2
M.D. Ritz et al. / Polyhedron 180 (2020) 114416
catalyst loading (3d and 3i) or adding electron-donating groups
(3 m) did result in an increase in observed yield. The hydrogena-
tion reaction was extended to substrates bearing cyclopentane
(3 g) and cyclopropane (3 h) functional groups. Furthermore, inter-
nal alkenes were also tolerated as seen by products 3n, 3o, and 3p.
Strikingly, substrates with electron withdrawing groups shut
down catalytic activity (See Supporting Information for all incom-
patible substrates). For example, 1,1-disubstituted olefins with
chloro, bromo, CF3, and CO2Et substituents gave no hydrogenation
products. Additional experiments were run to better understand
why certain substrates are not tolerated by this system. 2a and
an incompatible substrate, 2-(1-phenylvinyl)thiophene, were
added to the same reaction flask and subjected to the standard
hydrogenation conditions previously described. No product forma-
tion of either substrate was seen, suggesting that catalyst inhibi-
tion, from substrate binding, could be the cause of this inactivity.
Another competition reaction was run between 2a and an incom-
patible substrate, 1-bromo-4-(1-phenylvinyl)benzene. Here it
was seen that some product from hydrogenation of 2a, 1,1-diphe-
nylethane, had formed. This result suggests that the inability to
hydrogenate alkenes with some electron-withdrawing groups is
due to the substrates inhibiting the catalyst itself, either by coordi-
Scheme 1. Cobalt-catalyzed alkene hydrogenation catalysts.
air-sensitive adding an additional amount of rigor to the hydro-
genation reactions. Herein, we report a bench stable, novel, bisox-
azoline ligated cobalt(II) dichloride complex utilizing a one-step
synthesis to access enantioenriched ligand (Scheme 2). Starting
with cheap and commercially available building blocks, treatment
nation or by
Information).
a
chemical side-reaction (See Supporting
It is also important to note that while complex 1 is enantiopure,
no enantioinduction was observed in these hydrogenation reac-
tions (1–5% e.e., see Supporting Information). The lack of enantios-
electivity was surprising, in that hydroboration of arylethenes with
a similar pyridine-oxazoline-cobalt catalyst proceeds with good
enantioselectivity (up to 99% e.e.) [29]. Addition of a drop of mer-
cury to the reaction with 1,1-diphenylethylene does not inhibit the
reaction (quantitative hydrogenation), providing support for a
homogeneous catalyst. However, examination of the filtrate at
the end of the hydrogenation reaction by 1H NMR spectroscopy
shows the presence of free bis-oxazoline ligand. This observation
could indicate that the true catalyst forms from ligand dissociation,
but the mercury drop experiment argues against a nanoparticle
catalyst. It is more likely that ligand loss follows from catalyst
decomposition, as no species could be isolated or identified by
reaction of 1 with NaBEt3H. This methodology was extended to
the gram scale synthesis of 3c where a 90% isolated yield was
achieved. The ability to apply this chemistry to a large scale high-
lights the practicality of this methodology (See Supporting
Information).
of iminodiacetonitrile with
L
-valine in the presence of ZnCl2
afforded enantiopure ligand in 21% isolated yield. Reaction of
enantiopure ligand with anhydrous Co(II) dichloride in THF at
40 °C (to improve solubility) afforded paramagnetic complex 1.
The 1H NMR spectra of complex 1 show broadened and paramag-
netically shifted resonances (See Supporting Information). Slow
diffusion of ether into a THF solution containing 1 at ꢀ30 °C
afforded a crystal suitable for X-ray structure determination
(Fig. 1), which reveals a distorted trigonal bipyramidal geometry
(
s5 = 0.716) [28] around the cobalt center.
2.2. Hydrogenations with catalyst 1
We began our investigation by examining the hydrogenation of
1,1-diphenylethylene (2a) catalyzed by a combination of bisoxazo-
line ligated cobalt complex 1 and reductant in a solution of toluene
at room temperature under 1 atm of hydrogen. The use of sodium
triethylborohydride as the reductant of choice resulted in 99% yield
of hydrogenated product 3a (Table 1, entry 1). Examination of
alternative reducing agents did result in productive catalysis, how-
ever, poor conversion was observed (entry 2–3). Control reactions
showed that without catalyst (entry 4) or reducing agent (entry 5),
no product formation was observed.
3. Experimental
3.1. General information
With optimized reaction conditions, various alkenes were
tested (Table 2). In general, these conditions proved to be general
for a variety of alkenes with nonpolar and electron-donating
groups, with electron withdrawing groups not being tolerated
(See Supporting Information). Electron-donating groups such as
methoxy (3b, 3g-h, 3m, and 3p) and a tertiary amine (3c) were
well tolerated. Next, in order to assess steric hindrance a series
of diarylethenes methylated in the ortho, meta, and para position
were subjected to the optimized reaction conditions. Methylation
of the meta (3f) and para (3e) position gave hydrogenated product,
however, no product was observed upon ortho-methylation (See
Supporting Information).
Unless otherwise noted, all reagents were purchased from com-
mercial suppliers and used without further purification. Iminodi-
acetonitrile (Alfa Aesar) was recrystallized from ethyl acetate
before use. All procedures and routine manipulations were per-
formed under a nitrogen atmosphere (glovebox) or on a high-vac-
uum line using modified Schlenk techniques. Tetrahydrofuran
(THF), toluene, ether, pentane, and dichloromethane (DCM) were
obtained by from an Innovative Technology PS-MD-6 solvent
purification system. Solvents were further degassed by 3–6 free-
pump-thaw cycles and stored in a nitrogen-filled glovebox. 1H
and 13C{1H} NMR spectra were acquired on 400 and 500 MHz Bru-
ker NMR instruments. NMR chemical shifts are reported in ppm
and are referenced to the residual solvent peak for CDCl3
(d = 7.26 ppm, 1H NMR; d = 77.16 ppm, 13C{1H} NMR. Coupling con-
stants (J) are reported in Hertz. GC–MS spectra were recorded on a
Shimadzu QP2010 instrument with an SH-Rxi-5 ms column
Applying these conditions to N-heterocyclic containing
substrates proved to be successful (3d and 3 l). Additionally, boron
(3i) and ferrocene (3j) containing functional groups were tolerated.
For substrates that showed diminished yields, increasing the