Process development for a herbicide intermediate via catalytic carboxylation of an aromatic diazonium compound

Article abstract of DOI:10.1021/op025621m

A laboratory process is described for the preparation of 2-sulfo-4-methoxybenzoic acid (SMBA) via diazotization of 2-amino-4-methoxy-sulfonic acid followed by Pd-catalyzed carbonylation in the presence of water. The most important process parameters were the catalyst precursor and water content. Both a two-step procedure using isolated diazonium compound as well as a one-pot reaction proved to be feasible. A yield of > 95% SMBA was obtained after 3-4 h, using 1 mol % PdCl₂ at 8 bar and 60°C. This is the first technically feasible catalytic carbonylation of an arene diazonium compound.

Full text of DOI:10.1021/op025621m

Organic Process Research & Development 2003, 7, 429−431
Process Development for a Herbicide Intermediate via Catalytic Carboxylation
of an Aromatic Diazonium Compound
Urs Siegrist,*,† Thomas Rapold,‡ and Hans-Ulrich Blaser†
SolVias AG, R 1055.6, P.O. Box, CH-4002 Basel, Switzerland, and Syngenta Crop Protection Mu¨nchwilen AG,
CH-4333 Mu¨nchwilen, Switzerland
Abstract:
A laboratory process is described for the preparation of 2-sulfo-
4-methoxybenzoic acid (SMBA) via diazotization of 2-amino-
4-methoxy-sulfonic acid followed by Pd-catalyzed carbonylation
in the presence of water. The most important process param-
eters were the catalyst precursor and water content. Both a
two-step procedure using isolated diazonium compound as well
as a one-pot reaction proved to be feasible. A yield of >95%
Figure 1. Structures of target compounds.
SMBA was obtained after 3-4 h, using 1 mol % PdCl2 at 8
bar and 60 °C. This is the first technically feasible catalytic
carbonylation of arene diazonium salts (see Figure 3) were
carbonylation of an arene diazonium compound.
therefore of high interest to us. Since 2-amino-4-methoxy-
sulfonic acid was commercially available at reasonable cost,
a feasibility study was started. This note describes our efforts
Background
to develop a technically feasible catalytic process for the
large-scale preparation of SMBA.
2-Sulfo-4-methoxybenzoic acid (SMBA) is a key inter-
mediate in the synthesis of CGA 308 956, a development
herbicide of the former Ciba-Geigy AG (see Figure 1). Since
the compound was not available commercially1 and this
specific substitution pattern is not easily accessible by the
usual synthetic methods for o-sulfobenzoic acids such as gas-
phase oxidation of the corresponding toluenesulfonic acid,
a new and efficient synthesis had to be developed. The first
kilograms for field trials and toxicological tests were prepared
from 2,2′-disulfo-4,4′-dinitro stilbene, a dyestuff intermediate
readily available within Ciba-Geigy. SMBA was obtained
in the four-step synthesis depicted in Figure 2 via oxidation
of the CdC bond and substitution of the nitro group in a
nonoptimized overall yield of ca. 25%. However, it was
obvious from the start that with an atom efficiency of 4.8%2
this could not be a large-scale technical synthesis.
Feasibility Study. First experiments were carried out with
the isolated diazonium salt prepared by diazotization of
2-amino-4-methoxy-sulfonic acid under classical conditions
(H2O, HCl, NaNO2) which readily precipitated from the
reaction solution.6 This inner salt is surprisingly stable, and
according to DSC decomposes only at temperatures >160
°C. The original reactions conditions described by the
Kikukawa team (acetonitrile, NaOAc, 2 mol % Pd(OAc)2)
were, in principle, successful, but several modifications
proved to be beneficial to our system. Sodium acetate was
replaced by water as nucleophile (ca. 2% in acetonitrile, ca.
2 equiv/diazonium compound), and the catalyst was added
just before the reaction in the presence of CO to get
reproducible results. In addition, a reliable analytical method
had to be developed which in our experience is often difficult
for highly water-soluble compounds. Yields of SMBA as
determined by HPLC in the reaction solution were >80%,
but a rather long reaction time of 20 h was needed for good
conversions.
An alternative access to substituted benzoic acids would
be the Pd-catalyzed carbonylation of aryl iodides and
bromides as first described by Heck.3 However, besides the
low catalyst activity, a major problem is the high cost of
most aryl bromides, which in addition are often not available
on large scale. The reports by Kikukawa et al.4,5 on the
Process Optimization. Catalyst costs are certainly an
issue since rather high catalyst loadings are necessary.
Because palladium acetate is very expensive, several other
Pd precursors were tested. Fortunately, commercially avail-
† Solvias AG.
‡ Syngenta Crop Protection Mu¨nchwilen AG.
(1) SMBA (CAS 40567-33-7) is only described in the patent literature, Koike,
W.; Kimoto, T.; Matsui, S. (Ihara Chemicals). DE 2616611, 1977. Purified
samples melted at 122 °C (sintering between 90 and 100 °C); 1H NMR
(DMSO, 250 MHz) 3.94 ppm (s, 3H), 7.16 ppm (m, 1H), 7.47 ppm (m,
1H), 7.93 ppm (m, 1H).
(2) Trost, B. M. Angew. Chem., Int. Ed. Engl. 1995, 34, 259.
(3) Heck, R. F. Palladium Reagents in Organic Synthesis; Academic Press:
London, 1985.
(4) Kikukawa, K.; Kono, K.; Nagira, K.; Wada, F.; Matsuda, T. Tetrahedron
Lett. 1980, 21, 2877. Kikukawa, K.; Nagira, K.; Wada, F.; Matsuda, T. J.
Org. Chem. 1980, 45, 2365. Kikukawa, K.; Kono, K.; Nagira, K.; Wada,
F.; Matsuda, T. J. Org. Chem. 1981, 46, 4413.
(5) For a recent overview, see: Beller, M. In Applied Homogeneous Catalysis
by Organometallic Complexes, 2nd ed.; Cornils, B., Herrmann, W. A., Eds.;
Wiley-VCH: Weinheim, 2002; p 145.
(6) Experimental procedure for the diazotization: 48 g of 2-amino-4-methoxy-
sulfonic acid were suspended in 100 mL of water and 100 mL of 37%
hydrochloric acid and cooled to 10 °C. Sodium nitrite (10.4 g, dissolved
in 10 mL water) was added within 2 h under stirring, and the reaction
mass was stirred for an additional 1 h; 5.6 g of sulfamic acid was then
added. The precipitated product was filtered and washed with 20 mL of 2
N HCl, methanol, and diethyl ether, respectively. The light yellow crystals
were dried at rt; yield: 98%.
10.1021/op025621m CCC: $25.00 © 2003 American Chemical Society
Published on Web 03/13/2003
Vol. 7, No. 3, 2003 / Organic Process Research & Development
•
429

Figure 2. Kilogram-scale synthesis starting from dyestuff intermediate.
Figure 3. Typical conditions for the preparation of mixed anhydrides from diazonium salts.
Figure 4. Laboratory process for the production of SMBA.
Table 1: Optimized reaction conditions for the synthesis of
able solid PdCl2 proved to be just as effective, whereas the
cheaper PdCl2 solutions or Pd-dibenzylideneacetone com-
plexes gave significantly lower selectivities or yields or both.
First experiments showed that the palladium can be
separated from the reaction solution and recovered by adding
active carbon followed by mild reduction by hydrogen.7
Recovery of Pd was >95%, but under these nonoptimized
conditions, some acetonitrile was hydrogenated which can
lead to problems in the next reaction step. This aspect was
not investigated further.
Since acetonitrile was confirmed to be the solvent of
choice, solvent effects were studied only within narrow
limits. The reaction could be run in water but yields dropped
to 40%; in the absence of water, formation of the anhydride
could be expected (which would be an advantage for the
next reaction step), but the reaction stopped after about 20%
conversion. In the end, the optimal amount of water was
5-15 equiv relative to the amount for the diazonium
intermediate. Using 1 mol % PdCl2 under an initial CO
pressure of 8 bar (venting the gas phase 2-3 times to remove
the nitrogen) at 60 °C gave >95% of SMBA (HPLC) within
6 h (see Figure 4).
SMBA8
concentration of
diazonium compound
PdCl2
water content
p (CO), T
reaction time
active carbon (0.1 bar H2, rt)
20% w/w in acetonitrile
1mol % (s/c 100)
14 equiv/diazonium compd
8 bar (variant 2), 60 °C
3-4 h
4 g/mol diazonium compd,
reaction time 2 h
>95%
(ca. 85%, H2O, byproducts)
85-95%
isolated yield
(content in solid)
recovered Pd on active carbon
With the help of a heat-flow calorimeter to get information
on the course of the very exothermic reaction (ca. 330 kJ/
mol), two variants were studied for the removal of N2: (1)
Continuous dosage of CO via a massflow controller com-
bined with the slow release of a mixture of N2 and CO (total
pressure 8 bar). (2) Stepwise reaction by pressurizing to 8
bar CO, waiting for all CO to react (ca. 45 min), releasing
the nitrogen, and repeating the procedure four times. Both
procedures gave similar maximum heat flows and yields; in
variant 1 about 20-40% excess of CO was needed.
Towards a Pilot Process. In this development phase the
following aspects were investigated: removal of nitrogen
on a larger scale, optimization of the reaction variables such
as concentration of starting material, water content, and so
on, suitability of the product for the next reaction step, and
safety aspects.
(8) Experimental procedure using isolated diazonium intermediate: 100 g of
dry diazonium compound together with 277 g of acetonitrile and 55.8 g of
water were placed in an 850-mL glass autoclave. The reactor was flushed
three times with 8 bar argon. PdCl2 (0.83 g dissolved in 17 g of acetonitrile)
was added and the reactor immediately pressurized to 8 bar CO. The
reaction mixture was heated to 60 °C; after 30 min the autoclave was vented
to 3 bar and the pressure increased again to 8 bar with CO; this cycle was
repeated five times. After 3 h the pressure was completely released. After
evaporation the light-brown 2-sulfo-4-methoxybenzoic acid was dried at
60 °C: yield 97%.
(7) For a precedent, see: Baumeister, P.; Meyer, W.; Oertle, K.; Seifert, G.;
Siegrist, U.; Steiner, H. Chimia 1997, 51, 144.
430
•
Vol. 7, No. 3, 2003 / Organic Process Research & Development

We also investigated the possibility to carry out the
reaction without isolating the diazonium intermediate. A
“one-pot” procedure would, of course, be a real advantage
and would give considerable cost savings. A major problem
is the oxidizing character of NaNO2 which would interfere
with the Pd0 species assumed to insert into the C-N2+ bond
via an oxidative addition as a first step of the accepted cata-
lytic cycle of the carbonylation reaction.5 For this reason the
excess of nitrite was destroyed after the diazotization with
sulfamic acid; acetonitrile and PdCl2 were then added to the
reaction solution, and the reaction was started by pressuring
with CO. To our pleasant surprise, after a reaction time of 6
h we could obtain assay yields of SMBA up to 88%.
kawa et al., the replacement of NaOAc by water, the water
concentration, the use of PdCl2 as catalyst, and working at
60 °C had the most beneficial effect on yield and reaction
rate. The carbonylation route has an atom efficiency of 55%,
and to our knowledge it is the first technically feasible
catalytic carbonylation of an arene diazonium compound.
We have indications that the procedure is quite general for
the synthesis of 2-sulfo benzoic acids.
Acknowledgment
We thank M. Mu¨ller and J. Bru¨nisholz for valuable
discussions and W. Schmidt for careful experimental work.
Received for review December 23, 2002.
OP025621M
Summary and Conclusions
The optimal process conditions are summarized in Table
1. Compared to the original procedure described by Kiku-
Vol. 7, No. 3, 2003 / Organic Process Research & Development
•
431