Angewandte
Chemie
ing the order NTf2À < OTfÀ < HCO2 that parallel their
HCO2H from the amine-functionalized IL 5 proceeded more
À
increasing basicities, dielectric constants, and Kamlet–Taft
parameters b (Table 2, entries 1, 4, and 5).[10] Using the amine-
free IL 10 with a formate anion also resulted in high TONs
and TOFs (Table 2, entry 6). Notably, the TOF of 314 hÀ1
obtained in the 1:1 molar mixture of ILs 5 and 7 is one order
of magnitude higher than the highest activity previously
reported in ILs without addition of nonionic amines. The high
formic acid-to-amine ratios of up to 1:2 indicate that high
HCO2H loadings can be achieved in these media, a beneficial
property for potential extraction processes.
Immobilizing the stabilizing base on the surface of a solid
support was also shown to be effective for the first time. A
polystyrene resin functionalized with diethylamine groups
(QuadraPure-DMA, Johnson Matthey) suspended in non-
functionalized EMIM NTf2 (11) gave high initial TOFs above
200 hÀ1, together with a high formic acid-to-amine ratio of
approximately 1:2.
effectively, and after 95 h an extraction efficiency of 84% was
obtained, corresponding to initial extraction rates of at least
3.5 mghÀ1. Up to 94% mass recovery was reached after
longer times. Pure HCO2H without any cross-contamination
was extracted with scCO2 from all of the bases according to
1H NMR spectroscopic analysis of the extracts.
Based on these data, the IL-adjusted catalyst system 4 was
combined with selected nonvolatile bases as the stationary
phase for the continuous-flow CO2 hydrogenation process
with integrated product extraction. A high-pressure steel
autoclave (V= 10 mL) equipped with a window was charged
with a solution of catalyst 4 in the amine-functionalized IL 5
(V= 1.0 mL, c(cat.) = 2 mmolmlÀ1), heated to 508C and
pressurized to 200 bar with CO2. The continuous reaction
was started by passing the supercritical CO2/H2 flow (V-
(CO2) = 200 mLN minÀ1, V(H2) = 20 mLN minÀ1) through the
reactor under vigorous stirring. Extracts were collected by
passing the exiting stream through cooling traps filled with
water, which were periodically changed and analyzed by
1H NMR. The error in the determination of the HCO2H
concentration was determined to be in the range 2–10% by
calibration measurements (see the Supporting Information).
Already in the first sample after 5 h, a HCO2H concentration
corresponding to a TON of 28 was detected. The cumulative
TONs over time are depicted in Figure 2. The TON of
HCO2H produced after 211 h was 358. Pronounced catalyst
The extraction of free HCO2H from these nonvolatile
bases with scCO2 was assessed in a custom-built continuous-
flow setup with precise control over pressure, temperature,
and flow.[11] The solubility of HCO2H in scCO2 was deter-
mined to be at least 49 gLÀ1 at 508C and 100 bar. Figure 1
shows selected examples for the extraction efficiency (mass of
Figure 1. Extraction efficiency (mass of extracted HCO2H/mass of
initially added HCO2H) during extractions of HCO2H with scCO2 from
immobilized bases. Lines are only provided as a guide for the eye.
Conditions: T=508C, p=200 bar, V(IL)=1.0 mL, V-
Figure 2. Cumulative turnover numbers for catalyst 4 in IL 5 (ꢂ) and in
(CO2)=200 mLN minÀ1, V(H2)=20 mLN minÀ1, V(autoclave)=10 mL; 5
*
QuadraPure-DMA/IL (11, ) with respect to time. Lines are only
!
(ꢂ): m(HCO2H)=159 mg, acid/amine=1.2; 10 ( ): m-
provided as a guide for the eye. Conditions: T=508C, p=200 bar,
*
(HCO2H)=375 mg, acid/formate=1.3, after 26 h T=708C; 11 ( ):
V(IL)=1.0 mL, V(CO2)=200 mLN minÀ1, V(H2)=20 mLN minÀ1, V(auto-
m(resin)=0.5 g, m(HCO2H)=130 mg, acid/amine=1.1 calculated for
clave)=10 mL, c(cat.)=2 mmolmLÀ1
.
a theoretical amine capacity of the resin of 5 mmolgÀ1
.
deactivation was observed after 70 h in this experiment, which
could also be visually followed by color changes from bright
yellow to orange. The H NMR spectrum of IL 5 recovered
extracted HCO2H/mass of initially added HCO2H) over time
at initial formic acid-to-base ratios between 1.1–1.3 and
comparable scCO2/H2 flow values (see the Supporting
Information for details). Extraction from the formate IL 10
was sluggish, reaching an extraction efficiency of only 14%
after 74 h at an average rate of 0.9 mghÀ1. Increasing the
temperature from 508C to 708C did not significantly improve
the extraction rate. HCO2H extraction from the IL-suspended
resin QuadraPure-DMA 11 gave an extraction efficiency of
24% after 91 h at an average rate of 0.4 mghÀ1. Recovery of
1
from the reactor after 211 h on-stream was virtually
unchanged (see the Supporting Information), indicating
a reasonable stability of the IL under sustained reaction
conditions.
When the reactor was charged with a solution of catalyst 4
in EMIM NTf2 (V= 1.0 mL, c(cat.) = 2 mmolmLÀ1) with
QuadraPure-DMA 11 (0.5 g) suspended therein, the contin-
uous reaction conducted under the same conditions showed
Angew. Chem. Int. Ed. 2012, 51, 1 – 5
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3
These are not the final page numbers!