amine derivatives were limited to a single successful separa-
tion of Me-IAN enantiomers by preparatory chiral HPLC.
We describe here a practical synthesis of IAN amine 1. As
this is a “parent” IAN amine from which numerous others
can be synthesized, this development constitutes a general
route to N-substituted enantioenriched ꢀ-diketimines and
considerably extends access to potential chiral ꢀ-diketimi-
nate-based complexes.
Diastereomers 3a,b provided considerable flexibility to the
singular goal to separate them. For example, these diaster-
eomers are readily separated by preparative flash chroma-
tography (∆R ) 0.09). The more attractive approach is
f
fractional recrystallization that returns the (R)-IAN diaste-
reomer 3a selectively (>94% purity, HPLC). This diaste-
reomer is consistently a crystalline solid, whereas (S)-IAN
diastereomer 3b is routinely a foam (but will occasionally
crystallize upon standing when pure and solventless).
Resolutions are normally wasteful in nature as the dias-
tereomer of the chiral auxiliary is often useful only for
synthetic access to the enantiomer of the compound of
interest. Diastereomer 3b can be used in this manner to
provide (+)-IAN amine. Alternatively, we observed atropi-
somerization in alcohol solvent that is more rapid than the
same isomerization in aromatic hydrocarbons, which require
approximately 17 h at 145 °C (xylenes) to reach an
equilibrium ratio (1:1) of epimers 3a and 3b. Epimerization
of 3b for the purpose of ultimate convergence to one
enantiomer (1) could be achieved by warming the mother
liquor enriched in 3b in ethanol for 48 h (eq 1).
Extensive efforts to resolve various racemates in this chiral
pyridine family were uniformly unsuccessful, including
9
traditional and family approaches to chiral salt formation
and fractional recrystallization. These methods used a
selection of tartaric and mandelic acids that provided
crystalline salts, but without diastereomer enrichment in all
fractional recrystallization studies. The synthesis of racemic
ligands was readily scaled during this time, and many IAN
amine derivatives and their intermediates were routinely
crystalline solids. We therefore turned to R-methyl benzyl-
amine IAN 3 and a study of its suitability as a resolving
agent without primary recourse to chromatography.
1
0
Naphthol 2-triflate was coupled with commercially
available (S)-R-methylbenzyl amine using palladium cataly-
1
1
12
sis (Scheme 1). The desired aryl amine (2) was obtained
Scheme 1. Preparation of (-)-IAN Amine (1)
Subsequent cooling through fractional recrystallization re-
turns (R)-IAN 3a selectively (as a crystalline salt; 25% of
theoretical yield for each recycle) and ultimately enantio-
merically enriched (-)-1.
Removal of the auxiliary was examined using a variety
of conditions and strategies. Catalytic hydrogenation using
traditional palladium catalysts did not provide the desired
IAN amine (1) at pressures to 250 psi at room temperature.
Transfer hydrogenation required heating in methanol to
debenzylate the naphthylamine, but as noted above, atropi-
somerization under these conditions is favorable. We there-
fore turned to Lewis acid assisted debenzylation by ethane
thiol. Boron tribromide provided IAN amine 1 but was
complicated by additional products. By comparison, alumi-
num tribromide provided a substantially improved outcome,
returning IAN amine in 65% yield and 87% ee. This initial
result used amounts employed by others, which corresponded
14
3
to 32 equiv of AlBr and ethane thiol solvent (1821 equiv).
This protocol was improved by reducing the amount of AlBr
3
in 66% yield after direct recrystallization. 1-Chloroisoquino-
line is both commercially available and readily prepared.
Its coupling with naphthylamine 2 was accomplished using
trimethyl aluminum to deliver the desired IAN amine 3 as a
1
3
to 10 equiv while employing an amount of ethane thiol
solvent corresponding to 600 equiv. In this manner, the
desired IAN amine was obtained in 63% yield and 96% ee
reproducibly at the 1 g scale.
It is significant to note that our observations during this
optimization indicate that the amounts of Lewis acid and
mercaptan can be further lowered, but that the reaction
quench must be carefully executed in order to avoid
racemization. Specifically, 1 M NaOH must be added
rapidly to an efficiently cooled reaction flask at 0 °C.
Deviation from this protocol resulted in the formation of
1
:1 mixture of diastereomers. This C-arylation was entirely
1
regioselective to the limits of detection ( H NMR) and is
attributed to the nature of the intermediate aluminum anilide.
6
(
8) Cortright, S. B.; Yoder, R. A.; Johnston, J. N. Heterocycles 2004,
2, 223.
9) Vries, T.; Wynberg, H.; van Echten, E.; Koek, J.; ten Hoeve, W.;
6
(
Kellogg, R. M.; Broxterman, Q. B.; Minnaard, A.; Kaptein, B.; van der
Sluis, S.; Hulshof, L.; Kooistra, J. Angew. Chem., Int. Ed. 1998, 37, 2349.
(
10) Stang, P. J.; Hanack, M.; Subramanian, L. R. Synthesis 1982, 85.
2446
Org. Lett., Vol. 10, No. 12, 2008