On the racemization of chiral imidazolines

Article abstract of DOI:10.1021/jo8024384

(Chemical Equation Presented) Racemization of chiral imidazolines with base has been studied for the first time following an unexpected finding in the synthesis of chiral imidazoline ligands. Amine bases do not cause racemization. Strong inorganic bases can induce racemization, yet this occurs only when the nitrogen is unsubstituted, in agreement with a symmetry-allowed thermal disrotatory ring opening and closure from a diazapentadienyl anion. Surprisingly, even with electron-withdrawing N-substituents, no racemization is observed. Conditions which allow for the racemization-free manipulations of this important compound class have been defined and developed.

Full text of DOI:10.1021/jo8024384

On the Racemization of Chiral Imidazolines†
Carl A. Busacca,* Teresa Bartholomeyzik, Sreedhar Cheekoori, Nelu Grinberg,
Heewon Lee, Shengli Ma, Anjan Saha, Sherry Shen, and Chris H. Senanayake
Department of Chemical DeVelopment, Boehringer-Ingelheim Pharmaceuticals, Inc., 900 Ridgebury Road,
Ridgefield, Connecticut 06877
carl.busacca@boehringer-ingelheim.com
ReceiVed NoVember 3, 2008
Racemization of chiral imidazolines with base has been studied for the first time following an unexpected
finding in the synthesis of chiral imidazoline ligands. Amine bases do not cause racemization. Strong
inorganic bases can induce racemization, yet this occurs only when the nitrogen is unsubstituted, in
agreement with a symmetry-allowed thermal disrotatory ring opening and closure from a diazapentadienyl
anion. Surprisingly, even with electron-withdrawing N-substituents, no racemization is observed. Conditions
which allow for the racemization-free manipulations of this important compound class have been defined
and developed.
SCHEME 1. Racemization in SNAr Reaction
Introduction
The synthesis and use of imidazolines has expanded dramati-
cally in the past decade. This has been driven both by increased
applications in catalysis1 and especially by the development of
imidazoline-based drugs for various therapeutic targets. Human
imidazoline binding sites (IBS) have been identified, and the
nature of these receptor subtypes and their preferred ligands
have been explored in detail.2 Imidazoline drugs are now in
development for a remarkable array of disease states including
oncology,3 depression and neuroprotection,4 hypertension,5
inflammation,6 and analgesia.7,4a Many of these molecules are
chiral.
extensive racemization which occurred in the phosphide borane
SNAr reaction. A typical result using dicyclohexylphosphine
borane with KOH as base is shown in Scheme 1. We needed
therefore to study the parameters which control this process and
to gain an understanding of the mechanism involved.
We have recently reported the development of chiral phos-
phinoimidazolines ligands (BIPI ligands) that perform asym-
metric hydrogenations with near-perfect enantioselection.1a An
unexpected finding was made during this research, namely
Results and Discussion
We repeated the SNAr reaction of Scheme 1 using t-BuOK
in DMSO. Aliquots were periodically removed, quenched with
† Dedicated to the memory of the late Professor Albert I. Meyers.
9756 J. Org. Chem. 2008, 73, 9756–9761
10.1021/jo8024384 CCC: $40.75 2008 American Chemical Society
Published on Web 11/18/2008

On the Racemization of Chiral Imidazolines
required for racemization to occur, and we subsequently found
that the parent, unsubstituted system behaves similarly. It is also
clear from Figure 1 that if the SNAr reaction is not quenched at
the moment full conversion is reached, the ee of the isolated
product from the basic reaction mixture will continue to erode.
This fact then explained the range of optical purities we had
observed in different SNAr reactions prior to optimization.
Control experiments showed that the nucleophile in the SNAr
reactions, the phosphide borane anion, was also not a necessary
requirement for racemization. Racemization in DMSO or DMAc
with inorganic bases such as KOH, t-BuOK, or NaH themselves
occurred readily. We also examined the amine bases triethy-
lamine, DABCO, DMAP, and DBU. In no case did these bases
lead to imidazoline racemization, even after 24 h at 60 °C. With
the inorganic bases under these “superbasic”9 conditions,
however, the racemization which took place was always
accompanied by a significant amount of decomposition that was
not seen with the phosphide borane anion. We therefore
employed the phosphide borane as base in all subsequent studies
of the racemization. We chose to further simplify the system
by employing the unsubstituted chiral imidazoline (S,S)-3. By
removing the fluoro substituent, any substitution chemistry,
which was now known to be irrelevant to the racemization
process, could thus also be avoided.
Racemization of (S,S)-3 was studied at 60 °C using the
phosphide borane anion in DMSO, so that a baseline racem-
ization rate for the parent system could be established. In order
FIGURE 1. Racemization of reaction components.
(4) (a) Halaris, A.; Piletz, J. E. Ann. N.Y. Acad. Sci. 2003, 1009, 1 (Agmatine
and Imidazolines). (b) Cameron, O. G.; Abelson, J. L.; Young, E. A. Biol.
Psychiatry 2004, 56 (11), 875. (c) Lange, J. H. M.; Kruse, C. G.; Van
Stuivenberg, H. H. U.S. Pat. Appl. 2005187259. (d) Dahmani, S.; Paris, A.;
Jannier, V.; Hein, L.; Rouelle, D.; Scholtz, J.; Gressens, P.; Mantz, J.
Anesthesiology 2008, 108 (3), 457.
(5) (a) Shankar, P. R.; Subish, P. Pharm. ReV. 2007, 6 (31), 106. (b) Rayner,
B. Curr. Med. Res. Opin. 2004, 20 (3), 359. (c) van Zwieten, P. A. Am. J.
CardioVasc. Drugs 2001, 1 (5), 321. (d) Galley, G.; Groebke Zbinden, K.;
Hoener, M.; Kolczewski, S.; Norcross, R.; Stalder, H. WO Pat. Appl.
2007085557. (e) DeBernardis, J. F.; Kerkman, D. J.; Arendsen, D. L.; Buckner,
S. A.; Kyncl, J. J.; Hancock, A. A. J. Med. Chem. 1987, 30 (6), 1011. (f)
Biedermann, J.; Leon-Lomeli, A.; Borbe, H. O.; Prop, G. J. Med. Chem. 1986,
29 (7), 1183. (g) Schmitt, H. L.; Meyer, P.; Malen, C. E.; Dausse, J. P. R.;
Dabire, H.; Mouille, N. V. P. E. WO Pat. Appl. 8304025, 1983. (h) Wang, Z.;
Jonca, M.; Lambros, T.; Ferguson, S.; Goodnow, R. J. Pharm. Biomed. Anal.
2007, 45 (5), 720. (i) Geijo Caballero, F. A.; Prats Palacin, J.; Valles Plana,
J. M. ES Pat. Appl. 548382, 1986.
acid, and chromatographed to isolate both unreacted starting
material and product. The optical purities (% S,S) of each
component were then determined by chiral HPLC, and the
results were plotted as a function of conversion (Figure 1).
Both reaction components were observed to racemize and at
similar rates. We had initially suspected that the phosphine
borane functionality may have facilitated racemization if an
equilibrium existed between 2 and a species in which the borane
had transferred to the basic imidazoline nitrogen. There is clear
precedent for increased acidity of hydrogens at the R center in
such aminoboranes,8 yet in fact, the phosphine borane is not
(1) (a) Busacca, C. A.; Lorenz, J. C.; Grinberg, N.; Haddad, N.; Lee, H.; Li,
Z.; Liang, M.; Reeves, D.; Saha, A.; Varsolona, R.; Senanayake, C. H. Org.
Lett. 2008, 10 (2), 341. (b) Ramalingam, B.; Neuburger, M.; Pfaltz, A. Synthesis
2007, 4, 572. (c) Ma, K.; You, J. Chem.sEur. J. 2007, 13 (6), 1863. (d) Weiss,
M. E.; Fischer, D. F.; Xin, Z.-q.; Jautze, S.; Schweizer, W. B.; Peters, R. Angew.
Chem., Int. Ed. 2006, 45 (34), 5694. (e) Arai, T.; Mizukami, T.; Yokoyama, N.;
Nakazato, D.; Yanagisawa, A. Synlett 2005, 17, 2670. (f) Bhor, S.; Anilkumar,
G.; Tse, M. K.; Klawonn, M.; Doebler, C.; Bitterlich, B.; Grotevendt, A.; Beller,
M. Org. Lett. 2005, 7 (16), 3393. (g) Busacca, C. A.; Grossbach, D.; Campbell,
S. J.; Dong, Y.; Eriksson, M. C.; Harris, R. E.; James-Jones, P.; Kim, J.-Y.;
Lorenz, J. C.; McKellop, K. B.; O’Brien, E. M.; Qiu, F.; Simpson, R. D.; Smith,
L.; So, R. C.; Spinelli, E. M.; Vitous, J.; Zavattaro, C. J. Org. Chem. 2004, 69
(16), 5187. (h) Guiu, E.; Claver, C.; Benet-Buchholtz, J.; Castillon, S.
Tetrahedron: Asymmetry 2004, 15 (21), 3365. (i) Busacca, C. A.; Grossbach,
D.; So, R. C.; O’Brien, E. M.; Spinelli, E. M. Org. Lett. 2003, 5 (4), 595. (j)
Menges, F.; Neuburger, M.; Pfaltz, A. Org. Lett. 2002, 4 (26), 4713. (k)
Suginome, M.; Collet, S.; Ito, Y. Org. Lett. 2002, 4 (3), 351. (l) Busacca, C. A.
U.S. Patent 6 316 620, 2001. (m) Busacca, C. A.; Lorenz, J. C. Electronically
Tuned Ligands for Asymmetric Hydrogenation, WO2008067218.
(2) For reviews, see: (a) Dardonville, C.; Rozas, I. Med. Res. ReV. 2004, 24
(5), 639. (b) Head, G. A.; Mayorov, D. N. CardioVasc. Hematol. Agents 2006,
4 (1), 17. (c) Hudson, A. L.; Nutt, D. J.; Husbands, S. M. Pharm. News 2001,
8 (3), 26.
(3) (a) Buolamwini, J. K.; Addo, J.; Kamath, S.; Patil, S.; Mason, D.; Ores,
M. Curr. Cancer Drug Targets 2005, 5 (1), 57. (b) Vassilev, L. T.; Vu, B. T.;
Graves, B.; Carvajal, D.; Podlaski, F.; Filipovic, Z.; Kong, N.; Kammlott, U.;
Lukacs, C.; Klein, C.; Foutouhi, N.; Liu, E. Science 2004, 303 (5659), 844. (c)
Ding, Q.; Graves, B. J.; Kong, N.; Liu, J.-J.; Lovey, A. J.; Pizzolato, G.; Roberts,
J. L.; So, S.-S.; Vu, B. T.; Wovkulich, P. M. WO Pat. Appl. 2007063013. (d)
Tepe, J. J. US Pat. Appl. 20080114015.
(6) (a) Holt, A.; Smith, D. J.; Cendron, L.; Zanotti, G.; Rigo, A.; Di Paolo,
M. L. Mol. Pharmacol. 2008, 73 (2), 525. (b) Regunathan, S.; Feinstein, D. L.;
Reis, D. J. Ann. N.Y. Acad. Sci. 1999, 881, 410. (c) Tepe, J. J. WO Pat. Appl.
2007133790. (d) Dejovin, J. A.; Rossi, T. M. WO Pat. Appl. 2005115395. (e)
Ohshima, E.; Sone, H.; Kotera, O.; Luly, J. R.; Larosa, G. J. WO Pat. Appl.
2003087063.
(7) For a review, see: Khan, Z. P.; Ferguson, C. N.; Jones, R. M. Anaesthesia
1999, 54 (2), 146–165.
(8) For reviews, see: (a) Kessar, S. V.; Singh, P. Chem. ReV. 1997, 97, 721.
(b) Carboni, B.; Monnier, L. Tetrahedron 1999, 55 (5), 1197.
(9) Reichardt, C. SolVents and SolVent Effects in Organic Chemistry; Wiley-
VCH: Weinheim, 2003; pp 243-268.
(10) (a) Corey, E. J.; Kuhnle, N. M. Tetrahedron Lett. 1997, 38 (50), 8631.
(b) Braddock, D. C.; Redmond, J. M.; Hermitage, S. A.; White, A. J. P. AdV.
Synth. Catal. 2006, 348, 911. (c) Braddock, D. C.; Hermitage, S. A.; Redmond,
J. M.; White, A. J. P. Tetrahedron: Asymmetry 2006, 17, 2935. (d) Hunter, D. H.;
Sim, S. K. Can. J. Chem. 1972, 50, 669. (e) Hunter, D. H.; Sim, S. K. Can.
J. Chem. 1972, 50, 678. (f) Hunter, D. H.; Sim, S. K. J. Am. Chem. Soc. 1969,
91 (22), 6202. (g) Saigo, K.; Kubota, N.; Takebayashi, S.; Hasegawa, M. Bull.
Chem. Soc. Jpn. 1986, 59, 931.
(11) (a) Bates, R. B.; McCombs, D. A. Tetrahedron Lett. 1969, 10 (12),
977. (b) Shoppee, C. W.; Henderson, G. N. J. Chem. Soc., Chem. Commun.
1974, 14, 561. (c) Williams, D. R.; Reeves, J. T.; Nag, P. P.; Pitcock, W. H.;
Baik, M.-H. J. Am. Chem. Soc. 2006, 128 (37), 12339. (d) Habersaat, N.; Frohlich,
R.; Wurthwein, E.-U. Eur. J. Org. Chem. 2004, 12, 2567. (e) Williams, D. R.;
Reeves, J. T. J. Am. Chem. Soc. 2004, 126 (11), 3434. (f) Bates, R. B.; Brenner,
S.; Deines, W. H.; McCombs, D. A.; Potter, D. E. J. Am. Chem. Soc. 1970, 92
(21), 6345. (g) Braddock, D. C.; Redmond, J. M.; Hermitage, S. A.; White,
A. J. P. AdV. Synth. Catal. 2006, 348 (7 + 8), 911. (h) Laurent, M. A. Ann.
1837, 130. (i) Laurent, M. A. C. R. Acad. Sci. 1844, 19, 353.
J. Org. Chem. Vol. 73, No. 24, 2008 9757

Busacca et al.
SCHEME 3. Proposed Racemization Mechanism
SCHEME 2. N-Substituted Imidazolinesa
SCHEME 4. Route to DPEN, 7
a Key: (a) TsCl, TEA, DCM, 64%; (b) KHMDS, MeI, 67%.
Recent interest10a,c,g,11g in this venerable chemical reaction has
focused not on the imidazoline as a target but on the hydrolysis
of isoamarine to racemic 1,2-diphenylethylenediamine, 7 (DPEN,
Scheme 4), as an inexpensive entry to this building block and
its optical resolution. The kinetics and mechanism of formation
of amarine have been studied in detail,10d-f and those results
help drive our research as well. It was found that amarine, which
is thermodynamically less stable than 3, was formed under
thermal conditions from an intermediate anion and that 3 was
generated when the reaction was irradiated at 592 nm, the
absorbance maximum of that anion. These results were found
to be fully consistent with a thermally allowed disrotatory
cyclization to give amarine and a photochemically allowed
cyclization to give isoamarine.
This precedent certainly suggests that racemization of chiral
imidazolines in the current work, which clearly requires anion
formation to occur, likely involves the mechanistic reverse of
the forward process. We had not, however, determined whether
any of the meso-imidazoline 6 was generated in our racemization
experiments. We thus prepared 6 from meso-DPEN and
developed a chiral HPLC method capable of resolving (S,S)-3,
(R,R)-3, and meso-6. We then repeated the time-dependent
racemization experiment of Figure 2. No meso-6 was observed
at any time point, and spiking experiments showed that very
low levels of 6 (<1%) could readily be quantitated. This result
did not address the possibility that 6 had formed in the
racemization experiment yet isomerized to rac-3 too quickly
to be observed, however. It was also unclear whether the
isomerization process would be prevented by a N-substituent,
as was observed with racemization. To address the latter, we
alkylated meso-6 with MeI, as shown in Scheme 5. A 2.5:1
mixture of desired cis-8 to rac-5 was formed, showing the ease
with which isomerization can occur, even at ambient temper-
ature. Heptane trituration of this 2.5:1 mixture afforded a 93:7
mixture, favoring cis-8. We next subjected both 6 and 8 (as the
93:7 mixture) to the identical conditions used in the racemization
and plotted percent meso and cis as a function of time, as shown
in Figure 3. The racemization data for (S,S)-3 is included for
comparison purposes. It can be seen that the isomerization of 6
FIGURE 2. Racemization study of 3-5.
to further understand the mechanism of racemization, two
N-substituted chiral imidazolines, the sulfonamide 4 and methyl
species 5, were each prepared on one step (Scheme 2). Substrates
3-5 were then each independently subjected to the racemization
conditions of the parent using the phosphide borane anion at
60 °C. Sulfonamide 4 suffered from some sulfonyl cleavage
by phosphide borane under these conditions, yet the starting
material was easily reisolated chromatographically for optical
purity determination. The N-methyl species 5 did not degrade
under the racemization conditions. The results for all three
species are shown in Figure 2. Imidazoline 3 showed the
expected time-dependent racemization, similar to the results in
Figure 1. When the optical purities of isolated 4 and 5 were
measured, we found that these N-substituted species showed
no racemization under the identical conditions used for 3.
It was thus clear that introduction of any nitrogen substituent,
even an electron-withdrawing one, completely suppressed the
racemization process. Based on these results, we have therefore
proposed the following symmetry-allowed process as a possible
mechanism for imidazoline racemization (Scheme 3).
Deprotonation of imidazoline 3 by a strong base would
produce anion A, which could lead to ring opening, generating
diazadienyl anion B in a thermally allowed disrotatory process.
This species would then undergo a disrotatory closure to furnish
the observed mixture of enantiomers.10a,d-f,11 Racemic 3 is also
known as “isoamarine”, and its synthesis starting with benzal-
dehyde and ammonia and proceeding through the meso-
imidazoline 6 has been known for more than 150 years.11h,i
9758 J. Org. Chem. Vol. 73, No. 24, 2008

On the Racemization of Chiral Imidazolines
SCHEME 5. Synthesis of cis-8a
SCHEME 7. Crossover Experiment
a Key: KHMDS, MeI, 20 °C, 99%.
potential explanation for racemization can be imagined based
on the known syntheses of diamines via imine dimerization.
Alexakis,13 Smith,14 and Wotiz15 have utilized the coupling of
radical anions to make N,N′-disubstituted diamines from simple
imines with low valent metals. Critically, it has been observed14
that treatment of the dianions of these products, when the
diamines are unsymmetric, leads to the “crossover” products,
proving the intermediacy of radical anions. Any process such
as this, or the one depicted in Scheme 5, that leads to complete
scission of the imidazoline ring, should therefore be expected
to yield crossover products in an unsymmetrical system.
We have previously described the synthesis of optically pure,
non-C2-symmetric diamines and their incorporation into chiral
imidazolines.1g We therefore condensed one of these optically
pure diamines, (S,S)-11, with ethyl benzimidate to furnish (S,S)-
12 as a stable, crystalline solid, as shown in Scheme 7. Our
approach was to perform a crossover experiment with this
nonsymmetric imidazoline and search for the crossover products
3 and 13, the latter of which was readily prepared (one step)
as an authentic standard. We subjected (S,S)-12 to the
identical racemization conditions previously described. After
12 h at 60 °C with the phosphide borane anion, the optical
purity of (S,S)-12 had been reduced to 30% ee. Analysis of
the reaction mixture by both standard gradient C8 HPLC and
chiral HPLC showed that absolutely none of the crossover
imidazolines 3 or 13 were formed. These results, as well as
all of those reported herein, are consistent with the symmetry-
allowed process of Scheme 3.
FIGURE 3. Isomerization of 6 and 8.
SCHEME 6. Nitrile Ylide Formation
With a working mechanistic model in place, we next
examined the role of temperature on the racemization of
imidazoline 3 with phosphide borane. The optical purities at
six temperatures (25-65 °C) were determined under our
standard conditions with phosphide borane in DMSO at two
time points.
As shown in Table 1, there is an extreme temperature
dependence on the racemization. At 65 °C, only 7 h exposure
reduced the optical purity to 33%. When the reaction was held
constant just above ambient temperature, at 25 °C, however,
no decrease in optical purity was observed after 16 h. Opera-
from cis to trans with phosphide borane anion is more rapid
than the racemization process, yet it is far from instantaneous.
Fully 2 h was required to achieve complete isomerization. The
N-alkylated species 8 isomerized at a similar rate. Thus,
isomerization, unlike racemization, takes place whether a
N-substituent is present or not. Therefore, these two processes
are mechanistically distinct, and isomerization would not be
expected to follow the apparent symmetry-controlled process
of racemization. All of these results are therefore consistent with
racemization proceeding by the symmetry-allowed process
shown in Scheme 3.
Imidazolines have been prepared by cycloaddition of nitrile
ylides with imine derivatives.12 Fragmentation of the ring could
thus conceivably occur as shown in Scheme 6. Recombination
of fragments 9 and 10 might then lead to racemization. Another
(13) (a) Alexakis, A.; Aujard, I.; Mangeney, P. Synlett 1998, 873. (b)
Alexakis, A.; Aujard, I.; Mangeney, P. Synlett 1998, 875. (c) Alexakis, A.; Aujard,
I.; Kanger, T.; Mangeney, P.; Magomedov, N.; Hart, D. J. Org. Synth. 1999,
76, 23.
(14) Smith, J. G.; Ho, I. J. Org. Chem. 1972, 37 (4), 653.
(15) Wotiz, J. H.; Kleopfer, R. D.; Barelski, P. M.; Hinckley, C. C.; Koster,
D. F. J. Org. Chem. 1972, 37 (11), 1758.
(16) (a) Kim, H.; Nguyen, Y.; Yen, C. P.-H.; Chagal, L.; Lough, A. J.; Kim,
B. M.; Chin, J. J. Am. Chem. Soc. 2008, 130 (36), 12184. (b) Kim, H. J.; Kim,
W.; Lough, A. J.; Kim, B. M.; Chin, J. J. Am. Chem. Soc. 2005, 127 (48), 16776.
(12) (a) v. Bittner, G.; Witte, H.; Hesse, G. Liebigs Ann. Chem. 1968, 713,
1. (b) Bunge, K.; Huisgen, R.; Raab, R.; Sturm, H. J. Chem. Ber. 1972, 105 (4),
1307. For a closely related transformation, see: (c) Bowman, R. K.; Johnson,
J. S. J. Org. Chem. 2004, 69 (24), 8537.
J. Org. Chem. Vol. 73, No. 24, 2008 9759

Busacca et al.
TABLE 1. Racemization of 3, 25-65 °C
SCHEME 9. Utilization of Diamine Mixtures
entry
T (°C)
% ee
time (h)
1
2
3
4
5
6
25
45
50
55
60
65
99, 99
85, 77
72, 63
64, 53
47, 25
33, 10
7, 16
7, 16
7, 16
7, 16
7, 16
7, 16
with 2-fluorobenzaldehyde to generate the mixture of fluoro-
imidazolines meso-15 + rac-1 as shown in Scheme 9, using
the procedure of Fujioka and Kita.18 This mixture was then
reacted with the phosphideborane anion at 60 °C and the reaction
held at this temperature until substitution and isomerization were
both complete by HPLC (4.5 h). As expected, racemic phos-
phinoimidazoline 2 was produced in high overall yield as a
single geometric isomer. Optical resolution of this material by
a number of methods could then furnish optically pure
material.10b,c This methodology should prove particularly useful
when ligands derived from diamines that are not commercially
available, or are prohibitively expensive, are required, or when
both enantiomeric ligands are needed for catalyst development.
SCHEME 8. Ambient Temperature SNAr
tionally then, only brief exposures at temperatures above ∼30 °C
can be tolerated.
These results proved critical to our subsequent development
of a racemization-free SNAr reaction.1a After some experimenta-
tion, we found we could carry out our desired phosphide borane
SNAr at ambient temperature, provided 2 equiv of the phosphide
borane was employed (Scheme 8). In this way, more than 100 g
of phosphine borane 2 were produced with 0% racemization,
and SNAr reactions on many different substrates have now been
carried out successfully using this protocol, furnishing optically
pure products.
Conclusions
In summary, the unexpected base-promoted racemization of
optically pure imidazolines has been studied and found to
depend principally on the base used and the reaction temper-
ature. Amine bases cause no racemization, even at elevated
temperatures, while inorganic bases in dipolar aprotic solvents
such as DMSO and DMAc can lead to racemization. Chiral
imidazolines can still be manipulated without racemization under
these strongly basic conditions, provided that the temperature
is ∼30 °C or lower. Somewhat surprisingly, introduction of a
substituent on the imidazoline nitrogen atom acts to protect the
system from racemization, even at elevated temperature. These
and other results are consistent with a racemization mechanism
involving deprotonation followed by two symmetry-allowed
disrotatory processes. An understanding of these mechanisms
and the key process parameters surrounding racemization and
isomerization have led to to the development of robust,
racemization-free SNAr reactions to produce the BIPI ligands,
and these processes have been successfully demonstrated on
multigram scale.
One of the very significant advantages of the Alexakis13
diamine syntheses is that the initial 1:1 mixture of meso- and
rac-diamines produced is of no consequence, since the meso-
diamine is efficiently isomerized to the rac-diamine by the low-
valent metal. Simple resolution then provides the optically pure
diamine. Although we believe our imidazoline syntheses to be
mechanistically separate from the radical anion process, the
Alexakis work and our racemization and isomerization results
described here suggested an opportunity to effect a similar
transformation. Optically pure 1,2-diphenylethylenediamine
(DPEN) is readily available in bulk as both antipodes, and many
optically pure diamines are now commercially available via the
asymmetric diaza-Cope rearrangement of Chin.16 However,
many of these are not available in bulk, and more unusual chiral
diamines are often not commercially available. By contrast,
many diamines can be made as mixtures of meso- and rac- forms
by inexpensive imine dimerization, as mentioned.13a-c,17 We
therefore felt that employing a mixture of rac- and meso-
diamines in our imidazoline synthesis, followed by an SNAr
reaction that was deliberately performed at elevated temperature,
should provide the pure (racemic) phosphine borane product.
As a test of this concept, we therefore first condensed equal
quantities of meso-14 and rac-7 1,2-diphenylethylendiamines
Experimental Section
Imidazoline (S,S)-3. A 250 mL flask was charged with 7.40 g
of ethyl benzimidate (39.9 mmol, 1 equiv), 8.43 g of (S,S)-diamine
DPEN (39.9 mmol, 1 eqiuv), and 50 mL of abs EtOH in the order
given. The resulting slurry was stirred 1 h at rt and then heated to
reflux. After 1 h, the mixture was cooled to rt, and then EtOH was
removed in vacuo. To the resulting residue was added 100 mL of
1 N NaOH + 200 mL of EtOAc and the mixture stirred well for 5
min. The organic phase was washed with H2O (1 × 100 mL) and
then concentrated in vacuo to a light yellow solid. This solid was
recrystallized from ∼75 mL of boiling EtOAc. On cooling while
(17) For examples, see: (a) v. Betschart, C.; Schmidt, B.; Seebach, D. HelV.
Chim. Acta 1988, 71, 1999. (b) Kise, N.; Ueda, N. Tetrahedron Lett. 2001, 42,
2365. For reviews, see: (c) Lucet, D.; Le Gall, T.; Mioskowski, C. Angew. Chem.,
Int. Ed. 1998, 37 (19), 2580. (d) Saibabu Kotti, S. R. S.; Timmons, C.; Li, G.
Chem. Biol. Drug Des. 2006, 67, 101.
(18) (a) Fujioka, H.; Murai, K.; Kubo, O.; Ohba, Y.; Kita, Y. Tetrahedron
2007, 63 (3), 638. (b) Fujioka, H.; Murai, K.; Ohba, Y.; Hiramatsu, A.; Kita, Y.
Tetrahedron Lett. 2005, 46 (13), 2197.
9760 J. Org. Chem. Vol. 73, No. 24, 2008

On the Racemization of Chiral Imidazolines
standing, colorless needles were deposited. The solid was filtered,
and air-dried on the frit for 20 min to give 9.10 g of imidazoline
3 (76% first crop yield) as a white, crystalline solid: mp 176-177
(d), 127.1 (d), 127.0 (d), 126.9 (d), 78.6 (d), 77.8 (d), 34.9 (q);
HRMS [M + H]+ calcd for C22H20N2 313.1699, obsd 313.1698.
Racemization Studies on (S,S)-3, (S,S)-4, and (S,S)-5 (Figure 2). A
three-neck 100 mL flask containing a stir bar and equipped with a
closable inert gas valve was charged with 467 mg of dicyclohexyl-
phosphine borane (2.20 mmol, 2.2 equiv). The flask was then
evacuated and Ar filled (3×), and then the gas valve was closed
and the flask transferred to a N2 glovebox. The flask was then
charged with 248 mg of t-BuOK (2.20 mmol, 2.2 equiv), sealed,
and removed from the glovebox. The flask was again evacuated
and Ar filled (3×), and then 5.0 mL of DMSO was added via
syringe. The resultant mixture was stirred for 30 min at ambient
temperature, and then 298 mg of imidazoline (S,S)-3 (1.00 mmol,
1 equiv), contained in a 1 dram vial, was added neat, at once, by
quickly removing a septum. The flask was again evacuated and Ar
filled (3×), then placed in a pre-equilibrated 60 °C oil bath under
Ar. Aliquots (100 µL) were periodically removed via syringe and
quenched into 1 mL 5 M NH4Cl in 1 dram screw-cap vials. This
mixture was then extracted with ∼2 mL of MTBE by brief shaking.
The MTBE phase was then analyzed directly by chiral HPLC.
The identical stoichiometry, reaction concentration, and temper-
ature were employed in the racemization studies using (S,S)-4 and
(S,S)-5. For the study with (S,S)-5, however, some sulfonamide
hydrolysis occurred during the reaction. For this reason, further
purification prior to the chiral HPLC analyses was required: The
MTBE solution containing 5 was concentrated on the rotovap and
then spotted heavily across the bottom of a standard 5 × 7 cm
TLC plate and eluted in a TLC chamber using 1:1 hexane/EtOAc,
like a “mini prep plate”. The upper band (Rf 0.62), contained
sulfonamide 5, while the lower band (Rf 0.14) contained the
cleavage product 3. When the plate was dry, the upper band was
scraped off with a razor blade and the silica extracted with ∼1 mL
of MeOH by agitating in a 1 dram vial on a platform vibrator for
several min. The resultant mixture was then filtered using a 0.5
µm syringe tip filter directly into an HPLC vial for analysis.
Although dilute, chiral HPLC analyses were readily made by
increasing the injection volume to 50 µL.
Chiral HPLC analysis of 3: AGP, 4 × 150 mm, 5 µm particle
size, 30 °C, 0.9 mL/min, isocratic with 80:20 A:B, where A )
0.1% HOAc in H2O, buffered to pH 7 with NH4OH using a digital
pH meter, B ) MeCN. Analysis wavelength 224 nm, 2 µL injection
volume, retention times: (R,R)-3: 9.18 min, (S,S)-3: 10.86 min,
meso-6: 10.15 min.
Chiral HPLC analysis of 4: Chiralpak IA, 4.6 × 250 mm, 5 µm
particle size, 20 °C, 1.0 mL/min, isocratic with 75:25 n-heptane/
i-PrOH, analysis wavelength 220 nm, 5 µL injection volume,
retention times: (S,S)-4: 9.24 min, (R,R)-4: 10.10 min.
Chiral HPLC analysis of 5: Chiralpak AD-H, 4.6 × 250 mm, 5
µm particle size, 25 °C, 1.5 mL/min, isocratic with 90:10 A:B,
where A ) heptane, B ) n-PrOH. Analysis wavelength 254 nm,
5.0 µL injection volume, retention times: (R,R)-5: 4.64 min, (S,S)-
5: 5.69 min.
°C; [R]20 -31 (c 0.16, CH2Cl2); chiral HPLC (see below) shows
D
>99% ee; 1H (500 MHz, CDCl3) δ 7.95 (d, J ) 7.2 Hz, 2H), 7.72
(m, 1H), 7.46 (t, J ) 7.7 Hz, 2H), 7.36-7.39 (m, 4H), 7.03-7.33
(m, 6H), 5.50 (br s, 1H), 4.91 (s, 2H); 13C NMR (125 MHz, CDCl3)
δ 163.0 (s), 143.5 (s), 130.9 (d), 130.1 (s), 128.6 (d), 128.5 (d),
127.41(d), 127.35 (d), 126.5 (d), 74.4 (d); HRMS [M + H]+ calcd
for C21H18N2 299.1542, obsd 299.1550.
N-Tosylimidazoline (S,S)-4. A three-neck 50 mL flask was
charged with 1.00 g of (S,S)-3 (3.35 mmol, 1 equiv), 639 mg of
TsCl (3.35 mmol, 1 equiv), and 8.0 mL of CH2Cl2 in the order
given. The resulting slurry was cooled to ∼5 °C, then 0.81 mL of
TEA (6.70 mmol, 2 equiv) was added via syringe over ∼1 min,
giving a thinner slurry. The bath was then removed, and the mixture
allowed to warm to ambient temperature. After 4 h, HPLC showed
7% of 3 remained, so an additional 50 mg of TsCl was added.
After 1 h, the volatiles were removed in vacuo and the residue
partitioned between 50 mL of EtOAc and 25 mL of 0.5 N HCl.
The organic phase was washed with satd NaCl (1 × 30 mL) and
dried (MgSO4), and the solvents were removed in vacuo to give a
yellow oil. After standing for 24 h, the product crystallized. The
solid was then recrystallized from boiling EtOAc. The solid thus
obtained was filtered and air-dried on the frit for ∼20 min to give
0.96 g of (S,S)-4 (64% recrystallized yield) as a white solid: mp
136-137 °C; [R]20 +151 (c 0.25, CH2Cl2); chiral HPLC (see
D
below) shows >99% ee; 1H (500 MHz, CDCl3) δ 7.88 (d, J ) 7.2
Hz, 2H), 7.56 (t, J ) 7.5 Hz, 1H), 7.48 (m, 6H), 7.41 (m, 1H),
7.26 (m, 3H), 7.20 (t, J ) 7.7 Hz, 2H), 7.09 (d, J ) 8.2 Hz, 2H),
6.86 (d, J ) 7.3 Hz, 2H), 5.15 (d, J ) 4.5 Hz, 1H), 5.13 (d, J )
4.5 Hz, 1H), 2.42 (s, 3H); 13C NMR (125 MHz, CDCl3) δ 159.6
(s), 144.5 (s), 142.4 (s), 141.7 (s), 134.9 (s), 131.1 (d), 130.5 (s),
129.8 (d), 129.5 (d), 129.1 (d), 128.6 (d), 128.2 (d), 127.8 (d), 127.7
(d), 127.3 (d), 126.0 (d), 125.9 (d), 77.3 (d), 72.4 (d), 21.5 (q);
HRMS [M + H]+ calcd for C28H24N2O2S 453.1631, obsd 453.1649.
N-Methylimidazoline (S,S)-5. A three-neck 100 mL flask was
charged with 1.00 g of (S,S)-imidazoline 3 (3.36 mmol, 1 equiv)
and then evacuated and Ar filled. Dry THF (15 mL) was then added
via syringe, and the resulting solution was cooled to 5 °C under
Ar. To this was added 7.06 mL of 0.5 M KHMDS/PhMe (3.53
mmol, 1.05 equiv) dropwise via syringe over ∼5 min. A thick slurry
formed near the end of the addition. The bath was removed, and
the mixture allowed to warm to rt. After 1.5 h at rt, a yellow slurry
was present. MeI (0.22 mL, 3.53 mmol, 1.05 equiv) was added at
once via syringe, causing a white slurry to form. After 2 h at rt,
TLC (4:1 EtOAc/hexane) showed the reaction was complete to a
more polar streak. The volatiles were then removed in vacuo under
house vacuum to give a semisolid. These solids were then suspended
in 25 mL of EtOAc, stirred vigorously for 5 min, and then filtered
through a Celite pad to remove KI. The filtrate was evaporated
under high vacuum to give 1.09 g of a yellow oil. This oil was
chromatographed on silica gel eluting with EtOAc and then 10%
MeOH/EtOAc to give 0.70 g of 5 (67%) as a yellow oil, which
Acknowledgment. We thank Professor Scott Denmark,
University of Illinois, for very helpful discussions.
crystallized on standing: mp 60-61 °C; [R]20 +55 (c 0.14,
D
CH2Cl2); chiral HPLC (see below) shows >99% ee; 1H (500 MHz,
CDCl3) δ 7.79 (m, 2H), 7.50 (m, 3H), 7.31-7.43 (m, 10H), 5.01
(d, J ) 9.8 Hz, 1H), 4.30 (d, J ) 9.8 Hz, 1H), 2.77 (s, 3H); 13C
NMR (125 MHz, CDCl3) δ 167.0 (s), 143.8 (s), 141.5 (s), 131.3
(s), 130.0 (d), 128.8 (d), 128.5 (d), 128.41 (d), 128.36 (d), 127.8
Supporting Information Available: Experimental proce-
dures and characterization data for all compounds. This material
is available free of charge via the Internet at http://pubs.acs.org.
JO8024384
J. Org. Chem. Vol. 73, No. 24, 2008 9761