A comparison of catalysts to promote imidazolide couplings including the identification of 2-hydroxy-5-nitropyridine as a new, safe, and effective catalyst

Article abstract of DOI:10.1021/op0580062

Five catalysts were compared with respect to their safety and catalytic effectiveness for promoting imidazolide couplings. Reaction rate enhancement, shock sensitivity, and differential scanning calorimetry (DSC) data were considered in this analysis. 6-Chloro-1-hydroxybenzotriazole, which has been described in the literature as a safe catalyst, was found to be shock sensitive. 2-Hydroxy-5-nitropyridine is a new catalyst for this type of reaction and was found to be safe, effective, readily available, and similar in price to that of the 1-hydroxybenzotriazole, a common catalyst for promoting acylation reactions.

Full text of DOI:10.1021/op0580062

Organic Process Research & Development 2005, 9, 956−961
Full Papers
A Comparison of Catalysts to Promote Imidazolide Couplings Including the
Identification of 2-Hydroxy-5-nitropyridine as a New, Safe, and Effective
Catalyst
Peter J. Dunn,*,† Wilfried Hoffmann,† Ying Kang,†,‡ John C. Mitchell,‡ and Martin J. Snowden‡
Department of Chemical Research and DeVelopment, Pfizer Global Research and DeVelopment,
Sandwich, Kent CT13 9NJ, United Kingdom, and Medway Sciences, UniVersity of Greenwich,
Medway Campus, Chatham, Kent ME4 4TB, United Kingdom
Scheme 1
Abstract:
Five catalysts were compared with respect to their safety and
catalytic effectiveness for promoting imidazolide couplings.
Reaction rate enhancement, shock sensitivity, and differential
scanning calorimetry (DSC) data were considered in this
analysis. 6-Chloro-1-hydroxybenzotriazole, which has been
described in the literature as a safe catalyst, was found to be
shock sensitive. 2-Hydroxy-5-nitropyridine is a new catalyst for
this type of reaction and was found to be safe, effective, readily
available, and similar in price to that of the 1-hydroxybenzo-
triazole, a common catalyst for promoting acylation reactions.
Scheme 2
Introduction
The use of N,N1-carbonyldiimidazole (CDI) as a coupling
agent for amide bond formation has recently been applied
to the large-scale synthesis of a number of pharmaceutical
products, for example, sildenafil,1 darifenacin,2 and sunitinib.3
In addition a survey of 128 chemical processes recently
developed by AstraZeneca, GlaxoSmithKline (GSK), and
Pfizer revealed that CDI was the third most common reagent
used for N-acylation reactions and was used in 9 out of 84
examples.4 The advantages are that the byproducts, carbon
dioxide and imidazole, are innocuous and that the evolution
of carbon dioxide in forming the imidazolide 1 provides a
driving force for the reaction5,6 (Scheme 1). In addition the
imidazole byproduct can be removed by an acidic wash2 or
remains in solution in the organic solvent.1
One factor influencing the increased use of CDI on a large
scale is its relatively low price (around $8/mol for a large-
scale purchase). Although CDI is more expensive than
traditional amide-forming reagents, such as thionyl chloride
and isobutyl chloroformate, in some cases the high yields
and clean environmental conditions can justify its use.
A disadvantage of CDI is that the resulting imidazolide
1 is less reactive than the corresponding acid chloride, and
hence amide couplings with either hindered carboxylic acids
or weakly nucleophilic amines can be unacceptably slow.
One way around this issue is to use a catalyst. The first
reported example was the use of 1-hydroxybenzotriazole
(HOBt, 2) to catalyse the reaction of the hindered imidazolide
3 with the tyrosine derivative 4 via the active ester 57
(Scheme 2). However, the additional use of HOBt does have
some drawbacks. The material is known to explode when
heated beyond its melting point (around 160 °C).8 The
* To
whom
correspondence
should
be
addressed.
E-mail:
peter.dunn@pfizer.com.
† Department of Chemical Research and Development, Pfizer Global Research
and Development.
‡ Medway Sciences, University of Greenwich.
(1) Dale, D. J.; Dunn, P. J.; Golightly, C.; Hughes, M. L.; Levett, P. C.; Pearce,
A. K.; Searle, P. M.; Ward, G.; Wood, A. S. Org. Process Res. DeV. 2000,
4, 17-22.
(2) Dunn, P. J.; Newbury, T. N.; Matthews, J. M.; O’Connor, G. World Patent
WO 03/080599.
(3) Vaidyanathan, R.; Kalthod, V. G.; Ngo, D. P.; Manley, J. M., Lapekas, S.
P. J. Org. Chem. 2004, 69, 2565-2568.
(4) Carey, J. S.; Laffan, D.; Thomson, C.; Williams, M. T. Org. Biomol. Chem.
Submitted for publication.
(7) Dunn, P. J.; Hughes, M. L.; Searle, P. M.; Wood A. S. Org. Process Res.
DeV. 2003, 7, 226-236.
(5) Anderson, G. W.; Paul, R. J. Am. Chem. Soc. 1958, 80, 4423.
(6) Anderson, G. W.; Paul, R. J. Am. Chem. Soc. 1960, 82, 4596.
(8) Urben, P. G., Ed. Brethericks Handbook of ReactiVe Chemical Hazards,
6th ed.; Butterworth-Heinemann: Oxford, Boston, 1999; p 746.
956
•
Vol. 9, No. 6, 2005 / Organic Process Research & Development
10.1021/op0580062 CCC: $30.25 © 2005 American Chemical Society
Published on Web 08/27/2005

Scheme 3
Table 1. Comments on the speed of reaction between the imidazolide 10 and the various amines
transportation of HOBt on a large scale is subject to
restrictions in Europe because of this explosive potential.
6-Chloro-1-hydroxybenzotriazole (Cl-HOBt, 6) is com-
mercially available and has been reported9 to be safer than
HOBt. It also has a lower pKa (4.15 vs 4.60)10 which should
make it a more effective catalyst, but no safety data were
reported for this catalyst. 1-Hydroxy-7-azabenzotriazole
(HOAt, 7) is commercially available and has an even lower
pKa (3.47)10 and hence should be a very effective catalyst in
promoting imidazolide-coupling reactions. However HOAt
is reported to have significant safety drawbacks and is shock
sensitive.11 Finally 2-hydroxypyridine (HOPy, 8) was re-
ported to be a reasonable catalyst to promote imidazolide
couplings12 although this work involved proprietary com-
pounds, and only partial structures were reported. The
objective of this paper is to run a side-by-side comparison
of the safety and catalytic effectiveness of these four
catalysts. In addition as the pKa of the catalyst was thought
to be related to its catalytic effectiveness, we also decided
to study 2-hydroxy-5-nitropyridine (NO2-HOPy, 9) as a
potential catalyst as this compound will be significantly more
acidic than 2-hydroxypyridine.
Results and Discussion
The first part of the work (summarised in Scheme 3) was
to define the scope of the reaction and select a reaction for
the side-by-side comparison study, the results of which are
summarised in Tables 1 and 2.
The imidazolide 10 undergoes rapid reaction with aliphatic
amines, and amide formation proceeds at a reasonable rate
with aromatic amines; hence, these reactions do not need
catalysis. The reaction of 10 with weakly basic amines such
as 4-aminobenzonitrile and 4-chloroaniline are sluggish and
can be catalysed. However, as can be seen from Table 2,
the more hindered imidazolide 11 undergoes a very sluggish
(11) See MSDS, Applied Biosystems, July 2002.
(12) Bright, R.; Dale, D. J.; Dunn, P. J.; Hussain, F.; Kang, Y.; Mason, C.;
Mitchell, J. C.; Snowdon, M. J. Org. Process Res. DeV. 2004, 8, 1054-
1058.
(9) Marder, O.; Albericio, F. Chem. Oggi 2003, 35-40.
(10) Koppel, I.; Koppel, J.; Leito, I.; Viljar, P.; Grehn, L.; Ragnarsson, U. J.
Chem. Res. (S) 1993, 446-447.
Vol. 9, No. 6, 2005 / Organic Process Research & Development
•
957

Table 2. Comments on the speed of the reaction imidazolide 11 and various amines
Table 3. Percentage conversion of aniline to 2-methyl-2-phenylpropananilide (12, R ) Ph) with 0.5 equiv of various catalysts in
boiling ethyl acetate (78 °C)
Percentage (%) of 2-methyl-2-phenylpropananilide (12) relative to aniline
time (h)
uncatalyzed
HOPy (8)
NO2-HOPy (9)
HOBt‚H2O (2)
Cl-HOBt (6)
HOAt (7)
0
0.25
0.5
1
2
4
0
N/A
N/A
3.5
4.2
9.7
18
26.6
33.2
44
0
N/A
1.6
3.8
7.4
15.1
28.6
41
49.5
52.8
0
N/A
6.2
11.1
21.9
34.3
50.1
63
2.1
N/A
49.4
63.9
72.7
79.1
81.5
82.7
N/A
93.2
6.2
N/A
62.2
74.8
83
85.9
88.5
90.4
N/A
90.9
4.8
67.1
77.9
83
87.6
88.4
89.5
N/A
N/A
N/A
8
13
17
24
68.1
73.6
reaction with aromatic amines, and these reactions benefit
from the addition of a catalyst. Finally, the reaction of the
imidazolide 11 with 4-aminobenzonitrile is unacceptably
slow even in the presence of a catalyst. Hence, it appears
that the reactions of very hindered imidazolides and weakly
basic aromatic amines are probably outside of the scope of
this methodology.
From the reactions shown in Tables 1 and 2 the reaction
of imidazolide 11 with aniline was selected as the reaction
for the side-by-side comparison of the five catalysts. These
reactions were carried out in boiling ethyl acetate with some
representative samples taken at different time points over
24 h. The results were examined with GC/MS and are shown
in Table 3 and Figure 1. These results clearly show that the
order of catalytic effectiveness is:
rates. These results are summarised in Table 4 and Figure
2.
Process Safety Work. The five catalysts were evaluated
for shock sensitivity by drop hammer testing and for thermal
instability by DSC analysis. The results are summarised in
Table 5.
There were two reasons to think that the fused triazole-
based catalysts might have some shock sensitivity. The first
HOAt > Cl-HOBt > HOBt‚H2O > NO2-HOPy > HOPy
Perhaps not surprisingly the relative effectiveness of the
three fused triazole-based catalysts matches the pKa of the
hydroxy proton. For the two least effective catalysts 8 and
9 it was shown that increasing the amount of catalyst and
switching to a higher-boiling solvent gave good reaction
Figure 1. Percentage conversion of aniline to 2-methyl-2-
phenylpropananilide (12, R ) Ph) with 0.5 equiv of various
catalysts in boiling ethyl acetate (78 °C).
958
•
Vol. 9, No. 6, 2005 / Organic Process Research & Development

Table 4. Percentage conversion of aniline to
2-methyl-2-phenylpropananilide (12, R ) Ph) with 2 equiv of
catalyst in boiling n-propyl acetate (102 °C)
percentage (%) of
2-methyl-2-phenylpropananilide (12) relative to aniline
time (h)
uncatalyzed
HOPy (8)
NO2-HOPy (9)
0
0.5
1
2
4
0
N/A
6
13.6
23.5
39.8
55.6
61.7
70.1
0
0
50.5
65.3
77
86.3
91.5
93.5
100
100
11.5
23.4
41.2
60.5
74.4
82.2
85.5
86.3
Figure 3. Prediction of catalyst shock sensitivity by using the
Yoshida DSC/Shock sensitivity correlation.
8
13
17
24
sensitivity would also be observed with higher-energy
impacts.
6-Chloro-1-hydroxybenzotriazole and HOBt‚H2O lie just
above the line and hence would be predicted to be borderline
cases. We found 6-chloro-1-hydroxybenzotriazole to be
shock sensitive when subjected to a 30-H impact. In our
hands 1-hydroxybenzotriazole hydrate was not shock sensi-
tive when subjected to an impact of 50-100 J. However
some other workers have reported that this material is shock
sensitive;14 thus, caution should be taken when handling this
material.15
2-Hydroxy-5-nitropyridine lies below the line and hence
would not be expected to be shock sensitive. This was
confirmed by experiment. 2-Hydroxypyridine does not
exhibit an exotherm at any temperature up to 400 °C and
hence lies well below the line and is highly unlikely to be
shock sensitive. This compound was not tested.
Table 5. The DSC and shock sensitivity results of the five
catalysts
DSC results
shock
onset temp decomp. energy
catalysts
HOAt (7)
sensitivity
(°C)
(J/g)
positive (5 J)
positive (30 J)
190
185
160
293
-2400
-2100
-1740
-1820
æ
Cl-HOBt (6)
HOBt.H2O (2) negativea
NO2-HOPy (9) negative
HOPy (8)
not tested
No exotherm
up to 400 °C
a No audible detonation, at either 50 or 100 J, but some decomposition is
evident when subjected to a 100 J impact.
In conclusion, HOAt has the highest energy of decom-
position on the DSC and has strong shock sensitivity. This
was clearly the most hazardous of the five catalysts.
2-Hydroxypyridine is the safest catalyst, and this is
followed by 2-hydroxy-5-nitropyridine which is not shock
sensitive and has a relatively high onset of decomposition
in the DSC (293 °C). HOBt‚H2O and 6-chloro-1-hydroxy-
benzotriazole have similar safety profiles; choosing between
HOBt‚H2O or 6-chloro-1-hydroxybenzotriazole on a safety
basis would depend on a full assessment of the individual
reaction, although clearly the fact that 6-chloro-1-hydroxy-
benzotriazole is shock sensitive is a major consideration. In
summary the safety order is:
Figure 2. Percentage conversion of aniline to 2-methyl-2-
phenylpropananilide (12, R ) Ph) with 2 equiv of catalyst in
boiling n-propyl acetate (102 °C).
HOPy > NO2-HOPy > HOBt‚H2O ≈ Cl-HOBt > HOAt
was that HOAt had previously been reported to be shock
sensitive.11 The second is that shock sensitivity can be
predicted with reasonable accuracy by plotting DSC results
onto a Yoshida plot.13 Figure 3 shows such a plot; com-
pounds that lie above the line are expected to be shock
sensitive, whereas compounds that lie below the line may
not be. It should be stressed that this prediction should be
backed up by actual testing. HOAt lies above the line and
hence would be expected to be strongly shock sensitive. The
compound had previously been reported to be shock sensi-
tive, but no information was given on the degree of shock
sensitivity.11 This paper now reports that HOAt is shock
sensitive when subjected to a 5-J impact. Of course shock
Conclusions
Five catalysts have been assessed for their safety and their
catalytic effectiveness in promoting imidazolide couplings.
In terms of catalytic effectiveness the order was found to
be:
HOAt > Cl-HOBt > HOBt‚H2O > NO2-HOPy > HOPy
(14) Wehrstedt, K.; Heitkamp D. 1-Hydroxybenzotriazole (HOBt) and Similar
Compounds: Properties and Classification; OECD-IGUS Energetic and
Oxidising Substances (EOS) Working Group Meeting, Paris, France, March
2002. This information can be found on http://www.bam.de/english/
index4.htm. In addition it is reported that anhydrous HOBt is more dangerous
than the hydrate of HOBt.
(13) Yoshida, T. Kogyyo Kayaku 1987, 5, 311-316 (in Japanese).
Vol. 9, No. 6, 2005 / Organic Process Research & Development
•
959

In terms of their process safety the order was almost
exactly the reverse:
General Procedure for Reactions Reported in Tables
1 and 2. The imidazolide solution (of either 10 or 11) (50
mL, 0.39 M, 19.5 mmol) in ethyl acetate was heated to reflux
(78 °C). The imidazolide solution was treated with an amine
(19.5 mmol) (as specified in the table), followed by the
addition of catalyst, HOBt‚H2O 2 (9.75 mmol) (if specified
in the table). The reaction mixtures were stirred and heated
at reflux for 13 h. Representative samples were taken during
this period and analysed by GC/MS.
General Procedure for Reactions Reported in Table
3 (with 1-Hydroxybenzotriazole‚hydrate, 6-Chloro-1-
hydroxybenzotriazole, 5-Nitro-2-hydroxypyridine, and
2-Hydroxypyridine as Catalysts). A solution of 2-methyl-
2-phenylpropanoyl-1H-imidazole 11 (50 mL, 0.39 M, 19.5
mmol) in ethyl acetate was heated to reflux. Aniline (1.82
g, 19.5 mmol) was added, followed immediately by the
catalyst (9.75 mmol) as a single portion. The mixture was
heated at reflux for 24 h during which time representative
samples were taken and analysed by GC/MS.
2-Methyl-2-phenylpropananilide (12, R ) Ph): Pre-
pared Using 1-Hydroxy-7-azabenzotriazole as Catalyst.
CAUTION: 1-Hydroxy-7-azabenzotriazole (HOAt) is strongly
shock sensitive and a highly energetic molecule.11 Hence,
for this experiment the scale was reduced 5-fold and
performed behind a blast screen.16
A solution of 2-methyl-2-phenylpropanoyl-1H-imidazole
11 (10 mL, 0.39M, 3.9 mmol) in ethyl acetate was heated
to reflux. Aniline (364 mg, 3.9 mmol) was added, followed
immediately by 1-hydroxy-7-azabenzotriazole (263 mg, 1.95
mmol) as a single portion. The mixture was heated at reflux
for 8 h during which time samples were taken and analysed
by GC/MS.
HOPy > NO2-HOPy > HOBt‚H2O ≈ Cl-HOBt > HOAt
In searching for a balance of process safety and effective-
ness it seems that 2-hydroxy-5-nitropyridine (NO2-HOPy,
9) a new catalyst for promoting imidazolide couplings offers
the best combination. 2-Hydroxy-5-nitropyridine is less
effective than the fused triazole catalysts; however, it is safe
and can be very effective (as seen in Figure 2); in addition
it is readily available and comparatively inexpensive (price
similar to that of 1-hydroxybenzotriazole hydrate). When the
reaction is complete, it can easily be removed by washing
with base. A full experimental procedure detailing its use is
given in the Experimental Section.
Experimental Section
6-Chloro-1-hydroxybenzotriazole is available on a large
scale and was obtained from GL Biochem Company,
Shanghai, China. 1-Hydroxy-7-azabenzotriazole was pur-
chased from Acros, and 1-hydroxybenzotriazole hydrate was
purchased from Aldrich. All other chemicals were widely
available.
Reactions were carried out with Mettler-Toledo’s Mul-
timax. 1H spectra were recorded on a Varian (1H 300 MHz)
spectrometer, and melting points were measured with a Buchi
Melting Point B-545 apparatus. Flash chromatography was
carried out with Biotage Horizon Flash Collector. GC/MS
data was obtained using a Hewlett-Packard, 6890 series GC
system packed with capillary column (HP-5MS, 30.4 m ×
250 µm × 0.25 µm), with helium as carrier gas (flow rate
23.2 mL/min), and a mass spectrometer as a detector. DSC
were run on DSC 822e, Mettler-Toledo, the scanning rate
was 5 °C/min, with a temperature range up to 400 °C. Shock
sensitivity tests were carried out with a BAM Fall Hammer.
1-(2-Phenylpropanoyl)-1H-imidazole 10. A suspension
of N,N1-carbonyldiimidazole (CDI) (48.4 g, 0.298 mol) in
either ethyl acetate or n-propyl acetate (0.686 L) was treated
with 2-phenylpropanoic acid (40 g, 0.266 mol) over at least
a 10-min period. The initial suspension turned to a homo-
geneous solution as the acid was added. GC/MS confirmed
complete consumption of the starting material. The resulting
solution of the title compound was stored under nitrogen,
and aliquots of this bulk solution were used in subsequent
reactions.
1-(2-Methyl-2-phenylpropanoyl)-1H-imidazole 11. A
suspension of N,N1-carbonyldiimidazole (CDI) (46.7 g, 0.288
mol) in either ethyl acetate or n-propyl acetate (0.66 L) was
treated with 2-methyl-2-phenylpropanoic acid (42.2 g, 0.257
mol) over at least a 10-min period. The initial suspension
turned to a homogeneous solution as the acid was added.
GC/MS confirmed complete consumption of the starting
material. The resulting solution of the title compound was
stored under nitrogen, and aliquots of this bulk solution were
used in subsequent reactions.
General Procedure for Reactions Reported in Table
4. A solution of 2-methyl-2-phenylpropanoyl-1H-imidazole
11 (50 mL, 0.39 M, 19.5 mmol) in n-propyl acetate was
heated to reflux. Aniline (1.82 g, 19.5 mmol) was added,
followed immediately by the catalyst (39 mmol) as a single
portion. The mixture was heated at reflux for 24 h during
which time samples were taken and analysed by GC/MS.
2-Methyl-2-phenylpropananalide (12, R ) Ph): Pre-
pared Using 2-Hydroxy-5-nitropyridine as Catalyst. A
solution of 2-methyl-2-phenylpropanoyl-1H-imidazole 11 in
n-propyl acetate (50 mL, 0.39 M, 19.5 mmol) was heated to
reflux. Aniline (1.82 g, 19.5 mmol) was added, followed
immediately by 2-hydroxy-5-nitropyridine 9 (5.4 g, 39
mmol), both reagents being added in one portion. The
reaction was heated to reflux (102 °C) for 24 h. After cooling,
the reaction mixture was washed with 1 M sodium hydroxide
(2 × 30 mL) and brine, and the resulting solution was dried
over sodium sulphate and evaporated to give the title
compound as a solid (3.9 g, 84%). This material was
chromatographed with high recovery to give the title
compound as a pale-yellow solid, mp 99-99.5 °C (lit.17
1
100.5-101.5 °C), H NMR (CDCl3): 1.65 (6H, s, 2CH3),
6.80 [1H, s (br), NH (disappears following D2O shake)],
7.00-7.45 (10H, m, PhH). The NMR data was in agreement
(15) Unfortunately a further reminder of the potential hazards of handling this
material on a large scale was demonstrated by a recent explosion at Lacamas
Laboratories, Portland, Oregon, U.S.A., which was caused by 1-hydroxy-
benzotriazole. The incident took place on May 9, 2005.
(16) Note that this material has recently become available as a 0.5-0.7 M solution
in DMF from the Sigma-Aldrich company and other suppliers.
(17) Lyle, R. E.; Lyle, G. G. J. Org. Chem. 1953, 18, 1058-1064.
960
•
Vol. 9, No. 6, 2005 / Organic Process Research & Development

with that previously published.18 GC/MS 239 (amide, reten-
tion time 11.15 min), 119 (PhC(CH3)2).
published.18 GC/MS 239 (amide, retention time 11.15 min),
119 (PhC(CH3)2).
2-Methyl-2-phenylpropananalide (12, R ) Ph): Pre-
pared Using 1-Hydroxybenzotriazole Hydrate as Catalyst.
A solution of 2-methyl-2-phenylpropanoyl-1H-imidazole 11
in ethyl acetate (50 mL, 0.39 M, 19.5 mmol) was heated to
reflux. Aniline (1.82 g, 19.5 mmol) was added, followed
immediately by 1-hydroxybenzotriazole hydrate 2 (1.46 g,
9.5 mmol) both reagents being added in one portion. The
reaction was heated to reflux (78 °C) for 24 h. After cooling,
the reaction mixture was washed with 1 M sodium hydroxide
(2 × 30 mL) and brine, and the resulting solution was dried
over sodium sulphate and evaporated to give the title
compound as an oily solid (3.7 g, 80%). This material was
purified by chromatography to give the title compound as a
pale yellow solid, mp 99-100 °C (lit.17 100.5-101.5 °C),
1H NMR (CDCl3): 1.65 (6H, s, 2CH3), 6.80 [1H, s (br),
NH (disappears following D2O shake)], 7.00-7.45 (10H, m,
PhH). The NMR data was in agreement with that previously
Shock Sensitivity Testing. These tests were carried out
according to the method of the UN Recommendation on the
Transport of Dangerous Goods, the Official Journal of the
European Community as well as to the directive 84/449/
EEC and NF T 20-038 Test A.14 with the exception that
the four materials were tested as supplied and were not dried,
ground, or sieved before testing. A minimum of six tests
were performed. One positive result or more in six tests
indicated impact sensitivity.
Acknowledgment
We thank Mr. David Dale for helpful discussions with
regard to the chemical safety aspects of these catalysts and
for performing shock sensitivity testing.
Received for review May 16, 2005.
OP0580062
(18) Olsen, E. O. Acta Chem. Scand. 1975, B29, 953-962.
Vol. 9, No. 6, 2005 / Organic Process Research & Development
•
961