Dynamic biphasic counterion exchange in a configurationally stable aziridinium ion: Efficient synthesis and isolation of a koga C 2-symmetric tetraamine base

Article abstract of DOI:10.1021/op0602371

An efficient synthetic process for chiral tetraamine base (R,R)-5 is reported that leverages mechanistic understanding to enable control over key transformations. Specifically, a configurationally stable and observable aziridinium ion intermediate was found to undergo counterfoil exchange impacting the feasibility of the process. Mechanistic investigations revealed that both counterion exchange and trapping of the aziridinium ion were biphasic events and that the former could be suppressed at lower temperatures, facilitating the reaction on both small and large scales. The mechanistic insights gained have led to the development of an efficient one-pot process that enables preparation of multiple kilograms of (R,R)-5 as a crystalline solid without chromatography in excellent chemical and chiral purity.

Full text of DOI:10.1021/op0602371

Organic Process Research & Development 2007, 11, 215−222
Dynamic Biphasic Counterion Exchange in a Configurationally Stable
Aziridinium Ion: Efficient Synthesis and Isolation of a Koga C2-Symmetric
Tetraamine Base
Matthew J. Frizzle,† Sebastien Caille,‡ Teresa L. Marshall,† Kenneth McRae,‡ Kelly Nadeau,† Gary Guo,§ Steven Wu,§
Michael J. Martinelli,‡ and George A. Moniz*,†
Chemical Process R&D, Amgen, Inc., One Kendall Square, Building 1000, Cambridge, Massachusetts 02139, U.S.A.,
Chemical Process R&D, Amgen, Inc., One Amgen Center DriVe, Thousand Oaks, California 91320, U.S.A., and
Analytical Chemistry, Amgen, Inc., One Amgen Center DriVe, Thousand Oaks, California 91320, U.S.A.
Scheme 1
Abstract:
An efficient synthetic process for chiral tetraamine base (R,R)-5
is reported that leverages mechanistic understanding to enable
control over key transformations. Specifically, a configuration-
ally stable and observable aziridinium ion intermediate was
found to undergo counterion exchange impacting the feasibility
of the process. Mechanistic investigations revealed that both
counterion exchange and trapping of the aziridinium ion were
biphasic events and that the former could be suppressed at
lower temperatures, facilitating the reaction on both small and
large scales. The mechanistic insights gained have led to the
development of an efficient one-pot process that enables
preparation of multiple kilograms of (R,R)-5 as a crystalline
solid without chromatography in excellent chemical and chiral
purity.
In 1986, Koga et al. reported the use of chiral lithium
amide bases (2a-d, Scheme 1) for the synthesis of chiral
enolates from prochiral ketones via enantiotopic proton
abstraction.1 In a subsequent report, Koga demonstrated that
a related chiral base ((R,R)-5) could also direct the approach
of an electrophile to prochiral enolates (4a and 4b) with the
resulting alkylated R-tetralones (6a and 6b) being isolated
in high enantioselectivity.2,3
As part of an ongoing development program employing
such an enantioselective alkylation, we required a means of
preparing multi-kilogram quantities of tetraamine (R,R)-5.
In 2001, O’Brien reported a general method for synthesis of
Koga’s chiral amines (Scheme 2).4 The procedure involved
thermal ring-opening of the appropriate antipode of styrene
oxide (7) with piperidine to afford aminoalcohols 8a and
8b, which were isolated and then treated sequentially with
mesyl chloride and the appropriate amine in diethyl ether to
afford the desired Koga amine. Interestingly, displacement
of the corresponding mesylates (9a and 9b) was found to
proceed at the benzylic carbon only and with net retention
of stereochemistry, suggesting the intermediacy of aziri-
dinium ion 10. An acid/base extraction protocol afforded the
desired Koga amines (e.g., (R,R)-5) with good purity. Given
the apparent ease of this two-step protocol, we used this
report as the starting point for our own investigations.
* To whom correspondence should be addressed. E-mail: gmoniz@amgen.com.
† Chemical Process R&D, Massachusetts.
‡ Chemical Process R&D, California.
§ Analytical Chemistry.
(1) Shirai, R.; Tanaka, M.; Koga, K. J. Am. Chem. Soc. 1986, 108, 543.
(2) (a) Murakata, M.; Nakajima, M.; Koga, K. J. Chem. Soc., Chem. Commun.
1990, 1657. (b) Imai, M.; Hagihara, A.; Kawasaki, H.; Manabe, K.; Koga,
K. J. Am. Chem. Soc. 1994, 116, 8829. (c) Yamashita, Y.; Odashima, K.;
Koga, K. Tetrahedron Lett. 1999, 40, 2803.
(3) For reviews of asymmetric synthesis using chiral lithium amide bases, see:
(a) Cox, P. J.; Simpkins, N. S. Tetrahedron: Asymmetry 1991, 2, 1. (b)
O’Brien, P. J. Chem. Soc., Perkin. Trans. 1 1998, 1439. (c) Tomioka, K.
Synthesis 1990, 541. (d) Majewski, M. AdV. Asymm. Synth. 1998, 3, 39.
(4) (a) Poumellec, P.; O’Brien, P. Tetrahedron Lett. 1996, 37, 5619. (b) deSousa,
S. E.; O’Brien, P.; Poumelec, P. J. Chem. Soc., Perkin Trans. 1 1998, 1483.
(c) Cuthbertson, E.; O’Brien, P.; Towers, T. Synthesis 2001, 693. (d) Towers,
T. D.; O’Brien, P. J. Org. Chem. 2002, 67, 304.
Results and Discussion
In our hands, the reported procedure by O’Brien was
functional. Ring-opening of (R)-7 afforded a near 1:1 ratio
of aminoalcohols 8a and 8b, which were carried through
10.1021/op0602371 CCC: $37.00 © 2007 American Chemical Society
Published on Web 02/20/2007
Vol. 11, No. 2, 2007 / Organic Process Research & Development
•
215

Scheme 2
and chemistry, and thus a means of maintaining agitation
throughout the entire reaction would be necessary.
(2) Although reported by both Koga and O’Brien to be
a colorless oil, Koga amine (R,R)-5 solidified upon stand-
ing to a hard, amorphous solid that was very difficult to
break up. To ease isolation and handling, a means of crystal-
lizing (R,R)-5 to a flowing, manageable solid would be
required.
Our efforts to resolve the agitator seizure problem initially
focused on characterizing the reactivity that was the source
of the stirring difficulty. Isolation and characterization of the
thick, gummy precipitate revealed its identity to be Et3NH‚
OMs. This suggested that formation of aziridinium ion 10
did not take place gradually, but rather was a rapid event at
10-20 °C, and that the mesylate anion liberated by the
aziridinium formation underwent counterion exchange with
Et3N‚HCl. To confirm that such a counterion exchange event
was operative under these conditions, an experiment was
performed wherein (R)-8a and (S)-8b were treated with 1.1
equiv of MsCl at -10 °C. The supernatant reaction solution
was sampled following complete MsCl addition at -10 °C
and at intervals while warming to 22 °C, and the resulting
aliquots were analyzed rapidly by 1H NMR. As can be seen
from Figure 1, initially both regioisomeric mesylates 9a and
9b are observed (Spectrum A in Figure 1). As the reaction
progresses, both sets of mesylate signals are observed to
converge to a single isomeric â-chloroamine 12 (Spectrum
B), which is ultimately the sole constituent of the supernatant
solution (Spectrum C). This outcome revealed not only that
aziridinium ion 10a did undergo mesylate-to-chloride coun-
terion exchange but also that the chloride counterion was
sufficiently nucleophilic to attack the aziridinium ion at the
more reactive benzylic position, alleviating ring strain and
establishing a covalent bond (Scheme 3). When held at -10
°C, mesylates 9a and 9b remained as stable covalent species
with no chloride 12 visible even after 3 h, demonstrating
that conversion of 9a/9b to 12 via aziridinium ion 10a could
be suppressed at lower temperatures.
Further insight into the mechanism and driving force for
the counterion exchange was gained by evaluating the
solubilities of the reacting partners in the MTBE phase. As
can be seen from Table 1, these studies revealed that, while
â-chloroamine 12 was soluble in MTBE to 115 mg/mL,
aziridinium mesylate 10a was only soluble to 5 mg/mL9 and
Et3N‚HCl was soluble to less than 0.1 mg/mL. These data
point to a biphasic process wherein counterion exchange
occurs upon precipitation of 10a and â-chloroamine 12 is
re-extracted into the MTBE phase (Scheme 4). Importantly,
the equilibrium between covalently bound mesylates 9a/9b
and 10a was determined to lie extensively in favor of 10a
at room temperature when 9a and 9b were prepared using
methanesulfonic anhydride (to eliminate the possibility of
counterion exchange).9 This equilibrium, coupled with the
solubility-driven equilibrium between the aziridinium chlo-
the mesylation step at 0 °C in MTBE. The mixture was
warmed to room temperature and then treated with 1,3-
propanediamine (11) and water. Koga amine (R,R)-5 was
isolated in 99% chiral purity, lending further credence to
the intermediacy of aziridinium ion 10.5-7 There were,
however, two noteworthy observations made during these
initial experiments that would impact future efforts to scale
the synthesis of (R,R)-5 via the current procedure:
(1) As the reaction mixture was allowed to warm above
10 °C following mesylation and prior to charging 11, a
significant change in the reaction mixture consistency was
observed. Specifically, the white, flocculent precipitate of
Et3N‚HCl generated during the mesylation reaction, if the
mixture were allowed to warm to ca. 10 °C, changed rapidly
to a thick, gummy material that quickly settled to the bottom
of the reactor and seized the agitator.8 This stirring difficulty
was observed in a number of ether solvents as well as in
aromatic hydrocarbon solvents such as benzene and toluene.
Addition of 1,3-diaminopropane (11) was followed by
addition of water, which dissolved the gum and enabled
stirring to be resumed. However, on larger scales, even
temporary agitator seizure could be damaging to equipment
(5) For examples of aziridinium ions in synthesis, see: (a) Liu, Q.; Marchington,
A. P.; Rayner, C. M. Tetrahedron 1997, 53, 15729. (b) Rayner C. M. Synlett
1997, 11. (c) Okuda, M.; Tomioka, K. Tetrahedron Lett. 1994, 35, 4585.
(d) Gmeiner, P.; Junge, D.; Ka¨rtner, A. J. Org. Chem. 1994, 59, 6766. (e)
Freedman, J.; Vaal, M. J.; Huber, E. W. J. Org. Chem. 1991, 56, 670. (f)
Williams, D. R.; Brown, D. L.; Benbow, J. W. J. Am. Chem. Soc. 1989,
111, 1923. (g) Rosen, T.; Fesik, S. W.; Chu, D. T. W.; Pernet, A. G.
Synthesis 1988, 40. (h) Kametani, T.; Honda, T. AdV. Heterocycl. Chem.
1986, 39, 181. (i) Couterier, C.; Blanchet, J.; Schlama, T.; Zhu, J. Org.
Lett. 2006, 8, 2183.
(6) Chiral purity is reported relative to both the opposite enantiomer ((S,S)-5)
and the meso isomer (meso-5). Analytical standards of both (S,S)-5 and
meso-5 were obtained.
(7) For a review of Neighboring Group Participation in organic reactions, see:
Capon, B.; McManus, S. P. Neighboring Group Participation; Plenum
Press: New York, 1976.
(8) O’Brien also reported stirring difficulty at lower temperatures due to the
formation of Et3N‚HCl. Using mechanical agitation, stirring was easily
maintained with the initial precipitate of Et3N‚HCl. The pronounced change
in the form of the precipitate on warming was the event that seized the
agitator in our case.
(9) An authentic 1H NMR spectrum of aziridinium ion 10a is available as
Supporting Information. For NMR data of an unrelated aziridinium ion,
see: Liu, Q.; Marchington, A. P.; Boden, N.; Rayner, C. M. J. Chem. Soc.,
Perkin Trans. 1 1997, 511.
216
•
Vol. 11, No. 2, 2007 / Organic Process Research & Development

1
Figure 1. Conversion of mesylates 9a and 9b to â-chloroamine 12 as observed by H NMR (CDCl3).
ride 10b and 12, presumably drives the counterion exchange
towards the chloride.10
dinium ion formation, maintaining the covalent mesylate
intermediates 9a and 9b. Once mesylation was complete,
1,3-diaminopropane (11) and water would be added and the
mixture warmed only once all reacting components were
present. Key to the success of this approach would be the
rapid formation and subsequent trapping of the incipient
aziridinium ion 10 by 1,3-diaminopropane (11). However,
it was recognized that one potential liability of this approach
might be competing reactivities between the aziridinium ion
(10) and mesylates (9a and 9b) which would lead to mixtures
of regioisomers and stereoisomers.
Having characterized the counterion exchange process,
efforts focused on maintaining agitation throughout the
mesylation and subsequent aziridinium formation reactions.
Given that the aziridinium ion formation with subsequent
conversion to chloroamine 12 was observed to occur upon
warming, studies focused on maintaining the temperature in
the range -10 to -8 °C during the mesylation step to
maintain stirrability. It was expected based on the above
NMR studies that lower temperatures would suppress aziri-
To assess the feasibility of this approach, a study of the
composition of the biphasic reaction mixture was carried out
(Table 2). These studies revealed that 1,3-diaminopropane
(11) was located in the lower aqueous phase within the limits
of detection. Monoadduct 13 also had substantial concentra-
tion in the aqueous phase, while the product Koga amine
(5) was found to reside almost exclusively in the upper
(10) The intermediate â-chloroamine 12 derived from mesylates (R)-9a and (S)-
9b via aziridinium ion 10a gives rise to (R,R)-5 under the reaction conditions
reported by O’Brien. Thus, reversion of 12 to the aziridinium ion prior to
amine trapping is required for stereochemical competence. The formation
of 12 from 9a/9b has been observed previously; however neither the time
course and temperature sensitivity of this interconversion nor its mechanism
of formation were reported. See: Anderson, S. R.; Ayers, J. T.; DeVries,
K. M.; Ito, F.; Mendenhall, D.; Vanderplas, B. C. Tetrahedron: Asymmetry
1999, 10, 2655.
Vol. 11, No. 2, 2007 / Organic Process Research & Development
•
217

Scheme 3
Scheme 4
MTBE phase. These data suggested that aziridinium trapping
would occur principally in the aqueous phase where 1,3-
diaminopropane (11) resides and monoadduct 13 concentra-
tions are significant. The final Koga amine product is then
extracted back into the MTBE layer. Given such a mecha-
nism, displacement of the mesylate by 11 in the organic
phase would not be competitive.
In the event, mesylation proceeded smoothly at -10 °C,
preserving the white, flocculent precipitate of Et3N‚HCl
throughout the mesylation process. Once complete, 1,3-
diaminopropane (11) and water were added and the biphasic
mixture was warmed to 25 °C. Proceeding through the acid/
base workup utilized thus far afforded Koga amine (R,R)-5
in equivalent yield and chemical and chiral purity to those
obtained previously.
amine (R,R)-5 as a flowing, filterable solid was essential.
Current batches were isolated by rotary evaporation as a
thick, yellow oil that slowly solidified into a very hard solid
mass. Crystallization studies focused on solvent/antisolvent
combinations. Several combinations were screened, the
majority of which led to the precipitation of an oily species.
Ultimately, a protocol was established wherein crude (R,R)-5
was dissolved in IPA (5 mL/gram of crude solid). The
resulting solution was cooled to 0 °C, and the cold solution
was seeded with pure (R,R)-5 (0.1% w/w). This resulted in
the formation of a sizable seed bed to which water (0.9 ×
IPA volume) was charged.11 The formation of a seed bed
prior to antisolvent charge was critical to obtaining a
crystalline solid rather than an oil. The thick product slurry
was warmed to room temperature, filtered, and washed with
1.1:1 IPA/water to afford (R,R)-5 as a white, crystalline solid
in 99.9% chiral purity and 99.7 A% chemical purity by
HPLC.12
Having gained control over the counterion exchange
process that led to the agitator seizure issue, attention was
turned towards controlling the form of the final product. To
facilitate isolation on-scale, a means of crystallizing Koga
Table 1. Solubility of reacting species in MTBE phase at
22 °C
Analysis of the mother liquor by GC/MS revealed that
no (R,R)-5 remained in the mother liquor following crystal-
lization (see Figure 3). This was further confirmed by a
spiking experiment wherein 1c was added to the mother
liquor prior to GC/MS analysis. Identification of the major
(11) Water was charged portionwise. An exotherm of ca. 9 °C was observed
following each addition, and the mixture temperature was allowed to return
to 0 °C between additions.
(12) A% refers to the HPLC peak area % of the desired material versus all other
visible impurity peaks in the appropriate assay. Chiral purity refers to the
HPLC peak area % of the desired isomer (i.e., (R,R)-5) versus the enantiomer
((S,S)-5) and the meso-isomer (meso-5).
218
•
Vol. 11, No. 2, 2007 / Organic Process Research & Development

Table 2. Species partitioning between MTBE and aqueous phasesa
a Layer content determined by gas chromatography. See Supporting Information for details. b ND ) not detected.
Figure 2. GC/MS traces of crystallization mother liquor and mother liquor spiked with (R,R)-5.
impurities was also accomplished, and their structures are
indicated in Figure 2.
Efforts next focused on merging the two steps into a one-
pot protocol. Scheme 5 describes the through-process that
Vol. 11, No. 2, 2007 / Organic Process Research & Development
•
219

Scheme 5
Table 3. Change in concentration of (R,R)-5 in MTBE phase
versus time
concentration
of (R,R)-5
(mg/mL)
concentration
change/time
(mg/mL/h)
time (h)
1.5
2.5
3.5
4.5
21
64.5
69.0
70.8
71.9
79.8
79.2
0.0
4.5
1.8
1.1
0.48
-0.24
23.5
Figure 3. Plot of concentration of (R,R)-5 in MTBE phase
versus time
in the aqueous phase and the product is extracted into the
organic phase, obtaining a representative sample of the
reaction mixture that would contain consistent levels of
starting materials and product was not possible. To monitor
the progress of the reaction, we developed a protocol wherein
the concentration of (R,R)-5 in the organic phase was
monitored. The endpoint of the reaction was signaled by a
plateau in product concentration in the MTBE phase.
Concentration data collected during a representative 1 kg
reaction and the corresponding plot of concentration versus
time are provided in Table 3 and Figure 3, respectively.
Once complete, the reaction mixture was allowed to settle
to two phases and subjected to an aqueous workup consisting
of an ammonium chloride wash and brine wash. Solvent
exchange to IPA proceeded smoothly in two rounds of
addition and distillation. Residual solvent analysis revealed
that the resulting IPA solution contained an undetectable level
of MTBE and an acceptable level (0.01 wt %) of toluene.
Cooling to 0 °C and seeding (0.1 wt %) resulted in the
expected seed bed to which was charged water (4.5×) to
accomplish complete crystallization. Filtration afforded a
70% overall yield of (R,R)-5 as a white, crystalline solid
(99.9 wt %, 99.8 A% chiral purity).
resulted from these efforts. Conversion of (R)-Styrene oxide
((R)-7) to aminoalcohols 8a and 8b proceeded smoothly at
80 °C in 3 h. To ensure the success of the subsequent
mesylation step, thorough ethanol removal would need to
be accomplished. Thus, a solvent exchange protocol was
performed by first distilling the bulk of the ethanol and then
performing two azeotropic distillations with toluene. Fol-
lowing the second toluene distillation, MTBE was added.
Analysis of the resulting solution for residual solvents
revealed very low (0.01 wt %) ethanol content and acceptable
(2.6 wt %) toluene content. The mesylation step was
conducted at -10 °C to preserve the mesylate and suppress
aziridinium ion formation in accord with earlier experiments
(see above). As expected, mesylation proceeded smoothly
with no agitator interference due to the formation of
Et3NH‚OMs. Warming of the reaction mixture was
performed only after both diamine 11 and water had been
added.
In conclusion, we have developed an efficient one pot
process for the synthesis of Koga tetraamine (R,R)-5 as a
crystalline solid in excellent chemical and chiral purity.
Importantly, the developed process leverages mechanistic
understanding regarding the intermediacy of aziridinium ion
10a and the counterion exchange process that takes place
with this configurationally stable and observable intermediate
to enable efficient control over both of these events.
The biphasic nature of the aziridinium-opening/Koga
amine-formation step (i.e., 10f5, Scheme 5) posed an
analytical challenge in determining the degree of reaction
completion. Since the bulk of the reaction appears to proceed
220
•
Vol. 11, No. 2, 2007 / Organic Process Research & Development

Experimental Section
CAUTION: This is an exothermic reaction. Do not exceed
an internal reaction temperature of -10 °C to -8 °C during
this addition to maintain adequate stirring.
Materials and Methods. The vessel used was a 50 L
nonjacketed glass reactor containing an internal Teflon-
coated copper coil connected to a circulating chiller (operat-
ing range -60 to 150 °C). The 7-port reactor head with a 4
in. reagent charging port was fitted with a high efficiency
condenser (operating at -5 °C), a temperature probe, a nitro-
gen gas inlet, an off-gas outlet connected to an oil bubbler,
and a glass stir-shaft equipped with two paddles. The reactor
was verified clean prior to operation and was placed under
an atmosphere of nitrogen gas for 15 min prior to charging.
All commercially obtained reagents were used as received.
High performance liquid chromatography (HPLC) was
performed using Agilent 1100 systems. Gas chromatography
After the addition was complete the batch was stirred for
30 min at -10 °C. Adequate agitation was maintained by
keeping the temperature at or near -10 °C. An analytical
sample was pulled, and analysis indicated 98.6% conversion
of aminoalcohols 8a and 8b to mesylates 9a and 9b. Next,
1,3-diaminopropane (11, 354 mL, 4.2 mol, 0.50 equiv) was
charged from a clean self-equilibrating glass dropping funnel
followed by deionized water (4.1 L). The reaction mixture
was warmed to 20 °C over 1 h, during which time the
precipitated triethylamine salt dissolved. The mixture was
stirred at 20 °C for 18 h. Analytical samples were pulled at
1, 2, 5, and 18 h time points by halting agitation, allowing
the layers to settle, and sampling the upper (MTBE) layer.
Agitation was then reinitiated. HPLC analysis at each time
point provided the data in Table 3. Once deemed complete,
stirring was halted to allow phase separation (ca. 10 min).
The lower aqueous phase was drained. The remaining organic
fraction was washed with saturated ammonium chloride
solution (3.5 L) by agitating for 20 min followed by phase
separation. The lower aqueous layer was drained. The
remaining organic fraction was washed with brine (3.5 L)
by agitating for 20 min followed by phase separation. The
lower aqueous layer was drained. The reactor containing the
organic phase was fitted with a batch concentration receiver
flask. The reactor was placed under a vacuum (ca. 20 Torr),
and the internal temperature was increased to 30 °C. Distillate
was collected, and the reaction mixture was concentrated to
a total volume of ca. 3.5-4.0 L. Isopropyl alcohol (5.0 L)
was charged to the reactor, and the mixture stirred to
homogeneity. A sample was pulled to determine the residual
solvent levels which showed residual MTBE at 3.33%
(Pass: e 0.1%) and residual toluene at 0.13% (Pass: e
0.1%). The reactor was again placed under a house vacuum
(ca. 20 Torr), and the internal temperature was increased to
50 °C. Some forerun was observed at distillate temperature
29 °C (MTBE). Once the forerun ceased, distillate was
collected and the batch was concentrated to a volume of ca.
3.5 L (semiviscous oil). The reactor was again cooled to 22
°C and vented with nitrogen to atmospheric pressure.
Additional isopropyl alcohol (5.0 L) was charged to the
reactor, and the mixture was stirred to homogeneity. A
sample was again pulled to determine residual solvent levels
which showed residual MTBE below detectable levels and
residual toluene at 0.01% (Pass: e 0.1%). The internal
temperature was cooled to 0 °C, and seed (R,R)-5 (2.0 g)
was charged to the reactor. After 45 min, a thick seed bed
was present and water (4.5 L) was charged portionwise to
the reactor, allowing the internal temperature to return to 0
°C following each water charge. The fluid temperature was
set to -20 °C to control/abate the exotherms. Once internal
temperature reached 0 °C following the final water charge,
the mixture was held at 0 °C for 30 min and then warmed
to 22 °C and stirred for 12 h at 22 °C. The resulting mixture
was filtered through a medium-porosity sintered glass funnel.
The remaining solid in the reactor was rinsed out with 1.1:
1
(GC) was performed using Agilent 6890 systems. H and
13C NMR chemical shifts are reported as δ values relative
to internal chloroform (1H δ 7.27 ppm, 13C δ 77.0 ppm)
One-Pot Synthesis of Koga Amine (R,R)-5. To a 50 L
reactor was charged ethanol (Absolute, 2.0 L), and stirring
was commenced (ca. 200 rpm). (R)-Styrene oxide ((R)-7,
1.017 kg, 8.46 mol) was charged. Ethanol (2 × 1.0 L) was
used to rinse the residual (R)-7 from the container, and each
rinse was charged. Piperidine (1.00 L, 10.16 mol, 1.2 equiv)
was then charged to the reactor in a single portion. Ethanol
(2 × 500 mL) was used to rinse residual piperidine into the
reactor. The contents of the reactor were then heated to an
internal temperature of 80 °C ( 2 °C and held at this
temperature for 5 h. An analytical sample was pulled
following the 5 h reflux period and subjected to HPLC
analysis which showed 99.1% conversion to aminoalcohols
8a and 8b. The reactor was cooled to 22 °C and equipped
with a batch concentration receiver flask. The reactor was
placed under a vacuum (ca. 20 Torr) and then heated to an
internal temperature of 45 °C ( 2 °C, and distillate was
collected. The reaction mixture was concentrated until
approximately a 2.0 L total volume was obtained. Toluene
(3.0 L) was charged, and the mixture again was heated to
an internal temperature of 45 °C ( 2 °C. Distillate was
collected, and the batch again was concentrated to ap-
proximately a 2.0 L total volume (slurry). The reactor was
cooled to 22 °C, and additional toluene (3.0 L) was charged.
A second azeotropic distillation was performed. The reactor
was cooled to 22 °C, and MTBE (9.0 L) was charged. An
analytical sample was pulled and analyzed for residual
solvents which showed residual ethanol at 0.01% (Pass: e
0.5%) and residual toluene at 2.6%. To the MTBE solution
of aminoalcohols 8a and 8b was added triethylamine (4.37
L, 31.4 mol, 3.7 equiv). Stirring was commenced, and no
further additions to the reactor were made until the solids
had dissolved (ca. 10 min). The batch was cooled to -10
°C, and upon reaching this temperature, mesyl chloride (790
mL, 10.2 mol, 1.2 equiv) was charged to the batch from a
self-equilibrating glass dropping funnel at a rate of ap-
proximately 3.3 mL/min.
NOTE: Addition rate and chiller temperature were
regulated such that the internal temperature of the reaction
mixture was maintained between -8 and -10 °C during the
addition. Total addition time was 50 min.
Vol. 11, No. 2, 2007 / Organic Process Research & Development
•
221

1.0 IPA/water solution (2 × 850 mL), and this was applied
as a wash of the isolated solid on the funnel. The wet cake
was allowed to dry on the funnel for 3 h, after which it was
transferred to a vacuum oven and dried at 50 °C and 15 Torr
for 24 h with a nitrogen bleed to afford 1330 g of dried
(R,R)-5 (70% yield based on styrene oxide, 99.9% chiral
purity). Mp 63-64 °C; 1H NMR (400 MHz, CDCl3) δ 7.20-
7.35 (m, 10H), 3.71 (dd, J ) 3.5, 10.9 Hz, 2 H), 2.21-2.54
(m, 18H), 1.39-1.68 (m, 14H); 13C NMR (100 MHz, CDCl3)
δ 143.2, 128.2, 127.3, 126.9, 66.6, 60.1, 54.5, 46.2, 30.4,
26.1, 24.4; IR (thin film/NaCl) 3310 (w), 2929 (m), 2787
(m), 1452 (m), 1150 (m), 1116 (m), 914 (w), 748 (s). HRMS
(EI) m/z found: 449.3639 [calcd for C24H45N4 (M + H):
449.3644. Anal. Calcd for C24H44N4: C, 77.63; H, 9.88; N,
12.49. Found: C, 77.66; H, 10.08; N, 12.46. [R]D20 -112 (c
5.45, CHCl3).
Acknowledgment
The authors wish to thank Dr. M. Bhupathy and Ms.
RoseAnn Price for assistance with raw material procurement
and logistics.
Supporting Information Available
Analytical methods for each step including methods for
residual solvents and analytical data for intermediates, as well
as procedures and NMR data/spectra for intermediates. This
information is available free of charge via the Internet at
http://pubs.acs.org.
Received for review November 15, 2006.
OP0602371
222
•
Vol. 11, No. 2, 2007 / Organic Process Research & Development