Matrix isolation and AB initio theoretical investigations of carbon-based molecules of astrophysical interest / by Scott...

MISSING IMAGE

Material Information

Title:
Matrix isolation and AB initio theoretical investigations of carbon-based molecules of astrophysical interest / by Scott P. Ekern
Physical Description:
xv, 165 leaves : ill. ; 29 cm.
Language:
English
Creator:
Ekern, Scott P, 1958-
Publication Date:

Subjects

Subjects / Keywords:
Chemistry thesis, Ph. D   ( lcsh )
Dissertations, Academic -- Chemistry -- UF   ( lcsh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1996.
Bibliography:
Includes bibliographical references (leaves 154-164).
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 026429931
oclc - 36634589
System ID:
AA00017667:00001


This item is only available as the following downloads:


Full Text










MATRIX ISOLATION AND AB INITIO THEORETICAL INVESTIGATIONS OF
CARBON-BASED MOLECULES OF ASTROPHYSICAL INTEREST












By

SCOTT P. EKERN


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


1996


















TABLE OF CONTENTS

page


LIST OF TABLES ....................................... ........... v


LIST OF FIGURES ....................................... ........ viii


AB STR A C T ......................................... ........... xiv


CHAPTERS


1 INTRODUCTION ..............................


B background ................................
E xperim ental ..................................
O verview .....................................


2 SPECTROSCOPY AND PHOTOCHEMISTRY OF THE
C3.H20 COMPLEX ............. ..........


. ... ......... 1


.............. 12


Introduction .....................
Theoretical ...........
Results ...........
The C3.H20 Complex .........
The trans-HPD Isomer .........
The cis-HPD Isomer ..........
HPD Isomer Interconversion ..
Discussion ....................
Sum m ary .......................


3 THE C3H2O POTENTIAL SURFACE


Introduction .....................
Theoretical ..........
Results ............
D discussion ......................
Sum m ary .......................


............
............
............
............
............
............
............
............
............
















4 A PHOTOCHEMICAL MECHANISM FOR C30 FORMATION ........ 81

Introduction ...................................... .......... 81
Theoretical ....................................... ........ 83
Results ....................................... .......... 84
The C3H20 Potential Surface .............................. 89
Bimolecular Reactions Forming C30 ........................ 97
Sum m ary ........................................ ........99

5 EXPERIMENTAL OBSERVATION OF LONG CHAIN
CARBON ANIONS ...................................... 101

Introduction ....................................... ...... 101
Experim ental ............................. ..... .. ..... 102
R results . . . . . .. 104
E electronic ................................... .... ... 104
V ibrational .................................... ........ 118
Summary ................ ........... ......... ........ 120

6 THEORETICAL STUDIES OF CARBON ANIONS ................ 122

Introduction ................. ....... ............ ..... .. 122
Theoretical ............... ................ .... .... 123
Results and Discussion ........................................ 124
Linear Anions ..................................... ..... 124
Cyclic A onions ........................................... 133
Summary ......... .................. .......... 138

7 THEORETICAL STUDIES OF C3 PERTURBED BY AN
ARGON MATRIX ..................................... 139

Introduction ..................................... .......... 139
T theoretical ........................................ ...... 140
Results ................................ ........... 140
D discussion ......................................... ........ 147

8 FUTURE WORK ........................................ 148
















LIST OF REFERENCES .................................. ........ 154

BIOGRAPHICAL SKETCH ................................ .. ..... 165













LIST OF TABLES


Iaie page

2-1. Experimental frequencies and vibrational assignments for small carbon
clusters in their ground states .............. ................ 16

2-2. Comparison of experimental (Ar matrix) and calculated (HF/6-31G*)
IR frequencies and intensities for the C3,H20 complex. Frequencies
in cm" and intensities in km/mol. .............. ................ 20

2-3. The experimental (Ar matrix) and calculated (HF/6-31G*) IR
frequencies for the CCC asymmetric stretching mode of 1213C3,H20
complex. Frequencies in cm'1 ..................................... 22

2-4. The experimental (Ar matrix) and calculated IR frequencies (cm"-) and
intensities (km/mol) for trans-3-hydroxypropadienylidene (trans-HPD)
in its ground electronic state. ......... .. ...... .. ....... ...... 31

2-5. Experimental and calculated (HF/6-31G*) IR frequencies (cm"n) for
all isotopomers of cis and trans-HPD for the most intense CCC
asymmetric stretch, CCH+HCO bend (trans-HPD) and HCO bend
+ CO stretch (cis-HPD) modes. .................. ............... 35

2-6. The experimental (Ar matrix) and calculated IR frequencies (cm'1)
and intensities (km/mol) for cis-3-hydroxypropadienylidene (cis-HPD)
in its ground electronic state. ................. .................. 40

3-1. Total energies (hartrees), zero-point corrected relative energies
(kcal/mol), and zero-point energies (kcal/mol) of all minima
calculated at the HF/6-31G* level. Total energy (hartrees)
calculated at the QCISD(T)/6-31G*//HF/6-31G* level. HF/6-31G*
calculated zero-point corrections scaled by 0.90 ...... ............... 59








3-2. Total energies (hartrees), zero-point corrected relative energies (kcal/mol),
and zero-point energies (kcal/mol) of each transition state at the HF/6-3 G*
level. Total energy (hartrees) calculated at the QCISD(T)/6-31G*//
HF/6-31G* level. HF/6-31G* calculated zero-point corrections
scaled by 0.90. ....... ............... .... ...... ......... 60

3-3. Total energies (hartrees), zero-point corrected relative energies (kcallmol),
and zero-point energies (kcal/mol) of minima calculated at the
MP2/6-31G* level. MP2/6-31G* calculated zero-point corrections
scaled by 0.95. ........ ......................... .......... 61

3-4. Total energies (hartrees), zero-point corrected relative energies (kcal/mol),
and zero-point energies (kcal/mol) for all transition states found at the
MP2/6-31G* level. MP2/6-31G* calculated zero-point corrections
scaled by 0.95. ..................................... .......... 62

3-5. Rotational constants (GHz) of minima found at the MP2/6-31G* level.
Experimental values in parenthesis. ................. .............. 78

3-6. Vibrational frequencies (cm''), intensities (km/mol), and symmetries
of each C3H20 isomer found at the MP2/6-31G* level .................. 79

4-1. Total energies (hartrees), zero-point corrected relative energies (kcal/mol),
spin squared eigenvalues, and zero-point energies (kcal/mol) of minima
found at the MP2/6-31 G* level of theory ................... .. ...... 85

4-2. Total energies (hartrees), zero-point corrected relative energies (kcal/mol),
spin squared eigenvalues, and zero-point energies (kcal/mol) of
transition states found at the MP2/6-31G* level of theory. .............. 85

5-1. Experimental electronic absorption band positions (X in nm, v in cm"1)
and absorbances for anionic carbon clusters isolated in neon and argon
matricies. Band widths (FWHM, in cm') in parenthesis. All 0-0
transition energies are plotted in Figure 5-6 ......... ......... 110

6-1. Linear carbon anion total energies (hartrees), zero-point energies
(kcal/mol), rotational constants (GHz), spin squared values, in their
ground electronic states calculated at the B3LYP/6-31G* level of theory. ... 126

6-2. Vibrational frequencies (cm'l), IR intensities (km/mol), and symmetries
for linear carbon anions Cn" (n = 2-9) calculated at the B3LYP/6-31G*
level of theory ............................... ......... 127








6-3. A comparison of experimental and select calculated linear carbon
anion bands. Frequencies in cm-1, intensities in km/mol ................ 128

6-4. Energies of reaction (kcal/mol) between neutral linear carbon chains
(nd 9 atoms) calculated at the B3LYP/6-31G* level of theory.
Not zero-point corrected. ...................................... 130

6-5. Energies of reaction (kcal/mol) between linear carbon neutrals
abscissaa) and linear carbon anions ordinatee) yielding a longer linear
carbon anion. Calculated at the B3LYP/6-31G* level of theory.
Electronic states in parenthesis. ................. ................ 131

6-6. Symmetries, total energies (hartrees), spin squared values, zero-point
energies (kcal/mol), and the number of imaginary frequencies (iv)
found for each of the cyclic carbon anion species ..................... 135

6-7. Vibrational frequencies of the cyclic carbon anions calculated at the
B3LYP/6-31G* level. Frequencies are in cmn' and intensities are in
km/mol. ....... .............................. ..... 136

7-1. Total energies (hartrees), relative energies (kcal/mol), geometric
parameters (in angstroms and degrees), and Mulliken charges for the
central carbon--argon orientation as a function ofC3-Ar distance ......... 141

7-2. Total energies (hartrees), relative energies (kcal/mol), goemetric
parameters (in angstroms and degrees), and Mulliken charges for the
bridging carbon--argon orientation as a function ofC3-Ar distance ........ 142

7-3. Total energies (hartrees), relative energies (kcal/mol), goemetric
parameters (in angstroms and degrees), and Mulliken charges for the
terminal carbon--argon orientation as a function ofC3-Ar distance ........ 143













LIST OF FIGURES


Figure pag

1-1. Experimental set-up for matrix isolation of carbon anions produced
by electron impact ionization ............... ................... 5

1-2. Expander module and matrix window ............................. .. 6

1-3. Experimental set-up for matrix isolation of carbon anions produced
by dual laser vaporization of graphite and yttrium metal .................. 7

1-4. The experimental reaction scheme relevant to Chapters 2, 3, and 4 ......... 11

2-1. Structures of the intermediates hydroxyethynlcarbene (HEC, 1), cis-3-
Hydroxypropadienylidene (cis-HPD, 2), trans-3-hydroxypropadienylidene
(trans-HPD, 3), the C3oH20 complex (4), and the transition state
between the complex and the HPD isomers (5). .......... .. ..... 15

2-2. Portions of the 10K matrix infrared spectrum of 12C/Ar + H20
(0.1%) in the asymmetric stretch region (ca. 2000 cm'") and in the
OH asymmetric (v3) and symmetric (vl) region (ca. 3700 cm'1).
Monomer (M) and dimer (D) bands are marked; C3.H20 complex
bands are starred. Inset shows the v2 + V3 combination band of HO ....... 18

2-3. Portion of the annealed (32K) matrix infrared spectra of 12C/3C/Ar
([]:[13C] = 1:1) (top) and of 2C (12K) (bottom) in the asymmetric
CC stretching region. The C3 isotopomers are marked by triangles, while
the C3.H20 complex isotopomers are marked with dots. The marked
bands in the lower spectrum are due to C9 (1998 cm'), C3 (2039 cm'),
and the C3.H20 complex (2052 cm'1). ........................ 21

2-4. Portion of the 12K matrix infrared spectra of 2Cn/Ar + H20 (0.2%)
in the OH stretch region before photolysis (b) and after photolysis (a).
The monomer (M), dimer (D), and trimer (T) bands of water are marked
accordingly. The starred bands are the OH stretches in the C3.HzO
complex ................. .............................. .. 24









2-5. Portion of the 12K matrix infrared spectra of 2Cn/Ar + HO (0.2%)
before photolysis and after photolysis (X > 390 nm). Bands which have
grown upon photolysis are due to C30, propynal, and trans-HPD.
Note that the C3H20O complex band has decreased upon photolysis.
C3*(H20)n (n = 2 and 3) bands are marked by stars. ................ 25

2-6. Portion of the 12K matrix infrared spectra of '2Cn/Ar + H20 (0.2%)
before photolysis (b) and after photolysis (a) (X > 390 nm). All bands
in the difference spectrum are due to trans-HPD. ................... ... 26

2-7. Electronic spectra for 12C3 and 12C3H20O complex before photolysis
(upper) and after (lower). Band at 405.4 nm is assigned to the
12C3*H20 complex. .............. ................ ......... 28

2-8. Comparison of the experimental (upper) and theoretical (lower) spectra
of trans-HPD. Bands with one star, two stars, and dots are due to the
C3*H20 complex, propynal, and trans-HPD, respectively ................ 32

2-9. Portion of the 12K matrix infrared spectra of 2C/13C/Ar + H20 (0.15%)
([12C]:[13C] = 1:2) before photolysis and after 30 min of photolysis
(X > 390 nm). The negative peaks in the difference spectrum (triangles) are
the asymmetric CCC stretches of the '12'C3H20O complex isotopomers.
The positive peaks (dots) are the asymmetric CCC stretches of the 1'13C3
sustituted trans-HPD isotopomers. ................. ......... ....... 34

2-10. Portion of the 12K matrix infrared spectrum of 12C/13C/Ar + H2O
(0.15%) (['1C]:[13C] = 1:2) after 30 min photolysis (X > 390 nm).
Isotopomeric bands are due to trans-HPD. .. .... ................. 36

2-11. Portion of the 12K matrix infrared spectra of 12Cn/Ar + H20 (0.2%)
before and after 30 min of photolysis (. > 390 nm), but recorded during
photolysis. The 3245.2 and 3258.9 cm7' bands are the v, + V3
combination bands in C3 and the C3*H20 complex. ................... ... 38

2-12. Portion of the 12K matrix infrared spectra of 2C/13C/Ar + H20 (0.15%)
([ 2C]:['3C] = 1:2) before photolysis and after 30 min photolysis
(2 > 390 nm), but recorded during photolysis. The 1254.3 cm' band is
the COH bend + CO stretch of cis-HPD. ........................... 39








2-13. Portion of the 12K matrix infrared spectra of 12C/13C/Ar + H20 (0.15%)
([12C]:[lC] = 1:2) after 30 min of photolysis (% > 390 nm) and recorded
with the photolysis lamp on (bottom) and off(middle). In the difference
spectra (top) are shown the CCC asymmetric stretching modes of the
Isotopomers of cis-HPD (positive peaks) and trans-HPD (negative peaks). ... 41

2-14. Time dependence of the difference spectrum (At-A). The t = 0 and
and subsequent plots were scanned after the photolysis lamp was turned
off Each spectrum was recorded for ca. 10 sec, with A. scanned after
2 hrs. The 1992.5 cm-' (trans-HPD) band is at a maximum absorbance
at t = 2 hr, at which time the 1999.6 cm'1 (cis-HPD band is at its
maximum absorbance ........................................... 43

2-15. Dependence of lnIAt-A. vs. time after 30 min of photolysis (X > 390 nm)
of the 12C/Ar + H20 (0.15%) matrix for T = 11 and 20K. The A, and
A, absorbances are identical to the ones shown in Figure 2-14 ............. 45

2-16. Potential energy surface for the rotation of the H atom about the C-O
bond of HPD in its ground electronic state. See text for AE and AG
values. ................... ................... .............. 48

3-1. Unoptimized structures of C3H20 isomers studied. .................. .. 58

3-2. The HF/6-31G* calculated potential surface. Only lowest energy
isomers are shown. Zero-point corrected energies (kcal/mol)
relative to A in parenthesis ... ........ .. ............ ...... 63

3-3. The QCISD(T)/6-31 G*//HF/6-31 G* calculated potential surface.
Only lowest energy isomers are shown. Zero-point corrected energies
(kcal/mol) relative to A in parenthesis. ............... ......... 64

3-4. The MP2/6-31G* calculated potential surface. Only lowest energy
isomers are shown. Zero-point corrected energies (kcal/mol)
relative to A in parenthesis ................... ................... 65

3-5. Optimized structures of all minima found at the MP2/6-3 1G* level
oftheory ......... .......................................... 66

3-6. Optimized structures of all transition structures found at the
MP2/6-31G* level of theory .................................. 68

3-7. Portion of the 12K matrix infrared spectrum of '2C/Ar before and
after photolysis (X > 390 nm) with a water filter present ................. 74









3-8. Portion of the 12K matrix infrared spectrum of 12C/Ar before and
after photolysis (X > 390 nm) without a water filter ............... ... 75

3-9. The experimental reaction scheme. The large arrow shows the reaction
path described by the potential surface in Chapter 3. .................. 77

4-1. The experimental reaction scheme. The large arrow shows the
reaction path represented by the potential surface in Chapter 4 ............ 82

4-2. Optimized structures of all minima found at the MP2/6-31 G* level
oftheory .......... .... ........................ ............ 87

4-3. Optimized structures of all transition states found at the MP2/6-31 G*
level of theory ................... ............. .......... 88

4-4. Potential surface at the MP2/6-31G* level. Numbers are zero-point
corrected energies (kcal/mol) relative to the C3H20O complex ............. 90

4-5. Graph representing the singlet (squares) and triplet (circles) surfaces
for H atom dissociation from cis-HPD (C). Note that the triplet
surface approaches the molecule F + H atom total energies at infinite
separation (68.1 kcal/mol). Energies are relative to the fully optimized
cis-HPD molecule. ................. ................... .. .... 93

4-6. Graph representing the singlet (squares) and triplet (circles) surfaces
for H atom dissociation from propynal (E). Note that the triplet
surface approaches the molecule F + H atom total energies at infinite
separation (130.4 kcal/mol). Energies are relative to the fully optimized
propynal molecule. ............ .. .......................... 94

4-7. Graph representing the singlet (squares) and triplet (circles) surfaces
for H atom dissociation from trans-HPD (B). Note that the triplet
surface approaches the molecule G + H atom total energies at infinite
separation (98.0 kcal/mol). Energies are relative to the fully optimized
trans-HPD molecule. ................ ................... .... 96

5-1. Part of the vibrational (left) and electronic (right) spectra for the
Cn/Cn clusters in an Ar matrix at 12K, recorded before (upper) and
after photolysis (lower). Photolysis was for 0.5 hr using a 265 nm
cut-off filter. .......... ............................... 105








5-2. Part of the electronic spectra ofCn species vaporized with two
different laser beam densities at the vaporization point. In (a) a low
flux of photons is used and no plasma is observed. In (b) a high flux
of photons is used and a purple plasma is seen at the graphite surface.
Note the difference in C2 at 238.8 nm. ............................ 106

5-3. Part of the vibrational spectra of C species vaporized with two
different laser beam densities at the vaporization point. In (a) a low
flux of photons is used and no plasma is observed. In (b) a high flux
of photons is used and a purple plasma is seen at the graphite surface ..... 107

5-4. Spectrum showing absorption bands due to the electronic transitions
of the C16 and C1,, clusters isolated in an Ar matrix ................... 108


5-5. Spectrum showing absorption bands due to electronic transitions of
the C34 and C36 clusters isolated in an Ar matrix. Other known
bands are marked accordingly ................... ............. 109

5-6. Graph of 0-0 electronic transition energy as a function of the number
of carbon atoms in the anionic cluster ......................... 113

5-7. The infrared emission spectrum (UIR) recorded from planetary nebula
NGC 7027 and HD4179. The positions of the 0-0 electronic transitions
for chain carbon cluster anions Cn" (18 < n 36) from this work are
indicated. UIR emission spectra taken from reference 17 ............... 116

5-8. Portion of the infrared spectrum of Cn/Cn" species trapped in an Ar
matrix at 12K ........ ................................ 119

6-1. Linear carbon anion optimized geometries. Bond lengths in angstroms.
B3LYP/6-31G* lengths in bold type, ROHF/6-31G* lengths from
reference 156 in normal type. ............ ................. .. 125

6-2. Plot of the exothermicity of the reaction between neutral carbon cluster
chains ordinatee) and anionic carbon chains (in graph). Note the
larger exothermicity for reactions with even numbered neutral chains. ...... 132

6-3. Optimized structures of all stable cyclic anion structures found at the
B3LYP/6-31G* level of theory. ............................... 134

7-1. Central, bridging, and terminal geometries .......................... 140








7-2. Potential energy for the bending of C3 in the terminal orientation as a
function of C-Ar distance. Energies are relative to the linear (CCC
bond angle = 00). ................. ................... ..... ... 145

7-3. Interaction energies between C3 and Ar in the central, bridging, and
terminal orientations. Energies are relative to the completely optimized
C3*Ar complex. .................................. .. ...... 146

8-1. Infrared emission spectra of several unsubstituted PAH molecules.
Black bands at the bottom represent wavelengths where UIRs are
observed. Spectra were recorded using the single-photon infrared
emission spectroscopy (SPIRES) technique. Figure from reference 173. ... 151














Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy


MATRIX ISOLATION AND AB INITIO THEORETICAL INVESTIGATIONS OF
CARBON-BASED MOLECULES OF ASTROPHYSICAL INTEREST

By

Scott P. Ekem

December 1996

Chairperson: Martin T. Vala
Major Department: Chemistry

Carbon clusters and their products have been of interest to researchers for some

time. This interest has arisen because of the involvement of carbon clusters in such

diverse processes and products as soot, fullerenes, flames, nanotubes, and interstellar

species. In this dissertation, two major avenues of research into carbon clusters are

investigated. In the first, the complex formed between the C3 cluster and water is probed

experimentally and theoretically. Photolysis of the complex is shown to produce a stable

intermediate and the products propynal and tricarbon monoxide. Further photolysis is of

propynal yields acetylene and carbon monoxide. The structures of the intermediate and

products are proven by isotopic substitution. The mechanisms for the most probable

reaction pathways are investigated by ab initio calculations. The implication of these

results for the formation of these species in small interstellar dust grains is explored.








The second major avenue of research investigated concerns carbon cluster anions.

Relatively little work has been reported on these species compared to the neutral species.

New methods for their production and stabilization in rare gas matrices are introduced.

Density functional theory (DFT) has shown that the linear isomers are much more stable

than their cyclic counterparts. DFT also been used to show that the reaction between

neutral and anionic clusters is strongly exothermic. This is consistent with the

experimental results which provide evidence for even-numbered carbon cluster anions

from C4 through C36. The vibrational spectra of a number of these anions is also assigned.

The electronic spectra for the Cn series spans the spectral range from the visible through

the mid-infrared. It is shown that there is a reasonable match for the longer chain anions

with several of the unidentified interstellar infrared emission bands.














CHAPTER 1
INTRODUCTION



Background



Matrix isolation spectroscopy was introduced originally by Whittle, Dows, and

Pimentel in 1954.1 Since then numerous books and articles have been published on the

subject.2' This technique involves the codeposition of a host matrix gas and the vapor of

a guest molecule, which is the molecule of interest. The temperature of most matrices fall

within the 4 to 20K range, with the final temperature being dependent on the refrigeration

system and the melting temperature of the matrix host gas being employed. Typically, the

temperature is kept several degrees below the melting temperature of the host material to

minimize diffusion of guest species through the matrix. A matrix gas is chosen primarily

with two criteria in mind. First, the matrix material should be transparent in the

spectroscopic region of interest. Second, it should be as weakly interacting as possible

with other host atoms or molecules and the guest material. Even though N2 and 02 are

sometimes employed, the most widely chosen matrix materials are the noble gases: neon,

argon, krypton, and xenon. They are transparent from the vacuum ultraviolet region to

the far infrared. But most importantly, perturbation of the guest species is smallest when a

rare gas is used. In order to minimize the possibility of aggregates of guest molecules,










which could undergo chemical reactions (photolytically induced or not) or perturb the

spectrum, a large host-to-guest ratio is deposited. These ratios generally fall in the range

102 to 105:1, with most being 103:1.

Matrix isolation spectroscopy has the advantage of being able to trap ions, radicals,

weakly bound complexes, atoms, and unusual molecules for long periods of time. For

better or worse the trapped species are in their ground electronic and vibrational states.

And, like at higher temperatures, fluorescence and phosphorescence are seen only from

the lowest lying excited state. Careful control of temperature can influence the rate of

diffusion, and this can be used to change the concentrations and/or identities of many

species. There is even indication that molecules can be created in the matrix that could

not be made in the gas phase because the matrix itself prevents parts of molecules from

irreversibly separating after dissociation.4 While there are many points that make matrix

isolation spectroscopy a useful tool for interrogating charged or transient species,

operating in the condensed phase has several unfortunate consequences. Most manifest

themselves spectroscopically as some combination of line broadening, band splitting, line

shape changes, or gas-to-matrix frequency shifts. All of these phenomena arise from

guest/host interactions, and are, therefore, intrinsic to the technique. Since more attention

is paid to the frequency shifts, priority is given to minimizing this effect. Shifts are

approximately 1%, but many times this is enough to prevent positive band assignment.

Argon is the most commonly used matrix material; however, neon is preferred as it

exhibits the smallest shifts from gas phase values.










Matrix isolation spectroscopy is applied here to try to shed light on one of two

longstanding and (possibly) related astrophysical problems. The first is concerned with a

series of over 100 absorption bands spread throughout the visible and near infrared

regions commonly known as the diffuse interstellar bands (DIBs).5 The origins of these

bands, originally discovered in 1934,6 have been the subject of wide spread debate. Smith,

Snow, and York7 suggested some time ago that the carriers of the DIBs could be

molecular in origin rather than solid grains. Since then linear carbon chains,8 free carbene

molecules,9 polycyclic aromatic hydrocarbons (PAHs),'1 and the C60, ion" have been

proposed as molecular species contributing to the DIBs. These species were suggested

because they and/or their ions have absorption bands in the visible and ultraviolet regions

spanned by the DIBs. Recently a new hypothesis was forwarded which could conceivably

account for all of the DIBs.12 This model involves the nonlinear absorption of Lyman-

alpha and optical photons near a star by the H2 molecule. However, experiments have

challenged this hypothesis by only being able to reproduce a few of the observed DIBs.3

The other part of the astrophysical puzzle deals with the unidentified infrared bands

(UIRs). Discovered in 1973,14'15 the interstellar infrared emission bands are a series of

bands observed at 3.3, 6.2, 7.7, 8.8, and 11.3 gm. It has been proposed that the UIR

bands originate from IR emission of neutral and/or ionic polycyclic aromatic hydrocarbons

(PAHs).16'17 The reasoning behind this is simple. The 3.3 and 6.2 gm bands correspond

closely to an aromatic C-H stretch and C=C stretch in aromatic rings, respectively. The

7.7 pm band is also assigned to C-C stretch, however it is not seen in smaller aromatic

compounds and could be representative of larger species.'l The 8.8 and 11.3 pm bands










correspond to an in-plane aromatic C-H bend and an out-of-plane aromatic C-H bend,

respectively. The position of the C-H out-of-plane bend band depends highly on the

number of adjacent hydrogens on the ring, and can range from 11.0 to 13.3 im.19 Matrix

isolation IR spectroscopy is a valuable tool to investigate the species (especially ions)

suspected of contributing to the UIRs, largely due to its ability to isolate molecules at the

very cold temperatures characteristic of the interstellar medium.



Experimental



Matrix isolation spectroscopy is carried out in a high vacuum chamber containing

several windows and ports for the introduction of photons and gases (Figure 1-1). As

each experiment has its own unique setup only the general details common to each

experiment will be given here, and the specifics will be summarized in each chapter. The

electron gun assembly shown in Figure 1-1 is shown for clarity only and is not present in

each experiment.

The cryostat chamber is evacuated by an two inch air-cooled oil diffusion pump

(Alcatel PDR 250, rated at 250 //sec) backed by a mechanical roughing pump (Alcatel

2004, rated at 450 //sec) and baffled by a liquid nitrogen cold trap. The pressure is

monitored by an ion gauge (Kurt J. Lesker Co. Model G075K) and ion gauge controller

(Granville-Phillips 280) spanning the range 104 though 10'8 torr. Vacuum was maintained

between nonpermanent parts with rubber O-rings. Pressures inside the chamber with the

window cooled to ca. 20K and the nitrogen trap filled were typically 3x10-8 torr. At room










Cryostat 2nd Stage---



Cryostat Window


IR BEAM


Nd:YAG Laser
Beam


IV R
+6t0V R ino


Argon Gas Inlet

*


Graphite Rod


S Electrode (0) T T -- Tantalum Anode
0 Electrode (0) L---- Ceramic Insulati

A CAB
g .... m / i

T.L + Electron Gun
:7 1-


Figure 1-1. Experimental set-up for matrix isolation of carbon anions produced by electron impact ionization.


Cathode


on









Thermocouple_
and Electrical

He In ---


He Out --


I L


Window Frame
Matrix Window
Thermal
Shield


Figure 1-2. Expander module and matrix window.













Cryostat 2nd Stage A,


Cryostat Window



IR BEAM


Graphite Rod


Yttrium Metal


1064 nm 532 nm


Figure 1-3. Experimental set-up for matrix isolation of carbon anions produced by dual laser
vaporization of graphite and yttrium metal.


Argon Gas Inlet










temperature, with only the diffusion and forepumps operating, the pressure was slightly

more than a factor of 10 higher. Pressure during deposition never exceeded 1.2x10"7 torr,

this translates into deposition times of 1 to 3 hours depending on the guest/host ratio.

The refrigerator is a two stage closed-cycle helium cryostat consisting of a

compressor (Air Products Model 1RO4W) and an expander (Air Products Model DE

202). A copper radiation shield that partially surrounds the cold window is attached to the

first stage (Figure 1-2). A window sandwiched between two halves of a copper frame is

attached to the second stage. Matrix windows were BaF2 (International Crystal

Laboratories, transparent from 200 nm to 700 cm'). The cryostat expander can be

rotated through 3600 so the matrix window can face different external ports and windows.

Ionization of carbon clusters is achieved with two methods which take advantage of

the high electron affinity ofCn carbon clusters (1.98 to 4.38 eV for 3 < n < 9).20,21 The

first uses the electron gun assembly shown in Figure 1-1. The cathode is a 0.1 mm

tungsten wire rolled into a coil and centered on a 3 mm exit hole. Emitted electrons are

captured by carbon molecules and the negative carbon cluster ions are partially extracted

from the vaporization region by an electric field created between the +60V O-ring

electrode and -120V cathode of the electron gun, and then trapped in an Ar matrix. A

canted ring electrode positioned above and adjacent to the deposition window serves to

accelerate the ionizing electron beam and also acts as a monitor of electron production.

The second method involves dual laser beam plasma generation of carbon vapor and

high energy photons from yttrium metal (Figure 1-3). Vacuum UV radiation and soft X-

rays are produced when a pulsed Nd:YAG laser beam is focused onto metal targets such










as tungsten, molybdenum, and titanium.22 In these experiments the fundamental (1064

nm) and second (532 nm) harmonics of an Nd:YAG laser are separated spatially with a

prism. The second harmonic was used to vaporize graphite from a rod, while the

fundamental was tightly focused onto a piece of tungsten metal placed in the center of the

graphite rod. The two plasma plumes are generated simultaneously about 1.5 mm apart,

and, instead of being partially extracted with a voltage, the negatively charged carbon ions

(as well as other carbon material) are cryo-pumped onto the matrix window.

Two independent methods were used to measure window temperatures. A gold-

chromel (0.07% Fe) thermocouple was joined with indium metal and placed between the

window frame and second stage. The temperature was read digitally by a programmable

temperature controller (Scientific Instruments Model 9600).

Research grade argon (Matheson Gas Products, 99.9995% pure) was used as the

matrix gas throughout. Mixtures with CCI4 and H20 were made on a vacuum manifold to

the indicated dilution. All carbon species were produced by pulsed Nd:YAG laser

(Spectra Physics Model DCR-11, 532 and 1064 nm, 10 mJ/cm2) vaporization of a graphite

rod or pellet. Mixed isotopes of carbon were produced by laser vaporizing a pellet

pressed from powders of 12C (Alfa, 99.9995% pure) and 3C (Isotec, 98.4% atom

percent). The electronic (200-1000 nm) and vibrational (7000-700 cm'1) absorption

spectra of the trapped species were measured with a Cary 17 UV/VIS/IR

spectrophotometer (0.5-1.0 nm resolution) and a MIDAC (Model M2000) FTIR

spectrometer (0.5 cm'1 resolution), respectively. Photolysis of the matrices was carried

out using a medium pressure mercury lamp with a 5 cm long H20 heat-filter and color








10

filters (veb Jena) or with a IkW Xe lamp plus a 0.25m monochrometer (2 mm slit) tunable

over the 380-550 nm region.



Overview



This dissertation is grouped into three parts. Chapters 2, 3, and 4 are concerned with

the photochemistry of the C3*H20 complex. The experimental identification of the

C3*H20 complex and the hydroxypropadienylidene isomers is detailed in Chapter 2.

Chapters 3 and 4 outline two theoretically generated potential surfaces which explain the

presence of certain C3H20 isomers and other photoproducts observed experimentally.

The two mechanisms are shown schematically in Figure 1-4 as they pertain to each

chapter. Chapter 5 reports the experimental assignments of some newly-discovered

carbon anion electronic and vibrational frequencies. Chapter 6 is a density functional

theory modeling of linear and some cyclic carbon anion vibrational frequencies with

comparison to the experimental results in Chapter 5. Finally, Chapter 7 reports the results

ofab initio calculations on rare-gas/C3 interactions. The aim of this chapter is to attempt

to model (to a first approximation) how an argon atom might perturb a 'floppy' molecule

like C3 and how C3 might influence its packing environment of argon atoms.













C3-H20


H





t-HPD


X > 400 nm


H-CC--C
H
Propynal


280 > X > 320 nm


Figure 1-4. The experimental reaction scheme relevant to Chapters 2, 3, and 4.


C30














CHAPTER 2
SPECTROSCOPY AND PHOTOCHEMISTRY OF THE C3,H20 COMPLEX



Introduction



Interest in small carbon clusters has grown rapidly since the early pioneering work of

Weltner and coworkers on the vibrational, electronic, and electron spin resonance

spectroscopy of these species isolated in rare gas matrices.23-25 Much of the current

interest revolves around the structure, size, and widespread involvement of these clusters

in various processes. They have been implicated in the formation of the fullerenes, in

flame chemistry, in soot production, and in interstellar species.26-30

Carbon cluster work has proceeded rapidly in the past several years. The linearity

(or near-linearity) of gaseous small carbon clusters (n < 9) is now generally accepted. The

structure of C3 has been investigated intensely.31"34 The very low value of the bending

frequency (v2 62 cm'l) in the gas phase suggests that the molecule is floppy. Recent

calculations33'3s show that the bending potential of free linear C3 is essentially flat up to

1600. The v1 + v3 combination mode of all six 12C/13C isotopomers of C3 have recently

been measured,36 with the conclusion that in Ar and Kr matrices the molecule is bent with

a bond angle of 160.








13

Although the main focus in the recent past has been on the structures of the carbon

clusters, little attention has been given to their reactivities with other species. The reaction

of small carbon clusters (Cn, n < 5) with H0O in Ar matrices has, however, been

investigated experimentally by Ortman, Hauge, Margrave, and Kafafi (OHMK).37 These

researchers concluded that certain carbon species, such as atoms in their 3P ground states

or diatomic molecules in their '1Eg ground or 3g- excited states, do not react with H20.

However, atomic carbon in its 'D excited state is known to react with H20 to form CO,

H2, and formaldehyde.38 In their study of C3 with H20, OHMK assigned the 2052 cmf1

band to the CCC asymmetric stretch of a C3.H20 complex. It was originally suspected

that a different species was responsible,39 but here new evidence is presented which

confirms OHMK's assignment. During photolysis with radiation of wavelengths 400

nm, OHMK observed that new bands grew in at 1999.8, 1992.8, 1459.6, 1252.5, and

1061.1 cm'1. It was suggested that these bands originated from an intermediate,

hydroxyethynylcarbene (HEC) (1), a photoproduct of the C3,H20 complex. Later, Liu,

Zhou, and Pulay (LZP) reported40 ab initio results of 1 and the cis and trans isomers of 3-

hydroxypropadienylidene (HPD) (2) and (3). From a comparison of OHMK's

experimental spectra with their calculated IR frequencies and intensities, LZP suggested

that 2 and 3, not 1, were responsible for the above IR bands.

In this chapter, results of the investigation of the photochemistry of the C3*H20

complex are reported. In addition to the identification of the intermediate(s), there remain

a number of other unsettled problems. (1) Besides the CCC asymmetric stretching mode,

what are the other experimental IR frequencies of the C3HO20 complex? (2) What are the










true identities of the 1999.8 cm'1 and 1992.8 cm'1 bands? Is one a site band of the other,

as OHMK suggest, or are they the CCC asymmetric stretch modes of cis and trans-HPD,

as LZP suggest? (3) Can the cis and trans isomers be differentiated experimentally? (4)

Why do only wavelengths in the 400 nm region induce the C3.H20 photoreaction? (5)

What are the structures of the C3HO20 complex and the cis and trans-HPD isomers in an

Ar matrix? (6) Why are three IR bands predicted for the two isomers in the 1200-1300

cmn' region, yet only two were observed by OHMK?



Theoretical



Ab initio calculations were carried out on the C3*H20 complex (4), cis and trans-

HPD (2 and 3), and the transition state between the complex and HPD isomers (5) (see

Figure 2-1) using the GAUSSIAN 92 program package.42 The calculations were carried

out at the HF/6-31G* level with the optimized geometries shown in Figure 2-1. Other

starting CH20O geometries were tested in order to determine if multiple structures might

exist. Interaction of the oxygen with the terminal carbon, the hydrogen or oxygen with

the central carbon of C3, and an O-H bond parallel to a C-C bond (both orientations) were

used as starting geometries. In all cases the geometry ultimately optimized to the

hydrogen bridged structure shown in Figure 2-1. Vibrational frequencies were calculated

at the HF/6-31G* level to determine the nature of the stationary points and to help

confirm experimental isotopomer assignments. Scaling factors for the calculated

frequencies of the various isotopomers were determined by making a ratio of the











1
167.3
1.067 1.226 1439
H C.C 1.329
175.2 108.2 H'N /H
0 0.977

2
1.334 0.976
122.1 u
1.279 1.344 109.8
C-C C
173.7 122.1093
H
3 H
\ 0.982
107.8
124.7
1.279 1.349 1.328
C-C C
089
179.2 124.6 189
H


4
H
179.0 174.1 1038 /0.968
c Y.300 2.330-
1.306 -----H -971
171.8



5 H
0.984[
171.2 107.2

1.297 1 297
0 C'. 1.273
152.1 1.207 1



Figure 2-1. Structures of the intermediates hydroxyethynylcarbene (HEC, 1),
cis-3-hydroxypropadienylidene (cis-HPD, 2), trans-3-hydroxypropadienylidene
(trans-HPD, 3), the C3zH20 complex (4), and the transition state between the
complex and the HPD isomers (5).












Table 2-1. Experimental frequencies and vibrational assignments for small carbon clusters in their ground states.


Carbon Experimental Mode
Cluster Structure Frequecies Description Remarks
C3 ('Z+) linear in gas phase3334 2038.923 asym. stretch, v3 (o,) 2040.2 cm", gas phase46
bent in Ar matrix36 1214.036 sym. stretch, v, (ag) 1224.5 cm"l, gas phase47
~8036.45 bending, v2 (x1) 63.4165 cm-1, gas phase48
C4 (3,g) linear 1543.449 asym. stretch, v3 (o) 1548.9368 cmn', gas phase54'55
C5 (' Eg) linear 216450 asym. stretch, v3 (u) 2169.44 cmn1, gas phase29'51
1446.6s3 asym. stretch, v4 (ou)
C6 (3,g) linear 1952.5s2 asym. stretch, v4 (o0) 1959.85852 cm'-, gas phase54'55
1197.353 asym. stretch, vs (Ou)
C7 ('Eg ) linear or 2128.143,54,55 asym. stretch, v4 (Oa) 2138.1951 cm"-, gas phase56-58
near linear 189451'52 asym. stretch, v5 (an)
C, (3Eg) linear or near linear expected structure
C9 ('Eg ) linear or near linear 1998.459.60 asym. stretch, v6 (cu) 2014.277964 cm"f, gas phase61










experimental to calculated all '2C isotopomer frequencies for a particular mode, then

applying the ratio to the remaining calculated frequencies.



Results



The C-2H*0 Complex



The C3,H20 complex has C, symmetry and is predicted to be bound by a mere 1.4

kcal/mol (zero-point corrected). Figure 2-2 (right) shows the asymmetric CCC stretch

region for a number of carbon clusters formed in an Ar matrix at 11K. The initial

distribution for cluster sizes in the matrix depends on the power density of the laser beam,

matrix deposition temperature and the Ar gas pressure in the deposition chamber. The

clusters C3 and C2, and presumably C atoms, are favored when higher laser powers are

used. A much broader distribution of clusters is made when the beam is defocused or

lower laser powers are used. Aggregation of the clusters is readily achieved by annealing

the matrix. This results from diffusion of the Cn species in solid Ar which increases rapidly

with increasing temperature in the 20-36K range43 and from the exothermic heats of

formation of the carbon clusters.44 In earlier work many of the IR bands observed in this

region were assigned using matrix and/or gaseous samples. These assignments are

compiled in Table 2-1.

Water is ubiquitous in matrix isolation experiments. Its removal is exceedingly

difficult due to its high affinity for all surfaces. Its presence is however easily noted by the










R
to
C1 2 + 1/3
C17




CO
CD
Go
C12
to


00,



Cr
a"

*^
C> L C i




'I C6
CO, >* x
O CO a

C,0C CO


C.



-i

cd
I0
/I


0-
38(


3700


3600
WAVENUMBERS [cm -1]


2000


1900


Figure 2-2. Portions of the 10K matrix infrared spectrum of '2C/Ar + 120 (0.1%) in the asymmetric stretch region
(ca. 2000 cm'-) and in the OH asymmetric (v3) and symmetric (v,) region (ca. 3700 cm'1). Monomer (M) and dimer
(D) bands are marked; C3*H20 complex bands are starred. Inset shows the v2 + v3 combination band of H20.


CO7




12








*v "
SO


LO
vW -
u-4'
I -i


I :: I I *v


J
*,


CO
rC
C3


30










appearance of the IR bands in the OH stretch region (ca. 3700 cm-'). Figure 2-2 (left)

shows the IR bands of matrix-isolated noninteracting H20 monomers (M) and (D) for a

sample formed from a mixture of Ar/0.1% HO2/Cn clusters. The inset shows the region of

the v, + v3 combination bands of noninteracting H20. In addition to known bands

attributable to water and carbon clusters, the figure also shows two bands at 2052.3 cm'1

and 3598.0 cm'1 (starred). The 2052 cm'" band has been assigned previously by OHMK

to the asymmetric CCC stretch of the C3H20 complex.

The 2052.3 cm'l band. The vibrational frequencies and intensities for the C3H20O

complex are given in Table 2-2. Twelve modes are predicted but only five intraspecies

modes were observed. The calculated intermoiety modes and the C3 bending mode are

included in the table but were not observed since their frequencies are below the range of

the spectrometer employed. The predictions include three perturbed OH modes at 1626.8,

3604, and 3705.2 cm'1 and two perturbed CC stretches at 1216 and 2052.3 cm'1. The

2052 cm'1 mode is predicted to be the most intense, with three others about one-tenth as

intense and one practically zero. These bands are compared in the table to the

experimental bands assigned to the C3oH20 complex. A discussion of these assignments

follows below.

A mixed 12C/'3C isotope run aided in the assignments of the bands and helped in

establishing the geometry of the complex. For a symmetric complex (e.g., one with the

H20 bound symmetrically to the middle carbon) six isotopomer peaks are expected for the

asymmetric CCC stretch. However, for an asymmetrically bound complex (e.g., H20

bound to one end of the C3) eight isotopomeric peaks are predicted. The experimental













Table 2-2. Comparison of experimental (Ar matrix) and calculated (HF/6-31G*) IR
frequencies and intensities for the C3*H20 complex. Frequencies in cm"n and intensities
in km/mol.



Mode
vx Intensity (exp) vea" Intensity (calc)b Description


154.7 (0.11)

82.0 (0.06)

1450.5 (1.0)

117.4 (0.08)

2.9 (0.002)

168.7 (0.12)

74.3 (0.05)

7.9 (0.005)

29.9 (0.02)

1.29 (0.001)

8.3 (0.006)

0.2 (0.0)


I & .1 I


OH asym. stretch

OH sym. stretch

CC asym. stretch

HOH bend

CC sym. stretch


a Calculated frequencies scaled by 0.8856.
b Relative intensities in parenthesis.
C Expected value based on the 3258.9 cm'1 (v,
by H20 in the C3*H20 complex (see text).


+ v3) combination band of C3 perturbed


3712.2

3598.0

2052.3

1593.4

1214.2


(0.1)

(0.09)

(1.0)

(0.11)

(0.0)


3705.2

3604.0

2052.3

1626.8

1216.0

270.0

203.2

144.0

138.7

83.9

37.2

33.8













I 12c:13c12


01






SI12C 2039 cm-1 1998cm-1


2052cmri1
I _

2050 2025 2000 1975
/cm-1"

Figure 2-3. Portion of the annealed (32K) matrix infrared sprectra of '2C/3C/Ar
([2C]:["1C]=1:1) (top) and of 12C/Ar (12K) (bottom) in the asymmetric CC stretching
regoin. The C3 isotopomers are marked by triangles, while the C3.H20 complex
isotopomers are marked with dots. The marked bands in the lower spectrum are due
to C9 (1998 cm'), C3 (2039 cm'l), and the C3.H20 complex (2052 cm1).













Table 2-3. The experimental (Ar matrix) and calculated (HF/6-31G*) IR frequencies
for the CCC asymmetric stretching mode of '213C3,H20 complexes. Frequencies in
cm-.


Vcalc Vlcb
v (FG+force const. fit) (HF/6-31G*) Isotopomer

2052.3 2052.82 (+0.52) 2052.3 (0.0) 12-12-12-1-16-1

2040.0 2040.19 (+0.19) 2040.1 (+0.1) 13-12-12-1-16-1

2038.5 2038.87 (+0.37) 2038.2 (-0.3) 12-12-13-1-16-1

2026.0 2025.33 (-0.67) 2025.5 (-0.5) 13-12-13-1-16-1

2000.5 2000.69 (+0.19) 1998.9 (-1.6) 12-13-12-1-16-1

1988.2 1987.86 (-0.54) 1986.5 (-1.7) 13-13-12-1-16-1

1986.5 1986.26 (-0.24) 1984.6 (-1.9) 12-13-13-1-16-1

1973.3 1972.40 (-0.90) 1971.5 (-1.8) 13-13-13-1-16-1



a Program from reference 62; deviations from experimental frequencies given in
parentheses; RMS deviation of all 8 bands is 0.44 cm"1. Force constants found in
force constant adjustment calculation: ACI-C2) = 11.213;AC2-C3) = 10.932;
fC3,Hi) = 2.079;ftHi-O) = 9.483;fO-H2) = 5.331;f Ci-C2, C2-C3) = 1.239, and
bfH1-O, O-H) = -1.208 in units of 102 N m1.
b Scaled by 0.8856. Deviations from experimental frequencies given in parentheses.










spectrum of the 1213C3.H20 complex in the asymmetric CCC stretch region is shown in

Figure 2-3. The band positions, collected in Table 2-3, are compared to a normal

coordinate force constant adjustment calculation62 and the ab initio frequencies. Both the

force constant fit and the ab initio frequencies for a planar asymmetrical complex are close

to the experimental values, leading to the conclusion that the 2052.4 cm'1 band is due to

the asymmetric CCC stretch of the planar asymmetrical complex.

The 3598 cm' band. The spectra in Figures 2-4 to 2-6 give different regions of the

IR spectrum of a CJ/H20/Ar matrix before photolysis for a higher concentration (0.2%)

than in Figure 2-2 (0.1%). The intensities of the 2052.4 cm'' and 3597.6 cmn' bands both

increase with increasing water concentration while the 2038.9 cm'1 band (due to

uncomplexed C3) decreases. Forty different spectral runs, including photolysis and

annealing experiments, established a positive correlation (correlation coefficient = 0.97)

between the 3598 cmn1 band and the 2052.4 cm"' band intensities. The 3598 cm'1 band,

because of its proximity to the calculated frequency at 3604 cm'1 and the close match of

experimental and calculated relative intensities, is attributed to the OH symmetric stretch

of the C3'H20 complex.

The 3712 cm'1 band. The OH asymmetric stretch in the complex is predicted to

appear at 3705.2 cm"1. As Figure 2-4 shows the experimental spectrum is very crowded in

this region, with monomer, dimer, and trimer H20 bands appearing. Upon photolysis

changes are detected in this region. The changes are complicated by overlap with the

3711.5 cm-' band (1., 00 transition of the v3 H20 monomer mode) which is expected to

remain unchanged upon photolysis. There are several reasons for believing that the



















7 M i
I I Ir

Before Photolysis (b)

0-t I After Photolysis (a)


Difference (a-b)


3800 3700 3600 3500 3400
WAVENUMBERS [cm -1]

Figure 2-4. Portion of the 12K matrix infrared spectra of 'Cn/Ar + HO (0.2%) in the OH stretch region
before photolysis (b) and after photolysis (a). The monomer (M), dimer (D), and trimer (T) bands of
water are marked accordingly. The starred bands are the OH stretches in the C3*H20 complex.















.21 I A r

Photolysis

SI Photolysis











22o0 21e00 2o 0

WAVENUMBERS [cm -1]
Figure 2-5. Portion of the 12K matrix infrared spectra of 12Cn/Ar + H20 (0.2%) before photolysis and after
photolysis (X> 390 nm). Bands which have grown upon photolysis are due to C30, propynal, and trans-HPD. Note
that the C3.H20 complex band has decreased upon photolysis. C3.(H20)n (n = 2 and 3) bands are marked by stars.









U


t-HPD


X 0. 1 cS I I-

.2 CO

CO)
w. -W After Photoly,




.1 -Before Phot4




Difference (

0


1400 1200 1000
WAVENUMBERS [cm -1]

Figure 2-6. Portion of the 12K matrix infrared spectra of 2Cn/Ar + H20 (0.2%) before photolysis (b)
and after photolysis (a) (X > 390 nm). All bands in the difference spectrum are due to trans-HPD.


800


olysis (b)










negative peak in the difference spectrum contains a contribution from the temperature

dependent 3711.5 cmn' band and the 3712.2 cm'1 C3,H20 complex band. Namely, (1) the

intensity ratio of the 3755.7 cm'1 (Oo 1.1) band to the 3777.9 cm'1 (1.- 2.2) band of the

H20 monomer is larger in spectrum (a) than in spectrum (b), indicating that the

temperature is lower in (a). This is reasonable since (a) was recorded an hour later than

(b), after the matrix had more time to cool. (2) The intensity ratio of the 3777.9 cm"' and

3711.5 cm'1 bands is known63 to be 1.4 at 15K in an Ar matrix. While it is not possible to

determine this ratio here, it can be deduced from the v2 + v3 combination bands (Figure

2-2). The 5345.2 cm'1 (v2 + 3776.4 cm"') band is slightly more intense than the 5280.1

cm' (v2 + 3711.5 cm'1) band, which here is shifted to 3710.8 cm"1. It is expected that the

3776.4 and 3710.8 cm'1 bands should be in approximately the same ratio. But, as the

difference spectrum in Figure 2-4 shows, the intensity ratio is inverted. Thus it is

reasonable to conclude that another species is contributing to this band. Because of its

proximity to the predicted 3705.2 cm' band, the 3712.2 cm'1 band is assigned to the OH

asymmetric stretch of the C3,H20 complex.

The 1214.3 and 1593.4 cm'1 bands. A perturbed HOH bend is predicted for the

complex at 1626.8 cm"1 with an intensity intermediate between the two OH stretch modes.

A band at 1593.4 cm1', having the correct behavior can be assigned to this mode. With

photolysis this band decreases in intensity, in parallel with the 2052.2 cm'1 band, as

expected.

A symmetric CC stretch is predicted at 1216 cm'1 and is expected to have very low

intensity. Its presence can, however, be determined from the v, + v3 combination region.














.5 -


C 4 H 20 I






O 9


0 4


0-
I I I




370 380 390 400 410 420

WAVELENGTH [nm]


Figure 2-7. Electronic spectra for 12C3 and '2C3*H20 complex before photolysis
(upper) and after (lower). Band at 405.4 nm is assigned to the 12C3*H20 complex.










Two weak bands are observed in this region. One at 3245.2 cm', which has been

previously assigned36'37 to the v, + v3 combination band in C3, and the other at 3258.9

cm-' which we assign here to the v, + V3 combination band of the complex. To determine

the frequency of the v, mode, we use the anharmonicity factor of 7.7 cm' found36 for the

combination mode of C3 in Ar. If this value is correct for the complex also, it is expected

that the CC symmetric mode frequency in C3.H20 should be found at 3258.9 2052.4 +

7.7 cm'1 = 1214.2 cm"1. This is very close to the observed v1 value of 1214.0 cm"1 for C3

in Ar and to the 1216 cm"1 value calculated here for the complex. Furthermore, this band

decreases upon photolysis of the matrix, as expected. Based on this evidence the CC

symmetric stretch in the complex is assigned to the band at 1214.2 cm'.

Electronic spectrum. OHMK showed previously37 that photolysis with X = 420 nm

produces no reaction, but with X > 400 nm photochemical activity is observed. Since

photolysis in the 400-500 nm region destroys the C3.H20 complex, the electronic band of

the complex responsible for activation was sought by scanning this region before and after

photolysis. Figure 2-7 shows that a band at 405.4 nm disappears upon photolysis.

Furthermore this band correlates well with the 2052.3 cm' band (correlation coefficient =

0.95). The integral intensity ratio, I (405.4 nm)/I (2052.3 cm'1), is found to be ca. 70.

The 405.4 nm band we assign to the electronic band origin of the complex, a transition

analogous to the 'H 'Y transition ofuncomplexed C3 at 408.4 nm.23-25 The above

assignment demonstrates that the rearrangement of the C3*H20 complex to the isomers of

HPD and/or C30 requires the excess energy acquired from excitation to an excited

electronic state of the complex.










C30*HO to HPD transition state. Using ab initio calculations, the transition state

(TS) connecting the C.H20O complex and the trans-HPD isomer has been located. The

optimized TS geometry found is illustrated in Figure 2-1. The C3H20O complex is planar

with a long C-H hydrogen bridge (2.492A). The TS geometry is nonplanar with the

terminal carbon interacting with both the hydrogen and oxygen of H20. The C-H bond

length has decreased to 1.257A on its way to 1.075A in the HPD isomers. The calculated

barrier of 29.3 kcal/mol (MP2/6-31G* + zero-point correction) from complex to transition

state represents a substantial endothermic barrier and explains why instantaneous reaction

between C3 and HZO is not seen.



The trans-HPD Isomer



Photolysis of the matrix isolated mixture described above with X > 390 nm leads to a

decrease in the bands at 2066.8, 2056.2, 2052.4, 3597.6, and 3712.2 cm-1. The latter

three are due to the C3*H20 complex, while the former two are tentatively assigned here

to the C3 (H20)3 and C3(H20)2 complexes, respectively. As a result of this photolysis,

new bands appear at 2243.2, 2108.4, 1992.5, 1688.4, 1460.8, 1280.5, 1223.1, 1016.3,

and 940.1 cm'1. The 2243.2 and 2108.4 cm'1 bands have previously37'64 been assigned to

the asymmetric CCC stretch of C30 and CC stretch of propynal, respectively. The

remaining peaks must clearly also be photolysis products. It will be shown below that

they are due to trans-HPD.














Table 2-4. The experimental (Ar matrix) and calculated IR frequencies (cmf') and
intensities (km/mol) for trans-3-hydroxypropadienylidene (trans-HPD) in its ground
electronic state.



trans-HPD

HF/6-3 G* DZP+diffuseb
Mode
Vv (cm-) Vcac Intensitysity b Iensityc Description
3619.7 120.4 (0.11) 3669 127(0.10) OH str.

3051.2 10.3 (0.01) 3080 6(0.005) CH str.

1992.5 (1.0) 1974.0 1104.7 (1.0) 1940 1274 (1.0) CCC asym. str.

1460.8 (0.14) 1487.9 292.0 (0.26) 1482 355 (0.28) COH bend, CH str.

1280.5 (0.04) 1311.5 63.8 (0.06) 1309 105 (0.08) COH bend, CO str.

1223.1 (0.16) 1222.6 265.5 (0.24) 1216 284 (0.22) HCO, CCH bend

1016.3 (0.12) 1023.4 109.5 (0.10) 1018 91(0.07) CC, CO str.

997.0 0.4 (0.0004) 995 1 (0.0008) CH wag

655.5 200.3 (0.18) 657.5 193 (0.15) HOCC, HOCH tor.

589.7 40.8 (0.04) 596.0 46 (0.04) CCO bend

196.0 2.6(0.002) 267.2 1(0.0008) CCCO bend

177.2 3.8 (0.003) 211.4 3 (0.002) CCC bend


a Frequencies scaled by 0.897.
b From reference 40. Frequencies have been scaled by 0.9.
c Normalized relative intensities in parenthesis.











Expt. Difference Spectrum( Photolyzed Unphotolyzed)


= '- 4 I ')



Theoretical Spectrum (HF/6-31G*) for t-HPD
-JJ

.1 -




1o r, I I




2000 1500 1000
WAVENUMBERS [cm -1]
Figure 2-8. Comparison of the experimental (upper) and theoretical (lower) spectra oftrans-HPD. Bands
with one star, two stars, and dots are due to the C3*H20 complex, propynal, and trans-HPD, respectively.










LZP have shown40 on the basis of extensive ab initio calculations that one of the

major products of photolysis of the C3H20O complex is probably trans-HPD. In Table 2-4

are presented experimental frequencies and intensities, results from LZP's calculation and

ab initio calculations for trans-HPD from this work. A graphical comparison is given in

Figure 2-8. It can be seen that the correspondence for the bands at 1992.5, 1460.8,

1280.5, 1223.1, and 1016.3 cm'1 is quite good. Even at the SCF level the calculated

frequencies and intensities match reasonably well with experiment, the largest frequency

deviation being ca. 30 cm1'.

Further support for this assignment comes from '2C/'3C isotopic studies. Taking

into account all possible unique arrangements of 12C and 3C in the HPD molecule, eight

isotopomeric bands are expected. In Figure 2-9 are shown the 12C/13C mixed isotope

spectra before and after photolysis. The difference spectrum shows a series of eight

negative peaks, attributed above to the isotopomers of the C3.H20 complex. Also shown

in the difference spectrum are a series of eight positive peaks which are also listed in Table

2-5. The table gives the results of the ab initio calculations of the IR frequencies for the

eight isotopomers of trans-HPD. As confirmation of this assignment, the isotopomeric

band structure was sought for the next most intense trans-HPD band at 1223.1 cm'1. A

set of five weak bands, found in this region, are displayed in Figure 2-10. The energy

spread of these bands is much smaller than the spread in the 1992.5 cm'1 range (5.5 cm-1

vs. 84.6 cm'1). This is because the 1223.1 cm'l mode is a combination HCO/CCH bending

vibration while the 1992.5 cm' band is an asymmetric CCC stretching vibration. The

match between the observed and calculated peaks, given in Table 2-5, is quite good which















t-HPD
0 *@ 0 @ 0


to o co o / --
S N iCr iwM II



IDifference (A-B)

A AA A A A .


C3, H0,




A2c I


Before
Photolysis


After
Photolysis


2050


2000 1950
WAVENUMBERS [cm -1]


1900


Figure 2-9. Portion of the 12K matrix infrared spectra of 2C/1'C/Ar + H20 (0.15%)
([12C]:[13C]=1:2) before photolysis and after 30 min of photolysis (X > 390 nm). The
negative peaks in the difference spectrum (triangles) are the asymmetric CCC stretches of
the '2,3C3H20O complex isotopomers. The positive peaks (dots) are the asymmetric CCC
stretches of the '2,3C3-substituted trans-HPD isotopomers.


_









Table 2-5. Experimental and calculated (HF/6-31G*) IR frequencies (cm'') for all isotopomers of cis and trans-HPD for
the most intense CCC asymmetric stretch, CCH+HCO bend (trans-HPD) and HCO bend + CO stretch (cis-HPD) modes.


trans-HPD cis-HPD

Isotopmer Vexp vcl vexp cl vexp Vcal Vexp- Vcl-

12-12-12-1-16-1 1992.5 1989.7 (0.99)ab +2.8 1999.6 2000.5 (1.00)d -0.9
12-12-13-1-16-1 1984.8 1983.3 (0.95) +1.5 1991.3 1992.5 (0.96) -1.2
13-12-12-1-16-1 1974.7 1970.2 (1.00) +4.5 1982.0 1982.7 (1.00) -0.7
13-12-13-1-16-1 1965.8 1963.3 (0.96) +2.5 1973.1 1973.9 (0.96) -0.9
12-13-12-1-16-1 1937.6 1938.9 (0.95) -1.3 1950.2 1950.4 (0.95) -0.2
12-13-13-1-16-1 1923.4 1932.5 (0.91) -9.1 1942.2 1942.0 (0.91) +0.2
13-13-12-1-16-1 1918.5 1918.9 (0.96) -0.4 1932.8 1932.3 (0.96) 40.5
13-13-13-1-16-1 1907.9 1911.9(0.92) -4.0 1923.5 1922(0.91) +0.6
12-12-12-1-16-1 1223.1 1223.1 (0.85)cf 0.0 1254.3 1254.3 (0.97)0'f 0.0
12-12-13-1-16-1 1220.2 1220.1 (0.97) +0.1 1247.1 (0.73)
13-12-12-1-16-1 1221.7 1221.9 (0.87) -0.2 1254.2 (0.98)
13-12-13-1-16-1 1218.6 1218.7 (0.99) -0.1 1246.9 (0.75)
12-13-12-1-16-1 1221.7 1222.4 (0.86) -0.7 1253.7 (0.99)
12-13-13-1-16-1 1218.9 1218.7 (0.97) +0.2 1246.4 (0.74)
13-13-12-1-16-1 1221.7 1221.2 (0.88) -0.5 1253.7 (1.00)
13-13-13-1-16-1 1217.6 1218.0 (1.00) -0.4 1246.2 (0.77)

a All frequencies for this mode scaled by 0.904.
bNormalized relative intensities in parenthesis.
C Scaled by 0.897.
d Scaled by 0.903.
e Scaled by 0.879.
Due to the overlapped isotopomer bands, the assignment of the experimental frequencies for the CCH+HCO bend mode
is tentative.












.002- r- l


t-HPD : (HCO/CCH Bend)
Isotopomeric Structure (2 C / '" C)








0

1240 1230 1220 121
WAVENUMBERS [cm -1]

Figure 2-10. Portion of the 12K matrix infrared spectrum of 2C/'3C/Ar + H20 (0.15%)
([12C]:[13C]=1:2) after 30 min photolysis (X > 390 nm). Isotopomeric bands are due to trans-HPD.


0










leads to the conclusion that the set of peaks at 1992.5, 1460.8, 1280.5, 1223.1, and

1016.3 cm' originate from trans-HPD.




The cis-HPD Isomer



Normally, photolysis of a matrix occurs just prior to a scan. Occasionally, however,

photoproducts are produced whose lifetimes are so short that they are not observable

under such an experimental protocol. Therefore scans were taken during photolysis. The

matrix sample window was aligned at 450 to the photolysis beam and the IR probe beam.

Figures 2-11 and 2-12 show the results of spectra taken during photolysis. Two new

bands at 1999.6 and 1254.3 cm1' grow in, while the five bands, attributable to trans-HPD,

all decrease in parallel. With the photolysis light source off, the process is reversed. This

strongly suggests a photolysis induced process, with spontaneous reversion in the dark.

The species producing the 1999.6 and 1254.3 cm'1 bands will be shown to be cis-HPD.

LZP have predicted40 that HPD can also exist in the cis form. Their calculations of

the IR frequencies and intensities for cis-HPD, compared with this work, are given in

Table 2-6. The most intense bands are predicted to lie at 1987.2 and 1279.8 cm'1, the

former about twice as intense as the latter. This is the same pattern observed for the two

new bands at 1999.6 and 1254.3 cm-'.

Further support for the attribution of these bands to cis-HPD comes from '2C/13C

isotopic studies. Figure 2-13 gives the spectra of the 12C/13C isotopically substituted















U) CV3 CM
CM 0>
xx 88 I During Photolysis L
ihtl '0


Bef

S^^thwU~l^A^


I I
I I


Dif


C3. H 20


ore Photolysis I



ference (B-D)



C3* H 20O


C6I



oco
CO. I-
CO
MI
040
CM~
J ^


3250


2100 2050
WAVENUMBERS [cm -1]


2000


Figure 2-11. Portion of the 12K matrix infrared spectra of 12C/Ar + H20 (0.2%) before and
after 30 min of photolysis (X > 390 nm), but recorded during photolysis. The 3245.2 and
3258.9 cm"f bands are the v, + v3 combination bands in C3 and the C3.H20 complex.


.2-


3300


rv~v~vr~M'nlu ~r7rr` YYr










.02


0
0 .01






O


1400 1200 1000
WAVENUMBERS [cm -1]


Figure 2-12. Portion of the 12K matrix infrared spectra of 2C/3C/Ar + H20 (0.15%)
([12C]:['3C]=1:2 before photolysis and after 30 min photolysis (X > 390 nm), but recorded
during photolysis. The 1254.3 cm-' band is the COH bend + CO stretch ofcis-HPD.














Table 2-6. The experimental (Ar matrix) and calculated IR frequencies (cm-1) and
intensities (km/mol) for cis-3-hydroxypropadienylidene (cis-HPD) in its ground
electronic state.



cis-HPD

HF/6-31 G* DZP+diffuseb

Mode
Vec VCa Int Iensity b nten Description
3676.9 208.4 (0.19) 3714 221 (0.17) OH str.

3000.0 21.9 (0.02) 3020 17(0.013) CH str.

1999.6 (1.0) 1987.2 1120.0(1.0) 1951 1284(1.0) CCC asym. str.

1254.3 (0.35) 1507.3 82.3 (0.07) 1495 116(0.09) COH bend, CH str.

1279.8 494.9 (0.44) 1281 628 (0.49) COH bend, CO str.

1249.1 178.8 (0.16) 1240 104 (0.08) HCO, CCH bend

1011.5 2.2 (0.002) 1001 1(0.001) CC, CO str.

968.3 2.2 (0.002) 956.3 2 (0.002) CH wag

592.3 13.1 (0.01) 594.5 15 (0.01) HOCC, HOCH tor.

489.5 116.2 (0.10) 494.0 106 (0.08) CCO bend

210.5 15.2 (0.01) 267.8 32(0.02) CCCO bend

193.0 1.1 (0.001) 213.5 2(0.002) CCC bend


a Frequencies scaled by 0.897.
b From reference 40. Frequencies have been scaled by 0.9.
c Normalized relative intensities in parenthesis.












Difference (hv.-hvy.) x 5


S* t-HPD

After 30 min. Photolysis (hvo.)




After 30 min. Photolysis (hv.)

0
2050 2000 1950 1900
WAVENUMBERS [cm -1]
Figure 2-13. Portion of the 12K matrix infrared spectra of 2C/1C/Ar + H20 (0.15%) ([12C]:[13C]=1:2) after 30 min of photolysis
(X > 390 nm) and recorded with the photolysis lamp on (bottom) and off(middle). The difference spectra (top) are shown
the CCC asymmetric stretching modes of the isotopomers of cis-HPD (positive peaks) and trans-HPD (negative peaks).








42

species recorded with and without the photolysis lamp on. The difference spectrum shows

the negative peaks of trans-HPD, beginning at 1992.5 cm'' and stretching to 1907.9 cm"1.

Also shown is a set of eight positive peaks starting at 1999.6 cnm' and ending at 1923.5

cm-1, which are created as a result of photolysis. The negative peak due to trans-HPD at

1923.4 cm' is not seen in this spectrum due to overlap with the 1923.5 cm'1 isotopomer

peak ofcis-HPD. Table 2-5 lists the results of the ab initio calculations of the eight cis-

HPD isotopomeric bands. Isotopomeric bands in the 1254.3 cm"' region were searched

for, and none were detected. Presumably this is due to the expected weak intensity.



HPD Isomer Interconversion



The process by which the two isomers interconvert can be investigated by focusing

on similar vibrational modes in the two isomers and monitoring their intensity with time.

First, the kinetic behavior is tracked by looking at the CCC asymmetric stretching modes.

At long times after photolysis, the absorbance of the 1999.6 cm"' cis-HPD band is a factor

of 1.26x10'2 smaller than the absorbance of the 1992.6 cm'1 trans-HPD, which is at a

maximum. The difference spectra (A, Aj) for both isomers as a function of time are

plotted in Figure 2-14. The positive peak (1999.6 cm'1, cis-HPD) is at a maximum at the

photolysis cutoff oft = 0, and decreases with time. The negative peak (1992.6 cm-',

trans-HPD) is at a minimum at t = 0 and increases to a maximum at t = co. (An increasing

absorbance will appear as a negative peak decreasing in amplitude with time.) The fact

that the integral intensity of the 1999.6 cm'1 (cis-HPD) positive peak is approximately



























0 -*

- *-


0 5 10 15 20
TIME [min]


Figure 2-14. Time dependence of the difference spectrum (At A.). The t = 0 and
subsequent plots were scanned after the photolysis lamp was turned off. Each spectrum
was recorded for ca. 10 sec, with A scanned after 2 hr. The 1992.5 cm1' (trans-HPD)
band is at maximum absorbance at t = 2 hr, at which time the 1999.6 cm'n (cis-HPD)
band is at its minimum absorbance.


c-HPD


.01-


-.01


(xx


c~i

L.


,


t-HPD








44

equal to the integral intensity of the 1992.6 cm' (trans-HPD) negative peak indicates that

the product of the concentration and the transition strength for this mode is about the

same for both cis and trans isomers. Table 2-6 shows that calculations predict the same

transition strength of the CCC asymmetric stretch for both isomers.

Next, the behavior of the combined COH bend/CO stretch modes in the two isomers

is investigated. The difference spectrum shows that the integral intensity of the 1254.3

cm'1 band (cis-HPD) is approximately twice as intense as the 1223.1 cm'1 band (trans-

HPD), if one assumes that the concentration changes {A[cis-HPD] = A[trans-HPD]} in

the light induced process are the same for both isomers. The latter is a reasonable

assumption if the isomers interconvert and are not formed by any other process. Both

calculations predicted that this mode in cis-HPD should be approximately twice as intense

as in trans-HPD. The theoretical calculations predict both intensities and frequencies for

the two HPD isomers reasonably well.

The dark reaction relaxation time and (crudely) its temperature dependence may be

determined by using the data in Figure 2-14. Both data sets were fit to the first-order

expression



[At A,] = [At A] exp (-t / t),



where t is the relaxation (appearance or disappearance) time for either isomer. If one

isomer converts directly into the other, the relaxation times should be the same for both.

Figure 2-15 shows the plot of In IA, A. vs. time for both isomers at two different matrix












5 .5 . .. .. .


-6 . s^ T=11K .
Sk=1 (kSUx 10-3 o'

< -6.5- -







AE= 0.06kJ/mol I

0 5 10 15 20

TIME [min]

Figure 2-15. Dependence of lnl|A -A. vs. time after 30 min of photolysis (X > 390 nm) of the 2Cn/Ar + H20
(0.15%) matrix for T = 11 and 20K. The A, and A, absorbances are identical to the ones shown in Figure 2-14.










temperatures. The slopes (= -1/t) at a given temperature are the same for each isomer.

Hence, it can be concluded that the cis-HPD conformer relaxes completely to trans-HPD

with a relaxation time of = 10.8 min (k = 1.54x10"3 s') at 11K and T = 8.2 min

(k=2.03x10" s"') at 20K.

Although higher temperature measurements were attempted, aggregation of the

carbon clusters interferes at temperatures higher than 20K.41 In addition, other reaction

channels, probably different for the two isomers, open up. We observed different values

for the relaxation times for the two isomers at higher temperatures. In addition, we found

that the 1999.6 cmn' (cis-HPD) peak of(A- A.) was more intense than the 1992.6 cm"'

(trans-HPD) peak at elevated temperatures.

Despite this difficulty, it is possible to obtain a rough estimate of the activation

barrier for the cis to trans-HPD isomerization from the 11K and 20K runs. Using


AE
lnk= E +lnA
RT



we find AE, = 0.06 + 0.02 kJ/mol and A = 2.84x10'3 s'-. This very small value leads to

the question of whether there is a barrier at all. This value is equivalent to ca. 5 cm'1.

This is compared to the calculated barrier of 34.7 kJ/mol (2900 cm'). There are two

possible explanations for such a low value. First, it could simply represent a lower bound

on the activation energy since the energies used in the IR probe beam in this experiment

(7000-700 cm"-) are much greater than this value. Or, the process involved in the

conversion could involve hydrogen tunneling. (OHMK reported that photolysis








47

experiments involving a C3.D20 complex were slow and required four times as long as the

C3*H20 complex.)37

The difference in stabilization energies for the two isomers may be determined from

the intensity distribution of their CCC asymmetric stretching vibrations. At long times

after the photolysis is terminated, the 1992.6 cm'1 band is about 79 times more intense

than the 1999.6 cmn' band. Taking the integral intensities of these two bands to be equal,

the equilibrium constant for the two isomers at 11K is



[cis-HPD] =1.2X1O-2.
S[trans-HPD]


From this value the standard free energy difference for the two isomers can be obtained

from the expression


AGllK0 = -RT InK1lK = 0.4 + 0.15 kJ/mol,



where AG IKo is the standard free energy difference between cis-HPD and trans-HPD

isomers in equilibrium in an Ar matrix. Figure 2-16 shows a schematic diagram of the

potential energy curve involved in the isomerization process. The cis isomer is less stable

than the trans. The experimental value of 0.4 kJ/mol is comparable to the standard free

energy difference of 0.12 kJ/mol for the two hydroxy isomers of amino-hydroxy-9-

methylguanine in Ar at 12K.6 It may be compared also with the theoretical value of the

internal energy difference between the two isomers. LZP in their ab initio studies4" of the















H


C==C=-=C
\H


AGi11K


- -


CIS


TRANS
TRANS


Figure 2-16. Potential energy surface for the rotation of the H atom about the C-O
bond of HPD in its ground electronic state. See text for AE and AGo values.


-T
TRANS










HPD products predicted that the cis form should lie about 23.4 k/mol above the trans

form. This calculated difference in stabilization energies is practically independent of

method; RHF, MP2, MP3, or MP4(SD) give the same results to + 1.3 k/mol and are very

different from the experimental value. While the matrix is expected to influence this value

somewhat, this large discrepancy suggests the possibility of a different mechanism

involved in the conversion process.

LZP also located the transition state for the isomerization of cis to trans-HPD. They

calculated that the energy of the transition state lies 46 kJ/mol above the cis singlet ground

state (MP2/DZP level). The almost negligible value of 0.06 kJ/mol for the activation

barrier we find here is also very different from the calculated one. As LZP point out, the

interconversion could be occurring via hydrogen tunneling, for which there should be no

activation barrier.



Discussion



Our observations may be significant with respect to the mechanisms of chemical

reaction in the interstellar medium. Both C30 and propynal, the final products of

photolysis of the C3,H20 complex, have been observed64'65 in almost equal abundance in

the cold (10K) molecular cloud, TMC-1. Herbst, Smith, and Adams have postulated66

that the following ion-molecule reactions are the principal pathways of formation for C30

and H2C30:










C2H2" + CO H2C3O+ + hv (la)

H2C30- + e" C30 + H2 (Ib)

HC30 + H (Ic)

and

C2H3 + CO H3C3O+ + hv (2a)

H3C3O+ + e C30 + H2 + H (2b)

H2C30 + H (2c)

On the other hand, Brown and coworkers have suggested67 that the most important

pathway for C30 production is:



O + C4 -30 + C (3a)

C3 + CO (3b)



The estimates by Brown and coworkers67 and by Herbst and Leung6 for the formation of

C30 are in good agreement with observations. Irvine et al. point out65 that, although the

branching ratio in (2) is not known from laboratory studies, the fact that nearly equal

abundances of CO3 and propynal are observed, must mean that (2c) is a significant

reaction pathway.

Results presented here suggest a different mechanism. As we have shown, the

C3.H2O complex readily forms in low temperature environments and, when photolyzed,

forms the intermediates, cis and trans-HPD. Further photolysis results in C30 and










propynal formation. Desorption from the surface of the dust particles could then take

place by any of the mechanisms proposed by others previously.69 The matrix environment

used in our studies is obviously different from that of dust particles, but the conditions of

low temperature and exposure to visible and/or ultraviolet radiation are common to both.

The low activation energy measured for cis to trans isomerization of HPD might suggest

that hydrogen tunneling is present. Herbst has recently proposed70 that hydrogen

tunneling could be a viable low temperature process for the reaction between neutral

species of astrophysical importance.



Summary



In this chapter the existence of a complex between C3 and H20 in an Ar matrix has

been confirmed via '2C/13C isotopic labeling studies using FTIR spectroscopy in

conjunction with ab initio calculations. IR frequencies attributable to the complex have

been observed at 2052.3 (asym. CCC str.), 3598 (OH sym. str.), 3712.2 (OH asym. str.),

1593.4 (HOH bend), and 1214.3 cm"1 (CC sym. str.). An electronic band at 405.4 nm is

assigned to the perturbed 11 ,- 1F transition of the C3 portion of the complex. The

ground state geometry of the complex is determined to be planar with one hydrogen of the

H20 weakly bound to a terminal C of the C3 with an intermoiety angle of 175.

Photolysis of the C3.H20 complex results in the formation of C30 and cis and trans

3-hydroxypropadienylidene (HPD). IR peaks assigned to trans-HPD have been observed

at 1992.5, 1460.8, 1280.5, 1223.1, and 1016.3 cm-'. Further photolysis of trans-HPD










results in isomerization to cis-HPD, which exhibits IR bands at 1999.6 and 1254.3 cm'1.

Two sets of eight '2C/'3C isotopomeric bands have ben observed for the intense 1992.5

cmn1 trans-HPD and 1999.6 cm'~ cis-HPD asymmetric CCC stretching modes. All

observed bands are shown to match ab initio values reasonably well.

The dark relaxation rate of cis-HPD to trans-HPD has been found to be 1.54x103 s"1

at 11K and 2.03x10"' s" at 20K. A lower limit for the activation energy of this process

has been established as approximately 0.06 k/mol. In addition, the difference in standard

free energies of the two isomers has been determined to be 0.4 kJ/mol. Both of these

values are much smaller than theoretical estimates. This finding may be of importance for

other chemical reactions which occur in the interstellar medium. Indeed, the mechanism

for propynal formation in these experiments could be a viable method of propynal

formation in interstellar dust clouds.














CHAPTER 3
THE C3H20 POTENTIAL SURFACE



Introduction



Over one hundred molecular species exist in the interstellar medium.71 Two of

these, propynal and C30, have been observed in approximately equal abundances in the

molecular cloud TMC-1.65,67,72 The mechanisms by which these species are formed has

been the topic of widespread interest. Herbst, Smith, and Adams postulated66 that they

are formed via ion-molecule reactions involving the radiative association processes:


C2H2 + CO H2C3O' + hv

C2H3' + CO H3C3O' + hv


followed by the dissociative electron recombination processes:


H2C3O+ + e" C30 + H2

C30 + H2

H3C30+ + e" H2C30 + H2 + H

H2C0 + H








54

Alternatively, Brown and coworkers have proposed65 that the most important pathway for

C30 production in "young" interstellar clouds (-106 years) is reaction of atomic oxygen

with C4:

O+C4 C30+C

C3 +CO


These suggestions have now been investigated experimentally73 and theoretically74

by researchers in New Zealand. Petrie, Bettens, Freeman, and McEwan measured73 rate

coefficients and product distributions for the gas phase reaction ofHInC3O' (n = 0-3) and

C302+ with a number of neutral species including H20 using the selected ion flow tube

(SIFT) technique. They concluded that HnC3O' ions were notably unreactive toward the

major known neutral constituents of interstellar clouds. Very recently, Maclagan,

McEwan, and Scott reported74 ab initio G2 level calculations and showed that the

reaction of C2H3 and CO yields C2-protonated propadienone rather than O-protonated

propynal. They concluded that while this species could form neutral propadieneone via

dissociative electron recombination, propynal could not be formed in a similar manner.

However, the amount of internal energy available after electron recombination should be

sufficient to make other unimolecular dissociation channels available.

In this study a different mechanism, which is initiated by the formation of a complex

between the carbon cluster, C3, and water, is examined. In an important study Ortman,

Hauge, Margrave, and Kafafi were the first to report37 evidence for a complex between C3

and H20 in an argon matrix at 12K. When photolyzed with radiation of X > 400nm, the










complex was shown to rearrange to C30, propynal, and a third compound postulated to

be hydroxyethynylcarbene (HEC, J). Further photolysis with radiation between

2802 X> 360 nm caused the third compound's bands to decrease and those due to

propynal to increase. No new photoproducts were noted. Later, Liu, Zhou, and Pulay

(LZP) reported40 a comprehensive ab initio theoretical study of the expected infrared

spectral bands of HEC and an alternate intermediate, 3-hydroxypropadienylidene (HPD, B

and C). By comparison with the calculated experimental spectra, LZP proposed that B

and not J, was responsible for the intermediate bands.

In Chapter 2 it was shown from 12C/13C isotopic studies that the intermediate was

indeed B.75 It was also shown that further photolysis of B led to propynal. It was

suggested that small carbon clusters such as C3 could complex with H20 on the icy

mantles of small dust grains prevalent in interstellar clouds, and upon photolysis lead to

the formation of propynal. This solid state mechanism is very different from the gas phase

one proposed by Herbst and coworkers" and by Brown and coworkers.6

Given the conclusions reached by the New Zealand researchers on the unlikelihood

of the ion-molecule reaction route to propynal in the interstellar medium and the apparent

ease of formation of propynal and C30 from the photolysis of solid state C3.H20

complexes, this chapter will present the theoretical investigation of the entire C3H20

potential surface. The study in Chapter 2 was limited primarily to the C3.H20 complex

and the HPD intermediates B and C. From that study several questions arose. (1) Given

the photolytic conditions and the number of C3H20 isomers available, why were the HPD

intermediates (B and C), propynal (0) and C30 observed? (2) Why is propynal










apparently not photodepleted? (3) Two other C3H20 isomers, cyclopropenone (R) and

propadienone (S), are known to be very stable. Why are neither seen experimentally?



Theoretical



All calculations were carried out using the GAUSSIAN 92 program package.42 All

structures presented here are in their singlet ground states. Geometries were optimized

initially at the HF/6-31G* level. Then, to gauge the effects of electron correlation, single

point calculations at the QCISD(T)/6-31G* level were run on the HF optimized

geometries. The HF structures were then used as starting points for optimization at the

MP2/6-31G* level. Unless otherwise specified, the discussion in the following sections is

restricted exclusively to the surface at the higher MP2/6-31G* level of theory. Vibrational

frequencies for all HF and MP2 points were performed at their respective levels in order

to characterize the nature of the stationary points. In all cases the intrinsic reaction

coordinate (IRC) utility was employed to confirm that the transition state did indeed

connect to the correct minima. Each minima is connected to one or more other minima

through a transition state in which either one hydrogen atom migrates to an adjacent heavy

atom or a bond between heavy atoms (i.e., C and 0) either forms or breaks. There are,

however, three exceptions to this rule. TS1, which involves carbon inserting into a water

OH bond, and TS14 and TS15 which are concerned with the unimolecular dissociation of

R and S respectively into acetylene and CO. Transition states between conformers were

not investigated.














The structures of all intermediates and stable compounds studied in this work are

shown in Figure 3-1. Table 3-1 lists the total energies, relative energies, and zero-point

energies of each minimum at the HF/6-3 G* and QCISD(T)/6-31G*//HF/6-3 G* levels,

while Table 3-2 gives similar quantities for the corresponding transition states. Tables 3-3

and 3-4 give the equivalent values from the MP2/6-31G* level calculations. Figures 3-2

to 3-4 show these energies schematically as potential surfaces. The HF/6-31G* surface is

complete, incorporating all transition states and isomers (except for G, which is predicted

to be unstable at all levels of theory). The MP2 optimized minima are shown in Figure

3-5. Several of the minima and transition states studied were found to be unstable at the

MP2/6-31G* level of theory. In each case they followed a CCC ring cyclization path,

optimizing to another structure. Both isomers E and F optimized to H, while G optimized

to I. And, the very stable cyclopropenone molecule (R) was formed from the unstable

isomer P.

The transition state structures between the minima are indicative of the probable

mechanism ofinterconversion. Figure 3-6 shows the geometries of the MP2 optimized

transition states. All transition states represent adiabatic barriers which are the minimum

energies necessary to promote isomerization. Five transition states-TS4, TS5, TS7, TS12

and TS13-found at the HF level could not be located at the MP2 level, and are presumed

not to exist. This has serious consequences for the shape and content of the potential

surface. The absence of these transition states eliminates isomers H and N from the MP2










A
/H
C=C=C----H-0O


H
/ H
D) O

C-C


H-0
E
c= c


H



C C --H


C H
10//
L HC-C






H-CcC-C


H
B
/ 0

H


F H
C /C--
c= c


O


M C-C



1 /

H



H
H C /
C=C
P a


C
/0--H


H


G
C-0
c^/
c=cH


J H-CC-Co



K H-C=C-C

H
O".... C
N c=c
H H


? /H
H/
H NC
\0


0
R II
C

H -C CH


H S
C =C=C=0
H


Figure 3-1. Unoptimized structures of C3H20 isomers studied.











Table 3-1. Total energies (hartrees), zero-point corrected relative energies' (kcal/mol),
and zero-point energies (kcal/mol) of all minima calculated at the HF/6-31G* level.
Total energy (hartrees) calculated at the QCISD(T)/6-31G*//HF/6-31G* level.
HF/6-31G* calculated zero-point corrections scaled by 0.90.


Relative Relative
Isomer Sym HF/6-31G* Energy QCISD(T)/6-31G* Energy Z.P.E
III I I


-189.369 60
-189.465 66
-189.456 91
-189.406 15
-189.397 22
-189.41128


-189.485 80
-189.482 23
-189.465 45
-189.459 86

-189.486 53
-189.483 75
-189.438 69
-189.541 11
-189.495 84

-189.490 72

-189.533 98
-189.536 25


0.0

-55.9
-50.6

-20.3
-14.8

-23.4


-69.3
-66.2
-56.3
-53.0

-70.5
-68.7

-39.8
-103.5
-76.1
-72.9
-98.8
-101.0


-189.946 83

-190.044 28
-190.036 25
-189.972 95
-189.948 54
-189.973 27


-190.055 43
-190.051 95
-190.041 94
-190.037 74

-190.056 61
-190.053 12
-190.025 56
-190.12307

-190.066 06

-190.067 85
-190.11414
-190.122 46


0.0

-56.9
-52.0
-13.8
-11.1
-13.9


-64.5
-61.5
-55.9
-53.4

-66.0
-63.8
-45.9
-106.5

-71.7
-72.8
-100.7
-106.6


21.1

25.9
25.7
24.0
23.9

24.1


26.1
26.0
25.3
25.1

24.3
24.3

25.0
25.7

24.5
24.5

25.9
25.1


CO C.v -112.737 88 -113.036 12 3.5

C2H2 Dh, -76.817 83 -77.091 83 18.5


a Energy relative to the C3*H20 complex. b Not predicted to exist at the HF/6-31G*
level.












Table 3-2. Total energies (hartrees), zero-point corrected relative energies (kcal/mol),
and zero-point energies (kcal/mol) of each transition state at the HF/6-3 1G* level. Total
energy (hartrees) calculated at the QCISD(T)/6-31G*//HF/6-31G* level.
HF/6-31 G*calculated zero-point corrections scaled by 0.90.



Relative Relative
Sym HF/6-31G* Energy QCISD(T)/6-31G* Energy Z.P.E.
7 I


-189.285 27

-189.31851

-189.395 85

-189.326 29

-189.387 59

-189.461 04

-189.37731

-189.37995

-189.416 24

-189.491 96

-189.394 43

-189.435 81

-189.386 35

-189.459 18

-189.489 78


53.8

32.0

-15.2

28.2

-8.9

-56.6

-2.6

-6.2

-29.3

-75.2

-15.6

-38.6

-9.5

-54.9

-73.9


-189.892 84

-189.91204

-189.966 93

-189.916 97

-189.97607

-190.044 01

-189.964 20

-189.983 08

-190.006 30

-190.074 19

-189.988 37

-190.025 64

-189.95022

-190.045 62

-190.065 96


34.7

21.7

-11.4

19.7

-16.0

-60.2

-8.7

-22.4

-37.3

-77.3

-26.1

-46.4

-1.1

-60.7

-73.2


20.1

21.3

22.4

20.0

23.7

22.0

23.5

21.4

21.0

24.0

21.2

24.4

22.2

22.5

22.7


a Energy relative to the C3*H20 complex.


TS1

TS2

TS3

TS4

TS5

TS6

TS7

TS8

TS9

TS10

TS11

TS12

TS13

TS14

TS15


C,

C,

C,

C,

C,

C,

C,

C,

C,

C,

C,

C,

C,

C,

C,


- I I I I I i. -










Table 3-3. Total energies (hartrees), zero-point corrected relative energies' (kcal/mol),
and zero-point energies (kcal/mol) of minima calculated at the MP2/6-31G*level.
MP2/6-31G* calculated zero-point corrections scaled by 0.95.


Relative
Isomer MP2/6-31G* Energy Z.P.E.

A -189.895 21 0.0 19.9
B -189.997 78 -60.2 24.2
C -189.988 86 -54.7 24.1
D -189.913 79 -9.4 22.3
Eb
Fb
Gc
H -190.014 95 -70.9 24.3
I -190.011 32 -68.7 24.2
J -189.997 71 -64.3 23.1
K -189.993 01 -58.6 22.7
L -190.004 76 -66.4 22.3
M -190.001 51 -64.5 22.2
N -189.983 80 -52.2 23.5
O -190.086 63 -117.1 23.1
pd

Q -190.024 58 -78.2 23.1
R -190.077 69 -110.7 23.9
S -190.08631 -116.2 23.8

C2H2 -77.06679 16.1
CO -113.021 21 3.0

a Energy relative to the C3*H20 complex. b Not predicted to exist.
Reverts to structure H. c Not predicted to exist. Reverts to structure I.
d Not predicted to exist. Reverts to structure R.











Table 3-4. Total energies (hartrees), zero-point corrected relative energies
(kcal/mol), and zero-point energies (kcal/mol) for all transition states found
at the MP2/6-31G* level. MP2/6-31G* calculated zero-point corrections
scaled by 0.95.



Relative
MP2/6-31 G* Energy Z.P.E.


-189.848 43

-189.858 92

-189.905 82






-190.001 45



-189.946 22

-189.953 58

-190.021 03

-189.957 66






-190.005 41

-190.009 10


30.7

23.3

-5.5






-66.0



-32.5

-36.6

-76.3

-39.4






-67.3

-71.4


h h


18.4

19.3

21.0






20.5



19.4

19.8

22.6

19.7






21.8

19.9


a Energies relative to the C3*H20 complex.
the MP2/6-31G* level.


b Stationary point not found at


TS1

TS2

TS3

TS4b

TS5b

TS6

TS7b

TS8

TS9

TS10

TS11

TS12b

TS13b

TS14

TS15










A H
TSI >
C3*H20 C=C-c
*H20 (53.8)
(0.0) H


TS2
(32.0)


(-23.4) Hco/
c=<^


TS5
H j /(-8.9)
(H -
(-69.3) c H
^..-IH.


(-2.6)
TS7


TS4
(28.2)


TS3
(-15.2)


H-C C-C-H
(-56.3)


TS8
(-6.2)


F


N
(-39.8)


TS12
---386)
(-38.6)


H H
H *


TS13
(-9.5)


L
(-70.5)


H -O





TS6
(-56.6)






H-C13C-)

(-103.5)


R
(-98.8)


TS14
(-54.9)


TS10
(-75.2)


TS11
(-15.6)


"<,

(-76.1)
P


Figure 3-2. The HF/6-31G* calculated potential surface. Only lowest energy isomers are shown.
Zero-point corrected energies (kcal/mol) relative to A in parentheses.


B
(-55.9)


(-115.9)
C2H2
+
CO


TS15
(-73.9)


(-101.0)


TS9
(-29.3)










A H
TS1 >
C3"H20 C=C=C
C(34.7)c
/n n\ H


TS2
(21.7)


F H
(-13.9) C_/
c=C,


TS5
H (-16.0)
H ,"
(-64.5)


(-8.7)
TS7


(-55.9)


L
(-66.0)


TS4
(19.7)-
(19.7)


TS3
(-11.4)


TS12
(-46.4)


TS6
(-60.2)


TS8
(-22.4)


TS9
(-37.3)


H-C=C-

(-106.9)
O


N
(-45.9)


H H


TS13
(-1.1)


R
(-100.7)

I TS14
(-60.7)
vH


TS10
/ (-77.3)
/i


H(-72
(-72.8)


(-112.8)
C2H2
+
CO


TS15
(-73.2)


TSII
[ -------:C- 0
(-26.1)
(-106.6)
S


Figure 3-3. The QCISD(T)/6-31G*//HF/6-31G* calculated potential surface. Only lowest energy isomers are shown.
Zero-point corrected energies (kcal/mol) relative to A in parentheses.


B
(-56.9)


%v. )


c /H
-c









A H
TS1
C3*Hz20 c=c=
(30.7)
(0.0) H


TS2
(23.3)


(-110.7)


K TS14

HH (-67.3)


TS10
(-76.3)

V S1


TS11
(-39.4) /
(-116.7)


Figure 3-4. The MP2/6-31G* calculated potential surface. Only lowest energy isomers are shown.
Zero-point corrected energies (kcal/mol) relative to A in parentheses.


B
(-60.2)


D ,
(-9.4) o

C=c/C


L
(-66.4)
C H

0~


TS6
(-66.0)


TS3
(-5.5)


TS8
(-32.5)


(-120.9)


C2H2

+
CO


TS15
(-71.4)


(-64.3)
(-64.3)


H-1=C-C

(-117.1)


TS9
(-36.6)-
(-36.6)


.<
H
S/


(-78.2)














A (Cs)
H
179.0 174.1 103.8 0968

-1 -- -7 171.81
171.8 .91


0.976
C(Cs) O H
122.1 1/09.8
1.279 1.3 1.334
C Ci C
173.7 122.3109
H


H (Cs)


H
\0.982
B (Cs) 107.8(o

124.7/1.328
1.279 1.3497
C C C
179.2 1246 1.089
H


H
0.977
D (CI) 0 107.8
1.328
129.3
111.3 1.503
1.31
C110.5 C
110.5 -. 1.083


108.9

152.0
1.334


1.329


175.2


J (Cs)


I (CS)


169.3
H 1.067 1.226 1.443
111.5 1.322
176.2 m
113.5
K (Cs) 0.987
H


Figure 3-5. Optimized structures of all minima found at the
MP2/6-31 G* level of theory.




















L (Cs)


M (Cs)


1.518
C
61.9
93. 1
1.356 101.8 1.414
N(C1) C C
1.391 1.079
1.09207
131.9 134.4 H
n


Q (Cs)


o (Cs)


123.2


178.9


169.61


1.214
151.9
1.438
56.1
68.1 153.9
1084 C 5
144.1.08402 H354
H 144.2 H


R (C2v)


S (Cs)


Figure 3-5 continued. Optimized structures of all minima found at the
MP2/6-3 G* level of theory










H
0TS1 984
TS1 03 107.2
C-6.7," 1.273
188.8 152.1 1.207



TS3 1.317

C-C ..115 4\0.988
'.br125. 8
".9.3r H
1.828' 1.119
H




H
TS8
175.4 11.325 .
_65. 11 1.175
169.8 1433 -275-H


TS10


O
\ 1.179

TS14 151.4C


TS2 H
208.1 111.1 0.988
C1.298 1.3941221.9 ss
,53.5 1.316
HI '1.417



H
1.196/ '1.410
'169 122.8 1.224
c 1.45 -C
1.265
176.8 114.1 1.103
TS6 H



H TS9




H
114.0 ,
1.092 174.7










1.372 H 1.223 T
116.9 61 1 .3
1.357% 1 1.362



H










H / 1. 1 8
1429.85 1407.7 2.305










1.08 .0
1.372%% 1.223 TS11



C 1.150



1. 129.8 107.7 2.305
1TS15
1.092 111.5
H


Figure 3-6. Optimized structures of all transition structures
found at the MP2/6-3 1G* level of theory.











surface, but more importantly it forces the production of cyclopropenone (R) and

propadienone (S) to proceed through propynal (0). And, if propynal is indeed a trap, R

and S will not be seen experimentally.



Discussion



An understanding of the photochemical transformation of the C3*H20 complex to

propynal (and beyond) is greatly aided by consideration of the calculated MP2 potential

surface. The C3.H20 complex (A) has been studied experimentally37'75 and theoretically in

detail previously.40'75 There exists a large barrier (TS1) to the formation of the HPD

intermediates B and C which is calculated to be 52 kcal/mol at the HF level, 33.8 kcal/mol

at the QCISD(T) level, and 29.3 kcal/mol at the MP2 level. Clearly, this barrier prevents

the immediate reaction of C3 and water, as observed for atomic carbon and water.38 There

is a large discrepancy between the calculated barrier for TS 1 and the photolysis

wavelength (X-400 nm) necessary to promote reaction. This can be explained by the fact

that going from A to B and C proceeds through an excited electronic state. In fact, there

is no reaction until the 'I1, 'eg electronic transition in C3 at 405 nm is excited. The

trans or cis-HPD intermediates have been observed experimentally75 which indicates that

there are barriers sufficiently large to preclude their immediate photochemical destruction.

Ortman et al. showed37 that radiation between 320 and 380 nm does not deplete the

intermediate B, but radiation between 280 and 360 nm does. Thus, the energy necessary

to destroy HPD must lie between 280 to 320 nm (89 and 102 kcal/mol). The former limit











matches well the MP2 barrier of 87.0 kcal/mol from B to the transition state TS2.

Radiation with hX400 nm (71.5 kcal/mol) used to photolyze the C3,H20 complex to B

and C, is not sufficiently energetic to continue photoconversion over TS2. However, light

in the 280-320 nm range is. This would explain why the HPD isomers (B, C) are seen

experimentally.

The computed surface from B through intermediates D and J to propynal (0)

involves three consecutive H atom migrations along the CCCO framework. Transition

state TS2 shows that the migration involves formation of a three-membered ring with H

bridging two adjacent carbons in the chain. TS3 involves a similar structure as the H

migrates further down the chain. Finally, TS8, leading from J to O, involves the migration

of the hydroxy hydrogen to the adjacent carbon. Once photoconversion from B or C to D

occurs, the remaining barriers to propynal formation are relatively small (30 to 40 kcallmol

range) and easily surmounted by photons in the 280-320 nm range.

Propynal is predicted to be the global minimum of the CHO20 surface at the MP2

level of theory (117.1 kcal/mol more stable than the C3.H20 complex). It is pertinent to

ask whether propynal can act as a photochemical 'dead end' or if it is possible to deplete it

during photolysis to produce other photoproducts? The three calculated barriers which

deplete propynal (TS6, TS8, and TS9) are in the 80 to 90 kcal/mol range. Radiation with

X 320 nm (89 kcal/mol) is sufficient to scale all three barriers. Thus, propynal is

predicted to be depleted slowly by continuous irradiation at 320 nm or below, and this

depletion is in fact seen experimentally videe infra).










The destruction ofpropynal implies that one or more of the C3H20 isomers,

especially R (cyclopropenone) or S (propadienone), could be formed. Propynal (0),

cyclopropenone (R), and propadienone (S) are the most stable C3H20 isomers. A few

C3H20 isomers have been investigated by Radom and coworkers.7678 The order of

stability found by Komornicki, Dykstra, Vincent, and Radom78 is also found here. At the

HF level propynal is predicted to be more stable than propadienone by 3.0 kcal/mol and

cyclopropenone by 4.5 kcal/mol. At the MP2 level, the order remains the same but with

propynal now predicted to be more stable than propadienone by just 0.2 kcal/mol and

cyclopropenone by 5.6 kcal/mol. The relative order of these three isomers found by

Maclagan, McEwan, and Scott74 at the very sophisticated G2 level finds propadienone (S)

to be more stable than propynal (0) by 1.6 kcal/mol and cyclopropenone (R) by 8.4

kcal/mol.

The barriers to formation of the isomers propadienone and cyclopropadienone from

Q and the subsequent barriers to dissociation into acetylene and CO are smaller than the

photolysis energy used experimentally. The MP2 barrier from Q to propadienone (S) is

42 kcal/mol via TSl 1, while the barrier from Q to cyclopropenone (R) is only 2.3

kcal/mol via TS10. From R to C2H2 and CO requires 45.4 kcal/mol energy input, whereas

the decomposition of S via TS15 requires 46.6 kcal/mol. The latter pathway would yield

vinylidene (H2CC) and CO initially, but the vinylidene to acetylene rearrangement barrier



t An upper bound to the vinylidene zero point level of 15 525 cm'1 above acetylene has
been reported by Chen, Y.; Jonas, D. M.; Kinsey, J. J. L.; Field, R. W. J. Chem. Phys. 91,
3976, 1989; but the isomerization barrier between the two isomers has yet to be measured.










is calculated to require only 8.6 kcal/mol of activation energy.79'8, These results are

consistent with previously reported observations that the photofragmentation of

propadienone leads to acetylene and CO.1"83

From the presently computed surface the expectation is that, once formed, propynal

may be depleted slowly by crossing one or more of the transition states TS3 to 13 upon

photolysis. It appears likely that R and S should be formed and subsequently decomposed

to form acetylene and CO. Searching for evidence on the validity of this mechanism,

previous photochemical infrared data for evidence of CO and/or acetylene were re-

examined. Ortman et al. reported37 IR bands for both, but stated that CO was present in

their experiments as a trace gas and that acetylene was formed from the reaction of C, and

residual H2 from the cracking of pump oil on hot surfaces or outgassing of the graphite

sample. Our results, given in Figure 3-7, show that neither CO (2138 cm'") nor acetylene

(736 cm'1) bands are present in our unphotolyzed sample of Cn in Ar (with 0.15% H20).

However, after photolysis with a water filter, bands due to CO, acetylene, and propynal

(2108 cm-') are present (cf Figure 3-8). In addition, new unassigned bands at 2148.7,

2145.9, and 743.9 cm'1 are also seen. In an identical experiment with no water filter and

longer photolysis times, no propynal peak was found, but peaks due to acetylene and CO

are present. The conclusion arrived at from this evidence is that CO and acetylene are the

final products of the photolysis of the C3*H20 complex, and that propynal is not a

photochemical 'dead-end'. Propynal itself is slowly depleted by photolysis eventually

yielding CO and acetylene. Evidence for the existence of R and S has not been found in

either the matrix experiments or to date in the interstellar medium. To aid in their










identification, calculated rotational constants and IR frequencies at the MP2 level are

given in Tables 3-5 and 3-6, respectively, for all the C3H20 isomers studied here.

The major species discussed here-propynal, acetylene, CO, C3, and H20-are all

constituents of the interstellar medium.7 It is hypothesized that propynal may be formed

from the photolysis of the C3*H20 complex. This complex could be present in the icy

mantles of small dust grains in the interstellar medium. Other molecules are suspected to

originate in a similar manner on dust grains.'"86 There is evidence for C3 in the mantles of

dust grains near the protostar W33A. Tielens and Allamandola have observed"83 a band

at 2043 cm"1 which is close to the C3/argon matrix value of 2039 cm''. These workers had

tentatively assigned this band to OCS and another at 2165 cm'' to an 'XCN' species. This

2165 cm"' band is very close to the Cs/argon matrix band at 2164 cm'1. These (and

possibly other) Cn species could be accreted from matter lost from carbon rich stars. It is

speculated here that, especially on small grains where exothermic reactions can raise the

grain temperature many degrees, photolysis of the C3.H20 complex in icy layers of the

grain mantle may lead to the eventual expulsion of propynal from the grain into the

interstellar medium. The pronounced relative stability of propynal could make this a

viable process on small grains where the grain diameter is -10 gm or less. Long duration

photolysis of the C3.H20 complex in deeper-lying layers could result in further

photorearrangement to the acetylene and CO products, with eventual release to the

interstellar medium.



















o .035-
4) I


< .03-


Before photolysis
.025



2250 2200 2150 2100 2050

Wavenumber (cm-')
Figure 3-7. Portion of the 12K matrix infrared spectrum of 12CAr
before and after photolysis (X > 390 nm) with a water filter present.















After photolysis
without filter


0 .05


< .04


.03

Before photolysis
.02 -
2240 2220 2200 2180 2160 2140
Wavenumber (cm"')
Figure 3-8. Portion of the 12K matrix infrared specrtum of 2CnAr
before and after photolysis (X > 390 nm) without a water filter.










Summary



The potential surface for the rearrangement of nineteen C3H20 isomers has been

calculated at the HF/6-31G*, QCISD(T)/6-31G*//HF/6-31G*, and MP2/6-31G* levels of

theory. (The portion of Figure 1-4 that is represented by the potential surface in this

chapter is shown in Figure 3-9). The relative stabilities and even the existence of certain

minima and transition states is profoundly effected by the level of theory. The

nonexistence of the isomers H and N at the MP2 level forces the production of

cyclopropenone and propadienone to proceed through propynal, a condition not required

at the HF level. The observed photochemical pathway from the C3.H20 complex to

propynal through the hydroxypropadienylidene intermediates during photolysis with X

400 nm is shown to be consistent with the MP2 surface. The mechanism of isomer

interconversion is predicted to involve the migration of hydrogen atoms along a CCCO

backbone via transition states in which one H is bonded to adjacent heavy atoms (C or O).

If propynal is predicted to be photolytically depleted through Q, the isomers

cyclopropenone and propadienone may be formed. However, the barriers leading out

from these molecules are small and further decomposition to acetylene and CO is

predicted. Experimental evidence is presented which supports the validity of this

mechanism. It is proposed that the propynal observed in interstellar space is produced via

the photolysis ifC3*H20 complexes in the ice mantles of dust grains, and not in gas-phase

ion-molecule reactions, as previously suggested.












C3-H20


V
H



C=C=C\

t-HPD


S> 400 nm


0
H-C==C---C
H
Propynal


280 > X > 320 nm


Figure 3-9. The experimental reaction scheme. The large arrow shows
the reaction path described by the potential surface in Chapter 3.


C0
C30


I













Table 3-5. Rotational Constants (GHz) of minima found at the MP2/6-31G*
level. Experimental values in parentheses.



Isomer A B C


510.18

53.18

63.78

24.31

32.3

32.21

67.7

58.86

59.61

28.59

21.81

65.31 (68.04)a

33.55

31.81 (32.05)b

123.79 (149.83)c


1.74

5.07

4.89

7.63

7.65

7.63

4.82

4.92

5.1

6.39

13.03

4.77 (4.826)a

6.32

7.72 (7.825)b

4.41 (4.387)c


1.73

4.63

4.54

5.95

6.18

6.17

4.5

4.54

4.7

5.22

8.78

4.44 (4.500)'

5.32

6.21 (6.281)b

4.26 (4.258)C


SFrom reference 87. b From reference 88. c From reference 89.











Table 3-6. Vibrational frequencies (cm'1), intensities (km/mol), and symmetries of each
C3H20 isomer found at the MP2/6-31G* level.




A 3901.1 (153)a', 3765.5 (117)a', 2185.5 (421)a', 1751.6 (104)a', 1224.8 (2)a',
286.6 (174)a", 257.7 (87)a', 177.7 (0), 176.7 (18), 115.2 (3)a', 43.9 (8), 37.8 (1)

B 3647.8 (94) a', 3236.7 (7) a', 2133.1 (635) a', 1531.9 (184) a', 1384.9 (56) a',
1280.7 (191)a', 1048.9 (84)a', 1021.2 (18)a", 749.5 (148)a", 587.5 (36)a', 195.8
(10), 145.0 (4)

C 3722.5 (171)a', 3167.7 (18)a', 2140.3 (679)a', 1558.9 (59)a', 1329.7 (336)a',
1323.3 (130)a', 1048.8 (1)a', 978.1 (0)a", 608.8 (108)a", 592.7 (8)a', 205.4 (5)a",
172.7 (1)a'

D 3696.3 (154), 3291.0 (28), 1677.0 (128), 1372.6 (62), 1311.1 (232), 1044.1 (95),
945.0 (67), 828.2 (47), 676.8 (126), 395.4 (47), 258.7 (4), 167.0 (7)

H 3682.9 (85)a', 3326.4 (2)a', 1865.7 (151)a', 1462.0 (65)a', 1253.9 (159)a', 1212.8
(31)a', 981.8 (5)a', 873.1 (35)a", 825.4 (11)a', 593.6 (142)a", 481.2 (0)*a", 423.9
(38)a'

I 3707.0 (79)a', 3318.2 (2)a', 1843.0 (186)a', 1471.9 (44)a', 1266.6 (13)a', 1179.0
(185)a', 970.2 (12)a', 865.6 (12)a", 823.6 (6)a', 568.6 (136)a", 483.2 (4)a", 445.9
(0)*a'

J 3692.9 (150)a', 3516.5 (52)a', 2153.1 (53)a', 1392.0 (31)a', 1325.2 (321)a', 932.7
(45)a', 867.9 (80)a", 640.1 (40)a", 622.4 (2)a', 560.8 (49)a', 254.9 (5)a", 219.6 (9)a'

K 3512.7 (60)a', 3505.1 (8)a', 2132.1 (46)a', 1388.1 (7)a', 1343.2 (292)a', 910.2
(10)a', 876.0 (150)a", 638.4 (23)a", 638.2 (27)a', 565.6 (32)a', 239.8 (15)a",
193.3 (7)a'

L 3175.8 (40)a', 3069.0 (57)a', 1742.5 (152)a', 1713.3 (172)a', 1457.8 (4)a', 1093.3
(56)a', 1000.7 (1)a", 901.9 (43)a', 640.4 (45)a", 511.5 (12)a', 178.4 (27)a', 145.8
(22)a"


Table 3-6. Continued.












M 3152.9 (26)a', 3045.1 (97)a', 1781.2 (85)a', 1701.5 (245)a', 1455.2 (18)a', 1027.4
(54)a', 1000.9 (3)a", 944.5 (34)a', 574.9 (30)a", 559.6 (15)a', 175.5 (7)a', 117.7
(0)*a"

N 3354.3 (4), 3192.7 (12), 1468.6 (7), 1307.1 (23), 1281.8 (26), 1178.4 (39), 1017.4
(3), 987.4 (31), 835.0 (38), 761.7 (30), 609.0 (9), 439.6 (20)

0 3513.2 (50)a', 3055.0 (77)a', 2151.4 (63)a', 1723.5 (98)a', 1450.8 (10)a', 1012.8
(0)*a", 982.8 (100)a', 634.9 (32)a', 587.0 (18), 585.7 (43), 255.1 (3)a", 200.1
(4)a'

Q 3273.6 (1)a', 3072.2 (30)a', 2174.0 (521)a', 1404.3 (56)a', 1294.4 (13)a', 1070.7
(11)a', 1041.7 (35)a', 1005.8 (209)a", 670.5 (15)a', 484.1 (35)a', 448.3 (21)a",
242.2 (70)a'

R 3300.7 (1)a,, 3269.9 (0)*b2, 1980.5 (509)a,, 1559.7 (4)a,, 1188.7 (7)b2, 1069.8
(6)a,, 955.4 (0)a2, 870.4 (0)aj, 857.5 (47)b2, 735.4 (62)bi, 528.1 (0)*b2, 457.9
(16)b,

S 3251.2 (10)a', 3163.0 (22)a', 2206.8 (616)a', 1734.5 (6)a', 1529.6 (1)a', 1108.6
(22)a', 1021.8 (27)a", 960.5 (1)a', 710.5 (8)a", 514.8 (10)a', 278.7 (4)a", 204.9
(17)a'


*Calculated intensity < 1, rounded to zero.














CHAPTER 4
A PHOTOCHEMICAL MECHANISM FOR C30 FORMATION



Introduction



The C3*H20 complex was originally identified in 1990 by Ortman et al.37 Since that

time this laboratory has been interested in the photochemical mechanisms and products

which arise by photolyzing the C3*H20 complex. The experiment itself is simple. Laser

vaporized graphite and argon seeded with ca. 0.75% H20 are codeposited in a matrix.

Bands due to a complex between C3 and H20 are seen: 3598 cm'1 (OH sym. str.), 3712.2

cm"1 (OH asym. str.), 2052.3 cm-1 (asym. CCC str.), 1593.4 cm'1 (HOH bend), and 1214.3

cmn' (sym. CC str.). An electronic band at 405.4 nm arising from the C3 moiety is also

seen.75 The C3.H20 complex is then subjected to a two step photolysis procedure. The

first step photolyses the complex with radiation which spans the aforementioned 405.4 nm

electronic band. After photolysis new bands due to C30, propynal (E), and 3-

hydroxypropadienylidene (HPD) (B and C) appear while the C3*H20 complex bands

decrease. The second photolysis involves irradiating the matrix with light of wavelengths

280 < X < 320 nm. HPD bands are observed to decrease and bands due to propynal, CO,

and acetylene to increase. (The experiment is summarized schematically in Figure 1-4,

while the reaction route modeled in this chapter is shown in Figure 4-1).










C3-H20


H


HC=C
t-HPD


X > 400 nm


V

H-C-C-C
H
Propynal


280 > X > 320 nm


Figure 4-1. The experimental reaction scheme. The large arrow shows
the reaction path represented by the potential surface in Chapter 4.


C30


i










In Chapter 3 ab initio calculations were employed to characterize the mechanism

leading from the C3H20O complex to HPD to propynal and finally to CO and acetylene.90

The model involves initial insertion of the C3 molecule into the OH bond of water. The

remainder of the surface involves a series of hydrogen migrations followed by dissociation

of two isomers into CO and acetylene. While this mechanism describes the experimental

observations behavior quite well, several questions remain about the reaction scheme in

Figure 4-1. First, even though it was shown how propynal could form from HPD during

the second photolysis step, it was not at all clear how propynal was being formed in the

first photolysis step with lower photon energies. Second, how was C30 produced? Did it

arise from the C3*H20 complex? Third, why does C30 form only during the first

photolysis step and not during the second, higher energy step? Presented in this chapter

is a calculated potential surface which explains the observed one step photolysis of

C3H20O to propynal and C30.



Theoretical



All calculations were carried out at the MP2/6-31G* level using the GAUSSIAN 94

program package.91 The inclusion of electron correlation has been shown to be extremely

important in obtaining correct geometries76 and can have a profound effect on the

quantitative aspects and even the content of potential energy surfaces.90 Unless otherwise

stated, all minima and transition states were optimized in their ground electronic states

constrained to the symmetries given in Tables 4-1 and 4-2. Optimized structures of all








84

minima and transition states are shown in Figures 4-2 and 4-3, respectively. The surface

calculated in this study represents the minimum energies which must be supplied to

promote reaction. Vibrational frequencies were calculated in order to determine the

nature of the stationary point and to provide thermochemical data necessary for

corrections to reaction energies. Finally, internal reaction coordinate (IRC) calculations

were performed to confirm which minima connected to a particular transition state. The

potential surface shown in Figure 4-4 represents points along the H2C30 potential surface.

Those molecules that have one or no hydrogens have the MP2/6-31G* H or H2 energy

added in (at the infinite separation limit) so that the H2C30 total energy is maintained.



Results



Whatever surface is modeled it must ultimately lead to the creation of C30 and

propynal. We have shown previously that the precursor to propynal is the C3.H20

complex, whereas C30 may have several unique routes of formation. These include:



C3*H20 + hv C30 + 2H or H2 (1)

C2 + CO + hv C30 (2)

C2 + H20 C2*H20 (3a)

C2*H20 + hv C20 + 2H or H2 (3b)

C +CO + hv C20 (4a)

C + C20 + hv C30 (4b)










Table 4-1. Total energies (hartrees), zero-point corrected relative energies (kcal/mol),
spin squared eigenvalues, and zero-point energies (kcal/mol) of minima found at the
MP2/6-31G* level of theory.

Relative
Molecule Sym MP2/6-31 G* Energy Z.P.E.
A C, -189.895 21 0.0 0.0 19.9

B Cs -189.997 78 -60.3 0.0 24.2

C C, -189.98886 -54.8 0.0 24.1

D C, -189.997 71 -61.3 0.0 23.1

E C, -190.086 63 -117.1 0.0 23.1

F Cs -189.374 38 13.3a 1.167 18.9

G C, -189.343 49 30.9a 0.922 17.1

C (3p) -37.733 83 2.005 ---

C2 ('Eg) Dh -75.696 17 0.0 2.7

C2 (3n) D,h -75.688 20 2.005 2.7

CO (1) C._ -113.02121 0.0 3.0

C20 (E-) C.v -150.83492 2.059 5.6

C3O ('+) Cv -188.860 03 0.0 9.7

C30 (3) C, -188.740 66 2.521 10.2

C402 (3 ) D-h -301.89613 2.151 15.8
H -0.498 23 0.750 ---

H2 D h -1.144 14 0.0 6.5


a Relative energy composed of stationary point and hydrogen atom total energies with
zero point correction.




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EH2AGMZ0O_JLQK1O INGEST_TIME 2013-10-24T19:35:24Z PACKAGE AA00017667_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES