\ I . giiiij $32331. .
. v. 57.. 1 . .. . Ll 592.13.?! 1 >. i . L « .2. if: .y as: LIB!
.m. . r! I... I ZhflnMW’L351353lPS Z: 7....) u u... u .yM..5)55 5h...‘.llurl .h?dfin‘i§§f¢fi9>fl’.1hfilfll¥ufiHALUQhMHPngthgx.Allis):
. . .5. n‘i\.:hvfi\wn$ufiw§vi§haoihaL§.§? .:..L|\..th;flo>\.rnfn. h. ‘\h>?.)\i£
r u. h . ~ a \Lt‘us :U..).vs.s.=)afi (JVH’U.§3.9K§I5!.§§\£\R NIH-AL. .. . u) a .
..§VM.\2. . v I; - ON“ I diva...” 5|». 7:. I . r . . «at: “mehvuzktfi it... ‘H’Hvsxav..lnuh,u%.:nuhw ihfi.\...1lflflg. a “(yang-Is; 53%
’2: "7": 3"! n v I u v. 1.): . . u:.i’.lz\ - i
luv. .0. 7:..1!..5IVY9VNK~MM. ..... .. y . 7.“.257 ahuéfii l .i v: ((5 gig Fifi-32:;
w‘vnzr’.slyficwfivns: a pa... 7 u! that. .51. .. . . . . . I. E. Ehrgga :vguix .5 l... Pitt‘s. (221.1931...
2. 711.132... aaaaaa 3 .zxfiianz a: 2.. 1 .. 2 «i .5. . - u 5‘ 1973. lbs:\\)9 .. gsé )1\$b..rr~t. rXrafivlS.§.t.Lr/<tr7§_
Jflflw‘flflnnla a; s Sluufllveil.lu.li57§ . Nita. v ~ to y .ktcrsitwasnmu‘ t\~§vnlfi?§$z .Ihiihg v ‘1 s . :fiwl‘ttflfiiutetvx)! \
.. .139 f . .... .. . 0.. . I. . .I . . ....:. 5. .n I «gig L99... 1%... 5%
g..zu;51flmcrczlv.o£fi§rv Futilifiuzsh‘ VLF...) . . . .25...‘.I€v6« nn>t+hn.fivsflsturu9i.hht\$¢\fl { {IV ‘£.V§}.l¢z .
51.7.5. 7......9. .1) :fi. H: v... Ix.‘1:t~¥v.{13vruvr . .1745)?! 1. «3.55 L.) 7.....\ . 1.. IA {slings-(5W3 I159E.»!a..-\3 .‘Piilry\t22£§
11.43- 3;:L1571rt .51; flag 2:. I7.)..1...i.£...nt:a.ch .x :1 z . at}!§>hV1..1&? L2 32%.“? kl.ii§i-§§i1§ . 5 fi\...li’.£h5§.
:vi». 7. :stEE.‘ Irvugzh’ 1.1.2.2325 Z4133? ... .. 3... 3... 1).. 4?. I: .v A... M35: 3. £16312: i3 32$}intttsnlri.
{:19 ir‘.r:n7:!xtc.t. {79.35 in): £.;ta1l)t.iu . 2 ¢ \ tut-v.13... i; git-IA". ‘ Sikhs... Aiigz‘iiri «5‘4 If S ’58.??‘5‘?lf:tl—
:3: a £5.33: . . $125.91.; : 3:31.. . . ., .3: :....?:ihi£:z . u 53......»152: $98.11;}: .3::.
.02; 3'1. 5".5llflvv 1v 1 1.91.1! 1.3T: t,‘vi?tfi 3 5.5.3324 t I 25110. £52. 9-. 31.333.55.59 l:::§lli 5 tithé‘itx
021! :29) L4. .5...) .. dip-n: 25.1.1.5. i xiv-1&5 .. Jug ~2E..g e: I: G533».§.§15n.ifl~.:uiiii.§sg‘flfi xggfgé.“
, Q! r , ’5 I ‘ g. ! z;:;,§ ‘ t s ) it: ~
a viazvr 37 3%. L5... 31.2.: zlxvt-ztl : .... Nv 71m: 9 .. . ti 3 séiusxbf} .l u §.1£§ is}...
.1; .. Hialfiuwnmuvfflfiifi wan: 8%. kg?! .1315: {gauné EVLfio$PPn3§§FS§§ xiii”. fikitn .Sfifxuu... haiifa _ .
. . t .4 l... 28“ . g5
. z. .. .. 5.. ts $5.53.??? . g‘: int!
1.19.1.5 Lilith”?! 7‘1““ x .

~. a . )5tibsh!.fll\
’9‘ i}, \uvas‘gla h
{“Enl‘ ~ 9

. v. ..
.. . .g . . i...
3. .. . . as: v. I .
5.1.3.... iglzvzluammurz. . fl! .. . I}!!! n
3..» ..... L3:$.Svl.; .5... 9!. ngiafizuwhgflaalywhh 1.... $114.... 1 {1:83
1 . ~ . $1“: ‘9 z

   

 

 
 

   

   
        

                 

 

       

.
a . ~ .. I . fizéaas .1.) n 009:9..‘ullk‘39ti
«H»I~.lv;VJ-u.,.‘1!2!vSign .HaizfisnrwlnulIWIN’ I t . 7 x. ~ . .Iflfisgtvllvi.§w :ll'flHaiu-Lf 568.41.!1. )5
I I. 2c.li.§a'.wlA-.Esh.'flh‘tti 00-h" ...«w,hVQh—I~Au.fld.flu4i I; i 5!. . fin. . i532 .
1......ivnIJmatvltlJavIJ.....~.n..-.H.3N.tt‘$t y‘tv'vfi 2......tl1fn'nth‘2anflnZy7ldfiu .. “new“... lily IAN“! 1...”...55 .. z
.. . T u: Edvixgnv‘n Eviarfvairfiflci s. tintil.7‘.l. . 523.019.. tgfia 16 “Ian Eyizlnvflfltliultut
. 1.1.1.1.... :5... .3 1.... 1.1.. . 1.. 3L 33%, gay.
8 221...... ifivzydirgs; 6E§.t.zflz.h . lihiiiihvflflihnnflfiis. “$92-1. . 12.1%. i: xiii..1tn.i!¢i..s
, .. ..m.....i!.nu.tsv.sun§. «hum...» -gsffivmwabihn L... . . § .1831}...
. .. .. Bum... S:..In.3§ii .351 . g .11. .3434... .. (3.3.... LR:
. .i Izfignflu .. gguzizu 2.2.1.3. . :2 3.... u... 1.. 153.51....ng
. I: . 01.98! jib-’31 .2... .b’it‘lfl.‘ ‘5: avv‘nggilliiitll flinch“: g \ti‘§§§.égfi
,. - ...:.. . z .1... (twangfiq. . ..... 1c. szfiqflhfiflfib $1513....a4i $1.32 SEER... t . «Nana. . .
.1...‘ . :v: lttflJ-I‘: : §HQ.? A 151ng a 34.17 3 l: .
. a .. . R3503... 1.... $35th . .. .
l w.

.
p ‘i! 17% Q- i
u th’fi‘WF-fiv arguggaabu .mtaiadwtt. 1%....
v-w : ‘ ‘ 77 .
.1nu§.i+.§dnuiyfisvlvl§i 58.51.. Wiinsg .. :7
. . .. Elana 1...? E... 5.9....
. . r
.1 ii 1.9... .QWWWELI I'lfluntle.
1&1. Lu. :5
; f.
..

   

6
2
3

 

 

    

{€111th .

v1. 13.}.11‘5!" 57.! . . ‘3‘;
if.“ 131312;. 03? . Nahuunxvasiwlcvi
iiiilx .f~%%§3

 

              

   

. .siliqivis . .

v . 7R I . a . . . .

\1 \w . . . .

1%vii‘ é I114}! Jill-71.1% if). ‘1‘}..1511 .

12.1.: IL." .G‘i‘i} .. i1. . .
.\ . . .

. I
I... In».
..T‘

      

  
    

4:9):

    

Estfr?
.410 gle

A 214%.? fig. {’7}!!-

   

: a .. . . n . .

. .. . Eirbwzx. i; . . i . It§{.flf L...
........... n Bl§5tsjr~5i§lu v T c; . ‘1}! $5.. 3 II... 111%! iiiwf‘iiuiii . lhhsxrhu‘n
E; ii"! (.1 . .itfil‘ls._..os n... 13111.... ‘ .l 7111: Lil???)

a...) v' 5%? .I. 3 .l q :1 v.75} . a! $11. iii‘fi‘i . lllt.‘

IvoL 10% I giqui‘i‘tu. c. 11%! 1%" . i173§§ . 3.5.31!

19% 1... 1.19 kg Iii-Hz. r73.\\1‘..1§114§51i )i‘lxls‘ .‘anu. 12.3.”: 3351(“3: gtfivvi$iilliw . Isl. It?”

. {I 1 2. 1311-13.“ .\ ‘3 II. 13.... )2... 7‘ gi‘!)1vls%fifl‘\\ll.l 11‘): 13!»! g. . .

. #333. i H.) . .11}? 71“: .2... x d. ‘1‘ \ ‘11.. (1‘7... .2. $1....X.‘11111\\Y\)..1$}1 .. 1115\13111‘i\s‘\‘q I \ It‘ll. Egtx

. . ,‘\.1 ti...) , I .. .. . .J'. ... ‘ 1 § . . .1.v1\i ,..v.1| 1...... .41. .5 ~ . .7:.x.:. . . . It .y . . . \z

TH E315

 

 

This is to certify that the

dissertation entitled

FACTORS AFFECTING DIASTEREOSELECTIVE
REDUCTIONS OF 9 , 10-DIHYDRO-9 , l 0- (1 1-KETO-
12-ALKYLETHANO)-ANTHRACENES

presented by

BRUCE R. OSTERBY

has been accepted towards fulfillment
of the requirements for

Ph. D. CHEMISTRY

degree in

MR

Major prol'éssor
Date I”?! I 3)! 3k

MSU is an Afflmmlive Anion/Equal Opportunity Institution 0-12771

 

 

 

RETURNING MATERIALS:
IV1ESI_J Place in book drop to
LIBRARJES remove this checkout from
as. your record. FINES Will

 

 

V be charged if book is
returned after the date
stamped below.

 

 

 

 

FACTORS AFFECTING DIASTEREOSELECTIVE REDUCTIONS 0F

9 10-DIHYDRO-9 10- 11-KETO-12-ALKYLETHANO -ANTHRACENES

by

Bruce R. Osterby

A THESIS
Submitted to
Michigan State University

in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Chemistry

1982

 

ABSTRACT

FACTORS AFFECTING DIASTEREOSELECTIVE REDUCTIONS 0F

9.10-DIHYDEO-9L10-(ll-EETO-IZ-ALKYLETHANOz-ANTHRACENES

by

Bruce R. Osterby

A cornerstone in theories of diastereoselective reductions of
ketones by complex metal hydrides is that the reagent attacks the
least hindered side. Varech and Jacquesl’z, however, reported that
this axiom does not hold true for a series of bicyclic ketones, but
was in fact directly opposite to it. When 9,10 - dihydro - 9,10 - (11
- keto - 12 - alkylethano) - anthracenes (alkyl = methyl, ethyl, and
iggépropyl) were reduced by lithium aluminum hydride, attack of the
hindered (alkyl) side increased as the steric bulk of the alkyl group
increased. The percentage of attack of the alkyl side was 35, 51 and
80% for methyl, ethyl, and iggfpropyl, respectively. When the bulky
group was trbutyl or phenyl, attack of the hindered side was 23 and
0%, respectively.. These "surprising" results were attributed to an
anistrOpic inductive effect.3

I have examined the product ratios as a function of temperature
and calculated the parameters AAHI and AASI. The study was done for
lithium aluminum hydride reductions in diethyl ether and
tetrahydrofuran and for sodium borohydride reductions in isopropyl

alcohol. The AAH‘ component indicates that attack of the more

hindered side is due to enthalpy factors. This has been attributed to
product development control. The AAS:F component shows a strong
preference for attack of the least hindered side (to give the gig
alcohol). This preference is more pronounced for the trbutyl group
than for the other alkyl groups and is more pronounced in
tetrahydrofuran than diethyl ether. These results are most likely due
to solvent entropy effects. In the phenyl substituted case the
hydride attacked from the least hindered side in all cases studied.
This has been interpreted as the result of a strong coulombic
repulsion between the negatively charged nucleophile and the electron

density of the phenyl n system.

1. D. Varech and J. Jacques, Tetrahedron Lett., g_, 4443(1973).

 

2. M. J. Brienne, D. Varech and J. Jacques, Tetrahedron Lett., fig,
1233(1974).

3. M. C. Cherest, H. Felkin, P. Tacheau, I. Jacques and D. Varech,
Chem. Comm., 372(1977).

ACKNOWLEDGMENTS

I would like to express my deepest gratitude to Dr. Gerasimos I.
Karabatsos, under whose guidance this research was performed.

I also would like to acknowledge the faculty, staff and graduate
students of Michigan State University's chemistry department for their
help and friendship.

I would like to give a special thanks to my mother and father for
their unfailing support and love throughout my education.

Finally, I wish to express my appreciation and gratitude to my
wife, Meg, for her unselfish support and immense help in making this

all possible. Her effort and assistance will always be remembered.

ii

TABLE 93 CONTENTS
page
LI ST OF TABLES O O O O O O O O O O O O O O O 0 v

LIST OF FIGMS O O C O C O O O O O O O O O O 0 v i

INTRODUCTION

I. The Scope of the Problem . . . . . . . . . . . 1

II. The Stereochemical and Energy Relationships for
Carbonyl Addition Reactions . . . . . . . . . . 2

III. Models for Diastereoselective Additions to and

Reductions of Ketones (Factors Affecting

Asymmetric Induction) . . . . . . . . . . . . 6

A. Steric and Conformational Concepts . . . . . . 7
l. Acyclic Ketones and Aldehydes . . . . . . . 7
2. Cyclic Ketones . . . . . . . . . . . . 11

B. Molecular Orbital Models . . . . . . . . . . 16

IV. Hydride Reductions of 3 - Alkyl (and Phenyl)
Bicyclo[2.2.2]octan - ;,- ones and 9,10 - Dihydro -

9,10 - (11 - Keto - 12 - Alkyl (and Phenyl) ethano) -
Anthracenes . . . . . . . . . . . . . . . 19

RESULTS AND DISCUSSION

I. RESULTS . . . . . . . . . . . . . . . . . 23
II. DISCUSSION . . . . . . . . . . . . . . . . 33

A. The Composition of Lithium Aluminum Hydride in
Diethyl Ether and Tetrahyrdofuran . . . . . . . 33

B. The Mechanism of Lithium Aluminum Hydride Reduction
of Ketones and the Role of Alkoxyaluminohydrides . . 34

C. The Mechanism of Sodium Borohydride Reduction of
Ketones in Alcohols . . . . . . . . . . . 37

D. The Activation Parameters . . . . . . . . . 38

iii

1. The Role of Enthalpy . . . . . . .
2. The Role of Entropy . . . . . . . .

3. Conclusions . . . . . . . . . . .

EXPERIMENTAL

I. General Methods . . . . . . . . . . . .
A. Stock solutions of complex metal hydrides .
B. Temperature Control . . . . . . .
C. General Procedure for Reductions . . .
II. Syntheses . . . . . . . . . . . . . .
III. Product Ratio Analysis . . . .

IV. Error Analysis . . . . . . . . . . . .

APPENDIX

A MAGNESIUM-25 NMR STUDY OF ORGANOMAGNESIUM COMPOUNDS
AND MAGNESIUM HALIDES IN DIETHYL ETHER AND TETRAHYDROFURAN

REFERENCES . . . . . . . . . . . . . . .

iv

. 38

45

46

47

47

47

48

49

. 62

66

. 68

. 75

LIST 0_F TABLES

Percentages of alcohols from equations 2 and 3 (%trans
= 100% - %cis). Reactions were run at 20°C in diethyl

ether. 0 O C O O O O O O I O O O O 0

Product ratios from lithium aluminum hydride reductions

in diethyl ether. 0 O O O O O O O O O 0

Product ratios from lithium aluminum hydride reductions

in tetrahydrofuran. . . . . . . . . . .

Product ratios from sodium borohydride reductions in
isoprOpyl alcohol. . . . . . . . . . . .

Results from thermodynamic equilibration via the MPVO
reaction. Lithium tri - t_- butoxyaluminohydride
reductions. . . . . . . . . . . . . .

Activation parameters from Figure 11. Lithium aluminum

hydride in diethyl ether. . . . . . . . .

Activation parameters from Figure 12. Lithium aluminum

hydride in tetrahydrofuran. . . . . . . . .

Activation parameters from Figure 13. Sodium
borohydride in isopropyl alcohol. . . . . . .

stg nmr Data for some Grignard Reagents and Magnesium

Halides in Diethyl Ether and Tetrahydrofuran (THF).

page

. 25

. 27

. 28

10.

11.

12.

13.

14.

15.

LIST OF FIGURES

The three stereotopic relationships. . . . . . .

Enantioselective (left) and diastereoselective (right)
reaction profiles. . . . . . . . . . . . .

The open chain, cyclic and dipolar models (8, M, L are
small, medium, and large, respectively). . . . .

The two transition states proposed by Karabatsos.
Felkin's model for 1,2-asymmetric induction. The
preferred transition state is on the left (S, M, L
are small, medium, and large, respectively). . . .

Felkin's model for cyclohexanone systems. . . . .

Wertz and Allinger's model based on fewer gauche
vicinal H/H interactions. . . . . . . . . .

Felkin's torsional strain adapted to cyclopentanone
systems. I O I O O I O O O O O O O 0 O

The case of camphor and norcamphor. . . . . . .
Ahn's antiperplanar effect. . . . . . . . . .
Lithium aluminum hydride in diethyl ether. . . . .
Lithium aluminum hydride in tetrahydrofuran. . . .

Sodium borohydride in isopropyl alcohol. . . . .

4 - t - butylcylcohexanone and 2,2 - dimethyl - 4 - t -

butylcyclohexanone. . . . . . . . . . . .

The bicyclo[2.2.2] system. . . . . . . . . .

vi

page

. 10

. 15

. 19

. 31

. 41

INTRODUCTION

I. The Scope _f the Problem
When 4-trbutylcyclohexanone is reduced by lithium aluminum

hydride a pair of diastereomeric alcohols is formed in the ratio of

9:1 (Eq 1).1

0

L M IH H" H H
W 5204’ m

90% ' l0°Io

V

The reaction is stereoselective, which should not be surprising. What
is surprising is that the reason(s) for the observed selectivity is
not known for certain. Several theories have been postulated to
explain why axial attack is preferred, but no one theory has been
overwhelmingly satisfying. This reaction exemplifies, .perhaps more
than any other, the problem of stereochemistry in complex metal
hydride reductions of cyclic ketones.

The factors which control the stereochemistry of ketone (both
cyclic and acyclic) reductions have been speculated about for thirty
years. With almost an unlimited assortment of ketones, no single
generalization is applicable to all systems. Studies of acyclic
ketones and aldehydes have focused primarily on those with an
asymmetric center adjacent to the carbonyl group (1,2-asymmetric

induction). Studies of cyclic ketones have been divided into groups

depending on the conformation(s) of either rigid or flexible systems.

Aside from an intrinsic reason for studying factors which control
the stereochemistry of ketone reductions, there is a practical reason.
A large number of multistep total syntheses require at some point one
or more asymmetric centers. Very often the methodology used in
solving the stereochemical problem has been empirical. Understanding
the factors which control stereoselectivity have greatly increased the
efficiency of such syntheses. Such knowledge has also been useful for
configurational assignments and mechanistic studies.

There are a vast number of carbonyl addition reactions. One must
be careful in extrapolating data from one type of reaction to another.
As with complex metal hydride reductions and organo-metallic
additions, most of these reactions are kinetically controlled.
Therefore, the stereoselectivity is directly proportional to the
difference in transition state energies, AAGI. Since transition state
structures can only be surmised, determining the factors which control

stereoselectivity is, by nature, a very subtle problem.

II. The Stereochemical and Energy Relationships jg; Carbonyl Addition
Reactions

The terms enantiomeric and diastereomeric are commonly used to
describe the relationship between pairs of stereoisomers. In a
similar manner the relationship between portions of a molecule may be
compared. Recognizing the stereochemical relationships between parts
within a molecule is extremely valuable in solving stereochemical

problems.

Consider the two faces of a carbonyl double bond. The
stereoisomeric environments of the two faces may be equivalent,
enantiotOpic or diastereotopic. The classification depends upon the
symmetry e1ement(s) associated with the molecule.“ If the molecule has
a C3 axis along the C-0 bond, then, by definition, the faces are
equivalent. They are indistinguishable under all circumstances and
addition reactions would give no stereoisomers. If the faces can only

be interchanged by an S axis (mirror plane containing the sp2

1
carbonyl bonds), then the faces are enantiotopic. They are
indistinguishable under achiral conditions and distinguishable under
chiral conditions. Addition of an achiral reagent would give a
racemic mixture of enantiomers. If the reagent is chiral, then only
by chance would the mixture be racemic. Finally, if the faces are not
interchangeable by any symmetry operation, then they are
diastereotOpic. They are distinguishable under all circumstances and
addition reactions would give a pair of diastereomers in, most likely,
unequal amounts.

With the stereotOpic relationships for the two faces of a
carbonyl defined, consider now the two reaction paths. With achiral
reagents the reaction paths are related in the same manner as the
products (there may be certain exceptions with diastereotOpic faces
because of a preliminary reaction or coordination on one of the
sides). For ketones with diastereotopic faces the transition state
energies are most likely different. In a kinetically controlled

reaction this difference is directly pr0portional to the product

ratio. Such processes are termed diastereoselective. There is no

energy difference between the transition states for ketones with
enantiotopic faces using achiral reagents. A chiral reagent, however,
turns an enantiotomeric relationship into a diastereomeric one. Thus
ketones with enantiotopic faces go through diastereomeric transition
states with chiral reagents. The term enantioselective applies to
this type of reaction. The hydride reductions of the three ketones in
Figure 1 illustrate these points.

The energy relationships for enantioselective and
diastereoselective reactions are shown in Figure 2.

Although the terms enantioselective and diastereoselective have
been used recently to describe reactions similar to those just
discussed, the terms asymmetric synthesis or asymmetric induction are
more prevalent in the literature. Any reaction where the epimers of a
new asymmetric center are not formed equally is classified as an
asymmetric synthesis. An enantioselective reaction is an asymmetric
synthesis. A diastereoselective reaction is usually but not always an
asymmetric synthesis. The example of 4 - £_- butylcyclohexanone
illustrates this point. Nevertheless, most ketones with
diastereotopic faces have an asymmetric center incorporated in the
molecule and upon addition reactions form epimers of a new asymmetric
center (provided an identical ligand is not added to the carbonyl).

In studying the factors which affect diastereoselective
asymmetric syntheses it is next to impossible to use isomerically pure
enantiomers. This is no problem, however, since a racemic mixture
gives the same information. For example, consider an Optically active

ketone of S configuration. Upon reaction it gives a ratio

EQUIVALENT FACES

o H. OH
1C: ’2]
/ \ /"\
H3C CH3 H3C CH3
ENANTIOTOPIC FACES
liq, H 0 ti OH
I ks H / k, I
. / \ .. . / \ A/ \
H3C HM szCH3 H3C CH2CH3 HBL I CHZCH3
‘ k5 =i<r (achira/ reagent) KR)
k3 #k, (chiral reagenr,’
DIAS TEREOTOPIC FACES
HQ; ti 0 FL OH
ac, CH kss rid CH —9k’5 ,
_{ /\/ /\/3 /C\/CH3
3C C H3C q\ fiBC C
"J CH2CH3 H CHZCHB H CH2CH3
(3,8) (8) (9.3)

Figure l. The three stereotOpic relationships.

   

 

 

 

 

Figure 2. Enantioselective (left) and diastereoselective (right)

reaction profiles.

6
of diastereomers (S,R):(S,S). Symmetry principles require that the R
enantiomer give exactly the same ratio. Therefore the ratio of the
two enantiomeric pairs [(S,R)+(R,S)]:[(S,S)+(R,R)] is identical to the

ratio obtained from an isomerically pure enantiomer. For example:

 

(S.R) - (20%)-——-
(S)‘ketone

(S.S) - (80%)

(R.R) - (80%)
(R)-ketone

(R.S) - (20%)--

III. Models for Diastegeoselective Additions £2 and Reductions 2f

 

Ketones (Factors Affecting Asymmetric Induction)

The first person to recognize the concept of asymmetric induction
was Emil Fischer in 1894.3 In going from one sugar to its next higher
homolog, via the cyanohydrin reaction, he reported that the reaction
"proceeds in an asymmetric manner". It wasn't until the early 1950's,
however, that the foundation was laid' for interpreting the
stereochemistry of carbonyl addition and reduction reactions. The two
types of reactions most studied in this field are complex metal
hydride reductions and organo-metallic additions. This is due both to

their importance and to the large amount of data on such reactions.

A. Steric 53g Conformational Concepts

The most popular way of predicting asymmetric induction has been
and is based upon steric interactions as the nucleophile approaches
the carbonyl. Both SCF-LCGO-MO calculations4 and chemical intuition
predict that the nucleophile approaches the carbonyl somewhat
perpendicular to its plane. Because of the low activation energies
and the high exothermicity of these reactions, an early "reactant
like" transition state is assumed based on Hammond’s postulate.5"

1. Acyclic Ketones ppg Aldehydes

The first rule to correlate and predict the stereochemical
direction for 1,2-asymmetric induction was proposed by Cram.1 He
reasoned that a particular conformation will be preferred in the
transition state and the nucleophile will simply choose the least
hindered side to give the major diastereomer. He proposed ‘both an
open chain and a cyclic model (Figure 3). In the open chain model
vsteric bulk was assigned to the three ligands on the asymmetric
carbon. It was assumed that the carbonyl oxygen would orient itself
between the two least bulky groups. Cram’s reasoning was that the
oxygen effectively becomes the bulkiest group in the molecule because
of complexation to the metal. The reagent then prefers to attack the
least hindered side. This model predicts the major diastereomer in
most cases where the three groups are either alkyl, aryl, or hydrogen.
In cases where an oxygen or nitrogen is available for further
complexation (such as an amino or hydroxyl group) the cyclic model
applies. A third model was later proposed by Cornforth.a Called the

dipolar model, it included cases where a halogen, usually chlorine, is

8
one of the three ligands (Figure 3). Cornforth assumed the trans

dipolar conformation is more reactive.

~ R
.stgR’ L crew cum R.
\o” “o L 5 K o
4%
w") ( j i
3 R C! R
crcuc amoun

Figure 3. The open chain, cyclic and dipolar models (S, M, L are

small, medium, and large, respectively).

In 1967 a semiquantitative model for 1,2-asymmetric induction was
proposed by Karabatsos.’ This was based on experimental evidence of
the preferred conformations of ketones and aldehydes determined by
microwave and NMR spectroscopy. That evidence shows the stable
conformations are those with a single bond eclipising the carbonyl
double bond. Karabatsos assumed that the transition states were
similar. Using this assumption, the most stable diastereomeric
transition states were qualitatively predicted. The two transition
states are depicted in Figure 4. It was assumed that the incoming
reagent only attacks the least hindered side, where the smallest group
(usually a hydrogen) is situated. The relative stabilities of the two
transition states give the diastereomer ratio. A semiquantitative

prediction of these stabilities was made by considering the energy

difference of the two ground state conformations. Other interactions

 

 

 

9
are evident and difficult to predict, but to a first approximation

tend to cancel.’

M
O.
s =SMALL
---R’ lw=MEmuu
S =LARGE
L
R

 

Figure 4. The two transition states proposed by Karabatsos.

A third model was advanced by Felkin and co-workers which again
assumed an early transition state and was also applied to
cyclohexanone systems.1° They contend that there are two types of
interaction in the transition states. One involves torsional strain
(Pitzer strain) with the newly forming bond so as to prefer staggered
conformations. The other type considers the Steric interactions of
both the incoming group, R', and the achiral group R (Figure 5). This
differs from the models of both Cram and Karabatsos in which the
interaction of the carbonyl oxygen is important. Of the six possible
staggered conformations, three were qualitatively chosen to be of
lower energies. These three transition states are shown in Figure 5.

The combination of torsional strain and steric interactions was
used to explain metal hydride reductions of and Grignard additions to
cyclohexanones. In the absence of axial substituents at C3 and C5,

torsional strain between the axial hydrogens at Cz and C6 and the

10

Figure 5. Felkin's model for 1,2-asymmetric induction. The preferred

transition state is on the left (S, M, L are small, medium, and large,

respectively).

axda/ attack
preferred

 

H torsional
strain

steric
strain

Hu'la
.,. H
H 0
H
l

\ equatorial

attack preferred

Figure 6. Felkin's model for cyclohexanone systems.

 

11

newly forming bond is predominant. With a substituent in the axial
position at C3 and/or C5, steric interactions at those positions
become more important (Figure 6). This is similar to the explanation
Richer gave, except he labeled the affect of the C2 and C‘ axial
hydrogens as steric strain.11
2. MW

Steric affects in the reductions of, and additions to,
cyclohexanones have been and are, of particular interest. The problem
is well documented."'13 The first empirical observation for such

systems was made by D. H. R. Barton in 1953.14

He noted that, in
general, the metal hydride reductions of unhindered keto groups give
the equatorial epimer. The axial epimer is formed if it is hindered.

. . . 15
This is true to a certain extent. Dauben,

in 1956, thought the fact
that, for simple monocyclic ketones, the product from metal hydride
reductions is also the thermodynamically more stable isomer, was not
fortuitous. Since the equatorial epimer is favored in both cases, he
reasoned that product stability played an important role. By
examining a number of reactions he concluded that equatorial attack is
preferred sterically and that axial attack is preferred because of
product stability. At that time Dauben rationalized the observed
stereoselectivity as a function of "Steric Approach Control" versus
"Product Development Control". This is similar to Browns’s "Steric
Strain Control" and "Product Stability Control".1‘ On the other hand,
Felkin reasoned that the nature of the transition states should be

similar and using his model there is no reason to invoke product

stability.

12

The concept of product development control was criticized on two

17". First of all, in the

additional points by Eliel and Klein.
lithium aluminum hydride reduction of 4fpfbutylcyclohexanone, the more
stable isomer is formed in a greater proportion than in the
corresponding thermodynamic equilibrium of the aluminum alkoxide
derivatives. The pgppg to pig ratio of the products of the reduction
in diethyl ether at room temperature is 10:1 and for the corresponding
equilibrium in tetrahydrofuran (66°C) the ratio is 4.5:1. Eliel
suggested that such a difference could not support product stability
as the controlling factor. Furthermore, relative rate studies and
kinetic results show that putting substituents in the axial_ positions

at C and C5 drastically slows down axial attack, but has only a

3
nominal effect on the rate of equatorial attack. Thus, Eliel and
Klein ruled out product development control as an important factor.

Wertz and Allinger approached the problem from a slightly

different angle.1’

From molecular mechanics calculations they
concluded that the minimization of the number of gauche vicinal H/H
interactions is a strong driving force for a number of conformational
results. For example, they explained the preferred pppi conformation
for butane as a result of fewer gauche vicinal H/H interactions. In
the same vein they explained the preferred equatorial position of a
substituent on a cyclohexane ring as being a result of the energetic
preference for the tertiary hydrogen to be axial rather than
equatorial. Thus, they suggested the incoming hydride is less

hindered in axial attack than in equatorial attack, based on the fewer

number of gauche vicinal H/H interactions (Figure 7).

13

O...
H
(CH2). 0
H
.
hindered

Figure 7. Wertz and Allinger's model based on fewer gauche vicinal

H/H interactions.

This theory is actually very similar to Felkin's model. Wertz and
Allinger point this out and see no serious conflict between the two.

The concepts of torsional strain and steric strain in the
transition state have been fairly successful in aiding the
interpretion of stereochemical results. Ashbyzo has pointed out,
however, that while torsional strain may hinder equatorial attack (Cz
and C‘ axial hydrogens), no mention has been made of the torsional
strain involved with axial attack. If the carbonyl group is in the
plane of the equatorial C3 and C‘ hydrogens, then the incoming group
should encounter torsional strain from the C3—C3 and C5-C‘ bonds of
the ring. He points out that this may provide further evidence that
the carbonyl group lies below the plane of the C3 and C‘ equatorial
hydrogens, as depicted in Figure 7.

More recently an alternative explanation for cyclohexanone

 

14

systems has been presented by Wigfield.21 Based on kinetic evidence,
he suggested a long acyclic transition state for sodium borohydride
reductions. Taking this into account, he proposed that if the angle
of approach of the metal hydride is 1260 to the carbonyl (based on
theoretical calculations‘), rather than 90°, then interference with
the axial hydrogen at C4 may be severe. Wigfield rationalized that
the stereoselectivity may simply be explained by taking into account
the added interference of the axial hydrogen at C4. Furthermore, he
demonstrated that an axial methyl group at C4 does have a pronounced
affect on equatorial attack.23 It should be noted that the
stereoselectivity of sodium borohydride is essentially the same as
lithium aluminum hydride in these systems.

A short account of the stereochemistry of cyclohexanone
reductions would not be complete without mention of one of the more
perplexing problems. When 4-prbutylcyclohexanone is reduced by
lithium trimethoxyaluminohydride, axial attack is reduced to 60% (90%
with lithium aluminum hydride). This would seem in line with the
greater steric bulk of the reagent and the notion that axial attack is
sterically controlled. However, when lithium pgi 1 p -
butoxyaluminohydride is used the percentage of axial attack goes back
to 90%. No satisfactory explanation for this observation is at

23 that the observed

present available. It is the opinion of Ashby
ratios are probably a result of fundamental differences in the manner
by which these two alkoxy hydrides react. This demonstrates the need

to know more about the mechanisms of these reactions in order to

explain the stereochemistry.

 

 

 

15

Cyclopentanones have been studied far less, but follow a pattern
akin to cyclohexanones. In the case of 2-methylcyclopentanone,
lithium aluminum hydride and sodium borohydride attack gig to the
methyl group, giving 76 and 74% of the ppppp alcohol,
respectively." The methyl group is most likely in a quasi-equatorial
position and attack appears to be from the more hindered side (Figure
8). Morrison and Mosher have implied that the stereoselectivity can
be explained by invoking torsional strain.3’ Lithium pp; — p,-
butoxyaluminohydride and lithium trimethoxyaluminohydride give 72 and
56%, respectively, of the ppppp alcohol.3‘ It should be noted that in
analyzing these data, the contribution from the conformation with the
methyl group being quasi - axial is not known for certain. If a
methyl group is locked in a quasi - axial position at C3, as in pig -
3,4 - dimethylcyclopentanone, then pip attack is reduced to 10% (for
the lithium aluminum hydride reduction). This is presumably because

of steric hindrance," as is depicted in Figure 8.

7630/0 (LiA/H4) steric hindrance

   

torsional strain 90°lo (LiA/H4j

Figure 8. Felkin's torsional strain adapted to cyclopentanone

systems.

 

l6
Reductions of the bicyclic ketones camphor and norcamphor are
generally assumed to be largely controlled by steric hindrance.37 The
gyp C, methyl group of camphor severely blocks 252 attack, thus the
lithium aluminum hydride reduction of camphor gives 91% of the pgp
alcohol. In the case of norcamphor, ppgp attack is more hindered and
a similar reduction gives 91% of the gpdp alcohol. Organo-metallics

show even greater selectivity as one might expect. Figure 9 shows the

preferred faces of attack.

l

c
amphor norcamphor

Figure 9. The case of camphor and norcamphor.

B. Molecular Orbital Models

If the faces of a carbonyl are diastereotOpic, then on this basis
there is no reason for the electon density on each face to be the
same. In 1973 this idea was investigated by ° Ahn and
co-workers." They carried out ab initio (STD-3G) calculations on a
number of fixed conformations of simple aldehydes and ketones. Their
calculations showed that the n-electron cloud density is greater on
one diastereotopic face than on the other. Furthermore, if attack
occured at the more electropositive face, then the same

stereochemistry would result as predicted by steric control. It

 

17

should be noted that in the cases studied, the difference between the
two faces was always the same, that is, a methyl versus a hydrOgen.

At about the same time Ahn's work was published, Klein examined
the case of cyclohexanone by frontier orbital analysis.” An
electrophile is expected to attack the highest occupied molecular
orbital (HOMO) and a nucleoPhile the lowest unoccupied molecular
orbital (LUMO). By mixing the n orbital of the carbonyl with the eS
(symmetrical) orbitals of the C3-C3 and C5-C‘ bonds of the ring (B to
the carbonyl), Klein concluded that the n lobes of the ensuing HOMO
would be larger on the equatorial side. If all other considerations
were equal, electrophiles would be expected to react on this side
(equatorial attack). In fact, diborane does attack
methylenecyclohexanes equatorially.3° Repeating the same process with
the corresponding anti-bonding orbitals, the p orbital of the carbonyl
carbon is distorted to the axial side in the LUMO. Nucleophiles would
therefore prefer this side and axial attack is observed with complex
metal hydrides.

Klein has used another approach for the same problem.29 He
considers the interactions between the occupied n orbital and the ciS‘I
orbitals of the C3-C3 and C5-C‘ bonds. To secure better overlap, the
n lobes should distort to the equatorial side. In a similar manner,
the vacant r‘ orbital is stabilized by the as bonding orbitals of the

C -C and C5-C‘ bonds. Thus the n‘ orbital is distorted to the axial

z a
side.

Ashby and Boone are of the Opinion that Klein misinterpreted the

LUMO and that a better representation of it would in fact predict

18

equatorial attack for nucleophiles.31 They also suggest that the
choice of the BC-C bond is not straightforward. Klein had chosen the
BC-C bond because it is more polarizable than the BC-H bond. Ashby
points out, however, that the BC-H bond has a somewhat better geometry
to effect overlap with the carbonyl orbitals.3z

The most ambitious MO study was carried out by Ahn and Eisenstein
in 1977 on 2-chlor0propanal and 2-methylbutanal."3 Ab initio (STO-3G)
calculations were done of the transition state energies for the attack
of hydride (H-) on the two faces. A total of twelve conformations
were generated by rotating the C1-C2 bond by 300 increments. For each
molecule, therefore, twenty-four diastereomeric transition state
energies were calculated. Ahn and Eisenstein also took into account
the effect of the counterion and solvent. Their results produced six
low energy transition states, three for each diastereomeric face. The
geometries are very similar to those predicted by Felkin and
co-workers.

The stabilization of these low .lying transition states is
attributed to what Ahn and Eisenstein call the antiperiplanar effect.
This effect arises when the newly forming C1-H and the Cz-L bonds (L
is the large or polar substituent) are antiperiplanar (Figure 10).
Such an arrangement would allow for good overlap between the n.
orbital of the carbonyl and the of orbital of the Cz-L bond. The
principle orbital interaction in the transition state involves the
HOMO of the nucleophile and the n‘ orbital of the carbonyl. The a. of

the Cz-L bond is thus in a position to stabilize the n’ orbital of the

carbonyl and therefore lower the transition state energy. Ahn and

19

Eisenstein's calculations assess an extra 5-10 kcal/mole of
stabilization energy due to this effect.

Boone and Ashby have noted that Felkin's torsional strain,
Klein's orbital distortion theory and Ahn's antiperiplanar effect are
all electronic in nature, not steric, and suggest the term

"antiperiplanar factors" to include all three concepts."

attack

\
Q 0

0/e o

/

Figure 10. Ahn's antiperplanar effect.

IV. Hydride Reductions pf‘g - Alkyl (and Phenyl) Bicyclo|2.2,2|octan
- 2 - ones and 9,10 - Dihydro - 9 10 - (11 - Keto - 12 - Al kyl (and

Phenyl) ethano) - Anthracenes,

In 1973 Varech and Jacques reported "surprising" results for the
lithium aluminum hydride reduction of 3-alkyl (and phenyl)

bicyclo[2.2.2]octan-2-ones (Eq 2).33 Further studies were carried out

 

36,31

and extended to the analogous dibenzo system (Eq 3). Table 1

contains the results of the reductions.

H R H R H
Ll A lH (2)
c_i_s_
R H
is

I

a5Q M J; a! Cc»

g__ trans

 

Table 1. Percentages of alcohols from equations 2 and 3 ($trans =

100% - %cis). Reactions were run at 20°C in diethyl ether.

R = (Eg 2)_%cis (Eg él,%cis
ca, so 65
cazca, 32 49
ca<ca,)z 23 20
C(ca,)3 85 ' 77
c‘a, 95 100

The proportion of attack from the more hindered side is
unexpectedly very high. This proportion increases in the series
methyl, ethyl, isopropyl, just Opposite to what is expected on steric
grounds. It was proposed by the authors that an anisotropic inductive

effect was responsible. That is, the electron releasing ability of

 

21

the alkyl groups increases in the series, thus making the face of the
carbonyl opposite to those groups increasingly more negative. This in
turn would slow the rate of attack of anionic nucleophiles. The
pfbutyl group did not follow this trend because the "extra" methyl
group blocked the last reasonably easy pathway. The phenyl case is in
line with the theory since it is inductively an electron withdrawing
group and would have an activating effect.

My research is directed at probing these reactions further. The
bulk of this investigation is concerned with dissecting the
differences in the free energies of activation into their enthalpy and

entropy components.

41 _. _ :F
AAGAB - AAHih ThASAB

The difference in the free energies of activation is a function of the

product ratio (A/B).

AAGih = —RT 1n(A/B)
A systematic study of the changes in product ratios (A/B) with
temperature gave the necessary activation parameters TAAS; and AAHI
via the following equation.

ln(A/B) = -(AAH:B/R)(1/T) + AASih/R

This is significant since the models to date for asymmetric induction

 

22

are valid only for enthalpy controlled reactions. One should realize
that the activation entropies for complex metal hydride reductions are
generally between -20 and -40 e.u., thus accounting for up to 50% of
the free energy activation barrier at room temperature.‘

Solvent effects were also evaluated and a few different reagents
tested. In addition, the thermodynamic equilibration of the isomeric
alcohols was carried out.

I have chosen to study the dibenzo system (Eq 3) for two reasons:
(1) It is devoid of eclipsing interactions between vicinal hydrogens,
thus, there should be no small deformations of the bicyclic skeleton

and (2) analysis by NMR should be more easily interpreted.

 

 

RESULTS AND DISCUSSION

I. RESULTS

The results of complex metal hydride reductions of 9,10 - dihydro
- 9,10 - (11 - keto - 12 - alkyl (and phenyl) ethano) - anthracenes
are presented in Tables 2, 3, and 4. Results of reduction by lithium
£51 - .p - butoxyaluminohydride are presented in Table 5. Also in
Table 5 are the results of the thermodynamic equilibration of the
alcohol products, via the Meerwein - Ponndorf - Verley - Oppenhauer
reaction.

The changes in product ratio with temperature are indicative of
differences in A the enthalpies (AAHI) and entropies (AASI) of
activation. The following thermodynamic equations define the
relationship between the product ratio and the activation parameters

(Eqs 4,5 and 6).

AAG:F = -RT ln (pig/trans) (4)
AAG‘F = AAHt - TAAS'F (5)
ln (c_is_/trans) = -(AAH:F/R)(l/T) + hast/R (6)

It is asssumed that these are valid in the temperature range
studied." By plotting ln (gig/pgpps) versus 1/T one may derive the
activation parameters. The difference in the enthalpies of activation
was calculated from the slope of the line. The difference in the
entrOpies of activation was calculated from the intercept. These

plots are shown in Figures 11, 12 and 13. The activation parameters

23

 

24

are recorded in Tables 6, 7 and 8.

The product ratios were measured using NMR peak areas. This
procedure is described in the experimental section. The structures
were elucidated by examining the coupling constants and carrying out
homo - decoupling experiments. The use of a lanthanide shift reagent
helped to verify NMR peak assignments. .These assignments were
consistent with those reported by Varech and Jacques.”

All reductions were run with a large excess of reducing agent.
This was done to insure that the stereochemical outcome was the result

of the transfer of the first hydrogen only. Details are in the

experimental section.

Table 2.

diethyl ether.

-CH3CH3
‘-CH(CH3)3
_C(CH3)3

—C6ns

Table 3.

tetrahydrofuran.

% cis:trans

r»
k5

.ox
so
o:
on

50:50

22:78

80:20

see Error section

22°

k3

62:38

50:50

21:79

78:22

100: 0

% cis:trans

mfg
76:24
61:39
42:58

85:15

see Error section

2.2.“;
77:23
60:40
38:62

85:15

100: O

25

59:41

32:68

84:16

Product ratios from lithium aluminum hydride

Product ratios from lithium aluminum hydride

57:43

23:77

84:16

reductions

reductions

in

in

Table

4. Product

isopropyl alcohol.

2°C.

a

ratios

from

% cis:trans

.219
61:39
49:51
33:67

72:28

see Error section

22°

R:

62:38

50:50

29:71

73:27

100: 0

b The reaction was too slow.

26

sodium borohydride

A ratio of 74:26 was obtained at

51:49

25:75

74:26

reductions

in

Table

4. Product

isoprOpyl alcohol.

2°C.

a

ratios

from

% cis:trans

.232
61:39
49:51
33 :67

72:28

see Error section

22°

up

62:38

50:50

29:71

73:27

100: O

b The reaction was too slow.

26

sodium borohydride

A ratio of 74:26 was obtained at

25:75

74:26

17:83

reductions

in

27
Table 5. Results from thermodynamic equilibration via the MPVOb

reaction. Lithium tri - p - butoxyaluminohydride reductions.

i cis:trans (i 2%)8

 

.8 = M2199 .2 2
-CH3 6:94 63:37 90:10
-CH3CH, 3:97. 59:41 88:12
-CH(CH,)z (1% cis 54:46

-C(CH,), (1% cis 84:16

-C‘H5 1:99

a

see Error section
b Thermodynamic equilibration via the Meerwein - Ponndorf -

Verley - Oppenhauer reaction.

c Lithium tri - p - butoxyaluminohydride in tetrahydrofuran
from reference 37.
d Lithium ri - p - butoxyaluminohydride in diethyl ether.

28

Table 6. Activation parameters from Figure ll.a Lithium aluminum

hydride in diethyl ether.

‘3 = AAH; r6; callmole AAyI igggg_gg pp
-CH3 -33 +1.1
-CH2CH3 +200 +0.7
-CH(CH3), +966 +0.7
-C(CH,)3 +120 +3.0

a

see Error section.

Table 7. Activation parameters from Figure l2.a Lithium aluminum

hydride in tetrahydrofuran.

‘R = AAHI is; callmole AA§I rQy§§_pppy
-CH, -320 +1.4
-CH3CH3 +135 +1.3
-CH(CH3)2 +820 +1.7
-C(CH,)3 +25 +3.6

a

see Error section.

29

Table 8. Activation parameters from Figure l3.a Sodium borohydride in

isoprOpyl alcohol.

3 = AAHI r53 callmole AASI $0.33 e.u.
-CH3 -164 ' +0 .4
-CH3CH3 +190 +0.6
-CH(CH3)z +760 +0.8
-C(CH,), —140 +1.5

a

see Error section.

30

312“

LIA/H
__4.

2.0- fl‘Q “20

 

 

 

Ix?“

 

 

.0
Q
1

/

In (gig/trans)
I

R :
-2.0d

‘C(CH3)3 A\
-CH(CH3)2

[>000

-3. 0-

 

 

'400 I I I
0.0 1.0 2b .330 430 52
(l/T) x 10‘3

I

Figure 11. Lithium aluminum hydride in diethyl ether.

31

 

 

3-0 "l
2.0-
<><> —-<> @—
-i E //E
-13 /

In (us/trans)
I .
:o
[>
[>

 

O -c (CH3)3
i [:J ‘13ft3
‘200‘ O
A

‘3.C}—

 

 

H OH

trans

 

‘4‘0 ' F “n 1 I
00 L0 2.0 3.0 4.0

(l/T) x 10*3

I

Figure 12. Lithium aluminum hydride in tetrahydrofuran.

32

3.0
l R g
‘ O “C (CHJIJ
E] °C+t3
2.0"
C) ‘Cifi207fl3
. Z: ~Cii(Cit3)2
‘________________
1.0- -0 -0 -<>

 

 

00- “CD—Q “‘6 \o‘
‘23

if, A

t \A

‘Efi - ‘\\\\\\£>

8

is d ‘\\\\\\\\\\\‘z:

I

I“

<3
I

   

 

 

NaBH4
iFPrOH
“4.0 r r I I i I ' l f T
0.0 [.0 2.0 3.0 4.0 5.0
“/77 x 173

Figure 13. Sodium borohydride in isopropyl alcohol.

33

II. DISCUSSION

The composition of a complex metal hydride and the mechanism of
its reduction reactions is fundamental to any discussion concerning
the stereoselectivity of such reactions. For that reason, a brief
account of these tapics is given.

A. Thy Composition pf Lithium Aluminum Hydride ip Diethyl gfippg
ppd Tetrahyrdofuran

Most of the information concerning the composition of lithium
aluminum hydride in diethyl ether and tetrahydrofuran has come from
conductance and molecular association (ebullioscopic) studies. The
equivalent conductance of lithium aluminum hydride in diethyl ether is
very small and concentration independent. For the same species in
tetrahydrofuran, however, the equivalent conductance is substantially
larger and increases with increasing concentration.‘0 This indicates
that lithium aluminum hydride in diethyl ether and tetrahydrofuran
exists as contact ions and solvent-separated ions respectively.

Molecular association studies of lithium aluminum hydride in
diethyl ether and tetrahydrofuran were done by ebullioscopic molecular
weight studies. In tetrahydrofuran, lithium aluminum hydride has an
association value (defined as the ratio of experimentally determined
weight to formula weight) of 1.0 in dilute solution (0.05 m). This
value increases with increasing concentration to a limiting value of
1.8 (0.5 m and greater). The association value for lithium aluminum
hydride in diethyl ether is somewhat higher, starting at 1.7 in dilute

solution (0.05 m) and increasing to 2.3 at the highest concentration

(0.45 m).

 

 

 

34

Ashby4° has interpreted the results for tetrahydrofuran solutions
as being due to solvent-separated ion pairs in dilute solution, while
triple ions are formed at higher concentrations. Lithium aluminum
hydride in diethyl ether appears to exist as contact ion pairs in
equilibrium with higher aggregates. The better solvating ability of
tetrahydrofuran and its higher dielectric constant (7.6 versus 4.34)

is consistent with these interpretations.

B. The Mechanism pf Lithium Aluminum Hydride Reduction pf

The kinetics of ketone reductions by lithium aluminum hydride has
been followed spectrOphotometrically by observing the disappearance of
the n 9 n. band in the ultraviolet. In tetrahydrofuran the reduction
is first order in both ketone and lithium aluminum hydride.‘1 In
diethyl ether the reduction is first order in ketone, but about 1/2
order in lithium aluminum hydrids.‘° In diethyl ether this indicates
participation of both monomeric and dimeric species.

The role of the metal ion appears to be important but not
essential for reaction. For example, Ashby has reported that lithium
aluminum hydride is about ten times more reactive than sodium

1

borohydride.‘ He has also used tri - p - octyl -‘p - prepylammonium

aluminum hydride as an effective reducing agent. Chelating agents,
such as N, N, N', N", N", N. - hexamethyltriethylenetetraamine (for
lithium ion) and 18 - crown - 6 (for sodium ion), do not inhibit the
reduction. However, it appears that the metal ion is involved in the

transition state, especially since different metals can show marked

changes in the stereochemistry. The fact that mesityl phenyl ketone

35

is about 40 times more reactive in diethyl other than in
tetrahydrofuran can also be attributed to the metal ion. That is,
since tetrahydrofuran can solvate cations to a greater extent than
diethyl ether, the ketone can compete more favorably for the metal ion
in diethyl ether.

The enthalpies of activation for the reduction of mesityl phenyl
ketone by lithium aluminum hydride and sodium aluminum hydride are
10.5 and 18.1 kcal/mole respectively.‘1 The corresponding entropies of
activation are -26.2 e.u. for lithium aluminum hydride and -5.4 e.u.
for sodium aluminum hydride. This is consistent with cation
association of the carbonyl oxygen in the transition state. The
lithium cation can polarize the carbonyl to a greater extent, thus
lowering the enthalpy of activation relative to the sodium cation.
The more negative entropy of activation for lithium is consistent with
the ability of that cation to order the ketone and a number of solvent
molecules to a greater extent than the sodium cation. Ashby has also
suggested that the lithium cation may be associated with one of the
non - reacting hydrogens of lithium aluminum hydride, whereas the
sodium cation would not.

Kinetic isotOpe effects have been calculated for the reduction of
mesityl phenyl ketone in diethyl ether and tetrahydrofuran. In both

‘1’41 Other than demonstrating that the

cases the kH/kD was about 1.3.
hydride is transferred in the transition state, it is difficult to
interpret these results because of secondary isotope effects and the

uncertainty of the Al-H(D)-C bond angle.

Ashby has prOposed the following mechanism based largely on these

 

 

36

observations (Eq 7).‘1

S
\lfl AIH: + C=0 =———‘
./
S
S a
\ ..
/CK\ AIH. + 8 ——~
3 o\
\C/
/
“i
l s)
[3,3 8 H H
\C=0/§\S \ ./ \ /
, , - —‘ S + LI Al (7)
/: : H / \O/ \
H-A'I/ S I H
/\ C-—H
H H /\

 

 

The role of alkoxyaluminohydrides, has been investigated by
Smith‘3 and Brown.‘3'4‘ It has been established by Smith that the
transfers of the second, third, and fourth hydrogens from aluminum
hydride are slower than the first transfer. However, these rate
differences are not great enough to disregard the role of the
alkoxyaluminohydrides, and may be significant when the ketone/lithium
aluminum hydride ratio is close to one. Smith's studies were done on
tertiary butoxy derivatives. Brown has determined that secondary
alkoxyaluminohydrides (products from ketone reductions)
disproportionate rapidly to regenerate aluminum hydride plus the

tetraalkoxyaluminate. If the alkoxy group is not secondary, the

disproportionation is much slower.

 

37

C. (Th2 Mechanism pf Sodium Borohydride Reduction pf Ketones ip
Alcohols

The composition of sodium borohydride in hydrolytic solvents is
not known. It is generally assumed that it is strongly solvated and
it is known to react with some alcoholic solvents to form alkoxy
intermediates. Brown has reported that it is stable in isoprOpyl
alcohol at temperatures as high as 60°C.‘5

The rate of reduction has been determined to be first order in

“"’ Wigfield“, has

ketone and first order in sodium borohydride.
also determined the rate to be 3/2 order in alcohol. This has been
explained by assuming that an alkoxide (from ionization of the
solvent) and a solvent molecule are involved in the transition state.
Wigfield reported that the products of reduction are the free alcohol,
derived from the ketone, and the alkoxy borohydride, in which the

alkoxy group is derived from the solvent. He has proposed the

following mechanism (Eq 8).

RO ------ B ----- H ----- c--—--=o ----- H ----- 0R (.9)

 

 

The sodium ion is not included in the transition state. Studies“"’,

including the removal of sodium cation by cryptatesso, have shown that

this is so.

Unlike lithium aluminum hydride and its alkoxy derivatives, the

 

38

transfer of the second, third, or fourth hydrogen of sodium

46,51

borohydride is faster than the initial reduction step. Secondary

and tertiary alkoxy derivatives of sodium borohydride are fairly
stable to disproportionation.’z"3
D. 132 Activation Parameters
The activation parameters calculated define the relationships

- AH?

enthalpy favors attack from the least hindered side to give the cis

a: - a:
and (Ase-ii Astrans)°

(Aniis

rans) When R is a methyl group,

 

isomer. When R is an ethyl or isopropyl group, there is a crossover
and enthalpy favors attack of the more hindered side. Attack of the
more hindered side is favored more for isOpropyl than for ethyl. In
the case of pfbutyl, there is a slight preference for attack from the
more hindered side in diethyl ether, essentially no preference in
tetrahydrofuran, and a slight preference for attack from the least
hindered side in isopropyl alcohol.

In all cases studied, entrapy favors attack from the least
hindered side to give the pig isomer. When R is a prbutyl group, the
attack is favored strongly.

No activation parameters could be calculated for the phenyl
substituted case. The cis isomer is essentially quantitatively
favored in all three systems.

1. Ihg.gplg pf Enthalpy

It is interesting to note that the trend observed by Varech and
Jacques" and later by Felkin and co-workers37 is indeed due to
enthalpy factors. Felkin and co-workers proposed an anisotropic

inductive effect to rationalize the stereoselectivity. An effect such

.—__. §__._.._ __-

 

 

39

as this would be apparent in the enthalpies of activation.
Intuitively, one would expect the two faces of any of the ketones
studied to have slightly different electron densities. The importance
of this in the transition state would be difficult, at best, to
predict. It should be noted in this regard, that calculations of the
electron density on the diastereotopic faces of carbonyls have been
done." These calculations predict just the opposite to an anisotrOpic
inductive effect when comparing a methyl to a hydrogen. That is, the
face pppi to the methyl group has less electron density than the face
pppi to the hydrogen. It would also require that the transition state
be very "reactant like". In other words, the orbitals of the n system
must still be largely intact.

It is noteworthy to point out that the results in this study do
not agree with Ahn's antiperiplanar effect. An antiperiplanar effect
would also show up in the enthalpies of activation. The geometries of
the transition states in these ketones would most likely not be
identical to those proposed by Ahn, but would appear to be similar.

It would also seem unlikely that any small distortions in the
bicyclic skeleton could account for the stereochemistry. There are no
eclipsing interactions of substituents on the ring system, except for
the carbonyl oxygen and the bridgehead hydrogen. It would appear that
such an interaction would not interfere, since in the ground state
carbonyls prefer an eclipsing interaction.‘4

There are some features in this system (bicyclo[2.2.2]). however,

which set it apart from other ketones. First of all, the incoming

hydride must eclipse another bond in the transition state. This would

 

4O

certainly raise the enthalpy of activation compared to acyclic or
simple monocyclic ketones. Another feature is that the oxygen of the
carbonyl must, in the course of the reaction eclipse another bond.
Only if the carbonyl remained trigonal in the transition state, which
seems unlikely, could eclipsing be ignored. Again, this eclipsing
would add to the enthalpy of activation, especially since the oxygen
would need a metal counterion and/or a high degree of salvation.
Thus, the overall enthalpies of activation would, to a first
approximation, most likely be greater than in more flexible systems.
Even without eclipsing considerations, the system itself is fairly
hindered. It is reasonable to assume that the transition states would
be further along the reaction coordinate and thus resemble products
more so than for "normal" ketones.

A comparison of the eclipsing interaction of a hydride with a
methyl group versus a hydride with another hydrogen is in order.
Competitive rate studies of the two molecules in Figure 14 have been

carried out by Eliel and Senda.17

W1 W111

l.2£§ l.0

Figure 14. 4 -‘£ - butylcylcohexanone and 2,2 - dimethyl - 4 - p -

butylcyclohexanone.

They determined that equatorial attack of 4-pfbutylcyclohexanone

 

 

41

by lithium aluminum hydride is only 1.25 times faster than equatorial
attack of 2,2-dimethyl-4-prbutylcyclohexanone. If tetrahydrofuran is
used as the solvent instead of diethyl ether, this value is even
smaller (1.15). It is apparent that the axial methyl group does not
affect equatorial attack nearly as much as one might expect based on
steric hindrance. In fact, when lithium pgiéprbutoxyaluminohydride is
used to reduce these two compounds in tetrahydrofuran, equatorial
attack is still favored for 4-prbutylcyclohexanone by only a factor of
1.5.

Returning to the bicyclo[2.2.2] system, an analagous type of
situation is occurring. The torsional strain due to an eclipsing
hydrOgen is of the same order of magnitude as the steric hindrance due
to a methyl group. Furthermore, in the series methyl, ethyl,
isopropyl, the situation should not change significantly since there
will always be a pathway which would in a sense ignore the "extra"

methyl groups. This pathway is depicted in Figure 15.

 

Figure 15. The bicyclo[2.2.2] system.

On the side of the carbonyl opposite to the attack, the oxygen

begins to eclipse either a hydrogen or an alkyl group. There is no

 

 

 

42

question that it would prefer to eclipse a hydrogen. The
thermodynamic equilibration ratios bear this out (Table 5).
Furthermore, these values indicate that the "extra" methyl group in
the series is noticed. In other words, the reason that complex metal
hydrides attack the more hindered side is the unfavorable eclipsing
interaction of the oxygen with the alkyl group when the hydride
attacks from the least hindered side. This implies that the
interaction of the oxygen with the alkyl group is more significant
than the interaction of the hydride with the alkyl group. In fact
this is not unreasonable since complex metal hydrides invariably act
as "small reagents" when compared to other nucleOphiles. Equations 9

through 14 illustrate this point.55

Q)

00)

 

{l l)

02)

’ 0H
/0
W HCECMQBF CECH (l3)
EIZO
0H
CHJMgBr Ma’s (14)
57‘20

 

 

 

43

If the oxygen-alkyl group interaction is greater than the metal
hydride-alkyl group interaction then the transition state resembles
the products to a certain extent. There is evidence to suggest that
rehybridization of the carbonyl to sp3 is significant. For example,
Hammett studies have shown that reaction of sodium borohydride with

"57 and acetophenones” gives large positive

substituted fluorenones‘
p values of 2.65 and 3.06 respectively. The corresponding value for
substituted benzophenones with lithium aluminum hydride is 1.95.‘2
Lamaty and Geneste have argued that the high values for the sodium
borohydride reductions indicate that such reactions pass through
product-like transition states."

Wigfield has reported an interesting difference between the
sodium borohydride reductions of unhindered and hindered
cyclohexanones.‘7 Specific isokinetic plots for axial and equatorial
attack were constructed. (The enthalpy of activation for axial attack
was plotted against the entropy of activation for axial attack. The
same was done for equatorial attack.) These plots showed a clear cut
difference in isokinetic behavior between hindered and unhindered
cyclohexanones. It has been argued by Leffler that such a distinct
difference corresponds to a qualitatively different ‘transition
state.‘0 Isokinetic relationships have been criticized‘l, but
Wigfield's plots seem to indicate a different transition state for
hindered cyclohexanones versus unhindered ones.

Wigfield has also noted a similar phenomenon in a kinetic isotope

effect study.‘z Reduction of a number of ketones with lithium tri - p

- butoxyaluminohydride(deuteride) gave isotOpe effects ranging from

 

44

1.0 to 1.5. The corresponding isotope effects for two highly hindered
ketones, 3,3,5,5 - tetramethylcylcohexanone and 4 - acetyl - 3,3,5,5 -
tetramethylcyclohexanone are inverse, 0.79 and 0.70, respectively. An
inverse isotOpe effect can be explained by invoking a late transition
state since a weaker Al-H(D) bond is being replaced by a stronger

C-H(D) bond. The isotope effects for sodium borohydride reductions

59,63,64

are inverse (kn/kn = 0.59 - 0.77). The inverse isotope effects
are difficult to interpret since the primary effect is masked by three
secondary effects. Both an early and a late transition state have

”“ Lithium aluminum hydride has been reported to have

been proposed.‘
an isotope effect which is normal and small (kH/kD = 1.3).‘3 Again,
this result is difficult to interpret.

The trend in stereoselectivity for the bicyclo[2.2.2] system can
be explained by invoking a late transition state. Product stabilities
are reflected in the transition state. The prbutyl case is consistent
with this because the hindered side is effectively blocked and steric
approach control is dominant. However, even in this case a
significant prOportion of attack is from the hindered side, thus
product stability is still important.

The phenyl substituted system appears out of line since it favors
the .21; product almost exclusively. In examining the following two
systems (Eqs 15,16), however, it is apparent that a n system can

hinder attack.3"‘5

 

45

C)
LIAIH4 05)
EQO

@700»; 307°
H OH
O NaBH HO & (I 6)
/ i'—Pr0H /
. Iqu

Presumably this is due to a coulombic repulsion between the high
electron density of the n system with the negatively charged
nucleophile.
2. TE; 3215 pf Entropy

Entropy favors the pig product in all reductions. This effect is
more pronounced for the prbutyl group than for the other alkyl groups.
The difference in results for diethyl ether and tetrahydrofuran as
solvent is significant. These results are most likely due to solvent
entropy effects. That is, in the pig case the carbonyl oxygen starts
eclipsing the alkyl group; the disordering of solvent molecules around
the oxygen metal bond is greater than in the pgppg case. This
explains why the pfbutyl group has such a larger AASI: It cannot
rotate out of the way as ethyl and isopropyl can. This also explains
the difference between reductions in tetrahydrofuran and diethyl
ether. In tetrahydrofuran, the cation is solvated to a greater degree

and thus the effect is magnified. In isopropyl alcohol, the cation is

46

not involved in the transition state, thus salvation is apparently not
as important. This interpretation is consistent with a late
transition state.

It is difficult to assess the role of rotational entropy for the
alkyl groups. A late transition state would hinder rotation for both
routes of attack.

3. Conclusions

The results of this study show that the observed trend in
stereoselectivity in the bicyclo[2.2.2]-octanone system is due to
enthalpy factors. The enthalpy factors which control this selectivity
can be attributed to product stability. This implies that the
transition state resembles products to a certain extent. It does not
imply that this is true in all cases. This appears to be a special
case.

The results of this study also show the importance of considering

salvation in evaluating entropy effects.

 

EXPERIMENTAL

I. General Methods

A. fippgk solutions pf complex ppppl hydrides

Diethyl ether and tetrahydrofuran were dried over lithium
aluminum hydride, distilled and stored over molecular sieves (4A).
Is0pr0pyl alcohol was dried over calcium sulfate and distilled. Stock
solutions of lithium aluminum hydride were prepared as follows.
Lithium aluminum hydride (Aldrich, 95%) and the solvent were stirred
together for 1h under an argon atmosphere. The crude slurry was then
siphoned through a U-shaped glass tube, which contained a glass wool
plug, into a Schlenk filter tube packed with CeliteR. After
filtering, the clear solution was stored under argon. Lithium
aluminum deuteride and lithium ppitprbutoxyaluminohydride solutions
were made by stirring them, under an argon atmosphere, in solvent.
Sodium borohydride solutions were made by stirring the hydride in
isoprOpyl alcohol until all material was dissolved (several hours, a
drying tube was used to keep moisture out). The solutions made in
this manner were between 0.05 and 0.10 molar.

B. Temperature Control

Reductions done at -77°C were controlled by using a dry
ice-acetone bath. A dry ice-carbon tetrachloride bath and an ice
water bath were used for reductions at -22°C and 2°C, respectively.

Temperatures at reflux were controlled using an ordinary oil bath.

47

 

48

C. General Procedure jg; Reductions

Method A: A solution of the complex metal hydride (8-12 mL,
0.5-1.0 mmol) was put into a 25 mL round-bottomed flask equipped with
a magnetic stir bar, gas inlet valve and septum. The solution was
allowed to come to thermal equilibrium under an argon atmosphere. In
a separate 25 mL round-bottomed flask equipped with a gas inlet valve
and septum were combined the ketone (20-30 mg, 0.07-0.12 mmol) and 5
mL of solvent. If the reduction was run below room temperature the
ketone solution was also cooled to the reaction temperature. The
ketone solution was then slowly added (10-15 min) to the complex metal
hydride solution by using a syringe. If the reduction was run at
reflux, a reflux condensor was connected to the flask. The following
reaction times were sufficient for complete reduction: refluxing
diethyl ether, tetrahydrofuran and isopropyl alcohol (10 min); 22°C
diethyl ether, tetrahydrofuran and isopropyl alcohol (15 min); -22°C
diethyl ether and tetrahydrofuran (0.5h), isopropyl alcohol (1h);
-77°C diethyl ether and tetrahydrofuran (2h), isopropyl alcohol (6h)
and very slow for the p-butyl system. The solution was hydrolyzed
with 20 mL of water and extracted with three 30 mL portions of ether.
The ether extracts were combined and dried over anhydrous magnesium
sulfate, filtered and the solvent removed under reduced pressure. The
crude product mixture (>90% yield) was used for NMR spectra.

Method B: Using the identical experimental set—up as Method A,
the metal hydride solution was added to the ketone solution (inverse
addition). Care was taken to initially add the metal hydride very

slowly. Other than that, reaction times were the same.

 

49

Method A was used throughout the study. Method B was used
periodically to see if inverse addition would affect product ratios.

No such affect was detected.

II. Syptheses

2-Nitroethanol (1)

2-Nitroethanol was prepared according to the procedure of
Noland.“ Paraformaldehyde (75.0 g, 2.5 mol) and nitromethane (1.70
kg, 28.0 mol) were combined in a 3 L, three-necked round-bottomed
flask equipped with a mechanical stirrer, a dropping funnel and a
thermometer. The mixture was vigorously stirred and a solution of
potassium hydroxide in methanol (3 ‘3) was added dropwise until an
apparent pH of 8 (pH paper). The paraformaldehyde completely
dissolved and the solution temperature increased to 40°C. Stirring
was continued for 1h after which concentrated sulfuric acid (18 M) was
added drapwise until an apparent pH of 4 (pH paper). Stirring was
continued for an additional hour. Precipitates, formed during the
acidification, were removed by filtration and the nitromethane removed
under reduced pressure. The crude 2-nitroethanol was vacuum distilled
with an equal volume of diphenyl ether (heat dispersing agent), bp
59-650C (0.20 mm). The two phase distillate was separated and the
crude 2-nitroethanol (114.0 g, 50%) was used without further
purification; ‘H NMR (CDCl,) 5 4.60-4.25 (2a, br m), 4.20—3.80 (2a, br

m), 3.50 (1H, br s).

 

50

l-Nitro-Z- ro anol (2)

l-Nitro-Z-propanol was prepared according to the procedure of
Hard and Nilson." Nitromethane (183.0 g, 3.0 mol) and 155 mL of water
were combined in a 2 L, three-necked round-bottomed flask equipped
with a mechanical stirrer, drapping funnel and a reflux condensor. To
this was added enough sodium carbonate to give an apparent pH of 10
(pH paper). A solution of acetaldehyde (132.0 g, 3.0 mol) and 120 mL
of water was added dropwise to the solution over a 90 min period,
during which time the reaction temperature increased to 60°C. Sodium
carbonate was added periodically to maintain a pH of about 10. The
solution was stirred for 2h and then let stand for an additional 2-3h.
The contents of the flask were poured into a separatory funnel and
extracted with three 200 mL portions of ether. The combined ethereal
solution was dried over anhydrous sodium sulfate, filtered, and the
solvent removed under reduced pressure to afford the crude
1-nitro-2-propanol. Vacuum distillation gave pure l~nitro~2~propanol
(214.2 g, 2.04 mol, 68%), bp 76—77°c (1.5 mm). [1it" bp 86—89°c (8
mm)]; ‘H NMR (C001,) 6 4.60—4.25 (3H. br m). 3.20 (1H, br s), 1.21

(3H. d).

1-Nitr012-butppplqigl

1-Nitro-2-butanol was prepared in the same manner as described
for the preparation of 1-nitro-2-propanol.‘7 The crude nitro alcohol
was purified by vacuum distillation, 64% yield, bp 75-76°c (1.1 mm).
[lit‘7 98-99°c (9 mm)]: ‘H NMR (CDCl,) 8 4.47-3.93 (3a, br m). 3.02

(1H, br s). 1.80-1.32 (2H, m), 1.03 (3H, t).

 

51

3-Methyl-l-nitro-2-butanol (4)

Using the same reaction conditions as previously described for
1-nitro-2-pr0panol (2) and 1-nitro-2-butanol (i). 2-methylpr0panal
(iso-butyraldehyde) was condensed with nitromethane. The crude nitro
alcohol was purified by vacuum distillation, 56% yield, bp 79-80°C
(1.0 mm); 1H NMR (cnc1,) 8 4.48-4.00 (3a, br m). 3.05 (1H, br s).

1.85-1.36 (1H. m). 1.23 (3H. d). 1.12 (3H. d).

Nitroethylene (5)

Nitroethylene was prepared according to the procedure of Buckley
and Scaife.fl 2-Nitroethanol (100.0 g, 1.10 mol) and phthalic
anhydride (180.0 g, 1.22 mol) were combined in a vacuum distillation
apparatus with a short fractionating column and a stir bar. Under
reduced pressure (80 mm) the mixture was heated to approximately 170°C
(oil bath temp 170-175°C). A two phase distillate of water and
nitroethylene was collected. The nitroethylene was separated from the
water, dried over anyhdrous magnesium sulfate, filtered and
redistilled (34.5 g, 43%). bp 42-44°c (88 mm), [1it" hp 38—39°c (80
mm)]; 1H NMR (CDCl3) 6 7.55-6.68 (m). The nitroethylene was used

immediately.

1-Nitropropepp‘jgl

l-Nitro-Z-propanol was dehydrated as described for the
dehydration of 2-nitroethanol." The two phase distillate of crude
product and water was collected at an oil bath temperature of 180°C

(80-100 mm). The layers were separated and the crude product was

 

52

dried over anhydrous magnesium sulfate, filtered, and vacuum

1

distilled, 65% yield, hp 50-51°c (80 mm). [lit" hp 54°C (88 mm)]; H
NMR (CDC13) 6 7.60-6.67 (2H, m). 1.88 (3H, d). l-Nitropropene
prepared in this manner is predominantly the ppgpg isomer based upon
comparison with identical spectra published by Baskov and
co-workers." The pig isomer is present in about 5% by proton NMR

integration."0

pgpps-l-Nitro-l-bppgpg‘11)

ppppg-l-Nitro-l-butene was prepared in the same manner as
nitroethylene and l-nitrOpropene. Heating to an oil bath temperature
of 180°C (80-100 mm) resulted in only partial dehydration with
significant amounts of 1-nitro-2-butanol being collected. After
removing the water from the two phase distillate, the mixture of
product and starting material was resubjected to the dehydration
conditions. The crude nitro alkene was purified by vacuum
distillation, 47% yield, hp 58-59°c (75 mm); 1H NMR (CDCl,) 8
7.50-6.65 (2H, m). 2.58-2.02 (2H. m). 1.10 (3H, t). l-Nitro-l-butene
prepared in this manner is assigned the ppppp configuration since the
olefinic pattern in the proton NMR is similar to those published by

69
Baskov and co-workers.

trans-3-Methyl-l-pitro-l-butene (8)
The dehydratrion of 3-methyl-1-nitro-2-butanol was carried out as
described for the dehydration of l-nitro-Z-butanol. The nitro alkene

was purified by vacuum distillation, bp 54-55°C (55 mm); 1H NMR

53

(CDCl,) 6 7.28 (1H, d of d, J= 16 and 7 Hz), 6.98 (1H, d, J= 16 Hz),

2.64 (1H, octet). 1.00 (6H, d).

trans-B-Nitrostyrene (9)

pgppgrB-Nitrostyrene was prepared by the method of Worrall."1
Nitromethane (61.0 g, 1.0 mol), benzaldehyde (106.0 g, 1.0 mol) and
100 mL of methanol were combined in a 2 L, three-necked round-bottomed
flask equipped with a mechanical stirrer, dropping funnel and
thermometer. After cooling (0°C) and with stirring, a solution of
sodium hydroxide (42.0 g, 1.05 mol) and ice water containing crushed
ice (100 mL total volume) was slowly added at such a rate that the
temperature was kept at 10-15°C. During the addition a white
precipitate formed and stirring became difficult (20 mL of methanol
was added to aid stirring). After 15 min of standing, 600 mL of ice
water containing crushed ice was added. The cooled solution (0-5°C)
was then immediately added, with stirring, to 500 mL of hydrochloric
acid (7.2 M). A yellow crystalline mass was formed as the solution
came in contact with the acid. The solid mass was suction filtered
and washed several times with water. The crude nitrostyrene was
purified by two recrystallizations from ethanol (110.3 g, 0.74 mol,
74%), mp 57-58°c, (lit'1 mp 57-58°C): ‘H NMR (CDCl,) 5 7.29 (5H, m).

7.21 (1H, d, J= 14 Hz). 6.93 (1H, d, J= 14 Hz).

9,10-Dihydro-9,10-(trpps-ll-pitro-lZ-methylethpno)-anthracene (10)
9,10-Dihydro-9,lO-(trans-11-nitro-12-methylethano)-anthracene was

prepared according to the method of Noland and co--workers."z

54

Anthracene (30.0 g, 0.168 mol) and 50 mL of prdichlorobenzene were
combined in a 200 mL round-bottomed flask equipped with a magnetic
stir bar and addition funnel. The flask was cooled (0°C), evacuated
and purged with argon. 1-Nitropropene (3.7 g, 0.042 mol) and 15 mL of
pfdichlorobenzene were quickly added to the addition funnel under an
argon atmosphere. The solution was heated to reflux and with
stirring, the nitro alkene was added dropwise over a 5 min period.
The solution was refluxed for 80 min and after cooling the solvent was
removed under reduced pressure. The residue was dissolved in boiling
benzene and upon concentrating and cooling, anthracene was removed.
The crude residue was chromatographed on alumina (300 g) packed wet
with petroleum ether (60-70). Elation with 10% ether-petroleum ether
(60-70) removed anthracene. Elution with methanol gave the crude
product (6.95 g. 0.0262 mol, 62%), mp 115—119°c. Two
recrystallizations from ethanol gave pure adduct (6.50 g, 0.0245 mol,
58%), mp 122-123°c, (lit" mp 122.5-123.5°C); 1H NMR (CDCl,) 8
7.28-6.88 (8H, m). 4.80 (1H, d. J= 2.6 Hz). 4.10 (1H. d of d, J= 4.6
Hz and 2.6 Hz). 3.95 (1H, d, J= 2.2 Hz). 2.64 (1H, m), 1.02 (3H. d. J=

7.0 Hz).

 

 

55

9,10-Dihydro-9l10-5ll-Keto-lngethylethanoz-Anthracene.Llll

9,10 - Dihydro - 9,10 - (nggg - 11 - nitro - 12 - methylethano)
- anthracene (2.45 g, 9.24 mmol) in 250 mL of ether was combined with
225 g of activated basic silica gel"3 (chromographic grade). Ether
was removed under reduced pressure and the dry silica gel was warmed
to 60°C for 7-8h. After cooling, the silica gel was eluted with ether
to give 1.86 g of crude 9,10 - dihydro - 9,10 - (11 - keto - 12 -
methylethano) - anthracene. The solid was chromatographed on alumina
(55 g) with 50% ether-petroleum ether (60-70) to give the pure ketone
(1.44 g, 6.15 mmol, 67%). mp 2121-122°c, (11t” mp 122°C); 1H NMR
(CDC13) 5 7.34-6.88 (8H. m). 4.78 (H—lO, s). 4.28 (H-9, d, J= 2.4 Hz).
2.29 (H-12, d of q, J= 7.3 and 2.4 Hz). 0.94 (3H, d, J= 7.3 Hz). The
residue. which was left on the chromatography column was mostly the

starting nitro adduct which was recycled.

 

9IlO-Dihydro-9.10-(trags-ll-Nitro-lg-Ethylethanoz-Anthracene (12)

 

The Diels-Alders reaction of anthracene and
tgggg-l-nitro-l-butene was carried out in the same manner as described
for anthracene and 1-nitropropene. The crude adduct was
chromatographed, 7% yield, but not purified further; 1H NMR (CDCl,) 8

7.42-7.00 (33, m). 4.92 (H-10, d), 4.38-4.17 (H411 and H49,

 

 

 

56

superimposed), 2.54 (H-12, m), 1.24 (2H, quintet). 0.94 (3H, t).

H\,CH2CH3
HHCH‘QCHJ OZN H
at. + -— A
999 H a ,2 Q
.7 —

9,10-Dihzdro-9,19-(ll-Keto-l2-Eth lethano -Anthracene (13)

 

The oxidation of the nitro adduct to the ketone was carried out
on activated basic silica gel as described for the
11-nitro—12-methylethano adduct, 1;. 59% yield, dec. 9s—11o°c, mt”
dec. 95-1100C); 1H NMR (CD013) 8 7.44-7.00 (8H. m). 4.77 (H-lO, s).

4.46 (H-9, d), 2.10 (H-12, octet), 1.62 (2H, m), 1.03 (3H, t).

H C H2CH3

 

H CH(CH3)2

+ _ .
@@@ OzN > H a Q
8 I4

--' ”meg. ...,, -

57

9,10-Dih1dro-9,10-(traps-ll-Nitro-IZ-iso-Propylethano)-Anthracene (142

The reaction between anthracene and

 

trans-3-methyl-l-nitro—l-butene was carried out as described for
anthracene and 1-nitr0pr0pene. The crude adduct was chromatographed,
4% yield, but not purified further; 1H NMR (CDCl,) 8 7.40-6.95 (83.
m), 4.82 (H910, d), 4.32-4.13 (H911 and H99, superimposed), 2.52

(H912, m), 1.75 (1H, m), 1.10-1.00 (6H, two d superimposed).

9,10-Dihydro-9,10-gll-Keto-lz-iso-Propylethanoz-Anthracene (15)

The oxidation of the nitro adduct to the ketone was carried out
on activated basic silica gel as described for the
11-nitro-12-methy1ethano adduct, 1;, 55% yield, mp 128-129°c, (lit”
mp 129°C); 1H NMR (cnc1,) 5 7.35-6.92 (8H. m). 4.66 (H-lO, s), 4.48
(H99, d). 2.00 (H912, d of d). 1.40 (1H, m). 0.97 (3H, d). 0.76 (3H,

d).

 

 

9I10-Dihydro-9,10—(trags-ll-Nitro-lZ-Phen lethano -Anthracene (l6)

9,10-Dihydro-9,10-(trans-ll-nitro-lZ-phenylethano)-anthracene was

 

synthesized as described by Noland and co-workers.12 B-Nitrostyrene
(6.0 g, 0.040 mol), anthracene (14.6 g, 0.088 mol) and 100 mL of

xylenes were combined in a 250 mL round-bottomed flask equipped with a

 

58

magnetic stir bar and a reflux condensor. Under an argon atmosphere
the solution was refluxed for 20h. After cooling, the solvent was
removed under reduced pressure. The residue was extracted six times
with boiling ether (50 mL each). The combined extracts were
concentrated and the crude residue chromatographed on alumina (300 g)
with 10% ether-petroleum ether (60-70) to remove traces of anthracene.
Elution with methanol, followed by recrystallization from ethanol gave
pure _1_6 (4.91 g, 0.015 mol, 38%), mp 150-151°c. (111:3’ mp ISO-151°C);
‘H NMR (0001,) 6 7.40-6.78 (11H, m), 6.60-6.35 (2H, m), 5.10 (a—10, d,
J= 2.7 Hz). 4.90 (H911, d of d. J= 4.7 and 2.7 Hz). 4.30 (H99. d. J=

2.2 Hz), 3.91 (H-12, d of d, J= 4.7 and 2.2 Hz).

H CsHs
~06 H>=<CGH5 H
+ ——_x
5’ H a Q
9 I6

H C5H5 H C5H5
GEN F1 <3 ‘l'r
, - rR'

59

9 lO-Dih dro-9 10- 11-Keto-12-Phen lethano -Anthracene (11)

Using the same procedure as described earlier for 11, the
Diels-Alder adduct 16 was converted to 9.10 - dihydro - 9,10 - (11 -
keto - 12 - phenylethano) - anthracene. The pure ketone was recoved
in 65% yield, mp 159-160°c. (11t” mp 157—159°C); 1H NMR (c001,) 6
7.50-7.00 (11H, m), 6.58-6.49 (2H, m). 4.96 (H-10, s). 4.49 (H-9, d,

J= 2.3 Hz), 3.61 (H912, d, J= 2.3 Hz).

9,10-Dihydro-9,10-(11- itroethanoz-Anthracene (18)

9,10-Dihydro-9,10-(119Nitroethano)-anthracene was synthesized by
the same method as described for 19. Nitroethylene was freshly made
or distilled just before use. The crude adduct was purified by
chromatography and recrystallized as previously described for 1;, 39%
yield, mp 109-110°c, (11t7’ mp 109-110°C); 1H NMR (0001,) 6 7.40-7.05.
(8H. m). 5.06 (H-10. d. I= 3.0 Hz). 4.80 (H911. m). 4.41 (H-9, t, I=
2.7 Hz). 2.52 (H912, d of t, J= 13.7 and 3.0 Hz). 2.27 (H-12, d of d

of d, J= 16.5, 9.5 and 2.7 Hz).

@99 52:>:<: _‘

9 lO-Dih dro-9 10- 11-ketoethano -Anthracene (l9)

 

The conversion of the nitro adduct to the corresponding ketone

has been described, 11. The crude ketone was chromatographed on

 

60

alumina (50 g/ 1.5 g ketone) with 50% ether-petroleum ether (60-70) to
give pure 12. mp 150-151°C, (1it” mp 151°C; ‘3 NMR (CDC1,) 6
7.40-6.92 (8H, m). 4.70 (H910, s), 4.45 (H—9, t, J= 2.5 Hz). 2.27 (2H,

d, J= 2.5 Hz).

H H H

H
GEN F1 C>

g_© “ fl_©

9,10-Dihydro-9,10-(ll-Trimeth lsilox ethen 1 -Anthracene (20)

 

To a 25-mL round-bottomed flask, equipped with a rubber septum,
gas inlet valve and a stir bar were added 4.0 mL of a hexane solution
of grbutyllithium (1.56 g, 6.24 mmol) and 4 mL of dry hexane. After
cooling (0°C), dry diisopropylamine (0.90 mL, 0.65 g, 6.38 mmol) was
slowly injected by syringe into the reaction flask. The resulting
solution was stirred for 15 min at 0°C, after which the solvent was
removed under reduced pressure. To the resulting white powder was
added 7 mL of dry tetrahydrofuran. Ketone 12 (1.37 g, 6.24 mmol) in
4.5 mL of tetrahydrofuran was slowly added to the lithium
diisopropylamide solution by means of a syringe. The resulting deep
purple solution was stirred at 0°C for 10 min after which
chlorotrimethylsilane (0.77 mL, 0.71 g, 6.55 mmol) was slowly injected
into the reaction mixture by syringe. The resulting solution was
stirred at room temperature for 15 min, diluted with 50 mL of hexane

and washed twice with 30 mL of.10% cold aqueous sodium biéarbonate.

 

61

The organic layer was dried with anhydrous magnesium sulfate, filtered
and the solvent removed under reduced pressure to give crude 20; 1H
NMR (CDC13) 5 7.74-7.42 (4H. m). 7.28-7.15 (4H, m). 6.01 (H911, d of
d. I: 16 and 2 Hz). 5.08 (H-12, d, J= 16 Hz). 4.88 (H—9, d, J= 2 Hz).

0.08 (9H. s).

H

(mama-ob
H a (9

£2

 

9,10-Dihydro-9,10-(11:5eto-12-t-bgtylethanoz-Anthracene (2;)

The silyl enol ether, ‘20, was converted to 2;, using Chan's
£9buty1 alkylation procedure."4 Freshly distilled titanium
tetrachloride (2.27 g, 1.32 mL, 12.0 mmol), £9butylchloride (0.56 g,
6.0 mmol) and 15 mL of dry methylene chloride were combined, under
argon, in a 25 mL round-bottomed flask equipped with a magnetic stir
bar, gas inlet valve and dropping funnel. After cooling the solution
to -40°C and with stirring, the crude silyl enol ether, 29, (1.75 g,
5.99 mmol) in 5 mL of methylene chloride was added dropwise to it.
The mixture was stirred for 3h at -40°C and then hydrolyzed with 5 mL
of saturated sodium carbonate solution. The resultant mixture was
extracted with 50 mL of carbon tetrachloride and washed with three 25
mL portions of sodium carbonate solution. The organic layer was dried

with anhydrous magnesium sulfate, filtered and evaporatated under

reduced pressure. Two recrystallizations from ethanol gave pure 2;

m-.. _.q_..

62

(0.92 g, 3.48 mmol, 58%), mp 131—132°C, (1iu3’ mp 132°C); 1H NMR
(CDC1,) 6 7.45-7.05 (83, m), 4.80 (n+10, s), 4.69 (n+9, d), 2.16

(H-12, d).

H
H C(CH3)3
(CH3)aSi-O L o
32 ::J a :::’

Eguilibration _j the the Qiasteregmgric Algghglg.(ggyglrgagtiggl

The equilibration of the alcohols was carried out as described by
Eliel.1’ The mixture of diastereomeric alcohols (0.10 g), acetone
(0.08 g), aluminum isopropoxide (0.10 g) and iSOprOpyl alcohol (7 mL)
were combined in a 25 mL round-bottomed flask equipped with a magnetic
stir bar, reflux condensor and drying tube. The solution was refluxed
for 7d, after which it was quenched with 10 mL of 35% cold aqueous
sulfuric acid. The resulting solution was extracted with three 25 mL
portions of ether. The combined ethereal extracts were washed with
aqueous sodium bicarbonate, water, and brine and dried over anhydrous
magnesium sulfate. After filtering the solution the ether was removed

under reduced pressure. The residue was analyzed by NMR.

III. Product Ratio Analysis
The product ratios (cis/trans) were determined by proton NMR peak
integration. In the case of ketone 11, the ensuing diastereomers, 2;

and 2;, were assigned the following chemical shifts: 22 (CDCl,) 8

 

 

63

4.38 (H-lO. d. J= 2.7Hz). 4.12 (H-ll. d of d. I= 8.8 and 2.7Hz). 3.96
(H99, d, J= 1.5Hz). 2.21 (H912, m). 0.74 (’CHS, d); 2; 4.25 (H—lO, d.
J= 2.8). 3.91 (H-9, d, J=1.5Hz). 3.53 (H-ll. br t. J= 3 Hz), 1.60
(H-12, m). 0.90 (-CH3, d). The coupling constant of 8.8 Hz for H—ll

of 2; is consistent with cis stereochemistry. Likewise, the coupling

H CH3 H CH3 H CH3
<3 H (Ni PK) H

constant of 3 Hz for H-ll of 23 is in line with Egggg stereochemistry.
Reduction of 1; with lithium aluminum deuteride resulted in the loss
of H-ll peaks, while H-10 peaks became singlets. Using
homo-decoupling techniques, irradiation of 8 4.38 (H910, 22) produced
a doublet at 8 4.12 (H911, 22, I= 8.8 Hz) and irradiation of 8 4.25
(H-10, 2;) resulted in a doublet at 8 3.53 (H-ll, 23, I= 2.7Hz). When
the lanthanide shift reagent. tris(6,6,7.7,8,8,8 - heptafluoro - 2,2 -
dimethyl - 3,5 - octanedionato) europium, (Eu(fod),). (5-50 mol per
cent), was added step-wise with increasing concentration, the doublet
at 8 0.74 was chemically shifted downfield a factor of three times
that of the doublet at 8 0.90. Relative peak areas were consistent
with the chemical shift assignments. Varech and Jacques reported H-ll
chemical shifts for 22 and 23 of 8 4.12 and 3.5, respectively (no
other chemical shifts were reported by them).3’ The H-10 peak areas

were used to determine product ratios.

64

H CH2CH3 H CHpCHa H CHpCH3

Diastereomers 21 and 2§_ from 22 were assigned the following

chemical shifts: 21 (CDCl,) 8 4.38 (H910, d, J= 2.5Hz), 4.16 (H99 and

H-ll, superimposed; H-11, d of d, I= 9.0 and 2.5Hz; H—9, d, J= 1.5Hz),

1.98 (H912. m); g; 4.28 (H-10. d. J= 2.6Hz). 4.08 (H99, d, J 1.6Hz).
3.60 (H911, br m), 1.57 (H-12, br m). These assignments were found to
be consistent in the same manner as tested for 22 and 22, except no
lanthanide shift study was done. The chemical shifts for the
methylene hydrogens were superimposed between 8 1.40 and 1.10. The
two doublets for the methyl groups were superimposed between 8 1.09

and 1.01. The H—ll assignments were consistent with those reported.”

The H-lO peak areas were used to determine product ratios.

H CH(CH;)2 H CH(CH,)2 H CH-(CH,)2
o b ' H OH Ho H
i ,. Q g .. Q 5.. Q

Alcohols 2g and 21 were assigned the following chemical shifts:
(‘CDC1,) 2_6 6 4.43 (H-lO, d, 2.0 Hz) 4.28 (H-9, superimposed with 11-9

of 21), 4.23 (H-11,J= 8.0 and 2.0 Hz), 1.80 (H-12, br m); 21 4.32

 

65

(H910, d, J= 2.0Hz), 4.28 (H99, superimposed with H-9 of 2g). 3.54
(H—ll, br t, I= 2.5 Hz). 1.63 (H912, br m). Homo-decoupling
experiments and reaction with lithium aluminum deuteride were used to
verify these assignments. Relative peak areas were consistent. The
chemical shifts for the H-ll hydrogens are reported to be 4.23 and
3.54, for 2§_and 21, respectively.” The H-10 peak area of 2g and the

H—ll peak of 21 were used to determine product ratios.

H C(CH.). H C(CH3)3 H C(CH.),
o H OH Ho H

“)2, —° @_© @_©

The gig and $5232 alcohols, 22 and 21, were give the following
proton NMR assignments: (CDC13) 2§_8 4.31 (H911, br d, J= 9.0 Hz).
1.00 (-C(CH,)3, s); 22 3.91 (H911, br t, J= 3.5 Hz). 0.80 (-C(CH,),.
s). The bridgehead hydrogens (H99,10) from both isomers were
superimposed in the region 4.48 - 4.34. The H-12 hydrogen from 22 and
_2 were also superimposed, 1.70 - 1.60. Varech and Jacques reported
chemical shifts of 4.30 and 3.91 for the H-ll hydrogen of 2§_ and 22,
respectively.” The relative peak areas at 1.00 and 0.80 were used to
determine product ratios. For this reason the lanthanide shift
reagent Eu(fod)3 was used to help verify the assignments. The
step-wise addition of 5-50 mole per cent of Eu(fod)3 to the mixture
produced downfield chemical shifts for the peak at 1.00 which were a

factor of 3.5 times greater than the 0.80 peak. These assignments are

66

consistent with the relative peak areas for the H911 hydrogens.

H C6H5 H CGHS H C6H5

The following chemical shift assignments were made for 2g and 21:
(CDCl,) 19 8 4.62 (H910, d. J= 2.8Hz). 4.48 (H911, d of d. J= 10.0 and
2.8Hz). 4.43 (H-9, d, J= 1Hz), 3.44 (H-12, d of d, J= 10.0 and 1H2);
21 4.59 (H910, d, J= 2.6 Hz). 4.48 (H-9, d, J= 2.7 Hz). 4.20 (H911, br
t, J= 3.8), 3.20 (H-12, br t, J= 3.8 Hz). These assignments were
consistent with homo-decoupling experiments and reduction with lithium
aluminum deuteride. As in the previous cases, the H-11 chemical
shifts are consistent with those published.

The spectra were taken on either a Bruker WH9180 or Bruker WP-250
FT NMR spectrometer. The product ratios were calculated by cutting

and weighing peak areas.

IV. 2552; Analysis

The errors involved in the calculation of product ratios are due
to an error in the temperature and errors in the method of
intergrating peak areas. A temperature variation of t 20 contributes
approximately one per cent to the error of a product ratio. The error
in the integration of peak areas is dependent upon the signal to noise

ratio of the spectra and human error in cutting and weighing peaks

 

67

(the method used here). Since the signal to noise ratio was quite
high (solid baseline) and the results reproducible, this error was
estimated to be no greater than i 2 per cent. At least two
experiments were run at each temperature.

The error in calculating the difference in enthalpies of
activation was determined by using Equation 17.1"71

5 = 2RT'Ta/(T'-T) (u << 1) (17)

The symbols in the expression are defined as follows: 8 is the
error in AAH:F in cal/mole, R is the universal gas constant, T' and T
are the temperatures at the extreme of the temperature range and a
represents the maximum fractional error of the ratios. The maximum
fractional error for all data was assumed to be 0.03.

The maximum possible error (a) in AAS; is given by Equation

18 16,17

6 = fill/T + (T'—T)/2T'T] (u << 1) (18)

 

APPENDIX

—————_———_—

The composition of Grignards has been the subject of considerable
interest since their discovery in 1900. Through the years a number of
structures have been suggested for the complexes. Today there is no
doubt that a single composition can not represent Grignards. A number

of studies, primarily by Ashby"

, have shown that Grignards are best
represented by a complex equilibrium (Eq 17), referred to as the

extended Schlenk equilibrium.

 

etc . =R3Mg :Mng-——‘R3Mg + ngz= 2RMgX =(RMgX) ;— etc . (17)

From the organic chemisfs viewpoint, the various complexes and
equilibria should be identified so that mechanisms and transition
states can be more completely described. A better understanding of a
Grignard's composition would be useful in explaining its
stereoselectivity in addition reactions to ketones and aldehydes.

Ideally, one would want to probe the composition of Grignards
spectroscopically. A limited amount of knowledge has been obtained

using infrared”, 1H nmr and 13C nmr'1 spectroscopy. There have

been a few reports in the literature on z’Mg nmr spectroscopy'z"3,
however, no such report concerns Grignards. Reported here are some
z5Mg nmr results on Grignard systems. The chemical shifts (8) and

line widths (Ayl/z) are reported in Table 9. No 1’Mg resonances were

observed for Grignards in diethyl ether. There were also no

68 .

69

resonances observed for bis-(phenylethynyl) magnesium in
tetrahydrofuran or for dimethyl magnesium or diethyl magnesium in
either solvent.

The nmr linewidth of a quadrupolar nuclide, such as magnesium, is
generally determined by the interaction of the nuclide's quadrupolar
moment with electric field gradients at the nucleus. Since Mg+3 is
strongly hydrated in aqueous solution. electric field gradients at the
nucleus are most likely very small because of the highly symmetrical
arrangement of water molecules around the ion. Thus, for the
following discussion it is assumed that the quadrupolar relaxation
mechanism dominates the 3‘Mg relaxation thereby accounting for the
observed linewidths.

It is currently accepted that in diethyl other the Schlenk
equilibrium favors RMgX and aggregated species." These species would
be expected to have low electonic symmetry at the metal site and have
some degree of covalent bonding to the magnesium. Therefore, large
local electronic field gradients could be expected at the magnesium
nucleus. This, plus the dynamic nature of such systems broadens the
HMg lines beyond detection..

Magnesium bromide in diethyl ether would be expected to have
higher electronic symmetry at the nucleus than a Grignard. A
resonance for magnesium bromide in diethyl ether is observed (Table
9). Crystal structures of phenyl magnesium bromide“ and magnesium
bromideu (diethyl ether solvated) ‘ are consistent with this

interpretation based upon electronic symmetry.

 

70

In tetrahydrofuran, Grignard complexes are widely thought to
exist primarily as RzMg plus aggregated forms.7' One would expect that
the Rqu species (in whatever aggregated and solvated form) could be
characterized by the same types of arguments proposed for RMgX in
diethyl ether. No signals were observed for dimethyl magnesium,
diethyl magnesium or bis-(phenylethynyl) magnesium in the temperature
range 210 to 293°K. However, magnesium bromide was observed in
tetrahydrofuran (Table 9) and its linewidth indicates high electronic

symetry around the magnesium nucleus.

Table 9. “Mg nmr Data for some Grignard Reagents and Magnesium

Halides in Diethyl Ether and Tetrahydrofuran (THF).

Compound T(°K) 5 (PM!)a ATi/z

Magnesium Sulfate
1.0 g in Water 293 0 5

Magnesium Bromide
0.1 M in diethyl ether 293 47 131

Magnesium Bromide
0.1 _M in THF 293 9 20

Magnesium Bromide
0.1 M in THF ‘ 272 6 5

Magnesium Chloride
0.1 M in THF 293 9 230

Magnesium Chloride
0.1 M in THF 262 10 215

Phenyl Magnesium Bromide
0.1 M in THF 303 12 400

Phenyl Magnesium Bromide
0.1 M in THF 293 11 287

 

71

(Table 9 continued)

Phenyl Magnesium Bromide

0.1 M in THF 273 7 33
Phenyl Magnesium Bromide

0.1 M in THF 250 6 19
Phenyl Magnesium Bromide

0.1 g in m 2101’ 5 43
Ethyl Magnesium Bromide

0.1 M in THF 301 9 224
Ethyl Magnesium Bromide

0.1.! in THF 280 7 33
Ethyl Magnesium Bromide

0.1 M in THF 260 5 14
Phenylethynyl Magnesium

Bromide 0.1 M in THF 293 9 245
Phenylethynyl Magnesium

Bromide 0.1 M in THF 263 6 23
Methyl Magnesium Bromide

0.1 M in THF 293 no resonance observed
Dicyc10pentadienyl

Magnesium in THF 293 -39 43
(dilute)

a Positive values are deshielded.

b A white precipitate formed in the nmr tube.

The resonances observed for Grignards in tetrahydrofuran are most
likely due to magnesium bromide which may be more properly thought of
as solvated ions or ion pairs. This species is probably in
equilibrium with other species represented in the Schlenk equilibrium.
Crystallographic stuctures of magnesium bromide" and methyl magnesium
bromideH (tetrahydrofuran solvated) are consistent with the theory

that line widths are a function of electronic symmetry at the

 

72

magnesium nucleus.

It is of interest to note the large chemical shift difference
between magnesium bromide in diethyl ether versus tetrahydrofuran.
This no doubt is due to the differing solvating ability of the two
solvents and perhaps the proximity of the bromide ion. The crystal
structure of magnesium bromide derived from diethyl ether has two
solvent molecules per magnesium. In the case of the less sterically
demanding tetrahydrofuran there are four solvent molecules per
magnesium. _‘

Since the line width of a stg resonance is dependent upon the
electronic symmetry around the nucleus, a symmetrical organometallic
species was sought. 0f the available crystallograpic strutures
reported in the literature, the most symmetrical was
bis-(phenylethynyl) magnesium.u This compound has an octahedral
arrangement of ligands with four tetrahydrofuran molecules in a plane
and £522; phenylethynyl ligands. As mentioned, no resonance was
observed between 210 and 293‘s.

The case of dicyc10pentadienyl magnesium is interesting because
it represents the only case in which an organomagnesium species is
definitively detected. Here the line width is relatively narrow which
is consistent with a symmetrical "sandwich" compound. The fact that
the cyclopentadienyl rings are a bonded rather than a bonded to the
magnesium may also be significant. The upfield (shielding) chemical
shift is most likely due to a combination of the anistropic magnetic
properties. of the aromatic rings plus the high shielding provided by

the cyc10pentadienyl ligands.

73

Experimental

Magnesiumrzs nmr spectra were obtained at 11.0 MHz using a Bruker
WH+180 FT NMR. Twenty mm spinning sample tubes were used throughout
this study. Field/Frequency lock was maintained by locking on the 2H
resonance of deuterium oxide or d‘-acetone (for low temperatures)
which was contained in a co-axial 5 mm tube centered within the 20 mm
tube. Line widths were measured as the full width at half-height.

Diethyl ether and tetrahydrofuran were purified by refluxing and
distilling over sodium/benzOphenone and stored under argon and over
molecular sieves. Organic halides were distilled before use.
Sublimed magnesium. was obtained from the Dow Chemical Company,
Midland, Michigan.

The following is the general procedure for the preparation of
Grignards. Tb a dry 250 mL, three-necked round‘bottomed .flask
equipped with a dropping funnel, a magnetic stir bar, a condensor and
an air inlet valve was added sublimed magnesium (1.1 equiv/1.0 equiv
organic halide). The dropping funnel was equipped with a septum. The
flask was evacuated and purged with argon. Using a syringe, 5-10 mL
of solvent was added to the flask followed by 0.5 mL of the organic
halide. After a short induction period or with gentle heating the
reaction started. The remaining solvent and organic halide were
combined in the dropping funnel using a syringe. This solution was
added, with stirring, at a rate to maintain the reaction. After the
addition, the reaction mixture was refluxed for 30 minutes. The
mixture was allowed to cool to room temperature and settle. under an

argon atmosphere, the Grignard solution was siphoned through a

_... ~ aa—e-

 

74

U‘shaped tube equipped with a frit directly into the nmr tube. The
solutions were clear and colorless except for phenyl magnesium bromide
(which was only slightly discolored).

Magnesium halides were prepared from the corresponding mercuric
salts using an excess of magnesium as described by Ashby and Arnott.u
Dimethyl and diethyl magnesium were prepared from the corresponding
dialkyl mercury as described by Ashby and Arnott.u The dialkyl
mercury was prepared by the method of Gilman and Brown.’0
Phenylethynyl magnesium bromide was prepared by adding a slight excess

‘ Bis-(phenylethynyl)

of phenylacetylene to ethyl magnesium bromide.‘
magnesium was prepared by adding a slight excess of phenylacetylene to
dimethyl magnesium.u Dicyclopentadienyl magnesium was kindly provided

by Dr. Carl H. Brubaker.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

REFERENCES

E. L. Eliel, Rec. Chem. Progr., 3;, 129(1961).
K. Mislow and M. Raban, Top, Stereochem. ;, 1(1967).

E. Fischer, Ber., 21, 3231.

H. B. Burgi, J. M. Lehn and G. Wipf. IM,A!; Chem. Soc,, 26,
1956(1974).

G. 8. Hammond, I; Am, Chem. Soc,, 11, 334(1955).

For activation parameters see References 41,42 and 47. For
heats of reaction see R. E. Davis and J. Carter, Tetrahedron,
22, 495(1966).

D. I. Cram and F. A. Abd Elhafez, 1, Am, Chem, Soc., 11,
5828(1952); D. J. Cram and F. D. Greene, ibid., 1g, 6005(1953);
D. J. Cram and D. R. Wilson, ibid., §§, 1254(1963).

I. W. Cornforth, R. H. Cornforth and K. K. Mathews, 1, Chem,
Soc,, 112(1959).

G. J. Karabatsos, 14.52; Chem. Soc., §2, 1367(1967).

M. Cherest, H. Felkin and N. Prudent, Tetrahedron Lett.,
2201(1968).

 

J. Richer, I; Org. Chem,, 39, 324(1965).

1. R. Boone and E. C. Ashby, Top, Stereochem. 1;, 53(1979).

I. D. Morrison and H. S. Mosher, Asymmetric Organic Reactions,
Prentice-Hall, Englewoods Cliffs, N. J., 1971; paperback reprint,
American Chemical Society, Washington D. C. 1976.

D. H. R. Barton, M, Chem, Soc,, 1027(1953).

 

W. G. Dauben, G. I. Fonken and D. S. Noyce, I; Am. Chem. Soc.,
lfi. 2579(1956).

 

H. C. Brown and H. R. Deck. J'_. _A!_,_ M m. fl. 5620(1965).
E. L. Eliel and Y. Senda, Tetrahedron, 36, 2411(1970).

I. Klein and E. Dunkelblum, Tetrahedron M££;,, 6127(1968).

D. H. Wertz and N. L. Allinger, Tetrahedron, 39, 1597(1974).

E. C. Ashby and I. T. Laemmle, Chemical Reviews, 15, 521(1975).

75

 

 

21.
22.
23.
24.
25.
26.
27.

28.

29.
30.
31.
32.
33.
34.
35.

36.

37.

38.

39.

40.

41.
42.

43.

76

n. c. Wigfield and F. v. Gowland. L M... 21102.... 5;. 1108(1977).
n. c. Wigfield. an L 91.12;... a. 646(1977).

Reference 12, p. 77.

Reference 13, p. 119 and Reference 12, p. 91.

Reference 13, p. 120.

Reference 12, p. 86.

Reference 12, pp. 79-81 and Reference 13, pp. 129-130.

N. T. Ahn, O. Eisenstein, I. M. Lefour and M. R. Tran Huu Dan, 1;
Am. Chem, Soc., 2;, 6146(1973).

I. KLein, Tetrahedron Lg£;,, 4307(1973).

J. Klein and D. Lichtenberg, MM Q£Ba.§!2!;o.§2; 2654(1970).
E. C. Ashby and J. R. Boone. l; Q;g,.§hgg,,lfil, 2890(1976).
Reference 20, p. 544.

N. T. Ahn and O. Eisenstein, Mggz,ԣ&ԤMig&, ;, 61(1977).
Reference 12, p. 75.

D. Varech and J. Jacques, Tetrahedron ngg,, 1;, 4443(1973).

M. J. Brienne, D. Varech and J. Jacques, Tetrahedron Lett., 16,
1233(1974).

M. C. Cherest, H. Felkin, P. Tacheau, I. Jacques and D. Varech,
Chem. Comm., 372(1977).

Y. Izumi and A. Tai, Stereo-Differentiating Reactions, Konansha
Scientific Books, Tokyo, Japan, 1977, pp. 178-184.

D. Varech, M. J. Brienne and I. Jacques, 1; Chem. Res. 81302.,
12. 319(1979).

E. C. Ashby, F. R. Dobbs and H. P. Hopkins, Jr., 1; Am, Chem.
Soc,, 21, 3158(1975). A

E. C. Ashby and J. R. Boone, I. Am. Chem. Soc., 2;, 5524(1976).

K. E. Viegers and S. G. Smith, 1, Mg, Chem, Soc., 22, 1480(1977).

H. C. Brown and R. F. McFarlin, l;.é!a Chem. Soc., £9,
5372(1958).

44.

45.

46.

47.
48.
49.
50.

51.

52.

53.

54.

55.

56.
57.
58.
59.
60.
61.

62.

63.
64.

65.

77

H. C. Brown and C. F. Shoaf, J. Am, Chem. Soc., gg, 1079(1964).

 

H. C. Brown, E. J. Mead and B. C. Subba Rao, 1,,MM, Chem. Soc.,
11. 6209(1955).

H. C. Brown, 0. H. Wheeler and X. Ichikawa, Tetrahedron, ;,
214(1957).

1). c. Wigfield and F. w. Gowland. J_. ml @291. 5;. 2396(1976).
n. c. Wigfield and F. w. Gowland. .J_.. m. 9.112... 1.2.. 1108(1977).
H. C. Brown and K. Ichikawa, IM.A!;‘£MQQM.§22&..§§. 4372(1961).
I. L. Pierre and H. Handel, Tetrahedron LQEEL. 2317(1974).

H. C. Brown. E. J. Mead and C. I. Shoaf, I; Am. Chem. Soc., 18.
3616(1956).

H. Haubenstock and E. L. Eliel, 11.52; Chem, Soc,, £1,
2368(1962).

B. Rickborn and M. T. Wuesthoff,‘1, Am. Chem. Soc,, 22,
6894(1970).

G. J. Karabatsos and D. I. Fenoglio, Topics Mg Stereochemistgy,
é. 167(1970).

For Equations 9, 10, and 11 see Reference 13, p. 123. For
Equation 12 see Reference 12, p. 89. For Equations 13 and 14 see
Reference 20, p. 532.

G. G. Smith and R. P. Bayer, Tetrahedron, Mg, 323(1962).

J. A. Perry and K. D. Warren, ;, Chem. Soc,, 4049(1965).

K. Bowden and M. Hardy, Tetrahedron, 22, 1169(1966).

P

. Geneste and G. Lamaty, Bull. Soc. Chim. 25;, 669(1968).

 

J. E. Leffler, 1‘ Org, Chem., 29, 1202(1955).
R. C. Peterson, 1, Org. Chem,, 22, 3133(1964).

n. c. Wigfield, n. J. Phelps, 3. F. Pottie and a. Sander,
q, Ln, Chem. Soc., 21, 898(1975).

D. C. Wigfield and D. J. Phelps, Chem. Commun,, 1152(1970).
D. C. Wigfield and D. I. Phelps, Can. M, Chem., gg, 388(1972).

H. C. Brown and J. Muzzio, erAm. Chem. Soc., MM, 2811(1966).

66.
67.
68.

69.

70.
71.

72.

73.

74.

75.

76.

77.

78.

79.

80.

81.

82.
83.
84.
85.

86.

78

W. E. Noland, Org. Syp., Coll. Vol. 5, 833(1973).
C. D. Hurd and M. E. Nilson, l; Org. Chem.. 29, 927(1955).
G. D. Buckley and C. W. Scaife, M, Chem, Soc.. 1471(1947).

Y. V. Baskov, T. Urbanski, M. Witanowski and L. Stefaniak,
Tetrahedron, 29, 1519(1964).

The methyl doublet for the cis isomer is at 8 2.02.
D. W. Worrall, Org. 823,, Coll. vol. 1, 413(1941).

W. E. Noland, H. I. Freeman and M. S. Baker, 11.52; Chem. Soc..
1M, 188(1956).

E. Keinan and Y. Mazur, 1;,AQ& Chem. Soc,, 22, 3861(1977).

T. H. Chan, I. Paterson and J. Pinsonnault, Tetrahedron Lett.,
fifi. 4183(1977).

E. L. Eliel and S. H. Schroeter, l; Am. Chem. Soc.. 81.
5031(1965).

K. B. Wiberg, Physical Organic Chemistry, John Wiley and Sons,
Inc., New York, 1964. pp. 376-379.

R. C. Peterson, J. Markgraf and S. D. Ross, 1; Am. Chem. Soc..
£1. 3819(1961).

E. c. Ashby, Quart, Rev,, 2;, 259(1967). E. c. Ashby, Bull. Soc.
Chim, Fr,. 2133(1972).

 

R. M. Salinger and H. S. Mosher, leAm. Chem. Soc., 85,
118(1963).

G. E. Parris and E. C. Ashby, J; Am. Chem. Soc,, 23, 1206(1971)
and references therein.

D. Leibfritz, B. O. Wagner and J. D. Roberts, Leibigs Ann, Chem.
763. 173(1972).

L. Simeral and c. a. Maciel. L. 2222.. Elan“ £9. 552(1976).

P. H. Heubel and A. I. Popov, 1; Solution 9222,, g, 283(1979).

6. n. Stucky and a. E. Bundle. .1. _Am. £11.22... in... 132. 1002(1963).
H. Schibilla and M. T. LeBihan. Acta Egyg£&,{;§, 232(1967).

M. C. Perucaud and M. T. LeBihan, Acta Cryst,, Sect. B, 31,
1502(1968).

 

87.

89.

90.

79

M. Vallino, L, Organometal. Chem., LO, 1(1969).

M. C. Perucaud, M. J. Ducom and M. Vallino. _C_. M, Acad. Sci..
Paris, ser. C, 264, 571(1967).

 

 

E. C. Ashby and R. C. Arnott, l; Organometal. Chem,, 141,
1(1968) .

H. Gilman and R. E. Brown, L a, Chem. Soc,, 52, 3514(1930).

IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII

, llHlHfllllUlMflm[[tfllLUlllflllffllejfltW”