THE STEREOCHEMISTRY OF THE REDUCTION
OF AN ALICYCLIC UNHINDERED KETONE
WITH SOME BOROHYDRIDES
HUGH EDWARD WISE, JR.
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
To merely acknowledge the technical assistance provided by
Dr. W. .. Jones in the course of this investigation would not only
understate the case, but also hardly express the author's sincere
appreciation for a relationship which has made Graduate study possible
in an atmosphere of friendly informality. For the patience with which
Dr. Jones has extended many hours of personal time, the author will
almwys be gratefully indebted.
The author vould also like to acknowledge the invaluable support
of his parents in helping him to rise to his present educational
level. without their constant encouragement, faith and financial
assistance made available in crucial periods, it would have undoubtedly
been a much rougher road.
TABLE OF CONTENTS
LIST OF FIGURES iv
I. SCOPE OF THE STUDY I
11. INTRODUCTION 3
iX.1 SELECTION O THE REDUCTION SYSTEM AMD ETHOD OF
ANALYZING THE PRODUCTS 13
IV. RESULTS AND DISCUSSION 18
V. EXPERIMENTAL 38
VI. BIBLIOGRAPHY 61
BIOGRAPHICAL SKETCH 63
LIST r' FIGURES
1 Apparatus for Diborane Reduction3 45
2 Apparatus for the Production and Handling of
Aluminum Borohydride 51
3 Typical Gas Chromatograph of a Reduction Mixture 59
I. SCOPE OF THE STUDY
A study of the literature indicates the need for additional
stereochemicnl data to aid in further elucidation of the mechanism of
borohydride reductions, particularly in regard to the individual aspects
of each hydride transfer, This need becomes obvious when one attempts
to e::plain some of the stereoisomeric ratios obtained by previous inves-
tigators with reference to their generalizations concerning the nature
of the factors controlling these ratios. Other considerations, such as
the role played by the cation and the solvent, loom as distinct possi-
bilities in accounting for inconsistencies encountered in the appli-
cation of these generalizations, Insofar as the stereochemistry of
diborane reductions is concerned, there appears to be a decided lack of
information published to date.
With this situation in mind, it was decided to establish the
following objectives for this study.
(1) To evaluate the effect of the cation on the average stereochemis-
try of all four hydride transfers from the borohydride ion, as well as
the initial transfer separate from the remaining three.
(2) To obtain stereochemical data on sodium borohydride reductions in
diglyme (previously unavailable) and to use this information to support
or modify the currently suggested mechanism, particularly with respect
to the individual hydride transfers involved.
(3) To obtain stereochemical data on diborane reductions for purposes
of not only testing its consistency with the proposed mechanism in the
light of current stereochemical theories, but also to help clarify the
nature of the last three hydride transfers of the borohydrides.
The name "borohydride" has been adopted for compounds containing
the BH4I ion or its substituted derivatives,' even though they may
actually contain no hydrogen available as a hydride ion. The hydride
of boron, B2 6, has been assigned the name "diborane" and the B33
fragment "borine." Because the borne fragment, not diborane, appears
to be the species involved in the acid-base type complexes formed when
diborane acts as a Lewis acid, diborane is sometimes considered as a
dimerized version of borine.
History of the borohydrides and diborane
The real beginning of borohydride chemistry occurred in 1931, when
Schlesinger and Burg developed an improved method for preparing di-
borane.2 It wasn't until 1940 that this material was used in the synthe-
sis of the borohydrides of lithium, aluminum and beryllium by reaction
of the metal alkyl with diborane.3 Because the borohydrides of aluminum
and beryllium are volatile and because of their experience in working
with these materials, Schlesinger and Brown early in 1941 were requested
by the government to undertake an investigation of new volatile com-n
pounds of uranium.1 In the course of these investigations, improved
methods were developed for preparing starting materials and intermedi-
ates required for the preparation of the uranium borohydrides. As a
result of these efforts '"many new types of reactions were observed and
at least partially investigated, hitherto unknown compounds were dis-
covered, and the field of the chemistry of the hydrogen compounds of
boron was greatly enlarged." This study gave birth to the borohydrides
of sodium and potassium.4
Although the borohydrides were actually discovered before the
aluminohydrides,5 publication of their preparation and use as reducing
agents was stalled because the work was done under government contract
and had to be declassified following World War Il. Consequently,
lithium aluminum hydride became popular as a selective reducing agent
sooner than the borohydrides.6
Meanwhile, the behavior of diborane toward organic compounds conu
training a carbonyl group was noted in 1939.7 However, the development
of the metal borohydrides and aluminohydrides held off further investi-
gation of diborane as a reducing agent for organic compounds until the
In 1949, Chaikin and W. G. Brown reported the reduction of alde-
hydes, ketones and acid chlorides with sodium borohydride.9 The
specificity of the borohydrides for reduction of carbonyl groups in
compounds containing other reducible groups, their solubility in water,
or alcohol-rater mixtures, and the fact that reductions could be carried
out in simple stirred vessels at atmospheric pressure quickly es-
tablished uses in the pharmaceutical industry. The first of these uses
was in the synthesis of cortical hormones.
During the next few years, the price of the borohydrides dropped
as process improvements were made and markets developed.10 Research
was devoted to seeking methods of increasing the mild reducing power of
sodium borohydride. It was found that the addition of certain metal
halides end alloxy groups would effect this increase.11,12,13 The
elcectivity of the alkaline-earth and aluminum borohydrides was also
noted.14 As a result of these efforts, the reducing power of the
borohydrides was increased to include the reduction of esters, nitriles,
amides and carboxylic acids. These changes, together with diborane,
gave a series of reducing agents which vary from mild to acLivc, from
basic to acidic, to selectively reduce certain functional groups and
leave others unaffected.
The stereochemistry of reductions by complex meta. hydrides
Initial consideration of the factors governing the stereochemistry
of reductions by complex metal hydrides began as early as 1937415.
While studying the relative stabilities of-the epimeric cis-5-hydrindan-
ols, Huckel found that a mixture of alcohols of the came composition a$
that obtained by direct equilibration could be obtained by treating the
alcohol with sodium in refluxing xylene. One isomer was found to pre-
dominate over the other in such equilibria, so that this technique
could be used to study the relative stabilities of the two diastereomeric
alcohols. Several years later, Hughes and Ingoldl6 concluded that the
reduction of a carbonyl group is a bimolecular addition reaction and
should, therefore, be one in which steric factors are operative. The
numerous investigators which followed all used these two factors, often
competitive, to explain the stereochemical consequences of reduction by
complex metal hydrides of a variety of alicyclic and acyclic hetones, as
well as numerous ketosteroids.
In 1949, Trevoy and Brown1 showed that lithium aluminum hydride
(LAR) opens epoxide rings by a bimolecular inversion mechanism,
indicating the possible importance of steric effects They saw a
common storic element in the attack of a complex hydride and a catalyst-
hydrogen complex. In the same year, Jackmann, .t S,, 1 studying the
results of the reduction of a series of substituted cyclohexanones with
aluminum isopropoxide, noted that increasing the steric hindrance of
the carbonyl group by bulky groups on adjacent carbons tended to increase
the yield of axial opimer. They also reduced menthone with the aluminum
derivatives of four secondary alcohols and found that increasing steric
hindrance in the alkoxide group increased the proportion of axial
ihomer. From these results, they concluded that the chief factor de-
termininng the relative amounts of ciL- and tranc-epimere was steric
hindrance about the carbonyl group, but that the relative stabilities of
the two epimers might also be important in some cases.
Noyce and Denney19 in 1950, studied the reduction of a series of
substituted cyclohexanones with LAH and noted the predominance of the
equatorial epimesr Since LAR had already been shown to open epoxides by
an SN2 mechanism, it seems likely that similar addition of hydrogen to a
carbonyl group should also occur opposite to a hindering group to give
a cia_ hydroxyl group. They pointed out that random reduction might be
expected in an unhindered carbonyl group if it were not for the fact
that the activated intermediate complex partakes of the character of the
product, so that marked steric hindrance would be required to prevent
the predominance of the more stable isomer. Applications of these
ideas to ketooteroids was made in the same year by Shoppee and Summoers.20
They made a poor case in attempting to account for a preponderance of
bg~ea-epimer in the reduction of a hetosteroid with LAU by attributing
it to steric hindrance by an angular methyl group in the 4-position,
relative to the carbonyl group, which favored rearward attack by A1H4".
Several studies since then have shown that 4-substituted alkyl groups
make no steric contribution to the reduction of cyclohexanones, so that
this explanation leaves something to be desired.
A year later Nace and O'Conner21 offered additional refinements
to these ideas as a result of their study of the reduction of cholesta-
none with LAH and aluminum alkoxides. They attributed more hindrance on
the 'lpha side to the 5-hydrogen (also on the alpha side) and noted that
increasing the sise of the alkoxy group resulted in an increase in the
yield of alpha-cholestanol. Even though hindrance favored the alpha*
epineride, the beta-epimeride was usually produced in greater quantity
because it is more stable. They concluded from this that the alpha"
producing transition states may rearrange to form the thermodynamically
favored beta-producing transition states,
Meanwhile, the acyclic systems were not being neglected. Cram and
Elhafez22 in 1952, studied reductions of substituted butanones with com-
plex metal hydrides and formulated a rule of "steric control of
asymmetric induction" to predict stereochemical direction in reactions
of acyclic systems in which a new asymmetric center is created adjacent
to an old, "That diastereoisomer will predominate which would be formed
by approach of the entering group from the least hindered side of the
carbonyl group when the rotational conformation of the C-C bond is such
that the carbonyl group is flanked by the two least bulky groups
attached to the adjacent asymmetric center."
In 1953, Barton23 pointed out the greater stability of equatorial
over axial substituents in cyclohexane derivatives and repeated the
generalization of earlier investigators that, in unhindered hetones, the
thermodynamically more stable isomer will predominate in LAH reductions.
He also defined a keto group as being subject to steric hindrance when
it does not react readily with reagents like semicarbazide and the
Several publications on the stereochemistry of hydride reductions
appeared in 1956.24,25 Among these was one by Dauben, Fonken and
Noyce,2 which showed that, in the reduction of alkylcyclohexanones,
changing the reducing agent from LAH to Al(O-i-Pr)3 leads to an in-
creasing proportion of axial (unstable) isomer, They then introduced
new nomenclature by discussing the mechanism of the reduction step with
reference to the ease of formation of the initial metallo-organic comr
plex (steric approach control) and to the relative energetic of the
formation of products aoce the complex is formed (product development
In 1958, Hardy and Wicker27 became concerned about the results
reported by these early investigators and undertook a study involving
the reduction of substituted cyclohexanones with four different re*
agents (KBHI4, aBB4, sodium in ethyl alcohol and aluminum isopropoxide).
They found no support from these results for some of the generalizations
made by previous workers as to the probable mechanism of these re-
An excellent summary of the steric and stability factors and how
they influence the stereoisomeric content of reduction products was
published in 1959 in a study by Beckett, j ~ I.,28 on the reduction of
tropinone using a variety of complex metal hydrides in different solvents.
A recent study by Eliel and Rerick9 provides an example of almost
complete thermodynamic control. In the reduction of 4-t-butylcyclo-
he:anone with LAH-A1C13 (1:4), if additional ketone is added at the end
of the reaction, equilibration of the bulky aluminum-oxygen complexes
results in over 99 per cent of the equatorial isomer.
In surmarizing the stereochemical generalizations proposed by
previous workers it may be said that, in the reduction by complex metal
hydrides of ketones capable of forming diastereoisomeric alcohols, two
factors appear to be operating to determine the ratio of epimers ob-
tained. One involves the competitive attack of the reducing species
upon a hindered versus a less-hindered side of the carbonyl group in
which steric factors are operative. The other involves the relative
thermodynamic stabilities of the two possible orientations of an initial
complex formed by the metal hydride and the carbonyl group. This latter
factor appears to be the more important of the two, especially in
alicyclic systems, and apparently begins to influence the nature of the
product as early as the transition state when orbital hybridization is
changing from sp2 to sp3*
Mechanism of borohydride and dlborane reductions
Brown, Schlesinger and Burg7 first proposed the mechanism for the
reduction of organic compounds containing a carbonyl group with diborane,
This mechanism involves "(1) the intermediate formation of a complex
compound of borine, BH,, with the carbonyl oxygen, followed by (2) the
migration of a proton with a pair of electrons from the boron atom to
the carbon atom of the carbonyl group."
In 1949, two years after he introduced it as a convenient reducing
agent for carbonyl-containing organic compounds,6 W. G. Brown17 proposed
a mechanism for reductions involving ULA with the following comments.
"With few exceptions the known reactions of IAH with organic compounds
consists essentially of the displacement of a strongly electronegative
element (oxygen, nitrogen or halogen) on carbon b hydrogen. Ion-polar
groups, eg., isolated double bonds, are not affected. It is therefore
reasonable to suppose that the reaction proceeds by a polar mechanism.
Moreover, the hydride reactions thus far reported have certain features
in common with the more familiar nucleophilic displacement reactions on
carbon and a common basis in mechanism is indicated. It appears likely
that the reactant is actually a series of complex aluminohydride ions,
AlXnH4-n~ there n progresses from zero to four during the course of a
reaction, which act as carriers for hydride ions." Thus, the initial
hydride transfer is by an incipient A1H4 ion and subsequent stages of
the reaction involve hydride migration in a ketone-A1H3 complex. He
goes on to say that, "in comparison with LAH, the borohydrides of
lithium and sodium differ principally in their lesser aggressiveness. It
is highly probable that the mechanism of their action is essentially
similar, and that their lower reactivity 4 ts a consequence of the greater
stability of the borohydride ion and of the reluctance with which it
transmits a hydride fragment to an electron-deficient center."
Further support for this picture was provided in a kinetic study
by H. C. Brown in 1957.30 It had previously been demonstrated by Garrett
and Lyttle39 that the reduction of 3-alpha-hydroxy-llalpha-acetoxy-
pregnan-20-one by sodium borohydride was second order. to simplify his
study, Brown used isopropyl alcohol as a solvent because it exhibited no
measurable reaction with sodium borohydride over extended periods of time,
and acetone as the reducible carbonyl compound. Second order kinetics
were observed, first order in acetone and first order in sodium boro-
hydride. To account for these kinetics, four successive stages were
proposed. A four-step reaction with second order kinetics was then
R2CO + BH4" k-- H3BOCHR2
PICO + Hg3EOC" -- k2'O- B(OCHR2)2
R2CO + H2B(OCHR2)2" k2.. IB(OCHR2)3
R2CO + BB(00CHR)3" -kz B(OCRIR)4
explained to be "consistent with a slow rate-determining reaction for the
first stage, with successive stages being considerably faster.31 This
explanation was shown to be reasonable by the preparation of the inter-
mediate NaBH(OCHR2)3, and the demonstration that its reaction with
acetone is much faster than the reaction of the ketone with sodium boro-
hydride itself.13 The observed rate constants may either represent the
rate of the direct reaction of borohydride ion with the carbonyl group,
C=0 + BH4 C2-
or it may represent the product of an equilibrium constant K for the
association of borohydride ion and the aldehyde or ketone, and the rate
constant k for the subsequent transfer of a hydride ion to the carbonyl
C=-0 + BH4" ... ..*B~I -- -> -P-H
A choice between these possibilities does not appear possible at this
in summary, previous investigators have suggested a different
mechanism for reductions involving diborane than for the borohydrides.
These would appear to require different modes of hydride transfer in
each case. Diborane is a Lewis acid which functions best as a reducing
agent in attacking groups at positions of high electron density. In its
reaction with a carbonyl group, the initial stage is believed to be an
acid-base interaction of BHI3 with the oxygen atom. This results in an
electron deficiency at the carbonyl carbon and is followed by the
transfer of a hydride intramolecularly. The borohydride ion, on the
other hand, functions as a Lewis base which prefers to attack a group
at positions of low electron density. For the first hydride at least,
this would necessarily involve a direct intermolecular hydride transfer
as the borohydride ion (perhaps solvated) approaches the carbonyl carbon
atom, rendered electron deficient by the polarization of he carbonyl
group. The subsequent hydride transfers could then occur intra- or
intermolecularly, depending on whether the remaining hydrides are pro-
vided by the BH3 fragment or, as Brown suggests, by all:oxyborohydride
lZI. SELECTION OF THE REDUCTION SYSTEM AND
METHOD OF ANALYZING THE PRODUCTS
Selection of the carbonvy compound
It was desirable to have a model compound which would remain
conformationally homogeneous during the course of the reaction in order
that the dinstereoisomeric composition of the products reflect the true
stereochemistry of the reduction step. Uinstein and Holnesc32 have
shown 4-t-butylcylohexyl derivatives to exist preferentially in the
chair conformation and that "the energy quantity by which a _-butyl
group favors the equatorial position is sufficiently large to guarantee
conformational homogeneity," In addition, "the t-butyl group is so
distant from the reaction zone that polar effects due to the _t-butyl
group tend to be small or negligible." For these reasons, 4-t-butyl-
cyclohexanone seemed ideal as the ketone with which to study the stereo-
chemistry of reductions with the borohydrides and diborane.
Selection of the solvent
It was desirable to have a solvent in which the borohydrides would
be sufficiently soluble and with which they would undergo no significant
reaction for the duration of the reduction period. Brown, head and
Rao33 have shown icopropanol and the dimethyl ethers of di* and triethylene
glycol (diglyme and triglyme, respectively) to be solvents which would
satisfy these requirements. To eliminate the possibility of solvent
effects causing differences in the stereochemical consequences of the
various reductions to be undertaken with different reducing agents, it
was decided to use only one solvent throughout the study, Accordingly,
diglyme was selected as the solvent.
Temperature of the reductions
Nace and O'Conner21 found that the isomeric composition upon
reduction of cholestanone by LAN and aluminum alkoxides was greatly
affected by temperature, Trevoy and Brown17 suggested that lower
temperatures increased the "directive effect" on LAH reductions of
l:2-diketones. However, Beckett, aZl .,28 found that changes in the
temperature of reduction had negligible effect on the stereochemical
course of the reduction using sodium borohydride. It was decided,
nevertheless, to minimize any effects caused by temperature differences
by running all reductions at about the same temperature. Most of them
were carried out at 20 degrees in a constant-temperature water bath,
with some undergoing reaction at room temperatures ranging from 24-28
Method of analyzing the products
Wicker34 gives a rather substantial review of some of the methods
that have been used by various investigators for analysis of the
isomeric composition of the products of reduction. Some of the early
work involved the use of refractive indices and densities. Later
workers based their analyses on viscosities and optical rotation.
Considerable work was reported which involved the conversion of the
alcohols into solid esters, fractionally recrystallizing them and running
a thermal analysis (heat of combustion) on the ester fractions. This
latter method proved to be unreliable because of poor material balances.
Still other work was based on incorrect refractive indices given by
earlier investigators and so cannot be considered valid.
Wicker defines "the most reliable method of analysis as one
based on physical properties of the total product in which the values
of the two stereoisomers concerned show a reasonably measurable
difference." In analyzing the methyl cyclohexanols, he favored density
over refractive index because of larger differences between isomers for
the former property. Another method he considers fairly reliable is
based on melting-point composition diagrams, but even this is of
questionable value because it depends on all impurities, e.g., unre-
acted ketone, being absent.
Two methods used with success by more recent investigators28'35
are infrared and gas chromatographic analysis. Results by these methods
compare very favorably with each other and show reasonably good pre-
cision. Also they avoid the serious disadvantage of the other methods,
insofar as the presence of impurities or unreacted ketone does not
drastically interfere with the analysis, which can be carried out on the
total product. For this reason, and the fact that specific information
was available on column packing which had proven effective in the sepa.
ration of 4-t-butylcyclohexanols, it was decided to use gas chroma-
tography as the method of analysis.
Review of the stereochemical considerations involved in the reduction of
Before turning to a discussion of the results of this study, it
was felt that a review of the considerations involved in the reduction
of cyclohexanone-type compounds might be helpful to the reader unfamiliar
with the steric situation prevailing in such systems. The following
description refers to the figure below.
Axial Plane of sp2
Position /Hybrid Orbitals
S Carbonyl Oxygen
L .-- ---Equatorial
The view is through the plane of the cyclohexane ringS chair
conformation, with the carbon skeleton shaded. For purposes of illus-
tration, hydrogen atoms other than those in the 3,5- and 2,6-axial
positions have been omitted.
According to the mechanism of carbonyl reactions, any species
attacking a carbonyl group must do so from a direction perpendicular to
the plane of the sp2 hybrid orbital, one of which constitutes the siga
portion of the carbon-oxygen bonding, This being the case, the illus-
tration clearly shows that the 3,5-hydrogens are in a position to inter-
act more effectively with any group approaching from direction A than
the 2,6-hydrogens could with any group approaching from directionE B The
basis of this difference lies in the fact that the plane of the carbonyl
group forms an acute angle with the plane of the 3,5-carbon-hydrogen
bonds, while the analogous angle for the 2,6-carbon-hydrogen bonds is
obtuse. Side A is, therefore, said to be more sterically hindered than
side E, a situation which favors approach of an attacking species from
side E with the eventual result of an axially positioned hydroxyl group.
Similar reasoning prevails in accounting for the greater thermo-
dynamic stability of the equatorial position. The greater interaction
of the 3,5-hydrogens with the oxygen atom of the incipient alkoxy group
favors the equatorial position in any equilibrium existing between an
axial-producing and an equatorial-producing transition state. This inter-
action with the 3,5-hydrogens would, of course, be even greater if the
oxygen atom were part of a complex which would increase its bulk.
By way of additional nomenclature, the names cEi and trans are
often used for the axial and equatorial isomers, respectively. These
terms, borrowed from geometric isomerism, have reference to the re-
lationship which the hydroxyl and t-butyl groups would enjoy with respect
to each other if the cyclohexane ring were planar.
IV. RESULTS AND DISCUSSION
Reductions with the borohydrides
A series of reductions was carried out involving the borohydrides
of three different cations with 4-t-butylcyclohexanone in a ratio of
1:4. The objective of this series was two-fold. One was to evaluate
the effect of the cation on the stereochemistry of the reduction. The
other was to provide data on the overall stereochemistry of all four
hydride transfers by exhausting the hydride content of the borohydride
ion* These results had not been previously available for such reductions
carried out in diglyme. The results are summarized in Table 1.
Table 1 ,
Borohydride Average peak height ratio % r.cis-isomer
Sodium 0.295 21.0
Magnesium 0.275 20.4'
Lithium 0.210 15.3
*This value represents an average, as explained in the experimental
The reduction of the ketone with sodium borohydride in a 1:1 ratio
was also investigated. The result was the same as that for the 1:4
reduction above. This supports Bromn's mechanism,30 which maintains
that the last three hydride transfers are rapid and that the first is
the rate-determining step. If one of the hydride transfers subsequent
to the first was slow,- then in the presence of excess borohydride, all
of the hydride content of the borohydride ion presumably would not be
utilized. In this case one should observe stereochemical results
different from the 1:4 reductions. It may be objected that all four of
the reduction steps may have the same stereochemistry. This is hardly
likely, as will be shown to be the case later in the discussion, because
a reducing species of different stereochemical potential is involved in
each of the four hydride transfers.
If the cation is involved in the mechanism of borohydride re-
ductions, one might expect two possible consequences of this involvement.
These concern the reactivity and the stereochemistry respectively.
Previous investigators appear to have devoted their attention
exclusively to reactivity differences between borohydrides of different
cations. Brown,33 for example, ran a reaction involving ethyl benzoate
with lithium and sodium borohydride in diglyme. It was found that the
sodium borohydride reaction mixture lost less than 10 per cent of the
available hydrogen in 24 hours at 75 degrees, while lithium borohydride
reduced the ester easily under the same conditions. Both salts were
completely in solution. He concluded from these results that such
differences in reactivity do not have their origin in colubility differ-
ences of the two salts.
It has been shown that the addition of lithium and magnesium
salts to diglyme solutions of sodium borohydride results in a more
aggressive reduction than can be achieved with sodium borohydride alone..
One possible explanation for this difference in reactivity might be that
the lithium and magnesium ions, because of their larger charge per size
compared to that of sodium, interact with the carbonyl oxygen to increase
the polarization of the carbonyl group. Thi' would render the carbonyl
carbon more susceptible to nucleophilic attack and account for the
increased reactivity resulting from the presence of the lithium and
magnesium ions in the reduction mixture.
No reference appears to have been made to possible cation involve-
ment in the stereochemical aspects of the reduction step. It can be
seen from the results shown in Table I that the presence of sodium and
magnesium ions lead to essentially the same stereochemistry, whereas
the lithium ion resulted in less of the cjs-isomer. Now, if a cation
actually does interact with the carbonyl oxygen, as suggested earlier,
the complex formed by the metal ion and the oxygen would certainly be
more bulky than the oxygen alone. Thus, in the light of concepts
discussed in the introductory section, the transition state should be
subject, at least for the first hydride transfer, to more thermodynamic
control because the bulkier complex would favor the equatorial position
as shown below. This should result in less of the cis-isomer, as is the
case for the lithium borohydride reduction. The question then arises as
to why the magnesium ion had no such consequences on the stereochemistry
of the reduction, since it has been shown to influence the reactivity.
The most reasonable answer to this would seem to be that although the
presence of the magnesium ion is enough to amplify the polarization of
the carbonyl group, its interaction with the carbonyl oxygen is
apparently not sufficient to appreciably alter the effective size of
the oxygen atom. Thus, its stereochemical effect is negligible.
Reduction with the borohydrides in the presence of amine
One of the problems that has plagued investigators studying the
stereochemical aspects of borohydride reductions is that the product
obtained represents an average of the stereochemistry of all four re-
duction steps, It was recognized from the outset of this problem that,
if some means could be devised to study the stereochemistry of one of
these steps separate from the others, the nature of the reducing species
involved in each of the various hydride transfers might enjoy consider-
With this quest in mind, perusal of a publication by Brown33
revealed a possibility that, with suitable application, might afford a
method of stopping the reduction after the first step. Brown provided
evidence which suggested that the presence of triethylamine caused the
reduction to temporarily cease after the first hydride transfer. He
found that the reaction of sodium borohydride with acetone in diglyme
at 0 degrees was far from complete even after 96 hours. However, in the
presence of triethylamine under the same conditions, 25 per cent of the
available hydride was lost within five minutes with no further loss for
periods as long as rix hours. He concluded that the catalytic activity
of the triethylamine was due to the participation of the amine in the
first hydride transfer by a Lewis acid-base reaction with the boro-
H H CH3
1 \ I 6-
Et3N + BH4 + (0C3)2C00 -. Et3N...B..H...C-0 EtSVBH3 + (CHg3)2CHo
Brown goes further and attributes the halting of reduction after
the first hydride transfer to the conflicting steric requirements of
the alkoxyborohydride and the triethylamine. This explanation appears
to be inconsistent with his data and with the mechanism for these re-
ductions which he suggested in a paper published during the following
year, The mechanism indicates that the last three hydride transfers,
involving alkoxyborohydride ions, are rapid. If this is the case, it
is difficult to imagine how the amine could prevent rapid utilization
of the total hydride content once the rate-determining first step had
been accomplished. His data, however, show that only 25 per cent of
the hydride content had been used at the end of six hours. Furthermore,
his data also indicated that only 17.2 per cent of the hydride content
had been used after twelve hours in the absence of the amine. Now,
0.75 of this hydride contents or about 13 per cent, must have been
due to the loss of the last three hydrides, since this occurs rapidly
once the first hydride is lost. In the came period of time, the re-
duction mixture with the amine lost 32 per cent of its hydride con-
tent, only 7 per cent more than the 25 per cent lost within the first
few minutes of reaction. From these results, one can hardly conclude
that the alkoxyborohydride ion is present in the amine reduction
mixture, much loss its steric interaction with the amine.
A more reasonable explanation for the lack of further reduction
following the first hydride transfer in the presence of the amine is
that the borine fragment is completed by the amine. This complex is
then sufficiently stable under the reaction conditions to avoid the loss
of any additional hydride ions for several hours (at least six hours,
according to Brown's data).
Certain observations made during the course of the present
investigation strongly suggest the presence of the amine-borane complex
in these reduction mixtures. If the reduction mixtures are simply
hydrolyzed in acid solution (not subjected to the KIO3 workup explained
in the experimental section), infrared indicates a B-H stretch at 4.2-
4.3 microns for the pentane extract of this hydrolysis mixture. If the
hydride had not been rendered stable to acid attack by being completed,
these conditions would normally be severe enough to cause any hydrogen
bonded to boron to be liberated rapidly as molecular hydrogen. Synthe-
sis of the triethylamine-borane and comparison of its infrared spectrum
with that of the reduction mixture confirmed its presence in the re-
duction mixture beyond any reasonable doubt,
Further evidence that the amine is involved in a complex was pro-
vided by gas chromatographic analysis of this same pentane extract and
the observation of a peak corresponding to triethylamine (evaluated by
a separate injection of the pure amine). The acquisition of this infor-
mation was accidental and proved to be costly, because it resulted in
the gas chromatographic column being made useless for further analyses
of the alcohols.
The danger of ruining the gas chromatographic column by injecting
a reduction mixture containing residual amounts of the amine-borane
complex requires some additional comment. It was noted on the gas
chromatograph which recorded this incident that, although the peaks for
the alcohols were absent, a large peak of short retention time appeared
which was not recorded for the 1:4 reduction mixtures. To ascertain the
cause of this peak, some triethylamine-borane was synthesized and in-
jected into the vapor fractometer. An identical peal resulted. This
was followed by an injection of triethylamine, again with similar
results. It was concluded from this that the peak was due to tri-
ethylamine. Therefore, the amine-borane complex must thermally decom-
pose on contact with the hot injection block to produce the amine and
diborane. The diborane must then react with the column pacing causing
it to absorb the alcohols, because their peaks were never observed
thereafter using that particular column,
A series of reductions was carried out involving the borohydrides
of the same three cations used previously with 4-t-butylcyclohexanone
in the presence of a tertiary amine in a 1:1:1 ratio. The intent of
this innovation, to reiterate the considerations discussed above, was to
use one hydride only from each borohydride ion, the remaining BL,
fragment completing with the amine to form a so-called amine-borane
complex. Thus, the stereochemistry of the first hydride transfer could
R2CO + BHF R2CHO + B113
BH3 + (C2115) 3N-- (C25)3 311: B3
be observed independently of the remaining three. The results of this
series of reductions are summarized in Table 2.
S Table 2
Calc'd avg. % cis-isomer
Avg. pk. ht. for the last three
Borohydride ratio % ci-isomer hydride transfers
Sodium 0.242 16.4 22.5
Magnesium 0.209 15.2 22.1
Lithium 0.166 12.8 16.1
The results for sodium and magnesium are essentially the same,
at least within the experimental error inherent in the analytical
method. The lithium gave slightly less of the cis*isomero which is
consistent with the results of the lithium borohydride in the 1:4
reduction, In that reaction the average per cent cis-isomer for all
four hydride transfers was less than that of the other cations, so it is
reasonable to expect a proportionately smaller yield of cis-isomer for
the first hydride transfer of the lithium borohydride in this series for
.the same reason previously described.
It is reasonable to suspect that reduction may not cease entirely
with the formation of Et3N:BH3 in the presence of excess ketone, since
Brown's data indicates a decrease in hydride content of more than 25 per
cent after twelve hours. However, even though the reductions in this
series were run for as long as two days, insurance against further re-
duction after the first step was obtained by adding only sufficient
ketone to utilize one hydride from each borohydride ion, It has been
already pointed out that the amine may participate in the first hydride
transfer, If this possibility actually obtains, then a reducing species
which is bulkier than the borohydride ion may be anticipated. This should
lead to an increase in the cis-isomer content of the product because of
the increased steric requirements of a larger reducing species. Therefore,
the values in Table 2 represent a maximum value for the first hydride
These results allow calculation of the average stereochemistry of
the last three hydride transfers from the borohydride ion. This calcula-
tion is based on the following equation, where X is the average per cent
e-issomer for each of the last three reduction steps,
Per cent c.-isomer from first step + 3X/4 = per cent cis-isomer from
all four steps
These calculations have been made for all three of the borohydrides used
and are entered in the last column of Table 2.
An independent calculation of the stereochemistry of the first
hydride transfer was made using a calculated average of the last three
steps. The calculation of the average stereochemistry for the last three
hydride transfers was made by simulating the most probable situation
following the intermolecular transfer of the first hydride, 3 ketone
plus 0,9 alkoxide plus excess diborane. This reaction yielded 19,4 per
cent cis-isoner. Hydrolysis of a sample of the alkoxide showed it to
contain about 1-2 per cent c s-isomer, The calculation was as follows.
2.0 + 3X / 3.9 19.4 X = 24.5 per cent cislicomer
The 24.5 per cent doesn't correspond exactly to the other value of 22.5
per cent, but it is sufficiently close to provide support for a figure
somewhere in this range. This result and the result previously obtained
for all four hydride transfers (21.0 per cent) permit calculation of the
first reduction step as follows.
X + 3(24.5) / 4 = 21.0 X = 10.5 per cent SJ_-isomer
Now, because a 1 per cent error either way in the 24.5 or 21.0 per cent
figures can cause X to vary from 3.5-17.5 per cent, one cannot be certain
of the significance of this calculated value of 10.5 per cent. It is,
however, in the right direction to support the experimental results pre-
sented here which indicate that less ciS.isomer is produced in the first
reduction step than in the last three.
In conclusion, the borohydride ion was suggested as the reducing
species for the first reduction step. According to the principle of
steric control, it should give less of the cis-isomer than the more
bulky alkoxyborohydrido groups which follow in the last three reduction
steps, because it can approach the carbonyl group from an axial direction
with less steric interaction, Therefore, the stereochemical results of
this series are entirely consistent with both the previously suggested
mechanism for borohydride reductions and the partial control of
diastereomeric ratios by a steric factor.
Reduction of the ketone with NaBH4 in the presence of tertiary amines
In order to investigate the possibility that structural changes in
the tertiary amine used might change the stereochemical consequences of
its presence in the reduction mixture, amines other than triethylamine
were considered. The amine, ketone and sodium borohydride were combined
in the usual 1:1sl ratio to give the results in Table 3.
Amine Kb Avg4 pk. ht. ratio % cis-isomer
Triethylamine 5.65 x 10*4 0.242 16.4
Pyridine 1.4 x 10*9 0.261 18.6
N,N-Dimethylaniline 2.4 x 10*10 0.308 21.0
Triisopropylamine was also originally included in the series, but
its amine-borane complex proved to be so stable that it resisted all
attempts at its decomposition. Consequently, the stereochemistry of
reduction in its presence was not determined.
Although it is obvious from these results that there is no relation
between the stereochemistry of the reduction and the stereostructural
aspects of the amine, it will be noted that there appears to be almost a
direct correlation between the basic strength of the amine and the pro-
portion of cia-isomer obtained, This can be explained in the following
manner. It is reasonable to assume that the stability of the amine-
borane complex, formed after the first hydride transfer from the boro-
hydride ion, will be determined by the ability of the amine to function
as a Lewis base. Thus, the pyridine and N,N-dimethylaniline are weaker
bases than the triethylamine and their complexes would be anticipated to
be less stable* It has already been demonstrated that negligible re*
duction occurs after the first hydride transfer in the presence of tri-
ethylamine, so that the stereochemistry of this reaction can be con*
sidered a limiting case for the series. The fact that a higher pro-
portion of the cis-isomer was obtained for the pyridine and N,IU-
dimethylaniline is consistent with the possibility that their complexes
were not stable enough to prevent further reduction. If this is the case,
the stereochemistry of the reduction involving pyridine and N,N-
dimethylaniline would be expected to approach that of the borohydride ion
alone. This, of course, was observed.
Reduction with diborane
To our knowledge, the stereochemistry of the diborane reduction of
a carbonyl group has never been examined Geperimentally. The proposed
mechanism for the reaction postulates an initial interaction of the BB3
with the carbonyl oxygen, followed by intramolecular hydride transfer.
It was, therefore, of interest to obtain stereochemical data for this
reduction and evaluate its consistency with the proposed mechanism.
According to Burg, Brown and Schlesinger, excess diborane reacts
readily %with acetone to produce the corresponding dialkoxy derivative
of borine, (RO)2BH. they attempted to isolate the monoalkosy deriva-
tive, but were unable to do so, This indicates that the first two re*
duction steps with diborane must be rapid and the third step is slow
With an excess of the ketone the trialkylborate ester was formed, in-
dicating that the BRE was forced to utilize all three of its hydride.
When these two.diborane reductions were carried out in the present
investigation using 4-.-butylcyclohexanone, a difference ia stereo-
chemistry was observed# The results are shown in Table 4.
Type of reduction Avg. pk. ht. ratio % cig-isomer
Ketone + excess diborane 0.0813 7.6*
Diborane + excess ketone 0.180 12,6
*This is an approximate value, since the gas chromatographic standard
curve is not considered reliable in this range.
rhus, when only two of the hydrides are utilized, 7.6 per cent is ob-
tained, whereas use of all three hydrides yields 12.6 per cent of the
These results are consistent with a change in stereochesistry for
the last hydride transfer which may be calculated as follows.
2(7,6) + X / 3 12.6 X 22.6 per cent is*isomer
They also lead one to suspect an intermolecular hydride transfer for this
step. This is because an increase in the amount of cSO-isomer generally
means a larger reducing species being forced to approach the carbonyl
group more equatorially, transferring its hydride intermolecularly, to
give the axial hydroxyl group in the final product.
To determine if any equilibration of the isomers was taking place
under the reaction conditions, a mixture of the alcohols containing
more than 40 per cent of the cis-siomer was treated with diborane.
Analysis of the mixture following this treatment showed the isomeric
ratio to be unchanged, If equilibration had occurred, one would antici-
pate an increase in the amount of .trans-isomer.
At the same time, these results are somewhat puzzling in the light
of Bromw's suggested mechanism7 for diborane reductions, because one
would expect the dialkoxyborine to function as a Lewis acid, coordinating
initially with the electrons of the carbonyl oxygen and then transferring
its hydride intramolecularly. If this expectation actually obtains, it
becomes difficult to rationalize the increase in cis-isomer for the final
step of the diborane reduction because the dialkoxyborine-oxygen complex
should be forcing the transition state toward the equatorial position,
as shown below.
S+ .I ..---Equatorial position
The transfer of the hydride should then logically occur on the axial
side and should lead to an increase in the trans-isomer content of the
product. A way out of this dilemma presents itself in the postulation
of an intermolecular hydride transfer for the third step of the diborane
reduction and one is forced to seriously consider this possibility in
view of the stereochemical results. This transfer is structurally
represented as follows,
(RO) 2B + R2CO B (OR) 3
further elucdation of the mechanism of borohydride reductions from
the stereochemical results
Evidence has been presented thus far in this discussion to indi-
cate that the reducing 'species involved in the first hydride transfer
is the BH4" ion. Calculations made possible by experimental results
from the various reductions have established that the proportion of
c*ASisomer produced in each reduction step tends to increase with suc-
cecding steps. This has been the case for both borohydride and diborane
reductions and is consistent with the production of a progressively
bulkier reducing species following the transfer of each hydride.
the question then becomes, from what species are the last three
hydrides transferred? Two series of hydride donors become possibilities.
One is BH3 and its alkoxy derivatives, ROBH~ and (RO)2BH. The other is
the alkoxyborobydridee, ROBH3 (RO)2BHU and (RO)3BH suggested by
Brown. To distinguish between these, the stereochemistry of a reduction
in which the BH3 was forced to use all three hydrides was found to
average 12.6 per cent cis-isomer. It may be reasonably concluded from
this that the last three hydride transfers do not involve the alkoxy-
borine series because the results would be similar to those previously
calculated (22.5-24.5 per cent) for these steps. Therefore, in the last
three hydride .transfers* one and possibly all three involve an alkoxy-
boolhydride ion as the reducing species.
Brown't work7,30 indicates the relative reducing properties of the
hydride donors in the two series to be as follows.
BH3) ROBH2) (RO)2BH
(RO)3BH ) (RO)2BH- > ROBH3
The (RO)3BH ion has been demonstrated by Brown (for the case of tri-
isopropoxy-borohydride) to reduce more rapidly than the BI4" ion. The
(RO)2BH is unlikely as the hydride donor for the fourth step because it
is less reactive than its two predecessors. The (RO)3BE'" is therefore,
favored as the reducing species for the final step.
A further distinction in the nature of the reducing species involved
in the middle two steps should be possible if stereochemical results for
the last step could be obtained, These would be available from the re-
duction of 4-t-butylcyclohexanone by its trialkoxyborohydride ion. This
would allow the average stereochemistry of the second and third hydride
transfers to be calculated, since the stereochemical results of both the
first and all four reduction steps are known. There is good evidence
from the work of Brown, Burg and Schlesinger7 that the reaction of the
ketone with excess diborane involves the SBH and ROBB2. The average
stereochemistry of these two reduction steps has been experimentally
determined in this investigation. Therefore, by a comparison of the
stereochemistry of these two steps with that calculated for the second
and third steps of the borohydride reduction, one should be able to draw
some tentative conclusions as to whether steps two and/or three involve
an alkoxyborohydride ion.
Unfortunately, all attempts to prepare the sodium trialkoxy-
borohydride of 4-t-butylcyclohex-anol gave a product which on analysis
proved to be of uncertain identity. Most disappointing of all, perhaps,
was that this product had no apparent reducing properties. Even the
somewhat different tack of ni::ing a diglyme solution of the ketone with
NaH in the presence of the borate ester of the alcohol in 1:1:1 ratio
failed to achieve any obvious reduction. All this, of course, made
further elucidation of the mechanism of borohydride reduction by the
above approach virtually impossible.
Reduction with aluminum borohydride
It was originally intended to study the stereochemistry of aluminum
borohydride reduction as an example of the behavior of Group II1 boro-
hydrides, so that the results could be compared to those of Groups I and
II, in an effort to evaluate the effect of the cation on the stereo-
chemistry of these reductions in a homologous fashion. However, as the
acquisition of experimental results shown in Table 5 progressed, it be*
came increasingly apparent that they could be more profitably discussed
separately from the results obtained from the borohydrido reductions
involving the other cations.
The reductions of 4-t-butylcyclohexanone with aluminum borohydride
were of two types, those carried out in the presence of triethylamine
and those in which the amine was absent. The stereochemical results are
summarized in Table 5.
In the absence of amine, the reducing species may be the undis-
sociated Al(BH4)3, or it could conceivably be some dissociated form
which yields free borohydride ions. The results of these reductions
are so similar to comparable ones of the other borohydrides that, if
the latter case prevails, the stereoicomeric ratios could be explained
in the came manner.
Molar ratios in reduction mixture
Al(BH4)3 Ketone Et3N Avg. pk. ht. ratio % cis-isomer
1 12 0.299 20.3
1.5 12 0.295 19,9
1 2.2 2.5 0.342 23.5
1 3 8 0.453 32.5*
*The gas chromatography standard curve used did not extend as far as the
peak height ratio obtained for this product. Therefore, the standard
curve for another similar column was used, .which did cover this range,
and indicated a value of 34.5 per cent for thick peak height ratio.,
Since the slopes of the two curves were about comparable, comparison of
values in the range where they covered the same ratios showed a -2 per
cent difference. Hence, subtraction of 2 per cent from the value
obtained from the other curve gave 32.5 per cent.
However, if the former possibility obtains, the similarity of the
results would be simply coincidental because in all likelihood a larger
reducing species is involved in the first step which is considerably
different in character from the borohydride ion itself. Schlesinger
and Burg3 interpret its structure as containing three BH4" groups, which
are so extensively distorted by the small highly charged aluminum ion,
that the ionic character of the molecule has largely, if not completely,
been lost. An examination of the physical properties of aluminum
borohydride also indicates the bonding involved to be much more covalent
in nature than that of the other borohydrides,1 i.e., lower m.p., b.p.,
higher volatility, etc.36
In those reductions involving the amine, it will be noted that
increasing the concentration of the amine causes.a sharp increase in
the cis-isomer content of the product. Theoretically, each mole of
A1(BH4)3 should require three moles of amine to stop reduction after
the first hydride transfer from each of the three borohydride.groups.
However, if it is assumed that the aluminum ion itself, because of its
electronic configuration, is capable of acting as a Lewis acid toward
the amine, then a fourth mole of amine could conceivably be utilized by
the reducing species, Inclusion of a two-fold (4 x 2 = 8) excess of the
amine in the reduction mixture resulted in 32.5 per cent aci-ioamer,
which is probably the maximum value obtainable for this type reduction.
These results argue strongly in favor of the aluminum borohydride
functioning in an undissociated form, at least for the first reduction
step, because free borohydride ions have been previously shown to givo
only 16.4 per cent cis-isomer in the presence of triethylaminie.
The remaining question is why the two reductions involving amine
differed in their stereochemistry with a change in the relative amine
concentration. It will be noted that, in the reduction in which the
amine was not in excess, there is less than the four moles which could
be theoretically utilized by the reducing species~s In fact, there is
not even enough for the three borohydride groups. Hence, the 23.5 per
cent cS~-isooer obtained is probably a result of the reduction being
allowed to proceed beyond the first step because of insufficient amines,
If this is the case, the diastereomeric ratio could be anticipated to
approach that of the reduction in which no amine was involved. This
would account for the observed results.
Summary of conclusions from the results
Lithium borohydride was found to give less cSgjalcohol than either
sodium or r-,Gnesium borohydride when all four of the hydrides were
utilized. Differences in the degree to which the cation interacts with
the carbonyl oxygen was advanced as an e:-=plnation for these results.
Stereochemical data for aluminum borohydride reductions was obtained in
the absence and presence of triethylamine. This information is con-
sistent with the aluminum borohydride functioning as an undissociated
molecule for the first hydride transfer.
A means of investigating the stereochemistry of the first hydride
transfer from the borohydride ion was made available by reduction in the
presence of triethylamine, because of the formation of an amine-borano
complex following the first hydride transfer. This was found to give
less of the SIg-alcohol than the average of all four reduction steps.
Therefore, more cls-alcohol must be produced in the last three re-
duction steps than in the first, which is consistent with a progressively
more bulky reducing species being forced to approach l-ho carbonyl group
from an equatorial direction.
The influence of structural changes in the tertiary amino used on
the stereochemistry of the reduction was investigated by considering
amines other than triethylamine, Ho relation was found between the
stereostructural aspects of the amine and the stereochemistry of the re-
duction. However, there appeared to be an almost direct correlation
between the basic strength of the amine and the proportion of c9s-isomer
produced. The stability of the amine-borane complex formed was ad-
vanced as an explanation for this relationship.
Previously unavailable stereochemical data for diborane reductions
was obtained which indicated a change in stereochemistry for the last
reduction step. The direction of change was consistent with an inter-
molecular transfer of the last hydride from diborane.
The reaucrig species ,involved : the first hydride transfer of
borohydride reductions is logically the -Sia toa* 'The average stereo-
chemistry of the last three steps has beea calculated. When these
calculated results were comparOr to those obtained when diborane was:
forced to exhaust its hydride content, it was concluded that one or all
of the reducing specie's i the last three reduction step are alkoxy-
borohydride ions, as opposed t tothe involvement of borne and its alkoxy
Diglyme, Ether 141
NaH, 53 per cent dis-
persionn mineral oil
B?3 etherate complex
Metal Hydrides, Inc.
Metal Hydrides, Inc.
Amend Drug and
Ansul Chemical Company
K and K Laboratories
Dow Chemical Company
Metal Hydrides, Inc.
Procedure of Brown33
Mead and Rao. Reflux
and distill over CaH.
vacuum over WAII as
from pentane to bring
m.p. within 2 degrees
(free from mono)
Distilled over acetic
wver KOH pellets after
drying over anhydrous
None (reagent grade)
I er -*.ir *- k 'A-*1 I -1 -'l = Iw I .. -0^
M-t i l
Diglyme solutions of the borohydrides and hetone
Diglyme solutions of NaBH4, LBH4 and ketone were made up so as to
contain as close to one male per milliliter as possible.
Sodium ~broaydride solution. A 100 milliliter volumetric flask
was baked dry, flushed with dry nitrogen, corked and tared accurately.
It was then placed in a dry box under an atmosphere of nitrogen which
had been scrubbed with concentrated sulfuric acid and passed over
magnesium perchlorate. Because it is quite hygroscopic, the NaBH4 was
weighed approximately on a triple beam balance in the dry box. The
flask and its contents were then removed from the dry box and weighed
accurately. The flask was reintroduced into the dry box and freshly
distilled diglyme added until the solution was up to volume. Since the
solvation of NaBH4 is an exothermic process, the flask contained some-
what less than the correct volume when its contents cooled to room
temperature, so additional diglyme had to be added at that time. Even
after swirling for some time and allowing the solution to stand over-
night, a fine precipitate persisted on the bottom of the flask. The
supernatant liquid was withdrawn from the volumetric flask, still in the
dry box, with a long-needled hypodermic syringe and placed in a bottle
fitted with a rubber stopper suitable for multiple syringe withdrawals.
The solution was then removed from the dry box and one milliliter
aliquots titrated with standard KIO3 to determine the actual hydride
content. 'roma this the solubility of NaBH4 in diglyme was found to be
0.827 mole per milliliter,
Lithium borohydride solution. A 25 milliliter volumetric flask
was baked dry, flushed with dry nitrogen, corked and placed in the dry
box. Enough LBB4 was then weighed into the flask so that it exceeded
the solubility limits of the salt in 25 milliliters of diglyme (0.4493
grams),. It was made up to volume in diglyme and swirled, to promote
solution. This solvation process was also exothermic so that similar
precautions bad to be taken concerning the final makeup to the correct
volume, All of the UBIt4 did not dissolve because of the deliberate
excess of the solubility limits. One milliliter aliquota of the super*
natant liquid were withdrawn still in the dry box, and removed for
titration with standard KI03 for hydride content. From this the solu-
bility of LiUR4 in diglyme was found to be 0.5602 mvole per milliliter.
Ketone solution. Since the ketone is not hygroscopic, it was
weighed accurately outside the dry box into a 100 milliliter volumetric
flask which had been baked dry and flushed with dry nitrogen* The flask
was then introduced into the dry box and diglyme added up to about 5-
10 milliliters less than volume, The solvation of the ketone was an
endothermic process, so that on coming to room temperature, the volume
increased. The solution was then made up to volume, transferred to a
syringe bottle with the special rubber stopper and taken out of the dry
box. The ketone was extremely soluble in diglyma, so that no trouble
was encountered in getting 0.1 mole into the 100 milliliter of solvent.
Method of analyzing the borohydrides by titration
Determination of the tnoles of boron per gram of sample.7 The
sample was weighed (precautions taken against hydrolysis by moisture
in the air) and hydrolyzed by addition to water. Sodium ion was analyzed
as base by titrating to a methyl red end point with standard 0.1 N qHC
and boron as boric acid by titrating, after the addition of mannitol, to
a phenolphthalein end point with standard 0.1 N NaOH. hese two ti-
trations were run on the same sample. From this data, the mmoles.of
boron per gram of sample was calculated.
Determination of the moles of hydride tesr aram of sample.10 A
weighed sample was added to a known excess of standard KIO3 solution.
Water and an excess of IK were added after a few minutes of stirring.
The solution was then acidified with 10 per cent 0C1. If iodine pre-
cipitated, more EI was added until it all went into solution. After
neutralizing with NaBCO3 to prevent air oxidation of iodide to iodate,
the iodine solution was titrated with standard sodium thiosulfate
solution, using starch as an indicator, to a colorless end point. The
difference between the ieq. of K103 and Na2S203 indicated the number of
meq. of hydride reduced. Since there are two mteq, per mmole of hydride,
the number of meq, of hydride had to be divided by two to calculate the
number of roles of hydride per gram of the ample.
Calculation of the /lJ Since the mole of both hydride and boron
per gram of sample were now hInown, the ratio of moles hydride per mole
boron in the sample could be calculated.
General procedure for borohAdrde reductions of i the kot.on
A specially prepared 13 milliliter tube, sealed at one end, was
baked dry, flushed with dry nitrogen and capped with a rubber stopper
suitable for injections. The solutions were injected in appropriate
amounts using a hypodermic syringe of 2 ec. capacity, which had also
been baked dry. The tube was then placed in a water bath at 20 degrees
for two or three days, during which time it was removed periodically and
agitated. As far as specific quantities are concerned, the borohydrides
were usually used in 3 millimole quantities and the I:etone and amino
added in proportionate quantities depending on the ratio under con-
The general workup procedure for a 1:4 reduction mixture was to
pour it into 1004*150 milliliters of water, add several drops of con-
centrated IE1 and shalce thoroughly in a separatory funnel to hydrolyze
the alkoxyborohydride solid which formed in the reaction tube several
hours after injection of the reduction mixture. The alcohol usually
floated to the surface on standing after the shaking. Pentane was then
added to dissolve the alcohol and the entire mixture was shaken to
extract any remaining alcohol. The aqueous layer was then discarded and
the pentane layer washed twice with dilute HC1 and, finally, once with
water. The pentane solution was dried over anhydrous sodium sulfate and
concentrated on a steam bath to about 5 milliliters through a one foot
column packed with steel sponge.
The general procedure for working up a 1:1:1 amine reduction
mixture was to pour it into a mixture of 10 milliliters of isopropyl
alcohol and 10 milliliters of 10 per cent K103 solution which had been
mixed in such proportions with water (usually about 20 milliliters)
that a clear solution was obtained. The purpose of this was to dcccn-
pse the amine-borane complex which formed after the first hydride
transferred from the borobydride ion, After shaking this mixture in a
sepazatory funnel and allowing it to stand for about fifteen minutes,
pentane was added to extract the alcohol. After draining the aqueous
layer, the pentane layer was washed twice with dilute HC1 and once with
water. Sometimes iodine formed in the aqueous layer upon addition of
the acid. This was removed by washing with 10 per cent sodium thio-
sulfate solution followed by water. The pentane solution of the alcohol
was dried over aphydrous sodium sulfate and evaporated as in the case of
the 1:4 workup. Before this solution was injected into the column, it
was checked for B3H stretch (4.2*4.3 microns) by infrared to avoid
ruining the column, as explained in the discussion of results. If the
B-H stretch appeared in the spectrum, the workup procedure was repeated.
Magnecium borohydride reduction mixture
Kollonitsch, Fuchs and Gabor14 reported the preparation Of alkaline
earth borohydrides by the reaction of a3BH4 with the appropriate metal
chloride. These reactions were carried out in extremely dry alcoholic
solutions, giving about 80 per cent yield based on the amount of unre-
acted WaBH4. Application of this metathetical process for the production
of It(H4)2 to the problem at hand consisted of generating the magnesium
borohydride Z situ by mixing NaBI4 with anhydrous IgC12 in diglyme and
adding the ketone to this mixture. To ascertain whether the r~anesium
ion was actually involved in the reduction step, a series of four re-
ductions was set up in which the quantity of FMCl was varied from
stoichiometrically sufficient to 0.25 that amount. All of these re-
ductions gave essentially the same isomeric mixture. However, the
extent of reduction was directly dependent on the amount of 11C012 present
for equal reduction times. The proportions of the reactants were:
4 ketone + 1 NaBH4 + X KgCl2.
X 4 Avg. pk. ht. ratio % cis-isomer
0.500 0.275 19.6
0.375 0.297 21.2
0.250 0.290 20.7
0.125 0.297 21.2
i ~ ~ ~ ~ - II-lJI IlI
Zaducti. with diborane
Diborane was generatedd external to the reaction flask by adding
diglyme saturated with NaBH4 to 1F3 etherate. The generator consisted
of a 50 milliliter flask fitted with a stopcock. joint and oblique side
arm capped with a rubber serum bottle cap through which injection of the
NaBUH solution took place. The generation procedure was to flush the
system with nitrogen admitted through the side arm yv a hypodermic
needle. The stopcock Was then closed and the joint removed with the
nitrogen continuing to flow. Apprcximat .ly 25 milliliter of BF3
etherate was added to the flask amid voluminous clouds of fumes and the
stopcock joint reinserted. After additional flushing of the system,
injection of the NaBH4 solution was initiated and continued in increments
sufficient to create a steady flow of gas as evidenced by the bubbles
in the reaction flask. Nitrogen flow was stopped when generation of the
diborane began. Diborane which escaped reacting in the reaction mixture
was exited through an acetone trap and vented to the hood.
Previous to this operation, solid ketone and, in one reaction,
sodium salt of the alcohol had been weighed and added to the reaction
flask while it was being flushed with nitrogen. Enough diglyrm was
added to bring the end of the addition tube* fitted with a scintered
glass plug, just below the surface. The reaction flask contained a
magnetic stirrer which was used to agitate the contents from time to
time. The connecting joints Were all secured to their respective flasks
by means of wire and rubber bands to prevent the positive pressures
developed from within from dislodging them. The reaction flask was later
modified to a tube with a side arm and the magnetic stirrer discarded,
To acetone trap
Apparatus for Diborane Reductions
agitation being provided by the bubbling action. All apparatus, in-
cluding the hypodermic syringe, was dried in an oven prior to use. The
reaction tube and generator are illustrated in Figure 1.
The workup of these reduction mixtures was identical to that of
the 1:4 borohydride reduction mixtures described previously.
Preparation of the sodium salt of 4-Z-butylcyclohexanol
Several methods of preparing the alkoxide were attempted before a
usable product was obtained, These include tl.- -caction of the alcohol
with small cubes of sodium metal in dry hexane in a nitrogen atmosphere
and the reaction of the alcohol with sodium metal in a sealed tube
heated to about 110 degrees. Analysis of the product obtained by these
methods was carried out by hydrolyzing weighed samples and titrating
with standard acid. The results indicated contamination of the product
to be as much as 50 per cent, both high and low. Cunningham relates
a method of preparing an alcoholate of more than four carbons by re-
fluxing the alcohol with another alcoholate of lees than four carbons,
the two alkoxy groups mutually displacing one another and the more
volatile alcohol distilling off to shift the equilibrium in favor of
the alcoholate of more than four carbons. This method wan tried by
attempting to cause 4-g-butylcyclohexanol to react with sodium methylate
in diglyme. IHoever, no methanol was distilled over at temperatures up
to 100 degrees, so the procedure was abandoned. Synthesis of an alkoxide
product suitable for use was finally effected by stirring a mixture of
53 per cent NaH-mineral oil dispersion and alcohol in diglyme at a
temperature of about 130 degrees until evolution of hydrogen (volume
measured on a wet test meter) had ceased. The slurry was filtered under
nitrogen and the residue transferred in a dry box to a sublimation
apparatus where the remaining diglyme and alcohol were removed by
heating overnight under vacuum. Analysis of the product was as pre-
viously indicated and, in addition, the alcohol resulting from hydroly-
sis was recovered and weighed. The recovered alcohol was analyzed by
gas chromatography and found to be essentially pure trans-icomor (1*
2 per cent pis-isomer).
In analyzing the product, it was assumed that all of the alcohol
was removed from the reaction product by sublimation and that the
principle contamination is NaOH. Using two equations and the two un-
lmowns, the per cent alkoxide found in two samples was 78 and 77 per
cent. Calculating the per cent alkoxide from the recovered alcohol
of the same two samples gave 68 and 74 per cent respectively. A value
of about 70 per cent purity was used in calculations involving the
alkoxide. The analyses above would indicate this to be a valid esti-
mate of the minimum alkoxide content of the product.
Preparation of sodium trialkoxyborohydride and attempts to utilize it
in the reduction of the ketone
Into a 250 milliliter 4-necked flask fitted with a thermometer,
nitrogen inlet and wet-test meter was placed NaBiH and the commercial
mixture of the alcohols in a 1:3 ratio respectively. The system was
flushed with nitrogen and diglyme added to dissolve the contents of
the flask. The mixture was heated to 130 degrees with agitation by a
magnetic stirrer and maintained under these conditions for about five
days until no more hydrogen was evolved, as indicated by the wet-test
meter. Comparison of the volume of hydrogen obtained with the
calculated amount indicated the reaction to be 97 per cent completed.
A thick slurry gradually formed during the course of the reaction. The
solid was filtered under nitrogen i~ a special saintered glass filtering
apparatus and the residue dried by heating under vacuum.
Analysis of the product consisted of establishing its H/B by
titration and microanalysis for carbon end hydrogen. Two of the above
preparations were:undertaken. Titration analysis of the first gave a
H/B of 0.499, while that of the second wa. 0*209 Microanalysis:.;:
Calculated C 71.97; H, 11.68. Found C, 68,05; H# 10.81. tydrolyr-r
f, this product, gave the pure tran-alcohol as indicated by gas chroma-
S Two reductions were undertaken with this product. First, 1.1116
grams (2.224 millimnles,. assuming possession of NaBH(OR)3) of the product
was mixed tith 10 miltillters of diglyme and to this heterogeneous
mixture was added 2.224 millinoles of the ketone in solution. After..
24 hours, only slightreduction of the ketone was indicated. No esti-
mate of the isomeric ratio was possible from these quantities of the
alcohols produced in the reduction.
The second reaction attempted consisted of adding 0.7962 grams,
(1.593 millimolea) of the product to 10 milliliters of diglyme in a
tube fitted with a stirrer and sido-arm through which injections could
be made. To this was added 1.593 millimoles of ketone in solution.
After one week, only slight reduction of the ketone was indicated.
One last attempt to achieve the probable situation of the final
reduction step was made by preparing the borate ester Of the alcohol
and adding it to NaH and the ketone JAi a lt:i ratio. Preparation of
the borate ester42 was effected by the reaction of boric acid with
4-t-butylcyclohexanol in toluene solution and the use of a Dean-Stark
trap to allow removal of water from the toluene-water azeotrope. On
achieving approximately the theoretical quantity of water by refluxing
the reaction mixture overnight, the toluene was removed, first by
distillation at atmospheric pressure and finally under vacuum. Cold
acetone, which had been rendered reasonably dry by treatment with CaC12,
anhydrous sodium sulfate and distillation, was then added to the residue
and the mixture was triturated under an atmosphere of dry nitrogen,
washed on the filter with additional cold acetone and dried by low heat
under vacuum. The solid had a m.p. 113-115 degrees.
The 51 per cent NaH-mineral oil dispersion and the borate ester
were carefully weighed by the usual technique of transferring stoppered
tubes in and out of the dry box. The diglyme solution of the ketone
was then added by hypodermic syringe, additional freshly distilled
diglyme being added because the slurry in the tube was too thick for
effective mixing of the materials by shaking. After standing for three
days, with intermittent agitation, the mixtures were worked up for
analysis. A comparison of the peak height ratios of the hydrolysis
product of the borate ester and that of the reduction mixture indicated
Analysis of the precipitate from a 1:3 NaBH?-ketone reduction mixture
A reduction was carried out in the usual manner involving a 13
mixture of sodium borohydride and ketone respectively. After about two
days reduction time, the precipitate was filtered under nitrogen and the
residue dried by heating under vacuum.
Analysis by titration indicated the H/B was 0.0935. Microanaly-
sis: Calculated C, 71.97; HU 11.68. Found (duplicate analysis)
C, 71.91; Ho 11.76; end C, 72.12; H, 12.09.
P duction of aasnuM borohvdride and the procadre utilized to
introduce it into a reduction mixture
The procedure used for the production of aluminum borohydride is
that of Schlesinger, Brown and Hyde.43 Quantities of materials were
scaled down to 0,1 those used by these investigators in the preparation
described. The design of their vacuum system was modified slightly to
include a line connecting the manifold to the trap train, so that
manometric changes could occur independently of the traps. An outlet
to transfer the aluminum borohydride to a reaction tube and fo cali-
bration purposes was also added* These are illustrated in Figure 2.
Calibration of the system was effected with a bulb of known
volume (212.3 milliliters), fitted with a stopcock and 10/30 s.t.
male joint. The portion of the entire vacuum line defined as the
system was bounded by stopcocks As C, D, P and H (refer to Figure 2).
The procedure was to pump the system down with the bulb stopcockk
closed) connected to the outlet. The manometer reading at this point
was assigned a sero pressure value. After closing stopcock A, the air
in the bulb (at atmospheric pressure PI) was allowed to expand into
the system by opening the bulb stopcock. The manometer was again read
and the pressure differential between this reading and the former
constituted the final pressure P2 within the system. The volume of
the system was then calculated in the following way.
Apparatus for the Production and Handling of Aluminum Borohydride
Atmospheric pressure, P1 758.3 millime cr-
Average pressure differential,. P 369.2 millimeter,
Volume of bulb 212,3 milliliters
Volume of system + bulb: 212.3 x Z8 =c 436.0 milliliters
Volume of system: 436.0 212.3 223.7 milliliters
From this volume, the pressure differential required to put one milli-
mole of-gas into the system was estimated.
P = aRT/V (1x103)(62,360)(299)/ 223.7 = 82.8 millmeters per
mititaoles of gas
The molecular weight of the product was also evaluated by means of
the Duass method. The system was evacuated with the bulb (stopcock
open) attached to the outlet. Stopcock A was closed and the product,
whtch had been previously transferred to U-3, allowed to warm to room
temperature. When a pressure differential of approximately 200 milli-
meters was reached, stopcock B was closed along with the bulb stopcock.
0-3 was recoiled in liquid nitrogen, stopcock B opened and condensation
of the aluminum borohydride back into U-3 allowed to continue until no
further pressure decrease was noted. Then with stopcock E closed, the
bulb (previously tared) was removed with its stopcock closed and weighed
to determine the weight of the material in the bulb by difference. The
molecular weight of the material was then calculated as follows.
Sample weight 0.1532 grams
Pressure differential 201.0 millimeters
Temperature 301 degrees K
)N.W. = (wt.)RT/PW (0.1532)(62.360) (301)/(201.0) (212.3) = 67.38
Value calculated from atomic weights 71~57
A duplicate determination was made and gave a result of 67.37,
Reductions with aluminum borohydride were carried out in 13 x 130
millimeter tubes, sealed at one end and fitted with a stopcoch and
10/30 s.t. male joint at the other. Before placing any materials in
a tube, it was first evacuated ad flushed twice with nitrogen in-
troduced into the system ia stopcock G. Filled with nitrogen under
positive pressure, the tube was removed from the outlet with its
stopcock closed. Diglyme solution, amine, etc., were introduced into
the tube by means of a long hypodermic needle which extended down'm
through the stopcock. This minimized the amount of air present in the
tube. In one run, solid ketone was weighed into the tube before
sealing on the stopeock and diglym added later to give a solution more
concentrated than the stock diglyms solution. The tube was then re-
attached to the outlet and cooled in a dry-ice-acetone bath before
evacuating, so that the materials within wouldn't distill. After
thoroughly evacuating the system and reaction tube, the reaction-tube
stopcock and stopcock A were closed. U*3 was allowed to warm up until
the pressure rose sufficiently to introduce the appropriate amount of
aluminum borohydride into the system (see previous calculation). Wh en
this pressure differential was reached, stopcock B was closed, the
reaction tube placed in liquid nitrogen and its stopcock opened. Con-
densation of aluminum borohydride out of the system into the reaction
tube was allowed to continue until the pressure had been reduced to
its original reading. It was found in practice that exactly the same
original manometer reading was never quite achieved in these condensa-
tibns. If the deficit was as mach as 10 millimeters, more aluinum
borohydride was introduced into the system and condensed into the re-
action tube until the deficiency was made up. The reaction tube
stopcock was then closed, stopcock A opened and the system evacuated
and flushed with nitrogen before closing stopcock E and removing the
reaction tube, which was allowed to come to room temperature. This
precaution was necessary to remove all aluminum borohydride residue
from the system prior to exposing the outlet connecting joint to the
air. The aluminum borohydride explodes spontaneously and burns
rapidly (green, orange flames) on contact with air containing traces
of moisture. Although Brown points this out in his experimnntal dis-
cussion, it wasn't fully appreciated by this author until a small
amount of supposed by-product from the preparation gave more than ample
demonstration of this hazardous property.
The reduction took place over a period of two to three days at
Comparison of the relative volatility of the two isomeric alcohols
Some of the commercially available mixed 4-t-butylcyclohexanols
was sublimed. Pentane solutions of the sublimate, residue and the
starting mixture were analyzed by gas chromatography.
material Avg. pkh ht. ratio % cis-isomar
starting mixture 0.2714 19.3
Sublimate 0.4001 29.9
Residue 0.2977 21.2
These results indicate the c.s-alcohol to be the more volatile of the
two. The fact that the residue appears to contain more cis than the
starting material becomes somewhat difficult to rationalize in the
light of these results. However, the difference is only about 2 per
cent which may be approaching the limits of accuracy for the method of
analysis. This information was of interest in connection with a
possible method of separating the alcohols from the unrencted ketone
in the product prior to analysis.
Preparation of triethylamine-borane
Into triethylamine was bubbled diborane generated by addition of
NaBH4 in diglyme to BF3 etherate, as described previously, Any
unrencted triethylamine was removed by evaporation under vacuum"
Pentane was added to the residue and a solid which formed was filtered
off. This left a clear filtrate which was dried over anhydrous sodium
sulfate. Removal of the pentane under vacuum left an oily residue.
The product was analyzed with KIO3 and Na2S203 solutions for hydride
content and 96 per' cent purity was indicated..
Preparation of Pure g- and Trans-4-t-butylcyclohexanolo
(1) Preparation of tras-acid phthalate. This procedure was adapted
from that of Winsteein and Holness.32 Into a 250.milliliter flask fitted
uith a reflux condenser was placed 42 grams of the commercial mixture
of the alcohols, 42 grams of phthalic anhydride and 50 milliliters of
reagent grade pyridine. This was heated on the steam bath for five to
six hours. After cooling, the reaction mixture was poured into dilute
HC1 and extracted with ether. The ether extract was washed with dilute
acid and water, dried over anhydrous sodium sulfate and evaporated.
The residue was recrystallized three or four times from pentane to give
crystals melting 146-147 degrees.
'(2) Hydrolysis of _tran-acid phthalate. A 25 gram quantity of the
jtans-acid phthalate was dissolved in a solution of 20 grams of NaOH
in 100 milliliters of water and steam distilled, White clumps of solid
alcohol collected in the condenser which had to be allowed to warm up
periodically to avoid stoppage. The distillate was extracted several
times with pentane, the combined extracts dried over anhydrous sodium
sulfate and the solvent evaporated. The white solid residue had a m.p*
80-81 degrees and checked out satisfactorily as the trans-*alcohol using
(1) Preparation of &fn- butylcycloheyl-p*toluenesulfonate. This
procedure was also adapted from that of Winstein and Holness,. Since
that time, a more satisfactory method has been published utilizing the
formation of the 3,5-dinitrobenzoate.35 A solution of 9.60 grams of
the. trans-alcohol, 11I70 grams of p-toluene sulfonyl chloride and 60
milliliters of reagent grade pyridine was allowed to stand overnight
at room temperature. This solution was washed with dilute NatCO3
solution, dried and evaporated.- Recrystallization from pentane gave
98 grams of the tosylate m.p. 89-90 degrees (yield 51 per cent).
(2) Acetolyses 0of the tosylate. This procedure was adapted from that
of Stork and White.40 The entire 9.8 rrams of trans-tosylate was added
to 35 milliliters of glacial acetic acid containing 8 grams of anhydrouc
potassium acetate. This solution was kept on the steam bath for about
24 hours. It was cooled, poured into a liter of cold 10 per cent NaOH
and extracted with ether. The ether extract was dried over anhydrous
sodium sulfate and evaporated, The residue was hydrolyzed by refluxing
with 10 per cent NaOH for several hours and extracted with pentane. The
pentane extract was washed with water, dried and evaporated. The white
solid remaining melted at 81-82 degrees and was shown to be the qA*
alcohol by gas chromatography.
Gas chromatographic analyss ofthe. reduction product
A Perkin-Elmer Model 154*B vapor fractometer, operating at 140
degrees and 25-30 pounds per square inch and utilizing an 18-foot
coiled 1/4 inch copper tube packed with 30-60 mesh Tide detergent
(. and M Scientific Corporation) provided analysis of the reduction
Packing of the column was effected by allowing the correct length
of copper tubing to hang suspended down a stairwell, plugging up the
bottom end with glass wool and pouring in the detergent with sufficient
agitation of the entire tube by tapping until it accommodated no more.
After plugging up the other end, the tube was coiled about a pipe
approximately 2.25 inches o.d., with one end extending down through
the center of the helix whose coils were spread slightly after winding
so as to allow for a more complete heat equilibration during its
operation. Before the column was used for analysis, it was "baked"
in the heating chamber for about 18 hours at a temperature of 180-190
degrees and a helium pressure of 10 pounds per square inch to remove
excess liquid phase from the detergent packing, The liquid phase comes
off initially in a discontinuous fashion, finally leveling off to a
constant flow which continues throughout the operational life of the
column. If not removed prior to the column's use, this initial
effluent causes an erratic pattern of thermistor response which
interferes with the analysis.
The colta was standardized by making up samples of the two
isomeric alcohols of known composition and calculating the ratio of the
the cis peak (which occurs prior to the trang height to the sun of
the cis and trn peak heights for from three to five injections for
each sample, These ratios were averaged. A plot of per cent of cis
isomer versus average peak height ratio was then constructed from this
data, which is as follows.
Co.lumxn er 1 ..... Column twber 2..
% c~g-isomer Avg. pk. he. ratio % cis-isomer Avg. pk. ht. ratio
9.1 0.093 + 0.004
20.6 0.289 + 0.003
28.8 0.385 + 0.006
9.1 0.101 + 0.003
15.1 0.225 + 0.005
19.9 0.295 + 0.000
28.8 0.412 0.004
:! ~ "Z .r!I !H!
It can be seen that the average deviation from the average peak
height ratio ranges from 0.003*6, which is more than 1 per cent in
terms of the actual numbers involved. However, when this deviation is
evaluated in terms of per cent of cjl-isomer, the slope of the standard
curve is such that the error is less than 1 per cent. Therefore, the
analytical method appears to be easily within I per cent error. This
was interesting, for Eliel and Ro35 placed no greater confidence in
their results (measured the &ame way) than + 2 per cent. The gas chroma-
tograph for a typical reduction WAixurae is shown in Figure 3.
It was found that the value of this ratio would change for any
one sample as the absolute height of the peaks changed. Eowveer, when
the tallest peak (usually the .rans was approximately 12 centimeters
high, the ratio became reasonably constant. Therefore, the standard
curve and all analyses taken from it were derived fro ,trans peaks at
ca. 20 minutes retention time
Figure 3. Typical Gas Chromatograph of a Reduction Mixture
least 12 centimetersin height, Tto standard curves are included
because two columns were used in the overall studyy and each gave
slightly different values for the same standard samples. Peak heights
were measured from the base line ineodiately preceeding the peak, even
though this line was seldom the same for the i~d and tRas peaks,
The volume of the injections generally was 0.03-.05 milliliters,
depending on the concentration of the pentane solution containing the
isomeric mixture. In any event, volume sufficient to yield the retired
peak height was always used, trial injections sometimes being necessary
to determine this volume,
1, H. 1. Schlcsinger, H. C. Brown, t al., J, Am. Chem. Soc. J5,
.2. H. Schlesinger and A. B. Burg, iid. 4321 (1931).
3. H. I. Schlesinger, A. B. Burg, at_*, AidL l 3421 (1940).
4. H. I. Schlesinger, H, C. Brown, Hoekstra and Rapp,bkid. 75,
SA. A. r. inholt, A. C. Bon ad ad H. 1. Schlesinger, ibid. 69,
6. R. F. Nystrom and W. G. Brown, ibid, 9., 1197 (1947).
7. H. C. Brown, H. I* Schlesinger and A. B. Burg, ib.. aj, 673 (1939).
8. H. C. Brown and B. C, Subba Rao, J. Org. Chem. ~., 1135 (1957),
9. D. E. Chaikin and W. G. Brown, J. Am. Chem. Soc. 7fl 122 (1949).
10. Metal Hydrides Inc., BIH4-NaBB4 Brochure.
11. H. C. Brown and B. C. Subba Rao, J. Am. Chem. Soc. 7., 2582 (1956).
12. Park and 0. Fuchs, J. Org. Chem. 21, 1513 (1956).
13. H. C. Brown, E, J. Mead and C,J. hosaf, J. Am. Chem. Soc. 7, 3616
14. J. Kollonitsch, 0. Fuchs and V. Gabor, Nnture 175f 346 (1955).
15. W. Huckel, Ann. 533, 1 (1937).
16. Dostrovsky, Hughes and Ingold, J. Chem. Soc., 173 (1946).
17. L. W. Trevoy and H. G. Brown, J. Am. Chem. Soc. ~j, 1675 (1949).
18. L. M. Jackann, A. K. lncbeth and J. A. Mills, J. Chem. Soc., 2641
19. D. S. Noyce and D.B. Denney, J. Am. Chem. Soc. 72, 5743 (1950).
20. C. W. Shoppee and G. H. R. Summers, J. Chem. Soc., 687 (1950).
21, R. NUce and 0. L, O'Conner, J. Am. Chem. Soc. IJ 5824 (1951).
22. D. J. Cram and F. A. A. Ehbafez, ,d. ,As 5828 (1952).
23. D. H. I. Barton, J. Chem. Soc., 1027 (1953).
24. J. B. Uoland and M. J. Jefraim, J. Am. Chem. Soc. l, 2798 (1956).
25. . .Dauben, S ec., Ad. 18, 3752 (1956).
26. W* G. Dauben, 0, J. rFoken and D, S. oycea, ib. Z& 2579 (1956).
27. 4 D. Hardy and R. J. Wicker, bj d. ., 640 (1958).
28. A. H. Beckett, pj jo Tetrahedron 6, 319 (1959).
29. BE L. Eliel and M. H. Rcrick, J. Am. Chem. Soc. 1367 (1960).
30. H. C. Brown, Wheeler and Ichikawa, Tetrahedron 1, 214 (1957).
31. L, P. Hmoaett, Phys. Org. CheMin 331.
32. S. Winstein and N. J. tolness, J. Am. Chem. Soc. 2, 5562 (1955).
33. Hi C. Brown, BE J. Mand and B. C. Subba Bao, ,b. 77 6209 (1955).
34. R. J. Wicker, J. Cham. SoC., 2166 (1956).
35 E. L. Eliel and R. S. So, 3J Am. Chem. Soc. 79, 5992 (1957).
36. H, X. Schlesinger and H. C. Brown, ibid. 2, 3429 (1940).
37, H. C. Brown, H. I. Schlesinger#, S 4,. 75, 192 (1953).
38. H. C. Brown and W. Korytnyk, ibid. 2. 3866 (1960).
39. E. R. Garrett and D. A. Lyttle, ib.d. 6051 (1953).
40. G. Stork and W. N. White, id. 1., 4617 (1956).
41, G. L. Cuniagha3n, U. S. Patent 98,085 (1936),
42. H. Steinberg and D. L. Hunter, ind. Eng. Chem. .j, 174 (1957).
43. H. 1. Schlesinger, H. C. Brown and B. K. Hyde, J. Am. Chem, Soc.
^,i 209 (1953).
Hugh Edward Wise' Jr. was born October::12 1930, at Lafayette,
Indiana. In June, 1948, he was graduated from South Broward High
School, Dania, Florida, Ii Jrie, 1952, he received the degree of
Bachelor of Arts from Vanderbilt University. From 1953 until 1956,
M,. Wise played professional baseball during the sprii and sumer
and attended the Uaiversity of Florida in the fall semester. Since
that time, he has been on campus the year around. After being admitted
to the Graduate School in 1954, he was variously employed on contract
research, as a graduate assistant, teaching assistant, etc., in the
Department of Chemistry until 1960.0 e wtas self-employed during the
final year of his work toward the decree of Doctor of Philosophy.
He is a member of Kappa Sigma social krateinity, Qama Signm
Epsilon honorary chemical fraternity and The American Chemical Society.
This dissertation was prepared under the direction of the
chairman of the candidate's supervisory committee and has been approved
by all members of that committee. It was submitted to the Dean of the
College of Arts and Sciences and to the Graduate Council, and was
approved as partial fulfillment of the requirements for the degree
Doctor of Phiiosophy.
June $S 1961
Dean, College of Arts and Scfences
Dean, Graduate School
Internet Distribution Consent Agreement
1. In reference to the following dissertation:
AUTHOR: Hugh E.Wise, Jr.
TITLE: The stereochemistry of the reduction of an alicyclic unhindered ketone with
some borohydrides and diborane. (record number: 424019)
PUBLICATION DATE: 1961
I, Hugh E. Wise, Jr., as copyright holder for the aforementioned dissertation, hereby grant
specific and limited archive and distribution rights to the Board of Trustees of the University of
Florida and its agents. I authorize the University of Florida to digitize and distribute the
dissertation described above for nonprofit, educational purposes via the Internet or successive
This is a non-exclusive grant of permissions for specific off-line and on-line uses for an
indefinite term. Off-line uses shall be limited to those specifically allowed by "Fair Use" as
prescribed by the terms of United States copyright legislation (cf, Title 17, U.S. Code) as well as
to the maintenance and preservation of a digital archive copy. Digitization allows the University
of Florida or its scanning vendor to generate image- and text-based versions as appropriate and to
provide and enhance access using search software.
This grant of permissions prohibits use of the digitized versions for commercial use or profit.
SignaturFof Copyright Holder
_Hugh E. Wise, Jr.
Printed or Typed Name of Copyright Holder/Licensee
Personal information blurred
Date of Signature