Tsvaygboym PhD Thesis 2007 - BW

RICE UNIVERSITY

Photochemical Studies of

Single-Walled Carbon Nanotube Ozonides and -Azoxy Ketones

by

Konstantin Tsvaygboym

A THESIS SUBMITTED

IN PARTIAL FULFILLMENT OF THE

REQUIREMENTS FOR THE DEGREE

Doctor of Philosophy

APPROVED, THESIS COMMITTEE:

Paul S. Engel,

Professor of Chemistry

W. Edward Billups,

Professor of Chemistry

Michael R. Diehl,

Assistant Professor of Bioengineering

HOUSTON, TEXAS

APRIL 2007

Volume I of II

ABSTRACT

Photochemical Studies of

Single-Walled Carbon Nanotube Ozonides and -Azoxy Ketones

by

Konstantin Tsvaygboym

This thesis contributes to two disparate problems in chemistry: studying properties of

carbon nanotube ozonides and products of their decomposition and determining behavior

of -azoxy radicals.

This work demonstrates that interaction of ozone with single-walled carbon

nanotubes (SWNT) results in formation of 1,2,3-trioxolanes (SWNTO3). Their formation

rate was found to be on the order of subseconds at room temperature for diluted SWNT -

1% aqueous SDS suspensions. SWNTO3 decayed to SWNT epoxides (SWNTO) with

release of molecular oxygen. Gas evolution measurements performed on dry ozonated

SWNT showed oxygen release to follow a simple exponential rise with rates

approximately 1.5 – 2 min-1

at r. t. The lifetime of SWNTO3, with a dissociation

activation energy of approximately 0.7 eV, depends on temperature and SWNT type. At

room temperature, it is less than two minutes for small-diameter SWNTs suspended in

water. Ozonides exhibited extreme quenching of SWNT fluorescence and substantial

bleaching of NIR absorption. The maximum number of 1,2,3-trioxolanes forming on the

surface of SWNT at any given time was found to be less than 4% of the theoretical value,

indicating a saturation point. Reaction of ozonated nanotubes with excess ozone is

limited by the SWNTO3 decomposition rate. Thinner tubes exhibited faster ozonide

decay rates resulting in greater oxidation levels over time in excess of ozone. Ozonation

with small quantities of ozone did not result in a D-band increase in the Raman spectra,

both for solid and liquid state experiments, though substantial decrease of the G band was

observed. IR absorbance kinetics of SWNT films revealed exponential intensity drift over

time with rates close to those in fluorescence and NIR absorbance techniques. Ozonated

SWNTs were found to abstract electrons from amines and thiols, thus resulting in

covalent attachment of nucleophiles to the sidewall.

The azoxy functional group greatly stabilizes an attached carbon-centered radical,

but the chemistry of such -azoxy radicals is unclear. This work reports that generation

of -azoxy radicals by irradiation of -azoxy ketones PhCO-C(Me)2-N=N(O)-R causes

ketone rearrangement to azoester compounds PhCOO-C(Me)2-N=N-R. This study

proposes a mechanism for this rearrangement.

Acknowledgments

I am grateful to my advisor, Prof. Paul S. Engel for allowing me to work on an

exciting, cutting edge project revolving around carbon nanotube ozonides. I have been

honored to work with a number of faculty, post docs, graduate and undergraduate

students, who immensely deepened my understanding of scientific principles and fostered

my teaching skills. There is no doubt some of them will become leading figures in

science, technology and business.

I would like to thank friends and relatives who were very supportive throughout my

graduate studies. Your help and advice are much appreciated.

Table of Contents

Volume I

Title Page i

Abstract iii

Acknowledgments v

Table of Contents vi

List of Symbols and Abbreviations ix

Part I

Chapter 1. Spectral and physical characteristics of reference SWNT samples 2

Introduction 3

References and Notes 12

Chapter 2. Carbon nanotube ozonides: formation rates, oxygen evolution,

decomposition rates and activation energies, determination of

saturation limits and a comparison of spectral changes in

fluorescence and UV-Vis-NIR absorption

13

Introduction, Results and Conclusions 14

Experimental Part 82

References and Notes 88

Chapter 3. Influence of SWNT ozonation on D and G bands in Raman spectra 91

Introduction, Results and Conclusions 92

vii

Experimental Part 106

References and Notes 107

Chapter 4. IR studies of SWNT ozonides and of products of their reactions

with different classes of compounds

109

Introduction, Results and Conclusions 110

Experimental Part 134

References and Notes 137

Chapter 5. Reaction of ozonated SWNT with electron rich nucleophiles

(amines, thiols and other)

139

Introduction, Results and Conclusions 140

Experimental Part 170

References and Notes 174

Chapter 6. Trapping reactive centers on SWNTOn with electron rich

nucleophiles (amines, thiols)

177

Introduction, Results and Conclusions 178

Experimental Part 181

References and Notes 182

Chapter 7. Reactions between ozonated SWNT and different classes of

compounds studied by X-ray photoelectron spectroscopy

183

Introduction, Results and Conclusions 184

Experimental Part 197

References and Notes 201

viii

Part II

Chapter 1. Photorearrangement of -Azoxy Ketones and Triplet Sensitization

of Azoxy Compounds

203

Introduction, Results and Conclusions 204

Experimental Part 221

References and Notes 227

Volume II

Appendix A Mathematics for regression analysis of fluorescence and NIR

absorbance data

235

Appendix B Supporting Information for Part I, Chapter 5. 1H NMR spectrum 251

Appendix C Supporting Information for Part I, Chapter 7. XPS spectra for

reactions of ozonated SWNT with different classes of compounds

253

Appendix D Supporting Information for Part II, Chapter 1. Calculated isotropic

Fermi contact couplings, computed structures, ESR, UV and NMR

spectra

323

ix

List of Symbols and Abbreviations

a.u. absorbance units

abs absorbance

ATR FT-IR attenuated total reflectance Fourier transform infrared

C60 fullerene C60

ca. Latin word for approximately

DTT dithiothreitol

ESCA electron spectroscopy for chemical analysis

em emission

ex excitation

HipCo high pressure carbon monoxide method

HOMO highest occupied molecular orbital

Imax maximum intensity

I/Imax normalized value(s)

Imax/I quenching factor, a degree of quenching, inverted normalized value(s)

absmax

local absorption maximum (spectral)

emmax

local emission maximum (spectral)

LUMO lowest unoccupied molecular orbital

NIR near IR

(n,m) carbon nanotube indices

O3 ozone

PM3 parametric method No. 3

x

PTFE polytetrafluoroethylene

r2 coefficient of determination, same as correlation coefficient

RBM radial breathing mode

SDS sodium dodecyl sulfate

SDBS sodium dodecyl benzyl sulfonate

SWNT single-walled nanotube

SWNTO3 product(s) of ozonation of single-walled carbon nanotube

lifetime

TMPD N,N,N’,N’-tetramethyl-p-phenylenediamine

uL microliter(s)

Wurster reagent N,N,N’,N’-tetramethyl-p-phenylenediamine (same as TMPD)

XPS X-ray photoelectron spectroscopy (same as ESCA)

Part I

Chapter 1

Spectral and physical characteristics of reference SWNT samples

3

1.1. Introduction

Single walled carbon nanotubes (SWNTs), a graphene sheet rolled up into a tubular

shape, may turn out to be a promising material for electronics, field emission, heat

transfer, sensing, material reinforcement, imaging, medicinal and other applications.1-3

Research in the area of carbon nanotubes increased significantly in the last several years

and is highly competitive, partly due to possible commercialization of their unique

properties. This chapter provides a brief introduction to key aspects of the spectroscopic

measurements of single walled carbon nanotubes (SWNT) discussed throughout this

thesis. Spectroscopic changes of SWNT after functionalization may not have the same

behavior as would be expected for a small molecule. An interesting example of this can

be found in Chapter 4 discussing IR absorption changes of SWNT over time after

ozonation. Chapter 1 contains an interconversion table of wavelengths and wavenumbers

that will be of use in Chapter 3, describing the Raman measurements performed on

aqueous SWNT suspensions as well as for discussion of IR results. Also, SWNT

fluorescence spectra obtained with different excitation sources are shown deconvoluted.

A brief table summarizes how much each tube contributes to the observed fluorescence

intensity. Other aspects like nomenclature and 3D structure are discussed as well. The

following section provides UV and NIR absorption spectra and talks about work with

different batches from the HipCo reactor (Rice University).

1.2. Near-IR fluorescence spectra

Two lasers, 660 nm and 785 nm were used for excitation of single-walled carbon

nanotubes (SWNT), the former one utilized for the majority of the spectra presented.

Wavenumber and wavelength scales are used interchangeably in this work. Table 1

4

shows the relation of the two scales. Specific Raman shifts from the 669.9 nm excitation

source are also included.

Table 1. Interconversion of wavelengths and wavenumbers for Visible, NIR and IR

regions. Raman shifts from 669.9 nm excitation source are provided.

Range , nm , cm-1

Shift, cm-1

Visible 669.9 14928 0

700.0 14286

733.5 13633 1294 (D)

749.5 13342 1585 (G)

785.0 12739

811.0 12330 2597 (G’) 830.0 12048

NIR 900.0 11111

1000 10000

1100 9091

1200 8333

1300 7692

1400 7143

1429 7000

IR 1500 6667

1600 6250

2000 5000

2500 4000

3333 3000

5000 2000

8333 1200

9091 1100

10000 1000

11111 900

12500 800

Aqueous SWNT-SDS suspensions are known to fluoresce when excited with suitable

lasers. Spectra obtained after excitation with 660 and 785 nm lasers are shown in Figure

1. The spectra were deconvoluted and peaks of interest assigned (n,m) numbers according

to published data.4

5

Norm

aliz

ed F

luore

scence

0.00

0.25

0.50

0.75

1.00

ex = 660 nm8,3

7,5

10,2

7,6

9,5 10,3

8,7

11,1

Optical frequency (cm-1

)

75008500950010500

Norm

aliz

ed F

luore

scence

0.00

0.25

0.50

8,3 7,510,2

11,310,5

8,7

6,57,6

9,7

ex = 785 nm

Figure 1. Fluorescence of aqueous SWNT-SDS suspensions. Tubes of interest are

marked with (n,m) numbers. The same (n,m) tube is shown with the same color and

symbol on both graphs. Top: excited with 660 nm. Bottom: excited with 785 nm laser.

Fluorescence changes in spectra obtained with em 660 nm were examined at four

distinct wavelengths: 954, 1027, 1125 and 1250 nm. The major contributors to

fluorescence intensity at each wavelength are summarized in Table 2.

6

Table 2. Major contributors to fluorescence intensity at four distinct wavelengths.*

em, nm , cm-1

(n,m) type of

major

contributors

tube diameter,

nm

% of total

emission at em

955.6 10465 8,3 0.782 95.4

6,5 0.757 1.9

1027.6 9731 7,5 0.829 85.0

10,2 0.884 5.3

8,1 0.678 4.3

1124.6 8892 7,6 0.895 78.9

8,4 0.840 8.3

9,2 0.806 3.7

9,4 0.916 3.4

1250.1 8000 9,5 0.976 39.8

10,3 0.936 30.3

11,1 0.916 12.0

8,7 1.032 6.2

10,5 1.050 3.8

8,6 0.966 3.0

* Excitation source ex

max 660 nm.

Minor contributors were excluded from Table 2 for clarity. Tube (8,3) contributed

95% of peak intensity at 954 nm, as deduced from spectrum deconvolution.5

Analogously, 85 % of peak intensity at 1027 nm was from tube (7,5). Tube (7,6) gave

only 79% of peak intensity at 1125 nm. The peak at 1250 nm was from a combination of

tubes, none contributing more than 40 % of total intensity.

Assignment of numbers (n,m) for carbon nanotubes is summarized in Figure 2.

7

Basis vectors

1,0 2,0 3,0 4,0 5,0 6,0 7,0 8,0 9,0 10,0 11,0 12,0

1,1 2,1 3,1 4,1 5,1 6,1 7,1 8,1 9,1 10,1 11,1 12,1

2,2 3,2 4,2 5,2 6,2 7,2 8,2 9,2 10,2 11,2

3,3 4,3 5,3 6,3 7,3 9,3 11,3

4,4 5,4 6,4 7,4 8,4 9,4 10,4

5,5 6,5 8,5 10,5

0,0

6,6 8,6 9,6 10,6

7,7 8,7 9,7

Armchair

ZigzagChiral

angle

13,1

12,2

12,3

11,4

11,5

7,7

10,6

13,0

8,7 9,7 10,7

8,8 9,8

8,3

7,5

7,6

9,5

10,38,3n = 8

m = 3

Roll-up vector

Basis vectors

1,0 2,0 3,0 4,0 5,0 6,0 7,0 8,0 9,0 10,0 11,0 12,0

1,1 2,1 3,1 4,1 5,1 6,1 7,1 8,1 9,1 10,1 11,1 12,1

2,2 3,2 4,2 5,2 6,2 7,2 8,2 9,2 10,2 11,2

3,3 4,3 5,3 6,3 7,3 9,3 11,3

4,4 5,4 6,4 7,4 8,4 9,4 10,4

5,5 6,5 8,5 10,5

0,0

6,6 8,6 9,6 10,6

7,7 8,7 9,7

Armchair

ZigzagChiral

angle

13,1

12,2

12,3

11,4

11,5

7,7

10,6

13,0

8,7 9,7 10,7

8,8 9,8

8,3

7,5

7,6

9,5

10,38,3n = 8

m = 3

Basis vectors

1,0 2,0 3,0 4,0 5,0 6,0 7,0 8,0 9,0 10,0 11,0 12,0

1,1 2,1 3,1 4,1 5,1 6,1 7,1 8,1 9,1 10,1 11,1 12,1

2,2 3,2 4,2 5,2 6,2 7,2 8,2 9,2 10,2 11,2

3,3 4,3 5,3 6,3 7,3 9,3 11,3

4,4 5,4 6,4 7,4 8,4 9,4 10,4

5,5 6,5 8,5 10,5

0,0

6,6 8,6 9,6 10,6

7,7 8,7 9,7

Armchair

ZigzagChiral

angle

13,1

12,2

12,3

11,4

11,5

7,7

10,6

13,0

8,7 9,7 10,7

8,8 9,8

8,3

7,5

7,6

9,5

10,38,3n = 8

m = 3

8,3n = 8

m = 3

Roll-up vector

Figure 2. Construction of a nanotube from a graphene sheet. Numbers n and m determine

the final position of a roll-up vector. Rolling sheet to superimpose hexagons (0,0) and

(8,3) will result in tube (8,3) with roll-up vector being perpendicular to tube direction.

Tubes of interest are emphasized with thick hexagons.

The physical structures of tubes of interest are shown in Figure 3.

(8,3) (7,5) (7,6) (9,5) (10,3)

Figure 3. Tubes (n,m) with the highest fluorescence intensity in HipCo samples for 661

nm excitation source. Each tube is shown in two projections (top and bottom).

8

It is important to note that there is no linear relationship between (n,m) tubes’

relative concentrations and their emission intensities for any given ex . This is because

SWNT fluorescence intensity is dependent on the wavelength of incident light. For

example, tubes (8,3), (7,5) and (7,6) with the highest emission intensity in the ex 660 nm

spectrum (Figure 1) are only a small fraction of a bulk sample (Figure 4).

6,4

9,1

8,3

6,5

7,3

7,5

8,1

10,2

9,48,4

7,6

9,2

12,1

8,6

11,3

9,5

10,3

10,5

11,1

8,7

14,0

13,2

9,7

12,411,4

12,2

10,6 11,6

9,8

15,1

14,3

10,8

13,512,5

13,3

14,1

11,7 12,7

10,9 11,9

15,2

16,0

14,4

5,0 7,0 8,0 10,0 11,0 13,0

5,1 6,1

6,2 7,2

5,3

5,4

12,8

11,10

13,6

armchair

zigzag

Figure 4. Distribution of (n,m) species in HipCo SWNT sample calculated from emission

spectra with ex 660 and 785 nm.6 Thickness of a hexagon is linearly proportional to tube

abundance in the sample.

Relative abundances of tubes were estimated by recording two separate emission

spectra with ex 660 and 785 nm. The knowledge of (n,m) tube abundance is of great

importance for absorption studies where measurements are performed on a bulk sample.

For example, if the bulk sample has two types of species, A and B, which transform over

time, independently of each other, into species A’ and B’ with corresponding rates c and

9

d, an overall absorbance can be expressed with a first order equation

dtct ebeatAbs )(

where a and b are Arrhenius prefactors derived from tube abundances. Typical HipCo

SWNT samples are estimated to have over forty different semiconducting tubes and

about fifteen metallic tubes. This means that observed absorbance can be affected by as

many as fifty five different species in a sample. Knowing relative abundances of specific

(n,m) tubes may help interpret absorbance kinetics.

Since metallic tubes do not fluoresce, their number is only an estimate. Studies of

SWNT radial breathing modes (RBM) in Raman spectra served as a basic for the relation

of abundances of metallic and semiconducting tubes.

Discussion of the mathematics behind (n,m) tube relative abundance calculations,

based on fluorescence emission spectra, is beyond the scope of this work and is not

included.5

1.3. UV-Visible and Near-IR absorption spectra

UV-Vis absorption spectrum for SWNT (HipCo, batch 162.4, Rice University) is

provided below:

10

Wavelength (nm)

300 400 500 600 700

Ab

so

rban

ce

(a

.u.)

0.2

0.4

0.6

Figure 5. UV-Vis absorption spectrum of aqueous SWNT – SDS suspension.

Absorption peaks in the area 450-550 nm are commonly assigned to metallic tubes.

Peaks in the area 650-750 nm are commonly assigned to semiconducting tubes.

NIR absorption of SWNT is thought to be caused by a conjugated network of double

bonds. It is not clear if the conjugated acene system in SWNT can be considered truly

aromatic. Hückel molecular orbital (HMO) theory states planarity as one of the most

important prerequisites of aromaticity. Ozonation of SWNT sidewall results in significant

decrease of NIR absorption. NIR absorption spectrum of pristine SWNT is provided

below.

11

Wavelength (nm)

900 1000 1100 1200 1300

Absorb

ance (

a.u

.)

0.15

0.20

0.25

Figure 6. NIR absorption spectrum of aqueous SWNT-SDS suspension.

Note the difference in the vertical scale for the above two spectra

1.4. Other spectra

Other reference spectra of SWNT, like IR, solid and liquid Raman and ESCA will be

introduced throughout the text.

1.5. Properties of different batches of SWNT

Different batches of SWNT from the HipCo process (Rice University) were used in

this work. All batches had similar or identical spectroscopic properties. Batch 153.3 was

used for fluorescence studies of the reaction between 2-methoxyethylamine and ozonated

SWNT. Batches 162.4 and 162.8 were used for IR studies. Batch 161.1 was used for UV,

liquid Raman and fluorescence studies. The majority of SWNT samples in this work were

used as synthesized, without purification. Unless otherwise noted, tubes were pristine.

SWNT – 1 wt. % aq. SDS suspension was prepared by a standard procedure outlined in

12

the experimental part. SWNT bundles, carbonaceous matter, metal catalyst and other

impurities are thought to be removed from the final SWNT – SDS suspension. Unless

otherwise noted, all SWNT – SDS samples used in this work were prepared by the same

procedure. Typically a large stock of SWNT – SDS suspension was prepared and used

for a great number of experiments.

1.6. References and Notes

1. Baughman, R. H.; Zakhidov, A. A.; de Heer, W. A., Carbon nanotubes - the route

toward applications. Science 2002, 297, (5582), 787-792.

2. Avouris, P., Molecular electronics with carbon nanotubes. Accounts of Chemical

Research 2002, 35, (12), 1026-1034.

3. Calvert, P., Nanotube composites - A recipe for strength. Nature 1999, 399, (6733),

210-211.

4. Weisman, R. B.; Bachilo, S. M., Dependence of optical transition energies on

structure for single-walled carbon nanotubes in aqueous suspension: An empirical

Kataura plot. Nano Letters 2003, 3, (9), 1235-1238.

5. Deconvolution performed with software package that accompanied NS1

NanoSpectralyzer (Applied NanoFluorescence LLC.).

6. Applied NanoFluorescence LLC http://www.appliednanofluorescence.com/.

13

Chapter 2

Carbon nanotube ozonides: formation rates, oxygen evolution,

decomposition rates and activation energies, determination of

saturation limits and a comparison of spectral changes in fluorescence

and UV-Vis-NIR absorption

14

2.1. Introduction

A number of publications have been dedicated specifically to ozonation of carbon

nanotubes. Recently, Chen1, 2

reported that 9 wt. % O3 in O2 bubbled through SWNT

suspension in perfluoropolyether (PFPE) at r. t. for periods ranging from 1 to 8 hours,

followed by a 30 minute purge with oxygen, resulted in SWNT shortening. Simmons et

al. 3 studied ozonation as a possible tool to selectively decrease conductivity of SWNT on

a microfabricated chip upon UV/ozone exposure. Samples exposed for one hour at r. t.

were shown to form characteristic carbonyl and ether bonds (XPS data), and SWNT

electrical resistance increased. The provided Raman spectra show D and G bands at

different times. After ten minutes of UV/ozone exposure, the G band decreased ca. five

times, but the D band did not change. The authors concluded that sidewall oxidation by

ozone and molecular oxygen resulted in - conjugated network disruption. Banerjee et

al.4-6

conducted a series of studies on ozonation of carbon nanotubes. The author noted

that Raman spectra of carbon nanotubes are strongly resonance enhanced, and as a result

signals from the functionalizing moieties are rarely seen in Raman spectra.4, 7

In a

different study, SWNT sidewall was ozonated (ca. ~10% O3 in O2) in a methanolic

suspension (100 mg in 150 mL) at -78 C for one hour and reacted with “cleaving”

reagents (either sodium borohydride or dimethyl sulfide).5 The authors assumed

formation of ozonides, by an analogy with alkenes, pointing out that C60O3 has been

reported in the literature.8 The “cleaving” step was introduced to alter relative distribution

of products (ethers, carbonyls and esters). The authors concluded that SWNT ozonation

15

could be used as a nondestructive method of introducing oxygenated functionalities

directly onto the sidewall.

In another study6 Banerjee et al. demonstrated that after solution phase ozonolysis of

SWNT (ethanolic suspensions, 2 hours), Raman peaks corresponding to smaller diameter

tubes were relatively diminished in intensity when compared to the profile of larger

diameter tubes. The author found no chiral selectivity (i.e. dependence on tube “twist,”

Figure 3, Chapter 1) and concluded that tube curvature and -orbital misalignment are the

main reasons for the observed selectivity. A theoretical study providing activation

energies for a reaction of ozone with SWNT has been reported.9 Cai et al.

10 reported

ozonation of SWNT and their assembly on top of oligo(phenylene ethynylene) self-

assembled monolayers. Oxidation produced oxygenated functional groups like carboxylic

acids, esters and quinone moieties. Depending on the degree of ozonation, the electrical

resistance was found 20 to 2000 times higher than that of pristine SWNT. Oxidation was

performed on a dry “bucky” paper with UV/O3 generator in ambient air for 25 minutes to

5 hours. Ozonated SWNT absorption in the IR region was shown to stop changing after 3

hours of ozonation. An IR peak at 1580 cm-1

was assigned to the stretching mode

(C=C) of double bonds in the nanotube backbone near functionalized carbon atoms.11

Ogrin et al.12

estimated an approximate molecular formula of SWNT ozonated for 3

hours to be C6O, i.e. every third double bond had an epoxide. None of the mentioned

publications focuses on SWNT ozonides kinetics.

A number of articles have been published on ozonation of fullerenes, a short analog

of SWNT, and their properties.13-19

Chibante and Heymann determined products of

16

ozonation of C60 in toluene solution included structures C60On, with n ranging from 1 to

6, and insoluble tan-colored precipitates.20

Bulgakov et al.21

found that epoxides C60On

(n = 1 – 6) are accumulated within the first three minutes of continuous ozonation.

Further ozone/oxygen mixture bubbling resulted in formation of ketone and ester

functional groups. Heymann et al.8 found that at 23 C ozonide C60O3 had a lifetime ca.

22 minutes in toluene, 330 minutes in a dry state and 770 min in octane.

Razumovskii et al.18, 19

reported ozonide formation rates for C60O3 (8.8 104 M

-1s

-1

at 0 C) and C70O3 (5 104 M

-1s

-1 at 22 C) in CCl4 solvent. The authors found that the

reactions obeyed a bimolecular rate law. The reactivity of C60 with ozone decreased ca.

90 times after the formation of C60O3. A similar tendency was found for C70, where the

formation of the first ozonide was 6 – 8 times faster than the subsequent ones. Fullerene

C70 was shown to uptake only 12 molecules of ozone within the first 16 minutes of

continuous O3/O2 gaseous mixture bubbling. The authors concluded that the formation of

the ozonide exerts an electronegative inductive effect on the adjacent network of

conjugated double bonds, similar to ozonation of divinylbenzene.22

Kinetics of SWNT ozonides have not been published to date. Among the reasons,

there are: different production methods resulting in different (n,m) types of SWNT in a

batch sample, a presence of a large number of different tubes in each SWNT sample,

poor solubility of SWNT in solvents, the need for efficient purification from the metal

catalyst, different purification techniques affect differently chemical and physical

properties of SWNT. Measuring kinetics on SWNT is a challenge. This chapter will

describe some interesting research findings discovered while attempting to study kinetics

17

of SWNT ozonides. Topics like deoxygenation of SWNT ozonides, NIR fluorescence

quenching degree, influence of high and low load of ozone on SWNTO3 decomposition

rates, proposed electronic transitions in SWNT and SWNTO3, decomposition rate

dependence on tube diameter, saturation limits in excess ozone, comparison of NIR

fluorescence and NIR absorption kinetics, establishing an average decomposition rate by

UV, structural changes and decomposition activation energies will be discussed.

2.2 Results and Discussion

A set of experiments was designed to measure the amount of oxygen evolving from

the surface of ozonated SWNT. No such study has been reported to date, even though a

number of articles on SWNT ozonation have been published.

Results of the experiment are summarized in Figure 1.

18

Time (min)

0 2 4 6 8 10

Pre

ssure

(m

To

rr)

0

25

50

75

100

Pressure (P1)

Pressure (P21

)

Pressure (P22

)

NIR

Absorb

an

ce (

a.u

.)

1.44

1.46

1.48

1.50

1.52

NIR Abs (A)

Pre

ssure

(m

To

rr)

0

25

50

75

100

P1

P21

P22

A

Figure 1. (Top) Pressure change at r. t. due to oxygen release from 2 () and 4 mg (

and ) of ozonated dry SWNT-coated glass (upper three curves) and corresponding

system-leak references (lower two curves, and ). (Bottom) NIR Absorbance

recovery of ozonated SWNT in solid form monitored at 1450 nm and r. t. (upper curve,

) and pressure change at r. t. due to oxygen release from 2 () and 4 mg ( and ) of

ozonated SWNT in solid state (lower three curves) after system leak correction. Curves

and were measured after the first and the second ozonation of the same sample

correspondingly.

Slurry of 2 or 4 mg of SWNT (as noted in Figure 1) in benzene (ca. 10 mL) was

added to the reaction vessel and was kept rotating until all the solvent was evaporated.

19

Such circular motion resulted in a thin SWNT film along the entire reaction vessel. A

vacuum line was degassed overnight, then the vessel was cooled to 5 C and 10 mL of

O3/O2 gaseous mixture (ca. 3 v/v % ozone23

) was injected to the bottom of the cylinder,

the cap closed and the vessel was left at atmospheric pressure for one minute. The valve

on the vessel was opened to the vacuum system and the vessel was evacuated for 1.5 min,

after which the pump was cut off and data were acquired. Degassing for one and half

minutes was found sufficient to bring the vacuum in the entire system to below 1 mTorr.

Time t = 0 min in Figures 1 and 2 indicates the point when the pump was cut off from the

system.

Time (min)

0 2 4 6 8 10

Pre

ssu

re (

mT

orr

)

0

25

50

75

100

Pressure (P1)

Pressure (P21

)

Pressure (P22

)

P1

P21

P22

A

NIR

Ab

so

rba

nce

(a

.u.)

1.44

1.46

1.48

1.50

1.52

NIR Abs (A)

Figure 2. Regression curves for NIR Absorbance at abs 1450 nm and for pressure

changes after SWNT ozonation in a dry state. Curves P1 () and P21 () correspond to

first ozonation of 2 and 4 mg of SWNT respectively. Curve P22 () was measured after

the sequential ozonation of 4 mg sample.

Cutting off the vacuum pump was followed by removing the ice bath and warming

the reaction vessel to r. t. with a water bath. Data points were collected until the observed

20

deoxygenation rate decreased to below the system leak rate value (ca. 0.5 mTorr/min).

The second sample (4 mg SWNT) was ozonated two times with approximately one hour

interval between oxidations.

The highest amount of oxygen evolved after gaseous ozonation of solid SWNT was

estimated as 0.72 umol within a 20 min time period at room temperature. This

corresponds to 0.2% of carbon atoms (or to 0.1% of double bonds) of SWNT (4 mg)

oxidized with ozone, assuming that all carbon soot was indeed SWNT or had a fullerene-

like structure. Weighing error of SWNT could bring an error into the calculated value. It

is possible that the number of double bonds reacted with O3 was higher, though it would

still be significantly less than the 3 – 4 %, estimated in UV studies at 260 nm. (NIR

absorbance estimation was at ca. 4 – 5 %). A possible explanation for such a low yield of

oxygen is SWNT bundling, which physically prevented large surface areas of SWNT

from reacting with gaseous ozone.

Ozonation of SWNT flakes resulted in their immediate burning. Ice bath cooling and

SWNT deposited along the glass wall of the reaction vessel were found necessary to

prevent this highly exothermic reaction from overheating.

NIR absorption was fitted with formula F1, while the pressure curves were fitted

with the 5-parameter two exponential rise formula F2 (formula selection discussed in

Appendix A):

minmin1

y

ceaey

yy

n

bt

bt

final

(F1)

21

)1()1(0n

bt

bt eceayy (F2)

Regression results are summarized in Table 1 below.

Table 1. Regression results for pressure changes and for NIR Absorption at

em 1450 nm after SWNT ozonation.*

Data Set Oxygen gas release NIR Absorption

Parameter F2, 4 mg (P21) F2, 4 mg (P22) F2, 2 mg (P1) F1

b, min-1

2.07 1.45 1.88 1.540

n 9.63 11.45 9 14.02

r2 0.9991 0.9987 0.9995 0.9995

ymin - - - 0.0000

* Ozonation of SWNT film deposited on a glass surface. The formula number and SWNT

amount used for the experiment are written at the head of each column. Rates b are

expressed in [min-1

]. Active constraints used in analysis were n > 9 (2 mg SWNT

pressure curve) and ymin > 0 (NIR absorption).

Points at time zero were excluded from regression because those were acquired at

5 C; all subsequent points were acquired at or near r. t. Constraints n > 9 and ymin > 0

were introduced to generate a better fit to the experimental data. Limiting n to greater

than nine was needed to better describe the term n

bt

ce , a “slow” component, for

pressure curve P1. Parameter n describes how many times the slow component is slower

than the fast one.

Approximately the same amount of ozone (O3/O2 gaseous mixture) was injected into

the reaction vessel in each experiment. The first time ozonation (P21) yielded a slightly

higher rate than the subsequent one (P22). All rates were comparable to those observed by

NIR fluorescence recovery, indicating that decomposition of a single ozonide is likely to

increase fluorescence intensity. This result means that the smallest section of SWNT

22

needed for a tube to fluorescence can be loaded with no more than one or two ozonides

on its surface, at least in an aqueous suspension.

Fluorescence studies demonstrated that 1,2,3-trioxolanes on the surface of SWNT

prevented the tube from emitting in the NIR region. If the “minimal” section of SWNT

needed for fluorescence carried several ozonides, all of them would have to decompose

before this section would gain its ability to fluoresce. If that were the case, then true

ozonide decay rates would be several times greater than those observed by fluorescence.

Observation of similar rates in vacuum deoxygenation of SWNTO3 and in fluorescence

techniques implies that decomposition of nearly every ozonide results in a fluorescence

increase.

It was found difficult to quench SWNT fluorescence completely. The highest

quenching degree (Imax/I) was less than 1000 times and tubes were shown to quickly

recover from that state. Quenching 1000 times means that 0.1% of previously emitting

“sections” of SWNT continued to fluoresce. Full fluorescence quenching was not

observed. A study was performed to investigate the fluorescence quenching degree

(Imax/I) as a function of the volume of injected O3/O2 gaseous mixture (ca. 3 v/v %

ozone). After excluding the most extreme points (i.e. the lowest intensity point after

ozonation), even with large amounts of ozone, such as 2 mL of O3/O2 gaseous mixture,

fluorescence could not be quenched more than 140 times (Figure 3).

23

O3 / O2 mixture volume (mL)

0.0 0.5 1.0 1.5 2.0

Flu

ore

sce

nce Q

ue

nchin

g (

I max/I

)

0

20

40

60

80

100

954 nm

1027 nm

1125 nm

1251 nm

Figure 3. Dependence of fluorescence quenching degree (Imax/I) on the amount of O3/O2

gaseous mixture (ca. 3 v/v % ozone) injected.

Figure 3 demonstrates that injection of 0.3 mL of O3/O2 gaseous mixture decreased

fluorescence intensity of tube (8,3) with emmax

954 nm approximately 6 times. In

percent values it means that only 17% of all emitting “sections” were contributing to

fluorescence. One would expect that increasing the ozone load by 20 % could nearly

completely extinguish fluorescence from the tube (8,3). Interestingly, injection of 0.5 mL

of O3/O2 mixture quenched fluorescence only 16 times, i.e. 6% of SWNT was still

emitting. Further increase of the ozone load to 1.0 mL quenched emission only 41 times,

with 2.4% of emitters still left to be quenched. To conclude, increasing ozone load from

0.3 mL to 1.0 mL, i.e. by 330%, could not extinguish the remaining 17% of emitting

sections of SWNT. This observation meant that tubes are getting oxidized with ozone in

bands and not randomly.

24

Changes in SWNT fluorescence after oxidation with ozone

Wavelength (nm)

950 1050 1150 1250 1350

No

rma

lize

d F

luo

resce

nce

In

tensity

0.0

0.2

0.4

0.6

0.8

1.0

Before O3

1 min

3 min

9 min

ex = 660 nm

Figure 4. Addition of aqueous solution of ozone (50 uL, Abs (260 nm, 1 cm) = 1.25 a.u.)

to 0.5 mL SWNT-SDS aq. suspension. Used 660 nm laser for excitation. Fluorescence

emission quenching was followed by a slow recovery. Spectra recorded before, 1, 3 and 9

min after ozonation.

The overlaid spectra in Figure 4 show SWNT fluorescence change over time after

ozonation. The spectrum of pristine SWNT is provided for comparison.

SWNT oxidation was accomplished by an addition of a small volume of water

saturated with ozone. It was desired to prepare a saturated solution of ozone, thus

decreasing the volume of ozonated water needed for oxidation. A dilution of SWNT-SDS

suspension was a concern, since dilution could result in SWNT agglomeration, thus

leading to lower fluorescence intensity. In general, bubbling O3/O2 gaseous mixture

through the solution was of a greater benefit, since in that case there was no need to

25

worry about sample dilution. While dilution with 1% SDS decreases SWNT fluorescence

intensity, no comprehensive study was performed in this work to estimate the influence

of dilution on fluorescence. Aqueous SDS solution was not used for ozone accumulation

primarily because this surfactant is a known catalyst for conversion of ozone into

molecular oxygen.

For ozonation in the solution phase, fluorescence quenched to a different extent (20

to 100 times) had fairly close recovery rates as shown in logarithmic scale in Figure 5B.

26

I /

I ma

x

0.0

0.2

0.4

0.6

0.8

1.0

Time (min)

3 6 9 12 15

I ma

x / I

(lo

g s

cale

)

1

2

4

10

20

40

100 B

A

Figure 5. SWNT emission change at 1247 nm after ozonation. Used 661 nm laser for

excitation. Every fifth experimental point is shown with a symbol. (A) Change of

normalized fluorescence intensity (I/Imax) with addition of different amounts of ozone.

The higher level of ozone resulted in a lower fluorescence intensity (line ). The lowest

amount of ozone gave the highest intensity (line ). (B) Fluorescence quenching factor

(Imax/I) shown in logarithmic scale. Different oxidation degrees gave close decomposition

rates.

27

Time (min)

0 5 10 15

Invert

ed

No

rmaliz

ed F

luo

resce

nce

, I m

ax /

I

0

50

100

150

954 nm

1026 nm

1123 nm

1250 nm

A

B

C

D

0 1 2

5

10

15

20

A

Figure 6. Influence of ozone load on fluorescence quenching at different emmax

.

Inverted normalized fluorescence (Imax/I) at four distinct wavelengths is shown. Four

independent experiments (A-D) are shifted along the time axis for clarity. A zoom-in for

experiment A is provided in the upper left corner. Wider tubes ( em 1250 nm) were

quenched more at low ozone loads (A-B). With higher loads all tubes were quenched to

the same degree (C-D).

Large loads of ozone, typically above 1 mL of O3/O2 gaseous mixture (ca. 3 v/v %

O3), injected into 1 mL of SWNT – SDS suspension resulted in slower decay rates. Rates

obtained from samples with quenching degree (Imax/I) below 200 were reproducible.

Rates obtained from higher levels of ozonation were difficult to reproduce even with a

thermostated cuvette. The common problem was the curve deviation from simple

exponential decay.

Tubes emitting at longer wavelength, emmax

= 1250, typically with wider diameters,

were quenched to a higher degree within Imax/I range 20 to 130 (Figures 6A and B).

Higher ozonation loads resulted in all tubes getting quenched to the same degree (Figure

6C and D).

28

Time (min)

0 2 4 6 8 10

Inve

rte

d N

orm

aliz

ed

Flu

ore

sce

nce

0

25

50

75

100

2 4 6

5

9

13 A

B

C

D

em = 1026 nm

less O3

more O3

H

H

L

L

Figure 7. Regression fit for inverted normalized fluorescence at 1026 nm (formula F4).

Each regression curve represents an independent experiment and is shown with a solid

thin line. Samples were ozonated to a different extent; curves A and B correspond to a

low ozone load, while C and D to a high load. Arrows labeled L and H point to curve

deviation caused by slowly decaying ozonides. Curves B-D were shifted along the time

axis for clarity.

Emission kinetics at emmax

= 1026 nm for different ozone loads were fitted with

regression curves; the highest two points (after ozonation) on inverted normalized

fluorescence data sets were excluded from regression analysis. As described in Appendix

A equilibration periods should be excluded from ozonide decay regression.

Overall decomposition rates were found to be lower with higher ozone loads. Two

possible explanations for such a phenomenon are a) ozonides formed are either lateral or

longitudinal to tube axis (Figure 27), or b) closely situated ozonides affect decomposition

of nearby ozonides.

29

The ozonide decay rates for curves A – D (Figure 7) were calculated with the

following formula:

n

bt

bt

final ceaeyy (F4)

Regression results for Figure 7 are summarized in Table 2.

Table 2. Regression results calculated with formula F4 for inverted normalized

fluorescence data recorded at 1026 nm emission wavelength.*

Curves

Parameter A, n > 10 B, n > 10 C, n > 8 C, n > 10 D, n > 8 D, n > 10

a 28.45 58.43 102.4 195.1

b, min-1

1.96 1.59 1.27 1.24 1.36 1.32

yfinal 1.71 1.48 1.10 1.10

c 1.18 2.24 4.48 6.26

n 10 10.00 8.00 8.00

r2 0.9997 0.9997 0.9977 0.9954

* Curve one-letter symbol and a lower boundary for variable n are written at the head of

each column. Constraints used for regression were: 0 < a < 1000; 0 < b < 100; 0 < c < a;

n < 100; yfinal > 1.05. Rates b are in min-1

.

For regression purposes, 'tails' on inverted data sets were truncated to increase the

weight of points related to a fast decay. Rates were calculated with 5-parameter formula

F4. The formula has fast and slow exponential terms, the slow one being n times slower

than the fast one. (For details on mathematics behind regression see Appendix A.)

Parameter n was kept greater than 10 for low ozone load curves, since formation of

slowly decaying ozonides was minimal; n was set to be greater than 8 for high load

curves, since there was a greater number of slowly decaying ozonides. Decreasing n

value to less than 8 would increase influence of the slower component on regression

curve.

30

The higher load of ozone resulted in oxidation of sites with slower decay rates. Sites

that required higher activation energy for oxidation resulted in formation of more stable

ozonides, contributing to a slower component. In other words, double bonds that were

harder to oxidize gave slower 1,2,3-trioxolane decay rates.

Curves with lower ozone load were fitted well with n > 10 (i.e. small “slow”

component). Regression curves for higher ozone load had difficulty fitting to

experimental points and n value constraint was brought down to n > 8. Even such

adjustment did not help regression curve to fit D data set (Figure 7 D), Experiment D had

the highest ozonation degree. Arrows H and L point to deviation of experimental points

from regression line. (Rates for curves C and D were also calculated with n >10

constraint; see Table 2)

The main purpose of the introduction of the slow exponential component was to

improve correlation between normalized and inverted normalized data sets. Appendix A

explains this issue in great detail. Normalized experimental curves were found to have

slowly rising tails and required an introduction of a slow component. The assumption was

made that the slow component should be n times slower than the fast one.

Rates calculated for lower loads of ozone, curves A and B, were 1.96 and 1.59

accordingly. Rates for higher loads of ozone were 1.27 and 1.36 for curves C and D

accordingly. Curve D could not be fitted as well as other three curves. Higher degree of

ozonation resulted in lower coefficients of determination, r2

(SigmaPlot® software

package, used for regression analysis, defines r2 as a coefficient of determination).

31

SWNT oxidation with solvated ozone. Influence of ozone load on NIR absorption

and fluorescence.

Time (sec)

0 200 400 600 800 1000 1200

No

rma

lize

d I

nte

nsity

0.0

0.2

0.4

0.6

0.8

1.0

No

rma

lize

d I

nte

nsity

0.0

0.2

0.4

0.6

0.8

1.0

954 nm

1026 nm

1123 nm

1250 nm

F1

F2

A1

A2

Figure 8. Regression analysis of normalized absorbance (A1 and A2) and fluorescence

(F1 and F2) intensities of ozonated SWNT at four distinct wavelengths. SWNT sample

was oxidized with solvated ozone. Data points for absorption and fluorescence were

acquired sequentially with 1 sec delay. Points not used in regression are depicted with

dotted lines. Regression curves are shown with solid lines. Legends are the same for the

top and the bottom plots. Used 661 nm excitation source for fluorescence measurements.

Top: high load of trioxolanes, Bottom: low load of trioxolanes.

32

Curves in Figure 8 show NIR absorption and fluorescence change with introduction

of ozone into system. Intensities dropped down and then slowly recovered to sub initial

values. Formula used for regression on NIR absorption had six parameters (F1); formula

used for normalized fluorescence data sets had five parameters (F2).

minmin1

y

ceaey

yy

n

bt

bt

final

(F1)

n

bt

bt

final ceaey

y1

(F2)

In this particular experiment, water saturated with ozone was used instead of

bubbling gaseous ozone. Absorption points after reagent addition were adjusted to

compensate for dilution. Calculated ozonide decomposition rates are summarized in

Table 3.

33

Table 3. SWNT oxidation with solvated ozone. Regression results calculated with

formulas F1 and F2 for normalized NIR absorption and fluorescence data recorded at

four distinct emission wavelengths.*

Fluorescence NIR absorption

em, nm 954 1026 1123 1250 954 1026 1123 1250

High load F1 A1

b, s -1

0.0225 0.0146 0.0100 0.0061 0.0187 0.0110 0.0076 0.0058

b, min -1

1.35 0.88 0.60 0.37 1.12 0.66 0.46 0.35

n 10.57 10.00 10.00 10.00 20.00 20.00 20.00 20.00

r2 0.9995 0.9996 0.9996 0.9993 0.9652 0.9834 0.9964 0.9974

Low load F2 A2

b, s -1

0.0480 0.0428 0.0276 0.0132 - - 0.0302 0.0158

b, min -1

2.88 2.57 1.66 0.79 - - 1.81 0.95

n 15.00 15.00 12.62 10.00 - - 20.00 20.00

r2 0.9673 0.9895 0.9959 0.9985 - - 0.9822 0.9949

* Emission or absorption wavelength is written at the head of each column. Constraints

used for regression analysis were: 0 < a < 1000; 0 < b < 100; 0 < c < a; for fluorescence

10 < n < 15; for absorption n < 20; yfinal > 1.05. Abbreviations: r2 – coefficient of

determination, n – determines how many times slow exponential term is slower than the

fast one, b – 1,2,3-trioxolane decomposition rate.

Constraint 10 < n < 15 used in fluorescence regression was needed to prevent very

low n values, leading to a greater influence of the term n

bt

ce . Such reduction of n value

led to meaningless rates b, and it was imperative to keep 'slow' component as a small

contributor to the overall intensity change. Constraint n < 20 was set for NIR absorption

regression curve. With no upper constraint for parameter n, regression on NIR absorption

data set was attempting to set abnormally large n values for nearly straight 'tails'. Greater

n value resulted in a slower second term. When n values are abnormally high, regression

results in converting the curve into a straight line, which is not the case.

34

Fluorescence rates for low ozone load were found to be at least two times faster than

those for a high load of ozone. The same was true for NIR absorption rates.

Fluorescence and absorption rates obtained from the same sample were found to be

close, but not equal. For high load of ozone, fluorescence rates were slightly higher than

those for NIR absorption. For low load of ozone fluorescence rates were slightly lower.

Observation of a close relationship between fluorescence and NIR absorption growth

rates led to the following diagram of transition states (Figure 9).

En

erg

y

-4

-3

-2

-1

0

1

2

3

4

v1

v2

c1

c2

conduction

valence

v1

v2

c1

c2

Density of Electronic States

v1

v2

c1

c2

A B C

ozonideE

11 E22

E11

E11

E22

fluorescence abs

ozonide

NIR abs

Figure 9. Schematic density of electronic states for pristine and ozonated SWNTs. Thick

solid arrows depict optical excitation and emission transitions of interest; thin dashed

arrows denote nonradiative relaxation of the electron (in the conduction band) and the

hole (in the valence band) before emission. (A) Transitions of interest in pristine SWNT.

Diagram adopted from Science, 2002, 298, 2361-2366. (B) Transitions in ozonated

SWNT. Nonradiative relaxation c1 → ozonide → v1 is a major process and shown with

thick solid arrows. Fluorescence from c1 to v1 is a minor process and shown with a dotted

arrow. (C) NIR absorption of ozonated SWNT. Depletion of an electron density of band

v1 by ozonides resulted in a weaker NIR absorption v1 → c1 (v – valence band, c –

conduction band.)

35

SWNT oxidation with gaseous ozone (high load)

Time (sec)

0 200 400 600 800 1000 1200

Norm

aliz

ed Inte

nsity

0.0

0.3

0.6

0.9

954 nm

1026 nm

1123 nm

1250 nm LH

L

H

L

A

F

Figure 10. Regression analysis of normalized NIR absorption (A) and fluorescence (F)

intensities of ozonated SWNT at four distinct wavelengths. SWNT - SDS suspension was

bubbled with O3/O2 gaseous mixture (ca. 3 v/v % ozone). Data points for absorption and

fluorescence were acquired sequentially with 1 sec delay. Points not used in regression

are depicted with dotted lines. Regression curves are shown with solid lines. Each

wavelength is marked with an individual symbol. Labels L and H denote curve wobbling

above and below regression line. Used 661 nm excitation source for fluorescence

measurements. Upper curves: NIR absorption, Lower curves: fluorescence.

Regression results for Figure 10 above are summarized in Table 4.

36

Table 4. SWNT oxidation with gaseous ozone. Regression results were calculated with

formulas F1 for normalized NIR absorption and F2 for fluorescence data recorded at four

distinct emission wavelengths*

High load Fluorescence NIR absorption

em, nm 954 1026 1123 1250 954 1026 1123 1250

b, s -1

0.0211 0.0153 0.0113 0.0064 0.0180 0.0099 0.0079 0.0051

b, min -1

1.27 0.92 0.68 0.38 1.08 0.59 0.47 0.31

n 10.00 10.00 8.00 8.00 20.00 20.00 20.00 20.00

r2 0.9995 0.9998 0.9998 0.9979 0.9824 0.9918 0.9986 0.9987

* Emission or absorption wavelength is written at the head of each column. Constraints

used for regression were: 0 < a < 1000; 0 < b < 100; 0 < c < a; for fluorescence 8 < n <

10; for absorption n < 20; yfinal > 1.05. Abbreviations: r2 – coefficient of determination, n

– determines how many times slow exponential term is slower than the fast one, b –

1,2,3-trioxolane decomposition rate.

Formation rates for SWNT ozonides

Formation of SWNT ozonide is schematically shown in Scheme 1 below

Scheme 1

OO

O

SWNT SWNTO3

O3

Formation rates of 1,2,3-trioxolanes were measured at 25 C by monitoring

absorption at ozone absmax

= 260 nm (Figure 11).

37

Time (sec)

0 1 2 3 4 5

Ab

so

rba

nce

x 1

00

0

(a.u

.)

0

1

2

b1

b2

b3

Figure 11. Absorption kinetics at 25 ºC monitored at 260 nm after bubbling O3/O2

gaseous mixture through SWNT-SDS suspension. Absorption decrease represents ozone

consumption and 1,2,3-trioxolane formation rates. Curves are shifted along the vertical

axis to bring regression lines to approximate zero with t → . Curves are: ozonation of

4x diluted SWNT (line ), ozonation of 8x diluted SWNT (line ), and ozone bubbled

through preliminary ozonated 4x diluted SWNT (line ).

Each curve in Figure 11 represents a separate experiment. The same amount of gas

(0.5 mL) with the same concentration of ozone (ca. 0.5 v/v %) were used for all

injections. SDS was found to be unreactive with small amounts of ozone and its influence

on absorption was below detection limits. There is always a possibility that small percent

of impurities from SDS can affect absorption change, thus leading to misinterpretation of

kinetics results. This is thought not to be the case in the above-mentioned experiments

(Figure 11), because pre-ozonated SWNT gave a comparable ozone

consumption/trioxolanes formation rate. Additionally, all samples were bleached to levels

below original absorbance, indicating that some double bonds were no longer existing.

Based on these facts, it is believed that the measured kinetics curves are from chemical

38

reaction between ozone and SWNT and not from some unknown impurity. Dilution 4 and

8 times of stock SWNT – SDS suspension with 1 wt. % aq. SDS was necessary to acquire

sufficient data points for regression.

1,2,3-Trioxolane formation rates are summarized in Table 5.

Table 5. SWNT ozonide formation rates.

Curve b, s –1

r2 ccarbon, mg/L cdouble bond, mmol/L

4x diluted SWNT 1.44 0.8936 1.44 0.06

8x diluted SWNT 0.63 0.8498 0.72 0.03

4x dil. prelim. ozntd SWNT* 0.52 0.9475 1.44 0.06

* - Preliminary ozonated tubes were heated to 40 C for ca. 4 hours before their use in this

experiment; b – rate.

Rate for 8 times diluted SWNT – SDS suspension was found to be at least two times

slower than the one for 4 times diluted sample. Concentration of SWNT for 4 and 8 times

diluted suspensions are summarized in Table 5. Total concentration of double bonds in 4

times diluted sample was calculated 0.06 mmol/L, yielding the reaction rate constant

2.4·104 M

-1s

-1, which is of the same order of magnitude as the formation rate constants of

C60O3 (8.8 104 M

-1s

-1 at 0 C) and C70O3 (5 10

4 M

-1s

-1 at 22 C) in CCl4 solvent.

18, 19

The rate constant for 4 times diluted SWNT suspension is five orders of magnitude

slower than the diffusion rate constant 109 M

-1s

-1 for small organic molecules in hexane.

It is likely that water viscosity, solvation of ozone with water molecules, SDS

hydrophobic shell around SWNT, large molecular weight of the tubes and tubular

structure with large aspect ratio, all contributed to the rate decrease.

39

Establishing of a saturation limit with different amounts of ozone

Absorption changes after injections of different amounts of ozone into diluted

SWNT – SDS suspension are shown in Figure 12. SWNT - SDS suspension was diluted

eight times to decrease reaction rate between SWNT and ozone. Ozone concentrations

were approximately 0.5%, 0.6%, 0.75% and 1% in air stream (curves , , ,

accordingly). Ozone concentration was manipulated by dilution with air. A half milliliter

of O3 - air gaseous mixture was injected into 1 mL of SWNT – SDS suspension in each

case. All measurements were performed in a thermostated cuvette at 25.0 C.

Time (sec)

0 20 40 60 80 100 120

Absorb

an

ce x100 (

a.u

.)

7.6

7.8

8.0

8.2

8.4

3x

4x

5x

6x

O3

/ O2 mixture

dilution with air:

d8

saturation with ozonides

ozone consumption

Figure 12. Dependence of SWNT absorption on the amount of injected ozone. Each

curve represents a separate experiment. Kinetics curves were monitored at 260 nm and

25.0 C. Symbols are labeled with O3/O2 mixture dilution degrees. The difference

between initial and final absorbance is marked with d8.

Curves were shifted along the time axis to set injection point to zero seconds. Three

of the four curves were multiplied by the corresponding coefficients to bring the initial

40

absorbance to the same level (difference in absorbance before adjustment was very small;

initial absorbance ranged between 0.0737 and 0.0761 a.u.)

Spikes at 0 seconds are due to needle insertion and shown with dotted lines.

Exponential decrease of absorbance right after the injection represents ozone

consumption and 1,2,3-trioxolane formation.

Higher concentrations of ozone resulted in identical downward step (value d8 in

Figure 12), indicating that tubes got saturated with ozonides. Exponential decay is

schematically divided into two sections: saturation of SWNT with ozonides (left) and

ozone consumption by „emptied‟ sections of SWNT (right). Value d8 equaled to 0.0017

a.u. was found approximately 9 times smaller than downward step d in non-diluted

SWNT suspension (Figure 13). Addition of 0.5 mL of 6x diluted O3/O2 gaseous mixture

was not sufficient to saturate SWNT with 1,2,3-trioxolanes. Rate of reaction between

SWNT and the least concentrated ozone/air gas mixture (curve ) was calculated to be

b = 0.63 s-1

, corresponding to lifetime = 1.6 sec (see Figure 12 above for details.)

41

Influence of multiple ozone injections on SWNT saturation. Oxidation with 4 min

intervals

Time (min)

10 20 30

Ab

so

rba

nce

(a

.u.)

0.54

0.57

0.60

b6 b7b8

b5

b4

b3

b2

b1

d

Figure 13. Injections of 1.5 mL of O3/O2 gaseous mixture into 1.5 mL SWNT - SDS

suspension with 4 min time intervals. Suspension absorbance was monitored at 260 nm

and r. t. Upward spikes are due to needle insertion and are emphasized with arrows.

Difference in absorbance before and after the first injection is marked with letter d.

Difference in absorbance between before and after the first ozone injection was

d = 0.0154 a.u.(ca. 3%). This value is approximately 9 times larger than that for 8x

diluted SWNT suspension. Rates noted as b1 through b8 in Figure 13 are summarized in

Table 6.

Increase in rate b and decrease in the absolute value of step d (marked on Figure 13)

are believed to be associated with SWNT saturation. Experimental points in Figure 13

above, shown with circles, form simple exponential decay curves right after each

injection of ozone. This decay represents reaction of freely floating ozone with SWNT

and conversion of ozone to oxygen by collision with SDS and water molecules.

42

Table 6. Regression results for single exponential decay of absorbance after multiple

ozone injections into SWNT – SDS suspension.

Injection cycle*

1st 2

nd 3

rd 4

th 5

th 6

th 7

th 8

th

b, min-1

16.60 10.68 7.51 4.95 3.55 3.23 2.91 2.93

, sec 3.6 5.6 8.0 12.1 16.9 18.6 20.6 20.5

r2 0.9917 0.9971 0.9987 0.9987 0.9988 0.9994 0.9992 0.9988

* - Regression was performed with a single exponential decay formula. Variables: b –

decay rate, - lifetime, r 2

– coefficient of determination.

SWNT oxidation with different amounts of ozone. Estimation of a saturation limit

by NIR fluorescence.

Wavelength (nm)

950 1050 1150 1250 1350

Flu

ore

sce

nce

In

tensity

(nW

/nm

)

0.00

0.05

0.10

0.15

before O3

0.5 mL

1 mL

2 mL

3 mL

4 mL

5 mL

6 mL

8 mL

10 mL

1265

0.03

0.05

Figure 14. Fluorescence spectra after bubbling specific amounts of ozone through

SWNT-SDS suspensions. Each curve represents a separate experiment. One percent

aqueous SDS solution ozonated with 10 mL of O3/O2 gaseous mixture (ca. 3 v/v %

ozone) served as a baseline for all experiment. Samples were ozonated 3 days before

fluorescence acquisition. The 785 nm laser was used for excitation. Arrows point to the

saturation level. A zoom-in shows that saturation was reached with 3 mL of O3/O2

gaseous mixture (ca. 3 v/v % ozone; curve ). Tube (6,5) with emmax

977 nm was the

most difficult to oxidize.

43

SWNT suspension aliquots were bubbled with different amounts of O3/O2 gaseous

mixture (0.5 – 10 mL) and spectra overlaid (Figure 14). Gas was injected slowly in each

case. To avoid misinterpretations, spectra were recorded three days after ozonation,

which was plenty of time for ozonide decomposition and SWNT structural

rearrangements. One percent aqueous sodium dodecyl sulfate solution bubbled with

10 mL of O3/O2 gaseous mixture (ca. 3 v/v % ozone) served as a baseline for all curves in

Figure 14. Spectra overlay demonstrated that SWNT got saturated with ozone at 3 mL

O3/O2 gaseous mixture load (ca. 3 v/v % ozone). Curves for 8 and 10 mL have less

intense fluorescence than all other ones. It may be concluded that during slow O3/O2

gaseous mixture injection some of the ozonides have decomposed, thus allowing for a

greater amount of ozone to react with SWNT. Separate UV studies demonstrated that

after SWNT got saturated with 1,2,3-trioxolanes, which occurred at or below 3 mL load,

excess of ozone dissolved in aqueous media (Figure 13). This in turn provided an

additional supply of ozone for subsequent oxidation. Because collision of ozone with

SDS molecules leads to its conversion to oxygen (ozone decay rate in 1% aq. SDS is

about 0.43 min-1

at r. t.), slow addition of 8 and 10 mL provided sufficient amount of

ozone to overcome deactivation by SDS, hydroxyl and water species. Typically, injection

of 10 mL of O3/O2 gaseous mixture required more than a minute to complete.

Notably, tube (6,5) with emmax

977 nm (diameter = 0.76 nm) had the least ozonation

degree, i.e. was the most difficult to oxidize. This tube has a very small twist when

compared to other tubes in the experiment.

3D Structures of tubes of interest are shown in Figure 15.

44

(8,3) (6,5) (10,2) (11,3) (10,5)

Figure 15. Tubes (n,m) with well separated emission peaks on spectrum for 785 nm

excitation source. Each tube is shown in two projections (top and bottom). Tube number

is indicated below each pair. Tubes are drawn not to scale.

The physical and optical properties of tubes that were well separated in emission

spectrum with 785 nm excitation are summarized in Table 7.

Table 7. Summary of physical and optical properties of tubes with well separated peaks

in emission spectrum*

Tube (n,m) emmax, nm , cm

-1 diameter, nm

8,3 954 10486 0.78

6,5 977 10234 0.76

10,2 1056 9468 0.88

11,3 1201 8327 1.01

10,5 1253 7982 1.05

*- Used 785 nm excitation source.

The only tube that had difficulty getting ozonated with 10 mL of O3/O2 gaseous

mixture (ca. 3 v/v % ozone) was tube (6,5). It has the smallest “twist” out of all well

defined tubes in emission spectrum (Figure 14). It may be concluded that the tube “twist”

increases double bond reactivity with ozone. Tube (6,5) was also estimated to be the most

abundant in utilized HipCo sample (see Chapter 1 for tube abundance distribution). All

45

tubes had substantial difference in diameter and no conclusion could be made with regard

to dependence of ozonation degree on the tube diameter at 10 mL O3/O2 gaseous mixture

load (ca. 3 v/v % ozone). Particularly, tubes (8,3) and (6,5) had very close diameter, but

different oxidation degree, as evidenced by peaks at 954 and 977 nm (Figure 14).

SWNT oxidation with different amounts of ozone. Estimation of a saturation limit

by NIR absorption.

NIR absorption was measured on samples discussed above. For experimental details

see Figure 14 above and accompanying notes.

Wavelength (nm)

950 1050 1150 1250

Absorb

an

ce (

a.u

.)

0.15

0.20

0.25

before O3

0.5 mL

1 mL

2 mL

3 mL

4 mL

5 mL

6 mL

8 mL

10 mL

Figure 16. NIR absorption spectra after bubbling specific amounts of O3/O2 through

SWNT-SDS suspensions. Each curve represents a separate experiment. One percent

aqueous SDS ozonated with 10 mL of O3/O2 served as a baseline for all experiment.

Samples were ozonated 3 days prior to NIR absorption acquisition. Arrows point to the

saturation level, which was reached at 2 mL of O3/O2 gaseous mixture (ca. 3 v/v %

ozone; curve ). The area near 977 nm was the most difficult to oxidize and had the

least percent decrease.

46

Analogously to fluorescence spectra, the area near 977 nm had the least decrease in

absorbance as compared to percent values for all tubes. Saturation point was reached with

2 mL O3/O2 gaseous mixture (ca. 3 v/v % ozone; pointed with arrows in Figure 16). To

avoid misinterpretations, NIR absorption spectra were recorded three days after

ozonation.

Noisiness of spectra starting from 2 mL is presumed to be associated with production

of a number of nonequivalent sections of tubes. Notably, there are no peak shifts between

2 and 10 mL of O3/O2 gaseous mixture (ca. 3 v/v % ozone). This means that ozonation is

following a specific pattern, rather than a random one. An increase in the number of

peaks would be expected for random ozonation of SWNT. Despite significant decrease in

absorption (compare curves and ), the number of peaks and their absmax

were

preserved, thus indicating an ordered oxidation. Higher loads of ozone (within 10 mL

O3/O2 gaseous mixture) are thought to produce a greater number of „sections‟ of SWNT

ozonated with the same pattern.

SWNT oxidation with different amounts of ozone. Estimation of a saturation limit

by UV-Vis absorption.

UV-Vis spectrum of 1% aq. SDS was found to be unchanged in a region 235 - 800

nm after bubbling with 10 mL of O3/O2 gaseous mixture. Aqueous SDS solution purged

with 10 mL of O3/O2 gaseous mixture served as a baseline for all spectra in Figure 17.

47

Wavelength (nm)

250 350 450 550 650 750

Absorb

ance (

a.u

.)

0.2

0.4

0.6

245 285

0.55

0.59

0.63 before O3

0.5 mL

1 mL

2 mL

3 mL

4 mL

5 mL

6 mL

8 mL

10 mL

730

0.19

0.21

Figure 17. UV-Vis absorption spectra after bubbling specific amounts of ozone through

SWNT-SDS suspensions. Each curve represents a separate experiment. One percent

aqueous SDS ozonated with 10 mL of O3 served as a baseline for all experiment. Samples

were ozonated 3 days before UV-Vis absorption acquisition. Arrows point to the

saturation level, which was reached at or below 2 mL of O3/O2 gaseous mixture (ca. 3 v/v

% ozone; curve ).

SWNT spectra had smooth transition from 0 mL (curve ) spectrum to 10 mL one

(curve , Figure 17).

SWNT ozonation for a specific period of time. Influence of ‘saturation’ and

1,2,3-trioxolane decomposition rates on overall sidewall oxidation as monitored by

NIR fluorescence.

Ozonation of SWNT-SDS suspensions was conducted for specific periods of time,

ranging from 30 sec to 30 min. One percent aq. solution of sodium dodecyl sulfate,

bubbled with ozone for specified periods of time served as a baseline for each curve (i.e.

each curve had its own baseline). Interesting spectral changes were observed and are

summarized in Figure 18.

48

Wavelength (nm)

950 1050 1150 1250 1350

Flu

ore

sce

nce

In

ten

sity (

nW

/nm

)

0.00

0.04

0.08

0.12

0.16before O3

0.5 m

1 m

2 m

4 m

5 m

10 m

30 m

Figure 18. Influence of ozonation on SWNT fluorescence spectra. Each curve represents

a separate experiment. Symbol legends denote ozone bubbling times in minutes. Spectra

were recorded 3 days after ozonation. Ozone was bubbled through samples at room

temperature. The 785 nm laser was used for excitation. Arrows point to fluorescence

curves after 0.5 and 5 min of continuous bubbling of O3/O2 gaseous mixture (ca. 3 v/v %

ozone) through SWNT-SDS suspension. Curves for 5, 10 and 30 min and before O3 are

shown with thick lines.

Two sets of pairs of arrows in Figure 18 demonstrate location of curves after 30 sec

and 5 min of continuous bubbling of O3/O2 gaseous mixture. The gas flow rate was

approximately 26 mL/min. This means 0.5 and 5 min bubbling correspond to 13 and 130

mL of O3/O2 gaseous mixture.

The upper left arrow points to tube (6,5) which was found to be fairly robust to

ozonation with 10 mL of O3/O2 gaseous mixture (ca. 3 v/v % ozone). In terms of percent

values, intensities of all tubes except (6,5) were substantially bleached within 30 sec of

continuous bubbling. Notably, tube (8,3), with roughly the same diameter as (6,5), was

bleached more than (6,5) after 30 sec ozonation. (Tube properties are summarized in

49

Table 7). Interestingly, after 5 min of bubbling of O3/O2 gaseous mixture, tube (6,5) is

nearly gone, while the ones emitting at longer wavelengths are still present. Comparing

fluorescence intensities at 977 and 1255 nm for curves and in Figure 18, one may

believe that tube (6,5) had a high degree of aromaticity, or conjugation of double bonds,

thus making it difficult to oxidize. Once that conjugation was disrupted, oxidation of this

particular tube was comparable to those with similar diameters.

Fluorescence of SWNT with shorter emmax

was much weaker than those with longer

emission wavelengths after 2 minutes of continuous bubbling of O3/O2 gaseous mixture

(curve ). It is believed that 1,2,3-trioxolanes formed on tubes emitting in a range 900-

1100 nm decayed faster then those emitting at longer wavelengths (as evidenced by

comparison 30 sec and 2 min curves, and ).

After saturation limit of trioxolanes on the surface of SWNT was reached (curve )

tube (8,3) with emmax

954 nm decreased a lot more than those emitting at longer

wavelengths. It was concluded that tube diameter is one of the major factors affecting

rates of trioxolane decomposition, thus allowing for more ozonides to be formed on the

SWNT surface.

Faster decay rates lead to a greater number of ozonides formed, resulting in a higher

turnover within the same period of time. After 5 min of continuous bubbling fluorescence

at 954 nm is negligible, but at 1150 nm and longer wavelengths is only a little less than in

the spectrum for 30 seconds bubbling.

50

SWNT ozonation for a specific period of time. Influence of ‘saturation’ and 1,2,3-

trioxolane decomposition rates on overall sidewall oxidation as monitored by NIR

absorption.

NIR absorption was measured on samples that were bubbled with O3/O2 gaseous

mixture (ca. 3 v/v % ozone) for specific periods of time, as discussed in the section

above. In the same manner, aq. SDS solutions bubbled with ozone for noted periods of

time served as a baselines in each case (i.e. each curve had its own baseline).

Wavelength (nm)

950 1050 1150 1250

Absorb

ance (

a.u

.)

0.05

0.10

0.15

0.20

0.25

before O3

0.5 m

1 m

2 m

4 m

5 m

10 m

30 m

1 hr

Figure 19. Influence of ozonation on SWNT NIR absorption spectra. Each curve

represents a separate experiment. Symbol legends denote ozone bubbling times in

minutes. Spectra were recorded 3 days after ozonation. Ozone was bubbled through

samples at room temperature. Arrows point to curves after 0.5 and 10 min of continuous

bubbling of O3/O2 gaseous mixture (ca. 3 v/v % ozone) through SWNT-SDS suspension.

Bleaching of NIR absorption of ozonated SWNT was less informative when

compared to fluorescence spectra. Saturation of SWNT with trioxolanes was reached at

or before 30 sec of continuous bubbling of O3/O2 gaseous mixture (ca. 3 v/v % ozone).

51

Initial bleaching at shorter wavelengths was less than that at longer ones. As evidenced

from comparison of the curves before and after 30 sec of ozone bubbling (curves and

), the change in absorbance at 977 nm was only a fraction of that at 1123 nm or 1250

nm.

After 10 min of continuous bubbling, all characteristic peaks below 1100 nm

disappeared. This means that faster decomposition rates of 1,2,3-trioxolanes of thinner

tubes resulted their destruction, while thicker ones were still maintaining their

characteristic absorption. The peaks maxima did not shift after 0.5 min of ozonation and

the curve stayed smooth (curve ). This means that NIR absorption peaks in Figure 19

above are for a small section of a tube. Bubbling for one minute resulted in a very „noisy‟

curve. This meant that after one minute of continuous bubbling damage to tubes was so

great that it finally manifested itself in NIR spectra. It is still not clear how large these

sections were.

SWNT ozonation for a specific period of time. Influence of ‘saturation’ and

1,2,3-trioxolane decomposition rates on overall sidewall oxidation as monitored by

UV-Vis absorption.

UV-Vis absorption was measured on samples that were bubbled with O3/O2 gaseous

mixture (ca. 3 v/v % ozone) for specific periods of time, as discussed in two previous

sections. In the same manner, aq. SDS solutions bubbled with ozone for a specific period

of time served as a baseline for each experiment (each curve had its own baseline).

Below are overlaid 1% aq. SDS spectra after bubbling for specified periods of time.

With regard to baseline subtractions, SDS absorption increased in the range 250 – 330

52

nm to acceptable levels, i.e. it was sufficiently low. At wavelengths below 250 nm it

increased greatly and that region will not be discussed in this work.

Wavelength (nm)

300 400 500 600 700

Ab

so

rba

nce

(a

.u.)

0.2

0.3

0.4

0.5

0.6 before O3

0.5 m

1 m

2 m

3 m

4 m

5 m

10 m

30 m

1 hr

Figure 20. UV-Vis spectra of 1% aq. SDS bubbled with O3/O2 gaseous mixture (ca. 3 v/v

% ozone) for noted periods of time.

SDS was found to be sufficiently robust to oxidation with ozone and could be used

as baseline reference for wavelengths above 250 nm.

Below are overlaid UV-Vis spectra of SWNT ozonated for noted periods of time

(Figure 21). SWNT spectrum obtained before ozonation is provided for reference.

53

Wavelength (nm)

300 400 500 600 700

Ab

so

rba

nce

(a

.u.)

0.1

0.2

0.3

0.4

0.5

0.6 before O3

0.5 min

1 m

2 m

3 m

4 m

5 m

10 m

30 m

1 hr

310 330

0.30

0.35

0.40

0.45

0.50

Figure 21. Influence of ozonation on UV-Vis absorption spectra of SWNT. Each curve

represents a separate experiment. Symbol legends denote ozone bubbling times in

minutes. Spectra were recorded 3 days after ozonation. Ozone was bubbled through

samples at room temperature. Arrows point to curves 3, 4 and 5 min of continuous

bubbling of O3/O2 gaseous mixture (ca. 3 v/v % ozone) through SWNT-SDS suspension.

SWNT were found to give essentially the same UV-Vis spectra for 3, 4 and 5 min of

continuous ozonation. Three minutes of bubbling are equivalent of a slow injection of

about 78 mL of O3/O2 gaseous mixture (ca. 3 v/v % ozone). It is not clear why there was

a „freeze‟ in spectral changes after 3 min. With small ozone loads, saturation point for

SWNT was found to be around 0.5 min of bubbling as described in the sections above.

SWNT spectra had smooth transition from 0 min spectrum to 30 min one.

54

Ozone consumption as monitored by absorption at 260 nm

Ozone consumption after the first injection was monitored near its absmax

= 260 nm

(Figure 22).

240 250 260 270 280

Absorb

an

ce (

a.u

.)

0.52

0.54

0.56

0.58

0.60

0.62before O3

13.7 sec

27.5 s

41.2 s

55.0 s

68.7 s

Wavelength (nm)

240 250 260 270 280

0.00

0.02

0.04

Figure 22. UV spectrum change with ozone consumption. Left: spectra as recorded.

Right: spectra after subtraction of pristine SWNT spectrum (). Symbols are the same

for both graphs and denote times after the injection of O3/O2 gaseous mixture. One

percent aq. SDS served as a baseline. Curves , and are shown with thick lines.

One and half milliliter of O3/O2 gaseous mixture (ca. 3 v/v % ozone) was injected

into 1.5 mL SWNT – SDS suspension. Injection was done during 2 second interval

between curves and in Figure 22 (left side).

Decrease of ozone absorption was several times greater than that of SWNT (compare

curves , and in Figure 22). SWNT bleaching is the difference between absorbance

before and after SWNT ozonation. In absolute values SWNT bleaching at 260 nm was

0.0028 a.u. between curves and . Ozone absorbance change was approximately

0.047 a.u., or 17 times greater than SWNT bleaching. Sodium dodecyl sulfate, a

55

surfactant, also contributed to ozone decomposition (ozone decay rate in 1% aq. SDS is

about 0.43 min-1

at r. t.).

It was established that ozonide decomposition results in increase of SWNT

absorbance. At 260 nm this rise is barely noticeable. Effect on absorbance is more

pronounced with longer wavelengths. Typically, absorption curve plotted against time

would have a fast rising beginning and a slowly changing tail. Initial change in

absorbance is attributed to 1,2,3-trioxolanes decomposition. Tail part within the first 30

min after ozonation is attributed to a combination of structural rearrangements of SWNT

and decomposition of slowly decaying ozonides (section ii in Figure 23).

Ozonide decomposition was studied by UV-Vis at 25 and 40 C. To avoid possible

chemical transformations, the longer wavelength, 735 nm, was chosen for monitoring. A

time period of five hours at 25 C was not sufficient to reach the maximum absorbance.

Temperature was increased to 40 C and the experiment was repeated (Figure 23).

56

Time (min)0 100 200 300 400

Ab

so

rba

nce

at

73

5 n

m (

a.u

.)

0.17

0.18

0.19

0.20

0.21

5 6 7

0.17

0.18

0.19

0.20

0.21

50 150 250 350

0.198

0.199

0.200

0.201

i ii iii

Figure 23. SWNT absorbance change at 40 C after ozonation. The point of injection is

emphasized with an arrow. An upward spike at the time of injection is due to a needle

insertion. 2.5 mL of O3/O2 gaseous mixture was bubbled through 1 mL of SWNT

suspension over a period of 20 sec. A fast changing beginning and a slowly changing

„tail‟ are zoomed for clarity. An approximate place of absorbance maximum after the

injection is marked with a dotted vertical line. Plot is schematically divided into: (i) – fast

O3 decay, (ii) – slow O3 decay and rearrangements, (iii) – rearrangements.

Heating for 6 hours at 40 C was needed to bring absorbance at 735 nm to its

maximum value. It is presumed that the slow change is associated with SWNT structural

changes and migration of double bonds. Absorbance at 735 nm after recovery was 4%

lower than that recorded before ozonation. This result is in good agreement with SWNT

absorption decrease measured in UV region. Ozonation of SWNT resulted in 3 – 4 %

absorbance bleaching at 260 nm.

57

Establishing an average rate of ozone consumption

Time (min)

0 5 10 15 20 25

Absorb

an

ce (

a.u

.)

0.0

0.2

0.4

0.6

0.8 Abs (H2O/O3)

Abs (diluted H2O/O3)

b1

b2

SDS

SWNT - SDS

Figure 24. Estimation of an average rate of ozone consumption by SWNT. Absorption

was monitored at 260 nm and 25 C. Curves and represent two separate

experiments. A suspension of SWNT was added to an aqueous solution of ozone (curve

). The second experiment included an injection of an equivalent volume of an aqueous

SDS solution. The points of injection are marked with arrows. Expected absorbance level

for curve after dilution is marked with a horizontal dash line. Experimental points used

for regression analysis are shown with thick lines. Experimental points not used in

regression are shown with dotted lines. Regression lines are depicted with thin black

lines.

Figure 24 demonstrates a difference between rates of ozone decay in presence (curve

) and absence of SWNT (curve ). Two separate experiments were designed to

compare rates of ozone consumption obtained by UV and NIR fluorescence

measurements. The first experiment included injection of 700 uL of SWNT-1% SDS aq.

suspension into 800 uL of an aqueous solution of ozone. The initial absorbance of ozone

solution in the first experiment was above 0.8 a.u. Injection of SWNT suspension is

marked with an arrow (Figure 24, curve ). Dashed lines on the same figure are shown

58

to visually demonstrate the predicted absorbance after dilution. The selected points for

regression analysis are shown with thick solid lines. Resulted regression curves are

shown with thin black lines. Experiment was repeated with an injection of 700 uL of 1%

aq. SDS solution only (curve ). In the case of the second experiment initial ozone

absorbance was only 0.65 a.u.; this was sufficient for the purpose of the experiment.

Preparation of an ozone solution for the second experiment with the absorbance of 0.8

a.u. was not necessary. Sections of the curves before injection have gradual slopes,

common for slow ozone decay in water. Regression results are summarized in Table 8

below.

An upper boundary for absorbance after reagent injection was calculated in

accordance with the following formula: ofbid AAAkAA

where dA – absorbance after dilution, iA – absorbance at the time of injection, bA – SWNT

bleaching at saturation point ( bA = 0.008 a.u.; cf. Figure 13), fA – final absorbance, k – dilution

coefficient, k = 0.533, oA – absorption of consumed ozone.

Estimation of the upper boundary was needed to determine which points must not be

included in regression analysis. Due to rapidly changing exponential decay immediately

after ozonation, there was no clear division between the effects of dilution and

absorbance decay. The decrease of ozone absorption due to the formation of ozonides is

thought to be several times larger than SWNT bleaching (see Figure 22), but the exact

number is difficult to estimate.

59

Table 8. Regression results for single exponential decay of absorbance after injections of

SWNT – 1 wt. % aq. SDS suspension and of 1 wt. % aq. SDS solution into aqueous

solutions of ozone.

b, min-1

, min r2

dA fA

SWNT-SDS (b1) 1.65 0.61 0.9978 0.6981 – oA 0.265

SDS (b2) 0.29 3.45 0.99997 0.3602 – oA 0.015

b diff 1.36 0.74

* - Regression was performed with a single exponential decay formula. Variables: b –

decay rate, - lifetime, r 2

– coefficient of determination, dA – absorbance after dilution,

fA – final absorbance, oA – absorption of consumed ozone.

An amount of ozone in aqueous solution with absorbance 0.8 a.u. exceeds many

times what could react with SWNT at any given time. Bleaching of SWNT, i.e. decrease

of absorbance due to reaction with ozone, was estimated to be near 0.008 a.u. based on

the data obtained in previous studies (this chapter, cf. Figure 13). The final absorbance of

SWNT – SDS suspension after full consumption of ozone was found to be fA = 0.265

(Table 8). The final absorbance fA of SWNT-SDS suspension would be higher if water

had no oxidizer in it. This implies that only 0.008/0.265 = 3 % of a total number of

double bonds were consumed in excess of O3 (i.e. within ca. 3 min after the injection,

curve in Figure 24). The saturation point was reached within seconds after injection

(cf. formation kinetics in Figure 11). A number of consumed double bonds is likely to be

higher than 3%, but it is still a very small number. Vacuum deoxygenation of SWNTO3

gave an estimate of 0.2 % of double bonds were consumed by ozone within 1 min of

reaction at 5 C in solid state (this chapter, Figure 1). Fast decay of curve (Figure 24)

60

is mainly due to ozone consumption by SWNT or decomposition by collision with SDS,

and not because of SWNT bleaching (Figure 22).

An assumption was made that about the same number of collisions occurred between

molecules of ozone and SDS in both experiments, thus making it possible to extract the

rate of ozone consumption by SWNT via subtraction of rate b2 from b1. The difference is

b diff = 1.36 min-1

, which is somewhere in between the rates obtained for fluorescence

measurements at four distinct wavelengths: 954, 1026, 1123 and 1250 nm.

It is important to keep in mind that the decomposition of 1,2,3-trioxolanes is a rate

determining step in the formation of subsequent or “new” ozonides after the previous

ones underwent decomposition. This is always true when the amount of ozone in the

system exceeds the saturation limit. SWNT samples saturated with 1,2,3-trioxolanes were

found to have 3 to 4 % of the total number of double bonds converted to ozonides. The

decomposition rates obtained from fluorescence studies were employed in the

interpretation of UV studies. Particularly, the fluorescence recovery rate was found to

increase exponentially with a decrease in tube diameter (plot is not shown). Thus, the

faster decay rates were observed for thinner tubes.

Metallic tubes are believed to react with ozone in the same fashion the

semiconducting tubes do, though they do not fluoresce. The overall rate b diff = 1.36 min-1

indicates that metallic tubes may have similar distribution of tube diameters.

More in-depth investigation is needed to study differences between UV results

recorded at 260 nm and fluorescence data to make a better estimation of how many and

what types of metallic tubes are present in the sample. As noted earlier, metallic tubes do

61

not fluoresce, and their quantitative analysis needs significant improvements. Currently,

there is no fast and efficient way to estimate distribution of metallic tubes in a sample.

An influence of the injection period on SWNT saturation with 1,2,3-trioxolanes

Time (min)

2 10 18 26 34

0.54

0.57

0.60

0.63

16 min

Absorb

ance a

t 260 n

m (

a.u

.)

0.54

0.57

0.60

0.63

8 min

0.54

0.57

0.60

0.63

4 min

b4

b8

b16

d4

d8

d16

Interval:

3rd

3rd

3rd

Figure 25. An influence of the injection time interval on SWNT saturation with 1,2,3-

trioxolanes. Three curves (top, middle and bottom) are three separate experiments.

Reactions were monitored by absorption at 260 nm and 25 C. Experimental points are

shown with circles. Upward spikes are due to needle insertion into cuvette at the time of

injection of O3/O2 gaseous mixture. First two injection points are emphasized with thin

black arrows. All three experiments had multiple injections with intervals: 4 min (top,

), 8 min (middle, ) and 16 min (bottom, ).

62

The same amount of ozone was injected into SWNT – SDS samples with different

time intervals. Results are summarized in Figure 25. SWNT suspension was subjected to

multiple ozonation with different time intervals. Ozone consumption rates were

calculated for an injection at or near 34 min (Figure 25).

Nine injections of 1.5 mL of O3/O2 gaseous mixture into 1 mL of SWNT – SDS

suspension with 4 min intervals resulted in a small upward step (marked as d4;

d4 = 0.0002 a.u.) and ozone consumption rate b4 = 3.42 min-1

(Figure 25). Extending time

interval to 8 min gave a larger downward step, d8 = 0.0033 a.u., and faster ozone

consumption rate, b8 = 3.95 min-1

(Table 9). A sixteen minute interval further increased

the rate, b16 = 4.63 min-1

, and increased the step to d16 = 0.0068 a.u. These results indicate

that neither four nor eight minutes are sufficient to fully decompose all ozonides.

Additional experiments would be of benefit to determine if 16 min is long enough to

decompose all trioxolanes. Separate vacuum deoxygenation studies indicated that oxygen

from ozonide decomposition can evolve for 20 minutes before deoxygenation amount per

minute falls below system leak threshold. The downward step d18 after the third injection

(curve ) was larger than the third steps on two other curves ( and ).

63

Table 9. Results summary for multiple injections of O3/O2 gaseous mixture into SWNT –

SDS suspensions with different time intervals*

3rd

injection at 34 min

Interval, min d, a.u. d, a.u. b, min-1

, min r2

4 0.0061 (0.0002) 3.42 0.29 0.9991

8 0.0040 0.0033 3.95 0.25 0.9988

16 0.0068 0.0068 4.63 0.22 0.9969

* - Regression was performed with a single exponential decay formula. Variables: d –

SWNT bleaching step, b – decay rate, - lifetime, r 2

– coefficient of determination.

Saturation of SWNT with 1,2,3-trioxolanes by reaction with solvated ozone

Time (min)

0.0 0.5 1.0 1.5

Ab

so

rba

nce

at

26

0 n

m (

a.u

.)

0.4

0.5

0.6

s

s

a3

a2

a1

d1

d2

Figure 26. Additions of water saturated with ozone to SWNT – SDS suspension.

Absorbance was monitored at 260 nm. Points of injection are emphasized with arrows.

Upward spikes caused by mixture stirring are marked with letter s. An expected

absorbance change due to dilution is shown with levels d1 and d2 (dotted lines) for

comparison.

The first addition of 0.2 mL of water saturated with ozone (absorbance ca. 0.85 a.u.)

to SWNT suspension (1 mL) resulted in previously seen ozone consumption „tail‟,

64

indicating that amount of ozone exceeded SWNT saturation limit. Observed absorbance

after the first injection was a2 = 0.4811, which is 0.0162 a.u. less than the expected

absorbance d1 when accounted for dilution. In terms of percent values, SWNT

absorbance was bleached by 1 - a2/d1 = 3.3 % at 260 nm. This is a very small number of

double bonds. It is believed that electron withdrawing by 1,2,3-trioxolanes has a

profound influence on tube ability to react with molecules of ozone. The second injection

of aqueous ozone gave even slower rate of ozone consumption, indicating tube saturation

with ozonides. An overall SWNT bleaching after the second injection was 1 - a3/d2 = 3.1

%. The error in these calculations may be significant, due to the fact that ozone slowly

reacts with SDS with an increase in absorbance at 260 nm. The more SWNTs are

saturated, the higher the percentage of SDS molecules will react with ozone. While the

error estimation is a subject for further discussion, the fact that only a small percent of

double bonds could be converted into trioxolanes is undeniable.

Structural changes of SWNT after ozonation

HipCo SWNTs were found to release a large amount of heat when reacted with

ozone at room temperature. Typically, such heating results in SWNT burning. Enthalpy

estimation for SWNT-O3 and SWNT and O3 with PM3 RHF method supports this

experimental observation (data not provided). Below are 3D pictures of semiconducting

tube (8,0) with three ozonides next to each other (Figure 27):

65

Figure 27. PM3 RHF optimized lateral (C) and longitudinal (A and B) trioxolanes on the

surface of semiconducting tube (8,0). Directions of ozonides are shown with dotted lines

and marked a, b and c for corresponding ozonides. Tube diameters of pristine and

ozonated sections are marked with labels s1 and s2 correspondingly. (Top) Diagonal

projection. (Bottom) Side projection. Optimization results are courtesy of S. Ghosh and S.

Bachilo.

An optimization of oxidized SWNT yielded tube widening in the area of ozonides

due several sp3 hybridized carbons. A longitudinal trioxolane, optimized with PM3

method for tube (10,0), had sp3 carbons significantly above the surface of SWNT

(marked with distance d2 in Figure 28). Carbon-carbon bond in ozonide was found longer

than what would be expected for a single C-C bond; a further optimization may be

needed.

Figure 28. An optimized structure of semiconducting tube (10,0) with a longitudinal

1,2,3-trioxolane on its surface. PM3 method was used for calculations. Shown are two

different projections. Optimization results are courtesy of S. Bachilo.

66

Ozonide decomposition will ultimately lead to a formation of an epoxide.

Semiconducting tube (8,0) with a lateral epoxide was optimized with PM3 RHF method

(Figure 29)

Figure 29. An optimized structure of semiconducting tube (8,0) with a lateral epoxide on

its surface. An optimization yielded elevation of sp3 carbons of epoxide above the surface

of the tube (distance d2). Used PM3 method for calculations. Shown are three different

projections. Optimization results are courtesy of S. Ghosh and S. Bachilo.

Previously performed optimization of ozonated fullerenes24

gave essentially the same

elevation of sp3 carbon atoms of an epoxide. Two fullerene structures, with „open‟ and

„closed‟ epoxides are shown of Figure 30. In both cases ratio d2/d3 was found to be

greater than 1, indicating fullerene distortion, or a “squeeze.”

67

A B C

D E F

Figure 30. PM3 optimized open (oxidoannulene) and closed (epoxide) structures of

C60O. d2 – height, d3 – width; (A-C) – an open form of C60O; (D-F) – a closed form of

C60O. Three different projections are shown for both structures. Optimization results are

courtesy of S. Bachilo.

Interaction of visible and NIR light with SWNT

SWNTs are black in color and seemed to absorb electromagnetic waves throughout

the entire electromagnetic spectrum. Spectra presented in this work covered UV, NIR, IR

areas with frequencies 41000 – 400 cm-1

. Pristine SWNTs had high absorbance at low

wavenumbers, below 1000 cm-1

. This absorbance is attributed to metallic SWNTs.

Ozonation of SWNTs resulted in irreversible damage to tube structure and disappearance

of low frequency transitions (IR studies, Chapter 4).

For tube to be black in color, an absorbing chromophore must have a substantial

number of conjugated rings in acene system. Figure 31 shows possible interactions of a

diode laser beam with semiconducting tubes (8,0) and (10,0).

68

(8,3)

(10,3)

Figure 31. Possible ways for interaction of light with SWNT. Tubes are shown from a

side. Possible chromophores are shown with a darker color; their direction is marked with

double arrows a1, a2 and a3. Top: semiconducting tube (8,3); Bottom: semiconducting

tube (10,3).

A profound work by Clar,25

who synthesized and studied UV-Vis spectra of

hundreds of polycyclic hydrocarbons, revealed that a significant number of conjugated

rings is needed to make compound absorb at wavelengths above 700 nm (Figure 32).

p absmax 737 nm

dark green-black

7.8,15.16-dibenzoterrylene

p absmax 603 nm

dark blue

bisanthene

p absmax 582 nm

violet - blue

pentacene

p absmax 693 nm

dark green

hexacene

Figure 32. Bathochromic effect for separate and conjugated acene systems. Provided are

absmax

for p-band transitions. Conjugation of acene systems increases p absmax

.

69

Bathochromic effect is a result of “dilution” of an aromatic sextet (shown with

circles). One sextet is shared among annealed rings. Clar observed strong bathochromic

effect for compounds with multiple conjugation of acene systems (dibenzoterrylene and

bisanthene in Figure 32). Observed NIR absorption and fluorescence peaks of SWNT are

likely from fairly small segments. Such an interpretation of SWNT absorption absmax

and emission emmax

wavelengths would explain why SWNT with up to 3 % of double

bonds being quenched by ozonides do not show peak shifts for absmax

or emmax

after

ozonation.

Activation energies for decomposition of SWNT trioxolanes

Activation energies Eact were estimated from trioxolane decomposition rates, b.

Rates were obtained by regression on inverted normalized fluorescence data sets. Curves

tails were truncated as needed to increase „weight‟ of initial points. A primary goal was to

estimate activation energies of trioxolane decay, a fast process. Activation energies for

slow processes were not studied. Oxidation reactions were conducted in a thermostated

cuvette with temperatures set to 15.4, 20.1, 25.0 and 29.9 C. Amount of ozone bubbled

through an aqueous SWNT suspension was sufficient to substantially decrease SWNT

fluorescence intensity. It was found that at higher temperatures more ozone had to be

injected to get the same oxidation degree. Ozone is less stable at higher temperatures,

thus the observed phenomenon is expected. Constituents of reaction media, such as

sodium hydroxide and sodium dodecyl sulfate are known for their ability to speed up

ozone decomposition, and are expected to increase ozone decay rates due to a greater

70

number of collisions between ozone molecules and hydroxyls or dodecyl sulfate anions at

higher temperatures. Due to the nature of the experimental setup, air above an aqueous

SWNT suspension was purged right after bubbling of O3/O2 gaseous mixture (ca. 3 v/v %

ozone). This step was needed to avoid penetration of gaseous ozone accumulated above

the liquid back into suspension, thus decreasing an observed ozonide decay rate.

Surfactant sodium dodecyl sulfate formed foam after bubbling O3/O2 gaseous mixture

and helped in retaining unreacted ozone right above the aqueous media. Purging of air

above the liquid was primarily to destroy foam bubbles, thus releasing entrapped ozone.

Rates obtained from regression were plotted against reciprocal temperatures to

obtain decay activation energies (Figure 33).

1 / T (103 K

-1)

3.30 3.35 3.40 3.45

Ln

(b

)

-5.5

-5.0

-4.5

-4.0

-3.5

-3.0 954 nm

1026 nm

1123 nm

1250 nm

Figure 33. Arrhenius plot. SWNTO3 decomposition rates (b, s-1

) were measured at

different temperatures for 954, 1026, 1123 and 1250 nm emission wavelengths.

Many factors contributed to possible errors in trioxolane decomposition rates. Larger

error was observed for longer emission wavelengths. While 95% of emission at emmax

71

954 nm is coming from tube (8,3), fluorescence at emmax

1250 nm is a combination of

signals from tubes, none contributing more than 40% of observed intensity. Assuming

that ozonide decay of different tubes will be affected differently with temperature

changes, larger error was expected.

Generally, experiments were repeated many times with different loads of ozone to

obtain the most accurate rates. The most common problem observed with rates

measurements was having either too much ozone or too little ozone. Fluorescence of tube

(8,3) with emmax

= 954 nm was quenched less efficiently than that of tubes with longer

emission wavelengths, e.g. (9,5) and (10,3).

Too little ozone resulted in insufficient number of exponential decay points for tube

(8,3) with emmax

954 nm. Too much ozone resulted in a decay curve distortion, or

waviness, mainly for wider tubes, e.g. (9,5) with emmax

~ 1244 nm. Getting the right

amount was a challenge.

Obtained ozonide lifetimes and activation energies are summarized in Table 10.

72

Table 10. Activation energies and ozonide lifetimes for tubes emitting at 954, 1026, 1123

and 1250 nm wavelengths.

Lifetime , sec

T, C 954 nm 1026 nm 1123 nm 1250 nm

15.4 85 94 137 221

20.1 47 54 67 92

25.0 31 38 50 75

29.9 22 27 35 48

Eact (kJ/mol) 68.9 61.6 66.8 72.1

Eact (eV) 0.71 0.64 0.69 0.75

r2 0.9907 0.9886 0.9618 0.9289

* - Abbreviations are: Eact - activation energy for ozonide decay; - ozonide lifetime,

which is an inverse value of decomposition rate b, i.e. = 1/b; r 2

– coefficient of

determination.

SWNT fluorescence sensitivity to 1,2,3-trioxolanes on its surface

Fluorescence of tubes with larger diameters, mainly (9,5), (10,3) and (11,1),

represented by curve in Figure 34, was quenched four times stronger than that of tube

(8,3) (curve ). Such a difference in fluorescence quenching degrees is thought to be due

to varied sensitivity of tubes to similar number of ozonides on their surface.

73

Time (min)

0 10 20 30 40 50

Inve

rte

d N

orm

aliz

ed

Flu

ore

sce

nce

( I

ma

x /

I )

0

10

20

30

40

No

rma

lize

d F

luo

resce

nce

0.0

0.2

0.4

0.6

0.8

1.0

954 nm

1027 nm

1125 nm

1251 nm

A

B

Figure 34. Formation and decomposition kinetics of SWNT ozonides at four distinct

wavelengths. Ozone addition at ca. 2 min led to a sharp decline in fluorescence intensity

followed by partial recovery over time. (A) Normalized fluorescence, (B) inverted

normalized fluorescence (Imax/I). Symbols denote wavelengths at which kinetics were

monitored. The 660 nm laser was used for excitation.

Tubes of interest and their diameters are summarized in Table 11.

74

Table 11. Major contributors to fluorescence intensity at four distinct wavelengths*

em, nm , cm-1

(n,m) type of

major

contributors

tube diameter,

nm

% of total

emission at em

955.6 10465 8,3 0.782 95.4

6,5 0.757 1.9

1027.6 9731 7,5 0.829 85.0

10,2 0.884 5.3

8,1 0.678 4.3

1124.6 8892 7,6 0.895 78.9

8,4 0.840 8.3

9,2 0.806 3.7

9,4 0.916 3.4

1250.1 8000 9,5 0.976 39.8

10,3 0.936 30.3

11,1 0.916 12.0

8,7 1.032 6.2

10,5 1.050 3.8

8,6 0.966 3.0

* Excitation source ex

max 660 nm.

Tube surface area is l times greater than its diameter, where l – is a tube length.

Evidently, surface area increase is no more than 20% for tubes with larger diameters

( emmax

~ 1250 nm). All tubes listed in Table 11 have a chiral angle, meaning they are

“twisted”. 3D structures of major contributors to fluorescence intensity are shown below

(drawn not to scale).

(8,3) (7,5) (7,6) (9,5) (10,3)

75

The tube curvature is playing a key role in ozonide decomposition rates, though there

is no clear understanding why fluorescence quenching of ozonated tubes is four times

greater for (9,5) and (10,3) when compared to (8,3). One possible explanation of such

difference between quenching degrees (Imax/I) is a varied tube sensitivity to presence of

1,2,3-trioxolanes on its surface. While it is possible to suggest that the reaction rate

between ozone and tube (9,5) may be faster than that for (8,3), normalized fluorescence

intensity of tube (9,5) fifty three minutes after ozonation was ca. 10% lower than that for

tube (8,3) (Figure 34). The 10% difference in normalized intensities could be because of

a wider diameter of tube (9,5).

76

Wavelength (nm)

950 1050 1150 1250

Norm

aliz

ed F

luore

scence I

nte

nsity

0.0

0.2

0.4

0.6

0.8

1.0

Flu

ore

sce

nce Inte

nsity (

nW

/nm

)

0.0

0.5

1.0

1.5

2.0

53 min

2 min

before O3

A

B

951 954

0.9

1.0

Figure 35. An influence of SWNT ozonation on emmax

locations. The reaction was

monitored at r. t. Plots are: (A) Fluorescence spectra as measured, (B) normalized

fluorescence (I/Imax). Curves are: – pristine SWNT suspended in 1 wt. % aqueous

SDS, – SWNT two minutes after bubbling 2 mL of O3/O2 gaseous mixture (ca. 3 v/v

% of ozone), – spectrum 53 min after ozonation. Legends are the same for the top and

the bottom plots.

77

A very small blue shift was observed for tube (8,3) with emmax

~ 954 nm right after

ozonation (curve ). This phenomenon is thought to be due to a depletion of an electron

density of the -cloud. This shift disappeared over time (compare curves and , see

zoom-in in Figure 35B). Fluorescence of tube (8,3) recovered to about 70 % level of its

original intensity 53 min after ozonation. Regression performed on the curve (curve ,

Figure 35A) estimated that final recovery level would not get higher than 75 % of an

initial intensity. It is not clear why fluorescence peak maxima are so insensitive to

sidewall modification. It is reasonable to expect emmax

peaks to shift for functionalized

tubes, though it is not observed. At the same time, a variety of surfactants (CTAB, SDBS,

SDS, Brij 700) were shown to affect peak maxima.26

From these observations a

conclusion can be made, that fluorescence from modified sections of SWNT does not

recover and that an observed recovery is likely coming from an increase of an electron

density on SWNT after ozonides‟ decay.

PM3-level optimization of SWNT tube with several 1,2,3-trioxolanes on its surface

gave a “squeezed” tube, with tube getting wider at the place of functionalization (Figure

27). Similar results were obtained for ozonated fullerenes. Such distortion affects tube

dimensions, and thus can affect its emmax

. Tube distortion and disrupted conjugated

-system are likely to be two main reasons for lack of fluorescence from functionalized

sections of SWNT.

Overall, fluorescence peaks maxima were found to be very insensitive to side wall

modification with 1,2,3-trioxolanes, i.e. there was no significant shift observed. It is

believed that fluorescence after ozonation is coming from pristine sections of tubes only.

78

Fluorescence recovery is likely due to an increase of an electron density of a -cloud

after ozonides decomposition to SWNT epoxides and molecular oxygen.8 Fluorescence

from functionalized areas is thought to be irreversibly lost. Special treatment, like tube

baking at high temperature, would be needed to restore original SWNT structure.

Work with ozone: gaseous vs. aqueous ozone and parameters that affect decay rates

observed by fluorescence

Ozone diffusion through water was found to be somewhat inefficient. In a separate

experiment a slow addition of an equal amount of water to an aqueous solution of ozone,

with Abs 260 ~ 1, did not yield absorbance decrease to an estimated dilution level even

within several minutes after addition. Mechanical stirring was required to halve the

absorbance value. The same problem was encountered when water saturated with ozone

was added to SWNT-SDS aqueous suspension. Seemingly large quantities of ozone were

quenching SWNT fluorescence inefficiently, even with fast injections. Speed of

injections was found to affect the effectiveness of mixing. Stirring with a spatula was

shown to increase SWNT oxidation degree. Bubbling of known volumes of gaseous

ozone through SWNT suspension gave a much better reproducibility of fluorescence

quenching degrees (Imax/I) when compared to addition of aqueous ozone. The main

advantage of using gaseous O3/O2 mixture was thorough suspension stirring with gas

bubbles. On the flip side, gaseous ozone was found entrapped in foam bubbles above the

aqueous suspension after injections and required an additional purge of space above the

liquid to eliminate ozone penetration from the gaseous phase back into suspension.

79

Sodium dodecyl sulfate, a surfactant for SWNT, is known for its ability to foam. Ozone

has a long decay rate in the air. Purging of air above the liquid was primarily to destroy

foam bubbles, thus releasing entrapped ozone. Penetration of gaseous ozone back into the

liquid phase affected negatively rates observed by fluorescence, meaning actual ozonide

decay rates are faster then the ones observed. An experiment was performed by bubbling

the same amounts of ozone through two 1 wt. % aqueous SDS suspensions. Air above the

liquid in a second cuvette was purged and decay rates were measured by absorbance

change at 260 nm. The purged cuvette had a slightly faster ozone decay rate, 0.47 vs.

0.43 min-1

.

Decay rates of SWNT 1,2,3-trioxolanes measured with fluorescence technique are

also dependant on solution acidity and surfactant aggregation at low temperature.

Decreasing temperature to below 15 C may result in surfactant precipitation. Acidity per

se does not seem to affect 1,2,3-trioxolane decomposition, but it has a strong influence on

SWNT fluorescence. Increase in acidity would give a weaker emission signal. Keeping

suspension pH in the range 8 -9 was necessary.

To summarize, injection of gaseous ozone was found to be the most efficient way to

oxidize SWNT. Injections of gaseous ozone should be followed by a brief stirring with a

spatula and a thorough purge of the gaseous phase above the SWNT suspension. It is

desirable to have the same injection speed in all runs. Running experiments in a

thermostated cuvette was found to be necessary for obtaining reproducible kinetics

results. These guidelines were followed in obtaining 1,2,3-trioxolane decay rates.

Suspension acidity should be kept in a range pH 8 – 9.

80

Fluorescence intensity drift

SWNT fluorescence was found to be fluctuating. Such fluctuations are thought to be

dependant on a number of parameters. Two of the most important factors were

suspension temperature and the length of laser pulses. Even small changes in the

temperature were shown to affect fluorescence intensity of SWNT suspended in 1 % aq.

SDS solutions. Such fluctuation was less noticeable in sodium dodecyl benzyl sulfonate

(SDBS) surfactant. An attempt to use SWNT suspended in SDBS was not successful, as

it was found that ozone does not quench SWNT fluorescence efficiently in presence of

SDBS. SDBS itself is reacting with ozone, thus reducing amount of ozone available for

reaction with SWNT. The laser pulse length was found to affect fluorescence fluctuation.

The least drifts were obtained with > 10 sec intervals between 500 msec pulses.

Decreasing duty cycle to below ten seconds resulted in random intensity drifts, both

upward and downward relative to the initial fluorescence. The majority of experiments in

this work were repeated numerous times to ensure that the observed kinetics were

reproducible and not affected by fluorescence drifts.

Ozone loads

Abnormally high loads of ozone, typically greater than 2 mL of O3/O2 gaseous

mixture (ca. 3 v/v % ozone) were found to distort single exponential decay curves.

Particularly, three types of decay were observed with high loads of ozone: fast ozonide

decay, slow ozonide decay and some rearrangement processes that did not result in

oxygen release. All three processes were found to follow exponential decays. The fast

and slow ozonide decays could be regressed with 5-parameter two exponential decay

81

formulas. Oxygen release was observed during these two processes. The very slow

rearrangement process was always observed, but its rate was not studied. Particularly,

random fluorescence drifts were overshadowing slowly changing fluorescence intensity,

thus introducing a large error.

2.3. Conclusions

An interaction of ozone with single-walled carbon nanotubes (SWNT) resulted in the

formation of 1,2,3-trioxolanes (SWNTO3). Obtained formation rate was 2.4·104 M

-1s

-1 for

SWNT – 1% aq SDS suspension, which is of the same order of magnitude as the

formation rates reported for carbon tetrachloride solutions of C60O3 and C70O3. SWNTO3

decayed to SWNT oxides (SWNTO) with release of molecular oxygen. A vacuum

deoxygenation technique performed on dry ozonated SWNT showed oxygen release to

follow simple exponential rise with rates ca. 1.5 – 2 min-1

at r. t. The lifetime of

SWNTO3, was shown to depend on temperature and SWNT type, and at room

temperature was less than two minutes for small-diameter SWNTs suspended in water.

Ozonides exhibited an extreme quenching of SWNT fluorescence and a substantial

bleaching of NIR absorption. The maximum number of 1,2,3-trioxolanes forming on the

surface of SWNT at any given time was found to be less than 4% of the theoretical value,

indicating a saturation point. Reaction of ozonated nanotubes with excess ozone was

limited by the SWNTO3 decomposition rates. Thinner tubes exhibited faster ozonide

decay rates resulting in greater oxidation levels over time in excess of ozone.

82

2.4. Experimental Part

Setup for measurement of oxygen evolution from SWNT ozonated in solid state

A set of experiments was designed to measure the amount of oxygen evolving from

the surface of ozonated SWNT. A vacuum line, shown in Figure 36 was equipped with an

electronic pressure gauge.

Figure 36. (Left) A vacuum line for measurement of oxygen release. A pressure gauge is

in the left top corner; a valve to cut off an oil diffusion pump is at the lower left. (Right)

A reaction vessel with a cap at the top and a valve on a side.

The entire system with a reaction vessel attached was vacuumed for at least one day

before each experiment. Typically, such vacuuming resulted in a background leak of

0.5 mTorr/min. This was a sufficiently low leak for the purposes of measurements.

Ozone collection

An example of a syringe used for ozone collection is shown in Figure 37. Typically,

samples were collected for 1 minute in a 3 mL disposable syringe with the O3/O2 gaseous

mixture flow set to 25 mL per minute (ca. 3 v/v % ozone).

83

Figure 37. (A) Teflon tubing coming out of a corona discharge ozonator and a

homemade assembly for syringe attachment. (B) 3 mL disposable syringe outer core is

attached to an assembly for ozone collection.

It was found necessary to connect a syringe plunger to the outer core immediately

after it was taken off of the ozonation line. This was done to reduce air flows above the

syringe and made the amount of ozone reproducible with each collection.

Experimental determination of the amount of oxygen evolved from ozonated

SWNT.

The following procedure describes determination of the volumes of the reaction

vessel and the vacuum line. Calculations of the amount of evolved oxygen are also

provided.

The vacuum line was vented to the air to get 803 mTorr in the entire system. The

reaction vessel was cut off and the rest of the system evacuated. The pump was cut off

and the reaction vessel opened; gas from the vessel spread through the system and caused

pressure to change to 197 mTorr, indicating the ratio of volumes k = 803/197 = 4.08. The

reaction vessel was taken off the vacuum line, filled with water and its volume

determined, Vol (vessel) = 34.0 mL, from which volume of the system was determined

84

Vol (system) = 139 mL. Calculation of the number of moles of oxygen evolved after

ozonation was determined with the following equations:

obs

obsobs

stp

stpstp

T

VP

T

VP can be rewritten as

obsstp

stpobs

obsstpTP

TPVV

Parameters are:

Tstp = 273.15 K

Tobs = 293 K

Pstp = 760 Torr (1 atm)

Vm = 22.414 L/mol (at 0 C and 1 atm)

][1071.1760293

15.273139.0 4 TorrP

PLV obs

obsstp

The molar amount of oxygen was estimated with the following equation

414.222

stp

m

stp

O

V

V

V

The results are summarized in Table 12:

85

Table 12. Conversion of oxygen gas pressure into amount

O2), umol Vstp, uL Pobs, mTorr

0.23 5.1 30

0.30 6.8 40

0.38 8.6 50

0.45 10 60

0.54 12 70

0.62 14 80

0.67 15 90

0.72 16 95

Equipment

Near-IR fluorescence and absorbance in the range 900-1350 nm were recorded on a

NS1 Nanospectralyzer (Applied NanoFluorescence LLC, Houston, TX). Built-in lasers

660 and 785 nm were employed for excitation as noted in the text. 1% aq. sodium

dodecyl sulfate (Aldrich) solution served as a reference. Software that came with the NS1

Nanospectralyzer was used for spectra deconvolution. UV-Vis absorbance in a range

250-900 nm was recorded on a Cary 4E UV-Vis Spectrophotometer with 1 cm quartz

cuvette. Unless otherwise noted, 1 cm quartz cuvettes were used for UV-Vis and NIR

absorption measurements. Oxygen gas measurements were performed with a 275 Mini-

Convectron (Granville-Phillips Co.) pressure gauge.

Preparation of a surfactant coated SWNT aqueous suspension

Single-walled carbon nanotubes (60 mg, SWNT, HipCo, batch 161.1, raw,

unpurified, Rice University) were debundled in 1 wt. % aqueous solution (200 mL) of

sodium dodecylsulfate (Aldrich) with a bath sonicator (FS 14, Fisher Scientific) for

4 hours. Further dispersion was performed by applying intense ultrasonic agitation

(7 watt) with a tip sonicator (Microson XL2000) for 15 min. The sample was ultra-

86

centrifuged (Sorvall, Discovery 100SE) for 4 hours at 151514.2 g to obtain a clear, dark

grey decant, which was separated from the precipitate right after the centrifugation. The

SWNT concentration in the decant was approx 23 mg/L estimated by visible absorption.

(Absorbance values measured at 632 and 763 nm were divided by coefficients 0.036 and

0.043 L/mg respectively to calculate the SWNT concentration (mg/L).)27

The resulting

suspension was diluted four times with 1% aq. SDS solution and used for aqueous

experiments. An oxygen flow meter (Puritan-Bennett Corp.) was used at flow rates 1/16

and 1/32 L/min as noted.

Ozonation procedures

Ozone was generated by passing oxygen (industrial grade, Matheson Tri-Gas) with a

flow rate 1/32 L/min at r. t. through a high frequency corona discharge ozonator

(GE60FM, Yanco Industries Ltd, www.ozoneservices.com) set to a maximum output

(power level 10). The ozonator was idled before use for at least 7 min to reach its

maximum output as instructed by manufacturer. Three different ozonation procedures

used in this work are described below.

Direct ozonation of aqueous SWNT-SDS suspension. A gaseous mixture of O3/O2

(ca. 3 v/v % ozone) was bubbled through an aqueous SWNT suspension in a test tube for

a desired period of time.

Gaseous ozone/oxygen injection with a syringe. Ozone was collected in a plastic

syringe for 1 min (3 mL volume; cf. Figure 37B) or 2 min (5-10 mL volume), capped

with a plunger and the desired volume injected into a cuvette with SWNT suspension

followed by brief stirring with a spatula and purging of the air above the liquid.

87

Dry ozonation. The oxygen flow rate was kept at 1/16 L/min. A SWNT film was

ozonated for a desired period of time at r. t. in a small chamber by holding 1/16” ID

Teflon tubing above the dry SWNT film. Alternatively, a specific volume of O3/O2

gaseous mixture (ca. 3 v/v % ozone) could be squeezed from a syringe pointing the gas

stream at the SWNT film.

Oxygen evolution measurements

A corresponding amount of SWNT (2 or 4 mg, SWNT, HipCo, batch 162.8, raw,

unpurified, Rice University) was debundled in benzene (10 mL) with a bath sonicator

(Fisher Scientific, FS 60) and a slurry added to a cylindrical reaction vessel. The vessel

was tilted horizontally and rotated until all solvent was evaporated. Thin Teflon tubing

with a constant flow of nitrogen gas was inserted into the vessel to speed up the

evaporation process. The resulting SWNT-coated glass vessel had an evenly distributed

layer of nanotubes. The vessel was attached to a vacuum line and the entire system

evacuated for a day. When the system leak decreased to below 0.5 mTorr/min following

pump cut off, the vessel was cooled with an ice bath, an oxygen/ozone gaseous mixture

injected (10 mL, ca. 3 v/v % ozone), the vessel capped; waited for 1 min and then the

entire system was degassed for 2.5 min, then the pump was cut off and the pressure

monitored for 20 min. For the second ozonation, the system was thoroughly degassed,

reaction vessel cooled, the O2/O3 gaseous mixture (10 mL) was injected and the above

procedure followed.

88

References and Notes

1. Chen, Z. Y.; Hauge, R. H.; Smalley, R. E., Ozonolysis of functionalized single-

walled carbon nanotubes. Journal of Nanoscience and Nanotechnology 2006, 6,

(7), 1935-1938.

2. Chen, Z. Y.; Ziegler, K. J.; Shaver, J.; Hauge, R. H.; Smalley, R. E., Cutting of

single-walled carbon nanotubes by ozonolysis. Journal of Physical Chemistry B

2006, 110, (24), 11624-11627.

3. Simmons, J. M.; Nichols, B. M.; Baker, S. E.; Marcus, M. S.; Castellini, O. M.;

Lee, C. S.; Hamers, R. J.; Eriksson, M. A., Effect of ozone oxidation on single-

walled carbon nanotubes. Journal of Physical Chemistry B 2006, 110, (14), 7113-

7118.

4. Banerjee, S.; Hemraj-Benny, T.; Balasubramanian, M.; Fischer, D. A.; Misewich,

J. A.; Wong, S. S., Ozonized single-walled carbon nanotubes investigated using

NEXAFS spectroscopy. Chemical Communications 2004, (7), 772-773.

5. Banerjee, S.; Wong, S. S., Rational sidewall functionalization and purification of

single-walled carbon nanotubes by solution-phase ozonolysis. Journal of Physical

Chemistry B 2002, 106, (47), 12144-12151.

6. Banerjee, S.; Wong, S. S., Demonstration of diameter-selective reactivity in the

sidewall ozonation of SWNTs by resonance Raman spectroscopy. Nano Letters

2004, 4, (8), 1445-1450.

7. Dresselhaus, M. S.; Dresselhaus, G.; Jorio, A.; Souza, A. G.; Pimenta, M. A.;

Saito, R., Single nanotube Raman spectroscopy. Accounts of Chemical Research

2002, 35, (12), 1070-1078.

8. Heymann, D.; Bachilo, S. M.; Weisman, R. B.; Cataldo, F.; Fokkens, R. H.;

Nibbering, N. M. M.; Vis, R. D.; Chibante, L. P. F., C60O3, a fullerene ozonide:

Synthesis end dissociation to C60O and O2. Journal of the American Chemical

Society 2000, 122, (46), 11473-11479.

9. Liu, L. V.; Tian, W. Q.; Wang, Y. A., Ozonization at the vacancy defect site of

the single-walled carbon nanotube. Journal of Physical Chemistry B 2006, 110,

(26), 13037-13044.

10. Cai, L. T.; Bahr, J. L.; Yao, Y. X.; Tour, J. M., Ozonation of single-walled carbon

nanotubes and their assemblies on rigid self-assembled monolayers. Chemistry of

Materials 2002, 14, (10), 4235-4241.

11. Mawhinney, D. B.; Naumenko, V.; Kuznetsova, A.; Yates, J. T.; Liu, J.; Smalley,

R. E., Infrared spectral evidence for the etching of carbon nanotubes: Ozone

89

oxidation at 298 K. Journal of the American Chemical Society 2000, 122, (10),

2383-2384.

12. Ogrin, D.; Chattopadhyay, J.; Sadana, A. K.; Billups, W. E.; Barron, A. R.,

Epoxidation and deoxygenation of single-walled carbon nanotubes:

Quantification of epoxide defects. Journal of the American Chemical Society

2006, 128, (35), 11322-11323.

13. Cataldo, F., Polymeric fullerene oxide (fullerene ozopolymers) produced by

prolonged ozonation of C60 and C70 fullerenes. Carbon 2002, 40, (9), 1457-1467.

14. Cataldo, F.; Heymann, D., A study of polymeric products formed by C60 and C70

fullerene ozonation. Polymer Degradation and Stability 2000, 70, (2), 237-243.

15. Churilov, G. N.; Isakova, V. G.; Weisman, R. B.; Bulina, N. V.; Bachilo, S. M.;

Cybulski, D.; Glushchenko, G. A.; Vnukova, N. G., Synthesis of fullerene

derivatives. Physics of the Solid State 2002, 44, (4), 601-602.

16. Deng, J. P.; Mou, C. Y.; Han, C. C., Oxidation of fullerenes by ozone. Fullerene

Science and Technology 1997, 5, (5), 1033-1044.

17. Heymann, D.; Weisman, R. B., Fullerene oxides and ozonides. Comptes Rendus

Chimie 2006, 9, (7-8), 1107-1116.

18. Razumovskii, S. D.; Bulgakov, P. G.; Ponomareva, Y. G.; Budtov, V. P., Kinetics

and stoichiometry of the reaction between ozone and C70 fullerene in CCl4.

Kinetics and Catalysis 2006, 47, (3), 347-350.

19. Razumovskii, S. D.; Bulgakov, R. G.; Nevyadovskii, E. Y., Kinetics and

stoichiometry of the reaction of ozone with fullerene C60 in a CCl4 solution.

Kinetics and Catalysis 2003, 44, (2), 229-232.

20. Chibante, L. P. F.; Heymann, D., On the Geochemistry of Fullerenes - Stability of

C60 in Ambient Air and the Role of Ozone. Geochimica Et Cosmochimica Acta

1993, 57, (8), 1879-1881.

21. Bulgakov, R. G.; Nevyadovskii, E. Y.; Belyaeva, A. S.; Golikova, M. T.;

Ushakova, Z. I.; Ponomareva, Y. G.; Dzhemilev, U. M.; Razumovskii, S. D.;

Valyamova, F. G., Water-soluble polyketones and esters as the main stable

products of ozonolysis of fullerene C-60 solutions. Russian Chemical Bulletin

2004, 53, (1), 148-159.

22. Razumovskii, S. D.; Zaikov, G. E., Ozone and its reactions with organic

compounds. ed.; Elsevier Science Publishers: New York, NY, 1984; Vol. I, p.

403.

90

23. Concentration of ozone in oxygen stream was provided by the manufacturer of the

ozonator.

24. Weisman, R. B.; Heymann, D.; Bachilo, S. M., Synthesis and characterization of

the "missing" oxide of C60: [5,6]-open C60O. Journal of the American Chemical

Society 2001, 123, (39), 9720-9721.

25. Clar, E., Polycyclic Hydrocarbons. Academic Press: New York, 1964; Vol. 1, p.

487.

26. Moore, V. C.; Strano, M. S.; Haroz, E. H.; Hauge, R. H.; Smalley, R. E.; Schmidt,

J.; Talmon, Y., Individually suspended single-walled carbon nanotubes in various

surfactants. Nano Letters 2003, 3, (10), 1379-1382.

27. Moore, V. C. Single walled carbon nanotubes: Suspension in aqueous/surfactant

media and chirality controlled synthesis on surfaces. Rice University, Houston,

2005.

91

Chapter 3

Influence of SWNT ozonation on D and G bands in Raman spectra

92

3.1. Introduction

Studying disorder-induced D and tangential G modes of graphite crystallites in

Raman spectra can be tracked back to the pioneering work of Tuinstra and Koenig1

published in 1970. Kastner et al.2 compared Raman and infrared spectra of carbon

nanotubes to those of highly oriented pyrolytic graphite (HOPG). Raman active E2g in-

plane stretching mode, often designated as the G mode, was found at 1574 cm-1

for

carbon nanotubes. A disorder induced D-band was found at ca. 1350 cm.-1

The IR

spectrum of HOPG had lines at 868 (out-of-plane mode, A2u) and 1588 cm-1

(in-plane

mode, E1u), while that of carbon nanotube had the same A2u peak and a much broader

asymmetrical E1u peak at 1575 cm.-1

The IR spectrum of carbon nanotubes presented in

the article does not resemble the ones observed for pristine HipCo tubes in this work. The

peaks A2u and E1u were observed for ozonated SWNT samples, but not for pristine ones

(this work). Pimenta et al.3 studied disorder-induced D and D’ Raman features, as well as

the G’-band (the overtone of the D-band which is always observed in defect-free

samples). The authors determined G band (appearing near 1582 cm-1

in graphite) as a

doubly degenerate phonon mode (E2g symmetry) at the first Brillouin zone center that is

Raman active for sp2 carbon network. Pimenta noted that integrated intensity ratio ID/IG

for the D band and G band is widely used for characterizing the defect quantity in

graphitic materials. Expanding the work of Tuinstra and Koenig,1, 4

Cancado et al.5

provided the following formula for estimation of in-plane crystallite size La:

93

1

410104.2)(G

Dlasera

I

InmL , where crystallite size La and excitation

wavelength are expressed in nanometers. Parameters ID and IG are integrated intensities

of the G and disorder-induced D bands in Raman spectra. Thomsen and Reich6 linked

excitation energy dependence of the D mode to a double resonant process. Ferrari et al.7

examined relation of D, G and G’ bands in graphene, and found the dependence of G’

band shape and location on the number of graphene layers. A recent summary of IR and

Raman spectral changes of purified vs. pristine SWNT has been reported.8 An analysis of

Raman spectra of SWNT from different sources was published by Hennrich et al.9

It appears that the influence of small quantities of ozone on SWNT Raman features

has not been reported to date. This chapter will give a brief summary of the influence of

ozonation of aqueous and solid samples on Raman disorder (D) and tangential (G) modes

in SWNT. As will be demonstrated, small loads of ozone resulted in a significant

decrease of G band and had little or no influence on the disorder band.

94

3.2. Results and discussion

Wavelength (nm)

710 730 750 770 790 810

Counts

, x10

3

5

10

15

20

749.5

5 min

10 m

17 m

24 m

32 m

37 m

46 m

D

G

G'

Figure 1. Raman spectra acquired over time on an aqueous SWNT – SDS suspension

after bubbling O3/O2 gaseous mixture (ca. 3 v/v % ozone) for 3 min. The sample was

excited with 669.9 nm laser source. Consult Chapter 1 for conversion of wavelengths into

wavenumbers. Symbols are labeled with approximate times from the beginning of

ozonation.

Several Raman spectra of an aqueous SWNT – SDS suspension were measured over

time after its ozonation. Gas bubbles evolving during ozonides decomposition were found

to be interfering with Raman measurements causing random, not reproducible peaks in

the acquired spectra. An attempt was made to focus on SWNT spectral changes after

major portion of ozonides has already decomposed, thus decreasing chances of

misinterpreting the data. The first spectrum in Figure 1 was recorded ca. 2 minutes after

bubbling of gaseous ozone through a SWNT – SDS suspension.

As seen in Figure 1, the G-peak increased over time by 33% from its 5 min value.

Within the same timeframe the basis function of D-band did not change. A change in

95

fluorescence background between 5 and 10 min spectra is noticeable. Small random

peaks due to light interference with oxygen bubbles are seen in the region 710 – 725 nm

(curve ).

It is believed that multiple processes contribute to an increase in the intensity of the

G- band, two major ones being ozonide decay (a fast process) and structural

rearrangement of epoxides to oxidoannulenes (a slow process). By analogy with

fullerenes,10

an oxidation of SWNT is likely to yield multiple functional groups on its

surface. It was concluded from IR studies of ozonated SWNT films (see Chapter 4) that

epoxides, aryl ethers, -diketones, anhydrides and esters are likely to be dominant

functional groups on the SWNT surface after ozonide decay. G-band intensity growth in

Figure 1 is thought to be due to SWNT structural changes leading to a greater number of

Raman active double bonds.

Multiple injections of O3/O2 gaseous mixture (ca. 3 v/v % ozone) into a SWNT –

SDS suspension led to absorbance bleaching. Absorbance at 260 nm fell by

approximately 10% after eight injections with 4 min intervals between each (Chapter 2,

Figure 13). Raman spectra were acquired before and 15 hrs after ozonation. As was

demonstrated in the fluorescence studies (cf. Chapter 2 for details), extended time period

was needed for tubes to recover. As shown in Figure 2, G-peak recovered only to 62 %

level of its original intensity. A slight increase of background is thought to be due to

fluorescence of certain oxidized sections of SWNT. The basis function and the shape of

the D-band did not change after ozonation (Figure 2).

96

Wavelength (nm)

710 730 750 770 790 810

Coun

ts, x10

3

10

20

30

40before O

3

15 hr after

730 750

Cou

nts

, x1

03

3

6

D

G

G'

D

G

Figure 2. Raman spectra of SWNT-SDS aqueous suspension before and 15 hrs after

ozonation. Sample was excited with 669.9 nm laser source. Eight O3/O2 gaseous mixture

injections (ca. 3 v/v % ozone), 1.5 mL each, were made with 4 min intervals and

suspension kept at r. t. for 15 hours.

Though the D/G peak intensity ratio changed, this result does not support the

common point of view that D band intensity should increase with a higher degree of side

wall functionalization.

For a vibration to be Raman active, it needs a change in polarizability and no change

in a dipole moment. Carbon atoms participating in C-O bonds of SWNT epoxides and

oxidoannulenes can not be considered as truly sp3 hybridized, since bond angles do not

match those in a diamond. Furthermore, if oxidoannulene rearranges to diaryl ether,

carbon atoms become sp2 hybridized. Introduction of epoxides, oxidoannulenes,

anhydrides, esters and -diketones is likely to increase the intensity of asymmetric

stretches, contributing little or nothing to stretches with zero dipole. Theoretically

97

peroxides, epoxides and aryl ethers should have Raman active stretches in the region

1000-1200 cm-1

, near D-band, but experimentally no D-band growth could be observed

for an aqueous suspension of ozonated SWNT.

The ambiguity surrounding D-band growth during side wall modification led to a

series of UV measurements. As mentioned above, eight 1.5 mL injections of O3/O2

gaseous mixture (ca. 3 v/v % ozone) were bubbled through 1.5 mL of SWNT - SDS

aqueous suspension with 4 min intervals. The reaction of SWNT with ozone was

monitored at absmax

(O3) = 260 nm (Figure 3). Each injection led to an initial absorbance

increase followed by a fast exponential decay. The amount of ozone that could be

dissolved in 1% SDS solution was found to be limited. An absorbance jump in the

presence of SWNT was less than 0.1 a. u. An increase in the upward jump amplitude

(first to fifth injections, dotted line in Figure 3) and longer times needed to consume free

floating ozone was interpreted as a consequence of SWNT saturation with 1,2,3-

trioxolanes, also known as primary ozonides. A four minute time period at r. t. was found

to be insufficient to decompose all ozonides. This interpretation explains why downward

“step,” or SWNT bleaching, is getting smaller with each subsequent ozonation.

98

Time (min)

10 20 30

Ab

so

rba

nce

(a

.u.)

0.54

0.57

0.60

Figure 3. Absorbance of SWNT – SDS aqueous suspension at 260 nm and room

temperature. Eight injections of O3/O2 gaseous mixture (ca. 3 v/v % ozone) were made

with 4 min intervals (emphasized with arrows). Solid line represents absorbance change

due to needle insertion. Points above 0.62 a.u. are extraneous and were excluded from the

graph. A diagonal dotted line marks an increase of absorbance right after each ozone

injection.

The resulting suspension was kept at r.t. for 15 hrs and then the Raman spectrum was

recorded (Figure 2).

Oxidation of SWNT with different amounts of ozone as monitored by Raman.

Injection of specific volumes of ozone (0.5 – 10 mL) led to a decrease of the G-peak

intensity with no change in the D band (Figure 4).

A functionalization of SWNT side wall with 1,2,3-trioxolanes was accompanied by a

broad increase of background fluorescence intensity. The G-peak dropped approximately

26% from its initial value after an injection of 4 mL of O3/O2 gaseous mixture (ca. 3 v/v

% ozone) and reached a saturation point. An injection of 5 or 6 mL of ozone brought the

99

G peak to the same intensity level. Higher loads of ozone (8 and 10 mL) resulted in

increased background fluorescence (curves and in Figure 4).

Wavelength (nm)

710 730 750 770 790 810

Counts

, x10

4

1

2

3

4

749.5

Cou

nts

, x1

04

2.8

3.2

3.6

before O3

0.5 mL

1 mL

2 mL

3 mL

4 mL

5 mL

6 mL

8 mL

10 mL

734

Cou

nts

, x1

04

0.4

0.8

D

G

G'

D G

Figure 4. The influence of different volumes of ozone on G peak intensity. Specific

amounts of gas were injected into separate aliquots of aq. SWNT - SDS suspension. All

samples were heated to 40 C for 30 min before measuring Raman spectra. The 669.9 nm

laser was used for excitation.

To better understand how many 1,2,3-trioxolanes can form on SWNT at any given

time, aliquots of SWNT suspension were bubbled with ozone oxygen mixture (OOM) for

specific periods of time. The first bubbling time was set to 30 seconds, which would be

the equivalent of injecting 13 mL of O3/O2 gaseous mixture. Other times were 1, 2, 3, 4,

5, 10, 30 min and 1 hour. Bubbling O3/O2 gaseous mixture for one hour resulted in a

complete disappearance of D and G bands and a substantial increase in fluorescence. The

sample changed color from light grey to light brown. This spectrum will not be presented

here. Thirty minute ozonation gave a very strong fluorescence and small D and G peaks,

100

as expected. When corrected for fluorescence background, the basis function under D-

band preserved its size and shape (curve in Figure 5).

Wavelength (nm)

710 730 750 770 790 810

Counts

, x10

4

0

1

2

3

4before O3

30 s

1 m

2 m

3 m

4 m

5 m

10 m

30 m

D

G

G'

Figure 5. Bubbling O3/O2 gaseous mixture (ca. 3 v/v % ozone) through SWNT – SDS

suspension for noted periods of time. Spectra are shown as measured and overlaid for

comparison. The 669.9 nm laser was used for excitation. Each curve represents a separate

experiment. All samples were heated to 40 C for 30 min before measuring Raman

spectra.

A decrease in G band intensity was not accompanied by D band increase after

ozonation (Figure 5). D band was found unchanged within the first ten minutes of

bubbling O3/O2 gaseous mixture (curves and ).

Fluorescence of functionalized SWNT was found to be fairly strong. Interestingly,

the Raman spectrum of a phenylsulfonated SWNT aqueous suspension in 1% SDS (not

shown) resembled the one ozonated for 30 min (curve in Figure 5).

To simplify a visual comparison of D and G peaks for different curves, plots were

shifted along vertical axis to make D-band (734 nm) at the same intensity level as in

101

SWNT sample before ozonation. This is done to compensate for unequal fluorescence

background in all samples.

Wavelength (nm)

710 730 750 770 790 810

Counts

, x10

4

0

1

2

3

4

before O3

30 s

1 m

2 m

3 m

4 m

5 m

10 m

30 m

749.5D

GG

G'

Figure 6. Bubbling O3/O2 gaseous mixture (ca. 3 v/v % ozone) through SWNT – SDS

suspension for noted periods of time. Raman spectra were shifted along vertical axis to

make D-band (734 nm) at the same intensity level as in SWNT sample before ozonation.

The 669.9 nm laser was used for excitation. Each curve represents a separate experiment.

All samples were heated to 40 C for 30 min before measuring Raman spectra.

Figure 6 demonstrates that the basis function for G band stopped decreasing at or

before 2 min of bubbling of O3/O2 gaseous mixture (ca. 3 v/v % ozone), indicating a

saturation point. Two minutes of bubbling is equivalent to a slow injection of 50 mL of

O3/O2 gaseous mixture. It is clearly seen in Figure 6 that ozonation for the first 10 min

does not affect the D band. After an initial fast decrease of G band (curves , and )

curves for 2 through 5 min of continuous ozonation (, , and ) gave essentially the

same intensity of G band. Longer ozonation times resulted in further decrease of G band

(curves and ).

102

The influence of sidewall modification on Raman spectra. A comparison between

different reactions.

A comparison was made between Raman spectra of pristine, dodecylated,

phenylated, phenylsulfonated and ozonated SWNT. Both liquid (dispersed tubes) and

solid state (bundled tubes) spectra were acquired, showing close similarity. Water

suspensions of the above mentioned samples of SWNT were prepared in 1 aq. % SDS by

bath sonication. Raman spectra of suspensions were acquired with 669.9 nm laser source

and had slightly broader peaks (spectra not shown). Fluorescence from samples was

much stronger in liquid state and therefore only solid state spectra will be discussed.

Wavenumber (cm-1

)

500 1000 1500 2000 2500 3000

Co

un

ts

x1

04

0

2

4

6

8

10

12Pristine SWNT

dodecyl

Ph

C6H

4-SO

3H

epoxide

Modified SWNT:

G

G'D

1295 1595

1

2

3

4

5

D

G

Figure 7. Raman spectra of pristine and functionalized SWNT overlaid for comparison.

Solid SWNT samples were excited with 785 nm source. All intensities are shown in

counts, as measured. SWNT samples are: – pristine, – dodecylated, –

phenylated, – phenylsulphonated, – ozonated. Dodecylated and phenylsulfonated

SWNT samples are courtesy of Feng Liang.

103

Intensities of spectra in Figure 7 are shown in counts, as measured. Commonly used

normalization of different spectra at the G-peak frequency will not be used in this work

for reasons discussed below. Utilization of ratio of intensities of D and G bands (D/G

ratio) will be used instead to describe a degree of SWNT functionalization. There is no

doubt that D band changes in its shape and intensity after SWNT functionalization, but

quantifying those changes cannot be easily derived from D/G ratios. Comparison of

spectra in Figure 7 clearly demonstrates the complexity of the quantification problem.

There is a striking difference in intensities of peaks of common aromatic compounds

versus SWNT in Raman spectra. Spectra in Figure 7 were acquired with an attenuated

laser on a single scan. Recording a spectrum of picric acid with the same conditions gave

much weaker intensities. The highest peak on a spectrum of picric acid (not shown) had

approximately the same height as the D band in pristine SWNT. The high intensity of the

G peak, ascribed to sp2-hybridized carbon atoms, is thought to be due to resonance of

conjugated double bonds.

A disruption of conjugation and introduction of electron withdrawing groups during

oxidation with ozone led to a substantial decrease of the G peak intensity and, at the same

time, gave almost no change of D band (curve ). Some broadening of the D peak is

associated with formation of epoxides and other functional groups. Banerjee11

in his

studies of ozonated SWNT samples encountered exactly the same problem with weak

increase in D band. He proposed that Raman cross sections for sp2 and sp

3 hybridized

carbons are dramatically different.

Clearly, conversion of an sp2-hybridized carbon to an sp

3 carbon after ozonation did

not result in increase of D band intensity (Figure 7). This result can be explained by

104

distinguishing sp3 carbon atoms of SWNT framework formed by an attack of a carbon-

based reactive center (e.g. phenyl radical) and other reactive species (e.g. ozone). An

electron withdrawing character of an attacking species is likely to decrease the intensity

of symmetrical stretches (Raman active) and increase the intensity of asymmetrical

stretches (IR active) of sp3 carbon atom on SWNT. Thus, an introduction of epoxides

gave no intense symmetric stretches. IR studies of ozonated samples demonstrated that

some epoxides can rearrange to other functional groups or be further oxidized with

ozone. Particularly, aryl ethers, -diketones, lactones and anhydrides are thought to be

forming in addition to epoxides. Conversion of epoxides to other functional groups is

another reason why there was no substantial increase of intensity at or near D band.

Recording Raman spectra right after ozonation of an aqueous suspension of SWNT

with a small amount of ozone (~2 mL O3/O2 gaseous mixture, ca. 3 v/v % ozone) gave no

increase of D band, a clear indication that formed ozonides and epoxides do not produce

intense symmetric stretches. In fact, no additional peaks were observed in Raman, even

though the presence of ozonides within first several minutes was proved by changes in

UV, IR, fluorescence and by measuring an amount of evolved oxygen. Raman spectra

show a change in G peak intensity over time, i.e. G peak increased with ozonide

decomposition, but no additional peaks attributable to symmetric stretches of epoxides or

ozonides could be found.

Phenylated () and phenylsulfonated () SWNT samples gave different intensities

of D band. There are three parameters that affect D band intensity: a) quantity of SWNT

in the beam of excitation source, b) degree of functionalization, and c) nature of a group

105

covalently attached to SWNT framework. Disorder, or D band, of phenylsulfonated

SWNT was found to have lower intensity than that in phenylated sample. Due to the

nature of Raman microscope measurement, it is very difficult to have exactly the same

amount of a SWNT in the laser spot. While there is a possibility that the amount of

SWNT-(C6H4-SO3H)n sample in a laser spot was less than that in phenylated sample, the

presence of electron withdrawing group on the ring is likely to decrease intensity of

symmetric stretches of sp3 carbon atoms on SWNT. Both phenylated and

phenylsulfonated samples were prepared with roughly the same amount of SWNT per

studied surface area (for solid state Raman).

Intensity of D band in dodecylated sample (curve ) is thought to be a composite of

symmetric vibrations of sp3 carbons in dodecyl chains and in SWNT. There is no clear

explanation why symmetric stretches of methyl and methylene groups, expected to

manifest themselves at 2850-2950 cm-1

are not present in the spectrum (curve ). Tight

packing of SWNT bundles could have influenced the symmetry of CH3 and CH2

stretches. It should be noted that fluorescence of a dodecylated sample was much higher

than of any other sample in Figure 7. Ideally, fluorescence background should be

subtracted for more accurate comparison of basis functions of D and G peaks.

3.3. Conclusions

The major change in Raman spectra of SWNT after functionalization is a decrease of

G peak. The intensity change is directly proportional to the degree of disruption of

conjugated - system. The D peak was shown to increase, but the growth was only a

fraction of the intensity lost at the G peak. The presence of an electron withdrawing

106

group next to sp3 carbon on SWNT is likely to decrease intensity of sp

3-carbon

symmetric stretches, thus resulting in minimal or no growth of D band in functionalized

samples. Formation of sp3 carbon centers on SWNT during reaction with ozone gave no

additional peaks (for ozonides and for epoxides) and did not result in D band growth. The

shape of the D band changed slightly, likely due to appearance of different functional

groups on SWNT surface. In-plane crystallite size (La) decreased approximately 1.8 times

after two minutes of bubbling O3/O2 gaseous mixture (ca. 3 v/v % ozone). Qualitatively,

phenylsulfonated SWNT were found to have lower D band intensity when compared to

phenylated SWNT. Published cases of D band growth values based on D/G ratios are less

representative of D band growth and more representative of G band decrease. While there

is no doubt that the D band increases, at least in cases with phenylated and dodecylated

samples, this increase is minor when compared to G band decrease. In discussion of

factors influencing D band growth, Raman data of dodecylated SWNT samples should be

treated with caution, since dodecyl chains themselves have a peak at 1300 cm-1

, near the

D band, thus causing it to go to a higher intensity level.

3.4. Experimental Part

SWNT - SDS aqueous suspension was prepared by a standard protocol (see

Experimental Part for Chapter 2). UV-Vis absorption in a range 250 - 900 nm was

recorded on a Cary 4E UV-Vis Spectrophotometer with 1 cm quartz cuvette. Solution

Raman spectra were recorded on a Jobin Yvon Spex Fluorolog with a 4 x 4 mm quartz

cuvette and an external 669.9 nm diode laser. Solid state Raman spectra were recorded on

107

Renishaw Raman Microscope; each spectrum was obtained in one pass with 785 nm

excitation source.

3.5. References and Notes

1. Tuinstra, F.; Koenig, J. L., Raman Spectrum of Graphite. Journal of Chemical

Physics 1970, 53, (3), 1126.

2. Kastner, J.; Pichler, T.; Kuzmany, H.; Curran, S.; Blau, W.; Weldon, D. N.;

Delamesiere, M.; Draper, S.; Zandbergen, H., Resonance Raman and Infrared-

Spectroscopy of Carbon Nanotubues. Chemical Physics Letters 1994, 221, (1-2),

53-58.

3. Pimenta, M. A.; Dresselhaus, G.; Dresselhaus, M. S.; Cancado, L. G.; Jorio, A.;

Saito, R., Studying disorder in graphite-based systems by Raman spectroscopy.

Physical Chemistry Chemical Physics 2007, 9, (11), 1276-1291.

4. Tuinstra, F.; Koenig, J. L., Characterization of Graphite Fiber Surfaces with

Raman Spectroscopy. Journal of Composite Materials 1970, 4, 492.

5. Cancado, L. G.; Takai, K.; Enoki, T.; Endo, M.; Kim, Y. A.; Mizusaki, H.; Jorio,

A.; Coelho, L. N.; Magalhaes-Paniago, R.; Pimenta, M. A., General equation for

the determination of the crystallite size La of nanographite by Raman

spectroscopy. Applied Physics Letters 2006, 88, (16), 163106.

6. Thomsen, C.; Reich, S., Double resonant Raman scattering in graphite. Physical

Review Letters 2000, 85, (24), 5214-5217.

7. Ferrari, A. C.; Meyer, J. C.; Scardaci, V.; Casiraghi, C.; Lazzeri, M.; Mauri, F.;

Piscanec, S.; Jiang, D.; Novoselov, K. S.; Roth, S.; Geim, A. K., Raman spectrum

of graphene and graphene layers. Physical Review Letters 2006, 97, (18), 187401.

8. Kim, U. J.; Furtado, C. A.; Liu, X. M.; Chen, G. G.; Eklund, P. C., Raman and IR

spectroscopy of chemically processed single-walled carbon nanotubes. Journal of

the American Chemical Society 2005, 127, (44), 15437-15445.

9. Hennrich, F.; Krupke, R.; Lebedkin, S.; Arnold, K.; Fischer, R.; Resasco, D. E.;

Kappes, M., Raman spectroscopy of individual single-walled carbon nanotubes

from various sources. Journal of Physical Chemistry B 2005, 109, (21), 10567-

10573.

108

10. Weisman, R. B.; Heymann, D.; Bachilo, S. M., Synthesis and characterization of

the "missing" oxide of C60: [5,6]-open C60O. Journal of the American Chemical

Society 2001, 123, (39), 9720-9721.

11. Banerjee, S.; Hemraj-Benny, T.; Balasubramanian, M.; Fischer, D. A.; Misewich,

J. A.; Wong, S. S., Ozonized single-walled carbon nanotubes investigated using

NEXAFS spectroscopy. Chemical Communications 2004, (7), 772-773.

109

Chapter 4

IR studies of SWNT ozonides and of products of their reactions with

different classes of compounds

110

4.1. Introduction

Ozonation of alkenes is known to proceed at very low temperatures with formation

of primary ozonides.1 Formed 1,2,3-trioxolanes are typically not stable above – 100 C.

1-4

Hull et al.2 determined that the majority of 1,2,3-trioxolanes studied rearranged to 1,2,4-

trioxolanes at temperatures above – 100 C. Kohlmiller and Andrews1 and Samuni and

Haas5 assigned the following IR active bands to primary ozonide of ethylene: 846 cm

-1

( sym O-O-O stretch), 927 cm-1

(C-O stretch) and 983 cm-1

(C-O stretch). Hull et al.2

found characteristic bands for primary ozonides of different alkenes to be in 850-1050

cm-1

region. Andrews and Kohlmiller6 compared strong IR bands of 1,2,3-trioxolanes of

propene, trans-2-butene, 2-methylpropene, tetramethylethylene and trans-

diisopropylethylene. The C-O stretches for all ozonides were found in the region 850-

1000 cm.-1

Contrary to small molecule alkenes, ozonides of fullerene (C60O3) 7 and of carbon

nanotubes (SWNT(O3)n) (this work, Chapter 2) were found to be significantly more

stable. Heymann et al.7 found that at 23 C ozonide C60O3 has a lifetime ca. 22 minutes in

toluene. Lifetimes of SWNT(O3)n, depending on tube type, were found in the range 0.5 –

2 minutes at 20 C (this work, Chapter 2). Thus, it is reasonable to expect characteristic

IR active stretches of primary ozonides of C60 and SWNT at room temperature. This

chapter will discuss the influence of oxidation with ozone/oxygen gaseous mixture on IR

spectra of carbon nanotubes (SWNTs). A brief summary of published data on ozonation

of SWNT is discussed in the introduction of Chapter 2. The majority of articles discussed

had IR spectra of final products with carboxylic, ester, quinone, and other functional

111

moieties on SWNT, and none provided any spectral examination of SWNT ozonides. To

date, IR verification of the existence of SWNT ozonides has not been published. Kinetics

determined in this chapter will be compared to those discussed for fluorescence and UV

techniques (Chapter 2).

4.2. Discussion and results

An extensive IR study has been accomplished here in search of peaks that could be

assigned to 1,2,3-trioxolanes or products of their decomposition, like epoxides and

oxidoannulenes.8 For the majority of spectra presented in this section, constant purging of

the IR chamber with nitrogen was done to decrease the intensity of water and carbon

dioxide peaks.

The SWNT film was formed on a BaF2 window by adding several drops of SWNT

slurry freshly dispersed in benzene. Dry SWNT film was ozonated for 30 seconds with a

stream of O3/O2 gaseous mixture (ca. 1.5 v/v % ozone) and the absorption IR spectrum

measured (Figure 1).

No peaks attributable to ozonides or epoxides could be found. Slow absorbance drift

over time was observed. Peaks for ozonides and epoxides were expected in the area

800 – 1200 cm-1

. Kamaras et al.9 and Hu et al.

10 attributed absorbance depletion at

frequencies below 1000 cm-1

(see Figure 1) to covalent modification of metallic SWNT.

112

Wavenumber (cm-1

)

1500300045006000

Ab

so

rba

nce

(a

.u.)

0.15

0.20

0.25Before O3

2 min

5.4 min

15.8 min

Figure 1. Absorbance IR of SWNT film before and after purging with O3/O2 gaseous

mixture (ca. 1.5 v/v % ozone) for 30 sec. Curves are: - before ozonation, - two

minutes after ozonation, – 5.4 min after O3, - 15.8 min after O3.

The experiment was repeated with fullerene C60, since its ozonides have been

reported in the literature.7, 8

The T1u vibrations of C60 at 1182 and 1429 cm-1

observed in

this work (Figure 2) are in good agreement with published data.11

Fullerene ozonide

(C60O) lifetime was reported ~ 22 min at 23 C 8 and deemed long enough to be

detected by IR. It was reasonable to expect to see asymmetric stretches imparted by the

1,2,3-trioxolane moiety (Figure 2).

113

Wavenumber (cm-1)

1500300045006000

Ab

so

rban

ce

(a

.u.)

0.04

0.08

0.12

before O3

2 m

5.4 m

2 m

4.8 m

after 1st O3

after 2nd

O3

Figure 2. Absorbance IR of ozonated C60 film on BaF2 window before and after purging

with O3/O2 gaseous mixture (ca. 1.5 v/v % ozone) for 20 sec. Curves are: - before

ozonation, - two minutes after 1st ozonation, – 5.4 min after 1

st O3, - 2 min after

2nd

O3, - 4.8 min after 2nd

O3. Sharp peaks are 1182 and 1429 cm.-1

A small background shift was observed at frequencies above 4500 cm-1

after each

ozonation; no peaks attributable to ozonides could be found.

An attempt was made to run IR of SWNT buckypaper without IR windows. Though

such paper had high absorption below 1500 cm-1

, its transmission increased greatly right

after ozonation. Spectrum acquisition was performed on buckypaper stretched across a

4 x 6 mm opening as shown in the pictures below.

114

Figure 3. Buckypaper stretched across a 4 x 6 mm opening for windowless IR

measurement. (A-B) SWNT film attached to adhesive tape, (C-D) IR holder with 4 x 6

openning, (E-F) film inside the holder.

Buckypaper was flushed with O3/O2 gaseous mixture (ca. 1.5 v/v % ozone) for

30 sec and spectra recorded (Figure 4).

Wavenumber (cm-1

)

150030004500

Abso

rba

nce

(a.u

.)

1.3

1.7

2.1

2.5

2.9

before

2 m

10.4 m

2 m

10.8 m

2 m

45.1 m

2 m

11.8 m

after 1st O

3

after 2nd

O3

after 3rd

O3

after 4th O

3

Figure 4. Absorbance IR of SWNT buckypaper after four 30 sec purging with O3/O2

gaseous mixture (ca. 1.5 v/v % ozone). Symbols denote times after the beginning of each

ozonation. (Windowless IR.)

No peaks below 1500 cm-1

attributable to ozonides could be found. Spectra were

fairly noisy at frequencies below 1050 cm-1

. Absorbance drift over time was observed as

115

in previous experiments. Two of the most prominent peaks in the spectra, near 1220 and

1570 cm-1

, are likely to be from asymmetric stretches of single (C-C-EWG) and double

(C=C-EWG) bonds on SWNT surface located in the vicinity of electron-withdrawing

groups (EWG). Drifts of peaks’ maxima towards higher frequencies with subsequent

ozonations are due to an increase of the number of electron withdrawing groups on the

SWNT surface.

Raman and IR spectra overlay shows the proximity of D and G bands and two major

peaks in the IR spectrum of ozonated SWNT:

Absorb

ance (

a.u

.) (

IR)

0.48

0.52

0.56

0.60IR after O3

D

G

G'

Wavenumber (cm-1

)

0 500 1000 1500 2000 2500

Counts

x10

5 (

Ram

an)

0.0

0.4

0.8

1.2

Raman before O3

ab

Figure 5. Overlay of absorbance IR spectrum of ozonated SWNT with Raman spectrum

of pristine SWNT.

The experiment was repeated with a longer ozonation period to bring the absorbance

at 1000 cm-1

to below 1.5 a.u. (Figure 6).

116

Wavenumber (cm-1

)

1000200030004000

Ab

so

rban

ce

(a

.u.)

0.75

1.50

2.25

3.00

before

6 m

18.3 m

6 m

46.1 m

6 m

50.3 m

6 m

29.4 m

after 1st O3

after 2nd

O3

after 3rd

O3

after 4th O3

Figure 6. Absorbance IR of SWNT buckypaper after four 5 min purging with O3/O2

gaseous mixture (ca. 1.5 v/v % ozone). Symbols denote times after the beginning of each

ozonation. (Windowless IR.)

The signal to noise ratio increased dramatically for area below 1500 cm-1

(compare

Figures 4 and 6), but no peaks attributable to ozonides could be found. Absorbance drift

was substantial after the first 5 min of purging with O3/O2 gaseous mixture (compare

curves and ).

117

Wavenumber (cm-1

)

1000200030004000

Absorb

ance (

a.u

.)

0.6

0.8

1.0

1.2

after 4th

O3

17

65

34

70

12

22

15

72

13

71

11

44

Figure 7. Absorbance IR of SWNT buckypaper after fourth 5 min purging with O3/O2

gaseous mixture (ca. 1.5 v/v % ozone).

Additional peaks found in SWNT spectra after multiple oxidations were 3470, 1765,

1371 and 1144 cm-1

(Figure 7). The peak at 1765 cm-1

could be attributed to 1,2-

diketones, anhydrides, unsaturated esters or a combination of thereof; 1371 cm-1

to

epoxides and 1144 cm-1

to diaryl ethers. Absorbance kinetics were measured and are

presented below. Found peaks’ maxima were close to those observed by Cai et al.12

118

10 20

Absorb

ance (

a.u

.)

1.4

1.5

1.6

1.7

1.8

1.9

Time (min)

10 20 30 40 50

1.2

1.3

1.4

1.5

1.6

1.7

10 20 30 40 50

1.0

1.1

1.2

1.3

1.4

1.5 Abs at 1315 cm-1

Abs at 1625 cm-1

Abs at 2250 cm-1

A

A B

A B

after 3rd

O3after 2nd

O3after 1st O3

Figure 8. Absorbance IR kinetics of SWNT buckypaper after the first three 5 min

purging with O3/O2 gaseous mixture (ca. 1.5 v/v % ozone). For comparison purposes,

absorbance and time axes have the same scale on all three graphs. Each symbol is marked

with a distinct wavenumber at which kinetics was monitored. Schematic diagram: A –

ozonide decay, B – structural rearrangements. Left: after 1st O3, Middle: after 2

nd O3,

Right: after 3rd

O3.

Absorbance changes within the first ten minutes were observed after the first three

five-minute ozonations. The absorbance change is believed to be associated with ozonide

decay. An upward movement, i.e. absorbance increase, could be caused by an increase of

electron density on SWNT after ozonide decay.

119

Lack of success with finding characteristic ozonide peaks prompted the use of a KBr

window in place of BaF2. Pictures of SWNT film on KBr are shown below.

Figure 9. SWNT film on KBr window (Fisher Scientific) is shown from different angles.

Solid state Raman was recorded on KBr window (shown in Figure 9) before and

after ozonation (Figure 10; ozonation was monitored by IR). The intensity of D band was

found unchanged after extensive oxidation with ozone. The 785 nm laser was used for

excitation.

Wavenumber (cm-1

)

500 1000 1500 2000 2500

Counts

x10

5

0.0

0.4

0.8

1.2

before O3

after O3

D

G

G'

1292 1591

Counts

x10

3

1.7

7.0

D

G

Figure 10. Raman spectra of SWNT film on KBr window before and after several rounds

of ozonation. Zoom-in shows D band did not change. G and G’ bands were bleached

substantially. The 785 nm laser was used for excitation.

Raman spectra shown in Figure 10 correspond to curves and in Figure 11.

120

Wavenumber (cm-1

)

200040006000

Absorb

ance (

a.u

.)

0.2

0.4

0.6

0.8

1.0

before

1 m

8.9 m

1.5 m

23.6 m

1.5 m

19.5 m

1.5 m

9.3 m

1.5 m

5.8 m

after 1st O3

after 6th O3

after 7th O3

after 8th O3

after 10th O3

Figure 11. Absorbance IR spectra of SWNT film on KBr window after ten rounds of

ozonation. Oxidations 1 through 5 were done with 1 mL of O3/O2 gaseous mixture (ca.

1.5 v/v % ozone). Ozonations 6 through 10 were performed by blowing O3/O2 gaseous

mixture (ca. 1.5 v/v % ozone) onto SWNT film for 30 sec. Some spectra are not shown to

avoid clutter. Curve is the spectrum of a pristine SWNT film.

Spectra on KBr were found to swing around a pivoting point near 5000 cm-1

.

Absorbance drift over time was larger than what was observed for BaF2 and windowless

IR experiments. Compare curves and in Figure 11. An initial population of low

frequency vibrations (1000 cm-1

) and a depopulation of electronic transitions (7000 cm-1

)

drifted over time to repopulate electronic transitions. Such a swing is believed to be

associated with the charge transfer from ozonated SWNT to KBr or vice versa.

Below are zoomed-in spectra demonstrating a swing motion of the spectra.

121

Wavenumber (cm-1

)

1500300045006000

Absorb

ance (

a.u

.)

0.45

0.60

1 m

15.6 m

1 m after 5th O3

after 4th O3

Absorb

ance (

a.u

.)

0.50

0.65

1 m

11.2 m

1 m after 3rd

O3

after 2nd

O3

A

B

Figure 12. Absorbance IR spectra of SWNT film on a KBr window between 2nd

and 5th

rounds of ozonation. Oxidations were done with 1 mL of O3/O2 gaseous mixture (ca. 1.5

v/v % ozone). Legends to symbols denote times after the beginning of the corresponding

ozonation. (A) Spectra after 2nd

, before 3rd

and after 3rd

ozonation, (B) spectra after 4th

,

before 5th

and after 5th

O3. See note to Figure 11 for more details.

A depopulation of NIR absorbance was observed for SWNT – SDS aqueous

suspensions right after ozonation (Chapter 2, Figure 8). A similar tendency was observed

for a solid SWNT film near the 7000 cm-1

region in Figure 12A. Ozonation resulted in a

122

depopulation of electronic transitions in NIR, with a slow recovery over time (Figure

12A, curves and ). Swing motion was accompanied by an overall bleaching (Figure

12B).

A fast scanning (2 scan averaging) on BaF2 window was performed to measure the

kinetics right after an ozone injection. The O3/O2 gaseous mixture (ca. 3 v/v % ozone)

was injected directly into IR chamber with the needle pointed towards SWNT film. The

chamber was under a constant nitrogen flush. The first point was acquired 10 sec after the

injection.

Time (min)

0 3 6 9 12

Absorb

ance a

t 2250 c

m-1

(a.u

.)

0.64

0.65

A B

Figure 13. Absorbance IR kinetics of SWNT film on BaF2 window at 2250 cm-1

.

Oxidation was done with 5 mL of O3/O2 gaseous mixture (ca. 1.5 v/v of ozone). A – fast

decay; B – slow rearrangement.

A five parameter two-exponential decay formula was used for the regression. A

decay rate b was calculated to be 2.6 min-1

, which corresponds to ~ 0.4 min. Kinetics

123

measurement were repeated on a KBr window and a comparable rate was obtained

(Figure 4).

Time (min)

0 5 10 15 20

Absorb

ance a

t 2250 c

m-1

(a.u

.)

0.76

0.77

0.78 after 1st O

3

Figure 14. Absorbance IR kinetics of SWNT film on KBr window at 2250 cm-1

.

Oxidation done with 3 mL of O3/O2 gaseous mixture (ca. 1.5 v/v of ozone).

The decay rate obtained from regression was 2.1 min-1

, which corresponds to

~ 0.48 min.

The following conclusions can be made for IR spectra of ozonated SWNT. No

characteristic peaks could be found in the area typical for 1,2,3-trioxolanes or peroxides.

Commonly throughout all experiments, absorbance drift was observed. In the case with

KBr window used as a support, the change resembled a swing with a pivoting point in the

range 4000-5000 cm-1

. In the case with BaF2 absorbance decreased over time at

2250 cm.-1

In the case with windowless IR, the change led to an increase of absorbance in

the range 600 – 4000 cm-1

during the first ten minutes. Regression yielded decay rates

2.1 and 2.6 min-1

for KBr and BaF2 experiments. Ozonide peaks could not be found for

124

C60O3 either. There is no clear understanding why ozonides do not have characteristic

C-O bands in IR in the area 850 – 1050 cm-1

. The absorbance drift is thought to be

associated with ozonide decay. Interestingly, for films on KBr and BaF2 windows

absorbance decreased over time at 2250 cm-1

; for windowless IR on buckypaper the

absorbance increased during the first ten minutes. IR window may be involved in a

charge transfer to or from ozonated SWNT.

Influence of amines and solvents on ozonated SWNT

Ozonation of SWNT film with O3/O2 gaseous mixture to a moderate degree yields a

characteristic wavy curve (Figure 15, curve ). This curve was used as a reference for

reactions of ozonated SWNT with amines.

Wavenumber (cm-1

)

1500300045006000

Ab

so

rba

nce

(a

.u.)

0.1

0.2

0.3

0.4

before

after 1st O

3

after 2nd

O3

after 3rd

O3

after 4th

O3

f

10001500

a

c

b

b'

ed

d'

Figure 15. Absorbance IR spectra of SWNT film on BaF2 window ozonated four times.

Peak maxima values (cm-1

): a = 1760, b = 1550, b' = 1559, c = 1369, d = 1190, d' = 1206,

e = 1163, f = 3495.

125

A comparison was made between dry ozonated SWNT (Figure 15) and SWNT

ozonated in suspension. Three solvents used for this comparison were methanol, ethanol

and trifluoroethanol. Dilute suspensions of SWNT in solvents of interest were bath

sonicated for 1 hour, O3/O2 gaseous mixture (ca. 3 v/v % ozone) was bubbled through

suspensions for 1 min at r. t. and IR spectra recorded 1 hour after ozonation. Spectra were

overlaid for comparison and shown below (Figure 16).

Wavenumber (cm-1

)

1500300045006000

Absorb

ance (

a.u

.)

MeOH

EtOH

CF3CH

2OH

Figure 16. A solvent shielding effect on SWNT ozonation degree. Bubbling O3/O2

gaseous mixture (ca. 3 v/v % ozone) for 1 min at r. t. through SWNT suspension in

methanol (), ethanol () and trifluoroethanol () yielded spectra identical to those

obtained after dry ozonation. Spectra were recorded 1 hour after ozonation as a dry

SWNT film on BaF2 window.

All three solvents were found unreactive with ozonated SWNT. No IR active peaks

attributable to solvents’ residues were seen on spectra. Oxidation in ethanol resulted in

the lowest ozonation degree, while trifluoroethanol gave the highest oxidation degree. It

was concluded that each solvent possesses a unique shielding ability, preventing ozone

126

from reacting with SWNT. An alternative explanation would be a different degree of

ozone solubilization in each solvent. Chen et al.13

reported that perfluoroethers have a

good solubility of ozone.

Below is an example of a reagent, acetic acid, which does not react with ozonated

SWNT.

Wavenumber (cm-1

)

1500300045006000

Absorb

ance (

a.u

.)

0.00

0.15

0.30

0.45

0.60

before

after

reference

AcOH

Figure 17. Addition of acetic acid to ozonated SWNT film on BaF2 window did not

affect SWNT curve. Curves are: – SWNT film before O3, – reference, mixed

SWNT film with acetic acid for 1 min then dried with a heat gun, –SWNT film purged

with O3/O2 gaseous mixture (ca. 1.5 v/v % ozone), – ATR converted spectrum of an

authentic sample of acetic acid for comparison. Curves and were measured on the

same film. All curves except are absorbance IR spectra.

A few drops of n-butyl amine were added to SWNT plate treated first with ozone and

then with acetic acid. The BaF2 plate was dried with a heat gun and an IR spectrum

obtained (Figure 18). The arrows schematically show repopulation of electronic

transitions at 7000 cm-1

and depopulation of vibrational ones at 1000 cm-1

. A spectrum of

127

an authentic sample of n-butyl amine (curve ) is provided for comparison. It is believed

that the first step in the reaction between amine and SWNT is an electron transfer from

electron rich amine to ozonated SWNT. It is not clear why ozonated SWNT plays a role

of an oxidizer. As demonstrated by XPS (Chapter 7), such a reaction does not take place

between pristine SWNT and amines at r. t.

Wavenumber (cm-1

)

1500300045006000

Absorb

an

ce (

a.u

.)

0.00

0.15

0.30

0.45

0.60

before

after O3 & AcOH

added n-BuNH2

n-BuNH2

15003000

c

de

a b

c

ab

de

Figure 18. A reaction of ozonated SWNT film with n-butyl amine. A repopulation of

electronic transitions and a depopulation of vibrational ones are emphasized with arrows.

Curves are: – SWNT film before O3, – an ozonated SWNT film treated with acetic

acid, –added n-BuNH2 to SWNT treated with O3 and AcOH, reacted for 1 min and

dried with a heat gun, – ATR converted spectrum of an authentic sample of n-butyl

amine for comparison. Curves , and were measured on the same film. All curves

except are absorbance IR spectra.

128

SWNT films, ozonated for 30 seconds, were treated with n-butylamine forty minutes

and two days after ozonation. In both cases amines were shown to react with SWNT

(Figure 19). The sample with a forty minute delay was dried on a vacuum pump before

IR measurements. Sample with a two-day delay was dried with a heat gun.

Wavelength (cm-1

)

1500300045006000

Ab

so

rba

nce

(a

.u.)

0.00

0.15

0.30

after O3

2 day + amine

40 min + amine

n-BuNH2

Figure 19. A reaction of ozonated SWNT film with n-butyl amine 40 min and 2 days

after ozonation. Repopulation of electronic transitions and depopulation of vibrational

ones is emphasized with arrows. Curves are: – SWNT film after O3, – ozonated

SWNT film treated with n-butylamine 2 days after O3, –added n-BuNH2 to SWNT

forty minutes after O3, – an ATR converted spectrum of an authentic sample of

n-butylamine is overlaid for comparison. Curve was scaled down by a factor of 0.87 to

compensate for the higher load of SWNT on IR window. Curves and were

measured on the same film. All curves except are absorbance IR spectra.

129

Secondary and tertiary amines were also tested. The resultung spectra are shown

below. Triethyl amine was shown to react with SWNT, which further supports the

likelihood of an electron transfer between ozonated SWNT and amine.

Wavenumber (cm-1

)

1500300045006000

Ab

so

rba

nce

(a

.u.)

0.00

0.15

0.30

0.45

0.60

before

after

reference

Et3N

15003000

de

abc

a

de

Figure 20. A reaction of ozonated SWNT film with triethyl amine. Curves are: –

SWNT film before O3, – ozonated SWNT film treated with amine, – added amine

to pristine SWNT, reacted for 1 min and dried with a heat gun, – an ATR converted

spectrum of an authentic sample of triethyl amine is overlaid for comparison. Curves

and were measured on the same film. All curves except are absorbance IR spectra.

Reaction between triethyl amine and SWNT does not seem to be as effective as

between primary amines and ozonated SWNT. The product is thought to be an

ammonium salt. The first step is an electron transfer from amino group to ozonated

SWNT; the second step is thought to be a covalent attachment of radical Et3N+

to the

surface of SWNT. Similar types of reactions had been studied by Isobe et al. on a

fullerene C60 substrate.14

The author proposed for a reaction of fullerene C60 with

dialkylamine a formation of a long lived pair of aminium radical R2NH+ and C60

–

130

radical anion. Such pair existed only in the absence of oxygen. When the mixture was

exposed to the air, two radicals reacted immediately. The author observed a trace of

hydrogen peroxide, which was formed during oxygen reduction by an aminium radical.

Wavenumber (cm-1

)

1500300045006000

Absorb

ance (

a.u

.)

0.00

0.15

0.30

0.45

0.60

before

after

reference

Et2NH

Figure 21. A reaction of ozonated SWNT film with diethyl amine. Curves are: –

SWNT film before O3, – ozonated SWNT film treated with amine, – reference,

added diethyl amine to SWNT, reacted for 1 min and dried with a heat gun, – ATR

converted spectrum of an authentic sample of diethyl amine is overlaid for comparison.

Curves and were measured on the same film. All curves except are absorbance

IR spectra.

Reaction of ozonated SWNT with diethyl amine was essentially the same as with

n-BuNH2 and Et3N.

131

Reaction between 2-methoxyethyl amine and ozonated SWNT is shown below.

Wavenumber (cm-1

)

1500300045006000

Absorb

ance (

a.u

.)

0.00

0.15

0.30

0.45

0.60

before

after

reference

MeOCH2CH

2NH

2

ba

c

15003000

ba

c

de

f

de

f

Figure 22. A reaction of ozonated SWNT film with 2-methoxyethyl amine. Curves are:

– SWNT film before O3, – ozonated SWNT film treated with amine, – reference,

added amine to SWNT, reacted for 1 min and dried with a heat gun, – ATR converted

spectrum of an authentic sample of amine is overlaid for comparison. Curves and

were measured on the same film. All curves except are absorbance IR spectra.

Enrichment of amines and amides on the surface of SWNT (IR monitoring)

An enrichment of n-butyl amine on the surface of SWNT was successfully realized

by cycling ozonation with amine addition. The experiment was performed on the same

BaF2 plate. SWNT film ozonation was alternated with amine addition. Butyl amine, used

in the experiment, was chosen for its low boiling point. A heat gun was used to remove

traces of unreacted amine from the surface of SWNT. Distinct amide, C(=O)O, OH and

NH stretches were observed in the IR spectra (Figure 23).

132

Wavenumber (cm-1

)

1500300045006000

Ab

so

rba

nce

(a

.u.)

0.1

0.2

0.3

0.4

before

after 1st O3

added amine (1st cycle)

after 2nd O3

added amine (2nd cycle)

after 3rd O3

added amine (3rd cycle)

after 4th O3

Figure 23. Cycled reaction of ozonated SWNT film with n-butyl amine. Curves are:

– SWNT film before O3, – ozonated SWNT film, – ozonated SWNT film treated

with amine (1st cycle), – after second ozonation, - added amine (2

nd cycle),

– after third ozonation, - added amine (3rd

cycle), - after forth ozonation. Water

spikes in the area 1600 cm-1

were removed manually to improve clarity. All curves are

absorbance IR spectra.

SWNTs were debundled in ethanol by bath sonication, and then added as a slurry to

the surface of BaF2 window. The spectrum of pristine SWNT was measured and the film

was subjected to a stream of O3/O2 gaseous mixture (ca. 1.5 v/v % ozone) for 1 min at r.

t. The IR spectrum of ozonated SWNT was recorded and several microliters of n-butyl

amine added to the surface of BaF2. The plate was kept under the cover for 1 min and

then dried on vacuum pump for 3 min. Ozonation, IR measurements and reactions with

amine were repeated for three more cycles (Figure 23). A zoom-in of the region 4000-

700 cm-1

is shown in Figure 24.

133

Wavenumber (cm-1

)15003000

Absorb

ance (

a.u

.)

0.10

0.15

0.20

0.25

ab

c

d

e

f g

h

i

j

Figure 24. Cycled reaction of ozonated SWNT film with n-butyl amine. Curves are:

– SWNT film before O3, – ozonated SWNT film, – ozonated SWNT film treated

with amine (1st cycle), – after second ozonation, - added amine (2

nd cycle),

– after third ozonation, - added amine (3rd

cycle), - after forth ozonation. Water

spikes in the area 1600 cm-1

were removed manually to improve clarity. All curves are

absorbance IR spectra.

Peak data for Figure 24 are summarized in Table 1 below.

Table 1. IR peaks assignment for amine/amide enriched SWNT film

Abbreviation Wavenumber Peak assignment

a 3300 N-H stretch; amide or amine;

b 2982- 2829 C-H stretches (CH2, CH3)

c 1760 1,2-diketones, anhydrides, unsaturated esters

d 1672 amide C=O stretch

e 1573 C=C and N-H bending from amides/amines

f 1451 CH2

g 1375 C-N stretch (amide); CH2 scissoring; epoxides

h 1196 CH2 scissoring

i 1118 C-N stretch in amine

j 683 N –H wagging

134

The observed increase in peak intensities is thought to be directly linked to the

amount of amine or amide attached to the surface of SWNT.

4.3. Conclusions

IR monitoring of dry ozonated SWNT film at 2250 cm-1

revealed exponential

absorbance change over time. The rate of the change was found to be close to the rates

obtained with fluorescence and NIR absorbance techniques for ozonated SWNT. For

experiments performed on KBr window, ozonated SWNT demonstrated an initial

bleaching of absorbance in a wide range, above 5000 cm-1

. The change was repeatable

with each injection. Bleaching at frequencies greater than 5000 cm-1

was accompanied by

an absorbance increase at frequencies lower than 4000 cm-1

.

Ozonated SWNT films were shown to react with amines of different structure

(primary, secondary and tertiary). Reaction with amines was independent of time period

after ozonation. SWNT ozonated two days prior to addition of amine were shown to

undergo the same reaction as the ones that were ozonated right before amine addition.

An enrichment of amines/amides on the surface of SWNT was successfully realized

by cycling ozonation, addition of neat amine and vacuum drying.

4.4. Experimental Part

Ozonation procedures

Ozone was generated by passing oxygen (industrial grade, Matheson Tri-Gas) with a

flow rate 1/16 L/min at r.t. through a high frequency corona discharge ozonator

135

(GE60FM, Yanco Industries Ltd, www.ozoneservices.com) set to the maximum output

(power level 10). The ozonator was idled before use for at least 7 min to reach the

maximum output as instructed by the manufacturer. Two different ozonation procedures

used in this chapter are described below.

Continuous ozonation of SWNT suspensions in a solvent of interest. A gaseous

mixture of O3/O2 (ca. 3 v/v % ozone) was bubbled through SWNT suspension in a test

tube for the desired period of time.

Gaseous ozone/oxygen injection with a syringe. Ozone was collected in a plastic

syringe for 1 min (3 mL volume) or 2 min (5-10 mL volume), capped with a plunger and

the desired volume released by blowing on top of SWNT film.

Continuous ozonation of dry SWNT films. The oxygen flow rate was kept at 1/16

L/min. Unless otherwise noted, a SWNT film was ozonated for 1 min at r. t. in a small

chamber by holding 1/16” ID Teflon tubing above the dry SWNT film.

Routine IR spectra measurements

Bath sonicated SWNTs (HipCo Lab, batch 162.8 or 161.1, Rice University) in

benzene were used for IR measurements. The Fourier Transform Infrared Spectrometer

(JASCO FT/IR-660 Plus) chamber was flushed with nitrogen unless otherwise noted. All

measurements were performed on dry films at or near r. t.; no special temperature

monitoring was made inside the chamber during IR measurements. Experiments had

either attenuated total reflectance (ATR) or absorbance setup which is noted on each

figure; KBr and BaF2 windows were used for absorbance IR measurements.

Some of the spectra of carbon nanotube films, or buckypapers, were recorded

without IR windows. A special film holder with 3 x 5 mm rectangle opening was

136

constructed to hold the buckypaper. Films were prepared by a dropwise addition of a

concentrated SWNT suspension in benzene onto the surface of a filter paper followed by

careful SWNT film removal. Paper twisting and turning was required to visually detach

nanotube film from the filter paper before film removal. Extreme care should be observed

to ensure nanotube film integrity.

Some spectra have a carbon dioxide peak present right after ozonation. This is

caused by an introduction of air into the chamber during IR holder handling. Over time, a

constant nitrogen flush removed carbon dioxide from the chamber. Nearly all FT/IR

spectra were acquired with constant nitrogen gas purge to reduce intensity of carbon

dioxide and water peaks.

IR kinetic measurements

The spectrometer was purged with nitrogen before and during measurements. For

kinetic studies 3 mL of O3/O2 gaseous mixture (ca. 3 v/v % ozone) was injected directly

into the IR chamber with a needle pointed towards SWNT film on BaF2 or KBr window.

137

4.5. References and Notes

1. Kohlmiller, C. K.; Andrews, L., Infrared-Spectrum of the Primary Ozonide of

Ethylene in Solid Xenon. Journal of the American Chemical Society 1981, 103,

(10), 2578-2583.

2. Hull, L. A.; Heicklen, J.; Hisatsun.Ic, Low-Temperature Infrared Studies of

Simple Alkene-Ozone Reactions. Journal of the American Chemical Society

1972, 94, (14), 4856-4864.

3. Hisatsune, I. C.; Kolopajlo, L. H.; Heicklen, J., Low-Temperature IR Studies of

Some Chloroethylene-Ozone Reactions. Journal of the American Chemical

Society 1977, 99, (11), 3704-3708.

4. Mile, B.; Morris, G. W.; Alcock, W. G., Infrared-Spectra and Kinetics of

Decomposition of Primary Ozonides in the Liquid-Phase at Low-Temperatures.

Journal of the Chemical Society-Perkin Transactions 2 1979, (12), 1644-1652.

5. Samuni, U.; Haas, Y., An ab-initio study of the normal modes of the primary and

secondary ozonides of ethylene. Spectrochimica Acta Part a-Molecular and

Biomolecular Spectroscopy 1996, 52, (11), 1479-1492.

6. Andrews, L.; Kohlmiller, C. K., Infrared-Spectra and Photochemistry of the

Primary and Secondary Ozonides of Propene, Trans-2-Butene, and

Methylpropene in Solid Argon. Journal of Physical Chemistry 1982, 86, (23),

4548-4557.

7. Heymann, D.; Bachilo, S. M.; Weisman, R. B.; Cataldo, F.; Fokkens, R. H.;

Nibbering, N. M. M.; Vis, R. D.; Chibante, L. P. F., C60O3, a fullerene ozonide:

Synthesis end dissociation to C60O and O-2. Journal of the American Chemical

Society 2000, 122, (46), 11473-11479.

8. Weisman, R. B.; Heymann, D.; Bachilo, S. M., Synthesis and characterization of

the "missing" oxide of C-60: [5,6]-open C60O. Journal of the American Chemical

Society 2001, 123, (39), 9720-9721.

9. Kamaras, K.; Itkis, M. E.; Hu, H.; Zhao, B.; Haddon, R. C., Covalent bond

formation to a carbon nanotube metal. Science 2003, 301, (5639), 1501-1501.

10. Hu, H.; Zhao, B.; Hamon, M. A.; Kamaras, K.; Itkis, M. E.; Haddon, R. C.,

Sidewall functionalization of single-walled carbon nanotubes by addition of

dichlorocarbene. Journal of the American Chemical Society 2003, 125, (48),

14893-14900.

138

11. Chase, B.; Herron, N.; Holler, E., Vibrational Spectroscopy of C-60 and C-70

Temperature-Dependent Studies. Journal of Physical Chemistry 1992, 96, (11),

4262-4266.

12. Cai, L. T.; Bahr, J. L.; Yao, Y. X.; Tour, J. M., Ozonation of single-walled carbon

nanotubes and their assemblies on rigid self-assembled monolayers. Chemistry of

Materials 2002, 14, (10), 4235-4241.

13. Chen, Z. Y.; Ziegler, K. J.; Shaver, J.; Hauge, R. H.; Smalley, R. E., Cutting of

single-walled carbon nanotubes by ozonolysis. Journal of Physical Chemistry B

2006, 110, (24), 11624-11627.

14. Isobe, H.; Tanaka, T.; Nakanishi, W.; Lemiegre, L.; Nakamura, E., Regioselective

oxygenative tetraamination of 60 fullerene. Fullerene-mediated reduction of

molecular oxygen by amine via ground state single electron transfer in dimethyl

sulfoxide. Journal of Organic Chemistry 2005, 70, (12), 4826-4832.

139

Chapter 5

Reaction of ozonated SWNT with electron rich nucleophiles (amines,

thiols and other)

140

5.1. Introduction

Reactions of ozonated SWNT with electron rich nucleophiles like amines and thiols

have not been discussed in the scientific literature yet. This chapter will expand on the

subject of reacting ozonated SWNT and C60 with amines and thiols. Studies will be

supported by SWNT fluorescence, UV absorbance, XPS, NMR and visual color changes.

The greatest advantage of this methodology is the speed of the reaction. Reaction of

amines with ozonated C60 was on the order of subseconds for a diluted fullerene solution

in toluene. Reaction of amines with SWNT was found to be slower, with a rate ca.

2 min-1

for a diluted SWNT suspension in ethanol. The presented experiments were

conducted in different solvents (water, ethanol and toluene), at r. t. and required only

minutes for completion. Particularly, an experiment with N,N,N’,N’- tetramethyl–p-

phenylenediamine (TMPD) was shown to reach completion within ca. 10 min after

mixing of ozonated SWNT and TMPD. (This strategy has been successfully extended

onto a covalent attachment of amino acids to SWNT sidewall and is discussed in chapters

6 and 7. For IR monitoring of reactions between ozonated SWNT and amines see Chapter

4.)

Closely related work performed on fullerene C60 with four peroxide moieties

(–OOBut), covalently attached to it, has been studied and published very recently.

1

Particularly, primary and secondary amines were shown to get covalently attached to the

fullerene framework with peroxides still intact. Reactions were reported to proceed fast.

Peroxides significantly decreased the activation energy for reaction between C60(OOBu-

t)4(O)1 and amines. While the reader may find it unique that amines were not oxidized by

141

peroxy-t-butyl moieties on C60, researchers who conducted this study were able to obtain

X-ray structures of some of the compounds.1

Reactions of amines with underivatized and derivatized fullerenes have been studied

and reported in peer reviewed journals for more than a decade.1-9

A typical reaction of

C60 with amine requires very long times (days to weeks at r. t.) or heating.8

Methodologies developed in this work and the one published by Hu et al.1 have a great

advantage of being fast.

It should be noted that the lifetime of SWNT ozonides (1,2,3-trioxolanes) was

determined in this work to be in the range 20 – 200 sec at r. t. (see Chapter 2). In view of

this result, the term “ozonated SWNT” can mean both SWNT with ozonides still present

on its surface and SWNT with all ozonides decomposed. Several experiments in this

work were specifically designed to demonstrate that amines can react with ozonated

SWNT in the absence of ozonides (Chapter 4, Figure 19).

The only readily available example of reaction of ozonated SWNT with electron rich

nucleophile is the reaction of SWNTO3 with dimethyl sulfide (DMS) at -78 C.10

This

easily oxidizable reagent was added to SWNT ozonated for an extended period of time.

The author was assuming an analogy between the reduction of secondary ozonides of

small molecules and the reduction of primary ozonides on the surface of SWNT. The

goal was to preserve ketone and aldehyde groups on the surface of SWNT, while

converting dimethyl sulfide into dimethyl sulfoxide. Ozonide decomposition rates

obtained in this work (Chapter 2) clearly indicated that at temperature – 78 C, 1,2,3-

trioxolanes should be very stable. Also, formation of a secondary ozonide in place of a

142

primary one is not likely for rigid SWNT structure. Article did not mention any attempt

to covalently attach dimethyl sulfide to the surface of SWNT. It is likely that DMS

reacted not only with 1,2,3-trioxolanes, but also with SWNT sidewall. Provided XPS

spectra did not cover S 2s or S 2p regions, thus it is not clear if any sulfur was present on

the surface of SWNT. While conversion of organic sulfides and phosphines to

corresponding oxides in reactions with 1,2,4-trioxolanes is known,11-13

it is not clear what

will be the products of reduction of 1,2,3-trioxolanes formed on rigid SWNT framework.

Formation of 1,2,4-trioxolanes by the Criegee mechanism,12

as it happens with small

organic molecules, seems unlikely.

With regard to SWNT thiolation, Lim et al.14

reported the conversion of terminal

carboxylic moieties of 200 nm long SWNT to –CH2SH by a series of reduction,

chlorination (SOCl2) and thiolation. Nakamura et al.15

reported a sidewall attachment of

sulfur-containing functionalities by irradiation of aliphatic disulfides in presence of

SWNT. The reaction was conducted with a low pressure mercury lamp (60W; > 200

nm) for 4 hours at r. t. Liu et al.16

reported an attachment of a thiol through an amide

linkage. Another example of an indirect attachment of amines includes the work of Peng

et al.17

who reported SWNT sidewall functionalization through the reaction with succinic

or glutaric acid acyl peroxides in o-dichlorobenzene at 80-90 C. Subsequent treatment

with thionyl chloride and amines yielded amide functionalized carbon nanotubes.

The most practical functionalization of SWNT with amines can be accomplished

through the reaction of fluorinated SWNT (SWNTF).18

Interestingly, the IR spectrum of

SWNTF reported by Stevens et al.18

is nearly identical to those demonstrated in this work

143

for ozonated SWNT (e.g. Figure 4 in Chapter 4). A typical procedure requires heating of

fluoronanotubes in excess of amine (used as a solvent) to 100-170 C for 4 hours.

Holzinger et al.19

reported preparation of alkoxycarbonylaziridino-SWNTs by

reaction of nanotubes with azidoformates at 160 C in 1,1,2,2-tetrachloroethane.

Derivatized tubes had a much better solubility in DMSO than pristine ones, thus allowing

for their separation.

A number of derivatized fullerenes with indirectly attached amines, i.e. nitrogens are

not attached to C60 framework, have been reported.20, 21

This topic is of less interest and

will not be discussed in this chapter.

144

5.2. Results and discussion

The degree of fluorescence recovery of ozonated SWNT with and without addition

of electron rich nucleophiles (amines, thiols and other)

Side wall functionalization of SWNT with ozonides has been demonstrated in this

work by various techniques: oxygen evolution, fluorescence bleaching, UV absorbance

bleaching, NIR absorbance bleaching, Raman G-band bleaching, IR absorbance increase

over a wide range of vibrational frequencies, typically below 4000 cm-1

, and an increased

percent of oxygen atoms on SWNT surface as determined by XPS.

0.1

0.3

0.5

0.7

0.9

1.1

1.3

1.5

1.7

1.9

2.1

2.3

2.5

10001100

1200

1300

1400

1500

1600

02

46

8

Flu

ore

scence Inte

nsity (

nW

/nm

)

Wavelength (nm)

Time (min)

0.20.40.6

0.81.01.21.3

1.51.7

1.9

Figure 1. Formation of SWNT ozonides and their reaction with 2-methoxyethylamine

monitored by NIR fluorescence at r. t. Ozone addition at ca. 30 sec led to a sharp decline

in fluorescence intensity. An addition of amine at ca. 1.5 min resulted in a fast recovery.

The 660 nm laser was used for excitation.

A full fluorescence recovery was observed after an addition of 3-methoxyethylamine

to ozonated SWNT. 3D kinetics at different wavelengths is shown in Figure 1. As

145

demonstrated later in this chapter, intercalation of amines into SDS quasi-liquid shell can

increase SWNT fluorescence intensity, though not much. Observed fluorescence increase

to the level above the preozonated one is thought to be due to: a) reaction of ozonated

SWNT with amines and b) amine intercalation. SWNT fluorescence is thought to be

strongly dependent on electron density of a -cloud formed by a conjugated system of

double bonds. An initial injection of 2 mL of O3/O2 gaseous mixture (ca. 1.5 v/v %

ozone) resulted in substantial fluorescence intensity depletion. A time period of 1 min

was allotted to ensure there is no more free ozone in aq. suspension. Amine (6 uL,

9 umol, 13 v/v % in water) addition to 1 mL of SWNT – SDS suspension led to a fast

fluorescence intensity rise.

Below is a graph demonstrating the influence of SWNT sidewall modification with

1,2,3-trioxolanes and a subsequent reaction with 2-methoxyethylamine on emmax

locations (Figure 2).

Several nanometers blue shift of the largest peak, near 953 nm, is likely due to tube

- cloud depletion (Figure 2B). Addition of amine red shifted emmax

of tube (8,3) near

953 nm. It is not clear if the red shift is caused by amine covalent attachment or by an

SDS shell intercalation or both. As shown later in this chapter, at lower ozone loads there

was no blue shift observed for ozonated SWNT. Thus small emmax

blue shifts were

observed only at extreme conditions, such as a high load of ozone. XPS and IR studies

confirmed that no chemical modification occurred upon mixing pristine SWNT with

amine at normal pressure and room temperature (see Appendix C for XPS spectra; see

curve in Figure 21 in Chapter 4 for an IR example). Thus, an intercalation of SDS

146

shell with amine molecules brought the final fluorescence intensity to a level higher than

that in the initial SWNT sample.

Wavelength (nm)

950 1050 1150 1250 1350

No

rma

lize

d F

luo

resce

nce

0.0

0.2

0.4

0.6

0.8

1.0

Flu

ore

sce

nce

In

ten

sity (

nW

/nm

)

0.0

0.5

1.0

1.5

2.0

2.5

A

B

951 954 957 960

0.8

0.9

1.0

SWNT(O3)n + amine

pristine SWNT

SWNT(O3)n

Figure 2. Influence of SWNT modification on em

max location. (A) Fluorescence spectra

as measured, (B) normalized fluorescence and zoom-in for tube (8,3) with

emmax

= 954 nm. Curves are: – pristine SWNT suspended in 1 wt. % aqueous SDS,

– SWNT after bubbling 2 mL of O3/O2 gaseous mixture (ca. 3 v/v % of ozone), – a

spectrum recorded 10 min after addition of 2-methoxyethylamine to ozonated SWNT.

The 660 nm laser was used for excitation.

147

Generally, SWNT fluorescence recovery after ozonation has a much slower rate than

in the case with electron rich nucleophiles added (Figure 3).

0.1

0.3

0.5

0.7

0.9

1.1

1.3

1.5

1.7

1.9

2.1

10001100

1200

1300

1400

1500

1600

05

10

15

20

Flu

ore

scence Inte

nsity (

nW

/nm

)

Wavelength (nm)

Time (min)

0.20.40.6

0.81.01.21.3

1.51.7

1.9

Figure 3. Formation and decomposition kinetics of SWNT ozonides at r. t. Ozone

addition at ca. 2 min led to a sharp decline in fluorescence intensity followed by a partial

recovery over time. The 660 nm laser was used for excitation. A small amount of ozone

was used for an oxidation (0.17 mL O3/O2 gaseous mixture, ca. 3 v/v % ozone).

Fluorescence of tubes emitting at shorter wavelengths was quenched less and

recovered faster when compared to tubes at longer wavelengths. There was a much

slower recovery of ozonated tubes when amine was not added (compare to Figure 1).

148

Reaction of ozonated SWNT with N,N,N’,N’-tetramethyl-p-phenylenediamine.

Ozonated SWNTs were reacted with N,N,N’,N’- tetramethyl-p-phenylenediamine

(TMPD), also known as Wurster reagent, to investigate the radical nature of the first step

of an electron transfer from amine to oxidized SWNT. Reaction between amines and

ozonated SWNT was demonstrated to proceed even two days after ozonation (Chapter 4),

indicating that oxidation with ozone converted SWNT into some form of a mild oxidizer.

TMPD (5.1 umol) was added to a very dilute suspension of SWNT in ethanol right after,

10, 20, 40 and 60 min after ozonation, reacted for 15 min and UV-Vis spectra obtained

(Figure 4).

Wavelength (nm)

400 500 600

Absorb

an

ce (

a.u

.)

0.5

2.0

3.5

565 615

2.5

3.5

no O3

0 min

10 m

20 m

40 m

60 m

Figure 4. UV-Visible spectra of reaction of N,N,N’,N’- tetramethyl-p-phenylenediamine

with ozonated SWNT. The same amount of Wurster reagent was added to all samples.

Additions were made right after, 10, 20, 40 and 60 min after ozonation. Non-ozonated

SWNT in ethanol, containing Wurster reagent, served as a reference sample (curve ).

UV spectra were recorded approximately 15 min after the addition of Wurster reagent.

149

All ozonated suspensions gained a deep purple color regardless of when Wurster

reagent was added. Reaction mixtures were collected after kinetics measurements and

photographed (Figure 5).

a b c d e f

Figure 5. Reaction of N,N,N’,N’- tetramethyl-p-phenylenediamine with ozonated SWNT.

Samples b-f gained a deep purple color due to a production of TMPD+ radicals. The

same amount of Wurster reagent was added to all samples. Black suspended flakes seen

on all pictures are SWNT bundles. (a) A reference sample, an ethanolic mixture of

pristine SWNT and TMPD, (b) TMPD added right after ozonation, (c) TMPD added 10

min after O3, (d) TMPD added 20 min after O3, (e) TMPD added 40 min after O3, (f)

TMPD added 60 min after O3.

Black flakes seen in Figure 5 are SWNT bundles. All samples were prepared from a

single stock suspension of SWNT in ethanol. Analogously, Wurster reagent was prepared

as a stock solution and aliquots were drawn for each experiment. SWNT flakes in ethanol

tended to precipitate over time and care was taken to prevent disturbance of samples

during spectra acquisition. As expected, addition of Wurster reagent right after and 10

min after ozonation gave higher degrees of its conversion into radical species TMPD+.

Addition of reagent 20, 40 and 60 min after ozonation gave essentially the same level of

consumption of TMPD (see Figure 4).

Absorbance change was monitored at 565 nm, near the local maximum (Figure 6).

Kinetics acquisition started approximately 10 sec after addition of the reagent. The delay

was used for mixing TMPD with SWNT flakes. TMPD+ production kinetics for samples

150

with reagent added right after, 40 and 60 min after ozonation were overlaid for

comparison and shown in Figure 6.

Time (min)

0 2 4 6 8 10 12 14 16

Absorb

an

ce (

a.u

.)

1

2

30 min

40 m

60 m

Figure 6. Absorbance rise at 565 nm with TMPD+ radical formation by oxidized SWNT.

Wurster reagent was added to ozonated SWNT right after, 40 and 60 min after SWNT

ozonation.

All three curves (, and in Figure 6) were fitted with 3 – parameter

exponential rise formula (F1) with excellent r2 values.

)1(0

bteayy (F1)

Table 1. Regression parameters for TMPD+ radical formation by oxidized SWNT.

Curve y0 a b, min-1

r2

0 min () 0.6793 2.9385 0.5077 0.9998

40 min () 0.3692 3.1732 0.4649 0.9996

60 min () 0.3043 3.1059 0.5196 0.9998

Close TMPD+ radical production rates and r

2 ~ 1 indicate that TMPD was in large

excess and reaction was first order with respect to the number of „reactive‟ centers on

151

SWNT. Obtained rate constants b corresponded to TMPDconsumed lifetime ~ 2 min at r. t.

in the presence of ozonated SWNT. The dependence of concentration of SWNT or

TMPD on TMPD+ radical production rate was not studied. A greater amount of TMPD

+

radicals produced with no “waiting” time is attributed to an additional reaction between

1,2,3-trioxolanes and TMPD. An appearance of a strong purple color is associated with

an electron transfer from TMPD to oxidizing species. Though pH affects the coloration of

TMPD solution, it does not impart such an intense purple color. Pictures of TMPD

coloration in EtOH/H2O (1:1 v/v) at different pH values are shown in Figure 7 below.

a b c d e f

Figure 7. TMPD coloration at different pH. An acidity of a TMPD solution in

water/ethanol mixture (1:1 v/v) was adjusted with HCl or NaOH. pH values were (a) 0.3,

(b) 1.8, (c) 3.6, (d) 8.7, (e) 13.1, (f) 13.9.

As seen in Figure 7, protonated TMPD at pH 3.6 (sample c) had a light crimson

color. No purple color could be obtained by adjusting pH. Scheme 1 below summarizes

observed reactions between TMPD and ozonated SWNT.

152

Scheme 1

NN

pristine SWNT

EtOH

no reaction

colorless

Wursterreagent

OO

O

epoxide

1,2,3-trioxolane

EtOH

O

NN

solvdeep purple

O

+

solv

or

otherproducts+

A reasonable question to ask: if solution turns purple, does TMPD+ radical get

attached to SWNT surface? Could it be that radical produced is dissolved by the media

and not attached to SWNT surface?

To answer on this question, XPS spectra were acquired on samples (Teflon® coated

with SWNT) dipped into a solution of TMPD in ethanol, followed by washing with

ethanol (2 times) and water. The first sample was subjected to a stream of O3/O2 gaseous

mixture (ca. 1.5 v/v % ozone) for 1 min before dipping it into TMPD solution (2.8 mg in

500 uL of ethanol). A washing step was designed to remove traces of unreacted TMPD.

As seen in Figure 8, ozonated SWNT film got a small number of TMPD molecules

attached. A delocalization of an unpaired electron in TMPD+

and solvation with ethanol

molecules are thought to be the major reasons for turning solution color to a deep purple.

It is reasonable to expect a greater percent of radicals, generated from aliphatic amines,

getting attached to SWNT surface. Aliphatic amines do not have electron delocalization

seen in TMPD and are thought to be more reactive.

153

Binding Energy (eV)

02004006008001000

Co

un

ts /

se

c

x1

04

1

2

3

4

Co

un

ts /

se

c

x1

04

1

2

3

4

5

- O

KLL

- O

1s

- N

1s

- F

e 2

p3/2

- O

1s

- F

e 2

p3/2

- C

1s

SWNT + Wurster reagent

Ozonated SWNT + Wurster reagent

- C

1s

Figure 8. XPS spectra of SWNT reacted with an ethanolic solution of TMPD. (Top) a

product from reaction with pristine SWNT, a reference, (bottom) a product from reaction

with ozonated SWNT.

Fluorescence studies of reaction between dithiothreitol (DTT) and ozonated SWNT

Reaction between a thiol and ozonated SWNT has been studied by fluorescence

(Figure 9).

SH

OH

OH

HS

dithiothreitol

154

Time (sec)

0 200 400 600

Norm

aliz

ed F

luore

scence

0.0

0.3

0.6

0.9

954 nm

1026 nm

1123 nm

1250 nm

Figure 9. An addition of dithiothreitol to ozonated SWNT – SDS suspension. The sample

was excited with 661 nm laser source. Fluorescence of tube (8,3) at 954 nm recovered at

once to a 90% level of its initial intensity. Other peaks returned to lower levels. The

timing of DTT addition is emphasized with an arrow.

Fluorescence decreased at least 200 times upon reaction with ozone at all four

emission wavelengths. Partial ozonide decay brought fluorescence intensities up to 1 % at

954 nm and 2 % at 1250 nm of initial levels after which dithiothreitol was injected. An

immediate fluorescence recovery followed thiol addition.

Different tubes recovered to different levels. As observed in other experiments, tubes

with emmax

at shorter wavelengths recovered to higher percent levels. Dithiothreitol was

added 2.5 minutes after injection of ozone/oxygen gaseous mixture. UV studies

demonstrated (Chapter 2) that all freely floating ozone should be gone within less than 2

min at r. t. A fluorescence rise was another indication of the absence of ozone. It is

believed there was no more unreacted ozone in aqueous SWNT suspension at the time of

a dithiol injection.

155

Alkyl thiols, analogously to aliphatic amines, are thought to serve as electron donors

during the first step of interaction with ozonated SWNT. An increase of electron density

on SWNT is supported by the evidence of increased NIR absorbance of SWNT at

frequencies above 7000 cm-1

upon reaction with amines (IR studies, Chapter 4). A greater

number of electrons in a -cloud causes SWNT to absorb stronger in the NIR.

Noticeably, recovery of fluorescence after thiol addition was permanent and showed

no signs of equilibration after chemical addition (compare to Figure 10). The jump was

abrupt and did not change over time. There are three possible explanations for this

phenomenon. Ozonide decay by thiol is the most obvious one. The second one is a

covalent attachment of thiol to SWNT. A third possibility is that hydrophobic interaction

of SWNT with thiol led to an accumulation of thiols on the surface of tubes, in the same

manner as with 2-methoxyethylamine. All explanations are likely to be valid, though the

third one would not result in such a dramatic intensity change.

Injections of water-miscible 2-methoxyethylamine into SWNT-SDS aqueous

suspensions were shown to yield fluorescence increase (Figure 10). It is reasonable to

assume that thiols would result in a similar fluorescence intensity increase.

156

Time (min)0 10 20

Norm

aliz

ed F

luore

scence

1.0

1.2

1.4

954 nm

1026 nm

1123 nm

1250 nm

Figure 10. Fluorescence intensity change of pristine SWNT at four distinct wavelengths

after several additions of 2-methoxyethylamine. Injections are emphasized with arrows.

The 661 nm laser was used for excitation.

Larger diameter tubes, contributing to emission at 1250 nm, had no fluorescence

increase with second through fourth additions of amine. Spikes right after additions were

due to misbalance and are not meaningful. In contrast, thinner tubes, emitting at 954 nm,

had fluorescence increase after all four injections.

SDS shells around nanotubes are thought to be in quasi liquid state. The above

experiment demonstrated that: a) an increased number of amine molecules at the surface

of SWNT causes stronger emission and b) SDS shell around thinner tubes ( emmax

954

nm) could not be saturated with amine molecules as it was in the case with thicker ones

( emmax

1250 nm). It is concluded that SDS shells around thicker tubes are more rigid,

more hydrophobic and could be harder to penetrate for a heavily solvated reagent like 2-

methoxyethylamine.

157

Spikes right after chemical additions are thought to be due to an increased

concentration of amine at the surface of SWNT. System equilibration results in expulsion

of amine molecules from quasi-liquid shells of sodium dodecyl sulfate. Lack of an

“equilibration” period in Figure 9 suggests that thiol could get covalently attached to the

SWNT surface.

Wavelength (nm)

950 1050 1150 1250

Flu

uo

resce

nce

In

ten

sity (

nW

/nm

)

0.00

0.01

0.02

0.03

0.04

0.05

951 955 959N

orm

aliz

ed

0.7

0.8

0.9

1.0Pristine SWNT

1st

2nd

3rd

4th

amine addition

Figure 11. Dilution of SDS shell with 2-methoxyethylamine and its influence on SWNT

emission. The 660 nm laser was used for excitation. Equal amounts of amine were

injected each time. Main plot: fluorescence spectra as measured, zoom-in: normalized

fluorescence near 954 nm peak. Curves are: – pristine SWNT suspended in 1 wt. %

aqueous SDS; Symbols , , , and correspond to spectra after 1st, 2

nd, 3

rd and 4

th

injections of amine.

As seen in Figure 11, each subsequent injection of amine (4 umol, 5 uL,

c = 6.7 v/v % in 1 wt. % aq. SDS) resulted in fluorescence rise at peak maxima near 954,

1026 and 1123 nm. No change was observed for 1250 nm. Rise was greater with shorter

wavelengths, indicating higher degree of penetration of quasi-liquid SDS shell. No peak

shifts were observed. This is in contrast to small shift at 954 nm observed for reaction of

158

amines with ozonated SWNT. In general, peak shifts of SWNT should be treated with

caution, since different surfactants (SDS, SDBS, CTAB, Brij 700 and other) were shown

to affect peak maxima location.22

Dithiothreitol solubility in water is significantly lower than that of SDS;

15.4 g/L for dithiol vs. 200 g/L for sulfate. The argument could be made that DTT

preferentially intercalated in between SDS chains on the surface of SWNT. It is worth

noting that such high solubility of SDS is achieved through formation of micelles. In this

work 1 wt. % SDS (10mM) aqueous suspensions were used for all fluorescence and

liquid Raman measurements. Critical micelle concentration for SDS is about 8mM in

water.23

Encapsulation of DTT molecules within SDS micelles is yet another possibility.

Due to a great number of variables in the system, fluorescence intensity change alone is

not a proof of covalent attachment. An independent analytical method such as X-ray

Photoelectron Spectroscopy (XPS) is needed to establish elemental composition of atoms

on the surface of SWNT.

In conclusion, the fluorescence technique was useful in demonstrating how different

chemicals affect ozonide stability, but it provided no actual proof of covalent attachment

of such reagents to SWNT sidewall. Amines and thiols were shown to decompose

ozonides on the SWNT surface, thus affecting SWNT fluorescence intensity.

Fluorescence studies of reaction between 2-methoxyethylamine and ozonated SWNT

at different times after ozonation

The influence of amines on fluorescence of ozonated SWNT at different time

intervals after oxidation was investigated.

159

Time (min)

0 10 20 30

No

rma

lize

d F

luo

resce

nce

0.0

0.3

0.6

0.9

Imax

L1L2

a bc

d

e

em = 1250 nm

No

rma

lize

d F

luo

resce

nce

0.0

0.3

0.6

0.9

46 umol

35 umol

23 umol

4 umol

4 umol

4 umol

em = 1026 nmI

max

L1L2

a b

c

d

e

Figure 12. Addition of 2-methoxyethylamine to ozonated SWNT – SDS suspension.

Each curve represents a separate experiment and is marked with a symbol. All curves are

overlaid for comparison purposes. All reactions are run at 23.1 C. Symbol legends are

the same for the top and the bottom graphs. Imax is a SWNT fluorescence level before

bubbling ozone. Letters: (a) a fluorescence drop with an injection of 1 mL of O3/O2

gaseous mixture (ca. 3 v/v % ozone), (b-d) addition of 4 umol of amine at different times,

(e) an addition of 23, 35 and 46 umol of amine at approx. the same time. Top: emission at

1026 nm. Bottom: emission at 1250 nm.

160

A determination of ozonide decomposition rates made it possible to distinguish

between reactions of amines with SWNT epoxides and SWNT ozonides. Based on

oxygen evolution studies, all ozonides are thought to be decomposed within 20 minutes at

r. t. (Chapter 2, Figure 2). Normalized fluorescence spectra of oxidized SWNT treated

with different quantities of 2-methoxyethylamine and at different times are shown in

Figure 12.

2-Methoxyethylamine was considered a good choice for such studies, since it is

miscible with water and is not expected to have hydrophobic interactions, or adhesion to

SWNT surface. Two heteroatoms on a molecule greatly facilitate its solvation with water

dipoles.

Levels L1 and L2 in Figure 12 show schematically the dependence of the

fluorescence recovery on the amount of amine added. Larger amounts of amine resulted

in a slightly higher level of fluorescence (compare L1 to L2 in Figure 12) and faster

reaction rates as judged by curve slopes right after amine addition. Injections of 23 to 46

umol of amine gave steeper slopes than those after 4 umol. Addition of 4 umol of a

nucleophile at different times resulted in the same recovery levels (schematic line L1).

This result means that the amount of amine was in a large excess to the number of 1,2,3-

trioxolanes formed on the surface of SWNT after ozone bubbling.

In accordance with the Le Chatelier principle, an increased number of amine

molecules in the suspension resulted in a higher intercalation level of SDS shell with

amines, leading to steeper fluorescence recovery slopes (Figure 12).

Fluorescence jumped upward after addition of amine 24.3 min after ozonation (d of

Figure 12). It is thought that jump is associated with a) covalent attachment of amines to

161

ozonated SWNT and b) intercalation of SDS shell with freely floating amines. As seen in

Figure 10, addition of the same amount of amine to non oxidized SWNT gave

fluorescence increase by only 0.15 (at 1026 nm) and 0.05 (at 1250 nm). Separate X-ray

Photoelectron Spectroscopy measurements further confirmed that SWNT surface is

modified with nitrogen containing groups (Chapter 7).

Molar ratio of the amount of ozonated double bonds to that of added amine

A change in absorbance at 260 nm is directly proportional to the number of 1,2,3-

trioxolanes formed. Epoxides and 1,2,3 trioxolanes are not expected to absorb at such a

short wavelength. Whether SWNT absorbance changes linearly with decrease in the

number of double bonds is not known. Let‟s assume that at low levels of SWNT

ozonation it is linear. Bubbling ~ 1 mL of O3/O2 gaseous mixture (ca. 3 v/v % ozone) was

used for a low degree ozonation. Resting on the above assumptions, the following

calculations were made. In a separate experiment SWNT absorbance was bleached less

than 4 % at 260 nm after bubbling 1.5 mL of O3/O2 gaseous mixture through suspension

(Chapter 2, Figure 3). An approximate number of double bonds reacted with ozone can

be estimated from this result. A SWNT-SDS suspension had a tube concentration of 5.75

ug of carbon/mL. Four per cent of this amount corresponds to 0.23 ug/mL, or 0.019

umol/mL. Dividing this number by two yields the amount of double bonds reacted with

ozone, 0.01 umol/mL. This is ca. 400 times less than the lowest amount of amine injected

(see Figure 12).

162

Interestingly, a reaction between amine and ozonated SWNT can occur even in the

absence of 1,2,3-trioxolanes. Amines were shown to react with SWNT two days after

ozonation (Chapter 4, Figure 19).

Possible reaction mechanisms between electron rich nucleophiles and ozonated

SWNT are summarized in Scheme 2.

163

Scheme 2

OO

OO

OO

OO O

Nu

epoxide

O

oxidoannulene

O O

Nu

1,2,3-trioxolane

Nu = AlkSH, ArSH, R3N, RNH2, R2NH, PR3, guanidine, etc.

Nu

Nu

Nu

SWNT=

OO

epoxide

RNH2

RNH2 (xs)

RNH2

Route A

Route B

RNH3

RNH

RNH O

NH

R

RNH3

RNH2

O

NH

R

H

OO

ONu

NuOno attachment ofNu to sidewall

Route C

Reactions between amines and ozonated fullerene C60Ox

Reactions observed between ozonated SWNT and amines prompted a series of

r. t. experiments with ozonated C60Ox. Thus, a purple solution of C60 in toluene was

ozonated briefly for 5 sec with a stream of O3/O2 gaseous mixture (ca. 3 v/v % ozone),

purged thoroughly with argon gas and amine was added. Purging with inert gas was done

to ensure absence of ozone in toluene. An orange color of ozonated fullerene immediately

changed to a dark brown, indicating the reaction proceeded fast at r. t. Primary and

164

tertiary amines were used to demonstrate the electron transfer nature of this reaction.

Triethylamine, in the same manner as n-butylamine, reacted with ozonated fullerenes to

produce a highly polar brownish product, which oiled out of toluene over time. Results of

these reactions are summarized in Scheme 3.

Scheme 3

C60

BuNH2

Et3N

BuNH2

Et3N

light purpleorange

no reaction

no reactiontoluene

toluene

NEt3

toluene

dark brown

NH2Bu

dark brown

C60 On + m

C60

OO

O

O

n

m

C60 On + m

k

k

r.t.

r.t.

r.t.

otherproducts+

otherproducts+

Reaction mixtures were photographed and shown in Figure 13.

Figure 13. A solution color comparison of C60 and derived products. Solutions are in

toluene. All reactions were run for 5 min at r. t. Labels are: (A) fullerene C60, (B)

fullerene ozonated for several seconds C60(O3)x and (C) reaction product of ozonated

fullerene and n-butyl amine C60Ox(NH2Bu)y

Reaction of ozonated fullerene with triethylamine produced the same dark brown

color as in reaction with n-butyl amine, indicating that electron transfer from nitrogen to

C60(O3)x is likely to be the first step. An analysis of product from reaction with

triethylamine by 1H NMR and IR indicated the presence of triethyl ammonium salt.

Characteristic shifts in 1H NMR (MeOH-d4) were at 3.18 and 1.28 ppm for ethyl groups,

165

as would be expected for an ammonium salt. An authentic sample of

tetraethylammonium chloride in

MeOH-d4 had chemical shifts 3.35 and 1.33 ppm, which are very close to the ones in

the fullerene derivative.

To demonstrate the ability of amines to react with ozonated fullerenes even in the

absence of 1,2,3-trioxolanes, n-butyl amine was added to ozonated C60 twenty minutes

after oxidation. The same color change from orange to dark brown occurred (Figure 14).

This implied that the color change is likely due to a rearrangement of conjugated double

bonds on fullerene.

Figure 14. A solution color comparison of C60 and derived products. Solutions A and B

are in toluene; sample C was dried under vacuum and redissolved in MeOH. All reactions

were run for 5 min at r. t. Labels are: (A) fullerene C60, (B) mixture of C60 and n-BuNH2

and (C) amine added to C60Ox twenty minutes after ozonation

Cross-linking of aminated fullerenes has been reported in literature, which could be

the case here as well.2 IR spectrum and peak assignments for structure C60Ox(NH2Bu)y

are provided below.

166

Wavenumber (cm-1

)

100020003000

Absorb

ance (

a.u

.)

0.1

0.2

0.3

C60

Ox(NH

zBu)

y

a

b

c

d

e

f

g

h i

j

k

Fugure 15. IR spectrum of C60Ox(NH2Bu)y

Table 2. IR peak assignments for structure C60Ox(NH2Bu)y:

Symbol Position Assignment

a 3025 N-H stretch in R2NH2+; may be combined with peak from C60Ox

b 2958 b-d CH3 and CH2 stretches

c 2930 ditto

d 2872 ditto

e 1711 C=O from C60Ox

f 1641 C=C

g 1586 N-H bend

h 1465 CH3 as bend & CH2 sym bend (scissoring)

i 1378 CH3 sym bend

j 1081 C-N stretch

k 736 N-H wagging

167

IR spectrum of C60Ox(NEt3)y and peak assignments are provided below.

Wavenumber (cm-1

)

100020003000

Absorb

ance (

a.u

.)

0.05

0.10

0.15

0.20C

60O

x(NEt

3)y · n MeOH-d4

a

b

c

d

e

fg

h i

j

k

Figure 16. IR spectrum C60Ox(NEt3)y

Table 3. IR peak assignments for structure C60Ox(NEt3)y:

Symbol Position Assignment

a 2000-

3500

unassigned; it could be from C60Ox or C60Ox(NEt3)y(MeOH)n

b 2984 unassigned

c 2955 c-d CH3 and CH2 stretches

d 2917 ditto

e 2850 ditto

f 1732 C=O from C60Ox

g 1614 unassigned

h 1455 CH3 as bend & CH2 sym bend (scissoring)

i 1390 CH3 sym bend

j 1096 C-N stretch

k 804 unassigned

IR spectra had multiple bands coming from amine residues. Band assignments are

listed in the corresponding tables under the IR spectra. Amines are believed to be

168

covalently attached to oxidized C60 species. Samples changed color from purple to deep

orange during ozonation, and then from orange to dark brown upon amine addition.

Amines chosen for reaction had low boiling points and were expected to be completely

evaporated during vacuum drying (10-3

Torr). Products slowly aggregated and oiled out

of toluene as dark brown – black liquid drops. It is likely that formed species were

charged and solvated by the excess of freely floating amine molecules. Toluene was

evaporated and product redissolved in methanol-d4, forming a dark brown solution. C60 is

extremely insoluble in solvents like methanol. Obtained products C60Ox(NH2Bu)y and

C60Ox(NEt3)y had excellent solubility in methanol. Ability to redissolve C60Ox(amine)y in

methanol is an indication of a highly polar product. Change in color from orange to dark

brown is thought to be due to electron transfer from amine to C60Ox species. Exact

structures of products were not established. Repetition of experiment with longer

ozonation period resulted in a better solubility of C60Ox(amine)y in methanol, indicating

an increased number of amine residues on a single C60 skeleton.

Benzene oxidation with ozone is known.24

Some IR peaks may be coming from non-

volatile species formed by toluene oxidation. The nature of the very broad peak at

2000-3500 cm-1

in reaction with triethylamine is not clear. It is thought that methanol-d4

could get involved in protonation of the reactive species.

13C NMR was not performed on a crude mixture. MALDI-TOF characterization of

crude products gave complex spectra for both triethyl and n-butyl amines. It is believed

that there were multiple amine residues on C60Ox. Multiple charges, expected for

C60Ox(NEt3)y species, further complicated MS peaks‟ assignment. Oxidation of several

169

double bonds on C60 is well documented in the literature.25

The number of amines on

each C60 skeleton was not established.

Though reaction between amines and C60O3 was clearly manifested by changes in

color, solubility in methanol, characteristic peaks in 1H NMR and IR, thorough study is

needed for unambiguous structure assignments. Particularly, separation of products on

HPLC and MS and NMR characterization of each product is needed.

5.3. Conclusions

Ozonated SWNT suspended in aqueous SDS solution were shown to react with

electron rich nucleophiles, like amines and thiols. While reaction between SWNT and

amines is known to proceed with a very slow rate at normal conditions, the technique

presented here allowed SWNT side wall modification within minutes. The key to

decreasing activation energy of the process was oxidation of SWNT with ozone, resulting

in conversion of SWNT to a mild oxidizer, capable of abstracting electrons from thio and

amino moieties. Formed radicals are thought to be of a high energy and reacted with

SWNT in a matter of minutes. Reaction of Wurster reagent with ozonated SWNT was

monitored by UV and the rate of TMPD+ radical production found to be ca. 0.5 min

-1 for

a very dilute ethanolic suspension of SWNT. Reactions of amines with ozonated

fullerene C60 demonstrated similar behavior. n-Butyl and triethyl amines reacted with

C60On within seconds with corresponding color change from deep orange to deep brown.

Formed products were found to be well soluble in methanol, thus C60 solubility in

methanol was greatly increased.

170

5.4. Experimental Part

Reaction of ozonated SWNT with N,N,N’,N’-tetramethyl-p-phenylenediamine

Ethanol was used as a solvent (200 proof, Aaper Alcohol and Chemical Co.). A

preliminary sonicated for 40 min concentrated stock suspension of carbon nanotubes in

ethanol (Rice University, HipCo raw tubes, unpurified, batch 162.8) was used for

reactions. All sonication was done with a bath sonicator (Fisher FS60). The Ozone

Services Inc. ozonator (model GE60/FM 500) with a power level set to the maximum and

oxygen gas flow to 1/16 L/min was used for ozone production. Ozonator was warmed up

for 7 min per manufacturer instruction to reach its maximum output. UV measurements

were performed on Cary 4E spectrophotometer with 1cm quartz cuvette. Ethanol

spectrum served as a baseline. Wurster reagent was prepared by dissolution of N,N,N’,N’-

tetramethyl-p-phenylenediamine (11.5 mg) in ethanol (2 mL).

A very dilute suspension of SWNT (< 0.1 mg) was prepared in 20 mL of ethanol and

bath sonicated for 30 sec. Finely dispersed suspension was shaken to “crash out” SWNT

back to carbon nanotube flakes, and then distributed 2.5 mL aliquots into separate glass

vials. Suspension was swirled before each aliquot was taken out of a stock solution.

Wurster reagent (150 uL, 5.2 umol) was added to the first sample. Second sample was

ozonated by bubbling O3/O2 gaseous mixture (ca. 3 v/v % ozone) through the suspension

for 1 min at r. t. Air was blown off the top of the vial with a stream of argon for about 20

sec, then argon was bubbled through ozonated solution for about 30 sec with speed

approx. 1/16 L/min, and blowing air off of the top of the vial with argon was repeated as

before. Wurster reagent (150 uL) was added to ozonated sample. Other samples were

171

analogously ozonated and purged with argon, but waited for 10, 40 and 60 min before

addition of Wurster reagent.

Reaction of ozonated SWNT with dithiothreitol

A suspension of SWNT (HipCo, batch 161.1, Rice University) in 1 wt. % aq. SDS

was prepared by a standard procedure. Suspension pH was adjusted with NaOH to pH 8.

SWNT concentration was estimated at 5.75 mg/L. The O3/O2 gaseous mixture (1 mL, ca.

3 v/v % ozone) was bubbled through 1 mL of SWNT – SDS suspension, waited for 2.5

min and added dithiothreitol (5 uL, a diluted solution in water). Reaction was monitored

at 22.2 C by fluorescence. The 661 nm laser was used for excitation.

Influence of amine on pristine SWNT fluorescence intensity

A suspension of SWNT (HipCo, batch 161.1, Rice University) in 1 wt. % aq. SDS

was prepared by a standard procedure. Suspension pH was adjusted with NaOH to pH 8.

SWNT concentration was estimated at 5.75 mg/L. Multiple additions of 2-methoxyethyl

amine (4 umol, 5 uL, 6.7 v/v % in 1 wt. % aq. SDS) to SWNT – SDS suspension at 23.1

C were monitored by fluorescence. The 661 nm laser was used for excitation.

Reaction of 2-methoxyethylamine with ozonated SWNT

SWNT-SDS suspension was prepared by a standard procedure from HipCo 161.1

raw tubes (see Chapter 2 for details). pH was adjusted to 8 with 0.1N aq. NaOH,

suspension shaken and bath sonicated for 1 min. Solutions of 2-methoxyethylamine

(0.77M and 2.3M) in 1 wt. % aq. SDS were used for amine additions. All reactions were

run at 23.1 C in a thermostated quartz cuvette with 1 mL of SWNT – SDS suspension.

Spectra were recorded on NanoSpectralyzer NS1 with 661 nm excitation source. Points

172

were acquired every 10 sec with 500 ms excitation pulses. The O3/O2 gaseous mixture

(1 mL, ca. 3 v/v % ozone) was bubbled through SWNT suspension. In three separate

experiments amine (4 umol) was added at times 1.7, 9.5 and 24.3 min. In another three

different experiments 23, 35 and 46 umol of amine were added at approx. 5.5 min.

Resulted emission curves were normalized and overlaid for comparison.

Reaction of amines with ozonated C60

Purple-colored aliquots of C60 in toluene (0.25 mL; conc. not available) were

bubbled for 5 sec with O3/O2 gaseous mixture (ca. 3 v/v % ozone) followed by thorough

purging with nitrogen and addition of amine (5 uL). Waited for 5 min, evaporated toluene

and dried on vacuum for 5 min. Residue was redissolved in MeOH-d4 and a few drops

were added to BaF2 window. Plate was dried in vacuum for 15 min and IR measured.

Chamber was purged with nitrogen to remove carbon dioxide and moisture. The

remaining portion of solution was used for 1H NMR and MS. MeOH-d4 served as a

reference. 1H NMR (MeOH-d4), C60Ox(NR3)y 3.18 (q, J = 7.2 Hz) and 1.28 (dt, J1 =

7.7 Hz; J1 = 3.3 Hz); 1H NMR of n-BuNH2 was poorly resolved and data is not provided.

Reaction of 2-methoxyethylamine with ozonated SWNT (3D kinetics experiment)

SWNT – 1 wt. % aq. S DS suspension was prepared by a standard procedure from

HipCo 153.3 raw tubes (details in experimental part of Chapter 2). pH was adjusted to 8

with 0.1N aq. NaOH, sample shaken and bath sonicated for 1 min. The O3/O2 gaseous

mixture (2 mL, ca. 3 v/v % ozone) was injected into 1.0 mL of SWNT – SDS suspension

(c ~ 5.75 mg/L; 0.48 umol carbon atoms/mL), waited for 1 min and amine was injected

173

(6 uL, 9 umol, 13 v/v % in water). Kinetics was monitored by fluorescence. The 660 nm

laser was used for excitation.

SWNT ozonation (3D kinetics experiment)

SWNT – 1 wt. % aq. S DS suspension was prepared by a standard procedure from

HipCo 153.3 raw tubes (details in experimental part of Chapter 2). pH was adjusted to 8

with 0.1N aq. NaOH, sample shaken and bath sonicated for 1 min. The O3/O2 gaseous

mixture (0.17 mL, ca. 3 v/v % ozone) was injected into 1.0 mL of SWNT – SDS

suspension (c ~ 5.75 mg /L; 0.48 umol carbon atoms/mL). Kinetics was monitored by

fluorescence. The 660 nm laser was used for excitation.

Procedure for Wurster reagent (for XPS measurements)

A Teflon plate covered with SWNT film was ozonated for 1 min with a stream of

O3/O2 gaseous mixture (ca. 3 v/v % ozone). Plate was dipped into the solution of TMPD

(2.7 mg in 500 uL EtOH) for 15 min, and then washed two times in EtOH (1 mL) and

one time in water (1 mL) for 3 min each. The sample was dried in vacuo before XPS

measurements.

The pH adjustments

Millipore Milli-Q water (18.2 MOhm/cm at 25 C), ethanol, 0.1 and 1N aq. NaOH

(Fisher Scientific), 0.1 and 1N aq. HCl (Fisher Scientific) were used to prepare TMPD

solutions with different pH. Measurements were performed with a digital pH meter

(Fisher Scientific, Dual Channel pH meter AR 50).

174

5.5. References and Notes

1. Hu, X. Q.; Jiang, Z. P.; Jia, Z. S.; Huang, S. H.; Yang, X. B.; Li, Y. L.; Gan, L.

B.; Zhang, S. W.; Zhu, D. B., Amination of 60 fullerene by ammonia and by

primary and secondary aliphatic amines - Preparation of amino C60 fullerene

peroxides. Chemistry-a European Journal 2007, 13, (4), 1129-1141.

2. Manolova, N.; Rashkov, I.; Beguin, F.; Vandamme, H., Amphiphilic Derivatives

of Fullerenes Formed by Polymer Modification. Journal of the Chemical Society-

Chemical Communications 1993, (23), 1725-1727.

3. Nakamura, E.; Isobe, H., Functionalized fullerenes in water. The first 10 years of

their chemistry, biology, and nanoscience. Accounts of Chemical Research 2003,

36, (11), 807-815.

4. Isobe, H.; Tanaka, T.; Nakanishi, W.; Lemiegre, L.; Nakamura, E., Regioselective

oxygenative tetraamination of C60 fullerene. Fullerene-mediated reduction of

molecular oxygen by amine via ground state single electron transfer in dimethyl

sulfoxide. Journal of Organic Chemistry 2005, 70, (12), 4826-4832.

5. Isobe, H.; Tomita, N.; Jinno, S.; Okayama, H.; Nakamura, E., Synthesis and

transfection capability of multi-functionalized fullerene polyamine. Chemistry

Letters 2001, (12), 1214-1215.

6. Isobe, H.; Ohbayashi, A.; Sawamura, M.; Nakamura, E., A cage with fullerene

end caps. Journal of the American Chemical Society 2000, 122, (11), 2669-2670.

7. Isobe, H.; Tomita, N.; Nakamura, E., One-step multiple addition of amine to C60

fullerene. - Synthesis of tetra(amino)fullerene epoxide under photochemical

aerobic conditions. Organic Letters 2000, 2, (23), 3663-3665.

8. Schick, G.; Kampe, K. D.; Hirsch, A., Reaction of C60 Fullerene with Morpholine

and Piperidine - Preferred 1,4-Additions and Fullerene Dimer Formation. Journal

of the Chemical Society-Chemical Communications 1995, (19), 2023-2024.

9. Isobe, H.; Nakanishi, W.; Tomita, N.; Jinno, S.; Okayama, H.; Nakamura, E.,

Gene delivery by aminofullerenes: Structural requirements for efficient

transfection. Chemistry-an Asian Journal 2006, 1, (1-2), 167-175.

10. Banerjee, S.; Wong, S. S., Rational sidewall functionalization and purification of

single-walled carbon nanotubes by solution-phase ozonolysis. Journal of Physical

Chemistry B 2002, 106, (47), 12144-12151.

11. Kuwabara, H.; Ushigoe, Y.; Nojima, M., Synthesis and reaction of cyano-

substituted 1,2,4-trioxolanes. Journal of the Chemical Society-Perkin

Transactions 1 1996, (9), 871-874.

175

12. Bailey, P. S., Ozonation in Organic Chemistry. Academic Press: New York,

1978; Vol. 1, p. 272.

13. Bailey, P. S., Ozonation in Organic Chemistry. Academic Press: New York,

1982; Vol. 2, p. 497.

14. Lim, J. K.; Yun, W. S.; Yoon, M. H.; Lee, S. K.; Kim, C. H.; Kim, K.; Kim, S.

K., Selective thiolation of single-walled carbon nanotubes. Synthetic Metals 2003,

139, (2), 521-527.

15. Nakamura, T.; Ohana, T.; Ishihara, M.; Tanaka, A.; Koga, Y., Sidewall

modification of single-walled carbon nanotubes with sulfur-containing

functionalities and gold nanoparticle attachment. Chemistry Letters 2006, 35, (7),

742-743.

16. Liu, Z. F.; Shen, Z. Y.; Zhu, T.; Hou, S. F.; Ying, L. Z.; Shi, Z. J.; Gu, Z. N.,

Organizing single-walled carbon nanotubes on gold using a wet chemical self-

assembling technique. Langmuir 2000, 16, (8), 3569-3573.

17. Peng, H. Q.; Alemany, L. B.; Margrave, J. L.; Khabashesku, V. N., Sidewall

carboxylic acid functionalization of single-walled carbon nanotubes. Journal of

the American Chemical Society 2003, 125, (49), 15174-15182.

18. Stevens, J. L.; Huang, A. Y.; Peng, H. Q.; Chiang, L. W.; Khabashesku, V. N.;

Margrave, J. L., Sidewall amino-functionalization of single-walled carbon

nanotubes through fluorination and subsequent reactions with terminal diamines.

Nano Letters 2003, 3, (3), 331-336.

19. Holzinger, M.; Vostrowsky, O.; Hirsch, A.; Hennrich, F.; Kappes, M.; Weiss, R.;

Jellen, F., Sidewall functionalization of carbon nanotubes. Angewandte Chemie-

International Edition 2001, 40, (21), 4002-4005.

20. Yang, J. H.; Wang, K.; Driver, J.; Barron, A. R., The use of fullerene substituted

phenylalanine amino acid as a passport for peptides through cell membranes.

Organic & Biomolecular Chemistry 2007, 5, (2), 260-266.

21. Rouse, J. G.; Yang, J. Z.; Barron, A. R.; Monteiro-Riviere, N. A., Fullerene-based

amino acid nanoparticle interactions with human epidermal keratinocytes.

Toxicology in Vitro 2006, 20, (8), 1313-1320.

22. Moore, V. C.; Strano, M. S.; Haroz, E. H.; Hauge, R. H.; Smalley, R. E.; Schmidt,

J.; Talmon, Y., Individually suspended single-walled carbon nanotubes in various

surfactants. Nano Letters 2003, 3, (10), 1379-1382.

176

23. Esposito, C.; Colicchio, P.; Facchiano, A.; Ragone, R., Effect of a weak

electrolyte on the critical micellar concentration of sodium dodecyl sulfate.

Journal of Colloid and Interface Science 1998, 200, (2), 310-312.

24. Nakagawa, T. W.; Andrews, L. J.; Keefer, R. M., The Kinetics of Ozonization of

Polyalkylbenzenes. Journal of the American Chemical Society 1960, 82, (2), 269-

276.

25. Manning, T. J.; Olsen, K.; Hardin, L.; Purcell, J.; Ayers, T. M.; Duncan, M. A.;

Phillips, D., Extensive ozonation of C60: Degradation or polymerization? Ozone-

Science & Engineering 2006, 28, (3), 177-180.

177

Chapter 6

Trapping reactive centers on SWNTOn with electron rich nucleophiles

(amines, thiols)

178

6.1. Introduction

This chapter examines the possibility of trapping reactive centers on SWNT after it

was subjected to ozonation. Successfully trapped oxidative areas on “SWNT” should

loose their ability to withdraw electrons from N,N,N’,N’-tetramethyl-p-

phenylenediamine, thus preventing purple coloration. Such trapping methodology for

ozonated SWNT has not been described in the literature.

6.2. Results and discussion

Reactive centers on ozonated carbon nanotubes (SWNTOn) were efficiently trapped

with reagents of choice, thus preventing N,N,N’,N’-tetramethyl-p-phenylenediamine

(TMPD) from getting oxidized. TMPD undergoes a characteristic color change from a

pale tan to a deep purple due to radical cation generation when reacted with oxidizers. A

number of reagents were tested in reaction with ozonated SWNT: guanidine,

dithiothreitol, 2-methoxyethylamine and sodium salts of arginine, -aminobutyric acid,

lysine and cysteine. All reagents were expected to give up an electron in reaction with

ozonated SWNT. Reaction was run against a reference sample, which was prepared by

exactly the same procedure but did not include an ozonation step.

For reaction to proceed, an amino group of a reagent must be in its free form. As

seen in Figure 1, having charged amino groups, guanidine·HCl and

-aminobutyric acid were not able to quench reactive centers on SWNTOn. Subsequent

additions of TMPD turned suspensions to purple. Without NaOH added, lysine was found

to quench SWNTOn reactive centers inefficiently. Addition of TMPD yielded a faint

purple color.

179

The following is a brief summary of the procedure. A diluted stock suspension of

SWNT (less than 2.0 mg/70 mL) was prepared by bath sonication and aliquots were

drawn from each sample. Stock solutions were prepared for each nucleophilic reagent to

ensure that an equal amount of reagent was used for both reaction and reference samples.

SWNT suspensions were ozonated for 1 min, thoroughly purged with argon, then a

reagent of interest was added and Wurster reagent added 15 min after. The results of

these reactions are summarized in Figure 1.

Cysteine was found to have a superb affinity for ozonated SWNT. Addition of this

reagent to ozonated SWNT formed grey flakes that precipitated over time (pointed with

an arrow in Figure 1). Formation of a cross-linked network where both amino and thio

groups participate in reaction is likely. Cysteine in the reference sample did not show

such behavior.

180

a b c d e

a guanidine·HCl + 1 eq NaOH

b guanidine·HCl + 1 eq NaOH (reference)

c guanidine·HCl

d guanidine·HCl (reference)

e TMPD

f g h i j

f GABA + 1 eq NaOH

g GABA + 1 eq NaOH (reference)

h GABA

i GABA (reference)

j TMPD

k l m n o

k Lysine + 1 eq NaOH

l Lysine + 1 eq NaOH (reference)

m Lysine

n Lysine (reference)

o TMPD

p q r s t

p Cysteine + 2 eq NaOH

q Cysteine + 2 eq NaOH (reference)

r Dithiothreitol

s Dithiothreitol (reference)

t TMPD

u v w x y

u MeOCH2CH2NH2

v MeOCH2CH2NH2 (reference)

w Arginine + 3.5 eq NaOH

x Arginine 3.5 eq NaOH (reference)

y TMPD

p q pz qz

p Cysteine + 2 eq NaOH

q Cysteine + 2 eq NaOH (reference)

pz Zoom-in of p

qz Zoom-in of q

Figure 1. Trapping reactive centers on SWNTOn with electron rich amines and thiols

181

6.3. Conclusion

Amines and amino acids salts were found to be effective traps of oxidized sections of

SWNT. Addition of TMPD resulted in no coloration. The first step in trapping is an

electron transfer from amine to ozonated SWNT with subsequent attachment of aminium

radical to SWNT surface. Guanidine hydrochloride, having no available lone pair, was

unable to prevent formation of TMPD˙+ radical.

6.4. Experimental Part

Wurster reagent was prepared by dissolution of TMPD (55 mg) in EtOH (10 mL,

200 proof). A stock suspension of SWNT (pristine SWNT, HipCo, batch 162.8, Rice

University) was prepared in ethanol with concentration less than 2 mg/70 mL. Occasional

swirling of SWNT stock suspension was performed to ensure equal amounts of carbon

nanotubes were withdrawn during sample preparation. Stock solutions were prepared for

each of the reagents to ensure an identical concentration of reagents in reference and

reaction samples (see Table 1). SWNT suspension (less than 0.1 mg; 2.5 mL EtOH) was

bubbled with O3/O2 gaseous mixture (ca. 3 v/v % ozone) for 1min at r. t., sample was

purged with argon, then a solution of a compound of interest was added and the mixture

swirled. After 15 min, Wurster reagent (150 uL) was added. For the corresponding

reference sample a reagent of interest (see conditions in Table 1) and Wurster reagent

(150 uL) were added to SWNT suspensions without ozonation or purging.

182

Table 1. Reagent concentrations for trapping of SWNT(O)n reactive centers*

Compounds

FW

weight eq. of

1N

NaOH other

H2O

vol

total

vol

mg mmol NaOH vol, uL solvent needed uL

Cysteine 121.16 5 0.0413 2 82.5 17.5 100.0

Lysine 146.19 5 0.0342 1 34.2 65.8 100.0

Lysine- no NaOH 5 0.0 100.0 100.0

Arginine-HCl 174.2 5 0.0287 3.5 100.4 0 100.0

GABA 103.1 5 0.0485 1 48.5 51.5 100.0

GABA- no NaOH 5 0.0 100.0 100.0

Guanidine-HCl 95.53 10 0.1047 1 104.7 95.3 200.0

Guanidine-HCl -

no NaOH 10 0.0 200.0 200.0

Dithiothreitol 154.25 10 0.0648 0 0.0 EtOH 0.0 200.0

MeOCH2CH2NH2 10ul 0 0.0 EtOH 0.0 200.0

* - Double the amounts to prepare stock solutions of each reagent.

6.5. References and Notes

There are no references for this chapter.

183

Chapter 7

Reactions between ozonated SWNT and different classes of compounds

studied by X-ray photoelectron spectroscopy

184

7.1. Introduction

A covalent attachment of amines to ozonated SWNT allows extension of this

methodology to biologically active molecules with thio and amino moieties. This chapter

examines conditions needed to improve covalent attachment of amino acids to SWNT

sidewall. A particular problem to be addressed is the determination of the amount of

sodium hydroxide needed for the reaction.

The successful attachment of a single amino acid would imply that a number of

amino acids, or peptides, could be attached to the surface of SWNT, fullerenes and

similar species. The simplicity and the speed of the reaction between ozonated SWNT or

ozonated C60 and amines make it a great tool for synthetic chemistry.

7.2. Results and Discussion

A great number of reagents containing heteroatoms like N, S, Cl, Br and F have

being tried in reactions with ozonated SWNT (see Appendix C). Amines and thiols were

shown to be the most effective ones. Nearly all essential amino acids were shown to react

with ozonated SWNT at r. t. within minutes. XPS gave nitrogen concentrations in a range

2 – 3 atomic % on the surface of SWNT for a majority of compounds.

To ensure that the observed peaks in XPS spectra are not the result of a hydrophobic

interaction or some unknown side reaction, all reactions were run against their own

references. Reference samples were prepared by exactly the same procedure, but did not

include an ozonation step. A special ‘washing’ step was developed for a complete

removal of an unreacted reagent from SWNT surface. Introduction of the washing step

was imperative for complete removal of nonvolatile reagents like amino acids. Judged by

185

the absence of characteristic peaks (N, S, F), a vast majority of reference XPS spectra had

no unreacted reagents present.

Reaction is believed to proceed through an electron transfer from an electron rich

nucleophile (amine, thiol) to an ozonated SWNT. A conversion of an amino group to its

free form, i.e. RNH3+ to RNH2, was necessary for reaction to take place. Ammonium

salts, RNH3+, were shown to have no reaction with ozonated SWNT within 15 min after

addition (Chapter 6).

A thin film of SWNT, a buckypaper, deposited onto a Teflon® 0.2 um filter by

filtration (Figure 1), was employed for ozonation and subsequent reactions with reagents

of choice.

Figure 1. Plates for XPS measurements. SWNT film was deposited onto 0.2 um

Polypropylene backed PTFE filter by filtration. Ruler is provided for reference.

With respect to XPS spectra, Teflon ® surface was found to be of a greater benefit

when compared to indium foil. Particularly, obtained spectra did not have intense peaks

from the substrate commonly seen on spectra with indium support. From a sample

preparation standpoint, it was important that all samples had the same thickness of

SWNT film on a substrate. Whatman 0.2 um Polypropylene backed PTFE membrane

filter (WTP type) was ideal for such purpose. Bath sonicated slurry of SWNT in ethanol

was filtered through PTFE filter to obtain a uniformly distributed SWNT layer on the

186

surface of a membrane. SWNT film was found to have reasonable conductivity and XPS

measurement (with neutralizer turned on) did not result in charge accumulation on

SWNT surface.

Figure 2. Plate holder for XPS measurements. Zoom-in shows a SWNT film at the

bottom of the well and a Teflon® ‘lid’. (A) 24-well holder for XPS plate handling, (B) a lid to

prevent SWNT samples from flying around due to static, (C) 4 x 6 mm SWNT film deposited

onto PTFE 0.2 um filter at the bottom of the well.

Teflon ‘plates’ with SWNT film on it had dimensions 4 x 6 mm, and were found to

be very sensitive to static. Figure 2 shows a plate holder used for samples’ collection and

transportation.

Typical XPS spectra for reference and ozonated samples are shown in Figure 3.

187

Binding Energy (eV)

02004006008001000

Co

un

ts /

se

c

x1

04

1

2

3

4

5

Co

un

ts /

se

c

x1

04

1

2

3

4

5

- O

KLL

- F

KLL

- F

1s

- O

1s

- N

1s

- C

1s

- F

e 2

p3

/2

- O

KLL

- O

1s

- F

e 2

p3

/2

- C

1s

392400408

SWNT + Guanidine

Ozonated SWNT + Guanidine

NH

H2N NH2

Figure 3. XPS spectra of SWNT reacted with an aqueous solution of guanidine. Top: a

product from reaction with pristine SWNT, a reference. Bottom: a product from reaction

with ozonated SWNT. A zoom-in in the right upper corner shows N 1s peak. A fluorine

peak comes from PTFE substrate.

The characteristic peak for nitrogen at ca. 400 eV was observed for amines and a

majority of amino acids used in this work. The spectrum obtained on a sample treated

with ozone and guanidine was estimated to have ca. 6 atomic % concentration of

nitrogen. This number is about 2 – 3 times higher than those seen for amines. Such an

increase is thought to be due to the three nitrogen atoms in guanidine molecule. Also,

guanidine does not have bulky substituents to hinder it from incident X – rays during

spectrum acquisition.

188

A typical XPS spectrum for a product of reaction of an amino acid and ozonated

SWNT is shown in Figure 4.

Binding Energy (eV)

02004006008001000

Co

un

ts /

se

c

x1

04

1

2

3

4

5

Co

un

ts /

se

c

x1

04

1

2

3

4

5

- O

KL

L

- F

KL

L

- F

1s

- O

1s

- N

1s

- C

1s

- F

e 2

p3

/2

- O

KL

L

- O

1s

- F

e 2

p3

/2

- C

1s

392400408

SWNT + Argininate, Na salt

Ozonated SWNT + Argininate, Na salt

- N

a 1

s

- N

a K

LL

O

ONa

NH2

NH

H2N

NH

Figure 4. XPS spectra of SWNT reacted with an aqueous solution of argininate. Top: a

product from reaction with pristine SWNT, a reference. Bottom: a product from reaction

with ozonated SWNT. A zoom-in in the right upper corner shows N 1s peak. A fluorine

peak comes from PTFE substrate. Atomic concentrations (%) for bottom spectrum are:

C1s 72.9, O1s 20.0, N1s 3.4, Na1s 2.6, Fe2p3/2 1.0.

Sodium peaks were commonly observed in XPS spectra of samples that underwent

ozonation and treatment with sodium salts of amino acids. The peak is likely from the

carboxylate of an amino acid. Other possibilities, like formation of carboxylates during

ozonation and sodium anion exchange during chemical treatment should not be excluded.

189

Carbon nanotube ozonides were shown to decompose with lifetime < 3 min at r. t.

(Chapter 2). The ability of SWNT to react with amines at different time intervals after

ozonation was studied by XPS. Spectra were recorded at 0 min, i.e. right after, 10, 20, 40

and 60 min after ozonation.

Binding Energy (eV)

300400500300400500

Co

un

ts /

se

c

x1

04

1

2

3

4

300400500

- O

1s

- N

1s

- C

1s

- O

1s

- N

1s

- C

1s

- O

1s

- C

1s

0 min 60 minReference

Figure 5. XPS spectra of SWNT reacted with 2-methoxyethylamine. Labels N 1s mark

nitrogen peaks. Left: a product from reaction with pristine SWNT, a reference. Middle:

amine added right after SWNT ozonation. Right: amine added 60 min after ozonation.

Spectra for reactions of 2-methoxyethylamine with ozonated SWNT right after and

60 min after ozonation are shown in Figure 5. A corresponding reference spectrum is

shown for comparison. All samples were found to have amine attached to its surface.

This means that the covalent attachment of amine to SWNT is independent of the

presence of ozonides. The decomposition of 1,2,3-trioxolanes resulted in the formation of

a carbon nanotube SWNTOn , exhibiting unique oxidative properties not seen in pristine

SWNT.

190

As a part of the experiment, amine/amide enrichment of the SWNT surface was

accomplished by cycling ozonation and reagent addition. The results are summarized in

Figure 6.

Binding Energy (eV)

300500300500

Co

un

ts /

se

c

x1

04

1

2

3

4

5

300500 300500 300500

- O

1s

- N

1s

- C

1s

- O

1s

- N

1s

- C

1s

- O

1s

- O

1s

- N

1s

- C

1s

- O

1s

- N

1s

- C

1s

- C

1s B EDA C

Figure 6. SWNT surface enrichment with 2-methoxyethylamine. Shown are XPS

spectra. Samples C-E went through cycling ozonation - reagent addition. Each section

(A-E) represents a separate experiment. Labels N 1s mark nitrogen peaks; (A) after

reaction with pristine SWNT, a reference, cN 0 at. %; (B) after reaction with ozonated

SWNT, cN 2.8 atomic %; (C) 2 cycles of ozonation – amine addition, cN 4.5 at. %; (D) 3

cycles, cN 6.0 at. %; (E) 4 cycles, cN 7.8 at. %.

Four cycles were made; each one included: SWNT plate (4 x 6 mm buckypaper on

Teflon® substrate) ozonation for 1 min with a stream of O3/O2 gaseous mixture (ca. 1.5

v/v % ozone), a reaction with neat amine (8 uL) for 1 min and a vacuum drying for 3

min. XPS gave ca. a 2 % increase in atomic concentration of nitrogen on the surface of

SWNT with each subsequent cycle. Similar amine/amide enrichment work was done with

IR measurement, results of which are discussed in Chapter 4. Conversion of amines to

191

amides is expected for multiple ozonation cycles, and was demonstrated by formation of

characteristic amide bonds at 1680 cm-1

(Chapter 4). The possibility of nitrogen peak

splitting into two nonequivalent ones was not studied by XPS. More in-depth work is

needed to determine possible conversion of amines to amides during subsequent

ozonation steps. Values of nitrogen atomic % on the surface of SWNTOn were obtained

with survey scans and will need to be refined in future work.

Reagents found to have good reactivity with ozonated SWNT (SWNTOn) are shown

in Figure 7 (see XPS spectra in Appendix C.)

192

NH

H2N NH2

HON

N

NN

NH2

NHNH

O

ONaO

OONa

N

N

O

ONa

NH2

NaO

O

ONa

O

NH2

O

ONa

NH2

NaO

O

O

ONaH2N

HN

N NH2

ONa

O

NH2

ONa

O

H2N

NH2

OH

O

H2N

ONa

O

NH2

S

H2N

O

ONa

O

NH2

ONa

O

H2N

O

NH2

O

ONa

NH2

NH

H2N

NH

Argininate, Na+

Glutaminate, Na+

Asparaginate, Na+

Methioninate, Na+

Lysinate, Na+

Histidinate, Na+

Glycinate, Na+

Glutamate, 2Na+

Aspartate, 2Na+

Alaninate, Na+

Lysine

Wurster reagent

Folic acid, 2 Na+ salt

Guanidine

NH2

OH O

OH Threoninate, Na+

NH2HO

ONa

O

Tyrosinate, Na+

O

ONaH2N GABA, Na+ salt

ONH2

2-methoxyethylamine

Aqueous ammonia

H2NNH2

H2NOH

NH2Isoamylamine

ethanolamine

ethylenediamine

NH3aq.

N

N

HO

SONa

O

OHEPES, Na+

Figure 7. Reagents found to be reactive with ozonated SWNT. All compounds gave

characteristic N 1 s peak in XPS spectra.

193

Compounds with more than one nitrogen, e.g. lysinate, argininate, guanidine,

ethylene diamine, gave stronger signals for N 1s peak. In a separate study, ethylene

diamine was shown to crosslink SWNT, leading to its inability to get debundled upon

bath sonication in ethanol. Such behavior is another indication of a covalent attachment

of reagents to the surface of SWNT.

Possible reaction mechanisms are proposed on Scheme 1. Ozonides are thought to

oxidize amines and thiols, yielding no covalent attachment to SWNT surface.

Scheme 1

OO

OO

OO

OO O

Nu

epoxide

O

oxidoannulene

O O

Nu

1,2,3-trioxolane

Nu = AlkSH, ArSH, R3N, RNH2, R2NH, PR3, guanidine, etc.

Nu

Nu

Nu

SWNT=

OO

epoxide

RNH2

RNH2 (xs)

RNH2

Route A

Route B

RNH3

RNH

RNH O

NH

R

RNH3

RNH2

O

NH

R

H

OO

ONu

NuOno attachment ofNu to sidewall

Route C

194

Reactions with epoxides and oxidoannulenes were shown by XPS and IR to yield

reagent attachment to the SWNT sidewall.

Samples that had poor reactivity with SWNTOn included: isoleucinate, valinate,

serinate, cysteinate, aniline, triethylamine, thiophenol and p-fluorothiophenol. The latter

two compounds produce stabilized radicals, either PhS+ and F-C6H4S

+ or PhSH

+ and F-

C6H4SH+ or a combination of thereof and were expected to have poor attachment. Both

thiols were shown to have attachment to corresponding reference samples, either through

a side reaction or a non-covalent interaction.

Hydrophobic interaction or inefficient ‘washing’ steps are possible. Valinate,

isoleucinate and triethylamine had poor attachment probably due to steric hindrance.

Aniline, having its nitrogen lone pair in conjugation with an aromatic ring was shown to

give a weak signal for N 1s peak, as expected. The acidity constant of PhNH3+ is pKa ~

4.6 (in water) which is much lower than that of amines (pKa > 8).

Samples with uncertain reactivity towards SWNTOn included: uracil, adenine and

2-mercaptopyridine. All compounds have nitrogen atoms with delocalized lone pairs.

HN

NH

O

O

SNH

N

NNH

N

NH2

H

N SH

pKa ~ 9

pKa ~ 3.5

uracil adenine 2-mercaptopyridine

delocalized lone pair

The nature of uracil attachment is not clear. While uracil can be deprotonated at

pH > 9.2, its protonation would require a very strong acid. Lone pairs of nitrogen atoms

of amide-like groups can be protonated at pH ~ 0. In turn, this means that this lone pair is

195

not ‘available’ for an electron transfer to SWNTOn. Poor reactivity with SWNTOn for all

three compounds was expected.

Samples that had no reactivity with SWNTOn included: urea, NaCN, NaBr, TsCl and

CF3CH2OH. Lack of reactivity between urea and SWNTOn seemed reasonable, taking

into account that lone pairs on urea nitrogen atoms are delocalized onto the adjacent

carbonyl. Reasonably good nucleophiles, like CN –

and Br –

, were shown to be unreactive

with SWNTOn, indicating that nucleophilic attack on SWNTOn may not be a

requirement for amine or thiol attachment observed in other experiments.

Trifluoroethanol did not react with ozonated SWNT as expected.

Samples that were shown to have hydrophobic interaction with SWNT included

cysteine and phenylalanine (see spectra in Appendix C). Cysteine (10 mg/200 uL) was

shown to crash out of aqueous solution onto the surface of SWNT film. Apparently, a

thiol group had a superb affinity for carbon nanotubes (Figure 8).

Figure 8. Comparison of SWNT films dipped into aqueous solutions of: (A-B) cysteine

sodium salt (10 mg/200 uL plus 2 eq. of NaOH) and (C-D) cysteine (10 mg/200 uL).

Top: ozonated SWNT. Bottom: reference SWNT

Development of cysteine or thiol based surfactants for SWNT solubilization may be

an interesting area to look into. XPS of samples C and D verified that precipitate is

cysteine. Phenylalanine was observed on both reference and reaction spectra. It is

believed that hydrophobic interaction kept this acid attached to SWNT surface.

196

7.3. Conclusions

A new methodology for SWNT sidewall functionalization has been developed for

amines and thiols and was extended to a class of amino acids. The main requirement for

amino acids to be reactive with ozonated SWNT was the presence of a sufficient amount

of sodium hydroxide to convert all free carboxylic groups to sodium salts, thus freeing up

amino groups for an electron transfer to ozonated SWNT.

Nearly all amino acids got attached to ozonated SWNT, while no attachment was

observed for non ozonated SWNT. A borderline case, where attachment could be either

covalent or hydrophobic or both, included phenylalanine. Cysteine was shown to have a

very good affinity for ozonated SWNT. It precipitated out of an aqueous solution forming

a thin film on the surface of SWNT. Usage of cysteine terminated surfactants for SWNT

encapsulation could be an area for further research. As expected, nucleophiles like CN –

and Br – were not able to get attached to SWNT surface, indicating that electron transfer

is the first step.

197

7.4. Experimental Part

General information

A sonication bath (Fisher Scientific, FS 60) was used for all bath sonications.

MiniVortex (Fisher Scientific) was used mainly for solubilization of amino acids. X-ray

Photoelectron Spectroscopy measurements were performed on PHI Quantera. XPS

samples were prepared either on indium foil (Aldrich) or on polypropylene backed PTFE

membrane filters (Whatman, 0.2 um, WTP type).

Millipore Milli-Q water (18.2 MOhm/cm at 25 C) was used for experiments with

amino acids. A pH meter (Fisher Scientific, Dual Channel pH meter AR 50) and 1N aq.

NaOH (Fisher Scientific) were used for pH adjustments.

All -amino acids used in this work were natural. Adenine, uracil and the majority of

essential amino acids were from US Biological; asparagine, NH4OH, glycine and urea

were obtained from Fisher; cysteine, arginine hydrochloride and guanosine were from

Aldrich; glutamic acid was from Acros; benzene from EMD Chemicals Inc.; ethyl

alcohol (200 proof) was from AAPER Alcohol & Chemical Co.; kimwipes ® were from

Kimberly-Clark.

SWNT plate preparation

SWNT (2-4 mg, SWNT, HipCo, batch 162.8, raw, unpurified, Rice University) were

debundled in benzene (10 mL) with a bath sonicator (Fisher Scientific, FS 60) until there

were no more solid particles present in the suspension, then concentrated to a slurry and

added dropwise to the surface of PFTE filter.

198

Procedure for Wurster reagent

A Teflon plate covered with SWNT film was ozonated for 1 min with a stream of

O3/O2 gaseous mixture (ca. 1.5 v/v % ozone). The plate was dipped into a solution of

TMPD (2.7 mg in 500 uL EtOH) for 15 min, and then washed two times in EtOH (1 mL)

and one time in water (1 mL) for 3 min each.

Procedure for enrichment of 2-methoxyethyl amine on the surface of ozonated

SWNT

A Teflon plate covered with SWNT film was ozonated for 1 min with a stream of

O3/O2 gaseous mixture (ca. 1.5 v/v % ozone). Neat amine (8 uL) was added, SWNT plate

was covered with a jar to prevent amine evaporation, reacted for 1 min, the SWNT plate

was dried under vacuum for 3 min and this cycle was repeated as many times as

necessary.

Procedure for reaction of NaBr, NaCN and CF3CH2OH with ozonated SWNT

A Teflon plate with deposited SWNT film was ozonated for 1 min with a stream of

O3/O2 gaseous mixture (ca. 1.5 v/v % ozone). Neat trifluoroethanol (8 uL) was added,

SWNT plate covered with a jar, reacted for 1 min and dried under vacuum for 3 min

before XPS measurements.

For reactions with salts, SWNT plates were dipped into dilute aqueous solutions of

NaBr and NaCN for 1 min, then dipped two times into deionized water (3 min each),

wiped off excess of water with Kimwipes ® and dried under vacuum for 3 min before

XPS measurements.

199

A general procedure

Deionized water (MilliQ) was used for all aqueous solutions. Solvents and amounts

of reagents are noted in Table 1. To help with solubilization, bath sonicator and

MiniVortex were used as needed. Prepared solutions were stored at 10 C before use.

All reactions and washes were performed in 1.5 mL microcentrifuge tubes

(Eppendorf vials). Pristine SWNTs (HipCo, batch 162.8, Rice University) were bath

sonicated in ethanol for 40 min, solvent was filtered out on 0.2 um PTFE filter

(polypropylene backed, Whatman, WTP type) to form an equally distributed SWNT film

on its surface. Filter was cut into 4 x 6 mm sections for use in chemical reactions.

A 4 x 6 mm section of SWNT film on Teflon ®

substrate was subjected to a steam of

O3/O2 gaseous mixture (ca. 1.5 v/v % ozone) for 1 min. The film was removed from the

ozonation chamber and submerged into a solution of interest (see Table 1 for details) for

3 min. Film was taken out of a reaction solution, excess of liquid was removed with

Kimwipes.® Unless otherwise noted in Table 1, washed reaction plate two times in

deionized water (1 mL) for three minutes each. Plates were dried under vacuum for 15

min before loading into XPS machine.

200

Table 1. Preparation of reagents for XPS experiment

Reagent FW mg mmol NaOH

eq

1N NaOH

uL

Other

solvent

H2O

uL

Total

uL

Alanine 89.09 5 0.0561 1 56.1 43.9 100

Cysteine 121.16 10 0.0825 2 165.1 30.0 200

Cysteine No NaOH 121.16 10 0.0825 0 0.0 200.0 200

Aspartic acid 133.1 5 0.0376 2 75.1 24.9 100

Glutamic acid 147.13 10 0.0680 2 135.9 0.0 136

Phenylalanine 165.19 5 0.0303 1 30.3 69.7 100

Glycine 75.07 5 0.0666 1 66.6 33.4 100

Histidine 155.16 5 0.0322 1 32.2 67.8 100

Isoleucine 131.17 5 0.0381 1 38.1 61.9 100

Lysine 146.19 5 0.0342 1 34.2 65.8 100

Lysine No NaOH 146.19 5 0.0342 0 0.0 100.0 100

Methionine 149.21 5 0.0335 1 33.5 66.5 100

Asparagine 132.12 5 0.0378 1 37.8 62.2 100

Proline 115.13 5 0.0434 1 43.4 56.6 100

Glutamine 146.15 5 0.0342 1 34.2 65.8 100

Arginine-HCl 174.2 5 0.0287 2 57.4 42.6 100

Serine 105.09 5 0.0476 2 95.2 4.8 100

Threonine 119.12 5 0.0420 1 42.0 58.0 100

Valine 117.15 5 0.0427 1 42.7 57.3 100

Tyrosine 181.19 5 0.0276 2 55.2 44.8 100

GABA 103.1 5 0.0485 1 48.5 51.5 100

urea 60 10 0.1667 0 0.0 100.0 100

uracil a 112 5 0.0446 0 100.0 40 uL

CH3CN

- 140

adenine a 135.1 5 0.0370 0 100.0 40 uL

CH3CN

- 140

NH4OH aq. solution b 35 20 uL - 0 0.0 H2O 80.0 100

Guanidine·HCl 95.53 10 0.1047 1 104.7 95.3 200

2-Mercaptopyridine c 111.17 10 0.0900 0 0.0 CH3CN 0.0 200

HEPES 238.3 10 0.0420 1 42.0 158.0 200

Folic acid d 441.4 10 0.0227 2 45.3 154.7 200

a First wash 40 : 100 uL (CH3CN : 1N aq. NaOH), then 2x water 1 mL;

b - Dilute 5x NH4OH

conc; dip teflon plate into this solution; c - First wash 1 mL CH3CN then 2x H2O;

d - First wash

200 uL of 1N NaOH, then 2x H2O (1 mL).

201

7.5. References and Notes

There are no references for this chapter.

Part II

203

Chapter 1

Photorearrangement of -Azoxy Ketones and Triplet Sensitization of

Azoxy Compounds

204

1.1. Abstract

Ph

O

NN

O

R' O NN

R'Ph

O

hv

a R' = t-Bu

b R' = Ph4 8

A recent review on the radical chemistry of the azoxy group1 revealed that the nature

of -azoxy radicals remains unclear. This chapter presents the generation of -azoxy

radicals under mild conditions by irradiation of -azoxy ketones 4a,b. These compounds

undergo -cleavage to yield radicals 5a,b, whose oxygen atom then recombines with

benzoyl radicals to produce presumed intermediate 15. Formal Claisen rearrangement

gives -benzoyloxyazo compounds 8a,b, which are themselves photolabile, leading to

both radical and ionic decomposition. The ESR spectrum of 5a was simulated to extract

the isotropic hyperfine splitting constants, which showed its resonance stabilization

energy to be exceptionally large. Azoxy compounds have been found for the first time to

be good quenchers of triplet excited acetophenone. The main sensitized photoreaction of

7Z in benzene is deoxygenation. The principal direct irradiation product of 4bZ and

model azoxyalkane 7Z is the E isomer, whose thermal reversion to Z is much faster than

that of previously studied analogues.

1.2. Introduction

Despite its occurrence in a number of biologically active molecules,2-6

the azoxy

functional group has received less attention than its lower oxidation state analogue, the

205

aliphatic azo group.7 The latter one is a widely used source of free radicals, but as pointed

out in a recent review,1 the radical chemistry of azoxy compounds is sparse.

The generation of -azoxy radical 2, also known as hydrazonyl oxide 3, has been

proposed in the literature, but no product attributable to this radical could be found.

Perester decomposition yielded -azoxy radical 1, which in turn underwent fragmentation

to ethylene with a rate constant below 2 × 105 s

-1 at 120 C (Scheme 1).

8 On the other

hand, bromination of azoxy compounds is known to occur at the distal carbon (i.e. away

from azoxy oxygen), presumably via -azoxy radicals.3, 9

The goal of the present work

was to generate -azoxy radicals under mild conditions in hopes of learning more about

their chemistry.

Scheme 1

NN

O

NN

O

H2C CH2

NN

O

31 2

1.3. Results

Norrish Type I photochemical cleavage of -azoxy ketone 4, a previously unknown

structural type, was chosen as the path to -azoxy radicals. Photolysis of some phenones

and benzyl ketones yields alkyl radicals in solution;10-14

hence, it is not unreasonable that

4a-c might also exhibit -cleavage (Scheme 2).

206

Scheme 2

R1

O

NN

O

R2hv

a: R1 = Ph, R2 = t-Bu

b: R1 = Ph, R2 = Ph

4

NN

O

R2

R1

O

c: R1 = PhCH2, R2 = t-Bu

5

Photochemistry of 4a. The extinction coefficients of 4a at two photochemically

useful wavelengths are included in Table 1, and the spectra can be found in Appendix D.

Table 1. UV Extinction Coefficients of Azoxy and Azo Compounds

compd (313 nm) (366 nm)

4a 128 16

4bZ 830 22

8a 4.7 26

8b 212 80

6 32 1.9

7Z 262 7.7

PhCOMe 41 6.7

Ph

O

NN

O

NN

O

Ph

O

NN

O

Ph

NN

O

Ph

313 = 32 313 = 262

313 ~ 40

313 = 830 313 = 128

313 ~ 40

7Z6

207

The bichromophoric molecule absorbs twice as strongly as the sum of its two

individual chromophores, represented by acetophenone and azoxy-tert-butane 6.15

Irradiation of 4a in benzene at 313 nm and 25 C caused clean rearrangement to an

unexpected product, azoester 8a (Scheme 3), whose extinction coefficients are included

in Table 1. The quantum yield for appearance of 8a was 0.02 at 23, 60, and 100 C, and

the reaction even proceeded at -78 C ( was established with DBH actinometry).

Scheme 3

Ph

O

NN

O

R' O NN

R'Ph

O

hv

a R' = t-Bu

b R' = Ph

4 8

The disappearance rate of 4a was not diminished by the inclusion of 0.1 M

biphenyl16

as quencher, but 0.1 M 1,3-cyclohexadiene, a quencher of much lower triplet

energy than biphenyl, decreased the conversion of 4a to 8a by 18.1%. Table 2

summarizes triplet state energies of acetone, acetophenone, biphenyl and

1,3-cyclohexadiene (CHD) used in this work.

Table 2. Triplet energies and lifetimes for excited acetone, acetophenone, Ph-Ph and

CHD.

Compd ET, kcal / mol Lifetime T, us Reference

acetone 80 17

PhCOMe 73.6 0.86 this work

Ph-Ph 65.5 16

CHD 54 1.3 18, 19

208

To determine whether -cleavage of the ketone moiety was on the pathway from 4a

to 8a, the benzoyl radicals were trapped with tert-butyl thiol.20

Irradiation of 4a at 313

nm with 0.16 M tert-butyl thiol afforded benzaldehyde in 25% yield, as determined by

NMR. A corresponding reduction in the amount of 8a was observed. No product from

trapping of 5a could be identified.

The photolysis of 4a was also investigated by ESR spectroscopy. Irradiation of 4a in

toluene-d8 at + 8 C with a full 500 W mercury arc lamp gave a weak but structured ESR

signal. Suspecting that the signal was due to radical 5a, this intermediate was generated

independently by irradiation of di-tert-butyl peroxide21, 22

containing

tert-butyl(ONN)azoxy-2-propane 9 at -78 C. A stronger signal ~ 60 G wide containing

about 37 lines was observed and was very similar in appearance to the one from 4a (See

ESR spectra in Appendix D).

Scheme 4

The GC trace of the irradiated solution showed several peaks, none of which were in

the right region for the C-C dimer of 5a.8 To support the assignment of the ESR signals to

5a, the experiment was repeated with 9 at -50 C. The superior resolution (see Appendix

D) allowed isotropic simulation (coefficient of determination r2 = 0.926) with the

program WinSim,23

which led to the following splitting constants: N1 13.75 G, N2 1.91

G, 3H 5.64 G, 3H 5.12 G. A second radical was present in the spectrum but since its

intensity was much lower than that of 5a, this minor species was not studied further.

209

Photochemistry of 4bZ. As shown in Table 1, 4bZ absorbs about three times more

strongly at 313 nm than the sum of its two chromophores, which are represented by

acetophenone and 7Z.24-26

Irradiation of 4bZ at 313 nm both at 25 and -78 C gave the

intensely yellow 8b, but the major reaction was Z E photoisomerization about the

azoxy double bond.27, 28

The E isomer (4bE) reverted to the Z isomer 4bZ with a half-life

of 14 min at 50 C. This behavior supports the structural assignment of 4bE.

Scheme 5

The chemical shift changes in toluene-d8 upon Z-E isomerization (4bZ Me2 1.716

ppm vs 4bE Me2 at 1.377 ppm) are similar to those of 7Z 7E (1.472 vs 0.943 ppm).

The magnitude of the upfield shift upon Z E isomerization is larger than that in the

methyl and ethyl analogues.28

Irradiation of 4bZ at ambient temperature and 313 nm in

acetone-d6 as solvent and triplet sensitizer29-32

exhibited the same rearrangement quantum

yield as in toluene-d8. Irradiation of 4bZ in toluene at -53 C in the ESR cavity led to a

weak signal about 50 G wide consisting of 13 peaks, which were much broader than

those from 4a. This spectrum was definitely not due to phenyldiazenyl radicals33

arising

from secondary photolysis of 8b. Instead, the similarity of the major nitrogen coupling to

that of 5a suggests that this radical is 5b.

210

Irradiation of 4c. This compound was designed to decarbonylate photochemically

but it proved to be quite photostable. Thus 4c upon irradiation with an unfiltered 500 W

mercury lamp gave neither 8c nor the E azoxy isomer but instead slowly decomposed to a

mixture of many products.

Triplet Sensitization of Model Azoxy Compounds. Because the

photorearrangement of 4bZ 4bE in acetone-d6 is the first reported triplet sensitized

reaction of an azoxy compound, we examined the azoxy group as a triplet quencher.

Acetophenone was chosen as the donor because of its high triplet energy (73.6

kcal/mol),17

its structural similarity to the ketone moiety of 4, and because it is known to

phosphoresce in solution.34

Azoxy-tert-butane 6 (0.012 M) and phenylazoxy-tert-butane

7Z (0.0021 M) were found to be strong quenchers of acetophenone emission intensity.

The quenching rate constant in isooctane was obtained by quenching the triplet lifetime

of acetophenone, giving kq values of 9.8 × 108 M

-1 s

-1 for 6 and 4.4 × 10

9 M

-1 s

-1 for 7Z.

Since 6 and 7Z proved to be good triplet quenchers, the triplet sensitized

photochemistry of 7Z was investigated. Two degassed and sealed NMR tubes were

prepared, one containing 0.0191 M 7Z in toluene-d8 and the other containing the same

concentration of 7Z plus 0.437 M acetophenone. The high concentration of acetophenone

was needed to make it the main light absorber because the extinction coefficient of 7Z

exceeded that of acetophenone at 366 nm (cf. Table 1). Irradiating both tubes in parallel

for 4 h at 366 nm led to the same products but in different amounts (cf. Table 3).

211

Table 3. Relative GC Peak Areas in the Direct and Sensitized Photolysis of 7Z at 366 nm

in toulene-d8

a Conditions: -5 C, 4 h, 62% of 7Z left unreacted as determined by an internal standard.

b Conditions: PhCOMe, -5 C, 4 h, 58% of 7Z left unreacted.

Whereas Z E isomerization was dominant under direct irradiation, this process is

actually negligible with acetophenone present. We calculate that 4.6% of the incident

light was absorbed directly by 7Z in the sensitized experiment, as compared to 16% in the

direct irradiation, where the overall absorbance was only 0.073. Since the contribution of

direct photolysis in the sensitized reaction was 4.6 / 16 = 29%, the bulk of the 7E seen in

the latter case arose by direct irradiation of 7Z.

The photoisomer 7E was thermally labile, exhibiting a half-life for reversion to 7Z of

2.3 h at 25 C. Due to steric repulsion, 7E is much more labile than its ethyl analogue,

which reverts completely to the Z isomer in 5 h at 110 C.28

On the other hand, 7E is

slightly more stable than 10Z, whose half-life for conversion of 10E is 1.36 h at 25 C.35

The set 7E, 10Z provides a rare opportunity to compare the thermal isomerization rate of

a cis azoalkane with that of its related azoxy compound.

Deoxygenation was the main process observed under triplet sensitization but the fate

of oxygen is unknown. The structure of the deoxygenated product, phenylazo-tert-butane

10Z,E, was proven by comparison with an authentic sample.36, 37

The sensitized

212

irradiation was repeated in C6D6 to rule out hydrogen / deuterium abstraction from

toluene-d8, but the outcome was the same.

Because direct irradiation of azoxy compounds can led to oxadiaziridines27, 38-40

as

well as Z E isomerization,27, 28

it was important to verify the structure of 7E.

Surprisingly, no 1H NOE was seen for the tert-butyl protons and the ortho H's of the

aromatic ring. This negative result prompted for an examination of the NOE of a model

compound 10Z, whose trans-cis photoisomerization is well-known.35

Since 10Z also

failed to exhibit NOE, experimentally measured 15

N chemical shifts were compared to

those calculated theoretically,41

as shown in Table 4. The proximal and distal nitrogens

were assigned unambiguously by 1H-

15N HMBC.

Table 4. 15

N Chemical Shiftsa

7Z calc 7Z obs 7E calc 7E obs 12 calc

Nb 1.2 -20.2 30.8 1.4 -115.4

Nc -39.7 -52.1 -23.9 -42 -164.3

a ppm from nitromethane standard.

b Nitrogen proximal to the tert-butyl group.

c N-O.

Although the calculated shifts are 10-20 ppm downfield from the observed value for

7Z, the predicted direction and magnitude of the changes upon Z E isomerization lie

within 8 ppm of the observed value. The chemical shifts of oxadiaziridine 12 are

calculated to fall drastically upfield from those of 7Z,E, ruling out 12 as the photoisomer.

Additional support for the structure of 7E was provided by its independent synthesis

from 10Z. A solution of 10E in toluene-d8 was irradiated in an NMR tube with a 150 W

213

xenon lamp until the 10Z:10E ratio was 0.35. The mixture was oxidized in situ with

MCPBA, resulting in the clean conversion of all 10Z to 7E within 10 min. The tert-butyl

group of 7E ( = 0.943 ppm) exhibited a sizable upfield shift in toluene-d8 relative to that

of 7Z (1.472 ppm), just as in the azo precursors (10E 1.311; 10Z 1.090).28

After 10Z had

disappeared, the oxidation of 10E continued at a much slower rate and oxygen was

introduced proximal15

to the tert-butyl group to yield major product 11 accompanied by

~ 6% 7Z. The reversal of oxidation regiochemistry between 10E and 10Z is noteworthy

and apparently unprecedented since the only acyclic cis azoalkane oxidized previously

was symmetrical.27

Secondary Photolysis of Azoesters. As seen in Table 1, the azoester products 8a,b

absorb enough light to undergo secondary photoreactions and in fact these can complicate

the mechanistic interpretation of azoxy ketone photolysis. Irradiation of authentic 8a,b

yielded nitrogen plus a mixture of products, as shown below (Scheme 6). In 8b only,

decomposition was accompanied by isomerization to the cis azo isomer 13. Because it

was initially surprising to find benzoic acid as the major product of 8a,b, the irradiation

of 8a was repeated in MeOH-d4. The benzoyloxy group was found partially replaced by

CD3O to yield 14, consistent with ionic dissociation of the substituent to the azo

group.42, 43

214

Scheme 6

1.4. Discussion

Mechanism of 4 to 8. The simplest mechanism for the photorearrangement begins

with -cleavage to benzoyl radical plus -azoxy radical 5a. These radicals recombine at

azoxy oxygen to afford intermediate 15, which then rearranges to the final product 8

(Scheme 7). However, careful scrutiny of the reaction by low-temperature NMR failed to

reveal the presence of 15.

215

Scheme 7

A literature search revealed no published structures quite like 15 44-46

and an attempt

to generate 15 from acetone tert-butyl hydrazone and benzoyl peroxide at 0 C led

cleanly to 8a and unknown NMR peaks not consistent with 15a. Having failed to observe

this postulated intermediate by NMR, the activation energy for Claisen rearrangement

was calculated theoretically at the B3LYP/6-31G* level by a collaborator.41

Table 5

shows that the activation enthalpy lies in the same range as that for the Cope

rearrangement.47

Radical delocalizing substituents R1 and methyl groups R

2 and R

3

decrease H but not enough to explain an inability to detect 15. Homolytic dissociation

of 15 was calculated to require more energy than the figures in Table 5. Perhaps 15

absorbs UV light as does a peroxide or N-chloroamine and 15 8 occurs

photochemically.

216

Table 5. Calculated Activation Energy for Claisen Rearrangement of 15 to 8*

R1

R2

R3

HF, hartrees

TS,

hartrees H ,

kcal/mol

H H H -338.465 327 5 -338.419 204 8 28.9

H2C=CH- H H -415.868 578 8 -415.828 300 7 25.3

Ph H H -569.530 707 4 -569.490 234 4 25.4

H H Me -417.103 647 5 -417.069 106 5 21.7

H Me Me -456.418 324 7 -456.387 535 6 19.3

H2C=CH- Me Me -533.822 631 1 -533.789 195 6 21.0

* - calculation results courtesy of William B. Smith, Texas Christian University, Fort

Worth, Texas.

Lacking experimental evidence for 15, a verification was needed that -cleavage of

4a really is the first step of the reaction. Irradiation of this azoxy ketone with tert-butyl

thiol led to benzaldehyde20

but it was not clear initially whether it came from 4a or from

secondary photolysis of 8a. However, a plot of benzaldehyde versus time showed a

sizable initial slope while a similar plot of acetone, a decomposition product of 8, showed

an initial slope of zero. If benzaldehyde arose solely from secondary photolysis of 8a,

both plots would have exhibited an initial slope of zero. The ESR experiment with 4a

also supports initial -cleavage since essentially the same spectrum was observed from

4a and from irradiation of 9 with di-tert-butyl peroxide. The nitrogen splitting constants

extracted from the latter spectrum (13.75, 1.91 G) are in accord with those of known

heavily substituted hydrazonyl oxide radicals (e.g. for t-Bu2C=N2-N

1(O•)-Ph, a(N

1) =

12.1 G, a(N2) = 2.7 G

48 and for t-BuPhC=N

2-N

1(O•)-t-Bu, a(N

1) = 12.5 G, a(N

2) = 1.7

G.49

The proton splittings of 5a are exceptionally small, reflecting the high degree of

radical stabilization that one might expect for this nitroxyl structure.50

217

There remains the possibility that -cleavage is a side reaction not on the path from

4a to 8a. However, in a t-BuSH trapping experiment where 60% of 4a reacted, the yield

of PhCHO was 25% while that of 8a and its photolysis products was 21% and 15%,

respectively. Thiol trap t-BuSH was found diminishing the amount of 8a by an amount

close to the yield of PhCHO, thus placing PhCO• on the reaction pathway. Since t-BuSH

failed to trap 5a, the only evidence for this radical comes from the ESR experiment.

The low quantum yield for photorearrangement of 4a,b may be attributed to efficient

recombination of benzoyl radicals at the original site of attachment. However, the spin

density of 5a is higher at oxygen than at carbon, as shown by the calculated spin densities

below (calculated by a collaborator). Moreover, cage recombination of other "allylic"

radicals occurs at both ends.51

Another explanation for the low quantum yield is radiationless decay of excited 4a,b,

which could be accompanied by Z E isomerization of the azoxy group. To explore the

importance of these factors, a brief investigation of the azoxy group as a triplet energy

acceptor was conducted.

Triplet Energy Transfer to Model Azoxy Compounds. Azoxyalkanes were found

to be surprisingly rapid triplet quenchers. The quenching rate constant of acetophenone

triplets was in the range of 109 M

-1s

-1, which is similar to that for azo-tert-butane.

52 The

"triplet sensitized" deoxygenation of azoxybenzene53

was shown to be a case of chemical

218

sensitization;54

that is, ketyl radicals derived from the sensitizer reduced the azoxy

group.55

For this reason, acetophenone-sensitized deoxygenation of 7Z in toluene (cf.

Table 3) was initially suspected to follow the same pathway. However, the product

distribution was unchanged in C6D6 solvent, suggesting that triplet sensitized

deoxygenation56

is a true photoreaction of 7Z. In contrast, direct irradiation gave cis-trans

isomerization, already a well-established process.27, 28, 57

Investigating the triplet energy

of azoxy compounds and the fate of oxygen are potential topics for further research.

Although the azoxy group quenches ketone triplets intermolecularly, it is

inappropriate to discuss 4a,b in terms of individual chromophores because the enhanced

UV absorption (Table 1) indicates mixing of electronic states. The experimental facts are

that upon direct irradiation in benzene-d6 neither compound undergoes deoxygenation,

only 4bZ exhibits cis-trans isomerization, and both compounds undergo inefficient

-cleavage of the benzoyl group.

Photolysis of -Benzoyloxyazoalkanes. The initial products 8a,b are themselves

photolabile, giving rise to a product mixture dominated by benzoic acid and acetone. Part

of the likely mechanism, shown in Scheme 8, involves the usual nitrogen loss from

azoalkanes, presumably via thermolysis of the labile cis isomer7 followed by plausible

reactions of the formed tert-butyl and acyloxyalkyl radicals 16. Published information on

1-acyloxyalkyl radicals is sparse and mainly concerns their inter- and intramolecular

addition to alkenes.58, 59

However, Wille recently reported that photolysis of 17 gave

cyclohexanone in 33% yield by fragmentation of radical 18,60

exactly analogous to the

behavior of 16.

219

Scheme 8.

The major photolysis product of 8a,b is benzoic acid but Scheme 8 cannot explain its

presence. Levi and Malament reported that acyclic azoalkanes containing -chloro or

-acyloxy groups underwent heterolytic cleavage even in nonpolar solvents.42, 43

The

same unusual mechanism can be applied to formation of benzoate and the known highly

stabilized -azo cation 19.61

Most likely, benzoic acid arises by proton transfer from

adventitious water to benzoate, as the ortho protons of the acid were clearly visible by

NMR of the unopened, degassed, sealed tubes. The ionic mechanism cannot operate

exclusively, even though acetone could arise when 19 traps water, because it cannot

rationalize the obviously free-radical-derived products isobutane, benzaldehyde,

benzophenone, and isopropyl benzoate. Thus 8a undergoes competitive C-N homolysis

and C-O heterolysis. To support the ionic mechanism, 8a was irradiated in CD3OD and

monitored by 1H NMR (Scheme 9). The starting material disappeared over 19 h and gave

220

mainly 14, as proven by comparison with an authentic sample of the nondeuterated

analogue. On further irradiation, 14 also decomposed.

Scheme 9

Irradiation of 8b in C6D6 caused azo trans-cis isomerization, in accord with the

known photochemistry of phenylazoalkanes.36

Several of the products were similar to

those of 8a, again suggesting a competition between ionic and homolytic decomposition.

However, the yield of benzophenone and benzaldehyde from 8a was much higher than

that from 8b, indicating that that the C-N homolysis pathway is more important in 8a (cf.

Scheme 8). The -azo cation from 8b is surely more stabilized than 19, favoring ionic

decomposition of 8b.

1.5. Conclusions

Despite our failure to identify any products from 5a,b, their involvement in the

photorearrangement of -azoxy ketones 4a, 4bZ to 8a, 8b and their observation by ESR

shows that sterically unhindered -azoxy radicals are viable intermediates. The small

hydrogen hyperfine splittings in the ESR spectrum of 5a indicate a very high degree of

resonance stabilization. Direct or acetone sensitized irradiation of 4bZ also induces azoxy

Z E isomerization, but in model compound 7Z acetophenone triplets cause

deoxygenation without Z E isomerization. Azoxy compounds are surprisingly rapid

quenchers of acetophenone triplets. In both 4bZ and 7, steric repulsion causes thermal

221

reversion of the E azoxy isomer to be much faster than in previously reported

homologues. Azoesters 8a,b undergo photochemical C-O heterolysis in competition with

C-N homolysis to 1-acyloxy radicals, which fragment to ketones plus acyl radicals.

1.6. Experimental Part

2-tert-Butyl(ONN)azoxy-2-benzoylpropane, PhCOCMe2-N=N(O)-Bu-t (4a). A

solution of nitroso-tert-butane dimer (250 mg, 1.43 mmol, 0.8 equiv) in CH3CN (10 mL)

was stirred for 3.5 h at 25 C in the dark to allow dissociation to monomer. Meanwhile, a

solution of 2-benzoyl-2-aminopropane hydrochloride62

(795 mg, 3.98 mmol, 1 equiv) in

1.5 N aq HCl (12 mL, 4.5 equiv) was added dropwise to a suspension of Ca(OCl)2 (1.424

g, 5.97 mmol, 60 wt %, 1.5 equiv) in CH2Cl2 (24 mL) and water (24 mL) at 5 C. After

the mixture was stirred for 1 h at 5 C, the organic layer was separated and the water

layer was extracted with CH2Cl2 (2 × 8 mL). The combined organic phase was dried over

MgSO4, filtered, and concentrated to yellow oil (0.806 g, 87% yield). The crude N,N-

dichloroamine was light sensitive and was used immediately in the next step. 1H NMR

(CDCl3) 8.23 (m, 2H), 7.56 (m, 1H), 7.45 (m, 2H), 1.74 (s, 6H). Following Nelson et

al.,24

KI (298 mg, 1.79 mmol) was added to the above nitroso-tert-butane solution at 25

C, the temperature was lowered to 0 C, and the N,N-dichloroamine (416 mg, 1.79

mmol) in MeCN (5 mL) was added. The mixture was stirred for 2-3 h at 5 C, then the

temperature was gradually raised to 25 C and the mixture was stirred overnight. Water

(50 mL) and ether (25 mL) were stirred in and the water layer was separated and

extracted with ether (2 × 8 mL). The combined organic layer was extracted with

222

sufficient aq Na2S2O3 (~ 312 mg, 1.97 mmol) to change the color from dark brown to

light green. After drying over MgSO4 and removal of the solvent, the crystalline product

was purified by silica gel chromatography, eluting with 4:1 hexane:ethyl acetate, Rf 0.52.

Solvent evaporation yielded 273 mg (61%) of -azoxy ketone 4a, mp 70.5-71 C. Further

purification was effected by recrystallization from a small amount of MeOH. 1H NMR

(CDCl3) 7.84 (m, 2H), 7.46 (m, 1H), 7.35 (m, 2H), 1.59 (s, 6H), 1.31 (s, 9H). NMR

(C6D6) 7.99 (m, 2H), 7.08 (m, 1H), 6.99 (m, 2H), 1.64 (s, 6H), 1.07 (s, 9H). NMR

(toluene-d8) 7.89 (m, 2H), 7.10 (m, 1H), 7.01 (m, 2H), 1.59 (s, 6H), 1.09 (s, 9H). 13

C

NMR (CDCl3) 199.7, 135.0, 132.2, 127.78, 127.72, 76.1, 68.6, 27.6, 22.5. NMR

(toluene-d8) 198.2, 136.0, 131.8, 128.2, 127.7, 75.8, 68.7, 27.4, 22.6.

2-Phenyl(ONN)azoxy-2-benzoylpropane (4bZ), PhCOCMe2-N=N(O)-Ph.

2-Benzoyl-2-(N,N-dichloroamino)propane was reacted with 1.1 equiv of nitrosobenzene

in MeCN as described above, except that the nitroso dimer dissociation step was omitted

because nitrosobenzene exists largely as the monomer. The product was purified on silica

gel, eluting with 5:1 hexane:ethyl acetate, Rf 0.43. Vacuum drying afforded 325 mg

(68%) of product that was further purified by recrystallization from hot hexane (2 mL),

mp 56-58 C. 1H NMR (toluene-d8) 8.02 (m, 2H), 7.86 (m, 2H), 6.80-6.96 (m, 6H),

1.72 (s, 6H). 13

C NMR (toluene-d8) 197.5, 147.3, 135.5, 132.2, 131.6, 128.7, 128.2,

127.9, 122.2, 69.6, 22.9.

2-Amino-2-phenylacetylpropane Hydrochloride, PhCH2-COCMe2-NH3Cl.

To a mixture of 3-methyl-1-phenyl-2-butene (0.7 g, 4.79 mmol) and isopentyl nitrite

(0.77 mL, 5.75 mmol) cooled to 5 C was added dropwise concentrated HCl (0.96 mL,

223

11.5 mmol, 12 N). After the solution was stirred for 15 min, the nitroso chloride dimer

precipitated as a greenish solid. This product was washed with a small portion of warm

acetone, dried in vacuo, and used in the next step without further purification. Yield 0.50

g (49%) of colorless crystals, mp 133-134 C (lit. mp 136-137 C).63

The nitroso dimer

(0.25 g, 1.18 mmol) was mixed with MeOH:EtOH (3 mL:3 mL) and the solution was

stirred for 24 h under under 10 psi of NH3 at 45 C,64

causing eventual dissolution of the

dimer. The solvent was evaporated and the solid was dissolved in 50 mL of 6 M HCl by

heating to 50 C. This solution was extracted with ether, and then the aqueous phase was

made alkaline with sodium carbonate (note: CO2 evolution!). Upon raising the pH to 10,

the color changed from yellow to blue/green. The precipitate was dissolved in ether and

dried over Na2SO4, and the solvent was evaporated. The residue (190 mg) was found by

NMR to be a mixture of oxime and ketone. A solution of 6 N aq HCl (10 mL) was added

and the mixture was stirred for 1.5 h at 50 C. The solvent was evaporated and the

residue dissolved in a small amount of isopropyl alcohol by heating to 50 C. The salt

was precipitated by addition of ether, filtered, and dried in vacuo. Yield 91 mg (36%) of

PhCH2-COCMe2-NH3Cl, mp 134-141 C. 1H NMR (MeOH-d4) 7.23-7.34 (m, 5H),

1.95 (s, 2H), 1.64 (s, 6H). 13

C NMR (MeOH-d4) 207.0, 134.8, 130.9, 129.7, 128.3,

63.4, 43.1, 23.1.

2-tert-Butyl(ONN)azoxy-2-phenylacetylpropane, PhCH2-COCMe2-N=N(O)-Bu-t

(4c), was prepared in the same manner as PhCO-CMe2-N=N(O)-Bu-t. Purification on

silica gel eluting with 4:1 hexane:ethyl acetate (Rf 0.51) yielded 32 mg of clear, oily

product (42% based on PhCH2-CO-CMe2-NH3Cl). 1H NMR (toluene-d8) 7.00-7.21 (m,

224

5H), 3.45 (s, 2H), 1.34 (s, 6H), 1.29 (s, 9H). 13

C NMR (toluene-d8) 203.8, 137.5, 130.3,

128.4, 126.7, 76.1, 69.3, 43.5, 27.8, 21.5.

Azoxy-tert-butane was made according to the method of Freeman15

and fractionally

distilled twice, bp 66 C/40 mm. 1H NMR (C6D6) 1.40 (s, 9H), 1.35 (s, 9H).

13C NMR

(C6D6) 76.39, 57.83, 28.25, 25.69.

tert-Butyl(O,N,N)azoxy-2-propane, i-Pr-N=N(O)-Bu-t,8 was synthesized by a

modified literature procedure.65

Nitroso-tert-butane dimer (217.5 mg, 1.25 mmol) in

absolute EtOH (2.5 mL) was stirred for 3 h at 25 C in the dark. In a separate vessel,

i-PrNHOH·HCl (290 mg, 2.6 mmol) was added in one portion to a solution of KOH (154

mg, 2.75 mmol) in absolute EtOH (2.5 mL). After brief stirring, the ethanolic suspension

of i-PrNHOH was added to a dark blue solution of t-BuNO. The mixture was stirred for

2 h at 25 C and for 16 h at 38 C, after which it was diluted with 1 N aq HCl (5 mL),

extracted with pentane (10 × 3 mL), and dried over Na2SO4. The solvent was removed by

careful bulb-to-bulb distillation at 50 mmHg, cooling the receiver in a dry ice-2-propanol

bath. Since the azoxyalkane product was highly volatile, some loss was unavoidable. The

residue was distilled by reducing the pressure and the product was purified by preparative

HPLC on a silica gel column, eluting with 10:1 pentane:ethyl ether. Bulb-to-bulb

distillation at 50 mmHg removed the solvent, after which the residue was distilled at

0.001 mm to a trap at -196 C. Yield 55 uL. GC analysis of the collected solvent

indicated that it contained approximately 67 uL of i-Pr-N=N(O)Bu-t. 1H NMR (C6D6)

4.32 (septet, 1H, J = 6.4 Hz), 1.36 (s, 9H), 1.15 (d, 6H, J = 6.4 Hz). 13

C NMR (C6D6)

75.5, 51.0, 28.1, 19.5.

225

t-Bu-N=N(O)-Ph (7Z) was made according to Sullivan et al.26

and distilled, bp 44-

62 C / 0.001 mm. Further purification was effected by silica gel chromatography, eluting

with 9:1 hexane:EtOAc, Rf 0.54. The pure compound retains a yellow coloration. 1H

NMR (toluene-d8) 8.13 (m, 2H), 6.90-7.02 (m, 3H), 1.47 (s, 9H). 13

C NMR (toluene-d8)

149.6, 131.4, 128.9, 122.8, 59.2, 26.3. 15

N NMR ( vs MeNO2) -52.1 (N-O), -20.2. The

15N HMBC experiment was optimized for a long-range N-H coupling of 3 Hz.

t-Bu-N=N(O)-Ph (7E): 1H NMR (toluene-d8) 6.92-7.03 (m, 3H), 6.78 (m, 2H),

0.94 (s, 9H). 15

N NMR ( vs MeNO2) - 42.0 (N-O), 1.4.

2-Phenylazo-2-benzoyloxypropane (8b), PhCOO-C(Me)2-N=N-Ph. To freshly

distilled PhNHNH2 (3.85 g, 35.6 mmol) and water (29 mL) at 5 C was added acetic acid

(0.95 mL, 16.6 mmol), then dropwise 5.8 mL of acetone. After the solution was stirred

for 1.5 h, the air-sensitive hydrazone was quickly filtered off, washed twice with ice

water, and dried in vacuo. Yield 90%. NMR (C6D6) 1.04 (s, 3H), 1.75 (s, 3H), 6.30-

6.50 (br, 1H), 6.72-7.30 (m. 5H). To the freshly prepared hydrazone (1.446 g, 9.77 mmol)

in CH2Cl2 (8 mL) at -78 C was added dropwise t-BuOCl (1.219 g, 11.24 mmol). The

mixture was stirred for 2.5 h as the temperature rose to 25 C. The solvent and tert-butyl

alcohol byproduct were evaporated and the residual Ph-N=N-CMe2-Cl was dissolved in

benzene (5 mL). Because neat Ph-N=N-CMe2-Cl decomposes within 30 min at 25 C,

this freshly prepared solution was added dropwise to silver benzoate (1.79 g, 7.80 mmol)

in benzene (10 mL) at 5 C. The heterogeneous mixture was stirred for 1 h at 5 C, then

at 25 C for 3 days. The azoester suspension was filtered through a cotton plug and then

through filter paper and the filtrate was concentrated to deep yellow oil. The crude

226

product was purified on silica gel, eluting with 9:1 hexane/ethyl acetate, Rf 0.47. Yield

1.71 g (81%) of deep yellow solid with mp 48-49 C. The compound crystallized from

hexane after maintaining the solution at -20 C for several weeks. 1H NMR (C6D6) 8.26

(m, 2H), 7.79 (m, 2H), 7.03-7.16 (m, 6H), 1.78 (s, 6H). 13

C NMR (C6D6) 165.3, 152.3,

133.2, 132.2, 131.4, 130.5, 129.5, 128.9, 123.4, 102.6, 25.2.

2-tert-Butylazo-2-benzoyloxypropane, PhCOO-CMe2-N=N-Bu-t (8a). Crude 2-

tert-butylazo-2-chloropropane, t-Bu-N=N-CMe2-Cl,66

was treated with silver benzoate as

in the synthesis of 8b above. The product was purified on silica gel, eluting with 9:1

hexane:ethyl acetate, Rf 0.48. Yield 0.956 g (53% from acetone-tert-butyl hydrazone).

The azo compound crystallized at 25 C over time, yielding ice-like, transparent crystals

that melted at 29-30 C to light beige oil. 1H NMR (C6D6) 8.20-8.22 (m, 2H), 7.02-7.12

(m, 3H), 1.65 (s, 6H), 1.20 (s, 9H). 13

C NMR (C6D6) 164.75, 132.67, 132.02, 130.04,

128.42, 101.32, 67.09, 26.69, 24.59.

2-tert-Butylazo-2-methoxypropane, t-Bu-N=N-C(Me)2-OMe (protiated 14).67

To t-Bu-N=N-CMe2-Cl (0.934 g, 7.29 mmol) prepared as above and cooled to -78 C was

added 2 mL of MeOH.68

In a separate vessel, sodium methoxide (0.45 g, 8.39 mmol) was

dissolved in MeOH (4 mL) and the solution was added to the reaction mixture dropwise

at -78 C. The mixture was stirred at -78 C for 10 min and the temperature was allowed

to increase gradually to 25 C over 3 h. Water and CH2Cl2 were added, the mixture was

shaken, and the aqueous phase was extracted with CH2Cl2 several times. The combined

organic phase was dried over Na2SO4 and the solvent was evaporated. Removal of

227

residual solvent in vacuo at ~ 40 mmHg yielded 0.509 g (44% from t-Bu-NH-N=CMe2)

of volatile liquid. 1H NMR (MeOH-d4) 3.39 (s, 3H), 1.21 (s, 6H), 1.19 (s, 9H).

1.7. Supporting information

Calculated isotropic Fermi contact couplings for 5a, ESR spectra of 5a, UV spectra

of 4a, 6, 8a, 4bZ, 7Z, and 8b, computed structures of 15, 8, 5a, and NMR spectra are

provided in Appendix D.

1.8. References and Notes

1. Engel, P. S.; Duan, S. M.; He, S. L.; Tsvaygboym, K.; Wang, C. R.; Wu, A. Y.;

Ying, Y. M.; Smith, W. B., The free radical chemistry of the azoxy group.

Arkivoc 2003, (12), 89-108.

2. Brough, J. N.; Lythgoe, B.; Waterhouse, P., Macrozamin. IV. Positional and

Geometrical Isomerism in Mixed Aliphatic-Aromatic Azoxy-Compounds.

Journal of the Chemical Society 1954, (12), 4069-4079.

3. Matsumoto, H.; Nagahama, T.; Larson, H. O., Studies on Methylazoxymethanol

Aglycone of Cycasin - a Synthesis of Methylazoxymethyl Acetate. Biochemical

Journal 1965, 95, (2), 13C-14C.

4. Fujiu, M.; Sawairi, S.; Shimada, H.; Takaya, H.; Aoki, Y.; Okuda, T.; Yokose, K.,

Azoxybacilin, a Novel Antifungal Agent Produced by Bacillus-Cereus NR2991 -

Production, Isolation and Structure Elucidation. Journal of Antibiotics 1994, 47,

(7), 833-835.

5. Ohwada, J.; Umeda, I.; Ontsuka, H.; Aoki, Y.; Shimma, N., Synthesis and

Structure-Activity Relationships of a Novel Antifungal Agent, Azoxybacilin.

Chemical & Pharmaceutical Bulletin 1994, 42, (8), 1703-1705.

6. Parry, R. J.; Li, Y.; Lii, F. L., Biosynthesis of Azoxy Compounds - Investigations

of Valanimycin Biosynthesis. Journal of the American Chemical Society 1992,

114, (25), 10062-10064.

7. Engel, P. S., Mechanism of the Thermal and Photochemical Decomposition of

Azoalkanes. Chemical Reviews 1980, 80, (2), 99-150.

228

8. Engel, P. S.; He, S. L.; Wang, C. R.; Duan, S. M.; Smith, W. B., Radical

cyclization and fragmentation of azoxy compounds. Journal of the American

Chemical Society 1999, 121, (27), 6367-6374.

9. Taylor, K. G.; Clark, M. S., Aliphatic Azoxy Compounds. V. Functionalization of

(Z)-Phenylmethyldiazene 1-Oxide. Journal of Organic Chemistry 1976, 41, (7),

1141-1145.

10. Gould, I. R.; Zimmt, M. B.; Turro, N. J.; Baretz, B. H.; Lehr, G. F., Dynamics of

Radical Pair Reactions in Micelles. Journal of the American Chemical Society

1985, 107, (16), 4607-4612.

11. Lewis, F. D.; Magyar, J. G., Photoreduction and Alpha Cleavage of Aryl Alkyl

Ketones. Journal of Organic Chemistry 1972, 37, (13), 2102-2107.