{WWI-I |I'»)';;II;'II-‘. ','.""'i'hl'.-AFI .‘n-I .. ., ,Iwu-u "14,. r- . 1. . I I . I «I. I , ‘ ‘ , J . . . . -
‘:‘..V': .......,.y _ ..
..I II... , w “II-Inn.“ HIII Im -.,;.-II- . I ~ _."II .[ " '- I H . .1“ “II III." ‘nm AI. -. .

. -'”I“II’I‘I’WWWII‘MM’IWM’WI“IIW"‘

PART I

‘ STEIIEOICIIEMISTRY ARR IIRLERRIAR i

REARRAI‘IGEMENTS OF :
BIS‘I} -DIIIET€}PIAT€II

‘ DIGXOMOLYBDEIIIIMM) RRIRREEXER ‘ , .

PART II

f A TEST RE THE ROLE RE IRRRIEIIE ‘

EXTRAEOOIIDINATION 0F SILIGGN
. IN SILYIATIMN REACTIONS

RARE III 7 '

‘ SYNTHESIS AII‘R STEREGCIIEMIS‘I’R‘I‘ 0F ’

TRIMETHYLSILYL 6 KETOAIIIII‘IES

.‘Ihesis IRE the. Degree of P‘h1. DI. E
MICHIGAN STATE UNIVERSITY '
WILLIAM II. ELEMENTS
1974

 

 

This is to certify that the

thesis entitled
PART I. STEREOCHEMISTRY AND MOLECULAR REARRANGEMENTS OF BIS-
(B -D I KETONATO) D IOXOMOLYBDENUM (V I) COMPLEXES
PART II. A TEST OF THE ROLE OF INCIPIENT EXTRACOORDINATION
0F SILICON IN SILYLATION REACTIONS
PART III. SYNTHESIS AND STEREOCHEMISTRY OF TRIMETHYLSILYL
B-KETOAMINES presented by

WILLIAM R . CLEMENTS

has been accepted towards fulfillment
of the requirements for

Ph.D. Chemistry

degree in

RR...“

Major professor

 

 

 

Dateflw. 5/77
A/ J 6‘

0-7 639

 
 
     

  
  

. mg‘, Rs
HDAB & sous I
800K IIIIIIIIRI IND._

“LEE! :va'r;

“II On In 1-

ABSTRACT
PART I

STEREOCHEMISTRY AND MOLECULAR REARRANGEMENTS 0F
BIS(B-DIKETONATO)DIOXOMOLYBDENUM(VI) COMPLEXES

by

William R. Clements

A series of octahedral dioxomolybdenum(VI) B-diketonate complexes
of the type M002(dik)2 have been prepared by reaction of the free

diketone and M0042— in acidic aqueous solution. Proton magnetic

resonance studies show that the M002 moiety adopts exclusively a cis

donfiguration as illustrated by isomers I and II-IV fer the symmetric

(a-c) and asymmetric (d-e) diketonate derivatives, respectively.

 

CH3; dik = acetylacetonate

tert-C4H9; dik = dipivaloylmethanate
- CD(CH3)2; dik = diisobutylmethanate-d3

O
50
l

William Clements

 

 

 

 

 

 

 

 

O
0
II III IV
d. R1 = CH3, R3 = tert-C4H9; dik = pivaloylacetonate
R1 = tert-C4H9, R2 = CH(CH3)2; dik = pivaloylisobutylmethanate
R1 = CH3, R2 = CH(CH3)2; dik = acetylisobutylmethanate
g. R1 = C6H5; R2 = CH(CH3)2; dik = benzoylisobutyrylmethanate

The complexes are stereochemically non-rigid and undergo rapid
intramolecular rearrangements which lead to interchange of non-
equivalent R group environments in I and isomerization and inversion of
II, III, and IV. First order rate constants for the rearrangement
of Ib in dichloromethane were determined by fu11 pmr line shape
analysis. The values of the Arrhenius activation energy and frequency
factor are 17.0 :_l.0 kcal/mole and exp (12.56 :_0.57), respectively.
The kinetics data indicate that the size of the central metal ion is
the principal factor'influencing'the lability of the coordinate metal-

oxygen bonds in this and related nonrigid do and d10 metal complexes.

William Clements

Temperature dependent pmr studies of the benzoylisobutyrylmethanate
and deuterated diisobutyrylmethanate derivatives indicate that while
a twist mechanism is likely for the rearrangement process it does not

occur about one specific axis in the molecules.

PART II

A TEST OF THE ROLE OF INCIPIENT EXTRACOORDINATION
OF SILICON IN SILYLATION REACTIONS

The silylation of methanol by the segcis(V) and segtrans(VI)

isomers of trimethylacetylacetone has been investigated in order to

R Si-~-O
O/ 3 n _
I - \ / " \
CH H CH C='0
3 3 &
“3
V VI

test Klebe's hypothesis (Accts. Chem. Res., é, 299(1970)) that incipient

 

extracoordination in the ground state of certain triorganosilicon com-
pounds may facilitate their ability to undergo proton transfer reactions
via formation of six-coordinate transition states that are more stable
than the five-coordinate transition states that can be achieved by
organosilicon compounds which are not incipiently extracoordinated.
Compounds V and VI are ideally suited for a test of this hypothesis.

As the former isomer may achieve six-coordination in a silylation

reaction whereas the latter cannot (eq 2).

William Clements

 

 

 

 

 

R Si 0 e l
3, x\ R C q )
o /C-CH3
\Czc + R'OH -—> ——)R SiOR' + CH COCH COCH
/ \ R R 3 3 2 3
CH3 H
ROH
_ .J
v
CH
CH \35 -
\3 ,c-o
R Si ’C:C\ (eq 2)
3 \o /H R H
\C=C + R'OH —; -+R SiOR' + CH COCH COCH
/ R R 3 3 2 3
CH C—CH
3 H 3
O ROH
VI

The reaction of methanol with a mixture of V and VI has been monitored

by pmr spectroscopy and the ratio of [gig/trangj followed. The segcis to
segtrans ratio remains constant when the silylation reaction is carried
out in methylene chloride and chlorobenzene. Thus, both isomers react
with methanol at the same rate and the ability of the segcis form to
expand its coordination does not facilitate the proton transfer reaction
It is concluded that expansion of the coordination sphere of silicon

from four to five or from four to six is not the rate determining step

in the silylation reaction.

William Clements

PART III

SYNTHESIS AND STEREOCHEMISTRY OF TRIMETHYLSILYL 8-KETOAMINES

The first examples of silyl B-ketoamines have been prepared by
reaction of the potassium salts of three different anilinopentenones
and trimethylchlorosilane in hexane. Pmr and ir spectroscopic studies
indicate that the compounds are O-silylated and exist as four dia-

stereomers (VII-X) due to seqcis-seqtrans isomerization about the C=C

 

bond and syn-anti isomerization about the C=NR bond (R = C6 5’

m-C H Cl and p-C6H4F:

 

6 4
(C113)381\O CH (CH3)381 0 CH
\. /’ 3 \\ 1’. 3
C C
N H
/C\ 11‘ /H
R CH3 CH3
seqcis-anti (VII) segcis-syn (VIII)
(CH3)381\0 CH (CH3)381\0 CH
\C/ 3 \ / 3
,2 :4 R
\ / \ /
H CHC=N\R H C/c =N
3 H3

seqtrans-anti (IX) segtrans-szn (X)

William Clements

Rapid syn-anti isomerization is observed for both the segcis and
segtrans isomeric pairs. Based on tentative chemical shift assignments,
the interchange of syn-anti forms appears to be appreciably faster for
the segcis pair due to steric effects and, possibly, a l, S silyl group

migration to give a N-silylated intermediate (XI):

 

 

r l
0§ /CH3
C\C/H
VII; R331\ u (2’ VIII
/N-C
\
R CH
3J
L.
XI

Syn-anti isomerization of the segtrans pair undoubtedly involves a

conventional tortional or lateral shift mechanism.

PART I

STEREOCHEMISTRY AND MOLECULAR REARRANGEMENTS OF
BIS(B-DIKETONATO)DIOXOMOLYBDENUM(VI) COMPLEXES

PART II

A TEST OF THE ROLE OF INCIPIENT EXTRACOORDINATION
OF SILICON IN SILYLATION REACTIONS

PART III

SYNTHESIS AND STEREOCHEMISTRY OF TRIMETHYLSILYL B-KETOAMINES

BY

William R. Clements

A THESIS

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

DOCTOR OF PHILOSOPHY
Department of Chemistry

1974

To

Karen

ACKNOWLEDGMENTS

Dr. Thomas J. Pinnavaia has been a continual source of guidance,
wisdom and encouragement, for which I am sincerely thankful. He is
an able and patient teacher to whom I shall always be deeply indebted.

I would like to express my appreciation to Dr. Gordon A. Nelson
for being my second reader, and to the Chemistry Department for its
financial support which helped make these studies possible.

It would be impossible to express my gratitude to my wife, Karen,
for her constant love, encouragement and understanding because mere
words cannot express feelings of such magnitude. Suffice it to say that
I am forever grateful. I wish to thank my two sons, Billy and Mike,
for their love; it has made me a rich man. I also wish to thank my

mother, Elsie Vance, for her continual love and encouragement.

iii

II.

III.

IV.

II.

III.

TABLE OF CONTENTS
PART I

INTRODUCTION .
EXPERIMENTAL .

A Reagents and Solvents. . . . . . . . . . . . . . .
B. General Techniques .
C Syntheses.

l. cis- -Dioxobis(acetylacetonato)molybdenum(VI). .

2. C15- Dioxobis(dipivaloylmethanato)molybdenum(VI).

3. Pivaloylacetone, Pivaloylisobutyrylmethane,
Acetylisobutyrylmethane, Benzoylisobutyrylmethane,
and Deuterated Diisobutyrylmethane Derivatives of
DioxomolybdenumCVI).

RESULTS AND DISCUSSION .

A. Preparation and Characterization of Dioxobis-
(dipivaloylmethanato)molybdenum(VI). . . . .
B. Kinetics Study of Dioxobis(dipivaloylmethanato)-
molybdenum(VI) Rearrangement . .
C. Preparation and Study of Dioxobis(d3- -diisobutyryl-
methanato)molybdenum(VI). . . . . .
D. Preparation and Study of Dioxobis(benzoyl-
isobutyrylmethanato)molybdenum(VI).

SUMMARY.

PART II

INTRODUCTION .
EXPERIMENTAL .

A. Reagents and Solvents.
B. General Techniques .

C. Synthesis of Trimethylsilylacetylacetonate .

RESULTS AND DISCUSSION .

iv

Page

0‘

oou \lO‘O

. 12

. 12
. 17
. 29
. 35

. 45

. 48
. 51
. 51
. 51
. 51

. 52

PART III

Page
I. INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . 58
II. EXPERIMENTAL . . . . . . . . . . . . . . . . . . . . . . . . 64

A. General Techniques . . . . . . . . . . . . 64
B. Synthesis of 4-Ani1ino-3-—pentene-—2-—one . . . . . . . . . 64
C Synthesis of 4— -m Chloroanilino- 3- -pentene- 2- -one

and 4-p:Fluoroanilino- 3— —pentene- -2- -one. . . . . . . . . . 64
D. Synthesis and Characterization of Trimethyl-

(4- anilino- 3- -pentene-2- -one)si1ane. . . . . . 65
E. Synthesis of Trimethyl(4-m- Chloroanilino- 3- -pentene-.

2- one)silane and Trimethylyl(4-pffluoroanilino-

3- -pentene- -2- one)silane . . . . . . . . . . . . . . . . . 66

III. RESULTS AND DISCUSSION . . . . . . . . . . . . . . . . . . . 67

A. Preparation and Properties of Silyl B-Ketoamines . . . . 67
B. Characterization of Silyl B-Ketoamines . . . . . . . . . 68

IV. CONCLUSIONS. . . . . . . . . . . . . . . . . . . . . . . . . 78
BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . . . . 82
APPENDIX . . . . . . . . . . . . . . . . . . . . ... . . . . 86

S A. Ligand Abbreviations. . . . . . . . . 86

B. Analysis of Multisite Dynamic NMR Spectra . . . . . . 87

Table

II.

III.

IV.

LIST OF TABLES

PART I

Properties of Synthesized Dioxobis(B-diketonato)
molybdenumCVI) Complexes . . . . . . .

Infrared Assignments for MoOz(dpm)2

tert- -C4H9 PMR Line Shape Parameters and Kinetic Data for
M002(dpm)2 in Methylene Chloride . . . . . . . . . . . .

Activation Parameters for MoOz(acac)2. . . . . . . . .

PART II

Segcis to Segtranso Ratio of (CH3)35i(acac) During Reaction

With CH30H at -5. 3° C .

vi

Page

10

16

26

30

56

LIST OF FIGURES
Figure _ Page

PART I

1. Infrared spectrum of M002(dpm)2 as a nujol mull between
CsI salt plates . . . . . . . . . . . . . . . . . . . . . . . 15

2. Proton nmr spectrum (60MHz) of M002(dpm)2
chlorobenzene at ambient machine temperature. . . . . . . . . 19

3. Temperature dependence of tert-C4H9 resonance lines of
M002(dpm)2 in methylene chloride. . . . . . . . . . . . . . . 21

 

4. (A) Temperature dependence of tert- -C4H9 resonance lines of
M00 2(dpm)2 in methylene chloride, 60MHz spectra.
(B) Calculated pmr spectra for the mean lifetimes
indicated . . . . . . . . . . . . . . . . . . . . . . . . 25
5. Ln k vs: l/T plot for MoOz(dpm)2 in methylene chloride. . . . 28

6. Temperature dependence of iSOpropyl methyl lines of
MoOz(d3- -dibm)2 in 1, 2- dichloroethane; 100MHz spectra. . . . . 34

7. Isopropyl methyl lines of MoOz(bibm)2 in 1, 2- dichloro-
ethane, 100MHz spectra. . . . . . . . . . . . . . 38

8. Possible isomers of MoOz(bibm)2 . . . . . . . . . . . . . . . 41

9. Results of trigonal twist mechanism operating in
MoOz(bibm)2 . . . . . . . . . . . . . . . . . . . . . . . . . 44

PART II

10. Possible mechanism for the reaction of CH3OH with the seqcis

isomer (A) and segtrans isomer (B) of (CH3)35i(acac). . . . . 54
PART III
11. Possible isomers of silyl B-ketoamines. . . . . . . . . . . . 63

12. Ambient pmr spectra of: (1), (CH3 ) Si(papo); (2),
(CH3)3 Si(m- -capo); (3), (CH3)3 Si(pffapo); 60MHz. . . . . . . . 70

vii

Figure Page

13. Ligand methyl region of temperature dependent pmr
spectra of (CH3)3Si(papo) in chlorobenzene . . . . . . . . . 72

14. Silyl methyl region and ligand =CH- region of
temperature dependent pmr spectra of (CH3 ) 3Si(papo)
in chlorobenzene . . . . . . . . . . . . . . 74

15. Infrared spectrum of neat (CH3)3 Si(papo) between
CsI salt plates. . . . . . . . . . . . . . . . 77

viii

PART I

I. INTRODUCTION

Metal complexes, especially those of d0 and d10 metals, frequently
do not possess a static structure in solution, but rather they rapidly
alter the coordination about the central atom. Determination of the
rates and mechanisms of intramolecular rearrangement reactions of metal
chelate complexes has been a long-standing problem of considerable
importance in coordination chemistry.1-3 There are two types of
rearrangement mechanisms, twist and bond rupture. The twist mechanism
involves a rotation of the bonds in the ground state configuration to
give a transition state or intermediate of different symmetry but of the
same coordination number. In the bond rupture mechanism a metal ligand
bond breaks giving an activated species of reduced coordination. It
is difficult in most cases to determine which of these mechanisms is
Operating.

The B-diketonate complexes of dioxomolybdenum(VI) represent an
important class of d0 metal compounds, about which there was virtually
no dynamic stereochemical information available. The earliest known
complex of this type was prepared by F. Gach in 1900,4 but the compound
was incorrectly formulated as Mo(acac)2, where acac = acetylacetonate
ion. Morgan and CastellS in 1928 demonstrated that the compound prepared
by Gach was the same as that prepared and correctly formulated as
MoOz(acac)2 by Rosenheim and Bertheim6 in 1903. It has only been in
the last seven years that the structure of these complexes has been

elucidated.7-11

In principle, two possible isomers might exist for these complexes
when the ligand is symmetric. The molybdenyl oxygens may be Opposite

or trans(I) to one another or they may be adjacent or cis(II) to each

other.
H\ C/Ra
R c/ \
‘ O /R I 0
/C\O 1-0/c\ /C\O—"—O
H-——C\ I M0 | /C-H Rb I Mo I
.r'O O 0
,c T \R Rb\c/O—r
R O R \ O
c 2’
H’ \C\
R
a
(I) (11)

As late as 1966, Larson and Moore12 proposed that the oxygens of the
M0022+ moiety in M002(acac)2 possessed a trans (linear) configuration.
This proposal was based on infrared data which showed that the moly-
bdenum to oxygen absorption was 70-100 cm"1 lower than that found for
many other molybdenyl compounds possessing a gi§_configuration. They
reasoned that this low energy shift in the absorption was a result of

a "trans effect" similar to that observed for the trans ReOz+ group.
This proposal by Larson and Moore stirred the interests of other workers,
including ourselves. Atovmyan and Sokolova performed an x—ray crystal
structure determination of dioxobis(8-hydroxyquinolinato)molybdenum(v1)9
and found that the molybdenyl oxygens were cis. Additional work

by Moore soon demonstrated that his original proposal was wrong.

In fact, the results of all subsequent studies of octahedral dioxomoly-

bdenum(v1) complexes indicates that to date they all contain a cis-

M0022+ moiety.

The fact that these complexes adOpt a cis configuration has
important stereochemical implications. It means, for example, that

with symmetric ligands these complexes have C symmetry and that the

2
R groups exist in two different environments (cf, R8 and Rb in II).
As a result of some dynamic change in the coordination sphere of the
central metal atom the process of site exchange may occur, wherein R3
is interchanged with Rb. When these R groups contain hydrogen atoms
as in methyl, isopropyl, tertiary butyl, etc., then proton magnetic
resonance spectroscopy is an ideal tool for observing site exchange.
The shapes of the pmr lines arising from the nonequivalent environments
are a function of the rate of site interchange. Thus, from the depen-
dence of line-shape on exchange rate, it is possible to determine the
rate of molecular rearrangement.”-19
The significance of studying the dynamic stereochemistry of
B-diketonate complexes of dioxomolybdenumCVI) is more than simply
filling an important information void or determining whether they are
cis and/or trans isomers in solution, or even whether they rearrange
at all. Although these are certainly meaningful questions, their real
significance lies in the fact that they provide a unique opportunity
to determine the relative importance of metal ion size and charge on

the lability of his and tris diketonate complexes of d0 and d10 metal

ions. These were the principal questions addressed in this research.

At the time this research was initiated, several dynamic stereo-
chemical studies of d0 and d10 metal diketonate complexes had been
reported. Their relative labilities depend on the nature of the

metal ion as follows: Ge4+(0.50 K)20 < A13+(0.S3 R)21‘23 <

21-23 24,25 < Zr4+(0.80 X)’26 Sc3+(0.81 A),23

Ga3+(0.62 K) < Ti4+(0.68 X)
In3+(0.81 A),22 where the numbers in parentheses are the Pauling
crystal radii. Regardless of the coordination number of the central
ion, the labilities spanned a rather remarkable range, from Ge4+ which
rearranges with a half life on the order of several minutes to Zr4+,
Sc3+ and In3+ which rearrange with a half life of 10-5 seconds or less.
Based on the above lability sequence, it appears that ionic radius is
an important factor determining the rates of molecular rearrangement.
At the onset of this work the relationship between lability and ionic
radius had not been recognized. It is interesting to note that the
ionic charge, in particular the charge to radius ratio, does not appear
to play a significant role. This is important because if a bond
rupture mechanism operated in the rearrangements, then one would expect
the energy needed to open a M—O bond to increase and the rates to
decrease with increasing metal ion charge.

The dioxomolybdenumCVI) diketonate complexes offer the opportunity

to probe the relative importance of size versus ionic charge better

than any of the previously known d0 or d10 metal complexes. Molybdenum(VI)

has a Pauling radius of 0.62 A, the same as that of 633+ while

3+

having a charge to radius ratio twice that of Ga If the charge on

the central metal atom is going to exert its influence, it certainly

should do so in the case of Mo6+. Thus a study of the dynamic stereo-
chemistry of dioxomolybdenumCVI) diketonates was not only expected to
answer important questions about the nature of dioxomolybdenumCVI)
complexes, but also to improve our understanding of the dynamic stereo-

chemistry of all d0 and d10 metal complexes.

II. EXPERIMENTAL

A. Reagents and Solvents

Sodium molybdate and acetylacetone were obtained from Matheson,
Coleman and Bell, and were used without further purification.
Dipivaloylmethane was obtained from Aldrich Chemical Company and used
without further purification. Pivaloylacetone was prepared in our
laboratory by Jerry Howe, using the method of Adams and Hauser.27
Alan Schwartz prepared pivaloylisobutyrylmethane, acelylisobutyryl-
methane and deuterated disobutyrylmethane by the method of Adams and
Hauser. Benzoylisobutyrylmethane was also prepared by this method.
The purity of the B-diketones, including those purchased, was checked
by nmr spectroscopy.

Methylene chloride, carbon tetrachloride and chlorobenzene were
dried over calcium hydride for at least 24 hours and freshly distilled
before use. Benzene and hexane were dried over lithium aluminum hydride

for at least 24 hours and freshly distilled before use.

B. General Techniques

All glassware was dried at least 24 hours at 180°C and cooled in
a desiccator whenever possible. Hygroscopic reagents and products
were handled and transfers effected in a polyethylene glove bag under
an atmosphere of dry nitrogen. The B-diketones synthesized in our

laboratory were purified by vacuum distillation.

7

Proton magnetic resonance spectra were obtained on a Varian
A 56/60 D analytical spectrometer operated at a frequency of 60.000MHz
or a Varian HA-lOO analytical spectrometer at 100.000MHz. The probe
temperature was maintained at :_O.S°C with a Varian Model V-6040
temperature controller. Temperatures were determined by measuring
the chemical shift differences between proton resonances of methanol
(low temperatures) or ethylene glycol (high temperatures) and
substituting them into the equations of Van Geet.28 Magnetic field
sweep widths were calibrated by the audio-frequency side-band technique.
All spectra were recorded at field strengths well below the level
required to produce saturation. Chemical shifts were measured using
tetramethylsilane as an internal standard.

Theoretical pmr line-shapes for two site exchange were calculated
using the Gutowsky-Holm equation13 as modified by Rogers and Woodbrey14
(see appendix C ref. 29). Multisite exchange spectra were generated
using a modification of the Whitesides-Lisle EXCNMR computer program30
developed by Pinnavaia and Teets (see appendix ). Calculations were
performed on a CDC 3600 or CDC 6500 computer at Michigan State

University.

C. Syntheses

cis-Dioxobis(acetylacetonato molybdenum(VI)

 

cis-Dioxobis(acetylacetonato)molybdenum(VI) was prepared
according to the method of Gehrke and Vea1.31 Sodium molybdate
dihydrate (4.85 g, 0.02 moles) was dissolved in 50 ml of distilled

water. The solution was adjusted to a pH of ml with 6 N hydrochloric

acid and placed in a 100 m1 beaker. The beaker was placed on a stir

8

plate in a darkened hood, as the product was reported to be light
sensitive.31 Acetylacetone (6.0 g, 0.006 moles) was added to the
beaker with vigorous stirring. A pale yellow precipitate of MoOz(acac)2
formed almost immediately. The mixture was allowed to stir for approxi-
mately 15 minutes, and then it was filtered through a glass frit and
washed with cold ethanol (0°C). The pale yellow solid was dried in
vacuo at room temperature for several hours (yield, 49%). Identity

of the complex was confirmed by its IR spectrum, which matched exactly

the published IR data for MoOz(acac)2.31’32

The complex was used with-
out further purification. It can,however, be recrystallized from hot

acetylacetone:53

cisvDioxobis(dipivaloylmethanato)molybdenum(VI)
Attempts to prepare cis-dioxobis(dipivaloylmethanato)moly-
bdenum(VI) by the same procedure used for the acetylacetonato complex
were unsuccessful. Similar attempts to prepare it by ligand exchange

were also unsuccessfu1.

MoOZ(acac)2 + 2 Hdpm —+—f—->'M002(dpm)2 + Hacac (l)

The dipivaloylmethanate derivative was finally prepared using a
modification of M. Jones procedure for making the acetylacetone complex.34
Sodium molybdate dihydrate (2.42 g, 0.01 moles) was dissolved in 25 ml
of water in a beaker and the solution made very acidic with the addition
of 1.3 ml of 90% nitric acid. Dipivaloylmethane (5.53 g, 0.03 moles)
was added to the molybdate solution with vigorous stirring. Pale yellow

M002(dpm)2 precipitated within 3—4 hours. After approximately 24 hours

of stirring the mixture was filtered through a glass frit, and the

9

complex was washed several times with cold (0°C) hexane. The complex
was dried in_vacug_and purified by vacuum sublimation at 110°C.
cis-Dioxobis(dipivaloylmethanato)molybdenum(v1) has not been
previously reported in the literature. Its melting point was determined
as l3l.S°C using a Thomas—Hoover model 6406-H Capillary Melting Point
Apparatus. The complex behaved as a nonelectrolyte in nitrobenzene
(A = 3.62 x 10‘2 cmzohm-lmole-l at a concentration of 9.92 x 10'3 M)
and was found to be monomeric by a cryoscopic molecular weight deter-
mination in purified benzene. Recrystallized benzil was used as the
calibrating solute in the determination of the molal freezing point
depression constant of benzene (5.87 deg/molal). Freezing point
depressions were measured with a Beckman differential thermometer
graduated at 0.0l° internals and estimated to :_0.001° using a magnifying
thermometer reader. The experimental molecular weight was 475 g/mole,
which is within 5% of the actual value of 494 for a monomeric species.
Chemical analysis was performed by Galbraith Laboratories, Inc.,
Knoxville, Tennessee. Anal. Calculated for MoC H O ' Mo, 19.42;

22 38 6'
C, 53.47; H, 7.75. Found: Mo, 19.62; C, 53.26; H, 7.96.

Pivaloylacetone, Pivaloylisobutyrylmethane, Acetylisobutyryl-
methane, Benzoylisobutyrylmethane and Deuterated Diisobutylrylmethane
Derivatives of DioxomolybdenumjVI)

M0022+ complexes of pivaloylacetone, pivaloylisobutyrylmethane,
acetylisobutyrylmethane, benzoylisobutyrylmethane and deuterated
diisobutyrylmethane were prepared using the method outlined for the
preparation of cis-dioxobis(dipivaloylmethanato)molybdenum(VI). All
were pale yellow solids. Physical properties of these complexes are

given in Table I.

10

 

us: no :oAumNAAAmumAho
Ion Mom eczom uco>Aom
oAnmuASm oz .cofiuchEmucoo

A.HOM:WSw:oo ”A.Homcflcvfioo ”A.Hom.fimezomzo

 

 

 

 

 

mo uAswoh m on Ame AzoA o N ko NAEnAmeooz
AAAosmsco ocfioo mcflvoz .A. Now. Amy oo: o A. How. AmVAum : u ”A.Aom.AmVN Au mu
. . m o
AoNAAovoo oooN o>Am A How EVAU m u .AmA NAoo>ovNooz
noNgz venom moco>Aom oz .A. How. Amy :: u A. Aom. syn Home A How. euN Au :0
o
.om.m
AAzoAm N A A ANS N .OAA NAeoNe- ovNooz
Apo> owcmgoxo mszfihousma ”A.A0m.EUN Au v: u N A Am.A0mV Au :0 m
.I . N
ocoNcoooHoA: A. Homv NAon :: u- N A .I 3A Aom. >V20m : u.I.A. Aom. >WzoN :u .NmA NAsoevNooz
:A AAvoou momooEoooo A Aomw :: o A Aomvm Aozu A AomvN AuN :u
muco>Aom oApNEOHm :A I. m
momooeoooo ecu o>AoAmcom .IA. Aomv NAzoV :N u- N AI o.A Aom. >Uzom : uI o. A. Aom. >VzoN :6 .mwA NAomoovNoo:
oANAA mA onoEou A. How. Amy o: ouo A. How. Amvm Home A How. AvauN mu
ucAom
mthEom onAAAosAom NcAvoz .monoEou

moonanu AA>VescmnnAAoEAOumcouoan-mvmAnoonQ woNAmospcAm mo mowuhomoud

.A vomA

11

.mmcflxoxAm Show ow Hmoamm pap .oncooAm :A AAAvon Ahm> o>AommAw mmxmAmeou o

.AomCA u mAnsAom:A A.A0m.Am u mAnsAom AAuamAAm
A.A0m.E u oAnsAom AAmumnmon A.A0m n oAnsAom A.Aom.> u mAnsAOm Ahm> n

.Hvou zoAAoA mAwm m mum mmonmsoo on“ AA< m

 

mcoNcoDOHuAc
:A AAwwmma momomeoomn

mxamemm

A.Aom:Aszou
”A.Aom:AVozou ”A.A0chAAAUU ”A.Aom.eUAumzcu mmA NAEAADANOOE
.A. N v N - . . . m A . N N o .
. Aomv Au : u N A .A Aomu Ammo .A Aomv Au mu
unwom

d5:33ko maApAoz mongeOu

Aw.pcouv A mAAmA

III. RESULTS AND DISCUSSION

A. Preparation and Characterization of Dioxobis(dipivaloylmethanato)-

molybdenum(y1),

 

Prior to this work only a few complexes of dioxomolybdenum(v1) had
been reported.8 Furthermore, all of the known molybdenyl B-diketonate
complexes are not especially well suited for accurate kinetic studies by
nmr line broadening methods due in part to their limited solubility in
organic solvents and their tendency to decompose in the presence of

8,31

light. The acetylacetonate derivative, which is the most widely

4-9,11,12,31-40

studied and best characterized complex, is only sparingly

soluble in most common organic solvents and decomposes upon standing in
solvents such as benzene, methylene chloride and chloroform 8’12’31 to
give a molybdenum blue. Furthermore, upon exposure to light the complex
gradually turns blue—green.

Thus, the need for a new, more stable and soluble bis diketonate of
dioxomolybdenum(VI) was apparent. In the hOpe of fulfilling this need
we attempted the synthesis of MoOz(dpm)2, where dpm = dipivaloylmethanate,
by the method of Gehrke and Veal.31 This method was unsuccessful with
the molybdenum apparently undergoing gradual reduction to a molybdenum
blue.41 A modification of this procedure, wherein concentrated nitric
acid was used in place of dilute hydrochloric, was successful. Two
possible reasons for this may be that (1) an oxidizing acid is necessary

to prevent reduction of the molybdenum(VI) and (2) depolymerization of

the isopolymolybdates occurs in very strong acid solutions.41

Na M00 + 2Hdpm + 2HNO

2 4 + M002(dpm)2 + 2H 0 + 2NaNO (2)

3 2 3

12

13

The dioxobis(dipivaloylmethanato)molybdenum(VI) proved to be much
better suited for an nmr line broadening study than previously known
derivatives. It was relatively soluble (10-20 g per 100 ml solvent)
in solvents such as methylene chloride, chloroform and chlorobenzene.

In addition, the complex was much more stable in solution, decomposing
only after aging for several weeks. However, the compound decomposed

in only 3-4 days in benzene and 1—2 hours in nitrobenzene. Furthermore,
the dipivaloylmethanate derivative was not sensitive to light, and only
slightly sensitive to atmospheric moisture, gradually turning blue-green.

Infrared vibrational frequencies agree very well with those of the
analogous acetylacetonate complex.32 The molybdenyl stretch, vMo-O
(904 cm-1) is identical, while the ligand oxygen to metal stretches,
VMO-O (512, 481 cm'l) are shifted slightly to lower energy. The ligand
vC==0 (1584 cm.1 and VC:;C (1500 cm-1) are also in

exact agreement with the assignments for acetylacetonate complex.

ring stretches,

Figure 1 shows the infrared spectrum of dioxobis(dipivaloylmethanato)—
molybdenumCVI) in the region 1650-450 cm.1 and table II gives some

of the important assignments based on comparisons with the analogous
acetylacetonate complex, for which Soptrayanov, Nikolovski and Petrov32
have made spectral assignments.

Proton magnetic resonance spectroscopy clearly shows that M002(dpm)2
exists solely in the cis configuration. The pmr spectrum exhibits only
one ligand -CH= line indicating that only one of the two possible isomers
is present; while, the ligand tertiary butyl region exhibits two lines
of equal intensity, consistent with the C2 symmetry of the cis isomer.
The relative intensities of the lines are 1:9:9 thus, confirming a cis

dioxo structure. This pmr spectrum was observed in both aromatic and

14

Figure 1. Infrared spectrum of M002(dpm)2 as a nujol mull between
CsI salt plates. The 1601.8 and the 1028.3 cm'1
absorptions of polystyrene were used as references.

 

15

Ohm

oov

ooo

 

 

oom

 

 

 

 

 

.2 U
89

 

 

H enamwm

oom—

 

00V—

 

coo,

oom—

16

Table II. Infrared Assignments for MoOz(dpm)2

 

5 (cm-1) 1 Approximate Description
1584 V8 C220 stretch
1500 VS C23C stretch
1024 m Mo = 0 stretch
938 V8 C---CH:5 stretch + C::O stretch
904 V8 Mo = O stretch
806 S C-—H deformation (out-of—plane)
512 S

3 Mo-—O stretch
481 m

 

I = intensity; m, medium; S, strong; VS, very strong.

17

non-aromatic solvents. Figure 2 shows the pmr spectrum for M002(dpm)2

in chlorobenzene.

B. Kinetics Study of Dioxobis(dipivaloylmethanato)molybdenum(VI)

Rearrangement.

 

In methylene chloride solution below 35°C, the interchange of the
nonequivalent dipivaloylmethanate tert- C4H9 groups is sufficiently slow

to observe two well resolved tert- C4H9 proton resonance lines in the
pmr. Above 35°C the two lines broaden and coalesce into a very broad
line (ca. 50°C), which then sharpens above the coalescence temperature
(see Figure 3). The Egrt- C4H9 resonances did not exhibit temperature
dependent chemical shifts and the separation of the two lines in the
absence of exchange (low temperature) was 6.84 Hz. The natural line
width at half maximum amplitude was 0.55 Hz for both lines.

The weak central peak in Figure 3 corresponds to an approximately
1% impurity of free ligand. The amount of free ligand impurity could
be reduced to this level by recrystallization or sublimation, but it was
never completely removed. This small impurity did not interfere with
the line shape analysis and, in fact, served a useful purpose. It
proves that the rearrangement process is an intramolecular process and
not intermolecular (i;g:J complete ligand dissociation does not operate)
because the free ligand line remains unchanged throughout the exchange
region. If ligand dissociation were responsible for the observed site
exchange then we would expect to see the free ligand line become

averaged (coalesce) with the complex lines. This is not observed.

Further, evidence that the exchange process was exclusively intramolecular

18

Figure 2. Proton nmr spectrum (60MHz) of.MoOz(dpm)2 in chlorobenzene
at ambient machine temperature (22: 40°C); concentration
lOg/lOOml of solvent.

 

19

mm.w ww.w

 

N unnumm

AEQmVp

vw.m

fry

20

Figure 3. Temperature dependence of tert—C4H9 resonance lines of
MoOz(dpm)2 in methylene chloride; concentration is
log/100ml solvent.

 

 

21

 

 

35.1° ~-

Figure 3

22

was the fact that adding free ligand had no effect on the rate of inter-
change and that the rates were independent of complex concentration.
Gutowsky and Holm13 have shown that the mean lifetime, TA and TB’
of protons exchanging between two nonequivalent sites can be determined
from the nmr line shapes if the frequency separation between the
resonance lines in the absence of exchange, 6v, and the transverse
relaxation time T2, are known. This is done by matching the theoretical
spectra calculated using the Gutowsky-Holm equation as modified by

Rogers and Woodbrey14 (3) for different values of r (r = rArB/(TA+rB)

with experimental spectra.

 

P P
B A
-w,Mo P[l+r( + —-—-)] + QR}
{ T2A T213

 

 

 

 

 

= (3)
P2 + R2
1
6w P P
where P E T[-————- - sz + (-—-—)2] + B + A
T2A TZB 2 T2B T2A
Q = T[Aw-5—w(P-P)]
2 A B
R E Aw[l + 17(T1 + T1” + % (Tl— - TL) 4. £559. (FA-PB)
2A ZB 23 2A

v describes the proton absorption line shape (amplitude)

”1 describes the applied radiofrequency field

Mo is the static nuclear magnetization at thermal equilibrium
PA,PB are the proton fractions contributing to each component
TZA’TZB are the transverse relaxation times

Aw represents different values between the applied radiofrequency and
the frequency at the center of the two resonances

So is the separation of the components assuming no exchange and no
overlap of the components

23

Values of T forMoOz(dpm)2 in the region of exchange were
determined by a trial and error method. Theoretical spectra calculated
for different values of I were plotted and compared to the experimental
spectra of MoOz(dpm)2 at different temperatures. This process of
comparison was repeated until the best possible match of total line
shape was obtained for each experimental spectrum. Figure 4 shows the
experimental spectra with their matching theoretical spectra and T
values. Table III shows the nmr line shape parameters and kinetics data
for M002(dpm)2, where k = l/Zr.

Activation parameters were calculated using the first-order rate
constants determined from the nmr line shape analysis. A plot of In k
XE: l/T is shown in Figure 5. A linear least squares analysis (see
ref. 29 appendix B) was used to determine the $10pe and intercept,
which gave an Arrhenius activation energy and frequency factor of
17.0 :_l.0 kcal/mole and exp (12.56 :_0.57), respectively. The errors
in the activation parameters are estimated to the 95% confidence level.
An activation entropy of -3.0 :_2.6 eu at 25°C was calculated from the

frequency factor assuming that the Eyring equation (4) holds.

3!-

AS = R In A-R[1 + In 551; (4)

R = gas constant = 1.9872 defined cal/°K mole
A = frequency factor

T = temperature (°K)

k = Boltzmann constant = 1.3805 x 10.16 erg/°K
h

= Planck constant = 6.6254 x 10-27 erg sec

 

24

Figure 4. (A) Temperature dependence of tert-C4Hg resonance lines of

MoOz(dpm)2 in methylene chloride; concentration is lOg/100ml
solvent; 60MHz spectra. The weak central peak corresponds
to a 1% impurity of free ligand. (B) Calculated pmr

spectra for the mean lifetimes indicated. The values of

the frequency separation in absence of exchange (6v) and

the transverse relaxation time (T2) used in the calculations
were 6.84Hz and 0.58 sec, respectively.

 

25

v ohamwm

00m 02.0

no» Rood

com 0306

0mm ommod

<m oomod
A1 V

o
NIn

 

 

26

Table III. tert - C H9 PMR Line Shape Parameters and Kinetic
Data for MoOz(dpm)2 in Methylene Chloride

 

Temp, °C rE- W*23 Hz k, sec.1
35. 5.25 3.60
39. 3.12 5.20
39. 2.92 5.43
43. 1.92 7.43
46. 1.37 9.56
50. ”1.00 7.40 14.3
54. 5.80 17.9
57. 4.10 25.0

|U1W

Ratio of maximum to central minimum amplitude below coalescence.
Full line width at half maximum amplitude above coalescence.

 

27

Figure 5. Lnky_s_. l/T plot for M002 (dpm)2 in methylene chloride
solution.

 

28

Figure 5

1N6

:0 9N H O.” .. Newmwd
36302 H < co.
22:28.. 3 H9: N am

Cod
q

 

v0.0
q

1r..."

06

 

nun

29

These results are in accord with those of Wise, et al.,11 for the
MoOZ(acac)2 complex. Their results are summarized in Table IV which
is a reproduction of Table III of reference 11.

The results of our study are extremely interesting because they
show that this Mo6+ complex rearranges at approximately the same rate

21’22’23 Thus, the fact that Mo6+ has

as analogous Ga3+ complexes.
twice the charge to radius ratio of Ca3+ does not have any appreciable
effect on its relative lability. Quite clearly, this illustrates that
ionic radius is an important factor in determining the relative rates

of rearrangement for d0 and (110 metal B-diketonate complexes, and that

it is much more influencial than ionic charge. Since ionic charge

would be expected to have an important effect in bond rupture mechanisms,
whereas, ionic size should have the major effect in twist mechanisms, it
would appear that a twist mechanism may be the principal method of

rearrangement for the known d0 and d10 metal B—diketonates.

c, Preparation and Study of Dioxobis(dz-diisobutyrylmethanato)-

 

molybdenumgyll,

Dixoxbis(dS-diisobutyrylmethanato)molybdenum(VI) was prepared by
a modification of Gehrke and Veal's method described earlier fOr
dioxobis(dipivaloylmethanato)molybdenum(VI). The reaction was carried
out in deuterium oxide in place of water as a solvent. Although hydro-
gen ions were available as a result of the nitric acid used, this did not
constitute a problem since deuterium exchange is very slow. Proton
magnetic resonance spectroscopy confirmed that almost no deuterium

exchange occurred during the course of the reaction.

30

Table IV.

Benzene soln

 

E = 17.0 kcal/mol

A

AH1 = 16.4 kcal/mol

AS1 = 4.4 cal/deg mol
0(EA, AHt) = 2.4 kcal/mol
0(ASI) = 8 cal/deg mol

Activation Parameters for M002(acac)211

Chloroform soln

 

E = 13.0 kcal/mol

A

AB: = 12.4 kcal/mol

AS* = 14 cal/deg mol

0(EA, A°HI) 1.0 kcal/mol
o(AS*) = 4 cal/deg mol

31

The M002(d3—dibm)2 complex has some interesting properties. As
in the case of the dpm derivative the iSOprOpyl groups of'MoOz(d3-dibm)2
should exist in two different environments because of the C2 symmetry
of the complex. In addition, as a result of the dissymmetry (no mirror
plane) the methyls on each isopropyl group are diastereotopic (labeled
r and s in III) and as such should give separate lines in the pmr

spectrum. It should be noted that observation of diastereotopic effects

 

 

 

 

 

M.a(r)
a (s) D
b (”Me Me
o+
b (”MO
Mo
b(”Me
b(l’)M. .
a(s)M. Aha“)
III

in chiral metal chelates was, until very recently, relatively rare.“-46

This means that for this deuterated complex in the absence of exchange
there should be four isopropyl methyl lines in the pmr. This is, in
fact, what is observed. Since inversion of configuration is the only
way for the diastereotopic methyls to interchange environments, it

means that this is a complex in which one can compare the relative rates
of inversion and site exchange. This has important mechanistic impli-
cations because for twist mechanisms inversion always accompanies site

exchange (same rate), whereas, for bond rupture mechanisms inversion

32

may or may not accompany site exchange. Thus, if the rate of inversion
and the rate of site exchange are different then a bond rupture mechanism
must operate at least to some extent.

In 1,2-dichloroethane solution below 20°C the interchange of the
nonequivalent isopropyl methyl groups is sufficiently slow to observe
four well resolved methyl proton resonance lines in the pmr. Above
20°C the two low field lines coalesce and as the temperature is raised
to 47°C the two high field lines broaden and merge. These two lines,
which are already exchange broadened, then coalesce (ca. 60°C) into a
single line which becomes sharper as the temperature is raised (see
Figure 6).

An approximate determination of the rates of inversion and site
exchange was made by applying the simplified Gutowsky-Holm equation (5)
to the coalescence of the two high field lines and then to the
coalescence of the high field pair with the low field pair of lines.

The errors associated with the simplified equation and its region of

k = 5%.: (n//2)(6v2-6ve2)1/2 (5)
6v = frequency separation (Hz) of the lines in
the absence of exchange
éve = experimentally determined frequency separation

of the lines in the region of exchange

applicability have been determined by comparison with complete nmr

13,47,43

line shape analysis. It has been shown that, in general, the

error introduced by using equation (5) does not exceed 10% for systems

47

for which 6v Zr :_5. In the M002 (dS—dibm)2 system 6v Zr did not

exceed 1.5. No attempt was made to determine which coalescence

Figure 6.

 

“3*

33

Temperature dependence of isopropyl methyl lines of
‘MoOz(d3-dibm)2 in 1,2—dichloroethane; concentration is
lOg/lOOml solvent; 100MHz spectra.

 

34

 

 

 

37°

 

 

 

 

 

 

19°

 

Figure 6

 

3S

represented lines being averaged by site exchange and which represented
the inversion process. Application of equation (5) gave a first order
rate constant at 25°C for the averaging of the two high field lines of
8.58 sec-1, and for the averaging of the high field pair with the low
field pair of 8.33 sec-1. This means that within the accuracy of the
determination the rate of inversion is equal to the rate of site
exchange. These results strongly suggest that a twist mechanism is
operating since inversion always accompanies site-exchange for a twist,
but in the case of bond rupture site exchange may occur without inversion

of configuration.

0, Preparation and Study of Dioxobis(benzoylisobutyrylmethanato)-

 

molybdenum(VI).

 

Dioxobis(benzoylisobutyrylmethanato)molybdenum(VI) was prepared
by a modification of Gehrke and Veal's synthesis for M002(acac)2, which
has been described earlier for dioxobis(dipivaloylmethanato)molybdenum(VI),
Benzoylisobutyrylmethanate is an unsymmetrical ligand and as
such the complex M002(bibm)2 may exist in three different configurations

(isomers) with the molybdenyl oxygens cis. The isopropyl methyls are

1-P

H

¢

[M

is

 

O—

x
5

 

 

 

¢)\r-l NM 1.1).
i-Pr ¢
1v V VI

36

diastereotopic as they were for the diisobutyrylmethanate derivative,
except that in this case the ligand is not deuterated. Assuming that
all three isomers exist in solution, the pmr spectrum should be very
complicated. Isomer IV should give two isopropyl methyl lines for the
diastereotopic isopropyl groups, each of which is split by the proton
on the isopropyl group, resulting in a total of four lines. Similarly
isomer V should also give four lines in the pmr spectrum. In isomer VI
the isopropyl groups exist in two different environments, thus giving
rise to two groups of four lines or a total of eight lines. The pmr
then should consist of sixteen lines in the isopropyl methyl region,
assuming all three isomers are present and all the lines are resolvable.
Only twelve lines were resolvable even though the ligand -CH= region
indicated that all three isomers were present (see Figure 7).

The nmr studies of M002(bibm)2 were hampered because the complex
was not appreciably soluble in many organic solvents. Concentrations
of lOg/lOOml or higher could be obtained in methylene chloride, chloro-
form and 1,2-dichloroethane, but solubility was very low in aromatic
solvents. The pmr spectrum was essentially the same in all the
suitable solvents, and no more than twelve isopropyl methyl lines could
be resolved in any solvent at any temperature. The ligand -CH= exhibited
only one line at room temperature which would broaden at low tempera-
ture indicating three isomers were present, but never splitting out
sufficiently to resolve three separate peaks.

The compound was prepared in the hope of analyzing its temperature
dependent nmr spectra, using the adaptation of Whitesides-Lisle EXCNMR

program for multisite exchange as developed by Pinnavaia and Teets and

37

Figure 7. ISOprOpyl methyl lines of M002(bibm)2 in 1,2-dichloroethane;
concentration 20g/100ml solvent; 100MHz spectra. The lines
marked "x" are due to a small amount of free ligand impurity.

 

 

 

 

 

 

38

 

 

 

 

X
X
69’
X
58°
X 49'
40'
X
17 '

Figure 7

39

described in the appendix. Unfortunately because of poor solubility
in many solvents and lack of resolution it was not possible to make
unambiguous line assignments, which are essential to determining the
detailed method of rearrangement for these complexes. The pmr shift
reagent Eu(fod)3, where fod = 1,1,l,2,2,3,3-heptaf1uoro-7,7-dimethyl-
4,6-octanedione, was used unsuccessfully in an attempt to spread the
spectrum out and improve resolution.

The interpretation of the pmr spectrum in terms of a particular
mechanism is further complicated by the fact that the coalescence pattern
is very complex. It appears that all the isomers are being averaged
simultaneously (see Figure 7), which means that no one simple mechanism
is operating. This point can be illustrated better by a comparison with
the results found for Fe(III)50 and other metal dithiocarbamates,51
where the data indicates that intramolecular rearrangement occurs via
a trigonal twist mechanism (i.e. rotation about a real or pseudo C3
axis in the molecule). The dioxomolybdenum(VI) complexes do not
possess a real or pseudo C3 axis, but if rotation were to occur only

about that axis which most resembles a pseudo C axis (VII) then we

3

should see a very characteristic coalescence pattern. Figure 8 shows

1
VV

‘7 '

the possible isomers for a dioxomolybdenum(VI) complex with unsymmetrical

VII

ligands, such as M002(bibm)2. Although the trans isomers are depicted,
there is no evidence to indicate that they exist, except possibly as

short lived intermediates. Figure 8 also shows the enantiomeric forms

40

Figure 8. Possible isomers of M002(bibm)2

 

 

 

cl 3
_
z [-

 

\

°\.
mfg

a:
2 'U
U

 

C

 

 

 

 

:bg

42

of each isomer, because, while enantiomers can not be distinguished

in the pmr spectrum, nevertheless, interconversion (inversion) leads

to averaging of the diastereotopic isopropyl methyls of the M002(bibm)z
complex. Figure 9 shows what happens when "pseudo" trigonal twist (VII)
operates in each of the possible isomers. The overall effect then is

to interconvert isomers Cd,£ and T2d,£’ and to simply scramble the

environments within the Tld,£ isomers. Thus, if this mechanism alone
operated in M002(bibm)2 the four proton resonance lines of isomer IV
would coalesce with the four lines of isomer V giving only a spin coupled
doublet in the region of rapid exchange. Similarly the eight lines of
isomer VI would collapse to a doublet. This is not what is observed

(see Figure 7), and so we can rule out a simple Bailar (trigonal) twist
as the sole rearrangement pathway. Likewise, rotations solely about
other specific axes in the complex can be ruled out. The mechanism of
rearrangement of these complexes must then be a nontrivial one, in which

rotation about several axes in the complex occurs or in which bond

rupture occurs producing several of the possible intermediates.

”'1'“:- -

._-_ “A,

 

Figure 9.

43

Results of trigonal twist mechanism operating in
MoOz(bibm)2

 

 

 

 

Figure 9

IV. SUMMARY

The relative rates of rearrangement of neutral d0 and d10 his and

tris metal B-diketonate complexes appears to be influenced primarily by
the size of the metal ion. With the exception of Sn4+ (0.41 A), the
following order of increasing stereochemical lability with increasing
Pauling crystal radius of the metal ion is observed for the known neutral,
six-coordinated, d0 and d10 metal B-diketonates: Ge4+(0.50 A) <
A13+(0.53 R) < Ga3+(0.62 K), Mo6+(0.62 K) < Ti4+(0.68 K) < Zr4+(0.80 3),
Sc3+(0.81 A), In3+(0.8l A). Recently Jones and Fay49 have shown that
Sn(acac)2X2, where X = F, Cl, Br, I, rearranges slower than the

Ti(acac)X2 analoges. Nevertheless, the importance of metal ion radius

in the above series is clearly evident.

Pinnavaia has interpreted the rearrangements of both aluminum(III)
and gallium(III) diketonates in terms of a bond rupture mechanism,21
whereas Holm,8 has favored bond rupture for the former complexes and
twisting for the latter. The equal rates of inversion and site exchange
for the M002(d3-dibm)2 as well as the similar labilities of molybdenyl
and gallium(III) B-diketonates would appear to be explained adequately by

twisting mechanisms, because the energy required for the M0 core to

6
achieve trigonal prismatic configuration should be about the same in
both cases. It should be emphasized, however, that this interpretation
does not unequivocally preclude the possibility of bond rupture
mechanisms. Although the formal charge to radius ratio of Mo6+ is
twice that of Ga3+, the large multiple bond order (22: 2.48) for the
molybdenum-oxygen bonds of the M0022+ moiety should lower the effective

charge on the central metal atom and thus decrease the energy needed to

open a diketonate oxygen-molybdenum bond.

45

46

Recently Gould has reported that dioxobis(acetylacetonato)moly-

bdenum(VI) is a good catalyst in the epoxidation of cycloalkenes.33

The two possible mechanisms proposed imply that the active metal, in

MoOz(acac)2 .

+ ROZH

 

0+ ROH (6)

33,50

this case molybdenum, be substitution labile. Our studies10 as

well as those of Wise11 indicate that these complexes are indeed labile,

| Bu \1 /BU
:MOVI—O/ —--) —Mo-—-O‘ + Ho+ (7)
u\ (I) |\
4.
H"

Bu
-—Mo—-—O

\| /BU
———) _.Mo-—O_ + (3)
4}\ O—H \

/§S\

47

however, they do not preclude and in fact suggest that a twist

mechanism may operate. A twist mechanism would not give the substitution
lability required by Gould's mechanisms for the formation of a metal
hydroperoxide intermediate. Further, both our work and that of Wise

show that a dissociative mechanism clearly does not operate in these
complexes for the rearrangement process. It would appear, however,

that the intramolecular rearrangement of the molybdenum catalyst probably
does not play an important role in its catalytic activity, since the
epoxidation reaction is a bimolecular reaction and proceeds at a rate

of 10 to 100 times faster than the rearrangement.

 

 

PART II

I. INTRODUCTION

Silylation reactions and the factors affecting them have become
increasingly important in recent years. This is largely a result of
the fact that a wide variety of organic and biologically significant
molecules possessing an acidic hydrogen atom will reversibly form
silyl derivatives in which the hydrogen is replaced by an R38i group.
Silylation has the effect of reducing hydrogen bonding between
molecules and consequently increasing the solubility and volatility
of the parent molecule. Among the numerous examples which could be

cited,53’S4

the amino acid tryptophan which is high melting, insoluble
in most organic solvents and has a negligible vapor pressure can be
converted to the tris(trimethylsilyl) derivative (VII) which is

soluble in

- CH2(':HCOOSI (C113)3

HNSl(CH3)3

VII

’1“
31(CH3)3

organic solvents and capable of being analyzed by glpc.54 Addition of
water will regenerate the tryptOphan from the silyl derivative.

The silylamides (VIII) are thermodynamically and kinetically
superior to the silylamines (IX) or chlorosilanes (X) as silylating

agents for organic and biological molecules. KlebeSS has suggested

48

49

. ,0 . /R .
RSSl—w—c \R 12351-—N\R R3SIX
R
VIII IX x

that the remarkably high rates of silylation fer the amides are
intimately related to their ability to expand the coordination sphere
of the silicon atom from four to six. As a result of incipient extra—
coordination of silicon by the amide oxygenation, octahedral transition

states or intermediates (XI) are considered likely in the proton

transfer reaction:

 

 

 

 

 

 

0:8,R
R . N\R RCONHR
RCONRSiRs + R'OH‘—-—-& Si -—————9 -+ (9)
n ' °
R n R 081R
F 3
R'OH j
XI

In contradistinction, silylation by silylamines or silylhalides must

proceed through higher energy five-coordinate transition states:

HOR
I
. .tR .
I __ °- 1
R381X + R OH ————) R IIQR ————> R3310R + HX (10)
X

This hypothesis, while reasonable, is in most cases impossible
to test. However, silyl enol ethers constitute a group of organosilicon
compounds uniquely suited for studying the effects of valence shell
expansion on substitution rates. The compounds possess sequence cis

(XII) and sequence trans forms (XIII).56 Incipient extracoordination

50

3 _
R381-0\ /C\ RSSr—O\ ~ H
//C==C\\ R. C—-C
R H R C==O
R
XII XIII

. . . . . . 5
Is pOSSIble 1n the seqc15 Isomer, but not in the seqtrans form. 7

 

Thus if the ability of silicon to expand its coordination sphere from
four to five or from four to six are important factors in substitution
processes of organosilicon, then the rate of silylation of an alcohol
should be appreciably more facile for the seqcis than seqtrans form.
The objective of this project, therefore, was to test the importance of
incipient extracoordination in the silylation of methanol by the seqcis

and seqtrans forms of trimethylyilylacetylacetonate (R = CH3 XII and XIII).

NOTE = For an explanation of the terms seqcis and seqtrans see

reference 56.

II. EXPERIMENTAL

A. Reagents and Solvents

 

Methylene chloride and chlorobenzene were dried over calcium
hydride for at least 24 hours and freshly distilled before use.
Methanol was dried over calcium sulfate for at least 48 hours and

freshly distilled before use.

B. General Techniques

 

All glassware was dried at least 24 hours at 180°C and cooled
in a desiccator whenever possible. Transfers were effected in a poly-
ethylene glove bag under an atmosphere of dry nitrogen.

Proton magnetic resonance techniques have been described already

in the experimental section of Part I.

C. Synthesis of Trimethylsilyacetylacetonate

 

Trimethylsilylacetylacetonate was prepared and purified in our

laboratory by Jerry Howe according to the general method described by

West.58 Bp = 66-68° (4 torr), n 23°6 = 1.4507, lit.58

nD25 = 1.4546.

D Bp = 66-68 ,

Anal. Calculated for SiC8H1602: Si, 16.30; C, 55.77; H, 9.39;

M01 wt, 172. Found: Si, 16.51; C, 55.70; H, 9.21; mol wt, 184.

51

 

.191. . . 4.1... .a: 3.4....de

III. RESULTS AND DISCUSSION

Pinnavaia and Howe59 have confirmed that trimethylsilylacetyl-
acetonate, (CH3)38i(acac), exists in solution as two isomers, a
seqcis (XII, R = CH3) and a seqtrans (XIII, R = CH3). The seqcis to

seqtrans ratio in chlorobenzene is 0.34 :_0.04.57 Furthermore, they

have shown that the seqcis isomer undergoes a rapid degenerate rearrange-
ment which is believed to occur via a pentacoordinated silicon inter-

mediate or transition state. A similar process is not observed for

 

 

r' _
(2) 2
o§C-CH (2) O/C/CH3 0 C’CH3( )
1 3 "\ - \\
R381 0\ /C—-H {:2 R35i\ ‘C-H (r: R351 [C-H
~ 0:0

(1 (1) O‘Cx/cn (1) ‘CH (1)

CH3 3

' 7 (11)

XII

the seqtrans isomer because of restricted rotation about the carbon-carbon
double bond. No broadening of the pmr lines for the seqtrans acetylace-
tonate methyls was observed even at 120°C. This has important implications,
because it means that upon substitution the seqcis isomer of trimethyl-
silylacetylacetonate can readily expand its coordination Sphere to six,
whereas the seqtrans isomer can only expand its coordination sphere to
five. Figure 10 illustrates the difference in mechanism for the seqcis

and seqtrans isomers reacting with methanol.

52

 

I :1
. wa “Pi-It'd

 

 

Figure 10.

53

Possible mechanism for the reaction of CH3OH with the

seqcis isomer (A) and seqtrans isomer (B) of (CH3)3Si(acac),
where R is a methyl group.

 

54

OH museum

m
omomz m/Imx m “.3
+ AHHW .nmlm xulu\ /olemmm
mmnmom m.. . munmwwnw oe x: . Amy
\\0
O
mom a
omom: m \ m \m
$32 1 oi 13ml / . 3
uuuo
[o \

55

The reaction rates of the seqcis and seqtrans isomers of trimethyl-
silylacetylacetone with methanol were studied in methylene chloride and
chlorobenzene. In a typical kinetics run, 0.16g of (CH3)SSi(acac) was
dissolved in 0.7lg of CH2C12 in an pmr tube and the solution was allowed
to equilibrate at -5.3°C in the probe of the pmr spectrometer.
Approximately 60 minutes was allowed for temperature equilibration.

A solution of 0.02g of CHSOH in 0.25g of CHZCl2 was prepared and placed
in a vial in an ice bath (%0°C). To initiate the reaction the methanol
solution was extracted from the vial with a pipet and injected into the
pmr tube which remained in the probe. The rate of change of the

mole fractions of the seqcis and seqtrans isomers was determined by
electronic integration of the ligand methyl proton resonance signals.

The important variable being measured was the seqcis to seqtrans ratio as
a function of time. At -5.3°C the silylation reaction was sufficiently
slow that more than one hour was required for complete reaction, and

the seqcis to seqtrans ratio could be easily followed. In addition, it

should be noted that the rate of silylation is much faster than the

rate of seqcis-seqtrans isomerization, which is very slow and permits

 

separation of the isomers.6O Table IV shows the results obtained in

methylene chloride and chlorobenzene.

It is apparent from these results that at least for (CH3)SSi(acac)
incipient extracoordination does not play an important role in the
rate of its silylation reactions. The methanol reacts equally well
with both isomers. This would seem to indicate that expansion of
the silicon coordination sphere is not the principle factor affecting
the activation energy in organosilyl displacement reactions as Klebe

has postulated.

56

Table V. Segcis to Segtrans Ratio of (CH3)35i(acac) During Reaction

With CH30H -s.3°c.

 

 

Methylene Chloride Solvent Chlorobenzene Solvent
Time from [cis]/[trans] Time from [cis]/[trans]
mixing(min) mixing(min)
0 .64 O .31
3.0 .60 1.0 .30
5.8 .61 2.0 .32
6.5 .60 5.0 .32
9.4 .62 6.0 .31
10.5 .60 8.7 .31
13.3 .59 10.0 .31
14.3 .61 13.0 ' .31
16.3 .61 15.0 .30
20.5 .60
22.5 .58
26.0 .61

29.0 .63

 

 

S7

Sommer61 has pointed out the importance of electronic effects in
reactions of this type, whereby the effect on the rate closely parallels
the ability of the leaving group to carry a negative charge. This
would explain why silylamides are better silylating agents than silyl-
amines or silylhalides, because the leaving anion is resonance stabilized

for the amides but not for the amines or halides. Thus, our work

suggests that electronic effects rather than incipient extracoordination

exert greater influence on rates of organosilyl displacement reactions.

PART I I I

I. INTRODUCTION

In recent years there has been a good deal of interest in the

migration of R Si groups between electronegative centers. A general

3
feature of these molecular rearrangements is that they are rapid and
reversible. This may be illustrated by the equilibrium between (XIII)

and(XIV), where X and Y represent electronegative atoms such as oxygen

Y) R351 - Y)
R3Si - X c:::3 X (12)
XIII XIV

or nitrogen. When X and Y are identical, X\_/X in its unsilylated form
is a symmetric, uninegative ligand. The symmetry of the ligand means
that the molecular rearrangement is degenerate, but nonetheless
observable by temperature dependent nmr spectroscopy. Examples of this

growing class of compounds include silyl derivatives of B-dike-

57,59,62,63 64 63 65 66

tones, malonates, tropolone, pyrazoles, triazene,

. . 67 . . . . .
benzamidine, and hydra21ne anions.68 In contradistinction, when X and
Y are not the same, then X- and Y- silylated constitutional isomers are
possible and the intramolecular rearrangement results in their rapid

tautomerization. These types of compounds are exemplified by the silyl

55,69

anilides and amides.7O

Silyl group 1, 2 migrations have been observed between the two

65

nitrogen atoms in pyrazoles (XV). The rate of rearrangement increases

(CH3)3SI--z--E
\./ \
R5 ‘9 R3
R4

XV

58

59

as the substituents R3, R4 and R5 become more electron releasing. This

electronic effect has been explained as being due to increased electron
density, hence availability, at the free nitrogen.

Nitrogen to nitrogen, nitrogen to oxygen and nitrogen to sulfUr 1,3
migrations have been reported. The silyl triazene (XVI)66 and

benzamidene (XVII)67 exibit facile interchange of the R Si group

3

between nitrogen atoms. Rearrangement rates for the triazene derivatives

as
R351-N-N==N-CH3 (CH3)331-—N\C__ <:::>
CH3 CH N6
3
XVI XVII

increase as the electron withdrawing ability of the groups attached to

silicon increases. The R Si group migrates between oxygen and nitrogen

3
55,69

in the silyl amides (XVIII),7O anilides (XIX), 1

and benzimidates (XX).7

[(CH3)3S1]2N-fi-R % X
0 R—C—N-—
(CH3)351
XVIII XIX

(CH3138iN§t-_<g:§3{x
. /
(CH3)SSIO

XX

60

The rate of migration increases in the anilides and benzimidates as

electron releasing groups are replaced on the carbonyl carbon.69’71 This
trend parallels that found for substitution on the pyrazole ring. In
addition, electron withdrawing substituents increase the equilibrium
concentration of the oxygen silylated isomer.55 Itoh has reported a

l, 3 silyl migration between sulfur and nitrogen in the adduct of bis-

trimethylsilyl sulfide with phenyl isocyanate (XXI).71

81. (CH ) "
33‘N-c—0—SImH

Cf 3°

XXI

A l, 4 migration of the SiR group has been observed by Reich in

3
2-(2-butyldimethylsiloxy)cyclohepta-Z,4,6-trienone (XXII).63 He notes

- 0
(C”5)2?1 0‘\ 46
cuscnzcncn3

XXII

that while this and other silyl shifts (e.g. XII, XV, XVI) can be
considered sigmatropic rearrangements, the facility with which all of
these reactions proceed suggests that they are best considered as
internal nucleophilic displacements.

Oxygen to oxygen 1, 5 migrations are well documented for the silyl
carbomethoxyketene acetals72 or malonates (XXIII) and silyl enol ethers

57,59

of triacelylmethane (XXIV)62 and acetylacetone (XII). The

 

 

 

 

 

61

$“3 ?”3
(CH) 51-0 /c ,c\
33 \f/mm3 0/ \fi \0
I
O\c/C\R CH/C\OS' CH
. 5 1( :93
OCH3
XXIII XXIV

rearrangement is only observed for the seqcis form of (XXIII) and (X11)
and not for the seqtrans forms. Pinnavaia has observed dramatic sub-
stituentS7 and angle strain effects73 on the rate of R351 group migrations
in enol ethers. Sigmatropic shifts should be relatively insensitive to
these effects and, hence, it appears that these 1, 5 shifts are best
viewed as internal nucleophilic displacements.

The objective of this work was to prepare and investigate a new
class of silyl B-ketoamines with potential for 1, 5 migration of the
R331 group between oxygen and nitrogen. The silyl B-ketoamines
represent a complex system in which six different isomers might be

expected (see Figure 11), four oxygen silylated and two nitrogen silylated

isomers. The four 0- silyl isomers differ by virtue of seqcis-seqtrans

 

isomerization about the C=C bond and syn-anti isomerization about the
=NR bond. The N-silyl isomers can exist only as seqcis and seqtrans
forms. The existence of N- and O-silyl tautomers may be anticipated
since these tautomers are observed in silyl amides. One or both of

the seqcis O-silyl isomers (XXV and XXVI) should interchange rapidly

with the seqcis N-silyl form (XXIX) in presence of l, S- R Si migration,

3

but the remaining isomers should be stereochemically rigid.

In $31.01

I I 1.941.”.64

'1
“TWA

 

 

Figure 11.

62

Possible isomers of silyl B-ketoamines.

 

63

XXVII

XXIX

o-c /CH3
(CH ) 51/ 30.11
5 3 /
/N=C\
R CH
3
XXVI
0 c/CH3 CH
/
((113)351’ flea-c§ 3
H N—R
XXVIII
R
| /CH:5
(CH3) 351 \lc— c\\
H 0
XXX

Figure 11

II. EXPERIMENTAL

A. General Techniques

 

All general methods of purifying diketones and solvents and drying
of glassware were analogous to those described in Part I and II of this
thesis. Aniline was obtained from Fisher Scientific Company, m - chloro—
aniline from Matheson, Coleman and Bell, and p_— fluoroaniline from

Aldrich Chemical Company. The aniline and aniline derivatives were

used without further purification.

B. Synthesis of 4-Anilino-3-pentene-2-one

 

4-Ani1ino-3-pentene-2-one was prepared by using the general method
of Everett and Holm.74 Acetylacetone (30g, 0.30 moles) and aniline
(42g, 0.45 moles) were placed in a 100ml round bottom flask and heated
to ca, 100°C for 5 hours. The solution was then cooled and extracted
with three 50ml portions of ether. The ether solution was dried over
anhydrous sodium sulfate and then filtered through a glass frit.
Cooling the solution in dry ice produced 4-ani1ino-3-pentene-2-one.
The product was recrystallized three times from ether and then vacuum
dried, giving 40.7g (77.5% yield) of off-white crystals with a melting

75,76

point of 47°-48°C. Literature melting point, 47°-48°C.

C. Synthesis of 4-m-Chloroanilino-3—pentene-2-one and 4-p-Fluoro-

 

anilino-3-pentene-2-one

 

4-m-Chloroani1ino-3-pentene-2-one and 4ep-fluoroanilino—3-pentene-
2-one were prepared by the method described above for 4-anilino-3-

pentene-Z-one. A 50% mole excess of the primary amine was added directly

64

65

to the B-diketone, and the mixture was heated to £2: 100°C for several
hours. The products were then extracted and crystallized from ether or
hexane. The 4-m—chloroanilino-3-pentene-2-onethus prepared had a melting
point of 38°-39.S°C and the 4-p-f1uoroanilino-3-pentene-2-one melted at
48°-49.5°C. The literature value of the melting point of 47m-chloro-

anilino-3-pentene—2-one is 42°.79 The purity of 4-pefluoroanilino-3-

pentene-Z-one was verified by pmr spectroscopyifi
/

D. Synthesis and Characterization of Trimethyl(4-anilino-3-pentene-2—

 

one)silane

 

There are no previously reported preparations of silyl B-ketoamines.
4-Anilino-3-pentene-2-one (20g, 0.11 moles) was dissolved in SE: 400ml
of dry hexane and placed in a 500ml three necked round bottom flask
with a gas inlet tube and condenser attached. A gas outlet was attached
to the condenser and then to a mineral oil bubbler by means of tygon
tubing. The system was purged with dry nitrogen for SE: 15 minutes, and
then freshly cut potassium metal (4.4g, 0.11 moles) in small pieces was
added to the reaction flask. The mixture was stirred for SE: 12 hours
while passing a slow stream of dry nitrogen through the system. At the
end of 12 hours a very viscous mixture of hexane and potassium salt of
the B-ketoamine had formed. Chlorotrimethylsilane (13.0g, 0.12 moles)
was dissolved in 50ml of dry hexane and placed in a dropping funnel
(operations performed in dry box), which was then connected to the
round bottom flask replacing the glass stopper. The Chlorotrimethyl-
silane was then added dropwise to the hexane-potassium salt mixture.
After addition was complete the mixture was heated to SE: 50°C and

allowed to stir for 24 hours. The mixture was transferred in a nitrogen

66

filled glove bag into stOppered centrifuge tubes, and the finely divided
potassium chloride precipitate was removed by centrifugation. The
supernate was then placed in a flask on the spinning band distillation
column and the hexane was distilled off. A vacuum of 0.5 torr was applied
to the column and the product was vacuum distilled. 4-Anilino-3-pentene-
2-one was collected in four fractions boiling from 90° to 97°C at 0.5
torr. The purest fraction as determined by pmr was collected at a
temperature of 96°-97°C and was sent to Galbraith Laboratories for
chemical analysis.

Anal. Calculated for SiC H NO: Si, 11.35; C, 67.96; H, 8.56;

12 21
N, 5.66. Found: Si, 11.14, C, 67.78; H, 8.55; N, 5.47.

E. Synthesis of Trimethyl(4-m-chloroanilino-3jpentene-2-one)silane and

Trimethyl(4-p-f1uoroanilino-Sfpentene-Z-one)silane

 

Trimethyl(4tmfchloroani1ino-3—pentene-2-one)silane and trimethyl-
(4-p-f1uoroanilino-3-pentene-2—one)silane were both prepared using the
method described above fer trimethyl(4-anilino-3-pentene-2-one)silane.
They are both pale yellow liquids which boil at 113°C at 0.5 torr and
83°C at 0.1 torr, respectively. Both compounds hydrolyze readily,
precluding good chemical analyses. The fraction of hydrolysis products
was observed to increase with increasing ageing of the products in

sealed nmr tubes.

III. RESULTS AND DISCUSSION

A. Preparation and Properties of Silyl B-Ketoamines

 

The silyl B-ketoamines prepared in this study represent a new
class of silicon compounds. While B-ketoamines and their complexes with
many metal ions have been widely studied (for a survey see ref. 78) there
are no previous reports of silicon B-ketoamine complexes. These compounds
were formed quite easily with the aniline derivatives of the B-ketoamines,
by simply making the potassium salt with potassium metal (eq. 13) and then
allowing this to react with Chlorotrimethylsilane (eq. 14). However,
when a hydrogen or methyl group was bonded to the nitrogen in place of

the phenyl group in the aniline derivatives the reactions would not go.

O'TH‘N’R 0 N’R
K I M _—) K+ ~®’ + 1””2 (13)
R R R
0 N’R
+ ‘1':
K ~-I + (CH3)3SiC1 -—-? KCl + silyl B-ketoamines

(14)

The derivatives of anilino—, mfchloroanilino, and p-fluoroanilino-

3—pentene-2-one, abbreviated (CH3)35i(papo),* (CH3)3Si(m-capo) and

(CH3)3SiQp-fapo), respectively, are all high boiling pale yellow liquids.

 

I For a list of the ligand abbreviations used see Appendix A.

67

68

A11 hydrolyzed easily with the m-chloro and p-fluoroaniline derivatives
being somewhat more susceptible to hydrolysis than the unsubstituted
aniline derivative. All the silyl B-ketoamines were stable and miscible

in common organic solvents such as methylene chloride and chlorobenzene.

B. Characterization of Silyl B-Ketoamines

 

Trimethyl(4-anilino-3—pentene-2-one)silane, trimethyl(4-mgchloro-
anilino-3-pentene-2-one)silane and trimethyl(4jp—fluoroanilino-3-pentene-
2-one)silane all have very similar pmr spectra (see Figure 12).

Figure 11 shows the possible isomers that might be expected. Each isomer
should give rise to single silyl methyl and ligand C-H lines and two
ligand methyl lines as a result of its unsymmetric nature. It is
apparent then from the eight ligand methyl lines present that there are
present only four of the six possible isomers one might expect. It is
interesting to note that even the relative concentrations of isomers
appears to be approximately the same for all three compounds. Cooling
the solution of (CH3)3Si(papo) to ~64° did not resolve any new pmr

lines or reveal any new isomers. The high temperature spectra of
(CH3)3Si(papo) and (CH3)SSi(p-fapo) were recorded and they are almost
identical. Figures 13 and 14 show the temperature dependent pmr spectra of
(CH3)SSi(papo). The four isomers have been labeled A, B, C and D.

Ligand -CH= assignments were made on the basis of spin decoupling
experiments and silyl methyl lines were assigned on the basis of relative
intensities (graphical integration). It can be seen that isomers A and

B become exchange averaged first and then isomers C and D coalesce,

but even at 179°C isomers A and B do not interchange with isomers C

and D. Further heating causes irreversible decomposition of the silyl

B-ketoamines.

 

 

 

 

 

 

 
 

 

 

 

Figure 12.

69

Ambient pmr spectra of:

(CH3)3Si(m-capo); (3), (CH3)3Si(pffapo).
SE: 20g/100ml chlorobenzene.

(1). (CH3)3Si(papo); (2).

Concentrations

 

 

Figure 13.

71

Ligand methyl region of temperature dependent pmr spectra
of (CH3)3Si(papo) in chlorobenzene; concentration is

£23 SOg/lOOml solvent; 60MHz. The lines marked "x" are due
to a small amount of free ligand impurity.

72

Ma ohamwm
w

nnfi on.“ on."
b L h

 

I’D"

3.. on; an; n... n...

h b P h L

o o a < a > 44 < J. ..: u

g .3 ‘
.3. .

 

 

 

 

a .43

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 14.

73

Silyl methyl and ligand =CH— region of temperature
dependent pmr spectra of (CH3) Si(papo) in chlorobenzene;
concentration is $2: 50g/100ml solvent; 60MHz. The lines
marked "x" and "H" are due to a small amount of free ligand
and a hydrolysis product, respectively. (Note: The silyl
methyl region at 42°C was run at half the sweep width of
the other spectra.)

3 6.5»: a
86.. :6 26

I80

.516 u
. . own 36
h P

 

 

 

;
,: ..
.:
._,_
.4
:
.M ..._
1‘.
.113/ _\II)\II

 

 

 

 

 

1571.1 ..P,.,I‘. x... I. 0.:

75

The infrared spectrum of (CH3)SSi(p-fapo) (see Figure 14) exhibits
an intense absorption at 1030 cm.1 consistent with the Si-O stretch
observed in the silyl enol ethers.58 In addition a broad intense
absorption occurs in the 1570—1630 cm"1 which is the region expected

for both C=CS8

and C=N67 stretching vibrations. Furthermore, there is
no absorption in the 1700 cm-1 region where the stretching frequency of
an uncoordinated carbonyl group should be found.58 All of this indicates

that the four isomers present in these compounds are the four oxygen

bonded isomers XXV-XXVIII.

 

 

 

76

 

Figure 15. Infrared spectrum of neat (CH )Ssi(papo) between CsI salt
plates. The 1601.8 and 1028. cm'1 absorptions of poly-
styrene were used as references.

 

 

77

 

com

 

 

 

 

 

 

 

coop

 

 

 

 

oou. 75.0

—

 

 

 

 

mH «Aswan

00V—

 

 

coo—

 

 

 

 

 

 

comp

IV. CONCLUSIONS

The pmr data strongly indicate that the silyl B-ketoamines exist
exclusively as oxygen silylated isomers XXV-XXVIII. The line broadening
observed in the variable temperature pmr spectra is attributed to rapid
gynfgnti_isomerization about the C=NR bond and occurs in both the Seqcis
and segtrans forms which leads to XXV 2 XXVI and XVII Z XXVIII averaging.
The only other possible explanation of the variable temperature pmr data
must be based on the assumption that both oxygen and nitrogen silyl
tautomers are present but that Syg;§§£i isomerization in the O-bonded
forms is fast even at -64°C. This would account for the apparent
existence of four isomers. The two high temperature pmr broadening
phenomena could then result from seqcis t seqtrans isomerization about
the C=C bond in the oxygen and nitrogen bonded tautomers. This latter
interpretation is improbable, because it requires rotation about the

C=C bond to be more facile than a l, 5 migration of the R Si group. In

3

the silyl enol ethers of B-diketones it has been shown that 1, 5 RSSi

group migration is much more facile than seqcis-seqtrans isomerization.

 

Using space filling models it is possible to make tentative
assignments of some of the pmr lines to specific isomers. The models
clearly show that only one isomer, XXV, places the -Si(CH3)3 group
directly over the diamagnetic cone of the phenyl group, thus shifting

it well up field of the other silyl methyl lines. This means then that

if isomer XXV gives rise to the lines marked A in Figures 13 and 14 then
isomer XXVI must be responsible for the lines marked B. The ring current

of the phenyl group in isomer XXVII is also responsible for shifting the

78

79

ligand methyl adjacent to oxygen to higher field and the CH -C=NR methyl

3
to lower field relative to the other isomers. (Thus, isomer XXVII
produces the lines marked C and isomer XXVIII produces the lines marked
D in the spectra of Figures 13 and 14. The one area of uncertainty is in the
assignment of the C and D silyl methyl lines. The chemical shifts and
relative intensities of these lines are too similar to make positive
assignments.

Based on the above pmr line assignments, the lower temperature pmr

interchange process is assigned to XXV 1 XXVI isomerization and the

higher temperature process involves XXVII # XXVIII interchange:

Lower temperature process

 

 

(CH ) SiO-C"CH3 {2:22 (CH ) SiO-—C”CH3
3 3 as 3 3 1§
R C-H C-H
‘ =C< ,N=c<
CH3 R CH3
seqcis-syn (XXV) seqtrans—anti (XXVI)
Higher temperature process
R
CH) 30 c’CH3 1113:! CH 50 c’CH3 N R
( 3 3 1 s ¢ ( 3)3 1 ‘- 5 /
/ - C-C
H \CH 11’ CH
3 3
seqtrans-syn (XXVII) seqtrans-anti (XXVIII)

 

 

The onset of line broadening fer the lower temperature process
occurs at £3: 61°C, whereas for the higher temperature process line
broadening begins at £2: 92°C. Thus gyn:anti_isomerization is appreciably
more facile for the seqcis than for the seqtrans form. The difference

in rates may be due to differences in steric factors. Molecular models

l «I.

.. ....£;l....1l.sz drumming

... .. I

if 1.... .llllanh....141.rr. .d
. 0 . I .

 

 

QIELIIHD .lul‘flvEB. a N. ‘0 I. I. A. m.
1

80

indicate that in the seqcis isomers the §y2_conformation should be
destabilized by steric factors relative to the an£i_conformation. 0n
the other hand steric factors do not appear to influence the relative
stabilities of the syn:anti_conformations in the seqtrans form. Thus,
whether the mechanism for gyn:anti_isomerization involves tortional
motion about the C=NR bond or a "lateral shift" analogous to inversion
about nitrogen in amines,79"82 the larger rate for the seqcis form
could be due to destabilization of the §y2_conformation.

Another possible factor facilitating the synfanti_interchange in
the seqcis form is formation of a nitrogen bonded intermediate via a

l, 5 R Si group migration:

3
' CH 7

/CH3 I 3 /CH3

(CH3)3S1----O--C\.§ O==C\\ (CH3)SSi-O-C§§

-H (=3 C-H (:2 C-H (15)

R\ / || /
N=C\CH (CH3)351\ ,C\ R,N=<:\CH
3 /N CH:5 3

R

seqcis-syn (XXV) ,_ J seqcis-ati (XXVI)

 

 

 

 

In order for this to be a viable mechanism, the intermediate would have
to possess appreciable dfl-p1t Si=N character as illustrated in the

resonance ferm XXIX. This resonance form allows the Si=N-R plane to

CH
_ C=C: 3
,N-C\
R CH3

XXIX

 

 

81

be perpendicular to the plane of the C=C bond. A l, 5 migration of the
R38i group by internalnuc1e0philic attack of the C=O group would then

allow the developing RN=C bond to achieve either a syn or anti con-

figuration.

BIBLIOGRAPHY

10.

11.

12.
13.
14.
15.
16.

17.

18.

19.

F.

BIBLIOGRAPHY

Basolo and R. G. Pearson, "Mechanisms of Inorganic Reactions,"

2nd ed., Wiley, New York, N.Y., 1967, pp. 300—334.

J.

N.

L.

T.

1,

J. Fortman and R. E. Sievers, Coord. Chem. Rev., 6, 331(1971).

 

Serpone and D. G. Bickley, Prog. Inorg, Chem., 11, 391(1972).

 

. Gach, Monatsch., 21, 98(1900).

. F. Morgan and R. A. S. Castell, J. Chem. Soc., 3252(1928).

 

 

. Rosenheim and A. Bertheim, Z. Anorg. Chem., 34, 427(1903).

. W. Moore and M. L. Larson, Inorg. Chem., 6, 998(1967).

 

. W. Moore and R. E. Rice, Inorg. Chem., 1, 2510(1968).

 

O. Atovmyan and Y. A. Sokolova, Chem. Commun., 649(1969).

 

J. Pinnavaia and W. R. Clements, Inorg. Nucl. Chem. Letters,
1127(1971).

 

B. M. Craven, K. C. Ramey and W. B. Wise, Inorg. Chem., 19,

262
M.
H.
M.
F.
L.
E.
630
A.
J.

Nuc
N.

 

6(1971).

 

L. Larson and F. W. Moore, Inorg, Chem., S, 801(1966).

Gutowsky and C. Holm, J. Chem. Phys., 2%, 1228(1956).

 

Rogers and J. Woodbrey, J. Phys. Chem., 66, 540(1962).

 

Block, Phys. Rev., 70, 460(1946).

H. Piette and W. A. Anderson, J. Chem. Phys., 30, 899(1959).

 

Grunwald, A. Loewenstein, and S. Meiboom, J. Chem. Phys., 27,
(1957).

 

Loewenstein and S. Meiboom, ibid., 27, 1067(1957).

A. Pople, W. G. Schneide, and H. J. Bernstein, "High-resolution
lear Magnetic Resonance," McGraw-Hill Book Co., New York,
Y., 1959.

82

20.

21.
22.

23.

24.
25.
26.
27.

28.

29.

30.

31.

32.

33.
34.

35.

36.

37.
38.
39.

40.

83

T. J. Pinnavaia, L. J. Matienzo and Y. A. Peters, Inorg: Chem.,
2’ 993(1970).

 

D. A. Case and T. J. Pinnavaia, Inorg. Chem., 10, 482(1971).

 

R. C. Fay and F. S. Piper, Inorg. Chem., 3, 348(1964).

 

J. R. Hutchison, J. G. Gordon and R. H. Holm, ibid., IQ,
1004(1971).

R. C. Fay and R. N. Lowry, ibid., Q, 1512(1967).

D. C. Bradley and C. E. Holloway, J. Chem. Soc._(A), 282(1969).

 

T. J. Pinnavaia and R. C. Fay, Inorg. Chem., 1, 502(1968).

 

J. Adams and C. Hauser, J. Amer. Chem. Soc., 4S, 1220(1944).

 

 

(3)9 A. L. VanGeet, Anal. Chem., gg, 679(1970).
(b) ibid., gg, 2227(1968).

J. J. Howe, Ph.D. Thesis, Michigan State University, East Lansing,
Michigan, 1971.

G. M. Whitesides and J. S. Fleming, J. Amer. Chem. Soc., mg’
2855(1967); J. B. Lisle, S. B. Thesis, Massachusetts Institute

of Technology, 1968.

 

H. Gehrke and J. Veal, Inorg. Chim. Acta., 3, 623(1969).

 

B. Soptrajanov, A. Nikolovski, and J. Petrov, Spectrochim. Acta.,
24A, 1617(1968).

 

C. Su, J. W. Reed and E. S. Gould, Inorg. Chem., 12, 337(1973).

 

M. M. Jones, J. Amer. Chem. Soc., 81, 3188(1959).

 

W. C. Fernelius, K. Terada and B. E. Bryant, Inorg. Chem., 6,
147(1960).

0. P. Afanasev, A. N. Bantysh and D. A. Knyazev, Russ. J. Inorg.
Chem., 13, 182(1968).

 

J. Bernal, Chem. and Ind., 1343(1966).

 

J. P. McKaveney and H. Freiser, Anal. Chem., S, 801(1966).

 

D. Grdenic and B. Korpar-Colig, Proc. Chem. Soc., 308(1963).

 

R. D. Archer, ARL 63-160(1963).

41.

42.

43.

44.

45.

46.

47.

48.
49.

50.

51.

52.

53.

54.

55.

56.

57.

58.
59.

60.

61.

84

F. A. Cotton and G. Wilkinson, "Advanced Inorganic Chemistry,"
John Wiley 6 Sons, Inc., New York, N. Y., 1972, pp. 944-972.

J. G. Gordon and R. H. Holm, J. Amer. Chem. Soc., 92, 5319(1970).

 

J. R. Hutchison, J. G. Gordon, and R. H. Holm, Inorg. Chem., 10,
1004(1971).

 

S. S. Eaton, J. R. Hutchison, R. H. Holm and E. L. Muetterties,
J. Amer. Chem. Soc., 94, 6411(1972).

 

S. S. Eaton, G. R. Eaton, R. H. Holm and E. L. Muetterties,
J. Amer. Chem. Soc., 9S, 1116(1973).

 

B. Jurado and C. S. Springer, Chem. Commun., 85(1971).

 

A. Allerhand, H. S. Gutowsky, J. Jonas, and R. A. Meinzer,
J. Amer. Chem. Soc., 88, 3185(1966).

 

K. J. Dahlquist and S. Forsen, J. Phys. Chem., 69, 4062(1965).

 

R. W. Jones and R. C. Fay, Inorg. Chem., 12, 2599(1973).

 

B. S. Gould, R. R. Hiatt and K. C. Irwin, J. Amer. Chem. Soc.,
90, 4573(1968).

 

L. Que and L. H. Pignolet, Inorg. Chem., 13, 351(1974).

 

L. H. Pignolet, R. A. Lewis, and R. H. Holm, J. Amer. Chem. Soc.,
g3, 360(1971).

 

A. E. Pierce, "Silylation of Organic Compounds," Pierce Chemical Co.,
Rockford, Illinois, 1968.

J. F. Klebe, H. Finkbeiner, and D. M. White, J. Amer. Chem. Soc.,
88, 3390(1966).

 

J. F. Klebe, Accounts Chem. Res., 3, 299(1970).

 

J. E. Blackwood, C. L. Gladys, K. L. Loening, A. E. Petrarca and
J. E. Rush, J. Amer. Chem. 533., 90, 509(1968).

 

T. J. Pinnavaia, W. T. Collins, and J. J. Howe, J. Amer. Chem. Soc.,
92, 4544(1970).

 

 

R. West, J. Amer. Chem. Soc., 80, 3246(1958).

J. J. Howe and T. J. Pinnavaia, J. Amer. Chem. Soc., 91, 5378(1969).

 

J. K. Kusnezowa, K. Ruhlmann, and E. Grundemann, J. Organometal.
Chem., 41, 53(1973).

 

L. H. Sommer, "Stereochemistry, Mechanism and Silicon," McGraw-Hill,
New York, N.Y., 1965.

62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.

78.

79.

80.
81.

82.

E.

R.

85

. Shanan-Atidi and Y. Shvo, Tetrahedron Lett., 7, 603(1971).

 

. J. Reich and D. A. Murcia, J. Amer. Chem. Soc., fig, 3418(1973).

 

. N. Kuo, F. Chen, and C. Ainsworth, J. Chem. Soc;, D, 137(1971).

 

. A. O'Brien and C. P. Hrung, J. Organometal. Chem., 27, 185(1971).

 

. Wiberg and H. J. Pracht, Chem. Ber., 105, 1388(1972).
. J. Scherer and P. Hernig, ibid., 101, 2533(1968).

. West and B. Bechlmeir, J. Amer. Chem. Soc., 94, 1649(1972).

 

. Fukui, K. Itoh, and Y. Ishii, J. Chem. Soc., PII, 1043(1972).

 

. Pump and E. G. Rochow, Chem. Ber., 97, 627(1964).

. Itoh, M. Katsuda, and Y. Ishii, J. Chem. Soc., B, 302(1970).

 

. N. Kuo, F. Chen, and C. Ainsworth, J. Chem. Soc., D, 137(1971).

 

. J. Pinnavaia and J. A. McClarin, J. Amer. Chem. Soc., 96, 3012(1974).

 

. W. Everett and R. H. Holm, ibid., 81, 2117(1965).
. F. Holtzclaw, J. P. Collman, and R. M. Alire, ibid., 89, 1100(1958).

. Combes and C. Combes, Bull. Soc. Chim. Fr., 7, 778(1892).

 

Roberts and E. E. Furner, J. Chem. Soc., 1832(1927).

 

H. Holm, G. W. Everett, and A. Chakravorty, Prog. Inorg. Chem.,

 

z, 83(1966).

G.

E. Hall, W. J. Middleton, and J. D. Roberts, J. Amer. Chem.

 

Soc., 93, 4778(1971).

W.

C.

D.

G. Herkstroeter, ibid., 95, 8686(1973).
K. Sauers and H. M. Relles, ibid., 9S, 7731(1973).

Y. Curtin, E. J. Grubbs, and C. G. McCarty, ibid., 88, 2775(1966).

 

,.4 .3,....6l14!

 

APPENDIX
A

Abbreviation

 

acac
aibm
bibm
dibm
dpm

pibm
pvac

P3P°

m-capo

pffapo

Ligand Abbreviations

86

Acetylacetone
Acetylisobutyrylmethane
Benzoylisobutyrylmethane
Diisobutyrylmethane
Dipivaloylmethane
Pivaloylisobutyrylmethane
Pivaloylacetone

4-Phenylamino-3-pentene-
2-one

4-erhloroanilino-3-pentene-
2-one

4-pfFluoroanilino-3-pentene-
2-one

APPENDIX
B

 

 

 

 

Analysis of Multisite Dynamic NMR Spectra

by

Richard Teets and T. J. Pinnavaia

Anderson-Kubo-Sack Theories

 

According to the theories of Anderson,1 Kubo,2 and Sack,3 rapid
exchange NMR spectra can be calculated in the fellowing manner.
Suppose there are N magnetically inequivalent groups which give rise
to N lines in a slow exchange NMR spectrum. Let the resonance fre-
quencies and relaxation times for these N groups be designated as W1
and T2.. Let Kij be the probability per unit time to go from group i

1
to group j. And let Kii be chosen so that

Finally EQi is the relative population proportion for group i. This
includes both equilibrium concentrations and number of equivalent groups.
Suppose, for example, group 1 is a methyl group and group 2 is a
methylene group. Then EQ1 = 3 and E02 = 2. Suppose further that the
isomer containing the methyl group is present at twice the concentration
of the isomer containing the methylene group. Then EQ1 = 6 and EQ1 = 2.
(Note that fer low temperatures, EQi will give the relative heights of
the N lines in the spectrum). With the definitions given above, the
intensity, 1(w), as a function of frequency, may be calculated by

the following equation.

87

 

 

88

' . 1 I ‘1 1')
(W91)1 ' T2 " K111(12 K1N 1
X 1 - K 1
21 2N
I(w) = -Real [BQ1,EQ2,...EQN] K31 l
. l 1
L KNl'” (“"th ' 'i'zN + KNNJ N

 

 

 

 

In the above equation, the i = VI: and "Real" means take the real part
of the above complex expression.
The above expression may be simplified by the following definitions:
Let EQ equal a l by N row matrix with elements EQJ as defined
above. Let WR be a diagonal matrix (of order N by N) where the diagonal
elements are the W defined above. Let T2 be an N by N diagonal matrix

J

with the diagonal elements equal to 1/T2 as defined abOve. Let K be a

J
matrix with elements equal to the Kij described above. Note that the
diagonal elements are fixed so that the rows sum to zero. Finally, let
ID be the N by N identity matrix (the diagonal elements are ones) and

let [1] be the N by 1 column matrix with each element a 1. With the

previous definitions the Intensity can be written as

1(6) = -Real (EQ°[w-ID-WR)i - T2 + K]-1-[l]) (1)

In practice, it is convenient to rewrite K as (1/T)K' where K' is the
part of the matrix which depends only on the assumed mechanism and r is
a scalar representing the temperature dependence of the rearrangement
process. In other words, I is the mean lifetime before exchange which

is temperature dependent.

89

The above formula can be illustrated by an example due to

Whitesides.4 Consider:

The copper shifts around the ring, thus averaging the signals from the
protons in environments 1,2,3. Since environments 2 and 3 have two
protons each, while environment 1 has only one proton, the matrix
EQ = [1,2,2].

The resonant frequencies for each of the three types of proton
environments can be determined from a low temperature spectrum. Call

these frequencies w1,w2,w3 for groups 1,2, and 3 respectively.

ml 0 0
Then WR = 0 m2 0
0 0 “3

The T2 values were determined from the low temperature line widths. A

value of 0.034 seconds was used for each group. So the matrix T2 is

1/.034 0

T2 = 0 1/.034 0

0 1/.034
J

 

 

 

 

 

 

 

 

90

The kinetic matrix will depend on the mechanism chosen and on the
temperature at which the corresponding experimental spectrum was run.
Let r by the mean lifetime before a copper shift. Then l/r is the
rate of shifting from one environment to another. Now consider a
specific mechanism, a 1-3 shift with 1-2 shifts forbidden. Then the

probability to go from environment 1 to environment 2 is 0 so K 0.

1,2 =

Each time the copper shifts, the environment 1 will go to an environ-

ment 3 so K = 1/1. Since the row must sum to zero, K = -l/r.

1,3 1,1
Now, half of the time, the copper will shift to one of the "3" environ-
ments and the other half of the time, the copper will shift to the

other environment. When the 1-3 shift occurs, the other 3 will become

a 2. Therefore, K3,2 = 0.5/1 and K3,1 = 0.5/r so K3,3 = -1/T. Using
the same logic we find that
1 r
1 -1/T 0 l/T -l 0 1
K = 2 0 -0.5/T 0.5/r = l/T 0 -0.5 0.5
3 0.5/T 0.5/T -1/r ‘ 0.5 0.5 -1

 

 

Combining all the above matrices according to equation (1) we have:

(In-011M - 71337 - llt 0 ‘ 1/1
100) - -Real [1,2,2]- 0 (b-w2)1 - 1%;- - .5/1 0.5/1 1

 

 

 

 

l .
0.5/1’ 0.5/1' (u-w3)i - .—037 - l/‘r lj

 

91

(Note: the above matrix may appear different than that used by Whitesides.
The differences is merely one of notation. He used the slightly con-
fusing notations that is indicated on the left below, rather than the

more straightforward approach used above and below on the right.)

 

 

 

 

2 3 1 1 2 3
2 F 2 1 r '7 where 1,2,3 indicates
nonequivalent en-
3 2 vironments
1 3
L J L _
Whitesides' matrix This paper's matrix

Prggram KINET

 

Program KINET uses equation (1) to calculate the intensity as a
function of frequency. I(m) is calculated for evenly spaced frequencies
from a specified low frequency to a specified high frequency. These
(I(w), w) pairs are then plotted to give the calculated spectrum. The
user must therefore include in his input the parameters in equation (1)
plus various plot parameters.

The data must be entered following the format specified in Table 1.
In brief, the user must have four sets of data cards. The first set
contains only one card, which specifies the low and high frequency
limits, the number of (1(w), w) pairs to calculate, and the number of
resonating groups, N. There are N cards in the second card set, one for
each group, which specify the resonance frequencies (WR), the equilibrium
populations (EQ), and the linewidths which are used to calculate 1/T2 by

the relation l/T2 = n-(line width). (This assumes a lorentzian line

 

 

92

 

Table I
Card Contents and Explanations* Card Columns
A Format (12, 2F7.2, l4, F5.2)
(1 card) Size of Matrix (N): 1-2

This is the number of rows in the kinetic
matrix. It should be a two place integer
less than 20.

Low Frequency: 3-9
This is the frequency which corresponds to (Decimal)
the leftmost end of the plot. It is a 7
place decimal number, including the decimal
point and an optional + or — sign. Input
should be in Hz.

High Frequency: 10-16
This is the frequency which corresponds to (Decimal)
the rightmost end of the plot. Like the
low frequency, it is a 7 place decimal
number.

Number of calculated points (M): 17-20
This instructs the computer to calculate (Integer)

 

the intensity at M evenly spaced fre-
quencies between the low frequency and the
high frequency. These M points are then
plotted to give the theoretical spectrum.
This should be a feur place integer less

than or equal to 500. Since the computer

 

Card

93

Table I (cont'd.)

Contents and Explanations°

Card Columns

 

A cont.

B

(N cards)

 

calculation time is proportional to the

number of calculated points, M should be

kept to a minimum, allowing for reasonable

resolution on plots.

Width of plot in cm:
This should be a 5 place decimal number.
The maximum width of the plotter is 30

inches so the total width of all plots on

one line (including the 1 cm between plots)

must be less than 76 cm.
Format (F6.3, 2F7.2)

Each of the N cards contains the following

information for one of the N groups (1 group

per card)

Equilibrium Population Proportion:

This is a 6 place decimal number specifying

the relative amount of the group present.
It includes both concentration and the

number of equivalent absorbing nuclei.

This number is only relative to the popula-

tion of the other groups.

Resonant Frequency:

This is a 7 place decimal number specifying

the resonant frequency of the group (in Hz)

 

1-6

(Decimal)

7-13

(Decimal)

94

Table I (cont'd.)

 

Card Contents and Explanations Card Columns
B cont. in the absence of exchange. This may be
calculated by extrapolating the chemical
shift from the low temperature slow
exchange limit into the temperature region
of interest. The important thing here is
the difference in frequencies between
groups. Thus the frequencies entered can
be relative to any standard such as TMS.
Negative frequencies are therefore
acceptable.
Line Width: 14-20
This is the line width at half maximum (Decimal)
specified in Hz. It is a 7 place decimal
number. This is a relatively uncritical
parameter which can be determined from low
temperature spectra, and the best fit of
the curves to experimental spectra.
C Format (10F8.4)
2-N Cards This is the kinetic matrix, not including 1-8, 9-16
the mean lifetime parameter (r)** The 17-24 etc.
matrix is read one row per two cards. Each (Decimal)

 

matrix element is read as an 8 place
decimal number. After the N elements on

the two cards, the rest of the two cards is

 

 

95

Table I.(Cont'd.)

Card Contents and Explanations Card Columns

 

C cont. left blank. Thus a four by four matrix
would have 4 numbers on the first half of
the first card, the second half of this
card would be blank, the second card would
be blank, the third card would contain the

4 elements of the second row of the matrix,

 

the fourth card would be blank, and so on.
Regardless of the number read in for a
diagonal element, the computer will set the
diagonal = - (sum of other row elements) so

the diagonal may be left blank.

D Format (E8.6, F5.2, Il)
Tau (T)! 1-8
This is the mean lifetime before exchange. (Decimal - see
Each element in the Kinetic Matrix is text)
devided by T to give the matrix K in equa-
tion (1). The value of Tau should decrease
with increases in the temperature at which
the spectrum was run. Tau is in seconds.
This is an 8 place decimal number with an
option for a power of ten factor. The
number entered must be of the form

:_NNN.NNE :_MM. The first part (NNN.NNN)

 

 

Card

96

Table I (Cont'd.)

Contents and Explanations

Card Columns

 

D cont.

 

is multiplied by 10 raised to the :.MM
power. Thus 3 x 10';4 second could be
entered as 0.000300 or as 003.0E-4. The
total quantity entered must take 8 spaces,
blanks may be used in front of (to the left
of the number).

If Tau is positive, the computer will pro-
ceed to plot the spectrum using the other
data read in above. When the plot is
finished, the computer will read another
value of Tau. If this value is positive, a
new plot will be made using this value of
Tau and the other data read previously.

If Tau is zero (the card is blank) the pro-
gram will end.

If Tau is equal to -l.0, the computer will
read in a new set of resonant frequencies
(card set B), but use the previous Kinetic
matrix. The next card after set B should
be a new Tau value. If Tau is -2.0, the
computer will read in a new Kinetic matrix
(card set C), but not a new set of resonant
frequencies. The next card after the

matrix should be a new value of Tau.

 

 

 

 

 

97

Table I (Cont'd.)

Card Contents and Explanations Card Columns

 

If Tau is -0.5, the computer will restart
completely and read in a new card A, new
card set B, new card set C, and then a new
value of Tau.

Height of Plot in cm: 9-13
The plot will be scaled so that the highest (Decimal)
peak will be the specified height. This

should be a 5 place decimal number less

than 25 cm.
Plot Option: 14
This controls where the plot is made, with (Integer)

respect to the previous plot.
The first plot will be in the lower left
corner of the paper. For the rest of the
plots, the position of the plot is
determined according to the following:
Option 0 Start this plot at the left
side of the paper, 2cm above
the highest previous plot.
Option 1 Plot over the previous plot.
Option 2 Start plot 1cm to the right of
the previous plot. (Make sure
that the total length including

space between plots is less

 

 

than 75cm per line.)

Card

98

Table I (Cont'd.)

Contents and Explanations

Card Columns

 

D cont.

The total height of all plots including the

two cm between plots should not exceed

about 300cm;

Example

A-first plot-any option
B-second plot-Option 2
C-third plot-Option l

D-fourth plot-option 0

 

 

*Footnotes for Table I.

1.

2.

Blanks are the same as zeros.

The entire number including + or -
signs (+ sign optional) and the decimal
point should fit within the column
limits given in Table 1.

Numbers designated as integers must not
contain decimal points. They must be
right justified: the rightmost digit
should appear in the rightmost column
designated for that number. (This is
due to the computer interpreting blanks
as zeros. For example, the number N on
card set A is an integer which occupies
columns 1 and 2. If N=3, a 3 should be
punched in column 2 and a blank in
column 1.)

 

99

Table I (Cont'd.)

1
Card Contents and Explanations Card Columns

 

4. Numbers designated as decimals should
always contain a decimal point.

** Note on Kinetic matrix: the rows of the
kinetic matrix should be entered in the

same order as the cards for the resonance
frequencies.

 

 

 

100

shape.)5 The frequencies and line widths are in Hz. The third card
set contains 2 x N cards, 2 for each row of the kinetic matrix K'.

The fourth card set contains one card which specifies the mean lifetime
before exchange (T) in seconds. This is used to find K = l/r K'.

After these 4 sets are read in, the computer will plot the spectrum and

then read another value of tau.

Dealingwith Spin-Spin Coupled Systems

 

Whenever possible, chemical methods (such as use of deuterium)
should be used to avoid spin-spin coupling. However, for loosely
coupled systems split into doublets where the resonant frequencies and
splitting constants can be accurately determined, the following approach
can be used.6 The spectrum for N doublets each with the same value of
J can be thought of as a superposition of 2 identical spectra containing
N singlets, but with the two spectra offset by J Hz. Due to second order
splitting effects, the upfield spectrum should be multiplied by an
empirically determined weighting factor so that the doublets have the
correct height relative to each other.

Very minor modifications to the program are necessary to allow

treatment of the above case. Instructions for making the changes are

given in the comments in the program near statement 410 and statement 379.

Error Messages

 

Inversion of matrices involves many arithmetical operations which
can lead to an accumulation of round-off errors. In certain cases
these errors can be significant. The algorithm used for inversion of the
matrices should minimize the round-off errors. Also the relatively

large word of the CDC 3600 (48 bits) should vertually eliminate the

 

unrfi _

101

chance of round off errors causing erros in the computed spectra.
However, the progrm includes two safety checks. After the printout
for each Tau value, there will be a message "ERROR INDEX = ...." and
the computer will print a number times a power of ten. This number
should be less than about 10-8. (See the section "details of the
program" for the actual method of calculating the error index.) Also,
every fifth calculated point fer the spectrum, the computer calculated
another error index. If this error is less than 10-7, nothing is
printed, if it is greater than 10-7, the computer will print "Error
of .... at Freq. ...." It is not anticipated that any of these
messages will be printed.

For each plot, the computer wll print the computer calculation
time in missiseconds. This is merely for the sake of curiousity.

The computer may also print "Singular ..." and print a number
which will either be 2*N or N+l where N is the order of the matrix.
This means that the matrix does not have an inverse. This error
probably indicates that there was an error with the input data. If
the input data is correct, then there may be a hidden program error.
The program has been tested, and no errors have been feund, but it is

not possible to guarentee that the program is perfect.

Details of the Program

 

In an attempt to minimize computation time, the NMR spectrum is
calculated in a manner slightly different than implied in equation (I)
of the general background. Using the definitions of the general
background section let M = (w-ID-WR)i-T2 + K so that

1(0)) = - Real [EQ-M'l-[IJ]

I flu-V” fun ‘MI‘I‘

 

 

 

102

since 104'1 = ID it follows that MM'1[I] = ID[1] = [1]
Let AR + (AI)i = M-1[l] where AR and A1 are both real column
matrices.
Then I(w) = - Real [EQ-(AR + AIi)] = - EQ-AR

Also M(AR + AIi) = [1]
Thus to calculate the intensity, we must merely find AR. Write M out
completely, and expanding the product we have:
M(AR + Ali) = ((w-ID-WR)i - T2 + K) (AR + Ali) =

-(w-ID WR)~AI + (-T2 + K)AR + i[(m-ID-WR)-AR + (-T2 + K)-AI] = [1]

Thus -(w-ID-WR)°AI + (-T2 + K)AR = [l] (2)
and (onD-WR)°AR + (-T2 + K) AI = o (3)
Solving equation (3) for AI we get AI = - (-T2 + X)'1(o ID-WR)AR.

Putting this value for AI in equation 2 we get

[(m'ID-WR)(-T2 + K)-1(w-ID-WR) + (-T2 + K)]Ar = [I] = B'AR (4)

The whole problem resolves down to solving equation (4) for the column
matrix AR. Despite the apparent complexity of equation 4 it is easier
to use than the formula given in equation (1). The matrices (-T2 + K)
and (-T2 + K)"1 are constant and need be calculated only once for each
spectrum. Since (w~ID-WR) is a diagonal matrix the product
(w-ID-WR)(-T2 + K)'1(w-ID-WR) can be fermed efficiently. It should be
noted that AR can be calculated by triangularizing the matrix B in
equation 4, thus the matrix does not have to be completely inverted.
Finally, this approach uses only real numbers so that time is not
wasted in inefficient complex arithmetic. (Note the difference between

taking a real product and a complex product: a-c requires one

103

multiplication while (a+bi)°(c+di) = (ac—bd) + (bc+ad)i requires 4
multiplications and 2 additions.)

The quantity (-T2 + K)'1 is found by triangularizing (-T2 + K)
and then finishing the inversion. Pivotal condensation is used for all
triangularization.8

Pivotal condensation is explained by S. D. Conte in Elementary,

 

Numerical Analysis, McGraw-Hill Inc. 1965 p. 156-163. Essentially,

 

the idea is to use elementary row Operations to transform a matrix

into upper triangular form. At each step, the rows are interchanged

 

so that division by small numbers is avoided. Subscripted subscripts a
were used so that row interchange could be accomplished by merely
rearranging the subscript matrix rather than actually recopying the
rows.

Since the frequencies in equations 1-4 are in angular units, the
input frequencies are multiplied by 2-n to convert them to angular
units. The elements of the T2 matrix are l/T2 values for each group.
These are calculated from the line widths by the formula l/T2 = w-(line
width) which assumes a Lorentzion line shape.5

If any of the matrices is singular, the computer will print
"singular" and also print a number which will be 2*N if the matrix
(-T2 + K) is singular and will be N+l if the matrix B in equation 4
is singular. It is possible that if (-T2 + K) is singular, the
spectrum can be calculated using equation 1 and complex arithmetic.
However, I imagine that since these equations represent a physical
reality, the matrices will never be singular as long as the line

widths are nonzero.

104

The reliability of the inverse (-T2 + K).1 is tested by taking
(-T2 + X)'1(-T2 + K)-ID = El and (-T2 + K)(-T2 + K)-1-ID = 52. The
Error Index is the square root of the sum of the squares of each of
the elements of D1 and E2. Thus, theoretically, the Error Index
should be zero. However, due to an accumulation of round-off errors,
it will not be zero. If it is on the order of 10'8 or less, there
should be nothing to worry about, because experimental approximations
involved in the kinetic matrix will probably be much larger than 10"8

For every fifth calculation of the quantity AR in equation (4),
an error index is calculated by finding E3 = EQ°(B-AR-[l]).

The error index is found by taking the square root of the sum
of the squares of the elements of E3. If this error is greater

than 10-7, it is reported, otherwise, it is ignored.

The plotting is done using the MSU SUBROUTINE PLOT which plots

X, Y pairs on a calcomp plotter. For the first plot, the pen is moved

to the left 30 inches to ensure that it is on the left side of the

plotter paper.

105

References

Anderson, P. W., J. Phys. Soc. Japan, 2, 316(1954).

 

Kubo, R., J. Phys. Soc. Japan, 2, 935(1954).

 

Sack, Mol. Phys., 1, 163(1958).
Whitesides, G. M. and Fleming, JACS, SS! 2855(1967).

Emsley, Feeney, Sutcliffe, High Resolution NMR Spectrosc0py,
Pergamon Press, 1967, p.30.

 

Holm, Inorg. Chem., 19, 1004(1971) uses this approach on spin-spin
coupled tris octahedral chelate.

 

Saunders, M., article in:

A Ehrenburg, B. G. Malstrom, Mag. Resonance in Bio Systems.,
Pergamon Press, Oxford, 1967.

This reference contains a concise overview to computer calculation
of NMR spectra. It also includes references to work done on
closely coupled systems. These systems require about an hour of
computer time per plot. Note that Saunders' matrix which

 

corresponds to my equation 1 has a + sign in front of the l/T2 values

instead of a - sign. All other sources use the - sign.

Conte, S. D., Elementary Numerical Analysis, McGraw Hill, 1965,
p. 156-163.

 

Gutowsky, H. S., Vold, Wells, J. Chem. Phys., 4S, 4107(1965).

 

 

106

IJ080212016o51556092.00CLEMENTSoBILL
IFOROLON

523
10

503

522
30
501

521

PROGRAM KINET

DIMENSION AKIZOQZOIO AKIIZOQZO)QNIZOIQEO(20I9T2IZDIoSKIZDoZDI

DIMENSION AIZOOAOIOIIIZOIOANSISOOI

COMMON NOIIOAOTEMPQAMAXOIMAX

TAII3-oos

IFLAGID

READ ORDER OF KINETIC MATRIX. LON FREQUENCY IN H20 HIGH FREOUENCY IN HZ.
AND SIZE OF PLOT IN CM (LESS THAN ABOUT TSI

READ IOQNQHLO oNHIoMoCM

FORMATIIZOZFToZOIboFSoZI

PRINT SO30N0ULOoNHIoMoCM

FORMATIIDH ORDER OF MATRIX. OIJOAXQISH LON FREDUENCT'OFD.29AXOI6M

IHIGH FREDUENCT'9F89204XQISHNOo OF POINTS. QISOAXOIQNSIIE OF PLOT.
ZQFGoZe/I

READ RELATIVE POP. IN THE ABSENCE OF EXCHANGE. RESONANT FREOUENCYQ AND
LINENIDTH IN HZ.

DO 20 N'ION

READ 3OQEOIKIOMINIITZIKI

FORMATIF60392FTQZI

PRINT SOIOEOIKIONIKIOTZINI

FORMATIZAN POPULATION PROPORTIDN' OFToJOIXOZDHRESDNANT FREDUENCT'

IOFDOZOONOIEHLINE NIDTH- OFDQZI

IF ITAUOIISZIOOOOSZI
DO AD N'ION

READ KINETIC MATRIX ONE RON PER THO CARDS
READ SOOISKIKvIIOI'IOIOI
READ SOOISKIKOIIOI'IIOZOI

FORNATIIOFOoAI

PRINTSIO

FORMATIIOHOKINETIC MATRIX'o/I

IF IN-IOISOboSOGOSOT

DO 508 I'IeN

PRINT SODQISNIIOKION'IONI

GO TO 60

DO SI? I'ION
PRINT SORQTSKIIONIOK'IOIOI
PRINT SORQISKIIQKIoK'IIONI
FORMATIIDIIX9F9oAII
READ A PARAMETER TAU AND THE HEIGHT OF THE PLOT IN CM LESS TNAN 25
READ TOOTAUOHEIGHTOIOPT

FORNNT ‘E006OF5020E‘,

IOPT IS THE PLOT OPTION (SEE END OF PROGRAM)

TEST TAU TO DECIDE NEXT STEP

IF TAU IS POSITIVE THEN USE IT AS PARAMETER TO CALCULATE LINESHAPE
IF TAU IS 0 THEN END. IF TAU IS 'I THEN READ NEH SET OF RESONANCE FRED
IF TAU IS LESS THAN 'I THEN READ IN A NEH KINETIC MATRIK
IF TAU IS DETNEEN D AND -1 THEN RESTART COMPLETELY
IFITAUISZOQDAQOA
IFITAU’IoDISZIQSZZQSZS
TTIM'DO
TIMEAITIMEFITTIMI
PRINT SOAQTAUOHEIGHT
FORMAT(SHOTAU'OEIIvoSXoISHHEIGHT OF PLOT-choZI
DO 80 I'ION
nIG'Do
DO ODD K'IQN
AKIIONI'SNIIONI/TAU
OIG'DIGOAKIIOKI
SURTRACT ANGULAR LINE HIDTH FOR DIAGONAL ELEHENTS
ALL CALCULATIONS ARE IN ANGULAR UNITS SO MULTIPLT HIDTH TIMES PI
ALSO ENSURE THAT THE RONS SUM TO ZERO BY ADJUSTING DIAOONAL ELEMENTS
AKIIOII'AKIIOII-TZIIIO30IAISD -DIG

107

C INVERT REAL PART OF MATRIX BY PIVOTAL CONDENSATIDN
DO II I'ION
DO ZI K'ION
AIIONI'AKIIQKI
21 AIICKONI'OQ
AIIOIONI'IO
11 [1(1):]
C TO IMPLEMENT PIVOTINGO USE MATRIX OF SUDSCRIPTS
C NON TRIANGULARIZE MATRIX SO IT CAN BE INYERTED
CALL TRIANG(Z.NI
NNN'N-I
DO 52 I'IONNN
C NON FINISH INVERSION
L-N-IOI
AMAXIAIII(LIOLI
LLL'NOI
III'Z'N
DO 32 J'LLLQIII
32 AIIIILIQJIIAIIIILIoJI/AMAX
LLL'L-I
OO 52 K'IOLLL
TEMPIA(II(KIoLI
JJJ'NOI
KKK'Z'N
DO 52 J'JJJOKKK
52 AIII(KIoJIIAIIITKIoJI-TEMPRAIIIILIoJI
AMAX'AIIIIIIOII
JJJ'NOI
KKK'ZPN
DO 62 J'JJJOKKK
62 AIIIIIIQJIIAIIIIIIoJI/AMAI
C TRANSFER TO SAFE KEEPING
' 00 I2 I'IeN
DO I? J'ION
12 AKIIIOJIIAIIIIIIQJONI
ERRO'DoO
00 I3 I'ION
DO I3 J'ION
ERRIPOoD
ERRZ'DoO
DO 3I3 N‘IoN
ERRI'ERRIOAKIIQKI9AKIINIJI
313 ERRZ'ERRZOAKIIIOKIOANIKOJI
IFII'JIOIAOJISOSIA
315 ERRI'ERRI-Io
ERRZ'ERRz-Io
316 ERRO'ERROOERRI'ERRIOERRZPERRZ
13 CONTINUE
ERRD'SORTIERROI
PRINT 5369ERRO

53‘ FORMATI. ERROR INDEN.POEII0‘I
C ABOVE HAS CALCULATION OF ERROR INVOLVED IN INVERTINO REAL PART
C OF MATRIX. THE PROCESS USED HAS TO FIND THE PRODUCT OF AN TIMES AII
C AND THE PRODUCT OF AMI TIME AK AND TO COMPARE THEM TO THE IDENTITY MATRII
C THE NUMBER OUTPUTTED IS THE SUM OF THE SQUARES OF THE DEVIATIONS OF EACH
C OF THESE PRODUCTS FROM THE IDENTITY MATRII. (ACTUALLY THE SOUARE ROOT
C OF THIS OUANTITY IS OUTPUTTEDI
560 DO 4 K-IOM
C NON FIND ICU) FOR M VALUES OF N
NI'ULO OICUHI-NLOIPNI/M

DO 26 IIIoN

00 2A J'IIN
26 AIIOJIIAKICIOJI°IHI-NCIII.(II-U IJII'39.ATOOO9OAKII0JI
C 39.676gfl9 IS IZPI) SOUARED TO CONVERT TO ANOULAR FREQUENCY

O 36 'IoN

AIION’II'I.

36

66
56

706

562

«as:
«so
953

955
S52

99

6lI

35550553.“)

706
705
612
613
414

617
616

108

IIIII=I
CALL TRIANO(N’I)
NON SOLVE THIS SYSTEM OF SIMULTANEOHS EO. TO GET THE REAL INTENSITY
A(N0N°2)=A(IIINIONOII/AIIIINIQNI
AFTER TRIANGULATING MATRIX. SOLVE HY BACK SUHSTITUTION
LILzN-I
OO S6KK=IOLLL
J=N-KK
TFNP3A(II(JIQNOII
JJJ’J’I
OO 66 L=JJJON
TENP=TEHP-A(L9N02).A(TI(JIOL)
A‘JONOZI3TFHP/AIIT(J)0JI
ANSIKI 300
O0 706 I=ION
ANSIK)=ANS(KI.(EOTII'AIIONOZII
THE EOLLONING ERROR CHECK IS ”ONE ON EVERY FIFTH POINT
IF((K/5)’S-KISS?OS6ZO S5?
NON CHECK FOR ERRORS RY CALCULATING ERROR=SQRT(EOPEO’RESIOUALYRESIOUALI)I
TEHP=OO
DO 550 I=I0N
Tfiup13‘I 0
OO SSI J=I0N
TFHPIZTENPIOAIJCNOPIYIAKI(IOJ).(HI-H(I)I.(HI'HIJII.
I39o676OHROAKIIcJI)
TFNPsTEHPOTEHPI'TENPIPEOIIIYEOIII
ERROR8 SORTITEHPI
IFIERROR'oOODOOOIISS29S5205S3
PRINT SSSQFPRORv HI
FORHATIIIH ERROR OF oEIIo6oIOH AT FREQ. oFBoZ)
CONTINUE
CONTINUE
TIHEASTTNEFCTTIHI-TIHFA
PRINT RQOTIHEA
FORNATIIOH TIHE oFIO-2o/I
IFTIELAO’6IOO6IIO6IO
CALI. anT (0.0000002000060000)
MOVE THE PEN TO THE FAR LEFT OF THE PAPER
CALL pLOT (o.o-l.o2)
IFLAG=I
FIRST TINE ONLY SET UPPER BOUND TO ?60 HHICH IS SUPPOSED TO BE I20 INCHES
CALL PLOT(?6DOO.030200002000I
HI TSAV3HE IGHT
HITSAV3CH
IORT‘I
X90530.
TFHP'O.
OO 6I6 KzloH
Y=ANS(K)
TO DEAL WITH COUPLED SYSTEMS. (HUT NOT CLOSELY COUPLED SYSTEHSIO
CHANGE THE FOLIOHING COMMENTED INSTRUCTIONS TO REAL INSTRUCTIONS HY
RFHOVINO THE C FROH EACH CARO. ALSO REOEFINE THE VARIABLES IA AND HEIGHT
SO THAT IX=ITHE NHHRER OF CALCULATED POINTS UETHEEN DOUBLET PEAKSI AND
HEIGHTSITHE RELATIVE SIZE OF THE UPFIELO PEAKI
SEE ALSO THE CONHENTS AFTER STATEMENT 379
IX=36
HEIGH1=O.HQ
IF(K-TXI705070S9706
YSYOHEIGHTYANSIK‘IXI
TFITEHPOYI6I296I3Q6I3
TFMP=-Y
CONTINUE
CONTINUE
IFIHEIGHTOPS0I6I606IOQ6I7
HFIGHT=?S.O
YSCaznoo/(TEHP/(HEIGHT/2.56)I

.50"?

603

601

60?
607
60“
605
179

708
707

96

61

SI
6]

TO!

121
Ill

91
161

IT]

109

lRC=RflO./(“/(C“/2.96TT

OPTION 0 IQ MOVE DFN HP 2 CM AHOVE PRTVTOIK HlfiHEfiT POINT AND TO LFFT
unnfilu. THFN 910T. nntlnu | Is anr rpnn thSFNT DnfilTION.
ODTION ? [9 anr DFN a (N In uthT nr pnfvlnus PLOT AND 9L0! Funu THERF
COLL anT (009".000?OOo/?o5“Q)Ono/?OS“I
lF(IUpT-l)601-602.603
XPOS=XPOS0HIT<AV‘I.

CALL DLOT(O.9NITSAV‘I.02I
an TO 60?
CALL PLOTTHITSAV°?.o-XRO§.?)
XDOS=0.

HITSAv=CM

60 To bflS
TFTHITSAV-CMT607o607oAON
HTTSAv=CN
TF(H£TuHT-HTTSAVT179.179.609
HTTfiAv=HFlfiHT

CALL PLOT(0.~O..0~Y§C.XSCT
Y=-AN§(1T

CALL PLOT(YoOoo?I

X300

00 7 T320”

=-AN<(IT
FOR COUPIFD SYfiTEMfi. DEMOVE THE C FROM THE FOLLOUIVG 2 CARDS TO RAKE
THFM INTO DEAL TN§TDUCTION§
IETI’IXT7O79707070H
Y=V-IFIGHTOANS([-TXT

1:101.
CALL PLOTIYQXOII

CONTINUE

FALL PLOTTO..X.2)

CALL PLOT(0.o0.ol)

CALL PLnT(0.oo.o-TT
an T0 60
CHNTTNUF
FA”)
cunnoutlnE TRIANG(HN)
“TNENQTON A(20.a0).ll(?0)
CONHUN NQIIOAOTEHPoAHAXoIHAl
SURDUHTTNE TO CanFRT MATRIX A AND ITS ASSOCIATED CONSTANT
MATRIX Tn UPPER TRIANGULAR FORM
NNNzn-l
DO 7II=IONNN
AHAX=AHS(A(lltl)oT))

luax=lltl)

[11:10!
00 «I K=ITToN
TFMP:AHS(A(II(K)¢ITT
IFTTEnP-AMAXTSToSIofil

ANszTFNP

TI(T)=IT(KT

Tl(K)=lMAx

TMAX=ITTIT
CONTIMUF
CONTINUE
AHAX3AIIHAXQII

IETAHAXIHIoQIoHI

tFfiT FOP NONSlNhHIAP OR <TNGULAH
LLL=IOI
no 71 K=LLLoN
l=II(K)
TFNR:A(L0TT/AHAX
TFTTEMP)IOTolll.IOI
SKIP IF ATDFADY 0

ITK=IOI
nn :2: I=IIK.NM
ATLoJT=AtlolT-TFNDOATTVhloJ)
CONTINUF
FONTINUF

THIS FINISHFQHFQ TDTANG HIAHIIATION
TF(A(TI(N).NTTITI-ulvlll

ppINT lal.HM '
FORMATTION SINGHIAD .19)
QTOR

CONTIMHF
DrtuNN
fMIT

 

 

i—‘~ 4""