Development of an efficient process towards the benzimidazole BYK308944: A key intermediate in the synthesis of a potassium-competitive acid blocker

Article abstract of DOI:10.1021/op9002505

An entirely new synthesis of the important benzimidazole building block BYK308944 was elaborated using the Stobbe reaction as central element. BYK308944 constitutes an important intermediate for the preparation of 3,6,7,8-tetrahydrochromeno[7,8-d] imidazoles as potassium-competitive acid blockers. The new route relies on the hydroxymethylation of 1,2- dimethylimidazole as cheap starting material followed by oxidation of the corresponding alcohol and Stobbe condensation of the resulting aldehyde with diethyl succinate. All synthetic steps of this new approach were optimized particularly with the goal to establish a process amenable for large- scale preparation.

Full text of DOI:10.1021/op9002505

Organic Process Research & Development 2010, 14, 142–151
Development of an Efficient Process Towards the Benzimidazole BYK308944: A Key
Intermediate in the Synthesis of a Potassium-Competitive Acid Blocker
Matthias Webel,*,† Andreas Marc Palmer,*,‡ Christian Scheufler,† Dieter Haag,§ and Bernd Mu¨ller§
NYCOMED GmbH, Department of Process Chemistry Research, Byk-Gulden-Strasse 2, D-78467 Konstanz, Germany,
NYCOMED GmbH, Department of Medicinal Chemistry, Byk-Gulden-Strasse 2, D-78467 Konstanz, Germany, and NYCOMED
GmbH, Department of Process Chemistry DeVelopment, Robert-Bosch-Strasse 8, D-78224 Singen, Germany
Abstract:
An entirely new synthesis of the important benzimidazole building
block BYK308944 was elaborated using the Stobbe reaction as
central element. BYK308944 constitutes an important intermediate
for the preparation of 3,6,7,8-tetrahydrochromeno[7,8-d]imidazoles
as potassium-competitive acid blockers. The new route relies on
the hydroxymethylation of 1,2-dimethylimidazole as cheap starting
material followed by oxidation of the corresponding alcohol and
Figure 1. Potassium-competitive acid blocker 3,6,7,8-tetrahy-
drochromeno[7,8-d]imidazole - BYK405879.
Stobbe condensation of the resulting aldehyde with diethyl suc-
cinate. All synthetic steps of this new approach were optimized
particularly with the goal to establish a process amenable for large-
scale preparation.
debenzylation of 6, affording the desired target compound
BYK308944 (7).
As already highlighted in our recent publication, the original
medicinal chemistry route was strongly hampered by several
steps that in general are not amenable for a large-scale
manufacturing. Particularly, the use of hydrazine for the
reduction of the nitro group as well as the application of
Introduction
Recently we reported on the “Large scale asymmetric
synthesis of the 3,6,7,8-tetrahydrochromeno[7,8-d]imidazole
cyanoborohydride for the reductive methylation of the primary
BYK405879 - a promising candidate for the treatment of acid-
amine moiety mark a significant drawback of the synthesis due
related diseases” (Figure 1).1 In that article we also briefly
to safety and toxicity concerns. Furthermore, the carbonylation
described the preparation of the intermediate BYK308944
reaction poses a strong hurdle, because only few companies
following a protocol from medicinal chemistry. Herein, we wish
possess the necessary equipment to safely perform this reaction
now to cover our activities concerning a completely new route
due to the high toxicity of carbon monoxide. Additionally, the
to this material, particularly with respect to scale-up.
final debenzylation step is constrained by the use of high
The original medicinal chemistry protocol uses 2-amino-3-
amounts of palladium on charcoal to drive this sluggish reaction
nitrophenol (1) as an expensive starting material (Scheme 1).
Taking into regard the rather lengthy linear synthesis, this
to completion.
However, due to time pressure at the beginning of the
exhibits a strong cost-contributing factor. Applying this material,
development and the urgent need of material for early studies,
the synthesis commences with the preparation of the respective
we were forced to scale-up these procedures to multikilogram
benzyl-protected phenol followed by selective bromination. The
scale together with our external partner. Due to the problems
C2 building block required for the formation of the benzimi-
faced in the course of this synthesis, it quickly turned out that
dazole system was introduced by acetylation of the amino
a more practical and foremost safe alternative was needed. Thus,
function (2). In the next step, the nitro group was reduced, using
a mixture of ferrous chloride and hydrazine (3). After dimethy-
lation of the newly formed primary amino group (4), the
we started in parallel with our efforts to search for an alternative
route.
cyclization to the benzimidazole scaffold was carried out (5),
and the amide functionality was introduced by palladium-
catalyzed amidocarbonylation in the presence of dimethylamine,
leading to compound 6. Finally, the last step comprises the
Results and Discussion
Alternative Synthesis. Only few viable possibilities exist
for the construction of benzimidazoles beside the traditional
approach applied in the medicinal chemistry synthesis. There-
fore, we decided right from the start to follow a strategy that is
focused on the Stobbe reaction of aldehyde 10 with diethyl
succinate (Scheme 2).2,3 This reaction has been already exem-
plified in literature for a variety of heterocyclic aldehydes
* To whom correspondence should be addressed. Telephone: +49 7531 84
4563 (M.W.); +49 7531 84 4783 (A.M.P). Fax: +49 7531 849 4563 (M.W.);
+49 7531 849 2087 (A.M.P.). E-mail: matthias.webel@nycomed.com;
andreas.palmer@nycomed.com.
† Department of Process Chemistry Research, Konstanz.
‡ Department of Medicinal Chemistry, Konstanz.
§ Department of Process Chemistry Development, Singen.
(1) Palmer, A. M.; Webel, M.; Scheufler, C.; Haag, D.; Mueller, B. Org.
Process Res. DeV. 2008, 12, 1170–1182.
(2) Stobbe, H. Chem. Ber. 1893, 26, 2312.
(3) Johnson, W. S.; Daub, G. H. Org. React. 1951, 6, 1–73.
142
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development
10.1021/op9002505  2010 American Chemical Society
Published on Web 12/08/2009

Scheme 1. Original medicinal chemistry route towards BYK308944a
a Reagents and conditions: (i) FeCl3, charcoal, hydrazine hydrate, MeOH, 70-75 °C, 18 h, 82%; (ii) formaldehyde, NaBH3CN, MeOH/AcOH, RT, 2 h, 84%; (iii)
POCl3, 75 °C, 2 h, 86%; (iv) Pd(OAc)2, PPh3, Me2NH (2 M in THF), DMAP, 6 bar CO, DMF, 130 °C, 18 h, 95%; (v) Pd/C, H2, MeOH, 50 °C, 3-9 h, >78%.
Scheme 2. Alternative scale-up route towards BYK308944a
a Reagents and conditions: (i) H2CO (aq), NaOAc/AcOH, 90-96 °C, 43-50 h; (ii) maleic acid, 2-PrOH, 74-82 °C f 7-15 °C, 4-87 h, 40%; (iii) NaOCl/
TEMPO, NaI, NaOH (aq), 8-20 °C, 2.25-24 h; (iv) K3PO4, t-BuOH, 75-85 °C, 1-2 h; (v) MnO2, MsOH, t-BuOH, 75-85 °C, 14-18 h; (vi) diethyl succinate,
NaOEt/EtOH, MeCN, 50-55 °C, 1-12 h, (vii) Ac2O, MeCN, 75-85 °C, 1.75-3.5 h; (viii) HCl (aq), 50-70 °C, 0.5 h; (ix) HCl (aq), 102-112 °C, 3-5 h, 70%; (x)
SOCl2, cat. DMF, MeCN, 75-83 °C, 1-2 h; (xi) HNMe2 × HCl, NEt3, 10-30 °C, 1.5-3 h, 70%.
containing, e.g., a furane,4 pyrazole,4c thiophene,4c,5 pyrrole,6
or indole moiety.7 Thus, we envisaged applying this procedure
to our problem, i.e. the construction of the benzimidazole
BYK308944 (7).
Hydroxymethylation. In order to utilize the Stobbe reaction
for our purposes, it was necessary to gain access to aldehyde
10. As this aldehyde is commercially available only in milligram
to gram quantities, a synthesis had to be established. The
hydroxymethylation of 1,2-dimethylimidazole (8) with form-
aldehyde constituted the first step in the synthesis of 10.8 Not
surprisingly, this hydroxymethylation did not run regioselec-
tively, and a mixture of both regioisomers (9a and 16) along
with dihydroxymethylated product 17 and remaining starting
material 8 was obtained (Scheme 3). When the reaction was
driven to an almost complete consumption of starting material,
a rather long reaction time of 4-5 d was required that is not
feasible from a practical point of view. Furthermore, after a
reaction period of more than 2 d, only a minor increase of the
desired product 9a was noticed, whereas the amount of
dihydroxymethylated product 17 rose to 20%. Thus, a total
conversion of starting material had to be avoided. An optimum
(4) (a) Howard, T. T.; Lingerfelt, B. M.; Purnell, B. L.; Scott, A. E.; Price,
C. A.; Townes, H. M.; McNulty, L.; Handl, H. L.; Summerville, K.;
Hudson, S. J.; Bowen, J. P.; Kiakos, K.; Hartley, J. A.; Lee, M. Bioorg.
Med. Chem. 2002, 10, 2941–2952. (b) Abdel-Wahhab, S. M.; El-Assal,
L. S. J. Chem. Soc. Commun. 1968, 7, 867–869. (c) Simoni, D.;
Romagnoli, R.; Baruchello, R.; Rondanin, R.; Rizzi, M.; Pavani, M. G.;
Alloatti, D.; Giannini, G.; Marcellini, M.; Riccioni, T.; Castorina, M.;
Guglielmi, M. B.; Bucci, F.; Carminati, P.; Pisano, C. J. Med. Chem.
2006, 49, 3143–3152. (d) Gordaliza, M.; Miguel del Corral, J. M.;
Castro, M. A.; Salinero, M. A.; San Feliciano, A.; Dorado, J. M.; Valle,
F. Synlett 1996, 1201–1202.
(5) El-Rayyes, N. R. J. Prakt. Chem. 1973, 315, 300–306.
(6) El-Rayyes, N. R. J. Prakt. Chem. 1973, 315, 295–299.
(7) El-Rayyes, N. R. J. Prakt. Chem. 1974, 316, 386–390.
(8) Jones, R. G. J. Am. Chem. Soc. 1949, 71, 383–386.
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
143

Scheme 3. Hydroxymethylation of 1,2-dimethylimidazole (8)
additive, the reactions performed remarkably better. Advanta-
geously, methanesulfonic acid could be added in amounts as
low as 10%, allowing longer reaction times without inducing
the formation of side products as compared to sulfuric acid.
But even here, despite the longer reaction times, further oxidant
often had to be added several times in the course of reaction
(entries 5 and 6).
was found when the reaction was stopped after 50 h, at the
latest. In this case, not more than 5-7% of the respective
undesired isomers 16 and 17 were obtained in sum, leaving
approximately 40% of unreacted starting material behind.
Despite these flaws, we decided to further elaborate this
protocol, as both raw materials are inexpensive and com-
mercially available in large quantities. During the preparation
of our batches in the kilo-lab, at a reaction temperature of 90-95
°C, a noticeable amount of formaldehyde sublimed in the
condenser leading to the risk of clogging and to a potential
exposure problem through the open layout of the kilo-lab
equipment (air exhaust). Hence, our effort with respect to pilot-
plant batches was directed towards running the reaction prefer-
ably in a closed system. As our pilot-plant equipment is usually
only capable of resisting a pressure of up to 4-6 bar, some
batches were prepared in the autoclave to assess this problem.
Luckily, almost no pressure increase could be observed, and
therefore, the synthesis of further batches could be run safely
in a closed vessel.
The workup consisted of an extraction that worked best when
a pH value above 13 was maintained and n-butanol was used
as extractant.
To our delight, a convenient separation procedure leading
to the pure desired isomer 9a was found using a simple
precipitation step with maleic acid, as the maleate of the wrong
isomer 16 remained in solution together with untransformed
starting material 8. This way, our product could be consistently
isolated in yields around 40% as maleate 9b, also on larger
scale.
Oxidation with Manganese Dioxide. In the beginning of
the project, the oxidation of the alcohol 9a to the aldehyde 10
was elaborated using manganese dioxide (MnO2) as oxidant9
(Table 1). Initially, this reaction was sluggish or often com-
pletely stalled; thus, conversions were low and rarely exceeded
50% (entry 1). We quickly figured out that the quality of the
applied manganese dioxide seemed to be of great importance
(only technical grade MnO2 with a purity of >90% and without
specified mesh size was used). Particularly older batches of this
reagent tended to compromise the reaction, and we assumed
that this was attributed to a lowered activity of MnO2 after
longer storage times. Indeed, improvement was achieved either
by using commercially available activated MnO2 or MnO2 that
had been activated by prolonged heating in our laboratories. A
screening for possible additives revealed a beneficial effect of
acids, e.g., sulfuric acid (H2SO4) or methanesulfonic acid
(MeSO3H). However, applying H2SO4 as additive in an organic
solvent led mainly to decomposition (entry 2). The reaction in
pure sulfuric acid worked, albeit in low yields that could be
slightly improved when activated MnO2 was applied (entries 3
and 4). In contrast, when methanesulfonic acid was used as
The final breakthrough was achieved after identifying tert-
butanol as superior solvent. Using a combination of tert-butanol,
10% of methanesulfonic acid and 4 equiv of MnO2 (entry 7)
led to a clean and complete conversion.
At a later stage, for our pilot-plant batches, Hyflo Super Cel
was added, helping to prevent the precipitation of MnO2 during
the reaction and facilitating the filtration. Additionally, the
amount of MnO2 could be slightly reduced to 3 equiv.
Although this procedure worked well and reproducibly in
our kilo-laboratories and initially in the pilot plant, the high
amount of oxidant needed and the considerable waste stream
made the reaction not very viable for larger scale, particularly
from an environmental point of view. Furthermore, the oxidation
was only feasible using the free imidazole base 9a but not the
maleate 9b that was obtained during the separation process of
the product mixture of the hydroxymethylation reaction.
Unfortunately, the release of the maleate 9b proved to be
problematic, since an aqueous-based release and extraction
seemed to be impossible due to the good water solubility of
the imidazole 9a. Hence, we scrutinized the use of an alcoholate
or a solid anorganic base such as potassium carbonate (K2CO3)
in an appropriate alcoholic solvent. While with alcoholates some
side products occurred, K2CO3 required elevated temperatures
that in turn led to evolution of carbon dioxide. Finally, we found
the optimal conditions by applying potassium phosphate in tert-
butanol. This offered the additional advantage that the same
solvent could be used for the release of 9a and for the
subsequent oxidation reaction of 9a to 10.
Alternative Oxidation Methods. Taking into account all
the above-mentioned disadvantages, a better oxidation method
was undoubtedly needed. We investigated a couple of different
commonly used oxidation protocols (Table 2), including transi-
tion-metal catalyzed oxidations with sodium wolframate,10
ruthenium trichloride,11 and tetra-N-propylammonium perruth-
enate12 (TPAP) using hydrogenperoxide (H2O2)/N-morpholine-
N-oxide (NMO) as oxidants, respectively, and the trichloroiso-
cyanuric acid (TCCA)/2,2,6,6-tetramethylpiperidine-1-oxyl
(TEMPO) system.13 These protocols either did not work at all
(entries 1-3) or delivered only modest yields of 10, even in
the presence of high amounts of expensive catalyst (entries 4,
5). Therefore, we decided to switch to a hypochlorite-TEMPO-
based oxidation procedure.14
Hypochlorite-TEMPO-Based Oxidation. Starting the
optimization experiments with the free base 9a, the elaboration
(10) Sato, K.; Aoki, M.; Takagi, J.; Noyori, R. J. Am. Chem. Soc. 1997,
119, 12386–12387.
(11) Barak, G.; Dakka, J.; Sasson, Y. J. Org. Chem. 1988, 53, 3553–3555.
(12) (a) Griffith, W. P.; Ley, S. V.; Whitcombe, G. P.; White, A. D.
J. Chem. Soc., Chem. Commun. 1987, 1625–1627. (b) Ley, S. V.;
Norman, J.; Griffith, W. P.; Marsden, S. P. Synthesis 1994, 639–666.
(13) Jenny, C. J.; Lohri, B.; Schlageter, M. Eur. Pat. 775684, 1997.
(14) Anneli, P. C.; Biffi, C.; Montanari, F.; Quici, S. J. Org. Chem. 1987,
52, 2559–2562.
(9) Fatiadi, A. J. Synthesis 1976, 65-104 and 133-167.
144
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development

Table 1. Oxidation of the alcohol 9a to the aldehyde 10 with mangenese dioxide
a
entry
time, h
temp., °C
reflux
reflux
80
85
50
reflux
80
solvent
additives
MnO2
yield (10) %
b
1
2
3
4
5
6
7
16
1
3
3
15
18
15
DCM
MeCN
H2SO4 (10% aq)
H2SO4 (10% aq)
MeOH
DCM
t-BuOH
-
Cc
A
-
5% H2SO4 (10% aq)
-
dec
36
51
56
94
88
Cc
Bc
Bc
Bc
B
-
10% MeSO3H
10% MeSO3H
10% MeSO3H
a A ) MnO2 activated at 150 °C overnight in vacuo. B ) commercially activated MnO2. C ) unactivated MnO2. b Incomplete conversion. c Fresh equivalents of MnO2
added after filtration several times during reaction.
Table 2. Alternative oxidation methods of the alcohol 9a to the aldehyde 10
entry
oxidant
catalyst
conversion %
comments
1
2
3
4
5
H2O2
Na2WO4
RuCl3
-
no conv.
H2O2
-
no conv.
H2O2/NMO
TCCA
NMO
RuCl3
-
minimal conv.
TEMPO
[Pr4N]RuO4
∼30
-
30
isolated yield, 8% catalyst
seemed initially a little laborious, and first trials were not
successful, as under standard conditions (NaHCO3 with KBr)
mainly an ipso-substitution of the hydroxymethyl group with
bromine (compound 18) as well as a further aromatic substitu-
tion furnishing the dibrominated compound 19 was observed
(Figure 2, Table 3, entry 1). Working without KBr as additive
in DCM as solvent led exclusively to ipso-chlorination15
(compound 20, entry 2), while in ethyl acetate the reaction
remained incomplete with only >40% conversion together with
amounts of chlorinated byproduct (entry 3).
Finally, switching to another protocol, adding sodium iodide
(NaI) instead of KBr promised to be a better alternative. At the
beginning, 2.0 equiv of NaI were needed to prevent the
occurrence of chlorinated byproduct along with a minimum of
2.0 equiv of NaOCl to observe a noteworthy conversion (Table
4, entries 1 and 2). Lowering the amount of NaI induced a
distinct formation of the undesired chlorinated byproduct (e.g.,
20), also when the amount of hypochlorite was reduced at the
same time (entry 3). The optimum temperature range was
identified within 10-22 °C; below this temperature, accumula-
tion was risked; at temperatures higher than 25 °C, TEMPO
was deactivated14 (entry 4). The reaction appeared relatively
unsusceptible to the applied amount of TEMPO, as it worked
almost equally well with only 0.01 equiv of this reagent; using
0.03 equiv afforded an only marginally higher yield (compare
entries 5 and 6, Table 4). A considerable improvement could
be achieved by increasing the pH value by adding more sodium
hydrogen carbonate solution. Hereby, a nearly quantitative
conversion was reached virtually immediately after the complete
addition of the hypochlorite (entry 7). Moreover, it could be
shown that an additional organic solvent was not required but
the reaction could be carried out solely in water (entry 8).
Subsequently, these optimized conditions were applied to
the oxidation starting from the maleate 9b. In order to adjust
the proper pH value, NaOH solution was used. The reaction
ran smoothly and completely with a slight excess (5%) of
NaOCl (1.58 equiv) with respect to iodide and 0.01 equiv of
TEMPO, using water as the only solvent (entry 9).
A starting pH of >6 was required to ensure a sufficient
reaction rate, quickly ascending to 11-12 during the NaOCl
dosing, resulting in a particularly fast reaction in this pH range
(Figure 3). After completion of the NaOCl addition, the pH
decreased to pH ) 10 within 3 h and to pH ) 8 after further
15 h.
These observations suggested to start the oxidation directly
at pH 11-12 by adding the appropriate volume of NaOH from
the beginning. This led indeed to an improvement, especially
with regard to the amount of NaI that could be considerably
reduced without any noticeable ipso-chlorination. Thus, the
following optimized conditions were elaborated: pH ) 11-12
(adjusted with NaOH), 0.01 equiv of TEMPO, 0.10 equiv of
NaI, and 1.15 equiv of NaOCl at 8-20 °C (entry 10). Using
these conditions, the reaction went to completion in usually less
than an hour. The excess of oxidant was destroyed by sodium
thiosulfate, and after workup the product 10 was telescoped to
the next stage without isolation.
Stobbe Reaction. The following step, the Stobbe condensa-
tion, was initially elaborated as a fast-track solution using an
excess of 10 equiv of diethyl succinate and 5 equiv of sodium
ethylate. After completion of the reaction, the product 11 had
to be extracted into water in order to remove the remaining
diethyl succinate. Either a complete removal of the solvent or
an extensive azeotropic distillation was mandatory to satisfy-
ingly run the upcoming cyclization in the presence of the water-
sensitive acetic anhydride. Although this could still be conducted
in kilo-lab scale, it was obviously not very suitable for the pilot
plant, also under economical aspects. Hence, we concentrated
our efforts on the reduction of the applied amount of diethyl
succinate. Although it appeared at the beginning that the amount
of diethyl succinate could be reduced easily, we had to recognize
Figure 2. Byproducts of the hypochlorite-TEMPO-based
oxidation of the alcohol 9a.
(15) Sarkanen, K. V.; Dence, C. W. J. Org. Chem. 1960, 25, 715–720.
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
145

Table 3. Initial experiments on the hypochlorite-TEMPO-based oxidation of the alcohol 9a to the aldehyde 10
entry
time, h
temp., °C
0
0-RT
10-RT
solvent
oxidant
additives
yield (10) %
a
1
2
3
6
15
15
DCM
DCM
EE
NaOCl/TEMPO
NaOCl/TEMPO
NaOCl/TEMPO
NaHCO3/KBr
NaHCO3
NaHCO3
-
-
b
<40c
a Mixture of 18 and 19. b Mainly ipso-chlorinated product 20. c >50% starting material; chlorinated byproduct.
Table 4. Optimization of the hypochlorite-TEMPO-based oxidation of the alcohol 9a to the aldehyde 10
product composition
SM 9/
base
product 10/
entry
temp., °C
NaI equiv
NaOCla equiv
TEMPO equiv
(aqueous solution)
chlorin. byproduct %
1
2
3
4
5
6
7
8
9
10
15
15
22
30
22
22
22
22
15
15
2.0
2.0
1.0
1.0
2.0
2.0
2.0
2.0
1.5
0.1
1.15
2.0
0.02
0.02
0.02
0.02
0.01
0.03
0.01
0.01
0.01
0.01
7.5 vol. NaHCO3 (sat.)
7.5 vol. NaHCO3 (sat.)
7.5 vol. NaHCO3 (sat.)
7.5 vol. NaHCO3 (sat.)
7.5 vol. NaHCO3 (sat.)
7.5 vol. NaHCO3 (sat.)
15 vol. NaHCO3 (sat.)
15 vol. NaHCO3 (sat.)
3.0 mL NaOH (10%)
1.9 mL NaOH (40%)
65/21/0b
4/79/0b
21/57/9b
32/45/9b
5/78/0b
2/83/0b
0/90/1c
1.5
1.5
2.1
2.1
2.1
,
2.1
0/90/1c e
,
1.58
1.15
3/97/0d e
,
0/97/0d e
a Addition time 55 min. b Determined by HPLC 30 min after hypochlorite addition. c Determined by HPLC 10 min after hypochlorite addition. d Determined by HPLC
overnight after hypochlorite addition using maleate 9b as starting material. e No extra solvent used.
reaction was run using 2.5 equiv of diethyl succinate (entry 7)
and 1.4 equiv of sodium ethylate with an addition time for the
aldehyde 10 of 3-4 h for maximum yields.
A possible rationalization of these outcomes could derive
from the mechanism of the Stobbe condensation16 itself, as
depicted in Scheme 4. The striking attribute of this mechanism
is the generation of the γ-lactone 21. This lactone can either
decay to the desired product 11a or react with a second molecule
of aldehyde 10, leading to the formation of a double-alkylated
adduct 22. Although this kind of adduct is known from
literature,17 we were not able to clearly prove its existence in
our studies. We assume that any 22 formed in the course of
the reaction would be prone to polymerization. Indeed, in the
normal addition mode the side products were obtained as a tarry
mass not suitable for analytical investigation.
Figure 3. Conversion and pH value in the course of the
hypochlorite-TEMPO-based oxidation of the alcohol 9a to the
aldehyde 10.
Two possibilities exist to suppress the formation of the
unwanted side product 22, one consisting in the use of a
reasonable excess of succinate to shift the equilibrium to
γ-lactone 21, reducing the amount of “free” aldehyde 10 that
is able to undergo double-coupling, as it was realized in the
initial procedure. The other possibility comprises the slow
addition of the aldehyde 10 as final component with the same
goal, to keep the aldehyde concentration as low as possible.
Without doubt, the latter variant represented the better option.
The workup and isolation of the Stobbe condensation product
11a proved to be tedious. Initial attempts to acidify the reaction
mixture and to precipitate the product as a hydrochloride salt
after azeotropic distillation were successful on small scale but
did not work anymore in larger batches. This was due to
prolonged distillation times, resulting in esterification with
remaining ethanol or a transesterification with the present excess
that the decrease was limited to a minimum of 4.0 equiv (Table
5, entries 1 and 2).
The systematic screening of the reaction conditions included
the stoichiometry of the applied reagents, their mode of addition,
and possible solvents. These studies revealed that a good
conversion was already reached with only 1.1 equiv of diethyl
succinate and 1.1 equiv of sodium ethylate. The order and time
of addition were absolutely decisive for this remarkable reduc-
tion of reagents. If sodium ethylate was dosed as last component,
as done initially, the above-mentioned reduction of the stoichi-
ometry of diethyl succinate led to a considerable decrease of
conversion to 62% and an even greater loss in yield (entry 3).
On the other hand, when the aldehyde 10 was added to a
preformed mixture of diethyl succinate and sodium ethylate at
50 °C, the reaction proceeded in a quasi-instantaneous, dose-
controlled manner. The yield increased proportionally with the
addition time (entries 4-6). An optimum was reached with a
conversion of >90% after 4 h (>80% after 1 h). In practice, the
(16) Johnson, W. S.; McCloskey, A. L.; Dunnigan, D. A. J. Am. Chem.
Soc. 1950, 72, 514–517.
(17) Miguel del Corral, J. M.; Gordaliza, M.; Castro, A. M.; Salinero, M. A.;
Dorado, J. M.; San Feliciano, A. Synthesis 2000, 154–164.
146
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development

Table 5. Stobbe reaction of the aldehyde 10 with diethyl succinate
addition time
h
diethyl succinate
equiv
sodium ethylate
equiv
temp.
°C
entry
conversion %
1
2
3
4
5
6
7
0.5a
0.5a
2a
7.5
4.0
1.1
1.1
1.1
1.1
2.5
3.0
2.0
1.1
1.1
1.1
1.1
1.4
50
50
50
50
50
50
50
88c
75c
62
82
85
93
95
1b
2b
4b
3-4b
a Addition of sodium ethylate. b Addition of aldehyde 10. c Isolated yield.
Scheme 4. Mechanism of the Stobbe condensation
of diethylsuccinate, respectively, leading to the formation of
the corresponding diester of 11a. Therefore, we decided not to
isolate the Stobbe adduct but to telescope it directly to the next
step.
Intramolecular Cyclization. The starting point for the
cyclization of the Stobbe adduct 11 to the benzimidazole 12
was a kilo-lab procedure using the crude isolated condensation
residue (vide supra) in an excess of pure acetic anhydride (6.5
vol) and sodium acetate (Scheme 2).18 Evidently, this was not
viable for further scale-up, so we started the process improve-
ment using the basic product mixture obtained under the
optimized conditions.
In order to reasonably reduce the amount of acetic anhydride,
the mixture had to be free of remaining alcohols (tert-butanol
from the oxidation step, ethanol from the Stobbe reaction). This
was easily achieved by careful azeotropic removal of these
solvents with toluene. As a result, the stoichiometry of acetic
anhydride could be reduced to 2.2 equiv, which was added over
a period of 1-2 h to a warm toluene solution (75-85 °C),
ensuring a relatively good cyclization. Surprisingly, upon further
scale-up from 5 to 20 kg in the pilot plant a nonstirrable mass
was formed during the azeotropic destillation that even damaged
the slide ring sealing of the stirrer. The trouble shooting
consisted of the addition of acetonitrile that in following batches
effectively prevented agglutination. From that moment on
toluene was completely substituted by acetonitrile, already at
the stage of the azeotropic distillation. Although the process
appeared rather robust at this point, an unsteadiness in yield
was nevertheless noticed (yields 48-67%).
According to our observations, the reaction went consider-
ably smoother when acetic anhydride was added immediately
in larger portions compared to the originally longer addition
times. A putative explanation could be derived from Scheme
5.19 From a mechanistic point of view, for a complete cycliza-
tion, a minimum of 2 equiv of acetic anhydride is necessary;
otherwise the reaction would stop at at the stage of one of the
intermediates. These intermediates, particularly the conjugated
ester 23, are probably prone to side reactions or degradation
(polymerizations etc.). Thus, either the large excess of acetic
anhydride as applied for our initial kilo-lab batches, or presum-
ably an inverse addition regime should be able to circumvent
the problem.
Indeed, if the addition order was inverted, and the product
11 of the Stobbe condensation was added to a preheated mixture
of acetic anhydride and acetonitrile, the reaction seemed to be
much more consistent and reproducibly delivered the expected
benzimidazole 12 in higher yields.
In order to strengthen our hypotheses, an in-depth monitoring
of both addition variants via reaction-IR was performed. Figure
4 displays a stacked IR spectrum acquired under the original
conditions, i.e. adding acetic anhydride to a solution of starting
material 11 at 85 °C over a period of 2 h. The formation of an
intermediate at 1570 cm-1 is striking, ascending shortly after
starting of the acetic anhydride dosing, then slowly decreasing
during further stirring after the end of dosing. As this kind of
accumulation is not noticed under the inverted dosing regime,
this indeed seems to support our assumptions. That barely no
(19) Baraldi, P. G.; Cacciari, B.; Spaluto, G.; Romagnoli, R.; Braccioli,
G.; Zaid, A. N.; Pineda de las Infantas, M. J. Synthesis 1997, 1140–
1142.
(18) Borsche, W.; Kettner, S.; Gillies, M.; Kuhn, H.; Manteuffel, R. Liebigs
Ann. Chem. 1936, 526, 1–22.
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
147

Scheme 5. Mechanism of the intramolecular cyclization of the imidazole 11 to the benzimidazole 12
product peaks appear in this spectrum is due to the fact that
most of the product 12 has precipitated together with byproduct
in the reaction slurry and therefore cannot be monitored by IR.
On the contrary, the inverted addition delivered a clear solution
over the entire reaction course.
For the inverted addition (Figure 5) a typical IR output for
a classical dose-controlled reaction is obtained. At 1710 cm-1
the peak caused by the phenyl ester moiety of the cyclized
product 12 steeply rises right from the beginning of dosing,
whereas a second product peak at 1770 cm-1 is partly
overlapped by an absorption of acetic anhydride. The consump-
tion of acetic anhydride in the course of the reaction is noticeable
at 1820 cm-1.
The upcoming deacetylation followed by saponification of
the ester group was achieved with hydrochloric acid in two
stages without any special problems (Schemes 2 and 5).
For the deacetylation, the acid was added at 50-70 °C until
a pH value of 0.1-0.6 was reached. This was sufficient to
immediately cleave the acetyl group (13). For the saponification
to the carboxylic acid 14, the solvent was stripped at elevated
temperature under ambient pressure followed by subsequent
heating at 100-110 °C.
Amide Formation. Although the final amide formation as
a standard reaction was not expected to cause many problems,
it turned out to be troublesome (Scheme 2).
Preliminary experiments to achieve the amidation via the
acid chloride 15 seemed not very promising at the beginning.
Thus, we decided to postpone a comprehensive elaboration to
a later point in time. Furthermore, a direct conversion of the
ethyl ester 13 to the amide 7 with a number of amine sources,
irrespective of the nature of the applied amine and the reaction
conditions (up to 140 °C in various solvents), afforded either
no product at all or complex mixtures.
The reaction with carbonyldiimidazole (CDI) constituted a
quick solution. But even this conversion was challenging. More
than 2 equiv of CDI were mandatory due to the presence of
the free phenol group, consuming one equivalent of CDI. This
led to an increased evolution of carbon dioxide. Subsequently,
an aqueous solution of dimethylamine was added, the solvent
(DMF) was removed, and the product 7 was precipitated as
hydrochloride in only 50% yield.
Figure 4. Reaction-IR of the Stobbe cyclization of the imidazole
11 to the benzimidazole 12 - normal addition.
Concerning these drawbacks, a better alternative was ur-
gently needed. A more thorough screening of the conditions
for the formation of the acid chloride 15 showed that this could
be conveniently obtained in acetonitrile at 70-80 °C with 1.4
equiv of thionyl chloride. Unfortunately, the acid chloride 15
is only sparingly soluble in acetonitrile, thus forming a suspen-
sion that is not suitable for a quench with aqueous dimethy-
lamine solution. The application of either dimethylamine in THF
or dimethylamine hydrochloride followed by triethylamine
appeared to be feasible solutions. The latter variant was finally
chosen, mainly for economical reasons, and worked out for our
pilot-plant batches.
Figure 5. Reaction-IR of the Stobbe cyclization of the imidazole
11 to the benzimidazole 12 - inverse addition.
148
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development

Conclusion
In summary, a feasible large-scale access to the benzimi-
dazole BYK308944 (7) has been found. This new route allows
the use of much cheaper starting materials and exhibits a shorter
and also safer alternative compared to the original medicinal
chemistry synthesis.
melting point apparatus and are uncorrected. 1H NMR spectra
were recorded with a Bruker AV 300 FT-NMR spectrometer
at a frequency of 300 MHz, or a Bruker AV 400 FT-NMR
spectrometer at a frequency of 400 MHz. 13C NMR spectra
were acquired with a Bruker AV 400 FT-NMR spectrometer
at a frequency of 100 MHz. DMSO-d6 was used as solvent.
The chemical shifts were reported as parts per million (δ ppm)
with tetramethylsilane (TMS) or DMSO as an internal standard
(1H: δTMS ) 0.00 ppm; 13C δDMSO ) 39.50 ppm). High-
resolution mass spectra were obtained on an Agilent LC/MSD
TOF instrument using electrospray ionization (ESI positive).
Elemental analysis was performed on a Carlo Erba 1106 C, H,
N analyzer. Karl Fischer titration was performed on a Metrohm
KF Coulometer 756 KF (Metrohm 774 Sample Oven Proces-
sor). FT-IR spectra were recorded using an Analect ChemEye
from Hamilton Sundstrand-Applied Instrument Technologies
with a Fibreoptic MIR Diamond ATR-Probe from IFS Aachen
on a FlexyLab-reactor system by Systag.
All difficulties arising in this new route could be successfully
overcome: Although the hydroxymethylation of 1,2-dimeth-
ylimidazole (8) could not be accomplished in a selective manner,
the mixture (8, 9a, 16, 17) could be successfully separated by
precipitation with maleic acid. The upcoming oxidation of
alcohol 9a to the aldehyde 10 was initially elaborated utilizing
manganese dioxide as oxidant, switching later to a more
economic hypochlorite-TEMPO-based oxidation protocol. For
this optimization a variety of variables had to be taken into
account.
Due to their mechanistically complex nature, a lot of effort
had to be put into the challenging Stobbe reaction of aldehyde
10 with diethyl succinate and the subsequent cyclization of
imidazole 11 to the benzimidazole 12. Finally, the amide
formation could be achieved by transformation of the carboxylic
acid 14 to the acid chloride 15 and subsequent reaction of 15
with dimethylamine hydrochloride.
(1,2-Dimethyl-1H-imidazol-5-yl)methanol Maleate (9b).
Hydroxymethylation. At a temperature of 10-25 °C, 1,2-
dimethylimidazole (8) (10.0 kg, 104 mol) and sodium acetate
(17.1 kg, 208 mol) were dissolved under stirring in formalde-
hyde (78.0 L, 37% aqueous solution) and acetic acid (14.0 L,
99%). The reaction mixture was heated to 90-96 °C and stirred
for 43-50 h. Subsequently, the mixture was cooled to 10-50
°C and diluted with water (40.0 L). A pH of g13 was adjusted
by slow addition (0.5-1.0 h) of sodium hydroxide solution (65.0
L, 40% aqueous solution). 1-Butanol (55.0 L) was added at
10-25 °C, and the layers were separated. The organic layer
was collected, and the aqueous phase was washed with
1-butanol (2 × 27.0 L). The combined organic phases were
washed (15 min) with saturated aqueous sodium thiosulfate
solution (2 × ∼25.0 L) at 10-25 °C. The solvent (95-100 L)
was stripped under vacuum (50-75 °C) and the product
precipitated to some extent. The obtained crude mixture,
consisting of the two regioisomers 9a and 16, dihydroxylated
product 17, remaining starting material 8, and a residual amount
of 1-butanol (∼10 L), was directly used in the following step
(salt formation). Precipitation as Maleic Acid Salt. At a
temperature of 40-60 °C, the crude mixture obtained in the
formylation step was dissolved in 2-propanol (33.0 L), and
maleic acid (8.9 kg, 76.7 mol) was added. The solution was
heated to 74-82 °C, stirred for 10-60 min under these
conditions, cooled to 35-45 °C, and inoculated with crystals
of (1,2-dimethyl-1H-imidazol-5-yl)methanol maleate (9b). After
seeding, the mixture was cooled to 15-25 °C within 1-3 h,
and stirring was continued for 2-72 h. Finally, the suspension
was cooled to 7-15 °C and stirred for 1-12 h. At this
temperature, the suspension was centrifuged, and the product
was washed with 2-propanol (8-12 L) and dried overnight in
Vacuo at 50 °C, affording 10.1 kg (40% yield) of the pure title
compound 9b (stoichiometric ratio with respect to maleic acid
1:1): mp 105-107 °C. 1H NMR (DMSO-d6, 300 MHz): δ )
2.55 (s, 3 H, 2-CH3), 3.67 (s, 3 H, N-CH3), 4.53 (s, 2 H,
CH2OH), 6.05 (s, 2 H, maleic acid), 7.40 (s, 1 H, 4-H). Free
In order to facilitate the whole synthesis, most of the
reactions were elaborated to be telescoped to the appropriate
next stage. In fact, in this eight-step sequence only two
intermediates and the final product had to be isolated. Finally,
all stages were carried out in pilot-plant scale.
Experimental Section
General. All chemicals were purchased from the major
chemical suppliers and used without any further purification.
1,2-Dimethylimidazole (8) and 2-amino-3-nitrophenol (1) were
purchased from Merck (46 €/kg) and AK-Scientific (450 $/100
g), respectively. Manganese dioxide of the following qualities
has been applied: Screening: A ) Merck (105957, powder
LAB, >90%), B ) Aldrich (63548, technical, activated, >90%);
Pilot-plant batches ) Merck (8.05958, precipitated, active,
>90%). Hypochlorite solution was purchased from Hedinger
(S012121001001) and the assay measured before use. The
progress of the reaction was monitored on Macherey-Nagel
HPTLC plates Nano-SIL 20 UV254 (0.20 mm layer, nano silica
gel 60 with fluorescence indicator UV254) using dichloromethane/
methanol as solvent system. The reactions to 11, 12, and 14
were additionally monitored by HPLC on a Merck Hitachi
HPLC-system using the following conditions: Zorbax SB C8,
150 cm × 4.6 mm × 3.5 µm, flow 1.0 mL/min; wavelength )
275 nm; temperature 30 °C; eluent A ) MeCN, B ) 10 mmol
K2HPO4; injection volume: 10 µL of ∼0.3% solution in A/B,
1:1 v/v; gradient (0-5 min) A ) 2%, B ) 98%; (35 min) A )
60%, B ) 40%; (38-45 min) A ) 2%, B ) 98%. For the
detection of 9b and 10, the same HPLC method and conditions
were used with the exception of wavelength ) 229 nm. Column
chromatography was performed with Merck silica gel 60
(70-230 mesh ASTM) with the solvent mixtures specified in
the corresponding experiment. Spots were visualized by iodine
vapour or by irradiation with ultraviolet light (254 nm). Melting
points (mp) were taken in open capillaries on a Bu¨chi B-540
1
base 9a: H NMR (DMSO-d6, 400 MHz): δ ) 2.24 (s, 3 H,
2-CH3), 3.48 (s, 3 H, N-CH3), 4.37 (s, 2 H, CH2OH), 5.02
(bs, 1 H, CH2OH), 6.60 (s, 1 H, 4-H). 13C NMR (DMSO-d6,
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
149

100 MHz): δ ) 12.9, 29.9, 52.8, 125.0, 131.6, 144.7. Anal.
Calcd for C6H10N2O: C, 57.12; H, 7.99; N, 22.20. Found: C,
56.56; H, 7.87; N, 22.09. HRMS (ESI) m/z C6H11N2O [M +
H]+ Calcd: 127.0866. Found: 127.0872.
2-CH3), 3.80 (s, 3 H, N-CH3), 7.77 (s, 1 H, 4-H), 9.63 (s, 1
H, CHO). 13C NMR (DMSO-d6, 100 MHz): δ ) 12.6, 31.8,
131.6, 142.7, 152.7, 179.3. Anal. Calcd for C6H8N2O: C, 58.05;
H, 6.50; N, 22.57. Found: C, 56.59; H, 6.47; N, 22.04. HRMS
(ESI) m/z C6H9N2O [M + H]+ Calcd: 125.0709. Found:
125.0713.
1,2-Dimethyl-1H-imidazole-5-carbaldehyde (10). Variant
A. The reaction vessel was charged at room temperature with
(1,2-dimethyl-1H-imidazol-5-yl)methanol maleate (9b) (10.0 kg,
41.2 mol), sodium iodide (0.62 kg, 4.13 mol), TEMPO (39.0
g, 0.25 mol), and water (7.0 L). At 15-25 °C, sodium hydroxide
solution (8.5 kg, 40% aqueous solution) was slowly added
(0.5-1.5 h) under stirring until a pH value of 11.5-12.5 was
reached. At 8-20 °C, sodium hypochlorite solution (∼25 L,
∼47 mol, 1.15 equiv, ∼10% aqueous solution) was slowly
added (2-4 h) under stirring. The mixture was stirred at 8-20
°C until complete conversion of the starting material was
accomplished (15-45 min, 20 h max). Sodium thiosulfate (1.97
kg, 12.5 mol) was added at 10-25 °C and the reaction mixture
stirred for further 30-60 min under these conditions, inducing
a color change from green to yellow. The suspension was
filtered and the solid washed with acetonitrile (34 L). The
combined biphasic liquids were retransferred into the reaction
vessel and stirred for 15 min. The layers were separated, and
the aqueous phase was extracted with acetonitrile (2 × 34 L).
The combined organic layers were heated to 60-75 °C,
and the solvent was removed in vacuo (90-95 L). If the water
content of the residue was higher than 1%, more acetonitrile
(20 L) was added and stripped at 60-75 °C in vacuo. The
residue was filtrated and the filter cake washed with acetonitrile
(5 L). The crude product solution of 10 was directly submitted
to the next step (Stobbe condensation). Variant B. At room
temperature, the reaction vessel was charged with (1,2-dimethyl-
1H-imidazol-5-yl)methanol (9a) (500 g, 3.96 mol), manganese
dioxide (1.38 kg, 15.5 mol), methanesulfonic acid (25.7 mL,
0.4 mol), and tert-butanol (6.25 L). The reaction mixture was
heated to 75-85 °C and stirred for 18-22 h. After completion
of the reaction, the mixture was cooled to 35-45 °C and filtered,
and the filter cake washed with a mixture of methanol/
dichloromethane (∼1 L, 1:1). The filtrate was heated to 50-100
°C, the solvent was removed in vacuo, and the remaining solid
was dried overnight in vacuo at 50 °C, affording 490 g (99%
yield) of the pure title compound 10: mp 73-75 °C. Variant
C. An inertized reaction vessel was charged at room temperature
with (1,2-dimethyl-1H-imidazol-5-yl)methanol maleate (9b)
(10.0 kg, 41.2 mol), potassium phosphate (17.5 kg, 82.4 mol),
and tert-butanol (50.0 L). This mixture was heated to 75-85
°C for 1-2 h under stirring. Subsequently, it was cooled to
60-70 °C, filtrated over a pressure filter, and rinsed with hot
(60-70 °C) tert-butanol (30.0 L). The solution was transferred
into a second inertized vessel, and Hyflo Super Cel (10.0 kg),
manganese dioxide (10.0 kg, 124 mol), and methanesulfonic
acid (0.2 L, 3.1 mol) were added. The mixture was heated to
75-85 °C for 14-18 h under stirring, cooled again to 60-70
°C followed by filtration over a pressure filter. The filter cake
was washed with hot (60-70 °C) tert-butanol (40.0 L). All of
the filtrate was transferred back to the vessel and concentrated
in Vacuo at 45-65 °C. This concentrated solution was used
without further purification for the next step (Stobbe condensa-
4-(1,2-Dimethyl-1H-imidazol-5-yl)-3-(ethoxycarbonyl)but-
3-enoic Acid (11). A stirred mixture of diethyl succinate (17.5
kg, 100 mol) and sodium ethylate solution (19.2 kg, 56.5 mol,
20% in ethanol) was heated to 50-55 °C. At this temperature,
a solution of 1,2-dimethyl-1H-imidazole-5-carbaldehyde (10)
(5.0 kg, 40.3 mol, dissolved in 15.0 L acetonitrile) was slowly
added (3-4 h), and stirring was continued for 1-12 h under
these conditions. If the amount of remaining starting material
was higher than 2% (TLC), more sodium ethylate solution
(∼1.4 kg) was added over 15-30 min and the reaction time
extended for further 1-2 h. When complete transformation was
accomplished, the solvent was removed in Vacuo at 60-75 °C
and acetonitrile (42 L) was added. Another portion of solvent
(∼20 L) was removed in Vacuo at 50-70 °C. The resulting
solution of the title product 11 was directly transferred to the
next reaction step (cyclization). A product sample was isolated
by removing the solvent (rest amounts of acetic acid were
1
detected). H NMR (DMSO-d6, 400 MHz): δ ) 1.26 (t, J )
7.0 Hz, 3 H, OCH2CH3), 2.35 (s, 3 H, 2-CH3), 3.48 (s, 2 H,
CH2), 3.56 (s, 3 H, N-CH3), 4.20 (q, J ) 7.0 Hz, 2 H,
OCH2CH3), 7.15 (s, 1 H, CdH), 7.52 (s, 1 H, 4-H), 12.60 (bs,
1 H, COOH). 13C NMR (DMSO-d6, 100 MHz): δ ) 13.2, 14.1,
30.3, 34.3, 60.5, 122.4, 126.2, 127.3, 130.8, 166.8, 171.5.
HRMS (ESI) m/z C12H17N2O4 [M + H]+ Calcd: 253.1183.
Found: 253.1173.
Ethyl 4-(Acetyloxy)-1,2-dimethyl-1H-benzimidazole-6-
carboxylate (12). The vessel was charged with acetic anhydride
(9.9 kg, 97 mol) and acetonitrile (9 L), and the mixture was
heated to 75-85 °C. At this temperature, the residual solution
from the Stobbe condensation (4-(1,2-dimethyl-1H-imidazol-
5-yl)-3-(ethoxycarbonyl)but-3-enoic acid) (11) was added under
vigorous stirring over a period of 1-2 h. Stirring was continued
for 45-90 min under these conditions. If the amount of
remaining starting material was higher than 2% (TLC), more
acetic anhydride (∼0.4 kg) was added over 15-30 min and
the reaction time extended for further 1-2 h. The reaction
solution of the title compound was used without isolation for
the next reaction step (cleavage of the acetyl group). The
analytical data were obtained by taking a sample and removing
the solvent: mp: 166-169 °C. 1H NMR (DMSO-d6, 400 MHz):
δ ) 1.33 (t, J ) 7.1 Hz, 3 H, OCH2CH3), 2.37 (s, 3 H, 2-CH3),
2.56 (s, 3 H, COCH3), 3.81 (s, 3 H, N-CH3), 4.35 (q, J ) 7.1
Hz, 2 H, OCH2CH3), 7.52 (d, J ) 1.3 Hz, 1 H, 7-H), 8.04 (d,
J ) 1.3 Hz, 1 H, 5-H). 13C NMR (DMSO-d6, 100 MHz): δ )
13.5, 14.2, 20.6, 30.2, 60.7, 109.6, 114.9, 123.1, 137.3, 138.4,
140.0, 156.0, 165.6, 168.7. Anal. Calcd for C14H16N2O4: C,
60.86; H, 5.84; N, 10.14. Found: C, 60.82; H, 5.83; N, 10.06.
HRMS (ESI) m/z C14H17N2O4 [M + H]+ Calcd: 277.1183.
Found: 277.1189.
Ethyl 4-Hydroxy-1,2-dimethyl-1H-benzimidazole-6-car-
boxylate (13). At a temperature of 50-70 °C, water (18 L)
was added to the reaction solution of ethyl 4-(acetyloxy)-1,2-
1
tion). H NMR (DMSO-d6, 400 MHz): δ ) 2.38 (s, 3 H,
150
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development

dimethyl-1H-benzimidazole-6-carboxylate (12) and the pH
adjusted with hydrochloric acid (∼18 L, 20% aqueous solution)
to a value of 0.1-0.6. Under these conditions, the acetyl
protecting group was cleaved. The title compound was not
isolated.
under stirring. The mixture containing acid chloride 15 was
stirred at 75-83 °C for 1-2 h, cooled to 10-30 °C, and
dimethyl ammonium chloride (3.6 kg, 44.1 mol) was added in
two portions. Over 1.5-3 h, triethylamine (12.3 kg, 121.6 mol)
was added under cooling, maintaining a temperature of 10-30
°C. Finally, water (25 L) was added, the mixture heated to
60-80 °C and the solvent (∼47 L) removed in Vacuo. The
reaction mixture was cooled to 10-20 °C over 3 h and stirred
further for at least 1 h. The crystallized product 7 was
centrifuged and the filter cake washed with water (∼10 L). The
filter cake was dried overnight in Vacuo at 55 °C, affording 4.0
kg (70% yield) of the pure title compound 7: mp, dec at
239-241 °C. 1H NMR (DMSO-d6, 400 MHz): δ ) 2.51 (s, 3
H, 2-CH3), 2.96 (s, 6 H, N{CH3}2), 3.69 (s, 3 H, N-CH3),
6.53 (d, J ) 1.3 Hz, 1 H, 7-H), 6.98 (d, J ) 1.3 Hz, 1 H, 5-H),
9.95 (s, 1 H, OH). 13C NMR (DMSO-d6, 100 MHz): δ ) 13.3,
29.8, 34.9, 100.0, 105.7, 130.5, 132.0, 136.9, 147.7, 151.4,
170.9. Anal. Calcd for C12H15N3O2: C, 61.79; H, 6.48; N, 18.01.
Found: C, 61.85; H, 6.50; N, 18.01. HRMS (ESI) m/z
C12H16N3O2 [M + H]+ Calcd: 234.1237. Found: 234.1237.
4-Hydroxy-1,2-dimethyl-1H-benzimidazole-6-carboxyl-
ic Acid (14). At ambient pressure, the solvent (∼50 L) was
removed at 80-105 °C and the reaction mixture stirred at
102-112 °C for 1.5-2.5 h. Another portion of solvent (2 ×
∼5 L) was stripped at this temperature (without application of
vacuum) and the remaining mixture stirred for 1.5-2.5 h. If
the amount of remaining ethyl 4-hydroxy-1,2-dimethyl-1H-
benzimidazole-6-carboxylate (13) was higher than 2% (HPLC),
more hydrochloric acid (∼4 L) was added over a period of
5-15 min at 50-70 °C, the reaction time extended for further
1.5-2.5 h, and the solvent removed as described above. At
50-70 °C, water (23 L) was added over a period of 5-30 min.
A pH value of 4.5-5.0 was adjusted by addition of sodium
hydroxide solution (40% aqueous solution) over a period of
0.5-1 h, and a thick suspension was obtained. The suspension
was cooled to 15-30 °C over 1-4 h, to 8-15 °C over 0.5-2
h, stirred for 0.5-2 h and centrifuged. The filter cake was
washed with water (∼8 L) and dried overnight in Vacuo at 50
°C affording 5.8 kg (70% yield) of the pure title compound
14: mp, dec at >300 °C. 1H NMR (D2O + NaOD, 400 MHz):
δ ) 2.41 (s, 3 H, 2-CH3), 3.49 (s, 3 H, N-CH3), 6.94 (d, J )
1.4 Hz, 1 H, 7-H), 7.07 (d, J ) 1.4 Hz, 1 H, 5-H). 13C NMR
(D2O + NaOD, 100 MHz): δ ) 12.3, 29.4, 97.0, 111.0, 131.9,
136.0, 136.8, 152.2, 157.3, 176.9. HRMS (ESI) m/z C10H11N2O3
[M + H]+ Calcd: 207.0764. Found: 207.0772.
Acknowledgment
We are grateful to Ms. Melanie Schieferdecker, Mr. Oliver
Meyer, Mr. Jakob Habel, and Mr. Reinhard Geiger (Nycomed,
Konstanz) as well as Mr. Walter Palosch, Mr. Steffen Weingart,
¨
Mr. Ramazan Onder, Ms. Stefanie Ritzi, Ms. Christiane Hilpert,
Ms. Antje Dorner, and Ms. Tamara Baldus (Nycomed, Singen)
for technical assistance.
Supporting Information Available
4-Hydroxy-N,N,1,2-tetramethyl-1H-benzimidazole-6-car-
boxamide (7). The vessel was charged with 4-hydroxy-1,2-
dimethyl-1H-benzimidazole-6-carboxylic acid (14) (5.0 kg, 24.3
mol), acetonitrile (40 L), dimethylformamide (93 mL, 1.21 mol),
and the mixture heated to 75-83 °C. At this temperature,
thionyl chloride (4.0 kg, 33.6 mol) was slowly added (2-3 h)
Analytical data of byproducts 18-20. This material is
available free of charge via the Internet at http://pubs.acs.org.
Received for review September 23, 2009.
OP9002505
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
151