A versatile synthesis of phosphine-aminophosphine ligands for asymmetric catalysis

Article abstract of DOI:10.1016/j.tetasy.2005.05.014

A new and versatile synthesis of phosphine-aminophosphine ligands allows the incorporation of a wide range of nitrogen and phosphorus substituents into these ligands, several of which exhibit improved properties for rhodium-catalyzed asymmetric hydrogenation reactions. This synthesis also allows the preparation of mixed phosphine-phosphoramidite species.

Full text of DOI:10.1016/j.tetasy.2005.05.014

Tetrahedron:
Asymmetry
Tetrahedron: Asymmetry 16 (2005) 2063–2066
A versatile synthesis of phosphine–aminophosphine ligands
for asymmetric catalysis
Neil W. Boaz,* James A. Ponasik, Jr. and Shannon E. Large
Research Laboratories, Eastman Chemical Company, PO Box 1972, Kingsport, TN 37662, USA
Received 7 April 2005; accepted 6 May 2005
Abstract—A new and versatile synthesis of phosphine–aminophosphine ligands allows the incorporation of a wide range of nitrogen
and phosphorus substituents into these ligands, several of which exhibit improved properties for rhodium-catalyzed asymmetric
hydrogenation reactions. This synthesis also allows the preparation of mixed phosphine–phosphoramidite species.
ꢀ 2005 Elsevier Ltd. All rights reserved.
1. Introduction
for the asymmetric hydrogenation of a wide variety of
dehydroamino acids, itaconates, and a-ketoesters. In
addition, these species demonstrate outstanding air sta-
bility and are designed to avoid any extreme conditions
in their preparation. Thus, this ligand family combines a
number of characteristics that make them uniquely
attractive for asymmetric catalysis.
Asymmetric catalysis is a powerful technology for gener-
ating single enantiomer materials. Of the asymmetric
catalytic reactions, asymmetric hydrogenation is per-
haps the most powerful, as it utilizes an inexpensive re-
agent (hydrogen) with an often relatively inexpensive
unsaturated substrate to afford a high value single enan-
tiomer product. The catalysis is usually performed with
precious metals complexed to chiral ligands, which en-
force the enantioselectivity. There have been a large
number of chiral ligands prepared for asymmetric catal-
ysis.1 In general, the most easily designed and most
abundant ligands are C2-symmetrical bis-phosphines
due to their stereochemical redundancy. Significantly
lesser in number are other arrangements, such as biden-
tate bis-aminophosphines,2 bis-phosphites,3 bis-phos-
phinites,4 bis-phosphonites,5 and mixed ligands such as
phosphine-phosphite,6 aminophosphine-phosphinite,7
phosphine-phosphonite,8 phosphine-phosphoramidite,9
and phosphonite-phosphite10 systems.
R
N
PR'2
Fe PPh2
1
2. Results and discussion
It was noted at an early stage that a limited variation of
the substituents on the nitrogen and phosphorus atoms
of the aminophosphine portion of 1 resulted in signifi-
cant modulation of the reactivity characteristics of these
ligands. Unfortunately, a wider nitrogen and phospho-
rus substituent variation was not viable using the origi-
nal synthesis. Variation of the nitrogen substituent was
limited by the steric hindrance around the final interme-
diate 2, such that when R was larger than n-propyl, the
final reaction with a chlorodiarylphosphine failed,
affording none of the desired product 1 under the stan-
dard reaction conditions. Stronger conditions led only
to reagent decomposition. Variation of the phosphorus
substituents was more feasible, although there are few
commercially available chlorophosphines, and the prep-
aration of these types of species is tedious.
Perhaps the most noteworthy of the nonsymmetric chi-
ral bidentate ligands possessing a ferrocenyl backbone,
with a number of highly enantioselective species have
found prominence.11 Among these is a series of phosph-
inoferrocenyl-aminophosphine ligands that have re-
cently been reported as useful species for asymmetric
catalysis.12 The rhodium complexes of these BoPhozTM
ligands 1 exhibit high activities and enantioselectivities
*
Corresponding author. Tel.: +1 423 229 8105; fax: +1 423 224
7582; e-mail: nwboaz@eastman.com
0957-4166/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetasy.2005.05.014

2064
N. W. Boaz et al. / Tetrahedron: Asymmetry 16 (2005) 2063–2066
It was apparent that a modification of the synthetic
route was required to allow the preparation of a wider
variety of ligands 1. The most attractive strategy
involved exploiting the steric hindrance of 2 as an
advantage. This was achieved by utilizing the small elec-
trophile phosphorus trichloride in place of the larger
chlorodiphenylphosphine (or other disubstituted chloro-
phosphine). Under normal circumstances, a statistical
mixture of compounds with one, two, or three amines
on the phosphorus would be anticipated. However, the
steric bulk of 2 is such that with a 1:1 mixture of PCl3
and 2 only the monoadduct was formed, with no indica-
tion of diaminochlorophosphine (or products there-
from) or triaminophosphine present. Subsequent
reaction of the dichloroaminophosphine with a nucleo-
phile such as a Grignard reagent can afford the desired
species 1. The viability of this synthesis was initially
examined for the preparation of the parent species 1a
where R is methyl and R0 is phenyl. The reaction was
performed under moderate conditions (0 ꢁC to ambient
temperature) and afforded the desired 1a in 83% isolated
yield.13 With this proof of principle, a wide variety of
ligands were prepared with numerous substituents on
both the nitrogen and phosphorus (Table 1). The lati-
tude of this procedure is exemplified by the fact that
even the N-tert-butyl species 1d can be prepared, albeit
in moderate yield. The main impurities observed in these
reactions were aryl coupling compounds from the aryl
Grignard reagents. These were removed and the ligands
purified by either chromatography or crystallization.
R
N
R
PR'2
Rh
N
MeOH
or THF
PR'2
+
Rh
+
Fe PPh2
P
+
Fe
Ph2
-
-
CF3SO3
CF3SO3
1
Rh-1
Figure 1. Preparation of rhodium complex of 1.
O
O
H2
R1
OR2
R3
R1
OR2
R3
Y
Y
Rh-1
O
O
3
4
a: R1 = Ph, R2 = R3 = Me, Y = NH
b: R1 = Ph, R2 =H, R3 = Me, Y = NH
c: R1 = R2 = H, R3 = Me, Y = NH
d: R1 = 2-naphthyl, R2 = R3 = Me, Y = NH
e: R1 = cyclopropyl, R2 = Bn, R3 = Boc,Y = NH
f: R1 = H, R2 = Me, R3 = OMe, Y = CH2
g: R1 = R2 = H, R3 = OH, Y = CH2
Figure 2. Asymmetric hydrogenation reactions.
As indicated in Table 2, replacing the R group of 1 with
a substituent larger than methyl had in general a nega-
tive effect on the enantioselectivity of the asymmetric
hydrogenation of a variety of dehydroamino acids and
itaconates. These results validate the previous limited
study, which indicated that the N-methyl group was
superior to other nitrogen substituents such as ethyl or
n-propyl.12 The inclusion of a second center of asymme-
try in the substituent did not overcome the deleterious
effect of the added bulk, although moderate match and
mis-match effects were observed (ligands 1g and 1h).
Table 1. Ligands prepared
R
R
1) PCl3/NEt3
PhCH3
N
N
PR'2
H
Fe PPh2
Fe PPh2
2) R'MgBr
2
1
Compound
R
R0
Yield (%)
a
b
c
d
e
f
Me
i-Pr
Ph
Ph
83
67
13
48
77
64
75
83
73
52
61
81
35
83
84
85
77
58
Table 2. Nitrogen substituent effects on the asymmetric hydrogena-
tions of substrates 3 with Rh-1a
Neopentyl
t-Bu
Ph
Ph
Ligand
% ee
Ph
PhCH2
Ph
Ph
3a
3b
3c
3d
3e
3f
3g
1a
1b
1c
1d
1e
1f
99.1
91.8
70.7
60.4
89.8
88.3
94.2
73.0
99.4
17.4
81.2
75.6
94.0
93.4
82.8
86.0
96.1
68.3
93.6
93.4
93.4
13.6
95.4
79.6
98.1
90.0
47.6
57.0
88.9
86.3
92.6
70.9
98.6
32.9
26.8
38.4
67.8
70.6
46.8
19.4
94.0
50.2
5.5
97.4
90.9
16.4
21.6
66.4
83.3
11.4
21.7
g
h
i
(S)-Phenethyl
(R)-Phenethyl
Ph
Ph
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
4-MeOC6H4
4-MeC6H4
3,5-Me2C6H3
4-FC6H4
4-ClC6H4
3-FC6H4
3,4-F2C6H3
3,4-Cl2C6H3
3,5-F2C6H3
3,5-Cl2C6H3
15.8
0.8
j
k
m
n
o
p
q
r
44.0
20.6
28.6
1g
1h
a Hydrogenations were run at 0.1 M concentration in THF or metha-
nol with 1 mol % catalyst at 10 psig for 6 h.
s
Variation of the phosphorus substituents resulted in a
number of ligands that afforded very high enantioselec-
tivities for the asymmetric hydrogenation of dehydro-
amino acids and itaconates (Table 3), in many cases
matching and sometimes surpassing the parent ligand
1a. The data indicate that electron-withdrawing groups
Rhodium complexes of these new ligands were prepared
as indicated below (Fig. 1) and examined in situ for the
asymmetric hydrogenation of a variety of substrates
(Fig. 2).

N. W. Boaz et al. / Tetrahedron: Asymmetry 16 (2005) 2063–2066
2065
Table 4. Catalytic activities for the asymmetric hydrogenation of 3b
with Rh-1a
on the aromatic ring generally enhance the enantioselec-
tivity, while electron-donating groups have a slightly
negative effect. This may be due to an enhanced differ-
ence between a more electron-poor aminophosphine
and the electron-rich ferrocenyl phosphine, perhaps
resulting in more ordered and tighter substrate binding.
This is particularly noteworthy for the itaconate sub-
strates 3f and 3g, as a number of ligands with
electron-withdrawing substituents afford higher enantio-
selectivities with these substrates than the parent 1a.
Indeed, the best overall ligand 1o has 3-fluorophenyl
substituents on the phosphine, and affords hydrogena-
tion results (as the rhodium complex) that are compara-
ble or superior to 1a. This ligand provides
2-methylsuccinic acid in 99.0% ee, the best result
obtained to date for this substrate using ligands 1.
O
O
H2
Ph
OH
CH3
Ph
OH
CH3
Rh-1
HN
HN
O
O
3b
4b
TOF (hꢀ1 c
Ligand
ee (%)
Conversion (%)
)
1a
1i
98.8
91.5
94.8
97.6
98.6
97.1
99.2
98.5
99.0
99.0
98.4
99.6
99.8
99.0
99.5
99.8
99.8
99.9
99.8
99.8
99.8
99.5
49,900
25,000
16,100
46,700
42,000
9,100
1jc
1kc
1m
1n
1o
1p
1q
1r
27,900
17,800
20,600
17,800
14,500
Table 3. Phosphorus substituent effects on the asymmetric hydroge-
nations of substrates 3 with Rh-1a
1sb
Ligand
% ee
a Reactions run at 0.75 M concentration in methanol with 40 psig
hydrogen at a S:C ratio of 3000:1 unless otherwise noted.
b The methodology for determining the turnover frequency is as
described in Ref. 12b.
3a
3b
3c
3d
3e
3f
3g
1a
1i
99.1
97.8
97.7
98.2
98.7
98.7
98.8
99.0
98.9
98.8
98.7
99.4
98.3
98.8
82.9
98.8
99.0
99.2
97.2
99.2
99.1
99.1
96.1
94.0
84.3
50.2
96.7
95.6
97.6
94.8
43.0
15.2
20.0
98.1
96.7
97.5
97.9
98.3
98.5
98.5
98.5
98.6
98.6
98.4
98.6
98.1
98.2
95.3
98.8
98.4
99.0
99.4
98.8
99.2
98.8
94.0
92.2
91.6
96.2
96.8
93.3
97.4
94.2
95.4
94.3
96.5
97.6
94.1
96.0
98.9
97.6
97.5
99.0
98.4
98.6
98.6
98.8
c Reaction run at a S:C ratio of 5000:1.
1j
1k
1m
1n
1o
1p
1q
1r
1s
base) to afford mixed phosphine–phosphoramidites 5. A
variety of species were prepared encompassing electron-
donating and electron-withdrawing groups as well as
auxiliary chirality as indicated in Figure 3.
a Hydrogenations were run at 0.1 M concentration in THF or metha-
nol with 1 mol % catalyst at 10 psig for 6 h.
CH3
CH3
1) PCl3/NEt3
N
N
PhCH3
P(OR")2
H
Fe PPh2
Fe
PPh2
2) 2 R"OH
Et3N
The activity of homogeneous catalysts is of great practi-
cal consequence, as highly active species usually require
less of the expensive catalyst. The rhodium complex of
ligand 1a is particularly highly active, which compounds
its utility for asymmetric hydrogenation reactions.12
2
5
a: R" = Ph (79%)
b: R" = 4-MeOPh (65%)
c: R" = 4-CF3Ph (55%)
d: R" = R-1,1'-Binaphthyl-2,2'-diyl (71%)
e: R" = S-1,1'-Binaphthyl-2,2'-diyl (51%)
It was of interest to determine the activity of the rho-
dium complexes of these newly generated ligands 1, in
particular the ligands with varying phosphorus substitu-
ents, as they could provide insight into both steric and
electronic effects (and in general afford hydrogenation
products with high enantioselectivity). The turnover fre-
quencies of these species were determined from their ini-
Figure 3. Phosphine–phosphoramidite ligand preparation (isolated
yields in parentheses).
Asymmetric hydrogenation results using these species
are shown in Table 5, and indicate that the rhodium
complexes of phosphine–phosphoramidites 5 afford high
enantioselectivites in some cases, but are not nearly as
general as the phosphine–aminophosphines 1.
tial rates of hydrogenation of
a test substrate,
2-acetamidocinnamic acid (3b).12b There was no obvious
correlation of rate with either substituent electronic nat-
ure or steric constraints (Table 4). However, most of the
rhodium complexes of these ligands are highly active for
the asymmetric hydrogenation of 3b, with only 1n failing
to crest a turnover frequency of 10,000 hꢀ1. Thus, any of
these ligands will afford useful practical species for
asymmetric hydrogenation.
Of particular interest are the BINOL-substituted species
5d and 5e wherein the BINOL functionality spans both
P–O valencies. Reports from other research groups con-
cerning these species appeared concurrent with our
investigations, although the ligands were prepared using
slightly different methodologies.9b–d These reports indi-
cate that species 5d in particular affords high enantio-
selectivities for a variety of asymmetric hydrogenation
reactions. Our work agrees with these reports, particu-
larly for substrate 3b, wherein the (R)-BINOL derivative
The modified synthetic procedure outlined above also al-
lows the use of other nucleophiles in place of Grignard re-
agents. In particular, phenols can be used as nucleophiles
in the second step of the reaction (with triethylamine as

2066
N. W. Boaz et al. / Tetrahedron: Asymmetry 16 (2005) 2063–2066
Table 5. Asymmetric hydrogenations of substrates 3 with Rh-5a
5. Reetz, M. T.; Gosberg, A.; Goddard, R.; Kyung, S.-H.
Chem. Commun. 1998, 2077, and references cited therein.
6. Sakai, N.; Mano, S.; Nozaki, K.; Takaya, H. J. Am.
Chem. Soc. 1993, 115, 7033.
7. (a) Krause, H. W.; Kruezfeld, H.-J.; Schmidt, U.; Dobler,
Ch.; Michalik, M.; Taudien, S.; Fischer, C. H. Chirality
1996, 8, 173; (b) Mi, A.; Lou, R.; Jiang, Y.; Deng, J.; Qin,
Y.; Fu, F.; Li, Z.; Hu, W.; Chan, A. S. C. Synlett 1988,
847.
Ligand
% ee
3a
3b
3c
3e
nrb
3f
6.9
3g
5a
5b
5c
5d
5e
92.9
95.2
91.0
99.1
17.5
97.0
98.2
97.8
99.9
54.1
53.8
92.6
90.0
67.1
85.6
64.8
63.6
75.8
95.7
92.6
54.2
46.6
nrb
7.8
23.2
35.6
5.0
81.2
a Hydrogenations were run at 0.5 M concentration in THF or metha-
nol with 1 mol % catalyst at 10 psig for 6 h.
8. Reetz, M. T.; Gosberg, A. Tetrahedron: Asymmetry 1999,
10, 2129.
b No hydrogenation was observed.
9. (a) Francio, G.; Faraone, F.; Leitner, W. Angew. Chem.,
Int. Ed. 2000, 39, 1428; (b) Jia, X.; Li, X.; Lam, W. S.;
Kok, S. H. L.; Xu, L.; Lu, G.; Yeung, C.-H.; Chan, A. S.
C. Tetrahedron: Asymmetry 2004, 15, 2273; (c) Hu, X.-P.;
Zheng, Z. Org. Lett. 2004, 6, 2585; (d) Hu, X.-P.; Zheng,
Z. Org. Lett. 2005, 7, 419.
10. Reetz, M. T.; Pasto, M. Tetrahedron Lett. 2000, 41, 3315.
11. For reviews, see: (a) Richards, C. J. Tetrahedron: Asym-
metry 1998, 9, 2377–2407; (b) Colcacot, T. J. Chem. Rev.
2003, 103, 3101–3118.
5d affords near enantiomeric purity. There is also a sig-
nificant stereochemical match–mismatch effect with
most substrates between the diastereomers. The high
enantioselectivity generated from the rhodium complex
of 5d is likely due to the axial chirality of the BINOL
and not just the cyclic nature of the phosphoramidite
diol portion, as ligand prepared from the enantiomers
of hydrobenzoin and diethyl tartrate afforded consis-
tently poor enantioselectivities for asymmetric hydrogen-
ations. The kinetics of the hydrogenations utilizing the
rhodium complexes of 5d and 5e were examined with
substrate 3b as above, and indicated 5300 and 12,800
catalytic turnovers per hour, respectively, while main-
taining exceedingly high enantioselectivity with 5d
(99.8% ee). Thus, the rhodium complexes of these li-
gands, although substantially slower than the best phos-
phine–aminophosphines 1, show activities comparable
to many other ligands.12b Unfortunately, the exceed-
ingly high enantioselectivities with the rhodium complex
of 5d are not particularly broad, as the reactions with 3c
and 3f gave poor results.
12. (a) Boaz, N. W.; Debenham, S. D.; Mackenzie, E. B.;
Large, S. E. Org. Lett. 2002, 4, 2421; (b) Boaz, N. W.;
Mackenzie, E. B.; Debenham, S. D.; Large, S. E.; Ponasik,
J. A., Jr. J. Org. Chem. 2005, 70, 1872.
13. A typical synthetic procedure is exemplified for the
preparation of 1a: Toluene (10 mL) was added to a 100-
mL three-necked flask, which was cooled in ice to below
5 ꢁC. Phosphorus trichloride (0.26 mL; 3.0 mmol;
1.0 equiv) was added followed by triethylamine
(0.50 mL; 3.6 mmol; 1.2 equiv). (R)-N-Methyl-1-[(S)-2-
(diphenylphosphino)ferrocenyl]-ethylamine 2a (1.26 g;
3.0 mmol) dissolved in 10 mL of toluene was added over
about 5 min such that the temperature remained below
10 ꢁC. The reaction mixture was allowed to warm to
ambient temperature over 30 min and then stirred at
ambient temperature for 2 h. The reaction mixture was
cooled in ice to below 5 ꢁC and a 3.0 M ethereal solution
of phenylmagnesium bromide (3.5 mL; 10.5 mmol;
3.5 equiv) was added over ca. 10 min such that the
temperature remained below 10 ꢁC. The reaction mixture
was allowed to warm to ambient temperature overnight to
completely consume 2a according to TLC analysis. The
mixture was cooled in ice-water and saturated aqueous
sodium bicarbonate solution (20 mL) added at a rate such
that the temperature remained below 15 ꢁC. The layers
were separated, the aqueous solution was filtered to
remove insolubles, and it was further extracted with ethyl
acetate. The combined organic extracts were dried with
sodium sulfate and concentrated. The crude product was
filtered through a pad of neutral alumina and eluted with
1:9 ethyl acetate–heptane with 5% added triethylamine to
afford 1.52 g (83%) of 1a. 1H NMR (CDCl3) d 7.65 (m,
2H); 7.4–7.0 (m, 14H); 6.82 (m, 4H); 5.006 (m, 1H); 4.502
(br s, 1H); 4.40 (m, 1H); 4.15 (m, 1H); 3.798 (s, 5H); 2.148
(d, 3H, J = 3.30 Hz); 1.471 (d, 3H, J = 6.87 Hz). 13C NMR
(CDCl3) d 142.4 (d, JC–P = 9 Hz); 140.3 (d, JC–P = 24 Hz);
140.0 (d, JC–P = 10 Hz); 139.5 (d, JC–P = 7 Hz); 135.9 (d,
3. Conclusion
A new and general methodology has been developed to
prepare a wide variety of phosphine–aminophosphine
and phosphine–phosphoramidite ligands. Many of these
species afford excellent results as rhodium catalysts for
asymmetric hydrogenation reactions.
References
1. (a) For recent reviews, see: Ohkuma, T.; Kitamura, M.;
Noyori, R. In Catalytic Asymmetric Synthesis; Ojima, I.,
Ed., 2nd ed.; Wiley-VCH: New York, 2000, pp 1–110; (b)
Brown, J. M. In Comprehensive Asymmetric Catalysis;
Jacobsen, E. N., Pfaltz, A., Yamamoto, H., Eds.;
Springer: Berlin, 1999; Vol. I, pp 121–182; (c) Tang, W.;
Zhang, X. Chem. Rev. 2003, 103, 3029–3069.
2. Fiorini, M.; Giongo, G. M. J. Mol. Catal. 1980, 7, 411,
and references cited therein.
3. Reetz, M. T.; Neiberger, T. Angew. Chem., Int. Ed. 1999,
38, 179, and references cited therein.
JC–P = 22 Hz); 134.1 (d, JC–P = 22 Hz); 132.3 (d, JC–P =
17 Hz); 131.7 (d, JC–P = 18 Hz); 129.3 (s); 128.8 (s); 128.1–
127.2; 97.9 (dd, JC–P = 14, 28 Hz); 75.7 (d, JC–P = 14 Hz);
72.1 (d, JC–P = 5 Hz); 70.5 (s); 69.9 (s); 69.8 (s); 58.4 (dd,
4. (a) Cullen, W. R.; Sugi, Y. Tetrahedron Lett. 1978, 19,
1635; (b) Selke, R.; Pracejus, H. J. Mol. Catal. 1986, 37,
213; (c) Rajanbabu, T. V.; Ayers, T. A.; Halliday, G. A.;
You, K. K.; Calabrese, J. C. J. Org. Chem. 1997, 62, 6012;
(d) Chan, A. S. C.; Hu, W.; Pai, C. C.; Lau, C.-P.; Jiang,
Y.; Mi, A.; Yan, M.; Sun, J.; Lou, R.; Deng, J. J. Am.
Chem. Soc. 1997, 119, 9570.
JC–P = 9, 39 Hz); 30.6 (d, JC–P = 11 Hz); 18.5 (d, JC–P =
6 Hz). 31P NMR (CD2Cl2) d 58.8 (d, JP–P = 7.7 Hz); ꢀ25.3
(d, JP–P = 7.7 Hz). Chiral HPLC (250 · 4.6 mm Chiralpak
AD-H, 99:1 hexane–isopropanol, 1 mL/min, k = 254 nm):
tR (S,R-1a) 10.4 min, tR (R,S-1a) 11.6 min. HRMS m/z
calcd for C37H35FeNP2 (M+) 611.15942, found 611.16429.
24
½aꢁD ¼ ꢀ257 (c 0.96, toluene).