The continuous self aldol condensation of propionaldehyde in supercritical carbon dioxide: A highly selective catalytic route to 2-methylpentenal

Article abstract of DOI:10.1039/b819687g

The aldol reactions of propionaldehyde and butyraldehyde have been explored in supercritical CO₂, scCO₂, using an automated continuous flow reactor. The reaction was found to proceed over a variety of heterogeneous acidic and basic catalysts and with increased selectivity compared to using neat reactants.

Full text of DOI:10.1039/b819687g

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/greenchem | Green Chemistry
The continuous self aldol condensation of propionaldehyde in supercritical
carbon dioxide: a highly selective catalytic route to 2-methylpentenal†
James G. Stevens, Richard A. Bourne and Martyn Poliakoff*
Received 5th November 2008, Accepted 18th December 2008
First published as an Advance Article on the web 23rd January 2009
DOI: 10.1039/b819687g
The aldol reactions of propionaldehyde and butyraldehyde have been explored in supercritical
CO2, scCO2, using an automated continuous flow reactor. The reaction was found to proceed over
a variety of heterogeneous acidic and basic catalysts and with increased selectivity compared to
using neat reactants.
of the aldol intermediate, 4) only reached a maximum of 70%.
However, they were unable to produce 5 directly from 2 with
Introduction
The aldol reaction is one of the most widely used reactions
high selectivity.
for C–C bond formation,1,2 yet it has several drawbacks from
The work described in this paper aims to build on the work
an environmental standpoint. Some of these issues have been
carried out by Baiker and co-workers by studying the aldol
reaction further in scCO2, and by exploring whether the use
highlighted by Mestres3 and Kelly et al.4 Industrially, the aldol
synthesis has to overcome poor selectivity3 and the use of
of hydrogen can be avoided and, hence, whether unsaturated
stoichiometric quantities of aqueous base with neutralisation
aldols can be produced with high selectivity. Hydrogen acts as
and separation of the aldol product after completion of the
a driving force in Scheme 1 because 5 does not participate in
reaction. This neutralisation generates large volumes of aqueous
the acid catalysed equilibrium steps, thus forcing the reaction
inorganic waste which require disposal.4
The continuous synthesis of 2-ethylhexanal (5), a commer-
cially important aldol product, from crotonaldehyde (1) has
been elegantly demonstrated by Baiker and co-workers,5,6 who
performed the reaction continuously in scCO2 over a palladium
doped acidic resin (Amberlystꢀ 15) Scheme 1. Their process
combined the aldol condensation and the hydrogenation of the
unsaturated aldol product into a single process, hence driving
the reaction equilibrium. They achieved >99% hydrogenation
of 1 to 2 but the further conversion to 5 (via the hydrogenation
towards the saturated aldehyde product.
All our reactions were carried out on an automated contin-
uous reactor7 which is described briefly in the Experimental
section at the end of the paper. The reactor has on-line GC
analysis. This allows the operator to program a series of reaction
conditions, for example ramping of the temperature, while
analytical data are collected automatically under computer
control.
R
Proof of concept
The self aldol reaction of butyraldehyde (2) was explored over
R
Amberlystꢀ 15 in continuous flow without the addition of
hydrogen (Scheme 2) utilising scCO2 as the reaction medium.
This is the first stage of the reaction that Baiker and co-workers
reported to give only 70% of the saturated aldehyde 5. Fig. 1
shows the conversion of 2 and selectivity to 4 as a function of
reactor temperature at 8 MPa pressure with fixed flow rates of
both CO2 and 2. It can be seen that the conversion increases
over the whole temperature range studied. A marked increase
◦
Scheme
1
The hydrogenation5,6 of crotonaldehyde (1) to yield
in the selectivity is also observed from 50 to 70 C, but above
butyraldehyde (2), with subsequent self aldol condensation to form
2-ethylhexenal (4) and hydrogenation of the unsaturated aldol product
to form 2-ethylhexanal (5).
this temperature the selectivity for 4 remains constant and high,
>90%.
School of Chemistry, The University of Nottingham, Nottingham, UK
NG7 2RD. E-mail: Martyn.Poliakoff@nottingham.ac.uk; Fax: +44
(0)115 951 3508; Tel: +44 (0)115 951 3520
† Electronic supplementary information (ESI) available: Phase be-
haviour of propionaldehyde and 2-methylpentenal in supercritical
carbon dioxide. See DOI: 10.1039/b819687g
Scheme 2 Self aldol condensation of 2 showing both the aldol
intermediate (3), and the unsaturated dehydration product (4).
This journal is
The Royal Society of Chemistry 2009
Green Chem., 2009, 11, 409–416 | 409
©

View Article Online
Table 1 Maximum yield of 8 achieved during temperature ramp over
various catalysts from the self condensation of 6 at flow rates of 0.1 mL
min-1 6, 1.0 mL min-1 CO2 at 10 MPa (base catalysts in italics)
Catalyst
% Yield 8a
% Sel. 8
Temp/◦C
g-Alumina
78
76
10
49
39
48
63
70
89
88
83
77
95
84
81
83
210
135
180b
125
240
125
155
285
AmberlystꢀR 15
AmberlystꢀR A26
DeloxanꢀR ASP
Magnesium oxide
NafionꢀR SAC13
PuroliteꢀR CT175
Supported NaOH
a Determined by GC. b Catalyst undergoing rapid degradation.
Fig. 1 The formation of the unsaturated butyraldehyde aldol product
4 in scCO2 showing the temperature dependence of the conversion (᭜)
and selectivity (᭛) of the self aldol condensation of 2 at flow rates of
0.1 mL min-1 of 2 and 1.0 mL min-1 of CO2 at 8 MPa using a 10 mL
reactor filled with AmberlystꢀR 15.
Matsui et al.17,18 have explored the effect of a pressure of CO2,
in a batch reaction, on the selectivity of the aldol condensation
of 6 whilst using a MgO catalyst. At 12 MPa, under supercritical
conditions, 94% selectivity towards 8 was achieved, whereas at
the subcritical pressure of 5 MPa, 85% selectivity towards the
aldol product was observed.
This result indicates that it is indeed possible to conduct the
aldol reaction effectively in scCO2 using a heterogeneous acid
catalyst to produce the unsaturated aldol product in high yields.
Our aim was to use the continuous flow scCO2 apparatus7 to
screen rapidly a variety of both acidic and basic heterogeneous
catalysts, see Table 1, for their suitability for the selective self con-
densation of propionaldehyde across a range of temperatures.
Once a suitable catalyst has been found it would then be tested
under a wide variety of conditions to optimise the performance
of the reaction further. The high solubility of gaseous hydrogen
Reaction of propionaldehyde
Following this successful proof of concept experiment with
butyraldehyde, we have studied the self aldol condensation of
propionaldehyde (6) to yield 2-methylpentenal (8) (Scheme 3).
19
and organic compounds in scCO2 can be used in combination
with the aldol reaction to allow hydrogenation of the unsaturated
product in situ if desired.5,6
Catalyst screening
Scheme 3 The self aldol condensation of propionaldehyde (6) showing
both the aldol intermediate, 3-hydroxy-2-methylpentanal (7), and the
unsaturated dehydration product 2-methylpentenal (8).
For each catalyst system the reaction was performed using
identical reaction conditions at a pressure of 10 MPa. Further
details of the catalysts used in this study are outlined later in
Table 4 (Experimental section). Fig. 2 shows the conversion and
selectivity towards 8 for each catalyst across the temperature
range tested. The maximum yields are summarised in Table 1.
However, from the point of view of environmental sustainabil-
ity, selectivity towards the desired product, Table 2, is more
important than conversion, because of the ease of separation
of unreacted 6 from the product 8 by simple distillation due to
the large difference in boiling point of the two compounds; this
enables unreacted 6 to be easily repassed through the reaction
system.
Aims and strategy
Compound 8 is a commercially important compound that is
used widely in the fragrance, flavour,8 and cosmetic industries9
as well as being an important intermediate in the synthesis of
pharmaceuticals10,11 and plasticisers.12 Compound 8 is manufac-
tured industrially by the aldol condensation of 6 in the presence
of stoichiometric amounts of an aqueous base such as NaOH
or KOH. Under optimum conditions, 99% conversion of 6 is
achieved with 86% selectivity to 8. Additionally, it has been
demonstrated that the inclusion of an inert solvent can reduce
the formation of any unwanted by-products.12
Mehnert et al. have studied the base catalysed aldol condensa-
tion of 6 in ionic liquids. They reported 100% conversion at 80 ◦C
in 3 h with 83% selectivity for 8 using NaOH in [BMIM][PF6].13,14
The solvent-free condensation of 6 has also been explored using
a batch process at 100 ◦C using solid base catalysts by Sharma
et al.15 Their optimum results were achieved using an activated
Mg/Al hydrotalcite catalyst, giving 97% conversion of the 6 with
99% selectivity to 8. The aldol condensation of 6 has also been
carried out by Scheidt16 in the gas phase over lithium phosphate
at 270 ◦C. This gave 95% selectivity to 8 at 32% conversion of 6.
At temperatures <70 ◦C, with Brønsted acid catalysts such
R
R
as Amberlystꢀ 15 or Puroliteꢀ CT-175 (Fig. 2b and 2e), the
by-products consisted of the aldol intermediate, 7, and a variety
of other compounds including C9 aldehydes. These arise from
further aldol reactions, and Michael additions to the
a,b-unsaturated product 8. A typical product distribution can
be seen in the gas chromatogram, shown in Fig. 3.
A similar profile of conversion and selectivity was obtained
with the basic catalyst, NaOH/silica, although higher
temperatures were required to achieve this reactivity (Fig. 2h).
Magnesium oxide (Fig. 2f) also gave excellent selectivity
albeit with a poor conversion. The experiments suggested that
the MgO underwent rapid deactivation with a decrease in the
410 | Green Chem., 2009, 11, 409–416
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
Fig. 2 Summary of catalyst performance over a range of temperatures, showing rare examples of base catalysis in scCO2. The plots show the
temperature dependence of the conversion (ꢀ) and selectivity (ꢁ) of the self aldol condensation of 6 to 8 in the presence of a variety of catalysts,
each packed into a 10 mL reactor tube. The reactions were carried out at flow rates of 0.1 mL min-1 of 6 and 1.0 mL min-1 of CO2 at 10 MPa. The
catalysts are: Amphoteric: a, g-alumina; Brønsted acids: b, AmberlystꢀR 15; c, DeloxanꢀR ASP IV/6–2; d, NafionꢀR SAC-13; e, PuroliteꢀR CT-175; Bases:
f, magnesium oxide; g, AmberlystꢀR A26; h, supported NaOH. A moving average line has been drawn between data points to aid visualisation.
R
conversion as the temperature was increased over time. This
rapid deactivation may be due to the removal of catalytically
active hydroxide groups from the surface of MgO. The freshly
exposed MgO can then form inactive carbonates upon reaction
with CO2.20 Amberlystꢀ A26 had poor activity (Fig. 2g)
with only minor conversion over the temperature range studied.
This is probably due to the formation of carbamate species by
the reaction of CO2 with the amine groups of the catalyst.
This journal is
The Royal Society of Chemistry 2009
Green Chem., 2009, 11, 409–416 | 411
©

View Article Online
Table 2 Maximum selectivity achieved during temperature ramp over
various catalysts for the self condensation of 6 at flow rates of 0.1 mL
min-1 6, 1.0 mL min-1 CO2 at 10 MPa (base catalysts in italics)
Catalyst
% Conv. 6a
% Sel. 8
Temp/◦C
g-Alumina
46
43
7
50
22
47
37
41
96
99
97
89
>99
87
>99
97
158
95
AmberlystꢀR 15
AmberlystꢀR A26
DeloxanꢀR ASP
Magnesium oxide
NafionꢀR SAC13
PuroliteꢀR CT175
Supported NaOH
160b
146
184
100
116
174
a Determined by GC. b Catalyst undergoing rapid degradation.
Fig. 4 Demonstration of how the addition of CO2 reduces the amount
of the intermediate 7 in the product stream at lower temperatures. The
plot shows the temperature dependence of the selectivity for 7 during
the self aldol condensation of 6 over AmberlystꢀR 15 at 10 MPa, flow
rates of: ꢁ, 0.1 mL min-1 of 6, no CO2; ᭺, 0.5 mL min-1 of 6, no CO2;
ꢀ, 0.1 mL min-1 of 6 and 1.0 mL min-1 of CO2.
Fig. 3 GC chromatogram (see Experimental for details) taken from
Fig. 2b. This shows the ran◦ge of products from the self condensation of
6 over AmberlystꢀR 15 at 50 C. The main peaks are due to 6, 7 and 8; the
inset figure shows the region from 11.0–12.2 mins with expanded scales.
(Other minor peaks remain unidentified.)
R
Lower selectivity was achieved with Nafionꢀ SAC13 than with
R
R
Amberlystꢀ 15 possibly because the higher acidity of Nafionꢀ
accelerated the rate of side reactions. Proportionately more 7
was formed at <60 ◦C with up to 10% selectivity. Even at 140 ◦C
Fig. 5 Plot demonstrating the increase in selectivity for 8 in the presence
of CO2 during the self aldol condensation of 6 over AmberlystꢀR 15.
Carried out at 10 MPa, flow rates of: ꢁ, 0.1 mL min-1 of 6, no CO2; ᭺,
0.5 mL min-1 of 6, no CO2; ꢀ, 0.1 mL min-1 of 6 and 1.0 mL min-1 of
CO2.
R
there was up to 5% selectivity towards 7 with Nafionꢀ SAC13.
R
R
Overall, g-alumina, Amberlystꢀ 15, and Puroliteꢀ CT-175
appeared to have the greatest potential for further optimisation
whilst maintaining a high selectivity for 8.
As the temperature was increased, the conversion in scCO2
increased markedly and scCO2 has little effect on conversion
at ≥130 ◦C.
R
Optimisation of the reaction with Amberlystꢀ 15
It is postulated that the reaction is driven to completion by
the insolubility of the polar aldol intermediate 7 in scCO2.
Intermediate 7 will most likely exist in a liquid-like phase with a
higher affinity for the catalyst surface. It can then react forming
8 which is then removed from the catalyst due to its higher
solubilty in scCO2. See ESI† for details of phase behaviour of 6
and 8 in scCO2 – unfortunately it was not possible to isolate the
intermediate 7 in sufficient purity to analyse its phase behaviour
with CO2.
Because the selectivity is higher for the reaction in scCO2, the
overall yield of 8 is higher with CO2, than without, between 60
and 140 ◦C (see Fig. 6).
Entries 1–4 in Table 3 show the effect of altering the residence
time by varying the flow rates of CO2 and 6 whilst maintaining
a constant 6 mol% organic ratio at an isothermal reactor
temperature of 90 ◦C. As the residence time was decreased, a
R
Following the catalyst screening, Amberlystꢀ 15 was chosen
for optimisation due to its low operating temperature, high
selectivity and high yield compared to Puroliteꢀ and g-alumina
R
(Table 1). Optimisation of the reaction in scCO2 was carried out
by varying several reactor parameters. Firstly, the reaction was
carried out under the same conditions both with and without
CO2. Three reactions were performed to study the effect of CO2
on the system: 0.1 mL min-1 6 without CO2; 0.5 mL min-1
6
without CO2; 0.1 mL min-1 6 with 1.0 mL min-1 CO2, these results
are shown in Fig. 4, Fig. 5 and Fig. 6. The scCO2 reaction system
produced much lower amounts of 7, the aldol intermediate,
to 8.
Similarly, Fig. 5 shows that the presence of scCO2 increases the
selectivity toward the desired product 8. However, the conversion
at low temperatures was significantly higher without CO2 (Fig. 6).
412 | Green Chem., 2009, 11, 409–416
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
Fig. 7 Effect of temperature on the yield of 8 with: , a single reactor;
, two reactors in series (doubling the reactor volume). Carried out at
flow rates of 0.1 mL min-1 of 6 and 1.0 mL min-1 of CO2 at 8 MPa in
the presence of AmberlystꢀR 15. The addition of a second reactor had no
observable effect on the selectivity.
the reaction. Therefore, the benefits of using CO2 on selectivity
can be retained even when using high concentrations of 6.
The effect of pressure was also explored (Table 3, N = 9–12).
No definitive change in selectivity was observed as the pressure
was varied from 6 to 20 MPa.
The activity of the catalyst was monitored over time un-
der constant conditions of 90 ◦C and 8 MPa with no pre-
equilibration of the system. These conditions were chosen to
maximise the conversion with >99% selectivity. There was a
rapid reduction in conversion from 86% to 50% over the first 8
h with a small percentage of side products being observed. The
conversion then stabilised at 43% and was maintained at this
level up to 65 h with no observable side products (by GC) (Fig. 8).
A loss of activity was also observed by Baiker and co-workers5
when performing the aldol reaction in scCO2. The stabilisation
of conversion following an initial rapid drop suggests that a
Fig. 6 At lower temperatures the conversion of 6 is higher without
CO2, but the yield of 8 is higher with CO2 due to the significant increase
in selectivity. Temperature dependence of: (a), the conversion of 6 during
its self aldol condensation over AmberlystꢀR 15 and (b), the yield of 8
from the self condensation of 6 over AmberlystꢀR 15 is illustrated. Carried
out at 10 MPa, flow rates of: ꢁ, 0.1 mL min-1 of 6, no CO2; , 0.5 mL
min-1 of 6, no CO2; ꢀ, 0.1 mL min-1 of 6 and 1.0 mL min-1 of CO2.
marginal decrease in conversion was observed (51% → 44%, as
total flow rate was increased from 0.825 mL min-1 → 3.85 mL
min-1). To test further the effect of residence time a second
reactor was added downstream of the first. Fig. 7 shows that an
additional catalyst bed gives an extra 20% in conversion from 60
to 105 ◦C, accompanied by an increase in selectivity to >99%.
Entries 5–8 in Table 3 show that the molar ratio of CO2 : 6
has only a minimal effect on the conversion and selectivity of
R
partial desulfonation of Amberlystꢀ 15 has occurred, rather
than a coking process. We have previously observed a similar
desulfonation with the Friedel–Crafts alkylation of anisole.21
The in situ hydrogenation of the unsaturated product 8 is
also of interest, see Scheme 4. The hydrogenation was carried
R
out by replacing the Amberlystꢀ 15 with a 50 wt% mixture of
Table 3 Study on the effect of reaction parameters on the conversion and selectivity of 6 to 8
N
T/◦C
P/MPa
CO2 flow/mL min-1
6 flow/mL min-1
Mol% of 6
Conv. (%)
Sel. (%)
1a
90
90
90
10
10
10
10
10
10
10
10
6
0.75
1.00
2.25
3.50
0.97
1.04
1.10
1.15
1.00
1.00
1.00
1.00
0.075
0.100
0.225
0.350
1.027
0.645
0.335
0.077
0.100
0.100
0.100
0.100
6
6
6
51
48
46
44
56
57
50
59
66
71
62
68
95
92
96
95
90
88
88
87
64
65
68
76
2a
3a
4a
90
6
5b
120
120
120
120
120
120
120
120
40
28
16
4
6
6
6b
7b
8b
9c
10c
11c
12c
10
15
20
6
6
The parameters were varied as follows: a residence time; b concentration of substrate 6; c system pressure.
This journal is
The Royal Society of Chemistry 2009
Green Chem., 2009, 11, 409–416 | 413
©

View Article Online
hydrogenation of 8 was achieved. The conversion was slightly
R
elevated higher than that achieved with Amberlystꢀ 15 alone
(Fig. 7) suggesting that the hydrogenation of 8 to 10 may drive
the dehydration of 7 to some extent.
Conclusions
The self condensation of butyraldehyde and propionaldehyde
in scCO2 has been successfully conducted, yielding the
a,b-unsaturated aldol products with high selectivity. A series
of catalysts has been evaluated, using an automated continuous
flow reactor. Surprisingly, two basic catalysts were effective in
R
scCO2, but overall the experiments showed that Amberlystꢀ 15
was the most effective of the catalysts screened. The reaction
R
parameters with Amberlystꢀ 15 parameters were then studied
Fig. 8 Effect of catalyst activity over an extended time period on the
conversion (ꢀ) and selectivity (ꢁ) of the self aldol condensation of 6 to
8 in the presence of AmberlystꢀR 15. The catalyst was loaded into two
reactors in series and the reaction was carried out isothermally at 90 ◦C,
8 MPa with flow rates of 0.1 mL min-1 of 6 and 1.0 mL min-1 of CO2.
in more detail. The formation of the unsaturated aldol product,
8, was more selective in scCO2 than in neat propionaldehyde,
over the whole temperature range studied, an effect attributed
to the postulated insolubility of the polar aldol intermediate,
7, in scCO2. The additional selectivity which is observed in the
scCO2 system delivers higher yields of 8, whilst any unreacted
starting material could be easily recycled. Finally, it was possible
to perform an in situ hydrogenation of 8 by using a mixed
R
catalyst bed consisting of a 50 wt% mixture of Amberlystꢀ
15 and 2% Pd on silica/alumina (Johnson Matthey Type 31)
producing 2-methylvaleraldehyde in high yield. Thus, overall,
our experiments demonstrate that aldol condensations in scCO2
have even wider potential than reported by previous workers in
this field.
Scheme 4 Self aldol condensation of propionaldehyde (6) with possible
hydrogenation products: propan-1-ol (9), 2-methylvaleraldehyde (10)
and 2-methyl-1-pentanol (11).
Experimental
Safety hazard
R
Amberlystꢀ 15 and 2% Pd on silica/alumina (Johnson Matthey
Type 31) and dosing hydrogen into the system at 0.55 molar
equivalents (based on 6). Fig. 9 shows the conversion of 6 and
selectivity to 10 as a function of temperature. No hydrogenation
of the carbonyl group to 11 was observed even when complete
CAUTION! The experiments described in this paper involve the
use of relatively high pressures and require equipment with the
appropriate pressure rating.
Catalysts and reagents
Commercially available catalysts were employed: g-alumina (SI
R
R
Group), Amberlystꢀ 15 (Lancaster Synthesis), Amberlystꢀ A26
R
(Alfa Aesar), Deloxanꢀ ASP (Degussa Hu¨ls), magnesium oxide
R
R
(Acros), Nafionꢀ SAC 13 (Sigma Aldrich), Puroliteꢀ CT-175
(Purolite International Ltd). NaOH was used as NaOH on a
silica support, see Table 4. Commercially available propionalde-
hyde (Acros, 97%) and butyraldehyde (Acros, 97%) were used
without further purification.
Automated continuous flow apparatus and reaction procedure
Fig. 10 shows a schematic view of the experimental apparatus. A
programmable CO2 pump, CP (Jasco PU-1580-CO2) and HPLC
pump, OP (Jasco PU-980) were connected to a 1/4 inch tee-piece,
packed with glass beads, PH. The tee-piece acts as both mixer
and pre-heater, heated by cartridge heaters within an aluminium
heating block. Hydrogen was dosed into the system via a 6-port
Rheodyne, HD, and was premixed with the CO2 in a static mixer,
M. The reactor, R, consisted of a 10 mL 316 SS tube (100 mm ¥
12 mm OD), packed with catalyst and heated by cartridge heaters
Fig. 9 Plot of the temperature dependence of the conversion (ꢀ) and
selectivity (ꢁ) of the self aldol condensation and in situ hydrogenation
of 6 to 10 in the presence of AmberlystꢀR 15 mixed with 2% Pd on
silica/alumina (JM Type 31). The reactions were carried out at flow
rates of 0.1 mL min-1 of 6 and 1.0 mL min-1 of CO2 at 8 MPa with 0.55
equivalents of hydrogen. No hydrogenation of the carbonyl group to 9
or 11 was observed.
414 | Green Chem., 2009, 11, 409–416
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
Table 4 Catalysts employed in this study
Catalyst
Support
Active group
Catalyst loading
g-Alumina
N/A
N/Aa
–SO3H
N/A
AmberlystꢀR 15
Macroreticular polystyrene cross-linked with divinylbenzene
Macroreticular polystyrene cross-linked with divinylbenzene
Polysiloxane
4.7 eq/kg
0.80 eq/L
Unknown
N/A
AmberlystꢀR A26
DeloxanꢀR ASP IV/6–2
Magnesium oxide
NafionꢀR SAC13
–N(Me)3OH
–SO3H
N/A
N/Aa
Copolymer resin (of tetrafluoroethane and
perfluoro-2-(flurosulfonylethoxy)propyl vinyl ether)
Macroporous polystyrene cross-linked with divinylbenzene
Silica
–CF2CF2SO3H
0.13 eq/kg
PuroliteꢀR CT-175
Supported NaOH
–SO3H
NaOH
4.9 eq/kg
3.5 eq/kg
a The metal oxides, Al2O3 and MgO20 possess a variety of different active sites and modes of catalysis.
22
pressure was logged at several points throughout the system
R
on a computer, via a PicoLogꢀ data recorder.
In a typical experiment, the reactor was filled with catalyst and
sealed into the apparatus. After the equipment had stabilised
at the required pressure and flow rate of CO2, the reactor
was heated to the reaction temperature and the substrate
was pumped into the system. Experimental parameters were
programmed into the pumps, back pressure regulator and GC.
Unless otherwise stated, the standard reaction conditions were:
flow of CO2 1.0 mL min-1 at -10 ◦C and 5.8 MPa, flow of
the substrate 0.1 mL min-1, 10 MPa operating pressure. The
GC was fitted with an RTX-5 column (30 m, ID 0.32 mm,
film thickness, 0.25◦mm), held at 40 ◦C for 1 minute, ramped at
◦
17 C/min to 280 C and then held for 2 min. Quantification
was performed by integration of the peak areas; response factors
and conversions were calculated by the internal normalisation
method. The products were identified by comparing the GC
retention time with known standards and by GC–MS.
Acknowledgements
We thank the EPSRC and Thomas Swan & Co. Ltd. for support.
We also thank Dr P. N. Gooden, Miss K. A. Scovell, Dr P.
Licence and Dr S. K. Ross for scientific discussion and Mr M.
Dellar, Mr M. Guyler, Mr R. Wilson and Mr P. Fields for their
technical support at the University of Nottingham.
References
Fig. 10 Schematic of the automated supercritical fluid continuous flow
apparatus. The components are labelled as follows: CP, chilled CO2
pump; HD, H2 dosing unit; M, mixer; OP, organic pump; PH, pre-
heater and mixer; R, reactor; HP-SL, high pressure sample loop; BPR,
back pressure regulator; GLC, gas liquid chromatograph; W, organic
waste.
1 J. Clayden, N. Greeves, S. Warren and P. Wothers, Organic Chemistry,
Oxford University Press, Oxford, UK, 2001.
2 M. M. Green and H. A. Wittcoff, Organic Chemistry Principles and
Industrial Practice, Wiley-VCH, Weinheim, 2003.
3 R. Mestres, Green Chem., 2004, 6, 583–603.
4 G. J. Kelly, F. King and M. Kett, Green Chem., 2002, 4, 392–
399.
5 T. Seki, J. D. Grunwaldt, N. van Vegten and A. Baiker, Adv. Synth.
Catal., 2008, 350, 691–705.
6 T. Seki, J. D. Grunwaldt and A. Baiker, Chem. Commun., 2007,
3562–3564.
7 B. Walsh, J. R. Hyde, P. Licence and M. Poliakoff, Green Chem.,
2005, 7, 456–463.
8 T. Aida, T. Nagasawa and Y. Yamazaki, WO Pat., 065851,
2008.
9 S. Tanaka, K. Fukuda and T. Asada, WO Pat., 029033, 2004.
10 S. Aratake, T. Itoh, T. Okano, T. Usui, M. Shoji and Y. Hayashi,
Chem. Commun., 2007, 2524–2526.
11 R. H. Fan, D. M. Pu, F. Q. Wen, Y. Ye and X. L. Wang, J. Org.
Chem., 2008, 73, 3623–3625.
in an aluminium heating block, controlled via a programmable
heating controller (Eurotherm 2416). The mixture of products
and gases passed through a HPLC 4-port sample loop, HP-SL,
which injects the sample into the carrier gas flow of a GLC
(Shimadzu GC-17A), where product separation and analysis
was carried out.
The system pressure was maintained by a programmable
Back Pressure Regulator, BPR (Jasco BP-1580-81). Products
were collected in the waste bottle, W. The temperature and
This journal is
The Royal Society of Chemistry 2009
Green Chem., 2009, 11, 409–416 | 415
©

View Article Online
12 A. D. Godwin, R. H. Schlosberg, F. Hershkowitz, M. G. Matturro,
G. Kiss, K. C. Nadler, P. L. Buess, R. C. Miller, P. W. Allen, H. W.
Deckman, R. Caers, E. J. Mozeleski and R. P. Reynolds, US Pat.,
6,307,093, 2001.
17 K. Matsui, H. Kawanami, Y. Ikushima and H. Hayashi, Chem.
Commun., 2003, 2502–2503.
18 K. Matsui, H. Kawanami and H. Hayashi, in Carbon Dioxide Uti-
lization for Global Sustainability, Elsevier Science Bv, Amsterdam,
2004, pp. 363–368.
19 M. G. Hitzler, F. R. Smail, S. K. Ross and M. Poliakoff, Org. Process
Res. Dev., 1998, 2, 137–146.
20 H. Hattori, Chem. Rev., 1995, 95, 537–558.
13 C. P. Mehnert, N. C. Dispenziere and R. H. Schlosberg, US Pat.,
6,552,232, 2003.
14 C. P. Mehnert, N. C. Dispenziere and R. A. Cook, Chem. Commun.,
2002, 1610–1611.
15 S. K. Sharma, P. A. Parikh and R. V. Jasra, J. Mol. Catal. A: Chem.,
2007, 278, 135–144.
16 F. M. Scheidt, J. Catal., 1964, 3, 372–378.
21 R. Amandi, P. Licence, S. K. Ross, O. Aaltonen and M. Poliakoff,
Org. Process Res. Dev., 2005, 9, 451–456.
22 G. W. Kabalka and R. M. Pagni, Tetrahedron, 1997, 53, 7999–8065.
416 | Green Chem., 2009, 11, 409–416
This journal is
The Royal Society of Chemistry 2009
©