Ind. Eng. Chem. Res. 2000, 39, 583-591
583
KINETICS, CATALYSIS, AND REACTION ENGINEERING Oxidation of Aroclor 1248 in Supercritical Water: A Global Kinetic Study Gheorghe Anitescu† and Lawrence L. Tavlarides* Department of Chemical Engineering and Materials Science, Syracuse University, Syracuse, New York 13244
Supercritical water oxidation of extracted contaminants is the second step of a two-stage supercritical fluid technology proposed to remediate soils and sediments contaminated with polychlorinated biphenyls (PCBs) and/or polyaromatic hydrocarbons. In connection with the second step, the supercritical water oxidation rate of Aroclor 1248 (A1248), a mixture of ∼76 PCB congeners, is investigated at 25.3 MPa and temperatures of 723, 748, 773, and 823 K. The reactions are conducted in an isothermal, isobaric plug-flow tubular reactor, and GC/ECD, GC/ FID, GC/TCD, and GC/MS chromatographic methods are employed for product analysis. Experiments are conducted at a nominal A1248 feed concentration of 5.75 × 10-5 mol/L (reaction conditions) using a methanol solution of 5.245 g/L (5245 ppm) and H2O2 as an initial oxidant (providing ∼20 mol % excess of O2). Molar global conversion of A1248 varies from 36.06% (for residence time equal to 6.29 s at 723 K) to 99.95% (54.4 s at 823 K). The overall conversion follows apparent second order, and the rate constant calculated from the data leads to Arrhenius parameters of frequency factor A ) 1017.0(0.1 s-1 (mol/L)-1 and energy of activation Ea ) 186 ( 2 kJ/mol (44.43 ( 0.51 kcal/mol). The congener specific analysis indicates a buildup of intermediate congener byproducts, which also undergo oxidation decomposition. The identified reaction products are mainly biphenyl, low-chlorinated PCB congeners such as 2-chlorobiphenyl and 2,2′-dichlorobiphenyl, CO, and CO2. 1. Introduction One of the promising applications of supercritical water (SCW) as a reaction medium is in the waste treatment technology via supercritical water oxidation (SCWO). Two reviews made by Savage et al.1 and Savage2 give a comprehensive account of research in SCW up to 1999. Of all the articles published in this field in the last 11 years (more than 250), over half have appeared since 1995. SCWO of organic compounds is the process that has undoubtedly received the most attention. The technology takes advantage of the complete miscibility of most organic compounds and oxygen with SCW, eliminating the slow mass-transfer process that occurs in multiphase systems. Moreover, essentially complete conversion of organic carbon to carbon oxides occurs on the time scale of a few minutes. SCWO technology has been shown to be effective for destroying a large variety of industrial and high-risk wastes.3-6 The SCWO process is conducted at temperatures and pressures above the critical point of water (647 K and 22.1 MPa) and is considered applicable to aqueous streams containing 0-20% organics.7 In May 1994, the first commercial SCWO waste-processing facility was implemented in Austin, TX, by Eco Waste Technologies.8 The performance of this unit demon* Corresponding author. E-mail:
[email protected]. Fax: 315-443-1243. Tel.: 315-443-1883. † On leave from the Department of Physical Chemistry, Bucharest University, Bucharest 7034, Romania.
strates the large-scale viability of SCWO process. For wastewater-treated sludges, the effluent meets stringent environmental standards at costs considerably less than fluidized bed incineration.9 Different classes of organic compounds have been subjected to SCWO conditions including chlorinated aromatics.10-16 Oe17 reported a complete decomposition of polychlorinated biphenyls (PCBs) feed up to 7% using a bench-scale apparatus, and Hatakeda et al.18,19 reported SCWO of 3-chlorobiphenyl. Recently, Aki and Abraham20 made an economic evaluation of SCWO and a comparison with alternative waste treatment technologies. These results demonstrate the feasibility for SCWO destruction of environmentally harmful chemicals including PCBs. However, there is a gap in available data to optimally design SCWO reactors for PCBs in general and in particular for PCBs dissolved in soil extracts. Although the efficacy of the SCWO process has been proven,21 reaction kinetic studies of the process are still at a level of simple monocomponent systems such as methanol,22-26 phenol27 and its derivatives,10,11,28-30 C1 compounds,31,32 methylphosphonic acid,33 etc. Agreement between models and experimental data is far from expectations in most cases. Some SCWO experimental studies on model compounds were limited to a narrow range of temperature and residence time, preventing model development for optimal reactor design and defining conditions for high levels of conversion. These studies were made to test detailed chemical kinetic models or to test the applicability of combustion mechanisms at SCWO conditions. The temperature ranges
10.1021/ie990704l CCC: $19.00 © 2000 American Chemical Society Published on Web 02/03/2000
584
Ind. Eng. Chem. Res., Vol. 39, No. 3, 2000
Figure 1. Schematic flow diagram of continuous-flow SCWO reactor system. T ) thermocouple, P ) pressure indicator, RD ) rupture disk, MMV ) micrometric valve; FM ) flow meter, HPP-1 ) high-pressure ISCO syringe pump, HPP-2 ) Rainin pump.
were chosen to be around 773 K or higher, and residence times were in a very narrow range of 0-10 s with only few experiments over 10 s. Further, no kinetic studies have been reported on PCB oxidation in SCW to determine reactor optimum conditions and the nature and concentrations of reaction byproducts. This lack of suitable data led us to initiate research work in the field of SCWO. In addition, a two-stage supercritical fluid technology process of PCB extraction from soils and sediments with SC-CO2/methanol fluids followed by SCWO of the soil/sediment extracts is under investigation at Syracuse University. The second step requires oxidation of PCB/methanol extract. Therefore, a systematic kinetic study of this system was initiated with methanol oxidation26 over a wide range of conditions. That study was intended to assess the capabilities of this technique to obtain more accurate data in a reliable and quick manner and to serve as a bridge for PCB/MeOH SCWO. This study continues the work with the oxidation of Aroclor 1248 (A1248), a mixture of approximately 76 PCB congeners. Future SCWO studies will be conducted to understand the kinetics and reaction pathway of individual PCBs and to determine the conditions for complete oxidation of several single PCB congeners and other Aroclor mixtures in methanol. With appropriate models this information will permit evaluation of the feasibility and cost of SCWO of PCB extracts from soils. 2. Experimental Section 2.1. Apparatus. All of the A1248 oxidation experiments in supercritical water are conducted in a highpressure, isothermal plug-flow tubular reactor capable of continuous operation at temperatures up to 873 K and pressures up to 689 bar. Under the operating conditions, Reynolds numbers calculated for our reactor characterize a turbulent flow. The assumption of a plugflow reactor is therefore valid. As shown in Figure 1, the plug-flow reactor system consists of three major subsystems: pumps and preheaters, reactor, and cooling and separation. The experimental setup used is described in detail elsewhere,26 so only the most relevant
features will be given here. In the pump and preheating subsystem, PCBs/MeOH and oxidant (H2O2/H2O solutions) are delivered in separate lines by high-pressure feed pumps. The oxidizer preheated high-pressure tubing (1.6 mm i.d., 3.2 mm o.d., ∼4 m long) is Hastelloy C-276, while the organic feed line is stainless steel tubing (∼1 m long, 0.25 mm i.d.). The organic feed is mixed with supercritical oxidant in a mixing block designed to combine the two flows at a 45° angle of incidence. The combined flow exits the cylindrical block (3 cm long, 1.3 cm o.d. with a 1.6 mm tunnel) on the oxidant line direction. The temperature is monitored both at the inlet (in the center line of combined flows) and outlet (on the tubing wall) of the reactor. By measuring the wall temperature at the outlet of the reactor, a connection union with an undesired dead volume and the chance of PCB deposits is avoided. Temperature does not vary significantly inside and outside the reactor (maximum ( 1 °C). The variation of reactor residence time is achieved either by changing the reactor length (two sections of Hastelloy tubing of 5 and 16 m) while maintaining a constant feed rate or by changing the feed rate keeping the same reactor length. The reactor and the preheating coils are immersed in an isothermal fluidized sand bath. The cooling and separation subsystem consists of two water-cooled glass separators fabricated at Syracuse University. 2.2. Procedure. The oxidizer is fed to the system as a solution of H2O2/H2O. It is subsequently thermally decomposed in the preheating section to a high-pressure mixture of O2 in supercritical water. The preheated A1248/MeOH and oxidant streams are combined in a mixing block at the desired experimental temperature and then passed into the reactor. The pressure in the reactor is fixed by manually controlling the total flow rate of the exiting reactants/reaction products using a specially designed micrometric valve (Autoclave Engineers Inc.). The stream pressure is dropped to ambient conditions, and the gaseous and liquid phases are separated in either one of two separators connected in parallel. During the unsteady-state portion of the reaction, the cooled products are collected in the first separator and removed later. The products of the steadystate reaction (5-10 min) are depressurized, cooled, and separated in the second separator and further analyzed by chromatographic methods. The steady state is assumed to be attained after adequate time has transpired for reactants to flow through both the preheater and reactor and SCWO conditions (T, P, and flows) are stabilized. 2.3. Analytical Technique. Chromatographic data are obtained by three Hewlett-Packard 5890 series II gas chromatographs. The gaseous phase is captured into a 250 µL sample loop and analyzed by online GC/TCD. A portion of the liquid-phase product is diluted and analyzed by offline capillary GC/ECD, GC/FID, and/or GC/MSD to measure the amount of unreacted PCBs and some of the byproducts. GC separation is achieved on capillary columns HP Ultra-2 (25 m × 0.20 mm i.d., 0.33 µm film thickness), DB-1 (30 m × 0.32 mm i.d., 3 µm film thickness), and HP-5MS (30 m × 0.25 mm i.d., 0.25 µm film thickness), respectively. Helium is used as carrier gas with a flow rate of 2.5 mL/min (at 150 °C). For PCB analysis the oven temperature profile was 120-280 °C at 4 °C min-1. The injection port and FID/ ECD were operated at 250 and 300/330 °C, respectively. The GC was calibrated with standard solutions contain-
Ind. Eng. Chem. Res., Vol. 39, No. 3, 2000 585
ing certified concentrations of more then 100 PCB congeners in hexane. Chromatographic errors are found to be less than 4% through replicate analysis. Individual calibration curves are obtained for all of the 76 PCB congeners initially or/and finally present in A1248/ effluent streams over the full range of solute concentrations in the diluted sample solutions. To identify the oxidation reaction byproducts of A1248, the effluent liquid-phase samples were analyzed by GC/MS using the NIST library of spectra. 2.4. Reactants. The oxidant is O2, supplied for most of the experiments as a solution of hydrogen peroxide of 6 wt % concentration prepared from 30 wt % H2O2/ H2O solution (purum p.a., Fluka). A1248 mixture and the PCB standard solutions are received from SUNY at Albany (School of Public Health).34 The purity of methanol (Optima, Fisher Scientific) is min. 99.9%, and the methanol-A1248 feed is 0.88 mol % (1.53 wt %) in a balance of water (distilled and deionized). All reactants are used with no further purification. 2.5. Residence Time Calculation. Residence time (τ) is calculated by dividing reactor volume (V) by the volumetric flow rate at the entrance of the reactor at supercritical conditions (vSC,0):
τ ) V/vSC,0
(1)
On the basis of the mass balance of the materials in the liquid phase (L) and in the supercritical phase (SC), a simple equation to calculate residence time can be obtained:
τ ) V(FSC,0/FL)/vL
(2)
The residence time is in s, the volume of the reactor is in mL, the densities of the both phases are in g/mL, and the initial flow rate of the liquid reactants (vL) is in mL/ s. The density of water at reaction conditions is calculated using NBS Steam Tables.35 The fluid mixture density is assumed to be that of water as the concentration of A1248 and oxygen in the supercritical phase is less than 1.6 mol % and methanol density is sufficiently close to that of water. 3. Results and Discussion Table 1 shows the congener pattern and the molar composition (Yi) of A1248 used in SCWO experiments. The composition of A1248 has been obtained by GC/ECD from a solution with a total PCB concentration of 2100 ppb in methanol (Table 1). To identify the PCB congeners in A1248, a standard solution of 115 PCB congeners is employed. Seventy-three peaks are found from which 54 are considered, the remaining having too low concentrations to be accurately quantified. From the total of 76 congeners considered in these 54 peaks, 35 are positively identified and the other 41 are coeluted (Table 1). The averaged molar mass of A1248 is 292 g/mol,36 the same value as that of a tetrachlorobiphenyl. This could be a reason to consider A1248 as a pseudotetrachlorobiphenyl compound for the purpose of global kinetic studies. Moreover, among the Aroclor mixtures, A1248 contains the highest amount of 3,3′,4,4′-tetrachlorobyphenyl, one of the PCB congeners related to both high toxicity and resistance to decomposition. The main variable parameters in SCWO experiments with A1248 are the residence time and the temperature. The residence times of the experiments range from 4.53
Table 1. Congener Pattern and Molar Composition (Yi) of A1248 peak
IUPAC no.
congener structure
Yi (%)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55
1 4+10 7+9 6 5+8 19 18 15+17 24+27 16+32 26 31 28 33+53 51 22 45 46 52 49 47+48+75 44 37+42+59 41+64 40 74 70 66+95 91 56+60 92 84 90+101 99 83 97 87 85 77+110 82 123+149 118 153 105+132 141 138+163+164 187 185 174 180 170+190 199 196+203 194 206
2 22′ + 26 24 + 25 23′ 23 + 24′ 22′6 22′5 44′ + 22′4 236 + 23′6 22′3 + 24′6 23′5 24′5 244′ 2′34 + 22′56′ 22′46′ 234′ 22′36 22′36′ 22′55′ 22′45′ 22′44′ + 22′45 + 244′6 22′35′ 344′ + 22′34′ + 233′6 22′34 + 234′6 22′33′ 244′5 23′4′5 23′44′ + 22′35′6 22′34′6 233′4′ + 2344′ 22′355′ 22′33′6 22′34′5 + 22′455′ 22′44′5 22′33′5 22′3′45 22′345′ 22′344′ 33′44′ + 233′4′6 22′33′4 2′344′5 + 22′34′5′6 23′44′5 22′44′55′ 233′44′ + 22′33′46′ 22′3455′ 22′344′5′ + 233′4′56 + 233′4′5′6 22′34′55′6 22′3455′6 22′33′456′ 22′344′55′ 22′33′44′5 + 233′44′56 22′33′4566′ 22′33′44′5′6 + 22′344′55′6 22′33′44′55′ 22′33′44′55′6
0.000 0.327 0.069 0.148 0.732 0.197 3.395 1.593 0.111 1.529 0.328 5.237 3.567 1.960 0.473 1.297 1.246 0.660 5.652 3.192 2.985 5.662 2.342 5.164 1.351 4.800 10.28 5.026 0.472 17.41 0.244 0.618 1.478 0.924 0.182 0.736 1.116 0.613 2.731 0.606 0.130 1.456 0.104 0.862 0.033 0.184 0.039 0.035 0.028 0.055 0.020 0.023 0.018 0.016 0.007
s at 823 K to 62.2 s at 773 K. The flow rates of H2O2/ H2O and A1248/MeOH liquid solutions are varied from 5.0 to 10.0 mL/min and from 0.1 to 0.2 mL/min, respectively. When combined with the two reactors, these flow rates result in a significant overlap in residence times for each isotherm, which permits reproducibility of the data to be verified. The experiments are conducted isothermally at four different temperatures: 723, 748, 773, and 823 K. Since oxygen and organic reactants exist in a single supercritical phase, the reaction rate may be considered independent of the oxygen concentration if the latter is in excess. A reaction order of zero for oxygen was reported in most publications. Experiments using many different [H2O2]/[H2O] concentrations previously,26 and a few in the present work (5-12 and 6-10 wt %, respectively) show that conversion of PCBs/MeOH is
586
Ind. Eng. Chem. Res., Vol. 39, No. 3, 2000
Table 2. SCWO of A1248 (5245 ppm in MeOH): Global Molar Conversions at 25.3 MPa (CA,0 ) 5.75 × 10-5 mol/L; CMeOH,0 ) 0.079 mol/L; CO2,0 ) 0.144 mol/L; CH2O,0 ) 8.82 mol/L) τ (s)
XA
(XA)cal
RD (%) τ (s)
XA
(XA)cal
RD (%)
0.6691 0.7205 0.7724 0.8111
0.6738 0.7230 0.7632 0.8092
0.70 0.35 -1.19 -0.24
6.29 8.39 12.6 16.8
0.3606 0.4163 0.4998 0.5685
0.3496 0.4161 0.5143 0.5831
T ) 723 K -3.05 25.2 -0.05 32.2 2.90 40.2 2.57 53.7
5.65 7.53 11.3 15.1
0.5721 0.6403 0.7155 0.7709
0.5700 0.6361 0.7205 0.7725
T ) 748 K -0.37 22.6 -0.66 28.9 0.70 36.1 0.21 48.2
0.8303 0.8647 0.8899 0.9153
0.8325 0.8623 0.8852 0.9099
0.27 -0.28 -0.53 -0.59
5.18 6.91 10.4 10.4 10.4 13.8
0.7686 0.8184 0.8565 0.8546 0.8624 0.8830
0.7549 0.8019 0.8561 0.8561 0.8561 0.8857
T ) 773 K -1.78 17.8 -2.02 20.7 -0.05 24.8 0.17 31.1 -0.73 41.4 0.31 62.2
0.9112 0.9198 0.9341 0.9431 0.9547 0.9674
0.9077 0.9188 0.9305 0.9429 0.9556 0.9690
-0.39 -0.11 -0.39 -0.02 0.09 0.17
4.53 6.04 9.07 9.07 12.1 18.1
0.9059 0.9381 0.9548 0.9546 0.9695 0.9773
0.9338 0.9485 0.9640 0.9640 0.9722 0.9806
T ) 823 K 3.08 21.8 1.10 27.2 0.96 36.3 0.98 36.3 0.27 54.4 0.34 54.4
0.9791 0.9829 0.9967 0.9938 0.9995 0.9985
0.9836 0.9866 0.9897 0.9897 0.9928 0.9928
0.46 0.37 -0.71 -0.42 -0.67 -0.57
independent of initial oxygen concentration ranging from a stoichiometric molar ratio of 1.5 to an excess molar ratio of 3.0 to oxidize the bulk methanol to CO2. Thus, the reaction rate can be assumed to be independent of O2 concentration over the above concentration range. Accordingly, most of the experiments are conducted at a nominal [H2O2]/[H2O] concentration of 6 wt %, providing an excess of O2 of approximately 20 mol %. Initial O2 concentration is 0.144 mol/L or 1.59% at the reaction conditions. It is very important that the H2O2 be completely converted to O2 and H2O to ensure the reliability of the experimental results.25,37 Using the maximum oxidizer flow rate for our experiments of 10 mL/min, the residence time in the preheating line is approximately 7 s at 673 K. This high value of the preheating time allows for a complete decomposition of H2O2 before the oxidizer and organics are mixed.37 For lower flow rates and higher temperatures of the experiments, the H2O2 decomposition has better conditions to be complete. If occurring, the thermal decomposition of A1248 is included in the global kinetics of SCWO. The experimental results of the oxidation of A1248 in supercritical water at 25.3 MPa and 723, 748, 773, and 823 K for initial A1248 feed concentrations of 5.75 × 10-5 mol/L (reaction conditions) are included in Table 2 and plotted versus residence time in Figure 2. The second column of the Table 2 displays the molar conversion of A1248 (XA) expressed as the ratio of the reacted to the initial amount (moles) charged for steadystate run time, ∆t:
XA ) [A1248]reacted/[A1248]0 ) ([A1248]0 [A1248]τ)/[A1248]0 ) 1 - V∆tΣ(Ci/Mi)/V0Σ(Ci,0/Mi) (3) Here V∆t is the measured steady-state effluent liquid volume and V0 is the volume of the A1248/MeOH solution fed during the steady-state run time, ∆t. The PCB congener mass concentrations in the initial A1248/ MeOH solution and the condensed effluent aqueous
Figure 2. A1248 conversion versus residence time: s, calculated by eq 7; experimental (2, 723 K; 9, 748 K; b, 773 K, [, 823 K). CA,0 ) 5.75 × 10-5 mol/L; CMeOH,0 ) 0.079 mol/L; CO2,0 ) 0.144 mol/L; CH2O,0 ) 8.82 mol/L.
solutions are Ci,0 and Ci, respectively, and Mi represents the molar mass of individual PCB congeners. Some of the conversions of A1248 are obtained from replicate runs with standard deviations less than 0.53%. Overall, conversions range from 36.06% (6.29 s, 723 K) to 99.95% (54.4 s, 823 K) depending on residence time and temperature. The major products of the SCWO of A1248/MeOH are biphenyl, numerous low chlorinated PCB congeners such as 2-chlorobiphenyl (2-CB) and 2,2′-dichlorobiphenyl (2,2′-DCB), CO, and CO2. Some small amounts of CH4 are obtained at high temperatures (773 and 823 K) and short residence times. Biphenyl, CO, and CO2 were considered among the major products of A1248/ MeOH SCWO by their area peaks in the chromatograms (GC/FID/MSD and GC/TCD, respectively). A preliminary investigation has been conducted to identify the oxidation reaction byproducts of A1248. For this purpose, a GC/MS technique and NIST library of spectra are employed to analyze the effluent liquid-phase samples. The results positively show biphenyl as a major byproduct and several benzene derivatives for lower temperatures and residence time values. However, there are no dioxins among them at the level of GC/MS limit of detection (∼0.1 ppb). The positive identification and quantification of these compounds by further concentration of the hexane extracted of aqueous solutions have not been executed at this stage of the research and need to be further investigated. In the experiments with A1248/MeOH, as the bulk amount of methanol is oxidized (99.35% MeOH vs 0.65% A1248), it is difficult to obtain a carbon mass balance for A1248 because of analytical accuracy. Further, another reason we cannot obtain a carbon mass balance is that some of the effluent products such as unreacted methanol,
Ind. Eng. Chem. Res., Vol. 39, No. 3, 2000 587
Figure 3. Experimental yields of 2-CB expressed as the ratio of number of moles of 2-CB in the effluent stream (n1) to total number of moles of A1248 (n0) charged for steady-state run time, ∆t. (9, 723 K; b, 448 K; 2, 773 K; [, 823 K.)
formaldehyde, CO, CH4, and/or organic acids were not measured. Nevertheless, we have verified the reliability of our technique with our study on methanol oxidation in supercritical water, in which the carbon balances were closed within analytical accuracy.
The global reaction network for SCWO of A1248 based on our experimental data can be written in the broadest sense as consecutive dechlorination and oxidation reactions:
C12H10-mClm f ‚‚‚ f C12H10-m+nClm-n f ‚‚‚ f C12H10 f 6CO + 6CO2 + 5H2O (4) Chlorinated compounds typically have chlorine atoms removed under SCWO conditions. The global kinetics for SCWO of A1248 may be conveniently examined by assuming that the global rate of this reaction network (-rA) is proportional to the A1248 concentration in the reactor at a given time (5.75 × 10-5 mol/L at τ ) 0) and independent of the water, methanol, and O2 concentrations (8.82 mol/L or 97.53%, 0.079 mol/L or 0.87%, and 0.144 mol/L or 1.59%, respectively, at τ ) 0). The reaction of A1248 is assumed to be independent of the methanol concentration on the basis of two reasons: (a) methanol is in large excess compared to A1248 (99.5 wt %); (b) no methylated byproducts were detected by GC/ ECD, GC/FID, and GC/MSD analytical techniques. Therefore, the reaction rate can be written
-rA ) -dCA/dt ) kCA
k ) A exp(-Ea/RT)
(6)
To examine a power-low dependency of the conversion, the integrated form of eq 5 can be obtained:
4. Global Kinetics of Aroclor 1248 Oxidation
R
Figure 4. Experimental yields of PCB congeners at 723 K for which the production in the first step of the SCWO process exceeds the oxidation rate. (ni ) number of moles of each congener in the effluent stream for steady-state run time, ∆t; ni,0 ) number of moles of each congener charged to the reactor over ∆t.) The number designation for each symbol represents the GC peak number in Table 1 and associated PCB congener(s).
(5)
Here, the global reaction rate constant can be expressed in terms of Arrhenius frequency factor, A, and the energy of activation, Ea:
XA ) 1 - (1 + (R - 1)10b exp(-Ea/RT) [A1248]R-1τ)1/1-R (7) Here 10b exp(-Ea/RT) represents the rate constant (k) with the Arrhenius frequency factor (A ) 10b) and the energy of activation (Ea) as fitting parameters along with the order of the reaction (R). These parameters are determined over all temperatures by best fit of the experimental data. The regression technique is performed with the Microcal Origin.6.0 software package by minimizing the objective function defined as the sum of squares of the difference between the experimental and predicted conversions (eq 7). The value of R is found to be very close to a second-order reaction value: 2.09 ( 0.02. The values of the Arrhenius parameters corresponding to k are A ) 1017.0(0.1 s-1(mol/L)-1 and Ea ) 186 ( 2 kJ/mol (44.43 ( 0.51 kcal/mol). On the basis of the above results, the overall reaction rate expression can be written
-rA ) 1017.0(0.1 exp(-(186 ( 2)/RT)[A1248]2.09(0.02 (8) Two observations can be made regarding the generality of the above rate law. (1) The rate expression is obtained by using only a constant initial concentration of A1248 (5245 ppm in methanol). However, our work on SCWO of 3,3′,4,4′-tetrachlorobiphenyl38 indicates that a much lower PCB concentration (500 ppm in methanol) does not significantly change the rate-law
588
Ind. Eng. Chem. Res., Vol. 39, No. 3, 2000
Figure 5. Experimental conversions versus residence time of A1248 congeners for which oxidation rate is prevalent over all residence times: (a, top left) 723 K; (b, top right) 748 K; (c, bottom) 773 K; (d, bottom right) 823 K.
expression (R ) 1.96, A ) 1018.8, and Ea ) 192 kJ/mol). (2) When working with the constant initial feed concentration of an excess of oxidant, the fitting procedure cannot meaningfully discriminate between A and [O2]β as both are constant. For β ) 0, the expression is obviously valid. If β > 0, the value of [O2]β is lumped with A.
The calculated conversion values considering the fitting parameters in eq 7 are presented in Table 2 along with relative deviations defined as
RD ) 100 × (XAcal - XAexp)/XAexp
(9)
The absolute average relative deviations (AARD )
Ind. Eng. Chem. Res., Vol. 39, No. 3, 2000 589
Figure 6. Congener pattern shift of A1248 oxidation from initial composition to composition at 4.53 s reaction time and then to composition at 54.4 s reaction time at 823 K (see Table 1 for congeners associated with peak numbers). Relative abundance (%) of congeners is defined as 100× (molar fraction).
∑|RDi|/∑i) for the four sets of data are 1.38% at 723 K, 0.45% at 748 K, 0.52% at 773 K, 0.83% at 823 K, and 0.80% overall. The larger ARDs are observed for the short residence times. Good agreement between the calculated values and experimental data is observed. Congener specific analysis is made and the conversion patterns of the main components have been determined simultaneously in an attempt to explain the overall reaction trends. There is an indication of the competing phenomena of production and decomposition of intermediate byproducts with the former dominant in the earlier stage of the reaction and the latter dominant at the more advanced stages of the oxidation process. On the basis of the dominant phenomenon, two groups of congeners can be distinguished: congeners with initial concentrations increasing during the early stages of the reaction and congeners with initial concentrations continuously diminishing. Among the congeners in the first category, 2-CB is a case with special consideration. It is essentially not present in the initial composition of A1248 but is among the major components of the effluent streams. To show the production/oxidation behavior of this congener, we plot the ratio of moles of 2-CB (in the effluent stream) to the total moles of A1248 charged over steady-state run time ∆t versus residence time. Figure 3 shows a maximum in the amount of 2-CB produced by meta and para dechlorination of higher chlorinated PCBs as a function of residence time and temperature: ∼20 s at
723 K, ∼15 s at 748 K, ∼8 s at 773 K, and ∼5 s at 823 K. Other components of A1248 with generally lower IUPAC numbers (especially 2,2′-DCB and the congeners represented by the next 15 peaks in Table 1) exhibit similar trends with 2-CB at 723 K (Figure 4). We see here that the maximum 2,2′-DCB concentration reaches ∼39 times the initial feed concentration at ∼15 s reaction time. The number of these congeners gradually diminishes at higher temperatures (748 and 773 K) so that only four remain at 823 K. The amount of these congeners also diminishes with increasing temperature. There is a second set of higher chlorinated congeners with generally higher IUPAC numbers for which oxidation predominates during all reaction times at all temperatures. For these congeners, the experimental conversion values are plotted versus residence time for the four temperatures: 723 K (Figure 5a), 748 K (Figure 5b), 773 K (Figure 5c), and 823 K (Figure 5d). Because of the high density of points, these graphs show mainly the envelopes of conversions of the individual congeners with residence time as all congeners cannot be distinguished because of the overlap and a separate display of each congener is not appropriate here. Nonplanar congeners with predominant ortho Cl-substituted positions (2,2′,4,6-, 2,2′,5,6-, 2,2′,3,6-, and 2,2′,3,6-tetrachlorobiphenyls) are the lower limit and higher chlorinated congeners are the upper limit of the envelopes. These results may be explained by the production of the lowest chlorinated congeners (not significantly present in the
590
Ind. Eng. Chem. Res., Vol. 39, No. 3, 2000
initial composition of Aroclor 1248) by an incomplete dechlorination process of higher chlorinated congeners at lower residence times. In accordance with the above results, a general observation about the reaction pathway can be made. The massive production of 2-CB and 2,2′-DCB may suggest that chlorination in the ortho position for PCB congeners is the most resistant to SCWO. Similar resistance to reaction in biodechlorination is known for these types of congeners:39 2-CB and 2,2′- and 2,6-DCBs comprise over 65% of the biodechlorination products transported by waters of the Upper Hudson River.40 In contrast, Martino and Savage28-30 reported a higher reactivity for ortho-substituted phenol compared to meta and para derivatives. One possible explanation here may be due to differences in reactivity resulting from the two different coplanar/noncoplanar conformations of PCB congeners: ortho-substituted congeners exhibit a noncoplanar conformation while the non-ortho-substituted congeners exhibit a dioxin-like coplanar conformation.41 The meta and para chlorines from noncoplanar congeners appear to be removed first in the early stage of the reactions, resulting in an increase of lower ortho-chlorinated congeners. Therefore, a shift in the congener pattern of Aroclor 1248 occurs between the feed material and the product. This shift is exemplified in Figure 6. At 550 °C and 54.4 s, the highest value of 99.95% molar conversion for Aroclor 1248 feed of 5245 ppm in methanol is obtained. This conversion value is very high when one considers the remarkable chemical stability of PCB congeners in Aroclor mixtures such as A1248, as they were employed even as fire-retardant materials. Further, even higher conversions can be achieved with either increasing temperature or residence time. To determine the optimal conditions for a complete destruction of PCBs in methanol solutions, kinetic studies for a congener specific oxidation pathway are required. Knowledge regarding the effects of chlorine positions, degree of chlorination, and synergistic effects in Aroclor mixtures oxidation pathways is needed. To secure this information, studies on SCWO of PCB isomers (2-, 3-, and 4-chlorinated biphenyls) and of different chlorinated PCBs are being conducted at different temperatures and residence times. 5. Summary and Conclusions The oxidation of A1248 is studied over a temperature range of 723-823 K and residence times from 4.53 to 62.2 s. The A1248/methanol conversions are determined for an excess of oxygen of 20%. The reaction kinetics appears to be independent of the O2 concentration. The experimental results are correlated with an overall second-order kinetic model over the temperature range studied. Complex reaction pathways of this SCWO system have been observed. The global reaction network for SCWO of A1248/methanol is based on the consecutive reactions of dechlorination and oxidation with biphenyl, CO, and CO2 as ultimately the main products of the process. Competing phenomena of production and decomposition of intermediate byproducts occur with the former dominant in the earlier stage of the reaction and the latter dominant at the more advanced stages of the oxidation process. Acknowledgment The financial support from the National Institute of Environmental Health and Sciences Superfund Basic
Research Program, Grant P42 ES-04913, is acknowledged. The authors thank Dr. Anthony DeCaprio for providing the analytical calibration samples of A1248. The helpful comments of three anonymous reviewers are also acknowledged. Literature Cited (1) Savage, P. E.; Gopalan, S.; Mizan, T. I.; Martino, C. J.; Brock, E. E. Reactions at Supercritical Conditions: Applications and Fundamentals. AIChE J. 1995, 41, 1723. (2) Savage, P. E. Organic Chemical Reactions in Supercritical Water. Chem. Rev. 1999, 99, 603. (3) Modell, M. Processing methods for the oxidation of organics in supercritical water. U.S. Patent 4,543,190, 1985. (4) Modell, M. Supercritical Water Oxidation. In The Standard Handbook of Hazardous Site Treatment and Disposal; Freeman, H. M., Ed.; McGraw-Hill: New York, 1989. (5) Swallow, K. C.; Killilea, W. R.; Malinowski, K. C.; Staszak, C. N. The MODAR Process for the Destruction of Hazardous Organic Wastes-Field Test of a Pilot-Scale Unit. Site Management 1989, 9, 19. (6) Barner, H. E.; Huang, C. Y.; Johnson, T.; Martch, M. A.; Killilea, W. R. Supercritical Water Oxidation: An Emerging Technology. J. Hazard. Mater. 1992, 31, 1. (7) Tester, J. W.; Holgate, H. R.; Armellini, F. J.; Webley, P. A.; Killilea, W. R.; Hong, G. T.; Barner, H. E. Supercritical water oxidation technology: process development and fundamental research. In Emerging technologies in hazardous waste management III; Tedder, D. W., Pohland, F. G., Eds.; ACS Symposium Series S18; American Chemical Society: Washington, DC, 1993; p 35. (8) McBrayer, R. N. Design and operation of the first commercial supercritical water oxidation facility. First International Workshop on Supercritical Water Oxidation, Jacksonville, FL, 1995. (9) Modell, M.; Mayr, S.; Kemna, A. Supercritical water oxidation of aqueous wastes. In Official Proceedings of the 56th International Water Conference, 1995; p 479. (10) Li, R.; Savage, P. E.; Szmukler, D. 2- Chlorophenol Oxidation in Supercritical Water: Global Kinetics and Reaction Products. AIChE J. 1993, 39 (1), 178. (11) Lin, K. S.; Wang, H. P.; Li, M. C. Oxidation of 2,4Dichlorophenol in Supercritical Water. Chemosphere 1998, 36 (9), 2075. (12) Yang, H. H.; Eckert, C. A. Homogeneous Catalysis in the Oxidation of p-Chlorophenol in Supercritical Water. Ind. Eng. Chem. Res. 1988, 27, 2009. (13) Jin, L.; Shah, Y. T.; Abraham, M. A. The Effect of Supercritical Water on the Catalytic Oxidation of 1,4-Dichlorobenzene. J. Supercrit. Fluids 1990, 3 (4), 233. (14) Hatakeda, K.; Ikushima, Y.; Ito, S.; Saito, N.; Sato, O. Treatment of Chlorinated Aromatic Compounds Using Supercritical Water. In Proceedings of the International Congress of Pacific Chemical Societies; The Chemical Society of Japan: Sendai, Japan, 1995; Vol. 10, p 481. (15) Lee, D. S.; Gloyna, E. F.; Li, L. Efficiency of Hydrogen Peroxide and Oxygen in Supercritical Water Oxidation of 2,4Dichlorophenol and Acetic Acid. J. Supercrit. Fluids 1990, 3 (4), 249. (16) Houser, J. T.; Liu, X. Reactions of 1-Chloro-3-phenylpropane, 2-Chlorotoluene, and 4-Chlorophenol in Supercritical Water. J. Supercrit. Fluids 1996, 9 (3), 167. (17) Oe, T. Waste Water Treatment by Supercritical Water Oxidation Kami Pa Gikyoshy 1998, 52 (8), 1056. (18) Hatakeda, K.; Ikushima, Y.; Ito, S.; Saito, N.; Sato, O. Supercritical Water Oxidation of a PCB of 3-Chlorobiphenyl Using Hydrogen Peroxide. Chem. Lett. 1997, 245. (19) Hatakeda, K.; Ikushima, Y.; Sato, O.; Aizawa, T.; Saito, N. Supercritical Water Oxidation of Polychlorinated Biphenyls Using Hydrogen Peroxide. Chem. Eng. Sci. 1999, 54, 3079. (20) Aki, S. N. V. K.; Abraham, M. A. An Economic Evaluation of Catalytic Supercritical Water Oxidation: Comparison with Alternative Waste Treatment Technologies. Environ. Prog. 1998, 17 (4), 246. (21) Gloyna, E. F.; Li, L.; McBrayer, R. N. Engineering Aspects of Supercritical Water Oxidation. Water Sci. Technol. 1994, 30, 1.
Ind. Eng. Chem. Res., Vol. 39, No. 3, 2000 591 (22) Tester, J. W.; Webley, P. A.; Holgate, H. R. Revised Global Kinetic Measurement of Methanol Oxidation in Supercritical Water. Ind. Eng. Chem. Res. 1993, 32, 236. (23) Brock, E. E.; Oshima, Y.; Savage, P. E.; Barker, J. R. Kinetics and Mechanism of Methanol Oxidation in Supercritical Water. J. Phys. Chem. 1996, 100, 15834. (24) Alkam, M. K.; Pai, V. M.; Butler, P. B.; Pitz, W. J. Methanol and Hydrogen Oxidation Kinetics in Water at Supercritical States. Combust. Flame 1996, 106, 110. (25) Rice, S. F.; Hunter, T. B.; Ryden, A. C.; Hanush, R. G. Raman Spectroscopic Measurement of Oxidation in Supercritical Water. 1. Conversion of Methanol to Formaldehyde. Ind. Eng. Chem. Res. 1996, 35, 2161. (26) Anitescu, G.; Zhang, Z.; Tavlarides, L. L. A Kinetic Study of Methanol Oxidation in Supercritical Water. Ind. Eng. Chem. Res. 1999, 38 (6), 2231. (27) Gopalan, S.; Savage, P. E. A Reaction Network Model for Phenol Oxidation in Supercritical Water. AIChE J. 1995, 41 (4), 1864. (28) Martino, C. J.; Savage, P. E. Supercritical Water Oxidation Kinetics, Products, and Pathways for CH3- and CHO-Substituted Phenols. Ind. Eng. Chem. Res. 1997, 36, 1391. (29) Martino, C. J.; Savage, P. E. Supercritical Water Oxidation Kinetics and Pathways for Ethylphenols, Hydroxyacetophenones, and Other Monosubstituted Phenols. Ind. Eng. Chem. Res. 1999, 38, 1775. (30) Martino, C. J.; Savage, P. E. Oxidation and Thermolysis of Methoxy-, Nitro-, and Hydroxy-Substituted Phenols in Supercritical Water. Ind. Eng. Chem. Res. 1999, 38, 1784. (31) Brock, E. E.; Savage, P. E. A Detailed Chemical Kinetics Model for Supercritical Water Oxidation of C1 Compounds and Hydrogen. AIChE J. 1995, 41, 1874. (32) Steeper, R. R.; Rice, S. F.; Kennedy, I. M.; Aiken, J. D. Kinetics Measurements of Methane Oxidation in Supercritical Water. J. Phys. Chem. 1996, 100, 184.
(33) Bianchetta, S.; Li, L.; Gloyna, E. F. Supercritical Water Oxidation of Methylphosphonic Acid. Ind. Eng. Chem. Res. 1999, 38, 2902. (34) DeCaprio, A. The State University of New York at Albany, School of Public Health, personal communication, 1999. (35) Haar, L.; Gallagher, J. S.; Kell, G. S. NBS/NRC Steam Tables: Thermodynamic and Transport Properties and Computer Programs for Vapor and Liquid States of Water in SI Units; Hemisphere: New York, 1984. (36) Erickson, M. D. Analytical Chemistry of PCBs; Lewis: Chelsea, MI, 1992. (37) Croiset, E.; Rice, S. F.; Hanush, R. G. Hydrogen Peroxide Decomposition in Supercritical Water. AIChE J. 1997, 43, 2343. (38) Anitescu, G.; Tavlarides, L. L. Supercritical Water Oxidation (SCWO) of Individual Polychlorinated Biphenyl (PCB) Congeners and Aroclor 1248 PCB Mixture in Methanol Solutions. AIChE Annual Meeting, Dallas, TX, November 1999; Paper 65f. (39) Rhee, G. Y.; Sokol, R. C.; Bethoney, C. M.; Bush, B. Dechlorination of PCBs by Hudson River Sediment Organisms: Specificity to the Chlorination Pattern of Congeners, Environ. Sci. Technol. 1993, 27, 1190. (40) Bush, B.; Simpson, K.; Shane, L., Koblintz, R. PCB congener analysis of water and Caddisfly larvae (Insecta Trichoptera) in the Upper Hudson River by glass capillary chromatography. Bull. Environ. Contam. Toxicol. 1985, 34, 96. (41) Carpenter, D. O. PCBs: what they are, how they behave, and uncertainties that exist. Improved Communications workshop: Remediation of PCB-Contaminated Sites in New York and Massachusetts; School of Public Health: Albany, NY, October 1998.
Received for review September 23, 1999 Revised manuscript received December 6, 1999 Accepted December 8, 1999 IE990704L
