Angewandte
Communications
Chemie
we were able to verify that acid-sensitive functionalities, such
nificantly lower ee (ꢂ 88%, see chap. 2.3). Thus, the present
as Boc-carbamates, cyclic acetals, tert-butyl esters and
silylethers, which would be rapidly cleaved by HCl, are
compatible with the reaction conditions (Scheme 1C).
Noteworthy, in these cases no additional base as in
previous catalytic approaches[8] was required, which would
result in a depleted atom economy and poorer waste balance.
Moreover, a range of sterically hindered b-branched amino
alcohol derivatives was efficiently converted into the corre-
sponding chlorides (examples 221 to 226). As reported for
chlorinations with Vilsmeier–Haack-type reagents,[13a] steri-
cally encumbered cyclohexanol derivatives represent chal-
lenging substrates. Nevertheless, catalytic chlorination of
cholesterol 117 afforded the steroid analogue 217 (in this case
with retention of configuration)[8b] in 51% yield, which could
be improved to 74% by switching to DMF as the solvent (and
catalyst). Besides the excellent functional group compatibility
our method is also distinguished by a superior scalability as
demonstrated by the mol scale synthesis of geranyl chloride
(E-23) and the non-commercial chlorides Z-23 and 227
(Scheme 1B,E). Importantly, these experiments were con-
ducted utilizing standard laboratory equipment (volume
ꢂ 1 L).
method must be considered as the most efficient approach
available today for the stereoselective chlorination of non-
racemic alcohols, also with respect to operational simplicity
and waste-balance. Furthermore, on dilution of the crude
chlorination reaction mixtures with MeCN or MeOH fol-
lowed by addition of K2CO3 as a base and a suitable N-, O-, S-,
and C-nucleophile, we were able to obtain a range of amines,
azides, ethers, sulfides, and nitriles of type 8–12 as shown in
Scheme 2.
Actually, the synthesis of allylic chlorides such as Z- and
E-23, which are important building blocks for natural
products, is not trivial at all:[15] In the past specialized and
waste-intensive procedures had to be applied, as simple
reagents, such as SOCl2 or HCl, only afford complex mixtures
of linear and branched regioisomers and E/Z-diastereomers
(see chap. 2.1 and Table S4,S5).[1c–e] In contrast, our catalytic
chlorination method not only provides Z- and E-23 with
excellent regioselectivity but also displays a superior waste-
balance with economy factors (E-factors; that is the mass
ratio of waste/isolated product) down to 2. This meets the
typical range for industrial bulk chemical production on
a multi-ton scale (E-factors of 1–5).[5] According to Metzgerꢀs
environmental assessment tool for organic synthesis
(EATOS)[6] our synthesis of E-23 generates 89–94% less
waste than the conventional methods (e.g. under Appel-
conditions with CCl4/PPh3). And due to the low amounts of
solvent (or its complete omission) and the non-hazardous
nature of the reagents involved the overall environmental
impact (which also considers potential human and ecotoxico-
logical effects)[5,6] is minimized to ꢂ 7% (see Figures S1–S3).
In contrast, the majority of the previous catalytic methods
require rather large solvent amounts to suppress sideproduct
formation (e.g. oxalate esters in the case of the reagent oxalyl
chloride).[8] A comparative EATOS examination of our
method with recent catalytic approaches for the chlorination
of 1-phenylethanol (12) revealed that the waste amount could
actually be reduced to ꢂ 4% (Figure S4).
Scheme 2. One-pot chlorination and substitution (for detailed reaction
conditions see Supporting Information). All yields refer to isolated
products after chromatography. [a] Transformation 2!9–12 conducted
in the presence of 10 mol% of TBAI. [b] ee determined by [a]D25-value.
As starting material phenylethanol S-12 in 99% ee was utilized.
Advantageous, this practical one-pot route allows avoid-
ing the isolation of alkyl chlorides 2, which are prone to
(thermal or SiO2-induced) HCl elimination or hydrolysis in
many cases. Again, an excellent stereoselectivity was
observed (99%!95% ee), when the non-racemic alcohol S-
12 was employed as starting material to furnish the drug S-
Fendiline (S-82f) under overall retention (two-fold SN2-
inversion). To further demonstrate the applicability, racemic
Clopidogrel was efficiently prepared (from the racemic
precursor). The proposed mechanism for the chlorination
1!2 is shown in Scheme 3A.
First evidence for type II intermediates was gained by the
identification of formyl esters 13 as side products. On treating
aliphatic alcohols 1 with equimolar amounts of a formamide 3
and BzCl at room temperature (in CDCl3), iminium salts of
type II could be detected by NMR spectroscopy (Sche-
me 3B). While hydrolysis of iminium chlorides II provided
formiates of type 13, heating to 608C effected quantitative
conversion into chlorides
2 (Scheme S10). Comparison
experiments verified that esters of type 7 and 13 (side
products formed in trace amounts) do not occur as productive
intermediates in the catalytic chlorination cycle: Exposed to
an excess of HCl under the standard reaction conditions, 71
and 131, respectively, afforded chloride 21 only in negligible
quantities (Scheme 3C).
In conclusion, a highly efficient method for the trans-
formation of alcohols 1 into chlorides 2 utilizing BzCl as the
sole reagent in the presence of a simple formamide catalyst
We could also show that enantioenriched alcohols
(99% ee) are converted under inversion into the correspond-
ing chlorides with excellent levels of stereoselectivity (ꢀ
95% ee, Scheme 1D). Particularly remarkable is the chlori-
nation of 1-phenylethanol (S-12) to benzylic chloride (R-22)
with 95% ee. In this representative example other methods
such as the Appel-reaction, which is usually the first choice
for stereospecific chlorinations,[1c–e] afforded R-22 with sig-
Angew. Chem. Int. Ed. 2016, 55, 1 – 6
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3
These are not the final page numbers!