THE ROLE OF POLYCYCLIC AROMATIC HYDROCARBONS IN THE STUDY OF FULLERENE FORMATION

Item Type text; Dissertation-Reproduction (electronic)

Authors Zumwalt, Michael Crane

Publisher The University of Arizona.

Rights Copyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author.

Download date 29/09/2021 18:49:04

Link to Item http://hdl.handle.net/10150/187566 INFORMATION TO USERS

This manuscript ,has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer.

The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrougb, substandard margins, and improper alignment can adversely affect reproduction.

In the unlikely. event that the author did not send UMI a -;omplete manuscript and there are missing pages, these will be noted Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.

Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand comer and continuing from left to right in equal sections with small overlaps. Each original is also photograpbed in one exposure and is included in reduced form at the back of the book.

Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white pbotographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.

UMI A Bell & Howell Information Company 300 North Zeeb Road. Ann Arbor. MI48106·1346 USA 313/761-4700 800:521·0600

THE ROLE OF POLYCYCLIC AROMATIC HYDROCARBONS

IN THE STUDY OF FULLERENE FORMATION

by

Michael Crane Zumwalt

A Dissertation Submitted to the Faculty of the

DEPARTMENT OF PHYSICS

In Partial Fulfillment of the Requirements For the degree of

DOCTOR OF PHILOSOPHY

In the Graduate College

THE UNIVERSITY OF ARIZONA

1995 OMI Number: 9620438

UMI Microfonn 9620438 Copyright 1996, by UMI Company. All rights reserved.

This microfonn edition is protected against unauthorized copying under Title 17, United States Code.

UMI 300 North Zeeb Road Ann Arbor, MI 48103 2

THE UNIVERSITY OF ARIZONA GRADUATE COLLEGE

As members of the Final Examination Committee, we certify that we have Michael Crane Zumwalt read the dissertation prepared by------entitled The Role of Polycyclic Aromatic Hydrocarcons in the Study of Fullerene Formation

and recommend that it be accepted as fulfilling the dissertation requirement for the Degree of Doctor of Philosophy/Physics

M.B. Denton Date

R.H.?X~ Parmenter Date

Date

Date

Final approval and acceptance of this dissertation is contingent upon the candidate's submission of the final copy of the dissertation to the Graduate College.

I hereby certify that I have read this dissertation prepared under my direction and recommend that it be accepted as fulfilling the dissertation requirement.

, Di~~ Date Donald R. Huffman 3

STATEMENT BY AUTHOR

This dissertation has been submitted in partial fulfillment of requirements for an advanced degree at The University of Arizona and is deposited in the University Library to be made available to borrowers under rules of the library.

Brief quotations from this dissertation are allowable without special permission, provided that accurate acknowledgment of source is made. Requests for permission for extended quotation from or reproduction of this manuscript in whole or in part may be granted by the head of the major department or the Dean of the Graduate College when in his or her judgment the proposed use of the material is in the interests of scholarship. In all other instances, however, permission must be obtained from the author. 4

ACKNOWLEDGEMENTS

I greatly appreciate all of the people in both the Department of Physics and the Department of Chemistry at the University of Arizona who assisted me in this work one way or another. A special thanks to Professors Donald R. Huffman (physics) and M. Bonner Denton (chemistry). I am very privileged to have worked for both of these men. As my friend Lowell Lamb once told me, "Five minutes with either one of them is like three weeks in the library." I also thank them for their Willingness to let me go wild in their laboratories.

Two gentlemen I have the great privilege to know during my time here at the UofA, are Frank A. Tinker, PhD., and Lowell D. Lamb, PhD. Their friendship is beyond value. I have greatly benefited from their mature advice through the years.

F or the last eight years that I have known my wife, she has been a constant source of love and encouragement. God placed her in my life and led me to Him through her. This work is the result of the combined effort between Marie A. Zumwalt and myself. Our marriage has been blessed with three beautiful boys named Jonathan, Christopher, and Benjamin. Having a family has provided me with a real sense of purpose not only in this work, but in all of my studies at the UofA.

Most of all I thank my Lord and Savior, Jesus Christ. Without Him life can have no meaning. 5

TABLE OF CONTENTS

LIST OF FIGURES ...... 6 LIST OF TABLES...... 10 ABSTRACT...... 11 1 INTRODUCTION...... 12 2 THEORIES of FULLERENE FORMATION...... 20 2.1 "Pentagon Road" ...... 20 2.2 "Fullerene Road" ...... 22 2.3 Combination ofNaphthalenoctyJ Units ...... 23 2.4 "Ring Stacking"...... 24 2.5 Annealing of Large Carbon Rings...... 24 3 EXPERIMENT...... 26 3.1 Pyrolysis ofPAH's...... 26 3.2 Carbon-Arc Vaporization ...... 27 3.3 Techniques for Analysis...... 30 4 RESULTS - Pyrolytic Formation...... 32 5 RESULTS - Hydrogen Poisoning...... 40

5.1 Yield of C60 ••.•....•.••.•..•.•..•.....•.•.••••.•..• 40 5.2 Mass Spectrometry...... 41 5.3 Infrared Absorption...... 47 5.4 UltravioletlVisible Absorption...... 49 5.5 High-Performance Liquid-Chromatography...... 53 5.6 Gas Analysis ...... 54 5.7 Soot Analysis...... 55 5.8 A Closer Look ...... 59 6 DISCUSSION and CONCLUSIONS ...... 72 APPENDIX A- Anomolous Results ...... 77 LIST OF REFERENCES...... 89 6

LIST OF FIGURES

FIGURE 1.1, Extinction due to interstellar dust. Adapted from ref. 1, used by pennission...... 12 FIGURE 1.2, Fullerene synthesis by laser ablation of graphite. A plume of helium is injected across the graphite disk as a laser ablates molecular carbon from its surface. Fullerene molecules are fonned within this helium gas, which carries these newly fonned molecules out of this region and into a mass spectrometer. Adapted from ref. 20, used by pennission...... 13 FIGURE 1.3, Carbon-arc vaporization of graphite for producing fullerenes. Also

shown are two primary methods of C60 and C70 extraction from soot. Adapted from ref. 1, used by permission...... 14

FIGURE 1.4, Reactions ofC6o. Taken from ref. 4, used by pennission...... 16

FIGURE 1.5, Dominant fullerene family includes the primary isomers of C78 and

C84 • ••••••••••••••••••••••••••••••••••••••••••••••••• 17 FIGURE 1.6, Likely pathways offullerene formation? ...... 18 FIGURE 2.1, Number of dangling bonds in compact hexagon-only sheets, compared to the compact molecules that obey the "pentagon rule," as a function of the number of carbon atoms in the sheet. Taken from ref. 20, used by permission...... 21

FIGURE 2.2, Growth of C60 by successive additions of naphthalenoctyl (C IO)' Dotted lines are used to illustrate necessary bonding and consequent curvature to

close and form C60• •••••••••••••••••••••••••••••••••••.••• 23

FIGURE 2.3, Isomers of C40+ are collisionally heated to form a C38 fullerene

following a C2 loss. Taken from ref. 24, used by permission...... 25 FIGURE 3.1, Schematic of tube furnace setup for PAH pyrolysis...... 26 FIGURE 3.2, Carbon-arc chamber used for hydrogen poisoning of fullerene production...... 28 FIGURE 4.1, Molecules pyrolyzed in attempt to form fullerenes. Both hexagon-only (a) and pentagon (b) donors were tried...... 32 7

LIST OF FIGURES - Continued

FIGURE 4.2, Tiling of C60 using (a) six naphthalenoctyls, in black, or (b) twelve pentagons, in black...... 33

FIGURE 4.3, Mass spectrum demonstrating successful formation of C60 by pyrolysis of naphthalene along with evidence of the necessary intermediate molecules. . 38

FIGURE 5.1, Yield of C60 in harvested soot as a function of % H2...... 40 FIGURE 5.2, Percent mass yield of soluble products in soot...... 42 FIGURE 5.3a, Mass spectrum of 0% H2 soot extraction...... 42 FIGURE 5.3b, Mass spectrum of IS% H2 soot extraction...... 43 FIGURE 5.3c, Mass spectrum of SO% H2 soot extraction...... 43 FIGURE 5.4, Log plot of mass peak intensities vs. hydrogen poisoning...... 45 FIGURE 5.5, Sample ion chromatograms for particular rnJz values obtained from electron-impact mass spectrometry of a hydrogen-poisoning toluene-soxhlet extract. Arrow in top chromatogram illustrates method for determining peak intensities that were then normalized and plotted in figure S.4...... 46 FIGURE 5.6a, IR absorption spectrum of 0% H2 soot extract...... 48 FIGURE 5.6b, IR absorption spectrum of IS(I)% H2 soot extract...... 48 FIGURE 5.6c, IR absorption spectrum of SO% H2 soot extract...... 49 FIGURE 5.7, Log plot ofIR peak intensities vs. hydrogen poisoning...... 50 FIGURE 5.8a, UVNis absorption spectrum of 0% H2 soot extract...... 51 FIGURE 5.8b, UVNis absorption spectrum of IS% H2 soot extract...... 51 FIGURE 5.8c, UV Nis absorption spectrum of SO% H2 soot extract...... 52 FIGURE 5.9, Log plot ofUVNis peak intensities vs. hydrogen poisoning. . . .. 52 FIGURE 5.10, HPLC chromatograms ofO, IS, and SO% H2 soot extracts...... 53 FIGURE 5.11, Log plot of HPLC peak intensities vs. hydrogen poisoning. .... 54 FIGURE 5.12a, Raman spectrum of 0% H2 soot...... , 56 FIGURE 5.12b, Raman spectrum of 10% H2 soot...... 56 FIGURE 5.12c, Raman spectrum of IS% H2 soot...... 57 FIGURE 5.12d, Raman spectrum of 20% H2 soot...... 57 8

LIST OF FIGURES - Continued

FIGURE 5.12e, Raman spectrum of 50% H2 soot...... 58 FIGURE 5.12f, Raman spectrum of 80% H2 soot...... 58 FIGURE 5.13, Plot of soot C=C peak areas by Raman analysis as a function of%H2...... 59 FIGURE 5.14, Compact, hexagon-only PAH's believed to exist in hydrogen-poisoned carbon-arc soot. Asterisks denote molecules with published spectra...... 60 FIGURE 5.15, Compact, hexagon-pentagon only PAH's believed to exist in hydrogen­ poisoned carbon-arc soot. Asterisks denote molecules with published spectra. . 61 FIGURE 5.16, Bowl shaped series that follow "pentagon rule." ...... 63 FIGURE 5.17, Electron-impact mass spectrum of a 10% H2 run taken from a series reported earlier[35]...... 64 FIGURE 5.18, IR absorption spectrum of 80% hydrogen-poisoned soot extract. .. 66 FIGURE 5.19, UVNis absorption spectrum of 80% hydrogen-poisoned soot extract. 66 FIGURE 5.20, HPLC peak assignments using known PAH's to determine retention times in the 80% hydrogen-poisoned soot extract. "Peak #'s" were fraction-

collected and analyzed by mass spectrom~try...... 68 FIGURE 5.21, Time-of-flight mass spectra of the four peaks HPLC fraction collected from the 80% hydrogen-poisoned soot extract and labeled in Figure 5.20. ... 70 FIGURE 5.22, Plot of PAH mass vs. HPLC retention time of the 80% hydrogen­ poisoned soot extract using known PAH's and mass spectra of fraction-collected HPLC elution peaks...... 71 FIGURE 6.1, Qualitative sketch of overall picture concerning what is known of soot extracts...... 74 FIGURE A.l, Electron-impact mass spectrum of 0% hydrogen-poisoned soot extract...... 78 FIGURE A.2, Electron-impact mass spectrum of3% hydrogen-poisoned soot extract...... 78 9

LIST OF FIGURES - Continued

FIGURE A.3, Electron-impact mass spectrum of 10% hydrogen-poisoned soot extract. · ...... , 79 FIGURE A.4, Electron-impact mass spectrum of 15(1)% hydrogen-poisoned soot extract...... 79 FIGURE A.5, Electron-impact mass spectrum of 15(11)% hydrogen-poisoned soot extract...... 80 FIGURE A.6, Electron-impact mass spectrum of20% hydrogen-poisoned soot extract...... 80 FIGURE A.7, Electron-impact mass spectrum of50% hydrogen-poisoned soot extract. · ...... , 81 FIGURE A.8, Electron-impact mass spectrum of 80% hydrogen-poisoned soot extract. · ...... 81 FIGURE A.9, Infrared spectrum of 0% hydrogen-poisoned soot extract...... 82 FIGURE A.I0, Infrared spectrum of3% hydrogen-poisoned soot extract...... 82 FIGURE A.n, Infrared spectrum of 10% hydrogen-poisoned soot extract...... 83 FIGURE A.12, Infrared spectrum of 15(1)% hydrogen-poisoned soot extract. 83 FIGURE A.13, Infrared spectrum of 15(11)% hydrogen-poisoned soot extract. 84 FIGURE A.14, Infrared spectrum of20% hydrogen-poisoned soot extract...... 84 FIGURE A.15, Infrared spectrum of 50% hydrogen-poisoned soot extract...... 85 FIGURE A.16, Infrared spectrum of 80% hydrogen-poisoned soot extract...... 85 FIGURE A.17, Illustration of molecule that could be partly responsible for contamination in hydrogen-poisoning work. Replacing the t-Butyl's with hydrocarbon chains and analyzing the molecule by electron-impact mass

spectrometry sh~uld result in a fragmentation pattern witnessed in the 15(1)% and 20% H2 mass spectra...... 86 FIGURE A.18, Log plot of mass spectral peaks of known contaminants vs. hydrogen- poisoning run...... 87 10

LIST OF TABLES

TABLE 4.1, Results of attempts to fonn fullerenes from the pyrolysis ofPAH's. . 34 TABLE 5.1, Extracted mass yields of harvested soot in hydrogen poisoning. ... 41 TABLE 5.2, Calculated nonnalization constants for plotting MS, UV Nis, and HPLC peak, intensities in hydrogen poisoning...... 46 TABLE 5.3, Calculated nonnalization constants for plotting FTIR peak intensities in hydrogen poisoning...... '...... 50 TABLE 5.4, Infrared peaks of known P AB' s...... 65 TABLE 5.5, Ultra-violet/visible absorption peaks of PAB's that have known peaks in the IR...... 67 TABLE A.l, Infrared assignments of peaks unique to the 15(1)% and 20% samples. 87 11

ABSTRACT

Two approaches intended to elucidate the fullerene-formation mechanism are presented. The first of these involves pyrolytic synthesis of fullerenes from hydrocarbon ring structures known as polycyclic aromatic hydrocarbons (PAH's). Following work by

Taylor et al. (Nature 366, 728, 1993), C60 is be made by heating a naphthalene vapor/argon mixture to approximately 1000°C. The use of several precursor P AH' s, including naphthalene, is examined in this work. The second approach involves the intentional poisoning of carbon-arc fullerene production by the addition of hydrogen (H2) to the quenching atmosphere. By adding hydrogen in varying amounts one produces both PAH's and chain molecules, possibly representing interrupted steps of the pathway leading to fullerenes. Various analytical techniques are employed to examine both approaches. It is shown by mass spectrometry' that pyrolytic synthesis is not indicative of the fullerene-formation mechanism of the carbon-arc technique pioneered by Krlitschmer et al. (Nature 347, 354, 1990). In addition to mass spectrometry, Fourier-transform infrared and ultra-violet/visible absorption spectroscopy, high-performance Iiquid­ chromatography, and Raman-scattering spectroscopy are brought to bear in the analysis of the hydrogen-poisoning approach. From the analysis the PAH molecules formed in the hydrogen poisoning of the carbon-arc do not appear to comprise a pathway to fullerene formation. Howev~r, there is evidence indicating that chains, produced as a result of hydrogen contamination of the carbon-arc technique, are related to the formation of fullerene molecules. 12

CHAPTERl

Introduction

The discovery of fullerene molecules began with the investigation of the interstellar dust of galaxies in our universe. Galaxies contain stars, planets, gases, and dust particles. Such particles are believed to be ejected from supernovae and from certain cool stars. During the 1970's and 80's, two independent research teams tried to determine what the dust particles were made of. Both groups were looking at the interstellar extinction curve (Figure 1.1) for clues. The interstellar medium is made up primarily of gases, of which hydrogen is the largest constituent. A small fraction of this medium is made up of dust particles. The interstellar extinction curve represents the starlight which

Diffuse z Interstellar o Bands t; 2:

lKw

°O~~--~----~4~----6~----~8----~~--~~ PHOTON ENERGY (electron volts)

FIGURE 1.1: Extinction due to interstellar dust. Adapted from ref. 1, used by permission. 13

is scattered and absorbed by the dust particles.

There are two important features of Figure 1.1. The first is comprised of several peaks in the visible range known as the Diffuse Interstellar Bands (DIB's). The research team led by Harry Kroto, University of Sussex (England), and Richard Smalley, Rice University, believed these bands to be due to small carbon clusters. In an attempt to produce these clusters, the team employed laser-vaporization of graphite in the presence of an inert atmosphere of helium, as shown in Figure 1.2 [2]. A pulsed plume of helium swept the newly formed clusters into a time-of-flight mass spectrometer (TOFIMS).

~ To ~---•.~ Mass '-...... - Analyzer Je~

Vaporization Integration Laser -. Cup

FIGURE 1.2: Fullerene synthesis by laser ablation of graphite. A plume of helium is injected across the graphite disk as a laser ablates molecular carbon from its surface. Fullerene molecules are formed within this helium gas, which carries these newly formed molecules out of this region and into a mass spectrometer. Adapted from ref. 20, used by permission.

In 1985, Kroto and Smalley varied the experimental conditions over a wide range of helium pressures and laser intensities. By tuning the conditions just right, a mass peak at a mJz value of 720 stood out, implying the presence of a stable molecule. Since the molecule could be made up of only carbon atoms, they proposed a 60-carbon atom, icosohedrally-symmetric, soccerball-shaped molecule, and named it buckminster­ fullerene, after the famous architect of geodesic domes, Buckminster Fuller. For the next five years, further experimentation with this technique provided results consistent with 14 the proposed structure and associated properties, but absolute confinnation of the molecule's existence would had to wait until 1990.

The most prominent feature in Figure 1.1 is the absorption peak at 2200A. Immediately after its disovery in the late 1960's, it was proposed that the 220nm interstellar band was due to graphitic particles. To produce graphitic particles similar to the supposed interstellar particles, Donald Huffman, at the University of Arizona, and Wolfgang Kratschmer, at the Max Planck Institute (Heidelberg), began generating smoke particles of graphitic carbon by vaporizing graphite electrodes in a bell jar filled with helium, as shown in Figure 1.3.

10 vacuum I pump.ad ... bellum lource

r-C60 I c'o ntractiOD! by dissolvlag by lubllm.Uoa

soot C60 I c'o sublimed outoraool ~~~crucible lolueae ( "- brownlsb ~_HWP"' ~~~Uid poured

liquid~ evaponled leaviag C60 I C70 residue

FIGURE 1.3: Carbon-arc vaporization of graphite for producing fullerenes. Also shown are the two primary methods of C60 and C70 extraction from soot. Adapted from ref. 1, used by permission. 15

Between 1983 and 1990, Huffman and Kratschmer varied the experimental conditions over a wide variety of helium pressures and arc intensities[3]. Tuning the conditions just right resulted in the production of large amounts of the buckminsterfullerene, or C60. For the first time, the macroscopic quantities needed to positively confirm geometry and determine properties of C60 were produced. This production technique has subsequently become the standard for making fullerenes. It is important to note that although this is not a combustion process as there is no oxygen present, fullerene soot does resemble chimney soot.

Nearly as amazing as the discovery of C60-containing soot production utilizing a more than 100-year-old carbon-arc technology, was the discovery that C60 could be extracted from the soot because of its solubility in organic solvents like benzene and toluene. This solubility gave birth to fullerene chemistry. Figure 1.4 illustrates types of

C60 derivatives that have been formed as a result. For example, fullerols are water soluble, thus suggesting biological applications. Fulleranes are of interest for hydrogen storage and battery applications. The formation of a particular methanofullerene has even opened the door for fullerenes to the pharmaceutical arena[5].

In addition to solubility, the fullerenes can be extracted from the soot by heat. C60 sublimes at around 350°C, far below the nearly 4000 degree sublimation temperature of amorphous carbon found in the remaining soot. Films of C60 formed from the deposition of molecular C60 led to the discovery that at room temperature, solid C60 has an FCC lattice structure[6], making C60 a new form of solid carbon. The two other forms are are graphite and diamond. This relative ease of sublimation has resulted in potential electronic applications as well as diamond film growth. In addition, doping C60 films with alkali metals ha~ produced superconductors[7]. It turns out that other, all-carbon, cage molecules are formed in both the laser-ablation and carbon-arc techniques, and can 16

FIGURE 1.4: Reactions of C60• Taken from ref. 4, used by permission. be extracted. Mass spectrometry shows that a whole family of fullerene molecules, up to at least Csoo, exists[8). Typically, total fullerene yields were, are, 10-15% of the soot by mass. The six most abundantly formed fullerenes are shown in Figure 1.5.

Even before the chemistry and materials science of fullerenes were born in 1990, the hollow aspect of these molecules was exploited. The cavity of the smallest stable fullerene C60 is large enough to accommodate any element up to and including uranium [9) , which has been done.

By-products of the carbon-arc formation of fullerenes are found on the graphite cathode in the form of nanoparticles. Electron micrographs show that the particles can be multiwalled coaxial tubes[lO). With the proper catalyst, single-walled tubes are also 17 fonned and are found in spider-web deposits within the chamber[ll]. Encapsulating metals within the tubes may lead to applications in nanowire technology. In addition, other potential uses include nanoprobes and nanopipettes.

C70 C76 C60 ~ &' ~.: \t3Z .. ~ G

D2 - C 84 D3 - C 78 C2Y- C 78 @ 8 ~

FIGURE 1.5: Dominant fullerene family includes the primary isomers ofC78 and C84•

This Work

Fullerene synthesis is not limited to the carbon-arc or laser-ablation techniques. Very low yields « ppm) are found in .naturally occurring sources such as the Cretaceous-Tertiary (KT) boundary of the earth's crust[12], and in residues of some lightning strikes[13]. Hydrocarbon-based fullerene synthesis also is possible using benzene/oxygen flames[14], or pyrolysis of naphthalene[15). However, the most common technique for macroscopic production of C60 and the other fullerenes is by the vaporization of graphite in a helium atmosphere. And yet, as easy as the fullerenes and fullerene-related materials are to produce, the details of the process remain a mystery. Several mechanisms have been proposed in an effort to explain the high-yield fonnation of closed-caged carbon clusters out of the chaos of a condensing carbon vapor. Five of these mechanisms were recently critiqued by R. F. Curl[16], and are discussed in Chapter 2. 18

Generally accepted is the assumption that C I, C2, and C3 are the primary molecules vaporized from the graphite rod in the carbon arc, and formed in this vapor are fullerenes at up to 30% yields. Whatever the formation pathway, it must include some form of intermediate molecules. If the intermediate carbon clusters contain the hexagons and pentagons found in fullerenes, then they should resemble the carbon arrangements of a class of both curved and flat molecules called polycyclic aromatic hydrocarbons (PAH's). See Figure 1.6, for example. Of course the intermediate molecules leading up to fullerenes would not contain hydrogen, and they would be very reactive, especially with one another. By deliberately introducing hydrogen into the fullerene-making process, one might hope to form stable PAH's[17]. These could then be studied for clues regarding the pathway to the fullerenes. PAHCsY If? ~ I J Fullerene ? I ? C 1, C2 and C3 's @I _ • / V - I @ c9

FIGURE 1.6: Likely pathway offuJlerene formation?

As it now stands, kinetic-formation modeling of fullerenes is not well understood. There may be one pathway, or there may be several. In any case, there should be intermediate-sized molecules that lie in the pathway between vaporized carbon and the closed-cage, all-carbon molecules known as fullerenes. If it is possible to freeze out these intermediates by hydrogen attachment, isolate them from the soot, and analyze them;then the results may either promote a particular model, or not, or require a whole new model altogether. It is this very reasoning that provides the impetus for this work. 19

As a second approach, it should be possible to fonn fullerenes by starting with PAH's, and by way of pyrolysis, to strip off the hydrogens, and induce them to merge. Following the experimental-setup descriptions in Chapter 3, experimental results of PAH pyrolysis are found in Chapter 4.

In Chapter 5 of this work, results of hydrogen poisoning in the carbon-arc technique are described. Chapter 6 is the Discussion and Conclusions. Appendix A is used for displaying all of the mass and infrared spectra important to this work, and to discuss various anomalous results. 20

CHAPTER 2

Theories of Fullerene Formation

Several theories involving the fonnation pathways of fullerenes exist, of which five are considered here. One assumption underlies any theory that models the fonnation of the fullerene molecules and is based on experimental observations of subliming graphite. Many experiments throughout the 1950's and 60's, including Drowart et al.[IS], for example, showed that when graphite sublimes, the primary constituents are C), C2, and

C3. Therefore, whatever fonnation model turns out to be correct for fullerene fonnation, these three clusters of carbon are precursors.

2.1 "Pentagon Road"

The Pentagon Road idea of fullerene fonnation was initially proposed in 1987 by Heath et al.[19]. While growth of the all-carbon molecules in the 2-25 atom range are assumed to consist of linear chains and mono cyclic rings, it is proposed that near 30 carbon atoms graphitic sheets are fonning. As these sheets are ingesting smaller carbon species, the growing molecules seek out a lower-energy path to reduce the number of dangling bonds by including pentagons in their structures. This pathway requires not only pentagon fonnation along the molecule's periphery, but internal bond rearrangement to form the pentagons among the interior hexagons. The pentagons are necessary to induce the curvature leading to closure and eventual fonnation offullerenes.

In 1992, Smalley[20] modified this formation picture by placing more importance on the dangling bond reduction and less importance on the random nature of pentagon formation. In Figure 2.1, the number of dangling bonds in an all-carbon molecule is plotted as a function of the number of atoms in the molecule. The "hexagonal sheet" distribution is concerned with the growth of compact molecules consisting of hexagons 21

22

20 r.t.' !\ C;~ \8 . ~o~ ~t."'o .. 16 c "0 m 14 CI' .'.: 0- 12 c 0 Pentogon Rule 0 '0 10 ..v D 8 E ::J Z 6

4

2

0 0 20 40 60 80 Number of "toms

FIGURE 2.1: Number of dangling bonds in compact hexagon-only sheets, compared to the compact molecules that obey the "pentagon rule," as a function of the number of carbon atoms in the sheet. Taken from ref. 20, used by permission. only. Naphthalenoctyl, for example, is a bicyclic CIO molecule (two hexagons sharing a side) with eight dangling bonds. If these bonds are tied off by hydrogens, the molecule would be called naphthalene (CIOHg). The pentagon rule requires that as many pentagons as possible be incorporated into the graphite sheets such that no two are adjacent. In the region of20-30 carbon atoms following the pentagon road leads to a significant reduction in the number of dangling bonds compared to a growing graphitic sheet.

Conclusions !Bat Smalley draws from the dangling-bond distribution are that "The energetically most-favored form of any open graphitic sheet is one which (a) is made up solely of pentagons and hexagons and (b) has as many pentagons as possible, while (c) avoiding adjacent pentagons [20) ." As appealing as this formation model is, as yet there 22 is little experimental evidence to confirm it. Clearly, extracting intermediate molecules from fullerene soot that lie along the 20-50 carbon atom range and obey the pentagon-rule distribution would greatly bolster this model of the formation pathway. However, the only carbon cluster that satisfies Smalley's criteria, and has been observed in fullerene production as a hydrogenated molecule, is at the low-mass end of 20 carbon atoms[14]

(Corannulene -C2oH IO). Corannulene is a curved PAH molecule of five hexagons surrounding a pentagon.

2.2 "Fullerene Road"

In 1991, Heath bypassed any consideration of growing chains or rings and simply proposed that fullerenes first form from clusters with approximately 40 carbon atoms

[21]. From there they grow by the addition of small species, e.g. C2, until C60 is formed. A fullerene of 40 carbons has adjacent pentagons, so the inclusion of smaller species naturally results in pentagon-pentagon separations. The idea is plausible because adjacent pentagons are high-energylhighly reactive sites. When C60 is initially formed, it may be a fullerene or almost a fullerene. In the latter case, some bond rearrangement is necessary to reach the fullerene isomer. However, once the fullerene is formed it is highly unlikely that a small carbon molecule (e.g. C2) could be incorporated, leading to the formation of C70, for instance. On the other hand, larger clusters could react with the fullerene to form fullerenes like C70 upon expelling a smaller intermediate.

The facile production of metal-encapsulated fullerenes, represents a serious difficulty for this model as it would appear to be difficult to introduce a metal into a closed fullerene. In the pentagon road model, the metal could nestle into a growing bowl that curves over the top and closes. At first glance the Fullerene Road model seems to explain the C2 separation of adjacent peaks in mass spectrometry for the fullerenes and their fragments. However, in the case of fragments, C2-loss transitions may be entirely unrelated to C2-addition transitions. 23

2.3 Combination of Naphthalenoctyl Units

The two kinetic models presented next start the fullerene pathway by combining intermediate-sized molecules to form a C60 ball. The first of these was proposed by Goeres & Sedlmayr (1991), and involves the successive combination of six bicyclic C 10 clusters (naphthalenoctyl's)[22], as shown in Figure 2.2. The dotted lines represent bonding, and are used to preserve clarity in drawing what are actually curved molecules.

To produce C70 without any bond rearrangement, seven ClO's combine to form a ball containing six-, five-, and four-sided rings. This isomer is not seen experimentally, and other isomers built out of ClO's are not observed either. Their predicted yield of C60 is 0.2%, and that ofC7o is a factor of25 lower. However, Curl points out that, theoretically, the naphthalenic CIO is much higher in energy than mono cyclic C IO[16]. This implies that unless there is a high concentration of ClO's in the carbon-arc chamber, the proposed pathway will have a negligible contribution to fullerene formation. Worse still is the fact that the presence of other experimentally observed fullerenes is not explained.

, ,, , , , , , , ,, , , \<':~.>'~:~::::

'. .. eo ......

FIGURE 2.2: Growth ofC6o by successive additions ofnaphthalenoctyl (CIO)' Dotted lines are used to iIlustrate necessary bonding and consequent curvature to close and form C60•

However, in the case of Taylor et a1.[15], and confirmed in this work, naphthalene

(C IO Hg) can be sublimed and carried into an oven at -lOOO°C, stripped of its hydrogens, 24

and combined with like molecules to fonn C60 and C70. In this work, a yield of .05% is obtained, along with a C60:C70 ratio of 30:1. The yield and ratio compare favorably with the model's predictions mentioned in the last paragraph. In addition, the presence of C70 is explained by Taylor as being due to bond rearrangement.

2.4 "Ring Stacking"

Another model that combines intennediate molecules to fonn C60 was proposed in 1992 by Wakabayashi & Achiba[23J. This model also begins with a naphthalenoctyl molecule and joins to it a monocyclic CIS ring. Another CIS ring is stacked onto this bowl-shaped molecule, followed by a C I2 ring, and finally a C2 to close and fonn C60• The C70 molecule is stack-assembled by also adding the CIS ring to naphthalenoctyl, but is followed by C20, C 16, and 3 C2's to close. Other fullerenes are not considered, and no yields are predicted for C60 or C70 .

There is no direct experimental evidence to support this growth model. In addition, it suffers the same problem as the model based on CIO's. There must be a substantial number of each of these molecules to fonn the typical yield of C60 found in carbon-arc generated soot. However, favorable to this model, is that when a CIO caps the hole in the CIS ring, distortion commences and continues to the point of producing the C2S intennediate molecule that is believed to be a fullerene precursor.

2.5 Annealing of Large Carbon Rings

The last kinetic model of fullerene fonnation discussed here appears to have the most experimental evidence to support it. In 1993, the group of von HeIden, Gotts & Bowers published experimental results that supported their conception of fullerene production[24J. Figure 2.3 shows the possible schemes for converting 40 carbon-atom rings into a fullerene product, with the energy required for the bond rearrangement coming from buffer-gas collisions. In their experiment, the molecular-ring isomers are 25

produced in the laser vaporization of graphite and then mass selected before entering a region occupied by a buffer gas. The gas region has a voltage bias to accelerate the ions at the entrance. The rings absorb energy from collisions that induce annealing and

convert the planar rings into a fullerene, with C2 loss. The C2 loss is a consequence of rapid bond breaking and rearrangement during the ring-fullerene transition. o Energy Energy CO + 0-0

Energy ~ qf t ct; Full ...... pn>dact + c.,

FIGURE 2.3: Isomers ofC4o+ are collisionaly heated to form a C38 fullerene following a C2 loss. Taken from ref. 24, used by pennission.

During the same year, Hunter, Fye, Roskamp, and Jarrold produced similar annealing results, but extended the work to include 50-70 carbon-atom isomers[25]. For the first time synthesis of C60 was demonstrated using planar rings. The higher fullerenes

(C 70 and up) could probably be synthesized if large enough rings (> 70 carbon atoms) could be produced. Further, a recent result of Jarrold et a1.[26] shows that the endohedral

La@C60+ (ionized lanthanum-encapsulated C60) can be synthesized by the laser vaporization of a La03-doped graphite rod to produce 60-carbon ring structures with a lanthanum atom attached. 26

CHAPTER 3

Experiment

The experimental work is comprised of both the pyrolysis of polycyclic aromatic hydrocarbons in a tube furnace and the carbon-arc vaporization of graphite in the presence of hydrogen. Each experimental technique is discussed here in detail. In addition, the analytical techniques are discussed.

3.1 Pyrolysis of P AH' s .

A schematic diagram of the apparatus used for the pyrolysis of polycyclic aromatic hydrocarbons (PAH's) is shown in Figure 3.1. Solid and liquid samples of known PAH's were vaporized in small quartz boats placed inside quartz tubes of 3/4" Ld., 2- and 3-foot lengths. The boats were made by cutting 112" Ld. tubes in half along its axis of symmetry. The boat ends were fashioned by melting the tube ends and turning them up at 90 0 angles. Their lengths were 1-112" and .4", capable of holding up to 100 and 500 milligrams of sample, respectively, depending on solid sample density. Ultra-high purity argon flowed through the large quartz tubes, out the exit and through a cold trap (liquid N2) providing the carrier gas for the vaporized samples and. Flow rates between 20 ml/min and 10llmin were maintained by flow meters. quartz tube (2' L x 19mm on) .

argon (u.b.p.) so ml/min furnace -+ "'*'----_.... _------~~-===~~------~------+ sample boat ..... 1000 DC to cold trap

propane torcb (for sample vaporization)

FIGURE 3.1: Schematic of tube furnace setup for PAH pyrolysis. 27

Two tube furnaces were used for pyrolysis. The smaller of the two has an upper limit of 950°C. The larger furnace is capable of reaching l200°C. Three kinds of heating sources were used for vaporizing the PAH samples prior to their exposure to the high­ temperature furnaces. In most of the earlier runs a propane torch was employed. For a more controlled heating a small tube furnace (approximately 6" long) was used. In some cases it was satisfactory to simply place the sample in its holder inside the quartz tube near the edge of the furnace.

Once the entire sample was vaporized, the tube was removed from the furnace and allowed to cool while the argon gas continued to flow. In the case of the larger furnace the cold trap had to be disconnected so that the tube could slide out of the furnace. Once the tube was set out to cool, the cold trap connection was re-established to limit sample exposure to air. In the case of the smaller furnace, a lid was lifted to allow an easy removal of the tube without disconnecting any part of the flow system. Once the tubes were cooled, the pyrolysis products were harvested. Most of the sample which had not undergone pyrolysis (not blackened), was removed separately and discarded. Either toluene or acetone was used to rinse out the tube interior. A metal rod was employed to scrape off deposits that the solvents could not remove. The harvested sample was filtered and washed with acetone repeatedly until no more color eluted from the filter. This step was performed to remove as much P AH sample from the extract as possible. The remaining extract or residue in the filter was then added to CS2 to dissolve any fullerenes.

3.2 Carbon-Arc Vaporization

Carbon is vaporized from a 114" graphite rod in the presence of controlled mixtures of helium and hydrogen . gases. As illustrated in Figure 3.2, two hollow, stainless-steel electrodes are electrically connected to an AC power supply which is capable of producing up to 225 Amps at 50 Volts. By way of vacuum feedthroughs the electrodes extend into the vacuum chamber. A graphite cuff, one inch in diameter and approximate- 28

Pressure

Hydrogen .. Helium ..

Vacuum Feedtbrougb

~ Water 1±::=s:0= .-

FIGURE 3.2: Carbon-arc chamber used for hydrogen poisoning of fullerene production. ly one inch in length, is mounted on each electrode. The cuff, shown on the left of Fig 3.2, has a 114" hole drilled into it about 112" deep to hold the 114" diameter graphite rods that are eventually vaporized. The chamber vessel itself consists of a seven-inch diameter, twelve-inch long, stainless-steel cylinder capped off at each end by flanges. The flange through which the cuff-only electrode enters is brass and is fixed from above to a custom-built Unistrut (Wayne, MI) assembly. This flange also contains the pumping port. The cylinder and remaining flange are also supported from above by the Unistrut assembly, but are free to roll along one axis towards and away from the affixed flange. This freedom allows for easy access to the chamber's interior for rod mounting, soot harvesting, and cleaning. In fact, both electrodes and flanges, the shroud, and cylinder 29

are toluene washed and acetone rinsed between rod burns. This cleaning reduces the possibility of contamination from one run to the next.

The cylinder contains a viewport and a port leading to the vacuum gauges, the gas manifold, and a venting valve. It is through the gas manifold that both helium and hydrogen enter the chamber. In addition, products that remain in the gas phase following vaporization, are extracted from the chamber through a port (not shown) into a container which can be valved off, detached, and analyzed by a gas chromatograph mass spectrometer.

The movable flange, through which the rod-mounted electrode penetrates, also supports a three-inch diameter, six-inch long copper shroud that surrounds the arc region, with 114" diameter copper tubing coiled around and soldered to it. Water flows though this tubing to conduct excess arc-generated heat out of the chamber. The carbon soot deposited on the inside of the shroud remains exposed to the arc and is not harvested for analysis since the arc's heat and ultraviolet radiation are known to destroy fullerene molecules[27]. The electrodes are also water-cooled to help prevent melting at the very hot points where the cuffs are mounted.

The two flanges and .the cylinder are sealed to each other through the vacuum created by the mechanical pump. Only Viton O-rings are used in the assembly on the flanges and the electrode feedthroughs. Once 60 mTorr is attained, the chamber is backfilled to -100 Torr of helium and/or hydrogen. The overall apparatus for the hydrogen poisoning aspect of this work resembles the standard carbon-arc assembly except that a gas manifold is added to allow the introduction of both helium and hydrogen gases as well as a port for gas extraction.

In the presence of the helium buffer gas and at a total chamber pressure of 100· Torr, soot production runs were carried out with partial pressure fractions ranging from zero to 100% H2• Approximately 300 mg of soot were harvested after each run and the 30 toluene-soluble products were soxhlet-extracted. Since many P AH's are formed in this type of process[17], necessary precaution must be used in handling what are most likely carcinogenic substances.

A typical run followed an overnight pump-down of the chamber which had been cleaned with toluene and acetone. The final chamber pressure was always less than 60 mTorr. The chamber was then backfilled to 200 Torr with helium and pumped out three times. Once the pressure returned to less than 60mTorr, hydrogen (H2) filled the chamber to the fractional level for the particular run. Every run was at 100 Torr total pressure. If the hydrogen level was 20%, for example, then enough hydrogen was introduced into the chamber to increase the pressure from less than 60mTorr to 20 Torr. Helium was then introduced, resulting in an increase from 20 to 100 Torr. Vaporization of the graphite rod was carried out for five minutes USillg AC power. During this time approximately two inches of rod were consumed. The chamber was allowed to cool for 30 minutes. It was then opened and the soot from all areas of the chamber, except the shroud's interior, was harvested.

3.3 Techniques for Analysis

Various analytical techniques were applied in this work to the soot, extracts from the soot, and gaseous samples from the chamber. Mass spectrometry of the dissolved extracts from the hydrogen-poisoning and PAH-pyrolysis was carried out on a Model 90 double-focusing mass spectrometer (Finnigan MAT, Bremen, Germany) following thermal desorption and electron-impact (EI) ionization of the samples. This mass spectrometer is capable of monitoring ion intensities as a function of time for selected masses. The gas extractions from the hydrogen-poisoned carbon-arc runs were analyzed by a Model 5890 gas chromatography / Model 5970A-MSD mass spectrometer (Hewlett­ Packard, Santa Clara, Calif.). The mass analysis of samples fraction-collected in high­ performance liquid-chromatography (HPLC) was carried out on a nitrogen (N2) laser­ desorption, reflectron-equipped, time-of-flight, mass spectrometer that was assembled by 31 the author. Portions of the raw soot and the soxhlet-extracted residue of the hydrogen­ poisoning samples were dispersed in a KBr matrix, pressed into a pellet, and analyzed using an MB-Series Fourier Transform IR spectrometer (Bomem, Quebec, Canada). For the soot samples harvested directly from the carbon-arc chamber, -0.2 mg quantities of sample and -300 mg of KBr were used. For the samples extracted with toluene and then dried, -5 mg quantities were dispersed in -300 mg of KBr. Two hundred scans in the mid-IR region (4000 to 400 cm- I) were carried out on each pellet sample and referenced to a pure KBr pellet. A summation-averaging routine is built into the Bomem software to reduce random noise. Soot produced under hydrogen-poisoned conditions was also analyzed by CCD-based, Raman-scattering spectroscopy (Photometrics, Ltd., Tucson, AZ).

Toluene extracts were analyzed by absorption in the ultraviolet and visible wavelength regions on a Cary Model 118 spectrophotometer (Varian, S ugarl and , TX). This instrument is interfaced to a computer that runs the spectrophotometer, takes the absorption data, and determines the amount of C60, C70 and higher fullerenes in solution based on mass-normalized spectra of pure fullerene samples[28]. HPLC was carried out using a Trident-tri DNP (Regis Technologies, Inc., Morton Grove, IL), or Buckyclutcher I, stationary phase column and a 1:1, by volume, mobile phase of hexanes and toluene. Both hydrogen-poisoned and PAH-pyrolysis samples were analyzed by HPLC. The absorption detector monitored the eluted sample at 300 nm. 32

CHAPTER 4

Results - Pyrolytic Formation

This chapter presents the results of many attempts to form C60 and other fullerenes by the pyrolysis of polycyclic aromatic hydrocarbons (P AH's). Several commercially-available solid and liquid samples of PAH molecules were used as starting materials for pyrolysis. It is shown through both high-performance liquid-chromatography and mass spectrometry that of all the PAH's tested, naphthalene is the only molecule confirmed to form C60 by pyrolysis. This positive result agrees with the earlier work of Taylor et a1.[15].

Figure 4.1 shows the vanous PAH's that were vaporized in the tube furnace detailed in section 3.1. Naphthalene was used because Taylor had already demonstrated o benzene (7Samu) naphthalene (I2Samu) o CKJ] cyclopentane (7Damu) dlcyclopentadlene (132amu decalin (13Samu) , N-.: "<:N ~ , ~ 00~ ~ OO quinoline (I30amu) lsoquinoline (130amu)

~ azulene (17Damu) ferrocene (IS6amu) ~ anthracene (I7Samu)

(a) (b)

FIGURE 4.1: Molecules pyrolyzed in attempt to fonn fullerenes. Both hexagon-only (a) and pentagon (b) donors were tried. 33

successful formation of C60 and C70. Figure 4.2a illustrates the tiling of C60 from SIX naphthalenoctyls (naphthalene minus the hydrogen's), and how the fullerene can be formed without carbon-carbon bond breaking.

(a) (b)

FIGURE 4.2: Tiling ofC60 using (a) six naphthalenoctyls, in black, or (b) twelve pentagons, in black.

Along with naphthalene, other hexagon-donors also shown in Fig 4.1 a were used. The motivation for trying quinoline and isoquinoline is that a fullerene formed from them is expected to have nitrogen atoms replacing carbon in the cage. Though C60 cannot be tiled from benzene or anthracene without carbon-carbon bond breaking, it is of interest to try them in case such bond breaking actually occurs.

Illustrated in Figure 4.2b is the tiling of C60 by pentagons only. Analogous to the naphthalenic molecules, pentagon-donors shown in Fig. 4.1 b were tested. Table 4.1 is a listing of all the trials, both successful and unsuccessful, to detect fullerenes formed from the pyrolysis of PAH' s. In some cases, HPLC showed success while mass spectrometry did not. Since contamination was a real problem for the HPLC, such mixed results were counted as nonsuccesses. Of the pentagon-donating P AH's, ferrocene deserves further work because the quantities of C60 observed were greater than expected for contamination. 34

Flowrate Furnace Sublimation Fullerenes (doped?) (ml/min.) temperature(°C) source: confirmed? Date PAH Small boat Pyrex wool T=Torch. Sample size unless present unless F=Furnace HPLC MS noted: noted: N=No (OC. if known) L=Large

J()! 13/1)3 Anthracene 200( Ll l)()O' . i"on:h Y ~.~ Yes

900 T

1200(N) Furnace N N 35

12/14/93 Ferrocene 1000 900 F N NA 12/14/93 Ferrocene 5000 900 F N NA 12114/93 Ferrocene 10000 900 F N NA 12/15193 Ferrocene 50 900 F N NA 12/15193 Ferrocene 125 900 F N NA 12/15193 Ferrocene 200 900 F N NA 12117/93 Ferrocene 10000 900 F N NA 12117/93 Ferrocene 50 900 F Y NA 12/20193 Ferrocene 50 900 F N NA 12/20193 Ferrocene 50 900 F N NA 12/20193 Ferrocene 50 900 F N NA 12/21193 Ferrocene 50 700 F N NA 12/21193 Ferrocene 50 700 F N NA 12/21193 Ferrocene 50 700 F N NA 12/22/93 Ferrocene 50 500 F N NA 12/22/93 Ferrocene 50 500 F N NA 12/22/93 Ferrocene 50 500 F N NA 115194 Ferrocene 50 1200 F Y NA 115194 Ferrocene 1000 1200 F Y NA 116194 Ferrocene 50 1200 F Y NA 116194 Ferrocene 1000 1200 F Y NA 117194 Ferrocene 50 1200 F Y NA 1114/94 Ferrocene 50 980 F(250) N NA 1114/94 Ferrocene 50 980 F(250) N NA 1114/94 Ferrocene 250 980 F(250 N NA 1114/94 Ferrocene 250 980 F(250) N NA 1118/94 Ferrocene 50 980 F(400) N NA 1118/94 Ferrocene 50 980 F(400) N NA 1118/94 Ferrocene 250 980 F(400) N NA 1/18/94 Ferrocene 250 980 F(400) N NA 1119194 Ferrocene 50 1200 F N NA 1119194 Ferrocene 50 1200 F N NA 1124/94 Ferrocene 50 980 F(250) N NA 1124/94 Ferrocene 50 980 F(250) N NA 1124/94 Ferrocene 250 980 F(250) N NA 1124/94 Ferrocene 250 980 F(250) N NA 1/24/94 Ferrocene 1000 980 F(250) N NA 1124/94 Ferrocene 1000 980 F(250) N NA 1/24/94 Ferrocene 250 980 F(400) N NA 1124/94 Ferrocene 250 980 F(400) N NA 1/24/94 Ferrocene 1000 980 F(400) N NA 1124/94 Ferrocene 1000 980 F(400) N NA 1125194 Ferrocene 50 980 F(400) N NA 1125194 Ferrocene 50 980 F(400) N NA 11?194 Ferrocene 50 980 F(250) N N 11?194 Ferrocene 250 980 F(250) N NA 11?194 Ferrocene 1000 980 F(250) N NA 2/2/94 Ferrocene 50 980 T N NA 36

2/2/94 Ferrocene 50 980 T N NA 2/2/94 Ferrocene 50 980 T N NA 2/4/94 Ferrocene 50 980 F(l80) N NA 2/4/94 Ferrocene 250 980 F(l80) N NA 2/4/94 Ferrocene alot 980 F(l80) N NA .J/2(J1'}.J . Fcrroccllc 50 t)()()(N) : F Y y' 4/20/94 Ferrocene 50 900(N) F Y NA 4/21194 Ferrocene 50 900(N) F Y N 4/21194 Ferrocene 50 900(N) F Y N 1111193 Isoquinoline 50 900 T M N 4/21194 Isoquinoline 50 1200(N) F M N 10/16193 Narhthalcl1I.: 2()()(i.) t)()() 'I y. Y 10/16/93 Naphthalene 200(L) 600 T M NA 10/18/93 Naphthalene 200(L) 900 T N NA 10/18/93 Naphthalene 200 900 T N NA 10/19/93 Naphthalene 200 900 T N NA 10/19/93 Naphthalene 100 900 T N NA 10/19/93 Naphthalene 50 900 T M NA 10/21193 Naphthalene 50 900 T Y N 10/21193 Naphthalene 20 900 T N NA 10/21193 Naphthalene 100 900 T N NA 10/22/93 Naphthalene 50 900 T M NA 10/22/93 Naphthalene 50 900 T N NA 1114/93 Naphthalene 50 900 T N NA 1114/93 Naphthalene 50 900 T N NA 1114/93 Naphthalene 50 900 T N NA 1115/93 Naphthalene 125 900 T N NA 1115/93 Naphthalene 125 900 T N NA 1115/93 Naphthalene 125 900 T N NA 1119/93 Naphthalene 200 900 T N NA 1119/93 Naphthalene 200 900 T N NA 11/9/93 Naphthalene 200 900 T N NA 2/28/92 Naphthalene 200 950 T N NA 2/28/92 Naphthalene 200 950 T N NA 2/28/92 Naphthalene 50 950 T N NA 2/28/92 Naphthalene 50 950 T N NA 2/28/92 Naphthalene 200 800 T N NA 2/28/92 Naphthalene 200 800 T N NA 2/28/92 Naphthalene 200 800 T N NA 2/28/92 Naphthalene 50 800 T N NA 3/1194 Naphthalene 200 650 T NA NA 3/1194 Naphthalene 200 950 T N NA 3/1/94 Naphthalene 200 950 T N NA 3/1194 Naphthalene 50 950 T N NA 3/1194 Naphthalene 50 950 T N NA 3/3/94 Naphthalene 200 950 F N NA 3/3/94 Naphthalene 200 950 F N NA 3/3/94 Naphthalene 200 950 F N NA 37

3/3/94 Naphthalene 200 950 F N NA 3/9/94 Naphthalene 50 1200 F N NA 3/9/94 Naphthalene 50 1200 F N NA 3/9/94 Naphthalene 50 1200 F N NA 3/9/94 Naphthalene 50 1200 F N NA 3/10/94 Naphthalene 200 1200 F N NA 3/10/94 Naphthalene 200 1200 F N NA 3/10/94 Naphthalene 200 1200 F N NA 3/10/94 Naphthalene 200 1200 F N NA 3/14/94 Naphthalene 500 900 F Y NA 3/14/94 Naphthalene 500 900 F M NA 3/IS/I).J Narhthalene SOO ' , I)()O F Y Y 3/15/94 Naphthalene 500 900 F Y N 3/16/94 Naphthalene 500 900 F Y NA 3/16/94 Naphthalene 500 900 F Y NA 3/26/94 Naphthalene 100 1000 F Y NA 3/26/94 Naphthalene 50 1000 F Y NA 4/19/94 Naphthalene 100 1000 F Y NA 4/19/94 Naphthalene 60 1000 F N NA 4/19/94 Naphthalene 50 1000 F N NA 4/19/94 Naphthalene 20 1000 F N NA 10/26/93 Quinoline 50 900 F N NA 4/25/94 Quinoline 50 1200 F N NA

Table 4.1: Results of attempts to form fullerenes from the pyrolysis ofPAH's.

Though naphthalene pyrolysis has, f~r the most part, been unsuccessful in forming fullerenes, when it is successful, the mass spectra clearly show the intermediates as well. See, for example, Figure 4.3. Other peaks are present in the spectra and may be the result of carbon-carbon bond breaking. Taylor explains the presence of C70 as possibly due to an isomerization of seven joined naphthalenoctyls. Where the fullerenes are formed remains a mystery. They either form in the flowing vapor stream or they form directly on the surface of the quartz tube.

Furnace temperatures exceeding 900°C appear to be necessary for successful formation of fullerenes by pyrolysis. As for sample harvesting, it does seem helpful to remove the pyrolysis products from the tube with acetone. It is recommended that samples be filtered and rinsed repeatedly with acetone until the yellow color of the PAR's 38

disappears. The filtered reside may then be placed in CS2, which should dissolve any fullerenes present.

As yet, there seem to be no set conditions for the highest or most consistent yield of C60 in the pyrolysis of P AH's. Even the ratio of C60 to C70 can vary quite 76 BE+85

83 .~ 97

.~.,. rIJ 580 ....=~ .=.,. 588 .=Q.,.

m/z FIGURE 4.3: Mass spectrum demonstrating successful formation ofC6o by pyrolysis of naphthalene along with evidence of the necessary intermediate molecules. 39 substantially. In the case of naphthalene, HPLC has shown the highest yield to be .05% by mass of the starting material, and the C60 to C70 ratio to range from 0.5 to 30. 40

CHAPTERS

Results - Hydrogen Poisoning

The carbon-arc technique for producing fullerenes was carried out in an alternate atmosphere containing hydrogen (H2). A total pressure of 100 Torr was maintained with helium and some partial pressure fraction of hydrogen. The hydrogen percentage ranged from zero to 100%.

5.1 Yield ofC60

Figure 5.1 shows the yield of C60 in harvested soot as a function of %H2• by pressure. The data was obtained from a curve fit program[28] based on known mass-normalized

7

6 2 runs@3% It' 5 -, 0 4 IQ U -- c~ 3 - 2 runs@ 15% 2 It' -•

o +-----+-----r---~~--~~--_+----_.----~------~------~----~- o to 20 30 40 so 60 70 80 90 100 0loH2

FIGURE 5.1: Yield ofC60 in harvested soot as a function of%H2• 41

ultraviolet/visible (UVNis) absorption spectra of C60 and C70, the prominent fullerene products in the carbon-arc synthesis. Runs at 3% and 15% H2 were made to determine reproducibility. A portion of soot sample harvested from the PAH chamber following each run, was placed in a soxhlet for toluene extraction as mentioned in Chapter 3. Table 5.1 lists the percentage yields by mass of extractable material from harvested soot. This Table plays an important role in adjusting the peak intensity values obtained from all the analytical techniques used, so that they may be normalized with respect to each other. The values are also plotted in Figure 5.2. Only one of the two runs at the 3% level is listed.

% H2 (run # if more Harvested Soot Extracted Mass Percentage Yield than one) Mass (mg) (mg) (%)

0 481 27.69 5.76 3(11) 488 23.10 4.73 10 484 11.84 2.45 15(1) 484 10.88 2.25 15(11) 482 7.10 1.47 20 480 5.64 1.18 50 483 18.8 3.89 80 481 19.40 4.03

Table 5.1: Extracted mass yields of harvested soot in hydrogen poisoning.

5.2 Mass Spectrometry

The three electron-impact (EI) mass spectra that best represent the steady progression of larger ring structures with increasing H2 concentration, are shown in Figures 5.3a-c. As expected, the spectrum of the 0% H2 run shows that C60 is the primary constituent. The spectrum shown for 15% H2 demonstrates the effect of hydrogen by the presence of 42

6

5 • "0 -~ 4 • • >.w ~

--=~ 3 00= • • "l- 2 • •

o +------;------~------~------~------~------~------_r------__1 o 10 20 30 40 50 60 70 80 %H2 FIGURE 5.2: Percent mass yield of soluble products in soot.

729

-.,.~ = of ~ .,.a fI.) 149 91 I =~ 369 C ++ ..... I 60 .,.= It' ....=

699

FIGURE 5.3a: Mass spectrum of 0% H2 soot extraction. 43

399 IE+83

,-...... ,rI.l ..... 326 15%) H2 =. 279 .c,. 619 350 633 '-'= 276 I .....~ rI.l 256 I ...,=~ .....= .....=Q

FIGURE 5.3b: Mass spectrum of 15% H2 soot extraction.

39B

374 324 I 50% H2 226 I

.....~ 276 424 rIl 199 .e= I ..... = 2:12 449 I I ...=Q

mJz

FIGURE 5.3c: Mass spectrum of 50% H2 soot extraction. 44

intennediate mass peaks in the 200 - 400 m/z range. The mass distribution centered at 633 m/z could not be consistently reproduced. It is anomalous and is discussed later in this work. The spectrum for 50% H2 displays a family ofPAH's distributed about an m/z value of 350. Each prominent peak in this distribution is separated from the next by m/z

= 24 corresponding to two carbon atoms. The distribution of peaks ranging in mass from 55 to 199 appears to be made up primarily of doubly ionized molecules of larger mass, fragments, and 2-3 ring PAH's. The peak at 360 in Figure 5.3a is the doubly ionized buckyball, or C60 ++. The non-fullerene mass peaks of prominence are at m/z values of 202, 226, 252, 276, 300, 324, 350, 374, 398, 424, and 448. All of these peaks have been checked as potential fragment molecules of some larger parent molecules[29]. The reason for such a check is to eIimin~te mass peaks that are commonly known fragments and that should not be considered representative of a fullerene-pathway intermediate molecule. None of these peaks could be attributed to fragments of other known PAH's as a result of electron-impact ionization. A peak seen at m/z of 250 was given particular attention because it may have been due to a molecule called corannulene (C2oH 10). Corannulene is widely considered to be an excellent candidate precursor in fullerene synthesis. When stripped of hydrogen (i.e. C20), three of them can combine to form a buckyball without any carbon-carbon bond-br~aking. However, an rn/z value of 250 can also be attributed to a fragment of the 252 mass P AH.

In Figure 5.4 all mass peaks of interest are plotted as normalized ion intensity versus the hydrogen poisoning level. At the 15% level, the peak intensity plotted is an average of both results. Similar averaging is also carried out for the plots reSUlting from all the analytical techniques used in this work. "Normalized" means that the ion intensities of each peak are multiplied by a value that compensates for the amount of material extracted from the soot by soxhlet. At the 0% H2 poisoning level, this factor is unity. The relation used is

Ip (peak intensity plotted) = a * 1m (peak intensity measured), 45

where a. is the nonnalization constant taken from Table 5.2. The method for detennining the measured peak intensity (1m) varies from one analytical technique to the next. In mass spectrometry (MS) the measured peak intensity is simply the highest level of ion

6

202 5 226 176 151 300 150 4 314 350 C. 374 -t:::, t).[) 398 => 3 ....;! m414 360 - 448 2 710

o +-----~------~----~------~----~~----~------r_----~ o 10 20 30 40 50 80

FIGURE 5.4: Log plot of mass peak intensities vs. hydrogen poisoning.

current detennined from the ion chromatograms, which show how many ions of a particular mass are detected in the mass spectrometer as a function of time. An example of ion chromatograms is shown in Figure 5.5. For Fourier transfonn infrared (FTIR) and UV Nis absorption, peak intensities were simply represented by the heights of the bands. Eluted peak intensities in the high-perfonnance liquid-chromatography (HPLC) were measured in tenns of their computed areas.

Figure 5.4 shows that the presence of C60 decreases with increasing hydrogen poisoning while the presence of PAH molecules increases with increasing hydrogen poisoning. In fact, the amount of C60 drops by threee orders of magnitude from 0% to

80% H2• The last major drop in C60 presence occurs between 20% and 50% H2• As will be seen, these trends are consistent with trends of certain peaks in the infrared (IR), 46

18~./ZI258• ------maxlmumpeakvalue ------_J8E+83 8 (1m) 3.278

6 4

2 ~ ______~ __~==~~e~e~ ______' __ ~ 18[jL./ZI2S2 8E+I. rl • 838 ~ : I ~ ; • 0 j'f"" A r~~: '~l·/zI3e. r~~;::

-" , I 29"7~:';=;= 48 69 89 188 121 148 168 time (sec) FIGURE 5.5: Sample ion chromatograms for particular m/z values obtained from electron­ impact mass spectrometry of a hydrogen-poisoning toluene-soxhlet extract. Arrow in top chromatogram illustrates method for determining peak intensities that were then normalized and plotted in figure 5.4.

Decimal Equivalent of Normalization Constant

%H2 Percentage Yield (a) (Table 5.1)

0 .0576 .0576/.0576 = 1.00 3(11) .0473 .0473/.0576 = 0.82 10 .0245 .02451.0576 = 0.42 15(1) .0225 .02251.0576 = 0.39 15(11) .0]47 .0147/.0576 = 0.27 20 .0] 18 .0118/.0576 = 0.16 50 .0389 .0389/.0576 = 0.68 80 .0403 .0403/.0576 = 0.70

Table 5.2: Calculated normalization constants for plotting MS, UVNis, and HPLC peak intensities in hydrogen poisoning. 47

UVNis and HPLC.

Although not seen in other spectra, anomalous peaks in both MS and IR analysis do appear in the first of two 15% H2 extracts (15[1]% H2) and in the 20% H2 extract. Evidence for concluding that these peaks are a result of contamination in either the carbon-arc chamber or the extraction' procedure will be presented in Appendix A along with all spectra of the 0% - 80% H2 poisoning runs that were also analyzed by IR and UV Nis absorption.

5.3 Infrared Absorption

Infrared absorption spectra of the same extractions used for the mass spectra of Figure 5.3 are shown in Figures 5.6a-c. These spectra have been referenced to a pure potassium­ bromide (KBr) disk prepared in the same way as the sample disk. The four IR lines due to C60 and two of those belonging to C70 can be seen in the 0% and 15(1)% spectra.

I Smaller bands in the 500 to 1500 cm- region are due to C70 and to combination modes of C60• In the 50% hydrogen spectrum C60 is not detectable, but numerous other bands have grown in this region. They represent C-H bending modes in PAH molecules that are seen in the mass spectra. Table 5.3 lists the concentrations of extracted material in each KBr pellet used for IR analysis. Also listed are the normalization factors applied to the peak intensities for plotting into Figure 5.7 with respect to hydrogen poisoning.

Analogous to the plots of Figure 5.4 for mass spectra, Figure 5.7 shows IR peak trends with increasing hydrogen. The IR absorption peaks plotted for C60 are the most prominent at 527 and 577 cm- I . Peaks not found in the 0% H2 spectrum and believed to be representative of PAH's include 540, 594, 617, 687, 710, 751, 824, 842, 882, 914,

1921, and 3280 cm- I. Though many more will be more carefully examined later, these peaks are plotted to lend further support to the fact that as hydrogen presence is increased in the carbon-arc chamber, some type of molecules are increasingly being formed. 48

3~------~

* C6()

2

QJ s::C,J ell ,Q I..o en ,Q -<

o~------~------~ 1000 500 Wavenumbers (em-I)

FIGURE 5.6a: IR absorption spectrum of 0% H2 soot extract.

3~------~

2

QJ s::C,J ell ,Q I..o en ,Q -<

o+------~------~ 1000 500 Wavenumbers (em-I) FIGURE 5.6b: IR absorption spectrum of 15(1)% H2 soot extract. 49

3.------~

2 50%H2

GJ tI as ,&J= 0 /I)"" ,&J < a.... 1

O~------r_------~ 1000 500 Wavenumbers (em-I) FIGURE 5.6c: IR absorption spectrum of 50% H2 soot extract.

5.4 UltravioletlVisible Absorption

The UV Nis spectra of the same three extractions are shown in Figures 5.8a-c. The large peak at 334 run is due to C60. Its strength has decreased at 15% H2 concentration and is practically absent in the 50% H2 concentration, where several weaker peaks signal the probable presence of other molecules. For plotting the several peak intensities of the UV Nis analysis, the same normalization constants found in Table 5.2 are applied. The peaks found at 334, 379, and 405 run, are attributed to C60. On the other hand, those peaks not seen until the 15% H2 level is reached (290.5, 298, 302, 320, 336, 345, 386, and 430 run) are believed to indicate a growing P AH presence with increasing hydrogen in the harvested soot. These peaks are plotted in Figure 5.9.

The IR and UV Nis spectra give a more quantitative representation of the C60 concentration than the mass spectra. Although the mass spectrum of the 50% H2 sample 50

Normalization Mass of Material Used Normalization constant

%H2 Constant of Table in KBr for FTIR for Figure 5.7 5.2 Analysis (mg) (a.)

0 1.00 5.28 1 3(11) 0.82 4.82 0.82 x 4.82/5.28 = 0.75 10 0.42 5.34 0.42 x 5.34/5.28 = 0.43 15(1) 0.39 5.50 0.39 x 5.50/5.28 = 0.41 15(11) 0.26 5.68 0.26 x 5.68/5.28 = 0.27 20 0.20 4.13 0.20 x 4.13/5.28 = 0.16 50 0.68 5.15 0.68 x 5.15/5.28 = 0.66 80 0.70 5.25 0.70 x 5.25/5.28 = 0.70

Table 5.3: Calculated nonnalization constants for plotting FTIR peak intensities in hydrogen poisoning.

2 517

1.5

c:. 0.5 -t::!- eo Q ..:I 0

·0.5

·1

.1.5 0 10 20 30 40 so 60 70 80

FIGURE 5.7: Log plot ofIR peak intensities vs. hydrogen poisoning. 51

shows appreciable C60' it is quite possible that its presence could be due to contamination from previous runs. In addition, at H2 concentrations between 10% to 20% we have observed appreciable C60 in IR and UVNis spectra. Although specific P AH molecules

2 3340 1.8 " O%,H 1.6 2 1.4 1.2 0= 1 0.8 0.6 0.4 0.2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 on on on on on on on on on on on on on on on on on on on on on on on on on on 00 0\ 0 N M ..,. on \0 I' 00 0\ N ...... ,. on \0 I' 00 0\ 0 N M N N M M M M M M M M ..,. ..,. ..,. ..,. ..,. ..,. ..,. ..,. ..,. - ...... ~ - on -on on on wavelength (A)

FIGURE 5.8a: UVNis absorption spectrum of 0% H2 soot extract.

2 1.8 1.6

1.4 3340 1.2 / o= 0.8 0.6 0.4 4050 / 0.2 O+-~~~-r-+-+~~--~r-+-+-~~~~-r-+~~~~r-+-~ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 on on on on on on on on on on on on on on on on on on on on M ..,. on \0 I' 00 0\ 0 N M ..,. on \0 iQ 00 0\ 0 N M M M ..,. ..,. ..,. ..,. ..,. ..,. ..,. ..,. ..,. ..,. M M M M M - on -on on on wavelength (A)

FIGURE 5.8b: UVNis absorption spectrum of 15% H2 soot extract. 52

2 1.8 50%, H2 1.6 1.4 1.2 =0 1 0.8 0.6 0.4 0.2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 If) II") 11"\ 11"\ 11"\ 11"\ 11"\ 11"\ 11"\ on 11"\ on If) 11"\ on 11"\ on II") on on on on on on on DO Q\ 0 N ...... ,. on \0 j!? DO 0\ 0 N ..... II") \0 r-- DO Q\ 0 N ...... ,. ..,. ..,. ..,. ; ..,. ..,. ..,. ..,. ..,. on on 11"\ on N N .... -...... - - wavelength (A)

FIGURE 5.Se: UV Nis absorption spectrum of 50% H2 soot extract.

2

33<10 A 1.5

1.3360 .. 3450 3010 Co 0.5 2910 -'-" c=.o 3860 -Cl ~ 0 2905

11.000 -0.5 lit-

.)

-1.5 0 )0 20 30 40 so 60 70 80 %H2

FIGURE 5.9: Log plot ofUVNis peak intensities vs. hydrogen poisoning. have not been identified from the IR and UV Nis spectra, both sets of results demonstrate the decrease of C60 and C70 with increasing H2 concentration along with the growth of molecules associated with the hydrogen poisoning. 53

5.5 High-Performance Liquid-Chromatography

HPLC chromatograms of the same three extractions of 0%, 15% and 50% are shown in

Figure 5.10. The retention times of 1.74, and 2.40 minutes correspond to C60 and C70, respectively. Values from Table 5.2 are used for plotting the retention time peak intensities as a function of fractional hydrogen pressure. Along with the retention time of

C60, the retention times of 1.26, 1.43, 1.61, 1.85,2.08,2.18,2.33,2.45, 2.59, 2.88, 3.13, 3.72, 4.14, 4.68, and 5.88 minutes are plotted as normalized peak intensity versus hydrogen poisoning level in Figure 5.11. The sixteen HPLC peaks listed here indicate

It" C60 c60 It"

.9 Toluene B" Toluene e ...... Toluene ~ ~ ......

o dmc(mlD) o dmc(mID) o dmc(mla)

FIGURE 5.10: HPLC chromatograms of 0, 15, and 50% H2 soot extracts. 54 the presence of at least fifteen different molecules formed as a result of exposing the carbon-arc ablation of graphite to hydrogen. As noted previously, a major drop in the C60 peak height can be seen with increasing hydrogen poisoning. At the same time, many other peaks appear, again indicating the presence of other molecules. The peak for C60 reaches its lowest point at 20% and then grows slightly with increasing hydrogen. Since

C60 is not seen at the 80% level of H2 in any other analytical technique, it is believed that a new molecule causes the 1.74 minute retention peak plotted in Figure 5.11.

2

l.33 1.5 1.85 3.13 3.111.45 1.11 143 2.18' 1. 1.59 4,68 1.61 4.14 1.16 ~Q O.S .5.88 ~ 1.74 o

-0.5

-I +-----~------~----~----~------+------+------~----~ o 10 20 30 40 so 60 70 80 %H2 FIGURE 5.11: Log plot ofHPLe peak intensities vs. hydrogen poisoning.

5.6 Gas Analysis

Before taking a closer look at what particular molecules may actually be forming as a result of hydrogen poisoning, it should be noted that much work was done to determine if any potential fullerene precursors could be found outside the extracted materials .. For instance, the gas formed following carbon-arc burns at 0%, 15%, and 80% H2 levels, was analyzed by gas chromatography/mass spectrometry (GCMS). The major components of the gas were found to be acetylene and methane. While acetylene was not detectable at 55

0% H2, its presence rose from 6000 to 110,000 ppm in the 15% and 80% H2 levels, respectively. Methane was not detectable until the third run of 80% H2 was gas-analyzed to show an 800 ppm concentration.

5.7 Soot Analysis

Along with analysis of the gas, the harvested soots were analyzed to try to find any molecules that may have been left in the soot following extraction. A look at Figure 5.2 shows that the amount of material extracted from the soot follows a real trend. The valley created between the descending fullerene yield and the ascending PAH yield, with increasing %H2, raises the question of what is being created in this region. Since Figure 5.2 shows this region to contain more insoluble soot, it is concluded that a transition is occurring here between appreciable fullerene production and appreciable PAH production in which not much of either is being made. Further analysis of the raw soot was carried out by both mass spectrometry and infrared absorption techniques. However, neither analytical technique showed peaks representing anything of interest in the soot.

It was not until the analytical technique of Raman-scattering spectroscopy was brought to bear on the soot that the realization of what was preferentially forming in the lower region of the valley occurred. Soots produced at 0, 10, 15,20, 50, and 80% H2, deposited on a quartz window at the bottom of the chamber near the fixed flange were analyzed using Raman spectroscopy. Figures 5.I2a - 5.12f display the results of this analysis. Based on the literature (e.g. Ref. 30), a peak in the 2100-2150 cm-) region is interpreted as the presence of a carbon-based, chain-type molecule containing triple­ bonded carbon. Shown in Figure 5.13, this polyyne-type molecule appears to emerge out of the background at 10%, maximize its presence at 15%, and disappear by 50% H2. Other structures to note in the spectra are peaks associated with disordered graphite (d), graphitic bonding (g), and an underlying distribution due to amorphous carbon (see Ref. 31). Seen only in the 0%H2 spectrum is a peak associated with fullerenes (see ref. 32). 56

fullerene g d 1 i'.... = ..::,of ....,e. rIl= .....!=

500 1000 1500 2000 2500 reI. wavenumber FIGURE 5.12a: Raman spectrum of 0% H2 soot.

g d i'.... = of ..::, ....,e. rIl .!= .S

500 1000 1500 2000 2500 reI. wavenumber FIGURE 5.12b: Raman spectrum of 10% H2 soot. 57

g d

:i.... = of ~ C=C ....~ CIJ ~= ....=

500 1000 1500 2000 2500 rei. wavenumber FIGURE 5.12c: Raman spectrum of 15% H2 soot.

:i.... = of ~ ....~ CIJ = C=C ....i

500 1000 1500 2000 2500 rei. wavenumber FIGURE 5.12d: Raman spectrum of20% H2 soot. 58

g d -....:l = of ~ ....0 III= ...~=

500 1000 1500 2000 2500 rei. wavenumber FIGURE 5.12e: Raman spectrum of 50% H2 soot.

g d ...~ = of ~ ....0 III =~ ...-= c=C?

500 1000 1500 2000 2500 reI. wavenumber

FIGURE 5.12f: Raman spectrum of 80% H2 soot. 59

4 • /\~RamaD 3 • 2

o -----.

o 20 60 80

FIGURE 5.13: Plot of soot C=C peak areas by Raman analysis as a function of%H2•

5.8 A Closer Look

Figure 5.14 illustrates the hexagon-only portion of what we believe to be the most likely products in the hydrogen-poisoning of the carbon-arc technique. Figure 5.15 illustrates the molecules that contain both hexagons and pentagons. Both types of molecules shown were obtained by a literature search[34] to find all known PAH's of particular masses consisting of hexagons and· pentagons only. To the author's knowledge, only those shown with asterisks next to their molecule names have been characterized by absorption spectra found in the literature. The remaining molecules are shown to illustrate potential growth schemes of PAH's. All of these masses have been observed as peaks in the mass spectra.

Most of the mass spectral peaks seen in the 50% H2 sample are attributed to graphitic, or hexagonal ring structures. In Figure 5.14, the series of masses 202, 252, 276, 300, 350, 374, 398, 424, and 448, which constitute the most prominent series in the mass spectrum, represents successively larger compact groupings of hexagons. It is 60

l6l6J !QlQT202. Pyrene

252. o 2":200 000 a 252. a 00 rgrgrgr Benzo[eJpyrene~ ~Perylene Benzo[a]pyrene

0 276. 000 ~ a rgrgrgr 276. Ananthracene Benzo[ghi]perylene * 0300 • 000 00 0Coronene

Benzo[a]coronene

Dibenzo[bc,kl]coronen

Ovalene

Benzo[g]naphth[8,I,2-abc]coronene

Benz[ d]ovalene

FIGURE 5.14: Compact, hexagon-only PAH's believed to exist in hydrogen-poisoned carbon­ arc soot. Asterisks denote molecules with published spectra. 61

2~ 161 ~202* ©roo rgrgu Aceanthrylene Acephenanthrylene Fluoranthene00

J.6laL 226 * J.6laL226 * rQLJQJ QYQXl . Benzo[ghi]f1uoranthene Cyclopenta[ cd]pyrene

252* 252*~ ~5~*ldL ':5:"*n ©OO ~ rgTQI.JgJ rgLJQTQJ Benzo[a]f1uoranthene Benzo[jJf1uoranthene Benzo[k]f1uoranthene Indeno[2,1-a]phenalene 252 * 25!*ldL ...... IQTQUQl Benzo[b ]f1uoranthene Benzo[j]acephenanthry lene

27~ l6r"1dJ 27~*J6l6l [Q1JQT " rgLJQTQJ Indeno[I,2,3-cd]f1uoranthene Indeno[I,2,3-cd]pyrene

Cyclopenta[bc]coronene

3~ l6I6n6J rQUQTQT'V' Diindeno[l ,2,3,cd: l',2',3'-jk]pyrene

Dicyclopenta[ aJ]coronene

FIGURE S.lS: Compact, hexagon-pentagon only PAH's believed to exist in hydrogen-poisoned carbon-arc soot. Asterisks denote molecules with published spectra. 62

possible that the fall-off of intensity in the mass spectrum above about 450 represents a decreasing solubility for the large clusters in toluene. It may be that the hexagonal carbon-clusters, in the absence of H2, would grow and become incorporated into the nonsoluble graphite soot. Laser desorption/mass spectrometry used to analyze the soot was unsuccessful at finding these larger clusters.

Of more interest in connection with fullerene production are the molecules that contain pentagons, shown in Figure 5.15 and Figure 5.16. A few of the peaks seen in the mass spectra cannot be attributed to hexagon-only ring structures. Masses 226 and 324, for example, can be interpreted as an addition of a single pentagon to pyrene (202 amu) and coronene (300 amu), respectively. The beginning ofa series incorporating pentagons is corannulene (250 amu), shown at the top left of Figure 5.16. This molecule is bowl­ shaped with a pentagon surrounded by five hexagons. Successive addition of one to five pentagons around the periphery of corannulene leads to the mass series 276, 300, 324, 350, and 374. The last member is a bowl-shaped PAR containing 30 carbon atoms, which would form C60 if hydrogen's were not present and if it were merged with a second such cluster. Unfortunately, the uniqueness of this corannulene series is difficult to establish without further analysis since the ma§ses 276,300,350, and 374 are also present in the hexagon-only series in the Figure 5.14. In addition, seeing a mass peak at 250 is most likely the result of two-hydrogen loss from the more common PAR mass of 252. The mass 324 P AH in the corannulene series does not occur in the hexagon-only group, but it does occur as the one pentagon addition to coronene.

On the far right in Figure 5.16 is a proposed series starting with mass 262. We have further identified this molecule by GCMS as the aromatic ring structure sumanene. This peak was found in relatively high abundance at 10% H2 shown in Figure 5.17 from an earlier run series[35), but not reproduced in the run series of this work. An HPLC 63

sumanene~ 262 CbJI

276

300 300 ~

324

1 etc.

350

FIGURE 5.16: Bowl shaped series that follow "pentagon rule." chromatogram of this earlier 10% H2 sample containing the 262 was carried out and compared to the chromatogram of the 10% H2 sample used in this work, which does not show the presence of mass 262 in the mass spectra. Since the chromatograms both showed the same peaks, no additional ones could be ascribed to the possible presence of a 262 mass. If a new peak had been seen only in the mass 262-containing sample, then fraction collection of the peak followed by mass spectrometry could have verified the presence of sumanene. Fraction collection is further discussed later in this chapter.

As in the case of corannulene, additional rings to the sumanene can lead to molecules with mass peaks also seen in the mass spectra (e.g. 300 amu). Despite the considerable overlap of PAH's in hexagon-only and pentagon-containing molecules, there 64

are several candidates that seem to require pentagons. These are mass 226, corannulene (250 amu), sumanene (262 amu) and mass 324.

262 ~ Qg9l .....~..... =

91 97 149 49 174 I

399

609 m/z

FIGURE 5.17: Electron-impact mass spectrum ofa 10% H2 run taken from a series reported earlier[35J.

Of the many possible PAH's present in the extracts, as inferred from the mass spectra, those that have infrared data are listed in Table 5.4. For each PAH a list of the five known PAH's that may be hiding under such a broad peak of a sample mixture should be considered. Figure 5.18 shows the IR spectrum of the 80% H2 sample in the 500 to 1800 em-I region. This spectrum is used for peak matching since the most peaks are present at this level of poisoning. PAH's that have at least two matches of their strongest peaks in the IR are listed in Table 5.5 with their respective UVNis data that fall within the 285-600 nm range. If a particular P AH is present in the extracted mixture, its strongest absorption features should be noticeable. The same peak-matching technique is used as in the IR except that the tolerance is now 1 nm instead of 3 em-I. The UV Nis spectrum for 80% H2 also contains the most peaks for comparison and is shown in Figure 5.19. 65

Molecule Name Infrared Peaks (cm-) in Descending Strength

Fluoranthene 776*, 748*, 826*, 619*, and 1439* Pyrene 840*,710*,749*, 1185,and 1433 Benzo[ghi]fluoranthene 824*, 757, 664, 598, and 502 Benzo[ a]fluoranthene 736*, 751 *,620*, 1439*, and 876 Benzo[b]fluoranthene 778,745*, 736*, 898, and 1452* Benzo[jJfluoranthene 813*, 768, 1426*,739, and 1482 Benzo[k]fluoranthene 822*, 745*, 888, 777*, and 1484 Cyclopenta[ cd] pyrene 838*,687*, 759, 889, and 1384* Indeno[2, I-a]phenalene 817*,769, 742,1216, and 700 Benzo[ a ]pyrene 689*,874, 742, 761, and 753* Benzo[ e ]pyrene 745*,827*, 772*, 1440*, and 1408* Perylene 763,813*,1499,1381*, and 1216 Corannulene 835,659,547*, 1312*, and 1134* Indeno[ I ,2,3-cd]pyrene 732*,842*, 1444*, 774*, and 887 Benzo[ghi]perylene 845, 813*, 753*, 767, and 648 Anallthrene 878,757,698,812*, and 794 Coronene 843*, 1305*,767, 1126, and 1613 Ovalene 1462*,874, 1381 *,837*, and 962

Table 5.4: Infrared peaks of known P AH's.

Combining the evidence supplied by Tables 5.4 and 5.5, the two PAH molecules of pyrene and coronene seem likely to exist in the extracted mixture of the 80% H2 level. It should be noted that PAH's typically have strong absorption peaks below 285 nm. Since the samples were extracted in toluene, whose absorption cutoff is at 285 nm, Table 66

1.5

.5

2000 1500 1000 500 Wavenumbers (em-I)

FIGURE 5.18: IR absorption spectrum of 80% hydrogen-poisoned soot extract.

2 1.8 1.6 1.4 1.2 =0 1 0.8 0.6 4300 0.4 tI 0.2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 In In In In on In In In on In In In on In In In on In In In s: In In In ·on N M -.to In 10 lQ 00 0- 0 .... N on 10 00 0- 0 N M 00 0- 0 !;f! -.to -.to -.to 'If' on on on on N N M -M M M M M M M M M -.to 'If' -.to ::: ~ - wavelength (A)

FIGURE 5.19: UVNis absorption spectrum of80% hydrogen-poisoned soot extract. 67

5.5 is incomplete. This means that many of the molecules listed in Table 5.5 may be in the extracted sample mixture at 80% H2 poisoning, but can't be positively identified because many of their absorption peaks are less than 285 nm.

Molecule Name UVNis Peaks (nm) in Descending Strength

Fluoranthene 288*,359,342,324, and 309 Pyrene 335.5*,320*,332,306, and 295 Benzo[ a]fluoranthene 422, 294, 309,388, and 364 Benzo[k ]fluoranthene 308,296,303,402, and 379.5 Cyclopenta[ cd]pyrene 356, 375, 340, 368, and 310 Benzo[ e]pyrene 290*,333,318,307, and 367 Perylene 435,408,292,386*, and 368 Indeno[ 1,2,3-cd]pyrene 303.5,316,292,360.5, and 377.5 Coronene 302*, 290*; 339, 323, and 333

Table 5.5: Ultra-violet/visible absorption peaks ofPAH's·that have known peaks in the IR.

In another attempt to detennine what molecules may be fonned in the hydrogen­ poisoned carbon-arc process, HPLC was employed. Figure 5.20 shows a typical chromatogram of the 80% H2 extraction. By taking a known PAH, adding it to the 80% H2 extraction, and then injecting this sample onto the HPLC column, one of the chromatogram peaks should be greatly enhanced. This fonn of "spiking" the sample was used to detennine what molecules may already be in the extract according to retention time. Using this method, the peaks of Figure 5.20 are tentatively assigned as due to the molecules pictured. In order to confinn the presence of particular molecules, fractional collection of each peak must be perfonned to obtain enough sample for mass spectrometry. The mass spectrometry result is then compared in tenns of fragmentation 68

1Dl6J~ 9? rgJgr or \9d?l 00 It'" \

~

=~ -.s= ooo@a 00 000 00 ~ ~, * ~ or ~ ~ ~ .:.: =~ 6f Q.,

o time

FIGURE 5.20: HPLC peak assignments using known PAH's to determine retention times in the 80% hydrogen-poisoned soot extract. "Peak #'s" were fraction-collected and analyzed by mass spectrometry. 69 pattern to the mass spectrometry results of known P AH's with the same molecular weights.

The fraction samples collected from the HPLC were analyzed by LDTOF mass spectrometry which are shown in Figure 5.21. Collection was carried out by first noting where in time the peak of interest eluted. At 2 mllmin. flowrate it was estimated that -2.5 seconds elapse between the time that the absorbance detector responds to the sample, and the time when it actually leaves the detector exit port. At the port exit a vial is placed to collect the sample as it comes out. Since a majority of the collected volume is the mobile phase solvent, several collections must be performed. For this work, twenty collections of each sample were made. Following collection the concentrating is increased by evaporation of the solvent. The residue is then dissolved in a small amount of CS2 as preparation for mass spectrometry. Unfortunately, the samples collected do not seem to have any particular fragmentation patterns as a result of LDTOFIMS. Most, if not all, of the peaks at arrival times of 15 J.lsec, or less, can be attributed to background noise. Although the peaks labeled pertain to the samples, they are unable to aid in distinguishing one P AH from another one of the same mass.

Figure 5.22 shows a plot of mass versus HPLC retention time using the masses of the known P AH's that were purposely spiked in the 80% H2 chromatograms. Included are the masses determined from LDTOFIMS of the collected sample fractions. Extrapolating this curve out to the retention times of the last peaks of the chromatogram suggests that masses of at least 326 amu are present in the HPLC. 70

250 -

200 peak #1 150

50

0-

250

200 = =r

100

SO

04------~1------~------r------~------OS 2uS 4uS 6uS BuS time-of-fIigbt (J..ISec) 250

200 - peak #3 ISO

50 -

0-

250 -

200 - peak #4 150 -

100

50-

O~------~------~------_r------~------OS 2uS 4uS 6uS BuS time-of-fIigbt (J..ISec)

FIGURE 5.21: Time-of-flight mass spectra of the four peaks HPLC fraction collected from the 80% hydrogen-poisoned soot extract and labeled in Figure 5.20. 71

350 • 300 • • • corannulene ...... • • 250 •• • -~ 200 •• S • rIl c:rIl 150 ~ 100 0 2 3 4 5 HPLC Retention Time (min)

FIGURE 5.22: Plot ofPAH mass vs. HPLC retention time of the 80% hydrogen-poisoned soot extract using known PAH's and mass spectra of fraction-collected HPLC elution peaks. 72

CHAPTER 6

Discussion and Conclusions

Two methods were followed in this work to better understand fullerene formation. In the pyrolysis ofPAH's approach, presented in Chapter 4, the goal was to confirm the results of Taylor et al.lIS], and to find sources other than naphthalene. However, the author's results met with few successes. The success of naphthalene confirmed the idea that one can start with a possible intermediate molecule of the fullerene pathway, and through pyrolysis, form a fullerene molecule. It also confirmed the model of Goeres and Sedlymayr, which predicted the formation of fullerene molecules from the combination of naphthalenoctyl units as described in Chapter 2. However, this model was originally proposed to explain fullerene formation in the carbon-arc. Upon comparison of the mass spectrometry analysis of the pyrolytic formation of fullerenes using naphthalene (Fig. 4.3) with the mass spectrometry results from carbon-arc produced fullerenes, it appears that the two formation techniques are unrelated. Furthermore, the yield of .05% C60 in the pyrolysis of naphthalene is negligible compared to the carbon-arc yield of up to 30%. A few of the runs with ferrocene appeared to successfully form fullerenes. However, the mass spectra did not provide evidence of i~termediate molecules to illustrate how the fullerenes were formed. Therefore, the presence of fullerenes in the pyrolysis of ferrocene is probably due to contamination.

The goal of introducing hydrogen into the carbon-arc process of fullerene­ formation, was to "freeze out" molecules that lie along the pathway to fullerene formation. From the beginning of this work it was already known from the work of Wilson et al.ll7] that adding hydrogen to the carbon-arc vaporization of graphite would lead to the production of P AH' s. However, the previous work was limited in the amount of analysis and the extent of hydrogen poisoning. Both aspects were extended in the work reported here. The degree of analysis included mass spectrometry, Fourier- 73

transform infrared and ultra-violet/visible absorption spectroscopy, Raman-scattering spectroscopy, and high-performance liquid-chromatography, and the degree of hydrogen poisoning was extended to 100% of 100 Torr pressure.

The analytical results show that the fullerene yield, in the carbon-arc production, decreases with increasing hydrogen presence. The presence of hydrogen appears to play a role by affecting the carbon-carbon reactions required for fullerene formation. In addition, analysis shows that a class of hydrocarbon molecules known as P AH' s are produced in larger quantities and in larger molecular sizes, as the amount of hydrogen poisoning is increase.d. This result indicates that not only are the hydrogens preventing the formation of fullerenes, but they are also combining with the carbon atoms to form polycyclic aromatic hydrocarbons.

Hydrogen poisoning at low levels shows possible smaller P AH' s incorporating pentagons that may be consistent with a pentagon road pathway. On the other hand, at higher hydrogen poisoning levels, there is no evidence for bowl-shaped PAH's with many pentagons. The larger molecules require more pentagons for curvature leading to a fullerene. The lack of evidence is best explained as due to the hydrogen forcing the hydrocarbon growth into an entirely differenf pattern from the pathway taken in the absence of hydrogen. The Pentagon Road remains a possible valid model because the pathway taken in the absence of hydrogen may be comprised of growing polycyclic ring structures of hexagons and pentagons that are not allowed to form in the presence of hydrogen.

PAH's are not the only molecules formed in the H2 poisoning approach. Figure 5.13 shows strong evidence for the production of chains over a certain range of hydrogen poisoning. The Raman spectra indicate that these chains contain carbon triple-bonds (CsC), which further indicates the preferential production of polyyne-type chains in a region where fullerene and PAH production are relatively weak. A qUalitative sketch of 74

what has been determined to be in the soot is shown in Figure 6.1. While there is no basis for relating the fullerenes distribution to the PAH's one, it is quite plausible that the -. .el .§ fullerenes .e PAH's ~ .0 .-s::CIl .sQ)

100

FIGURE 6.1: Qualitative sketch of overall picture concerning what is known of soot extracts.

decrease in intensity of fullerenes is related to the increase of chains, as a function of increasing hydrogen. As fewer chains are able to continue growing into fullerenes, fewer fullerenes are produced. Past the 15% H2 mark, the decrease in intensity of chains could be due to the presence of hydrogen inhibiting t!'te developing growth of chains.

The increasing presence of acetylene (C2H2), as hydrogen poisoning is increased, may give some support to this idea. Acetylene is itself a linear molecule in which the carbons are triple-bonded to each other (H-C=C-H). Its increasing concentration may be a consequence of the rapid tying-off of polyyne chains in the presence of increased hydrogen.

The production of polyyne-type chains would seem to fit best into the Annealing of Large Carbon Rings model presented in Chapter 2. The large carbon rings are thought to be polyyne carbon-chains[25] that attach themselves at their own ends to form planar mono-, bi-, and poly-cyclic ring structures. At some point in the growth of these rings the conditions are such that collapsing in a coiling manner to form fullerenes is favorable. 75

The rings are large enough that the carbon-carbon bonding remains chain-like. Whether the presence of hydrogen serves to freeze out the large rings or simply reduces their chances of fonning, the Raman scattering analysis demonstrates the resulting carbon bonding ofpolyyne-type carbon clusters.

Since the presence of carbon clusters in interstellar space has long been a motivating factor for our laboratory studies, the present results on P AH molecules were examined for any clues to the solutions of various astrophysical problems. For instance, P AH's are considered the likely sources of the observed infrared emission features known as the Unidentified Interstellar Bands (VIB's). Their peak locations in the infrared are 3030, 1610, 1230, 1163, and 885 cm-I. The spectrum of 80% H2 shown in Figure 5.18 does not indicate the presence of molecules that would give rise to these five peaks. Although the hydrogen-poisoned carbon-arc technique produces a wide variety of hydrocarbon molecules, none seem to match the long sought-after molecules of interstellar space.

Another interesting result to note is that none of the small, presumably stable hydrogenated fullerenes[36] are detected as peaks in the mass spectra. The proposed molecules include C2oH20 (260 amu), C2sH4 (340 amu), C32H2 (386 amu), C36HI2 (444 amu) and CsoHIO (610). Mass 300 (C24H I2) is of course present in our mass spectra, but is more naturally explained in the way depicted in Figure 5.14.

Further Work

Continued pursuance of the sumanene molecule is one worthwhile extension of this work. To find, extract and c.haracterize this molecule would not only be of interest in tenns of newly discovered molecules, but the fact that only three such molecules are required to fonn C60, compared to six naphthalenes, makes sumanene a potentially more efficient feed material that naphthalene. 76

In addition, the author recently learned that corannulene has joined naphthalene as a source for fullerene production by pyrolysis[37j. This result is not too surprising when one realizes that corannulene lies farther along the pathway to fullerenes than naphthalene, and the corresponding probability to combine and form fullerenes is higher. However, if naphthalene can produce fullerenes, why not quinoline? Quinoline is naphthalene with a single nitrogen replacing a carbon. The resulting fullerene should then be a doped molecule. Along with quinoline, more work should be done with ferrocene. To make fullerenes from pentagons only is intriguing, and some of the positive results obtained by the author could not be explained away as contamination. However, there was no consistency in the results and no intermediates would present themselves to show the steps that ferrocene took to become the observed C60, C70, and

CS4 molecules.

One of the last experiments done in this work was the Raman spectroscopy of the soot, which indicated the presence of carbon chains in the soot. Since this late result fits in nicely with current ideas of fullerene formation from chains that form large rings in the carbon vapor, this direction should be pursued. It would be particularly useful to investigate the solubility of the chains, so that other analyses such as HPLC, UVNis, and mass spectrometry may be explored.

The most direct approach to the study of intermediates between C I' C2, and C3, and fullerenes would be in-situ spectroscopy of the fullerene arc. In the ultraviolet, for example, one could use imaging spectroscopy with a spectrograph and CCD camera to provide spectra at a variety of positions in the near vicinity of the arc. One might be able to identify emission or absorption from carbon clusters in the process of growing fullerenes. 77

APPENDIX A

Anomalous Results

The mass distribution centered at m1z of 633 shown in the 15(1)% H2 spectrum of Figure

5.3b will be discussed in this App~ndix. As will be shown, a similar distribution occurred at the 20% level, but could not be reproduced at another 15% level run [15(11)%]. To explain what the author believes the source, or sources, of this distribution to be, results from mass spectrometry, and infrared for all levels of hydrogen poisoning in the carbon-arc process that were analyzed by these techniques, will be included here. Although UVNisible absorption spectroscopy and high-perfonnance liquid­ chromatography were also perfonned, the corresponding spectra are not included for both the sake of brevity and because in neither technique did peaks appear only in the first of two 15% level runs [15(1)%] and the 20% H2 level runs.

Figures A.l - A.8 show sample EI mass spectra for the 0, 3, 10, 15(1), 15(11), 20, 50, and 80% H2 poisoning levels in the carbon-arc. Figures A.9 - A.16 show the corresponding infrared absorption spectra, using KBr for the pellet matrix. Both analytical techniques were discussed in Chapter 3. As mentioned above, the mass peak distributions about an m1z of 633 are seen only in the 15(1)% and 20% runs. Also in these runs are notably high levels of known contaminants in samples analyzed by mass spectrometry. The m1z values of these known contaminants are 57, 91, 149, 167, and 279, all of which are attributable to the family of molecules known as phthalate esters. The 605 peak, being the lowest mass of the distribution with no smaller fragments is likely an aromatic molecule. The 619 - 661 peaks are the additions of one to four CH2 molecules, respectively. An additional CH3 gives the parent molecule at rn/z of 676. One may imagine a PAH molecule with hydrocarbon chains attached. The mass spectrometer simply knocks the chains off, link by link until the aromatic remains. The literature shows the mass spectral fragmentation pattern of such an aromatic with t-butyl groups attached[38]. Replace those groups with the hydrocarbon chains, and one should 78

729 aE+B3 C + 60 Jtt" .-...... I'I.l .=- ..c=. ~ '-'= C 149 .-I'I.l 91 I ....=~ .-= .-=Q

FIGURE A.1: Electron-impact mass spectrum of 0% hydrogen-poisoned soot extract .

57 • E+83

.-i' = ..c=. ~ !:- C .-I'I.l .e= c + = 60 .- Jtt" .-=Q

FIGURE A.2: Electron-impact mass spectrum of3% hydrogen-poisoned soot extract. 79

720

II -.... c + 60 = Itt' of $ ....~ =CIl ....~= ....=

rnJz

FIGURE A.3: Electron-impact mass spectrum of 10% hydrogen-poisoned soot extract.

309

CIl 326 ...... 279 -= = 619 .Q =.. 633 --~ .-CIl 2S6 I .....=~ .-= ....=

FIGURE A.4: Electron-impact mass spectrum of 15(1)% hydrogen-poisoned soot extract. 80

~7

...-....rIJ = ,.Q= .. 720 1 c + = 60 ...--.c It'' rIJ ....=~ .S ...=0

rn/z 880

FIGURE A.S: Electron-impact mass spectrum of lS(II)% hydrogen-poisoned soot extract.

91 .E+04

20% H2 ...-....rIJ = 149 = I of = ...--.c rIJ

=~ ...... = "=0 167 276 ... I 1308 326 r;~oI, 374 633

208 408 6BO rn/Z eBB lOBO

FIGURE A.6: Electron-impact mass spectrum of20% hydrogen-poisoned soot extract. 81

39B

300 I 374 .--.:l.,. 324 ., == 226 = I of $ p 424 ..,. rI.I 199 I == i.,. 149 . 252 I ..,.Q== 729 I

FIGURE A.7: Electron-impact mass spectrum of 50% hydrogen-poisoned soot extract.

226 324 I 389 I 350 .i'.,. == 91 = I of $ .p.,. 276 rI.I 159 280 ~== , 258 ..,.== .,.Q== 398 . I

mJz soo

FIGURE A.S: Electron-impact mass spectrum of 80% hydrogen-poisoned soot extract. 82

4~------,

*c60

(Ij .CJ l: * as 'f2 o .0"' < *

O~------T------r------.------~ 3000 2000 1000 Wavenumbers (em-I)

FIGURE A.9: Infrared spectrum of 0% hydrogen-poisoned soot extract.

4~------,

(Ij CJ l: as 'f2 o .0 <"'

O~------r------r------~------~ 3000 2000 1000 Wavenumbers (em-I)

FIGURE A.tO: Infrared spectrum of3% hydrogen-poisoned soot extract. 83

*c60

0.1...-----.------.... ,----.------....-----' 3000 2000 1000 Wavenumbers (em-I)

FIGURE A.ll: Infrared spectrum of 10% hydrogen-poisoned soot extract.

CII tJ I: <1:1 'f2 o en ~

3000 2000 1000 Wavenumbers (em--I)

FIGURE A.12: Infrared spectrum of 15(1)% hydrogen-poisoned soot extract. 84

4~------rr------.

* C60

O~------r------r------r------~ 3000 2000 1000 Wavenumbers (em-I)

FIGURE A.13: Infrared spectrum of 15(11)% hydrogen-poisoned soot extract.

4~------~------~

III (J d as of 2 o (/) .c <

O~------r------r------~------~ 3000 2000 1000 Wavenumbers (em-I)

FIGURE A.14: Infrared spectrum of20% hydrogen-poisoned soot extract. 85

4~------~

O~------r------r------r------~ 3000 2000 1000 Wavenumbers (cm-l)

FIGURE A.IS: Infrared spectrum of 50% hydrogen-poisoned soot extract.

4~------~

QJ =() 'f!2'" c I/) .&J <

O~------~------r------~------~ 3000 2000 1000 Wavenumbers (cm-l)

FIGURE A.16: Infrared spectrum of 80% hydrogen-poisoned soot extract. 86

expect to see a fragmentation pattern similar to the one centered about rnJz of 633. Figure A.17 shows a picture of the corresponding molecule. The fragmentation pattern for this molecule has only a few peaks in its mass spectrum. The smaller peaks are at rnJz's of 676 and 661. The peak at 619 is off scale.

Bu-t

Bu-t

MeO 3,S-Dimethoxy-4,7-di-(4-methoxy-3, 5-di-t-butylphenyl)-benzo[ c ]cinnoline

t-Bu OMe

FIGURE A.I7: Illustration of molecule that could be partly resposible for contamination in hydrogen-poisoning work. Replacing the t-Butyl's with hydrocarbon chains and analyzing the molecule by electron-impact mass spectrometry should result in a fragmentation pattern witnessed in the 15(1)% ami 20% H2 mass spectra.

In addition, relatively strong peaks in the infrared are observed only in these runs of 15(1)% and 20% H2 levels. Table A.I lists the corresponding IR peak values and assignments. Using the methodology outlined in Chapter 5, the mass spectral peak intensities as a function of %H2 level are shown in Figure A.IS. However, points for both 15% runs are plotted, unlike the plots in Chapter 5 that used averages, to demonstrate the inconsistency of seeing such peaks at this level. Since ion chromatograms of the mass distribution surrounding rnJz of 633 could not be obtained, the corresponding peak intensities are not plotted. 87

IRPeaks of IRPeaks of

15(1)% H2 20%H2 IR Assignments (em-I) (em-I)

704.9 702.0 CH, NH rocking. 742.5 743.5 1039.5 1039.5 . 1072.3 1071.3 C-C, C-N, C-O stretch. 1120.5 . 1120.5 1273.8 1272.9 C-O, C-N stretch. OH bend. 1286.4 1287.3 1378.0 1379.0 CHbend 1579.5 1580.5 NH bend. C=N, C=C stretch. 1599.0 1599.8 1725.2 1727.1 C=O stretch.

Table A.I: Infrared assignments of peaks unique to the 15(1)% and 20% samples.

6

mlz91

5 mlz91

.49

4 57 149 -e Co 167 btl 3 179 179 Q 167 .::l

2

o +,------~------~------~------~------~------~ o 3 10 15(1) 15(11) 20 so 80 %Hz FIGURE A.IS: Log plot of mass spectral peaks of known contaminants vs. hydrogen-poisoning run. 88

Interestingly, the known contaminant peak intensities follow the same pattern as those centered about rnJz of 633 in the mass spectra. The local maxima are observed in the IR as well and seem to indicate that the known contaminants are likely sources. Perhaps the higher mass distribution corresponds to some combination of the smaller sized contaminants. Whatever the source, it is unlikely that this molecule, or molecules, could have been created in the carbon-arc chamber. The presence of attached chains would seem to preclude this. The author believes that the observed, but unknown, contaminant results from the presence of air during the soxhlet extractions. This would explain where both the oxygen and nitrogen originate, but it doesn't explain why the products are not observed at all levels. 89

LIST OF REFERENCES

[1) D.R Huffman, Physics Today 11, 22 (1991).

[2) H.W. Kroto, J.R. Heath, s.c. O'Brien, R.F. Curl, and RE. Smalley, Nature 318, 162 (1985).

[3) W. Kratschmer, L.D. Lamb, K. Fostiropoulos, and D.R Huffman, Nature 347, 354 (1990).

[4) R Taylor & D.RM. Walton, Nature 363, 685 (1993).

[5) S.H. Friedman, D.L. DeCamp, R.P. Sijbesma, G. Srdanov, F. Wudl, and G.L. Kenyon, J. Am. Chem. Soc. 115, 6506 (1993).

[6) RM. Fleming, T. Siegrist, P.M. March, B. Hessen, A.R Kortan, D.W. Murphy, RC. Haddon, R Taylor, G. Dabbagh, A.M. Mujsce, M.L. Kaplan, and S.M. Zahurak,in Clusters and Cluster-Assembled Materials, edited by RS. A verback, J. Bernholc, and D.L. Nelson (Mater. Res: Soc. Symp. Proc., 206, Pittsburg, PA, 1991), p. 691.

[7] A.F Hebard, M.J. Rosseinsky, RC. Haddon, D.W. Murphy, S.H. Glarum, T.T.M. Palstra, A.P. Ramirez, and A.R Kortan, Nature 350,600 (1991).

[8) H.S. Shinohara, H. Sato, Y. Saito, A. lzuoka, T. Sugawara, H. Ito, T. Sakrai, and T. Matsuo, Rapid Comm. Mass Spec. 6, 413 (1992).

[9) T. Guo, M.D. Diener, Y. Chai, M.J. Alford, RE. Haufler, S.M. McClure, T. Ohno, J.H. Weaver, G.E. Scuseria, and RE. Smalley, Science 257, 1661 (1992).

[10) S. Iijima, Nature 354,56 (1991).

[11] D.S. Bethune, C.H. Kiang, M.S. deVries, G. Gorman, R Savoy, J. Vazquez, and R. Beyers, Nature 363,605 (1993).

[12) D. Heymann, L.P.F. Chibante, RP. Brroks, W.S. Wolbach, and RE. Smalley, Science 265, 645 (1994).

[13] T.K. Daly, P.R Buseck, P. Williams, and C.F. Lewis, Science 259, 1599 (1993). 90

[14] J.B. Howard, J.T. McKinnon, Y. Makarovsky, A.L. Lafleur, and M.E. Johnson, Nature 352, 139 (1991).

[15] R. Taylor, G.J. Langley, H.W. Kroto, and D.R.M. Walton, Nature 366, 728 (1993).

[16] R.F. Curl, Phil. Trans. R. Soc. Lond A 343, 19 (1993).

[17] M.A. Wilson, L.S.K. Pang, R.A. Quezada, KJ. Fisher, I.G. Dance, and G.D. Willett, Carbon 31, 393 (1993).

[18] J. Drowart, R.P. Burns, G. DeMaria, and M.G. Inghram, Chem. Phys. 31, 1131 (1959).

[19) J.R. Heath, S.C. O'Brien, R.F. Curl, H.W. Kroto, and R.E. Smalley, Comments ' Condo Mat. Phys. 13, 119 (1987).

[20] R.E. Smalley, Acc. Chem. Res. 25,98 (1992).

[21] J.R. Heath, in Fullerenes: Synthesis, Properties, and ChemiStry ofLarge Carbon Clusters, edited by G.S. Hammond & V.J. Kuck (Am. Chern. Soc. Symp. Ser., 481, Washington, D.C., 1991), p. 1.

[22] A. Goeres & E. Sedlmayr, Chem. Phys. Lett. 184,310 (1991).

[23] T. Wakabayashi & Y. Achiba, Chem. Phys. Lett. 190, 465 (1992).

[24] G. von HeIden, N.G. Gotts, and M.T. Bowers, Nature 363, 60 (1993).

[25) J.M. Hunter, J.L. Fye, E.J. Roskamp~ and M.F. Jarrold,J. Phys. Chem. 98, 1810 (1994).

[26] D.E. Clemmer, KB. Shelimov, and M.F. Jarrold, Nature 367, 718 (1994).

[27] R. Taylor, J.P. Parsons, A.G. Avent, S.P. Rannard, T.J. Dennis, J.P. Hare, H.W. Kroto, and D.R.M. Walton, Nature 351,277 (1991).

[28] F.A. Tinker, Ph.D. Dissertation, University of Arizona, 1995.

[29] Eight Peak Mass Spectra, Vol. 3, Part 3; Compiled by The Mass Spectrometry Data Centre, in collaboration with Imperial Chemical Industries PLC; The Mass Spectrometry Data Centre, The Royal Society of Chemistry: The University, Nottingham, NG7 2RD, UK, 1983, p. 6020. 91

[30] M. Nakamizo, R Kammereck, and P.L. Walker, Carbon 12,259 (1974).

[31] J. Robertson, Adv. in Phys. 35, 317 (1986).

[32] D.S. Bethune, G. Meijer, W.C. Tang, H.J. Rosen, W.G. Golden, H. Seki, C.A. Brown, and M.S. de Vries, Chem. Phys. Lett. 179, 181 (1991).

[33] Spectrometric Identification ofOrganic Compounds, 5th ed.; RM. Silverstein, G.C. Bassler and T.e. Morrill; John Wiley & Sons; New York, 1991.

[34] Ring Systems Handbook, Chemical Abstracts Service; American Chemical Society: Columbus, Ohio, 1993.

[35] M.C. Zumwalt & D.R Huffman, "Fullerenes and Polycyclic Aromatic Hydrocarbons," in Recent Advances in the Chemistry and Physics ofFullerenes and Related Materials, K.M. Kadish & R.S. Ruoff, Eds., 94-24, 360-370, The Electrochemical Society, Pennington, N.J. (1994).

[36] H.W. Kroto and D.RM. Walton, Chem. Phys. Lett. 214,353 (1993).

[37] R Taylor, Nature, submitted, 1995.

[38] The Wiley/NBS Registry ofMass Spectral Data, Vol. 6; F.W. McLafferty & D.B. Stauffer, Ed.; John Wiley & Sons: New York, 1989.