Sie sind auf Seite 1von 11

Chemical Engineering Science 58 (2003) 257 267

www.elsevier.com/locate/ces

Liquid phase catalytic hydrodechlorination of 2,4-dichlorophenol over carbon supported palladium: an evaluation of transport limitations
Guang Yuan, Mark A. Keane
Department of Chemical and Materials Engineering, University of Kentucky, Lexington, KY 40506, USA Received 16 July 2002; received in revised form 5 September 2002; accepted 12 September 2002

Abstract The catalytic hydrodechlorination (HDC) of aqueous 2,4-dichlorophenol (2,4-DCP) solutions over Pd/C catalysts (110% w/w Pd) has been investigated at 303 K in a stirred slurry reactor. The experimental results have shown that 2,4-DCP is converted to phenol quantitatively and 2-chlorophenol (2-CP) is the only intermediate product within detect limitations ( 0:2 mM). The system is 100% selective in terms of dechlorination and phenol hydrogenation only proceeds once complete dechlorination has been attained. The reaction pathway is illustrated and HDC progress is related to pH changes in solution. The mass-transfer limitations have been evaluated experimentally using the diagnostic criteria associated with varying hydrogen ow rate, stirring speed, catalyst concentration and particle size. Experimental results combined with parameter estimation have revealed the in uence of mass transfer at the liquid/solid interface and intraparticle di usion in limiting HDC rate. These e ects can be minimized for the less active 1% w/w Pd/C catalysts where the stirring speed 1000 rpm, hydrogen ow 150 cm3 min1 , catalyst concentration 0:5 g dm3 and particle sizes 45 m. The selectivity trends associated with 1% w/w Pd/C were the same whether the system operated under physical transport or chemical control. The selectivity with respect to 2-CP was however limited by mass-transfer processes in the HDC reaction using higher Pd loadings. ? 2002 Elsevier Science Ltd. All rights reserved.
Keywords: Environment; Multiphase reactions; Reaction engineering; Transport processes; Hydrodechlorination; Palladium/carbon

1. Introduction Chlorophenols (CPs) are commercially important chemicals, used as end products and intermediates in the manufacture of herbicides, dyes and plant growth regulators (Buchel, 1984; Tsyganok, Yamanaka, & Otsuka, 1999; Dabo et al., 2000; Juhler, Sorensen, & Larsen, 2001). By their nature, they are highly toxic and poorly biodegradable compounds that are now established as a class of priority environmental pollutants (Svenson, Kjeller, & Rappe, 1989; Wang, Hsieh, Chou, & Chang, 1999; Shin & Keane, 2000; USEPA, 2000). There is a denite need for e cient methods of dechlorination that are suitable for eliminating CPs from both concentrated industrial e uents and diluted polluted groundwater. At present, the available techniques

Corresponding author. Tel.: +1-859-257-8028; fax: +1-859-323-1929. E-mail address: makeane@engr.uky.edu (M. A. Keane).

used to detoxify chlorophenolic waste either require high temperatures and/or pressures (incineration, wet oxidation) (Suzdorf, Morozov, Anshits, Tsiganova, & Anshits, 1994; Shin & Keane, 1998; Qin, Zhang, & Chuang, 2001), complex equipment in the case of photochemical (Skurlatov et al., 1997; Dionysiou et al., 2000; Nensala & Nyokong, 2000; Ormad, Ovelleiro, & Kiwi, 2001) and sonophotochemical degradation (Shirgaonkar & Pandit, 1998) or are limited to a narrow concentration range (bioremediation) (Svenson et al., 1989; Janssen, Oppentocht, & Poelarends, 2001). Oxidative degradation, as a destructive methodology, can produce even more toxic by-products such as dioxins (Erickson, Swanson, Flora, & Hinshaw, 1989). Catalytic hydrodechlorination (HDC) represents a promising alternative non-destructive approach that can operate under mild conditions and is suitable for the treatment of both concentrated and diluted CP streams (Chon & Allen, 1991; Hoke, Gramiccioni, & Balko, 1992; Felis, Bellefon, Fouilloux, & Schweich, 1999a; Shin & Keane, 1998, 1999, 2000). The HDC products may be reused or recycled and the catalytic

0009-2509/02/$ - see front matter ? 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 0 0 9 - 2 5 0 9 ( 0 2 ) 0 0 4 7 6 - 1

258

G. Yuan, M. A. Keane / Chemical Engineering Science 58 (2003) 257 267

reactor can be readily coupled with separation units (Felis, Fouilloux, Bellefon, & Schweich, 1999b; Matatov-Meytal & Sheintuch, 2000). In a very recent study, Bengtson, Scheel, Theis, and Fritsch (2002) have demonstrated the viability of coupling a semi-permeable membrane to a chemical reactor as a means of simultaneously concentrating and reacting 4-CP in aqueous solution. The study of the catalytic HDC of CPs was initiated in the 1960s (Rylander, 1967) but the rst comprehensive report of the liquid phase exhaustive HDC of CPs to phenol over Pd/C only appeared in 1992 (Hoke et al., 1992). At present, the available literature relates to Pd=Al2 O3 , Pd/C, RhPt/C and Ru/C catalysts (Hoke et al., 1992; Shuth & Reinhard, 1998; Felis et al., 1999a; Matatov-Meytal & Sheintuch, 2000; Ukisu, Kameoka, & Miyadera, 2000). The (activated) carbon supported systems have been shown to exhibit high HDC activities where the carbon substrate may serve as a source of reactive spillover hydrogen (Shindler, Matatov-Meytal, & Sheintuch, 2001). The catalytic HDC of 4-chlorophenol (4-CP), 2,4-dichlorophenol (DCP) and 2,4,5-trichlorophenol (TCP) has been investigated in the liquid phase, but because of the low solubility of CPs in water, a complete kinetic analysis of HDC in aqueous solution has been limited to 4-CP as one of the more soluble (ca. 2:7 g per 100 cm3 water at 293 K) isomers (Felis et al., 1999a; Shindler et al., 2001). One critical issue associated with liquid phase HDC is the appreciable catalyst deactivation caused by the HCl by-product (Urbano & Marinas, 2001). Due to the weak acidity of CPs, the addition of base can serve to increase solubility as well as limiting HCl poisoning (Felis et al., 1999a). With the addition of base, catalyst deactivation is largely governed by HCl solubility/transport and the nature of the basic species in the catalyst matrix. Hoke et al. (1992) have reported higher HDC rates and enhanced catalyst stability in an ethanol/water mixture when compared with pure ethanol. In marked contrast, Shindler et al. (2001) obtained a constant HDC of 4-CP in water over Pd supported on carbon cloths for up to 3 h without the addition of any base. In practical applications, industrial waste e uents typically contain an array of poly-chlorinated phenols and their isomers. In order to develop more e ective catalysts and build a kinetic database on which to establish a working process, there is an urgent demand for a fundamental understanding of the catalytic HDC of mono and poly-chlorinated phenols. Although the feasibility of exhaustive aqueous phase dechlorination of poly-chlorophenols to phenol over Ru/C and Pd/C has been established (Felis et al., 1999a; Hoke et al., 1992), a detailed kinetic study is still not forthcoming. The possibility of structure sensitivity and the nature of the reaction mechanism in the conversion of poly-chlorophenols have been considered only to a limited extent in the literature (Felis et al., 1999a; Hoke et al., 1992; Urbano & Marinas, 2001; Matatov-Meytal & Sheintuch, 2002). DCP is a priority pollutant used in the production of the commercial pesticide, 2,4-D, of which

ca. 26; 300 tons were produced in the US in 1995 (USEPA, 1998). Contact with concentrated 2,4-DCP can prove fatal to humans (Ormad et al., 2001; USEPA, 2000) and even in very diluted aqueous solution (solubility: ca. 0:5 g per 100 cm3 water at 293 K) this compound inhibits the activity of many micro-organisms used in waste water treatment (Wang, Lee, & Kuan, 2000). In addition, this isomer is the intermediate product in the HDC of 2,4,6- and 2,4,5-TCP, also established highly toxic compounds. The preceding makes 2,4-DCP a suitable model compound for a HDC kinetic study. Although the gas phase HDC of 2,4-DCP over Ni=SiO2 has been considered (Shin and Keane, 1999), detailed studies of the liquid phase HDC of 2,4-DCP have yet to be reported. The work presented herein focuses on the liquid phase HDC of 2,4-DCP over Pd/C catalysts where the Pd loading spans the range 110% w/w. The high activity of Pd/C leads to signicant mass-transfer limitations in liquid phase operation; these transport limitations have been evaluated. 2. Experimental 2.1. Chemicals Chlorophenols (2-CP (99+%), 4-CP (99+%), 2,4-DCP (99+%)) and NaOH (99+%) were purchased from the Aldrich Chemical Company and used as received. The Pd on activated carbon catalysts with the nominal loadings (1%, 3%, 5% and 10% w/w) were also supplied by Aldrich with the following information: 1% Pd/C (12% w/w H2 O, 1.04% w/w Pd), 3% Pd/C (no moisture content, 3.1% w/w Pd), 5% Pd/C (1% w/w H2 O, 5.17% w/w Pd), 10%Pd/C (9.80 10.20% w/w Pd). All the catalysts were used as received but sieved (ATM ne test sieves) into batches of varying particle diameter range: 100 mesh (150 m), 100 200 mesh (150 75 m), 200 325 mesh (75 45 m), 325 400 mesh (45 37 m) and 400 mesh (37 m). The mean particle diameters in the batches as supplied were 89 m (1% w/w Pd/C), 75 m (3% w/w Pd/C), 75 m (5% w/w Pd/C) and 56 m (10% w/w Pd/C); these were determined on the basis of a distribution of particle sizes by weight. Stock reaction solutions were prepared with deionized water. 2.2. Catalytic procedure The liquid phase HDC reaction was carried out in a stirred Pyrex vessel 10 cm high 6 cm i.d., equipped with a H2 supply at a constant volumetric ow rate; a schematic for the reactor is provided in Fig. 1. The H2 ow was delivered using a mass ow controller (Brooks) and adjusted from 20 to 500 cm3 min1 . The optimum liquid content of the reactor is 100 cm3 . A four-bladed, 45 pitched glass impeller provided e ective agitation with a stirring speed in the range of 0 2000 rpm (Eurostar digital). A water circulating

G. Yuan, M. A. Keane / Chemical Engineering Science 58 (2003) 257 267

259

Fig. 1. Schematic diagram of the slurry reactor system, showing: (1) He cylinder; (2) pressure regulator; (3) needle valve; (4) hydrogen cylinder; (5) mass- ow controller; (6) water bath and recirculator; (7) thermocouple probe; (8) sampling tubing; (9) gas sparge tubing; (10) impeller; (11) jacketed reaction vessel; (12) pH probe; (13) gas inlet; (14) sampling port (with lter); (15) cooling water inlet; (16) cooling water outlet; (17) gas vent; (18) pH data acquisition; (19) inlet of recirculated water; (20) outlet of recirculated water.

jacket connected to a water bath and recirculator (Julabo) was used to stabilize the temperature to within 0:25 K. At the beginning of each experiment, 100 cm3 of stock aqueous chloroarene solution (0.0475 and 0:09 mol dm3 ) containing NaOH ([Cl]/[OH] slightly above 1 mol mol1 ) was charged with catalyst. The suspension was agitated in a He ow (50 cm3 min1 ) and the temperature allowed to stabilize at 302:5 0:25 K. Hydrogen was then introduced (time t = 0 for reaction) and the pH of the reaction mixture was monitored continuously using a Dow-Corning pencil electrode coupled to a data logging and collection system (Pico Technology Ltd.). A non-invasive liquid sampling system via in-line lters allowed a controlled syringe removal of aliquots (0:5 cm3 ) of reactant/product(s) with no loss of catalyst from the system. Before gas chromatographic analysis, the basic solution samples were neutralized with dilute acetic acid (ca. 0:2 M). 2.3. Analysis and calculation The gas chromatograph (Perkin-Elmer Auto System XL) with an auto sampler was equipped with an FID and a DB-1 J& W Scientic capillary column (i:d: = 0:2 mm, length = 50 m, lm thickness = 0:33 m). The relative peak area % was converted to mol% using regression equations based on detailed calibration and the detection limit typically corresponded to a feedstock conversion 0:4 mol%: overall analytic reproducibility was better than 5%. The concentration of organic species (2,4-DCP, 2-CP and phenol) in the bulk liquid phase was determined using the total mass balance in the reaction mixture where the organic species were taken to be non-volatile and the e ect of uptake on the carbon support was negligible (Shindler et al., 2001). The HCl produced and hydrogen consumption during reaction (mol dm3 ) has been calculated from the molar balance based on GC analysis of organic content. The (overall) selectivity (as a percentage) with respect to 2-CP

(S2-CP ) is dened as the mol% 2-CP in terms of the total moles of product(s) formed, i.e. S2-CP % = [2-CP] 100: [2-CP] + [Phenol] (1)

The observed initial rate of 2,4-DCP consumption was determined using pseudo-rst-order linear regression from temporal concentration proles. The observed maximum H2 consumption rates were likewise estimated from the temporal HCl concentration response. In each instance, HDC was accompanied by an induction period associated with catalyst activation and hydrogen transport. A time lag (6 4 min) was employed to adjust the reaction time scale and the data within this time delay were neglected.

3. Results and discussion 3.1. Hydrodechlorination pathway and pH response The HDC of 2,4-DCP can proceed in a stepwise and/or concerted fashion with 2-CP and 4-CP as partially dechlorinated products (see Fig. 2). Typical reactant/products distributions as a function of time are shown in Fig. 3(a), taking the 10% w/w Pd/C as a representative catalyst. The results of two separate catalytic runs (denoted by solid and open symbols) are provided to illustrate the degree of raw data reproducibility. The reaction was 100% selective in terms of dechlorination in that there was no evidence of any ring reduction before dechlorination was complete. At reaction times in excess of 20 min, further conversion to cyclohexanone and cyclohexanol was observed. In every instance, the mol fraction of 4-CP in the reaction mixture was below detection limit ( 0:4%, ca. 0:4 mM); the selectivity with respect to 4-CP approaches zero within the range of dechlorination reaction conditions considered in this study.

260

G. Yuan, M. A. Keane / Chemical Engineering Science 58 (2003) 257 267

the point at which the pH change was at a maximum (ca. 11) corresponds to the decline in selectivity with respect to the intermediate partially dechlorinated 2-CP. When all the chloroaromatic (2,4-DCP+2-CP) was converted to phenol, a minimum value of pH (ca. 7.8) was observed. There was a slight increase of pH at extended reaction times due to the further hydrogenation of phenol to cyclohexanone and cyclohexanol. 3.2. Mass transport resistances In order to assess the intrinsic activity of the catalysts in the HDC of 2,4-DCP, the reaction conditions must rst be established wherein mass-transfer limitations are negligible and the reactor is operated in the reaction-limited region. In this semi-batch slurry type reactor, the overall rate of chemical transformation is governed by a series of reaction and mass-transfer steps that proceed simultaneously (Keane, 1997): (a) H2 di usion through the gas lm at the gas/liquid interface; (b) H2 di usion through the liquid lm at the gas/liquid interface; (c) chloroaromatic and H2 diffusion through the liquid lm at the liquid/solid interface; (d) reaction and di usion at the external/internal surface of catalyst; (e) the reverse transport of products into bulk solution. As undiluted H2 gas has been used in the reaction, the gas lm mass-transfer resistance can be neglected and the observed consumption rate (RA , mol dm3 min1 ) of species A (i.e. H2 ) in bulk liquid can be expressed from lm theory as (Carberry, 1976) RA = dCAL = kL a(CAL CAL ) dt (2) (3) (4)

Fig. 2. Schematic of the HDC pathway for 2,4-DCP.

100

80

mol %

60

40

20

0 0 5 10 Time (min) 15 20

(a)

13 12 11 pH 10 9 8 7 0
(b)

= kS aS (CAL CAS ) = i RV ; AS

10

15 Time (min)

20

25

30

Fig. 3. HDC of 2,4-DCP over 10% w/w Pd/C: H2 feed rate = 250 cm3 min1 ; stirring speed = 1100 rpm; initial [Cl]=[Pd] = 200 mol mol1 . (a) Liquid phase composition vs. time in terms of mol% 2; 4-DCP(; ); 2-CP( ; 4) and phenol(; ), (b) pH change in bulk solution with time. Note: solid and open symbols represent two separate runs with di erent samples from the same batch of catalyst.

The temporal pH of the bulk solution is shown in Fig. 3(b). The initial pH value was always above 12.4 for [Cl]=[OH] = 1 mol mol1 . The decrease of pH was initially slight but accelerated with the consumption of 2,4-DCP and

where Eq. (2) describes the mass transfer at the gas/liquid interface (G/L), C AL is the equilibrium concentration of species A in the liquid lm and CAL the concentration of species A in bulk liquid. Eq. (3) represents the mass transfer at the liquid (L)/solid (S) interface and CAS is the concentration of species A at the external surface of the catalyst particle. Eq. (4) integrates the internal e ectiveness factor (i ) with an arbitrary volumetric reaction rate expression (RV ), so that it can describe the overall reaction/di usion AS rate at the external liquid/solid interface. The HDC reaction rate under ambient pressure is controlled by the concentration of H2 at the catalyst surface (H2 solubility in water at 293 K, 1 bar = ca. 0:8 mmol kg1 water (Shindler et al., 2001)). The G/L interface mass-transfer resistance can be minimized by increasing the volumetric mass-transfer coefcient (kL a) and so the H2 concentration in bulk liquid can approach the equilibrium value (CAL = CAL = 0:8 mM). In this study, three established diagnostic criteria (Butt, 2000), i.e. H2 feed rate, stirring speed and catalyst concentration

G. Yuan, M. A. Keane / Chemical Engineering Science 58 (2003) 257 267


10

261

10

10 C 2,4-DCP (mol dm )

8
102 C2,4-DCP (mol dm-3)

-3

6 4

0 0 10 20 30 40 50

0 0 10 20 30 40 50

(a)
100

Time (min)

(a)
100

Time (min)

80
80
S2-CP %

60

40

S2-CP %
0.0 0.2 0.4 X 2,4-DCP 0.6 0.8 1.0

60

40

20
20

0
0 0 0.2 0.4 X 2,4-DCP 0.6 0.8 1

(b)

Fig. 4. (a) Variation of 2,4-DCP concentration with time during the HDC of 2,4-DCP over 10% w/w Pd/C with varying H2 feed rate: 50 cm3 min1 ( ); 100 cm3 min1 (4); 150 cm3 min1 ( ); 200 cm3 min1 (); 250 cm3 min1 (). (b) Selectivity with respect to 2-CP (S2-CP %) as a function of 2,4-DCP conversion (X2; 4-DCP ), symbols as above. Stirring speed = 1100 rpm; initial [Cl]=[Pd] = 400 mol mol1 .

(b)

Fig. 5. (a) Variation of 2,4-DCP concentration with time during the HDC of 2,4-DCP over 10% w/w Pd/C as a function of stirring speed: 500 rpm ( ); 700 rpm ( ); 900 rpm (4); 1000 rpm (); 1100 rpm (). (b) Selectivity with respect to 2-CP (S2-CP %) as a function of 2,4-DCP conversion (X2; 4-DCP ), symbols as above. H2 feed rate=250 cm3 min1 ; [Cl]=[Pd] = 400 mol mol1 .

have been employed to evaluate the G/L mass-transfer limitations experimentally. Hydrogen mass-transfer limitations can rst be minimized by an increase in H2 feed rate, which can enhance the gas holdup in a three phase slurry conguration and therefore increase the total area of the G/L interface (Butt, 2000). An adjustment of the H2 feed and stirring speed can at least ensure a steady state H2 consumption in the bulk liquid, and the attainment of equilibrium at the G/L interface. The in uence of H2 feed rate on the progress of the reaction is shown in Fig. 4(a). It can be seen that there was a signicant increase in the rate of 2,4-DCP consumption when the H2 ow was increased from 50 to 100 cm3 min1 . However, the reaction was essentially insensitive to any increase in H2 feed rate above 150 cm3 min1 . The H2 feed rate had no signicant in uence on the selectivity with which 2-CP was formed during the reaction (Fig. 4(b)) and each selectivity vs. activity prole overlapped. A more e ective agitation

of the reaction mixture can enhance the rate of mass transfer at both gas/liquid and liquid/solid interfaces. The in uence of stirring speed on 2,4-DCP consumption is shown in Fig. 5(a). The decline in 2,4-DCP concentration increased as the stirring speed was raised from 500 to 900 rpm but remained largely unchanged when the stirring speed exceeded 1000 rpm. It can be seen in Fig. 5(b) that variations in stirring speed also had no signicant e ect on the selectivity with respect to 2-CP, i.e. both the H2 feed rate and stirring speed do not in uence the conversion/selectivity relationship, which can be considered to operate under surface control. A third verication that the apparent reaction rate is not under gas/liquid interface mass-transfer control lies in the in uence of catalyst concentration (gPd dm3 solution). At higher values of the latter, the rate of gas adsorption can be controlling due to a higher apparent HDC rate but an invariant H2 adsorption rate that is limited by mass transfer at G/L.

262
10

G. Yuan, M. A. Keane / Chemical Engineering Science 58 (2003) 257 267

8 10 [HCl] (mol dm )
-3

0 0 0.02 0.04 0.06 0.08 0.1 0.12

(a)
100

Specific time (min.gPd)

(Carberry, 1976). The data plotted in Fig. 6(b) show that the selectivity trends are the same in the catalyst-controlling region but deviate to some extent under gas adsorption control (at catalyst concentration = 1:0 g dm3 ). Repeated sets of runs are again included to illustrate experimental reproducibility. The L/S interface mass-transfer and intraparticle di usion limitations are more di cult to assess separately, because both are related to the intrinsic catalyst activity and particle size. However, the importance of external L/S mass-transfer and internal di usion limitations can be evaluated primarily by calculation. Assuming the reaction is fast (CAS 0), the maximum consumption rate of species A (RAmax ) under L/S mass control can be deduced from Eq. (3) as dCAL = kS aS CAL ; (5) RAmax = dt where the reaction operates under catalyst or chemical control, the liquid phase mass-transfer resistance can be neglected and the concentration of species A in the liquid phase can be taken to be that at equilibrium (C AL CAL ). The maximum rate of conversion under L/S interface mass-transfer control can be obtained from Eq. (4). Taking the observed conversion rate of A (RAobs ), in the absence of interphase concentration gradients in an isothermal system, the ratio RAobs =RAmax = Ca (Carberry number) should 0:1 for a rst-order reaction (Butt, 2000), i.e. RA RAobs 0:1: (6) (Ca)A = obs = RAmax kS aS C AL The volumetric surface area of the catalyst particles aS (m2 m3 liquid) can be estimated using Eq. (7) 6:0mC ; (7) aS = S dS where mC is the mass of catalyst per unit volume of solution (kg m3 liquid) and S (kg m3 catalysts) the bulk density of the solid catalyst (experimental estimation is 275 kg m3 for Pd/C) and the particle diameter (dS ) is determined on a distribution by weight basis (see Experimental section). The mass-transfer coe cient kS can be estimated from empirical correlation. Mass transfer to particles submerged in a uid is generally described by equations based on boundary layer theory, which leads to a correlation equation of the form (Beenackers & Van Swaaij, 1993) Sh = 2:0 + C Ren Scm with the Sherwood number (Sh) dened as k S dS Sh = : D (8)

80

S2-CP %

60

40

20

0 0 0.2 0.4 X 2,4-DCP 0.6 0.8 1

(b)

Fig. 6. (a) Concentration of HCl produced as a function of specic time during the HDC of 2,4-DCP over 10% w/w Pd/C where catalyst concentration = 1:0 g dm3 (; ); 0:5 g dm3 ( ; ); 0:33 g dm3 (4); 0:25 g dm3 ( ). (b) Selectivity with respect to 2-CP (S2-CP %) as a function of 2,4-DCP conversion (X2; 4-DCP ), symbols as above. Initial concentration of 2; 4-DCP = 0:0475 mol dm3 ; H2 feed rate = 250 cm3 min1 ; stirring speed = 1100 rpm; initial [Cl]=[Pd] = 100400 mol mol1 . Note: solid and open symbols represent two separate runs with di erent samples from the same batch of catalyst.

In order to ensure that the uid phase is saturated with H2 and the only possible rate-limiting step is reaction/di usion about (or within) the catalyst particles, a series of HDC reactions at di erent catalyst concentrations were examined and the experimental results are shown in Fig. 6(a): the specic time (plotted on the abscissa) is dened as the product of time and mass of Pd. This plot allows us to compare the 1 specic 2,4-DCP consumption rate (mol dm3 min1 gPd ) at di erent catalyst concentrations. It is apparent that when the initial Cl/Pd ratio (mol mol1 ) is above 200 (catalyst concentration 0:5 g dm3 ), the HCl production rates with respect to unit Pd mass overlap, demonstrating that the apparent (or global) HDC rates are directly proportional to the mass of catalyst. This suggests that when the catalyst concentration is below 0:5 g dm3 , reactor operation is removed from the gas adsorption-controlling region

(9)

Because of the ne solid particle size, we can assume that the relative velocity between particles and uids is quite low and the Reynolds number (Re) approaches zero. The Sherwood number in Eq. (8) approaches 2 (Deen, 1998) and the calculated maximum consumption rate of species A under L/S external mass-transfer limitation (RAmax )

G. Yuan, M. A. Keane / Chemical Engineering Science 58 (2003) 257 267

263

can be calculated from a combination of Eqs. (7) and (9) to give RAmax 12:0DAL mC C AL = ; 2 S dS (10)
10 C 2,4-DCP (mol dm )
-3

where DAL is the di usivity of species A in the liquid phase. The di usion coe cient of H2 in water has been reported to equal 4:5 109 m2 s1 at 298 K (Cussler, 1997) and the di usivity of 2,4-DCP in water may be estimated using the Wilke-Chang equation, which is an empirical modication of the StokesEinstein relation, represented by (Cussler, 1997) 7:4 1012 ( ML )0:5 T DBL = (m2 s1 ); 0:6 L VB (11)

0 0 30 60 Time (min) 90 120 150

(a)
100 80 60 40 20 0 0 0.2

where B corresponds to the solute (2,4-DCP), L the solvent (water), an association factor (=2:26 for water), ML the molecular weight of water (g mol1 ), T the absolute temperature (K), L water viscosity (cP) and VB the critical molar volume of solute B. The calculated di usivity of 2,4-DCP in water at 303 K is 2:4109 m2 s1 . The calculated maximum consumption rate of hydrogen (RAmax ) and 2,4-DCP (RBmax ) under conditions of transport control and the observed initial consumption rates of hydrogen (RAobs ) and 2,4-DCP (RBobs ) consumption over 1%, 3%, 5% and 10% w/w Pd/C catalysts are compared in Table 1. The importance of internal (intraparticle) di usion was evaluated by the calculation of the Weisz-Prater criterion (Lee, 1985) where, in the absence of concentration gradients in an isothermal spherical particle for a rst-order reaction, the following applies: (
S )A

S2-CP %

0.4 X 2,4-DCP

0.6

0.8

(b)

RAobs (dS =6)2 1: C AL DAe

(12)

Fig. 7. (a) Variation of 2,4-DCP concentration with time during the HDC of 2,4-DCP over 1% w/w Pd/C as a function of catalyst particle size (dS ; m): 125 75 m (); original batch (weighted average dS = 89 m) ( ); 75 45 m (); 45 37 m (4); 37 m ( ). (b) Selectivity with respect to 2-CP (S2-CP %) as a function of 2,4-DCP conversion (X2; 4-DCP ), symbols as above. Initial concentration of 2; 4-DCP = 0:0475 mol dm3 ; H2 feed rate = 250 cm3 min1 ; stirring speed = 1100 rpm; mC = 0:5 g dm3 .

The parameter DAe is the e ective di usivity of species A in the catalyst matrix where, in the absence of any information concerning the porosity and tortuosity associated with these catalysts, an order of magnitude estimation DAe = 0:1DAL was employed as discussed by Butt (2000). The calculated values of S with respect to H2 (A) and 2,4-DCP (B) are also listed in Table 1. It can be seen that the observed maximum rates of H2 consumption are greater than the calculated values (for L/S mass control). Moreover, the values of the Ca number with respect to H2 are all greater than unity. In contrast, the Ca number with respect to 2,4-DCP is signicantly less than 1. The results suggest that each catalytic systems operates under H2 transfer limitations at the L/S interface but the interphase mass transfer of 2,4-DCP/2-CP can be considered to be of little importance. The calculated values of S indicate that the H2 intraparticle di usion also limits the HDC rate for each catalyst with the exception of the 1% w/w Pd loading. The e ects of intraparticle di usion of the organic species can, however, be neglected in each case.

An increase in H2 solubility (elevated pressures), lower Pd loading and reduction in the catalyst particle size can serve to minimize the contribution to HDC rate due to intraparticle mass transfer and transport at the L/S interface. From a consideration of Eqs. (10) and (12), the most effective strategy is a reduction in the size of the catalyst particle. Taking the catalyst that exhibits the lowest activity (1% w/w Pd/C), the e ect of particle size on HDC is shown in Fig. 7. A signicant enhancement in 2,4-DCP consumption is evident when the catalyst size was reduced from that associated with the original batch (dS = 89 m) to less than 45 m, diagnostic of L/S mass-transfer control for larger particle sizes. The apparent rate (in terms of 2,4-DCP consumption) exhibited little dependence on size for particles smaller than 45 m. Taking particles 45 m the calculated Weisz-Prater criterion with respect to H2 = 0:27, indicative of a limited contribution to overall rate by internal di usion considerations. We can conclude that in the case

264

G. Yuan, M. A. Keane / Chemical Engineering Science 58 (2003) 257 267

Table 1 A comparison of calculated (max) with observed (obs) consumption rates of hydrogen (A) and 2,4-DCP (B) with an evaluation of liquid/solid interphase and intraparticle mass-transfer limitations using Carberry (Ca) and Weisz-Prater ( S ) parameters

Pd mC (w/w%) (g dm3 ) 10 5 3 1 0.5 0.5 0.83 0.5

RAmax (mol min1 dm3 ) 15 104 8:5 104 14 104 6:1 104

RBmax (mol min1 dm3 ) 4:5 102 2:6 102 4:4 102 1:9 102

RAobs (mol min1 dm3 ) 80 104 86 104 86 104 12 104

RBobs (mol min1 dm3 ) 4:9 103 6:0 103 6:4 103 0:8 103

(Ca)A 5.3 10 6.1 2.0

(Ca)B 0.11 0.23 0.15 0.04

S )A

S )B

1.9 3.7 3.7 0.7

0.04 0.08 0.09 0.02

of 1% w/w Pd/C with sizes 45 m, the reaction largely falls within the kinetic controlled region. Once again, the selectivity/activity trends coincide in both H2 transport limited and kinetically controlled regimes, as shown in Fig. 7(b). Taking the 1% w/w Pd/C, even at larger particle sizes, the mass-transfer limiting reagent is solely H2 (Table 1) and the constraints due to the transport of 2,4-DCP and 2-CP are not signicant with common associated selectivity/conversion proles. However, at higher Pd loadings, the liquid/solid interphase and intraparticle mass transfer may impact not only on HDC rate but may also in uence the liquid phase composition. If we consider the HDC of 2,4-DCP to phenol as a series/parallel process (see Fig. 2) involving pseudo-rst-order irreversible steps (assume constant H2 concentration at the catalyst surface) and denoting the constituent H2 , 2,4-DCP, 2-CP and phenol as A, B, D and E, respectively, 2; 4-DCP(B) + H2 (A) 2-CP(D); 2-CP(D) + H2 (A) Phenol(E); 2; 4-DCP(B) + 2H2 (A) Phenol(E):
k3 (CA ;CB ;CE ) k2 (CA ;CD ;CE ) k1 (CA ;CB ;CD )

Yd =

RDobs i 1 k1 (i2 k2 )(i1 Da1 ) = RBobs i 1 k1 + i 3 k3 (i1 k1 + i3 k3 )(1 + i2 Da2 ) 1 + i1 Da1 + i3 Da3 CDL i2 k2 ; (i1 k1 + i3 k3 ) 1 + i2 Da2 CBL (21)

where Yd is the observed di erential yield with respect to the partially dechlorinated 2-CP product. As (Da1 =Da2 =Da3 ) = (k1 =k2 =k3 ), Eq. (21) can reduce to Yd = RDobs i 1 k1 = RBobs i 1 k1 + i 3 k3 1 (i2 Da2 ) (1 + i2 Da2 ) (22)

i2 k2 1 + i1 Da1 + i3 Da3 CDL (i1 k1 + i3 k3 ) 1 + i2 Da2 CBL

and the initial di erential 2-CP yield can be estimated from the intercept of a plot of YD vs. CDL =CAL Ydin = 1 i 1 k1 : i1 k1 + i3 k3 (1 + i2 Da2 ) (23)

(13) (14) (15)

Then, the observed rate of consumption of 2,4-DCP (RBobs ) and rate of formation of 2-CP (RDobs ) in the bulk liquid phase can be expressed as RBobs = (kS )B aS (CBL CBS ) = (i1 k1 + i3 k3 )CBS ; RDobs = (kS )D aS (CDS CDL ) = i1 k1 CBS i2 k2 CDS ; (16) (17)

Representative 2-CP di erential yields and overall selectivity dependences are plotted in Fig. 8 for an array of Pd/C particle sizes and Pd loadings; the observed rates RBobs and RDobs (Fig. 8(a)) at time tj are calculated from the second-order nite di erential as RBobs = RDobs = dCBL dt
t=tj

CBLj1 CBLj+1 ; 2 t

(24) (25)

where (kS )B and (kS )D are the L/S interface mass transfer coe cients for 2,4-DCP (B) and 2-CP (D), respectively, i is the internal e ectiveness factor and the subscript number represents each reaction step (1, 2, 3). Because (kS )B =(kS )D = DB =DD 1, then the L/S mass-transfer coefcient kS = (kS )B = (kS )D . Designating Damkohler numbers Da as Da1 , Da2 and Da3 for each respective step, we nd that (Carberry, 1976) Da1 = k1 =kS aS ; Da2 = k2 =kS aS ; Da3 = k3 =kS aS ; (18) (19) (20)

dCDL dt

t=tj

CDLj+1 CDLj1 : 2 t

The parameters CBLj+1 and CDLj+1 represent the bulk liquid concentration of species B/D at time tj+1 , CBLj1 and CDLj1 refer to time tj1 and t is the pertinent experimental time increment. From a consideration of Fig. 8(a), the proles associated with 1% w/w Pd/C exhibit a linear dependence where the initial di erential 2-CP yield is close to unity. This suggests that the apparent HDC reaction kinetics coincide with a rst-order homogenous consecutive reaction scheme (k1 =k3 =19 1), assuming a constant H2 concentration at the surface. The overlap of di erential yield and selectivity proles for the 1% w/w Pd/C of varying particle size (Fig. 8) indicates that the external/internal di usions of 2,4-DCP, 2-CP and phenol (Da1 , Da2 and Da3 0) and the

G. Yuan, M. A. Keane / Chemical Engineering Science 58 (2003) 257 267


1

265

0.5
y = -1.0538x + 0.95 1%Pd/C, ds<37micro y = -1.4739x + 0.85 10% Pd/C, ds<37micro

and 10% w/w Pd loadings (Fig. 8). The curvature of the di erential yield proles for the higher Pd loaded catalysts means that the application of a solely apparent rst-order kinetic treatment is not feasible. These results are suggestive of di erences in the intrinsic HDC behavior of Pd/C catalysts with di erent Pd content. An explicit assignment of chemical structure sensitivity is, however, outside the remit of this paper and will be the subject of a future report.

Yd

-0.5

-1 0 0.5 1
C2-CP/C2,4-DCP

1.5

4. Conclusions The liquid phase HDC of 2,4-DCP over Pd/C is e ective and exhaustive: 2,4-DCP is converted to phenol quantitatively and 2-chlorophenol (2-CP) is the only intermediate product within detection limit ( 0:2 mM). The system is 100% selective in terms of dechlorination and phenol hydrogenation only proceeds once complete dechlorination has been attained. The high HDC activity of Pd/C induces a series of mass-transfer limitations. Only in the case of the less active 1% w/w Pd/C with particle sizes 45 m, stirring speeds 1000 rpm, H2 feed rate 150 cm3 min1 and an initial Cl/Pd ratio 200 mol mol1 can we minimize the in uence of mass-transfer processes on the overall reaction rate. Under these conditions, we consider that HDC (at 303 K) is in the reaction-controlled region. The mass-transfer limitations show little in uence on the reaction selectivity prole for 1% w/w Pd/C. However, the selectivity with respect to 2-CP is limited by the L/S interphase and intraparticle di usion in the case of Pd/C catalysts with a higher Pd loading.

(a)

100 80 60 40 20 0
0 0.2 0.4 X2,4-DCP 0.6 0.8 1

Fig. 8. (a) Di erential 2-CP yield (Yd = Robs (2-CP)=Robs (2,4-DCP)) as a function of the concentration ratio (C2-CP =C2; 4-DCP ) in bulk liquid: original batch of 10% w/w Pd/C, weighted average dS = 56 m ( );10% w/w Pd/C; dS 37 m ( ); original batch 5% w/w Pd/C, weighted average dS = 75 m (); original batch 1% w/w Pd/C, weighted average dS = 89 m (); 1% w/w Pd/C, dS 37 m (O). (b) Overall selectivity with respect to 2-CP (S2-CP %) as a function of 2,4-DCP conversion (X2; 4-DCP ), symbols as above. Initial concentration of 2; 4-DCP = 0:0475 mol dm3 ; mC = 0:5 g dm3 ; H2 feed rate = 250 cm3 min1 ; stirring speed = 1100 rpm.

S2-CP %

(b)

Notation a aS A B Ca CA CAL C AL C dS D Da DAL DA e specic gasliquid contact area, m2 m3 liquid volumetric surface area of catalyst particles, m2 m3 liquid refers to hydrogen refers to 2,4-DCP Carberry number, dimensionless concentration of species A, mol dm3 bulk concentration of species A in liquid, mol dm3 concentration of species A in liquid at gasliquid interface, mol dm3 constant, dimensionless particle diameter, m refers to 2-CP Damkohler number, dimensionless molecular di usion coe cient of diluted species A in liquid solvent L, m2 s1 e ective molecular di usion coe cient of diluted species A in catalyst matrix, m2 s1

intraparticle di usion of H2 are not rate limiting (i1 , i2 and i3 1). Consequently, a slope of Yd vs. CDL =CBL ( k2 =k1 ) close to unity suggests that the initial HDC of 2,4-DCP and 2-CP formation are essentially equivalent for the Pd dilute catalyst. In the case of the 10% w/w Pd/C, the dechlorination yield/selectivity exhibited a decided dependence on particle size, where 2-CP production was enhanced over smaller catalyst particles (Fig. 8). This demonstrates that a partial dechlorination selectivity is also subject to mass-transfer limitations over Pd concentrated catalysts. The initial differential 2-CP yield approached unity for the smallest particle size (k1 =k3 6), which delivered the best intraparticle e ectiveness (i1 ) and lowest L/S interphase mass-transfer resistance (Da2 ) but both e ects still contribute to the liquid phase composition (see Eq. (23)). The 5% w/w Pd/C exhibited dechlorination behavior in terms of L/S contribution that is intermediate between that recorded for the 1%

266

G. Yuan, M. A. Keane / Chemical Engineering Science 58 (2003) 257 267

E G j kG kL kS (kS )A L m mC ML n RA RV AS Rin Robs Rmax Re S Sc Sh S% T VA X Yd

refers to Phenol gas phase subscript: refers to reaction steps (i.e. 1, 2, 3) gas-lm mass-transfer coe cient, m s1 liquid-lm mass-transfer coe cient, m s1 liquid-to-solid mass-transfer coe cient, m s1 liquid-to-solid mass-transfer coe cient with respect to species A, m s1 liquid phase constant, dimensionless catalyst concentration, g dm3 liquid solvent molecular weight, g mol1 constant, dimensionless global rate of reaction with respect to species A, mol dm3 min1 volumetric reaction rate expression, mol dm3 min1 initial rate of reaction, mol dm3 min1 observed reaction rate, mol dm3 min1 calculated maximum reaction rate, mol dm3 min1 Reynolds number, dimensionless solid phase Schmidt number (= = D), dimensionless Sherwood number (=kS dS =D around particles), dimensionless overall reaction selectivity (%) temperature, K critical molar volume of solute A, m3 mol1 conversion, dimensionless di erential yield with respect to 2-CP, dimensionless

Greek letters i
L L S S

internal e ectiveness factor, dimensionless viscosity of liquid, cp liquid density, kg m3 particle density, kg m3 association factor, dimensionless Weisz-Prater criterion: ( S )A = RAobs (dS =6)2 = C AL DAe , dimensionless

References
Beenackers, A. A. C. M., & Van Swaaij, W. P. M. (1993). Mass transfer in gasliquid slurry reactors. Chemical Engineering Science, 48, 31093139. Bengtson, G., Scheel, H., Theis, J., & Fritsch, D. (2002). Catalytic membrane reactor to simultaneously concentrate and react organics. Chemical Engineering Journal, 85, 303311. Buchel, K. H. (1984). Political, economic, philosophical aspects of pesticide use for human welfare. Regulatory Toxicology and Pharmacology, 4, 174191.

Butt, J. B. (2000). Reaction kinetics and reactor design (2nd ed.) (pp. 592 607). New York: Marcel Dekker Inc. Carberry, J. J. (1976). Chemical and catalytic reaction engineering (pp. 580 584). New York: McGraw-Hill Inc. Chon, S., & Allen, D. T. (1991). Catalytic hydroprocessing of chlorophenols. A.I.Ch.E. Journal, 3, 17301732. Cussler, E. L. (1997). Di usion: mass transfer in uid systems (2nd ed.) (pp. 112120). Cambridge, UK: Cambridge University Press. Dabo, P., Cyr, A., Laplante, F., Jean, F., Menard, H., & Lessard, J. (2000). Electrocatalytic dehydrochlorination of pentachlorophenol to phenol or cyclohexanol. Environmental Science and Technology, 34, 12651268. Deen, W. M. (1998). Analysis of transport phenomena (pp. 411 429). New York: Oxford University Press. Dionysiou, D. D., Khodadoust, A. P., Kern, A. M., Suidan, M. T., Baudin, I., & Laine, J. (2000). Continuous-mode photocatalytic degradation of chlorinated phenols and pesticides in water using a bench-scale TiO2 rotating disk reactor. Applied Catalysis B: Environmental, 24, 139155. Erickson, M. D., Swanson, S. E., Flora, J. D., & Hinshaw, G. D. (1989). Polychlorinated dibenzofurans and other thermal combustion products from dielectric uids containing polychlorinated biphenyls. Environmental Science and Technology, 23, 462470. Felis, V., Bellefon, C. D., Fouilloux, P., & Schweich, D. (1999a). Hydrodechlorination and hydrodearomatisation of monoaromatic chlorophenols into cyclohexanol on Ru/C catalysts applied to water depollution: In uence of the basic solvent and kinetics of the reactions. Applied Catalysis B: Environmental, 20, 91100. Felis, V., Fouilloux, P., Bellefon, C. D., & Schweich, D. (1999b). Detoxication of water containing parachlorophenol. Recent Progres en Genie des Procedes, 13(70), 303308. Hoke, J. B., Gramiccioni, G. A., & Balko, E. N. (1992). Catalytic hydrodechlorination of chlorophenols. Applied Catalysis B: Environmental, 1, 285296. Janssen, D. B., Oppentocht, J. E., & Poelarends, G. J. (2001). Microbial dehalogenation. Current Opinions in Biotechnology, 12, 254258. Juhler, R. K., Sorensen, S. R., & Larsen, L. (2001). Analysing transformation products of herbicide residues in environmental samples. Water Research, 35, 13711378. Keane, M. A. (1997). A kinetic treatment of heterogeneous enantioselective catalysis. Journal of the Chemical Society, Faraday Transactions, 93, 20012007. Lee, H. H. (1985). Heterogeneous reactor design (pp. 139 142) Boston: Butterworth Publishers. Matatov-Meytal, Yu., & Sheintuch, M. (2000). Catalytic regeneration of chloroorganics-saturated activated carbon using hydrodechlorination. Industrial and Engineering Chemistry Research, 39, 1823. Matatov-Meytal, Yu., & Sheintuch, M. (2002). Hydrotreating processes for catalytic abatement of water pollutants. Catalysis Today, 75, 6367. Nensala, N., & Nyokong, T. (2000). Photocatalytic properties of neodymium diphthalocyanine towards the transformation of 4-chlorophenol. Journal of Molecular Catalysis A: Chemical, 164, 6976. Ormad, M. P., Ovelleiro, J. L., & Kiwi, J. (2001). Photocatalytic degradation of concentrated solutions of 2,4-dichlorophenol using low energy light: Identication of intermediates. Applied Catalysis B: Environmental, 32, 157166. Qin, J., Zhang, Q., & Chuang, K. T. (2001). Catalytic wet oxidation of p-chlorophenol over supported noble metal catalysts. Applied Catalysis B: Environmental, 29, 115123. Rylander, P. N. (1967). Catalytic hydrogenation over platinum metals (pp. 405 411). New York: Academic Press. Shin, E.-J., & Keane, M. A. (1998). Gas phase catalytic hydrodechlorination of chlorophenols using a supported nickel catalyst. Applied Catalysis B: Environmental, 18, 241250.

G. Yuan, M. A. Keane / Chemical Engineering Science 58 (2003) 257 267 Shin, E.-J., & Keane, M. A. (1999). Detoxication of dichlorophenols by catalytic hydrodechlorination using a nickel/silica catalyst. Chemical Engineering Science, 54, 11091120. Shin, E.-J., & Keane, M. A. (2000). Gas phase catalytic hydroprocessing of trichlorophenols. Journal of Chemical Technololgy and Biotechnology, 75, 159167. Shindler, Yu., Matatov-Meytal, Yu., & Sheintuch, M. (2001). Wet hydrodechlorination of p-chlorophenol using Pd supported on an activated carbon cloth. Industrial and Engineering Chemistry Research, 40, 33013308. Shirgaonkar, I. Z., & Pandit, A. B. (1998). Sonophotochemical destruction of aqueous solution of 2,4,6-trichlorophenol. Ultrasonics and Sonochemistry, 5, 5361. Shuth, C., & Reinhard, M. (1998). Hydrodechlorination and hydrogenation of aromatic compounds over palladium on alumina in hydrogen-saturated water. Applied Catalysis B: Environmental, 18, 215221. Skurlatov, Yu. I., Ernestova, L. S., Vichutinskaya, E. V., Samsonov, D. P., Semenova, I. V., Rodko, I. Ya., Shvidky, V. O., Pervunina, R. I., & Kemp, T. J. (1997). Photochemical transformation of polychlorinated phenols. Journal of Photochemistry and Photobiology A: Chemistry, 107, 207213. Suzdorf, A. R., Morozov, S. V., Anshits, N. N., Tsiganova, S. I., & Anshits, A. G. (1994). Gas phase hydrodechlorination of chlorinated aromatic compounds on nickel catalysts. Catalysis Letters, 29, 4955.

267

Svenson, A., Kjeller, L. O., & Rappe, C. (1989). Enzymatic chlorophenol oxidation as a means of chlorinated dioxin and dibenzofuran formation. Chemosphere, 19, 585587. Tsyganok, A. I., Yamanaka, I., & Otsuka, K. (1999). Dechlorination of chloroaromatics by electrocatalytic reduction over palladium-loaded carbon felt at room temperature. Chemosphere, 39, 18191827. Ukisu, Y., Kameoka, S., & Miyadera, T. (2000). Catalytic dechlorination of aromatic chlorides with noble-metal catalysts under mild conditions: Approach to practical use. Applied Catalysis B: Environmental, 27, 97104. Urbano, F. J., & Marinas, J. M. (2001). Hydrogenolysis of organohalogen compounds over palladium supported catalysts. Journal of Molecular Catalysis A: Chemical, 173, 329345. USEPA (1998). The inventory of sources of Dioxin in the United States. EPA/600/P-98/002Aa. USEPA (2000). Chemical advisory and notice of potential risk: Skin exposure to molten 2,4-Dichlorophenol (2,4-DCP) can cause rapid death. Advisory Document No. 8EHQ-14302. Wang, K., Hsieh, Y., Chou, M., & Chang, C. (1999). Photocatalytic degradation of 2-chloro and 2-nitrophenol by titanium dioxide suspensions in aqueous solutions. Applied Catalysis B: Environmental, 21, 18. Wang, C. C., Lee, C. M., & Kuan, C. H. (2000). Removal of 2; 4-dichlorophenol by suspended and immobilized Bacillus insolitus. Chemosphere, 41, 447452.

Das könnte Ihnen auch gefallen