In situ FTIR study and scale-up of an enolization - Azidation sequence

Article abstract of DOI:10.1021/op900299c

A key step in the synthesis of an optically active aminoalcoholcontaining active pharmaceutical ingredient (API) involved the diastereoselective introduction of an azido functional group on a functionalized chiral oxazolidinone. This was accomplished via a low-temperature enolization, followed by a quench with triisopropylbenzenesulfonyl azide. To enable scale-up of this process, the enolization temperature had to be increased from the original <-65 °C to approximately -40 °C. In situ FTIR was used to study the enolization and quench stages of the reaction. The half-life of the enolate at -45 °C was estimated to be 12 h on the basis of in situ FTIR profiling. Examination of the in situ FTIR data also provided evidence that the reaction between the enolate and triisopropylbenzenesulfonyl azide was instantaneous and demonstrated that accumulation of triisopropylbenzenesulfonyl azide did not occur. A combination of in situ FTIR experiments and traditional parameter ranging experiments resulted in a process that was successfully run at -40 °C without an appreciable erosion of facial selectivity or yield.

Full text of DOI:10.1021/op900299c

Organic Process Research & Development 2010, 14, 466–469
In Situ FTIR Study and Scale-Up of An Enolization-Azidation Sequence†
Terrence J. Connolly,* Eric C. Hansen, and Michael F. MacEwan
Chemical DeVelopment, Wyeth Research, 401 North Middletown Road, Pearl RiVer, New York 10965, U.S.A.
Abstract:
Further processing constraints were added by Wyeth’s Chemical
Development Material Operations Group, who informed the
team that trisyl azide was more easily and economically sourced
as a 30 wt % solution in toluene than as an isolable solid.
A key step in the synthesis of an optically active aminoalcohol-
containing active pharmaceutical ingredient (API) involved the
diastereoselective introduction of an azido functional group on a
functionalized chiral oxazolidinone. This was accomplished via a
low-temperature enolization, followed by a quench with triisopro-
pylbenzenesulfonyl azide. To enable scale-up of this process, the
enolization temperature had to be increased from the original
Results and Discussion
The process initially used by Chemical Development in-
volved addition of 1 equiv of 0.5 M potassium hexamethyl-
disilazane (KHMDS) in toluene to a solution of 2 in 6.5 volumes
of THF, while the internal batch temperature was maintained
below -65 °C, followed by a solution of trisyl azide in 4
volumes of THF, then acetic acid and finally water. The high
dilution of substrate and base had a detrimental impact on
throughput in the 50-L vessel. The maximum batch size that
could be processed was approximately 2.8 kg of 2, based on
<
-65 °C to approximately -40 °C. In situ FTIR was used to study
the enolization and quench stages of the reaction. The half-life of
the enolate at -45 °C was estimated to be 12 h on the basis of in
situ FTIR profiling. Examination of the in situ FTIR data also
provided evidence that the reaction between the enolate and
triisopropylbenzenesulfonyl azide was instantaneous and demon-
strated that accumulation of triisopropylbenzenesulfonyl azide did
not occur. A combination of in situ FTIR experiments and
traditional parameter ranging experiments resulted in a process
that was successfully run at -40 °C without an appreciable erosion
of facial selectivity or yield.
Vmax being below 48 L following the acetic acid addition.
Significant improvements were made with respect to through-
put by reducing the amount of solvent to dissolve 2 from 6.5
to 2.5 volume equivalents, which still kept 2 completely in
solution even at the low temperatures required for the process,
and by the use of more concentrated 0.91 M KHMDS in THF.
These two changes along with use of a 30 wt % solution of
trisyl azide in toluene allowed the batch size to be increased to
5 kg of 2 based on a Vmax of 48 L.
With process throughput improved, efforts were then focused
on increasing the operating temperature to the -40 to -50 °C
range requested by the kilogram laboratory team. A series of
experiments were performed that addressed the operating
temperature and robustness of the process by extension of
addition times and evaluation of a series of hold times. The
results of this study are shown in Table 1.
Introduction
The evolution of a synthetic process from initial discovery
synthesis to the successful preparation of multikilogram quanti-
ties of compound 1 has been communicated previously.
Installation of the chiral center bearing the nitrogen relied on
methodology developed by Evans that involved low-temper-
ature enolization of oxazolidinone derivative 2 followed by
conversion to azido intermediate 5 with triisopropylbenzene-
sulfonyl azide (trisyl azide) and reduction (Scheme 1).
The original Discovery conditions involved performing the
enolization and trisyl azide addition sequence below -65 °C.
Although these conditions were suitable for the first Chemical
Development batch of 1, these low-temperature conditions could
not be achieved in our 50-L Hastelloy vessel. In order to allow
for heat of mixing and a slight exotherm for the reactions, the
kilolab team recommended that the process be evaluated in the
laboratory at an internal reaction temperature of ∼-40 °C.
1
2
3
This study indicated that increasing the enolization operating
temperature did not have a significant detrimental impact on
the conversion of 2 to 5 within the processing times examined,
suggesting higher temperatures were acceptable. During this
4
work, an intermediate assumed to be 4 was observed by LC.
Intermediate 4 did not decompose to 5 until acetic acid was added.
To gain insight into the kinetics of the reaction and to further
evaluate the stability of intermediates (in parallel with the
parameter ranging experiments reported above) in situ FTIR
was evaluated as a process analytical technology (PAT) method.
Several recent reports have highlighted the utility of using PAT
to gain better process understanding, and in situ FTIR has been
†
Wyeth became part of Pfizer on October 16, 2009.
*
Address correspondence to this author at the above mailing address, attention
B222/2125, or use e-mail: connolt@wyeth.com.
(
1) Alimardanov, A.; Nikitenko, A.; Connolly, T. J.; Feigelson, G.; Chan,
A. W.; Ding, Z.; Ghosh, M.; Shi, X.; Ren, J.; Hansen, E.; Farr, R.;
MacEwan, M.; Tadayon, S.; Springer, D.; Kreft, A. F.; Potoski, J. R.
Org. Process Res. DeV. 2009, 13, 1161–1168.
5-8
established as an important tool.
We established that the wavelength region between ap-
(2) Evans, D. A.; Britton, T. C.; Ellman, J. A.; Dorow, R. L. J. Am. Chem.
Soc. 1990, 112, 4011–4030.
-1
proximately 2200 and 1550 cm contained no interfering peaks
(
3) Safety screening performed in-house generated results in line with those
reported previously by Merck workers for triisopropylbenzenesulfonyl
azide. See: Tuma, L. D. Thermochim. Acta 1994, 243, 161–167. A review
of the thermal behavior of several sulfonyl azides is available: Bollinger,
F. W.; Tuma, L. D. Synlett 1996, 407-413.
from the solvent (THF) and the starting material 2 had a strong
-1
stretching band at 1825-1775 cm due to the oxazolidinone
carbonyl group (Figure 1). As KHMDS (0.91 M in THF) was
4
66
•
Vol. 14, No. 2, 2010 / Organic Process Research & Development
10.1021/op900299c  2010 American Chemical Society
Published on Web 01/26/2010

Scheme 1. Synthesis of 1
added to 2, the peak representing 2 disappeared, and a new
that the reaction between the enolate 3 and trisyl azide to form
4 was fast enough at this temperature to prevent any accumula-
tion of trisyl azide. It was only at the end of the reaction that a
small increase occurred at 2150-2090 cm due to a slight
overcharge of trisyl azide. A timecourse plot of the peaks of
interest is shown in Figure 2.
-1 9
peak for the enolate 3 grew in at 1740-1715 cm . After a
0 min hold time, a solution of trisyl azide in THF was added
and led to the disappearance of the peak for 3 and reappearance
3
-1
-1 10
of the carbonyl band at 1825-1775 cm . It was interesting
to note that, although trisyl azide had a very strong stretching
-1
band at 2125 cm , no adsorption at this frequency was
If the enolate was held at -45 °C for 4 h before the solution
of trisyl azide was added, the enolate signal decreased by
11
observed during the addition of trisyl azide. This indicated
Table 1. Summary of parameter ranging experiments
base addition
time (min)
enolate hold
time (min)
temperature of trisyl
azide solution (°C)
hold time after trisyl
azide addition (min)
entry
temperature (°C)
assay yield (%)a
1
2
3
4
5
6
7
-78
-50
-30
-78
-40
-40
-10
30
10
10
2
10
30
5
45
45
40
30
45
40
30
-78
-78
-78
-78
-78
25
2
2
90
87
86
97
92
98
94
2
60
2
b
75
25
15
a
Determined using HPLC. b Trisyl azide was added over 75 min in this example.
Figure 1. IR spectral region of interest.
Vol. 14, No. 2, 2010 / Organic Process Research & Development
•
467

Figure 2. In situ FTIR profile at -45 °C for KHMDS and trisyl azide additions.
Figure 3. In situ FTIR profile at -45 °C for KHMDS and trisylazide additions.
1
2
approximately 16% (Figure 3). Extrapolation assuming linear
decay provided an estimate of approximately 12 h for the half-
life of the enolate at -45 °C. Performing the enolization at -10
°C and holding the solution overnight led to complete decom-
position of the enolate and suggested a half-life of 2.5-3 h
(
4) The Evans group has isolated this type of intermediate and determined
that it exists as a mixture of two tautomers. We have only shown one
form here. For further details on the structural assignment of a
sulfonylated triazine as an intermediate, see ref 2.
(Figure 4). In this case, addition of trisyl azide after the overnight
hold led to an immediate appearance of the azide stretch since
none of the reagent was consumed.
(
5) Argentine, M. D.; Braden, T. M.; Czarnik, J.; Conder, E. W.; Dunlap,
S. E.; Fennell, J. W.; LaPack, M. A.; Rothhaar, R. R.; Scherer, R. B.;
Schmid, C. R.; Vicenzi, J. T.; Wei, J. G.; Werner, J. A. Org. Process
Res. DeV. 2009, 13, 131–143.
These in situ FTIR experiments provided some insight into
enolate stability and led us to scale up entry 6 in Table 1 to 3
L. Although the experiment summarized as entry 7 in Table 1
gave excellent results on a small scale, the potential liabilities
of the 3 h half-life of enolate 3 at -10 °C were viewed as being
too large. Scale-up to approximately 100 g of 2 afforded a 93%
assay yield of 5. Aqueous acetic acid was added, and the
(
(
6) Baenziger, M.; Mak, C.-P.; Muehle, H.; Nobs, F.; Prikoszovich, W.;
Reber, J.-L.; Sunay, U. Org. Process Res. DeV. 1997, 1, 395–406.
7) Horv a´ th, A.; De Smet, K.; Ormerod, D.; Depr e´ , D.; P e´ rez-Balado,
C.; Govaerts, T.; Van den Heuvel, D.; Schorpion, I. Org. Process Res.
DeV. 2005, 9, 356–359.
(
(
8) Rein, A. J.; Donahue, S. M.; Pavlosky, M. A. Curr. Opin. Drug
DiscoVery DeV. 2000, 3, 734–742.
9) This peak also corresponds to the oxazolidinone carbonyl group.
Enolization to 3 is expected to result in a lower stretching frequency
for the carbonyl group. The stretching frequencies of the oxazolidinone
carbonyl groups of N-acyl and N-methyl oxazolidinone occur at 1795
(11) This stretch was assigned as the asymmetric stretch due to the N
3
group on the basis of vapor phase data published recently by the
Novartis group. See: Wiss, J.; Fleury, C.; Onken, U. Org. Process
Res. DeV. 2006, 10, 349–353.
-
1
and 1735 cm , respectively.
(12) The solution of 3 remained homogenous over the hold period so the
decrease in signal strength was not due to precipitation of the enolate
species.
(
10) In this case, THF was used for dilution of the azido transfer reagent
to remove any interference from another solvent.
4
68
•
Vol. 14, No. 2, 2010 / Organic Process Research & Development

Figure 4. In situ FTIR profile at -10 °C for KHMDS and trisylazide additions.
reaction mass was then warmed to ambient temperature to affect
applied. Each in situ FTIR experiment was performed in a 25
13
conversion of 4 to 5. Heptane was added and the organic layer
washed with aqueous potassium phosphate to remove the
sulfinic acid byproduct. The organic solution was exchanged
under vacuum to ethanol, and the reduction to 6 was performed
mL three-neck round-bottom flask. Compound 2 (1.50 g) was
dissolved in THF (8 mL) and cooled to the desired reaction
temperature, which was measured using a thermocouple inserted
into the reaction vessel. KHMDS (4.5 mL, 0.91 M in THF)
was charged via syringe pump over 15 to 30 min. The mixture
was held for the specified period of time at the reaction
temperature, and trisyl azide (1.12 g in 3.5 mL THF) was added
over 30 min, followed by HOAc, and the reaction mixture was
warmed to ambient temperature.
1
using conditions already reported. The diastereomeric ratio of
product 6 was comparable to results achieved when the
enolization was performed at -65 °C, indicating that facial
selectivity was not affected by performing the enolization at
the higher temperature.
Preparation of 5. A solution of THF (350 mL) and 2
(
-
106.95 g at 93.5 wt %, 242 mmol) was maintained at -40 to
50 °C as KHMDS (266 g, 0.91 M in THF, 266.1 mmol) was
added over 20 min. The resulting mixture was held between
40 and -50 °C for 25 min and a precooled (∼-65 °C)
Conclusion
In situ FTIR studies were used in combination with
traditional parameter ranging experiments in the development
of an enolization-azidation sequence and to establish the
temperature range for acceptable enolate stability.
-
solution of trisyl azide (249.5 g, 30 wt % in toluene, 242 mmol)
was added via cannula over 7 min as the temperature was
maintained between -50 and -40 °C. The flask and cannula
were rinsed forward with THF (30 mL), and the reaction
mixture was stirred for 15 min at -45 °C. Acetic acid (66 g,
Experimental Section
General. Reaction monitoring was carried out on an Agilent
1
200 series HPLC with PDA monitoring (210 nm) equipped
with an Agilent Extend C-18 column (3.5 mm × 100 mm) at
5 °C. Elution was performed with a flow rate of 1.0 mL/min
and a gradient method using acetonitrile (mobile phase A) and
water containing 0.1% H PO (mobile phase B) (Initial: 50%
4
.6 equiv) in water (70 mL) was added over 2 min, raising the
temperature from -47 to -25 °C. The reaction mixture was
warmed to ambient temperature and stirred until in-process
HPLC indicated <1% of 4. Assay of the resulting reaction
mixture indicated clean conversion to 5 (approximately 101 g,
3
3
4
A, 12 min 90% A, 15 min 90% A). Observed retention times
were as follows: 6 (2.37 min), toluene (4.34 min), triisopropy-
lbenzenesulfinic acid (4.76 min), 5 (8.34 min), 2 (7.17 min),
trisyl azide (12.09 min).
In situ FTIR experiments were performed using a Mettler-
Toledo ReactIR system (ReactIR4000 with K6 conduit and 16
mm SiComp probe). A SiComp probe was used because the
9
3%). Heptane (250 mL) was added to the mixture, and the
organic layer was separated. The organic layer was washed with
an aqueous solution of potassium phosphate tribasic (3 × 250
mL of 0.25 M solution) at 30 °C. The organic layer was
concentrated to an oil.
Acknowledgment
-1
area of interest to be examined extended above 2000 cm
unsymmetrical azide stretching). IR spectra were recorded
The Pearl River Kilogram Laboratory Team and Material
Operations Group are acknowledged for their support of this
project.
(
-1
every 2 min at 8 cm resolution, and spectral data were
collected over the 4000-650 cm range. Peak heights were
-1
measured from the zero baseline after a baseline offset was
Received for review November 12, 2009.
OP900299C
(
13) Addition of acetic acid as a neat liquid caused some stirring problems
that were eliminated by adding the acetic acid as an aqueous solution.
Vol. 14, No. 2, 2010 / Organic Process Research & Development
•
469