Ultraselective epoxidation of butadiene on Cu{111} and the effects of Cs promotion [6]

Full text of DOI:10.1021/ja994125j

J. Am. Chem. Soc. 2000, 122, 2381-2382
2381
Ultraselective Epoxidation of Butadiene on Cu{111}
and the Effects of Cs Promotion
James J. Cowell, Ashok K. Santra, and Richard M. Lambert*
Department of Chemistry, UniVersity of Cambridge
Cambridge CB2 1EW, England
ReceiVed NoVember 29, 1999
Heterogeneously catalyzed alkene epoxidation is of great
interest from both academic and technological points of view.
Ethene epoxidation, for which Ag is an apparently unique catalyst,
is a strategically important large-scale industrial process. As a
result, elucidation of the mechanism of Ag-catalyzed epoxidation
has attracted a great deal of academic research and many key
features of the reaction may be regarded as well understood.
Studies on single-crystal surfaces of silver have provided most
of the fundamental insight, including the identification of oxygen
adatoms, as opposed to adsorbed dioxygen, as the key epoxidizing
agent.1 Similarly, the mode of action of adsorbed chlorine1 and
alkali2 promoters has been determined. A recent review is
provided by ref 3. It is desirable to broaden this chemistry to the
selective oxidation of certain higher terminal alkenes whose
epoxides are valuable and versatile intermediates: propene4 and
butadiene are important examples. As part of this effort, one
should also seek alternative catalysts to silver. Recently, using
styrene as a model terminal alkene, we showed that single crystal
surfaces of Cu can act as highly effective epoxidation catalysts.5
Here we provide the first demonstration that the clean Cu-
{111} surface can catalyze the epoxidation of a desired target
molecule, butadiene, to 3,4-epoxybut-1-ene with extremely high
selectivity: in this respect Cu far outperforms Ag. We also show
that coadsorbed Cs strongly promotes the epoxidation of buta-
diene, without detracting from selectivity, in contrast with the
behavior of alkalis in Ag-catalyzed ethene epoxidation.2 The
mechanistic implications of these findings are discussed. Epoxy-
butene is important because the molecule is highly functional,
with each carbon atom chemically distinct. This allows manu-
facture of such diverse products as hydroxy ethers, glycols, and
amino alcohols as well as intermediates for pharmaceutical and
agricultural chemicals.
Temperature-programmed reaction (TPR) and X-ray photo-
electron spectroscopy (XPS) measurements were carried out in
an ultrahigh vacuum apparatus, described elsewhere.6 Gas expo-
sures are given in Langmuirs (1 L ) 1 × 10-6 Torr s-1) and the
TPR data were acquired at a heating rate of 5 K s-1. Reaction
products were identified on the basis of their mass spectral
fragmentation patterns. Cs7 and O8 coverages (θCs and θO) were
estimated from XPS intensities, by reference to calibrations
available in the literature. These quantities are specified in
fractional monolayers (ML).
Thermal desorption spectra obtained following butadiene
chemisorption at 170 K exhibited a maximum at ∼230 K and a
poorly resolved shoulder at ∼350 K. Temperature-programmed
reaction data were obtained by preadsorbing varying amounts of
oxygen at 170 K, followed by a saturation dose of butadiene
Figure 1. The conversion of butadiene to epoxybutene as a function of
oxygen coverage over Cu{111}. Inset: Representative mass spectrum
of butadiene and epoxybutene for θO ∼ 0.04 ML.
(2 L) at 170 K. The resulting spectra showed the formation of
epoxybutene, detected at m/z 42, and representative results are
shown in the inset to Figure 1. No CO2, H2O, furan, formaldehyde,
2-butenal, propenal, 2,5-dihydrofuran, nor any other products were
detected under any conditions, indicating extremely efficient
epoxidation with selectivity of ∼100%.
The yield of epoxide passed through a maximum as the oxygen
pre-coverage was increased; it was highest for an oxygen pre-
dose of 1L (θO ∼ 0.04 ML), reactivity eventually being quenched
at θO ∼ 0.5 ML (Figure 1). This latter condition corresponds to
coverage of the surface by a disordered Cu2O-like phase.9 The
reactively formed epoxybutene exhibited a broad maximum at
∼250 K tailing to higher temperatures. Note that the uptake of
butadiene was not significantly affected by the amount of
preadsorbed oxygen, so that the falloff in butadiene production
was not due a decrease in alkene coverage.
These results indicate that minimum necessary and sufficient
conditions for butadiene epoxidation on Cu{111} are butadiene
molecules adsorbed on Cu metal sites in the vicinity of oxygen
adatoms. The “oxidic” oxygen formed at high coverages is inef-
fective, even though plenty of the alkene is available. The impli-
cation is that the chemical state of oxygen is critically important,
oxidic oxygen being an ineffective electrophile with respect to
the adsorbed alkene. In this regard the epoxidation chemistry
parallels that observed on Ag where the valence charge density
on oxygen determines selectivity.1 On Ag, the charge density on
oxygen can be diminished by coadsorbed chlorine,1 thus creating
a better electrophile and favoring epoxidation. Consistent with
this, alkalis have the opposite effect2 favoring combustion over
epoxidation. Cu, unlike Ag, readily forms the metal oxide, even
under ultrahigh vacuum conditions. If, as seems likely, the tran-
sition from chemisorbed oxygen to Cu oxidation is accompanied
by a significant change in the valence charge density on oxygen,
the present results may be understood on a similar basis to that
proposed for the case of Ag.1 That is, oxygen chemisorbed on
Cu is a good electrophile for the π-adsorbed alkene whereas the
more highly charged oxidic oxygen (formally O2-) is a poor
electrophile, and therefore ineffective for epoxidation.
* Address correspondence to this author. E-mail: RML1@cam.ac.uk.
(1) Grant, R. B.; Lambert, R. M. J. Catal. 1985, 92, 364.
(2) Grant, R. B.; Lambert, R. M. Langmuir 1985, 1, 29.
(3) Serafin, J. G.; Liu, A. C.; Seyedmonir, S. R. J. Mol. Catal. A 1998,
131, 157.
(4) Hayashi, T.; Tanaka, K.; Haruta, M. J. Catal. 1998, 178, 566.
(5) Cowell, J. J.; Santra, A. K.; Lindsay, R.; Lambert, R. M.; Baraldi, A.;
Goldoni, A. Surf. Sci. 1999, 437, 1.
(6) Horton, J. H.; Moggridge, G. D.; Ormerod, R. M.; Kolobov, A. V.;
Lambert, R. M. Thin Solid Films 1994, 237, 134.
(7) de Carvalho, A. V.; Woodruff, D. P.; Kerkar, M. Surf. Sci. 1994, 320,
315.
(8) Niehus, H. Surf. Sci. 1983, 130, 41.
(9) Jensen, F.; Besenbacher, F.; Stensgaard, I. Surf. Sci. 1992, 269/270, 400.
It is of interest to examine the effects of coadsorbed alkali on
this system. Recall that in the case of Ag, added alkali strongly
decreases selectivity toward epoxide formation, both for ethene
10.1021/ja994125j CCC: $19.00 © 2000 American Chemical Society
Published on Web 02/24/2000

2382 J. Am. Chem. Soc., Vol. 122, No. 10, 2000
Communications to the Editor
Figure 4. Reaction scheme proposed for butadiene epoxidation over Ag.
To understand how Cs modifies the catalytic behavior of the
copper/oxygen system, we carried out XPS measurements to
examine the chemical state of the system as a function of Cs
coverage. This was achieved by depositing varying amounts of
Cs on a pre-oxidized Cu2O-like surface and monitoring the
changes observed in the O 1s region. The Cu2O-like surface was
prepared by exposing the clean surface to 105 L of O2 at 670 K.9
Figure 3 shows the resulting curve-fitted O 1s spectra for different
Cs loadings. The Cu2O-like surface (spectrum a) exhibited a single
peak at 529.8 eV. Addition of 0.25 ML and then another 0.25
ML of Cs to this surface led to the progressive growth of two
new features at 528.7 and 531.3 eV, while the intensity of the
529.8 eV feature decreased (spectra b and c). At θCs ) 0.75 ML
the feature at 529.8 eV was undetectable, while the other two
features became more prominent. O 1s emissions at 528.7 and
531.3 eV are associated with CsxO and Cs2O, respectively.
Thus the results clearly show that Cs leads to the reduction of
oxidic copper sites to metallic copper sites, yielding a mixture of
Cs2O and CsxO.
Figure 2. Conversion of butadiene to epoxybutene as a function of oxy-
gen coverage over caesiated Cu{111}, θCs ∼ 0.07 ML. Inset: Representa-
tive mass spectrum of butadiene and epoxybutene for θO ∼ 0.04 ML.
We may therefore write:
Cu2O + Cs f Cu + Cs2O + CsxO
It thus appears that Cs acts to prevent the formation of copper
oxide, which according to our results is catalytically inert.
It is interesting to compare Cu with Ag in regard to butadiene-
selective oxidation. Single-crystal studies on Ag{110} showed
no evidence for epoxide formation, the preferred products being
dihydrofuran and furan formed via a 1,4-cycloaddition.11 Buta-
diene can be selectively oxidized to epoxybutene over supported
silver catalysts12 but at only 1% conversion and 50% selectivity.
The failure to observe epoxidation in the single-crystal case was
attributed12 to further conversion of the epoxide; the relevant
processes are summarized in Figure 4.
We conclude that under TPR conditions butadiene may be
selectively oxidized to epoxybutene over Cu{111} at high
conversion and with close to 100% selectivity. The reaction occurs
between oxygen adatoms and butadiene molecules adsorbed on
Cu metal sites. Over-oxidation of the metal suppresses catalytic
activity and Cs acts to enhance the reaction yield, without
decreasing selectivity, by a clean-off reaction which regenerates
Cu metal sites.
Figure 3. O 1s X-ray photoelectron spectra showing conversion of Cu2O
to Cu as a function of Cs dose: (a) no Cs; (b) 0.25 ML Cs; (c) 0.5 ML
Cs; and (d) 0.75 ML Cs.
itself2 and for the model alkene styrene.10 The clean surface was
pre-dosed with varying amounts of oxygen, followed by a fixed
∼0.07 ML dose of Cs and then a (saturating) butadiene dose of
2 L. The results are illustrated in Figure 2 along with typical TPR
spectra, which are shown in the inset. Again, no products other
than epoxybutene were detected. Thus in contrast with Ag-
catalyzed epoxidation, alkali has no effect on selectivity, which
remains very high (effectively 100% within the accuracy of our
measurements). Instead, for any given oxygen coverage, Cs
strongly increases the amount of butadiene conversion to epoxy-
butene (cf. Figures 1 and 2). For θO ∼ 0.1, ∼90% of the adsorbed
butadiene is epoxidized. It is important to note that the increased
yield of epoxide is not due to increased butadiene adsorption:
this is hardly affected by to the presence of ∼0.07 ML of Cs.
For oxygen pre-coverages >0.08 ML, conversion decreased
progressively, being fully quenched at 0.5 ML, just as in the case
of the alkali-free surface. However, as inspection of Figures 1
and 2 shows, for any given oxygen pre-coverage the effect of Cs
was to (approximately) double the yield of epoxide.
Acknowledgment. J.J.C. acknowledges the award of an EPSRC
CASE studentship and additional support from BP-AMOCO Chemicals
plc. A.K.S. acknowledges support from BP-AMOCO Chemicals plc.
Financial support from the UK Engineering and Physical Sciences
Research Council under Grant GR/M76706 is gratefully acknowledged.
JA994125J
(10) Hawker, S.; Mukoid, C.; Badyal, J. P. S.; Lambert, R. M. Surf. Sci.
1989, 219, L615.
(11) Roberts, J. T.; Capote, A. J.; Madix, R. J. J. Am. Chem. Soc. 1991,
113, 9848.
(12) Monnier, J. R. from 3rd World Congress on Oxidation Catalysis;
Grasselli, R. K., Oyama, S. T., Lyons, J. E., Eds; Elsevier Science B.V.: New
York, 1997.