Multiscale Reaction Pathway Analysis of Methyl tert-Butyl Ether

Ind. Eng. Chem. Res. 2002, 41, 1-8

1

APPLIED CHEMISTRY Multiscale Reaction Pathway Analysis of Methyl tert-Butyl Ether Hydrolysis under Hydrothermal Conditions Joshua D. Taylor,†,§ Federico A. Pacheco,† Jeffrey I. Steinfeld,‡,§ and Jefferson W. Tester*,†,§ Department of Chemical Engineering, Department of Chemistry, Building 2-221, Energy Laboratory, Building E40-445, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139

Two decomposition mechanisms for methyl tert-butyl ether (MTBE) under hydrothermal conditions were analyzed: a unimolecular decomposition pathway and an acid-catalyzed hydrolysis pathway. Ab initio quantum chemistry methods were employed to account for solvent effects on the activation energy of the unimolecular decomposition pathway over a range of compressed fluid conditions from 150 to 600 °C at 250 bar. Applying this correction resulted in a local minimum in the reaction rate below the critical temperature followed by a local maximum above the critical temperature. However, this reaction pathway contributed minimally to the overall decomposition, except at temperatures above 550 °C. Experiments were conducted under acidic and basic conditions to determine the effect of pH on the reaction rate. These results confirmed that the primary reaction path is acid-catalyzed hydrolysis. Experimentally determined reaction orders for the concentration of H+ ranged from 0.57 to 0.83 ((0.05) from 200 to 450 °C, below the reaction order of 1.0 predicted by the proposed mechanism. The discrepancy between measured and predicted values was most likely caused by the increasing importance of the reverse reaction in the acid-catalyzed hydrolysis pathway under more acidic conditions. Background and Motivation The physical properties of water change dramatically as the temperature and pressure are increased beyond their critical values (Tc ) 374 °C and Pc ) 221 bar), causing shifts in solubility, reaction rate, and mechanistic pathway. Below its critical temperature, liquid water behaves primarily as a polar solvent in which ionic reaction pathways readily occur. Above its critical point, water behaves as a nonpolar solvent, showing complete miscibility with gases and many nonpolar organic compounds such as CH4, C6H6, and C6H14. In the supercritical region, nonionic pathways, including free-radical reactions, tend to dominate. In an earlier paper, we reported experimental observations of the hydrolysis of methyl tert-butyl ether (MTBE) under hydrothermal conditions from 150 to 600 °C at 250 bar.1 The experimentally measured reaction rate for the decomposition of MTBE reached a local maximum around 350 °C followed by a local minimum near 400 °C. Two possible explanations for the decrease in reaction rate with increasing temperature near the critical point were proposed and evaluated. The first explanation assumed that two different reaction pathways occurred at isobaric conditions of 250 bar. With this assumption, an ionic hydrolysis mecha* To whom correspondence should be addressed. E-mail: [email protected]. Fax: (617) 253-8013. † Department of Chemical Engineering. ‡ Department of Chemistry, Building 2-221. § Energy Laboratory, Building E40-445.

nism dominates below the critical temperature where water is a polar solvent. Above the critical temperature, where water behaves as a nonpolar solvent, a nonionic pathway becomes the primary decomposition mechanism. Two separate rate expressions were fit to the data in the distinct regions, and comparable activation energies were determined for both temperature/density domains. However, the activation energy determined for the postulated nonionic pathway (at supercritical temperatures) was 98 kJ/mol whereas the gas-phase unimolecular decomposition value reported in the literature was 247 kJ/mol.2 This large difference in the activation energies suggested that a different mechanism might be dominant in the supercritical region. The second explanation for the unique behavior was that the reaction occurred by an acid-catalyzed mechanism (shown in Figure 1) over the entire temperature domain. According to this hypothesis, the decrease in reaction rate with increasing temperature near the critical point was attributed to the sharp drop in the concentration of H+ as the ion dissociation product of water (Kw ) aH+aOH-) decreased from 10-11 to 10-18. An empirical rate expression with explicit first-order dependence on [H+] was fit to the data and was consistent with the measured rate constants within experimental uncertainty over the entire temperature range.1 Although the acid-catalyzed mechanism was consistent with the experimentally measured rates, other mechanisms could also fit the experimental decomposition rate. For example, a base-catalyzed mechanism would show behavior identical to the acid-catalyzed mechanism, assuming that [H+] ) [OH-] (i.e., the

10.1021/ie010495g CCC: $22.00 © 2002 American Chemical Society Published on Web 01/02/2002

2

Ind. Eng. Chem. Res., Vol. 41, No. 1, 2002

chemistry methods were employed to determine the effects of the solvent on the activation energy. By doing so, the contribution of the unimolecular decomposition over the entire temperature range could be determined. The aim of the second part of this study was to validate the acid-catalyzed hydrolysis pathway. According to the proposed reaction mechanism, the rate of MTBE decomposition should be directly proportional to the concentration of H+ in solution. Experimental measurements of the hydrolysis rate in the presence of strong acid or base were conducted to identify the effect of pH on the decomposition kinetics. By comparing the reaction rate at several different pH values, the dependence on the concentration of H+ was determined and compared to that predicted by the mechanism. Review of Prior Studies

Figure 1. Proposed acid-catalyzed mechanism for the decomposition of MTBE.

acidity or basicity of the products and intermediates did not significantly alter the pH). Furthermore, ionic hydrolysis pathways in the supercritical region have not yet been experimentally validated. Therefore, further experiments and calculations were necessary to confirm the hypothesized mechanism. Objective The ultimate objective of this study was to characterize the reaction mechanism(s) contributing to MTBE decomposition in both the sub- and supercritical water. Both ab initio quantum chemistry methods and experimental measurements were employed to achieve this goal. Two different reaction mechanisms were considered to determine the relative importance of each pathway. These two pathways were as follows:

(1) The unimolecular decomposition CH3OC(CH3)3 f CH3OH + i-C4H8 (2) The acid-catalyzed hydrolysis 8 CH3OH + (CH3)3COH CH3OC(CH3)3 + H2O 9 + H

The approach in the first part of this study was to calculate the effect of the solvent on the gas-phase unimolecular decomposition rate of MTBE. In our previous paper,1 this pathway was considered minor except at temperatures greater than 550 °C because of its large activation energy (247 kJ/mol) in the gas phase.2 Because the transition state of this pathway has a significantly larger dipole moment than MTBE, however, the effective activation energy is reduced as the solvent becomes more polar (at lower temperatures), causing the rate to increase. Computational quantum

Ab Initio Calculations of Solvent Effects on Reactions in SCW. Several authors have used ab initio computational tools to model solvent effects on reactions in SCW. Johnston and co-workers3,4 have looked extensively at the effects of solvation on the reaction of a chloride ion with methyl chloride using molecular dynamics (MD) tools as well as continuum models. The authors found that the change in solvation of ions in supercritical water, relative to ambient water, resulted in reaction rates that were 9-12 orders of magnitude greater than those in ambient, liquid water. Pomelli and Tomasi5 later showed that they could reproduce the results by Johnston and co-workers using the polarized continuum model. This indicated that computations could be calculated much more efficiently than using MD to model individual water molecules. Akiya and Savage6 used a combination of density functional theory and molecular dynamics for calculations on the dissociation of hydrogen peroxide in supercritical water. In their study, ab initio calculations were performed at vacuum conditions, and the resulting species (reactant and transition state) were then placed into the MD simulation. The authors found that solvation effects in supercritical water reduced the activation barrier for the dissociation of H2O2 by 2.1 kJ/mol relative to the gas-phase reaction. In our group, Marrone and co-workers7 used ab initio computational tools to model solvent effects on the rate of methylene chloride hydrolysis in sub- and supercritical water. Here, we found the rate-determining step to be the nucleophilic attack of H2O on CH2Cl2 and calculated the location of the transition state. Because the transition state was very polar, the activation barrier increased as the solvent became more polar. Thus, it was possible to develop a correction to the activation energy, relative to ambient water, using ab initio tools. Calculated results were in good agreement with subsequent experimental measurements in our laboratory.8 Effects of pH on Hydrothermal Reactions. Many authors have discussed acid-catalyzed reaction mechanisms for chemical transformations in subcritical water. Siskin and Katritzky’s review article9 lists many of the reactions studied in hydrothermal water and specifies which reactions are reported as being acid-catalyzed. The goal of these studies was to show the range of possible chemical reactions and qualitatively discuss mechanisms, but not to quantitatively investigate effects of acidity on reaction kinetics. Several authors

Ind. Eng. Chem. Res., Vol. 41, No. 1, 2002 3

have performed detailed kinetic studies on acid-catalyzed hydrolysis reactions in subcritical10-12 and supercritical12,13 water. In the study by Krammer and Vogel,12 the authors concluded that, near the critical temperature, the reaction mechanism changed from acid catalysis (below Tc) to direct nucleophilic attack (above Tc). Penninger et al.14 studied the effects of added NaCl on the hydrolysis rate of diphenyl ether in supercritical water (415-480 °C). Upon addition of small amounts of NaCl, the hydrolysis rate decreased because of the basicity of Cl- under supercritical water conditions. However, at higher salt concentrations, the hydrolysis rate increased proportionally to the square root of the NaCl concentration. This was explained by a mechanism where hydrated Na+ clusters provided the catalytic protons. Antal and co-workers15-18 have studied the acidcatalyzed dehydration of several alcohols in subcritical water (T ) 225-320 °C) both with and without the addition of strong acid or base. In their initial study of the dehydration of tert-butyl alcohol to form isobutene, the authors described the kinetics of hydrolysis by an acid-catalzyed mechanism with two accessible reaction pathways. 16 In one pathway, the protonated alcohol was dehydrated, leaving a tertiary carbocation that then underwent elimination to form isobutene. In the second pathway, the carbocation reacted with another alcohol to form a protonated ether, which deprotonated and then underwent a unimolecular decomposition to form one isobutene molecule and one tert-butyl alcohol molecule. In a study of the hydrolysis of 1- and 2-propanol,18 the primary alcohol was reported to react via direct acidcatalyzed dehydration, whereas 2-propanol reacted by an acid-catalyzed ether formation followed by an uncatalyzed, unimolecular decomposition reaction. We note that these uncatalyzed unimolecular decomposition pathways of ethers are contrary to the mechanism we have proposed for MTBE decomposition and further motivate our multiscale analysis of hydrolysis pathways. Computational and Experimental Methods Ab Initio Calculations. Gaussian 98 (G98)19 was employed for quantum chemical calculations on the unimolecular decomposition of MTBE. To model the solvent effects, the self-consistent isodensity polarized continuum model (SCI-PCM) was used. In this model, the shape of the molecule is approximated by a threedimensional surface mapped at a normalized electron cloud density of 0.0004 (default value in G98). Then, the charge density is determined over this surface, and the solvent is modeled as a polarized continuum with a dielectric strength specified by the user. In these calculations, the bulk value for the dielectric strength of pure water was used. Although several authors have suggested that the “local” dielectric constant (around a solute) of water near its critical point is greater than the bulk value, this effect is not well understood and was therefore not incorporated into these calculations. The calculations were performed by first estimating the geometry of MTBE with empirical methods. The geometry of MTBE was then optimized using density functional theory (B3LYP) with the 6-31g(d) basis set in a vacuum. The SCI-PCM solvent model was then added and the geometry optimized for each value of dielectric strength (), beginning with the smallest values of (highest temperature) and using the result-

ing geometry as the initial guess for the subsequent value of . Once the geometry was determined for each value of , the vibrational frequencies and thermochemistry were calculated with B3LYP/6-31g(d). Finally, a larger basis set (6-311++g(3df,2p)) was used for singlepoint calculations at each value of to obtain more accurate estimates of the energies. Once the calculations of MTBE were completed, the geometry of the transition state for the unimolecular decomposition was investigated. The reaction coordinate of the unimolecular decomposition was assumed to be hydrogen transfer from a methyl group to the oxygen atom. The transition state geometry was determined by guessing an initial structure and allowing G98 to perform an optimization. In the optimization, the atoms were iteratively moved until the total energy was minimized for the “molecule” except along the reaction coordinate, where the energy was maximized (saddle point). A frequency calculation was then performed to ensure that the proper reaction coordinate was determined. After optimizing the geometry of the transition state in a vacuum conditions, we then followed the same procedure as before for calculating the geometry, energy, and thermochemistry at each value of . pH Dependence of MTBE Hydrolysis Rate. The objective of these experiments was to determine the effects of pH on the hydrolysis rate of MTBE. The experimental apparatus has been described previously and was used with the same operating procedure.1 However, in place of deionized water, solutions of HCl were used for acidic conditions, and solutions of NaOH were used for basic conditions. Solutions were prepared by first placing a measured quantity of deionized water into the feed bottle. Helium was then bubbled through the water for 5 min to remove any dissolved oxygen or carbon dioxide. After the water was properly degassed, a measured quantity of 1.0 M HCl (or 0.1033 N NaOH) was injected with a syringe through a septum and continually mixed with a stir bar on a magnetic plate. A head pressure of 10 psig (1.7 bar) helium was maintained over the solutions to exclude O2 and CO2 from the feed. In the subcritical temperature region (at 250 bar), the hydrolysis rate of MTBE was measured at 200 and 250 °C at five different pH values. Two acidic solutions, two basic solutions, and one neutral solution (pure deionized water) were used to determine the dependence of the reaction rate on the concentration of H+ (catalyst). The acidic solutions had ambient pH values of 3.0 and 3.7, and the basic solutions had pH values of 10.0 and 11.0. The pH of each solution was verified with titration prior to the experiments. The pH of the effluent from each experiment was measured with pH paper ((0.2 pH units) to ensure that the products did not change the pH significantly. For subcritical temperatures, these prepared solutions were sufficiently dilute to ensure complete dissociation of HCl and NaOH. The concentration of H+ at the reaction conditions (subcritical temperature and high pressure) was calculated as follows. For acidic conditions, the concentration of H+ was determined by correcting the concentration (from the completely dissociated HCl) for the change in density from ambient conditions to reaction conditions,

[H+]T,P ) [H+]ambient

F(T, P) Fambient

(1)

4


Figure 2. Molecular geometries of MTBE and the unimolecular decomposition transition state calculated at the B3LYP/6-31g(d) level.

where F(T,P) is the density of the solution at the temperature and pressure of the reactor and Fambient is the ambient density of the solution (assumed to be that of pure water). For neutral conditions, the concentration of H+ was calculated from the literature values of the dissociation of water, 20

[H+]T,P )

x

Kw(T,P) γ2((T,P)

(2)

where γ( is the mean ion activity coefficient at T and P relative to a reference state of 1 m ionic strength (extrapolated from dilute solution behavior). γ( was estimated using an extended Debye-Hu¨ckel model and found to vary from 1.0 to 1.1 over the conditions studied. For basic conditions, the concentration of OH- was determined in the same manner as H+ was in eq 1,

[OH-]T,P ) [OH-]ambient

F(T,P) Fambient

(3)

and then [H+] was calculated from this value of [OH-] as

[H+]T,P )

Kw(T,P) 2 γ((T,P) [OH-]T,P

(4)

Experiments in the supercritical region were conducted at 450, 550, and 600 °C and 250 bar with the same range of pH solutions. However, under supercritical conditions, HCl and NaOH do not dissociate significantly. Estimates for the concentration of H+ for these experiments will be discussed in the following section. In all experiments, the concentrations of organic compounds in the liquid-phase effluent were measured by direct injection into a gas chromatograph (HP 6890) with a flame ionization detector, as described previously.1 Results Unimolecular Decomposition Pathway. The ab initio calculations found the transition state for the unimolecular decomposition pathway as an intermediate point in the transfer of a hydrogen atom from a methyl group to the methoxy group. The geometries of MTBE and the transition state are shown in Figure 2, where the reaction coordinate involves the hydrogen

Figure 3. Dipole moments of MTBE and transition state versus bulk dielectric strength ().

atom traveling from one group to the other. The completion of the reaction would involve the departure of the methanol molecule, leaving isobutene as the other product. The calculated activation energy for the unimolecular decomposition in a vacuum was 251 kJ/mol. The experimentally measured activation energy under verylow-pressure-pyrolysis conditions (P < 10-5 bar) was 247 kJ/mol.2 This type of discrepancy is not unusual for ab initio calculations because of their finite-sized basis sets. For the purposes of this study, then, the activation energy in a vacuum was set to the value measured experimentally (in low-pressure gas-phase conditions). Calculated changes to the activation energy due to solvation effects are used as a correction to the literature value. Calculations were performed successfully for all of the supercritical temperatures, but geometry optimizations for the transition state failed to converge for dielectric strengths greater than 7.0 (T ) 381.5 °C, P ) 250 bar). While this may limit the absolute applicability to the subcritical region, trends can be observed in the successful calculations and used to make approximations over the entire temperature range. Furthermore, the results will show that this pathway does not significantly contribute to the decomposition of MTBE at lower temperatures. A plot of the dipole moment of MTBE and the transition state are shown as a function of the bulk dielectric strength () in Figure 3. According to our calculations, MTBE has a dipole moment of 0.47 D (Debye) in a vacuum, which increases with dielectric strength and finally levels off at ≈0.6 D as becomes large (>5). The transition state has a dipole moment in a vacuum of 0.95 D, which initially increases more rapidly with , and then levels off around 1.4 D. The transition state’s dipole moment is approximately twice as large as that of the reactant. The larger dipole moment results in greater stabilization of the transition state as the polarity of the solvent is increased, effectively decreasing the activation barrier. The calculated activation energy as a function of is shown in Figure 4. The decrease appears approximately linear over this range of dielectric strength. One would expect that the activation energy would eventually level off as the polarity increases further. However, assuming Ea continues to decrease with the same slope should provide the most conservative estimate of the effective activation energy as a function of . Applying this slope

Ind. Eng. Chem. Res., Vol. 41, No. 1, 2002 5

Figure 4. Activation energy of unimolecular decomposition as a function of bulk dielectric strength ().

Figure 6. MTBE hydrolysis rate constant versus pH at 200 °C and 250 bar. Table 1. Summary of Measurements at 200 °C and 250 bar ambient pH 3 3.7 7 (neutral) 10 11

Figure 5. Comparison of measured MTBE disappearance rate constant with calculated contribution due to the unimolecular decomposition versus 1000/T.

as a correction to the rate expression measured in the gas phase, one obtains,

kuni ) A exp

[

]

-Ea - f() ) RT 1013.9 exp

[

]

-246860 + 1250 (5) 8.314T

where A is the pre-exponential factor in units s-1, f() is a correction factor to the activation energy, and R is the universal gas constant (8.314 J/mol‚K). A comparison between the unimolecular decomposition rate constant calculated with eq 5 and the assumedfirst-order rate constant measured in sub- and supercritical water is plotted versus 1000/T in Figure 5. The behavior shown confirms that the unimolecular decomposition pathway is insignificant compared to the total reaction rate, except at temperatures greater than 550 °C. The expression for the rate constant in eq 5 gives a lower bound to the rate that MTBE will decompose at these temperatures. Because the observed rate is significantly faster than that predicted in the gas phase, one can conclude that water must behave as more than just a dielectric medium (i.e., as a reactant and/or catalyst) for the overall reaction. This fact suggests that acid-/base-catalyzed hydrolysis may play an important role. Acid-Catalyzed Hydrolysis Pathway. Subcritical Experiments. In the subcritical region, experiments were performed at 200 and 250 °C at 250 bar. A summary of the experiments conducted at 200 °C and 250 bar is shown in Table 1. The only measured products from

calculated pH (200 °C, no. of residence conversion 250 bar) points times (s) (%) 3.1 3.8 5.6 7.2 8.2

6 5 6 3 3

10-120 15-180 30-720 30-9900 300-6000

11-92 13-94 5-77 0-15 0-1

kexpt (s-1) 2.1 × 10-2 1.7 × 10-2 2.0 × 10-3 1.7 × 10-5 1.4 × 10-6

these experiments were methanol, tert-butyl alcohol, and isobutene. This indicates that no new decomposition pathways leading to additional products are triggered by the addition of acid or base. Furthermore, the relative amounts of each compound produced did not change as a function of pH nor did the pH (measured in the effluent) change as a function of conversion. According to the proposed acid-catalyzed mechanism (Figure 1), the experimentally measured rate constant (kexpt) for the first-order disappearance of MTBE can be written as follows,

kexpt ) k2K1(γ([H+])

(6)

-log10(kexpt) ) -log10(k2K1) + pH

(7)

or

where K1 is the equilibrium constant of MTBE protonation (step 1 in Figure 1), k2 is the forward rate constant of step 2 in Figure 1 (rate-limiting step), and (γ([H+]) is the activity of H+. Therefore, a plot of -log10(kexpt) versus pH (where pH ) -log10(aH+) = -log10[H+]) should have a slope of 1.0. Such a plot is shown in Figure 6 for the experiments conducted at 200 °C. Although the data can be linearly correlated over most of the pH range, the dependence of kexpt on pH is lessened as acidity increases. The leveling off at low pH is probably because the reverse reaction is also catalyzed by H+. At low pH, the reverse reaction becomes competitive with the forward reaction, and the apparent rate of decomposition is slower than expected. The slope obtained from a least-squares fit of all the data is 0.83 ( 0.35 (at 95% confidence). If we eliminate the data point at the lowest pH (where the reverse reaction is expected to be most significant), the slope increases to 0.94 ( 0.5, which agrees with the predicted value within experimental uncertainty. Similar experiments were conducted at 250 °C and 250 bar and are summarized in Table 2. Again, the only products detected were methanol, isobutene, and tert-

6


Figure 7. MTBE hydrolysis rate constant versus pH at 250 °C and 250 bar. Figure 8. MTBE hydrolysis rate constant versus pH at 450 °C and 250 bar.

Table 2. Summary of Measurements at 250 °C and 250 bar ambient pH 3 3.7 7 (neutral) 10 11


4 5 7 4 4

5-20 5-30 30-540 30-5400 60-5400

84-98 61-98 23-93 1-74 0-40

kexpt (s-1) 1.5 × 10-1 1.3 × 10-1 5.5 × 10-3 2.5 × 10-4 3.4 × 10-5

butyl alcohol. Under neutral and acidic conditions, comparable amounts of tert-butyl alcohol and isobutene were produced. However, under basic conditions, significantly more tert-butyl alcohol was produced than isobutene. In all cases, the effluent pH, measured at ambient conditions, did not change over the course of the reactions, indicating that the products do not significantly alter the pH of the solution. A plot of -log10(kexpt) versus pH is shown in Figure 7. Similar to the results at 200 °C, the data are linear over most of the pH range, with somewhat less dependence as acidity increases. The slope obtained using all of the data points was 0.75 ( 0.19. If the lowest pH data point is omitted, the slope increases to 0.83 ( 0.15. These values are slightly lower than expected, but given the assumptions used, are still consistent with the proposed acid-catalyzed mechanism. Supercritical Experiments. In the supercritical region, experiments were conducted at 450, 550, and 600 °C at 250 bar. Calculation of the concentration of H+ under supercritical conditions is not straightforward because of incomplete dissociation of HCl and NaOH. Another complication is that dissociation constants under these conditions inherently have large uncertainties because of the difficulty of conducting electrochemical measurements in supercritical solutions. For the experiments in supercritical water, the dissociation constants for HCl and NaOH were calculated using correlations found in the literature.21 The K values used for 450 °C and 250 bar (F ) 109 kg/m3) were KHCl ) 10-13.5 and KNaOH ) 10-7.0. The value for Kw used at these conditions was 10-18.36. 22 With the use of these values, the concentration of H+ for the acidic conditions was found as follows,

[H+] )

x

KHCl [HCl] Kγ,HCl

(8)

where Kγ,HCl is the ratio of activity coefficients in the equilibrium expression for the dissociation of HCl. Similarly, the concentration of H+ in the basic experi-

Table 3. Summary of Measurements at 450 °C and 250 bar ambient pH 3 3.7 7 (neutral) 10 11


5 5 18 4 4

7-45 5-60 5-60 30-3600 30-1800

74-90 54-90 25-77 1-44 0.5-18

kexpt (s-1) 3.6 × 10-2 2.9 × 10-2 2.3 × 10-2 2.0 × 10-4 1.1 × 10-4

ments was calculated as follows,

[H+] )

Kw Kγ,H2O

x

KNaOH [NaOH] Kγ,NaOH

(9)

where Kγ,H2O and Kγ,NaOH are the ratio of activity coefficients in the equilibrium expressions for H2O and NaOH dissociation, respectively. A summary of the experiments, with calculated [H+] and pHs, is shown in Table 3. It is important to notice that the initial conversion measured increased greatly at pH values 221 bar), the unimolecular decomposition rate sets the lower limit on the total measured reaction rate. As the concentration of H+ decreased, the relative importance of this “minor pathway” increased and became the dominant mechanism of decomposition. By the addition of excess base, the rate of acid-catalyzed pathways could be slowed to the point where gas-phase reaction pathways could be measured. Such limiting behavior verified the applicability of using gas-phase kinetic expressions to model reactions in supercritical water. Conclusions Reaction pathways of MTBE decomposition in suband supercritical water were analyzed to establish the importance of acid-catalyzed hydrolysis in comparison to unimolecular decomposition. Ab initio computational tools were employed to correct the activation energy of the gas-phase unimolecular decomposition pathway. Although correcting for solvent effects decreased the activation barrier from the value in the gas phase, the pathway did not contribute to the overall reaction rate

8


for temperatures

Multiscale Reaction Pathway Analysis of Methyl tert-Butyl Ether

Recommend Documents