Environ. Sci. Technol. 2005, 39, 5427-5435
High-Temperature Liquid Water: A Viable Medium for Terephthalic Acid Synthesis JENNIFER B. DUNN† AND PHILLIP E. SAVAGE* Department of Chemical Engineering, University of Michigan, Ann Arbor, Michigan 48109-2136
We report new information concerning the effect of oxygen concentration and catalyst concentration and identity [MnBr2, CoBr2, ZrBr4, and Mn(OAc)2] on the partial oxidation of p-xylene in dense water from 250 to 380 °C. Water is a more environmentally benign replacement for the acetic acid solvent used commercially. We used a 440 mL Hastelloy batch reactor for all experiments and monitored O2 consumption and product (including COx) formation. Increasing the catalyst concentration at 300 °C significantly increased terephthalic acid yields. MnBr2 was the most active catalyst of those we assessed. Increasing the initial O2 concentration beyond a modest excess did not significantly alter the terephthalic acid yield, but it increased the CO2 yield. Injecting supplemental O2 midreaction, however, did cause the terephthalic acid yield to increase. The highest terephthalic acid yields (>80%) occurred at 300 °C, [p-xylene]0 ) 0.02 M, [O2]0 ) 0.10 M, [Br] ) 0.014 M, and t ) 5-15 min. These yields are the highest reported to date from this reaction in high-temperature liquid water. Moreover, under these conditions and t ) 15 min, COx yields were below 2% and reaction intermediates were not detected.
Introduction Investigations of dense, high-temperature water (HTW) as an environmentally benign medium for organic chemical reactions are increasing and encompass, for example, Beckman rearrangements (1), crossed aldol condensations (2), ene reactions (3), and bisphenol A cleavage (4). The unique properties of HTW when compared to ambient water, including a decreased dielectric constant and elevated ion product, are among the underlying causes of its successful application as a solvent for organic reactions (5). In addition to its intriguing technical applications, HTW-based chemistry is attractive because it is consistent with several of the goals of green chemistry (6) and green engineering (7). Both philosophies promote the use of inherently nonhazardous materials, such as water, in products and processes. Our initial interest in the synthesis of terephthalic acid in HTW via the selective partial oxidation of p-xylene stemmed from a demonstration of this reaction (8). The economic and environmental benefits that could result from replacing acetic acid, the medium used in commercial terephthalic acid * Corresponding author phone: (734) 764-3386; fax: (734) 7630459; e-mail:
[email protected]. † Present address: Department of Mechanical and Industrial Engineering, University of Illinois at Chicago, Chicago, IL 606077022. 10.1021/es048575+ CCC: $30.25 Published on Web 06/10/2005
2005 American Chemical Society
manufacturing processes, with HTW provided additional motivation. An HTW-based process operating at 300 °C with air separation prior to compression would likely be both economically competitive with and less energy- and pollutant-intensive than the current technology (9), if terephthalic acid can be produced in high yields at 300 °C (or below). Hamley et al. (10) reported terephthalic acid yields in excess of 90% from p-xylene partial oxidation in supercritical water at 400 °C at residence times less than 1 min. They designed a flow reactor to minimize catalyst (MnBr2) exposure to hydrothermal conditions. A cold, aqueous MnBr2 stream mixes with the heated solvent, p-xylene, and oxidant just before entering the reactor. With this technique, they prevented overoxidation and precipitation of the Mn. When the catalyst had been heated along with the solvent, the authors observed MnO2 in the preheater and hypothesized that MnO2 precipitation was the likely cause of low carbon balances. Shielding the catalyst from hydrothermal conditions during heat-up of other reactor constituents was crucial to the success of the reaction. Subsequent experiments (11) under the reaction conditions that optimized terephthalic acid yields were done for other xylene and trimethylbenzene isomers in supercritical water. p-Xylene produced a terephthalic acid yield of 90%, but the other xylene isomers formed dicarboxylic acids in yields from 16% to 66%. The partial oxidation of mesitylene and pseudocumene produced tricarboxylic acids in 78% and 50% yield, respectively. We too have previously reported on p-xylene partial oxidation in HTW (12, 13). Specifically, we determined the effects of catalyst identity, supercritical water density, and initial p-xylene, oxidant, and catalyst concentrations on the yield of terephthalic acid and reaction intermediates. Reactions occurred isothermally from 250 to 380 °C in stainless steel mini-batch reactors. The highest terephthalic acid yield obtained under supercritical conditions was 57% ( 15% at 380 °C and a batch holding time of 7.5 min (12). The highest yield under subcritical conditions was 49% ( 8% at 300 °C and a batch holding time of 30 min (13). Note that we recently reported revised values for the yields of terephthaldicarboxaldehyde, 4-carboxybenzaldehyde, and 4-methylbenzyl alcohol (reaction intermediates that formed in under 10% yield each) (14). These revised values do not affect any of the conclusions drawn from the experimental results. Herein, we present new results concerning the effects of catalyst identity, catalyst concentration, oxygen concentration, and reaction temperature on terephthalic acid production from p-xylene in HTW. We used a specially designed stirred 440 mL Hastelloy autoclave batch reactor system. A syringe pump rapidly introduces the metal bromide catalyst to the reactor along with the p-xylene when the oxidant and water are at the reaction temperature. This mode of operation avoided catalyst degradation during heat-up and gave a sharp t ) 0. We report the first successful application of air as the oxidant for metal-bromide-catalyzed p-xylene partial oxidation in HTW and the first measures of COx formation and O2 consumption from this reaction in HTW. Measuring the COx yields is important because if more than a few percent of the carbon burns, the process economics become unfavorable. We also report results from reactions with the previously unexplored (in HTW) catalyst combination of CoBr2 and Mn(OAc)2 with a metal:bromide ratio of 1:1 and with Zr as a cocatalyst (added as ZrBr4). Furthermore, we report herein the highest terephthalic yields to date for synthesis under subcritical conditions. While some contend (11) that subVOL. 39, NO. 14, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5427
critical HTW is an undesirable medium for the autoxidation of methyl aromatics, these results demonstrate its viability.
Experimental Section Materials and Procedure. We purchased all chemicals from Sigma-Aldrich in high purity and used them as received. All experiments were conducted in a 440 mL Hastelloy FC series autoclave batch reactor fabricated by PPI (Warminster, PA). An experiment begins with loading distilled, deionized water into the reactor. For experiments at subcritical temperatures, we add enough water such that the liquid phase occupies approximately 95% of the reactor volume at the reaction temperature. The system pressure at the reaction temperature will be at least the vapor pressure of water. For experiments at supercritical temperatures, the desired water density under the reaction conditions determines the water loading. Once the reactor has been loaded with water and sealed by bolting the base to the head, we add 500 psi of nitrogen to the reactor and stir the reactor contents. After an hour, the reactor is vented to expunge gases (CO2, O2) that had been dissolved in the water. Next, we add air at a desired pressure and heat the reactor. When the reactor reaches the set point temperature, we withdraw a sample to determine the initial O2 concentration and the amounts of residual COx that are present. An aqueous catalyst solution (30 mL) is pressurized in a syringe pump (Isco model 260D) to a pressure above the reactor pressure. The catalysts were MnBr2, CoBr2, Mn(OAc)2, and ZrBr4. This solution travels through a 1 mL sample loop filled with p-xylene before entering into the reactor. The p-xylene and catalyst thus enter the preheated, pressurized reactor together. Their introduction defines t ) 0 for the reaction. As the reaction progresses, we withdraw samples from the liquid phase into stainless steel sample bulbs, with a nominal internal volume of 0.8 mL. Weighing each sample bulb before and after sampling provides the mass of the sample collected. Organic Product Analysis. We recover organic products from a sample bulb, the contents of which are at a high pressure, by first attaching it to an expansion tube (∼1 mL). The expansion tube terminates in a septum, into which we insert a 10 mL syringe. The valve on the sample bulb is then opened and closed very quickly. Gas and some liquid expands from the sample bulb and enters the syringe. We discharge the syringe contents into dimethyl sulfoxide (DMSO) and repeat the procedure until the bulb is depressurized. We then rinse the bulb with DMSO and combine all liquid recovered from the bulb in a 25 mL volumetric flask for analysis via HPLC by a method described previously (13). This HPLC analysis provides the concentration of each compound in the sample bulb, which then admits calculation (14) of product molar yields (moles of product formed per mole of p-xylene loaded into the reactor). We verified that this sample collection and analysis procedure provides accurate yields of liquid-phase organic compounds. In one test, we loaded known amounts of terephthalic acid and water into the reactor, heated it to 200 °C, and collected liquid-phase samples. Analysis of samples indicated that the liquid phase contained 93% ( 3% of the expected concentration of terephthalic acid. We ran similar experiments with p-xylene, adding it to the reactor along with N2 and water. If all of the p-xylene were in the aqueous phase, its concentration in these experiments would be 0.12 M, which is below the saturation concentration (15). We heated the reactor to 280 °C and pressures of 1650 and 3000 psi and analyzed the liquid phase. We also conducted complementary vapor-liquid equilibrium (VLE) calculations with AspenPlus (16) using several thermodynamic property methods that apply to systems with polar molecules at high temperatures and pressures to determine the amount of 5428
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 14, 2005
FIGURE 1. Percent p-xylene in the liquid phase at 280 °C. p-xylene that resides in the vapor phase. Dunn (14) provides more details regarding these calculations. Figure 1 shows that the experimental data agree well with the results of the phase equilibrium calculations. Both show that an appreciable portion of the p-xylene loaded into the reactor resides in the vapor phase at equilibrium. Of course, as the reaction progresses, the consumption of p-xylene in the aqueous phase would lead to more vapor-phase p-xylene going into solution. The values shown for the VLE calculation are means calculated from the results for different thermodynamic property models. We considered the possibility that ring-opening products might form in oxidation reactions, especially under supercritical conditions. We analyzed several liquid samples, drawn after the reactor had cooled, with an HPLC technique used to detect aliphatic carboxylic acids (17). Three of these samples were from reactions at 380 °C and one was from a reaction at 300 °C. Each sample contained an unidentified compound with a UV spectrum similar to that of oxalic acid, suggesting that the compound might be an aliphatic carboxylic acid. The elution time of the unknown compound, which was identical in all the samples, did not match those of succinic, lauric, glyoxylic, propionic, formic, oxalic, maleic, or acetic acids, compounds that form during complete oxidation of aromatics in SCW (17). Without knowing the identity of the compound, there was no way to quantify its yield. We note, though, that the peak area of the unidentified compound was roughly 5-7 times greater in samples from supercritical experiments than in the sample from a subcritical experiment. Therefore, if this compound is a ringopening product, such reactions likely occur to a greater extent at supercritical conditions. Gaseous Product Analysis. This paper provides the first report of CO2 and CO yields from p-xylene partial oxidation in HTW. In experiments under subcritical reaction conditions, we cannot directly sample the gas phase because it occupies only about 5% of the reactor volume. Rather, we sample the liquid phase and infer the gas-phase composition from analysis of gases dissolved in the high-temperature liquid water. In experiments at supercritical conditions, the reactor contents are in one phase. We use a gas chromatograph (GC) equipped with a thermal conductivity detector and a method described previously (18) for gas analysis. Prior to analysis, the sample bulb resides in a freezer for at least a day to minimize the amount of water that enters the GC column. A large water peak can obstruct a CO2 peak on the chromatogram. We attach a cold sample bulb to the GC and open the valve so that gas fills the GC sample loop. Dunn (14) explains in detail the calculation of molar yields from the GC peak areas. Molar yields of CO and CO2 were calculated as moles of each product formed per mole of carbon loaded into the reactor. These yields exclude any residual COx present prior to the reaction. Mass/Atom Balances. For each experiment, we calculated mass, carbon, and oxygen balances based on the amount of each present in the p-xylene and air loaded into the reactor
at t ) 0. With a few exceptions, we report data only from samples with mass balances above 60%. The exception was made for samples withdrawn at short reaction times that contained an appreciable amount of p-xylene. Since p-xylene partitions between the vapor and liquid phases, a low mass balance at short times in these experiments is likely due to some p-xylene residing in the vapor phase, rather than loss of carbon during product recovery. The mean values of the mass, carbon, and oxygen balances for 45 independent samples under subcritical conditions were 91%, 90%, and 86%, respectively. Estimation of Uncertainty. In some experiments, we collected liquid samples at different times after the reaction had reached completion. The yields of organic products in each of these samples were largely time-invariant. Therefore, these samples can be viewed as replicates. Data from these samples permit calculation of 95% confidence intervals for the organic product yields. On the basis of this analysis, the relative uncertainty (at the 95% confidence level) for the molar yield of any of the organic products is likely less than 10% of the reported value. To determine the uncertainty associated with the yields of gaseous products, we again used replicate samples. We loaded water (15.3 mol) and air (0.19 mol) into the reactor, collected several liquid-phase samples at 300 °C, and measured the number of moles of O2, N2, CO, and CO2. This analysis (14) revealed that we can measure the number of moles of a gaseous compound in the reactor to within about half a millimole. This translates to a (0.7% absolute uncertainty in COx yields. The uncertainty in some data reported herein may, on occasion, exceed the estimates given above. There were a few instances where unphysical trends (e.g., CO2 yield decreases with time) were observed. We attribute these trends to the random error that accompanies challenging experiments in high-pressure reaction systems.
FIGURE 2. Temporal variation of product yields from p-xylene partial oxidation in HTW at 300 °C under base-case conditions ([p-xylene]0 ) 0.02 M, [O2]0 ) 0.10 M, [MnBr2] ) 0.003 M). (a) Organic compounds; (b) gaseous compounds.
Results and Discussion The partial oxidation of p-xylene yielded intermediate products observed in previous work (12, 13) as well as terephthalic acid. In addition to reporting the yields of these organic products, we also report, for the first time, yields of CO2, CO, and CH4. This section describes the effects of catalyst concentration and identity, O2 concentration, and reaction temperature on the partial oxidation of p-xylene in HTW. The figure captions in this section use the following abbreviations: 4-MBA is 4-methylbenzyl alcohol, 4-HMBA is 4-hydroxymethylbenzoic acid, 4-CBA is 4-carboxybenzaldehyde, and TPA is terephthalic acid. For clarity, we omit minor products from several figures. Dunn (14) provides complete results for all of the experiments. Base-Case Experiment. We conducted an initial, basecase experiment in HTW at 300 °C at conditions similar to those used in our previous work (13). We used less oxidant, however, to more closely mimic conditions needed commercially. Air compression will be a primary source of both capital and operating cost in a commercial-scale HTW-based process (9), so one desires to minimize the amount of oxidant in the reactor. Figure 2 displays the molar yields of major products from the base-case experiment. p-Xylene disappears quickly and p-tolualdehyde is the product present in the highest yields at short times. The tolualdehyde yield then decreases as it is oxidized to toluic acid, which subsequently forms 4-carboxybenzaldehyde. The yield of 4-CBA increases to a maximum around 3 min and then decreases as 4-CBA is oxidized further to TPA. The benzoic acid yield (not shown) is low, which indicates there is little decarboxylation of TPA occurring. The highest terephthalic acid yield, 49%, is consistent with the terephthalic acid yield observed under
FIGURE 3. Temporal variation of terephthalic acid (2, 4), p-toluic acid (9, 0), and p-tolualdehyde (b, O) yields from two p-xylene partial oxidation experiments in HTW at 300 °C ([p-xylene]0 ) 0.02 M, [O2]0 ) 0.1 M). Solid symbols indicate base-case experimental data ([MnBr2] ) 0.003 M); open symbols indicate data from the second experiment ([MnBr2] ) 0.002 M). similar conditions in mini-batch reactor experiments (13). At a batch holding time of 8-9 min, the two major reaction intermediates, p-toluic acid and 4-carboxybenzaldehyde, were present in 6% total yield, O2 conversion was nearly complete, and CO2 and CO had formed in 7% and 4% yields, respectively. We conducted a second experiment near the base-case conditions to assess reproducibility. The only difference is that the catalyst concentration was 33% lower in the second experiment. Figure 3 shows that the molar yields of key products from the second experiment possess the same trends as those from the base-case experiment, thereby demonstrating the level of reproducibility in these experiments. VOL. 39, NO. 14, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5429
The next sections report results from an extended investigation of the effect of catalyst concentration on the reaction. MnBr2 was used as a catalyst alone and in combination with CoBr2, Mn(OAc)2, and ZrBr4 at 300 °C. Catalyst Concentration and Addition Strategy. MnBr2, the catalyst, can form MnO2, which is insoluble in HTW. In the first step of the catalytic cycle, which we describe elsewhere (12), Mn(II) is oxidized to Mn(III). The Mn(III) species can be oxidized in reaction 1 to Mn(IV), which then reacts with water to form MnO2 (reaction 2), which can then precipitate:
2Mn(III) T Mn(IV) + Mn(II)
(1)
Mn(IV) + 2H2O f 4H+ + MnO2V
(2)
MnO2 formation has been observed in high-temperature water (10) and acetic acid reaction media (19-21). To test whether catalyst deactivation or precipitation as MnO2 affected the reaction, we conducted a set of experiments with identical reaction conditions but different catalyst addition strategies. In the first experiment, we injected supplemental catalyst into the reactor at a batch holding time of 10.2 min. The initial catalyst concentration was 0.004 M; the final catalyst concentration was 0.007 M. Figure 4 displays the product yields from this experiment. Upon injection of the fresh catalyst (vertical dashed line in Figure 4), the terephthalic acid yield showed no statistically significant change when compared to the point at 5 min. We take the scatter in the TPA yields in Figure 4 to be a manifestation of random error. No unreacted intermediates were detected at a 30 min batch holding time (Figure 4b), indicating that the reaction reached completion. CO2 formed in at most 5% yield, while CO yields did not rise above 2% (Figure 4c). Notably, the O2 conversion at the end of the run was 78%, whereas it had exceeded 90% in the previous experiments. The results obtained at 30 min are significant. They demonstrate synthesis of terephthalic acid in subcritical water in high yield in the absence of intermediate products and with a yield of CO and CO2 of less than 5%. These criteria are highly desired for commercial adoption of an HTW-based process. Next, we conducted an experiment with an initial catalyst concentration of 0.007 M, which is identical to the final catalyst concentration in the previous experiment. Figure 5 shows the yields of organic and gaseous products from this experiment. At a batch holding time of 2.9 min, the terephthalic acid yield had reached 61% as opposed to 35% in the previous experiment. It rose to 94% within 5.5 min and ranged between 70% and 85% for the duration of the experiment. The O2 conversion of less than 75% was comparable to that in the previous experiment. The last sample (t ) 15 min) contained no reaction intermediates. Again, this last sample demonstrated a high yield (89%) of terephthalic acid, no detectable intermediates, and low yields (80%) in HTW, with less than 5% of the carbon fed being lost to total oxidation. The highest terephthalic acid yield that we observed in this study was 5434
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 14, 2005
94% at 300 °C, [p-xylene]0 ) 0.02 M, [O2]0 ) 0.10 M, [MnBr2] ) 0.007 M, and t ) 5.5 min. These yields are the highest reported to date for this reaction in HTW under subcritical conditions. While the yields of intermediates were below detection limits at longer times in some experiments (Figures 4a and 5a), they were typically higher than one would desire. 4-Carboxybenzaldehyde is an intermediate of particular concern because it is difficult to separate from terephthalic acid. (2) It shows that there are multiple sets of reaction conditions and processing strategies that lead to yields of terephthalic acid g70%. Figure 14 compares terephthalic acid yields from five different experiments that produced high yields. The different strategies employed in these experiments included using a high initial catalyst concentration (either MnBr2 or a Co/Mn/Br combination), using a low initial catalyst concentration and adding more midreaction (with either one or two discrete additions), and adding more oxygen midreaction. The ultimate terephthalic acid yields from each of these experiments are all about the same. Therefore, terephthalic acid synthesis in HTW seems to enjoy at least a modest amount of processing flexibility. Adding O2 midreaction, using a high initial MnBr2 concentration, and injecting fresh catalyst midreaction are all possible routes to high terephthalic acid yields. Augmenting catalyst levels may be preferable, however, because O2 addition can promote burning. (3) It elucidates the effects of different catalyst-related variables, such as concentration, stability, identity, and addition strategy. The catalyst concentration strongly affects the terephthalic acid yield, O2 consumption, and total yield of reaction intermediates. We saw no evidence of MnO2 formation or Mn precipitation in reactions at 300 °C. Additionally, a Co/Mn/Br catalyst is not more active than MnBr2 alone in HTW. Equivalent terephthalic acid yields are attained if the bromide concentrations are equivalent. The O2 conversion and total yield of reaction intermediates, however, may vary depending on the identity of the metals. Rather than increasing catalytic activity, Zr inhibits the reaction in HTW. This work is the first to assess Zr as a cocatalyst for this reaction in HTW. (4) It determined the effect of O2 concentration on the reaction. A modest amount of excess O2 (about 60%) appears to be preferable. Less oxygen (about 15% excess) reduced TPA yields, and more oxygen (>200% excess) increased COx yields and did not promote additional terephthalic acid formation. Midreaction injection of O2 at a high oxygen conversion produced terephthalic acid yields higher than those produced by having all of the oxidant present at t ) 0. COx yields, however, may increase at higher initial O2 concentrations.
(5) It demonstrated that 300 °C is a good temperature for terephthalic acid synthesis in HTW. The reaction was much slower at 250 °C and produced only lower terephthalic acid yields. All reactions that we conducted at supercritical conditions also produced lower terephthalic acid yields. This work, together with previously published studies, indicates that high-temperature liquid water is a technically, economically (9), and environmentally promising medium for terephthalic acid synthesis.
Acknowledgments Dr. Valeria Dreyer and Dr. Phil Nubel (both of BP Research) provided technical assistance, insightful comments, and XRD and ICP analysis of solid products. We received financial support from BP, the National Science Foundation (CTS9985456), and the Petroleum Research Fund (34644-AC9), which is administered by the American Chemical Society. J.B.D. acknowledges support from a U.S. Environmental Protection Agency STAR Fellowship (2003-2004) and a University of Michigan Rackham Graduate School predoctoral fellowship (2003).
Literature Cited (1) Ikushima, Y.; Sato, O.; Sato, M.; Hatakeda, K.; Arai, M. Innovations in Chemical Reaction Processes Using Supercritical Water: an Environmental Application to the Production of -Caprolactam. Chem. Eng. Sci. 2003, 58, 935-941. (2) Comisar, C. M.; Savage, P. E. Kinetics of Crossed Aldol Condensations in High-Temperature Water. Green Chem. 2004, 6, 227-231. (3) Laitenen, A.; Takebayashi, Y.; Kylanlahti, I.; Yli-Kauhaluoma, J.; Sugeta, T.; Otake, K. Ene Reaction of Allylbenzene and NMethylmaleimide in Subcritical Water and Ethanol. Green Chem. 2004, 6, 49-52. (4) Hunter, S. E.; Felczak, C. A.; Savage, P. E. Synthesis of p-Isopropenylphenol in High-Temperature Water. Green Chem. 2004, 6, 222-226. (5) Savage, P. E. Organic Chemical Reactions in Supercritical Water. Chem. Rev. 1999, 99, 603-621. (6) Anastas, P. T.; Warner, J. C. Green Chemistry: Theory and Practice; Oxford University Press: New York, 1998. (7) Anastas, P. T.; Zimmerman, J. B. Design through the 12 Principles of Green Engineering. Environ. Sci. Technol. 2003, 37, 95A101A. (8) Holliday, R. L.; Brenton, Y. M.; Kolis, J. W. Organic Synthesis in Subcritical Water Oxidation of Alkyl Aromatics. J. Supercrit. Fluids 1998, 12, 255-260. (9) Dunn, J. B.; Savage, P. E. Economic and Environmental Assessment of High-Temperature Water as a Medium for Terephthalic Acid Synthesis. Green Chem. 2003, 5, 649-655. (10) Hamley, P. A.; Ilkenhans, T.; Webster, J. M.; Garcia-Verdugo, E.; Venardou, E.; Clarke, M. J.; Auerbach, R.; Thomas, W. B.; Whiston, K.; Poliakoff, M. Selective Partial Oxidation in Supercritical Water: the Continuous Generation of Terephthalic Acid from para-Xylene in High Yield. Green Chem. 2002, 4, 235-238.
(11) Garcia-Verdugo, E.; Venardou, E.; Thomas, W. B.; Whiston, K.; Partenheimer, W.; Hamley, P. A.; Poliakoff, M. Is it Possible to Achieve Highly Selective Oxidations in Supercritical Water? Aerobic Oxidation of Methylaromatic Compounds. Adv. Synth. Catal. 2004, 346, 307-316. (12) Dunn, J. B.; Savage, P. E. Terephthalic Acid Synthesis in Supercritical Water. Adv. Synth. Catal. 2002, 344, 385-392. (13) Dunn, J. B.; Savage, P. E. Terephthalic Acid Synthesis in HighTemperature Liquid Water. Ind. Eng. Chem. Res. 2002, 41, 4404465. (14) Dunn, J. B. The Partial Oxidation of p-Xylene in HighTemperature Water. Ph.D. Thesis, University of Michigan, Ann Arbor, MI, 2004. (15) Knauss, K. G.; Copenhaver, S. A. The Solubility of p-Xylene as a Function of Temperature and Pressure and Calculated Thermodynamic Quantities. Geochim. Cosmochim. Acta. 1995, 59, 2443. (16) AspenPlus, Version 11.1; Aspen Technology. (17) Thornton, T. J.; Savage, P. E. Phenol Oxidation in Supercritical Water. J. Supercrit. Fluids 1990, 3, 240-248. (18) Henrikson, J.; Savage, P. E. Water-Density Effects on Phenol Oxidation in Supercritical Water. AIChE J. 2003, 49, 718-726. (19) Cheng, Y.; Li, X.; Niu, J. Mechanism of Manganese Catalyst Precipitation in the Oxidation of p-Xylene. J. Zhejiang Univ., Eng. Ed. 2003, 37, 482-486. (20) Partenheimer, W. The Effect of Zirconium in Metal/Bromide Catalysts on the Autoxidation of p-Xylene Part II. Alternative Metals to Zirconium and the Effect of Zirconium on Manganese(IV) Dioxide Formation and Precipitation with Pyromellitic Acid. J. Mol. Catal. A 2003, 206, 131-144. (21) Jhung, S. H.; Park, Y. S. Precipitation of Manganese in the p-Xylene Oxidation with Oxygen-Enriched Gas in Liquid Phase. Bull. Korean Chem. Soc. 2002, 23, 369-373. (22) Partenheimer, W. Methodololgy and Scope of Metal/Bromide Autoxidation of Hydrocarbons. Catal. Today. 1995, 23, 69. (23) Partenheimer, W. The Effect of Zirconium in Metal/Bromide Catalysts During the Autoxidation of p-Xylene Part I. Activation and Changes in Benzaldehyde Intermediate Formation. J. Mol. Catal. A 2003, 206, 105-119. (24) Fisher, R. W. Oxidation of Alkyl-Substituted Aromatic Compounds with Air. In Applied Homogeneous Catalysis with Organometallic Compounds; Cornils, B., Herrmann, W. A., Eds.; Wiley-VCH: Weinheim, Germany, 2000. (25) Bhore, N. A.; Klein, M. T.; Bischoff, K. B. The Delplot Techniques a New Method for Reaction Pathway Analysis. Ind. Eng. Chem. Res. 1990, 29, 313. (26) Bhore, N. A.; Klein, M. T.; Bischoff, K. B. Species Rank in Reaction PathwayssApplication of Delplot Analysis. Chem. Eng. Sci. 1990, 45, 2109. (27) Savage, P. E.; Dunn, J. B.; Yu, J. Recent Advances in Catalytic Oxidation in Supercritical Water. Combust. Sci. Technol. 2005 (in press).
Received for review September 13, 2004. Revised manuscript received May 3, 2005. Accepted May 13, 2005. ES048575+
VOL. 39, NO. 14, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5435
