Studies of gas-phase ion/molecule reacations in relation to a proposed ionic mechanism of soot formation


Material Information

Studies of gas-phase ion/molecule reacations in relation to a proposed ionic mechanism of soot formation
Physical Description:
viii, 196 leaves : ill. ; 28 cm.
Öztürk, Feza
Publication Date:


Subjects / Keywords:
Soot -- Environmental aspects   ( lcsh )
Combustion deposits in engines   ( lcsh )
Ionization of gases   ( lcsh )
Hydrocarbons -- Combustion   ( lcsh )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1988.
Includes bibliographical references.
Statement of Responsibility:
by Feza Öztürk.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001123837
notis - AFM0889
oclc - 20071107
sobekcm - AA00004821_00001
System ID:

Full Text







Digitized by the Internet Archive
in 2011 with funding from
University of Florida, George A. Smathers Libraries with support from LYRASIS and the Sloan Foundation


This work is the final product of the efforts and

encouragement of many. First, I must thank my colleagues,

Bryan Hearn, Dr. Cliff Watson and Dr. Steve Bach, in the

Eyler group, including Dr. GOkhan Baykut and Dr. Mehdi Moini

who are no longer with the group, for their help and

suggestions. Each one deserves particular thanks for their

special efforts and friendship throughout these last years.

I have greatly benefitted particularly from their technical

experience and skills whenever questions and difficulties

arose in the lab.

Next, I wish to acknowledge my research advisor, Dr.

John Eyler, whose guidance and encouragement has enabled me

to experience the intellectual satisfaction and enjoyment of

scientific research. With his research funds, I had the

opportunity to spend my time exclusively doing research and

attend the annual meetings of the American Society of Mass

Spectrometry which provided the best setting for scientific

communication. I also would like to mention his editing

skills which have always assisted me and had a major role in

putting this manuscript to its final form.


I would like to acknowledge Dr. Floyd Wiseman from the

Environics Division of Tyndall Air Force Base for his

interest and efforts in this research project. His

contribution to the work by the kinetic modeling study

deserves an important credit for providing a better

understanding of experimental results.

Special thanks are extended to Dr. William Weltner, Dr.

Robert Hanrahan, Dr. Merle Battiste, Dr. Willis Person and

Dr. Charles Proctor for serving as committee members. I

particularly wish to thank Dr. Calvin VanderWerf, Dr. Robert

Hanrahan, Dr. Kathryn Williams, and Dr. William Weltner for

their efforts and willingness to be supportive in every way

throughout these long years.

Another special person deserving particular thanks is

my friend, Zekiye Onsan. She has been an endless source of

encouragement and never failed to give her help whenever

needed. I also wish to thank Taghl Alizadeh Yekani for

sharing the difficult times with me with patience and

sincerity in my earlier years of study.

Finally, there is my five-year old, Ellf. She made it

worhwhile by just being there for me.



ACKNOWLEDGEMENTS ................................. ii

ABSTRACT ........................................ vii


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

2 INSTRUMENTATION ............................. 15

3 ION/MOLECULE REACTIONS ..................... 29

Theory of Ion/Molecule Reactions ........... 29
Experimental Study of Ion/Molecule Reactions
Using FTICR ....... ........................ 38


Introduction ................................ 45
Experimental ................................. 48
Results .................................... 50
Discussion ..................................... 64


Introduction ................................ 72
Experimental ..... ............................ 73
Results ....................................... 74
Discussion ................................. 93

AND DIACETYLENE ............................ 101

Introduction ............................... 101
Experimental ................................. 104
Results ...................................... 107
Discussion ................................. .. 123


Introduction ............................... 132
Experimental ............................... 136
Results .................................... 138
Discussion .................................. 148



SPECTROMETRY ............................... 167


BIBLIOGRAPHY .............................. ....... 187

BIOGRAPHICAL SKETCH ................................ 196


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




August, 1988

Chairman: John R. Eyler
Major Department: Chemistry

The reactions of small hydrocarbon ions such as C3H3+,

C5H3+, C5H5+, and C7H7+ with acetylene and diacetylene have

been investigated using Fourier Transform Ion Cyclotron

Resonance (FTICR) mass spectrometry to provide qualitative

and quantitative information about a proposed ionic

mechanism of soot formation. Ion/molecule reaction pathways

and rate coefficients have been determined for several

isomers of each of the ions listed above, formed from a

variety of precursors, and reacting with precursors,

acetylene and diacetylene. Further understanding of

reaction mechanisms of some of the reactions proposed

experimentally was obtained from kinetic modeling studies.

All of the small hydrocarbon ions studied reacted with

diacetylene extensively while almost no production of larger


ions was observed with acetylene. Linear C3H3 was formed

by charge transfer ionization of propargyl Iodide using Xe+,

and was found to isomerize to the cyclic form of C3H3+ in

reactions with both acetylene and diacetylene. The

isomerization was shown to take place via a long-lived

C5H5+* complex by isotope exchange reactions between linear

C3H3+ and deuterated acetylene. The reaction rate

coefficients for the reaction of C3H3+ with deuterated

acetylene and diacetylene were (4.5 + 1.9) x 10-10 cm3/s and

(1.4 + 0.7) x 10-9 cm3/s, respectively.

While different structures could be attributed to

C3H3 C5H3 and C7H7 Ions produced from different
precursors on the basis of reactivity, this could not be

done for C5H5 ions, whose reactivity with acetylene and

diacetylene was similar within experimental error regardless

of precursor. Presence of two structural isomers of C5H3

was determined using different precursors and several

structures for these isomers were proposed. The reactive

isomer was observed to react with diacetylene with a rate

constant of (5.6 + 1.7) x 10-10 cm3/s. Rate coefficients

for the ion/molecule reactions of C7H7+ were found to be

different when it was formed from various precursors, which

implied the Involvement of different C7H7 structures. All

experimental results are discussed in relation to

theoretical work involving structures of the ions studied

and a proposed ionic route to soot formation.



Soot is a combustion product which has undesirable

effects both on human health and on the efficiency of fuel-

powered engines due to increased heat transfer to critical

engine components. Considering the fact that the use of

alternative fuels such as shale or coal-derived gas in the

near future would lead to increased soot emissions, search

for an effective way to reduce soot has been going on for a

long time. The design and development of advanced engines

to achieve soot suppression requires a sufficient

fundamental understanding of the mechanisms governing soot

formation. Soot formation in hydrocarbon flames involves a

succession of fast processes that occur within a few

milliseconds during the combustion of the hydrocarbon.

Although many investigations have been undertaken in

relation to the process of soot formation in fuel-rich

hydrocarbon flames, understanding of this process is still

limited. Experimental techniques such as mass spectrometry

(Calcote, 1963; Michaud et al., 1981), laser-light

scattering (Kent et al., 1981; D'Alessio et al., 1977),

laser induced ionization (Smith and Mallard, 1981), and

laser induced fluorescence (DiLorenzo et al., 1981) have


been used to study the molecular species and particles in


Soot collected from flames consists of chainlike

aggregates of spherical units having diameters of 10-50 nm

with a carbon/hydrogen atom ratio in the range 8:1-12:1

(Calcote, 1981). Three distinct steps of soot formation

have been recognized over many years of research (Calcote,


1. Nucleation- a phase change from molecular species where

chemical reactions dominate to incipient soot particles

where physical processes dominate;

2. Growth to spherical particles of 10-50 nm in diameter;


3. Aggregation or agglomeration of the spherical units to

form chains.

Among these, the first step, nucleation, is the least

understood in the process. Although many diverse theories

have been put forward to explain the nucleation process,

only two of these have received quantitative support:

1. Mechanisms involving neutral and free radical species;


2. Ionic mechanisms.

Neutral and Free Radical Mechanisms. A large number of

neutral and free radical mechanisms have been proposed to

describe the transition from molecular species to soot


(Palmer and Cullis, 1965; Lahaye and Prado, 1978; Gaydon and

Wolfhard, 1979).

One of the proposed paths involves polyacetylenes

where acetylene reacts to form higher species (Bonne et al.,


C4H2 C6H2 C8H2

j-H j-H I-H
C2H2 1 C2H2 C2H2 I
C2H2 -----> C4H3 (+H) -----> C6H3 (+H2) -----> C8H3 (+H2)

It was suggested that the continuation of this series

leads to larger radicals which further react with each other

and with higher polyacetylenes forming even larger

molecules. However, because this mechanism leads to

formation of a giant chain molecule instead of a polycyclic

carbon structure, it was not accepted as being responsible

for soot formation (Homann, 1967). To account for the

problem of ring formation, Homann and Wagner (1967, 1968)

suggested that radicals such as C2H attack the

polyacetylenic species forming side chains which lead to

branched polymerization, and presumably ring closures. The

experimental support of this scheme comes from the

observation that in the oxidation region of the flame, the

concentration of polyacetylenes decreases while that of the

polycyclic aromatics increases. But the growth of

polyacetylenes was considered too slow by Calcote (1981), to


account for the rapid formation of soot particles.

Furthermore, as pointed out by Bonne et al. (1965), Thomas

(1965) and Cullis (1976), the rearrangement of a

polyacetylene to an aromatic graphite-like structure would

also be a slow process.

Another hypothesis proposed by Glassman (1980)

emphasized the importance of strongly conjugated molecules

such as butadiene in the formation of ring structures with

side chains to which butadiene-type molecules keep adding.

This mechanism did not receive much support on the basis of

concentration profiles in which the butadiene concentration

is observed to fall very rapidly even before the initiation

of soot formation (Calcote,1981).

Some models considered C2, C2H and C3 radicals as

initial nuclei for soot formation which undergo condensation

reactions with each other or with acetylene (Jensen, 1974;

Smith, 1940; Carbannes, 1956). The main objection to these

mechanisms based on C2 and C3 condensation reactions was the

presence of a very strong C2 radiation in rich cyanogen-

oxygen flames which do not form soot.

Finally, direct condensation of aromatic rings was

proposed as being the most favored thermodynamic path to

soot formation (Bonne et al., 1965; Graham et al., 1975;

Stein, 1978). However, by a detailed analysis of the

concentration and the flux profiles of the species, Bittner

and Howard (1981a, 1981b) were able to show that soot is


produced via nonaromatic hydrocarbon intermediates rather

than via intact aromatic rings. But after a consideration

of many species, only a few intermediates were found to

favor the rearrangement to a ring structure over other

decomposition channels.

To conclude the discussion above, it is clear that all

the proposed neutral species mechanisms for soot formation

have serious problems with the nucleation step, and thus

none can be considered completely satisfactory in

understanding the soot formation process. Particularly it

is difficult in any of the proposed neutral schemes to

produce small cyclic and polycyclic species with a

sufficiently high rate to model the quite rapid soot

formation seen in flames. Recently, however, a

neutral/radical soot formation model has been proposed which

contains many steps and incorporates quite recently obtained

shock tube kinetic data (Frenklach et al., 1985). This

model is capable of yielding small ring compounds at a

higher rate than previously suggested radical mechanisms,

although there is some question as to the correct rate

constant for one or two of the crucial steps in formation of

the first ring.

Ionic Mechanisms. A number of experimental

observations in flames have led to a consideration of the


importance of ions In soot formation. Such observations in

flames include (Olson and Calcote, 1981a; Calcote, 1981):

1. Presence of a unit charge on soot particles;

2. A positive correlation between the growth of large ions

and the appearance of soot;

3. The increase both in the concentration of large ions and

the onset of soot formation by the addition of electrophilic


4. Identical location for the peak concentrations of

molecular ions and soot precursors;

5. Production of larger particles by an increase in the

residence time of the positive ions; and

6. High ion concentrations in hydrocarbon pyrolysis.

As a consequence of the above observations, several

ionic mechanisms have been suggested (for references, see

Calcote, 1981). The postulated general growth of soot

particles from primary ions is shown in Figure 1.1. Calcote

(1962) proposed the chemlionization reactions (1.1) and

(1.2) for the formation of primary ions:

CH + 0 ----> CHO+ + e~ (1.1)

CH* + C2H2 ----> C3H3+ + e- (1.2)

The C3H3+ ion is the dominant ion observed in fuel-rich

flames (Knewstubb and Sugden, 1959; Goodings et al., 1979;

Olson and Calcote, 1981b; Michaud et al., 1981) and its

Primary flame Neutral flame
ions species

C 2H2 5H3
C3H + C4H2 -.CH5, etc.
etc. C7H3
C5H;+ etc.

CzH ions

Larger ions

CXHy species
Large aromatic


RECOMBINATION ncipient soot particles


e e Small particles


Neutral and small
charged particles



-;.. '*^ 4.. .A:;t r ; ;.:r ,a/.< ES 2

Figure 1-1. Growth of Soot Particles from Primary Molecular


concentration falls very rapidly at the critical equivalence

ratio of soot formation where the concentration of the

larger ions starts to increase (see Figure 1.2)

(Calcote, 1981). Unfortunately, the mechanism of C3H3+

formation is still not clear. An alternative mechanism

(Calcote, 1972) for C3H3+ formation is shown by reactions

(1.3) and (1.4):

CHO+ + C2H20 ----> C2H30+ + CO (1.3)

C2H30 + C2H2 ----> C3H3 + CH20 (1.4)

Reaction (1.5) is postulated (Calcote, 1972) to account for

the dominant H30+ observed in near stoichiometric and lean


CHO + H20 ----> H30+ + CO (1.5)

The C2H3+ on was also considered as a possible nucleating

center for soot formation in flames and its formation is

shown by reaction (1.6) as proposed by Vinckler et al.


CHO+ + C2H2 ----> C2H3+ + CO (1.6)

The primary ions mentioned above are proposed to react

with CxHy molecules by fast ion/molecule reactions (Calcote,



10-11 C13.

1012- t
1.6 2.0 2.4 2.8

Figure 1.2. Effect of Equivalence Ratio on Peak Ion
Currents for 2.0 kPa Acetylene-Oxygen Flames. The shaded
area indicates the minimum equivalence ratio for soot

Reprinted by permission of Elsevier Science Publishing Co.,
Inc. from the article by H. Calcote (1981). Copyright 1981
by the Combustion Institute.

Combustion Institute.

1981; Olson and Calcote, 1981a). The typical set of
reactions for C3H3+ are shown below:

C3H3 + C2H2 ----> C5H3 + H2 (1.7)
C3H3. + C2H2 ----> C5H5 (1.8)
C3H3 + C2H ----> C5H3 + H (1.9)
C3H3+ + C4H2 ----> C5H3 + C2H2 (1.10)
C3H3 + C4H2 ----> C7H5 (1.11)

The next step is the further addition of neutral
building blocks, acetylene and diacetylene to the product
ions producing larger ions with increased C:H ratio.
Following are a few reactions representative of this step.

C5H3+ + C2H2 ----> C7H5 (1.12)
C5H3 + C4H2 ----> C9H5 (1.13)
+ 4
C7H5 + C4H2 ....> C9H5 + C2H2 (1.14)

This series of reactions is suggested to lead to the
formation of polycyclic structures such as C13H9 and

C19"11 4-

PQ) W0@
igj LlO




The problem of rearrangement of linear structures to

polycyclic structures encountered in the neutral mechanism

is overcome in the Ionic mechanism because of the general

observation that gaseous Ions usually rearrange upon

formation to their most stable structure (Calcote, 1981).

Michaud et al. (1981) have made an alternate suggestion

that direct reactions of C3H3+ with aromatic neutrals such

as benzene, toluene, naphthalene, methylnaphthalenes, and

indene may be more important in forming polycyclic ions than

sequential reactions involving acetylene and diacetylene.

In fact, recent ion cyclotron resonance (ICR) studies of

reactions of C3H3+ with aromatic neutrals showed (Baykut et

al., 1986) that these reactions are fast enough to be

considered as possible bypass channels in Ionic soot

formation pathways.

The main objection to the Ionic mechanism is the lower

concentration of ions in flames compared to that of neutral

species. Figure 1.3 (Calcote, 1981) shows the concentration

profiles of primary neutrals and Ions observed in flames as

a function of distance above the burner. One of the most

important features of the figure is that the soot

concentration is lower than that of the large positive Ions

which implies that there are probably enough of these large

ions to produce the observed levels of soot. Several orders

of magnitude higher concentrations of neutral species

compared to those of ions in flames can be explained as a



1016 H

10 s CsH

1014 CgH2 -

101 02


IONS. 300-1000
1010 -

10 '
0 1 2 3 4 5 6 7

Figure 1.3. Number Densities of Neutral and Ionic Species
Found within Flames.

Reprinted by permission of Elsevier Publishing Co., Inc.
from the article by H. Calcote (1981). Copyright 1981 by
the Combustion Institute.


requirement for the ion/molecule reactions to proceed fast

enough to account for the rapid formation of soot.

In this study, the reactions of primary Ions with flame

neutrals have been investigated to identify different

reactant ion structures, reaction products and mechanisms.

One of the goals of this study was to determine how fast

these ion/molecule reactions proceed under laboratory

conditions, which hopefully leads to a better understanding

of the complex phenomena occurring under flame conditions.

Fourier Transform Ion Cyclotron Resonance (FTICR) mass

spectrometry, used in this study, is well suited to

investigate ion/molecule reactions due to its inherent

qualities which differ from other mass spectrometers. These

qualities include very clean isolation of the mass of

interest from a complicated reaction mixture and accurate

rate constant measurements along with reaction mechanism and

branching ratio studies. The basics of this technique are

discussed in Chapter 2 along with an explanation of the

determination of some experimental pressure measurement

parameters such as Baratron factors and the system factor

inherent to the instrument. The use of FTICR mass

spectrometry for studying ion/molecule reactions is

discussed in Chapter 3 along with some theoretical models

developed for ion-molecule reactions. The main body of the

experimental work, reported in Chapters 4-7, deals with

proposed nucleation steps of the ionic mechanism of soot


formation. An overall conclusion of the experimental

results is included in Chapter 8.


Fourier Transform Ion Cyclotron Resonance (FTICR) Mass

Spectrometry was introduced in the mid-1970s by Comisarow

and Marshall (1974a, 1974b, 1974c, 1975). It involves the

application of Fourier multiplex concepts (Griffiths, 1978;

Marshall, 1982) to ICR mass spectrometry. High speed, high

sensitivity, computerization, ultra-high mass resolution,

and wide mass range are some of the advantages of Fourier

Transform over continuous-wave spectrometry (Marshall,

1985). Basic principles and the vast applications of this

relative newcomer to mass spectrometry are summarized in

several recent review articles (Johlman et al., 1983; Gross

and Rempel, 1984; Comisarow, 1985; Marshall, 1985; Baykut

and Eyler, 1986; Laude et al.,1986). In this chapter, the

basic theory of operation of ICR which led to the concepts

and development of the FTICR technique will first be

discussed. Following that is a description of the

instrument along with typical experimental parameters and

the general sequence of operation.

Theory. An ion of charge, q, moving at velocity, v

when put in a uniform magnetic field, B, is subject to the

Lorentz force,

F ma = q (v x B) (2.1)

which acts perpendicular to the direction of ion motion.

Under the influence of this force, the ion follows a helical

path (Lawrence and Livingston, 1932) which, when projected

into a plane perpendicular to the magnetic field gives a

circle with a radius, r proportional to the velocity of the

ion, as shown in Figure 2.1. Substitution of the

centripetal acceleration in terms of angular frequency, *,

into equation (2.1) gives

IFI = mial = mv2/r = mro2 = qorB (2.2)
a = qB/m (nks units) (2.3)

which relates the ion's characteristic cyclotron frequency

to its q/m (charge/mass). Equation (2.3) is called the

cyclotron equation and provides the basic principle of mass

measurement in terms of cyclotron frequency in the ICR mass

spectrometer. For a magnetic field of 3.0 Tesla, singly

charged ions with masses in the range of 18-5000 a.m.u. have

cyclotron frequencies in the radiofrequency range (10 kHz-

2.6 MHz), within which frequency can be measured with high


Operation. A schematic representation of the

commercial Nicolet FTMS-1000 mass spectrometer with a

superconducting solenoid magnet (3 Tesla) is shown in Figure

Figure 2.1. Motion of an Ion with Initial Velocity v in a
Magnetic Field B.


2.2. All data acquisition and processing and the selection

of various experimental parameters (except emission current)

are under computer control. Figure 2.3 shows a simplified

block diagram of an FTICR mass spectrometer. The simplicity

of operation results from the ability to form, excite, react

and detect ions in the same 1-inch cubic stainless steel

cell shown in Figure 2.4. This analyzer cell is located in

the center of a vacuum chamber which passes through the bore

of the magnet. The background pressure of 109 torr is

achieved by the use of a four-inch diameter oil diffusion

pump and by baking out the system at 2506C for several hours

each night.

An inlet system which is evacuated with a three-inch

diameter diffusion pump permits the introduction of gas and

liquid samples into the high vacuum through precision leak

valves. The liquid samples are used after multiple freeze-

pump-thaw cycles to remove non-condensible gases. A solids

insertion probe is used for the introduction of solid

samples into the vacuum chamber and can be heated up to

250C to promote sublimation of the solid.

Ion formation is typically achieved in 5 ms by

collisions of electrons accelerated to 10-70 eV with

neutrals leaked into the vacuum chamber. The number of Ions

is controlled by monitoring the emission current, measured

by means of an electron collector located behind the trap

plate opposite to the filament. When Ions are formed, they


I a 1

uu I 0 e

t s
a II I.
oIiJ I O

L.. ,-I I. I

Si /I
I(CM 4.)

-----------n---1-0I I- S

@ CJ. .

Figure 2.3. Simplified Block Diagram of an FTICR Mass



trap ing
plate L

Figure 2.4. Cubic FTICR Cell.


start cyclotron motion at their characteristic frequency

with a random phase. The trapping plates shown in Figure

2.4 constrain the ion motion in the direction parallel to

magnetic field lines. Positive and negative ions are trapped

by the application of, typically, positive and negative 1

volt potentials to the trapping plates, respectively.

To detect the ions, a fast sweep of oscillating voltage

in the radiofrequency range is applied to the transmitter

plates shown in Figure 2.4. Ions absorb energy at their

cyclotron frequency and are driven into coherent motion with

a larger radius, thus inducing an image current oscillating

at the frequency of their cyclotron motion on the receiver

plates (see Figure 2.5). The image current is amplified,

digitized and stored in the computer's memory as a time-

domain signal which contains the superimposed frequency

information of all the ions with different masses in the

analyzer cell. A fast Fourier Transform (Cooley and Tukey,

1965) is applied to the time-domain data to produce the

frequency-domain spectrum which can be plotted in terms of

the ion masses. Figure 2.6a shows a time-domain spectrum

of two superimposed sine waves of frequencies 1l and e2.

The decay of the signal results from the dephasing of ions

due to collisions with the neutral molecules. Fourier

transformation of this spectrum gives the mass-domain

spectrum shown in Figure 2.6b. In this procedure, the

lowest mass detected is limited by the sampling frequency of

o****** ...._.._ 1 0 ...

(a) (b) (c) (d)

Figure 2.5. Excitation and Detection of Ion Motion.
(a) Ion motion in magnetic field only, (b) motion during
radio-frequency pulse applied to the transmitter plates,
(c) larger radius of ion motion after excitation, and
(d) generation of an image current in the receiver plates by
a rotating Ion "clump".








O .



1O 20



30 40 50 60


70 80 O
70 80 90

Figure 2.6. Ion Detection by Fourier Transform. (a) Time-
domain spectrum (Abscissa is time; ordinate is ion signal).
(b) Frequency-Domain Spectrum after Fast Fourier Transform
of the time-domain data (Abscissa is frequency; ordinate is
ion signal).


i ; *,~-.


the analog-to-digital converter, which should be twice the

signal frequency according to the Nyquist criterion. Since

the highest cyclotron frequency detected is one-half the ADC

frequency, for a magnetic field of 3 Tesla used in this

work, the singly charged ion mass corresponding to this

frequency is 18 a.m.u., which sets the low mass limit.

A typical experimental sequence is shown in Figure 2.7.

Ions can be manipulated in various ways between the

formation and the excitation pulses. Ion ejection is used

to clean the cell of all ions except one of interest. Ions

are ejected by the same principle as excitation. The only

difference is that a higher amplitute of excitation voltage

is used. In this way ions achieve large enough radii to

strike the cell walls and be neutralized. Application of

either a single ejection pulse at a single cyclotron

frequency or a swept pulse covering a range of frequencies

is determined according to the user's purpose.

After selection of the ion of interest, different kinds

of ion chemistry can be performed. Among these, ion/molecule

reactions, collisionally activated dissociation, and

photodissociation are the most commonly used processes in

characterization of ion structures. Since the number of

stages of ion ejection and manipulation is not limited by

hardware or software, (MS)n (multiple stages of mass

selection and manipulation) experiments can be performed by

repeating the same processes for the various generations of

, 0

r. --------


x I 0
0 I
I 0
.- E

I e0
O )

0) i -.4

Q) c Q)
*- 0, x>

o ---


offspring ions. Laser desorption of solid samples and laser

ionization are other techniques used in conjuction with


Experimental Parameters. In this study, the FTICR

technique was used mainly to determine ion/molecule reaction

rates and mechanisms. Ion formation was accomplished by

dissociative ionization using various charge transfer agents

instead of electron impact ionization, in order to form the

ions with well-defined internal energies. Reaction pathways

were delineated by using the ejection capabilities of FTICR

which make it possible to eject one Ion from a complicated

reaction mixture to determine its contribution to the mass

spectrum of all the other ions.

Reaction rate coefficients were determined by

monitoring the intensity of the reactant ion as a function

of time (typically for at least 2s) after ejection of all

other Ions from the analyzer cell. Neutral gas pressures

were measured with an ionizing gauge. Ionization gauge

readings were then corrected by constructing calibration

curves of ionization gauge vs. capacitance manometer (MKS-

Baratron) readings in the 1 x 10-6 to 1 x 10-4 Torr range.

In order to correct for the fact that the ionization gauge

and capacitance manometer were located at different points

on the vacuum system, somewhat removed from the FTICR

analyzer cell, a correction factor was required. This was


obtained by determining the rate coefficient of a well-

studied reaction (C2H4+ + C2H4 ----> products), where the

ionization gauge pressure readings were corrected by using

the capacitance manometer. This experimentally determined

rate coefficient was then compared with the average of

published values, kay = (1.0 + 0.3) x 10- cm3/s (Herod and

Harrison, 1970; Gross and Norbeck, 1971; Sieck and Ausloos,

1972; Warneck, 1972; Le Breton et al., 1975) and the ratio

of the measured value to the published, which was 0.3 + 0.1,

was used as a correction factor. This factor was used in

calculating the absolute rate coefficients reported in this

work. The large uncertainties (95% confidence limits)

reported for the correction factor and rate coefficients

calculated using it are primarily due to the wide range of

reported values for the C2H4+ + C2H4 reaction used to

determine the correction factor. The even more widely

studied "standard" calibration reaction of CH4 with CH4

could not be employed because of the lowest accessible mass

limit in the FTICR instrument mentioned earlier. All

calculations of rate coefficients and 95% confidence limits

were performed with a menu-driven Fortran computer program

(given in Appendix I).


In this chapter the various theoretical models which

have been developed for ion/molecule reactions will first be

discussed. Then follows a description of practical aspects

of studying such reactions by the use of FTICR mass


Theory of Ion/Molecule Reactions

A number of both classical and statistical ion/molecule

collision theories have been introduced during the last two

decades to provide an adequate model explaining the

experimental observations of ion/molecule reactions.

Classical treatments mainly include the pure polarization

(ion-induced dipole) theory first developed by Langevin

(1905) and the various ion-dipole theories.

A. Pure Polarization Theory. Detailed discussions of

this model can be found in a number of articles (Gioumousis

and Stevenson, 1958; McDaniel, 1964; Futrell and Tiernan,

1968; Henglein, 1970). Langevin's model assumes that the

neutral has no permanent dipole moment, and that both the

ion and the neutral molecule are point particles with no

internal energy. The classical charge-induced dipole

potential at an ion-molecule separation r is

V(r) = aq2/2r4 (3.1)

where q is the charge on the ion and a is the polarizability

of the neutral. Since the energy of relative rotation,

Erot(r) of the particles is associated with an outwardly

directed centrifugal force, the effective potential of the

ion/molecule system can be given by

Veff(r) = V(r) + Erot(r) (3.2)


Veff(r) = -(q2a/2r4) + (L2/2pr2) (3.3)

where L is the classical orbiting angular momentum of the

two particles and p is the reduced mass. The total relative

energy of the system is a sum of the translational energy

and the effective potential energy.

Er = Etrans(r) + Veff(r) (3.4)

Figure 3.1 shows a plot of Veff(r) versus r at constant Er

for three different values of the impact parameter, b. When

b=O, since there is no centrifugal potential energy, the

effective potential is attractive at all ion/molecule

separations leading to a collision between two particles.

For b > 0, there is a critical value of the impact


-- i i -- i i -- -



0 2 4 6 8 10 12 14 16
-4 -


r/ cm

Figure 3.1. Plot of Veff vs r from Equation (3.3) for N2
Colliding with N2.


paramater, bc at which the particles orbit around the

scattering center with a constant radius, rc. At this

ion/molecule separation, there is no contribution from the

attractive potential and Veff is equal to centrifugal energy

which creates a "centrifugal barrier" to a capture

collision. For all b < bc, a capture collision occurs,

whereas it is precluded by the centrifugal barrier for all

b > bc.

Capture cross section is defined as the area of a

circle with radius bc perpendicular to the line of collision

at infinite ion/molecule separation and a capture collision

occurs for all the ions that pass through the circle when

approaching the neutral molecule. Capture cross section can

be derived for a given relative velocity in terms of the

charge of ion, q, polarizability of the neutral, a, and the

reduced mass, p as shown in equation (3.5).

Oc(v) 2xq(a/p)1/2/v (3.5)

Thus the collision rate constant is given by

kc = vac = 2xq(a/p)1/2 (3.6)

The rate coefficients obtained from this expression are

generally good for some simple low energy ion/molecule


reactions but underestimate the rate constants of most

ion/polar molecule collisions.

B. Ion/Dipole Theory. Several different models were

proposed to describe the theory of ion/dipole collisions.

Among these are "The Locked Dipole Approximation" (Moran and

Hamill, 1963), "The Frozen Rotor Approximation" (Dugan and

Magee, 1966), "Ion/Dipole Trajectory Calculations" (Dugan

and Magee, 1967), "The Average Dipole Orientation (ADO)

Theory" (Bowers and Laudenslager, 1972), "Barker-Ridge (BR)

Model" (Barker and Ridge, 1976) and "The Average Dipole

Orientation Theory with Conservation of Angular Momentum

(AADO)" (Su et al., 1978).

Both the "Locked Dipole" and "The Frozen Rotor"

approximations assume that the polar molecule is fixed at

angle 9 (0 = 0 for the "Locked Dipole" model) with respect

to the line of centers of the collision. The resulting rate

constants provide upper limits to the ion/dipole capture

collisions and overestimate the dipole effect (Su and

Bowers, 1973).

Ion/Dipole Trajectory Calculations provide numerical

solutions for the equations of motion for the collision of

an ion with a rotating polar molecule. As a result of these

calculations, the numerical capture cross section was found

to depend on relative translational energy. When compared

with experimental results, this numerical approach seems to


be useful in the investigation of nonreactive ion/dipole

phenomena such as energy transfer, but does not give a good

estimation of capture collision rates.

ADO theory assumes that there exists an overall average

orientation of the dipole with respect to the ion and the

extent of locking of the dipole increases as the dipole

moment of the molecule increases. Su and Bowers (1973) have

parametrized the ADO theory to calculate the ADO capture

rate constants. The results indicate that the ADO theory

gives the best prediction of rate constants in most cases

with respect to the other theories.

BR theory uses a simple statistical approach to study

the ion/dipole collisions and is based on the Langevin

model. BR results overestimate the dipole effect by almost

a factor of 2 for charge transfer and proton transfer as

well as momentum transfer reactions.

Su et al. (1978) modified the ADO theory by the

application of the conservation of angular momentum with the

assumption that there is no net angular momentum transfer

between the rotating molecule and the ion/molecule orbital

motion. This modified version of ADO theory is termed the

AADO theory. The capture rate constants obtained from AADO

model are larger than the ADO predictions, providing better

agreement between theory and experiment.


C. Ion/Quadrupole Theory. This theory makes

theoretical predictions of ion/quadrupole interactions for

the molecules having D, symmetries. Similar to ADO theory,

the average quadrupole orientation theory, AQO, was

developed by Su and Bowers (1975). This model predicts

significant quadrupole effects for molecules with high

quadrupole moments and improves the agreement between theory

and experiment.

Statistical Treatments of ion/molecule reactions

include the statistical theory and the orbiting transition

state theory. Statistical Theory is divided into three

categories according to the type of reaction dynamics as

shown in Figure 3.2.

A. Transition State Theory. This theory is developed

for direct bimolecular reactions (Pelzer and Wigner, 1932;

Evans and Polanyi, 1935; Eyring, 1935). This method was

later refined by Pechukas and McLafferty (1972) and by

Miller (1974). The microcanonical rate constant is given by

k(E) = Wt(E-E0)/hp(E) (3.7)

where Wt(E-E0) s1 the total sum of states vibrationall and

rotational) of the transition state complex with nonfixed

energy less than or equal to E-EO, where E is the total

A I - - -

(o) (b)
---i --------
A+B -- ............-- l-4 ---



Figure 3.2. Schematic Potential Surfaces for Various
Ion/Molecule Reactions. (a) Direct bimolecular, (b)
Unimolecular, and (c) Complex formation reactions.


system energy, and p(E) is the total density of states of

the reactants. For a given potential surface, transition

state theory provides a rigorous upper limit to the true

rate constant.

B. Unimolecular Theory. This theory involves two

approaches, RRKM and QET (quasi-equilibrium theory) which

arrive at similar results for the unimolecular reaction of

an isolated energized molecule. These methods were

developed independently by Marcus and Rice (1951) using the

assumptions of transition state theory and by Rosenstock et

al. (1952) using the quasi-equilibrium hypothesis.

Experimental evidence suggests that most unimolecular

systems appear to behave according to the quasi-equilibrium

hypothesis which assumes a random distribution of excess

energy independent of the pattern of initial energization

prior to unimolecular decomposition.

C. Complex Formation Theory. This theory involves the

statistical treatment of reactions which form a long-lived

complex assuming quasi-equilibrium for the intermediate

complex (Keck, 1958; Light, 1967). Thus, the rate constant

for decomposition to each available channel is assumed to be

proportional to the flux through the transition state for

that channel.


The Orbiting Transition State Theory is developed to

overcome the problem of determining the energy distribution

of reactants that leads to the formation of a given

transition state complex. In this method the transition

state is located at the maximum of the long range effective

potential where the reactants retain their original

identity. The orbiting rate constants for both bimolecular

reactions (Miller, 1976; Light, 1967), and unimolecular

reactions (Chesnavich and Bowers, 1977) can be obtained as a

function of translational, rotational, and vibrational

energy of the reacting system. Comparison with experiment

in general reveals that orbiting unimolecular rate constants

are unrealistically high while the orbiting bimolecular rate

constants are similar to experimental results at thermal

energies. This model also predicts the probability of a

given translational energy release (Chesnavich and Bowers,

1976) which in most cases is in good agreement with

experimental results for reactions with no reverse

activation energy.

Experimental Study of Ion/Molecule Reactions Using FTICR

Rate Coefficients. ICR techniques cannot be used for

experimental tests of statistical theories since the

application of statistical models requires a detailed

knowledge of energy states of the system, which cannot be


obtained from ICR. On the other hand, theoretical

bimolecular rate constants using Langevin's model are

usually found to agree with the experimental results

obtained for low energy ion/molecule reactions using ICR

technique. Reactant ions with energies close to thermal can

be prepared by charge transfer ionization of a precursor

molecule followed by collisional relaxation of the ions by

inert gas molecules. Thus, this method provides the

determination of thermal bimolecular rate constants which

can be used to test the basic assumptions of the Pure

Polarization method.

The first step in the experimental determination of

rate coefficients by FTICR involves the ejection of all ions

except the one of interest from the analyzer cell following

charge transfer ionization. Next the decay of the ion

signal is monitored as a function of time long enough to

permit the ion of interest to react away completely.

Method of Data Analysis. The ion/molecule reactions

studied in FTICR most often involve bimolecular collisions

and follow second order kinetics. The rate expression for a

simple second order reaction is given by

-d[I]/dt = k2[l][R] (3.8)

where I is the ion and R is the reactant neutral. Since the

ion number density created in the FTICR instrument is at


least three orders of magnitude lower than the neutral

number density, a "pseudo-first order" rate coefficient,

ki can be substituted for the term k21R] in equation (3.8).

-d[I]/dt = k1 [I] (3.9)

After integration, the rate expression giving the number

density of the reactant at any given time is obtained:

[I](t) = [I](0) exp(-ki t) (3.10)

where [I](0) is the initial number of ions present. Thus,

ki can be determined from the slope of the simple In(ion

signal) vs time plot. The true rate coefficient, k2 can

then be obtained from k1 by dividing ki by the number

density of the reactant neutral. The number density of the

neutral is determined using the expression, (n/V) P/RT,

where V and T are the volume and the temperature of the

analyzer cell respectively, and P is the pressure of the

reactant neutral measured by a capacitance manometer.

In an experimental situation, there is often a second

neutral present in the cell, which is the precursor neutral

of the ion of interest. Thus, the decay of the ion signal

includes ion/molecule reactions with both the precursor and

the neutral molecules of interest. In this case the rate

expression becomes

-d[I]/dt = (k2[R] + k3[P])[I] (3.11)

where k3 is the bimolecular rate coefficient for the

reaction of the ion with the precursor molecule, P. The

integrated form is

[I](t) = [I](0) exp [-(k2[R] + k3[P])t] (3.12)

In order to determine the rate coefficient of reactions of

the ion with the neutral reactant, it is necessary first to

monitor the reactions of the ion with precursor neutral with

no reactant neutral present in the analyzer cell and then to

subtract the rate coefficient for this reaction from the

total rate coefficient observed in the presence of both the

precursor neutral and the reactant of interest. If the

precursor neutral pressure is kept constant for the two

cases, only the pseudo-first order rate constants are used

in the subtraction step.

When two or more isomers of a particular ion are formed

from the same parent neutral, a single ion signal is

observed in the mass spectrometer whose intensity is the sum

of the intensities of all the isomers. For a simple case of

only two ion structures,


[I](t) = [A](t) + [B](t)


Thus, the integrated form of the rate expression is given by

[I](t) = [A](0) exp[-(k2[R] + k3[P])t]

+ [B](0) exp[-(k2'[R] + k3'[P])t] (3.14)

where A and B are the isomers of ion I. Consequently, the

decay curve of the ion signal is composed of segments with

two different exponential time constants, each slope

corresponding to the reaction rate of a different ion

structure. The sum of the rate coefficients (for reactions

with precursor and reactant neutrals) of the slower reacting

structure is calculated from the tall of the double decay

curve while that of the faster reacting structure is

determined by subtracting the amount of the slow reacting

isomer at each point in the initial part of the curve using

Equation (3.14). Often, one of the structures is completely

unreactive, while the other is reactive as shown in Figure

3.3, simplifying the data analysis for structure


Reaction Mechanisms. Reaction pathways leading to

final product ions from a ion/molecule reactant system can

be studied using the ejection capabilities of FTICR. The












C) i



-- 01

I- 7IVN99S pNoI


method involves the ejection of each first generation

product ion continuously during the reaction time to see its

effect on the intensity distribution of the mass spectrum of

the final product ions. This method also can be used to

calculate the branching ratios for the product ions.

Studying reaction pathways and branching ratios of the ions

produced in different ways (either from different precursors

or different methods of ionization) provides an alternative

means to rate coefficient determination-for the

differentiation of structural isomers of the ion of




The C3H3+ ion has received considerable attention in

recent years as a possible soot precursor because it is

found in quite high abundance in fuel-rich and sooting

flames (Goodings et al, 1979; Olson and Calcote, 1981b;

Michaud et al., 1981). Although substantial uncertainty

remains as to C3H3+ formation mechanisms in flames (Calcote,

1981), the ion is postulated (Calcote, 1981; Olson and

Calcote, 1981a) to react with neutrals such as acetylene,

diacetylene, and C2H in rapid, sequential condensation and

condensation/elimination reactions, forming successively

larger ions, which can rearrange to cyclic species during

the ion/molecule reaction chain.

Two isomeric structures of C3H3+ shown in Figure 4.1

are important in discussing the role of C3H3+ in

ion/molecule reaction mechanisms. The first is the

cyclopropenylium ion, recognized as the most stable isomer,

which has a theoretically calculated heat of formation of

253 kcal/mol (Radom et al., 1976), in quite good agreement

with the 256 kcal/mol determined by experiment (Lossing,

1972). A second and potentially more important C3H3


t -C,-C-H


(4.1) (4.2)

Figure 4.1. The Most Stable C3H3 Structures:
Cyclopropenylium (4.1) and Propargyllum (4.2).


structure is that of the linear propargylium ion with a

calculated (Radom et al., 1976) heat of formation 31 to 34

kcal/mol higher than that of the cyclopropenylium ion, in

fair agreement with the 25 kcal/mol difference found

experimentally by Lossing (1972). Recent calculations have

shown several other stable isomeric structures of the CH3 3

ion with higher heats of formation relative to the

cyclopropenylium and propargylium ions.

The importance of different precursors in affecting the

reactivity of C3H3+ ions was reported in a thermochemical

study by Holmes and Lossing (1979). In an ICR study of

C3H3 + reactions, Ausloos and Lias (1981) showed that

significant fractions of the linear isomer can be produced

by charge transfer reactions of small ions (Ar*, Xe CO+,

Ne+, etc.) with propargyl chloride and bromide. Later it

was reported (Baykut et al., 1986) that even higher

proportions of the propargyllum isomer relative to the

cyclopropenylium isomer can be obtained with propargyl

iodide either by electron impact or charge exchange using


A study by Smyth et al.(1982) demonstrated that the

cyclopropenylium ion was relatively unreactive toward simple

hydrocarbon fuels whereas the linear, propargylium ion was

quite reactive. In particular, reaction of propargylium

ions with acetylene was reported to produce C5H3 and C5H5

ions with an overall C3H3+ disappearance rate coefficient of


1 x 10-9 cm3/s. In the work discussed in this chapter,

results substantially different from those reported earlier

were found and the research was thus expanded to investigate

possible production of C5H5+ Ions from ionic sources other

than C3H3+ present in the reaction media using the ejection

capabilities of Fourier transform ion cyclotron resonance

(FTICR) mass spectrometry. Reactions of propargylium ions

with C2D2 were also studied in order to delineate further a

proposed mechanism for the C3H3+/acetylene interaction.

Diacetylene is another important flame neutral which

has been postulated to react with C3H3+ ions in an

Ion/molecule soot formation mechanism (Olson and Calcote,

1981a). Reaction pathways and the rate coefficient for the

reaction of propargylium ions with diacetylene near room

temperature were thus also investigated. In this chapter,

results for the reactions of C3H3+ ons with acetylene,

deuterated acetylene, and diacetylene are reported and

discussed in relation to previous work involving C3H3

reactions (Ausloos and Lias, 1981; Smyth et al., 1982;

Baykut et al.,1986).


Reactive and non-reactive C3H3 ions were produced by

charge transfer reactions of various precursors with Xe+,

formed with an electron beam pulse of 5 ms duration at an

ionizing electron energy of ca. 15 eV. Propargyl iodide was


used as a precursor for C3H3 in studies of the reaction of

this ion with acetylene, deuterated acetylene, and

diacetylene, while a number of different precursors

(propargyl iodide, propargyl bromide, propargyl chloride,

propyne and allene) were used to investigate the reaction

mechanisms leading to C5H5+ Ion formation. In some studies

C3H3+ ions were produced directly from the above-mentioned

precursors by electron impact to compare the effect of

ionization technique on the ratio of reactive to unreactive


Propargyl iodide was prepared from propargyl chloride

by a halogen exchange reaction (March, 1977). The details

of purification are given elsewhere (Baykut et al., 1986).

Propargyl chloride, propargyl bromide, allene, propyne and

acetylene were obtained commercially and their purity was

checked by obtaining wide mass range spectra. In the case

of acetylene, some production of protonated acetone was

observed at long delay times indicating the presence of a

small amount of acetone as an impurity. Propargyl bromide

was distilled before use in order to remove toluene which

was present as a stabilizing agent. Deuterated acetylene

was prepared from D20 and CaC2. Diacetylene was prepared by

the method of dehydrochlorination of 1,4-dichloro-3-butyne

in aqueous potassium hydroxide/dioxane solution (Snow,

1985). All the samples were used after multiple

freeze-pump-thaw cycles.



C4H+ Reactions with Acetylene. Despite an earlier

report (Smyth et al., 1982) that C3H3 Is quite reactive

with acetylene, only very low intensities of C5H3+ and C5H5+

produced via this reaction could be found in this work.

Experimental conditions of the earlier study were duplicated

as closely as possible, and then varied substantially with

respect to relative pressures of neutrals (from 1:1 to 8:1

C2H2:C3H3I) and overall system pressure (from 5 X 10-7 to 3

x 10-5 torr). The C3H3+ ions were formed from propargyl

chloride, bromide, and iodide by both charge transfer using

Xe+ and electron impact. In order to determine other

possible sources of C5H5 observed under the earlier

reaction conditions, binary mixtures of acetylene and one of

the C3H3 precursors reported (Ausloos and Lias, 1981; Smyth

et al., 1982; Baykut et al., 1986) earlier were used.

Intensities of C5H5+ and C5H3+ were first measured after a

125 ms reaction time. Then the parent ion, C2H2+, and C3H3

were each ejected separately during the 125 ms reaction

period to assess their contribution to C5H5+ and C5H3


For each different neutral precursor, the sources and

amounts of C5H5 ions produced were found to be different.

Propyne and allene were similar in producing large amounts

of C5H5+ and no C5H3+ ion. However, the C3H3+ + C2H2


reaction was not responsible for C05H5 formation. The main

reactions leading to C5H5+ were

C2H2+ + C3H4 ----> C5H5 + H (4.1)

C3H4 + C2H2 ----> C5H5 + H (4.2)

in both cases. On the other hand, when propargyl iodide,

propargyl bromide and propargyl chloride were used as

precursors, relatively smaller amounts of C5H5+5 on

formation were observed along with C5H31 ion production. In

the propargyl chloride case

C3H3C1+ + C2H2 ----> C5H5+ + Cl (4.3)

C2H2 + C3H3Cl ----> C5H5+ + Cl (4.4)

were the major reactions leading to C5H5+ formation. For

propargyl bromide the

C3H3Br + C2H2 ----> C5H5+ + Br (4.5)

reaction was the only source of C5H5+ ions observed. Any

contribution to C5H5+ formation from linear C3H3 was less

than the experimental uncertainty. Finally, very little

(almost negligible) amounts of C5H54 ions were observed when

propargyl iodide was used as a precursor and the reactions

the reactions

C2H2+ + C3H3I ----> C5H5 + I (4.6)

C3H3+ + C2H2 ----> C5H5 (4.7)

were the major contributors in this case. An upper limit

for the rate constant, k, for reaction (4.7) was estimated

as 5 X 10-12 cm3/s by assuming that the very small C5H5

signal observed resulted from this reaction, and using the

expression [C3H3+](t) = [C3H3+](0) [CgH5+](t) =

[C3H3+](O)e-nkt, where n is the C2H2 number density.

Overall results for the production of C5H5+ and C5H3+ ons

in mixtures of acetylene and various neutrals used as the

precursors of C3H+ are summarized in Table 4.1.

Because propargyl iodide was shown to produce the

highest reactive/unreactive ratio of C3H3+ ons both in

earlier (Baykut et al., 1986; Holmes and Lossing, 1979) and

the present work (see Table 4.2), it was used as a precursor

for C3H3+ ions in these reaction kinetics studies. Since

the precursor neutral molecule was always present in the

FTICR analyzer cell, it was a competitor with the reactant

neutral of interest in ion/molecule reactions involving

C3H3+ In order to determine the rate coefficient of

reactions of C3H3+ with the neutral reactant, it was

necessary first to monitor the reactions of this ion with

C3H31 and then to subtract the rate coefficient for this

reaction from the total rate coefficient observed in the

presence of both the precursor neutral and the reactant of


Production of C5H5* and 05H13 Ions in Mixtures of
Various Neutrals and Acetylenea

Neutral Ionic sources Ratio of Ionic sources Intensity Percent
of C5H5+ after C5H5+ prod. of C5H3+ of C5H5+ reactive
Xe+ charge relative to vs. C5H3+ C3H3'
transfer that in ions
ionization of allene case
a mixture of
the neutral
and acetylene

Allene C2H2+(40%) 1.0 <5

Propyne C2H2(40%) 0.75 -- 30

Propargyl C2H2+(40-50%) 0.25 C3H3Cl+(20%) 3.0 15
chloride C3H3Cl+(50-60%) C3H3+(20%)

Propargyl C3H3Br"(90-1001) 0.08 C2H2+(70%) 2.0 85
bromide [C3H3 3


Propargyl C2H2+(40%) (0.02 C2H2C(50%) 1.7 90
iodide C3H3+(60%) C3H3+(50%)

? All ions were produced by chemical ionization charge transfer from Xe+.
o Percentages show the relative contributions to C515" and C5i3 production ns
determined by double resonance experiments and have an estimated uncertainty
of 10%.
c Neutral reactants all had the same pressure (7 X 10-7 torr) .is measured by
the ionization gauge. Xenon and acetylene pressures were 5.d 10- .and l.n
X 10-o torr. respectively.



Percentagesa of reactive C3H3+ found from various precursors
by various ionization techniques (monitored by observing
reaction with the precursor neutral).


Ionizing Propargyl Propargyl Propargyl
Technique iodide bromide chloride

Electron impact (15eV) 90 40 10

Chemical ionization charge 90 85 15
transfer with Xe+

aEstimated error is +5%.


interest. Reactions of C3H3+ with propargyl iodide were

monitored as a function of time following charge transfer

chemical ionization of C3H3I by Xe+ and ejection of all ions

but C3H3+ from the analyzer cell. Results obtained were

identical to C3H3+ reaction channels with propargyl iodide

which have been reported elsewhere (Baykut et al., 1986).

Isomerization of Linear C3Hj. In addition to the

absence of any significant C5H3+ and C5H5+ formed by

reaction of linear C3H3+ with C2H2, it was also observed

that C2H2 led to the isomerization of linear C3H3 ions to

their cyclic form, thus rendering them unreactive toward

their parent neutral (C3H3I) as well as toward C2H2. This

isomerization was followed as a function of C2H2 pressure

and a direct pressure dependence was found, as can be seen

in Figure 4.2.

_CH+ Reactions with C2D2. To achieve a better

understanding of the isomerization of linear C3H3+, C2D2

instead of C2H2 was used as the neutral reactant. The

following isotope exchange reactions were observed:

C3H3+ + C2D2 ----> C3H2D+ + C2DH (4.8)

C3H3+ C2D2 ----> C3HD2 + C2H2 (4.9)

C3H2D+ + C2D2 ----> C3HD2 + C2DH (4.10)

C3H2D+ + C2D2 ----> C3D3+ + C2H2 (4.11)

Deactivation of 1- C31-1 by acetylene at different pressures

0.5 1.0 1.5



Figure 4.2. Isomerization of Linear C3H3 Ions at Different
Pressures of C2H2. C3H3 ions were produced by charge
transfer reactions with Xe. p(C3H3I) = 1.1 x 10-7 torr;
p(Xe) was adjusted to maintain a constant total pressure of
2.6 x 10-6 torr as measured on the ionization gauge. (All
pressures are capacitance-manometer corrected.)



o 0.5





C3HD2'+ + C2D2 ----> C3D3 + C2DH (4.12)

Using FTICR ejection capabilities, it was found that

reactions (4.9) and (4.10) contribute equally to the

production of C3HD2+ while reaction (4.12) produces more of

C3D3+ (80%) than reaction (4.11) (20%).

Ion intensity vs. time curves for the C3H3 /C2D2

reaction are shown in Figure 4.3. The overall rate

coefficient for the disappearance of C3H3+ was calculated by

subtracting the observed rate coefficient for the reaction

with propargyl iodide from the total observed rate

coefficient in the presence of C2D2. This observed rate

coefficient was then converted to the true rate coefficient

using the corrected pressure of C2D2. A value of (4.5+1.9)

x 10-10 cm3/s was found at a cell temperature of 373 K for

the disappearance of C3H3q (reactions (4.8) and (4.9)). In

Figure 4.4, ion intensity vs. time curves of C3H3+ are

compared for reactions with and without C2D2.

CH + Reactions with Diacetylene. After ejection of

all ions except C3H3+ following charge transfer chemical

ionization by Xe1 of a mixture of diacetylene and propargyl

iodide, the ion/molecule reactions as a function of time

were monitored. Consecutive C2 and C4H2 addition reactions

were observed:

C3H3 + C4H2 ----> C7H5


Reactions of L- C3H+ with C D2


0 40 -


0 A -r l i i I I I I
0 0.5 1.0
Time /s

Figure 4.3. Isotope Exchange Reactions of C3H3 with C2D2.
Disappearance of C3H3 Ion includes reactions with propargyl
iodide. Note that the sum of all isotopic forms of C3H3
remaining at the end of the reaction with C2D2 approximately
equals the total unreactive C3H3 when C2H2 is used as a
neutral reactant at the same pressure (see Fig. 4.2).
p(C3H3I) = 1.1 X 10-7 torr p(C2D2) = 1.2 x 10-6 torr; p(Xe)
= 6.2 x 10 torr.

Reactions of LI- C3H with C3H3I

+ PO



and C2D2

0 0.5 1.0 1.5


Time /s

Figure 4.4. C3H3 Ion Decay Curves for Reaction with C3H31
and C2D2. (Pressures are the same as given for Fig. 4.3.)


C3H3 + C4112 ---- C5H3 + C2H2 (4.14)

C5H3 + C4H2 ----> C9H5 (4.15)

C5H3 + C4H2 --> C7H3+ + C2H2 (4.16)

CH3+ + C4H2 ----> C11H5 (4.17)

C9H5+ + C4H2 ----> C11H5+ + C2H2 (4.18)

C9H5+ + C4H2 ---> C13H7 (4.19)

Some of these product ions were seen to react further

with propargyl iodide by displacement of atomic iodine:

05H3 + CHI1 ----> C8H6 + 1 (4.20)

C7H3 + C3H3I --> C10H6+ + 1 (4.21)

C7H5 + C3H3 ----> C10 + I (4.22)
7+ +
C8H6 + C3H3I ----> C11H9+ I (4.23)

C9H5 + 0C3H3 ---> 12H8 + I (4.24)

Ion intensity vs. time curves for the C3H3+/C4H2

reaction are shown in Figure 4.5. The rate coefficient for

the disappearance of C3H31 (reactions (4.13) and (4.14),

Figure 4.6) was calculated as described earlier, and a value

of k (1.4 + 0.7) x 10~ cm3/s was found. Propargyl

iodide, bromide, and chloride were all used as precursors of

C3H3+ in studying its reactions with diacetylene. For each

precursor, both electron impact and charge transfer chemical

ionization techniques were used. The percentages of

reactive isomer in the reaction with diacetylene are shown

Reactions of 1-C H with C4H

+ C H +
O :ih Cz3
40 C H
0: CsH+

S% : C9H+

S20 -

0 0.5 1.0 1.5 2.0
Time /s

Figure 4.5. Reactions of C3H3+ with C4H2. Disappearance of
C3H3+ and product ions include reactions with propargyl
iodide. p(C3H3I) = 1.1 x 10-7 torr; p(C4H2) = 4.8 x 10-7
torr; p(Xe) = 6.2 x 10-6 torr. (All pressures are
capacitance-manometer corrected.)

Reactions of 1-C3 H with C3 H3 I and C4H2

0 1.0



Figure 4.6. C3H3 Decay Curves for the Reactions with C3H3I
and C4H2. (Pressures are the same as given for Fig. 4.5.)


Percentagesa of reactive C3H3+ observed in the reaction with


Propargyl Propargyl Propargyl
iodide bromide chloride


Electron impact (15eV) 75 30 5

Chemical ionization charge 75 65 5
transfer with Xe+

aEstimated error is +5% bP(C4H2) = 4.8 x 10- torr
a(CHs tlmated0 or


in Table 4.3. When these percentages of reactive isomer

were compared to those in the absence of C4H2 (see Table

4.2), it was clear that some isomerization of the reactive

linear C3H3 ion, as well as reactions (4.13) and (4.14),

had taken place (see also Fig. 4.6). This isomerization was

followed as a function of C4H2 pressure and a direct

pressure dependence was found, as can be seen in Table 4.4.

The reactions of 1-C3H3+ with propargyl iodide and with

both acetylene and diacetylene have also been followed at

several elevated temperatures up to 500 K. All the rate

constants were found to be similar to their room temperature

value within experimental error.


Effect of Different Precursors. Different percentages

of reactive C3H3 were found from three different

precursors, propargyl iodide, propargyl bromide and

propargyl chloride as shown in Table 4.2. To explain the

differences observed in reactivity, schematic potential

energy surfaces for these precursors are shown in Figure

4.7. Experimental thermochemical data reported by Holmes

and Lossing (1978) were used in the generation of the

potential surfaces. Reverse activation energies for c-C3H3

formation from propargyl bromide and chloride were

determined by the difference between the experimental and

the calculated appearance potentials. Since the appearance



Changes in CH3+ reactivitya at different pressures of

C4H2 pressure/10-7 Torr % unreactive C3H3

0.8 16
1.6 17
4.8 25
7.2 32
8.0 35
9.6 40

al-C3H3 ions were produced from propargyl iodide by
chemical ionization charge transfer with Xe+. (p(C3H3I)
1.1 X 107 torr; pXe was ad usted to maintain a constant
total pressure of 2.6 X 10 torr as read on the ionization
All pressures are capacitance-manometer corrected.

1.13 ev

c-C3H3 *


10.48 eV

9.10 ev



.12.13 eV
SI 'x


10.68 eV


C H Cl

Figure 4.7. Schematic Potential Energy Surfaces for
C3H3+/C3H3X+ System from Different Precursors. Propargyl
Iodide (a), Propargyl Bromide (b). and Propargyl Chloride



potential of C3H3+ produced from propargyl iodide very

closely corresponds to the calculated threshold for 1-C3H3+

rather than c-C3H3+, the dissociation to the latter is

assumed to have a significant energy barrier. Thus, reverse

activation energy for the dissociation channel giving

unreactive c-C3H3+ decreases in the order Eiodo > Ebromo >

Echloro as shown in the figure. Production of almost

exclusively reactive 1-C3H3+ by both electron impact and Xe+

charge transfer ionization from propargyl iodide suggests

that Eiodo is so large that the fragmentation channel

leading to l-C3H3+ becomes the lowest energy channel. In

the case of propargyl bromide, there is enough excess energy

to dissociate to both 1-C3H3+ and c-C3H3+. Production of

80% 1-C3H3 by Xe+ chemical ionization suggests no

significant energy barrier for the 1-C3H3+ channel. It is

interesting to note that dissociation to l-C3H3+ reduces by

a factor of two when electron impact ionization is used,

which demonstrates the effect of a large distribution of

electron energies from an electron impact ionization source

on the relative abundances of two channels almost equally

accessible energetically. The very small reactive 1-C3H3+

percentage produced from propargyl chloride by both electron

impact and Xe+ charge transfer Ionization suggests at least

a small energy barrier for this dissociation channel, as

indicated schematically in Figure 4.7c.


Internal Energy of the Ions in Relation to Rate

Coefficient Measurements. In studying reactions it is

desirable to have knowledge of the internal energy

distribution of the reactant ions. The reactions studied

here are bimolecular addition reactions followed by

unimolecular decomposition. As shown in Figure 4.7a, when

1-C3H31 is formed from propargyl iodide by Xe+ charge

transfer chemical ionization, 1.5 eV of excess energy is

available. Much of this excess energy will be converted

into translational motion of the heavy Xe and I neutrals

resulting from the charge transfer in the collision process.

Under typical experimental conditions (ca. 2 3 x 10-6 torr

total pressure), about 125 as was allowed for the charge

transfer process and for ejecting intermediates. For these

conditions the C3H3+ ions collided a number (10-15) of times

with the excess Xe present in the FTICR cell, leading to

near thermalization of internal energy before the

ion/molecule reactions were monitored. Since the reaction

time scale was on the order of seconds, it can be assumed

that any slight initial deviation from Boltzmann behavior

presents no serious error. On the other hand, the

observation that the rate constants for the reactions of 1-

C3H3 with propargyl iodide and with both diacetylene and

acetylene are temperature independent implies that

thermalization of the ions is not complete under the

conditions reported and there is still some internal energy


in 1-C3H3 which is comparable in magnitude to that

contributed from the range of temperatures studied.

Reactivity of l-C3H3 with Acetylene. Although the

results of C3H3 + C2H2 reaction are not in agreement with

the earlier report (Smyth et al., 1982) of C3H3 /C2H2

reactivity, the discrepancy is most likely due to

limitations of the older pulsed ICR (Smyth et al., 1982)

instrumentation for studying ion/molecule reaction pathways

in complicated systems when compared to newer FTICR

capabilities. Facile ejection of all ions except the one

whose ion/molecule reactions are being investigated offers a

very clean monitoring opportunity for product-parent

relationships even in complicated consecutive and

competitive reaction systems. Various alternative pathways

for the production of C5H3+ and C5H5g which have been

described above probably contributed significantly to the

intensities of these ions seen in the earlier work.

Additional support for the low reactivity of C3H3+ with C2H2

is found in a recent report (Anicich et al., 1986) of the

rate coefficient for this reaction as less than 0.01 X 10~

cm3/s, although the isomeric form of C3H3+ was not given.

It is also possible that the highest pressures used in this

work did not reach those of the earlier study due to

differences in the location of capacitance manometers,

ionization gauges, etc. Thus third body stabilization of


C5H5+ collision complexes might have been occurring to some

extent in the earlier work and not in that reported here.

In fact, such collisional stabilization of the association

complexes for the reactions of C3H3+ and C4H4 with C2H2 has

been shown to occur in higher pressure SIFT studies (Smith

and Adams, 1987; Knight et al., 1987).

The most likely mechanism of the observed isomerization

of C3H3+ ions by collisions with acetylene is a "reactive"

aather than a "non-reactive" one. That is, it results from

an intimate encounter of the ion and neutral in the C5H5+

collision complex. This hypothesis is confirmed by the fact

that deuterated forms of C3H3+ were produced when C2D2 was

the neutral reactant (see Figure 4.4). Kinetic modeling

studies (discussed in Chapter 5) indicate that in some cases

the C5H5' collision complex dissociates to give the cyclic,

unreactive, C3H3+ somer, instead of the reactive, linear

form which reacted initially. The possibility of

non-reactive collisional isomerizatlon of linear C3H3+ to

the cyclic isomer has been ruled out because experiments at

elevated pressures of xenon (to ca. 1 X 10-5 torr) showed no

interconversion. Similar interconversion of C4H4+ ons from

a linear to cyclic form has also been reported (Jarrold et

al., 1984) in the reaction with C2H2 and has also been shown

to take place via complex formation by using isotopically

labeled C2H2. To confirm the hypothesis that energetically

less stable, reactive, (linear) C3H3+ ions interconvert to

more stable, unreactive ones, cyclic C3H3 ions were also

reacted with C2D2 and no isotope exchange reactions were


Reactivity of 1-C31H3+ with Diacetylene. Plots of C3H3+

ion intensity vs. time for reaction with diacetylene (C4H2)

(Figure 4.5) indicate a 10-12% increase in the intensity of

the unreactive isomer relative to the reaction when the

parent precursor only is present. Isomerization of reactive

C3H3+ was also seen when different precursors were used

(compare Tables 4.2 and 4.3). A similar mechanism involving

complex formation may be responsible for this isomerization

as well, although it was not investigated in any detail.



As reported in the last chapter, bimolecular reactions

of the propargylium form of C3H3+ with acetylene most often

result in an isomerization to the cyclopropenylium isomer.

To help understand this isomerization process, C3H3+

reactions with deuterated acetylene were investigated.

These studies showed that the isomerization proceeds via the

C5H5+ ion/molecule reaction complex, which is sufficiently

long-lived under the experimental conditions employed that

deuterium exchange, as well as isomerization, takes place.

Thus with time the reactive propargylium C3H3+ isomer is

converted to both reactive and unreactive species containing

one, two, and three deuterium atoms. There is no evidence,

either experimental or theoretical, that the propargylium

cation converts into the cyclopropenylium cation in the

absence of the C5H5+ reaction complex. In order to better

understand the isomerization which converts the reactive

to the unreactive form of C3H3+, kinetic modeling studies of

the ion intensity vs. time curves reported in Chapter 4 were


carried out.1 It was also hoped that fitting procedures

would produce improved ion/molecule reaction rate

coefficients. Quantum mechanical calculations2 on C3H3+ and

C5H5 structures and reactivity were used to guide the

modeling effort.


A mixture (predominantly propargylium) of the C3H3+

isomers was in most cases formed by charge transfer from

propargyl iodide to Xe+, produced by 15 eV electron

ionization of Xe (present at pressures > 10x those of other

gases). Other conditions such as neutral partial pressures,

pulse sequences and reaction times, and the sources of

chemicals were kept as close as possible to those reported

in Chapter 4 for the duplicate kinetics experiments reported

and modeled here. Any significant deviations are given in

the text, table headings, or figure captions. All pressures

reported in this chapter were determined by a capacitance

manometer, and then multiplied by a "system factor" of 0.30

which corrects for the fact that the pressure read by the

capacitance manometer is not the same as that in the FTICR

1The kinetic modeling studies were performed in the
Environics Division of Air Force Engineering and Services
Center, Tyndall Air Force Base, Florida by P. Wiseman using
multiple experimental data sets produced at identical
conditions to those reported in Chapter 4.
2A. Cameron, J. Leszczynski, M. C. Zerner and B Weiner,
submitted. J. Feng, J. Leszczynski and M. C. Zerner, submitted.


cell. Non-linear least-squares fitting routines employing

Marquardt's algorithm (Annino and Driver, 1986), implemented

on two different computers3, were used for kinetic modeling.

Complete analytical solutions were obtained from the

chemical models developed below for the systems C3H3+ + C2H2

and C3H3 + C4H2. A complete analytical solution was not

possible when an isotope effect was included in the chemical

model for the C3H3 + C2D2 system. Numerical integration

used the finite difference method (Annino and Driver, 1986).


Models of CH4+ + C2H2 Reaction. As reported in

Chapter 4, collision of the propargylium cation, 1-C3H3

with acetylene forms the cyclopropenylium cation, c-C3H3,

which is unreactive on the time scale of the experiments,

given the pressures attainable in the FTICR cell.

Experiments with C2D2 showed that an encounter complex which

allows for isotopic scrambling is formed. Hence, whatever

the isomeric form of this complex, a structure having the

chemical formula C5H5+ can be postulated. Since no species

of m/z 65 is observed in the mass spectrum, the (C5H5+)

species must be in steady-state and of low concentration.

The simplest scheme which takes into account this

information is given in Figure 5.1.

3Tektronix Model 4054 and Hewlett Packard Model 150.

L-C3H3+ + C2H2 -- > C5H+
L-C3H3+ + C3H3I --> sink

C5H5 k > &-C3H3+ + C2H2
C5H5 > c-C3H3 + C2H2

Figure 5.1. Reaction Scheme Postulated for the Kinetic
Modeling of the Reaction of linear C3H3+ with Acetylene.


Applying steady-state kinetics to this scheme yields

I(t) = 10 kp'I10 (1 e et)/ (5.1)

in which I1 and I(t) are Ion intensities initially and as a

function of time, respectively, Il refers only to the linear


S= kpPC3H3I (5.2)

S= [kfkcPC2H2/(kl + kc)] + kp' (5.3)

Equation (5.1) was fitted to several kinetic runs reported

in Chapter 4. Table 5.1 shows results of th se fits. A

plot of 8 vs. PC H should be linear, as implied by equation

(5.3), and this is demonstrated in Figure 5.2. Results

yield kfkc/(kl + kc) = 2.3(.2) x 106 torr-1l-1 and kp'

1.4(0.2) s-1.

Models of C3H_+ + C2D2 Reaction. Reaction of C3H3 +

C2D2 is complicated by the observation that isotopic

scrambling occurs and isotope effects are possible. Several

models were tried, including those which allowed for

complete isotopic scrambling and those which allowed only

partial scrambling. The simplest model allowed for complete

scrambling and no isotope effects4

4Lampe and Field (1959) studied the reaction of CD4 +
C2H4, and observed the following yields: C3HD4 : 1/10,
C3HD2 : 2/5 C3H2D3 : 2/5, and C3HD : 1/10. Statistical
yields with no isotope effect would have been: C3HD :
1/14, C3H5D2+ : 3/7, C3H2D3 : 3/7, and C3HD4+ 114, very
35114 very3


Results of fits of equation (5.3) to kinetic data for the
C3H3+ + C2H2 reaction.a

PC2H2 /torr

k IO/Arb. units s-1

0 1.53(.06) 1.58(.08)







2.0 x 10-6 1.71(.11) 6.33(.38)

aThe standard error of estimate computed by the fitting
program is shown in parentheses.

close to the observed values. Hence, shece the C3L8+
complex exhibited almost complete scrambling, it is
reasonable to expect the C5L5g complex modeled here to
undergo complete, or nearly complete scrambling.


4.4 x 10-7

1.0 x 10-6

1.6 x 10-6

6.2 -

5.2 -

\ 4.2 -

3.2 -

2.2 -

1.2 '-
0 .13 .26 .39 .52 .65

PC2H2/(1 0-6torr)

Figure 5.2. The Plot of 0 as a Function of Acetylene
Pressure (see Equation (5.3)).


Complete scrambling occurs when fragmentation of the

complex. C5L5+ (L = H, D), yields precursors having a

statistical distribution of hydrogen and deuterium atoms.

C5H3D2 will then yield the following ratios of precursors:

C3H3+ = 1/10, C3H2D+ = 3/5, and C3HD2+ = 3/10. Using these

statistics for obtaining the isotopic distribution in the

C3L3 precursors, we obtain the scheme shown in Figure 5.3.

With the assumption again that all four C5L5+ complexes are

in steady-state, a full analytical solution is possible for

the set of kinetic differential rate equations. The

solutions to the set of equations are given in Appendix

II.A. Figure 5.4 shows the best fit curves to a typical

data set. Table 5.2 shows the fitted parameters, errors,

and residual sum of squares from fits of a typical data set.

An examination of Figure 5.4 shows that the model given

by Figure 5.3 does not adequately explain the production of

C3HD2+. In an attempt to examine this, incomplete, or

partial isotopic scrambling was next assumed. To do this

correctly requires a detailed knowledge of the chemistry of

the system, which is not available. A somewhat crude

application of isotope effects applies multiplicative

factors to the individual rate constants and this procedure

requires but a nominal knowledge of the structure of the

t-C3H3 + C2D2 -> CSH302

t-C3H3 + C3H31 > sink

t-C3H2D + C2D2 --> C5g2D3+

-C3H2D + C31131 --> sink

I-C3HD2 + C2D2 --> C5HD4

-C3HD2 *+ C3131 -P-> sink

1-CD + CD -- > C5D
3 3 2 2 5 5

i -C3D3 CH31 -- -> sink

+ tk + 1 3 +

+ kc 1 + 1 3 +
C H3D2 0 c3CH3 +10 C202 + c-C3H2D

3 3" + 3
5 C2HD + C-C3HD + -O CH2
5 2 10 3 2 10 22

5 2 10 310 1 0 22H

C52D3 L-3 + 7D + 2 C + -C3 D2

SC2 HD + 3 H2D + 3

CCD 2DC 1 +C22

3 3 WCH2D2 32

SC2HD + 2 10 -2D2

C HD4 > -C HD2 + C2D2 -C3D3 + C2HD

CRD4 > c-CHD2 C D c-C3D3+ + C
55 22 5 2+35C 5 C2X

kt 4

+ kc +
C5DS -- > -C3D3 + C2D2

C D ->c-C 0 CD,

Figure 5.3. Reaction Scheme Postulated for the Kinetic
Modeling of the Reaction of linear C3H3+ with Deuterated
Acetylene Assuming Complete Scrambling and No Isotope


0 1C3H20+

.6 + IC3H3+



.2 -


0 L -
0 .5 1 1.5



= .08

< .06
S.04 a
.02 7A C3D3+

0 O 0C3HD2+
0 .5 1 1.5

Figure 5.4. Model Fit (using the scheme of Figure 5.3) to a
Typical Data Set for C H+ + C2D2 Reactions. Ion Intensity
vs. time curves for (a C3H3 and C3H2D+ and (b) C3HD2 and

ons. Ion Intensity
vs. time curves for (a C3H3 and C3H2D+ and (b) C3HD2 and












C 0
,e >

b Q)



-0 U


0. -
0 U C











to to C-
a i I
0 0 0


eq '* eq

r. 0. 0.

a. & m
































































C 0

o u
a. a.


species involved5. This simple scheme was applied as

described below.

Quantum mechanical calculations6 indicate that C2H2

does not readily react with the cyclopropenylium cation, but

does react with the propargylium cation (1-C3H3+) without

barrier with formation of the four possible products shown

in Figure 5.5. Thermodynamically, only Structure (5.1) is

stable with respect to decomposition to c-C3H3+ + C2H2.

However, since the 1-C3H3+ + C2H2 ----> Structure (5.1)

reaction is at least 60 kcal/mol exothermic6 in the

absence of stabilizing collisions this energy then permits

many different isomeric forms to be sampled before

decomposition back to C3H3 + C2H2. If the sampling of all

isomers is fast, complete isotopic scrambling is expected.

Given the uncertainty in C5H5 structures, total

equivalence of carbon atoms in the complex is assumed. For

ease of understanding, a cyclic C5H5 complex, in which all

C atoms are sp2 hybridized, might be visualized. With these

assumptions, it is a simple matter to determine which carbon

atoms undergo hybridization changes during the reaction.

For instance, in the attack of C2H2 on 1-C3H3+, three carbon

5Although somewhat crude, applying a multiplicative factor
to the rate constant for each deuterated site is in keeping with
experimental evidence. For instance, in the acetolysis of some
identical tosylates, each deuterium atom substitution changed the
rate constant by ca. 0.84 (Streitweiser et. al., 1958).

6Details of these studies are given in: J. Feng, J.
Leszczynski and M. C. Zerner, submitted; and J. Leszczinski, M.
C. Zerner and F. Wiseman, submitted.




1.3018 r J.oJ


Figure 5.5. C5H5 Structures which are Proposed to Form by
the Reaction of Propargylium Ion with Acetylene Without an
Energy Barrier.


atoms change from sp to sp2 hybridized. Upon fragmentation,

some carbon atoms remain sp2 hybridized; others become sp


In general, for a-secondary isotope effects, an

increase in hybridization in going from the reactant state

to the transition state yields an inverse isotope effect (kH

< kD), whereas a decrease in hybridization yields a normal

isotope effect (kH > kD) (Dreuth and Kwart, 1980). In this

system the bimolecular addition reaction will have an

inverse isotope effect, and the fragmentation reaction

should have a normal isotope effect. Figure 5.6 shows how

the isotope effects for the reaction of l-C3H3+ + C2L2 can

arise. For the forward addition reaction, EFH > EFD and for

the fragmentation reaction, ERH < ERD, in which the

subscripts "F" and "R" refer to forward and reverse,

respectively, and "H" and "D" refer to protonated and

deuterated species, respectively. Since the bimolecular

addition is very exothermic for the formation of most C5L5+

isomers, it might be expected that the "average" transition

state structure might closely resemble the reactants and not

any of the C5L5 isomers. This in turn implies that ERD -

ERH > EFH EFD (zero-point effects). However, the excess

energy in the reaction will allow longer sampling times for

the more energetic CL5+ isomers. The less energetic

isomers which are sampled will undoubtedly be in higher

rotational and vibrational levels. The overall effect Is to



"CC-CH + L-C C-L (sp) ERD

(s --> sp2 for forward;
sp --> sp for reverse.)


nC 2)


Figure 5.6. A Schematic Representation showing the
Qualitative Differences in the Zero-Point Vibration Energy
Levels for the Reactants. Transition-State, and a
Representative C5L5 Isomer for the Reaction of linear C3H3
with C2H2/C2D2. (EFH, EFD, ERH, and ERD are explained in
the text.)


lessen the normal isotope effects expected in the

fragmentation of the C5L5c isomers, unless there are

sufficient collisions to stabilize the isomers prior to


In the model development outlined below, only

a-secondary isotope effects will be considered important.

$-secondary isotope effects, arising predominantly from

hyperconjugation, can sometimes be important (Melander,

1960), but will be assumed here to be minor compared to the

a-effects. It is also possible that hydride/deuteride

shifts may be occurring in the transition- state. However,

hydride transfers often exhibit small isotope effects

(Melander, and Saunders, 1980) and hydride shifts, if they

occur at all, will be assumed here to give negligible

contributions to the isotope effects.

Even though the different reactions in Figure 5.3 will

have different isotope effects, the introduction of an

independent fitting parameter for each type of reaction is

not justified. Though crude, only one additional parameter

was introduced into the model to account for all potential

a-secondary isotope effects. The method for introducing

this parameter is outlined as follows.

If a carbon atom bearing a deuterium atom undergoes a

hybridization change from sp to sp2 (force field becoming

stronger), the "isotope effect factor", F, is introduced as

a multiplicative factor in the rate constant. For two


deuterium atoms, F2 is the multiplicative factor, etc. If

the deuterium atom is attached to an atom changing from sp2

to sp hybridized, the rate constant is divided by F, for two

deuteriums, F2, etc. Introducing the same factor for both

addition and fragmentation reactions implies a constraint

which is at best only qualitatively correct. Applied in the

numerator the factor corrects for a single deuterium atom

(F2 for two, etc.) attached to a center undergoing

hybridization change from sp to sp2 in the transition-state

complex. Applied in the denominator, it corrects for a

change from sp2 to sp hybridization. Since the

transition-state complex has a stronger force field than the

reactant state (1-C3gL + C2L2) at these centers, the

a-secondary isotope effect will be "inverse" and F should

therefore be greater than unity.

Using the structural notation,

H\ H
1-HDC3H+ = [ C-CC-H]H, 1-H2C3D [ C-C=C-D]+,

H\ D
1-HDC3D+ = [ C-C=C-D]+, and 1-D2C3 = [ C-C=C-H]+,
D/ D/

the full kinetic scheme is shown in Figure 5.7. The neutral

fragmentation products, C2H2, C2HD, and C2D2, have not been

included for brevity. It is assumed that l-C3L3 undergoes

the same kind of hybridization changes when reacting with

C3H3I as it does with C2L2. As shown in Chapter 4, the

-3H3 + C2D2 -- CSH3D2+

L-C3H3 C3I -> sink

CH3D2 > t-C -F L-HDC H + -H 2C3D + L-HDC3D + Lt-D2CH
2 i 2 + 1 + 1 ^aca^

t-HDC H + + C H- sink

L-H2C3D + C D1 ---> sink
L-H2C3D C C33 c-CD cC sDk

C-H k L-C3+D + I-l-HDC3D + +- H + L -HDCH + + L-H2C3D
3S 2 3 F 32 10 33

C 8H20 DE- C3D3 + I C-C3D + -- C-C3H2D
L-HDC 3' CD ---f CUHD4

S 22 53 -fc-CHD2 C HD 2
t-HDC D + C2D2 -- > sink

52 522 52

-D kC + 2DC D # CkS/F2
CsHDCD 1 CF D --L D> sink 3-C33D3C I 2 -> C D 3
CU k + + 1 1*4 /

CSD ---> 33-CD 4. t-HDC D 4. --D2C 4 C-LLDS D3

CU4 -> C3D3 c- c.C-CHD CSD -> c-CD3

Acetylene Assuming Complete Scrambling and a-secondary
Isotope Effects.
Isotope Effects.

3 4

Figure 5.7. Reaction Scheme Postulated for the Kinetic
Modeling of the Reaction of linear C33H3 with Deuterated
Acetylene Assuming Complete Scrambling and a-secondary
Isotope Effects.


products of the l-C3L3 + C3H3I reaction are of higher mass

and do not enter further into the kinetic schemes modeled


Steady-state conditions are applied for all four C5L5+

isotopic species as before (without regard for isomeric

differentiation) and the differential rate equations for

scheme shown in Figure 5.7 are given in Appendix II.B.

Table 5.2 shows fitting results for several data sets using

the same kinetic scheme. Figure 5.8 shows plots of the best

fit of this model to the same data set as fitted in Figure


Some experiments were conducted where certain ions were

ejected from the analyzer cell using FTICR double resonance

techniques (Comisarow et. al., 1978) as they formed. Among

the ions ejected were C3H2D+ and C3HD2. Without further

fitting, this model was used to predict the behavior of the

kinetic system if these ions were ejected. Figure 5.9 shows

predicted results and data points.

Models for CSH+ + C4H2. The reaction of C3H3+ with

C4H2 (diacetylene) is kinetically more complicated than that

of C3H3 with C2H2. There are more isomeric possibilities,

and ion/molecule reaction products of higher m/z are

detected. Several models were tried in attempts to fit the

experimental data, and the best of these made the

assumptions that C7H5+ and the excited forms of C9H5+ and


o IC3H2D+

.6 + IC3H3+




0 .5 1 1.5



.12 -

.12- /

0c Dat0 si AC2 a
z 0 .

.08 -
< .02 -

< / --- MODEL FIT

0 A 'C3HD2+
0 .5 1 1.5


Figure 5.8. Model Fit (using the scheme of Figure 5.7) to a
Typical Data set for linear C3H+ + C2D2 Reactions. Ion
intensity vs. time curves for (a) C3H3 and C3H2D and (b)
C3HD2 and C3D3 .

5.8. Model Fit (using the scheme of Figure 5.7) to a
Typical Data set for linear C3H+ + C2 D Reactions. Ion
intensity vs. time curves for (a) C3H3+ and C3H2D and (b)
C3HD2 and C3D3 .

t .122




0 .3 .6 .9 1.2


12 A IC3D3+

n .09 o



0 .3 .6 .9 1.2


Figure 5.9. Data from Ejection Studies and Model
Prediction. (a) C3HD2+ ion Is ejected, (b) C3H2D+ ion is
ejected. (Poor signal/noise ralos of experimental data
points are due to the effect of the ejection pulse on the
neighboring ion.)