With the optimized reaction conditions in hand, we investigated the
substrate scope of γ ,γ -aryl/alkyl-substitutedN -phthaloyl allylamines with the (Z )-configuration at
first. All the hydrogenated products, regardless of the electronic
properties of R1 and the steric hindrance of
R2, were obtained in excellent yields and
enantioselectivities (Scheme 2). The electron-donating 4-methyl- and
4-ethyl-substituted amides 2b and 2c were obtained in
98% ee, while 4-tert -butyl substituted product 2d was
obtained with an excellent enantioselectivity of >99.9%
ee. For substrates bearing
Scheme 2. Substrate scope of (Z )-1. Conditions:(Z)-1 (0.2 mmol), (R )-SKP (0.0021 mmol),
[Rh(cod)2]SbF6 (0.002 mmol),
H2 (50 atm), EtOAc (2 mL), rt, 12 h. Isolated yields
were recorded. The ee values of 2 were determined by HPLC using
chiral columns.
an electron-withdrawing 4-halogen group, the corresponding products2e and 2f were obtained in 99% and 98% ee,
respectively. The substrates bearing phenyl and nitro groups gave
complete conversions and perfect enantioselectivities
(>99.9% ee for 2g and 2i ), while another
product 2h , with a trifluoromethyl at the para -position,
was obtained with 99% ee. When the electron-donating methyl group was
substituted at the meta -position, the enantioselectivity was
maintained at 98% (2j ), whereas the presence of a methyl group
at the ortho -position resulted in a reduced enantioselectivity of
92% (2l ). The substrates with a fluoro group at themeta -position provided the products with an excellent
enantioselectivity of 99% (2k ). Disubstituted substrates
bearing a 1-naphthyl group gave the corresponding product 2qwith good enantioselectivity. Other disubstituted substrates bearing
electron-withdrawing groups gave their desired products with better ee
values than those bearing electron-donating groups
(2m -2p ). Additionally, substrates bearingn -butyl as an R2 substituent, which were
synthesized from n -butyl lithium, have also been tested. The use
of n -butyl instead of a methyl group led to an increase in
enantioselectivity compared with the model product (2r vs2a ), while the enantioselectivities of other products bearing
different substituents remained at 99% (2s -2u vs2k , 2n , 2p ). Furthermore, substrates
processing heteroaryl groups, such as 2-thienyl and 3-thienyl, were also
amenable to this catalytic system, affording 2v and 2wquantitatively with 96% and 95% ee values, respectively. To our
delight, an alkyl-substituted substrate also accommodate this catalytic
system, providing 2x with an excellent enantioselectivity of
98%. The absolute configurations of γ -chirogenic amines in
Scheme 2 were considered to be the same as 2g whose
configuration was assigned to be R by XRD analysis.
Subsequently, we studied the asymmetric hydrogenation ofγ ,γ -aryl/alkyl-substituted N -phthaloyl allylamines
with the (E )-configuration. All the reduced products, regardless
of the electronic properties of R1 and the steric
hindrance of R2, were obtained in excellent yields and
good to excellent enantioselectivities (Scheme 3). The electron-donating
4-alkyl-substituted amides 2b′ and 2y′ were obtained
with 90% ee, while the electron-withdrawing 4-halogen-substituted
products 2e′ , 2z′ and 2f′ were obtained with
ees of 92%. The 4-CF3-substituted product 2h′was synthesized with an enantioselectivity of 92%, while the
4-Ph-substituted amide 2g′ showed a better ee of 94%. The
substrates with methyl and fluoro group at the meta -position
provided the products 2j′ and 2k′ with complete
conversions and better enantioselectivities (91% and 93% ee,
respectively) than those bearing corresponding para -substituents.
In contrast to the allylamides with a (Z )-configuration, the
(E )-substrates have better enantioselectivities when the
substituents are at the ortho -position. More specifically, the
2-methyl substituted amide 2l′ was obtained with 97% ee, while
the 2-halogen-substituted amides 2aa′ and 2ab′ were
both obtained in 98% ee. Disubstituted substrates bearing
3,5-dimethylphenyl, 3,5-dichlorophenyl, 1,3-benzodioxol-5-yl or
2-naphthyl moieties gave enantioselectivities between 90-93%
(2o′ , 2ac′ , 2ad′ and 2ae′ ), while
1-naphthyl-substituted product 2q′ was obtained with an
excellent enantioselectivity of 97%. Additionally, substrates bearing
different R2 substituents have also been tested. The
use of larger alkyl substituents led to an increase in
enantioselectivity (2af′ -2ah′ ). However, the
application of a benzyl group gave the product 2ai′ with a
relatively lower ee of 87%. A substrate bearing γ -2-naphthyl andγ -phenyl groups was also subjected to hydrogenation to give the
desired product 2aj′ with 86% ee. Furthermore, substrate
processing a heteroaryl group such as a 2-thienyl ring was also amenable
to this catalytic system, affording 2v′ with 73% ee. To our
delight, alkyl substituted substrate are also ammenable to this
catalytic system, providing the desired product 2x′ with 92%
ee. The absolute configuration of the chiral γ -branched amides in
Scheme 3 were opposite to the corresponding products in Scheme 2
according to the HPLC charts.
Scheme 3. Substrate scope of (E )-1. Conditions:(E)-1 (0.2 mmol), (R )-SKP (0.0021 mmol),
[Rh(cod)2]SbF6 (0.002 mmol),
H2 (50 atm), EtOAc (2 mL), rt, 12 h. Isolated yields
were recorded. The ee values of 2′ were determined by HPLC
using chiral columns.
Organofluorine compounds possess unique physicochemical properties and
are highly important in medicinal, agricultural and material chemistry.
Approximately 30% of all agrochemicals and almost 20% of all
pharmaceuticals contain fluorine. However, the asymmetric introduction
of fluorine into organic molecules is challenging. Hence, the
development of new, simple and efficient methods to access a diverse
array of novel chiral fluorinated derivatives, has become a highly
prioritized research area.[16] Transition
metal-catalyzed asymmetric hydrogenation of alkenyl fluorides is one of
the most promising methods.[17] The relatively
strong electron withdrawing behaviour of the fluorine together with the
easy-to-break C–F bond, making these types of alkene substrates
difficult to be hydrogenated. With the above-mentioned limitations in
mind, we set out to develop the first efficient asymmetric hydrogenation
of γ -aryl-γ -fluoro-substituted N -phthaloyl
allylamines for the highly efficient synthesis ofγ -aryl-γ -fluoro-substituted chiral amines.
Scheme 4. Substrate scope of γ -fluoro-substituted
allylamides. Conditions: 3 (0.2 mmol), (R )-SKP (0.0021
mmol), [Rh(cod)2]SbF6 (0.002 mmol),
H2 (10 atm), EtOAc (2 mL), -10 °C, 12 h. Isolated yields
were recorded. The ee values of 4 were determined by HPLC using
chiral columns.
Applying our previously optimized conditions (50 atm H2,
rt) to substrate 3a gave only 30% of
the desired product 4a with
96% ee together with 70% de-fluorinated by-product. In order to solve
the problem of de-fluorination, we further optimized the conditions. To
our delight, decreasing the hydrogen pressure and reaction temperature
to 10 atm and -10 °C respectively, successfully afforded the desired
product 4a in 95% yield and with 96% ee, and reduced the
amount of de-fluorinated by-product to 5%. With the adjusted reaction
conditions in hand, we investigated the substrate scope ofγ -fluoro allylamides with the (Z )-configuration (Scheme
4). The substrates with electron-withdrawing groups on the phenyl rings
were well tolerated and generated the desired products in high isolated
yields and with excellent enantioselectivities. The
4-halogen-substituted products 4b and 4c were obtained
in 99% ee. The 4-CF3-substituted product 4dalso gave an excellent ee of 99%. The 3-halogen-substituted products4e and 4f were obtained with slightly lower
enantioselectivities of 95% and 98%. For substrates bearing a fluorine
at the 2-position, the desired product 4g was obtained with
96% ee. Disubstituted substrates bearing electron-withdrawing groups
also gave their desired products with excellent ee values
(4h -4k ). More specifically, the product 4iwith 3,4-dichloro substituents was obtained with a perfect
enantioselectivity of >99.9% ee. The product 4jbearing 3,5-difluoro substituents and the product 4k bearing
3,5-dichloro substituents both gave their products with 99% ee. The
absolute configurations of the chiralγ -aryl-γ -fluoro-substituted amides were considered to be
the same as 4i whose configuration was assigned to be Sby XRD analysis.
Scheme 5. Substrate scope of γ -aryloxy-substituted
allylamides. Conditions: 5 (0.2 mmol), (R )-SKP (0.0021
mmol), [Rh(cod)2]SbF6 (0.002 mmol),
H2 (50 atm), EtOAc (2 mL), rt, 12 h. Isolated yields
were recorded. The ee values of 6 were determined by HPLC using
chiral columns.
γ -Aryl-γ -aryloxy-substituted chiral amines are widely
present in natural products and chiral pharmaceuticals, making the
asymmetric catalytic synthesis of such compounds very attractive. As far
as we know, no asymmetric hydrogenation of γ -aryloxy-substituted
allylamides has been studied, probably due to difficulties related to
the low activities of ene ethers and the poor stereocontrol of the
reaction.
Using the previously optimized reaction conditions, we investigated the
substrate scope of γ -aryloxy-substituted allylamides with a
(Z )-configuration (Scheme 5). All the reduced products,
regardless of the electronic properties and the steric hindrance of
R1, were obtained in excellent yields and
enantioselectivities. γ -Phenyloxy-γ -phenyl-substituted
amide 6a was obtained in 95% ee. For a substrate bearing an
electron-donating methyl substituent at the 4-position of aryloxy, the
desired product 6b was obtained in 97% ee.
4-Halogen-substituted compounds showed an increasing tendency towards
better enantioselectivity with the corresponding products6c -6e being obtained with 98%, 99% and
>99.9% ee values, respectively. The product 6fbearing a 4-phenyl group was produced with 99% ee, while the
4-CF3-substituted product 6g was obtained with
96% ee. A similar trend was observed for the products6h -6k possessing 3-substituents with that of the
4-substituted ones. The product 6l bearing a 2-fluorine group
was also produced with 98% ee. Disubstituted substrates bearing
electron-withdrawing groups gave the desired products with better ee
values than those bearing electron-donating groups
(6m -6p ). The absolute configurations of the chiralγ -aryloxy-γ -phenyl-substituted amides were considered to
be the same as 6e whose configuration was assigned to beS by XRD analysis.
To demonstrate the potential utility of the protocol for the synthesis
of γ -chirogenic amines, the hydrogenation was carried out on a
gram scale and under a high substrate/catalyst (S/C) ratio, and the
hydrogenated products were further converted to several important
pharmaceutical compounds (Scheme 6). The hydrogenation catalyzed by
(R )-SKP/[Rh(cod)2]SbF6 under
20000 S/C afforded the desired product 2a and 4m in
high yields and excellent enantioselectivities. Compound 2a was
smoothly hydrazinolyzed to the corresponding γ -chirogenic amine7 , which can be further derivatized to a bioactive amide8 in high yield with no loss in enantioselectivity. The product4m was hydrazinolyzed, ring closed and salified to generate
1,2,3,4-tetrahydro-4-methylquinoline hydrochloride 11 with
perfect ee. For another example, the hydrogenation of 5a at
5000 S/C produced the γ -fluoro-substituted amide in 93% yield
and with 96% ee. In particular, the hydrogenation of 5g could
also be conducted in EtOAc under 50 atm H2 at room
temperature catalyzed by
(S )-SKP/[Rh(cod)2]SbF6 under
50000 S/C, affording the desired product 6g in high yield and
excellent enantioselectivitiy. After changing the N -substituent
from phthaloyl to methyl, the antidepressant drug Fluoxetine was
obtained in 83% yield from 5g and with >99.9%
ee.
To gain insight into the reaction meachanism, deuterium experiment using
D2 instead of H2 was conducted fort he
hydrogenation of substrate (E)-1a , resulting in an
exclusive deuterium addition to the vinyl group and with no alkene
isomerization occurring (Scheme 7).
Scheme 6. Scale-up and applications.
Scheme 7. Deuterium experiment.
Computational calculations were conducted in order to gain further
insight into the catalytic mechanism and stereocontrol. Unlike many
other bisphosphine-Rh-catalyzed asymmetric hydrogenations,SKP-Rh does not demonstrate any capacity to form stable
catalyst-substrate complexes with either E - orZ -substrates. The corresponding minima for SKP-Rh-Subcomplexes can be located, but they are extremely unstable (more than 10
kcal/mol endergonic) and do not involve any coordination between the
catalyst and the double bond of the substrate. On the other hand, the
formation of molecular hydrogen complexSKP-Rh-H2 and further dihydride complexSKP-Rh-HH via TS-OA are strongly exergonic. Actually,
it is also possible to hydrogenate the extremely unstableSKP-Rh-Sub complexes with essentially the same result. The
resting state consists of SKP-Rh-HH and uncoordinated
substrate. It should be noted that the dihydride
RhIII-complex bears a P-Rh-P bite angle of 171.7° and
does not contain a hydride trans to a phosphorus atom, that
features in the way of its further reaction with the substrate (Figure
2). So far as we know, a similar form of coordination occurs only in
pincer-metal complexes, and has not been reported in bidentate
bisphosphine-metal complexes.[18]
Due to the configuration of the dihydride SKP-Rh-HH with
extremely crowded space around Rh, the calculation also does not show
coordination of the substrate’s double bond to Rh in theSKP-Rh-HH-Sub complexes. The coordination is strongly
endergonic and the complex is maintained by weak intramolecular
interactions. This is completely different from the traditional
coordination-promoted activation mode in the Rh-catalyzed
hydrogenation.[12] Due to specific features of the
substrate coordination, and flexibility of the conformations of the
phenyl rings of the catalyst, there are numerous possibilities for
substrate coordination, and hence for the following stages (see
Supporting Information for details). The most possible pathways for theE -substrate have been shown in Figure 3. Migratory insertion
proceeds with reasonable activation barriers for both S- andR -pathways, yielding the corresponding monohydridesSKP-Rh-H-IntS and SKP-Rh-H-IntR . Further
reductive elimination gave the SKP-Rh-Prod complexes and
dissociation of the substrate completes the catalytic cycle. According
to the free energy profile, both migratory insertion and reductive
elimination contribute to the enantioselectivity and a high barrier for
reductive elimination may switch the stereoselection. The difference in
Gibbs free energies between TS-MIS andTS-MIR is 7.0 kcal/mol in favor of the former.TS-RER is more stable than TS-RES by 3.8
kcal/mol. Hence, the whole catalytic cycle for the E -substrate is
S-stereoselective with a free energy difference of 3.2 kcal/mol.
It also can be seen that the substrate selectively adopts a specific
configuration to approach the Rh-hydride and thus forms a relatively
stable transition state with more weak interactions between catalyst and
substrate. Important weak intramolecular interactions in these
transition states are shown in Figure 4. The red numbers show the
position of the transferred hydride relative to the Rh and carbon atoms.
Blue numbers show distances between the protons and centers of the
phenyl rings characterizing the corresponding C-H…π interactions.
There are other weak intramolecular interactions in both structures
which roughly compensate each other. From TS-MIS andTS-MIR we can roughly estimate that the energy of two
single C-H…π interactions is approximately 7.0 kcal/mol. FromTS-RES and TS-RER we can roughly
estimate that the free energy of a single C-H…π interaction is
around 3.8 kcal/mol. The two results taken together give a ∆G ≈ 4
kcal/mol for a C-H…π interaction of approximately 2.5 Å length.
The authors realize that these evaluations are rough and approximate,
but are strictly convinced that acquisition of such data is very
important for quantification of weak interactions between the catalyst
and substrate and are an ultimate goal of the mechanistic studies of
enantioselective catalytic cycles.
Figure 2. Oxidative addition.
Figure 3. Migratory insertion and reductive elimination.