VIBRATIONAL ANALYSIS AND AB INITIO

STUDIES OF PROPIOLIC ACID

by

EDMUND MOSES NSO NDIP, B.S., M.S.

A DISSERTATION

IN

CHEMISTRY

Submitted to the Graduate Faculty of Texas Tech University in Partial Fulfillment of the Requirements for the Degree of

DOCTOR OF PHILOSOPHY

Approved

August, 1987 ACKNOWLEDGEMENTS

My educational experiences at Texas Tech University have been very rewarding and along the way, I have had the good fortune of being associated with a lot of people. Consequently, I am indebted to them and wish to express my sincere gratitude for their help. First of all, I wish to thank Professor R. L Redington, my research mentor, for his ever enduring support, guidance and patience throughout the course of my study. My special thanks also go to Professor R. E. Wilde, Jr. for his friendship throughout the years. To the members of my committee, I say thank you for your patience. My thanks also go to Dr. J. L Mills, Dr. R. D. Larsen, and Dr. Jim Liang and his associates at the University of Utah, Chemistry Department. I do recognize here the goodwill of Prof. Josef MichI of the University of Texas at Austin (formerly of the University of Utah). Financial support was received from Texas Tech University and the Robert Welch Foundation. I must thank the government and people of Cameroon for the scholarships given me throughout the many years of my education. To my kids, Edmund, Jr. and Laura, I say thanks for the smiles throughout the difficult moments that we all shared. I especially thank my wife, Grace Manyi Ndip, for the sacrifices she has made over the years and for her help in preparing parts of this dissertation. TABLE OF CONTENTS

ACKNOWLEDGEMENTS ABSTRACT LIST OF TABLES LIST OF FIGURES LIST OF SCHEMES

I INTRODUCTION TO THE STUDY OF PROPIOLIC ACID Literature Study and Review Propiolic Acid~An Attractive Model Scope of the Present Work II EXPERIMENTAL TECHNIQUES Matrix Isolation Technique Matrix Materials and Properties Chemicals Instrumentation Sample Preparation Sample Deposition III ASSIGNMENT TECHNIQUES IV THEORETICAL BACKGROUND TO MOLECULAR MECHANICS Introduction Fundamentals of the Molecular Mechanics of Propiolic Acid V VIBRATIONAL ASSIGNMENTS: INTERPRETATIONS AND DISCUSSION Introduction

Vibrational Assignments and Interpretations Conclusions VI MATRIX EFFECTS Introduction Theories of Matrix Shifts Matrix Materials Crystal Data Matrix Enviromental Effects Results and Discussion Conclusions VII THEORETICAL BACKGROUND TO AB-INITIO CALCULATIONS FOR PROPIOLIC ACID Total Energy Evaluation Basis Sets Evaluation of the Total Molecular Energy Geometry Optimization Force Constant Evaluation and Vibrational Analysis Evaluation of One - Electron Properties Unimolecular Reactivities-Calculation of the Potential Energy Surface Electron Correlation and Configuration Interaction

VIII MOLECULAR ORBITAL STUDIES OF PROPIOLIC ACID

Computational Details iv Results Further Discussion

Overall Conclusions from MO Studies IX OVERALL CONCLUSIONS REFERENCES APPENDICES A1. CALIBRATED VIBRATIONAL FREQUENCIES OF PROPIOLIC ACID (PA) A2. CALIBRATED VIBRATIONAL FREQUENCIES OF DIDEUTERATED PROPIOLIC ACID (PA-D2) A3. CALIBRATED VIBRATIONAL FREQUENCIES OF MONODEUTERATED PROPIOLIC ACID (PA-OD) B1. OBSERVED WATER FREQUENCIES (CM-"") ABSTRACT

A vibrational analysis and ab initio studies of propiolic acid have been carried out in a two part study. In the first part, infrared matrix isolation spectra of propiolic acid isolated in solid argon, carbon monoxide, nitrogen and neon at 11-35K have been recorded in the range 4000 - 400 cm"^. Spectra were also recorded for the isotopically labeled O and H isotopomers isolated in argon and nitrogen matrices. Spectra have been interpreted using isotopic splitting patterns, correlations with spectra of related molecules, and MO normal coordinate analysis at the ab initio (6-31G* basis set) level and semi-emperical MINDO / 3 level using standard basis set. Computational studies using MINDO / 3 and GAUSSIAN 82 have been carried out to determine geometries, energies, dipole moments, rotational constants, vibrational frequencies, force constants, and one- electron properties. Comparisons have been made with experimental data to check the accuracy of computed molecular parameters. Various unimolecular decomposition channels have been investigated and possible bimolecular decomposition channels postulated. The computed structures and energies of various intermediates have been determined. The internal rotation (torsional) barrier for hydroxyl group relative to the C-0 bond has been determined at various levels and the conversion from the cis to trans conformer has been identified as a prerequisite for some decomposition channels.

VI LIST OF TABLES

1.1 Rotational Constants and Centrifugal Distortion Constants (MHz) 1.2 Moments of Inertia and Hydrogen Atom Coordinates 1.3 Stark Effect Measurements 1.4 Spectroscopic Constants of Propiolic Acid 1.5 lETS Vibrational Data for Chemisorbed Propiolic Acid 1.6 Theoretical Total Energies 2.1 Selected Matrix Properties 2.2 Electrical Properties of Matrix Materials 2.3 Thermal Properties of Matrix Materials 2.4 Site Diameters of Matrix Materials 4.1 Cartesian Coordinates (3N) of Propiolic Acid 4.2 Internal Coordinates (R) 4.3 Internal Symmetry Coordinates for Propiolic Acid 5.1 Fundamentals of Normal Propiolic Acid Compared 5.2 Fundamental Frequencies of Deuteriopropiolic Acid (PA-D2) 5.3 The Fundamental Vibrational Modes of Propiolic Acid

5.4 S/^(HC=C) ^°'' Pi'op'Of^y' Chloride and Propionyl Fluoride 5-5 ^(P-C=C) ^ ^(H-C=C) Ratios for the H-C^C Deformation

5.6 Summary of Assignments for PA-D2 in the 680 - 400 cm-1 Region

5.7 Fundamental Frequencies of Matrix isolated Propiolic Acid Monomers in Argon vii 5.8 Fundamental Frequencies of Monomer Propiolic Acid Isolated in Nitrogen 129 5.9 Fundamentals of Monomer Propiolic Acids Isolated in Carbon Monoxide 130 5.10 Comparison of Observed and MINDO / 3 Calculated Frequencies 131 5.11 Comparison of Calculated versus Observed Frequencies for Ne / PA 132 6.1 Site Diameters 163 6.2 Oxygen-18 Congeners of HCCCOOH 6.3 "• ^O Congeners for Fundamentals of DCCCOOH 6.4 Frequencies for DCCCOOD Oxygen-18 Congeners 6.5 HCCCOOD Oxygen-18 Congeners 6.6 Matrix Shifts for the HCCCOOH Fundamentals in Various Matrices 6.7 Comparison of Matrix Isotope Effects between HCCCOOH and DCCCOOD (VHCCCOOH ' ^DCCCOOD) 6.8 Comparison of Matrix Isotope Effects between HCCCOOH and DCCCOOH (VHCCCOOH " ^DCCCOOH) 6.9 Comparison of Matrix Isotope Effects between DCCCOOH and DCCCOOD (VQCCCOOH " ^DCCCOOD) 6.10 Comparison of Matrix Isotope Effects between HCCCOOH and HCCCOOD (VHCCCOOH " ^HCCCOOD) 6.11 Comparison of Matrix Isotope Effects between HCCCOOD and DCCCOOD (VHCCCOOD " ^DCCCOOD) 7.1 One - Electron properties

8.1 Geometry of Cis Propiolic Acid, Structure (I)

8.2 Geometry ofTrans Propiolic Acid, Structure (II)

8.3 Dipole Moments of Cis Propiolic Acid, Structure (I)

viii 8.4 Dipole Moments of Trans Propiolic Acid, Structure (II) 218 8.5 Rotational Parameters [(Calculated versus Experimental (GHz))] 219 8.6 Moments of Inertia 220 8.7 Cis / Trans Relative Stabilization Energies 221 8.8 Symmetry Coordinates of Propiolic Acid 226 8.9 Computed Diagonal Force Constants for Stnjcture (I) 227 8.10 Deviations in Calculated Diagonal Force Constants 228 8.11a Force Constants for Propiolic Acid (cis conformer -

8.11b 6-31G*) 229 8.11c Force Constants for Propiolic Acid (cis): 6-31G 230 8.12 Force Constants for Propiolic Acid (trans): 6-31G 231 4-31G Diagonal Force Constants for the Carboxyl 8.13 Group Modes Compared 234 Calculated Harmonic Frequencies (cm"^) for Cis Propiolic Acid, Structure (I) 235 8.14 Calculated Harmonic Frequencies (cm-1) for Trans Propiolic Acid, Stmcture (II) 236 8.15 Thermochemistry from Frequency Calculations for Cis Propiolic Acid, Structure (I) 239 8.16 Thermochemistry from Frequency Calculation for Trans Propiolic Acid, Structure (II) 240 8.17 Stabilization Energies and the Effects of Electron Correlation (kcal / mol.) 242 8.18 Total Atomic Charges for Propiolic Acid cis Propiolic Acid 243 8.19 ST0-3G Optimized Parameters for Cis PA Inversion to Trans PA 245 8.20 3-21G Optimized Parameters for Propiolic Acid Inversion Barrier 246 ix 8.21 6-31G* Optimized Parameters for Propiolic Acid Inversion Barrier 247 8.22 Molecular Parameters for Description of Intrinsic Reaction Coordinate 259 8.23 Molecular Parameters for Local Minima^nd Maxima on Reaction Coordinate (RC = Co - Hy) at the 6-31GSCF level 263 8.24a Calculated Harmonic Frequencies of (V) 267 8.24b Geometry, Energy, Dipole Moment of (V) at 6-31G level 268 8.25 Optimized Geometries and Corresponding Energies for HCCHCOO (T^) at 6-31G SCF level 275 8.26 Decomposition of trans - HCCHCOO (T-|) Complex, 4^2 277 8.27 Decomposition of cis - HCCHCOO (T-|) 284 8.28 Optimized Geometries and Energies for Decarbonylation Reaction Intermediate(VII) 290 8.29 Molecular Parameters for Stationary Points on Reaction Coordinate at the 6-31G SCF level 291 8.30 Harmonic Frequencies (cm"^) and Thermochemistry of the Metastable Reaction Intermediate (VII) 294 8.31 Comparison of Atomic Charges for Decarbonylation Intermediates at the 6-31G SCF level 295 LIST OF FIGURES

1.1 Microwave Model Structure of Propiolic Acid 7 2.1 Schematic Diagram of a Matrix Isolation Experiment 23 2.2 Sample Preparation Setup 30

2.3 ^®0-lsotope Exchange Reaction Vessel 32 2.4 Sample Deposition Setup 34

3.1 Effects of 50% '^ ^O-Enrichment in Oxygen Containing Molecule with Equivalent Oxygen Atoms 39 5.1 The Two Possible Structures of Propiolic Acid: I and II 59 5.2 Survey Spectrum of Propiolic Acid Isolated in solid Argon Matrix (M / S = 4500 /1) 60 5.3 Survey Spectrum of Propiolic Acid Isolated in solid Carbon Monoxide Matrix (M / S = 3000 /1) 61 5.4 Infrared Spectra of the OH and H-C Stretching regions for PA Isolated in Argon Matrix 64 5.5 Spectra of VQH ^"^^1 VQ|_| Stretching regions of Ar / PA-'' ^O 65 5.6 The OH and C-H stretching regions for PA Isolated in Nitrogen (A); Neon (B); and Carbon Monoxide (C) 66 5.7 The OH Stretching region of ^^0-labeled Propiolic Acid Isolated in Nitrogen Matrix at 11.9K and 18.3K(M/S = 3000/1) 67 5.8 The OH, C-H, OD, and C=C Stretcing regions of Ar / PA-OD at 11.4K (M / S = 3500 /1) 69 5.9 The Spectrum of N2 / PA-OD iM / S = 4000 /1) at 11.4K in the region 4000 - 2000 cm'^ 70

5.10 Spectra of N2 / PA-OD (M /S = 4000 /1) at 18.9K 71

5.11 Spectra of Ar / PA-D2 (M / S = 500 /1) in the region 4000-2000 cm"' 72 xi 5.12 Detail Spectra of CO / PA-D2 (M / S = 1000 /1) in the 4000 - 2000 cm"'showing the VQH (DCCCOOH) and VQD (DCCCOOD) bands at 12K 73 5.13 Detail Spectrum of No / PA-P2 (M / S = 1000 /1) at 12K in the region 4000 - 20O0 cm"' 74 5.14 "'^O-labeled Spectra of PA-D2 in the 0-D region 76 5.15 The y^Q=Q Stretching region of Propiolic Acid Isolated in Vanous Matrices 78 5.16 The VQ=Q region of PA-D2 Isolated in Various Matrices 79 5.17 The Carbonyl Stretching region of Propiolic Acid Isolated in Argon (A) and Neon (B) at 11.4K 80 5.18 Spectra of the VQ_O and VQ.Q Stretching regions of Ar / PA-OD at 1 f;4K (M / S = 4000 /1) 81 5.19 The Mid-frequency region of Propiolic Acid Isolated in Nitrogen 82 5.20 The Carbonyl region of N2 / PA at 10K and 20K 84 5.21 The Carbonyl Stretching region of N2 / PA-OD (M/S = 3500/1)at(a) 11.6Kand(b) 18.8K 85 5.22 The Carbonyl region of Propiolic Acid Isolated in Carbon Monoxide Matrix 86 5.23 Spectra of the ^^Q^Q region for PA-D2 87 5.24 Spectra of PA-^ ^O Isolated in Nitrogen and Argon for the Carbonyl region 88 5.25 Infrared Spectra of Ar / PA in the Mid-frequency region 90 5.26 Infrared Spectra of Ne / PA in the Mid-frequency region 91 5.27 The vn n Stretching region of N2 / PA (M / s = 3500 /1)

at lljSKand 18.8K 93

5.28 ^ ®0 - Spectra of Propiolic Acid Isolated in Ar at 11K 94

5.29 Ar / PA Spectmm at 11.4K in the region of 680 -1200 cm"'' 96

XII 5.30 Spectra in the 6QOH / VQ.Q region of Propiolic Acid Isolated in Argon and Neon Matrices 97 5.31 Spectra of Ar / PA in the 900 - 680 cm""* region 98 5.32 Fundamental Modes of Ar / HCCCOOD (M / S = 3500 /1) in the region 1200 - 680 cm"'' 99 5.33 Spectra of Ne / PA in the region 850 - 600 cm"'' 100 5.34 The Mid-frequency region of N2 / PA Spectra at 11.4K 101 5.35 Spectra of N2 / PA-OD (M / S = 3500 /1) in the 1200 - 680 cm-1 region at 11.6 and 18.8K 102 5.36 Spectrum of CO / PA (M / S = 500 /1) in the 1220 - 680 cm"' region (m = monomer, d = dimer) 103 5.37 Spectrum of CO / PA-D2 in the 1220 - 680 cm"'' region 105 5.38 The 6QQH / VQ-Q Spectra of'' ^0-labeled Propiolic Acid Isolateoin Argon (A) and Nitrogen (B1 and B2) at 11.4Kand18.3K 106 5.39 Spectra of Ne / PA in the 600 - 500 cm'"' region 110 5.40 Spectra of Ar / PA-OD in the 680 - 580 cm"'' region 111 5.41 Spectra of Ar / PA-OD in the 580 - 455 cm"'' region 112 5.42 Spectra of Ar / PA-D2 (M / S = 500 /1) in the 800 - 400 cm"'' region 113 5.43 The 680 - 500 cm"'' region of N2 / PA at 10K 114

5.44 No / PA-OD (M / S = 3500 /1) Spectra in the 680 - 400 cm"' region 115 5.45 Spectrum of No / PA-D2 (M / S = 500 /1) in the 800-400 cm'''^region at 12K 116 5 46 Spectrum of No / PA-D2 (M / S = 500 /1) in the 800-400 cm"''^region at 20K 117 5 47 The 900 - 400 cm-1 region of Propiolic Acid Isolated in solid Carbon Monoxide Matrix (M / S = 3000 /1) 118

XIII 5.48 The 800 - 400 cm"'' region of PA-D2 Isolated in Carbon Monoxide Matrix (M / S = 500 /1) 119 5.49 Spectra in the 680 - 400 cm"'' region of Ar / PA 121

5.50 ''^O - Spectrum of Propiolic Acid in the 680 - 500 cm"'' region Isolated in Argon Matrix at 12K 122 5.51 ''^O - Spectrum of Ar / PA-''80D2 in the 680 - 500 cm"'' region 123 5.52 Comparison of Calculated versus Observed Frequencies in Argon Matrix 133 5.53 Comparison of Calculated versus Observed Frequencies in Nitrogen Matrix 134 5.54 Comparison of Calculated versus Observed Frequencies in Carbon Monoxide Matrix 135 5.55 Ar / PA (M / S = 4000 /1) Spectrum in the 680 - 400 cm"'' region at 30K 137 5.56 Spectrum of Ar / PA (M / S = 4500 /1) in the 1200 - 680 cm"'' region at 30K 138 5.57 The Carbonyl and Mid-frequency regions of Ar / PA (M/S = 4500/1) at 30K 139 5.58 The High Frequency region of Ar / PA (M / S = 4500 /1) at 30K 140 5.59 Survey Spectrum of CO / PA (M / S = 3000 /1) at 12K 141 5.60 Survey Spectrum of CO / PA (M / S = 3000 /1) at 24K 142 5.61 Survey Spectrum of CO / PA (M / S = 3000 /1) at 35K 143 5 62 The 2000 -1200 cm"'' region of PA-D2 Isolated in solid Argon Matrix (M / S = 500 /1) 145 5 63 Spectra of CO / PA (M / S = 3000 /1) and CO / PA-D2 (M/S = 500/1) at 12K 146 5.64 Spectrum of CO / PA-D2 (M / S = 500 /1) in the 1220 - 680 cm"' region (m = monomer, d = dimer) 147

XIV 5.65 Details of N2 / PA-Do (M / S = 500 /1) Spectra in the 2000 -1200 cm"^ region 148

5.66 The 1200 - 680 cm"'' region of Ar / PA-D2 (M / S = 500 /1) 149

6.1 Close-packing of Spheres showing Three Layers; a, b, c, and Tetrahedral and Octahedral holes 161 6.2 Tetrahedral and Octahedral holes between Layers of Close-packed Spheres 162 6.3 The Size of Monomeric Propiolic Acid 164

6.4 Correlation of VQIH Frequencies with Polarizabilities of Matrices 180

6.5 Correlation of VQ_O Frequencies with Polarizabilities of Matrices 181

6.6 Correlation of SQQQ and xoH Frequencies with the Polarizabilities otYhe Matrices 182

6.7 Correlation of yQ.n I SrOH Frequencies with the Polarizabilities of tne Matrices 183

6.8 Correlation of "^(^.Q, Frequencies with the Polarizabilities of Matrices 184 6.9 Cage Model for Propiolic Acid Showing Idealized Spherical Cage with Neighboring Matrix Atoms (M) 187 7.1 Overall Logical Structure of a Typical Ab Initio MO Program 209 8.1 The Two Conformers of Propiolic Acid 211 8.2 Ab Initio Calculated Model Structure (6-31G**) 223 8 3 Definition of Internal Valence Coordinates for Propiolic Acid 225 8.4 Energy Profiles for the Conversion of Cis to Trans Propiolic Acid 248 8.5 Definition of the Z - Matrix (Distance Matrix) Parameters for Propiolic Acid 258

XV 8.6 Pictorial Desciption of the Intrinsic Reaction Coordinate (IRC) atthe 3-21G level 260 8.7a Energy Profile for the Unimolecular Decomposition of Propiolic Acid at the 6-31G level 261 8.7b Definition of Molecular Parameters for Intermediate Structures 262 8.8 Structure of the Metastable Intermediate (V) at the 6-31G level 265 8.9 a: TS Structure for Vinylidene - Acetylene Isomerization b: Energetics of the C2H2 Rearrangement 269 8.10 Total Atomic Charges for (V) and TS at 6-31G level 270 8.11 Structure of 4a and 4h and the Transition State for their Interconversion 274 8.12 Energy Profile and Transition State Structure for the Interconversion of HCCHCOO (T^) 276 8.13 Pictorial Representation of the Progress of Decomposition 279 8.14 Definition of Geometrical Parameters in the Decomposition of 4i> 280 8.15 Energy Profile for Decomposition of ^ (Ti) 282 8.16 Energy Profile for the Decomposition of 4a (Ti) 283 8.17 Summary of the Energetics of the Decarboxylation Process 286 8 18 Structure of Decarbonylation Reaction Intermediate (6-31G level) 289 8.19 Pictorial Representation of the Progress of Decarbonylation 292 8.20 Energy Profile for Decarbonylation of Propiolic Acid 293

XVI LIST OF SCHEMES

8.1 a Unimolecular Decomposition of Propiolic Acid 253 8.1 b Bimolecular Decomposition of Propiolic Acid 254 8.2 The Decarbonylation of Propiolic Acid 255 8.3 Unimolecular Decomposition of Propiolic Acid on the T-| Potential Energy Surface . 272 8.4 Other Possible Decay Channels in Propiolic Acid 297

XVII CHAPTER I INTRODUCTION TO THE STUDY OF PROPIOLIC ACID

The simpler carboxylic acids have been studied for a variety of reasons. For example, the dimers of acetic and formic acids provide attractive models for the study of proton tunneling [1,2,3]. In addition, there are numerous studies of their monomers along with their dimers. But even for these 'well studied' acids there are a number of unresolved problems. In addition, both acetic acid and formic acid do not possess two functional groups separated by a carbon-carbon single bond. Propiolic acid is one of the simplest molecules that satisfies this condition. Propiolic acid is a seven-atom molecule. It is soluble in a variety of solvents. It has a molecular weight of 70.09g, and a density of 1.1380 g/cc. It has a refractive index of 1.4306 and a large acid dissociation constant compared to other three-carbon acids. For example, at 20°C the pKa for propionic acid is 1.35 x 10"^ while that of propiolic acid is 1.36 x 10"^. This difference is attributed to the strong electron-withdrawing ability of the ethynyl group relative to that of other hydrocarbon substituents. A survey of all the work dealing with propiolic acid is out of the scope of the present study. The work that will be reviewed here deals with aspects relevant to its molecular geometry, molecular properties, molecular spectroscopy and where appropriate its molecular reaction dynamics. The review will encompass studies dealing with the vibrational analysis of related molecules and derivatives of propiolic acid. The review has been divided into the following general headings: 1) Infrared and Raman studies of Related Molecules 2) Infrared, Raman and other Spectroscopic Studies of Propiolic Acid 3) Thermodynamic Data 4) Review of Pertinent Ab Initio Calculations It will be obvious after this review why propiolic acid is a good model for the studies that have been carried out in the present study.

1 2 Literature Study and Review Infrared and Raman Studies of Related Molecules The first work ever done involving propiolic acid was carried out by Wilson and Wenzke. Although the carboxyl group is quite complex, they tried to determine the effects of the triple bond on the moments of a series of propiolic acids. It was concluded from this study that the closeness of the carbon-carbon triple bond had a marked effect on both the electric moment and the ionization constants of the acids studied. They concluded that the presence of the triple bond causes the carboxyl group to become more positive In character. This increased polarity of the 0-H bond is manifested by an increase in the resultant moment for the carboxyl group and an increased ionization constant for the acetylenic acids. The relative conjugative aptitudes of acetylenic and ethylenic bonds was investigated by Kochi and Hammond In 1953 [5]. Their study concluded that acetylenic groups showed low conjugative aptitudes or poor electrical transmission toward electron deficient systems. On the other hand they may exhibit acceptor properties by interaction in addition to exerting a strong dipole Influence. The tests for such interactions was investigated by Petrov and Yakovleva [6]. They showed that it is possible for acetylenic compounds to bond with amines, a fact supported by observed spectral shifts of the acetylenic group frequencies. The importance of this study and others [7] is that shifts in the H-C=C stretching frequencies can be used as proof for hydrogen bonding involving this group. A variety of infrared solution, vapor, and liquid phase spectra along with Raman spectra have been obtained for several acetylenic compounds. Nyquist et al. [8], and Nyquist [9], have proposed assignments for the fundamentals of the ethynyl portion of these molecules. Assignments have been made for the fundamentals of deuteriopropargyl chloride (D-C=C-CH2CI) based on vapor phase, liquid phase and solution infrared spectra obtained in the range 3800 - 45 cm"''. Raman data was obtained in the range 3400 -100 cm"''. Propargyl alcohol spectra have also been assigned. Brand and Watson in 1960 [10] measured the infrared spectrum of propynal in the range 3600 - 200 cm"''. A complete vibrational analysis 3 was proposed. Similar studies were performed by King and Moule in 1961 [11]. In a much more elaborate study they obtained infrared spectra for gaseous propynal and its deuterated analogues. Liquid Raman spectra were also recorded. Vibrational assignments were given and correlated for absorptions in the range 4000 - 380 cm"'' • Job and King [12] studied the infrared spectra of cyanoacetylene in the vapor, liquid and solid phases. Cyanoacetylene is an important molecule in astrophysics as it has been observed in interstellar clouds. This study was motivated by an earlier study which provided an incomplete assignment of vibrational modes. The spectra of gaseous trifluoromethyl acetylene (CF3C^H) and CF3C=CD were recorded in the range 4000 - 75 cm"'' by Berney et al. in 1963 [13]. Liquid Raman data was obtained for CF3C=CH in the same study. The first matrix isolation study on any acetylenic compound was performed by Sanborn [14] in 1967. He used the matrix isolation technique along with the temperature dependence of gas phase spectral features to investigate the acetylenic stretching bands of trifluoromethyl acetylene (TFMA). He showed that the multi-Q branched stretching infrared bands in TFMA vapor were the result of a combination of "hot" bands and Fermi resonance. In 1971, Guilleme et al. [15] studied self-association in a-ethylenic and a-acetylenic carboxylic acids. While dimers are usually thought of as being the predominant form in which these acids exist, they postulated the possible presence of linear as well as cyclic dimers. They also deduced a method for calculating equilibrium constants. For propiolic acid, the self-association constant was determined to be 725 mole"^ .1 at 25°C. The free hydroxyl stretching frequency was found to be 3518.8 cm"''. The free carbonyl stretching frequency was observed at 1743.5 cm'^ while the bonded C=0 stretch was observed at 1699.3 cm"''. In addition, a pKa of 1.89, a -AG of 2.68 kcal / bond , a -AH of 3.5 kcal / bond and an entropy , -AS of 2.8 cal / deg - bond were also determined. Williams [16] studied both the ground and first excited electronic states of propynal and its deuterated analogues. A normal-coordinate analysis was performed using the quadratic force field for the ground state and first electronic singlet states of the molecules. He used semi-empirical SCF-MO 4 theory (CNDO / 2) to correlate changes in geometry and electronic structure of excitation as well as accounting for the remaining observed electronic bands of propynal. Another theoretical study was carried out by Bournay and Marechal [17]. This study involved the dynamics of of proton tunneling in hydrogen bonded systems of propiolic and acrylic acids along with their deuterated analogues. Theoretical dimer frequencies were calculated and spectra generated. Some of the more recent spectroscopic studies include those of Augdahl et al. [18] for tetrolyl and propiolyl chloride; Balfour et al. [19] for propiolyl and deuteriopropiolyl fluoride and Brown et al. [20], who carried out lETS (Inelastic Tunneling Spectroscopy) studies on carboxylic acids. In the lETS studies surface phenomena of carboxylic acids and related species chemisorbed on plasma-grown aluminium oxide were observed. The results deduced were mixed. Theoretical vibrational analysis carried out in the present study for the propiolate ion were motivated in part by this study. More spectroscopic studies were carried out by Balfour et al. [21] in 1979 in which they recorded and measured infrared and Raman spectra for propiolyl (PC) and tetriolyl (TC) chlorides. Fundamentals were assigned. The relationship between conjugation and vibrational frequencies was investigated. They also examined the effects of free internal rotation of the methyl group on the band contours of several infrared.bands. Travert et al. [22] recorded and analysed the vibrational spectra of propargyl alcohol and its deuterated analogues. By comparing bands in the 1500 - 600 cm"'' range with combination bands involving the (OH) bending vibration between 5000 - 4500 cm"'', they were able to identify bands with OH-bending character. This study also permitted a complete assignment of all the bands in the 4000 - 200 cm"'' range. A normal coordinate analysis for the propiolyl halides and their deuterated analogues was carried out by Balfour and Phibbs [23]. As a result of their work, improved assignments for several of the COX (X=halides) bending modes were deduced. Noteworthy is a comparison of possible shifts that can occur In different environments. For example, in CH3COCI the C=0 wag is observed at ca. 500 cm"^. On the other hand, it is shifted to about 650 5 cm"'' in PC and TC. Based on a force constant calculation they showed that this vibrational mode is fairly well localized. Microwave spectra for four isotopic species of PC were measured by Davis and Gerry in 1982 [24] for the ground and vg = 1 excited vibrational states. Values were obtained for the rotational constants along with quartic centrifugal distortion constants and some molecular moments. An approximate harmonic force field was calculated by combining the centrifugal distortion constants with vibrational frequencies from the literature. Similar microwave studies have been carried out for acid halide derivatives of the simple carboxylic acids [25 - 32]. Methyl propynoate is the methyl ester of propiolic acid . It's infrared and Raman spectra were recorded by Katon et al. in 1983 [33]. They obtained liquid phase spectra although conformational and spectroscopic studies had been carried out earlier by Williams et al. [34] and by Lin et al. [35] respectively. Similar spectra were also recorded for dimethyl 1,4-butynedioate (DMBD). The infrared spectra of both compounds were also recorded in the solid phase. Based on a C2h symmetry a vibrational assignment was proposed for DMBD. Their data indicated the existence of two conformers. Methyl propynoate spectra were also assigned. Hamada et al. [36] recorded and measured the gas phase infrared absorption spectra and the low temperature argon matrix spectra of propargylamine. They performed a complete vibrational analysis with the aid of ab initio MO calculations. They also derived an effective force field which reproduced the observed frequencies. Electronic absorption spectra in neon matrices were obtained for N=C-C=N+, HC=CC=N"'", and CH3C=CC=N+ by Fulara et al. in 1985 [37]. Their vibrational interpretation led to the assignment of some fundamentals of the cations in their excited electronic states. Infrared, Raman and Other Spectroscopic Studies on Propiolic Acid. The first experimental spectroscopic study of propiolic acid was carried out by Katon and McDevitt in 1965 [38]. They obtained Raman and IR spectra for propynoic acid and its sodium salt in the range 4000 -100 cm"''. Partial spectra were also obtained for the deuterated analogues. They assigned fundamental frequencies based on spectra recorded in solution, liquid and vapor phases. This study provided the first example of a spectroscopic study of a carboxylic acid with two functional groups in such close proximity. It also raised a number of important questions regarding the stability and spectroscopy of propiolic acid. Some of these questions will be answered by both the experimental and theoretical work contained in the present study. The microwave spectra of propiolic acid and its two monodeuterated analogues were first reported in 1972 by Lister and Tyler [39]. They found that propiolic acid is a planar molecule with the hydroxyl proton adopting a cis-geometry with respect to the carbonyl group. A structure consistent with the observed moments of inertia was proposed (Figure 1.1). The dipole moment was measured to be 1.59 ± 0.03D and lying almost parallel to the carbonyl bond. They found no evidence for the existence of two rotameric forms. The microwave data from their study is summarized in Tables 1.1 and 1.2. The calculated dipole moments were obtained from Stark effect measurements (Table 1.3). The structure and conformation of propiolic acid were also discussed at great length. Since no other data relevant to the structure of propiolic acid is available, the model stnjcture proposed in Figure 1.1 has been adopted as the structure of propiolic acid in this dissertation. The molecular parameters (geometry) have been used as the starting geometrical parameters in the ab initio MO studies. From Table 1.3, it is seen that the dipole moment of propiolic acid is slightly larger than the dipole moment of formic acid but it is also a lot smaller than the dipole moment of either formaldehyde or propynal. Because the other molecules were not studied in the present work no attempts will be made to correlate these differences in dipole moment to infrared frequencies and intensities. An improved study of the microwave spectroscopy of propiolic acid was carried out by Wellington Davis and Gerry in 1976 [43]. They measured the microwave spectra of propiolic acid and propiolic acid-d up to the J=30 line. This enabled them to accurately determine rotational and centrifugal distortion constants for each isotopomer. To accomplish the above, they used Watson's Hamiltonian [44] with terms up to the sixth degree in the angular momentum : Figure 1.1. Microwave Model Structure of Propiolic Acid [39]. 8 Table 1 J: Rotational Constants and Centrifugal Distortion Constants (MHz) @

HCCCOOH HCCCOOD DCCCOOH

A 12110.09 11858.32 12109.93

B 4146.94 4015.69 3819.71

C 3084.49 2995.58 2899.63

'^aaaa -0.1476 -0.0172 -0.0355

'^bbbb -0.0076 +0.0013 -0.0101

'^cccc -0.0025 +0.0028 -0.0060

%bab -0.0426 -0.0563 -0.0160

(S): reference [39] Table 1.2 : Moments of Inertia and Hydrogen Atom Coordinates [39]

A(amu A'

HCCCOOH* 41.7447 121.9052 163.8929 0.2430

HCCCOOD* 42.6308 152.8897 168.7590 0.2369

DCCCOOH* 41.7452 132.3484 174.3432 0.2496

Model 42.1126 120.9109 163.0235

Hydroxyl H Acetylenic H a b a b(AO)

Observed 1.993 •0.968 •3.244 -0.024

Model 1.963 •1.022 •3.232 -0.052

*Calculated from rotational constants in Table 1.1 corrected for the contribution from x^i^aij and using the conversion factor 5.05531 x 10^ a.m.u.A°2 MHz. 10 Table 1.3 : Stark Effect Measurements [39]

H-CCCOOH Av/E2MHz(V/cm)"2x10"'*

Observed Calculated ''01^^00 M=0 1.76 1.79 32i<-3-|2 M=1 2.30 2.26 M=2 4.22 4.20 M=3 7.37 7.49

[IQ = 0.80 ± 0.02D, ^ib = 1.38 ± 0.02

H-CCCOOD Av/E2MHz(V/cm)"2x10"'^ Observed Calculated

^12^^01 M=0 1.69 1.69 32i<-3i2 M=3 7.34 7.30

[IQ = 0.76 ± 0.02D, ^t> = 1.40 ± 0.02D [i= 1.59 ±0.03

Compound Dipole Moment Ref.

Formaldehyde 2.33 D [40] Propynal 2.47 D [41] Formic Acid 1.40 D [42] 11 H =HR + HD + H'D (1.1)

HR = 1/2(B+C)P2 + [A-1/2(B+C)]Pa2 + 1/2(B-C)(Pb2-p2)

HD = -AjP4 • ^JKP^Pa^ - AKPa^ - (Pb^ - Pc2)[5jp2 + eKPa^]

- [5jp2 + 6KPa2](Pb2 - Pc^)

H'D = Hjp6 .. HjKP4Pa2 ^ HKjP^Pa^ + HKPa^ + (Pb^ - Pc^)

X [hjP4 + hjKP2p22 ^ h^Pa^] + [hjP^ + hjKP^Pa^ + hKPa^l

X (Pb^ - Pc^).

Each of the terms in Equation (1.1) consists of complex terms involving rotational contants (A, B and C), the components of the angular momentum (Pa, Pb' sncl PQ), the quartic distortion constants (Aj, Aj^, Aj<, dj, and d^), and the sextic constants (Hj, Hj|^, H^^j, H^, hj, hj^, and hj^). The distortion and rotational constants were manipulated accordingly to give good values for the derived molecular constants shown in Table 1.4 [43]. A study that bears closely to the vibrational analysis of propiolic acid was carried out by Brown et al. in 1978 [20]. Using lETS to observe surface phenomena for propiolic, propenoic, and 3-methyl- but-2-enoic acids on plasma-grown aluminium oxide, spectra were obtained for the corresponding carboxylate ions. The spectra for HC^C-COO", H2C=CH-C00', and (CH3)2C=CH-COO" were recorded and assigned. For the propynoate ion, the C2v symmetry was adopted in assigning the fundamentals in Table 1.5. The work cited in [20] gives complimentary data for assigning some of the fundamental vibrational modes of propiolic acid. It is also responsible in part for motivating the theoretical study of propiolic acid in the ground and first excited states. The C-COO rocking motion is assigned as the absorption at 732 cm"^. This is also the region in which one would expect to find the H-C=C acetylenic out of plane linear bending fundamental. Contrary to the convention discussed by Herzberg [45], the 12 Table 1.4 : Spectroscopic Constants of Propiolic Acid [43]

HC^-COOH HC^-COOD

A 12110.0172 ±0.0046 11858.4445 ±0.0050 B 4146.9388 ±0.0014 4015.7137± 0.0015 C 3084.4861 ±0.0012 2995.5965 ±0.0011

AjxIO"^ 5.376 ±0.146 5.205 ±0.140 AjKXl02 2.132 ±0.010 1.891 ±0.015 AKXI 0^ -7.715 ±0.045 -4.026 ±0.070

djx10^ 1.605 ±0.043 1.593 ±0.070 d^xlO^ 1.219±0.013 1.106±0.016

HjxIO^ -2.20 ± 0.63 -0.73 ±0.81 HjKXlO^ 6.88 ±1.58 4.99 ±2.64 HKJXIOS -1.99 ±0.41 -1.27 ±0.65 H^xlO^ 1.37 ±0.29 0.94 ±0.49

hjxIO^ 1.17±0.31 0.76 ±0.56 hjKXlO^ -5.90 ± 1.36 -2.91 ±2.21 h^xlO^ -5.86 ±1.19 3.95 ± 1.93 13 Table 1.5 : lETS Vibrational Data for Chemisorbed Propynoic Acid [20]

H-C=C-COO- Symmetry Class Description

lETS Na+Salt 3600 v(O-H) 3272 3278 ai ^1 v(C-H) 3082 2967 2921 2878 2828 2108 2095 ai V2 v(C^) 1631 1588 1600 b2 ^9 Vasym(C02') 1447 Vsy(C02") 1427 1364 1382 ai ^3 Vsym(C02") 1270 1056 953 928 ' 877 (sh) 891 ai ^4 v(C-C) 832 776 789 ai ^5 5(C02") 684 704 b2 ^10 6(C-H) 630 652 bi ^6 7(C-H) 585 582 b2 V11 6(CC0) 511 375 489/440 bl V7 7(0C0) 300 lattice phonon 252 256 b2 vi2 6(C^C) 206 bl ^8 7(C^C) 14 in-plane linear bending modes asssociated with the H-C=C-C group have been assigned to absorptions at higher frequencies than the out-of-plane vibrations. These and other discrepancies will be discussed in the sections dealing with the assignment of propiolic acid fundamental modes.

Thermodynamic Data There is very little thermochemical data available on heats of formation for gaseous substituted acetylenes [46, 47]. Flitcroft and Skinner [48] reported heats of hydrogenation in solution. Pedley and Rylance [47] reported heats of formation in the liquid state for several conjugated acetylenic esters and acids. There is, however, a disparity of at least 15 kcal / mol on the stabilization and destabilization effects for some substituents. Furthermore, gaseous state corrections are unknown and therefore can only be crudely estimated [49]. No heats of formation have been reported for propiolic acid. Heats of formation are usually calculated as part of the results in MINDO/3 calculations.

Ab Initio and Semi-Empirical(MINDO / 3) Calculations Hehre et al. [50] have published results of ab initio calculations at the 4-31G level for some acetylene derivatives. More results at the ST0-3G level have been published by Dill et al. [51]. While a variety of calculations have been performed on a number of relatives, the computational chemist has ignored propiolic acid. Transition states and stabilization energies have been calculated for propynal [52]. Hamada et al. [36] used ab initio frequency calculations in the vibrational analysis of propargylamine, where observed and calculated frequencies were used to obtain a force field which reproduced observed frequencies with good accuracy despite the inherent overestimation of parameters by ab initio methods. A number of ab initio calculations have been carried out that are significant to certain structural units in propiolic acid. The relationship between bond lengths and the corresponding quadratic stretching force constants has been studied extensively for diatomics. It has been difficult to 15 effect studies of this kind for polyatomics largely because of the lack of sufficient experimental data [53-56]. Extensions of these studies have been carried out for several systems. Bock et al. [57] calculated the geometry of the C-C(H)=0 group, stretching force constant f^Q.Q, and the coupling constant ^c=0,C-C using the unsealed 4-31G basis with full geometry optimization for various substituted carbonyl compounds. The same authors had previously investigated the influence of substituents and intramolecular hydrogen bonding on these carbonyl compounds [58]. They also calculated the harmonic and anharmonie force fields and the fundamental vibrational frequencies for performic acid [59]. On formic acid, an LCAO-MO description of self association types had been carried out as early as 1968 [60]. Harmonic force constants and the uniqueness of force fields derived from vibrational data have been studied by Tae-Kyu Ha et al. [61]. But while various studies have been carried out for the derivatives and relatives of propiolic acid, the only study to date on propiolic acid is that of Furet et al. [62]. They studied substituent effects on acetylene stability with the GAUSSIAN 80 series of programs. Ab initio calculations using ST0-3G, 6-31G, 6-31G** and 6-311G** basis sets and standard geometries were used. In addition, geometry optimizations were performed for a number of acetylene derivatives. The results of total energy determinations for propiolic acid and other molecules of interest are summarized in Table 1.6 [62]. Although total energies have been determined, other aspects dealing with structures, electrostatic properties, and unimolecular reactivities associated with the ground and first excited electronic states have not yet been studied. The ab initio studies of the present work have been designed to evaluate and determine most of these properties.

Conclusions from Literature Survey The above review, although inexhaustive, reveals several points : 1) The molecular spectroscopy of molecules and fragments related to propiolic acid has been explored to great lengths. Assignments for both Raman and infrared data have been discussed in various cases. Spectral 16 Table 1.6: Theoretical Total Energies [62]

Molecule (Energies in hartrees)

ST0-3G 6-31G 6-31G//6-31G 6-31G" 6-311G"

H-C^-H 75.85339 76.79261 76.79279 76.82138 76.84052

H-C=C-C=N 166.40853 168.47859 168.47912 168.54944 168.58601

H-C=CCHO 187.08088 189.45788 189.45910 189.54170 189.58578

H-C=C-COOH 260.93646 264.30564 264.30977 264.43093 264.49499

HC^C-COOCHg 299.51751 303.30614 303.45238

H-C=C-CH20CH3 226.85089 229.63427 17 information has been obtained for various derivatives in the liquid, solid, and vapor phases. Experimental force fields have been determined in certain cases. 2) There is no experimental force field based on a normal coordinate analysis for propiolic acid although a harmonic force field has been determined from microwave data. There are simply not enough fundamentals assigned with certainty for one to do a complete vibrational analysis of propiolic acid. 3) Propiolic acid has been observed and is known to be unstable. It decomposes easily to acetylene and carbon dioxide. The dynamics and energetics of this decomposition have not been studied either experimentally or theoretically. Theoretical force fields which could provide plausible guesses to the actual force field have not been determined both for the ground and first excited states. In addition, there is no knowledge of its one-electron properties apart from its dipole moment. The determination of various moments is important in the determination of molecular structure and the study of intermolecular forces. 4) Thp nature of the hydrogen bonding and the possible existence of dimers and higher polymers have not been determined either experimentally or theoretically although studies of this nature have been carried out for formic acid. Results of such studies would provide information regarding the stability of these polymeric species. 5) The question of transferability of molecular parameters including force constants between structurally related molecules has not been investigated for propiolic acid.

Propiolic Acid-An Attractive Model The proceeding sections show that there is a lot of information available in the literature on the spectroscopy of molecules and ions related to propiolic acid. Practical experience has shown that it is a difficult molecule to handle. The technique of matrix isolation is designed to handle both stable and unstable molecules and ions. Propiolic acid is a suitable molecule for the application of this technique. 18 The stnjctural relationship between acetylene, formic acid and propiolic acid cannot be overstated. It is readily seen that propiolic acid can be derived from formic acid by replacing the formyl proton with the ethynyl group. The spectra of both acetylene and formic acid have been analysed by several authors through a combination of theory and experiment. While the fundamental vibrational modes of both these molecules have been assigned to satisfaction, no such definitive assignments exist for propiolic acid. As a result of this scarcity of spectroscopic information, it is fair to assume that a study of propiolic acid will provide a common ground for correlating spectroscopic information from acetylene, formic acid, and propiolic acid. This would also provide a more concrete basis for determining the possible rules involved in transferring molecular parameters between related molecules. A number of additional points can be made to justify a study of propiolic acid. First, several fragments and molecules related to propiolic acid have been observed in interstellar clouds. Cyanoacetylene [63,64], is an important molecule in astrophysics. Formic acid [65], the ethynyl radical [66], and propyne [67] have all been observed. It is logical to assume that propiolic acid would be a possible candidate for observation. Secondly, the rather scanty literature available on propiolic acid suggests it has been a less than popular subject for experimental vibrational spectroscopy. In order to carry out any qualitative or quantitative analysis, one must have a self consistent pool of data. Such a basis can be obtained by providing a set of observed frequencies which can be assigned in a definitive manner to the different normal modes of the molecular system. It is clear that such a basis does not exist for propiolic acid. Thirdly, the determination of force fields requires that the data required be readily available. There are usually more force constants to be determined from a smaller number of observed frequencies. Redington [68] used isotopic labelling to get 163 observed frequencies for formic acid. Propiolic acid, with seven atoms, has 66 independent force constants and as such they cannot all be determined from 15 frequencies. To alleviate this problem, somewhat, experimental frequencies are combined with quantum mechanically determined force constants. Quantum mechanical 19 calculations of this nature have been carried out for small systems with a great deal of certainty. Propiolic acid is a large molecule by ordinary quantum calculation standards. Therefore, performing these type of calculations for such a large molecule would test the efficiency of ab initio methods for determining harmonic frequencies and other molecular properties. The literature involving quantum mechanically determined properties of propiolic acid is scarce. The authenticity of determined parameters will be ascertained by comparing values for computed parameters with experimentally determined values where available. The force fields, geometries and other structurally related parameters will be combined with experimental data to obtain a satisfactory and reproducible force field and model stmcture. Propiolic acid decomposes easily to acetylene and carbon dioxide. It is not known whether this decomposition is a thermal or photochemical process. The dynamics of this process are not known. Calculations providing transition state barriers, inversion barriers, product energies and possible decomposition channels would present a more complete picture in the understanding of the chemistry of propiolic acid. Quantum mechanical calculations have already been performed on hydroxyacetylene and its other isomers [69]. Calculated information would provide data to help experimentalists in their search and design of suitable experiments for studying these molecules.

Scope of the Present Work The work performed for the present study is divided into two parts: 1) Experimental spectroscopy using the matrix isolation technique; and 2) Computational studies using the ab initio MO program - GAUSSIAN 82, and the Semi-Empirical Quantum Mechanical Program- MINDO / 3. The objectives here are two-fold: First, the experimental studies would provide the pool of frequencies needed to determine an experimental force field based on experimental frequencies. Secondly, the computational studies will provide estimates for the force field, vibrational frequencies, and 20 information concerning the molecular reactivities of propiolic acid. Quantum mechanical force constants will be used as an initial estimate for determining a more plausible force field based on observed molecular parameters. The relationships between experimentally determined frequencies and the theoretically determined values will be explored so that a more accurate assignment of the fundamental vibrations will be achieved. In the first part of this research, the matrix isolation technique has been employed in obtaining detailed infrared spectra of propiolic acid and its isotopomers in argon, carbon monoxide and nitrogen matrices. The deuteriopropiolic acid used in the study was made in this laboratory. Isotopic substitution spectra using labelled compounds were recorded for propiolic acid and deuteriopropiolic acid isolated in argon and nitrogen matrices. The analysis of monomer spectra is reconsidered and experiments for various M / S (matrix-to-sample) ratios are performed in order to identify bands associated with dimers and other polymeric species of propiolic acid in argon and nitrogen matrices. The fundamental frequencies obtained for the monomer will be used to determine the experimental force field. In the second part, computational studies using the GAUSSIAN 82 program [70] have been performed. Optimized geometries along with corresponding total energies have been calculated for propiolic acid using various basis sets. Vibrational frequencies and force constants are also determined for propiolic acid monomer. One-electron-molecular properties are calculated for the two rotamers. The inversion barrier between the two rotamers is determined. Transition states and intermediate stnjctures for various probable decomposition channels are determined. These results have all been combined to obtain a clearer picture of the reaction dynamics. Normal coordinate analysis for both the ground singlet and lowest triplet states were carried out. A single calculation using the Force method in MINDO / 3 has been used to provide comparable information for geometries, dipole moments, total energy, vibrational frequencies, force constants and rotational spectroscopic constants. For comparison, the force constants and fundamental frequencies along with the geometries of the propiolate ion have been calculated in the 21 ground electronic state. This calculation was prompted in part by results of lETS experiments on carboxylic acids and by anotherstudy for the formate ion. The focus of these calculations is to obtain data needed to make comparisons between various systems and propiolic acid in an attempt to better understand the various aspects of propiolic acid spectroscopy and reactivity that have been object of the present study. Overall, calculated parameters are then compared with experimental data for related molecules where it exists. The use of HF-SCF single determinant MO methods and observed molecular parameters have been used to test for anharmonie effects. CHAPTER II EXPERIMENTAL TECHNIQUES

Matrix Isolation Technique The technique of matrix isolation is well known and has been described by several experimentalists [71-73]. A more comprehensive discussion on the technique and other relevant aspects have been described by Cradoek and Hineheliffe [74]. The field of matrix-isolation spectroscopy has been reviewed by several authors [75]. The general features of a matrix isolation experiment are shown in Figure 2.1. This technique involves the rapid cooling of a mixture of an absorbing species (sample) and a diluent gas (matrix). Such cooling results in the formation of a solid matrix at low temperatures (10 - 20K). Under these conditions, the sample can then be examined using any intended spectroscopic technique. In the present work, infrared spectroscopy (conventional and Fourier transform) has been used. An advantage of this technique is that under the conditions described above, the rotational and translational motions of the sample are frozen leaving only the vibrational motions for observation. In addition, the use of fairly large matrix-to-sample ratios removes effects caused by interactions due to the presence of adjacent sample molecules. Also, the rigidity of the matrix prevents diffusion of reactive molecules which could otherwise lead to reactions with other sample molecules or the matrix. The matrices usually used are either the noble gases (Ar, Ne, Kr, and Xe) or the molecular gases (CO, CO2. O2. and N2). In the present work, argon, carbon monoxide, nitrogen and neon have been used. The sample is propiolic acid and its isotopomers. With low M / S (matrix to sample) ratios, sample-sample interactions become important. In some studies, this is desirable as this method can be used to investigate dimers and formation of other polymeric species. Such sample-sample interactions will manifest themselves in the infrared spectrum in the form of changes in position, shape and intensity of sample absorption frequencies. In principle, all of these effects are capable of yielding information concerning both the intramolecular and intermolecular

22 23

U5

c (D E CD Q. X LU c g H o

CO "o E

CD b

"ca E a> x: o CO

C\J

Li. 24 forces present in the system. The freezing of rotational and translational degrees of freedom (leaving 3N-5 or 3N-6 ) in matrix-isolation often yields spectra with sharp, distinct, and often times well resolved bands. The advantage is that it enhances sensitivity and resolution of closely spaced bands including small spectral shifts often present as a result of heavy atom effects. The assignment of monomer absorption frequencies have been made by observing the relative intensities of absorption bands, the splitting patterns of isotopically labelled species, and ab initio calculated frequencies. A more detailed description of the techniques used is discussed in the next chapter.

Matrix Materials and Properties Research grade matrix gases were obtained from the following companies: N2 - Liquid Carbide Corporation, Chicago, Illinois, purity = 99.9995%. Ar - Scientific Gas Products, Purity = 99.9995%. CO - Linde Co. and Matheson Gas Products. All matrix gases had a stated purity of at least 99.9%. Tables 2.1 to 2.4 contain the relevant information with regards to spectral changes due to matrix effects. It must be noted that the matrix isolation technique involves a combination of several distinct technologies, each of which interacts with the others. The most basic factor is the low temperature needed to give rigid matrices. This requires cryogenic technology, which in turn requires the use of high vacuum techniques without which the low temperatures cannot be maintained. These are described in the instrumentation section.

Chemicals Propiolic Acid, H-C=C-COOH. Propiolic acid (PA) was bought from Aldrich Chemical Company, Milwaukee, Wisconsin. It is a liquid (b.p. 102 / 200 mm.) at room temperature. Before being used, it was purified through a series of 25 Table 2.1: Selected Matrix Properties [74]

Matrix Ar CO ^2

Properties

Diameters (A°) 3.75 3.0*^ 0.3Tm (°K) 25 20 19 0.5Tm (°K) 42 34 32 T(P=10"^orr)-°K 33 33 29 T(P=10"^orr)-°K 39 38 34 Cooling power to 20K/mW 78 -81 77 Cooling power to 0.3Tm/mW 78 81 77 First resonance transition 104 - - Internal vibrations (cm"'') - 2140(R,IR) 2330 (R) Lattice modes - 86(IR),50(IR) 83(R),69(IR). 49(IR) 47(R) 37(R),33(R)

a: Nitrogen molecular diameter, b: Nitrogen atom diameter. 26 Table 2.2 : Electrical Properties of Matrix Materials [74]

Ar CO N'

Dipole moment: p.x10^^ (esu.cm) 0 0.112 0 Quadrupole moment: 8x10^^ (esu.cm^) 0 -2.5 -1.4 Polarisability: a(10 A°3) 16.3 19.5 17.6

Table 2.3: Thermal Properties of Matrix Materials [74]

Td (°K) 35 35 30 m.p. (°K) 83.3 68.1 63.2 b.p. (°K) 87.3 81.6 77.4

Table 2.4: Site Diameters of Matrix Materials [74]

Ar CO N2

Diameter (nm) 37.55 39.99 39.91 Oh 15.6 Td 8.5 Approx. site dim. 46.1x34.8x34 45.2x34.2x34.2 Crystal stmcture fee fee fee Reference. [76] [77] 178] 27 fractional distillations followed by a number of trap-to-trap distillations. It was noticed that propiolic acid is unstable as its color changed from transparent to deep brown through yellow when the sample was exposed to light. These color changes were still noticed even during storage in a refrigerator or in a resealable container in the dessieator.

Water- ''^O, H2''^0 Lots of 1 gm samples were purchased from Prochem Company with 99.0% oxygen -18 enrichment. Samples were used without further purification.

Deuterium Oxide, D2O Samples were purchased from a variety of sources: Aldrich Chemical Company with 99.8% D-enriehment; KOR ISOTOPES with 99.75 atom %D; and Stohler Isotope Chemicals with 99.8%D. Samples used were subjected to a series of freeze-thaw cycles and evacuations before use.

Sodium Deuteroxide (NaOD/D20 solution) Samples were purchased in 10 g lots from Aldrich Chemical Company, Milwaukee, Wisconsin as a 40 wt. % solution of NaOD in D2O (99+ %D). It is a liquid at room temperature. It was used in the preparation of deuteriopropiolic acid (PA-D2).

Deuterated ''^O-water (D2''^0) Sample was purchased from Merck Company Incorporated of Canada. Except subjecting samples through a series of freeze-thaw cycles, sample was used without further purification. 28 Instrumentation Spectrophotometers Spectral data was obtained with a Beekman IR-9 spectrophotometer and a Nicolet Model 7199 FT-IR spectrometer. The Beekman IR-9 is equipped with a potassium bromide prism and a pair of gratings. Under maximum scale expansion, the spectrophotometer records 5 cm"'' per inch in the range 400-2000 cm"'' and 10 cm"'' per inch in the range 2000-4000 cm"''. Spectra were usually obtained in the range 400-4000 cm"''. Without rigorous calibration, the IR-9 is accurate to within ±1 cm"'' in the 400-2000 cm"'' range and to within ±2 cm"'' in the 2000-4000 cm""* range. The instrument was calibrated periodically using atmospheric absorptions for water and carbon dioxide and the compilations prepared by Plyler et al [79]. The Beckmann IR-9 is a double beam instmment. The Nicolet Model 7199 FT-IR was used courtesy of the Instrumentation Division of the Chemistry Department, University of Utah at Salt Lake City. This is a single beam instrument. Rather than use a monochromator, the FT-IR uses a Michelson interferometer, a device that preserves both frequency and intensity information under conditions of the experiment. The NIC 7199 is accurate to within ± I cm"'' in the range 400-4000 cm"''. The theory and practice including applications of FT-IR have been discussed more thoroughly by Bell [80], Griffiths [81], and in the series edited by Ferraro and Basile [82].

Cryogenics The basic component is the refrigerator or refrigerant required to produce the very low temperatures needed to form a solid matrix in which the sample is stabilized. Such temperatures were achieved by using an Air Products Displex Model 202 closed-cycle refrigerator equipped with a Csl (cesium iodide) cold window enclosed in a vacuum shroud. A high vacuum (10'^ torr or better) was maintained around the cold surfaces at all times. Such pressures were achieved by a combination of diffusion and mechanical pumps and the pressures were measured using a discharge gauge. 29 Sample Preparation Argon and nitrogen were used as the principal matrices in this investigation. Some spectra were obtained with carbon monoxide and neon matrices. Once the desired sample or labeled sample had been purified, mixtures with matrix to sample ratios between 500/1 and 4000/1 were prepared. The sample species involved were all liquids at room temperature.

Propiolic Acid A 2 liter sample bulb and a smaller calibrated vessel (221.0 cc) were evacuated to below one micron (1^) pressure at the manifold of an all-glass vacuum line. The evacuation is achieved by a two-stage pumping sytem with a liquid nitrogen-cooled trap which gave ultimate pressures of about 10"^ mm Hg. The arrangement used consisted of a mechanical (rotary) pump, which provided an ultimate pressure of about 10"^ mm Hg backing an oil-diffusion pump which reduced the ultimate pressure to 10"^ mm Hg and increased the speed of pumping near the lower pressure limit of the mechanical pump. The manifold and the calibrated and sample containing bulbs were flushed several times with propiolic acid vapor. After the last flushing, the 2 liter bulb was closed and the calibrated vessel filled with the desired amount of acid vapor monitored by an octoil manometer. The calibrated vessel was then closed and the rest of the manifold evacuated. The vapor in the calibrated vessel was then transferred to the 2-1 bulb by trapping with liquid nitrogen. The acid was usually transferred over a period of half an hour to one hour so as to ensure complete sample transfer. At the end of the transfer, the sample bulb was closed and then left to warm up to room temperature. The desired volume of matrix material was then added with the pressure being monitored by a mercury manometer (Figure 2.2). Although the octoil manometer fluid was not changed frequently, care was taken to make sure that it was completely degassed every time a new sample was to be prepared. This was achieved by applying the heat gun whenever the manifold was being evacuated. 30

Pressure Vgauge

Propiolic Acid container Calibrated LN2trap ^^^ filter vessel bulb Oil manometer

diffusion pump

Figure 2.2. Sample Preparation Setup. 31 Propiolic Acid -Di Only the hydroxyl hydrogen of the carboxylic acid function is exchanged. The deuteration was effected through vapor-phase H / D exchange. The vapor phase exchange was carried out on the vacuum line. As in the preparation of propiolic acid samples, a measured amount of propiolic acid was transferred to a 2-1 bulb that had been flushed several times with D2O vapor. The manifold was then evacuated and then flushed several times with D2O vapor. The manifold was then evacuated slightly and the calculated amount of D2O added. The bulb was then closed and the entire line evacuated. To encourage exchange, the bulb was placed over an oven at about 40° C overnight. Matrix gas was then added as desired. M/S ratios prepared varied from 450 /1 to 4000 /1. The vapor phase exchange method applied above had the advantage in that excess water could be pumped off easily. Band intensities of residual water contamination were usually weak and were easily identified from known water spectra. Preparations using liquid phase exchange were not very successful although spectra indicated that H / D exchange had occurred. The presence of excess D2O was a problem. In the vapor phase exchange method D2O was used in slight excess and the extent of deuteration could be easily ascertained by comparing band intensities for the VQH ^^^ VQD absorptions.

Preparation of ''^0-Congeners These samples were prepared using micromolar liquid samples prepared in turn by condensing H2^®0 and acid vapors into the sealed stem of a 4mm stopcock attached to a 1 or 2-liter sample bulb as shown in Figure 2.3. The vapors were separately measured into the 2-liter bulb using a manometer (generally 1.0 to 1.5 torrs) and then trapped through the stopcock into the small reaction vessel. Catalytic amounts of hydrogen chloride gas or vapors from concentrated hydrochloric acid were then added and samples were left to exchange overnight. The matrix isolation samples were then prepared by vaporizing the liquid sample from the reaction vessel into the 2-liter bulb where matrix gas was added. This 32

2 liter bulb

reaction vessel

Figure 2.3. ''^O- Isotope Exchange Reaction Vessel. 33 method has the advantage that it is economical in terms of H2''^0 consumption. The '' ^O /'' ^O ratio was usually chosen to be unity. The procedure described above gave four congeners of propiolic acid, namely: HC=CCOOH, HC=CC*OOH, HC=CCO*OH, and HC=CC*0*OH. The extent of '°0-enrichment of the samples was deduced from the spectra in combination with known splitting patterns [73]. In order to clarify band assignments, various sample concentrations were prepared. M/S ratios of 1000 /1, 2000 /1 and 3000 /1 were prepared and spectra recorded.

Synthesis and Preparation of Deuteriopropiolic Acid (PA-D2) A 10% solution (2.3 ml of PA in 23 mis of D2O) was neutralized with 31-32 ml of 40% NaOD/D20 to basic pH (pH > 12) in an ice bath. The solution was left to sit with stirring in the bath for another two hours. The resulting solution was neutralized while cold with 6.5 - 8 ml of concentrated HCI. The PA-D2 was then extracted with 100 ml of ether. The ethereal solution was dried with anhydrous calcium sulphate, filtered and the ether removed under reduced pressure. The PA-D2 was then purified by a series of trap-to-trap distillations. The C-D enrichment was greater than 97% while there was 100% H / D conversion for the hydroxyl. The preparation of samples for spectra was the same as for PA, deuterated or PA-OD, and ''^0-congeners. The D2^®0-congeners were prepared using the same procedure as for the H2^^0-congeners.

Sample Deposition In matrix isolation experiments, sample deposition can be accomplished either using pulsed flow techniques or the conventional steady flow deposition method. For this work, the conventional slow spray (steady flow or thermal effusive) deposition technique was used. The set up for sample deposition is shown in Figure 2.4. The cryostat was evacuated to pressures better than 10"^ torr. The entire deposition set up was evacuated to pressures better than 10'^ torr. The overall vacuum in the deposition system was maintained by using a liquid nitrogen trapped oil 34

OL

CD CO C EE g "ion Q. CD O a, E i CO CO Q. E cvi •D (D C a. CO c o ^. o 't/i o D "*- 5= u. 35 diffusion pump. Such a vacuum was maintained throughout the course of an experiment. The system was checked for leaks and then the refrigerator turned on. It took about an hour and fifteen minutes for the refrigerator to cool the cold window from ambient temperature to 11K, the temperature at which spectra were recorded. Once this temperature had been achieved the refrigerator, which at this point has its cold window parallel to the deposition port, was rotated through 90 degrees to bring the cold window perpendicular to the deposition port. The entire system was then flushed a couple of times with the gaseous matrix / sample mixture. The deposition rate is controlled by manipulating a teflon needle valve. The progress of deposition is monitored by observing changes in pressure of the gaseous mixture using a mercury manometer. Depending on the M / S ratio, 30 to 160 torrs of the gaseous mixture were deposited over a period of one to two hours. A temperature controller - indicator (Air Products Model APQ-G) provided ± 0.1 K temperature stability of the sample between 11K and 300K. The vacuum shroud used is a cylindrical stainless steel vessel of 70 mm i.d. and 170 mm length, with a pair of Csl windows through which IR spectra of a sample on the cold window can be measured. At the end of deposition, the vacuum shroud was rotated 90 degrees so that the cold window became perpendicular to the IR beam axis. Spectra were then recorded using an IR-9 spectrophotometer for group I experiments and a NIC Model 7199 FT-IR spectrophotometer for group II experiments. All PA and PA-OD, spectra in Ar and N2 are group I experiments. All PA-D2 spectra belong to group II. PA / CO spectra also belong to group II with PA+H2^®0, PA+D2^®0 in Ar and N2. All spectra were recorded in the range 400-4000 cm"''. Annealed sample spectra were also recorded in some experiments with temperatures being varied up to 35K. Although spectra were recorded for the entire range, in most cases spectra recorded for regions with suspected thermal dependence (e.g., OH stretch, C=0 stretch and other carboxylic acid group modes). In an experiment where a deuterated sample was to be studied, the deposition line was saturated with D2O vapor several times prior to use. CHAPTER III ASSIGNMENT TECHNIQUES

The spectra obtained had to be assigned and this constitutes a fundamental problem associated with infrared spectroscopy. The assignment of individual peaks to a particular species has only been accomplished in the smaller molecules where may be-other spectroscopic methods provide additional useful information. In other cases bands must be assigned empirically using the methods or techniques described below. The most general technique for assigning a band to a particular vibrational transition is to compare its frequency with known "characteristic frequencies." This is also referred to as the "group frequency method." It involves the use of certain general correlations between groups present in a molecule and the infrared bands observed in its spectrum. Fundamental to this method is the assumption that certain functional groups give rise to more or less the same frequency regardless of the parent compound. For example, one may expect to find C-H stretching vibrations in the region 3300-2700 cm"^ or 0-H stretching vibrations in the region 3600-3000 cm"''. Unfortunately, these correlations are less reliable for deformation vibrations largely because of coupling between vibrational modes which might turn to weaken or strengthen bonds. Another disadvantage of this method is that it does not allow for the distinction between different combinations of products containing similar groups in a matrix study where more than one species may be present. Characteristic frequencies are most helpful in regions where there are very few bands. In crowded regions, their utility is greatly diminished as bands turn to overiap each other. Related to the use of characteristic frequencies is the technique of using information from related molecules. There is less ambiguity with these comparisons because their spectra tend to be similar. In this \Nork, comparisons of spectra of PA, PA-Di, and PA-D2 and isotopomers were very helpful. Beneficial comparisons were made with spectra of other carboxylic acids and acetylenic compounds. 36 37 The isotopic substitution method is a most powerful method for assigning vibrational bands. These different isotopically substituted isotopomers will give rise to species that have different masses and hence different vibrational frequencies. Bands that do not shift in frequency when a given atom is isotopically substituted must arise from a mode that does not include the motion of that atom. The usefulness of this method is limited by the fact that in large molecules some vibration frequencies may be affected only slightly by a change of mass of a particular atom. In matrix isolation spectra where bands are sharp, it is possible to measure shifts of the order of 1 cm"'' or better. The magnitudes of the frequency shifts of the various bands show which atoms are principally involved in each vibration. Isotope effects are based on Hooke's Law model. In a diatomic molecule the frequency of vibration is given by

co=(1/27i)(Vk/^i) (3.1)

where k = force constant and \i is the reduced mass of the atoms involved [83]. If it is assumed that no change in force constant occurs with substitution, the ratios of the frequencies of the two isotopic molecules is given by

CO/co ' = V^/^'. (3.2)

The generalized expression for polyatomic molecules is given in the Teller-Redlich product rule [83]:

nco'k / C0k=n(mi / m'i)1/2(M' / M)3/2(r^ryr^ / Ixlylz)^/^ . (3.3)

This can be applied generally to either estimate an isotopic shift or to help confirm the assignments of co^ and co'^. In the expression above, the primes refer to the isotopic species. The m's, M's, and I's refer to atomic masses, molecular masses and principal moments of inertia respectively. In this work, shifts produced by H / D exchange for acetylenic and hydroxyl hydrogens and shifts resulting from ''SQ-labeling were studied in detail. To 38 a lesser degree shifts arising from the formation of dimers and other polymeric species were noted. The OH stretch shows large frequency shifts as a result of dimerization. Other modes are less affected and sustain much smaller shifts, for example H-C stretch and C^C stretch. Isotopic splitting patterns provide additional information. It is usually nice to have equal proportions of the two isotopes. If mixing occurs, species with one atom of the element concerned give 1:1 doublets, species with two atoms in equivalent positions give 1:2:1 triplets and species with three equivalent atoms 1:3:3:1 quartets. This bionomial distribution is observed when isotopes are randomly distributed (Figure 3.1) [73]. Sometimes the weakest component may not be observed, especially when other bands are present in the same region of the spectrum. Where vapor phase spectra are available, the band shapes can be useful tools for the interpretation of spectra. Ueda and Shimanouchi [84] calculated the shapes of band envelopes for several types of asymmetric rotors. To use this information, one would need to know the rotational constants and principal moments of inertia of the molecule. In cases where these are not known, they can be calculated by applying the formulas of classical mechanics [85]. Also the calculation of P- and R- branch separations may prove useful and empirical formulas have been provided for asymmetric tops [86]. Molecular vibrations, in addition to infrared absorptions, also give rise to Raman scattering. The activity of a vibrational mode can be determined by applying Group Theory. The depolarisation ratio is applied to help determine the symmetry of a particular mode. In practice , it is often seen that a vibration which is weak in the infrared will be very strong in the Raman [87]. Some vibrations of propiolic acid fall into this category. Raman data necessary for a complete spectral data analysis was not obtained in this study, but reference is made to existing data [38]. Perhaps the most sophisticated method for assigning observed infrared bands is the normal coordinate analysis. It involves computations of theoretical spectra from assumed structures and force constants which are then brought into coincidence with the observed bands by adjusting the assumed force constants. Ab initio force constants and force constants 39

X-O stretch Xl6o X'ISQ XO2 (symmetric stretch) 16-18

16-16 18-18

1:1 doublet as in CO 1:2:1 triplet as in CO2

16-16-16 16-16-18 16-18-18 18-18-18

XO3 (symmetric stretch) 1:3:3:1 quartet as in peracids.

Figure 3.1. Effects of 50% ^^O- Enrichment in Oxygen Containing Molecules with Equivalent Oxygen Atoms [74]. 40 transferred from related molecules will be applied in the future. While this method can be used to achieve satisfactory results, a drawback is that it assumes purely harmonic motion. In some cases, anharmonicity corrections can be made. A satisfactory application is the explanation of "anomalous" isotope effects. The methods discussed above were applied at varying extents to assign the fundamental vibrational frequencies for propiolic acid and its isotopomers. While, some vibrational frequencies were assigned without difficulties, others have received ambiguous assignments. These are discussed with relevant spectra wherever such uncertainty exists. Some of the problems encountered result from insufficient exchange for the ^^0-labeled species, and from interference by dimer formation resulting from concentrated samples. CHAPTER IV THEORETICAL BACKGROUND TO MOLECULAR MECHANICS

Introduction The determination of the forces that hold atoms together in a molecule (intramolecular) or molecules in a molecular system (intermolecular) is one of the key aspects of vibrational spectroscopy. Such information can be derived from data obtained in infrared, Raman or a combination of both spectroscopies. The sum of these forces for any molecule is generally referred to as the molecular force field and its evaluation is referred to as molecular mechanics. In mathematical terms, it is described by the potential energy function, V:

3N 3N 3N V(q) = Vg +1 (av/aqi)eqj +1/2 1 I (a2v/aqiaqj)eqiqj + higher terms. (4.1) i=1 1=1 j=1

In this expression, qj represents some type of general displacement coordinate related to the molecular prameters. In order to determine V(q), force constants fy must be evaluated. These evaluations have been carried out to satisfaction only for the smaller systems - diatomics, triatomics, and some four-atom molecules such as acetylene and ammonia. The problem becomes more difficult for systems with four atoms or more as the number of independent force constants to be evaluated becomes very large. For propiolic acid, there are Pj(Pj+1 )/2 independent force constants to be evaluated for each symmetry class of dimension pj. Thus for propiolic acid, there are 76 independent force constants to be evaluated. E. Bright Wilson [82] developed a method which applies the techniques of group theory in solving the molecular secular equation:

41 42 |FG - EXI = 0 (4 2) where F represents a symmetrical matrix whose elements are linear combinations of elementary force constants derived from the potential energy expression (to be developed later). G represents a symmetrical matrix whose elements are also linear combinations of geometrical parameters (bond lengths and bond angles). E is a unit matrix and X represents a diagonal matrix with elements X\ = 47c2c2coj2. jhe coj's are either observed vibrational frequencies from experimental spectroscopy or calculated frequencies from semi-empirical and ab initio methods. In the present work, attempts have been made to accurately assign the observed frequencies that will be used to determine force constants obtained by combining this data with force constants calculated using ab initio methods. One of the problems with this determination is the necessity to correct the observed frequencies for anharmonicity. Such corrections may be appreciable and can amount to as much as 180 cm"^ for frequencies involving the vibrations of hydrogen atoms [88]. Such a possibility exists here for modes involving the acetylenic and hydroxyl hydrogens. It should be noted that the calculated force fields are based on the harmonic approximation derived from the Hooke's Law model for molecular systems. Introduction to the theory and applications of molecular vibrations can be found in a number of books which adequately cover the field [89-93]. More exhaustive texts on molecular vibrations and the calculations of the forces which govern them are to be found in the original treatise by Wilson, Decius, and Cross [83] and in specialist chapters by Mills [94], Shimanouchi [95], Duncan [96], and a monograph by Burkert and Alllnger [97]. The central problem of any vibrational analysis is the determination of a force field from the vibrational frequencies, and In this process the normal coordinates are obtained automatically. Our approach employs both methods. As already mentioned, the main stumbling block to a direct determination of the molecular force field from the frequencies in vibrational analysis is the large number of force constants. While there are Pj normal coordinates, the corresponding number of independent force constants is 43 Pi(Pi+1 )/2. Therefore, a complete determination of all the force constants requires the analysis of the spectra of isotopically substituted molecules. The underlying phenomenon is the fact that more frequencies are obtained without introducing any additional force constants. A literature survey would reveal that there are many problems associated with molecular force field determinations. To alleviate some of these problems, spectroscopists have introduced a number of systematic simplifications of the complete force field. In these assumptions, conditions imposed lead to the fact that as many off-diagonal terms as possible are set equal to zero. Each set of assumptions leads to a different type of force field. In the present study, the simple force field which best fits ordinary chemical ideas about the nature of interatomic forces has been chosen. This is the General Valence Force Field (GVFF). It is based on internal displacement coordinates: bond distances, bond angles and torsion angles. A very rough approximation is to neglect all off-diagonal force constants. The result is a force field with Hooke's law harmonic potentials (Equation 4.3).

V = 1/2 E frj(ri - rQj) 2+ 1/2 Z fe,k(qk" qo,k)2 + 1/22 fp|( x, - TQ i)^ • (4.3) i k I

The true force field, however, will include interaction force constants. The magnitude of these constants will depend on the extent of coupling between the normal modes. In order to evaluate the intramolecular force field chosen above, the following procedure,which is generally used, has been employed [98]. This includes : 1) Choice and construction of appropriate coordinates; 2) Constmction of the F matrix; 3) Constmction of the G matrix; 4) Determination of the force constants; and 5) Determination of the valence force field potential function. The highlights of the different steps involved will be discussed here in the context of propiolic acid. 44 Fundamentals of the Molecular Mechanics of Propiolic Add Choice and Construction of Appropriate Coordinates The interatomic potential for a known geometry is written based on Cartesian coordinates as (Equation 4.4):

3N 3N 2v = 2Vo + 2 I (av / axj)oXj + I (a^v / axjaxj)oXjXj + ... (4.4) i=1 i,j=1

The VQ term is removed by a suitable shifting of the origin. The linear term becomes zero since for independent coordinates at equilibrium (av / axj) = 0. The Taylor series expansion of the potential function, as written above, is always truncated at the second order term. To make the interpretation of force constants more meaningful internal coordinates are used [83]. These are bond lengths, bond angles, and torsional angles. The problem is further simplified by transforming these into internal symmetry coordinates. The corresponding vibration potential function becomes:

2v = 2 s (av / aRi)o Rj +1 (a^v / aRjaRj)o RJRJ. (4.5) i "J

In matrix notation, the above equation becomes :

2V = R pR R (4.6)

where F^J: = (a^V / aRiaR:)o . The advantage here is that pf^y offer a more direct physical interpretation between force constants and other molecular properties. Symmetry coordinates S are then finally defined by another linear transformation. One notes that the number of internal symmetry coordinates gives the number of vibrational modes. The different coordinate systems for propiolic acid are given in Tables 4.1, 4.2 and 4.3 based on the model structure obtained from microwave data. 45 Table 4.1: Cartesian Coordinates (3N) of Propiolic Acid

Center Atom Atomic Cartesi an Coordinates Number Name Weight X Y Z

1 H 1.0080 1.4628 -1.1588 0.0

2 0 16.0000 0.4948 -1.2485 0.0

3 C 12.0000 0.0000 0.0000 0.0

4 C 12.0000 -1.4450 0.0000 0.0

5 C 12.0000 -2.6540 0.0000 0.0

6 H 1.0080 -3.7090 0.0000 0.0

7 0 16.0000 0.6745 0.9949 0.0 46 Table 4.2: Internal Coordinates (R)

Coordinate Number Definition

R1 D(O-H) R2 D(H-C) R3 D(C^) R4 D(C=0) R5 D(C-O) R6 D(C-C) R7 0(H-CsC) R8 0(C=C-C) R9 0(C-C=O) RIO 0(O=C-O) R11 0(C-C-O) R12 0(C-O-H)

R13 n(H-c^) R14 n(C^-C) R15 T(C-O-H) R16 n(C-c=0)

*: D refers to bond length, 0 refers to bond angle, and n and x refer to torsional angles. 47 Table 4.3: Internal Symmetry Coordinates for Propiolic Acid

Symmetry Coordinates

A'- Block Si ARi S2 AR2 S3 AR3 S4 AR4 S5 AR5 Se A06 S7 ARg Ss A01 S9 A04 = (1/V2(A03-A05) -A05) S10 1/V6(2A04-A03- S11 A02

A"- Block Si 2 AHi Si3 An3 Si4 Ax A02 Si 5

The cyclic relationship between the angles of the carboxyl group results in one reduntant coordinate. This coordinate is defined by: S = 1 / V3 (A03 +A04 + A05). 48 In Table 4.3 above, the number of internal coordinates is larger than the number of vibrational degrees of freedom. There is a redundant coordinate. While the handling of the redundancy has been the subject of controversy in the literature [97-99], it seems to be well established that the redundancy relation among Wilson's coordinates is linear. The use of symmetry coordinates allows maximum factorization of the G and F matrices thereby making the calculations easier.

Construction of the F matrix Since a choice of appropriate coordinates was made in 1) (this being internal symmetry coordinates), the potential can now be rewritten in the form:

2V = 2S(av / aSi)oSi+1 (a^v / aSiaSj)oSjSj. (4.9) i i.i

The expression reduces to

2V = IFijSjSj (4.10) ij where the Fj.'s are the elements of the F matrix consisting of linear combinations of the fy's. An equivalent way of expressing the result above (4.10), in matrix notation is

2V = S'FS, (4.11)

where S is a column matrix and S' a row matrix, in terms of the internal coordinates; S' is the transpose of S. It is important that V be expressed in terms of S for maximum symmetry factoring and in terms of fij's for chemical and physical interpretation. The relationship between the fy's and F is

F = UFU' (4-13) 49 where U and its transpose U' are used to symmetry-factor the f matrix with elements fy. The f matrix is generated by constructing a multiplication table of the internal coordinates (Equation 4.14).

ARi. AR2. AR3 ARn ARI AR2 AR3 (4.14)

ARn

The U matrix is simply the coefficients of Equation 4.14, and its transpose is the inverse U"^, since U describes a linear orthogonal transformation. The F matrix can, therefore, be factorised appropriately.

Constmction of the G matrix Its construction is similar to that of the F matrix since

G = U'gU' (4.15)

where g is the matrix that contains elements in terms of the atomic masses and the geometrical parameters. In calculating the G matrix, the procedure used in [83] is followed: Calculate e vectors; calculate S vectors and then using the expression

N gtt' = ^ ^^aSta-Sfa (^•^^' 00=1 50 where the fs refer to the intemal coordinates, ^i is the inverse mass and a refers to the atoms and the S's refer to the displacement vectors for each atom. The dot represents the scalar product of the two vectors. The elements of the G matrix can be obtained directly from the expression for the kinetic energy

2T = S'G-1S (4.17) where G"'' is the inverse G matrix, S is the time derivative of the internal symmetry coordinate matrix, and S' of the transpose. The generation of the G elements in Equation (4.17) for general vibration motions have been summarized in Appendix VI of Wilson, Decius, and Cross [83].

Determination of the Force Constants Having obtained the force-constant matrix, F and the G matrix elements (from atomic masses and molecular geometry in terms of the basis coodinates), they are then substituted into the secular determinant,

|GF->.E| = 0. (4.18)

In the present study vibration frequencies to be used are embedded in X (characteristic roots related to the corresponding normal vibration wave numbers by ^ = ^K^C^CO^). Examples of the actual breakdown of the solution of the secular determinant can be found in various spectroscopy texts and in the references cited in this section. While Equation (4.18) represents the fundamental idea of the calculation, adjustments can be made to account for the various computer programs that can be used to solve the equation. A modification of the Wilson technique is used in this study. In this technique the eigenvalue equation to be solved is of the type:

GFL=LA (4.19) 51 where G and F have their usual meanings, L is the matrix of the eigen - vectors which provide the relative amplitudes of oscillation, and L is the diagonal matrix of the eigenvalues which provide the normal or fundamental frequencies. In order to obtain satisfactory solutions, the "Jacoby-diagonalisation" method was used. Since the method handles only symmetrical matrices and the GF matrix of Equation (4.19) is not symmetrical, manipulations are required. The various ways in which such manipulations have been accomplished can be found in [95,96]. In the present work, efforts have been made to obtain experimental frequencies from which force constants will be calculated. The number of independent force constants, unfortunately, is larger than the number of experimental frequencies from a single molecule. To solve this problem isotopic frequencies were used to provide additional experimental data. As usual, least squares procedures were introduced [99-101]. The fundamentals of least squares calculations are the following. It was shown [102] that the relation between the characteristic frequencies X\ and the force constants Fy can be expanded in a Taylor series:

AXj = KaXj / aFjk)AFjk +1 /2 xi(a2?ij / aFj,,aF|m)aFjkaF,r^+... (4.20)

Only the first term of Equation (4.20) is retained. From a known G matrix and an initial F° matrix the eigenvalue equation is solved.

GF°L° = L°A° (4.21)

A° is the diagonal matrix of the calculated frequencies resulting from the starting guess of F°. The starting guess F° was obtained from ab initio calculations to be discussed later. L° is the matrix of corresponding eigenvectors. While the methods are out of the scope of this study, it has been shown elsewhere using Equation (4.20) that

AXj = I[(L°ji)2AFjj] + 2I[L°jiL°kiAFjk] (4.22)

which in matrix notation is written as 52 M, = JAF (4 23) where /SX is a column matrix of A\, AF is a column matrix of the elements AFy and is a rectangular matrix containing the products of L\. elements, calculated in Equation (4.21). One then moves on to least squares calculations in trying to compute from Equation (4.23) the AF corrections to F° such that the errors AXj = Xf^^^ - X^j^a' are minimised. From Equation (4.23) one has

JPAX=(JPJ)AF (4.24)

AF =(JPJ)-''JPA2. (4.25)

and one calculates the corrections to be applied to F° such that the sum of weighted squares of the residuals is minimised. A new force constant matrix E = F + AF is then generated and this iterative procedure is followed several times until AF becomes very small. P is a diagonal matrix comprising the statistical weights which may be given to the observed quantities ^j^*^^. The least squares calculation outlined above can be extended to accomodate data from isotopic molecules. The transferability of force constants from one isotopic molecule to another is in this case assumed to be exact. This procedure permits us to now calculate force constants using the additional frequencies from isotopic molecules. CHAPTER V VIBRATIONAL ASSIGNMENTS: INTERPRETATIONS AND DISCUSSION

Introduction The previous chapter dealt with the theoretical foundations of a vibrational analysis. The work of the present chapter deals with interpreting the spectra that have been recorded so as to obtain the self consistent pool of satisfactorily assigned fundamentals to be used in the determination of the molecular force field of propiolic acid. Spectra of propiolic acid and propiolic acid isotopomers isolated in solid argon, carbon monoxide, neon and nitrogen matrices were recorded in the range 400 - 4000 cm "^. In agreement with microwave studies of the related ester, studies on propiolic acid indicate that the cis conformer is the most abundant conformer. The interpretation of spectra is based on a Cg symmetry for propiolic acid. This is based on a planar HCCCOOH skeleton in which the hydroxyl hydrogen is cis relative to the carbonyl. One of the difficulties in assigning carboxylic acid spectra is the presence of bands due to hydrogen-bonded dimers and higher oligmers in the spectra of these compounds. To alleviate this problem, spectra were recorded for matrix-to-sample ratios varying from 4000 /1 (dilute) to 500 /1 (concentrated). Spectra for very dilute samples show relatively fewer absorptions than spectra for more concentrated samples. This concentration dependence can be used to differentiate between monomer and polymer absorptions. Unfortunately, in dilute sample spectra ( M / S > 3500), it is difficult to detect weak bands. The vibrational analysis of propiolic acid spectra was greatly aided by isotopic substitution: replacement of acetylenic and acid protons by deuterium and oxygen-exchange on the carboxyl group. Comparison of the spectra of the protonated acid and deuterated acids helped to ensure that peaks were assigned correctly. It was found that, although deuteration of the hydroxyl proton was nearly 100% efficient in the vapor phase due to 53 54 exchange with hydrogen in the sample handling equipment, the effective percentage of deuterated molecules was reduced to about 50% in matrix - isolation experiments. Emphasis has also been placed on assignments for related molecules, and on interpretations of ab initio calculated frequencies and normal modes for both cis and trans conformers. Related molecules referenced include the acetylenes, propargyl halides, propynal, methyl propiolate, and their various deuterated analogues; and the more well studied carboxylic acids (formic acid, acetic acid, trifluoracetic acid) and various other studies involving the carboxyl group. Assignments have been checked with results of a quantum mechanical normal coordinate analysis.

Vibrational Assignments and Interpretations Propiolic Acid Monomer Propiolic acid has the Cg point group which contains symmetry elements E (identity) and a^ (horizontal plane ).

^ E q^

A' Tx, Ty; R^ a XX a yy a ^z axy

A" T^; Rx. Ry otyz ^x

It is clear from the character table that no vibrational modes are exclusively Raman active. Propiolic acid has 15 fundamentals which can be classified as 11 A' + 4A", where A' fundamental vibrations are those symmetric with respect to the molecular plane and the A" modes are antisymmetric with respect to the plane. The A' fundamentals are actually parallel to the plane while the A" modes are perpendicular to the plane. The propiolic acid spectra obtained in this laboratory reflect the difference between matrix 55 isolation technique spectra, and solution and liquid phase spectra obtained previously by Katon and McDevitt [38]. All of the deuterio propiolic acid spectra were recorded with the NIC 7199 FT-IR. This data was used to provide confirmatory evidence for assignments of the principal isotopomers. The spectra of PA-D2 indicated that at least a 97% H / D exchange of the acetylenic hydrogen had been achieved. This observation is confirmed by intensity data for the acetylenic H-C stretching frequency. This conversion could be reversed easily by the presence of HCI vapor from concentrated hydrochloric acid. This was noted in ''°0-exchange spectra where concentrated hydrochloric acid was used as a source for hydrogen chloride vapor. In view of this fact, the assignment of propiolic acid - d2 spectra was found to be a more difficult task as four isotopic species were expected, namely, D-C=C-COOD, D-C=C-COOH, H-C=C-COOD, and H-C=C-COOH. The OH / OD conversion was 100% complete although the effectiveness was reduced when recording spectra by exchange with sample handling equipment. For a complete vibrational interpretation, the results of the present study and those of Katon and McDevitt [38] have been combined in Tables 5.1 and 5.2 for normal propiolic acid (PA) and deuterio propiolic acid (PA-D2) respectively. Tables 5.1 and 5.2 show that in previous studies, there were not enough fundamentals assigned to be able to determine the experimental force field for propiolic acid. In addition to spectra obtained for deuterio propiolic acid, isotopic substitution spectra were obtained for PA-D2 isolated in argon and nitrogen matrices. The differentiation between monomer and dimer bands is based on a comparison of spectral features (intensities, position, and number of bands) for M / S varying from 3500 - 4000 /1 to < 500 /1. The former is more likely to show mostly monomer absorptions while the latter will show both monomer and dimer absorptions. Calibrated spectra for PA, PA-OD, and PA-D2 at various matrix- to- sample ratios are listed in Appendix A. To discuss the assignment of the spectra, the 400 - 4000 cm"'' range has been divided into four spectral regions: high frequency (4000 - 2000 cm"'') spectral region, the carbonyl and mid-frequency regions (2000 -1200 cm"''), the 1200 - 680 cm"'' region and the 680 - 400 cm"''. The four spectral 56 Table 5.1: Fundamentals of Normal Propiolic Acid Compared

Present Wori< Reference 38

Ar CO N2 IR Sol. IR vap Raman Assignment type

3550 3438.8 3520 3588 3581 (A) VI 3568 3315.6 3302.5 3307 3306(m) 3329 3297(m,b) V2 3324(B) 2140 2139.6 2138 2130(m) 2137 2129(vs,P) ^3 2125(B) 1753.7 1746.2 1748.5 1698(s) 1676(s,b,P) V4 1428(vw) 1303 1342 1327.5 1400(m) 1373 1401(vw) V5 1362(A) 1354 1151 1178.8 1153 1263(s) 1310 1269(w,P) ^6 1300(A) 1290 817.5 826.3 820 859(w) 869(S,P) ^7 695.1 699.0 708 720(w) 690 723(m) ^8 683(B) 653.1 659.4 665 752(w-m ) 753(vw,sh) ^9 587.5 586.5 586.3 487(w) ^10 240(w-m ) 245(m) VII

755.2 762 758.5 693(w) 693(w) vi2 572.4 612.5 607 600(w) 603(w) VI3 529.5 541 533.8 900(w) 823 V14 812(A) 805 218(s) VI5

s=strong, m=medium, w=weak, v=very, b=broad, P=polarized, sh=shoulder, v=stretch,' 5 = in-plane bend, n =not plane bend (y ). 57 Table 5.2: Fundamental Frequencies of Deuteriopropiolic Acid (PA-D2)

Present \Nork Reference 38 Ar CO N2 IR(Soln) Raman(Liq)

2590.6(m ) 2588.0(m ) 2600.0(m) 2365 VI 2356 2215 2614.3(m) 2628.0(m) 2615.8(m) 2631 (w,b) 2610(s) V2 2600(m) 1977.5(m) 2089.0(m) 1974.2(m) 1987(s) 1980(s) V3 1715.1(s) 1738.0(s) 1738.0(s) 1700(s) 1682(m) V4 1276.4(m) 1275.0(m) 1276.4(m) 1327(s) V5 932.3(m) 933.0(m) 934.0(m) 1050(m) V6 850.8(mw)1 851.0(mw) 851.0(mw) 861 (w) 860(w) V7 647.7(m) 660.5 (ms) 661.4(ms) V8 612.5(m) 612.0(mw) 612.0(m) 600(w) V9 (483.6) (483.6) 483.6(m-w) vio 179.0 179.0 179.0^ VII

612.5(m) 612.0(m) 612.0(m) 575(w) VI2 618.6(mw) (618.6) (618.6) vi3 421.9 460.4(m) 637(m) vi4 217.0 217.0 217.0^ V15

a: from reference 19. 58 regions enumerated above account for thirteen of the fifteen fundamental modes. The two fundamentals unaccounted for occur in the region below 400 cm"''. These are the deformational modes of the carbon skeleton. In assigning the fundamental modes of propiolic acid, it is important to recognize that there are two possible rotational isomers (I) and (II). Previous interpretations of the spectra of propiolic acid used the cis-carbonyl, weakly intramolecular H-bonded structure (I). Their data could not distinguish between structures (I) and (II). Structure (II) is obtained by rotation around the C-0 bond. According to the microwave studies of Lister and Tyler [39], structure (I) is the most abundant conformer. Calculations of the stabilization energy of both conformers carried out in the present study support this conclusion. The cis conformer, structure (I) was found to be the most stable form for propiolic acid monomer. It is more stable by about 5 kcal / mol compared to the trans conformer, structure (II). Assuming a Boltzmann distribution, the abundance of the trans conformer, stmcture (II) at room temperature would be about 5% and it might just be possible to detect this conformer in matrix isolation experiments. .It would be assumed in assigning fundamental modes that the most intense absorption bands in the spectra are due to fundamentals of the cis conformer, stmcture (I), which is clearly present in the highest abundance. Bands of medium to weak intensity are assigned to fundamentals, overtone, and combination modes of structure (I) and to the more intense fundamental modes of the trans conformer, structure (II). Both conformers belong to the point group Cg and have fifteen fundamental vibrational modes (Figure 5.1). The impurity peaks (largely waters - H2O, HDO, D2O) can be easily identified [103 -105]. The calibrated spectra of propiolic acid are tabulated in Appendix A. Survey spectra of normal propiolic acid isolated in argon (using the Beekman IR-9 spectrophotometer) and carbon monoxide (using the Nicolet Model 7199 Series FT-IR) are shown in Figures 5.2 and 5.3. Propiolic acid spectra isolated in nitrogen were also recorded with the Beekman IR-9 and do reflect the overall features of the NIC 7199 FT-IR spectra. The fifteen fundamental vibrational modes to be assigned are given in Table 5.3. 59

Figure 5.1. The Two Possible Structures of Propiolic Acid: I and II. 60

I20©'>

' t »—J 20(3a^

3800

I IL-i 1 4OOOI

Figure 5.2. Survey Spectmm of Propiolic Acid Isolated In solid Argon Matrix (M/S = 4500/1). 61

o o o CO

CO

•g x O c o c o JO CO o "o

O

•g o < g "g 'Q. o

o E t3 CD Q. CO CD

CO

CO lO 05 g) ij_ 62 Table 5.3: The Fundamental Vibrational Modes of Propiolic Acid

Symmetry and Fundamental mode Description

A' - block

V1 O-H(D) stretch V2 C-H(D) stretch V3 C=C stretch V4 C=0 stretch V5 C-0 / C-O-H(D) deformation V6 C-O-H(D) / C-0 deformation V7 C-C stretch V8 H-C=C in - plane bend V9 0-C=0 scissor Vio C-COOH group deformation V11 C=C-C in-plane bend

A" - block

V12 H-C=C out - of - plane bend vi3 C=0 out - of - plane rock V14 C-O-H(D) torsion VI5 C=C-C out - of- plane bend 63 I: Hioh-frecuencv renion (4000 -2000cm-1^ Four fundamental vibrational modes of propiolic acid are expected to occur in the high frequency region: VQH.VQH, VQD. and VQ=Q. This region contains few bands. It does, however, provide seminal evidence for the presence of the trans conformer of PA in the matrix isolation samples.

The OH / OD Stretching Modes Besides the weak to medium intensity water bands at 3780.0, 3758.1, 3710 / 3712.0, and 3599.5 (vw) em"'', three bands consistently appeared at 3574.0, 3551.0 and 3531.0 em"'' in the spectra of PA isolated in Ar (Figure 5.4). The bands at 3551.0 and 3531.0 cm"'' are typical of the OH stretching frequency for an intramolecular hydrogen - bonded carboxylic acid mononer [106]. Furthermore, the oxygen -18 spectrum of this region shown in Figure 5.5 reveals that the oxygen -18 splittings of about 11.5 cm"'' occur for both bands. The band at 3574.0 cm"'' could arise from water (vi, H2O dimer). Nonetheless, it is attributed here to the free OH stretching mode of the trans conformer-strueture (II). The difference between the 3574 cm"'' band and either the 3551.0 or 3531.0 cm"'' bands in an Ar matrix of 23 - 43 cm"'' reflects the red-shift of the hydrogen-bonded OH stretching mode. Figure 5.5 reveals that the 3574 cm'^ bond of figure 5.4 shows an almost normal oxygen -18 splitting pattern with bands at 3573.0 (A), 3566.0 (B), 3563.0 (C) and 3552.5 (D) em"''. Bands E (3533.8), F (3528.2) and G (3524.4) would constitute an unusual oxygen -18 splitting pattern. Those could not be analyzed in more detail because of the poor background. In spectra of Ar / PA at less concentrated M / S ratios, the 3531.0 cm"'' band is more intense than the 3551 em"'' band. In such spectra, the 3574 cm"'' band is observed to be very weak. This variation in intensity was used as an aid in assigning the OH stretching fundamental in the other matrices. The OH and CH stretching regions of propiolic acid isolated in nitrogen (A), neon (B), and carbon monoxide (C) are shown in Figure 5.6. Figure 5.7 shows the OH stretching region of N2 / PA-''^0. Furthermore, the oxygen -18 spectra of this region reveals that oxygen -18 splittings of 10 -11 cm"'' occur for both bands. The spectra of propiolic acid isolated in 64

X

"(0 c o

<

T3

O cn < a.

CA Cg

k. CO c o "S

O

c CO

"o 2 t3 0} Q. CO T3 0) V- CO

in CD 65

Figure 5.5. Spectra of VQH ^^^ VQH Stretching regions of Ar / PA-'' ®0. 66

Rgure 5.6. The OH and C-H Stretching regions for PA Isolated in Nitrogen (A); Neon (B); and Carbon Monoxide (C). 67

Rgure 5.7. The OH Stretching region of ^ ^O-labeled Propiolic Acid Isolated in Nitrogen Matrix at 11.9K and 18.3K (M / S = 3000 /1). 68 neon show that the OH stretching mode occurs at a much higher frequency, that is, at 3580 cm"'' with a weak shoulder at 3569.0 cm"''. Very weak bands are observed to occur at 3578.5/3784.0 and 3500 cm"'' . The former are attributed to matrix isolated water and the latter could be assinged as an overtone band. There is a very weak shoulder observed at 1749. cm"'' (2 X 1749.7 = 3499.4 cm"''). The OH stretching mode of propiolic acid isolated in solid carbon monoxide matrix has been assigned as the strong band at 3438.8 em"''. This red shift of at least 100 - 141 cm"'' compared to the other matrices is typical of the hydrogen-bonded OH stretching mode. It could also reflect the extent of sample / matrix interactions and their dependence on the polarisabilities of the various matrices. For monodeutererated propiolic acid (PA-OD) the OD stretching mode is observed clearly at 2602 cm"'' with a weak shoulder at 2608 cm"'' for the cis conformer (data for Ar matrix). A very weak band at 2622 cm"'' has been tentatively assigned as the OD stretching mode for the trans conformer, structure (II) as shown in Figure 5.8. In spectra of PA isolated in N2 matrix, the OD stretch is observed as a doublet at 2601 and 2609 em"'' (11.4K). A single band is observed to occur at 2607 cm"'' when the spectrum was recorded at 18.9K (Figures 5.9 and 5.10). The ratio of intensities of cis propiolic acid bands of PA-OD compared to PA-OH is about 2:1. This may reflect a higher H-bonded stability for PA-OD compared to PA-OH. No spectra were recorded for PA-OD isolated in Ne and CO matrices although it would be a minor component in the PA-D2 spectra isolated in CO matrix. In dideuterated propiolic acid (PA-D2 ) spectra, there are two principal species: DC^CCOOD and DC^CCOOH. Figures 5.11-13 show spectra of PA-D2 isolated in argon, carbon monoxide, and nitrogen. The relationship between VQH 0^ DCCCOOH and VQD of DCCCOOD is clear. The stretching frequency for DCCCOOH is observed to occur at 3515.6 cm"'' , 3495/3463 cm"'', and at 3523.2 cm"'' in argon, carbon monoxide, and nitrogen matrices respectively. The corresponding bands for DCCCOOD monomer are observed to occur at 2590.6 cm"^, 2588 / 2578 cm"'', and at 2602 / 2600 cm"^ in argon, carbon monoxide, and nitrogen matrices 69

o in CO II CO

CO a o < Q.

c g £ CD C o CD ^. CO o III o •o c CO Q" O I o O sz H 00 in (D

^.

IT 70

a>

(0 o o o

CO o O < QL CSJ

E .

® u Q. o CO o

I- ' O oi § ID Tj- 2 ^ D -2 Ui C3) il £ 71

T3 CD •XD CO sz c O w c CO Q O >

CD 00

O o o II CO

Q O <

CSJ Z "o

® CO in 2 3 72

E o o o o CSJ I o o o

o o lO Q CO o o O O cso 9 Q °- O ^ > < • • o CO o 18

lO o

.g>. 73

E c o o o CVJ i£ 1 CSJ o T— o •*-^ o CO "^ (A

•— Q 2 O > CVJT3

G1 c < m 0. X "^ COo oo "o o CQ) ^^2^ *- -* C) (D X CI U) >o — H) CO ^ CD '*"' Q CD c csi ^ o • lO x{/:)

E o o o o CVJ • o o o "^ g a> (D

CVJ

CO o o o II CO oCV J I < Q. CVJ

E 5 ts CD C^ CO B o CO in CD k_ 3 m 75 respectively. Figure 5.14 shows ''^O-spectra for the stretching regions of monomer and dimer PA-D2 isolated in argon and nitrogen matrices.

Acetylenic Group Stretching Modes (vc-H(D).^ VQ=Q) The C-H stretching mode is expected to occur between 3300 and 3350 cm"''. In structure-related molecules [8,9] and in a previous study [38], the CH stretch is observed in the region 3300 - 3334 cm"''. In the present study, it is observed to occur in the region 3307 - 3328 em"'' . It exhibits two important characteristics: (1) constancy in position; and (2) the presence of a weak shoulder on the high frequency side contrary to a weak shoulder generally observed on the low frequency side [8]. This is only a minor difference as matrix spectra indicate a probable dependence on the matrix material. In the present matrix isolation study of propiolic acid, the CH stretching mode is abserved to occur at 3315.6 em"'', (3324.6, shoulder), 3307.0 em"'' (3310.0, shoulder), 3325/3328.0 em"'' (3349, shoulder), and at 3302.5 cm"'' (3308.8, shoulder) in argon, nitrogen, neon and carbon monoxide matrices respectively, (Figures 5.4 and 5.6). The presence of a weak shoulder has been explained in part as the result of Fermi resonance involving the various stretches in the H-C=C-C group. The CH vibrational mode is a relatively weak absorption in the infrared but strong absorption in the Raman [38]. This property was found useful in assigning the C-D stretch in PA-D2. In the Raman spectra of DCCCOOD, the C-D stretch is observed as a strong band at 2610 cm"'' [38]. In the present study, it is observed as a medium to strong band at 2614.3 em"'', 2643 / 2628 cm"'', and at 2618 / 2615.8 cm"'' in Ar, CO, and N2 matrices respectively (Figure 5.11-13). It is slightly higher than expected because it participates in Fermi resonance with the other modes of the acetylenic group. The CH / CD frequency ratio can be correlated in Ar, CO, and N2 with the magnitude being 1.246,1.252, and 1.264 in argon, carbon monoxide and nitrogen respectively. The C=C stretching vibration (V3 ) is known to occur near 2100 cm"'' from studies of simple molecules. Katon and McDevitt assigned a medium intensity band at 2130 cm"'' (IR solution), a very stong Raman polarized 76

Figure 5.14. "'^o-iabeled Spectra of PA-D2 In the 0-D region. 77 band at 2129 cm"'' and a type B contour band at 2125 - 2137 cm"'' to this mode. In the present study, it is assigned as the medium to strong band observed at 2138.0 cm"'', 2140.0 cm"'', and 2139.5 cm"'' in nitrogen, argon and neon matrix spectra respectively (Figure 5.15). However, substitution of C-D for C-H in H-C=C-COOH lowers VQ=Q by approximately 153 -161 em"'' in argon matrix spectra, and by 156-164 em"''in the spectra of PA-D2 isolated in nitrogen matrix (Figure 5.16). It is observed to occur at 1985.4 / 1977.5 em"'' in Ar matrix spectra, 2093.0 / 2089 cm"'' in CO matrix spectra, and at 1982.0 /1974.2 em"^ in nitrogen matrix spectra. This frequency behavior of VQ=Q with substitution of C-D for C-H has been attributed to coupling between the yQ=Q and VQ.Q modes, since VQ.Q occurs at a higher frequency and "^0=0 at a lower frequency than expected [9]. In addition to the assignments discussed above, bands associated with monomer and dimeric species of H2 O, D2 O, and HDO were identified by comparison to assignments from known water spectra and various other studies [103-105]. A summary of the various water species identified in the present study is given in Table Bl for Ar, N2 , CO and Ne matrix spectra (Appendix B). In some spectra (PA-D2 spectra) the 0-D stretch of the DCCCOOD dimer is observed to occur at 2340.6 em"'' in Ar matrix spectra; 2346.0 cm"'' in CO matrix spectra, and at 2348.0 em"'' in N2 matrix spectra. In N2 / PA-D2 spectra this mode is associated with a multiplet stmcture as other absorptions occur at 2328 em"'' and 2351 em"''. In PA spectra, a sometimes weak to medium intensity absorption observed at 2340 - 2350 cm"'' is due to matrix isolated CO2. In CO / PA and CO / PA-D2 spectra, a number of bands (usually very intense) are observed in the region 2000 - 2100 cm"''. These are associated with CO matrix absorptions [103].

II: The Carhonvl and Minfrpguencv Regions V£=Q Stretch Matrix isolation spectra of propiolic acid monomer in the carbonyl region are somewhat complex. The number of bands observed in this region is very dependent on the M/S ratio. Figures 5.17, and 18, 5.19 - 78

Figure 5.15. The vC^C Stretching region of Propiolic Acid Isolated in Various Matrices. 79

Figure 5.16. The VQ=Q region of PA-D2 Isolated In Various Matrices. 80

c T "D CD ^M^ CO O V) •D O < O o CL O i_ DL >»_ O c o CD CD k_ CD C , SI ^ ^ »o • . CD T— k. T— CO "cO >N „-^ C CD

•*—^ o c o CD Car b CD 2 •o h- CO • ^l^m^ r^ < •• lO c CD o 3 p CD < LL.E 81

Rgure 5.18. Spectra of the VQ^Q and VQ.Q Stretching regions of Ar / PA-OD at 11.4K (M / S = 4000 /1). 82

Figure 5.19. The Mid-frequency region of Propiolic Acid Isolated in Nitrogen. 83 5.21, and 5.22 show spectra of the carbonyl region of PA isolated in Ar, Ne and CO matrices under various conditions. The VQ^Q stretch is generally observed to be the most intense band in the entire spectmm. Figure 5.17 shows the principal absorption bands of monomeric propiolic acid isolated in argon and neon. The carbonyl stretching fundamental is observed to occur at 1753.7 cm"'' in Ar and at 1759.0 em"'' in Ne matrix spectra. It shows a weak shoulder on the low frequency side in both matrices (1750.5 cm"'' in Ar, and 1756.0 em"'' in Ne). The band at 1741.3 (Ar matrix) is too intense to be an overtone or combination band. It is attributed to the carbonyl stretch of a ketone as in glyoxylic acid [107]: The medium intensity band at 1725.9 cm"'' is associated with the C=0 stretching mode for propiolic acid dimer. In N2 matrix spectra, the carbonyl stretching fundamental is observed as a very intense band at 1748.5 -1752.0 cm"''. It is assigned here as the intense band at 1748.5 cm"'' ( Figure 5.19). The medium intensity band at about 1595-1598 cm'^ is associated with the V2 mode of monomeric H2 O. Figures 5.20 and 5.21 show the effect of annealing spectra of propiolic acid isolated in N2. At 11.4K, a general case of doubling is observed for the carboxylic acid group modes - VOH(OD)' VC=0 ^nd S/QQH) / VQ.Q- On warming the sample to 18.9 - 20K, this doubling disappears. In the carbonyl region, for example. Figure 5.21 shows that the bands at about 1752.5 and 1749.5 coalesce to give a more intense band at about 1748.5 em"^. Cooling the system back to 11.4K, Figure 5.21 b, shows that the slow annealing process is irreversible. The same behavior is observed for the vc=0 stretch of PA-OD in which the bands at 1740.0 and 1743.5 em""' coalesce to give one band at about 1740.0 cm'^. In the spectra of propiolic acid isolated in CO matrix the carbonyl stretching fundamental is assigned as the strong band at 1746.2 cm"'', (Figure 5.22). The carbonyl stretch is little affected by complete deuteration of propiolic acid. In PA-D2 isolated in Ar matrix spectra the carbonyl stretching frequency is assigned as the strong band at 1715.1 cm"'' (DCCCOOD) and 1721.3 cm"'' (DCCCOOH). In spectra of PA-D2 isolated in N2, the corresponding assignments are 1738.0 cm"'' and 1750 cm"'' for DCCCOOD and DCCCOOH respectively (Figure 5.23). Figure 5.24 shows 84

o CVJ T3 C CO o

CO < Q.

CVJ

o c g CD CD

c o JQ CO o

CO o o lO CO II CO

Q O <

QL

CSJ g CD CD CD c 'sz o k_

">» c o CO k- 00 CO O CD

•D c CSJ lO CO

3 CD^ 86

•D CD ^^ CO o C/) •D O < o o CL O ^. CL »»_ o c o CD CD k_ ">c , o X JD L_ k- CU cn ^ oCD CD -o hx-: X o . c- CSJ o CSJ s lO o(_ a ur e CD CO Li-O 87

Figure 5.23. Spectra of the VQ^Q region for PA-D2. 88

1 1 1 1 -I— 1 1 1 1 1 1750 1720 169 D

W 1739.2 A / 1704.8 \/ I70a2 ^"^^'"^"'"'^-''^^ 17482

1715.8 I A 1 s

\^ j\ yVv^'^^^'Ar/PA^ I / I74a5

1750.5

1760 , 1730 Il7[6.8. '"^^ , 1

Figure 5.24. Spectra of PA-''8o Isolated in Nitrogen and Argon for the Carbonyl region. 89 the complicated stmcture of the C=0 stretching modes of oxygen-18 substituted propiolic acids isolated in nitrogen and argon. The oxygen-18 splitting of the VQ^Q "^ocle can be explained without much trouble. The vc=''^0^"^the VQ^18o modes are assigned to the bands at 1748.2 cm"'' and 1715.8 em"'' in N2 matrix spectra, and at 1750.5 /1749.0 cm"'' and 1716.8 cm'"' in Ar matrix spectra respectively. The isotopic shifts are 32.4 cm"' and 33.7 em"'' in N2 and Ar matrices respectively. These shifts are comparable to shifts observed for this same mode in glyoxylic and pyruvic acids [107]. The assignments of the various carbonyl group modes for the different oxygen isotopic congeners are summarized in the next chapter.

The Mid-frequency Region This region contains the C-0 stretching vibration, overtones of the acetylenic group deformation modes and bands usually associated with carboxylic acid dimer species. Propiolic acid spectra for this region are shown in Figures 5.25 and 5.26 for argon and neon matrices. Figure 5.18 showed spectra of PA-OD isolated in solid argon matrix. The C-0 stretching frequency is rather difficult to locate in the vapor phase because of the numerous band envelopes in this area. Feairheller and Katon placed the frequency at 1413 cm"'', which is typical for some acids [108]. For example, in trifluroacetic acid (TFAA), the C-0 frequency is located at 1415 em"''. However, matrix isolation experiments Indicate that the C-0 stretching vibration actually has a frequency value closer to 1330 em"''; which is between the frequency for TFAA and acetic acid whose C-0 stretching frequency comes at 1259 cm"''[2]. In the present study, there are a number of medium to strong intensity bands in the region 1200 to 1500 em"''. In Ar/ PA and Ne / PA spectra, a strong band is observed to occur at 1303.0 cm"'' and 1301.8 cm"'' respectively. In addition to the above bands, there are other weak to medium intensity bands at 1235,1248.8, 1262.8, 1277, 1288, 1362, 1384.8, 1406.2 /1411.7 and 1427.5/1432.5 cm"'' in Ar/ HCCCOOH spectra. 90

c g CD CD k_ >s oc

CD

<

O 2

CO T3

in CVJ in (D k. 3 LL 91

c g 'CD

c (D c3 r

CD

•D

< Q.

(D

O CO tks_ a.CD CO T3 CD k. CO

CO CVJ in 2 3 CD IE 92 In Ar/PA-OD spectra (Figure 5.18), a very strong band is observed to occur at 1275.8 cm"'' while a weak to medium intensity band is observed at 1299.8 cm"''. In N2 / PA-OD spectra (Rgure 5.27), a strong band is observed to occur at 1277.2 cm"'', and a weak band at 1313.2 . It is well known that in carboxylic acids, the C-0 and COH modes mix very strongly. In the IR vapor spectrum, the C-0 mode is associated with absorptions in the range 1354 to 1373 cm"''. The isotopomer pattern of Figure 5.28 suggests that the band at 1362 (Ar / PA spectra) or 1360 (Ne / PA spectra) should be assigned as a combination mode between vy (C-C stretch), and a fundamental near 500 cm"''. There is a band at 529.5 em"'', assigned as vi4 (out-of-plane COH angle bend) with undetermined but small 0-18 isotope effects. Figure 5.28 shows that the assignment of 1362 (Ar / PA spectra) favors the combination mode assignment although its assignment as a fundamental cannot be unequivocally ruled out using the available data. Therefore, with 1362 and 1360 assigned as vy + V14, the nominal C-0 / COH mode,V5 , is chosen as the intense band at 1303.0 cm"'' in Ar / PA spectra. In N2 / PA and CO / PA spectra, the corresponding assignments for V5 are 1328.4 cm"'' and 1342.0 em"'' respectively. In Ar / PA spectra, there is a strong band at 1235 cm"''. This band is too strong to be either a combination or overtone band. It is associated here with the COH angle bend in propiolic acid dimer. The medium intensity band at 1277 cm"'' shows up in all the spectra. It is particularly intense in PA-OD spectra. It shows a recognizable oxygen -18 pattern. It Is thus assigned as the C-0 stretching fundamental in PA-OD. It is possible that it is the C-0 stretch in trans propiolic acid, structure (II). The weak bands occuring in the region 1240 -1270 are associated with either the overtone band of H-C=C or combination bands involving bands in the 600-400 em"'' region.

Ill- The 1200-680 cm"'' Rggipn Figures 5.29 -5.36 show spectra of either PA or PA-OD isolated in Ar, Ne, N2, and CO matrices. Figure 5.29 is a survey spectrum of Ar / PA in which matrix gas and PA vapor were codeposited (in situ-mixing) on the 93

00 00 "D c

CO

CD CO o in CO II CO

< Q.

CVJ

o c g 'CD CD ^. O) c sz o CD k. CO o oI > CD sz I- CVJ in CD k- 3 O) Li. 94

CO

< c T3 CD

O tn •g "o < o o a. o &. *o 2 XJ (D CL CO I O 00

00 CVJ in CD k_ 3 Li. 95 cold window. The other spectra were recorded from premixed samples with specific M / S ratios as indicated. The principal absorption bands in this region can be assigned to five fundamentals: the COH(D) in-plane angle bend, the C-C stretching fundamental, and the in- and out-of-plane deformation modes for the acetylenic hydrogen.

^^^ ^COH(D) ^ VQ.Q Fundamental Figure 5.30 shows that there is an intense band always present in the spectra in the region 1145 -1180 cm"''. In some of the spectra, a doublet is observed to occur in this region of the spectmm. From the previous section, it is clear that the band at 1263 cm"'' cannot be consistently attributed to combination or overtone modes of the weakly intramolecular H-bonded cis propiolic acid conformer, structure (I). In addition there is a weak to medium intensity band at 1112 cm"^ that cannot also be assigned to any modes of cis propiolic acid. These frequencies (1263 ± 2 and 1112 ± 2 em"^) are typical for nominal CO / COH and COH / CO deformations of the free carboxyl trans propiolic acid conformer, structure (II). The COH / CO and C-0 / COH modes are at 1259 and 1181 in acetic acid monomer [2,109] and at 1226 and 1120 cm"'' in formic acid monomer [110]. With the 1303.0 cm"'' (Ar) assigned as V5, the nominal COH / C-0 mode, Vg , is chosen as the intense band at 1148.5 cm"'' in Ar / PA spectra, 1149.0 cm"'' in Ne / PA spectra, 1153 cm"'' in N2 / PA spectra, and finally at 1178.9 em"'' in CO / PA spectra. The variable intensity band at 1158.5 (Ar), or 1161.5 (Ne) is attributed to the vg mode for the trans propiolic acid conformer. In Ar / PA-OD spectra, a single band associated with this deformation mode is observed to occur at 1159 em"''. The corresponding 6QQQ/VQQ band is observed to occur at 1015 cm"'' as shown in Figure 5.32. Figure 5.35 shows spectra of PA-OD isolated in nitrogen matrix in the region 1200 - 680 cm"''. The effects of temperature variation alluded to in an eariier section for carboxylic acid group modes is also very evident in this case. Both the COH / CO and COD / CO deformation modes are observed to be doubled at 10 -11.6K. On warming the sample to 18.8 - 20K, the doublets coalesce to form a single band at 1153.3 em"'' for the 96

E u o o CVJ

o 00 CO B c g 'CD 2 CD

CO E 2 B CD Q. CO < Q.

CD CVJ in o k_ 3 CD IT 97

"U CD •-- CO o tn "^ T3 O < O o CL O k_ CL «^_ O c o CD CD i» o o > ".«.. X o o CO CD sz CO a> c C3

CO tr i tws. sCO CD Q. c CO o CD C> Z CO X3 • C" m CO c 3 o CD B LI < 98

c g

CD

E o o 00 CO o o CD (D

< Q.

o 2 ts CD Q. CO CO in 2 3 Li. 99

o o LO CO

CO

Q O O O O O

o (/) CD •D O • ^' 1

CO Ch) •^ C O CD CO E CD CO 1 •D o L. o _) CM LL , c CSJ CO o CD

in ai_>^ CD k_ CD 3 SZ D) •*-' Li. _c 100

E o o o CO • o in 00 g CD CD

< CL a>

o 2 ts CD a. CO CO CO in CD i- 3 CD IT 101

Figure 5.34. The Mid-frequency region of N2 / PA Spectra at 11.4K. 102

CO c g CD CD

E o o CO CD I o o

CSJ

CD

o o to CO

CO

Q o I < CL

CSJ o CO ts CD vJ ^^ 00 to •D CO LO c CD 3 CO LJ. C-JD- 103

E o o CO CD

O CSJ CSJ

CD SZ o o lO

CO

< , CL '^ CD O b O O ••— II o •D E 1^

3 CILU O ) ect r a i_ CO O E CD II E, 5. 3 CD k- c 3 g CD CD 11 CD 104 COH / CO, and 1011 cm"'' for the COD / CO deformation modes. Throughout this study, temperature effects of this nature have been used as an aid in the assignment of fundamental vibrations associated with the carboxylic acid group modes. No spectra were recorded for PA-OD isolated in Ne and CO matrices although CO / PA-D2 spectra for this region indicate the presence of bands associated with PA-OD (Figure 5.37). The weak band at 1137 em"'' in Ar and Ne spectra is assigned here as the overtone band for the COH torsional mode,vi4 ^oreis propiolic acid. Figure 5.38 shows 18-0 spectra for the 6QQ|-J / VQ.Q mode isolated in argon and nitrogen matrices. All the bands show a blue shift from the normal propiolic acid bands. Assignments for all four oxygen -18 congeners have been summarized in the next chapter.

The VQ.Q Fundamental Mode It is probably a misnomer to call vj the "C-C" stretching mode, as isotopomer spectra show that the vj normal coordinate (from ab initio normal coordinate analysis) incorporates appreciable contributions from the C-0 and C=0 internal coordinates. This complex behavior of the C-C stretch must involve the intramolecular H-bonded conformer, stmcture (I). The C-C stretching mode shows remarkable constancy in both its intensity and position in the various matrices. It is a weak to medium intensity band observed to occur at 817.5, 815.2, 821.0, and 826.3 em"'' for Ar / PA, Ne / PA, N2 / PA and CO / PA spectra respectively. It is worth noting the relationship between the VQ.Q frequency and the polarisability of the matrix. It is clear, at least in the present study, that the C-C stretching frequency increases with the polarisability of the matrix. It is clear from spectra of this region that it shifts to higher frequencies when the molecule is either partially or completely deuterated. For example, in Ar / PA-OD spectra it is observed to occur at 820 cm"^ , while in Ar / PA-D2 spectra, it occurs at 825 / 850 cm"^. This same pattern is noted for PA-OD and PA-D2 spectra in N2 and CO matrices. 105

CoO CD a J* r* • Oc CD o 2 CaD E o o 00 o CO CO 1 o CVJ CVJ C

00 Rg u 106

"3W 1114.%

llSS-9

"•»Ml165 ^ •!» 115a

1172

11650 II4T4 ||:s4« il2«».S ims 171.1

iieo 11(0 . r L 111f!__0 l '1 WAVfiHlP6£(^S

Figure 5.38. The SQQH / VQ.Q region of ''^O-labeled Propiolic Acid Isolated in Argon (A) and Nitrogen (Bl and B2) at 11.4K and 18.3K. 107 The assignment of the acetylenic group bending modes in the present study differs greatly from assignments of the same mode for previous studies [38] and for related molecules [34,35]. In agreement with the ab initio normal mode analysis, the out-of-plane mode is assigned a higher frequency than the in-plane mode. This ordering pattern is in agreement with results of the well-studied acetylene reported in the literature and other texts [43]. It does however disagree with assignments given by Nyquist et al. for related molecules [8,9].

The TC and SH-C=C Modes The 7C|_|.Q=Q mode, and 5|-|.Q=Q modes are observed anywhere between 690 cm"'' and 762 cm"''. In Ar / PA spectra, the 7C|-|.Q=Q mode, vi 2, is assigned as the band at 755.2 cm"'', while 5H-Q=Q, VQ is assigned as the band at 694.0 cm"'' (Figure 5.31). The origin of the band at 736.7 cm"' is undetermined at this time since it is not consistently observed in various spectra. An initial guess would be that it is associated with degenerate modes of the H-C=C group in trans propiolic acid conformer, structure (II). The position of n[^.Q=Q seems to be independent of deuteration of the hydroxyl group as it is observed to occur at 756.0 em"'' in Ar / PA-OD spectra (Figure 5.32). In Ne / PA spectra ^Figure 5.33), Vi2 is observed as a strong band at 754.6 em"^ while vg is assigned as the weak band at 690.6 cm"^. In N2 / PA and CO / PA spectra, the corresponding assignments for V12 are 758.5 cm"'' and 762.5 cm"'' respectively (Figures 5.34 and 5.36). In N2 / PA-OD spectra, it is shifted slightly to lower frequencies being observed at 757.5 cm"^ (Figure 5.35). The assignments for VQ and V12 given above are based on the following argument: Substituted acetylenes with axial symmetry show a single strong band in the region 600-700 cm"'' arising from the H-CC deformation. However, the attachment of a planar group may result in splitting of the degeneracy and the occurrence of two bands corresponding to the out-of-plane and in-plane motions [10,11,18,19] Table 5.4 shows results of a normal cordinate analysis of propiolyl halides [21b, 23]. A helpful hint in the assignment of the deformation modes is obtained by 108 Table 5.4: 6/7C(HC=C) ^^^ Propionyl Chloride and Propionyl Fluoride*

Approx. H-C=C-COCI D-C=C-COCI H-C=C-COF D-C=C-COF mode vCal ^obs ^eal ^obs ^eal ^obs ^cal ^obs (cm"'')

^HC^ 691 696 530 522 692 695 528 511

^HC^ 706 703 547 547 735 737 570 (546)

*: The results presented here are based on a normal coordinate analysis carried out by Balfour et al. [ 23].

Table 5.5: V/Q.Q^Q) / V/|_|_Q=Q \ Ratios for the H-C=C Deformation Modes

Deformation Propiolyl Chloride Propiolyl Fluoride Propiolic Acid Mode Ar CO N2

5HQ=Q 0.75 0.735 0.876 0.871 0.860

TC HQ=Q 0.778 0.741 0.808 0.802 0.806 109 calculating the ratio for v D-C=C / H-C=C ^O'' ^^^ propiolyl halides and for the various assignments for propiolic acid in the three matrices studied (Table 5.5). Furthermore, in propynal [10], both the H-C^C and D-C=C deformation modes are associated with very strong bands at 662 and 520 cm"''. Spectra for PA and PA-D2 recorded in the present study do not give the same pair of bands in the regions described. It would seem that both bands are shifted to higher frequencies by about 100 em"''. In spectra of PA-D2 isolated in the various matrices, a strong band is consistently observed in the vicinity of 612 cm"''. Its assignment as TC / 6D-Q=Q will be discussed in the next section. PA-D2 spectra in the various matrices show no bands or very weak bands in the 690-800 em"'' region. The absence of any bands in this region is assumed to be indicative of the extent of deuteration of the acetylenic hydrogen.

IV: The 680-400 cm"'' Region This region of the spectra is by far the most difficult to assign for two reasons: 1) it is very crowded; and 2) bands are usually of weak to medium intensity coupled with the possibility for band overiap. Having accounted for nine of the fifteen fundamentals and knowing that two of the remaining six modes occur in the region below 400 em"^, four fundamentals are therefore expected in this region. Figure 5.32 and Figure 5.39 show spectra of Ne / PA in the region 680-400 em"''. Spectra of PA, PA-OD and PA-D2 isolated in solid argon matrix are shown in Figures 5.40-5.42. Figures 5.43-5.46 and 5.47 - 48 show spectra in this region for N2 and CO matrices respectively. The assignment of fundamentals In this region assumes the ordering pattern established from ab initio normal coordinate analysis. In Ar / PA spectra, the most intense bands are observed to occur at 652.2 / 656.5, 587.5, 572.4 / 568.5 and 529.5 cm"''. There are additional weak bands at 637.8 and 621.5 em"''. In Ne / PA spectra, the candidates for assignment as fundamentals are observed to occur at 656.0, 570.5, 573.3, and 530.0 cm"^. There are in addition to the above, a number of weaker bands at 615, 609.2, 545, and 482 em"''. 110

o > m ; «•

o *CD CD 1—i"^ <^ - ^^•^ N9/HCCCOO H E o o 1 I o in o — o

55 0 CO CD 1 I

< Q.

CD 1 573 3 O 2 ^ o TS CD C^ CO

CD CO in CD 60 0 ^. 3 O) LL Ill

Rgure 5.40. Spectra of Ar / PA-OD in the 680 - 580 cm"'' region. 112

c g

CD

E u in m o 00 in CD

Q O < Q.

O CO B CD Q. CO in CD k. 3 LL 113

o o ^ 1 E o 10 o CVJ s* o 1 o 0o0 ian J* CD SZ •*—' _c in -—* r» •*" ^ "*^ O O lO O o II in CO ••^«

:^«M«Ki^ CSJ

52 5 Q < a. o "~~^ in in < «^» o CO in N in BCD Q. CO

8 42 . 10 • lO

o

< CL

CVJ o c g CD CD

E o o o m o 00 CO sz

CO in 2 3 LL 115

g

CD

E o o o "^ I o 00 CD CD SZ

CO ^_ B CD Q. CO

O O lO CO

CO

Q o I <

CSJ

lO CD 1— 3 g) Li. 116

in £} m a jf O M O -^ o *> (D CoO J« CD SZ ••—' c O ^..^ cn T— m ^^ O O lO ii> II K 0) CO ^>^

^—.2^ a CSJ CNJ Q (O 1 < CL ^.^ CSJ !3 2 ta o E o 3 »i h- B0) CS^J Cl.^ ^H tn ^ s in TT.O r* "•lO* CD^ CD »- v_ 3V o .g^E a 117

800 755 710 665 620 575 530 165 ^tO

Figure 5.46. Spectrum of N2 / PA-D2 (M / S = 500 /1) in the 800 - 400 cm"^ region at 20K. 118

900 B50 BOO 750 700 650 600 550 500

Figure 5.47. The 900 - 400 em"'' region of Propiolic Acid Isolated in solid Carbon Monoxide (M / S = 3000 /1). 119

c o

CO O

•D CD ^—• JO o CSJ Q < Q. B c g 'O) CD

E o o o o o in '^ II CO o o 00 CD ^B 00 2 '^ CD •g lO "x CD o 3 C O) O 120 Katon and McDevitt assigned a medium intensity band at 752 em"'' in IR solution spectra as the OCO in-plane bending motion of the carboxyl group. From previous assignments for formic and acetic acids [111], three fundamental vibrations associated with the carboxyl group of monomeric acids are predicted to occur in the 400-700 em"'' range. Because it is difficult to completely eliminate self association in carboxylic acids, the spectra in this region would be expected to be somewhat more complex. Wilmshurst [111] has described these vibrations for acetic acid monomer as : (1) the 0-C=0 bending motion at 654 em"'', (2) the CO2 in-plane rocking motion at 536 cm"'' and, (3) the CO2 out-of-plane rocking motion at 582 cm"''. In spectra obtained from propiolic acid in the various matrices, bands are generally observed at about the same frequencies as those suggested above for acetic acid monomers. In TFAA [109], the 0-C=0 deformation mode is assigned to a band at 663 em"''. The assignments suggested below are based on the above argument. Figure 5.49 shows spectra of propiolic acid isolated in solid argon matrix. In this experiment, the sample vapor and matrix gas were deposited simultaneously. In assigning these spectra for Ar / PA, attention has been given to the method used in preparing the sample. Considering the similarity in electronic properties between the ethynyl group in propiolic acid and the CF3 group in TFAA, the strong band at 652.2 cm"'' (Ar) is assigned as the 0-C=0 scissoring fundamental in propiolic acid. Figures 5.50 and 5.51 show oxygen -18 spectra of propiolic acids isolated in argon matrix at 12K. Corresponding assignments for this mode in propiolic acid isolated in Ne, N2 , and CO are 651.5 / 650.0, 665, and 655 cm"'' respectively. It is possible that there are apppreciable contributions from the other carboxyl group modes to the general motion of this normal cordinate as indicated from plotting the general motions determined from ab initio vibrational analysis. The complex behavior observed for most of the carboxyl group modes must be associated with the intra - molecular H-bonded species (cis propiolic acid). The assignment above is consistent with assignments for the same mode in glyoxylic and pyruvic acids [107], formic acid [110] and acrylic acid [112]. There is a weak to medium intensity band in the vicinity of 586 cm"'' in the various spectra. 121

Rgure 5.49. Spectra in the 680 - 400 cm"'' region of Ar / PA. 122

00

g *CD CD

E o o o ^ lO o CO CD CD

•g -Ou> < o 01 9

1 "o •Qo . .ro o 616 . E o 3 -^ B CD u> Q. . CO ^ o CSJ 00 CO •^ 5165 0 X 40 u> o lO B ^00 66 3 lo:^ ^KT " 2 c / ^ L. < 123

c g 'CD 2

1. E o o o m • o 00 CO CD

CVJ o o 00 I < Q.

E E B CD CL CO I O 00

in in 2 3 CD 124 One of its characteristics is that it is insensitive to deuteration. Because of its overall constancy in position and considering a similar assignment proposed by Bently and Ryan, the 586.0 (Ne) 587.5 (Ar), 586.3 (N2 ) and 586.8 (CO) cm"'' band is assigned as the deformation mode (rocking) of the entire carboxyl group (PQ.QQQH )• Comparison of the acid spectra with those previously obtained for ketones and aldehydes [113], suggests that a strong band near 500 em"'' could be better described as a C-C=0 in-plane bending motion. In the present study, one of the more intense bands in this region is observed to occur at 573.3 cm"'', 572.4 / 568.5 cm"'', 607 em"'' , and 612.5 em"'' in spectra of propiolic acid isolated in neon, argon, nitrogen and carbon monoxide respectively. It is assigned as the TC Q^Q (carboxyl rocking motion), v-13 of the carboxyl group in the present study. Ab initio calculations indicate appreciable contributions from the ethynyl group deformation modes to the general motion of this normal corrdinate. In Ar / PA-OD spectra, a new band is observed at 508 cm"^. It cannot possibly be attributed to any other fundamental as no bands are observed at 572 cm"^. In N2 / PA-OD spectra, there is a new pair of bands at 517 and 507.5 em"'' (Figure 5.45). It is the contention of this study that this is the corresponding mode, that is, (TCQ^Q) in PA-OD species. The last unassigned fundamental is the COH torsion. For pyruvic and oxalic acid monomers, the OH torsional mode is observed as an intense band. It is also intense for glyoxylic acid [109]. In propiolic acid spectra, an intense band is observed to occur at 530.0 (Ne), 529.5 (Ar), 533.8 (N2 ) and at 541.0 (CO) cm"'' . In Ar / PA-OD and N2 / PA-OD spectra, deuteration results in a medium intensity band at 472.6 and 462.5 cm"'' respectively. The TCOD/TCOH band intensity ratio is about 2:1. It is apparent that the OH torsional mode responds in the same manner to Ar and N2 matrix environments. In addition to the assignments proposed above, there are a number of bands that are inconsistently observed in various spectra. A comparison of Figures 5.41 and 5.50 shows that bands at 623.5 and 617.8 are observed to be weak in the former but intense and well resolved in the latter. These bands may be due to the preseenee of the hydrated molecule. There is. 125 however, no general evidence to support the consistent presence of hydrated species in the matirix isolation spectra analysed. In PA-D2 spectra, a key assignment is that of the deformation modes of the acetylenic group upon deuteration of the acetylenic hydrogen. Spectra of PA-D2 isolated in N2 and CO do not show many peaks in the 680-400 cm"'' region. On the other hand, this region is more crowded for Ar matrix spectra. In H-C=C-CH OH, the IR vapor phase spectra show bands at 628 and 650 cm"''. The substitution of C-D for C-H, however, results in a single band being observed at 500 cm"'' [9c]. Assignments have already been suggested for the out-of- and in-plane deformation modes of the H-C=C group. In spectra of PA-D2 Isolated in Ar, N2-, and CO, a band is consistently observed at 612.5 cm"'' with a weak shoulder at 609 cm"''. The D-C=C deformation modes are generally observed to occur in the 500-540 cm"'' range in related molecules. The band at 612.5 / 609.4 (sh) does not show any oxygen -18 effects. It is assigned on the basis of the above argument as the TC- and 5D_Q=Q for D-C=CCOOH and D-C=CCOOD in analogy to a similar situation discussed above for the HC=CCH2 OD / DC=CCH2 OD pair. The constancy in position and intensity of this mode in the various spectra must be noted . Table 5.6 presents a summary of the assignments for PA-D2 spectra in the 680-400 em"'' region. The previous sections have provided assignments for thirteen of the fifteen fundamental frequencies of propiolic acid. The two unassigned modes are associated with the deformation of the carbon skeleton. These are expected to occur in the region below 400 cm"''. Katon and McDevitt [38] assigned bands at 245 (medium) and 218 (strong) to the in-plane and out-of-plane deformation modes of propiolic acid from liquid Raman data. The predicted pattern from ab initio calculations and that observed for related molecules suggests that the higher frequency should be associated with the out-of-plane skeletal deformation mode. The assignments for these modes adopted in the present study are those for H-C=C-C(0)F [19]. This tentative assignment is based on the fact that the hydroxyl group and fluorine are isoelectronic and have just about the same mass. Therefore, such substitution of F by OH in H-C=C-COF to give H-C=C-COOH would little affect the vibrational energy and hence the vibrational frequencies. 126 Table 5.6: Summary of Assignments for PA-D2 in the 680 - 400 cm"'' Region

Ar CO N2 Assignments

401.5 406.3 405.3 409.0 YQQQ torsion in DCCCOOD 410.2 YQCO'" DCCCOOD 421.9 460.4 '^ODin DCCCOOD 483.6 YQCO'" DCCCOOH 532.5 OCO wag in

- PA-D2 587.5(vw) PC-COOH in D-CCCOOH 609.4(sh) 609.0(sh) 609.0(sh) ^c-c=o /^c=o 612.5 612.0 612.0 S/T^D-C^C 647.7 0-C=0 wag 650.2 654.7 0-C=0 scissor 659.0 660.5(w,sh) 661.4* n 667.2 679.7 127 This is clearly illustrated by comparing carboxyl group frequencies in the present work with assignments for H-C=C-C(0)F [19]. The in-plane C=C-C deformation has been assigned the frequency 189 cm"'' in H-C^COOH(D) isotopomers and 179 cm"'' in D-C^COOH(D) isotopomers. The out-of-plane frequencies are 229 and 217 em"'' respectively. This completes an assignment for all fifteen fundamentals of propiolic acid monomer. With the low frequency fundamentals, we can examine more possibilities for assigning weak bands in the various matrices as either overtones, combination or difference bands. In Ne / PA spectra, the weak band at 658.5 can be assigned as the y QQQ mode at 482 plus the 6Q=Q.Q mode. The previous sections have dealt with the assignment of the fundamental frequencies of the monomers of the principal propiolic acid isotopomers. These assignments are summarized in the following tables (Tables 5.7 - 5.9).

V: Comparison of Ab Initio MO Calculated and Observed Frequencies To aid in the assignment of fundamentals, ab initio frequencies were calculated using the MO program GAUSSIAN 82 (to be discussed subsequently) using the 6-31G* basis set. These calculated frequencies have been compared with the observed frequencies in Tables 5.10, and 5.11 and in Figures 5.52 - 5.54 for the fundamentals of PA observed in Ar, N2, and CO. It should be recalled that ab initio frequencies calculated are based on the harmonic approximation and they are usually higher. However, most of them could be linearly connected with the observed frquencies by an empirical conversion factor cocalc^^abs ^^'^^ ^^'''^^ ^^^^ "^^^'''^ ^° "^^^"^ but is overall consistent. The conversion factors are shown in Table 5.10 to be 1.110+0.06 for argon, 1.12 ±0.07 for carbon monoxide, and 1.11 ±0.06 for nitrogen matrix. The expected empirical conversion factor is 1.11. The agreement between the observed and theoretical conversion factors is 128 Table 5.7: Fundamental Frequencies of Matrix isolated Propiolic Acid Monomers in Argon

Symmetry HCCCOOH HCCCOOD DCCCOOH DCCCOOD Mode v(cm"'') v(cm''') v(cm'-l) v(em"'')

A' - block V1 3550.0 2602.0 3515.6 2590.6 VO-H(D) V2 3315.6 3312.0 2614.3 2614.3 VC-H(D) V3 2140.0 2130.0 1985.4 1977.5 vc^c ^4 1753.7 1742.0 1721.3 1715.1 vc=o 1276.4 ^5 1303.0 1275.8 1352.4 vC-0/S(COH) 932.3 ^6 1151.0 1015.0 1127.0 SC0H/V(C0) 850.8 ^7 817.5 820.0 825.0 V(C-C) 612.5 612.5 ^8 695.1 693.0 Sc=C-H(D) 654.7 647.7 ^9 653.1 650.5 SQCO 587.5 483.6 Vio 587.5 586.5 PC-COOH(D) 179.0 179.0 ^11 189.0 189.0 Sc=c-c^^

A"- block vi2 759.0 756.0 612.5 612.5 ^C=C-H(D) Vi3 571.9 507.4 650.2 618.6 7^0=0/^0=0 471.5 534.4 421.9 '^COH(D) ^14 534.4 229.0 217.0 217.0 ^C^C-C"''' ^15 229.0

v: represents a bond stretch, 6: represents an inplane angle bending, p: represents a rocking motion, TC: represents an out of plane angle bending motion, and x or y represent a torsional or twisting motion; co: represents a wagging motion. ++: values taken from reference 19. 129 Tabte 5.8: Fundamental Frequencies of Monomer Propiolic Acid Isolated in

Symmetry HCCCOOH HCCCOOD DCCCOOH DCCCOOD Mode"

vlcm-l) v(cm-'') v(cm-1) v(cm-1)

A'- block

VI 3520.0 2607.0 3523.2 2600.0 VQ-H 3307.0 3307.0 2618.0 V2 2615.8 VC-H(D) V3 2138.0 2135.0 1982.0 1974.2 VC^C 1748.5 1740.0 V4 1743.0 1738.0 VC=0 1327.5 1277.2 V5 1354.6 1276.4 vC-0 1153.0 V6 1012.0 1127.2 941.0 SCOH(D) V7 820.5 820.0 848.0 851.0 VC-C 708.0 Vg 708.0 612.0 612.0 ^C=C-H(D) V9 668.0 664.0 661.4 661.4 ^0-0=0 VlO 588.0 588.0 587.5 483.6 PC-COOH(D) vil 189.0 189.0 179.0 179.0 Sc^c^^

A"- block

V12 759.0 757.0 612.0 612.0 ^C=CH(D) vi3 609.0 507.0 650.2+ 618.6+ ^C=0/«C=0 Vl4 535.5 463.5 535.5 _ 460.4 '^COH(D) VI5 229.0 229.0 217.0 217.0 ^C=CC "^"^

**. These descriptions are the same as those of Table 5.7. +: Argon matrix value. ++: values taken from HCCC(0)F (reference 19). 130 Table 5.9: Fundamentals of Monomer Propiolic Acids Isolated in Carbon Monoxide

Symmetry HCCCOOH DCCCOOH DCCCOOD Mode V (cm"'') V (cm"'') V (cm"^)

A'-block 3438.8 vi 3495.0 2588.0 vo-H(D) 3302.5 V2 2643.0 2628.0 VC-H(D) 2139.6 V3 2093.0 2089.0 VC^C 1746.2 V4 1743.0 1738.0 VC=0 V5 1342.0 1354.0 1275.0 vC-0 V6 1178.8 1125.0 933.0 SCOH(D) v? 826.3 847.0 851.0 vC-C Vg 699.0 612.0 612.0 SC=C-H(D) V9 659.4 659.0 660.5 ^0-0=0 Vio 586.8 587.5 (483.6) PC-COOH(D) vil 189.0 179.0 179.0 Sc=c-c^^

A"-block V12 762.5 612.0 612.0 ' ^C=C-H(D) VI3 612.5 (650.2) (618.6) ^C=0/^C=0 VI4 541.2 541.2 465.3** /427.3 '^COH(D) VI5 229.0 217.0 217.0 ^ C=C-C"*"^

+: No spectra were obtained here for the HCCCOOD isotopomer. ***: Same definition for symbols used in Tables 5.7 and 5.8; **: very very weak, difficult to say. **Ar matrix value, ( ): N2 matrix value. ++: values taken from reference 19. 131 Table 5.10: Comparison of Observed and MINDO / 3 Calculated Frequencies

Symmetry ^cale/vobs Calculated Frequencies

Ar CO N2 MlNDO/3 0-9wealc

A'- blOCK VI 1.323 1.365 1.334 4695 4223

V2 1.144 1.145 1.145 3790 3411

V3 1.1 1.1 1.101 2354 2299

V4 1.099 1.104 1.101 1928 1735

V5 1.098 1.069 1.081 1435 1292

V6 0.922 0.913 0.919 1069 962

V7 1.220 1.211 1.218 1000 900

V8 1.046 1.040 1.027 727 654

V9 1.021 1.012 0.999 667 600

Vio 0.713 0.794 0.793 466 419 vil — — — 176 158

1.06±0.10 1.12±0.19 1.11 ±0.09 1.11

A"- block

VI2 1.026 1.021 1.026 779 701 vi3 1.080 1.008 1.015 618 556 vi4 0.968 0.958 0.966 518 466

— — — VI5 257 231 132 Table 5.11: Comparison of Calculated versus Observed Frequencies for Ne / PA

Symmetry

VNe/PA ^MINDO/3/VNe ^6-31G*/VNe

A'- block VI 3580.0 1.312 1.131 V2 3328.0 1.139 1.101 V3 2139.5 1.100 1.136 V4 1759.0 1.096 1.154 V5 1301.8 1.055 1.109 V6 1147.2 0.932 1.163 V7 815.2 1.227 1.097 V8 690.6 1.053 1.196 V9 651.5 1.024 1.019 Vio 586.0 0.795 0.988 vi1 189.0 0.931 1.175

1.109±0.06

A"- block VI2 754.6 1.032 1.169 V13 573.3 1.078 1.511 V14 530.0 0.977 1.222 vi5 229.0 1.122 1.266 133

•c B :2 c o CD

cn g 'o c CD 3 cr CD w. LL T3 CD cn JO O CA 3 U) CD > T3 CD B 3 jO CO o B c o cn •c CO Q. E o O cvj in in CD 3 £ 134

X •c B c CD o> O

/" y (A 2 - _^-..y^j^-^"o 9r c CD 3 CT CD u. •o tCD CD cn y J3 / / O y cn / o 3 / o, cn y CsJ k- y > T3 0) / B 3 / io / CO / Oc o cn 'k_ y CO C3. E O O o o o O o o IS. o CsJ CO r-K CO in CO c CO U O o in CD

CO

CD •g "x o c o c o oCO

cn 2 *o c CD c3 r CD

CD a: CD cn o cn 3 cn CD > TD CD B 3 jO oCO c o cn •c CO CL E o O

"^ in in CD w. 3 CD 136 remarkable. This consistency is observed for both in- and out -of-plane vibrational fundamentals. This comparison provides further evidence for the correct assignment of fundamentals in the present study. The MINDO / 3 calculated frequencies have also been compared to the observed frequencies. There is no doubt that they predict the correct ordering of the fundamentals. The calculated frequencies seem to be especially good for the very low vibrational modes. The MINDO/3 method however overestimates the v-] and V2 modes by at least 32%. Electron correlation methods can be used to improve the effects of anharmonicity which generally tend to lower frequencies. This overestimation is not unexpected. It reflects basis set effects. The MINDO/3 method uses STO basis sets. The 6-31G* basis set includes polarization functions. This is a larger basis set and results of calculated parameters usually tend toward the HF limit yielding better agreement with experimental data.

VI: Discussion In making the assignments above attention has been given to the behavior of propiolic acid in the different matrices. A case in point is the effect of temperature on the spectra of propiolic acid isolated in argon, carbon monoxide and nitrogen. When annealing temperatures were kept below T^, the temperature at which diffusion first becomes important (35K for Ar), there are really no significant changes in the appearance of the spectra. For Ar / PA spectra, bands at 535, 567 and 617 cm"'' disappear. More significant is the definitiveness of bands at 930 - 935 cm"'', and 1727 cm"''. The former bands cited could quite possibly be associated with fundamental modes of the trans propiolic acid conformer. The latter bands are usually associated with carboxylic acid dimers and no alternative assignments will be suggested in this ease. Figures 5.55 - 5.58 show spectra of Ar / PA recorded at 30K. Survey spectra of CO / PA recorded at 12, 24 and 35K are shown in Figures 5.59-5.61. Figure 5.61 shows that at temperatures more or less equal to T^-j, the diffusion temperature (35K for CO) the spectaim lost most of its features. The observed temperature dependence is irreversible. The 137

Figure 5.55. Ar / PA (M / S = 4000 /1) Spectrum in the 680 - 400 cm"'' region at 30K. 138

c o CD CD

E o o CO CD I o o CSJ

CD x:

o o lO

II CO

< Q_

o E 3 B CD Q. CO CD lO lO CD ^ 3 O g)2 LL B 139

o CO B

o in

II CO

< CL

O cn g O) CD o c CD c3 r CD

T3 T3 C CO *>c% o

CO O CD

m in CD k_ 3 CD IT 140

o CO B o in

II CO

< QL o c g "CD 2 >> co CD c3 r CD

X

CD

00 m in (D k. 3 g> LL 141

CSJ

CO o o o CO II CO

< CL o o B E 3 B CD CL CO CD

3 CO cr> lO lO CD k_ 3 LJ- 142

CSJ B o o o CO II CO

< CL O O B E

o CD CL CO >% CD

3 CO b CD LO CD v_ 3 143

LO CO B

o o o CO II CO

< CL o o B E o CD CL CO

3 CO

CD lO CD w_ 3 144 corresponding intensity changes for some bands observed in the spectrum suggests that population changes for various species caused by sample diffusion or isomerization may have occurred during the annealing process. Temperature dependent behavior of spectra of propiolic acid isolated in nitrogen matrix were found more useful. One general observation is that this behavior was particulariy important for -COOH group modes - OH, C=0, COH / C-0. Spectra for regions containing these fundamentals have already been discussed. This behavior is considered to be both a matrix effect and also an indication of the presence of other species. Matrix effects are discussed in the next chapter. In addition, no set intensity patterns were observed that seemed to suggest population changes for various species as a result of sample diffusion or isomerization during the gentle annealing process. Isomerization (internal rotation about the C-0 bond) would result in an increase in the population of the trans propiolic acid conformer. This would support observations from microwave and present data, that the cis propiolic acid conformer is the most abundant conformer at ordinary temperatures althought the possibility exist for the presence of both conformers in the vapor phase.

Propiolic Acid Dimers The emphasis of the discussions in the previous sections has been on the assignment of fundamental vibrational modes of monomeric propiolic acid. The presence of propiolic acid dimer absorptions in some of the spectra has been noted in those areas where their occurrence is important in differentiating between monomer fundamentals and fundamentals of polymeric species, particularly, dimers. Such differences are particularly important in spectra of regions containing the carboxylic acid group modes. Figures 5.62 - 5.66 show spectra of PA - D2 isolated in Ar, N2, and CO in the regions 2000 - 1200 cm"'', and 1200 - 400 cm"''. The OH / OD dimer absorptions have already been discussed in the assignment of these same modes for the monomeric species. The emphasis of this section is on the C=0, and C-0 / COH(D) modes for the dimers. The carbonyl stretch is observed to have shifted by at least 100 145

a ^ — cDn 50 0 II CO "*.. Q :)x u s Q T-l B S c o a CD o < ID P r-l o

• CD sol a T-^ oCS i 1 < Q. w p D o o a o o Bc o o - r- o CJ CD TH CD CD o CD n p k. y— • Q E CD a O O - oo CSJ t-» T—

1 o o o Q CVJ a

- cn Th e TH • CVJ CO in CD u_ • 3 CD _L IL 146

CSJ Q < CL O O •D C CO

o o o CO II CO

< CL o O

co$^ o CO CD CL CO c^8 CD t?) lO II CD k_ CO 3 • CD. 147

o oo CD

D j« r- o CSJ o CSJ a •»~ 00 CD ^ •<-^ c o , ^ (D •^- CO -.». O O 1 LO o II CSJ CD CO . •^». ^- Si a CSJT3 oo en 9 "

a CVo i ^^ CD a SZ o cn _c »H CO ^^r^.f o D CD a Q. j« CO r^ ^—.V ^ -^... o o a in o T^ lOII CO ^»^ a 2 CO ' «-oi CSJ Q o < a CL h- •—. H CSJ Z c M- O o o'a cn CD s — ^- »-« CO^ CD 'c Q E o o 19 0 .65 . 20 0 lO 1- Q) " D 3O 8 ,g>8 CO LL CSJ 149

o o m II CO

CvJ Q < CL

c g *CD CD

E o o 00 CO oI o CSJ CD SZ

CO CO in CD k_ 3 CD 150 cm"''. It is observed to occur as the very strong band at 1592.2, 1599.8, and 1598.0 cm"''in argon, cariDon monoxide, and nitrogen matrix spectra, respectively. Figure 5.65 shows the contrasting behavior of the carbonyl stretch in PA (M / S = 3000 /1) and PA -D2 (M / S = 500 /I). The C-0 stretch, which is usually difficult to locate in PA spectra, is observed to be a very intense band in the vicinity of 1400 cm"^. It is observed to occur at 1403.0,1407.0, and 1405.9 cm"'' in argon, carbon monoxide, and nitrogen matrix spectra, respectively. The COH deformation mode is also observed in spectra recorded for low M / S ratios. Figures 5.36(00 / PA) and 5.38(N2 / PA - D2), and Figures 5.63 and 5.66 show that the COH deformation is sensitive to dimer formation. It is observed to occur as a medium to strong band at 1050, 1098, and 1098 cm"'' in Ar, CO, and N2 matrix isolated spectra respectively. Assignments for some dimer fundamentals are indicated in Appendix A2.

Conclusions Infrared spectra of propiolic acid monomer with hydroxyl, and acetylenic D (for the first time), have been recorded in Ar, CO, N2 and Ne (normal propiolic acid monomer only) matrices and analysed. Isotopic labeling experiments have been performed to obtain confirmatory evidence for the assignments of the various fundamentals of propiolic acid. Infrared spectra of general 0-18 labeling are reported for Ar and N2 matrix isolated samples. The data demonstrate that large, heavy atom isotope effects occur several of the vibrational modes. The data also demonstrate small isotope effects for some of the vibrational modes. No "• ^O - exchange spectra were recorded for neither PA nor PA-D2 isolated in CO and Ne matrix. Values for all fifteen fundamental frequencies of the cis conformer, Structure(l), have been proposed for the four principal isotopomers (HCCCOOH, HCCCOOD, DCCCOOD, and DCCCOOH). The frequencies correlate with results for structurally related molecules. A normal coordinate analysis for the cis conformer will be feasible when data for the modes observed to occur in the region below 400 cm"'' and results 151 of more detailed labeling experiments for the PA - D2 monomer are available. The assignments of the protonated acid were compared with ab initio calculated frequencies (6-31G*) as additional evidence for the correctness and consistency of the assignments proposed in this study. The overall fit has been found to be quite consistent. Having obtained a plausible self-consistent pool of observed frequencies, this would then be used as a source for data in determining the experimental force field. The assignments discussed can be divided into two groups. The first group consists of fundamentals in the 4000 - 680 em"'' region. Based on spectral evidence obtained herein and considering the various assignment techniques, the assignments proposed for bands in this region are believed to be correct. The second group includes absorptions in the 680 - 400 cm"^ range. Attempts have been made here to assign these as unambiguously as possible based on corraborative evidence obtained in the present study and by comparison to other published data for similar modes in structurally related molecules. The assignments of the acetylenic deformation modes and the CO2 bending mode are different from previous assignments. In lETS studies [20] the symmetric 6o-C=0 fundamental is assigned as the absorption band at 776 cm"^. Katon and McDevitt [38] give the assignment as the weak to medium intensity absorption at 752 cm"^. In TFAA, the same mode is assigned as the absorption at 781 cm"''. No characteristic splitting pattern was observed for this mode in ''®0-exehange spectra. Any splitting that would occur in this case, it is contended would be the result of matrix effects and self-association. Also the complete disappearance of this absorption in D-C=C spectra does not agree with assigning this absorption as a CO2 mode. While internal rotation about the C-0 bond is possible, no evidence has been noted in the present study from IR / Ml spectra for the existence of two comformers in premixed samples. There are, however, indications from infrared spectra in which sample vapor and matrix gas (Ar) are codeposited that the trans conformer, Stmcture (II). might be present in minor amounts. Matrix isolation sampling is capable of trapping a high-energy vapor-phase 152 conformer before it can relax to the low-energy form favored by cryogenic temperatures [114]. This would make propiolic acid an attractive candidate for conformational studies involving matrix sample deposition from carefully thermostatted effusion cells. In the present work, however, the relative populations of the two conformers could not be studied by temperature variations. For the most part, the doubling observed for some fundamentals is believed to be strictly a matrix effect., A similar conclusion was arrived at by Lister and Tyler [39] from results of microwave study. A theoretical study of this inversion barrier in this dissertation reveals that a high barrier exists for this conversion. Consequently, it only can be achieved through photochemical excitation. Furthermore, although it is difficult to obtain the vapor phase spectrum of propiolic acid, the technique of matrix isolation can be used effectively to obtain quality spectra. One of the problems associated with propiolic acid is its easy decomposition to acetylene and carbon dioxide. It is believed that such a decomposition would not be possible at the temperatures at which spectra were recorded. Certain vibrations of propiolic acid monomer are highly sensitive to the matrix environment. They appear differently in Ar, CO, and Ne versus N2 matrix isolation spectra. The strongest matrix dependence is shown by the hydroxyl, carbonyl, and COH in-plane bending fundamentals. The overall temperature dependence of spectra of PA and isotopomers isolated in N2 matrix reveal the following patterns: 1) for both stretching and bending modes, low temperature spectra show equal intensity for both absorptions, 2) at high temperatures (18K and up), for the stretches the low frequency absorption disappears, for bending modes, the high frequency absorption disappears. The slow annealing process in nitrogen matrix is irreversible. Several relatively intense overtone and combination modes, which are probably participating in resonance interactions with nearby fundamentals, are observed in matrix isolation spectra of PA. Oxygen isotope exchange reactions between propiolic acid and water were found to occur readily in concentrated mixtures, and it was found that 153 the exchange could not take place in the absence of catalytic amounts of hydrogen chloride vapor ( actually hydrochloric acid). A comparison of spectra for M / S > 3000 /1 and M / S < 500 /1 showed the presence of dimers in the latter. The identification of dimer absorption bands is based on a comparison with assignments for similar studies for formic and acetic acids found in the literature[115]. A number of bands associated with dimers of propiolic acid have been identified and assigned accordingly. CHAPTER VI MATRIX EFFECTS

Introduction The handling and easy decomposition problems associated with propiolic acid have been alluded to in an eariier section of this dissertation. It is for problems of this nature, stabilizing radical species and other "reactive" molecules such as monomers of hydrogen-bonded compounds in the solid phase for as long as needed, so that their spectroscopy can be examined that the technique of matrix isolation derives its greatest value. Under the circumstances, molecular rotations and translations along with a reduction in intermolecular interactions results in a sharpening of solute absorptions. In heavy atom labeling work, this allows the spectroscopists to measure the small frequency shifts usually associated with such labeling. This provides more observed frequencies needed to strengthen normal coordinate calculations and getting a more realistic force field for the molecule being studied. Despite all the advantages associated with this technique, one setback involves the fact that band origins in IR / Ml technique are usually displaced form gas phase values principally due to the interactions between the solute and solvent (matrix). In this work, the frequency shifts of propiolic acid vibrations caused by the different matrices have been observed. It is not uncommon to see spectral multiplets for some bands. Such variations in frequency are usually referred to as matrix effects. Infrared spectroscopy is one of the ways by which the nature of matrix isolated species can be established. In the present study, the various techniques by which IR spectroscopy can be applied for such a study have been previously discussed in the chapter on assignment techniques. However, it must not be forgotten that this technique does not imply isolation of molecules as would be the case in the gas phase. Therefore, it is indeed possible that the observed fundamentals of a molecule being

154 155 studied may be perturbed by the matrix environment. The results of such perturtDations by the matrix environment can be manifested in a variety of ways including: the number of absorptions; the position of the absorption and the shapes of these absorptions. The first effect is usually treated as matrix splitting, that is, the appearance of several closely spaced absorptions where only one is expected. Matrix splitting effects are usually the result of the existence of several possible sites in the matrix for the molecule. The effect of the matrix on peak position is usually referred to as matrix shifts. The third way in which matrix effects are manifested depends on the spectroscopic method used. In the present case as it is generally for infrared studies, absorption bands in the spectrum tend to be much narrower in the matrix than in the gas phase. The reason for this is the loss of rotational fine stnjcture in the case of matrix isolated molecules.

Theories of Matrix Shifts Til61 The treatment of matrix environmental effects is complicated even for atomic and diatomic species. The focus of the present study is the result of matrix splitting and matrix shifts. A matrix shift is defined by the expression (Equation 6.1)

^v = vmatrix - Vgas phase- (^•'')

The effect of the matrix on the energy levels, and hence the observed vibrational energies of a molecule may be treated as a perturbation of the harmonic potential by the interaction of the vibrating molecule with the cage that contains it. In the case of solvent shifts, a number of theoretical models have been used to explain these. For example, solvent shifts in solution have been explained by the Kirkwood-Bauer - Magat model [117]

Av/v =C(e-1)/ (2e-Hl ), (6.2) 156 and the Buckingham model [118] for nonpolar

Av / V = C-, -f-1/2 (C2 + C3) (e - 1) / (2e-h 1) (6.3a)

and polar solvents

Av/v = C-| +C2(e- l)/(2e+ 1) + C3 (TI^ .1)/(2TI2 + I) (6.3b)

David and Hallam[119]

Av / v = C-i' + C2' (e -1) / (e + 2) + C3' (T]2 -1) / (ri^ + 2). (6.4)

The above models assume the solute is a point dipole in a spherical cavity within the solvent medium of uniform dielectric constant e', and refractive index T|. The theories developed to deal with vibrational shifts in solution could be extended to cover matrices, the only difference being that in matrices the solute molecule is held in a rigid "cage," a situation in which repulsive forces may play an important role. In the above equations, C's are constants, e is the dielectric constant and ri is the refractive index of the solvent. Barnes has recently reviewed the extension of these models to matrix isolated species [116]. For a diatomic molecule, an additional term is added to the harmonic potential resulting in a modified potential of the form:

V = 1/2 k(r - re)2 + b(D - r)"" . (6.5)

The last term in (6.5) is a repulsive term. The effect of the matrix on the potential is determined by the nature of last term in Equation 6.5 (that is, whether it is replusive or attractive potential). The qualitative effects of the nature of the perturbation term have been more fully discussed by Cradoek and Hinchliffe [74]. If an attractive potential exist between the trapped molecule and the cage, it will have the effect of: i) increasing the equilibrium bond distance, ii) lowering the potential energy of the minimum, iii) decreasing the vibration frequency, 157 iv) increasing the anharmonicity. The converse is true if the potential is repulsive. While a diatomic molecule presents a simple picture, the situation is more complex for polyatomic molecules where the concepts of attractive and repulsive potentials superimposed on the molecular potential could be assumed. Some general conclusions have been derived from studies of bands for a number of compounds. For absorption bands below 1000 em"'' shifts are most often increased in the matrix whereas for absorption bands above 1000 cm"^ shifts are usually decreased in the matrix. The shift defined in Equation (6.1) is basically an intermolecular interaction problem which usually involves the determination of various potentials (electrostatic, inductive, dispersive and repulsive). Attempts have been made to analyze matrix-induced frequency shifts by direct summation of the interactions between the point polar solute molecule and the polarisable molecular cage. The intermolecular potential energy may be expressed as the sum of four terms, arising from electrostatic, inductive, dispersive, and repulsive contributions [120], and thus the shift is given by

^v = vm - Vg = Aveiec + ^Vj^^ + Av^js + Av^ep • (6.6)

The electrostatic contribution is zero for nonpolar matrices such as argon. In such matrices, it arises from the interaction between the permanent charge distributions of the solute and matrix molecules. Even when the matrix molecule has a quadrupole, as in nitrogen, the electrostatic shift is nullified because of the symmetry of the f.e.c. lattice. The assumption here is that the solute occupies a substitutional site. The inductive effect arises form the interaction between the permanent charge distribution of one molecule; and the moments induced in the other molecule. Based on a point polarisable atom model [121] the shift is given by

^Vjnd = "'^6ab^(^^a^) / ^^^0^ ^^''^^ do is the diameter of the substitutional site occupied by the solute molecule 158 in the matrix lattice, Ng is the effective number of nearest neighbors for an r "^ law (14.454 for a f.c.c. lattice [122]), [i^ is the dipole moment of the solute molecule and a^ is the polarisability of the matrix molecule. On the basis of a cavity model, the shift is given by:

Avjnd = - [8(e' - 1) / (2e + DhcdQ^ ] AC^i^^) (6.8)

In the models cited above, terms involving solute moments higher than the dipole are neglected. Both models are unrealistic as the point polarisable atom model underestimates, while the cavity model overestimates the short - range interactions with nearest neighbors in acutal contact with the solute molecule. Expressions for the dispersive shift (due to attraction between the instantaneous charge distributions of the solute and matrix molecules), were given by Friedmann and Kimel using a Lennard-Jones-Devonshire cell model [123]:

Avdis = " 4N6 e(x)o / he * [a(x)o / d^f F(a) (6.9)

and for the replusive shift (due to overiap of electronic charge distributions):

Avrep = 4Ni 2 e(x)o[^(x)o / dol^ ^ / he * [F(^i) + F(a)] (2+y /1 +y) (6.1 Oa)

where

F(a) =[ aa(x)v. aa(x)o] / aa(x)o, and

F(^i) =[^a(x)o V2e(x)o^^(x)o][^^a(>^)v" ^^a(x)o (^- "• O*^'

e{x)y and G(X)^ are v-dependent (v is the vibrational state of the solute molecule) Lennard-Jones parameters, y is a molecular parameter, and N-|2 = 12.132 for a f.c.c. lattice. 159 Matrix Materials [124] Three principal matrices (Ar, CO and N2) were used in this study. Some of the appropriate matrix properties for these materials were listed in the experimental techniques section. Obviously not every material can be used as a matrix material although the number of potential substances is large. However, there are several properties that a substance must have to be a successful matrix material. These include: i) purity: the matrix material must be of a high degree of purity. The non-reproducibility of some spectra has been attributed to the presence of impurities in the lattice. ii) Vibrational Spectrum: The matrix material should not have absorptions overlapping those of the isolated sample species particularly if these have not been previously identified. Ill) Inertness: It is a necessity that the matrix not react with solute although in specific experiments this might be a desirable effect. Generally, the limitations of this criterion will depend upon the nature of the solute and experiment. iv) Rigidity: Rigidity of matrices is a very critical requirement. This is necessary to prevent diffusion of solute molecules within the matrix cage which could lead to aggregation and formation of oligomeric species of the sample to be studied. This property is usually determined by T^j (the temperature at which diffusion first becomes important). The problem of diffusion in solid matrices has been discussed by Pimentel [125]. v) Volatility: The matrix material must have a sufficiently high vapor pressure at room temperature to allow it to be handled in the vacuum line since the matrix material and solute are usually mixed in the gas phase, and then sprayed on to the cold window in the sample cell. A detailed discussion of all the requirements and properties can be found in [124]. The choice of matrix in the present study was based on availability and their vibrational spectra. The noble gases and nitrogen have no absorptions in the near infrared and for most purposes are chemically inert. Argon, carbon monoxide and nitrogen were the matrices of choice. 160 Crvstal Data [124] The most commonly used matrix materials, the noble gases, are spherical atoms. The common molecular matrices, such as, nitrogen and carbon monoxide, are characterised by a nearly spherical electron cloud. The result of this spherical shapes is that these materials tend to adhere to the "principle of closest packing," crystallizing either in hexagonal close-packed structure (h.c.p.) or cubic close-packed structure (c.c.p.) as shown in Figure 6.1. Consequently, they maximize the number of nearest neighbors and the intermolecular van der Waals forces. In the h.c.p. and c.c.p. arrangements, each sphere has 12 nearest neighbors, six in its own plane, three in the plane above and three in the plane below. In a c.c.p. arrangement, the unit cell is a face-centered cubic (f.c.c.) lattice. The noble gases cyrstallize in this lattice. In the c.c.p. lattices, there are three possible guest sites. These are, one substitutional site (S) in which a sample (guest) molecule replaces a matrix (host) molecule, and two interstitial sites (tetrahedral (T) and octahedral (O)), Figure 6.2. It is obvious from Table 6.1 that none of these sites are large enough to accomodate the propiolic acid monomer. The geometry of propiolic acid is known from the microwave work done by Lister and Tyler [39]. The molecular size of propiolic acid can be estimated using van der Waals radii of atoms and bond lengths [127]. The molecular size of PA can be estimated to be approximately 4.495 to 5.30AO ^g shown in Figure 6.3. The propiolic acid monomer will most likely occupy a distorted double or triple substitutional site in Ar, CO, and N2 lattice if it is isolated within the crystal instead of at a grain boundary or other site.

M^^friy Fnvironmental Effects [1241 There are six general effects, all of which may combine to modify a vibrational band shape, intensity and frequency and, in some cases, to cause a single vibrational mode to have a multiplet structure. These include: mutliple trapping sites, molecular rotation, medium effect (matrix shift) aggregation, coupling with lattice vibrations and phonon bands. All of 161

(b)

(c)

Figure 6.1. Close-packing of Spheres, Showing Three Layers-a b r and Tetrahedral and Octahedral Holes. ' ' • ^'^^^a 162

T

(a)

O

(b)

Figure 6.2. Tetrahedral and Octahedral Holes between Layers of Close-packed Spheres. 163 Table 6.1: Site Diameters

Substitutional Interstitial site site(AO) Oh(AO) T^, (A^) Ref

Ar(4K) 3.755 1.56 0.85 [126]

(20K) 3.760

Matrix Crystal Transition Mean site Approx. site shape Ref

Stmcture Temperature Diameter A°

N2(4K) f.c.c 35.6 3.991 4.52x3.42x3.42 [78]

(20K) 4.004

CO(23K) f.c.c 61.6 3.999 4.61x3.48x3.48 [77] 164

5.3 A

Figure 6.3. The Size of Monomeric Propiolic Acid. 165 these effects may complicate the analysis of matrix spectra. Therefore, a knowledge of the nature of these effects is a necessary prerequisite for a complete interpretation of the experimental data obtained by the matrix isolation method. It is generally assumed that a guest molecule is usually trapped in a substitutional site, formed by the removal of one or more matrix molecules. The intermolecular forces between matrix and guest molecules will be different for each site and the resulting perturbations of the energy levels may lead to two sets of frequencies for a polyatomic guest. Trapped species are so tightly held at the low temperatures that only the pure vibrational modes will appear as sharp bands. Under conditions of perfect Isolation, the guest molecule is subject only to solute-matrix interactions. These will perturb the guest molecule's vibrational energy levels, and be reflected in a frequency shift. True isolation is achieved only at very high matrix / guest molecule ratio, usually greater than 1000. At low M / S ratios molecular aggregates may be formed and trapped in addition to monomers. Molecular association is usually greatest for guest molecules capable of hydrogen-bonding. The multiple features due to self association are usually readily identified from their concentration dependence and from warm-up experiments in which monomers may diffuse to form dimers and higher oligomers.

ppf^ijlt.q and Discussion While the Taylor series expansion of the intramolecular potential energy for a diatomic molecule is simple, the counterpart for a polyatomic solute molecule is more complicated:

V = Ve + 5:(5V / 5Qj)o Qj +1/2 ^^ (S^V / 5Qj5Qk)oQj Q k+ - (6.11) j i k where the Q; are normal coordinates. A qualitative approach has been 166 adopted here to discuss the marix shifts observed in the systems of matrix-isolated propiolic acids. The matrix effects mentioned in the present work are divided into two categories: (i) the matrix-induced frequency shifts of the different vibrational modes of propiolic acid (especially VQH . "^Q^Q^ ^QQH ^ vc=0» 5o.c=0' and TQOH "T^ocles; (ii) the matrix induced effect on the isotope shifts of PA systems. Tables 6.2 - 6.5 show the vibration frequencies of the stretching, bending, deformation and torsional modes of the -COOH frequencies for HCCCOOH, HCCCOOD, DCCCOOH, DCCCOOD isolated in various matrices (principally Ar and N2). The frequency shifts tabulated above do not include shifts on non-oxygen modes which were noted to show distinct patterns. Such patterns were particularly important in assigning the fundamental modes of DCCCOOH and DCCCOOD. In both (D,H) and (D,D)Jsotopomers isolated in Ar and N2, the higher frequency for modes in the same area have been assigned to the (D,H) isotopomers while the lower frequency is correspondingly associated with the (D,D) isotopomer. Based on the method of preparation of the sample,the statistical distribution favored the (D,D) isotopomer whose absorptions were observed to be more intense. Aside from the frequency shifts for oxygen modes, a multiplet structure was observed for some other modes as well. In Ar / PA+D2^ ®0 spectra, the C=C stretching absorption is observed as a multiplet structure with bands at 2138 cm"'' and 2127 cm"''. This same multiplet stmcture is observed for the mode assigned in this study as the out-of-plane H-C=C deformation. Absorptions are observed to occur at 755 cm"'', and 747.5 cm"'' and another pair at 736.2 and 734.5 cm"''. In corresponding spectra for samples isolated in N2 matrix for the v.|2 mode, absorptions are observed at 760.0 cm"'' and 757.5 cm"''. In N2 / PA-D2 + D2^ ^O spectra the VQ=Q stretch shows a multiplet structure. Multiplet structures can be associated with the occurrence of multiple trapping sites in the matrix. Such effects are particularly more pronounced in N2 matrix spectra which as previously noted show a change in the number of absorption peaks with temperature at which spectra were recorded. 167 Table 6.2: Oxygen-18 Congeners of HCCCOOH

The OH stretching frequencies for oxygen congeners of HCCCOOH isolated in different matrices.

-COOH -C^SOOH -CO^SQH -C''8O^8OH 6jn 6out*

Vapor"*" 3581.0 Ne"'"+ 3580.0 CO** 3438.8 Ar 3550.1 3550.1 3538.0 3538.0 12.1 12.1 N2 3534.0 3534.0 3524.0 3524.0 10.0 10.0

+: Data from Reference 38. *: 6in and 5out "^ean the frequency spacings between the inner two frequencies and outer two frequencies respectively. ++: data from neon matrix obtained by R.L. Redington (unpublished). **: from CO / PA spectra analysis.

Table 6.2 (continued) The carbonyl stretching frequencies (VQ^Q )

-COOH -C1800H -CO^SOH -CISQ^SOH Sjn So^t

Vapor"*" — Ne"*""*- 1759.0 CO** 1746.0 Ar 1750.5 1716.8 1749.0 1716.8 32.2 33.7 1715.8 32.4 32.4 N2 1748.2 1715.8 1748.2 168 Table 6.2 (continued) The nominal CO / COH mode,V5 for HCCCOOH

-COOH -C''S00H -CO''SOH -C18O18OH Sjn 5 out

Vapor"*" 1362.0 Ne+-*- 1301.8 CO** 1342.0 Ar 1302.3 1297.0 1266.9 1262.5 30.1 39.8 N2 1328.4

Table 6.2 (continued) The nominal COH / CO mode, VQ for HCCCOOH

-COOH -CISQOH -CQISQH -CISQ^SOH 5in b^^x

Vapor"*" 1300.0 Ne"*"*- 1149.0 CO** 1178.8 Ar 1158.9 1152.4 1139.5 1134.8 12.9 24.1 N2 1153.0 1147.6 1134.0 1129.5 13.6 23.5 169 Table 6.2 (continued) The ^0-0=0 Bending Frequencies

-COOH -C''S00H -CO''SOH -C18O''8OH 6in 5"ou t

Vapor"*" — Ne"^-*- 651.5 CO** 655.0 Ar 663.0 656.0 662.0 650.0 6.0 13.0 N2 653.0

Table 6.2 (continued) The carboxyl group rocking mode, pC-COOH of HCCCOOH

•COOH -C''S00H -CO^^oH -CI^QI^OH 5jn 5out

Vapor"*" Ne"*""*" 586.0 CO** 586.8 Ar 585.4 580.4 585.4 580.4 5.0 5.0 N2 586.3 170 Table 6.2 (continued) The TQH torsional mode frequencies

-COOH -C''800H -CQI^OH -C''8O''8OH

Vapor"*" 812.0 Ne"^"^ 530.0 CO** 541.0 Ar 533.3 529.0 533.3 529.0 4.3 N2 536.0 528.5 536.0 528.5 7.5

Table 6.2 (continued) The nQ_o rocking mode

-COOH -C''S00H -CQI^OH -C''8O''8OH

Vapor"*" — Ne"^"^ 573.3 CO** 612.5 Ar 571.6 566.0 571.6 566.0 5.6 N2 612.0 608.0 612.0 608.0 4.0 171 Table 6.3: ''^O- Congeners for Fundamentals of DCCCOOH

The VQH stretching mode for DCCCOOH

-COOH -Cl^oOH -CQI^OH -CISQI^OH 5

Vapor CO* 3495.0 Ar 3511.4 3511.4 3501.9 3501.9 9.5 N2 3505.5

Table 6.3 (continued) The C-0 stretching mode for DCCCOOH

-COOH -C^SOOH -CQI^OH -CISQI^OH 5^ 5out

CO 1354.0 Ar 1374.4 1371.2 1361.8 1353.0 9.4 21.4 N2 1405.8 1397.1 1378.5 1351.9 18.6 53.9

': ''^O- frequencies for dimer. 172 Table 6.3 (continued) The COH - bending mode frequencies for DCCCOOH

-COOH -C''800H -C0''80H -C''8o''80H 6in 6 in °out

CO 1125.0

Ar 1132.8 1127.5 1125.5 1119.0 7.3 8.5

No 1127.5 1125.5 1116.9 1111.9 10.6 14.6

Table 6.3 (continued) The 0-0=0 bending mode

-COOH -C^SQOH -CO^SQH -CISQI^OH 6jn SQUI

CO 659.0

Ar 658.5 656.0 652.5 650.5 3.5 8.0

N2 665.0 657.0 665.0 657.0 8.0 8.0 173 Table 6.4: Frequencies for DCCCOOD Oxygen -18 Congeners

The VQD stretching frequencies for DCCCOOD

-COOH -C''800H -C0''80H -C''8o18oH Sin 5 in "out

CO 2588.0 Ar 2591.3 2586.4 2576.2 2572.0 10.2 15.1 N2 2620.5 2602.5 2620.5 2602.5 18.0 18.0

Table 6.4 (continued) The VQ_O stretching frequencies

-COOH -C^SOOH -CO^SQH -C^SO''SOH 6jn SQUI

CO —

Ar 1723.0 1700.3 1717.1 1695.0 16.8 28.0

N2 1700.3 1695.0 1684.0 1679.0 11.0 21.3 174 Table 6.4 (continued) The COD-in plane bending frequencies

•COOH -C''800H -CQI^OH -C18O18OH 5jn 5out

CO 941.3 Ar 934.8 N2 940.0

Table 6.4 (continued) The 5o-C=0 bending frequencies

-COOH -C^SOOH -C0''80H -Cl8o''8oH 6jn 5 out

CO 660.5

Ar 660.5 658.3 656.0 652.2 2.3 8.3

N2 662.2 656.9 655.6 652.4 1.3 9.8 175 Table 6.5: HCCCOOD Oxygen -18 Congeners

The VQD stretching frequencies in different matrices

•COOH -C''800H -CO^SOH -C'l^ol^OH 6in 5 in "out

CO 2577.0

Ar 2624.4 2620.2 2615.4 2604.0 9.0 16.2

No 2622.5 2609.4 2605.0 ' 2600.0 9.4 17.5

Table 6.5 (continued) The VQ_O stretching frequencies

-COOH -C''S00H -CO^^OH -CI^QI^OH 5in SQUI

Ar 1740.8 1718.0 1740.8 1718.0 22.8 22.8

No 1746.3 1731.3 1738.8 1721.3 15.0 17.5 176 Table 6.5 (continued) The VQ.Q stretching frequencies

-COOH -C^SOOH -CQI^OH -C^^OISOH Sin 6 in "out

Ar (1277.3) (1266.9) (1274.3) (1266.9) 7.4 7.4 N2 1275.0

Table 6.5 (continued) The 5QOD bending frequencies

-COOH -C^SoOH -CQISQH -CI^OISOH 5jn 5out

Ar 1004.9 1001.0 990.5 985.3 9.5 19.6 N2 1018.1

Table 6.5 (continued) The So-c=0 bending frequencies

-COOH -C''800H -CO''SOH -ClSolSOH 6in 6out

Ar 639.0 627.5 634.0 624.5 3.0 5.0 N2 666.1 661.9 666.1 661.9 4.2 4.2 177 Table 6.5 (continued) The PC-COOH(D) rocking mode frequencies

-COOH -C''800H -COI8OH -CI80I80H 6in 5 n "out

CO 483.6 Ar 586.0 580.0 586.0 580.0 6.0 6.0 N2 590.0 586.5 590.0 586.5 3.5 3.5

Table 6.5 (continued) The 7^0=0 / "C=0(acid) frequencies

-COOH -Cl^OOH -CQISQH -CISQISOH 6jn 6out

Ar 571.0 566.0 571.0 566.0 5.0 5.0 N2 611.8 607.5 611.8 607.5 4.3 4.3

Table 6.5 (continued) The XQOD torsional Frequencies

-COOH -C^QOOH -CQISQH -CI80I80H 6jn 6out

CO 465.3 Ar 462.0 457.0 462.0 457.0 7.0 7.0 N2 461.9 452.5 461.9 452.5 9.4 9.4 178 Table 6.6 shows matrix shifts observed in the various matrices. The (-) and (+) sign are based on whether a frequency is blue- or red-shifted from the neon matrix value. The largest shifts are observed for the v-| mode where shifts range from 50 cm"''to 141.2 cm"''. The vj, V2, V4and, v 5 modes are red-shifted, while the other fundamentals are blue-shifted. It is also clear that whether blue- or red-shifted, the shifts are all small compared to that of the v-| mode. Generally, shifts are large in the more reactive matrices (CO and N2). Figures 6.4 to 6.8 show shifts for some normal modes (VQH' VC=0> VC-0 ^ ^COH' vc-C' ^OCO' '^COH) P'otted against the polarisabilities (a ) of the different matrices. These figures clearly show the trends of the vibrational shifts that occur in these matrices; the stretching modes, v(C=0) and v (OH) generally have a red shift for (D,H) and (H,H) isotopomers; and a blue shift for the v(C=0) mode for the (D,D) isotopomer; the bending ( 5(C0H) / v(C-O) and torsional mode ( x(COH) has a blue shift for the reactive matrices N2 and CO. In order to explain the observed trends in matrix effects and matrix shifts, it has been assumed that the shifts arise from two principal sources: one is from geometric distortions, and the other is from the force field changes induced by the different matrices. The forces that exist between the propiolic acid and its environment are most likely not strong enough to distort the propiolic acid molecules to an appreciable extent. Thus, it has been assumed that contributions from geometric distortions will be less important in explaining the observed frequency shifts than the force field perturbation and other previously mentioned intermolecular interactions. As a first approximation the force field perturbation can be expressed as a function of the polarisability. The vibrational secular equation can be separated to incorporate the two possible sources, with geometric distortions featured in the G matrix and force field perturbation in the F matrix (Equations 6.12 and 6.13).

GC^) = G(V) + AG (6.12)

F(M)= F(V) + 1/2CXAF. (6.13) 179 Table 6.6: Matrix Shifts for the HCCCOOH Fundamentals in Various Matrices

Symmetry / Description Ne Ar N2 CO

A'-block v (0-H) 3580.0 50.0 60.0 141.2 v(C-H) 3325.0 11.0 16.0 16.2 V(CHC) 2139.5 -0.5 1.5 - 0.1 v (0=0) 1759.0 5.0 9.5 12.8 v(C-O) 1301.0 -2.0 -27.4 -40.0 6(C0H) 1147.2 - 3.8 - 5.8 -24.1 v(C-C) 815.2 - 4.8 - 5.8 -11.1 5(H-C=C) 690.6 - 4.4 -17.4 -8.4 5( 0=0-0) 650.0(1.5) -2.0 (.5) -16.5 -7.9 PC-COOH 586.0 - 1.5 - 2.0 -0.8 ** 5(C=C-C) 189.0 N/A N/A N/A

A" - block 7r( H-C=C) 754.6 -4.4 - 4.4 -5.4 7l( 0=0) 573.3 1.4 -33.7 -39.2 T(COH) 530.0 0.5 -3.8 -11.0 7c(C=C-C) 229.0 N/A N/A N/A

*: Shifts are with reference to Neon value (A = V|sjg - V(^ ). **: N / A , from HCCC(0)F liquid Raman spectra. 180

+ DCCCOOH 0 HCCCOOH

Figure 6.4. Correlation of VQH Frequencies with Polarizabilities of Matrices. 181

DCCCOOH A DCCCOOB HCCCOOH 20 ,C0 .N2

16 -Ar AQ

12 ro o 8

-Ne

1600 1700 Id _L_ JO CMC-I)

Figure 6.5. Correlation of VQ_O Frequencies with Polarizabilities of Matrices. 182

HCQCOOH(^COH^ HCCCOOH (SQ^Q)

20 \

03 \ 16 \ \

12

•< O a

4-

550 625 700 _l L J 1_ J 1 1 CM (-1)

Figure 6.6. Correlation of SQCO ^"^ "^OH 'frequencies with the Polarizabilities of the Matrices. 183

DCCCOOH H:-0 (DCCCOOH) I HCCCOOH DCCCOOD 'COH

20 _ \ CO a ii

16 LAr CI

12 _

Ss

.Ne

"

iigo 1200 1300 CMH)

Rgure 6.7. Correlation of VQ.Q I SQQH Frequencies with the Polarizabilities of the Matrices. 184

DCCCOOH

20 HCCCOOHi' CO C) y Q C) y 16 UAr y 0

J2-

§8

_Ne

800 825 850 _! I I CM(-I)

Rgure 6.8. Correlation of vc.c Frequencies with the Polarizabilities of Matrices. 185 The Wilson FG formation [83] is then applied to obtain

G(M)P(M)L(M)^L('^)^(M) jg^^j

The resulting frequencies obtained will be incorporated in A, in the form X\^ ) = Xi(v) + ^X\. Substituting appropriately for the various terms in Equation 6.14 yields

G(M)F(M) = (G(V) + AG)(F(V) + i/2aAF)

= (G(V).F(V) + AG.F(V) + 1/2aG(V). AF -h 1 /2cxAG . AF. (6.15)

If it is assumed that LC^) = L(V), and that the term AG . AF is negligible, the secular equation becomes

I AG . F(V) + 1 / 2aG(V) • AF - A>.E| =0. (6.16)

One can then view AG . F(^) -I- AF . G(^) as the perturbation arising from the previously mentioned sources. In principle, observed matrix shifts can be used to calculate AF by an iterative procedure. The straight line correlations obtained for the data in figures 6.4 to 6.8 can be interpreted in part as expressing the linear dependence assumed between the force field matrix and the polarisability implied in the development of Equation (6.16). The only large deviation from linearity in a for the stretching data occurs for VQH '" the CO matrix. CO has a permanent dipole moment. The bending modes form linearity in a roughly according to the matrix quadmpole moments. The internal coordinates associated with small force constants such as SQQH ^rid ^QQQ tend to be distorted relatively easier than VQ_Q. This difference in behavior can be explained using the "cage" formalism proposed by Pimentel and Charies [128]. This means that bending, rocking, and low frequency stretching tend to act as though the molecule is trapped in a tight cage yielding positive shifts. On the other hand, high force constant coordinates such as VQH- VQU, VQ=Q, and VQ^Q act as if the molecule is in a "loose" cage resulting 186 in negative shifts. For propiolic acid, it is clear that the points of contact between the molecule and the matrix surrounding are the high frequency coordinates VQH. "^0=0 ' VH-C=C-C ( Figure 6.9). The matrix-induced effect on isotopic shifts, is summarised in Tables 6.2 - 6.5 by *in and *out (the frequency spacings between oxygen 16-16 and oxygen 18 -18 congeners). They do not vary much for the different matrices. The frequency spacings are collected for the various modes in Ar and N2 (principally) and CO in Tables 6.2 - 6.5. In summary, it would appear that the normal modes of PA and PA-D2 congeners are not particularly sensitive to the environment due to the different matrices, although the frequencies themselves are sensitive to the matrix materials. Table 6.6 summarizes somewhat the dependence of the frequencies on the matrix materials. Tables 6.7 - 6.11 below list Av = V|_| - VQ isotope shifts of the observed frequencies for the fundamental modes of the pairs HCCCOOH and DCCCOOD, HCCCOOH and DCCCOOH, HCCCOOH and HCCCOOD, DCCCOOH and DCCCOOD and HCCCOOD and DCCCOOD. Usually, comparisons are made between gas phase and matrix data. Gas phase spectral data is available only for the normal propiolic acid monomer [38]. The comparisons that have been made under the present circumstances are meant to provide possible predictions from the product rule. It is clear that very large shifts are observed for all three matrices for the HCCCOOH / DCCCOOD, HCCCCOOH / HCCCOOD and DCCCOOH / DCCCOOD pairs. This can be predicted from the product rule. V2 also shows very large shifts for the HCCCOOH / DCCCOOD and HCCCOOH / DCCCOOH pairs. At least for V2 , these shifts could also have been predicted from the product mle since this mode is directly involved in the H / D isotope effect. In general all the matrices show comparable shifts for any particular mode for the pair in question.

Conclusions. The frequency shifts for several modes of propiolic acids isolated in different matrices were collected and analyzed. These were correlated as a 187

Figure 6.9. Cage Model for Propiolic Acid Showing Idealized Spherical Cage with Neighboring Matrix Atoms (M). 188 Table 6.7: Comparison of Matrix Isotope Effects between HCCCOOH and DCCCOOD (Vj-iccCOOH " VDCCCOOD)

Ar CO N'

VI 1189.4 1092.8 1172.0 V2 656.0 680.8 692.0

V3 162.5 50.6 163.8 V4 161.8 147.4 154.0 -78.4 V5 -100.0 -65.3 223/219.0 V6 218.7 152.1 -30.0 V7 -30.8 -24.7 96.0 V8 82.5 87.0 6.6 V9 5.4 -1.1 103.2 103.4 Vio 103.9 10.0 Vil 10.0 10.0 150.5 147.0 vi2 146.5 -6.1 -9.6 V13 -46.5 75.9 75.1 V14 112.5 12.0 12.0 V15 12.0 189 Table 6.8: Comparison of Matrix Isotope Effects between HCCCOOH and DCCCOOH (VHQQQOOH " VDCCCOOH)

Ar CO _ N2

V1 13.9 -25.2 -30.0

V2 656.0 665.8 689.0

V3 154.6 46.6 156.0 V4 31.0 146.2 132.0

V5 -44.1 -12.3 -25.3 V6 35.1 -32.0 25.8 V7 -5.0 -20.7 -27.0 96.0 V8 82.5 87.0 6.6 V9 -1.6 0.4 -0.7 0.5 Vio 0.0 V11 10.0 10.0 10.0 151.0 147.0 V12 146.5 -37.7 -41.2 V13 -78.3 0.0 0.0 V14 0.0 12.0 12.0 vi5 12.0 1 90 nrrr^n9°"^P^'''^°" °^ '^^^'''^ '^°^°P® ^^^^^ts between DCCCOOH and DOGCOOD (VDCCCOOH " VDCCCOOD)

Ar CO N2

vi 1175.0 1118.0 1202.0 V2 0.0 15.0 3.0 V3 7.9 4.0 7.8 V4 30.8 0.2 22.0

V5 -51.9 -53.0 -53.1

V6 183.6 184.0 187.2

V7 -25.8 -4.0 -3.0

V8 0.0 0.0 0.0

V9 7.0 -1.5 0.0

Vio 103.9 103.9 103.9

Vil 0.0 0.0 0.0

V12 0.0 0.0 0.0

V13 31.6 31.6 31.6

Vl4 112.5 75.9 75.1

VI5 0.0 0.0 0.0 191 Table 6.10: Comparison of Matrix Isotope Effects between HCCCOOH and HCCCOOD (VHQQQOOH " VHCCCOOD)

Ar N'

V1 928.0 913.0

V2 2.0 2.0

V3 10.0 3.0

V4 12.0 5.0

V5 37.0 49.5 161(143.0) V6 146(140.0) 1.0 V7 0.0 0.0 V8 2.0 4.0 V9 2.6 0.0 Vio 1.0 Vil 0.0 0.0 2.0 V12 3.0 102.0 VI3 64.5 72.0 vi4 62.9 0.0 V15 0.0 192 Table 6.11: Comparison of Matrix Isotope Effects between HCCCOOD and DCCCOOD ( VHQQQOOD " VDCCCOOD)

Ar N2

VI 261.4 259.0

V2 654.0 690.0

V3 152.5 160.8

V4 149.8 149.0

V5 -127.0 -127.9

V6 81.7 72.0(66.0)

V7 -30.8 -31.0

V8 80.5 96.0 2.6 V9 2.8 104.4 Vio 102.9

Vil 10.0 10.0 145.0 VI2 143.5 -111.2 -111.6 V13 3.1 VI4 49.6 12.0 12.0 V15 193 function of the polarisabilities of Ar, CO, N2 and Ne. Table 6.6 shows that in most cases, the neon matrix behaves similarly to the more studied matrices (Ar, CO and N2). Generally, stretching modes showed red shifts while the bending modes showed blue shifts. The effect of the matrix induced shifts have been explained as the result in perturbation of the force field of the molecule. The "cage" formalism of Pimentel and Charies can be used to explain an across the board behavior of the various modes. Multiple band absorptions observed for some modes can be explained either as the result of the presence of multiple trapping sites or aggregation. In the present study, temperature dependent studies have been used to determine that the presence of multiple trapping sites is responsible for the multiplet stmcture of some modes. Dimer formation, as shown in spectra of concentrated samples, has been shown to be a contributing factor to the occurrence of multiplet structure for some modes. Figures 6.4 to 6.8 show a distinct pattern. The shifts observed for the various modes depend on whether or not the matrix is reactive (CO and N2) or non reactive (Ar and Ne). This is shown in the plots by the dashed lines. CHAPTER VII THEORETICAL BACKGROUND TO AB-INITIO CALCULATIONS FOR PROPIOLIC ACID

The ab initio MO calculations carried out for propiolic acid can be divided into the following subheadings: total energy evaluations, geometry optimizations, force constants determination and frequency calculations, one electron properties (electrostatic properties), torsional barrier calculations, molecular reaction dynamics - unimolecular decomposition channels for propiolic acid, effects of electron correlation, and configuration interaction calculations. The fundamental ideas that govern the calculations cited above are discussed below. The calculations overall are based on the Hartree-Fock Self Consistent Field method for open and closed shell systems as programmed in the GAUSSIAN 82 program by Pople et al. [69].

Total Energy Evaluation The exact Schrodinger Equation for a molecular system can be written as:

A H^(r,R) = EH'(r,R) (7.1) where E is the total internal energy of the molecule. The wavefunction 4'(r,R) depends on both the electronic coordinates r and the nuclear coordinates R. H is the nonrelativistic molecular Hamiltonian operator, which when written for a molecule with N nuclei and n electrons takes the form:

AN n n N N N H = -1/2IVk2/ Mk-1/2IV^2.^ XZk/r^k-^I Z^Z, / r^i + 11 / r^^ l^_1 |x=1 ^L=1k=1 k

(7.2)

194 195 All coordinates used in the system are based on the center of mass system. The first term represents the kinetic energy of the nuclei with M^ the mass of the k^^ nucleus expressed in atomic units. The second term is the electrons' kinetic energy. The third term is the electron-nucleus-attractive potential energy while the fourth term represents the nuclear-nuclear repulsive potential energy. The fifth term represents the electron-electron repulsive potential energy. Compared to atoms, there are several components to the total molecular energy. Also molecules have various motions all coupled in a rather delicate fashion. Thus, the evaluation of the total molecular energy is difficult to achieve. To simplify this evaluation, certain assumptions have to be made. The most important of these assumptions is the Born-Oppenheimer Approximation. The Born-Oppenheimer Approximation permits the solutions of the Schrodinger Equation to be expanded in a power series in W^^^. M is the total mass of the system. The approximation is based on the fact nuclei are more massive than electrons and they move a lot slower. In this regard, electrons can be considered as being in quasi-stationary states during the course of nuclear vibrations. The Born - Oppenheimer Approximation allows the breakdown of the total wavefunction into two parts:

4'(r,R) = ^R('')'2'('^) ^^-^^

^p{(r) is the electronic wave function for fixed nuclear positions. 0(R) is the nuclear wavefunction. The molecular Hamiltonian, on this basis, can be rewritten in the form:

H = TR-Fh + V (7-4)

where TR is the kinetic energy operator of the nuclei, h is a sum of operators each of the form

h^ = -V^2/2-l Zk/r^k (7.5) 196 and V is the potential-energy operator for nuclear-nuclear repulsions and electron-electron repulsions. The operator (h + V) is called the electronic Hamiltonian. The molecular Schrodinger Equation is rewritten as

(h + V)VFR(r) = E(R)H'R(r) (7.6)

ER is usually referred to as the molecular energy in the fixed-nuclei approximation. It is this energy that is usually evaluated in ab initio programs given a convenient choice of wavefunction (basis set). The potential energy curve for a diatomic is obtained by plotting E(R) versus the internuclear distance R. For polyatomic molecules, one would expect more complicated potential energy surfaces in view of the more complicated functional dependence of energy on molecular geometry.

Basis Sets [129] In using ab initio programs, a choice of basis sets has to be made. While the quantum-mechanical procedure defines the level of theory for which calculations are being carried out, basis sets define the model for which the theory is being applied. This defines the molecular orbitals which describe the electronic structure of molecules. These functions are usually multicentered and delocalized over the entire molecule. The analytic molecular orbitals are approximated by a linear combination of atomic orbitals. Such combinations are defined by a set of standard coefficients and exponents for stored basis sets or can be read in with the input for a nonstandard basis. In the present study, all calculations were performed using stored (standard) basis sets. Once the choice of basis set is indicated, the MO calculation then involves finding the combinations of the atomic orbitals that have proper symmetries and give the lowest (most negative) electronic energy. The procedures including other theoretical aspects involved can be found in texts dealing with computational chemistry [129-131]. 197 The choice of basis sets or choices of atomic orbitals almost always employed by ab initio methods are the Gaussian type orbitals (GTO's). In these bases, each atomic orbital is made up of a number of Gaussian probability functions. The advantage in using these is that they are faster computationally when employed in the evaluation of one- and two-electron integrals. Several optional GTO basis sets of varying-size are included with the program. Basis set selection for MO calculations was reviewed recently by Davidson and Feller [132]. There are several types of optional basis sets: i) STO-nG: These are the simplest of the optional basis sets. STO-nG stands for Slater Type Orbitals simulated by n Gaussian functions each. Each atomic orbital consists of gaussian functions added together. The coefficients of the Gaussian functions are selected so as to give as good a fit as possible to the corresponding Slater Type Orbitals. The most commonly used STO-nG is the ST0-3G. It is referred to as a "minimal basis" set. This definition implies that it has only as many orbitals as are necessary to accommodate the electrons of the neutral atom. The ST0-3G basis set is also very economical. ll) Split-valence or double zeta basis set: These provide a solution to some of the weaknesses and problems encountered in minimal basis sets. In these bases atomic orbitals are split into two parts, an inner-compact orbital; and an outer-more diffuse one. Coefficients for the two parts can be varied independently during construction of the molecular orbitals in the SCF-procedure. In this way, the size of the atomic orbital that contributes to the molecular orbital can be varied within the limits set by the inner and outer basis functions. While only the valence orbitals are involved in this case of split-valence basis sets, double-zeta basis sets involve core orbitals too. Double-zeta implies two different exponents. The most commonly used basis set in this category is the 4-31G. The nomenclature means that the core orbitals consist of 4 and the inner and outer valence orbitals of 3 and 1 Gaussian functions, respectively. Calculations using this procedure have been carried out in the present study. Basis sets involved are the 3-21G, 4-31G and the 6-31G. 198 III) Polarization basis sets: These represent an improvement because of the addition of d-orbitals. Their purpose is to allow a shift of the center away from the nucleus. The most commonly used standard basis set in this category is the 6-31G*. It has been used for some calculations in the present work. It uses six primitive Gaussians for the core orbitals, a three/one split for the s- and p-orbitals, and a single set of six d-functions *indicated by an asterik). Basis sets involving further development of the 6-31G*, although available, were not used in this study because of the amount of computer time required. Finally, the choice of basis set is important for two reasons. First, the number of basis functions the program can handle is limited. Secondly, the computer time required is roughly proportional to the fourth power of the number of basis functions. Generally, the number of basis functions rises rapidly with increasing sophistication of the basis set.

Evaluation of the Total Molecular Energy [130] The total molecular energy is determined by solving the molecular Schrodinger Equation. The wavefunction used in the process is simplified by the application of the Born-Oppenheimer approximation. The electronic Hamiltonian for a system with 2M electrons may be written as

A A 2M M(2M-1) H° = H°(1,2,. . ., 2M) = I h^ + 1 q^v (") }i^1 |i>v

To determine the energy, we solve the expectation value equation on substituting for^° and H; that is,

E = <^°|H|^°>

On integrating out the spin variables, the total molecular energy has the form 199 M MM E = 2Ih0pp + 11 (2J0pq - K0pq) (7.9) p p q where Jpq and Kpq symbolize the Coulomb and Exchange integrals respectively. The superscript 0 indicates that these matrix representations are over the MO basis. In more advanced theory, the expressions above can be manipulated to obtain the so-called Fock operator; F. Solving the Schrodinger Equation is then reduced to solving equations of the form:

'^^p = '^p^p- (7.10)

The elements of the diagonal matrix e are the molecular orbital energies and their sum gives the total molecular orbital energy;

M N N E = 2 lej - Z S pjj(2Jjj - Kjj) = Zep. (7.11) 1=1 i=1j=1

The difficulty of calculating the operator required in the Hartree-Fock equation is overcome by solving the eigen-problem equation in an iterative manner. The process used is referred to as the Self Consistent Field (SCF) method. A description of the SCF method is outside the scope of this study but a flow chart description can be found in Appendix C [129].

Geometn/ Optimization [130] One of the most important applications of the single determinant LCAO SCF MO theory is the determination of molecular structures. Although the geometries of many stable molecules are well known from experimental studies, to obtain consistent theoretical results it has become commonplace to carry out geometry optimizations. The results obtained are basis set dependent and only in the limit of very large basis sets does the calculated geometry closely approach the experimental geometry. 200 To every geometry, defined by a set of internal coordinates {Rj}, there is an associated energy value E(R).

E(R) = E(Ri,..., R3N.6) (7.12)

There is a minimum energy value associated with the optimum geometry.

Eopt. = E(Rl°P^' R2°P^ • • •. R3N-6°P^) (7-13)

The search for the optimum geometry, that is, the choice of {Rj°P^} from an infinite number of sets of {Rj}, is termed geometry optimization. For a diatomic molecule, the problem is simple, but it becomes extremely complicated for polyatomics and E(R) has to be expressed as a quadratic function of a series of 3N-6 variables measured from the minimum:

(Rl - Rl°P^). (R2 " R2°P^). • • • . (R3N-6 - R3N-6°P^) (7-14)

The potential approximated as a quadratic function close enough to the minimum can be rewritten in the generalized vector form as:

E(R) = EoPt + (R - ROP^)A(R • R°PV (7-15)

The matrix elements of A are also related to force constants. Each of the diagonal elements is associated with the force constant of one mode only while the off-diagonal elements are related to the interaction force constants (Equation. 7.16 and 7.17) d2E(R)/a(Rj-RiOP^)2 = 2ajj = kjj (7.16) a2E(R)/9(Ri-RjOP^)a(Rj-Rj°P^) = 2aij = kjj (7.17)

In a geometrty optimization, E°P^ and all the different elements of matrix A are determined so that E(R) of Equation (7.16) if fitted to the true hyper- 201 surface, E(R), that is, equation 7.13. For a fit, one needs to compute a minimum of 1 + 1/2(3N-6)2 + l/2(3N-6) points on the energy hypersurface "close enough" to the minimum.

Force Constant Evaluation and Vibrational Analysis [130] Expressions have already been written for force constants in the previous section. A force constant Fjj is defined as the second partial derivative of the molecular energy E(R) with respect to the nuclear coordinates Rj and R: at the equilibrium configuration:

Fjj = (a2E(R)/aRjaRj)o - (7.18)

Force constants can be calculated in three ways which differ in the manner in which the differentiations in (7.12) are carried out: (A) Numerically twice; (B) Analytically twice; (C) First analytically and then numerically. The procedures and other aspects of the various methods have been discussed by Pulay [133] and by Hess et al. [134]. Once the force constants have been obtained, they may be transformed by Wilson's FG matrix formalism to obtain fundamental vibrational frequencies. In the present work, method (C) above has been used to obtain force constants and the FG matrix formalism applied to obtain fundamental frequencies. Method (C) involves two steps: first, the energy is differentiated analytically to obtain the force acting in the direction of Rj, that is, fj=-aE(R)/aRj (7.19)

The equilibrium geometry is determined by allowing the nuclear coordinates to relax until the net forces on the atoms vanish. Secondly, the negative derivative of f; with respect to Rj is then determined numerically, carrying out calculations for different Rj's near the equilibrium configuration. Method (C) is often referred to as the force method.

Fjj = -(afj/aRj) (7.20) 202 Since the energy expression is

E = <0|H|cD> (7 21) and assuming F is real, then fj = -aE/aRj = - - 2 (7.22)

The first term in (7.22) is the Hellmann-Feynmann force. Since aH/aRj is a one - electron operator, it can be calculated easily. The second term is the wavefunction force (WFF) and it is a lot more difficult to evaluate. This force depends on the choice of basis set as this will determine the nature of the wavefunction to be used. The results that will be presented are based on internal valence coordinates. The force constants calculated fall into five classes: i) diagonal stretching terms F^ ii) interactions between stretching F^^, III) interactions between a stretching and bending F^^^, iv) diagonal bending terms F^^, (in-plane and out of plane), v) interactions between two bending F^^Q^- (in-plane and out of plane sets). The force constants obtained from ab initio calculations have been used as plausible guesses in the determination of the experimental force field for propiolic acid. The vibrational frequencies will be used as an aid in the assignment of experimental data and also in the calculation of fundamental frequencies for other isotopomers. Bowman recently reviewed the SCF approach to polyatomic vibrations [135].

Fvalnation of One-Electron Properties [130] These are computed from one-electron operators (P) and molecular wavefunctions. The molecular wavefunction is as previously discussed a single Slater determinant:

A 0 = A[0i(1)a(1)0i(2)p(2). . . 0M(2M-1)a(2M-1)cDM(2M)(3(2M)] (7.23) 203 The explicit form of the one electron operator is:

A A A Q = Q° + «i (7.24) i

The summation in Equation (7.24) is over all the electrons. In order to evaluate these electrostatic properties, the expression for the expectation value of a quantum mechanical observable Equation (7.25) is used.

Q =«D|Q|cD> (7 25)

Substituting appropriately in (7.24) gives

M Q=Q°-h2 X

where the summation is over all occupied MO's and the factor 2 is for the double occupancy of each MO. In the present study, the dipole moment, second moment, third moment, fourth moment, electric field potential and electric field gradient have been evaluated for propiolic acid conformers using a variety of basis sets. The corresponding one-electron properties that can be calculated are summarized in Table 7.1. It should be noted that multiple moments are usually computed relative to the center of mass of molecules. Generally, one electron properties computed at the HF limit are expected to nearly reproduce their corresponding experimental values. Whether or not the magnitudes are basis set dependent will be ascertained. 204 Table 7.1: One - Electron Properties

Property Symbol(^) Operator available Nuclear Electronic

Dipole moment MN) V V Second moment Qa(3(C) V V

Third moment RapT^C) V V

Quadrupole moment 0ap(C) V V

Octupole moment ^aP7(^) V V Potential V(N) V V

Electric field Ea(N) V V

Electric field gradient PaaC^) V V Diamagnetic susceptibility XAv(C) - V Diamagnetic shielding aAv(N) ^f

(a) The complete table can be found in [130]. 205 Unimolecular Rp;^rtivities-CalnNlatinn of the Potential Energy .S^'rff^Ce A chemical change can be fully characterized by analyzing its energy hypersurface. This energy hypersurface is the total molecular energy expressed as a function of 3N-6 independent variables:

E(R) = E(Ri,R2 R3N.6) (7.27) where N is the number of constituent atoms. Ideally, we can describe important points on the surface corresponding to I (initial state or reactant), M (an intermediate) and F (the final state or products). The inter - relationship between any two of the minima represents a thermodynamics problem and the energy separation between any of the pairs measures the thermodynamic stability of the species of the pair against the other species of the same pair. In addition to the thermodynamic analysis of a chemical change, a kinetic analysis of the same problem can be carried out. This requires some knowledge (geometry and energy) about the transition states as implied in the scheme:

I => TS1 <=> M TS2 => F

Therefore the generation of a hypersurface would give all the information about a chemical system that undergoes a conformational or reactive unimolecular process. In line with the above discussion, a few aspects of the hypersurface determination have been studied. This include: the torsional potential for converting propiolic acid (cis) to propiolic acid (trans); and then transition states and minimum energy structures on the potential energy surface for the probable unimolecular decomposition channels. In general total energies are computed as a function of molecular geometries. Besides the minima there are other important points on the surface, particularly the saddle point associated with the transition state of a conformational change. Information derived from such an analysis leads us to theoretical 206 stereochemistry. The various points can be characterized in terms of the force constant with respect to some coordinate Rj. For a minimum: (a2E(R) / aPj^jRj = Rjmin = fjj > 0 1 < i <(3N-6) for all coordinates. For a saddle point: (a2E(R) / aRj2)Rj = Rjsad = fj j> 0 1 < i < (3N-7) for all coordinates except one-which may be taken to be the last coordinate. (a2E(R) / aRj2)Ri=Rjsad = fjj < 0 I = 3N-6 the force constant is negative corresponding to an imaginary vibrational frequency. The appearance of a saddle point is either a minimum or a maximum and depends on the chosen cross-section of the hypersurface. It is important to remember that the position of a minimum or saddle point is dependent upon the level of theory employed. Effects of improved basis set and / or correlation levels may be extreme and may even result in changes of the qualitative conclusions.

Electron Correlation and Configuration Interaction [131] Electron correlation effect methods are applied in order to improve the theoretical model chosen for description. While the basic solution of the Schrodinger deals with problems in which electrons take only one configuration, the methods of electron correlation assume a more accurate wavefunction in which contributions from other possible electronic configurations are included. The more accurate wavefunction takes the form:

"¥ = ao^o + ^I'^l+^l'^Z + ^n'^n (7-28)

where ^j represent wavefunctions for all the possible configurations and the linear coefficients, aj represent the contributions of each configuration. These coefficients are calculated in the process. Within the framework of a given basis set, inclusion of wavefunctions for all possible electronic configurations is termed "full configuration interaction." 207 Theoretical models may be improved in two ways: improvement of the basis set and improvement of the correlation technique. In the present study attempts have been made to use both methods. Particular attention has been paid to the effects of such improvements on energies, dipole moments, electrostatic properties, and vibrational frequencies. Two correlation methods have been used : CID(Configuration Interaction with Double Substitutions), and MPp (Moller-Plesset Perturbation Method). Moller- Plesset models are formulated by introducing a generalized electronic Hamiltonian H;^.

H^ = Ho + Xy (7.29)

in which the perturbation , Xy is defined by

xy = >.(H - HQ) (7.30)

H is the correct Hamiltonian and >. is a dimensionless parameter. In MP theory, the zero-order Hamiltonian, HQ is taken to be the sum of the one-electron Fock operators. ^;^ and E;^, the exact ground-state wavefunction (within a given basis set) and energy for a system described by H;^, may be expanded in powers of X:

WX = 4^(0) +X'¥i^) + ?i2^(2) +

E;^ = E(0) + >£(^) + ^2E(2) + (7.31)

Practical correlation methods are formulated by setting the parameter X=^, and by truncation of the series in Equation (7.31) to give various orders. In Equation (7.31), 208 ^( 0) = XJ/Q (VJ/Q = Hartree - Fock wavefunction).

E(0) = lej (ej = one-electron energies) (7.32)

Therefore, the MP energy to first order is the Hartree-Fock energy. The fundamental concepts discussed here can be generalized. The work presented in this dissertation used the ab initio program GAUSSIAN 82 [70]. The logical structure of a typical ab initio program is given in Figure 7.1 and the executional sequence in a calculation for some procedures can be found in [131]. Results for calculations performed are reported and discussed in a subsequent chapter. While no attempt has been made to follow succintly the order in which the basic theory has been presented, results presented fall in three general areas: geometries and vibrational analysis including force fields; one electron properties and variation with basis sets, and molecular reaction dynamics including conformational changes. On a lighter note, the theory and methods of MINDO / 3 have been recently reviewed by Lewis [136 ]. MINDO / 3 is a semiempirical self-consistent-field (SCF) molecular-orbital method (MO) approach to calculating electronic structures. The acronym MINDO / 3 stands for the Third Version of the Modified Intermediate Neglect of Differential Overiap. In view of its simplicity, it is a more attractive program to organic and medicinal chemists. It is an all-valence electron method. In this method, an approximation for the core Hamiltonian is maintained while the valence electrons are assigned a minimum basis set atomic orbital (AO) wavefunctions 0j. Energies, geometry, dipole moments, rotational constants, moments of inertia, force constants and normal coordinates have been calculated using this technique. 209

Specification of molecular geometry

Calculation of symmetry infonmation

Specification of basis set

Calculation of initial guess

iJolution of SCF equations CONTROL / MA<^<^ PROGRAM 1 \JSTORA 1 V^l li^VGJ Population analysis

Calculation of energy gradient

Calculation of correlation energy

Plotting of molecular orbitals and / or electron density

Figure 7.1. Overall Logical Structure of a Typical Ab Initio MO Program. Program modules are indicated by rectangles, mass storage by circles. There is two-way communication between the individual modules and the control program or mass storage files. The sequence of execution of individual modules is dictated by user input and is controlled by a single master program labeled control program [131]. CHAPTER VIII MOLECULAR ORBITAL STUDIES OF PROPIOLIC ACID

The previous chapters have dealt with obtaining and interpreting spectra of propiolic acid isolated in different solid matrices. It has been noted that one of the difficulties associated with obtaining gas-phase spectra for propiolic acid is the fact that it decomposes easily to acetylene and carbon dioxide. Unimolecular reaction channels that are conceivable for the propiolic acid monomer include rotational isomerization, intramolecular proton exchanges and several dissociation reactions. In view of the amount of work that has been done on related molecules, it is surprising that little is known experimentally or theoretically about propiolic acid. The only other known studies on propiolic acid are the spectroscopic studies of Katon and McDevitt [38] and the MO calculations of Furet et al. [62]. Lister and Tyler [39] and Wellington Davis and Gerry [43] interpreted their microwave data based on Structure I, the intramolecular hydrogen bonded conformer with Cg symmetry (Figure 8.1). The character of presumably higher energy conformers, e.g.. Structure II, and the nature of conformational transitions are not known. This section of this dissertation addresses some of these properties using molecular orbital theory with minimal, intermediate and relatively large basis sets. In addition, unimolecular dissociation of the molecule is addressed since propiolic acid is unstable and no information is available experimentally or otherwise about the energetics of these decompositions. To date the single ab intio calculation previously cited [62] provided energies of the weakly intramolecular hydrogen-bonded conformer (cis propiolic acid) at the ST0-3G, 4-31G, 6-31G, 6-31G*, and 6-31G** basis set level. The absence of an experimental force field based on a normal coordinate analysis prompted the need for some of the work done here. We have used such results as an aid in assigning experimental frequencies. In the present chapter, the results of semi - empirical (MINDO / 3) and ab initio (GAUSSIAN 82) calculations carried out with a variety of basis sets are

210 21 1

Figure 8.1. The Two Conformers of Propiolic Acid. 212 presented. In the first part of the discussion, results of MINDO/3 and GAUSSIAN 82 computations have been combined to obtain information on the conformational aspects of propiolic acid. Such a picture has been discussed based on energies, geometries, rotational constants, dipole moments, and moments of inertia. The calculations were performed to provide reasonable theoretical estimates of the various molecular parameters. A quantum mechanical normal coordinate analysis has been carried out to identify a suitable initial estimate for the experimental force field and to help ensure that experimental frequencies have been correctly assigned. Fundamental, though, is the question of how well HF-SCF MO methods reproduce experimentally determined parameters. This study is concluded with calculations meant to provide reasonable theoretical estimates for the geometries and energies of the more general conformations of propiolic acid, and the energetics of decomposition channels. The results bear strongly on the interpretations of Structure (I) as the most abundant conformer existing at ordinary temperatures.

Computational Details For the MINDO / 3 calculation, the standard basis set (1 s for H, 1 s, 2s, 2p's for C, O) was used. The Fortran IV version of the program [137] was mn on an IBM 3033 computer with the assistance of Dr. Raghuveer. The GAUSSIAN 82 code with its standard optimization routines for a VAX II / 780 computer were used for all calculations. The runs utilized the builtinST0-3G, 3-210,3-21 G*, 4-310,6-31 G, 6-31G* and 6-31G** basis sets. Full and partial geometry optimizations have been carried out to determine geometries, energies, dipole moments and'the other molecular parameters. Single point calculations were carried out with some basis sets for both conformers and at other critical points to assess the effects of electron correlation and configuration interaction on relative energies and other molecular properties. Electron correlation effects were incorporated at the second (MP2) and third (MP3) levels of the Moeller-Plesset (MP) perturbation theory with the 6-31G and 6-31G* basis sets and the corresponding optimized geometries. 213 Results The calculations that were performed have been divided into two groups so as to facilitate the presentation and discussion of the results. The first part of the results deals with results meant to establish which of the two conformers is more stable. In addition to geometries, energies and moments of inertia, a calculation of the electrostatic properties has been carried out to provide more evidence for the computed structures of the two conformers. Also included in Section A are the results of a quantum mechanical normal coordinate analysis. Having established the geometry of propiolic acid conformers, the second part. Section B deals with the molecular reactivities of the two conformers in the SQ and Ti electronic states on the propiolic acid hypersurface.

A: Conformational Studies of Propiolic Acid Equilibrium Geometries and Quantum Mechanical Model Structure Tables 8.1 and 8.2 show calculated equilibrium geometries for structures (I) and (II). It is clear that the inclusion of polarization functions into the basis sets causes a revision of the angles around the atoms with lone pairs of electrons. It is also clear that this causes the skeletal bonding to become too short. Comparisons between ab initio and MINDO / 3 geometry relaxational effects show that the MINDO / 3 method fails to predict correctly the lone pair effects in bonded interactions. The case in point here is the 0-H bond which is underestimated by about 22%. Tables 8.1 and 8.2 also show that in most cases the bond lengths are underestimated by about 0.01 -0.04A°. MINDO / 3 also overestimates the COH bond angle by about 19%. But the overall agreement is good with bond lengths being predicted to within ± 0.04A°, and bond angles to within ± 3° on the average. It can also be seen that improvement to double zeta or split valence basis sets leads to some improvement in bond lengths and deficiences in bond angles. On the other hand, the addition of polarization functions to split valence bases improves predictions about angles and the general features of a model such as HF/6-31G** appears to be close to the HF limit [138]. 214 Table 8.1: Geometry of Cis Propiolic Acid, Structure (I)

Molecxjiar Parameters MO Calculated Parameters

a MW MINDO/3 ST0-3G 3-21G 3-21G* 4-31G 6-31G 6-31G* 6-31G"

BoncJ lengths R(O-H) 0.972 0.758 0.991 0.969 0.969 0.956 0.955 0.953 0.949 R(C-O) 1.343 1.330 1.391 1.351 1.350 1.344 1.347 1.323 1.322 R(C=0) 1.202 1.211 1.219 1.201 1.201 1.206 1.210 1.184 1.184 R(C-C) 1.445 1.447 1.481 1.437 1.438 1.432 1.436 1.456 1.457 R(C^) 1.209 1.206 1.173 1.185 1.185 1.187 1.191 1.184 1.185 R(H-C) 1.055 1.073 1.067 1.051 1.051 1.052 1.054 1.058 1.058

Bond angles

-E(HF) 260.93744 262.9499 262.94991 264.03698 264.31007 264.43051 264.43910

a: experimental data from reference 39. 215 Table8.2: Geometry of Trans Propiolic Acid, Structure (II)

Molecular Basis set RarameleisP.

STO-3G 3-21G* 4^1G &31G 631G* 6-31G*

Bond lengths R(C-H) 1.067 1.052 1.053 1.055 1.058 1.058 R(C^) 1.173 1.186 1.189 1.929 1.186 1.186 R(C-C) 1.489 1.451 1.443 1.447 1.466 1.466 R(C=0) 1.217 1.195 1.197 1.201 1.177 1.178 R(-C-O) 1.397 1.355 1.350 1.354 1.327 1.327 R(O-H) 0.988 0.965 0.952 0.952 0.949 0.945

Bond angles

-E(HF) 260.940377 262.941634 264.030286 264.300269 264.423017 264.431752

@: bond lengths are in A°, and bond angles are in degrees. * energies are in hartrees. 216 Dipole Moments. Moments of Inertia. and Rotational Constants Tables 8.3-8.6 show the dipole moments, rotational constants and moments of inertia for the two conformers. Tables 8.3, 8.5a, and 8.6a also show the experimentally determined parameters for Structure (I) (PA-cis). Clearly, the trans conformer has a higher dipole moment. Experimental dipole moments were determined by Lister and Tyler using Stark effect measurements. The orientation of the dipole is determined by the cosine of the ratio (|ix / [ij)- The experimental value is an angle -60° to the a-inertial axis and it is almost parallel to theC=0 bond. The calculated orientations are in the last columns of Tables 8.3 and 8.4 shown above. The best approximations for PA-cis confomer are given by the intermediate to large split valence basis sets (4-31G and 6-31G). For the PA-trans conformer, there is a marked basis sets dependence of the dipole moment orientation. The agreement with the experimental orientation is good to within 9% when the polarization functions are added to the large split valence or double zeta basis sets. The orentiation is -66° for the 6-31G** basis set which includes d functions for 0,0 and p orbitals for hydrogen. The agreement between calculated and experimental rotational constants is overall good. Again the MINDO / 3 method failed to predict these parameters with satisfactory accuracy. At the ab initio level the agreement is excellent (0.5% - 4%). The experimentally calculated inertial defects (A a.m.u.°A) is 0.2430. The calculated inertial defects are 0.0. The computations agree that the molecule Is planar. There are no experimentally observed parameters for the trans conformer but its predicted geometry and other structural parameters seem to follow the same patterns noted for the cis conformer. The calculated energies for the cis and trans conformers have been compared in Table 8.7. The cis conformer is used as the reference by virtue of its lower energy. The ST0-3G basis predicts the trans conformer is more stable than the cis. This is the reverse pattern where it is observed that structure (I) is on the average 5.0 kcals / mol lower in energy than structure (II). This prediction agrees with microwave observations that at ordinary temperatures, only one conformer is present and hence the adopted cis hydroxyl oriented structure for propiolic acid. 217 Table 8.3: Dipole Moments of Cis Propiolic Acid, Stnjcture (I)

Observed dpole moments, D

Ma Mb Mc [ty ofiertation(G) 0.80 1.38 0.00 - 1.599 60

Calculated dipole rTX)ments, D

Mx My Mz ^*r 0*

MINDO/3 -1.6411 -1.8149 -0.0000 2.4468 48 MP3/STO-3G 0.6457 -0.4294 0.0000 0.7755 34 HF/ST0-3G -0.5978 0.0003 -0.4179 0.7294 35 HF/3-21G 1.0333 -1.3360 0.0000 1.6889 52 HF/3-21G* 1.0476 -1.3468 0.0000 1.7062 52 MP2/3-21G* 1.2771 1.0579 0.0000 1.6583 40 HF/3-21G* 0.9930 -1.2146 0.0000 1.5688 51 HF/4-31G 1.1755 -1.6328 0.0000 2.0119 54 MP2/4-31G 1.0399 -1.4529 0.0000 1.7867 54 HF/4-31G(OPT) 1.0399 -1.4529 0.0000 1.7869 54 HF/6-31G(SP) 1.2091 -1.6400 0.0000 2.0375 54 MP2/6-31G 1.4614 -1.4451 0.0000 1.7930 54 HF/6-31G(OP"n 1.0614 -1.4451 0.0000 1.7930 54 HF/6-31G* 1.3192 -1.2273 0.0000 1.8018 43 MP2/6-31G* 1.3399 -1.2635 0.0000 1.8417 35 HF/6-3lG*(SP) 1.2456 -1.3588 0.0000 1.8434 47 HF/6-31G" 1.3751 -1.2545 0.0000 1.8613 43 HF/6-31G"(OPT) 1.2639 -1.3504 0.0000 1.8496 47 MP2/6-31G** 1.2650 -1.3515 0.0000 1.8512 47 MP3/6-31G** 1.2650 -1.3515 0.0000 1.8512 47

*: Orientation of dipole moment: [i^/^ij to a - inertial axis. 218 Table 8.4: Dipole Moments of Trans Propiolic Acid, Stnjcture (II)

Calculated dipole moments, D

^lx ^iy ^z ^T

HF/ST0-3G 1.0350 -3.1298 0.0000 3.2965 71.7^(72°)

HF/3-21G* 1.5839 -4.9992 0.0000 5.2441 72.0°

HF/3-21G*(OPT) 1.7493 -4.9573 0.0000 5.2569 71.0°

HF/4-31G 1.5934 -5.3947 0.0000 5.6251 74.0°

HF/4-31G{OPT) 1.8645 -5.3461 0.0000 5.6619 71.0°

HF/6-31G(OPT) 1.8896 -5.3938 O.OODO 5.7153 71.0°

CID/6-31G* 1.7775 -4.7356 0.0000 5.0582 69.0°

HF/6-31G 1.9783 -4.5060 0.0000 4.921 66.0(3)°

HF/6-31G** 2.0116 -4.4710 0.0000 4.9027 66(65.8)° 219 Table 8.5: Rotational Parameters [(Caculated vs Experimental (GHz))]

Basis Sets B

8.5a: Cis PropiolicAcid, Stmcture (I)

[39] 12.11009 4.14694 3.08449 [43] 12.11002 4.14694 3.08449 MINDO/3 11.42250 4.35044 3.15052 ST0-3G 12.23388 4.19657 3.12471 3-21G 12.16960 4.19704 3.12075 3-21G* 12.16668 4.19699 3.12054 4-31G 12.23153 4.19766 3.12516 6-31G 12.17139 4.16880 3.10523 6-31G* 12.53374 4.18733 3.13873 6-31G**(opt) 12.540850 4.187263 3.13914 6-31G**(prop=opt) 12.165159 4.151485 3.09521

8.5b: Trans Propiolic Acid, Structure (II)

3-21G* 11.72355 4.23411 3.11066 4-31G 11.55308 4.29129 3.12904 6-31G 11.69917 4.21872 3.10063 6-31G**(prop=opt) 11.54643 4.29261 3.12925 220 Table 8.6: Moments of Inertia

'A 'B "C

8.6a: Cis PropiolicAcid, Stmcture (I)

MODEL 42.1126 120.9109 163.0235 MINDO/3 44.2575 116.2022 160.4597 [39] 41.7447 121.9052 163.8929 ST0-3G 147.5201 430.0915 577.5720+ 3-21G* 148.2992 430.0040 578.3033+ 4-31G 147.5484 429.9398 577.4881 + 6-31G 148.2775 432.9164 581.1938+ 6-31G* 143.9907 431.0004 574.9912+ 6-31G** — — —

8.6b: Propiolic Acid (trans)

3-21G* 153.9416 426.2390 580.1806+ 4-31G 156.2131 420.5596 576.7726+ 6-31G 154.2624 427.7935 582.0559+ 6-31G*

moments of inertia are in atomic units. 221 Table 8.7: Cis/Trans Relative Stablization Energies -

Basis set AErei (=Eas-Etrans)

ST0-3G -hi.8679 kcal / mol *

3-21G

3-21G* -5.2084 kcal/mol**

4-31G -5.9614 kcal/mol**

6-31G -6.0869 kca 1/mol **

6-31G* -4.0977 kcal / mol **

6-31G** -4.5809 kcal/mof

': trans more stable; **: cis more stable. 222 In the evaluation of one-electron properties at the ab initio 6-31G** level, a very slight bent geometry for the HC=C-C chain (Figure 8.2) is obtained for propiolic acid. We believe that polarization functions apparently play a role in stabilizing conformations with lower local symmetry at the oxygen atom. In hydrazine, for example, a twisted structure is obtained with the ST0-3G basis and hydroxylamine gives a Cg structure with the OH bond trans to the direction of the lone pair on nitrogen [138 ]. The constructed model agrees with the computed model that while being planar, propiolic acid is not completely linear. It is thus not surprising that the energetics and structural requirements favor the intramolecularly bonded conformer (structure (I)). This value, however, is also compatible with the assumption that for experimental purposes and more so in theinterpretation of propiolic acid vibrational spectra, structure (I) (cis conformer) is the most abundant of the conformers of propiolic acid.

Quantum-Mechanical Normal Coordinate Analysis There is no report in the literature on either the experimental or theoretical force field for propiolic acid. The calculation of harmonic force fields of polyatomic molecules is important because they play an important role in the interpretation of vibrational spectra and the prediction of other vibrational properties. It is well established that completely empirical force fields cannot be determined from experimental data for molecules containing more than a few atoms without resolving to making numerous assumptions [139]. On the other hand, adequate sophisticated and purely theoretical quantum mechanical calculations cannot be carried out except for very simple molecules. Propiolic acid, with seven atoms is not a small molecule for such effective calculations. To achieve acceptable levels of accuracy it is necessary to compute force fields using large basis sets plus good treatments of electron correlation. Recently, this problem has been addressed by calculating the so called "scaled quantum mechanical (SQM) force field". The calculations that have been carried o'ut in the present study provide a first step in addressing this problem in the context of propiolic acid. 223

Rgure 8.2. Ab Initio Calculated Model Structure (6-31G**). 224 Force constants reported here were evaluated numerically from energy second derivatives using a standard option in the GAUSSIAN 82 program (FREQ = NUMER). The quadratic force constants reported here are for stmcture (I) (6-31G*) and stmcture (II) (6-31G). In this study, each atom was displaced around the equilibrium geometry by 0.01 A° or ±1 degree along the cartesian coordinates. Figure 8.3 defines the internal valence coordinates of propiolic acid (cis conformer). The internal symmetry coordinates of propiolic acid are shown in Table 8.8. Table 8.9 lists the computed harmonic diagonal force constants for the internal coordinates in the various basis sets which were used in the calculations. Table 8.10 lists the % deviation of force constants for every coordinate. The complete valence force field for the cis conformer is given in Table 8.11. Table 8.9 shows that the consistency of force constants for basis sets larger than the minimal basis set is good. The ST0-3G basis set overestimates all the force constants. The average deviation of the calculated force constants for the stretching modes is 5 - 25%. For the in- plane bending modes it is 2 -12%, and for the out-of-plane bending coordinates the % deviation is 6 - 36%. These large deviations are largely due to the ineffieciency of the ST0-3G basis set. The worst deviation is observed for the C=0 rocking / wagging coordinate. It is not surprising because the force constant for the 0=0 stretching coordinate is grossly miscalculated. It is well known that ab initio force constants could be overestimated by as much as 22%. In making a comparison between the quadratic force fields calculated ab initio and possible experimental fields, it should be remembered that the geometry employed in the ab initio calculations is different from the experimental geometry. Secondly, quadratic force constants obtained by ab initio calculations exhibit certain systematic deviations with respect to the experimental force field, particularly in overestimating the diagonal force constants [140]. Experience has shown that at the 4-31G and comparable levels [139] the calculated values overestimate diagonal stretching constants by up to 10% depending on the geometry employed (whether optimized or experimental). In-plane diagonal bending constants are overestimated by about 10-15%. With planar 7c-systems, diagonal out-of-plane deformation constants are consistently overestimated by up to 225

Rgure g.3. Definition of Internal Valence Cordlnates for Propiolic Acid. 226 Table 8.8: Symmetry Coordinates of Propiolic Acid

A'-block

0-Hstr Si=AR6 H-Cstr. S2 = ARi C^str. S3 = AR2 0=0 str. S4 = AR4 C-Ostr. S5 = AR5 COH bend Sg = A0i C-C stretch S7 = AR3 H-C^ bend S3 = AQQ 0-C=0 bend Sg = 1 / V 2(A04-A03) C-COOH rock Si Q = 1/ V 6(202 - A03-A04) C^-C bend (in-plane) Si 1 = A05

A" -block

H-C=C (out-of-plane) ^12 = ^^1 * 0=0 wag / rock Si 3 = A7U3 COH torsion Si4 = Ax 0=0-0 (out-of-plane) ^15 = A7:2

*.: Tc's are used here to define the out-of-plane torsional coordinates each involving four atoms. 227 Table 8.9: Computed Diagonal Force Constants for Stmcture (I)

8.9a : Stretching force constants (mdyn /AO)

H-C 0^ C-C C=0 0-0 0-H

6-31G* 7.116 20.404 6.151 16.578 8.140 9.182 6-31G 7.133 20.002 6.767 14.247 7.430 8.917 4-31G 7.104 20.356 6.711 14.633 7.414 8.810 3-21G* 7.069 20.422 6.035 15.152 7.298 8.405 3-21G 7.070 20.478 6.046 15.135 7.302 8.401 ST0-3G 8.769 23.195 8.231 18.460 10.269 12.125

8.9b: In-plane bending force constants (mdyn A°)

H-C^ 0=0-0 C-C=0 0-0-0 -C-O-H 0-C=0 6-31G* 1.427 1.253 9.378 9.934 3.195 2.110 6-31G 1.685 1.494 9.336 9.808 3.038 2.019 4-31G 1.678 1.776 9.456 9.951 3.034 2.059 3-21G* 1.733 1.826 9.022 9.745 3.155 2.027 3-21G 1.514 1.558 9.027 9.737 3.157 2.026 ST0-3G 1.690 1.642 8.923 9.205 3.638 1.985

8.9c: Out-of-plane bending and torsional force constants (mdyn A°)

^H-C=CC ^HC=C-C ^^0-0-0=0 '^C-COH

6-31G* 1.242 1.412 0.364 0.657 6-31G 1.645 1.712 0.034 0.606 4-31G 1.476 1.690 0.031 0.605 3-21G* 1.498 1.550 0.066* 0.585 3-21G 1.513 1.558* 0.076* 0.587 ST0-3G 1.587 1.691 0.045 0.541 228 Table 8.10: Deviations in Calculated Diagonal Force Constants

Internal Coordinate Average force constant* % deviation

Stretching force constants

C-H 7.410.6 8.4 0=0 20.8 ±1.1 5.2 C-C 6.7 ±0.8 11.5 0=0 15.7±1.4 20.0 C-0 8.0 ±1.1 13.4 0-H 9.3 ±1.3 - 13.9

In-plane angle bending force constants

H-C=C 1.6 ±0.2 6.8 0=0-0 1.6 ±0.2 11.9 0-0=0 9.2 ±0.2 2.2 C-C-0 9.7 ±0.3 2.6 COH 3.2 ±0.2 6.4 0=0-0 2.0 ±0.1 1.9

Out-of-plane torsional force constants H-C=CC 1.5 ±0.1 8.5 HC=C-C 1.6 ±0.1 6.7 CC-C=0 0.1 ±0.0 35.6 C-COH 0.6± 0.0 5.8

*: Standard deviation is for average of values for the six basis sets used. 229 Table 8.11 a: Force Constants for Propiolic Acid (cis conformer - 6-31G*)

8*

7.0688 -0.0121 20.4300 0.0246 -0.0539 6.0350 -0.0036 -0.0097 0.7943 15.1516 - -0.0084 0.0092 0.5278 1.1733 7.2981 0.0009 -0.0076 -0.-0139 -0.0458 0.0632 8.4052 0.0040 -0.0252 0.0034 -0.0251 0.0112 0.0017 1.4980 0.0040 -0.0252 0.0034 -0.0251 0.0112 0.0017 0.2340 1.5504 0.0005 0.0777 -0.0159 -0.0758 -0.1057 0.0026 -0.0604 -0.4070 -0.0188 0.0860 0.9583 -0.4917 -1.1118 0.2092 0.0646 0.3201 -0.0170 0.1102 1.2080 -1.5569 0.3307 0.3554 0.0207 0.0385

10 11 12 13 14 15 16"

9.0221 5.8625 9.7446 0.6472 1.0702 3.1544 0.0000 -0.0000 0.0000 1.7330 1.8262 0.0873 0.0065 0.2511 -0.0112 2.0272 0.0451 0.0022 -0.3341 -0.5849 230 Table 8.1 lb: Force Constants for Propiolic Acid (cis): 6-31G

8*

7.1325 -0.1738 20.0015 0.0265 -0.0023 6.7655 -0.0079 -0.0229 0.7746 14.2470 -0.0081 -0.0153 0.4988 1.3008 7.4300 0.0017 -0.0086 -0.0140 -0.0464 0.0647 8.9173 0.0047 -0.0405 0.0042 -0.0299 0.0196 -0.0036 1.4937 -0.0003 -0.3336 -0.0067 -0.0511 0.0554 0.0204 0.2004 1.7117 -0.0179 0.1959 0.9474 -0.0437 -.1233 0.1818 -0.0621 -0.3898 -0.0140 0.0500 1.1375 -1.5205 0.2407 0.3146 0.0841 0.2368 0.0030 0.0017 0.0584 0.0321 0.7521 0.5180 0.0140 0.0537 0.0000 0.0000 -0.0000 -0.0000 -0.0000 -0.0000 -0.0000 -0.0000

10 11 12 13 14 15 16*

9.3361 5.8012 9.8078 0.5958 0.9977 3.0378 -0.0000 0.0000 -0.0000 1.6847 0.0911 0.7906 -0.0072 -0.0254 0.0034 0.0489 0.1692 -0.0047 2.0193 -0.0047 -0.0366 0.0014 -0.3629 0.6056 231 Table 8.11c: Force Constants for Propiolic Acid (trans): 6-31G

8*

7.1136 -0.1781 0.1992 0.0220 -0.0408 6.2348 -0.0076 0.0007 0.8443 14.8676 -0.0017 0.0162 0.5561 1.2446 7.2981 0.0026 -0.0198 -0.0154 -0.0453 -0.0577 9.0571 0.0047 0.1163 0.0051 -0.0097 0.0042 -0.0016 1.5574 0.0019 -0.3814 -0.0120 -0.0556 0.0182 0.0079 -0.4199 1.5102 -0.0169 0.1843 1.020 -0.0609 -1.6586 -0.0929 0.0730 -0.3693 -0.0018 0.0468 1.1720 -1.2170 -0.5068 -0.2142 -0.0258 0.2681 0.0087 0.0062 0.0230 0.1217 0.8870 0.3397 0.0171 -0.0182 0.0 0.0 -0.0 -0.0 -0.0 -0.0 -0.0 0.0

10 11 12 13 14 15 16*

9.1939 5.5494 9.6789 •0.4338 -0.4404 2.9212 0.0000 0.0000 -0.0000 1.6273 0.7224 0.0010 0.0032 0.1393 -0.0020 0.6129 0.0067 0.0012 -0.0504 0.3090

**: The force constant matrix calculated at the HF 6-31G* and 6-31G levels for the two conformers of propiolic acid - structures (I) and (11) is given in Table 8.11. The following is a coordinate definition of the elements of the F matrix: 232 1 C-H

2 CHC 3 C-C 4 0=0 5 C-0 6 0-H 7 H-C=C in-plane bend 8 C=C-C in-plane bend 9 C-C=0 in-plane bend 10 C-C-0 in-plane bend 11 C-O-H in-plane bend 12 C=C-C out-of-plane bend 13 H-CC out-of-plane bend 14 C=0 out-of-plane deformation 15 0-C=0 in-plane bend 16 C-O-H torsion

The out-of-plane matrix elements (force constants) are actually given as torsions in the output data obtained from the calculations. In the tables that follow, the units employed are: stretch-stretch force constant - mdyn / A° ; stretch-bending force constant - mdyn, and for bend-bend force constant - mdyn . A°. 233 25% due to nondynamical effects [138]. The results of calculations of the force field of formic acid are given in [141]. The common modes between formic acid and propiolic acid are those of the carboxyl group. To ensure correct association and establish more ground rules for transferring force constants between related molecules, some diagonal force constants have been compared for formic acid and propiolic acid. While this is not the complete force field, it is fair to assume that force constants for the carboxylic acid group could be transferred between these related molecules, and that scaling factors can be used in determining the experimental force field of propiolic acid. The differences in magnitude of force constants could be the result of differences in electronic effects. The ethynyl group attached to the carboxylic acid group in propiolic acid is electron withdrawing and more so compared to the single hydrogen in formic acid. The presence of such a group results in reduced electron density on the carboxyl carbon. One point of disagreement is the SQQH' ^COH ^°''^® constant. It is more than doubled in propiolic acid. In Table 8.12, the values in brackets are projected force constants calculated for propiolic acid based on comparisons between the calculated and observed force constants for formic acid. Tables 8.13 and 8.14 list the calculated harmonic frequencies for the cis and trans conformers of propiolic acid respectively. Table 8.13 shows that the HF-SCF method gives frequency ordering of the vibrational modes in agreement with observed frequencies. The differences in absolute values of the wave numbers is attributed to the neglect of anharmonie corrections to the observed frequencies and to the fact that the HF-SCF method overestimates quadratic force constants. A configuration interaction calculation, or any other method which accounts for electron correlation along with the use of larger basis sets will help in correcting these deficiences. These effects generally tend to lower frequencies. In Table 8.14, it is noted that the ST0-3G and 6-31G* basis give a reversed order of the vibrational modes for V7 and V12. The ST0-3G, 4-31G and 6-31G give reversed order of vibrational modes between V3 and V13. The reversal of the ordering pattern between modes of the same molecule for different basis sets has been seen before in the case of 234 Table 8.12: 4-31G Diagonal Force Constants for the Carboxyl Group Modes Compared

formic acid [141] propiolic acid projected fc.

te=o / 0=0 14.837 14.633 (13.508)

^C-0 / c-0 7.447 7.414 (6.170) tan / OH 7.960 8.810 (8.520) toco / 000 2.324 2.059 (2.044)

^HOC / HOC 1.005 3.034 (2.140) ton / xOH 0.506 0.605

In Table 8.12 above, the projected force constants were determined by taking the ratio of the calculated to the observed force constants for formic acid multiplied by the corresponding calculated force constant for propiolic acid. 235 Table 8.13: Calculated Harmonic Frequencies (cm"^) for Cis Propiolic Acid, Stmcture (I).

ST0-3G 3-21G 3-21G* 4-31G 6-31G 6-31G* MINDO/3

A' - block

VI 4659 3873 3874 3967 3991 4049 4695 V2 4068 3659 3659 3666 3673 3664 3790 ^3 2607 2432 2430 2436 2421 2430 2354 ^4 2178 1943 1944 1918 1897 2029 1928 ^5 1636 1491 1490 1493 1492 1508 1435 ^6 1274 1235 1234 1259 1265 1335 1069 V7 956 937 936 933 935 894 1000 ^8 858 867 866 890 890 826 727 ^9 674 684 684 697 693 664 667 Vio 485 566 566 577 571 579 466 Vil 179 217 217 232 232 222 176

A"-lbiQC K vi2 867 1028 1028 1003 1002 882 779 VI3 742 877 877 873 859 866 618 VI4 635 626 624 650 651 648 518 327 326 290 257 VI5 329 327 327 236 Table 8.14: Calculated Harmonic Frequencies (cm"'') of Trans Propiolic Acid, Structure (II)

3-21G* 4-31G 6-31G

A' - block

VI 3911 3732 4026 V2 3654 3632 3667 V3 2414 2308 2404 V4 1990 1970 1949 V5 1341 1532 1357 V6 1275 1326 1280 V7 931 924 923 V8 868 902 897 V9 678 718 686 Vio 575 591 578 Vil 207 287 222

A" - block

V12 1026 984 978 V13 876 868 843 VI4 456 730 473 VI5 332 386 319 237 hydroxylamine [140]. In the trans conformer of propiolic acid, this reversal of frequency ordering patterns is observed for the V3 and vi 3 modes for the 3-21G*, and 4-31G and 6-31G basis set. Further reversal of patterns is also observed for vi 2 and vi 3. The frequencies predicted for both conformers using various basis sets are listed in Tables 8.13 and 8.14 above. In comparing the experimental and calculated frequencies for the cis conformer, several points may be noted: i) In the case of the cis conformer where a comprehensive comparison between theory and experiment is possible, the frequencies calculated using the 6-31G* basis set are in much better agreement with experiment. ii) All the calculated in-plane frequencies are too large, but never more than 10% for all the basis sets. The worst and best agreements are noted by the magnitude of the ratio w^^i^ / VQ^^g reported for the various matrices at the 6-31G* level. For the various matrices, the overall agreement is good for the In-plane modes with empirical conversion factors of 1.11 ±0.06 for Ar, 1.12 ±0.07 for CO, and 1.11 ±0.06 forN2 matrices, respectively. In the neon matrix, the corresponding empirical conversion factor is 1.11 ± 0.06. The conversion factor is generally observed to be 1.11 [138] which would suggest excellent agreement between the assignments and experimental frequencies, based on a quantum mechanical normal coordinate analysis. A comparison of MINDO / 3 calculated frequencies and observed frequencies supports the fact that the correct assignments may have been made since the calculated empirical conversion factor agrees well with the observed empirical conversion factor. iii) The out-of-plane modes show higher empirical conversion factors in all the instances but consistent in all the matrices. A comparison of the values calculated in all the matrices reveals a number of differences that may be of structural significance. i) Both conformers show reversal in the ordering patterns in some basis sets. ii) A comparison of the calculated frequencies for the out-of-plane modes of the two conformers shows that these frequencies are higher for the cis conformer than they are for the trans conformer. Although out-of-plane frequencies are not in particularly good agreement for the cis 238 conformers the difference between values predicted for the torsional frequency vi 4 for the cis and trans conformer, 107 cm"'' (at the most) is in good agreement with experimental evidence from related molecules. iii) Whereas all the in-plane frequencies calculated for the trans conformer, with the exception of the 0-H stretch, are smaller than those for the cis conformer, the out-of-plane frequencies are all smaller. iv) Since no frequency calculations were carried out for the trans conformer with the 6-31G* basis set we are tempted to make comparisons for calculations in similar basis set. For the 6-31G basis set, the 0-H frequency is larger in the trans than in the cis conformer leading to a smaller 0-H stretching frequency and a free OH in the trans. The same explanation would account for the higher VQ.Q I SQQH Tiode in the trans conformer compared to the cis. Overall cis-conformer frequencies are higher than trans conformer frequencies. Thermochemical information is also usually obtained from frequency calculations. Tables 8.15 and 8.16 summarize such information for the cis and trans conformers of propiolic acid. Included in such information are zero-point energies, heat capacities and absolute entropies at 298.15K obtained from the calculated harmonic frequencies and moments of inertia within the rigid-rotor harmonic oscillator approximation [73]. This information is necessary for discussion of thermodynamic properties of propiolic acid which have never before been calculated. The MINDO / 3 method is parametrized to optimize geometries and heats of formation. From MINDO / 3 calculations, the total energy of the cis propiolic acid conformer is computed to be -38.062995 au (-23892.522 kcal / mol), the binding energy is -1.402405 a.u. (-880.304 kcal / mol) and the heat of formation was calculated to be -144.312 kcal / mol. The drawback here is the fact that the contributions due to electron correlation are not known as this is included in the parametrization of experimental data. In the ab initio calculations, the zero point vibrational energy correction is about the same for both conformers 29 - 30 kcal / mol. It is also apparent that conformational changes do not result in any significant variations in the total energy, heat capacity (0^) and absolute entropies of propiolic acid. 239 Table 8.15: Thermochemistry from Frequency Calculations for Cis Propiolic Acid, Stmcture (I)

321-G 3-21G* 431G 6-31G 6-31G*

Total Enerov 32.451 32.446 32.651 32.626 32.673 Electronic 0.000 0.000 0.000 0.000 0.000 Translational 0.889 0.889 0.889 0.889 0.889

Rotational 0.889 0.889 0.889 0.889 0.889 Vibrational 30.674 30.668 30.873 30.848 30.896 C^.(ca\/mo\-K) 14.499 14.505 14.368 14.403 14.779

Electronic 0.000 0.000 0.000 0.000 0.000 Translational 2.981 2.981 2.981 2.981 2.981 Rotational 2.981 2.981 2.981 2.981 2.981 Vibrational 8.538 8.543 8.406 8.817 8.849 Srcal/mol-K^ 68.843 68.842 68.638 68.678 69.039

Electronic 0.000 0.000 0.000 0.000 0.000 Translational 38.655 38.655 38.655 38.655 38.655 Rotational 25.114 25.115 25.108 25.126 25.082 Vibrational 5.074 5.073 4.875 4.897 4.897 ZPE 29.677 29.671 29.907 29.877 29.857

T(K) 298.15 298.15 298.15 298.15 298.15

ZPE stands for zero point vibrational energy expressed in kcal / mol. 240 Table 8.16: Thermochemistry from Frequency Calculations for Trans Propiolic Acid, Stmcture (II)

3-21G* 4-31G - 6-31G

Total Energy 32.208 32.488 32.292 Electronic 0.000 0.000 0.000 Translational 0.889 0.889 0.889 Rotational 0.889 0.889 0.889 Vibrational 30.431 30.711 32.515

Heat Capacity 14.897 13.902 14.906 Electronic 0.000 0.000 0.000 Translational 2.981 2.981 2.981 Rotational 2.981 2.981 2.981 Vibrational 8.936 7.941 8.945

Entropy 69.300 67.886 69.207 Electronic 0.000 0.000 0.000 Translational 38.655 38.655 38.655 Rotational 25.146 25.141 25.155 Vibrational 5.499 4.090 5.398

ZPE 29.358 29.864 29.453

Stands for zero point vibrational energy. 241 These calculated thermochemical data would not mean much generally but in the absence of experimental data, predicted frequencies can obviously be used to estimate zero-point vibrational energies - energy terms that are essential in the comparison of experimental with theoretical reaction heats. For the cis conformer, the theoretical zero point energy is 29-30 kcal / mol. The corresponding value for the trans conformer is 30 - 32 kcal / mol.

Effects of Electron Correlation The effects of electron correlation on some molecular parameters was investigated. Calculations were carried out using the Moeller-Plesset Perturbation theory to third order, MP3. In some cases, configuration interaction with double subsitutions (CID) calculations were carried out for both the cis and trans conformers of propiolic acid. Table 8.17 shows results of calculations carried out for the cis conformer. Equation 8.1 below defines the correlation energy. It can be seen that inclusion of electron correlation increases the stabilization energy by at least 186 kcal / mol for the ST0-3G basis to 462 kcal / mol for the 6-31G** basis. In Equation (8.1) below, the subscript n is the order of the perturbation. The correlation energy is almost completely accounted for by the n = 2 term. Table 8.18 lists total atomic charges with and without inclusion of electron correlation. CID calculations carried out for the trans conformer of propiolic acid with the 4-31G and 6-31G* basis showed that while the energy improved (E(CID)= -264.4734263 hartrees) compared to (E(HF) = -264.0244123 hartrees), there was a rearrangement of the overall electronic population distribution but the net charge on the molecule was unchanged.

Barriers to Internal Rotation It is usual to infer from thermodynamic data that a potential barrier has to be surmounted when turning a molecule from one configuration to another. In general, there is a dependence of the potential, V, on the orientation of the parts of a molecule which can rotate relative to one 242 Table 8.17: Stablilization Energies and the Effects of Electron Correlation (kcal / mol.)

Basis set Method

RHF(OPT) MP2 MP3 A(CE)@

ST0-3G -260.937441686 -261.23512331 -261.25048058 186.801 196.438

3-21G -262.949907068

3-21G* -262.949913561 -263.45674927 318.050

4-31G -264.03511204 -264.55673071 327.326

6-31G -264.306987315 -264.82150745 322.872

6-31G* -264.427611378 -265.15321090 -265.15399525 455.328 455.820

6-31G** -264.436020107 -265.17165513 -265.17299185 461.626 462.465

(2): A(CE) = E(MPn) - E(HF) = Correlation Energy (8.1 243 Table 8.18: Total Atomic Charges for Cis Propiolic Acid

Hi ^2 C3 C4 O5 OQ H7

HF/ST0-3G 0.880 6.080 6.056 5.677 8.241 8.278 0.789

MP3/STO-3G 0.880 6.079 6.053 5.664 8.254 8.293 0.776

HF/3-21G 0.632 6.314 6.076 5.152 8.558 8.694 0.575

MP2/3-21G* 0.630 6.302 6.102 5.151 8.558 8.680 0.577

MP2/4-31G 0.655 6.296 6.094 5.177 8.534 8.691 0.553

MP2/6-31G 0.623 6.360 6.094 5.202 8.492 8.683 0.547

MP2/6-31G* 0.683 6.380 5.909 5.293 8.529 8.683 0.522

MP2/6-31G** 0.723 6.358 5.907 5.288 8.521 8.574 0.630 244 another. The minima in the curve correspond to more stable configurations of the molecule. The differences in potential energy between neighboring minima and maxima are known as barrier heights, while the potential energy differences between the various minima are termed energy differences between isomers. For propiolic acid, it is possible to rotate the hydroxyl (0-H) about the 0-0 bond. In this section, results of the conversion of the cis (Stmcture (I)) to trans (Structure (II)) conformer of propiolic acid have been investigated. The optimized geometrical parameters, energy, and dipole moments obtained with the ST0-3G, 3-21G and 6-31G* basis set for propiolic acid are listed in Tables 8.19, 8.20, and 8.21. The cis conformer corresponds to 0=0°, the transition state would be in the neighborhood of 0-90°, and the trans conformer corresponds to 0=180°. Sampling intervals varied from basis to basis. For the ST0-3G basis where calculations could quickly be done, 10-20° sampling intervals were used; for the 3-21G basis set, the sampling interval is 45 degrees; and for the 6-31G*, the sampling interval is 90° but an extra point was calculated at 98°. The various potential surfaces are shown in figure 8.4 for the ST0-3G, 3-21G, and 6-31G* basis set respectively. A least-squares fit of the usual cosine series to the calculated points was performed and yielded the following expressions for the potential function in the different models:

STO-3G V = 1.385 (1 -Cos0) + 3.498 (1 -Cos20) + 0.40 (1 -Cos30) (8.2)

3-21G V = 2.379 (1 -Cos0) + 4.038 (1 -Cos20) -»• 0.22 (1 -Cos30) (8.3)

6-31G* V = 3.190 (1-COS0) + 4.180 (1-Cos 20) + 0.790 (1-Cos30) (8.4) where 0 is the O = C-O-H dihedral, the units are kcal / mol. The primary maxima and ©^ax ^^^ shown in the plots (Figure 8.4). Calculations for the barrier at the MP2 / 3-21G* level yield a primary maximum (13.71 kcal / mol) at 90 degrees. The relative stabilization energy between the cis and trans 245 Table 8.19: ST0-3G Optimized Parameters for Cis PA Inversion to Trans PA

Dih. angle 0 40 80 90 130 150 180

Bond lengths (A°) R(Hi-C2) 1.067 1.067 1.067 1.067 1.067 1.067 1.067 R(C2-C3) 1.173 1.173 1.1733 1.173 1.173 1.173 1.173 R(C3-C4) 1.482 1.484 1.487 1.488 1.490 1.490 1.489 R(C4-05) 1.219 1.219 1.218 1.217 1.211 1.217 1.217 R(C4-06) 1.391 1.402 1.414 1.414 1.406 1.401 1.397 R(06-H7) 0.991 0.990 0.991 0.992 0.990 0.989 0.965

^^ 0.5072 0.6252 1.2816 1.4918 2.3337 2.6591 1.0224 Hy -0.5396 -0.8844 -1.3530 -1.3725 -1.0524 -0.6885 -3.1389 Mz 0.0000 0.6077 0.7308 0.8096 1.2674 1.4893 0.0000 ^^T 0.74 1.24 2.00 2.18 2.86 3.13 3.30 e 46.8 59.8 50.2 46.9 35.2 31.7 72.0

-E(a.u) 260.94606 260.93941 260.93234 260.93207 260.93584 260.93850 260.94037

AErel® 0.0 4.17 8.61 8.78 6.42 4.74 3.57

^ : The relative energy is in kcal / mol. 246 Table 8.20: 3-21G Optimized Parameters for Propiolic Acid Inversion Barrier

Dihedral angle 0° 45.0 90.0 135.0 180.0

RIH1-C2) 1.051 1.051 1.051 1.052 1.052 R(C2-C3) 1.185 1.186 1.186 1.186 1.186 R(C3-C4) 1.438 1.439 1.442 1.448 1.457 R(C4-05) 1.201 1.198 1.195 1.194 1.195 R(C4-06) 1.350 1.364 1.376 1.366 1.355 R(06-H7) 0.969 0.968 0.970 0.968 0.965

^^x 1.0476 1.4210 1.7436 2.4056 1.7491 ^y -1.3468 -1.3835 -1.9565 -1.3629 -4.9575

\^Z 0.0000 -1.5028 -2.6877 -3.9374 0.0000 ^T 1.71 2.49 3.75 4.81 5.26 0 52.1 55.2 62.3 60.0 70.6

-E(a.u.) 262.949914 262.941173 262.932905 262.936905 262.941634

AEre|@ 000 ^'"^^ ''0-67 8.16 5.20

@: The relative energy is in kcal / mol. 247 Table 8.21: 6-31G* Optimized Parameters for Propiolic Acid Inversion Barrier

Dihedral angle 0 90.0 98.0 180.0

R(O-H) 0.953 0.951 0.951 0.949 R(C-O) 1.323 1.350 1.349 1.328 R(C=0) 1.184 1.177 1.176 1.177 R(C-C) 1.456 1.460 1.461 1.466 R(C=C) 1.184 1.186 1.186 1.186 R(H-C) 1.058 1.058 1.058 1.058

^^x 1.3192 -1.8968 1.9783 \^y -1.2273 1.7186 -4.5060 \^z 0.0000 -2.3969 0.0000 n 1.80 3.51 4.92 0 42.93 42.18 66.30

-E (a. u.) 264.430475 264.410820 264.410882 264.417800

AEre|@ 0.00 12.33 12.30 7.95

@: The relative energy is in kcal / mol. ^^

248

7SE ,A ST0-3G B 3-2IG C 6 316* 15-

180 90 0 90 180 1T?ANS • CIS TRANS

Rgure Q.4. Energy Profiles for the Conversion of Cis to Trans Propiolic Acid. 249 conformers is 6.08 kcal / mol. All the calculated relative stabilization energies are so high that a Boltzmann distribution would strongly favor the cis conformer at room temperature. But high barriers can be conquered through photochemical excitation. From the studies of the barrier at various levels, it is clear that the conversion from the cis conformer to the trans conformer involves a high barrier. At room temperature, kT-0.6 kcals / mol. The conclusion from this study is that at room temperature only one conformer, the cis conformer, exists with appreciable population. This result is in agreement with the results of microwave studies. Furthermore, it supports the fact that absorptions in the infrared spectra of propiolic acid should be assigned as absorptions of the cis conformer. The fact that the cis conformer is the stable conformer suggests that the decomposition of propiolic acid may not be a unimolecular process. But propiolic acid decomposes rather easily to acetylene and carbon dioxide [38]. This suggests that heating the sample to a high enough temperature would probably trigger a cis / trans conversion and hence decomposition of propiolic acid would have the trans conformer as the reactive intermediate.

Geometry and Energy of the Transition StatefTS^ for Internal Rotation The predicted position of the TS (transition state) is basis set dependent. In going from the minimal ST0-3G to the extended 6-31G* basis, the predicted dihedral angle of the transition state ranges from 92.4° (ST0-3G), 96.98 (3-21G), 90.0 (3-21G*), to 92.92° (6-31G*). The TS stmcture is thus very close to that with the 0-H bond perpendicular to the molecular plane. The barrier heights to internal rotation from the predicted most stable rotamer, obtained by the ST0-3G, 3-21G, and 6-31G* optimizations, are 8.8 kcal / mol, 10.8 kcal / mol and 12.4 kcal / mol respectively. For the 3-21G* basis with electron correlation taken into consideration (MP2), the dihedral angle is 90 degrees and the barrier height is 13.7 kcals/mol. There are some differences in the geometries of the conformers. In the ST0-3G model, lengths in the cis and trans conformers are about the 250 same for bonds not associated with the rotation center. For bonds involved in the rotation, there are remarkable changes. The 0-H bond length is 0.03A° longer in the cis than in the trans. The 0-0=0 bond angle is smaller in the trans conformer by about 2°. At the 3-21G level, the COH angle is larger in the trans by at least 3° compared to the cis. The same trend is noted for the 6-31G* model. Although, the barrier seems high it is comparable to similarly calculated barriers for formic acid by Schwartz et al. [143]. In formic acid the ground cis conformer is 8.1 kcals/mol below the trans conformer and an intermediate barrier form with the hydrogen twisted up 90° is at 13.0 kcals above the cis conformer.

Conclusions from Conformational Studies MINDO / 3 and GAUSSIAN 82 have been used to establish and confirm the stmctures of the two conformers of monomer propiolic acid. Based on the results of the calculations of structure - related molecular parameters, the cis conformer (structure (I)) is the most stable conformer of propiolic acid. At room temperature, the cis conformer is the most abundant species. The energy difference is about 5 kcals / mol. The most important difference between the ab initio calculated geometry variations and those resulting from the MINDO force field calculation ( see Table 8.1 ) is seen in the behavior of the carboxylic acid group geometrical parameters - <0-C=0,

B: Molecular Reactivities Introduction Isomers with the composition C2H2O occur as products or intermediates in a variety of chemical reactions. The most important pathways to these isomers and to their higher homologues are the thermolysis or photolysis of a-diazoketones, the reaction of oxygen ('' D) with acetylene, the photolysis of ketenes and the reactions of carbenes with CO. The photochemical decomposition of propynal and the thermal decomposition of acetic and formic acids have been investigated by several authors [143,144]. There are a number of detailed studies involving the photophysical processes in the first excited singlet and triplet states in propynal that have been reported recently [145 -147]. While there is almost no polyatomic molecule for which our understanding of the primary photodissociation processes has reached a satisfactory level, such 252 information is not available for propiolic acid. It is, therefore, difficult to make any comparisons as to the probable nature of the decomposition of propiolic acid. In the absence of experimental data, the role of the theoretician and of his calculations are twofold: a) to guide the experimentalist in his research; b) to predict theoretical properties and the modes of formation and decay of unknown species. In the present study, calculations have been carried out to determine the energetics of various decomposition channels on the propiolic acid hypersurface. Some controversy exists as to whether the dissociation of formaldehyde to molecular products proceeds from the SQ or Ti state [148 - 149]. The isomerization of propiolic acid has been shown in the last section to occur on the SQ surface. The decomposition and energy profiles determined in this study are for the SQ and Ti states of propiolic acid. The present study also examines a number of minimum energy and transition state stmctures on the ground state, (SQ) and lowest triplet, (Ti) surfaces of propiolic acid. Two decomposition channels were investigated. The first of these is the decarboxylation process leading to molecular products. In decarboxylation there are two possibilities (Scheme 8.1). a) Unimolecular decarboxylation: Path (a) involves the formation of a four- center reaction complex resulting from a 1,3H-shift of the H7 atom. The geometry of the critical structure (III) has been optimized and its decompostion studied. b) Bimolecular decarboxylation: Path (b) involves a nine - center transition reaction complex. In the light of the present calculations it is thought that a key step in this mechanism would require the interaction of two acid molecules, yielding in a first step, a complex-(IV) that involves most of the atoms on the acetylenic and carboxylic acid groups, which subsequently decomposes into a ketene-like molecule (V) and propiolic acid. The second process that has been investigated is the unimoleculardecarbonylation of propiolic acid (Scheme 8.2). The decarbonylation occurs from the cis conformer via a 3-center reaction complex, VII. The decarbonylation process involves most bond lengths and bond angles of the propiolic acid frame. 253

\n'^ C=C-C

H-OC-H t (J . H-C-C^-C^

HI '^

Scheme 8.1a. Unimolecular Decomposition of Propiolic Acid. a

254

II

IH-IL

H

PAID HCCH-fC02

Scheme 8.1b. Bimolecular Decomposition of Propiolic Add. 255

H-CsQ H

1 0 H-oc<-.^/ ^/H Vf I H-0«.-rw..--.- /

Scheme 8.2. The Decarbonylation of Propiolic Add. 256 The effects of electron correlation on stabilization energies were studied by application of the MP perturbation theory to second order with the 6-31G basis set for all critical and transition state structures. All geometries have been optimized using SCF gradient methods and the 3-21G* and 6-31G basis sets. The natures of stationary state structures (minima, transition states) were verified by calculating the harmonic vibrational frequencies using the (FREQ = NUMER) option. Higher level calculations in terms of electron correlation (MP2 / 6-31G) were carried out with the 6-31G geometries. Potential energy surface features examined include: propiolic acid (SQ and Ti), cis and trans conversion (ie the internal rotation interconverting the singlet propiolic acid to trans and cis HCCHCOO, the molecular dissociation of cis and trans HCCHCOO, and their transition states, the rearrangement from HCCCOOH to HCCOHOO, the dissociation of HCCOHOO to molecular products; all of these on the Ti surface. The HCCCOOH rearrangement to H2CCCOO is postulated as a possible bimolecular process on the SQ surface.

Discussion The results of studies on the molecular reactivities of the various propiolic acid conformers have been divided into two parts: 1) The first part deals with the various aspects of the decarboxylation process on the SQ potential energy surface. In this part, an intrinsic reaction coordinate (IRC) is identified. 2) The second part of the discussion concentrates on the decarboxylation and decarbonylation processes occurring on the Ti potential energy surface. The discussion is concluded with some insight into other possible chemical decay channels for propiolic acid. In the discussion of the decomposition channels on the Ti surface, emphasis has been placed on the decarboxylation via a four-center transition state. The treatment of the decarbonylation process is much more general. The present study also assumes that the decarbonylation of propiolic acid can take place on the SQ potential energy surface. 257 SQ Potential Energy Surface (1-A' Electronic State) The decarboxylation of propiolic acid on the SQ potential energy surface can take place via one of two processes: a) Unimolecular decarboxylation: The unimolecular decarboxylation of PA would take place via a concerted mechanism. The concerted mechanism of decarboxylation implies at first a rotation of the hydroxyl group from the most stable conformation (cis) to the less stable conformation (trans). The barrier for this rotation has already been calculated to be 10.8 kcals / mol at the 3-21G level. From this point, the spontaneous decarboxylation is seen as a 1,3 - shift of the H7 hydrogen atom to the C3 carbon atom of the acetylenic group yielding acetylene and carbon dioxide. The concertedness of the mechanism is further defined by the formation of a four - center transition state 111 in which the formation of the C3 - H7 bond is simultaneous with the breaking of the Og - H7 and C3 - C4 bonds. The variations of these bond lengths in the predicted transition state with respect to their equilibrium values either in the reactant or in the products seem to suggest a complex reaction coordinate. Figure 8.5 shows a Z - matrix (distance matrix) description of the geometrical parameters used in describing the various stmctures involved in the concerted process. Table 8.22 shows results of point by point calculations carried out to define the intrinsic reaction coordinate for this process. A schematic representation of the results of Table 8.22 is shown in Figure 8.6. In the light of the present information and in comparison to results obtained in the decarbonylation of propynal [52,143] and the dehydration of acetic acid [144] the pertinent features of the unimolecular decarboxylation of propiolic acid are summarized. As displayed in Figure 8.7, the lowest channel ( energy threshold < 100 kcal) involves a number of transition structures with a maximum barrier height of 90 kcal / mol (A) and a minimum barrier height of 19 kcal / mol (0). There is a local maximum at 32 kcal / mol (B). This reaction path bears some similarities with that of formaldehyde decay H2CO => H2 + CO, the activation barrier of which has been the subject of numerous calculations [147], its most recent value being 80.9 kcal / mol [148]. The geometrical parameters for the various transition state structures on the energy surface are given in Table 8.23. 258

z

Rgure 8.5. Definition of the Z- matrix (Distance Matrix) Parameters for Propiolic Acid 259 Table 8.22: Molecular Parameters for Description of Intrinsic Reaction Coordinate

RC = 2.3415 2.1415 1.9415 1.7415 1.5415 A+

Rl 1.052 1.052 1.052 1.052 1.052 0.000 R2 1.186 1.187 1.187 1.187 1.189 0.003 R3 1.450 1.447 1.448 1.451 1.459 0.009 R4 1.194 1.192 1.189 1.186 1.183 -0.011 R5 1.355 1.355 1.357 1.361 1.367 0.012 O6H7 0.965 0.965 0.968 0.977 0.996 0.031

Al 179.89 179.81 179.49 178.89 177.66 2.23 A2 178.64 179.80 178.94 178.39 179.66 1.10 A3 125.51 128.00 130.44 132.76 134.90 9.39 A4 112.89 108.48 104.03 99.58 95.20 -17.69 A5 125.30 120.39 116.13 113.17 112.11 -13.19 0=0-0 121.60 123.52 125.57 127.66 129.90 8.30

^^x 1.5890 1.0219 0.4818 -0.0129 -0.4546 -5.0202 -5.1653 -5.1968 -5.1514 -5.0802 ^y ^z 0.0000 0.0000 0.0000 0.0000 0.0000 ^^T 5.27 5.26 5.22 5.15 5.10 0.17

-E(HF) 262.941453 262.937561 262.927704 262.910165 262.882468 0.05899

-E(UMP2) 263.455080 263.452018 263.443885 263.429354 263.406774 0.04831

AErel* 0.0 2.44 8.63 19.63 37.01 37.02

E(C02) ^^ 3-21G level is -186.561256 hartrees; E(PA) at 3-21G is -262.9499071 hartrees. * AErel is in kcal / mol and with respect to PA trans. +: A is change in molecular parameters compared to ground state values. 260

Figure 8.6. Pictorial Description of the Intrinsic Reaction Coordinate (IRC) atthe 3-21G level. 261

I00_ 90 t 80 \ \ I 1 I I 70 I

\! 60 c ::^5o

o \ ^40 \ \ 30 ' \ 20 \

c \

10 \

2.0 „ .,. I 'f ^ 0 yo 180 RIC3H7)A*' CiS TRANS

Figure 8.7a. Energy Profile for the Unimolecular Decomposition of Prooiolic Acid at the 6-31G level. 262

IR:

®

Rgure 8.7b. Definition of Molecular Parameters for Intermediate Structures. 263 Table 8.23: Molecular Parameters fqr Local Minima and Maxima on Reaction Coordinate (RC = C3 - H7) at the 6-31G SCF level

RC=A° 1.5556 1.5408 1.8666 1.9207

Rl 1.0545 1.0539 1.0516 1.0530 R2 1.1753 1.1776 1.1890 1.1825 R3 1.4671 1.4737 1.3996 1.4830 R4 1.2131 1.2082 1.1781 1.1777 R5 1.3828 1.3749 1.3971 1.3765 O6H7 0.9260 1.5280 0.8973 1.2102

Al 179.8466 179.8564 175.7256 179.3488 A2 179.6735 168.2833 172.2367 176.1115 A3 123.1479 125.5266 138.6166 130.631 A4 103.9356 104.7098 99.7526 102.4024 A5 124.5895 113.1499 107.4444 113.9581 OCO 132.9165 129.7636 121.6309 126.9666

•E(HF) 262.850554 262.797179 262.910892 262.889827

AErel 5^-0^ 90.53 19.18 32.40

*: from HF / 3-21G SCF model calculations. E is expressed in hartrees and AE is expressed in kcal / mol. Table also shows asynchronous nature of the decarboxylation process as there are no set patterns in variations in the magnitude of molecular parameters along the reaction coordinate. 264 In summary, the mechanism of the unimolecular decarboxylation of propiolic acid yielding C2H2 and CO2 first involves a conformational change from the most stable cis conformation of propiolic acid, 1 to the less stable trans conformer U. Then, the concerted decarboxylation proceeds from that intermediate via a four - center transition state (A) in which the breaking of the C3 - C4 and Og - H7 bonds takes place simultaneously with the formation of the C3 - H7 bond. Table 8.22 shows variations of the molecular parameters defining the progress of the reaction. It also shows that they are interdependent, and that, while the decarboxylation reaction is concerted, it is an asynchronous process. Comparisons of results of geometry optimization for the stable conformers and calculations of the concerted process at the ST0-3G and 6-31G levels showed that the overall description of the process is effectively represented by the results discussed here for the 3-21G model.

Details of the Postulated Bimolecular Decarboxylation of Propiolic Acid ( SQ Potential Energy Surface). The conversion of cis propiolic acid conformer to trans-propiolic acid conformer has been described previously as providing the suitable configuration for a 1,3 hydrogen shift. But, in the light of the present calculations it is thought that a key step in this process would require the interaction of two acid molecules, yielding in a second step a nine-centre transition state, which then decomposes rapidly into a ketene-like intermediate (V) and propiolic acid. The geometry of the critical structure V has been optimized and the transition state for its decomposition to vinylidene and carbon dioxide identified at the 6-31G level. Structure V lies in a well 34 kcal. above ground state propiolic acid. A comparison of its geometry and energy to those of products reveals that it is a van der Waals complex. The intermediate stmcture, V, can decompose into either singlet or triplet vinylidene and carbon dioxide. The vinylidene molecule is so reactive ( suggested lifetime of the order of 10'''0 s [149]) that, it rearranges almost immediately into acetylene. Figure 8.8 shows the geometry of V and the transition state for its decomposition. To further determine the nature of the TS, harmonic vibrational frequencies have been 265

Rgure 8.8. Structure of the Metastable Intermediate (V) at the 6-31G level. 266 calculated at the 6-31G level. No imaginary frequencies were obtained (Table 8.24a, 8.24b). This confirms the fact that V is a metastable intermediate. Dykstra and Schaefer [150] have studied the vinylidene - acetylene rearrangement (Equation 8.2):

H2C = C:=»H-C = C-H. (8.2)

In addition to their study, there have been a number of theoretical [151] and spectroscopic [149,152, 153] studies on the singlet vinylidene rearrangement. The barrier height is somewhere between 8.6 kcal (DZ -i- P-SCEP) and 26.0 kcal (DZ - SCF). The transition state stmcture and the energetics of the C2H2 rearrangement were not determined in the present study but the results of Dykstra and Schaefer from DZ + P-SCEP studies are given in Figure 8.9 [150]. Figure 8.10 shows total atomic charges obtained for the present study. This concludes the discussion of the decomposition of propiolic acid on the SQ potential energy surface. In the unimolecular decomposition channel a concerted mechanism can be used to explain product formation from the four - centre transition state. Although molecular products may be thermodynamically accessible relative to the low lying states, there is a very large barrier to this dissociation [154,155]. In the light of this possibility for formaldehyde and in view of the results of the present work for propiolic acid, it is tentatively concluded that the unimolecular decomposition to molecular products will very likely not occur from the SQ potential surface. In the bimolecular decomposition channel, product formation results from the decomposition of a metastable intermediate formed from the decomposition of a nine - center transition state. The geometries and energies of the key intermediates in both the unimolecular and the postulated bimolecular process have been calculated. The connection between IV and the vinylidene - acetylene rearrangement study has been made and used to explain the formation of acetylene. Calculated energies for the individual product molecules show that (V) could be better described as a van der Waals complex of vinylidene [ E( DZ-SCF) = -76.74019 [150], calculated E (6-31G) = - 76.74082 [present work] and CO2 [ E ( 6-31G) = ^^

267 Table 8.24a: Calculated Harmonic Frequencies of (V)

Symmetry Description

A' - block 3415 vC-H (asym. stretch) 3327 vC-H (sym. stretch) 2378 vC02 1834 vC-C 1418 6CH2 1407 5002 694 6HCC / 5CH2 654 6CO2 176 5C=C-C deformation 98 skeletal deformation 38 H2CCC deformation

A" - block 1011 YHCC/'yCH2 662 7CO2 118 TCH2 71 H2CCC deformation 268 Table 8.24b: Geometry, Energy, and Dipole Moment of (V) at 6-31G level

Molecular (V) (TSf parameters

C1-C2 3.0491 3.0487 C2-O3 1.1613 1.1613 C2.O4 1.1613 1.1613 C1-C5 1.2993 1.2992 C5-H6 1.0755 1.0754 C5-H7 1.0755 1.0754

C1C2C5 179.9977 179.9976 C1C2O3 91.3027 89.9822 Ci C2O4 91.3027 89.9822 CiCsHg 120.7970 120.7108 C1C5H7 120.7970 120.7108 O3C2O4 177.3946 177.3946

^lx 0.0001 0.0001 ^ly -2.7396 -2.7387 ^2 0.0000 0.0000 ^ij 2.74 2.74

E(hartrees) 264.255767 264.255767 -E(MP2) 264.759269

AErei@ 34.03 kcal/mol AEYe|@ ^9-0^ kcal/mol

(2): energy differences calculated relative to ground state propiolic acid. 269 H A

^/ 57.2- Yo

/ \ / \ 1.079?

l^)

C2H2 Reorrongement a.u. kcal •77.00 H50

40 -77.03 30

20 -77.06 - 10

0 -77.09

L I 90- I80** :c«c a — -CBC" Ib^

Figure 8.9. a: TS Structure for Vinylidene - Acetylene Isomerization. b: Energetics of the C2H2 Rearrangement [150]. 270

(V) TS

C1 6.855 5.855 02 5.114 5.114 03 8.449 8.449 04 8.449 8.449 05 6.626 6.626 H6 0.753 0.753 H7 0.753 0.753

Rgure 8.10. Total Atomic Charges for (V) and TS at 6 -31G level. 271 -187.5149520 [144]]. A divergent curve of the type shown for the C2H2 rearrangement [150] is observed for the decomposition of C2H2CO2.

Ti Potential Energy Surface ( 3-A' electronic state). As a consequence of the discussion of the previous paragraphs the emphasis of the following sections is on unimolecular processes occurring on the Ti potential energy surface. The decomposition channels investigated are the unimolecular decartDoxylation and decarbonylation of propiolic acid. As previously mentioned, the unimolecular decomposition channels have been investigated in the present study at the 3-21G* and 6-31G levels. The effects of electron correlation have been included using the Moller-Plesset perturbation theory to second order (MP2 / 6-31G // 6-31G).

Details of the Unimolecular Decarboxylation of Propiolic Acid As in the unimolecular process investigated on the SQ potential energy surface, the unimolecular decarboxylation of propiolic acid on the Ti surface yielding C2H2 and CO2 first involves a conformational change from the most stable cis conformer, 1 to the less stable trans conformer, H. Then, the decarboxylation proceeds from that intermediate via a planar four - center transition state, 2 in which the breaking of the'06-H7 bond leads to the formation of the metastable intermediate, 4_. The metastable intermediate 4 can exist in either the cis conformation, 4a, or as the trans conformer, 4^2. The formation of molecular products then proceeds from either 4a or 4i2 as a result of the mpture of the C3 - C4 bond. The whole reaction is a 1,3- transfer of the H7 hydroxyl hydrogen to the C3 carbon atom of the acetylenic group combined with a mpture of the system into two product molecules (Scheme 8.3). The geometry of the metastable transition state 4 (4a and 4^) has been optimized at the 6-31G (4a and 4b) level using the optimized geometrical parameters from the 3-21G* model. The decomposition to products and the various structures involved have been characterized 272

D

V-H^ I

f^CCHCOO]___p '^'^'SJ d i—> 4b TRANS)

Potn'i;arEnerg;SS." '^^°"^'^°^'"°" °' "-P'^''^ Acid on the Ti 273 at the 6-31G level. Figure 8.11 shows the optimized geometry of the metastable transition state, 4 at the 6-31G SCF levels. The 3-21G* model gives a structure ^ in which the carbon frame is essentially linear. The 6-31G gives a model with a bent carbon frame. The possibility exists for isomerization between the cis and trans conformers of 4 (HCCHCOO), as can be seen from figure 8.11. The potential energy surface for this internal rotation about the C3-C4 bond has been investigated. The trans form was found to be more stable than the cis conformer by about 0.65 kcals / mol. This low barrier suggests that at room temperature there is a 2 :1 ratio in relative abundance in favor of the trans conformer, 4ti. Results for the calculated geometries and energies of 4a and 4^2 along with the corresponding parameters for some points on the interconversion potential surface are given in Table 8.25. The stmcture of the cis / trans interconversion transition state, 5 is shown in Figure 8.11. The transition state for the cis / trans conversion occurs at 81.4 kcal / mol relative to ground state propiolic acid (Figure 8.12). In the 1,2 - hydrogen shift from triplet vinylidene to triplet acetylene, a large classical barrier of - 55 kcal / mol is observed and the TS is a non-planar structure. The magnitude of the barrier in this case is higher but the nature of the transition structure is observed to be the same as in the case cited [152]. The relatively small energy difference between 4a and 4i2 carbene complexes is In agreement with the same small energy difference between cis and trans triplet acetylenes computed by Wetmore and Schaefer [153 ]. These carbenes are also planar with the cis having a C2v symmetry and the trans a C2h symmetry. The gauche configuration is the transition structure with a C2 symmetry and it has the hydrogen perpen -dicular to the molecular plane. The decomposition to products has been investigated from both the cis and trans conformers of HCCHCO2.

Decompostion of trans - HCCHCOO. 4b (Table 8.26). Figure 8.13, shows a pictorial view of the progress of decomposition of reaction intermediate 4t2- The energy profile for bond rupture leading to products and the structure of the transition state for the decomposition of the trans conformer, 4^2 are shown in Figure 8.14. The transition state occurs at a 274

c ^ o

c> § c

(D

c o '^ '55 lO c i h- a>

Tc3 CO

c (0

e ts a 03

§1 CsO 275 Table 8.25: Optimized Geometries* and Corresponding Energies for HCCHCOO (Ti) at 6-31G SCF level

Dihedral angle 0.0 45.0 90.0 180.0

Hi.02 1.0680 1.068 1.0678 1.0688 C2-C3 1.3476 1.3627 1.3949 1.3481 C3-C4 1.459 1.4785 1.548 1.4585 C4-O5 1.2157 1.3729 1.372 1.3729^ C4-O6 1.3736 1.2125 1.206 1.2186^ C3-H7 1.0768 1.0826 1.1064 1.0737

-E(HF) 264.262440 262.245001 264.18026 264.263474

AEi(rel)** 23.74 34.68 75.31 23.09

AE2(rel)*** 29.85 40.79 81.41 29.20

*bond lengths in Angstroms; bond angles in degrees ** energy difference relative to the HCCCOOH trans conformer (ground state) ***energy difference relative to the HCCCOOH cis conformer (ground state) a,b have been switched to reflect conformational changes. 2766

o c g

CD > C

o

o

S^- o 3 1- CO 0)

CO c g w c (D I- •D C CO

o. 8> CD

c\iO (|OUJ/ 'TD0>j)3V^ OOQ CD X DO U-X 277

Table 8.26: Decomposition of trans - HCCHCOO {Ti) Complex, 4^

RC= A° 1.4585 1.5585 1.6585 1.7585 1.9585 2.1585

Rl 1.0688 1.0684 1.0679 1.0673 1.0652 1.0609 R2 1.3481 1.3401 1.3336 1.3275 1.3097 1.2772

R3 1.3729 1.3662 1.3613 1.3588 1.3657 1.3792 R4 1.2186 1.2091 1.2026 1.1978 i.191 1.1854 R5 1.0737 1.0726 1.0715 1.0703 1.0673 ' 1.063

Al 136.6783 137.0028 137.4327 138.0766 141.6308 151.3127 A2 125.2387 124.0549 122.8334 121.5395 118.7243 115.4614 A3 115.2628 113.6994 112.2071 110.8155 108.6727 107.1434 A4 124.8887 124.6742 124.6464 124.8666 126.3488 128.4567 A5 120.9988 123.2911 125.6147 128.0396 133.8714 142.7905 O=C-O119.0 121.6264 123.1464 124.318 124.9785 124.3998

^l-p 3.2800 3.2083 3.1180 3.0435 2.9685 2.8270

•E(UHF)264.26347 264.25861 264.24719 264.23265 264.20333 264.18588

AE' 23.09 26.14 33.31 42.43 60.83 71.78 AE" 29.24 32.29 39.46 48.58 66.98 77.93

E* refers to trans PA (ground state) E" refers to PA cis ground state. E(UHF) in hartrees 278 Table 8.26 continued.

C3 - 04= A° 2.3585 2.4585 2.3585^

Rl 1.0549 1.0542 . 1.0556 R2 1.223 1.2111 1.2355 R3 1.3825 1.3805 1.3826 R4 1.1823 1.1813 1.1832 R5 1.0558 1.054 1.0576

Al 174.7241 177.2953 170.3187 A2 107.7641 105.2343 110.1841 A3 103.3269 99.2124 105.2443 A4 132.4564 136.5606 130.5193 A5 162.9032 168.9035 156.8093 OCO 124.2167 124.227 124.2364

^iD 2.6151 2.3682 2.7771

-E(HF)a.u. -264.18398 -264.18483 -264.18263

AE' 72.97 72.44 73.82 AE" 79.12 78.59 79.97

These are transition state parameters( Figure 8.14). Figure 8.13. Pictorial Representation of the Progress of Decomposition. 280

5y 8.14. Oe«n,„on o, Geo.e.nca, Pa.a.eters In ..e Deco ^position 281 03 - C4 separation of 2.2931 A° with an activation barrier of 50.7 kcal / mol. The effects of electron correlation have been considered by UMP2 / 6-31G / / 6-31G calculation based on the equilibrium geometry of the transition state. The activation barrier relative to ground state propiolic acid is 79.94 kcal / mol (-80 kcals). The energy profile for the decomposition of 4l2 is shown in Figure 8.15. When electron correlation is accounted for, this barrier becomes 117.98 (118) kcal / mol.

Decomposition of cis- HCCHCOO (Ti) Complex. The two lowest cis triplet states of acetylene are the ^A2 and ^62 [157].

HvJC)94 "SOSH K I2Z8^.

A2 ^62

It is thought that in the light of the closeness of the cis and trans energies, that decomposition from the cis carbene should occur instantaneously at a C3-C4 distance of 2.059 A° giving an activation barrier with respect to ground state propiolic acid of 78.5 kcal / mol. This barrier is much higher than that computed for the trans carbene intermediate (50.7 kcal / mol). The energy profile for the decomposition of the cis complex is shown in Figure 8 .16 and Table 8.27. It is somewhat difficult to discuss this particular decomposition because the complex was constrained to a planar geometry and it might be possible that internal rotation of the CO2 group about the C3 - C4 bond 282

AE t 90

80 •0

70

60

50 - 40

30 -

20 - 2a24Kcal 10 6.15 I. _j .1 I t I i^^ >^ ' ' t ' » ' ' » 14 20 ^6 iL RCIAI

Rgure 8.15. Energy profile for decomposition of 4b (Ti). 283

AE '0 8 i

70

60

50 78.50 kcal

40

30

''23.74 kcal 20

0 u V- i J—I 1 1—I—,J,,^—I 1 1-4 ZO RCIA")

Rgure 8.16. Energy Profile for the Decomposition of CIs-HCCHCOO (Ti). 284 Table 8.27: Decomposition of Cis -HCCHCOO (Ti)

RC= A° 1.459 1.559 1.759 1.959 2.059* opt=ts

H1C2 1.0680 1.068 1.0677 1.0659 1.061 1.068 02^ 1.3476 1.3404 1.3298 1.314 1.2887 1.3477 04^^ 1.3736 1.3667 1.3581 1.3647 1.3808 1.3737 C4O6 1.2157 1.2071 1.1973 1.1904 1.187 1.2156

C3H7 1.0768 1.0754 1.0728 1.0692 1.0663 1.0768

H^C2C3 134.7053 135.2349 136.5677 141.1678 154.2267 134.7374 C2C3C4 121.8104 120.2745 117.2306 114.5898 113.1773 121.8378 C3C4O5 112.557 111.2534 109.0091 107.5074 107.3093 112.5782

03^4^6 127.4582 127.1117 126.6285 127.4439 128.7008 127.4787 C2C3H7 121.0112 123.3047 128.1087 134.4308 140.5543 121.0529 O5C4O6 119.9841 121.6349 124.3634 125.0476 123.9899 119.9432

^^D 4.1628 4.0804 3.9327 3.9001 3.7789 4.1628

-E(HF) 264.26244 264.25741 264.23018 264.19745 264.18499 264.26244

AE.,. 23.74 26.89 43.98 64.52 72.35 23.74

AE2** 29.89 33.04 50.13 70.67 78.50 29.89

*: Calculations do not converge after this point; Starting molecular parameters are obtained. It would appear that product formation has occurred giving an upper energy limit for dissociation of 78.5 kcal / mol. 285 takes precedence over its decomposition. In the decomposition of the trans complex in the previous section, the product orientation would be expected to have the triplet acetylene and carbon dioxide parallel to each other and parallel to the plane of the molecule. The cis complex on the other hand should yield products parallel to each other but perpendicular to the molecular plane. The geometries of the resulting triplet acetylenes have been discussed. A number of studies have been carried out on the geometries of acetylene in the ground and low lying excited states. In this study the possibility exist for having four acetylene structures for the low lying triplet states. In agreement with the results of Abramson et al. [158 ] we found the cis triplet acetylenes would be more stable than the trans acetylenes. We attibute this difference in stabiltity to the result of steric effects caused by the presence of the bulky carbon dioxide group. The bond lengths, in general, for the acetylenes are shorter than those of the predicted acetylene structures by about 0.01 A° . The bond angles show a much larger variation when molecular parameters are compared to the experimental geometries of acetylene. We believe these differences in geometry are simply due to basis set effects. The literature studies cited used larger basis sets and accounted for the effects of electron correlation. The results for the molecular parameters of the acetylenes of this study are comparable to the acetylene ground state molecular parameters: rg(C = 0) is 1.1213 (DZ + P), 1.230 (DZ); TQ (0 - H) is 1.066 (DZ + P) and 1.071 (DZ); ©e (H - 0 = 0) is 180.0° (DZ-H P) and 180° (DZ). But while ground state (SQ) acetylene is linear, ttie triplet acetylenes are planar and bent. The energetics of the bimolecular decomposition of propiolic acid (SQ) and unimolecular process on the Ti surfaces are summarized in Figure 8.17. In Figure 8.17 the transition states to products from the cis (4a) and trans (4b) conformers of HCCHCOO have been labeled as TSC and TST respectively. The various triplet acetylenes constituting the products are represented by ^62. ^A2 for the cis structures and by ^A^ and ^B^ for the trans stmctures. The stability order of the resulting acetylenes is evident. 286

AE T 120

100

CIS JRANS 80 JS «~^ -ff.9l4

E ^B2+C02 \tCO2 60 62.13 eas ^.Bu+CO

40

HpC JKJO^ 4a kb 39L67 20 T I 47. ^ ^^' 35.72 29^ 29.2

T HCClT?C02

Rgure 8.17. Summary of the Energetics of the Decarboxylation Process. 287 Decabonylation of Propiolic Acid This reaction is expected to occur via a three - center reactive intermediate (Scheme 8.2) as is the case in propynal [ 52, 143,145 -146]. The chemical decay behavior of propiolic acid implies the elimination reaction.

H-C=C -COOH => 0 O -h HO =C-OH (cis, trans).

The various isomers of the resulting hydroxyacetylene have been studied in more detail by Tanaka and Yoshimine [ 69] in both the ground singlet and excited singlet and triplet states. The isomers of HC=C-OH have been studied at various levels for various reasons. A more related study is the one by Basch [ 159] who was interested in the photodecomposition of the ketone (C2H2O) to CH2 + CO. He pointed out that the ^Bi and ^Bi (triplet and singlet) states dissociate directly to CH2 -•• CO without a barrier. While the decomposition of HC=C-OH (hydroxyacetylene) has not been studied here, this study reports geometries and energies along with vibrational frequencies for the reaction intermediate leading to hydroxyacetylene and carbon monoxide. The precursor is propiolic acid. In the case of propynal, spectral and dynamic data has indicated that the Si and Ti potential surfaces are nondissociative [160 -161]. It was also inferred that the dissociation takes place on the SQ potential surface. Extensive ab initio calculations at the 01 level were performed to map the SQ surface and to examine chemical decay channels. The calculations predicted the lowest reaction path for decarbonylation being a concerted mechanism involving the migration of the aldehyde H atom in the molecular plane. The barrier height was calculated to be 69.9 kcal/mol [161]. The suggested mechanism involves excitation into the Si vibrational ground state at 74.8 kcal / mol followed by internal conversion. An upper limit for the reaction has been estimated at 72 kcal / mol [162]. In the present work, extensive ab initio calculations at the HF SCF level have been performed to map the Ti potential surface and to examine the chemical decay channels of propiolic acid. The lowest reaction path predicted by calculations is decarbonylation via a concerted mechanism 288 which involves the migration of the OH group in the molecular plane. Taking into account the zero point energies, the barrier height has been calculated to be 89.7 kcal / mol. Figure 8.18 shows the stmcture of the three-center reaction intermediate (VII). Tables 8.28 and 8.29 show data for optimized geometries and energies for the reaction intermediate and other stationary points on the surface. Data for ground state propiolic acid has been included in Table 8.28 to show variations in geometrical parameters as a result of decarbonylation. Figure 8.19 shows a pictorial view of the progress of decarbonylation as the C3-C4 bond is being stretched. The potential energy profile for decarbonylation is shown in Figure 8.20. Table 8.30 shows vibrational analysis data for the reaction intermediate. Table 8.31 shows total atomic charges for ground state propiolic acid, the metastable reaction intermediate, and other stationary points on the reaction coordinate. In the light of the results presented in this section, and in comparison to the more detailed studies of propynal, the pertinent features of the decarbonylation reaction of propiolic acid are summarized. As displayed in Figure 8.20, the lowest channel involves a metastable intermediate and a number of secondary minima with a maximum activation barrier height of 170.36 kcal / mol. The calculations indicate that the decarbonylation channel involves a concerted process in which the 03-05 bond is being formed while the C4-O6 and C3-C4 bonds are being broken. Table 8.31 shows that VII is a carbene. It is also a shallow minimum on the energy hypersurface. Since the possibilty exists for cis and trans conformers, it is likely that this metastable carbene-like intermediate could have two exit channels although both would lead to the same products. The first exit channel is a stretching of the C3-C4 bond and possibly a saddle point to yield HC=C -OH (cis) and CO. The second channel leads to the transformation of MO to yield HC=COH(trans) and CO. This concludes the section on the molecular reactivities of propiolic acid conformers. The emphasis of the previous sections has been to provide fundamental information necessary for the discussion of the more theoretical aspects of the dynamics and photophysical processes occuring in propiolic acid. To this end, geometries, energies and frequencies have been calculated and the various energy surfaces constmcted. Transition 289

(e'aiG lelel).''"'''"^ °' Decarbonylation Reaction Intermediate 290 Table 8.28: Optimized Geometries and Energies for Decarbonylation Reaction Intermediate(VII)

Molecular parameters Cis PA (3-21G*) 3-21G* 6-31G//3-21G*

H1C2 1.0514 1.0672 1.069 C2C3 1.1852 1.3536 1.3641 C3C4 1.4377 1.4408 1.438 C405 1.2011 1.2094 1.2152 C406 1.3504 1.3803 1.3806 06H7 0.9686 0.965 0.9502

H1C2C3 179.73 135.23 135.13 C2C3C4 177.23 120.89 123.41 C3C405 126.09 130.34 129.67 C2C306 125.59 123.92 ^306'^7 112.96 115.27 ^40306 113.53 112.68

^^x 1.0476 1.1383 1.1068 -1.3468 4.0796 4.3493 ^y ^^z 0.0000 0.0000 0.0000 HT 1.71 4.24 4.49

-E(HF) 262.949901 262.863995 264.230469

E(MP2) -263.456749 -263.294702 -264.660064

AE(kcal / r nol) 0.00 53.90 48.02 (with ZPE;1 0.00 28.28 24.05 291 Table 8.29: Molecular Parameters for Stationary Points on Reaction Coordinate at the 6-31G SCF level.

RC= A° 1.638 1.938 2.038

Rl 1.0693 1.0715 1.0722 R2 1.346 1.3305 1.3267 R3 1.3706 1.3545 1.3504 R4 0.9515 0.9544 0.9551 R5 1.1913 1.1584 1.1498

Al 135.291 133.5599 132.9287 A2 127.6504 131.2464 131.7667 A3 115.4689 116.3874 116.9276 A4 121.429 118.8256 118.191 A5 126.8513 120.5467 117.9664

^^x 0.6736 0.3693 0.2994 4.4145 3.9313 3.6335 ^ ^^z 0.0000 0.0000 0.0000 ^J 4.4656 3.9486 3.6458

- E(HF)* 264.213200 264.173659 264.168379

AErel (kcal) ^8.85 83.66 128.67

': energy expressed in hartrees 292

Rgure 8.19. Pictorial Representation of the Progress of Decarbonylation. 293

Figure 8.20. Energy Profile for Decarbonylation of Propiolic Acid. 294 Table 8.30: Harmonic Frequencies (cm"'') and Thermochemistry of the Metastable Reaction Intermediate (VII)

Symmetry 3-21G* 0.9vj 6-31G 0.9vj Description vi(A') 3903 3512.7 4044 3639.6 VQH V2(A') 3433 3089.7 3438 3094.2 VH-C 1572 V3(A') 1414.8 1571 1413.9 vc=0 1462 1315.8 V4(A') 1498 1348.2 VC-Q/SCOH V5(A') 1425 1282.5 1422 1279.8 VC-C V6(A') 1175 1057.5 1214 1092.6 ^COH V7(A') 967 870.3 970 873.0 V8(A') 932 838.8 944 849.6 V9(A") 707 636.3 663 596.7 ^H-C=C Vio(A') 630 567.0 633 569.7 Vil (A") 514 462.6 490 441.0 Vi2(A') 452 406.8 462 415.8 Vi3(A") 291 261.9 303 272.7 Vi4(A') 262 235.8 274 246.6 Vi5(A") 196 176.4 162 145.8 ^HOCC=0

ZPE 25.62 25.86

Total Energy 28.855 29.120

Cv(cal / °K-mol) 16.933 16.993

S(cal / °K-mol) 73.707 74.006 295 Table 8.31: Comparison of Total Atomic Charges for Decarbonylation Intermediates at the 6-31G SCF level.

PA cis MC(Ti) A B

Hi 0.623 0.805 0.804 0.813 0.816

02 6.360 6.084 6.102 6.150 6.174

C3 6.094 5.817 5.823 5.746 5.709

C4 5.202 5.558 5.530 5.581 5.602

O5 8.492 8.420 8.429 8.402 8.387

Og 8.683 8.741 8.739 8.738 8.741

H7 0.547 0.576 0.573 0.571 0.570 296 state structures have been identified and corresponding activation barriers determined.

Further Discussion The formation of radicals although not discussed cannot be discounted. It is seen as a possible route to the formation of hydroxyketene (Scheme 8.4a) below. In contrast, anyway, to the decarboxylation reaction, the reaction intermediate for decarbonylation at both the 3-21G* and 6-31G SCF levels is a minimum energy stmcture. In decarboxylation, the reaction intermediate at the 3-21G* level is a saddle point. Also energy differences computed relative to ground state propiolic acid show that the metastable carbene-like intermediate is more stable than the corresponding reaction intermediate for decarboxylation. Hydroxyacteylene is very unlikely to occur in the Wolff rearrangement because of high barriers against its formation from either formylmethylene or from the ketone. However, once formed it should be stable. It is surprisingly stable to rearrangement into ketene, althout it is the enol form. Therefore, it might very well occur in dense clouds. Pacansky and Bargon have calculated its rotational constants and considered the possiblity of its formation from the photolysis of perpropiolic acid in an argon matrix [163 ]. We seek here to establish the alternative path way of using propiolic acid instead of the more dangerous compound perpropiolic acid. Propiolic acid could yield hydroxyacetylene according to the following scheme (Scheme 8.4b)

Overall Conclusions from MO Studies The vjork that has been discussed in this section is meant to compliment the work done in the first part, in an attempt to better understand both the vibrational spectroscopy and other physical aspects of propiolic acid chemical dynamics. In the eariier sections, molecular parameters relevant to the structure of propiolic acid were calculated. These have been 297

a) 0 .yO // C H-C-^'C ^+1 ^ H-C-p •Q Y H

H\ :=C=C=0 Hc /

b)

H-C=C—a'^'^+OH

H2C=C=0 ^ H-C=C—OH -f- CO

Scheme 8.4. Other Possible Decay Channels in Propiolic Acid. 298 used to propose a quantum mechanical model comparable to the structure determined from microwave data. The HF SCF method with various basis sets has been used to give vibrational frequencies and force constants for propiolic acid for the first time. The results of the 6-31G* basis set have been used as an aid in assigning experimental frequencies determined in the first part. Although the HF SCF method gives frequencies higher than the observed ones by as much as 15%, the calculated normal coordinates are reasonable, practical and acceptable. The ab initio MO method is useful particularly when we attempt to understand or predict some systematic deviation or similarity of molecular parameters between related molecules. The force constants calculated in this study, for example, have been compared with similariy calculated parameters in more well studied systems (e.g., formic acid) in an attempt to better understand the rules for transferring molecular parameters such as those between stmcturally related molecules. The various unimolecular decomposition channels of propiolic acid have been investigated. The transition and metastable structures involved in the two principal decomposition channels - decarboxylation and decarbonylation - have been identified and characterised by obtaining optimized geometries and energies, dipole moments, and rotational constants where applicable. Frequency calculations have been performed to differentiate between transition state structures and local minima on the hypersurface. The bulk of the work done here has been largely on the lowest triplet state potential energy surface for propiolic acid. The barrier to interconversion between the cis and trans conformers has been determined at various levels of SCF theory and shown to be comparable to barriers of the same process in other systems. The barrier to internal rotation calculated here for propiolic acid has a maximum value of 13.2 kcal / mol. The cis / trans energy difference varies from 3.5 to 8.1 kcal / mol. This high barrier is in agreement with microwave studies, suggesting that ground state propiolic acid exists almost exclusively in the cis form. In view of the high barriers involved in various decomposition channels, it is suggested that the unimolecular decomposition of propiolic acid is very likely to be achieved through photochemical means. Such a 299 process could be important in obtaining molecules such as hydroxy - acetylene in the laboratory and characterizing them. This would provide guidelines to astronomers on what chemical and otherwise properties to look for in interstellar clouds in an attempt to locate these species although such observation would be very unlikely. Overall, these molecular orbital calculations are meant to provide guidelines for interpreting the behavior of the propiolic acid monomer and suggests possible avenues for future experimental work, e.g. using the propiolic acid monomer to produce hydroxyacetylene. Such experiments could be carried out using the matrix isolation technique or molecular beam sampling conditions. In more elaborate experiments, the techniques of comparative laser pyrolysis could be applied to the molecular elimination reactions discussed here. CHAPTER IX OVERALL CONCLUSIONS

In this dissertation the methods of infrared matrix isolation spectroscopy and quantum mechanics have been used to obtain information about propiolic acid. In the first part of this disseration attention was focused on obtaining a self-consistent pool of frequencies necessary for the determination of the force field of propiolic acid. The technique of matrix isolation has been used appropriately to obtain such information. The fundamental vibrational frequencies of propiolic acid have been obtained and assigned. Experiments concentrated on obtaining spectra of propiolic acid isolated in solid matrices (Ar, CO, N2, and Ne). ^^0-labeled isotopomer spectra have been recorded for samples isolated in solid argon and nitrogen matrices. For the first time, matrix isolation technique spectra have been recorded for deuteriopropiolic acid (PA-D2). Spectra were recorded for PA-D2 isolated in solid argon, carbon monoxide and nitrogen. ^^0-labeled spectra were also recorded. The results of fundamentals for the principal isotopomers have all been combined to provide a better understanding of the vibrational spectroscopy of propiolic acid. A comparison of H/D isotope exchange leading to HCCCOOH and DCCCOOD strongly support the contention of this work that the out of plane acetylenic hydrogen deformation mode occurs at a frequency higher than the in plane mode. Accordingly, it has been assigned as the peak obseved at 760 ± 4 cm ""I. It is believed that thirteen of the fifteen fundamental vibrational modes of propiolic acid are in the region of 4000 - 400 cm"'. Spectra were recorded at temperatures ranging from 11K to 35K. Effects of temperature variations have been deduced. They are more important in spectra recorded in solid nitrogen matrix. Such effects were used as an aid to assigning some fundamentals. Spectra recorded in solid argon and cartDon monoxide matrices did not show any appreciable

300 301 temperature dependence. Spectra were also recorded in argon and nitrogen matrices for various M / S ratios varying from.4000 /1 to 500 /1. Variations in the resulting spectra were found useful in differentiating between monomer absorptions and absorptions due to the formation of dimers and higher oligomers. The second part of this dissertation deals with results of quantum mechanical calculations using MINDO / 3 and GAUSSIAN 82. Geometries, energies, dipole moments, rotational constants, force constants and vibrational frequencies have been calculated. All of these parameters support the fact that the cis conformer is the most stable form of propiolic acid. Based on these parameters and calculated moments of inertia, a model consistent with the results has been postulated. This model agrees with microwave data in that propiolic acid is planar but disagrees with the fact that it is completely linear. One electron properties have been calculated. These can be used to describe the nature and various other electrostatic aspects of the interaction of propiolic acid with other charged or uncharged species. The force constants calculated agree with similar calcualtions for related molecules. It is possible under the circumstances to use scaling factors employed in the calculation of the force field for formic acid in similar calculations for propiolic acid. This avenue will be employed in the future to obtain the experimental force field for propiolic acid. Unimolecular decomposition channels have been investigated for propiolic acid. Decarboxylation can occur in two ways. It can occur via a four-center reaction intermediate (unimolecular) or an eight or nine-center transistion state (bimolecular) involving a hydrogen transferring agent. This decarboxylation can occur from the SQ or from the Ti electronic states. We have shown that in the SQ state, decarboxylation via a bimolecular process is the lowest energy decay channel, whereas in the Ti it is a unimolecular process. The decarbonylation of propiolic acid occurs through a three-center transition state. It involves the decomposition of a metastable conformer on the potential energy surface. The situation in propiolic acid is very similar to that observed in propynal. Results of decarboxylation and decarbonylation compare energetically to results of similar processes in other acids. These results 302 have been combined to provide the possibility that hydroxyacetylene and hydroxy ketene can be derived from propiolic acid. Overall, the work of this disseration presents the fundamentals that can be used in the future for more sophisticated studies both experimentally and theoretically. REFERENCES

1. S. D. Peyerimhoff and R.J. Buenker, J. Chem. Phys.,5Q, 1846 (1969).

2. 0. V. Berney. R. L Redington and K.C. Lin, J. Chem. Phys. 53, 1713(1970).

3. R. L Redington and K. 0. Lin, J. Chem. Phys., £4. 4111 (1971). 4. 0. J. Wilson and H. H. Wenzke, J. Chem. Phys. ^, 1265 (1935). 5. J. K. Kochi and G. S. Hammond, J. Am. Chem. Soc, 7^, 3452 (1953).

6. A. A. Petrov and T. V. Yakovleva, Opt. i Spektr., F. 817 (1959). 7. R. Fuchs, J. Am. Chem. Soc, 3209 (1963). 8. R. A. Nyquist and W. J. Potts, Spectrochim. Acta., ig. 419 (1960); J. 0. Evans and R. A. Nyquist, ibid, !£, 1153 (1963). 9a. R. A. Nyquist, T. L. Reder, G. R. Ward and G.J. Kallos, Spectrochim. Acta, 2ZA. 541 (1971). 9b. R. A. Nyquist, T. L. Reder, F. F. Stec and G. J. Kallos, Spectrochim. Acta, 27A, 897(1971). 9c. R. A. Nyquist, Spectrochim. Acta 27A. 2513 (1971). 10. J. 0. D. Brand and J. K. G. Watson, Trans. Faraday. Society, ^6, 1582(1960). 11. G. W. King and D. Moule, Spectrochim. Acta, 12, 286 (1961). 12. V. A. Job and G. W. King, Can. J. Chem. 41. 3132 (1963). 13. 0. V. Berney, L R. Cousins and F. A. Miller, Spectrochim. Acta 19, 2019(1963). 14. Russell H. Sanborn, Spectrochim. Acta, 23A .1999 (1967).

15. J. Guilleme, M. Chabanel and B. Wojtkowiak, Spectrochim. Acta, 2ZA, 2355(1971). 16. D. R. Williams, J. Chem. Phys., ^,4578 (1971).

17. J. Bournay and Y. Marechal, J. Chem. Phys. 55.1230 (1971). 303 304 18. E. Augdahl, E. Kloster-Jensen and A. Rogstad, Spectrochim. Acta, 2M. 399(1974).

19. W. J. Balfour, D. Klapstein and S. Visaisnik, Spectrochim. Acta, 21A. 1085(1975). 20. N. M. D. Brown, W. J. Nelson, and D. G. Walmsley, J. Chem. Soc. Faraday Trans. II, Z5.32 (1979). 21. W. J. Balfour K. Beveridge, and J. 0. M. Zwinkels, Spectrochim. Acta, 25A, 163(1979). 22. J. Travert, J. 0. Lavalley and D. Chenery, Spectrochim. Acta, 35A. 291(1979). 23. W. K. Balfour and M. K. Phibbs, Spectrochim. Acta, 35A. 385 (1979). 24 R. Wellington Davis and M. 0. L Gerry, Can. J. Chem. £Q, 679 (1982). 25. R. F. Miller and R. F. Curi, Jr., J. Chem. Phys. M. 1847 (1961). 26. H. Takeo and 0. Matsumura, J. Chem. Phys., M, 4536 (1976). 27. L Pierce and LO. Krisher, J. Chem. Phys. 21, 875 (1959). 28. K. M. Sinott, J. Chem. Phys. M. 851 (1961). 29. H. N. Volltrauer and R.H. Schwendeman, J. Chem. Phys., M, 268 (1971). 30. K. P. R. Nair and J. E. Boggs, J. Mol. Stmct. 22, 45 (1976). 31. O. L Stiefrater and E. B. Wilson, Jr., J. Chem. Soc, 4761 (1956).

32. H. Karisson, J. Mol. Struct., 22. 227 (1976). 33. J. E. Katon and Th. B. Ben Kinney, Spectrochim. Acta, 2M. 877(1983). 34 G. Williams and N. L. Owens, Trans. Faraday Society II, £Z. 922(1971). 35 0. Y. Lin, A. Krantz, T. D. Goldfarb, J. Am. Chem. Soc, M. 9282 (1972). 36 Y Hamada, M. Tsuboi, M. Nakata and M. Tasumi, J. Mol. Spectres., mZ, 269(1984). 305 37. J. Fulara, S. Leutwyler, J.P. Maier, and U. Spittel, J. Phys. Chem., 2^,3190(1985).

38. J. E. Katon and N. T. McDeVitt, Spectrochim. Acta., 21, 1717 (1965).

39. D. G. Lister and J. K. Tyler, Spectrochim. Acta, 22A. 1423 (1972). 40. K. Kondo, J. Phys. Soc. Japan, 15, 307 (1960).

41. J. A. Howe and J. H. Goldstein, J. Chem. Phys., £2. 1223(1955). 42. H. Kim, R. Kellar and W. D. Gwinn, J. Chem. Phys. 2Z, 2748 (1962).

43. R. Wellington Davis and M. 0. L Gerry, J. Mol. Spectres., 5^. 407 (1976). 44. J. K. G. Watson, J. Chem. Phys., 42, 1935(1967); 42.4517 (1968). 45. G. Herzberg, "Infrared and Raman Spectra," Van Nostrand Reinhold Company, New York, NY (1945). 46. J. D. Cox and G. Pilcher, "Thermochemistry of Organic and Organometallic Compounds," Academic Press, New York, NY (1970). 47. J. B. Pedley and J. Rylance, Sussex-N.P.L. Computer Analysed thermochemical data: Organic and Organometallic Compounds. University of Sussex, Brighton, England (1977).

48a. T. L Flitcroft and H. A. Skinner, Trans. Faraday Soc, M, 47 (1958). 48b. J. E. Bennett, Ph. D. Thesis, University of Manchester, 1959. 48c E. Schjanberg, Z. Phys. Chem., 42, 317 (1938). 49. R. Fuchs and LA. Peacock, J. Phys. Chem.,_S2, 1975 (1979). 50a. W. J. Hehre, R. Ditchfield, L Radom, and J. A. Pople, J. Am. Chem. Soc, £2,4796(1970). 50b. M. R. Peterson and I. G. Cslzmadia, J. Mol. Struct., 22, 323 (1983). 51. J. D. Dill, A. Greenberg, and J. F. Liebman, J. Am. Chem. Soc, IM, 6814(1979). 52. H. Stafast, H. Pfister and J. Robert Huber, J. Phys. Chem., 22, 5074(1985). 306 53. J. A. Ladd, W. J. Orville-Thomas and B. 0. Cox, Spectrochim. Acta. 2Q. 1771 (1964).

54. J. A. Ladd, and W. J. Orville Thomas, Spectrochim. Acta, 22, 919(1966).

55. T. H. Thomas, J.A. Ladd, V. I. P. Jones, and W. J. Orville-Thomas, J. Mol. Struct., 2, 49(1969). 56. D. R. Herschbach and V. W. Laurie, J. Chem. Phys., 22. 458 (1961). 57. 0. W. Bock, M. Trachtman and P. George, Chem. Phys., 68. 143(1982).

58. 0. W. Bock, M. Trachtman and P. George, Chem. Phys., 22, 303(1981). 59. 0. W. Bock, M. Trachtman and P. George, J. Mol. Spectrosc, 24, 256(1980). 60a. P. Schuster and Th. Funck, Chem. Phys. Lett. 2, 587 (1968). 60b. K. Morokuma, H. Kato, Y. Yonezawa, and K. Fukui, Bull. Chem. Soc. Japan, 22, 1263 (1965). 61. Tae-Kyu. Ha, R. Mayer and Hs. H. Gunthard, Chem. Phys. Lett., 22, 17(1978). 62. P. Furet, G. Hallak, R.L Matcha and R. Fuchs, Can. J. Chem., 22. 2990(1985). 63. LM. Tack and S.G. Kukolich, J. Chem. Phys., Z2. 6512 (1983).

64. B.E. Turner, Astrophys. J, 122, L35 (1971). 65. B. Zuckerman, J. A. Ball and 0. A. Gottlier, Astrophys. J. 122. L41 (1971). 66. K. D. Tucker, M. L Kutner, and P. Thaddeus, Astrophys. J, 122, Ll 15 (1974). 67. L E. Snyder and D. Bhul, Bull. Amer. Astron., Soc.,J^, 388 (1971). 68. R. L Redington, J. Mol. Spectres. 22.171 (1977). 69 K. Tanaka and M. Yoshimine, J. Am. Chem. Soc, 122. 7655 (1980). 307 70. J. S. Binkley, R. A. Whiteside, K. Raghavachari, R. Seeger, D. J. DeFrees, H.B. Schlegel, M. J. Frisch, J. A. Pople, and L R. Kahn, "GAUSSIAN 82 Release H," Carnegie-Mellon Univeristy, Pittsburgh, 1982.

71. H. E. Hallam, Spectres. Proc Cent. 6th, 182 (1976).

72. L. Andrews, Adv. Infrared Raman Spectros._Z, 59 (1980). 73. A. J. Barnes, K. Szczepaniak, and W. J. Orville -Thomas, J. Mel. Stmct., 22, 39(1980). 74. S. Craddeck and A. J. Hineheliffe, "Matrix Isolation: A Technique for the Study of Reactive Inorganic Species," Cambridge University Press,Cambridge, UK, 1975. 75. D. W. Green, Gerald T. Reddy in "Fourier Tranferm Infrared Spectroscopy", eds. John R. Feraro and Louis J. Basile, Vel.1, Ch.1, Academic Press, New York, N. Y. 1978. 76. O. G. Peterson, D. N. Batcheller, and R. O. Simmons, Phys. Rev., 122,703(1966). 11. 0. S. Barrett and L Meyer, J. Chem. Phys., 42. 3502 (1965). 78. W. F. Giangue and J. O. Clayton, J. Am. Chem. Sec. 22, 4875 (1933). 79. E. K. Plyler, A. Danti, L. R. Blaine and E. D. Tidewell, J. Res. N.B.S.,24A,29(1960). 80. Robert J. Bell, "Introductory Fourier Transform Spectroscopy", Academic Press, New York, N. Y. 1972. 81. Peter R. Griffiths, "Chemical Infrared Fourier Transform Spectroscopy", Wiley-Interscience, New York, N.Y. 1975. 82. John R. Ferraro and Louis J. Basile (eds.), "Fourier Transform Infrared Spectroscopy: Application to Chemical Systems, Vol. 1, Academic Press, New York, N.Y. 1978.

83. E. B. Wilson, J. 0. Decius, and P. 0. Cross, "Molecular Vibrations," McGraw Hill Book Company Inc., New Yori<, N.Y. 1955.

84. T. Ueda and T. Shimanouchi, J. Spectrosc, 22, 350 (1968).

85. H. Goldstein, "Classical Mechanics", Addison-Wesley Pubishing Co. Inc., Cambridge, Mass (1950). 308 86. W. A. Seth Paul and G. Dykstra, Spectrochim. Acta, 23A. 2861(1967).

87. P. Mirone and P. Chiboli, Spectrochim. Acta, 1^, 1425 (1962).

88. J. Pliva , V. Spirko and D. Paporisek, J. Mol. Spectrosc, 22, 331 (1967).

89. K. Nakamoto, "Infrared Spectra of Inorganic and Coordination Compounds", Wiley, London, 1970, 2nd edition. 90. P. Gans, "Vibrating Molecules: "An Introduction to the Interpretation of Infrared and Raman Spectra," Chapman and Hall, London, 1971.

91. D. Steele, "Theory of Vibrational Spectroscopy," Saunders, London 1971. 92. L. H. Jones, "Inorganic Vibrational Spectroscopy," Vol. 1, Dekker, New York, N.Y. 1971. 93. L. A. Woodward, "Theory of Molecular Vibrations and Vibrational Spectrosocpy," Clarendon Press, Oxford, 1972. 94. I. M. Mills, "Force Constant Calculations for Small Molecules", in "Infrared Spectroscopy and Molecular Structure," ed. M. Davies, Elsevier, Amsterdam, 1963, Oh.5. 95. T. Shimanouchi, "The Molecular Force Field" in "Physical Chemistry: An Advanced Treatise," Vol. IV, ed. H. Eyring, D. Henderson, and W. Jost, Academic Press, New York, N. Y., 1970, Oh. 6. 96. J. L Duncan, "Force Constant Calculations in Molecules," in Mol. Spectrosc,3, 104 (1975), Ch.2. 97. U. Burkert and N. Allinger, "Molecular Mechanics," ACS Monograph 177, ACS, Washington, D.C., (198.2) Oh. 1,2. 98. William A. Guillory," Introduction to Molecular Structure and Spectroscopy," Allyn and Bacon, Inc., Boston, Mass., 1977.

99. B. Crawford and J. Overend, J. Mol. Spectry, 42, 307 (1964).

100. M. Gussoni and G. Zerbi, Chem. Phys. Letts., 2.145 (1968); Rend. Accad. Nazionale Lincei, Serie VIII. Vol. XL, fasc. 6 (1966).

101. I.M. Mills, Chem. Phys. Letts., 2, 267 (1969). 102. J. H. Schachtschneider, Shell Report. No. 57-65 (1966). r

309 103. W. Hagens and A. G. G. M. Tielens, J. Chem. Phys., Z2.4198 (1981).

104. R. M. Bentword, A. J. Barnes, and W. J. Orville-Thomas, J. Mol. Spectr.,24..391 (1980).

105. E. B. Whipple, J. H. Goldstein, L Mandell, G. S. Reddy, and G. R. McClure, J. Am. Chem. Soc, 21, 1321 (1959). 106. A. Schellenberger, W. Beer, and G. Oehme, Spectrochim. Acta, 21. 1345(1965).

107. Liang, Jim Chin-Kang, "Ph.D. Disseration, 1981", Texas Tech University, Lubbock (1981). 108. J. E. Katon and W. R. Feairheller, Jr., J. Chem. Phys., M. 1248 (1966). 109. R. L. Redington, Spectrochim. Acta. 21A, 1699 (1975). 110. R. L Redington, J. Mol. Spectr. 22,171 (1977). 111. J. K. Wilmhurst, J. Chem. Phys.,22,1171 (1956) 112. Kennedy, John R. , "M.S. Thesis", Texas Tech University, Lubbock (1974). 113. F. F. Bentley and M. T. Ryan, Spectrochim. Acta, 20A. 685 (1964). 114. P. Felder and Hs. H. Gunthard, Spectrochim. Acta, 22, 223 (1980) 115. S. Pinchas and I. Laulicht, "Infrared Spectra of Labelled Compounds", Academic Press, New York (1971). 116. A. Barnes, "Theoretical Treatment of Matrix Effects," in Vibrational Spectra of Trapped Species, H.E. Hallam, ed., Wiley, London, 1974. 117. E. Bauer and M. Magat, J. Phys. Radium 2, 319 (1938). 118. A. D. Buckingham, Proc Roy., Soc. A 248. 169 (1958): ibid. A255. 32 (1960); Trans. Faraday Soc, 22, 753 (1960). 119. J. G. David and H. E. Hallam, Spectre. Chim. Acta, 22A, 593 (1967). 120. J. O. Hirschfelders, 0. F. Curtiss, and R. B. Bird, Molecular Theory of Gases and Liquids, John Wiley and Sons, Inc., New Yori<, N.Y. 1954. r

310 121. A. J. Barnes and H. E. Hallam "Infrared Studies of Matrix Isolated Species", Quart. Rev. Chem. Soc, 22, 392 (1969). 122. J. E. Lennard-Jones and A. E. Ingham, Proc. Roy. Soc, 107A. 636 (1925).

123. H. Friedmann and S. Kimmel, J. Chem. Phys., 42, 3925 (1965). 124. H. E. Hallam and G. F. Scrimshaw in "Vibrational Spectroscopy of Trapped Species: Infrared and Raman Studies of Matrix-Isolated Molecules Radicals and Ions," H. E. Hallam (ed.), John Wiley and Sons, 1974.

125. G. C. Pimentel in H. E. Hallam in M. J. Wells (ed.). Spectroscopy, Proc 3rd Int. Petroleum Hydrocarbon Research (jroup Conference, Inst, of Petroleum, London, 1962.

126. D. N. Batchelder, D. L Losee, R. O. Simons, Phys. Rev., 122, 767(1967). 127. L. Pauling, "Nature of Chemical Bond", Cornell University Press, 1948, 2nd edition, p. 189 128. G. C. Pimentel and S. W. Charles, Pure Appl. Chem., [, III (1963). 129. T. Clari<, "A Handbook of Computational Chemistry: A Practical Guide to Chemical Stmcture and Energy Calculations," Wiley-Interscience, 1985. 130. I. G. Cslzmadia, "Theory and Practice of MO Calculations in Organic Molecules," Elsevier (1976). 131. W. H. Lehre, L. Random, P.v.R. Schleyer, and J.A. Pople, "Ab Initio Molecular OriDital Theory," Wiley-Interscience, New York, N.Y., 1985. 132. E. R. Davidson and D. Feller, Chem. Rev., 22, 681 (1986). 133. P. Pulay, Mol. Phys.. H, 197 (1969). 134. B. Andes Hess, Jr., L J. Schaad, P. Carsky, and R. Zahradnik, Chem. Rev.,22, 709(1986). 135. J. M. Bowman, Ace Chem. Res., 12, 202 (1986).

136. F.V. David Lewis, Ace Chem. Res., 22, 1111 (1986).

137. M. J. S. Dewarand D. H. Lo, J. Am. Chem. Soc. ,24, 5296 (1972). 138. "Modern Theoretical Chemistry: Applications of Electronic Structure Theory," Vol. 4, ed. H. F. Schaefer III. F

311

139. P. Pulay, G. Forgassi, G. Porgor, J. E. Boggs, and A. Vargha, J. Am. Chem. Soc, 122, 7037 (1983).

140. P. Pulay, in "Applications of Electronic Theory" (H.F. Schaefer III, ed.), pp. 153-185, Plenum, New York, 1977.

141. C.W. Bock and M. Trachtman, P. George, J. Mol. Spectrosc, 22, 131 (1980).

142. N. Tanaka, Y. Hamada, M. Tsuboi, Chem. Phys., 24, 65 (1985). 143. P. Russegger and H. Lischka, Chem. Phys., 86, 21 (1984). 144. P. Ruelle, Chem. Phys., 110.263(1986). 145. U. Bmhimann and J. R. Huber, Chem. Phys. Letts., 22, 405 (1982). 146. U. Bruhlmann, J. R. Huber and P. Russeger, Chem. Phys. Letts., Z2. 179 (1980); 24, 479(1981). 147a. M. E. Schwartz, E. F. Hayes and S. Rothenberg, J. Chem. Phys., 22,2011 (1970). 147b. 0. B. Moore and J. 0. Weisshaar, Annu. Rev. Phys. Chem., 24, 525 (1983) 148. M. Dupuis, W. A. Lester Jr., B. H. Lengsfield III, and B. Liu, J. Chem. Phys., 22, 6167 (1983) 149. P.S. Skell, F. A. Fagone and J.K. Klabunde, J. Am. Chem. Soc, 24,7862(1972). 150. C. E. Dykstra and H. F. Schaefer III, J. Am. Chem. Soc, 100. 1378 (1978). 151. Y. Osamura, H. F. Schaefer III, S. K. Gray, and W. H. Miller, J. Am. Chem. Soc, 122. 1904(1981) 152. 0. Reiser, F. M. Lussier, 0. 0. Jensen, and J. I. Steinfeld, J. Am. Chem. Soc, 121. 350 (1979). 153. 0. Reiser and J. I. Steinfeld, J. Phys. Chem., 24. 680 (1980) 154. A.E. Bmno, R.P. Steer, J. Chem. Phys. , Z2, 6600 (1983).

155. J.D. Goddard, Can. J. Chem.,22.1910 (1985). 156. M.P. Conrad and H.F. Schaefer III, J. Chem. Phys..100. 7820 (1978). r^

312

157. R. W. Wetmore and H. F. Schaefer III, J. Chem. Phys., 22. 1648 (1978) 158. E. Abramson, R. W. Field, D. Imre, K. K. Innes, and J. L Kinsey, J. Chem. Phys.. 22. 453 (1985). 159. H. Basch. Theor. Chim. Acta, 22, 151 (1973). 160. H. Stafast, H. Bitto, and J. R. Huber, J. Chem. Phys., 79, 3660 (1983); H. Bitto, H. Stafast, P. Russegger, and J. R. Huber, Chem. Phys., 24. 249 (1984); J. Muhlbach, M. Dubs, H. Bitto, and J. R. Huber, Chem. Phys. Lett., HI, 288 (1984).

161. P. Russegger and H. Lischka, Chem. Phys.,22. 31 (1983); P. Russegger and J. R. Huber, Chem. Phys., 22. 33 (1984). 162. U. Bruhlmann, P. Russegger, H. Stafast, and J. R. Huber, Ber. Bunsenges. Phys. Chem., 22. 261 (1985). 163. J. Pacansky and J. Bargon, J. Am. Chem. Soc, 2Z. 6896 (1975) and references therein. r

APPENDICES

Al: CALIBRATED VIBRATIONAL FREQUENCIES OF PROPIOLIC ACID (PA).

A2: CALIBRATED VIBRATIONAL FREQUENCIES OF DIDEUTERATED PROPIOLIC ACID (PA-D2).

A3: CALIBRATED VIBRATIONAL FREQUENCIES OF MONODEUTERATED PROPIOLIC ACID (PA-OD).

Bl: OBSERVED WATER FREQUENCIES.

313 r

314 Appendix Al: Calibrated Vibrational Frequencies of Propiolic Acid (PA)

Ar CO N2 Ne Assign.

482.0(w) xOco(Ne/PA) 529.5(m) 530.0(s) '^COH 533.8(w-m) '^COH 537.6(vw) 540.5(w) TCOH 545.0(vw) 568.5(w,sh) 572.4(s-vs) ^0=0 573.3(s-vs) ^0=0 579.5(w) 1 586.0(w) PC-COOH 586.3(w) 586.3(w) PC-COOH 1 587.5(w) PC-COOH

607.0(m-s) 609.7(w,sh) 609.2(vw) 612.5(w) 615.0(vw) 615.8(vw) 620.8(vvw) 621.5(vvw) 621.7(vvw) 637.8(w) 645.5(w,sh) 650.0(m-s) ^0 652.2(m-s) 651.5(m-s) SQCG 656.5(vw,sh) 656.2(vw-w) 659.4(m-s) ^000 661.5(w,sh) 663.5(m) 665.0(m) ^000 666.0 668.2(vw) 690.6(w-m) ^-c=c 693.8(m) 699.0(w) 708.0(w.b) 720.0(vw-w) 736.7(m) 748.5(vw) 751.3(w-m) 754.6(S) ^c^ 755.2(s) '^HC^ 757.0(w) 758.7(vw,sh) 758.5(s) ^0^ 762.0 315 Appendix Al continued

Ar CO N2 Ne Assign.

762.4(s) 763.0(vw.sh) n^c^ 774.0(w) 815.2(m-s) VQ-C 817.5(m-s) 817.8(vw.sh) ^C-C 820.5(w-m) ^C-C 823.0(vw) 825.0(vw.b) 826.3(m) 'C-C 835.0(w,b) 841.2(w.b) 928.5(vw,b) 1090.0(vw,vb) 1093.1 (s) ^COCCHdme r 1103.5(w,b) 1110.5(w,sh) 1112.0(w-m) 1112.2(m) 1120.5(w-m) 1137.7(vw) 1137.0(w) 1147.2(vs) vc-O C'S PA 1148.5(vs) 1153.0(m^) VQ^OSPA 1154.0(s) 1158.5(w-vw) 1161.4(m) 1163.0(s) VQ^ trans PA 1166.4(w) 1170.0(w) 1171.3(w-m) 1172.6(w-m) 1174.7(vw) 1178.8(s-vs) ^CO/COH 1181.0(w-vw) 1187.0(w-m,b) 1212.4(m-s) 1218.0(vw,sh) 1235.0(S-vs) 8(X>HPAciTier 1241.0(m-s) ScoiPAdimer 1243.5(w) 2v 8 1248.8(vw) 1251.0(vw) 1257.5(vw) 1262.9(w-m.b) 1276.4(m) 1277.0(m) ^CQPACD 316 Appendix Al continued

Ar CO N2 Ne Assign.

1281.8(vw) 1281.9(m-s)

1288.0(vvw) 1289.0(w) 1290.0(m-s) VQ_Q (trans) 1291.4(w-m) 1297.1 (w.sh) 1297.0(w.sh) 1300.5(vw.sh) 1301.8(S-VS) VQ.Q (cis) 1303.0(s) VQ.o(cis) 1312.0(w)

1315.0(w) 1323.8(vw) 1326.8(sh) 1328.4(s) vc.o(cis) 1342.0(m) VQ.Q (cis) 1350.5(w) 1362(m) 1360.0(s) V7 + V.14 1370.3(w) 1381.5(w) 1384.8(w,b) 1406.2(w) 1406.0(m-s) 1400.6(w-m) V3,HDO 1411.7(w) 1411.0(m) 1414.8(w) 1427.5(vw) 1432.5(vw) 1510(vw) 1513.0(b) 1519.0(b) 1557.0(w-m) 1573.6(m) 1589.5(w,sh) 1593.0(m) 1599.5(vw) 1600.0(s) V3, H2O 1609.5(m) 1619.0(vw) 1625.0(m) 1630.5(w) 1638.5(vw,vb) 1662.6(vw.b) 1695.8(w) 1702.0(vw) 1707.0(m) 1700.8(m,b) 317 Appendix Al continued

Ar CO N' Ne Assign.

1713.2(vw) 1715.9(vw) 1725.9(s) ^C=Oc*ner 1727.3(w) 1730.0(vw) 1734.6(w) 1734.5(w.sh) 1738.5(w) 1739.0(w) 1742.3(w) 1741.3(S) ^0=0 1746.2(vs) ^0=0 1748.5(vs) ^0=0 1750.5(sh) 1752.7(s) ^0=0 1753.7(vs) 1756.0(vw,sh) ^c=o 1758.2(w,b) 1759.0(vvs) 1767.0(w) 1779.0(vw) 1781.0(w,b) 1787.0(w) 1794.0(m) 1815.0? 1817.0? 1994.0 ? 1996.0 ? 2030.0 2113.0(vw) 2117.0 2124.0 2127.0(m) ^c-c 2129.0 2130.0 2136.0 2140.0(S-vs) 2139.6(ws) 2139.5(vs) ''C^C 2144.5(w,sh) 2147.0(vw,sh) 2147.0(w) 2150.0(vvw,sh) 2152.0(vw) 2290.0(vw) 2328.0 2332.0 2350(vs) 2348.0(s) CO matrix 2352.0(vs) 2541.0(vvw) 2670.0(vw) 318 Appendix Al continued

Ar CO No Ne Assign.

3111.5(w,b) 3157.0(vw,vb) 3184.8(vw,vb) 3242.0(w-m) 3260.8(w-m) 3289.0(vw) 3302.5(w-m) VQ.|_| 3308.8(vw,sh) 3307.0(m) VQ.^ 3310.0(sh,b) 3315.6(m-s) VQ.|_| 3324.6(sh) 3325.0(s) VC.H 3328.0(s,sh) 3340.0(w) 3409(w.b) 3438.8(s) VQ.H (Cis) 3497.8(w) 3504.0(vw) 3517.5(vw) 3520.5(m-s) VQ.H (cis) 3531.0(w) 3531.5(m-s) VQ4_| (trans) 3551.0(s) vo.H(cis) 3569.0(vw,sh) 3574.0(m) VQi^Orans) 3580.0(s-vs) VQ-H (cis) 3599.5(vw,vb) 3599.0 (vw) 3636.0 ? 3706.0(m) 3711(m,b) 3729.0(vw,vb) 3758.1 (w-m) 3780.0(w-m) 3782.5(vw) 319 Appendix A2: Calibrated Vibrational Frequencies for Dideuterated Propiolic Acid (PA-D2)

Nitrogen Argon Carbon monoxide Assignment

406.3 410.2 421.9 436.4(w) 440.0(w) 453.5(w) 460.4(w) XQD DCCCOOD 478.3(w) 480.5(w) 483.6(w-m) PC-COOD 532.5 OCO wag PA-D2 587.5( WW) PC-COOH 609.0(w,sh) 609.4( vw.sh) 609.0(w,sh) "c=c/^cc=o 612.0(m-s) 612.5(m) 612.0(m) g/TtD-C^C 647.7 OCO wag 649.0(w) 650.2 654.7( w.sh) OCO scissor 659.0(w) OCO scissor 660.0(w) 661.4(s) OCO scissor 667.0(m-s) 679.7( w-m) 718.0( w-m) 729.7( w-m) 741 (ww) 825.0( m) VQ.Q DCCCOOH 833 (WW) 847.0(m) ^C-C DCCCOOD 847.5(m) VQ,Q DCCCOOD 850.7(s-vs) 851.0(w) VQ,Q DCCCOOH 851.3(m) VQ.Q DCCCOOH 932.3( m-s) SQOD DCCCOOD 933.0(m) SQOD DCCCOOD 933.5(m) ^COD DCCCOOD 941.3(m) SQOD DCCCOOD 941.0(m) SQOD DCCCOOD 1042.0(vs) SQOD HCCCOOD 1044.0(m-s) 1044.0(s) ScOD HCCCOOD 1046.0(m) SQOD HCCCOOD 1047.9(s) SQOD HCCCOOD 1050.0(vvw,sh) 1088.9(vw,sh) 5COH(DCCCOOH)2 1090.0(m,vb) 1090.0(m) 5QOH(DCCCOOH)2 r

320 Appendix A2 continued.

Nitrogen Argon Carbon monoxide Assignment

1112.0(w) 1112.8(vb) 1115.9(vb) 1116.0(w) iii6.5(w-m) 1121.0(m) 6QQ_, DCCCOOH 1125.0(m) ^COH DCCCOOH 1127.2(s) 6QOH DCCCOOH 1141.3(vw) 1144.0(w) 1146.6(vw) 1148.0(w) 1159.0(w) 1160.0(w-m) 1160.6 1174.7 1174.7 V2 , D2O 1179.5(vs) 1181.0(m) V2, dimer D2O 1193.0(m) 1199.0(w) 1206.0(w) 1215.0(w) 1276.4 v Q.O DCCCOOD 1281.0(w) 1343.2(w) 1344.1 (m) C-O DCCCOOH 1350.0(w) 1352.8(m) 1352.8(w) v Q.0 DCCCOOH 1354.0(w) v Q.0 DCCCOOH 1358.0 1360.5(w) 1366.1 (w-m) 1368.1 (w) 1377.0(w) V2Ar/HD0 1384.0(w) V2CO/HDO 1389.7(w-m) 1398.8(s) V c-0 (PA-D2)2 1403.3(S) V Q.0 (PA-D2)2 1405.9(S) V C-0 (PA-D2)2 1407.0(s) V c-0 (PA-D2)2 1413.0(w-m) 1419.0(w) 1427.2(w-m) 1431.0(w) 1445.8(vw) 1445.0(w) 1475.0(w) 1494.0(w) 1494.5 r

J- Appendix A2 continued.

Nitrogen Argon Carbon monoxide Assignment

1591.2(m-s) vc=o (PA-D2)2 v2, Ar/H20 1598.0(vs) vc=0 (PA-D2)2 1600.0(m) 1604.0(s,sh) vc=0 (PA-D2)2 1607.0(w) 1618.0(m) 1622.8(vvb) 1629.0(m) 1637.0(m) 1658.0(m) 1715.1 (m-s) VQ^O DCCCOOD 1721.3(w-m) VQ^O DCCCOOH 1738.0(s) VQ^O DCCCOOD 1743.0(s,sh) VQ^O HCCCOOD 1750.0(m-s) VQ_O DCCCOOH 1753.6(m.sh) VQ^O HCCCOOH 1971.0 1974.2(w-m) 1975.0(m) VQ^ DCCCOOD 1977.5(s,sh) \Q-Q DCCCOOD 1982.3(m-s) VC=C DCCCOOD 1985.4(ws) VQ^ DCCCOOH 1986.0(m,b) \Q^ DCCCOOH 2021.3(vw) 2089.0(w) 2093.0(w) 2111.0(w) 2132.5(vw-m) 2138.0(w) 2140 .0(ws) y^Q=Q dimer 2143.0(s) '^Q=Q dimer 2147.9(m-s) ^C^C <^if^6r 2200.0(b.m-w) 2275.0(vvw,vb) 2282.2(w) 2327.8(m-s) voo(DCCCOOD)2 2340.7 Voo(DCCCOOD)2 2332.0(m-s) voo(DOCCOOO)2 2345.0(w) 2346.0(m) 2348.0(vvs) voo(DCCCOOD)2 2351.1 (vw.sh)) 2456.0(vw.vb)) 322 Appendix A2 continued.

Nitrogen Argon Carbon monoxide Assignment

2461.6(w) 2481.0(m) 2485.0(w,vb) 2489.7(m) 2532.0(m) 2566.0(m) 2578.0(m) VQD DCCCOOD 2588.4(m) VQD DCCCOOD 2590.6 VQQ DCCCOOD 2596.0 v^ Ar/D20 2599.0(m) VQD DCCCOOD 2600.5(w) 2607.0(w-m) 2614.3(m-s) ^C-D 2616.8(w,sh) 2620.2(m) ^C-D 2628.0(m) ^C-D 2643.0(m) ^C-D 2649.0(w) 2654.3(m) V2, N2/D2O 2676 .0(w,sh) 2692.0(m) 2697.5(w) 2698.0(m) 2700.9(w) 2704(6).5(vs) V2, N2/HDO 2716.0(m) 2725.0(w) 2726.0(m) 2733.0(w) 2738.6(w-m) 2751.0(m) 2757(8).4(m) 2765.0(vs) 2785.0(vw) 2784.0(w) 2808.0(vvw,b,sh) 2818.0(w) 2850.0(s.sh) 2858.5(w-m) 2862.6(w-m) 2864.0(w) 2872.0(w) 2874.3(w-m) 2882.0(w) 2885.9(m) 323 Appendix A2 continued.

Nitrogen Argon Carbon monoxide Assignment

2903.0(w) 2916.6(m.b) 2916.7(m) VQ_|(DCCC0CH)2 2936.0(w) 2943.0(w) VOH(DCCCOOH)2 2945.0(m) VQQ(DCCC0CH)2 2981.7(m) VOH(DCCCOOH)2 2982.3 2993.4(m) VQ_|(DCCC0CH)2 3013.8(w) 3148.0 3217.0 3226.0 3282.0 3330.0(vw) 3359.0(w) 3362.5(w,b) 3440.0(w) 3464.0(w) 3495.0(w) VQH DCCCOOH 3501.1(w.b) 3503.8(w) 3509.6(w) 3511.7(w) 3515.6(m) VQH DCCCOOH 3523.2(m-s) VQH DCCCOOH 3528.0(w) 3534.0(m) 3550.1 (m) 3564.5(m) 3595.0(w) v.|, H2O dimer 3627.5(m) 3632(3).5(s) v.|, H2O dimer 3670.3(m-s) 3694.0(w) V3, Ar/HDO 3703.0(w) V3, Ar/H20 3704.5(m-s) vi,N2/H20 r

324 Appendix A3: Calibrated Vibrational Frequencies of Monodeuterated Propiolic Acid (PA-OD).

Nitrogen (M/S = 3500/1) Argon (M/S = 4000/1) Assignment

463.5(m-s) 471.0(m) ^COD PA-OD 507.5(w) (506.5 at 18.9K) 508.0(m) ^C=0 i 510.0(w) 517.0(w) 522.5(m) 535.0(w) 535.5(w) ^COH PA 586.5(w) PC-COOD(H) 587.0(w) 608.5(w-m) ®C=0 617.3(w.b) 623.5(w-m,b) 652.5(m) ^OCO 660.0(w?) 664.0(w?) 668.0(m) 692.5(w,b) ^HC^C 696.0(w) 708.0 b 708.0 ^HC^C 737.0(w-m) 753.3 w.sh 757.0(m) (757.5 at 18.9K) 756.5(m-s) '^HC^C 760.3(w,sh) 767.8(w,b) 770.0(ww) 774.0(vvw) 820.5(w.b) ^C-C 820.5(w) (819.5 at 18.9K) ^C-C 834.5(w) (835.5 at 18.9K) VQ,Q PA-OD 1006.0 (vvb.sh) 1014.0 m SQOD PA-OD 1011.0(s) (1010.0 at 18.9K) SQOD PA-OD 1016.5(m) 1121.0(w-m) 1115.5(m) 1152.5(m) (1153.3 at 18.9K) 1152.5(vvw,vvb) SCOH PA 1159.5(m-s) ^COH PA 1178.8(S-vs) V3. N2/D2O 1181.1? V2, Ar /H2O dimer 1241.7(w-m) 1242.5(m) 1257.7(w,b) 1260.5(w-m,b) r

325 Appendix A3 continued

Nitrogen (M/S = 3500/1) Argon (M/S = 4000/1) Assignment

1263.5(m) 1275.7(s) ^C-0/PA-OD 1277.2(vs) ^C-0/PA-OD 1298.8 1313.2(vw-w) 1330..0(w-m,vb) 1339..0(m,vb) 1349.0 (vb.sh) 1354.5 (m) 1385.0(vvb) 1399.0(m) V2. Ar/HDO 1406.0(s) 1408.3(m-s) V2, HDO dimer 1418.5(w,b) 1432.8(m) 1594.1 (w) V2, H2O dimer 1599.3 (m) (1598.5 at 18.9K) 1600.0(m) V2, H2O dimer 1613.5(vw,vb)? 1701.3(vw,vb) 1714.4(vw,b) 1720.0(w.b,sh) 1723.8(w,b) 1727..4(m) 1740..0(s-vs) (1738.5 at 18.9K) VQ^O- PA-OD 1740.5(s-vs) vc=0' PA-OD 1743.2(s,sh) 1749.5(m) 1751.1 (m-s) vc=0' PA 1752.5 (w-m) (1748.5 at 18.9K) vc=0- PA

1763.5(w,vb) 1768.0(vw.vb) 1860(vw) 2128.6(w,sh) 2138.0 (m) VQ^Q PA-OD 2138.0(m) 2142.0 (vw.sh) ^CHC PA-OD 2601.0(m) ^OD/PA-OD 2602.0(m) ^OD/PA-OD 2608.0(w,sh) 2609.5(m)(2607at 18.9K) ^OD/PA-OD 2622.0(w,b) V1.D2O 2656.0 (w-m) 2662.0(w-m) 2707.5(m) 2711.6(m) v.,. HDO 2728.0(b,w?) 2740.0(w,sh) ''

326 Appendix A3 continued

Nitrogen (M/S = 3500/1) Argon (M/S = 4000/1) Assignment

2743.4(w) V3, D2O dimer 2760.0(vw.b.sh) 2765.5(m) 2770.6(m-s) v^. D2O 3307.0(m) ^H-C=C PA-OD 3308.5(vw.sh) 3313.0(m) ^H-C-C PA-OD 3325.6(w.b,sh) 3524.0(m.b) VO-H PA 3532.0(m) ^0-H PA

3682.5 3687.2(m) V3, H2O dimer 3730.0 3732.2(w-m) V3. H2O r

327 APPENDIX B

Appendix Bl: Observed Water Frequencies (cm"^).

Assignment vi V2 V3

Matrix / Species

Ar/H20 3574.5 1594.1 3732.2

Ar/HDO 2711.6 1399.0 3687.2

Ar/D20 2662.0 1190.0 2770.6

N2/H2O 3730.0 1598.5

N2/HDO 3682.5 2707.5 1406.0

N2/D2O 2766.5 2657.5 1178.8(1179.2)

CO/H2O 3598.6(dimer) 1599.5 3705.7