Exploring the Ring-Opening Pathways in the Reaction of Morpholinyl

Jun 11, 2012 - Department of Chemistry, Brookhaven National Laboratory, Upton, New York 11973, United States. J. Phys. Chem. A , 2012, 116 (26), ...
0 downloads 0 Views 1MB Size
Article pubs.acs.org/JPCA

Exploring the Ring-Opening Pathways in the Reaction of Morpholinyl Radicals with Oxygen Molecule Pratik P. Dholabhai* and Hua-Gen Yu Department of Chemistry, Brookhaven National Laboratory, Upton, New York 11973, United States S Supporting Information *

ABSTRACT: Quantum chemistry calculations using hybrid density functional theory and the coupled-cluster method have been performed to investigate the ring-opening pathways in the oxidation of morpholine (1-oxa-4-aza-cyclohexane). Hydrogen abstraction can form two different carbon-centered radicals, morpholin-2-yl or morpholin-3-yl, or the nitrogen-centered radical, morpholin4-yl, none of which are found to have low-energy pathways to ring-opening. Extensive exploration of multiple reaction pathways following molecular oxygen addition to these three radicals revealed two competitive low energy pathways to ring-opening. Addition of O2 to either carbon-centered radical, followed by a 1,4-H shifting mechanism can yield a long-lived cyclic epoxy intermediate, susceptible to ring-opening, following further radical attack. In particular, the second pathway begins with O2 attack on morpholin-2-yl, followed by a 1,5-H shift and a unimolecular ring-opening without having to overcome a high barrier, releasing a significant amount of heat in the overall ring-opening reaction. The calculations provide valuable context for the development of mechanisms for the low temperature combustion chemistry of nitrogen and oxygen-containing fuels.

1. INTRODUCTION In the near future, the use of nontraditional transportation fuels with high concentration of cyclic, and/or O- and N-containing compounds is expected to increase considerably. These fuels are produced largely from nontraditional sources such as oil shales, sand oils, and biomass.1−5 Designing clean and efficient combustion engines that use such fuels requires their thermodynamic and kinetic data, which is currently limited. Recently, morpholine has received attention as a model fuel since it is one of the simplest heterocyclic secondary amines that contains both nitrogen and oxygen heteroatoms.1,6−8 Moreover, morpholine is an exceptionally versatile chemical compound with several important industrial applications.1,6−8 Lucassen and co-workers have provided a detailed account of the combustion chemistry of morpholine by using it as a model fuel to investigate decomposition and oxidation pathways, with particular attention to the formation of harmful intermediates and products.7,8 A majority of the results pertaining to likely pathways for the decomposition and oxidation of morpholine are obtained experimentally [for instance, see refs 1 and 6−8 and the references therein]. However, the combustion kinetics of morpholine is still not well understood. In particular, the ring-opening process is a vital step in the combustion chemistry of morpholine, as without the ring-opening, these cyclic molecules and/or radicals can contribute to the formation of soot. In the present article, we report results for the ring-opening reaction pathways in morpholine using a hybrid density function theory (DFT) and CCSD(T) method. The formation © 2012 American Chemical Society

and stability of morpholinyl radicals morpholin-2-yl, morpholin-3-yl, and morpholin-4-yl via H-abstraction from three distinct positions is briefly addressed. We then consider the reaction of morpholinyl radicals with an oxygen molecule, a crucial step at the initial O2- and fuel-rich stage. Its kinetics can affect both the fuel efficiency and the pollutant reduction. To the best of our knowledge, these reactions have not yet received consideration from a theoretical perspective.

2. COMPUTATIONAL METHOD All calculations were performed using the Gaussian 09 program.9 Geometry optimizations and harmonic vibrational frequencies of stationary points on the potential energy surface of the system of interest were carried out using B3LYP,10,11 a hybrid DFT functional in conjunction with the 6-311+G(d,p) basis set.12,13 In order to refine the energetics of various reaction channels, a two-point basis set extrapolation scheme was employed to obtain the energy of every stationary point at the complete basis set (CBS) limit.14−18 In addition, to corroborate our results with a higher level theory, single point energy calculations at the CCSD(T)/aug-cc-pVTZ19−22 level of theory were performed for the optimized geometries. It was found that the CCSD(T)/aug-cc-pVTZ results are in good agreement with the B3LYP/CBS extrapolation values. However, only CCSD(T)/aug-cc-pVTZ energies are reported. Received: February 13, 2012 Revised: June 8, 2012 Published: June 11, 2012 7123

dx.doi.org/10.1021/jp3014496 | J. Phys. Chem. A 2012, 116, 7123−7127

The Journal of Physical Chemistry A

Article

3. RESULTS AND DISCUSSION Geometry optimization of morpholine yields two stable conformations: with the N−H bond either axial or equatorial. The equatorial conformation (see Figure 1a) is found to be

Table 1. B3LYP Electronic Energies (E), Scaled ZPEs, and Scaled Imaginary Frequencies (ν for Transition State Structures) for Stationary Points on the Potential Energy Surface of the Reaction between Morpholinyl Radicals and O2a

Figure 1. Optimized B3LYP/6-311+G(d,p) geometries of stationary points (a) Mor (b) 2• (c) 3• (d) 4•. Red, gray, blue, and white balls correspond to O, C, N and H atoms, respectively. Black dot corresponds to the site where the radical is formed.

marginally more stable (≈2 kcal/mol), consistent with previously reported results.23 Throughout this work, morpholine will be abbreviated as Mor, and a generic morpholinyl radical, obtained by hydrogen abstraction will be abbreviated as Mor•. The three specific isomeric radicals obtained by hydrogen abstraction will be denoted by 2• for morpholin-2yl, 3• for morpholin-3-yl, and 4• for morpholin-4-yl, as illustrated in Figure 1. Listed in Table 1 are electronic energies and scaled zero-point energy (ZPE) corrections for the molecules and stationary points encountered in this study and single point energies at the CCSD(T)/aug-cc-pVTZ level of theory for the optimized geometries. Scaled imaginary frequencies for the transition state structures are also included in Table 1. Reaction barrier heights (ZPE-corrected), enthalpies (0 K) of formation and transition states are denoted by Ea, ΔH, and TS, respectively. The respective scaling factors for ZPE and harmonic vibrational frequencies are 0.9877 and 0.9679.24 3.1. Formation of Morpholinyl Radicals. The initial step in morpholine combustion, like any other saturated fuel, is likely to be H-abstraction by assorted small radicals.7 We have verified computationally that direct attack by molecular oxygen on morpholine is unlikely. The reaction Mor + O2 → Mor• + HO2 is endothermic for all three isomeric Mor• isomers (ΔH ≈ 40 to 45 kcal/mol; see Figure S1, Supporting Information). Another plausible direct oxygen attack is the Mor + O2 → O(CH2)2NH(CH)2 + H 2O2 reaction. This reaction is exothermic (ΔH = −5.75 kcal/mol), but with a high barrier (Ea = 42.87 kcal/mol), making this reaction also unlikely. As discussed by Lucassen et al., the formation of morpholinyl radicals can occur via hydrogen abstraction by small, active radicals such as H, O, and OH.7 The Mor + OH reaction has recently been studied by SenGupta et al.25 and by Lucassen and co-workers.6−8 Our present calculations on the Mor + OH →

species

E (au) B3LYP/ 6-311+G(d,p)

E (au) CCSD(T)/augcc-pVTZ

ZPE (kcal/mol)

ν (cm−1)

O2 OH H2O C2H4 Mor 2• 3• 4• R3O2 Q2OOH c-QO (CHO)2NH R2O2 Q3OOH c-QO* c-PO o-PO Q6OOH o-RCOH TS1 TS2 TS3 TS4 TS5 TS6 TS7 TSH

−150.370417 −75.762337 −76.458462 −78.615512 −287.882326 −287.219863 −287.227505 −287.223854 −437.645414 −437.624058 −361.886311 −283.279848 −437.641666 −437.630112 −361.883846 −361.299699 −361.271271 −437.620589 −361.911142 −437.588978 −437.603939 −361.793971 −437.601434 −437.625673 −437.595971 −437.594523 −361.252935

−150.140942 −75.645579 −76.342305 −78.443659 −287.347279 −286.682148 −286.688698 −286.686384 −436.886782 −436.867182 −361.247711 −282.797241 −436.884607 −436.873081 −361.244914 −360.649749 −360.619069 −436.866437 −361.261677 −436.827729 −436.836729 −361.143473 −436.840127 −436.868069 −436.838518 −436.814489 −360.596023

2.31 5.24 13.19 31.48 83.50 74.94 74.99 74.65 80.82 79.62 71.85 33.98 80.83 79.58 71.58 63.69 59.86 79.46 68.44 77.13 77.58 76.68 77.26 78.72 77.07 76.73 60.64

1817i 636i 1782i 1469i 218i 1653i 980i 584i

a

The structure labels correspond to the optimized geometries shown in Figures 1−5. Single point energies at the CCSD(T)/aug-cc-pVTZ level of theory for the optimized geometries are also provided.

Mor• + H2O reactions give enthalpies of reaction, ΔH, to be −21.74 kcal/mol, −26.48 kcal/mol, and −24.54 kcal/mol for the formation of three isomeric radicals 2•, 3•, and 4• (Figure S2, Supporting Information). Barrierless paths connect reactants and products in each case, and we were unable to find locally bound complexes or TS structures for any of these three pathways. In contrast, SenGupta et al. found an optimized TS structure on the H-abstraction path from Mor + OH → Mor• + H2O, as a shallow, submerged barrier separating a OH−Mor complex from the steeper downhill path to H2O + 2•.25 No other TS structures were identified in the SenGupta work, and our calculated value for ΔH (−21.74 kcal/mol) is in very good agreement with theirs (−21.44 kcal/mol) for the Habstraction leading to 2•.25 Among the three isomeric morpholinyl radicals, 3• is found to be the most stable, in accord with previously reported results.23,25 However, since the H-abstraction reactions to form any of these radicals occur without a significant activation barrier,25 all must be considered in assessing plausible subsequent reactions. We have not investigated in any detail the isomerization barriers between the morpholinyl radicals, but proceed now to a discussion of ringopening pathways in these radicals. 3.2. Ring-Opening in Morpholinyl Radicals. The combustion products of saturated cyclic fuels are highly 7124

dx.doi.org/10.1021/jp3014496 | J. Phys. Chem. A 2012, 116, 7123−7127

The Journal of Physical Chemistry A

Article

dependent upon the ring-opening reactions,26,27 as they influence the competition between cracking to smaller molecules and radicals and the formation of aromatics and soot. Our calculations of direct unimolecular ring-opening processes for the three morpholinyl radicals show endothermic processes with large barrier heights (Ea ranging from 22 to 63 kcal/mol) owing to the cleavage of one strong C−X (X = C, N, and O) bond. While these endothermic ring-opening isomerizations, followed by H-loss or β-scission, can play a role in the high-temperature combustion of morpholine or other saturated cyclic fuels,5,7,8,27 our primary interest in the current work has been on the low energy ring-opening pathways mediated by O2 addition that can play a key role in the low-temperature oxidation chemistry, as summarized, for example, in a recent review article by Zádor et al.28 3.3. Oxygen Addition to Morpholinyl Radicals: Cyclic Epoxide Pathways. Pathways to ring-opening that begin with molecular oxygen addition to morpholinyl radicals are illustrated in Figures 2−5, based on the calculations summarized in Table 1. Subsequent energies quoted in the text and pathways depicted in Figures 2−5 are for the CCSD(T)/aug-cc-pVTZ level of theory but with the B3LYP ZPE corrections. Figure 2 shows the energy diagram for a ring-

exothermicity of ring-opening. This suggests that this cyclic epoxy intermediate may accumulate in a combustion environment including morpholine, as related compounds have been reported in low-temperature combustion of cyclic hydrocarbons.28 A similar reaction pathway is displayed in Figure 3 for the reaction of 2• with O2. In this case, the initial addition to form

Figure 3. Relative energy (kcal/mol, including ZPE corrections) diagram at the CCSD(T)/aug-cc-pVTZ level for the reaction of 2• with O2 via the formation of epoxy intermediate (c-QO) and leading to the ring-opening process.

R2O2 is barrierless and exothermic by 35.02 kcal/mol. The isomerization via TS4 to Q3OOH and the OH elimination via TS5 occur with substantially lower barriers than the analogous reactions of 3•, enough so that the Q3OOH radical would be extremely short-lived. The OH elimination from Q3OOH provides an alternate path to a cyclic epoxide, c-QO. A pair of chiral carbon centers in the epoxy ring due to the distinct O and N heteroatoms in the morpholine ring gives the cyclic epoxide two enantiomeric structures, with a random chirality determined by the initial approach of O2 to the 2• or 3• radicals. In addition, we find two conformational minima in the c-QO structures (denoted as c-QO and c-QO* in Table 1), differing in energy by about 1 kcal/mol, and separated by a small torsional barrier. The behavior of all the cyclic epoxide isomers should be similar in the combustion environment. Since the cyclic epoxy species are long-lived intermediates and their direct ring-opening is improbable, we performed a preliminary investigation for their ring-opening mechanism via a chemical reaction with O2 and OH. It was found that a favorable pathway is the H-abstraction reaction at the α-N position of the epoxy structure by OH. As shown in Figure 4, the overall ring-opening reaction has no apparent activation energy, implying that the reaction is fast. In addition, the cyclic structure c-PO is more stable than the open structure o-PO, suggesting the possibility that this pathway may not lead to ring-opening very efficiently. However, the possibility of an alternative pathway to ring-opening for this case has not been probed. In addition, a similar reaction process is also studied for O2 addition to the nitrogen atom in 4•. The initial addition reaction is endothermic (8.76 kcal/mol) as compared to the previous two exothermic cases. Later, the H-shifting isomerization reaction leading to the formation of O-

Figure 2. Relative energy (kcal/mol, including ZPE corrections) diagram at the CCSD(T)/aug-cc-pVTZ level for the reaction of 3• with O2 via the formation of epoxy intermediate (c-QO) and leading to the ring-opening process. The stationary points are labeled, and the atomic color scheme is consistent with Figure 1. The energy zero is set at the reactant asymptote.

opening pathway beginning with O2 + 3•. Along the minimum energy path, there are three intermediate structures: R3O2, Q2OOH, and c-QO. We use a notation where the superscript in RnO2 denotes the original radical, and the superscript in QnOOH denotes the source of the internally abstracted H atom and the location of the resulting radical center. The R3O2 intermediate is formed by the addition of O2 to the 3• radical, in a barrierless association reaction, exothermic by 32.34 kcal/ mol. A metastable hydroperoxy radical Q2OOH is produced in an intramolecular hydrogen abstraction via TS1, 33.36 kcal above R3O2, but only 1.02 kcal/mol above the energy of the O2 + 3• reactants. The Q2OOH radical can in turn lose an OH radical over TS2 with a barrier height 17.07 kcal/mol to form a highly stable cyclic epoxy intermediate, c-QO. As shown in Figure 2, the ring-opening barrier (TS3) of this cyclic epoxy intermediate is very high (70.24 kcal/mol) despite the large 7125

dx.doi.org/10.1021/jp3014496 | J. Phys. Chem. A 2012, 116, 7123−7127

The Journal of Physical Chemistry A

Article

barrier height is about 1 kcal/mol higher than TS4. Unlike the other QOOH structures we have investigated, Q6OOH has a small barrier TS7 (29.84 kcal/mol) to the very exothermic OH loss with ring-opening to produce NH(CH2CHO)2. We have performed intrinsic reaction coordinate (IRC) calculations to confirm that TS7 is a true transition state separating Q6OOH and the ring-opened products. Pulling OH farther from the TS7 structure synchronously weakens and breaks the 1−2 O−C bond. The observed spontaneous O−C bond breaking is consistent with the presence of a very small barrier reported for the ring-opening of a cyclic hydrocarbon oxy-compound.29 The overall reaction from 2• + O2 to OH + NH(CH2CHO)2 produces a large amount of heat (56.38 kcal/mol) through a sequence of steps where only one transition state is slightly (4.85 kcal/mol) above the reactant energy. This ring-opening pathway, not available for pure hydrocarbons, may play an important role in the low-temperature combustion of morpholine.

4. CONCLUSIONS As direct unimolecular ring-opening in morpholinyl radicals is found to be improbable in the low-temperature regime, their reaction with O2 has been systematically investigated using a hybrid DFT and CCSD(T) method to comprehend the ringopening mechanism. Results show that in combustion conditions, the morpholinyl radicals can be produced by the reaction of morpholine with OH, but the reaction with O2 is not favorable. The morpholinyl radicals are found to be energetically stable. The radicals have a strong reactivity with O2, as the addition reactions of O2 to morpholinyl radicals have no apparent activation barriers, and are considerably exothermic. In addition, two distinct reaction mechanisms for the reaction of morpholinyl radicals with O2 have been revealed. The 1,4-H shifting isomerization reaction is likely to produce a stable cyclic epoxy intermediate, whereas the 1,5-H shifting isomerization reaction leads to the ring-opened products. These two isomerization reactions will compete owing to their comparable barrier heights along the reaction pathway. In particular, the 1,5-H shifting reaction pathway involves an energy barrier that is merely 4.85 kcal/mol higher than the reactant dissociation limit, and the overall ringopening reaction can produce a significant amount of heat. These novel findings provide valuable insights for the development of mechanisms for low-temperature combustion of morpholine and similar cyclic O- and N-containing fuels.

Figure 4. Relative energy (kcal/mol, including ZPE corrections) diagram at the CCSD(T)/aug-cc-pVTZ level for the reaction of cyclic epoxy intermediate c-QO with OH and subsequent ring-opening.

(CH2)2NOOHCHCH2 and the simultaneous separation of O(CH2)2NOCHCH2 and OH is found to be exothermic (ΔH = −22.22 kcal/mol), but has to overcome a very high energy barrier of Ea = 42.63 kcal/mol. As this isomerization reaction is highly unfavorable as compared to the H-shifting isomerization reactions in the previous two cases, this reaction pathway was not investigated further. 3.4. Oxygen Addition to Morpholinyl Radicals: Direct QOOH Ring-Opening. We have identified an additional RO2−QOOH isomerization for R = 2•, with a low-energy path to ring-opening, as illustrated in Figure 5. In this mechanism, the 2• + O2 adduct undergoes a 1,5-H shifting isomerization via TS6, containing a six-membered ring, to form Q6OOH. Two different internal H-abstractions can compete in R2O2 via TS4 and TS6 to create two different isomeric QOOH radicals. Both barrier heights are below the energy of 2• + O2, and the TS6



ASSOCIATED CONTENT

S Supporting Information *

Information for the geometries, rotational constants, and harmonic frequencies of the species listed in Table 1. This material is available free of charge via the Internet at http:// pubs.acs.org



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.



Figure 5. Relative energy (kcal/mol, including ZPE corrections) diagram at the CCSD(T)/aug-cc-pVTZ level for the reaction of 2• with O2 with the same reactants as in Figure 2, but through a low barrier pathway for the ring-opening reaction.

ACKNOWLEDGMENTS This work was performed at the Brookhaven National Laboratory under the Contract No. DE-AC02-98CH10886, 7126

dx.doi.org/10.1021/jp3014496 | J. Phys. Chem. A 2012, 116, 7123−7127

The Journal of Physical Chemistry A

Article

and used the resources of the National Energy Research Scientific Computing Center (NERSC) under the Contract No. DE-AC02-05CH11231, with the U.S. Department of Energy and supported by its Division of Chemical Sciences, Office of Basic Energy Sciences. The authors are grateful to Dr. Gregory E. Hall for stimulating discussions and critical reading of the manuscript.



REFERENCES

(1) Nau, P.; Seipel, A.; Lucassen, A.; Brockhinke, A.; KohseHöinghaus, K. Exp. Fluids 2010, 49, 761−773. (2) Wu, C.; Tree, D.; Baxter, L. Proc. Combust. Inst. 2007, 31, 2787− 2794. (3) Stubenberger, G.; Scharler, R.; Zahirović, S.; Obernberger, I. Fuel 2008, 87, 793−806. (4) Di Nola, G.; de Jong, W.; Spliethoff, H. Fuel Process. Technol. 2010, 91, 103−115. (5) Wang, Z.; Lucassen, A.; Zhang, L.; Yang, J.; Kohse-Höinghaus, K.; Qi, F. Proc. Combust. Inst. 2011, 33, 415−423. (6) Kohse-Höinghaus, K.; Oßwald, P.; Cool, T. A.; Kasper, T.; Hansen, N.; Qi, F.; Westbrook, C. K.; Westmoreland, P. R. Angew. Chem., Int. Ed. 2010, 49, 3572−3597. (7) Lucassen, A.; Oßwald, P.; Struckmeier, U.; Kohse-Höinghaus, K.; Kasper, T.; Hansen, N.; Cool, T. A.; Westmoreland, P. R. Proc. Combust. Inst. 2009, 32, 1269−1276. (8) Lucassen, A.; Labbe, N.; Westmoreland, P. R.; Kohse-Höinghaus, K. Combust. Flame 2011, 158, 1647−1666. (9) Frisch, M. J. et al., Gaussian 09, revision A.1; Gaussian, Inc.: Wallingford, CT, 2009. (10) Lee, C. T.; Yang, W. T.; Parr, R. G. Phys. Rev. B 1988, 37, 785− 789. (11) Becke, A. D. J. Chem. Phys. 1993, 98, 5648−5652. (12) McLean, A. D.; Chandler, G. S. J. Chem. Phys. 1980, 72, 5639− 5648. (13) Krishnan, R.; Binkley, J. S.; Seeger, R.; Pople, J. A. J. Chem. Phys. 1980, 72, 650−654. (14) Dunning, T. H., Jr. J. Chem. Phys. 1989, 90, 1007−1023. (15) Feller, D. J. Chem. Phys. 1992, 96, 6104−6114. (16) Ř ezác,̌ J.; Hobza, P. J. Chem. Theory Comput. 2011, 7, 685−689. (17) Jensen, F. Introduction to Computational Chemistry; John Wiley & Sons, Ltd.: West Sussex, England, 2007. (18) Varandas, A. J. C. J. Chem. Phys. 2007, 126, 244105(1)− 244105(15). (19) Kendall, R. A.; Dunning, T. H., Jr.; Harrison, R. J. J. Chem. Phys. 1992, 96, 6796−6806. (20) Woon, D. E.; Dunning, T. H., Jr. J. Chem. Phys. 1993, 98, 1358− 1371. (21) Pople, J. A.; Head-Gordon, M.; Raghavachari, K. J. Chem. Phys. 1987, 87, 5968−5975. (22) Bartlett, R. J.; Purvis, G. D., III. Int. J. Quantum Chem. 1978, 14, 561−581. (23) Wayner, D. D. M.; Clark, K. B.; Rauk, A.; Yu, D.; Armstrong, D. A. J. Am. Chem. Soc. 1997, 119, 8925−8932. (24) Andersson, M. P.; Uvdal, P. J. Phys. Chem. A 2005, 109, 2937− 2941. (25) SenGupta, S.; Indulkar, Y.; Kumar, A.; Dhanya, S.; Naik, P. D.; Bajaj, P. N. J. Phys. Chem. A 2010, 114, 7709−7715. (26) Fernandes, R. X.; Zádor, J.; Jusinski, L. E.; Miller, J. A.; Taatjes, C. A. Phys. Chem. Chem. Phys. 2009, 11, 1320−1327. (27) Silke, E. J.; Pitz, W. J.; Westbrook, C. K.; Ribaucour, M. J. Phys. Chem. A 2007, 111, 3761−3775. (28) Zádor, J.; Taatjes, C. A.; Fernández, R. X. Prog. Energy Combust. Sci. 2011, 37, 371−421. (29) Yu, H. -G. Chem. Phys. Lett. 2007, 441, 20−24.

7127

dx.doi.org/10.1021/jp3014496 | J. Phys. Chem. A 2012, 116, 7123−7127