C-H Arylation of Heterocyclic N-Oxides Through in Situ Diazotisation of Anilines without Added Promoters: A Green and Selective Coupling Process

Article abstract of DOI:10.1021/acs.oprd.6b00117

A green and selective method for the generation of biaryl compounds through C-H arylation of heterocyclic N-oxides, in which the addition of ascorbic acid as a promoter is not required for either the generation of an aryldiazonium species or the subsequent arylation, is presented. Reaction conditions were optimized through multivariate data analysis, including orthogonal projections to latent structures (OPLS) and design of experiments (DoE) methodologies, resulting in further sustainability improvements, and were then applied to a range of substrates to establish the scope and limitations of the process. The reaction was studied using in situ infrared spectroscopy and a mechanism is presented that accounts for the available data from this and previous studies. The reaction was also performed on a multigram scale, with calorimetry studies to support further scale-up of this promoter-free transformation.

Full text of DOI:10.1021/acs.oprd.6b00117

Subscriber access provided by UCL Library Services
Full Paper
C – H Arylation of Heterocyclic N Oxides Through In-Situ Diazotisation Of
Anilines Without Added Promoters: A Green And Selective Coupling Process
Aymeric Pierre Colleville, Richard A.J. Horan, Sandrine Olazabal, and Nicholas Charles Oliver Tomkinson
Org. Process Res. Dev., Just Accepted Manuscript • DOI: 10.1021/acs.oprd.6b00117 • Publication Date (Web): 30 May 2016
Downloaded from http://pubs.acs.org on May 31, 2016
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 40
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
C – H Arylation of Heterocyclic N Oxides Through In-Situ Diaꢀ
zotisation Of Anilines Without Added Promoters: A Green
And Selective Coupling 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
Aymeric P. Colleville,†,‡ Richard A. J. Horan,†,* Sandrine Olazabal† and Nicholas C. O. Tomkinson‡
†API Chemistry, GlaxoSmithKline Research and Development Ltd., Gunnels Wood Road, Stevenage,
Hertfordshire, SG1 2NY, U.K.
‡WestCHEM, Department of Pure and Applied Chemistry, Thomas Graham Building, University of
Strathclyde, 295 Cathedral Street, Glasgow, G1 1XL, U.K.
Corresponding Author
*Email: richard.a.horan@gsk.com
ACS Paragon Plus Environment

Organic Process Research & Development
Page 2 of 40
Table of Contents Graphic
1
2
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
ACS Paragon Plus Environment

Page 3 of 40
Organic Process Research & Development
ABSTRACT: A green and selective method for the generation of biꢀaryl compounds through C—H aryꢀ
lation of heterocyclic N-oxides is presented in which the addition of ascorbic acid as a promoter is not
required for either the generation of an aryldiazonium species or the subsequent arylation. Reaction
conditions were optimised through Multivariate Data Analysis, including Orthogonal Projections to Laꢀ
tent Structures (OPLS) and Design of Experiments (DoE) methodologies resulting in further sustainabilꢀ
ity improvements, and were then applied to a range of substrates to establish the scope and limitations
of the process. The reaction was studied using in-situ infraꢀred spectroscopy and a mechanism is preꢀ
sented that accounts for the available data from this and previous studies. The reaction was also perꢀ
formed on a multigram scale, with calorimetry studies to support further scaleꢀup of this promoterꢀfree
transformation.
1
2
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
Keywords: Radical arylation, Aryldiazonium salts, Green Chemistry, Nꢀoxide heterocycles, Design of
Experiments, Calorimetry
ACS Paragon Plus Environment

Organic Process Research & Development
Page 4 of 40
INTRODUCTION
1
2
3
4
5
6
7
8
9
Palladium catalysed crossꢀcoupling reactions have become a staple of organic synthesis since
their discovery in the 1970s and have been extensively studied in the intervening decades.1 Partly as a
consequence of their wide use, a number of drawbacks with these methods have been highlighted inꢀ
cluding: the preparation of the aryl halide and aryl metal coupling partners; the use of someꢀtimes exotꢀ
ic palladium – ligand combinations and the requirement for elevated reaction temperatures in many
cases. Each of these considerations can have a significant impact on the length, cost and sustainability
of any new synthetic route, especially during scaleꢀup, and can reduce the substrate scope when highꢀ
ly functionalised or sensitive substrates are to be employed.
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
New methodologies focusing on the insertion of a metal complex into aromatic C—H bonds
have recently emerged in an attempt to address some of these issues but the majority still require the
use of a transition metal catalyst,2–9 the longꢀterm sustainability of which, particularly in the case of the
platinumꢀgroup metals, is an area of considerable concern.10
A potentially more sustainable approach is the direct arylation of aromatic rings employing an
aryl radical species generated in-situ.11,12 Aryl radical species 1 can be obtained from the corresponding
arylboronic acids 2,13–16 arylhalides 3,17 aryldiazonium salts 418–21 and diaryliodonium salts 522–24 (Figꢀ
ure 1).
Figure 1. Radicalꢀmediated CꢀH arylation
ACS Paragon Plus Environment

Page 5 of 40
Organic Process Research & Development
Addition of the aryl radical to a second aryl species (Ar—H) followed by oxidation / rearomatisaꢀ
1
2
tion then leads to a biaryl product 6. Amongst the methods for the generation of biaryls, Gomberg and
Bachmann described one of the first radicalꢀmediated arylations of aromatic rings through the in-situ
generation of an aryl radical by reaction of an aryldiazonium salt with a base in water.25–27 Although
conceptually simple, a major limitation of this transformation arises when the aryl radical is reacted with
complex aromatics, leading to chaotic additions and mixtures of regioisomeric products.28 For example,
whereas unsubstituted heterocyclic substrates such as furan, thiophene and pyrrole exclusively led to
the 2ꢀarylated product,18,20,22,23 reactions with pyridine provided a mixture of regioisomers,28 albeit in
reasonable yields. Along with problems arising due to the high reactivity of the aryl radical intermediꢀ
ates, an additional challenge is the handling of the potentially unstable aryldiazonium precursors. Altꢀ
hough these high energy species are routinely handled on large scale,29–32 safety concerns33 still inhibit
their adoption among the wider synthetic community.
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
Recently, Carrillo and coꢀworkers described an elegant method telescoping a diazotisation and
a radicalꢀmediated arylation reaction using ascorbic acid as the sole promoter for both steps (Scheme
1).34
Scheme 1. Telescoped diazotisation / arylation
The in-situ generation of the aryldiazonium salt from a cheap and readily available aniline helps
to address safety concerns around the isolation of these species, while the use of an organic promoter
(ascorbic acid) provides an intriguing alternative to standard metalꢀcatalysed couplings. As well as subꢀ
ACS Paragon Plus Environment

Organic Process Research & Development
Page 6 of 40
strates such as furan 8 which undergo addition of an aryl radical at the 2ꢀposition to give the biaryl
product 9, they showed that pyridine Nꢀoxide 11 also underwent a regioselective addition of the aryl
radical derived from aniline 10 to give the product 12 (Scheme 1). This single example of such an imꢀ
portant heterocyclic structure is significant and represents a major advance in this class of coupling reꢀ
action. Attracted by the reactivity of an aryl radical species with pyridine Nꢀoxide 11 and in line with our
interest in aryldiazonium compounds,32 we sought to expand this methodology to other pharmaceuticalꢀ
ly relevant nitrogen heterocycles. Within this paper we show that quinoline Nꢀoxides and their derivaꢀ
tives are effective coupling partners for arylation with aryldiazonium species and that ascorbic acid is
not required as an additive in this reaction, contrary to previous reports. We also demonstrate how inꢀ
dustryꢀstandard statistical and experimental tools can be used to rapidly optimise the reaction and proꢀ
vide additional mechanistic understanding.
1
2
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
RESULTS AND DISCUSSION
As a starting point to our research we examined a series of nitrogen containing heterocycles of releꢀ
vance to the pharmaceutical industry to establish if the reaction was also applicable to these scaffolds
(Table 1). For the initial screening process the reaction conditions reported by Carrillo and coꢀworkers34
were adapted, reducing the excess of tertꢀbutyl nitrite to 1.2 equivalents and omitting the DMSO cosolꢀ
vent. The reaction was found to be remarkably substrate specific; whilst pyridine Nꢀoxide 11 and pyraꢀ
zine Nꢀoxide 13 could be arylated with outstanding regioselectivity but disappointing yield (Table 1, enꢀ
tries 1 and 2), only traces of the possible regioisomeric adducts 21 and 22 were detected by LC/MS
when pyridazine Nꢀoxide 14 and pyrimidine Nꢀoxide 15 were evaluated as potential substrates (Table
1, entries 3 and 4). Although single ring nitrogen heterocycles did not react in high yield, we were deꢀ
lighted to observe that fused ring systems displayed significantly more promise. For example, quinoline
Nꢀoxide 16 and quinoxaline Nꢀoxide 17 were both successfully arylated under our standard conditions
in 60% and 71% yield respectively (Table 1, entries 5 and 6). Surprisingly, the isomeric isoquinoline
Nꢀoxide 19 gave only traces of the arylated products 25 under these standard conditions (Table 1, enꢀ
try 7).
ACS Paragon Plus Environment

Page 7 of 40
Organic Process Research & Development
Table 1: Screening of heterocyclic Nꢀoxides
1
2
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
Entry
1
Substrate
Product
Yielda
22%
2
19%
3
4
Tracesb
Tracesb
5
6
60%
71%
7
Tracesb
aIsolated yield after purification by column chromatography. A mixture of regioisomers was observed
b
by LC/MS.
ACS Paragon Plus Environment

Organic Process Research & Development
Page 8 of 40
Whilst not exhaustive, this initial series of heterocyclic Nꢀoxides suggested this transformation
should have significantly broader scope than the single example reported by Carrillo and coꢀworkers34
and provided a number of interesting trends that warranted further investigation. These promising preꢀ
liminary results encouraged us to screen different organic solvents in order to evaluate the impact of
the reaction medium on the yield through OPLS analysis.35 The solvent screening results required by
this statistical method were obtained through the evaluation of inexpensive benzocaine 7 and commerꢀ
cially available quinoline Nꢀoxide 16 as model substrates for this transformation in a series of common
organic solvents (Table 2) selected from a PCA model (Figure 2).36
1
2
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 2: PCA solvent model
Lꢀascorbic acid (10.0 mol%)
tBuONO (1.2 equiv)
NH2
N
O
O
O
N
O
Solvent (50 vol)
rt, 17 h
OEt
OEt
7
16
23
ACS Paragon Plus Environment

Page 9 of 40
Organic Process Research & Development
These reaction media were selected according to their difference in terms of physicoꢀchemical
1
2
properties, also referred to as solvent descriptors. These inherent descriptors were identified and reꢀ
ported by Carlson and coꢀworkers36 to describe each solvent through numerical values in nine dimenꢀ
sions: melting point, boiling point, dielectric constant, dipole moment, refractive index, normalised
ReichardtꢀDimroth parameter, lipophilicity (logP) and water solubility. The spatial position of a solvent in
the PCA solvent model in Figure 2 is representative of the combination of those descriptors.
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
Table 2: Solvent screening for the radical arylation of quinoline Nꢀoxide 16
Entry
Solvent
Heptane
Toluene
1,4ꢀDioxane
THF
HPLC yielda
0%
Isolated yieldb
1
2
ꢀ
1%
ꢀ
3
1%
ꢀ
4
1%
ꢀ
5
EtOAc
1%
ꢀ
ꢀ
6
DMSO
EtOH
4%
7
7%
ꢀ
8
Water
9%
ꢀ
9
DMF
16%
31%
42%
50%
19%
33%
40%
60%
10
11
12
Acetone
CH2Cl2
MeCN
aHPLC yield calculated with calibration curve. bIsolated yield after purification by column chromatogꢀ
raphy.
ACS Paragon Plus Environment

Organic Process Research & Development
Page 10 of 40
The reaction did not proceed well in heptane, toluene, 1,4ꢀdioxane, tetrahydrofuran, ethyl acetate,
DMSO, ethanol and water as only low yields were observed by HPLC (Table 2, entries 1ꢀ8). DMF, aceꢀ
tone and dichloromethane gave 16%, 31% and 42% yield of the biaryl product 23 respectively (Table 2,
entries 9–11). Among all the solvents screened, acetonitrile gave the product in 60% isolated yield
(Table 2, entry 12). As mentioned previously, these screening results were used to build the OPLS reꢀ
gression model describing reaction yield as a function of solvent properties in order to increase process
understanding. This regression method aims to find the multidimensional direction in the solvent deꢀ
scriptors space that explains the maximum variance in the yield (Figure 3).
1
2
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 3: OPLS regression model
Descriptors highlighted in green show a statisticallyꢀsignificant positive correlation with reaction yield;
those in red show a statisticallyꢀsignificant negative correlation with reaction yield; and those in dark
blue/grey show a correlation that is not statistically significant (95% confidence interval). The model
(R2X=63.9%, R2Y=74.8%, Q2=43.7%) revealed that solvents with high dipole moment, dielectric conꢀ
stant and normalised ReichardtꢀDimroth37 parameters resulted in high yield whereas increased lipoꢀ
ACS Paragon Plus Environment

Page 11 of 40
Organic Process Research & Development
philicity of the solvent was detrimental to the yield (Figure 3). Accordingly, this model highlights the
clear affinity of the reaction for polar, polarisable solvents, which is consistent with the large number of
charged intermediates involved in the transformation (vide infra). Although the PCA model used conꢀ
tains a wide variety of “green” alternatives, the only solvents predicted to give similar or better yields
than acetonitrile were nitromethane and trifluoroethanol, which are both less desirable from a sustainaꢀ
bility perspective. Therefore acetonitrile was chosen for the further development of this transformation
through Design of Experiment methodology.
1
2
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
Design of Experiments is a multifactorial analysis technique allowing the identification of key reꢀ
action parameters within a previously defined space. With suitable solvent and model substrates estabꢀ
lished, it was decided to evaluate the impact of other reaction parameters (Table 3), including the
ascorbic acid loading (A), the number of equivalents of quinoline Nꢀoxide 16 (B), the solvent volumes
(C) and the temperature (D) through 20 runs using a full factorial model where HPLC yield was used as
the response (see supporting information).
Table 3: Design of Experiment summary with optimised conditions
Solvent
volumes
(vol)
Temperature
Ascorbic acid
(mol %)
(A)
Quinoline
Nꢀoxide 17 (equiv)
(B)
Entry
Settings
(°C)
(D)
(C)
1
2
3
4
Minimum
Centre
1%
10.5%
20%
ꢀ
1.0
2.0
3.0
2.0
10
0 °C
55
100
10
20 °C
40 °C
40 °C
Maximum
Selected
The results were interpreted using DesignꢀExpert (version 7.1.1, StatsꢀEase, Inc) software withꢀ
out transformation in order to obtain the halfꢀnormal plot (Figure 4) of this accurate but not predictive
model due to the significant curvature detected (R2 = 0.9104, Adj R2 = 0.8760, Pred R2 = 0.7795).
ACS Paragon Plus Environment

Organic Process Research & Development
Page 12 of 40
Figure 4: Halfꢀnormal plot using HPLC yield as response
1
2
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
Standardised Effects
Orange: positive effects; Blue: negative effects; Green: error from replicates
The halfꢀnormal plot revealed that the temperature (D) had the greatest influence as reactions
performed at elevated temperatures gave the best results (Figure 4). This study also revealed that the
interaction between the amount of quinoline Nꢀoxide 16 (B) and the temperature (D) also had a signifiꢀ
cant impact on the reaction (Figure 5): whilst a high loading of quinoline Nꢀoxide 16 was beneficial to
the reaction at high temperature (red line), it was found to be detrimental at low temperature (black
line). Although this result does not seem intuitive, it was reproducible and was believed to be symptoꢀ
matic of the mechanistic complexity of the reaction. A response surface design to estimate the curvaꢀ
ture would be required in order to fully understand the actual effect of the curvature on this interaction.
ACS Paragon Plus Environment

Page 13 of 40
Organic Process Research & Development
1
2
3
4
5
6
Figure 5: Quinoline Nꢀoxide equivalents / temperature interaction plot
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
Temperature : 40 °C
Centre points
Temperature : 0 °C
Number of equivalents of quinoline Nꢀoxide 16 (equiv)
While the amount of quinoline Nꢀoxide 16 (B) and the temperature (D) were found to have a proꢀ
nounced impact on the yield, the number of solvent volumes (C) and more surprisingly, the loading of
ascorbic acid (A) had no impact on the reaction outcome within the ranges studied and as a conseꢀ
quence of these results, minimum quantities of these reagents could be used to render this metalꢀfree
process even more sustainable. A compromise consisting of two equivalents of the Nꢀoxide heterocycle
at 40 °C in 10 volumes of solvent was selected as exhibited in Table 4, entry 4. The unexpected result
obtained in this study with regards to the impact of ascorbic acid (A) loading on this transformation
ACS Paragon Plus Environment

Organic Process Research & Development
Page 14 of 40
raised the intriguing possibility of reducing the loading still further and warranted additionnal investigaꢀ
tion of its role in this radical arylation.
1
2
3
4
5
6
7
8
Carrillo and coꢀworkers34 hypothesised the possible role of ascorbic acid in this radical arylation
of heterocycles as shown in Figure 6.
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 6: Hypothetical role of ascorbic acid
In this proposed mechanism, aniline 26 is diazotised in the presence of tertꢀbutyl nitrite and
ascorbic acid to form the aryldiazonium ascorbate salt 27 which reacts to generate the key intermediate
2838 which was believed to generate aryl radicals through spontaneous dediazotisation and generation
of a stable ascorbate radical.38 In order to probe this hypothesis, we elected to prepare the compound
36 according to the method published by Doyle and coꢀworkers39 and to study its behaviour under the
reaction conditions (Scheme 2).
ACS Paragon Plus Environment

Page 15 of 40
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Scheme 2: Tentative synthesis of diazoether 36
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 reaction of 4ꢀ(trifluoromethyl)benzenediazonium tetrafluoroborate 34 with ascorbic acid 35 in
aqueous acetonitrile led to the formation of a precipitate which was isolated by filtration. While standard
analysis of the precipate by 1H NMR, 13C NMR and mass spectrometry were consistent with the reportꢀ
ed structure 36,39 the IR spectrum exhibited three distinct bands in the carbonyl region
(1781 cmꢀ1, 1760 cmꢀ1 and 1693 cmꢀ1) which did not match the single carbonyl functionality of 36. While
this result was unexpected, a detailed study40 by Ley and coꢀworkers has shown that rather than 36,
the isolated product was 37, arising from a JappꢀKlingemann rearrangement (Figure 7).
Figure 7: JappꢀKlingemann rearrangement for hydrazine formation40
ACS Paragon Plus Environment

Organic Process Research & Development
Page 16 of 40
1
2
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 proposal from Carrillo and coꢀworkers34 involving the formation of a diazoether intermediꢀ
ate38 was inconsistent with this observation and we decided to monitor the reaction with an infrared
probe and LC/MS in order to identify potential reaction intermediates. Optimal reaction conditions preꢀ
viously established through DoE were thus adapted with minimal amounts of ascorbic acid used
(1.0 mol%) in a reaction where the product could be easily isolated by precipitation and filtration
(Scheme 3).
Scheme 3: Reaction for IR and LC/MS monitoring
The infrared trace of the reaction mixture was analysed with the software iC IR (version 4.3,
Mettler Toledo) with the attribution of representative wavenumber for every component involved in this
arylation, except for the product 44 as no characteristic band could be identified to monitor its formation
(see supporting information).
The identification of characteristic wavenumbers for each reagent allowed the monitoring of
their behavior over the course of the reaction (Figure 9). As expected, 4ꢀ(trifluoromethyl)aniline 42,
6ꢀmethoxyquinoline Nꢀoxide 43 and tertꢀbutyl nitrite were consumed while the formation of
ACS Paragon Plus Environment

Page 17 of 40
Organic Process Research & Development
4ꢀ(trifluoromethyl)benzenediazonium intermediate 34 was observed. Further analysis of the infrared
spectra collected during this experiment allowed the identification of a potential intermediate with a
99.2% fit through the predicative function ConcIRT (pink line, Figure 8). It is noteworthy that the trend
observed for this predicted intermediate is consistent with the behavior of a reaction intermediate as it
is formed and consumed at the same time. In addition to the prediction of possible intermediates inꢀ
volved in the transformation, a predicted infrared spectrum was also generated for this intermediate
(Figure 9).
1
2
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 8: Wavenumbers monitoring and potential intermediate identification
Lꢀascorbic acid (1.0 mol%)
OMe
OMe
tBuONO (1.2 equiv)
NH2
N
N
F3C
MeCN (10 vol)
40 °C
O
O
F3C
42
43
44 (40%)
Entry
Component
Wavenumber
ACS Paragon Plus Environment

Organic Process Research & Development
Page 18 of 40
1
2
3
4
5
4ꢀ(Trifluoromethyl)aniline 42
4ꢀ(Trifluoromethyl)benzenediazonium 34
6ꢀMethoxyquinoline Nꢀoxide 43
tertꢀButyl nitrite
1125 to 1097 cmꢀ1
725 to 711 cmꢀ1
1232 to 1199 cmꢀ1
1656 to 1540 cmꢀ1
Calculated
1
2
3
4
5
6
7
8
9
Simulated intermediate
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
Moreover, a compound exhibiting a similar behavior to the intemediate identified through ReactIR analꢀ
ysis was also detected by LC/MS with a molecular weight of 333 g.molꢀ1. As similarities were observed
between the infrared spectrum of the intermediate and those of the aniline 42, and the aryldiazonium
35 (Figure 10, bands at 1616 cmꢀ1, 1320 cmꢀ1, 1168 cmꢀ1, 1116 cmꢀ1 and 1062 cmꢀ1), a structure derived
from these scaffolds was envisaged for this intermediate. In addition to this observation, the molecular
weight obtained by LC/MS allowed us to postulate that this intermediate was the triazene 45, arising
from the reaction of aryldiazonium 34 with the aniline 42. This theory was supported by the the ready
preparation and isolation of 45 using conditions similar to those in the coupling reaction (Scheme 4).
Scheme 4: Preparation of triazene intermediate 45
Comparison of the infrared spectrum of isolated triazene 45 (see supporting information) with the calcuꢀ
lated spectrum of the predicted intermediate showed a high degree of similarity (Figure 9).
ACS Paragon Plus Environment

Page 19 of 40
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Figure 9: Comparison of triazene 45 with other reaction components by IR
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
ACS Paragon Plus Environment

Organic Process Research & Development
Page 20 of 40
1
2
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
Aniline
174
Triazene
264
Entry
Aryl diazonium 185
Simulated intermediate
1
2
3
4
5
1614 cmꢀ1
1324 cmꢀ1
1159 cmꢀ1
1109 cmꢀ1
1064 cmꢀ1
N.A.
1614 cmꢀ1
1322 cmꢀ1
1167 cmꢀ1
1118 cmꢀ1
1064 cmꢀ1
1614 cmꢀ1
1317 cmꢀ1
1152 cmꢀ1
1101 cmꢀ1
1063 cmꢀ1
1318 cmꢀ1
1150 cmꢀ1
N.A.
1064 cmꢀ1
Selected among the similarities exhibited by the spectra, a significant medium band at 1614 cmꢀ
1 corresponding to the N—H bend and a strong band around 1320 cmꢀ1 corresponding to the aromatic
amine C—N stretch were observed. This result provides strong evidence of a triazene intermediate
which could either be involved on the catalytic pathway (Figure 10, Pathway A) or simply acts as an
aryldiazonium reservoir (Pathway B).
Figure 10: Postulated reaction pathways from the triazene compound 45
ACS Paragon Plus Environment

Page 21 of 40
Organic Process Research & Development
1
2
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
Hence, in Pathway A, triazene 45 acts as a radical initiator in an analogous manner to methodꢀ
ologies involving triazene intermediates41 to generate a reactive aryl radical 46 and a stabilised aniline
radical 47, the release of nitrogen providing a driving force of this process (Figure 10). These radical
species may then be involved in the arylation shown in Figure 6 to generate the product 44, with the
aniline radical 47 being reduced during the process. Alternatively the triazene may act as an offꢀcycle
reservoir of aryldiazonium 34 which can react with the Nꢀoxide 43 through a yet unknown mechanism
(Pathway B), in which ascorbic acid may or may not be involved. While we feel that Pathway B is most
likely to be radical in nature, an ionic mechanism cannot be ruled out at this stage. The difficulty of inꢀ
vestigating these reactions is compounded by the fact that a radical mechanism would require only a
trace amount of an initiator, following which the chain propogating species may be intermediates in the
mechanism from 43 to 44 as shown in Figure 6 (radical cycle).
During their investigations using furan as the coupling partner, Carrillo and coꢀworkers34
showed that ascorbic acid had a benificial impact on the reaction when starting from a preꢀformed diaꢀ
zonium salt, but no control reaction was reported starting from the aniline. It is possible that the N-
ACS Paragon Plus Environment

Organic Process Research & Development
Page 22 of 40
oxide functionality itself can somehow promote formation of an aryl radical such as 46 when ascorbic
acid is not present (Figure 10, Pathway B). We therefore elected to conduct control reactions with furan
as the coupling partner using the conditions reported by Carrillo and coꢀworkers in the presence and
absence of ascorbic acid (Table 4).34
1
2
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
Table 4: Control reactions
Entry
Conditions
Literature34
Yield
64%
52%
53%
1
2
3
With ascorbic acid
Without ascorbic acid
While the yields were lower than those reported in the literature, identical results were obtained
regardless of the presence or absence of ascorbic acid and confirm that neither the ascorbic acid, nor
the N-oxide functionality are required to promote the generation of an aryl radical intermediate. To add
to the confusion, a number of groups have reported reactions involving the reaction of aryldiazonium
salts in the presence of ascorbic acid where control reactions without this reagent show significantly
inferior results, although some product is often observed.42 What is clear is that the mechanism of theꢀ
se reactions warrants further investigation, and is likely to be significantly more complex than first
thought.
The true test of any new green methodology often comes during scaleꢀup where the sustainabilꢀ
ity advantages become more important. We therefore decided to study this radical arylation using reacꢀ
tion calorimetry in order to demonstrate the synthetic potential of this methodology while generating
data which could support our mechanistic hypothesis. Reaction of 4ꢀ(trifluoromethyl)aniline 42 with
6ꢀmethoxyquinoline Nꢀoxide 43 under optimised conditions (Scheme 5) was carried out on a 5 gram
ACS Paragon Plus Environment

Page 23 of 40
Organic Process Research & Development
scale in a reaction calorimetry vessel equipped with a gas flow monitor with the aim to monitor the enꢀ
ergy of the system and to identify potentially hazardous exothermic and gasꢀevolving events. It is
noteworthy that the product of this arylation precipitated spontaneously from the reaction mixture and
was isolated by filtration in 39% yield, thus delivering 3.9 grams of the target compound 44 (Scheme 5).
1
2
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
Scheme 5: Reaction for scaleꢀup and calorimetry study
The dropwise addition of 4ꢀ(trifluoromethyl)aniline 42 to a mixture containing 6ꢀ
methoxyquinoline Nꢀoxide 43 in acetonitrile at 40 °C proved to be very mildly exothermic with a 3 °C
predicted adiabatic temperature rise. Subsequent controlled addition of tertꢀbutyl nitrite to the reaction
mixture over 10 minutes correlated with a strongly exothermic event and substantial gas evolution (Figꢀ
ure 11) with a calculated 81 °C adiabatic temperature rise. Accumulation (of reagent or intermediates)
was also a concern with 88% of the total heat output occurring after the addition of tertꢀbutyl nitrite was
complete, implying that the reaction was intrinsically slower than the rate of addition. This issue could
be addressed by increasing the rate of reaction (for example by increasing reaction temperature) or by
reducing the rate of tertꢀbutyl nitrite addition and will be examined prior to further scaleꢀup.
Figure 11: Gas and heat output of experiment shown in Scheme 5
ACS Paragon Plus Environment

Organic Process Research & Development
Page 24 of 40
10
9
8
7
6
5
4
3
2
1
0
1
2
3
4
5
6
7
8
2400
2200
2000
1800
1600
1400
1200
1000
800
tertꢀbutyl nitrite addition
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
Heat output
Gas output
600
400
200
0
1.70
2.70
3.70
4.70
5.70
Experiment time [h]
In addition to these results, the chart presented in Figure 11 is consistent with the formation of the inꢀ
termediate discussed previously as two distinct peaks were observed during the calorimetry study. We
suspect that the sharp peak between 1.9 and 2.5 hours corresponds to the direct reaction of the aryldiꢀ
azonium species, while the second peak, between 1.9 and 4.7 hours, which was broader than the first,
could be attributed to the consumption of the triazene intermediate 45. However, given the relatively
modest yield of the reaction, it is also possible that the two peaks correspond to the desired reaction
and an as yet unknown side reaction.
The previous studies allowed the identification of optimal reaction conditions excluding ascorbic
acid which were applied to a range of functionalised anilines and Nꢀoxide heterocycles in order to deꢀ
fine the scope and the limitations of this reaction (Table 5).
Table 5: Scope of the CꢀH arylation of Nꢀoxide heterocycles
ACS Paragon Plus Environment

Page 25 of 40
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Yield
without
35a
Yield
with
35a
Nꢀoxide
Product
Entry
Aniline
heterocycle
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
52%
52%
55%
53%
54%
51%
2
3
4
5
44%
52%
58%
51%
6
7
52%
53%
50%
51%
ACS Paragon Plus Environment

Organic Process Research & Development
Page 26 of 40
Yield
without
35a
Yield
Nꢀoxide
Product
1
2
3
Entry
Aniline
with
heterocycle
35a
4
5
6
7
8
8
70%
65%
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
9
74%
15%
43%
35%
N.A.
22%
43%
34%
10
11
12
aIsolated yield after chromatography.
The electronic properties of both the aniline and the N-oxide heterocycle did not appear to affect
the reaction as all combinations gave similar results with yields around 50%. Nevertheless it is noteꢀ
worthy that aminopyridine derivatives were much less efficient coupling partners compared to their othꢀ
erwise similar aniline counterparts (Table 5, entries 1 and 12). This observation is consistent with our
experience of pyridyldiazonium species exhibiting lower stability than the corresponding phenyldiazoniꢀ
um compounds.32 In contrast, the arylation proved to be compatible with a representative range of funcꢀ
tional groups used in medicinal chemistry such as amide (Table 5, entry 11), esters (Table 5, entries 1,
ACS Paragon Plus Environment

Page 27 of 40
Organic Process Research & Development
2 and 12), a NꢀBoc protected amine (Table 5, entry 4) and halides (Table 5, entries 5 and 7), which
provide significant opportunity for further elaboration. In addition, steric encumbrance of the N-oxide
coupling partner was well tolerated (Table 5, entry 5) with the presence of an orthoꢀmethyl group havꢀ
ing no apparent impact on the yield of the transformation. This study also confirmed the observation of
regiospecificity for the 2ꢀposition of the Nꢀoxide heterocycle as only traces of the potential regioisomeric
adducts were observed by LC/MS during this work. The close similarity between the yields obtained
with and without ascorbic acid provides strong evidence that this additive is not required for the reacꢀ
tion to proceed.
1
2
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
While arylated heterocyclic substrates could be prepared in moderate to good yields through
this radical CꢀH arylation reaction, we elected to apply this methodology to the pharmaceutically releꢀ
vant substrate SB222200 71, a selective, reversible and competitive antagonist of human NKꢀ3 recepꢀ
tor (Figure 12).43
Figure 12: Formal synthesis of SB222200
The CꢀH arylation developed in this work required the preparation of the N-oxide substrate 56
which was achieved in 67% yield by oxidation of methyl 4ꢀcarboxyꢀquinoline 66 using mꢀCPBA. This
example showed the ease of introduction of the Nꢀoxide group on quinoline substrates as this comꢀ
ACS Paragon Plus Environment

Organic Process Research & Development
Page 28 of 40
pound was obtained in good yield with unoptimised reaction conditions. Subsequent arylation of
Nꢀoxide heterocycle 56 was carried out in 70% yield and the methyl ester of 62 was directly converted
to the corresponding amide 69 (85%) in the presence of DABAlꢀMe3 67 and chiral amine 68.44 The
N-oxide group was then easily removed through unoptimised hydrogenation conditions catalysed by
rhodium on activated charcoal and delivered the arylated quinoline 70 in 93% yield. The conversion of
70 to 71 has been reported in the literature45 so our work represents a formal total synthesis of
SB222200.
1
2
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
CONCLUSION
In conclusion, we have developed a sustainable CꢀH radical arylation of quinoline Nꢀoxide deꢀ
rivatives using a variety of industryꢀstandard tools (Multivariate Data Analysis, Design of Experiments,
IR monitoring and reaction calorimetry) to accelerate the research while simultaneously providing valuꢀ
able mechanistic insights. The understanding provided by the careful use of these tools can be used to
support decision making processes when choices and compromises need to be made and we hope
that this demonstration will stimulate their wider adoption within the chemistry community. In our hands,
the use of ascorbic acid as a putative promoter of the reaction is not necessary as very similar yields
were observed in the presence or absence of this material. The use of in-situ monitoring and supportꢀ
ing analytical techniques has allowed the identification of a triazene intermediate which could potentialꢀ
ly act as a precursor to the active radical species and has highlighted the need for further detailed inꢀ
vestigation of the reaction. Our optimised process releases nitrogen, tꢀbutanol and water as the only
coꢀproducts; is compatible with a wide range of functional groups and was successfully performed on a
multigram scale. While the N-oxide functionality is required for high regioselectivity, it is unlikely to be
present in the final product and can be removed under catalytic conditions to afford the corresponding
2ꢀarylquinoline, providing expedient access to this pharmaceutically important family of compounds.
Current work is focused on expanding our understanding of the mechanistic course of this intriguing
reaction and we will report on our findings in due course.
EXPERIMENTAL SECTION
ACS Paragon Plus Environment

Page 29 of 40
Organic Process Research & Development
General Procedure for the screening of N-oxide heterocycles and solvents (Tables 1 and 2): To a
round bottom flask containing ethyl 4ꢀaminobenzoate 7 (250 mg, 1.513 mmol), Nꢀoxide heterocycle
(3.03 mmol) and Lꢀascorbic acid 35 (27 mg, 0.151 mmol) under a nitrogen atmosphere was added deꢀ
gassed solvent (12.5 mL), followed by tertꢀbutyl nitrite (0.240 mL, 1.816 mmol) and the reaction mixture
was stirred for 17 hours at 20 °C. After this time, reaction mixture was evaporated to dryness and the
crude product was purified by silica gel chromatography eluting with either heptane/ethyl acetate or
CH2Cl2/MeOH mixtures to yield the target compound.
1
2
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
2ꢀ(4ꢀ(Ethoxycarbonyl)phenyl)pyridine Nꢀoxide (19): Orange solid (81 mg, 22% yield), Mp 119–
1
120 °C. H NMR (DMSOꢀd6, 400 MHz): δ 8.39ꢀ8.38 (m, 1H), 8.07 (d, 2H, J = 8.6 Hz), 8.00 (d, 2H,
J = 8.6 Hz), 7.72ꢀ7.67 (m, 1H), 7.45ꢀ7.48 (m, 2H), 4.37 (q, 2H, J = 7.1 Hz), 1.36 (t, 3H, J = 7.1 Hz)
13
ppm. C NMR (DMSOꢀd6, 100 MHz): δ 165.3 (CIV), 146.6 (CIV), 140.2, 137.2 (CIV), 130.2 (CIV), 129.6,
128.7, 127.7, 126.1, 125.5, 60.9, 14.1 ppm. HRMS (ESI+): calculated for C14H14NO3 [M + H]+ 244.0960,
found 244.0968.
2ꢀ(4ꢀ(Ethoxycarbonyl)phenyl)pyrazine Nꢀoxide (20): Orange solid (69 mg, 19% yield), Mp 154–
1
159 °C. H NMR (DMSOꢀd6, 400 MHz): δ 8.83 (s, 1H), 8.55 (d, 1H, J = 4.1 Hz), 8.48 (d, 1H,
J = 4.1 Hz), 8.08 (d, 2H, J = 8.5 Hz), 8.01 (d, 2H, J = 8.5 Hz), 4.36 (q, 2H, J = 7.1 Hz), 1.35 (t, 3H,
J = 7.1 Hz) ppm. 13C NMR (DMSOꢀd6, 100 MHz): δ 165.2 (CIV), 148.4, 146.9, 142.5 (CIV), 134.5, 133.7
(CIV), 130.9 (CIV), 129.7, 128.9, 61.0, 14.1 ppm. HRMS (ESI+): calculated for C13H13N2O3 [M + H]+
245.0924, found 245.0921.
2ꢀ(4ꢀ(Ethoxycarbonyl)phenyl)quinoline Nꢀoxide (23): Pale pink solid (264 mg, 60% yield), Mp 196–
200 °C. 1H NMR (DMSOꢀd6, 400 MHz): δ 8.67 (d, 1H, J = 8.8 Hz), 8.19 (d, 2H, J = 8.6 Hz), 8.15 (d, 1H,
J = 8.7 Hz), 8.12 (d, 2H, J = 8.6 Hz), 8.05 (d, 1H, J = 8.5 Hz), 7.90 (dd, 1H, J1 = 8.8 Hz, J2 = 1.5 Hz),
7.81ꢀ7.78 (m, 2H), 4.39 (q, 2H, J = 7.1 Hz), 1.38 (t, 3H, J = 7.1 Hz) ppm. 13C NMR (DMSOꢀd6,
100 MHz): δ 165.3 (CIV), 142.8 (CIV), 141.5 (CIV), 137.8 (CIV), 130.8, 130.3 (CIV), 129.9, 129.7 (CIV),
128.9, 128.7, 128.6, 124.9, 123.4, 119.4, 61.0, 14.2 ppm. HRMS (ESI+): calculated for C18H16NO3
[M + H]+ 294.1127, found 294.1125.
ACS Paragon Plus Environment