chemically and biological inert, photocatalytically stable, easy to produce and use, and cheap and not harmful to the environment and to humans. When TiO2 is illuminated with the light of energy greater than its band-gap energy, an electron is trans* To whom correspondence should be addressed. Fax: (852) 2334-6389. E-mail: [email protected].

ferred to the conduction band leaving a vacancy in the valence band. The electron and hole can migrate to the surface of the TiO2. At the interface, electron transfer can occur either from the conduction band (CB) to an acceptor or a donor to the valence band (VB). The generation of •OH is generally believed to be the primary oxidant in photocatalytic reactions (eqs 1 and 2).11

TiO2 + hυ f e-CB + h+VB

(1)

h+VB + OH- f •OH

(2)

Excited-state conduction-band electrons (e-CB) and valence-band holes (h+VB) can recombine and dissipate the input energy as heat, get trapped in stable surface states, or react with electron donors or electron acceptors on the surface of the semiconductor. In the absence of suitable electrons and hole scavengers, the stored energy is dissipated within a few nanoseconds by recombination. To promote an efficient photocatalysis, it is necessary to prevent the accumulation of e-CB and h+VB on the surface as such an accumulation could increase the recombination rate and lower the quantum yield. It has been pointed out that an increase in the number of trapped electrons could prevent the recombination from occurring.12 The chemical o-chloroaniline was chosen as the probe compound in this study. It is used in the manufacture of dye, petroleum solvents, rubber, and fungicides.13 Such a chemical may be released into the environment from process/waste emissions involved in the production or use and may also enter the environment as a product

10.1021/ie0506419 CCC: $30.25 © 2005 American Chemical Society Published on Web 09/29/2005

Ind. Eng. Chem. Res., Vol. 44, No. 22, 2005 8185

in the degradation of various pesticides14 or from food residues.15 Chloroanilines bind with humic materials in water and soil. The Environmental Agency of Japan reported that there was a maximum of 0.56 mg L-1 o-chloroaniline in the country’s coastal waters in 1990.16 However, the methods for treating or removing this compound are still very limited. Also, it has been proven that dechlorination cannot be observed for the photosensitization of o-chloroaniline under sun-simulated light while m- and p-chloroanilines gave degradation rates of 3.4 × 103 and 6.5 × 103 min-1, respectively, in a 250-min treatment.17 Therefore, an examination of the photocatalytic degradation of o-chloroaniline is valuable and of great interest. Experimental Section Reagents Used. The TiO2 catalyst used was Degussa P25 (80% anatase and 20% rutile) with a BET surface area of 50 m2/g and a density of 3.85 g/cm3. It has an average aggregate size of 200 nm and is made up of 30 nm primary particles.18 A high percentage of the anatase phase in P25 TiO2 is more photoactive than in the brocktile or rutile phases. The probe o-chloroaniline (99.8%; C6H6ClN) with a molecular weight of 127.57 M and water solubility of 0.5 g/100 at 20 °C was purchased from Fluka. Acetonitrile (HPLC grade), acetone (reagent grade), and H2O2 (30%) were purchased from Tedia, while Rose Bengal was purchased from Aldrich. Doubly ionized water was used throughout the experiment. The water used in the preparation of all of the solutions was obtained from a Millipore Waters Milli-Q water purification system. All of the chemicals were used without further purification. Photoreaction Experiment. The photodegradation experiments were conducted in a Rayonet RPR-200 photochemical reactor manufactured by the Southern New England Ultraviolet Co. To ensure a thorough mixing, 150 mL of solution was dispensed into a 300 mL quartz cylinder with mechanical stirring before and during the illumination. Eight phosphor-coated lowpressure mercury lamps at 300 nm having a total photon intensity of 5.6 × 10-6 einstein L-1 s-1 were installed on the photoreactor. The pH level of the solution was adjusted via H2SO4 and NaOH, respectively. Samples were withdrawn intermittently at a predetermined schedule and were filtered through a 0.45 µm membrane to keep the TiO2 free from the solution before quantification. Methanol and sodium azide (NaN3) were used as the quencher for the oxidants H2O2 and Rose Bengal1 to prevent subsequent oxidation. Control experiments without a photocatalyst and UV illumination were also performed for comparison. Analytical Technique. The amount of probe that remained in the treated samples was analyzed by HPLC and compared to the original (unreacted) sample. The system was comprised of a high-pressure pump, a 100 µL-loop injector port, a Restek pinnacle octylamine (5 µm, 0.46 × 25 cm) column, and a UV detector. A mixture of 60% acetonitrile and 40% water was used as the mobile phase running at a flow rate of 1 mL/min. Adequate degassing of the mobile phase prior to injection was performed to inhibit the generation of gas bubbles during the analysis. Results and Discussion Effect of Different Initial Concentrations of o-ClA. The photodecay of o-ClA mainly follows a pseudo

Figure 1. (a) Photodegradation of o-ClA at different initial concentrations (eight lamps at 300 nm were used, and [TiO2] was fixed at 0.1 g L-1). (b) L-H plot of o-ClA photodecay.

first-order decay,19 where the resulting decay rate constants are strongly dependent on the initial concentration of o-ClA:

[ClA]t ) [ClA]0e-k1t

(3)

where [ClA]t (mM-1) is the concentration of o-ClA at time t, [ClA]0 is the initial concentration of o-ClA (mM-1), and k1 is the pseudo first-order decay rate constant (min-1), which is the slope calculated from the plot of ln([ClA]t/[ClA]0) versus reaction time, t. The photodecay of o-ClA ranging from 0.095 to 0.570 mM in the presence of 0.1 g L-1 TiO2 under a 300 nm UV irradiation is shown in Figure 1a. For a 40-min illumination, it was observed that 96% of the o-ClA was removed as the initial [o-ClA] was low (0.095 mM), while the performance was significantly reduced to about 50% as the initial [o-ClA] was increased to 0.570 mM. Generally speaking, a higher initial o-ClA yielded a lower removal rate under the same reaction conditions. Because the conventional pseudo first-order method (eq 3) is not capable of incorporating the variation of initial [o-ClA] concentrations to the corresponding rate constants, an alternative approach was used in this study for further analysis. The Langmuir-Hinshelwood (LH) kinetic was used to quantitatively delineate substrate preadsorption in both solid-gas and solid-liquid

8186

Ind. Eng. Chem. Res., Vol. 44, No. 22, 2005

Figure 2. Comparison of photodecay rate of o-ClA at various TiO2 dosages ([o-ClA]0 ) 9.5 × 10-5 M; eight lamps were used at 300 nm).

reactions.20 It is also possible to use the LH kinetic in heterogeneous photocatalysis systems only when homogeneous reaction is assumed to be insignificant.21,22 The initial rate for the first 30-min, R (mM min-1), can be expressed as:

R)-

kK[ClA]0 d[ClA]0 ) dt 1 + K[ClA]0

(4)

where K is the equilibrium absorption constant of o-ClA on the surface (mM-1) of the catalyst (i.e., TiO2) and k represents the limiting reaction rate at maximum coverage (mM min-1). In a pseudo first-order kinetic, the incorporation of -d[ClA]0/dt ) k1[ClA]0 (derived from eq 3) into eq 4 gives

k1[ClA]0 )

kK[ClA]0 1 + K[ClA]0

(5)

The above equation can be linearized by taking the reciprocal on both sides. After rearranging, the equation becomes

1 1 1 ) [ClA]0 + k1 k kK

(6)

The plot of 1/k1 versus [ClA]0 resolves 1/k and 1/kK from the slope and intercept, respectively, and the results were shown in Figure 1b. The degradation of o-ClA at different initial concentrations fits well to the LH model with a linear regression (r2) of 0.993. The two constants, k and K, were calculated to be 0.0112 mM min-1 and 102.6 mM-1 from the slope and intercept, respectively. The good correlation reveals that the adsorption of o-ClA to the surface of TiO2 is likely one of the major mechanisms governing the decay of o-ClA in such a photocatalytic process. Effect on TiO2 Dosage. A constant [o-ClA] concentration of 0.095 mM was used exclusively in the next tests. Different [TiO2] doses were added to the sample with continuous stirring under UV illumination, and the remaining o-ClA is shown in Figure 2. A dark reaction with sole TiO2 was also included for comparison. No obvious degradation in a dark TiO2 system was observed, implying that an external light source is necessary to initiate the process. There was a loss of o-ClA of less than 10% due solely to adsorption on the TiO2

Figure 3. First-order decay of o-ClA with several TiO2 dosages at 30-min illumination ([o-ClA]0 ) 9.5 × 10-5 M; eight lamps were used at 300 nm).

surface, and no further adsorption was observed after 24 h. The adsorption effect is regarded as minimal when compared to the TiO2/UV system, which reached a rate of removal of above 90% within 30 min in most of our tests. It is therefore reasonable to believe that the photocatalytic process with TiO2 is the dominant reaction pathway contributing to the removal of o-ClA. In addition, with the increment of [TiO2] doses from 0.017 to 1.6 g L-1, a faster and greater degradation of o-ClA was observed as shown in Figure 2. All of the degradations follow the pseudo first-order decay with an r2 of above 0.98 (see Figure 3). As compared to the direct photolysis (i.e., in the absence of TiO2), an improvement in the reaction rate of about 3 times was observed with the minimum addition of 0.017 g L-1 TiO2. This increase in the reaction rate should be the result of the increase in activated oxidants in the system by the introduction of TiO2 particles. More surface active sites are available for photon absorption, as described in eqs 1 and 2, and therefore more •OH and/or h+VB was generated in the system. It was interesting to note that the performance of the photocatalytic degradation of o-ClA was not a simple linear correlation with the [TiO2] doses. At lower [TiO2] doses, the pseudo first-order rate constants increased significantly (and linearly) with the increment of [TiO2] doses. As the [TiO2] doses increased above 0.1 g L-1, however, another linear correlation was observed at a much lower rate of improvement, indicating that the process was gradually retarded. The initial reaction rates were plotted against the [TiO2] doses, as shown in Figure 4. This is possibly due to the increase in the opacity of the solution with an excessive amount of TiO2 in the reaction, causing a reduction in the penetration of light and a slow in further rate increments.23 Thus, an unlimited increase in photocatalyst does not always guarantee a beneficial effect to the photoreaction. An optimum dose of photocatalyst is surely one of the critical design parameters for such a slurry system in any practical application. The pH Effect on the UV/TiO2 System. The pH levels of the solution could be another critical factor affecting the UV-catalytic process. A study of initial pH levels was therefore conducted at a range of 2-11, where the extension of the study to very low pH levels

Ind. Eng. Chem. Res., Vol. 44, No. 22, 2005 8187

Figure 4. Change of photoreaction rate with TiO2 dosage.

Figure 6. (a) Change of photodecay rate at different initial pH levels. (b) Proportion of o-ClA at different pH levels.

Figure 5. Effect of initial pH on o-ClA photodecay ([o-ClA]0 ) 9.5 × 10-5 M).

was due to the consideration of the possible effect of the pKa of o-ClA. The pKa of o-ClA was reported to be 2.6 (eq 7).24

The 30-min photodecay at various initial pH levels was plotted in Figure 5. All of the data followed the pseudo first-order reaction except the one at a very low pH level. At a pH level of 2, the slowest initial decay was observed and only 48% of the o-ClA was removed at the end of the reaction. This is compared to the situation at other pH levels, where a removal rate of more than 90% can be obtained. In fact, the decay was severely hindered and the reaction ceased after 10 min of illumination. This is likely due to the change in the molecular and particle structure of o-ClA and TiO2, respectively, under high concentrations of proton, which will be discussed later in detail. By comparing the initial reaction rates among various initial pH levels as shown in Figure 6, the initial decay rates were generally found to increase from a low to a high pH, but to level off to about 0.1 min-1 when the pH became greater than 7.

This feature is likely due to the surface charge of TiO2. The point of the zero charge (pzc) of TiO2 is reported to be pH 6.8, implying TiO2 carries a positive charge below that value (in an acidic medium) while it is negatively charged above pH 6.8 (in alkaline).25 Positive charges on the surface (pH < 6.8) favor the migration of photoproduced electrons and suppress electron-hole recombination, but produce comparatively fewer oxygenate radicals. On the contrary, negative charges on the surface (pH > 6.8) enhance the transportation of holes to the surface, which react with OH- and H2O and generate •OH radicals; hence, the rate of photooxidation increases. The effect is magnified when pKa is taken into account in the process at an extremely acidic medium. Recall that the cationic o-ClA is dominant at pH 2 (about 80% are [o-ClA]+, as shown in Figure 6b). The repulsive force is well established between the charged [o-ClA]+ molecules and [TiO2]+ particles, which dramatically hinders the photodegradability of the probe at extremely acidic conditions. Use of Additives in the Photocatalytic System. The oxidant H2O2 and sensitizer Rose Bengal (RB) were selected as the additives in the UV/TiO2 system to examine the possibility of rate improvement. Appropriate quenchers (as indicated in the Experimental Section) were added to each collected sample to prevent sequential reactions. A wide range of H2O2 and RB were applied, and rate acceleration (conditional) and deceleration were observed, respectively. (i) UV/TiO2/H2O2 System. The degradation rate of o-ClA tends to increase at low dosages of H2O2, and a maximum enhancement of 1.8 times was observed at

8188

Ind. Eng. Chem. Res., Vol. 44, No. 22, 2005

Figure 8. The first-order decay of o-ClA in UV/TiO2/Rose Bengal. Figure 7. Reaction rates of o-ClA at various H2O2 dosages under a fixed UV/TiO2 system (eight lamps at 300 nm were illuminated for 30 min, [TiO2] ) 0.1 g L-1).

10 mM (see Figure 7). It is suggested that the enhancement is mainly brought by the additional free radicals generated from the cleavage of O-O bond of photoexcited H2O2.26 Also, being an electron acceptor, H2O2 has a higher activity and efficiency in this role than oxygen for the titania-excited electrons. This would reduce the chances of an electron-holes recombination, which is an undesirable process in the photocatalytic process. In addition, the presence of oxygen in the water (i.e., dissolved oxygen) may induce the formation of H2O2 and, subsequently, of •OH. Therefore, the addition of H2O2 may shorten the subreactions and accelerate the process directly. Equations 8-11 show the formation of •OH radicals, of which eqs 10 and 11 represent both generated and additional H2O2:27

e-CB + O2 f •O2-

(8)

O2- + 2H+ + e- f H2O2

(9)

H2O2 + hυ f 2•OH

(10)

H2O2 + eCB- f •OH + OH-

(11)



However, the further addition of H2O2 (above 10 mM) to the solution inhibits the reaction rate. As the [H2O2] dose was increased gradually, retardation of the process was observed simultaneously. The pseudo first-order decay constant of o-ClA at 0.4 M H2O2 drops back to almost the same level as that of the sole UV/TiO2 system. This is because the H2O2 present in excess amount will consume the valuable •OH (like a scavenger) that previously formed in the solution and subsequently generate less-reactive HO2• radicals. These subreactions are shown in eqs 12-14:6

H2O2 + •OH f HO2• + H2O

(12)

HO2• + •OH f H2O + O2

(13)

H2O2 + 2h+VB f HO2• + 2H+

(14)

In addition, H2O2 may react with the photogenerated holes at elevated dosages, which will partly contribute to the rate retardation at higher [H2O2] levels.28

(ii) UV/TiO2/RB. Rose Bengal (RB) is one of the most popular anionic sensitizers used in water solutions. Its excited triplet state is efficiently quenched by oxygen showing a high quantum yield of singlet oxygen formation.29 However, it was unusual to see the occurrence of rate retardation as the dose of [RB] was increased, as shown in Figure 8. For example, with 10 mg L-1 RB in the UV/TiO2 solution, the reaction was reduced by about 4.7 times as compared to the situation in the absence of an RB sensitizer. This was partly due to the poor light absorption character of RB in the solution. The UV adsorptivity of RB is high at 450-550 nm and below 270 nm, while the sensitivity of absorbance was quite low at 300 nm. The proposed singlet oxygen production by RB was possibly suppressed under the light with 300 nm, which significantly reduced the sensitization ability. Unfortunately, the 300 nm is the best wavelength for UV/TiO2 applications. In addition, the decrease in the number of TiO2 active sites could retard the degradation process as well. RB, being an organic sensitizer, can become a competitor with o-ClA for free radicals and for the active site on the TiO2 surface (which reduces the adsorption probability of the probe compound onto the TiO2 surface). Therefore, it is believed that both the low sensitivity of RB at 300 nm and the competition for the active site are the dominant mechanisms to inhibit the photocatalytic process. Conclusion The photoinduced degradation of o-ClA in the presence of TiO2 and with various additives was studied. TiO2, one of the popular photocatalysts, generally enhances the decay of o-ClA through the generation of powerful hydroxyl radicals. However, a continuous increment of [TiO2] doses will scatter the light and slow the photoreaction. An optimal dosage of TiO2 at 0.1 g L-1 is recommended. The photodegradation of o-ClA is sensitive to the system pH. The reaction is obviously very much inhibited at strong acidic conditions (pH 2). Solution pH keeping from neutral to alkali conditions is more preferable. Care should be taken with the initial pH of the influent to ensure pH does not cause retardation of the process. The combined effects of oxidant (H2O2) and/or sensitizer (Rose Bengal) with UV/TiO2 led to totally different outcomes. Hydrogen peroxide can promote the photocatalytic process as long as the [H2O2] is not overdosed (i.e.,