Communication
RSC Advances
ꢁ0.5 seconds. It is worth noting, that the residence time can be
accurately controlled using a water quench before depressur-
ization and collection of the reaction mixture. As small amounts
of precipitates were formed either during the carboxylation, or
aer adding H2O a 1 mL coil was installed directly aer the
quench. During the time in which the reaction mixture passes
Fig. 1 Carboxylation of alkynes and heterocycles using organolithium this residence unit all solids dissolved, allowing the reaction
bases and CO2.
mixture to smoothly pass the pressure-regulating unit.
With the optimized conditions in hand, we next evaluated
our setup using several terminal alkynes (Table 1). It has to be
noted that in all examples, small amounts of the substrate were
still present in the reaction mixture as analyzed by HPLC-UV.20
Furthermore, we realized that 1–10% of a byproduct was formed
during the reaction which could be identied as a trimethylsi-
lane derivative apparently from a reaction of the base and the
alkyne moiety. Electron-rich (1a, 1b) and electron poor (1c, 1d,
1f) aromatic alkynes were transformed into their corresponding
carboxylic acids resulting in good yields aer extraction of the
byproducts followed by acidication and crystallization.
Unfortunately, it was not possible to convert the phenol deriv-
ative 1e and 3-ethynylpyridine (1h) due to precipitation of either
the organometallic intermediate or, in case of 1h, the carboxy-
late, resulting in a clogged reactor. However, we could expand
the synthetic scope of our methodology by heterocyclic (1g) and
aliphatic (1i, 1j) substrates without further modications.
The propiolic acids obtained are valuable synthetic inter-
mediates for a range of pharmaceutically or industrially relevant
molecules, including polymers, coumarins, avones, spi-
robenzofuranes, spiroindoles or vinyl suldes.21
In order to broaden the synthetic horizon of the lithiation–
carboxylation ow reactor concept, we decided to test the same
methodology for its suitability to convert heterocycles into their
corresponding carboxylic acids. Initially we decided to use the
lithiation of thiophene (3a) by lithium diisopropylamide (LDA)
with subsequent carboxylation for a feasibility study. In an
initial experiment the reactor clogged immediately aer the
water quench which prompted us to change to a mixture of
water and acetic acid (10 : 1) to overcome this hurdle. Gratify-
ingly, thiophene-2-carboxylic acid (4a) could be isolated by
extraction aer this minor modication in satisfying yields
without any further re-optimization of the continuous method
(Fig. 3). An additional experiment using signicantly higher
amounts of CO2 (1.9 equiv.) resulted in only slightly higher
amounts of the desired target molecule. Methyl substituted
thiophenes (3b, 3c) as well as benzofuran (3f) provided
moderate yields. Further optimization studies using 3b and
either a higher excess of LDA (2 M, 2.2 equiv.) or larger amounts
of carbon dioxide (30 mLN minÀ1, 1.9 equiv.) did not result in a
signicant improvement. In case of 1-phenylpyrazole the
reactor immediately clogged aer mixing with the organome-
tallic base as the intermediate is apparently completely insol-
uble in THF. However, the use of electron poor thiophenes
(4d, 4e) expanded the synthetic potential of the presented
continuous ow methodology.
with lithium bis(trimethylsilyl)amide (LiHMDS) and CO2 in a
gas/liquid continuous ow regime. We initially optimized the
parameters required for the carboxylation step by mixing a
solution of LiHMDS and 1a in THF with the gaseous reagent in a
T-mixing unit.17 Aer an intensive evaluation of various reaction
parameters, including gas and liquid ow rates, residence time,
temperature as well as back pressure regulation it was estab-
lished that the carboxylate is formed in sufficient amounts
within ꢁ0.5 seconds at room temperature. The next logical step
was to combine the gas/liquid carbon dioxide xation with an
on-demand continuous ow generation of the reactive inter-
mediate.17 Importantly, this could be easily implemented using
an additional mixer in combination with a 0.1 mL residence
time, resulting in the 4-feed approach shown in Fig. 2.18
In the nal setup, a liquid stream of a commercially available
1 M solution of LiHMDS in THF was mixed with the alkyne
dissolved in THF via a T-mixer. The ow rates and substrate
concentration were set in order to obtain a slight excess of the
lithium amide (1.1 equiv.). Aer a residence time of ꢁ3 seconds
the lithium alkynate was mixed with the gaseous reagent in a
second T-mixer resulting in a completely homogenous gas/
liquid mixture at a back pressure of 10 bar. It has to be pointed
out that the CO2 ow rate of 19 mLN minÀ1 corresponds to a
comparably low excess (1.2 equiv.) and therefore constitutes an
almost quantitative consumption of the greenhouse gas.
However, we additionally realized that the performance of the
CO2 mass ow controller had to be stabilized by preheating the
gas to 65 ꢀC in a stainless steel coil.19 High conversions of
the organolithium intermediate to the corresponding carboxy-
late could be achieved within a short residence time of only
In summary, we have developed a fast and efficient process
for the continuous lithiation and subsequent carboxylation with
carbon dioxide of terminal alkynes and heterocycles. The
Fig. 2 Continuous flow set up for the synthesis of carboxylic acids
with CO2.
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 13430–13433 | 13431
Pieber, Bartholomaeus
