Leng Wenhua, Tong Shaoping, Cheng Shao^, an, Zhang Jianqing, Cao Chunan
(Department of Chemistry, Yuquan Campust, Zhejiang University, Hangzhou, 310027,  China)

Received Jul. 30, 2000; Supported by the National Natural Science Foundation of China (29877024).

Abstract  In the present study, the photocatalytic decomposition of p-chloroaniline (PCA) in water was explored using titanium dioxide as catalysts immobilized on porous nickel with binder. The results showed that PCA can be destroyed effectively by this method in the presence of oxygen. The degradation rate is strongly dependent on the initial PCA concentration, pH and oxygen concentration. The initial quantum yield is 4.54% for PCA 8.65×10^-5 mol·dm^-3. Total mineralization required much longer illumination time than the disappearance of PCA. Without oxygen, the photocatalytic reaction could be performed via imposing external anodic bias. The degradation kinetic of PCA can be described by Langmuir-Hinshelwood equation, and the main immediate products of PCA were aniline, 4-chloronitroaniline, azobenzene, 4-4' - dichloroazobenzene and nitrobenzene, which were mineralized to NH₄⁺、Cl^-、NO₃^- and CO₂.
Keywords Photocatalytic decomposition; Titanium dioxide; Immobilization; P-chloroaniline

Titanium dioxide photocatalysis has been demonstrated a promising method ^{[1, 2]} for the pollutant treatment mainly due to its capability of complete mineralizing or at least partly destroying a variety of organic pollutants. However, much attention in this area has focused on the use of slurry system, about which the need to separate the spent fine catalysts particles and to keep the semiconductor suspended is usually necessary. Recently, some researches have tried to minimize these problems by immobilizing TiO₂ on various materials for photocatalytic reaction, such as glass ^{[3, 4]}, sea sand and gel ^[5], conductive glass ^{[6, 7]}, titanium ^[8-10] and stainless steel ^[11], etc. Most of those methods used for the adhesion of photocatalysts on supports need heating. This may bring some disadvantages: (a) some porous structure gets lost, thus inherently a decrease in the surface area available for reaction and an impedance to the transportation of solutes arises, consequently the efficiency of photocatalytic reaction will decrease; (b) it requires to control strictly the heating temperature to preserve photocatalytic activity.
    We have explored a new simple fixation technique of TiO₂ bound by binder on a porous nickel support for the photocatalytic degradation of sulfosalicylic acid that is strongly adsorbed to TiO₂ and nickel, the results is satisfactory ^[12]. Porous nickel was chosen as support in view of its following characteristics: (a) high porosity suitable for catalyst fixation and the transportation of solutes; (b) good conductivity capable of imposing an external potential bias on it to improve the photogenerating carries separation, and (c) low cost and wide application in industry, for example as current collector in batteries.
    PCA is commonly produced as by-product of some petroleum and chemical industries, is highly toxic. The studies of its degradation have been investigated by many methods such as hydrogen peroxide ^[13] and electro-oxidation ^[14] and the photocatalytic degradation in slurry system ^[15]. In this study, we report its photocatalytic degradation over immobilized TiO₂. The effect of the initial PCA concentration, pH, oxygen concentration, temperature and external potential on the photocatalytic reaction rate was investigated. The quantum yield of degradation reaction and the formation of Cl^-, NO₃^-, NH₄⁺ ions and CO₂ were also measured. The degradation kinetic and mechanism were also discussed.

1 EXPERIMENTAL
1.1 Materials
TiO₂ (C. P. 99.9%, Shanghai) was used as received. Polyvinyl alcohol (PVA 124) used as binder was purchased from Japan. Porous nickel sheet (thickness about 0.2 cm) was used as support. PCA were analytical grade and used without further purification; KNO₃ were governmental reagent; H₂O₂ were 29 wt. %. The other chemicals and solvents were reagent grade and used without further purification. Deionized and double distilled water was used throughout the work.
1.2 Immobilization of catalysts
The procedures of fixation of TiO₂ were similar to that described in ref.12, i. e., TiO₂ was mixed with 3-wt. % PVA at the ratio of 3 to 2. It was coated on one side of a 15×16 cm porous nickel sheet. The coated sheet (denoted hereafter as TiO₂/Ni) was dried at 60^°C for 4 hour and packed to thinner sheet in the thickness of 0.09 cm. Then it was rinsed with saturated Na₂CO₃ to remove grease and/or photooxidation products and washed with water to release the loosely bound particles and dried by heating at 60^°C prior to use or reused. In photoelectrochemical experiments, the nickel double sides covered with TiO₂ was fixed on the outer surface of the borosillicate glass tube as working photoelectrode (Blank porous nickel as counter electrode, a double-junction saturated calomel electrode as reference, 0.001 mol dm^-3 NaClO₄ as support electrolyte). The bare nickel in same thickness was used as blank tests. The difference in weight between the coated and bare films gave the TiO₂ weight as listed per gram nickel in the following experiments. We may not obtain the same weight of TiO₂, surface area, and etc on different TiO₂/Ni sheet, so the results can not compare with each other in the text. But for comparison the contrast experiments were conducted on the same TiO₂/Ni sheet.
1.3 Characterization of TiO₂/Ni
X-ray diffraction analysis (XRD) was carried out with CuKα radiation on a Rigaku D/max 3B diffractometer (working at 45 kV, 40 mA, at a speed of 10 deg min^-1 and a step 0.02 deg. ).
    The scanning electron microscopy (SEM) was carried out on a low vacuum SEM Hitachi (S-570) with an energy dispersive x-ray (EDX) analysis attachment.
1.4 Photoreactor and light source  
Photoreactor was an annulus reactor with a F6.5 cm cylindrical borosillicate glass inner tube (Thickness about 3 mm) around which the TiO₂/Ni was wrapped. The light source (4×6 W UV lamp, Emax=365nm) was fixed on the central axis of the tube. The number of the incident light (300-400 nm) inside the photoreactor in TiO₂/Ni free solutions measured employing potassium ferrioxalate actinometry [16] was 1.4×10^-7 mol·s^-1.
1.5 procedure and analyses    
The solution (180 cm³) was placed in the annular region (width about 0.5 cm) between the tube and the double-walled glass outer jacket (cooling at 25+1^°C), magnetically stirred and kept purged at the side of photoreactor with air at a rate of 16 dm^-3 h^-1 and initial pH 7.0 unless otherwise stated. The PCA concentration in the solution was estimated colorimetrically by the modified method ^[17] (coupling time was 2 hour). The degradation products NO₃^- and NH₄⁺ were estimated by indophenol blue ^[18] and phenate method ^[18], respectively. Chloride ion concentrations were measured with a chloride ion-selective electrode and double-junction reference electrode by withdrawing 20cm³ solution from the reactor and returning after measurement. HClO₄ and/or NaOH adjusted the pH of solution before reaction.
    Organic solutions were analyzed by gas chromatography/mass spectrometry (GC/MS, HP 6890/5973) with a HP5 column (30 m×0.25 mm). The oven was programmed as follows: isothermal at 45^°C for 2 min, from 45^°C -250^°C at 6^°C min^-1, and isothermal at 250^°C for 10 min. In the electron impact experiments, the electron energy was set at 70 eV and the electron current was set at 1 mA.

2 RESULTS AND DISCUSSION
2.1 Characterization of TiO₂/Ni      
XRD analysis showed that TiO₂ powder was almost anatase phase and the average particle size was about 114 nm calculated by Scherrer formular. The specific surface of TiO₂ was 14.98 m² g^-1 (measured on Omnisorp 100 CX, Coulter, USA). The texture and morphology of the porous nickel and TiO₂/Ni can be observed in Fig. 1. SEM analysis showed that the nickel substrate was meshy and high porous (95%), the diameter of pores was in the range of from 5mm to 50m m. Thus it is favorable for the fixation of TiO₂. TiO₂ was dispersed in the top of the pores of the nickel sheet but there remained many pores inside of the TiO₂/Ni, The diameter of its pore was less than 10 mm.
2.2 Influence of the initial concentration
It is found that a negligible decrease in the concentration of PCA for 8h was observed both in the presence of TiO₂ without UV light irradiation (3% degradation) and without TiO₂ but with illumination (8% degradation). However, PCA can be degraded quickly over illuminated TiO₂ (as shown in Fig. 2), it resulted obviously from photocatalytic oxidation. The rate increase with increasing PCA concentration as indicated in Fig. 2b and the photocatalytic degradation reaction follows a quasi-first-order in the range of a studied PCA concentration, which is concluded from the straight-line relationship of ln (c_o/c) vs. irradiation time as shown in Fig. 2a. The slope of the plot gives the apparent first-order rate constant k_app as listed in Table 1. It is clear that the apparent rate constants decrease with increasing PCA concentration. This can be explained by assuming that the photoproducts were competition for the surface site of TiO₂ with PCA. Table 1 also listed the half-lives and quantum yields of the photodegradation of PCA as a function of initial concentration.

    
                 (a)                                               (b)
Fig.1 SEM micrographs for (a) the porous nickel and (b) TiO₂/Ni

	(a)
	(b)

Fig.2 (a) Pseudo-first-order kinetics for PCAs in photocatalytic degradation (b) Linearized reciprocal kinetic plot for the photocatalytic degradation of PCA（Experimental conditions: TiO₂, 1.48g/g (Ni)

Table 1. Apparent first-order rate constants k_app, half life t_1/2, t_1/2', quantum yield j and linear coefficient r² for the degradation of PCA at different initial concentrations (Experimental conditions see Fig. 2)

2.3 Influence of the initial pH
Fig. 3 shows the dependence of degradation of PCA on the initial pH solution. The increasing r₀ with increasing pH can be attributed to the increase number of OH^- ions at the surface of TiO₂ ( OH^· radical can be formed by trapping photoproduced holes). Also the dissociation of PCA probably changes its reactivity. Claire [19] found that OH^· radicals were the sole oxidant under the condition of pH 11 and its role was larger in this case than in neutral and acid medium. So r₀ increases rapidly with increasing pH higher than 11. But at the lowest pH (to avoid Ni dissolving, lower than pH 4 not investigated) the rate is influenced at least by two factors. One in the case it is lack of OH^- ions, thus produce less OH^· radicals, the rate decreases; another it can provide more H⁺ that contributes the formation of HO₂^{- [20]}, thus the rate increase. In the experiments the rate increases slightly that may be attribute to the dominant role of the latter. In general, the results indicate the efficiency of the process is not much affected over a wide moderate range of pH, which is quite satisfactory in view of applications.

Fig.3   Effect of the initial pH on the initial rate of PCA in photocatalytic degradation (Experimental conditions: C0, 11.04 mg dm-3; TiO2= 1.48g/g (Ni); UV 8h)

2.4 Influence of the oxygen
The influence of oxygen on the degradation of PCA is shown in Table 2. It is obvious that the oxidation rate of PCA increases with increasing oxygen concentration. This has well been demonstrated early by many studies ^{[1, 2]}. Oxygen is often as an electron acceptor without which the photocarriers will be totally combined and it may be also an oxidant in the further process of the hydroxyl of organic substance. In addition, adsorptive oxygen at the surface of particles is another source of OH^· radical via a series of chemical process ^[20].
    It also can be observed from the Table 2 that the formation rate of chloride ions is larger than the disappearance of PCA both in the case of air and oxygen purge. Similar results were obtained in the degradation of chloramben ^[21]. They explained that hydroxyl radicals usually attack the aromatic substrates to form the corresponding OH^· adducts still adsorbed onto the oxide particles where they can interact with the trapped electrons to release chloride.
2.5 Influence of the temperature
Table 3 lists the apparent rate constant as a function of temperature for PCA degradation 8h. The Arrhenius equation allowed a calculation of the activation energy of the PCA degradation of 6.12 kJ·mol^-1 (r=0.993), which is less than the value of 10 kJ mol^-1 obtained for phenol ^[20], 11 kJ mol^-1 for salicyclic acid ^[3], 13 kJ mol^-1 for oxalic ^[24]. This indicated that PCA is easily degraded.

Table 2. Effect of oxygen content on the half life of PCA (t_{1/2, PCA}) and chloride removal (t_{1/2, cl}) for the PCA photocatalytic degradation (C_{0, PCA} =112.41mg dm^-3; TiO₂= 1.98 g/g (Ni))

Table 3. Apparent first-order rate constants k_app, as a function of temperature (T) for the PCA degradation (Experimental conditions: C₀, 11.47 mg dm^-3; TiO₂=1.93 g/g (Ni); UV 8h)

2.6 Formation of nitrate ions and ammonium ions
Photoproducts may be more toxic than the solute itself, so totally destroying the pollutants is necessary. For evaluate what the extent of PCA mineralization performed, photocatalytic degradation was conducted and the results are shown in Fig.4. It shows that the PCA was completely disappeared when illuminated for 10 h, while only 51.0% and 20.0% of NH₄⁺ and NO₃^- was formed respectively. The formation of NH₄⁺ ions was rather quick in initial stage and then became slow after reaction for 7h and 30min, while the formation of NO₃^- increased slowly all the time with increasing illumination time. These indicated that the remaining of nitrogen exists in others compounds and the oxidation of nitrogen to NO₃^- may occur via the intermediate formation of NH₄^{+ [15]}. Similar results were found in photocatalytic degradation of PCA in slurry system when the illumination time is not longer enough ^[15]. Fig.4 also shows that the degradation of COD is a more slow process. All these indicated total mineralization requires a much longer illumination time than the disappearance of PCA in the experiment. It may be attributed to that the photoproducts were difficult to degrade and the light intensity was rather poor.
2.7 Quantum yield
The quantum yield is an important parameter for photocatalytic process in practical application and it can be defined as the ratio of the number of molecules of PCA reacting to the number of photons supplied. Assuming that all the incident light was absorbed by the titanium dioxide, under PCA 8.65×10^-4 mol·dm^-3 (other experimental conditions see Fig. 2), the initial rate of PCA was found to be 3.81×10^-7 mol·dm^-3 min^-1. Therefore the initial quantum yield for the degradation of PCA is 4.54%.
2.8 Influence of the external potential bias
The photocatalytic reaction cannot proceed in the presence of nitrogen purge. For this case, a test of photocatalytic degradation of PCA by applying an external anodic bias was conducted and the results are shown in Fig. 5. It was found whether in the presence of UV and external potential or with external potential alone, the reactions followed a quasi-first-order expression. But the apparent first-order rate constant k_app of the former is larger than that of the latter. For example, by exerting anodic 700 mV the apparent quantum yield was 13.75% (not including the part of direct electro-oxidation). The trend of between apparent first-order rate constants k_app and pH under potential bias 600 mV as shown in Table 4 is consistent with that of no potential bias (see Fig. 3). These indicate that the anodic bias can be efficiently improved photocatalytic reaction. The reason is that bias potential applied to the photocatalyst film had separated the photogenerated carriers thereby improving the quantum yield ^{[6, 7]}.
    So far we can see that this approach has the advantage of not only the ease of treating the photocatalyst after use but also capable of applying a bias potential to improve the efficiency of PCA photodegradation with nitrogen on TiO₂/Ni. By this to study the photoelectrochemistry behavior of organic substrates in details is under way in our laboratory.

	Fig.4 Ions formation, CODcr and PCA degradation (Experimental conditions: C0, 22.70 mg dm-3; pH 8.70; TiO2= 2.05 g/g (Ni); oxygen purge)
	Fig.5 Effect of applied potential (Vs SCE) on the photocatalytic degradation of PCA (Experimental conditions: C0, 9.80 mg dm-3; TiO2=3.95 g/g (Ni); N2 purge)

Table 4 Apparent first-order rate constants k_app and linear coefficient r for the PCA degradation 3h
at different pH with potential bias 600 mV (N₂ purge, other experimental conditions see Fig. 8)

2.9. Kinetic and mechanism aspects
Langmuir-Hinshelwood (L-H) equation is often used to express the heterogeneous photocatalytic reaction ^{[1-3, 12, 25]}. For the conditions in our experiments, then it is given by
r =(+) (1)
where r (mg dm^-3 h^-1）is the reaction rate for the oxidation of PCA, c（mg dm^-3 ）is the concentration of PCA, k_r（mg dm^-3 h^-1）is the specific reaction rate constant, k_a (dm³ mg^-1) is the equilibrium adsorption constant of PCA, c_w (mg dm^-3 ) is the concentration of solvent, k_w (dm³ mg^-1) is the equilibrium adsorption constant of solvent, c_i（mg dm^-3）is the concentration of products, k_i (dm³ mg^-1) is the equilibrium adsorption constant of products and t (h) is the reaction time.
    By integrating Eqn. (1), we get：
t = (2)
If c₀ is small, then the adsorption of its product is negligible compare to reactant, i.e., . By substituting c=c₀ into Eqn.（1）and (2), Eqn.（3）and (4) can be obtained, respectively.

（3）

（4）

As above mentioned,  c₀ is small, so the second item in equation (4) can be neglected; for common system,  c_w is a constant and it is much more larger than  c, thus Eqn. (4) can be written into: （5）

When it was plotted by and from equation（3）and (5), straight line could be obtained, respectively. In addition, while by substituting   c₀ c =c₀/2 into equation (4), we can get:
（6）
Straight line can also be obtained when it was plotted by vs c_o from Eqn.（6）.
    For our results as indicated in Fig.2a, the straight-line relationship of ln (c₀/c) versus irradiation time were observed. The plot of the reciprocal initial rate r₀^-1 as a function of the reciprocal initial concentration C₀ yields a straight line (see Fig.2b). The linear transform of this expression yields k_r=8.622 mg dm^-3 h^-1 and k_a=4.225×10^-2 dm³ mg^-1（r =0.9996）. By substituting them into Eqn. (6), the estimated half-lives are obtained and listed in Table 1. The dependence of on the initial concentration of PCAs is also a straight line: , (r²=0.995). So the photocatalytic oxidation of PCA in the system can be well expressed by L-H model.
    It has been proved ^{[1, 2, 20]} that photogenerated holes oxidize water or adsorbed OH^- at the surface of semiconductor to hydroxyl radicals. These highly reactive radicals can then be used to mineralize or at least partially degrade most organic pollutants. At the same time the holes can also react with PCA directly to produce PCA cation radical. Due to the complexity of the radical-induced reactions occurring in photocatalytic process, it is difficult to indicate an exhaustive reaction scheme explaining the formation of all detected intermediates. However, a relatively low number of rather abundant aromatic transients have been recognized during the photocatalytic reaction, so that a tentative and simple scheme can be drawn, with considering into the usual transformation processes of other aromatics. We have obtained 5 intermediates during the initial photocatalytic degradation of PCA by GC/MS as listed in Table 5. Azobenzene (4- 4'-dichloroazobenzene) was formed from the coupling of neutral free radical produced through the deprotonation of the aniline (PCA) cation radical which was probably generated by reacting with hole^[14]. Generation of NH₄⁺ is considered to be produced by the hydrolytic degradation of imino intermediates leading to the corresponding benzoquinone ^[14]. But we did not detect this compound by GC/MS. Brillas et al. ^[13] also did not find it detected by GC/MS and HPLC during the electrochemical oxidation of PCA. They did not think that this compound could be ruled out as intermediate due to its rapidly decomposed by different paths. Hydroxyl radicals attack the aromatic compound should be an important step to produce hydroxyaromatic intermediates in photocatalytic reaction, for example, hydrohydroquinone is the main intermediate product detected for phenol ^[20] and hydroxyaniline for aromatic amines ^{[15, 25]}. However, they are usually unstable and difficult to identify during the initial steps of the treatment.
    So based on the above GC/MS results we preliminary proposed the reaction scheme as shown in Fig. 6.

Fig. 6 Supposed photocatalytic pathway of PCA

Table 5 Identification of photocatalytic immediates of PCA
(PCA 50 mg dm^-3; pH12.5; oxygen purging; UV 1h)

2.10 Further remarks about the TiO₂/Ni
The coming off of powdered TiO₂ from the substrate was not observed during the process of photocatalytic degradation. Ni²⁺ concentration in solution measured by atomic absorption was less than 1 mg L^-1 when pH higher than 7.0 but less than 70mg L^-1 at pH 1.8 due to the dissolving for 7 days of the porous nickel. This is because of the inhibition of TiO₂ on bare nickel. Contrast tests were conducted in slurry system (not shown here) and the results showed that there was no catalysis of Ni²⁺ in solution for photocatalytic reaction. The physical chemistry properties of the Ni/TiO₂ were stable. Little change was observed even after using for 80 h and the reproducibility of the parallel experiments was tested to be satisfactory (less than 5%).
    Unfortunately, PVA is also a sacrificial electron donor ^[26]. Thus it reduces the number of photoactive substances for its competition reaction to OH^· or positive hole. Therefore, it is not a stable and good binder in practical applications. We have ever used polytetrafluoroethylene (PTFE) as a binder but the results were not satisfactory due to its poor performance of binding. It appears to be necessary to probe more suitable binders and this is under way in our laboratory.

3 CONCLUSIONS    
It has been shown that it was possible to destroy PCA over TiO₂ immobilized on porous nickel by binder in photocatalytic reaction. This is a fairy simple method and potential used as immobilized catalysts working in batch photoreactor from the point of application with advantage of eliminating the final filtration of TiO₂ fine particles in suspension. In the absence of oxygen, external anodic potential bias can make photocatalytic reaction proceed. These may provide a quite desirable approach to degrade organic substrates. It is worthy to point out that photocatalysis produces more toxic photoproducts such as nitrobenzene and 4-chloronitroaniline during the reaction, so it needs to take measures to totally eliminate the photoproducts in its applications.

REFERENCES  
[1] Hoffman M R, Martin S T, Wonyong C et al. Chem. Rev., 1995, 19: 69.
[2] Oills D F. Environ. Sci. Technol., 1985, 19: 480.
[3] Matthews R W. J. Phys. Chem., 1987, 91: 3328.
[4] Preis S,   Krichevskaya M, Kharchenko A. Wat. Sci. Technol., 1997, 35: 265.
[5] Matthews R W, Mcevoy S R. Solar Energy., 1992, 49: 507.
[6] Vinodgopal K, Hotchandani S,  Kamat P V. J. Phys. Chem., 1993, 97: 9040.
[7] Hidaka H, Asai Y,   Jincai Z et al. J. Phys. Chem., 1995, 99: 8244.
[8] Pelegrini R, Peralta-Zamora P,  Andrade A R D et al. Appl. Catal. B: Environ., 1999, 22: 83.
[9] Butterfield I M, Christensen P A, Hamnett A et al. J. Appl. Electrochem., 1997, 27: 385.
[10] Byrne J A, Eggins B R. J. Electroanal. Chem., 1996, 457: 61
[11] Fernandez, Lassaletta G, Jimenez V M et al. Appl. Catal. B: Environ., 1995, 7: 49.
[12] Liu Hong, Cheng Shao'an, Zhang Jiangqing et al. Chemosphere, 1999, 38: 283.
[13] Brillas E, Bastida R M,  Uosa E. J. Electrochem. Soc., 1995, 142: 1733.
[14] Hand R L, Nelson   R F. J. Electrochem. Soc., 1978, 125: 1059.
[15] Pramauro E, Prevot A B. Analyst, 1995, 120: 237.
[16] Pruden A L, Ollis D F. J. Catal., 1983, 82: 404.
[17] Daniel J W. Analyst, 1961, 86: 640.
[18] Analysis method for environmental pollutants, Vol.1: inorganic substance (in Chinese). 1987. Beijing: Kexue Press.
[19] Claire R, Boule P. J. Photochem Photobiol A: Chem., 1991, 60: 235.
[20] Ken-ichi Okamoto, Yasunori Yamamoto, Hiroki Tanaka et al. Bull. Chem. Soc. Jpn., 1985, 58: 2015.
[21] Prevot A B, Vincenti M, Bianciotto A et al. Appl. Catal. B: Environ., 1999, 22: 149.
[22] Abdullah M, Lew G KC, Matthews R W. J. Phys. Chem., 1990, 94: 6820.
[23] Lew G KC, MeEvoy S R,  Matthews R W. Environ. Sci. Technol., 1991, 25: 460.
[24] Herrmann J M, Mozzanega M N,   Pichat P. J. Photochem., 1983, 22: 333.
[25] Sánchez L, Peral J, Domènech X. Electrochim. Acta, 1997, 42: 1877
[26] Brown G T, Darwent J R, Fletcher P D I. J. Am. Chem. Soc., 1985, 107: 6446

[ Back ] [ Up ] [ Next ] Mirror Site in USA  Europe  China  CSTNet ChinaNet