O
o
a
5
M
O
8
no

 

W In
III}
1 1.2.3:? ‘ 13“”);

  
 
 
 
   

u! 1 [1|] )‘431 .3“

uh

«I ' My QJL‘ .-

I , I I» .
3s: 1 II!’ . ."~1‘“:'q"ql'.)3gl

' I 'IIM .333.‘ III IL} 3 531-3..“

\I' 3H K?"
'_' Y 3II'I‘3. - Iu‘v ' I .333: V W :33
.. I I. , I. ,,. (III

     
       
 
  
 
 
 
 
 
  

   

  
  

\III'.?5- ..‘.."E ’I04I3l~' .‘l. 711.”, '. ..
3'II' II, ”It: ”his. ."'!'I: ~.r.I.3IIII:.g;, MRI;
‘3‘» ‘ $34 3‘ I- “4‘3" thy-13“. ,l‘.. .. a

  
   
 
 
 
  

    

‘3’1'33313‘3’1333‘23 I I\ :v n 333,333.“) ‘5‘ ,I,‘,. 3‘ :3‘
3'3 I 3 ilfiV“. 33131373 MJ ‘ I :3 “"39"". *3. “3.3.33 n.“ 3331333. (3.7. ‘/:I Jikwfit‘ 3
I *l'l ‘ "31 3‘ '-.'."' .3 3 3“ :(ewiql. 3533.:{2- ‘,
,3; my ..I..: i..qg'-t,(:.,. I,;I3I}33fi';¢flb.I,III,.
I!" I‘ ' | 3, I.' ' .3 3..'I I ‘ 31 'II' 3
. «3-3 “.4, I 3113 333W 33. IIII‘:4;fiA‘fi.i I 3"...
I. II!“ ‘I‘ .':| I ',I( I.. . .l H M“.
I.l JII 3 ' : 3 ' 3 I. ' . I pg .‘
‘ ‘ 1 ‘I:IJ‘|'|I '. I I,3"1II .‘l-y.“
_ f . 21““ ." ‘ 34:.("1- ‘ I' "'
3“ ‘ 3* 3' I.“ 313.332‘3” 3 .3 3:33;... 1"" p 0." I l
3‘» , < 3. ‘Iwu- " .l‘ ' II” — I ' , ‘ |
'0 L "(,’f\g1l"'3 ' ,‘ I “m 3| ’4, l‘l l'
- ' o I ‘ I I ‘ t.
.‘ ’:"1I.'.,II'II': ‘II I... .--I ..
3 t-I-‘III. '~‘ I " a: ;I'-.I'. W-uI .
a - ‘34"I'I""II . I 'I .i‘l' "I '3’ _ 3.1 3II- . ‘
- 33‘ ‘ o -. 3; .: 3; ."'3i "I 3 1334“ IIlu" $1 33:336..

—“.‘ '

Y ' ' I“: ”H .
‘3‘ 3333 3.333. --33'Ht3 film (4'93: .3

L. .;'( "J
«3“ f- : Jr'-

W931? -:,~ “I’M. .’
3 ”E3333 21”?" ”13.3; if?) 1-" “"3? Wynn»:

#33 III.
(”II 1W» 3—“ . .._..t

4“ .bs
‘ T3.
JT

2."

  
  

.
nu-

dll‘sut t.

      

 

.‘,"r!_.'

.IIi.

     
   

“4%

    

‘3 I!“
In”!

'K
o'
p.

 

I‘ ‘. In”: 3‘ “II-1

I?

' c
[#4333393
“I ‘I
#13“;
‘— ._z.
“4%
"I
f
7;?

 

 

 

 

I

’

This is to certify that the I
dissertation entitled '

The Gas Phase Chemistry of Metal and Metal-

Containing Ions with Amines and Alkanes

presented by
Barbara Diane Radecki

has been accepted towards fulfillment
of the requirements for

Ph.D. degreeinflhfimm

 

$1 m

Major professor

Date /$’ 1%; 35-

MS U is an Waive Action/Equal Opportunity Institution 0-12771

 

 

MSU

LIBRARIES
-_

 

 

RETURNING MATERIALS:
Place in book drop to
remove this checkout from
your record. FINES will
be charged if book is
returned after the date
stamped below.

 

 

 

 

 

THE GAS PHASE CHEMISTRY OF METAL AND.METAL-CONTAINING
IONS WITH AHINES AND ALKANES

by

Barbara Diane Radecki

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY

Department of Chemistry
1985

ILJ‘ (5 u.-‘J

ABSTRACT

THE GAS PHASE CHEMISTRY OF METAL AND METAL-CONTAINING
IONS WITH AMINES AND ALKANES
by

Barbara Diane Radecki

The gas phase chemistry of metal and metal-containing
ions with amines and alkanes is presented. The three step
mechanistic sequence, metal insertion/fi-hydrogen shift/
competitive ligand loss, is used to explain the formation of
the observed product ions in these reactions.

The reactions of cobalt and cobalt—containing ions with
secondary and tertiary amines serve to illustrate the
‘utility of'nmtal ions as chemical ionization reagents. The
product ions observed in the reactions of the various amines
are found to be dependent upon the structure of the amine
being analyzed. Molecular weight information is apparent
from ligand substitution reactions.

The metal insertion step as it occurs in the reactions
of Fe+, Co+ and Ni+ with linear alkanes is examined. The
relationship between the nmmal insertion intermediates
formed in the first step of the mechanism and final product
distributions is presented. Preferential insertion of W

into specific C-C bonds in alkanes is noted. A correlation

between preference for insertion into a particular C—C
skeletal bond and the ionization potentials of the alkyl
radicals formed when the C-C bond is cleaved to form R-W-R'

is discussed. This correlation suggests that as the

IP(CnHZn+l) decreases, D (M+’CnH2n+l) increases, implying
that the insertion intermediates involving those in which
the sum of the two M‘l'-R bonds is largest are formed
preferentially.

Finally the chemistry observed with linear and branched
alkanes when one or two ligands such as CO or PF3 are bound
to the metal ion is examined. Ion/induced dipole
interaction energy calculations are presented. The
feasibility of an intact alkane displacing a CO or PF3
ligand is established. Differences in the chemistry of W
and ML+ (n=l,2) are explained on the basis of steric
effects. Parallels between gas phase and condensed phase

alkane chemistry are discussed.

This dissertation is dedicated
to my husband, Pat, my three
beautiful children, Alicia,
Kathleen and Michael and to my

Mother and Father.

ii

ACKNOWLEDGMENTS

First, the author gratefully acknowledges
the assistance and guidance of her research
advisor, Dr.John Allison. John has been
helpful and conscientious as a research
director as well as a good friend. His

efforts have always been appreciated.

I would also like to thank Tony, Mac, D.B.
and Rich for the friendship and support
they provided to me during my ICR years

at M.S.U.

iii

LIST OF
LIST OF
LIST OF
CHAPTER
CHAPTER
CHAPTER

CHAPTER
CHAPTER

- CHAPTER

CHAPTER

CHAPTER
LIST OF

TABLE OF CONTENTS

PACE
TABLES ....................................... v
FIGURES ................................... viii
SCHEMES ................ L .................... ix
1 - INTRODUCTION ........ .~ .................... 1
2 - PRINCIPLES OF OPERATION OF ICR.‘ .......... A
3 - OVERVIEW OF CAS PHASE TRANSITION

METAL STUDIES ........................... 15
a - EXPERIMENTAL .................. ' .......... 26
5 - CAS PHASE CHEMISTRY OF Co(CO)3NO

WITH SECONDARY AND TERTIARY AMINES ...... 30
6 - THE PREFERENCE OF Fe+. co+ and N1+

FOR INSERTION INTO THE C-C BONDS,,,,

OF LINEAR ALKANES ....................... 51
7 - LICAND EFFECTS .......................... 97
8 - CONCLUSION ............................. 147
REFERENCES ................................. 151

iv

Table

10

11

12

13

14

15

LIST OF TABLES

Reactions Observed for Secondary Amines,

Tertiary Amines and Analogous Bthers .......... 31
Product Distributions Observed in

2:1 Mixture Methane to ML“ ..................... 56
Produét Distributions Observed in

2:1 Mixture Ethane to MLn ...................... 56
Product Distributions Observed in

2:1 Mixture Propane to ML“ ..................... 57
Product Distributions Observed in

2:1 Mixture Butane to ML“ ............ , ........ 58
Product Distributions Observed in

2:1 Mixture Pentane to ML“ ..................... 59
Product Distributions Observed in

2:1 Mixture Hexane to MLn ..................... 60
Product Distributions Observed in

2:]. Mixture Heptane to ML!‘1 ..................... 62
Product Distributions Observed in

2:1 Mixture Octane to MLn ..................... 63
Product Distributions Observed in

2:1 Mixture Nonane to ML“ ..................... 65
Product Distributions Observed in

2:1 Mixture Decane to MLn .- ................... 67
Product Distributions Observed in

2:1 Mixture Z-Methylpropane to MLn ............ 69
Product Distributions Observed in

2:1 Mixture 2,2-Dimethylpropane to MLn ........ 70
Product Distributions Observed in

2:1 Mixture 3-Methy1butane to MLn ,,,,,,,,,,,,, 71
Product Distributions Observed in

2:1 Mixture 2,3-Dimethy1butane to MLn ,,,,,,,, 73

16
17
l8
19
20
21

22

23
24
25

‘26
27

28
29

30

Table Page

Product Distributions Observed in
2,2-Dimethylbutane to ML“: ........................ 74

Product Distributions Observed in
2:1 Mixture 2,2,3,3-Tetramethylbutane to ML“. ...... 75

Product Distributions Observed in
2:1 Mixture 3-Methylpentane to ML“ ................ 76

Product Distributions Observed in
2:1 Mixture 2,4-Dimethy1pentane toMLn ............ 77

Product Distributions Observed in
2:1 Mixture 2,3,4-Trimethy1pentane to MLn ......... 78

Product Distributions Observed in
2:1 Mixture 2,2,4-Trimethy1pentane to ML“ ......... 80

Per Cent (1,2) vs. (1.4) Dehydrogenation in
the Reactions of M+ with Deuterated Butane,

Pentane and Hexane. ................................ 84
Preference of Fe+, Co+ and Ni+ for C-C
Bond Insertion in Linear Alkanes .................. 87

Alkyl Radical Ionization Potentials and Bond
Strengths of Various Atomic Cations ................ 90

Results of Polarizability and Energy Calculations
for Linear Alkanes Interacting with a Point Charge107

M+-L Bond Strengths .............................. 109

Bond by Bond Breakdown of U(r) for "Arrow Length"
of 1 A ........................................... 119

Bond by Bond Breakdown of U(R) for "Arrow Length"
of 2 K ........................................... 120

Sum of the Energies of Interaction Between a
Charge and the three C-C Bonds of Butane ,,,,,,,,,, 122

Interaction Energy BetweenM+ and the Carbon
and Hydrogen Atoms of Butane ...................... 126

vi

Table Page

31

32

33

Relative Rates in 2:1 Mixtures of Alkane to

Co(CO)3NO ..................................... 138
Relative Rates in 2:1 Mixtures of Alkane to
Ni(CO)4 ........................................ 139
Relative Rates in 2:1 Mixtures of Alkane to
Ni(PF3)4 ....................................... 140

vii

LIST OF FIGURES

Figure Page

1 Block Diagram of ICR Spectrometer for

Single Resonance Experiment ..................... 6
2 The ICR Cell .................................... 6
3 The Drift Velocity ........ '.' .................... 1o

4 Correlation Between Priority of Attack and
the Ionization Potential Sums ,,,,,,,,, ' ......... 91
5 Plot of -U(r) vs. r, Including D(M+-L) ,,,,,,,, 110

6 Relationship Between Charge Location and

C-C Bonds of Butane ........................... 116

7 Angle and Distance Measurements for Charge
Located at B ............... . .................. 117
A 8 Ligand Substitution Patterns ,,,,,,,,,,,,,,,,,, 127
9 Definition of the Cone Angle ,,,,,,,,,,,,,,,,,, 134

viii

LIST OF SCHEMES

Page
SCHEME 1 .......................................... 38
SCHEME 2 ............................................ 43
SCHEME 3 ....................... ‘. .................... a9
SCHEME 4 ............................................ 82
SCHEME 5 ................................... . ........ 100

ix

Chapter 1

Introduction

Within the past several years, the gas phase chem-
istry occurring between several transition metal ions and
various organic molecules has been reported.1’5 The
majority of this work has involved the characterization
of the reactions of bare (+1) metal ions with polar
organic molecules. Out of this early wOrk came the for-
mulation of a mechanism involving a metal ion insertion
IB-H shift/ competitive ligand loss sequence» This
mechanism has been used to explain the formation of the
majority of the product ions. Its feasibility has been
established through both labellingéi7 and energetic
studies.8

Overall, the significance of this work is threefold.
In terms of the mechanistic implications, these gas phase
reactions allow one to probe the rearrangements of
organic molecules on metal centers in the absence of
complicating solvent molecules. Secondly, thermochemical
information such as bond strengths and metal ion reaction
energetics can be inferred from the results. Finally, in
terms of the analytical applications, these metal ions
can be assessed as chemical ionization (CI) reagents.

1

Note that ideal CI reagents allow for the determination
of both structural and molecular weight information in
regard to the analyte species.

To gain further insights into these gas phase
processes, Ion Cyclotron Resonance (ICR) spectrometry is

used to study the reactions of metal and metal-containing

ions with.amines and alkanes. First, the reactions of'

the electron impact fragments of Co(CO)3NO with select
secondary and tertiary amines will be discussed. This
work represents a continuation in the characterization of
the reactions of metal ions with polar organic molecules.
These results will be compared to the gas phase reactions
of cobalt and cobalt-containing ions with primary amines9
and to the chemistry observed for amines in condensed
phaseslo. Next, the chemistry observed between the
electron impact fragments of Co(CO)3NO, Ni(CO)4. Ni(PF3)z,
and Fe(CO)5 and a series of linear and branched alkanes
will be examined. The discussion of the alkane chemistry
will be divided into two parts. To gain a better
understanding of the insertion process, the first part
will examine the metal insertion step as it occurs in
linear alkanes. An ordering and explanation of the
preference of M+ (M+-Fe+,Co+.Ni+) for the C-C bonds of
alkanes will be established. Finally, changes observed

in the chemistry of bare (+1) metal ions with alkanes

when ligands such as CO, NO or PF3 are bound to the metal
will be discussed. Parallels to condensed phase
chemistry will be made.

Overall, the product ions formed in these gas phase
studies yield both.structura1 and molecular‘weight infor-
mation. Structural information becomes apparent from a
knowledge of the mechanism of interaction between metal
ions and organic molecules. Molecular weight information
can be obtained from ligand displacement reactions.
However, before presenting the results, the basis for
their interpretation must be established ”in an
understanding of the basic principles of operation of an
ICR spectrometer. This understanding, coupled with a
knowledge of the pertinent research already carried out
in the areas of both gas phase and condensed phase
organometallic chemistry, will enable one to extract .the
structural, mechanistic and thermodynamic information

which is inherent to the results.

Chapter 2

Principals of Operation of ICR

ICR 'spectrometry' ‘was developed as a ‘mass
spectrometric technique for the study of bimolecular
reactions in the gas phase (at low pressures. These
bimolecular reactions proceed as a result of ion-neutral
collisions which occur when the pressure is increased
above a low "collision free" pressure of ~2x10"6 torr
(for a reaction time of ulmsec). The ionic products of
these reactions manifest themselves as "new" peaks in the
mass spectrum. The study of these reactions under "mass
spectrometric conditions" allows for an examination of
reactive systems in the absence of complicating solvent
interferences.

There are several features of ICR spectrometry which
make it particularly well suited for the study of these
bimolecular or ion/molecule reactions. First of all, the
ICR is operated routinely at pressures of less than
1x10'5 torr. These low pressures, coupled.with reaction
times which are on the order of milliseconds,ensure that
the predominant reactions taking place are bimolecular in

nature. Furthermore these bimolecular reactions are
4

5

believed to occur within an order of magnitude of the
collision frequency (~1xlO"10 cm3 molecule ’ls'l) which
implies‘ that the observed reactions proceed with
essentially no activation energy. Furthermore, the
observed processes are assumed to arise from exothermic
or thermoneutral processes. This assumption allows one
to determine limits on heats of formation of products and
on bond strengths. Finally, unique to this technique is
ion cyclotron double resonance which provides a means of
positively identifying predursor ions of reaction
products.

Figure 1 represents a block diagram of an ICR spec-
trometer. Ion formation, focussing, acceleration and
detection occur in the ICR "cell". The cell, illustrated
in Figure 2, consists of three sections which include an
ion source, analyzer, and collector. The ICR cell is
housed in a stainless steel vacuum chamber and is

' located between the poles of an electromagnet.

Thorough reviews of the theory involved in ICR spec- '
trometry have been presented in the literature11'16. A
brief review of these principles will be provided here to
describe the dynamics of charged particles in electric
and magnetic fields.

Ions are formed in the source by electron impact
(70eV). Due to the presence of a magnetic field, B, the
iOns are constrained to circular orbits in a plane which

is normal to B. For an ion of mass m, acceleration a,

 

 

DRIFT‘E
TRAPPING
VOLTAGES

 

 

 

 

IAMPLIFIER

 

 

 

  
   

 

 

 

MODULATION

    

 

 

 

 

 

 

 

 

 

 

 

 

REFERENCE POWER
lumen , ENERGY 5.
osc . mlssmn . i SUPPLY _
‘ * CURRENT ' - 1
EASE- l CONTROL SNEEP -
SENSITIVE n RECORDER CONTROL
‘ DETECTOR 1 i J L ‘

 

Figure 1.

Block Diagram of ICR Spectrometer for Single Resonance
Experiment.

Source :Analyzer ' Collectorl

rf from . ' l
Marginal Oscillator I . l
l
l
- l
I

Drift Plate
Ukudyzer)

t

on Current

u
t
-

Drift.Plate
(Source)

Collector -flll
y I

 

1&7?~f

 

 

 

T;apping‘Voltage
Figure 2.

The ICR Ce1111

7

charge q, and velocity V which is perpendicular to 'B', the
force acting on the ion is presented in equation (1).
F = ma - q(§LE;§).e 923' (1)
(where V is the velocity :ormal t: B, 117] 8 v)
This force is perpendicular to B and v. Furthermore, due
to the Circular motiOn of the ions
a - PE (2)
T
where r is equal to the radius of the ion path.
Substituting equation (2) into (1) and rearranging, an
expression for the angular frequency, 3, of the ion in
rad/s is obtained (equation (3)). r
ma - my} - qvb
“c r' X ' SE (3)
1“ m
we is the natural cyclotron frequency of an ion in the
presence of a magnetic field. For example, the cyclotron
' frequency of a Co+ ion in a magnetic field strength of

1.0 Tesla is:

- S a
- he a 1.6x10 19C x 1.0? :- 2.6x10 2
”c I? 6.28 x 59u x 1.66x10"'kg7u

Note that cc is independent of velocity. Furthermore,
for a constant “c. the mass to charge ratio of the ion
varies linearly with B.

The motion of ions in the presence of a magnetic

field is not restricted in the plane parallel to B.

Thus, to constrain the ions in the cell in the XY plane.

8

a trapping potential, Vtrapv is applied to the trapping
jplates. The electric field established by this potential
is parallel to B and restricts the motion of the ions in
the plane perpendicular to B'by creating a potential
energy minimum near the center of the cell. Therefore,
when detecting positive ions, small positive voltages
(<O.5V) are applied to plates 2 and 4 in Figure 2.

To move the ions from the source into the analyzer,
a potential difference, Vdrift, is applied between the
plates which are located above and below the electron
beam (plates 1, 5, 3 and 6 in Figure 2). The electric
field .produced by the drift plate -potentials is
perpendicular to the magnetic field (Ex B). Under these
conditions, the ions move at right angles to both E and
B. The force acting on the iOn under these circumstances
is given in equation (4).

F “21.2.! e of +5123 (4)
at c

Assuming that E and B are constant, the velocity, V, .
of the ion consists of two components,‘V‘, associated
with the cyclotron motion and‘V'd, the drift velocity.
The equation for velocity is presented in equation (S).

7-? +3? -'\T' + cExB (5)
d '3?-

(Note that 73 is equal to (cExB)/B2). Substitution of
equation (5) into equation (4) yields a new expression

for F: F = m E8 4167' x j; (6)
. at c

9

From equation (6), it is evident that v‘ is independent
of E. Therefore, one can conclude that this component of
the velocity is associated with the cyclic motion of the
ion occurring about the magnetic field lines, with the
ion moving at its cyclotron frequency. The other
component of velocity Va is perpendicular to both Eand
'B'. The drift velocity can be simply evaluated12 by
equation '(7) where adrift is the magnitude of the

electric field. The electric

Vd I" Egrift (7)
B

field can be obtained by dividing 'the potential
difference between the drift plates by the distance
separating the plates. Note that the drift velocity is
independent of mass and charge. Typically, the distance
between the plates is 1.0 cm and the potential difference

between the plates is 0.5 V at B - 1.0T. The drift
- velocity of an ion calculated by substitution of these
parameters into equation (7) is 50 m/s. Considering that
the distance between the electron beam (point of

ionization) and the far end of the analyzer region is

approximately 8.6 cm, we can determine that an ion

spends about 2x 10'3 s in the cell at a magnetic field
strength equal to lT. Furthermore, the total distance
travelled by the ion during this time period is

approximately 1m. An illustration of the drift velocity
is provided in Figure 3.

10

U” .. mm

 

 

U W
m ABSENCE _ 9
or s 3 Va

Figure 3. Drift Velocity.

A marginal oscillator with phase sensitive detection
is connected to the top and bottom drift plates in the
analyzer region. The marginal oscillator serves as the
detection system for the ions once they pass into the
analyzer region. The output voltage of the marginal
oscillator is varied sinusoidally at a frequency which is
dependent upon the inductance and capacitance of a tank
circuit. The ICR cell serves as a capacitive element in
the tank circuit of the marginal oscillator. Detection
of an ion occurs when we of the ion is the same as the
frequency of the marginal oscillator. When these
frequencies match, the ion is said to be in resonance
with the applied field. When in resonance, an ion absorbs
energy from the radio-frequency, rf, field produced by
the marginal oscillator and in doing so generates an
electric signal in the circuit.

The power absorbed by an ion from the rf field in the

ll

resonance region can be determined. Since the rf signal
is an electromagnetic wave, it has an amplitude of trf.
When interacting with a moving charge, the rf signal can
be resolved into two contrarotating circular components,
each with an amplitude of erf/Z. When an ion is in
resonance, its orbit tracks only one of these circular
components. Thus, the frequency of the ion is the same
as that Of the rf field. When in resonance, the force

acting on the ion is given by equation (8).

F - me - erf q (8)
T
Solving equation (8) for a, equation (9) is obtained.
argues (9)
Zn:

Integration of equation (9) with respect to time yields
an expression for velocity. Therefore, after an ion has

been in resonance for a time period of t, its velocity is

t
v-fadtsczggt (10)
O .

2m

 

The instantaneous power absorbed,A(t), by an ion is a_
product of the force acting on the ion (equation 9) and

its velocity (equation 10).

A(t) - Fv - rfz qz t (11)
‘tm
To calculate the average power absorbed by an ion during

the time, ‘t (O < t < t), the ion spends in the detector,

equation (11) is integrated with respect to time from 0

toT. 1

A 1' 2 Mt s l f A(t)dt a c 2 21 (12)
( ) ) ‘1’ o ‘lé—‘mq.

12

Equation (12) yields the power absorbed by one ion at
resonance. Power absorption is inversely proportional
to mass.

During the course of a typical single resonance
experiment, the magnetic field strength is scanned from
0 to 1.8 T (181(6) while the frequency, ”m.o.» of the
marginal oscillator is held constant and chosen such that
one atomic mass unit corresponds to a change of 0.01 T

(100 Gauss):
,m o . “m o g 1,6021x10-19c x 0.01 T - 1.5357x105Hg

2n 6.2832x1uxl.6604x10-27Kg/u

 

 

Recall that from equation (3), the m/q ratio varies
linearly with B at a constant frequency. Thus by setting
the marginal oscillator frequency at 153.57 kHz and
scanning B, each ion is brought into resonance, one at a
time, with the marginal oscillator and detected. A plot
of absorption versus B, which is linear in mass, is
' obtained.

To enhance the signal-to-noise ratio, SIN, the ion
signal is modulated by applying a 27Hz square wave on one
of the trapping plates (plate 2, Figure 2). When the
potential on the plate is negative, all of the positively
charged ions are swept out of the cell. As a result, the
rf level of the marginal oscillator varies at this same
frequency. Using a lock-in amplifier in conjunction with
the marginal oscillator, the SIN is greatly improved.

Under these circumstances, the marginal oscillator is

13

modulated at the same frequency as the square wave on the
trapping plate. Thus, the marginal oscillator generates
a signal only when ions are present in the cell.
Theoretically, a marginal oscillator can detect as few as
14 ions.12

Following the analyzer is a collector region. The
collector plates are all connected to ground. When the
ions drift into these plates, an ion current is
produced. The total ion current can be measured using an
electrometer. The total ion current is on the order of
10"12 Amperes. '

Ion-neutral collisions resulting in the formation of
products occur when the pressure in the cell is increased
above a "collision free" pressure of ~2x10"6 torr. To
identify the reaction sequences involving the produet
ions formed in ion—molecule reactions, ion cyclotron
double resonance (ICDR) is utilized. This technique is
unique to ICR spectrometry and it allows one to
unambiguously assign the precursors of a product ion.

During ICDR, the frequency of a variable rf
oscillator is scanned in the source thus adding energy to
potential reactant ions. At the same time, the product
ion under consideration is kept in resonance with the
marginal oscillator in the analyzer. Thus, its intensity
is continuously being monitored. When the we of a

potential reactant ion equals the frequency of the

14

secondary oscillator in the source, the ion absorbs
energy and is swept from the cell. If the ejected ion
corresponds to a precursor, a decrease in.the intensity
of the peak representing the product ion being monitored
is observed. Since the magnetic field strength is kept
constant during ICDR, the following relationship holds:

3 ' “product m‘product C 3 "precursor mprecursor C (13)

q q

Therefore, the mass of the precursor is readily

 

 

determined from equation (13) when the frequency ratio is

known.

” roduct ' mprecursor
P

 

“precursor mproduct

Detailed descriptions of the ICDR experiment 11,12 and
thorough discussions concerning the interpretation of
double resonance results16o17 can be found in the

literature.

Chapter 3

Overview of Gas Phase

Transition Metal Studies

Early work13a19 done with transition metal compounds
in gas phase studies inyolved volatile carbonyl
containing species, M(CO)x (where M = Ni,x - 4; M - Fe, x
- 5; M 8 Cr, Mo, W, x - 6). Transition metal carbonyls
were of interest due to the fact that metal carbonyls
play an important structural role in many organometallic
complexes. Gas phase studies allowed for an examination
of the intrinsic properties of a metal complex in the
absence of complications from the interference of sol-
vent. The fragmentation patterns observed for these metal
carbonyls following ionization via electron impact was
used to determine the ionization potential of MCOx and
the appearance potentials of the remaining fragment ions.
Using this information, the energetics involved. in
ionization and dissociation of the ions was determined.

The first ion/molecule reactions observed for
mononuclear organometallic compounds were reported by
Schumaker and Taubenest.20 They observed peaks
corresponding to the fbrmation of bimetallic ions in

their mass spectrometer. They used reaction (1) to
15

16

account for their results

NICp+ + Nisz 4 Nich3+ (1)
(where Cp 8 C5H5).
Another early investigation by Miller21 revealed that
polynuclear metal complexes were formed from
ion/molecule reactions in CpV(CO)4, Cr(CO)6, Ni(PF3)4,
Cr(06H6)2, CpCrCONO and (C6H6)Cr(CO)3. Furthermore,
MUller reported ion/molecule reactions occurring between
ionized transition metal complexes such as CpCo(CO)2+ or
CpNiNO‘l’ and a number of c and a donor molecules including
H20, N113 and benzene. Emerging from these investigations
was the realization that ion/molecule reactions of
organometallic complexes in the gas phase yielded useful
information regarding the formation and stability of
metal-metal and metal-ligand bonds.

During the time in which Muller was conducting his
studies, the ICR technique was being developed. The
utility of ICR in the study of gas phase ion/molecule
reactions was soon realized.22 Foster and Beauchamp were
the first to use this technique to study the reactions of
ions formed by electron impact on Fe(CO)5 with neutral
Fe(CO)5 and in binary mixtures with molecules such as
CH3F,H20,NH3 and HCl.23 They observed the formation of
binuclear iron species. Using ICDR, they were able to
unambiguously determine that this binuclear iron species

‘was formed from Fe+ and Fe(CO)5:

Fe+ + Fe(CO)5 9 Fe2(CO)4+ + co

17

Results from the experiments involving binary mixtures
enabled them to determine bond energies and the relative
rates of ligand substitution in Fe(CO)n+ (n -= 0-5). From
these studies, it was evident that ICR could serve as a
useful tool in the characterization of the thermodynamic
and intrinsic chemical properties of ions and molecules
in the absence of solvent.

In addition to the formation of polynuclear clusters
and ligand substitution reactions, many transition metal
complexes were found to exhibit rich chemistries with
organic molecules in the gas phase.2“"7.5 For example,
Ridge g£_gl. were the first to report the formation of
metal to carbon bonds in the reactions of the electron
impact fragments of Co(CO)3NO and Fe(CO)5 with alkyl
halides7 and alcohols}.6 Tricarbonylnitrosylcobalt(O)
and other metal carbonyl compounds were chosen for study
because of their volatility and because CO and NO serve
as important ligands in transition metal chemistry. The
formation of metal to carbon bonds was explained by the
oxidative addition of 11+, formed from electron impact on
MLn (M 8 Fe, Co,Ni,Cr), to C-X bonds (where X = halide or
OH) in saturated polar organic neutral molecules.

Essentially all of the polar organic molecules that
have been studied to date react with transition metal
ions such as Fe+, 00+ or Ni‘l' in the gas phase. Reaction

(2) illustrates the dominant mechanism believed to be

18

operative in these gas phase studies. This mechanism

involves a metal insertion IB-H shift/ competitive ligand

loss sequence.

c3+ >—c1 —> c3—c1—>/||--clJ—CI —>—
H H

/ll--Co”--CIH /|] —--c&‘ + HO! ‘2’
Cot-eon + J

-

Thus, Co+ inserts into the polar C—Cl bond followed by
the shift of a B-hydrogenZB. Isopropyl chloride is
"rearranged" on the ionic metal center into HCl and
propane. HCl and propane then compete as ligands on the
metal. Similarly for alcohols, Ridge and coworkers
reported that the metal ion inserts into the C-OH bond to
eliminate H20.1 In general, Fe'l', Co"’ and Ni+ were all
found to dehydrate alcohols and dehydrohalogenate alkyl
halides},6

Many other polar organic molecules have been studied
with various metal containing compounds. Ridge and

coworkers also studied mixtures of T1014 with small

olefins30 and oxygen containing organic compounds.33

These reactions were of interest due to the catalytic

19

participation of TiC14 in the Ziegler-Natta scheme for
the polymerization of olefins in solution. Ridge et al.
found that the electron impact fragments of TiClz,

(Including T1+. TIC1+. T1C12+, TIC13+ and TIC14+)
exhibited specific chemistry with organic compounds.30:33

Furthermore, different chemistry was observed depending
upon how many chlorines were bound to Ti‘l'. For example,
T1012+ and TiCl3+ were both found to eliminate HCl and
smaller olefins from larger olefins, Ti‘l' was found only
to eliminate H2, and TiClA'l’ was unreactive.

Beauchamp m1. have reported the decarbonylation
of aldehydes by (n5-C5H5)Ni+ in the gas phase.27 The
mechanism they proposed involved a reaction intermediate
such as

~ chit-co
RH

from which the competitive elimination of RH or CO could
occur. Decarbonylation of aldehydes in solution and on
solid surfaces has been reported in the literature.76
Beauchamp and his coworkers at Cal Tech have made
other significant contributions to the area of gas phase
transition metal chemistry. They have studied cobalt and
chromium carbenes2»“2 in an effort to characterize these

metal carbenes in the gas phase. Transition metal

carbenes are of interest because they may serve as

intermediates in the Fischer-Tropsch synthesis and in

20

olefin metathesis. Furthermore, Beauchamp eg_g1. have
examined the basicity of iron pentacarbonyl,2“
ferrocene25 and nickeloceneZGin the gas phase using ICR
and have reported the proton affinities of a number of
organotransition metal complexes!“ They have determined
the metal-ligand bond dissociation energies of some
n-donor ligands to cyclopentadienylnickel cation79 and
have described the methodology involved in the
determination of metal-hydrogen, metal-carbon and
metal-ligand bond dissociation energies in (the gas
phase.37o77-78 The purpose of these studies was to gain
information relating to the intrinsic reactivity of
transition metal complexes in the absence of solvent.
Freiser and' coworkers at Purdue reported the
cleavage of C-C, C-H, and C-0 bonds in the gas phase
reactions of Fe'l’ with ketones and ethers.’t8 They
observed the formation of alkyl-, acyl- and alkoxide Fe‘l'
species arising from oxidative addition reactions. They
also observed decarbonylation for a few small ketones.
Dehydrogenation was reported for unbranched ketones and
reductive elimination of methane occurred in ketones
which were branched at the a carbon. They also reported
the gas phase chemistry of Cu+ with esters and ketones.38
The gas phase chemistry of Cu+ with alkyl
chlorides“0 was reported by Staley M. Halide

transfer onto Cu+ followed by dehydrochlorination was

21

observed. The mechanism they used to account for their
results once again involved oxidative addition of Cu+ to
the polar carbon-halide bond.

Work done in the Allison lab at MSU has involved the
characterization of the reactions occurring between iron
or chromium containing ions and polyethers.6o The
chemistry occurring between the 70eV electron impact
fragments of Co(CO)3NO and various bifunctional organic
molecules has also been reported.5’59 These results
indicated that bifunctional molecules exhibit (a much
richer chemistry than monofunctional compounds due to the
interaction of the metal ion with two polar functional
groups instead of one. Also, the reactions of
nitroalkanes and cobalt-containing compounds have been
reported recently by Allison g£_g1.72

From the brief overview presented above, it is
evident that the chemistry of many transition metal ions
and transition metal complexes has been studied. The
results‘ illustrate the rich chemistry occurring with
various types of organic molecules in the gas phase.
Furthermore,the mechanism which has been found to be
operative in these reactions is one involving an
oxidative addition of the metal ion followed by the shift
of a hydrogen and competitive ligand 1033.697

In an effort to characterize the reactions of metal

ions with select nitrogen containing compounds, a study

of the reactions of cobalt-containing ions with secondary

22

and tertiary amines was carried out as the first part of

this Ph.D. project. This particular part of the project
represents a continuation of work done earlier with
primary amines by this author for a Master's thesis.

In investigating the chemistry of metal ions with
organic compounds containing C-N bonds, it was determined
in our lab that nitroalkanes react with Co+ to yield
products indicative of initial 00* insertion into the C-N
bond. For example, reaction (3) was reported for Co'l’ and

2-nitropropane. ,
Co+ + >—N02 -) y cO+-N02 a Co-HNOz‘l’ + J (3)
H ,

The trend established by the earlier work done with polar
compounds (reaction(l)) and by reaction (3) suggested
that primary amines would undergo similar reactions.
Therefore, for the reaction of Co+ with isopropylamine,

reactions (4) and (5) were expected:

Co+ + >-NH2 -» 2—Co‘l’ -NH2 -I: CoNE3+ +J| (4)
H
C0122“ + NH3 (5)

However, products indicative of Co+ insertion into the
C-N bond of primary amines were never observed. Instead,
only products indicative of metal ion insertion into C-C
and C-H bonds were observed. Thus, Co+ reacted with
primary amines to induce the elimination of H2, small

alkanes and small alkenes.”

23

These results were not totally unexpected since they were
similar to results observed by Mfiller for metal-containing
ions such as C5H6V+ and C5H5NiNO+ with diethylamine
and dimethylamine. 30' 31 He reported reaction products
corresponding to the elimination of one or more molecules
of hydrogen or of small hydrocarbons.

In Chapter 5 of this dissertation, the chemistry of the
Coma); (x-o-z) and Co(CO)NO+ (Ran-3110113 formed by
electron impact on Co(CO)3NO with secondary and tertiary
amines will be discussed. It will become apparent, that
the reactions observed for secondary amines parallel those
observed for primary amines and those observed by MUller,
whereas the products formed in the reaction of metal ions
with tertiary amines are suggestive of a mechanism
involving metal 'ion insertion into C-N bonds. These
results will be compared to those observed in condensed
phases.

Chapters 6 and 7 will be devoted to the gas phase
chemistry of metal—containing-ions and alkanes.43 In 1979,
first row transition metal ions were reported to interact
with alkanes in bimolecular gas phase reactions. These
initial results indicated that it was energetically
feasible for ions such as Fe*, Co+ and Ni+ to insert into
any of the skeletal bonds of most alkanes in an
exothermic process. 43:52:57 For example, in the
reaction of Col"with n-butane, products indicative of

initial insertion into both the terminal and internal C-C

24

bond were observed:‘3 (reactions (6) and (7))

 

co.
0
CH3CH2CH2CH3
CH3-C0‘433H7 Csz-COI-Czfls
l 3
1 11-11 mm 1
COC3H5’ CoCzl'l‘I
. 0
CH4 (bl C2"; (7)

Twelve percent of the products of this reaction were
attributed to CoC3H6+. The CoC2H4+ product represented
fifty-nine per cent of the products. The remainder of the
products (twentyenine per cent) arose through hydrogen
elimination resulting in the formation of CoC4H3.

The product distributions observed in the reaction of Co+
and butane, along with subsequent studies of the mechanisms
involved in H2 elimination 51052 suggested that the
formation of products from intermediate 3 are favored over
intermediate,1. A number of attempts were made to explain
why some products are formed in greater abundance than
others. (These attempts will be discussed in Chapter 6).
We believe that the insertion process plays a dominant
role in determining the final product distributions.
Based on work performed in our laboratory, along with

product distributions reported previously in the litera-

25

ture,6:35:43o52.53.57 a direct correlation between the C-C
bonds of normal alkanes preferred by Hi'for insertion and
the ionization potentials (IP's) of the alkyl group bound
to M+'(M=Pe. Co, Ni) upon its insertion is reported and
rationalized in Chapter 6. It is suggested that this

correlation is indicative of reactions where the most

stable insertion intermediate is formed preferentially thus
leading to the greatest abundance of products. Therefore,
the distribution of insertion intermediates reflects their
relative stabilities.

Finally, Chapter 7 will examine the changes which occur
in _the chemistry of Fe Co and Ni with straight-chained

and branched alkanes as up to two ligands (including
CO, NO or PF ) are bound to the metal. Previous stu-
d1686136043052I53I57 have focussed on the reactions of the
bare (+1) metal ions with alkanes. However, in condensed
phases, the organometallic reactions of interest involve

metal complexes.35137

 

In the gas phase, with the addition of ligands, a number
of pathways may be taken, depending upon the number and
nature of the ligands. These pathways will be diScuss-
ed and compared to those observed in condensed phase

chemistry.

Chapter 4

Experimental

This chapter will consist of three parts. The first
part will describe the specifications of the ICR
spectrometer used in this study. The second part will
list the chemicals utilized along with their
manufacturers. The last part. will briefly describe how

the data were acquired.

ICR Specifications

All experiments were performed on an ICR
spectrometer of conventional designs“ which was
constructed at Michigan State University. The ICR cell
consists of a source, analyzer and collector region. The
overall dimensions of the cell are 0.88 inch by 0.88inch
by 6.25 inches. The source region is 2.00 inches long,
the analyzer region is 3.75 inches long, and the
collector is 0.50 inch long. Ions are formed in the
source by electron impact with 70eV electrons. The
electron filament is emission regulated. Both the
filament controller and plate voltage controller for the

ICR cell were designed and constructed at MSU. The

26

27

instrument is operable in either the pulsed or drift ICR
modes. All datafor this study were obtained under
normal-drift-mode conditions by using trapping voltage
modulation and phase sensitive detection. A marginal
oscillator detection system based on the design of
Warnick, Anders and Sharp85 was used. Ion cyclotron
double resonance experiments11 were performed using a
Wavetek Model 144 sweep generator as the secondary
oscillator. The ICR cell is contained in a stainless
steel vacuum system and sits between the polecaps of a
Varian 12 inch electromagnet (1.5 in. gap). A varian
V-7800, 13 kW power supply and a Fieldial Mark I Magnetic
Field Regulator are used to operate the electromagnet.

The instrument is pumped by a 4-inch diffusion pump
with a liquid nitrogen cold trap and an Ultek 20 liter
per second ion pump. Samples are admitted by Varian
951-5106 precision leak valves. The approximate sample
pressures are measured with a Veeco RGlOOO ionization

gauge.

Chemicals

Tricarbonylnitrosylcobalt(O) and tetrakis(trifluoro-
phosphine)nickel were obtained from Alfa Inorganics. Iron
pentacarbonyl, 2,2,3,3-tetramethylbutane and butane were
purchased from Aldrich Chemical Co. Piperidine, pyrr-
olidine, tetrahydrofuran, diethyl ether, diethylamine

and triethylamine were obtained from Chem-Service Inc.

28

Tetrahydrofuran-dg was purchased from Norell Chemical
Co., Inc. Ethylbutylamine, propane, Z-methylpropane,
2,2-dimethylpropane, butane, 2,3-dimethy1butane, 2,2-di-
methylbutane, pentane, 3-methy1pentane, 2,4-dimethy1pen-
tane , 2 , 3 , 4-trimethylpentane , 2 , 2 ,4-trimethy1pentane ,
hexane, heptane, octane, nonane and decane were obtained
from Fluka Chemical Corporation. Methane and ethane were
purchased from Matheson Company, Inc. All samples were
degassed three times by freeze-pump-thaw cycles. After
degassing, the vapor pressure of a liquid sample (typi-
cally around 6 torr) was admitted into an evacuated glass

bulb and was used without further purification.

Data Acquisit1gg

Data were acquired in the following manner. High-and
low-pressure (1x10"’5 torr versus 1x10"6 torr) spectra “of
each compound were obtained. Any ions formed by
ion/molecule reactions in the organic compound alone and
in the metal compound alone were determined. Next, a 1:2
mixture (by pressure) of metal compound to alkane or
amine was introduced into the cell at a total pressure of
~8xlO"6 torr. Spectra were taken up to a mass to charge
ratio (m/z) of 260 and all product ions characteristic of
the mixture were identified.

Conventional double resonance techniques were used

29

to identify the ionic precursors of the ion/molecule
reaction products. (Note, that if A‘l’ -9 B‘l’ and B'l’ -) C+,
then the precursors of CI as determined by ICDR will be
A+ and BT). Branching ratios were calculated to determine

product distributions and were taken to be accurate to

within 1 101. The reported reactions are facile and occur
within an order of magnitude of the collision frequency

(k s 1 -10x10'10 cm3 molecule‘ls’1)o An approximate

lower limit for the rate constant for processes not
observed is 5 x 10'11 cm3 molecule‘ls'l.

(Note that when working with metal carbOnyls in
these experiments caution must be used in interpreting
results and suggesting product compositions. This is due
to the fact that both Csz, and CO have a nominal mass of
28u. Yet, following a careful analysis of all precursors
of each product and a comparison of these results to the
results of similar studies, the processes reported in
this dissertation are thought to represent a

self-consistent set).

Chapter 5
Gas Phase Chemistry of Co(CO)3NO

with Secondary and Tertiary Amines

In this chapter, the reactions of the 70eV electron
impact fragments of Co(CO)3NO (Co+,CoCO+,CoNO+,Co(CO)2+,'
CoCONo+, Co(CO)2NO+ and Co(CO)3No+) with diethylamine.
pyrrolidine, piperidine, ethylbutylamine and triethyla-
mine will be presented. A summary of these reactions, in
terms of the neutral losses, is presented in Table 1
along with the branching ratios. Also included in Table l
are the reactions observed for a few analogous oxygen
containing compounds. (Note that labelled tetrahydrofuran
was used to unambiguously identify products because in

the reactions of unlabelled tetrahydrofuran, loss of

- ‘ (H20+CO) could not be distinguished from loss of (C2H50)

(CZHA and CO both correspond to 28u)). The discussion of
these results will consist of four parts. The first part
will examine the formation of CoH as a neutral product
and will present evidence indicating that metal-induced
H2 elimination from amines occurs with Co"’ insertion into
an 11-11 bond as a first step. The second part will discuss
intermediates involving metal-amide bonds. A comparison

to the chemistry occurring in condensed phases will be

30

31

3:38 .3393: oz u .z... c

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

a... N N N N.N,. N N... .e.~.8.8
N N N N N N N .SN.8.8
... ... ... .1 .1 ... _ ... .288
8. . ..3.....-x.8.8 l a . .9288 "2°38... 8.28538
.33 SN .83 SN .83 SN .83 8N .83 SN 53 SN
.83 8 ...3 8 .53 8 .83 .. 8 .83 8 8.: 8 .23 8 .N.8.8
.23 :8 .83 8
.83 8 .83 8 . N8
.83 8 . szNu .23 8 .33 8 . N... .33 8 .83 8
.33 8 . ._.N8 .33 N=N .83. 8 . 8N8 .33 8 .33 8 .83 8 . N:
.23 oxNN .83 8 . N: ....3 N: .23 8 . 8N: .83 N: .53 8. . {Nu 3.3 N: .88
.83 :8
.23 BN8 .83 :8
.83 .8 .23 ._.NO .33 .8 see as?
.83 {N8 .83 ._.Nu .83 :8 .83 9N8 .83 {NO
.83 ...N8 .33 N: ...3 NzN .83. 8N: .8... =8 .83 {N8 .3: =8 .8
\IV\/ \)2\/\/ . D a 0 \IO\/ \In\/ .953: S. 2.38...
z I u o I 2823.

.33... .32.! 838...... 3. ‘88.... 25.883 from 88...: 2. 8...... D.B.... .855 8888 to. 8:88 23.88..

 

.~ 03-»

32

made. The third part will examine the differences with
the addition of CO and NO ligands on the metal. ‘The last
part will compare and contrast the reactions of cobalt
and cobalt-containing ions with analogous oxygen
containing compounds.

The data presented in Table l coupled with the
results previously reported for primary amines71 suggests
that both the N-H bond and the a-C-H bond of amines are
attacked by Co+. Attack of the o-C-H bond results in the
loss of the CoH neutral and formation of the (amine-1T+
ion. Insertion of Co+ into the N-H bond leads to H2
elimination. '

Consider first the formation of CoH. Co‘l' reacts
with all of the secondary and tertiary amines in Table l
to form COH. Co+ also reacts with methyl-, ethyl-,
propyl—, isopropyl-, butyl-, isobutyl-, . and
sec-butylamine to form CoH.71 It does g2; react with
tert-butylamine to form CoH. Thus, this process occurs
for triethylamine (which does not contain an N-H bond)
and for methylamine (which contains only a-C-H bonds and
N-H bonds) but not for tert-butylamine which has no
a-C-H. These results suggest that the reaction is
predominantly an a-hydride abstraction. A general

mechanism for this process is provided in reaction(8).

33

 

R R’ l
\l ,
Co’oR-CH-NH2—9 C-H‘--Co
II ' I
R NHZ. R,.
3 —-> [Cu-Cori (8)
3 RN“:
\ -Cd-H -CoH
1 ,
L - 91H

 

I

R-g-NHepiNH
2 R

Thermodynamic conclusions based on this reaction type are
difficult to make due to the fact that accurate
thermodynamic data on the possible (amine-l)+ ions are
not available. Also, bond strengths for Co-H reported in
the literature have values ranging from 40 to 70
kcal/mol.88 However, the driving force of this reaction
may be the stability of the product ion. For example,
when Co'l' reacts with CH3NH2 to abstract a hydrogen, CH4N+
is formed. While the structure of CHAN” formed in this

manner is unknown, available thermochemical data32

suggests that the protonated amide, CH4N+, is a highly

stable ion.

34

Table 1 also shows that H2 is eliminated in the
reactions of Co+ with piperidine and ethylbutylamine.
Furthermore, it is evident that Co+ does not react with
triethylamine to induce the elimination of H2. Note that
all of the primary amines previously studied, except
tert-butylamine, reacted with 00+ by elimination of H2.
Overall, .these results suggest that H2 elimination occurs
following Co+ insertion .into the N-H bond. H2
elimination is illustrated for ethylbutylamine in
reaction(9). Several observations support this
mechanism. First, H2 elimination occurs in the reaction
of 00+ and methylamine in which only a CtN double bond
can be formed (for a 1,2 elimination). Also, H2
elimination is not observed for tert-butylamine. This
negative result is expected based on the proposed
mechanism since, following Co+ insertion into the NeH
bond, no 9-H is available to shift. Furthermore, in

terms of primary amines, labelling experiments with

- +-
co" + (angst-Icarus -- cal-15": Co I-l

[cs-H
CH30H20H2 H (9)
H N -I-I c H N
c2 5.....co" .3 2 5II---CO*("
c“ OH H

I I
mam-12c"2 °"s°"'2°"'2

35

CszNDz are consistent with the proposed mechanism in
that HD' elimination predominates. Finally, Table 1 shows
that H2 elimination occurs for secondary amines which
contain an N-H bond but not for triethylamine which has
no N-H bond. If H2 elimination occurred following
insertion into the a-C-H bond, then H2 elimination
resulting in the formation of a C-C double bond would be
expected for triethylamine. Since this is not observed,
this suggests that insertion into the NBHIbond leads to
H2 elimination.

Note that H2 elimination is not observed in the
reaction of Co"' and pyrrolidine, a secondary amine
consisting of a five-membered ring. Based on the
proposed mechanism, H2 elimination would result in the
formation of a double bond in the ring. The 120° bond
angles associated with a double bond are much larger than
the 105° angles associated with a stable five-membered
ring (based on cyclopentane).39 Compression of these 120°.
bond angles to satisfy the angle requirements of the ring
would result in tremendous strain. A product ion
involving this strained configuration would not be
expected and was not observed.

Interestingly, in the case of piperidine, up to three
molecules of hydrogen can be eliminated to form a

"metal-pyridine" complex (reaction (10)).

36

N H -H -H -H N
e: + 0 —»< Eu-cd-H 4.448.133 (10)

The driving force of this reaction may be the formation
of a conjugated system which can complex strongly with
the cobalt ion. Note that the proposed product structure
involves pyridine which is known to be very stable.
Similarly, in primary amines, when the skeletal chain
consists of four or more atoms, the elimination of two
molecules of hydrogen is observed. This results in the
formation of the metal-butadiene like complex illustrated
in reaction(ll).71 Again, the driving force of
reaction (11) may be the formation of the butadiene-like
ligand which can strongly complex to Co‘l'.
cs“ + m" TEL-m ilk-n m,‘

47' \tl
Co NH C 0 NH

~

A comparison of the chemistry of diethylamine and
ethylbutylamine shows that Co+ reacts with
ethylbutylamine to form a larger variety of products than
when it reacts with diethylamine. The only product
formed in the reaction of Co+ and diethylamine is CoH.
Products indicative of initial Co‘l’ insertion into C—C or
N-H bonds of diethylamine are not observed. These
results suggest that the products observed for

ethylbutylamine arise from initial Co+ insertion into the

37

C-C bonds of the bggyl portion of the molecule. These
reactions are illustrated in Scheme 1 (omitting, for
simplicity, H2 elimination and hydride abstraction).
Thus, Co'l’ inserts into a C-C bond, afi-hydrogen shifts
onto the metal and finally two molecules compete as
ligands on the metal. Previous work indicates that the
bond strengths between Co"’ and a variety of ligands can
be correlated to the proton affinities of the ligands.5
Therefore, the larger the proton affinity, the stronger
the bond to the metal ion. In general, proton affinities
take the following order:83

Amines, Imines ) Alkenes > Alkanes'
Thus,in each reaction sequence in Scheme 1, the more
weakly bound ligands are lost to form the final products.
Therefore, we believe that the products observed for
ethylbutylamine are 933 indicative of initial insertion
into the C-N bonds. These results are consistent with
those reported for primary amines.71 For example, the
reactions of ethylamine are typical of primary amines
(reactions (12) and (13))

00* 4' CZHSNH2 1:4 Co(C2H5N)+ + H2 (12)

Co(CH3N)+ + CH4 (13)

Products resulting from Co"’ insertion into N-H bonds
(reaction (12)) and into C-C bonds (reaction (13)) are

observed. No reactions appear to occur via initial

insertion into the C-N bond.

38

 

Scheme 1
Co‘ + WNHV
Co'i/YNHW /\_ Cok'rfi
. l .

 

(CTHA °+ (C4 “11”) (ETHACJKAH’ N)
cc*(c,H,,N) C3(C.H.N)
+ +
CZH, C2 H6

 

# _

(c,n,)c3(c,H,N)

c3 (C3H7N)
+

C, "s

39

In contrast to the results for primary and secondary
amines, triethylamine gggg form products indicative of
Co+ insertion into the C-N bond. These reactions are
shown in (14) and (15).

Co+ + (C2H5)3N -» Czns-Co+-N(C2R5)z —. Co(Cz,H11N)+ + C2Hz,(14)
\Co(C(,H9N) + + c2116 (15)
A comparison of the structures which would result from Co"’
insertion into the C-N bond of primary, secondary and
tertiary amines reveals an interesting parallel to
condensed phase chemistry.
R-Co+-NH2 R-Co+-NHR’ R-Co+-NRR"
(I) (II) (III)
Structure (I) would result from insertion of Co‘l’ into the
C-N bond of primary amines and represents a primary
amide. Co+ insertion into the C-N bond of secondary
amines would give rise to structure (II) which. is
analogous to a secondary amide. Structure (III)
corresponds to Co+ insertion into tertiary amines and
represents a tertiary amide. Products indicative of the
formation of structures (I) and (II) in the gas phase are
never observed. However, products indicative of the
formation of structure (III) in the gas phase are
observed for triethylamine. Note that in condensed
phases, transition metal primary amides and secondary
amides are rare while tertiary amides of cobalt are

common.10 The increasing stability of amides with

40

increasing N-alkyl substitution is not well understood.

Primary and secondary amides may condense with the loss

of NH3' or NHz—R to yield

R M M
I . V
on
/ \ l
M M M

unless there is steric
hindrance contributing to the "kinetic stabilization" of
the metal amide species.1°. Alternatively. primary and
secondary amides may be difficult to isolate in condensed
phases due to an ability to undergo "B decompositions" in
which M-H bonds are formed in a process similar to that
occurring for metal alkyls.Jlo
Furthermore, in condensed phases, transition
metal-tertiary amide bonds are considered moderately
strong:10
M—R < M-NRR’< MOR < MF
It is interesting to note that in the series C2H5NH2,
(02H5)2NH, (CzH5)3N the strength of the C-N bond increases
from 77 to 89 kcal/mol.82 Despite this, Co‘l’ inserts into
the C-N bond of triethylamine. This may be interpreted
as an indication of an increasing metal-amide bond
strength as the extent of N-alkyl substitution increases.
In general, the gas phase results suggest that, just as
in solution, the formation of a tertiary amide is
preferred over the formation of secondary and primary

amides. Also, it is interesting to note that in

41

solution, amines prefer to react to form amides when the
oxidation state of the metal is low10 and in this study

gaseous Co+ was utilized.

Interesting ligand effects are observed in the
the reactions of COCO+ with diethylamine and pyrrolidine.
Recall that Co" reacts with these secondary amines only
to form CoH. However, 'CoCO‘l’ reacts with diethylamine and
pyrrolidine to induce the elimination of H2. One
possible explanation of these results involves the
formulation of a mechanism in which £939: inserts into
the C-N bond of diethylamine and pyrrolidine. This
explanation is based on a series of reactions observed
for ethylamine and confirmed by labelling experiments:71

Coco+ + CZHSNHZ -» CO(C2H7N)"’ + CO (16)

00(0237N)"’ + c21151132 -» Co(NR3)(c2H5an)+ + CzHa (17)

By what is formally considered to be a ligand
substitution process, a complex between Co+ and
ethylamine is formed in reaction (16). Note that the
further reactions of a product ion are useful in gaining
insights into its structure.1 Thus. the further reaction
of the product in (16),reaction (17), implies that the
product of (16) exists ,as:

(0211:.) 33+ (N113)

Structure IV apparently is not formed when Co" reacts
with ethylamine. The formation of IV would require the
insertion of Co+ into,a C-N bond of this primary amine.

In reaction (16) however, the reactant ion is CoCO+. This

42

seemingly contradictory set of reactions can be
‘understood by considering the potential set of processes
which can occur between COCO+ and ethylamine if’CoCO+

inserts into C-C, N-H and C-N bonds. These processes are

shown in Scheme 2.

Assuming that all four intermediates (VI V", VI:
IV) are formed, it is evident that a variety of ligands
are attached to Co+ before the dissociation to final
products occurs. Proton affinities suggest that the bond
strength between 00* and" the following series of
molecules takes the following order:33

NH3 > CzHa > co > CH4 > H2

Thus in each reaction sequence in Scheme 2, the more
weakly bound ligands are lost to form the final products.
The product formed via intermediate IV" contains two
strongly bound ligands. NH3 and Czfla, which are retained
while CO is lost. Other possible structures for
Cc(C2H7N)+ such as Co(C2E5N)(H2)+ and Co(CH3N)(CH4)+
would not be expected since the intermediate through
which they would be formed would lose H2 or CH4 rather
than retain these electron-poor ligands. Thus, the
Co(C2H7N)+’ion formed in reaction (16) has the structure
Co(C2H4)(NH3)+, and 1c reacts further in (17) by a ligand
substitution in which a second molecule of ethylamine
displaces the more weakly bound CzHa from the metal

center .

43

m 338.

 

 

8
... V
8 N: . N... oN_._NO IO
n ...? N n+N n N ..._. I? N +
... :5. z 38 ..z #38 ..z x “.838 +.N..2N.IO.8 +£28.88
n # V N N _ n N % N A 0
. 12:3... I O. . ....me I O. . N:zN:3 88 NzNO. .xz 13.6w. :3
8 8 _— 8
x
...H. / ...E .. H. ..H.
\MJO \:O \1 TO All :2 (v
szuwwuNIO 2.8-12O On: szNzo c.O: 1w folio...” IO
cm 8.— » o 8
. _
szfNO
+

+88 .

44

Based on this analysis for ethylamine, we believe
that CoCO+ inserts into the C-N bond of diethylamine and
pyrrolidine to induce the elimination of H2. Reaction

(18) illustrates this mechanism for diethylamine. Note

0
H C H
CoCd+‘\,Nv-h-$H,CH,-éo-N§HCH3
" H H
ram of (18)
28-H’s
C? N _2,‘H ' cp N
+
u--- :3 u |"'C°"i'
H/ )1 He CH3 ”CC“:

that the dehydrogenation process occurring in reaction
(18) involves the shift of two hydrogens which are
located 3 to the metal ion. H2 elimination resulting
from the shift of two B-H's onto the metal in a manner
similar to that illustrated in reaction (18) has been
previously reported in the gas phase reactions of Fe",
Co" and Ni"' with alkanes.51v52v57v53 The dehydrogenation
reaction occurring for pyrrolidine is illustrated in
reaction (19). This reaction involves the insertion of
Co+ into a C-N bond thus forming a six membered ring

followed by the cleavage of this ring:

H C0
+ H C9. )4 I + ‘H2 CO\ 0..."
COCO+ N -—)- Co —>H-Co\hl' -aI- Cp
< > \" \u‘
\ (19)

The question that remains to be answered is: why
would CoCO+ insert into C-N bonds in primary and

secondary amines when 00* does not? Note that in the

45

insertion process

hfl'+*AB-§(AMBY+
the sum of the strengths of the two bonds formed, CM+-A)
and (AM+-B), must be greater than the sum of the strength

of the bond which is broken, (A-B) and the promotion
energy (PE)9O of M* for the process to be exothermic.
Since products indicative of Co+ insertion into C-N bonds
in primary and secondary amines are never observed, one
can conclude that it is not energetically feasible for
Co... to insert into the C-N bond because the energy
associated with the formation of the bonds RHéC-Co+ and
Co+ -NHR’is not sufficient to overcome the prommtion

energy needed to place 00* in its proper configuration
for bonding and the energy required to break the C-N
‘bond. Perhaps Coco+ forms stronger bonds to RHZC- and/or
~NHR’than does 004' thus making the insertion of CoCO+

into C—N possible.

There are several observations which lend support to
the claim that COCO+ forms stronger bonds than Co+.
First, data reported in the literature indicates that
this claim is true for the bond between a metal ion and
co. D(Mco+-00) > D(M+-c0) for similar first row
transition metal ions where.M t Cr. Fe and Ni and for M -
{4.32 Also, it has been reported recently for scandium,

that D(LSc+ - c113) > D(Sc+-CH3) for L - H,CH3. Thus ScL+

forms a stronger bond to methyl than Sc+ alone.91

46

Furthermore. 60* reacts with ethylamine and
ethylbutylamine to eliminate Hz via insertion into N-H.
Note that the N-H bond in diethylamine is approximately
15 kcal/mol stronger than the N-H bond in both ethylamine
and ethylbutylamine.82 No reaction is observed for Co“
with diethylamine except hydride abstraction. However. H2
elimination occurs when CoCO+ reacts with diethylamine.
This suggests that higher energy intermediates can be
formed with Coco+ than with Co+. Thus, the sum of the
bonds formed in IV’is larger than the C-N bond strength
in ethylamine and the insertion becomes energetically
accessible. . 8

Unfortunately. RCO+-NHZ and RCo+-NHR ' bond
strengths are not available. Thus a precise evaluation
of the thermodynamic influence on insertion into such C-N
bonds by (20* cannot be made. Nonetheless, these
experimental results indicate that CoT-NHZ and CoT-NHR
bonds may be weak. and D(Co+-NR2) is largely dependent
upon the extent of N-alkyl substitution.

Also reported in Table l are the reactions observed
for (20030»;+ and Co(CO)xNO+ ions with secondary and
tertiary amines. In general. as the number of carbonyls
present on the metal center increase, the observed
chemistry involves almost exclusively ligand substitution

reactions in which one or more 00's are displaced by an

amine. For example, the following ligand substitutions

47

were observed for all of the amines studied. (Note: "Am"

stands for the parent amine).

COCO+ + Am -) CoAm+ + CO (20)
Co(CO)2+ + AmfiCOAm-F + CO (21)
' CoAm+ +2 co (22)

With the presence of a nitrosyl ligand on the metal.

the only reactions observed are ligand substitutions.

Cocouo+ + Am -» CoNOAm+ + co (23)
Co(c0)2No+ + Am -) CoNOAm+ + no (24)
Co(c0)3No+ + Am Cocouomn+ + 2 co (25)

.1: Comm". * ”O (26)

Note that all ligand substitutions involved only the
displacment of carbonyl ligands. Never was the nitrosyl
ligand displaced. The energy needed to displace the 00's
comes from the energy released in the formation of the
new bond formed between the metal and the complexing
amine. Therefore. the number of 00's displaced is a rough
indication of the cobalt-ligand bond strengths.
Finally,the products observed in the chemistry of
Co"’ with secondary amines are much different than those
observed for analogous ethers. Oxygen is a very good
"director" for the site of metal attack in the organic
molecule; only insertion into the C-0 bond occurs. Thus.
. diethyl ether and tetrahydrofuran react to form products

indicative of exclusive (10+ insertion into the C-0 bond.

48

These products are listed in Table l. The processes
leading to the formation of products for 60+ and
tetrahydrofuran are indicated in Scheme 3. H20
elimination is shown to occur by first inducing the
isomerization of the ether to the ene-ol. Following Co+
insertion into the C-0 bond of tetrahydrofuran, there may
be _attack of the carbon which is bonded to the s carbon
to facilitate CHZO elimination. These results are similar
to those reported for small alcohols in which only
insertion into the C-OH bond is observed. For example,
isopropyl alcohol reacts with 00+ by H20 elimination but
not by cm, or H2 elimination pathways.1-6

The results observed for the analogous ethers
contrast with those observed for diethylamine and
pyrrolidine. These two secondary amines only react with
60+ to form hydride abstraction products. Furthermore.
the products observed for ethylbutylamine appear to be
typical H2 and small-hydrocarbon eliminations as a result
of metal insertion into C-C and N-H bonds. These results
also contrast to those for the ethers. Interestingly,
triethylamine does react with 00+ to form products
analogous to diethyl ether. Co+ readily inserts into the
polar C-N bond of triethylamine.

In conclusion, this study has allowed for an
examination of the basic chemistry of the isolated Co+

metal center with a series of amines and a determination

49

Scheme 3

 

 

+ c6”), H-ca" .
(C3H3)Co (CHzo) <— ' \
.J
shows) + CH20 c +.-.‘ '
+ \Ji
Flo-COW

i

4- o'\ 4.
H20 4- Cd) <——— (H O)Co (C H )
:>, 2 ‘4 s

50

of how this chemistry changes as the available orbitals
on the metal became occupied with CO and NO ligands.
Furthermore, this study has shown that the chemistry
which occurs for secondary and tertiary amines depends
heavily upon the structure of the amine. .A mechanism
involving a metal insertionIs-hydrogen shift and
competitive ligand loss was proposed consistent with the
observed results. Ligand substitution reactions were
reported. Since structural information is obtained by
knowledge of the mechanism of interaction and molecular
weight information can be obtained from ligand
substitution reactions. this study has illustrated the

potential of Co(CO)3NO as chemical ionization reagent.

Chapter 6
The Preference of Fe*. Co+ and Nf+
for Insertion into the C-C Bonds of Linear Alkenes

If each step in the metal insertion Is-H shift/
competitive ligand loss sequence were fully understood.
products and branching ratios could be predicted _a_
2.12213.- This would render metal ions as ideal reagents in.
chemical ionization mass spectrometry. The least
understood step in this sequence is the metal insertion
process, yet it appears to be the dominant factor
affecting product distributions.“3 To better understand
the insertion process, a specific ordering of the
preference of H'" (M - Fe". 00*, 81*) for insertion into
the C-C bonds of linear alkanes up to and including

fidecane will be determined. To achieve this end. first

other attempts which have been made to rationalize
product distributions will be presented. Next, a
procedure for establishing an ordering of the preference
of M" for C-C bonds in alkanes will be presented followed
by that ordering. Finally, an explanation of the factors
which render certain C-C bonds in nonpolar alkanes more
susceptible to metal ion attack than others will be

provided.

51

52

Several attempts have been made to explain why some
products are formed in greater abundance than others.
Most of these explanations have focused on the various
thermochemical aspects of the mechanism. Before going
into these thermochemical aspects. one must be aware that
evidence indicative of random or nonselective metal
insertion is never observed.57 If insertion were
predominantly a random process, then for the reaction of
00* with butane, products arising from intermediate ‘1
would outnumber products arising. from intermediate 3
Co+ + cn3cnzcazcng I n3c-cf-cuzcnzcng —» pdts (6)

cn3ca;-c6" -CH2ca3 Pdts (7)
by a frequency of two to one. However. recall from
Chapter 3. that the majority of the products observed in
the reaction of 00+ and butane are attributed to

intermediate 2 and not 2;. Product distributions‘

' . indicative of a random trend have never been observed for

alkanes or for any other class of compounds.

In terms of the thermochemistry involved in these gas
phase reactions. one might expect the stability of the
final products to be the dominant factor in determining
product distributions. In the reaction of 60+ and
butane. this stability can be assessed upon consideration

of the energies involved in the organic decomposition of

53

butane and the metal-ligand bond strengths in the
products. The energies required for the necessary

organic decompositions are shown here:82

AH kcal/mol
n-Chfilo CH1, + C3116 17.2 (27)

R2114 4- C236 22.4 (28)
Unfortunately. the necessary metal-ligand bond strengths
have not been determined experimentally. However, we can
apply the correlation5 between metal ligand bond
strengths and the proton affinities83 of the ligands.
The proton affinity of propene is 185 kcallmol. The
proton affinity of ethylene is 164 kcallmol. Thus. since
less energy is required to form propene from butane than
ethylene from butane. and because the resulting
Co+-propene bond is stronger than the COT-ethylene bond

(based on the proton affinities) one might expect that

-.reaction (6) would be favored over reaction (7).

However. as we have seen earlier. the experimental
results indicate that the products of reaction (7) are
formed in a greater abundance than those of reaction (6).
Thus. this analysis, as well as similar analyses for many
other reactions. shows that the final product stabilities
for these thermoneutral/exothermic reactions do not
determine the product distributions.

Another approach used to explain product distributions

focuses on the metal insertion step. Recall that for a

54

metal to successfully insert into a bond

14““ + R1R2 —9 R1 41" -Rz (29)
both the energy required to cleave that bond D(R1-R2) and
the energy of the two H+-R bonds formed must be
considered. Since accurate thermodynamic information is
available only for the M" -CH3 bonds (M - first row
transition metal)77o78, discussions have focused on a
correlation of product distributions with the bond
strength of the C-C bond which is cleaved. A
consideration of n-butane reveals that the terminal C-C
bond strength is ~3 kcallmol greater than that for the
internal C-C bond.82 This suggests that -M'* selectively
inserts into the weakest skeletal bond of butane to yield
the largest fraction of products. However. a difference
in bond strength of only 3 kcal/mol may not be expected
to produce such a dramatic selectivity. Furthermore. in

the larger alkanes, the difference in energy of the

' . internal C-C bonds becomes negligible. Therefore, based

on this type of analysis, the trends observed for large
alkanes cannot be rationalized. Thus. no correlation
between D(R1-R2) and product distributions can be
accurately made for the metal ion/alkane reaction results
in the literature.

The thermochemistry involved in 221:}; bond breaking
and bond making has also been considered to explain the
failure of M+ to insert into terminal C-C bonds of longer

alkanes. Freiser et al. suggest that the higher energy of

55

the terminal C-C bond coupled with an increase in the
metal-alkyl bond energy as the length of the alkyl chain
increases, results in selective insertion into internal
bonds in alkanes.57 They suggest that the increase in
metal-alkyl bond energies is due to "electronic
considerations".S7 ‘

We suggest here that, since the c-c bond strengths
in the n-alkanes are all approximately equal.82 the W—R
bonds formed in reaction (29) determine which insertion
intermediates are most stable”. Based on work performed
in our laboratory along with product distributions
reported previously5o6v43v35o51v57, we report here a
direct correlation between the C-C bonds of normal
alkanes preferred by N'" for insertion and the ionization
potentials (IP's). of the alkyl groups bound to M" upon
its insertion. We suggest that this correlation is

indicative of reactions where the most stable insertion

'- intermediate is formed preferentially.

Before a procedure for establishing an ordering of
the preference of M'" for the C-C bonds in alkanes can be
presented, both the experimentally determined product
distributions and the mechanism of interaction leading to
product formation must be known. The products (both
neutral and ionic and branching ratios obtained in the
reactions of Co(CO)+h, Ni(PF3)+n: N1<¢°>n+ and Fe(CO)+h[n
- 0.1.2] with linear and branched alkanes are presented

in Tables 2 thru 21. These branching ratios are

56

I .. .m.z T mmmo ... ”$.3sz

n u .m.z Al meawo + ”622

o.” u 8.... wmwooo All OINO + ”888

 

perm: 0.... fig OBUKHS HuN CH 8:008 gflgfihumflo ”US .0 Hugh.

u n .m.z Ann! $6 + Emmavflz

6 + ”.85.

E
I'
§
§
T
”as”

”8.8

 

cm: o... «ensue: «sauna... H8 5 8388 8033383 ”688.. .m mama...

S7

o.H

0.H

mo.o

mmeo

vm.o

Hm.o

mm.o

o.H

B
. a...

man. 75 +wImOmaaaz

name + memo?

C
mmmc+ vmo + meowz mmmu ++Ammmvwz

83...: + mmmoBETl
8: + mmmuaz al!

8: + v.8 + wmmofiz T

 

 

 

 

8: +~I + NImOazal.
8: + «IN + wmmowz 4 mImO + ”80:2
8: + ammoooalh
C
00: + Nm + wmmooox/ ammo +1088

 

FE 0» 0280.5 859:: .235 025085 5.033333 yon—comm .v mama.

58

Den

O.H

 

N"

«5.0

moeo
mweo

Hm.o

mH.o

mh.o

NN.o

0H.o
Haeo
mu.o

v0.0

$0.0

mm.o
moeo

Nm.o
.mmm.

me: 75 +ow=vommaez AIJ
mam: + omvaaz AI.

m N wv
...—act. II. +I22AI.

 

name + OI~O + NINBz AIIIII OHva + ”Annexe.

8. 75 +owmvo8az A.|

Ca V

m Uuz ATII

8c++

85 w: + .9sz All

 

m N

08+ IO+ wzwgz Alrl cameo + ”.832

8. 75 +owmvo8oo AII.
8... + owmvooo All
8: ... Na + WIVUOU T

00: + vmu + wmmooo Arll

 

00c + meU + WENURU Trul

cameo + ”8.8

 

\

ca: 0» ocmusm unsung: Haw cw co>uenno acofiuonauunwo anemone .m mamas

59

 

m NHm

 

 

 

 

I 04 I E: + +I 32 al.l

I I ~00 name + mm + owing. All

I I 3.0 _ mam: ...mmmo + mmmgz All

I .. 8.0 meat +0200 + ”Imus. All «ammo + ”.0832
04 .. I _ 8.75 +~HIm08az all

.. ..v.0 I 8: + «Humvee. T,

I 00.0 8.0 00: + «I ... 0HImowz AIJ

I 010 3.0 . 00: + OINO + mImOaz All.

I I 2.0 80+ ammo + ”smog. AIIIIII 3:00 + ”.832
0..” I I 02.75 + NHmmUBOO All

I 00.0 I . 00: +~mmmoooall

I v0.0 m~.0 000 + «I Elm—.0081;

I 8.0 00.0 00: + OINO + mImooo All

I IHII. NE 8: + mmmu + wmwoou TIII wammo + ”80.00
Nut Hut 0»:

 

c
as 3 33:8 30“..tz Haw 5 @0338 3033333 uosmoum .m mums...

60

hm.o

nn.o

no.o

mm.c

 

NI:

no.0

~n.o

N~.o

ao.o

ao.o

 

HI“

ce.o

cm.o

NN.o
aN.o
nH.c
wa.o

 

on:

89L: ... I+IO O8HIAJ

+<H= o

8: + IOIIAII
N ... N

8:... I... I0+ MIIOIIIJ

8.. + IInO ... OInOIIII

82-5 + eHIeoooeohg

8a ... .2 IIOOOAnJ
8s + NI. + NH IIOO
e N

UaTI.

8a+ I 0 ++ I IeonaII

ooa + N: + omwo + eroou AI.

8: + IIMO + eInOeoAl

 

a
II /\/\/+ +88;

 

I“ </\/+ H8800

 

an: on scene: ousutz Hun sH ve>uoano acoHusAHuuan uosvoum N mam<9

61

OH

 

NIG

 

 

 

I I . «IISIE + «HIIOMIIEAII
3.0 I 0&0 ... «HIOOIIAI.

I 2.0 meme + NI ... NHIeoazAI.
2.0 00.0 0.3: + 0I00 + 0HI00HIT_

I 3.0 . «II... + 0I~O + MIIOEAI

I 8.0 0.80 + NI + 0I~O ... NIIOIIT

I 00.0 0.30 + 0I00 I. MIMOEAIII /\/\/+ ”A0832
HI: . ass .

AvoaaHu600v

5 mafia.

62

04 04 I . 0&0 + WHIIIOII

 

 

 

 

 

I I 00.0 nae... NI + ”H.102

I I 00 .0 0.5:. IIMO + +0Ivoaz >\/\/ + ”Ame: Hz
0.” I I 82L: ._.WHI008IIAII

I 00.0 I 8: + WHIHOIIAI

I ”0.0 00.0 8: + ~I + HHIIOIIAII

I I 3.0 . 000 + 0I~O + WHImOHzAl

I I 3.0 .. 00:... mImu + WIIOHIAI

I I 3.0 00: + ~It. 0I00 + +0Ivoazml

I I mmd 00: +0HI¢O + +0Imoa24|ll >\/\/+ ”.832
04 «0.0 I 8: + 3.1000?!

I 3.0 2.0 8: + NI + «MI008AII

I 00.0 2.0 8: + OI~O+ mHI008AII

I 010 3.0 8: + IImO + mIvoooT

I 2.0 8.0 8: +0HIvO + mImOooAIII \/\/\/+ ”.808
.~uc n: on:

 

“a: Cu 2300a gawk“: HuN Cd 8:038 gfigfihunfia “Dag .0 mum‘s...

63

o.H

mm.o

mc.o

 

NIC

nn.o

n¢.o

w~.o
NH.o
nu.o
nH.c
mH.o

 

HUG

oe.o
MH.D
mn.o

no.c
OH.o
NN.o
nH.o

Ie.c

 

CI“

8.. + II ..IIOIIAII
IIIO ... 0IIIOII All
8: + 0HIIO + WIIOIIT

80 ... NHIIO + ”Inge—ml.

.8: ...

8: + IHII I08 TI

80 + NI + IHIIonfl

80 + II~0+ NH ...IIonAIl

8: + IIIO + 0H ImonT

8a ... 0HIIO + +I IoooTl,

III In
8:... I0++I onAIL

 

D
II + +88 «2

 

II </\/\/+ H8800

 

nu: ou ocauoo ousuxHx.H”N 6H vu>uompo IsoHuanHuuan nonvoum .m mqm<H

64

H0.0

0H.0

 

NI-h

no.0
0H.0

No.0

uH.0

Is.0
IH.0

0H.0

 

HI—H

0H.0
«0.0
HH.0

N~.0

8.0~

HH.0

00.0
In.0
0H.0

sn.O
Ola

 

8

_ 83L: + I.I+IIO8IIAI

0H 0
8: + +I 0.931
8: + NI + IHII

oo:_+ ammo +.0H¢n

OeIAII
oomAIl
80 + II + IIIO + IIIOIIAI
0IIIO + OaIAI.

uumxTII

0+

on: +

00a + «m +.0Hmeo +

.+

no
MI
0

0+9

80 + NIIIO +

«H I

m uflmAll

~II+

8e+ I O+ INOIIAI

+

I II I
III I. +I OIIII.
III: + NI + IHII

IIIa+II 0II

III: + 0III I I

UHZATI.

III: + NIIIO ... ”IIOII AI
AvoscHuaoov

0 mandfi

m U +. +fl UHZATII

o + +r UHZ.AII

 

II<<</+ H83:

 

II a I
<<</+ +I .52

65

00.0

00.0

00.0

00.0

00.0

00.0

 

NI“

0H.0

IH.0

00.0
00.0
IH.0
0H.0
0N.0
50.0
50.0
0H.0
NH.0

0H.0

06.6
HUG

 

00.0
50.0
00.0
00.0
0~.0

00.0
00.0
00.0
00.0
0H.0
Hn.0
00.0

0N.o
Ola

 

0275 + 0+I 008II¢
80 + +00 +IIOIIA|

80 + NI + II IIOIIT
000 + IIIO + II ...IIOIIAJ
80 + IIIO + II +IIOIIAII
.000 + 0IIIO + S +IIOIIT

80 + NIIIO + WIIOII AI

 

80 +IIIIO + +IIOII AII
8III5 + 00 IM0800AIJ
8: ++0... ..IIOooAI

80 + II + IIIIOSAI

80 + IIIO + II II000 AI...

80 + IIIO + «I +I0000A]

80 + 0IIIO + 3 I000 AI!

80 + NIII0+ IIoeo Al.

80 + II + NIIIO +

«H 0

II000AII

a
.II \/\/\/\/+ +8032

 

Ineeet

con + I u +

+~o-+~o-+%

III >.\/>\/+ ”80000

 

A!

00 23:02 0.3003: HIN .5 00:00.5 23303330 Ion—Iona .0H mania

665

00.0
00.0

05.0

0N.0

 

NI“

55.0
0H.0
H0.0

00.0

«0.0

00.0

00.0

00.0

00.0

00.0

0H.0

 

HUG

«0.0
NH.0
00.0
00.0

N~.0

00.0
00.0
«0.0
0H.0
0H.0
5H.0
0~.0

HH.O
Ola

 

IIIIIIE + 0II00IIIIIJ

III: + 0II00IIAI

0.50 + II + IIIIOIIAI

0.30 + III IIO + II IIIOIIAI

III. + IIIO + IMI I

II: + 0III0+ 0III0IIAI

III: + IIIIO + IIIOIIAI

IOIIAI
I

 

0.30 + III—00+ IIIOII AI

uncooMATI

+0I IoooIAI

83.5 + 8.. I.

80 +
80 + II + III.II00.IAI
80 + IIIO + III.II00IAI

80 + IIIO + II IIOIIIAI

e I
III>>>\/+ +I IIIII

 

80 + 0III0+ 0III00IAI

8s + IIII0+ III IOIIAI

80 + II + IIIIO ++ 0 IIOIIAI

8: + III00 + MIIOIIIIP

80 + IIIIO + NII IOIIAP
AvoaaHuaoov

0H mam<fi

II >>>>+ use...

67

a.

«N 0H

 

 

 

 

 

II.0 . - 003.5 + +I 0002 T
II.0 II.0 .. 8.. + I.I+I0I0II II:

. II.0 I0.0 00.. + II + 0II.I0I0II AI!

u I0.0 II.0 00: + III0 + III.II0II T

- - II.0 .00.. + 0III0 + IHII0II TL

- II.0 II.0 00.. + IIII0 + 0HII0II T1

. . I0.0 00.. + IIII0 + MII0II TI

u II.0 0I.0 00: + IIII0 + ”Inga Tl /\/\/\/\/+ ”8002
II.0 I0.0 .. I 83.3 + IMI0I000o0 T.

II.0 II.0 I0.0 00.. + IMI0I000 4n.

. 0I.0 I0.0 00.. + II + 0.I+I0I0o0 T

- I0.0 I0.0 00.. + III0 + I.I+II0o0 «l

. I0.0 I0.0 000 + III0 + I.I+II0o0 .1

- II.0 II.0 00.. + 0III0 + IHII0o0 T

- 0I.0 II.0 000 + IIII0 + 0II.II000 T

- III0 II.0 00: + IIII0 + .I+II08 T

. I0.0 II.0 00: + IIII0 + NII08 +|I + ”80000
«I: an: 0|:

 

5mm cu acuoon unsung: HIN a“ vo>uonno uaoquSAquaun unavoum .Ha mam<H

0.H

68

0c.0

H0.0

 

NI:

5H.0

00.0

m0.0

 

HI:

0H.0
II.0
NH.0
00.0
IN.0
0H.0

nH.0

NH.0
«H.0
ma.0
0N.0

He.0
OI:

 

003-5 + I.I+I0I000~I all

‘

00: + I.I+I0I0II TL

00: + II + 0.I+I0I

II III
+I

oun.¢ll

oum.TlJ

00: + 0III0 + I.I+II0u..I All
II I 0I I
+I

ova + m o +

00c + m U +

IIII0 + II.II0II T:

02. T

00: +

 

IIII0 + III0oI T

a
00 + +

IIIIILII + I.I+I0I0II.III T

mum: + NMI0H

III: + III0 + IJIrII0II T
III: + 0III0

0&9: + Nfimmo +

III: + IIII0 + wI

III: + IIII0 + ”III0; L

+ I.I+II0II III.
0.I+II0II II,

c

AconcaucouV

HH mnm<fi

Uwz.$lls

Uwz fill!

r /\/>\/\/+ ”80va

 

l a I
/\/\/\/\/+ +I IIIII

69

 

0.H 00.0
I 55.0

I «H.0

I 0N.0

I 00.0

I 0n.0

I n5.0

I mN.0

 

NI“ HI:

mN.0
m5.0

 

DIG .

m 0H

IIIIIS +

nhma+ +

.IIII + I

82-5 + I...

CD: +
I

ooc+ so +

003.5 +

CD: +
N

8=+I+

oo:.+ e

:0 +

:0 +

meomhmHz

0I II0II

osmon

I II000II

+0H mIon

III0II

OUT—000000 AJ

+0I II0o0 T

+II0o0.IJ

as L

+MI+I+

a I
+ +I IIIII

G
+ +800 Iz

 

ll /_\ + ”8800

 

:4: cu uaumoumHmcuuxlu ouauxwz Hun :I vu>hmnpo Icowuanfiuumwn uuavoum.~H NHO<H

7O

NHmn

0.I I I IIIIIIS + +I 0 IIII
I II.0 I . III: + III.II0II
I II.0 0.I III: + II0 + ”II0; /T + ”IIIIIII

 

 

 

II.0 I I .. ‘ 00III5 + IHII000II
I0.0 II.0 I 00.. + IIIIII0IIW

I II.0 0.I 00: + II0 + ”II0; /*\ + H8002
II.0 I I 0275 + I.I+II00000
II.0 II.0 I 00III5 + II0 + ”II000o0w

I II.0 0.I 00: + II0 + .I+II0o0 /‘\ + ”8008
II: II: . 0n: 0

 

PE 3 oaomoumHhmuuaIaININ 0&5sz HIN 5 3:33 2.5393330 ans—03m .2 Hands

71

 

 

o.H I

I 5.20
I 0H.o
I nn.0
o.H I

I 5......0
I 910
NI: HI“

nH.0
0N0
00.0

0H.0.

HN.0
00.0

0.30

 

OI:

003-5 + I.I.II000II T.

NHn

00.. + +I III II,

N 0Hm

00: + a + +5 UHZT,

00: + N: + «:0 + Na.»

00.. + III0 + III0IIII

00 N

I-r+

00.. + I 0 + I 0IzI|.I.I

...

+I
II I
00.. + +I

N 0Hm

003-8 + II I0008J

000T

006+ m... +m UOUAII,

00..+II+I

\D

I

mu... m 000T

00.. + III0 + II0o0TI

+04-

00% All,

 

 

G
+ +800 E

fl
0
+ +88 0

 

55 cu mcouISthuszN 9:503: H8 .5 vmzumno IcouuanwuumIn

nos—00.5 ..IH M305.

72

0.H

 

NIH"

no.0

nN.0

«H.0

 

HIHH

«H.0
5H.0
NH.0
00.0

0N.0

 

Ola

NH: mm

IIIIIIII + I0 IIIIIJ

III... + II IIIIIIII

III: + II + 0III0IITJ

mum: + «so +

mum: + N3 + emu +

m 0 N

SUHZ.¢II

MQQ

mUHZ.¢II

\D'+\D‘+‘D+

m

hmn_+ m U + m UHz.¢ll,

+

IIII + III0 + I 0IIIIIL

+.:

AvuaaIIIOUV

0H man<fi

 

.. I
+ +I IIIII

73

M50

5N.0

 

NI:

0H0
II0.0
00.0

I~5.0

00.0

0n.0

 

.mIC

HH.0

00.0

00.0
nH.0
0N.0_

5n.0

0H0
MH.0
nN.0

0.30

 

GIG

.Ich

m
...EAH av ... ...: U mmwz AIJ

n IIH0

II: + .+I 0Iz.TII

mum: + Na + NHEOUIZAIIL

name + 0:09 + WxMOHzT

00IIIII +
I

000II T
S I
+I

II I
+I
00.. + I0 +

00=+II+I

w

I0 + II0IIAII.

....

I

\D

8: ... moN +

.I
con + 0:no + EMUHZAIIII

w “0sz

+~o+

:0 + snoou AIJ

00.. + II + II0 + II0..0 A]
II0o0 T
I

I

O
H

006 +

+

I

+I+I

00: + SUN +

Co: + 0:00 + m 000T

+0

OHZT.

 

§ ++I IIIII

 

a
I § + +8002
VI fl + H8008

 

 

INS cu ucquII—HhcuaaIvInJ 0.305”: H" N III vmtouno 28395.33: uusvoum .mH mung.

74

 

NI:

II.H.0

no.0

No.0

0H0

0H.0

0n.0

IIH.0

0N.0

 

HI:

05.0

5N.0

50.0
0H.0
nn.0

00.0

II.0.

00.0
.IH.0
HH.0
Hn.0
0N.0

00 . 0
OIQ

 

0

0mm: .... omN0 ...

2.. + II + III0 +

IIH=0

+::+::

002;: +

00.. + +IIII0IIAIJ

00II + «IIO + 0Hm m.

00..+II+I II

0N 0.I

00:... m0... +m0HZAl,

00II + Nu + 0mN0 + 0

N...

00.. + 0III0+ III 02 IL

++
002;: + II0 + 0II I

00.. + II + +III. II0o0AII

00.. + II0 + 0II I

00..+II+I

000 All

...m

$0 + 0000 All

00II + 0=N0 +

00.. + II + III0 +

0n

II0o0 AIL
II0o0 II
I

0+H~o+co+co+

0.3—?!

:0 + ...: 0HZAII,

...mIIUHz Al,

”MBWZUI
II0II g/.. + +I IIIIII

000.;— All_

 

00000 All.

6
II /\7 + +800Iz

 

G
00..... I0+ I000ALII <7 ++80Io0

4.

55 cu ocauapHNIfiuEII—INJ 0.25sz HIN III wok—025 IIIOIIIEIIIIIG you—Ion

m .0H H.540.

75

 

 

 

I me.0 I

I NN.0 Nm.0
I 00.0 00.0
I 00.0 I

I 0e.0 0.H
I NH.0 I
00.0 I I

I 0N.0 I

I 0H.0 I

I 0H.0 50.0
Nm.0 0N.0 I

I 0H.0 . 00.0
NIII HIII 0|:

III... +

mum—I +

00:,+

00II +

00III=I I.

00.. + II + III. II000 Al.

00.. +
027.0 +

00..I +

0H: c

n

mm: +

II0I

0Hfl<0+

II0I

0H «0

00:.+

II0I

0Hm¢

0 + +2 00000ITI

I

II II0Iz
+ S II0II
I II0II /_\_/+ ”IIIIIII

0H :00

00.. + II
+ 0I II0Iz
I0 +III0II /_\_/ + ”800:.

II +II0008AI

+0H 20000ATJ

+ 0.I+I I000 II

0 c

 

mm.» G
0 + + I000ALII /_\—/ + +8008

 

:A! 00 wcuuanhnuwaauuoaIndnqqu ouauxwz HIN :« vu>uuuno IcquaaquII0 uusvoum .5H man<a

76

0.H

 

NI...—

00.0

5m.0

no.0

MN.0
0N.0
mm.0
0H.0
nN.0
0H.0
0H.0

H<.0

 

HI:

0H.0
no.0

00.0

no.9

mn.0

HH.0
mn.0
0e.0

«H.0
OI:

 

IIIIIIII + III. II0IIIIIAJ
I

III: + II + II II0IIAII

III0 + III0 + WII0IIA|

mum: + Na + 00N0... 0m<0HzITII

«H: 0

. 83.5 + 000IIA|

00.. + +II II0IIA|

N NH :0

00II + m +

00.. + III0+ I II0IIAII
00.. + II + III0 + III0IIA|L
00.. + II II0o0AII

N I on

00: + m +_ £0 +

00: +_0:N0 +

N 0 N

00»I + m + w 0 +

mN

00II + 0Hme0 + 0000 AIL

I+-c‘+~:‘+¢:+

I.I. + +IIII0IIT

:0H2I4l!

+I 0004!
00.. + II + II II0o0AIL.
00000AII,
II0o0 AI...

 

..I
I. \</ ++I IIIII

 

a
.l g + +8003.

 

II. \</ +H80000

 

:4: cu manucmthcuszm wuautz HIN cH 0m>uuaao IcoHuanHuumHn uuavoum .0H 00049

77

No.0

mm.o

 

NI:

cw.o
om.o

no.0

an.o

mn.o
nH.o
ON.c
aN.o

 

HI:

mN.o
no.0

NH.0

9H.o

III.

om.o

cH.c
mH.o
mm.c

Hn.o
OI:

 

I.I.IIIIE +

m

0H mnom

n +cH I

IE.» + ...: UHzAlll

III: I. II + III IIIIAII

III: + III0+ +II.II 2241
II: + oIIIo + ”In IIIAL
6275 + II IIoouIzAll
8: + +II +IIoIIAl.

8: I. II + II IIIIIIIAII
8n + III0+ .I+II SIT
8n + IIIIo + NI IIIIAIL
83-5 + II I8I I884»!

8: + +II +IIoouA|

8a+II+IIII

8: + III I.

:UOUATII

NHmooou AI.

8: + III0 +
II I

I084]

I 08 T

0+

”MIN
.7
0m
a
00 + +m

IEHWI All.

 

L/L/ II
I + +I I.IvIz

 

L/\F a
I. +882

 

I. \FL/ + H888

 

:4: 0» unaucunHfi sumaInII. N wunuxwz H. N :I vo>uunpo maowusawuumwn unavoum .mH m4m<9

78

0.H

 

NIC

«c.O
He.o

QH.o

cH.o
oo.o
on.o
on.o

ao.o

 

HI“

nH.o
nN.o

«0.0

no.c
0H.0
5N.0
HH.°
mN.c

QH.o

 

OI—u

can I. II I. IHIIIIIT
8: + III I. IHIIoIzAII

8: + III0 + IHIIIIIT

8: + III0 + III + ”IIIII ml,

NHmn o n

8: I. o I. +I III»...

8: + IHIIoooAII

8: I. II + IHIIooo AI

8: + III ..I I.IIIIooo AI

8.. I. III0 I. IHIIooo ml

8: I. oIIIu + MIIooo All

8: ... IIIu + III + ”IIooo AI
8: + IIIIo + IIIoooml

+

 

fl
III + +Aoosz

 

G
[I O
+ +88 0

 

:4: cu manucwmflmnuoaIIBIIIMIN oususz HIN aw vm>ummno IGOIusnquIIa unavoum.cN ugm<a

79

o.H

 

NIH—

o¢.o
mm.o

ao.o

No.0

 

HIG

mH.c
Hm.o
mH.o

on.o

 

CI—H

I - II I I
IIII 8 + +I I IIIIIJ

n wH I

I3: + m UHZAII

.7

III: I. II + IHIIIIIAII

n OH m

E: + IIII + I.I IIIAI

III: + IIII I. III +

IIII I. IIIII I. IIIIIIm.L

IIIII AI

‘+~o-+~o

Avuacwuaoov
ON mam—SH.

 

a m
I. +I III II

80

0.H

No.0

 

NI:

OI.o
hI.o

nH.o

Hm.o
oc.o
Ic.o
nI.o
9H.o
no.0
mo.o
«c.O
Ic.o
HH.o
no.0

 

HI:

no.6
nn.o

Io.o

oh.o

INA.h

sc.o

00.0

HI.o

oI.O
OI:

 

M

Q
v-I

IE: + m o;— All.

+

nhmu.+ N: +

mama + OHEIU +

‘D
H

w UIZATI.

IIIII Al.
I

 

IIII + IIIII I.

N

M

m UfizATII
UIZITII
whoqz All

IIIII A]
I

+'¢>+-¢I+'

\D
H
on

ooa.+
. I
DUB + :0 +

on: +.omNUN +

OH I

m + m

I+I+§S+

ooa + m o + m UdZA-l,

+.

III I. IIIII ... IIIIIIIL

II I
I.I

-+~o

IIIIIII I. IIIII All

III I. IHIIIII Al
III + II + III.IIIoI T
I II I

III + II + I 2.on

‘-
IIIIIII I. IIIII + MIIIIIIIAII

III + IIIII + ”IIIII All

\_/\_/ I. ...I IIIII

 

F \_/\./ + ”IIIIII

 

III + IIIII I. ”IIIII Al

I L/\_/ + H888

 

cg: ou ocuucumthuoaIuhquNqM.

unausz HIN nu vo>hoano IGOIuanHuuan nonvoum.HN mam<a-

81

consistent with those previously reported5-5o43o35o51v57
and represent the product distributions. Knowledge of the
mechanism allows for correlation between specific
insertion intermediates and specific products. When
mechanistic information is coupled to product
distributions, a direct relationship between product and
intermediate abundances can be established.

Consider the reaction of Co" and heptane as an
example. Heptane is composed of three pairs of chemically

indistinguishable bonds, labelled here as a, b and c:

15 c

In the reaction of Co+ with heptane, Co+ has a choice of
inserting into three "distinct" C-C bonds: a, b and c.
Based on Scheme 4. which involves the accepted metal
insertion ls-H shift/ competitive ligand loss sequence,
and the branching ratios taken from Table 8. an ordering‘
'-of the preference of Co“' for a. b and c can ”be
determined. Thus, disregarding Hz elimination,
intermediate :2 accounts for zero per cent of the
products. 1 accounts for seventeen per cent of the
products and § accounts for seventy-two per cent of the
products, suggesting bond (c) is preferred over (b) which
is preferred over (a) by 00*.

This initial analysis has assumed that H2

elimination occurs exclusively through 00* insertion into

C-H bonds followed by the shift of a s-hydrogen (1,2

Scheme 4

Co+ +WV

Intermediates

CH3-Co+-csn13

3
N

czns-Co+-csnu

3

, . C3H7-CO+-C4H9

5
~

82

 

 

 

H—shift
H-Shift r’
,
H-shift f
L.

H—shift

I

 

' mow-c.7111 5

(a

VProducts

Cocsulz” + CH4

COC5H10+ 4' C2H6

CoC2H4+ + 051112

COC4Hg+ 4‘ C3Hg
COC4HG+ + H2
COC3H6+ 4’ 041-110

COC7H14+ 4' H2

695

I7%
0%

16%
0%
56%

10%

83

dehydrogenation). However, the literature reports that Hz
elimination can also be induced by N'” insertion into C-C
bonds fbllowed by the shift of mo hydrogens which are 3
to the meta151v57 (1.1: dehydrogenation). Therefore. to
accurately assess the preference of M'“ for the C-C bonds
of alkanes. differentiation between H2 elimination
resulting from initial insertion into C-H and into c-c
bonds must be established.

Two approaches have been used to distinguish between
1.2 and 1,4 dehydrogenation.“ These include labelling
experiments53 and collision-induced dissociation (CID)
experiments.57 By examining the branching ratios
associated with H2. HD and D2 elimination in the
reactions of H" and strategically deuterated alkanes, the
percentage of W” insertion into each of the various C-H
and C-C bonds .should become apparent. The reactions of

specifically deuterated butane. pentane and hexanes with

. ‘ Fe+, Co‘" and 81* were studied.53 The percentages of both

1.2 and 1.4 dehydrogenation observed in these reactions
are sumarized in Table 22. For example, the results for
the reaction of Co"' with CD36H2C320D3 suggest that
fifty-one percent of the H2 elimination products result
from initial insertion into the internal c-c bond and
forty-nine per cent from insertion into 04! bonds. The
results obtained for the reactions of Fe+ and 00* with
deuterated pentane and hexane are inconclusive. The

products observed in these cases can be attributed to

84

.IoIIIIomoIIEwI :1: I0 III .653 3 RIBIIIII I.I III 383% .8388 I0 IqusnoI .n
.mm .umm II III «Hams II ncqusnIIunIo uuscoum co owmmm .I

 

 

 

 

 

 

 

 

 

I u m I I
I I I I m
. . .
n H I III " I n n I III m I III m I +II
. .
n u . u n
I . . n
_ . u u .
I u I III " III. I u n I " III II I II +8
. . .
u . I m u u
I I I I I
I . I III. ” III I u n I “ III II .I II +II
I . I u .
m I m I I
A I u I I J. I I u s I u I I ... s +2
I I.I” II III II III ”II I.I II III” II III II I;” II III II IIIHII II
. I IIIIII 05.2 838

 

0 05802 05 gang é mkuflhfluflflo 20%; 2 HO EOdUUGOm

 

.n> N H ucmo uwm.- mamas

 

85

either 1.2 or 1.4 dehydrogenation or both. Furthermore.
the authors note that caution must be used in
interpreting the results of these labelling experiments
due to potential deuterium isotope effects which may be
significant in all or some of the individual steps of the
mechanism. 53

Also, some products observed in the reactions of Fe+
with deuterated alkanes are indicative of the occurrence
of B~methyl shifts.53 However. these products are
representative of only a small fraction of the total
reaction products. Therefore, the ordering of the
preference of Fe+ for the C-C bonds of alkanes should be
unaffected by the occurrence of B-methyl shifts.

Jacobson and Freiser have utilized CID to probe the
structures of the ionic products formed in the
dehydrogenation reactions occurring between 11'" and select

alkanes.57 Their results suggest that thirty per cent of

' - the dehydrogenation occurs via insertion into C-C bonds

and seventy per cent by insertion into C-H bonds in
alkanes for Fe+. For 00+. ninety per cent of the
dehydrogenation occurs through a 1,4 process and ten per
cent by insertion into C-H bonds. Ni" appears to insert
exclusively into C-C bonds to eliminate H2.

When dehydrogenation can result from 34+ insertion
into more than one C-C bond of a particular alkane,
structurally different product ions of the same mass to

charge ratio may be formed. The CID results of Jacobson

86

and Freiser further suggest which of the C-C bonds in
hexane and heptane are preferred for insertion by 00* or
Ni+ in dehydrogenation reactions. For example. of the
dehydrogenation products resulting from the reaction of
Co+ with heptane, sixty-two per cent would suggest a
Co+(propene)(butene) complex and thirty-eight per cent a
Co+(ethylene)(pentene) complex.S7 These results indicate
that bond c is preferred over b for insertion by Co"' to
eliminate Hz from heptane. Reassessing the abundances of
intermediates 3,4,5 and b in Scheme 4 based on this
information. the overall ordering of the preference of
Co"’ for insertion into heptane is c )-b > a. Note
further, that according to the results in Tables 5 thru
11, as the length of the alkyl chain increases from
butane to decane, the total percentage of Hz elimination
significantly decreases. Thus consideration of H2

elimination resulting from insertion into C-C bonds in

' . determining the abundances of the various intermediates

becomes unnecessary for alkanes larger than heptane.
Using the product distributions provided in Tables 5
thru 11 in conjunction with Freiser's results for 1,4
dehydrogenation occurring in hexane and heptane,57 an
ordering of the preferred sites of Co"’ and Ni“’ insertion
into the C-C bonds of normal alkanes from butane to
decane is presented in Table 23. Note that the trends
observed for these two metal ions are identical.

Furthermore, an‘analysis of the product distributions

87

Table 23. Pufm I of Pe", Co+ and an (or c-c Bond insertion
‘ , {unmet-Alkenes"

a. . c
£00233.
£33393. 1

a 339.. .32..
£333.. 0%

~ 33333333.
3399399353.:

a. Asdsterminedbythe nethoddscuadln the text. mowezlfilvoustmt with
ICR results from at: laboratory and previouly reported resul ' lthh experimental
unc mmmmmm

b. Double circles lndlcate that a different preference {or these two bench was observed
{or the metal ions studied. In the case of decane. Fe‘ exhibited equal preference for

88

observed in the reactions of Fe" and n-CnHZn+2 (n - 4-7)
previously reported in the literature52o53o57 along with
the Fe+ data.provided in Tables 9 thru 11 indicates that
Fe+ exhibits the same trends as Co+ and Ni+. In general.
terminal C-C bonds have the lowest priority for insertion
in each case.

Why are some C-C bonds in nonpolar alkanes more
susceptible to insertion than others? As we have seen.
for successful insertion to occur. the energy associated
with bond breaking (plus the promotion energy of 11*)”
must be compensated for by the energy released during
bond formation. Of these two processes. direct
correlation between C-C bond strengths in alkanes and the
preferred sites of M“‘ insertion is indeterminate.
.Pertinent experimental data regarding the heats of
formationof alkyl radicals which are needed to
calculate C-C bond strengths are either lacking or else
are indicative of differences which are quite small
(<Skca1)’and.may be insufficient to explain the dramatic
selectivity observed. Any correlation between the bond
making process and the ordering could be determined
through a knowledge of M'"~R bond strengths. Experimental
values of these bond strengths are unavailable for R>CH3.
However, we believe they can be assessed upon

consideration of polarizabilities.

89
A comparison of the experimentally determined values

of D(H+-ll) and D(N‘"-Cli3) reveals that It" forms a stronger

bond toca methyl group than to a hydrogen53v53- Thi'
trend is opposite to that observed for the neutral metal

species and has been explained on the basis of
polarizability. Since CH3 is more polarizable than H,
083 can better stabilize the positive charge on the metal
and in doing so forms a stronger bond to the metal ion.
Building on this example. the polarizability of an alkyl
group (CnH2n+1) would be expected to increase as n gets
larger. We suggest that an increase in polarizability
parallels an increase in D(lf"—Cnll2n+1).

Unfortunately, experimentally determined alkyl group
polarizabilities are not available. However, ionization
potentials. which in part reflect polarizabilities, are
available. In general, the smaller the IP. the larger the
polarizability. Table 24 lists the IP's of cash” from‘
'.n - 4 to n - 10. Note that the IP's decrease as n
increases. Therefore, we suggest that the polarizability
increases leading to an increase in D(H"’ —CnH2n+1) as n
gets larger. This is apparently true for D°(Hg+-R),
n-(01+--R) and D°(xo+—s) as 1: shown in Table 24. As
IP(CnH2n+1) decreases. each of these bond strengths
increases. (This does not appear to be the case for the
analogous neutral species).9 Furthermore, as IP(CnH2n+1)

decreases. the H0M0(Cnflzn+1) -LUMO(M"') gap may decrease

90

e-

..IzIIYIJI:
._ou :Iwcobmvcom .m do: 5qu coy—3 Sun a 6:330 Io Ion-Ea: ..m....> 333 £5 E :02» m3 9.: I306. 05 o
«III e 5qu Iona—omega 9 .8. do: £96 533 5.5 0 .3 Jon E9; :33 3.5 I. ._m .uom 5qu 5:3 3.5 I

III: V 22.6-..

IIIII 2:5...

III I I.III 2:5-..

II; I III: 2:6-..

I: I I.I: :23-..

I.III I.III III I Ii: .55...

.93 I.I: «.2: III I I.III III0-..

I.I... I.I: I.III «I... I I.III I=Io

I.I... I.III I.I: «I... I I.III Izo
o...§.8...§-+u5.c 225.8... 3. +055 22.58“. .3... +8.5 28:8... a. 75 I83.

I.I—«no 2.83 sat; 8 aegis: ....- aszfiaoa 8:58. .868. 5.: .«I 3......

.91
thus facilitating bonding.96

Upon consideration of the bond making process,
direct correlation is found between the sum of the IP's
of the alkyl groups bound to I!" after insertion and the
ordering of the preference of M+ for the C-C bonds of
alkanes. Figure 4 graphically illustrates this
correlation. Note that within the experimental error of
the thermodynamic information available. the preference
of attack is largest when the sum of the corresponding
alkyl IP's is lowest. Priority decreases as the sum of
the IP's increases. 'For example in heptane. the sums of
the IP's associated with the alkyl groups involved in the
various intermediates can be ranked.as follows:

[IP(Cfia)+IP(CoHn)]>[IP(Czfis)+IP(CsH11)
> [ IP( 0311-, )+IP('CI.H9) ]
Based on the experimental results, the smallest sum is
associated with intermediate 0 which accounts for the‘
- ' majority of the products. Thus. the third C-C bond in
heptane. designated as c. is given a priority of attack
of Q. The largest sum corresponds to A which accounts
for the smallest number of products. Therefore. the
terminal C-C bond in heptane is given a priority of @in
Figure A. This correlation suggests that when more than
one metal insertion intermediate of the type R-Mh—R' is
possible, the one in which the M+—-R bonds are strongest

is formed preferentially. We believe that the

’92

 

420 . I . .l .

4:5 ' .
4:0 I . r . . -
405 . . .
460 . . .
395 ~ .

t 1
a“ r- ' d
380 u- ' a

I! "‘1 .
’ / u .

U

U

21PiR-l.kcol/mols

.370 -

 

 

G) O *3 0 6)
Priority of Attack

Figure4. Correlation between priority of attack and the radical ionization potential
sums. The C-C bond for which insertion by M” M 8 Fe, Co, Ni) occurs
to the greatest extent is given a priority of . The bond next most
favored is priority ® , etc. When a C-C bond is cleaved, two alkyl
radicals are formed. The sum of the ionization potentials of these two
radicals correlates with the extent to which M" inserts into that bond.
The correlation is shown for butane ( C ), pentane (O), hexane (I),
heptane (D), octane (A), nonane ( A) and decane ( G).

93

distribution of insertion intermediates reflects their

relative stabilities. For heptane, then, we suggest that

Q is favored over g which is favored over g because:

[D°(M+-C3H7) 4' D°(C3H7M+-C4H9)l > [D°(M+-C2H5) 4' D°(C285M+'C5H11)l
>[D’(M+-CH3) + D°(CH3M+-05H13)l

A slight discrepancy is seen in the trend observed
for decane in Figure 4. The insertion intermediate
involving the lowest sum of IP's is ranked third
(priority III) in preference: This discrepancy can be.
understood on the basis of the uncertainty associated
with the experimental values of the alkyl group IP's.
This uncertainty causes overlap in the values of the sums
of the IP's associated with the three most preferred c-c
bonds in decane by W. Furthermore, as the number of
carbons in the alkyl chain increases. the difference in

the IP of (Cn32n+l) and (Cn+1iiz(n+1)+1) decreases until:

’ . it eventually becomes negligible. Thus. it would appear

that. as in the case of D°(Cl+-R), the difference in bond
strengths. [D°(M*—Cnfizn+1) -D'(M+-C(n+1)H2(n+1)+1)l.
should also decrease with increasing n. Thus, metal
ion-alkyl bond strengths would be expected to converge
when n is sufficiently large. This convergence would
contribute further to overlap in the sums of the IP's as
the chain length of the alkyl groups bound to M+ get

larger. Perhaps then decane marks the transition where

94

direct correlation between the sum of the IP's and
priority of attack cannot be established.

Using this model. the fact that M" rarely inserts
into terminal C-C bonds53o57 can be explained on the
basis of the relatively high IP of the methyl radical
which in turn suggests that the M‘*-Cli3 bond may be the
weakest of the Fri-alkyl bonds. Also, it must be noted
that this_model does not appear to hold for the reactions
of transition metal ions and polar molecules (RX). The
functional group, (X), in polar molecules may have a
strong effect in "directing" the.site of attack of the
metal ion to certain skeletal bonds which would not be
attacked in the corresponding alkanes.(Rl-i).19

Two major problems arise when one attempts to apply
this type of analysis to branched alkanes. First. a
significant number of IP's of the radicals associated

with branched alkanes are not available. Secondly, due

'. to the presence of the branched sites. metal ion

insertion into several chemically nonequivalent C-C bonds
can result in the loss of the same small neutral
molecule. For example, in the reaction of Pi“ with
2,3,4-trimethylpentane, there are two insertion
intermediates which could be responsible for the
observed methane elimination. Reactions (30) and (31)

illustrate this point.

95

H C"
+
'1 I l 3 4

CHal-I

+ _ (30)
‘° WK 3 ..
' ‘ems-g—f—Sfl—a—Glz; —-> io'jr‘t 014

l
(3!)

 

H fo‘ I1.
CH3

Both of the ionic products in reactions (30) and (31)
correspond to an m/z - 157 and therefore cannot be
distinguished in the mass spectrum. Due to complications
such as these, a similar analysis for branched alkanes
. ' will not be attempted.

Finally, we realize that there may be other
explanations for the observed correlation. We believe
that the distribution of observed products reflects the
distribution of intermediates which are formed "early" in
the reaction. This prompts us to ask how the initial
distribution of insertion intermediates is established.
There are at least two possibilities. Perhaps rapid

insertion!"deinsertion" occurs. allowing the metal ion to

- 96
"sample" all possibilities. A second explanation may
involve the formation of the most stable ion-alkane

complexes. Thus, for butane, consider intermediates 2

and 8.

Co” 90*
sarong-carom carom-carom
7 3
~ ~

A consideration of alkyl polarizabilities may predict
that 8 is more stable than 7. Therefore, the
distribution of initial complex structures may determine
which bonds are attacked, rather than the stabilities of

the insertion intermediates which follow.

Chapter 7
Ligand Effects

Several studies have appeared in the literature yielding
useful insights into the reactions of bare (+1) metal ions
and alkanes. As discussed in Chapter 6, product formation in
these gas phase reactions can be rationalized by a mechanism
involving the metal insertion.ls~H shift/ competitive ligand
loss mechanistic sequence. Furthermore metal insertion into
the C-C bonds of alkanes was shown to be favored by Fe+, 60+,
and Ni+ over insertion into C-H bonds.

In condensed phases, organometallic chemistry involves
metal complexes.36v37v97’1°3 Ligands such as CO, N0 and PF3
are frequently utilized in these metal complexes and serve as
important ligands in organometallic chemistry.104 Therefore,
'.it would be useful to investigate the changes which occur in
the chemistry of Fe+, Co"' and Ni+ with alkanes when one or
two of these ligands are bound to the metal, and compare this
chemistry to that observed for organometallic complexes in
condensed phases. Thus, in this chapter, the chemistry of
Fe(CO)n+, Co(c0)n+, ooN0+, Ni(CO)n+, and Ni(PF3)n+ [n . 1,2]
with a series of linear and branched alkanes will be
examined. The products and branching ratios associated with
these reactions are summarized in Tables 2 through 21. These

results will be compared to those observed for the bare metal

97

-98

ions. Parallels to solution chemistry will be discussed.
With the addition of ligands to the metal, several
different types of reactions are observed depending upon the
number and nature of the ligands. FeCO+,CoCO+ and NiPF3+ all
react with alkanes to form some of the same products as the
corresponding bare metal ion. In this case, the addition of
a ligand has no apparent effect on the reactivity of the
metal. we refer to this situation as one where the ligand
acts as a spectator. For example, both 604' and 0060* react
with Z-methylbutane to form CngH5+:
Co+ + A/ -b CoC3H6+ + C2116 (32)
0000+ + /1\/'-OCOC3H6+ + CzH6 + CO I (33)
In contrast to this is the CoNO+ ion which is totally
unreactive. The presence of the nitrosyl ligand apparently

deactivates the metal ion. In addition, FeCO+, 0060+, NiCO+
and NiPF3+ all react with alkanes to form "ligand

substitution" products. This process is illustrated for
. NiPF3+ and 2-methylbutane in reaction (34).

NiPF3+ +b-r N1C5H12+ + PF3 (34)
A metal ion complex containing two ligands only reacts with
alkanes to form "ligand substitution" products in which one
or both ligands are lost. For hexane and Ni(CO)2+, reactions
(35) and (36) are observed.

N1(c0)2+ um ENICOC6H14+ + co (35)
N1C6H14+ + ZCO (35)

99

Thus, in general ML+ reacts to form many of the same products
as W”; both ML+ and ML2+ are involved in ligand substitution
reactions.

There are several considerations which must be taken into
.account in order to evaluate any ligand effects which may be
associated with these reactions. First, it would be useful
to know at which point in the mechanism the ligand is lost in
these reactions. Assuming that the same mechanism operative
for H" also applies to ML+. Scheme 5 illustrates a potential
series of reactions occurring between CoCO+ and hexane. Note
that there are at least four different points in.Scheme S
where CO could be lost. Furthermore. it is evident from
Scheme 5 that the structure of the ligand substitution
product depends upon where in the overall reaction sequence
the CO is lost. Thus, if there was sufficient energy
associated. with intermediate .3” the initially’ formed
CoCO(alkane)‘+ complex, simple ligand substitution cduld
' occur. In this case the structure of complex $2, the
ligand substitution product would involve a Co+ and an intact
hexane molecule. Alternatively. if it is energetically
feasible. CoCO+ could insert into a C-C bond of hexane to
form an intermediate such as AI}. In this case, loss of CO
could occur either before _the alkane has a chance to
rearrange on the metal center (intermediate‘Af) or after
alkane rearrangement (intermediates i3 and \J’Z)- Note
that if there is any excess energy present in these ligand

substitution products, they may form products which are

100

 

 

 

._ J ._. J z a
_
m m I m .. I m .. \
=8 ..fiatl zu+8_ AI Ions/J
8 8
S. 4 £3
ozmu I ...8 All N_imu. 8&1"... LGu- 8\
+ + W. + / J
. 8 8
2 2 Z .2
.2 = : S a

 

on .. 2-. Z- -2 I; so 3...
:9... =mullzu+8n=\lll:u+8 191+1934+19881r :u++88
8

m uzuzum

101

identical to those observed for Co+ alone. These processes
would account for the spectator type reactions observed for
CoCO+ and hexane. Before correlating the branching ratios
for CoCO+ with intermediates as we did with 60+ in Chapter 6,

it is necessary to determine whether

the ligand substitution
product is best represented by a complex such as A9
(wt-1212') or by an insertion complex such as )5 (R-W-R').
Note that previous studies in the area of gas phase
organometallic chemistry have concentrated heavily on the
reactions of metal and metal containing ions with polar
organic molecules such as isopropyl cthride. It has been
determined, that Co"' reacts with isopropyl chloride to form

an intermediate in which propene and HCl compete as ligands

on the metal: 1

. oo*+ )—cu *kci-c1»/Ij--c'o*-c1—-— /l[—--co*---cm
H H I8
~
J—-Co‘--—cm -'--> Ju-cci +ch
H ' '
‘V
Following its formation, Intermediate 2'3 loses HCl to form
the final product. It has been noted in these reactions that
when two ligands such as 1101 and propene compete as ligands
on the metal, the ligand with the largest proton affinity

(PA) is preferentially retained on the metal. According to

102

the PA rules, the larger the PA of a ligand, the better its
electron donating ability. Thus, since W is election
deficient, it retains the ligands which are the best bases in
these reactions. Since the PA of HCl is less than that of
propene, HCl is lost in the above reaction. Furthermore, the
PA rule also applies to the ligand.substitution reactions
observed for these polar organic molecules. Thus when COCO+
reacts with a polar organic molecule such as methyl iodide
Coco+ + c1131 —-» Co+cocn31 --t Co+cn31 + co
CH3I and CO compete as ligands on the metal. Since the PA of
CH3I is approximately 171 kcal/mol and that of C0 is 142
kcal/m0133, CO is preferentially lost in this reaction to
form the ligand substitution product.

Apparent ligand substitution products are also observed in
the reactions of metal-containing ions and alkanes. The
following reaction is observed for methane:

0000* + CH4 —: Com-14+ + co
--However, the PA of methane is only 132 kcal/mol which is°less
than the PA of CO. Therefore, the observed loss of CO in
this reaction is in violation of the PA rule. This would
suggest that either the PA rule does not apply to the
‘reactions of metal-containing ions and alkanes or the ligand"
substitution product formed in the above reaction with
methane does not involve an intact alkane molecule. If the
latter were the case, the general process presented in

reaction (37) would illustrate the apparent ligand

. 103
substitution process occurring for alkanes, where the final
product involves a metal ion bound to two alkyl ligands or in

the case of CH4, the product would be (HCo+Cl-i3).

/co
Coco+ + RR' —+ c6” ——9 Co+-R' + co (37)
I \

R R' a

To evaluate the feasibility of the formation of a metal
ion-intact alkane complex such as 38 in Scheme 5, the first
part of this discussion will examine the energy associated
with the interaction of a positively charged metal ion and a
polarizable alkane. Next, a comparison of the chemistry of M‘"
and ML+ with alkanes will be made; their "reactivities" will
be compared and contrasted. (Note that reactivity is
frequently defined in such studies by the number of different
products which are formed, since this is generally indicative
of the number of bonds into which a given metal will insert
in an organic molecule). Product distributions will‘ be
-evaluated to determine whether the observed ordering of
preference of ML+ for the C-C bonds of alkanes parallels that
observed for W. Differences in this ordering will be
assessed in terms of both steric effects and the bond
strengths associated with the ligands on the metal. Obvious'
steric effects may provide evidence supporting the existence
of a metal ion complex such as 9.3, in Scheme 5. For
example, if the ML" insertion complexes involving the least
amount of steric hindrance are formed preferentially, then

the number of bonds into which ML+ inserts may be limited.

104
Under these circumstances, the reactivity of ML+ would be
expected to be different than that observed for M+. In
comparison, one would expect any steric effects associated
with the formation of a complex such as .13 to be less
significant than those associated with the ML+ insertion
complex (’13) since L is lost before the insertion step.
Finally, parallels between the gas phase chemistry observed

and solution chemistry will be made.

ALKANE POLARIZABILITIES

If the energy released during-the initial interaction of a
positively charged metal ion (ML"’) and a polarizable alkane
is greater than the M+-L bond strength, then the formation of
a metal ion-intact alkane is feasible and reactions such as

Coco+ + c611“, -) Co “1114+ + so

can be considered to be ligand substitutions. To calculate
this energy, alkane polarizabilities must be knewn.
'.Therefore, methods of obtaining alkane polarizabilities
followed by the energy 'calculations based on these
polarizabilities will now be presented. (Note that we are
assuming here that the presence of L on W does not alter the
charge on W. Thus we are assuming that the energy"
associated with the interaction between ML+ and the alkane is

the same as that for W and an alkane).

105

Empirical methods have been established for calculating
average molecular polarizabilities)”:106 The values of the
alkane polarizabilities, a, which are utilized in this study
can be calculated using equation (14) which was formulated by

Miller and Savchik. 105

«0.53%

N in this equation equals the number of electrons in the
molecule and ‘A represents the atomic hybrid component (ahc) .'
of polarizability for each atom, A, in the molecule. .Each'
atom in the molecule is represented by one value of Q; in
equation (14). These values have been determined by Miller
and Savchik and are provided in reference 105. The value of 7:

coinciding with a particular atom depends upon the
hybridization of that atom. For example. there are three
different values of 1A for carbon because carbon can be

either sp3,sp2 or sp hybridized.

106

The polarizabilities calculated from equation (14) have
been found to be within one per cent of the experimentally
determined values for most alkanes.105 Therefore, the
advantages of using this method lie in its simplicity --only
a few parameters are necessary, and in its accuracy. However,
one of the major shortcomings of this approach lies in the
fact that TA does not depend on the atoms which are bonded to
.A. As a result, the polarizabilities calculated for isomers
such as pentane and neopentane are identical. For the purpose
‘of this discussion, the values calculated for a are taken to
be the average molecular polarizablities of the linear
alkanes.105 These values are listed in Table 25.

The potential, U, associated with a positively charged ion
separated from a polarizable alkane by a distance of r can

now be calculated using equation (15) where zs is the charge

U(r): _a2 6 <15)

of the ion. (Note that equation (15) is valid for r>> radius
of the molecule).107 For simplicity, let us first assume
that we are calculating the potential energy of a point

charge separated from a spherically symmetric polarizable

107

L

 

 

 

 

 

 

 

ann.o ca.m m.on oaa awe Imam In.HH ocsooa
mwu.c Ho.o o.mn cow men omou oe.c~ osmcoz
sm~.c NH.< o.~n Hod won enmu mn.a accuse
a-.c no.n c.o~ III use scum I~.a saunas:
oau.c ea.n N.e~ MNH Ian Gama «I.I cacao:
oo~.o no.~ c.o~ non nun anon Ho.o ocmucom
mnH.c ca.~ I.IH n.ew cow mend om.e samusm
cc~.o se.H a.NH ~.wo can mean ms.n usadonm
nnso.o oH.H oo.o c.oc med was me." museum
~neo.¢ «II.0 nn.n o.sN «.mw «me sm.H census:
«3:. «an» «nus «Tn «main «Tu 38:53 1%
Ios\IauI have-
omuono acuom s can: meduosuoucn unsung: accede you
ucoduoasoHou amuse» one huaaaaouancnom Io muasnom .mn canes

 

108

point by distances between 1 and 10 angstroms. The results of
these energy calculations are tabulated in Table 25 along
with the' polarizabilities. It is apparent from the average
molecular polarizabilities that, the larger the alkane, the
greater the polarizability. Furthermore, it is evident from
U(r) that these ion/molecule interactions fall off rapidly
with distance and thus cannot be considered long range.
However, at distances less than 3A, interaction energies
greater than 100 kcallmol are calculated for alkanes larger
than butane.

Table 26 lists the M+-L bond strengths which have been
reported in the literature. (Note based: on the available
heats of formation reported for Ni+, Ni-CO+ and CO, D(Ni+-CO)
was calculated to be only 6.6 kcallmol. This seems
unreasonably low compared to the bond dissociation energies
of the other metal carbonyls presented in Table 26 which were
determined experimentally. Nonetheless, this value is
' included in the Table. Furthermore, the trends reported 77
for D(M"’—-CH3) and D(M+-li) [where m - re+, Co+, Ni+] would
suggest that

[D(Fe+-CO)] > [D(Co+-CO)] > [D(Ni+-CO)]
[Trends for M+-li may not be relevant to M+—C0 depending on
the nature of the M+-CO bonding which is currently not well
understood]. Therefore, D(Ni+-CO) is probably less than 40
but greater than 6.6 kcallmol]. The information provided in
Figure 5, which graphically illustrates the information

contained in Tables 25 and 26, suggests that at least one CO

109

Table 26. tf"-L Bond Strengths

 

li - L . Bond Strengthfltcaljmou
Fe+-CO ' 1638 '
Co+-CO 40"

Ni+-CO 6.6c

Ni+-PF 75‘1

a. Taken from Ref. 29
b. Taken from Ref. 108
c. Taken from Ref. 82
d. Taken from Ref. 109

110

I0,000

A A44Al

. Decane
none at.
Octane
Neptune
Heretic.
Pentane
But‘fl‘
Propane.
Ethane
.mflnnc.

..IODIOODO

   

 

 

 

 

 

l‘ (A)

4.
Figure 5. Plot of U(r) vs. r, Including D(M - L).

111
or PF3 ligand on P!" can be displaced by any alkane larger

than propane at distances of separation of approximately 2A.
Unfortunately,the above analysis involves two major
oversimplifications. First, an ion is not a dimensionless
point charge. It has a radius which will limit its minimum
distance of separation from an alkane. For example, the
radius of Co"' is approximately 0.7A110 and that of hydrogen
is at least 0.5A.111 The sum of these radii would suggest
that the minimum distance possible between Co"’ and a hydrogen
atom in an alkane is ~1.2A.- Therefore, the energies
calculated for separations of 1A in Table 25 have little
physical significance. Furthermore, alkanes larger than
methane cannot realistically be represented as symetrical
polarizable spheres. Each alkane has a specific geometry
associated with its lowest energy configuration. If we
consider these geometries in three dimensional space, it is
apparent that the distance between a "fixed" charge and ‘the
"_various portions of the alkane molecule cannot be represented
by a single value. That is, a charge situated 2A away from a
hydrogen located on a terminal carbon in hexane will be more
than 2A away from the terminal carbon on the other end of the
molecule. Therefore, to more realistically assess the energy
involved in the interaction of an ion and an alkane, we need
to evaluate the average molecular polarizability as a series
of components, where each component represents the
polarizability of a certain portion of the molecule. Then,

the individual interactions of the charge with the various

112

polarizable parts of the molecule located at various
distances from the charge can be calculated.

One method making this possible involves the calculation
of the average molecular polarizability as a function of bond
polarizabilities.106 According to this method, each bond in
the molecule has a specific polarizability associated with
it. The polarizability of the bond in the direction
perpendicular to the bond differs from the polarizability in
the direction parallel to the bond. The parallel component
of polarizability is represented by the symbol cha. oi
represents the . perpendicular component of bond
polarizability. an and on have. been reported by Denbigh for
various bonds.112 For example, 0‘" for the C-C bond equals
1.13cm3/mol and.ai equals l.20xlO"2 cm3/mol. If an electric
field makes an angle 9 with a chemical bond, the
polarizability of the bond, 09, can be calculated using
equation (17).

2 .
a. = GHCOS 9 'l- a‘Slnzo (17)

Averaging a. over all orientations leads to an average value

of the bond polarizability, a:

113

a: 1(au +2ai) (18)

3

The average polarizability of the molecule is then taken as
the sum of the average polarizabilities of its bonds. Note
that the average polarizabilities obtained using this
method1132 are in agreement with those calculated using
equation (14).

One of the limitations of this method lies in.the fact
i that it cannot be applied successfully to molecules involving
resonance structures.

In the simplest application of this method, one can first
imagine a molecule in three dimensional space and assign to
each atom in the molecule three coordinates. Second, one can
define the location of a charge, placed near the molecule, by
three coordinates. Third, one can calculate the angle 9
established by the field lines associated with the charge and
the center of the polarizable bond. Once these angles are

known, equation (17) can be used to calculate a.. Fourth, one

114

can calculate the various distances between the "fixed"
charge and the various bonds of the molecule. These
distances can be used in equation (15) along with the
respective a.'s to calculate the specific U(r) associated
with the charge and a particular bond. Finally, the sum of
all of the U(r)'s can be obtained. This sum represents the
total energy of interaction associated with that particular
molecular configuration and charge location.

It is evident from the number of variables which must be
defined in order to perform this calculation, that the larger
the alkane, the more potentially complex this analysis. It
was not within the scope of this study to perform exhaustive
calculations. Rather, we wish to determine whether it is
feasible for an intact alkane to displace a CO or PF3 ligand
from a metal ion. Thus, calculations were performed for
various configurations of butane. Each alkane larger than
butane consists of at least one "butane unit" plus additional
' C-C bonds. Thus, if the calculations reveal that it is
theoretically possible for butane to displace L,then it will
be assumed that alkanes larger than butane can also displace
L. 1

The calculations were carried out assuming a "boat"
configuration for butane. (Note that the "chair"
configuration of butane actually corresponds to a lower
energy configuration. However, we assumed that the energy
barrier to rotation about the central C-C bond separating

the "chair" from the "boat" configuration was approximately 3

115
kcallmol. In using the chair configuration we were able to

maximize the number of interactions between the charge and
the six terminal C-H bonds of butane). According to Figure 6,
the skeletal chain of butane is situated in the plane of the
page. Furthermore, A, B, C and D represent four different
locations in the plane of the paper where the positive charge
can be placed. As a reference, the field lines set up by the
charge at either location A, B, or C will make a 90’ angle
with the C-C bond directly in "front" of the charge. The
field lines set up by the charge at point D will make a 0"
angle to the C-C bond "below" it. The arrows in Figure 6
point to these respective C-C bonds. Each carbon and hydrogen
atom is numbered for future reference. Calculations of U(r)
were carried out for "arrow lengths" of l or 2A. All other
distances and angles were measured relative to these
separations of 1 and 2A. For example, Figure 7 shows all of
the angle and distance measurements pertinent to the
. calculation of U(r) for a charge located at point B, 2A away
from the internal C-C bond. (Recall that in actuality
separations of 1A are physically impossible based on atomic
and ionic radii. However, these calculations at short
distances should serve as a gauge of interaction, i.e., if
the energies calculated for these short distances are less
than D(M+-L), then at longer distances there is no chance of
an intact alkane displacing L). A breakdown of the energy of
interaction between a charge located at either position A, B,
C or D and the individual C-C and C-H bonds in butane is
provided in Tables 27 and 28 for "arrow lengths" of 1 and 23.

116

 

 

Figure 6. Relationship between Charge Location

and c-c Bonds of Butane

117

 

 

 

Figure 7. Angle and Distance Measurements

For Charge Located at B

118

Thus, if a charge is located at point B in Figure 7, its
energy of. interaction with either the 01-02 or C3-C4 bonds of
butane, according to Table 27 is -l3.1 kcallmol. However, its
overall energy of interaction with all of the C-C and C-H
bonds of butane is -124 Kcal/mol. It is evident from Tables
27 and 28 that the overall energies calculated at short
distances are greater than the M+-L bond strengths provided
in Table 26. However, once the charge is moved back an
additional Angstrom, only one out of the four locations, B,
provides for an interaction energy (124 kcallmol) which is
greater than the M+-L bond strengths indicated in'Table 26.
In general, a consideration of the energies associated with
the interaction of MT and the individual C-C and C-H bonds in
Tables 27 and 28 reveals that the strongest interactions
occur with the C-H bonds which are located closest to the
charge.

In terms of the chemistry which is observed for butane,
perhaps those bonds which experience the strongest
interactions with 14+ are weakened in the process thus
facilitating their cleavage. Since C-C bonds are lower in
energy than C-H bonds, let us consider the energy of
interaction associated with the C-C bonds in butane using
the values in Table 27. There are various approaches which
could place- the metal ion near the terminal or internal C-C
bonds of butane. A, B, C and D represent these various

approaches. The strongest interaction when the ion is

Table

119

- U(r) kcallmol

For “Arrow Length" of 18

Bond by Bond Breakdown of U(r)

 

 

 

BOND A B C D
Cl-C2 3.3 4.6 18.9 0.2
Cz-C3 19.0 3.3 3.3 1.5
C3-C4 0.8 4.6 18.9 33.2
01-81 72.4 10.5 '1.3 0.9
01-82 21.6 15.9 1.3 0.9
01-83 21.6 15.9 1.3 0.9
C2-84 21.6 18.7 28.0 0.6
Cz-H5 21.6 18.7 28.0 0.6
€3-36 1.4 18.7 28.0 1.7
03-117 1.4 18.7 28.0 1.7
C4-88 0.6 10.5 1.3 120.7
04-H9 0.6 15.9 1.3 120.7
C4-H10 0.6 15.9 1.3 120.7
Total U(r) 186.5 172 160.9 404.3

.Table 28.

120

- U(r) kcallmol

Bond by Bond Breakdown of U(r)

For “Arrow Length“ of 23

 

 

noun A s c 0
01-c2 0.2 13.1 2.3 0.2
c2-c3 4.6 0.2 0.2 0.5
C3-C4 0.3 13.1 2.3 5.4
cl-H1 13.1 ‘9.9 0.5 0.4
cl-H2 '3.8 15.8 0.5 0.4
C1-H3 3.8 15.8 0.5 0.4
cz-u4 3.8 3.6 3.2 0.2
CZ-HS 3.8 3.6 3.2 0.2
C3-H6 0.5 3.6 3.2 0.6
c3-a7 0.5 3.6 3.2 0.6
ca'Hs 0.2 9.9 0.5 7.8
C4-H9 0.2 15.8 0.5 7.8
era10 0.2 15.8 0.5 7.8
Total U(r) 35 124 20.6 32.3

121

located at point A occurs with the 02-03 bond. Weakening of
this bond may result in the formation of the [C2H5-m-C2H5]
insertion intermediate. This intermediate leads to ethane
(or H2) elimination in butane. When the ion is located at
point B, there are no significant interactions with any of
the C-C bonds. However, the ion interacts strongly with both
the Cl-Cz and C3-C4 bonds when located at point C. This
interaction with both terminal bonds may cause a subsequent
weakening in the 02-03 bond leading to the formation of an
insertion intermediate which again can be represented as
[Csz-M+-C2H5]. 'Finally,the'strongest interaction that
occurs with a C-C bond when a charge is located at point D
occurs with the 03-04 bond. Weakening of this bond may
result in the formation of the [C3H7—M+-CH3] insertion
intermediate which can subsequently undergo a 8-H shift and
eliminate methane. However, at location D there is an even
greater interaction between the charge and the three terminal
I C-H bonds closest to it. This interaction may facilitate the
formation of an intermediate such as [CaHg-M'I-H] which could
undergo H2 elimination.

Since D(C-H) is larger than D(C-C), a larger energy of
interaction may be necessary to break the stronger C-H bond.
Finally, we can use Table 27 to obtain the sums of the
energies of interaction between the charge and all three of
the C-C bonds in butane for a particular charge location.
These results are summarized in Table 29. For example, when

located at point A, the total energy of interaction between

122

Table 29. Sum of the Energies of Interaction Between

a Charge and the three C-C Bonds of Butane

Total Energy of Interaction Between

 

 

Location Charge and C-C Bonds (kcallmol)!
A . 23.1
B ‘ . 12.5
C 41.1
D . 34.9

a. Based on values in Table 27.

123

the charge and the three C-C bonds of butane is 23.1
kcallmol. A comparison of these sums reveals that location C
provides for the largest energy of interaction. Recall that
the insertion intermediate proposed earlier for this
interaction was [C2H5-M+-CZH5]. From Table 23 in chapter 6,
the C-C bond most preferred by M" for insertion into butane
is the internal 02-03 bond. Location D provides for the next
largest energy of interaction. The insertion intermediate
proposed for this interaction is [C3H7-M+-CH3]. According to
Table 23, the Cl-Cz bond in butane was the next most
preferred bond by M+ for insertion. The interaction at point
A is weaker than that at point D.' Furthermore, location B
provides for the weakest energy of interaction between the
charge and the three C-C bonds of butane. Interestingly,
these results suggest that there is a correlation between the
bonds most preferred by M+ for insertion (as presented in
Table 23) and their energies of interaction.

Overall, these energy calculations, which take into
account the size of the metal ion and the alkane molecule
geometry, do suggest that it is feasible for an intact alkane
larger than propane to complex with a metal ion and displace
at least one CO or PF3 ligand.

In addition to polarizability calculations, we note that
the electron density associated with the carbon and hydrogen
atoms of alkanes can be partitioned into charges on these
atoms.113 Therefore, we must also assess the significance of

the interaction between the charges of the atoms of the

124

alkane (dipolar bonds) and the charge of the metal ion since
ion-dipole interactions are longer range than ion-induced
dipole interactions and could be of relatively high energy.
Equation (19) can be used to calculate the potential energy,

U(r),between two charges separated by a distance, r (where

U(r) - z 2 £2
1 2
r

21 is the charge of the carbon or hydrogen atom and 22 is the
charge of the metal atom). Values for the atomic charges
associated with a select set of alkanes can be found in the
literature.113 These values are derived from ab initio
calculations. A description of the methodology behind these
ab initio calculations is beyond the scope of this discussion
and will not be presented here. For our purposes, it is
necessary to know that the atomic charge of a primary
hydrogen in butane is -32.4 millielectrons (me) [lme - 10'3
electron] that of a primary carbon is +89.2 me, that of a
secondary carbon is 84.8 me and that of a secondary H is
-38.4 me.113 Using these values in equation (19), we can
determine that the potential energy between a (+1) metal ion
located at point B in Figure 7 and butane is 20.6 kcallmol.

The individual interaction energies between M'“ and the C and

125

H atoms of butane under these circumstances are presented in
Table 30. The overall energy here represents a repulsive
interaction in contrast to the polarizability results.
Therefore, this model predicts that the approximate
interaction energy of M+ and butane in Figure 7 is 103.4
‘kcal/mol. Since 103.4 kcallmol is greater than the M+-L bond
strengths in Table 26,it is still feasible for an alkane
larger than propane to displace L while remaining intact.
Further evidence in support of the significance of alkane
polarizabilities in these gas phase reactions can be found in
the reactions of alkanes with.ML2+.. Theoretically, iflkd
is sufficiently large, an intact alkane will be capable of
displacing more than one ligand on the metal ion. Note, as
the length of the alkyl chain increases, ligand substitution
patterns for ML2+ change. Thus, alkanes larger than heptane
always displace two L's. Alkanes smaller than hexane only
displace one L. These results are presented graphically in
' Figure 8. For example, pentane reacts with Co(CO)2+ and

displaces only one carbonyl ligand.
Co(CO)2+ + Pentane -—9 COCOC5H12+ + CO

Apparently, the energy obtained in the interaction of pentane
and Co+-(CO)2 is only large enough to cleave one Co+-CO bond.
IHowever, when hexane reacts with Co(CO)2+, either one or two

of the carbonyls are displaced.

CO(C0)2+ + Hexane ‘[: COCOC6H14+ + CO
COC6H14+ + 200

Furthermore, when alkanes larger than hexane react with

V 126

Table 30. Interaction Energy Between H‘

and the Carbon and Hydrogen atoms of Butane

 

 

ATOM U(r) kcallmol
CI 21.1
C2 13.2
.C3 13.2
C4 . 21.1
Bl -4.4
Hz -5.1
H3 -5.1
H4 -4.7
H5 -4.7
H6 -4.7
H7 -4.7
H8 -4.4
Hg -5.1
H10 -5.1

 

Total U(r) - 20.6

MAXIMUM
NUMBER OF
LIGANDS
DISPLACED
IN M(00)§

Us: :]

MAXIMUM
NUMBER OF
LIGANDS
DISPLACEDI*
IN N1(PF3)2

 

127

 

 

 

 

 

 

 

Figure 8.

Ligand Substitution Patterns

7
2.. . - - 1 _1
l 7 “ ‘ f 5 -
II
I
I
I
0 fun]! I 1 I I I I I l
l 2 3 4 5 b 7 B 9 H)
n [CnHZn+1]
1 . - - -
o I V I I I 1 T r 1 l
I 2 3 4 5 6 7 a 9 u)
n [an2n+l]

128
Co(CO)2+, loss of both carbonyls is observed.

06(00)2+ + Heptane —§ COC7H16+ + 200

06(00)2+ + Octane —+ c6c3813+ + 200
These results suggest that the greater polarizability of a
larger alkane allows for a stronger interaction with M+,
resulting in a release of energy capable of breaking both
Co+-CO bonds. (Recall from Table 25 and Figure 5 that the
average molecular polarizabilities of linear alkanes do
increase as the alkyl chain length increases thus
substantiating the above claim). Also, recent work reported
by Larsen and Ridge suggests that the structure of the ligand
substitution product formed in reaction (38)

Fe(CO)2+ + Butane --)FeC5H100+ + CO (3%)

involves a Fe+y CO and an 13529; butane molecule. The major
peaks they observed in the collision induced decomposition

spectrum of FeC5H100+ correspond to Fe+,FeCO+ and FeC4H10+.68

Thus, their results also suggest that an intact alkane‘can
displace a ligand on the metal.

Metal ions containing more than two ligands are found to
be totally unreactive in terms of cleaving bonds in the
alkane. Perhaps the presence of three or more ligands on the
metal center prohibits the metal center from getting
sufficiently close to the alkane. Under these circumstances,
the metal center and the alkane have no opportunity to
interact. Note that similar reasoning has been used to
explain the lack of reactivity observed for Fe(CO)n+ [n =
3-5] and 12-crown-4.6o In this case, Fe+, FeCO+ and Fe(C0)2+

129
react with lZ-crown-4 to formiproducts whereas Fe(CO)n+ [n =

3-5] does not.
COMPARISON or THE REACTIVITY or M" AND ML+

Now that we have determined that it is possible for an
alkane to displace a CO or PF3 ligand before metal ion
insertion, let us compare the reactivity of M+ andlflfi'in
terms of the number of different products which are formed by
each metal species. to determine any steric effects which may
be indicative of limited MLI ihsértion into alkanes. An
inspection of Tables 4 through 21 reveals that the reactivity
of ML+ is less than that of M". However, in the reactions
where M‘" and ML+ are precursors to the same product ions, it
can be shown that ML+ and MI exhibit the same preferences for
the C-C bonds of alkanes using the same type of analysis

presented in Chapter 6. Thus, Table 23 of chapter 6 which
- indicates the ordering of the preference of 14+ for the'C-C
bonds of linear alkanes also applies to ML+. The decrease in
the reactivity of ML+ can be attributed to the fact that the
metal ion-alkane complex formed by ML+ is lower in energy
than the metal-ion alkane complex formed initially in the
reaction of 11+ and an alkane by an amount equal to D(M+-L).
Thus, the data in Tables 4 through 21 further reveal that
NiPF3 is less reactive with alkanes than NiCO+. Reactions
(3?) through (42) with pentane serve to illustrate this

point.

_ 130

NiCO+ + Pentane I‘liczgi‘lg+ “’ CZHB + C0 (37)
NiC5H10+ + H2 4 co (2,0)
New“ + CO (41)
+ N'C H + + PF
NiPFa + Pentane -* 1 5 12 3 (42)

Thus, NiCO+ forms three products and NIPF3 is a precursor for
only one product ion. In addition, according to Table '26,
D(Ni+-PF3) is greater than D(NiT-CO), ‘ These results suggest
that there is a correlation between the reactivity of ML+ and
the M+-+L bond strength -- the stronger the M+L bond, the
less the reactivity of ML+ relative to W. Therefore, the
NiT-pentane complex formed by NiPF3+ is less energetic than
the Nit-pentane complex formed by NiCO+ which is less
energetic than the Ni+-pentane complex formed by Ni+. As a
result, the following trend in reactivities is both expected
and observed: -
Ni+ > Nico+ > NiPF3+ (43)

In general, the lack of obvious steric effects in these
reactions may indicate that insertion of ML+ into C-C bonds
is a less favored process than insertion of 14+ into alkanes

following the displacement of L by an alkane.

13 1
STERIC EFFECTS

One interesting trend which we do attribute to steric
effects is observed in the reactions of 1100* and alkanes.
MCO+ reacts with alkanes larger than hexane to eliminate more
H2 than M". For example, thirty-one per cent of NiCO+ reacts
with heptane to eliminate H2 whereas only nine per cent of

Ni+ reacts to eliminate H2:

97.
111+ + Heptane —) NiC7H14+ + H2 (44)
317.
Nico+ + Heptane —-) NiC7H14+ + H + co (45)

These results suggest that reactions (44) and (45) involve
two different processes for H2 elimination. We suggest that
H2 elimination in reaction (46) results from the initial
insertion of MCO+ into the terminal C-H bonds of these larger
- alkanes. This is unlike M’" which preferentially inserts into
C-C bonds. We believe that insertion into C-C bonds by MCO+
is prohibited due to steric effects. Consider the following

reaction:

A C0
1‘53 / ‘\
1800* + A-B -> A M* SE5 ——> Products (46)

 

132

In this reaction SE; refers to the energy associated with
the steric effects produced by two adjacent ligands. The
energy, H insertion, required for insertion in reaction (46)

can be written as:

Hingeption = D(A-B)-[D(-M+CO-A) + D(AH +CO-B)] + Es (47)

where ES 3251:, . [Note that reaction (47) involves the
energy associated with bond making, bond breaking and the
steric effects.) Based on chapter 6, if the
thermo dynamically most stable insertion intermediate is
formed in reaction (46). M+-C and M'I—H bond, strengths suggest
that insertion into C—C bonds is preferred. However, if the
steric effects (Es) associated with the C-C insertion
intermediate are more significant than E3 associated with the
C-H insertion intermediate, MCOT may show a preference for

insertion into C-H bonds. Furthermore, insertion into a
terminal C-H bond should minimize steric effects in the
intermediate in reaction (46). Thus, for heptane, we suggest
that Ni+ insertion into C-C bonds leads to more stable
intermediates than insertion into C-H bonds. Therefore, in

terms of stability:
AHf(C3H7-Ni+-C4H9) >Anf(h-Ni+-07H15) (48)
This implies:

[DOW-cam) + ntcanqmt-cwgn > [0(Ni+-H)+D(nNi+-C7H15)l

133

NiCO+, on the other hand, reacts with heptane to form an
intermediate which favors increased H2 elimination. This

suggests that the formation of intermediate 19

 

 

is preferred over the formation of intermediate 13
re (a.
3H7-Ni
\ 2
C439

- J
implying that 85 (18) < E5 (13) such that

 

 

iD(Ni+co-H) + D<HNico+-c7nls)-ss(pl >

[D(Ni+CO-C3H7) + 0(C3H7Ni+-coc,H9)—ssqg)l

ExperimentalLy, steric effects are assessed through bond
strengths because SE5 cannot be
determined individually. Thus, based on the results ob-
served for Ni+ and N100... with heptane we can write:
[D(C3H7-Ni+-H)] > [D(07H15C0Ni+ -H)]
and
[13(c7li1 581+ 41)] Q- [0(c7H15m~:i+-H)]

These trends in bond strengths suggest that there is a

134

greater steric effect associated with the insertion of NiCO+
into a C-C bond than into a C-H bond. Thus, we believe that
NCO+ preferentially inserts into the terminal C-H bonds of
large alkanes to eliminate H2. There is only one large
alkane that NiPF3 reacts with to eliminate more H2 than Ni+.
This alkane is nonane. Thus, we suggest that insertion of
NiPF3+ into the C-H bonds of alkanes may be less favored due
to the steric effects which result when a bulky PF3 ligand is
adjacent to a bulky alkyl group on Ni+. Note that steric
effects can be correlated with an easily measured parameter
known as the cone angle 0.104 The cone angle for symmetrical
ligands is illustrated in Figure 9. 9 is defined as the apex
angle of a cylindrical cone which just touches the van der
Waals radii of the outermost atoms in the ligand. For
comparative purposes, the value of the ligand cone angle for.
PF3 is 104’, whereas 6 - 95° for 00.114 Furthermore, for H,

9 -7S°, for CH3, 6 8 90’ and for Csz, the cone angle is
‘ 102°.114

 

Figure 9. Definition of the Cone Angle

135

One interesting point raised by MCO+ insertion is: why
would NCO+ insert into a C-H bond if calculations of the
average molecular polarizability and U(r) indicate that it
should be easy to displace a CO before insertion? From Table
6 it is apparent that the energy of interaction U(r), between
1!" and butane is dependent upon the location of 14+ relative
to the C-C and C-H bonds of butane. Thus, certain locations
of 16" result only in weak interactions with the alkane which
are too small to sufficiently compensate for the dissociation
energy of the M+-L bond. Under these circumstances it is
possible for MCO+'to remain "intact", and insert into a C-H
bond. Interestingly, the gas phase reactions recently
reported for MH+56 and MCH3 and alkanes63 indicate that these
two metal ion-ligand complexes insert exclusively into the
C-H bonds of alkanes unlike W. For example, in the reaction
of NiD+ and ethane, NiCsz+ and HD are formed as the products
(reaction (49)). ‘

NiD+ 021:6 —. NiCzH5+ + m) (49)

These results suggest that insertion of NiD+ into a C-H bond
occurs as a first step followed by the shift of a 8-H onto
the metal. HD is eliminated in the last step. Note that H2
elimination is not observed in the reaction of Ni+ and
ethane. Apparently the presence of the H ligand on Ni+ opens
up new pathways for reaction just as the presence of CO on M"
opens up new pathways for H2 elimination.

Recall from chapter 5 that the possibility of 0000*

136

forming stronger bonds to amines than 00* was raised.
Similarly, MCO" may form stronger bonds to alkanes than 11'"
making insertion into the higher energy C-H bond favorable.
[Note that D(C-—H) -1o3 kcallmol and 0(0-0) . 85 kea1./m61].82
Promotion energies9o indicate the amount of energy required
by M+ to achieve an electron configuration conducive to
two-bond formation. Perhaps less promotion energy is required
for MCO+ relative to M+ due to the presence of the ligand
which causes the electron configuration of NCO+ to be

"inherently" more conducive towards bonding to other ligands.
This might explain why CoCO+ reacts with methane and ethane
(Tables 2 and 3) and Co+ does not react at all. The ionic
products of reactions (51)and (53) may involve a Co+ bonded

to an H and an alkyl group.

c6+ + CH4 -—>N.a. _ (50)
Coco+ + CH4 ->Cocn4+ + (:0 (51)
06* + cans —->N.R. (52)

Coco+ + CZHG —> CoC2H5+ + C0 (53)

Furthermore, perhaps the concept of promotion energy can be
used to explain the lack of reactivity observed for CoNO+ and
alkanes.

Interestingly, 2,2-dimethylpropane,2,2 dimethylbutane and
2,2,4,4-tetramethy1butane all react with CoCO+ to form a

product in which. the CO ligand is retained while a smaller

137

neutral alkane is eliminated. These are the only examples in
which C0 retention in the product ion is observed. Each of
these branched alkanes contains at least one neopentyl group.
These results suggest that it is possible for CoCO+ to insert

into the C-C bond of branched alkanes containing a neopentyl

group.
06(00)+ + xix _ -—-r 06051180+ + (1:4 (54)
06(00)+ + N ——-3 06051418 0 + 041110 (56)

Thus, once CoCO+‘inserts into a C-C bond, a s-hydrogen
shifts, followed by the elimination of.a small neutral
alkane. Note that Co+’reacts with these branched alkanes to

eliminate the same small neutral alkanes as CoCO+ does.
RELATIVE RATES

Reactivity can also be considered in terms of relative
rates. Here we will define the reactivity of ML,{" [11 = 0—2]
in terms of the relative rate at which they react (as opposed
to defining reactivity as the number of different products
MLn+ can form). The relative rates calculated for 14*, 1411*,
and MLZ+ in 2:1 mixtures of alkane to metal compound are
presented in Tables 30 through 32. These relative rates, K,
were estimated from the relative abundances of the reactant
product ion, I (Product), in the mass spectra of the reaction

mixtures using equation (20):

138

Table 31. Relative Reaction Rates

in 2:1 Mixtures of Alkanes to Co(CO)3N0

00(C0); + Methane

Ethane

Propane

Butane

Pentane

Hexane

Heptane

Octane

Nonane

Decane

Z-Methyl propane
2,2-Dimethy1propane
Z-Methylbutane
2,2-Dimethy1butane

2 ,3-Dimethyl butane
2,2,3,3-Tetramethylbutane
3-Methy1pentane

2 ,4-Di methyl pentane

2 ,2 ,4-Tri methyl pentane
2,3,4-Trimethy1pentane

n-l

 

1.0
1.0
1.0
0.27
0.78

’ 0.70

0.93
0.21
0.92
1.0

0.19
0.69
0.71

1.0

0.91
1.0
1.0
0.55
1.0
1.0

n-Z
'0

 

0.65
1.0
0.71
0.12
1.0
0.76
0.71
0.74
1.0
1.0

0.79

1.0

0.45
0.17

139

Table 32. Relative Reaction Rates

in 2:1 Mixtures of Alkanes to Ni(CO)4

"-0 21. is.
111(00): + Methane _6' 0 0
Ethane 0 0 0
Prepane 1.0 0.24 0.14
Butane 1.0 0.37 0.05
Pentane 0.32 1.0 0.81
Hexane ' 1.0 0.87 0.77
Heptane 1 .0 0.95 0.58
Octane 1.0 0.19 0.08
Nonane 0.91 1.0 0.53
Decane 1.0 0.82‘ 0.89
Z-Methylpropane 0.15 0.19 1.0
2,2-Dimethy1propane 0.93 1.0 0.88
Z-Methylbutane 0.76 1.0 0.82
2 ,2-Dimethy1 butane 1 .0 0.94 0 . 44
2 ,3-01methy1 butane 0. 75 1.0 0.48
2,2,3,3-Tetramethy1butane 0.49 1.0 0
3-Methy1pentane 0.42 1.0 0.50
2,4-0imethy1pentane 0,63 0.77 1.0
2,2,4-Trimethy1pentane 0.62 1.0 0
2,3,4-Trimethy1pentane 0.99 1.0 0.05

.140

Table 33. Relative Reaction Rates

in 2:1 Mixtures of Alkanes to Ni(PP3)‘

+
Ni(PF3)n + Methane

Ethane

Propane

Butane

. Pentane

Hexane

Heptane

Octane

Nonane

Decane

Z-Methylpropane
2,2-Dimethy1propane
Z-Methylbutane
2,2-Dimethylbutane
2,3-0imethy1butane
2,2,3,3-Tetramethylbutane
3-Methy1pentane
2,4-Dimethy1pentane
2,2,4-Trimethy1pentane
2,3,4-Trimethy1pentane

n=0

0

0
0.10
0.38
1.0
0.44
0.40
0.61
1.0
0.45
0.24
0.72
0.61
1.0
0.62
0.66
0.54
0.47
0.32
0.29

n-1

 

0

0
1.0
1.0
0.70
1.0
1.0
1.0
0.88
0.51
1.0
1.0
1.0

0.72
1.0
1.0
1.0
1.0
1.0

n-2

0.49
0.43

0.33

0.52
0.85
1.0

0.90
0.97
0.93

1.0

0.93
0.85
0.12
0.91

141

ill .
Z: 1(Product 1)
i=1
K °‘ ... <20)
loin“) + 21(Pmduct 1)
1-1

 

The fastest reactions are arbitrarily assigned a magnitude of
1.0 in Tables 30 through 32. These results suggest the
following trends in reactivity for cobalt- and.

nickel-containing ions for ML“+ reacting with an alkane:

mh++Alkane e'rroduct1

 

k
m Product m
+ .. + +
m1: ) - K(&)(D )> K(CO((D)2 ) (57)
K(Ni+) ’-‘-' K(NiCO+)> K(Ni(CD)2+) (58)
xmirrf) > Km?) > xmicprpzb <59)
tdmmelii 2: k1

From reactions (58) and (59), it follows that
K(NiPF3+) > K(Ni+) 9 K(Nim+)

Overall, the M'" and 1400* ions react at approximately the same
rate with alkanes. However, when a PF3 ligand is present on
Ni+, NiPF3+ is found to react faster than Ni+. In all cases,

ML+2 reacts most slowly with alkanes.

142
SOLUTION CHEMISTRY PARALLELS

Compared to other chemically active molecules containing
multiple bonds or lone pairs of electrons, alkanes, which
consist of stable C-C and C-Ho bonds, have been found to be
relatively inert in condensed phase chemistry. However,
since alkanes are of fundamental importance in the chemical
industry, recently there has been a drive towards devising
systems capable of selectively activating the C-H or C—C
bonds of alkanes.‘ Many examples of the activation-of C-H
bonds in polar organic molecules by metal complexes in
solution can be found.104 Apparently, polar substituents
make facile activation of the C--H bonds in these molecules
possible.10‘*:115 Alkanes obviously contain no polar
substituents. However, the oxidative addition of a metal to

the C-H bonds of alkanes has been reported recently in’ the

fliterature. For example, Crabtree et al. reported the

stoichiometric and selective dehydrogenation of cyclic
alkanes. Apparently, the two acetone ligands in an iridium
complex (which also contained two hydride and two
triphenylphosphine ligands) are so weakly bound to Ir, they
can easily be displaced bydhydrocarbon molecule. Crabtree
suggests that the initial step in this reaction involves the
insertion of Ir into a C-H bond of the alkane.37 This
reaction is carried out in 1,2-dichloroethane, a

noncoordinating solvent. In addition, Bergman et al. have

143

reported the discovery of another iridium complex which adds
oxidatively to the C-H bonds in saturated hydrocarbons
leading to the formation of stable hydridoalkyl-metal
complexes. This reaction proceeds at room temperature in a
homogeneous solution. This process is illustrated for

cyclohexane in reaction (60):86

I
Meap—l‘ . . .H UV > Maap—‘r. . .H + H2

" 0

Note that this iridium complex only involves four ligands.
Bergman suggests that upon photolysis the metal complex loses
both of its hydrides as dihydrogen. As a result, the iridium
complex is coordinatively unsaturated and capable of
inserting into a C-H bond of cyclohexane. Furthermore,

Bergman has recently reported that a similar rhenium complex

. 144
activates the C-H bonds in alkanes.116 As these examples may
suggest, in solution the C-H bonds of alkanes are found to be
preferentially activated over the C-C bonds by metal
complexes. Apparently, the activation of C-C bonds is
unfavored due to the steric hindrance that would result from
the interaction of a metal complex and a C-C bond.115 Recall
that similar reasoning was used to explain the preferential
insertion of NCO+ into C-H bonds in alkanes in the gas phase.

Furthermore, Poliakoff and Turner117 suggest that upon
photolysis of Fe(CO)5 in a low-temperature methane matrix,
Fe(CO)5CH4 is formed. Their data indicate that the methane
ligand in this metal complex is intact. These results are
similar to those we presented earlier for the reactions of
ML2+ and alkanes in which we suggested that intact alkanes
were capable of displacing L.

Furthermore, in a study of the correlation between the
activity of platinum crystal surfaces and the catalytic
- reactions of hydrocarbons, Somorjai gt_:__a_l_. reported 113 .that
highly coordinated atoms are the key components of catalysis
with high turnover rates. Interestingly,their results suggest
that metal-ligand catalysts will break and form H-H and C-H
bonds quite readily. However, unlike a highly uncoordinated
atom in a metallic cluster, metal ligand catalysts lack the
ability to break c-c bonds.118

Thus, preferential attack of C-H bonds in alkanes by
metal-ligand complexes is observed on platinum surfaces. In

addition, Somorjai suggests that highly uncoordinated metal

145

atoms in metal clusters are capable of breaking C-C bonds.
These results parallel those observed in the gas phase
reactions of bare metal ions and alkanes where C-C bonds are
preferentially attacked.

Other examples of the activation of C-C bonds by

uncoordinated metal atoms can be found in the literature. For

 

example, zirconium atoms have been reported to add
oxidatively to the C-C (and C-H) bonds of methane, ethane,
propene, and propane at cryogenic temperatures.119
Furthermore, cleavage of the C-C bonds in pentane by nickel
clusters at low temperatures has been reported by Davis and
Klabunde.102 Thus parallels between the observed reactions
for M" in the gas phase and the reactions observed for metal

atoms in condensed phases can also be made.
CONCLUSION

In conclusion, the ligand effects presented in this chapter
have yielded some interesting insights into the gas phase
chemistry of MLn+ (n 8 0,1,2) and alkanes. Evidence
indicating that intact alkanes are capable of displacing CO
or PF3 ligands was presented and rationalized on the basis of
ion-induced dipole and energetic calculations. Also, the
preferential insertion of 1111* into C-H bonds resulting in H2

elimination was discussed in terms of steric effects and

146

contrasted to the gas phase chemistry observed for 18*.
Finally condensed phase reactions, paralleling the gas phase
reactions, were presented.

Note that, up to this point in time, work in the area of
organometallic gas phase chemistry has centered on the
characterization of the reactions occurring between various
metal-containing ions and organic molecules. The purpose of
this study has been to go beyond characterizing reactions by
gaining new insights into the nature of the interaction
occurring between metal ions and organic molecules in the gas
phase. However, caution must be used in interpreting the
results of the energetic calculations due to the numerous
assumptions (already cited in the text) associated with this
analysis. Furthermore, the author notes that in this
analysis, no consideration of the dynamics associated with

these ion/molecule reactions has been made.

CHAPTER 8

CONCLUSION

Chemical Ionization (CI) mass spectrometry is a
technique in which sample molecules undergo ion/molecule
reactions with reagent ions and thus become ionized. Since
the energy transferred to the sample molecule in an
ion/molecule reaction is lese than that transferred upon
ionization via conventional 70eV electron impact, CI is
considered a "soft" ionization technique. 'As a result, the
analyte molecule is found to undergo less fragmentation.
Usually the molecular weight of the analyte can be readily
determined from C1 mass spectra. Furthermore, unlike the
70eV non-selective electrons in electron impact ionization

reagent ions which react specifically with analyte

' molecules can be chosen for study. Thus CI reactions

frequently yield structural information concerning the'
analyte molecule.

Previously in the area of gas phase organometallic
chemistry, the focus has been on the characterization of
the reactions of metal and metal—containing ions with
organic molecules. A mechanism involving a metal ion
insertion/ B-hydrogen shift/competitive ligand loss

sequence has been used to explain the formation of the

147

148

majority of the product ions observed in these reactions.
If each step in this mechanism were fully understood,
products and branching ratios could be predicted a 23121;;
making metal ions ideal CI reagents. The metal insertion
step appears to be most important in determining final
product distributions.“3 However, it is the least un-
derstood step of this mechanism. Thus, although previous
studies have provided useful insights into organometallic
gas phase chemistry by revealing that metal ions react
specifically with organic molecules, there is still much
that is not understood about the nature of the interactions
between metal ions and organic molecules. The purpose of
this study was to gain a better understanding of the nature
of this interaction.

The first part of the discussion, Chapter 5, provided a
characterization of the gas phase reactions of cobalt and
cobalt-containing ions with secondary and tertiary amines.
This chapter illustrated the utility of Co(CO)3NO as a CI
reagent. It was shown that Co+ produced from 70eV electron
impact reacted specifically with amines. The mechanism
consistent with the formation of product ions involved the
oxidative addition of 00+ to C-C and C-H bonds followed by
a B-hydrogen atom transfer to the metal and then the
elimination of a small neutral molecule such as H2 or CZH6.
Knowledge of this mechanism allowed one to interpret the

mass spectra of amines assigning structures to all of the

149

product peaks. Thus, specific structural information could
be formulated since the observed chemistry depended upon
the structures of the reactant amines. Furthermore, a
characteristic series of ligand substitution reactions was
observed yielding molecular weight information.

Metal and metal-containing ions were also found to
react specifically with alkanes. Analogous to the amine
chemistry, the observed product ions in alkane mixtures
were indicative of initial metal ion insertion into C-0 and
C-H bonds. The data suggested that metal ions prefer-
entially inserted into the C-C bonds of alkanes rather
than into C-H bonds.

The purpose of Chapter 6 was to gain a better
understanding of the insertion process as it occurs in
alkanes. (Note that the functional group, X, in polar
molecules, RX, may have a strong effect in "directing" the
site of attack of the metal ion to certain skeletal bonds
which may not be attacked in the corresponding alkane
(RH).19 In the gas phase chemistry of M+(M-Fe, Co, Ni) with
alkanes, the metal ion can insert into a number of C-0
bonds to form intermediates of the type R-M+-R'. As
indicated by the product distributions, insertion into
certain C-C bonds is preferred over others. To explain
this observation, a correlation between the preference by
14* for insertion into a particular skeletal bond and the

ionization potentials of the alkyl radicals formed upon

150

insertion was noted. This correlation suggests an inverse
relationship between IP(CnH2n+1) and D(M+-CnH2n+1),i.e. ,
IP decreases and bond energy increases as 1': increases. A
similar relationship between IP and bond strengths of alkyl
radicals to other cations such as Hg+ was presented
supporting this suggestion. Thus, this correlation may
suggest that the preferred intermediates resulting from M‘"
insertion into the C-0 bonds of alkanes involve those in
which the (W—R) bonds which are formed are the strongest.

Finally in Chapter 7, ligand effects observed in the

reactions of metal-containing ions and alkanes were
discussed. With the addition of one or two ligands, three
types of behavior were observed:

1. Ligand acted as a "spectator" and ML'" reacted with

an alkane just as 15" did.

2. As the number of ligands on the metal exceeded
one, ligand substitution products were favored.

3. The amount of H2 elimination increased for NCO": as
the size of the alkane increased implying that
insertion into C-H bonds is favored by MCO+.

Differences in the chemistry of W and MLn'" (nsl,2) with
alkanes was explained on the basis of steric effects.
Furthermore, evidence indicating that intact alkanes are
capable of displacing CO or PF3 ligands was presented and
rationalized on the basis of ion/induced dipole energetic
calculations. Parallels to condensed phase chemistry

were made .

List of References

11.

12.

13.
14.

15.

16.

17.

18.

19.

1.18 t of References

J. Allison, D.P. Ridge, J. Am. Chem. Soc., 12;, 4998 (1979).
P.B. Armentrout, J.L. Beauchamp, J. Chem. Phys., 15, 2819 (1981).
RIW. Jones, R.H. Staley, J. Phys. Chem., 22, 1387 (1982).

C.J. Cassady, B.S. Freiser, J. Am. Chem. Soc., 121, 1566 (1985).
A. Tsarbopoulos, J. Allison, Organometallics, 2, 86 (1984).

J. Allison, D.P. Ridge, J. Am. Chem. Soc., 22, 7445 (1976).

J. Allison, D.P. Ridge, J. Organometal. Chem., 22, (811 (1975).

P.B. Armentrout, J.L. Beauchamp, J. Am. Chem. Soc., 222, 1736
(1980).

B.D. Radecki, Master's Thesis, Nichigan State University, 1982.
N.F. Lappert, P.P. Power, A.R. Sanger, R.C. Srivastava, ”Natal
and Hetelloid Amides;” Ellis Horwood Limited; West Sussex,
England, 1980.

J.L. Beauchamp, Ann. Rev. Phys. Chem., 22, 527 (1971)

T.A. Lehman, N.N. Bursey, ”Ion Cyclotron Resonance Spectroscopy,"
J. Wiley and Sons, New York (1976).

J.D. Baldeschwieler, Science, 122, 263 (1968).

J.D. Baldeschwieler, S. S. Woodgate, Acc. Chem. Res., 5, 114
(1971).

C.J. Drewery, 0.0. Goode, K.R. Jennings, ”Ion Cyclotron Mass
Spectrometry, ”NTP International Review of Science-Physical
Chemistry," Series One, Volume Five, Haas Spectrometry,

A. Nacoll, Ed., University Park Press, Baltimore, 183 (1972).
J.N.S. Henis, ”Ion Cyclotron Resonance Spectrometry,"
Ion-Molecule Reactions, Vol. 2, J. L. Franklin, Ed.,

Plenum Press, New York, 395 (1972).

0.0. Coode, A.J. Ferret-Carrels, K.R. Jennings, Int. J. Mass
Spec. Ion Phys., 2, 229 (1970).

G.A. Junk, H.J. Svec, Z. Naturforsch,22, 1 (1968).

R.E. Winters, R.W. Riser, J. Inorg. Chem., 2, 699 (1964).

151

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.
30.

31.

32.

33.

34.

35.

36.

37.

38.

152

E. Schumacher, R. Taubenest, Helv. Chim. Acts, 21, 1525 (1964).
J. Huller, ”Ion Molecule Reactions of Organometallic
Complexes,” Advances in Mass Spectrometry, Vol. 6, A. R. West,
Ed., Appl. Sci. Publishers, Ltd. England, 823 (1974).

R.C. Dunbar, J.F. Ennever, J.P. Fackler, Inorg. Chem., 12, 2734
(1973).

H.S. Foster, J.L. Beauchamp, J. Am. Chem. Soc., 22, 4924
(1971).

N.S. Foster and J.L. Beauchamp, J. Amer. Chem., Soc., 21, 4808
(1975).

N.S. Foster and J.L. Beauchamp, J. Amer. Chem. Soc., 21, 4814
(1975).

R.R. Corderman and J.L. Beauchamp, Inorg. Chem., 12, 665
1976).

R.R. Corderman and J.L. Beauchamp, J; Amer. Chem. Soc., 22,
5700 (1976).

R.D. Bach,A.T. Weibel, J. Patane and L. Kevan, J. Amer. Chem.
Soc., 22, 6237 (1976).

R. V. Hodges and J. L. Beauchamp, Anal. Chem., 22, 825 (1976).
J. Allison, D.P. Ridge, J. Am. Chem. Soc., 22, 35 (1977).

C.H. Waddle, J. Allison, D.P. Ridge, J. Amer. Chem. Soc., 22, ,
105 (1977).

c. Dietz, 0.5. Chatelier, D.P. Ridge, J. Am. Chem. Soc., 199,'
4905 (1978).

J. Allison and D.P. Ridge, J. Am. Chem. Soc., 122, 163 (1978).

J. Allison, R. Kinser, T.G. Dietz, H. deAngelis, D.P. Ridge, J.
Amer. Chem. Soc., 122, 2706 (1978).

R.C. Burnier, T.J. Carlin, W.D. Reents,Jr., R.B. Cody, R.K.
Lengel, B.S. Freiser, J. Am. Chem. Soc., 221, 7127 (1979).

J. Allison, R.B. Freas, D.P. Ridge, J. Amer. Chem. Soc., 121,
1332 (1979).

J.L. Beauchamp, A.E. Stevens, R.R. Corderman, Pure & Appl.
Chem., 21, 967 (1979).

R.C. Burnier, G.D. Byrd, B.S. Freiser, Anal. Chem., 22, 1641
(1980).

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

51.

52.

53.

54.

55.

56.

57.

153

P.B. Armentrout, J.L. Beauchamp, Chem. Phys., 29, 37 (1980).

R.U. Jones and R.B. Staley, J. Am. Chem. Soc., 122, 3794
(1980).

J.S. Uppal, R.B. Staley, J. Am. Chem.Soc., 122, 4144 (1980).

L.F. Halle, P.B. Armentrout, J.L. Beauchamp, J. Am. Chem. Soc.,
122, 962 (1981).

P. B. Armentrout, J. L. Beauchamp, J. Am. Chem. Soc., 103,784
(1981)

.A.B. Stevens, J.L. Beauchamp, J. Am. Chem. Soc.,122, 190 (1981).

P.B. Armentrout, L.F. Halls, J.L. Beauchamp, J. Am. Chem. Soc.,
122, 6624 (1981).

P. B. Armentrout, J. L. Beauchamp,. J. Am. Chem. Soc.,103 6628
(1981) .

J.S. Dppsl, D.B. Johnson, R.B. Staley, J. Amy Chem. Soc., 122,
508 (1981).

C. Burnier, G.D. Byrd, B.S. Freiser, J. Am. Chem. Soc., 192.4360
(1981).

0.0. Byrd, R.C. Burnier, B.S. Preiser, J. Am. Chem. Soc., 122,
3565 (1982).

G.D. Byrd, B.S. Freiser, J. Am. Chem. Soc., 125, 5944 (1982). ‘

L.F. Halle, R. Houriet, H.M. Kappes, R.fl. Staley, J.L.
Beauchamp, J. Am. Chem. Soc. 122, 6293 (1982).

L.P. Belle, P.B. Armentrout, J.L. Beauchamp, Organometallics,
1, 963 (1982).

R. Houriet, L.F. Belle, J.L. Beauchamp, Orgsnometallics, 2,
1818 (1983).

D.B. Jacobson, B.S. Freiser, J. Am. Chem. Soc., 192, 7492
(1983).

D.B. Jacobson, B.S. Freiser, J. Am. Chem. Soc., 122, 7484

T.J. Carlin, L. Sallans, C.J. Csssady, D.B. Jacobson, B.S.
Freiser, J. Am. Chem. Soc., 122, 6320 (1983).

D.B. Jacobson, B.S. Freiser, J. Am. Chem. Soc., 105, 5197
(1983).

58.

59.

60.
61.
62.
63.
64.
65.

66.

67.

68.
69.
70.
71.
72.

73.

74.

75.

76.

77.

78.

1511

D.B. Jacobson, B.S. Preiser, J. Am. Chem. Soc., 19;, 736 (1983).

H. Lombarski, J. Allison, Int. J. Mass Spectron. Ion Phys., 92,
281 (1983).

3.x. Huang, J. Allison, Organometallics, 2, 883 (1983).

C.J. Cassady, B.S. Preiser, J. Am. Chem. Soc., 192, 6176 (1984).
D.B. Jacobson, B.S. Preiser, J. Am. Chem. Soc., 192, 5351 (1984).
D.B. Jocobson, B.S. Preiser, J. Am. Chem. Soc., 192, 3891 (1984).
D.B. Jacobson, B.S. Freiser, J. Am. Chem. Soc., 192, 3900 (1984).
D.B. Jacobson, B.S. Freiser, Organometallica, 2, 13 (1984).

I.C. Jackson, D.B. Jacobson, B.S. Preiser, J. Am. Chem. Soc.,
192, 1252 (1984).

D.B. Jacobson, 6.0. Byrd, B.S. Preiser, Inorg. Chem., 22,553
(1984).

B.S. Larsen, D.P. Ridge, J. Am. Chem Soc., 192, 1912 (1984).
R.B. Press, D.P. Ridge, J. Am. Chem. Soc., 192, 825 (1984).
J. Wronka, D.P. Ridge, J. Am. Chem. Soc., 192, 67 (1984).
B.D. Radecki, J. Allison, J. Am. Chem. Soc., 192, 946 (1984).

C.J. Cassady, B.S. Freiser, S.W} McElvany, J. Allison, J. Am.
Chem. Soc., 192, 125 (1984).

H.B. Wise, D.B. Jacobson, B.S. Freiser, J. Am. Chem. Soc., 192
1590 (1985).

C.J. Cassady, B.S. Freiser, J. Am. Chem. Soc., 192, 1566
(1985).

D.B. Jacobson, B.S. Preiser, J. Am. Chem. Soc., 192, 1581
(1985).

J. Balpern, in "Organotransition Chemistry," Y. Ishii and
H. Tsutsui, Plenum Press, New York, p. 109, 1975.

P.B. Armentrout, L.F. Halls, J.L. Beauchamp, J. Am. Chem.
Soc., 192, 190 (1981).

A.E. Stevens, J.L. Beauchamp, J. Am. Chem. Soc., 192, 190
(1981).

155

79. R.R. Corderman, J.L. Beauchamp, J. Am. Chem. Soc., 22, 3998
(1976).

80. J.Mu11er, W. Coll, Chem. Ber., 192, 1129 (1973).

81. J. Muller, W. Bolzinger, W. Kalbfus, J. Organomet.
Chem., 22, 213 (1975).

82. Bond strengths calculated by using the data from:
a) J.L. Franklin, J.G. Dillard, B.M. Rosenstock, J.I. Herron,
R. Draxyl, Nat. Stand. Ref. Data Ser., Nat. Bur. Stand., 22
(1969). b) H.M. Rosenstock, K. Draxl, B.W. Steiner, J.T. Herron,
J. Phys. Chem. Ref. Data, 2_(1977). c) D.P. McMillen, D.N.
Golden, Ann. Rev. Phys. Chem. 22, 493 (1982). d) J.D. Cox,
G. Pilcher, ”Thermochemistry of organic and Organometsllic
Compounds,” Academic Press, New York (1970).

83. D.A. Aue, N.I. Bowers, "Gas Phase Ion Chemistry,” v. 2, N.T.
Bowers, Bd.,.Academic Press, New York (1979).

84. D. Wobschall, ”Ion Cyclotron Resonance Spectrometer," Rev. Sci.
Instr., 22, 466 (1965).

85. A. Warnick, L.R. Anders, T.E. Sharp, Rev.Sci.Instr.,§2, 929
(1974).

86. A.H. Janowicz, R.C. Bergman, J. Am. Chem. Soc., 192, 3521 (1982).

87. R.H. Crabtree, J. Quirk, J. Nihelcic, J. Am. Chem. Soc., 192,
107 (1982).

88. See, for example: a) P.B. Armentrout, J.L. Beauchamp, J. Am.
Chem. Soc.,192, 1736 (1980). b) J.A. Connor in "Topics in
Current Chemistry. No. 71. Inorganic]Chemistry/Netlearbonyl.
Chemistry," F.L. Boschke, Bd., Springer-Verlag, New York (1977).

89. J.B. fiendrickson, D.J. Cram, 0.8. Hammond, ”Organic Chemistry," '
McCraw-Hill Book Company, New York (1970).

90. S.J. Babinec, J. Allison, J. Am. Chem. Soc., 192, 7718 (1984).
91. P.B. Armentroout, N. Aristov, R. Georgiadis, L. Sunderlin, "33rd
Annual Conference on Mass Spectrometry and Allied Topics,”

San Diego, 1985, paper ROAS.
92. P.A. Houle, J.L. Beauchamp, J. Am. Chem.Soc., 191, 4067 (1979).
93. J.M. Williams, W.N.Hamill, J. Chem. Phys., 22, 4467 (1968).

94. F.P. Lossing, A. Macoll, Can. J. Chem., 22, 990 (1976).

'156

95. A. Tsarbopouloa, B.D. Radecki, J. Allison, ”32nd Annual
Conference on Mass Spectrometry and Allied topics," San Antonio,
1984, paper TOD 10.

96. I. Fleming, ”Frontier Orbitals and Organic Chemical Reactions,"
John Wiley and Sons, New York, 1976.

97. H.J. Wax, J.H. Stryker, J.H..Buchanan, C.A. Kovac, R.C.
Bergman, J. Am. Chem. Soc., 192, 1121 (1984).

98. G.W. Parshall, Acc. Chem. Res., 2, 113 (1975).

99. T.H. Tulip, J.A. Ibers, J. Am. Chem. Soc., 191, 4201 (1979).

100. H.P. Schaefer, Acc. Chem. Res., 19, 287 (1977).

101. L. Abis, A. San, J. Halpern, J. Am. Chem. Soc., 199, 2915 (1978).
102. S.C. Davis, R.J. Rlabunde, J. Am. Chem. Soc., 199, 5973 (1978).
103. P.L. Watson, D.C. Res, J. Am. Chem. Soc., 192, 6471 (1982)

104. F.A. Cotton, G. Wilkinson, ”Advanced Inorganic Chemistry,”
John-Wiley & Sons, New York, (1980).

105. R.J. Miller, J.A. Savcik, J. Am. Chem. Soc., 191, 7206 (1979).

106. J.O. Hirschfelder, C.F. Curtias, R.B. Bird, ”Molecular Theory of
Gases and Liquids," John-Wiley & Sons, New York (1964).

107. S.W. Benson,”The Foundation of Chemical Kinetics,” NcGraw-Hill
Book Company,lnc., New York (1960).

108. L.P. Halls, W.B. Crowe, P.B. Armentrout, J.L. Beauchamp,
Organometsllics, 2, 1694 (1984).

109. R.W. Riser, N.S. Rrassol, J. Am. Chem. Soc., 22, (1967).

110. J.B. Huheey, ”Inorganic Chemistry," Harper and Row, Publishers,
New York (1978).

111. F. Daniels, R.A. Alberty, "Physical Chemistry," John-Wiley &
Sons,Inc., New York (1975).

112. R.C. Denbigh, Trans. Faraday Soc., 22, 936 (1940).

113. S. Fliszar, ”Charge Distributions and Chemical Effects,"
Springer-Verlag, New York (1983).

114. C.A. Tolman, Chem. Rev.,12, 313 (1977).

115. A.B. Shilov, ”Activation of Saturated Hydrocarbons by Transition
Metal Complexes,” D. Reidel Publishing Company, Boston (1984).

157

116. R.C. Bergman, P.B. Seidler, T.T. Wenzel, J. Am. Chem. Soc., 1_l,
4358 (1985).

117. H. Poliakoff, J.J. Turner, J.C.S. Dalton, 2277 (1974).
118. C.A. Somorjai, D.W. Blakely, Nature, 222,580 (1975).

119. R.J. Remick, T.A. Asunta, P.S. Skell, J. Am. Chem. Soc., 1_1,
1320 (1979).