Application of kinetic modeling and competitive solvent hydrolysis in the development of a highly selective hydrolysis of a nitrile to an amide

Article abstract of DOI:10.1021/op4003054

A combination of mechanism-guided experimentation and kinetic modeling was used to develop a mild, selective, and robust hydroxide-promoted process for of a nitrile to an amide using a substoichiometric amount of aqueous sodium hydroxide in a mixed water and N-methyl-2-pyrrolidone solvent system. The new process eliminated a major reaction impurity, minimized overhydrolysis of the product amide by selection of a solvent that would be sacrificially hydrolyzed, eliminated genotoxic impurities, and improved the intrinsic safety of the process by eliminating the use of hydrogen peroxide. The process was demonstrated in duplicate on a 90 kg scale, with 89% isolated yield and greater than 99.8% purity.

Full text of DOI:10.1021/op4003054

Subscriber access provided by TAIPEI MEDICAL COLLEGE
Full Paper
Application of Kinetic Modeling and Competitive Solvent Hydrolysis in
the Development of a Highly-Selective Hydrolysis of a Nitrile to an Amide
Jeffry K. Niemeier, Roger R Rothhaar, John A Werner, and Jeffrey T Vicenzi
Org. Process Res. Dev., Just Accepted Manuscript • DOI: 10.1021/op4003054 • Publication Date (Web): 11 Feb 2014
Downloaded from http://pubs.acs.org on February 13, 2014
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Organic Process Research & Development is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 21
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Application of Kinetic Modeling and Competitive
Solvent Hydrolysis in the Development of a Highlyꢀ
Selective Hydrolysis of a Nitrile to an Amide
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Jeffry K. Niemeier*, Roger R. Rothhaar*, Jeffrey T. Vicenzi, and John A. Werner
Lilly Research Laboratories, Eli Lilly and Company, Lilly Corporate Center, Indianapolis IN,
46285
ABSTRACT
A combination of mechanismꢀguided experimentation and kinetic modeling was used to develop
a mild, selective, and robust hydroxideꢀpromoted process for conversion of a nitrile to an amide
using a substoichiometric amount of aqueous sodium hydroxide in a mixed water and Nꢀmethylꢀ
2ꢀpyrrolidone solvent system. The new process eliminated a major reaction impurity, minimized
overꢀhydrolysis of the product amide by selection of a solvent that would be sacrificially
ACS Paragon Plus Environment
1

Organic Process Research & Development
Page 2 of 21
1
2
3
4
5
6
7
8
9
hydrolyzed, eliminated genotoxic impurities, and improved the intrinsic safety of the process by
eliminating the use of hydrogen peroxide. The process was demonstrated in duplicate on a 90 kg
scale, with 89 % isolated yield and greater than 99.8 % purity.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
INTRODUCTION
The conversion of a nitrile to an amide by hydrolysis can be carried out under a variety of
conditions. The most straightforward methods involve simple acidꢀmediated or baseꢀmediated
hydrolysis. However, these methods are often complicated by overꢀhydrolysis to the
corresponding carboxylic acid. Alternative conditions with improved selectivity for amide
formation include: metalꢀmediated hydrolysis,1 use of substituted hydroxylamine additives,2
enzymatic hydrolysis,3,4 and hydrogen peroxideꢀbased chemistry.5 This paper details the
development of a highlyꢀselective, aqueous baseꢀcatalyzed process for the preparation of amide
2, a transforming growth factorꢀ β (TGFꢀβ) type I receptor antagonist6,7 that is currently in
clinical development. Reaction selectivity was achieved by careful control of reaction conditions
and use of a solvent with a competitive rate of hydrolysis relative to the amide. Optimization
was aided by kinetic modeling to predict the robustness of the process under manufacturing
conditions.
Early process development efforts indicated a hydrogen peroxideꢀmediated hydrolysis would
be superior to simple acidꢀ or baseꢀcatalyzed processes due to better control of undesired overꢀ
hydrolysis to the acid impurity. The initial process, using K2CO3, 35% H2O2, H2O, DMSO
(Scheme 1, condition “a”), afforded amide 2 in 87% yield at 25 kg scale with little of the
undesired acid 3 formed.
ACS Paragon Plus Environment
2

Page 3 of 21
Organic Process Research & Development
1
2
3
4
Scheme 1. Reaction products observed in the hydrolysis of 1 to 2.
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Reaction Conditons: a) K2CO3, 35% H2O2, H2O, DMSO, carbon treatment (87% yield in Campaign 1) or b) NaOH,
35% H2O2, H2O, NMP, DMSO (87% yield in Campaign 2)
This process was further refined for scale up by using a mixed NMP/DMSO8 solvent system to
reduce sulfur emissions,9 and substituting NaOH for K2CO3 to enhance base solubility (condition
“b” in Scheme 1). This process afforded amide 2 with an average yield of 87% on a scale up to
50 kg. None of the acid hydrolysis product 3 was formed during the reaction.
While this process was relatively robust, discovery of mutagenic Nꢀoxide impurities,10 safety
concerns involving hydrogen peroxide,11 oxygen generation during the process, and increased
concern regarding incineration of sulfurꢀcontaining waste streams on commercial scale prompted
exploration for an alternative, peroxideꢀfree process. However, hydrolysis of nitrile 1 with an
aqueous hydroxide base, such as aqueous NaOH, in a waterꢀmiscible organic solvent gave not
only the expected product 2 and impurity 3, but also compound 4, which resulted from the anion
of amide 2 reacting with nitrile 1. This impurity was particularly problematic as minimal
ACS Paragon Plus Environment
3

Organic Process Research & Development
Page 4 of 21
1
2
3
4
rejection occurred during isolation of 2 by crystallization. Therefore our process optimization
5
6
efforts needed to target conditions which controlled both impurities 3 and 4.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
RESULTS AND DISCUSSION - Development of the Hydroxide-
Mediated Hydrolysis Process
Solvent Selection
Polar aprotic solvent screening for the hydroxideꢀpromoted hydrolysis was focused on solvents
other than DMSO to reduce SOx emissions during waste solvent incineration. Water miscible
solvents were given priority in screening partly because water could be added for direct
crystallization of the product from the reaction mixture. Alcoholic solvents gave either
incomplete conversion or unacceptably high levels of acid 3 depending upon reaction conditions.
N,NꢀDimethylacetamide (DMAC) and 1,3ꢀdimethylꢀ3,4,5,6ꢀtetrahydroꢀ2(1H)ꢀpyrimidone
(DMPU) proved to be quite reactive with NaOH under the reaction conditions, and gave either
low conversion of starting materials, or required large amounts of base to give full conversion.
Both Nꢀmethylꢀ2ꢀpyrrolidone (NMP) and 1,3ꢀdimethylꢀ2ꢀimidazolidinone (DMI) gave full
conversion under the conditions screened (0.5 equiv. NaOH, 100 ºC, 1:4 water to organic solvent
at 5 mL/g of 1), with similar impurity profiles. NMP was chosen over DMI due to availability
and lower cost.
Controlling Acid Impurity 3 – Use of k Ratio Analysis
Minimizing formation of acid 3 was one of the primary goals of our process optimization
work. Some authors have described selective conversion of nitriles to amides as difficult due to
ACS Paragon Plus Environment
4

Page 5 of 21
Organic Process Research & Development
1
2
relatively fast hydrolysis of the amide,12 while others have shown it can be accomplished with
high yield.13,14 Under proper conditions selectivity is very good. Solvent choice is one of the
key factors.15,16
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The acid impurity is formed by what is known as “seriesꢀparallel” or “consecutiveꢀ
competitive” reactions (Scheme 2).
Scheme 2. General Scheme for ConsecutiveꢀCompetitive Reactions
k1
A + B
P + B
P
I
k2
ki = rate constant for reaction i
In our case, the product formation reaction did not consume reactant B (hydroxide), as shown
in Scheme 3.
Scheme 3. Conversion of Nitrile to Amide and Acid by ConsecutiveꢀCompetitve Reactions17
For a reaction scheme of this type, the level of impurity formed will be a strong function of the
percent conversion of the starting material and the ratio of the reaction rate constants k1 and k2 (k
ratio; defined as k1/k2 for purposes of this paper). A general mathematical solution for
ACS Paragon Plus Environment
5

Organic Process Research & Development
Page 6 of 21
1
2
consecutiveꢀcompetitive reactions with various k ratios is shown in Figure 118,19. This
relationship between the k ratio and the amount of product and impurity can be shown to also
hold for this case in which the first reaction is catalytic in reactant B (hydroxide), and there is an
additional reaction consuming some of reactant B; for example, by hydrolysis of the solvent
NMP, as discussed below.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1 illustrates that one can’t determine preferred reaction conditions by simply
comparing the amount of impurity I formed in two experiments, unless they are at the same
percent conversion. Instead, it is necessary to utilize the solution shown in Figure 1 (or the
equation on which it is based) to determine the k ratio under different reaction conditions.
The shaded region in Figure 1 shows the acceptable range of in situ concentrations of the
nitrile 1 (< 0.3%) and acid 3 (< 3%), which were established by measurement of the rejection
efficiencies of both compounds in the crystallization process. This allowed us to establish a
target k ratio of greater than 200 for the process. A process with this level of selectivity allows
for the simultaneous control of both 1 and 3 to acceptable levels at the end of the reaction. At k
ratio of 150 or less at least one of the two components would be outside the acceptable range.
ACS Paragon Plus Environment
6

Page 7 of 21
Organic Process Research & Development
1
2
3
4
5
6
7
8
4.0%
3.5%
3.0%
2.5%
2.0%
1.5%
1.0%
0.5%
0.0%
9
Desired post-reaction concentrations
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
k1/k2 = 150
k1/k2 = 200
k1/k2 = 300
k1/k2 = 500
k1/k2 = 1000
0.0%
0.5%
1.0%
1.5%
2.0%
wt% unreacted starting material (1) remaining
Figure 1. Effect of k ratio and % conversion on acid impurity formation for a consecutiveꢀ
competitive reaction.
Controlling Acid Impurity 3 – Use of Competitive Solvent Hydrolysis
It is well known that basic hydrolysis of a nitrile is a catalytic process, wherein hydroxide is
not consumed. This creates an extra challenge in the control of the reaction endꢀpoint, since
given sufficient time, all the hydroxide would be consumed by hydrolysis of the amide product,
leading to increased impurity and decreasing the yield of the reaction. This issue was addressed
by selection of NMP as a cosolvent, which hydrolyzes at a rate comparable to the product amide
2, thereby protecting it. This strategy would allow use of relatively high concentrations of
ACS Paragon Plus Environment
7

Organic Process Research & Development
Page 8 of 21
1
2
3
4
5
6
7
8
9
hydroxide to give rapid product 2 formation, but prevent excessive overꢀhydrolysis, because the
hydroxide is largely consumed by hydrolysis of the solvent rather than of our desired product.
We found that NMP hydrolyzed at an appropriate rate to serve this purpose while also providing
the necessary solubility for our substrate in the presence of aqueous base (Scheme 4).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 4. Hydroxideꢀcatalyzed hydrolysis of 1 in the presence of NMP
k 1
k 2
k 3
(1)
(2)
(3)
NMP
4-( N-methyl-amino)-butyric acid
(5)
Reaction Temperature
The effect of temperature on the reaction rate and selectivity is shown in Table 1.
Table 1. Hydrolysis Reaction Temperature Screening (6 hr time point)
k1/k2 ratioc
Entrya
Temp (ºC)
1 (%)b
92.2
22.2
0.1
2 (%)b
7.7
3 (%)b
0.0
4 (%)b
0.0
1
2
3
4
25
50
ꢀꢀd
270
420
ꢀꢀd
77.0
97.7
97.0
0.3
0.3
75
2.2
0.0
100
0.0
3.0
0.0
ACS Paragon Plus Environment
8

Page 9 of 21
Organic Process Research & Development
1
2
3
4
5
a) Reaction Conditions: NaOH (0.4 equiv), 22 wt% H2O in NMP (4ꢀ6 vol), 4 h
b) HPLC area%
6
7
8
c) these k ratios were determined using data points near the end of the reaction and the equation shown in endnote
19
9
d) it was not possible to estimate a k ratio for these experiment due to low measured concentrations of either 1 or 3
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
A temperature of 75 °C was chosen as it gave an acceptable reaction rate, nonꢀdetectable levels
of 4, and good selectivity.
Base Stoichiometry Studies
With the knowledge that the hydrolysis of NMP as a side reaction would consume base as the
desired catalytic process proceeded, a screen of NaOH stoichiometry was carried out from 0.25ꢀ
0.60 equivalents, relative to 1. Reasonable reaction times (3ꢀ4 hours) and low levels of both
impurities 3 and 4 were obtained with 0.25 equivalents.
Optimization of NMP/Water Ratio
With preliminary reaction conditions well understood (NMP as solvent, approximately
20 wt % water,20 0.25 equiv. NaOH, 75 °C), the final key parameter, the NMP/water solvent
ratio, was optimized. The NMP amount was held at about 4 volumes relative to starting material
throughout the development effort, as this was sufficient to give full solubility of both starting
material and product under the reaction conditions.
The data in Table 2 clearly demonstrates that as the water content of the reaction was
increased, the final area % of 4 decreased. At the extreme ends of the screening, the reaction rate
for the desired reaction slowed. With regard to k ratio (and hence minimization of acid impurity
3 formation), the optimal amount of water is about 22 wt %. Studies showed impurity 4 is both
ACS Paragon Plus Environment
9

Organic Process Research & Development
Page 10 of 21
1
2
3
4
5
6
7
8
9
formed and degraded during the reaction, but is degraded to a greater extent under the reaction
conditions when higher amounts of water are present. Ultimately, a water content of
approximately 24 wt % was chosen to balance acid impurity 3 formation and coupled impurity 4
degradation rate. This gave an acceptable reaction time (3ꢀ4 h) and good yield.
Table 2. NMP/Water Ratio Screening Results (6 hr time point)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
k1/k2 ratioc
Entrya
Water (wt %)
1 (%)b
37.8
11.8
1.8
2 (%)b
53.0
79.4
96.3
98.3
98.2
98.1
97.4
94.9
3 (%)b
0.5
4 (%)b
7.5
7.0
1.2
0.1
0
1
2
3
4
5
6
7
8
1.1
5.3
12
22
25
27
41
49
60
0.4
320
390
400
350
310
210
140
0.8
0.1
1.4
0.0
1.8
0.0
1.9
0
0.2
2.4
0
3.3
1.7
0
a) Reaction Conditions: NaOH (0.4 equiv), X wt% H2O in NMP (4ꢀ6 vol), 6 h at 75 oC
b) HPLC in situ area%.
c) These k ratio values are averages calculated using several data points near the end of reaction and the equation
shown in endnote 19.
The optimized process afforded amide 2 with a yield of 89% and at least 99.8% purity in each of
two 90 kg batches. The acid hydrolysis product 3 was successfully controlled to 0.12ꢀ0.14 % in
the isolated product. Compound 4 was observed at below 0.05 % by the end of processing and
was undetectable in the isolated product.
DEVELOPMENT AND UTILIZATION OF A KINETIC MODEL
Although the kinetic model previously discussed was useful for calculating k1/k2 ratios from
experimental results, a more rigorous kinetic model appropriate for the stage of development was
developed to provide a deeper understanding of the process, and guide additional development
and optimization.
ACS Paragon Plus Environment
10

Page 11 of 21
Organic Process Research & Development
1
2
3
4
5
6
Development of the Kinetic Model
7
8
9
Table 3 shows the reactions that were included in the model. Although there are often many
steps involved in a reaction mechanism, generally the best approach is to only include reactions
that are rate limiting steps or are important to obtain an accurate mass balance. This is the
approach we used to develop a model.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Reactions for formation and degradation of the coupled impurity 4 were not included in the
model because under all conditions near those of the final selected conditions, the coupled
impurity concentrations were < 1% during the reactions, and less than 0.05 area % at the end of
the reaction.
Although qualitative analytical data showed that NMP hydrolysis was occurring, it proved
challenging to develop a reproducible quantitative method. Therefore, the rate of NMP
hydrolysis was determined indirectly by an overall best fit of the reaction data.
For collection of kinetic data, four reactions were run at small scale (2ꢀ40 g) under the selected
reaction conditions (NMPꢀwater coꢀsolvent system with 24 wt % water), with concentrations
determined by HPLC analysis.21 These reactions differed only in scale and temperature. The
data were fit using the Dynochem® program,22 and the resulting parameters are shown in Table
3. The activation energies were fit using data sets from separate reactions run at 65 ºC and 85
ºC. The 95% confidence intervals for the fitted parameters were less than 15%, indicating a
good fit.
Figure 2 provides a comparison of the experimental data and the prediction from the kinetic
model. Notice the hydroxide concentration is expected to decrease significantly over the course
of the reaction due to consumption by NMP hydrolysis.
ACS Paragon Plus Environment
11

Organic Process Research & Development
Page 12 of 21
1
2
3
4
Table 3. Modeled Reactions and Fitted Kinetic Parameters for NMPꢀWater Coꢀsolvent with
5
6
24 wt % Water23
7
8
9
Ea
(kJ/mol)
Reaction
k at 75 ºC
9.27 Eꢀ3a l/(mole*s)
1.00 E3b
l/(mole*s)
k units
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Hydrolysis
step 1
1 + OHꢀ ꢀ RC(OH)=Nꢀ
RC(OH)=Nꢀ + H2O ꢀ 2 + OHꢀ
87.1 a
Hydrolysis
step 2
100 b
74.5 a
77.0 a
Acid
impurity
formation
2 + OHꢀ ꢀ RCO2 + NH3
3.63 Eꢀ5 a l/(mole*s)
1.91 Eꢀ5 a l/(mole*s)
-
NMP
hydrolysis
NMP + OHꢀ ꢀ MeNH(CH2)3CO2
ꢀ
a) fit using DynoChem software22 b) assigned value for nonꢀrate limiting step
0.5
Nitrile (1), experimental
0.45
Amide(2), experimental
0.4
Acid Impurity (3), experimental
Nitrile (1), predicted
0.35
Amide(2), predicted
0.3
Acid Impurity (3), predicted
0.25
Hydroxide, predicted
0.2
0.05
0.15
0.1
0.05
0
0
100
200
300
400
500
Time(minutes)
0
0
100
200
300
400
500
Time (minutes)
Figure 2. Example of model fit for 24 wt % water at 75 ºC for 40ꢀg run.
ACS Paragon Plus Environment
12

Page 13 of 21
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
For consecutiveꢀcompetitive reactions it is advantageous to add the reagent (in this case
hydroxide) to the substrate rather than the converse in order to minimize impurity formation.24 If
substrate is added to reagent there is an increased tendency to have a high concentration of
product in contact with excess reagent, and this leads to a higher rate of impurity formation. For
this system the mixing time is fast relative to the reaction time, and our practice has been to add
the hydroxide quickly.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Utilization of the Model
Effect of solvent hydrolysis
The hydrolysis of NMP provides a builtꢀin control on the maximum amount of acid impurity
formed regardless of the reaction time, thereby giving a more robust process. The modelꢀ
predicted concentrations of nitrile 1 and acid 3 for the process, both with and without NMP
hydrolysis, are shown in Figure 3. The desired reaction endpoint is less than 3% acid impurity
with less than 0.3% remaining starting material (1). If another solvent was found that provided
the same kinetics for the nitrile and amide hydrolysis reactions, but wasn’t prone to hydrolysis
itself, there would be a relatively narrow window of about 30 minutes in which to stop the
reaction (e.g., by cooling or neutralization). The process with NMP hydrolysis is much more
forgiving, and only requires the reaction time be sufficient (about 3 hours) to reduce the amount
of nitrile 1 to an acceptable concentration. Due to the competitive hydrolysis of NMP, the level
of acid 3 is always lower than the acceptable level of 3%, even at extended reaction times.
ACS Paragon Plus Environment
13

Organic Process Research & Development
Page 14 of 21
1
2
3
4
5
6
7
8
9
5.0
4.5
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
Acid impurity
(3), assuming
no NMP
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
hydrolysis
Nitrile (1),
assuming no
NMP hydrolysis
Acid impurity
(3), assuming
NMP hydrolysis
Nitrile (1),
assuming NMP
hydrolysis
0
50
100
150
200
250
300
350
400
450
500
Time (min)
Figure 3. Comparison of process with and without NMP hydrolysis for 75 ºC reaction
temperature, 0.25 equivalents NaOH. If NMP hydrolysis did not occur, acceptable times would
be between the orange lines. With NMP hydrolysis, time is only constrained on the low end,
with all times greater than 180 min being acceptable, as shown by the green lines.
Simulation of process robustness
The kinetic model was used to study process robustness for deviations that could occur in
manufacturing. The following singleꢀfailure scenarios were among those evaluated.
• Temperature control events (desired reaction T ± 12 ºC25)
• Various NaOH charging events
• Nitrile 1 charging events (± 6% relative to target)
ACS Paragon Plus Environment
14

Page 15 of 21
Organic Process Research & Development
1
2
3
4
• Reaction time events (± 1 hour stir time)
5
6
7
8
9
Only the ꢀ12 ºC temperature control event (reaction carried out at 63 ºC instead of 75 ºC ) gave
results that were out of process specifications, due to excess remaining starting material.
However, a simulation indicated this event could be rectified by reꢀheating the mixture to 75 ºC
(instead of 63 ºC) for one hour, or stirring at 63 ºC for a total of 8 hours. Note that none of the
scenarios resulted in predicted acid impurity 3 above the target in-situ level of 3%.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
CONCLUSIONS
Development of a simple aqueousꢀbase reaction system to carry out the conversion of 1 to 2,
while utilizing NMP as a sacrificial solvent to minimize product loss due to hydrolysis, has
provided a robust, scalable process. A number of safety and process concerns associated with
use of hydrogen peroxide have been eliminated, including the risk of thermal decomposition, inꢀ
process oxygen generation, handling of an oxidizer, and potential formation of genotoxic Nꢀ
oxides in the penultimate step. The process also avoids use of DMSO, thereby eliminating SOx
formation during incineration of process waste streams. Careful control of solvent/water
composition and reaction temperature controlled the final concentration of impurity 4, which is
not rejected during product isolation.
The development of this process was enabled by a crossꢀfunctional approach leveraging both
mechanistic organic chemistry and kinetic modeling. Analysis of the ratio of the kinetic rate
constants (k ratio) for the formation of amide 2 and acid 3 provided a quantitative means for
rapidly identifying reaction conditions that simultaneously controlled both compounds to
acceptable levels. Finally, a kinetic model allowed exploration of process robustness over
ACS Paragon Plus Environment
15

Organic Process Research & Development
Page 16 of 21
1
2
3
4
5
6
various process operating ranges and failure modes without the need for extensive
experimentation.
7
8
9
The optimized process was successfully run twice at 90 kg scale. In each case, amide 2 was
obtained in 89% yield and greater than 99.8% purity. Impurities 3 and 4 were controlled to 0.12ꢀ
0.14% and < 0.01% respectively, demonstrating the robustness of the process.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
EXPERIMENTAL SECTION
General. All reactions were run under a nitrogen atmosphere, unless otherwise specified.
Reagents were used as received from commercial vendors unless otherwise noted. Proton NMR
spectra were obtained at 400 MHz and carbon NMR spectra were obtained at 100.6 MHz. NMR
chemical shifts are reported in δ units referenced to residual proton signals in the deuterated
solvent. Yields are corrected for chemical purity of both the limiting reagent and the product
(i.e., yield = (weight of product X purity/MW of product)/(weight of limiting reagent X
purity/MW of limiting reagent) X 100). If the purity of the product is not specified, it is greater
than 99%. Analytical methods used for reaction monitoring and purity determination as well as
other analytical tests are described in the Supporting Information.
4-(2-(6-Methylpyridin-2-yl)-5,6-dihydro-4H-pyrrolo[1,2-b]pyrazol-3-yl)quinolin-6-
carboxamide monohydrate (2). 4ꢀ(2ꢀ(6ꢀMethylpyridinꢀ2ꢀyl)ꢀ5,6ꢀdihydroꢀ4Hꢀpyrrolo[1,2ꢀ
b]pyrazolꢀ3ꢀyl)quinolineꢀ6ꢀcarbonitrile (1) (90.0 kg, 256 mol) was dissolved in Nꢀ
methylpyrrolidinone (NMP), (368.9 kg, 359 L) in a 500ꢀgal glassꢀlined reactor and was heated to
73ꢀ77 ºC with stirring. To this mixture was added 1 N aqueous sodium hydroxide solution (64
L, 64.0 mol). The resultant solution was stirred at 73ꢀ77 ºC and was monitored for reaction
ACS Paragon Plus Environment
16

Page 17 of 21
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
completion by HPLC analysis (typically 4 samples were taken in the first 3 hours to obtain
kinetic data, with an additional sample taken at 4 h reaction time). When the reaction was
confirmed complete by HPLC (< 0.4 area % 1 and < 0.1 area % 4), it was cooled to 50ꢀ55 ºC.
Water (44.1 L) was added to the solution and the resultant mixture was seeded with 2 (90 g, 0.1
wt %). The seed slurry was stirred at 50ꢀ55 ºC for 40 min, then water (180 L) was added at 50ꢀ
55 ºC over 1 h. The resultant product slurry was cooled to 20ꢀ25 ºC over 1 h and was stirred at
that temperature for 12 h. The slurry was filtered over a singleꢀplate filter. The wet cake was
washed at room temperature with a mixture of NMP/water (46.3 kg NMP/135 L water) followed
by water (540 L). The solids were dried in vacuo at 45 ºCꢀ55 ºC and 28 – 30 mm Hg to give 2 as
the desired monohydrate crystal form (88.0 kg, 89%) as an offꢀwhite solid, mp 137 oC, and an
HPLC purity of 99.8% (HPLC method found in Supporting Information). Anal. Calcd for
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
C22H19N5O·H2O: C, 68.20; H, 5.46; N, 18.08. Found: C, 68.18; H, 5.34; N, 17.90. H NMR
(DMSOꢀd6: δ) 1.74 (s, 3H), 2.63 (m, 2H), 2.82 (br s, 2H), 4.30 (t, J = 7.2 Hz, 2H), 6.93 (m, 1H),
7.37 (s, 1H), 7.41 (d, J = 4.4 Hz, 1H), 7.56 (m, 1H), 7.58 (m, 1H), 8.04, (s, 1H), 8.04 (d, J = 4.4
Hz, 1H), 8.12 (dd, J = 8.8, 1.6 Hz, 1H), 8.25 (d, J = 2.0 Hz, 1H), 8.87 (d, J = 4.4 Hz, 1H). 13C
NMR (DMSOꢀd6: δ 22.56, 23.24, 25.58, 48.01, 109.36, 117.74, 121.26, 122.95, 126.73, 127.16
(2C), 129.01, 131.10, 136.68, 142.98, 147.20, 148.99, 151.08, 151.58, 152.13, 156.37, 167.47.
IR (KBr): 3349, 3162, 3067, 2988, 2851, 1679, 1323, 864, 825 cmꢀ1. HRMS (m/z M+1): Calcd
for C22H19N5O: 370.1653. Found: 370.1662.
ASSOCIATED CONTENT
The Supporting Information section includes:
• effect of solvent ratio, temperature, and equivalents of NaOH on mixture homogeneity
• studies on alternative bases
ACS Paragon Plus Environment
17

Organic Process Research & Development
Page 18 of 21
1
2
3
4
5
6
• analytical methods
This material is available free of charge via the Internet at http://pubs.acs.org.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ꢁ
TOC Graphic
ꢁ
AUTHOR INFORMATION
Corresponding Author
* To whom correspondence should be addressed. Jeffry K. Niemeier: Telephone: (317) 276ꢀ
2066. Eꢀmail: niemeier_jeffry_k@lilly.com. Roger R. Rothhaar: Telephone: (317) 433ꢀ3769. Eꢀ
mail: rrothhaar@lilly.com.
ꢁ
ABBREVIATIONS
DMAC, N,Nꢀdimethylacetamide; DMPU, 1,3ꢀdimethylꢀ3,4,5,6ꢀtetrahydroꢀ2(1H)ꢀpyrimidone;
NMP, Nꢀmethylꢀ2ꢀpyrrolidone; DMI, 1,3ꢀdimethylꢀ2ꢀimidazolidinone; HPLC, high performance
liquid chromatography;
ACS Paragon Plus Environment
18

Page 19 of 21
Organic Process Research & Development
1
2
3
4
5
REFERENCES
6
7
8
9
1 Pombeiro, A. J. L.; Kukushkin, V. Y. Inorg. Chim. Acta 2005, 358, 1ꢀ21.
2 Ma, X.; Lu, M. J. Chem. Res. 2011, 35 (8), 480ꢀ483.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3 Hopmann, K. H.; Guo, J.; Himo, F. Inorg. Chem. 2007, 46 (12), 4850ꢀ4856.
4 Feng, Y. S.; Lee, C.; Chen, L. Int. Biodeterioration Biodegradation 2007, 59 (3), 211ꢀ215.
5 Katritzky, A. R.; Pilarski, B.; Urogdi, L. Synthesis, 1989, 12, 949ꢀ950.
6 Bueno, Lorea; de Alwis, Dinesh P.; Pitou, Celine; Yingling, Jonathan; Lahn, Michael; Glatt,
Sophie; Troconiz, Inaki F. Eur. J. Cancer, 2008, 44, 142ꢀ150.
7 Mundla, Sreenivasa Reddy; Preparation of 2ꢀ(6ꢀmethylpyridinꢀ2ꢀyl)ꢀ3ꢀ[6ꢀamidoquinolinꢀ4ꢀ
yl)ꢀ5,6ꢀdihydroꢀ4Hꢀpyrrolo[1,2ꢀb]pyrazole monohydrate as a transforming growth factorꢀβ
(TGFꢀβ) inhibitor. US 20100120854, 2010.
8 The use of NMP may be regulated in the European Union under REACH Registration,
Evaluation, Authorisation and Restriction of Chemicals (REACH) legislation. The status of
NMP and other solvents continues to evolve.
9 While sulfurꢀcontaining wastes can be incinerated, control of sulfur emissions to regulatory
limits can present a challenge. In addition, DMSO can cause process safety concerns due to the
potential for autocatalytic decomposition and decomposition catalyzed by impurities. See Zhe
Wang, Z.; Richter, S. M.; Gates, B. D.; Grieme, T. A., Org. Process Res. and Dev., 2012, 16(12),
1994ꢀ2000.
ACS Paragon Plus Environment
19

Organic Process Research & Development
Page 20 of 21
1
2
3
4
5
6
7
8
10 Low levels of two Nꢀoxides were formed, with oxidation at either the pyridine or quinoline
nitrogens. Both compounds gave positive results in the Ames assay.
9
11 Process safety concerns with hydrogen peroxide include: the potential for runaway reaction
with gas evolution, initiated by heat and/or catalytic impurities such as inappropriate metals; and
the potential for oxidation reactions during waste handling in the event of inadequate quenching.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
12 Devarajan, T. S.; Pintauro, P. N. Ind. Eng. Chem. Res. 1991, 30, 581ꢀ585.
13 Bendale, P. M.; Khadilkar, B. M. Synth. Commun. 2000, 30 (10), 1713ꢀ1718.
14 Dotani, M.; Ookawa, T. Jpn. Kokai Tokkyo Koho 1994, JP 06116221 A 19940426. |
15 Hall, J. H.; Gisler, M. J. Org. Chem. 1976, 41 (23), 3769ꢀ3770.
16 Kojo, I, K; Awazu, S.; Hanano, M. Chem. Pharm. Bull. 1974, 22 (4), 864ꢀ870.
17 The first reaction is expected to occur in two steps as shown in Table 3.
18 For derivation of the solution, see one of these references: (a) Hill, C.G. Introduction to
Chemical Engineering Kinetics and Reactor Design; John Wiley and Sons: New York, 1977;
Equation 9.3.9. or (b) Levenspiel, O. Chemical Reaction Engineering, 3rd ed.; John Wiley and
Sons: New York, 1999; p 188.
19 The only assumption in the derivation of this solution is that the reactions are irreversible
and first order in the concentration of A, B, and P and that no other side reactions serve to
consume significant amounts of these species. The plot also assumes the initial concentrations
ACS Paragon Plus Environment
20

Page 21 of 21
Organic Process Research & Development
1
2
3
4
5
6
of product P and impurity I were zero. The following equation describes the relationship
7
8
between the product and starting material concentrations, and is the basis of Figure 1.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2 / k1
k






C p




CA
CA
1
=
−

k2
CA0
CA0
C

A0 

1−

k1
20 Relative to total solvent.
21 Experiments were run in virtuallyꢀfull sealed Teflon® tubes to minimize loss of water to the
headspace. Teflon was selected over glass to eliminate the potential for reaction of hydroxide
with glass at elevated temperature. Several samples were collected for HPLC analysis during the
course of the experiments.
22 DynoChem® software website is www.scaleꢀup.com. The program’s fitting routine
minimizes the error between experimental and model values.
23 For purposes of modeling the reactions were presumed to be irreversible, bimolecular, and
not transport limited. The carboxylate forming steps are irreversible because hydroxide is a
much stronger base than the carboxylic acids.
24 Levenspiel, O. Chemical Reaction Engineering, 3rd ed.; John Wiley and Sons: New York,
1999; p 188
25
Seibert, K. D.; Sethuraman, S., Mitchell, J. D.; Griffiths, K. L.; McGarvey, B. J Pharm
Innov 2008, 3, 105–112. This paper describes the rationale behind the selection of temperature
control events of up to 12 oC.
ACS Paragon Plus Environment
21