1208
Ind. Eng. Chem. Res. 2009, 48, 1208–1214
Catalytic Supercritical Water Oxidation for the Destruction of Quinoline Over MnO2/CuO Mixed Catalyst Mauricio J. Angeles-Herna´ndez,* Gary A. Leeke, and Regina C. D. Santos School of Chemical Engineering, UniVersity of Birmingham, Edgbaston, Birmingham, B15 2TT, United Kingdom
This article presents the results of the oxidation at supercritical conditions of a nonbiodegradable and highly stable nitrogen containing heterocyclic organic compound (quinoline) over a mixed catalyst in a tubular fixedbed reactor. The effect of operating conditions namely temperature, pressure, oxygen concentration, and initial quinoline concentration were studied to evaluate their effect on the removal of the organic compound. The results showed that the catalytic reaction depended strongly on temperature and pressure; nevertheless it was very efficient at milder operating conditions close to the critical point of water. A power-law kinetic model was proposed to quantify the catalytic oxidation of quinoline. Three Langmuir-Hinshelwood-Hougen-Watson reaction rates models were also explored to fit the experimental data. However, the kinetic data were better represented by the power-law kinetic model. The catalyst was able to maintain its activity, and thus it can be used as an alternative to reduce the severity of the process. Introduction The destruction of harmful compounds in aqueous streams has become of great importance in meeting statutory regulations. When common processes for wastewater treatment (i.e., biological oxidation) fail to destroy complex and/or stable compounds or mixtures of them, it is necessary to seek other suitable alternatives. Supercritical water oxidation has been explored as a media for the complete oxidation of toxic compounds in aqueous wastewater. Oxidation at supercritical conditions offers two advantages over similar well-known treatment processes: it occurs at much faster reaction rates than wet air oxidation and produces less harmful byproducts than incineration for its lower operating temperatures. The solvating power of water above its critical point (647.1 K and 22.05 MPa) makes it an exceptional solvent for organic compounds and gases like oxygen. The process is therefore single-phase and removal efficiencies of around 99.99% are achievable. However, because of their high stability many heterocyclic hydrocarbons are only partially oxidized, and the production of intermediates is inevitable. The byproducts produced (e.g., ammonia) are sometimes more stable than the original compound, and therefore these are harder to remove at conventional operating conditions (823-853 K and 25.0 MPa).1 As a result, in the search to improve the elimination of the organic matter with no intermediate production, a catalytic process has been envisaged. The gaslike viscosity and diffusivity of supercritical water facilitates the common interphase limitations found in heterogeneous reactions. On the other hand, the catalyst will also reduce the severity of the process allowing operation conditions close to the critical point of water without affecting the process efficiency.2 Milder operating conditions minimize the energy consumption and reduces the stress and corrosion in the equipment.3 Owing to the difficulty of destruction of nitrogen containing heterocyclic compounds, quinoline was chosen as a model compound. Under a noncatalytic supercritical water oxidation process it has been proven that operating conditions higher than 823 K and 25.0 MPa are required to efficiently affect its * To whom correspondence should be addressed. E-mail: mauricio.
[email protected]. Tel.: +44 (0) 121 414 6965. Fax: +44 (0) 121 414 5324.
removal.4 It is believed that in the presence of a catalyst these conditions can be reduced. In the pharmaceutical industry, quinoline is part of the structure of many antiseptics and antibiotics, although it is also used to produce dyes, herbicides, and paints. It is a mutagen agent which also attacks the human respiratory system; continuous exposure causes liver damage and can create allergic responses.5,6 In the present work catalytic supercritical water oxidation of quinoline was performed to investigate the removal of nitrogen containing heterocyclic compounds in the presence of a MnO2/CuO mixed catalyst (Carulite 300). Experimental Section Catalytic supercritical water oxidation of quinoline experiments were conducted in a tubular fixed-bed reactor. A schematic diagram of the laboratory scale catalytic supercritical water oxidation (CSCWO) rig is shown in Figure 1. Oxygen used for the reaction was produced by thermal decomposition of hydrogen peroxide (H2O2) in the preheating section. Hydrogen peroxide was chosen as oxidant because aqueous solutions of hydrogen peroxide can be easily prepared and pumped into the reaction system. Water, solutions of quinoline (C9H7N) and hydrogen peroxide were delivered by two Gilson 305 and one Jasco PU-1586 liquid chromatography pumps. The solution of hydrogen peroxide was kept on ice to avoid any H2O2 decomposition. Oxidant solution and water were preheated separately in two coiled sections each having dimensions of 7 m made of stainless steel (SS) 316 tubing with 6.25 mm o.d. and 2.1 mm i.d. The solution of the organic compound was mixed with supercritical water at a short distance to the inlet of the reactor, so hydrous pyrolysis of quinoline is reduced. The mixing of the streams occurs in the pipeline after the two tees and a 10 µm sintered filter. The complete mixing is assured once the reacting mixture went through the stainless steel porous disk and entered the reactor. The turbulence created in the microchannels of the disk assures a high quality mixing of the streams. The catalytic reactor is a SS 316 tubing section; it has dimensions of 100 mm length, 14.28 mm o.d. and 4.76 mm i.d. After intensive operation under supercritical conditions SS can be corroded causing operational problems. As a conse-
10.1021/ie8006402 CCC: $40.75 2009 American Chemical Society Published on Web 11/21/2008
Ind. Eng. Chem. Res., Vol. 48, No. 3, 2009 1209
Figure 1. Catalytic supercritical water oxidation rig. Table 1. Variables Studied parameter investigated
operational range
WHSV, s-1
temperature, K pressure, MPa initial oxygen concentration, SR initial quinoline concentration, mmol/L
673–773 23.0–30.0 0.5–10.0 0.1–0.6
0.04-0.125 0.04-0.125 0.04-0.125 0.04-0.125
quence, the equipment should be periodically inspected or replaced. At these operating conditions nickel base alloys are preferred as construction materials. The selection of stainless steel was based on internal safety regulations. For this specific process (oxidant and corrosive atmosphere under high pressure and temperature) and independently of the construction material, the lifetime of the rig was limited. After a series of experiments parts of the rig must be totally replaced, and therefore SS was cheaper for the construction. The selection of a specific material could influence the reaction path because metals or metallic oxides in the alloy have been identified to act as catalyst. Segond et al.,7 reported that the SCWO reaction of ammonia occurs via a parallel homogeneous and heterogeneous mechanism in the wall of the reactor. They have concluded that the reaction in the wall catalyzed by the stainless steel reactor was significant. Meanwhile, Webley et al.,1 pointed out that in the case of Inconel 625 the contribution of the wall reaction was lower than the homogeneous reaction. However, to our understanding this effect has not been reported for the oxidation of organic compounds. Porous SS discs (10 µm sintered filter; Mott Corporation) were place at both ends of the reactor to confine the catalyst within the reactor. The preheating system and the reactor were enclosed within an air heated electric furnace (AEW, Hampshire). Once the reacting mixture left the furnace, it was cooled by passing through a heat exchanger then expanded through a ball and micrometering valves to ambient pressure. Temperature along the experimental rig was measured by four thermocouples, located at the mixing point of the solutions just before entering the reactor, at the reactor exit, before entering the heat exchanger, and after the set of two valves, respectively. Pressure was monitored at the inlet of the oxidant feed stream and after the heat exchanger (see Figure 1). The catalyst particle sizes and weight hourly space velocities (WHSV) were conveniently chosen to avoid pressure drops higher than 0.3 MPa. After each experiment all feed lines were flushed for 1 h at 1 mL/min to clean the system.
The quantitative analysis of the liquid effluent was performed by liquid chromatography using an Agilent 1100 series equipped with a UV detector and a Phenomenex Luna 5 µm C-18(2) 150 × 4.6 mm column, and the total organic carbon (TOC) content was measured using a Shimadzu TOC 5050A. Hydrogen peroxide quantification was carried out by titration with a potassium permanganate (Sigma-Aldrich) solution. Complete decomposition of hydrogen peroxide has been verified experimentally at the outlet of the preheating section by feeding into the reactor only solutions of hydrogen peroxide in the absence of catalyst. Experiments were designed to cover the range of temperature, hydrogen peroxide concentrations, and reacting mixture flow rates used in the present work. Quinoline 99% (Acros-Organics) and hydrogen peroxide 50 wt % solution in water (Sigma-Aldrich) were used without any further treatment. The catalyst evaluated was a commercialmixed catalyst Carulite 300 (Carus Chemicals). This catalyst is an unsupported mix of manganese dioxide and copper oxide. The catalyst was chosen because both metals have exhibited good performance under supercritical conditions.8 During the evaluation of the key operating reaction conditions, a total of 1.0 g of catalyst with 45-63 µm particle-size was packed with silica particles (Acros-Organics) of 250-300 µm to a volumetric ratio of 1:1. However, different amounts of catalyst and particles sizes were used to study the external and internal mass transfer limitations. Results and Discussion Although, oxidation reactions are highly exothermic, isothermal operation was assumed based on the low concentration of quinoline in the feed stream and also because water removes the heat generated by the oxidation reaction. Moreover, the dilution of the catalyst with inert material assures an even distribution of the temperature along the reactor and prevents the presence of hot-spots in the catalytic bed.9 The fluctuation of temperature was (1 K. Pressure varied within (0.3 MPa and the maximum pressure drop in the system was (0.3 MPa and consequently isobaric conditions were assumed. Previous hydrolysis studies have demonstrated that quinoline was not decomposed at temperatures below 753 K and only less than 1.2% was removed at 773 K.4 Thus the effect of the hydrolysis reaction of quinoline was negligible, and it assures that only the oxidation reaction is responsible for its removal.
1210 Ind. Eng. Chem. Res., Vol. 48, No. 3, 2009
Figure 2. Reproducibility tests of quinoline oxidation (23 MPa, 673 K, initial quinoline concentration of 0.3 mmol/L, and SR ) 1).
Figure 3. Effect of temperature on the removal of TOC and quinoline (23 MPa, initial quinoline concentration of 0.3 mmol/L, and SR ) 1).
Catalytic supercritical water oxidation experiments were carried out to observe the effect of the process variables; temperature, pressure, concentration of oxygen, and initial concentration of quinoline (a summary of the operating conditions studied is shown in Table 1). The range of concentration of quinoline studied mimics the values found in pharmaceutical wastewater streams. Concentration of oxygen was calculated from the stoichiometric reaction
any byproduct is present the TOC removal value will be lower than quinoline value. The reaction proceeds slower than other similar nitrogencontaining organic compounds previously studied, and this shows that quinoline has a higher stability.10,11 Effect of Temperature. Quinoline has been the subject of study in previous work and it was shown that it was hardly oxidized under noncatalytic conditions under temperatures of 773 K.5 The addition of a catalyst aims to lower the severity of the reaction and its temperature, and therefore the energy consumption, and to improve the oxidized product. Temperature has been proven to be the main controlling variable of the reaction; essentially the higher the temperature, the more effective the oxidation reaction is in terms of the removal of the organic compound and the production of intermediates. However, the severity of the reaction can accelerate the corrosion of the reactor. Therefore, the inclusion of a catalyst is envisaged as a mean to diminish the thermal stress as well. A series of experiments were performed at 23.0 MPa, with an initial quinoline concentration of 0.3 mmol/L and stoichiometric ratio of 1. The temperature was varied from 673 to 773 K, and liquid samples were taken and analyzed. Figure 3 shows the effect of temperature at different space velocities. The reaction proceeded nearly to completion at WHSV ) 0.04 s-1, and temperature did not have any effect on the removal of TOC and quinoline. At the lowest temperature value studied the removal reached a value close to 99% for TOC and 98% for quinoline. The closeness of the values indicated that the catalyst also lowered the production of any intermediates. As the temperature increased at higher WHSV than 0.04 s-1, the conversion was higher as the temperature increased. Effect of Pressure. During the experiments temperature was maintained at 673 K and the concentration of quinoline was 0.3 mmol/L with a stoichiometric ratio of 1. The density of the reacting mixture varied from 133.8 to 357.1 kg/m3. The results are depicted in Figure 4. WHSV ) 0.04 s-1 allowed the reaction to near completion, and the effect of pressure was largely unnoticed. Higher space velocities showed clearly the dependency on pressure. As the pressure increased the amount of remaining TOC and quinoline decreased. When the pressure rose from 23.0 to 30.0 MPa at a constant space velocity of 0.125 s-1; the removal improved by 34% for TOC and 42% for quinoline. It is said that the solvation power of supercritical fluids is intimately related to their density and that density can be adjusted to improve reaction rates at supercritical condi-
2C9H7N + 21.5O2 f 18CO2 + 7H2O + N2
(1)
From this equation the stoichiometric ratio (SR) of oxygen to quinoline is obtained to calculate the required oxygen concentration as a function of the initial quinoline concentration: SR )
νQCO2 νO2CQ
(2)
Removal or conversion will be used indistinctly to refer to the efficiency of the process. Removal is given in terms of TOC and quinoline, and it is defined in terms of the initial and final concentrations: removalTOC )
CTOC0 - CTOC 100 CTOC0
removalQ )
CQ0 - CQ 100 CQ0
(3) (4)
Reproducibility. Tests were performed to evaluate the experimental error during the catalytic study. A series of five experiments were carried out under the same operating conditions of 23.0 MPa and 673 K, using a concentration of quinoline of 0.3 mmol/L and a SR of 1. During each experiment samples were taken at 1/WHSV of 8, 10, 15, 20, and 25 s. The average and standard deviation of the removal in terms of TOC and quinoline were calculated from the five samples obtained at each residence time. The results of reproducibility test are shown in Figure 2. The maximum values of standard deviation were 5.7 and 4.8% of removal in terms of TOC and quinoline removal, respectively. The values provide an estimate of the reliability of the experimental results. An interesting result is that as the reaction proceeds toward higher spatial times values of removal in terms of TOC and quinoline were very close to each other and indistinguishable. TOC removal was associated with the production of intermediates during the reaction. If the remaining TOC value depended solely on the remaining quinoline in the outlet stream both removal values would be equal. However, if
Ind. Eng. Chem. Res., Vol. 48, No. 3, 2009 1211
Figure 4. Effect of pressure on the removal of TOC and quinoline (673 K, initial quinoline concentration of 0.3 mmol/L, and SR ) 1).
Figure 5. Effect of oxygen concentration on the removal of TOC and quinoline (23 MPa, 673 K, and initial quinoline concentration of 0.3 mmol/L).
tions.12 In addition, the TOC and quinoline removal values maintained the same trend by keeping a constant distance from each other, which showed that pressure did not affect the production of byproduct. Effect of Initial Oxygen Concentration. The stoichiometric ratio of oxygen to quinoline was evaluated from 0.5 to 10 according to eq 2, meanwhile quinoline concentration was kept constant at 0.3 mmol/L. Pressure and temperature of the system was maintained at 23.0 MPa and 673 K, respectively. Oxygen concentration rapidly improved the removal of the organic compound in the outlet stream when it was increased from 0.5 to 4.0 (Figure 5). Above a SR of 4 the elimination was not significantly improved. However, concentration of oxygen did have an effect on the amount of TOC produced. Above SR 2 the values of TOC and quinoline removal tend to overlap, which indicates elimination values close to each other. Consequently, excess of oxygen can be introduced as an additional parameter to control to some extent the oxidation reaction and the elimination of the intermediates. Effect of Initial Concentration of Quinoline. The oxidation reaction was evaluated in terms of the amount of quinoline concentration that can be efficiently oxidized. For this purpose the concentration of quinoline was varied from 0.1 to 0.6 mmol/L and oxygen was supplied to a stoichiometric ratio of 1 in each case. Pressure of the system was 23.0 MPa and temperature 673 K. Figure 6 presents the findings of the experiments. In general higher concentrations of the quinoline
Figure 6. Effect of quinoline initial concentration in the oxidation reaction (23 MPa, 673 K, and SR) 1).
led to an improvement in the removal. For example, an inlet concentration of 0.6 mmol/L (approximately 600 ppm of quinoline at atmospheric conditions) reduced its TOC content to 99% at a WHSV ) 0.04 s-1. The reaction was therefore able to cope with higher concentrations of quinoline more efficiently without affecting the reactor performance. At space velocity of 0.04 s-1, the effect was barely noticeable above a concentration of 0.2 mmol/L; higher concentrations brought the reaction close to completion and just traces of both TOC and quinoline were found in the stream. The effect of concentration in a tubular fixed bed reactor has been previously studied by Krajnc and Levec,13 where they also proved that higher concentration of the acetic acid (the organic compound studied) promoted faster reaction rates. Evaluation of Concentration Gradients. The presence of external (interphase) and internal (intraphase) concentration gradients were experimentally assessed to evaluate their effect on the reaction. Both tests were carried out following common procedures which rely on the evaluation of conversion dependency on the former to the superficial velocity and the latter to the particle size.9,14 External Concentration Gradients. A series of experiments were performed at pressure of 23.0 MPa and temperature of 673 K. For the experiments quinoline concentration was maintained at 0.3 mmol/L with a SR of 1. Catalyst particle size was 212-250 µm, and the amounts of catalyst in the reactor were 0.6, 0.8, 1.0, and 1.2 g. Thus flow rate will proportionally increase to keep a constant space velocity. For the test, space velocities of 0.04 and 0.125 s-1 were selected to investigate whether there is a change in the removal caused by the increment of the flow rate of the reacting mixture. Figure 7 illustrates the effect of the flow rate at the two different space velocities studied. At WHSV ) 0.04 s-1 the conversion was kept constant, and thus it was independent of the flow rate of the reacting mixture. At 0.125 s-1 there is a slight increment in conversion; however, it was considered that the change was not significant enough to have any influence in the reaction. Oshima et al.,15 have also proved the absence of interphase concentration gradients at almost the same reaction conditions; however in their study the organic compound studied was phenol. Internal Concentration Gradients. Once the external concentration gradients have now been discarded, the next step was to quantify the effect of the particle size on the removal. The operating pressure and temperature and reactants concentration were the same as for the evaluation of external concentration
1212 Ind. Eng. Chem. Res., Vol. 48, No. 3, 2009
Figure 7. Effect of intraphase limitations (23 MPa, 673 K, initial quinoline concentration of 0.3 mmol/L, and SR ) 1).
Figure 9. Fitting of experimental data to the power-law kinetic model.
likely described by a power-law kinetic model, so eq 5 is transformed to dXQ ) kCQa COb 2 d(W ⁄ FQ0)
Figure 8. Evaluation of internal concentration gradients (23 MPa, 673 K, initial quinoline concentration of 0.3 mmol/L, and SR ) 1).
gradients. For these experiments 1.0 g of catalyst was placed in the reactor and the range of the particles sizes of catalyst used were 45-63, 150-212, and 212-250 µm. The effects of the particle sizes on the conversion are shown in Figure 8. At any given space velocity and particle size the removal was unchanged and consequently the intraphase concentration gradients were also discarded. After external and internal concentration gradients were assessed and discarded, it was concluded that the system was under chemical kinetic control. Thus kinetic parameters evaluated will not be affected by any concentration gradient, and thus it can be considered as a pseudohomogeneous model. Reaction Kinetics of Catalytic Supercritical Water Oxidation of Quinoline. The fitting of experimental data was performed using the integral analysis method proposed by Froment and Hosten.16 In the analysis the experimental data were fitted to the continuity equation of a tubular reactor: dXQ ) -rQ d(W ⁄ FQ0)
(5)
The differential equation is subjected to following initial conditions: XQ(0) ) XQ0
when
W ⁄ FQ0 ) 0
(6)
The problem is simplified to assume a suitable model for the reaction rate of quinoline (rQ) and compare the predicted outcome values of the proposed model to the experimental data. The most common approach is to assume that the reaction is
(7)
where the kinetic constant (k) and reaction orders (a and b) are calculated from experimental data. The problem is then reduced to find those optimum values that minimize the difference between experimental and predicted data. The minimization problem was solved by the implementation of the downhill simplex or Nelder-Mead algorithm and nonlinear least-squares based on a modification of the Levenberg-Marquardt algorithm.17 Simplex routine was used to estimate the initial values of the parameters to ensure the convergence of the LevenbergMarquardt algorithm to a reasonable solution. The routines were implemented in Python18 using the scientific library SciPy.19 The best fitting values found after solving the minimization with a confidence level of 95% are shown in the following equation that expresses the reaction rate of quinoline in the reactor: -rQ ) 0.288 ( 0.12C Q0.43(0.23C O0.23(0.09 2
(8)
Figure 9 shows a comparison of the fraction reacted between the experimental data set and those values found using the fitted reaction rate. In most of the cases, the values lie within a (10% of the fraction of quinoline reacted, showing a good agreement of the power-law kinetic model for the prediction of the experimental values. Pinto et al.,20 have reported reaction orders for the noncatalytic oxidation of quinoline and it is interesting to compare how the reaction rate changes with respect to the heterogeneous reaction. The reaction orders reported in their study were 0.8 with respect to quinoline and 0.3 with respect to oxygen. The values present in this study showed smaller reaction orders with those reported, although the difference with respect to oxygen was smaller. The reaction rate showed a similar dependency on oxygen concentration although the influence of quinoline was lowered in the catalytic reaction. The noninteger values of the reaction orders on both reflects the complexity of the process, which goes beyond the simple interaction of quinoline and oxygen as it is shown in eq 1 and implies the occurrence of side reactions.21Some authors have proposed that more traditional reaction rates for heterogeneous reactions that comprise adsorption and desorption steps of the chemical species could be more appropriate to explain the CSCWO reaction. The reaction models proposed are in majority the type of Langmuir-Hinshelwood-Hougen-Watson (LHHW), although a Mars-van Krevelen mechanism has also been proposed.22,23 Three LHHW
Ind. Eng. Chem. Res., Vol. 48, No. 3, 2009 1213
this time the activity of the catalyst was slightly affected until the end of the experiment. A similar finding was reported by Yu and Savage,23 for the oxidation of phenol. The lost of activity of the catalyst could be due to the transformation of the crystal structure of both oxides under supercritical conditions which were less active.33,34 In addition, it is worth mentioning that the unsupported mix catalyst used here lacks the advantageous properties provided by a support catalyst, mainly, the resistance to any thermal or mechanical shock. At supercritical operating conditions these factors are prevalent, and they are likely to lessen the catalyst lifetime. However, it is very promising that catalyst formulations based on MnO2 and CuO can be successfully exploited as alternatives to reduce the severity of the process and can be used for a wider range of organic compounds. Figure 10. Catalyst stability test (23 MPa, 673 K, 0.3 mmol quinoline/L, and SR ) 1).
reaction rates models were also explored to fit the experimental data. The LHHW reaction models were taken from previous research works, the first one assumes adsorption of reactants on different catalyst sites, the second presumes adsorption of the species on the same site, and the last model comprises adsorption of one of the species on a site and then a dissociative adsorption of the second species on a different site.15,24-26 However, the kinetic data in the present work were better represented by the power-law kinetic model. This fact has been pointed out before when Aki and Abraham,27 could not justify the use of any LHHW models with their experimental data, arguing that there is no clear evidence of the precise mechanism of the CSCWO. Additionally, Krajnc and Levec,13 compared the power-law and LHHW reaction rates and concluded that although both represented appropriately their experimental results, because of the uncertainty generated by the mechanistic approach, they recommended the use of power-law kinetic models instead. This is supported by the fact that different reaction mechanisms have been proposed for the catalytic oxidation of phenol at the same reaction conditions.15,26 CSCWO could be better explained in terms of elementary reactions mechanisms that can be used to gain insight into the reaction process. Nonetheless, only noncatalytic reactions have been modeled by this approach at operating conditions that allow the assumption that water only acts as a collision partner and discard any other type of molecular interaction. Although, the models have identified the importance of production of some free radical during the reaction, they were not able to accurately predict either product distribution or reagent disappearance.28,29 This suggests that the reaction indeed could be more complex, and the presence of a side mechanism could be expected.30 In addition, models should also take into account the solute-solvent interactions present in supercritical fluids.31 Thus a simpler model that could account for more complex reaction steps like the power-law model is preferred for its practicality.32 Catalyst Stability. An experiment was performed to evaluate the activity of the catalyst at supercritical conditions. In the test a solution of 0.3 mmol/L of quinoline was pumped through the reactor at WHSV ) 0.04 s-1. The stoichiometric ratio was maintained at 1 and the operation conditions were 23.0 MPa and 673 K. The liquid stream was sampled to monitor the stability of the catalyst in terms of TOC and quinoline removal. The catalyst activity decreased considerably within the 0.5 h of operation and after 1.0 h, the removal of TOC and quinoline was reduced by 25 and 20%, respectively, from its initial value (Figure 10) and the production of intermediates occurred. After
Concluding Remarks In this work a series of experiments were performed in a tubular reactor to evaluate the catalytic supercritical oxidation of quinoline over a mixed MnO2/CuO catalyst. Although, it was shown that the removal of quinoline depended strongly on the operating temperature and pressure conditions, the reaction proved to proceed efficiently at operating conditions close to the critical point of water, where lower energy consumption is required and less thermal and mechanical stresses exist within the process equipment. The initial concentration of oxygen had a positive influence on the oxidation reaction; oxygen in excess is preferred to accelerate the catalytic oxidation reaction and to control the production of intermediates. Close values of quinoline and TOC removal were found during the experiments, this indicates that the catalyst reduced the production of undesirable intermediates. In addition, the removal of TOC and quinoline was not affected by the amount of the initial concentration of the organic compound. Based on its stoichiometry and the operation at supercritical conditions, the reaction involves a complex mechanism. Nevertheless, the reaction was satisfactorily represented using a power-law kinetic model. Catalysts based on formulations of MnO2 and CuO can be successfully used for the destruction of a wide type of organic compounds in water. The study demonstrated the feasibility of transition metal oxides to destroy a highly stable and nonbiodegradable organic compound commonly found in the wastewater discharges of pharmaceutical companies. These metal oxides offer good activity at the severe operating conditions of supercritical water oxidation. Nomenclature a ) reaction order of quinoline b ) reaction order of oxygen CO2 ) oxygen concentration in the bulk [mmol/L] CQ ) quinoline concentration in the bulk [mmol/L] CQ0 ) initial quinoline concentration in the bulk [mmol/L] CTOC ) concentration of TOC [mmol/L] CTOC0 ) initial concentration of TOC [mmol/L] FQ0 ) initial molar flow rate of quinoline [mmol/s] k ) kinetic constant [mmol/L]-(a+b)+1/s removalQ ) removal expressed in terms of quinoline [%] removalTOC ) removal expressed in terms of TOC [%] rQ ) reaction rate of quinoline [mmol/kg catalyst/s] W ) weight of the catalyst [g] WHSV ) weight hourly space velocity [s-1]
1214 Ind. Eng. Chem. Res., Vol. 48, No. 3, 2009 Xq ) fraction converted of quinoline νQ ) quinoline stoichiometric coefficient νO2 ) oxygen stoichiometric coefficient
Acknowledgment Mauricio Angeles expresses his deep gratitude to Consejo Nacional de Ciencia y Tecnologı´a (CONACYT), Me´xico (Studentship No. 204623), and Secretarı´a de Educacio´n Pu´blica (SEP) through its Direccio´n General de Relaciones Internacionales, Me´xico, for the financial support provided for this research. The authors would like to thank Dr. Serafim Bakalis for his useful suggestions in the fitting of the experimental data. Literature Cited (1) Webley, P. A.; Tester, J. W.; Holgate, H. R. Oxidation kinetics of ammonia and ammonia-methanol mixtures in supercritical water in the temperature range 530-700°C at 246 bar. Ind. Eng. Chem. Res. 1991, 30, 1745. (2) Savage, P. E.; Dunn, J. B.; Yu, J. Recent advances in catalytic oxidation in supercritical water. Combust. Sci. Technol. 2006, 178, 443. (3) Bermejo, M. D.; Cocero, M. J. Supercritical water oxidation: A technical review. AIChE J. 2006, 52, 3933. (4) Pinto, L. S. D. Supercritical water oxidation of nitrogen-containing organic compounds: Process operating conditions and reaction kinetics. Ph.D. Thesis. University of Birmingham, Birmingham, U.K., 2004. (5) Pinto, L. D. S.; Freitas dos Santos, L. M.; Al-Duri, B.; Santos, R. C. D. Supercritical water oxidation of quinoline in a continuous plug flow reactorsPart 1: Effect of key operating parameters. J. Chem. Technol. Biotechnol. 2006, 81, 912. (6) Nicolaescu, A. R.; Wiest, O.; Kamat, P. V. Mechanistic pathways of the hydroxyl radical reactions of quinoline. 1. Identification, distribution, and yields of hydroxylated products. J. Phys. Chem. A 2005, 109, 2822. (7) Segond, N.; Matsumura, Y.; Yamamoto, K. Determination of ammonia oxidation rate in sub- and supercritical water. Ind. Eng. Chem. Res. 2002, 41, 6020. (8) Zhang, X.; Savage, P. E. Fast catalytic oxidation of phenol in supercritical water. Catal. Today 1998, 40, 333. (9) Perego, C.; Peratello, S. Experimental methods in catalytic kinetics. Catal. Today 1999, 52, 133. (10) Angeles, M. J.; Al-Marzouqui, A. H.; Al-Duri, B.; Santos, R. C. D. Catalytic supercritical water oxidation for the destruction of DBU. I Iberoamerican Conference on Supercritical Fluids, ProSCiba, April 1013th, Iguassu Falls, Brazil, 2007. (11) Aki, S. N. V. K.; Abraham, M. A. Catalytic supercritical water oxidation of pyridine: kinetics and mass transfer. Chem. Eng. Sci. 1999, 54, 3533. (12) Eckert, C. A.; Chandler, K. Tuning fluid solvents for chemical reactions. J. Supercrit. Fluids 1998, 13, 187. (13) Krajnc, M.; Levec, J. The role of catalyst in supercritical water oxidation of acetic acid. Appl. Catal., B 1997, 13, 93. (14) Dautzenberg, F. M. Ten guidelines for catalyst testing In Characterization and Catalyst DeVelopment; Bradley, S. A., Galhuso, M. J.,
Bertolacini, R. J., Eds.; ACS Symposium Series: Los Angeles, CA, 1989; Vol. 411, p 99. (15) Oshima, Y.; Tomita, K.; Koda, S. Kinetics of the catalytic oxidation of phenol over manganese oxide in supercritical water. Ind. Eng. Chem. Res. 1999, 38, 4183. (16) Froment, G. F.; Hosten, L. H. Catalytic Kinetics: Modelling. In Catalysis: Science and Technology; Anderson, J. R., Boudart, M., Eds.; Springer-Verlag: Berlin, 1990. (17) Press, W. H.; Teukolsky, S. A.; Vetterling, W. T.; Flannery, B. P. Numerical Recipes; Cambridge University Press: Cambridge, U.K., 2007. (18) Langtangen, H. P. Python Scripting for Computational Science; Springer-Verlag: Berlin, 2008. (19) Jones, E.; Oliphant, T.; Peterson, P.; Millman, J.; Kern, R.; van der Walt, S.; Hunter, J. D.; Perez, F.; Certik, O.; Kitterman, S. et al. SciPy: Open Source Scientific Tools for Python, http://www.scipy.org, 2001. (20) Pinto, L. D. S.; Freitas dos Santos, L. M.; Santos, R. C. D.; Al-Duri, B. Supercritical water oxidation of quinoline in a continuous plug flow reactorsPart 2: Kinetics. J. Chem. Technol. Biotechnol. 2006, 81, 919. (21) Hill, C. G. An Introduction to Chemical Engineering and Reactor Design; Wiley: New York, 1977. (22) Yu, J.; Savage, P. E. Kinetics of catalytic supercritical water oxidation of phenol over TiO2. EnViron. Sci. Technol. 2000, 34, 3191. (23) Yu, J.; Savage, P. E. Phenol oxidation over CuO/Al2O3 in supercritical water. Appl. Catal., B 2000, 28, 275. (24) Savage, P. E. Heterogeneous catalysis in supercritical water. Catal. Today 2000, 62, 167. (25) Krajnc, M.; Levec, J. Oxidation of phenol over transition-metal oxide catalyst in supercritical water. Ind. Eng. Chem. Res. 1997, 36, 3439. (26) Yu, J.; Savage, P. E. Catalytic oxidation of phenol over MnO2 in supercritical water. Ind. Eng. Chem. Res. 1999, 38, 3793. (27) Aki, S.; Abraham, M. A. Catalytic supercritical water oxidation of pyridine: Comparison of catalysts. Ind. Eng. Chem. Res. 1999, 38, 358. (28) 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. A 2000, 104, 10576. (29) Holgate, H. R.; Tester, J. W. Oxidation of hydrogen and carbon monoxide in sub- and supercritical water: Reaction kinetics, pathways and water-density effects. 2. Elementary reaction modeling. J. Phys. Chem. 1994, 98, 810. (30) Bu¨hler, W.; Dinjus, E.; Ederer, H. J.; Kruse, A.; Mas, C. Ionic reactions and pyrolisis of glycerol as competing reaction pathways in nearand supercritical water. J. Supercrit. Fluids 2002, 22, 37. (31) Kruse, A.; Dinjus, E. Hot compressed water as a medium and reactant. 2. Degradation Reactions. J. Supercrit. Fluids 2007, 41, 361. (32) Levenspiel, O. Chemical reaction engineering. Ind. Eng. Chem. Res. 1999, 38, 4140. (33) Tomita, K.; Oshima, Y. Stability of manganese oxide in catalytic supercritical water oxidation of phenol. Ind. Eng. Chem. Res. 2004, 43, 7740. (34) Yu, J.; Savage, P. E. Catalyst activity, stability, and transformations during oxidation in supercritical water. Appl. Catal., B 2001, 31, 123.
ReceiVed for reView April 20, 2008 ReVised manuscript receiVed September 11, 2008 Accepted September 15, 2008 IE8006402
