Reductive alkylation of thioamides with Grignard reagents in the presence of Ti(O i Pr)4: Insight and extension

Article abstract of DOI:10.1021/om500603v

The reductive alkylation of thioamides by Grignard reagents in the presence of Ti(OiPr)₄ is the subject of a study involving 20 different substrates. The influence of various parameters has been evaluated, showing notably that the yields of this moderately efficient process can be improved in several cases by applying a slow addition of the Grignard reagent. The results presented in this contribution also provide new insight into the reactivity of the proposed key intermediates, namely, a metalated iminium species and, ultimately, an α-metalated amine. Interestingly, by control of the temperature and the amount of Grignard reagent engaged, the reaction can be directed toward the selective formation of the former titanium intermediate complex. This represents an extension of the original method, allowing the synthesis of various previously inaccessible substituted amines by subsequent addition of a nucleophilic reagent. This role can be played not only by organomagnesium compounds but also by alkyllithium reagents, alkyltitanium(IV) complexes, and lithium aluminum hydride. The properties of the α-metalated amine final intermediate have also been explored, demonstrating that this complex is a poor nucleophile but can act as a radical precursor, which is especially evidenced when the resulting radical species are stabilized. Overall, this chemistry thus proves unexpectedly rich and can plausibly lay the basis for the development of new applications in the future.

Full text of DOI:10.1021/om500603v

Article
pubs.acs.org/Organometallics
Reductive Alkylation of Thioamides with Grignard Reagents in the
Presence of Ti(OiPr)4: Insight and Extension
Fabien Hermant,† Ewelina Urbanska,†,‡ Sarah Seizilles de Mazancourt,† Thomas Maubert,†
́
Emmanuel Nicolas,§ and Yvan Six*,†
†Laboratoire de Synthes
‡Faculty of Chemistry, Nicolaus Copernicus University, Gagarina 7, 87-100 Torun
§Laboratoire de Chimie Molec
ulaire (LCM), UMR 9168 CNRS/Ecole Polytechnique, 91128 Palaiseau Cedex, France
̀
e Organique (DCSO), UMR 7652 CNRS/Ecole Polytechnique, 91128 Palaiseau Cedex, France
, Poland
́
́
S
* Supporting Information
ABSTRACT: The reductive alkylation of thioamides by Grignard
reagents in the presence of Ti(OiPr)4 is the subject of a study
involving 20 different substrates. The influence of various
parameters has been evaluated, showing notably that the yields
of this moderately efficient process can be improved in several
cases by applying a slow addition of the Grignard reagent. The
results presented in this contribution also provide new insight into
the reactivity of the proposed key intermediates, namely, a
metalated iminium species and, ultimately, an α-metalated amine.
Interestingly, by control of the temperature and the amount of
Grignard reagent engaged, the reaction can be directed toward the
selective formation of the former titanium intermediate complex.
This represents an extension of the original method, allowing the
synthesis of various previously inaccessible substituted amines by subsequent addition of a nucleophilic reagent. This role can be
played not only by organomagnesium compounds but also by alkyllithium reagents, alkyltitanium(IV) complexes, and lithium
aluminum hydride. The properties of the α-metalated amine final intermediate have also been explored, demonstrating that this
complex is a poor nucleophile but can act as a radical precursor, which is especially evidenced when the resulting radical species
are stabilized. Overall, this chemistry thus proves unexpectedly rich and can plausibly lay the basis for the development of new
applications in the future.
Scheme 1. Generation of Titanacyclopropane Species from
Ti(OiPr)4 and Grignard Reagents with β Hydrogen Atoms
INTRODUCTION
■
The development of new synthetic transformations based on
the combined use of Ti(OiPr)4 (or sometimes another
alkoxytitanium(IV) species) and Grignard reagents has
emerged from the pioneering work of the group of Kulinkovich
at the end of the 1980s/beginning of the 1990s.1 These new
methods now constitute a whole family of fairly general
reactions,2 the most prominent of which are the cyclo-
propanation of esters (the Kulinkovich reaction),1,2 the
synthesis of aminocyclopropanes from amides (the de Meijere
variation)3 or from nitriles (the Bertus−Szymoniak variation),4
and the functionalization of alkynes via titanacyclopropene
species (the Sato method).5 Typically, all of these methods use
Grignard reagents having hydrogen atoms at the β position.
They are thought to involve the initial formation of a
titanacyclopropane complex A (Scheme 1). In particular, both
nPrMgX and iPrMgX are assumed to give A1, while cC5H9MgX
and cC6H11MgX are precursors of A2 and A3, respectively.
In this context, we recently began to study the trans-
formations of thioamides 1 under Kulinkovich-type conditions
and found that they typically undergo a reductive alkylation
process (Scheme 2).6 To explain this transformation, we put
forward a ligand exchange reaction of the intermediate A with
the thiocarbonyl group, generating a thiatitanacyclopropane
complex B. This species could be in equilibrium with a
metalated iminium complex C. Attack of the Grignard reagent,
Special Issue: Catalytic and Organometallic Chemistry of Earth-
Abundant Metals
Received: June 6, 2014
Published: July 28, 2014
© 2014 American Chemical Society
5643
dx.doi.org/10.1021/om500603v | Organometallics 2014, 33, 5643−5653

Organometallics
Article
Scheme 2. Reductive Alkylation of Thioamides by Grignard
Reagents in the Presence of Ti(OiPr)4: General Equation
and Proposed Mechanism
used in excess, would then form the organotitanium species D,
protonolysis of which would account for the production of the
reductive alkylation products 2.
Thioamides fitted with a terminal alkene function separated
from the nitrogen atom by a two-carbon chain represent a
special category of substrates that preferentially undergo
intramolecular aminocyclopropanation rather than reductive
alkylation (Scheme 3). This process has been the subject of a
Scheme 3. Intramolecular Cyclopropanation of Thioamides
Fitted with a Suitable ω-Alkenyl Group at the Nitrogen
Atom
Figure 1. Thioamide substrates 1a−t.
Choice of the Solvent. Overall, ether solvents such as
tBuOMe, THF, Et2O, and cC5H9OMe can be employed, with
some variations in the results (see Table 5 for results in the
latter two solvents). One example run in cyclohexane gave a
significantly poorer result than in tBuOMe or THF (synthesis
of 2eb, Table 1). CH2Cl2 is not a suitable solvent (synthesis of
2oc, Table 3).
Substrate Scope. The yields obtained from non-alkenyl
acyclic substrates are typically moderate, and thioformamides
generally tend to perform somewhat better than thioacetamides
(Table 1, 2ab vs 2eb, 2bb vs 2gc). Secondary thioamides were
expected to be unsuitable substrates because of the reactivity of
the N−H bond. This point was verified by the reductive
alkylation of N-benzylthioacetamide (1d), which proceeded in
10% yield only. As mentioned in the Introduction, thioamides
fitted with a CHCH2 olefin function separated from the
nitrogen atom by a two-carbon chain undergo intramolecular
cyclopropanation via the intermediate complex E. With an
additional substitution grafted onto the alkene group, the yield
of this process typically drops to 5−10%. The reductive
alkylation products are then obtained in around 30% yield from
the 1,2-disubstituted alkene substrates 1i and 1j but in only 6%
yield from the 1,1-disubstituted alkene 1k (Table 2).
Satisfactory yields can be restored by increasing the length of
the spacer, even in the case of monosubstituted olefins. This is
illustrated by the conversion of the N-hex-5-enyl substrate 1m
into the tertiary amine 2mb, which proceeds in 48% yield. In
the case of thiolactams, good results can be achieved, especially
in the thiopyrrolidine series 1n−q, with yields of up to 81%
(Table 3). The reaction appears to be significantly less efficient
with six-membered-ring thiolactams (2sb vs 2ob). It is
interesting to note that starting from the β-lactam 1t, the
corresponding azetidine 2tb can be obtained, albeit in only 28%
yield.
separate study.7 It is proposed that the intermediate E is formed
in this case, possibly via B. Ring opening to afford a zwitterionic
complex F followed by the formation of the three-membered
ring according to an SE2(back) process would give the
aminocyclopropane product. This mechanism is thus analogous
to the accepted pathway for the intramolecular Kulinkovich−de
Meijere reaction of the corresponding carboxylic amide
substrates.3
The present contribution focuses on the reductive alkylation
process and reports the results of new studies aimed at
supporting or challenging the mechanism displayed in Scheme
2 as well as at extending the possibilities offered by this
chemistry.
RESULTS AND DISCUSSION
■
Thioamide Substrates. The structures of the thioamides
used in this work are presented in Figure 1. These substrates
include nonchiral and chiral compounds, acyclic thioamides,
and thiolactams with various ring sizes.
Results under the Initially Established “Standard”
Conditions. The results of the transformations of these
thioamides 1a−t under the initially established standard
conditions, typically involving the use of 1.5 equiv of Ti(OiPr)4
and 4.0 equiv of Grignard reagent, are presented in Tables 1, 2,
and 3. Several aspects are worthy of comment.
Effect of a Coordinating Group. The presence, at the
nitrogen atom, of a substituent fitted with an ether group that is
able to coordinate to the titanium center was anticipated to
favor the ligand exchange process generating the intermediate B
5644
dx.doi.org/10.1021/om500603v | Organometallics 2014, 33, 5643−5653

Organometallics
Article
a b
,
Table 1. Reductive Alkylation of Acyclic Thioamides 1a−h by Grignard Reagents in the Presence of Ti(OiPr)4
a
b
Reactions were typically run using 1.5 equiv of Ti(OiPr)4 and 4.0−4.5 equiv of Grignard reagent added over 5−30 min. Yields were estimated by
c
NMR analysis of the crude products. Yields in parentheses are for the isolated products. A dimeric byproduct 3cb was formed in 30−36% yield. See
the text for details. A secondary amine 4h formally resulting from hydrolysis of the thioamide was produced in 58% yield.
d
ab
,
Table 2. Reductive Alkylation of N-Alkenylthioamides 1i−m by Grignard Reagents in the Presence of Ti(OiPr)4
a
b
Reactions were typically run using 1.5 equiv of Ti(OiPr)4 and 4.0 equiv of Grignard reagent added over 3−8 min. Yields were estimated by NMR
c
analysis of the crude products. Yields in parentheses are for the isolated products. A secondary amine 4i formally resulting from loss of the
thiocarbonyl group was produced in 48% yield. The secondary amine byproduct 4k was produced in 36% yield. The secondary amine byproduct 4l
was produced in 33% yield in tBuOMe and 20% yield in THF. 40% unreacted starting material was observed in the crude product.
d
e
f
(see Scheme 2). Indeed, this idea was implemented successfully
in related reactions involving ligand exchange with olefins.8 In
the present transformation, the facilitating effect is not so clear
and moderate at best. For instance, with cyclohexylmagnesium
chloride in tBuOMe, the ether compounds 1f and 1m are
converted into the amines 2fb and 2mb in 36 and 42% isolated
yield, respectively, which is to be compared with the reaction of
thioamide 1e devoid of any ether function, which proceeds in
slightly lower yield (30%) under the same conditions. The
same reductive alkylation of the much bulkier thioamide 1h
proceeds in only 24% yield in spite of the presence of an ether
group. In the thiolactam series, the best yields are achieved with
the benzyloxy substrate 1o, but 1n, 1q, and 1r perform better
than the ethers 1p and 1s.
Choice of the Grignard Reagent. First, it is worth noting
that starting from thioamide 1a in THF, essentially no reaction
takes place when the Grignard reagent engaged is MeMgBr; no
trace of the reductive alkylation product 2ae is observed. Since
only Grignard reagents with β-hydrogen atoms can lead to the
generation of titanacyclopropane species A, this result strongly
supports the essential role of this type of intermediate complex
in the reductive alkylation process (Scheme 2). The other
Grignard reagents employed (cyclopentylmagnesium chloride,
cyclohexylmagnesium chloride, isopropylmagnesium chloride,
and n-propylmagnesium chloride) all lead to the corresponding
tertiary amine products 2.
Diastereoselectivity. The presence of a stereogenic center
directly attached to the nitrogen atom of the thioamide
substrate 1 does not generally induce high levels of
5645
dx.doi.org/10.1021/om500603v | Organometallics 2014, 33, 5643−5653

Organometallics
Article
a b
,
Table 3. Reductive Alkylation of Thiolactams 1n−t by Grignard Reagents in the Presence of Ti(OiPr)4
a
b
Reactions were typically run using 1.5 equiv of Ti(OiPr)4 and 4.0−4.5 equiv of Grignard reagent added over 2−30 min. Yields were estimated by
c
NMR analysis of the crude products. Yields in parentheses are for the isolated products. Only the structure of the major diastereoisomer is
displayed. This product was not isolated because it underwent a rapid cyclization by an intramolecular SN2 reaction (see Scheme 5). Yield after an
additional step consisting in an intramolecular SN2 reaction (see Scheme 5).
d
e
diastereoselectivity. In the cases of the chiral pyrrolidine-2-
thiones 1o and 1p and of the chiral piperidine-2-thione 1s, the
introduction of the alkyl group proceeds with a moderate
preference for the formation of the cis diastereoisomers (Table
3).9 Overall, the selectivities observed range from 50:50 to
72:28, the latter result being obtained in the synthesis of 2oc in
THF.
Scheme 4. Possible Competitive Pathways Leading to the
Formation of the Amine Byproducts 4
Competitive Processes. The undesired secondary amines 4,
formally resulting from loss of the thiocarbonyl group, are
systematically observed coproducts. In cases where the
reductive alkylation reaction affords a poor yield, 4 may be
produced in extensive amounts and even become the major
product. For instance, starting from the “bad” substrates 1g−l,
the corresponding amines 4h−l are produced in 20−58% yield.
A straightforward explanation for the formation of these
undesired compounds is a direct attack of the Grignard reagent
at the carbon atom of the thiocarbonyl group. This is supported
by the presence of the methyl ketone 5 resulting from
hydrolysis of the tetrahedral intermediate G (Scheme 4), and
the stench detected during workup of the reactions is consistent
with the production of H2S. However, experimental evidence
suggests that this is not the only pathway leading to 4, as the
observed amount of 5 is uniformly lower than the quantity of 4.
For instance, while 4h is produced in 58% yield from 1h using
cyclohexylmagnesium chloride in tBuOMe, only a 15% yield of
cyclohexyl methyl ketone is detected in the crude product.
Moreover, as already mentioned, the thioamide 1a is left
essentially unreacted when MeMgBr is used as the Grignard
reagent under the standard conditions in THF. Only a 3% yield
of amine 4a is produced in this reaction.
Another possibility resides in the aerobic oxidation of the
final intermediate D at the stage of the workup, instead of
protonolysis. This would lead to the peroxide species I via a
radical mechanism. Such oxidations of C(sp3)−Ti bonds into
C−O bonds have been reported,10 and in the present study the
formation of dimers in some cases is consistent with the
generation of a radical H at the α position relative to the
nitrogen atom (vide infra).11 The production of 4 by final
hydrolysis of remaining unreacted intermediate B or C was also
envisaged, but this hypothesis was invalidated by a comple-
mentary experiment (see Table 7, entry 10).
Rotamers. Thioamides are known to have a larger rotational
barrier than the corresponding carboxylic amides.12 Two
batches of crystalline thioformamide 1b were synthesized:
5646
dx.doi.org/10.1021/om500603v | Organometallics 2014, 33, 5643−5653

Organometallics
Article
Scheme 5. Reactions of N-(4-Chlorobutyl)thiolactam 1r
Scheme 6. Reaction of 1b Using Cyclohexylmagnesium Chloride with a D2O Quench at Low Temperature
1
Figure 2. H NMR spectra of (top) 2bb and (bottom) 2bb-d. The presence of two diastereoisomers of the latter compound, in approximately
equimolar amounts, is clearly shown by the splitting of the quartet signal (CH, gold color) and of the AB system (CH2, green color) as well as by the
two distinct CH doublet signals (salmon-pink color) instead of the AB part of an ABX system in the case of the nondeuterated molecule.
one consisting of a 72:28 mixture of rotamers and the other
containing a single rotamer. Having verified that the isomer-
ization of these rotamers is slow in solution,13 we compared the
results of reactions performed on the two batches. The
difference of behavior was found to be minor. The reductive
alkylation process is perhaps slightly more efficient starting
from the single rotamer, with somewhat less dibenzylamine
(4b) being produced.9
Special Case of the N-(4-Chlorobutyl) Substrate 1r. When
the standard protocol is used with isopropylmagnesium
chloride, the product isolated from thiolactam 1r is not the
corresponding tertiary amine 2rc but instead is the spirocyclic
quaternary ammonium salt 6rc (Scheme 5). The experimental
evidence indicates that this molecule is formed by an
intramolecular SN2 reaction of 2rc after the workup of the
reaction, at the stage of the concentration of the crude product
by rotavapor distillation at 40 °C. Indeed, in another
experiment conducted using n-propylmagnesium chloride, this
transformation could be completely suppressed by the addition
of HCl into the solution just before rotavapor distillation. The
hydrochloride thus formed (2rd·HCl) could then be converted
into the corresponding spirocyclic quaternary ammonium salt
6rd by simple stirring in tBuOMe in the presence of NaOH. In
addition, from a mechanistic point of view, the same
titanacyclopropane intermediate complex A1 (Scheme 1) is
expected to be formed from isopropylmagnesium chloride and
from n-propylmagnesium chloride. The fact that two different
products, the isopropyl- and n-propyl-substituted molecules 6rc
5647
dx.doi.org/10.1021/om500603v | Organometallics 2014, 33, 5643−5653

Organometallics
Article
and 6rd, are exclusively obtained using the corresponding
Grignard reagents provides convincing evidence that the
organometallic species attacking the metalated iminium
intermediate Cr is not A1.
Further Studies. On the basis of the first set of results,
further experiments were conducted with the aims of extending
the possibilities offered by the reductive alkylation reaction,
improving its efficiency, and gaining new insight into the
mechanistic pathways involved. Several topics were investigated
and are discussed hereafter.
beneficial (entry 8). Nonetheless, the 70% yield achieved for
the transformation of 1b into 2bb constitutes a significant
improvement over our initial result. The same kind of positive
effect has been observed recently in several cases of
intramolecular cyclopropanation reactions, performed under
similar conditions, starting from N-alkenyl thioamides.7
Amounts of Ti(OiPr)4 and Grignard Reagent. On the basis
of the mechanisms put forward in Schemes 1 and 2, the
quantitative formation of the reductive alkylation product 2
requires 1 equiv of A and 1 equiv of Grignard reagent.
Therefore, n ≥ 1 mmol of Ti(OiPr)4 and p ≥ 2n + 1 mmol of
Grignard reagent should be engaged per millimole of 1. Various
combinations complying with this rule were explored starting
from the substrates 1a, 1b, and 1e in THF with cyclo-
hexylmagnesium chloride as the Grignard reagent (Table 5).
Two reactions of 1b were also run with a smaller amount of
Ti(OiPr)4 (entries 5 and 6; vide infra). In full agreement with
Trapping of the Final Intermediate. The presence of a
carbon−metal bond in the proposed final intermediate D
(Scheme 2) is supported by the results of reactions quenched
with D2O. When cyclohexylmagnesium chloride was used as
the Grignard reagent with the thioamides 1a, 1b, and 1n, the
corresponding deuterated amines 2ab-d6, 2bb-d, and 2nb-d
were synthesized with ∼90% deuterium incorporation. The
case of 2bb-d is interesting because two diastereoisomers of this
molecule may be produced. In practice, the diastereoisomeric
ratio was found to be close to 50:50, even when the addition of
D2O was performed at −70 °C (Scheme 6 and Figure 2). If it is
assumed that the deuteriolysis of the final intermediate Dbb
proceeds with retention of configuration, this result can be
explained by poor diastereoselectivity of the Grignard reagent
addition onto the zwitterionic intermediate complex Cb and/or
by configurational instability of Dbb.
In another experiment, the formation a carbon−carbon bond
from Dab was attempted using benzaldehyde as an electrophilic
trap (75 min at 20 °C then 30 min at reflux of THF). However,
the formation of the anticipated amino alcohol product was not
observed. The major product was still the reductive alkylation
adduct 2ab (44% yield), suggesting that the nucleophilicity of
the complex Dab is not sufficient for it to react with
benzaldehyde. A similar negative result was obtained starting
from the thioacetamide 1e. Other reactions were attempted
without success: with an activated Michael acceptor, with the
radical trap TEMPO, with anisaldehyde in the presence of
CuBr2, and with allyl bromide in the presence of CuCN·2LiCl.9
Rate of Addition of the Grignard Reagent. The influence of
the rate of addition of the Grignard reagent was briefly
examined using cyclohexylmagnesium chloride, starting from
the substrates 1a and 1b in THF (Table 4). In both cases, an
addition time of about 2 h consistently gave better results than
a 5−30 min addition (entry 3 vs entries 1 and 2 and entry 7 vs
entry 4). It should be pointed out that the magnitude of this
effect is moderate, and further reducing the addition rate is not
Table 5. Effect of the Amounts of Ti(OiPr)4 and
a
cC6H11MgCl on the Reactions of 1a, 1b, and 1e
mmol of
Ti(OiPr)4 per
mmol of 1
mmol of
cC6H11MgCl
per mmol of 1
b
entry substrate product
yield
44%
c
1
2
1a
1b
2ab
2bb
1.25
1.5
1.5
1.5
0.5
0.75
1.0
1.0
1.5
1.5
1.2
1.5
4.0
c
4.0
46−48%
54% (43%)
55%
c
3
4.3
c
4
5.0
d
5
4.0
14%
d
6
4.0
55%
d
7
3.0
36%
d
8
4.0
65%
d
9
4.0
70%
d
10
11
12
4.5
66%
e
f
1e
2eb
4.0
39%
e
g
4.0
37%
a
b
Reactions were run in THF, unless otherwise stated. Yields
estimated by NMR analysis of the crude products. The yield in
c
parentheses is for the isolated product. The Grignard reagent was
d
added over 26−45 min. The Grignard reagent was added over 100−
e
f
120 min. The Grignard reagent was added over 7−8 min. 31% yield
g
in tBuOMe or Et2O; 34% yield in cC5H9OMe. 43−48% yield in
tBuOMe.
the proposed mechanism, with 1.5 mmol of Ti(OiPr)4/mmol
of 1, little change in the yield is observed when the amount p of
the Grignard reagent is varied in the interval ranging from 4.0
to 5.0 mmol/mmol of 1 (entries 2−4, 9, and 10). Moreover,
when the value of p is fixed at 4.0 mmol/mmol of 1, the effect
of a reduction of the amount n of Ti(OiPr)4 to 1.25, 1.2, or
even 1.0 mmol/mmol of 1 is marginal as well (entry 1 vs entry
2, entry 8 vs entry 9, and entry 11 vs entry 12). However, when
n is further reduced to 0.75 and down to 0.5 mmol/mmol of 1,
the conversion is not complete and the yield of 2bb decreases
dramatically (55% and 14%, respectively, vs 65% with n = 1.0;
entries 5 and 6 vs entry 8). This is again consistent with the
proposed mechanism, since an insufficient quantity of the
intermediate A3 would then be generated. It should also be
noted that with the theoretical minimal amounts of reagents (n
= 1.0 and p = 3.0 mmol/mmol of 1), an unsatisfactory 36%
yield is obtained (entry 7).
Table 4. Influence of the Rate of Addition of the Grignard
a
Reagent
b
entry
substrate
product
addition time (min)
yield
51%
1
2
3
4
5
6
7
8
1a
2ab
7
28
54% (43%)
55%
110
7
1b
2bb
58%
30
59%
60
64% (55%)
66−70%
63%
110
260
a
In contrast, rather unexpected results were observed when
more “irrational” proportions of reagents were engaged
(Table 6). In particular, the results presented in entries 4 and
Reactions were run in THF using 1.5 equiv of Ti(OiPr)4 and 4.0−4.5
b
equiv of cC6H11MgCl. Yields estimated by NMR analysis of the crude
product. Yields in parentheses are for the isolated products.
5648
dx.doi.org/10.1021/om500603v | Organometallics 2014, 33, 5643−5653

Organometallics
Article
Table 6. Reactions Run with Unusual Proportions of Reagents Compared with Reactions Performed under the Standard
a
Conditions
b
b
entry
1
substrate
product
reagents (mmol per mmol of 1)
yield
dr
1e
2eb
Ti(OiPr)4 (1.5)
cC6H11MgCl (4.0)
Ti(OiPr)4 (2.0)
cC6H11MgCl (4.0)
Ti(OiPr)4 (1.5)
cC6H11MgCl (4.0)
Ti(OiPr)4 (2.4)
cC6H11MgCl (3.6)
Ti(OiPr)4 (1.5)
iPrMgCl (4.0)
43−48% (30%)
−
c
2
3
4
5
6
48%
−
1o
1o
2ob
2oc
81% (51%)
68:32
75:25
67:33
74:26
d
69%
79%
e
Ti(OiPr)4 (3.0)
iPrMgCl (4.5)
57%
a
b
Reactions were run in tBuOMe. The Grignard reagent was added within 5−8 min in each case. Yields and diastereomeric ratios (dr) estimated by
c
NMR analysis of the crude products. Yields in parentheses are for the isolated products. N,N-Dibenzylamine (22%) and starting material (19%)
were also observed in the crude product. 21% unreacted starting material 1o was detected in the crude product. 32% 1o was detected in the crude
d
e
product.
Scheme 7. Simplified Mechanistic Proposition Explaining the Reductive Alkylation of 1 When Ti(OiPr)4 and the Grignard
Reagent Are Employed in a 2:3 Ratio
6 were obtained using a 2:3 ratio of Ti(OiPr)4 and Grignard
reagent RMgX. Under these conditions, an equimolar amount
of A and organotitanium species RTi(OiPr)3 J are expected to
be produced (Scheme 7), and the addition of a significant
amount of Grignard reagent onto the metalated iminium
intermediate complex C is unlikely.14 Nonetheless, although
the conversion is not complete, the reductive alkylation process
remains efficient and the diastereoselectivity of the formation of
2ob and 2oc is improved (entries 4 and 6 vs entries 3 and 5).
These results can readily be understood if it is the alkyltitanium
complex J rather than the Grignard reagent that attacks the
zwitterionic intermediate species C. Complementary experi-
ments presented in the next two subsections strongly support
the validity of this hypothesis. Since complexes such as J are
expected to be moderately nucleophilic, some activation may
operate, for instance via a transition state such as K (Scheme
7).15 The formation of an intermediate ate complex, where the
negatively charged sulfur atom of C would coordinate to the
metal center of J before intramolecular delivery of the alkyl
group, could be proposed as well.
With respect to the titanium(IV) alkoxide reagent, ClTi-
(OiPr)3 led to a comparatively good result in the synthesis of
the amine product 2ab. This compound was formed in 57%
yield from 1a, compared with 51% using Ti(OiPr)4 under
otherwise identical conditions (Scheme 8 top). This is the best
Scheme 8. Comparison of the Reactions of 1a using
Ti(OiPr)4, ClTi(OiPr)3, MeTi(OiPr)3, and Me2Ti(OiPr)2
Generated in Situ
Other Reagent Systems. Replacement of the Grignard
reagent with an organolithium reagent was briefly examined.
After 1a was treated with Ti(OiPr)4 (1.5 equiv) and nBuLi (4.0
equiv) in THF, only small amounts of the expected alkylation
product N,N-dibenzylpentylamine (2af) were detected, and the
main reaction product was N,N-dibenzylamine, isolated in 56%
yield. Such reaction conditions are thus not suitable.
5649
dx.doi.org/10.1021/om500603v | Organometallics 2014, 33, 5643−5653

Organometallics
Article
yield achieved to date for this reaction. In contrast, in the case
of the synthesis of 2fb from 1f, a run using Cl2Ti(OiPr)2
instead of Ti(OiPr)4 in tBuOMe led to a more complex mixture
of products and a reduction in the yield from 47 to 33%.
A particularly interesting alternative to the use of Ti(OiPr)4
is MeTi(OiPr)3, as put forward and demonstrated by de
Meijere and co-workers.3c,16,17 Indeed, while the formation of
the key titanacyclopropane intermediate A3 from Ti(OiPr)4
requires the consumption of 2 equiv of cC6H11MgCl, the same
species can in principle be generated from MeTi(OiPr)3 using 1
equiv of cC6H11MgCl. In the present case, employing
MeTi(OiPr)3 seemed an attractive way of proving that
alkyl(triisopropyloxy)titanium complexes J can alkylate the
intermediate C. Indeed, the amine 2ae resulting from reductive
methylation is obtained as the major product from 1a, albeit
with low conversion (Scheme 8 middle). Another experiment,
run by adding cC6H11MgCl to a preformed equimolar mixture
of MeTi(OiPr)3 and Me2Ti(OiPr)2 in the presence of 1a, leads
to an improved result, suggesting that Me2Ti(OiPr)2 can also
add efficiently onto Ca (Scheme 8 bottom).
Table 7. Extension to the Generation and Trapping of the
Metalated Iminium Complex C
a
Extension to the Introduction of Other Nucleophiles.
Finally, the more general possibility of controlling the
formation of the putative key zwitterionic intermediate C in
order to trap it with various nucleophiles was investigated. A
potential problem was an attack on this metalated iminium
species, in the course of its generation, by the organotitanium
intermediate complex J (see Scheme 7). Since our results
suggested that this process is slower in THF than in tBuOMe
and also slower at 0 °C than at 20 °C, the best chances of
success seemed to lie in the use of THF at low temperature.
Good and reliable results are indeed obtained using 1.5 equiv of
Ti(OiPr)4 and 3.0 equiv of cyclohexylmagnesium chloride
added at −30 °C in THF, with the reaction mixture being
allowed to warm to 0 °C over a 30 min period prior to trapping
with several nucleophiles (Table 7).18 As expected, the
following order is observed for the introduction of a methyl
group: MeLi > MeMgBr > MeTi(OiPr)3 (entries 1−5).
Notwithstanding the low yield obtained with the last of these
reagents (35%), its ability to add onto the species Ca is
demonstrated unequivocally. It is interesting to note that, as
anticipated, the addition of MeTi(OiPr)3 is more facile in
tBuOMe than in THF (entry 5 vs entry 4). Presumably the
transition state K can be reached more easily in the less-
coordinating solvent tBuOMe. In contrast, with MeLi, a better
result is obtained in THF than in tBuOMe (68%, entry 1 vs
44%, entry 2).
With PhLi, the dimeric diamine 3ah is produced almost
exclusively (entry 7). This perhaps surprising result is discussed
in the next subsection. A hydride or deuteride anion can also be
delivered using LiAlH4 or LiAlD4 (entries 8 and 9). In contrast,
the following reagents proved unsuitable: NaBD4, 2-phenyl-
ethynyllithium, 1-phenylvinyloxylithium, vinylmagnesium bro-
mide (with or without subsequent trapping with anisaldehyde),
(trifluoromethyl)trimethylsilane (in the presence or absence of
nBu4NF), phenol, and diethyl malonate deprotonated with
NaH. Rather surprisingly, extensive amounts of starting
material 1a were observed in several of these cases.9 An
experiment involving the direct aqueous workup of a solution
of Ca was then carried out. The crude product contained
essentially pure 1a (entry 10). Further work is needed to fully
understand this result and the overall reactivity of the putative
species Ca, but it appears that no reaction (or a very slow one)
takes place with water and decomplexation eventually occurs,
a
b
Unless otherwise stated, the reactions were run in THF. Yields
estimated by NMR analysis of the crude products. Yields in
parentheses are for the isolated products. This reaction was run in
tBuOMe. The substrate 1a was the main component of the crude
product (90−93%). A small amount of amine 2ah (4%) was also
observed. 3ah was produced as a 57:43 meso/dl mixture of
diastereoisomers. Only the meso compound was isolated. The
reaction was quenched with D2O. Deuterium incorporation: 71% d2,
26% d1, 3% d0. The hydrolysis was carried out on half of the reaction
mixture. The other half was successfully reacted with LiAlH4 to
produce 2ai. 55:45 dr (measured in the crude product). cC6H11MgCl
was added at 0 °C over 6 min, and MeLi was then added after 2 min of
stirring at 0 °C. 57:43 dr (measured in the crude product).
c
d
e
f
g
h
i
probably from Ba, which is assumed to be in equilibrium with
Ca. It is also interesting to note that no trace of the cyclohexyl-
containing amine 2ab is produced under these conditions.
The selectivity of the nucleophilic addition process is good in
several cases, with 2ab being produced as a very minor side
product (less than 7% yield; entries 1, 5, and 7−9). The result
of the reaction with water discussed above (entry 10) suggests
that 2ab is formed after the introduction of the nucleophilic
reagent, presumably when the reaction mixture is allowed to
warm to 20 °C. However, starting from the thiolactam substrate
1q, the formation of rather large amounts (20%) of cyclohexyl
byproduct 2qb is not avoided (entries 11 and 12). A solution to
this problem, in this particular case, consists of adding 3.0 equiv
of MeLi at 0 °C to a THF solution of 1q and cC6H11Ti(OiPr)3
(generated in situ from 1.5 equiv of Ti(OiPr)4 and 1.5 equiv of
cC6H11MgCl at −10 °C). Under these conditions, the
production of 2qb is suppressed and the yield of 2qe is not
affected.9 The absolute configuration of the carbon center
bearing the newly introduced methyl group, on the slightly
major diastereoisomer, is assigned as S on the basis of literature
data.19
An additional experiment aimed at quickly evaluating the
generation of Cb from 1b by ligand exchange from A (R = Et),
preformed by the reaction of 2.0 equiv of Ti(OiPr)4 with 3.6
5650
dx.doi.org/10.1021/om500603v | Organometallics 2014, 33, 5643−5653

Organometallics
Article
equiv of nBuLi in THF at 0 °C,20,21 gave an unsatisfactory
result after trapping with cyclohexylmagnesium chloride: the
expected product 2bb was formed in 35% yield but was
accompanied by the production of the n-pentylamine derivative
2bf in 18% yield.
Formation of Dimers. The unexpected dimeric compounds
3cb and 3ah are both produced as a mixture of two
diastereoisomers. The major diastereoisomer of 3ah was
shown to be the meso molecule by X-ray crystallography
(Figure 3). These products are likely to result from
radical species from the intermediates Dcb and Dah is thus
comparatively much more favorable thermodynamically than in
the general situation, and their dimerization is also especially
plausible.
In an attempt to ascertain whether 3cb could be formed at
the stage of the workup of the reaction mixture through a
pathway in which the radical Lcb would be generated after
single electron transfer from oxygen to Dcb, an experiment was
conducted in which the aqueous workup was preceded by
exposure of the reaction mixture to dry air for 30 min. Under
these conditions, the yield of the “normal” reductive alkylation
product 2cb dropped drastically from 33 to 5%, which is
consistent with the consumption of the final intermediate Dcb
and the generation of the radical Lcb. However, no increase in
the amount of 3cb was observed,23 and a new compound was
produced. Although the latter could not be isolated and has not
been unequivocally identified, it is reasonable to assume that it
is the hydroperoxide resulting from trapping of Lcb by
molecular oxygen.24 This result suggests that the diamines
3cb and 3ah are formed before the workup of the reaction
mixture.
It is worth pointing out that similarly stabilized radicals could
be generated in reactions performed from the substrates 1i and
1k, although the corresponding dimeric diamine products were
not observed. It is reasonable to assume that if formed, these
radicals would add intramolecularly onto the neighboring
alkene group, resulting in the production of other compounds
that were not identified.
CONCLUSIONS
■
The reductive alkylation of thioamides 1 by Grignard reagents
in the presence of Ti(OiPr)4 is a fairly general reaction. The
main limitations of this transformation are the necessity to use a
tertiary thioamide and, in the simplest version of the reaction, a
Grignard reagent having hydrogen atoms at the β position.
However, the yields are often rather moderate or low. Our new
investigations show that improved results can be achieved and
that the parameters especially worth considering for opti-
mization are the choice of the solvent, the nature of the
titanium(IV) reagent, in particular ClTi(OiPr)3 instead of
Ti(OiPr)4, and the rate of addition of the Grignard reagent.
From a mechanistic point of view, although our initial
proposition proved to be essentially correct, the actual picture
is more complicated, with a rich variety of possible pathways.
All of our results are consistent with the initial generation of a
titanacyclopropane complex A, as widely accepted when such a
combination of reagents is engaged. A ligand exchange process
with the thiocarbonyl group leads to the formation of the
metalated iminium species C. This complex behaves as a
moderately electrophilic reagent that has been demonstrated to
react with LiAlH4, alkyllithium, alkylmagnesium, and
alkyltitanium(IV) reagents. In particular, as C is being gradually
generated, trapping by the first intermediate RTi(OiPr)3
produced by the reaction of RMgX with Ti(OiPr)4 can be a
significant pathway, notably in tBuOMe. This process is
believed to involve coordination of RTi(OiPr)3 by the sulfur
atom of C.
Figure 3. Single-crystal X-ray diffraction structure of the major meso
diastereoisomer of 3ah. Ellipsoids are displayed at the 50% probability
level.
dimerization of the radicals Lcb and Lah generated by
homolysis of the carbon−titanium bond of the corresponding
intermediates Dcb and Dah (Scheme 9). Although the
mechanism of such a homolysis event is not clear (e.g., under
the action of light22 or after single-electron oxidation or attack
by a heteroatom-centered radical), one can point out that the
radicals Lcb and Lah are stabilized by resonance; Lah is even
stabilized by the captodative effect. The generation of these
Scheme 9. Proposed Pathway Leading to Diamines 3
Depending on the conditions applied, in particular the
proportions of the reagents, the actual reaction pathway is thus
flexible. When MeTi(OiPr)3 is used rather than Ti(OiPr)4, the
reaction can even be diverted toward the formation of reductive
methylation products that were not accessible initially.
Moreover, by reducing the amount of Grignard reagent
5651
dx.doi.org/10.1021/om500603v | Organometallics 2014, 33, 5643−5653

Organometallics
Article
General Procedure for the Generation of the Intermediate
Complex C, Followed by Trapping with a Nucleophilic
Reagent. Ti(OiPr)4 (1.5 mmol/mmol of 1) is added to a stirred
solution of thioamide substrate 1 in the reaction solvent (v mL). The
Grignard reagent (p mmol/mmol of 1) is then introduced (typically at
−30 °C) over 3−7 min. During the addition, the solution typically
turns yellow. The reaction mixture is then allowed to warm to 0 °C
over 30 min (orange color). The nucleophilic reagent (typically 1.5
mmol/mmol of 1) is added at 0 °C. After 5 min the cold bath is
removed, and the reaction mixture is stirred at 20 °C for 60 min. Then
25% NH3 aqueous solution (0.50 mL/mmol of substrate) is added,
and after 30 min of stirring (the formation of a white precipitate is
observed), the mixture is exposed to air and filtered through a short
pad of sand, Na2SO4, Celite, and sand (from bottom to top), rinsing
with Et2O. The resulting clear solution is concentrated under reduced
pressure to afford the crude product. The latter is analyzed by NMR
spectroscopy and the products are typically isolated by flash column
chromatography. Typically, the solvent is THF (v = 10), the Grignard
reagent is cC6H11MgCl (p = 3.0) and is added at −30 °C. The
reactions were performed starting from 0.50 to 1.0 mmol of 1.
employed and by controlling the temperature, it is possible to
selectively generate the above-mentioned metalated iminium
intermediate C, which can then react with a nucleophile
introduced at this stage. This represents a substantial extension
of our original method.
The final α-metalated amine intermediate D behaves as a
base but does not exhibit nucleophilic properties. However, it
can play the role of a radical precursor. In particular, when the
corresponding radicals are sufficiently stabilized, diamines
resulting from dimerization of these species can be produced.
Another possible reaction that was observed experimentally is
the formation of hydroperoxide compounds by aerobic
oxidation of D.
Future work on this topic in our laboratory will be devoted to
new extensions of the reductive alkylation reaction and further
elucidation of some aspects of this chemistry.
EXPERIMENTAL SECTION
■
General Information. Titanium(IV) isopropoxide (VERTEC
TIPT) was purchased from Alfa Aesar, distilled under reduced
pressure, and stored under nitrogen for several months. The Grignard
reagents (methylmagnesium bromide, n-propylmagnesium chloride,
isopropylmagnesium chloride, tert-butylmagnesium chloride, cyclo-
pentylmagnesium chloride, and cyclohexylmagnesium chloride) were
purchased from Sigma-Aldrich or Acros as solutions (typically ∼2 M in
Et2O) and titrated once a month according to methods described
earlier.10e,25 The organolithium reagents (methyllithium, n-butyl-
lithium, sec-butyllithium, and phenyllithium) were purchased from
Sigma-Aldrich or Acros as solutions (typically 1.7−2.5 M in hexanes)
and titrated once a month according to literature methods.25,26 Other
commercial reagents were used as received without purification.
Tetrahydrofuran, diethyl ether, toluene, and dichloromethane were
purified using an MB SPS-800 solvent purification system
(MBRAUN). tert-Butyl methyl ether and cyclopentyl methyl ether
were purchased from Acros or Alfa Aesar and used as received.
Petroleum ether refers to the 40−60 °C fraction. This and all of the
other solvents were purchased from VWR and used without
purification. All of the reactions were carried out under a stream of
nitrogen or argon, unless specified otherwise. The temperatures
mentioned are the temperatures of the cold baths used. Concentration
under reduced pressure was carried out using rotary evaporators at
40 °C.
ASSOCIATED CONTENT
* Supporting Information
■
S
Experimental procedures, additional results, characterization
data and spectra, and crystallographic data (CIF). This material
is available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*Fax: (+33) (0)1 6933 5972. E-mail: yvan.six@polytechnique.
■
edu.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
We warmly thank Ewelina Augustowska, Antoine Boiron,
me Deffit, Natalia Szwankowska, Aleksander Przybyła,
Kamila Piotrowska, and Natalia Witkowska for their contribu-
tions to this work. We are also grateful to Michel Levart for the
■
Jer
́
o
̂
́
MS analyses of our compounds and to the Ecole Polytechnique
for financial support, in particular for the research grant
attributed to F.H. Funding from the CNRS is also acknowl-
edged. Finally, Y.S. especially thanks Prof. S. Z. Zard for his
continuous encouragement and valuable support.
Flash column chromatography was performed on Merck silica gel
60 (40−63 μm). NMR spectra were recorded with Bruker AM 400
and AVANCE 400 spectrometers (1H at 400.2 MHz, 13C at 100.6
MHz). Infrared spectra were recorded with a PerkinElmer 2000 FT-IR
spectrometer. Melting points were determined using a Buchi 535
̈
REFERENCES
■
apparatus. Low-resolution mass spectra were recorded on a Hewlett-
Packard Quad GC−MS engine spectrometer via direct injection. High-
resolution mass spectrometry was performed on a JEOL GCmate II
spectrometer.
(1) (a) Kulinkovich, O. G.; Sviridov, S. V.; Vasilevskii, D. A.;
Prityskaya, T. S. Zh. Org. Khim. 1989, 25, 2244−2245; Russ. J. Org.
Chem. 1989, 25, 2027−2028. (b) Kulinkovich, O. G.; Vasilevskii, D.
A.; Savchenko, A. I.; Sviridov, S. V. Zh. Org. Khim. 1991, 27, 1428−
1430; Russ. J. Org. Chem. 1991, 27, 1249−1251. (c) Kulinkovich, O.
G.; Sviridov, S. V.; Vasilevskii, D. A.; Savchenko, A. I.; Pritytskaya, T. S.
Zh. Org. Khim. 1991, 27, 294−298; Russ. J. Org. Chem. 1991, 27, 250−
254. (d) Kulinkovich, O. G.; Sviridov, S. V.; Vasilevski, D. A. Synthesis
1991, 234.
(2) Selected reviews: (a) Kulinkovich, O. G.; de Meijere, A. Chem.
Rev. 2000, 100, 2789−2834. (b) Sato, F.; Urabe, H.; Okamoto, S.
Chem. Rev. 2000, 100, 2835−2886. (c) Wolan, A.; Six, Y. Tetrahedron
2010, 66, 15−61. (d) Wolan, A.; Six, Y. Tetrahedron 2010, 66, 3097−
3133. (e) Cha, J. K.; Kulinkovich, O. G. In Organic Reactions, Vol. 77;
Denmark, S. E., Ed.; Wiley: New York, 2012; Chapter 1.
General Procedure for the Reductive Alkylation of Thio-
amides. The alkoxytitanium(IV) reagent (n mmol/mmol of 1) is
added to a stirred solution of thioamide substrate 1 (typically 0.50−1.0
mmol) in the reaction solvent (v mL/mmol of 1). The Grignard
reagent (p mmol/mmol of 1) is then added at 0 °C over a time t. The
cold bath is removed, and the reaction mixture is stirred at 20 °C for a
time T. After this time, 25% NH3 aqueous solution (0.50 mL/mmol of
substrate) is added. After 30 min of stirring (the formation of a white
precipitate is observed), the mixture is exposed to air and filtered
through a short pad of sand, Na2SO4, Celite, and sand (from bottom
to top), rinsing with Et2O. The resulting clear solution is concentrated
under reduced pressure to afford the crude product. The latter is
analyzed by NMR spectroscopy. The products are typically isolated by
flash column chromatography. Typically, the alkoxytitanium(IV)
reagent is Ti(OiPr)4 (n = 1.5), p = 4 for the Grignard reagent, the
solvent is tBuOMe or THF (v = 10 to 20), t ranges from a few to 120
min, and T is 60 min.
(3) Initial report: (a) Chaplinski, V.; de Meijere, A. Angew. Chem.
1996, 108, 491−492; Angew. Chem., Int. Ed. Engl. 1996, 35, 413−414.
Dedicated review: (b) de Meijere, A.; Kozhushkov, S. I.; Savchenko, A.
I. J. Organomet. Chem. 2004, 689, 2033−2055. Also see: (c) de
Meijere, A.; Chaplinski, V.; Winsel, H.; Kordes, M.; Stecker, B.;
5652
dx.doi.org/10.1021/om500603v | Organometallics 2014, 33, 5643−5653

Organometallics
Article
Gazizova, V.; Savchenko, A. I.; Boese, R.; Schill (nee
́
Brackmann), F.
Chem. 2001, 624, 229−238. (c) Eisch, J. J.; Owuor, F. A.; Otieno, P. O.
Organometallics 2001, 20, 4132−4134.
Chem.Eur. J. 2010, 16, 13862−13875.
(4) Initial report: (a) Bertus, P.; Szymoniak, J. Chem. Commun. 2001,
1792−1793. Dedicated review: (b) Bertus, P.; Szymoniak, J. Synlett
2007, 1346−1356.
(22) Cp2TiCl2 undergoes homolysis of a C−Ti bond when exposed
to UV irradiation. See: Brindley, P. B.; Davies, A. G.; Hawari, J. A.-A. J.
Organomet. Chem. 1983, 250, 247−256 and references cited therein.
(23) However, it must be said that the solubility of the dimers 3cb
and 3ah in common solvents is low, which certainly affects the
accuracy of the yields obtained and tends for them to be
underestimated.
(5) Initial reports: (a) Harada, K.; Urabe, H.; Sato, F. Tetrahedron
Lett. 1995, 36, 3203−3206. (b) Nakagawa, T.; Kasatkin, A.; Sato, F.
Tetrahedron Lett. 1995, 36, 3207−3210. (c) Urabe, H.; Hata, T.; Sato,
F. Tetrahedron Lett. 1995, 36, 4261−4264. Reviews: (d) Sato, F.;
Urabe, H.; Okamoto, S. Synlett 2000, 753−775. (e) Sato, F.; Okamoto,
S. Adv. Synth. Catal. 2001, 343, 759−784. (f) Reichard, H. A.;
Micalizio, G. C. Chem. Sci. 2011, 2, 573−589.
(6) Augustowska, E.; Boiron, A.; Deffit, J.; Six, Y. Chem. Commun.
2012, 48, 5031−5033.
(7) Hermant, F.; Nicolas, E.; Six, Y. Tetrahedron 2014, 70, 3924−
3930.
(8) Selected examples: (a) Savchenko, A. I.; Kulinkovich, O. G. Zh.
Org. Khim. 1997, 33, 913−915; Russ. J. Org. Chem. 1997, 33, 846−848.
(b) Quan, L. G.; Kim, S.-H.; Lee, J. C.; Cha, J. K. Angew. Chem. 2002,
114, 2264−2266; Angew. Chem., Int. Ed. 2002, 41, 2160−2162.
(c) Isakov, V. E.; Kulinkovich, O. G. Synlett 2003, 967−970.
(d) Lysenko, I. L.; Kim, K. K.; Lee, H. G.; Cha, J. K. J. Am. Chem.
Soc. 2008, 130, 15997−16002. (e) Bobrov, D. N.; Kim, K.; Cha, J. K.
Tetrahedron Lett. 2008, 49, 4089−4091. (f) Madelaine, C.; Ouhamou,
N.; Chiaroni, A.; Vedrenne, E.; Grimaud, L.; Six, Y. Tetrahedron 2008,
64, 8878−8898. (g) Astashko, D.; Lee, H. G.; Bobrov, D. N.; Cha, J. K.
J. Org. Chem. 2009, 74, 5528−5532. (h) Astashko, D.; Cha, J. K.; Rao,
N. N.; Parida, B. B. Eur. J. Org. Chem. 2014, 181−187.
(9) See the Supporting Information for details.
1
(24) The H NMR signals assigned to this compound disappeared
when NaBH(OAc)3 was added to a CDCl3 solution of a crude product
where it had been identified as a constituent. See the Supporting
Information for further details.
(25) Lin, H.-S.; Paquette, L. A. Synth. Commun. 1994, 24, 2503−
2506.
(26) Kofron, W. G.; Baclawski, L. M. J. Org. Chem. 1976, 41, 1879−
1880.
(10) Especially relevant examples: (a) Nakamura, E.; Oshino, H.;
Kuwajima, I. J. Am. Chem. Soc. 1986, 108, 3745−3755. (b) Lee, J.; Ha,
J. D.; Cha, J. K. J. Am. Chem. Soc. 1997, 119, 8127−8128. (c) Sung, M.
J.; Lee, C.-W.; Cha, J. K. Synlett 1999, 561−562. (d) Six, Y. J. Chem.
Soc., Perkin Trans. 1 2002, 1159−1160. (e) Six, Y. Eur. J. Org. Chem.
2003, 1157−1171. (f) Cadoret, F.; Retailleau, P.; Six, Y. Tetrahedron
Lett. 2006, 47, 7749−7753.
(11) Moreover, experiments conducted in the frame of the present
study, where dry air was introduced at the end of the reaction of the
thioamide 1c, strongly support the hypothesis of the aerobic oxidation
of Dc into a peroxide compound (see the Supporting Information for
details). For this reason, we recommend maintaining the nitrogen (or
argon) flow during the aqueous workup of the reactions.
(12) Wiberg, K. B.; Rablen, P. R. J. Am. Chem. Soc. 1995, 117, 2201−
2209 and references therein.
(13) As evaluated by NMR spectroscopy in CDCl3 solution at 20 °C,
the rotamer ratio changes from 100:0 to 96:4 after 2 h. After total
times of 4 and 8 h, it becomes 93:7 and 88:12, respectively.
(14) For other applications of the use of Ti(OiPr)4 and a Grignard
reagent in a similar ratio, see: (a) Reference 10f. (b) Cadoret, F.; Six,
Y. Tetrahedron Lett. 2007, 48, 5491−5495. Also see: (c) Lack, O.;
Martin, R. E. Tetrahedron Lett. 2005, 46, 8207−8211 (2 equiv of
MeTi(OiPr)3 and 1 equiv of EtMgBr).
(15) With the standard procedure and proportions of reagents, it is
not excluded that the addition of the Grignard reagent onto C
proceeds via an analogous transition state.
(16) (a) Chaplinski, V.; Winsel, H.; Kordes, M.; de Meijere, A. Synlett
1997, 111−114. (b) de Meijere, A.; Winsel, H.; Stecker, B. Org. Synth.
2005, 81, 14−25.
(17) Also see: Lee, J. C.; Sung, M. J.; Cha, J. K. Tetrahedron Lett.
2001, 42, 2059−2061.
(18) Performing the reaction entirely at 0 °C gives somewhat inferior
results, and so does the use of tBuMgCl instead of cC6H11MgCl (see
the Supporting Information for details).
(19) Nugent, T. C.; Ghosh, A. K. Eur. J. Org. Chem. 2007, 3863−
3869.
(20) Rassadin, V. A.; Six, Y. Tetrahedron 2014, 70, 787−794.
(21) This combination of reagents was introduced by Eisch and co-
workers. See: (a) Eisch, J. J. J. Organomet. Chem. 2001, 617−618, 148−
157. (b) Eisch, J. J.; Gitua, J. N.; Otieno, P. O.; Shi, X. J. Organomet.
5653
dx.doi.org/10.1021/om500603v | Organometallics 2014, 33, 5643−5653