<%BANNER%>

Vibrational Spectroscopy of Astrophysical Species

Permanent Link: http://ufdc.ufl.edu/UFE0024354/00001

Material Information

Title: Vibrational Spectroscopy of Astrophysical Species
Physical Description: 1 online resource (122 p.)
Language: english
Creator: Wang, Yun
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: astrophysical, dft, matrix, metal, pah, vibrational
Chemistry -- Dissertations, Academic -- UF
Genre: Chemistry thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Identifications of the species responsible for the unidentified interstellar infrared (UIR) emission bands and the diffuse interstellar absorption (DIB) bands are the two of the major challenges in astrochemistry today. Polycyclic aromatic hydrocarbons (PAHs) have been proposed as the carriers of both signals. Carbon chain clusters and metals have both been detected in the interstellar medium. In this dissertation, reactions of iron with PAHs, and metals (copper, silver and gold) with carbon clusters were investigated. Matrix isolation spectroscopy coupled with density functional (DFT) calculations have been employed throughout this research. Laser ablated iron atoms and evaporated or sublimed benzene, naphthalene, fluorene, pyrene, or coronene were trapped together in solid Ar at 12K. Neutral Fe(benzene), Fe(benzene)2, Fe(naphthalene), Fe(fluorene), Fe(pyrene) and Fe(coronene) complexes were formed in the experiments and their infrared absorption spectra obtained. Theoretical calculations of the equilibrium geometries, stabilities, and harmonic vibrational frequencies of these complexes have been carried out using density functional theory. The calculations show that the dissociation energies (D0) of neutral Fe(PAH) complexes are substantially smaller than their cationic counterparts, indicating that the neutral complexes are less tightly bonded. Reactions of laser-ablated metal (copper, silver and gold) atoms with carbon clusters were investigated in excess argon matrices. Fourier transform infrared absorption spectra, with the assistance of 13C-isotopic substitution experiments and comparison with theoretical calculation, have led to the identification of near-linear CuC3, AgC3 and AuC3 clusters. Photo-induced isotopic scrambling was observed in the Cu1213C3 clusters and explained via a computed potential energy surface (PES) of this reaction. The mechanism for the photoscrambling is shown to involve the formation of a bicyclic CuC3 isomer. The formation of small metal-carbon clusters, CumCn and AgmCn (m=1, 2; n=1-3) were also explored theoretically.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Yun Wang.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Vala, Martin T.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024354:00001

Permanent Link: http://ufdc.ufl.edu/UFE0024354/00001

Material Information

Title: Vibrational Spectroscopy of Astrophysical Species
Physical Description: 1 online resource (122 p.)
Language: english
Creator: Wang, Yun
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: astrophysical, dft, matrix, metal, pah, vibrational
Chemistry -- Dissertations, Academic -- UF
Genre: Chemistry thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Identifications of the species responsible for the unidentified interstellar infrared (UIR) emission bands and the diffuse interstellar absorption (DIB) bands are the two of the major challenges in astrochemistry today. Polycyclic aromatic hydrocarbons (PAHs) have been proposed as the carriers of both signals. Carbon chain clusters and metals have both been detected in the interstellar medium. In this dissertation, reactions of iron with PAHs, and metals (copper, silver and gold) with carbon clusters were investigated. Matrix isolation spectroscopy coupled with density functional (DFT) calculations have been employed throughout this research. Laser ablated iron atoms and evaporated or sublimed benzene, naphthalene, fluorene, pyrene, or coronene were trapped together in solid Ar at 12K. Neutral Fe(benzene), Fe(benzene)2, Fe(naphthalene), Fe(fluorene), Fe(pyrene) and Fe(coronene) complexes were formed in the experiments and their infrared absorption spectra obtained. Theoretical calculations of the equilibrium geometries, stabilities, and harmonic vibrational frequencies of these complexes have been carried out using density functional theory. The calculations show that the dissociation energies (D0) of neutral Fe(PAH) complexes are substantially smaller than their cationic counterparts, indicating that the neutral complexes are less tightly bonded. Reactions of laser-ablated metal (copper, silver and gold) atoms with carbon clusters were investigated in excess argon matrices. Fourier transform infrared absorption spectra, with the assistance of 13C-isotopic substitution experiments and comparison with theoretical calculation, have led to the identification of near-linear CuC3, AgC3 and AuC3 clusters. Photo-induced isotopic scrambling was observed in the Cu1213C3 clusters and explained via a computed potential energy surface (PES) of this reaction. The mechanism for the photoscrambling is shown to involve the formation of a bicyclic CuC3 isomer. The formation of small metal-carbon clusters, CumCn and AgmCn (m=1, 2; n=1-3) were also explored theoretically.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Yun Wang.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Vala, Martin T.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024354:00001


This item has the following downloads:


Full Text

PAGE 1

1 VIBRATIONAL SPECTROSCOPY OF ASTROPHYSICAL SPECIES By YUN WANG A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2009

PAGE 2

2 2009 Yun Wang

PAGE 3

3 To my parents

PAGE 4

4 ACKNOWLEDGMENTS First and foremost, I would like to express my sincerest gratitude to my supervisor, Dr. Martin Vala, who has supported me thr oughout my graduate work at UF. He has been incredibly helpful, knowledgeable, and has been very patient allowing me to work in my own way. His attitudes to res earch and teaching have benefited me immensely. Were it not for his hard work and support I coul d never have accomplished what I have. I am really lucky and honored to be his last student. I especially thank Dr. Jan Szczepanski for his selfless aid, patient instruction and helpful discussions. He not only guided me on many topics of science but also shared his joy of life with me. Many results presented in this dissertation are a consequence of efficient teamwork between Jan and me. My thank fulnes s also goes to my Ph. D. committee members, Dr. John Eyler, Dr. So Hirata, Dr. David Powell and Dr. Char les Telesco. Your ambition for science and passion for research have intrigued my enthusiasm for scientific exploration. I would like to thank my parents for supporting me throughout my life. They let me chase my dream, even when it led me further and furt her from home. Nonetheless, their support was endless and, as such, so is my gratitude. I also acknowledge all my friends, for their support, understanding and encouragement throughout the past years.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................................... 4 LIST OF TABLES ................................................................................................................................ 8 LIST OF FIGURES ............................................................................................................................ 10 ABSTR ACT ........................................................................................................................................ 14 CHAPTER 1 ASTROPHYSICAL BACKGROUND ...................................................................................... 16 The Interstellar Medium ............................................................................................................. 16 Unidentified Interstellar Infrared Emission Bands ................................................................... 19 Diffuse Interstellar Absorption Bands ....................................................................................... 23 Metal Depletion in the ISM ........................................................................................................ 25 2 MATRIX ISOLATION SPECTROSCOPY METHOD ........................................................... 27 Fundamentals of Matrix Isolation Spectroscopy ....................................................................... 27 Experimental Method .................................................................................................................. 28 Experimental Setup .............................................................................................................. 28 Laser Ablation ...................................................................................................................... 31 Generation of PAHs ............................................................................................................. 31 3 VIBRATIONAL SPECTROSCOPY OF NEUTRAL COMPLEXES OF IRON AND POLYCYCLIC AROMATIC HYDROCARBONS ................................................................. 32 Introduction ................................................................................................................................. 32 Computational and Experimental Details .................................................................................. 34 Computational Details ......................................................................................................... 34 Experimental Details ........................................................................................................... 35 Fe(benzene) and Fe(benzene)2 complexes ................................................................................. 36 Fe(naphthalene) complexes ........................................................................................................ 44 Fe(fluorene) complexes .............................................................................................................. 48 Fe(pyrene) complexes ................................................................................................................. 50 Fe(coron ene) complexes ............................................................................................................. 53 Effect of complexation on IR spectra ........................................................................................ 55 Summary ...................................................................................................................................... 57

PAGE 6

6 4 COPPER CARBON CLUSTERS: STRUCTURE, INFRARED FREQUENCIES AND ISOTOPIC SCRAMBLING ....................................................................................................... 59 Introduction ................................................................................................................................. 59 Computational and Exp erimental Details .................................................................................. 60 Experimental methods ......................................................................................................... 60 Computational methods ....................................................................................................... 61 Experimental Infrared Spectra .................................................................................................... 62 Equilibrium geometries and vibrations for the Cum Cn (m =1,2, n=1, 2, 3) clusters .............. 65 The CuC Copper Carbon Cluster ....................................................................................... 68 The CuC2 Copper Carbon Cluster ...................................................................................... 6 8 The Cu2C Copper Carbon Cluster ...................................................................................... 68 The Cu2C2 Copper Carbon Cluster ..................................................................................... 68 The CuC3 Copper Carbon Cluster ...................................................................................... 69 The Cu2C3 Copper Carbon Cl uster ..................................................................................... 72 12/13C Isotope Scrambling in the Near Linear CuC3 Clusters ................................................... 72 Longer CuCn (n = 4 9) clusters ................................................................................................ 76 Summary ...................................................................................................................................... 80 5 SILVER CARBON CLUSTER: STRUCTURE AND INFRARED FREQUENCIES ......... 82 Introductio n ................................................................................................................................. 82 Computational and Experimental Details .................................................................................. 83 Experimental Methods ........................................................................................................ 83 Computational Methods ...................................................................................................... 84 Experimental Infrared Spectra .................................................................................................... 85 Equilibrium Geometries and Vibrations for AgmCn (m=1, 2; n=1 3) Clusters. ...................... 87 The AgC Silver Carbon Cluster ......................................................................................... 88 The AgC2 Silver Carbon Cluster ........................................................................................ 89 The AgC3 Silver Carbon Cluster ........................................................................................ 89 The Ag2C Silver Carbon Cluster ........................................................................................ 92 The Ag2C2 Silver -Carbon Cluster ....................................................................................... 92 The Ag2C3 Silver -Carbon Cluster ....................................................................................... 92 Summary ...................................................................................................................................... 93 6 GOLD CARBON CLUSTER: STR UCTURE AND INFRARED FREQUENCIES ............ 94 Introduction ................................................................................................................................. 94 Computational and Experimental Details .................................................................................. 95 Experimental Methods ........................................................................................................ 95 Computational Methods ...................................................................................................... 95 Experimental Infrared Spectra ................................................................ .................................... 97 Theoretical Calculations for Gold Carbon AuC3 Cluster ....................................................... 101 Summ a ry .................................................................................................................................... 106

PAGE 7

7 7 CONCLUSI ONS AND FUTURE WORK .............................................................................. 108 Iron and PAH Complexes ......................................................................................................... 108 Metal Carbon Clusters .............................................................................................................. 108 Future Work ............................................................................................................................... 110 LIST OF REFERENCES ................................................................................................................. 113 BIOGRAPHICAL SKETCH ........................................................................................................... 122

PAGE 8

8 LIST OF TABLES Ta ble page 1 1 List of identified interstellar molecules ................................................................................ 18 1 2 Astrophysically abundant elements and their depletion fac tors.43 ...................................... 26 3 1 Computed properties for Fe(PAH) complexes ..................................................................... 39 3 2 IR absorption spectra of Fe(C6H6) and Fe(C6D6) complexes. ............................................. 41 3 3 IR absorption spectra of Fe(C6H6)2 and Fe(C6D6)2 complexes. .......................................... 41 3 4 Comparison of present work with previous IR band assignments for Fe(C6H6) complex. .................................................................................................................................. 42 3 5 Comparison of present work with previous band assignments for Fe(C6H6)2 complex. ... 43 3 6 IR absorpt ion spectra of Fe(C10H8) and Fe(C10D8) iron(naphthalene) complexes. ........... 47 3 7 IR absorption spectra of Fe(C13H10) iron(fluorene) complex. ............................................. 48 3 8 IR absorption spectra of Fe(C16H10) iron(pyrene) complex (isomer F from Figure 31). ............................................................................................................................................. 52 3 9 IR absorption spectra of Fe(C24H12) iron (coronene) complex. .......................................... 55 4 1 Vibrational frequencies (cm1) and integral intensities (km/mol) for electronic ground states of CumCn (m = 1, 2; n = 1, 2, 3) clusters (displayed in Figure 4 3), calculated using B3LYP and MPW1PW91 functionals ....................................................................... 66 4 2 Calculated total energies, EZPE (Hartrees), corrected for zero point vibrational energies, and estimated dissociation energies, D0 (eV) or isomerization energy barriers, EIso (eV) for CumCn (n = 1, 2; m = 1, 2, 3) isomers. .............................................. 67 4 3 Experimental (Ar matrix, 12K) and calculated isotopomer frequencies (integral intensities) for the asymmetric and symmetric CC stretch fundamenta l modes of fully optimized equilibrium geometry of near linear 63Cu12/13C ( G, Figure 4 3). Proposed band assignments marked in Figure 4 2 are given in the first column. .............................. 71 4 4 Calculated (at M PW1PW91/6311++G(3df)) harmonic vibrational frequencies and their integral intensities (in parentheses) for the CuCn (n = 4 carbon clusters (displayed in Fig. 4 5). ............................................................................................. 77 5 1 Vibrational Frequencies (cm1) and Integral Intensities (km/mol) for Electronic Ground States of AgmCn (m = 1, 2; n = 1, 2, 3) Clusters (displayed in Figure 53), Calculated Using MPW1PW91/SDD Functional/basis sets. ............................................... 89

PAGE 9

9 5 2 Comparison of Experimental and Calculated Isotopomer Frequencies (Integral Intensities) for the Asymmetric C=C Stretch and Symmetric C=C Stretch Fundamental Modes of Near Linear 107Ag 12/13C3. ............................................................... 91 6 1 Vibrational frequencies (cm1) and integral intensities (km/mol) for electronic grou nd states of near linear AuC3 clusters (displayed in Figure 6 4), calculated using different functional/basis sets. ............................................................................................. 102 6 2 Comparison of Experimental and Calculated (at BPW91/SDD and BPW91/La nL2DZ level)Isotopomer Frequencies (Integral Intensities) for the Asymmetric C=C Stretch and Symmetric C=C Stretch Fundamental Modes of Near Linear 197Au 12/13C. ............................................................................................................... 103 6 3 Comparison of Experime ntal and Calculated (at MPW1PW91/SDD, MPW1PW91/SDDAll and MPW1PW91/LanL2DZ level) Isotopomer Frequencies (Integral Intensities) for the Asymmetric C=C Stretch and Symmetric C=C Stretch Fundamental Modes of Near Linear 197Au 12/13C. .............................................................. 104 6 4 Comparison of Experimental and Calculated (at MPW1PW91/6 311++G(3df)/SDD and MPW1PW91/6311++G(3df)/LanL2DZ level) Isotopomer Frequencies (Integral Intensities) for the Asymmetric C=C Stretch and Symmetric C= C Stretch Fundamental Modes of Near Linear 197Au 12/13C. .............................................................. 105 6 5 Comparison of Experimental and Calculated (at BPW91/6 311++G(3df)/SDD (peudopotential) and B3LYP/6 311++G(3df)/LanL2DZ level) I sotopomer Frequencies (Integral Intensities) for the Asymmetric C=C Stretch and Symmetric C=C Stretch Fundamental Modes of Near Linear 197Au 12/13C. ........................................ 106

PAGE 10

10 LIST OF FIGURES Figure page 1 1 The emission spectra of the post -AGB objects, IRAS 162794757, the Red Rectangle, and the planetary nebula, NGC 7027.16 ............................................................. 19 1 2 Comparison of UIR bands wit h PAH model. Top: (a) emission spectrum from the proto planetary nebula IRAS 22272+5435 compared with the (b) absorption spectrum produced from a mixture of neutral and cationic PAHs. Bottom: the (c) emission spectrum from the Orion ionization ridge compar ed with the (d) absorption spectrum produced from a mixture of fully ionized PAHs.20 .............................................. 21 1 3 Plot of the electronic transition wavelength for even-numbered carbon cluster anions observed in an Ar matrix at 36 K versus the number of carbons in the cluster chain.21.... 23 1 4 The synthetic spectrum of the hot B0 II star BD+63 1964.32............................................. 24 1 5 The elemental composition of the solar system. The abundance of hydrogen is arbitrarily set to 1012 so that the smallest abundance in the graph is about 1.42 ................. 25 2 1 Illustrati on of the principle of matrix isolation. The inert host matrix isolates guest particles from each other and prevents reaction. .................................................................. 28 2 2 Experimental setup for Fe -PAH experiments. ...................................................................... 29 2 3 Experimental setup for copper -carbon experiments. ........................................................... 30 3 1 Lowest energy stable structures for the complexes of iron with benzene (C6H6)(A ), bis -benzene (C6H6)2 (B), naphthalene (C10H8) (C ), fluorene (C13H10) ( D ), pyrene (C16H10) ( E), pyrene (second isomer, F ), and coronene ( G ), all optimized at the MPW1PW91/6 31+G(d, p) level of theory. ......................................................................... 36 3 2 IR absorption spectrum of benzene (C6H6) only (a), Fe codeposited with benzene (b), after matrix UV -visible photolysis (c) and after matrix annealing at 35 K (d), all in solid Ar at 12K. ...................................................................................................................... 37 3 3 Synthetic experimental and calculated infrared absorption spectra at indicated spin state multiplicities for the Fe(C6H6) complex. ..................................................................... 38 3 4 Synthetic experimental and calculat ed infrared absorption spectra for the Fe(C6H6)2 complex. The inaccessible energy region overlapped with the absorption band of benzene is marked by the horizontal line (a). ....................................................................... 40 3 5 IR absorpti on spectrum of: naphthalene (C10H8) only (a), Fe codeposited with naphthalene (b), after matrix UV visible photolysis (c), all in solid Ar at 12K. The star marked band at 873.1 cm1 is unassigned. ..................................................................... 44

PAGE 11

11 3 6 Calculated IR spectrum of Fe(C10D8) (a), experimental IR absorption spectrum of naphthalene -d8 (C10D8) only (b), experimental IR absorption spectrum of Fe codeposited with naphthalene -d8 (c), experimental IR absorption spectrum of Fe codeposite d with naphthalene -d8 after matrix UV -visible photolysis (d), all in solid Ar at 12 K. .............................................................................................................................. 45 3 7 Synthetic experimental and calculated infrared absorption spectra with different spin multipliciti es for the Fe(naphthalene) complex. ................................................................... 46 3 8 Calculated IR spectrum of Fe(C13H10) (a), experimental IR absorption spectrum of fluorene (C13H10) only (b), experimental IR absorption spectrum of F e codeposited with fluorene (c), experimental IR absorption spectrum of Fe codeposited with fluorene after matrix UV -visible photolysis (d), all in solid Ar at 12K. ............................. 50 3 9 Calculated IR spectr um of Fe(C16H10) (structure E) (a), calculated IR spectrum of Fe(C16H10) (structure F ) (b), experimental IR absorption spectrum of pyrene (C16H10) only (c), experimental IR absorption spectrum of Fe codeposited with pyrene (d), experimental IR absorption s pectrum of Fe codeposited with pyrene after matrix UV visible photolysis (e), all in solid Ar at 12K plotted in two energy regions [a] and [b]. ... 51 3 10 Calculated IR spectrum of Fe(C24H12) (a), experimental IR absorption spectrum of coronene (C24H12) only (b), experimental IR absorption spectrum of Fe codeposited with coronene (c), experimental IR absorption spectrum of Fe codeposited with coronene after matrix UV -visible photolysis (d), all in solid Ar at 12K. ........................... 54 3 11 Comparison of summed IR absorption spectra calculated at the MPW1PW91/631+G(d,p) level for a 1:1:1:1 mixture of naphthalene, fluorene, pyrene and coronene (PAH panel ), for their complexes with iron (Fe(PAH) panel), for similar mixture of cations (PAH+ panel) and for their cationic complexes with iron (Fe(PAH)+ panel). We note that the PAH, PAH+ and Fe(PAH)+ spectra are prepared at similar way to those of ref. 72,72 bu t for comparison purpose include also spectra of coronene, coronene+ and Fe(coronene)+. ............................................................................................... 56 4 1 Infrared absorption spectra of products of laser ablation of graphite (12C (99%) + 13C (1%)) (spe ctrum a) and products of two beam laser ablation of graphite and copper (spectrum b, enlarged twofold). The spectra were recorded after matrix annealing to 35 K then cooling back to 12 K. The major bands due to pure carbon clusters and their reaction produ cts with copper at 1830 and 1250.5 cm1 are indicated. Carbon monoxide and its product with copper are also marked. ..................................................... 63 4 2 Infrared spectra of reaction products from laser ablation of Cu and 12C (spectrum a) displayed in two energy regions. The bands at 1830.0 and 1250.5 cm1 are due to a common carrier containing Cu and 12C. Spectrum b was collected from runs similar to spectrum a, but with a 13C -enriched sample. The band carriers marked by ve rtical dashed lines are due to: C3H (1824.4 cm1),100,101 C2H+ (1820.2 cm1),102 and C12 (1817.9 cm1).103 The fractionations of b d as well as of e g isotopomers via the proposed 12/13C isotopic scrambling in nl Cu12/13C3 (see text) are marked. ....................... 64

PAGE 12

12 4 3 Optimized equilibrium structures for the CuC, CuC2, Cu2C, Cu2C2, CuC3, and Cu2C3 clusters. The bond lengths () and angles () calculated at B3LYP/6 311++G(3df) (italic type, top) and at MPW1PW91/6 311++G(3df) (normal type) are marked. The relative isomer energies are indicated. .................................................................................. 65 4 4 Total energy plot for the near linear CuC3 (2A 3 (2A 2A /2A1) clusters calculated at the MPW1PW91/6 311++G(3df) level by incrementing the (CCC) angle in the range of 77 the four remaining geometrical parameters were full y optimized. Note that the 63 121213 cluster reactant (left) rearranges to the 63 13 1212 lower energy isotopomer product by breaking the 12C Cu bond in structure H then passing over the TS and relaxing to the G product (right) in the forward reaction pathway. ................. 75 4 5 Optimized equilibrium structures for the CuCn (n = 4 9) clusters. The bond lengths () and angles () calculated at MPW1PW91/6 311++G(3df) are marked. ...................... 78 5 1 Infrared absorption spectra of products of laser ablation of graphite (spectrum a) and products of two -beam laser ablation of graphite and silver (spectrum b). The spectra were recorded after matrix annealing to 35K then cooling back to 12 K. The major bands due to pure carbon clusters and their reaction products with silver at 1827.8 and 1231.6 cm1 are indicated. ............................................................................................... 86 5 2 Infrared spectra of react ion products from laser ablation of Ag and 12C (spectrum a) and from laser ablation of Ag and 12/13C (spectrum b). The bands marked by vertical dashed lines are tentatively assigned to isotopomeric partners of the 1231.6 cm1 band. The bands marked by vert ical dotted lines are due to: C3H (1824.4 cm1),100,101 C2H+ (1820.2 cm1),102 C12 (1817.9 cm1),103 and C6 (1197.2 cm1).134 ............................... 87 5 3 Equilibrium structures for the AgC, AgC2, AgC3, Ag2C, Ag2C2, and Ag2C3 clusters. The bond lengths () and angles () calculated at MPW1PW91/SDD are marked. The relative isomer energies are indicated. .......................................................................... 88 6 1 Infrared absorption spectra of products of l aser ablation of graphite (spectrum a) and products of two -beam laser ablation of graphite and gold (spectrum b). The spectra were recorded after matrix annealing to 35K then cooling back to 12 K. The major bands due to pure carbon clusters and their react ion products with gold at 1845.2 and 1275.7 cm1 are indicated. ...................................................................................................... 98 6 2 Infrared spectra of reaction products from laser ablation of Au and 12C (spectrum a) and from laser ablation of Au and 12 /13C (spectrum b). The bands marked by vertical dashed and dotted lines are tentatively assigned to isotopomeric partners of the 1275.7 cm1 band. The bands marked by vertical dotted lines are due to: C3H (1824.4 cm1),100,101 C12 (1817.9 cm1),103 ........................................................................................... 99 6 3 Infrared spectra of reaction products from laser ablation of Cu and 12/13C (spectrum a), Ag and 12/13C (spectrum b) and Au and 12/13C (spectrum c) in 17501850 cm1

PAGE 13

13 region. The bands marked ar e assigned to isotopomeric bands for Cu12/13C3, Ag12/13C3 and Au12/13C3. ....................................................................................................... 100 6 4 Optimized equilibrium structure for the AuC3 cluster. The bond lengths () and angles () calculated at MPW1PW9 1/LanL2DZ (italic type, top) and at MPW1PW91/SDD (normal type) are marked. ................................................................... 105 7 1 Experimental Setup for Ion Mass Selection and Trapping. ............................................... 111

PAGE 14

14 Abstract of Di ssertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy VIBRATIONAL SPECTROSCOPY OF ASTROPHISICAL SPECIES By Yun Wang May 2009 Chair: Mart in Vala Major: Chemistry Identifications of the species responsible for the unidentified interstellar infrared (UIR) emission bands and the diffuse interstellar absorption (DIB) bands are the two of the major challenges in astro chemistry today. Polycyclic aromatic hydrocarbons (PAHs) have been proposed as the carriers of both signals. Carbon chain clusters and metals have both been detected in the interstellar medium. In this dissertation reaction s of iron with PAHs, and metal s (copper, silver and gold) w ith carbon clusters were investigated. Matrix isolation spectroscopy coupled with density functional (DFT) calculations have been employed throughout this research Laser ablated iron atoms and evaporated or sublimed benzene, naphthalene, fluorene, pyrene or coronene were trapped together in solid Ar at 12K. Neutral Fe(benzene), Fe(benzene)2, Fe(naphthalene), Fe(fluorene), Fe(pyrene) and Fe(coronene) complexes were formed in the experiments and t heir infrared absorption spectra obtained. Theoretical calcu lations of the equilibrium geometries, stabilities, and harmonic vibrational frequencies of these complexes have been carried out using density functional theory. The calculations show that the dissociation energies (D0) of neutral Fe(PAH) complexes are su bstantially smaller than their cationic counterparts indicating that the neutral complexes are less tightly bonded.

PAGE 15

15 Reactions of laser ablated metal (copper, silver and gold) atoms with carbon clusters were investigated in excess argon matri ces Fourier transform infrared absorption spectra, with the assistance of 13C isotopic substitution experiments and comparison with theoretical calculation, have led to the identification of near linear CuC3, AgC3 and AuC3 clusters. Photoinduced isotopic scrambling w as observed in the Cu1213C3 clusters and explained via a computed potential energy surface (PES) of this reaction. The mechanism for the photoscrambling is shown to involve the formation of a bicyclic CuC3 isomer. The formation of small metal -carbon cluste rs, CumCn and AgmCn (m=1, 2; n=1 3) were also explored theoretically.

PAGE 16

16 CHAPTER 1 ASTROPHYSICAL BACKGR OUN D For decades, astronomers have been intrigued and puzzled by what exist s in the extraterrestrial space. There has been accumulation of evidences that the molecules necessary for the evolution of life on earth were formed in the depths of space.1 To understand the chemistry of interstellar species and ultimately, the origin of life we must first identify the molecules found in space Determining the chemical composition of the materials in the space, especially the interstellar medium (ISM) has therefore become an active area of research in the field of astrochemistry. Th e molecules found so far have been determined by either radio astronomy infrared spectroscopy or visible/ultraviolet spe ctroscopy.2 Of primary interest to spectroscopists are the signals from the interstellar space, called the unidentified interstellar infrared emission bands (UIRs)35 and the diffuse interstellar absorption bands (DIBs).6,7 Identification of the carriers of the UIRs and DIBs is still a major challenge for spectroscopists and astrochemists despite considerable efort The I nterstellar M edium O uter s pace is not a complete vacuum. T he material which occupie s the space between the stars is called the interstellar medium (ISM) The ISM contains ordinary matter, cosmic rays and magnetic fields Approximately 99% (by mass) of the interstellar medium is mad e up of gas and the remaining 1% i s dust.8 The interstellar gas consists of neutral atoms and molecules as well as ions and electrons. Hydrogen and helium are the major components of interstellar gas, which constitute about 75% and 24% of the total interstellar mass. Other chemical elements, such as carbon, nitrogen, oxygen and some heavier elements (metals) make u p the remaining 1%.2 The density of the interstella r gas is very low, with an average density of 1 atom per cubic centimeter, which is comparable to the best vacuum that we can achieve in laboratory. Although the

PAGE 17

17 interstellar medium is extremely dilute it still accounts for 20 30% of the mass of our galaxy because the amount adds up over the vast distance between stars .1 The interstellar medium is a dynamic system: old stars die and spew matter and energy back in to the ISM Eve ntually, new stars form from the recycled interstellar material. The interstellar medium is not homogeneous. A wide range of densities and temperatures co -exist in the interstellar space. Densities denoted by the number of hydrogen nuclei nH, vary from 103 to 108 cm3; and temperatures found in different regions can range from 10 K up to 107 K.9 The interstellar medium contains three different phases defined by their densities and temperatures : the cold molecular clouds warm diffuse clouds an d the hot ionized gas .10 The cold molecular clouds (T ~ 10 100 K) are also called dark clouds because they block the lights from background stars. The densities of t he molecular clouds range from 102 to 106 cm3. Therefore, t he cold molecular clouds contribute roughly half of the interstellar mass although they occupy only a small portion (~1 2%) of the space The diffuse clouds consist of cold atomic gas (T ~ 100 K) and are almost transparent to background starlight except at some specific absorption lines. They have moderate densities (nH ~ 101 cm3) and temperatures (T ~ 104 K) and therefore provide ideal sightlines for absorption measurements in the ultraviolet an d visible ( UV/Vis) region against background stars. The hot ionized clouds are very diluted, with densities only 103 cm3 and temperatures as high as 106 K .10 Most molecules in the ISM have been identified via radio or microwave methods. However, this technique can only detect molecules with permanent dipole moments. Non -polar m olecules, such as H2, C5 and C6H6, were ide ntified using infrared, UV/Vis spectroscopy.11 The first molecule that has been detected in the ISM was the methylidyne radical (CH) in 1937.12 To date, over 140 molecules have been identified in the ISM with number of atoms ranging from 2

PAGE 18

18 to 13.11 Gly cine, the simplest amino acid necessary for life, was claimed to be observed in the interstellar medium in 2003,13 but its discovery was accompanied by some controversy.14,15 The molecules identified so far in the interstellar medium are listed in Table 1 1. Over 75% of these interstellar molecules are carbon -bearing species. Table 1 1. List of identified interstellar molecules Number of atoms Molecular Formula 2 H 2 CO CSi CP CS NO NS SO HCl NaCl KCl AlCl AlF PN SiN Si O SiS NH OH C 2 CN HF FeO LiH CH CH + CO + SO + SH O 2 N 2 CF+ 3 H 2 O H 2 S HCN HNC CO 2 SO 2 MgCN MgNC NaCN N 2 O NH 2 OCS CH 2 HCO C 3 C 2 H C 2 O C 2 S AlNC HNO SiCN N 2 H + SiNC c SiC 2 HCO + HOC + HCS + H 3 + OCN HCP CCP 4 NH 3 H 2 CO H 2 CS C 2 H 2 HNCO HNCS H 3 O + SiC 3 C 3 S H 2 CN c C3H l C3H HCCN CH 3 C 2 CN C 3 O HCNH + HOCO + C 3 N 5 CH 4 SiH 4 CH 2 NH NH 2 CN CH 2 CO HCOOH, HCC -CN HCC NC c C 3 H 2 l -C 3 H 2 CH 2 CN H 2 COH+, C 4 Si C 5 HNCCC C 4 H C 4 H-, HCOCN 6 CH 3 OH CH 3 SH C 2 H 4 HC 4 H CH 3 CN CH 3 NC HCONH 2 HCC COH HC 3 NH + HC 4 N C 5 N C 5 H H 2 CCCC H 2 CCNH c H 2 C 3 O 7 CH 2 CHOH c C 2 H 4 O HCOCH 3 H 3 C CC H CH 3 NH 2 CH 2 CHCN HCC CC CN C 6 H C 6 H 8 CH 3 COOH HCOOCH 3 HOCH 2 COH H 3 C CC CN H 2 C 6 HC 6 H C 7 H H 2 C=CH COH CH 2 CCHCN H 2 NCH 2 CN 9 (CH 3 ) 2 O CH 3 CH 2 CN CH 3 CH 2 OH CH 3 C 4 H HCC CC CC CN C 8 H CH 3 CONH 2 C 8 H CH 3 CHCH 2 10 (CH 3 ) 2 CO HOCH 2 CH 2 OH H 3 C CH 2 COH CH 3 C 4 CN 11 HC 9 N CH 3 C 6 H 12 C 6 H 6 13 HC 11 N

PAGE 19

19 Unidentified Interstellar I nfrared E mission B ands The unidentified interste llar infrared emission bands (UIRs) were discovered in 1973.3 They are a series of emission bands observed from a variety of objects in space, including planetary nebulae, H II regions, post asymptotic giant branch (AGB) stars, and some regions of the Milky Way and other galaxies All of the above regions contain interstellar matter and are full of ultraviolet radiation. Figure 1 1 shows the infrared spectra of three different objects in the interstellar medium.16 Figure 1 1. The emission spectra of the post -AGB objects, IRAS 162794757, the Red Rectangle, and the planetary nebula, NGC 7027.16

PAGE 20

20 Although the intensity distribution of each individual bands is different for spectra from a variety of interstellar environments, the band positions are very similar; they all have dominant infrared emission features at 3.3, 6.2, 7.7, 8.7 (3030, 1610, 1280, 1150 and 890 cm1).35 These features could no t be matched to any single or mixture of atomic lines. The identity of the carriers of these interstellar bands has become a topic of active and long-standing interest in the astrochemistry community Polycyclic aromatic hydrocarbons, also known as PAHs, are believed to be present in the interstellar medium and have been proposed as the carriers of the UIR bands. The proposal that the UIR bands are originated from gas -phase, neutral PAHs was first suggested by Leger and Puget in 1984.17 Shortly thereafter, Allamandola et al. argued that PAH cations c ould be the source of these bands.18 Later i t was proposed that the UIR bands cannot be explained solely on the basis of neutral PAH species, but that cation ic PAHs are a significant component. I n other words, a mixture of neutral and ionic PAHs is responsible for the interstellar bands.19,20 Figure 1 2 shows that the laboratory infrared spectra of a mixture of neutral and cationic PAHs match the ob served UIR bands quite well.20 The PAH model entails seve ral major parts : first, the interstellar emission is due to infrared fluorescence from gas phase molecules excited by the absorption of single ultraviolet and visible photons instead of thermal emission from a solid material ;17 seco nd, the carriers are carbon rich because the fraction of the total infrared energy that is emitted through these features is closely related to the density of carbon; third, the carriers must be stable enough to survive under the very harsh environment of the interstellar medium; also the relative band intensities can be different but the features are correlated, which indicates that the carriers belong to a single class of chemical species because the vibrational transitions, such as C -C or C -H stretchin g or

PAGE 21

21 bending modes, are similar for most PAH molecules ; and last but not least, the proposed carriers have an infrared spectrum that match es well with the UIR bands both in positions and relative band intensities. The PAH hypothesis is now widely accept ed. However, the exact mixture of PAHs and the mechanism of formation of interstellar PAHs are still unknown. Figure 1 2. Comparison of UIR bands with PAH model. Top: (a) emission spectrum from the proto planetary nebula IRAS 22272+ 5435 compared with th e (b) absorption spectrum produced from a mixture of neutral and cationic PAHs Bottom: the (c) emission spectrum from the Orion ionization ridge compared with the (d) absorption spectrum produced from a mixture of fully ionized PAHs.20 Another Hypothesis is that long chain carbon clusters should also be conside red as the carriers of the UIR bands.21 Although researchers suggests that only a millionth of the cosmically

PAGE 22

22 available carbon exist in the form of carbon chains,22 small carbon clusters such as C3 and C5 have still been detected in the circumstellar cloud of the carbon star IRC+10216 using IR detection methods.23,24 The emission mechanism of the long carbon chain model is fundamentally different from the PAH model. It states that after UV or visible absorption into electronic excited states the excess energy in the carbon chain internally converts to the lowest lying exited electronic state. Then electronic emission from this excited state to the ground state occurs This differs from the PAH model in that the emission is electronic in nature, not vibrational. The length of the carbon chain dictates where in the infrared region the emission takes place. Figure 1 3 shows the position of the 0 0 transitions of some even Cn (18 < n < 36) carbon cluster bands coincide with some of the UIR bands.21 For instance, the band a ssigned to C24 and C36 lies at 3.2787 and 6.1 respectively, which roughly fit with the broad UIR band at 3.29 and 6.2 Extension of the plot in Figure 13 indicates that C24 and C52 could be responsible for the 7.8 and 11.3 UIR bands. Th e long carbon chain model has been challenged on the grounds that carbon chain anions might not form and survive under the harsh conditions of the interstellar environment.25,26 H owever, the negative molecular ion s C6Hand C8Hhave been identified in the Galactic molecular source TMC 1 very recently .2729 C10Hwas calculated and anticipated to be present at comparable abundance to C6Hand C8H-.30 Even larger anions could possibly be observed because the abundance is expected to change little as the length of the carbon chain increases. Therefore, the long carbon chain model should gain more attention in the future.

PAGE 23

23 Figure 1 3. Plot of the electronic transition wavelength for even numbered carbon cluster anions observed in an Ar matrix at 36 K versus the number of carbons in the cluster chain.21 Diffuse Interstellar Absorption Bands The diffuse interstellar absorption bands (DIBs) are a series of absorption bands that are observed in the spectral range extending from the UV to the near IR (400 1300 nm). First discovered in 1922, the DIB bands have become one of the longest -standing astronomical spec troscopic mysteries .31 A synthetic spectrum of DIBs towards the hot B0 II star BD+63 1964 is displayed in figure 1 4 .32 This object is now commonly used as a reference target in the search of new DIBs. DIB bands exhibit a large diversity of band profiles : the full bandwidth at half maximum (FWHM) range f rom 0.06 to 4 nm;33 and the intensities of most DIB bands are not correlated to each other, which means there must be many carriers rather than just one carrier. To date, over 300 DIBs have been detected in the interstellar medium, but no definitive

PAGE 24

24 identification of the carriers of these bands has been reported yet Carbon clusters or chains,34,35 carbon rings or fullerenes,36,37 and PAHs have all been suggested as potential carriers for some DIB bands.25,38 However, some recent research indicates that PAH s having 30 or more carbon atoms especially in their cationic forms, are the strongest candidates as DIB carriers.39 Since such large PAHs are experimentally difficu lt to vaporize and keep in the gas phase as cations, usually solid -state spectroscopy is employed in studying the spectra of these PAH cations .40,41 However, only gas -phase spectroscopy can provide the information of specific molecules that are respon sible for the DIB bands. Figure 1 4. The synthetic spectrum of the hot B0 II star BD+63 1964.32

PAGE 25

25 Metal Depletion in the ISM Other than the most abundant elements hydrogen (H) and helium (He), there are many heavier elements present in the interstellar medium, such as carbon (C), nitrogen (N), oxygen (O), silicon (Si), iron (Fe) and copper (Cu). Figure 1 5. shows the elemental composition of our solar system.42 The heavier elements were produced through stellar nucleosynthesis. Figure 1 5. The elemental composition of the solar system. The abundance of hydrogen is arbitrarily set to 1012 so that the smallest abundance in the graph is about 1.42 Since the interstellar medium is quite well mixed, it is expected that the abundances of these interstellar element is similar to th e cosmic abundance in the Sun s atmosphere or in meteorites. However, observations of interstellar absorption lines show that a number of elements are depleted from the gas -phase by large fact ors. Table 1 2. lists the abundance and depletion factors of some interstellar elements. The depletion factor for a certain element (i) is defined relative to its cosmic abundances by ] ) / /( ) / [(s H i H i iN N N N where iN is the average number of the element, HN is the average number of hydrogen atoms and the notation s refers to the numbers in our solar system.43 Th e abundance relative to H in Table 1 2 w as arbitrarily cut off at 108. The abundances of the heavy elements relative to H add up to 1.60 x 103. The depletion factors vary from around 0.5 (C, O), to values around 0.1 (Cu), 0.01 (Fe) or even lower to ar ound 0.002 (Ca, Ti). Also, depletion tends to be more severe in regions with

PAGE 26

26 higher density and lower temperature.44 It has been proposed that the high depletion of interstellar iron could be due to iron complexes with various interstellar PAHs.45,46 Table 1 2. Astrophysically abundant elements and their deplet ion factors.43 Element Abundance Depletion Factor Element Abundance Depletion Factor H 1.000 1.0 Cl 1.86 ( 7) 0.5 He 0.0977 1.0 Ar 3.63 ( 6) 1.0 C 3.63 ( 4) 0.5 K 1.32 ( 7) 0.5 N 1.12 ( 4) 0.6 Ca 2.29 ( 6) 0.003 O 8.51 ( 4) 0.6 Ti 9.77 ( 8) 0.002 F 3.02 ( 8) 0.5 V 1.00 ( 8) ? Ne 1.23 ( 4) ? Cr 4.68 ( 7) 0.03 Na 2.14 ( 6) 0.25 Mn 2.45 ( 7) 0.07 Mg 3.80 ( 5) 0.2 Fe 4.68 ( 5) 0.01 Al 2.95 ( 6) 0.05 Co 8.32 ( 8) ? Si 3.55 ( 5) 0.1 Ni 1.78 ( 6) 0.04 P 2.82 ( 7) 0.6 Cu 1.62 ( 8) 0.1 S 1.62 ( 5) 0.6 Zn 3.98 ( 8) 0.6

PAGE 27

27 CHAPTER 2 MATRIX ISOLATION SPE CTROSCOPY METHOD Matrix isolation spectroscopy and theoretical calculations are two of the most popular a nd powerful tools in the study of astrochemistry. In this dissertation, both two techniques were employed to study the molecules of interest. Matrix isolation spectroscopy technique involved in this work is discussed in this chapter. Details of Density Functional Theory ( DFT ) calculation for each project will be provided in the following chapters individually. Fundamentals of Matrix Isolation Spectroscopy The matrix isolation method was first reported by Whittle, Dows, and Pimentel in 1954.47 Numerous books and articles have been published on this subject since then. The original idea of matrix isolation is to trap some free radicals and other unstable substances in a solid matrix of inert material at very low temperature (4 30 K) so that the matrix will inhibit diffusion of the trapped molecules, thus holding the active molecules effectively immobile in a nonreactive environment.47 The principle of matrix isolation is illustrated in Figure 2 1 The guest partilces can be generated by thermal evaporation, pyrolysis, microwave discharge, photolysis or laser ablation. The matrix materials are usually chemically inert and optically transparent from far infrared to ultraviolet Therefore, full range of IR or electronic spectra can be recorded. For the reasons mentioned a bove some noble gases are often used as matrix, such as argon, neon, xenon because they have no absorptions in the infrared range and they seldom react with other species Nitrogen, although not rare gas, can also be chosen as matrix material Matrix isol ation spectroscopy has become a powerful tool in the study of astrophysical species. The low temperature and high vacuum are two essential conditions for matrix isolation spectroscopy, which can simulate the interstellar environment. Matrix isolation spectroscopy is

PAGE 28

28 also very efficient for studying some unstable species which may possibly exist in the interstellar medium. Host Matrix Guest Particles (Molecules, Atoms, ions, etc) Figure 2 1. Illustration of t he princip le of matrix isolation. The inert host matrix isolates guest particles from each other and preve nts reaction. Experimental Method Experimental Setup Matrix isolation experiments were conducted in a high vacuum chamber at very low temperature. The chamber was inserted into a FT IR spectrometer (MIDAC M2000 or NICOLET Magna 560), allowing spectra colle ction during or after deposition The chamber was pumped by a diffusion pump, which was connected to a rotary pump. Before cooling down, the pressure in the chamber is on the order of 106 torr. After the system is cooled down to about 12 K the pressure d rops to around 107 torr, which is called the background pressure for the experiments. The deposition surface was a cesium iodine (CsI) window mounted in a copper holder, which is cooled down to 12 K by a closed-cycle helium cryostat (APD Cryogenics Inc.). The window temperature was monitored by a thermocouple temperature controller via an irongold thermocouple attached to the top of the copper holder. During deposition, the species of interest in this research was produced using different techniques, such as laser ablation and thermal evaporation. Argon gas or a mixture of argon and other gases was introduced into the chamber and directed towards the deposition window through a stainless steel tube with a needle

PAGE 29

29 valve. The gas flow was controlled by adjust ing the needle valve. The mixture of the reaction products and argon gas were trapped onto the CsI window cooled to 12 K. After 2 5 hours deposition, infrared absorption spectra in 700 4000 cm1 region were collected with 0.5 1.0 cm1 resolution Mat rix annealing to 35 K followed by recooling back to 12 K and matrix photolysis by applying UV -visible medium pressure 100W Hg lamp were also involved in most runs. A diagram of the experimental setup for the Fe(PAH) experiments is shown in Figure 2 2. Figure 2 2. Experimental setup for Fe PAH experiments. For copper -carbon, silver -carbon and goldcarbon experiments, the experimental setup was basicly the same as the Fe -PAH experiments but with small modification, which was displ ayed in Figure 2 3. The pulsed Nd YAG laser beam was split into two beams. One beam (with intensity of approximately 60%) was directed by prism P1 and focused by lens L1 on surface of piece of Cu /Silver/Gold metal to ca. 0.5 mm spot diameter while second beam (with intensity Gold Iron TC CsI Window Copper Window Holder, 12K Ar Fe Target UV Vis Hg Lamp PAH Oven Nd YAG Laser Resistive Heater 12 K Helium Cycle C ryostat FT IR

PAGE 30

30 approximately 40%) was focu sed on the pressed pellet of mixture of 12C and 13C powders (ISOTEC, Inc.)) to spot of 2 3 mm diameter The distance two spots was approximately 4 5 mm and they were located close to the holder of the CsI window. The beamsplitter used was an ordinary glass plate. The intensity ratio of reflected beam from surface of this plate to the intensity of transmitted beam can be change by rotating the plate ( i.e change the incidence angle). For such simple beamspl itter, if the incidence angle is set to about 80 degrees the intensities of reflected and transmitted beams are equal ( i.e. 50% each) as is governed by Fresnels formula for TE light polarization.48 Both L1 and L2 lenses were mounted on the micrometer screw driven tables allowed periodically to repostion the focu ssed laser beams over fresh regions of carbon and copper surfaces. Figure 2 3. Experimental setup for copper -carbon experiments. Gold Iron TC 12K CsI Window Ar gas 12 C / 13 C Sample UV Vis Hg Lamp Nd:YAG Laser Helium Cycle Cryostat FT IR 63 Cu / 65 Cu Sample Beamsplitter P2 P1 L1 L2

PAGE 31

31 Laser Ablation Laser ablation was employed to generate metal (Fe, Cu, Ag and Au) atoms and carbo n clusters. A pulsed Nd:YAG laser (1064 and 532 nm, 0.2 0.5 W, 10 Hz) was used to vaporize metal target and pressed pellet samples which may contain 12C or 12C/13C mixture. For copper carbon, silver -carbon and gold -carbon experiments, synchronized dual -b eam laser ablation of metal and carbon sample was performed while single -beam laser ablation of iron sample was used for Fe -PAH experiments. Standard carbon and metal samples were used without further purification: 12C (powdered graphite, natural abundance 12C (98.9%) and 13C (1.1%)), 13C (99%, ISOTEC), Cu (natural abundance: 63Cu (69.2%) and 65Cu (30.8%), SPEX), Ag (natural abundance: 107Ag (51.8%) and 109Ag (48.2%), SPEX), Au (197Au (99.9%), Kurt J. Lesker). Generation of PAHs Since t he vapor pressures of PAHs decrease with increasin g molecular size, different methods were necessary to introduce PAHs into the cryostat chamber Benzene vapor was mixed with argon (1% benzene) and injected through a needle valve. Naphthalene was introduced directly into the chamber through a leak valve. Fluorene was introduced similar as naphthalene but with heating. Pyrene and coronene were sublimed from a resistively -heated oven located inside the cryostat chamber, close to the iron sputter target.

PAGE 32

32 CHAPTER 3 V IBRATIONAL S P ECTROSCOPY OF NEUTRAL COMPLEXES OF IRON AND POLYCYCLIC AROMATIC HYDROCARBONS Introduction Recently polycyclic aromatic hydrocarbons (PAHs) and their ions have garnered enormous interest because of their possible involvement in interstellar chemistry. First proposed by Leger & Puget (1984) and Allamandola et al (1985) to account for the unidentified interstellar infrared (UIR) emission bands, PAHs are now believed to be present in many regions of the interstellar medium (ISM). The UIR bands have been observe d from reflection and planetary nebulae, H II regions, post -AGB objects, plus some regions of the Mil ky Way and other galaxies .17,18 Though their presence is now widely accepted, specific PAHs have no t yet been identified as carriers of the UIR bands, principally due to band width and frequency overlap problems. Iron, another important component in the interstellar medium, is one of the most abundant elements .49 Only H, He, O, C, N, S, Mg, and Si are more abundant.50 It has been established observationally that the cosmic abundance of iron i n the ISM is depleted,5153 but the reason for its depletion is not known. While the depletion factor, taken as the ratio of the average elemental abundance in our solar system to that in the ISM, is about 1 3 for carbon, oxygen, nitrogen and sulfur; for magnesium, silicon and iron, however, it is substantially larger: 34, 42, and 100, respectively.50 French researchers have proposed a possible mechanism for the depletion of iron. Serra et al. proposed that since collisions between iron and PAHs probably occur f requently in the ISM, such collisions could lead to stable iron PAH complexes. The sequestration of iron in a complexed form could then explain the unusual depletion of elemental iron.54 In addition, Chaudret and co -workers have also proposed that ionic Fe(PAH)+ complexes could be responsible for the UIR bands.55 This naturally raises the companion question: could neutral

PAGE 33

33 iron PAH complexes could also act as contributors to the UIR bands? The goal of this study was to determine whether neutral iron -PAH complexes were stable, whether they could account for the depletion of iron in the ISM, and whether they might be responsible for the UIR bands. While some studies have been reported on Fe -P AH complexes, by far the bulk of previous studies have dealt with the Fe -benzene system. Timms was the first to study the ironbenzene system and reported a complex of unknown stoichiometry.56 Efner et al. studied iron and benzene in argon matrices and reported the formation of the Fe(C6H6) complex.57 Aleksayan and Kurtikyan, and Shobert et al. studied iron and benzene in matrices as well and assigned the newly observed infrared abs orption bands to the Fe(C6H6)2 complex .58,59 Morand and Francis also studied the iron -6C6H64-C6H6)Fe complex.60 Parker and Peden co-deposited iron and benzene in krypton matrices and used Mssbauer spectroscopy to identify a number of iron-benzene complexes, including Fe(C6H6) and Fe(C6H6)2.61 Ball et al. studied the iron -benzene system in argon matrices by infrared spectroscopy and deduced the formation of Fe(C6H6), Fe(C6H6)2, and possibly Fe2(C6H6), complexes.62 However, i n none of the above studies was any theoretical work reported. Boissel and Marty et al. studied the reaction cross -sections and dissociation rates for complexes of Fe+ and various PAHs via Fourier transform mass spectroscopy.63,64 Caraiman and Bohme used selected ion flow tube studies to determine the reactivity of iron complexed to coronene.65 Using time -of -flight mass spectrometry, Duncan and coworkers studied the photodissociation dynamics of metal complexes with different ligands, in cluding PAHs.6670 Elustondo et al. a nd Morand and Francis also studied the iron-PAH system in cryogenic Ar matrices via infrared and UV visible absorption spectroscopy.60,71 The gas -phase experimental

PAGE 34

34 vibrational spectra of cationic iron complexes with benzene, naphthalene, and fluorene have been very recently reported.72 In this chapter an experimental and theoretical i nvestigation of the vibrational spectroscopy of neutral Fe(PAH) complexes was presented Infrared absorption spectra of neutral complexes of iron with benzene, naphthalene, fluorene, pyrene, and coronene in solid Ar at 12K have been obtained. Supporting ca lculations of the equilibrium geometries, stabilities, and harmonic vibrational frequencies of these complexes have been carried out using density functional theory (MPW1PW91/6 31+G(d,p) method) using a modified Perdew -Wang exchange and correlation functio nal/basis set. Computational and Experimental Details Computational Details The geometries, energies, and harmonic vibrational frequencies of Fe(PAH) complexes were calculated using the Gaussian 03 suite of programs .73 Global searches for t he most stable structures of singlet, triplet and quintet ground state multiplicities for the complexes were carried out using the MPW1PW91/631+G(d,p) method with modified Perdew -Wang exchange and correlation functional/basis set.74,75 Previous work has shown that the MPW1PW91 functional is generally preferred over the frequently used B3LYP approach for equilibrium structur es, harmonic frequencies and other spectroscopic parameters of metal -containing molecular systems.76 Computed harmonic midIR vibrational frequencies were scaled uniformly by a factor of 0.972. This factor is the ratio of the experimental gas -phase absorption frequency (672 cm1) of the strongest CH out -of -plane mode in neutral benzene to its calculated (MPW1PW91/6 31+G(d,p)) frequency (691.3 cm1).72 As done previously,72 the scaling factors for the strongest computed CH stretching modes of Fe(benzene), Fe (benzene)2, Fe(naphthalene), Fe(fluorene), Fe(pyrene) and Fe(coronene) were taken as 0.944, 0.944, 0.952, 0.949, 0.944 and 0.955,

PAGE 35

35 respectively. These values differ substantially from previously used mid -IR scaling factors due to the large CH stretching mod e anharmonicities.72 Dissociation energies, D0, of the Fe(PAH) complexes were determined from D0 ZPE(PAH) EZPE(Fe(PAH)) (1) where EZPE is the electronic energy corrected for the zero -point vibrational energy. Here it was assumed t hat Fe retains the spin multiplicity of the complex, e.g., triplet for the most stable complex. Thus, the electronic energy of the 5D4 quintet ground state of iron, E(Fe), was increased by 1.485 eV, which is the difference between the Fe atomic 3F4 and 5D4 states.77 Neither the geometry nor the electronic energy of the PAH ligand in the complex was found to be modified greatly compared to uncomplexed PAH. For example, the ZPE -corrected electronic energy of Fe -complexed naphthalene, pyrene and coronene is lower than that of the ligands alone by only 0.07, 0.06 and 0.01 eV. These values increase D0 by similar values. The small values for the correction factors vali dates the use of the EZPE(PAH) energy of the separated ligands in (1). A similar formula for D0 was used previously in our study of cationic Fe(PAH)+ complexes.72 Experimental Details The experimental setup has been described in details in Chapter 2. Briefly, a pulsed Nd: YAG laser (1064 and 532 nm) was focused on a solid Fe target. The ablated iron vapor was mixed with PAH vapor and gaseous Ar and trapped on a CsI window cooled to 12 K. After 24 hours deposition, infrared absorption spectra were collected at 1 cm1 resolution using a MIDAC M2000 FT IR spectrometer. Most runs also involved subsequent UV -visible photolysis with a medium pressure Hg lamp, after which the IR spectra were also recorded. Since vapor pressures of the PAHs decrease with increasing molecular size, different methods were necessary to introduce the PAHs into the cryostat chamber. Benzene vapor was mixed with argon and injected directly via a needle valve. Both naphthalene and fluorene vapors

PAGE 36

36 were introduced in a similar fashion, but fluorene required some heating. Pyrene and coronene were sublimed from a resistively-heated oven located inside the cryostat chamber, close to the iron sputter target. A. Fe(C6H6) (3A2) B. Fe(C6H6)2(3A1g) C. Fe(C10H8) (3A) D. Fe(C13H10) (3A) E. Fe(C16H10) (3A) F. Fe(C16H10) (3A) G. Fe(C24H12) (3A) G. Fe(C24H12) (3A) (top view) Figure 3 1. Lowest energy stable structures for the complexes of iron with be nzene (C6H6)( A ), bis -benzene (C6H6)2 (B), naphthalene (C10H8) (C ), fluorene (C13H10) ( D ), pyrene (C16H10) ( E), pyrene (second isomer, F ), and coronene ( G ), all optimized at the MPW1PW91/6 31+G(d, p) level of theory. Fe(benzene) and Fe(b enzene)2 complexes T he computed equilibrium geometries for the most stable complexes are shown in Figure 3 1. Experimental infrared spectra of iron co-deposited with benzene are shown in Fig ure 3 2 Seven new bands (715.9, 761.3, 811.4, 952.9, 967.5, 982.8, and 1442.5 cm1) w ere observed after deposition. The bands at 761.3, 811.4, 952.9 and 982.8 cm1 decreased after photolysis and/or annealing. They are assigned here to the Fe(benzene) complex. The other three bands (715.9, 967.5, and 1442.5 cm1) increased with matrix photolysis and annealing, and are assigned to the Fe(benzene)2 complex. Fe(benzene)2 bands do not decrease with photolysis probably because (1)

PAGE 37

37 the radiant input energy is distributed over a larger number of modes in the bis -complex than in Fe(benzene), such th at the energy is not sufficient to break the complexes Fe ring bond; (2) slight annealing always accompanies photolysis, with the result that benzene aggregates are preferentially produced during matrix annealing.78 Thus the Fe(benzene) + benzene Fe(benzene)2 reaction is likely to be o perative. 700 800 900 1000 1400 0.0 0.1 AbsorbanceWavenumbers (cm-1) 811.4 761.3 952.9 982.8 715.9 967.5 1442.5 d b c a Fe(C6H6) Fe(C6H6)2 Figure 3 2. IR absorption spectrum of benzene (C6H6) only (a), Fe codeposited with benzene (b), after matrix UV -visible photolysis (c) and after matrix annealing at 35 K ( d), all in solid Ar at 12K. Figure 3 3 displays the synthetic experimental (experimental band positions and relative intensities with artificial 1 cm1 bandwidths) and calculated spectra for different spin multiplicities (singlet, triplet and quintet) of the Fe(benzene) complex. It can be seen that the experimental band positions and relative intensities for the triplet multiplicity matches the

PAGE 38

38 calculated vibrational frequency and band intensity patterns best. Furthermore, the triplet is calculated to be th e most stable of the complexes (see Table 3 1). Our calculations also show that in the most stable mono-complex structure the iron atom is centered 1.49 above the benzene ring ( A in Fig ure 3 1). The benzene ring is bent slightly out of -plane (C2v symmetr y), as found previously for the Fe(benzene)+ complex.72 400 600 800 1000 1200 1400 1600 0 50 100 150 0.0 0.5 1.0 0 50 100 150 0 50 100 Wavenumbers (cm-1)Relative Intensity Integral Intensity (km/mol) Fe(C6H6)Exp. Synthetic Calc., Singlet Calc., Triplet Calc., Quintet Figure 3 3. Synthetic experimental and calculated infrared absorption spectra at indicated spin state multiplicities for the Fe(C6H6) complex. Fig ure 3 4 shows the synthetic experimental an d calculated spectra for the Fe(benzene)2 complex. The experimental spectrum fits the calculated spectrum quite well except for the bands at 1008.7/1008.9 cm1, which are predicted but not observed. They were probably overlapped by the benzene band at 1010.9 cm1. Support of this comes from the experimental observation that

PAGE 39

39 the 1010.9 cm1 band intensity increased during matrix annealing in a similar fashion to the other Fe(benzene)2 complex bands (marked by triangles in Fig ure 3 2), while at the same time pure benzene bands (unmarked in Fig ure 3 2 ) decreased slightly. According to our calculations, the Fe(benzene)2 complex should also be a triplet with D6h symmetry and the two rings separated by 3.48 Similar 6 6 type coordination in the Fe(C6H6)2 complexes was earlier suggested by Ball et al.62 T able 3 1. Computed properties for Fe(PAH) compl exes Complex a State Sym Do b kcal/mol(eV) Rel EZPE c kcal/mol(eV) rFe ring d A. Fe(C6H6) 1A1 C2v 42.7 (1.85) 3A2 C2v 49.1 (2.13) 0.0 (0.00) 1.49 5A2 C2v 9.2 (0.40) B. Fe(C6H6)2 3A1g D6h 27.2 (1.18) 0.0 (0.00) 1.74 C. Fe(C10H8) 1A' Cs 36.5 (1.58) 3A" Cs 44.1 (1.91) 0.0 (0.00) 1.55 5A" Cs 1.0 (0.04) D. Fe(C13H10) 1A C1 43.4 (1.88) 3A C1 47.5 (2.06) 0.0 (0.00) 1.51 E. Fe(C16H10) 3A" Cs 34.82 (1.51) 8.5 (0.37) 1.35 5A' Cs 15.9 (0.69) F. Fe(C16H10) 3A Cs 43.4 (1.88) 0.0 (0.00) 1. 55 5A" Cs 15.8 (0.69) G. Fe(C24H12) 3A C1 12.0 (0.52) 0.0 (0.00) 1.53 a The most stable complexes as optimized at MPW1PW91/6 31+G(d,p) level with equilibrium geometries shown in Figure 3 1. b Dissociation energies calculated using formula (1), s ee chapter 2. c Relative total electronic energy corrected for zero -point vibrational energy. d Iron ring distance

PAGE 40

40 400 600 800 1000 1200 1400 1600 0 50 100 0.0 0.5 1.0 Integral Intensity (km/mol) Wavenumbers (cm-1)Relative Intensity Fe(C6H6)2Exp. Synthetic Calc., Triplet a Figure 3 4. Synthetic experimental and calculated infrared absorption spectra for the Fe(C6H6)2 complex. The inaccessible energy region overlapped with the absorption band of benzene is marked by the horizontal line (a). Observed and calculated Fe(C6H6), Fe(C6D6) and Fe(C6H6)2, Fe(C6D6)2 frequencies and relative intensities are listed in Table 3 2 and 3 3, respectively. These results are com pared to previous work in Table 3 4 and 3 5. As can be noted, our results do not agree with some previous assignments. Our band assignments are based on observation of band intensity changes after UV photolysis and matrix annealing, plus comparison with vi brational frequency calculations. The match between observed band energies and relative intensities and computed frequencies and intensities for the both monoand bis -Fe -benzene complexes is quite good.

PAGE 41

41 Table 3 2. IR absorption spectra of Fe(C6H6) and F e(C6D6) complexes. Fe(C6H6) Fe(C6D6) Mode a cal b /cm 1 exp / cm 1 cal b / cm 1 exp / cm 1 320.8 (0.02) 305.3 (0.03) 321.4 (0.02) 306.1 (0.03) 333.9 (0.03) 758.1 (1.00) 761.3 (1.00) 582.8 (1.00) 584.8 (1.00) 799.9 (0.05) 802.8 (0.05) 811.4 (0.13) 619.4 (0.09) 621.3 (0.09) 631.1 (0.17) 982.7 (0.16) 982.8 (0.21) 779.8 (0.15) 983.0 (0.16) 780.2 (0.15) 785.9 (0.38) 952.1 (0.01) 952.9 (0.03) 909.0 (0.04) 1255.4 (0.09) 1257.8 (0.12) 1255.6 (0.09) r 3048.6 (0.22) 2254.2 (0.15) r 3049.0 (0.22) 2254.3 (0.15) a -plane -of -ring vibration, and is the ring breathing mode. b Frequencies are scaled uniformly by scaling factors of 0.972 and 0.944 for midIR and C H stretching modes, respectively. The integral intensity for the 758.1 and 582.8 cm1 bands are 139.1 and 75.8 km/mol, respectively. Onl y modes with relative intensities equal to or larger than 0.01 are listed. Table 3 3. IR absorption spectra of Fe(C6H6)2 and Fe(C6D6)2 complexes Fe(C 6 H 6 ) 2 Fe(C 6 D 6 ) 2 Mode a cal b / cm 1 exp / cm 1 cal b / cm 1 exp / cm 1 250.4 (0.71) 246.0 (1.00) 708.1 (1.00) 715.9 (1.00) 524.0 (0.51) 533.1 (0.51) 826.9 (0.03) 641.5 (0.02) 828.9 (0.03) 642.9 (0.02) 972.0 (0.26) 967.5 (0.38) 927.1 (0.36) 922.3 (0.3 9) 1008.7 (0.12) c 791.9 (0.12) 796.1 (0.31) 1008.9 (0.12) 792.2 (0.12) 1442.8 (0.06) 1443.4 (0.06) 1442.5 (0.27) r 3057.6 (0.36) 2262.6 (0.26) r 3057.7 (0.36) 2262.7 (0.26) r 3064.8 (0.08) a Notation used: R and r are CC -plane -of -ring movement, and is rings breathing mode (anti phase).

PAGE 42

42 b Frequencies are scaled uniformly by scaling factors of 0.972 an d 0.944 for midIR and C H stretching modes, respectively. The integral intensity for the 708.1 and 246.0 cm1 bands are 127.1 and 91.4 km/mol, respectively. Only the modes with relative intensities equal to or larger than 0.01 are listed. c This band was not observed due to the absorption band of benzene in this region. Table 3 4. Comparison of present work with previous IR band assignments for Fe(C6H6) complex. This work Ball et al. 62 Efner et al. 57 Mode a cal b / cm 1 exp/ cm 1 exp/ cm 1 exp/ cm 1 320.8 (0.02) 321.4 (0.02) 366 678.1 (m) 746.3 758.1 (1.00) 761.3 (1.00) 760.8 (s) 762 799.9 (0.05) 802.8 (0.05) 811.4 (0.13) 812.4 (m) 812 952.1 (0.01) 952.9 (0.03) 952.7 (w) 953 982.7 (0.16) 983.0 (0.16) 982.8 (0.21) 983.6 (w) 983 993 1010 1179 1246 1392 1421.0 (w) 1430 r 3048.6 (0.22) r 3049.0 (0.22) a -plane -of -ring movement, and is ring breathing mode. bFrequ encies are scaled uniformly by scaling factors of 0.972 and 0.944 for midIR and C H stretching modes, respectively. The integral intensity for the 758.1 band is 139.1 km/mol. Only modes with relative intensities equal to or larger than 0.01 are listed.

PAGE 43

43 Table 3 5. Comparison of present work with previous band assignments for Fe(C6H6)2 complex. This work Ball et al.62 Aleksanyan et al.58 Skobert et al.59 Mode cal b / cm 1 exp/ cm 1 exp/ cm 1 exp/ cm 1 exp/cm 1 250.4 (0.71) 609 600 664.1 (m) 685 708.1 (1.00) 715.9 (1.00) 715.7 (s) 709 715 725 780 775 826.9 (0.03) 828.9 (0.03) 859 855 930 972.0 (0.26) 967.5 (0.38) 967.6 (w) 968 965 978 989 1008.7 (0.12) 1008.9 (0.12) 1011 1035 1148 1175 1175 1310 1412 1442.8 (0.06) 1443.4 (0.06) 1442.5 (0.27) 1441.0 (w) 1438 1435 1478 r 3057.6 (0.36) r 3057.7 (0.36) r 3064.8 (0.08) a -plane -of -ring movement, and is ring breathing mode (out -of phase). bFrequencies are scaled uniformly by scaling factors of 0.972 and 0.944 for midIR and CH stretching modes, respectively. The integral intensity for the 708.1 cm1 band is 127.1 91.4 km/mol. Only modes with rela tive intensities equal to or larger than 0.02 are listed.

PAGE 44

44 Fe(naphthalene) complexes Shown in Figure 3 5 the experimental IR spectra of iron co -deposited with naphthalene exhibit bands at 625.7, 732.3, 736.7, 821.5, 982.9, 1072.2, 1350.5, 1410.7 and 1462. 9 cm1, of which the 1350.5 cm1 band is the strongest. These bands all decreased upon photolysis and annealing, and are assigned here to the Fe(naphthalene) complex. 600 800 1000 1200 1400 0.04 0.06 0.08 0.10 0.12 0.14 Wavenumbers (cm-1) Absorbance 821.5 821.5 732.3 625.7 736.7 982.9 982.9 1072.2 1072.2 1350.5 1350.5 1410.7 1410.7 1462.9 1462.9 Fe(C10H8) Fe(C10H8)c a b Figure 3 5. IR absorption spectrum of: naphthalene (C10H8) only (a), Fe codeposited wi th naphthalene (b), after matrix UV visible photolysis (c), all in solid Ar at 12K. The star marked band at 873.1 cm1 is unassigned. Table 3 6 shows the comparison of the experimental data with calculated vibrational frequencies. Although the strongest absorption band predicted for the Fe(naphthalene) complex lies at 773.3 cm1, no band was observed here because of the strong absorption of naphthalene in

PAGE 45

45 this region. The rest of the calculated bands for the triplet Fe(naphthalene) complex fit the experimen tal ones quite well. To confirm these assignments, isotopic substitution using naphthalene -d8 was tried. The experimental IR spectra of iron codeposited with naphthalene d8 is compared to the calculated spectrum of triplet Fe(naphthalene-d8) in Fig ure 3 6 The strongest band of the Fe(naphthalene d8) complex (at 618.5 cm1) is now observable. Its band position shows good agreement with the predicted one at 609.8 cm1. 618.5 816.2 858.5 1075.4 1251.2 1299.5 1350.9 600 800 1000 1200 0.05 0.10 AbsorbanceWavenumbers (cm-1) Fe(C10D8) Fe(C10D8)d c b a Figure 3 6. Calculated IR spectrum of Fe(C10D8) (a), experimental IR absorption spectr um of naphthalene -d8 (C10D8) only (b), experimental IR absorption spectrum of Fe codeposited with naphthalene -d8 (c), experimental IR absorption spectrum of Fe codeposited with naphthalene -d8 after matrix UV -visible photolysis (d), all in solid Ar at 12 K. In uncomplexed naphthalene, all the CH out of -plane modes have similar energies and form a strong band at 773.3 cm1. When Fe is complexed to naphthalene, the degeneracy of those

PAGE 46

46 CH out -of -plane modes is lifted and two different energy modes are observed, with an energy splitting of the 46.7 cm1 (see Figure 3 7 and Table 3 6). The strongest one involves vibrations mainly on the ring closest to Fe and the second, lower energy mode involves vibrations mostly on the other ring. A similar, though higher energy, splitting of 70.3 cm1 was found for similar modes in the cationic Fe(naphthalene)+ complex.72 Other modes in the mid IR region of the Fe(naphthalene) spectrum have larger relative intensities compared to uncomplexed naphthalene. The intensity distribution in the CH stretching region (ca. 3050 cm1) in the Fe(naphthalene) spectrum is very similar to that in the naphthalene ligand, signifying little influence of the iron on these modes. 0 50 100 150 0.0 0.5 1.0 0 50 600 800 1000 1200 1400 0 50 100 Integral Intensity (km/mol) Relative Intensity Wavenumbers (cm-1) Fe(C10H8)Exp. Synthetic Calc., Singlet Calc., Triplet Calc., Quintet Figure 3 7 Synthetic experimental and calculated infrared absorption spectra with different spin multiplicities for the Fe(naphthalene) complex.

PAGE 47

47 Table 3 6. IR absorption spectra of Fe(C10H8) and Fe(C10D8) iron(naphthalene) complexes. Modea Fe(C10H8) Fe(C10D8) cal b / cm 1 exp/ cm 1 cal b / cm 1 exp/ cm 1 252.5 (0.03) 240.0 (0.04) 309.6 (0.03) 289.4 (0.04) 330.2 (0.02) 311.8 (0.02) 382.4 (0.03) 347.1 (0.04) 457.9 (0.09) 400.7 (0.21) 619.9 (0.04) 625.7 (0.04) 591.9 (0.06) 732.2 (0.14) 732.3 (0.44) 570.4 (0. 02) 601.2 (0.08) 738.6 (0.08) 736.7 (0.10) 777.3 (1.00) c 609.8 (1.00) 618.5 (1.00) 782.5 (0.05) 826.4 (0.02) 821.5 (0.02) 644.7 (0.02) 759.3 (0.03) 985.8 (0.14) 982.9 (0.23) 799.5 (0.03) 813.2 (0.10) 8 16.2 (0.07) 1017.0 (0.02) 1060.6 (0.15) 1072.2 (0.14) 846.8 (0.08) 858.5 (0.07) 867.5 (0.05) 1062.4 (0.08) 1075.4 (0.04) 1230.4 (0.05) 1349.1 (0.35) 1350.5 (0.35) 1254.7 (0.10) 1251.2 (0.11) 1409.1 (0.04) 1410. 7 (0.02) 1300.2 (0.13) 1299.5 (0.03) 1349.5 (0.52) 1350.9 (0.41) 1463.9 (0.03) 1462.9 (0.04) 1377.1 (0.02) 1423.0 (0.03) r 3047.4 (0.04) r 3062.4 (0.13) 2271.4 (0.25) r 3069.1 (0.21) 2267.4 (0.05) r 3077.2 (0.18) 2278.4 (0. 08) r 3079.7 (0.12) 2283.5 (0.10) a -plane -of -ring movement. b Frequencies are scaled unifor mly by scaling factors of 0.972 and 0.952 for midIR and CH stretching modes, respectively. The integral intensity for the 758.1 and 582.8 cm1 bands are 131.3 and 78.8 km/mol, respectively. Only modes with relative intensities equal to or larger than 0.02 are listed. c No bands due to Fe(C10H8) are observed in the 775.4 791.8 cm1 region due to overlap by the strong band of neutral naphthalene at 783.4 cm1. The relative band integral intensities in experimental spectrum were scaled with the 1350.5 cm1 band intensity set to 0.35, a value equal to the calculated 1349.1 cm1 band intensity.

PAGE 48

48 Our calculation shows that in the most stable structure for the Fe(naphthalene) complex the iron atom lies above one of the six-membered carbon rings, rather than cent ered above the whole molecule. This is also the case for the Fe(naphthalene)+ complex.72 The Fe(naphthalene) complex has Cs symmetry. Calculations also show that the triplet Fe(naphthalene) complex has the lowest energy. Since the experimental spe ctrum fits the calculated triplet spectrum the best (Fig ure 3 7 ), we conclude that the Fe(naphthalene) complex has a triplet ground state. Fe(fluorene) complexes The experimental infrared spectra of iron codeposited with fluorene is compared with the calcu lated spectrum of triplet Fe(fluorene) in Fig ure 3 8 The eight bands observed after deposition (at 718.0, 757.1, 776.5, 971.4, 1026.2, 1270.4, 1466.3 and 1509.1 cm1) all decreased after photolysis, in similar fashion, as noted above, for the mono complexes of Fe with benzene and naphthalene. The strongest band, observed at 776.5 cm1, is predicted to lie at 769.4 cm1. Table 3 7, which compares the experimental data with the calculated vibrational frequencies, shows that good agreement is reached with th e calculated triplet spectrum. Our calculations also show that 1) the iron atom in the Fe(fluorene) complex resides 1.51 above one of the six-membered carbon rings, instead of above the middle five -membered one (and thus possesses C1 symmetry), and 2) the Fe(fluorene) complex with triplet multiplicity is the most stable (Table 3 1). For these reasons, we conclude that the Fe(fluorene) complex is also a triplet, the same as the Fe(benzene) and Fe(naphthalene) complexes. Table 3 7. IR absorption spectra of Fe(C13H10) iron(fluorene) complex. Modea cal b/ cm1 exp/ cm1 Modea cal b/ cm1 exp/ cm1 244.2 (0.07) 1220.8 (0.02) 307.1 (0.06) 1273.5 (0.19) 1270.4 (0.19) 309.1 (0.05) 1296.2 (0.03) 407.1 (0.04) 1351.8 (0.10)

PAGE 49

49 Table 3 7 Continued. Modea cal b/ cm1 exp/ cm1 Modea cal b/ cm1 exp/ cm1 421.8 (0.02) 1375.4 (0.03) 614.8 (0.03) 1393.2 (0.02) 618.6 (0.08) 1423.2 (0.05) 710.6 (0.20) 718.0 (0.29) 1447.1 (0.03) 732.4 (0.03) 1469.3 (0.12) 1466.3 (0.13) 752.2 (0.83) 757.1 (0.98) 1499.5 (0.64) 1509.1 (0.54) 769.4 (1.00) 776.5 (1.00) 1630.8 (0.36) 800.0 (0.06) r 2898.6 (0.30) 822.4 (0.03) r 2937.3 (0.09) 829.1 (0.05) r 3040.1 (0.09) 942.3 (0.03) r 3046.5 (0.05) 970.7 (0.20) 971.4 (0.13) r 3054.0 (0.07) 992.0 (0.03) r 3057.9 (0.21) 1029.9 (0.11) 1026.2 (0.10) r 3061.8 (0.25) 1044.3 (0.14) r 3069.3 (0.51) 1096.0 (0.03) r 3069.8 (0.07) 1191.6 (0.03) r 3069.8 (0.07) a -plane -of -of -plane ring movement. bFrequencies are scaled unif ormly by scaling factors of 0.972 and 0.949 for midIR and C H stretching modes, respectively. The integral intensity for the 769.4 cm1 band is 80.9 km/mol. Only modes with relative intensities equal to or larger than 0.02 are listed.

PAGE 50

50 800 1000 1200 1400 0.05 0.10 AbsorbanceWavenumbers (cm-1) 718.0 757.1 776.5 971.4 1026.2 1270.4 1466.3 1509.1 Fe(C13H10) Fe(C13H10)d c b a Figure 3 8. Calcu lated IR spectrum of Fe(C13H10) (a), experimental IR absorption spectrum of fluorene (C13H10) only (b), experimental IR absorption spectrum of Fe codeposited with fluorene (c), experimental IR absorption spectrum of Fe codeposited with fluorene after matri x UV -visible photolysis (d), all in solid Ar at 12K. Fe(pyrene) complexes The Fe(pyrene) complex has two possible isomers, with the optimized equilibrium structures E and F displayed in Fig ure 3 1. Structure E has the iron atom located above the carbon ri ng that shares four carbon atoms with other rings. In F the iron atom sits above the carbon ring that shares three carbon atoms with other rings. Both structures have Cs symmetry.

PAGE 51

51 700 800 0.06 0.08 0.10 0.12 0.14 0.16 0.18 AbsorbanceWavenumbers (cm-1) 651.2 736.8 792.4 795.5 832.2 e d c b a Fe(C16H10) Fe(C16H10)[a] 1100 1200 1300 1400 1500 0.06 0.08 0.10 0.12 0.14 0.16 0.18 AbsorbanceWavenumbers (cm-1) 1026.4 1223.3 1384.9 1392.2 1421.6 1488.6 Fe(C16H10) Fe(C16H10)e d c b a [b] Figure 3 9. Calculated IR spectrum of Fe(C16H10) (structure E) (a), c alculated IR spectrum of Fe(C16H10) (structure F ) (b), experimental IR absorption spectrum of pyrene (C16H10) only (c), experimental IR absorption spectrum of Fe codeposited with pyrene (d), experimental IR absorption spectrum of Fe codeposited with pyrene after matrix UV visible photolysis (e), all in solid Ar at 12K plotted in two energy regions [a] and [b].

PAGE 52

52 Figure 3 9 shows two regions of the experimental IR spectra of iron codeposited with pyrene. The calculated spectra for the Fe(pyrene) E and F comple xes are also given. Eleven bands were observed upon deposition, at 651.2, 736.8, 792.4, 795.5, 832.2, 1026.4, 1223.3, 1384.9, 1392.2, 1421.6 and 1488.6 cm1, with the strongest at 832.2 cm1. All these bands decreased with photolysis, and are here assigned to Fe(pyrene). While it is clear from Fig ure 3 9 that the calculated spectrum for isomer E fits the observed spectrum poorly, the experimental spectrum shows good agreement with the calculated triplet spectra for isomer F Furthermore, F is more stable by 0.37eV than E (Table 3 1) and the observed frequencies compare favorably with the calculated vibrational frequencies for this isomer, as shown in Table 3 8. We therefore conclude that the most stable geometry for the Fe(pyrene) complex is structure F and that the complex has a triplet ground state. In their work on the Fe -pyrene system, Elustondo and coworkers observed two sets of bands, at 790 and 831 cm1, and at 1221 and 1393 cm1.71 Their 790 cm1 band probably corresponds to our 792.4 and 795.5 cm1 bands, while their 831, 1221, 1393 cm1 bands very likely are our 832.2, 1223.3, 1392.2 cm1 bands, respectively. Elustondo attributed their bands to two different carriers on the basis of different annealing behavior.71 We have however found that all our bands decreased at similar rates with matrix photolysis and/or annealing. There is thus no indication from our work that these bands belong to two different carriers. Furthermore, according to our computational results in Table 3 8, all these bands can be assigned to the Fe(pyrene) complex. The band at 1966 cm1 reported by Elustondo et al.71 was not observed in our work. Table 3 8. IR absorption spectra of Fe(C16H10) iron(pyrene) complex (isomer F from Fig ure 3 1). Mode a cal b / cm 1 exp / cm 1 Mode a cal b / cm 1 exp / cm 1 224.2 (0.08) 1122.1 (0.03) 293.6 (0.03) 1172.1 (0.03)

PAGE 53

53 Table 3 8 Continued. Mode a cal b / cm 1 exp/ cm 1 Mode a cal b / cm 1 exp/ cm 1 318.3 (0.13) 1184.2 (0.08) 3 40.5 (0.03) 1207.0 (0.21) 1223.3 (0.26) 412.1 (0.02) 1293.5 (0.17) 468.1 (0.03) 1349.3 (0.22) 496.6 (0.02) 1380.4 (0.06) 1384.9 (0.09) 538.3 (0.03) 1382.2 (0.29) 1392.2 (0.10) 645.4 (0.24) 651.2 (0. 06) 1419.3 (0.19) 1421.6 (0.21) 741.2 (0.15) 736.8 (0.11) 1486.0 (0.03) 756.1 (0.09) 1493.5 (0.13) 1488.6 (0.10) 779.6 (0.16) 792.4 (0.18) 1617.5 (0.04) 784.2 (0.27) 795.5 (0.18) 1619.7 (0.46) 803.4 (0.03) r 3021.6 (0.03) 806.5 (0.06) r 3021.7 (0.02) 820.3 (1.00) 832.2 (1.00) r 3033.6 (0.09) 972.8 (0.04) r 3041.0 (0.56) 1029.8 (0.13) 1026.4(0.12) r 3049.5 (0.40) 1083.1 (0.10) r 3050.0 (0.06) a Notation used: R and r a -plane -of -ring movement. b Frequencies are scaled uniformly by scaling factors of 0.972 and 0.944 for mid IR and CH stretching m odes, respectively. The integral intensity for the 820.3 band is 113.0 km/mol. Only modes with relative intensities equal to or larger than 0.02 are listed. Fe(coronene) complexes Fig ure 3 10 pr esents the experimental infrared spectra of iron codeposited with coronene along with the calculated spectrum for the triplet Fe(coronene) complex. Eight bands were observed after deposition, at 791.8, 815.7, 846.1, 1096.8, 1302.8, 1306.1, 1366.9 and 1438.9 cm1, with the strongest at 846.1 cm1. All bands decreased after UV photolysis, as found for the Fe(PAH) complexes described above, but at a greater rate. From Fig ure 3 10 it is clear that the photodissociation yield of Fe(coronene) complex is larger than for Fe(naphthalene), Fe(fluorene) or Fe(pyrene). The reaso n may be the significantly smaller binding energy for Fe(corenene) than for the other complexes: Do[Fe(coronene)] = 0.52 eV compared to the 1.91, 2.06, and 1.88 eV

PAGE 54

54 for Fe(naphthalene), Fe(fluorene), and Fe(pyrene), respectively (Table 3 1). The good agre ement between the experimental and calculated band energies and intensities, shown in Fig ure 3 10 and Table 3 9, leads us to assign the above bands to the Fe(coronene) complex with a triplet electronic ground state. 800 1100 1200 1300 1400 0.16 0.18 0.20 0.22 AbsorbanceWavenumbers (cm-1) 791.8 1438.9 1366.9 1306.1 1302.8 1096.8 846.1 815.7 Fe(C24H12) Fe(C24H12)a d c b Figure 3 10. Calculated IR spectrum o f Fe(C24H12) (a), experimental IR absorption spectrum of coronene (C24H12) only (b), experimental IR absorption spectrum of Fe codeposited with coronene (c), experimental IR absorption spectrum of Fe codeposited with coronene after matrix UV -visible photol ysis (d), all in solid Ar at 12K. Calculations also show that the iron sits over one of the outside carbon ring of the coronene, instead of the middle one, for C1 symmetry. The ironring distance was calculated as 1.53 very similar to the 1.55, 1.51, an d 1.55 values found for the Fe(naphthalene), Fe(fluorene) and Fe(pyrene) triplet complexes, respectively.

PAGE 55

55 Table 3 9. IR absorption spectra of Fe(C24H12) iron (coronene) complex. Mode a cal b / cm 1 exp / cm 1 Mode a cal b / cm 1 exp / cm 1 112.2 (0.02) 1295.1 (0.24) 1302.8 (0.11) 196.8 (0.04) 1342.7 (0.05) 308.4 (0.06) 1368.4 (0.05) 342.1 (0.12) 1372.2 (0.42) 1366.9 (0.15) 365.1 (0.04) R 1382.6 (0.12) 368.8 (0.04) 1390.6 (0.16) 507.2 (0.11) 1399.9 (0.11) 555.6 (0.04) 1417.7 (0.15) 618.5 (0.08) 1437.8 (0.03) 734.3 (0.04) 1450.9 (0.17) 1438.9 (0.15) 753.1 (0.03) 1467.6 (0. 03) 757.2 (0.05) 1535.7 (0.05) 758.1 (0.02) 1559.5 (0.04) 785.6 (0.38) 791.8 (0.15) 1585.8 (0.12) 807.3 (0.33) 815.7 (0.17) 1635.8 (0.04) 845.4 (1.00) 846.1(1.00) r 3056.4 (0.02) 980.7 (0.03) r 3057. 7 (0.05) 1013.0 (0.07) r 3058.5 (0.02) 1083.3 (0.42) 1096.8 (0.15) r 3074.4 (0.05) 1119.4 (0.02) r 3074.8 (0.11) 1139.1 (0.03) r 3075.3 (0.10) 1146.8 (0.05) r 3076.3 (0.76) 1290.1 (0.11) 1306.1 (0.07) r 3077.0 (0.55) -plane -of -ring vibration. bFrequencies are scaled uniformly by scaling factors of 0.972 and 0. 955 for midIR and CH stretching modes, respectively. The integral intensity for the 845.4 cm 1 band is 114.4 km/mol, respectively. Effect of complexation on IR spectra Figure 3 11 gives a comparison of the calculated IR absorption spectra for the PAHs st udied here, and their neutral and cationic complexes with iron. Comparing the Fe(PAH) to PAH spectra indicates that the CH out -of -plane bending modes (in the 800 cm1 energy region) are shifted to lower energy in the complex. The lone exception is the Fe(fluorene) complex which contains a five -membered ring (for detailed band shifts, see Figure 3 8, 3 9 and

PAGE 56

56 3 10). Presumably, the presence of Fe in the neutral Fe(PAH) complex induces a flatter CH out of -plane bending mode potential which results in shifts in the low energy vibrations. The opposite is observed in the Fe(PAH)+ complexes: the CH out -of -plane bending mode shifts to higher energy compared to the PAH+ bands (Fig ure 3 11).72 600 1000 1400 2800 3200 Wavenumber ( cm-1) 6.3 7.3 13.2 11.7 10.2 8.7 6.9 8.2 6.1 3.24 6.5 7.0 7.7 8.7 11.5 12.2 13.0 12.7 12.2 10.2 9.2 8.3 7.7 7.3 6.7 6.2 3.45 3.25 13.5 12.8 11.7 8.8 7.6 6.9 6.1 3.44 3.25Fe(PAH )+PAH+Fe(PAH ) PAH 600 1000 1400 2800 3200 Wavenumber ( cm-1) 6.3 7.3 13.2 11.7 10.2 8.7 6.9 8.2 6.1 3.24 6.5 7.0 7.7 8.7 11.5 12.2 13.0 12.7 12.2 10.2 9.2 8.3 7.7 7.3 6.7 6.2 3.45 3.25 13.5 12.8 11.7 8.8 7.6 6.9 6.1 3.44 3.25Fe(PAH )+PAH+Fe(PAH ) PAH Figure 3 11. Comparison of summed IR absorption spectra calculated at the MPW1PW91/6 31+G(d,p) level for a 1:1:1:1 mixture of naphthalene, fluorene, pyrene and coronene (PAH panel), for their complexes with iron (Fe(PAH ) panel), for similar mixture of cations (PAH+ panel) and for their cationic complexes with iron (Fe(PAH)+ panel). We note that the PAH, PAH+ and Fe(PAH)+ spectra are prepared at similar way to those of ref. 72,72 but for comparison purpose includ e also spectra of coronene, coronene+ and Fe(coronene)+. In the mid IR part of the Fe(PAH) spectrum, the highest energy CC stretching mode (6.2

PAGE 57

57 (Fig ure 3 11). Again, this is a reversed from what is found in the Fe(PAH)+ and PA H+ systems. the PAH with Fe. In summary, as expected, there is a weaker interaction between neutral Fe and PAH ligands compared to the cationic complexes. This is manifest itself in smaller shifts in band positions and relative intensities in the Fe(PAH) complex IR spectra. Summary In this chapter the experimental infrared spectra for Fe(benzene), Fe(benzene)2, Fe(naphthalene), Fe(fluorene), Fe(pyrene) and Fe(coron ene) complexes have been obtained. The spectral features showed good agreement with the harmonic vibrational frequencies calculated from density functional electronic structure calculations at the MPW1PW91/631+G(d, p) level. The energy calculations and th e comparison of experimental with calculated spectra reveal that the most stable structures for these complexes are triplets. The binding energies for these complexes decrease with increasing PAH size which is opposite to that observed previously for the F e(PAH)+ cationic complexes.72 Equilibrium geometry calculations show that the iron atom always sits over one of the six-membered carbon rings in the complex, closer to the naphthalene, fluorene, pyrene and coronene ligands by 0.12, 0.17, 0.17, and 0.19 respectively, when compared to similar iron ring distances in the respective cationic complexes.72 The estimated dissociation energies for the neutral Fe(PAH) complexes of Fe(naphthalene), Fe(fluorene), Fe(pyrene) and Fe(coronene) are low er than their cationic quartet spin state counterparts by 0.84, 0.82, 0.92 and 2.35 eV, respectively.72 Since the IR spectra of Fe(PAH) complexes display only small band shifts and intensity changes when compared to PAH free ligands, it is possibl e that PAHs do.19 However, because of the low dissociation energies of the Fe(PAH) complexes, it is

PAGE 58

58 considered unlikely that these complexes would survive in the harsh environment of ISM. Their contribution to UIR bands is thus expected to be minimal or nonexistent.

PAGE 59

59 CHAPTER 4 COPPER CARBON CLUSTERS: STR UCTURE, INFRARED FRE QUENCIES AND ISOTOPIC SCRAM BLING Introduction Transition metals may be important components in the chemistry of the interstellar medium. Iron, with its high nuclear stability, is the most abundant metal in interstellar space, but d espite its abundance, it is known to be depleted ab out 100-fold compared to our solar system. Complexation of iron with polycyclic aromatic hydrocarbons (PAHs) and/or carbon clusters has been proposed a s the reason for this depletion.51 53,72 Other transition metals are also present in the interstellar medium, although at lower concentrations .42,53 T heir participation in interstellar chemistry is largely unknown, primarily because of a dearth of information on their spectral properties. However, recent studies by Graham and Rittby and coworkers on the infrared s pectroscopy of matrix -isolat ed transition metal C3 clusters such as linear GeC3Ge fanlike TiC3, linear CrC3 and CoC3, and floppy NiC3Ni have begun to provide some of this much -needed information.7983 Using satellite based measurements, Jenkins, Savage, and Spitzer analyzed Cu+ ion column densities along various lines of si ght for early -type stars, and fou nd a substantial depletion of Cu+ compared to our solar s ystem.53 L aboratory studies of copper reactivity with interstellar molecules such as carbon clusters are therefore important initial steps in understanding its interstellar chemistry. In this paper we report our laboratory study on the formation of Cu -carbon clusters, and present evidence for the existence of CuC3. Previous laboratory work on copper -carbon clusters has been limited. Cationic CuCn + clusters (n = 1 3), generated in spark discharges of Cu and graphite, have b een observed in mass spectrometric studies.84 Copper acetylide (CuC acetylenic compounds with copper salts in liquid ammonia85, and is widely used as a catalyst in

PAGE 60

60 the production of copper powder.86 Other CuCn compounds, including copper acetylide, have been observed in oxyacetylene/ hydrogen/ copper flames.87 Time -of -flight mass spectrometric studies of Cu C met allo car bohedrenes (m et -cars) identified the clusters CunC2 + (n = 2k+1, k = 1 7) and CunC4 + (n = 2k+1, k = 2 4) and speculated on their possible structures.88 In this chapter we report on a matrix isolation in frared spectroscopic study of laser ablated Cu -carbon clusters and corresponding theoretical work on small stable copper -carbon systems, CumCn (m = 1, 2; n = 1, 2, 3). The equilibrium geometries, vibrational harmonic frequencies, and stabilities for these clusters have been calculated using density functional theory. When compared with our experimental data these results show that only the CuC3 cluster was observed in our experiments. Photo -induced 12/13C isotopic scrambling in Cu12/13C3 isotopomers has als o been observed and is proposed to occur via a bicyclic CuC3 intermediate. Computational and Experimental Details Experimental methods The experimental apparatus used for the generation and trapping of copper -carbon clusters in solid Ar is presented in ch apter 2 and shown in Figure 2 3 Briefly, t he output from a pulsed Nd YAG laser (1064/532 nm, 0.2 0.5W, 10Hz) was split into two beams, with one beam (intensity ca. 60%) directed by prism P1 and focused by lens L1 to a ~ 0.5 mm diameter spot on a piece o f Cu metal (SPEX, natural abundance: 63Cu (69.2%) and 65Cu (30.8%)). The second beam (intensity ca. 40%) was focused to a 2 3 mm diameter spot on a pressed pellet of 12C and 13C (ISOTEC). The Cu and C samples, ca 4 5 mm apart, were positioned close to the CsI sample window. By rotating the glass plate beam splitter, the ratio of the intensity of the reflected beam to the intensity of the transmitted beam could easily be changed. With a ~ 80 angle of incidence, the intensities of the reflected and tran smitted beams are equal, as expected from the Fresnel formula.48 Both L1 and L2 lenses were mounted on micrometer screw -driven

PAGE 61

61 mounts which allowed for periodic re positioning of the focused laser beams onto fresh regions of the carbon and copper samples. Infrared spectra were scanned using a NICOLET Magna 560 F T IR spectrometer (0.5 cm1 resolution) after a 2 3 hrs deposition of the Cu/Cn/Ar mixture on a CsI sample window held at 12K by a closed-cycle helium cryostat (APD Displex). Annealing to 35K followed by recooling to 12K, as well as photolysis with a me dium pressure 100W Hg lamp were used as needed to induce secondary reactions. Computational methods Calculation of molecular equilibrium geometries and associated vibrational harmonic mode frequencies and dissociation energies were calculated using densit y functional theory with the Gaussian 03 program package73 using Beckes three -parameter hybrid functional combined with the nonlocal correction functional of Lee, Yang, and Parr (B3LYP)89 and a 6 311++G (3df) basis set which contains three sets of d functions and one set of f functions. For systems containing transition metals, the B3LY P functional has been shown to occasionally predict incorrect equilibrium geometries, as determined by a comparison of experimental and calculated vibrational spectra.76,9092 So, in addition, we used the MPW1PW91 functional, a modified Perdew -Wang exchange and correlation functional, along with a 6 3 11++G(3df) basis set to verify the B3LYP optimized structure.74,75 This approach has been recommended by Wiberg,90 Du nbar,91 and Oomens et al.76 in their studies of C H -, and metal containing systems. The MPW1PW91 functional was used in our recent studies of iron complexed with cationic and neutral polycyclic aromatic hydrocarbons.72,92 We also tested th e BPW91 functional (on the near -linear CuC3 cluster) by using the 6 311++G(3df) basis set for the carbons and SDD pseudopotential for copper. Such a functional

PAGE 62

62 has b een recommended for copper, silver and gold 93 and has been used in calculations of various reaction products of these metals with hydrogen.94 From our 63Cu12/13C3 isotopomer frequency calculations (BPW91/6311++G(3df) [carbons]/ SDD pseudopotential [Cu]) we find the maximum difference (after scaling) between predicted and observed vibrational frequencies is large, 7.3 cm1, but this value drops to 2.4 cm1 when the MPW1PW91 functional is used with the same basis sets. Finally, when the MPW1PW91 functional and the 6 311++G(3df) basis set is used for all atoms, this maximum differe nce drops to 1.6 cm1, an acceptable value for 13C labeling assignments. The potential energy surface of the isomerization reaction for the near linear CuC3 cluster was plotted using the MPW1PW91/6 311++G (3df) level of theory exclusively. The transition state (TS) identified from this plot was verified using traditional transition state optimization with the Berny optimization algorithm, as incorporated in the Gaussian 03 package.73 Experimental I nfrared S pectra Fig ure 4 1 displays the in frared absorption spectra in the 12001260 cm1 and 17502200 cm1 range s for species formed by the laser ablation of graphite (spectrum a) and by the synchronized dual laser ablation of copper and graphite (spectrum b). The majority of bands in spectrum a have previously been assigned to various neutral carbon clusters (C3 C12 ) by us and others.95,96 Several bands in spectrum b, not due to neutral carbon clusters, such as t he anionic cluster s, C5 (1831.8 cm1) and C6 (1936.5 cm1)97,98 are also present; the C5 band is very weak and masked by the shoulder of the strong 1830.0 cm1 band and is therefore not marked in the figure. Also present are i mpurities such as CO and H2O which are known to be more abundant whe n high laser ablation powers are used.81 Additional reaction products are also seen, namely CuCO (2009.5 cm1)99 and C3H (1824.4 cm1).100,101

PAGE 63

63 2150 2100 2050 2000 1950 1900 1850 1800 1250 1200 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20 Wavenumbers (cm-1)Absorbance (b). Cu/12C/Ar (a). 12C/Ar1830.0 1817.9C121894.3C71936.5C6 -1952.5C6 1946.0C11CuCO1998.0C92039.0C32078.1C92127.9C72138.5C O2164.0 2009.5 C5C111856.6 1250.5 Figure 4 1. Infrared absorption spectra of products of laser ablation of graphite (12C (99%) + 13C (1%)) (spectrum a) and products of two beam laser ablat ion of graphite and copper (spectrum b, enlarged twofold). The spectra were recorded after matrix annealing to 35 K then cooling back to 12 K. The major bands due to pure carbon clusters and their reaction products with copper at 1830 and 1250.5 cm1 are indicated. C arbon monoxide and its product with coppe r are also marked. Two new bands in spectrum b at 1830.0 and 1250.5 cm1, not found in spectrum a, were found to be dependent on both copper and carbon concentrations. To determine whether these two bands belong to the same species, a large number of expe riments were performed under different experimental conditions, for example, different Cu/C concentration ratios (using different ablating laser beam intensities ), matrix annealing up to 35K and, finally, UV -visible photolysis (up to 1 hr). We found that t he ratio of the integral intensities of the 1830 and 1250.5

PAGE 64

64 cm1 bands is constant under these different conditions (and 5.7 + 0.3 ), s upport ing the conclusion that both bands arise from a common carrier. 1840 1820 1800 1780 1760 1260 1240 1220 1200 0.10 0.12 0.14 0.16 0.18 Wavenumbers (cm-1)Absorbance 1830.0 1802.4 1788.1 1783.8 1764.4 1759.5 1825.5 1250.5 1807.6 1246.3a f d b e g h a c c(a). Cu/12C/Ar (b). Cu/12C/13C/Ar [12C] : [13C] = 3 Figure 4 2 Inf rared spectra of reaction products from laser ablation of Cu and 12C (spectrum a) displayed in two energy regions. The bands at 1830.0 and 1250.5 cm1 are due to a common carrier containing Cu and 12C. Spectrum b was collected from run s similar to spectrum a, but with a 13C -enriched sample The band carriers marked by vertical dashed lines are due to: C3H (1824.4 cm1),100,101 C2H+ (1820.2 cm1),102 and C12 (1817.9 cm1).103 The fractionations of b d as well as of e g isotopomers via the proposed 12/13C isotopic scrambling in nl Cu12/13C3 (see text) are marked. The results of isotopic (13C) substitution are shown in Figure 4 2 (top panel, spectrum b). Inspection of the 13C labeled spectrum bui lt on the 1830 cm1 band reveals eight isotopomeric bands, ( a -h ). A very similar isotopomer band pattern was previously observed for near -linear 12/13C3H2O complexes 104 and linear 12/13C3Cr 83 and 12/13C3Co 81 clusters. Such a band pattern is

PAGE 65

65 characteristic of the CC asymmetric stretching mode in linear or near linear 12/13C3 with one end bonded to a ligand. In the next section, we explore whether linear or near linear CuC3 clus ters can account for these results. 1.2781.962B cCuC2, C2v 1.817 2.324C cCu2C, C2v1.8961.278 E l Cu2C2, D h1.227 1.824 F bc Cu2C2, C2v2.021 114.3 1.286 G nl CuC3, Cs 1.825177.21.3131.266161.3 H bc CuC3, C2v1.9791.3431.571 71.6 A CuC C vD l Cu2C, C v2.314 1.829 1.365K crown Cu2C3, C2v1.9951.9531.5001.839 I w Cu2C3, C2v1.293166.5132.51.839 J cCu2C3, C2v2.407 1.357 1.873 76.1 L cCu2C3, C2v1.2741.8021.8530.00 eV 0.00 eV 0.00 eV 1.35 eV 0.34 eV 0.00 eV 0.66 eV 0.00 eV 2.10 eV 0.50 eV 0.25 eV 0.00 eV Figure 4 3. O ptimized equilibrium structures for the CuC, CuC2, Cu2C, Cu2C2, CuC3, and Cu2C3 clusters. The bond lengths () and angles () calculated at B3LYP/6 311++G(3df) (italic type, top) and at MPW1PW91/6 311++G( 3df) (normal type) are marked. The relative isomer energies are indicated. Equilibrium geometries and vibrations for the Cum Cn (m =1,2, n=1, 2, 3) clusters Previous IR studies have identified metal -carbon clusters of the type MC3 and MC3M (M = transition metal atom).7983 Here we explore such clusters theoretically for M = Cu, and, in addition, explore the smaller MC, MC2, and MC2M species. Figure 4 3 shows the stable equilibrium geometries found for the Cum Cn clu sters (m = 1, 2; n = 1, 2, 3) with displayed geometry and energy parameters predicted by MPW1PW91 and B3LYP functionals. Predicted

PAGE 66

66 h armonic vibrational frequencies and intensities are collected in Table 4 1. T otal energies (corrected for zero point vibrati onal energ ies) and dissociation energies are given in Table 4 2. Table 4 1. Vibrational frequencies (cm1) and integral intensities (km/mol) for electronic ground states of CumCn (m = 1, 2; n = 1, 2, 3) clusters (displayed in Figure 4 3 ), calculated using B3LYP and MPW1PW91 functionals. Cu m C n Isomer B3LYP/6 311++G(3df) MPW1PW91/6 311++G(3df) A CuC ( X 4 ) 572.5 (10) 585.6 (14) B c CuC 2 (X 2A 1 ) a 1 1750.3 (7), a 1 436.4 (5), b 2 291.1(0) a 1 1768.7 (5), a 1 449.3 (5), b 2 321.9 (0) C c Cu 2 C ( X 3 A 2 ) a 1 528.3 (6), b 2 389.7 (5), a 1 210.2 (1) a 1 534.1 (6), b 2 392.4 (7), a 1 218.9 (1) D l Cu 2 C (X 3g) 665.0 (7), 240.3 (2x26), 232.7 (5) 574.5 (0), 509.1 (474), 210.1 (1) 67.3 (0) E l Cu 2 C 2 (X 1g) g 2093.1 (0), u 659.3 (61), g 255.7 (2x0 ), g 254.9 (0), u 98.2 (2x50) g 2112.0 (0), u 662.2 (65), g 260.5 (2x0), g 257.5 (0), u 98.4 (2x56) F bc Cu 2 C 2 (X 1A1) a 1 1703.4 (10), b 2 497.2 (17), a 1 313.2 (21), a2 140.1 (0), b1 133.4 (9), a1 57.8 (14) a 1 1713.8 (8), b 2 518.3 (23), a 1 324.9 (2 3), a2 201.9 (0), b1 166.1 (10), a1 51.1 (16) G nl CuC 3 (X 2A a 453.6 (21), a (9), a a (2), a 84.5 (17) H bc CuC 3 (X 2A1) a 1 1595.8 (25), b 2 1208.2 (19), a 1 634.5(22), a1 351.6 (13), b1 241.5(7), b2 189.7 (2) a 1 1618.3 (13), b 2 1221.9 (21), a 1 729.5 (5), a1 372.2 (12), b1 240.1(7), b2 220.2 (1) I w Cu 2 C 3 (X 1A1) b 2 1776.1 (404), a 1 1299.8 (9), a 1 567.2 (3), b2 536.0 (35), a1 314.8 (7), b1 272.9 (3), a2 241.6 (0), b 2 1 05.5 (135), a 1 50.2 (7) b 2 1776.4 (437), a 1 1313.1 (7), a 1 570.3 (2), b2 548.6 (21), a1 322.9 (11), b1 275.6 (4), a2 240.4 (0), b2 87.8 (137), a1 47.5 (8) J c Cu 2 C 3 (X 1A1) a 1 1488.7 (601), b 2 1287.2 (5), a 1 447.1 (26), b2 413.2 (18), a1 381.0 (112), b1 356.0 (5), a2 190.0 (0), a 1 186.7(6), b 2 168.6 (18) a 1 1521.2 (552), b 2 1267.8 (11), a 1 469.9 (126), a1 457.6 (136), b2 411.5 (8), b1 348.3 (6), a2 198.5 (0), a1 171.8 (15), b 2 144.8 (23) K crown Cu2C3 (X 1A1) a 1 1471.1 (26), a 1 842.9 (134), b 2 725.9(3 7), b2 390.1 (55), a1 300.6 (9), a2 258.8 (0), b2 221.6 (0), b2 179.8 (18), a1 90.3 (2) a 1 1498.4 (11), a 1 906.8 (127), b 2 804.3(3), b2 425.1 (42), a1 319.4 (14), b2 258.3 (0), a2 256.8 (0), b1 176.4 (17), a1 97.2 (2) L c Cu 2 C 3 (X 1A1) a 1 1748.1 (87), b 2 638.3 (2), a 1 592.8 (9), a1 545.4 (20), b2 539.9 (2), a2 239.6 (0), b1 210.3 (23), a1 129.9 (1), b2 109.7 (44) a 1 1766.6 (90), b 2 658.9(2), a 1 605.6 (10), a1 559.1 (21), b2 557.6 (3), a2 228.0 (0), b1 206.4 (26), a1 120.6 (1), b2 70.2 (46)

PAGE 67

67 Table 4 2. Calculated total energies EZPE (Hartrees) corrected for zero point vibrational energies, and estimated dissociation energies, D0 (eV) or isomerization energy barriers, EIso (eV) for CumCn (n = 1, 2; m = 1, 2, 3) isomers. Cu m C n Isomer B3LYP/6 311++G(3df) MPW1PW91/6 311++G(3df) A CuC ( X 4 ) E ZPE = D0 (Cu (X 2S1/2) + C (X 3P)) = 1.99 eV E ZPE = D0 (Cu (X 2S1/2) + C (X 3P)) = 2.20 eV B c CuC 2 (X 2A1) E ZPE = D0 (Cu (X 2S1/2) + C2 (X 1g +)) = 2x2.33 eV E ZPE = 1716.594 551 D0 (Cu (X 2S1/2) + C2 (X 1g +)) = 2x2.47 eV C c Cu 2 C (X 3A2) E ZPE = D0 (Cu (X 2S1/2) + CuC ( X 4 )) = 2.29 eV E ZPE = D0 (Cu (X 2S1/2) + CuC ( X 4 )) = 2.21 eV D l Cu 2 C (X 3g) E ZPE = D0 (Cu (X 2S1/2) + CuC ( X 4 )) = 0.28 eV E ZPE = D0 (Cu (X 2S1/2) + CuC ( X 4 )) = 0.86 eV E l Cu 2 C 2 (X 1g) E ZPE = D0 (2Cu (X 2S1/2) + C2 (X 1g +)) = 2x4.09 eV E ZPE = D0 (2Cu (X 2S1/2) + C2 (X 1g +)) = 2x4.23 eV F bc Cu 2 C 2 (X 1A1) E ZPE = D0 (2Cu (X 2S1/2) + C2 (X 1g +)) = 4x1.85 eV E ZPE = D0 (2Cu (X 2S1/2) + C2 (X 1g +)) = 4x1.95 eV G nl CuC 3 (X 2A E ZPE = D0 (Cu (X 2S1/2) + C3 (X 1g +)) = 2.07 eV E ZPE = 655 782 D0 (Cu (X 2S1/2) + C3 (X 1g +)) = 2.26 eV H bc CuC 3 (X 2A1) E ZPE = EIso (bc -CuC3 ( X 2 A1) nl -CuC3 ( X 2 A a E ZPE = EIso (bc -CuC3(X 2 A1) nl -CuC3(X 2 A b I w Cu 2 C 3 (X 1A1) E ZPE = 164 221 D0 (Cu(X 2S1/2) + nl -CuC3 (X 2A = 1.86eV E ZPE = D0 (Cu(X 2S1/2) + nl -CuC3( X 2A = 1.85 eV J c Cu 2 C 3 (X 1A1) E ZPE = D0 (Cu(X 2S1/2) + bc -CuC3 (X 2 A1)) =2.00 eV E ZPE = D0 (Cu(X 2S1/2) + bc -CuC3 (X 2 A1)) = 1.94 eV K crown Cu2C3 (X 1A1) E ZPE = D0 (Cu(X 2S1/2) + bc -CuC3 (X 2 A1)) = 1.50 eV E ZPE = D0 (Cu(X 2S1/2) + bc -CuC3(X 2 A1)) = 1.70 eV L c Cu 2 C 3 (X 1A1) E ZPE = D0 (c -Cu2C ( X 3A2) + C2 ( X 1g +)) = 2x 3.06 eV E ZPE = D0 (c -Cu2C ( X 3A2) + C2 ( X 1g +)) = 2x 3.14 eV a The relative energy of TS(2A 311++G(3df) level. b See the PES of Figure 4 4

PAGE 68

68 The CuC Copper -Carbon Cluster MPW1PW91 /6 311++G(3df) calculations predict that the lowest energy multiplet of diatomic CuC is a quartet with bond length = 1.817 The doublet spin state is higher by 0.266 eV. B3LYP/6 311++G(3df) calculations find that the doublet state is only marginally higher (0.049 eV) than the quart et. Both theoretical levels predict very low vibrational integral intensities, 14 and 10 km/mol, respectively which accounts for CuC not being observed in our experiments. The CuC2 Copper -Carbon Cluster Our calculations predict that singlet dicarbon (C2) reacts spontaneously with Cu (X 2S1/2) to form a doublet cyclic cluster, c CuC2 (structure B). The predicted Cu C bond strength is 2.33 eV (2.47 eV) [B3LYP (MPW1PW91)], only slightly larger than in quartet CuC, (1.99 eV (2.20 eV )). Table 4 1 shows that the strongest infrared mode in c CuC2 lies at 1750.3 cm1 (1768.7 cm1) (CC stretch), but since its calculated intensity is only 7 km/mol (5 km/mol), the absence of this band in our spectra is understandable. The Cu2C Copper -Carbo n Cluster The lowest energy isomers of this cluster are triplets of C and D The D (linear) isomer is higher in energy by 1.35 eV (MPW1PW91/6311++G(3df)). T he highest IR integral intensity for the more stable cyclic isomer C is only 7 km/mol and in an ene rgy range inaccessible by our FT IR. The Cu2C2 Copper -Carbon Cluster Copper acetylide, l -CuC2Cu (E), is well known in catalytic chemistry.86 P redicted to appear at 662 cm1, this species is also absent in our spectra, presumably also because of its low concentration and intensity (65 km/mol). The second isomer bc Cu2C2 (F) is higher in e nergy by 0.66 eV (Figure 4 3) and also has a very low IR intensity of 23 km/mol (Table 4 1).

PAGE 69

69 The CuC3 Copper -Carbon Cluster Two sta ble structures were found for the CuC3 cluster: near -linear, nl CuC3 (G ), and bicyclic, bc CuC3 (H ). The former is more stab l e by 0.64 eV (0.34 eV) in B3LYP (MPW1PW91) calculations (Table 4 2). The highest predicted band intensity for H is 25 km/mol for the 1595.8 cm1 band. This species is not observed here. The predicted CuC bond length in nl -CuC3 ( G ) is 1.838 (1.825 ), cl ose to the experimental value of 1.8296 in CuCN.105 Unscaled p redicted vibrational frequenc ies (and integral intensities) for nl -CuC3 are 1900.5 (172 km/mol ) (asymmetric CC stretch) and 1281.5 cm1(31 km/mol) (symmetric CC stretch). The calculated ( MPW1PW91 ) isotopomer frequencies for the CC asymmetric and symmetric modes in nl Cu12C3 were scale d by factors (listed in Table 4 3 ) so they match the observed 1830 and 1250 cm1 bands. Although the MPW1PW91 predictions match experimental intensities [1830.0 (1.0) and 1250.5 (0.17) cm1], the B3LYP calculation fails to predict the correct intensity ratio for these bands (Table 4 1). The intensities predicted by B3LYP are 134 and 9 km/mol, (1.0 and 0.067 relative intensities), respectively. The geometry predicted for G is substantially different using B3LYP vs MPW1PW91, especially for the CuCC angle, 153.0 vs 161.3 degrees, respectively. In addition, some predicted mode symmetries are different for the two calculational levels, even though the structures and electronic ground states are the same (Table 4 1). The failure of the B3LYP functional for the G isomer is likely related to the large difference (0.3 eV) in predicted energy stability of the G and H isomers. To confirm the assignment of the observed bands to the near linear species, isotopic substitution experiments were run. Figure 4 2 shows bands due to the 12, 13C -isotopomers built on the 1830.0 and 1250.5 cm1 bands. Table 4 3 gives the comparison between the observed isotopomer bands and the predicted (and scaled) nl -63Cu12/13C3 frequencies. The frequencies

PAGE 70

70 predicted by the MPW1PW91/6 311++G(3df ) calculations support the assignment of these bands to nl CuC3 (X 2A G ). However, the B3LYP/6 311++G(3df) calculation gives considerably poorer results. The maximum differences between experimental and predicted isotopomer frequencies, expsc, are 4.3 cm1 (B3LYP/6 311++G(3df)) and 1.6 cm1 (MPW1PW91/6 311++G(3df) ). The average values of these differences for all observed isotopomers (for both modes) are 2.1 and 0. 96 cm1, respectively. Although the 1.6 and 0. 96 cm1values (MPW1PW91) are typical for structures assigned using isotopic13C labeling, the 4.3 and 2.1 cm1 (B3LYP) values are unacceptably large. Figure 4 2 shows that in the 1200 cm1 range only two isotopomer band s (labeled a and c ) appear while eight appear for the higher frequency mode. To understand this requires looking at the integral intensities of the c and d iso topomers for both modes (Table 4 3 ). MPW1PW91 calculations predict that the integral intensities for the asymmetric and symmetric C=C stretching modes in the c isotopomer are 158 and 32 km/mol, while in d they are 176 and 26 km/mol, respect ively. Thus, for the c isotopomer, the intensity ratio of the upper to the lower frequency band is expected to be 1788.1cm1 to 1246.3 cm1 bands. But, for the d isotopomer the predicted ratio is (=176/26) so the lower frequency band should be substantially weaker I ts absence is thus understandable. Information about the mechanism of formation of CuC3 can be obtained from a comparison of the isotopomeric band pattern observed for Cu C3 vs that for C3. The C3 isotopomers (and relative band intensities) observed (but not displayed) in the present work were: 12 1212 (1.0), 12 1213 (0.46), 13 1213 (0.09), 1213 12 (0.28), 13-1312 (0.20) and 13 1313 (0.11). The CuC3 isotopomers observed (and shown in Figure 4 2 and listed in Table 4 3)

PAGE 71

71 were: a (63 1212 12) (1.0), b (63 12 121 3 ) ( d (63 1312 12) ( 0.29), f (63 1312 13) (0.08), c (63 1213 12) (0.29), e (63 12 1313) (0.06) + g (63 13 1312) (0.12), and h (63 1313 13) (0.11). After summing the indicated band absorbances (because the same assumed 12/13C3 reactant leads to the ( b d ) and ( e g ) product pair s ), the observed isotopomer band pattern of Cu12/13C3 is very sim ilar to the intensity pattern in 12/13C3. We can thus conclude that Cu atoms react with already -formed C3 molecules to form CuC3. Table 4 3. E xperimenta l (Ar matrix 12K) and calculated isotopomer frequencies (integral intensities) for the asymmetric and symmetric CC stretch fundamental modes of fully optimiz ed equilibrium geometry of near linear 63Cu12/13C ( G Figure 4 3 ). Proposed band a ssignments marke d in Figure 4 2 are given in the first column. Isotopomer exp /cm1 sc a /cm 1 (km/mol) exp sc /cm 1 sc b /cm 1 (km/mol) exp sc /cm 1 B3LYP/6 311++G(3df) MPW1PW91/6 311++G(3df) Asymmetric CC stretch mode a 63 12 12 12 1830.0 1830.0 (134) 0.0 1830.0 (171) 0.0 b 63 12 12 13 1825.5 1824.2 (131) 1.3 1825.5 (168) 0.0 c 63 12 13 12 1788.1 1785.2 (125) 2.9 1786.6 (158) 1.5 d 63 13 12 12 1807.6 1810.4 (134) 2.8 1807.7 (176) 0.1 e 63 12 13 13 1783.8 1779.5 (123) 4.3 1782.2 (154) 1.6 f 63 13 12 13 1802.4 1803.8 (132) 1.4 1802.6 (172) 0.2 g 63 13 13 12 1764.4 1764.5 (125) 0.1 1763.1 (162) 1.3 h 63 13 13 13 1759.5 1758.0 (123) 1.5 1758.0 (158) 1.5 Symmetric CC stretch mode a 63 12 12 12 1250.5 1250.5 (9) 0.0 1250.5 (31) 0.0 b 63 12 12 1 3 1224.6 (9) 1224.2 (32) c 63 12 13 12 1246.3 1248.7 (9) 2.4 1247.8 (32) 1.5 d 63 13 12 12 1228.4 (7) 1229.7 (26) e 63 12 13 13 1222.3 (10) 1220.8 (33) f 63 13 12 13 1202.9 (8) 1203.9 (27) g 63 13 13 12 1249.0 (7) 1227.9 (27) h 63 13 13 13 1222.6 (8) 1201.4 (28) a Frequencies are scaled uniformly by a scaling factor of 0.9627 for the asymmetric mode and 0.9826 for the symmetric mode. b Frequencies are scaled uniformly by a scaling factor of 0.9629 for the asymmetric mode and 0.9 758 for the symmetric mode.

PAGE 72

72 This conclusion is supported by annealing and photolysis results. Matrix annealing (to 35 K) followed by cooling back to 12 K increases the intensities of the 1830.0 and 1250.5 cm1 bands by 27 %. This suggests that Cu (X 2S1/2) reacts with C3 ( X 1g +) as a result of diffusion at high er matrix temperature s Photolysis for 1 hr with photon energies of h reduced both the 1830.0 and 1250.5 cm1 bands by 13%. The predicted lowest energy photodissociation path is nl -CuC3 + h l -C3, with a dissociation energy of 2.07 eV (2.26 eV) (cf ., Table 4 2 ). This is close to the calculated Cu C bond ene rgy of 1.99 eV (2.20 eV) [B3LYP (MPW1PW91)] for diatomic CuC. The Cu2C3 Copper -Carbon Cluster Four stable C2v isomers were found on the singlet Cu2C3 potential surface. The l owest energy one is w -shaped ( I) and the other less stable ones ( J K and L) ar e either cyclic or crown -shaped Again, the large difference in stability between I and K for B3LYP and MPW1PW91 functionals is reflected in large frequency differences predicted for the three highest energy modes, and in the mismatch of the fundamental mo de symmetries of K as listed in Table 4 1. The most intense mode predicted for I is 1776.4 cm1 (437 km/mol), the asymmetric CC stretch. No bands assignable to I or J, K, L cluster s have been observed in this energy region. 12/13C -I sotope S crambling in the Near Linear CuC3 C lust ers Since the b (63 1212 13) and d (63 13 1212) iso topo mers originate from the same singly -substituted isotopic C3 precursor ( 13 1212), Cu might be expected to bond to either chain end with equal probability. If so, i t follows th at the b and d isomers should have about the same infrared intensities. Theoretical predictions roughly affirm this expectation. The d isomer is predicted to be about 5% more intense than b But, as can be seen in Figure 4 2 this is not what

PAGE 73

73 is observed. The intensity of the d band is approximately twice the intensity of the b band. A similar argument can be made for the doubly -substituted isomers, e (63 1213 13) and g (63 13 1312) (see Figure 4 2 ). g is observed to be approximately twice as intense as e This anomaly is the result of photoscrambling. As background, we first consider a similar observation in 12/13C3 that was also ascribed to isotopic photoscrambling.106 In 12/13C3, scrambling require s photon energies in the 2.75 3.54 eV range (covering the 3u (3 1u 1g + tr ansition).106 The maximum barr ier to the forward and backward reactions along the 3u (3 3A 3A 2 3A 3u (3 reaction pathway has been computed (using full configuration interaction) by Fueno and Taniguchi 107 as 2.82 eV. Visible/UV p hotolysis of C3 induce s intramolecular rearrangement to a cyclic C3 intermediate This is f ollowed by specific bond breaking which leads to a gain in the 1213 12 and 1213 13 isomer concentrations. This increase results because these isome rs have lower zero point energies (ZPE) than their precursors (12 1213 or 1312 13) The observation of isotopic scrambling in C3 supports the existence of the 3A 3 (3A 2) structure predicted by Fueno and Taniguchi.107 We now consider the photoinduced isotopic scrambling in CuC3 clusters. Because photoscrambling in C3 involves a cyclic intermediate a reaction pathway between G and H was sought theoretica lly The calculated PES (MPW1PW91/6 311++G(3df)) for the Cu 1212 13 and Cu 13 1212 isomers is displayed in Figure 4 4 The angle (C1C2C3) was incremented from 71.6 to 180 and, at each step, all other structural parameters fully optimized. Starting f rom the left G isomer, the PES rises 1.25 eV to a transition state, TS, and then falls to the H isomer, which lies 0.34 eV above G The structural parameters for G and H are given in Figure 3 and, for the TS are: R(C1C2) = 1.2975 R(C2C3) = 1.3238 R( C3Cu) = 1.8537

PAGE 74

74 C1C2C3 and ( CuC3C2) = 144.3. At = 77, the 2A in Figure 4 4 ) intersects with the H isomer distortion PES (large filled circles). The latter has a stable minimum at = 71.6. For the two curves calculated between > 77 and 93, the lower one (2A) almost preserves the C2v symmetry of the H isomer, although it was calculated using Cs symmetry. The bond lengths found show that R (C1C2) (C2C3) and R(CuC1) 3). However, for the hig her energy curve (2A (C1C2) 2C3) = 0.01 0.03 and the CuC bond lengths differences become large. At = 180, CuC3 is linear but exhibits one imaginary frequency ( 61cm1). The calculated barrier to linearity for the G isomer is so much smaller (39 cm1) than its computed vibrational frequencies (cf Table 4 1), that the nl -CuC3 ( G ) isomer can be characterized as quasilinear. The isotopic scrambling can most easily be tracked by following a speci fic nl -CuC3 isotopomer on the PES in Figure 4 4 First, consider the 63 1212 13 isotopomer G as the reactant. To drive the photoscrambling reaction, the G isomer must first absorb a photon and become electronically excited. From the doublet excited state, reversion to the ground state occurs partially by energy dissipation into the matrix bath modes and partially by internal conversion to the ground electronic state. After internal conversion, the G isomer presumably retains sufficient thermal energy to su rmount the ground state TS barrier, thus converting to the H isomer. T he 13C atom will now be located on the right hand side of H (see Figure 4 4 ). Further irradiation may either convert it back to the left G isomer (by breaking the just -formed Cu -13C bond ) or lead it to the right G iso topo mer (by breaking the Cu -12C bond). Either reaction is equally probable. If the latter reaction occurs, the 63 13 1212 isotopomer will be formed, and the 13C isotope position will have been scrambled. If the former react ion occurs, no scrambling will have occurred, and reversion to the initial reactant ensues. Backward reactions from the right

PAGE 75

75 isotopomer (through H to the left G ) could then occur. The relative rates of the forward and backward rates are dependent on the e nergy differences of the two stable G isotopomers and may be calculated. 12 13 TS (1.25 eV ) TS (1.25 eV ) 12 12 13 12 13 13 12 12 12 12 H (0.34 eV ) G (0.0 eV ) G ( 0.0008 eV ) 13 12 12 0.0 1.5 0.5 2.0 180.0 71.6 180.0 Rel Energy ( eV )degrees degrees2A2A1 Forward, kfForward, kf Backward, kb kf/kb 2.2 at 12K 2A Figure 4 4. To tal energy plot for the near linear CuC3 (2A 3 (2A 2A /2A1) clusters calculated at the MPW1PW91/6 311++G(3df) level by incrementing the (CCC) angle in the range of 77 the four remaining geometrical parameters were fully optimized. Note that the 63 121213 cluster reactant (left) rearranges to the 63 13 1212 lower energy isotopomer produc t by breaking the 12C Cu bond in structure H then passing over the TS and relaxing to the G product (right) in the forward reaction pathway. The ratio of the forward -to -backward reaction rate constants, under thermal equilibrium conditions, can be described by kf/kb= exp( G/kT) (1)

PAGE 76

76 where G is the difference in the Gibbs free energy of products and reactants, k is the Boltzmann constant and T is the sample temperature.108 At low temperatures (as in a matrix), it may be assumed that the changes in entropy are sm all, thus G = EZPE. Equation (1) can then be rewritten kf/kb= exp( EZPE /kT) (2) T he energy of the 63 1312 12 isotopomer is lower than the 6312 1213 isotopomer by 6.6 cm1(MPW1PW91/6311++G (3df)). Thus EZPE = 1and the ratio of the f orward and reverse rate constants, estima ted from (2), is concentration of the 63131212 isotopomer d will increase at the expense of the 6312 1213 isotopomer b as observed in Figure 4 2 A similar procedure can be applied t o the 6313 1312 product and 63 1213 13 reactant isotopomers. Again, EZPE was found to be = 1 and kf / kb 131312 isotopomer g is expected to increase, while 63 1213 13 isotopmer e should decrease, again as obser ved The photo induced 12/13C isotopic scrambling effect in CuC3 could be applicable to those clusters potentially formed in interstellar space, both in the gas phase and/or trapped on cold grain surfaces. The scrambling reactions could be driven by the ab sorption of stellar UV -visible radiation or cosmic radiation which penetrates even to dark and cold nebular interiors. Such scrambling will yield isotopic fractionation in 63Cu12/13C3 clusters via the reactions described here. Longer CuCn (n = 4 9 ) clust ers As indicated in Figure 4 1, there are many larger carbon clusters are formed in the present experiment s (i.e., larger than C3). There are also many bands left unassigned (in spectrum b of

PAGE 77

77 Fig ure 4 1 ), particularly in the 1800 cm1 region Some of these may be due to carbon and copper bearing clusters. To investigate this possibility, we calculated equilibrium geometries and vibrational harmonic frequencies for clusters with stoichiometry CuCn (n = 4 9) using MPW1PW91/6 311++G(3df) level of theory. The most stable geometries found are given in Fig ure 4 5 and the predicted vibrational frequencies in Table 4 4 Table 4 4. Calculated (at MPW1PW91/6311++G(3df)) harmonic vibrational frequencies and their integral intensities (in parentheses) for the CuCn (n = 4 carbon clusters (displayed in Fig. 4 5 ). Cluster Calc. (unsc.) freqs /cm 1 (km/mol) Calc.(sc.) freqs a /cm 1 (km/mol) CuC 4 C s 2139.9(48), 1870.7(160), 1021.5(21), 500.7(0), 497.7(1), 362.9(10), 198.3(54), 148.7(14), 88.1(1) 1801.3 (1 60) CuC 5 C s 2041.6(0), 1945.1(1023), 1515.3(62), 870.2(23), 604.0(6), 518.8(1), 381.5(1), 344.2(7), 324.1(4), 177.6(24), 162.8(16), 62.8(2) 1872.9 (1023) CuC 6 C 2169.6(207), 2100.1(246), 1917.8(329), 1275.1(0), 750.3(18), 648.8(1), 616.2(1), 499.2(0) 452.9(9), 305.3(8), 274.4(1), 265.6(0), 156.5(21), 144.6(33), 57.5(1), 56.2(0) 2089.1 (207) 2022.2 (246) 1846.6 (329) CuC 7, C 2139.0(111), 2056.6(1), 1883.2 (2106), 1617.7(264), 1138.8(3), 697.4(3), 669.2(19), 584.0(0), 580.3(0), 491.6(0), 377.8(3), 351.2(10), 285.6(7), 226.7(0), 216.1(0), 123.8(21), 119.9(19), 44.8(0), 44.2(0) 1813.3 (2106) CuC 8, C 2205.7(1015), 2125.6(220), 2070.6(8), 1923.0(501), 1409.3(66), 1014.5(15), 692.4(0), 656.2(1), 599.3(17), 587.9(1), 573.2(0), 467.2(0), 432.5(0), 293.5( 17), 288.5(7), 268.0(6), 198.0(1), 185.2(0), 104.4(17), 102.1(21), 37.7(0), 37.0(0) 2123.9 (1015) 1851.7 (501) CuC 9, C 2178.6(30), 2145.3(168), 2028.0(209), 1818.0(3419), 1666.5(558), 1293.6(71), 925.5(29), 734.7(4), 648.1(0), 642.6(0), 563.2(0), 549.0( 18), 538.4(0), 475.5(0), 376.9(0), 351.4(2), 253.7(4), 251.7(13), 248.4(10), 162.9(1), 161.4(0), 86.4(15), 85.8(15), 30.5(0), 30.0(0) 1750.6 (3419) a Most intense C -C asymmetric stretch modes, only, they are scaled uniformly by 0.9629 factor, transferred from CuC3-nl cluster.

PAGE 78

78 1.827 1.242 1.323 1.278 1.827 1.242 1.310 1.271 1.296 1.829 1.234 1.326 1.242 1.311 1.278 1.828 1.235 1.322 1.249 1.290 1.279 1.290 1.829 1.231 1.329 1.239 1.304 1.257 1.282 1.283 1.286 1.829 1.231 1.330 1.236 1.311 1.247 1.306 1.278 179.7179.9 178.6 179.9 179.8 179.97179.9CuC4, CsCuC5, CsCuC6, C vCuC7, C vCuC8, C vCuC9, C v Figure 4 5. O ptimized equilibrium structures for the CuCn (n = 4 9 ) clusters. The bond lengths () and angles () calculated at MPW1PW91/6 311++G(3df) are marked. The lowest energy equilibrium geometries for the CuCn(n = 4 9 ) c lusters are near linear Cs structures. While for the CuC4 and CuC5 clusters the EZPE (C v) EZPE(Cs) differences are 116 and 80 cm1, respectively, for larger clusters they are smaller than 30 cm1 and decreases with the size of the cluster. For this reas on, and because very similar mode frequencies and their integral intensities are found in both point group symmetr ies the CuCn (n = 6 Fig ure 4 5 as linear ones. Similar to the pure linear carbon clusters some alternation in bond leng ths is noted also in CuCn (n = 4 9) clusters, particularly on Cu side of the cluster. But, generally the carbon -carbon bonding is different in CuCn (n = 4 9) clusters than in Cn (n = 4

PAGE 79

79 9) clusters, with some t riple C 5 ). The bonding differences and mass effect are reflect ed in big changes in the infrared spectra when compared to pure Cn clusters. The most intense C = C asymmetric stretch vibrations in the Cn (n = 3 9) clusters occur in the 2165 (C5) 1544 cm1(C4) regi on, while for the CuCn (n = 3 9) clusters this mode is located in the narrow 1873 1751 cm1 region ( Note: uniform scaling for the harmonic calculated frequencies from nl -CuC3 was applied for all clusters (see Table 4 5 )). It is interesting to note that only the CuCn clusters with an even number of carbons have strong bands in the high energy region. Such clusters have also different C -C bonding than odd number carbons CuCn clusters. This is indicated by increasing number of triple CC bonds in even -numbe red clusters (see Figure 4 5). The most intense calculated ( and scaled) mode frequencies in Table 4 5 can be used to find any energy coincidences with the observed extra bands in the Cu/C/Ar experiment of Fig ure 4 1. T he frequencies of all the strongest ba nds predicted for the CuCn (n = 4 9) are grouped in the narrow energy range of 2124 2022 cm1 and 1873 1751 cm1, they overlap some of the extra bands that were observed in the 2117 2061cm1 and 1896 1787 cm1 range, respectively. Thus, it is ver y likely that those weak experimental bands observed in Figure 4 1 in both regions are due to the CuCn (n = 4 9) clusters. However, these extra bands were very weak, so that 13C isotope labeled experiments did not help in this case because the isotopomer bands are usually much smaller than the all -12C bands. In order to make further assignments of these bands, we tried to produce copper -carbon clusters and trap them in nitrogen matrices instead of the solid argon matrices. Longer carbon chains Cn (n > 3) have been observed to form in nitrogen matrices. For instance, linear C11 clusters were formed and studied in nitrogen matrices. The intensities of the C11 bands were much larger in nitrogen matrices compared to argon matrices so

PAGE 80

80 that 13C isotopic substitu tion experiments were employed to successfully assign the C11 bands.109 Therefore, we ran expriments in nitrogen matrices, expecting to form larger linear carbon clusters and therefore larger CuCn clusters, such as CuC5 and CuC7. However, this approach was not very successful and the bands were still too small to assign. Only theoretical work of longer CuCn (n = 4 9) clusters were presented in this dissertation. Summary After the ablation and trapping of copper and carbon in an argon matrix, new infrared absorption bands wer e observed at 1830.0 and 1250.5 cm1. These bands have been assigned to the two most intense modes, i.e, the asymmetric and symmetric C=C stretches, of the Cu carbon cluster, nl -63CuC3. This assignment was supported by calculations using density functional theory with MPW1PW91 functional and 6 311++G (3df) basis set The MPW1PW91/6311++G (3df) approach predicts the relative integral intensities for these modes quite well. However, the BPW91/6 311++G(3df) [carbons]// SDD pseudopotentials [Cu] and B3LYP/63 11++G(3df) calculations failed to correctly predict the 13C labeled infrared isotopomer frequencies and the relative integral intensities for those modes. The photo induced isotopic scrambling in 63Cu12/13C3 isotopomers was observed and explained via a co mputed (MPW1PW91/6 311++G(3df)) PES for this reaction. The forward reactions of G G* H G indicates that scrambling in the isotopomers 63 121213 (b ) 13 1212 ( d ) and 63 1213 13 (e ) 131312 ( g ) occurs via the bi cyclic H isomer. A t 12 K the estimated forward rate constant (kf) is 2.2 times larger than the backward rate constant (kb), so the accumulation of 63 1312 12 and 6313 1312 isotopomers is expected during matrix photolysis, exactly as observed. A number of pure carbon clu sters larger than C3, seen in Figure 4 1 we re formed in the present experiment s Many bands, mainly in the 1775 1900 cm1 range, in Figure 4 1 b have

PAGE 81

81 been left unassigned. It is likely that many are due to products of reactions of larger carbon clusters with copper. Theoretical calculation using MPW1PW91/6311++G(3df) level of theory predicts that the lowest energy equilibrium geometries for the CuCn(n = 4 9 ) clusters are near linear Cs structures. However, for CuCn(n = 6 9 ) clusters, the energy dif ference between the linear structure and near linear structure are very small (less than 30 cm1) and the difference decreases as the length of the clusters increases. The frequencies of the strongest bands predicted for the CuCn (n = 4 9) are grouped in the range of 2124 2022 cm1 and 1873 1751 cm1.

PAGE 82

82 CHAPTER 5 SILVER C ARBON CLUSTER: STRUC TURE AND INFRARED FR EQUENCIES Introduction The interaction of metal s with carbon has been a topic of long-standing interest. In part, this is the result of the p otential involvement of metal -carbon molecules in catalysis .110,111 Ever since the discovery of the novel metallocarbohedrenes (met -cars) by C astleman and coworkers, metal carbon research has accelerated.85 Met -cars have been formed with the stoichiometry M8C12 where M = Ti, V, Zr, Nb, Mo, Hf, Cr, or Fe but they can al s o be formed with two metals with stoichiometry Ti8 xMxC12, where M is Si, Y, Zr, Nb, Mo, Hf, Ta or W.112 115 Other types of clusters have also been studied. These have included the anionic clusters VmCn (m =1 4, n = 2 8) Co2Cn (n = 2, 3), and Nb2Cn (n = 4 9) studied by photoelectron spectroscopy and density functional theory (DFT) calculation s .116118 Wang and coworkers investigated the first row transition metal C3 clusters MC3 (with M=Sc, V, Cr, Mn, Fe, Co, Ni) as w ell as FeCn (n=3, 4), NbCn (n=2 7) and TiCn (n=2 5) also using photoelectron spectroscopy.119122 Similar studies on MCn clusters (with M=Sc, Y, La; n=5 20) were reported by Kohno et al.123 CunC2 + (n = 2k+1, k = 1 7) and CunC4 + (n = 2k+1, k = 2 4 ) clusters were studied using time -of -flight mass spectrometry.88 Cationic CuCn + clusters (n = 1 3) were generated in spark discharges of Cu and graphite and observed in mass spectrometric studies.84 Matrix isolation vibrational spectroscopy studies of GeC3Ge, TiC3, CrC3, CoC3, AlC3, AlC3Al and NiC3Ni have been reported by Graham and coworkers.7983,124 Copper and Silver polyynides (Cu and Ag capped carb ynes) were characterized using Raman spectroscopy.125 Other properties of metal -carbon clusters including trends in ionization potentials and electronic affinities have been investigated as well.126,127 Studies of CuC3 and its photo-induced isotopic scrambling were

PAGE 83

83 recently reported by us.128 Since silver is similar in electron configuration to copper, it is interesting to see whether similar, or different, silver carbon clusters are formed. Only a few reports on silver -carbon clusters have appeared Sil v er acetylide (AgC one of the oldest organometa llics was synthesized by reacting acetylenic compounds with an ammoniacal silver nitrate solution.129,130 Cationic AgCn + clusters (n = 1 3), generated in a radio frequency spark ion source, were investigated using mass spectrometric methods.84 In this chapter we report the first vibrational spectroscopic study of a silver -carbon cluster trapped in solid argon at 12 K. Corresponding theoretical work on small stable silver -carbon clusters, AgmCn (m = 1, 2; n = 1, 2, 3) using dens ity functional calculations has been carried out in the search for the equilibrium geometry and vibrational frequencies of the cluster gi ving rise to the two new bands observed. Both experiments and calculations indicate that the near linear AgC3 cluster is responsible. Computational and Experimental Details Experimental M ethods Silver -carbon clusters were generated by two -beam laser ablat ion of silver and graphite, or, in the case of the isotopic studies, a pressed pellet of 12C and 13C (ISOTEC). The experimental apparatus used for the generation and trapping of silver -carbon clusters in solid argon is similar to that described in Chapter 2 and Chapter 4 Briefly, the output of a pulsed Nd: YAG laser (1064/532 nm, 0.2 0.5W, 10Hz) was split into two beams with one beam focused on a small piece of silver sample (SPEX), and the other beam focused on the carbon sample. The reaction products were co deposited with argon gas onto a 12K CsI window cooled by a closed -cycle helium cryostat (APD Displex). After 2 3 hours deposition, infrared absorption spectra were collected using a NICOLET Magna 560 FT IR spectrometer (0.5 cm1 resolution). Anneal ing of

PAGE 84

84 the matrix ( heating to 35 K and recooling to 12 K), as well as photolysis with a medium pressure 100W Hg lamp, were also performed to induce secondary reactions. Computational M ethods All calculations were carried out using the Gaussian 03 suite of programs.73 The equilibrium geometries, harmonic vibrational frequencies, and dissociation energies were calculated using the MPW1PW91functional, a modified Perdew -Wang exchange and correlation functional ,74,75 with a SDD (the Stuttgart/Dresden ECPs and D95V basis set for silver and carbon, respectively).131 In previous work, the SDD basis set was chosen to calculate M -NH3 (M = Cu, Ag) and M+-H2S (M = Cu, Ag, Au) systems.132,133 The MPW1PW91 functional was recommended by Wiberg,90 Dunbar,91 and Oomens et al.76 for the calculation of C, H, and metal containing systems. The MPW1PW91 functional was also used in our recent studies of iron complexed with cationic and neutral polycyclic aromatic hydrocarbons .72,92 In our previous work on the CuC3 cluster,128 we compared the results calculated by the MPW1PW91 functional and the B3LYP ( Beckes three -parameter hybrid functional combined with the nonlocal correction functional of Lee, Yang, and Parr ),89 respectively. It showed that c alculations using the MPW1PW91 functional supported the experimental data whereas the B3LYP functional did poorly in predict ing the 13C lab e led isotopomer frequencies and the relative integral intensities of 63Cu12/13C3. Since silver and copper atoms have similar electronic configurations (Ag([Xe]4d105s1) and Cu([Ar]3d104s1)), we tested the MPW1PW91/SSD level of calculation on the near linear CuC3 cluster by comparing them to the MPW1PW91/6311++(3df) results and to the experimental 63Cu12/13C3 isotopomer frequencies.128 The comparison revealed that the maximum isotopomer frequency differences (after scaling) between MPW1PW91/SSD and experiment is 2.3 cm1, with an average difference of 1.3 cm1, comp ared to the 1.6 cm1 and 0.96 cm1 values when the MPW1PW91/6311++(3df) was used.128 Although the

PAGE 85

85 MPW1PW91/SDD numbers are a little higher than the MPW1PW91/6311++(3df) ones, they are still acceptable for isotopic 13C isotopomer frequency matching. For this reason, we used the MPW1PW91/SSD functional/basis set in the presen t work. Experimental Infrared Spectra The infrared absorption spectra in the 1180 1250 cm1 and 17502200 cm1 range for the species laser ablated from graphite and for the species formed by the synchronized dual laser ablation of silver and graphite are displayed in Figure 5 1 in panels a and b, respectively. Various neutral carbon clusters (from C3 C12) were observed in both spectrum a and b.95,96 Bands at 1936.5 cm1 in spectrum b is assigned to C6 .98 Impurities such as CO and H2O are also present in the experiments, due to high laser ablation power.81 A weak band at 1824.4 cm1, assigned to C3H,100,101 was also observed. Two new bands were observed in spectrum b, at 1827.8 and 1231.6 cm1. The higher frequency band i s about three times more intense than the lower frequency one. Both were found to be dependent on silver and carbon concentrations, and are thus attributed to a species containing both silver and carbon. Numerous experiments were performed to determine wh ether the bands belong to the same species. Different Ag/C ratios were affected by varying the ablating laser intensities. The two new bands decrease in intensity with lower ablation power and increase with higher power. Matrix annealing up to 35 K and rec ooling to 12 K increases the intensity of both bands by about 25%. UV -visible photolysis with a medium pressure mercury lamp up to 1 hour decreases both bands about 15%. Under all these different conditions, the integral intensity ratio of the 1827.8 and 1 231.6 cm1 bands remained constant (= 3.03 + 0.3), indicating that they belong to the same species.

PAGE 86

86 2150 2100 2050 2000 1950 1900 1850 1800 1250 1200 -0.05 0.00 0.05 0.10 0.15 0.20 0.25 0.30 AbsorbanceWavenumbers (cm-1) (b). Ag/12C/Ar (a). 12C/Ar1827.8 1817.9C121894.3C71936.5C6 -1952.5C6 1946.0C111998.0C92039.0C32078.1C92127.9C72138.5C O2164.0C5C111856.6 1231.6 1197.2 C6 Figure 5 1. Infrared absorption spectra of products of laser ablation of graphite (spectrum a) and products of two -beam laser ablation of graphite and silver (spectrum b). The spectra were recorded after matrix annealing to 35K then cooling back to 12 K. The major bands due to pure carbon clusters and their reaction products with silver at 1827.8 and 1231.6 cm1 are indicated. Isotopic (13C) substitution is a powerful means of identifying the structure of unknown molecular system. Figure 5 2 shows the spectrum of the reaction products of the laser ablation of Ag and 12C (spectrum a) compared to the spectrum of reaction products of laser ablation of Ag and 12/ 13C mixture (spectrum b). There are clearly eight isotopomeric bands ( a h ) built on the 1827.8 cm1 band in the 13C -labeled spectrum. In the lower energy region (1231.6 cm1 band), eight very weak bands were also observed. They are here tentatively assigned to various isotopomers (see Table 5 2). A similar isotopomer band pattern was previously observed for the near -linear 12/13Cu C3 cluster,128 the near linear 12/13C3H2O complex,104 and the linear 12/13C3Cr

PAGE 87

87 and 12/13C3Co clusters .81,83 This strongly suggests that the 1827.8 cm1 band should be assigned to the CC asymmetric stretching mode in linear or near linear 12/13C3 isotopomers bonded at one end to Ag. This preliminary conclusion is confirmed in the next section by computation of the equilibrium geometries and harmonic frequencie s of a number of silver -carbon clusters, including Ag12/13C3. 1840 1820 1800 1780 1760 1240 1220 1200 1180 0.06 0.08 0.10 0.12 0.14 AbsorbanceWavenumbers (cm-1) 1827.8 1823.3 1805.6 1800.5 1785.5 1781.5 1762.4 1757.6 1231.6 1228.2 1211.9 1209.6 1205.8 1201.7 1185.3 1183.9 a b d f c e g h a b c d e f g h(a). Ag/12C/Ar (b). Ag/12C/13C/Ar [12C] : [13C] = 3 : 1 Figure 5 2. Infrared spectra of reaction products from laser ablation of Ag and 12C (spectrum a) and from laser ablation of Ag and 12/13C (spectrum b). The bands marked by vertical dashed line s are tentatively assigned to isotopomeric partners of the 1231.6 cm1 band. The band s marked by vertical d otted lines are due to: C3H (1824.4 cm1),100,101 C2H+ (1820.2 cm1),102 C12 (1817.9 cm1),103 and C6 (1197.2 cm1).134 Equil ibrium Geometries and Vibrations for AgmCn (m=1, 2; n=1 -3) Clusters. A number of metal -carbon clusters of the type MC3 and MC3M (M = transition metal) have been studied previously.79 83,124,128 Here we explore theoretically the AgC, AgC2, AgC3, Ag2C,

PAGE 88

88 Ag2C2 and Ag2C3 clusters. Figure 5 3 shows the stable equilibrium geometries and relative energies pre dicted (MPW1PW91/SDD) for these clusters. Harmonic vibrational frequencies and integral intensities are displayed in Table 5 1. 0.00 eV A AgC Cv 2.051 2.009 1.247 2.0281.2911.340152.2175.9 2.0272.041 1.301 2.0491.318130.6163.7 2.668 2.150 1.374 92.7 2.616 2.078 2.6482.204 2.000 1.2482.2411.310 0.00 eV B l AgC2, CvC c AgC2, C2vD nl AgC3, Cs0.00 eV 0.04 eV 0.90 eV 0.00 eV 0.00 eV G l Ag2C2, D hE c Ag2C, C2vF l Ag2C, CvH w Ag2C3, C2v0.00 eV 0.91 eV I c Ag2C3, C2v3.23 eV J c Ag2C3, C2v2.051 Figure 5 3. Equilibrium structures for the AgC, AgC2, AgC3, Ag2C, Ag2C2, and Ag2C3 clusters. The bond lengths () and angles () calculated at MPW1PW91/SDD are marked. The relative isomer energies are indicated. The AgC Silver -Carbon Cluster The ground state of diatomic AgC is calculated to be a quartet with a bond length of 2.051 The doublet spin state is higher in energy than the quartet by 0.31 eV. The calculation indicates a very low vibrational integral intensity (9 km/mol), which is probably the reason why AgC was not observed in the experiments.

PAGE 89

89 Table 5 1. Vibrational Frequencies (cm1) and Integral Intensities (km/mol) for Electronic Ground States of AgmCn (m = 1, 2; n = 1, 2, 3) Clusters (displayed in Figure 5 3), Calculated Using MPW1PW91/SDD Functional/basis sets. Ag m C n Isomer MPW1PW91/ SDD A AgC ( X 4 ) 477.2 (9) B l AgC 2 (X 2 ) 2018.5 (34), 425.0 (21), 97.2 (2x2) C c AgC 2 ( X 2 A 1 ) a 1 170 2.2 (16), a 1 369.8 (11), b 2 179.0 ( 0 ) D nl Ag C 3 (X 2A a 860.0 (1 17 ), a 24.1 ( 14 ), a 21.9 ( 35 ), a 35.8 ( 9 ), a 6 (5), a 101.4 (17) E c Ag 2 C ( X 3 A 2 ) a 1 388. 7 ( 2 ), b 2 255.7 ( 4 ), a 1 156.1 ( 0 ) F l Ag 2 C (X 3g) 417.7 (4 ), 158.7 (1 ), 110.7 (1 ), 42.5 (0 ) G l Ag 2 C 2 ( X 1 g ) g 2 092.4 (0), u 6 00.3 ( 78 ), g 2 39 .5 (2x0), g 183.9 (0), u 83.9 (2x5 1 ) H w Ag 2 C 3 (X 1A 1 ) b 2 1761. 1 ( 639 ), a 1 1265.7 ( 11 ), a 1 547.3 ( 0 ), b 2 449.7 ( 44 ) b 1 2 89.9 (2 ), a 1 261.5 (9 ), a 2 2 15.9 (0), b 2 115.0 (1 55), a 1 36.4 (4 ) I c Ag 2 C 3 ( X 1 A 1 ) a 1 13 90.4 (7 21 ), b 2 1 297.4 ( 0 ), a 1 355 .3 ( 1 2 0 ), b 2 335.4 ( 30 ), b 1 334.2 (4), a 1 3 17.3 (1), a 1 150.3 ( 2 ) b 2 134.8 (1 0 ), a 2 120.7 (0) J c Ag 2 C 3 ( X 1 A 1 ) a 1 1 705.6 ( 103 ), a 1 519.1 (1 8 ), b 2 510.0 ( 0 ), a 1 4 65.3 ( 7 ), b 2 4 52.1 ( 0 ), a 2 233.3 (0), b 2 1 87.7 ( 4 3) b 1 174.0 ( 22 ), a 1 1 01.3 ( 0 ) The AgC2 Silver -Carbon Cluster Two structures, B and C in Figure 5 3, are predicted for AgC2. Linear l -Ag C2 (B) is marginally more stable by 0.04 eV than cyclic c -Ag C2 (C ). Table 5 1 shows that the strongest infrared modes in l Ag C2 and c -Ag C2 lie at 2018.5 and 1702.2 cm1, respectively. However, since the calculated integral intensities are relatively low, 34 and 16 km/mol, the absence of these two bands i s understandable. The AgC3 Silver -Carbon Cluster Only one stable structure ( D ) is found for the AgC3 cluster: the near linear one, as displayed in Figure 5 3. The bond lengths and angles calculated using MPW1PW91/SDD level are marked in the figure. The n l -AgC3 cluster has a doublet spin multiplicity. Searches for

PAGE 90

90 doublet bicyclic -AgC3 and cyclic -C3 bonded to Ag each resulted in one imaginary frequency. Predicted vibrational frequencies (unscaled) and integral intensities for nl -AgC3 are 1860.0 cm1 (117 km/mol) (asymmetric CC stretch mode) and 1224.1 cm1 (14 km/mol) (symmetric CC stretch mode). The intensity ratio of the two bands is relatively high (8.36) compared to experiment (3.03 + 0.3). The B3LYP functional with SDD basis set was also tested on nl -A gC3. The predicted vibrational frequencies (and integral intensities) are 1864.9 (133) and 1211.4 cm1 (9 km/mol), an even worse prediction of the intensity ratio. To confirm the assignment of the observed bands to the near linear species, isotopic substitution experiments were run. Figure 5 2 shows bands due to the 12, 13C -isotopomers built on the 1827.8 and 12 31.6 cm1 bands. T he comparison between the observed isotopomer bands and the predicted (scaled) nl -107Ag12/13C3 frequencies are listed in Table 5 2 As found in the study of CuC3 clusters, t he harmonic vibrational frequencies predicted by the MPW1PW91/ SDD calculations is in good agreement with the assignment of these bands to nl Ag C3 (X 2A (structure D ) while the B3LYP/ SDD match is again poorer. The maximum differences between experimental and predicted isotopomer frequencies, expsc, are 4.3 cm1 for B3LYP/SDD and 1.8 cm1 for MPW1PW91/ SDD The average values of these differences for all observed isotopomers (for both modes) are 1.3 (B3LYP) and 0. 63 cm1 (MPW1PW91). Although the 1.8 and 0. 63 cm1values are typical for structures assigned using isotopic 13C labeling, the 4.3 and 1.3 cm1 B3LYP values are too large to be acceptable Thus, based on the comparison with the MPW1PW91 calculation, we conclude that nl -Ag C3 (structure D ) was formed in our experiments and is responsible for the two bands at 1827.8 and 1231.6 cm1.

PAGE 91

91 T able 5 2 Comparison of Experimental and Calculated Isotopomer Frequencies (Integral Intensities) for the Asymmetric C=C Stretch and Symmetric C=C Stretch Fundamental Modes of Near Linear 107Ag 12/13C3. Isotopomer exp a /cm1 sc b /cm 1 (km mol 1 ) exp sc / cm 1 sc c /cm 1 (km mol 1 ) exp sc / cm 1 B3 LYP/SDD MPW1PW91/SDD Asymmetric C=C stretch mode a 107 12 12 12 1827.8 1827.8 (133) 0.0 1827.8 (117) 0.0 b 107 12 12 13 1823.3 1822.0 (132) 1.3 1823.2 (115) 0.1 c 107 12 13 12 1785.5 1782.9 (125) 2.6 1784.1 (108) 1.4 d 107 13 12 12 1805.6 1808.1 (132) 2.5 1805.7 (118) 0.1 e 107 12 13 13 1781.5 1777.2 (124) 4.3 1779.7 (106) 1.8 f 107 13 12 13 1800.5 1801.6 (131) 1.1 1800.5 (116) 0.0 g 107 13 13 12 1762.4 1762.3 (125) 0.1 1760.9 (110) 1.5 h 107 13 13 13 1757.6 1755.9 (123) 1.7 1755.8 (108) 1.8 Sy mmetric C=C stretch mode a 107 12 12 12 1231.6 1231.6 (9) 0.0 1231.6 (14) 0.0 b 107 12 12 13 1205.8 1205.3 (9) 0.5 1204.8 (15) 1.0 c 107 12 13 12 1228.2 1230.0 (9) 1.8 1229.2 (14) 1.0 d 107 13 12 12 1211.9 1210.4 (8) 1.5 1211.7 (12) 0.2 e 107 12 13 13 1201.7 1203.2 (9) 1.5 1201.9 (15) 0.2 f 107 13 12 13 1185.3 1184.5 (8) 0.8 1185.3 (12) 0.0 g 107 13 13 12 1209.7 1209.3 (8) 0.4 1210.0 (12) 0.3 h 107 13 13 13 1183.9 1183.1 (8) 0.8 1183.2 (13) 0.7 a Experimental band energies for symmetric mode are tentative due to their weak intensities. b Frequencies are scaled uniformly by scaling factor of 0.9801 for asymmetric mode and by 1.0167 for symmetric mode. c Frequencies are scaled uniformly by scaling factor of 0.9795 for asymmetric mode and by 1.0061 for symmetric mode. When silver bonds to either end of the same singly-substituted isotopic precursor, i.e., 12/13C3 (12 1213), both b (1071212 13) and d (10713 1212) isotopomers may be formed. Figure 5 2 shows that the b and d bands have compara ble intensities, as expected. Similarly, when Ag attaches to the C3 (12 1313) precursor, both e (10712 1 3 13) and g (107131312) isotopomers are formed. Again, Figure 5 2 shows that the e and g bands have similar intensities. In our previous study of C u C3, it was found that the analogous sets of isotopomeric bands were not of equal intensity. It was shown that this resulted from photoinduced isotopic scrambling of 13C and 12C isotopes in the cluster. The present observation that the bands in the b d s et and in

PAGE 92

92 the e g set retain their equal intensities leads to the conclusion that no photo induced isotopic scrambling takes place in AgC3. The photoscrambling in CuC3 and C3 involves a cyclic intermediate,106,107,128 but such a stable cyclic structure could not be found for AgC3 It is interesting to compare the structural parameters and binding energies of the nl -AgC3 and nl -Cu C3 clusters, both calculated at the MPW1PW91/SDD level. The metal -carbon bond lengths (BL) in nl -AgC3 are much longer than in nl CuC3. The BL ( Ag C) = 2.028 vs BL (Cu C) = 1.810 On the other hand, the carbon -carbon BLs are similar. In AgC3, the CC BLs are 1.291 and 1.340 while in CuC3, they are 1.287 and 1.342 The bond angle AgC C in nl AgC3 is significantly smaller (152.2 ) than in nl CuC3 (160.6 ). The longer bond lengths in nl AgC3 reflect its lower dissociation energy (1.77 eV) compared to 2 .38 eV for nl -CuC3. The Ag -C bond is weaker than the Cu C bond by 0.61 eV, reflecting the larger size of the silver atom. The Ag2C Silver -Carbon Cluster The lowest energy isomers of Ag2C cluster are triplets of E and F Our calculation using MPW1PW91/SDD shows t he F (linear) isomer is higher in energy by 0.90 eV No Ag2C was observed in our experiments because t he highest IR integral intensity for the more stable cyclic isomer E is only 4 km/mol and all predicted frequencies are out of the energy range accessible by our FT -IR instrument The Ag2C2 Silver -Carbon Cluster Th e linear Ag C2Ag cluster (G ), also known as silver acetylide, is the only stable structure found. T his species is p redicted to appear at 6 00.3 cm1 with 78 km/mol integral intensity, but is not observed. The Ag2C3 Silver -Carbon Cluster Three stable C2v isomers were found for Ag2C3; all are singlets The w-shape d one (H ) has the lowest energy, and the other two with cyclic structures (I and J ) are less stable by 0.91 and

PAGE 93

93 3.23 eV, respectivel y. Table 5 1 shows the harmonic vibrational frequencies and integral intensities for all three. T he most intense mode predicted for H is 17 61.1 cm1 (639 km/mol), the asymmetric CC stretch. Although the integral intensities for structures H I and J are a ll very large, we did not observe any bands assignable to these clusters Summary New silver -carbon species were sought by simultaneous dual laser ablation of silver and carbon followed by trapping in solid argon matrices at 12 K. Two new bands were observed at 1827.8 and 1231.6 cm1. Based on the results of isotopic 13C -substitution experiments and density functional theory calculations with the MPW1PW91/SDD functional/basis set, these two bands were assigned to the asymmetric and symmetric CC stretching m odes in the near linear AgC3 cluster. The calculated dissociation energy at the MPW1PW91/SDD level for the nl -AgC3 is 1.77 eV, which is lower by 0.61 eV than in nl CuC3. The weaker binding in nl -AgC3 is reflected in its longer metal carbon bonds compared to the ones in nl -CuC3. No photo induced isotopic scrambling in 107Ag12/ 13C3 isotopomers was observed as was found in Cu12/13C3, a finding consistent with the fact that no stable cyclic structure of doublet AgC3 could be found theoretically. Such a struct ure was previously determined to be essential to the photoscrambling found in CuC3. The equilibrium structures and vibrational frequencies for AgC, AgC2, Ag2C, Ag2C2 and Ag2C3 were also calculated using MPW1PW91/SDD, but none were observed in our experimen ts. Though there remain bands unassigned in our spectra, mainly in the 1750 1900 cm1 range, it is likely that they are due to larger clusters of AgCn (n > 3).

PAGE 94

94 CHAPTER 6 GOLD CARBON CLUSTER: STRU CTURE AND INFRARED F REQUENCIES Introduction Gold is a c lassic noble metal and seldom reacts with other molecules or atoms. Therefore, reactions of gold and small molecules are always of great interest. For example, extensive efforts have been made to study the reaction of gold with carbon monoxide due to its potential application in catalysis. The Andrews group has reported Au(CO)n (n=1, 2) and Au(CO)n + (n=1 4) clusters trapped in a neon matrix.135 Xu and coworkers further st udied the reaction of Au and CO and discovered Aun(CO) (n=1 5) and Aun(CO)2 (n=1, 2) clusters .136 A photoelectron spectroscop ic and quasi relativistic density functional theory (DFT) study of comp lexes of the Au6(CO)n and Au6(CO)n (n=0 3) complexes was reported by Wang et al.137 Interaction of CO with cationic gold clusters in the gas phase was studied using vibrational spectroscopy by the Rayner group.138,139 Recently, Duncan and coworkers produced Au+(CO)n complexes in the gas phase and investigated them using IR photodissociation spectroscopy.140 Infrared spectra of analogous gold thiocarbonyl complexes, Au(CS)n (n=1, 2) and Au2CS were also studied.141 In their investigation of the reactions of gold with hydrogen, Wang et al. reported that gold hydride anions are stable.94,142 The g old atom is found to be in three different oxidation states in the gold dihalides AuX2 /0/+ ( with X=Cl, Br).143 Other reactions, such as gold with nitrous oxide N2O were also studied in excess argon using matrix infrared spectroscopy.144 Our work on copper -carbon and silver -carbon clusters was discussed in the two previous chapters Since gold is similar in electron configuration to copper and silver i t was of interest to see whether gold react s similarly with carbon or form s other, more novel molecules or clusters. Gold -carbon clusters, such as A u +, Au=C=Au2+ and XAuC (where X = F, Cl, Br, I) have been previously studied theoretically .145 147 However, experimental work is severely

PAGE 95

95 lacking. To our knowledge, the study of the photodissociation of AuCn + by Duncan and coworkers is the only experimental investigation published to date .148 In this chapter we present a vibrational spectroscopic study of a gold carbon cluster trapped in solid argon at 12 K, also present complementary theoretical work on gold -carbon clusters. The e quilibrium geometry, vibrati onal frequencies and 13C lab e led isotopomer frequencies have been calculated using d ens ity functional theory and compared to the experimental value s Computational and Experimental Details Experimental M ethods Gold -carbon clusters were generated by two -be am laser ablation of gold and graphite. The experimental apparatus used for the generation and trapping of gold -carbon clusters in solid argon is similar to that used previously for producing copper -carbon clusters and silver -carbon clusters: The output of a pulsed Nd: YAG laser (1064/532 nm, 0.2 0.5W, 10Hz) was split into two beams with one beam focused on a small piece of gold sample ( Lesker ), and the other beam focused on the graphite sample or a pressed pellet of 12C and 13C (ISOTEC) in the case of t he isotopic studies. The reaction products were codeposited with argon gas onto a 12K CsI window cooled by a closed -cycle helium cryostat (APD Displex). After 2 3 hours deposition, infrared absorption spectra were collected using a NICOLET Magna 560 FT IR spectrometer (0.5 cm1 resolution). T he matrix was annealed to 35 K and recool ed to 12 K and was photolyzed with a medium pressure 100W Hg lamp, to induce secondary reactions. Computational M ethods Using the Gaussian 03 suite of programs,73 the equilibrium geometries, harmonic vibrational frequencies and dissociation energies were calculated using DFT theory, including the MPW1PW91, BPW91 and B3LYP functional, along with different basis sets. The MPW1PW91functional, a modified Perdew -Wang exchange and correlation functional ,74,75 was

PAGE 96

96 used with a SDD basis set.131 The MPW1PW91 functional was recommended by Wiberg,90 Dunbar,91 and Oomens et al.76 for the calculation of C, H, and metal -containing systems. The MPW1PW91 functional was also used in our recent studies of iron complexed with cationic an d neutral polycyclic aromatic hydrocarbons .72,92 The SDD basis set combines the Stuttgart Dresden effective core potential (ECP) for gold with the D95V basis set for carbon.131 In previous work, the SDD basis se t was chosen to study the M NH3 (M = Cu, Ag) and M+H2S (M = Cu, Ag, Au) systems.132,133 The MPW1PW91/S DD level of theory successfully predicted the 13C lab e led isotopomer frequencies in our recent study of the silver -carbon cluster, AgC3.149 Since gold, silver and copper atoms have similar electronic configurations (Au([Xe]4f145d106s1), Ag([Kr]4d105s1) and Cu([Ar]3d104s1)), we first calculate d using the MPW1PW91/SSD level of calculation on the near linear AuC3 cluster and compared them to the experimental 197Au12/13C3 isotopomer frequencies. The LanL2DZ basis set (D95 on carbon and Los Alamos ECP plus DZ on gold)150 152 was also used in calculations involving silver and gold.135 The MPW1PW91/LanL2DZ method was also tested. We also use the MPW1PW91 with SDDAll (Selects Stuttgart potentials for both gold and carbon atoms) basis set for calculations of the 13C lab e led isotopomer frequencies for near -linear AuC3 clusters. We also t ested the BPW91 functional with SDD and LanL2DZ basis set s on the near linear AuC3 culster. Such a functional has been recommended for calculations of copper, silver, and gold compounds by Wang et. al.94 who calculate d the reaction products gold with hydrogen us ing BPW91 density functional, 6 311++G(d,p) basis set for hydrogen and SDD pseudopotentials for the metal atom. In conclusion we used the BPW91/6 311++G(3df)/SDD pseudopotentials level of theory (6 311++G(3df) for carbon and SDD pseudopotentials for gold)

PAGE 97

97 Finally, different combinations of MPW1PW91, BPW91, B3LYP functional with 6 311++G(3df) basis set for carbon, SDD or LanL2DZ basis sets for gold (without pseudopotentials) were used to calculate 13C -lab e led isotopomer frequencies Experimental Infrared Spectra The infrared absorption spectra in the 1250 1300 cm1 and 17502200 cm1 range for the species laser ablated from graphite and for the species formed by the synchronized dual laser ablation of silver and graphite are displayed in Figure 6 1 in pane ls a and b, respectively. A number of bands observed in spectrum a and b have been previously assigned to the C3, C5, C6, C7, C9, C11 and C12 clusters .95,96 The 1936.5 cm1 band, assigned to C6 .98 was observed in spectrum b but not in spectrum a. Due to high laser ablation power, impurities such as CO and H2O are also pre sent.81 A we ak band at 1824.4 cm1, assigned to C3H,100,101 was also observed. Two new bands were observed in spectrum b, at 1845.2 and 1275.7 cm1. The higher frequency band is about five times more intense than the lower frequency one. Both were found to be dependent on silver and carbon concentr ations, and are thus attributed to a species containing both silver and carbon. In order to determine whether these two bands belong to the same species, a number of experiments were performed under different experimental conditions. First, different Au/C ratios were applied by varying the ablating laser intensities. The two new bands decrease in intensity with lower ablation power and increase with higher power. Second, the intensities of both bands increase by about 25% after matrix annealing up to 35 K and recooling to 12 K. And last, UV visible photolysis with a medium pressure mercury lamp up to 1 hour decreases both bands about 15%. The integral intensity ratio of the 1845.2 and 1275.7 cm1 bands remained almost constant (= 7.25+ 0.3) under these diffe rent conditions, supporting the conclusion that they belong to the same species.

PAGE 98

98 2150 2100 2050 2000 1950 1900 1850 1800 1300 1250 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 AbsorbanceWavenumbers (cm-1) (b). Au/12C/Ar (a). 12C/Ar1845.2 1817.9C121894.3C71936.5C6 -1952.5C6 1946.0C111998.0C92039.0C32078.1C92127.9C72138.5C O2164.0C5C111856.6 1275.7 Figure 6 1. Infrared absorption spectra of products of laser ablation of graphite (spectrum a) and products of two -beam laser ablation of graphite and gold (spectrum b). Th e spectra were recorded after matrix annealing to 35K then cooling back to 12 K. The major bands due to pure carbon clusters and their reaction products with gold at 18 45.2 and 1275.7 cm1 are indicated. Isotopic (13C) substitution experiments were perform ed again to identify the structure of the unknown species, similar to what was done in the study of copper carbon clusters and silver carbon clusters. Figure 6 2 shows the spectrum of the reaction products of the laser ablation of Au and 12C (spectrum a) c ompared to the spectrum of reaction products of laser ablation of Au and 12/13C mixture (spectrum b). Eight isotopomeric bands ( a h ) built on the 1845.2 cm1 band were clearly observed in the 13C -labeled spectrum. In the lower energy region (1275.7 cm1 band), five very weak bands were also observed. They are here tentatively assigned to va rious isotopomers (see Table 6 2 through 6 5 ).

PAGE 99

99 1840 1820 1800 1780 1280 1260 1240 1220 0.22 0.23 0.24 0.25 0.26 AbsorbanceWavenumbers (cm-1) 1816.2 1845.2 1840.9 1820.9 1804.1 1800.2 1778.5 1774.8 1275.7 1271.0 1254.9 1251.3 1226.2 a b d f c e g h a c d g h(a). Au/12C/Ar (b). Au/12C/13C/Ar [12C] : [13C] = 3 : 1 Figure 6 2. Infrared spectra of reaction pr oducts from laser ablation of Au and 12C (spectrum a) and from laser ablati on of A u and 12/13C (spectrum b). The bands marked by vertical dashed and dotted lines are tentatively assigned to isotopomeric partners of the 1275.7 cm1 band. The bands marked by vertical d otted lines are due to: C3H (1824.4 cm1),100,101 C12 (1817.9 cm1) ,103 Figure 6 3 shows the infrared spectra of reaction products from laser ablation of Cu and 12/13C (spectrum a), Ag and 1213C (spectrum b) and Au and 12/13C (spectrum c) in the 17501850 cm1 energy region, showing very similar isotopomer band pattern s with only small energy shift s We note that similar isotopomer band p attern was also previously observed for the near linear 12/13C3H2O complex,104 and the linear 12/13C3Cr and 12/13C3Co clusters .81,83 S uch a band pa t t ern is characteristic of the C=C asymmetric stretch mode in 12/13C3 with its one end bonded to the ligand in a near -linear or linear cluster geometry. This strongly suggests that the 1845.2 cm1

PAGE 100

100 band should be assigned to the CC asymmetric stretching mode in linear or near -linear 12/13C3 isotopomers bonded at one end to Au. 1840 1820 1800 1780 1760 0.16 0.18 0.20 0.22 AbsorbanceWavenumbers (cm-1) (c). Au/12C/13C/Ar [12C] : [13C] = 3 : 1 (b). Ag/12C/13C/Ar [12C] : [13C] = 3 : 1 (a). Cu/12C/13C/Ar [12C] : [13C] = 3 : 1 Figure 6 3 Infrared spectra of reaction pr oducts from laser ablation of Cu and 12/13C (spectrum a), Ag and 12/13C (spectrum b) and Au and 12/13C (spectrum c ) in 17501850 cm1 region. The bands marked are assigned to isotopomeric bands for Cu12/13C3, Ag12/13C3 and Au12/13C3. After inspection of the 13C labelled spectra of Figure 6 2 we discuss why the lower frequency mode (1275.7 cm1 band) we see only five isotopomer bands (labeled a, c, d, g and h), while there are eight istopomeric bands in the higher energy region (1845.2 cm1) Since the intensity of the band at 1275.7 cm1 is about one -fifth of the 1845.2 cm1 band, it is expected t hat the intensities of the isotopomer bands built on the 1275.7 cm1 band will also be about one -fifth of those built on 1845.2 cm1 band. But the isotopomer bands b, f, e are already very small.

PAGE 101

101 Therefore, the absence of the isotopomer bands in the lower frequency region is understandable. In the next section, we discuss the agreement between the experimental resul ts and theoretical calculation. When gold bonds to either end of the same singly substituted isotopic precursor, i.e., 12/13C3 (12 1213), b (19712 1213) and d (197 131212) isotopomers are formed. Figure 6 2 shows that the b and d bands have comparabl e intensities, as expected. Similarly, when Au attaches to the C3 (12 13 13) precursor, e (10712 1 3 13) and g (10713 13-12) isotopomers are formed. Again, Figure 6 2 shows that the e and g bands have similar intensities. This is consistent with our previ ous study of AgC3, which indicates that no photo induced isotopic scrambling takes place in AuC3.149 Theoretical Calculations for Gold -Carbon AuC3 Cluster The near linear structure is the only stable structure found for the AuC3 cluster, as displayed in Figure 6 4. The bond lengths and a ngle s calculated using MPW1PW91/SDD and MPW1PW91/LanL2DZ functional/basis sets are marked in the figure. The near linear AuC3 cluster has a doublet spin multiplicity. Predicted vibrational frequencies (unscaled) and intergral intensities for near linear AuC3 u sing different functional/basis sets are listed in Table 6 1. For all calculations, the most intense band is in 18501900 cm1 region (asymmetric C=C stretch mode); A m uch weaker band is predicted in the 1200 cm1 region. This is consistent with our experi mental data: a strong band appears at 1845.2 cm1 and a weaker band at 1275.7 cm1. The intensity ratio of the two bands is relatively high (20.492.3) compared to experiment value (7.25) independent of the functional/basis sets used. T he MPW1PW91/SDD, MPW 1PW91/LanL2DZ and MPW1PW91/SDDAll functional/basis sets gave the best predictions of the relative intensity ratio of the two bands (29.1, 23.6 and 20.4, respectively). The BPW91/6 311++G(3df)/SDD(peudopotential) level of calculation gave the worst results (92.3).

PAGE 102

102 Table 6 1. Vibrational frequencies (cm1) and integral intensities (km/mol) for electronic ground states of near linear Au C3 clusters (displayed in Figure 6 4 ), calculated using different functional/basis sets Functional/basis sets Near linear Au C 3 ( X 2 A MPW1PW91/SDD a 911.6 ( 204 ), a 64.3 ( 7 ), a 511.7 ( 36 ), a 56.7 ( 2 ), a 66.7 ( 4), a 138.5 ( 4 ) MPW1PW91/LanL2DZ a 916.4 ( 212 ), a 71.4 ( 9 ), a 525.0 ( 45 ), a 68.0 ( 3 ), a 73.6 ( 4), a 141.9 ( 6 ) MPW1PW91/SDDALL a 960.3 ( 204 ), a 12 80.7 ( 10 ), a 501.1 ( 34 ), a 43.1 ( 4 ), a 29.8 ( 3), a 135.9 ( 3 ) BPW91/SDD a 885.0 ( 343 ), a 18.4 ( 4 ), a 515.4 ( 16 ), a 37.8 ( 2 ), a 30.5 ( 4), a 132.8 ( 0 ) BPW91/LanL2DZ a 889.5 ( 326 ), a 27.2 ( 6 ), a 531.9 ( 26 ), a 50.5 ( 2 ), a 40.0 (4), a 137.9 (1 ) MPW1PW91/ 6 311++G(3df)/SDD a 943.1 ( 252 ), a 97.8 ( 5 ), a 523.7 ( 33 ), a 43.9 ( 6 ), a 195.1 ( 9), a 137.6 ( 3 ) MPW1PW91/ 6 311++G(3df)/LanL2DZ a 941.8 ( 268 ), a 99.4 ( 5 ), a 523.5 ( 38 ), a 43.4 ( 7 ), a 02.1 (9), a 141.0 (4 ) BPW9 1/ 6 311++G(3df)/SDD(peudo) a 922.8 ( 369 ), a 54.3 ( 4 ), a 527.0 ( 21 ), a 28.3 ( 4 ), a 173.1 ( 9), a 131.6 ( 1 ) B3L YP/ 6 311++G(3df)/LanL2DZ a 949.9 ( 315 ), a 86.4 ( 7 ), a 513.9 ( 34 ), a 36.7 ( 4 ), a 08.2 ( 9), a 135.4 ( 2 ) We also cal culated the 13C -lab e led isotopomer frequencies for the near linear AuC3 cluster using all the functional/basis sets mentioned above. The comparison between the observed isotopomer bands and the predicted (scaled) near -linear 197Au12/13C3 frequencies are li sted in Table s 6 2 through 6 5. In the silver -carbon clusters and copper carbon clusters, calculations with the MPW1PW91/SDD level of theory showed good agreement with experimental data. However, none of the calculation methods predict 13C -lab e led isotopom er frequencies for the near -linear AuC3 cluster very well. The MPW1PW91/SDD, MPW1PW91/LanL2DZ functional/basis sets give relatively good results. But the maximum differences between experimental and predicted isotopomer frequencies, expsc, are 7.0 cm1 for MPW1PW91/SDD and 6.3 cm1 for MPW1PW91/LanL2DZ, which is still too large to be accepted. Typical value of expsc is usually within 2.0 cm1. Using different basis sets on carbon atom and gold atom does not help

PAGE 103

103 either, as the maximum differenc es between experimental and predicted isotopomer frequencies, expsc, are 9.7 cm1 for MPW1PW91/6 311++G(3df)/SDD and 9.6 cm1 for MPW1PW91/6 311++G(3df)/LanL2DZ, which is an even worse prediction. T able 6 2. Comparison of Experimental and Calculated (at BPW91/SDD and BPW91/LanL2DZ level) Isotopomer Frequencies (Integral Intensities) for the Asymmetric C=C Stretch and Symmetric C=C Stretch Fundamental Modes of Near Linear 197A u 12/13C. X Isotopomer exp /cm1 sc a /cm1 exp sc /cm 1 sc b /cm1 exp sc /cm 1 BPW91/SDD BPW91/LanL2DZ Asymmetric C=C Stretch Mode a 197 12 12 12 1845.2 1845.2 (343) 0.0 1845.2 (326) 0.0 b 197 12 12 13 1840.9 1837.0 (339) 3.9 1837.4 (322) 3.5 c 197 12 13 12 1804.1 1798.9 (329) 5.2 1799.3 (313) 4.8 d 197 13 12 12 1820.9 1828.6 (336) 7.7 1827.9 (317) 7.0 e 197 12 13 13 1800.2 1790.7 (325) 9.5 1791.5 (309) 8.7 f 197 13 12 13 1816.3 1819.7 (332) 3.4 1819.4 (314) 3.1 g 197 13 13 12 1778.6 1781.6 (321) 3.0 1781.2 (304) 2.6 h 197 13 13 13 1773.8 1772.6 (317) 1.2 1772.6 (301) 1.2 Symmetric C=C Stretch Mode a 197 12 12 12 1275.7 1275.7 (4) 0.0 1275.7 (6) 0.0 b 197 12 12 13 1251.8 (4) 1251.9 (5) c 197 12 13 12 1271.1 1274.4 (4) 3.3 1274.2 (5) 3.1 d 197 13 12 12 1254.9 1250.0 (4) 4.9 1250.1 (6) 4. 8 e 197 12 13 13 1250.0 (4) 1250.0 (5) f 197 13 12 13 1226.7 (4) 1226.8 (5) g 197 13 13 12 1251.3 1249.3 (4) 2.0 1249.3 (6) 2.0 h 197 13 13 13 1226.2 1225.5 (3) 0.7 1225.6 (5) 0.6 a Frequencies are scaled uniformly by scaling factor of 0.9788 fo r asymmetric mode and by 1.0470 for symmetric mode. b Frequencies are scaled uniformly by scaling factor of 0.9765 for asymmetric mode and by 1.0395 for symmetric mode. The Andrews group used the BPW91 density functional, and 6 311++G(d,p) for hydrogen an d SDD peudopotentials for gold to successfully calculate the gold hydrides AuHn, and their anions. However, a similar level of theory, BPW91/6311++G(3df)/SDD (pseudopotential) gave the worst results in our study of the near linear AuC3 clusters. The maxi mum difference between experimental and predicted isotopomer frequencies, expsc, is as large as 11.5 cm1.

PAGE 104

104 T able 6 3. Comparison of Experimental and Calculated (at MPW1PW91/SDD, MPW1PW91/SDDAll and MPW1PW91/LanL2DZ level) Isotopomer Frequencies (Integral Intensities) for the Asymmetric C=C Stretch and Symmetric C=C Stretc h Fundamental Modes of Near Linear 197A u 12/13C. X Isotopomer exp /cm-1 sc a /cm -1 exp sc /cm -1 sc b /cm -1 exp sc /cm -1 sc c /cm -1 exp sc /cm -1 MPW1PW91/SDD MPW1PW91/SDDALL MPW1PW91/LanL2DZ Asymmetric C=C Stretch Mode a 197 12 12 12 1845.2 1845.2 (204) 0.0 1845.2 (180) 0.0 1845.2 (212) 0.0 b 197 12 12 13 1840.9 1838.5 (202) 2.4 1838.3 (178) 2.6 1838.8 (210) 2.1 c 197 12 13 12 1804.1 1799.7 (194) 4.4 1799.4 (171) 4.7 1800.1 (210) 4.0 d 197 13 12 12 1820.9 1826.4 (200) 5.5 182 6.8 (175) 5.9 1825.7 (207) 4.8 e 197 12 13 13 1800.2 1793.2 (192) 7.0 1792.6 (170) 7.6 1793.9 (200) 6.3 f 197 13 12 13 1816.3 1819.0 (198) 2.7 1819.3 (174) 3.0 1818.6 (205) 2.3 g 197 13 13 12 1778.6 1779.9 (190) 1.3 1780.2 (167) 1.6 1779.6 (197) 1.0 h 197 13 13 13 1773.8 1772.6 (188) 1.2 1772.6 (166) 1.2 1772.6 (195) 1.2 Symmetric C=C Stretch Mode a 197 12 12 12 1275.7 1275.7 (7) 0.0 1275.7 (10) 0.0 1275.7 (9) 0.0 b 197 12 12 13 1250.5 (7) 1250.2 (10) 1250.6 (8) c 197 12 13 12 1271.1 127 4.0 (7) 2.9 1274.1 (10) 3.0 1273.8 (9) 2.7 d 197 13 12 12 1254.9 1251.9 (7) 3.0 1251.8 (10) 3.1 1252.0 (8) 2.9 e 197 12 13 13 1248.2 (7) 1248.2 (10) 1248.0 (9) f 197 13 12 13 1227.2 (7) 1226.9 (9) 1227.3 (8) g 197 13 13 12 1251.3 1250.8 (7) 0.5 1250.8 (10) 0.5 1250.7 (8) 0.6 h 197 13 13 13 1226.2 1225.6 (7) 0.6 1225.5 (9) 0.7 1225.6 (8) 0.6 a Frequencies are scaled uniformly by scaling factor of 0.9653 for asymmetric mode and by 1.0090 for symmetric mode. b Frequencies are scaled uniformly by scaling factor of 0.9413 for asymmetric mode and by 0.9961 for symmetric mode. c Frequencies are scaled uniformly by scaling factor of 0.9628 for asymmetric mode and by 1.0034 for symmetric mode.

PAGE 105

105 T able 6 4. Comparison of Experimental and Calculated (at MPW1PW91/6-311++G(3df)/SDD and MPW1PW91/6311++G(3df)/LanL2DZ level) Isotopomer Frequencies (Integral Intensities) for the Asymmetric C=C Stretch and Symmetric C=C Stretch Fundamental Modes of Near Linear 197A u 12/13C. X Isotopomer exp /cm1 sc a /c m1 exp sc /cm1 sc b /cm1 exp sc /cm1 MPW1PW91/ 6 311++G(3df)/SDD MPW1PW91/ 6 311++G(3df)/LanL2DZ Asymmetric C=C Stretch Mode a 197 12 12 12 1845.2 1845.2 (252) 0.0 1845.2 (268) 0.0 b 197 12 12 13 1840.9 1837.1 (250) 3.8 1837.2 (265) 3 .7 c 197 12 13 12 1804.1 1798.6 (240) 5.5 1798.7 (255) 5.4 d 197 13 12 12 1820.9 1828.9 (247) 8.0 1828.8 (262) 7.9 e 197 12 13 13 1800.2 1790.5 (237) 9.7 1790.8 (252) 9.6 f 197 13 12 13 1816.3 1820.0 (245) 3.7 1820.0 (260) 3.7 g 197 13 13 12 1778. 6 1781.5 (235) 2.9 1781.4 (249) 2.8 h 197 13 13 13 1773.8 1772.6 (233) 1.2 1772.6 (247) 1.2 Symmetric C=C Stretch Mode a 197 12 12 12 1275.7 1275.7 (5) 0.0 1275.7 (5) 0.0 b 197 12 12 13 1251.0 (5) 1251.1 (5) c 197 12 13 12 1271.1 1274.5 (5) 3.4 1274.5 (5) 3.4 d 197 13 12 12 1254.9 1250.6 (5) 4.3 1250.7 (5) 4.2 e 197 12 13 13 1249.5 (5) 1249.5 (5) f 197 13 12 13 1226.6 (5) 1226.6 (5) g 197 13 13 12 1251.3 1249.9 (5) 1.4 1250.0 (5) 1.3 h 197 13 13 13 1226.2 1225.6 (5) 0.6 1225.5 (5) 0.7 a Frequencies are scaled uniformly by scaling factor of 0.9496 for asymmetric mode and by 0.9830 for symmetric mode. b Frequencies are scaled uniformly by scaling factor of 0.9503 for asymmetric mode and by 0.9818 for symmetric mode. 1.9401.2961.329145.6173.3 1.9301.2951.330146.3174.2 Figure 6 4. Optim ized equilibrium structure for the A uC3 cluster The bond lengths () and angles () calculated at MPW1PW91/ LanL2DZ (italic type, top) and at MPW1PW91/SDD (normal type) are marked. Overall, the results calculated with MPW1PW91 functional are better than th ose with BPW91 or B3LYP functional. Therefore, we believe that MPW1PW91 is still a better method

PAGE 106

106 when dealing with heavy metals, such as copper, silver and gold. To get better results of 13C lab e led isotopomer frequencies a n improved basis set than either the SDD and LanL2DZ or the application of relativistic methods is apparently needed. T able 6 5. Comparison of Experimental and Calculated (at BPW91/6 311++G(3df)/SDD (peudopotential) and B3LYP/6 311++G(3df)/LanL2DZ level) Isotopomer Frequencies (Integral Intensities) for the Asymmetric C=C Stretch and Symmetric C=C Stretch Fundamental Modes of Near Linear 197A u 12/13C. X Isotopomer exp /cm1 sc a /cm 1 exp sc /cm 1 sc b /cm 1 exp sc /cm 1 BPW91/ 6 311++G(3df)/SDD(peudo) B3LYP/ 6 311++G(3df)/LanL2DZ Asymmetric C=C Stretch Mode a 197 12 12 12 1845.2 1845.2 (369) 0.0 1845.2 (315) 0.0 b 197 12 12 13 1840.9 1835.8 (36 5) 5.1 1836.4 (312) 4.5 c 197 12 13 12 1804.1 1798.2 (353) 5.9 1798.3 (301) 5.8 d 197 13 12 12 1820.9 1830.5 (360) 9.6 1829.9 (307) 9.0 e 197 12 13 13 1800.2 1788.7 (349) 11.5 1789.5 (298) 10.7 f 197 13 12 13 1816.3 1820.4 (357) 4.1 1820.3 (305) 4. 0 g 197 13 13 12 1778.6 1782.9 (344) 4.3 1782.3 (293) 3.7 h 197 13 13 13 1773.8 1772.7 (341) 1.1 1772.6 (290) 1.2 Symmetric C=C Stretch Mode a 197 12 12 12 1275.7 1275.7 (4) 0.0 1275.7 (7) 0.0 b 197 12 12 13 1252.2 (4) 1251.2 (6) c 197 12 13 12 1271.1 1274.7 (4) 3.6 1274.8 (7) 3.7 d 197 13 12 12 1254.9 1249.4 (4) 5.5 1250.2 (7) 4.7 e 197 12 13 13 1250.9 (4) 1249.9 (4) f 197 13 12 13 1226.5 (4) 1226.3 (7) g 197 13 13 12 1251.3 1248.6 (4) 2.7 1249.6 (7) 1.7 h 197 13 13 13 1226.2 1225.6 (4) 0.6 1225.5 (6) 0.7 a Frequencies are scaled uniformly by scaling factor of 0.9596 for asymmetric mode and by 1.0171 for symmetric mode. b Frequencies are scaled uniformly by scaling factor of 0.9463 for asymmetric mode and by 0.9917 for symmetric mode. Summ a ry A n ew gold carbon species w as produced by simultaneous dual laser ablation of gold and carbon followed by trapping in solid argon matrices at 12 K. Two new bands were observed at 1845.2 and 1275.7 cm1. The 13C isotopic labeling experiments sh ow that these two bands have a very similar isotopomer band pattern compared to the near linear AgC3 cluster and CuC3 cluster.

PAGE 107

107 Therefore, these two bands are concluded to be due to the asymmetric and symmetric CC stretching modes in the near linear AuC3 cl uster. The predictions of 13C lab e led isotopomer frequencies for the near linear AuC3 cluster of using density functional theory calculat ions do not match the experiment al data very well. Calculation with MPW1PW91/SDD, MPW1PW91/LanL2DZ functional/basis se ts gave relatively better results, but they are still not with an acceptable range. No photo induced isotopic scrambling in 197Au12/ 13C3 isotopomers was observed, a similar finding with the study of Ag12/13C3, which is consistent with the fact that no stab le cyclic structure of doublet AuC3 could be found theoretically. Such a structure was previously determined to be essential to the photoscrambling found in near linear Cu12/13C3 cluster.

PAGE 108

108 CHAPTER 7 CONCLUSIONS AND FUTU RE WORK T he presence of polycyclic a romatic hydrocarbons (PAHs) and their cations and carbon clusters in the interstellar medium is now widely accepted. Metals, such as iron and copper are also present in the interstellar space. However, very little research has been done on the reaction of metal s with these astrophysical species. In this work, we studied the reaction products of iron with polycyclic aromatic hydrocarbons, and metal s with carbon clusters using matrix isolation spectroscopy along with theoretical calculations. Iron and PAH C omplexes Infrared absorption spectra of neutral complexes of iron with benzene, naphthalene, fluorene, pyrene, and coronene in solid Ar at 12K have been obtained. Supporting calculations of the equilibrium geometries, stabilities, and harmonic vibrational frequencies of these complexes have been carried out using density functional theory (MPW1PW91/631+G(d,p) method) using a modified Perdew Wang exchange and correlation functional/basis set. Our results indicate that the spin multiplicities of the complexe s electronic ground states are triplets. The calculations show that the iron atom is situated over the six-membered carbon ring of the polycyclic aromatic hydrocarbon (PAH) ligand. Calculated dissociation energies (D0) range from 0.52 eV for Fe(coronene) to 2.06 eV for Fe(fluorene). All are substantially less tightly bound than their cationic counterparts. Cationic Fe(PAH) complexes are therefore thought to be more plausible candidates for the carriers of the unidentified interstellar infrared (UIR) emission bands. Metal Carbon Clusters Copper -carbon clusters, silver -carbon clusters and gold-carbon clusters were formed by dual Nd/YAG laser vaporization of metal and carbon target s trapped in solid Ar at 12K and investigated by infrared spectroscopy. Densit y functional (DFT) calculations of a number of

PAGE 109

109 possible molecular structures for Cu, Ag and Au carbon clusters have been performed and their associated vibrational harmonic mode frequencies and dissociation energies were determined. 13C isotopic substitution experiments were performed to assist the identification of the new bands observed in each experiment. Two new bands at 1830.0 and 1250.5 cm1 were observed in the copper -carbon experiments. They were assigned to the asymmetric and symmetric C=C stretchi ng modes, respectively, in the near -linear CuC3 ( X 2A This assignment was supported by the DFT calculation at the MPW1PW91/6 311++G(3df) level and by 13C isotopic substitution experiments. The MPW1PW91/6311++G(3df) level of calculation successf ully predict s the relative integral intensities and the 13C labeled isotopomer frequencies for these modes while the B3LYP/6 311++G(3df) approach failed to do so. In the study of the silver -carbon system, we observed two extra bands at 1827.8 and 1231.6 c m1. Based on the results of isotopic 13C -substitution experiments and density functional theory calculations with the MPW1PW91/SDD functional/basis set, these two bands were assigned to the asymmetric and symmetric CC stretching modes in the near -linear A gC3 cluster. The results of calculation at MPW1PW91/SDD and B3LYP/SDD level were compared to the experimental data. Again, the MPW1PW91/SDD method made better predictions on both relative integral intensities and the 13C labeled isotopomer frequencies for the two modes. Therefore, we believed that MPW1PW91 is a better method when dealing with metal containing systems. After dual ablation and trapping of gold and carbon in the argon matrix, new infrared absorption bands were observed at 1845.2 and 1275.7 cm1. A new gold -carbon species was believed to be responsible for the two new bands. The 13C isotopic labeling experiments showed

PAGE 110

110 that these two bands have similar isotopomer band pattern compared to the near -linear assigned AgC3 cluster and CuC3 cluster. Th erefore, these two bands are very likely due to the asymmetric and symmetric CC stretching modes in the near linear AuC3 cluster. However, the density functional calculation using MPW1PW91, B3LYP and BPW91 functional, along with SDD, LanL2DZ or using diffe rent basis sets on gold and carbon atom, all failed to predict the 13C labeled isotopomer frequencies for the two bands. Hence, the assignment of the two new bands is tentative. Many pure carbon clusters larger than C3 were formed in the metal (Cu, Ag, Au ) -carbon experiments. There are a number of bands left unassigned in the spectra, mainly in the 17001900 cm1 region. These bands are likely due to the products of reactions of larger carbon clusters with metal. Although e fforts were made to produce and i dentify the copper and longer carbon chain clusters, none could be determined unequivocally due to the experimental difficulty of 13C isotopic labeling of larger carbon clusters (i.e., too low intensities for the isotopomeric bands) Photo induced 12/13C isotopic scrambling in Cu12/13C3 clusters has also been observed. The potential energy surface (PES) of the isomerization reaction for the near linear CuC3 cluster was plotted using the MPW1PW91/6311++G(3df) level of theory. The mechanism for the photoscr ambling is shown to involve the formation of a bicyclic CuC3 isomer. However, no photo induced isotopic scrambling was observed in the near linear AgC3 clusters or AuC3 clusters, probably because no stable cyclic structure of AgC3 and AuC3 is possible Suc h a structure was found to be essential to the photoscrambling procedure found in CuC3. Future Work The possible contribution of various PAH+ cations to the unidentified interstellar infrared emission bands (UIRs) has been well established in previous work But comparable broad studies

PAGE 111

111 of Metal (PAH)+ complexes are not yet available. In future work such studies could be extend ed for the most abundant interstellar metals, such as Al, Ni, and Cu To dif ferentiate between the various m etal -PAH cationic complex es deposited in the rare -gas matrix, one could mass -select them to deposition. F igure 7 1 shows such apparatus. Ar / PAH) in Ion Beam Wien Filter+ Wien Filter+ Matrix Gas In 12 K Cryostat Window Decelerator Decelerator Nd YAG Laser Beam Modified RGA 300 Mass Analyzer Ion Extractor and Acceler Mass Selector Ion Source Anode Hot Cathode and / or Metal Target Figure 7 1. Experimental Setup for Ion Mass Selection and Trapping. The PAH will be sublimed into the ion s ource the high repetition Nd/YAG laser will be use d to ablate and ionize a metal which will react with PAHs in the ion source The metal ions will collide with sublimed neutral PAHs forming cationic m etal PAH+ complexes. After extraction in the ion extractor, the Metal PAH+ beam will be passed through the m ass selector decelerated, and then mass analyzed with a modified RGA 300 Mass Analyzer. Once the mass signal of the desired fragment has been optimized, the quad turner wi ll be set to direct the beam in the 90 direction where it will be deposited with the matrix gas on a rotatable cryostat sample

PAGE 112

112 windo w. Although neutral complexes may also form, their trajectory will be different from the ionic complex beam; they will no t be focused by the einzel lenses in the ion extractor and will not be turned by the quad turner. Mass analysis will again be performed and after redirection toward the cryostat window, deposition in a matrix will occur. Ne matrices will be used since thes e matrices perturb electronic transitions less than other matrices. Depending on the ion currents achievable we anticipate that lengthy depositions will probably be necessary to collect enough material for spectral studies. Infrared absorption spectra for ions deposited will be scanned using modified MIDAC FT IR spectrometer in reflect ion mode. Such an experimental apparatus can also be used to study the reactions of metal with the fragments of the PAHs. These kind s of reactions could possibly occur in the interstellar space. The PAHs could potentially be fragmented by cosmic radiation. O ther future work would be to further study the interaction of metal s with longer carbon chain clusters. Argon and nitrogen matrices were used in the present work, and did not produce large enough amount s of metal longer carbon clusters. An overcoating approach could be tried. First, metal and carbon clusters will be formed by dual laser ablation and trapped in an argon matrix as usual. Then krypton gas will be introduced to the chamber and form a top coat on the already -formed matrix. Since krypton has a higher melting point than argon, the matrices can be maintained when annealed to a higher temperature. A lot of longer carbon chain cluster will form during the annealing process, thus increasing the chances of metal atom bonding to the longer carbon chain clusters. Once large signals of 12C metal -carbon clusters are obtained, the 13C labeling isotopic experiments can be performed to assist in the identification of the mo lecules.

PAGE 113

113 LIST OF REFERENCES (1) Ehrenfreund, P.; Charnley, S. B. Annu. Rev. Astron. Astrophy. 2000, 38, 427. (2) Schmidt, T.; Sharp, R. G. Aust. J. Chem. 2005, 58, 69. (3) Gillett, F. C.; Forrest, W. J.; Merrill, K. M. Astrophys. J. 1973, 183, 87. (4) Russell, R. W.; Soifer, B. T.; Willner, S. P. Astrophys. J. 1977, 217, L149. (5) Russell, R. W.; Soifer, B. T.; Willner, S. P. Astrophys. J. 1978, 220, 568. (6) Merrill, P. W. Astrophys. J. 1934, 79, 183. (7) Merrill, P. W. Astrop hys. J. 1936, 83, 126. (8) University of New Hampshire Expreimental Space Plasma Group The Interstellar Medium: an online tutorial ; http://www -ssg.sr.unh.edu/ism/intro.html 08/2008. (9) Ehrenfreu nd, P.; Krafft, C.; Kochan, H.; Pirronello, V. Laboratory Astrophysics and Space Research; Kluwer Academic Publishers, 1999. (10) Ferriere, K. M. Rev. Mod. Phys. 2001, 73, 1031. (11) NASA The Cosmic Ice Laboratry Interstellar Molecules ; http://www 691.gsfc.nasa.gov/cosmic.ice.lab/interstellar.htm 06/2008. (12) Swings, P.; Rosenfeld, L. Astrophys. J. 1937, 86, 483. (13) Kuan, Y. J.; Charnley, S. B.; Huang, H. C.; Tseng, W. L.; Kisiel, Z. Astrophysical Journal 2003, 593 848. (14) Snyder, L. E.; Lovas, F. J.; Hollis, J. M.; Friedel, D. N.; Jewell, P. R.; Remijan, A.; Ilyushin, V. V.; Alekseev, E. A.; Dyubko, S. F. Astrophysical Journal 2005, 619, 914. (15) Wilson, E. Chemical & Engineering News 2005, 83, 44. (16) Tielens, A. G. G. M.; Hony, S.; Van Kerckhoven, C.; Peeters, E. Interstellar and circumstellar PAHs. In European Space Agency, [Special Publication] SP 1999; Vol. SP 427; pp 579. (17) Leger, A.; Puget, J. L. Astron. Astrophys. 1984, 137 L5. (18) Allamandola, L. J.; Tielens, A. G. G. M.; Barker, J. R. Astrophys. J. 1985, 290, L25. (19) Szczepanski, J.; Vala, M. Nature 1993, 363, 699.

PAGE 114

114 (20) Allamandola, L. J.; Hudgins, D. M.; Sandford, S. A. Astrophys. J. 1999, 511 L115. (21) Szczepanski, J.; Fuller, J.; Ekern, S.; Vala, M. Spectrochim. Acta, Part A 2001, 57, 775. (22) Allamandola, L. J.; Hudgins, D. M.; Bauschlicher, C. W.; Langhoff, S. R. Astron. Astrophys. 1999, 352 659. (23) Hinkle, K. W.; Keady, J. J.; Bern ath, P. F. Science 1988, 241, 1319. (24) Bernath, P. F.; Hinkle, K. H.; Keady, J. J. Science 1989, 244 562. (25) Leger, A.; Dhendecourt, L. Astron. Astrophys. 1985, 146, 81. (26) Ruffle, D. P.; Bettens, R. P. A.; Terzieva, R.; Herbst, E. Astrophys. J. 1999, 523, 678. (27) McCarthy, M. C.; Gottlieb, C. A.; Gupta, H.; Thaddeus, P. Astrophys. J. 2006, 652, L141. (28) Brunken, S.; Gupta, H.; Gottlieb, C. A.; McCarthy, M. C.; Thaddeus, P. Astrophys. J. 2007, 664, L43. (29) Remijan, A. J.; Hollis, J. M.; L ovas, F. J.; Cordiner, M. A.; Millar, T. J.; Markwick -Kemper, A. J.; Jewell, P. R. Astrophys. J. 2007, 664, L47. (30) Millar, T. J.; Walsh, C.; Cordiner, M. A.; Ni Chuimin, R.; Herbst, E. Astrophys. J. 2007, 662 L87. (31) Heger, M. L. Lick Observatory b ulletin 1922, 337, 141. (32) Tuairisg, S. O.; Cami, J.; Foing, B. H.; Sonnentrucker, P.; Ehrenfreund, P. Astron. Astrophys. Suppl. Ser. 2000, 142, 225. (33) Herbig, G. H. Annu. Rev. Astron. Astrophy. 1995, 33, 19. (34) Douglas, A. E. Nature 1977, 269, 1 30. (35) Fulara, J.; Lessen, D.; Freivogel, P.; Maier, J. P. Nature 1993, 366, 439. (36) Ehrenfreund, P.; Foing, B. H. Astron. Astrophys. 1996, 307, L25. (37) Foing, B. H.; Ehrenfreund, P. Astron. Astrophys. 1997, 317, L59. (38) Vanderzwet, G. P.; Alla mandola, L. J. Astron. Astrophys. 1985, 146, 76. (39) Le Page, V.; Snow, T. P.; Bierbaum, V. M. Astrophys. J. 2003, 584, 316.

PAGE 115

115 (40) Salama, F.; Bakes, E. L. O.; Allamandola, L. J.; Tielens, A. G. G. M. Astrophys. J. 1996, 458, 621. (41) Brechignac, P.; P ino, T. Astron. Astrophys. 1999, 343, L49. (42) The NASA Imagine Team. In What is your Cosmic Connection to the Elements ; 1997: http://imagine.gsfc.nasa.gov/docs/te achers/elements/imagine/contents.htmal 09/2008. (43) Shull, J. M. Phys. Scr. 1993, T47 165. (44) Van Steenberg, M. E.; Shull, J. M. Astrophys. J. 1988, 330, 942. (45) Klotz, A.; Marty, P.; Boissel, P.; Serra, G.; Chaudret, B.; Daudey, J. P. Astron. A strophys. 1995, 304 520. (46) Marty, P.; deParseval, P.; Klotz, A.; Serra, G.; Boissel, P. Astron. Astrophys. 1996, 316 270. (47) Whittle, E.; Dows, D. A.; Pimentel, G. C. J. Chem. Phys. 1954, 22, 1943. (48) Fowles, G. R. Introduction to Modern Optics ; Dover Publications Inc.: New York, 1975. (49) Snow, T. P.; Rachford, B. L.; Figoski, L. Astrophys. J. 2002, 573, 662. (50) Escalante Ramirez, V. Re: Space matter, energy, anti -matter 2003; http://www.madsci.org/posts/archives/Mar2003/1046619096.As.r.html 11/2008. (51) De Boer, K. S.; Lamers, H. J. G. L. M. Astron. Astrophys. 1978, 69, 327. (52) Savage, B. D.; Bohlin, R. C. Astrophys. J. 1979, 229, 136. (53) Jenkins, E B.; Savage, B. D.; Spitzer, L. Astrophys. J. 1986, 301, 355. (54) Serra, G.; Chaudret, B.; Saillard, Y.; Lebeuze, A.; Rabaa, H.; Ristorcelli, I.; Klotz, A. Astron. Astrophys. 1992, 260, 489. (55) Chaudret, B.; Lebeuze, A.; Rabaa, H.; Saillard, J. Y.; S erra, G. New J. Chem. 1991, 15, 791. (56) Timms, P. L. J. Chem. Soc. D -Chem. Comm. 1969, 1033. (57) Efner, H. F.; Tevault, D. E.; Fox, W. B.; Smardzewski, R. R. J. Organomet. Chem. 1978, 146, 45. (58) Aleksanyan, V. T.; Kurtikyan, T. S. Koord. Khim. 1977 3 1548.

PAGE 116

116 (59) Shobert, A. L.; Hisatsune, I. C.; Skell, P. S. Spectrochim. Acta, Part A 1984, 40, 609. (60) Morand, P. D.; Francis, C. G. Organometallics 1985, 4 1653. (61) Parker, S. F.; Peden, C. H. F. J. Organomet. Chem. 1984, 272, 411. (62) Ball, D. W.; Kafafi, Z. H.; Hauge, R. H.; Margrave, J. L. J. Am. Chem. Soc. 1986, 108, 6621. (63) Boissel, P. Astron. Astrophys. 1994, 285, L33. (64) Marty, P.; deParseval, P.; Klotz, A.; Chaudret, B.; Serra, G.; Boissel, P. Chem. Phys. Lett. 1996, 256, 669. (65) Caraiman, D.; Bohme, D. K. Int. J. Mass spectrom. 2003, 223, 411. (66) Buchanan, J. W.; Reddic, J. E.; Grieves, G. A.; Duncan, M. A. J. Phys. Chem. A 1998, 102 6390. (67) Buchanan, J. W.; Grieves, G. A.; Flynn, N. D.; Duncan, M. A. Int. J. Mass s pectrom. 1999, 187 617. (68) Buchanan, J. W.; Grieves, G. A.; Reddic, J. E.; Duncan, M. A. Int. J. Mass spectrom. 1999, 183 323. (69) Foster, N. R.; Grieves, G. A.; Buchanan, J. W.; Flynn, N. D.; Duncan, M. A. J. Phys. Chem. A 2000, 104, 11055. (70) J aeger, T. D.; Duncan, M. A. J. Phys. Chem. A 2004, 108, 11296. (71) Elustondo, E.; Dalibart, M.; Deroult, J.; Mascetti, J. Phys. Chem. Earth Part C 1999, 24, 583. (72) Szczepanski, J.; Wang, H. Y.; Vala, M.; Tielens, A. G. G. M.; Eyler, J. R.; Oomens, J. Astrophys. J. 2006, 646, 666.

PAGE 117

117 (73) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. J. A.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennuc ci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, 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.; Ochterski, 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 03; Revision B. 05 ed.; Gaussian, Inc.: Wallingford, CT, 2003. (74) Perdew, J. P.; Wa ng, Y. Phys. Rev. B 1992, 45, 13244. (75) Adamo, C.; Barone, V. J. Chem. Phys. 1998, 108, 664. (76) Oomens, J.; Moore, D. T.; von Helden, G.; Meijer, G.; Dunbar, R. C. J. Am. Chem. Soc. 2004, 126, 724. (77) Values taken from the NIST atomic spectra data base. http://physics.nist.gov/PhysRefData/ASD/index.html (78) Brown, K. G.; Person, W. B. Spectrochim. Acta, Part A 1978, 34, 117. (79) Robbins, D. L.; Rittby, C. M. L.; Graham, W. R. M J. Chem. Phys. 2001, 114, 3570. (80) Kinzer, R. E.; Rittby, C. M. L.; Graham, W. R. M. J. Chem. Phys. 2006, 125 (81) Bates, S. A.; Rhodes, J. A.; Rittby, C. M. L.; Graham, W. R. M. J. Chem. Phys. 2007, 127 (82) Kinzer, R. E.; Rittby, C. M. L.; Graha m, W. R. M. J. Chem. Phys. 2008, 128 (83) Bates, S. A.; Rittby, C. M. L.; Graham, W. R. M. J. Chem. Phys. 2006, 125. (84) Datta, B. P.; Sant, V. L.; Raman, V. A.; Subbanna, C. S.; Jain, H. C. Int. J. Mass Spectrom. Ion Processes 1989, 91, 241. (85) Guo B. C.; Kerns, K. P.; Castleman, A. W. Science 1992, 255 1411. (86) Kosolapova, T. Y. Carbide ; Plenum Press: New York, 1971.

PAGE 118

118 (87) Katskov, D. A.; Kruglikova, L. P.; L'Vov, B. V.; Polzik, L. K. Zhurnal Prikladnoi Spektroskopii;J. Appl. Spectry. [USSR] 1 976, 25, 1459. (88) Yamada, Y.; Castleman, A. W. Chem. Phys. Lett. 1993, 204, 133. (89) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (90) Wiberg, K. B. J. Comput. Chem. 1999, 20, 1299. (91) Dunbar, R. C. J. Phys. Chem. A 2002, 106, 7328. (92) Wang, Y.; Szczepanski, J.; Vala, M. Chem. Phys. 2007, 342, 107. (93) Legge, F. S.; Nyberg, G. L.; Peel, J. B. J. Phys. Chem. A 2001, 105, 7905. (94) Wang, X. F.; Andrews, L. Angew. Chem. Int. Ed. 2003, 42, 5201. (95) Weltner, W., Jr.; Van Zee, R. J. Chem. Rev. 1 989, 89, 1713. (96) Van Orden, A.; Saykally, R. J. Chem. Rev. 1998, 98, 2313. (97) Szczepanski, J.; Hodyss, R.; Vala, M. J. Phys. Chem. A 1998, 102, 8300. (98) Szczepanski, J.; Auerbach, E.; Vala, M. J. Phys. Chem. A 1997, 101, 9296. (99) Gutsev, G. L. ; Andrews, L.; Bauschlicher, C. W. Chem. Phys. 2003, 290, 47. (100) Jacox, M. E.; Milligan, D. E. Chem. Phys. 1974, 4 45. (101) Jiang, Q.; Rittby, C. M. L.; Graham, W. R. M. J. Chem. Phys. 1993, 99, 3194. (102) Andrews, L.; Kushto, G. P.; Zhou, M. F.; Willson, S. P.; Souter, P. F. J. Chem. Phys. 1999, 110, 4457. (103) Ding, X. D.; Wang, S. L.; Rittby, C. M. L.; Graham, W. R. M. J. Chem. Phys. 2000, 112 5113. (104) Szczepanski, J.; Ekern, S.; Vala, M. J. Phys. Chem. 1995, 99, 8002. (105) Grotjahn, D. B.; Brewster, M. A.; Ziurys, L. M. J. Am. Chem. Soc. 2002, 124, 5895. (106) Szczepanski, J.; Vala, M. Eur. Phys. J. Special Topics 2007, 144, 27. (107) Fueno, H.; Taniguchi, Y. Chem. Phys. Lett. 1999, 312, 65. (108) Herbst, E. Space Sci. Rev. 2002, 9 6 1. (109) Lapinski, L.; Vala, M. Chem. Phys. Lett. 1999, 300, 195.

PAGE 119

119 (110) Liu, P.; Rodriguez, J. A.; Muckerman, J. T. J. Phys. Chem. B 2004, 108, 15662. (111) Liu, P.; Rodriguez, J. A.; Muckerman, J. T. J. Phys. Chem. B 2004, 108, 18796. (112) Cartier S. F.; May, B. D.; Castleman, A. W. J. Chem. Phys. 1994, 100, 5384. (113) Cartier, S. F.; May, B. D.; Castleman, A. W. J. Am. Chem. Soc. 1994, 116, 5295. (114) Leskiw, B. D.; Castleman, A. W. Comptes Rendus Physique 2002, 3 251. (115) Pilgrim, J. S.; Duncan, M. A. J. Am. Chem. Soc. 1993, 115, 9724. (116) Tono, K.; Terasaki, A.; Ohta, T.; Kondow, T. Chem. Phys. Lett. 2002, 351, 135. (117) Knappenberger, K. L.; Jones, C. E.; Sobhy, M. A.; Iordanov, I.; Sofo, J.; Castleman, A. W. J. Phys. Chem. A 2006, 110, 12814. (118) Knappenberger, K. L.; Clayborne, P. A.; Reveles, J. U.; Sobhy, M. A.; Jones, C. E.; Gupta, U. U.; Khanna, S. N.; Iordanov, I.; Sofo, J.; Castleman, A. W. Acs Nano 2007, 1 319. (119) Wang, L. S.; Li, X. J. Chem. Phys. 2000, 112, 3602. (120) Zhai, H. J.; Liu, S. R.; Li, X.; Wang, L. S. J. Chem. Phys. 2001, 115, 5170. (121) Fan, J. W.; Lou, L.; Wang, L. S. J. Chem. Phys. 1995, 102, 2701. (122) Wang, X. B.; Ding, C. F.; Wang, L. S. J. Phys. Chem. A 1997, 101, 7699. (123) Kohno, M.; Suz uki, S.; Shiromaru, H.; Kobayashi, K.; Nagase, S.; Achiba, Y.; Kietzmann, H.; Kessler, B.; Gantefoer, G.; Eberhardt, W. J. Electron. Spectrosc. Relat. Phenom. 2000, 112, 163. (124) Bates, S. A.; Rittby, C. M. L.; Graham, W. R. M. J. Chem. Phys. 2008, 128, 234301. (125) Cataldo, F. J. Raman Spectrosc. 2008, 39, 169. (126) Sakurai, H.; Castleman, A. W. J. Phys. Chem. A 1998, 102, 10486. (127) Wang, L. S.; Li, S.; Wu, H. B. J. Phys. Chem. 1996, 100, 19211. (128) Szczepanski, J.; Wang, Y.; Vala, M. J. Phys Chem. A 2008, 112, 4778. (129) Vogel, A. I. Practical Organic Chemistry 3rd ed.; Longmans: London, 1967. (130) Bahr, G.; Burba, P. In Methoden der Organishen Chemie ; Houben Weyl Thieme Verlag: Stuttgart, Germany, 1970; Vol. 13, Part I; pp 767.

PAGE 120

120 (131) Andrae, D.; Haussermann, U.; Dolg, M.; Stoll, H.; Preuss, H. Theor. Chim. Acta 1990, 77, 123. (132) Chan, W. T.; Fournier, R. Chem. Phys. Lett. 1999, 315, 257. (133) Hamilton, I. P. Chem. Phys. Lett. 2004, 390, 517. (134) Kranze, R. H.; Graham, W. R. M. J. Chem. Phys. 1993, 98, 71. (135) Liang, B. Y.; Andrews, L. J. Phys. Chem. A 2000, 104, 9156. (136) Jiang, L.; Xu, Q. J. Phys. Chem. A 2005, 109, 1026. (137) Zhai, H. J.; Kiran, B.; Dai, B.; Li, J.; Wang, L. S. J. Am. Chem. Soc. 2005, 127 12098. (13 8) Fielicke, A.; von Helden, G.; Meijer, G.; Pedersen, D. B.; Simard, B.; Rayner, D. M. J. Am. Chem. Soc. 2005, 127, 8416. (139) Fielicke, A.; von Helden, G.; Meijer, G.; Simard, B.; Rayner, D. M. J. Phys. Chem. B 2005, 109, 23935. (140) Velasquez, J.; N jegic, B.; Gordon, M. S.; Duncan, M. A. J. Phys. Chem. A 2008, 112, 1907. (141) Kong, Q.; Zeng, A.; Chen, M.; Xu, Q.; Zhou, M. J. Phys. Chem. A 2004, 108, 1531. (142) Wang, X.; Andrews, L. J. Phys. Chem. A 2002, 106, 3744. (143) Schroder, D.; Brown, R.; Schwerdtfeger, P.; Wang, X. B.; Yang, X.; Wang, L. S.; Schwarz, H. Angew. Chem. Int. Ed. 2003, 42, 311. (144) Jiang, L.; Kohyama, M.; Haruta, M.; Xu, Q. J. Phys. Chem. A 2008, 112, 13495. (145) Barysz, M.; Pyykko, P. Chem. Phys. Lett. 1998, 285, 398. (146) Pyykko, P.; Patzschke, M.; Suurpere, J. Chem. Phys. Lett. 2003, 381, 45. (147) Puzzarini, C.; Peterson, K. A. Chem. Phys. 2005, 311, 177. (148) Ticknor, B. W.; Bandyopadhyay, B.; Duncan, M. A. J. Phys. Chem. A 2008, 112, 12355. (149) Wang, Y.; Szcz epanski, J.; Vala, M. J. Phys. Chem. A 2008, 112, 11088. (150) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270. (151) Wadt, W. R.; Hay, P. J. J. Chem. Phys. 1985, 82, 284.

PAGE 121

121 (152) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299.

PAGE 122

122 BIOGRAPHICAL SKETCH Yun Wang was born in Shanghai, China in July 1981. She grew up in Shanghai and graduated from Shixi High School in June 1999. Yun continued her education at Fudan University, majoring in chemistry, and received her Bachelor of Science degree in Ju ne 2003. In August 2004, Yun enrolled in the Department of Chemistry at the University of Florida to pursue a Doctor of Philosophy degree in physical chemistry. From then on, she worked in Dr. Martin Valas research group, and dedicated herself to vibratio nal spectroscopy of astrophysical species.