Lifestyle

A new kinetic model for titanium dioxide mediated heterogeneous photocatalytic degradation of trichloroethylene in gas-phase

Description
A new kinetic model for titanium dioxide mediated heterogeneous photocatalytic degradation of trichloroethylene in gas-phase
Categories
Published
of 14
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Transcript
  A new kinetic model for titanium dioxide mediatedheterogeneous photocatalytic degradation of trichloroethylene in gas-phase Kristof Demeestere a , Alex De Visscher b , Jo Dewulf  a , Maarten Van Leeuwen b ,Herman Van Langenhove a, * a  Department of Organic Chemistry, Faculty of Agricultural and Applied Biological Sciences, Research Group of Environmental Organic Chemistry and Technology (EnVOC), Ghent University,Coupure Links 653, B-9000 Ghent, Belgium b  Department of Applied Analytical and Physical Chemistry, Faculty of Agricultural and Applied Biological Sciences,Ghent University, Coupure Links 653, B-9000 Ghent, Belgium Received 24 March 2004; received in revised form 25 June 2004; accepted 28 June 2004Available online 21 August 2004 Abstract This paper focuses on the kinetics of photocatalytic removal and carbon mineralization of gaseous trichloroethylene (TCE) on near-UVirradiated TiO 2  DegussaP25. Experiments were carried out in a flat-plate photoreactor at TCE inlet concentrations of 100–500 ppmv, relativehumidities (RH) of 0–62% and gas residence times of 2.5–60.3 s. Gas residence time distribution (RTD) curves revealed an axial dispersedplug flow in the photoreactor with Peclet numbers above 59.4. For all experimental conditions, the carbon mineralization efficiency (5.1–73.0%) was lower than the removal efficiency (8.6–99.9%) and dichloroacetylchloride (DCAC) was detected as a gas-phase degradationproduct. TCE removal efficiencies increased with lower TCE inlet concentrations, lower RH and higher gas residence times. Evaluatingdifferentkineticmodelsbyleastsquaresanalysis,itwasshownthattheLangmuir–Hinshelwood(LH)modelcouldnotgiveanadequatefittingto the experimental results. A new kinetic model, explicitly taking into account electron–hole pair reactions, was developed based on linearTCE adsorption–desorption equilibrium and first order reaction kinetics. The new kinetic model described the experimental results in a moreaccurate way, as exemplified by a more randomly distributed set of residuals and by a reduction of the sum of squares (SSQ) by a factor 1.7–8.5. The effect of TCE gas-phase concentration, RH and light intensity on adsorption–desorption kinetics, electron–hole concentrations andchemical conversion rates is discussed. # 2004 Elsevier B.V. All rights reserved. Keywords:  Photocatalysis; Titanium dioxide; Degussa P25; Volatile organic compounds; VOCs; Trichloroethylene; Gas-phase; Relative humidity; Kinetics;Modelling 1. Introduction Heterogeneous photocatalysis has become an intensivelyinvestigated technology for the purification, decontamina-tion and deodorization of waste gas streams and indoor air[1]. Particularly, photocatalytic degradation of volatileorganic compounds (VOCs) is gaining more and moreinterest, since the emission of these compounds contributeto environmental problems, such as tropospheric ozoneformation, stratospheric ozone layer depletion and globalwarming [2]. Moreover, many VOCs are toxic and/or car-cinogenic and can cause serious odour nuisance [3].Photocatalytic oxidation of chlorinated organic com-pounds, such as trichloroethylene (TCE) deserves specialattention because of their toxicity and biodegradationresistance. TCE has been widely used in industry and iscommonly found in emissions from chemical processing, www.elsevier.com/locate/apcatbApplied Catalysis B: Environmental 54 (2004) 261–274* Corresponding author. Tel.: +32 9 264 59 53; fax: +32 9 264 62 43. E-mailaddress: herman.vanlangenhove@ugent.be(H.VanLangenhove).0926-3373/$ – see front matter # 2004 Elsevier B.V. All rights reserved.doi:10.1016/j.apcatb.2004.06.020  landfills, hazardous waste disposal, groundwater remedia-tion sites, and even in indoor air [4]. Various treatmenttechnologies such as absorption, adsorption, condensation,biofiltration and incineration are under investigation and/oroperation to remove TCE from waste gases. However, allthese traditional technologies show several drawbacks andlimitations: (1) absorption and adsorption methods do notreally degrade pollutants but only transfer them into asecondary liquid and solid waste stream, respectively;(2) the efficiency of condensation and biofiltration islimited for volatile and hardly biodegradable compoundslike TCE; and (3) incineration of Cl-containing compoundscan result in hazardous by-products formation, e.g. dioxins[5]. Therefore, the development of advanced oxidationprocesses like heterogeneous photocatalysis is requiredto bring forward new and efficient end-of-pipe technolo-gies.In the last decade, the mechanism of heterogeneousphotocatalysis has been investigated by many researchers.According to Fu et al. [6], a photocatalytic reaction proceedson the surface of semiconductors via several steps: (1)production of electron–hole pairs by irradiating the semi-conductor with light having an energy content higher thanthe band gap energy of the semiconductor; (2) separation of the photogenerated electrons and holes due to trapping byspecies that are adsorbed on the semiconductor; (3) redoxreactions between the trapped electrons and holes and theadsorbates; (4) desorption of reaction products and recon-struction of the surface. Of all different semiconductorphotocatalysts tested, TiO 2  is the most commonly used,with TiO 2  Degussa P25 having become an efficient researchstandard [7].Research reports on photocatalytic degradation of TCEmainly focus on: (1) preparation methods of catalysts; (2)identification of intermediates; and (3) kinetic discussions.Partially because of the large variety of photoreactor con-cepts, lamps, experimental conditions and analytical meth-ods used, there are some ambiguous points in literature [8].For example, many authors report that photocatalytic degra-dation of TCE proceeds through the formation of highlyreactive radicals (e.g.   OH,   Cl or oxygen radicals) [9–13].However, direct oxidation of TCE by valence band holes hasalso been suggested [14]. Although there is still some debateabout whether the radicals and/or pollutants react on thesurface or in the bulk fluidum phase, the most commonlysuggested mechanism is the reaction between adsorbedspecies [15,16]. Since recombination of photogeneratedelectron–holepairsoccurswithinafraction ofananosecond,interfacial carrier trapping is only kinetically competitive if trapping species are already adsorbed prior to electron–holepair generation by a photon [2,17]. By consequence, kineticmodels describing photocatalytic degradation of pollutantsshould consider both adsorption–desorption and chemicalconversion kinetics. In this context, the Langmuir–Hinshel-wood (LH) model is the most frequently used kinetic modelto formulate rate equations in heterogeneous photocatalyticreactions [2,14,18–28]. Some authors extended the widelyused LH model for single compound degradation by con-sidering different adsorption site types [21–25] or compe-titive adsorption on one adsorption site type [26–28]. Forexample,Wangetal.[28]evaluatedtheapplicabilityofthreedifferent LH models for photocatalytic degradation of TCEin gas streams and concluded that a bimolecular LH model,considering competitive adsorption between TCE and watermolecules on one adsorption site type, resulted in the mostadequate fitting to their experimental results. Recently,Kutsuna et al. [29] investigated both dark adsorption andphotodegradation of three methyl perfluoroalkyl ethers onTiO 2  at concentrations between 0.001 and 0.016 mol m  3 .They concluded that photodegradation kinetics could not berepresented adequately by a LH model. In contrast, it wasfound that gas–solid adsorption–desorption equilibriumcould be described by the Henry type relation and thatthe initial photodegradation rate was proportional to thesquare root of the adsorbed substrate concentration in thedark. This shows that, in spite of numerous investigationson gas–solid heterogeneous photocatalysis, adsorption–desorption equilibrium and reaction kinetics are not com-pletely understood so far.The scope of this work was to gain new insights intophotocatalytic degradation kinetics of gaseous TCE on near-UV irradiated TiO 2  Degussa P25. A new kinetic model isdeveloped based on recently reported linear adsorption–desorption equilibrium data [30] and first order reactionkinetics. The adequacy of LH and the new kinetic model tofit the experimental results is compared and the effect of TCE gas-phase concentration, RH and light intensity onmodel parameters and TCE removal and carbon mineraliza-tion rates is discussed. 2. Experimental 2.1. Materials Trichloroethylene (TCE,  > 99.5%, A.C.S. reagent) anddichloroacetylchloride (DCAC, 97%) were purchased fromAldrich and Acros, respectively, and were used withoutfurther purification. Clean and dry air ([H 2 O]  <  3.0 ppmv;[CO 2 ]  <  1.0 ppmv; C  x H  y  <  0.5 ppmv) and dry air ([H 2 O]  < 3.0 ppmv) containing a calibrated amount of 350 ppmv CO 2 were provided by L’Air Liquide.The catalyst powder titanium dioxide (P25; 80% anatase,20% rutile) was obtained from Degussa–Hu¨ls Benelux andhad a purity of at least 99.5%. Stated impurities includeAl 2 O 3  (  0.3%), HCl (  0.3%), SiO 2  (  0.2%) and Fe 2 O 3 (  0.01%). The particles were reported to be non-porous andspherical, with an apparent density of 130 kg m  3 and aspecific surface area of 50  15 m 2 g  1 . Before all degrada-tion experiments, TiO 2  samples were heated at 105  8 C for1 hin order to remove possibly sorbed chemicals. No furthermodifications of the catalyst powder were done. K. Demeestere et al./Applied Catalysis B: Environmental 54 (2004) 261–274 262  2.2. Photoreactor and photocatalytic degradationconditions A scheme of the experimental set-up is shown in Fig. 1.The main part consists of the rectangular Plexiglas flat-platephotocatalytic reactor (length (  L  ): 20 cm, width ( W  ): 10 cm,height (  H  ): 3 cm), in which 13 Plexiglas baffles (  L   W    H  :9  0.2  3 cm) and four small Plexiglas structures (  L   W    H  : 3.5 cm    1.5–2.5 cm    3 cm) were provided tominimize the presence of dead volumes in the photoreactorand to increase the degree of turbulence. The effective gasvolume (V) and bottom area (A) of the photoreactor were0.422 L and 140.6 cm 2 , respectively. Near-UV irradiationwas provided by a blacklight blue 18W UV lamp (340 nm  < l  <  410 nm; emission maximum at 365 nm, Philips Light-ing), positioned just above the 2.2 mm thick borofloat glass-plate closing the reactor and covered by a cylindricalreflector. By means of potassium ferrioxalate actinometry[31–33], light intensity at the catalyst surface was deter-mined to be 2.34 mW cm  2 . TiO 2  powder (3.5 g) was dis-tributed uniformly over the bottom area of the reactor. Thetemperature in the reactor was controlled at 25.0    0.1  8 Cby immersing the bottom of the reactor (2.5 cm) in athermostatic water bath.The inlet gas stream was prepared by mixing two dif-ferent air streams, each of them provided from a compressedgas cylinder. The main flow (0.42–10.00 L min  1 ) consistedof clean and dry air and could be humidified by bubbling itthrough two impingers filled with deionized water andplaced in series in a thermostatic water bath at the desiredtemperature. The relative humidity (RH) of the air streamwas measured by a RH-sensor Testo 452 (Testo NV). Asecond gas stream (0.45–30.00 mL min  1 ) was enrichedwith TCE by bubbling it through another impinger filledwith pureliquid TCEandplaced inathermostatic waterbathat 25.0  0.1  8 C. The flow rate of both streams was adjustedby mass flow controllers (MFC Brooks, 5851S and 5850S).Prior to all photocatalytic degradation experiments, thephotoreactor was flushed with clean dry or humidified air forat least 12 h, to obtain the desired RH throughout the wholereactor. Subsequently, TCE contamined air stream waspassed through the reactor in the absence of UV irradiationuntil gas–solid adsorption–desorption equilibrium wasestablished (typically within 1 h), i.e. until TCE in- andoutlet concentrations were not significantly different fromeach other ( t  -test, significance level  a    0.05). Finally, theUV-lamp was switched on and TCE and CO 2  in- and outletconcentrations were measured for at least 75 min. AverageTCEremoval( h remov )andcarbonmineralizationefficiencies( h miner ) were calculated from at least six in- and six outletTCE and CO 2  concentrations, respectively, measured understeady state conditions, typically reached within 15 minafter switching on the lamp. 2.3. Analytical methods Gas samples(1 mL) were takenatthe in-and outlet ofthereactor using a 1 mL gastight pressure-lock precision ana-lytical syringe (Series A, Alltech Ass.). TCE concentrationswere measured with an Interscience 8000 Top gas chroma-tograph (GC), equipped with a flame ionization detector(FID, 30 ml min  1 H 2 ;  300 ml min  1 air). A 30 m DB-5bounded phase column (J&W Scientific Inc., 95% dimethyl-5%-diphenylpolysiloxane, internal diameter 0.53 mm, filmthickness 1.5 m m) with He as carrier gas (head pressure20 kPa) was used. The GC oven, injector and detector K. Demeestere et al./Applied Catalysis B: Environmental 54 (2004) 261–274  263Fig. 1. Schematic representation of the experimental set-up used for photocatalytic degradation experiments.  temperatures were kept constant at 70, 220 and 250  8 C,respectively. Peak areas were integrated by Chrom-CardThermoquest integration software. Carbon dioxide concen-trations were determined with an Agilent 6890 Series GC,equipped with a split/splitless injector and a thermal con-ductivity detector (TCD, reference flow: 5 ml min  1 He,make-up flow: 5 ml min  1 He). In the GC oven, two col-umns (Hewlett Packard) in series were provided: the firstPLOT Q column (divinylbenzene–polystyrene mixture,length 30 m, internal diameter 0.53 mm, film thickness40 m m) was chosen to retain the VOCs. The first fractioneluting from that column contained N 2  /O 2  and was sent tothe second PLOT molecular sieve 5 A˚ column (length 30 m,internal diameter 0.53 mm, film thickness 25 m m) via a six-way valve. By swithing that valve after 1.4 min, the mole-cular sieve column was excluded from the system and theCO 2  containing fraction eluting from the PLOT Q columnwas sent directly to the TCD detector. After 2.5 min, thevalve was switched again and N 2  /O 2  separation and detec-tion was performed. Injector and detector temperatures wereset at 250  8 C; GC oven temperature was kept at 50  8 C. Thecarrier gas (He) flow rate was 8 mL min  1 . Peak areas wereintegrated by Agilent software.Identification of intermediates was done by a GC-MSmethodology, adapted from Dewulf et al. [34]. Briefly, asampling tube (outer diameter 1/4 in.), filled with 200 mg of TENAX TA and flushed with He (30 mL min  1 ), was loadedwith 150 m L of reactor outlet gas through a heated GCinjector (110  8 C) connected with the sampling tube. After2.5 min, the sampling tube was disconnected from the injec-tor and organic compounds were desorbed in an UnityThermal Desorption system (Markes Int.) at 250  8 C andrefocused on a microtrap filled with TENAX TA, cooledat   10  8 C. After flash-heating of the microtrap, analyteswere injected onto a 60 m CP-SIL 5 CB Low Bleed/MSbounded phase capillary GC column (Chrompack, 100%dimethylpolysiloxane,internaldiameter0.32 mm,filmthick-ness 0.25 m m). He was used as a carrier gas (head pressure40 kPa)andtheGC oven temperaturewas rampedfrom 27 to70  8 C at a heating rate of 2  8 C min  1 . Masses from  m  /   z  35 to300 were recorded on a Trace DSQ Quadrupole GC-MS(Thermo Finnigan) with an electron impact energy of 70 eV. 3. Results and discussion 3.1. Characterization of the gas flow pattern in the photoreactor  To characterize the gas flow pattern in the photoreactor,the gas residence time distribution (RTD) curve was deter-mined at flow rates of 0.42 and 0.84 L min  1 , correspondingto theoretical gas residence times ( t  ) of 60.3 and 30.1 s,respectively. Therefore, a pulse input  m 0  = 4.88 mg of gaseous TCE was loaded to the reactor at time  t   = 0 s.TCE outlet concentrations [TCE] g,out  (mg L  1 ) were mea-sured as a function of time, divided by the pulse inputconcentration [TCE] g,input  =  m 0 V   1 (mg L  1 ) and plottedversus reduced time  u   =  t   /  t  (  ). Fig. 2 shows the RTD curvethus obtained at 0.42 L min  1 and reveals that the flowpattern in the photoreactor can be described adequatelyby an axial dispersed plug flow model [35,36]. E  ¼½ TCE  g ; out ½ TCE  g ; input ¼  ffiffiffiffiffiffiffiffiffi Pe L 4 pu  r   exp   Pe L ð 1  u  Þ 2 4 u  " # s  2 u   ¼ 2 ð Pe L Þ  1 þ 8 ð Pe L Þ  2 with Pe L  ¼  Lu D (1)where  D ,  L  ,  u  and  Pe L  represent the axial dispersion coeffi-cient (m 2 s  1 ), characteristic length (m), flow velocity(m s  1 ) and Peclet number (  ), respectively. From thedispersed plug flow model,  Pe L  numbers of 67.7 and59.4, effective average gas residence times of 29.4 and59.4 s and variances on  u   ( s  2 u  ) of 0.031 and 0.036 werecalculated at  t   values of 30.1 and 60.3 s, respectively. Sincecalculated  Pe L  numbers are rather high ( Pe L  >  40) [36], the degree of axial dispersion is low, indicating a turbulent flowregime in the photoreactor, even at the lowest flow rate. Thisis confirmed by the disagreement between the experimentaland laminar flow RTD curve [35] (Fig. 2). By consequence, diffusion of TCE molecules from gas-phase to the TiO 2 surface isexpected tobe quick, so that gaseous mass transferresistance may be not rate limiting in the photocatalyticdegradation process. 3.2. Photocatalytic degradation kinetics of TCE:experimental data Photocatalytic degradation and carbon mineralization of TCE was investigated at inlet concentrations of 100–500 ppmv (relative standard deviations R.S.D.  <  5%),relative humidities (RH) of 0–62% (R.S.D.  <  7%) andtheoretical gas residence times ( t  ) of 2.5–60.3 s (R.S.D. <  2%). Results are presented in Table 1. Preliminaryexperiments revealed that for all experimental conditions,no significant degradation of TCE was obtained by directphotolysis (only UV, no TiO 2 ) at the used wavelengthspectrum ( t  -test,  a    0.05).For all experimental conditions,  h miner  was lower than h remov , indicating the formation of stable intermediates.DCAC was identified by GC–MS analysis, being in accor-dance with previously reported TCE photocatalytic degra-dation pathways [9–12,37,38].The presence of water vapour in the gas stream reduced h remov  and  h miner , although the effect on  h miner  is onlysignificant at RH  >  24%. This is in agreement with resultsobtained by other authors [39,40]. Hegedu¨s and Dombi [38] recently reported a negligible effect of water vapour on theTCE degradation rate. Considering the photocatalytic reac-tion mechanism postulated by Yamazaki et al. [10,41] inwhich the important role of    OH radicals is emphasized, thenegative effect of water vapour on  h remov  and  h miner  is rather K. Demeestere et al./Applied Catalysis B: Environmental 54 (2004) 261274 264  surprising. It should be noted, however, that other authors[13] investigated the TCE degradation mechanism by O-atom isotope-labeled techniques and concluded that surfaceadsorbed water is not involved in the photocatalytic oxida-tion mechanism of TCE. In a previous work  [30], wedemonstrated that the adsorption–desorption equilibriumcoefficient of TCE on TiO 2  Degussa P25 decreased from0.403 to 0.0158 L g  1 with increasing RH from 0 to 90% ( T  K. Demeestere et al./Applied Catalysis B: Environmental 54 (2004) 261274  265Fig. 2. Experimental gas residence time distribution (RTD) curve at a theoretical average residence time of 60.3 s, compared with RTD curves according to anaxial dispersed plug flow and laminar flow model.Table 1TCE removal ( h remov ) and carbon mineralization ( h miner ) efficiencies (%) obtained at TCE inlet concentrations of 100–500 ppmv, RH = 0–62% and t   = 2.5–60.3 s t   (s)  C   = 100 ppmv  C   = 300 ppmv  C   = 500 ppmv h remov  (%)  h miner  (%)  h remov  (%)  h miner  (%)  h remov  (%)  h miner  (%)RH    0.8%2.5 40.3    3.4 14.5    0.1 n.d. n.d. n.d. n.d.2.7 n.d. a n.d. 21.1    2.0 5.6    0.9 n.d. n.d.5.0 52.1    1.5 19.9    0.5 29.8    2.2 10.9    0.8 23.2    2.8 7.7    0.310.0 72.9    0.7 31.4    2.9 56.1    1.9 18.6    0.8 40.9    4.1 12.4    0.120.1 90.4    0.6 41.8    0.8 78.2    1.9 29.4    1.3 68.2    1.3 20.2    0.930.1 97.0    0.2 47.9    3.3 91.8    0.2 37.6    1.2 80.3    1.4 24.8    1.445.2 99.4    0.1 63.0    1.8 99.1    0.1 45.7    2.3 94.7    0.5 37.9    0.560.3 99.7    0.1 64.9    1.6 99.9    0.0 47.5    3.1 99.5    0.1 40.3    0.9RH = 24.4    1.7%2.5 27.3    2.6 14.6    1.9 n.d. n.d. n.d. n.d.2.7 n.d. n.d. 12.5    0.8 10.6    1.1 n.d. n.d.5.0 43.9    1.7 19.3    1.0 24.7    2.5 13.7    0.4 18.0    1.7 9.1    0.810.0 54.0    2.4 28.4    0.9 46.4    0.8 20.0    0.7 35.1    1.6 16.4    1.220.1 65.0    2.7 34.5    0.2 59.2    2.5 29.2    1.0 49.4    2.0 21.1    1.330.1 70.8    0.4 49.4    0.6 70.1    1.8 34.5    0.7 60.6    2.1 24.8    2.245.2 80.9    1.0 68.7    1.5 86.0    0.4 41.7    0.2 75.5    3.3 36.9    1.560.3 90.0    0.6 73.0    3.2 92.4    0.9 46.0    0.9 82.5    0.6 44.2    2.2RH = 61.7    3.3%2.5 13.5    2.9 11.5    1.7 n.d. n.d. n.d. n.d.2.7 n.d. n.d. 8.6    0.9 5.1    0.4 n.d. n.d.5.0 20.8    2.2 15.8    0.2 16.1    0.6 6.7    0.2 9.9    1.4 6.0    0.310.0 34.5    2.5 20.6    1.1 24.5    2.4 9.0    0.1 18.5    0.7 8.5    0.520.1 44.9    2.6 24.9    1.7 44.9    3.6 13.9    0.2 32.0    2.4 14.5    0.930.1 52.5    2.5 31.9    1.5 51.8    3.9 20.9    0.8 43.8    1.8 17.2    0.345.2 60.4    2.4 37.5    1.1 62.8    2.3 23.7    0.1 56.6    3.0 19.2    0.260.3 71.0    1.8 46.3    0.2 72.3    1.7 26.4    0.8 66.9    1.3 24.3    0.5 a n.d.: not determined.
Search
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks