Citation
Laboratory Studies of Astrophysical Molecules

Material Information

Title:
Laboratory Studies of Astrophysical Molecules
Creator:
WANG, HAIYAN ( Author, Primary )
Copyright Date:
2008

Subjects

Subjects / Keywords:
Absorption spectra ( jstor )
Anions ( jstor )
Atoms ( jstor )
Carbon ( jstor )
Electronics ( jstor )
Infrared spectrum ( jstor )
Laser spectroscopy ( jstor )
Photolysis ( jstor )
Spectral bands ( jstor )
Vibrational spectra ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Haiyan Wang. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
12/31/2006
Resource Identifier:
443621810 ( OCLC )

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

LABORATORY STUDIES OF ASTROPHYSICAL MOLECULES By HAIYAN WANG A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2005

PAGE 2

Copyright 2005 by Haiyan Wang

PAGE 3

This dissertation is dedicated to my family, especially my parents and my wife for their wholehearted support.

PAGE 4

iv ACKNOWLEDGMENTS There are many individuals I would like to thank for their sup port during my four years of studies at UF. First of all, I would like to expre ss my great appreciation to my supervisor, Dr. Martin Vala, fo r his patient instruction, help ful advice and discussions, continuous support and encouragement during th e past four years. His attitudes to research and his earnest pursu it of the truth of science have benefited me immensely. Special thanks are given to Dr. Jan S zczepanski who not only instructed me on many topics but shared his wisdom of life and science. Several results presented in this dissertation are a consequence of effi cient teamwork between Jan and me. Dr. Philip Brucat was a scientific mentor and experimental motivator for me. He gave me insight about theoretical modeli ng and laser spectroscopy. His ambition for science and passion for research have intrigue d my enthusiasm for scientific exploration. My thankfulness also goes to Dr. John Eyler. Part of my research has been conducted using his instruments, which coul d not have been accomplished without his constant willingness to shar e his experience and knowledge. I also acknowledge my family and my friends for their support during all my school years. Especially, I want to thank my parents for their inspir ation and sacrifice. Finally, I thank my wife, Xihong Wu, for he r love, encouragement, and understanding throughout the past years.

PAGE 5

v TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES...........................................................................................................viii LIST OF FIGURES..........................................................................................................xii ABSTRACT....................................................................................................................xvi i CHAPTER 1 ASTROPHYSICAL BACKGROUND........................................................................1 The Interstellar Medium...............................................................................................1 Unidentified Interstellar Infrared Emission Bands.......................................................4 Diffuse Interstellar Absorption Bands..........................................................................6 2 THEORETICAL AND EXPE RIMENTAL METHODS.............................................8 Computation Details.....................................................................................................8 Geometry Optimization and Vibrational Frequencies...........................................8 Transition State Structure......................................................................................9 Vertical Excitation Energies for PAHs................................................................10 Experimental Details..................................................................................................11 Laser Ablation.....................................................................................................12 Electrical Discharge.............................................................................................12 Electron Ionization of Dibenzo[b,def]chrysene..................................................13 3 INFRARED ABSORPTION SPECTROSCOPY OF CARBON SULFUR CLUSTERS................................................................................................................15 Introduction.................................................................................................................15 The Linear C7S Carbon-Sulfur Cluster.......................................................................17 Theoretical Predictions........................................................................................17 Experimental Results...........................................................................................21 Band assignment for the 2( ) vibration......................................................22 Isotopomer band assignments for the 3( ) vibration..................................25 Isotopomer band assignments for the 5( ) vibration..................................28 Theoretical Calculations fo r Carbon-Sulfur Clusters CnS and SCnS ( n 29)............31

PAGE 6

vi Energies and Structures.......................................................................................31 Vibrational Properties..........................................................................................36 Dissociation Channels.........................................................................................42 The Linear CnS ( n = 2, 6) and CnS2 ( n = 7, 9, 11, 13, 15) Carbon-Sulfur Clusters....45 CnS and CnS2 Vibrational Frequenc y Coincidence Plots....................................46 The C2S Carbon-Sulfur Cluster...........................................................................52 The C6S and C7S2 Carbon-Sulfur Clusters..........................................................56 The CnS2 ( n > 7) Carbon-Sulfur Clusters............................................................59 The C9S2 carbon-sulfur cluster.....................................................................61 The C11S2 carbon-sulfur cluster....................................................................61 The C13S2 carbon-sulfur cluster....................................................................62 The C15S2 carbon-sulfur cluster....................................................................62 Summary.....................................................................................................................64 4 INFRARED ABSORPTION SPECTROSCOPY OF THE XENON CARBON CLUSTERS................................................................................................................67 Introduction.................................................................................................................67 Theoretical Results.....................................................................................................68 Experimental Results..................................................................................................74 Infrared Absorption of 12/13C2Xe in Argon Matrices...........................................75 Infrared Absorption of 12/13C3Xe in Argon Matrix..............................................77 Generation of CnXe ( n = 3, 5, 7, 9) Species........................................................79 Summary.....................................................................................................................81 5 REACTION OF TRICARBON CLUS TER WITH BENZENE AND AMMONIA..83 Introduction.................................................................................................................83 Reaction of C3 with Benzene......................................................................................85 Experimental IR Spectra......................................................................................85 Stable Structures..................................................................................................87 Predicted Infrared Frequencies............................................................................91 Possible C3 + C6H6 Reaction Pathways...............................................................94 Photolysis Experiments.......................................................................................98 Formation of the 1:2 Complexes.......................................................................102 Reaction of C3 with Ammonia..................................................................................105 Results and Discussion......................................................................................105 Summary...................................................................................................................112 6 VIBRATIONAL AND ELECTRONIC ABSORPTION SPECTROSCOPY OF DIBENZO[B,DEF]CHRYSE NE AND ITS IONS...................................................114 Introduction...............................................................................................................114 Structure and Energetics...........................................................................................115 Neutral DBC.............................................................................................................116 Infrared Absorption Spectra..............................................................................116 Electronic Absorption Spectra...........................................................................119

PAGE 7

vii DBC Cation and Anion.............................................................................................122 Infrared Absorption Spectra..............................................................................122 Electronic Absorption Spectra...........................................................................128 Astrochemistry Implications.....................................................................................130 The Unidentified Infrared (UIR) Emission Bands............................................131 The Diffuse Interstellar Bands (DIBs)..............................................................133 Summary...................................................................................................................134 7 ANHARMONICITY OF POLYCYCL IC AROMATIC HYDROCARBONS.......135 Introduction...............................................................................................................135 Anharmonicity of Naphthalene.................................................................................136 Infrared Active Vibrational Absorption............................................................137 Raman Active Vibrational Absorption..............................................................143 Anharmonicity of Phenanthrene and Anthracene.....................................................144 Summary...................................................................................................................147 8 CONCLUSION AND FUTURE WORK.................................................................148 Carbon Sulfur Clusters and Reac tion of Small Carbon Clusters..............................148 Polycyclic Aromatic Hydrocarbons..........................................................................149 Future Work..............................................................................................................150 LIST OF REFERENCES.................................................................................................156 BIOGRAPHICAL SKETCH...........................................................................................171

PAGE 8

viii LIST OF TABLES Table page 1-1 List of average chemical abundances pe r trillion atoms in several places of the Universe.....................................................................................................................2 1-2 List of identified interstellar molecules......................................................................3 1-3 Comparison of the electronic transi tions of PAH cations with DIBs........................7 3-1 Energies E (Hartrees), zero point energies ZPE (kcal/mol), dipole moments µe (Debyes), bond lengths (Å ), rotational constants Be (GHz), harmonic vibrational frequencies (unscaled, cm 1) and integral IR intensities (km/mol, in parentheses) calculated for linear C7S at various levels of theory...........................18 3-2 Energies E (Hartrees), dipole moments µe (Debyes), bond lengths (Å), and rotational constants Be (GHz), calculated for linear C7S at various levels of theory, all with a cc-pVDZ basis set........................................................................19 3-3 Comparison of experimental (Ar ma trix) and most intense calculated (B3LYP/cc-pVDZ) IR mode frequencies (cm 1) for linear CnS ( n = 1–9) and SCnS ( n = 1–9) carbon-sulfur clusters in their electronic ground states..................20 3-4 Comparison of observed (Ar matrix, 35 K) and calculated (B3LYP/6-311G* and B3LYP/cc-pVDZ) isotopomer frequencies (cm 1) of 2 mode for all-12C, singly-13C and all-13C substituted 12/13C7S linear carbon-sulfur clusters.............................23 3-5 Comparison of observed (Ar matrix, 35 K) and calculated (B3LYP/6-311G* and B3LYP/cc-pVDZ) isotopomer frequencies (cm 1) of 3 mode for all-12C, singly-13C and all-13C substituted 12/13C7S linear carbon-sulfur clusters.............................26 3-6 Comparison of observed (Ar matrix, 35 K) and calculated (B3LYP/6-311G* and B3LYP/cc-pVDZ) isotopomer frequencies (cm 1) of 5 mode for all-12C, singly-13C and all-13C substituted 12/13C7S linear carbon-sulfur clusters.............................29 3-7 Energies En (Hartrees), zero point energies ZPEn (kcal/mol), spin contamination and rotational constants Be (GHz) for CnS and CnS2 clusters, plus dipole moments µe (Debyes) for CnS clusters, calculated at the B3LYP/6-311G(d) level.......................................................................................................................... 31

PAGE 9

ix 3-8 Energies E (Hartrees), dipole moments µe (Debyes), bond lengths (Å), and rotational constants Be (GHz), calculated for linear C2S, C6S, and C7S2 clusters at B3LYP/6-311G(d), B3LYP/cc-pVDZ, B3LYP/aug-cc-pVTZ, MP2/cc-pVDZ and CCSD(T)/cc-pVDZ levels of theory.................................................................33 3-9 Calculated (B3LYP/6-311G(d )) vibrational frequencies (cm 1, unscaled) for linear CnS ( modes; n = 1–29) carbon-sulfur clusters in thei r electronic ground states......................................................................................................................... 38 3-10 Calculated (B3LYP/6-311G(d )) vibrational frequencies (cm 1, unscaled) for linear SCnS ( u modes; n = 1–29) clusters in thei r electronic ground states............39 3-11 Comparison of experimental (Ar ma trix) and calculated (B3LYP/6-311G(d)) most intense IR mode frequencies (cm 1) for linear CnS ( modes; n = 1–7) and SCnS ( u modes; n = 1–7) carbon-sulfur clusters in their electronic ground states......................................................................................................................... 40 3-12 Comparison of experimental (Ar ma trix) and calculated (B3LYP/6-311G(d)) most intense IR mode frequencies (cm 1) for linear CnS ( modes; n = 1–17) and SCnS ( u modes; n = 1–17) carbon-sulfur clusters in their electronic ground state.......................................................................................................................... .49 3-13 Comparison of observed (Ar matrix, 35 K) and calculated (B3LYP/6-311G(d) and B3LYP/cc-pVDZ) isotopomer frequencies (cm 1) of 1, and 2 modes for all-12C, singly-13C and all-13C substituted 12/13C2S linear carbon-sulfur clusters, as well as 4, 7, 10, and 12 modes for all-12C and singly-13C substituted 12/13C6S, 12/13C7S2, 12/13C13S2, and 12/13C15S2 clusters, re spectively...........................54 4-1 Rotational constants Be (GHz), dipole moments µe (Debyes), harmonic vibrational frequencies (unscaled, cm 1), integral IR intensities (km/mol, in parentheses), energies E (Hartrees), ze ro point energies ZPE (kcal/mol), and binding energies Eb (kJ/mol), calculated for CnXe ( n = 2, 3) at MP2 level with LJ18-Xe/6-311++G (2d,2p)-C basis set...................................................................70 4-2 Comparison of observed (Ar matrix, 20K) and calculated (MP2//LJ18-Xe/6311++G (2d,2p)-C) isotopomer frequencies (cm–1) for all-12C and singly-13C substituted 12/13CnXe ( n = 2, 3).................................................................................75 5-1 Zero-point-corrected relative energies, Erel (kJ/mol) and zero point energies, ZPE (kJ/mol) of various minima and transition states found on the C3+C6H6 singlet potential surface............................................................................................90 5-2 Comparison of observed (in solid Ar at 12 K) vibrationa l frequencies (in cm 1) of 3 mode of 12/13C3 isotopomers and 12/13C3 isotopomers pertubated by benzene in the 12C6H6•12/13C3 complexes (structures A , I Both Formed in reactions with no entrance barrier predicted) and structure M (a most sisomer found) of the complex from Figure 5-3) with the calculated harmonic frequencies at B3LYP/6-31G(d,p) Level.........................................................................................93

PAGE 10

x 5-3 Proposed assignment of the photoproduct IR absorp tion bands observed after UV-Visible photolysis (h < 5.2 eV) of the C6H6/Ar and C6H6/Cn/Ar matrices....100 5-4 Zero-point-corrected relative energies, Erel (kJ/mol), zero point energies, ZPE (kJ/mol) and vibrational frequencies of CC asymmetric stretch vibration in C3 unit calculated at B3LYP/631G(d,p) level of theory fo r various sstructures of the 1:2 C3/benzene and 2:1 C3/benzene complexes displayed in Figure 5-8.........104 5-5 Experimental (in solid Ar, 12 K) and calculated (B3LYP/6-311++G(d,p), unscaled) vibrational frequencies for the C3•NH3 and C3•ND3 complexes...........107 5-6 Observed (in solid Ar at 12K) a nd calculated IR band energies for the 12/13C3 and 12/13C3•NH3 complexes...........................................................................................108 6-1 Comparison of the calculated and expe rimental (Ar matrix, 12 K) infrared absorption bands (in cm 1) of neutral Dibenzo[b,def]chrysene in the 1Ag electronic ground state...........................................................................................117 6-2 Calculated and observed ve rtical excitation energies, , and oscillator strengths, f , for neutral Dibenzo[b,def]chrysene....................................................................120 6-3 Comparison of the calculated (B3LYP/6 -31G(d,p)) and experimental (Ar matrix, 12 K) infrared absorption bands (in cm 1) of dibenzo[b,def]chrysene radical cation in the 2Bg electronic ground state................................................................124 6-4 Comparison of the calculated (B3LYP/6 -31G(d,p)) and experimental (Ar matrix, 12 K) infrared absorption bands (in cm 1) of dibenzo[b,def]chrysene radical anion in the 2Au electronic ground state.................................................................125 7-1 Comparison of the calculated (B3LYP ) and experimental infrared active absorption bands (in cm 1) including their relative IR intensities of neutral naphthalene in the 1Ag electronic ground state.......................................................138 7-2 Comparison of the calculated (B3L YP) and experimental Raman active absorption bands (in cm 1) including their relative Ra man intensities of neutral naphthalene in the 1Ag electronic ground state.......................................................138 7-3 Comparison of the calculated (B3LYP ) and experimental infrared active absorption bands (in cm 1) including their relative IR intensities of cationic naphthalene in the 2Au electronic ground state.......................................................139 7-4 Comparison of the calculated (B3L YP) and experimental Raman active absorption bands (in cm 1) including their relative Ra man intensities of cationic naphthalene in the 2Au electronic ground state.......................................................139 7-5 Comparison of the calculated (B3LYP/6 -31G(d,p)) and experimental infrared active absorption bands (in cm 1) including their relative IR intensities of neutral and cationic phenanthrene in the electronic ground state.......................................145

PAGE 11

xi 7-6 Comparison of the calculated (B3LYP/6 -31G(d,p)) and experimental infrared active absorption bands (in cm 1) including their relative IR intensities of neutral and cationic anthracene in the electronic ground state...........................................146

PAGE 12

xii LIST OF FIGURES Figure page 1-1 The elemental composition of the solar system.........................................................2 1-2 PAH model for UIR bands.........................................................................................5 1-3 Carbon cluster model for UIR bands..........................................................................6 2-1 Matrix isolation spectroscopy instrumental setup....................................................13 3-1 Schematic representation of the atomic motion in the 2( ), 3( ), and 5( ) fundamental vibrationa l modes of linear C7S..........................................................22 3-2 IR absorption spectrum of the 2( ) mode region of neutral C7S and its 13C isotopic partners produced by laser ablation of a 1.0:0.15:1.0 mixture of 12C : 13C : 32S isolated in an Ar matrix at 35 K.................................................................23 3-3 IR absorption spectra of the 2( ) mode region of neutral C7S and its 13C isotopic partners recorded after tra pping the products in Ar at ca. 12 K, followed by annealing at 35 K......................................................................................................24 3-4 IR absorption spectrum of the 3( ) mode region of neutral C7S and its 13C isotopic partners produced by laser ablation of a 1:0.15:1 mixture of 12C: 13C: 32S in an Ar matrix at 35 K.......................................................................................26 3-5 IR absorption spectra of the 3( ) mode region of neutral C7S and its 13C isotopic partners recorded after tra pping the products in Ar at ca. 12 K, followed by annealing at 35 K......................................................................................................27 3-6 IR absorption spectrum of the 5( ) mode region of neutral C7S and its 13C isotopic partners produced by laser ablation of a 1.0:0.15:1.0 mixture of 12C : 13C : 32S isolated in an Ar matrix at 35 K.................................................................29 3-7 IR absorption spectra of the 5( ) mode region of neutral C7S and its 13C isotopic partners recorded after tra pping the products in Ar at ca. 12 K, followed by annealing at 35 K......................................................................................................30 3-8 The S0–C1 (A), C1–C2 (B), and C2–C3 (B) bond lengths for linear CnS and SCnS clusters as a function of number of carbon atoms, calculated at the B3LYP/6311G(d) level............................................................................................................35

PAGE 13

xiii 3-9 Vibrational frequencies of the mo st intense (C–C) mode for linear Cn, CnS and SCnS clusters computed at B3LYP/6-311G(d) level................................................37 3-10 Number of carbon atoms n is plotted as a function of stretching frequencies of and u mode for CnS (A) and SCnS (B) clusters calculated at B3LYP/6-311G(d) level, respectively.....................................................................................................41 3-11 Dissociation energies of CnS (A) and SCnS (B) clusters computed at B3LYP/6311G(d) level for the loss of S, CS, C2S, C, and C2 units........................................43 3-12 Schematic of the atomic motion in the C2S, C6S, C7S2, C13S2, and C15S2 linear carbon-sulfur clusters...............................................................................................45 3-13 Part of infrared absorption spectra (dis played in A, B and C energy regions) of laser ablated graphite products tra pped in solid Ar at 35 K (spectrum a ) and products of laser ablated mixture of graphi te/sulfur (4/1 molecular ratio) isolated in Ar at 12 K (spectrum b ).......................................................................................47 3-13 Continued................................................................................................................. 48 3-14 Part of infrared absorption spectra (d isplayed in A, and B energy regions) of diacetylene (C4H2) (0.25%)/ Ar electrical discharg e products trapped in solid Ar at 35 K and annealed up to 35 K (bottom spectrum a ) and products of C4H2 (0.25%)/CS2 (0.3%)/Ar electrical discharge is olated in Ar at 12 K (middle spectrum b )...............................................................................................................50 3-15 Coincidences of calculated (B3LYP/6 -311G(d,p) scaled, solid and empty dotes) vibrational frequencies of fundamental modes with experimentally observed band frequencies (listed in the column in cm 1 and marked as vertical dashed lines) for CnS (A) and CnS2 (B) linear carbon-sulfur clusters..................................51 3-16 Infrared spectrum of the products of 12C2H2/12CS2/Ar (spectrum a ) and the 13C2H2/12CS2/Ar (spectrum b ) discharges plotted in two energy regions of the 1 and 2 predicted fundamental modes frequencies...................................................53 3-17 Infrared spectra of the products of laser-ablated 12C2H2/12CS2/Ar (spectrum a ) and the 13C2H2/12CS2/Ar (spectrum b ) samples recorded after matrix annealing up to 35 K and re-cooling to 12 K............................................................................57 3-18 Infrared absorption spectra of the laser ablation products of 12C/32S/Ar (spectrum a ) and the 12/13C/12CS2/Ar (spectrum b ) recorded after matrix annealing (to 35 K) and re-cooling (to 12 K)...........................................................................................63 4-1 Ground state equilibrium geometries for the C2Xe and C3Xe species optimized at the MP2/LJ18 (xenon)/6-311+ +G (2d,2p) (carbon) level....................................69

PAGE 14

xiv 4-2 Plot of the ground state potential surface of linear C3Xe from MP2/LJ18 (xenon)/6-311++G (2d,2p) (carbon) calcu lations with varying C–Xe bond length, R, (plot A) or C–C–Xe bond angle, plot B)...........................................71 4-2 Continued.................................................................................................................7 2 4-3 Plot of the ground state potential surface of bent C3Xe from MP2/LJ18 (xenon)/6-311++G (2d,2p) (c arbon) calculations as a function of C–Xe bond length, R, (plot A) or C–C–Xe bond angle, plot B)...........................................73 4-4 Part of the IR absorption spectra of discharge products of mixtures of 12C2H2/13C2H2/Ar (spectrum A) and 12C2H2/13C2H2/Xe/Ar (0.15%/0.04%/1.5% in Ar) (spectrum B), all tra pped in solid Ar at 12 K................................................76 4-5 Part of the IR absorption spectra of discharge products from mixtures of 12C2H2/13C2H2/Xe/Ar (0.15%/0.04%/1.5% mixture), trapped in solid Ar at 12 K...78 4-6 Part of the infrared absorption spectrum of Cn and CnXe species ( n = 3, 5, 7 and 9) produced by laser ablation and isolated in solid Ar at 12 K (spectra a–c and A–C, respectively)....................................................................................................80 4-7 Experimental C–C asymmetric stretching frequency shifts for Cn clusters compared to CnXe complexes as a function of the number of carbon atoms...........81 5-1 Part of infrared absorption spectrum of C6H6 isolated in solid Ar at 12 K (spectrum a) and the carbon clusters (Cn, n = 3, 6 and 9) deposited with the mixture of Ar/C6H6 (0.5%) at 12 K (spectra b-d).....................................................86 5-2 Infrared absorption spectrum of the 12/13C3 carbon cluster (isotopomeric bands marked by triangles) and 12/13C3 perturbed by benzene in the 12/13C3 12C6H6 complex (bands marked by circles) (s pectrum a) and after 3 min. matrix photolysis using full spectral output of me dium pressure Hg lamp (spectrum b) all isolated in solid Ar at 12 K and di splayed in the C–C asymmetric stretch vibration energy region............................................................................................87 5-3 Structures (A-N) of stable minima of the C6H6 (X1A1g) +C3(g +) reaction products found (B3LYP/6-31G(d,p) theo ry) on the singlet potential surface..........88 5-4 Predicted (B3LYP/6-31G(d,p)) infrar ed absorption spectra of the products (whose structures are given in 5-3) of the reaction of benzene with C3...................92 5-5 Structures of the first or der transition st ates of the C6H6 (X1A1g) +C3 (g +) reaction products found on the singlet potential surface (B3LYP/6-31G(d,p)).......95 5-6 Schematic diagram of the singlet potential surface of reaction of C3 (X1g +) carbon cluster with C6H6 (X1Ag) benzene showing stable minima ( A M ) with the connected transition states ( TSn, n = 1-17)........................................................96

PAGE 15

xv 5-7 Formation mechanism for the PAH (inde ne-like form, Structure N) from the reaction of the ground state benzene and C3............................................................97 5-8 Structures of stable minima of the 1:2 C3/benzene ( O, P ) and 2:1 C3/benzene ( R ) complex found on the singlet pote ntial surface (B3LYP/6-31G(d,p))...................103 5-9 IR spectra of the C3 (X 1g +) + NH3 (X 1A1) reaction products trapped in solid Ar at 12 K and displayed in selected energy regions: without UV-visible irradiation (upper spectrum), after 3 min. irradiation (m iddle spectrum) and after an additional 15 min. irradiation (lower spectrum)................................................106 5-10 13C-labelled IR spectra for the C-C stretching mode of the 12/13C3 carbon cluster (triangles) and the 12/13C3•NH3 complexes (dots) ( cf. Table 5-6 for band assignments)...........................................................................................................107 5-11 Fully optimized ground state equili brium geometry for the singlet C3•NH3 (X 1A’, Cs) complex as calculated at the MP2/ 6-311++G(d,p) level of theory...........109 5-12 Sketch of the singlet pot ential energy surface for the C3 (X 1g +) + NH3 (X 1A1) reaction at the MP2/6-311++G(d,p) level..............................................................111 6-1 Equilibrium geometry of Dibenzo[ b,def]chrysene (DBC) calculated at the B3LYP/6-31G(d,p) level........................................................................................116 6-2 Experimental and calculated IR absorption spectra for neutral DBC....................118 6-3 Electronic absorption spectrum of neut ral DBC (Ar, 12 K) with band positions marked in nm..........................................................................................................121 6-4 Observed IR absorption spectra of neutral and ionic DBC (Ar, 12 K)..................123 6-5 Synthetic experimental and calculated IR absorpti on spectra for DBC cations and anions...............................................................................................................126 6-6 Experimental IR spectra for DBC spec ies (Ar, 12 K) under different photolysis time.........................................................................................................................12 7 6-7 Observed optical absorption spectra of neutral and ionic DBC in Ar at 12 K.......128 6-8 Experimental optical spectra (corresp onding to 6-6 IR spectra)) for DBC species trapped in Ar at 12 K under different photolysis times..........................................129 6-9 Comparison of UIR bands from the re flection nebula NGC 7023 with a mixture of DBC species (26% neutral, 7% anion and 67% cation).....................................132 7-1 Synthetic experimental and calcula ted IR absorption spectra for neutral naphthalene.............................................................................................................140

PAGE 16

xvi 7-2 Difference ( h a) of harmonic and anharmonic frequencies as a function of IR active vibrational modes of neutral naphthalene....................................................140 7-3 Synthetic experimental and calcula ted IR absorption spectra for cationic naphthalene.............................................................................................................141 7-4 Difference ( h a) of harmonic and anharmonic frequencies as a function of IR active vibrational modes of cationic naphthalene..................................................141 8-1 Photodissociation spectra of iron water complex cation........................................153 8-2 Predicted spectra for naphthalene and iron naphthalene complex cation..............154

PAGE 17

xvii Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy LABORATORY STUDIES OF ASTROPHYSICAL MOLECULES By Haiyan Wang December 2005 Chair: Martin Vala Major Department: Chemistry There is growing evidence that the molecu les necessary for the evolution of life on earth arrived from the interstellar medium. The study of these molecules is therefore of great current interest. Two major types of signals from interstellar space, so-called unidentified interstellar infrar ed emission bands and the diffuse interstellar absorption bands, have intrigued and puzzled astroche mists for decades. This work has been concentrated on how to contribute to an unders tanding of the origins of these perplexing signals from space and help identify other molecules that may exist in outer space. Matrix isolation spectroscopy (infrared and ultr aviolet-visible) combin ed with theoretical calculations has been employe d throughout this research. Fourier transform infrared absorption spectroscopic measurements, aided by theoretical calculations and 13C-isotope shifts, have led to the identification of eight heretofore unknown CnSm clusters: C2S, C6S, C7S, C7S2, C9S2, C11S2, C13S2, and C15S2. Infrared absorption studies of xenon polycarbon clusters aid in understanding the special electronic structure and reactivity of carbon clusters, which might be associated with the

PAGE 18

xviii formation mechanism of Buckyball (C60). Reaction of C3 with benzene and ammonia might be involved in the formation of more complex molecular structures, including polycyclic aromatic hydrocarbons (PAHs) and biomolecules such as the amino acids. High resolution vibrational and electronic spectra of neutral dibenzo[b,def]chrysene and its ions in 12 K argon matrices have b een recorded. Spectral assignments were supported by high level theoretic al calculations. A mixtur e of the neutral and ionic infrared spectra of dibenzo[b,def]chrysene re sembles the unidentified IR bands in the reflection nebula NGC 7023. Anharmonic freque ncy calculations for neutral and cationic naphthalene, phenanthrene and anthracene usi ng density functional theory have been carried out for the first time, and the results reveal that anharmonic analysis is significant for the C–H stretching modes of neutral PAHs, but harmonic computation is precise enough for the infrared active vibrationa l analysis of PAH cations. Anharmonic computation might be important for the Raman active vibrational modes.

PAGE 19

1 CHAPTER 1 ASTROPHYSICAL BACKGROUND With the advance of modern science and technology over the last decades, we have learned more about space. Determination of the chemical composition of space, especially the so-called interstellar medium, is a highlight of astrochemistry and is still far from complete. Most species existing in the interstellar medium are usually determined via radioastronomy, infrared and ultraviolet/visib le spectroscopy.1 There is growing evidence that these interstellar molecu les have contributed to the evolution of our Universe, galaxy and life.2 Two types of signals from space, the unidentified interstellar infrared (UIR) emission bands a nd the diffuse interstellar absorption bands (DIBs), have challenged astrochemists for many years.3-6 The main goal of this dissertation is to study possible carriers for these bands and their formation in space. The Interstellar Medium The material in space found between the stars in a galaxy is defined as the interstellar medium (ISM). The ISM is mainly made up of gas and some dust grains that are immersed in radiation, magnetic fields and cosmic ray particles. The ISM gas is mostly hydrogen (90%) and helium (10%) with small amounts of heavier elements, such as carbon, sulfur, and iron.7 The elemental composition of th e solar system is shown in Figure 1-1, in which the abundance of hydr ogen atoms is arbitrarily set to 1012.8 The density of interstellar gas is extremely low. For example, the density of the interstellar medium near the Sun is around 1.7 hydrogen atoms per cubic centimeter.7 Even with such a low density, the ISM still acc ounts for 20–30% of the mass of our galaxy.2 The

PAGE 20

2 density and composition vary from place to pla ce. Table 1-1 is a short list of the average number of atoms for different elements pe r trillion atoms in several places of the universe.9 It can be seen that our solar system has a smaller number of helium atoms but a higher concentration of hydr ogen and heavier elements compared to the average composition of the interstellar gas. Also, the interstellar medium has a completely different elemental distribution pattern from the earth’s crust and our human body. Figure 1-1. The elemental composition of the solar system. The abundance of hydrogen is arbitrarily set to 1012 so that the smallest abunda nce in the graph is about 1.8 Table 1-1. List of average chemical abundances per trillion atoms in several places of the Universe. Element a Interstellar Gas Sun Earth’s Crust Human Body Hydrogen (H) 9.00×1011 9.27×1011 2.88×1010 6.06×1011 Helium (He) 1.00×1011 7.20×1010 Carbon (C) 1.48×108 3.45×108 5.60×108 1.07×1011 Nitrogen (N) 6.70×107 1.07×108 7.00×107 2.44×1010 Oxygen (O) 4.63×108 6.27×108 6.04×1011 2.57×1011 Sodium (Na) 1.90×105 1.61×106 2.55×1010 7.50×108 Magnesium (Mg) 9.30×105 3.20×107 1.76×1010 1.10×108 Aluminum (Al) 1.20×103 2.30×106 6.25×1010 Silicon (Si) 7.60×105 3.20×107 2.05×1011 Phosphorus (P) 1.90×104 2.50×105 7.90×108 1.30×109 Sulfur (S) 7.60×106 1.50×107 3.30×108 1.30×109 Calcium (Ca) 4.20×102 1.80×106 1.88×1010 2.30×109 Iron (Fe) 2.50×105 2.30×107 1.86×1010 a Reference 9

PAGE 21

3 Table 1-2. List of identified interstellar molecules. Number of atomsa 2 3 4 5 6 7 8 9 10 AlCl AlNC C2CN C4H C2H4 C6H C2H6 (CH3)2O (CH3)2CO AlF C2H C2H2 C4S C5H c -C2H4O C6H2 C2H5CN C6H6 C C2O C3O C4Si C5N CH2CHCN C7H C2H5OH HC10CN CH C2S C3S C5 C5O CH2CHOH CH3COOH C8H HC9CN CH+ C2Si C3Si c -C3H2 C5S CH3NH2 H3C-CC-CN CH3C4H HOCH2CH2OH CN C3 c -C3H CH2CN CH3CN HC4-CN HCOOCH3 HC6CN CO CH2 CH2D+ CH2CO CH3NC HCC-CH3 HOCH2COH CO+ CO2 CH3 CH2NH CH3OH HCOCH3 CP H2O H2CN CH4 CH3SH CS H2S H2CO H2COH+ HC4H CSi H3 + H2CO+ HC2-CN HC3OH FeO H-CN H2CS HC2-NC HCONH2 H2 HCO H3O+ HCOOH HCl HCO+ HC-CN HNCCC HF HCS+ HCNH+ l -C3H2 KCl HNC HNCO NH2CN LiH HNO HNCS SiH4 NaCl HOC+ HOCO+ NH MgCN l -C3H NO MgNC ND3 NS N2H+, NH3 OH N2O PN NaCN SH NH2 SiN OCS SiO SiCN SiS SO2 SO SO+ a Reference 1 The ISM is composed of three types of clouds, classified according to their temperature: cold (10’s K), warm (100 to 1000’s K) and hot (millions K).10 It is clear that the composition of the ISM is dependent on the temperature. Cold temperatures are better suited for the formation and st abilization of molecu les and radicals. So some cold clouds (for example, the Sagittarius giant molecu lar cloud, Sgr B2(M)) cons ist of a relatively high concentration of molecules and are distin ct from molecular clouds that have much higher densities (~ 102–106 atoms/cm3) with temperatures around 10–30 K.10 Although molecular clouds occupy less than one percent of space, they contribute almost half the mass of the ISM. To date, more than 120 different chemical compounds have been

PAGE 22

4 identified in the ISM. Table 1-2 lists the interstellar molecules identified thus far.1 Most of these have been detected by radio or microwave methods, although some of them like C3, C5 and C6H6 have been identified by infrared (IR ), visible (Vis) and ultraviolet (UV) spectroscopy.1 They range in complexity from the simplest diatomic, H2, through familiar ones like water and acetylene to carbynes (long carbon chains) like HC11N. Interestingly, more than 75% of interstellar molecules are carbon-bearing species. Moreover, it has been accepted that polycyclic aromatic hydrocarbons may exist in space, but there are still no specific assignments.11,12 Unidentified Interstellar Infrared Emission Bands The unidentified interstellar infrared emission (UIR) bands, discovered in 1973, are a series of emission bands with in the wavelength region 3–15 µm.5 Dominant features are seen at 3.3, 6.2, 7.7, 8.6, and 11.2 µm w ith several underlying continua.5,6 UIR spectra from different interstellar environments are di stinguishable by the in tensity distribution of individual bands, but the general features, like band positions, are basically similar to one another. Possible carriers for these perplexi ng signals have intrigued astrochemists since 1973. Heretofore, no specific carrier of UIR bands has been assigned but two models have been proposed to interpret these emissions. Vibrational emission of polycyclic ar omatic hydrocarbons (PAHs) was first suggested by Leger and Puge t to explain the UIR bands.11 Since then, the study of the structure and spectra of PAHs has beco me a vigorous field in astrochemistry.11-14 This hypothesis has been widely accepted and P AHs are now commonly thought to be the carriers of the UIR bands, although the mechan ism of formation of the PAHs remains unsolved. PAHs are regarded as the most abundant free organic molecules in space. Neutral, dehydrogenated a nd ionized species, including their coordination compounds,

PAGE 23

5 may be present in space.13 Recently, it has been recomme nded that ionic PAHs or a mixture of neutral and ionic PAHs contribu te to the dominant features of the UIR bands.14 Figure 1-2 is an example of usi ng PAHs to explain the UIR signals.14 From the comparison, it can be seen that the UIR bands match well with the la boratory observation when a rational PAH mixture is applied. Figure 1-2. PAH model for UIR bands. Top: (a) emission spectrum from the proto– planetary nebula IRAS 22272+5435 comp ared with the (b) absorption spectrum produced from a mixture of neutral and cationic PAHs. Bottom: the (c) emission spectrum from the Orion i onization ridge compared with the (d) absorption spectrum produced from a mixture of fully ionized PAHs.14 Another possible UIR carrier is long chain carbon clusters, which is shown in Figure 1-3.15 It is clear that spec ific lengths of carbon chai ns might contribute to individual UIR bands. Besides, heteroat om-doped carbon clusters possess similar electronic transition wavelengths with a certain red or blue shift. This type of carbon cluster has been observed in space; for instance, HC11N has been identified in the cold

PAGE 24

6 dust cloud TMC-1.16 Therefore, carbon clusters and their derivatives should be considered as one of the possible candidate s for the UIR bands, despite many arguments in this field. Figure 1-3. Carbon cluster model for UIR bands . The electronic transition wavelength for even-numbered carbon cluster anions observed in an Ar matrix at 36 K versus the number of carbon atoms in the cluster chain.15 Diffuse Interstellar Absorption Bands The diffuse interstellar absorption bands (DIBs) were first recorded by Merrill many decades ago.3,4 They are a series of absorpti on bands occurring in the wavelength range from 400 to 1300 nm. The DIBs are notab le due to their weak, broad and shallow spectral features. The full bandwidth at half -maximum (FWHM) of DIBs can vary from 0.06 to 4 nm.17 To date, over 200 DIBs have been identi fied. It is obvious that more than one carrier is probably present. Recent rese arch supports the idea that PAHs and carbon clusters, especially their ionic forms, might also contribute to the DIBs.18-22 Table 1-3 shows the comparison between the absorption of some PAH cations and several DIBs.23 It is clear that the electr onic transitions of specific P AH cations might correspond to individual interstellar diffuse bands.

PAGE 25

7 Table 1-3. Comparison of the electronic transi tions of PAH cations with DIBs. Electronic transition of PAH cations were measured in neon matrices. PAH+ a peak (nm) DIBs (nm) Pyrene (C16H10 +) 439.5 442.9 1-Methylpyrene (CH3-C16H9 +) 444.2 442.9 4-Methylpyrene (CH3-C16H9 +) (457.7) 482.8 757.5 482.4 758.1 Naphthalene (C10H8 +) 652.0 674.2 652.0 674.1 Phenanthrene (C14H10 +) 856.8 898.3 857.2 Tetracene (C18H12 +) 864.7 864.8 Benzo[ghf]perylene (C22H12 +) 502.2 755.2 758.4 794.3 503.9 755.8; 756.2 758.1; 758.6 793.5 Coronene (C24H12 +) 459.0 946.5 459.5 946.6 a Reference 23 In this dissertation, matrix-isolation spectroscopy and theoretical modeling are employed to study chemical and physical propert ies of existing or possible astrophysical species. The aims are to discover possible car riers of the UIR and DIB bands, to interpret the formation mechanism of those carriers, a nd to provide reference spectra for future astrophysical observations. This dissertation consists of eigh t chapters. Experimental and theoretical methods are described in Chapte r 2. The identification of carbon sulfur clusters is detailed in Chapter 3. Chapters 4 and 5 are concerned with the reactions between small carbon clusters and xenon, a mmonia, and benzene. Vibrational and electronic spectra of dibe nzo[b,def]chrysene are reported in Chapter 6. Chapter 7 attempts to determine theoretically the a nharmonicity of PAHs. A brief conclusion and future prospects are included in Chapter 8.

PAGE 26

8 CHAPTER 2 THEORETICAL AND EXPE RIMENTAL METHODS Since potential applications of outer space, like extraterrestrial mining, are approaching the realm of po ssibility, astrochemistry is becoming more and more attractive but still undergoing challenge because of technical obstacles. Thus far, astrochemistry cannot be done. It can only be observed and simulated. Matrix isolation spectroscopy (infrared a nd ultraviolet-visible) and theore tical calculations are two of the most popular and feasible tools to predict and interpret astrophysical phenomena. This chapter will discuss the deta ils of these methods as employed in the research. Computation Details Most calculations were carri ed out using the Gaussian 98 or Gaussian 03 suite of programs.24,25 Specific computational problems us ing ACES II and NWCHEM software suites were performed as well.26,27 The geometry optimization, vibrational frequencies and energetics were calculated for all species of interest. Vertical excitation energies for PAHs and transition states for the reaction of tri-carbon with ammonia and benzene are also reported. Geometry Optimization a nd Vibrational Frequencies The geometry for each molecule was optim ized to obtain the minimum potential energy. The initial structure fo r individual molecules was constructed based on previous results or widely-accepted assumptions. For example, carbon sulfur clusters possess a linear structure and benzene has D6h symmetry. The geometry optimization was carried out both with symmetry restraint and wit hout, followed by a vibrational frequency

PAGE 27

9 calculation. Non-negative freque ncies obtained during the vibr ation analysis could certify a local minimum for the corresponding struct ure. Anharmonic frequencies for PAHs can be only predicted using a Gaussian 03 protocol.25 Most of calculations were computed usi ng density functional theory (DFT), like B3LYP theory (BeckeÂ’s three-parameter hybrid functional combin ed with Lee-Yang-Parr correlation functional),28,29 with various basis sets, such as DunningÂ’s correlation consistent double (cc-pVDZ),30 6-31G(d,p), 6-311G(d), and 6-311++G(d,p) basis set. In order to judge computational accuracy, ab in itio calculation results for some molecules were also obtained and compared with DFT. For instance, the energies and equilibrium geometries of C2S, C6S and SC7S were also calculated at the MP2/cc-pVDZ and CCSD (T)/cc-pVDZ levels. The choice of theory level and basis set to be used considered both reliability and computational cost. For exam ple, either the 6-31G(d,p) or 6-311+G(d,p) basis sets were applied to neutral dibe nzo[b,def]chrysene (DBC) using B3LYP theory, while only the 6-31G(d,p) basis set was em ployed for the DBC ions. It should be mentioned that due to the elec tronic structure of carbon xeno n complexes, geometries and vibrational frequencies for C2Xe or C3Xe were optimized using ab initio MP2 theory with a LJ18 basis set for Xe31 and a 6-311++G (2d, 2p) basis set for carbon. Finally, for some organic exotic compounds, Wiberg Bond I ndices (WBI) were pr edicted by adding the BNDIDX subroutine to the natura l bond orbital (NBO) analysis.32,33 Transition State Structure The transition states connecting the stable minima in the reaction and photolysis of C3 with ammonia and benzene were also investig ated using the Gaussian 98 or Gaussian 03 protocol. The synchronous transit-guided quasi Newton (STQN) me thod implemented in the Gaussian suites is a precise appro ach for locating transition structures.34,35 This

PAGE 28

10 method is able to converge efficiently when provided with an empirical estimate of the Hessian and suitable starting structures. Three st ep calculations were carried out to search for appropriate transition structures. The initial transition structure was generated and optimized by running the QST2 (and/or QS T3) computation followed by the harmonic frequency calculation. An imaginary freque ncy was expected for a possible transition structure that was then confirmed by the intrinsic reaction coordinate (IRC) calculation.36,37 Vertical Excitation Energies for PAHs Vertical excitation energies and oscillator strengths of neutral DBC and its ions were computed by the TDDFT method using the B3LYP functional with the 6-31G(d,p) basis set. In order to monitor the accuracy of the vertical excita tion energies using the Gaussian 98 platform, all excited-state calcul ations were also carri ed out with massively parallel DFT and TDDFT implementations of the NWCHEM quantum chemistry software suite,27 now widely used in modeling the excited states of PAHs.38 Vertical excitation energies and oscillator strengths of the neutral, ca tionic, and anionic species of DBC were computed by TDDFT using th e Becke-Lee-Yang-Parr (BLYP) and asymptotically-corrected Becke3-Lee-Y ang-Parr [B3LYP(AC)] functionals in conjunction with the 6-31G(d,p) basis set. Geometries were optimized in the ground states using the B3LYP func tional and the 6-31G(d,p) basi s set (for the neutral and cation) and the 6-31++G(d,p) basis set (for the anion). Generally, TDDFT is capable of reproducing the excitatio n energies of low-lying valence excited states of closedand open-shell PAHs within 0.3 eV and the corresponding oscillat or strengths only qualitatively.38 The asymptotic correction algorithm39 employed in this work automatically adjusts the depth of the exch ange-correlation potentials in the valence

PAGE 29

11 region before splicing a –1/ r asymptotic tail to it, rely ing on a phenomenological linear relationship between the ionization potential and highest occupied orbital energies of B3LYP DFT calculations.40 Excited-state roots were sought in each calculation using the Davidson trial vector algorithm. Twenty r oots were obtained for neutral DBC and only ten roots for its cation and anion. Experimental Details Matrix isolation spectroscopy (infrared a nd ultraviolet-visible) has been widely used in the study of astrophysical species. C oupled with theoretical calculation, it is an efficient tool for the study of radicals whic h may not be stable under “normal” conditions but may exist in extraterrestrial space. The experimental setup is shown in Figure 2-1. The details of the experimental apparatus are described in previous studies from the Vala group.41 Briefly, the species of concern in this research were produ ced using different techniques, such as laser ablation, electri cal discharge and electron ionization. The reaction products were trapped with di fferent isolant gases on a CsI or BaF2 (used for PAH samples only) window cooled to 12 K by a closed-cycle helium cryostat (ADP Displex). The selection of isolant gases that could be Ar or Kr or Ar mixed with other gases (Xe: 0.5–4%; NH3 or ND3: 0.1–1%; C6H6 or C6D6: 0.5–1%; CCl4: 0.3%) was dependent on the experimental interest. After ~ 2–5 hour deposition, infrared absorption spectra were recorded with a NICOLET Magna 560 (used for carbon sulfur clusters and reaction of carbon clusters with xenon, ammonia, and benzene) or MIDAC M2000 (used for PAH samples) Fourier transform infr ared (FT-IR) spectrometer in the 700–4000 cm 1 region at 0.25 cm 1 or 1.0 cm 1 resolution, respectively. Th ereafter, the electronic absorption spectrum (900–200 nm) of the same matrix (only for PAH samples) was collected with an IBM 9420 UV-visible sp ectrophotometer at 0.28 nm resolution. A

PAGE 30

12 correlation between the resulting IR and UV-visi ble spectral band intensities was used to confirm the identity of cations and ani ons. Secondary reactions were noted after annealing (to 35 K) or photolysis of the matr ix. Photolysis was accomplished with the use of an ultraviolet/visible medium pressu re 100W Hg lamp. Fo r the reaction of C3 with ammonia and benzene, different filters (Veb-Jena) were used to determine the photodissociation threshold of th e complex. In case of the C3 / benzene complex it was necessary to block the 532 nm radiation used in ablation si nce it was found to effectively destroy the complex. Laser Ablation Laser ablation was usually employed to generate carbon and carbon sulfur clusters. A pulsed Nd:YAG laser (1064 and 532 nm) was used to vaporize pressed pellet samples which may contain 12C, 13C, and S or their mixture. St andard carbon and sulfur samples were used as received: 12C (powdered graphite, natural abundance 12C (98.9%) and 13C (1.1%)); 13C (99%, Cambridge Isotope Laboratorie s, Inc.)), and S (natural abundance 32S (96%) and 34S (4%)). Electrical Discharge The electrical discharge device was the same as described previously by the Bondybey group.42 The gas mixture (~1 atm, 800 µ s) exited the pulsed valve into a 1400 V discharge channel consisting of two Al di sks separated by a 2 mm teflon disk, through which a 2 mm diameter hole was drilled. The pl asma products were trapped on a cryostat window (12 K). Different gas mixtures were us ed to generate differe nt species. Carbon sulfur species were generated in a pulsed-jet discharge (1.2 1.7 kV) of 12C2H2, (or 13C2H2) (0.1 0.3%)/CS2 (0.1 0.3%)/Ar and 12C4H2 (0.25%)/12CS2 (0.25%)/Ar mixtures. Carbon disulfide (Fisher Co.), acetylen e (purified, Matheson, Inc.), and 13C2H2 (99%,

PAGE 31

13 Cambridge Isotope Laboratories, Inc) were used as received. The CnXe species were produced in a pulsed-jet discharge of argon seeded with a mixture of 12C2H2 (0.02–0.5%), 13C2H2 (0.02–0.5%), and Xe (0.5–3%). Figure 2-1. Matrix isolation spec troscopy instrumental setup. Electron Ionization of Dibenzo[b,def]chrysene Solid neutral DBC powder was vaporized from a quartz oven heated to around 300°C. Using a homemade electron gun located ca. 3 cm from the BaF2 sample window,

PAGE 32

14 DBC ions were generated and deposited with the neutral species in an argon matrix. The electrons were accelerated toward the de position region by a +90 V O-ring electrode located 5 mm in front of the cryosta t window. The electron beam current ( ca. 50–70 µ A) was monitored using a microammeter connected in series with the ring electrode. Cation formation proceeds via Penning ionization with Ar* metastable species and Ar+ DBC charge transfer complexes, while the hi gh electron affinity of DBC makes anion formation possible. To obtain th e highest possible ratio of cati ons to neutral species, an electron scavenger, carbon tetrachloride CCl4 (Kodak, Spectrograde), was seeded into the argon gas (99.7% Ar, 0.3% CCl4). A lengthy deposition (typi cally 4–5 hrs) was required to achieve comparable concentrations of i ons and neutrals. Ther eafter, the electronic absorption spectrum (900–200 nm) of the same matrix was collected with an IBM 9420 UV-visible spectrophotometer at 0.28 nm resolution.

PAGE 33

15 CHAPTER 3 INFRARED ABSORPTION SPECTROSCOPY OF CARBON SULFUR CLUSTERS Introduction During the past decade, the polycarbon sulf ides have been studied extensively, both theoretically and experimentally, primarily b ecause of their importance as components of the interstellar medium (ISM) and building blocks in material sciences. Some carbonsulfur clusters exhi bit unusual electrical properties in amorphous organic semiconductor films.43 Small CnS ( n = 1–3, 5) species have been obser ved in the envelopes of several stars.44-47 Absence of an electric dipole prec ludes the interstellar detection of the symmetric SCnS species via microwave detection methods, but they may be observable by their IR absorption. A number of sulfur-containing carbon cl usters have been produced in the laboratory. C2S and C3S were initially detected in the laboratory by millimeter and submillimeter spectroscopy45,46 and then observed by radio frequency spectroscopy in the dark molecular cloud, TMC-1,44 and in the circumstellar en velope of the carbon star, IRC +10216.47-50 Using radioastronomy, both of these species have been observed in the vicinity of various ISM objects.47 Using infrared measurements, Maier and coworkers observed CnS ( n = 1, 3) and CnS2 ( n = 2–5) clusters in Ar matrices.51-54 Later, C2S2 and C3S2 clusters were observed by Andrews and coworkers using a discharge through a CS2 / Ar mixture.55 Recently, Vala’s group generated CnS and CnS2 ( n = 3–5) clusters by laser ablation of C / S pellets and identified th em by Fourier transform infrared (FT-IR) spectroscopy, supported by density f unctional theory calculations.56 Rotational constants

PAGE 34

16 for the linear clusters CnS ( n = 1–9) have been determined by Fourier transform microwave spectroscopy.45,46,57-59 In addition, electronic abso rption spectra of small C2S, C2S, C4S C5S, C6S and C6S clusters have been reported.60,61 Finally, CnS+ and CnS2 + clusters up to C27S2 + were determined using mass spectrometry.62,63 Early theoretical work was focused on small CnS species with n = 2–9.64-71 Later computational results on the geometries of CnS clusters, with n up to 20, were presented by Pascoli’s group.72 Except for C18S, which was predicted to have a monocyclic S capped structure, the CnS species ( n = 1–20) were found to be linear with the S atom terminating the carbon chain.72 Recently, other calculations on CnS and CnS2 clusters supported their linear structures.60,73,74 Although Cn, CnS, and SCnS ( n = 1–9) clusters are ge nerally linear in their electronic ground states , non-linear species do exist. Fr om theoretical predictions, Cn clusters with n 10 appear to prefer bent , cyclic or spherical stru ctures, but mostly linear isomers have been detected in the ga s phase and solid rare gas matrices.75-78 Electronic absorption spectra of linear Cn clusters ( n = 3–15, 17, 19 and 21) have been recorded recently.78-82 In addition, resonance Raman spectro metry has proven the existence of odd linear Cn isomers up to C29 in Ar matrices.15 Given the numerous large, linear, pure carbon clusters known, the generation of large linear CnS and SCnS clusters should be possible, either by laser ablation of carbon / su lfur mixtures, or by the reaction of linear Cn ( n 29 ) carbon clusters with S and / or CS species , with trapping in solid Ar matrices. In this chapter, observation of C7S species in argon matr ices is reported. Unknown carriers of some bands shown in the spectra wh ich were recorded from the experiments to identify the C7S cluster led to the theoretical computation for CnSm clusters ( n = 1–29, m

PAGE 35

17 = 1–2) to explore their energi es, structural parameters, a nd vibrational frequencies. Sequentially, the infrared bands were assigned to specific CnSm clusters with the aid of the theoretical results. The Linear C7S Carbon-Sulfur Cluster The linear C7S carbon-sulfur cluster was produced by the vaporization of a carbon and sulfur mixture, and studied via FT-IR spectroscopic met hods in argon matrices. Isotope shifts were also inves tigated. DFT and ab initio calcul ations were used to identify three new vibrational modes. Theoretical Predictions In its 1+ ground state linear C7S has 34 valence electrons, 16 of which are distributed in 8 doubly degenerate orbitals. Table 3-1 lists the C7S state energies, dipole moment, bond lengths, rotational constant and vibrational fr equencies, obtained at various levels of theory. Although primarily cumulenic in bonding, C7S still shows a small alternation in its C C bond lengths, a fact previously noted for smaller CnS ( n = 1– 5) clusters.56 For example, the C4 C5 bond length in C7S is shorter than the C3 C4 bond length (by ca. 0.016 Å). Different methods and basis sets predict similar carbon-carbon bond lengths and show the same alternation. The average C C bond length is about 1.28 Å, which is close to the experimental value a nd also close to other theoretical results. The calculated C S bond length, however, appears to vary with theoretica l approach. In C5S, this bond length is about 1.547(6) Å.59 Because of the similarities with other CnS clusters, it is expected that in C7S it should be close to 1.55 Å.65 Determining the C S bond length through the effect of isotopic substitution on the rotational co nstant is model dependent ( cf. , Table 3-1).

PAGE 36

18 Table 3-1. Energies E (Hartrees), zero point energies ZPE (kcal/mol), dipole moments µe (Debyes), bond lengths (Å ), rotational constants Be (GHz), harmonic vibrational frequencies (unscaled, cm 1) and integral IR intensities (km/mol, in parentheses) calculated for linear C7S at various levels of theory. C7S B3LYP /6-311G* B3LYP /cc-pVDZ B3LYP /aug-cc-pVTZ CASSCF(4,6) /cc-pVDZ E 664.7483 664.6965 664.7887 662.3804 ZPE 20.3672 21.7178 20.9487 22.3260 µe 5.7301 5.7005 6.1616 5.8285 S C1 1.5579 1.5668 1.5542 1.5454 C1 C2 1.2779 1.2875 1.2764 1.2780 C2 C3 1.2699 1.2799 1.2675 1.2635 C3 C4 1.2811 1.2905 1.2795 1.2894 C4 C5 1.2656 1.2755 1.2625 1.2680 C5 C6 1.2929 1.3022 1.2915 1.2989 C6 C7 1.2817 1.2911 1.2763 1.2658 Be a 0.4141 0.4081 0.4156 0.4149 Be b 0.4042 0.3984 0.4057 0.4050 1( ) 2229.8(231) 2238.6(799) 2218.9(339) 2339.0 2( ) 2209.2(5750) 2223.3(5089) 2191.2(5956) 2295.5 3( ) 2009.8(2859) 2018.5(2607) 2003.8(3107) 2054.5 4( ) 1712.4(139) 1719.2(93) 1710.5(141) 1811.7 5( ) 1284.8(284) 1289.7(236) 1284.5(259) 1368.6 6( ) 854.9(86) 858.3(75) 855.7(90) 901.5 7( ) 431.9(3) 433.7(3) 432.5(3) 458.0 8( )c 615.2(6) 700.3(6) 573.1(5) 669.7 9( )c 449.7(1) 577.2(0) 522.7(0) 591.6 10( )c 308.4(4) 453.5(1) 445.7(0) 500.3 11( )c 233.2(0) 268.7(3) 246.4(2) 240.0 12( )c 118.7(8) 150.5(4) 138.9(4) 140.3 13( )c 31.5(0) 55.0(2) 51.4(2) 52.2 a Rotational constant for C7 32S; experimental constant is 0.4144 GHz (Reference 59). b Rotational constant for C7 34S; experimental constant is 0.4045 GHz (Reference 59). c Doubly degenerate bending ( ) modes. Although both B3LYP and CASSCF (4, 6) methods predict the structure of C7S reliably, it was also studied by the CCSD(T ) and MP2 methods. The theoretical state energies, zero point energies (ZPEs), bond lengt hs and rotational constants are listed in Table 3-2. The rotation constants obtained by CCSD(T) calculations are worse than those using DFT, despite the fact that it pr ovides for precise electronic correlation.

PAGE 37

19 Table 3-2. Energies E (Hartrees), dipole moments µe (Debyes), bond lengths (Å), and rotational constants Be (GHz), calculated for linear C7S at various levels of theory, all with a cc-pVDZ basis set. C7S(1+) MP2 CCSD(Full) QCISD(T) CCSD(T) CCSD(T)(Full) E 663.3229 663.3265 663.3863 663.3829 663.4050 µe 5.0215 5.4842 4.4692 4.6856 4.7090 S C1 1.5701 1.5622 1.5806 1.5776 1.5758 C1 C2 1.3041 1.2982 1.3009 1.3018 1.3006 C2 C3 1.2904 1.2818 1.2958 1.2938 1.2927 C3 C4 1.3045 1.3013 1.3031 1.3042 1.3029 C4 C5 1.2882 1.2789 1.2926 1.2908 1.2897 C5 C6 1.3162 1.3130 1.3156 1.3165 1.3152 C6 C7 1.3104 1.2964 1.3138 1.3118 1.3107 Be a 0.4005 0.4050 0.3990 0.3995 0.4005 a Experimental rotational constant for C7 32S is 0.4144 GHZ (reference 59). Note that the predicted Be values are in poorer agreement with experiment than the Be values calculated at the B3LYP and CASSCF(4,6) levels listed in Table 3-1. As a test of the stability of the gr ound state wavefunction an EOM-CCSD (ACES II)26 calculation was performed on C7S with the same basis set and geometry as in the B3LYP calculations. Since only positive excitation energies were found, it was concluded that the predicte d structural and spectrosc opic constants found at the B3LYP/cc-pVDZ level should be reliable. Theoretical vibrational fre quencies can often be used to identify new molecular species. Table 3-3 presents the results of the B3LYP/cc-pVDZ level calculations on two carbon-sulfur cluster series, so me members of which have not yet been observed. For clusters that have been observed, compar ison of the calculated frequencies (with appropriate scaling factors, cf. Table 3-3) with the predic ted vibrational frequencies shows reasonably good agreement. The maximum discrepancy in the CnS series was found for the 5 mode of C7S ( ca. 30 cm 1), while the worst discrepancy in the SCnS series was found for the 5 mode of SC5S ( ca. 27 cm 1). These discrepancies arise either from mixing of similar modes not explicitly co nsidered or from the uniform scaling used

PAGE 38

20 for all C C and C S vibrational frequencies. For instance, the 5 mode of C7S is primarily a C C stretching mode, but the C S vibration also contri butes to this mode. Since the common scaling fact or used for these modes doe s not account for this mixing, the scaled 5 frequency does not match the experimental value very well. Table 3-3. Comparison of experimental (Ar matrix) and most intense calculated (B3LYP/cc-pVDZ) IR mode frequencies (cm 1) for linear CnS ( n = 1–9) and SCnS ( n = 1–9) carbon-sulfur clusters in their electronic ground states. Relative intensities are given in parentheses. CnS Mode exp (cm 1) cal a (cm 1) SCnS Mode exp (cm 1) cal a (cm 1) CS(1+) 1( ); C–S 1275.1(1.0)c,d 1275.1(1.0) SCS(1g +) 2( u); C–S 1528.2(1.0)c,d 1533.9(1.0) C2S(3) 1( ); C–C 1636.2(1.0) SC2S(3g ) 3( u); C–S 1179.7(1.0)d 1159.3(1.0) C3S(1+) 1( ); C–C 2( ); C–S 3( ); C–S 2047.6(1.0)c,d 1533.2(0.1)d 725.6(0.009)d 2047.6(1.0) 1540.8(0.04) 726.6(0.009) SC3S(1g +) 3( u); C–C 4( u); C–S 2078.5(1.0)c,d 1024.6(0.18)c,d 2081.5(1.0) 1022.7(0.11) C4S(3) 2( ); C–C 1746.8(1.0)c 1727.3(1.0) SC4S(3g ) 4( u); C–C 5( u); C–S 1872.1(1.0)d 897.7(0.117)d 843.7(0.056)d 1855.0(1.0) 870.3(0.18) C5S(1+) 1( ); C–C 3( ); C–C 2124.5(1.0)c 2140.0(1.0) 1580.0(0.15) SC5S(1g +) 4( u); C–C 5( u); C–C 6( u); C–S 2104.7(1.0)d 1687.9(0.36)d 783.5(0.04)d 2118.5(1.0) 1661.2(0.24) 796.4(0.04) C6S(3) 2( ); C–C 4( ); C–C 1995.4(1.0) 1344.8(0.18) SC6S(3g ) 5( u); C–C 6( u); C–C 2031.5(1.0) 1458.4(0.29) C7S(1+) 2( ); C–C 3( ); C–C 5( ); C–C 2088.1(0.67)b, e 1913.6(1.0)b 1256.1(0.07)b 2114.5(1.0) 1919.7(0.51) 1226.6(0.05) SC7S(1g +) 5( u); C–C 6( u); C–C 7( u); C–C 2088.5(1.0) 1952.7(0.88) 1341.4(0.14) C8S(3) 1( ); C–C 4( ); C–C 6( ); C–C 2072.2(1.0) 1760.6(0.43) 1092.9(0.06) SC8S(3g ) 6( u); C–C 7( u); C–C 8( u); C–C 2073.8(1.0) 1822.6(0.48) 1211.5(0.16) C9S(1+) 2( ); C–C 3( ); C–C 5( ); C–C 2099.1(0.09) 2019.5(1.0) 1658.3(0.095) SC9S(1g +) 7( u); C–C 8( u); C–C 9( u); C–C 1994.3(1.0) 1697.5(0.14) 1127.8(0.04) a Scaled uniformly by 0.9510 factor for m odes with primarily C–C character and by 0.9824 factor for modes with primarily C–S character. b This work. c Reference 56. d References 51, 53-55. e Tentative assignment. Since no imaginary frequencies were found for C7S ( cf. Table 3-1), its linear geometry is associated with a local minimum on the potential energy surface. Calculations at different le vels all predict similar vi brational frequencies for C7S. Its seven stretching ( ) modes and six doubly degenerate bending ( ) modes are all infrared active. Generally modes exhibit high intensities while modes possess relatively low intensities. Since the values of 1 and 2 are very close, a mixt ure of these two modes via

PAGE 39

21 Fermi resonance is possible. Isotopic shift cal culations reveal that the intensity of the 2 mode is redistributed into the 1 mode, which makes it difficult to predict the integral intensity of the 2 mode solely from theoretical calculations. Mode mixing could explain the fact that although the most intense mode calculated for C7S is 2, the 3 mode is the strongest observed. Th e ratio of experimental integral intensities for the 3 and 5 modes is ca. 14.3, of the same order of magnitude as the computed values of 10.0 (B3LYP/6311G(d)), 11.0 (B3LYP/cc-pVDZ) and 12.0 (B3LYP/aug-cc-pVTZ), cf. Table 3-1. But, the analogous ratio for the 2 and 3 modes is 0.67, much smaller than the calculated ratios of 2.01, 1.95, and 1.92, respectively ( cf. , Table 3-1). Experimental Results Three new bands were observed at 2088.1, 1913.6 and 1256.1 cm 1, and assigned here to the 2, 3 and 5 stretching modes of linear C7S, respectively, cf. , Figure 3-1. Figure 3-2 shows the infrared spectrum of the 2 mode for neutral C7S and its 13C isotopic partners (from a S / 13C mixture). Two other spectra for 12C / 32S and 12C / 13C mixtures, under the same experimental conditions, are also included and confirm these assignments. Trapping the products (in solid Ar, ca. 12 K) and annealing (to 35 K), leads to the IR spectra (in Figure 3-3) of the C7S 2 mode and its 13C isotopomeric bands. The corresponding spectra for the 3 and 5 modes are presented in Figure 3-4, Figure 3-5 and Figure 3-6, Figure 3-7, respect ively, where band assignments to pure carbon or other small carbon sulfur cl usters are indicated.56, 83-85 For example, the 1244.0 cm 1 band ( cf. Figure 3-6c) is due to the 10 ( u) mode of 12C12 (X3g), based on recent work.85 Each new assignment is discussed in turn below.

PAGE 40

22 Figure 3-1. Schematic representation of the atomic motion in the 2( ), 3( ), and 5( ) fundamental vibrationa l modes of linear C7S. Band assignment for the 2( ) vibration Frequency calculations for the 2 mode of C7S give 2114.5 cm 1 ( cf. , Table 3-3). With an assumed calculation error of ±50 cm 1, an observed band at 2088.1 cm 1 has been assigned tentatively to this mode. Seve ral reasons lie behind th is attribution. First, experiments under different conditions (such as matrix temperature and/or laser photon flux) showed that this band tracks the intensities of the two bands (1913.6 and 1256.1 cm 1) also assigned to linear C7S ( vide infra ). Second, of the small and mid-sized CnS and SCnS clusters given in Table 3-3, the 2088.1, 1913.6 and 1256.1 cm 1 bands fit only the frequency pattern predicted for C7S. However, as stated earlier, the mixing of 1 and 2 modes can lead to uncertainties in the pred icted intensities and frequencies of the all-12C and its single 13C substituted isotopomers. In fact, in the 13 -12-12-12-12-12-12-32, 12-12-12-12-1213 -12-32 and 12-12-12-12-12-1213 -32 isotopomers, which show the largest deviations in differe nce frequency values (i.e., exp – cal), there is a very large intensity flow from the 2 to the 1 mode. Therefore, the 2085.6 and 2063.6 cm 1 band assignments for the 2 mode in these isotopomers (in Fi gure 3-2 and Figure 3-3 and Table 3-4) are considered tentative. The annealing spectra of Figure 3-3 support this statement.

PAGE 41

23 Table 3-4. Comparison of observed (Ar matr ix, 35 K) and calculated (B3LYP/6-311G* and B3LYP/cc-pVDZ) isotopomer frequencies (cm 1) of 2 mode for all-12C, singly-13C and all-13C substituted 12/13C7S linear carbon-sulfur clusters. Isotopomer exp a 6-311G* b cc-pVDZ c exp6-311G* expcc-pVDZ 12-12-12-12-12-12-12-32 2088.1 2088.1 2088.1 0.0 0.0 13 -12-12-12-12-12-12-32 2088.1 2088.1 2087.8 0.0 0.3 1213 -12-12-12-12-12-32 2085.6 2087.4 2085.1 1.8 0.5 12-1213 -12-12-12-12-32 2068.0 2067.3 2067.2 0.7 0.8 12-12-1213 -12-12-12-32 2060.5 2058.7 2060.2 1.8 0.3 12-12-12-1213 -12-12-32 2056.5 2056.5 2055.7 0.0 0.8 12-12-12-12-1213 -12-32 2063.6 2067.4 2063.7 3.8 0.1 12-12-12-12-12-1213 -32 2085.6 2087.0 2085.5 1.4 0.1 a Tentative assignment. b Scaled by 0.9392 factor. c Scaled by 0.9452 factor. Figure 3-2. IR absorption spectrum of the 2( ) mode region of neutral C7S and its 13C isotopic partners produced by laser ablation of a 1.0:0.15:1.0 mixture of 12C : 13C : 32S isolated in an Ar matrix at 35 K. Two other spectra for 12C/32S and 12C/13C mixtures under the same experime ntal conditions are included to confirm the band assignments. All-12C and singly-13C substituted isotopomeric bands are marked with dot s. Overlapped bands are marked with two dots. Bands marked by empty tria ngles and circles are assigned to 12/13C9 and 12Cn clusters, respectively.

PAGE 42

24 Figure 3-3. IR absorption spectra of the 2( ) mode region of neutral C7S and its 13C isotopic partners recorded after trapping the products in Ar at ca. 12 K, followed by annealing at 35 K. An analysis of the spectra in Figure 3-2 and Figure3-3 raises the question: why is the intensity of the 2085.6 cm 1 band as intense as it is relative to the 2088.1 cm 1 band? The carrier of this band could be a singly 13C substituted isotopomer of 12/13CnS, a 12/13CnS2 cluster or a doubly 13C substituted group of similar clusters, all of which might have a frequency close to 2085 cm 1. If only cluster sizes with n < 10 are considered, calculations show none of these possibilities are tenable. One further possibility is some band intensity comes from the 1 mode of some singly 13C substituted isotopomers of 12/13C7S. The B3LYP/cc-pVDZ (unscaled) frequencie s (with intensities in parentheses) of the 13 -12-12-12-12-12-1232, 12-12-12-12-1213 -12-32 and 12-12-12-12-12-1213 -32 isotopomers are 2235.7 (900), 2234.2 (3400) and 2234.1 cm 1 (2220 km/mol). When grouped together, these bands have a total integral intensity comparable to the 2 mode of

PAGE 43

25 12C7S ( cf. Table 3-1). A 0.9335 scaling factor fits these calculated frequencies to the 2085.6 cm 1 experimental band well ( cf. , Figure 3-2). Considering all the above difficulties in determining the singly 13C substituted isotopomer frequencies and intensit ies, the assignment of the 2088.1 cm 1 band to the 2 mode of C7S must be considered tentative. Isotopomer band assignments for the 3( ) vibration In the past the absorption at 1913.6 cm 1 has been assigned to the 4 mode of the 1213 13 -12-12-12 isotopomer of C6 (Figure 3-4(c)). There ar e, however, several reasons that it could alternatively be attributed to the 3 stretching mode of linear C7S. First, as shown in Figure 3-4(b), this peak, whose inte nsity is much larger than that of the 4 mode of the 1213 13 -12-12-12 C6 isotopomer, appears among the products from an ablated 12C / 32S mixture. Second, if this band were only produced by the 4 mode of a 12/13C isotopomer, its intensity should be similar under the same experimental conditions. However, even though the same 12C / 13C ratio was used in the mi xture, the intensity of the 1913.6 cm 1 band is dramatically higher ( cf. Figure 3-4a). Therefore, the 1913.6 cm 1 band is assigned to the 3 stretching mode of linear C7S. Comparisons of observed and calcula ted isotopomer frequencies of the 3 mode for all-12C, singly-13C substituted and all-13C substituted 12/13C7S linear carbon-su lfur clusters are shown in Table 3-5. The de viation of calculated and expe rimental values varies only from 1.4 to 1.3 cm 1, which indicates the reliability of these new assignments. This is confirmed by the good agreement of the is otopic shifts computed at different computational levels.

PAGE 44

26 Table 3-5. Comparison of observed (Ar matr ix, 35 K) and calculated (B3LYP/6-311G* and B3LYP/cc-pVDZ) isotopomer frequencies (cm 1) of 3 mode for all-12C, singly-13C and all-13C substituted 12/13C7S linear carbon-sulfur clusters. Isotopomer exp 6-311G* a cc-pVDZ b exp6-311G* expcc-pVDZ 12-12-12-12-12-12-12-32 1913.6 1913.6 1913.6 0.0 0.0 13 -12-12-12-12-12-12-32 1904.7 1904.5 1904.7 0.2 0.0 1213 -12-12-12-12-12-32 1884.4 1885.5 1885.8 1.1 1.4 12-1213 -12-12-12-12-32 1913.6 1912.9 1913.0 0.7 0.6 12-12-1213 -12-12-12-32 1902.2 1901.9 1901.8 0.3 0.4 12-12-12-1213 -12-12-32 1911.7 1912.8 1912.8 1.1 1.1 12-12-12-12-1213 -12-32 1895.9 1896.5 1896.9 0.6 1.0 12-12-12-12-12-1213 -32 1902.2 1901.3 1901.3 0.9 0.9 13 13 13 13 13 13 13 -32 1840.1 1838.8 1838.8 1.3 1.3 a Scaled by 0.9521 factor. b Scaled by 0.9480 factor. Figure 3-4. IR absorption spectrum of the 3( ) mode region of neutral C7S and its 13C isotopic partners produced by laser ablation of a 1:0.15:1 mixture of 12C: 13C: 32S in an Ar matrix at 35 K. Two other spectra for 12C/32S and 12C/13C mixtures under the same experimental conditions are shown to confirm the band assignments. All-12C and singly-13C substituted isotopomeric bands are marked with dots, and overlapped bands are marked by two dots. Peaks with filled squares are due to all-12C and singly-13C substituted of linear C7. Bands marked by an empty square, black pent agon, empty star, plus sign and black triangle are assigned to 12-13-13-12-12-12 (12/13C6), 12CnSm, 13Cn, 12CnSm, and 13Cn clusters, respectively.

PAGE 45

27 Figure 3-5. IR absorption spectra of the 3( ) mode region of neutral C7S and its 13C isotopic partners recorded after trapping the products in Ar at ca. 12 K, followed by annealing at 35 K. Although it is easy in theory to distinguish between two close-lying vibrational frequencies, this is usually difficult in pract ice. In some cases more than one isotopomer may contribute to a band envelope. For example, the 12-1213 -12-12-12-12-32 isotopomer (1913.0 cm 1, cf. , Table 3-5) is predicted to lie within the envelope of the total 12C isotopomer observed at 1913.6 cm 1. Furthermore, the net absorption of the singly-13C-substituted isotopomer is usually larger than that of a doubly-13C-substituted peak. Figure 3-5 makes clear that , after annealing, not only do es the signal for the singly-13C-substituted isotopomer get stronger , but the intensities for the doubly-13C-substituted species are barely visi ble above the baseline . Since the calculated energy band separations for many doubly-13C-substituted C7S isotopomers are smaller than the exp – cal energy differences, a reliab le assignment of these bands becomes problematic.

PAGE 46

28 Interestingly, the intensities of the 12/13C7 isotopomer bands do not change during annealing. In previous work on pure carbon clusters, 36 it was found that the C7 IR band also does not change upon annealing. This was interpreted to mean that the formation of C7 from smaller carbon cluste rs and the reaction of C7 to form larger clusters occur with about equal rates. In the present case, if C7S were formed from C7 and S, it would be expected that the C7 peaks would decrease due to this new reaction channel. Since no band intensity decrease was observe d, it can be concluded that C7S probably forms from smaller clusters such as CnS and C7nS; n = 1–6, vide infra . Isotopomer band assignments for the 5( ) vibration Ablation and trapping of a 12C / 13C mixture leads to only a single band in 1235– 1260 cm 1 region, and this has been assigned to 12C12 ( cf. Figure 3-6c).35 For a 12C / 32S mixture, only one peak at 1256.1 cm 1 is observed ( cf. Figure 3-6b), which is here assigned to the 5 stretching mode of C7S. With a mixture of 12C / 13C / 32S, a number of bands between 1245 and 1250 cm 1 are seen in addition to the strong band at 1256.1 cm 1 and another due to13CS. Assignments for these new p eaks emerge from theoretical predictions of the isotop ic shifts of linear C7S. Table 3-6 lists all the observed and calculated isotopomer 5 mode frequencies of all-12C, singly-13C, and all-13C substituted linear 12/13C7S clusters. The experimental frequencie s are consistent with the predicted ones from different theoretical levels, with discrepancies of not more than 0.6 cm 1. The 5 frequencies of the 1213 -12-12-12-12-12-32 and 12-12-12-1213 -12-12-32 isotopomers are exactly the same as that of the all-12C isotopomer observed at 1256.1 cm 1. As can be seen in Figure 3-1, this result is due to the fact that this vibrational mode has a node at the position of the 13C substitutional site in these two isotopomers.

PAGE 47

29 Table 3-6. Comparison of observed (Ar matr ix, 35 K) and calculated (B3LYP/6-311G* and B3LYP/cc-pVDZ) isotopomer frequencies (cm 1) of 5 mode for all-12C, singly-13C and all-13C substituted 12/13C7S linear carbon-sulfur clusters. Isotopomer exp 6-311G* a cc-pVDZ b exp6-311G* expcc-pVDZ 12-12-12-12-12-12-12-32 1256.1 1256.1 1256.1 0.0 0.0 13 -12-12-12-12-12-12-32 1246.9 1246.8 1246.8 0.1 0.1 1213 -12-12-12-12-12-32 1256.1 1256.1 1256.1 0.0 0.0 12-1213 -12-12-12-12-32 1246.3 1245.8 1245.8 0.5 0.5 12-12-1213 -12-12-12-32 1246.9 1246.5 1246.5 0.4 0.4 12-12-12-1213 -12-12-32 1256.1 1256.1 1256.1 0.0 0.0 12-12-12-12-1213 -12-32 1248.5 1248.3 1248.3 0.2 0.2 12-12-12-12-12-1213 -32 1245.0 1244.7 1244.6 0.3 0.4 13 13 13 13 13 13 13 -32 1210.7 1210.2 1210.1 0.5 0.6 a Scaled by 0.9777 factor. b Scaled by 0.9739 factor. Figure 3-6. IR absorption spectrum of the 5( ) mode region of neutral C7S and its 13C isotopic partners produced by laser ablation of a 1.0:0.15:1.0 mixture of 12C : 13C : 32S isolated in an Ar matrix at 35 K. Two other spectra for 12C/32S and 12C/13C mixtures under the same experime ntal conditions are included to confirm the band assignments. All-12C and singly-13C substituted isotopomeric bands are marked with dot s, and overlapped bands are marked with two dots.

PAGE 48

30 Figure 3-7. IR absorption spectra of the 5( ) mode region of neutral C7S and its 13C isotopic partners recorded after trapping the products in Ar at ca. 12 K, followed by annealing at 35 K. Figure 3-7 shows that th e absorbance of the 5 mode (for all-12C and singly-13C substituted isotopomers) increases with te mperature during annealing. The simultaneous decrease in the 13CS absorption peak may be associ ated with the formation of C7S. Although ( cf. Figure 3-11) there is no direct information on how C7S forms in the matrix, B3LYP/cc-pVDZ calculations show that the lowest energy pathway to C7S occurs via reaction of C6 and CS. Since (CS)2 dimers have been reported in Ar matrices, 55 the CS moiety is probably mobile in solid Ar. The CS fragment is known to take part in the production of some CnS clusters,56 so it is reasonable to assume that CS may be a precursor for linear C7S. Other formation channels are, however, also possible, and could involve the mobile C and S species.

PAGE 49

31 Theoretical Calculations for Carbon-Sulfur Clusters CnS and SCnS ( n 29) Energies and Structures Table 3-7 shows the calcu lated total energies (En), zero point vibrational energies (ZPEn), spin expectation values , and rotational constants (Be) for neutral linear CnS, and SCnS clusters in their electronic ground states. Di pole moments ( µe) for the CnS clusters are also presented. Experi mental rotational constants for CnS ( n = 1–9) are also given for comparison.45,46,57-59 Table 3-7. Energies En (Hartrees), zero poi nt energies ZPEn (kcal/mol), spin contamination and rotational constants Be (GHz) for CnS and CnS2 clusters, plus dipole moments µe (Debyes) for CnS clusters, calculated at the B3LYP/6-311G(d) level. CnS SCnS n En ZPEn Be a µe En ZPEn Be 1 436.24615 1.86030 0.00000 24.39410 1.8395 834.55047 4.30747 0.00000 3.24669 2 474.27265 4.45174 2.01558 6.42558 2.8779 872.59243 6.94178 2.01177 1.55975 3 512.40260 8.18942 0.00000 2.87986 3.4971 910.71833 10.36653 0.00000 0.91329 4 550.45638 10.64824 2.02801 1.51406 4.0314 948.77566 13.24851 2.02049 0.57648 5 588.57426 14.29799 0.00000 0.92092 4.6446 986.89099 16.97687 0.00000 0.39575 6 626.63787 16.62491 2.04042 0.59618 5.0346 1024.95655 19.29428 2.03023 0.28171 7 664.74829 20.36717 0.00000 0.41406 5.7301 1063.06561 23.06634 0.00000 0.21017 8 702.81776 22.29385 2.05281 0.29738 6.0300 1101.13601 25.44852 2.04056 0.16023 9 740.92341 27.16976 0.00000 0.22267 6.8011 1139.24105 29.08632 0.00000 0.12590 10 778.99664 29.66980 2.06535 0.17029 7.0051 1177.31491 31.37436 2.05139 0.10038 11 817.09915 33.68150 0.00000 0.13391 7.8332 1215.41698 35.10090 0.00000 0.08173 12 855.17537 35.97567 2.07821 0.10683 8.0112 1253.49330 37.70984 2.06274 0.06725 13 893.27545 39.75851 0.00000 0.08693 8.8789 1291.59319 41.31972 0.00000 0.05621 14 931.35356 42.16937 2.09147 0.07152 8.9766 1329.67152 43.83912 2.07454 0.04735 15 969.45181 45.68840 0.00000 0.05972 9.9490 1367.76969 47.51548 0.00000 0.04038 16 1007.53180 48.42325 2.10506 0.05027 9.9864 1405.84956 49.96121 2.08682 0.03465 17 1045.62861 52.22259 0.00000 0.04281 10.9889 1443.94619 53.75679 0.00000 0.03002 18 1083.70965 54.57750 2.11923 0.03670 10.9623 1482.02740 55.60133 2.09955 0.02614 19 1121.80534 57.88523 0.00000 0.03176 12.0738 1520.12279 59.62620 0.00000 0.02295 20 1159.88746 60.19595 2.13375 0.02762 11.9707 1558.20510 61.98519 2.11270 0.02022 21 1197.98232 64.62252 0.00000 0.02421 13.2798 1596.29931 67.05173 0.00000 0.01794 22 1236.06540 66.24455 2.14860 0.02132 13.0325 1634.38215 69.14877 2.12636 0.01597 23 1274.15909 70.12173 0.00000 0.01889 14.4010 1672.47637 73.90505 0.00000 0.01430 24 1312.24307 72.37502 2.16418 0.01680 14.0322 1710.56047 74.01673 2.14057 0.01284 25 1350.33626 76.18697 0.00000 0.01502 15.6000 1748.65276 78.91588 0.00000 0.01159 26 1388.42076 79.19313 2.17987 0.01348 15.1178 1786.73826 79.84397 2.15526 0.01048 27 1426.51329 82.32376 0.00000 0.01215 16.7954 1824.83021 84.90048 0.00000 0.00952 28 1464.59848 84.41037 2.19686 0.01098 16.1545 1862.91590 85.96655 2.17036 0.00867 29 1502.69034 88.36385 0.00000 0.00996 18.0414 1901.00738 91.21811 0.00000 0.00792 a B0, experimental rotational constants for CnS ( n = 1–9) are 24.49550, 6.47775, 2.89038, 1.51916, 0.92270, 0.58251, 0.41443, 0.29781, and 0.22272GHz, respectively.45,46,57-59

PAGE 50

32 The predicted electroni c spin expectation values s how that the ground states of these two cluster series a lternate between singlets (1+) and triplets (3), as expected. Odd n clusters have fully-occupied orbitals, while the highest-lying orbital in even n clusters is half-occupied. The calculated el ectronic spins for the triplet ground states show only a very small degree of contamination. Energies of the geometry-optimized CnS and SCnS clusters are linearly dependent on chain length. The difference between the energy gap of CnS and Cn +1S clusters and SCnS and SCn +1S clusters is small, and close to the value of E [Cn] E [Cn +1], indicating that the sulfur atom does not dramati cally change the properties of the Cn chains. This reflects the similar ionization energies a nd electronegativities of carbon and sulfur.86 To test the reliability of the geometries predicted at the B3LYP/6-311G(d) level, bond lengths and rotational constants were also calculated at the B3LYP/cc-pVDZ, B3LYP/aug-cc-pVTZ, MP2, and CCSD(T)/cc-pVDZ levels for C2S, C6S and C7S2 ( cf. , Table 3-8). The predicted Be values for C2S at the MP2 and CCSD(T) levels are in poorer agreement with experiment than the ones calculated at the B3LYP level ( cf. , Table 3-8). On the other hand, the CCSD(T) Be value for C6S differs more from the experimental B0 value than the B3LYP and MP2 values do. It is interesting to note that B3LYP/aug-ccpVTZ predicts the best Be value for C2S, while MP2 gives the best values for C6S. However, the B3LYP/6-311G(d) predicted Be values for C2S and C6S, i.e., 6.42558 GHz and 0.59618 GHz, respectively, ( cf. , Table 3-7) differ from the best predicted Be values ( cf. , Table 3-8) by only 0.6% and 2.6%. Table 3-8 shows that the S–C1 and Ci–Cj bond lengths computed using DFT theory are always shorter than thos e obtained with ab initio th eory. Furthermore, the DFT Ci–Cj

PAGE 51

33 bond lengths in C6S show a noticeable alternation, indicating a slight bond conjugation. However, MP2 and CCSD(T) results show that the Ci–Cj bond lengths in C6S are almost identical, corresponding to a cumulenic structure. Table 3-8. Energies E (Hartrees), dipole moments µe (Debyes), bond lengths (Å), and rotational constants Be (GHz), calculated for linear C2S, C6S, and C7S2 clusters at B3LYP/6-311G(d), B3 LYP/cc-pVDZ, B3LYP/aug-cc-pVTZ, MP2/cc-pVDZ and CCSD(T)/cc-pVDZ levels of theory. CnSm Theory E µe S C1C1 C2C2 C3C3 C4C4 C5 C5 C6Be a C2S B3LYPb 474.2727 2.8779 1.5752 1.3137 6.4256 B3LYPc 474.2552 2.8350 1.5840 1.3244 6.3417 B3LYPd 474.2920 3.0645 1.5711 1.3086 6.4655 MP2 473.4542 2.9619 1.5863 1.3361 6.2872 CCSD(T) 473.5035 2.9434 1.5930 1.3400 6.2407 C6S B3LYPb 626.6379 5.0346 1.5667 1.2747 1.2811 1.2717 1.2908 1.2938 0.5962 B3LYPc 626.5936 5.0279 1.5756 1.2844 1.2908 1.2815 1.3002 1.3032 0.5877 B3LYPd 626.6743 5.3880 1.5632 1.2728 1.2791 1.2691 1.2889 1.2884 0.5986 MP2 625.2895 6.3227 1.5850 1.2951 1.2971 1.2893 1.3073 1.3057 0.5810 CCSD(T) 625.3825 6.3808 1.5840 1.2983 1.3028 1.2949 1.3141 1.3187 0.5768 C7S2 B3LYPb 1063.0656 0 1.5661 1.2731 1.2768 1.2748 0.2102 B3LYPc 1063.0057 0 1.5741 1.2832 1.2867 1.2844 0.2072 B3LYPd 1063.1144 0 1.5627 1.2713 1.2750 1.2726 0.2109 MP2 1061.0898 0 1.5789 1.2991 1.2980 1.2981 0.2036 CCSD(T) 1061.1483 0 1.5851 1.2976 1.3003 1.2987 0.2031 a B0, experimental rotational constants for C2 32S and C6 32S are 6.4778 and 0.5825 GHZ, respectively.45,59 Note that the predicted Be values at MP2 and CCSD(T) levels for C2S are in poorer agreement with experiment than the Be values calculated at the B3LYP level. Also, the Be (CCSD(T)) value in C6S departs more from the experimental value than the Be (B3LYP) and Be (MP2) values. b Calculated at B3LYP/6-311G(d) level. c Calculated at B3LYP/cc-pVDZ level. d Calculated at B3LYP/aug-cc-pVTZ level. The dipole moments of C2S and C6S from MP2 computations are similar to those from CCSD(T); both are larger than the valu es obtained from DFT at different levels, except for C2S from B3LYP/aug-cc-pVDZ. Since theoretical rotational constants are in good agreem ent with experiment, the structural parameters computed by DFT are appropriate for sulfur-doped carbon clusters. As chain length increases, rotational consta nts decrease almost exponentially, especially for clusters with larger n . As n increases, the differences in Be ,n decrease dramatically. Dipole moments for the CnS clusters increase almost linearly with increasing chain

PAGE 52

34 length, but for the larger clusters, small fluctu ations with an odd/ev en alternation appear. The dipole moments for SiCn ( n = 2–10) clusters also s how a similar fluctuating behavior.87 This is probably related to the singlet and triplet alternation of heteroatomcontaining carbon clusters. The sulfur atoms in both CnS and SCnS clusters bear a positive charge, varying from 0.25 e to 0.14 e , while the adjacent C atom possesse s a small negative charge. For CnS clusters, the charge on the terminal carbon atom is always close to –0.1 e , while the next-closest carbon bears the highest absolute negative charge. In SCnS species, most of the charge is located at the center of the chain, especially for the longer chains. All S–C and C–C bonds are weakly polarized, which is attributable to the small electronegativity differences between C and S. Figure 3-8 shows the S0–C1, C1–C2, and C2–C3 bond lengths for the CnS and SCnS clusters as a function of nu mber of carbon atoms. The S0–C1 bond lengths for CnS and SCnS clusters exhibit the same odd/even alte rnation. With increasing number of carbons, the S0–C1 bond length increases monotonically for odd n , while decreasing for even n . The odd/even difference gets smaller as the chain length increases. The S0–C1 lengths in CnS are much smaller than in SCnS for the same n , indicating that th e introduction of one more sulfur atom decreases the S0–C1 bond strength. This is also reflected in the lowering of the S0–C1 bonding strengths and stre tching frequencies in SCnS vs. CnS clusters ( vide infra ). No appreciable variation in the C–C bond le ngths is seen along the carbon chain, which suggests mostly cumulenic bonding in CnS and SCnS. However, the bond lengths are affected by the multiplicity of the gr ound state. As Figure 3-9B shows, C–C bond

PAGE 53

35 Figure 3-8. The S0–C1 (A), C1–C2 (B), and C2–C3 (B) bond lengths for linear CnS and SCnS clusters as a function of number of carbon atoms, calculated at the B3LYP/6-311G(d) level. lengths show a strong odd/ even alternation with n as S0–C1 bond lengthens. But the lengths of the C1–C2 bond adjacent to S exhibit a smaller odd/even effect. Also, the oddeven bond lengths differences are smaller fo r two sulfurs in the chain than for one. Interestingly, for the SCnS clusters with odd n , the C1–C2 bond lengths are almost

PAGE 54

36 identical, while for even n clusters the bond lengths incr ease a little with increasing n . But for CnS clusters, the odd-even C1–C2 length variation is reversed. The C1–C2 bond lengths in even clusters have similar va lues, while odd clusters display a decreasing pattern. When observing the C2–C3 bond lengths, a similar vari ation can be seen for CnS and SCnS clusters. Therefore, it appears that the addition of a sulfur atom decreases and reverses the variation of the adj acent C–C bond lengths in the carbon chain. Vibrational Properties No imaginary frequencies were found for any of the molecules studied, indicating that the linear geometries are indeed at local potential energy minima. Figure 3-9 shows the vibrational frequencies of the highest intensity mode for the Cn, CnS and SCnS clusters, as a function of n , while Figure 3-10A and Figure 3-10B present all the infraredactive stretching frequencies for the CnS and SCnS clusters ( n = 1–29). Figure 3-9 illustrates that the frequencie s of the most intense modes of the Cn, CnS, and SCnS clusters with odd n reach a maximum at n = 5 and then decrease with increasing numbers of carbon atoms. For clusters with the same number of carbons (with n > 5), band frequencies decrease in order Cn > CnS > SCnS. Even n clusters also obey the same trend, but the frequencies reach a maximum at n = 10, and then decrease. Although the intensities of these modes generally increase as the carbon chain increases, the pattern is not a simple one. Some bands have, howev er, very large computed intensities. For example, the C29S band at 1535.82 cm 1 has an intensity of 1.9×105 km/mol, while the SC29S band at 1495.31 cm 1 has a slightly larger intensity, 2.2×105 km/mol. Although the absorption intensity increases as a function of the size of the linear chain, such large intensities are probably not physically reas onable and may be due to wave function instabilities. However, it is stil l clear that observation of thes e species should be possible,

PAGE 55

37 but may be difficult due to unknown cluster abundances in any IR measurements, terrestrial or celestial. Figure 3-9. Vibrational frequencies of the most intense (C–C) mode for linear Cn, CnS and SCnS clusters computed at B3LYP/6-311G(d) level. Table 3-9 and Table 3-10 present the calculated infrared-active stretching frequencies for the CnS and SCnS clusters in their ground states . In both tables, the lowest frequencies represent the C–S stretching modes, while all th e others are the various C–C chain stretching modes. It can be seen th at the C–S vibrational frequencies decrease almost exponentially as the number of carbon atoms increases. Absolute intensities are also included in the tables , although they are notoriously difficult to measure experimentally. All the vibrational stretching frequencies, obtained at the B3LYP/6-311G(d) level, are plotted as a function of the number of carbon atoms in Figure 3-10. More vibrational fundamentals are squeezed into the high-en ergy region as the nu mber of carbon atoms

PAGE 56

38 Table 3-9. Calculated (B3LYP/6-311G (d)) vibrational frequencies (cm 1, unscaled) for linear CnS ( modes; n = 1–29) carbon-sulf ur clusters in their electronic ground states. Absolute integral IR intensities (km/mo l) for predicted frequencies are in parentheses. n Vibrational Frequencies & Integral Intensities 1 1301 (87) 2 860 (14), 1724 (43) 3 738 (13), 1566 (66), 2145 (1528) 4 612 (7), 1223 (5), 1810 (530), 2117 (33) 5 543 (7), 1091 (80), 1655 (623), 2088 (0), 2239 (4003) 6 476 (4), 938 (32), 1409 (275), 1837 (8), 2090 (1498), 2133 (38) 7 432 (3), 855 (86), 1285 (284), 1712 (139), 2010 (2859), 22091 (5750), 2230 (231) 8 390 (3), 765 (51), 1145 (171), 1517 (28), 1843.72 (1150), 2041 (15), 2097 (1), 2170 (2665) 9 359 (2), 708 (94), 1058 (172), 1406 (146), 1735 (1375), 1994 (9), 2109 (13747), 2197 (904), 2261 (48) 10 330 (1), 648 (61), 966 (120), 1281 (66), 1580 (667), 1851 (0), 1995 (15), 2029 (2971), 2170 (215), 2177 (3256) 11 307 (1), 606 (94), 903 (122), 1199 (168), 1484 (720), 1 759 (312), 1956 (6652), 2074 (16949), 2129 (126), 2233 (504), 2250 (119) 12 286 (1), 563 (64), 837 (89), 1110 (99), 1375 (465), 1627 ( 73), 1851 (1855), 1917 (4), 2019 (3), 2114 (8294), 2169 (841), 2204 (35) 13 268 (0), 530 (93), 789 (92), 1046 (188), 1298 (507), 1540 (248), 1767 (2874), 1953 (140), 1990 (31509), 2103 (93), 2192 (3), 2224 (2), 2246 (1933) 14 252 (1), 497 (67), 739 (68), 979 (123), 1216 (354), 1444 (108), 1657 (1124), 1836 (8), 1867 (0), 1989 (5061), 2105 (9156), 2117 (629), 2195 (863), 2207 (59) 15 238 (0), 471 (94), 702 (71), 930 (203), 1155 (393), 1374 (263), 1580 (1470), 1782 (894), 1908 (33271), 1965 (11982), 2067 (22), 2169 (1143), 2182 (81), 2234 (396), 2242 (2532) 16 225 (0), 445 (69), 663 (54), 877 (141), 1090 (284), 1297 (1 43), 1495 (842), 1682 (138), 1774 (11), 1855 (2934), 1983 (2), 2047 (17828), 2105 (164), 2168 (14), 2199 (5), 2211 (2350) 17 214 (0), 424 (92), 632 (57), 837 (213), 1040 (313), 1239 (288), 1430 (1059), 1612 (482), 1783 (9677), 1875 (51406), 1931 (85), 2049 (237), 2127 (0), 2158 (0), 2204 (5169), 2228 (1303), 2245 (205) 18 203 (0), 403 (70), 600 (43), 795 (154), 987 (234), 1177 (175), 1360 (671), 1535 (168), 1697 (1409), 1707 (349), 1859 (17), 1959 (11254), 2023 (13984), 2073 (58), 2153 (2006), 2168 (48), 2208 (41), 2214 (3115) 19 194 (0), 386 (92), 575 (45), 761 (219), 946 (260), 1128 (312), 1305 (832), 1476 (455), 1635 (3438), 1791 (20859), 1816 (54674), 1927 (93), 2027 (3), 2101 (66), 2119 (2289), 2191 (1334), 2197 (6403), 2234 (460), 2242 (311) 20 186 (0), 368 (70), 5496 (35), 726 (163), 903 (198), 1077 (2 03), 1246 (560), 1411 (201), 1566 (1339), 1636 (18), 1717 (222), 1854 (5309), 1959 (1267), 1967 (26602), 2059 (489), 2123 (0), 2153 (6), 2184 (6183), 2209 (1560), 2219 (110) 21 178 (0), 353 (92), 527 (36), 698 (222), 868 (220), 1036 (333), 1200 (686), 1359 (461), 1511 (2276), 1656 (1748), 1748 (90369), 1808 (5292), 1917 (18), 2014 (1068), 2056 (0), 2105 (8), 2154 (10837), 2184 (1742), 2216 (13), 2232 (13), 2240 (2101) 22 171 (0), 339 (71), 505 (28), 669 (168), 832 (170), 993 (225), 1150 (480), 1304 (234), 1452 (1094), 1569 (91), 1595 (184), 1726 (2755), 1859 (134), 1913 (36790), 1960 (1385), 2039 (10), 2111 (673), 2116 (3022), 2177 (142), 2186 (8672), 2214 (429), 2221 (195) 23 164 (0), 326 (94), 487 (29), 645 (225), 802 (189), 957 ( 352), 1110 (585), 1259 (480), 1403 (1728), 1542 (1125), 1665 (33331), 1705 (86984), 1806 (164), 1913 (390), 1995 (2), 2026 (11), 2079 (4513), 2140 (12268), 2151 (602), 2203 (828), 2214 (509), 2233 (240), 2239 (2730) 24 158 (0), 314 (72), 468 (23), 620 (170), 771 (149), 921 (243) , 1068 (418), 1212 (265), 1351 (927), 1485 (318), 1507 (8), 1614 (2010), 1737 (483), 1849 (21483), 1887 (25505), 1943 (42), 20288 (2013), 2074 (2), 2108 (1), 2149 (11635), 2177 (3004), 2199 (46), 2217 (31), 2221 (2183) 25 152 (0), 303 (96), 452 (24), 599 (226), 746 (166), 890 (369), 1033 (510), 1173 (509), 1309 (1416), 1441 (970), 1564 (10017), 1644 (135298), 1688 (2483), 1810 (18), 1906 (16), 1980 (203), 1990 (2072), 2069 (72), 2103 (19856), 2140 (231), 2180 (5), 2200 (3), 22 19 (4022), 2231 (1597), 2239 (542) 26 147 (0), 292 (74), 436 (18), 578 (173), 719 (129), 859 (259) , 996 (371), 1132 (292), 1264 (807), 1392 (332), 1443 (52), 1515 (1634), 1633 (392), 1743 (5660), 1835 (51322), 1860 (391 ), 1938 (735), 2012 (0), 2052 (49), 2083 (6228), 2143 (5726), 2147 (8797), 2190 (1039), 2203 (38.0), 2218 (24), 2223 (3126) 27 142 (1), 283 (99), 422 (20), 560 (229), 697 (145), 832 ( 384), 965 (454), 1097 (538), 1226 (1214), 1351 (924), 1471 (5630), 1581 (87161), 1596 (85055), 1700 (299), 1811 (32), 1904 (1256), 1939 (7), 1984 (2), 2047 (10847), 2084 (15389), 2117 (70), 2169 (1367), 2174 (320), 2210 (225), 2218 (6433), 2233 (432), 2239 (615) 28 137 (0), 273 (76), 408 (13), 541 (172), 673 (110), 804 (265) , 934 (346), 1061 (336), 1186 (703), 1308 (3586), 1380 (65), 1426 (1385), 1540 (393), 1647 (3539), 1752 (2384), 1791 (61936) , 1861 (16), 1930 (14), 2005 (3868), 2016 (24), 2075 (9), 2112 (21361), 2140 (1301), 2172 (15), 2194 (17), 2205 (4969), 2220 (1979), 2224 (278) 29 133 (1), 265 (102), 396 (17), 525 (229), 654 (129), 781 (396), 906 (409), 1031 (570), 1152 (1066), 1272 (927), 1386 (4026), 1498 (10990), 1536 (194838), 1608 (2290), 1708 (2), 1814 (637), 1893 (6), 1912 (43), 1971 (4443), 2042 (3255), 2048 (25298), 2106 (196), 2143 (0), 2163 (0), 2191 (7219), 2208 (2889), 2224 (26), 2232 (53), 2237 (2423)

PAGE 57

39 Table 3-10. Calculated (B3LYP/6-311G (d)) vibrational frequencies (cm 1, unscaled) for linear SCnS ( u modes; n = 1–29) clusters in thei r electronic ground states. Absolute integral IR intensities (k m/mol) for predicted frequencies in parentheses. n u Vibrational Frequencies & Integral Intensities 1 1555 (736) 2 1179 (150) 3 1037 (386), 2177 (3288) 4 882 (173), 1943 (925) 5 808 (295), 1737 (1742), 2213 (6280) 6 721 (175), 1527 (700), 2128 (2188) 7 671 (251), 1403 (1190), 2041 (7096), 2182 (7561) 8 615 (168), 1269 (665), 1908 (1855), 2170 (3514) 9 577 (2290), 1181 (975), 1775 (3251), 2077 (19808), 2202 (8) 10 537 (156), 1089 (625), 1641 (1355), 2063 (5014), 2166 (5529) 11 508 (196), 1023 (854), 1548 (2315), 1967 (20561), 2052 (21132), 2233 (1082) 12 477 (147), 955 (591), 1446 (1274), 1891 (3055), 2107 (11506), 2174 (0) 13 453 (178), 906 (774), 1371 (1925), 1789 (7258), 1953 (38344), 2121 (138), 2238 (2381) 14 429 (137), 854 (559), 1291 (1193), 1694 (2158), 2013 (10369), 2089 (10820), 2200 (1451) 15 410 (165), 815 (713), 1229 (1705), 1621 (4554), 1879 (55330), 1977 (1422), 2177 (2491), 2229 (2761) 16 390 (129), 773 (530), 1165 (1128), 1542 (2025), 1880 (5150), 2027 (212700), 2120 (293), 2210 (2747) 17 374 (159), 740 (673), 1114 (1598), 1478 (3632), 1785 (37987), 1847 (38548), 2067 (937), 2195 (7802), 2224 (78) 18 357 (123), 707 (503), 1063 (1078), 1411 (1891), 1725 (3331), 1965 (27269), 2017 (3784), 2163 (3415), 2209 (2937) 19 343 (147), 680 (623), 1021 (1471), 1357 (3103), 1662 (14157), 1773 (84030), 1946 (73), 2128 (4898), 2182 (7000), 2233 (926) 20 329 (118), 652 (4766), 977 (1041), 1300 (1797), 1600 (3081), 1871 (11820), 1945 (28606), 2077 (1521), 2183 (8568), 2208 (279) 21 318 (140), 630 (592), 943 (1390), 1254 (2844), 1546 (8704), 1710 (113144), 1819 (868), 2031 (2244), 2143 (14271), 2185 (11), 2236 (2014) 22 306 (114), 605 (451), 906 (1004), 1205 (1723), 1490 (2837), 1744 (5648), 1893 (45411), 1978 (62), 2126 (6151), 2178 (8053), 2216 (900) 23 296 (136), 586 (560), 876 (1344), 1165 (2638), 1441 (6833), 1647 (131889), 1709 (11809), 1928 (932), 2086 (10854), 2128 (9235), 2206 (1939), 2232 (2307) 24 285 (111), 565 (429), 844 (967), 1122 (1656), 1392 (2659), 1639 (4820), 1842 (51977), 1885 (4856), 2045 (3429), 2146 (16168), 2177 (250), 2220 (2354) 25 276 (133), 548 (536), 818 (1293), 1087 (2535), 1349 (5786), 1577 (97331), 1621 (78972), 1819 (269), 2004 (4342), 2087 (21882), 2147 (152), 2215 (6086), 2230 (221) 26 267 (109), 530 (411), 790 (9358), 1051 (1608), 1305 (2531), 1544 (4287), 1755 (15898), 1814 (53366), 1957 (1722), 2092 (11348), 2136 (12204), 2194 (2208), 2220 (2737) 27 259 (132), 514 (517), 767 (1248), 1020 (2477), 1267 (5292), 1 496 (46420), 1554 (165507), 1721 (1), 1915 (2313), 2046 (25565), 2079 (4145), 2175 (3230), 2209 (6054), 2233 (746) 28 251 (108), 499 (394), 743 (905), 988 (1569), 1229 (2438), 1459 (3936), 1667 (9679), 1769 (73017), 1873 (458), 2020 (6222), 2104 (25112), 2146 (21), 2205 (6943), 2220 (619) 29 244 (131), 485 (503), 723 (1226), 961 (2416), 1195 (4971) , 1416 (28353), 14951 (222506), 1633 (244), 1821 (1252), 1982 (10672), 2031 (27039), 2114 (1091), 2186 (10788), 2206 (256), 2234 (1901) increases. With an additional sulfur atom, the SCnS species have ( n + 1) stretching fundamentals. But, because of their high molecular symmetry, only a portion of these frequencies are infrared active, cf. , Figure 3-10B. For SCnS with even n , the number of u vibrational frequencies is equal to n /2, while the number of in frared-active stretching fundamentals for odd SCnS is ( n + 1)/2. In Figure 3-10, each line corresponds to a particular vibrational mode. In the low-en ergy region, these lines are well separated.

PAGE 58

40 Although carbon-sulfur clusters exhibit strong od d / even parity, vibrational lines for odd and even clusters fall on the same smooth e xponential curves, similar to those expected from the harmonic oscillator model. When observing the relatively high-energy region (higher than 1900 cm 1), the situation becomes more co mplicated. The trend lines cross, and the vibrational fundamentals are denser , making experimental assignments more difficult. Table 3-11. Comparison of expe rimental (Ar matrix) and calculated (B3LYP/6-311G(d)) most intense IR mode frequencies (cm 1) for linear CnS ( modes; n = 1–7) and SCnS ( u modes; n = 1–7) carbon-sulfur clusters in their electronic ground states. Relative IR intensities for pr edicted and experimental frequencies in parentheses. CnS SCnS n Mode cal a / cm 1 exp / cm 1 n Mode cal a / cm 1 exp / cm 1 1 (1+) 1; C–S 1284 (1.0) 1275.1 (1.0)b,c 1 (1g +) 2; C–S 1534 (1.0) 1528.2 (1.0)b,c 2 (3) 1; C–C 2; C–S 3 1647 (1.0)e 848 (0.33)e 254 (0.16)d,e 2 (3g ) 3; C–S 1163 (1.0) 1179.7 (1.0)c 3 (1+) 1; C–C 2; C–C 3; C–S 2050 (1.0) 1497 (0.043) 729 (0.0085) 2047.6 (1.0)b,c 1533.2 (0.1)c 725.6 (0.009)c 3 (1g +) 3; C–C 4; C–S 2080 (1.0) 1023 (0.12) 2078.5 (1.0)b,c 1024.6 (0.18)b,c 4 (3) 1; C–C 2; C–C 4; C–S 2023 (0.062) 1729 (1.0) 603 (0.013) 1746.8 (1.0)b 4 (3g ) 4; C–C 5; C–S 1856 (1.0) 870 (0.19) 1872.1 (1.0)c 897.7 (0.117)c 843.7 (0.056)c 5 (1+) 1; C–C 3; C–C 4; C–S 2140 (1.0) 1581 (0.16) 1076 (0.020) 2124.5 (1.0)b 5 (1g +) 4; C–C 5; C–C 6; C–S 2115 (1.0) 1660 (0.28) 797 (0.047) 2104.7 (1.0)c 1687.9 (0.36)c 783.5 (0.04)c 6 (3) 1; C–C 2; C–C 4; C–C 2039 (0.025) 1998 (1.0) 1346 (0.18) 6 (3g ) 5; C–C 6; C–C 7; C–S 2034 (1.0) 1459 (0.32) 712 (0.080) 7 (1+) 2; C–C 3; C–C 5; C–C 2111 (1.0) 1921 (0.50) 1228 (0.049) 2088.1 (0.67)c,f 1913.6 (1.0)c 1256.1 (0.07)c 7 (1g +) 5; C–C 6; C–C 7; C–C 2085 (1.0) 1951 (0.94) 1341 (0.16) a Scaled uniformly by 0.9556 for modes with primarily C C character and by 0.9866 for modes with primarily C S character. b Reference 56. c References 51-54, 88. d CCS bending vibration mode. e Harmonic frequencies for C2S in gas phase are 1708.2 ( 40), 862.5 (19), and 269.3 (13) cm 1 for 1, 2, and 3, respectively.89 f Tentative assignment. To test the reliability of the harmoni c frequency calculati ons at the B3LYP/6311G(d) level in CnS and SCnS clusters, the predicted and observed vibrational frequencies are compared in Table 3-11. Com puted relative absorption intensities can be

PAGE 59

41 Figure 3-10. Number of carbon atoms n is plotted as a function of stretching frequencies of and u mode for CnS (A) and SCnS (B) clusters calculated at B3LYP/6311G(d) level, respectively. Frequencie s are scaled uniformly by an optimum scaling factor of = 0.9556 (see text) for modes with primarily C C stretch character and by = 0.9866 for modes with primarily C S stretch character.

PAGE 60

42 seen to match experimental values well, ex cept for the high-energy modes. All available experimental data up to C7S are included.51-54,56,88,89 Harmonic frequencies (B3LYP/6311G(d)) were scaled unif ormly by a factor of = 0.9556 for C C stretching modes and by = 0.9866 for C S stretching modes ( cf. , Figure 3-10). The optimum scaling factors, , were generated by minimizing the residuals 2 expn i i theo i for all known linear carbon-sulfur clusters. Here i theo and exp i are the theoretical (B3LYP/6-311G(d)) and experimental (Ar matrix) ith fundamental mode frequencies, respectively. This procedure is similar to one desc ribed by Scott and Radom.90 A standard deviation of only 1 cm 1 between calculated and experimental C C stretches was found ( cf. , Table 3-11). Based on the above discussion, it is c oncluded that pure carbon and sulfur-doped carbon clusters not only possess similar struct ures and odd-even alternation effects, but also have similar vibrational properties. Th is finding may be important in a comparison of predicted and observed vibr ational frequencies of longer carbon-sulfur clusters, when more experimental results become available. Dissociation Channels The relative stabilities and/or reactivities of carbon-sulfu r clusters can be evaluated from the associated dissociation ener gies from the generalized reaction: CnSm CiSj + Cn – iSm – j ( n = 1–29; m = 0–2; i = 0–2; j = 0–1) With a sulfur atom at the end of the carbon chain, the possible dissociation pathways for CnS and SCnS clusters are more complex than for Cn clusters. In matrix isolation experiments, it is well known that sm all species like C, S and CS are relatively

PAGE 61

43 mobile during the annealing process. For example, C7S can be formed from the reaction of C6 with CS.88 (A) (B) Figure 3-11. Dissociation energies of CnS (A) and SCnS (B) clusters computed at B3LYP/6-311G(d) level for the loss of S, CS, C2S, C, and C2 units.

PAGE 62

44 Figure 3-11 presents the computed dissociation energies of CnS and SCnS clusters for five dissociation channels. These channels can be divided into two groups: 1) losing C or CS, and 2) losing S, C2, or C2S fragments. Figure 3-11 sh ows that the dissociation energies for the first two cha nnels alternate strongly with n , with the alternation decreasing as the number of carbon atoms incr eases. While for the latter pathways, the dissociation energies are not mu ch different from each other, particularly in the medium to longer chains. Other disso ciation pathways have also been calculated, but their dissociation energies are alwa ys greater than the loss of CS. For example, carbon-sulfur clusters losing a C3 unit exhibit the same trend as the loss of a single carbon. Both CnS and SCnS species favor the loss of CS, probably because of the special stability of CS. A similar dissociation pattern was found for CnSe clusters observed in tandem time-of-flight mass spec trometry (TOF-MS) studies.91 It is concluded that the carbon chain is the origin of the odd-even alte rnation effect seen here. The added sulfur atom(s) increase the comple xity of the dissociation process by weakening the bond strength between its neighbor ing carbon atom and other carbon atoms. The dissociation energies of the CnS and SCnS clusters with odd n are always larger those of even n clusters. Odd clusters are thus much more stab le than even ones. This is consistent with the fact that odd clusters are easier to obtain experi mentally. Interestingly, odd n CnP clusters, isoelectronic with the CnS clusters, have been found experimentally and theoretically to possess higher struct ural stability than similar even n species.92-94

PAGE 63

45 The Linear CnS ( n = 2, 6) and CnS2 ( n = 7, 9, 11, 13, 15) Carbon-Sulfur Clusters Infrared absorption spectrometric study wa s carried out for carbon-sulfur clusters isolated in argon matrices at cryogenic temperatures. Density functional theory calculations, reported in the pr evious section, complement th e experimental observations and allow the identification of the C2S, C6S, C7S2, C9S2, C11S2, C13S2, and C15S2 clusters and ascribe certain vibrational frequencie s to them. Several new bands observed at 1662.6, 857.2, 1377.9, 1368.8, 1832.4 and 1796.7 cm 1 were assigned below to the linear C2S (1, C–C stretch), C2S (2, C–S stretch), C6S (4, C–C stretch), C7S2 (7, C–C stretch), C13S2 (10, C–C stretch) and C15S2 (12, C–C stretch) carbon-sulfur clusters, respectively. The fundamental modes are displayed in Figure 3-12. Other bands at 2017.8, 2056.7, 1938.2, 1962.1, 1678.7, 1903.2, 1832.4 and 1504.5 cm 1 were tentatively assigned to the linear C6S (2, C–C stretch), C7S2 (5, C–C stretch), C7S2 (6, C–C stretch), C9S2 (7, C–C stretch), C9S2 (8, C–C stretch), C11S2 (9, C–C stretch), C13S2 (10, C–C stretch) and C15S2 (13, C–C stretch) carbon-sulf ur clusters, respectively. Figure 3-12. Schematic of the atomic motion in the C2S, C6S, C7S2, C13S2, and C15S2 linear carbon-sulfur clusters. The observed (13C isotopomer proved) vibrational modes ar e displayed, only.

PAGE 64

46 CnS and CnS2 Vibrational Frequency Coincidence Plots Figure 3-13 (A-C) shows the infrared abso rption spectra for pure carbon clusters (reference spectrum a ) and for carbon and carbon-su lfur clusters (spectra b and c ) in the 1970 1200 cm 1 region. Spectrum c was recorded after annealing the matrix originally characterized by spectrum b . Since both a and c were recorded on a nnealed matrices (35 K), the new bands in c are most probably due to CnS and CnS2 clusters. Figure 3-14 (A-B) spectra are generated by electrical discharge through mixtures of C4H2/Ar (spectrum a ) and C4H2/CS2/Ar (spectrum b and c ). The new bands are marked by black dots and listed in Figure 3-15 (right column). This figure is an important aid in finding coincidences between predicted mode frequencies in CnS and CnS2 and any new bands. (Hereafter this figure is referred to as the “coincidence plot”.) Calculated (and scaled) harmonic frequencies ( cf. Figure 3-10) are marked by filled dots (for n odd) and by empty dots (for n even). The error bars represent 2 the variance. The standard deviation ( was calculated for all known experimental C C stretching frequencies for CnS and CnS2 clusters (Ar matrices, cf. , Table 3-12) and the corres ponding calculated (and scaled) frequencies ( cf. , Table 3-11), and found to equal cm 1. The dashed vertical lines in Figure 3-15 are the frequency positions of all new bands observed (They are also lis ted in the right hand column of the figure). The analysis below excludes the 2170 1970 cm 1 region, because as Figure 310 shows, reliable mode assignments are very difficult in this range due to the high density of vibrational modes from different clusters. In this region, many clusters have similar frequencies and attribution of a particular band to a specifi c cluster is problematic. Some modes (of a specific cluster) in this region may also mix since they are symmetry allowed, resulting in

PAGE 65

47 unpredictable intensity changes. In additi on, this region also includes many modes for pure carbon clusters and they fr equently overlap bands of CnS or CnS2 clusters, also making band assignments difficult. Figure 3-13. Part of infrared absorption spectra (displayed in A, B and C energy regions) of laser ablated graphite products trapped in solid Ar at 35 K (spectrum a ) and products of laser ablated mixture of gr aphite/sulfur (4/1 molecular ratio) isolated in Ar at 12 K (spectrum b ). Spectrum c was recorded after matrix (spectrum b ) annealing up to 35 K and recooling to 12 K. Newly observed bands due to carbon-sulfur clusters are marked by so lid circles with their proposed assignments. Other known bands at 1952.0 cm 1 (C6),83 1945.7 cm 1 (C11),95 1915.6 cm 1 (C10),85 1913.6 cm 1 (C7S),88 1894.3 cm 1 (C7),84 1856.6 cm 1 (C11),95 1844.2 cm 1 (C8cyc),96,97 1818.0 cm 1 (C12),98 1710.5 cm 1 (C8),99,100 1746.8 cm 1 (C4S),56 1695.0 (C6cyc),101,102 1688.8 cm 1 (C5S2),56 1446.5 cm 1 (C5),103,104 1275.2 cm 1 (CS),55 1316.8 cm 1 (C7S)88 and 1197.3 cm 1 (C6)83 are marked as well. The st arred bands are due to singly 13Csubstituted isotopomers of 12/13C6.83

PAGE 66

48 Figure 3-13. Continued

PAGE 67

49 Table 3-12. Comparison of expe rimental (Ar matrix) and calculated (B3LYP/6-311G(d)) most intense IR mode frequencies (cm 1) for linear CnS ( modes; n = 1–17) and SCnS ( u modes; n = 1–17) carbon-sulfur clus ters in their electronic ground state. Relative IR intensitie s for predicted and experimental frequencies are give n in parentheses. CnS Mode cal a / cm 1 exp / cm 1 SCnS Mode cal a / cm 1 exp / cm 1 1 (1+) 1; C–S 1284 (1.0) 1275.1 (1.0)c,d 1 (1g +) 2; C–S 1534 (1.0) 1528.2 (1.0)c,d 2 (3) 1; C–C 2; C–S 3 1647 (1.0)g 848 (0.33)g 254 (0.16)f,g 1662.6 (1.0)b 857.2 (0.28)b 2 (3g ) 3; C–S 1163 (1.0) 1179.7 (1.0)d 3 (1+) 1; C–C 2; C–C 3; C–S 2050 (1.0) 1497 (0.043) 729 (0.0085) 2047.6 (1.0)c,d 1533.2 (0.1)d 725.6 (0.009)d 3 (1g +) 3; C–C 4; C–S 2080 (1.0) 1023 (0.12) 2078.5 (1.0)c,d 1024.6 (0.18)c,d 4 (3) 1; C–C 2; C–C 4; C–S 2023 (0.062) 1729 (1.0) 603 (0.013) 1746.8 (1.0)c 4 (3g ) 4; C–C 5; C–S 1856 (1.0) 870 (0.19) 1872.1 (1.0)d 897.7 (0.117)d 843.7 (0.056)d 5 (1+) 1; C–C 3; C–C 4; C–S 2140 (1.0) 1581 (0.16) 1076 (0.020) 2124.5 (1.0)c 5 (1g +) 4; C–C 5; C–C 6; C–S 2115 (1.0) 1660 (0.28) 797 (0.047) 2104.7 (1.0)d 1687.9 (0.36)d 783.5 (0.04)d 6 (3) 1; C–C 2; C–C 4; C–C 2039 (0.025) 1998 (1.0) 1346 (0.17) 2017.8 (1.0) b,e 1377.9 (0.25)b 6 (3g ) 5; C–C 6; C–C 7; C–S 2034 (1.0) 1459 (0.32) 712 (0.080) 7 (1+) 2; C–C 3; C–C 5; C–C 2111 (1.0) 1921 (0.50) 1228 (0.049) 2088.1 (0.67)d,e 1913.6 (1.0)d 1256.1 (0.07)d 7 (1g +) 5; C–C 6; C–C 7; C–C 2085 (1.0) 1951 (0.94) 1341 (0.16) 2056.7 (0.26)b,e 1938.2 (1.0)b,e 1368.8 (0.15)b 8 (3) 1; C–C 4; C–C 6; C–C 2073 (1.0) 1762 (0.43) 1094 (0.064) 8 (3g ) 6; C–C 7; C–C 8; C–C 2074 (1.0) 1823 (0.53) 1213 (0.19) 9 (1+) 2; C–C 3; C–C 5; C–C 2099 (0.066) 2015 (1.0) 1658 (0.10) 9 (1g +) 7; C–C 8; C–C 9; C–C 1985 (1.0) 1696 (0.16) 1128 (0.049) 1962.1 (1.0)b,e 1678.7 (0.11)b,e 10 (3) 1; C–C 3; C–C 6; C–C 2080 (1.0) 1939 (0.91) 1510 (0.20) 10 (3g ) 7; C–C 8; C–C 9; C–C 2070 (1.0) 1972 (0.91) 1569 (0.25) 11 (1+) 4; C–C 5; C–C 7; C–C 1982 (1.0) 1869 (0.39) 1418 (0.043) 11 (1g +) 8; C–C 9; C–C 10; C–C 1961 (1.0) 1879 (0.97) 1479 (0.11) 1903.2 (1.0)b,e 12 (3) 2; C–C 3; C–C 6; C–C 2073 (0.10) 2020 (1.0) 1769 (0.22) 12 (3g ) 9; C–C 10; C–C 11; C–C 2013 (1.0) 1807 (0.27) 1382 (0.11) 13 (1+) 1; C–C 5; C–C 7; C–C 2146 (0.061) 1902 (1.0) 1689 (0.091) 13 (1g +) 8; C–C 10; C–C 11; C–C 2138 (0.062) 1867 (1.0) 1710 (0.19) 1832.4 (1.0)b,e 14 (3) 4; C–C 5; C–C 8; C–C 2011 (1.0) 1900 (0.55) 1583 (0.12) 14 (3g ) 10; C–C 11; C–C 12; C–C 1996 (1.0) 1924 (0.96) 1619 (0.20) 15 (1+) 1; C–C 6; C–C 7; C–C 2142 (0.076) 1877 (0.36) 1824 (1.0) 15 (1g +) 9; C–C 12; C–C 13; C–C 2130 (0.050) 1795 (1.0) 1549 (0.082) 1796.9 (1.0)b 1504.5 (0.11)b,e 16 (3) 1; C–C 5; C–C 7; C–C 2113 (0.13) 1956 (1.0) 1773 (0.16) 16 (3g ) 10; C–C 12; C–C 13; C–C 2112 (0.13) 1937 (1.0) 1797 (0.24) 17 (1+) 3; C–C 8; C–C 9; C–C 2106 (0.10) 1792 (1.0) 1704 (0.19) 17 (1g +) 10; C–C 12; C–C 13; C–C 2097 (0.20) 1765 (1.0) 1706 (0.99) a Scaled uniformly by 0.9556 for modes with primarily C C character and by 0.9866 factor for modes with primarily C S character. b This work. c Reference 56. d References 51, 53-55, 88. e Tentative assignment. f CCS bending vibration mode. g Harmonic frequencies for C2S in gas phase are 1708.2 ( 40), 862.5 (19), and 269.3 (13) cm 1 for 1, 2, and 3, respectively.89

PAGE 68

50 Figure 3-14. Part of infrared absorption spectra (displayed in A, and B energy regions) of diacetylene (C4H2) (0.25%)/ Ar electrical discha rge products trapped in solid Ar at 35 K and annealed up to 35 K (bottom spectrum a ) and products of C4H2 (0.25%)/CS2 (0.3%)/Ar electrical discharge is olated in Ar at 12 K (middle spectrum b ). Top spectrum c was recorded after matrix (spectrum b ) annealing up to 35 K and recooling to 12 K. Newly observed bands due to carbon-sulfur clusters are marked by solid circles with their proposed assignments. In addition to know bands observed in Figure 3-13 the following previously reported bands are observed at 1956.6 cm 1 (C6H),105 1936.5 cm 1 (C6 ),106,107 1831.8 cm 1 (C5 ),107,108 1827.9 (1826.3) cm 1 (C4H2 +),109 1734.8 cm 1 (C7 ),107,108 1721.8 cm 1 (C3 ),107,110 1699.9 cm 1 (C4 ),111 1686.7 cm 1 (C9 )107,108 and they are marked as well.

PAGE 69

51 Figure 3-15. Coincidences of calculated (B 3LYP/6-311G(d,p) scaled, solid and empty dotes) vibrational frequencies of funda mental modes with experimentally observed band frequencies (listed in the column in cm 1 and marked as vertical dashed lines) for CnS (A) and CnS2 (B) linear carbon-sulfur clusters. The standard deviation for previo usly reported experimental ( cf. , Table 312) and predicted band frequencies is 17 cm 1, thus the bar error of 2 is marked for each predicted mode frequencies.

PAGE 70

52 The C2S Carbon-Sulfur Cluster A recent gas phase fluorescence spectral study of C2S reported three fundamental bands at 1708.2, 862.5 and 269.3 cm 1.89 Another report on the laser-induced photolysis of matrix-isolated C3S2 or C3OS produced infrared absorption bands in argon matrices at 1666.6 and 862.7 cm 1 which were attributed to C2S.112 Both bands were supported by DFT theory calculations, but 13C-substituted species were not observed.112 Although C2S is formed in secondary r eactions after discharging CS2, its abundance is expected to be low. Furtherm ore, because the most intense calculated infrared vibration for C2S is only 43 km/mol (B3LYP/6 -311G(d)), the likelihood of observing C2S is small. However, Figure 3-15 shows that two of the observed bands, at 1650.6 and 1662.7 cm 1, could be candidates for one of the fundamental modes in C2S. But C7S and C9S clusters are also candidates. (No CnS2 clusters are expected to have bands in this region.) Howeve r, the previously-reported C7S bands at 1913.6 and 1256.1 cm 1 show poor correlation with the 1650.6 and 1662.7 cm 1 bands. Thus, C7S can be eliminated as a potential carrier of either of these bands. 13C-labeling is an important method of confirming band assignments for unknown, relatively small, carbon-containing clusters. 12/13C-substitution in C2S should yield four different isotopomers giving rise to four IR bands. Similar substitution in C9S should yield a large number of isotopomeric bands, making distinction between C2S and C9S straightforward. Figure 3-16 show s infrared spectra recorded on the trapped products of discharges through 12C2H2/12CS2/Ar (spectrum a ) or 13C2H2/12CS2/Ar (spectrum b ). The regions plotted are located where the 1 and 2 fundamental modes of C2S are expected

PAGE 71

53 Figure 3-16. Infrared spectrum of the products of 12C2H2/12CS2/Ar (spectrum a ) and the 13C2H2/12CS2/Ar (spectrum b ) discharges plotted in two energy regions of the 1 and 2 predicted fundamental modes frequencies. Spectrum c is recorded for matrix ( b spectrum) UV photolyzed, annealed up to 35 K and recooled to 12 K of matrix. The filled dot ma rked bands are assigned to the 12/13C2S isotopomers ( cf. , Table 3-13). The star mark ed bands are due to bending vibration in matrix isolated H2O, while empty circle, triangle and square marked bands are due to other carriers.

PAGE 72

54 Table 3-13. Comparison of observed (Ar ma trix, 35 K) and calculated (B3LYP/6311G(d) and B3LYP/cc-pVDZ) isotopomer frequencies (cm 1) of 1, and 2 modes for all-12C, singly-13C and all-13C substituted 12/13C2S linear carbonsulfur clusters, as well as 4, 7, 10, and 12 modes for all-12C and singly-13C substituted 12/13C6S, 12/13C7S2, 12/13C13S2, and 12/13C15S2 clusters, respectively. Mode Isotopomer exp 1 a 2 b exp1 exp2 C2S 12 12 32 1662.6 1662.6 1662.6 0.0 0.0 1 (C C) 13 12 32 1644.1 1644.0 1644.1 0.2 0.1 12 13 32 1620.6 1618.8 1618.8 1.8 1.8 13 13 32 1602.0 1599.4 1599.4 2.6 2.6 C2S 12 12 32 857.2 857.2 857.2 0.0 0.0 2 (C S) 13 12 32 840.7 840.1 840.1 0.6 0.6 12 13 32 852.4 853.2 853.2 0.8 0.8 13 13 32 836.5 837.0 836.9 0.5 0.4 C6S 12 12 12 12 12 12 32 1377.9 1377.9 1377.9 0.0 0.0 4 (C C) 13 12 12 12 12 12 32 1367.8 1366.6 1366.9 1.2 0.9 12 13 12 12 12 12 32 1376.1 1376.0 1376.3 0.1 0.2 12 12 13 12 12 12 32 1361.0 1362.3 1362.6 1.3 1.6 12 12 12 13 12 12 32 1374.2 1373.8 1374.1 0.4 0.1 12 12 12 12 13 12 32 1372.9 1372.5 1372.9 0.1 0.5 12 12 12 12 12 13 32 1359.9 1360.4 1360.8 0.5 0.9 C7S2 32 12 12 12 12 12 12 12 32 1368.8 1368.8 1368.8 0.0 0.0 7 (C C) 32 13 12 12 12 12 12 12 32 1355.2 1355.0 1355.2 0.2 0.0 32 12 13 12 12 12 12 12 32 1365.0 1364.9 1364.9 0.1 0.1 32 12 12 13 12 12 12 12 32 1367.8 1366.7 1366.7 1.1 1.1 32 12 12 12 13 12 12 12 32 1355.2 1355.5 1355.6 0.3 0.4 C13S2 32 12 12 12 12 12 12 12 32 1832.4 1832.4 0.0 10 (C C) 32 13 12 12 12 12 12 12 32 1832.0 1832.2 0.2 32 12 13 12 12 12 12 12 32 1826.8 1825.2 1.6 32 12 12 13 12 12 12 12 32 1824.8 1823.2 1.6 32 12 12 12 13 12 12 12 32 1828.0 1827.2 0.8 32 12 12 12 12 13 12 12 32 1829.5 1829.8 0.3 32 12 12 12 12 12 13 12 32 1819.8 1819.8 0.0 32 12 12 12 12 12 12 13 32 1816.0 1816.4 0.4 C15S2 32 12 12 12 12 12 12 12 12 32 1796.9 1796.9 0.0 12 (C C) 32 13 12 12 12 12 12 12 12 32 1795.3 1795.4 0.1 32 12 13 12 12 12 12 12 12 32 1787.2 1790.1 2.9 32 12 12 13 12 12 12 12 12 32 1794.5 1794.8 0.3 32 12 12 12 13 12 12 12 12 32 1795.3 1796.1 0.8 32 12 12 12 12 13 12 12 12 32 1792.3 1791.9 0.4 32 12 12 12 12 12 13 12 12 32 1785.4 1783.4 2.0 32 12 12 12 12 12 12 13 12 32 1789.3 1790.9 1.6 32 12 12 12 12 12 12 12 13 32 1793.3 1794.3 1.0 a Calculated at B3LYP/6-311G (d) and scaled by 0.9651, 0.9976, 0.9745, 0.9705, 0.9381, and 0.9566 factor for 1 and 2 modes of C2S, 4 mode of C6S, 7 mode of C7S2, 10 mode of C13S2, 12 mode of C15S2, respectively. b Calculated at B3LYP/cc-PVDZ and Scaled by 0.9664, 0.9965, 0.9779, and 0.9753 factor for 1 and 2 modes of C2S, 4 mode of C6S, 7 mode of C7S2, respectively.

PAGE 73

55 ( cf ., Table 3-12). Spectrum c is recorded on a matrix originally characterized by spectrum b but subjected to UV photolysis, followed by annealing (to 35 K) and re-cooling (to 12 K). Spectrum c shows bands (dotted) due to a comm on carrier. Two of the dotted bands, at 1662.6 cm 1 and 857.2 cm 1, arise from an isomer containing only 12C atoms. The presence of four 12C/13C isotopomer bands ( cf ., Figure 3-16) suggests strongly that the carrier of the 1662.6 cm 1 band is C2S, and not C9S. To support this conclusion, frequency shifts for the four 12/13C2S isotopomers were calculated (at B3LYP/6-311G(d) and B3LY P/cc-pVDZ levels) by adopting force constants from the 12C2S calculations ( cf. , Table 3-8). These results are compared in Table 3-13 to the observed frequencies ( cf. , Figure 3-16). The predicted isotopomer frequencies differ from the experi mental ones by a maximum of 2.6 cm 1. This is acceptable agreement based on previous 13C isotopomer frequency studies of C3S, C4S, C5S and C7S.56,88 For the latter clusters, maximum isotopomer frequency deviations of ca . 3 cm 1 were reported. In conclusion, isotopic substitution confirms the 1662.6 and 857.2 cm 1 band assignments to the 1 (C C stretch) and 2 (C S stretch) modes in linear C2S. Some isotopomer bands attributed to the same species change relative intensities upon matrix photolysis/annealing ( cf. , Figure 3-16A and 3-16B, spectrum c ). For example, during annealing the 1620.6 and 852.4 cm 1 bands (both assigned to the 12 13 32 isotopomer) gain in inte nsity, while the 1644.2 and 840.7 cm 1 bands (both assigned to the 13 12 32 isotopomer) decrease. This inte resting effect can be explained by the formation and depletion of these isot opomers as a result of exothermic reactions initiated by photolysis or variable diffusion of 12/13C, S, 12/13CS species upon annealing. Not only do the laser ablation and discharg e experiments probably produce different

PAGE 74

56 concentrations of the initial reactants ( cf. , Figure 3-13, Figure 314, and Figure 3-16), but different reaction pathways are e xpected to form and destroy C2S. Both of these can lead to varying yields of formation and destru ction of the different isotopomers and thus different band intensities. The C6S and C7S2 Carbon-Sulfur Clusters The two most intense vibrational modes of C6S are predicted to fall at 1346 cm 1 ( , and 1998 cm 1 ( (mode designation and rela tive computed intensities in parentheses, cf. , Table 3-12). From the coinciden ce plot (Figure 3-15A), the 1346 cm 1 4 mode might be ascribed to one of several experimental bands, either c (1316.8), d (1332.3), e (1368.8) or f (1377.9 cm 1). To determine which of these is the correct choice, use was made of 12C/13C isotopic substitution. Rather than try to observe all 62 isotopomeric bands for C6S, the isotopic ratio [12C] / [13C] was chosen to be 6, so that only the all 12C and singly 13C-substituted isotopomer bands w ould be expected to appear. Figure 3-17 shows the spectra of laser-ablated mixtures of 12C/S (spectrum a ) and 12C/13C/S (spectrum b ) in the 1350 1380 cm 1 region. Isotopomeric bands built on the weak 1377.9 cm 1 band (marked by triangles) and the cm 1 band (marked by dots) can be seen in spectrum a . By testing the isotopomeric band patterns expected for C6S, C7S2, and other possible clusters, the 1377.9 cm 1 band is here assigned to the 4 mode of C6S. (The cm 1 band is discussed later). The calculated (B3LYP/6-311G(d) and B3LYP/cc-pVTZ) harmonic isotopomer frequenc ies for this mode are given in Table 313 where they are compared to the observed frequencies. The maximum difference between prediction and observation is 1.6 cm 1, making this assignment quite reliable. The relative strength of the 1355.2 cm 1 band is consistent with the fact that the

PAGE 75

57 calculation predicts two overlapping bands at this frequency. The 32 13 12 12 12 12 12 12 32 and 32 12 12 12 13 12 12 12 32 isotopomers have predicted bands at 1355.0 and 1355.5 cm 1, respectively ( cf. , Table 3-13). Also, a small intensity contribution to the 1355.2 cm 1 band profile comes from an unknown 12CnSm cluster, whose outline is apparent in the lower trace of Figure 3-17. Figure 3-17. Infrared spectra of the products of laser-ablated 12C2H2/12CS2/Ar (spectrum a ) and the 13C2H2/12CS2/Ar (spectrum b ) samples recorded after matrix annealing up to 35 K and re-cooling to 12 K. The triangle marked bands are assigned to the 12C-all and 13C-singly substituted isotopomers of linear 12/13C6S, while the dot marked bands are assigned to the 12C-all and 13C-singly substituted isotopomers of linear 12/13C7S2 ( cf. , Table 3-13). For the C7S2 cluster, the three strongest vibrat ional bands are predicted to lie at 2085 ( 5, 1.0), 1951 ( 6, 0.94), and 1341 ( cm 1 ( cf. , Table 3-12). Figure 3-15B shows that the experimental z (1938.2), (1939.8), (1960.5), and (1962.1) cm 1 bands are possible candidates for the 6 mode assignment and the c (1316.8), d (1332.3), and e (1368.8) cm 1 bands candidates for the 7 assignment. Before any two bands can be assigned to the 6 and 7 modes, a sufficient condition must be fulfilled, i.e., that the

PAGE 76

58 intensities of both bands correlate well under various different expe rimental conditions. After reviewing many spectra, taken with di fferent ablation laser fluxes, discharge voltages, and sample-to-matrix gas ratios, only the 1938.2 and 1368.8 cm 1 band pair satisfies the correlation requirement. Thus, it is likely that these bands should be assigned to the 6 and 7 modes of the C7S2 cluster, respectively. Th e use of isotopic labeling confirms this conclusion. The 13C-labeled spectrum built on the 1938.2 cm 1 band is difficult to analyze since the region for singly 13C-substituted isotopomer bands is overlapped by at least eight different carbon and carbon-sulfur clusters ( cf. , Figure 3-13). However, the region predicted for the 7 isotopomer bands is relatively clear, with only two ba nds appearing at 1377.9 and 1368.8 cm 1. As detailed above, the former band has been ascribed to the 4 mode of C6S. The computed isotopomeric band pattern for C7S2 is given in Table 3-13 and there compared to the observed bands from Figure 3-17. The maximum difference between calculated and experi mental bands is just 1.1 cm 1. Thus the 1368.8 cm 1 band is here assigned to the 7 mode of C7S2. It has become clear from comparison of the C C stretching mode isotopomer frequency shifts for C6S, C7S2, and C7S88 that each pattern is different, and if the molecule giving rise to the patter n is chosen correctly, only small exp theo frequency differences are found. Thus, 13C-labeling in infrared spectra is a sensitive molecular “fingerprint” tool, enabling clear differentia tion between different molecular clusters. The strongest band predicted for C6S is the mode at 1998 cm 1, while for C7S2, it is the mode at 2034 cm 1 ( cf. , Table 3-12). Searches in the region around 2000 cm 1 show that the 2017.8 and 2056.7 cm 1 bands are possible candida tes. Intensity correlation

PAGE 77

59 with already-assigned bands showed that the 2017.8 cm 1 band tracks the 1377.9 cm 1 (C6S) band, while the 2056.7 cm 1 closely follows the 1938.2 and 1368.8 cm 1 (C7S2) bands. Both sets were compared over many matrix spectra recorded under different experimental conditions. From this comparison, the 2017.8 and 2056.7 cm 1 bands are here tentatively assigned to the and modes of C6S and C7S2 linear clusters, respectively. The observed relative intensity of the high energy C C stretching mode ( 2000 cm 1) is much lower than predicted by DF T calculations, an observation also found previously for C9,82,100 C9Si,87 and C7S.88 However, the relative experimental intensities for other modes are generally consis tent with the calculated values ( cf. , Table 3-12). The CnS2 ( n > 7) Carbon-Sulfur Clusters A number of prominent bands in Figure 3-13 remain unassigned. To determine their origin, reference was again made to th e coincidence plot of Figure 3-15. Despite the narrowing of choices that this plot provides, the number of possibl e cluster choices for each observed band is still too large to reliably choose one. Insight based on the probability of formation of certain cluste rs over others was therefore sought. Up to the present, only those carbon-sulfur CnSm clusters with n < 7 and m =1 and 2 have been experimentally identified. While th ere are gaps in the knowledge even for this set of clusters, some vibrational frequencie s for most of these species are now known ( cf. , Table 3-12). Therefore, the atten tion was turned to clusters with n 8. As found previously, the formation of Cn =oddSm =1,2 clusters is energetically preferred over Cn =evenSm =1,2, for reactions involving the addition of C or CS to a pure carbon chain. For instance, the C8S + C and C8 + CS reactions forming C9S are more exothermic by ca. 1 eV than the C7S + C and C7 + CS reactions forming C8S ( cf. , Figure 3-11A). Larger

PAGE 78

60 reactants, such as C2S and C2, might also form C9S or C8S, but are expected to have smaller mobilities in solid Ar and therefore contribute less to the total yield. Moreover, the C9S + C and C9S + CS reactions which deplete C9S are less exothermic by ca. 1 eV than the analogous reactions which deplete C8S. Thus, C9S is expected to form preferentially over C8S. A similar argument can be made for the preferential formation of C9S2 over C8S2 ( cf. , Figure 3-11A and Figure 3-11B). Fu rthermore, the predicted integral infrared intensities for Cn =oddSm =1,2 are substantially larger than for Cn =evenSm =1,2 clusters (see Table 3-9 and Table 3-10), making observation via IR also more probable for odd clusters. In fact, the stro ngest reported band for C4S2 (in Ar) at 1872.1 cm 1 ( cf. Table 312) can not be seen in Figure 3-13 or Figur e 3-14. Furthermore, no experimental band could be found in Figure 3-13 that could be associated with the relatively strong bands predicted for C6S2. However, bands due to C2S (weak), C4S (relatively strong), and C6S (weak) are observed in Figure 3-13 and 3-14. Overall, the observed Cn =oddSm =1,2 cluster bands in Figure 3-14 are stronger than the Cn =evenSm =1,2 cluster bands appearing in the same figure, as predicted. At higher relative con centration of sulfur vs. carbon, laser ablation generates CnS2 clusters in higher yields. In the IR spectru m obtained with S/C = 1 (before annealing, not displayed), the most prominent bands are 2078.5 (C3S2), 2104.7 (C5S2), 1962.1 (assigned here to C9S2, vide infra ), 1938.2 (assigned in this work to C7S2), and the relatively strong 1903.2 cm 1 band (assigned here to C11S2, vide infra ). For example, in this spectrum the intensity ratio of the 1938.2 cm 1 (C7S2) band to the 1913.6 cm 1 (C7S) band is ca. 9, while in Figure 3-13A (where S/C = 0.25) the ratio is only ca. 0.25, using similar ablation laser settings. Thus, this strong variation in IR band intensities with S/C ratios is a helpful

PAGE 79

61 method to differentiate CnS2 from CnS (same n ) in Figure 3-13 and Figure 3-14. Because Cn =oddS2 clusters are preferred energetically over Cn =evenS2 clusters, and, because the predicted integral infrared intensities for the former clusters are substantially larger than for the latter, the most likely candidates for the as-yet unassigned strong bands in Figure 3-13 and Figure 3-14 are Cn =oddS2 clusters. To assign the remaining dotted bands in Figure 3-13 and Figure 3-14 to larger carbon-su lfur clusters, the frequency coincidence plot of Figure 3-15 is used and, where possible, 13C-labeling. The C9S2 carbon-sulfur cluster Focusing on C9S2, the coincidence plot of Figur e 3-15B shows that the strong (1962.1 cm 1) and weak (1960.5 cm 1) experimental bands are possible candidates for the strongest 7 vibrational mode of C9S2, predicted at 1985 cm 1, while the m (1678.7 cm 1) and n (1691.7 cm 1) bands are in the proximity of the predicted 1696 cm 1 ( 8) mode. Only the 1962.1 cm 1 and 1678.7 cm 1 bands correlate well in numerous spectra taken under many different experimental conditions. Another factor favoring the assignment of these bands to C9S2 is their intensity ratio. The predicted integral intensity ratio [I( 8) / I( 7)] of C9S2 is 0.16 ( i.e. , 3205 km/mol / 19808 km/mol), whereas the experimental ratio is 0.11. Fo r these reasons, the 1962.1 and 1678.7 cm 1 bands are tentatively assigned to the 7 and 8 C C stretching modes of linear C9S2. The C11S2 carbon-sulfur cluster Laser ablation spectra (with C / S = 1) suggest that the 1903.2 cm 1 band is due to a CnS2 cluster, with n = odd, as discussed above. From the coincidence plot, only the C11S2 and C13S2 clusters are good candidate s for this band assignment. Both have very strong bands predicted at this frequency and bot h have an odd number of carbons. Since 13C-

PAGE 80

62 labeling leads to the assignment of the 1832.4 cm 1 band to the C13S2 cluster ( vide infra ), the 1903.2 cm 1 band is thus tentatively assigned to the 9 C C stretch fundamental mode in C11S2. The C13S2 carbon-sulfur cluster The experimental u (1832.4 cm 1), v ( 1867.2 cm 1) and w (1903.2 cm 1) bands (Figure 3-15B) are possible choices for the predicted very strong 1953 cm 1 band of C13S2. 13C-labeling makes the correct choice simple. Single 13C-substitution yields a frequency pattern that is in agreement with only one experimental band, the 1832.4 cm 1. The spectrum which shows 8 isotopomeric ba nds is presented in Figure 3-18 and the band assignments given in Table 3-13. The ve ry small differences between predicted and observed isotopomer frequencies (1.6 cm 1 maximum) indicate that the assignment of the 1832.4 cm 1 band to the 10 C C stretching fundamental in C13S2 is quite definite. The C15S2 carbon-sulfur cluster The strongest IR vibrational mode (the 12) in the C15S2 cluster, predicted via B3LYP/6-311G(d) to lie at 1795 cm 1 (after common scaling), possesses a huge computed intensity of 55,330 km/mol. In Figur e 3-13 and Figure 3-14 the strongest band observed in this region is at 1796.9 cm 1. 13C isotopic labeling yields the spectrum displayed in Figure 3-18. The observed and pr edicted isotopomer bands are collected in Table 3-13. All 12/13C isotopomer bands are assignable to predicted bands with reasonable frequency differences (2.5 cm 1 maximum difference). The re latively large intensity of the observed 1795.3 cm 1 band is nicely explained by two overlapping isotopomeric bands. Two singly 13C-substituted isotopomers (32 13 12 12 12 Â… 12 32 and 32 12 12 12 13 Â… 12 32) give rise to two bands at 1795.4 and 1796.1 cm 1, respectively, which are practically overlapped. In addition, the 1787.2 cm 1 isotopomer

PAGE 81

63 band is partially overlapped by an unknown 12CnS or 12CnS2 cluster, which shows a small absorption in the lower 12C / S / Ar spectrum. Figure 3-18. Infrared absorption spectr a of the laser ablation products of 12C/32S/Ar (spectrum a ) and the 12/13C/12CS2/Ar (spectrum b ) recorded after matrix annealing (to 35 K) and re-cooling (to 12 K). The all 12C and singly 13Csubstituted isotopomeric bands of linear C13S2 are marked by triangles and of C15S2 by solid circles, with assignments in Table 3-13. The bands marked by stars are assigned to linear 12/13C12 carbon clusters.98 The 1796.9 cm 1 band correlates well intensity-w ise with the experimental 1504.5 cm 1 band. Furthermore, it falls well within the limits of the 13 mode of C15S2 in the coincidence plot. Also, the predicted intens ity ratio of 0.08 for these two bands is in concert with the observed ratio of 0.11. Thus, the 1796.9 cm 1 and 1504.5 cm 1 bands are assigned to the 12 and 13 C C stretching vibrations, re spectively, of the common carrier, C15S2.

PAGE 82

64 Summary The following conclusions were reached: (1) 12/13C7S carbon-sulfur clusters have been pr oduced by laser ablation of mixtures of graphite and sulfur, trappe d in solid Ar at 12 K, and studied with FT-IR absorption spectroscopic methods. Theoretical calculations at di fferent levels of theory show that C7S is linear in Ar matrices. Newly obser ved IR absorption bands at 2088.1, 1913.6 and 1256.1 cm 1 have been assigned to 2, 3 and 5 C C stretching modes of linear C7S, respectively. The mechanism of formation of linear C7S is most likely a simple addition reaction of C, CS and S species to already-formed carbon-sulfur CnS and pure carbon Cn clusters in the matrix. (2) The equilibrium geometries of two se ries of linear carbon-sulfur clusters, CnSm ( n = 1–29, m = 1–2), have been calculated at the B3LYP/6-311G(d) level. Their energies, structures, and vibrational fundamentals exhi bit a strong odd/even parity. To verify the accuracy of the computational results, the C2S, C6S, and C7S2 clusters were recalculated at higher theoretical levels, such as B3LYP/aug-cc-pVTZ, CCSD(T)/cc-pVDZ, etc . Unfortunately, results at these levels are not helpful for the present purposes. Values of the rotational constant Be for the CnS ( n < 9) clusters calculated at the B3LYP/6-311G(d) level differ from experimental ones by an average of only 0.5%. It is reasonable to conclude that this method could be used at a similar confidence level for the prediction of Be for longer carbon-sulfur cluste rs. Vibrational frequencies fo r the carbon-sulfur clusters are predicted within an acceptable error ( = 17 cm 1) when compared to known IR band frequencies. Finally, reaction of the CS radical with already-formed Cn or CnS clusters is

PAGE 83

65 predicted to be the dominant pathway in th e secondary processes occurring during matrix annealing and / or photolysis experiments. (3) Carbon-sulfur clusters of varying co mpositions and sizes have been produced by the laser ablation of pressed carbon/sulfu r pellets and by the ga seous discharge of mixtures of acetylene / carbon disulfide / ar gon or diacetylene / carbon disulfide / argon. In both approaches the products were trapped in argon matrices and studied via Fourier transform infrared absorption spectroscopy. Dens ity functional theory calculations were performed to aid in the assignment of the sp ectra to specific clusters. A “coincidence plot”, in which all new bands observed in th is study are plotted ag ainst the vibrational frequencies predicted for the CnSm clusters (with n < 17, m = 1, 2), was used to narrow the choices for possible assignments. Isotopic 13C substitution confirmed a number of these assignments. (4) New bands were observed at 1662.6, 857.2, 1377.9, 1368.8, 1832.4 and 1796.9 cm 1 and assigned to the linear carbon-sulfur clusters C2S (1, C–C stretch), C2S (2, C–S stretch), C6S (4, C–C stretch), C7S2 (7, C–C stretch), C13S2 (10, C–C stretch) and C15S2 (12, C–C stretch), respectively. These assi gnments were guided by B3LYP/6-311G(d) harmonic frequency predictions and supported by 13C-isotopomer band frequency shifts. (5) Assignments of additional bands at 2017.8 cm 1for C6S (2, C–C stretch); 2056.7 cm 1 for C7S2 (5, C–C stretch); 1938.2 cm 1 for C7S2 (6, C–C stretch); 1962.1 cm 1 for C9S2 (7, C–C stretch); 1678.7 cm 1 for C9S2 (8, C–C stretch); 1903.2 cm 1 for C11S2 (9, C–C stretch); and 1504.5 cm 1 for C15S2 (13, C–C stretch) are proposed, based on intensity correlations with bands assigned by 13C-labeling. These correlations were verified using three different, independent experiments: 1) la ser ablation with variable S /

PAGE 84

66 C ratios, 2) C2H2/CS2/Ar discharge, and 3) C4H2 / CS2 / Ar discharge, both with and without matrix annealing. Since no 13C-labeled spectra are presented for these latter bands, their assignments should be viewed as tentative. (6) Temperature-controlled diffusion reacti ons take place in solid Ar during matrix annealing. The addition of C, S and CS species to linear Cn and CnS clusters is proposed as the main mechanism of formati on of the longer lin ear chains of Cn +1S and Cn (or n +1)S2. In such a mechanism, the formation of any large cyclic CnS or CnS2 structures is unlikely. The Cn =oddS2 X (1g +) closed shell systems are observed here at preferential yields, as predicted. The C11S2, C13S2, C15S2 linear clusters reported in this work have predicted overall lengths (B3LYP/6-311G(d)) of 1.589 nm, 1.844 nm, 2.099 nm, respectively. They could be classified as nanowires based on th ese lengths and their ex pected high stability. These are attractive systems for further study b ecause their potential unique properties as related to material science and because of their possible astrophysical importance.

PAGE 85

67 CHAPTER 4 INFRARED ABSORPTION SPECTROSCOPY OF THE XENON CARBON CLUSTERS Introduction Because their outermost electronic sh ells are filled, the noble gases had traditionally been portrayed as chemically in ert. But this view was shattered in 1962 when Bartlett reported the first chemical species involving a nobl e gas atom, Xe·[PtF6].113 Since then other rare gas-containing compounds have been reported. Most recent studies have focused on the reaction of noble gases w ith strongly electronegative atoms, such as F and O.113-116 Very recently, rare gas atoms in cr yogenic matrices have been shown to insert in H–X bonds to form hydrides like HRgX , where X is O, H, or halogen or pseudohalogen (NCO, OH, CN, etc .) and Rg is Ar, Kr or Xe.117-138 By photoexciting Xe with a UV laser via two-photon absorption, G. Maier and coworkers detected the formation of C2Xe in a solid xenon matrix.139 Recently, Xeinserted hydrocarbons (HXeCCH, HXeCC, etc. ) have also been generated with 193 nm radiation or with fast elec tron dissociation in solid Xe.128,140 Xenon has also been shown to form weakly bound complexes in solid rare gases with other compounds, such as HCl, CN, dimethyl ether, benzene, oxirane, and difluorovinylidene.141-145 Theoretical studies on xenon-containing co mpounds have been reported over the past decade. H-Xe-R (R = C6H5 or other hydrocarbons) was pr edicted to be a stable species.146 Calculations on other Xe-inserted molecules, such as HXeCN and HXeNC, support the experimental observations well.136 Following the observation of the C2Xe

PAGE 86

68 species, high level ab initio calculations for C2Xe were reported and showed that C2Xe has a linear structure with a rather flat potential surface.147 This flatness causes an uncertainty in the length of the C–Xe bond. Considerable research has been devoted to carbon clusters since the discovery of buckminsterfullerene.148 Vibrational frequencies of carbon clusters have been investigated in solid matri ces for a number of years.78,149 Therefore, as an extension of this work, it is of interest to inves tigate the vibrational properties of the CnXe ( n 2) species. In this chapter, the de tection and characterization of CnXe ( n = 2, 3, 5, 7, 9) compounds will be discussed and possible mechanisms for their formation outlined. Theoretical Results The ground electronic state of the linear C2 and C3 is 1g +. For linear C2Xe and C3Xe, it is 1 . Bent C3Xe has a 1A ground state. Figure 4-1 shows the ground state equilibrium geometries for C2Xe and C3Xe, optimized at the MP2/LJ18 (xenon)/6311++G (2d,2p) (carbon) level of theory. Ta ble 4-1 lists the ca lculated rotational constants, dipole moments, harmonic vibra tional frequencies, energies, zero point energies, and binding energies. Th e C–C and C–Xe bond lengths in C2Xe are both shorter than in C3Xe implying stronger binding in C2Xe. Other characteristics also indicate the greater stability of C2Xe. The binding energy of C2Xe is four times that of C3Xe ( cf. , Table 4-1). The di pole moment in C3Xe is much smaller than in C2Xe. Although bent C3Xe has a larger dipole moment, a shorter C–Xe bond length, and is a slightly more stable than linear C3Xe, their zero point-corrected binding energies differ by only 0.13 kJ/mol.

PAGE 87

69 At the MP2/6-311++G(2d, 2p) level, th e C–C bond length is lengthened from 1.259 Å in C2 to 1.265 Å in C2Xe. The corresponding C–C stretching frequency is 61.9 cm 1 smaller in C2Xe, which is in concert with experiment.139 Therefore, it can be concluded that the C–C bond strength is weaken ed due to interaction with Xe. Figure 4-1. Ground state equilibr ium geometries for the C2Xe and C3Xe species optimized at the MP2/LJ18 (xenon)/6 -311++G (2d,2p) (carbon) level. However, in C3Xe, the interaction of C3 and Xe is weaker than C2 and Xe in C2Xe. The C–C bond lengths in C3 and C3Xe are all close to 1.299 Å. The most intense C–C stretching mode of C3 is predicted at 2121.4 cm 1, only ~ 1 cm 1 lower than in linear C3Xe and ~ 2 cm 1 higher than in bent C3Xe. The calculated electronic charge distribution on Xe also reflect s the weaker binding in C3Xe. In C2Xe, xenon possesses a +0.105 e charge, but in linear C3Xe, it is only +0.014 e and in bent C3Xe, only +0.018 e . Figure 4-2A and Figure 4-2B show the co mputed potential energy of the linear C3Xe complex (MP2 level) as a function of the C–Xe bond length, R, and the C–C–Xe bond angle, respectively. Figure 4-2C represents the potential surface of linear C3Xe

PAGE 88

70 Table 4-1. Rotational constants Be (GHz), dipole moments µe (Debyes), harmonic vibrational frequencies (unscaled, cm 1), integral IR intensities (km/mol, in parentheses), energies E (Hartrees), ze ro point energies ZPE (kcal/mol), and binding energies Eb (kJ/mol), calculated for CnXe ( n = 2, 3) at MP2 level with LJ18-Xe/6-311++G (2d,2p)-C basis set. CnXe Parameters B 1.8413434 µe 2.1801 : 123.5 (1), 1807.1 (430) d: 245.6 (110) E 201.9940455 ZPE 3.46202 C2Xe (linear) X1 Eb a 12.58 B 0.6132243 µ 0.0255 : 37. 5 (0), 1188.2 (0), 2122.2 (548) d: 12.3 (0), 176.4 (26) E 240.0091962 ZPE 5.32564 C3Xe (linear) X1 Eb a 3.21 B 16.4473847 0.9654028 0.9118788 µ 0.3476 a': 20.9 (0), 41.0 (0), 180.8 (26), 1191.4 (0.0), 2119.3 (505) a": 202.85 (24.1) E 240.0093147 ZPE 5.36973 C3Xe (bent) X1A' Eb a 3.34 a Binding energy of CnXe is calculated by Eb(CnXe) = E(CnXe) + ZPE(CnXe) E(Cn) ZPE(Cn) E(Xe) d Doubly degenerate bending ( ) modes as a function of R and . In plot Figure 4-2A, with the C–C–Xe bond angle ( fixed at 180°, the relative energy increa ses sharply when Xe approach es closer than 3.0 Å to the C2 fragment while it increases slowly a nd monotonically as the C–Xe bond elongates. C3Xe has a rather flat potential surface. For example, when the C–Xe bond length changes from 3.6 Å to 4.6 Å, the potential ener gy difference is less than 1 kJ/mol. As shown in Figure 4-2B, when the bond angle varies (with R fixed), three local minima

PAGE 89

71 are found, but their energy differences are ve ry small. Tunneling between these three states is possible. The relative energies at 2 = 95 and 3 = 265 are slightly lower than at equilibrium ( 1 = 180 ), which is due to the similarity of these two states to bent C3Xe. In fact, optimization of C3Xe at the 2 and 3 angles results in lo calization to bent C3Xe. Therefore, C3Xe with an angle of 2 (or 3) is probably the trans ition state between linear and bent C3Xe. Figure 4-2C shows clearly the flat potential surface of C3Xe, and points out the difficulty in determining the structural parameters of C3Xe. (A) Figure 4-2. Plot of the ground stat e potential surface of linear C3Xe from MP2/LJ18 (xenon)/6-311++G (2d,2p) (carbon) calcu lations with varying C–Xe bond length, R, (plot A) or C–C–Xe bond angle, plot B). The equilibrium bond length R0 and the local minima at 1, 2, and 3 are marked. Plot C shows the ground state potential surface of linear C3Xe (MP2/LJ18 (xenon)/6-311++G (2d,2p) (carbon)) as a function of th e C–Xe bond length (R) and C–C–Xe bond angle ( . The isoenergetic lines are marked in kJ/mol.

PAGE 90

72 (B) (C) Figure 4-2. Continued

PAGE 91

73 (A) (B) Figure 4-3. Plot of the ground stat e potential surface of bent C3Xe from MP2/LJ18 (xenon)/6-311++G (2d,2p) (c arbon) calculations as a function of C–Xe bond length, R, (plot A) or C–C–Xe bond angle, plot B). The equilibrium bond length R0 and the local minima at 1, 2, and 3 are marked.

PAGE 92

74 Figure 4-3 presents the potential surface of bent C3Xe as a function of the C–Xe bond length R ( A ) or C–C–Xe bond angle B ). Both plots show a similar trend to that seen in Figure 4-2. Although th e potential surface for bent C3Xe as a function of R and is not included in this wo rk, its contours should be si milar to those of linear C3Xe. Calculations at other levels of theory , such as density functional B3LYP/LJ18 (xenon)/6-311++G (2d,2p) (carbon) were also carried out, but th e results failed to predict the experimental observations on C2Xe. Compared to the more precise calculation of C2Xe,147 the calculation successfully predicts the vibrational frequenc y shift and binding energy rather well. Experimental Results Xenon-containing carbon clusters CnXe ( n = 3, 5, 7, 9) have been generated using either laser ablation or electrical discharge. Laser ablation favors the production of larger sized complexes such as C3Xe, C5Xe, etc. C2Xe may be generated via laser ablation, but its signal is typically much smaller compared to other clusters. On the other hand, small xenon-carbon compounds are preferentially pr oduced in the electrical discharge. C2Xe exhibits a large IR signal, while the C3Xe signal is moderate. It was, however, difficult to observe C7Xe or larger carbon-xenon clusters in the discharge. Furthermore, matrices formed during laser ablation are cloudier than matrices formed during discharges, especially during isotopic 12/13CnXe experiments. As a result, isotopomeric frequencies for 12/13C2Xe were only investigated in the discharge experiments. Figure 4-4 shows the newly-observe d isotopomer bands due to the 12C–13C stretching vibration in 12/13C2Xe clusters generated by an el ectrical discharge. Analogous isotopomer bands in 12/13C3Xe clusters from laser ablation are shown in Figure 4-5.

PAGE 93

75 Comparison of observed and calculated is otopomer frequencies for both types of 12/13CnXe complexes ( n = 2, 3) are tabulated in Table 42. Finally, Figure 4-6 presents the infrared absorption spectra of Cn and CnXe ( n = 3, 5, 7 and 9) species produced by laser ablation. The experimental frequency shifts for Cn clusters compared to CnXe ( n = 3, 5, 7, 9) are plotted as a function of th e number of carbon atoms in Figure 4-7. Table 4-2. Comparison of observed (Ar matr ix, 20K) and calculated (MP2//LJ18-Xe/6311++G (2d,2p)-C) isotopomer frequencies (cm–1) for all-12C and singly-13C substituted 12/13CnXe ( n = 2, 3). Isotopomer exp cal cal * exp– cal * 12-12-132 1774.2 1807.1 1774.2 0.0 13 -12-132 1739.3 1771.3 1739.1 0.2 1213 -132 1740.8 1772.7 1740.4 0.4 Linear C2X e 13 13 -132 1705.3 1736.2 1704.6 0.7 12-12-12-132 ~ 2033.3 2122.2 2033.3 ~ 0.0 13 -12-12-132 ~ 2020.7 2108.8 2020.5 ~ 0.2 1213 -12-132 ~ 1981.9 2067.1 1980.5 ~ 1.4 Linear C3Xe 12-1213 -132 ~ 2020.7 2108.8 2020.4 ~ 0.3 12-12-12-132 ~ 2033.3 2119.3 2033.3 ~ 0.0 13 -12-12-132 ~ 2020.7 2105.9 2020.4 ~ 0.3 1213 -12-132 ~ 1981.9 2064.3 1980.5 ~ 1.4 Bent C3Xe 12-1213 -132 ~ 2020.7 2105.9 2020.5 ~ 0.2 *Scaling factors for linear C2Xe, linear C3Xe, and bent C3Xe are 0.9818, 0.9581, and 0.9594, respectively. Infrared Absorption of 12/13C2Xe in Argon Matrices The infrared absorption spectra of 12C2Xe and 13C2Xe in xenon matrices measured previously showed peaks at 1767.0 cm 1 and 1698.6 cm 1, respectively.139 Bands from the singly-13C-substituted species have not yet been reported, so this assignment has not been fully confirmed. An elec trical discharge of acetylen e is known to produce a large concentration of C2 clusters.150 Therefore, if a small amount of Xe gas were seeded in argon gas along with 12C2H2 and 13C2H2, 12/13C2Xe might be produced as a discharge product of this mixture. Figure 4-4 shows the re sults of such an attempt. Four new bands

PAGE 94

76 are generated at 1774.2, 1740.8, 1739.3, and 1705.3 cm 1. An Ar-to-Xe matrix shift of ca. 7 cm 1 is found for the 1774.2 cm 1 (12C2Xe) and 1705.3 cm 1 (13C2Xe) bands. Because all four bands increase in the same ratio upon photolysis (1/2 hr), they are most probably due to the same species. It can be c oncluded that these four bands arise from the 12/13C2Xe isotopomers. This is supported by theo retical calculations . In Table 4-2, the deviation of computed and experi mental values is less than 1 cm 1, which confirms the assignments. Finally, the possibility that hydroge n could be attached to the species can be ruled out because the 1774.2 cm 1 band is also observed in th e laser ablation experiments. Figure 4-4. Part of the IR absorption spectra of discharge products of mixtures of 12C2H2/13C2H2/Ar (spectrum A) and 12C2H2/13C2H2/Xe/Ar (0.15%/0.04%/1.5% in Ar) (spectrum B), all trapped in solid Ar at 12 K. Spectrum C was recorded after 0.5 hr UV photolysis of the matrix characterized by spectrum B. The newly observed isotopomer bands in spectrum B and C are due to the C–C stretching vibration in 12/13C2Xe clusters (marked by filled circles). Note that the C–C vibration in C2 is IR silent and does not appear in A spectrum. The other bands marked by an empty triangle, a filled triangle, an empty star and a filled star are assigned to 12/13C3H (ref. 153), 12/13C3 (ref. 110), 13C2H (ref. 152), and 12C2H (ref. 154), respectively.

PAGE 95

77 Interestingly, the IR band from the doubly-13C-substituted C2Xe at 1705.3 cm 1 is stronger than that of the singlysubstituted species at 1740.8 or 1739.3 cm 1. This can be understood as follows. In order to form the 12C13CXe species, the CC triple bond in 13C2H2 and 12C2H2 must be broken which requires mu ch more energy than breaking the C–H bond. Thus, the production of the mixe d isotopic species is less favorable. In the discharge experiment, C2H is another important product and may be associated with the formation of C2Xe. C2H could lose a hydrogen and bond to xenon, or Xe can attack C2H to form C2Xe. Infrared Absorption of 12/13C3Xe in Argon Matrix In the IR spectrum of the produc ts from the laser ablation of a 12C / 13 C mixture in Ar seeded with a few percent Xe, there is always a small broad peak (at 2033.3 cm 1) on the low energy side of the C3 absorption. Moreover, this absorption band behaves differently from the absorption of C3 in a pure Xe matrix at 2023.0 cm 1.151 After annealing, all absorption bands of 12/13C3 decrease (Figure 4-5), while the bands (marked by dots) remain unchanged. These broad abso rption peaks also diminish in pure solid argon. But their intensities rise as the xenon c oncentration (in argon) increases. For these reasons, these bands are assigned to the 12/13C3Xe species. In Table 4-2, it can be seen that singly 13C-substituted isotopomer frequencies of either linear C3Xe or bent C3Xe are consistent with the experimental data. The largest disparity is only ca. 1.4 cm 1. In fact, the absorption of doubly 13C-substituted C3Xe and 13C3Xe can be also estimated. The observed isotopomer frequencies for 12/13C3Xe are 2008.1 ( 13 -1213 -132), 1968.8 ( 13 13 12-132), and 1955.5 ( 13 13 13 -132) cm 1, and the difference between these values and the predicted ones is less than 2 cm 1. However, due to the weak binding of C3 to Xe in

PAGE 96

78 C3Xe, 13C substitution in 12/13C3Xe generates isotopomer fre quencies not much different for bent or linear structures ( cf. Table 4-2). Thus, for larger CnXe complexes where the Cn–Xe binding energies expected are even smaller (based on the experimental C–C stretch vibrational frequency shifts of Figure 4-7) the 13C isotopomer frequencies are not likely to be sensitive to the geometry of the 12/13CnXe isomers either. Additionally, C3Xe has a rather flat potential surface with seve ral local minima, which may account for the broad absorption of C3Xe. In other words, these minima could be accessible by the trapped C3Xe, since in the matrix the C3Xe·Ar potential cage is blended with the C3–Xe potential. Figure 4-5. Part of the IR absorption spectra of discharge products from mixtures of 12C2H2/13C2H2/Xe/Ar (0.15%/0.04%/1.5% mixture) , trapped in solid Ar at 12 K. Spectrum b was recorded after annealing (to 35 K) the matrix which produced spectrum a. The newly observed isotopomer bands in spectra a and b are due to the C–C st retching vibration in 12/13C3Xe (marked by filled circles). The mechanism for the formation of C3Xe is more complicated than for C2Xe. As Figure 4-4 shows, C3H can be produced from an acetylene discharge. In analogy to the

PAGE 97

79 above mechanism for the formation of C2Xe, C3H could be a possible intermediate in the generation of C3Xe. But collision between C atoms, C2, Xe, and C2Xe species are probably the preferred pathwa ys to the formation of C3Xe in the laser ablation experiments. Generation of CnXe ( n = 3, 5, 7, 9) Species The infrared spectra of Cn clusters, produced by laser ablation and trapped with pure Ar, are shown in Figure 4-6( a-c). It can be seen that each cluster gives rise to a single sharp band within its region. Upon annealing, the concentrations of C3 and C5 decrease significantly, the signal for C7 remains unchanged, while the C9 signal increases ( cf. , Figure 4-6b). After photolysis (2 hrs), C7 and C9 are unaffected, C5 decreases slightly, and C3 decreases greatly. The C3 frequency also shifts to lower wavenumbers. This may be associated with the fact that long duration photolysis also warms the matrix, and smaller clusters may interact more strong ly with the Ar cage after restabilization during annealing. After long dura tion irradiation and subseque nt cooling, free cluster rotation in solid argon may be frozen in seve ral local minima resulting in the broadening and shifting of the C3 band. When a small amount of xenon is seeded in the argon isolant gas, additional bands appear on the low energy side of the fundamental band of each Cn cluster ( cf. , Figure 4-6 (A–C)). These new bands are assigned here to the CnXe species ( n = 3, 5, 7, 9). Figure 47 shows the IR frequency shifts for the Cn carbon clusters compar ed to the corresponding CnXe species. A smooth descending curve as a function of the number of carbon atoms can be seen. Unfortunately, because of the broadness of the CnXe bands, isotopomer frequencies for these species could not be observed. The assignment for the CnXe ( n = 5, 7, 9) species should therefore be rega rded as tentative at this point.

PAGE 98

80 Figure 4-6. Part of the infrar ed absorption spectrum of Cn and CnXe species ( n = 3, 5, 7 and 9) produced by laser ablation and isolat ed in solid Ar at 12 K (spectra a–c and A–C, respectively). Spectra b and B were scanned after matrix annealing up to 34 K and recooling to 12 K while sp ectra c and C were recorded after 2 hrs photolysis of the matrix (characteri zed by spectrum b and B) using the full spectral output from a 100 W medium pressure Hg lamp. The bands marked by filled circles are due to the C–C stretching vibrations in CnXe ( n = 3, 5, 7 and 9) species while higher energy bands at 2164.5, 2128.0, 2038.9 and 1998.0 cm 1 (not marked) are due to C–C stretches in C5, C7, C3 and C9 linear carbon clusters, respectively.47,100,155,156 The spectra in Figure 4-6 (A–C) also show that carbon clusters exhibit the same annealing and photolysis beha vior in the xenon-containing argon matrices as in pure argon matrices. But the CnXe bands show different behavior upon annealing and photolysis. First, the ratios of CnXe to Cn IR band intensities ar e different after warming and irradiating the matrices. Af ter annealing, the ratio of C3Xe to C3 is greater than 1, C5Xe is comparable to C5, C7Xe to C7 is increased, and C9Xe has a larger absorbance than C9. After photolysis, the bands for C7Xe and C9Xe are unchanged from those observed after annealing. However, the bands for C3Xe and C5Xe decreased.

PAGE 99

81 Figure 4-7. Experimental C–C asymmetric stretching frequency shifts for Cn clusters compared to CnXe complexes as a function of the number of carbon atoms. Finally, it should be reiterated that C2Xe undergoes a large frequency shift (~ 60 cm 1) compared to C2, but the shift for the CnXe ( n = 3, 5, 7, 9) species is below 10 cm 1. This difference in IR shifts is related to the fact that the Xe–C binding energy in C2Xe is much larger than for the CnXe ( n = 3, 5, 7, 9) species. Summary CnXe ( n = 2, 3, 5, 7, 9) clusters have been produced by laser ablation of compressed graphitic pellets and from a pulsed discharge through acetylene/Ar/Xe mixtures. The products have been trapped at 12 K in argon matrices and studied via FTIR absorption spectroscopic methods. Theoreti cal calculations at the MP2 level of theory show that C2Xe is linear in Ar matrices, while C3Xe probably exists as both linear and bent structures. Newly-observed IR absorption bands at 1774.2 and 2033.3 cm 1 have been assigned to the C–C stretching mode of C2Xe and C3Xe, respectively. This

PAGE 100

82 assignment is confirmed by the excellent match between theoretically predicted frequencies for the C2Xe and C3Xe isotopomers and their corresponding experimental values. C2H and C3H radicals are possible precursors for the formation of C2Xe and C3Xe, but other pathways cannot be ruled out . Tentative assignment of three new IR bands at 2161.4, 2126.6 and 1997.2 cm–1 to the CnXe ( n = 5, 7, 9) species is made here.

PAGE 101

83 CHAPTER 5 REACTION OF TRICARBON CLUS TER WITH BENZENE AND AMMONIA Introduction Carbon clusters are known to be reactive w ith a large number of heteroatoms (O, S, Se, N, H, Fe, Al, etc.). But the reacti on between carbon clusters and other stable molecules has rarely been studied. Recently, complexation of carbon clusters with water at cryogenic temperature wa s observed in argon matrices.107 In this chapter, reactions of the C3 carbon cluster with benzene and ammoni a are discussed. They were studied by FT-IR spectroscopy and substantia l theoretical calculations. Polycyclic aromatic hydrocarbons (PAHs), thought to be important constituents of the interstellar medium (ISM),157-166 have been suggested as the carriers of the unidentified interstellar infrared emission bands11,12 and of some of the diffuse interstellar visible absorption bands.167-169 Despite the potential importa nce of these species in the ISM, relatively few studies have been re ported on the mechanism of their formation under conditions that mimic interstellar envi ronments, such as molecular clouds or the outflow of carbon stars. Several gas phase chemical reaction networks have been developed, all of which proceed via benzene or phenyl radical formation.170-173 Other schemes, originally from combustion studies , have been reported which suggest the reaction of two propargyl radicals or the reaction of acet ylene with the C4H5 radical to form aromatic rings.174-176 The reaction of triplet diacetyle ne with 1,3-butadiene to form benzene and/or phenyl acetylene has also been proposed as a route to PAH formation in hydrocarbon-rich planetary environments.177 Finally, the synthesis of PAHs via solid-

PAGE 102

84 state reactions on carbon-rich grains or hydr ocarbon ices by high-energy cosmic-ray ions has also been suggested.178 Recently, in a crossed beam experiment , supported by theoretical investigation, Kaiser et al. showed that ground state carbon atoms react with benzene exothermically, with no entrance channel barrier.179-181 The reaction proceeds via formation of a complex and produces a cyclic C7H5 radical plus atomic hydrogen. In a second reaction channel the adduct product of C7H6 was observed as well. The authors conclude that because of the absence of an entrance barrier and because of the ubiquitous nature of the benzene unit in PAH-like materials, the gas phase reaction of carbon atoms with PAHs may be important in the formation of larger PAH species, both in carbon star outflows and in oxygen-poor combustion processes.181 This important conclusion suggests the possibility that other small carbon clusters, known to exist in the ISM,182,183 may also react with PAH-like materials. In this chapter, one such reaction is explored. Specifically, in solid Ar (a low temperature matrix) C3 reacts with benzene, a pparently with no entrance barrier, to produce a weakly bonded complex. The presence in the interstellar medium (ISM) of nitrogen-containing carbon chain molecules, such as HC3N and its isomer, C3NH, has been known for many years. Their mechanisms of formation are, however, stil l not completely unders tood. A variety of ionmolecule and neutral-neutral reactions have been proposed.49,184-186 In this chapter, another possibility is reported, the low te mperature formation of a complex between a small carbon chain molecule (C3) and ammonia (NH3). It is shown that under stepwise photolysis the complex rear ranges to first give C3NH and then its isomer, HC3N. Both the product molecules, C3NH (3-imino-1,2-propadienylidene) and HC3N (cyanoacetylene),

PAGE 103

85 and the reactant molecules, C3 and NH3, have been detected in th e vicinity of a number of interstellar clouds,49,185,187-192 such as the Taurus molecular cloud, TMC-1,185,193 the Sagittarius giant molecular cloud, Sgr B2(M),194 and in circumstellar envelope of the cool carbon rich red giant, IRC+10216.182,195-198 Such molecular sources are estimated to have temperatures as low as 10 100 K and are partially accessibl e to ultraviolet (UV)-visible and cosmic ray irradiation.199 Reaction of C3 with Benzene Experimental IR Spectra Figure 5-1a shows a portion of the infrared (IR) spectrum of the benzene deposited in argon at 12 K while Figure 5-1b show s the IR spectrum of a mixture of 12C3 and benzene in argon at 12 K. Four of the fi ve major bands observed in 1b are readily assigned to known carbon cluste rs or benzene: 2039.0 cm 1 (C3),149 1998.1 cm 1 (C9),78 1956.6 cm 1 (C6H6),200 and 1952.5 cm 1 (C6).78 The fifth band at 2034.5 cm 1 is ascribed here to the asymmetric C C stretch of 12C3, shifted to lower energy because of its complex formation with benzene. The 2034.5 cm 1 band is sensitive to benzene concentration: its relative intensity (vs. the C3 band) increasing with increased C6H6 concentration. Isotopic studies (12C/13C), shown in Figure 5-2, confirm the assertion of complex formation. The six 12/13C3 isotopomeric peaks (marked with black triangles) are apparent. Another set of peaks, approximately 4.5 cm 1 to lower energy of each 12/13C3 peak, can also be noted. An asymmetric C3•C6H6 complex (e.g. one end of 12/13C3 complexed to C6H6) is expected to yield eight isotopomer ic peaks, as observed previously for the C3•H2O complex.201 The six isotopomeric bands observed in Figure 5-2 suggest a symmetric orientation of C3 and benzene.

PAGE 104

86 Figure 5-1. Part of infrared absorption spectrum of C6H6 isolated in solid Ar at 12 K (spectrum a) and the carbon clusters (Cn, n = 3, 6 and 9) deposited with the mixture of Ar/C6H6 (0.5%) at 12 K (spectra b-d). Spectrum c was recorded after 3 min. photolysis of the matrix (characterized by spectrum b) using full spectral output from a 100 W medium pressure Hg lamp, while spectrum d was scanned after matrix annealing up to 34 K and recooling to 12 K. The 2034.5 cm 1 band (C3 C6H6 ) is also overlapped by the weak band of C3 trapped in a minor site of Ar.151 Note a group of photoproduct bands in the 1960 cm 1 energy region (marked p) as well as combination bands of C6H6 (marked by single open circles) and (C6H6)n polymer benzene band (marked by double open circles). It was not possible to extract the benzene mode vibrations in the complex with C3 (cf, spectrum in Figure 5-1b), even though the absorbance of the 2034.5 cm 1 band is ca. 0.1. The predicted (B3LYP/6-31G(d,p)) integral intensity for the 2034.5 cm 1 band is ca. 500 km/mol, but the predicted intensity fo r the most intense vibrational mode ( 4) in benzene complexed to C3 is only 90 km/mol. Moreover, theory predicts the frequency for this mode in the complex to lie at 680 cm 1 (scaled), on the low energy shoulder of the very strong, uncomplexed benzene band at 675.2 cm 1. The 680 cm 1 band also overlaps with polymeric benzene bands (with multiple sites) at 680.6 (photosensitive), 679, 677.8, 677.1 (photosensitive) and 676.7 cm 1.

PAGE 105

87 Figure 5-2. Infrared abso rption spectrum of the 12/13C3 carbon cluster (isotopomeric bands marked by triangles) and 12/13C3 perturbed by benzene in the 12/13C3 12C6H6 complex (bands marked by circles) (s pectrum a) and after 3 min. matrix photolysis using full spectral output of medium pressure Hg lamp (spectrum b) all isolated in solid Ar at 12 K and displayed in the C–C asymmetric stretch vibration energy region. The photoproduct ba nds in spectrum b are marked p. The frequencies and isotopomeric band assignments of 12C3 and 12/13C3 12C6H6 are listed in Table 5-1. The concentra tion of benzene in Ar was 1.0%. Note the small frequency band shift to lowe r energy compared to Figure 5-1 where a lower benzene concentration was used. Stable Structures Thirteen stable structures for the singlet C3 / C6H6 system were found using B3LYP/6-31G(d,p) and B3LYP/6311+G(d,p) theory. These stable forms, displayed in Figure 5-3, correspond to the following five init ial lines of approach: 1) out-of-plane approach of C3 along the benzene six-fold axis ( A , B and F ), 2) in-plane approach of C3 perpendicular to a benzene C C bond ( C , D , E and I ), 3) colinear approach of C3 with a benzene C H bond (no stable structure found202), 4) end-on approach of C3 to a benzene carbon atom ( L and M ), and 5) in-plane approach of C3 parallel to a benzene C C bond (initial structure switches to A ).

PAGE 106

88 A (–8.25) B (–8.20) Separate C3+ C6H6(0.0) C (–51.7)D (–202.6)E (–277.5)F (25.0) G (–48.6)H (–43.1)I (–51.6)J (–100.6) K (–234.0)L (–251.0)M (–281.3)N (–238.5)1.297 1.297 1.2973.145 3.145 3.151 3.136 1.3961.3931 . 4 0 01 . 4 0 01 . 4 0 01.3931 . 4 0 00.031 . 4 11.450.031 . 4 11.95 1.941.44 0.033 0.0341.297 1.961.94 1.451 . 4 11 . 4 11.454 1.5961 . 3 4 91 . 4 8 01 . 4 6 91.318 1.2871.15 0.861 . 7 81 . 0 70 . 9 62.13 1.711.3691 . 4 2 71 . 3 6 91 . 4 4 71.3721.273 1.41 2.311 . 2 41 . 6 11 . 1 81.59 1.3681 . 4 4 91 . 3 6 51 . 4 6 81 . 45 51 . 3 2 91 . 3 4 01.426 1.2111. 681 . 2 01 . 7 01 . 0 71 . 6 91 . 8 11. 131.09 2.821 . 2 9 81.5951 . 5 2 71 . 5 2 71.335 1.901 . 0 01 . 0 01 . 8 70.891 . 4 6 11 . 3 0 81 . 4 6 21 . 3 3 81 . 3 3 21 . 4 5 21 . 3 6 91 . 4 4 71 . 3 6 51 . 2 01 . 6 91 . 7 01 . 8 01 . 0 81 . 1 41 . 6 71 . 0 61 . 9 81.4631 . 4 6 31 . 3 0 71 . 3 6 61 . 4 4 71 . 3 6 91 . 4 5 31 . 3 3 01 . 3 3 71 . 2 11 . 6 91 . 0 81 . 7 01 . 8 01 . 1 31 . 6 81.052 . 0 01.454 1.5961 . 3 4 91 . 4 8 01 . 4 6 91 . 3 1 81.2871.15 0.861 . 7 81 . 0 72.131 . 7 10 . 9 61 . 50 51 . 3 4 71 . 4 8 71 . 3 3 11 . 2 8 21. 031 . 9 01 . 0 41 . 7 61 . 9 11.304 1.3081 . 4 5 11 . 3 8 21 . 4 1 01.406 1.370 1.45 1.281 . 4 11 . 4 01 . 0 21.77 1.96 1.390 1.388 1.4261 . 4 1 51 . 4 1 91 . 4 0 11 . 3 9 81 . 3 0 41 . 2 8 81.48 1.46 1.201 . 4 21 . 4 11 . 3 01 . 3 21 . 8 32 . 0 51.398 1.411 1.395 1.434 1.42 1.32 1.46 1.301 . 3 9 11 . 4 1 0 1 . 4 011 . 3 8 7 1 . 4 5 01 . 3 4 5 1 . 2 8 11 . 4 6 1. 4 8 1. 161 . 3 41 . 3 41 . 6 1 2 . 2 11. 40 2 1. 38 41 . 3 8 8 1.4051. 41 1 . 471 . 4 1 1 . 3 91 . 4 6 51 .5 5 51 . 5 1 41 . 2 3 3 2 . 7 11 . 0 31 . 0 01 . 0 9 A (–8.25) B (–8.20) Separate C3+ C6H6(0.0) C (–51.7)D (–202.6)E (–277.5)F (25.0) G (–48.6)H (–43.1)I (–51.6)J (–100.6) K (–234.0)L (–251.0)M (–281.3)N (–238.5)1.297 1.297 1.2973.145 3.145 3.151 3.136 1.3961.3931 . 4 0 01 . 4 0 01 . 4 0 01.3931 . 4 0 00.031 . 4 11.450.031 . 4 11.95 1.941.44 0.033 0.0341.297 1.961.94 1.451 . 4 11 . 4 11.454 1.5961 . 3 4 91 . 4 8 01 . 4 6 91.318 1.2871.15 0.861 . 7 81 . 0 70 . 9 62.13 1.711.3691 . 42 71 . 3 6 91 . 4 4 71.3721.273 1.41 2.311 . 241 . 6 11 . 1 81.59 1.3681 . 4 4 91 . 3 6 51 . 4 6 81 . 45 51 . 3 2 91 . 3 4 01.426 1.2111. 681 . 2 01 . 7 01 . 0 71 . 6 91 . 8 11. 131.09 2.821 . 2 9 81.5951 . 5 2 71 . 5 2 71.335 1.901 . 0 01 . 0 01 . 8 70.891 . 4 6 11 . 3 0 81 . 4 6 21 . 3 3 81 . 3 3 21 . 4 5 21 . 3 6 91 . 4 4 71 . 3 6 51 . 2 01 .6 91 . 7 01 . 8 01 . 0 81 . 1 41 . 6 71 . 0 61 . 9 81.4631 . 4 6 31 . 3 0 71 . 3 6 61 . 4 4 71 . 3 6 91 . 4 5 31 . 3 3 01 . 3 3 71 . 2 11 . 6 91 . 0 81 . 7 01 . 8 01 . 1 31 . 6 81.052 . 0 01.454 1.5961 . 3 4 91 . 4 8 01 . 4 6 91 . 3 1 81.2871.15 0.861 . 7 81 . 0 72.131 . 7 10 . 9 61 . 50 51 . 3 4 71 . 4 8 71 . 3 3 11 . 2 8 21. 031 . 9 01 . 0 41 . 7 61 . 9 11.304 1.3081 . 4 5 11 . 3 8 21 . 4 1 01.406 1.370 1.45 1.281 . 4 11 . 4 01 . 0 21.77 1.96 1.390 1.388 1.4261 . 4 1 51 . 4 1 91 . 4 0 11 . 3 9 81 . 3 0 41 . 2 8 81.48 1.46 1.201 . 4 21 . 4 11 . 3 01 . 3 21 . 8 32 . 0 51.398 1.411 1.395 1.434 1.42 1.32 1.46 1.301 . 3 9 11 . 4 1 0 1 . 4 011 . 3 8 7 1 . 4 5 01 . 3 4 5 1 . 2 8 11 . 4 6 1. 4 8 1. 161 . 3 41 . 3 41 . 6 1 2 . 2 11. 40 2 1. 38 41 . 3 8 8 1.4051. 41 1 . 471 . 4 1 1 . 3 91 . 4 6 51 .5 5 51 . 5 1 41 . 2 3 3 2 . 7 11 . 0 31 . 0 01 . 0 9 Figure 5-3. Structures (A-N) of stable minima of the C6H6 (X1A1g) +C3(g +) reaction products found (B3LYP/6-31G(d,p) theo ry) on the singlet potential surface. The bond lengths are in , while the zero point-corrected energies, with respect to the separated products, (in kJ /mol) are in parentheses. The Wiberg Bond Indices are given in bold type. Ne gative energies indicate that the energy of the product structure lies belo w the total energy of the separated reactants. The total energies of the fourteen A-N systems were minimized with all geometry optimizations permitted. Only singlet spin potential surfaces were considered since ground state C3 (1g +) and C6H6 (1Ag) are both singlet. The triplets of the calculated systems lie at much higher energies. In matrix experiments, it is likely that the reaction between benzene and C3 occurs during landing of the re actants on the matrix surface. Both reactants are thus expected to be in th eir electronic ground st ates and to have low kinetic energies. It is conceivable that a small fraction of C3 species could be deposited in

PAGE 107

89 the metastable 3 state (at 2.10 eV) via intersystem crossing from the excited 1 state (at 3.04 eV) if UV radiation from the vaporizati on plasma were present. But this is considered unlikely in the presen t experiments since no blue Ar/Cn plasma was observed during vaporization. Many interesting stable structures for the C3/C6H6 system are predicted. These include nonplanar I and C, in which C3 is bonded to two benzene C atoms, structure K with a C3H2 unit bonded in a similar way, the nine-membered J ring structure, and structure N with a five-membered ring fused to a be nzene ring (an indene-like structure). I is reminiscent of the structure found by Hahndorf et al .181 from a similar calculation on the C6H6/C (3P) system, while D is structurally similar to the stable seven-membered ring of the C6H6/C system.181 Wiberg Bond Indices (WBI) of ca. 1.0 for a “pure” single bond (C C in ethane), ca. 2.0 for a double bond (C=C in ethylene) and ca. 3.0 for a triple bond (C C in acetylene) are useful in predicting the CC bond characters in unknown structures.203 The WBI and bond length values for the A – N structures were determined and are displayed in Figure 53. The B3LYP/6-31G(d,p) ZPE corrected en ergies are collected in Table 5-1. A and B are very similar. In B , C3 is rotated a little relative to benzene ( cf. Figure 5-3). This small structural difference induces a small ( ca. 0.05 kJ/mol) difference in the binding energy of the complex. C and I have similar bond lengths. However, in C the C C C (C3) C (benzene) dihedral angle is 58 , while in I it is 92 . The WBI values of 1.71 (2.13) for C ( I ) indicate that the CC bonds in the C3 unit are less cumulenic than they are in free C3 (1.95). Since the WBI values in th e CCC-triangle of these structures are below one (0.86, 0.96), the structures s hould be considered to be complexes.

PAGE 108

90 Table 5-1. Zero-point-correct ed relative energies, Erel (kJ/mol) and zero point energies, ZPE (kJ/mol) of various minima and transition states found on the C3+C6H6 singlet potential surface. The Erel and ZPE were calculated at B3LYP/631G(d,p) (and/or 6-311+G( d,p)) levels of theory. B3LYP B3LYP 6-31G(d,p) 6-311+G(d,p) 6-31G(d,p) Stable Structure Erel a ZPE Erel a ZPE Transition State Erel a ZPE A 8.25 289.6 4.4 287.6 TS1 7.5 288.8 B 8.20 289.5 4.2 287.7 TS2 53.7 291.8 C 51.7 295.7 45.6 293.8 TS3 47.7 293.0 D 202.6 302.7 199.4 301.0 TS4 53.2 282.4 E 277.5 298.9 234.1 295.0 TS5 53.9 293.4 F 25.0 296.2 29.6 295.4 TS6 53.9 293.4 G 48.6 290.1 46.1 288.3 TS7 53.75 291.8 H 43.1 289.2 39.9 287.5 TS8 54.0 282.1 I 51.6 295.6 45.7 293.8 TS9 32.6 287.4 J 100.6 293.1 102.5 292.2 TS10 43.5 292.7 K 234.0 295.9 261.8 296.7 TS11 2.2 291.4 L 251.0 293.7 249.3 293.0 TS12 4.6 287.2 M 281.3 295.6 274.8 299.1 TS13 29.6 289.7 N 238.5 302.2 228.8 301.2 TS14 83.9 282.8 TS15 85.8 274.6 TS16 48.1 289.1 TS17 45.5 293.0 a Negative energies are below the sum of zer o-point-corrected ener gies of separated C3 and C6H6. b Zero-point corrected energies fo r reactants (in Hartrees) are: C3; 114.037 45 (B3LYP/6-31G(d,p)), 114.068 85 (B3LYP/6-311+G(d,p)) and C6H6; 232.157 59 (B3LYP/6-31G(d,p)), 232.211 10 (B3LYP/6-311+G(d,p)). Another predicted complex, F , has an inter-moiety (C3-C6H6) WBI index of 0.89 and a long “bond length” of 1.595 Å, but has a po sitive total energy and therefore is not a stable structure with respect to A or B. Of the seven-membered ring structures ( D , E , G and H), E is the most stable with a stabilization energy of 277.5 kJ/mol. The rings in these structures are formed by the incorporation of one of the C3 carbons into the benzene ri ng and migration of one hydrogen along the C3 unit. The most stable structure is the one with hydrogen on the end carbon of C3 ( i.e. , E ).

PAGE 109

91 In the cyclic J structure, the carbon framework is almost planar and the hydrogens out-of-plane. In M, the most stable structure globally, the six-membered ring is predicted to form from L, via a H migration from one end of the C3 unit to the other, but this requires a calculated energy of 299 kJ/mol ( vide infra ). Finally, N, a relatively stable, closed shell indene-like structure, is a doubly dehydrogenated polycyclic aromatic hydrocarbon (PAH) containing fused fiveand six-membered rings. A high WBI value of 2.71 is calculated between the C atoms in the five-membered ring with no attached HÂ’s indicating a triple bond between these atoms. Although structures D , E , K , L , M, and N would, at first glance, appear to be the most stable for the C3/C6H6 system, and therefore could be considered as likely products in these experiments, a comparison of the calculated and observed matrix CC asymmetric stretch frequencies indicates otherwise ( vide infra ). Predicted Infrared Frequencies Inspection of the computed infrared abso rption spectra (B3LYP /6-31G(d,p)) for the A N structures ( cf. Figure 5-4) shows that the relativ ely strongly bonded systems such as C N possess a C C asymmetric stretch vibration shifted to lower frequencies by more than 100 cm 1. (The cyclic J structure shows a weak band shifted by only 63 cm 1). These shifts are too large compared to the observed 4.5 cm 1 shift of the 2034.2 cm 1 band. Therefore, the C N species are unlikely to be responsible for the 2034.2 cm 1 band. On other hand, the weakly bonded A and B structures have relatively small calculated frequency shifts (22 cm 1) and are the strongest candi dates for this assignment.

PAGE 110

92 Figure 5-4. Predicted (B3LYP/6-31G(d,p)) infr ared absorption spectra of the products (whose structures are given in Figure 53) of the reaction of benzene with C3. The IR experimental spectrum of unperturbed C3 and C3 C6H6 clusters are also plotted. All band frequencies are scaled uniformly by 0.942. In the past, comparison of predicted and observed isotopomer frequencies has been used as a guide to the correct geometry of the system under study. Such a comparison for

PAGE 111

93 Table 5-2. Comparison of observed (in solid Ar at 12 K) vibrational frequencies (in cm 1) of 3 mode of 12/13C3 isotopomers and 12/13C3 isotopomers pertubated by benzene in the 12C6H6•12/13C3 complexes (structures A , I Both Formed in reactions with no entrance barrier predicted) and structure M (a most stable isomer found) of the complex from Fi gure 5-3) with the calculated harmonic frequencies at B3LYP/6-31G(d,p) Level. Isotopomer Observed /cm 1 Calculated a /cm 1 Scaled a /cm 1 Difference /cm 1 C3 Cluster 12-12-12 2038.7 2164.3 2038.7 0.0 13-12-12 2025.4 2150.5 2025.7 0.3 13-12-13 2012.6 2136.3 2012.3 0.3 12-13-12 1987.2 2108.1 1985.8 1.4 13-13-12 1974.1 2094.0 1972.5 1.6 13-13-13 1960.1 2079.4 1958.7 1.4 Structure A C6H6 12-12-12 2034.2 2140.4 2034.2 0.0 C6H6 13-12-12 2021.5 2126.9 2021.4 0.1 C6H6 13-12-13 2008.1 2112.7 2007.9 0.2 C6H6 12-13-12 1982.5 2084.9 1981.5 1.0 C6H6 12-13-13 1969.7 2071.0 1968.2 1.5 C6H6 13-13-13 1956.1 2056.4 1954.4 1.7 Structure I C6H6 12-12-12 2034.2 2061.4 2034.2 0.0 C6H6 13-12-12 2021.5 2047.6 2020.6 0.9 C6H6 12-12-13 2021.5 2047.3 2020.3 1.2 C6H6 13-12-13 2008.1 2032.4 2005.6 2.5 C6H6 12-13-12 1982.5 2011.1 1984.6 2.1 C6H6 13-13-12 1969.7 1996.4 1970.1 0.4 C6H6 12-13-13 1969.7 1996.9 1970.6 0.9 C6H6 13-13-13 1956.1 1981.0 1954.7 1.4 Structure M C6H6 12-12-12 2034.2 2047.6 2034.2 0.0 C6H6 13-12-12 2021.5 2039.9 2026.6 4.1 C6H6 12-12-13 2021.5 2028.7 2015.4 6.1 C6H6 13-12-13 2008.1 2020.4 2007.2 0.9 C6H6 12-13-12 1982.5 1995.8 1982.7 0.2 C6H6 13-13-12 1969.7 1987.9 1974.9 5.2 C6H6 12-13-13 1969.7 1976.4 1963.5 6.2 C6H6 13-13-13 1956.1 1967.7 1954.8 1.3 a Harmonic frequencies calculated at B3LYP/631G(d,p) and scaled uniformly by a factor of 0.9420 and (0.9504, 0.9868 and 0.9935) for 12/13C3 and 12C6H6•12/13C3 (structure A , I and M ) isotopomers, respectively. b Isotopomer frequencies calculated for B isomer are very similar to those of isomer A with the maximum differences of 0.2 cm 1 and are not listed. Note the eight isotopomeric bands are predicted for I and M structures.

PAGE 112

94 the present system is given in Table 5-2 for A , B , I (a moderately bonded system), and M (the most stable isomer). As a calibration of acceptable differences, non-complexed C3 isotopomeric frequencies are also given. Th e table shows that only for structures A and B are the frequency differences in the accep table range. Despite their greater overall stability, isomers I and M do not adequately account for the observed isotopomeric IR peaks, whereas isomers A and B do. Thus, only the A and B complexes are capable of predicting the pattern an d spacing of the isotopic IR bands, as well as the small frequency shift observed in the C3 asymmetric stretch. Possible C3 + C6H6 Reaction Pathways The structures and energies of the trans ition states connecting the stable minima on the singlet potential surface are displayed in Figure 5-5. Figure 5-6 sketches part of the C3 + C6H6 singlet potential surface from B3LYP/631G(d,p) calculations. Three structures A , B and I are formed directly from adduct reactions of C3 + C6H6 with no entrance barriers. A and B are connected by the TS1 transition state at 7.5 kJ/mol. All reactions from A and B have energy barriers of ca. 54 kJ/mol above the total energy of the separated C3 + C6H6 reactants. Thus, C , D , E, and I are not likely to exist under the present experimental conditions from A and B reaction pathways, but may be accessible in reactions from I, as shown in Figure 5-6b. Comput ations also predict that F , G and H will not be produced from A or B . Therefore, observation of only A (and/or B ) is also supported by the potential surface calc ulations which predict that the C3 + C6H6 A (and/or B ) reaction will stop if the reactants are in their ground electronic states. One of the more important predicted reacti on pathways is the formation of indenelike structure N . The C3 + C6H6 I TS12 K TS13 N branch of this reaction

PAGE 113

95 TS1 (–7.5)TS2 (53.7)TS3 (–47.7)TS4 (–53.2)TS5 (53.9) TS6 (53.9)TS7 (53.75)TS8 (54.0)TS9 (–32.6) TS10 (43.5)TS11 (–2.2)TS12 (–4.6)TS13 (–29.6) TS14 (83.9)TS15 (85.8)TS16 (48.1)TS17 (–45.5) 3.747 3.094 TS1 (–7.5)TS2 (53.7)TS3 (–47.7)TS4 (–53.2)TS5 (53.9) TS6 (53.9)TS7 (53.75)TS8 (54.0)TS9 (–32.6) TS10 (43.5)TS11 (–2.2)TS12 (–4.6)TS13 (–29.6) TS14 (83.9)TS15 (85.8)TS16 (48.1)TS17 (–45.5) 3.747 3.094 Figure 5-5. Structures of the first order transition st ates of the C6H6 (X1A1g) +C3 (g +) reaction products found on the singlet potential surface (B3LYP/6-31G(d,p)). Bond lengths are in , while the zero point-correct ed energies (in kJ/mol) are in parentheses. Negative energies indi cate that the energy of the product lies below the total energy of the separated reactants. is a low energy pathway for which all the transition state energies lie below the sum of the reactant energies. Previous workers have reported the production of indene in a 6001000 V (100-200 mA) benzene/Ar gaseous discharge.204 Previous unpublished matrix experiments from Vala’s laboratory with a 1000 V (50 mA) benzen e/Ar jet discharge revealed IR bands due to small car bon and hydrocarbon clusters with C3 most prominent. Thus, in the C6H6/Ar discharge of ref. 204, C3 and C6H6 in the plasma could react exothermically, following the reaction scheme of Figure 5-7, to form the indene-like N structure as well as indene itself in an exothermic doubl e hydrogen addition reaction.

PAGE 114

96 C3+C6H6TS253.7 A–8.25 B–8.20 TS1–7.5 TS753.75 C–51.7 TS553.9 TS854.0 TS653.9 F25.0 D–202.6 E–277.5 TS3–47.7 TS4–53.2 G–48.6 H TS9–32.6 I–51.6a C3+C6H6TS253.7 A–8.25 B–8.20 TS1–7.5 TS753.75 C–51.7 TS553.9 TS854.0 TS653.9 F25.0 D–202.6 E–277.5 TS3–47.7 TS4–53.2 G–48.6 H TS9–32.6 I–51.6a C3+C6H6TS1043.5 I–51.6 J–100.6 TS11–2.2 TS12–4.6 C–51.7 TS1483.9 TS1585.8 TS1648.1 D–202.6 E–277.5 TS3–47.7 TS4–53.2 K–234.0 M–281.3 TS13–29.6 L–251.0 N–238.5 b TS17–45.5 C3+C6H6TS1043.5 I–51.6 J–100.6 TS11–2.2 TS12–4.6 C–51.7 TS1483.9 TS1585.8 TS1648.1 D–202.6 E–277.5 TS3–47.7 TS4–53.2 K–234.0 M–281.3 TS13–29.6 L–251.0 N–238.5 b TS17–45.5 Figure 5-6. Schematic diagram of the singl et potential surface of reaction of C3 (X1g +) carbon cluster with C6H6 (X1Ag) benzene showing stable minima ( A M ) with the connected transition states ( TSn, n = 1-17). The zero point-corrected energies are given in kJ/mol. The A , B and I are predicted to form exothermically with no entrance barrie rs. The reaction pathways starting from A (or B ) and I are given in panels a and b , respectively. Negative energy values indicate that the energy of stru cture lies below the total energy of the separated reactants.

PAGE 115

97 +C6H6C3(0.0) I ( 51.6) K ( 234.0) N ( 238.5) TS12 ( 4.6) TS13 ( 29.6) TS1 ( 7.5) TS2 (53.7) A ( 8.25) B ( 8.20) +C6H6C3(0.0) I ( 51.6) K ( 234.0) N ( 238.5) TS12 ( 4.6) TS13 ( 29.6) TS1 ( 7.5) TS2 (53.7) A ( 8.25) B ( 8.20) Figure 5-7. Formation mechanism for the PAH (indene-like form, Structure N) from the reaction of the ground state benzene and C3. The zero point-corrected energies (kJ/mol) are listed for the stable A , B , I , K and N structures and for the TS1 , TS2 , TS12 and TS13 transition states. This poses the question of why no evidence is seen for I or its products E , J and N in the matrix experiments? This could be expl ained by the fact that there is a significant Ar cage potential that was not consid ered in the calculations. Producing I requires significant changes in the bond distances and angles of the benzene unit but not for A or B . In fact, in I the benzene unit is highly nonplanar ( cf. Figure 5-3). The two hydrogens closest to C3 are pushed drastically out-of-plane in the opposite direction from the C3

PAGE 116

98 unit. The CC bond lengths in the C3 ring are long: 1.596 Ã… ( vs. 1.396 Ã… for the bond originally in benzene) an d 1.469 Ã… for the other two that connect benzene to the C3 chain. The CC bond lengths of the C3 moiety are close to the original lengths in free C3, with the outer one having contracted sli ghtly to 1.287 Ã… and the inner one elongated slightly to 1.317 Ã… (from 1.297 Ã… in free C3). However, for the adduct forms of A and B, only marginal bond length changes vs. free benzene were calculated ( cf. Figure 5-3). It is postulated, therefore, that isomer I is not observed and complexes A and B are, because the former has a non-negligible en trance barrier and the latter do not. Photolysis Experiments Matrix photolysis (medium pressure Hg lamp, full spectral output, h < 5.2 eV, 3 min.) destroys the 2034.5 cm 1 complex band almost completely ( cf. Figure 5-1c and Figure 5-2b). Simultaneously with this decrea se, several new IR bands grow in. They are located at 1963.4, 1961.2, 1959.2 and 1953.8 cm 1 (all displayed in Figure 5-1 and labelled by the letter p), 1515.7 (broad), doubl et 1278.3, 1277 (not displayed) and doublet 832.9, 832.2cm 1 (not displayed) . Although the 1963.4 cm 1 band has been ascribed previously to propadienylidene (C3H2 isomer), with one-third the intensity of the 1952 cm 1 band (also assigned to the same species),205 this attribution s hould be questionable. If the 1963.4 cm 1 band were also due to propadienylidene, the 1952.0 cm 1 band should be able to be seen, but it is not. Previously, only the 1952.0 cm 1 band had been observed in an Ar matrix in this region and was assigned to propadienylidene.206 The 1959.2 cm 1 photoproduct band in Figure 5-1 can be tent atively assigned to the CC stretch in HCCH=C, another isomer of C3H2. A band at 1959.5 cm 1 band (with secondary trapping site bands at 1957.4 and 1960.5 cm 1) has previously been re ported for this isomer.206

PAGE 117

99 Small red shifts (several tenths of cm 1) were seen in the experiments when using a relatively high concentration of ben zene in Ar (or Kr). The 1278.3 and 1277 cm 1 bands are likely due to C3H2 (cyclopropenylidene), based on proximity to the 1278.6 and 1277.7 cm 1 bands already assigned to this molecule.206 In Kr, photofragment bands are located at 1959.3, 1957.8, 1952.8, 1949.4, 1513.7 and 1275.0 cm 1. The photodissociation experi ments carried out for the 12/13C3•C6H6 complexes show no change in these positions compared to the 12C3•C6H6 complexes ( cf. Figure 5-1 and Figure 5-2). This implies that these photoproducts come from benzene and not from 13C-substituted C3. Upon benzene deuterium substitution, the complex C3 asymmetric stretching frequency shifts 0.3 cm 1 to lower frequency, compared to a 0.6 cm 1 shift predicted by theory for the A structure. Photolysis experiments on the C6D6/12Cn/Ar system produced broad photoproduct bands centered at 1957.7 and 1501.7, and 1499.2 cm 1 (double site). These bands are likely associated with the respective 1959.2 and 1517.3, 1515.7 cm 1 photoproduct bands observed from the C6H6/12Cn/Ar experiments. Longer photolysis (0.5 hr) generates additional weak bands at 743.4, 771, 926.8, 1089.9, 1194.9, 1343.3, 1488.8, 1800.2, 1815.3, 1936.5, 2064.1, 2080.4, 2092.5, and 2158.1 cm 1. To identify the bands associated with isomerization and/or photodissociation products of benzene, an independent C6H6/Ar matrix photolysis experiment was carried out . A short UV-visible photolys is (3min) generated no observable photoproducts. However, more ex tensive UV photolysis (1hr) produced low intensity product bands at 707, 736, 742.8, 771, 926.8, 1089.8, 1313, 1343.3, and 1488.8 cm 1. The assignment of these bands, proposed in Table 5-3, is based on proximity to

PAGE 118

100 literature energy values and supporting calculations. The photoisomerization of benzene to fulvene and benzvalene, observed here, is consistent with the observations of Johnston and Sodeau.207 Table 5-3. Proposed assignment of the photopr oduct IR absorption bands observed after UV-Visible photolysis (h < 5.2 eV) of the C6H6/Ar and C6H6/Cn/Ar matrices. C6H6/Ar matrix, 12 Ka C6H6/Cn/Ar matrix, 12 Kb exp/cm 1 Assignment exp/cm 1 Assignment 707 (7) Phenyl, ref. 208 736 (4) Benzyne, ref. 209 742.8 (12) Benzvalene, ref. 207 743 (8) Benzvalene (calc. 766)209 771 (25) Fulvene, ref. 207 771 (6.5) Fulvene (calc. 801)206 833 (11) 926.8 (15) Fulvene, ref. 207 926.8 (2) Fulvene (calc. 961)32 1089.8 (5.5) Benzvalene, ref. 207 1089.9 (3.5) Benzvalene (calc. 1130)175 1194.9 (4) 1277 (6.1) 1278.3 (1.3) Cyclopropenylidene (calc.1316)205,206 1313 (2) Benzvalene, ref. 207 1343.3 (13) Fulvene, ref. 207 1343.3 (3.8) Fulvene (calc. 1383)177 1488.8 (5) Fulvene, this work 1488.8 (2.0) Fulvene (calc. 1551)181 1515.7 (27) 1517.3 (23) 1800.2 (2.4) C8 anion, T 1815.3 (5.1) C12 anion, T 1936.5 (9.3) C6 anion, ref. 106 1953.8 (23) 1959.2 (79) 1961.2(12) HCCH=C, ref. 206, T Secondary site 1963.4 (42) 2064.0 (4) C8 anion, ref. 106 2080.4 (3.4) 2092.5 (2) C10 anion, T 2158.1 (9.5) a One hour photolysis of the C6H6/Ar (1:400) matrix using a medium pressure Hg lamp (100 W) with full spectral output. b Half hour photolysis of the matrix containing small carbon clusters Cn and C6H6/Ar (1:200) using a medium pressure Hg la mp (100 W) with full spectral output. Band absorbances in experimental spect ra are listed (in parentheses) in 10 3 units. B3LYP/6-31G(d,p) calculated (uns caled) frequencies and integr al intensities (in km/mol, in brackets) of most intense modes of described molecules are listed in last column, this work. T, stands for tentative assignment.

PAGE 119

101 In addition, the band at 1488.5 cm 1 is assigned to fulvene since its intensity correlates to other fulvene ba nds and theory (B3LYP/6-31G (d,p)) predicts a relatively strong vibration in this region (1551 cm 1, unscaled with 27 km/mol integral intensity). Observation of the Cn ( n = 6, 8, 10 and 12) carbon cluste r anions from photolysis of the C6H6/12Cn/Ar matrix ( cf. Table 5-3) requires a few co mments. First, these anions are generated from photolysis of the 12Cn/Ar matrix as well. During matrix photolysis, electrons are detached from the anions whos e photodetachment energies are smaller than ca. 5.2 eV (accessible photon energy from Hg lamp) including Cn ( n < 6 and nall odd ) carbon anions. The Cn ( neven 6) possess large electron gas phase electron affinities (EA), ranging from ca. 4.2 eV for C6 to ca. 4.4 eV for C12.78 A stabilization energy ca. 1 eV should be added to the gas phase EA valu es to estimate the photodissociation energy threshold of anions in the matrix.203 Thus, for Cn ( neven> 6) anions in Ar, the photodissociation energy is just above 5.2 eV. Concen trations of these anions, therefore, are expected to build-up during Hg lamp photolysis, as observed ( cf. Table 5-3). At the same time, intensities of the 1952.5 cm 1 (C6), 2071.5, 1710.4 cm 1 (C8), and 1915.7 cm 1 (C10) bands decrease by ca. 10, 27, and 35%, respectively. This observation supports the idea that the larger even carbon clusters are e ffective electron scavengers in solid Ar. For photolysis times longer than 0.5 h, the concentration of C6 starts to decline, presumably because not many free electrons are left in the matrix and because photodissociation processes start to compete with electron attachment events for this cluster. Energy transfer from metastable 3 state of C3 to the benzene unit may occur and could explain the large difference in photol ysis time required to disrupt the benzene skeleton. From the photolysis products observed, it can be concluded that pumping either

PAGE 120

102 the A and/or B system does not result in matrix reactions traversing the TS2 and TS5 TS8 transition states ( cf. Figure 5-6a). The geometries of A ( B ) and TS2 ( TS5 TS8 ) ( cf. Figure 5-5) show that large stru ctural deformations, involving C3 rotation and benzene skeleton deformation, are required for such re actions. Presumably, such deformations are not allowed by the rigid matrix environment. The photon threshold necessary to initiate bleaching of the complexes was investigated using a set of optical filters with progressively higher cutoffs. It was found that a minimum energy of ca. 1.77 eV (700 nm) is required to initiate bleaching in Ar. This value, of course, include s the stabilization energy of th e complex in the Ar matrix cage. Although unknown for the C3 C6H6 /Ar system, for other molecular systems, even those isolated in a low perturbing matrix like Ne, matrix stabilization energies as high as ca. 1 eV have been found.203 Formation of the 1:2 Complexes The band at 2036 cm 1 is photosensitive ( cf. , Figure 5-1) and grows during matrix annealing. It is arguable that this band could be due to the larger complexes C3 (C6H6)2, C6H6 C3 C6H6, or C3 C6H6 C3. The formation of polymeric benzene in Ar matrices has been observed earlier200 and is indicated in the present work as well. Brown and Person200 have assigned the strong band at 675.2 cm 1 to the 4 (a2u) strongest fundamental mode of monomeric benzene. IR spectra recorded for benzene deposit ed in Ar at concentration ratios of 1:200 and 1:4000 indicate that the 679.0 and 680.6 cm 1 satellite bands of 4 (a2u) grow with benzene concentration as well as during matrix annealing, wh ile, at the same time, the monomer benzene bands decrease in intensity. From this pictur e it can be surmised that

PAGE 121

103 benzene is mobile in the Ar matrix and form s benzene polymers. If so, it is also likely that mobile benzene may be r eacting with already-formed C3 C6H6 complexes to generate 1:2 complexes. The increased mobility of benzene during matrix annealing is also indicated by the growth of the 2034.5 cm 1 band ( cf. , Figure 5-1d), here assigned to the C3 C6H6 complex. O (–11.5) P (–14.1) R (–14.9) 3.321 3.328 3.330 3.5503.071 3.2103 . 8 4 73.186 3.186 3.186 3.186 1.295 1.2991.2971 . 2 971 . 2 9 7 O (–11.5) P (–14.1) R (–14.9) 3.321 3.328 3.330 3.5503.071 3.2103 . 8 4 73.186 3.186 3.186 3.186 1.295 1.2991.2971 . 2 971 . 2 9 7 Figure 5-8. Structures of stable minima of the 1:2 C3/benzene ( O, P ) and 2:1 C3/benzene ( R ) complex found on the singlet potentia l surface (B3LYP/6-31G(d,p)). The zero point-corrected energies with re spect to the separated products (in kJ/mol) are in parentheses.

PAGE 122

104 Table 5-4. Zero-point-correct ed relative energies, Erel (kJ/mol), zero point energies, ZPE (kJ/mol) and vibrational frequencies of CC asymmetric stretch vibration in C3 unit calculated at B3LYP/ 6-31G(d,p) level of theory for various stable structures of the 1:2 C3/benzene and 2:1 C3/benzene complexes displayed in Figure 5-8. The observed IR frequencies for C3 carbon cluster and for C3 C6H6 complex (structure A ) are given for comparison. Frequencies/cm 1 B3LYP/6-31G(d,p) Stable Structure Erel a ZPE Unscaledb Scaledc Observed Ar Matrix, 12K C3 2163.6[770] 2039.0 2039.0 A –8.25 289.6 2140.4[473] 2017.1 2034.5 O –11.5 555.0 2140.4[347] 2017.1 P –14.1 555.5 2138.0[555] 2014.9 R –14.9 314.5 2142.5[400] 2142.9[550] 2019.1 2019.5 2036.0 a Negative energies are below the sum of zer o-point-corrected ener gies of separated C3 and C6H6. Zero-point corrected energies fo r reactants (in Hartrees) are: C3; 114.037 45 and C6H6; 232.157 59. b Calculated integral intensities in km/mol are in brackets. c Frequencies scaled unifor mly by scaling factor of 0.9424. To aid in the prediction of the carrier of the 2036.0 cm 1 band, series of calculations (B3LYP/6-31G(d,p)) were pe rformed on the singlet ground states of C6H6 C3 C6H6, C3 (C6H6)2 and C3 C6H6 C3 complexes. The stable structures of these systems are displayed in Figure 5-8, with en ergy and IR frequency properties listed in Table 5-4. Following previous work on the benzene,210 three stable geometries were considered for the dimer: sandwich, T-sh aped and PD (parallel-displaced)-shaped structures. Assuming that the A type interaction between C3 and benzene was most likely, the out-of-plane approach of C3 along the six-fold axis of one benzene was considered in the initial search for the st able structures of 1:2 C3/C6H6 (or C6H6/C3) systems. The three stable structures ( O , P and R) were found to have calculate d stabilization energies of 11.5, 14.1 and 14.9 kJ/mol, respectively, lower than the sum of the reactant energies. The (unscaled) harmonic frequencies fo r the CC asymmetric stretch in the C3 unit are not very different from the computed fr equency calculated for the structure A of the 1:1

PAGE 123

105 complex. They are 2140.4 ( O) , 2138.0 ( P) and a 2142.5, 2142.9 cm 1 doublet ( R ) compared to 2140.4 cm 1 in the 1:1 complex. In the expe rimental spectrum of Figure 5-1, the 2036 cm 1 band is shifted 1.5 cm 1 to higher energy of the A complexes’ band. While this higher energy shift mimics what is calculated for R, all the calculated frequencies for the1:2 complexes are so close that to make a decisive attribution to a particular structure appears impossible at the pres ent. In conclusion, the calcul ated frequencies for the CC vibration in the C3 unit of O , P and R complexes are in accord (a fter scaling) with the experimental band at 2036 cm 1, but a specific assignment is not possible. Reaction of C3 with Ammonia Results and Discussion Ammonia readily forms complexes w ith many molecules, including such interstellar species as water,211 acetylene,212 and benzene.213 C3 has been observed to complex with water201 and benzene.214 In the reaction of C3 (X 1g +) + NH3 (X 1A1) described here, three reaction steps have been identified. First, the C3•NH3 complex is formed in solid argon at 12 K. Second, HNC3 is formed by photolysis of the complex followed by loss of H2. Finally, a UV photo-initiated hydr ogen atom migration occurs along the NC3 backbone to form cyanoacetylene, HC3N, a very stable interstellar molecule. From CEPA-1 calculations the HC3N is significant lower in energy (by 225 kJ/mol) than its HNC3 isomer.215 For each reaction step, the reactants and products were identified by observation of their infrared spectra. In addition, support for the proposed mechanism comes from ab initio MP2/6-311 ++G(d,p) theoretical cal culations. Infrared absorption spectra of selected energy re gions are displayed in Figure 5-9. Newly observed bands at 3316.4, 1902.4 and 1326.4 cm 1 are assigned to the H N symmetric stretch, C C stretch and H N H umbrella mode vibrati ons, respectively, of the C3•NH3

PAGE 124

106 complex. The weaker IR bands of the C3•NH3 complex observed in different energy regions (not displayed) are listed in Table 5-5 with th eir proposed assignments. Growth of these bands parallels the ammonia concentr ation in Ar matrices. The band assignments are supported by calculations of harmonic vibrational frequencies, including 13C isotope energy shifts. IR bands observed for the complex using 13C isotopically-labeled C3 agree well with predicted bands ( cf . Figure 5-10 and Tabl e 5-6). Generally, the 12/13C3 isotopomer experimental frequency pattern for the C C stretching mode is very sensitive to the geometry and bonding of any carbon mol ecular system, and is consistent with the predicted one only if the equilibrium geometry is correct. In the case of 12/13C3, the exp calc maximum deviation is only 1.6 cm 1. A similar 1.7 cm 1 value is found for the 12/13C3•NH3 complexes ( cf. , Table 5-6), providing str ong support for the calculated Figure 5-9. IR spectra of the C3 (X 1g +) + NH3 (X 1A1) reaction products trapped in solid Ar at 12 K and displayed in selected energy regions: without UV-visible irradiation (upper spectrum ), after 3 min. irradiation (middle spectrum) and after an additional 15 min. irradiation (low er spectrum). In the first step, the C3•NH3 complex is formed (the 3316.4, 1902.4 and 1326.4 cm 1 bands); in the second step, HNC3 is photogenerated from the complex (the (3563.6, 3562.4) and (2209.0, 2205.4) cm 1 doublet bands) and, in the third step, HC3N is photoproduced from HNC3 (the 3314.6 and 2268.6 cm 1 bands). Other unmarked bands are due to ammonia and carbon clusters.

PAGE 125

107 Table 5-5. Experimental (i n solid Ar, 12 K) and calcu lated (B3LYP/6-311++G(d,p), unscaled) vibrational frequencies for the C3•NH3 and C3•ND3 complexes. The relative integral intensities are in parentheses while absolute integral intensities (in km/mol) are in brackets. C3•NH3 C3•ND3 exp (cm 1) calc (cm 1) exp (cm 1) calc (cm 1) Assignment 3344.3 (0.29) 3523.1 [78] (0.20) 2496.3 (0.26) 2596.4 [37] (0.10) H N asymm. stretch 3316.4 (0.11) 3485.1 [72] (0.19) 2474.8 (0.23) 2561.0 [38] (0.10) H N symm. stretch 1902.4 (1.00) 1985.1 [387] (1.00) 1901.3 (1.00) 1984.8 [392] (1.00) C C stretch 1570.1 (0.28) 1633.1 [38] (0.10) 1629.4 [39] (0.10) 1147.7 (0.11) 1178.1 [17] (0.04) 1179.3 [15] (0.04) H N H bend H N H bend 1326.4 (0.44) 1361.6 [83] (0.21) 1011.9 (0.22) 1029.0 [56] (0.14) H N H umbrella 922.3 [29] (0.07) 756.0 (0.10) 764.4 [37] (0.09) C N H bend Figure 5-10. 13C-labelled IR spectra for the C-C stretching mode of the 12/13C3 carbon cluster (triangles) and the 12/13C3•NH3 complexes (dots) ( cf. Table 5-6 for band assignments). Spectra were obtained after laser ablation of a [12C]/[13C] = 3 sample (upper spectrum) and ca. [12C]/[13C]= 99 graphite sample (lower spectrum), mixed with NH3(0.4%)/Ar gas and trapped on a 12 K CsI cryostat window. The 1998.1 and 1894.4 cm 1 bands are due to the 12C9 and 12C7 carbon clusters, respectively.78,84 structure of the complex given in Figure 5-11. Moreover, each of the six isotopomers of 12/13C3 is involved in formation of the 12/13C3•NH3 complexes. Eight total isotopomer complexes are expected, since the 12C 12C 13C and 13C 13C 12C isotopomers with NH3 at both ends will generate four IR bands.

PAGE 126

108 Table 5-6. Observed (in solid Ar at 12K) and calculated IR band energies for the 12/13C3 and 12/13C3•NH3 complexes. Isotopomer exp (cm 1) calc a (cm 1) exp calc a (cm 1) calc b (cm 1) exp calc b (cm 1) 12C 12C 12C 2039.1 2039.1 0.0 2039.1 0.0 12C 12C 13C 2026.3 2026.3 0.0 2026.2 0.1 13C 12C 13C 2012.9 2012.8 0.1 2012.8 0.1 12C 13C 12C 1987.6 1986.2 1.4 1986.1 1.4 12C 13C 13C 1974.5 1972.9 1.6 1972.9 1.6 13C 13C 13C 1960.7 1959.1 1.6 1959.1 1.6 12C 12C 12C NH3 1902.4 1902.4 0.0 1902.4 0.0 12C 12C 13C NH3 1896.1 1895.7 0.4 1895.8 0.3 13C 12C 12C NH3 1884.5 1884.7 0.2 1884.4 0.1 13C 12C 13C NH3 1877.5 1877.4 0.1 1877.4 0.1 12C 13C 12C NH3 1854.8 1853.5 1.3 1853.6 1.2 12C 13C 13C NH3 1848.4 1846.8 1.6 1846.9 1.5 13C 13C 12C NH3 1836.4 1835.1 1.3 1835.0 1.4 13C 13C 13C NH3 1829.5 1827.9 1.6 1827.8 1.7 a B3LYP/6-311++G(d,p) harmonic band energi es scaled uniformly by 0.950 58 and 0.958 35 scaling factors for the 12/13C3 and 12/13C3•NH3 isotopomers, respectively. b MP2/6-311++G(d,p) harmonic band energi es scaled uniformly by 0.950 85 and 0.943 45 scaling factors for the 12/13C3 and 12/13C3•NH3 isotopomers, respectively. When comparing the 12/13C isotopomer intensity pattern in 12/13C3•NH3 vs. 12/13C3, the intensity of the 13C 12C 12C•NH3 (1884.5 cm 1) band should be added to the 12C 12C 13C•NH3 (1896.1 cm 1) band and the 12C 13C 13C•NH3 (1848.4 cm 1) band should be combined with the 13C 13C 12C•NH3 (1836.4 cm 1) one. Doing so, the 12/13C isotopomer band intensity patterns in C3•NH3 and C3 become very similar. This is further confirmation that the carrier of th e eight bands (between 1902.4 and 1829.5 cm 1, cf. Figure 5-10) is formed in an adduct reaction with 12/13C3. As a further test, deuteriumsubstituted ammonia (ND3) was used to identify the carrier of the 3316.4, 1902.4 and 1326.4 cm 1 bands of Figure 5-9 as well as other C3•NH3 bands collected in Table 5-5. Replacement of H by D led to a shift of the strongest 1902.4 cm 1 band to 1901.3 cm 1 and the 1326.4 cm 1 band to 1011.9 cm 1. Acceptable agreement with the predicted

PAGE 127

109 (B3LYP/6-311++G(d,p) values of 1901.5 cm 1 (C C stretch) and 1002.4 cm 1 (H N H umbrella mode) for the C3•ND3 complex (scale factors transferred from C3•NH3) further confirm that the carrier is C3•NH3 and that its structure is as given in Figure 5-11. Figure 5-11. Fully optimized ground state e quilibrium geometry for the singlet C3•NH3 (X 1A’, Cs) complex as calculated at the MP 2/6-311++G(d,p) level of theory. The bond lengths (in ngstroms) are: R1(C1, C2) = 1.2882, R2(C2, C3) = 1.3488, R3(C3, N4) = 1.5982, R4(N4, H5) = 1.0216, R5(N4, H6) = 1.0216, R6(N4, H7) = 1.0282. The angles (in degrees) are: (C1, C2, C3) = 174.2, (C2, C3, N4) = 104.9, (H7, N4, C3) = 113.9, (H5, N4, H6) = 107.0, (H6, N4, H7) = 109.9, (H7, N4, H5) = 109.9, (C2, C3, N4, H7) = 0.0, (H5, N4, C3, C2) = 122.3, (H6, N4, C3, C2)= 122.3. In Ar matrices, the C C stretching vibration in non-complexed C3 is observed at 2039.1 cm 1, while the H N H umbrella mode in free NH3 is located at 974 cm 1. Upon complexation, energy shifts of 136.7 and 352.4 cm 1 for the C C stretch and H N H umbrella modes, respectively, indicate large forc e constant changes. Th is is reflected in the change in the calculated atom ic charge distribution in the C3 and NH3 moieties when complexed. At equilibrium a portion of the 0.417 negative charge (at MP2/6311++G(d,p)) on the N atom flows to the C3 cluster , producing an electric dipole moment in C3•NH3 in such a way that the total charge for the ammonia unit is positive while the total charge of the C3 unit is negative. This char ge redistribution accounts in

PAGE 128

110 part for the binding of C3 to ammonia. The estimated zero point-corrected non-BSSE (non-basis set superposition error correction) binding energy of C3 to ammonia is 14.39 kJ/mol (3.44 kcal/mol) at the e quilibrium geometry for the C3•NH3 complex (displayed in Figure 5-11). Thus, in the first reaction step, the C3•NH3 complexes are formed on the cold surface of Ar/C3/NH3 film. At 12 K, the reactants ar e expected to be electronically relaxed and in their singlet electronic ground states. From the potential energy surface (PES) mapped via MP2/6-311++G (d,p) level calculations for the C3 (X 1g +) + NH3 (X 1A1) complex reaction ( cf . Figure 5-12), it was found out that complexation proceeds exothermically (without an entrance barrier) to form C3•NH3. In the second reaction step, initiate d by UV-visible irra diation, new bands (doublets) at (3563.6, 3562.4) and (2209.0, 2205.4) cm 1 appear (marked with squares in Figure 5-9). Simultaneously with the growth of these bands, the 3316.4, 1902.4 and 1326.4 cm 1 C3•NH3 bands reduced in intensity (by ~ 57%). The four former bands are assigned to HNC3 based on similar band positions re ported in Ar matrices (12 K).216 The observed doublet band structure may arise from trapping of HNC3 in two solid Ar sites. After matrix annealing (28 K) the higher energy bands in dou blets vanished while their intensities are added to the 3562.4 and 2205.4 cm 1 band intensities. The equilibrium geometry of HNC3 is bent with a predicted (CCS D(T)/cc-pVQZ) barrier height to linearity of 331.6 cm 1 (3.97 kJ/mol).217 Formation of HNC3 from C3•NH3 requires the removal of H2 (or 2H) from the complex. The calculat ed potential energy surface for this reaction ( cf. Figure 5-12) indi cates that the TSAB transition state, connecting the C3•NH3 complex with the (HNC3 + H2) products, requires 194.1 kJ /mol. Experimentally, the

PAGE 129

111 barrier determined via various filters is 221.9 kJ/mol, and is readily surmounted photolytically. Figure 5-12. Sketch of the singlet potential energy surface for the C3 (X 1g +) + NH3 (X 1A1) reaction at the MP2/6-311++G(d,p) level. The stable structures ( A E ), with connected transitio n state structures ( TS ), are displayed. The zero pointcorrected transition energies are in kJ/mol. The reaction mechanism proposed here is for the formation of the C3•NH3 complex (structure A ), HNC3, 3imino-1,2-propadienylidene (structure B ), and HC3N, cyanoacetylene (structure E ), observed in Figure 5-9. After photogeneration of HNC3 , further UV-visible photolys is, in the third reaction step, yielded new bands at 3314.6 and 2268.6 cm 1 (marked by stars in Figure 5-9) which grew concurrently with the disappearance of the HNC3 bands. These new bands are assigned to HC3N (cyanoacetylene) based on proximity to literature values and supporting vibrational fr equency calculations.215,217,218 Formation of HC3N from C3NH requires migration of the H atom to the opposite end of the C3N chain. The PES in Figure 5-12 shows that HNC3 and its HC3N photoproduct are connected through three transition states, TSBC, TSCD and TSDE, and two stable structures, C and D . The predicted energy barrier from B to TSBC is 420.4 kJ/mol; using low pass filters it has been found that the barrier is between 435.1 and 531.6 kJ/mol. Upon reaching TSBC by irradiation, the

PAGE 130

112 system can be expected to quickly drain its excess energy via the TSBC C TSCD D TSDE E pathway. With only an activatio n barrier of 61.6 kJ/mol (from C to TSCD), a low yield for both the C and D isomers might be expected after such relatively high excitation energy. This is probably the reason why C and D were not detected in the experiments. The difference between the zero point-energy corrected minimum energies of HC3N ( E ) and HNC3 ( B ) is 233.7 kJ/mol (55.8 kcal/mol, at the MP2/6-311++G(d,p) level), which is good agreement with the 225.1 kJ/mol (53.8 kcal/mol) value from CEPA1 calculations.215 Summary For the reaction between C3 and benzene, it can be c onclude that: (1)The ground state of C3 has been shown to react in low temper ature Ar and Kr ma trices with ground state benzene (with no apparent entrance barrier) to yiel d a weakly bonded complex. (2) Good agreement was found between the IR abso rption spectra predicted at the B3LYP/631G(d,p) level for the complex in whic h the plane of the benzene ring and C3 lie parallel and atop each other and the experimental spectra using 13C-substitution in the C3 portion of the complex. (3) From theo retical calculations of the singlet potential surface of C3 + C6H6, it was found that the reaction likely ends at A and/or B, under the present experimental conditions. However, in gas phase molecular beam experiments, with selected reactant energies, the F, E and H stable structures may be generated from A and/or B . (4) In addition, isomers J , M and N are predicted to form from isomer I . One of these products, N , is a doubly dehydrogenated polycyclic aromatic hydrocarbon (PAH) with an indene-like framework. The present prediction of the formation of a PAH from

PAGE 131

113 the C3•C6H6 complex, may be an important, though heretofore unexplored, route to PAHs in the interstellar medium using com ponents already known to exist in space.182,219 The reaction pathways for the formation of HNC3 and HC3N described here could be applicable to inters tellar ices containing NH3 and C3 which are exposed to weak UV radiation.220 Both products could be pa rtially released into the gas phase from hotter parts of the ice surface in desorption processes sim ilar to that proposed for other interstellar ice-trapped molecules.199 Although other low temperature ne utral-neutral reactions such as C2H2 + CN HC3N + H have been proposed previously,184,186 the present proposal is an attractive alternative sin ce both reactants and products ha ve been observed in dark molecular clouds.

PAGE 132

114 CHAPTER 6 VIBRATIONAL AND ELECTRONIC ABSORPTION SPECTROSCOPY OF DIBENZO[B,DEF]CHRYSENE AND ITS IONS Introduction Polycyclic aromatic hydrocarbons (PAHs) ar e stable compounds composed of six (and/or five)-membered carbon rings. Due to their adverse effects on human health, considerable effort has been devoted to studi es of PAHs in the an alytical, biological and medical sciences. Techniques for monitoring PAHs in the environment have been developed, and mechanisms for their in fluence on the human body have been proposed.221-224 Recently, the spectra and structure of PA Hs have undergone intense study because of their probable importance in the field of as trochemistry. This interest arose because of the suggestion by Leger and Puget, and Allama ndola and coworkers, that PAHs might be responsible for the unidentified infrar ed (UIR) interstellar emission bands.11,225 This idea has now been widely accepted, although th e mechanism of PAH formation in space remains uncertain. Indeed, PAHs are regard ed as among the most abundant free organic molecules in space. They have also been proposed, together with unsaturated carbon chains and rings, as the carriers of the DI B bands. Neutral, deh ydrogenated and ionized PAH species, including their coordination compounds, may also be present in space.225 However, to date, no individual PAH species has been positively identified in space, except for the single ring species, benzene.226 Although spectral data are available only for a limited number of PAHs, it is generally believed that the UIR (and DIB) bands can

PAGE 133

115 be attributed to emission (or absorption) of hydrocarbon compounds containing between 20 and 100 (or more) carbon atoms.11,227 Constraints due to bandwidths, anharmonicities, and isotopic shifts from laboratory data hinder further astrophys ical identification.226 Current understanding holds that the UIR bands originate from a mixture of PAHs (a PAH “soup”), which makes their indivi dual identification and/or simulation problematic.12 Dibenzo[b,def]chrysene (DBC, Dibenzo[a ,h]pyrene), the third most carcinogenic PAH, has been studied previously using a number of techniques.228-230 However, the vibrational and electronic spect ra of neutral DBC and its i ons, coupled with high level theoretical calculations, have not yet been re ported. In this chapter, the infrared and electronic spectra of neutral, cationic, and anionic DBC tra pped in solid argon at 12 K are reported. Spectral assignments are supporte d by DFT calculations of vibrational harmonic frequencies and time-depende nt electronic transition energies. Structure and Energetics Figure 6-1 presents the optimized geometry found for neutral DBC (C2h symmetry) calculated at the B3LYP/6-31G(d,p) level. DBC ions still retain C2h symmetry, with similar bond lengths and angles. Th e C–C (C–H) bond length is ~ 1.4 (1.1) , while the C–C–C and C–C–H bond a ngles are close to 120 , with deviations smaller than 5 . These parameters are typical of six-membered P AH rings. The values shown in the figure are C–H bond energies (eV) for DBC neutral [and th e cation (in brackets)], calculated at the B3LYP/6-31G(d,p) level. The smallest C– H bond energy in neutral DBC is ~ 4.7 eV, which is only 0.1 eV lower than for the other C–H bonds. The C–H bond in the DBC cation is stronger, with energies rang ing between 5.47 and 5.92 eV. The dehydrogenation

PAGE 134

116 energies of neutral or cationic DBC are typical for the H atom loss from sp2 carbons, but much higher than from sp3 carbons.231 Thus, removal of hydrogens from neutral or cationic DBC is expected to be more difficu lt than from 2,3-benzofluorene or fluorene, where hydrogen atoms were relatively easily removed from the sp3 carbon.231 Dibenzo[b,def]chrysene(DBC), C24H14 4.78 4.69 4.78 4.81 [5.92] 4.79 4.79 4.69 [5.59] [5.59] [5.87] [5.47][5.57] [5.82]Dibenzo[b,def]chrysene(DBC), C24H14 4.78 4.69 4.78 4.81 [5.92] 4.79 4.79 4.69 [5.59] [5.59] [5.87] [5.47][5.57] [5.82] Figure 6-1. Equilibrium geometry of Dibe nzo[b,def]chrysene (DBC) calculated at the B3LYP/6-31G(d,p) level. Dehydrogenati on zero-point correc ted energies (in eV, B3LYP/6-31G(d,p)) for the neutral DBC and its cation (in brackets) are displayed. Neutral DBC Infrared Absorption Spectra The experimental vibrational spectrum of neutral DBC isolated in Ar matrix at 12 K is displayed in Figure 6-2. Calculated spectra are also included for comparison. Experimental and calculated band frequencies, relative intensities, and mode descriptions, are listed in Table 6-1. To account for vibrational anharmonicity and basis set deficiencies, the B3LYP frequencie s were scaled by a factor of 0.978.232 Only those bands with relative intensities larger than 2% are tabulated.

PAGE 135

117 Table 6-1. Comparison of the calculated and experimental (Ar matrix, 12 K) infrared absorption bands (in cm 1) of neutral Dibenzo[b,def]chrysene in the 1Ag electronic ground state. Rela tive IR intensities for pr edicted and experimental frequencies are give n in parentheses. Symmetry Mode Descriptiona cal b cal c exp bu (C-C-C) 162.3 (0.02) 162.1 (0.02) au (C-C-C) 164.1 (0.05) 161.0 (0.08) bu (C-C-C) 471.4 (0.07) 470.1 (0.08) au (C-C-C) 475.6 (0.13) 470.5 (0.23) bu (C-C-C) 512.5 (0.13) 512.4 (0.12) bu (C-C-C) 635.4 (0.31) 635.2 (0.25) au (C-C-H) + (C-C-C) 675.3 (0.14) 667.6 (0.26) au (C-C-H) 738.6 (0.97) 736.0 (1.57) 740.6 (1.31) au (C-C-H) 751.5 (0.86) 748.1 (0.75) 756.4 (0.87) bu (C-C-C) 752.4 (0.06) 752.1 (0.04) bu (C-C-C) 803.5 (0.20) 802.0 (0.21) 809.6 (0.12) au (C-C-H) 842.4 (0.25) 839.4 (0.24) 844.1 (0.32) au (C-C-H) 873.4 (1.00) 873.4 (1.00) 878.8 (1.00) bu (C-C-C) 873.4 (0.07) 873.5 (0.06) au (C-C-H) 927.1 (0.09) 945.4 (0.06) 947.9 (0.06) bu R(C-C) + (C-C-C) 1031.5 (0.17) 1023.7 (0.27) 1022.7 (0.09) bu (C-C-H) 1139.6 (0.03) 1133.3 (0.05) 1135.1 (0.03) bu (C-C-H) 1158.8 (0.04) 1150.0 (0.03) 1151.0 (0.03) bu (C-C-H) 1175.5 (0.06) 1170.0 (0.05) 1171.2 (0.05) bu (C-C-H) 1205.1 (0.23) 1198.6 (0.23) 1203.1 (0.17) bu (C-C-H) + R(C-C) 1240.3 (0.05) 1232.2 (0.03) 1234.7 (0.02) bu (C-C-H) 1272.0 (0.09) 1268.1 (0.08) 1273.7 (0.04) bu (C-C-H) + R(C-C) 1315.2 (0.18) 1305.9 (0.19) 1314.6 (0.09) bu R(C-C) + (C-C-H) 1368.0 (0.17) 1355.4 (0.15) 1365.0 (0.07) bu (C-C-H) + R(C-C) 1467.6 (0.17) 1455.1 (0.10) 1463.3 (0.13) bu (C-C-H) + R(C-C) 1490.6 (0.07) 1478.3 (0.04) bu R(C-C) + (C-C-H) 1538.4 (0.08) 1523.7 (0.10) 1524.1 (0.04) bu R(C-C) + (C-C-H) 1601.8 (0.06) 1587.0 (0.06) bu R(C-C) + (C-C-H) 1611.6 (0.05) 1595.0 (0.03) bu R(C-C) + (C-C-H) 1639.7 (0.25) 1622.4 (0.16) d bu r(C-H) 3103.7 (0.08) 3088.3 (0.06) 2861.5 (0.05) bu r(C-H) 3108.4 (0.12) 3092.5 (0.07) bu r(C-H) 3112.0 (0.44) 3096.1 (0.33) bu r(C-H) 3118.3 (0.79) 3101.9 (0.63) bu r(C-H) 3132.8 (1.54) 3116.2 (1.06) 2933.3 (1.08) bu r(C-H) 3143.4 (0.34) 3127.9 (0.18) bu r(C-H) 3158.1 (1.25) 3143.5 (0.77) 2966.1 (0.55) a Notation used: R and r are stretching modes, and are in-plane be nding modes, and and are out-of-plane vibrations. b Calculation was carried out at B3LYP/6-31G( d,p) level, and all frequencies were scaled by factor of 0.978. Only bands with relative intensities equal to or higher than 0.02 are listed. The integral intensity predicted for the 873.4 cm 1 band is 49.55 km/mol. c Calculation were at the B3 LYP/6-311+G(d,p) level, and all frequencies were scaled by a factor of 0.978. Only bands with relative intensities equa l to or higher than 0.02 are listed. The integral intensity predicted for the 873.4 cm 1 band is 57.72 km/mol. d Overlapped with water bending vibration bands.

PAGE 136

118 Figure 6-2. Experimental and calculated IR absorption spectra for neutral DBC. (a) IR absorption spectrum of DBC trapped in solid argon at 12 K. (b) IR spectrum calculated at the B3LYP/6-311+G(d,p) level (scaled by 0.978). (c) IR spectrum calculated at the B3LYP/ 6-31G(d,p) level (scaled by 0.978). All observed bands could be assigned to predicted frequencies, except for two, marked with stars (in Figure 6-2). In the 700–1400 cm 1 region, band positions predicted via B3LYP/6-311+G(d,p) calcula tions fit slightly better, with less than a 10 cm 1 discrepancy. Although the predicted fre quencies (using B3LYP/6-31G(d,p)) exhibit slightly larger differences, they are reliable and accurate enough to make assignments. In the C–H stretching region (2800–3100 cm 1), however, a discrepancy of ~ 190–250 cm 1 is invariably seen. Because of the broadness of the band profiles in this region, these assignments should be considered tentative. As Table 6-1 shows, the C–C–H out-of-plane bending vibrations are stronger than other vibrational modes. The strongest experimental absorption is observed at 740.6 cm 1. Theoretically, the C–C–H out-of-plane bending vibration at 873.4 cm 1 is very close to the C–C–H in -plane bending vibration at 873.5 cm 1. But the intensity of former is much stronger than the latter. On the other hand,

PAGE 137

119 experimentally these two bands merge into the one. Two bands (1483.2 and 1040.7 cm 1, starred in Figure 6-2) probably arise from contaminants or thermal decomposition products since they do not track the DBC absorption band intensity in different experiments, and their intensities, re lative to the DBC bands, depend on sample temperature. The IR spectrum of neutral DBC was previ ously recorded usin g a linear dichroism (LD) technique in a thick polyethylene sheet.230 Most bands observed in this work were reproduced in the matrix experiments, except for two, at 769 and 976 cm 1. Theoretical calculations predict no absorption close to th ese positions, casting doub t on the origin of these peaks. There are, of course, small fre quency differences between the LD and matrix peaks. Some features could not be observed in the LD experiments because of substrate absorption, viz , the band at 1135.1 cm 1 and all higher bands ( i.e. , > 1250 cm 1). Electronic Absorption Spectra Neutral DBC has strong absorption bands in the UV-visible region ( cf. Figure 6-3). The observed and theoretical vertical excita tion energies and osci llator strengths of neutral DBC are compared in Table 6-2. Th e electronic transitions were assigned by comparing the observed absorption bands to TDDFT-calculated vertical excitation energies and oscillator strengths ( cf. Table 6-2 and Figure 6-3). The observed energy for the S1 S0 transition is 2.83 eV (438.8 nm), which is close to the predicted value of 2.6–2.9 eV. Vibrational struct ure is observed on the first electronic absorption band ( cf. the enlargement in Figure 6-3). The 433.7, 420.2, and 414.1 nm bands are spaced 268, 1009, and 1359 cm 1 from the 438.8 nm (0 0) transition, respectively. Assuming no substantia l conformational change in its S1 state, the above intervals are well-described by ground state Raman-active frequencies (scaled by 0.978).

PAGE 138

120 Table 6-2. Calculated and observe d vertical excitation energies, , and oscillator strengths, f , for neutral Dibenzo[b,def]chrysene. NWCHEMa Gaussian 98a B3LYP (AC) BLYP B3LYP Experiment State Transitionb /eV f /eV f /eV f /eV f 1Bu –1 1* 2.90 0.3461 2.61 0.2358 2.69 0.2409 2.83 (438.8)c 2.77 (448)d 0.16c [0.18]d 1Bu –1 2* 3.26 0.0050 2.94 0.0032 3.25 0.0018 1Ag –1 3* 3.62 inactive 3.02 inactive 3.61 inactive 1Ag –1 4* 3.97 inactive 3.50 inactive 3.89 inactive 1Ag –3 1* 4.15 inactive 3.69 inactive 4.06 inactive 1Ag –4 1* 4.31 inactive 3.88 inactive 4.22 inactive 1Bu –2 1* 4.35 1.4226 3.89 0.8626 4.12 1.1853 4.06 (305.5)c 4.03 (308)d 0.54c [1.00]d 1Ag –2 3* 4.50 inactive 4.03 inactive 4.49 inactive 1Bu –5 1* 4.57 0.0071 3.92 0.0480 4.57 0.0011 1Ag –6 1* 4.81 inactive 4.10 inactive 4.81 inactive 1Bu –1 5* 4.94 0.4512 4.22 0.0361 4.83 0.6432 4.68 (265.1)c 4.66 (266)d 0.029c [0.24]d 1Ag –1 6* 4.94 inactive 4.25 inactive 4.90 inactive 1Bu –2 2* 5.00 0.6111 4.53 1.0414 4.88 0.0337 4.95 (250.6)c 4.88 (254)d 0.019c [0.26]d 1Ag –3 4* 5.21 inactive 4.62 inactive 5.20 inactive 1Bu –3 3* 5.31 0.4491 4.69 0.0113 5.21 0.3694 5.19 (239)d [0.22]d 1Ag –7 1* 5.43 inactive 4.75 inactive 5.41 inactive 1Bu –3 4* 5.45 0.3118 4.77 0.0607 5.41 0.0715 1Bu –5 2* 5.61 0.0086 5.00 0.0111 5.61 0.0039 1Ag –4 2* 5.67 inactive 5.00 inactive 5.50 inactive 1Ag –6 2* 5.91 inactive ··· ··· ··· ··· a All calculations were performed using 6-31G (d,p) standard basis set. The ground-state wave function transforms as the 1Ag irreducible repres entation of the C2h point group. b The orbitals are numbered in the orde r of increasing orbital energies with –1 and 1* being the highest occupied and lowe st unoccupied orbitals, respectively. c This work, wavelengths in nanom eters are given in parentheses. d Reference 233, wavelengths in nanometers are given in parentheses, and the number listed in the oscillator strength column repr esents the relative UV absorption intensity. The 257 (56 Å4/amu), 1032 (61 Å4/amu) and 1365 cm 1 (522 Å4/amu) bands match well with the observed intervals in the S1 state and are thus cons idered part of the S1 S0 electronic transition. The abso rption peaks at 428.8 and 423.6 nm are separated by ~ 2 × 268 (531) and ~ 3 × 268 (818) cm 1 from the 0-0 transition at 438.8 nm, and are taken as the 1st and 2nd overtones of the 268 cm 1 mode. Similarly, the 391.4 nm band is probably the 1st overtone of the 1359 cm 1 mode, since there is no corresponding Raman-active vibrational mode. The two re maining intervals associated with bands at 409.7 and 387.5 are assigned as combination modes: 1619 268 + 1359 cm 1 and 3017 268 + 2 × 1359

PAGE 139

121 cm 1. The strongest absorpti on at 305.5 nm (4.06 eV) is assigned to the S7 S0 transition, and its two neighbor bands (293.1 nm and 281.6 nm), 1385 and 2778 ( 2 × 1385) cm 1 from the 305.5 nm (0 0) band, can be ascribed to the predicted 1386 cm 1 (2490 Å4/amu) Raman-active mode. The final tw o bands are close to the excitation energies of the S11 S0 and S13 S0 transitions. The transitions missing in the observed spectrum (e.g., S3 S0) are either due to symmetry restrictions or expected low oscillator strengths ( cf . Table 6-2). Figure 6-3. Electronic absorption spectrum of neutral DBC (Ar, 12 K) with band positions marked in nm. Insert spectrum shows the S1 (1Bu) S0 (1Ag) transition. The band assignment is based on the 268, 1009 and 1359 cm-1 fundamental mode energies in S1 state (see text). The calculated vertical excitation energies and oscillator strengths for the first ten excited states are independent of the protocol used, Gaussian or NWChem ( cf. Table 6-2). However, for the higher excited states, the results differ. Assignments for S11 and higher transitions should thus be regarded as tentative. The oscillator strength ( f ) for an electronic transition can be calculated from the molar absorptivity (UV ) and the full width at ha lf maximum absorption ( 2 1 ) for each

PAGE 140

122 band: 2 1 9 910 33 . 4 10 33 . 4 UV UVd f . For the same sample/matrix, the molar absorptivity can be calculated from the relative IR and UVvisible band absorbance ( A ) using IR IR UV UVA A . The IR molar absorptivity (IR ) can be obtained from the theoretical integral intensity, 2 1303 . 2 IR abI . For example, the well-isolated band at 878.8 cm 1 has a predicted (B3LYP /6-31G(d,p)) integral intensity of 50 km/mol. Sin ce the observed bandwidth ( 2 1 ) is 2 cm 1, the estimated IR is 1086 M 1 cm 1. The measured absorbance of the 878.8 cm 1 IR band is 0.0043, while the measured absorbance of the 438.8 nm band in the same matrix is 0.41, and bandwidth is 105 cm 1. Therefore, the expected molar absorptivity,UV , for the 438.8 nm band is found to be 103.5 × 103 M 1 cm 1. After summing the vibron ic band contributions ( cf. Figure 6-3), the oscillator strength ( f ) for the S1 S0 transition of neutral DBC is found to be 0.16. f values for other electronic transitions of neutral and ionic DBC were determined similarly. The error in the f values comes mainly from the accuracy of the calculated integral IR intensities. DBC Cation and Anion Infrared Absorption Spectra The infrared absorption spectra of neutral DBC, and its cations and anions, are given in Figure 6-4. Bands originating sole ly from neutral DBC are shown in the lower spectrum (c). The middle spectrum (b) shows the results after elec tron bombardment of the gas mixture (Ar / CCl4). Bands corresponding to CCl3, CCl3 +, HAr2 +, and the CCl3•Cl complex are assigned based on previous studies.234-237 The new bands are mostly due to DBC cations. Because of its high electron affinity (2.0 eV),238 CCl4 acts as an electron scavenger with the result that the electrons from i onized DBC are trapped by CCl4 and its

PAGE 141

123 by-products, such as CCl3 (2.6 eV),239 CCl2 (2.5 eV),239 Cl (3.6 eV),240. All these byproducts have higher electron affinities than DBC (1.1 eV). Almost no DBC anion bands are observed in this spectrum. Without CCl4 in the gas mixture, absorption due to DBC anions (empty triangles), cations (filled triangles), and neutrals can be seen in the upper spectrum (a). The experimental frequenc ies and theoretical values for DBC cations and anions, computed at the B3LYP/6-31G (d,p) level (and scaled by 0.978), are compared in Tables 3 and 4, respectively. Figure 6-4. Observed IR absorption spectra of neutral and ionic DBC (Ar, 12 K). Bands marked with filled triangles are assigned to DBC cations, while DBC anions are labeled with empty triangles. (a) IR spectrum of DBC neutrals, cations and anions. (b) IR spectrum of DBC neutrals and cations. (c) IR spectrum of DBC neutrals. Bands marked with filled stars (1483.2, 1040.7 cm 1) are due to impurities. Bands marked by an empty star, empty circle, filled diamond, and empty diamond are attributed to CCl3 + (1036.5 cm 1), CCl3•Cl (1019.3 cm 1, 926.5 cm 1), CCl3 (898.0 cm 1), and HAr2 + (903.3 cm 1) radicals, respectively.234-237 Like other PAH ions,231,241,242 the intensities of the C–C stretching and C–C–H inplane bending modes in the 1000–1600 cm 1 region are stronger th an other vibrational

PAGE 142

124 modes. This is very different from neutral DBC whose most intense bands lie in the C–H wagging region. No band attributable to the DBC cation is observed in the C–H stretching region, because its strongest predic ted relative intensity is only 0.02. On the other hand, DBC anions have two C–H stretc hing modes in this region with computed relative intensities of 0.31 and 0. 25. But, because of the band overlap in this region, peaks attributable solely to the an ion could not be extracted from the experimental spectrum. Table 6-3. Comparison of the calculated (B3LYP/6-31G(d,p)) and experimental (Ar matrix, 12 K) infrared absorption bands (in cm 1) of dibenzo[b,def]chrysene radical cation in the 2Bg electronic ground state. Re lative IR intensities for predicted and experimental freque ncies are given in parentheses. Symmetry Mode Descriptiona cal b exp au (C-C-H) + (C-C-C) 467.2 (0.02) bu (C-C-C) 561.2 (0.03) au (C-C-H) 757.1 (0.20) au (C-C-H) 821.5 (0.02) au (C-C-H) 863.3 (0.02) au (C-C-H) 902.9 (0.07) 910.3 (0.13) bu (C-C-H) 1186.0 (0.34) 1178.9 (0.69) bu (C-C-H) 1207.9 (0.05) bu (C-C-H) + R(C-C) 1260.3 (0.03) bu R(C-C) + (C-C-H) 1304.6 (0.04) bu R(C-C) 1352.9 (1.00) 1348.7 (1.00) bu R(C-C) 1371.6 (0.14) 1364.9 (0.19) bu (C-C-H) 1428.1 (0.12) 1421.3 (0.28) bu (C-C-H) + R(C-C) 1434.4 (0.02) bu R(C-C) + (C-C-H) 1463.6 (0.04) 1451.8 (0.08) bu R(C-C) + (C-C-H) 1498.3 (0.34) 1496.4 (0.25) bu R(C-C) + (C-C-H) 1539.1 (0.11) 1529.9 (0.11) bu R(C-C) 1563.0 (0.17) bu R(C-C) 1565.9 (0.54) 1562.8 (0.47) bu R(C-C) 1607.5 (0.20) bu r(C-H) 3151.8 (0.02) bu r(C-H) 3157.4 (0.02) bu r(C-H) 3171.4 (0.02) a Notation used: R and r are stretching modes, and are in-plane be nding modes, and and are out-of-plane vibrations. b Frequencies scaled by factor of 0.978. The bands with relative intensities equal to 0.02 or higher are listed only. The integral intensity predicated for the 1352.9 cm 1 band is equal to 500.23 km/mol.

PAGE 143

125 Table 6-4. Comparison of the calculated (B3LYP/6-31G(d,p)) and experimental (Ar matrix, 12 K) infrared absorption bands (in cm 1) of dibenzo[b,def]chrysene radical anion in the 2Au electronic ground state. Re lative IR intensities for predicted and experimental freque ncies are given in parentheses. Symmetry Mode Descriptiona cal b exp bu (C-C-C) 466.5 (0.06) au (C-C-H) 709.8 (0.05) bu (C-C-C) 746.7 (0.04) au (C-C-H) 752.6 (0.08) au (C-C-H) 794.6 (0.07) au (C-C-H) 824.2 (0.02) bu R(C-C) + (C-C-H) 986.1 (0.03) bu R(C-C) + (C-C-H) 1033.9 (0.05) bu (C-C-H) 1152.0 (0.03) bu (C-C-H) 1174.6 (0.22) 1161.6 (0.35) bu (C-C-H) 1187.8 (0.02) bu (C-C-H) + R(C-C) 1245.5 (0.02) bu R(C-C) + (C-C-H) 1289.2 (0.06) bu R(C-C) 1323.1 (0.29) 1316.8 (0.32) bu R(C-C) 1336.2 (1.00) 1327.8 (1.00) bu R(C-C) + (C-C-H) 1407.0 (0.12) 1398.0 (0.23) bu R(C-C) + (C-C-H) 1416.2 (0.02) bu R(C-C) + (C-C-H) 1450.7 (0.24) 1437.6 (0.42) bu R(C-C) + (C-C-H) 1488.4 (0.04) bu R(C-C) + (C-C-H) 1527.5 (0.04) bu R(C-C) 1558.0 (0.54) bu R(C-C) 1563.8 (0.03) bu R(C-C) 1592.9 (0.08) bu r(C-H) 3073.7 (0.02) bu r(C-H) 3082.5 (0.10) bu r(C-H) 3090.0 (0.19) bu r(C-H) 3102.7 (0.31) bu r(C-H) 3114.6 (0.06) bu r(C-H) 3128.0 (0.25) a Notation used: R and r are stretching modes, and are in-plane be nding modes, and and are out-of-plane vibrations. b Frequencies scaled by factor of 0.978. The bands with relative intensities equal to 0.02 or higher are listed only. The integral intensity predicated for the 1336.2 cm 1 band is equal to 676.86 km/mol. To show clearly the correlation between e xperimental and theore tical bands of the DBC ions, synthetic “experimental ” spectra were constructed. A synthetic band, with half Gaussian and half Lorentzian character and a 3 cm 1 bandwidth, was situated at every

PAGE 144

126 Figure 6-5. Synthetic experime ntal and calculated IR absorp tion spectra for DBC cations and anions. Experimental IR spectr a for ions were synthesized from frequencies and relative intensities observed in Ar matrix at 12 K of Figure 64. The predicted IR spectra were co mputed at B3LYP/631G(d,p) level and scaled by 0.978. Strong water absorpti on region is marked by the horizontal solid line. frequency observed for a DBC ion. Theoretical spectra were generated in the same way, with a scaling factor of 0.978. As shown in Figure 6-5, th e observed frequencies are in good overall agreement with predictions. The pr edicted strong anionic absorption at ~ 1558 cm 1 is absent in the experimental spectru m, perhaps because of the overlap with

PAGE 145

127 the strong water absorption in this range. In addition, some of the DBC anion bands are photosensitive, making DBC band assignments in this region difficult. Figure 6-6. Experimental IR spectra for DBC species (Ar, 12 K) under different photolysis time. The DBC+ and DBC bands are marked by filled and empty triangles, respectively. Notation for othe r bands is same as in Figure 6-4. It is well known that photolys is results in a decrease in ion absorption in matrices. The effects of photolysis on the DBC ions are shown in Figure 6-6. Before photolysis, the [DBC+] / [DBC] ratio is about 1.2. After 1 min photolys is, the intensity ratio rises to ~ 2. On further photolysis (5 min) the ratio increa ses to 4. After still longer UV exposure (45 min), the DBC anions have disappeared while the DBC cations remain at ~ 30% of their original intensity. Overall, DBC anions ar e more sensitive than cations to UV light. Electrons photodetached from DBC are partially scavenged by cations in the matrix (DBC + and impurities such as HAr2 +) and by species with nonzero electron affinities (DBC and neutral impurities such as OH (EA = 1.8 eV)243). Only a portion of the electrons photodetached from DBC are recaptured by DBC + radicals because of various

PAGE 146

128 competitive electron capture processes. Ov erall, the photodissociation yield of DBC is larger than for DBC +. Electronic Absorption Spectra Electronic and vibrational spectra were recorded on the same sample/matrices. Figure 6-7 shows the optical spectra and Figure 6-4 shows the corresponding infrared spectra. Calculated and obser ved electronic transition en ergies for DBC cations and anions are compared in Tables 5 and 6, re spectively. The effect of UV photolysis on DBC ion concentrations is displayed in Figure 6-8. Figure 6-7. Observed optical ab sorption spectra of neutral an d ionic DBC in Ar at 12 K. Bands marked with filled triangles are assigned to DBC cations, while DBC anions are labeled with em pty triangles. (a) Optical spectrum of DBC neutrals, cations, and anions. (b) Optical spectr um of DBC neutrals and cations. (b) Optical spectrum of DBC neutrals. The carriers of the bands marked by stars are unknown. Electron bombardment of a DBC / CCl4 / Ar mixture produced several new electronic transitions (compare Figures 7b a nd 7c) at 805.0 nm (1.54 eV), 648.3 nm (1.91 eV), 626.3 nm (1.98 eV), and 589.3 nm (2.10 eV ), all of which are here assigned to DBC

PAGE 147

129 cation transitions. The first band (1.54 eV), assigned to the D2 (2Au) D0 (2Bg) transition, is within 0.06 eV of the theoretical prediction ( cf. Table 6-5). The strong progression starting at 648.3 nm (and including bands at 626.3 and 589.3 nm) is assigned to D3 (2Au) D0 (2Bg). A similar strong absorpti on was reported by the Shida228,244 group for DBC + at 653 nm in t-butyl chloride. Gr ound state fundamental frequencies of 544 cm 1 (231 Ã…4/amu) and 1564 cm 1 (1284 Ã…4/amu), calculated at the B3LYP/631G(d,p) level, are close to the observed D3 state intervals of 542 and 1544 cm 1. Figure 6-8. Experimental opti cal spectra (corresponding to Figure 6-6 IR spectra)) for DBC species trapped in Ar at 12 K under different photolysis times. Comparison of Figure 6-7a and Figure 67b shows three new bands at 683.4 nm (1.81 eV), 506.8 nm (2.45 eV), and 499.1 nm (2.48 eV). The 683.4 nm band is assigned to the D4 (2Bg) D0 (2Au) transition of the DBC anion, while the origin of two other bands is unknown. The latte r two bands (506.8 nm and 499.1 nm) show up only when DBC anions are created. They are not sensit ive to the UV photolysis . Figure 6-8 shows that only after longtime UV matrix irradiat ion do the ratios of these two peaks vary

PAGE 148

130 slightly. As shown in Figure 6-8, the intens ities of the DBC ionic electronic absorption vary with the same pattern as the infrared inte nsities in Figure 6-6. As mentioned above, prolonged ( i.e. , 45 min) photolysis results in a D BC cation signal decrease of 70% and a complete disappearance of the DBC ani on signal. This observation supports the vibrational band assignments to the two ions made from the analysis of the infrared spectra, and vice versa . Shida previously observed a band at 442 nm (2.81 eV) in solid 2methyltetrahydrofuran (at 77 K) and assigned it to the DBC an ion. The intensity of this band was even greater than the transition at 690 nm.228,244 However, in the experiments, this band is overlapped by the strong absorp tion of neutral DBC. Although there is a predicted electronic transition [D8 (2Bg) D0 (2Au)] at ~ 2.8 eV, its oscillator strength is much smaller than for the D4 D0 transition. Thus, there is not enough information from these experiments to confirm this assignment. Astrochemistry Implications The formation yields of P AH ions are dependent on thei r ionization energies (IE) and electron affinities (EA). Experimentally, the IE of DBC lies between 6.8 eV and 7.4 eV.245,246 An adiabatic IE of 6.19 eV and a vertical IE of 6.24 eV (B3LYP/6-31G(d,p) level) have been calculated. This IE is typical of many PAHs, and indicates the high probability of DBC ionization upon exposure to UV-visible [or cosmic ray radiation]. The calculated adiabatic EA for DBC is 0. 89 eV and the vertical EA is 0.79 eV (B3LYP/6-31G(d,p)). An experimental EA of 1.1 eV has been reported.247 The DFT approach usually gives comput ed electron affinities within 0.3 eV of the experimental value.248 Diffuse functions are not really im portant in improving estimated electron affinities for large PAH systems.249 With such a high electron affinity, DBC anions

PAGE 149

131 should easily form in the elec tron-rich conditions of inters tellar space, such as the envelope of stars. The Unidentified Infrared (UIR) Emission Bands Infrared absorption band energies and in tensity distributions for DBC cations and anions are quite different from neutral D BC. Many of the infrared absorption bands observed for the DBC ions are in line with th e unidentified interstellar infrared (UIR) emission features. Figure 6-9 shows that a mixture of DBC species (26% neutral, 7% anion and 67% cation) resembles reasonabl y well the UIR bands observed from the reflection nebula NGC 7023.250,251 The strongest absorption bands at 1348.7 cm 1 (7.4 µ m) and 1327.8 cm 1 (7.5 µ m) for the DBC cation and anion, respectively, correlate reasonably well with the most intense UIR broad band at 7.7 µ m. The, 910.3 cm 1 (11.0 µ m) DBC cation band is close to the 11.0 µ m emission feature. The 1178.9 cm 1 (8.5 µ m) peak may contribute to the 8.6 µ m UIR band, while the 1562.8 cm 1 (6.4 µ m) band may contribute to the 6.2 µ m UIR peak. The DBC anion absorption at 1161.6 cm 1 (8.6 µ m) may account for the 8.6 µ m emission feature. The 1316 cm 1 (7.6 µ m), 1398.0 cm 1 (7.2 µ m), and 1437.6 cm 1 (7.0 µ m) bands fit within the broa d UIR envelope between 6.2 µ m and 7.6 µ m. The predicted band at 1558.0 cm 1 (6.4 µ m) may correspond to the 6.2 µ m UIR feature. While the DBC IR bands fall in close proximity to the UIR bands, the values are, of course, from Ar matrices and thus subject to some shifts compared to gas phase values. What are the magnitudes of the expected shifts? Recently, IR spectra in the C H stretching region were reported252 for a few small PAHs seeded into supersonicallyexpanded carrier gases with low rotational ( 10 K) and vibrational (<50 K) temperatures. These results show that red shifts averaging 0.5–4.4 cm 1, going from the gas phase to

PAGE 150

132 Figure 6-9. Comparison of UIR bands from the reflection nebula NGC 7023 with a mixture of DBC species (26% neutral, 7% anion and 67% cation). a) UIR bands from NGC 7023 (adapted, with pe rmission, from refs. 250 and 251); b) Mixture of DBC species ( i.e. , sum of spectra c-e); c) DBC cation IR spectrum; d) DBC anion IR spectrum; e) neutra l DBC IR spectrum. Spectra c-e were constructed from the experimental spect rum of each species (neutral, cation or anion) in the 600–2000 cm 1 range, by increasing every peak bandwidth (FWHM) to 75 cm 1. The latter value was chosen since it is the experimental bandwidth of the most intense UIR band at 7.7 µ m (1298 cm 1). Ar matrices, can be expected . Based on Jacox’s compilation253 of gas-to-Ar matrix shifts for numerous neutral and ionic species, drama tic changes in the absorption energies of PAH ions in the mid-IR region are not expe cted with change of phase. Thus, the Ar

PAGE 151

133 matrix-to-gas shifts are expected to be too sm all to affect the two si gnificant digits quoted in the UIR band positions listed above. In summary, a mixture of DBC neutrals, ca tions, and anions does a passable job of mimicking the UIR features of NGC 7023, wh ich lends credence to the conclusion that such a mixture may contribute to the UIR emission features of this and other UIR sources. The Diffuse Interstellar Bands (DIBs) PAHs and their ions have also been pr oposed as possible carriers of the visible diffuse interstellar ab sorption bands (DIBs).254 A number of large PAHs, such as dicoronylene (C48H20), have been studied and compared with the DIB spectra.254 Although neutral DBC possesses no electronic transitions that match the DIB bands, its ions do. The most intense DCB cation elec tronic band at 648.3 nm (Ar/12K) and anion band at 683.4 nm (Ar/12K) are in close proxi mity to the DIB bands at 649.4 nm and 683.4 nm, respectively, observed from HD 183143.17 Matrix-to-gas shifts must be considered. The electronic abso rption bands of PAH cations us ually red-shift in matrices, compared to the absorption in the gas phase. For example, the shift is 120 cm 1 for naphthalene,255 218 cm 1 for acenaphthene,255 393 cm 1 for pyrene,255 289 cm 1 for fluorene,256 and 266 cm 1 for anthracene.257 Adopting an average gas-to-Ar matrix shift of 250 cm 1, the electronic bands for DBC cations and anions should appear in the gas phase at 638.0 nm and 676.5 nm, respectively. Compared to the DIB bands measured in the interstellar source HD 183143,17 DBC cations may thus c ontribute to the diffuse bands at 636.7, 637.6, and 637.9 nm, while DBC anions could be the carrier of the DIB bands at 676.8, 676.9, 677.0, and 677.9 nm. Ther efore, the DBC ions should be viewed as potential contributors to the DI B bands as well as the UIR bands.

PAGE 152

134 Summary (1) The vibrational and electronic spectra of neutral DBC were recorded in argon matrices at 12 K. Band assignments were made by comparison with theoretical calculations. Out-of-plane C–C–H be nding vibrations at 740.6 and 878.8 cm 1 were among the strongest modes observed. Five elec tronic transitions have been observed for neutral DBC, with the first [assigned to the S1 (1Bu) S0 (1Ag) transition] lying at 438.8 nm, and the strongest [assigned to the S7 (1Bu) S0 (1Ag) transition] locat ed at 305.5 nm. Vibrational structure for these two bands were also observed and assigned. (2) Infrared and optical spectra of the DBC cation and anion are reported for the first time. The electronic and vibrational spectra were recorded on the same matrix, allowing an estimate of the oscillator strength s of the observed electronic transitions. The IR band assignments were supported not onl y by theoretical pred ictions, but also by intensity correlation wi th visible band intens ities. The strongest IR bands of the DBC cations and anions lie at 1348.7 cm 1 and 1327.8 cm 1, respectively. The electronic band at 648.3 nm has been assi gned to the DBC cation D3 (2Au) D0 (2Bg) transition, and the band at 683.4 nm to the DBC anion D4 (2Bg) D0 (2Au) transition. The photodissociation rate for DBC ani ons is higher than for cations. (3) Comparison of the observed infrared a nd visible bands of the DBC cations and anions with the interstellar UIR and DI B bands reveals that DBC ions may be contributors to both sets of bands.

PAGE 153

135 CHAPTER 7 ANHARMONICITY OF POLYCYCLIC AROMATIC HYDROCARBONS Introduction Polycyclic aromatic hydrocarbons (PAHs), as a group, are possible carriers of the UIR bands and the DIBs. They have been ex tensively studied both experimentally and theoretically.38,258-270 Specific assignments of particular PAHs have not been made for the UIR or DIB bands due to two limitations, that is, the limited data available for the PAHs and the constraint of anharmonic ity from laboratory observations.226,258,259 The anharmonicity of the PAH vibrations is subs tantial, especially for the CH stretching modes, which broadens this absorption regi on and makes the assignment for these modes tentative.258,259 No experimental method to date can be used to entirely evaluate the anharmonic effect. Recently, jet-cooled infrared cavity ringdown spectroscopy (CRDS)252 and infrared multiphoton dissociation (IRMPD)260 spectroscopy have been used to investigate this effect to some extent. On the other ha nd, theoretical calculati ons are an alternative way to study the anharmonici ty without the influence of band overlapping that is common in spectral observation. However, anharmonic frequency calculations require complicated mathematical models and entail a high computational cost. Earlier trials to evaluate the anharmonic force field for naphthalene were based on a simple anharmonic local mode or tight-bin ding model, coupled with an adiabatic switching procedure, instead of ab initio or DFT calculations.261 The latest release of Gaussian 03 protocol not only improves the e fficiency of all computations but it also

PAGE 154

136 incorporates anharmonic frequency calcula tions without any limit on the number of atoms, a common defect for other computational programs.25 Recently, the DFT functionals in Gaussian 03 have been used to derive the anharmonic force field of uracil and 2-thiouracil. Satisfactor y agreement has been observed between calculated and matrix experimental fundamentals using the B3LYP functional, but larger errors were observed using the BLYP method.262 The harmonic frequencies of numerous aromatic compounds like naphthalene, phenanthrene, a nd anthracene have also been calculated using B3LYP functional over the decades, and in all cases, they have been shown to be quite reliable in the 700–1600 cm 1 region.263 Therefore, it seems reasonable to expect that the B3LYP density functi onal should produce reliable resu lts in the calculation of the vibrational anharmonicity of aromatic hydrocarbons. Anharmonicity of Naphthalene Harmonic and anharmonic vibrational frequencies for neutral and cationic naphthalene were computed at the B3LYP level of theo ry using 6-31G(d,p) and 6311G(d,p) basis sets. The geometry of naphthalene was restricted to D2h. Calculations using other basis sets, such as 4-31G, 6-31+G(d,p), 6-31G++G(d,p), 6-311+G(d,p), 6311++G(d,p) were carried out as well. The resu lts using the 4-31G ba sis set are close to the experimental values, but worse than 631G(d,p) or higher basi s sets. When diffuse functions (+ or ++) are used with 6-31 G(d,p) and 6-311G(d,p) basis sets, harmonic frequencies can be predicted correctly, but se veral vibrational modes were predicted with negative anharmonic frequencies, which is nonp hysical. The reason for this is due to the instability of perturbation of wavefunctions in Gaussian 03 th at lead to a mistake in the optimization of the anharmonic potential surf ace. Fortunately, previous studies have

PAGE 155

137 revealed that diffuse functions are not impor tant for large size organic molecules like PAHs.249 Therefore, no diffuse functions were used in the present calculations. Harmonic ( h) and anharmonic ( a) vibrational frequencie s and their relative intensities (IR) are reported and compared to experimental values ( e) in Table 7-1 through Table 7-4. A severe limitation of Gau ssian 03 is that absolute intensities of anharmonic frequencies can not be evaluated yet. So the intensities of the anharmonic frequencies were considered to be the sa me as harmonic ones. Harmonic frequencies at the 6-31G(d,p) and 6-311G(d,p) levels we re scaled by 0.978 and 0.985 respectively, while there is no scaling for anharmonic frequenc ies. Table 7-1 lists all the infrared active frequencies of neutral naphthalene and shoul d be compared to the experimental data observed by Hudgins et al. and Szczepanski et al.266,267 Anharmonic frequencies of naphthalene predicted by Parneix’s gr oup (VPB) using a tight-binding molecular dynamics model are reported as well for comparison.261 Raman active vibrational frequencies are presented in Table 7-2 a nd compared to the Raman shift frequency standards by McCreery fo r naphthalene pellets.271 Predicted IR frequencies of naphthalene cation are comp ared to experimental va lues by Hudgins et al.267 and Szczepanski et al.264 in Table 7-3, while theoretical Raman frequencies are compared to results from Sheng et al.272 and Szczepanski et al.273 in Table 7-4. Infrared Active Vibrational Absorption From Table 7-1 and Table 7-3, it is clear that IR activ e vibrational fundamentals of neutral and cationic naphthalene can be divided into two regions. Region A, from 400 to 1700 cm 1, is basically due to aromatic ring stru cture deformation, C–C stretching, and CH out-of-plane vibrations and region B, above 3000 cm 1, is normally due to C–H stretching. The scaling factor for harm onic frequencies of neutral and cationic

PAGE 156

138 Table 7-1. Comparison of the calculated (B 3LYP) and experimental infrared active absorption bands (in cm 1) including their relative IR intensities of neutral naphthalene in the 1Ag electronic ground state 6-31G(d,p) 6-311G(d,p) Hudgins et al.267 Szczepanski et al.266 VPB261 irrep h a IR h a IR e IR e IR a au 967.3 975.5 0.00 977.5 979.1 0.00 au 834.8 838.2 0.00 836.5 829.6 0.00 au 619.8 625.6 0.00 625.2 626.7 0.00 au 185.4 184.5 0.00 182.9 180.2 0.00 b1u 3122.7 3055.7 0.74 3128.7 3039.0 0.56 3050-3071 0.21 3028-3112 0.37 2976 b1u 3105.2 3074.0 0.08 3111.3 3041.1 0.05 2960 b1u 1619.7 1618.9 0.05 1617.5 1605.3 0.04 1599-1604 0.04 1601 0.05 1645 b1u 1393.6 1397.0 0.04 1397.0 1390.0 0. 04 1388-1395 1391, 1394 0.05 1532 b1u 1263.3 1280.3 0.09 1268.2 1276.6 0. 06 1267-1275 0.07 1269, 1272 0.05 1355 b1u 1129.2 1138.8 0.04 1134.0 1134.8 0. 04 1127-1142 0.07 1129, 1131 0.05 1202 b1u 789.8 798.8 0.00 796.9 799.2 0.00 863 b1u 356.1 364.4 0.01 360.1 365.6 0.01 379 b2u 3134.3 3088.0 0.61 3140.0 3062.8 0.43 3071-3083 0.09 3028-3117 0.37 2984 b2u 3107.0 3037.3 0.01 3113.6 3028.7 0.01 2976 b2u 1528.0 1526.1 0.09 1526.4 1513.6 0. 08 1506-1520 0.12 1513, 1515 0.07 1589 b2u 1377.1 1379.8 0.01 1371.2 1365.1 0.01 1360-1363 0.02 1361 0.02 1371 b2u 1213.6 1223.0 0.01 1214.6 1214.5 0. 01 1209-1216 0.03 1212, 1214 0.02 1250 b2u 1152.2 1161.4 0.01 1152.0 1152.1 0.01 1145 b2u 1020.0 1025.9 0.06 1020.8 1018.7 0. 06 1009-1018 0.07 1012, 1016 0.07 1108 b2u 621.0 629.0 0.04 626.7 630.1 0.03 617-622 0.04 621 0.04 653 b3u 947.1 960.0 0.04 959.6 965.2 0. 03 956-963 0.02 958, 960 0.03 1057 b3u 785.7 789.9 1.00 786.2 787.1 1. 00 780-790 1.00 783, 788 1.00 823 b3u 481.1 483.7 0.14 480.5 478.8 0.19 472-483 0.15 474 0.17 484 b3u 172.5 173.5 0.02 170.2 171.3 0.02 169 Table 7-2. Comparison of the calculated (B3LYP) and experimental Raman active absorption bands (in cm 1) including their relative Raman intensities of neutral naphth alene in the 1Ag electronic ground state 6-31G(d,p) 6-311G(d,p) McCreery271 irrep h a IR h a IR e IR ag 3135.4 3067.7 1.00 3141.2 3048.9 1.00 3056.4 0.32 ag 3110.4 3046.3 0.42 3117.0 3029.5 0.42 ag 1592.7 1591.7 0.06 1590.7 1577.1 0.05 1576.6 0.16 ag 1469.5 1475.1 0.01 1468.6 1466.4 0.10 1464.5 0.12 ag 1385.0 1382.5 0.24 1379.0 1366.3 0.25 1382.2 1.00 ag 1163.1 1177.0 0.01 1167.5 1171.4 0.00 ag 1030.4 1040.1 0.05 1031.2 1033.2 0.05 1021.6 0.11 ag 759.1 764.7 0.06 762.1 761.7 0.06 763.8 0.30 ag 509.1 516.2 0.03 512.4 515.7 0.03 513.8 0.29 b1g 930.8 942.7 0.00 942.5 946.0 0.00 b1g 718.7 723.0 0.01 717.9 716.4 0.00 b1g 388.2 389.4 0.01 389.5 387.2 0.00 b2g 975.1 979.5 0.00 984.7 980.9 0.00 b2g 879.4 884.4 0.01 883.3 882.5 0.00 b2g 767.1 774.2 0.01 774.9 773.7 0.00 b2g 470.5 473.7 0.00 472.0 471.8 0.00 b3g 3121.5 3054.3 0.30 3127.4 3044.5 0.29 b3g 3103.5 2980.1 0.03 3109.5 2997.1 0.03 b3g 1649.2 1644.0 0.02 1647.2 1630.1 0.01 b3g 1468.5 1473.0 0.11 1469.8 1464.1 0.01 b3g 1244.8 1252.5 0.01 1251.3 1247.3 0.01 b3g 1151.4 1159.6 0.00 1154.5 1154.7 0.01 1147.2 0.06 b3g 927.1 937.4 0.00 936.9 939.3 0.00 b3g 506.4 512.6 0.01 511.3 513.2 0.01

PAGE 157

139 Table 7-3. Comparison of the calculated (B 3LYP) and experimental infrared active absorption bands (in cm 1) including their relative IR intensities of cationic naphthalene in the 2Au electronic ground state 6-31G(d,p) 6-311G(d,p) Hudgins et al.267 Szczepanski et al.264 irrep h a IR h a IR e IR e IR au 1001.9 1008.1 0.00 1013.0 1014.3 0.00 au 857.6 857.9 0.00 863.7 854.8 0.00 au 550.5 550.5 0.00 553.2 551.2 0.00 au 179.7 178.3 0.00 179.0 176.3 0.00 b1u 3151.7 3099.8 0.00 3154.5 3081.4 0.00 b1u 3136.9 3080.0 0.00 3141.4 3064.0 0.00 b1u 1525.4 1531.7 0.41 1526.3 1522. 4 0.40 1519 0.10 1519 0.08 b1u 1405.6 1405.8 0.11 1411.6 1402. 2 0.11 1401 0.04 1401 0.04 b1u 1278.6 1287.0 0.05 1284.7 1283.4 0.04 b1u 1102.0 1109.8 0.02 1107.3 1106.2 0.02 b1u 790.2 797.4 0.00 797.4 798.4 0.00 b1u 349.8 358.0 0.00 354.0 359.4 0.00 b2u 3161.6 3116.7 0.01 3164.4 3097.3 0.01 b2u 3137.4 3075.1 0.00 3142.2 3057.4 0.00 b2u 1544.0 1540.4 0.12 1544.4 1530. 9 0.11 1526 0.29 1525 0.16 b2u 1398.0 1398.8 0.10 1393.0 1387.8 0.09 b2u 1214.5 1221.0 1.00 1216.2 1213.7 1.00 1215 1218 0.20 1.00 1218 1.00 b2u 1169.7 1181.1 0.08 1174.9 1176.7 0.07 b2u 1020.2 1032.7 0.07 1022.0 1027.2 0.06 1016 1023 0.05 1016 1023 0.20 0.06 b2u 595.2 601.8 0.04 600.9 603.2 0.04 b3u 980.1 985.1 0.01 989.5 985.5 0.01 b3u 761.8 765.2 0.42 767.3 766. 5 0.48 758.7 0.27 b3u 419.0 420.6 0.08 420.3 417.8 0.11 b3u 156.1 156.7 0.01 155.0 156.0 0.01 Table 7-4. Comparison of the calculated (B3LYP) and experimental Raman active absorption bands (in cm 1) including their relative Raman intensities of cationic naphthalene in the 2Au electronic ground state 6-31G(d,p) 6-311G(d,p) Sheng et al.272 Szczepanski et al.273 irrep h a IR h a IR e e ag 3161.7 3093.5 1.00 3164.5 3075.7 1.00 ag 3139.4 3078.2 0.44 3144.2 3060.6 0.43 ag 1598.6 1586.2 0.02 1599.2 1574.1 0.02 ag 1475.8 1478.7 0.02 1477.5 1470.9 0.02 ag 1388.9 1381.8 0.32 1384.4 1367.6 0.33 1398 1398 ag 1180.7 1194.5 0.01 1187.2 1189.1 0.01 ag 1048.1 1053.8 0.05 1050.0 1048.2 0.05 ag 760.1 764.6 0.08 763.3 761.5 0.09 769 766 ag 503.8 510.0 0.10 507.1 509.5 0.10 511 507 b1g 957.7 964.2 0.01 969.1 964.7 0.00 b1g 738.5 741.4 0.00 744.3 740.8 0.00 b1g 369.2 367.7 0.00 370.0 365.8 0.00 b2g 1005.7 1010.5 0.00 1014.3 1017.8 0.00 b2g 918.8 924.8 0.00 922.7 930.1 0.00 b2g 723.5 739.3 0.01 727.6 759.1 0.00 b2g 430.7 431.7 0.00 431.6 431.2 0.00 b3g 3151.5 3093.3 0.36 3154.3 3075.6 0.34 b3g 3135.0 3079.2 0.00 3139.4 3062.4 0.00 b3g 1484.1 1496.1 0.16 1484.6 1488.7 0.14 b3g 1441.7 1447.5 0.12 1446.7 1443.2 0.11 b3g 1234.8 1245.4 0.01 1241.2 1241.7 0.01 b3g 1098.7 1110.6 0.04 1101.4 1101.7 0.05 b3g 919.2 925.9 0.02 928.4 927.7 0.02 b3g 462.2 467.7 0.00 465.9 468.1 0.00

PAGE 158

140 Figure 7-1. Synthetic experi mental and calculated IR ab sorption spectra for neutral naphthalene. Experimental IR spectrum was synthesized from frequencies and relative intensities observe d by Szczepanski et al.266 The predicted anharmonic IR spectra were computed at B3LYP/6-31G(d,p) and B3LYP/6311G(d,p) level w ithout scaling. Figure 7-2. Difference ( h a) of harmonic and anharmonic frequencies as a function of IR active vibrational modes of neutral naphthalene. Modes (a, b Â… x) refer to the fundamental bands listed in Table 71 from the first row to the last row respectively.

PAGE 159

141 Figure 7-3. Synthetic experi mental and calculated IR ab sorption spectra for cationic naphthalene. Experimental IR spectrum was synthesized from frequencies and relative intensities obser ved by Hudgins et al.267 The predicted anharmonic IR spectra were computed at B3LYP/631G(d,p) and B3LYP/ 6-311G(d,p) level without scaling. Figure 7-4. Difference ( h a) of harmonic and anharmonic frequencies as a function of IR active vibrational modes of cationic naphthalene. Modes (a, b Â… x) refer to the fundamental bands listed in Table 73 from the first row to the last row respectively.

PAGE 160

142 naphthalene was chosen to obt ain the best fit of theoreti cal harmonic frequencies to experimental values within region A. For neutral naphthalene, anharmonic fre quencies without sc aling in region A have a satisfactory agreement with experimental obser vation, which is shown in Figure 7-1. No matter which level of theory was performed, the predicted spectra matched well with experimental one. The deviation between th eoretical and experimental frequencies is always below 15 cm 1. The higher 6-311G(d,p) basis se t might yield slightly more accurate band positions and intensities, but the 6-31G(d,p) basis set is large enough to capture the spectroscopic features of neutra l naphthalene, which is consistent with previous results.263 From the comparison shown in Figure 7-2, it is clear that scaled harmonic frequencies in region A are similar to anharmonic fre quencies without scaling, which indicates that anharmonicity for neutral naphthalene in region A is similar and harmonic frequencies can be easily calib rated by a uniform factor to offset the anharmonic effect. The scaling factor depends on the basis sets , for 6-31G(d,p), 0.978 factor was used for anharmonicity correction. When examining region B (modes e, f, m, and n) in Figure 7-2, harmonic frequencies still overestimate the vibrational energies even after scaling, indicating a much stronger anha rmonic effect for these modes. Due to strong anharmonic interaction in region B, absorption bands are always broad and overlapped with each other, and thus difficult to make assignment s. Since anharmonic calculations in region A matched well with experiments, agreement can be reasonably expected as well for region B. It can be seen in Figure 7-2 that basis se ts also affect the difference between harmonic and anharmonic frequencies. For the 6-31G(d,p ) basis set, the difference varies from 30

PAGE 161

143 to 70 cm 1, while the difference is kept at around 75 cm 1 for 6-311G(d,p) basis set. Overall, a different scaling factor has to be considered for harmonic frequencies of C–H stretching modes. The anharmonicity for these modes is around 3.5–4.5% of their harmonic energy level, but is also associated with the level of theory. In addition, anharmonic analysis using DFT theory yields more precise results than using other theoretical models proposed by Parneix’s group (VPB).261 Experimental observation for the naphthalen e cation is more complicated than for neutral naphthalene, but the observed absorption band pos itions are still close to the computed anharmonic frequencies, as presente d in Figure 7-3. Additionally, both levels of theory fail to predict their relative intensities. Some bands are missing in the experimental spectrum, which may be due to the instrumental det ection limit and overlap of different bands. With an appropriate sc aling factor, the harm onic frequencies are similar to the anharmonic ones in Region A ( cf. , Figure 7-4). The di fference is always within ±20 cm 1, while a bigger gap still exists for region B. However, since the intensities for the absorption bands in region B are very low, it is not really critical to evaluate their anharmonicity for naphthalen e cation. Therefore, for the naphthalene cation, a harmonic calculation with a proper scaling factor will be sufficient to investigate its spectroscopic properties, with modera te CPU time consumption compared to an anharmonic analysis with its much higher computer cost. Raman Active Vibrational Absorption Similar to infrared active vibrational fre quencies, for vibrational modes below the 1700 cm 1 region, both anharmonic frequencies and harmonic frequencies after scaling of neutral and cationic naphthalene ma tch the experimental data well ( cf. , Table 7-2 and Table 7-4). Around the 3000 cm 1 region, anharmonic frequencies, but not harmonic

PAGE 162

144 data, are in concert with the experimental band positions of neutral naphthalene. There are no data available for the naphthalene cation in this region. More Raman data will be necessary in order to evaluate the efficien cy of the anharmonic computation. For both regions, the anharmonic frequencies for naphtha lene differ from experimental values by less than 20 cm 1. However, harmonic calculations at both levels were unsuccessful in predicting the relative band intensities for three reasons. First, the experimental data were recorded in the solid phase, but theoretical values were calculated for the gas phase. Second, anharmonicity may have a strong e ffect on the band inte nsity distribution. Finally, there is the limitation of theoretical model, like the basis set and functional used. An anharmonic frequency calculation will improve the accuracy of the predicted Raman band positions, but more data are needed to test the reliability of the calculation. Anharmonicity of Phenanthrene and Anthracene A calculation of the anharmonic frequencies for phenanthrene and anthracene was carried out at the B3LYP/631G(d,p) level of theory. Since the computation for naphthalene revealed that B3 LYP/6-31G(d,p) is large enough to predict th e vibrational properties of PAHs, no attempt at using a hi gher level of theory for phenanthrene and anthracene was attempted. Harmonic (scale d by 0.978) and anharmonic frequencies of neutral and cationic phenanthrene together with relative intensities are tabulated and compared with the experimental data by Hudgins et al.268,270 in Table 7-5, while Table 76 presents the corresponding values for anth racene neutral and cati on. Neutral anthracene was observed in argon matrix by Szczepanski et al.,265 and anthracene cation was investigated by bot h Hudgins et al.269 and Szczepanski et al.265 From Table 7-5 and Table 7-6, it can be concluded that: 1) the absorption band positions match well with the anharmonic vibrational frequencies; 2) the difference

PAGE 163

145 Table 7-5. Comparison of the calculated (B3L YP/6-31G(d,p)) and experimental infrared active absorption bands (in cm 1) including their relative IR intensities of neutral and cationic phenanthren e in the electronic ground state Neutral Phenanthrene Cationic Phenanthrene B3LYP/6-31G(d,p) Hudgins et al.270 B3LYP/6-31G(d,p) Hudgins et al.268 h a IR e IR h e IR e IR 100.4 101.9 0.01 100.5 102.0 0.01 228.2 229.3 0.04 211.9 212.9 0.03 243.2 247.5 0.01 405.3 408.0 0.04 402.7 409.3 0.01 482.0 491.2 0.01 430.2 433.8 0.07 584. 5 592.4 0.21 582.0 0.22 436.7 443.8 0.03 688. 4 702.7 0.13 694.5 0.09 494.2 504.7 0.01 494.0 0.04 757.1 762.5 0.21 756.2 0.07 498.3 502.7 0.05 836. 3 841.0 0.19 836.0 0.06 545.0 553.1 0.01 884.9 891.8 0.03 617.6 624.7 0.09 617.5, 618.8 0.06 859.4 866.2 0.03 709.4 720.2 0.03 965.8 975.9 0.01 710.9 719.4 0.05 710.0, 714.5 0.04 981.8 990.1 0.11 735.6 743.7 1.00 735.0 1. 00 1039.9 1050.6 0.03 813.7 822.5 0.78 812.8 0. 69 1044.0 1049.2 0.02 824.1 836.1 0.00 833.0 0. 02 1135.3 1150.1 0.01 865.0 873.0 0.16 864.9 0. 12 1141.8 1153.9 0.01 869.9 879.7 0.03 877.6 0. 02 1144.4 1154.6 0.68 934.4 952.4 0.05 948.2 0. 03 1178.0 1193.0 0.01 994.1 1005.6 0.02 1002.5, 1003.7 0.02 1219.1 1224.1 0.06 1043.1 1051.7 0.08 1039.9, 1044.3 0.06 1226.1 1233.3 0.26 1258.7 0.06 1046.2 1054.5 0.01 1261. 4 1267.6 0.17 1264.7 0.02 1094.5 1103.6 0.02 1094.5, 1095.9 0.02 1267.0 0.16 1144.8 1154.7 0.02 1144.1 0.03 1277.5, 1282.5 1.00 1167.3 1182.2 0.01 1165.4 0. 02 1286.8 1298.4 0.02 1207.4 1215.4 0.04 1202.7 0.03 1323.3 1316.6 1.00 1299.0 0.11 1220.5 1227.0 0.01 1223.8 0. 01 1423.5 1428.9 0.01 1245.5 1258.9 0.17 1245.8, 1250.6 0.11 1432.5 1439.4 0.24 1303.5 1308.7 0.03 1302.9 0. 04 1438.7 1443.0 0.19 1358.5 1357.4 0.02 1351.4 0.01 1514.9 1511.7 0.04 1513.0 0.02 1431.5 1434.5 0.03 1419.0 0. 01 1529.2 1527.3 0.03 1427.5 0.01 1532.1 1526.9 0.05 1430.7 0.02 1558.1 1555.8 0.52 1551.0 0.07 1449.0 1453.3 0.05 1436.2 0. 01 1558.2 0.03 1447.9 0.01 1577.7 1577.4 0.93 1565.0 0.54 1467.6 1471.1 0.22 1460.4 0. 14 1613.2 1570.4 0.10 1513.8 1515.0 0.10 1504.7, 1505.9 0.09 3161.1 3100.8 0.01 1541.4 1540.4 0.03 1530.1, 1531.8 0.02 3164.9 3097.0 0.01 1622.2 1622.5 0.07 1597.9 0.01 1602.8 0.01 1633.6 1630.8 0.01 1639.1 1636.9 0.01 3106.4 3034.8 0.07 3107.1 3050.0 0.04 3117.2 3058.5 0.25 3122.7 3055.5 0.68 3129.7 3069.1 0.78 3133.0 3075.2 0.02 3139.4 3079.0 0.38 3149.8 3078.1 0.46

PAGE 164

146 Table 7-6. Comparison of the calculated (B3L YP/6-31G(d,p)) and experimental infrared active absorption bands (in cm 1) including their relative IR intensities of neutral and cationic anthracene in the electronic ground state Neutral Anthracene Cationic Anthracene B3LYP/6-31G(d,p) Szczepanski et al.265 B3LYP/6-31G(d,p) Hudgins et al.269 Szczepanski et al.265 h a IR e IR h e IR e IR e IR 91.6 92.9 0.01 86.3 86.3 0.01 230.2 234.4 0.02 438.0 439.2 0.08 432 0.07 474.0 475.9 0.22 468, 470 0.29 596.0 604.6 0.01 603.9 609.8 0.14 603 0.14 751.6 755.6 0.25 748.3 0.26 646.6 654.2 0.01 817.5 823.8 0.01 729.9 735.1 1.00 726, 729 1.00 908. 9 914.0 0.10 912.0 0.09 912 0.15 874.8 884.0 0.89 878.5 0.68 975.0 981.3 0.02 896.1 907.3 0.03 908 0.02 1034. 1 1053.9 0.02 1030 0.20 945.9 958.3 0.10 955, 958 0.07 1191. 7 1200.4 0.54 1183.3 0.01 1188 0.98 1011.4 1018.5 0.07 1001 0.07 1188.6 0.70 1143.6 1158.0 0.03 1149, 1151 0.04 1281. 4 1292.6 0.06 1290.4 0.06 1291 0.07 1152.0 1162.1 0.08 1294.3 1298.0 0.03 1314.6 0.06 1169.5 1178.5 0.0 1167, 1169 0.03 1360. 9 1358.5 1.00 1341.0 1.00 1341 1.00 1264.9 1278.7 0.13 1272 0.05 1352.6 0.31 1316.1 1326.5 0.08 1318 0.12 1364.4 0.04 1361.0 1371.7 0.06 1346 0.01 1414.9 1419.0 0.50 1406.1 0.02 1410 0.09 1398.7 1396.6 0.01 1409.5 0.11 1461.1 1463.0 0.03 1450 0.05 1418.4 0.86 1464.0 1468.8 0.02 1460 0.04 1430.2 0.01 1559.0 1554.3 0.08 1540, 1542 0.04 1462. 1 1465.8 0.19 1456.5 0.07 1457 0.05 1648.2 1649.4 0.11 1627 0.12 1459. 8 1468.3 0.07 3101.1 3022.2 0.15 3017, 3022 0.06 1548. 3 1542.3 0.28 1539.9 0.15 1540 0.04 3105.8 3031.7 0.23 3032 0.05 1592.0 1591.1 0.19 1586.4 0.14 3122.9 3052.5 1.17 3055, 3062 0.25 3134.6 3063.7 1.18 3067-3068 0.43 between anharmonic frequencies and ha rmonic frequencies for modes under 1700 cm 1 is less than 15 cm 1, so that the anharmonicity fo r these modes is all equal to ca. 2% of their vibrational energies; and 3) stronger anharmonicity can be seen for C–H stretching modes, but the magnitude is non-uniform. So scaling factors for these vibrations should be considered separately. Phenanthrene and anthracene cations have no observable absorption in this region. Therefore, a nharmonic calculations will not be necessary. Finally, relative intensities for neutral phe nanthrene and anthracene agree well with DFT predictions, but the DFT approach does not succeed in predicting the correct relative intensities for cations.

PAGE 165

147 Summary An anharmonic vibrational analysis has been performed for neutral and cationic naphthalene, phenanthrene and anthracene. The results reveal that anharmonic frequency calculations are important and necessary for neutral PAHs, especially when C–H stretching modes are considered. Harmonic analys is with an appropriate scaling factor is precise enough to predict observable infrared absorption band positions for PAH cations, but anharmonic frequencies for Raman active modes might be still important for PAH cations, although more experimental da ta are needed to verify this.

PAGE 166

148 CHAPTER 8 CONCLUSION AND FUTURE WORK Matrix isolation spectroscopy (infrared a nd ultraviolet-visibl e) and theoretical calculations have been empl oyed to study astrophysical sp ecies in this work. This research has focused mainly on two types of astrophysical molecules, carbon clusters and their derivatives, and polycyclic aromatic hydrocarbons. The conclusion and prospects for future work will be discussed in detail below. Carbon Sulfur Clusters and Reacti on of Small Carbon Clusters Recently, chains of carbon atoms and thei r derivatives, such as carbon clusters doped with sulfur, nitrogen, etc., have been studied extensively, primarily because of their importance as components of the interste llar medium (ISM) and building blocks in material sciences, such as fullerenes (like C60), new nanomaterials (like nanowires), and novel semiconductor materials. Carbon-sulfur clusters have been generated by laser ablation, electron-bombardment, or plasma pulse discharge. The products were then trapped in solid argon at 12 K. IR absorption spectroscopy and 13C-isotope shift measurement coupled with coincidence anal ysis and DFT computation have confirmed the assignment of the fundament al vibrational modes for C2S, C6S, C7S, C7S2, C13S2, and C15S2 clusters. Tentative assignment for so me carbon sulfur clusters such as C9S2 and C11S2 clusters was made without isotopomer frequency observation. Some of these molecules have not only astrophysical interest but they are also recognized as novel nanowires, such as C15S2 molecule with a linear structur e of length of 2.1 nanometers.

PAGE 167

149 Polycarbon xenon clusters (or complexes) were studied in argon matrices at 12 K. The charge transfer complexes C2Xe and C3Xe were characterized with extensive theoretical calculations and isotopic labe ling. Formation of other polycarbon xenon clusters was confirmed, but their isotope shift measurement is still a challenge. In summary, the unique electronic structures of CnXe species were revealed and may help in understanding the overall react ivity of carbon clusters. Reactions of small carbon clusters with ot her interstella r species might be involved in the formation of more complex molecule s, including biomolecules. Amino acids have been found in meteorites and, very recently , in interstellar sp ace. Reactions of C3 with benzene or ammonia (all three ar e interstellar molecules) were investigated using FTIR and theoretical computations. C3 was generated from laser ablation of graphite and trapped in argon seeded with a small frac tion of benzene or a mmonia. The results revealed that the reaction of C3 + C6H6 may account, at least in pa rt, for the formation of polycyclic aromatic hydrocarbons in space, and the reaction of C3 + NH3 leads to the formation of cyanoacetylene (HC3N), which is the possible pr ecursor of some important biological molecules, like uracil. Polycyclic Aromatic Hydrocarbons Polycyclic aromatic hydrocarbons (PAHs) a nd their ions are key molecular species in many branches of chemistry, such as interstellar, medical, environmental, and materials chemistry. These species are strong candidates for the carriers of interstellar infrared (UIR) emission features and diffuse interstellar visi ble absorption bands (DIB). They are also primary intermediate species th at form in combustion processes and are the most ubiquitous environmental contaminants from natural and ma nmade sources with various mutagenic and ca rcinogenic activities.

PAGE 168

150 Dibenzo[b,def]chrysene (C24H14, dibenzo[a,h]pyrene), th e third most carcinogenic hydrocarbon, and its ions were studied via spectroscopic observation and extensive calculations. High resolution vi brational and electronic spec tra of neutral DBC and its ions in argon matrices have been recorded . Spectral assignments were supported by high level theoretical calculations. The results are important for the furt her understanding of the electronic structure of DBC which might clarify his interaction with DNA. Moreover, comparison of the absorption bands of ionic dibenzo[b,def]chrysene with the interstellar features of the UIR and DIB bands shows a satisfactory agreement. And a mixture of vibrational bands of neutra l and ionic DBC exhibits a reasonably good match with the ISOCAM + CVF spectrum of th e reflection nebula NGC 7023. Anharmonicity for PAHs is important fo r a comparison of the signals from space and laboratory data. Anharmonic frequencie s for neutral and cationic naphthalene, phenanthrene, and anthracene, calculated at the B3LYP level, were compared with various experimental data. Satisfactory agreem ent was achieved for a ll three species. An anharmonic correction depends on the level of theory and the vibra tional energy range. A strong anharmonic effect was calculated for the C–H stretc hing modes of PAHs. Since PAH cations have very weak infrared absorpti on in this region, the harmonic frequencies, after scaling, can be used to predict the band positions. An anharmonic analysis of the Raman active vibrational modes is expected to improve the accuracy of the predictions. Future Work Matrix isolation spectroscopy is a powe rful tool in the investigation of astrophysical molecules, but limitations and di sadvantages do exist fo r this technique. First, there is no direct chemical compositi on information from spectroscopic data, which makes it difficult to analyze spectra of mixtures contai ning various unknown molecules

PAGE 169

151 or radicals. Second, the inters tellar medium exists in the ga s phase, but matrix data are obtained in solid phase. Third, although inert gases are comm on matrices, there is still a matrix shift and a site effect. Finally, not al l interstellar species can be explored using matrix isolation spectroscopy, especially ions that are difficult to generate and require special sample handling. Recently, mass spectrometry coupled with laser spectroscopy has emerged as an attractive alternative for future astrochemist ry research. With its high sensitivity and mass solving power, a mass spectrometer may serve as a reaction chamber and a mass indicator. And the subsequent spectroscopic investigation with theoretical computation can provide the electronic and structural prop erties for the species of concern. Several advantages and disadvantages ar e inherent in this technique. It is easy to investigate the ions and reaction of ions in the gas phase, but the temperature of the ion beam is impossible to measure and difficult to adju st. Spectroscopic explor ation can be done on specific radical ions, thus excluding any interference am ong different species in the spectrum. But the resolution of spectrum vari es significantly for di fferent instrumental designs. The colder the ion beams, the highe r the resolution. Alt hough the instruments are expensive, cumbersome, and noncommercial, more and more in terest is being displayed in this approach. Therefore, in the near fu ture, mass spectrometry combined with laser spectroscopy may be expected to blossom in the field of astrophysical molecular research. Recently, some exciting research a bout the reaction of iron with interstellar molecules such as carbonyl, water and P AHs has been accomplished. These reactions might account for the unusual depletion of ir on in outer space. And iron-PAH complexes might also be considered as possible UIR and DIB carriers.

PAGE 170

152 Time-of-flight (TOF) mass spectrometry integrated with laser photodissociation spectroscopy is employed in Dr. Brucat’s lab. First, ions are produced by the vaporization of a rotating transition metal rod, and ligand molecules like CO and H2O, carried by helium gas through a pulsed valve. Then ions are supersonically inj ected into a large vacuum chamber (500 liters) and pass through two skimmers. The ions are cooled down and then extracted into the acceleration region. Cations ar e first extracted and then accelerated 90° out of the molecu lar beam into a 2.45 m flight tube, then focused with the use of deflectors and einzel lenses. If performing photodissoci ation, tunable light from a Nd:YAG pumped dye laser is focused just befo re the entrance slit of an electrostatic sector, and photodissociates the mass-selected parent ions. Finally, pr ecursor or fragment ions enter into a 180° hemisphe rical electrostatic sector that serves as an ion kinetic energy analyzer. Mass-to-charge ratios (m/z ) can be determined by a microchannel plate (MCP). With this instrument, a cold ion b eam can be generated and a high resolution vibrational-rotational spectrum obtained. Figure 8-1 is an example of th e dissociation spectra of FeH2O+, FeHDO+ and FeD2O+.274 Spectroscopic constants such as the dissociation limit, fundamental vibrational frequencies, and first-order even second-order anharmonic constants of ironwater cation can be determined from these spec tra. Isotope effect is also clear in the spectra, and can be used to evaluate the prope rties of the complex cations. If the scanning speed of dye laser is decreased, the rotati onal profile may be observed for a certain vibrational transition. And it can be simulate d with an instrument al function (Gaussian, FWHM = 0.15 cm–1). Then a parameterized rotationa l band contour can be obtained and the rotational constants of th e corresponding electronic states derived. Since the rotational

PAGE 171

153 constant is an inverse function of bond length, the average bond lengths for the complexes in their ground state and excited states can be determined. Figure 8-1. Photodissociation spectra of iron water complex cation. The spectra were observed using TOF mass spectrometry combined with laser spectroscopy. Theoretical calculation has confirmed th e existence of transition metal PAH complexes. Figure 8-2 presents the infrar ed spectra for naphthalene and an ironnaphthalene complex cation with differing spin multiplicities.275 It is clear that the absorption of naphthalene is distinguishable from the absorption of the complex cations. With different spins, complex cations also possess different absorption band positions and intensity distributions. Thus, experimental infrared spectra for metal-PAH complexes can be used to determine not only their struct ure but also the spin of metals with which they are complexed. Matrix isolation spectroscopy is not well-su ited for this type of study. Recently, an FT-ICR mass spectrometer coupled to a free electron laser has been successful in recording the infrared spectra for iron-PAH comp lex ions. In this set up, iron cations were

PAGE 172

154 generated by the Nd:YAG laser ablation of the me tal, they then drifted into the ICR cell, and reacted with PAH vapor to form iron-P AH complex ions. Partic ular complex cations were trapped in the cell and dissociated with the IR laser beam from the free electron laser. An infrared active spectrum is obtai ned by plotting the disso ciation yield as a function of frequency of the IR laser beam. However, the resolution of spectra obtained using this technique is usually low due to the strong broadening caused by the high internal energies of the i ons, by multiphoton dissociation, an d by the inherent line-width of the free-electron laser. Figure 8-2. Predicted spectra for naphthal ene and iron naphthalene complex cation. Calculations were carried out at BPW 91 level with 6-31G(d) basis set for Fe and 4-31G(d,p) basis set for C and H. All frequencies were scaled by 0.95.

PAGE 173

155 Briefly, laser spectroscopy-assisted mass spectrometry is becoming more and more promising for the research of astrophysical species and the react ions among them. The drawback of mass spectrometry is that it is designed for the detection of ions, which makes it nearly impossible to study neutral spec ies. Nevertheless, it is expected that many interesting results will be forthcoming using this approach in the near future.

PAGE 174

156 LIST OF REFERENCES 1. Schmidt, T. W.; Sharp, R. G. Aust. J. Chem. 2005, 58 , 69 and references therein. 2. Ehrenfreund, P.; Charnley, S. B. Ann. Rev. Astron. Astrophys. 2000, 38 , 427. 3. Merrill, P. W. PASP 1934, 46 , 206. 4. Merrill, P. W. Astrophys. J. 1936, 83 , 126. 5. Gillet, F. C.; Forrest, W. J.; Merrill, K. M. Astrophys. J. 1973, 183 , 87. 6. Russel, R.; Soifer, B.; Willner, S. 1977, 217, L149; 1978, 220 , 568. 7. Unicersity of New Hampshire Experimental Space Plasma Group The Interstellar Mediu: an Online Tutorial 2000, http://www-ssg.sr.unh.edu/ism/index.html 08/2005. 8. The NASA Imagine Team What is your Cosmic C onnection to the Elements 1997, http://imagine.gsfc.nasa.gov/docs/teach ers/elements/imagine/contents.html 08/2005. 9. Escalante-Ramírez, V. Re: Space matter, energy, anti-matter 2003, http://www.madsci.org/posts/a rchives/Mar2003/1046619096.As.r.html 08/2005. 10. Spaans, M.; Ehrenfreund, P. In Laboratory Astrophysics and Space Research , eds, Ehrenfreund, P.; Krafft, K.; Kochan, H.; Pirronello, V. (Kluwer, Dordrecht), 1999. 11. Leger, A; Puget, J. L. Astron. Astrophys. 1984, 137 , L5. 12. Allamandola, L. J.; Tielens, A. G. G. M.; Barker, J. R. Astrophys. J. 1985, 290 , L25. 13. Allamandola, L. J. In Solid State Astrophysics , eds. Bussoletti, E.; Strazzulla, G. (North Holland, Amsterdam), 1991. 14. Allamandola, L. J.; Hudgins, D. M.; Sandford, S. A. Astrophys. J. 1999, 511 , L115. 15. Szczepanski, J.; Fuller, J.; Ekern, S.; Vala M. Spectrochim. Acta A 2001, 57 , 775. 16. Bell, M. B.; Matthews, H. E. Astrophys. J. 1985, 291 , L63. 17. Herbig, G. H. Ann. Rev. Astron. Astrophys. 1995, 33 , 19.

PAGE 175

157 18. Salama, F.; Bakes, E.; Allamandola, L. J.; Tielens, A. G. G. M. Astrophys J. 1996, 458 , 621. 19. Brechignac, P.; Pino, T. Astron. Astrophys. 1999, 343 , L49. 20. Freivogel, P.; Fulara, J.; Maier, J. P. Astrophys J. 1994, 431 , L151. 21. Yulej, M.; Kirkwood, D. A.; Pachkov, M.; Maier, J. P. Astrophys J. 1998, 506 , L49. 22. Foing, B. H.; Ehrenfreund, P. Astron. Astrophys. 1997, 317 , L59. 23. The Astrochemistry Lab at NASA Ames Research Center The Diffuse Interstellar Bands 2005, http://www.astrochem.org/DIBS.html 08/2005. 24. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; M ontgomery Jr., J. A.; Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V. ; Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochtersk i, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz, J. V.; Baboul, A. G.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Andres, J. L.; Gonzalez, C.; Head-Gordon, M.; Replogle, E. S.; Pople, J. A. Gaussian, Inc., Revision A.11, Pittsburgh PA, 1998. 25. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, Jr., J. A.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishi da, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ocht erski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas , O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A.; Gaussian , Inc., Revision B.05, Wallingford CT, 2003.

PAGE 176

158 26. Stanton, J. F.; Gauss, J.; Watts, J. D.; Nooijen, M.; Oliphant, N.; Perera, S. A.; Szalay, P.G.; Lauderdale, W. J.; Kucharski, S. A.; Gwaltney, S. R.; Beck, S.; Balková A.; Bernholdt,D. E.; Baeck, K. K. ; Rozyczko, P.; Sekino, H.; Hober, C.; Bartlett, R. J. ACES II is a program product of the Quantum Theory Project, University of Florida. 27. Straatsma, T. P.; Aprà, E.; Windus, T. L.; Dupuis, M.; Bylaska, E. J.; Jong, de W.; Hirata, S.; Smith, D. M. A. M.; Hackler, T.; Pollack, L.; Harrison, R. J.; Nieplocha, J.; Tipparaju, V.; Krishnan, M.; Auer, A. A.; Brown, E.; Cisneros, G.; Fann, G. I.; Fruchtl, H.; Garza, J.; Hirao, K.; Kenda ll, R.; Nichols, J. A.; Tsemekhman, K.; Valiev, M.; Wolinski, K.; Anchell, J.; Ber nholdt, D.; Borowski, P. ; Clark, T.; Clerc, D.; Dachsel, H.; Deegan, M.; Dyall, K. ; Elwood, D.; Glendening, E.; Gutowski, M.; Hess, A.; Jaffe, J.; Johnson, B.; Ju, J.; Kobayashi, R.; Kutteh, R.; Lin, Z.; Littlefield, R.; Long, X.; Meng, B.; Naka jima, T.; Niu, S.; Rosing, M.; Sandrone, G.; Stave, M.; Taylor, H.; Thomas, G.; Lenthe, van J.; Wong, A.; Zhang, Z. NWCHEM, A Computational Chemistry Pack age for Parallel Computers, Pacific Northwest National Laboratory, Richland, WA, 2003. 28. Becke, A.D. Phys. Rev. A 1988, 38 , 3098. 29. Lee, C.; Yang, W. and Parr, R.P. Phys. Rev. B 1988, 37 , 785. 30. Dunning, T. H. Jr. J. Chem. Phys. 1989, 90 , 1007. 31. Lundell, J.; Cohen, A.; Gerber, R. B. J. Phys. Chem. A 2002, 106 , 11950 and references therein. 32. Glending, E. D.; Weinhold, F. J. J. Comp. Chem. 1998, 19 , 628. 33. Wiberg, K. B. Tetrahedron 1968, 24 , 1083. 34. Peng, C.; Ayala, P. Y.; Sc hlegel, H. B.; Frisch, M. J. Comp. Chem. 1996, 17 , 49. 35. Peng, C.; Schlegel, H. B. Israel J. Chem. 1994, 33 , 449. 36. Gonzalez, C.; Schlegel, H. B. J. Chem. Phys. 1989, 90 , 2154. 37. Gonzalez, C.; Schlegel, H. B. J. Phys. Chem. 1990, 94 , 5523. 38. Hirata, S.; Head-Gordon, M.; Szczepanski, J.; Vala, M. J. Phys. Chem. A 2003, 107 , 4940 and reference therein. 39. Hirata, S.; Zhan, C. G.; Aprà, E.; Windus, T.; Dixon, D. A. J. Phys. Chem. A 2003, 107 , 10154. 40. Zhan, C. G.; Nichols, J. A.; Dixon, D. A.; J. Phys. Chem. A 2003, 107 , 4184.

PAGE 177

159 41. Vala, M.; Chandrasekhar, T. M.; Szczepanski, J.; Van Zee, R.; Weltner, W. Jr. J. Chem. Phys. 1989, 90 , 595. 42. Thoma, A.; Wurfel, B. E.; Schlachta, R.; Lask, G. M.; Bondybey, V. E. J. Phys. Chem. 1992, 96 , 7231. 43. Spiro, C. L.; Banholtzer, W. F.; McAtee, D. S. Thin Solid Films 1992, 220 , 122. 44. Kaifu, N.; Suzuki, H.; Ohishi, M.; Miyaji, T.; Ishikawa, S.; Kasuga, T.; Morimoto, M. Astrophys. J. 1987, 317 , L111. 45. Saito, S.; Kawaguchi, K.; Yamamoto, S.; Ohishi, M.; Suzuki, H.; Kaifu, N. Astrophys. J. 1987, 317 , L115. 46. Yamamoto, S.; Saito, S.; Kawaguchi, K.; Kaifu, N.; Suzuki, H.; Ohishi, M. Astrophys. J. 1987, 317 , L119. 47. Cernicharo, J.; Guelin, M.; Hein, H.; Kahane, C. Astron. Astrophys. 1987, 181 , L9. 48. Fuente, A.; Cernicharo, J.; Barcia, A.; Gomezgonzalez, J. Astron. Astrophys. 1990, 231 , 151. 49. Suzuki, H.; Yamamoto, S.; Ohishi, M.; Kaifu, N.; Ishikawa, S.; Hirahara, Y.; Takano, S. Astrophys. J. 1992, 392 , 551. 50. Hirahara, Y.; Suzuki, H.; Yamamoto, S.; Kawaguchi, K.; Kaifu, N.; Ohishi, M.; Takano, S.; Ishikawa, S. I.; Masuda, A. Astrophys. J. 1992, 394 , 539. 51. Maier, G.; Schrot, J.; Reisenauer, H. P.; Janoschek, R. Chem. Ber. 1990, 123 , 1753. 52. Maier, G.; Reisenauer, H. P.; Schrot, J.; Janoschek, R. Angew. Chem. Int. Edit. 1990, 29 , 1464. 53. Maier, G.; Schrot, J.; Reisenauer, H. P. Chem. Ber. 1991, 124 , 2613. 54. Maier, G.; Schrot, J.; Reisenauer, H. P.; Janoschek, R. Chem. Ber. 1991, 124 , 2617. 55. Bohn, R. B.; Hannachi, Y.; Andrews, L. J. Am. Chem. Soc. 1992, 114 , 6452. 56. Szczepanski, J.; Hodyss, R.; Fuller, J.; Vala, M. J. Phys. Chem. A 1999, 103 , 2975 and references therein. 57. Hirahara, Y.; Ohshima, Y.; Endo, Y. Astrophys. J. 1993, 408 , L113. 58. Kasai, Y.; Obi, K.; Ohshima, Y.; Hirahara, Y.; Endo, Y.; Kawaguchi, K.; Murakami, A. Astrophys. J. 1993, 410 , L45. 59. Gordon, V. D.; McCarthy, M. C. ; Apponi, A. J.; Thaddeus, P. Astrophys. J. Suppl. S. 2001, 134 , 311.

PAGE 178

160 60. Riaplov, E.; Wyss, M.; Maier, J. P.; Panten, D.; Chambaud, G.; Rosmus, P.; Fabian, J. J. Mol. Spectrosc. 2003, 222 , 15. 61. Riaplov, E.; Maier, J. P. J. Phys. Chem. A 2003, 107 , 8856. 62. Sulzle, D.; Beye, N.; Fanghanel, E.; Schwarz, H. Chem. Ber. 1990, 123 , 2069. 63. Burnin, A.; BelBruno, J. J. J. Phys. Chem. A 2003, 107 , 9547. 64. Lee, S. Chem. Phys. Lett. 1997, 268 , 69. 65. Kim, K. H.; Lee, B.; Lee, S. Chem. Phys. Lett. 1998, 297 , 65. 66. Peeso, D. J.; Ewing, D. W.; Curtis, T. T. Chem. Phys. Lett. 1990, 166 , 307. 67. Murakami, A. Astrophys. J. 1990, 357 , 288. 68. Xie, Y. M.; Schaefer, H. F. J. Chem. Phys. 1992, 96 , 3714. 69. Cai, Z. L.; Zhang, X. G.; Wang, X. Y. Chem. Phys. Lett. 1993, 213 , 168. 70. Seeger, S.; Botschwina, P.; Flugge, J. ; Reisenauer, H. P.; Maier, G. J. Mol. Struc. – Theochem 1994, 109 , 213. 71. Martin, J. M. L.; Francois, J. P.; Gijbels, R. J. Mol. Spectrosc. 1995, 169 , 445. 72. Pascoli, G.; Lavendy, H. Int. J. Mass. Spectrom. 1998, 181 , 11. 73. Tang, Z. C.; BelBruno, J. J. Int. J. Mass. Spectrom. 2001, 208 , 7. 74. Wang, H. Y.; Lu, X.; Huang, R. B.; Zheng, L. S. J. Mol. Struc. – Theochem 2002, 593 , 187. 75. Watts, J. D.; Gauss, J.; Stanton, J. F.; Bartlett, R. J. J. Chem. Phys. 1992, 97 , 8372. 76. Watts, J. D.; Bartlett, R. J. Chem. Phys. Lett. 1992, 190 , 19. 77. Martin, J. M. L.; Elyazal, J.; Francois, J. P. Chem. Phys. Lett. 1995, 242 , 570. 78. Van Orden, A.; Saykally, R. J. Chem Rev 1998, 98 , 2313 and reference therein. 79. Arnold, D. W.; Bradforth, S. E.; Kitsopoulos, T. N.; Neumark, D. M. J. Chem. Phys. 1991, 95 , 8753. 80. Forney, D.; Fulara, J.; Freivogel, P.; Jakobi, M.; Lessen, D.; Maier, J. P. J. Chem. Phys. 1995, 103 , 48. 81. Freivogel, P.; Fulara, J.; Jakobi, M.; Forney, D.; Maier, J. P. J. Chem. Phys. 1995, 103 , 54.

PAGE 179

161 82. Freivogel, P.; Grutter, M.; Forney, D.; Maier, J. P. Chem. Phys. 1997, 216 , 401. 83. Kranze, R. H.; Graham, W. R. M. J. Chem. Phys. 1993, 98 , 71. 84. Kranze, R. H.; Rittby, C. M. L.; Graham, W. R. M. J. Chem. Phys. 1996, 105 , 5313. 85. Banisaukas, J.; Szczepanski, J.; Vala, M. Manuscript in preparation. 86. CRC Handbook of Chemistry and Physics [CD-ROM]; CRC Press: Boca Raton, FL, 1999. 87. Ding, X. D.; Wang, S. L.; Rittby, C. M. L.; Graham, W. R. M. J. Chem. Phys. 1999, 110 , 11214. 88. Wang, H. Y.; Szczepanski, J.; Brucat, P. J.; Vala, M. T. J. Phys. Chem. A 2003, 107 , 10919. 89. Schoeffler, A. J.; Kohguchi, H.; Hoshina, K.; Ohshima, Y.; Endo, Y. J. Chem. Phys. 2001, 114 , 6142. 90. Scott, A. P.; Radom, L. J. Phys. Chem. 1996, 100 , 16502. 91. Wang, H. Y.; Huang, R. B.; Chen, H.; Lin, M. H.; Zheng, L. S. J. Phys. Chem. A 2001, 105 , 4653. 92. Huang, R. B.; Wang, C. R.; Liu, Z. Y.; Zheng, L. S.; Qi, F.; Sheng, L. S.; Yu, S. Q.; Zhang, Y. W. Z. Phys. D Atom Mol. Cl. 1995, 33 , 49. 93. Liu, Z. Y.; Huang, R. B.; Tang, Z. C.; Zheng, L. S. Chem. Phys. 1998, 229 , 335. 94. Pascoli, G.; Lavendy, H. J. Phys. Chem. A 1999, 103 , 3518. 95. Lapinski, L.; Vala, M. Chem. Phys. Lett. 1999, 300 , 195. 96. Wang, S. L.; Rittby, C. M. L.; Graham, W. R. M. J. Chem. Phys. 1997, 107 , 7025. 97. Wang, S. L.; Rittby, C. M. L.; Graham, W. R. M. J. Chem. Phys. 2000, 112 , 1457. 98. Ding, X. D.; Wang, S. L.; Rittby, C. M. L.; Graham, W. R. M. J. Chem. Phys. 2000, 112 , 5113. 99. Forney, D.; Freivogel, P.; Grutter, M.; Maier, J. P. J. Chem. Phys. 1996, 104 , 4954. 100. Szczepanski, J.; Ekern, S.; Chapo, C.; Vala, M. Chem. Phys. 1996, 211 , 359. 101. PresillaMarquez, J. D.; Sheehy, J. A.; Mills, J. D.; Carrick, P. G.; Larson, C. W. Chem. Phys. Lett. 1997, 274 , 439.

PAGE 180

162 102. Wang, S. L.; Rittby, C. M. L.; Graham, W. R. M. J. Chem. Phys. 1997, 107 , 6032. 103. Szczepanski, J.; Vala, M. J. Phys. Chem. 1991, 95 , 2792. 104. Kranze, R. H.; Graham, W. R. M. J. Chem. Phys. 1992, 96 , 2517. 105. Doyle, T. J.; Shen, L. N.; Rittby, C. M. L.; Graham, W. R. M. J. Chem. Phys. 1991, 95 , 6224. 106. Szczepanski, J.; Auerbach, E.; Vala, M. J. Phys. Chem. A 1997, 101 , 9296. 107. Szczepanski, J.; Ekern, S.; Vala, M. J. Phys. Chem. A 1997, 101 , 1841; Dibben, M.; Szczepanski, J.; Wehlburg, C.; Vala, M. ibid. 2000, 104 , 3584. 108. Szczepanski, J.; Hodyss, R.; Vala, M. J. Phys. Chem. A 1998, 102 , 8300. 109. Szczepanski, J.; Wang, H. Y.; Jones, B.; Arrington, C. A.; Vala, M. Phys. Chem. Chem. Phys. 2005, 7 , 738. 110. Szczepanski, J.; Wehlburg, C.; Vala, M. J. Phys. Chem. A 1997, 101 , 7039. 111. Szczepanski, J.; Vala, M.; Shen, L. N.; Withey, P. A.; Graham, W. R. M. J. Phys. Chem. A 1997, 101 , 8788. 112. Maier, G.; Reisenauer, H. P. Advances in Carbene Chemistry; Elsevier: Amsterdam, 2001. 113. Holloway, J. H.; Hope, E. G. Adv. Inorg. Chem. 1999, 46 , 51. 114. Seppelt, K.; Lentz, D. Prog. Inorg. Chem. 1982, 29 , 167. 115. Burger, H.; Kuna, R.; Ma, S.; Breidung, J.; Thiel, W. J. Chem. Phys. 1994, 101 , 1. 116. Burger, H.; Ma, S.; Breidung, J.; Thiel, W. J. Chem. Phys. 1996, 104 , 4945. 117. Bihary, Z.; Chaban, G. M.; Gerber, R. B. J. Chem. Phys. 2002, 116 , 5521. 118. Cohen, A.; Niv, M. Y.; Gerber, R. B. Faraday Discuss. 2001, 118 , 269. 119. Khriachtchev, L.; Lundell, J.; Petterss on, M.; Tanskanen, H.; Rasanen, M. J. Chem. Phys. 2002, 116 , 4758. 120. Khriachtchev, L.; Pettersson, M.; Isoniemi, E.; Rasanen, M. J. Chem. Phys. 1998, 108 , 5747. 121. Khriachtchev, L.; Pettersson, M.; Jol kkonen, S.; Pehkonen, S.; Rasanen, M. J. Chem. Phys. 2000, 112 , 2187.

PAGE 181

163 122. Khriachtchev, L.; Pettersson, M.; Lignell, A.; Rasanen, M. J. Am. Chem. Soc. 2001, 123 , 8610. 123. Khriachtchev, L.; Pettersson, M.; Lundell, J.; Rasanen, M. J. Chem. Phys. 2001, 114 , 7727. 124. Khriachtchev, L.; Pettersson, M.; Lundell, J.; Tanskanen, H.; Kiviniemi, T.; Runeberg, N.; Rasanen, M. J. Am. Chem. Soc. 2003, 125 , 1454. 125. Khriachtchev, L.; Pettersson, M.; Rune berg, N.; Lundell, J.; Rasanen, M. Nature 2000, 406 , 874. 126. Khriachtchev, L.; Pettersson, M. ; Tanskanen, H.; Rasanen, M. Chem. Phys. Lett. 2002, 359 , 135. 127. Khriachtchev, L.; Saarelainen, M. ; Pettersson, M.; Rasanen, M. J. Chem. Phys. 2003, 118 , 6403. 128. Khriachtchev, L.; Tanskanen, H.; Lunde ll, J.; Pettersson, M.; Kiljunen, H.; Rasanen, M. J. Am. Chem. Soc. 2003, 125 , 4696. 129. Khriachtchev, L.; Tanskanen, H.; Petterss on, M.; Rasanen, M.; Ahokas, J.; Kunttu, H.; Feldman, V. J. Chem. Phys. 2002, 116 , 5649. 130. Nemukhin, A. V.; Grigorenko, B. L.; Khri achtchev, L.; Tanskanen, H.; Pettersson, M.; Rasanen, M. J. Am. Chem. Soc. 2002, 124 , 10706. 131. Pettersson, M.; Khriachtchev, L. ; Jolkkonen, S.; Rasanen, M. J. Phys. Chem. A 1999, 103 , 9154. 132. Pettersson, M.; Khriachtchev, L.; Lignell, A.; Rasanen, M.; Bihary, Z.; Gerber, R. B. J. Chem. Phys. 2002, 116 , 2508. 133. Pettersson, M.; Khriachtchev, L.; Lunde ll, J.; Jolkkonen, S.; Rasanen, M. J. Phys. Chem. A 2000, 104 , 3579. 134. Pettersson, M.; Khriachtchev, L.; Lundell, J.; Rasanen, M. J. Am. Chem. Soc. 1999, 121 , 11904. 135. Pettersson, M.; Lundell, J.; Khriachtche v, L.; Isoniemi, E.; Rasanen, M. J. Am. Chem. Soc. 1998, 120 , 7979. 136. Pettersson, M.; Lundell, J.; Kh riachtchev, L.; Rasanen, M. J. Chem. Phys. 1998, 109 , 618. 137. Pettersson, M.; Lundell, J.; Rasanen, M. J. Chem. Phys. 1995, 103 , 205. 138. Pettersson, M.; Lundell, J.; Rasanen, M. J. Chem. Phys. 1995, 102 , 6423.

PAGE 182

164 139. Maier, G.; Lautz, C. Euro. J. Org. Chem. 1998, 769. 140. Feldman, V. I.; Sukhov, F. F.; Or lov, A. Y.; Tyulpina, I. V. J. Am. Chem. Soc. 2003, 125 , 4698. 141. Brupbacher, T.; Makarewicz, J.; Bauder, A. J. Chem. Phys. 1994, 101 , 9736. 142. Kotting, C.; Sander, W.; Breidung, J.; Th iel, W.; Senzlober, M.; Burger, H. J. Am. Chem. Soc. 1998, 120 , 219. 143. Rutkowski, K. S.; Melikova, S. M. J. Mol. Struc. 1998, 448 , 231. 144. Velino, B.; Favero, P. G.; Caminati, W. J. Chem. Phys. 2002, 117 , 5688. 145. Favero, L. B.; Velino, B.; Millemaggi, A.; Caminati, W. ChemPhysChem 2003, 4 , 881. 146. Brown, E. C.; Cohen, A.; Gerber, R. B. J. Chem. Phys. 2005, 122 , 171101 and references therein. 147. Breidung, J.; Thiel, W. J. Mol. Spectrosc. 2002, 216 , 424. 148. Kroto, H. W.; Heath, J. R.; Obrien, S. C.; Curl, R. F.; Smalley, R. E. Nature 1985, 318 , 162. 149. Weltner, W.; Vanzee, R. J. Chem. Rev. 1989, 89 , 1713 and references therein. 150. Smith, A. M.; Agreiter, J.; Hartle, M.; Engel, C.; Bondybey, V. E. Chem. Phys. 1994, 189 , 315. 151. Lakhlifi, A.; Dahoo, R.; Vala, M.; Szczepanski, J.; Ekern, S. Chem. Phys. 1997, 222 , 241. 152. Milligan, D. E.; Jacox, M. E.; Abouafma.L. J. Chem. Phys. 1967, 46 , 4562. 153. Jacox, M. E.; Milligan, D. E. Chem. Phys. 1974, 4 , 45. 154. Andrews, L.; Kushto, G. P.; Zhou, M. F.; Willson, S. P.; Souter, P. F. J. Chem. Phys. 1999, 110 , 4457. 155. Weltner, W.; Walsh, P. N.; Angell, C. L. J. Chem. Phys. 1964, 40 , 1299. 156. Kranze, R. H.; Rittby, C. M. L.; Graham, W. R. M. J. Chem. Phys. 1996, 105 , 5313. 157. Szczepanski, J.; Vala, M. Nature 1993, 363 , 699. 158. Parisel, O.; Berthier, J.; Ellinger, Y. Astron. Astrophys. 1992, 226 , L1.

PAGE 183

165 159. Langhoff, S. R.; Bauschlicher, C.W. Jr .; Hudgins, D. M.; Sandford, S. A.; Allamandola, L. J. J. Phys. Chem. A 1998, 102 , 1632. 160. Jourdain, M.; d’Hendecourt, L. B.; Geballe, T. R. Astron. Astrophys. 1990, 227 , 526. 161. Mendoza-Gómez, C. X.; deGroot, M. S.; Greenberg, M. M. Astrn. Astrophys. 1995, 295, 479. 162. Snow, T. P.; Page, V.; Kehe yan, Y.; Bierbaum, V. M. Nature 1998, 391 , 259. 163. Bernstein, M. P.; Sandford S. A.; Allamandola, L. J. Astrophys. J. 1996, 472 , L127. 164. Beintema, D. A.; van den Ancker, M. E.; Waters, L. B. F. M. Astron. Astrophys. 1996, 315 , L369. 165. Pauzat, F.; Talbi D.; Ellinger, Y. Astron. Astrophys. 1997, 319 , 318. 166. Ekern, S. P.; Marshall, A. G.; Szczepanski, J.; Vala, M. J. Phys. Chem. A 1998, 102 , 3498. 167. Salama, F.; Allamandola, L. J. Astrophys. J. 1990, 395 , 301. 168. Salama, F.; Allamandola, L. J. J. Chem. Soc. Fara. Trans. 1993, 89 , 2277. 169. Cossart-Magos, C.; Leach, S. Astron. Astrophys. 1990, 233 , 559. 170. Cherchneff, I.; Barker, J. R. Astrophys. J. 1992, 394 , 703. 171. Allain, T.; Sedlmayr, E.; Leach, S. Astrophys. Space Sci. 1995, 224 , 417. 172. Allain, T.; Sedlmayr, E.; Leach, S. Astron. Astrophys. 1997, 323 , 163. 173. Cadwell, B. J.; Wang, H.; Feigelson, E. D.; Frenklach, M. Astrophys. J. 1994, 429 , 285. 174. Kazakov, A.; Frenklach, M. Comb. Flame 1998, 112 , 270. 175. Chan, S. H.; Yin, J. Q.; Shi, B. J. Comb. Flame 1998, 112 , 445. 176. Wang R.; Cadman, P. Comb. Flame 1998, 112 , 359. 177. Arrington, C. A.; Ramos, C.; Robinson, A. D.; Zwier, T. S. J. Phys. Chem. A 1998, 102 , 3315. 178. Kaiser R. I.; Roessler, K. Astrophys. J. 1997, 475 , 144. 179. Kaiser, R. I.; Asvany, O.; Lee, Y. T. Planet. Space Sci. 2000, 48 , 483.

PAGE 184

166 180. Beltinger, H. F.; von Schleyer, P.; Schaefer, H. F.; Schreiner, P. R.; Kaiser, R. I.; Lee, Y. T. J. Chem. Phys. 2000, 113 , 4250. 181. Hahndorf, I.; Lee, Y. T.; Kaiser, R. I.; Ver eecken, L.; Peeters, J.; Bettinger, H. F.; Schreiner, P. R.; von Schleyer, P.; Allen W. D.; Schaefer, H. F. III J. Chem. Phys. 2002, 116 , 3248. 182. Hinkle, K. H.; Keady J. J.; Bernath, P. F. Science 1988, 241 , 1319. 183. Cernicharo, J.; Heras, A. M. ; Gonzalez-Alfonso, L. B. F. M.; Hony, S.; Yamamura, I.; Guelin, M.; Neri, R.; Dartois, E.; Pe rez-Martinez, S.; Cox, P.; DeJong, T.; Tielens, A. G. G. M.; Najarro, F. European Space Agency , [Special Publication] SP, 1999; SP-427 (Universe as Seen by ISO, Vol. 1), 1999, 285. 184. Woon, D. E.; Herbst, E. Astrophys. J. 1997, 477 , 204. 185. Ohishi, M.; Kaifu, N. Faraday Discuss. 1998, 109 , 205. 186. Huang, L. C. L.; Asvany, O.; Chang, A. H. H.; Balucani, N.; Lin, S. H.; Lee, Y. T.; Kaiser, R. I.; Osamura, Y. J. Chem. Phys. 2000, 113 , 8656. 187. Moran, J. M.; Papadopo, G. D.; Burke, B. F.; Schwartz, P. R.; Thacker, D. L.; Johnston, K. J.; Knowles, S. H.; Reisz, A. C.; Shapiro, I. I. Astrophys. J. 1973, 185 , 535. 188. Bell, M. B.; Avery, L. W.; Watson, J. K. G. Astrophys. J. Suppl. S. 1993, 86 , 211. 189. Hermsen, W.; Wilson, T. L.; Walmsley, C. M.; Henkel, C. Astron. Astrophys. 1988, 201 , 285. 190. Giesen, T. F.; Van Orden, A. O.; Cruzan, J. D.; Provencal, R. A.; Saykally, R. J.; Gendriesch, R.; Lewen, F.; Winnewisser, G. Astrophys. J. 2001, 551 , L181. 191. Maier, J. P.; Lakin, N. M.; Walk er, G. A. H.; Bohlender, D. A. Astrophys. J. 2001, 553 , 267. 192. Lacy, J.H.; Faraji, H.; Sandford, S. A.; Allamandola, L. J. Astrophys. J. 1998, 501 , L105. 193. Kawaguchi, K.; Takano, S.; Ohishi, M.; Is hikawa, S.;, Miyazawa, K.; Kaifu, N.; Yamashita, K.; Yamamoto, S.; Saito, S.; Ohshima, Y.; Endo, Y. Astrophys. J. , 1992, 396 , L49. 194. Churchwell, E.; Walmsley, C. M.; Winnewisser, G. Astron. Astrophys. 1977, 54 , 925. 195. Turner, B. E. Astrophys. J. 1971, 163 , L35. 196. Morris, M.; Turner, B. E.; Palmer, P.; Zuckerman, B. Astrophys. J. , 1976, 205 , 82.

PAGE 185

167 197. McGee, R. X.; Balister, M.; Newton, L. M. MNRAS 1977, 180 , 585. 198. Gensheimer, P. D. Astrophys. Space Sci. 1997, 251 , 199. 199. Schutte, W. A.; Greenberg, J. M. Astron. Astrophys. 1991, 244 , 190. 200. Brown, K. G.; Person, W. B. Spectrochim. Acta A 1978, 34 , 117. 201. Szczepanski, J.; Ekern, S. P.; Vala, M. J. Phys. Chem. 1995, 99 , 8002. 202. Interestingly, the calculated C3•HC6H5 bonding energy of –2.3 kJ/mol (with B3LYP/6-31G(d,p)) approaches zero as the ba sis set size is increased. Furthermore, the C3 asymmetric stretch in the complex’s ge ometry is computed to shift to higher frequencies, opposite to e xperimental observation. 203. Forney, D.; Freivogel, P.; Jacobi, M.; Lessen, D.; Maier, J. P. J. Chem. Phys. 1995, 103 , 43. 204. Güthe, F.; Ding, H.; Pino, T.; Maier, J. P. Chem. Phys. 2001, 269 , 347. 205. Maier, G.; Reisenauer, H. P.; Schwab, W.; Carsky, P.; Hess, B.A. Jr.; Schaad, L. J. J. Am. Chem. Soc. 1987, 109 , 5183. 206. Huang, J. W.; Graham, W. R. M. J. Chem. Phys. 1990, 93 , 1583. 207. Johnstone, D. E.; Sodeau, J. R. J. Phys. Chem. 1991, 95 , 165. 208. Jacox, M. E. J. Phys. Chem. 1982, 86 , 670. 209. Laboy, J. L.; Ault, B.S. J. Photochem. Photobiol. A: Chem. 1993, 74 , 99 and references therein. 210. Sinnokrot, M. O.; Valeev, E. F.; Sherrill, C. D. J. Am. Chem. Soc. 2002, 124 , 10887 and references therein. 211. Nelander, B.; Nord, L. J. Phys. Chem. 1982, 86 , 4375. 212. Fraser, G. T.; Leopold, K. R.; Klemperer, W. J. Chem. Phys. 1984, 80 , 1423. 213. Rodham, D. A.; Suzuki, S.; Suenram, R. D.; Lovas, F. J.; Dasgupta, S.; Goddard, W. A.; Blake, G. A. Nature 1993, 362 , 735. 214. Szczepanski, J.; Wang, H.; Vala, M. Chem. Phys. 2004, 303 , 165. 215. Botschwina, P.; Horn, M.; Seeger, S.; Fluegge, J. Chem. Phys. Lett. 1992, 195 , 427. 216. Kolos, R.; Sobolewski, A. L.; Chem. Phys. Lett. 2001, 344 , 625. 217. Botschwina, P. Phys. Chem. Chem. Phys. 2003, 5 , 3337.

PAGE 186

168 218. Guennoun, Z.; Couturier-Tamburelli, I.; Pietri, N.; Aycard, J. P. Chem. Phys. Lett. 2003, 368 , 574. 219. Cernicharo, J.; Heras, A.M.; Tielens, A.G. G. M.; Pardo, J. R.; Herpin, F.; Guelin, M.; Waters, L. B. F. M. Astrophys. J. 2001, 546, L123. 220. Kaiser, R. I. Chem. Rev. 2002, 102 , 1309. 221. Dipple, A., ACS Symposium Series 1985, 283 , 1. 222. Bostrom, C. E.; Gerde, P.; Hanberg, A.; Jernstrom, B.; Johansson, C.; Kyrklund, T.; Rannug, A.; Tornqvist, M.; Victorin, K.; Westerholm, R. Environ. Health Perspect. 2002, 110 , 451. 223. Lopes, W. A.; deAndrade, J. B. Quimica Nova 1996, 19 , 497. 224. Zimmermann, R.; Heger, H. J.; Kettrup, A.; Nikolai, U. Fresenius J. Anal. Chem. 2000, 366 , 368 225. Allamandola, L. J.; Tielens, A. G. G. M.; Barker, J. R. Astrophys. J. Suppl. Ser. 1989, 71 , 733. 226. dÂ’Hendecourt, L.; Ehrenfreund, P., Life Sciences: Complex Organics in Space 1997, 19 , 1023. 227. Rapacioli, M.; Joblin, C.; Boissel, P. Astron. Astrophys. 2005, 429 , 193 and reference therein. 228. Shida, T.; Iwata, S. J. Am. Chem. Soc. 1973, 95 , 3473. 229. Barone, P. M. V. B.; Camilo, A. Jr.; Galvao, D. S. Phys. Rev. Lett. 1996, 77 , 1186. 230. Andersen, K. B.; Waluk, J.; Thulstrup, E. W. Photochem. Photobio. 1999, 69 , 158. 231. Banisaukas, J.; Szczepanski, J.; Vala, M.; Hirata, S. J. Phys. Chem. A 2004, 108 , 3713. 232. Szczepanski, J.; Dibben, M. J.; Pearson, W.; Eyler, J. R.; Vala, M. J. Phys. Chem. A 2001, 105 , 9388. 233. Karcher, W.; Fordham, R.J.; Dubois, J. J.; Glaude, P. G. J. M.; Ligthart, J. A. M. Spectral Atlas of Polycyclic Aromatic Compounds , D. Reidel Publ. Co., Dordrecht, Boston, Lancaster, 1985. 234. Jacox, M. E.; Milligan, D. E. J. Chem. Phys. 1971, 54 , 3935 235. Andrews, L.; Prochaska, F. T. J. Phys. Chem. 1979, 83 , 368. 236. Andrews, L. J. Phys. Chem. 1967, 71 , 2761.

PAGE 187

169 237. Bondybey, V. E.; Pimentel, G. C. J. Chem. Phys. 1972, 56 , 3832. 238. Lacmann, K.; Maneira, M. J. P.; Moutinho, A. M. C.; Weigman, U. J. Chem. Phys. 1983, 78 , 1767. 239. Scheunemann, H. U.; Illenberger, E.; Baumgartel, H. Ber. Bunsen-Ges. Phys. Chem. 1980, 84 , 580. 240. Martin, J. D. D.; Hepburn, J. W. J. Chem. Phys. 1998, 109 , 8139. 241. Szczepanski, J.; Wehlburg, C.; Vala, M. Chem. Phys. Lett. 1995, 232 , 221. 242. Szczepanski, J; Banisaukas J.; Vala,M.; Hi rata, S.; Bartlett, R. J.; Head-Gordon, M. J. Phys. Chem. 2002, 106 , 63. 243. Schulz, P. A.; Mead, R. D.; Jone s, P. L.; Lineberger, W. C. J. Chem. Phys. 1982, 77 , 1153. 244. Shida, T. Electronic Absorption Spectra of Radical Ions Elsevier: Amsterdam, 1988. 245. Clar, E.; Schmidt, W. Tetrahedron 1979, 35 , 1027. 246. Clar, E.; Schmidt, W. Tetrahedron 1976, 32 , 2563. 247. Briegleb, G. Angew. Chem. 1964, 76 , 326. 248. Xie, Y.; Schaefer, H. F., III; Wang, Y.; Fu, X.; Liu, R. Mol. Phys. 2000, 98 , 279. and reference therein. 249. Gonzales, J. M.; Barden C. J.; Brown S. T. ; Schleyer P. von R.; Schaefer H. F. III; Li Q. S. J. Am. Chem. Soc. 2003, 125 , 1064. 250. Cesarsky, D.; Lequeux, J.; Abergel, A.; Pe rault, M.; Palazzi, E.; Madden, S.; and Tran, D. Astron. Astrophys. 1996, 315, L305. 251. Sellgren, K. Spectrochim.Acta 2001, 57 , 627. 252. Huneycutt, A. J.; Casaes, R. N.; McCall, B. J.; Chung, C. Y.; Lee, Y. P.; Saykally, R. ChemPhysChem 2004, 5 , 321 and references therein. 253. Jacox, M. E. Chem. Soc. Rev. 2002, 31 , 108. 254. Ruiterkamp, R.; Halasinski, T.; Foing, B. H.; Allamandola, L. J.; Schmidt, W.; Ehrenfreund, P. Astron. Astrophys. 2002, 390 , 1153, and references therein. 255. Biennier, L.; Salama, F.; Allama ndola, L. J.; Scherer, J. J. J. Chem. Phys. 2003, 118 , 7863.

PAGE 188

170 256. Pino, T.; Brechignac, P.; Dartois, E.; Demyk, K.; d'Hendecourt, L. Chem. Phys. Lett. 2001, 339 , 64. 257. Sukhorukov, O.; Staicu, A.; Diegel, E.; R ouille, G.; Henning, Th.; Huisken, F. Chem. Phys. Lett. 2004, 386 , 259. 258. Barker, J. R.; Allamandola, L. J.; Tielens, A. G. G. Astrophys. J. 1987, 315 , L61. 259. Peeters, E.; Hony, S.; Van Kerckhoven, C.; Tielens, A. G. G. M.; Allamandola, L. J.; Hudgins, D. M.; Bauschlicher, C. W. Astron. Astrophys. 2002, 390 , 1089. 260. Oomens, J.; Von Helden, G.; Meijer, G. Springer Proc. Phys. 2004, 9 , 545. 261. Kjaergaard, H. G.; Robinson, T. W.; Brooking, K. A. J. Phys. Chem. A 2000, 104, 11297; Van-Oanh, N.; Parneix, P.; Brechignac, P. ibid. 2002, 106, 10144. 262. Barone, V.; Festa, G.; Grandi A.; Rega N.; Sanna N. Chem. Phys. Lett. 2004, 388, 279. 263. Langhoff, S. R. J. Phys. Chem. 1996, 100, 2819 and reference therein. 264. Szczepanski, J.; Roser, D.; Personette, W.; Eyring, M.; Pellow, R.; Vala, M. J. Phys. Chem. 1992, 96 , 7876. 265. Szczepanski, J.; Vala, M.; Talbi, D.; Parisel, O.; Ellinger, Y. J. Chem. Phys. 1993, 98 , 4494. 266. Szczepanski, J.; Vala, M. Astrophys. J. 1993, 414 , 179. 267. Hudgins, D. M.; Sandford, S. A.; Allamandola, L. J. J. Phys. Chem. 1994, 98 , 4243. 268. Hudgins, D. M.; Allamandola, L. J. J. Phys. Chem. 1995, 99 , 3033. 269. Hudgins, D. M.; Allamandola, L. J. J. Phys. Chem. 1995, 99 , 8978. 270. Hudgins, D. M.; Sandford, S. A. J. Phys. Chem. 1998, 102 , 329. 271. The McCreery Group Raman Shift Frequency Standards 1997, http://www.chemistry.ohio-state.edu/ ~rmccreer/freqcorr/images/nap.html 08/2005. 272. Sheng, S. J.; Hug, G. Chem. Phys. Lett. 1978, 57 , 168. 273. Szczepanski, J.; Vala, M. Manuscript in preparation. 274. Wang, H. Y.; Brucat, P.; Vala, M. Manuscript in preparation. 275. Wang, H. Y.; Szczepanski, J.; Brucat, P.; Vala, M. Manuscript in preparation.

PAGE 189

171 BIOGRAPHICAL SKETCH Haiyan Wang was born in Changkeng Xi ang, Anxi County, Fujian Province, China, on September 10, 1976. He grew up in Changkeng Xiang and graduated from Chongde High School in July, 1994. Haiyan continued his education at Xiamen University, majoring in chemistry, and receiv ed his Bachelor of Sc ience degree in July, 1998. Then he was accepted into the Gradua te School at Xiamen University with a waiver of entrance examination and started research under the inst ruction of Dr. Lansun Zheng, concentrating on carbon clusters us ing time-of-flight mass spectrometry and theoretical modeling. He ear ned his Master of Science de gree in inorganic chemistry from Xiamen University in July, 2001. Haiy an enrolled in the Department of Chemistry at the University of Florida to pursue a Doctor of Philosophy degree in physical chemistry in August, 2001. From then on, he worked in Dr. Martin ValaÂ’s research group, and dedicated himself to laboratory studies of astroph ysical molecules. Since December, 2001, Haiyan Wang has been married to his wife, Xihong Wu, who is currently a PhD candidate in analytical chem istry instructed by Dr. James Winefordner.