Contrasting reactions of ketones and thioketones with alkyllithiums: A coordinated experimental and computational investigation

Article abstract of DOI:10.1021/ja210847n

The reaction of ketones with organolithium reagents generally proceeds by addition of the organometallic to the electrophilic carbon of the C=O group to give the lithium salt of a tertiary alcohol. The seemingly analogous reaction of thioketones with organolithiums is a fundamentally different process: such reactions typically afford a variety of products, and addition of the organolithium to carbon of the C=S group to give a thiol is a relatively unimportant component. Reactions of the stable thioketone, adamantantanethione (1), with several alkyllithiums (MeLi, n-BuLi and t-BuLi) in a variety of solvents have been studied in the first comprehensive investigation of the reactions of organolithiums with a representative alkyl-substituted thione. Reactions of 1 with n-BuLi or t-BuLi afforded 2-adamantanethiol (2) as the major product. In an effort to explain the marked difference in behavior of ketones and thioketones in reactions with organolithiums, transition states for both the addition and reduction reactions have been located at the B3LYP/6-311+G* level using acetone and thioacetone as model substrates. The transition states for the addition of dimeric MeLi to the C=O and C=S carbons of acetone and thioacetone were significantly different as a result of the small bond angles preferred by divalent sulfur, and this accounts for the much slower addition to a C=S carbon vis-a-vis a C=O group. Transition states for reduction of acetone and thioacetone by EtLi were similar, but the greater exothermicity of the reduction of the thioketone results in an earlier transition state and lower activation energy for this process than that for the reduction of a ketone. The possible role of radical-mediated processes in this chemistry is also discussed.

Full text of DOI:10.1021/ja210847n

Article
pubs.acs.org/JACS
Contrasting Reactions of Ketones and Thioketones with
Alkyllithiums: A Coordinated Experimental and Computational
Investigation
,
†
†
,§
William F. Bailey,* Ashley L. Bartelson, and Kenneth B. Wiberg*
†Department of Chemistry, University of Connecticut, Storrs, Connecticut 06269-3060, United States
§Department of Chemistry, Yale University, New Haven, Connecticut 06520-8107, United States
*
S Supporting Information
ABSTRACT: The reaction of ketones with organolithium
reagents generally proceeds by addition of the organometallic
to the electrophilic carbon of the CO group to give the
lithium salt of a tertiary alcohol. The seemingly analogous
reaction of thioketones with organolithiums is a fundamentally
different process: such reactions typically afford a variety of
products, and addition of the organolithium to carbon of the
CS group to give a thiol is a relatively unimportant
component. Reactions of the stable thioketone, adamantanta-
nethione (1), with several alkyllithiums (MeLi, n-BuLi and t-
BuLi) in a variety of solvents have been studied in the first comprehensive investigation of the reactions of organolithiums with a
representative alkyl-substituted thione. Reactions of 1 with n-BuLi or t-BuLi afforded 2-adamantanethiol (2) as the major
product. In an effort to explain the marked difference in behavior of ketones and thioketones in reactions with organolithiums,
transition states for both the addition and reduction reactions have been located at the B3LYP/6-311+G* level using acetone and
thioacetone as model substrates. The transition states for the addition of dimeric MeLi to the CO and CS carbons of
acetone and thioacetone were significantly different as a result of the small bond angles preferred by divalent sulfur, and this
accounts for the much slower addition to a CS carbon vis-a-vis a CO group. Transition states for reduction of acetone and
̀
thioacetone by EtLi were similar, but the greater exothermicity of the reduction of the thioketone results in an earlier transition
state and lower activation energy for this process than that for the reduction of a ketone. The possible role of radical-mediated
processes in this chemistry is also discussed.
INTRODUCTION
ingly, the origin of the disparate behavior of thioketones when
treated with a main-group organometallic remains poorly
understood. Indeed, there are remarkably few reports of
investigations of the reactions of main-group organometallics
■
The addition of a main-group organometallic compound, such
as a Grignard reagent or an organolithium, to a carbonyl group
is arguably the most important method for formation of a
4
−11
1
with thioketones.
This state of affairs is attributable largely
carbon−carbon bond. This fundamental reaction involves
to the fact that aliphatic thioketones are prone to fairly rapid
addition of the organometallic to the electrophilic carbon of the
12
2
oligomerization, and those bearing α-hydrogens readily form
CO group which, in the case of an aldehyde or ketone,
13
the corresponding thioenols. Consequently, the majority of
the few mechanistic studies of the reaction of main-group
organometallics with thioketones involved investigations of the
behavior of aromatic substrates such as thiobenzophenone and
allows for the ready synthesis of secondary or tertiary alcohols.
In striking contrast to the reactions of main-group organo-
metallics with a carbonyl group, the seemingly analogous
reaction of these reagents with a thioketone does not lead to a
thiol by addition of the organometallic to the carbon of the
CS group; rather, the reaction of organolithiums or
alkylmagnesium compounds with thioketones usually proceeds
either by thiophilic addition to the sulfur atom of the CS unit
to give a sulfide or by reduction of the CS group to give a
5
,9,11
related species.
Several mechanistic scenarios, ranging from nucleophilic
5
addition to the sulfur atom of a CS group to various free-
9
,11
radical mediated processes,
have been proposed to account
for the outcome of the reactions of organolithiums or
alkylmagnesium compounds with thioketones. Nonetheless,
there is no consensus as to the etiology of the contrasting
behavior of ketones and thioketones in their reactions with
3
thiol.
The addition of main-group organometallics to aliphatic
ketones has been investigated in detail. Such reactions are
1
generally considered to be ionic processes although radical
intermediates, generated via SET, have been implicated in
reactions involving some aromatic ketones. Somewhat surpris-
Received: November 17, 2011
Published: January 13, 2012
©
2012 American Chemical Society
3199
dx.doi.org/10.1021/ja210847n | J. Am. Chem. Soc. 2012, 134, 3199−3207

Journal of the American Chemical Society
Article
main-group organometallics. In an effort to more fully
understand the strikingly different behavior of ketones, which
add organolithiums to give tertiary alcohols, and thioketones,
which are either reduced by organolithiums to afford secondary
thiols or undergo thiophilic addition to deliver sulfides, we have
investigated the reactions of a representative aliphatic
thioketone, adamantanethione (1), with a selection of organo-
lithiums and have explored both the addition reactions and the
comparison of its GC retention time and mass spectrum to
17
that of an authentic sample, and a novel product, 2-
adamantylthio-2-methylthioadamantane (3), along with recov-
ered 1. Significantly, no product attributable to attack at the
carbon atom of the CS group was detected in any of the
reactions. The results of these experiments are summarized in
Table 1.
The major product of the reaction of 1 with MeLi in either
Et O or THF as solvent was 2-adamantylthio-2-methylthioa-
2
14
reduction reactions using computational methods.
damantane (3). Indeed, 3 was virtually the exclusive product of
the reactions conducted in THF (Table 1, entries 3 and 4) and
an analytical sample of 3 was prepared in this way. The reaction
of MeLi with 1 in n-pentane was very slow; the majority of the
reaction mixture consisted of unreacted starting material after a
reaction time of 1 h (Table 1, entry 5). The formation of 3 is
reminiscent of the generation of analogous compounds in the
reactions of thiobenzophenone with main-group organo-
RESULTS AND DISCUSSION
■
Reactions of Adamantanethione with Organo-
lithiums. Adamantantanethione (1) was readily prepared, as
shown below, by treatment of adamantanone with phosphorus
pentasulfide in pyridine as a solvent following the method of
1
5
Greidanus. This solid, orange, nonenolizable aliphatic
thioketone is relatively stable but, as is the case with all
5
1
6
metallics. The mechanism illustrated in Scheme 1, analogous
thiones, it is oxidized by molecular oxygen to the ketone.
Precautions, such as storing 1 under an inert atmosphere and
away from light, were taken to avoid such oxidation. However, a
small amount (≤3%) of adamantanone was present in samples
of 1 that were used in the studies described below. We were
unaware of the rapidity of the oxidation at the inception of this
study; consequently, the adamantanethione (1) used in initial
studies contained more of the ketone (∼3%) than did samples
used in later studies (<∼2%).
5
to one proposed by Beak and Worley in a similar context,
involving thiophilic addition of MeLi to the sulfur atom of the
thione followed by a second addition of the same sort, nicely
18
accounts for the generation of 3.
The presence of small quantities (2−4%) of 2-adamantane-
thiol (2) as a product of the reactions of 1 with MeLi was
unexpected. Clearly, direct reduction of the thione by MeLi is
implausible: the formation of 2 is likely due to reaction of 3
with MeLi. A separate experiment, involving addition of MeLi
to a solution of 3 in Et O, revealed that MeLi reacts with 3,
2
albeit very slowly, to afford of 2. A plausible mechanism that
accounts for this reaction is illustrated in the Scheme 2.
The reactions of adamantanethione (1) with n-BuLi in Et O,
2
THF, or n-pentane solution were also explored, and the results
of these experiments are summarized in Table 2. Three
products, 2-adamantanethiol (2), n-butyl-2-adamantyl sulfide
The reactions of representative alkyllithiums (MeLi, n-BuLi,
and t-BuLi) with 1 were conducted under an atmosphere of
argon at −78 °C by dropwise addition of the alkyllithium to 0.1
(
4), and 2-adamanthythio-2-butylthioadamantane (5), ac-
M solutions of 1 in the appropriate solvent (Et O, THF, or n-
counted for essentially the total material balance. Sulfide 4
2
pentane). During the course of the reactions, typically 10 min
at −78 °C, the orange color of the thione solutions faded.
Reaction mixtures were then quenched with oxygen-free
MeOH, and the crude product mixtures were analyzed by
both capillary GC and by GC-MS, affording baseline separation
of the products.
was identified by comparison of its mass spectrum to that
1
0,19
reported for this material
and the structure of 5 was
assigned by analogy to that of compound 3 generated in the
reaction of 1 with MeLi.
The reactions of 1 with n-BuLi in n-pentane as solvent
(Table 2, entries 7 and 8) as well as those run in Et O solution
2
The reactions of adamantanethione (1) with MeLi afforded
small quantities of 2-adamantanethiol (2), identified by
(Table 2, entries 1−4) gave 2-adamantanethiol (2) as the major
product along with smaller amounts of n-butyl-2-adamantyl
Table 1. Reactions of Adamantanethione (1) with MeLi
products, ,
ab
%
c
entry
solvent
MeLi
time (min)
1
2
3
1
2
Et O
1.1
1.5
1.1
1.5
1.5
10
60
10
10
60
21.5
10.3
5.3
3.0
3.2
1.5
1.5
1.6
75.5
86.5
92.3
96.3
2
Et O
2
d
3
THF
THF
4
2.2
e
5
n-C H
91.5
5
12
a
b
Yields were determined by capillary GC analysis of reaction mixtures. A small quantity of adamantanone was detected due to inadvertent oxidation
c
d
of unreacted 1 in the GC samples prior to analysis. Molar equiv of MeLi. The reaction mixture contained 0.6% and 0.3% of two unidentified
compounds. The reaction mixture contained 6.6% of an unidentified compound.
e
3
200
dx.doi.org/10.1021/ja210847n | J. Am. Chem. Soc. 2012, 134, 3199−3207

Journal of the American Chemical Society
Article
Scheme 1
Scheme 2
state of t-BuLi in the various solvents: t-BuLi exists
25
22
predominantly as a dimer in Et O, a monomer in THF,
2
26
and a tetramer in pentane.
The results summarized above demonstrate that the
reactions of organolithiums with ketones and thioketones
proceed in fundamentally different ways. Whereas addition of
the organolithium to the carbon of the CO group is the
major product in the reaction with ketones, reduction or
thiophilic addition to the sulfur of the CS group is the
outcome of reactions with thiones. In an effort to understand
these contrasting behaviors, the addition reactions and the
reduction reactions were investigated using computational
methods.
Computational Results. At the outset of the investigation,
it was assumed that reactions of ketones and thioketones with
organolithiums could be modeled as ionic processes. The
possibility that reactions of thioketones are initiated by single-
electron transfer is considered in the sequel.
sulfide (4) and, in the case of reactions conducted in ether,
some 5. Compound 5, presumably generated by initial attack of
the organolithium on the sulfur atom of the CS group
followed by reaction of the sulfur-stabilized intermediate with 1
as found for the reactions of MeLi with the thione (Scheme 1),
was the major product from reactions in THF (Table 2, entries
5
and 6). The reduction of 1 to thiol 2 almost certainly involves
concomitant oxidation of n-BuLi to 1-butene via transfer of a β-
hydrogen from the organometallic to the CS group, a well
9
characterized reaction, but no effort was made to identify the
1-butene coproduct. The variation in the product ratios among
reactions run in various solvents is likely related to the degree
Addition Reactions. The addition reactions were modeled
using acetone and thioacetone as substrates and MeLi as the
reagent. These reactions most likely involve alkyllithium
of aggregation of n-BuLi in the three solvents that were
2
0
21
examined: n-BuLi is a tetramer in Et O, a temperature-
2
22
27
dependent equilibrium of the dimer and tetramer in THF,
aggregates. Although it seems very unlikely that monomeric
23
and a hexameric aggregate in hydrocarbons. It should be
noted that no product attributable to addition of n-BuLi to the
carbon of the CS unit was detected in any of these
MeLi could participate in these processes, we studied its
reactions in order to gain information on the relative reactivity
of ketones and thioketones. In the gas phase, the reactions of
polar molecules frequently involve formation of a prereaction
complex that is stabilized by electrostatic effects. Therefore, the
reactions were studied computationally in both the gas phase
2
4
experiments.
The outcome of reactions of adamantanethione (1) with t-
BuLi, summarized in Table 3, were similar to those of n-BuLi.
The major product observed under all reaction conditions
studied was 2-adamantanethiol (2); the balance of the product
mixtures consisted of sulfide 6 generated by thiophilic addition
to the sulfur atom of the CS group. The slight variation in
product ratios is attributable to dfferences in the aggregation
28
and in THF solution using the IEFPCM model in Gaussian.
The intermediate complexes for the reaction of MeLi
monomer with both acetone, to give the lithium salt of the
tertiary alcohol, and thioacetone, to give the corresponding salt
of the tertiary thiol, were readily located, and calculations were
Table 2. Reactions of Adamantanethione (1) with n-BuLi
products, ,
ab
%
c
entry
solvent
n-BuLi
time (min)
1
2
4
5
1
2
3
Et O
1.1
1.5
1.5
1.5
1.1
1.5
1.1
1.5
10
10
20
10
10
10
10
10
7.7
3.3
4.2
1.2
5.1
4.8
2.7
1.3
70.8
70.0
80.9
70.6
33.4
30.7
88.5
90.8
17.5
13.0
11.9
14.4
13.7
8.3
4.0
13.7
3.0
2
Et O
2
Et O
2
d
4
Et O
2
13.8
47.8
56.2
5
6
7
8
THF
THF
n-C H
8.8
5
12
n-C H
7.9
5
12
a
b
Yields were determined by capillary GC analysis of reaction mixtures. The reaction mixtures contained ∼1.5% of 2-n-butyl-2-adamantanol
c
d
generated by addition of n-BuLi to adamantanone present in the starting material. Molar equiv of n-BuLi. 1 was added to n-BuLi in Et O (reverse
2
addition).
3
201
dx.doi.org/10.1021/ja210847n | J. Am. Chem. Soc. 2012, 134, 3199−3207

Journal of the American Chemical Society
Article
Table 3. Reactions of Adamantanethione (1) with t-BuLi
a
products,
%
b
entry
solvent
t-BuLi
time (min)
1
2
6
c
1
Et O
1.1
1.5
2.0
1.1
1.5
1.1
1.5
10
10
10
10
10
10
9.4
5.1
65.3
77.3
74.5
63.8
69.0
49.7
45.0
24.1
17.6
20.0
20.2
16.3
32.8
32.4
2
2
3
4
5
Et O
2
Et O
2
5.4
THF
THF
16.0
14.7
10.9
d
6
n-C H
5
12
e
7
n-C H
10
10.8
5
12
a
b
c
Yields were determined by capillary GC analysis of reaction mixtures. Molar equiv of t-BuLi. The reaction mixture contained 1.3% of an
d
e
unidentified compound. The reaction mixture contained 6.4% of an unidentified compound. The reaction mixture contained 3.9% and 7.8% of two
unidentified compounds.
carried out at the B3LYP/6-311+G* level in which the distance
reoptimization at each step. It was found that the intermediate
complexes unfolded, leading to the six-membered ring
transition states that we previously found starting with a
different initial premise. These structures were optimized to
the final transitions states, and they were found to have just one
imaginary frequency. Using the frequency data determined at
the B3LYP/6-311+G* level for all structures, the energies were
corrected for the zero-point energies and for the changes in
energy on going from 0 K to 298 K. The relative energies of the
species involved in the reactions are summarized in Table 4; full
between the CH and the CO or CS carbon was reduced
3
stepwise with optimization at each step. The highest energy
species thus found were optimized to transition states both in
the gas phase and in THF solution. The activation free energies
14
⧧
⧧
were quite small (ΔG = 5.1 kcal/mol for acetone and ΔG =
4
.1 kcal/mol for thioacetone; see Supporting Information) with
⧧
thioacetone having the lower ΔG . Because MeLi in solution
reacted slowly with thioacetone and did not add to the carbon
of the CS (Table 1), the results of these calculations are not
in accord with the experimental observations, showing that the
reaction with monomeric MeLi is not a viable mode.
Table 4. B3LYP/6-311+G* Energies for the Addition of
a
Given this result, the reactions of acetone and thioacetone
with the dimer of MeLi were studied in a similar fashion
starting with the initial complexes illustrated in Figure 1. The
MeLi Dimer to Acetone and Thioacetone
gas phase
ΔH ΔG
0.0 0.0
THF solution
ΔH ΔG
0.0 0.0
compound
acetone reactants
acetone initial complex
acetone TS
−13.2
−4.2
−49.2
0.0
−5.4
6.8
−4.7
2.6
5.5
14.8
−29.7
0.0
acetone products
−36.8
0.0
−43.2
0.0
thioacetone reactants
thioacetone initial complex
thioacetone TS
−7.7
3.8
1.2
−0.6
7.5
9.3
16.1
−41.0
21.0
−39.2
thioacetone products
−53.3
−55.1
a
In kcal/mol; ΔH and ΔG values are corrected for both differences in
ZPE and the change in enthalpy on going from 0 K (corresponding to
the calculations) to 298 K.
data are available in the Supporting Information. In each case,
the transition states were followed in the forward direction and
were found to give the expected products.
The results of the calculations for the addition reactions of
dimeric MeLi with both acetone and thioacetone are shown
graphically in Figure 2. Addition to the carbonyl carbon of
acetone involves a pronounced initial complex in the gas phase
that disappears on going to THF solution; the initial complex
in the analogous reaction with thioacetone has an energy near
that of the reactants in the gas phase, but again, the energy
increases on going to THF as the solvent.
Figure 1. Initial complexes between dimeric MeLi and acetone (top
structure) and thioacetone (bottom structure).
Both of the addition reactions are found to be quite
exothermic; ΔG° = −30 and −39 kcal/mol for the acetone and
thioacetone reactions in THF, respectively: the computed
distance between one methyl group and the CO or CS
carbon of the substrate was reduced in 0.1 Å steps with
3
202
dx.doi.org/10.1021/ja210847n | J. Am. Chem. Soc. 2012, 134, 3199−3207

Journal of the American Chemical Society
Article
Figure 2. Comparison of the B3LYP/6-311+G* calculated free energy changes (ΔG) for reaction of dimeric MeLi with acetone (left graph) and
thioacetone (right graph) in both the gas phase (gp) and in THF solution. The points on the plots are the relative free energies of the reactants, the
initial complexes, and the products: the lines are drawn to connect these points and have no further significance. The intermediate complexes in
THF were stationary points, and they must be preceded by a small transition state. This is of no consequence for the reaction and was not studied.
energy changes during the course of the reactions are illustrated
in Figure 3. However, despite the fact that addition to
addition to the carbonyl carbon, thioketones do not readily add
main group organometallic reagents to the CS carbon.
What then is the etiology of this effect? It has long been
recognized that if two reactions proceed via a similar rate-
determining step, the more exothermic reaction generally has
29
the lower activation energy. This is clearly not the situation
for the reactions of MeLi with acetone and thioacetone (Figure
3
): the more exothermic addition, that involving thioacetone,
has the higher activation energy. The fact that the curves cross
in Figure 3 suggests that there is some fundamental difference
between the transition states for these two reactions; indeed,
the calculations find just such a difference. An examination of
the transition state structures for the additions to acetone and
thioacetone, illustrated in Figure 4, demonstrates that the
principal difference resides in the geometry about sulfur in the
thioacetone transition state. The addition of dimeric MeLi to
acetone leads to a transition state in which the carbonyl carbon
of the acetone fragment is well aligned to accept the methyl
group; the computed Li−O−C angle is 169.7°. However, in the
transition state for addition to the CS carbon of thioacetone,
the analogous fragment is rotated, making it difficult to
complete the transition; the Li−S−C angle is 110.2°. This
result is, at least in part, a consequence of the well-known
preference for divalent sulfur to favor relative small bond angles
Figure 3. Comparison of the B3LYP/6-311+G* calculated free energy
changes (ΔG) for reaction of dimeric MeLi with acetone (blue) and
thioacetone (red) in THF solution. The points on the plots are the
relative free energies of the reactants, the initial complexes, and the
products: the lines are drawn to connect these points and have no
further significance.
(
e.g., the H−S−H angle in H S is 92°) whereas oxygen prefers
2
larger angles (e.g., the H−O−H angle in H O is 105°).
2
Reduction Reactions. Reduction of the CS group to
give a thiol is the major outcome of reactions of
adamantanethione (1) with either n-BuLi (Table 2) or t-BuLi
thioacetone is a significantly more exothermic process than is
⧧
the addition to acetone, the activation free energy (ΔG ) for
(
Table 3). In an effort to understand the origin of the facile
⧧
the addition of dimeric MeLi in THF to acetone (ΔG = 15
reduction of thioketones, but not ketones, by main-group
organometallics having a β-hydrogen, the reactions were
modeled using acetone and thioacetone as substrates and
monomeric EtLi as the reagent, which delivers a hydrogen to
the CO or CS group with the formation of ethylene and
the lithium salt of the alkoxide or thiol, in a fashion similar to
that described above for the addition reactions. The EtLi was
complexed to the heteroatom of the substrate, and the distance
from the β-hydrogen of the reagent to the CO or CS
kcal/mol) is some 6 kcal/mol lower than that for the analogous
⧧
addition to the CS compound (ΔG = 21 kcal/mol). This 6
kcal/mol difference in activation free energy (Figure 3)
corresponds to a difference in relative rates for the additions
of approximately 25 000 at room temperature. The computa-
tional result is consistent with the experimental observations
presented above that, whereas addition of an organolithium to a
ketone gives a high yield of the tertiary alcohol expected from
3
203
dx.doi.org/10.1021/ja210847n | J. Am. Chem. Soc. 2012, 134, 3199−3207

Journal of the American Chemical Society
Article
Figure 5. B3LYP/6-311+G*-calculated transition state structures for
reduction of acetone (top structure) and thioacetone (bottom
structure) by EtLi in THF solution.
Figure 4. B3LYP/6-311+G*-calculated transition state structures for
addition of dimeric MeLi to acetone (top structure) and thioacetone
(
bottom structure) in THF solution.
carbon of the substrate was reduced in 0.1 Å steps followed by
geometry optimization at the HF/3-21G level. The highest
energy structures were then optimized to a transition state at
this level and in each case had one imaginary frequency. The
resulting structures were reoptimized at the B3LYP/6-311+G*
level in the gas phase and in THF as the solvent, frequencies
were determined at this level, and free energies at 298 K were
computed from these data. The calculated relative energies are
summarized in Table 5, and the transition state structures are
illustrated in Figure 5.
Table 5. B3LYP/6-311+G* Energies for the Reduction of
a
Acetone and Thioacetone by Monomeric EtLi
gas phase
ΔH ΔG
0.0 0.0
THF solution
ΔH ΔG
0.0 0.0
Figure 6. Comparison of the B3LYP/6-311+G*-calculated free energy
changes (ΔG) for reduction of acetone (blue) and thioacetone (red)
by EtLi in THF solution. The points on the plots are the relative free
energies of the reactants, the initial complexes, and the products: the
lines are drawn to connect these points and have no further
significance.
compound
acetone reactants
acetone initial complex
acetone TS
−18.9
−9.9
−26.8
0.0
−11.1
3.1
−5.0
4.3
3.2
16.9
−22.9
0.0
acetone products
−24.8
0.0
−21.5
0.0
thioacetone reactants
thioacetone initial complex
thioacetone TS
‡
−13.7
−9.7
−36.5
−3.7
3.5
−1.3
0.3
7.5
than does the less exothermic ketone reduction (ΔG ≈ 17
12.4
−40.6
kcal/mol). This difference in activation free energy corresponds
to a difference in relative rates for the reactions of
approximately 2000 at room temperature favoring reduction
of the thioketone, a result that is fully in harmony with the
experimental observations.
thioacetone products
−34.0
−42.1
a
In kcal/mol; see footnote in Table 4.
The computed free energy changes during the course of the
reductions in THF solution are shown graphically in Figure 6.
The transition state structures displayed in Figure 5 reflect
the energy differences illustrated in Figure 6: the reduction of
thioacetone displays a less advanced transition state than does
the reduction of acetone. The developing π-bond in the EtLi
fragment that transfers a hydrogen is significantly shorter in the
transition state for reduction of acetone (1.438 Å) than is the
same distance in the transition state for reduction of
Both reduction reactions are found to be exothermic; ΔG° =
−
23 and −41 kcal/mol for the acetone and thioacetone
reactions, respectively. In contrast to the addition reactions
discussed above, the reduction reactions conform to the Bell−
29
Evans−Polanyi principle: the more exothermic thioketone
‡
reduction has a lower activation energy (ΔG = 12.4 kcal/mol)
3
204
dx.doi.org/10.1021/ja210847n | J. Am. Chem. Soc. 2012, 134, 3199−3207

Journal of the American Chemical Society
Article
thioacetone (1.474 Å). Similarly, the distance between the
hydrogen being transferred and the β-carbon of the EtLi
fragment from which it departs is longer in the acetone
reduction (1.238 Å) than it is in the thioacetone reaction
Table 6. Energy Changes Attending Addition of a Lithium
Atom to Ketones and Thioketones
a
B3LYP/6-311+G*
CBS-QB3
ΔH
compound
acetone
ΔH
ΔG
ΔG
(
1.150 Å).
In the case of the addition reactions discussed above, it
−18.4
−16.0
−34.7
−30.0
−38.0
−46.3
−11.8
−8.4
−27.6
−22.6
−30.3
−37.6
−16.2
−10.4
−35.7
−30.3
−31.1
−45.2
−8.6
−3.9
−28.7
−22.5
−23.8
−38.7
became apparent that the smallest MeLi aggregate participating
in the reaction would be the dimer. In the case of reduction
using via hydrogen atom transfer from the β-carbon of EtLi,
there seems to be less of a structural requirement for a second
EtLi in forming the transition state. In order to check this, we
have carried out a similar study using the dimer of EtLi as the
reactant. The transition states thus located (see Supporting
Information) had the second EtLi appended in a nonfunctional
manner to the transition states shown in Figure 4 for reduction
via the EtLi monomer; thioacetone was again calculated to be
more reactive than acetone.
adamantanone
benzophenone
thioacetone
adamantanethione
thiobenzophenone
a
In kcal/mol at 298 K.
none is the most exothermic among the ketones. Undoubtedly,
this result is a consequence of resonance stabilization of the
diaryl ketyls produced in these reactions. The substantial
difference between the exothermicity of the addition of a
In short, these computational results provide a reasonable
explanation of the difference in the outcome of reaction of
alkyllithiums with acetone and thioacetone, and by extension,
they elucidate the origin of the disparate behavior of other
ketones and thiones in reactions with main-group organo-
metallics.
̀
lithium atom to acetone or adamantanone vis-a-vis benzophe-
none makes it very reasonable to conclude that dialkyl ketones
react with main-group organometallics via ionic mechanisms
whereas benzophenone, and similar substrates, may react via
SET-initiated processes. The distinction among the thioketones
is not as obvious as is the case with the ketones. However, the
CBS-QB3 results indicate that reaction of thiobenzophenone
with a lithium atom is some 15−16 kcal/mol more exothermic
than the analogous reactions with acetone or adamantane-
thione. This result is consistent with the observation that
reaction of diaryl thiones with main-group organometallics may
be initiated by SET while dialkyl substrates likely react via ionic
mechanisms as suggested by the results presented above.
Radical-Mediated Reactions. Thus far, only ionic
processes have been considered in the computational
investigation of the reactions of ketones and thioketones with
alkyllithiums. However, there is good evidence that the
reactions of thiobenzophenone and related substrates with
11
alkyllithiums involve radical intermediates generated via SET.
Could such a mechanism also be involved in reactions of alkyl-
substituted ketones?
It is difficult to construct a complete model that might be
investigated computationally for such an SET-initiated process.
Presumably, a reaction of this sort is initiated by transfer of a
lithium atom from an organolithium aggregate to the thione
substrate to generate a lithium ketyl and an electron-deficient
aggregate residue. In order to simplify calculation of the relative
energy changes that might be involved in the generation of the
lithium ketyl, we have modeled this initial step by examining
the addition of a lithium atom to ketone and thioketone
substrates. Nonetheless, it must be remembered that the
removal of a lithium atom from an alkyllithium aggregate is an
endothermic process. Notwithstanding this caveat, the energy
required to remove the lithium atom from a given aggregate
will be a constant because the aggregate less one lithium atom
might reasonably be independent of the carbonyl or
thiocarbonyl compound to which it is transferred. In short,
the relative energy changes accompanying addition of a lithium
atom to ketones or thioketones provide the information needed
to evaluate the probability of SET-initiated reactions of such
substrates.
The energy changes attending addition of a lithium atom to
representative ketones and thioketones to give the lithium ketyl
were calculated at two theoretical levels in order to ensure that
the results are reliable. They include the model chemistry CBS-
QB3 and B3LYP/6-311+G* level, giving the results summar-
ized in Table 6. For the gas phase, the CBS-QB3 results should
be the more reliable; in any case, the ΔH values from the CBS-
QB3 and B3LYP calculations are linearly related. All of these
reactions are found to be quite exothermic (Table 6).
Not surprisingly, addition of a lithium atom to thiobenzo-
phenone is the most exothermic reaction among the
thioketones, and the corresponding reaction with benzophe-
CONCLUSIONS
■
The outcome of the reactions of alkyllithium reagents with
adamantanethione (1) in a variety of solvents, detailed in
Tables 1−3, represents the first comprehensive investigation, of
which we are aware, involving organolithiums and a
representative alkyl-substituted thione. The results presented
above clearly demonstrate that ketones and thioketones react
with organolithiums in fundamentally different ways: sterically
unencumbered alkyllithiums react rapidly with ketones via
addition to the carbonyl group; thioketones, in contrast, are
either reduced upon reaction with simple alkyllithiums or
undergo thiophilic addition of the alkyl group to the sulfur
atom of the CS unit. Indeed, no products attributable to
addition of an alkyllithium to the carbon atom of the CS
group in 1 was observed in any of the experiments.
The addition reactions and the reduction reactions of
ketones vis-a-vis thioketones with alkylithiums were modeled
̀
computationally as ionic processes using acetone and
thioacetone as substrates. The results of these studies, which
are in full agreement with the experimental results, indicate the
following: (1) the failure to observe addition of an alkyllithium
to the carbon atom of the CS group is likely the consequence
⧧
of a relatively high ΔG for the process caused by a transition
state geometry that makes it difficult to transfer the alkyl group
to the thiocarbonyl carbon; (2) the reduction of thioketones on
reaction with alkyllithiums bearing a β-hydrogen is more facile
than is reduction of a carbonyl group, and a six-membered
transition state accounts nicely for the difference in rate of
reduction of ketones and thioketones; (3) calculations that
modeled SET-initiated reactions by investigation of the reaction
of a lithium atom with representative ketones and thioketones
3
205
dx.doi.org/10.1021/ja210847n | J. Am. Chem. Soc. 2012, 134, 3199−3207

Journal of the American Chemical Society
Article
suggest that, whereas the reactions of thiobenzophenone and
related substrates with an alkyllithium may be initiated by SET,
the analogous reactions of dialkyl thiones, such as 1, likely
involve ionic processes.
ASSOCIATED CONTENT
Supporting Information
Complete ref 28, a summary of the calculations, including
computed energies, imaginary frequencies for transition states,
free energies in THF solution, and computed atomic
coordinates for transition states. This material is available free
of charge via the Internet at http://pubs.acs.org.
■
*
S
EXPERIMENTAL SECTION
■
General Procedures. Reactions involving alkyllithiums were
performed in flame-dried glassware under an atmosphere of argon.
Alkene-free n-pentane was purified by repeated washings of
commercial technical grade pentane with concentrated sulfuric acid
until the acid layer remained colorless, followed by washing with water
and saturated sodium bicarbonate, drying over magnesium sulfate, and
AUTHOR INFORMATION
■
Corresponding Author
william.bailey@uconn.edu; kenneth.wiberg@yale.edu
distillation from sodium/benzophenone/tetraethylene glycol. Et O
2
Notes
and THF were freshly distilled from sodium/benzophenone.
Commercial alkyllithium solutions were titrated prior to use with a
standard solution of sec-butanol in xylene using 1,10-phenanthroline as
The authors declare no competing financial interest.
3
0
ACKNOWLEDGMENTS
the indicator, as described by Watson and Eastham. Literature
procedures were followed for the preparation of adamantanethione
■
We are grateful to Dr. Christopher Wolterman of FMC,
Lithium Division, for generous gifts of the alkyllithiums used in
this investigation. The work at UCONN was supported by a
grant from Procter & Gamble Pharmaceuticals, Mason, OH.
15
17
(
1) and 2-adamantanethiol (2).
Product mixtures (Tables 1-3) were analyzed by GC on a 15-m ×
.25-mm × 1 μm MXT-200 capillary column using temperature
0
programming (35 °C for 10 min, increasing 5 °C/min to 240 °C for
2
5 min) and by GC-MS on a 25-m × 0.2-mm × 0.33-μm HP-5
REFERENCES
capillary column using temperature programming (100 °C for 2 min,
increasing 10 °C/min to 250 °C for 20 min).
■
(1) (a) Smith, M. B.; March, J. March’s Advanced Organic Chemistry,
6th ed.; Wiley & Sons: New York, 2007; pp 1300−1309. (b) Carey, F.
A.; Sundberg, R. J. Advanced Organic Chemistry Part B: Reactions and
Synthesis, 5th ed.; Springer: New York, 2007, pp 619−667.
(c) Kharasch, M. S.; Reinmuth, O. Grignard Reactions of Nonmetallic
Substances; Prentice Hall: New York, 1954.
(2) Note that reduction of the carbonyl group is a well documented
competing reaction particularly when sterically hindered ketones and/
or sterically congested organometallics are used (ref 1).
General Procedure for the Reactions of Adamantanethione
(1) with Organolithiums. Approximately 0.1 M solutions of
adamantanethione (1) in the appropriate dry and oxygen-free solvent
Et O, THF, or n-pentane; Tables 1−3) were cooled to −78 °C under
(
2
an atmosphere of argon, 1.1 or 1.5 molar equiv of an organolithium
solution (MeLi in Et O, n-BuLi in hexanes or t-BuLi in n-pentane) was
2
added dropwise, and the resulting solutions were stirred for a period of
time at −78 °C (specific conditions are given in Tables 1−3), during
which time the orange color of the thione faded. The reactions were
then quenched with methanol, the cooling bath was removed, and the
reaction mixtures were allowed to warm to room temperature. The
resulting solutions were washed with water and brine, dried (MgSO4),
and analyzed by GC and by GC-MS. In most cases, the products were
identified by comparison of their GC retention times and mass spectra
to those of authentic samples or by comparison of their mass spectra
to those reported in the literature.
(
3) (a) Wakefield, B. J. Organolithium Methods; Academic Press: New
York, 1988; pp 101−105. (b) Wakefield, B. J. Organomagnesium
Methods in Organic Synthesis; Academic Press: New York, 1995; pp
209−210. (c) Duus, F. In Comprehensive Organic Chemistry; Jones, D.
N., Ed.; Pergamon Press: Oxford, 1979; Vol. 3.
(4) Schonberg, A.; Rosenbach, A.; Schutz, O. Justus Liebigs Ann.
̈
Chem. 1927, 37, 454.
1
0,19
The structure of 3 obtained in
(5) (a) Beak, P.; Worley, J. W. J. Am. Chem. Soc. 1970, 92, 4142.
(b) Beak, P.; Worley, J. W. J. Am. Chem. Soc. 1972, 94, 597.
(6) (a) Paquer, D.; Vialle, J. C. R. Acad. Sci. 1972, 275, 589.
(b) Paquer, D. Bull. Soc. Chim. Fr. 1975, 1439.
reactions of 1 with MeLi (Table 1) was assigned on the basis of the
spectroscopic data obtained from an analytical sample prepared as
described below; the structure of 5 generated in reactions of 1 with n-
BuLi was assigned by analogy to that of 3.
(7) (a) Dagonneau, M.; Vialle, J. Tetrahedron Lett. 1973, 32, 3017.
(b) Dagonneau, M. J. Organomet. Chem. 1974, 80, 1.
(8) Paquer, D.; Vazeux, M. J. Organomet. Chem. 1977, 140, 257.
(9) (a) Ohno, A.; Nakamura, K.; Uohama, M.; Oka, S. Chem. Lett.
1975, 983. (b) Ohno, A.; Nakamura, K.; Uohama, M.; Oka, S.;
Yamabe, T.; Nagata, S. Bull. Chem. Soc. Jpn. 1975, 48, 3718.
(10) Rautenstrauch, V. Helv. Chim. Acta 1974, 57, 496.
(11) Alberti, A.; Benaglia, M.; Macciantelli, D.; Marcaccio, M.;
Olmeda, A.; Pedulli, G. F.; Roffia, S. J. Org. Chem. 1997, 62, 6309 and
references therein.
(12) For a review, see: Mayer, R.; Morgenstern, J.; Fabian, J. Angew.
Chem., Int. Ed. Engl. 1964, 3, 277.
(13) (a) Paquer, D.; Vialle, J. Bull. Soc. Chim. Fr. 1969, 3595.
(b) Lipscomb, R. D.; Sharkey, W. H. J. Polym. Sci., Part A: Polym.
Chem. 1970, 8, 2187.
(14) A preliminary account of a portion of this investigation has
appeared, see: Wiberg, K. B.; Bartelson, A. L.; Bailey, W. F.
Tetrahedron Lett. 2011, 52, 2169.
(15) Greidanus, J. W. Can. J. Chem. 1970, 48, 3530.
(16) (a) Gattermann, L.; Schulze, H. Chem. Ber. 1896, 29, 2944.
̈
(b) Schonberg, A.; Mustafa, A. J. Chem. Soc. 1943, 275. (c) Pushkara
Rao, V.; Ramamurthy, V. Tetrahedron 1985, 41, 2169 and references
therein.
2-Adamantylthio-2-methylthioadamantane (3). A solution of
833 mg (5.01 mmol) of adamantanethione (1) in 50 mL of dry,
oxygen-free Et O was cooled to −78 °C under an atmosphere of
argon, and 5.78 mL of a 1.57 M solution of MeLi (9.07 mmol) in Et O
2
2
was added dropwise. The solution was stirred for 4 h at −78 °C,
during which time the orange color of the thione faded, and 1 mL of
methanol was then added to quench the reaction. The cooling bath
was removed, the reaction mixture was stirred for an additional 15 min
and then washed with water, brine, and 5% aqueous NaOH solution.
The organic layer was dried (MgSO ) and concentrated to afford a
4
white paste that was recrystallized from ethanol−water to yield 423 mg
1
(48%) of the title compound as a white solid: mp 87−92 °C; H NMR
(CDCl , 400 MHz) 3.26 (broad s, 1 H), 2.54 (t, 4 H), 2.11 (broad s, 1
3
H), 2.08 (broad s, 1 H), 2.00 (s, 3 H), 1.89−1.84 (m, 14 H), 1.73
(
broad s, 2 H), 1.70 (broad s, 2 H), 1.63 (broad s, 2 H), 1.60 (broad s,
2
3
H); 13C NMR (CDCl , 100 MHz) δ 71.7, 51.2, 39.4, 39.2, 37.8, 36.1,
3
5.8, 33.9, 33.8 (two carbons), 27.9, 27.6 (two carbons), 27.4, 11.2;
+
+
GC-MS m/z (% relative intensity) 348 (1, M ), 301 (78, M − SCH ),
3
+
1
81 (100, M − SC H ), 167 (14, SC H ), 135 (68, C H ), 91
10
15
10 15
10 15
(
30), 79 (34), 67 (30); HRMS-ESI (m/z): [M − SCH ] calcd for
3
C H S, 301.1990; found, 301.2005.
20
29
Calculations. All of the ab initio calculations were carried out using
28
Gaussian-09.
(17) Greidanus, J. W. Can. J. Chem. 1970, 48, 3593.
3
206
dx.doi.org/10.1021/ja210847n | J. Am. Chem. Soc. 2012, 134, 3199−3207

Journal of the American Chemical Society
Article
(
18) In this connection, it might be noted that Paquer and Vazeux
ref 8) have reported that 2-methylthioadamantane is generated in
0% yield when 1.25 molar equiv of MeLi is added to an ethereal
(
7
solution of 1 at 25 C; they did not report the formation of 3. It may be
that the room temperature reaction leads to fairly rapid proton
abstraction from the Et O solvent by the sulfur-stabilized organo-
2
lithium formed by thiophilic addition to 1.
(
19) Law., K. Y.; de Mayo, P. J. Am. Chem. Soc. 1979, 101, 2351.
(
20) Gessner, V. H.; Das
̈
chlein, C.; Strohmann, C. Chem.Eur. J.
2
009, 15, 3320.
(
21) West, P.; Waack, R. J. Am. Chem. Soc. 1967, 89, 4395.
(
22) (a) McGarrity, J. F.; Ogle, C. J. Am. Chem. Soc. 1985, 107, 1805.
(
b) Bauer, W.; Winchester, W. R.; Schleyer, P. v. R. Organometallics
987, 6, 2371.
23) (a) Brown, T. L. Acc. Chem. Res. 1968, 1, 23. (b) Bywater, S. J.
Phys. Chem. 1975, 79, 2148.
24) In our preliminary report (ref 14), the GC peak corresponding
1
(
(
to the small quantity of 2-n-butyl-2-adamantanol present in the
product mixtures from reactions of n-BuLi with 1 (Table 2, footnote
b) was incorrectly attributed to 2-butyl-2-adamantanethiol.
(
25) (a) Bates, T. F.; Clarke, M. T.; Thomas, R. D. J. Am. Chem. Soc.
1
1
988, 110, 5109. (b) Kottke, T.; Stalke, D. Angew. Chem., Int. Ed. Engl.
993, 32, 580.
(
26) Thomas, R. D.; Clarke, M. T.; Jensen, R. M.; Young, T. C.
Organometallics 1985, 5, 1851.
27) Sapse, A. M.; Schleyer, P. v. R. Lithium Chemistry: a Theoretical
and Experimental Overview; Wiley: New York, 1995.
28) Frisch, M. J., et al. Gaussian-09; Gaussian, Inc., Pittsburgh PA,
998.
29) (a) Carroll, F. A. Perspectives on Structure and Mechanism in
Organic Chemistry, 2nd ed.; Wiley & Sons: New York, 2010; pp 360−
70. (b) Evans, M. G.; Polanyi, M. Trans. Faraday Soc. 1938, 34, 1.
c) Bell, R. P. Proc. R. Soc. London, Ser. A 1936, 154, 414.
30) Watson, S. C.; Eastham, J. F. J. Organomet. Chem. 1967, 9, 3210.
(
(
1
(
3
(
(
3
207
dx.doi.org/10.1021/ja210847n | J. Am. Chem. Soc. 2012, 134, 3199−3207