9432 J. Phys. Chem. A, Vol. 105, No. 41, 2001
Glebov et al.
conclusion. The results of the experiments #1-#3 (Table 1) are
in agreement with the measured solubility of maleic anhydride
in supercritical carbon dioxide (Figure 5).
for this discrepancy. These experiments are described in the
next section.
Analysis of Possible Sources of Errors. The following
possible sources of experimental errors were identified and
assessed:
Additional experiments (experiments #4 and #5 in Table 1)
were designed to verify whether a single phase or two phases
are present under the experimental conditions of the work of
Paulaitis and Alexander.6,7 The experiments were performed at
60 °C, 133 atm, and the loaded mole fraction of maleic
anhydride 2.5 × 10-3 (the set of the experimental conditions
of work7 closest to the experimental conditions of the current
work). In the binary system, maleic anhydride/carbon dioxide
(experiment #4) two phases (gaslike and liquid) were unam-
biguously observed during the whole observation time (2 h).
This infers that the solubility of maleic anhydride in supercritical
carbon dioxide at 60 °C and 133 atm is lower than the mole
fraction of maleic anhydride used in the experiments of Paulaitis
and Alexander (2.5 × 10-3,7). However, this observation is not
sufficient to unambiguously prove that the experiments in refs
6,7 were performed in a two-phase system because the second
reactant (isoprene) might serve as a cosolvent. To check this
possibility, experiment 5 (Table 1) was performed in the ternary
(maleic anhydride/isoprene/carbon dioxide) system. The ex-
perimental conditions were the same as the experimental
conditions of one of the experimental points measured at 60
°C and 133 atm of Paulaitis and Alexander.7 Droplets of a liquid
phase were observed during 2 h of the observation period (in
the experiments of Paulaitis and Alexander the first sample was
withdrawn 1 h after the loading of isoprene7). In conclusion,
the current experiments with the view tube unambiguously
demonstrate that at least at some experimental conditions used
in the previous works6,7 both the binary (maleic anhydride/
carbon dioxide) and ternary (maleic anhydride/isoprene/carbon
dioxide) mixtures are in the two-phase region.
1. Carbon Dioxide Composition in the Reactor. Supercritical
carbon dioxide obtained from the cylinders pressurized with
helium contains small amount of helium, which is dissolved in
the liquid carbon dioxide in the cylinders. Carbon dioxide
withdrawn from the cylinders was analyzed using mass-
spectrometry. The mole fraction of helium was determined to
be 2.5%. This small helium impurity is expected to have only
minor effects both on the media properties and the rates of
chemical reactions.
2. Impurities in Isoprene. No impurities in isoprene were
found by the GC analysis. Vacuum distillation of isoprene has
no effect on the reaction rate. No dependence of the apparent
rate constant on the concentration of isoprene was found within
the experimental accuracy (10%).
3. Influence of Water. It could be speculated that traces of
water (if exist) in the reaction mixture might have an effect on
the reaction rate via the mechanism of acid catalysis. Addition
of 0.01 mole fraction of water to the reaction mixture did not
affect the reaction rate within the experimental accuracy 10%.
4. Effect of Maleic Acid or other Impurities in Maleic
Anhydride. Maleic acid was the only impurity found in maleic
anhydride. The content of maleic acid in the commercial reagent
was about 3%. It was purified to better than 0.4% using the
procedure described earlier.
Maleic acid might affect the measured rate of reaction 1 in
two ways. First, it could affect the phase coexistence boundary.
Second, it could have a catalytic effect on reaction 1. Catalysis
of Diels-Alder reactions by Lewis acids was observed in the
liquid phase39 as well as in supercritical carbon dioxide.8,10
No effect of the addition of maleic acid (up to 3 wt %) on
the solubility of maleic anhydride in supercritical carbon dioxide
was observed in the current work. Purification of the commercial
reagent (3 wt % maleic acid) to less than 0.4 wt % had no effect
on the measured solubility of maleic anhydride in supercritical
carbon dioxide.
Another important issue is whether the homogeneous phase,
which fills the whole reactor is a true supercritical or a liquid
phase. Recently, Lin and Akkerman26 estimated the critical
conditions of reactive mixtures used in several earlier studies
of Diels-Alder reactions in supercritical carbon dioxide using
the group contribution method.37 According to their estimates,
under the conditions used in these previous studies9,10,14,19,29 the
shift of the critical point due to the presence of the reactants
could be large (e.g., 60 degrees Celsius9,10). On the basis of the
estimates performed Lin and Akkerman26 concluded that a
number of earlier studies were presumably performed either in
a two-phase region9,10 (which was confirmed by the visual
observation of the second phase in refs 14,26) or in a single
liquid-phase region, but not in the supercritical fluid.14,19,29
We used the recommendations of Reid et al.38 to evaluate
the critical parameters of both the binary and the ternary systems
formed by carbon dioxide, maleic anhydride, and isoprene.
When the mole fractions of the solutes are smaller than 0.0007
(the concentration range in which the measured apparent rate
constant of reaction 1 is independent of the concentration of
the reactants), the estimated differences between the critical
parameters (Tc, pc) of the mixtures and of pure CO2 are less
than 4 °C and 1 atm, respectively. Therefore, the single phase
observed in our experiments at 60 °C and 100 atm over this
concentration range is the supercritical phase.
No effect of either maleic anhydride purification (from 3 wt
% to < 0.4 wt %) or maleic acid addition (to 10 wt %) on the
apparent rate constant of reaction 1 was observed at “high”
concentrations (xMA ≈ 0.01), where the apparent rate constant
has maximum. At low concentrations (xMA < 0.001), purification
of maleic anhydride also has no effect on the reaction rate.
However, at low concentrations (xMA < 0.001) addition of large
amounts of maleic acid (up to xacid ≈ 0.1 xMA) lead to the
acceleration of the reaction by a factor of 2. This allows to
estimate an upper limit on the possible effect of maleic acid on
the reaction rate in the current experiments as < 4%.
Comparison of the Current and Previous6,7 Works. On
the basis of the discussion above the following conclusions were
derived. Only the experiments that were performed at suf-
ficiently low concentrations of maleic anhydride (<0.0007)
reflect the true supercritical phase reaction rate constant.
Averaging the rate constants values obtained at low mole
fractions of maleic anhydride (xMA < 0.0007, the five low
concentration points in Figure 4) yielded kx ) (17.0 ( 1.3) hr-1
.
The current kinetic measurements are consistent with both
the solubility measurements and the visual observations of the
phase behavior in this work. The disagreement with the previous
studies6,7 is apparent and significant. Beyond the explanation
given above, a number of hypothetical sources of errors were
initially investigated in an attempt to find another explanation
This is the value for the rate constant of reaction between
isoprene and maleic anhydride in SCF CO2 at 60 °C and 100
atm measured in the current work.
The earlier experiments6,7 have been performed in a two-
phase system. The initial concentrations of maleic anhydride