Comparison of Unimolecular Decomposition Pathways for Carboxylic

Article pubs.acs.org/JPCA

Comparison of Unimolecular Decomposition Pathways for Carboxylic Acids of Relevance to Biofuels Jared M. Clark, Mark R. Nimlos, and David J. Robichaud* National Bioenergy Center, National Renewable Energy Laboratory, Golden, Colorado 80401, United States S Supporting Information *

ABSTRACT: Quantum mechanical molecular modeling is used [M06-2X/6-311+ +G(2df,p)] to compare activation energies and rate constants for unimolecular decomposition pathways of saturated and unsaturated carboxylic acids that are important in the production of biofuels and that are models for plant and algae-derived intermediates. Dehydration and decarboxylation reactions are considered. The barrier heights to decarboxylation and dehydration are similar in magnitude for saturated acids (∼71 kcal mol−1), with an approximate 1:1 [H2O]/[CO2] branching ratio over the temperature range studied (500−2000 K). α,β-Unsaturation lowers the barrier to decarboxylation between 2.2 and 12.2 kcal mol−1 while increasing the barriers to dehydration by ∼3 kcal mol−1. The branching ratio, as a result, is an order of magnitude smaller, [H2O]/[CO2] = 0.07. For some α,β-unsaturated acids, six-center transition states are available for dehydration, with barrier heights of ∼35.0 kcal mol−1. The branching ratio for these acids can be as high as 370:1. β,γ-Unsaturation results in a small lowering in the barrier height to decarboxylation (∼70.0 kcal mol−1). β,γ-Unsaturation also leads to a lowering in the dehydration pathway from 1.7 to 5.1 kcal mol−1. These results are discussed with respect to predicted kinetic values for acids of importance in biofuels production.

■

INTRODUCTION Carboxylic acids are important in the processing of biomass into liquid fuels and value-added chemicals. Acetate groups on hemicellulose biopolymers are released during acid pretreatment and pyrolysis to form acetic acid, shown below.1 Acids are also formed from the decomposition of sugars, which are released during deconstruction of polysaccharides, and they can be formed from the thermal decomposition of polyhydroxyalkanoates, such as polyhydroxybutyrate shown below, carbon storage polymers for many micro-organisms.2−4 Finally, fatty acids are an important potential source of renewable carbon from algae,5−9 bacteria,10 oil crops,11 and animal fats.12,13 Examples of such organic acids are shown in Figure 1. The thermal deoxygenation of carboxylic acids is an important step in the conversion of biomass into aliphatic hydrocarbons suitable for use in biofuels and as petrochemical replacements. Decarboxylation, a primary decomposition pathway under pyrolysis conditions, represents an ideal conversion process for this purpose because it eliminates two atoms of oxygen for every carbon atom removed. The dehydration of organic acids represents a direct pathway to ketenes (RR′CCO). The major industrial use of ketene is in the production of acid anhydrides and diketene derivatives, which are themselves important precursors in the synthesis of a wide range of valueadded chemicals.16−18 Ketenes are also used in the preparation of β-lactones and β-lactams via [2 + 2] cycloaddition reactions with ketones and imines, respectively.19 Thus, it is critical to understand the branching ratios between decarboxylation and dehydration. The unimolecular decomposition of carboxylic acids occurs through two competing mechanisms, decarboxylation, leading to the formation of aliphatic hydrocarbons and carbon dioxide, © 2013 American Chemical Society

reaction 1, and dehydration, leading to the formation of water and a ketene derivative, reaction 2. Reactions 1 and 2 are demonstrated below for acetic acid.20−23

For acids containing α-hydrogens, dehydration may proceed through a concerted reaction mechanism, reaction 2, with a single transition state, or through a two-step process in which a hydrogen atom transfers from the α-carbon to the carbonyl oxygen to form an 1,1-enediol, as shown in reaction 3a. This reaction can then be followed by the elimination of a water molecule via a four-centered pericyclic transition state, reaction 3b, to form a ketene derivative.

Received: September 24, 2013 Revised: December 2, 2013 Published: December 2, 2013 260

dx.doi.org/10.1021/jp4095485 | J. Phys. Chem. A 2014, 118, 260−274

The Journal of Physical Chemistry A

Article

measured the same pathways in a shock tube at 1300−1950 K and measured the activation energies of the two pathways to be equal, having a value of 72.7 kcal mol−1. Theoretical work by Duan and Page26 in the mid-1990’s reported activation energies of 71.8, 76.4, and 73.1 kcal mol−1 for reactions 1−3, respectively, performed at the CASSCF 10-in-10/6-31G* level. These relatively high barriers to decomposition are due to the strain in the four-membered transition states involved in each process, as shown above in reactions 1−3. The presence of α,β- and β,γ-unsaturation in organic acids may allow for mechanisms involving larger, less strained geometries (i.e., six-membered transition states). Starting in the early 1950’s, work by Arnold et al.27−30 demonstrated that β,γ-unsaturated acids thermally decompose almost exclusively through decarboxylation. In 1964, Smith and Blau28 measured the activation energy for the decarboxylation of vinylacetic acid [CH2CHCH2C(O)OH], the simplest β,γ-unsaturated acid, in a closed reactor, for which they reported a value of 39.3 kcal mol−1, markedly lower (∼30 kcal mol−1) than that for formic or acetic acid. In the following year, Bigley and Thurman30 measured a value of 37.7 kcal mol−1 for the decarboxylation of 2,2-dimethyl-4-phenylbut-3-enoic acid. In subsequent work, Bigley31,32 reported the activation energies for a series of substituted vinylacetic acid analogues with values ranging from 28.8 to 39.3 kcal mol−1. The lower activation energies of these β,γ-unsaturated acids were first explained by Arnold et al.27 to be the result of a six-centered pericyclic transition state, as shown in reaction 4 for vinylacetic acid.

Figure 1. Natural carboxylic acids of interest in the conversion of biomass to liquid fuels. Bond energies for acetic acid are provided14,15 for later comparison with unimolecular decomposition pathways.

Similarly, the dehydration of the α,β-unsaturated acids possessing conveniently situated γ-hydrogens (i.e., isocrotonic acid, Z-CH3CHCHC(O)OH) are also able to pass through a six-centered transition state, either directly to form water and a ketene derivative (reaction 5) or indirectly through the formation of a 1,1-enediol intermediate (reaction 6). These six-centered reactions are illustrated below for isocrotonic acid.

While reactions 1−3 are in competition with each other, it is important to note that the enthalpy of reaction (ΔrH) for decarboxylation is negative (exothermic), while those for dehydration are positive (endothermic). For example, ΔrH for the decarboxylation of acetic acid is −7.3 kcal mol−1, while that for dehydration is 35.2 kcal mol−1, a difference of 42.5 kcal mol−1. This would indicate that at elevated temperatures, reactions 2−3 may become significant in determining the product branching ratio for the thermal decomposition of organic acids. Decarbonylation, the direct loss of CO and the formation of a Cn−1 alcohol, represents a third decomposition pathway for organic acids. While this pathway may represent a major decomposition pathway for organic acids in the presence of heterogeneous catalysts, our calculated gas phase barrier heights, 93−109 kcal mol−1, indicate that decarbonylation, although accessible, is not likely to be a competitive mechanism and will not be discussed further. Experimental and theoretical work involving the study of these pathways has largely been limited to small organic acids (i.e., formic and acetic). Acetic acid [CH3C(O)OH] has been investigated experimentally since the 1940’s. Bamford and Dewar24 first measured the activation energies for the competing decarboxylation and dehydration pathways in a flow reactor at 1068−1218 K and found them to be 62.0 and 67.5 kcal mol−1, respectively. Later, Mackie and Doolan25

The importance of six-centered transition states and the associated lower reaction barriers in the decomposition of organic acids cannot be overstated. In those acids for which they are available, they could be dominant unimolecular decomposition pathways. The work presented herein considers the competing decomposition pathways (reactions 1−6) for the family of possible saturated and unsaturated C2−C4 acids in order to estimate branching ratios between the decarboxylation and dehydration pathways. The activation energies and 261



Article

of the transition state, QAcid is the molecular partition function of the acid, ΔE0 is the difference between E0 (E0 = Eelect + EZPVE) of the transition state and reactants, R is the ideal gas constant, and κ(T) is the asymmetric Eckart tunneling factor.42−45 The partition functions, Q, as used here, are composed of translational (qtrans), electronic (qelect), rotational (qrot), and vibrational (qvib) components, as shown in eq 2.

predicted kinetic values for acids of importance to biofuels are presented, as well as guidelines toward the prediction of the fate of a chosen organic acid based on the presence of unsaturation proximate to the carboxylic acid functionality.

■

COMPUTATIONAL METHODS The Gaussian09 suite of programs33 was used to perform all of the electronic structure calculations described in this paper. The energies and structures of all species were calculated and fully optimized through the use of the hybrid meta-GGA functional M06-2X of Truhlar and Zhao34 with the 6-311++G(2df,p) basis set as well as at the B3LYP/CBSB7 level as part of the complete basis set (CBS) extrapolation method, CBS-QB3.35,36 Convergence was guaranteed by the use of tight convergence criteria and ultrafine pruned (99,590; 99 radial shells and 590 angular points per shell) integration grids. The M06-2X method has been shown to perform exceptionally well against databases involving thermochemistry (TC177), diverse barrier heights (DBH76), and noncovalent interaction energies (NCIE53), with mean errors of 1.3, 1.2, and 0.37 kcal mol−1, respectively.37 The CBS-QB3 method by Peterson and coworkers35,36 provides accurate results for energetics and thermochemistry, with a standard error that is within ∼1 kcal mol−1 for most species described in the G2/97 test set, and has been shown by several groups to provide accurate results for transition-state calculations.38−40 The robust systematic search (RSS) technique of Clark et al.41 was used to optimize all reactant and product conformations employing the M06-2X/6-311++G(2df,p) method. The optimized geometries resulting from the RSS search were used as input for the CBS-QB3 complex energy calculation. The RSS method searches the potential energy surface of a molecule via systematic bond dihedral rotations and geometry optimizations. These dihedral rotations result in a total of nm possible optimizations, where m is the number of bond dihedrals scanned and n is the number of unique positions (bond rotations) for each bond. The molecular structures of each carboxylic acid and the appropriate reaction products discussed in this work were generated by rotating each C−C or C−O bond dihedral by 60°. All transition states were verified as first-order saddle points via the presence of a single imaginary frequency, which corresponds to the motion along the reaction coordinate. Intrinsic reaction coordinate (IRC) calculations with a step size of 0.1 amu0.5 bohr were performed to verify the connection of each transition state with the appropriate reactants, intermediates, and/or products. Vibrational frequencies were modeled using the rigid rotor harmonic oscillator (RRHO) approximation. Lowfrequency vibrations corresponding to internal rotations were omitted from the RRHO analysis and treated as 1-D hindered rotors where appropriate. Potential energy scans about the rotors were also carried out at the appropriate levels of theory for each method used to determine the barrier to rotation, the rotational symmetry, and the number of rotational minima. The reaction rate constants for the various decomposition pathways of each acid were calculated over a range of temperatures, from 500 to 2000 K in 50 K increments, using conventional transition state theory (TST), as detailed in eq 1 k(T ) = κ(T )

kBT Q‡ −ΔE0 / RT e h Q Acid

Q = qtransqelecqrotqvib

(eq 2)

TST corrected for tunneling effects has been successfully used previously to calculate the rate coefficients for hydrogen abstraction reactions.46−51 Partition functions were corrected for 1-D hindered rotation, where appropriate, using the method outlined by Pfaendtner et al.52 by means of code developed internally53 (written in the Python programming environment). All of the kinetics presented in this work were determined at the M06-2X/6-311++G(2df,p) level of theory. The calculated rate constants, k(T), were then fitted to a standard Arrhenius expression, eq 3, to obtain the kinetic rate parameters A and Ea

k(T ) = A e−Ea / RT

(eq 3)

The calculated rate constants for the different decomposition pathways were then used to simulate the overall decomposition of each organic acid. The branching ratio [H2O]/[CO2] was determined at each temperature as a measure of the propensity of a given acid to undergo decomposition or dehydration as a function of temperature. This information becomes useful when developing specific strategies for the conversion of acids into hydrocarbon targets or into synthetically important ketene intermediates. Kinetic simulations were performed using Python. The performance of the employed computational methods in describing four-center and six-center proton-transfer transition states was benchmarked against two well-known chemical decompositions, the dehydration of ethanol (reaction 7) and the decomposition of ethyl acetate (reaction 8), which proceed through four- and six-centered transition states, respectively.

Table 1 collects the activation energies and pre-exponential factors for ethanol and ethyl acetate thermolysis. The activation energy for the dehydration of ethanol was calculated at the M06-2X/6-311++G(2df,p) level of theory to be 67.2 kcal mol−1, with a pre-exponential factor of 4.03 × 1013 s−1. As seen in Table 1, the calculated values for ethanol dehydration are in good agreement with the experimentally determined values reported by Tsang et al.54 and by Marinov et al.55 The computed activation energy is within 1.5 kcal mol−1 of the experimental values, and the pre-exponential factors are within a factor of 4 of experimental measurements. Similar treatment, vis-à-vis the decomposition of ethyl acetate, results in calculated rate parameters of 50.2 kcal mol−1 and 3.95 × 1012 s−1 for the activation energy and pre-exponential factor, respectively. Again, these values are in excellent agreement with the experimental data of McMillen et al.56 and Keller et al.,57

(eq 1)

where kB is Boltzmann’s constant, T is temperature in Kelvin, h is Planck’s constant, Q‡ and is the molecular partition function 262



Article

M06-2X/6-311++G(2df,p) computed values, 0.9 ± 0.6 kcal mol−1, and are not included in Table 2. For reference, the geometric differences between M06-2X/6-311++G(2df,p) and CBS-QB3 calculated structures were determined to be 2 > 1 placement of methyl groups: β > α Whereas α,β-unsaturation works to lower the barrier to decarboxylation, it has the opposite effect on the barriers to dehydration. For concerted dehydration, the number of methyl groups bound to the α,β-double bond affects the extent to which the barrier to concerted H2O production is raised. For acrylic acid, the barrier increases by 7.5 kcal mol−1. The addition of a methyl group to form crotonic acid and isocrotonic acid increases the barrier height to 5.7 kcal mol−1, 1.8 kcal mol−1 less than the increase for acrylic acid. This trend is continued by a second methyl group that increases the barrier by 4.4 kcal mol−1. Interestingly, the presence of additional methyl groups has no apparent effect on the formation of the enediol intermediate in the stepwise loss of H2O. The average barrier height increase for all of the α,β-unsaturated acids capable of dehydration is 3.7 kcal mol−1. β,γ-Unsaturated Acids. β,γ-Unsaturation results in only a minor reduction in the average decarboxylation barrier height, ∼2.2 kcal mol−1. This would imply that the effect of this type of unsaturation on decarboxylation will be small. β,γ-Unsaturation affects the dehydration of acids differently for reactions 2 and 3. For reaction 2 (concerted dehydration), the average barrier height is calculated to be 71.3 ± 1.3 kcal mol−1. For reaction 3, the reduction in barrier height is dependent on whether the unsaturation is a double or triple bond and whether α-methyl branching is present. For these acids, the reduction in barrier height is smaller for acids with α-methyl branching than that for those without. The barrier heights for the decomposition of γ,δ-unsaturated acids (i.e., allylacetic acid, CH2CHCH2CH2C(O)OH) are similar to those for the saturated acids, and it is likely that unsaturation beyond the β,γ-position will not play a significant role in affecting the reaction dynamics of the thermal decomposition of carboxylic acids. Unsaturated fatty acids, for example, with unsaturation beyond the β,γ-position, will likely thermally decompose as if they were saturated acids, resulting in a ∼50/50 production of H2O and CO2, assuming the absence of radical formation. Kinetics. Kinetic modeling regarding the gas-phase, thermal decomposition of carboxylic acids presented in this section excludes those reactions leading to the formation of radical species. The most facile homolytic fission of acetic acid, as a 267



Article

Table 4. Calculateda Activation Energies (kcal mol−1) and Pre-Exponential Factors (s−1) for Decarboxylationb and Dehydrationc Reactions, Expressed as Ea(A) reaction 2

reaction 3a

reaction 3bd

concerted dehydration

enediol formation

enediol dehydration

reaction 1 acid Saturated CH3C(O)OH CH3CH2C(O)OH CH3CH2CH2C(O)OH CH3CH(CH3)CH2C(O)OH CH3C(CH3)2CH2C(O)OH CH3CH(CH3)C(O)OH CH3CH2CH(CH3)C(O)OH CH3CH(CH3)CH(CH3)C(O)OH CH3C(CH3)2CH(CH3)C(O)OH CH3C(CH3)2C(O)OH CH3CH2C(CH3)2C(O)OH CH3CH(CH3)C(CH3)2C(O)OH CH3C(CH3)2C(CH3)2C(O)OH α,β-Unsaturated Alkenoic CH2CHC(O)OH CH2C(CH3)C(O)OH E-CH3CHCHC(O)OH Z-CH3CHCHC(O)OH E-CH3CHC(CH3)C(O)OH Z-CH3CHC(CH3)C(O)OH CH3C(CH3)CHC(O)OH CH3C(CH3)C(CH3)C(O)OH Z-CH3CHCHC(O)OHf Z-CH3CHC(CH3)C(O)OHf CH3C(CH3)CHC(O)OHf CH3C(CH3)C(CH3)C(O)OHf Alkynoic CHCC(O)OH CH3CHCC(O)OH β,γ-Unsaturated Alkenoic CH2CHCH2C(O)OH CH2C(CH3)CH2C(O)OH CH2CHCH(CH3)C(O)OH CH2C(CH3)CH(CH3)C(O)OH CH2CHC(CH3)2C(O)OH CH2C(CH3)C(CH3)2C(O)OH CH2CHCH2C(O)OHf CH2CHCH(CH3)C(O)OHf CH2C(CH3)CH2C(O)OHf CH2CHC(CH3)2C(O)OH CH2C(CH3)CH(CH3)C(O)OHf CH2C(CH3)C(CH3)2C(O)OHf Alynoic CHCCH2C(O)OH CHCCH(CH3)C(O)OH CHCC(CH3)2C(O)OH CHCCH2C(O)OHf CHCCH(CH3)C(O)OHf CHCC(CH3)2C(O)OHf γ,δ-Unsaturated CH2CHCH2CH2C(O)OH

decarboxylation

e

70.3 71.4 70.4 70.6 71.6 70.6 70.3 72.3 70.7 72.1 71.9 74.1 72.5

(3.10 (2.24 (1.70 (5.11 (6.82 (2.39 (2.64 (1.52 (2.38 (3.21 (2.68 (6.64 (7.63

× × × × × × × × × × × × ×

1013) 1013) 1013) 1013) 1013) 1013) 1013) 1013) 1013) 1013) 1013) 1013) 1013)

75.7 74.9 74.8 74.7 73.7 77.0 76.7 77.6 76.6

68.7 68.1 66.7 65.4 65.3 64.1 63.4 60.1

(3.96 (4.07 (6.78 (1.45 (4.77 (6.13 (1.40 (1.24

× × × × × × × ×

1013) 1013) 1013) 1013) 1013) 1013) 1014) 1014)

83.2 (3.07 × 1014)

74.9 (4.91 × 1013)

45.6 (6.99 × 1012)

81.7 (6.09 × 1014) 80.8 (9.06 × 1013)

74.1 (1.00 × 1014) 74.4 (1.07 × 1013)

44.9 (8.54 × 1012) 45.0 (6.89 × 1012)

79.9 (5.46 × 1014)

74.0 (4.70 × 1013)

44.3 (2.53 × 1013)

(4.93 (6.91 (5.08 (1.06 (7.66 (6.44 (6.16 (4.84 (1.18

× × × × × × × × ×

1013) 1013) 1013) 1014) 1013) 1013) 1013) 1013) 1014)

68.9 69.4 69.1 69.3 68.6 72.0 71.9 72.7 72.5

(5.54 (1.54 (1.04 (1.95 (1.46 (1.07 (1.00 (4.77 (8.09

× × × × × × × × ×

1012) 1013) 1013) 1013) 1013) 1013) 1013) 1012) 1012)

57.0 58.2 57.7 56.4

(1.35 (6.56 (1.15 (2.77

× × × ×

1012) 1012) 1013) 1012)

34.6 35.3 34.6 32.9

(4.36 (2.68 (3.67 (1.25

× × × ×

1011) 1012) 1012) 1012)

(2.87 (3.84 (1.30 (3.19

× × × ×

1013) 1013) 1013) 1013)

64.0 65.0 66.1 67.3

(8.54 (1.35 (3.32 (7.92

× × × ×

1012) 1013) 1012) 1012)

45.1 45.1 45.0 44.0 44.0 44.7 44.5 45.4 42.5

(1.04 (9.02 (4.02 (1.64 (2.30 (6.40 (3.05 (4.90 (1.13

× × × × × × × × ×

1013) 1012) 1012) 1013) 1013) 1012) 1013) 1012) 1013)

59.5 (1.14 × 1014) 58.7 (1.25 × 1014)

69.3 69.9 69.2 69.3 70.1 71.0 39.1 39.2 36.2 37.2 34.3 34.5

(3.18 (5.49 (2.40 (3.56 (2.78 (3.09 (1.81 (1.58 (7.59 (2.58 (6.01 (4.05

× × × × × × × × × × × ×

1013) 1013) 1013) 1013) 1013) 1013) 1012) 1012) 1012) 1012) 1012) 1012)

70.3 70.8 72.1 72.7

67.9 68.0 67.7 38.0 36.7 35.5

(6.76 (5.38 (3.43 (4.46 (4.08 (3.52

× × × × × ×

1013) 1013) 1013) 1012) 1012) 1012)

71.2 (4.97 × 1013) 73.9 (7.64 × 1013)

65.4 (9.80 × 1012) 64.6 (1.56 × 1013)

47.7 (1.50 × 1013) 42.2 (3.47 × 1012)

70.8 (2.83 × 1013)

74.9 (1.01 × 1014)

69.1 (1.50 × 1013)

45.1 (9.14 × 1012)

45.5 44.0 44.9 42.4

(1.03 (8.31 (5.38 (8.56

× × × ×

1013) 1012) 1012) 1012)

a

Calculated using M06-2X/6-311++G(2df,p). bReaction 1, except where noted. cReactions 2 and 3, except where noted. dBarrier height relative to the enediol intermediate. eBarrier height relative to the starting carboxylic acid. fSix-centered transition state, reactions 4−6. 268



Article

Table 5. Kinetic Model for Butyric Acid, CH3CH2CH2C(O)OH

reaction # 1 2 3a 3b

equation acid ↔ acid ↔ acid ↔ enediol

CO2 + hydrocarbon H2O + ketene enediol ↔ H2O + ketene

Table 6. Kinetic Model for β,γ-Unsaturated Alkynoic Acids for But-3-ynoic Acid, CHCCH2C(O)OH

Figure 4. Kinetic model for the thermal decomposition of trimethylacrylic acid (A) and 2,3-dimethylvinylacetic acid (B). Barrier heights for each reaction are shown in parentheses and are given in kcal mol−1.

Scheme 1

reaction # 1 2 3 3b 4a a

equation acid ↔ acid ↔ acid ↔ enediol acid ↔

CO2 + hydrocarbon 1 H2O + ketene enediol ↔ H2O + ketene CO2 + hydrocarbon 2

and geminal enediol (68.3 kcal mol−1), as trimethylacrylic acid, albeit via 1,3-(C,O)-hydrogen shift mechanisms. Senecioic Acid/3-Methylvinylacetic Acid. Figure 5 depicts the kinetic model that involves senecioic acid (CH3C(CH3)CHC(O)OH) and 3-methylvinylacetic acid (CH2C(CH3)CH2C(O)OH). Similar to the model in Figure 4, this model includes the interconversion of the two acids through a common 1,1-enediol intermediate, Scheme 1, 2-ethylene-2-methyleneethenediol (CH2C(CH3)CHC(OH)2), accessible via a 1,3-(O,C)-hydrogen shift from 3-methylvinylacetic acid (65.8 kcal mol−1) and a 1,5-(O,C)-hydrogen shift from senecioic acid (34.9 kcal mol−1). The common ketene product, 2-ethylene-2-methylketene (CH2C(CH3)CHC(O)), is attainable through similar H-shifts and the loss of a water molecule (70.1 and 56.7 kcal mol−1, respectively). The dehydration of senecioic acid through 1,3(O,C)-hydrogen shift mechanisms leads to the formation of the geminal enediol, allenediol (C(CH3)2CC(OH)2, 73.6 kcal mol−1), reaction 3, and alleneketene (C(CH3)2CC(O), 78.2 kcal mol−1), reaction 2, products. The dehydration of senecioic acid represents one of the major differences between the model in Figures 4 and 5. In Figure 4, it can be seen that trimethylacrylic acid does not possess 1,3-(C,O)-hydrogen shift

Pathway involving a six-center transition state.

molecule. Geminal enediol formation (33.1 kcal mol−1), produces 2-isopropylene-2-methylethenediol (CH2C(CH3)C(CH3)C(OH)2). With such a large difference in barrier height, geminal enediol formation is the most likely reaction pathway for this acid. Once formed, the geminal enediol can either dehydrate to form 2-isopropylene-2-methylketene (42.3 kcal mol−1), revert to trimethylacrylic acid (9.0 kcal mol−1), or convert to 2,3dimethylvinylacetic acid through a 1,3-(C,O)-hydrogen shift over a barrier of 47.0 kcal mol−1. The interconversion of α,β- and β,γunsaturated acids through a common geminal enediol via 1,5-(C,O)and 1,3-(C,O)-hydrogen shifts, respectively, is characteristic of such acids, as seen in Scheme 1. 2,3-Dimethylvinylacetic can decarboxylate via either 1,3-(C,O)(69.0 kcal mol−1) or 1,5-(C,O)-hydrogen (34.0 kcal mol−1) shifts, reactions 1 and 4, to produce 2-methyl-1-butene (CH2C(CH3)CH2CH3) and 2-methyl-2-butene, respectively, and CO2. The 1,5-(C,O)-hydrogen shift with loss of CO2 is typical of all β,γ-unsaturated acids and represents the major unimolecular decomposition channel for such acids. 2,3-Dimethylvinylacetic acid shares the same dehydration products, ketene (72.1 kcal mol−1) 269



Article

Figure 5. Kinetic model for the thermal decomposition of senecioic acid (A) and 3-methylvinylacetic acid (B). Barrier heights for each reaction are shown in parentheses and are given in kcal mol−1.

Figure 6. Kinetic model for the thermal decomposition of tiglic acid (A), angelic acid (B), and 2-methylvinylacetic acid (C). Barrier heights for each reaction are shown in parentheses and are given in kcal mol−1.

decomposition pathways as it has no α-hydrogens. The 1,3(O,C)-H shift decarboxylation of either acid, reaction 1, leads to the loss of CO2 and the formation of isobutylene (CH2C(CH3)2), 62.7 kcal mol−1 for senecioic acid and 69.7 kcal mol−1 for 3-methylvinylacetic acid. An additional decarboxylation pathway is available to 3-methylvinylacetic acid, a 1,5-(O,C)-hydrogen shift. This pathway leads to the formation of isobutylene as well but lies much lower in energy, 33.7 kcal mol−1, a 48% reduction in barrier height compared to the 1,3-(O,C)-H shift mechanism. Tiglic Acid/Angelic Acid/2-Methylvinylacetic Acid. Figure 6 represents the kinetic model used for the subset of acids that includes tiglic acid (E-CH3CHC(CH3)C(O)OH), angelic acid (Z-CH3CHC(CH3)C(O)OH), and 2-methylvinylacetic acid (CH2CHC(CH3)C(O)OH). In addition to the isomerization of Scheme 1, in which angelic acid interconverts via the geminal 2-ethylene-2methylethenediol (CH2CHC(CH3)C(OH)2), additional higher-energy isomerization possibilities exist to interconvert the acids in the model found in Figure 6. The first of these involves the two α,β-unsaturated acids, which are interconverted via a rotation of the α,β-double bond. The transition state to double bond rotation is described by a singlet, diradicaloid structure in which the two electrons of the breaking π bond occupy different orbitals but are antiferromagnetically coupled. During the rotation, the double bond elongates slightly as it passes through an energy maximum at a C0−Cα−Cβ−Cγ dihedral of 90°. As a measure of the reliability of the M06-2X functional to describe such a transition state, the transition state for the

double bond rotation of 2-butene was modeled. The rotational barrier for the isomerization of trans-2-butene to cis-2-butene is calculated to be 62.4 kcal mol−1, which is in excellent agreement with experiment (61.9−62.9 kcal mol−1).69,70 The barrier to the reverse process is 61.2 kcal mol−1. The difference of 1.3 kcal mol−1 in the barrier height indicates that the trans isomer is generally more stable than the cis isomer. The isomerization of tiglic acid and angelic acid has been calculated to be 55.8 kcal mol−1 from tiglic acid and 54.3 kcal mol−1 from angelic acid. As with 2-butene, the barrier to isomerization is higher for the trans isomer as the trans isomer is the more stable of the two by 1.5 kcal mol−1. The second isomerization is between the trans-α,βunsaturated acid and the β,γ-unsaturated acid through a 1,3(C,O)-hydrogen shift. Both isomerization barriers are high (79.8 and 85.7 kcal mol−1), but that faced by the trans-α,βunsaturated isomer is higher as the displacement of the double bond to the β,γ-position breaks the conjugation of the α,β-double bond with the carbonyl. This isomerization pathway is not likely to play a large role but has been included for completeness. In the absence of isomerization, the only decomposition route available to tiglic acid is a 1,3-(C,O)-hydrogen shift and the loss of a CO2, reaction 1, to form cis-2-butene. Isomerization to angelic acid, which has a lower barrier height than the decarboxylation, will result in the production of water through the dehydration of angelic acid via direct conversion to 2-ethylene-2-methylketene (CH2CHC(CH3)C(O)) and water through a 1,5-(C,O)-hydrogen shift or through a similar 270



Article

hydrogen shift mechanism to form the 1,1-enediol, 2-ethylene2-methylethenediol. The latter product can either dehydrate through a 1,3-(C,O)-hydrogen shift with the loss of a water molecule, convert back to angelic acid, or become 2-methylvinylacetic acid through a 1,3-(C,O)-hydrogen shift. Both angelic and 2methylvinylacetic acid can decarboxylate to form trans-2-butene, the former through a 1,3-(C,O)-hydrogen shift and the later via a 1,5-(C,O)-hydrogen shift. 2-Methylvinylacetic acid may also decarboxylate in a 1,3-(C,O) fashion to form 1-butene and CO2 or dehydrate in a concerted 1,3,(C,O)-hydrogen shift and loss of a water to form 2-ethylene-2-methylketene. Crotonic Acid/Isocrotonic Acid/Vinylacetic Acid. The thermal decomposition of crotonic acid (trans-CH3CHCHC(O)OH), isocrotonic acid (cis-CH3CHCHC(O)OH), and vinylacetic acid (CH2CHCH2C(O)OH) composes the final model studied in this work. This model has been the subject of a number of experimental studies by Bigley et al.32,67 Bigley’s conclusions regarding the isomerization relationship among the three acids, Scheme 2, of this model are consistent with the barrier heights shown in Figure 7. Scheme 2

Pressure−time curves obtained via pyrolysis experiments by Bigley, measuring the isomer distributions obtained by starting from each pure acid, indicate the validity of Scheme 2. Bigley rightly suggested that direct conversion of crotonic acid to vinylacetic acid does not occur at moderate temperatures (∼600 K). The barrier heights shown in Figure 7 associated with crotonic acid ↔ vinylacetic acid conversion are much higher (∼15 kcal mol−1) than the barriers to the isomerization of crotonic acid to isocrotonic acid. The results of these simulations are summarized in Figure 8, which depicts the branching ratio, [H2O]/[CO2], as a function of temperature. The branching ratio for formic acid (HC(O)OH, red circles) shows the expected behavior as referenced against experiment71,72 in that the rate of dehydration increases with increasing temperature. The branching ratio resulting from the thermal decomposition of formic acid has an average value of 6.3 above 1300 K, consistent with the experimentally determined value of 10, as reported by Hsu et al.59 Butyric acid (CH3CH2CH 2C(O)OH, blue circles) is representative of the saturated acids studied in this work and with acids containing unsaturation beyond the β,γ-position. These acids have a [H2O]/[CO2] of ∼1 that seems to show little temperature dependence. Saturated fatty acids and the majority of unsaturated fatty acids fall into this group. As these materials are of interest for the direct formation of liquid fuels from biosourced feed stock, pyrolysis in the absence of catalysts will produce a ∼50:50 mixture of Cn−1 hydrocarbons and ketene products, an undesirable outcome. Crotonic acid (CH3CHCHC(O)OH, green squares) is typical of α,β-unsaturated acids that do not have six-centered dehydration transition states. As the data in Figure 8 indicate, the primary decomposition pathway for crotonic-like organic acids is the loss of CO2, with little water produced, even at elevated temperatures (2000 K). This can be an important consideration as the hydrocarbon products of crotonic-like

Figure 7. Kinetic model for the thermal decomposition of crotonic acid (A), isocrotonic acid (B), and vinylacetic acid (C). Barrier heights for each reaction are shown in parentheses and are given in kcal mol−1.

acids are alkenes, useful intermediates in petrochemical and other synthetic processes. Vinylacetic acid (CH2CHCH2C(O)OH, yellow diamonds) reflects all of the β,γ-unsaturated acids studied. Similar to crotonic-like organic acids, vinylacetic acid and its derivatives produce alkenes almost exclusively. This is attractive as the production of vinylacetic and crotonic acids from glycerol has recently been reported.73 Glycerol represents an important source of carbon in the conversion of biomass to liquid fuels via the production of alkenes. Glycerol is produced directly from the hydrolysis of glycerides and represents a major byproduct of biodiesel production, accounting for approximately 1 kg of byproduct for every 9 kg of biodiesel produced. The method of conversion of glycerol reported in ref 66 is applicable to higher polyols (CnHn+2(OH)n), opening the door to the production of larger and possibly more branched alkenes. The α,β-unsaturated acids with methyl groups cis to the carboxylic acid functional group are plotted separately in Figure 8 as the associated branching ratios vary significantly. At low to moderate temperatures, these acids produce large amounts of H2O and conjugated ketenes. In principle, conjugated ketenes can undergo both [2 + 2], typical of most ketenes, and [4 + 2] cycloadditions, in the latter case participating as the diene.74 271



Article

acids in order to describe the importance and relationship of the competing decarboxylation and dehydration mechanisms and to understand the impact of unsaturation and branching on those mechanisms. Our results show that, within the unimolecular decomposition regime, decarboxylation and dehydration are equally likely for acetic acid and all saturated acids, over the temperature range studied, regardless of chain length or branching. The [H2O]/[CO2] branching ratio for these acids ranges from 0.6 at 500 K to near unity at temperatures over 1500 K. These results imply that at low acid concentration, in the absence of bimolecular mechanisms, a catalyst must be employed if one wishes to favor one potential decomposition outcome over another. For example, mononuclear [MO3(OH)]− and binuclear [M2O6(OH)]− oxo-anions (M = Mo, W)23 have been developed to selectively favor the formation of ketene (dehydration) from acetic acid to the exclusion of methane (decarboxylation). This work has also shown that the presence and location of unsaturation may play an important role in directing the decomposition of organic acids. In most instances, α,βunsaturation favors the loss of CO2, as is the case for crotonic acid. If six-centered transition states are available, α,βunsaturation leads almost exclusively to dehydration, especially at temperatures below 1000 K. β,γ-Unsaturation involving fourcentered transition states is shown to have little effect on decarboxylation but tends to enhance to routes to the production of H2O. However, this enhancement is found to be inconsequential in so much as all of the β,γ-unsaturated acids studied have six-centered decarboxylation pathways open to them, resulting in the predominance of the decarboxylation mechanism for these types of acids. These results indicate that molecules, such as unsaturated fatty acids, that have unsaturation close to the carboxylic acid, will predominately decompose to produce CO2 and the relevant hydrocarbon. However, if there is a β-methyl group cis to the carboxylic acid group, then dehydration will be favored, especially at temperatures below 1000 K. Unsaturation past the γ-position will have no result on the outcome of thermal decomposition of unsaturated fatty acids, within the scope of the reactions described in this work. Such acids will behave as though they are saturated molecules.

Figure 8. Simulated branching ratios, [H2O]/[CO2], for the series of acids represented in this work. Butyric acid (CH3CH2CH2C(O)OH, blue circles) is representative of the saturated acids and acids with unsaturation beyond the β,γ-position. Crotonic acid (CH3CHCHC(O)OH, green squares) is typical of α,β-unsaturated acids that do not have six-centered dehydration transition states. Vinylacetic acid (CH2CHCH2C(O)OH, yellow diamonds) reflects all of β,γ-unsaturated acids studied. The α,β-unsaturated acids with methyl groups cis to the carboxylic acid functional group are plotted separately as the associated branching ratios vary significantly.

The production of H2O for these acids via 1,3-(C,O)-hydrogen shifts will be a minor decomposition pathway for these acids, especially at lower to moderate temperatures (

Comparison of Unimolecular Decomposition Pathways for Carboxylic

Recommend Documents