The development of aldehyde selective organoaluminum reagents for organic synthesis


Material Information

The development of aldehyde selective organoaluminum reagents for organic synthesis
Physical Description:
x, 149 leaves : ill. ; 29 cm.
Abedi, Vahak, 1961-
Publication Date:


bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1991.
Includes bibliographical references (leaves 145-148).
Statement of Responsibility:
by Vahak Abedi.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001709810
notis - AJC2095
oclc - 25541170
System ID:

Full Text







to my

parents and sister


At this time, I would like to thank my advisor and

mentor professor Merle Battiste for his graciousness and

generosity with time and knowledge on my behalf. I would

also like to thank my parents and sister, without the love

and support of whom this achievement would have been

impossible. As for Velva, though there might not have been

many words said, the gourmet meals, the friendship, and

knowing she was there was appreciated. Many thanks to all

the Perles, Curt, Jim, Johnny, Mapi and Radi for all the

help and company they have provided through out these

years. And lastly, this acknowledgment will not be

complete without thanking Sev for his unconditional
























GENERAL . .. .

. iii

* vii

. ix

. 41
. 50

. 62

. 77

. 80

. 80
. 82

* *

BIBILIOGRAPHY. . . .... .145

BIOGRAPHICAL SKETCH ............... ..149


1-1 pKa values of Me3Si-substituted phenols. 7

2-1 Reactions of R"R'2A1 (R" = Me, TMSCH2, 29
R' = Me, Et) with selected carbonyl

2-2 Reactions of R"2R'Al (R' = Me, TMSCH2, .... 31
R" = TMSCH2) with selected carbonyl compounds

2-3 Reactions of (TMSCH2)2Et2Al-Li+ with 39
selected carbonyl compounds.

3-1 Reactions of (TMSCH2)3A1 (3-la) with .... .54
selected aldehydes.

3-2 Reactions of (TMSCH2)3Al (3-la) with ... .56
selected ketones.

4-1 Reaction of aromatic aldehydes with ... 70

4-2 Reactions of B3 and B4 with aromatic ... .72

4-2a: Reactions with benzaldehyde
4-2b: Reactions with 4-Tolualdehyde

6-1 GC analysis of the MPV reduction of. ... .140
benzaldehyde via its reaction with the product
of the reaction of Me3Al with 1-phenylethanol.

6-2 GC analysis of the MPV reduction of. ... .141
benzaldehyde via its reaction with the product
of the reaction of Me3Al with l-phenyl-2-

6-3 GC analysis of the MPV reduction of. ... .143
benzaldehyde via its reaction with the product
of the reaction of Me2BHTAl.OEt2 with

6-4 GC analysis of the MPV reduction of. 144
benzaldehyde via its reaction with the product
of the reaction of Me3Al with 1-phenylethanol.


BHT-H 2,6-di-tert-butyl-4-methylphenol

BHT 2,6-di-tert-butyl-4-methylphenoxy

tBu tert-butyl

t-BuO tert-butoxy

Cat catechol

cond. condensation

DEDTA diethylbis(trimethylsilylmethyl)aluminate

DETA diethyltrimethylsilylmethylaluminum

DMBAE 2,6-di-tert-butyl-4-methylphenoxydimethyl-


DMBA 2,6-di-tert-butyl-4-methylphenoxydimethyl-


DMTA dimethylbis(trimethylsilylmethyl)aluminum

equiv. equivalent

Et ethyl

hrs hours

M molarity

Me methyl

MDTA methylbis(trimethylsilylmethyl)aluminum

MPV Meerwein-Pondorff-Verley

PDC pyridinium dichromate
















room temperature


staring material



N, N, N', N'-tetramethylethylenediamine




tris(trimethylsilylmethyl)aluminum, the

LiBr salt, [(TMSCH2)3A1.3LiBr]


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





Chairman: Merle A. Battiste
Major Department: Chemistry

In the present study aluminum has been used as the

metal center in the modification of the Peterson reagent

[(CH3)3SiCH2M, M = Li, MgBr] for the selective methylen-

ation of aldehydes. The modification was approached

through variation of the substituents on the organo-

aluminum. The effect of alkyl, alkoxy, and halide groups

as coligands on aluminum on the transfer of the (CH3)3SiCH2

(TMSCH2) group to the carbonyl carbon has been examined.

Alkoxy and halide substituents deactivate the aluminum

reagent to carbonyl addition, whereas alkyl groups (ethyl

and methyl) compete with TMSCH2 in addition reactions with

carbonyls. The deactivation of the organoaluminums by the

alkoxy and halide ligands is believed to be the result of

the strong bridging capacity and aggregating capability of


the heteroatoms in addition to their inductive electron

withdrawing effect. The competitive transferal of the

methyl and ethyl substituents is contrary to the anion

stability criteria established for the preferred delivery

of carbanions on aluminum. The competitive addition of the

simple alkyl nucleophiles can be rationalized by the

exceptional bulk of the Me3Si group which outweighs its

anion stabilizing effect. The substitution of three TMSCH2

groups on aluminum affords [(CH3)3SiCH2]3Al (TTMA) which

has emerged as an aldehyde-selective TMSCH2 transferal

agent. The superior selectivity of this reagent is

demonstrated by its relative ease of reaction with aldehyde

vs. ketones and, more convincingly, with keto-aldehydes

under moderate reaction conditions. In addition, the

previously reported methylation and oxidation reactions of

BHTMe2Al:OEt2 (BHT = 2,6-di-tert-butyl-4-methylphenoxy)

with aromatic aldehydes has been reinvestigated. The

results of this study support an alternative scheme wherein

trimethylaluminum, liberated in low steady state

concentration from the solution of BHTMe2Al:OEt2, is the

major methylating agent in these reactions.



The area of organoaluminum chemistry has been under a

constant growth and development since discovery of the

direct synthesis of trialkylaluminums and the demonstration

of their application to the polymerization of olefins by

Ziegler and coworkers.1 As the result of the unique

properties of the carbon-aluminum bond, such as the

tendency to form bridged complexes with heteroatoms and

other metals, and their various modes of reactions with

olefins and other nucleophiles, organoaluminums have

attracted considerable interest in both areas of industry

and laboratory.

Additionally, organoaluminum compounds, due to their

lower reactivity and high oxophilicity, have found a

variety of applications in organic synthesis as selective

nucleophilic reagents. These applications include

diastereo-selective opening of epoxides,2 selective

alkylation and reduction of aldehydes3 and ketones4 and

selective reduction of alkenes, alkynes,5 and carboxylic

acid derivatives.6

The main subject of this dissertation, the development

and modification of the novel organoaluminum reagents, is

concentrated on the selective reactions of trialkyl- and

mixed-trialkylaluminums with carbonyls. While many

characteristic properties of trialkylaluminums have been

explored,' their employment in organic syntheses, as

nucleophilic compounds has not been fully recognized.

The development of an aldehyde selective methylenating

reagent by appropriate modification of the Peterson

olefination reagent was a major goal of this investigation.

These modifications involved incorporating the TMSCH2

substrate on a less reactive metal center. While there are

a variety of metals suitable for this purpose, the previous

experience in this laboratory with selective aluminum

reagents8 and the inherent selective properties of

organoaluminum reagents, focused attention mainly on


One of the more attractive properties of mixed-organo-

aluminum compounds is their ability to deliver one of their

substituents, generally exclusively, to a variety of

electrophiles. For example, Rathke alanes9 are among a

variety of other examples where the more stable carbanion,

the acetate group, is transferred preferentially over the

alkyl substituents (figure 1-la). Other examples include

dialkylalkynyl- (1-lb),10 dialkylalkenyl- (l-lc),1 and

dialkylindanylaluminums.12 In all of these cases the more

stable anion is transferred preferentially. An additional


-550 C, DME

Et2 Al CCH13
Toluene, 250 C


Me3 SiHC-==CHA1Et 2

THF, -78 oC r.t.
65% Si










Figure 1-1

characteristic of trialkylaluminums which makes them

especially suitable for an aldehyde selective reagent is

their inherent aldehyde and ketone selectivity in the

presence of esters and other less reactive electrophiles.13

The selective alkylation of aldehydes in the presence of

esters, by manipulation of the reaction conditions, has

been successfully employed in several synthesis (l-lc).



Selective Olefination Reactions

The olefination reaction of carbonyl compounds is an

important strategic transformation in organic synthesis.

There are many reagents available for the olefination

reaction of carbonyl compounds. The Wittig reaction14 and

its modifications15 and the Peterson reagents16'17 are some

of the better known and more popular reagents employed for

this purpose (figure 1-2). These reagents are highly

effective olefinating reagents and react readily with a

2 Ph3P-CHCO2CH2CH (-
O' OH Benzene / reflux OH

(0 (C2 2 2 P CHPh (1-2b)
O =_ NaH, DME

Li p (1-2c)


Figure 1-2

variety of carbonyl compounds. However, there are numerous

occasions in organic synthesis where there is a need for

olefination of one carbonyl in presence of other reactive

electrophilic centers, which often introduce a laborious

series of protection and deprotection steps. An alternate

solution to this problem would be the use of, for example,

a methylenating agent which would discriminate between two

or more carbonyl centers to react selectively with only


The selective methylenating reagents known to this

point have differing applications and properties in their

reactions with carbonyl compounds (figure 1-3). Some

reagents are selective towards aldehydes (l-3a),18 while

H hexanes


.-H (1-3a)



CH2(A1Br )2
THF, reflux


i. 2


THF, -780C







Figure 1-3


others are only compatible with nonenolizable carbonyls

(1-3b)'19 and a few are useful for readily enolizable

carbonyls (1-3c20,d21).

The selectivity of these reagents is the result of one

or a combination of the following factors: a) higher

reactivity of one carbonyl compared to the other, b) the

variations in the steric requirements of the two, and c)

the lower reactivity of the reagent. As is the case with

most selective reagents, it is the lower reactivity of the

reagent that leads to the distinguishing of the two

differing carbonyl groups. Hence, a common practice in

development of a selective methylenating reagent is the

modification of the parent reagent by either altering or

replacing the metal center.

The Peterson Olefination Reagent

The Peterson reagent22 (TMSCH2M, M = Li, MgX) is a

convenient and versatile olefinating agent that reacts

readily with aldehydes, ketones and esters at -78C. The

Peterson reagent takes advantage of the silyl group in both

addition and elimination steps of its reaction with

carbonyls. The silyl group stabilizes the a-anion in the

organometallic reagent and it also facilitates the removal

of the hydroxy group in the addition product, via its well

known p-effect.23 The B-silyl group offers a strong

stabilizing effect through electron donation to an adjacent


Figure 1-4. The p-effect of the silyl group

electron deficient center. The mechanism for this electron

donation is believed to be through hyperconjugation24 (p-a

conjugation, figure 1-4); the high degree of polarization

of the carbon-silicon bond, due to lower electronegativity

of the silicon, ensures a high electron density on carbon,

and enhances its ability to stabilize an adjacent

electron-poor center. The a-anion stabilization by the

trialkylsilyl group is believed to be due to a weak

electron accepting capability of the silyl group; this

effect is detectable with p-trimethylsilyl substituted

Table 1-1. The pKa values for para-trimethylsilyl substituted
phenols and anilines.

Arylsilane pKa at 25 C

R H 10.85
SiMe3 10.64

+ R H 4.62
R H SMe3 4.36
\-- SiMe3 4.36


phenols as they are more acidic than their unsubstituted

analogues25 and 4-trimethylsilyl-substituted anilines are

less basic (table 1-1).

The elimination of the trimethylsilyl and the hydroxy

group is possible under both acidic and basic conditions.

Under acidic condition the elimination occurs through an E

antiperiplanar transition state, whereas the base cat-

alyzed elimination occurs in a syn manner20 (figure 1-7).

These different modes of elimination offer a convenient

route to the stereoselective synthesis of cis and trans

double bonds.

SiMe3 H
H. ~ Pr n

Prn OH
SBF.Et2O or H*

SiMe3 H
Me3Si- o- Si

H W Pr Hr H
Pr H Pr

n n n
Pr Pr Pr


Figure 1-5

Organoaluminum Compounds: Structure and Reactivity

Trialkylaluminum compounds are almost all colorless

pyrophoric liquids at room temperature that are generally

dimeric in solution. The proton NMR of trimethylaluminum

shows only one peak, due to the rapid exchange of the

bridging methyl groups; however, at -75* C this process is

sufficiently slow for separate resonances to be observed27

(figure 1-6). An increase in the size of the alkyl

Me,,, Me_ ,,, _e
2 MeAl A AlMe AlMe

Figure 1-6

substituent, especially the branching at the p-position,

hinders the formation of the dimers and leads to monomeric

species, e.g. tris-2,4,4-trimethylpentylaluminum28 and


The dimeric nature of the trialkylaluminums results in

scrambling of the mixed trialkylaluminums. Studies on

this subject have demonstrated the dimeric nature of the

mixed trialkylaluminums, with the bridging positions

occupied preferentially by the smaller alkyl groups. NMR

studies30 of a 4:1 mole ratio mixture of monomeric

tri-isobutylaluminum and dimeric trimethylaluminum show the

exchange of methyl and isobutyl groups to form

di-isobutylmethylaluminum, which exists as a methyl-bridged

dimer (figure 1-7).

1 Me ,, Bu
I BuBu

Figure 1-7

The reaction of trialkylaluminums with carbonyl

compounds exhibits three possible reaction types: 1) alkyl

addition, 2) p-hydride transfer (reduction), and 3) enol-

ization of the carbonyl compound when a-hydrogens are

present (figure 1-8). The p-hydride reduction and the

alkyl addition reactions become competitive when a

Et3A1 + R1(RCH2)CO

Et Et
Et Et Et
S-' Al0 C

R 1 CH2R

H3O0+ tH3O+


R, R, = alkyl, aryl, H
Figure 1-8

p-hydrogen is present on the aluminumalkyl substituent.

The yields of reduction are increased by branching at the p

position of the alkyl substituent, or by employing low

ratios of the organoaluminum to carbonyl compound31 (:1:1).

The enolization reaction is also affected by similar

factors and can become a major side reaction with readily

enolizable carbonyls. With the exception of Me3Al and

Ph3Al, these characteristics limit the alkylation ability

of most other trialkylaluminums.

The less common side reactions of trialkylaluminums

with aldehydes are the Meerwein-Pondorff-Verley (MPV)





0- C.

.R 0R"

4 fR


+ R R"






R ""** R"
_- AR H

N** R"

R, R', R" alkyl, aryl

Figure 1-9

reduction32 and the Oppenauer oxidation33 reactions

(figure 1-9). These reactions usually arise from the

presence of excess carbonyl compound, or the low reactivity

of the trialkylaluminum. In situations where an excess of

aldehyde is present, after the addition reaction is

complete, the MPV reduction and the Oppenauer oxidation

take over to yield the reduced aldehyde and the oxidized

sec-alcohol. The slower oxidation and reduction processes

also become competitive with the addition reaction when the

trialkylalane is sluggish in transferal of the alkyl


Another limitation of the trialkylaluminums is the

diminished reactivity of the alkylaluminum after the

transferring of the first alkyl group to a carbonyl

compound. Studies conducted by Mole and coworkers34 show

the formation of stable hemialkoxides after the reaction of

two moles of trimethylaluminum with one mole of ketone or

Me .Me Me
"Al Al"
RCHO + Me Al2 MeO Me

R -Me

Figure 1-10

aldehyde (figure 1-10). This observation then relates the

single alkyl transferal ability of trialkylaluminums to the

deactivating effect of oxygen and other heteroatoms on


The lack of nucleophilicity of dialkylaluminum

alkoxides is explained by the inductive electron

withdrawing effect of oxygen, and the formation of oxygen

bridged aggregates. These aggregates inhibit the access of

the carbonyl group to the aluminum thus hindering the

formation of the carbonyl aluminum complex which leads to

the transition state of the alkyl addition. Other

heteroatom substituents have a similar effect on organo-

aluminum compounds. Note that the bridging capability35 of

heteroatoms with aluminum is stronger and more pronounced

than that of alkyl substituents. An excellent example of

the more favorable bridging ability of heteroatoms is the

thermal redistribution of a equimolar mixture of tri-iso-

propoxyaluminum and trimethylaluminum. This method is

employed in preparation of dimethylaluminum isopropoxide36

which exists as a dimer with bridging alkoxy groups (figure

1-11). This enhanced bridging capability arises from the


M_ 0... ....Me
2 i-PrO3Al + 4 Me3Al -- 3 J.AI A.I
Me^ A0' ^Me


Figure 1-11

association of the Lewis acidic aluminum center with a lone

pair of electrons from the more electronegative atom, and

usually leads to the formation of higher aggregates of

alkoxy, aryloxy and amine substituted organoaluminums.37

Preparation of monomers of organoaluminum aryloxides

is also possible when bulky phenoxides are employed.38 One

of the more popular phenols employed is the 2,6-di-tert-

butyl-4-methylphenol (BHT-H). Substitution of BHT for two

of the methyls of trimethylaluminum leads to the formation

of the monomeric BHT2AIMe, which has found a variety of

applications in stereoselective alkylation of ketones and


The successful reaction of trialkylaluminum with

carbonyl compounds often requires more than one equivalent

of the aluminum reagent. Generally, in alkylations of

carbonyls 1.5 to 2 equivalents of the organoaluminum are

used; however, in cases where p-hydride reduction competes

with the addition, up to four equivalents of the reagent

have been employed.31 Based on these results and related

kinetic studies,40 two moles of trialkylaluminums are

believed to be involved in their addition reactions to

carbonyls. Two mechanisms have been proposed41 based on

this criteria (figure 1-12): 1) transfer of the alkyl

substituent through a six membered cyclic transition state,

where the dimeric trialkylaluminum is completed to the

carbonyl oxygen (l-12a); 2) delivery of the alkyl group

through a four membered transition state, by one of the two
trialkylaluminums completed to the carbonyl oxygen (1-12b).




R1, R2 alkyl, aryl, H
R alkyl


6 R
/ R

or R-,1,



Figure 1-12

The Approach

As previously stated, the main subject of the
investigations reported in this dissertation involve the
development of a selective aluminum-based Peterson reagent.
The fulfillment of this task required the preparation of a
variety of organoaluminum reagents containing the TMSCH2
substituent, together with the study of their reactions
with carbonyl compounds. The lack of reactivity and
competitive transferring of alkyl substituents (ethyl and

methyl) on aluminum led to a more detailed study of the

relative reactivity of mixed trialkylaluminums (Chapter 2).

From the investigations on mixed trialkylaluminums,

(TMSCH2)3Al emerged as the aldehyde-selective modified

Peterson reagent. The superior aldehyde-selectivity of

(TMSCH2)3Al is demonstrated by its reactions with

aldehydes, ketones, and keto-aldehydes (Chapter 3). A

tangent to this project was the study of the methylation

of aromatic aldehyde using BHT (2,6-di-tert-butyl-4-methyl-

phenoxide) substituted alkylaluminums (Me2BHTAl:OEt2),

which surfaced from the future outlook for development of

stereoselective methylenating reagents (Chapter 4).



The first step in developing a selective organo-

metallic reagent involves defining a suitable metal center

with its corresponding ligands. As our initial attempt in

modification of the Peterson olefination reagent titanium

was chosen as the selective metal center in consideration

of its oxophilic character. In this area of selective

organometallic chemistry, reagents of titanium have proven

extremely effective. Various titanium reagents have been

used in stereo- and enantio-selective alkylations of

aldehydes42 (figure 2-la) and chemoselective alkylation of

tertiary halides43 (2-1b). In particular, (i-PrO)3TiCH3

has shown excellent selectivity in methylating aldehydes in

presence of ketones. Based on this information, an

extrapolation from the highly selective methylating reagent

(i-PrO)3TiCH3, resulted in choosing (i-PrO)3TiCH2TMS, which

appeared to be a good starting point for a selective TMSCH2

transferal agent.

The requisite reagent [(i-PrO)3TiCH2TMS] was prepared

by reaction of (i-PrO)3TiCl (in hexanes) with TMSCH2Li (in


0H CH3
i-PrO TiMe (2-1a)
0 88% ds OH

87% (2-1b)

Figure 2-1

pentane) and was employed (in situ) immediately after

preparation (figure 2-2). Preliminary examination of the

reactivity of this reagent with phenylacetaldehyde proved

fruitless, as GC analysis showed unreacted aldehyde and

aldol condensation as major products of these reactions.

Subsequently, it was discovered that this modified Peterson

reagent had also been studied by Maycock, yielding a

similar outcome.44

-78o C to r.t.
1) (i-PrO)3TiCl + TMSCH2Li -- (i-PrO)3TiCH2TMS + LiCI

-78o C to r.t.
2) PhCH2CHO + (i-PrO)3TiCH2TMS -----. s.m. + Condensation

Figure 2-2

The next obvious choice of a metal center for our

investigation of an aldehyde selective organometallic

reagent was aluminum; this choice was mainly based on the

well known oxophilic behavior of aluminum and our

laboratories familiarity with organoaluminum reagents.

Rathke alane (R2AlCH2CO2t-Bu, R = Me, Et) and its

regioselective opening of allylic epoxides had been the

subject of continuing interest and detailed studies in our

group8 (Chapter 1).

As it was pointed out previously, dialkylalkynyl- and

dialkylalkenylaluminum reagents have also been employed as

selective alkynyl and alkenyl transferring agents in

addition to aldehydes,45 and in stereo- and regio-selective

opening of epoxides46 (Chapter 1). The above reagents are

typically prepared from the corresponding alkynyl or

vinyllithium compounds and the diethyl- or the dimethyl-

chloroalane. Exchange of lithium with aluminum reduces the

reactivity of the organometallic reagent thereby increasing

its selectivity. The high regio- and stereoselectivity of

the aluminum reagents is thus explained by the lower

nucleophilicity and higher oxophilicity of organoaluminum

reagents, compared to lithium or magnesium reagents..

Considering the studies above, replacement of the

acetate group in the Rathke alane by a TMSCH2 group would

appear to be a reasonable choice for development of a

suitable TMSCH2 group transferring agent (figure 2-3).


Figure 2-3

Based on the stability of the TMSCH2 anion, the preferred

transferal of this substituent over the ethyl or the methyl

would also be expected. The higher stability order of the

TMSCH2 anion compared to the analogous hydrocarbon anions

has been established by gas phase studies conducted by

Brauman;47 other evidence for this higher stability include

the preparation of TMSCH2Li by direct lithiation of

tetramethylsilane using n-BuLi and TMEDA.48

In spite of these considerations, the reaction of

Me2AlCH2TMS (DMTA) with phenylacetaldehyde offered very

small amounts of the TMSCH2 addition product and more of

another which was later identified as the methyl addition

product (figure 2-4). The use of Et2AlCH2TMS (DETA) adduct

did not effect the outcome of the reaction, and yielded the

ethyl addition as the major product.

Hexanes |
0 0^0 0 C to r.t. O
major product
R Et, Me
Figure 2-4

An alternate solution to the unexpected preferred

transferal of the methyl and the ethyl group was the

replacement of the alkyl groups on aluminum with other less

nucleophilic substituents, which would have a stronger

bonding to aluminum, e.g. halogens and phenoxy groups. The

lowered nucleophilicity of the organoaluminum reagent, due

to exchange of the alkyl groups with the more electro-

negative substituents was expected; however, the extent of

this effect was not predictable.

In investigating the reactivity of these reagents,

both AlC13 and AlBr3 were used to prepare X2AlCH2TMS and

XAl(CH2TMS)2. The preparation was performed by adding one

or two equivalents of TMSCH2Li to the hexane solution of

the corresponding aluminum halides (figure 2-5). The

spontaneous reaction of the aluminum halide and the alkyl

lithium was accompanied by precipitation of LiBr and

slight warming of the reaction solution.

ALX3 + 00 C, Hexanes
X C1, Br

0 OH
--------- TMS

R aryl, alkyl <10%

Figure 2-5

As in the cases above, reaction of these reagents with

selected carbonyls (cyclohexanone, heptanal, benzaldehyde)

proved unsatisfactory, yielding small amount (10% <) of the

desired product and aldol condensation or unreacted

starting material as the major products (figure 2-6).

The use of catechol (catechol = Cat) as a potential

substituent on aluminum was also investigated. Catechol

substrates of aluminum were prepared by reaction of

equimolar solutions of catechol (THF or CH2C12 solution)

and dimethyl chloroalane (hexanes solution), yielding the

evolution of methane and the production of CatAlC1 (figure

2-6). CatAlCl (2-6a) can be subsequently reacted with

TMSCH2Li producing CatAlCH2TMS (2-6b) and LiCl. The

reaction of this reagent with heptanal resulted in small

amounts of reduced aldehyde (up to 25%) and mostly

condensation products with no evidence of any TMSCH2

addition products.

-78 oC to r.t.
OH CH2C12 or THF
R Me, Et

A1C1 -78 ------C to r.t.
AC -78 C to r.t.
~ `0

0 A1C1 + 2 RH


o/ A1CH2TMS + LiC1


\ ?R'CHO
SA1CH2TMS R CHO Condensation + RCH OH
0/ -78 oC to r.t.
CH2C12 or THF

R' alkyl
Figure 2-6

The presence of the reduced aldehyde in this set of

reactions in the absence of any addition or oxidized

addition products, rules out the MPV reduction and the

Oppenauer oxidation processes. A closer inspection of the

GC/MS analysis shows condensation products of heptyl

heptanoate with heptanal, which is an evidence for the

Tishchenko reaction49 (figure 2-7). While the Tishchenko

reaction is similar to the Cannizzaro reaction in that the

aldehyde is both oxidized and reduced, the two differ in

the conditions and the outcome of their reactions. The

trialkoxy aluminum catalyzed Tishchenko reaction is not

limited to non-enolizable aldehydes and yields an ester as

the major product.


R alkyl, aryl

Figure 2-7

The exact structure of the catechol based aluminum

reagent is not known, since the high oxophilicity of

aluminum (Chapter 1) could cause the formation of dimers or

higher aggregates with the alkoxide substituents. The

formation of aggregates with both the catechol and halide

substituted aluminum reagents together with the inductive

electron withdrawing effect of the oxygen and the halogens

on aluminum are the most likely reasons for their observed

lack of reactivity.

The unsatisfactory results in these experiments with

various substituents on aluminum, directed at modifying the

Peterson olefination reagent, led to a more detailed study

of the mixed trialkylaluminums and their reactions with

carbonyls. The purpose of this study was to obtain a

better understanding of the reactivity of the mixed

trialkylaluminums, and perhaps be able to determine the

criteria and the factors which govern the preferred

transferal of the substituents on aluminum. The factors to

be considered are a) statistical factors, i.e., the ratio

of ethyl and methyl to TMSCH2, b) steric reasons, i.e., the

bulk of TMSCH2, c) electronic reasons related to the

stability of the different alkyl anions or d) perhaps a

combination of all or some of the factors above.

The reactions of five aluminum reagents with selected

carbonyl compounds were used for this study. The carbonyl

compounds included three aldehydes (heptanal, benzaldehyde,

phenylacetaldehyde) and two ketones (cyclohexanone and

fluorenone). The aluminum reagents employed were diethyl-

trimethylsilylmethylaluminum (Et2AlCH2TMS, DETA), diethyl-

bis(trimethylsilylmethyl)aluminate [Li(Et2Al(CH2TMS) 2],

DEDTA), dimethyltrimethylsilylmethylaluminum (Me2AlCH2TMS,

DMTA), Lithium methylbis(trimethylsilylmethyl)aluminate

[Li+(MeAl(CH2TMS) 2-], MDTA), trimethylaluminum (Me3Al,

TMA), and tris(trimethylsilylmethyl)aluminum (Al(CH2TMS)3,

TTMA). Each reagent was prepared (with the exception of

Me3Al) by adding the required amount of TMSCH2Li (1.0 M in

pentane) to the corresponding hexanes solution of the

dialkyl chloroalane, alkyl dichloroalane or aluminum

trihalide (figure 2-8). Once prepared the reagents were

used immediately. The product distribution of these

reactions was determined by GC analysis using an internal

standard (tridecane), the yields were calculated as %

conversions and are listed in tables 2-1, 2-2, and 2-3.

-78 C to r.t.
R2A1C1 + TMSCH2Li 78 to r R2AICH2TMS + LiC1

R2A1C1 + 2 TMSCH2 Li [R2A1(CH2TMS)2]-Li+ + LiCI

RAIC12 + 2 TMSCH2Li RA1(CH2TMS)2 + 2 LiCI

RAIC12 + 3 TMSCH2Li -- [RAI(CH2TMS)3]-Li+ + 2 LiCI

00 C to r.t.
A1C13 + 3 TMSCH2Li CH C A1(CH2TMS)3 + 3 LiCI

R Me, Et

Figure 2-8

Because of the facile alkyl exchange capability of

mixed trialkyl aluminum, it is not possible to determine

the exact structure of the alkylating agent. As was

discussed in Chapter 1, mixed trialkyl aluminum exist

predominantly in a dimer structure where the unbranched

smaller alkyl groups occupy the bridging positions. Thus,

in the cases of DMTA and DETA the most likely structures

for the equilibrated mixed alanes are depicted in the

following equations (figure 2-9).

Both DMTA and DETA can exist as dimers of the original

monomers (2-9a) or could equilibrate to produce a new dimer

2-9c and R3Al as illustrated in equations 2-9b and 2-9c.

If the equilibration 2-9b dominates, then the majority of

the ethyl and methyl addition products would be the result

of R3Al alkylation. It is possible for R3Al to be the

major alkylating agent even if its concentration is low in

the solution. The facile reaction of R3Al with carbonyls

could lead to steady production of this aluminum reagent by

shifting the equilibrium 2-9b to the right. The slow

production of (TMSCH2)3Al and its inability for association

into a more stable dimer50 however, is expected and was

shown to disfavor the equilibrium suggested by equation

2-9d (Chapter 4).

.": ... A R ..N
.RAl R AlS (2-9a)
-fR R % 'CIL2TMS

2 R2A1CH2TMS k--

-1 RA1(CH2TMS)2 + R3A1 (2-9b)

2"*...A1 Al (2-9e)
00 R *. (2CH TMS
2 RAI(CH2TMS)2 k2

2 A1(CH2TMS)3 + R2A1CH2TMS (2-9d)

R Me, Et

Figure 2-9

The structure for MDTA can be compared to that of

(i-Bu)2MeAl where its NMR studies have concluded that this

species exists in a dimeric form (2-9c) with the majority

of the methyl groups occupying the bridge positions30

(Chapter 1).

The role or effect of the lithium halide salts present

in the reaction mixtures, due to in situ preparation of the

aluminum reagents, is not well known, and will not be

elaborated on in this chapter.

Reactions of TMA. DMTA. DETA. and MDTA with
Aldehydes and Ketones

The simplest of all trialkyl aluminum, Me3Al, though

an effective alkylating agent for aldehydes and ketones,

has not found widespread use in organic synthesis (Chapter

1). The selectivity of TMA allows its use in presence of

lactones and esters under controlled conditions.3 This

carbonyl selectivity of TMA combined with its impressive

yields of addition products(table 2-1), illustrate its

potential for employment in organic synthesis. The best

yields for reactions of Me3A1 with carbonyl compounds are

obtained in hydrocarbon solvents at 0 OC, where the

reactions are usually complete within ten minutes.

An overview of the reactions of DMTA, the next

trialkylaluminum in table 2-1, with all of the carbonyls

shows a larger than statistical (2:1) ratio of methyl to

TMSCH2 addition. These ratios hints to the presence of

other factors that are governing the preferential

transferal of one group over the other.

The Me:TMSCH2 ratio of addition is similar in all of

the DMTA examples except for fluorenone. Fluorenone is

both less reactive (conjugation of the carbonyl group with

the two aromatic rings) and sterically more demanding

towards nucleophilic additions. The steric limitations

caused by the two ring hydrogens at 1 and 8 positions of

fluorenone particularly reduce the access of large nucleo-

philes, such as the TMSCH2 anion. This lower reactivity is

clearly evident in all reactions of fluorenone with TMSCH2

containing alanes, notably so in the reactions with DMTA

and TTMA.

Cyclohexanone, the other ketone studied, is more

reactive and sterically less demanding than fluorenone,

however, because of the presence of its a-hydrogens, it can

be enolized. The less reactive and highly Lewis acidic

aluminum reagents in this study facilitate the enol-

ization, which is responsible for significant amounts of

unreacted starting material and aldol condensation

products. The results from the reaction of cyclohexanone

with DMTA show lower yields than that of the aldehydes;

nevertheless, the Me:TMSCH2 addition ratio (8.2) is

comparable if not higher than that of the aldehydes. The

same argument holds true for the reaction of cyclohexanone

with DETA. Here once again, the overall yield is much

lower than others, but the Et:TMSCH2 addition ratio (3.2)


TABLE 2-1. Reaction of R"R'2A1 (" Me, TMSCH2, R' Me, Et) with

selected carbonyl compounds.

R"R' Al


+ R2-7.

R" R' PR R' RI "
Me Me 99b
O CHO ------- ------------------ ------
TMSC Me 71 9 ( l^ J.---p --- .--- -- -----. ----. ------
TMSC Et 85 <1 142 7[12]

-- --- --- ---
STMSC 'Me 43 7.3 5.9 -- 6SM, 2C

TMSC Et 38 1 12 13[3] 3SM, 2C

Me Me 87b -- -- -- 3SM,3C

CHO TMSC Me 59 9 6.7 -- 2SM

TMSC Et 58 2 29 6[10]

SMe M e 80b -- -- -- 20SM

S TMSC I Me 45 6 8.2 -- 13C

TMSC Et 26 8 3.2 -- 5C

O Me 'Me 97b -- -- 4SM

S TMSC Me 73 3 24 -- 8SM

TMSC a Et 73 <1 146 15[5]

TMSC TMSCH2, S.M. starting material, C condensation products.
a: % methyl ketone, b: uncorrected GC results. RED. reduction
[ ] R / RED.. All yields calculated as percent conversion, based
on GC analysis using tridecane as an internal standard.


is surprisingly smaller than that for the aldehydes. The

reaction of cyclohexanone with MDTA, holds yet another

surprise; the overall yield is low as usual, but the

Me:TMSCH2 addition ratio is identical to benzaldehyde

(2.8). In all of the above cases the lower overall yields

from cyclohexanone can be attributed to the enolization

problem, and the similar Me:TMSCH2 addition ratios of

cyclohexanone and benzaldehyde indicate a similar steric

environment. Since cyclohexanone is the only sterically

less hindered ketone studied it would be difficult to

arrive at or propose any other conclusions about its

similar reactivity to that of benzaldehyde.

The lower reactivity of benzaldehyde compared to the

other aldehydes observed in this set of reactions is

probably due to the conjugation of the carbonyl to the

aromatic ring which decrease the electrophilicity of the

carbonyl group. The product ratios from the reaction of

benzaldehyde with DMTA is similar to that of the other

aldehydes, with a slightly larger Me:TMSCH2 addition ratio

(7.9). A slight difference in the outcome of the reaction

of benzaldehyde with DMTA vs. other aldehydes is the result

of small amounts of oxidation and reduction products which

are undetectable in other cases. The oxidized and reduced

side products are due to the Meerwein-Pondorff-Verley (MPV)

reduction of the starting aldehyde and Oppenauer oxidation

of the alcohol product.

TABLE 2-2. Reaction of R"2R'A1 (R" TMSCH2, R' TMSCH2, Me) with
selected carbonyl compounds.


R2R 2




+ -R


CARBONYL R2-R'A R271', R2 ). /R" Red. OTHER
R" R R' R' E R

MCHO TMSC Me 54 19 2.8 6

TMSC TMSC -- 45 (19)a -- 33 6C

H TMSC Me 19 24 0.79 2 9SM,1C
-- - -
TMSC 'TMSC -- 14--
I 17C

TMSC Me 18 29 (9)a 0.62 13 6C
-'---CHO ----------------------- -----
TMSC TMSC -- 77 -- <1 8C

0 '
TMSC Me 29 10 2.8 2SM
TMSC TMSC -- 12 -- -- 8C

O TMSC Me 42 3 13 -- 56SM

T C T H C - --- 4--- -- -
TMSC TMSC -- 4 -- -- 86SM

RED. reduction.

SM starting material, C condensation products.
a: % methyl ketone, also included in the total yield.

All yields were calculated as % conversion, based on GC analysis, using
tridecane as an internal standard.

Reaction of benzaldehyde with DETA (table 2-1)yields a

majority of ethyl addition with very small amounts of

TMSCH2 addition; the ratio of the two products being

noticeably larger (142) than the other carbonyl compounds,

except for that of fluorenone. The significant amount of

reduced aldehyde benzyll alcohol, 7%) in this reaction

which is greater than the TMSCH2 addition product (1%<)

results from p-hydride reduction. The reaction of MDTA

with benzaldehyde (table 2-2) gives a Me:TMSCH2 addition

ratio of 2.8 which is unlike the product ratios of other

aldehydes and is identical to cyclohexanone.

In general, reactions of phenylacetaldehyde with

trialkylaluminums and even alkyllithiums give lower yields,

with unreacted aldehyde and aldol condensation products as

the majority of the side products. This problem arises

from the acidic hydrogen of phenylacetaldehyde, which is

alpha to both the aromatic ring and the carbonyl group, and

facilitates the enolization of the aldehyde. The reaction

of phenylacetaldehyde with DMTA and DETA (table 2-1) is

typical of the reactions of the other aldehydes, with

similar Me:TMSCH2 and Et:TMSCH2 addition ratios of 5.9 and

12 respectively. The Me:TMSCH2 addition ratio from the

reaction of phenylacetaldehyde with MDTA (table 2-2) (0.79)

comes close to the statistical ratio of 0.5, a sign of

higher reactivity of the aldehyde and thus, lower

selectivity of the reagent with this carbonyl compound.

Heptanal is as reactive as phenylacetaldehyde with a

lesser tendency for enolization, and its reactions with

DMTA and DETA (table 2-1) result in outcomes similar to

that of phenylacetaldehyde. Heptanal also affords a close

to a statistical ratio of Me:TMSCH2 addition ratio (0.69)

in its reaction with MDTA (table (2-2), indicative of its

higher reactivity, and thus lower selectivity.

Reactions of TTMA with Aldehydes and Ketones

The results listed in table 2-2 for the reaction of

TTMA with carbonyl compounds are not an accurate measure of

this reagent's reactivity or usefulness. However, these

results expose some of the problems of the aluminum

reagents, and give an insight into their reactivities.

The basis for the above comments is the variety of ways

that TTMA can be prepared and used. For this set of

reactions the reagent was prepared (in situ) by adding

three equivalents of TMSCH2Li to one equivalent of

anhydrous 99.99% ALC13 (methylene chloride solution) with

stirring at room temperature for 2.5 hours.51 Since other

similar preparations call for extended reflux times, it is

not clear if in the preparation above, the formation of

TTMA is complete before its use (details on other methods

for preparation and use of this reagent will be discussed

in Chapter 3).

Reaction of benzaldehyde with TTMA shows three major

products, TMSCH2 addition (26%), benzyl alcohol (33%) and

acetophenone (19%). The benzyl alcohol is the product of

the MPV reduction of the aldehyde, which is concurrent with

the Oppenauer oxidation of the R-hydroxysilane, producing

the p-ketosilane. Acetophenone is believed to be the

result of the loss of TMS group from the p-ketosilane,

which is sensitive to enolization and desilylation. The

desilylation could occur either during the reaction through

enolization of the ketone, or during initial steps of the

work up from the direct desilylation of the p-ketosilane.

The extensive oxidation and reduction in these

reactions seem to be partially due to the lower reactivity

of the bulkier aluminum reagent. The slow addition of the

reagent combined with unreacted aldehyde gives the ideal

conditions for this oxidative and reductive processes

(Chapter 1).

The low reactivity of TTMA seems to lead to a larger

degree of enolization in its reaction with phenylacet-

aldehyde, and hence to larger amounts of aldol condensation

and recovered starting material.

Heptanal gave the best TMSCH2 addition results, with a

very small amount of reduction, in its reaction with TTMA.

This higher yield of addition product is most likely due to

the higher reactivity and lower enolization tendency of


In reactions of TTMA with cyclohexanone, poor yields

of addition and large amounts of aldol condensation

products were noted together with recovered starting

material. The results from this reaction were predictable

considering the lower reactivity and ease of enolization of

cyclohexanone compared to other carbonyls.

The effect of the bulk of the TTMA reagent on its

reactivity becomes more apparent in its reaction with

fluorenone, where after similar reaction time (4 hours) as

other carbonyls, only 4% of addition product is obtained.

Results and Discussion

A better insight to the reactivity and nature of these

reagents (DETA, DMTA,and MDTA) is obtained by an overall

look at the experimental results. In comparing the product

ratios from the reactions of DETA and DMTA one sees: first

a larger ratio of simple alkyl addition than TMSCH2

addition with DETA, and second, a comparable amounts of

reduction with DETA vs. TMSCH2 addition with DMTA. The

difference in reactivity of these two reagents could be due

to the bulk of the reagent and/or the variety of the path

ways for their reactions. For DMTA there are two paths of

reaction, either transferring of the methyl or the TMSCH2

group. In the case of DETA there are three paths of

reaction a) ethyl addition, b) TMSCH2 addition, and c)

p-Hydride reduction. The transferal of the methyl group

from DMTA seems to be much more facile than that of the

TMSCH2 group, which also holds true for DETA, however with

DETA there is an additional reaction pathway the

p-hydride reduction, that could compete with the transferal

of the TMSCH2 group. The p-hydride reduction indeed

competes with the TMSCH2 transferal as is evident by the

greater amount of reduction vs. addition observed in the

reactions of DETA with carbonyls.

Reactions of MDTA with carbonyls shows a significantly

larger amount of TMSCH2 addition than any of the previous

reactions. However, only in cases of heptanal and phenyl-

acetaldehyde does the Me:TMSCH2 ratio (0.62 and 0.79) come

close to the statistical one (0.5). There are also larger

amounts of oxidation and reduction products that could be

explained by the larger size and hence more sluggish

reactivity of MDTA (as discussed for TTMA); the effect of

the p-silyl group on the oxidation rate of the alcohol is

also an important factor which will be discussed in Chapter

3. This MPV reduction of the starting material is

possible only with aldehydes and is detected in all of the

three examples. However, the Oppenauer oxidation product

which should accompany the reduction is not detectable by

GC analysis of all of the product mixtures. The absence of

the oxidized products from GC chromatogram could be due to

three factors; first, the retention time of the alcohol and

its oxidized adduct, the ketone, are very close in most

cases so the identification of the two is difficult,

second, the ketone adducts could readily undergo multiple

aldol condensations and not be detected on GC, and lastly,

the possibility of the Tishchenko reaction exists which

could lead to esters and their subsequent crossed-Claisen

condensations, all of which would also be undetectable by


The explanation for the preferred or faster addition

of the simple alkyl groups vs. TMSCH2 group could be

steric, electronic, or a combination of both.

Electronically, according to the literature and the

research done in our group; the more stable carbanion on

aluminum is transferred more readily, and in most cases

exclusively, to the electrophile. In these studies the

TMSCH2 anion is the more stable carbanion when compared to

the methyl or ethyl anions, and thus should be the

preferred transferal group.

Sterically, larger more bulkier substrates on

aluminum would inhibit or slow down the complexation of the

aluminum and the carbonyl oxygen. These types of complexes

are believed to be the first intermediates in the formation

of addition products and are assumed to be on the pathway

leading to the transition state for successful transferring

of the alkyl groups (Chapter 1).

Preferential addition of the methyl and ethyl groups

in our studies of the mixed trialkylaluminums do not

support the correlation between carbanion stability and

preferential transferring of that group. This contra-

diction, nevertheless, can be explained by the size of the

TMSCH2 group. In this instance the steric factors seem to

outweigh the electronic ones. Even though TMSCH2 is the

more stable anion, the bulk of the group and the steric

requirements of the transition state inhibit its

preferential delivery. Even when tris(trimethylsilyl-

methyl)methylaluminate [(TMSCH2)3MeAl-Li+], prepared from

the reaction of 1:3 ratio of methylaluminum dichloride to

TMSCH2Li, was allowed to react with equimolar amounts of

benzaldehyde, only 42% of TMSCH2 addition compared to 46%

of methyl addition was observed (figure 2-10). Even though

hexanes IF
[(TMSCH2)3MeAL-Li+] + PhCHO 0 C r. PhCHCH3 + PhCHCH TMS
46% 42%

Figure 2-10

the aluminum species in this example is an anionic one (the

ate complex), and it does not possess the oxophilic

character of the aluminum reagents discussed previously,

nevertheless, it shows the greater facility of methyl

addition vs. the TMSCH2 addition.

In the reactions of [(TMSCH2)2Et2Al-Li+, (BTDEA)]

with carbonyls, the preferential delivery of the two

substituents can be compared without the interference of

the statistical considerations (table 2-3). In all three

examples of the reactions of BTDEA with carbonyls, ethyl

addition was the major pathway observed, and due to the

TABLE 2-3. Reaction of (TMSCH2)2Et2A1-Li* with selected carbonyl


R- CH3


TMSC TMSCH2, S.M. starting material, yields
as % conversion, and based on GC analysis using
internal standard.

were calculated
tridecane as an

lower oxophilicity of this aluminum species, no p-hydride

reductions of the carbonyls were observed. Based on the

reactions of tetraalkylaluminates discussed above, the

alkylating species is most likely the tetracoordinated

aluminum rather than the dissociated alkyl anion (figure

2-11). The most favorable dissociation of the aluminate

should be due to the loss of the larger and more stable

TMSCH2 carbanion to give a less crowded aluminum center and

a more stable dissociated carbanion. However, if this were

true, then the major product should have resulted from the


( 96 < 1 193

S58 2 31 12

1'0CHO 63 9 7 -

[Et2A(CH2TMS),]-Li+ -- W-- Et2A1CH2TMS + TMSCH2Li

Figure 2-11

TMSCH2 addition, not ethyl addition as was observed in all

of the examples. Information available in literature

related to our investigations is the study on the stability

of tetraalkylaluminates which demonstrates that lithium

tetramethylaluminate undergoes exchange reactions less

readily than trimethylaluminum.52 This study then, is also

in support of the mechanism of alkyl addition to carbonyls

via the associated tetraalkylaluminate.

The most important outcome of the reactions of BTDEA

with carbonyl compounds, is its evidence for much faster

addition of the ethyl group when statistically there is an

equal probability of addition by both substituents.

Furthermore, these results illustrate that, even with

possibly two different mechanisms of addition (ate species

vs. the uncharged alane), the smaller alkyl group transfers

much more readily than the larger TMSCH2 group.

A conclusion relevant to the development of a select-

ive Peterson methylenating reagent is the relationship

between the reactivity of the trialkylaluminum and its

size. This correlation is especially noticeable in

reactions of fluorenone, where as the bulk of the

organoaluminum increases (TMA < DMTA a DETA < MDTA < TTMA),

the reactivity and the alkyl addition yield decreases.



Preparation of (TMSCH213A1

The most relevant ramification from the study of the

mixed trialkylaluminums, for our pursuit of a selective

Peterson reagent, was the immergence of (TMSCH2)3Al as an

aldehyde-selective alkylating agent. Preliminary

investigations of the reactivity of TTMA showed close to a

three fold selectivity in the alkylation of benzaldehyde in

presence of cyclohexanone. This encouraging result led to

further investigation of both the reactivity and various

preparative methods of TTMA.

The synthesis and isolation of TTMA offered one of the

more challenging tasks of this research project, where even

the various methods of its preparation offered no help in

carrying on this task. In order to obtain enough TTMA for

a complete study of its reactions with carbonyl compounds,

all but one of the preparation methods had to be

investigated. As mentioned in chapter 2, there were three

known methods for preparing TTMA that employed different

organometallic reagents, and a proposed new method which

utilized trialkylborane and trialkylaluminum exchange

process (figure 3-1).

The first reported synthesis required a tedious

fourteen day reflux period of a mixture of Hg(CH2TMS)2,

aluminum foil and toluene (3-la).53 In this report

(TMSCH2)3A1 was characterized as a colorless pyrophoric

liquid (b.p. 51 OC at 0.08 mm Hg), which appears to give a

mixture of monomeric and dimeric species in benzene. The

proton NMR of TTMA in toluene showed two single peaks at

0.34 ppm (Me3Si-) and -0.22 ppm (-CH2-) with a 4.5:1 ratio.

110 OC
3(Me3SiCH2)2Hg + 2 Al/Hg 1 2 (MeSiCH2)3Al + 3 Hg (3-la)
14 days

hexanes, reflux
3 TMSCH2Li + AlBr3 hexanes, (TMSCH2)3A1 + 3 LiBr (3-1b)
12 hrs

3 TMSCH2Li + AICI3 r.t., 45 min. (TMSCH2)3A1 + 3 LiCI (3-1c)

figure 3-1

The newer and more convenient synthesis (3-1b)

involved a 12 hour reflux of a 3:1 mixture of TMSCH2Li (1.0

M, pentane solution) and aluminum bromide (suspended in

hexane).54 Another similar reported method (3-ic)51

required only a thirty minute stirring of anhydrous AlC13

and TMSCH2Li (three equivalents) in 1,2-dichloroethane at

room temperature. The TTMA reagent mixture was then used

in situ for chemoselective synthesis of allyltrimethyl-

silanes via the coupling of TTMA with vinyl triflates

(figure 3-2).


O | 1.4 eq.TTMA, Pd(0) SiMe3


figure 3-2

In order to obtain the pure TTMA in any of these

synthesis, separation of the resulting metal reagent from

the metal salts by vacuum (<.01 mm of Hg) distillation was


For our preliminary studies, the synthesis of TTMA was

carried out using the AlBr3 synthesis which was simple and

straight forward, however, the distillation and isolation

of the TTMA proved to be extremely difficult and gave low

yields. In the literature, the conditions for the

preliminary distillation (separation of liquid TTMA from

LiBr salt) did not include temperature of the heating bath,

the exact pressure (ca. high vacuum), or the amount of time

required for the complete distillation. During our

distillation attempts, even after six to seven hours of

heating at bath temperatures of 80-90C and a pressure of

0.01 mm (at best), lower than 45% isolated yields were

obtained (85% reported yield). Heating the distillation

flask over 1400C resulted in the extensive decomposition of

the solid mixture, which was apparent by its yellow

discoloration. Initially, the non availability of a high

vacuum source was thought to be the major cause of the

difficulties and the low yields encountered. Further

attempts using this synthetic and distillation procedure,

even at lower pressures, suggested that other factors such

as the purity of the starting materials were also affecting

the outcome of the reaction.

Unsatisfied with the above synthesis, we examined

other possible routes for preparation of TTMA. A less

important, but novel preparation of trialkylaluminums

involves the facile exchange of alkyl groups between boron

and aluminum.s5 The preparation depends upon the

progressive displacement of exchange equilibrium by

distillation of the most volatile trialkylborane (figure

3-3). The recent availability of (TMSCH2)3B from Aldrich

Me3Al + (TMSCH2)3B ) (TMSCH2)3A1 + Me3B (3-3a)

Figure 3-3

chemical company made this procedure even a more attractive

one. The exchange of the TMSCH2 groups on boron

[(TMSCH2)3B] with the methyl groups on aluminum (Me3A1,

b.p. 1300 C) would produce Me3B (b.p. -220 C), which should

easily distill from the reaction mixture, driving the

equilibrium to the aluminum side. According to the

available literature on the transalkylations between

trialkyl borons and aluminums,8 high reflux temperatures

which exceeded the b.p. of the solvent (1400C) were needed.

This information translated to the need for the neat

reagents, which posed the experimental problems of

obtaining and handling the highly pyrophoric neat TMA. To

avoid this difficulty, the 1.0 M hexane solution of Me3Al

was employed, and after addition of the neat (TMSCH2)3B to

this solution, the solvent was distilled off leaving a neat

mixture of the aluminum and the boron reagents. After the

mixture was refluxed (ca. 1100C) for 12 hours, NMR analysis

of the reaction mixture showed what appeared to be a

mixture of (TMSCH2)3B and Me(TMSCH2)2A1. The attempts to

separate the boron and the aluminum species through vacuum

distillation, also proved unsuccessful. The difficulty in

replacing the last methyl group on aluminum, verified by

the outcome of this experiment, is one more bit of evidence

for the kinetically slow conversion of MDTA to TTMA, as

pointed out previously (Chapter 2).

The AlC13 version of TTMA's synthesis was also

investigated; the in situ synthesis required the addition

of a 3:1 ratio of TMSCH2Li solution to AlCl3 (anhydrous

99.9%, in CH2C12), and stirring of the mixture for 2.5

hours before addition of the carbonyl compounds. As

mentioned previously, the TTMA used for the reactions

discussed in Chapter 2 was prepared using the above

procedure. However, the results of these reactions did not

seem to be the optimal outcomes expected from reactions of

TTMA. Also, it was not clear if the formation of TTMA

under these reaction conditions was complete or not. In

hopes of obtaining pure TTMA using this procedure, the

reaction was repeated on a larger scale, and was allowed to

stir for 22 hours at room temperature. As before, the

attempted distillations of this reaction mixture were most

unsatisfactory. Also, the reaction of the distillate (the

solid reaction mixture), after the distillation attempt,

with benzaldehyde afforded no TMSCH2 addition products.

At that time, the only hope for obtaining reasonable

amounts of neat TTMA, was to improve the yields of the

AlBrs procedure. The literature procedure for the

preparation of TTMA required freshly sublimed AlBr3 and

TMSCH2Li, which was not performed in our previous attempts.

Aluminum bromide was originally purchased from Alpha

(99.9%, white crystalline powder), and afforded reasonable

yields of distilled TTMA (up to 45%). However, the

subsequent purchase of AlBr3 from Aldrich (98.8%, large

yellow and brown crystals), and reactions from this sample

gave very low or no yields of neat TTMA. One of the

concerns in using the unpurified AlBr3, was the presence of

aluminum oxides that could cause the formation of large

aluminum aggregates and hinder the successful distillation

of the TTMA. Hence, the purification of the starting

materials seemed to be the best solution for improving the

yields of neat liquid TTMA.

The sublimation of the AlBr3 was performed in a large

sublimation apparatus at low pressures (0.25 mm of Hg), in

a 900C oil bath using a cold finger cooled by a dry ice

acetone bath. The sublimation yielded a fine white powder

which produced a white smoke during its weighing in the dry

box. However, we were unable to purify the TMSCH2Li, since

it required a vacuum of 10-5 mm that was not available to

us. Thus, the 1.0 M pentane solutions of TMSCH2Li,

purchased from Aldrich, had to be employed.

The sublimation of the AlBr3 immediately before use

improved the outcome of the distillation dramatically. The

distillation of the TTMA from the LiBr salt started

smoothly at a bath temperature of 700C and a pressure of

0.01 mm, resulting in 56% yield of neat TTMA; all of the

neat TTMA was subsequently smoothly redistilled at 45-48 C

(0.02 mm of Hg). Even though this yield was higher than

the previous distillation attempts had afforded, it was

still well below the literature yield of 85%. The 'H NMR

spectra of TTMA in benzene-d6 showed two singlets, one at

0.18 ppm (9 H) and other at -0.38 ppm (2 H), belonging to

the nine methyl and the three methylene groups,

respectively. The 13C NMR spectra also showed two

singlets; the larger (Me3Si-) at 2.89 ppm and the smaller

(-CH2-) at 5.15 ppm. This improvement in the reaction

yield suggests that starting materials of high purity are

essential for acceptable conversion yields, and perhaps a

better quality of TMSCH2Li could improve the yields even


In the early stages of this research the difficulty in

separation of the TTMA from the lithium salt through

distillation led us to attempt the separation by filtration

of the reaction mixture. The filtration was carried out

using the cannula filtration method, and proved to be a

slow and time consuming process.

The filtration of the liquid from the mixture removed

the yellow coloration of the solid, but the 1H NMR of the

concentrated filtrate, and also its reaction with

benzaldehyde, verified that the filtrate contained only

very small amounts of TTMA. This suggested that almost all

of the reagent was tied up completedd) with the lithium

salt. The complete removal of the solvent from the solid

under vacuum produced a white solid (TTMAs) which could be

powdered and stored in the dry box for an indefinite

periods of time (figure 3-4). This powder also proved to

be a potent TMSCH2 transferal agent in its reactions with

1) reflux,
12 hrs
3 TMSCH2 + ABr3 (TSCH2)3A1.3LiBr
2) removal of
solvent White powder

Figure 3-4

carbonyl compounds, further consequences of which will be

discussed in the reactions of TTMA.

The formation of such strong complexes of LiBr with

trialkylaluminums is not documented in the literature.

Furthermore, both lithium and bromide ions are believed to

have a lesser tendency and ease of complex formation with

trialkylaluminums.56 Thus, it is possible that three

equivalents of LiBr are needed for the formation of the

white powder we had isolated. Preliminary studies on this

hypothesis was performed by successive addition and then

reflux of 1, 2, and 3 equivalents of anhydrous LiBr to 1

equivalent of the neat liquid TTMA (solution in hexane) and

analyses of the mixture after each addition. Evaporation

of solvent after reflux with 1 and 2 equivalents of LiBr

gave a mixture of oil and powder; however, after the third

equivalent of LiBr was added a pasty mixture was obtained.

While, the results from this preliminary study are not

conclusive, it appears that three equivalents of LiBr are

needed for the formation of the solid complex .

The LiBr complex of TTMA (TTMA,) is mostly soluble in

THF and is accompanied by a slight warming of the resulting

solution. However, the proton NMR of TTMA, in THF-d4 shows

a large multiple in the vicinity of 0.0 ppm which is

difficult to interpret. This broadening of peaks is

presumed to be the result of the complexation of THF with

aluminum. Attempts at determining the stoichiometry of

TTMA, by utilizing its solubility in THF-d4 and employing

anhydrous anisole as an standard also proved inconclusive.

While the results of this experiment showed the presence of

approximately 60% neat TTMA in the LiBr salt mixture, the

reaction of TTMA, with aldehydes indicated a minimum of 77%

TTMA. These anomalies are probably the result of the

presence of adventitious water in both THF and anisole.

Additional investigations on this subject are clearly

necessary for understanding and utilization of this

property of the trialkylaluminums.

Reactions of (TMSCH2,3Al

In order to investigate the selectivity of the

modified Peterson reagent (TTMA) in its reactions with

carbonyl compounds, there was a choice of two different

species of TTMA; the neat TTMA and its lithium bromide

complex (TTMA,). However, because of our desire to

compare the reactivity of the two species, it was decided

to explore the reactions of both reagents with carbonyl

compounds. For the purpose of this study six more aromatic

aldehydes and ketones were added to the previous list of

carbonyls. The new aldehydes consisted of 4-tolualdehyde,

4-chloro-, 4-methoxy-, and 4-trifluoromethylbenzaldehydes;

the new ketones included acetophenone and 1-indanone.

The reaction conditions and reagent ratios for these

reactions were determined according to the reactions of

other trialkylaluminums. The consideration of these

factors is necessary, since they do have a considerable

effect on the course of the reactions of trialkylaluminums

with carbonyl groups. The major factors include the

solvent, the ratio of reagent to reactant, and the

temperature. While the temperature plays a less important

role in reactions of trialkylaluminums, large changes in

temperature do effect their reactivity. For example,

trimethylaluminum is unreactive with carbonyl groups at

-780C, but the best yields are obtained when reactions are

carried out at 0C. A better selectivity seems to be

obtained when reactions of TTMA are performed at room

temperature, nonetheless, TTMA remains reactive at 0C.

According to our experience, hydrocarbon solvents are best

suited for the reactions of trialkylaluminums with

carbonyls. Halogenated solvents are also compatible with

trialkylaluminums, but the more polar oxygenated solvents

(THF and ether) tend to lower both the reactivity of the

reagent and the yield of the reaction.

The ratio of the trialkylaluminum to the carbonyl is

one of the more important factors in the outcome of their

reactions, where an excess of the reagent is generally

required for a clean and a high yield reaction. Published

reports on reactions of trialkylaluminums with carbonyls

and mechanistic considerations show that for optimal

results 1.3 to 2 fold excess of the organo- aluminum is

usually required. 7

The reaction of neat liquid TTMA with excess

benzaldehyde in hexane shows the importance of the reagent

to carbonyl ratio, where the presence of excess aldehyde

causes a variety of major side reactions. The additional

products were identified by GC/MS analysis of the reaction

mixture as the MPV oxidation/reduction and crossed-aldol

condensation products (figure 3-5). The mass spectrum


(3-5a) 21%

(3-5b) 6%


(3-5c) 30%

(3-5d) 20%




Figure 3-5

confirms the structure of the condensation product as 3-5e

or 3-5f, which is the result of the crossed-aldol

condensation of benzaldehyde with p-silylketone. The silyl

group could be positioned either on the oxygen as the silyl

enol ether (3-5e) or p to the carbonyl (3-5f); however,

distinguishing the two isomers is difficult from analysis

of the mass spectra. The origin of the other side products

have already been discussed in Chapter 2.

Considering the importance of the reaction conditions,

all the reaction of TTMA and TTMA, were carried out in

hexanes at room temperature. There were four exceptions

with the reactions of TTMA,; 4-methoxy- and 4-trifluoro-

methylbenzaldehydes, acetophenone, and 1-indanone; here the

aldehydes were added to the cooled reagent mixture

(immersed in ice bath) and were warmed to r.t. within 30

minutes of the addition. The neat TTMA and TTMA, were

previously prepared and stored in the dry box. Prior to

each reaction the reagents were weighed into the reaction

flasks inside the dry box. Outside the dry box, they were

diluted with dry hexanes under an argon atmosphere.

Approximately 1.3:1 ratio of reagent to carbonyl was used

for each reaction. In the case of TTMA,, since the exact

stoichiometry of the salt was unknown, the formula weight

of (TMSCH2)3A1.3LiBr (549.2 g/mole) was used in measuring

1.3 equivalent of this reagent. The products and their

distributions, from the reactions of TTMA and TTMA, with

the aldehydes and ketones have been listed in tables 3-1

and 3-5 respectively.

An overall look at the reaction of the aldehydes with

the neat TTMA reveals the extent of oxidation and reduction

involved. All aldehydes, except for phenylacetaldehyde,

show at least a 10% reduction of starting material

TABLE 3-1. Reaction of (TMSCH2)3Al (TTMA) with selected




CHO A 53 (21) 27

o B 100 --

CHO A 63 (21) 35

B 79 -- -- 19


F B 61- -
O A 38 (20) 10 -- 31d

B 100

CHO A 46 (26) [13] 33 4 10

Me B 35 [31] 2 6 --

S A 41 -- 2 5
B 38 -- 15 27

A 57 (17) 24
000 CHO ---
B 83 (3) 4 -- 2

A : neat (TMSCH2)3Al,
( ): % of elimination
addition yield listed
[ ]: % methyl ketone,

B (TMSCH2)3A1.3LiBr salt.
product, also included in the total
out side the paranthesis.
also included in the total addition

RED. : reduction, COND. : condensation S.M. : starting material
All yields were calculated by GC analysis using tridecane as an
internal standard.


accompanied by its corresponding oxidation product (the

methyl ketone). Ideally the amounts of oxidation and

reduction products should be the same, but, as was

discussed in Chapter 2, the percentage of the Oppenauer

oxidation products detected by GC are generally slightly

less than the MPV reduction ones.

The GC/MS analysis of the reactions of neat TTMA with

4-methoxybenzaldehyde show some products that are not

detected with reactions of most other carbonyls. These

include the elimination product (-TMSOH) and the trimethyl-

silyl protected 4-methoxybenzylalcohol. The silylated

alcohol, also detected in reaction of 4-chlorobenzaldehyde

with neat TTMA, could be either due to GC decomposition or

an actual product of the reaction which survived the

hydrolytic workup (unlikely). Elimination products are

also detected in reactions of 4-chlorobenzaldehyde,

acetophenone, and 1-indanone. The presence of elimination

products with these carbonyls results from the effect of

the substituents on the aromatic ring or a to the alcohol

functionality. These substituents further enhance the

stability of the benzylic cation produced upon loosing the

hydroxy group, hence facilitating the elimination process.

The polymerization of the reactive eliminated products

(substituted styrenes) is probable, it would be undetect-

able on GC, and would lead to inaccurate addition yields.

TABLE 3-2. Reaction of (TMSCH2)3A1 (TTMA) elected




O A 26 -- 56

B 39 3 38

0| A 7 4 71
)I B 8 (5) 2 59

A -- -- 93

B 9 (6) 2 36

0 A 17 -- 55

B 43 51

A: neat (TMSCH2)3A1.
B: (TMSCH2)3A1.3LiBr salt.
c: % of elimination product included in this number.
S.M.: starting material, COND.: condensation products.
( ): % elimination, also included in the total yield.
Yields based on GC analysis using tridecane as an
internal standard.

The impact of substituent effects on the addition

reactions of aromatic aldehydes with TTMA and TTMA, is only

pronounced with 4-methoxybenzaldehyde. Here, the yields of

addition products are inferior to those of other aldehydes;

however, the low yields may have resulted from the

elimination of the TMSCH2 addition product (-TMSOH) and its

subsequent polymerization which is not detectable on GC.

The selectivity of TTMA and TTMA, becomes evident in

its reactions with ketones. In all of these reactions

starting material had been the major constituent of the

product mixture, and the highest yield of addition product

(43%) was obtained with fluorenone. The low reactivity of

TTMA (or TTMA,) with ketones is due to the bulk of the

reagent which hinders the complexation of it with the

sterically more demanding carbonyl groups. Once the

complexation has been achieved, the transferal of the large

TMSCH2 substituent also seems to be slow, and it leads to

the enolization of the ketone rather than addition of the

bulky group. The enolization due to the high Lewis acidity

of TTMA is also a problem with readily enolizable aldehydes

such as phenylacetaldehyde.

A comparison of the reactions of TTMA, and the neat

TTMA show that in general, the reactions of TTMA, are much

cleaner and give higher yields than that of the neat TTMA.

This difference could be attributed to a several speculated

factors. One factor could be the observed higher stability

of the TTMA,. The neat liquid TTMA is spontaneously

flammable upon exposure to air, and produces a white smoke

even on introducing deoxygenated and dried argon gas. The

lithium salt (TTMA,) however, is more stable when exposed

to air; turning yellow and decomposing slowly (note that

the salt is still flammable when exposed to moisture).

Even in the dry box, the lower stability of the neat TTMA

during its continuous usage seem to cause the formation of

aluminum oxides. The presence of these oxides during

reactions with carbonyls appear to increase the rate of the

oxidation and reduction side reactions which are greater in

reactions of the neat TTMA. Another factor is the

complexation of the bromide anion with aluminum which could

slightly decrease the oxophilicity of the alane (similar to

the ate complex), hence decreasing the unwanted oxidation-

reduction reactions after the addition of the TMSCH2 group.

And lastly, it is possible that the reactivity of the

carbonyl group is enhanced by its complexation with the

lithium cations present in the solution.

An indirect measure of the aldehyde selectivity of

TTMA can be obtained by analysis of the tabulated results

of its reactions with aldehydes vs. ketones. However, a

direct measurement of this selectivity can only be achieved

when both functional groups, ketone and aldehyde, are

present on one substrate. This comparison, thus requires

the synthesis of keto-aldehydes for a further demonstration

of the TTMA's high selectivity.

\ (3-6a)

LiAl(t-BuO)3H 33% PDC
\ THF, 4 0C + CH2C1


HO--- CH
Figure 3-6

The synthesis of the first keto-aldehyde was

accomplished by the lithium tri-tert-butoxyaluminohydride

[LiAl(t-BuO)sH] reduction of 4-carbonylchloride-9-

fluorenone (figure 3-6). The reduction afforded 33% of the

target 4-carboxaldehyde-9-fluorenone (3-6a), and 66% of the

corresponding alcohol (3-6b). The aldehyde and the alcohol

were separated by flash column chromatography, and

subsequently the alcohol was oxidized, using PDC, to the

corresponding aldehyde.

The second keto-aldehyde was synthesized by simple

reductive ozonolysis of (1R)-(+)-a-pinene, to give the

expected cyclobutane tethered keto-aldehyde (figure 3-7).

The reaction of 9-fluorenone-4-carboxaldehyde with

neat TTMA gave a mixture of many products which were not


1) 3. CH C12HCH

2) Me2S y

Figure 3-7

identified; however, reaction of the keto-aldehyde with

TTMA, (r.t.) afforded the aldehyde addition product (3-8a)

exclusively in less than 20 minutes (94% homogeneous by

GC). The reaction of TTMA, with the 3-acetyl-2,2-dimethyl-

cyclobutane acetaldehyde under the same conditions within

15 minutes produced 76% of the aldehyde addition product

(3-8b) and less than 4% of what appeared to be the ketone

addition product (uncorrected GC yields, figure 3-8).

It is important to note that none of the reactions

reported in this chapter have been optimized. Thus, an

optimization of the reaction conditions, such as larger

ratios of TTMA, should increase the observed yields even


The high selectivity of TTMA, with 4-carboxaldehyde-

9-fluorenone is somewhat expected, since the ketone moiety

is sterically more crowded (fluorenone) than the aldehyde.

Nevertheless, in 3-acetyl-2,2-dimethylcyclobutane acet-

aldehyde, even though the steric requirements of the two

functionalities are not as varied, a 19 to 1 aldehyde

selectivity is observed. The results from the reactions of


/ (TMSCH2)3A1 /
\ ^Hexanes, r.t.
<20 min.

SiMe3 OH
(3-8a) 94%

0 SiMe3 O
H (TMSCH2)3A1l 1
Hexanes, r.t.
<15 min. OH

(3-8b) 76%
The LiBr salt
Figure 3-8

TTMA, with the two keto-aldehydes further demonstrate its

high selectivity and potential use in organic synthesis.



In investigations pertaining to the mixed

trialkylaluminums and their reactions with carbonyl

compounds, it was noticed that the increase in the TMSCH2

addition products was concurrent with an increase in the

oxidation and reduction products. Oppenauer and Meerwein-

Verley-Pondorff processes, which involve a hydride transfer

through a six membered cyclic transition state, were

probably responsible for these oxidation and reduction

processes (figure 4-1). Additionally, the apparent

increase in rates of these processes for the intermediate

TMSCH2 aluminum alcoholates (4-la and 4-1b) was

preemptively accounted for by a p-silyl effect of the

trimethylsilyl substituent. Hydride transfer during the

Oppenauer oxidation of the p-silylalcohol should reduce the

electron density on the carbon p to the silyl group, and

this deficiency is compensated for by hyperconjugative

electron release from the silyl group, which in turn

facilitates the rate of the hydride transfer.

The validity of this hypothesis was examined by a

comparative rate study of the Oppenauer oxidation of a

p-silyl substituted alcohol vs. the unsubstituted analog.

It is worth noting that this rate study was aimed at

comparing the rates of the oxidations of the two alcohols

under normal reaction conditions, and was not meant to be

an exact determination of the oxidation rates.

,, _hcHo 0.-) Si-e3
R2A1CH2TMS (4-la) hO ,j i ? SiMe3

PCHO -- Ph
OA1R, y H V^

R Me, TMSCH2 0

Figure 4-1. The beta Silyl effect and its rate enhancement of the
MPV reduction.

Trimethylaluminum was added to separate solutions (0.1

M in hexanes, 0C) of 1-phenylethanol and 1-phenyl-2-

trimethylsilylethanol. An equimolar amount of benzaldehyde

was then added to each solution at room temperature (figure

4-2). The GC analysis of the reactions, monitored at

periodic intervals, showed that half of the p-silylalcohol

was oxidized within the first 15 minutes of the addition,

whereas 1-phenylethanol required close to 2.5 hours. This

difference translates to a 10 fold rate enhancement in the

oxidation of p-silylalcohol under normal reaction

conditions. Even though the rate enhancement is not

enormous, its effects on the reaction course is significant

and apparent in tabulated reactions of TMSCH2 substituted

alanes (Chapters 2 and 3).

H jAIMe2 h
SPhCHCH3 i- PhCCH3 + PhCh2OH

MeA1 + OH2

k2/k, a 10

Figure 4-2

Shortly after completion of the above investigation,

there appeared an interesting report by A. Barron that

attracted our close attention.58

This report, and a series of others by Barron,59,60

involved a study of the properties, structures, and

reactions of BHT (2,6-di-tert-butyl-4-methylphenoxy)

substituted alkylaluminums. The catalytic features of the

di-BHT and analogous optically pure binaphthol substituted

alkylaluminums have recently been examined by other

groups61 as Lewis acid activators of pericyclic reactions

involving oxygen centers. Additionally, the exceptionally

bulky di-BHT substituted alkylaluminums have shown

excellent diastereofacial selectivity in carbonyl

H C 0 0 CH3





t-Bu 1) AD t- OH + t- R
2) MeMgBr
84% 99 1

Figure 4-3

alkylation with reactive organoalkyl reagents,39 e.g.

methyl magnesium halide (figure 4-3).

However, contrary to expectations based on organo-

aluminum chemistry to that point, Barron had reported a

novel methyl addition by (BHT)Me2Al:OEt2 (DMBAE) to

aromatic aldehydes to yield directly the corresponding

methyl ketone (figure 4-4). In a subsequent report there

was a suggestion of a rather intriguing but speculative

mechanism for this, at first glance, extraordinary reaction

(figure 4-4).





X Cl. CH, NO2, H

.0..... si H
MBH... 'A' Ph

B T O......... *. Ph
e "Me

Me \Me-- .. -A l.. H
4-4a Me\

4-4d Me


BHT## ........
Me e .. BHT
Me Atl-*

BH~' ".... >H .. H
Me A 'Me- BHT
Me- Me

Figure 4-4

The methylation of an aldehyde by a dimethylaluminum

alkoxide, to the best of our knowledge, was the first one

of its kind reported; however, no attempts were made by the

authors to explain the mode of this addition. Until this

point, it was generally believed that the substitution of

even one alkoxide on an alkylaluminum deactivates the

aluminum in its alkyl additions to aldehydes or ketones.

The addition of only one out of the three alkyl groups on

trialkylaluminums to carbonyls, further supported this


The deaggregating effect of the BHT substituent on

aluminum, a key factor in the unexpected methylation

ability of DMBAE, is demonstrated by the low temperature

(-800C) H NMR study reported by Ittel et al..62 The study

shows that, contrary to previous reports (eq. 4-5a), the

species present in solutions of AlMe2BHT (DMBA) undergo two

concurrent disproportionations ( eqs. 4-5b and 4-5c). This

type of a dissociation is not known for similar dialkyl-

aluminum alkoxides bearing smaller alkoxy substituents.


1/2 Al2Me6 + A1MeBHT2



3 AIMe BHT <


1/2 Al2Me6 + AlMe2BHT

Al2Me BHT + AlMeBHT2

Me.,, O Me
M Al Ale
Me MeMe

Figure 4-5

The larger bulk of the BHT substituent inhibits the

formation of the stable and unreactive hemialkoxide, and

leads to its disproportionation and higher reactivity.



From the information that is available in the

literature, the identification of the actual methylating

species is not possible. Trimethylaluminum is the best

candidate for the possible methylating agent present in the

solutions of DMBA or DMBAE; however, acetophenone reacts

with AlMe3 but it is not methylated in solutions of

A1Me2BHT or its etherate. The selection of any other

methylating agent at this point would be speculative in

nature. Nevertheless, with the information presented here,

we will also try to shed some light on the identity of the

methylating agent.

By far the most interesting and peculiar result

reported by Barron and coworkers was the oxidation of an

alcohol by the reduction of a dialkylaluminum aryloxide.

In our investigations of the reactions of aldehydes with

dialkyl alkoxy and trialkylaluminums, discussed previously,

we had also encountered reduction and oxidation products.

However, the concurrent presence of the reduced aldehyde

and the oxidized product was clearly suggestive of MPV and

Oppenauer type of processes. In the light of our studies

and the well known reduction ability of alkoxyalumino-

hydrides, the quantitative conversion of the aromatic

aldehydes to the methyl ketone using DMBAE through an

irreversible hydride transfer to aluminum did not seem

plausible. Thus, as a consequence of the anomalies

discussed above, we were led to embark on our own investig-

ation of the reactions of DMBAE with aromatic aldehydes.

The preparation of DMBAE is easily performed by mixing

equimolar hexane solutions of trimethylaluminum and BHT-H

(at r.t., under argon), and subsequently adding a slight

excess of anhydrous ethyl ether. DMBAE can be prepared on

large scale and stored in the dry box, or it can be

prepared in situ and used fresh (figure 4-6).

BHT-H + MeAl Hexane BHTMe2A + CH4

BHTMe2Al + xs Et20 D BHTMe2A1:OEt2

Figure 4-6

The reactions of DMBAE with benzaldehyde were carried

out using a variety of reagent to aldehyde ratios (up to

3:1), using different solvents (hexane and toluene), and

using different orders of addition (adding aldehyde to

reagent and vice versa). However, all the reactions

afforded only close to a 50/50 mixture of benzyl alcohol

and acetophenone (table 4-1). Also, when BHTMe2Al was

reacted with benzaldehyde in the absence of ether,

considerable amounts of 1-phenylethanol was also detected

illustrating the important and unclear roll of the ether in

the oxidation-reduction processes involved. In all of the

reactions of benzaldehyde with DMBAE a small amount of

2-phenylisopropanol, a product of a possible subsequent

methyl addition to acetophenone, was also detected.

Table 4-1. Reaction of aromatic aldehydes with DMBA and DMBAE

and Ir.t. PhCH2OH + PhCCH3
Me BHTAl:OEt2 1 9

+ PhC(CH3)2 + PhCHO + PhCHCH3

RCHO 1 2 3 4 5

DMBAEb Ph 1.5 Hex. 26 20 2 3 --

DMBAE Ph 2 Hex. 45 32 9 2

DMBAE Ph 3 Hex. 34 34 6 2 9

DMBAEa 4-Tol. 1.5 Hex. 29 28 -- 37 2

DMBAE Ph 1.5 Tol. 42 33 4 3 -

DMBAEb Ph 1.5 Tol. 40 33 5 3

DMABc Ph 1.5 Hex. 19 28 -- -- 29

DMBA Me2BHTA1, DMBAE [Me2BHTAL]OEt2, Ph phenyl, 4-Tol. 4-Tolyl
Hex. hexanes, Tol. toluene, Equiv. equivalent.
a reaction time was only ten minutes, the rest for 3.5 to 4 hours.
b reverse addition, added the reagent to the aldehyde. All yields
based on GC analysis, using tridecane (c: BHT was used as standard) as
an internal standard.

Our findings at this point were neither in accord with

Barron's results nor supportive of the reported oxidation

mechanism. Thus further investigations were required to

specifically examine the validity of the suggested

mechanism. This task was accomplished by approaching the

suggested reaction intermediates by an alternative route:

addition of one or two equivalents of 1-phenylethanol to

one equivalent of DMBAE should afford the correct

stoichiometry and structural composition postulated as the

reaction intermediate 4-4c by Barron (figure 4-7)


(B3) 1 mmole 1 mmole hexanes, r.t5 t2
1 3.5 hrs Q \
j: OEt2

hexanes, r.t. t..,,
(B4) 2 mole 1 mmole
1 3.5 hrs p .*BHT


Figure 4-7

Stirring two mmoles of DMBAE with one mmole of

1-phenylethanol in hexane for three hours did not result in

any detectable oxidation of the latter, as it was

implicated for the proposed intermediate 4-4c. The

subsequent addition of 0.5 mmole of benzaldehyde to this

mixture resulted in the rapid (less than one hour)

methylation and reduction of benzaldehyde (0.14 mmoles),

along with a small amount of methylated acetophenone (0.03

mmoles), (table 4-2). Reaction of a 1:1 mixture of

1-phenylethanol and DMBAE (1 mmole) with benzaldehyde

Table 4-2. Reactions of B3 and B4 with aromatic aldehydes.
4-2a: Reactions with benzaldehyde
4-2b: Reactions with 4-tolualdehyde




0.5 B4

PhCH +




'hCH2 +









CH3 +








0.5 B4

ArCH2 +





rCHCH3 +



The values in tables (4-2a) and (4-2b),reported as millimoles.
Reactions of B3 were run for 18-19 hours, and reactions of B4 for
5.5 hours. Ar 4-Tolyl.



LrCCH3 +













(0.50 mmoles) afforded, slowly, the reduction of the

aldehyde (0.31 mmoles) and the oxidation of 1-phenyl-

ethanol to acetophenone (0.40 mmoles). Even though both

experiments clearly disprove the proposed oxidation

mechanism by showing the need of a hydride acceptor, the

interpretation of the observed product distribution is not

straight forward for two reasons: a) the methylation

product of benzaldehyde, 1-phenylethanol, is also used as a

starting material which obscures the origin of the alcohol

in the final product ratios, b) 1-phenylethanol and

acetophenone exhibit very close GC retention times

(separated by only 0.03 minutes) which results in peak

overlap at certain concentrations and perturbs the accuracy

of the GC analysis. To alleviate these difficulties, the

above reactions were repeated using 4-tolualdehyde for

which the GC retention time of methylated adduct and its

oxidized form are well separated.

The reaction of reagent B3 (figure 4-7) with 4-tolu-

aldehyde is a slow process which takes more than 5 hours to

complete and yields only the oxidation and the reduction

products (similar to benzaldehyde, table 4-2). Since the

above reaction is very slow, unlike the reaction of DMBAE

with benzaldehyde, and there are no methylated 4-tolu-

aldehyde products, then it is clear that the species of

aluminum involved in this reaction are not responsible for

the oxidation or the methylation processes of DMBAE.

The reaction of reagent B4 (figure 4-7) with 4-tolu-

aldehyde is a much faster process (less than 1 hour) and is

similar to the reaction with benzaldehyde with the same

reagent B4. The products from this reaction include the

methylation and reduction of 4-tolualdehyde along with the

oxidation of the methylation product (table 4-2). The

difference between this reaction and that of benzaldehyde

is the apparent consumption of the acetophenone in the

latter reaction by the condensation with benzaldehyde.

The most striking observation of the product

distribution for the 4-tolualdehyde reaction with B4 is the

exclusive reduction, oxidation, and methylation of the

4-tolualdehyde and almost complete recovery of the starting

1-phenylethanol. An important conclusion drawn from this

observation is that the methylation and the oxidation occur

so rapidly that there is no possibility of a substituent

exchange. This conclusion also implies that no 1-phenyl-

ethoxides are present on the methylating and oxidizing

aluminum species. In light of this outcome, our earlier

hypothesis, of trimethylaluminum as the methylating reagent

appears more certain. To further justify our hypothesis it

was decided to do a comparative rate study of the Oppenauer

oxidation of 1-phenylethanol using trimethylaluminum and

DMBAE (figure 4-8).

This experiment was carried out by mixing 1-phenyl-

ethanol (0.02 M solution) with separate equimolar solutions

of TMA and DMBAE (1:1 ratio). Subsequently, a solution of

benzaldehyde (1 equivalent, r.t.) was added to each

mixture. The monitoring of these reactions by GC showed

that the oxidation of the alcohol with the TMA-l-phenyl-

ethanol adduct was half complete within two hours, whereas,

even after 5 hours, the oxidation of the DMBAE-l-phenyl-

ethanol adduct did not reach one half-life of reaction.

MeAl |
H PhCHO ||
PhCHCHr + PhCCH1 + PhCh2OH

Figure 4-8

The faster oxidation of the dimethylaluminum alkoxide

vs. the BHT substituted aluminum alkoxide, the rapid

methylation and oxidation observed with reactions of DMBAE

with aldehydes, and the reaction of A with 4-tolualdehyde

all support the argument that trimethylaluminum is the

methylating agent in reactions of DMBA and DMBAE. The non-

reactivity of DMBAE with acetophenone discussed previously

would not appear to be in accord with this hypothesis;

however, it can be explained by the strong complexation of

DMBA with acetophenone.59 Furthermore the addition of

ether to DMBA should diminish the disproportionation

process of DMBA (eq. 4-5b) and thus reduce the concen-

tration of TMA in solution. The subsequent addition of

acetophenone to such a solution would result in only a

small amount of methylation before the complexation of the

remainder of the acetophenone with DMBA is complete. The

latter complexation would further impede the production of

TMA, and would explain the DMBAE's lack of reactivity

towards acetophenone.

From the above discussion ether appears to have a two

fold effect on the methylation and oxidation reactions of

DMBAE with aldehydes: a) it decreases the concentration of

TMA in solution which decreases the methylation rate, and

allowing sufficient time for the slower oxidation process

to take place, b) it helps the break up of the

hemialkoxides formed after transference of the methyl

group, thereby facilitating the complexation of another

molecule of aldehyde which leads to faster oxidation and

reduction rates.



The results presented in this dissertation, reveal

that unlike the previously known mixed organoaluminum

reagents, the preferred transferal of the more stable

carbanion is not always possible. While the preferred

delivery of the simple alkyl substituents (methyl and

ethyl) over the TMSCH2 group, can be rationalized by the

exceptional bulk of the latter, the preferred addition of

large substituents (1-indanyl) over smaller alkyl groups to

sterically demanding ketones (fluorenone), has also been

reported. It is possible though, that the steric

requirements of the TMSCH2 substituent outweigh the effect

of its carbanion stability, which would lead to its lack of

reactivity. There could also be an argument that the lower

anion stability of the TMSCH2 anion compared to ethyl and

methyl anions would also lead to the observed preferred

addition of the smaller alkyl groups; however, this

possibility is highly unlikely. An interesting correlation

is observed in comparing the size and the reactivity of

trialkylaluminums in their reactions with carbonyls. This

correlation is especially noticeable in reactions of

fluorenone where, as the bulk of the organoaluminum

increases (TMA < DMTA a DETA < MDTA < TTMA), the reactivity

and the product yield of alkyl addition decreases.

In contrast of the reactions of mixed trialkyl-

aluminums with aldehydes and ketones, tris(trimethylsilyl-

methyl)aluminum (TTMA) is highly selective towards

aldehydes. Two forms of TTMA can be employed in reaction

with carbonyls; the neat liquid TTMA and TTMA,

[(TMSCH2)3A1.3LiBr]. Reactions of TTMA, with most

aldehydes afford good to excellent yields of TMSCH2

addition products, whereas with ketones the yields are low

and mostly unreacted starting material is recovered. The

selectivity also holds true for the neat TTMA; however, the

yields are not as impressive as with TTMA,. The physical

characteristics (powder vs. liquid), higher air stability,

and ease of preparation of TTMA, make its use and

manipulation more convenient than the use of the neat

liquid TTMA. In comparing neat TTMA to TTMA,, it is clear

then, that TTMA, is the preferred reagent for the selective

methylenations of aldehydes.

In consideration of these results, the task of

modifying the Peterson reagent for selective methylenation

of aldehydes is accomplished by the development of the

TTMA, reagent. The high selectivity of this reagent has

been proven by its tabulated reactions with aldehydes vs.

ketones, and more convincingly, by its reactions with the


And lastly, the results of the investigations on

reactions of DMBAE and DMBA show that trimethylaluminum

liberated from disproportionations of DMBAE or DMBA is the

most likely methylating agent in the reactions of these

aluminum reagents with aromatic aldehydes. This hypothesis

is supported by the faster Me3Al assisted MPV oxidation of

1-phenylethanol vs. the similar DMBAE assisted oxidation.

Additionally, we have been unable to duplicate Barron's

reported high yields of methylation/oxidation reactions of

DMBAE with aromatic aldehydes, and have no explanations for

his published results.





All solvents were freshly distilled before employing

them in reactions, and were transferred using syringe

techniques. Hexanes, THF, toluene, diethyl ether, and

pentane were distilled from Na and benzophenone ketyl;

methylene chloride was dried over CaH2 and then distilled.

Reagents and Chemicals

All organometallic reagents were purchased from

Aldrich in sure seal bottles (Trimethylsilylmethyllithium

1.0 M in pentane, trimethylaluminum 2.0 M in hexanes,

dimethylaluminum chloride 1.0 M in Hexanes, methylaluminum

dichloride 1.0 M in hexanes, diethylaluminum chloride 1.0 M

in hexanes, MeMgBr 3.0 M in diethyl ether, MeLi 1.4 M in

diethyl ether, LiAlH(OBut)3 1.0 M in THF). AlBr3 was

purchased from Aldrich (98%) and Alpha (better quality,

99.9% or 98%), and it was sublimed before using. Heptanal,

cyclohexanone, benzaldehyde, 4-tolualdehyde, anisaldehyde,

acetophenone and phenylacetaldehyde were all distilled and

stored over activated 3A molecular sieves, under nitrogen

in septum fitted bottles. 1-indanone, fluorenone, and

4-chlorobenzaldehyde were recrystallized before use.

Spectra and Instruments

1H and 13C NMR spectroscopy was carried out on either

the General Electric QE-300 or Varian VXR-300

spectrometers. GC/MS and electron impact/low resolution

mass spectra were obtained on a Finnigan MAT 4500 mass

spectrometer. A Finnigan MAT 95 spectrometer was used for

high resolution electron impact and chemical ionization

exact mass determination. Infra-red spectra were run on a

Perkin-Elmer Model 1600 FT-IR spectrophtometer. GC

analysis was performed on HP5880A (crosslinked methyl-

silicon [HP] high performance capillary column). Melting

points were taken on a Thomas-Hoover capillary melting

point apparatus.

GC Standards

For a quantitative analysis of the reaction products

tridecane was used as an internal standard. Analytical

solutions of authentic reaction products and tridecane were

prepared and the relative response factors of the GC

detector to tridecane and the products were calculated. A

known amount of tridecane was added to the reaction

mixtures either during or after quenching of the reactions;

using this amount, the areas from the GC chromatogram and

the relative response factors, the corrected percentages

and yields were determined. The % error in yields

resulting from the use of the standards is assumed to be

4%. It was estimated by multiple runs of known standard

and substrate mixtures.

Apparatus and Technique

All glassware used for air-sensitive reactions was

flame dried under vacuum and filled with an inert

atmosphere of argon by successive purging and charging

using a dual manifold vacuum line; glassware was also dried

in oven for over 24 hours (110 OC) and then purged with

argon, as described previously, before use. Standard

syringe techniques were used for the introduction of liquid

reagents and solutions to reaction vessels; all syringes

were dried in oven (above conditions) and before use were

cooled in a desiccator or under a nitrogen atmosphere.

Experimental Procedures

Preparation of Authentic Samples of Trimethylsilylmethyl-

l-Phenyl-2-trimethylsilylethanol.63 In a dry round

bottom flask, equipped with a magnetic stirring bar, rubber

septum, and purged with argon was placed 2.50 ml of 1.0 M

pentane solution of trimethylsilylmethyllithium. To this

solution was added 4.0 ml of dry THF (this prevents the

crystallization of the reagent when cooled to -78 OC), and

the flask was immersed in a dry ice-acetone bath.

Benzaldehyde (0.24 g, 2.3 mmoles) dissolved in 4.0 ml of

dry THF was then added to the cooled stirred solution (a

slight yellow tint was observed upon addition of the

aldehyde; and after 10 minutes GC examination showed 97%

conversion to product). The reaction was allowed to stir

with warming to 0 OC (ca. 20 minutes), and was quenched by

adding 5.0 ml of 1.0 M HC1. The two layers were separated,

the aqueous layer was extracted with 3 x 5 ml of ether, and

the combined organic layers were washed with 5 ml of

distilled water followed by 5 ml of brine. The organic

layer was dried (MgSO4) for two hours, filtered, and

concentrated in vacuo to yield 0.25 g (1.3 mmoles, 58%

yield, 97% homogeneous by GC) of a light yellow oil. 1H

NMR (CDC13), trimethylsilyl 8 0.00 [(ref.), (s, 9 H)],

1.35 (-CH2a: ddd, JHz = 14.32, 7.37, 0.44,1 H), 1.26

(-CH2p: ddd, JHz = 14.32, 7.69, 0.47, 1 H), 2.0 (-OH:

s,broad, 1 H), 4.92 (-CH: dd, JHZ = 7.6, 7.5, 1 H),

7.40-7.65 (Ar: m, 5); 13C NMR: 8 Me3Si -1.1, -CH2 28.4, -CH

72.9, Ar 125.85, 127.6, 128.5, 145.3. GC/MS, m/e (70 ev):

191 (0.08), 179 (2), 107 (13), 104 (43), 77 (15), 75 (100),

73 (20), 45 (14).

1-(4-Chlorophenvl)-2-trimethylsilylethanol.4 Same

procedure as above: 6.20 ml (6.20 mmole) of trimethyl-

silylmethyllithium, 5 ml of THF, and 0.85 g (6.05 mmoles)

of 4-chlorobenzaldehyde; solution changed from mirky to

clear to orange yellow by the end of the addition. After

concentration, 1.31 g (5.6 mmoles, 97% homogeneous by GC,

92% yield) of a light yellow oil was obtained. 1H NMR:

(CDC13), Trimethylsilyl 8 0.00 [(ref.) s, 9 H)], 1.18

(-CH2a: dd, JHz = 14.26, 7.72, 1 H), 1.27 (-CH2p: dd, JHz =

14.26, 7.34, 1 H), 2.57 (-OH: s, broad, 1 H), 4.83 (-CH:

t, JHz = 7.54), 7.29-7.39 (Ar: m, 4 H); 13C: 8 Me3Si -1.2,

-CH2 28.4, -CH 72.0, Ar (127.2, 128.4, 132.9, 144.9). GC/HR

(CI, methane); 228 (0.2), 219 (100), 211 (34), 131 (22),

calculated for C11H17ClOSi: found 228.074, calculated



Same procedure as above: 6.20 ml (6.2 mmoles) of trimethyl-

silylmethyllithium, 5 ml of THF, and 1.05 g (6.03 mmoles)

of 4-triflouromethylbenzaldehyde; solution changed from

colorless to greenish blue, almost black. After

concentration, 1.78 g of a yellow oil (94% homogeneous by

GC) was obtained. The oil was dissolved in methylene

chloride, treated with activated charcoal, acid washed

alumina, filtered, concentrated in vacuo, and was placed

under higher vacuum (0.01 mm Hg, 30 minutes), to yield 1.10

g of light yellow oil (4.2 mmoles 99% homogeneous by GC,

69% final yield). H NMR: (CDC13), Trimethylsilyl 8

0.00 [(ref.), s, 9 H], 0.89 (-CH2a: dd, JHz = 14.40, 6.35,

1 H), 0.98 (-CH2p: dd, JHz = 14.40, 8.04, 1 H), 2.20 (-OH:

s, broad, 1 H), 4.93 (t, JHz = 7.45), 7.21 (Ar: d, JHZ =

8.15, 2 H), 7.34 (Ar: d, JHz = 8.15 2 H); 13C: 8 Me3Si

-1.1, -CH2 28.6, -CH 72.2, -CF3 124.2 (q, JHz = 272), Ar

[C3 125.3, (q, JHZ = 3.7), C2 126.0, C4 129.55 (q, JHz =

32.3), C1 150.6. HR/MS (CI, methane); calculated for

C12Hi6F30Si 261.0922 [(M H)+, (0.8)], measured 261.091,

fragmentation: 263 (0.3), 262 (0.2), 243 (41), 153 (100).

1-(4-Tolyl)-2-trimethylsilylethanol.64 Same procedure

as above: 10.0 ml (10.0 mmoles) of trimethylsilylmethyl-

lithium, 15 ml of THF, and 1.15 g (9.8 mmoles) of

4-tolualdehyde; color of solution changed to light

yellow. Upon concentration, 2.00 g (9.59 mmoles, 97%

homogeneous by GC, 95% yield) of a light yellow oil was

obtained. 1H NMR (CDC13), trimethylsilyl 8 0.00 [(ref.),

(s, 9 H)], 1.22 (-CH2a: ddd, JHz = 14.25, 7.94, 1.23, 1 H),

1.31 (-CH2p: ddd, JHZ = 14.13, 7.18, 1.15, 1 H), 2.42

(-CH3: s, 3 H), 2.48 (-OH: s, broad, 1 H), 4.92 (-CH: t,

JHz = 7.57, 1 H), 7.20 (Ar: d, JHZ = 7.93, 2 H), 7.30 (Ar:

d, 7.81, 2 H); 13C NMR: 6 Me3Si -1.2, -CH3 21.0, -CH2 28.1,

-CH 72.4, Ar (125.73, 128.91, 136.9, 143.5). GC/HR (CI,

methane); calculated for C12H20OSi 208.1283 measured

208.128 (2), fragmentation: 207 (4), 191 (100).

2-Phenvl-l-trimethvlsilvl-2-propanol.65 Same

procedure as above: 6.40 ml (6.40 mmole) of trimethyl-

silylmethyllithium, 5 ml of THF, and 0.72 g (6.0 mmoles) of

acetophenone; color of solution changed from clear

colorless to light yellow by the end of the addition. Upon

concentration, 1.14 g (74% conversion to product, and

unreacted starting material) of a yellow oil was obtained.

The product mixture was purified by flash chromatography

(80:20, hexane:ethyl acetate mixture) to give 0.81 g of a

colorless oil (3.9 mmoles, 95% homogeneous by GC, 65% final

yield). 1H NMR: (CDC13), Trimethylsilyl 6 0.00 [(ref.),

s, 9 H), 1.52 (-CH2: s, 2 H), 1.76 (-CH3: s, 3 H), 1.99

(-OH: s, 1 H), 7.34 [(Ar (ortho): ddm, JHZ = 7.08, 5.07 2

H)], 7.45 [(Ar (meta): ddm, JHz = 7.75, 7.15, 2 H)], 7.60

[(Ar (para): dm, Jnz = 7.08, 1 H)]; 13C: 8 Me3Si: -0.12,

-CH3 33.5, -CH2 35.0, CH 75.0, Ar (124.5, 126.4, 128.0,

149.7). GC/MS, m/e (70 ev): 207 (1<), 193 (22), 121 (33),

103 (43), 75 (100), 43 (43).

9-Trimethylsilylmethyl-9-fluorenol. Same procedure as

above: 6.40 ml (6.40 mmole) of trimethylsilylmethyl-

lithium, 10 ml of THF, 1.08 g (6.00 mmoles) of fluorenone;

the color of solution changed from clear colorless to a

dark red wine color by the end of the addition. Upon

quenching of the reaction mixture with 7 ml of saturated

ammonium chloride, the color of the mixture changed to a

bright cherry red color. After concentration, 1.77 g (75%

conversion to product) of a yellow powder was obtained.

The product mixture was purified by washing the yellow

solid with small amounts of hexane to remove the yellow

coloring, and then it was recrystallized from hexane. The

recrystallization resulted in isolation of 1.11 g (4.14

mmoles, 69% yield, 88% homogeneous by GC) of white

crystals. Even though GC showed the presence of impurities

with higher retention times than that of the product,

proton NMR, carbon NMR, and CHN analysis were indicative

of a single product. It was concluded then that the

impurities present in the GC chromatogram were due to the

decomposition of the product, either at the injection port

or in the column of the GC. 1H NMR: (CDC13), Trimethyl-

silyl 8 0.00 [(ref.), S, 9 H), 2.37 (-CH2: s, 2 H), 3.03

(-OH: s, 1 H), 7.92 (Ar: H2,7, ddd, JHZ = 7.32, 7.32, 1.26,

2 H), 7.99 (AT: H3,6, ddd, JHZ = 7.38, 7.46, 1.38, 2 H),

8.10 (Ar: Hi,8 dm, JHz = 7.20, 2 H), 8.24 (Ar: H4,5 dm, JHZ

= 7.95, 2 H); 13C: 8 Me3Si: -1.5, -CH2 29.3, -CH 81.6, Ar

(119.9, 123.8, 127.9, 128.8, 138.9, 149.6); HR/MS m/e (70

ev) 268.13 (13), 250.99 (4), 181.07 (100), 178.08, (46),

152.06, (18), 75.02 (38), calculated for C17H2oOSi

268.1284, measured 268.1286.

l-Trimethylsilyl-2-octanol.66 Same procedure as

above: 17.5 ml (17.5 mmole) of trimethylsilylmethyl-

lithium, 10 ml of THF, 1.94 g (17.0 mmoles) of heptanal.

After concentration, 2.75 g (61% conversion to product, and

condensation products) of a yellow oil was obtained. The

product mixture was purified by vacuum distillation using a

small fractional distillation column (14/20 joint, and

vigreux column) at a pressure of 0.5 mm Hg with the

receiving flask cooled in a dry ice-acetone bath. The

product distilled at close to room temperature (ca 20-230

C). 1H NMR: (CDC13), trimethylsilyl 8 0.00 [(ref.), 9 H,

s], 0.84 (overlapping -CH3 and R3SiCH2-: m, 5 H),

1.16-1.48 (-CH2(chain): m, 10 H), 1.56 (-OH, v. broad s, 1

H), 3.69-3.81 (-CH: m, 1 H); 13C NMR: Me3Si -0.76, 14.0,

22.6, 25.7, 26.6, 29.3, 31.8, 40.8, -CH 70.0. GC/HR (CI,

methane); calculated for [(M H)*, (1)] C11H250Si

201.1675, measured 201.168, fragmentation: 186 (20), 185

(100), 131 (25), 117 (36), 75 (28), 73 (51).

3-Phenyl-l-trimethylsilyl-2-propanol.66 Same procedure

as above: 10.5 ml (10.5 mmole) of trimethylsilyl methyl-

lithium, 25 ml of THF, and 1.20 g (10.0 mmoles) of

phenylacetaldehyde; color of the solution changed from

clear colorless to yellow by the end of the addition. Upon

concentration, 1.67 g of a yellow oil was obtained. GC

analysis of the final product showed only 23% of addition

product along with 45% unreacted aldehyde and 20% of aldol

condensation. The attempts of purification through

distillation were unsuccessful, thus GC/MS was employed in

identification and characterization of the product. GC/MS:

m/e (70 ev), 207 (<0.1), 191 (0.6), 117 (46), 101 (3), 92

(15), 91 (19), 75 (56), 73 (100), 45 (18).

1-Trimethvlsilvlmethvl-l-indanol. Same procedure as

above: 6.4 ml (6.4 mmole) of trimethylsilylmethyllithium,

10 ml of THF, and 0.76 g (5.8 mmoles) of 1-Indanone; color

of the solution changed to yellow by the end of the

addition. Upon concentration, 1.00 g of a yellow oil was

obtained. GC analysis of the final product showed only 45%

of addition product along with 46% of unreacted ketone.

The purification of the mixture by flash chromatography

caused the elimination of TMSOH and H20 yielding a mixture

of products. Due to the inability to isolate of the TMSCH2

addition product, GC/MS was employed for identification and

characterization of the alcohol product. GC/MS: m/e (70

ev), 203 (39), 202 [-18, (100)], 188 (29), 187 (94), 129

(26), 128 (72), 74 (51), 73 (91), 59 (42), 45 (36).

1-(4-Methoxvphenvl)-2-trimethylsilylethanol. Same

procedure as above: 6.2 ml (6.2 mmole) of trimethylsilyl-

methyllithium, 5 ml of THF, and 0.82 g (6.02 mmoles) of

4-methoxybenzaldehyde; solution changed from mirky to clear

with a yellow tint by the end of the addition. Upon

concentration, 1.14 g of a yellow oil was obtained. GC

analysis of the final product showed only 14% of addition

product along with 53% of the 4-methoxy styrene

(elimination of TMSOH). The ease of the elimination

reaction prevented the isolation of the alcohol product,

hence, GC/MS was employed for identification and

characterization of the TMSCH2 addition product. GC/MS:

m/e (70 ev), 224 (15), 209 (22), 207 (14), 191 (8), 137

(100), 134 (27), 115 (16), 109 (21), 91 (12), 75 (83), 73

(39), 45 (16).

1-Trimethylsilylmethylcyclohexanol. 6 Same procedure

as above: 15.0 ml (15.0 mmole) of trimethylsilylmethyl-

lithium, 25 ml of THF, and 1.47 g (15.0 mmoles) of

cyclohexanone; color of the solution changed from clear

colorless to light yellow by the end of the addition. Upon

concentration, 2.25 g (81% conversion to product, and

unreacted starting material) of a yellow oil was obtained.

The product mixture was purified by vacuum distillation

using a short path distillation column at pressure of 0.05

mm Hg. The first fraction was collected in the receiving

flask cooled in a dry ice-acetone bath (starting material

and some product), with the product distilling at

33.5-340C. Crystallization began in the receiver, but

after a few minutes the distillate started crystallizing in

the condenser. At this point the condenser was heated

using a heat gun to prevent the clogging of the tube. A

total of 1.45 g of white crystals was obtained (m.p.

35-360C, 7.72 mmoles, 99% homogeneous by GC, final yield of

52%). 1H NMR: (CDC13), trimethylsilyl 8 0.00 [(ref.), 9 H,

s], 0.90 (s, 2 H), 1.19 (v. broad, 1 H), 1.33-1.87 (m, 10