Oxidation of Biphenyl in Supercritical Water: Reaction Kinetics, Key

---

1226

Ind. Eng. Chem. Res. 2005, 44, 1226-1232

Oxidation of Biphenyl in Supercritical Water: Reaction Kinetics, Key Pathways, and Main Products Gheorghe Anitescu and Lawrence L. Tavlarides* Department of Chemical Engineering and Materials Science, Syracuse University, Syracuse, New York 13244

We have shown that biphenyl, the central structural block in polychlorinated biphenyl (PCB) molecules, is also a key component of PCB reaction networks in supercritical water (SCW). This study provides information on biphenyl thermolysis (SCWT) and oxidation (SCWO) in SCW, proves the high chemical stability of biphenyl to SCWT/SCWO conditions, and exposes the complexity of the opening-ring reactions involved in overall PCB reaction networks. The decomposition reactions of biphenyl are investigated at 25 MPa and 673-823 K in SCW in the presence of methanol, with and without oxygen (from H2O2/H2O solutions). The experiments with biphenyl, delivered to a tubular reactor both as solutions in methanol (2-6 g/L and 5 wt %) and neat (melted) reactant, are conducted isothermally at plug flow conditions. GC-FID/ TCD/MSD chromatographic methods are employed for product analysis. Global conversions for biphenyl disappearance under SCWO conditions vary from 3% (673 K and 3.8 s) to 97% (823 K and 14.2 s) when delivered at 2-6 g/L and follow overall second-order kinetics with Arrhenius parameters A ) 1024(0.3 (mol/L)-1 s-1 and Ea ) 294 ( 5 kJ/mol. The global kinetic model developed from these data is also valid for the SCWO of biphenyl/MeOH at 5 wt %, but it fails in the case of neat biphenyl. These findings suggest an important contribution of methanol in the reaction mechanisms, products, and pathways. The main primary reaction products are all three hydroxybiphenyl isomers. However, more than 50 other minor reaction intermediates (yields less than a few tenths of a percent) have been identified under different SCWO conditions. No significant reactions are observed under SCWT conditions. 1. Introduction Research on supercritical water oxidation (SCWO) and thermolysis (SCWT) of simple organic compounds has been conducted as these model compounds are either the rate-limiting steps in the oxidation of more complex organics, simulants for hazardous waste compounds, or characteristic pollutants.1 Among these selected chemical species are methane,2-4 methanol,5-8 ethanol,9,10 benzene,11 naphthalene,12 and phenol.13,14 In a series of previous studies, our group gathered comprehensive information on supercritical fluid extraction (SCFE)15 and SCWO/SCWT of individual polychlorinated biphenyl (PCB) congeners and mixtures (e.g., Aroclor 1248).16-19 The studies devoted to PCB destruction do not provide a detailed and exhaustive analysis of the entire reaction network from a starting PCB reactant to the final mineral products. Yet, it is shown that biphenyl, the building block for PCBs, is a stable intermediate, and beyond this intermediate there is an even more complex reaction network than that upstream of this point (Figure 1).16 Accordingly, this work expands the previous research, which was focused on SCWO/SCWT reaction kinetics of PCB dechlorination, to biphenyl in methanol solutions and as a neat reactant (melted). Biphenyl is an aromatic hydrocarbon, (C6H5)2, that is naturally occurring and is a common combustion product. Physicochemical properties indicate that below 344 K, biphenyl is a solid with low volatility (boiling point of 527 K and vapor pressure of 11.9 Pa at 298 K) and is relatively insoluble in water under ambient * To whom correspondence should be addressed. Tel: (315) 443-1883. Fax: (315) 443-1243. E-mail: [email protected].

conditions (7.88 mg/L). At this time there are no reported studies on SCWO reactions of biphenyl, but it is known that biphenyl is one of the most thermally stable of all organic compounds. It is combustible at high temperatures producing carbon dioxide and water when combustion is complete. Partial combustion produces carbon monoxide, smoke, soot, and low molecular weight hydrocarbons. Similar to benzene,11,20 biphenyl is expected to be less reactive under SCWO conditions than its substituted derivatives. Because biphenyl is solid under normal conditions and has very low solubility in liquid water, it is necessary to use an organic solvent to deliver this compound to the reactor. Our group successfully used methanol to charge insoluble PCBs to the SCWO reactor,16-18 which also showed other positive effects such as enhancing the reaction rate via free radicals produced during its partial oxidation. To be consistent with the previous studies of PCB/MeOH, we also used methanol as a solvent for biphenyl. The ability of an oxygenate additive to enhance the oxidation rate of targeted organics and mixtures is well-known in combustion,21,22 but only recently have several studies showed similar effects in SCWO reactions.16,17,23,24 The majority of the SCWO experiments on harmful organics have been focused on achieving high conversion values rather than kinetic data, which require assumed incomplete conversions. These studies confirm only that SCWO can destroy these compounds and do not address the issue of the formation of harmful intermediates. Conditions under which toxic products can be formed and minimized are not well defined. For example, in the presence of methanol, the reaction pathways are significantly different when compared with neat or water delivered organics. In the former case, production of harmful compounds such as dibenzofurans and dioxins

10.1021/ie049566c CCC: $30.25 © 2005 American Chemical Society Published on Web 09/16/2004

Ind. Eng. Chem. Res., Vol. 44, No. 5, 2005 1227

Figure 1. Biphenyl importance in the PCB reaction network under SCWO conditions. (ROP stands for ring-open products).

is avoided or minimized, and, as such, this situation leads to potentially clean and environmentally safe oxidation systems operating under optimum, less severe conditions. One goal of this study is to show that significantly different effects occur with respect to reaction products formed and reaction pathways developed when conducting reactions in SCW in the presence of co-oxidants. Several experiments on the oxidation in SCW of neat biphenyl in the feed have been conducted under similar conditions of biphenyl/MeOH systems. Accordingly, this paper reports data on biphenyl conversions both neat and in the presence of methanol. In accordance with the high chemical stability of the biphenyl, the influence of methanol on its decomposition rates is expected to be significant, thus facilitating, by comparison, an easy assessment of the effect of methanol on these reactions. More studies are in progress regarding the role of a cooxidant on reaction rates, mechanisms, and pathways. 2. Experimental Section 2.1. Apparatus. The experiments were conducted in a plug flow reactor system that has been used for previous studies in our laboratory.8,16-18,23 The experimental apparatus, described in detail elsewhere,8,25 consists of three major modules: pumps and preheaters, reactor, and cooling and separation. Two reactors were used in these experiments made from 2- and 12-m tubing of Hastelloy (HC-276, 1.6-mm-i.d., 3.2 mm o.d.) and housed within a fluidized sand bath. In the pump and preheating subsystem, organic and oxidant (H2O2/ H2O) solutions are delivered in separate lines by highpressure feed pumps (ISCO 100-D). The organic feed is contercurrently mixed with supercritical oxidant in a mixing tee. A special micrometric valve is used to maintain the desired pressure within the reactor. Temperature profiles for the reactor are obtained using type-K thermocouples that are connected at the inlet, middle, and outlet of the reactor streamlines. Due to the low organic concentrations employed, temperature does not vary significantly along the reactor (max. ( 2 K). The variation of residence time in the two reactors is achieved by changing the feed rates. The cooling and separation modules consist of a special water-cooled glass separator for steady-state effluents and a coiled, air-cooled, heat-exchanger, connected to a waste collecting bottle for the nonsteady-state portion of the reaction. 2.2. Procedure. The preheated biphenyl/MeOH or the melted biphenyl and aqueous streams are pumped

in the reactor at the desired temperature, pressure, and flow rates. The temperature and pressure are controlled by the adjustable heaters located in the sand bath and by the special designed micrometric valve positioned at the end of the reactor, respectively. The stream pressure is dropped to ambient conditions, and the gaseous and liquid phases are then directed in either one of two separators connected in parallel. During the nonsteadystate portion of the reaction, the liquid products are collected directly in a waste bottle and removed occasionally while the gases are vented. The products of the steady-state reaction 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 parameters are stabilized. The conditions at the end of the reactor and in the separator were carefully established such that no significant amount of organic reaction products are lost after the steady-state effluent content is removed for sampling. These include a thorough rinsing of the steady-state separator and both inlet and outlet pipes. 2.3. Experimental Conditions. Biphenyl thermolysis and oxidation experiments are all performed under isothermal and isobaric conditions with varying concentrations of biphenyl. The main reaction parameters considered are the residence time, temperature, and the biphenyl and oxygen initial concentrations. All the biphenyl-methanol experiments are performed in pairs of replicates, and the results show an overall standard deviation of less than 9%. The residence time of the experiments ranged from 1.8 s at 823 K to 30.3 s at 673 K and was determined by the flow rates of the oxidant and organic solutions at room conditions (5.0-10.0 mL/min and 0.1-0.2 mL/ min, respectively). When combined with the two reactors used for these sets of experiments, these flow rates resulted in five-point isotherms. By assuming a constant volumetric flow rate through the reactor and based on 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 ) (FSC/FL)V/vL

(1)

The residence time is in seconds, the volume V of the reactor is in mL, the densities of the both phases are in g/mL, and the total flow rate of the liquid reactants (vL) is in mL/s. The density of water at reaction conditions

1228

Ind. Eng. Chem. Res., Vol. 44, No. 5, 2005

is calculated using NBS Steam Tables.26 The fluid mixture density is assumed to be that of the bulk water, and the assumption is valid when checked with the Peng-Robinson equation of state. Pressure and temperatures of the SCWO experiments were 25 MPa and 673, 723, 773, and 823 K, while the SCWT experiments were conducted at 673, 723, and 773 K. The SCWT investigations carried out at lower temperatures of 673 and 723 K showed that biphenyl barely decomposed under these conditions. These results can be used to calculate biphenyl mass balances, and closures of 100 ( 3% were within the error limits of the overall method ((9%). Initial biphenyl concentrations were selected at values of 2-6 g/L in methanol and resulted in different concentrations under the reaction conditions (0.02-0.13 mM). This range of concentrations was selected to match that of the previous PCB experiments in order to facilitate comparisons between the two systems. A set of experiments was conducted with 5 wt % biphenyl in methanol (3 mM in the reactor) at 773 K and higher residence times (30-60 s), while another set was conducted with 13.6 mM neat biphenyl at 723 K and short residence times (2.5 and 5.0 s). Initial oxygen concentration for most of the experiments was provided by [H2O2]/[H2O] concentration of 6 wt %, providing 20 mol % excess oxygen. A set of experiments at higher concentrations of oxidant (12 wt % [H2O2]/[H2O] or more than 100 mol % excess oxygen) was also conducted. It is important that H2O2 be completely converted to O2 and H2O to ensure the reliability of the experimental results.27 The high values of the preheating time in our experiments allowed for a complete decomposition of H2O2 before the oxidizer and organics are mixed. 2.4. Analytical Techniques. Reliable analytical procedures are essential to secure reaction kinetics and pathways information for biphenyl SCWO. Reaction products were analyzed by gas chromatography (GC) with three Hewlett-Packard 5890 series II gas chromatographs. The gaseous phase was only occasionally captured into a 250-µL sample loop and analyzed by on line GC-TCD (CROMPACK, 25 m × 0.53 mm i.d., 25 µm film thickness). A portion of the liquid-phase product was first extracted in hexanes and then diluted or concentrated prior to analysis by off line capillary GCFID/ECD and GC-MSD to identify and measure the amount of unreacted biphenyl and the reaction products. GC separation was achieved on capillary columns: Ultra 2 (25 m × 0.2 mm i.d. × 0.33 µm film thickness) for the former techniques and HP-1MS (30 m × 0.25 mm i.d. × 0.25 µm film thickness) for the latter. The GC response factors were obtained by running standard solutions containing certified concentrations of biphenyl and other major reaction intermediates. Chromatographic errors were found to be less than 5% through replicate analysis. To identify the oxidation reaction products of biphenyl, the effluent liquid-phase samples were concentrated and analyzed by GC-MSD using both NIST and Wiley libraries of spectra. For this purpose, standard solutions containing the major reaction products were also analyzed. When a compound was identified under similar analytical conditions both in the effluent samples and standard solutions, it was considered positively identified (double match of GC retention times and library spectra). For the remaining products, careful attention was given to the identification procedure such that only very high library matches (for the vast

Figure 2. Biphenyl conversions versus reaction residence time under SCWT conditions ([,673 K; 9,723 K; 2,773 K) and [O2] effect on biphenyl conversions (], 20 mol % excess O2 and O, 100 mol % excess O2) at 25 MPa, 723 K, and [biphenyl]/MeOH of 2 g/L.

majority, >80%) and expected structural relationships with positively identified reaction products were considered. The response chromatographic factors obtained from the standard solutions were used in the calculations for the positively identified products. For the other reaction products, estimated values of the response factors were determined by similarities of molecular structures with the former compounds. Products with retention times shorter than those of hexanes were not analyzed. 2.5. Reactants. Biphenyl (99% purity), methanol (99.93%), and standard solutions of the main reaction products for GC analysis were supplied by Aldrich. The oxidant was oxygen, supplied for the experiments as solutions of hydrogen peroxide of 3-15 wt % concentration prepared from 30 wt % H2O2/H2O solution (purum p.a., Fluka) by dilution with deionized water. All reactants were used without further purification. 3. Results and Discussion 3.1. Biphenyl Conversion. The conversion is expressed as the ratio of the reacted to the initial number of moles of biphenyl charged for steady-state run time, ∆t:

X ) ([biphenyl]0 - [biphenyl]τ)/[biphenyl]0 ) 1 - V∆tC/V0C0 (2) Here, V∆t is the measured steady-state effluent liquid volume, while V0 is the volume of the organic solutions (both biphenyl/MeOH and neat biphenyl) fed during the steady-state run time, ∆t. The biphenyl concentrations in the initial organic solutions and the condensed effluent aqueous solutions are C0 and C, respectively. All of the biphenyl conversions were obtained from replicate runs with calculated standard deviations less than 9%. 3.1.1. Biphenyl/MeOH Thermolysis in Supercritical Water. In SCWT experiments, the conversions of biphenyl were negligible below 773 K and 20 s (Figure 2, lower sets of points). Under these conditions a maximum of 4.6% conversion was obtained, which was in the limits of the overall errors of the experiments. However, these results permitted the reliability of our technique to be verified. The experimental conditions and the chemical thermostability of biphenyl practically inhibit any significant reactions from occurring at 673 K and permitted accurate calculations of the biphenyl

Ind. Eng. Chem. Res., Vol. 44, No. 5, 2005 1229 Table 1. Conversions of Biphenyl Oxidation in SCW at 25 MPa with 20% Molar Excess Oxygena T ) 673 K

T ) 723 K

T ) 773 K

T ) 823 K

τ (s) (X)C1 τ (s) (X)C1

(X)C2 τ (s) (X)C1

(X)C2 τ (s) (X)C1

30.0 30.0 21.6 21.6 15.1 15.1 7.6 7.6 3.8 3.8

0.548 0.538 0.482 0.456 0.397 0.385 0.309 0.292 0.229 0.218

0.922 0.931 0.899 0.907 0.888 0.873 0.817 0.801 0.728 0.737

a

0.101 0.121 0.098 0.086 0.072 0.057 0.056 0.037 0.025 0.033

19.8 19.8 14.2 14.2 9.9 9.9 5.0 5.0 2.5 2.5

0.396 0.385 0.357 0.344 0.279 0.270 0.181 0.172 0.129 0.113

16.4 16.4 11.7 11.7 8.2 8.2 4.1 4.1 2.0 2.0

0.844 0.829 0.821 0.818 0.795 0.786 0.708 0.692 0.568 0.556

14.2 14.2 10.2 10.2 7.1 7.1 3.6 3.6 1.8 1.8

0.970 0.959 0.958 0.941 0.935 0.922 0.890 0.869 0.813 0.788

C1 ) 2 g/L and C2 ) 6 g/L biphenyl/MeOH.

mole balances. Such data sets are shown in Figure 2, with all the values at 673 K being in the range of 100 ( 3%. 3.1.2. Biphenyl SCWO. The experimental results of the SCWO of biphenyl/MeOH solutions (0.02-0.13 mM) at 25 MPa and 673, 723, 773, and 823 K with 20 mol % excess oxygen are shown in Table 1. A set of data at 723 K with 100 mol % excess oxygen is shown for an easy comparison with the above similar data in Figure 2. The conversions are displayed as biphenyl disappearance calculated with eq 3. Overall, conversions range from 3 to 11% (3.8 to 30 s residence time at 673 K) and from 80 to 97% (1.8 to 14.2 s at 823 K), depending on residence time, temperature, and initial reactant concentration. The conversion increases rapidly with temperature and is above 80% at 773 K and 24.5 s. This destruction level of biphenyl is very high when one considers the remarkable chemical stability of this compound but surprisingly lower than that of PCBs.16,17 An explanation can be based on different reaction mechanisms and pathways for PCB congeners17,23,28 and biphenyl. Further discussion is provided in section 3.4. In the experiments with the above solutions of biphenyl/MeOH (molar ratio of ∼0.01), as the bulk amount of methanol is oxidized, it is difficult to obtain a carbon mass balance for biphenyl since the final oxidation gas products are undistinguishable. Further, another reason we cannot obtain a carbon mass balance is that the effluent products with small molecular masses such as unreacted methanol, formaldehyde, CO, CH4, and some organic acids were only occasionally measured. Nevertheless, we have verified the reliability of our technique with the SCWT runs where no significant reactions occur (Figure 2). For the case of 5 wt % biphenyl/MeOH (∼3 mM in the reactor) oxidized at 773 K and higher residence times, the conversions are significantly higher than those of the above systems. For example, at 30 s, the biphenyl conversion is 0.993 compared to 0.926 for biphenyl/MeOH at 0.068 mM and 16.4 s. These results and those above at lower biphenyl/MeOH concentrations strongly suggest a reaction order >1. Several experiments were conducted at 723 K with melted biphenyl to see the differences in conversion, reaction products, and pathways when compared with SCWO of biphenyl/MeOH solutions. As expected, the overall conversion has been found to be slightly higher than that of the latter system (e.g., 0.381 vs 0.301 at 5 s) due to higher initial biphenyl concentration in the reactor (13.6 mM compared to 0.082 mM). Also, for the case of neat biphenyl, the concentrations of both unreacted biphenyl and reaction products are higher in

the effluent streams than those of biphenyl/MeOH solutions under similar SCWO conditions. 3.2. Global Kinetics. Biphenyl properties are not simply the summation of the benzene properties. Due to a free rotation around the inter-ring C-C bond, this species exhibits dynamic stereochemical structures undergoing competitive, multistep reactions for which detailed reaction mechanisms have not been pursued at this time. Further, when the intermediate reaction products are numerous free radicals, it may not be possible to perform independent experiments to determine the rate law parameters. Consequently, these parameters can be deduced from changes in the distribution of the reaction products with feed conditions. Under these circumstances, the information to be obtained from these SCWO studies is global kinetics, reaction networks, and the identities and yields of the products of incomplete oxidation. For a global kinetic analysis, the dependence of the reaction rate on the initial reactant concentrations has to be established. Regarding this dependence on the [O2]0, our previous experiments with different [H2O2]/ [H2O] concentrations (0-15 wt %) show that conversion of PCBs/MeOH is only slightly dependent on initial excess oxygen employed to oxidize both PCBs and the bulk methanol to CO2. Comparative biphenyl conversions for 20 and 100% molar excess oxygen are shown in Figure 2 for two sets of SCWO data at 773 K and 2 g/L biphenyl/MeOH as the initial reactant. A possible explanation of this behavior could be that increasing oxygen concentrations up to stoichiometric amounts enhance the overall conversion by increasing the rate of free radical generation in the reaction with methanol. Additional increase in oxygen has a marginal effect as it competitively reacts with free radicals from methanol. The SCWO reactions proceeding via free radicals released from methanol oxidation are faster than direct reactions with O2, and the reaction rate can be assumed to be quasi independent of excess O2 concentration. Similar behavior has been reported for the CH4/MeOH system24 and can be extended to multicomponent organic systems where the co-oxidation of a reactive species decisively contributes to the oxidation of the more resilient compounds. Under these circumstances, the oxygen contribution is seen to be indirect for the more stable reactants, while the fast reactions with the former species provide the reactive free radicals for the latter. Regarding a potential dependence of biphenyl conversion on methanol concentration, even though it appears that the SCWO of biphenyl proceeds through free radicals from methanol oxidation, we are not able at this stage of the research to explicitly account for the effect of [MeOH]0 in the reaction rate since the reaction mechanism is still unknown. Accordingly, the global kinetics of biphenyl SCWO may be conveniently examined by assuming that the global rate of this reaction network (-ri) is proportional to the biphenyl concentration in the reactor at a given time and independent of the water, methanol, and O2 concentrations (8.8 mol/L or 97.5%, 0.08 mol/L or 0.9%, and 0.14 mol/L or 1.6%, respectively, at τ ) 0). Therefore, the reaction rate can be written as

-ri ) Ci,0dXi/dτ ) k(Ci)R

(3)

Here, the global reaction rate constant can be expressed in terms of the Arrhenius frequency factor, A, and the energy of activation, Ea:

1230

Ind. Eng. Chem. Res., Vol. 44, No. 5, 2005

k ) A‚exp(-Ea/RT)

(4)

To examine a power-low dependence of the conversion, the integrated form of eq 3 can be obtained for the case of the plug flow reactor29 and for a reaction order higher than unity:

X) 1 - (1 + (R - 1)10bexp(-Ea/RT)[biphenyl]R-1τ)1/1-R (5) Here, 10bexp(-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 > 1). These parameters are determined over all temperatures and biphenyl concentrations by fitting 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 5). The values of the fitted parameters for the case of initial [biphenyl]/MeOH of 2 and 6 g/L are R ) 2.02 ( 0.05, b ) 24.4 ( 0.3, and Ea ) 294 ( 5 kJ/mol (uncertainties are for 95% confidence intervals, not standard deviations). The partial reaction order with respect to biphenyl, R, was found to be similar to those of all of the three monochlorinated biphenyl isomers, which also have been delivered into the reactor in methanol under similar SCWO conditions.17 Similarly, the Arrhenius parameters A are also close to those of the PCB congeners (1024.1-1024.8 s-1(mol/L)-1), while the activation energy is higher (Ea ) 281-292 kJ/mol for PCBs studied),17 showing that higher conditions of temperature must be used in the case of biphenyl for comparable conversions. This is indeed what we expected based on the accumulation of biphenyl when PCBs were reacted. An observation can be made regarding the generality of the above rate-law. If the oxygen initial concentration is to be considered when working with a 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 biphenyl conversion values considering the fitting parameters in eq 5 are presented in Figure 3. A good agreement between the calculated values and experimental data is observed except for the two low residence times of each isotherm. This aspect suggests a strong dependence of the reaction rates, especially at the early stages of the reactions, on the free radicals produced by methanol oxidation along with the dependence on biphenyl initial concentrations. Unfortunately, this dependence is yet unclear, but it is obviously determined by the reaction mechanism. Most of the differences between experimental and calculated conversion values are within (5% deviation, indicating that the fitting parameters in eq 5 are appropriate to represent the SCWO process of biphenyl decomposition. Larger underprediction and overprediction of data are observed for the lowest residence times. For the two other cases of significantly higher biphenyl concentrations (5 wt % in methanol and neat), we checked if the above model is still valid. In the case of 5 wt % biphenyl/MeOH, the calculated conversions agree well with those obtained experimentally. For

Figure 3. Biphenyl conversions as a function of reaction residence time and initial concentrations under SCWO conditions. ([, 673 K; b and O, 723 K; 2 and 4, 773 K; and 9, 823 K). Solid and open symbols are for the initial biphenyl/MeOH concentrations of 2 g/L and 6 g/L, respectively. Table 2. Reaction Products of a SCWO Experiment at 25 MPa, 723 K, 5 s, and [Biphenyl]/MeOH of 6 g/L compound 1. benzaldehydea 2. benzene acetaldehyde 3. acetophenonea 4. 2-butylbenzene 5. 1-phenyl-2-propen-1-onea + 2-phenylpropenalb 6. 1-phenyl-1-propanonea 7. 2-phenylfuran 8. 3-phenylfuran 9. 2-methylbenzofuran 10. 2,3-dimethylbenzofuran 11. 1-benzoyl-1-propene 12. 3-phenyl-2-butenal 13. biphenyla 14. 1(3H)-isobenzofuranone 15. o-hydroxybiphenyla 16. dibenzofurana 17. 4-phenyl-4-cyclopentene1,3-dione 18. c 19. x-methyl-y-naphthalenol or 1-phenylcyclohexene 20. x-OH-y-naphthalenecarboxaldehyde 21. 1-phenyl-1-penten-3-one 22. x-OH-y-naphthalene carboxaldehyde 23. 4-biphenylcarboxaldehyde 24. 1/2-ethoxynaphthalene 24. m-hydroxybiphenyla 25. p-hydroxybiphenyla

retention time (min)

molecular formula

library match

106 120 120 134 132 132 134 144 144 132 146 146 146 154 134 170 168 172

C7H6O C8H8O C8H8O C10H14 C9H8O C9H8O C9H10O C10H8O C10H8O C9H8O C10H10O C10H10O C10H10O C12H10 C8H6O2 C12H10O C12H8O C11H8O2

90 73 97 68 90 83 85 83 90 84 87 79 82 90 70 95 91 88

31.43

160 158 158 172

C11H10O C11H10O C11H8O2

79-84 73 84-87

31.52 32.23

160 172

C11H12O C11H8O2

90 81-83

33.09

182

C13H10O

90

33.64 33.83 34.10

172 170 170

C12H12O C12H10O C12H10O

86 94 94

5.38 7.60 8.40 9.74 11.67 12.12 14.65 15.06 16.09 17.75 18.08 20.60 21.28 23.15 26.56 26.70 28.60 29.07 29.89

M

a Positively identified reaction products. b Coeluted. c An unknown reaction product.

example, at 773 K and 30 s the calculated conversion with eq 5 is 0.9994, while the experimental value is 0.9928. However, in the case of SCWO of biphenyl without methanol, the calculated conversions by eq 5 do not agree with those experimentally determined. For example, at 723 K and 5 s, a value of 0.975 is calculated and is far greater than that of 0.381 obtained experimentally. 3.3. Reaction Products. A thorough investigation has been conducted to identify the oxidation reaction products of biphenyl delivered both in methanol solutions and neat. The most reactive biphenyl positions are ortho and para, while the most stable to oxidation is meta, as

Ind. Eng. Chem. Res., Vol. 44, No. 5, 2005 1231

Figure 4. The yields of the biphenyl reaction major products versus residence time: (1)-673 K; (2)-723 K; (3)-773 K; and (4)823 K. (9, o-hydroxybiphenyl; 0, m-hydroxybiphenyl; O, p-hydroxybiphenyl; 2, dibenzofuran; [, acetophenone). The lines connect data sets of product yields.

expected from chemical reactivities of substituted aromatic compounds. Both positively and tentatively identified reaction products for a representative experiment are shown in Table 2 (723 K, 25 MPa, 6 g/L biphenyl/ MeOH, 5 s, 20 mol % excess O2). The temporal profiles of the main components of the effluent streams are shown in Figure 4. With oxygen fed to the reactor, the only major products in the liquid samples of the SCWO of biphenyl/MeOH are o-, m-, and p-hydroxybiphenyl and unreacted biphenyl. Other major reaction products positively identified by further concentration of the hexane extracts of aqueous solutions are several benzene derivatives such as acetophenone and dibenzofuran. The molar yields of the major reaction products strongly depend on temperature and residence time (Figure 4). The maxima exhibited by the reaction intermediates at higher temperatures prove that these compounds undergo further oxidation upon formation with different reaction rates. The temporal profiles also clearly show that the hydroxybiphenyls are primary reaction products, while the more stable acetophenone and dibenzofuran are produced by successive reactions from the former intermediates. The symmetry of the o-hydroxybiphenyl and dibenzofuran

profiles is a strong suggestion that the latter is produced from the former and subsequently destroyed. More than 50 other minor products have also been identified in much smaller amounts under various SCWO conditions. No dioxins were found at the level of GC/MS limit of detection (∼1 ppb) in the case of biphenyl/MeOH reactions. However, dibenzo-p-dioxin has been produced in SCWO reactions of neat biphenyl and qualitatively detected in the reaction products of these experiments. This result emphasizes the positive effects of methanol in avoiding the formation of harmful reaction products. This very harmful compound has also been found in the SCWO of 4-chlorobiphenyl congener when delivered in benzene to the same reactor under similar conditions.23 A possible reaction to produce dibenzo-p-dioxin is the condensation of less stable intermediates such as 2-dihydroxybenzene (pyrocatechol) and/or 2-chlorinated phenols. At higher temperatures, neat biphenyl exhibits a coking tendency. Polycyclic aromatic hydrocarbon (PAH) and dioxin formation in this case illustrates that the undesirable SCWO reaction pathways develop under these conditions. 3.4. Overall Main Reaction Pathways. In the course of investigations of both PCB and biphenyl systems, it was determined that most of the reaction products behave differently, and it is very helpful to the scientific and regulatory communities that individual product pathways be available. One of the focal points of this study is the identification and quantification of the yields of the stable products from experiments in which incomplete oxidation is programmed to occur in order to examine whether any harmful, stable products can be formed. The identification of the reaction products under different reaction conditions permits an outline of general reaction pathways for the biphenyl reaction network in this SCWO process. Figure 5 shows the key, positively identified reaction intermediates and the most likely routes of the reaction network. The oxidation of biphenyl is certainly accelerated by the oxidation of methanol through higher concentrations of reactive intermediates (e.g., OH and HO2 radicals). The quantified reaction products suggest a first step of biphenyl oxidation by OH free radicals to hydroxylated derivatives followed by keto-enol tautomerism leading to open-ring carbonylic compounds and finally to mineral products. In a future study details of these reaction pathways will be presented.

Figure 5. Principal reaction pathways in the biphenyl SCWO reaction network. All of the basic reaction intermediates are positively identified.

1232

Ind. Eng. Chem. Res., Vol. 44, No. 5, 2005

4. Summary and Conclusions This study reports the results of experiments conducted on SCWO and SCWT of biphenyl both in methanol solutions (2-6 g/L) and as a neat reactant at 25 MPa, 673-823 K, and 0, 120, and 200 mol % O2, based on the reaction stoichiometry. The SCWO results show conversions higher than 80% at residence times beyond 2 s, while under SCWT conditions the biphenyl conversion is negligible below 773 K. However, these results confirm previous data on the higher chemical stability of biphenyl compared to that of PCB congeners under similar SCWO/SCWT conditions. When delivered to the reactor in methanol solutions, the overall oxidation rates of biphenyl follow secondorder reaction kinetics with the Arrhenius parameters A ) 1024(0.3 (mol/L)-1 s-1 and Ea ) 294 ( 5 kJ/mol. Here the concentrations of oxygen and methanol are essentially constant. This model is not valid for the neat biphenyl oxidation, proving that in this case other reaction pathways occur compared to those in the presence of methanol. The reaction products for the two cases also strengthen the above conclusion. For the case of biphenyl/MeOH solutions, among the reaction products of the designed incomplete conversions (more than 50) are the three hydroxybiphenyls, dibenzofuran, open-ring products (e.g., acetophenone, benzaldehyde, 1-phenyl-2-propen1-one, 1-phenyl-1-propanone), phenol, benzene, and a few naphthalene oxygenates. For the case of neat biphenyl reactions under similar conditions, besides the above products there are dibenzofuran derivatives (e.g., dibenzofuranols), more naphthalene oxygenates and other PAHs, and even dibenzodioxin. For this case there is a strong tendency of coking, with PAH formation proving this tendency. These results indicate that optimized values of the process variables and the addition of selected reactive co-oxidants can lead to the avoidance/minimization of these harmful products. More work is in progress regarding the reaction pathways and mechanisms for biphenyl SCWO. Acknowledgment The financial support from the National Science Foundation, Grant CTS-0115650, is acknowledged. Literature Cited (1) Rice, S. F.; Steeper, R. R. Oxidation rates of common organic compounds in supercritical water. J. Hazard. Mater. 1998, 59, 261. (2) Webley, P. A.; Tester, J. W. Fundamental Kinetics of Methane Oxidation in Supercritical Water. Energy Fuels 1991, 5, 411. (3) 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. (4) Savage, P. E.; Yu, J.; Stylski, N.; Brock, E. E. Kinetics and mechanism of methane oxidation in supercritical water. J. Supercrit. Fluids 1998, 12, 141. (5) 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. (6) 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. (7) 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.

(8) Anitescu, G., Zhang, Z.; Tavlarides, L. L. A Kinetic Study of Methanol Oxidation in Supercritical Water. Ind. Eng. Chem. Res. 1999, 38, 2231. (9) Schanzenbacher, J.; Taylor, J. D.; Tester, J. W. Ethanol oxidation and hydrolysis rates in supercritical water. J. Supercrit. Fluids 2002, 22, 139. (10) Rice, S. F., Croiset, E. Oxidation of Simple Alcohols in Supercritical Water III. Formation of Intermediates from Ethanol. Ind. Eng. Chem. Res. 2001, 40, 86. (11) DiNaro, J. L.; Tester, J. W.; Howard, J. B.; Swallow, K. C. Experimental Measurements of Benzene Oxidation in Supercritical Water. AIChE J. 2000, 46 (11), 2274. (12) Xu, S.; Fang, Z.; Kozinski, J. A. Oxidation of naphthalene in supercritical water up to 420 °C and 30 MPa. Comb. Sci. Technol. 2003, 175, 291. (13) Gopalan, S.; Savage, P. E. A Reaction Network Model for Phenol Oxidation in Supercritical Water. AIChE J. 1995, 41, 1(4), 1864. (14) Krajnc, M.; Levec, J. On the Kinetics of Phenol Oxidation in Supercritical Water. AIChE J. 1996, 42, 1977. (15) Zhou, W.; Anitescu, G.; Tavlarides, L. L. Desorption of Polychlorinated Biphenyls from Contaminated St. Lawrence River Sediments with Supercritical Fluids. Ind. Eng. Chem. Res. 2004, 43 (2), 397. (16) Anitescu, G.; Tavlarides, L. L. Methanol as a Cosolvent and Rate-Enhancer for the Oxidation Kinetics of 3,3′,4,4′-Tetrachlorobiphenyl Decomposition in Supercritical Water. Ind. Eng. Chem. Res. 2002, 41, 9. (17) Anitescu, G.; Munteanu, V.; Tavlarides, L. L. Decomposition of Monochlorinated Biphenyl Isomers in Supercritical Water in the Presence of Methanol with and without Oxygen: A Global Kinetic Study. AIChE J. 2004, 50 (7). (18) Anitescu, G.; Tavlarides, L. L. Oxidation of Aroclor 1248 in Supercritical Water: A Global Kinetic Study. Ind. Eng. Chem. Res. 2000, 39, 583. (19) Zhou, W.; Anitescu, G.; Rice, P. A.; Tavlarides, L. L. Supercritical Fluid Extraction-Oxidation Technology to Remediate PCB Contaminated Soils/Sediments: An Economic Analysis. Environ. Prog. 2004, in press. (20) DiNaro, J. L.; Howard, J. B.; Green, W. H.; Tester, J. W.; Bozzelli, J. W. Elementary Reaction Mechanism for Benzene Oxidation in Supercritical Water. J. Phys. Chem. 2000, 104, 10576. (21) Inal, F.; Senkan, S. M. Effects of Oxygenate Additives on Polycyclic Aromatic Hydrocarbons (PAHs) and Soot Formation. Comb. Sci. Technol. 2002, 174 (9), 1. (22) Choi, C. Y.; Reitz, R. D. An experimental study on the effects of oxygenated fuel blends and multiple injection strategies on DI diesel engine emissions, Fuel 1999, 78, 1303. (23) Anitescu, G.; Munteanu, V.; Tavlarides, L. L. Co-Oxidation Effects of Methanol and Benzene on the Decomposition of 4-Chlorobiphenyl in Supercritical Water. J. Supercrit. Fluids 2004, in press. (24) Savage, P. E.; Rovira, J.; Stylski, N.; Martino, C. J. Oxidation kinetics for CH4/methanol mixtures in supercritical water. J. Supercrit. Fluids 2000, 17, 155. (25) Zhang, Z. Supercritical water oxidation of 4-chlorobiphenyl: reaction kinetics, destruction efficiency, and byproducts, Ph.D. Dissertation, The Department of Chemical Engineering and Materials Sciences, Syracuse University, Syracuse, NY, 1998. (26) 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. (27) Croiset, E., Rice, S. F.; Hanush, R. G. Hydrogen Peroxide Decomposition in Supercritical Water. AIChE J. 1997, 43, 2343. (28) Cioslowski, J.; Liu, G.; Moncrieff, D. Energetics of the Homolytic C-H and C-Cl Bond Cleavages in Polychlorobenzenes: The Role of Electronic and Steric Effects. J. Phys. Chem. 1997, 101, 957. (29) Fogler, H. S. Elements of Chemical Reaction Engineering; Prentice-Hall: NJ, 1999.

Received for review May 20, 2004 Revised manuscript received July 30, 2004 Accepted August 6, 2004 IE049566C