Environ. Sci. Technol. 2002, 36, 5229-5237
Environmental Remediation by an Integrated Microwave/UV Illumination Technique. 3. A Microwave-Powered Plasma Light Source and Photoreactor To Degrade Pollutants in Aqueous Dispersions of TiO2 Illuminated by the Emitted UV/Visible Radiation SATOSHI HORIKOSHI AND HISAO HIDAKA* Frontier Research Center for the Global Environment Protection (EPFC), Meisei University, 2-1-1 Hodokubo, Hino, Tokyo 191-8506 Japan NICK SERPONE* Department of Chemistry and Biochemistry, Concordia University,1455 de Maisonneuve Boulevardd West, Montre´al, Que´bec, Canada H3G 1M8
The characteristic features of a novel double-quartz cylindrical plasma photoreactor (DQCPP) were assessed by examining the photodegradation of rhodamine-B dye (RhB+) in aqueous TiO2 dispersions irradiated simultaneously by both microwave radiation and UV/visible radiation emitted from a microwave-powered (MW, 2.45 GHz) electrodeless mercury lamp. The features of the DQCPP lamp are given and discussed in terms of the experimental output UV energy in the wavelength ranges 210-300 and 310-400 nm for applied MW powers from 74 to 621 W. The DQCPP and a watercooled DQCPP reactor absorbed more than 50% MW radiation (50-88 and 50-75%, respectively). The emitted light irradiance scaled sublinearly with applied MW power. Relative to the DQCPP lamp, loss of irradiance by the watercooled DQCPP lamp was ∼28-46% at 250 nm and ∼4158% at 360 nm in the range of MW power used. The smallest loss occurred at 178.9 W at which the degradation of RhB+ was subsequently examined by UV/visible spectroscopy and by total organic carbon analyses. Highly intense mercury lines were seen at 365, 404, 435, 546, and 579 nm (those below 365 nm were more than 10 times weaker). About 80% of the RhB+ solution was photomineralized after 60 min of irradiation of the aqueous RhB+/TiO2 dispersion with the DQCPP lamp; no UV/ visible spectral features of RhB+ were evident at wavelengths below 250 nm after 30 min. Possible effects of microwave radiation and temperature on the degradative process are described.
Introduction Water-soluble pollutants are easily decomposed in wastewater treatment by such oxidative species as •OH radicals * Corresponding authors. E-mail:
[email protected];
[email protected]. 10.1021/es020506c CCC: $22.00 Published on Web 10/23/2002
2002 American Chemical Society
generated in suitably irradiated aqueous TiO2 dispersions. Several excellent reviews (1-7) have described details of the photocatalytic degradation of such organic pollutants as chlorinated and phenolic compounds. By comparison, hydrophobic materials such as endocrine disruptors (8, 9), plastics (10-12), cationic surfactants (13), and cationic dyes and stable symmetrical substances possessing the triazine skeleton (e.g., atrazine (14)) are relatively nondegradable even in highly active photocatalytic systems. The photodegradation of the latter materials can be improved significantly using other technologies in concert with TiO2 photocatalytic degradation. For example, TiO2 photocatalytic degradation occurring in supercritical water represents a useful tandem combination of two different technologies to degrade two otherwise difficult substances such as atrazine and cyanuric acid (15). The solubilization and ultimate decomposition of cationic dyes in nonionic surfactant micelles can be improved by contact with positively charged TiO2 particles (16-18). The rate of degradation of rhodamine-B (RhB+) is significantly enhanced when the aqueous TiO2 dispersion is exposed to both microwave and UV light radiation (19). In general, the photocatalytic degradative process is made more effective whenever adsorption of simple and diluted polluting components of a feed (e.g., aldehydes and carboxylic acids) is facilitated on the particle surface of the TiO2 photocatalyst. Typically, relatively high concentrations of oxygen and low concentrations of organic substrates in the aqueous waste solution are an essential requirement in wastewater treatment that uses traditional lamps as the source of UV radiation. Among the various factors that affect the degradative process, the following three points are significant in the construction of a practical water treatment device: (1) the extent of adsorption of the organic contaminants on the TiO2 surface at the high concentrations typically encountered in highly loaded waste streams; (2) the relatively slow permeation of the pollutants in the stream; and (3) the limitation of UV light to penetrate and irradiate the photocatalyst. The latter is due to the inherent presence of other extraneous components in a muddy stream and the effect on the TiO2 particle surface in bulk water as the degradative process takes place without significant movement of the organic contaminants. Disposal of polluted wastewaters and the need for drainage are also important considerations in environmental remediation. In the concerted system of microwave radiation and UV/TiO2 photocatalysis, photodegradation occurs even under experimental conditions that use a small amount of the TiO2 catalyst, a relatively weak UV light source, and low oxygen concentrations (19). Accordingly, a method that can treat larger quantities of pollutants in wastewaters is conceivable by a hybrid combination of the microwave method and the photocatalytic technology. Microwave plasma is commonly employed as a useful etching tool for materials in chemical vapor deposition (CVD) and in surface processing of polymers and semiconductors, among others (see, e.g., refs 20 and 21). It has even found a niche in organic synthesis and degradation processes (22), in the removal of NOx (23) and in the degradation of chloro and nitro aromatics in an electrohydraulic discharge (24). In part 1 of this series (19), we reported one possible hybrid device, which could illuminate an aqueous system simultaneously by both microwave radiation and UV light. The degradation of a number of pollutants from carboxylates to surfactants, dyes, phenols, alcohols, and polymers was VOL. 36, NO. 23, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5229
FIGURE 1. (a) Experimental setup of a double-quartz cylindrical photoreactor (DQCPP) used in the photocatalytic decomposition of RhB+ dye in aqueous TiO2 dispersions using integrated UV/visible and microwave (MW) radiations; (b) similar experimental apparatus equipped with a UV/visible light source {a 250-W Hg lamp} and a source of MW radiation. examined in a batch reactor irradiated externally by UV light or microwaves from a magnetron. The integrated microwave(MW)/UV illumination method (PD/MW) was compared to the TiO2 photocatalytic method alone (PD), the microwave irradiation in the absence of TiO2 particulates (MW), and the thermally assisted TiO2 photocatalytic method (PD/TH). Not only did the PD/MW technique prove superior to the other three methods, its greater efficacy was also observed under limited quantities of molecular oxygen and low radiant excitance of the light source. In the present study, we examined the characteristic features of a novel photoreactor device different from that used earlier (19, 25). It uses an electrodeless double-quartz cyclindrical plasma photoreactor (DQCPP) lamp embedded inside a microwave oven in which the microwave radiation activates the lamp. The emitted UV/visible radiation in tandem with the microwaves drives the chemistry in such 5230
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 23, 2002
a photoreactor. We investigated the photodecomposition of
the cationic RhB+ dye to test the practicality and usefulness of this DQCPP device against a UV/visible light source positioned outside the microwave oven without worrying at this stage about the quantum efficiencies or engineering efficiencies and, thus, the economics. These will have to await further work. Possible effects of the microwave radiation and
TABLE 1. (a) Characteristics of the Microwave-Powered DQCPP light intensity (mW/cm2) relative light intensity power (W)
GIB with DQCPP (mA)a
MW-abs of DQCPP (mA)b
at 250 mm
at 360 mm
(250 mm) with MW power (×10-5)c
(360 mm) with MW power (×10-5)d
(250 mm) with GIRe
(360 mm) with GIRf
87.8 126.8 15.9 191.9 243.9 269.9 296.0 335.0 374.1 400.1 439.0 543.0 621.0
0.040 0.060 0.065 0.070 0.080 0.100 0.110 0.110 0.120 0.110 0.120 0.160 0.150
0.020 (50) 0.040 (67) 0.045 (88) 0.050 (71) 0.032 (40) 0.065 (65) 0.065 (59) 0.050 (46) 0.055 (46) 0.063 (57) 0.075 (63) 0.111 (69) 0.105 (70)
0.013 0.016 0.018 0.021 0.024 0.027 0.030 0.034 0.038 0.042 0.046 0.048 0.052
0.0183 0.0225 0.0253 0.0285 0.0322 0.0379 0.0437 0.0484 0.0554 0.0612 0.0669 0.0715 0.0753
14.806 12.618 10.233 10.943 9.840 10.004 10.135 10.149 10.158 10.497 10.478 8.840 8.374
20.844 17.745 14.383 14.853 13.202 14.042 14.764 14.448 14.809 15.296 15.239 13.168 12.126
0.325 0.267 0.277 0.300 0.300 0.270 0.273 0.309 0.317 0.382 0.383 0.300 0.347
0.458 0.375 0.389 0.407 0.403 0.380 0.397 0.440 0.462 0.556 0.558 0.447 0.502
(b) Characteristics of the Microwave-Powered-DQCPP under Circulated Water relative light intensity light intensity (mW/cm2) power (W)
GIB with DQCPP (mA)a
MW-abs. of DQCPP (mA)g
at 250 mm (mW/cm-5) and decrease rateh
at 360 mm (mW/cm-5) and decrease rateh
(250 mm) with MW power (×10-5)c
(360 mm) with MW power (×10-5)d
(250 mm) with GIRe
(360 mm with GIRf
74.7 126.8 178.9 204.9 230.9 243.9 283.0 335.0 361.0 413.0 439.0 517.0 595.4
0.035 0.050 0.050 0.060 0.070 0.080 0.090 0.100 0.090 0.100 0.110 0.140 0.180
0.015 (43) 0.030 (60) 0.030 (60) 0.040 (67) 0.022 (31) 0.045 (56) 0.045 (50) 0.040 (40) 0.025 (28) 0.053 (53) 0.065 (59) 0.091 (65) 0.135 (75)
0.007 (46) 0.009 (44) 0.013 (28) 0.014 (33) 0.017 (29) 0.017 (37) 0.020 (33) 0.021 (38) 0.021 (45) 0.023 (45) 0.025 (46) 0.027 (44) 0.029 (44)
0.0089 (51) 0.0105 (53) 0.0150 (41) 0.0153 (46) 0.0170 (47) 0.0215 (43) 0.0236 (46) 0.0240 (50) 0.0245 (56) 0.0265 (57) 0.0283 (58) 0.0363 (49) 0.0363 (52)
9.371 7.098 7.267 6.833 7.361 6.970 7.067 6.269 5.817 5.569 5.695 5.222 4.871
11.914 8.281 8.385 7.467 7.361 8.815 8.339 7.164 6.787 6.416 6.446 7.021 6.097
0.200 0.180 0.260 0.233 0.243 0.213 0.222 0.210 0.233 0.230 0.227 0.193 0.161
0.254 0.210 0.300 0.255 0.243 0.269 0.262 0.240 0.272 0.265 0.257 0.259 0.202
a GIR, gap between incident radiation minus reflected radiation. b (GIR with DQCPP) - (GIR without DQCPP), and % ) 100(MW-abs of DQCPP)/ (GIR with DQCPP). Values in parentheses are percent. c {mW cm-2 at 250 nm per watt of MW power}. d {mW cm-2 at 360 nm per watt of MW power}. e {mW cm-2 at 250 nm per GIR with DQCPP and water (mA)}. f {mW cm-2 at 360 nm per GIR with DQCPP and water (mA)}. g (GIR with DQCPP) - (GIR without DQCPP and water),and water)/(GIR with DQCPP and water). Values in parentheses are percent. h % ) 100(100 - (light intensity with DQCPP and water)/(light intensity with DQCPP)). Values in parentheses are percent.
ensuing heat consequences on the overall degradative process are inferred.
Experimental Conditions Materials. The chloride salt of the cationic rhodamine-B dye was supplied by Tokyo Kasei Co. Ltd. The TiO2 catalyst specimen was a gift from Degussa {P-25; BET surface area, 53 m2 g-1; particle size, ∼20-30 nm by TEM; crystallinity, 83% anatase and 17% rutile by X-ray diffraction}. Photocatalytic Degradation of RhB+ Using the DQCPP Lamp. Shikoku Instrument Co. Ltd supplied the microwave irradiation equipment. It consisted of a microwave generator (2.45 GHz; maximum power, 1.5 kW), a three-stub tuner, a power monitor, and an isolator made by Shibaura Mechatronics Co., Ltd. The internal dimensions of the stainless steel MW oven were 49 cm (width) × 30 cm (height) × 44 cm (depth). To obtain maximal power, the input voltage of the MW generator was 200 V and the current was 13.5 A. However, the actual utilizable power of the MW generator was 1.5 kW, since 44.4% of the power was converted to heat. The incident and reflected MW radiations were determined using a NR-1000 data acquisition monitor (Keyence Co. Ltd.) connected to a personal computer (PC). The optimal conditions of low reflection of the MW radiation intensity were achieved using the three-stub tuner as illustrated in Figure 1a. The DQCPP [dimensions: 19 cm (length) × 7 cm (external diameter) × 1 cm (internal diameter)] contained mercury gas plus a very small quantity of Ar gas introduced as a purge gas in the DQCPP lamp device after bringing the system to vacuum (∼10-7 Torr). The 19-
cm quartz pipe photoreactor was placed in the center of the DQCPP device. It was connected to a Teflon tube (2 m long and i.d. 9 mm) through which the aqueous RhB+/TiO2 dispersion was circulated by employing a peristaltic pump (150 mL; flow rate, 2000 mL min-1; RhB+ concentration, 0.05 mM; TiO2 loading, 300 mg; pH 5.51). The temperature of the RhB+ solution/TiO2 particles was maintained at ∼20-22 °C with a cooling water circulator located outside the MW oven. The DQCPP device was illuminated by continuous microwave radiation, and unless noted otherwise, the MW power was 178.9 W (see below and Table 1). Photocatalytic Degradation of RhB+ by Microwave and Hg Lamp Irradiation. The effect of the DQCPP device on the photodegradation of rhodamine-B was examined by comparing it with the hybrid combination of a normal Hg lamp and microwave radiation only insofar as to test an external UV light source. The contents in the quartz pipe photoreactor were microwave-irradiated from the same right side as the waveguide. The UV source (Ushio 250-W mercury lamp; power, ∼0.25 kW ) 41.7 V × 6 A) was located on top of the MW oven (see Figure 1b). It delivered a light flux of 1.298 mW cm-2 in the wavelength range 210-300 nm and 1.969 mW cm-2 between 310 and 400 nm (emission detected at 250 and 360 nm, respectively). The microwave power was 178.2 W, and the temperature of the aqueous solution was 22 ( 1 °C (see above). The UV/visible wavelength spectral outputs of the DQCPP lamp and the Hg lamp were recorded using a Fastevert S-2400 UV spectrophotometer (Soma Optics Ltd.). The temporal decrease of the absorption spectral features during the VOL. 36, NO. 23, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5231
FIGURE 2. Variations in the incident and reflected radiations with increasing MW power for (a) a device that includes a MW oven (control experiment), (b) a naked DQCPP system located in the MW oven, and (c) a water-cooled DQCPP system. (d) Percent absorption of microwave radiation as a function of applied MW power at the three conditions above. degradation of RhB+ was monitored on a JASCO model V-560V UV/Vis/NIR spectrophotometer. Total organic carbon (TOC) analyses were performed with a Shimadzu TOC-5000A analyzer. The prevailing temperatures in the photoreactor were measured using a Thermal Vision Laird 3A radiation thermometer (Nikon Co. Ltd.). Microsoft Excel 2000 was used to calculate the correlation coefficients.
Results and Discussion Characteristics of the Microwave-Powered DQCPP Device. The overall features of the MW device, the DQCPP, and of the corresponding water-cooled DQCPP/H2O configuration are illustrated in Figure 1a. The resulting quantitative characteristics are summarized in Table 1 for an applied microwave power ranging from 74 to 621 W. Figure 1b illustrates a similar setup when the more traditional Hg lamp arrangement was used to supply the UV/visible wavelengths coupled with the microwave radiation. The magnitude of the applied microwave radiation to power the DQCPP and the DQCPP/H2O systems differed slightly. Typically, the extent of absorption of microwave radiation by the DQCPP reactor was greater than 50% (range 50-88%) at all the applied MW powers, except for 243.9 (40% absorbed), 335.0 (46%), and 374.1 W (46%). Column 3 of Table 1a also reveals that the greatest absorption of MW radiation occurred at 175.9 W. By comparison, the quantity 5232
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 23, 2002
FIGURE 3. Ultraviolet and visible wavelengths emitted by the microwave-powered electrodeless DQCPP lamp at applied MW powers of 87.8-621 W and by a traditional 250-W Hg lamp (see top of figure). of microwave absorption for the water-cooled DQCPP/H2O system was also greater than 50%, except when the applied microwave power (range 74.7-595.4 W) was 74.7 (43%), 230.9 (31%), 335.0 (40%), and 361.0 W where it was 28%. For the cooled DQCPP, the greatest MW absorption was observed at an applied MW power of 595.4 W (75%); see column 3 of Table 1b. The irradiance of the UV light emitted by the DQCPP lamp system was monitored and quantified in the wavelength ranges 210-300 and 310-400 nm with maximal emission seen at 250 and 360 nm, respectively. The light irradiance increased with increase in applied microwave power at both 250 and 360 nm (columns 4 and 5, Table 1a). The light irradiance emitted by the electrodeless Hg lamp at these two wavelengths in the water-cooled DQCPP/H2O system also scaled with the applied microwave power (columns 4 and 5 in Table 1b). However, it was constantly lower than for the non-water-cooled DQCPP system by nearly 40% ((7%) at 250 nm and ∼50% ((7%) at 360 nm. In fact, the lowest decrease ratio in light irradiance from the DQCPP/H2O system to the light irradiance of DQCPP was observed for an applied
MW power of 178.9 W (columns 4 and 5 of Table 1b): 28% loss at 250 nm and 41% loss at 360 nm, respectively. The relative light irradiances with respect to the applied MW power (i.e., per watt of applied power) are listed in columns 6 and 7 of Table 1. In both cases, the lowest relative light irradiance was seen for the largest applied power. The tendency for this lower irradiance was especially significant for the water-cooled DQCPP/H2O system. The reasons for this decline are the likely decrease in temperature in the latter system and the concomitant absorption of microwave radiation by the cooling water. If the differences (GIRssee footnote to Table 1) between the incident (InR) and reflected (RfR) microwave radiation were also taken into account, the relevant relative light irradiances at both wavelengths (last two columns of Table 1) tended to be largest at the lowest applied microwave power. However, perusal of these data indicates that the relative light irradiances neither increased nor decreased monotonically with increase in microwave power. Nonetheless, the smallest loss of light irradiances at 250 and 360 nm occurred at 178.9 W in the water-cooled DQCPP/H2O configuration of Figure 1a (28 and 41%, respectively). Accordingly, the most suitable applied microwave power to use in the photodegradation experiments was 178.9 W for the DQCPP/H2O system. Both the InR and RfR microwave radiation scaled linearly with applied microwave power for the naked MW oven (Figure 2a), for the DQCPP setup (Figure 2b), and for the water-cooled DQCPP/H2O configuration (Figure 2c). The rate of increase of InR was slightly smaller for the MW oven alone (5.69 × 10-4 mA W-1; correlation coefficient, r ) 0.992) than for the equipped oven [(6.09-6.13) × 10-4 mA W-1; r ) 0.992]. By comparison, the rate of increase of RfR was greater for the MW oven alone (4.71 × 10-4 mA W-1; r ) 0.994). The rates of absorption of microwave radiation, described as percent radiation per watt of applied MW power [)100 - (100(Rr/Ir)], for the three configurations of Figure 1a are illustrated in Figure 2d. The slopes (rates) are rather small for the MW device, the DQCPP system, and the water-cooled DQCPP/H2O setup (-0.011, -0.014, and 0.003% W-1), respectively; the corresponding correlation coefficients are 0.35, 0.46, and 0.12, indicating a rather large noise in the data. The UV/visible spectral patterns of the DQCPP light source (Hg/Ar lamp) for applied microwave powers in the range 87.8-621.0 W are depicted in Figure 3 for the wavelength range 300-600 nm. Also illustrated is the output pattern of the more traditional Hg light source, which displayed mercury lines from 313 to 577 nm (top of Figure 3). Relevant mercury lines for the electrodeless lamp appeared at wavelengths 365, 404, 435, 546, and 579 nm. The light irradiances below 365 nm for the DQCPP lamp were more than 10 times smaller. Accordingly, they played only a minor (if any) role in the photochemical process and are not reported. To the extent that the band gap of anatase TiO2 is 3.2 eV (corresponds to λ ) 387.4 nm), the mercury line at 365 nm was available to irradiate the TiO2 particulates and to drive the titania-assisted photodegradation process. Mechanistic Considerations in the Photodegradation of the RhB+ Dye. The rhodamine-B dye significantly absorbs visible wavelengths, so that in the present context, the degradation of RhB+ was likely to proceed following two competitive pathways (A and Bssee below). (A) TiO2 Photodegradation Mechanism under UV Light Illumination (λ < 387.4 nm). Titania particles absorb UV light at energies greater than the band gap energy of 3.2 eV
(λ ) 387.4 nm) to generate electron/hole pairs (eq a1).
TiO2 + hv(λ 387.4 nm). Under these illumination conditions, the TiO2 semiconductor is not photoexcited. Zhao and co-workers (16-18, 27) reported that the chemisorbed RhB+ is excited at wavelengths longer than 470 nm to produce singlet and triplet excited states (eq b1; RhB+ads*). Subse-
RhB+ads + hv f RhB+ads*
(b1)
RhB+ads* + TiO2 f RhB•2+ads + TiO2 (e-CB)
(b2)
TiO2(e-CB) + O2 f O2•-
(b3)
O2•- + H+ f •OOH
(b4)
OOH + O2•- + H+ f O2 + H2O2
(b5)
H2O2 + TiO2(e-) f OH- + •OH + TiO2
(b6)
•
RhB•2+ads + O2 (and/or O2•-, and/or •OH) f intermediates f f f mineralized products (b7) quently, the excited RhB+* species injects an electron into the conduction band of the semiconductor TiO2 followed by the conversion of RhB+* to the radical RhB•2+ads dication. The electron can then be involved in reactions similar to eqs a3-a6 above (namely, eqs b3-b6) to generate the •OH radicals. Consequently, the degradation of RhB+ can also occur using the DQCPP visible light source wavelengths. However, with visible illumination only, deethylation of RhB+ is the primary process, which terminates as the deethylated rhodamine species no longer absorbs visible light (16). The importance of the a1-a6 pathway should be lessened somewhat because the RhB+ substrate must adsorb onto the TiO2 photocatalyst particles for it to photodegrade efficiently by pathway A above. At the prevalent conditions, both the VOL. 36, NO. 23, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5233
FIGURE 4. (a) Changes in mercury line intensities at 365, 404, 435, 546, and 579 nm from radiation emitted by the DQCPP light source with increasing applied microwave power; (b) light intensities normalized to the intensity at 365 nm for all five wavelengths. photocatalytic pathway (eqs a1-a6) and the visible pathway (eqs b1-b7) must overcome the Coulombic interactions between the cationic RhB+ substrate and the positively charged TiO2 surface at pHs below the point of zero charge of TiO2 (pzc ) 6.3). This adsorption is crucial (16-18, 27) in photocatalytic processes involving titania particles, since the oxidizing •OH radicals are poised to react on the particle surface. (Note that the pH of the dispersion was hardly affected from the initial pH of 5.51 during the degradation of RhB+ with the configuration setup of Figure 1a; pH was 5.62 after 60 min of irradiation.) Test of the MW/DQCPP and MW/Hg Lamp Systems in the Photodegradation of RhB+. The variations in the intensity of the mercury lines of Figure 3 with increase in microwave power at the five wavelengths (viz., 365, 404, 435, 546, and 579 nm) are illustrated in Figure 4a. Though the dependence on MW power was sublinear, the correlation coefficients were rather reasonable (r ) 0.993-0.995). The line intensities at the five wavelengths were normalized to the intensities at 365 nm with variations in applied microwave power (Figure 4b). The results indicate that the rate of change of the intensities diminishes with increase in microwave power relative to the data at 365 nm. The drop in rate was greatest at 435 nm, less at 546 and 404 nm, and least at 579 nm. Consequently, we deduce that the TiO2photocatalyzed degradation of rhodamine-B via pathway a1a6 is relatively most effective when the DQCPP lamp is used at the highest applied microwave power compared to lower MW powers. The temporal changes of the UV spectral features in the degradation of RhB+ in aqueous TiO2 dispersions were assessed using the DQCPP reactor exposed to MW radiation (TiO2/DQCPP/MW system). The changes are compared to those observed when the quartz pipe photoreactor was illuminated with a Hg lamp (TiO2/Hg lamp) and when this photoreactor was irradiated concomitantly with the Hg lamp and MW radiation (TiO2/Hg lamp/MW system)ssee Figure 5. The power of the Hg lamp was 250 W. The total power 5234
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 23, 2002
FIGURE 5. (a) Temporal variations in the UV/visible spectral features during the degradation of the RhB+ solution (0.05 mM in aqueous media) in aqueous TiO2 dispersions using the TiO2/DQCPP/MW system after irradiation for 15, 30, and 60 min. (b) Decrease of the UV/visible spectral pattern in the degradation of RhB+ using the TiO2/Hg lamp and TiO2/Hg lamp/MW systems under irradiation for 60 min. used for the TiO2/Hg lamp/MW system was 428.2 W (250 + 178.2 W). The difference (GIR) in intensity between the incident and reflected MW radiation in the TiO2/Hg lamp/ MW system was 0.04 mA; the corresponding difference for the TiO2/DQCPP/MW system was 0.05 mA (see column 2 of Table 1b). In other words, the rate of absorption of MW radiation by water and the quartz pipe photoreactor was greater than in the water-cooled DQCPP/H2O system. The temperature of the solution in the TiO2/DQCPP/MW, TiO2/ Hg lamp/MW, and TiO2/Hg lamp systems was maintained at 20-22 °C, using the cooling water circulator located outside the MW oven (see Figure 1). Approximately 51% of the RhB+ solution was degraded after 15 min of irradiation in the TiO2/DQCPP/MW system. After 30 and 60 min, the corresponding loss of RhB+ was 82 and 87%, respectively. As expected, losses of RhB+ in the TiO2/ Hg lamp and TiO2/Hg lamp/MW systems were significantly lower at 32 and 29%, respectively, after 60 min of irradiation. This was caused by less of the dispersion being irradiated when the setup configuration of Figure 1b was utilized. Variations in the fading of the RhB+ dye between the TiO2/ Hg lamp and the TiO2/Hg lamp/MW system were insignificant. Clearly, this configuration of having an external irradiation source to supply the UV/visible wavelengths is not useful for practical considerations. The spectral loss of RhB+ was confirmed by the TOC measurements. The initial 18.6 ppm TOC of the RhB+ solution decreased to 11.5 (38% loss of the RhB+), to 5.75 (69% loss), and to 3.68 ppm (80% loss) when the RhB+ dye was photodegraded with the TiO2/DQCPP/MW system for 15, 30, and 60 min, respectively (see Figure 6). With the lesser effective TiO2/Hg lamp and TiO2/Hg lamp/MW systems, TOC was 15.2 (18% loss) and 15.5 ppm (17% loss) after 30-min irradiation, respectively. After 60 min of irradiation the TOC remaining in the RhB+ solution was 12.6 (32% loss) and 13.5 ppm (28% loss), respectively. The average loss of the RhB+ solution with
FIGURE 6. Temporal decrease of total organic carbon (TOC) during the decomposition of RhB+ solution (initial TOC concentration, 18.6 ppm) at the conditions indicated (see also text). the TiO2/DQCPP/MW system was 87% after 60 min (Figure 5a), whereas for the TiO2/Hg lamp and TiO2/Hg lamp/MW systems, the percent loss was 32 and 29% (Figure 5b). It is significant to note that, after 30 min of irradiation of the aqueous RhB+/TiO2 dispersion dye with the DQCPP reactor, all the spectral features at wavelengths longer than 250 nm had totally disappeared (Figure 5a), indicating nearly complete conversion of RhB+ to the various intermediates and products (16). In a typical photocatalytic reaction, the disappearance of UV absorption features by an aromatic substrate points to cleavage of the aromatic ring and formation of photooxidized intermediates, followed by the evolution of CO2 gas. Accordingly, the decrease of TOC refers to evolution of CO2 gas from the reaction system. From the congruence between the loss of TOC (Figure 6) and loss of UV absorption (Figure 5a), the degradation of RhB+ that occurred when the DQCPP technique was used infers a pathway otherwise identical to the one suggested for the photocatalytic method (eqs a1a6). Note that deethylation of the RhB+ species via the visible pathway B above (eqs b1-b7) would occur by a stepwise monodeethylation process and would be exhibited by blueshifts of the UV absorption features at 450-600 nm (16). No such shifts were seen in the current study (Figure 5a). We recognize that the TiO2/DQCPP/MW system and the TiO2/Hg lamp/MW system are different only from the positioning of the light source (Figure 1) and from the nature of the two lamps used. The smaller light irradiance emitted by the DQCPP lamp should negatively affect the photodegradiative process relative to the use of the traditional Hg lamp. For the latter, the light irradiance emitted in the wavelength range 220-300 nm (emission detected at 250 nm) was 1.298 mW cm-2 and was 1.969 mW cm-2 in the range 310-400 nm (emission detected at 360 nm). They were measured at 55 cm from the lamp, identical to the distance of the quartz pipe photoreactor from the lamp (see Figure 1b). However, in the case of the TiO2/DQCPP/MW system, the RhB+ solution was circulated through the center of the DQCPP lamp. The relative irradiance emitted by the Hg lamp was 5.192 × 10-3 mW cm-2 at 250 nm and 7.876 × 10-3 mW cm-2 at 365 nm (per watt of input power, 250 W). By comparison, the relative light irradiance emitted by the DQCPP lamp was 0.0727 × 10-3 and 0.0839 × 10-3 mW cm-2 (per watt of input MW power, 178.9 W; see Table 1b) at these same wavelengths. Effect of MW Radiation and Temperature in the Degradation of RhB+. Although the irradiances of the Hg lamp were greater than those of the DQCPP light source, it is important to recall that only a small portion of the quartz pipe reactor was irradiated by using the fiber optic to deliver
FIGURE 7. Experimental setup having both a UV light source (Hg lamp) and a MW generator to examine the direct effect(s) of UV and MW radiations on the decomposition of the cationic RhB+ dye. the UV/visible wavelengths and the system configuration of Figure 1b. Experiments were nonetheless carried out to examine the consequences of the absorption of MW radiation and the ensuing temperature variations on the photodegradation of the rhodamine-B dye with the modified system configuration illustrated in Figure 7. With the setup employed earlier (19), microwave radiation had no effect on a pristine aqueous solution of the RhB+ dye and the heat released by absorption of the microwave radiation had little (if any) effect on the TiO2-photocatalyzed degradation of RhB+. The microwave generator and the peripherals used for this purpose were identical to those used in the abovementioned method. Note that in this case the whole Pyrex reactor was irradiated by the radiation emitted by the Hg lamp (Toshiba 75-W mercury lamp). The input MW power was 300 W, and the light irradiance emitted by the source in the wavelength range 310-400 nm (emission detected at λ ) 360 nm) was 0.4 mW cm-2. A 30mL aqueous RhB+ solution (0.05 mM; 150 mL) in the presence of TiO2 particles (loading, 300 mg) was contained in a 250mL Pyrex cylindrical reactor (i.d. ) 45 × 290 mm; Taiatsu Techno Co.; maximum pressure, 1 MPa; temperature, 150 °C). The dispersion was irradiated simultaneously using both microwave radiation and UV/visible radiation under magnetic agitation. A peristaltic pump circulated the aqueous RhB+/TiO2 dispersion at a flow rate maintained at 2000 mL min-1. The temperature increased to 65 ( 2 °C on absorption of MW radiation by the aqueous medium. When the dispersion was cooled by the water-cooling device located outside the MW device, denoted TiO2/Hg lamp/MWcool, the temperature stayed at ∼26 °C. The influence of the MW radiation on the degradation of RhB+ was examined using the TiO2/Hg lamp and the TiO2/ Hg lamp/MW systems, whereas the effect of temperature was determined using the TiO2/Hg lamp/MW and TiO2/Hg lamp/MWcool systems. In all cases, the effects were assessed by monitoring the temporal behavior of the UV/visible spectral features as the degradation of RhB+ progressed (see Figure 8). After 2 h of irradiation, a greater quantity of RhB+ degraded on using the TiO2/Hg lamp/MWcool ( ∼45% more) and TiO2/ Hg lamp/MW ( ∼60% more) systems compared to the TiO2/ Hg lamp ( ∼10%) system. Also, there appeared to be a MWVOL. 36, NO. 23, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5235
FIGURE 8. Spectral variations in the absorption patterns following the degradation of RhB+ in aqueous TiO2 dispersions irradiated for 2 and 5 h using the TiO2/Hg lamp system at 26 °C. The temperature of the TiO2/Hg lamp/MW system was 65 ( 2 °C, and for the watercooled TiO2/Hg lamp/MWcool system, it was 26 °C. The light irradiance was 0.4 mW cm-2; the microwave power output was ∼300 W. induced heat effect, however small (∼30%), in degrading the RhB+ substrate as evidenced by comparing the TiO2/Hg lamp/ MW to the TiO2/Hg lamp/MWcool system. At longer irradiation times (5 h; Figure 8b) the decrease of the UV/visible spectral features of RhB+ with microwave radiation coupled to the TiO2/Hg lamp systems was even more significant. The corresponding degradations of RhB+ were, respectively, about 70, 85, and 40%, whereas the heat effect was ∼50%. Note that these results cannot be compared to those of Figure 5b where coupling the microwave radiation to the TiO2/Hg lamp system showed no effects in degrading RhB+, inasmuch as the experimental conditions were entirely different. For example, the system configuration was different (Figure 7 versus Figure
1b), a larger portion of the reactor was irradiated in the configuration of Figure 7 than in Figure 1b, and the temperature control was different. Temperature variations between the naked DQCPP reactor and the water-cooled DQCPP/TiO2/H2O system were examined using the radiation thermograph positioned as in Figure 1a. The initial temperature was 25 °C. Position A in the photographs of Figure 9 refers to the surface (side view) of the DQCPP reactor, whereas position B is the center (top view) of the DQCPP system. The photographs were taken after 5 (a-i, b-i), 15 (a-ii, b-ii), and 60 min (a-iii, b-iii) for the naked DQCPP reactor (Figure 9a) and for the water-cooled DQCPP/TiO2/H2O system (Figure 9b). The temperature recorded graphically at position B reflects the temperature at the center of the aqueous TiO2 dispersion. The surface temperatures of the naked DQCPP system were 68 °C at 5 min, 84 °C at 15 min, and 92 °C at 60 min of MW irradiation. The temperatures at the center were slightly smaller by only a few degrees (see Figure 9) with the difference in temperature between the surface and the center of the reactor becoming smaller at longer MW irradiation times: +5 °C at 5 min, +3 °C at 15 min, and less than 1 °C at 60 min. Evidently, microwave radiation is absorbed to a greater extent at the surface of the DQCPP reactor. The increase in temperature averages out across the dispersion at the longer times as the dispersion is circulated through the DQCPP system. In the case of the water-cooled DQCPP/TiO2/H2O system, the increases in temperature at the surface of the quartz reactor were more modest: 33 °C at 5 min, 46 °C at 15 min, and 59 °C at 60 min of microwave irradiation. However, the differences in temperature between the surface and the center of the reactor were more pronounced: -7, -18, and -9 °C, respectively. Contrary to the naked DQCPP reactor, the temperatures at the surface (position A) are lower than in the center (position B). These differences are caused by the absorption of microwave radiation by the water coolant in
FIGURE 9. Thermographic images indicating the effective temperatures in (a) a naked DQCPP device and (b) a DQCPP system containing the aqueous TiO2 dispersion. Both were microwave-irradiated for 5, 15, and 60 min. 5236
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 23, 2002
the clearance between the DQCPP lamp and the quartz pipe reactor. The temperature of the coolant determined with a thermometer through an optical fiber was 54 °C at 5 min, 79 °C at 15 min, and 86 °C after 60 min of MW irradiation. The water-cooling circulator device (Figure 1) maintained the average temperature of the dispersion at ∼25 °C. The temperature of the solution at the surface of DQCPP was greater owing to the incident mercury plasma illumination. For the TiO2/Hg lamp/MW experimental setup (Figure 1b), the temperature of the aqueous dispersion also increased on absorption of MW radiation, but not to the extent it was in the TiO2/DQCPP/MW setup. In summary, usage of the TiO2/DQCPP/MW setup of Figure 1a led to an enhancement of the rate of decomposition of RhB+ in the aqueous TiO2 dispersions. We deduce that the faster dynamics were caused by a greater absorption of incident UV and visible radiation on the basis of the following considerations. The entire 19-cm length of the cylindrical DQCPP quartz reactor was illuminated at all possible angles (360°) in the setup of Figure 1a. By contrast, the radiation from the external Hg lamp reached the dispersion through the light guide (fiber optic), which focused the radiation only on a small portion of the dispersion. Accordingly, the quantity of dispersion that was irradiated with the setup of Figure 1b was much smaller than the quantity that was possible with the DQCPP light source. As a result, the extent of degradation of RhB+ was greater when irradiated with the DQCPP light source, even though this source emitted significantly less radiation. Measurements of quantum efficiencies were precluded by the nature of our devices. Nonetheless, the DQCPP reactor does provide for a greater degradation efficacy under the conditions and configuration setup used (28).
Acknowledgments We are grateful to the Frontier Research Promotion Foundation of Japan (H.H.) and to the Natural Sciences and Engineering Research Council of Canada (N.S.) for financial support of our work. We also express our sincere gratitude to the personnel of the Shikoku Instrument Co. Ltd. for the loan of and technical assistance with the use of the microwave device.
Literature Cited (1) Pelizzetti, E.; Minero, C.; Maurino, V. Adv. Colloid Interface Sci. 1990, 32, 271. (2) Ollis, D. F., Al-Ekabi, H., Eds. Photocatalytic Purification and Treatment of Water and Air; Elsevier: Amsterdam, 1993. (3) Fox, M. A.; Dulay, M. T. Chem. Rev. 1993, 93, 341. (4) Serpone, N., Pelizzetti, E., Eds. Photocatalysis: Fundamentals and Applications; Wiley-Interscience, New York, 1989. (5) Bahnemann, D. W. In Photochemical Conversion and Storage of Solar Energy; Pelizzetti, E., Schiavello, M., Eds.; Kluwer: Dordrecht, 1991; p 251. (6) Matthews, R. W. In Photochemical Conversion and Storage of Solar Energy; Pelizzetti, E., Schiavello, M., Eds.; Kluwer: Dordrecht, 1991; p 427. (7) Ollis, D. In Photochemical Conversion and Storage of Solar Energy; Pelizzetti, E., Schiavello, M., Eds.; Kluwer: Dordrecht, 1991; p 593.
(8) Zaleska, A.; Hupka, J.; Wiergowski, M.; Biziuk, M. J. Photochem. Photobiol. A: Chem. 2000, 135, 213. (9) Hidaka, H.; Jou, H.; Nohara, K.; Zhao, J. Chemosphere 1992, 25, 1589. (10) Hidaka, H.; Suzuki, Y.; Nohara, K.; Horikoshi, S.; Pelizzetti, E.; Serpone, N. J. Polym. Sci., Part A: Polym. Chem. 1996, 34, 1311. (11) Horikoshi, S.; Hidaka, H.; Hisamatsu, Y.; Serpone, N. Environ. Sci. Technol. 1998, 32, 4010. (12) Cho, S.; Choi, W. J. Photochem. Photobiol. A: Chem. 2001, 143, 221. (13) Hidaka, H.; Yamada, S.; Suenaga, S.; Zhao, J. J. Mol. Catal. 1990, 59, 279. (14) Pelizzetti, E.; Maurino, V.; Minero, C.; Carlin, V.; Pramauro, E.; Zerbinati, O.; Tosato, M. L. Environ. Sci. Technol. 1990, 24, 1559. (15) Horikoshi, S.; Hidaka, H.; Serpone, N. Chemosphere, to be submitted. (16) Liu, G.; Li, X.; Zhao, J.; Hidaka, H.; Serpone, N. Environ. Sci. Technol. 2000, 34, 3982. (17) Qu, P.; Zhao, J.; Zang, L.; Shen, T.; Hidaka, H. J. Photochem. Photobiol. A: Chem. 1998, 138, 39. (18) Zhao, J.; Wu, T.; Wu, K.; Oikawa, K.; Hidaka, H.; Serpone, N. Environ. Sci. Technol. 1998, 32, 2394. (19) Horikoshi, S.; Serpone, N.; Hidaka, H. Environ. Sci. Technol. 2002, 36, 1357. (20) Boulos, M. J. Therm. Spray Technol. 1992, 1, 33. (21) M. I. Boulos: “The inductively coupled radio frequency plasma”. High Temp. Mater. Process. 1997, 1, 17. (22) Chemat, S.; Aouabed, A.; Bartels, P. V.; Esveld, D. C.; Chemat, F. J. Microwave Power Electromagn. Energy 1999, 34, 55. (23) Daito, S.; Tochikubo, F.; Watanabe, T. Jpn. J. Appl. Phys. Part 1 2001, 40, 2475. (24) Willberg, D. M.; Lang, P. S.; Hoechemer, R. H.; Kratel, A.; Hoffmann, M. R. Environ. Sci. Technol. 1996, 30, 2526. (25) Horikoshi, S.; Hidaka, H.; Serpone, N. J. Photochem. Photobiol. A: Chem., in press (2002). (26) Radiation Chemistry Data Center, Notre Dame Radiation Laboratory, University of Notre Dame, Notre Dame, IN; website, http://allen.rad.nd.edu/RCDC/RCDC.html. (27) (a) Liu, G.; Li, X.; Zhao, J.; Horikoshi, S.; Hidaka, H. J. Mol. Catal., A 2000, 153, 221. (b) Wu, T.; Lin, T.; Zhao, J.; Hidaka, H.; Serpone, N. Environ. Sci. Technol. 1999, 33, 1379. (c) Wu, T.; Liu, G.; Zhao, J.; Hidaka, H.; Serpone, N. J. Phys. Chem. B 1999, 103, 4862. (d) Qu, P.; Zhao, J.; Shen, T.; Hidaka, H. J. Mol. Catal., A 1998, 129, 257. (e) Wu, T.; Liu, G.; Zhao, J.; Hidaka, H.; Serpone, N. J. Phys. Chem. B 1998, 102, 5845. (28) At this stage of our work, it is possible to speculate on what the effect(s) of microwave radiation might be on the TiO2-photocatalyzed degradation of pollutants. We suggest that there are thermal and nonthermal effects. In the present instance, we have seen some heat effects in the degradation of the RhB+ dye in the configuration of Figure 7. The nonthermal effects are those that the microwave radiation causes on the TiO2 photocatalyst proper. It is attractive to suggest that the MW radiation affects the surface of the TiO2 particles in such a way as to render the surface more photocatalytically active: (a) by inducing active defect sites that enhance formation of a greater quantity of the highly oxidizing •OH radicals as attested to by the ESR-DMPO technique (19) and (b) by suppressing electron/ hole recombination.
Received for review January 7, 2002. Revised manuscript received May 10, 2002. Accepted August 23, 2002. ES020506C
VOL. 36, NO. 23, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5237
