disubstituted enone, the material may simply form the conju-
gated diene-one product or the industrially significant, cyclic
isophorone. The conjugated nature of the enone product allows
for continued additions, forming higher molecular weight
species that can accumulate rapidly in an all-nanoporous
catalyst, thus restricting access to the active sites within the
nanopores and shutting down reaction.
fragments. This makes sense since the protons that are active for
base extraction are a to the carbonyl or conjugated. The non-
cyclic trimer has 12 of these active protons and the isophorone
has only 3 that are not sterically hindered. Hence the key to
reducing heavies formation and deactivation is to promote ring
formation over oligomerization. Since the heat of formation and
the activation energy to form the non-cyclic products are higher,
a reduction in temperature should favor the isophorone product.
Also, ring formation should have a lower (more negative) DS‡,
therefore, a decrease in temperature should again favor its
formation. Secondly, any variable that reduces the overall
conversion or catalyst basicity should reduce heavies formation.
Thus, less basic metals such as potassium or sodium, lower
acetone concentrations in the feed, and shorter contact times in
the reactor should favor isophorone production.
Caesium, vapor-deposited into the carbon and then oxidized,
is more active than Cs deposited from a salt in aqueous solution.
The Cs hydrous oxide that results from oxidation of the vapor-
deposited element is certainly a different phase; by examining
the X-ray diffraction patterns of the two materials we were able
to confirm this. The samples prepared with the CsOH solution
showed patterns indicative of large hydroxide crystals. In
contrast, the used vapor-deposited Cs/NPC material displayed
an X-ray diffraction pattern with sharp lines, but the lines did
not match any of the known caesium hydroxides or oxides. This
set of results points to a new approach in the synthesis and use
of solid Brønsted bases for continuous-flow organic synthesis.
By preparing a carbon with both nanopores and mesopores, we
can vacuum deposit cesium in a highly dispersed state, convert
it to an active form via oxidation, and provide the necessary
avenues for rapid molecular ingress and egress.
Experiments were carried out in a gas-phase flow reactor.† A
steady argon flow (150 cm3 min21) carried acetone vapor from
a room-temperature gas sparger through the heated catalyst bed.
Unreacted acetone and the heavier products were condensed
and analyzed quantitatively by gas chromatography and mass
spectrometry. The major product was isophorone along with
mesityl oxides and higher-molecular-weight species that were
lumped into a fraction called ‘heavies’. Acetone conversion and
selectivity to isophorone obtained over the vapor-deposited Cs/
NPC catalyst for various temperatures are displayed in Table 1.
Acetone conversion and selectivity to isophorone both maxi-
mize at 225–250 °C. At 20 °C isophorone selectivity drops due
to increased mesityl oxide formation. At temperatures above
250 °C the activity of the catalyst dropped even though the
selectivity to isophorone remained high. This was most likely
due to the accumulation of heavy products in the pores. Each
catalyst dropped to < 20% of its initial activity within 5 h and
gained approximately 10% of its initial weight. At 225 and
250 °C the turnovers (mol of acetone converted/mol Cs) before
catalyst deactivation were at a maximum of ≈ 10 (Table 1).
Significantly, the effect of air exposure on the catalyst before
reaction was to increase its initial activity. Catalyst samples that
were not air-exposed before reaction did not produce iso-
phorone in the first hour on stream. After this induction period,
however, they behaved as the air exposed samples did.
When compared to catalysts prepared by vapor deposition of
elemental Cs, those samples that were prepared with CsOH
solution‡ were less active and less selective. The selectivity to
isophorone over the catalyst prepared by vapor deposition was
ca. 60% throughout the experiment, whereas the selectivity for
the solution-impregnated sample began at 45% and declined to
30%. Both materials gained approximately 10% of their initial
weight in non-volatile compounds.
The vapor deposition of Cs led to catalysts that maintained
their high selectivity to isophorone even as their activities
dropped. The selectivity to isophorone remained 60% over a
conversion range from 0–20%. As the acetone conversion
approached 20% the loss in selectivity due to heavies formation
became more significant. Attempts made to achieve higher
conversions were not successful since the catalyst deactivated
more rapidly due to non-volatile product accumulation.
Finally, as control experiments, we tested the NPC support
alone without Cs and a sample of Cs-loaded NPC containing
only nanopores.§ Neither material led to any significant acetone
conversion. The Cs on all-nanoporous carbon catalyst, however,
did increase in mass by an amount equal to 10% of its initial
mass.
This work was supported by the Department of Energy,
Office of Basic Energy Sciences; the Delaware Research
Partnership; and the E. I. duPont deNemours and Co., Inc.
Notes and references
† A Tylan mass flow controller regulated an argon (Matheson Grade)
carrier-gas flow. The argon flow served two purposes: to purge the reactor
system of air and to carry acetone at its room-temperature vapor pressure
1
from a bubbler to the reactor. The reactor consisted of a ⁄2B o.d. quartz tube
heated by heat tape, the temperature of which was regulated by an Omega
PID temperature controller. The catalyst was placed in the heated zone of
the reactor between two plugs of glass wool. After leaving the reactor, the
acetone–argon stream was cooled and condensed into a catch pot by a shell
and tube condenser, through which flowed 210 °C chilled glycol. Samples
were taken by removing all the liquid from the catch pot with a syringe
inserted through a serum cap, weighed, spiked with a biphenyl standard, and
analyzed on a Varian GC with an FID detector.
‡ A sample of CsOH/NPC was prepared by stirring the NPC with 50%
CsOH solution for 1 h, filtering, and vacuum drying at 350 °C for 24 h.
§ NPC prepared by the pyrolysis of poly(furfuryl alcohol) alone at 850 °C
for 8 h under helium produces a purely nanoporous material with no pores
larger than 2 nm.
1 M. G. Stevens and H. C. Foley, Chem. Commun., 1997, 519.
2 M. G. Stevens, K. M. Sellers, S. Subramoney and H. C. Foley, Chem.
Commun., 1998, in press.
3 H. C. Foley, M. S. Kane and J. F. Goellner, in Access in Nanoporous
Materials, ed. J. J. Pinnavaia and M. F. Thorpe, Plenum, New York,
1995.
4 D. S. Lafyatis, J. Tung and H. C. Foley, Ind. Eng. Chem. Res., 1994, 30,
865.
5 M. S. Kane, L. C. Kao, R. Mariwala, D. F. Hilscher and H. C. Foley, Ind.
Eng. Chem. Res., 1996, 35, 3319.
6 G. Malinoski, W. H. Bruning, J. Am. Chem. Soc., 1967, 89, 5063.
7 M. G. Stevens, S. Subramoney and H. C. Foley, Chem. Phys. Lett., 1998,
292, 352.
8 K. Schmitt, Chem. Ind., 1966, 18, 204.
9 K. Othmer, Encyclopedia of Chemical Technology, Wiley, New York,
4th edn., 1995, vol. 14 p. 1000.
The GCMS data obtained indicate that the higher-energy,
non-cyclic trimer was the species that led to heavies formation;
none of the heavy materials identified contained cyclic
Table 1 Conversion, carbon selectivity towards isophorone and Cs0
turnovers at various temperatures. Argon flow rate: 150 cm3 mol21
saturated with acetone at 25 °C
Conversion (%)
(per g Cs)
Isophorone
selectivity (%)
T/°C
Cs Turnovers
200
225
250
275
300
6.1
11.9
11.4
5.2
54
61
60
57
57
5.34
10.43
10.04
4.56
3.8
3.38
Communication 8/08249I
276
Chem. Commun., 1999, 275–276