Microwave-initiated living free radical polymerization: optimization of the preparative scale synthesis of Rasta resins

Article abstract of DOI:10.1016/j.tetlet.2006.12.040

Microwave heating of high-loading TEMPO-methyl resin with functionalized styrenyl monomers in an ADVANCER preparative scale microwave synthesizer affords larger resin beads (>375 μm) via living free radical polymerization (LFRP). Unlike homogeneous MAOS r

Full text of DOI:10.1016/j.tetlet.2006.12.040

Tetrahedron Letters 48 (2007) 1497–1501
Microwave-initiated living free radical polymerization:
optimization of the preparative scale synthesis of Rasta resins
a,
a
a
a
*
Joseph M. Pawluczyk, Ray T. McClain, Chris Denicola, James J. Mulhearn, Jr.,
a
a,b
Deanne Jackson Rudd and Craig W. Lindsley
a
Department of Medicinal Chemistry, Merck Research Laboratories, PO Box 4, West Point, PA 19486, USA
b
VICB Program in Drug Discovery, Departments of Pharmacology and Chemistry,
Vanderbilt Medical Center, Nashville, TN 37232, USA
Received 27 October 2006; revised 7 December 2006; accepted 8 December 2006
Available online 16 January 2007
TM
Abstract—Microwave heating of high-loading TEMPO-methyl resin with functionalized styrenyl monomers in an ADVANCER
preparative scale microwave synthesizer affords larger resin beads (>375 lm) via living free radical polymerization (LFRP). Unlike
homogeneous MAOS reactions, the heterogeneous LFRP reaction is not directly scalable and required significant optimization.
Under modified conditions, high-loading Rasta resins (>3.8 mmol/g) are reproducibly obtained in 40–100 g quantities.
Ó 2006 Elsevier Ltd. All rights reserved.
Recently, microwave-assisted organic synthesis (MAOS)
has had a profound impact on the way organic and
medicinal chemists practice organic synthesis in both
same direct scalability of homogeneous MAOS reac-
tions employing a Biotage ADVANCER , a batch mul-
TM
timode microwave synthesizer with a 300 mL reaction
1
–4
9
academic and industrial laboratories.
Hundreds of
cavity. For instance, our laboratory recently reported
manuscripts have been published highlighting acceler-
ated reaction rates, increased yields with diminished side
product formation and increased reaction generality as a
on an MAOS protocol for the rapid synthesis of diverse
3,5,6-trisubstituted-1,2,4-triazines 3 by the condensation
of an acyl hydrazide 1 and a diketone 2 in HOAc with
1
–5
10
result of MAOS. Indeed, our laboratory relies heavily
on MAOS during the lead optimization (LO) phase of
excess NH OAc at 180 °C in 5 min (Table 1). Impor-
4
tantly, these conditions provided 3 in the same purity
and yield on a 0.05–3.0 mmol scale in a single-mode
synthesizer as observed on a 30 mmol scale in the
5
,6
drug discovery. In the LO phase, we employ single-
mode microwave synthesizers wherein temperature is
accurately measured by IR and delivers milligram to
gram quantities of compound (reaction vessels from
TM 11
multimode ADVANCER .
0
.2 to 20 mL). A critical issue for the pharmaceutical
These data led us to investigate the feasibility of a pre-
parative, heterogeneous microwave-initiated living free
radical polymerization (LRFP) protocol for the scalable
synthesis of Rasta resins. The LFRP process is initiated
by heating TEMPO-methyl resin 4, a solid-supported
radical initiator, with functionalized styrenyl monomers
5 conventionally at 130 °C for an average of 20 h or in a
single-mode microwave synthesizer for 5 min at 185 °C
industry has been one of scale, that is, would medicinal
chemistry routes employing MAOS in the lead optimiza-
tion phase be translatable to MAOS preparative scale
routes for development purposes?
In response to this key issue, numerous companies have
developed preparative scale microwave synthesizers
employing batch, batch flow, or continuous flow para-
digms. Early data suggests that there is scalability of
MAOS reactions from single-mode to multimode paral-
1
2–15
to deliver high-loading ‘Rasta resins’ 6.
Scheme 1
7
illustrates the architecture of a generic Rasta resin
depicted by a cartoon structure 7 in which hair-like
appendages represent new block polymer growth or
alternatively by 8, wherein the shaded inner circle repre-
sents the original cross-linked polystyrene (PS) core and
the outer clear circle represents new polymer growth.
This new class of resins was shown to have unique
8
lel batch reactors. In our hands, we have observed the
*
Corresponding author. Tel.: +1 215 652 7216; fax: +1 215 652
310; e-mail: joe_pawluczyk@merck.com
7
0040-4039/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2006.12.040

1498
J. M. Pawluczyk et al. / Tetrahedron Letters 48 (2007) 1497–1501
Table 1. Scalable MAOS 1,2,4-triazine synthesis
O
NH
4
OAc (xs)
HOAc
O
Ph
Ph
N
N
3
N
Ph
NHNH
2
+
Ph
80 oC, uw
5 min
1
N
O
N
1
2
a
Vial (mL)
Volume (mL)
Amount (mmol)
Theo yield (mg)
Yield (%)
Purity (%)
0
0
1
5
.2–0.5
.5–2.0
0–20
0.25
1.00
10.0
0.05
0.20
3.00
30.0
15.6
62.1
931.0
9310.0
88
90
92
92
86
92
93
92
0–300
100.0
a
Purity refers to purity of crude reaction by LCMS at 214 nm and ELSD.
R
Br
Br
R
O
n
N
O
N
5
9
O
n
N
4
1
~
30 o
C
20 h
185 o
10 min, uw
C
10
or
(1.00 g)
4
6
(6.95 g)
1
85 oC
10 min, uw
TM
Scheme 2. Large scale (20 mL reaction vessel) Optimizer run.
Cl
R
8
7
1
1
Scheme 1. Rasta resin.
polymeric
mass
4
1
80 o
C
1
0
min, uw
(10 g)
macromolecular architecture, wherein linear block poly-
mers emanate from the original cross-linked PS core. In
our previous MAOS LRFP protocol, we employed
TM
Scheme 3. Initial ADVANCER run.
2
00 mg of 4 to produce up to 1.5 g (P/I, product/initia-
tor ratios of >6) of various functionalized, >550 lm
Rasta resins 6 with loadings in excess of 5 mmol/g. Sig-
nificantly, there was excellent batch-to-batch reproduc-
ibility in the 0.5–2 mL reaction vials.15 However, even
with parallel processing, this protocol provided rela-
tively small quantities of Rasta resins that were insuffi-
cient for wide-spread application in our laboratory.
Moreover, shortly after our initial report, Aldrich com-
mercialized our Rasta resins ($8–50/g); therefore, the
development of a preparative MAOS protocol was of
cavity with 10 g of 4 (18.6 mmol, resin loading of
1.86 mmol/g, representing a 50-fold increase over the
standard protocol), 35 M equiv of 11 as a solvent, and
heated the reaction for 10 min at 180 °C (no stirring
and instrument was set to passively cool the reaction).
After 2 min of heating, the temperature spiked to above
250 °C and the system underwent a ‘crash cool’ and shut
down. Upon inspection of the cavity after cooling, the
LRFP protocol delivered, instead of resin beads, a poly-
meric mass.
1
6
a paramount concern.
Clearly, the heterogeneous LRFP protocol, unlike the
homogeneous 1,2,4-triazine protocol (Table 1), would
not be directly scalable. However, there are also instru-
ment differences. First, the LFRP protocol was devel-
oped on a single-mode microwave synthesizer, and the
Having recently acquired the 20 mL reaction vial capa-
bility on our single-mode synthesizer, we examined the
MAOS LRFP protocol on a larger scale. As shown in
Scheme 2, the scale of the LRFP reaction could be
increased 5-fold, such that 1.0 g of 4 would deliver
TM
ADVANCER is a multi-mode microwave. Secondly,
6
.95 g of bromo resin 10 with a P/I ratio of 6.9 and a
the single-mode units measure temperature with IR,
while the ADVANCER employs a fiber optic thermo-
TM
loading level of 5.2 mmol/g (42% Br) under the same
reaction conditions.
1
7
couple in the reaction cavity. One explanation for the
temperature spike is that the resin polymerized around
the thermocouple, thus insulating it from accurately
recording the temperature. Only when temperatures ex-
ceeded 200 °C, did the resin melt, allowing the thermo-
Naively, we anticipated that the LRFP protocol would
be directly scalable on the ADVANCER system. In
the event (Scheme 3), we charged the 300 mL reaction
TM

J. M. Pawluczyk et al. / Tetrahedron Letters 48 (2007) 1497–1501
1499
couple to record the temperature spike. This undesired
polymerization and temperature spike was observed at
all temperatures examined when using neat styrene as
a solvent.
Historically, LFRP is a solvent free suspension poly-
merization; however, for subsequent experiments, we
reduced the temperature and decided to add NMP as
1
2–15
a spectator co-solvent.
Subsequent trials employed
1
0 g of 4, a 35 M excess of either 9 or chloromethyl sty-
rene 11, and 90 mL of NMP (Scheme 4). The reaction
was heated at 140 °C for 60 min with a slow ramp of
0
.5 °C/s, no stirring and with passive cooling. Under
these conditions, the temperature was tightly maintained
and there was no temperature spike. The addition of
CH Cl , filtration through a sintered-glass frit and
2
2
washing with five cycles of CH Cl and MeOH delivered
2
2
free flowing, spherical resin beads 10 and the Rasta Mer-
rifield resin 12. For 10, the P/I ratio was 4.6, producing
4
(
4
6 g of resin beads with a loading level of 3.8 mmol/g
30.5% Br) and the P/I ratio for 12 was 4.6, affording
6 g of resin beads with a loading level of 5.30 mmol/g
1
7
(
18.55% Cl). This protocol is reproducible with varia-
tions in P/I and loading levels of less than 5% from
batch-to-batch. Figure 1 shows the photographs of
unswollen 4, 10, and 12 at equal magnification, note that
the Rasta resins remain spherical, but are smaller (370–
4
50 lm vs >550 lm) than obtained under the original
1
5,18
MAOS LFRP protocol.
While initially disappoint-
ing, these beads were found to be more resistant to
mechanical breakage than the >550 lm beads while
maintaining high-loading levels.
Further optimization studies focused on decreasing the
Figure 1. Resin photographs (equal magnification): (a) resin 4; (b)
resin 10; (c) resin 12.
6
0 min reaction time and identifying the maximum
18
allowable temperature for LFRP. All temperatures
above 160 °C led to temperature spikes, so efforts cen-
tered on minimizing the reaction time at 160 °C. As
shown in Table 2, 30 min proved to be the optimal reac-
tion time at 160 °C, with diminutions in P/I ratios and
loading levels as time decreased. These conditions affor-
ded the same P/I and loading level as those shown in
Table 2. Large scale optimization
Cl
Cl
O N
n
11
4
Br
Br
1
60 ºC
12
(10 g)
time, uw
a
Time (min)
Resin wt. (g)
P/I ratio
Loading level
mmol/g)
(
O N
n
9
4
30
15
41.0
37.2
32.6
28.9
4.1
3.7
3.2
2.9
5.3
5.4
5.2
4.8
o
10
1
40 C
1
0
(
10 g)
60 min, uw
(46 g)
5
a
Cl
Cl
Loading based on % Cl; 18.54, 19, 18.3, 16.9, respectively.
Scheme 4 at 140 °C for 60 min, and afforded an excellent
batch-to-batch reproducibility.
O N
11
4
n
Our attention then focused on determining the maxi-
mum scale possible in the ADVANCER unit. In the
event, 30 g of 4 (55.8 mmol, a 3-fold increase of Table
o
12
(46 g)
1
40 C
TM
(10 g)
60 min, uw
Scheme 4.
2 and a 150-fold increase in scale over our earlier work)

1500
J. M. Pawluczyk et al. / Tetrahedron Letters 48 (2007) 1497–1501
Cl
Cl
Table 3. Large scale synthesis of Rasta amines
Cl
(m,p)
1 2
HNR R
NMP
1 2
NR R
O
n
N
O N
n
2
00 ºC
11
4
mw, 30 min
40 o
60 min, uw
12
(116 g)
1
C
12 (10 grams)
(30 g)
Entry HNR
1
R
2
Analysis (%)
C 82.99; H 9.68; N 6.3; Cl 0.8
C 81.08; H 9.50; N 4.59; Cl 5.07 3.3
Loading
Scheme 5. Large-scale run.
(mmol/g)
1
2
HNEt
2
4.5
was diluted with a 24 M excess of chloromethyl styrene
1 and 100 mL of NMP (Scheme 5). The reaction was
heated at 140 °C for 60 min with a slow ramp of
HN
1
2
0
.5 °C/s, no stirring and with passive cooling. Under
3
4
5
HN
HN
HN
C 83.56; H 9.62; N 6.31; Cl 0.9
C 77.2; H 8.57; N 6.22; Cl 0.7
4.5
4.5
3.2
these conditions, the temperature gradually increased
to 153 °C then returned to 140 °C within the first
O
1
4 min of the reaction, at which then the instrument
applied microwave energy to maintain the set temperature.
NH C 77.93; H 8.44; N 8.9; Cl 3.2
Upon opening the reaction vessel, a small solid mass was
observed in the core of the vessel. This was removed and
the remaining resin was washed. The addition of
merization protocol for the reproducible batch-to-batch
synthesis of 40–100 g quantities of Rasta resins with
loading levels >3.8 mmol/g and with particle sizes in ex-
CH Cl , filtration through a sintered-glass frit, and
2
2
washing with five cycles of CH Cl and MeOH delivered
2
2
free flowing, spherical Rasta Merrifield resin 12 with a
TM
cess of 380 lm employing a multi-mode ADVANCER
P/I ratio of 3.9, affording 116 g of resin beads with a
microwave synthesis unit. Moreover, this study found
that not all MAOS reactions are directly scalable, and
that heterogeneous reactions benefit from the presence
of a spectator solvent. Importantly, this new, scalable
protocol finally affords multigram quantities of custom
Rasta resins, which should increase the frequency of
their application in solid phase and parallel synthesis.
1
7
loading level of 5.30 mmol/g (18.54% Cl).
With a scalable synthesis of 12, efforts centered on scal-
ing up the synthesis of functionalized Rasta resins for
use in parallel synthesis. In our earlier work, we were
able to generate high-loading (4–5 mmol/g), functional-
ized Rasta amine resins on ꢀ100 mg scale in the 0.5–
2
.0 mL reaction vessels (Scheme 6). Once again, this
scale was not practical for production and wide-spread
TM
adoption in our laboratory. In the ADVANCER , we
References and notes
increased the scale 100-fold, producing 10 g quantities
of functionalized Rasta amine resins (Table 3) with con-
sistent loading levels of 4.5 mmol/g (particle sizes
1
. Strauss, C. R.; Trainor, R. W. Aust. J. Chem. 1995, 48,
665–1692.
. Lindstrom, P. Tetrahedron 2001, 57, 9225–9283.
1
>380 lm) for entries 1–4, and diminished loading for
2
the Rasta piperazine (entry 5) due to internal cross-link-
ing. This directly scalable protocol provided sufficient
quantities of such important solution phase reagents as
Rasta-DIEA (entry 2), Rasta-NMM (entry 4), and Ras-
ta piperazine (entry 5) for routine solution phase parallel
synthesis in the laboratory.
3. Larhed, M.; Hallberg, A. Drug Discov. Today 2001, 6,
406–416.
4. Kappe, C. O. Angew. Chem., Int. Ed. 2004, 43, 6250–6284.
5
. Lindsley, C. W.; Wolkenberg, S. E. Discov. Dev. 2004, 4,
–5.
2
6
. Lindsley, C. W.; Zhao, Z.; Leister, W. H.; Robinson, R.;
Barnett, S. F.; Defeo-Jones, D.; Jones, R. E.; Huber, H.
E.; Hartman, G. D.; Huff, J. R.; Huber, H. E.; Duggan,
M. E. Bioorg. Med. Chem. Lett. 2005, 15, 76–764.
In summary, we have developed a preparative, hetero-
geneous microwave-initiated living free radical poly-
7
8
9
. Wolkenberg, S. E.; Shipe, W. D.; Lindsley, C. W.; Guare,
J. P.; Pawluczyk, J. M. Curr. Opin. Drug Discov. Dev.
2
005, 8, 701–708.
. Stadler, A.; Yousefi, B. H.; Dallinger, D.; Walla, P.; Van
der Eycken, E.; Kaval, N.; Kappe, C. O. Org. Proc. Res.
Dev. 2003, 7, 707–715.
Cl
. For information on the ADVANCERTM and other
1 2
HNR R
Biotage
instruments
and
details,
see:
http://
DMF
1 2
NR R
www.biotage.com.
O
n
N
2
00 o
0 min, uw
C
10. Zhao, Z.; Leister, W. H.; Strauss, K. A.; Wisnoski, D. D.;
3
Lindsley, C. W. Tetrahedron Lett. 2003, 44, 1123–1127.
12
4- to 5 mmol/g
1
1
1. Previously unpublished data from our lab and Biotage.
2. Hodges, J. C.; Harijrishnan, L. S.; Ault-Justus, S. J.
Comb. Chem. 2000, 2, 80–89.
(100 mg)
Scheme 6. Rasta amine resins.

J. M. Pawluczyk et al. / Tetrahedron Letters 48 (2007) 1497–1501
1501
13. Lindsley, C. W.; Hodges, J. C.; Filzen, G. F.; Watson, B.
M.; Geyer, A. G. J. Comb. Chem. 2000, 2, 550–559.
14. McAlpine, S. R.; Lindsley, C. W.; Hodges, J. C.; Leonard,
D. M.; Filzen, G. F. J. Comb. Chem. 2001, 3, 1–5.
15. Wisnoski, D. D.; Leister, W. H.; Strauss, K. A.; Zhao, Z.;
Lindsley, C. W. Tetrahedron Lett. 2003, 44, 4321–4325.
16. In Aldrichim. Acta 2005, 38, 17.
17. C, H, N, Cl, Br analysis provided by Quantitative
Technologies Inc., PO Box 470, Salem Ind. Park, Bldg.
5, Rt. 22E, Whitehouse, NJ 08888, USA.
18. The pictures were captured using an Olympus BX51
Polarizing Microscope equipped with an Insight FireWire
Color Mosaic Camera. Spot 4.0.5 software was used for
picture manipulation.