Article pubs.acs.org/JPCA
Formation of Furan along with HO2 during the OH-Initiated Oxidation of 2,5-DHF and 2,3-DHF: An Experimental and Computational Study Hariprasad D. Alwe, Asmita Sharma, Mohini P. Walavalkar, Suresh Dhanya,* and Prakash D. Naik Radiation and Photochemistry Division, Bhabha Atomic Research Centre, Trombay, Mumbai, 400 085, India S Supporting Information *
ABSTRACT: Experimental characterization of products during OH-initiated oxidation of dihydrofurans (DHF) confirms the formation of furan accompanied by the formation of HO2 to be a significant channel in 2,5-DHF (21 ± 3%), whereas it is absent in 2,3-DHF. Theoretical investigations on the reaction of OH with these molecules are carried out to understand this difference. All possible channels of reaction are studied at M06-2X level with 6-311G* basis set, and the stationary points on the potential energy surface are optimized. The overall rate coefficients calculated using conventional TST with Wigner tunneling correction for 2,5-DHF and 2,3-DHF are 2.25 × 10−11 and 4.13 × 10−10 cm3 molecule−1 s−1, respectively, in the same range as the previously determined experimental values. The branching ratios of different channels were estimated using the computed rate coefficients. The abstraction of H atom, leading to dihydrofuranyl radical, is found to be a significant probability, equally important as the addition of OH to the double bond in the case of 2,5-DHF. However, this probability is very small in the case of 2,3-DHF because the rate coefficient of the addition reaction is more than 10 times that of the abstraction reaction. This explains the conspicuous absence of furan among the products of the reaction of OH with 2,3-DHF. The calculations also indicate that the abstraction reaction, and hence furan formation, may become significant for OH-initiated oxidation of 2,3-DHF at temperatures relevant to combustion.
1. INTRODUCTION HOx radicals (OH and HO2 radicals) are important species in the tropospheric chemistry. During the tropospheric oxidation of many VOCs such as alkanes and alkenes, secondary reactions of peroxy and alkoxy radicals regenerate these HOx species,1 thus helping to maintain the oxidative capacity of the troposphere. Although OH radical is the key oxidative species, the large sequence of tropospheric reactions includes several steps leading to the fast interconversion of HO2 to OH and vice versa.1 The reactivity with OH as well as the reaction mechanism varies within the same class of molecules, depending on number of factors. Recently, during our study on the tropospheric degradation of dihydrofurans (DHFs), the rate coefficient of reaction of 2,3-DHF with OH at 298 K was found to be approximately 2 times higher than that of 2,5-DHF.2 Although the dominant atmospheric oxidation of 2,3-DHF occurs via its reaction with ozone, the reaction with OH is also expected to occur to an extent of about 10%.2 2,5-Dihydrofuranyl radical, the primary radical formed by the H-abstraction reaction of OH with 2,5-DHF, is reported to lead to the formation of furan along with HO2,3 the reaction sequence being the same as that of the formation of benzene during the reaction of cyclohexadiene with OH.3,4 However, there are no similar reports on the reaction of OH with 2,3-DHF. There are reports on the reactions of OH with a number of alkyl substituted 2,3-DHFs, important intermediates in the tropospheric oxidation of alkanes;5,6 however, there is no © XXXX American Chemical Society
mention of observing substituted furans among the reaction products.7 Formation of furan was noted during our study on the Cl atom initiated oxidation of 2,5-DHF but not in 2,3DHF.8 However, the conditions employed for studying the reactions of 2,3-DHF with Cl were complex, with dominating dark reactions, and hence, the results are not conclusive. In the case of similar reactions in cyclohexadienes, the formation of benzene was observed for both 1,4-cyclohexadiene and 1,3cyclohexadiene but with different yields.3 Because the formation of aromatic furan and the associated formation of HO2 are significant processes in tropospheric chemistry, it is important to know and quantify this possibility. Thus, the present study is undertaken to confirm if the reaction of OH with 2,3-DHF also leads to the formation of furan as in the case of 2,5-DHF and to quantify the yield if it is formed. Since reaction of ozone is also an important tropospheric pathway for these molecules, the possibility of any furan formation during this reaction, via secondary OH generation, is also investigated. Theoretical studies on the reactivity of both these unsaturated ethers with OH are also carried out to understand in depth the similarities and differences in the reaction mechanism of these two molecules, which differ in the position of the double bond. The details of the mechanism of their reactions with OH are Received: December 12, 2014 Revised: February 20, 2015
A
DOI: 10.1021/jp5122454 J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry A
for the second step, under steady state approximation, the overall reaction rate coefficient can be written as
also relevant to combustion chemistry, as these unsaturated ethers are important components of future biofuels.9
2. EXPERIMENTAL AND COMPUTATIONAL METHODS OH radicals were generated in situ by the photolysis of H2O2 at 254 nm (UV lamp from Sankyo Denki) in a mixture consisting of DHF (300−350 ppm), H2O2 (∼1.5 Torr), and air, maintaining the total pressure as 800 Torr. The quantum yield of OH is known to be 2 at this wavelength, with negligible generation of H and HO2.10,11 The photolysis of DHFs and dark reaction of DHF with H2O2 were found to be absent in our experimental conditions. The products of the reaction of OH with the DHFs were identified using GC−MS (ShimadzuGCMS-QP2010) by comparing the fragmentation pattern of the product with standard mass spectral library (NIST-05 and Wiley) data. The amounts of reactants and furan formed as a product were quantified using the calibration curve method. Calibration was performed by injecting different volumes of standard mixtures. Standard mixtures were prepared for furan (product) and DHFs (reactant) by injecting a measured quantity of pure liquid (1−4 μL) in a quartz bulb of known volume and diluting to 800 Torr with N2. Sufficient time was provided for ensuring a uniform distribution of the compounds in the standards. Ultrapure air (Zero grade; Chemtron Science Laboratories Mumbai, India) was used as the buffer gas. High purity oxygen (>99.9% from Alchemie Gases & Chemicals Pvt. Ltd., Mumbai, India) was used to generate the ozone/oxygen mixture. 2,5DHF (97%) and 2,3-DHF (99%) from Sigma-Aldrich were stored in glass containers under vacuum and were used after a number of freeze−pump−thaw cycles. Ab initio theoretical studies were carried out using computational methods to calculate the rate coefficients of the reactions of OH with 2,3- and 2,5-DHFs, using transition state theory, as discussed earlier.12 Here, the geometries of the reactants, prereactive complexes, transition states (TSs), and products of abstraction and addition channels were optimized at M06-2X level of theory, with 6-311G* basis set. The harmonic vibrational frequencies were calculated for each structure to verify the nature of the stationary point. The TS was identified by confirming the presence of only one imaginary frequency. All other stationary points were with real frequencies. Additionally, the TSs corresponding to the addition and abstraction reactions were also confirmed by the normal mode of analysis, that the imaginary frequencies in the TS correspond to the stretching modes of the CC, along with forming O−C bonds (in addition reactions) and the stretching modes of the breaking C−H with forming O−H bonds (in abstraction). All structures and normal modes of vibrations were viewed in MacMolPlt.13 The existence of the optimized TSs on the potential energy surface was further ascertained by intrinsic reaction coordinate (IRC) calculations. All optimizations and IRC calculations were conducted using the GAMESS14 program package. The rate coefficient for each transition state was computed using conventional transition state theory.15 The negative energies of the TSs indicate formation of prereactive complexes (RC), in equilibrium with the reactants, followed by the formation of products, as shown below.
k=
k1k 2 k −1
provided the equilibrium steps are fast and the second step is the rate-determining reaction. The equilibrium constant of the fast pre-equilibrium between the reactants and the RC may be derived in terms of the partition functions (Q) and activation energies (E) by applying basic statistical thermodynamic principles: ⎛Q ⎞ Keq = ⎜⎜ RC ⎟⎟e[(ER − ERC)/(RT )] ⎝ QR ⎠
By application of the classical TST, k2 may be calculated as k 2 = Γ(T )
kBT ⎛ QTS ⎞ [(ERC − ETS)/(RT )] ⎜⎜ ⎟⎟e h ⎝ Q RC ⎠
The effective rate coefficient is obtained as
k = Keqk 2 k(T ) = Γ(T )
kBT ⎛ QTS ⎞ [−ΔE0⧧ /(RT )] ⎜⎜ ⎟⎟e h ⎝ QR ⎠
where ΔE⧧0 is the molar zero-point energy inclusive of the barrier height, i.e., (ER − ETS), Γ(T) is the tunneling correction factor, ⧧ represents the transition state, kB is the Boltzmann constant, and h is Planck’s constant. T is the temperature in kelvin, QTS and QR are the partition functions for the TS and the reactants, and R is the universal gas constant. The above expression was used for calculating the rate coefficients for each TS. The tunneling corrections were calculated by using Wigner’s potential16 and Eckart’s symmetrical and unsymmetrical potentials.17,18 Although this expression is identical to that of a reaction where the prereactive complex is not considered, the energy barrier with respect to the prereactive complex, ETS − ERC, is necessary for calculating the correct tunneling factor Γ(T), as shown by Alvarez-Idaboy et al.19 This was considered while calculating Eckart’s symmetrical tunneling correction, and the energy of the postreactive complex was also considered in the case of Eckart’s unsymmetrical tunneling correction. The partition functions, Q, were calculated based on the optimized geometry under harmonic oscillator (HO) approximations. The electronic partition function of the OH radical was evaluated by considering the splitting of 139.7 cm−1 in the 2Π ground state of OH 20 into account.
3. RESULTS AND DISCUSSION 3.1. Products of Reaction of DHFs with OH. GC−MS analysis was conducted to identify the products of the reaction of OH with 2,3-DHF and 2,5-DHF. Figure 1A shows the ion chromatograms obtained with Q-Plot column. With 40% depletion of the parent molecule, furan (a) was the only major product observed for the reaction of 2,5-DHF with OH. In the case of 2,3-DHF, with the same extent of reaction ethyl formate (b) was the only major product observed. The formation of furan as a product of ozone reaction was also explored as shown in Figure 1B. Although furan was observed in the case of 2,5DHF, only ethyl formate (b) and 2-propenal (c) were observed in the case of 2,3-DHF.
reactants ⇌ RC → product
If k1 and k−1 are the rate coefficients for the forward and backward reactions of the equilibrium step and k2 is rate coefficient B
DOI: 10.1021/jp5122454 J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry A
acetate was observed as the major product formed via the alkoxy radical, and a similar mechanism is expected to lead to the ringopened product, 3-oxopropyl formate, in the case of 2,3-DHF. Under the present experimental conditions, where self-reactions of the hydroxyperoxy radicals dominate, cyclic C4 hydroxycarbonyl and diols are also expected, along with the equivalent ring-opened products.21 Ethyl formate, the only product observed under the present experimental conditions, might be formed by part of the radicals undergoing ring opening followed by fragmentation. The expected major products of ozone reaction6 also could not be observed in our present study. Ethyl formate observed in the case of ozone reaction with 2,3-DHF could be a product of the reaction of secondary OH generated during reaction of ozone, or both ethyl formate and propenal could be formed by further fragmentation of products generated from the ozonide. The limited number of observed products does not allow the elucidation of the complete reaction mechanism. Nevertheless, the results clearly show that furan, which is formed with a yield of about 20% during the reaction of OH with 2,5-DHF, is absent as a product of the reaction of OH with 2,3-DHF. To further understand this difference in the reaction mechanism, theoretical studies were conducted to estimate the rate coefficients of different channels of reactions of OH with these molecules, as discussed below. 3.2. Theoretical Calculations. In addition to the difference in the products, the reactions of 2,5-DHF and 2,3-DHF with OH have different rate coefficients, being almost 2 times higher in the case of 2,3-DHF. The theoretical study is directed toward identifying all possible channels of reactions and calculating their rate coefficients by identifying the corresponding TSs and associated energetics and by comparing their salient features. 3.2.1. Optimized Geometries and Energies of Stationary Points.
Figure 1. Ion chromatograms obtained during GC−MS analysis of products of (A) OH initiated oxidation and (B) ozone initiated oxidation of 2,5-DHF and 2,3-DHF: (1) 2,5-DHF + OH; (2) 2,3-DHF + OH; (3) 2,5-DHF + O3; (4) 2,3-DHF + O3; (5) 2,5-DHF + O3 (background). Identified products are (a) furan, (b) ethyl formate, and (c) 2-propenal. Concentration of DHFs is 325 ppm.
Further experiments on the reaction of ozone with 2,5-DHF were carried out in the presence and absence of cyclohexane as an OH scavenger to understand the origin of furan among the ozone reaction products. The quantitative analysis of DHF and furan showed a decrease in the depletion of DHF in the presence of OH scavengers (see Table 1), suggesting that Table 1. Estimation of Number of Molecules of 2,5-DHF Consumed and Furan Formed in the Absence and Presence of Cyclohexane experiment no. no. of molecules of 2,5-DHF depleted × 10 without cyclohexane with cyclohexane (10%) no. of molecules of furan formed × 10−17 without cyclohexane with cyclohexane (10%) yield of furan from reaction of OH
1
2
3
4
14.6 7.90
35.2 24.7
24.4 16.1
39.4 30.5
1.75 0.29 0.21
3.12 0.95 0.21
2.41 0.96 0.17
2.41 0.34 0.23
−17
The numbering of atoms used in the present study is represented in the structures depicted above. The optimized structures of 2,5-DHF and 2,3-DHF, with some important bond distances, are represented in Figures 2 and 5, respectively. Optimized structures of other stationary points in the potential energy surfaces of all the possible reactions with OH are included in these figures and in Figures 4 and 6. The details are discussed below. The optimized geometry of 2,5-DHF with lowest energy was found to be symmetric and planar, whereas 2,3-DHF was found to be deviating marginally from planarity, with dihedral angles of C2−C3−C4−O5, C3−C2−C1−O5, and C4−C3−C2−C1 as 2.42°, 16.92°, and 11.88°, respectively, similar to those calculated earlier.22 2,5-DHF being a symmetric structure, two sets of equivalent sites are available for the addition of OH. Additionally, the abstraction of two symmetrically placed sets of H atoms is also possible, the sets being (a) H6, H7, H10, H11and (b) H8, H9. Optimization of transition states was attempted for all these reactions. The optimized geometries are shown in Figure 2, and the corresponding energies are given in Table 2.
additional depletion is due to the reaction of OH generated during the reaction of ozone with 2,5-DHF. The ratio of the quantity of furan formed to the extra depletion in the absence of OH is 0.21 ± 0.03, similar to the yield of furan from OH reaction (18.3%) with 2,5-DHF, previously estimated by Ohta.3 This further confirms that the furan formed during the reaction of ozone is due to the reaction of secondary OH generated as a product of ozone reaction. In the case of 2,3-DHF, the yields of the products, as compared from the areas of peaks in ion chromatograms, were not significantly affected by the presence of the OH scavenger. In addition to furan, formed via the radical from the abstraction reaction, the hydroxydihydrofuranyl radicals formed by OH addition are expected to give peroxy radicals, which may lead to alkoxy radicals. The reactions of structurally similar 4,5dihydro-2-methylfuran have been studied previously, using FTIR and atmospheric pressure ionization tandem mass spectrometer, coupled with SPME sampling technique.6 3-Oxopropyl C
DOI: 10.1021/jp5122454 J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry A
As observed in several unsaturated molecules, all these reactions, involving both addition and abstraction reactions, were observed to proceed through a common prereactive complex, shown as RC in Figure 2. This complex is planar, and the O−H group placed above the ring is almost symmetrical, ∠OHO being 148° and the dihedral angle of O with respect to the ring being 54−56°. The distance of the H atom from the O atom on the ring (1.873 Å) and the marginal stretch of the O−H bond in the hydroxyl radical from 0.973 to 0.98 indicate a hydrogen bond interaction, which is possibly responsible for the high stability of RC (energy of the RC is −7 kcal/mol; see Table 2). All TSs and RC were confirmed by IRC calculations. A typical IRC plot in the case of the addition of OH is shown in Figure 3. The transition state for the addition of OH with 2,5-DHF is shown as AddTS in Figure 2. The oxygen atom is marginally shifted from the C−C−C−C plane. Depending on the point of addition and orientation of O−H bond, the dihedral angle of O5−C4−C3−C2 is found to be 10° to one side and 20° to the other side O5−C1−C2−C3. Two types of TS optimized for the abstraction of H atoms are AbsTS1 for H6, H7, H10, and H11 and AbsTS2 for H8 and H9, H8 and H9 being H atoms attached to double bonded carbon atoms. The ring structure in AbsTS1 remains almost planar, with C−C−C−O dihedral angle of 6.7° on the side of the H atom being abstracted and C−C−C−O dihedral angle of 4.7° on the other side. The AbsTS1 lies lower than the reactants, with the H atom of O−H oriented toward the O atom on the ring. AbsTS2 is energetically very high, and abstraction of H8 and H9 is almost improbable. The exit side of the potential energy surface also shows the presence of complexes lying lower in energy than the products. The geometries of the product complex (PC) and the products (hydroxyldihydrofuranyl radical and dihydrofuranyl radical) are presented in Figure 4, and the corresponding energies are provided in Table 3. As expected, the abstraction of H8 and H9 is an endothermic process, whereas the ΔH values of other abstraction channels and addition reaction are negative, with addition having a marginally higher negative value for ΔH. In the case of 2,3-DHF, the addition sites as well as the H atoms that are abstracted are not equivalent and each channel was investigated separately. Although the TSs for all the abstraction channels could be optimized, only one could be optimized for addition, the one at the carbon adjacent to the O atom, C4. Similar to the addition of OH, the energies of all the TSs, except the ones for the abstraction of H10 and H11, are negative, indicating the presence of a prereactive complex. From the final geometries obtained from the IRC calculations starting with optimized TSs, a common prereactive complex, shown as RC in Figure 5, was obtained for abstraction reactions. The ring structure of RC remains almost the same as that of 2,3-DHF. However, no complex could be obtained for the addition channel. The hydrogen atoms on the same carbon atoms were found to be equivalent, with a marginal difference
Figure 2. Optimized geometries of the reactants, prereactive complex, and transition states during the reaction of 2,5-DHF with OH.
Figure 3. Typical IRC plot for the reaction of OH with 2,5-DHF (addition reaction).
Table 2. Energies of Reactants, Different Transition States (TS), and Corresponding Reaction Complex Calculated for OH + 2,5-DHF Reaction 2,5-DHF OH RC AddTS AbsTS1 AbsTS2
imaginary frequency
energy (hartree)
ZPE (hartree)
corrected energy (hartree)
ΔE (kcal/mol)
410.1i 979.3i 1556.4i
−231.1694 −75.7141 −306.8992 −306.8926 −306.8867 −306.8685
0.0937 0.0085 0.1058 0.1058 0.1017 0.0993
−231.076 −75.7056 −306.7934 −306.7868 −306.7850 −306.7692
−7.61 −3.49 −2.35 7.55
D
DOI: 10.1021/jp5122454 J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry A
in the energy, as shown in Table 4. For the TS for the abstraction of H6 and H7, the H atoms α to O atom are similar in energy with those for abstraction of H8 and H9. The optimized geometries and energies of the complexes on the product side (PC) and the products are shown in Figure 6 and Table 5. Although the TS for addition at C3 could not be identified, the geometry of the product of addition at C3 could be optimized and was observed to be 1.2 k cal mol−1 higher than the radical corresponding to C4 addition. The energies of all the stationary points in the reactions of OH with 2,5-DHF and 2,3-DHF are depicted in Figures 7 and 8, respectively. 3.2.2. Calculation of Rate Coefficients. The rate coefficients for all channels of reactions of OH with 2,5-DHF and 2,3-DHF at 298 K, calculated using the TS energies at M06-2X levels, are summarized in Tables 6 and 7, respectively. Tunneling corrections using Eckart’s symmetrical and unsymmetrical potentials were applied for 2,5-DHF along with Wigner correction, as prereactive complex was located for all TSs. These corrections were applied for 2,3-DHF also, assuming that the prereactive complex for addition is the same as that for abstraction and that the addition forms the product radical directly without involving any product complex. The calculated rate coefficients of 2,3-DHF are given in Table 7. Because of the uncertainty of these assumptions, the rate coefficients considering Wigner’s tunneling correction are only attempted for comparison, as highlighted in both tables. It may be noted that the difference in the tunneling correction factors, whether Wigner’s or Eckart’s, is only marginal in these cases (Tables 6 and 7), and the above assumptions may not affect the calculated values significantly. The calculated rate coefficients of both 2,5-DHF and 2,3DHF are 2.25 × 10−11 and 4.13 × 10−10 cm3 molecule−1 s−1, respectively. They are in the same order as the experimental values, 6.45 × 10 −11 for 2,5-DHF and 1.2 × 10 −10 cm3 molecules−1 s−1 for 2,3-DHF. It is approximately 3 times lower for 2,5-DHF and 2 times higher for 2,3-DHF. While the rate coefficients of H atom abstraction of 2,5-DHF and 2,3DHF are comparable, the calculated rate coefficient of addition at one carbon atom is approximately 2 orders of magnitude higher in the case of 2,3-DHF than that of 2,5-DHF. Thus, even without inclusion of the rate coefficient of addition at C3, which could not be calculated without the optimized structure and energy of TS, the addition of OH to 2,3-DHF is observed to be more efficient than that to 2,5-DHF. The energy of the TS and the exothermicity of the reaction are found to be marginally more negative for the reaction of OH with 2,3-DHF. If we compare the rate coefficients of addition and abstraction in the case of 2,5-DHF, given in Table 6, it is clear that the possibility of the abstraction of H atom by OH is comparable to the possibility of the addition reaction in the case of 2,5-DHF (about 50%). The reaction of oxygen with dihydrofuranyl radical, formed by H atom abstraction, is known to form furan along with HO2 as given below.3
Figure 4. Optimized geometries of the product complex and products during the reaction of 2,5-DHF with OH.
Table 3. Energies of the Product Complex and Different Products Calculated for OH + 2,5-DHF reaction
H2O PC PAdd PAbs1 PAbs2
energy (hartree)
ZPE (hartree)
corrected energy (hartree)
ΔE (kcal/mol)
−76.3994 −306.9479 −306.9434 −230.5349 −230.4798
0.0217 0.1048 0.1086 0.0798 0.0812
−76.3777 −306.8431 −306.8348 −230.4551 −230.3986
−38.79 −33.58 −32.37 3.07
However, in the case of 2,3-DHF (Table 7), this possibility is negligible at 298 K with the total H abstraction contribution of 4 H atoms being approximately 5%. Although the TS for OH
Figure 5. Optimized geometries of the reactants, prereactive complex, and transition states during the reaction of 2,3-DHF with OH. E
DOI: 10.1021/jp5122454 J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry A Table 4. Energies of Reactants, TS, and Reaction Complex for OH + 2,3-DHF Reaction imaginary frequency
energy (hartree)
ZPE (hartree)
corrected energy (hartree)
ΔE (kcal/mol)
963.4i 966.7i 944.0i 933.2i 1470.5i 1656.2i 204.3i
−231.1753 −75.7141 −306.9010 −306.8915 −306.8915 −306.8914 −306.8914 −306.8739 −306.8734 −306.9008
0.0942 0.0085 0.1051 0.1016 0.1015 0.1016 0.1015 0.1000 0.0992 0.1056
−231.081 −75.7056 −306.7959 −306.7899 −306.7900 −306.7898 −306.7900 −306.7739 −306.7742 −306.7952
−5.77 −2.04 −2.09 −1.98 −2.06 8.01 7.85 −5.36
2,3-DHF OH RC AbsTSH6 AbsTSH7 AbsTSH8 AbsTSH9 AbsTSH10 AbsTSH11 AddTSC4
Figure 7. Energy level diagram for all channels of 2,5-DHF. Energy is in kcal mol−1 (not marked to scale for clarity).
Figure 6. Optimized geometries of the product complex and products during the reaction of 2,3-DHF with OH.
Table 5. Energies of Different Product Complexes and Corresponding Products Calculated for the Reaction of OH with 2,3-DHF
H2O PAbsH6 PAbsH7 PAbsH8 PAbsH9 PAbsH10 PAbsH11 PAddC4 PC1(H6/H7) PC2(H8/H9) PC3(H10) PC4(H11)
energy (hartree)
ZPE (hartree)
corrected energy (hartree)
ΔE (kcal/mol)
−76.3994 −230.5180 −230.5180 −230.5349 −230.5349 −230.4819 −230.4862 −306.9546 −306.9302 −306.9479 −306.8923 −306.8954
0.021669 0.0799 0.0799 0.0798 0.0798 0.0818 0.0813 0.1087 0.1043 0.1050 0.1062 0.1057
−76.3777 −230.4381 −230.4381 −230.4551 −230.4551 −230.4001 −230.4049 −306.8459 −306.8259 −306.8429 −306.7861 −306.7897
−18.29 −18.31 −28.99 −28.98 5.54 2.55 −37.16 −24.6 −35.31 0.38 −1.89
Figure 8. Energy level diagram for all channels of 2,3-DHF. Energy is in kcal mol−1 (not marked to scale for clarity).
addition at C3 could not be obtained in the present study, there is a finite possibility of occurrence of this reaction, which will further reduce the percentage contribution of abstraction F
DOI: 10.1021/jp5122454 J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry A
Table 6. Computed Rate Coefficients for Abstraction and Addition Pathways for the Reaction of OH with 2,5-DHF, at M06-2X Level, at 298 Ka k total = 2kAddTS + 4kAbsTS1 + 2kAbsTS2 HO AddTS AbsTS1 AbsTS2 ktotal
3.83 1.75 2.95 1.47
× × × ×
WTC
10−12 10−12 10−19 10−11
4.45 3.39 9.91 2.25
× × × ×
ESTC
10−12 10−12 10−19 10−11
4.54 3.57 1.02 2.35
× × × ×
10−12 10−12 10−18 10−11
EUTC 4.51 4.38 2.95 2.65
× × × ×
10−12 10−12 10−19 10−11
a The unit of k is cm3 molecule−1 s−1. HO: harmonic oscillator, without any tunnelling correction. WTC: Wigner’s tunneling correction. ESTC: Eckart’s symmetrical tunneling correction. EUTC: Eckart’s unsymmetrical tunneling correction.
Table 7. Computed Rate Coefficients for Abstraction and Addition Pathways for the Reaction of OH with 2,3-DHF, at M06-2X Level, at 298 Ka
k total = kAddTSC4 + kAbsTSH6 + kAbsTSH7 + kAbsTSH8 + kAbsTSH9 + kAbsTSH10 + kAbsTSH11 HO AddTSC4 AbsTSH6 AbsTSH7 AbsTSH8 AbsTSH9 AbsTSH10 AbsTSH11 ktotal
3.72 1.99 2.28 4.16 5.54 2.28 1.33 3.86
× × × × × × × ×
10−10 10−12 10−12 10−12 10−12 10−19 10−18 10−10
WTC 3.87 3.79 4.35 7.77 1.03 7.09 4.90 4.13
× × × × × × × ×
10−10 10−12 10−12 10−12 10−11 10−19 10−18 10−10
ESTC 4.09 4.07 4.68 8.33 1.10 7.30 5.05 4.37
× × × × × × × ×
10−10 10−12 10−12 10−12 10−11 10−19 10−18 10−10
EUTC 3.79 4.83 5.55 9.65 1.26 2.28 1.33 4.12
× × × × × × × ×
10−10 10−12 10−12 10−12 10−11 10−19 10−18 10−10
The unit of k is cm3 molecule−1 s−1. HO: harmonic oscillator, without any tunnelling correction. WTC: Wigner’s tunneling correction. ESTC: Eckart’s symmetrical tunneling correction. EUTC: Eckart’s unsymmetrical tunneling correction. Eckart’s correction is applied considering the same RC as that of abstraction reaction.
Figure 9. Variation of the computed rate coefficients at M06-2X with temperature for (A) 2,5-DHF without any correction, (B) 2,5-DHF with Wigner’s tunneling correction, (C) 2,3-DHF without any correction, and (D) 2,3-DHF with Wigner’s tunneling correction: (□) abstraction, (■) addition, and (●) total rate coefficients.
reaction. This explains the complete absence of furan among the products of reaction of OH with 2,3-DHF, which can be formed only from the dihydrofuranyl radical, the product of H abstraction reaction. Among the 4 H atoms that can be abstracted, the abstraction of H atoms at C2, the carbon adjacent to the double bond, is 2 times more probable than the abstraction of H atoms from the carbon adjacent to the O atom (C1). If formed, both these dihydrofuranyl radicals from 2,3-DHF, with radical centers at C2 and C1, are expected to undergo a reaction similar to reaction 2, leading to furan formation. The estimated contribution of the abstraction in 2,5-DHF is higher than the observed yield of furan, 0.21, which is expected because part of the dihydrofuranyl radicals may form peroxide radicals, which may lead to the formation of other alcohols and keto products. The presence of ethyl formate as a product of OH reaction with 2,3-DHF indicates that the radical formed by addition may undergo ring opening reactions, as well as fragmentation. The variation of calculated rate coefficients with temperature is shown in Figure 9. The total abstraction rate coefficient, considering all the possible channels of H abstraction, is plotted for both the molecules. Although the abstraction and addition rate coefficients are almost same for 2,5-DHF at room temperature (7.65 and 7 × 10−12 cm3 molecule−1 s−1), the abstraction rate coefficient becomes almost 5.9 times higher than the addition rate coefficient at 1000 K (1.54 × 10−12 and 2.59 × 10−13 cm3 molecule−1 s−1, respectively). In the case of 2,3-DHF, the addition rate coefficient, which is 15 times higher than the
total abstraction rate coefficient at room temperature, becomes one-third of the abstraction rate coefficient at 1000 K. Thus, furan formation, which does not occur during the atmospheric oxidation of 2,3-DHF, may become a significant pathway under the conditions of combustion of biofuels. Similar difference in the reactivity of 2,5-DHF and 2,3-DHF has been observed in the case of other reactions also. Formation of furan is observed to be an important channel of thermal decomposition of 2,5-DHF,23,24 but not in the case of 2,3-DHF.25 Similarly, catalytic dehydrogenation was observed to be more prominent from 2,5-DHF at temperatures as low as 320 K, whereas 2,3-DHF can undergo hydrogenation at that temperature.26 DFT studies have assigned the difference between the two molecules to high barrier in the dehydrogenation pathways in the case of 2,3DHF.27 The present computational results do not show the rate coefficient of the H abstraction reaction of 2,3-DHF to be lower than that of 2,5DHF. The absence of the abstraction pathways in 2,3-DHF is due to the large rate coefficient of the addition reaction, resulting from a low-lying TS. The reason for the lower energy of AddTS in the case of 2,3-DHF compared with that of 2,5DHF is not clear. The presence of the O atom in the adjacent position in the ring may facilitate the addition of OH to C4 in the case of 2,3-DHF, via a weak hydrogen bond interaction between the H of the incoming OH and the O atom of the ring. Such an interaction may not be very important in the case of 2,5-DHF, where the O atom is away from the double bond. However, the optimized structure of AddTSC4 in Figure 5 shows the H atom of OH to be directed away from the O atom
a
G
DOI: 10.1021/jp5122454 J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry A
(4) Tuazon, E. C.; Aschmann, S. M.; Nguyen, M. V.; Atkinson, R. Hatom abstraction from selected C-H bonds in 2,3-dimethylpentanal, 1,4-cyclohexadiene and 1,3,5-cycloheptatriene. Int. J. Chem. Kinet. 2003, 415−426. (5) Atkinson, R.; Arey, J.; Aschmann, S. M. Atmospheric chemistry of alkanes: review and recent developments. Atmos. Environ. 2008, 42, 5859−5871. (6) Martin, P.; Tuazon, E. C.; Aschmann, S. M.; Arey, J.; Atkinson, R. Formation and atmospheric reactions of 4,5-dihydro-2-methylfuran. J. Phys. Chem. A 2002, 106, 11492−11501. (7) Lim, Y. B.; Ziemann, P. J. Kinetics of the heterogeneous conversion of 1,4-hydroxycarbonyls to cyclic hemiacetals and dihydrofurans on organic aerosol particles. Phys. Chem. Chem. Phys. 2009, 11, 8029−8039. (8) Alwe, H. D.; Walavalkar, M. P.; Sharma, A.; Pushpa, K. K.; Dhanya, S.; Naik, P. D. Rate coefficients for the gas-phase reactions of chlorine atoms with cyclic ethers at 298 K. Int. J. Chem. Kinet. 2013, 45, 295−305. (9) Tran, L. S.; Sirjean, B.; Glaude, P.-A.; Fournet, R.; Battin-Leclerc, F. Progress in detailed kinetic modeling of the combustion of oxygenated components of biofuels. Energy 2012, 43, 4−18. (10) Thiebaud, J.; Aluculesei, A.; Fittschen, C. Formation of HO2 radicals from the photodissociation of H2O2 at 248 nm. J. Chem. Phys. 2007, 126, 186101. (11) Atkinson, R.; Baulch, D. L.; Cox, R. A.; Crowley, J. N.; Hampson, R. F.; Hynes, R. G.; Jenkin, M. E.; Rossi, M. J.; Troe, J. Atmos. Chem. Phys. 2004, 4, 1461−1738. IUPAC. Task Group on Atmospheric Chemical Kinetic Data Evaluation. Evaluated Kinetic Data. http://iupac.pole-ether.fr. (Vol. 2, accessed July 10, 2014). (12) Alwe, H. D.; Walavalkar, M. P.; Sharma, A.; Dhanya, S.; Naik, P. D. Reactivity of Cl atom with triple bonded moleculesA study with alcohols. J. Phys. Chem. A 2014, 118, 7695−7706. (13) Bode, B. M.; Gordon, M. S. MacMolPlt: a graphical user interface for GAMESS. J. Mol. Graphics Modell. 1998, 16, 133−138. (14) Schmidt, M. W.; Baldridge, K. K.; Boatz, J. A.; Elbert, S. T.; Gordon, M. S.; Jensen, J. J.; Koseki, S.; Matsunaga, N.; Nguyen, K. A.; Su, S.; et al. General atomic and molecular electronic structure system. J. Comput. Chem. 1993, 14, 1347−1363. Laidler, K. J. Chemical Kinetics, 3rd ed.; Pearson Education: Delhi, India, 2004. (15) Laidler, K. J. Chemical Kinetics, 3rd ed.; Pearson Education: Delhi, India, 2004. (16) Wigner, E. P. Ü ber das überschreiten von potentialschwellen bei chemischen reaktionen. Z. Phys. Chem. B 1932, 19, 203. (17) Eckart, C. The penetration of a potential barrier by electrons. Phys. Rev. 1930, 35, 1303. (18) Johnston, H. S.; Heicklen, J. Tunnelling corrections for unsymmetrical Eckart potential energy barriers. J. Phys. Chem. 1962, 66, 532−533. (19) Alvarez-Idaboy, J. R.; Mora-Diez, N.; Boyd, R. J.; Vivier-Bunge, A. On the importance of prereactive complexes in molecule−radical reactions: hydrogen abstraction from aldehydes by OH. J. Am. Chem. Soc. 2001, 123, 2018−2024. (20) Ogura, T.; Miyoshi, A.; Koshi, M. Rate coefficients of H-atom abstraction from ethers and isomerization of alkoxyalkylperoxy radicals. Phys. Chem. Chem. Phys. 2007, 9, 5133−5142. (21) Orlando, J. J.; Tyndall, G. S. Laboratory studies of organic peroxy radical chemistry: an overview with emphasis on recent issues of atmospheric significance. Chem. Soc. Rev. 2012, 41, 6294−6317. (22) Billes, F.; Böhlig, H.; Ackermann, M.; Kudra, M. A vibrational spectroscopic study on furan and its hydrated derivatives. J. Mol. Struct.: THEOCHEM 2004, 672, 1−16. (23) Chowdhury, P. K.; Mittal, J. P. IR multiphoton dissociation dynamics of 2,5-dihydrofuran. Time-resolved observation of concerted furan formation. Chem. Phys. Lett. 1995, 242, 421−426. (24) Lifshitz, A.; Bidani, M.; Bidani, S. Thermal reactions of cyclic ethers at high temperatures. 4. Pyrolysis of 2,5-dihydrofuran behind reflected shocks. J. Phys. Chem. 1986, 90, 6011−6014.
of the ring, ruling out this interaction to be responsible for the large rate coefficient of addition for 2,3-DHF. Although the difficulty in locating the TS for addition at C3 does not rule out the existence of this pathway, the results of the previous experimental study on reactions of C15 alkyl-substituted DHFs with OH suggest the dominance of addition at C4.28
4. CONCLUSION The formation of the aromatic molecule furan, along with HO2, is found to be absent during the OH-initiated oxidation of 2,3-DHF, unlike that of 2,5-DHF. A detailed theoretical study has been carried out with optimization of geometry and calculation of energies of all the stationary structures in the potential energy surface for the reaction of OH with both molecules. The rate coefficients of each possible channel are calculated by applying transition state theory. The calculated overall rate coefficients at 298 K are in the same range as those determined experimentally. The results show the reaction of OH to be dominated by addition at the carbon adjacent to O atom in the case of 2,3-DHF. In 2,5-DHF, the abstraction of H atom is also found to be as significant as the addition of OH to the double bond. This clearly explains the complete absence of the channel of formation of furan via dihydrofuranyl radical in the reaction of OH with 2,3-DHF. Calculation of energy at the CCSD(T) level and computation of rate coefficient using variational transition state are desirable to obtain more accurate values of the rate coefficients. However, the changes are expected to be marginal, and this is not expected to affect the above conclusion that the addition predominates the reaction of OH with 2,3-DHF, unlike the case of 2,5-DHF. Thus, this study establishes that the formation of furan and the direct generation of HO2 from the reaction of dihydrofuranyl radical with atmospheric oxygen are insignificant during the tropospheric degradation of 2,3-DHF.
■
ASSOCIATED CONTENT
S Supporting Information *
Structural parameters in the form of Cartesian coordinates of all the optimized geometries and mass spectra. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. Phone: 0091-22-25593760. Fax: 91-22-25505151. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS The authors acknowledge the support and encouragement from the Department of Atomic Energy, India, during the course of the present work.
■
REFERENCES
(1) Ehhalt, D. H. Photooxidation of trace gases in the troposphere. Phys. Chem. Chem. Phys. 1999, 1, 5401−5408. (2) Alwe, H. D.; Walavalkar, M. P.; Sharma, A.; Dhanya, S.; Naik, P. D. Tropospheric oxidation of cyclic unsaturated ethers in the day-time: comparison of the reactions with Cl, OH and O3 based on the determination of their rate coefficients at 298 K. Atmos. Environ. 2014, 82, 113−120. (3) Ohta, T. Reactivity of allylic hydrogen in cyclohexadiene towards OH radicals. Int. J. Chem. Kinet. 1984, 16, 1495−1503. H
DOI: 10.1021/jp5122454 J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry A (25) Lifshitz, A.; Bidani, M. Thermal reactions of cyclic ethers at high temperatures. 5. Pyrolysis of 2,3-dihydrofuran behind reflected shocks. J. Phys. Chem. 1989, 93, 1139−1144. (26) Horiuchi, C. M.; Medlin, W. Adsorption and reactivity of 2,3dihydrofuran and 2,5-dihydrofuran on Pd(111): influence of the CdC position on the reactivity of cyclic ethers. Langmuir 2010, 26 (16), 13320−13332. (27) Dubnikova, F.; Lifshitz, A. Molecular hydrogen elimination from 2,5-dihydrofuran, 2,3-dihydrofuran, and 2-methyl-2,5-dihydrofuran: quantum chemical and kinetics calculations. Int. J. Chem. Kinet. 2001, 33, 685−697. (28) Lim, Y. B.; Ziemann, P.J. Chemistry of secondary organic aerosol formation from OH radical-initiated reactions of linear, branched, and cyclic alkanes in the presence of NOx. Aerosol Sci. Technol. 2009, 43, 604−619.
I
DOI: 10.1021/jp5122454 J. Phys. Chem. A XXXX, XXX, XXX−XXX