Kinetics and Products from o-Cresol Oxidation in Supercritical Water


Kinetics and Products from o-Cresol Oxidation in Supercritical Waterhttps://pubs.acs.org/doi/pdf/10.1021/ie00045a003Simi...

2 downloads 69 Views 1MB Size

Ind. Eng. Chem. Res. 1996,34, 1941-1951

1941

Kinetics and Products from o-Cresol Oxidation in Supercritical Christopher J. Martino, Phillip E. Savage,* and John Kasiborski Department of Chemical Engineering, The University of Michigan, Ann Arbor, Michigan 48109-2136

Dilute aqueous solutions of o-cresol(2-methylphenol)were oxidized in a tubular flow reactor a t near-critical and supercritical conditions. The power-law rate expression that best correlates the kinetics of o-cresol disappearance is rate = 105.7exp(-29700/RT)[o-creso11°~57[0210~22[H~011~44. The power-law rate expression that best correlates the experimental results for the conversion of organic carbon to C02 is rate = 106.8exp(-34000/RT)[TOClo~34[O~lo~'3[H~O11~18. All concentrations are in moles per liter, the activation energy is in calories per mole, and the rate is in moles per liter per second. The most abundant products from o-cresol oxidation were typically phenol, 2-hydroxybenzaldehyde, 1,3-benzodioxole, indanone, CO, and C02. 2-Hydroxybenzaldehyde was the major primary product. A reanalysis of published kinetics data for the oxidation of two other ring-containing compounds (pyridine and 4-chlorophenol) in supercritical water revealed that the rate laws previously reported for these two compounds do not provide the best correlation of the experimental data. We report the new rate laws, which are similar to those for o-cresol, 2-chlorophenol, and phenol in that the global reaction orders are between 0.55 and 0.9 for the organic compounds and between 0.2 and 0.5 for oxygen.

Introduction Supercritical water oxidation (SCWO)is an emerging technology for the ultimate destruction of organic wastes. Organic compounds and oxygen can be intimately mixed in a single homogeneous aqueous phase at supercritical conditions (T,= 374 "C, P , = 218 atm). Thus, the rapid oxidation reactions are unhindered by interphase transport limitations that could occur at subcritical conditions where multiple phases exist. Savage et al. (1995) describe current research into SCWO reactions in their review of reactions a t supercritical conditions. The rational design, optimization, control, and analysis of SCWO processes requires a knowledge of SCWO kinetics and potential byproducts formed from the oxidation of real pollutants. Our research group has focused on the oxidation of phenolic compounds (Thornton and Savage, 1992a,b;Li et al., 1992, 1993; Gopalan and Savage, 19951, and in this paper we report the kinetics and products formed from SCWO of o-cresol(2methylphenol). o-Cresol can be found in wastewaters from coal-conversion processes (Jevtitch and Bhattacharyya, 1986) and from the production of phenolic resins and pesticides. Aside from our group's previous work, the only other reports of SCWO global rate laws, products, or pathways for aromatic or cyclic compounds are for 4-chlorophenol (Yang and Eckert, 19881, phenol (Wightman, 1981; Rice et al., 19931, and pyridine (Crain et al., 1993).

Experimental Section The reactor used in this study of o-cresol SCWO nominally operated isothermally, isobarically, and in plug flow (Thornton, 1991). A description of the reactor and analytical methods is given below. Reactor Assembly and Operation. Aqueous solutions of 0 2 and o-cresol were prepared separately and used as the reactor feed streams. The 0 2 solution was prepared by saturating 3 L of distilled, deionized,

* Corresponding author. E-mail:

[email protected].

0888-588519512634-1941$09.00/0

degassed water a t 10-20 "C with oxygen at pressures of 600-1900 psi in a Parr Model 4551 4 L stirred pressure vessel. One to two liters of aqueous o-cresol solutions ranging in concentration from 100 t o 1000 mg of organic compound per kg of water (ppmw) were loaded in a separate 1gal pressure vessel and blanketed with 500 psi helium. The o-cresol was obtained from Aldrich Chemical Co. in a nominal purity of 99% and used as received. The major impurities we detected were phenol (0.25 mol %) and the mlp-cresol isomers (0.33 mol %). The feed streams were pressurized and pumped through the reactor using two Eldex Model AA100-S micrometer pumps. The o-cresol and oxygen streams were preheated in separate coils of 2 and 4 m lengths, respectively, of 0.0625 in. (1.59 mm) outer diameter (0.d.) by 0.010 in. (0.25 mm) wall thickness (WT) Hastelloy C-276 tubing. These preheater lines meet in a mixing union, where the temperature was measured by a thermocouple and the mixed feed streams enter the 4 m long by 0.125 in. (3.18 mm) 0.d. by 0.035 in. (0.89 mm) WT Hastelloy reactor. For experiments wherein low residence times were desired, a similar reactor assembly, consisting of 2 m long preheaters and a 1m long reactor tube, was used. The preheater lines, mixing tee, and reactor coil are housed in an isothermal Techne Model SBL-2 fluidized aluminum oxide bath equipped with a Techne Model TC-8D temperature controller. The reactor effluent was cooled in two consecutive tube-in-tube heat exchangers and decompressed in a Tescom Model 44 back pressure regulator. The exiting stream was separated into gas and liquid phases (at ambient conditions) in a liquid trap. The gas stream was sent t o an on-line gas chromatograph (GC) equipped with a thermal conductivity detector (TCD),and the gas flow rate was measured with a bubble meter at the outlet of the system. The liquid flow rate was measured, and samples of the liquid phase were retained for analysis. Measuring these flow rates when only the 02/ HzO stream was flowing through the reactor and then again when both streams were flowing through the reactor allowed us to calculate the 0 2 concentration and the o-cresol concentration at the reactor entrance at 0 1995 American Chemical Society

1942 Ind. Eng. Chem. Res., Vol. 34, No. 6, 1995

reaction conditions. These measured flow rates, along with the density of the reaction mixture, were also used t o calculate the reactor residence time. Because the typical reaction mixture contained over 99 mol %water, the density of the mixture was taken to be that of pure water a t the system temperature and pressure. Analytical Chemistry. The gaseous reactor effluent was analyzed with a Hewlett-Packard Model 5890 Series I1 GC-TCD fitted with a 10 ft x 118 in. 0.d. stainless steel column packed with 100/120 mesh Supelco Carboseive S-11. A 10-port Valco valve injected a 0.5 mL sample into the column, and helium flowing at 20 mumin served as the carrier gas. The initial oven temperature was 35 "C for 7 min, and then the temperature increased 16 "C/min to 225 "C. Detector response factors for CO and C02 were determined experimentally. A reverse-phase high-performance liquid chromatograph (HPLC) with a Supelco LC-18 column was used to determine the concentration of o-cresol and other phenolics in the liquid effluent samples. Analyses were performed isocratically with a mobile phase of water/ acetonitrile in a 5:2 v/v ratio flowing at 1mumin. The UV absorbance a t a wavelength of 210 nm was monitored by a Spectra Physics detector, and detector response factors were determined experimentally. Before additional products in the liquid phase were analyzed, the samples were concentrated because many products were present in low concentrations. Twenty milliliters of the reactor effluent was extracted with three successive 10 mL aliquots of dichloromethane.The 30 mL volume of this organic phase was subsequently reduced to 1mL using a Kuderna-Danish concentrator in a water bath at 50 "C. Ten microliters of a dichloromethane solution containing l-naphthol was added to each sample prior to concentration. The added l-naphtho1 (about 0.170 mg) served as an internal standard during the GC analyses. Reaction products in these concentrated samples were identified with a HewlettPackard Model 5890 GC equipped with a Model 5970 mass selective detector. The reaction products were quantified with a Hewlett-Packard Model 5890 GC with a flame ionization detector (FID). When a suspected reaction product was available commercially, we positively identified that product by matching both the mass spectrum and retention time with those of the authentic sample. The FID response factor was then determined experimentally for these compounds. Other suspected products, for which the authentic compound was not available commercially, were tentatively identified by inspecting the mass spectra and matching them to spectra stored in the gas chromatography-mass spectrometry (GC-MS) computer database. The response factors for these products were assumed to be equal to those determined experimentally for chemically similar compounds. Both GCs were equipped with identical 12 m x 0.2 mm 0.d. HP-1 capillary columns and operated in an identical fashion. Helium served as the carrier gas, and 1pL of the sample was injected in the splitless mode. The oven temperature program we used had an initial temperature of 35 "C held for 5 min, a temperature increase of 4 "C/min to 120 "C, a second increase of 2 "C/min to 160 "C, and a final increase of 10 "C/min to 250 "C. The final temperature was maintained for an additional 5 min. The o-cresol conversion (X)was calculated as X = 1 - [o-cresolY[o-cresollo,where the subscript 0 refers to the concentration at the reactor entrance at reaction

conditions. The molar yields of liquid-phase organic products were calculated as the molar flow rate of a given product in the effluent divided by the molar flow rate of o-cresol into the reactor. The molar yields of CO and C02 were calculated similarly, but were normalized by 1/7 to reflect the stoichiometry for the formation of these single-carbon compounds from o-cresol, which has seven carbon atoms. The gas stream was assumed t o be in equilibrium with the liquid stream as the two phases were separated. Thus, molar yield calculations accounted for the liquid phase being saturated with CO and COZ. Additional details about the equipment and procedures used in the present study of o-cresol SCWO are available in the accounts of our group's previous studies of phenol (Thornton and Savage; 1990; Thornton et al., 1991; Gopalan and Savage, 1995) and 2-chlorophenol (Li et al., 1993).

Experimental Results Fifty-one oxidation experiments and three pyrolysis experiments were performed for o-cresol in the flow reactor. These included experiments in the near-critical as well as the supercritical regime. The reaction conditions ranged from pressures of 200 to 300 atm, temperatures of 350 to 500 "C, residence times of 0.5 to 46.3 s, and initial o-cresol concentrations of 2.2 x to 5.9 x M. Oxygen was always present in at least 100% stoichiometric excess for the oxidation experiments. Oxygen was largely excluded for the pyrolysis experiments, which resulted in low o-cresol conversions (