The initial stages of dichloromethane pyrolysis C2 and C3 product formation

Get Complete Project Material File(s) Now! »

THEORETICAL TREATMENT OF NON- RADICAL CHLOROETHYLENE / ACETYLENE RECOMBINATION

Concerns were outlined in the preceding chapter regarding the current understanding of mechanisms of chlorinated C4 product formation from C2 precursors. Early kinetic models estimated the rates of chlorinated vinyl radical decomposition, which, more accurate measurements suggest, may have been too slow. Thus, the importance of growth via such radicals may have been overemphasised. A series of novel ethylene‐acetylene and acetylene‐acetylene recombination channels have been considered in response to these deficiencies. We show such routes explain product yields in chlorinated systems well, have energies similar to those of competing ethynyl‐radical based reactions and become increasingly more facile with chlorine substitution.

Review of Nonradical C4 Growth Mechanisms

Non-radical recombination has primarily been assumed to proceed, if at all, via vinylidene addition to acetylene. The formation of vinylidene congeners from acetylenes was discussed in Chapter 4.1.3; it is with the mechanism of vinylidene addition, and rearrangement of subsequent adducts, that this section is concerned with. The role of other molecular addition reactions will also be discussed here. Hehre and Pople studied the minima on the C4H6 PES computationally, and found that a number of non-cyclic (for example, 1- and 2-butyne) and cyclic (cyclobutane and methylenecyclopropane, in particular) isomers are relatively stable, with trans-1,3-butadiene the global minimum;330 more recent works tend to confirm the validity of their results.258,331-333 As such, these compounds appear as probable candidates for the intermediates in the reaction of vinylidene with ethylene; in particular, the stability and structural similarity has led to much speculation regarding the role of methylenecyclopropane in addition reactions. Indeed, the experimental work of Davison et al.334 argues that the decomposition of methylenecyclopropane to ethylene and acetylene proceeds via a vinylidene mechanism, and their kinetic data suggests that vinylidene addition is almost barrierless. Supporting this, in what we believe are the only studies to date concerning chlorinated congeners, Lu and Wang have considered the reaction between dichlorovinylidene, Cl2C=C:, and C2H4 at the CCSD(T)/6-31G*//MP2/6-31G* and CCSD(T)/6-31G*//DFT/B3LYP/6-31G* levels of theory.335 Their results indicate that an initial adduct forms in a barrierless reaction and easily isomerises to dichlorinated methylenecyclopropane and cyclobutene isomers. They also note that these reactions are predicted to proceed more easily than in the non-chlorinated analogue.336 Further, methylenecyclopropanes readily lead to the C4H6-structures that we observe, albeit in relatively low abundances; alongside ethylene and acetylene, methylenecyclopropane pyrolysis has been shown to produce 1,3-butadiene.337 The C4H4 PES has also been subject to extensive study. A number of computational studies agree that vinylacetylene is the global minimum, with butatriene, methylenecyclopropane, and cyclobutadiene (in order of decreasing stability) representing the most stable isomers;330,338-340 by analogy with the C4H6 surface, these should represent the likely rearrangement intermediates. Early work into the vinylidene/acetylene reaction by pyrolysis of C2H2 and C2D2 show only a minor kinetic isotope effect when measuring the rate of acetylene dimerisation, consistent with the addition of a reactive isomer (presumably formed by a prior migration of an H atom) to H-C≡C-H in the formation of vinylacetylene. This is strong evidence toward vinylidene existing in equilibrium with acetylene; calculation of pre-exponential factors suggests that addition involves an asymmetric transition state in which the reacting species mostly retain their original geometries.341 Experiments examining the reverse reaction, vinylacetylene decomposition to acetylene, also finds little evidence of H-atom radical chain reactions; however, the authors also find that the activation energies determined appear consistent with a 2,3H migration in concert with the breaking of the C-C single bond in vinylacetylene, yielding H-C≡C-H and H2C=C:,287,342 and as such, reaction does not appear to pass through any of the stable isomers quoted earlier. In contrast to this, a number of much more recent matrix isolation studies into the addition of (usually fluorinated) vinylidene to acetylene has provided very strong evidence for the presence of intermediate species.343,344 Vinylidenes produced photolytically in matrices at approximately 10 K after UV irradiation ( λ < 248 nm) undergo rapid addition to the remaining acetylenic molecules upon annealing at 3040 K to form methylenecyclopropene derivatives. Irradiation at 420 nm generally revealed formation of vinylacetylene and butatriene signals at the expense of methylenecyclopropene. The reaction of difluorovinylidene with acetylene, for example, yields (difluoromethylene)cyclopropene, and following irradiation was shown to give 1,1-difluorobut-1-ene-3-yne.343 Methylenecyclopropene decomposition may be direct, although there is some suggestion it leads first to allenylcarbene, H2C=C=CR-CR: (R =H, F), which yields butatriene prior to vinylacetylene formation.345 Irrespective of the exact route, intermediate species are clearly feasible in the molecular dimerisation of acetylene. Maier and Lautz have also performed a series of UV-photolyses on matrix isolated acetylene mixtures; however, their approach involves excitation of the matrix, not the reagent, and thus inducing non-photolytic reaction.346 Irradiation of acetylene mixtures in a Xe matrix, particularly with longer wavelength radiation (248 nm), also yields the acetylene dimer, vinylacetylene, without the presence of radicals; however, in contrast with vinylidene-based processes, cyclobutadiene is observed in place of methylenecyclopropene. The authors suggest a mechanism via a 1,4-biradical, .HC=CH-CH=CH., as proposed by Benson,306 which either cyclises to cyclobutadiene, or forms vinylacetylene in a 1,3-H-shift; the rearrangement of cyclobutadiene to vinylacetylene also undoubtedly plays a role in high temperature systems. However, a rigorous treatment of non-radical acetylene dimerisation processes not involving vinylidene does not appear to have been undertaken. Examining the exhaustive searches of C4H4 PESs330,338-340 in detail reveals that dissociation of the cycloprop-3-ene-methylene carbene (which we give the label 1m when we introduce it in Figure 5.13) to two acetylene molecules, appears to be the only route, other than methylenecyclopropene decomposition to vinylidene and acetylene, leading from a C4H4 isomer to two C2 fragments. To our knowledge, however, formation of this carbene during acetylene pyrolysis as an alternative mechanism to molecular growth has not been subject to serious consideration given the high barrier of dimerisation, ~200 kJ mol-1, and low reverse barrier due to its instability. However, chlorinated congeners do not appear to have been considered. Given the observed increasing reactivity of acetylenes with increased chlorine substitution, a feature not readily explainable by vinylidene-based routes given a relative insensitivity of rearrangement and addition barriers with increased chlorination, and the apparent need for non-radical C4-formation routes (see Chapter 4.4.3), this chapter focuses largely on routes based on congeners of 1m.

Benchmarking Calculations Relevant to Molecular C2 Adducts

Achieving chemical accuracy without incurring impractical computational expense is of great importance. We present a number of benchmarking calculations to assess which model will serve us the best in these investigations. Diels-Alder cyclisations, H and Cl migrations, and HCl eliminations will all be considered. We will first find the optimum computational method to search the PESs; we will then test the G2MS level against these benchmarking reactions to ascertain its reliability for later high level calculations to support our results.

Benchmarking Calculations for Searching C2 Dimerisation PESs

Owing to the presence of a number of heavy atoms, the only methods accounting for correlation energy while remaining computationally tractable are DFT and MP2 approaches. Within the DFT framework, all structures were optimised using the B3LYP functional.157,156 6-31G* basis sets have been utilised for optimisation with both DFT and MP2 approaches, also with a view to reducing computational expense. ZPECs have been included using DFT/B3LYP/6-31G* (henceforth referred to as DFT-opt) frequencies exclusively, and have been scaled by 0.9806.347 Various single point energy schemes have been investigated, restricted to the same method employed for the optimisation but with a larger basis set. DFT methods have received substantially more attention due to the expected importance of unimolecular HCl elimination processes in small chlorinated hydrocarbons;348 as has been noted in numerous cases, DFT methods tend to reproduce HCl elimination reactions better than other methods. Ferguson et al.349 found the DFT/B3PW91/6-31G(d’,p’) method reproduced HCl eliminations from CH3CH2Cl, CH3CH2CH2Cl, and CF3CH2CH2Cl to within 2 kJ mol-1. While analysing 1,2-HF, HCl and ClF eliminations from CH2FCH2Cl, Rajamumar and Arunan169 found that, while B3LYP methods underestimate HCl elimination barriers by approximately 20 kJ mol-1 for not only the title compounds, but also chloroethane, they did note very good agreement for 1,2-dichloroethane. Further, higher level methods (for example, CCSD and MP2) tended to overestimate experimentally determined barriers by approximately 20 to 30 and 40 to 50 kJ mol-1 respectively. Preference for DFT over MP2 methods is justified in Table 5.1 where we have considered the deviations of energy barriers from experimental activation energies, or barriers from very high-level calculations, relative to values predicted from a variety of benchmarking methods. The reactions were chosen on the basis of their similarities to processes expected in our studies, and are shown explicitly in Figure 5.1. In Table 5.1, columns in italics show the energy barriers of structures fully optimised at the quoted level of theory; all other columns represent single point energy calculations based on these structures. We see that our chosen method must be the DFT/B3LYP/6-31+G*//DFT/B3LYP/6-31G* (henceforth referred to as DFT-SPE) level. Despite only a modest increase in the size of the basis set, this level has both the smallest maximum and average deviations from the published data of all of the methods considered; furthermore, for all reactions considered, DFT clearly outperforms MP2 calculations. It is clear from Table 5.1 that DFT-SPE calculations reproduce the 3- and 4centre HCl elimination barriers from trans-C2H2Cl2 [reactions (5.c) and (5.d)] calculated at the QCISD(T)/6-311+G(d,p)// MP2/6-31G* level of theory204 very closely (as well as matching the Cl and H migrations of reactions (5.a) and (5.b), calculated at the same level of theory). The experimental values for reaction (5.e) are less well reproduced. However, the DFT-opt level results provide a much better estimate; in fact, they reproduce exactly the value measured for reaction (5.e) by Rajakumar et al.352 who also quote 241.8 ± 8.4 kJ mol-1. Also of great importance is reproducing Diels-Alder cycloaddition barriers [reaction (5.f)], as most adducts considered are formed via steps leading to 3-membered rings; there is very good agreement with our results and those determined experimentally, with an activation energy of 100.8 ± 12.9 kJ mol-1.

READ WHAT IS CERVICAL CANCER?

Benchmarking High Level G2MS Calculations

Compound methods are approaches that can achieve near-chemical accuracy without incurring impractical computational expense. The first such model to attain widespread use and success was the Gaussian-2 (G2) approach,353,354 where MP2/6-31G(d) geometries are subject to QCISD(T) SPE calculations with the relatively small 6-311G(d) basis set. A number of MP4 SPE calculations are also performed with this and several other larger basis sets in order to try and correct for errors in the QCI energy arising from the use of small basis sets. HF-level ZPEs are also employed. Although highly accurate, and far less expensive than QCISD(T) approaches with large basis sets, some of these steps may still be prohibitively time-consuming and memory-intensive for application to larger systems. To this end, the analogous G2MS method has been developed: E(G2MS) = E[CCSD(T)/6-31G(d)] + E[MP2/6-311+G(2df,2p)] – E[MP2/6-31G(d)] + HLC G2MS + ZPECDFT All energies are obtained from single point corrections on DFT/B3LYP/6-31G(d) geometries, with the (scaled)347 ZPECDFT coming from the same calculations. HLCG2MS is an empirically determined high-level correction factor, which is commonly deemed negligibly small. This is a far less expensive approach than the G2 method, but also attains near-chemical accuracy.357,358 Benchmarking calculations have again been performed, with the results in Table 5.2. Both the maximum and average deviations in energy are very comparable to the DFT-SPE approach; thus, G2MS energies will be a very useful supplement to the DFT-SPE values, although will be employed only sparingly given their far higher computational expense.

Computational Results I: Ethylene/Acetylene Adducts

We begin searches for vinylidene-free, non-radical, vinyl- and diacetylene (and, to a lesser extent, butadiene) formation routes operative in chloroethylene pyrolysis systems by considering the ethylene/acetylene reaction. After locating the initial bimolecular addition routes, Chapter 5.3.1, rearrangements yielding our observed products (passing through intermediates such as methylenecyclopropane and cyclobutene congeners where necessary) will be considered in Chapters 5.3.2 to 5.3.5, with an overview presented in Chapter 5.3.6. The ethylene-acetylene reactions considered here are, to our knowledge, novel processes.

Ethylene/Acetylene Adducts: The initial adduct

Attempts to optimise chloro-congeners of ethylene/acetylene analogues of the .CH=CH-CH=CH. biradical suggested by Benson306 invariably failed. However, the novel transition state structures, TS1a-h (a-h denotes the congener pair considered) connecting minima associated with the ethylene/acetylene pair and methylenecyclopropane-based biradical congeners 1a-h were located instead. The energies and geometries of the eight ethylene/acetylene pairs a-h considered are given in Table 5.3 and Figure 5.2 respectively.

Table of Contents
ABSTRACT
ACKNOWLEDGMENTS
TABLE OF CONTENTS
TABLE OF FIGURES
TABLE OF TABLES
GLOSSARY
CHAPTER 1.DICHLOROMETHANE PYROLYSIS: ITS IMPORTANCE AND APPLICATIONS
1.1 .SIGNIFICANCE OF DCM COMBUSTION AND PYROLYSIS STUDIES
1.2 .THE MECHANISMS OF PAH AND FULLERENE PRECURSOR FORMATION: AN  EXPERIMNTAL AND THEORETICAL INVESTIGATION
CHAPTER 2. EXPERIMENTAL TECHNIQUES AND COMPUTATIONAL METHODOLOGY
2.1 EXPERIMENTAL TECHNIQUES AND ANALYTICAL PROCEDURES
2.2 QUANTUM CHEMISTRY: THEORY AND APPLICATION
CHAPTER 3.THE INITIAL STAGES OF DICHLOROMETHANE PYROLYSIS: C2 AND C3 PRODUCT FORMATION
3.1 LITERATURE REVIEW OF THE INITIAL DEGRADATION STEPS OF DCM
3.2 PREVIOUS IR LPHP STUDIES OF DCM
3.3 FURTHER IR LPHP STUDIES OF DCM
3.4 CONCLUSIONS DRAWN FROM DCM PYROLYSES AND THE DIRECTION OF THIS WORK CHAPTER 4.THERMAL DECOMPOSITION AND RECOMBINATION OF CHLORINATED C2 HYDOCARBONS
4.1 PREVIOUS WORK ON C2 DECOMPOSITION AND C4 FORMATION
4.2 PRELIMINARY EXPERIMENTAL RESULTS ON C2 DECOMPOSITION AND C4 GROWTH
4.3 RESULTS OF LOW TEMPERATURE STUDIES OF C2 DECOMPOSITION AND TRAPPING EXPERIMENTS
4.4 DISCUSSION OF EXPERIMENTAL RESULTS CONCERNING CHLOROETHYLENE DECOMOSITION AND GROWTH OF C4 PRODUCTS
4.5 EXPERIMENTAL STUDIES OF C4 PRODUCTS: CONCLUDING REMARKS
CHAPTER 5.THEORETICAL TREATMENT OF NON-RADICAL CHLOROETHYLENE/ACETYLENE RECOMBINATION
5.1 REVIEW OF NONRADICAL C4 GROWTH MECHANISMS
5.2 BENCHMARKING CALCULATIONS RELEVANT TO MOLECULAR C2 ADDUCTS
5.3 COMPUTATIONAL RESULTS I: ETHYLENE/ACETYLENE ADDUCTS
5.4 COMPUTATIONAL RESULTS II: ACETYLENE/ACETYLENE RECOMBINATION
5.5 DISCUSSION OF RADICAL AND NONRADICAL ROUTES IN C4 GROWTH
5.6 ON THE FORMATION OF C4 PRODUCTS IN CHLORINATED SYSTEMS: CONCLUSIONS
CHAPTER 6.
6.1 PREVIOUS WORK INTO THE FORMATION OF AROMATIC RINGS IN HIGH TEMPERATURE SYSTEMS
6.2 EXPERIMENTAL STUDIES INTO CHLORINATED BENZENE FORMATION
6.3 THEORETICAL STUDIES INTO CHLORINATED BENZENE FORMATION
6.4 KINETIC MODEL OF AROMATIC FORMATION IN CHLORINATED SYSTEMS
6.5 DISCUSSION AND CONCLUDING REMARKS
CHAPTER 7.THE FIRST FUSED RINGS: PRODUCTION OF PHENYLACETYLENES AND NAPHTHALENES
7.1 LITERATURE REVIEW OF THE C8 AND C10 FORMATION PATHWAYS IN HIGH TEMPERATURE SYSTEMS
7.2 EXPERIMENTAL INVESTIGATIONS REGARDING CHLORINATED NAPHTHALENE FORMATION
7.3 COMPUTATIONAL STUDIES OF CHLORINATED C8 AND C10 FORMATION
7.4 PROBABILISTIC KINETIC MODELS OF CHLORINATED C8 AND C10 FORMATION
7.5   TRICB CO-PYROLYSES: FURTHER MODEL AMENDMENTS AND CONFIRMATION OF THE HACA MECHANISM
7.6 CHLOROETHYLENEONLY PYROLYSES: THE EFFICACY OF THE HACA AND C4CL4ADDITION MECHANISM IN A GENERAL SYSTEM
7.7 C8 AND C10 FORMATION FROM CHLORINATED PRECURSORS CONCLUSIONS
CHAPTER 8.IMPORTANT MODELS IN PAH/FULLERENE GROWTH: PROTOTYPES BASED ON C12 SPECIES
8.1 LITERATURE REVIEW OF C12 FORMATION IN HIGH TEMPERATURE SYSTEMS
8.2 THE CHEMISTRY OF HEPTACHLOROACENAPHTHYLENE: 5MEMBERED RING SHIFTS AND 3C BAY ACETYLENE ADDITIONS
8.3 THE CHEMISTRY OF HEXACHLOROBIPHENYL: PAH DIMERISATION AND 4C BAY ACETYLENE ADDITIONS
8.4 –   PROTOTYPICAL STEPS IN PAH/FULLERENE FORMATION CONCLUSIONS
CHAPTER 9 .CONCLUSIONS AND FUTURE WORK
CHAPTER A2. EXPERIMENTAL TECHNIQUES AND COMPUTATIONAL METHODOLOGY
CHAPTER A6. THE FIRST AROMATIC RINGS: PYROLYTIC FORMATION OF THE CHLOROBENZENES
CHAPTER A7. THE FIRST FUSED RINGS: PRODUCTION OF PHENYLACETYLENES AND NAPHTHALENES
CHAPTER A8. IMPORTANT MODELS IN PAH/FULLERENE GROWTH: PROTOTYPES BASED ON C12 SPECIES
REFERENCES