Acid catalysts based on Cu/Ru alumina: Conversion of butyraldehyde to dibutyl ether

Article abstract of DOI:10.1006/jcat.1996.0327

A system made by combining two nonalloying metals, ruthenium and copper, using alumina as the oxide support was studied. This bimetallic supported catalyst has been used mainly in hydrogenolysis, dehydrogenation, and oxidation reactions of hydrocarbons. The preparation of such materials has been proposed to effect the selectivity and activity of a highly active metal by inclusion of a second less active metal. The samples were characterized by electron paramagnetic resonance spectrometry (EPR), X-ray diffraction (XRD), surface area, and surface acidities. The techniques EPR and XRD are ideal for studying the electronic and structural changes of the samples at different temperatures and concentrations. The primary reaction involved in this study was the hydrogenation of an aldehyde to the corresponding alcohol. A secondary reaction occurred as well. The acid catalyzed, substitution or bimolecular dehydration of the alcohol to the dibutyl ether was observed under certain catalytic conditions. These catalysts appeared to act as acid/base. Therefore this reaction to produce the ether is of special importance. A correlation between the electronic, structural and catalytic properties has been made to understand molecular processes' role in catalytic phenomena.

Full text of DOI:10.1006/jcat.1996.0327

JOURNAL OF CATALYSIS 163, 262–270 (1996)
ARTICLE NO. 0327
Acid Catalysts Based on Cu/Ru Alumina: Conversion
of Butyraldehyde to Dibutyl Ether
ꢀ
ꢀ
ꢀ
Susan Jansen, Michael Palmieri, Maria Gomez, and Steve Lawrence†
ꢀ
Department of Chemistry, Temple University, Philadelphia, Pennsylvania 19122; and †Department of Chemistry,
Saginaw Valley State University, University Center, Michigan 48710
Received July 6, 1995; revised May 15, 1996; accepted May 24, 1996
supported on oxide substrates, usually silica or alumina. As
was anticipated, the catalytic properties of the mixed metal
systems varied widely depending on the preparation, sup-
port characteristics, and catalytic application; i.e., varying
effects on activity and selectivity were observed for oxida-
tion and reduction.
A system made by combining two nonalloying metals, ruthenuim
and copper, using alumina as the oxide support was studied. This
bimetallic supported catalyst has been used mainly in hydrogenoly-
sis, dehydrogenation, and oxidation reactions of hydrocarbons. The
preparation of such materials has been proposed to effect the selec-
tivity and activity of a highly active metal by inclusion of a second
less active metal. The samples were characterized by electron para-
magnetic resonance spectrometry (EPR), X-ray diffraction (XRD),
surface area, and surface acidities. The techniques EPR and XRD
are ideal for studying the electronic and structural changes of the
samples at different temperatures and concentrations. The primary
reaction involved in this study was the hydrogenation of an alde-
hyde to the corresponding alcohol. A secondary reaction occurred
as well. The acid catalyzed, substitution or bimolecular dehydra-
tion of the alcohol to the dibutyl ether was observed under certain
catalytic conditions. These catalysts appeared to act as acid/base.
Therefore this reaction to produce the ether is of special impor-
tance. A correlation between the electronic, structural and catalytic
properties has been made to understand molecular processes’ role
In this study we focused on electronic features in
Cu/Ru/Al2O3 systems that define the catalytic properties
of a fairly simple bimetallic catalyst. The metals were cho-
sen for several reasons. (i) Copper aluminates, which adopt
the spinel structure, are known to form in preparation of
ceramic materials (7). The inclusion of a secondary metal
may affect the formation of the stable spinel structure and
the metal oxide surface state. (ii) Formation of ruthenium
oxides or related integrated ruthenium-copper oxides are
expected to exhibit modified electronic properties. (iii) The
structural data obtained will be of utility for the assessment
of catalytic properties of mixed metal systems. To date, at-
tempts at correlation of the catalytic or electronic proper-
tieswith the structuralor electronicphasesproduced during
preparation or surface catalytic species have proven diffi-
cult. This is in part because direct characterization of the
structural phases has been limited. Here, a variety of struc-
tural, electronic and chemical tools were applied to assess
the surface and bulk properties of the copper-ruthenium
system.
Catalytically, ruthenium dioxide is primarily uased for
oxidation of hydrocarbons and for hydrogenation and de-
hydrogenation of alkanes (8). Copper oxide is primarily
used for oxidation of hydrocarbons, dehydration, dehydro-
genation of alcohols, and reduction of aldehydes (9). The
bimetallic system is typically used for hydrogenolysis, de-
hydrogenation, and oxidation of hydrocarbons (10). This
work will provide an extensive characterization of the cata-
lytic properties of a copper-ruthenium system supported
on alumina. The hydrogenation reaction of butyraldehyde
was used to study the chemistry of the Ru–Cu/Al2O3 cata-
lysts. The active species in this reaction was expected to
be the cupric sites from copper oxide which is primarily
used for the reduction of aldehydes to alcohols and the
in catalytic phenomena.
ꢁc 1996 Academic Press, Inc.
INTRODUCTION
In the past 20 years there have been multiple reports on
mixed metal catalysts (1–5). The general motivation for the
preparation and study of such materials has been to affect
the selectivity of a highly active metal by inclusion of a sec-
ond less active metal. The primary metal is supposed to be
the “active” component while the secondarymetalischosen
to induce a pronounced selectivity in the catalytic reaction.
Furthermore, the metallic components were selected based
on certain electronic and packing characteristics including
metals capable of forming alloys in the metallic state such
as nickel and copper, copper and gold, and ruthenium and
copper. Nickel and copper and copper and gold are known
to form alloys in the bulk while it has been determined that
copper can grow pseudomorphically atop the (0001) sur-
face of ruthenium with very little tensile strain. Copper and
ruthenium, however, show no significant propensity to alloy
(6). Catalytic applications require that these materials be
262
0021-9517/96 $18.00
c
Copyright ꢁ 1996 by Academic Press, Inc.
All rights of reproduction in any form reserved.

ACID CATALYSTS BASED ON Cu/Ru ALUMINA
263
oxidation of alcohols to aldehydes. The oxygen atom of the than MTBE would be advantageous particularly if crude
RCOH and ROH is understood to be the site of attach- olefinic refinery streams could be used as feedstock (16–18).
ment for the reactant at a copper/CuO site on the surface.
The catalytic reduction of the aldehyde to its correspond-
ing alcohol, typically proceeds with the adsorption of the
EXPERIMENTAL
aldehyde onto the catalyst and then dissociative adsorption
==
I. Preparation of Materials
of H2 and H addition across the C O bond. With addi-
The materials were prepared by the incipient wetness
method (19), in which two different procedures were
employed. These are termed (i) “simultaneous” and (ii)
“stepwise” depositions: (i) In simultaneous deposition the
impregnating solutions containing the two metal salts
[CuNO3 ꢂ 6H2O (0.0393 M) and RuCl3 ꢂ xH2O (0.0247 M)
were added to the alumina simultaneously. The two metal
salts were purchased from Aldrich Chemical Company and
the alumina support (150 m2/g, 80–200 mesh) was pur-
chased from Fisher Chem Alert. (ii) The step-wise depo-
sition method divides the addition of the impregnating so-
lutions into two steps. First, one metal solution was added
to the alumina, dried (100ꢃC for 10 h), and then calcined
(300ꢃC for 12 h). The second metal solution was then added
and the final heat treatments were performed. These cat-
alysts all contained 1% rethenium by weight with varying
amounts of copper (0.15 to 1.1 wt% ).
==
tion of the hydrogen across the C O bond the aldehyde is
reduced/hydrogenated to the corresponding alcohol.
The alcohol may further react to produce certain ethers.
It is well known that the synthesis of symmetrical dialkyl
ethers (dibutyl ether) can occur by bimolecular dehydra-
tion from simple unhindered primary alcohols (1-butanol).
Under acidic dehydration conditions two reactions com-
pete: elimination to give an alkene and substitution to give
an ether. Substitution to produce the ether is a bimolecu-
lar dehydration and the elimination reaction to form the
alkene is a unimolecular hydration (11).
In general, the later transition metals like ruthenium
(RuO2) and copper (CuO) on oxide supports have not been
thought of as having sufficient acidic properties to affect
ether synthesis. The earlier transition metal oxides such as
tungsten (WO3) and molybdenum (MoO3) on an oxide sup-
port (TiO2) have been shown to have greater surface acidi-
ties than the support and have been used in many reactions
such as oxidation of methanol to formaldehyde and hydra-
tion of propylene (12–14). Not surprisingly Ru/Al2O3 and
Cu/Al2O3 systems have rarely been used as acid catalysts
and have been used mainly for hydrogenation, dehydro-
genation of alkanes, oxidation of alcohols, and reduction
of aldehydes. Even though there are few papers discussing
the catalysis of the Ru–Cu/Al2O3 mixed metal system, the
catalytic applications usually parallel the individual uses of
RuO2 and CuO on alumina. In this work it will be shown
that ruthenium and copper supported on an oxide substrate
has measurable surface acidity and can be used as an acid
catalyst, as in the formation of dibutyl ether as well as other
dialkyl ethers. Their potential as bifunctional or multifunc-
tional catalysts will be discussed in this paper.
The actual metal content/loading was determined by
atomic absorption (AA). The AA measurements were per-
formed on a BUCK Scientific (Model 200A) instrument.
The theoretical and actual amounts of metal content are
reported in Table 1.
II. Electron Paramagnetic Resonance (EPR)
EPR measurements were performed on a Bruker
ER-200D spectrometer, fitted with an ER035 gaussme-
ter to provide accurate magnetic field measurement to
0.001 G/1000 G. For variable temperature measurements,
temperature control was maintained by a VT-4114 unit
TABLE 1
Due to environmental concerns and new stringent en-
vironmental laws, there is a shift toward developing new
“cleaner” burning fuels. Ethers, such as methyl tertiary
butyl ether (MTBE), ethyl tertiary butyl ether (ETBE),
and tertiary amyl ether (TAME) have been shown to be
Materials, Percentage Loadings, Surface Areas and
Hammett Acidities
Mole
Surface
Catalyst
Ru (wt% )
Cu (wt% ) Cu–Ru areas (m2/g)
H0
cleaner burning fuels but MTBE is the most widely used of Cu/Al2O3
0.00
0.00
1.0(0.99)
14.6(14.55)
1.0(0.97)
0.00
0.15(0.16)
1.1(1.13)
0.15(0.14)
1.1(1.14)
0.00
—
—
—
0.25
1.50
0.25
1.50
0.00
222
212
197
204
209
204
209
152
ꢄ NA
ꢄ + 2
ꢄ + 1
ꢄ + 2
ꢄ + 2
ꢄ + 2
ꢄ + 2
ꢄ + 1
Cu/Al2O3
Ru/Al2O3
the ethers. ETBE and TAME have not found wide appli-
cations due to the relative high cost of ethanol compared
to methanol and the limited availability of iso-pentene rel-
ative to iso-butene (15) as they offer volatility advantages
Ru–Al2O3–Cu 1.0(1.01)
Ru–Al2O3–Cu 1.0(1.05)
Ru–Cu/Al2O3 1.0(1.06)
to MTBE. Developing better and more cost effective pro- Ru–Cu/Al2O3 1.0(1.03)
a
Al2O3
0.00
cesses for making these higher ethers is important. Catalysis
will play a significant role in developing new routes for these
more environmentally safe fuels. Therefore new inexpen-
sive catalytic routes to alternative ether components other
Note. The numbers in the parentheses are the actual weight percents
determined by AA.
a Support.

264
JANSEN ET AL.
hol (1-butanol) with the possible formation of the subse-
O
OH
ꢀ
ꢁ
quent ether (dibutyl ether). These two reactions are shown
in Schemes I and II. These catalysts were studied by a gas-
solid reaction which was carried out under continuous-flow
conditions in a glass tubular reactor (29). The reactor tube
was mounted in a tube furnace and the feed stream was
100% hydrogen. Since the products were in the gas phase,
GC was the method employed to study the reaction. The
column used was a 0.10% SP-1000, 80/100 carbopack, and
60 ꢇ 1/800 in size. The reaction conditions used for the reduc-
tion of butyraldehyde to 1-butanol was a temperature of
150ꢃC, a hydrogen flow of 10ml/min, and 160 mg of catalyst.
ꢀ
ꢁ
Ru–Cu/Al2O3
HCCH2CH2CH3 + H2 ꢅꢅꢅꢅꢅꢅꢅ→ H2CCH2CH2CH3
SCHEME I. Reduction/hydrogenation of aldehyde to alcohol.
150ꢃC
which maintains temperature within 0.5ꢃC. A series of EPR
measurements wwere made on materials in which the over-
all loading was relatively low, 1–3% by weight. For these
samples, the theoretical coverage limits were less than one-
tenth of a monolayer and as such, these materials provided
the potential for analysis of isolated metal-oxide interac-
tion. The potential for aggregation can also be addressed, as
any cooperative metal-metal/oxide interaction will appear
rather pronounced in the EPR characterization at such low
loadings and coverages. A complementary study at high
loading, 1/2 and full monolayer, i.e., 15–35% by weight was
also undertaken to determine the structural limits affecting
catalytic and electronic properties.
RESULTS AND DISCUSSION
XRD
The structural characterization of the calcined catalysts
was performed by XRD on the highly loaded and low
loaded materials. The highly loaded pure copper on alu-
mina showed crystalline CuO, Al2O3, and AlOOH peaks,
whereas the low loaded pure copper on alumina showed
Al2O3 and AlOOH peaks, only. It has been determined
from the XRD that the major components of the alumina
are the Al2O3 and AlOOH peaks. Using a peak finding pro-
gram (JCPDS-ICDD), it was established that these peaks
had an 85% correlation to the gamma phase. In the mixed
metal catalysts the dominant species present in the XRD
was RuO2. CuO was not present in the X-ray diffraction
spectra because it was either in a non crystalline form or
present primarily at the surface. Figures. 1a and 1b show
the XRD spectra.
As stated earlier, copper was chosen because it forms
copper aluminates which adopt the spinel structure. The
spinel structure consists of a close-packed lattice of oxide
ions with one-third of the metal ions in tetrahedral sites
and two-thirds of the metal ions in octahedral sites. In a
“normal” spinel, the M (+2) ions occupy tetrahedral sites
and the M (+3) ions occupy the octahedral sites. In an
“inverse” spinel, part of the M (+3) ions exchange positions
with the M (+2) ions, therefore, some of the M (+2) ions
are in octahedral sites (30). Copper aluminate (Cu/Al2O3)
is a partially inverse cubic spinel (31). The spinel structure is
shown in Fig. 2. It has been shown that under certain prepa-
ration conditions copper supported on high surface area
alumina forms crystalline aluminates, whereas at high level
of metal on a lower surface area support, CuO was predom-
inant (bulk material). This aluminate phase can be detected
even on material calcined as low as 300ꢃC (7, 32). In this
III. Catalyst Characterization
The calcined materials were characterized by XRD.
XRD measurements were performed on samples which
ranged from 1 to 2% in copper and ruthenium. Surface
area measurements for the highly loaded/low dispersion
and low loaded/high dispersion calcined materials were
determined using a liqud N2 BET adsorption appara-
tus. Acidities were assessed by surface state titration.
The surface acidity of these samples was evaluated using
the n-butylamine titration method established by Tanabe
(20–24). The total acidity is related to the number of
sites having acid strengths with H0 ꢆ +2.0. The compo-
sition, surface areas and relative Hammett acidities are
reported in Table 1. Infrared spectra recorded during
temperature-programmed decomposition (TPD) measure-
ments of ammonia, monomethylamine (MMA), dimethy-
lamine (DMA), and trimethylamine (TMA) have shown
that alumina and alumina supported metal catalysts con-
tain Lewis—as well as Brønsted—acid sites (25–28).
IV. Catalysis: Analysis of Reaction Mixture
In order to understand the interaction of copper (cop-
per oxide) and ruthenium (ruthenium dioxide) on the sup-
port, a study was performed by choosing a reaction that
was suitable for copper (CuO) states. This enabled us to
study the effects that ruthenium would have on copper.
The reaction of interest was the hydrogenation (reduction)
of an aldehyde (butyraldehyde) to its corresponding alco-
OH
ꢁ
ꢁ
2H2CCH2CH2CH3 + H+ → H3CCH2CH2CH2–O–CH2CH2CH2CH3
SCHEME II. Bimolecular dehydration of alcohol to dialkyl ether.

ACID CATALYSTS BASED ON Cu/Ru ALUMINA
265
FIG. 1. (a) XRD of highly loaded and low loaded Cu/Al2O3. (b) XRD of highly loaded and low loaded Ru–Cu/Al2O3.
work, a moderately high surface area alumina (150 m2/g) the copper centers which are observable in the EPR. This
was used and by XRD, the spinel structure copper alumi- maybe due to isolation of species or “interference” from
nates (CuAl2O4) were detected at a calcination tempera- electronic factors contributed by ruthenium. Similarly, cop-
ture of 500ꢃC. The XRD is shown in Fig. 3. Spinel formation per aluminates (CuAl2O4) are EPR silent (35). Since the
provides a possible mechanism for Cu/Ru aggregation on only EPR signal observed for our catalysts is from a tetrag-
alumina.
onally distorted octahedral coordination, this suggests that
some of the copper(II) ions are occupying the octahedral
sites of the spinel. Due to these “defect” features, an
EPR signal can be detected at room temperature. It has
been shown in other work that the concentration of these
“defect” states are proportional to the surface area (36).
There is an inverse correlation between the amount of
copper loaded and the spin concentration of copper (para-
magnetic species); i.e., the more copper present on alumina
the lower the spin concentration (33, 34, 37). This is due to
the increasing formation of CuO domains. Therefore, the
highly loaded/low dispersion materials have fewer spin ac-
tive species than the low loaded/high dispersion materials.
EPR
In copper(II) oxide, copper II has a d (9) electron con-
figuration (S = 1/2, I = 3/2) and is usually present in a
tetragonally distorted octahedral coordination. In this work
“CuO” was the species being analyzed but it is “defect”
states that give rise to the EPR signal. When copper is
loaded and calcined on alumina copper oxides are formed.
If significant CuO domains form, the copper centers are
antiferromagnetically coupled to other copper centers and
therefore are EPR silent (33, 34). The defect states are

266
JANSEN ET AL.
FIG. 2. Spinel structure.
The EPR spectra for the “differently” prepared catalysts,
i.e., simultaneous, stepwise depositions and high and low
loaded, are shown in Fig. 4.
The EPR signal for the copper species is nearly axial,
where gxx = gyy = gzz. The gand hyperfine couplingtensors
can be studied and information as to coordination effects
can be extracted from qualitative molecular orbital theory
(38). Because the g factor is extremely sensitive to change in
the chemicalenvironment, information about the electronic
environment for the copper environment as a function of
preparation can be studied. Since the signal is anisotropic,
FIG. 4. The EPR spectra of the differently prepared catalysts.
copper oxide. The theory of Maki and McGarvey for EPR
relates the molecular states of the cupric ion to g and g
k
⊥
of copper(II) complexes (38). This information can also be
used to calculate the delocalization of the unpaired electron
and the ꢀ bond density, see Eqs. [1] and [2]:
g (gzz), g (gxx , gyy) and A (hyperfine coupling) can be ob-
k
⊥
k
ꢂ
ꢃ
ꢄ
ꢅ
g = ge ꢅ 8ꢁ0ꢂꢃ1/1xy ꢂꢃ1 ꢅ 1/2ꢂ0 1 ꢅ ꢃ2 1/2T (n)
[1]
tained through simulation (39). These features can give in-
sight on how the chemical environment is changing for the
k
1
g
= ge ꢅ 2ꢁ0ꢂꢃ/1xz,yz[ꢂꢃ ꢅ 1/(2)1/2ꢂ0(1 ꢅ ꢃ2)1/2T (n)].
⊥
[2]
Here ꢂ2 is related to the delocalization of the unpaired elec-
tron, 1ꢅꢂ2 is the ꢀ bond strength between the central atom
and the ligands, and ꢂ is the molecular orbital coefficient
for the in plane ꢀ bonding of the copper (dx2 ꢅ y2) orbital.
The dx2 ꢅ y2 is the orbital where the unpaired electron “re-
sides” in a tetragonal distorted octahedral or in a square
planar complex and the 1E is approximately 15,000 cmꢅ1
(40). These parameters can be calculated from (41)
ꢂ2 = A /P + (g ꢅ 2) + (3/7)(g ꢅ 2) + 0.04.
[3]
k
k
⊥
The EPR parameters used in the above equation were cal-
culated using the EPR simulation program by Neese and
Nilges (39). Table 2 shows a list of the EPR features for the
“differently” prepared catalysts.
The EPR data suggest that there was a sizeable shift in
g
for the dominant magnetic species. Further evaluation
⊥
FIG. 3. XRD of the CuAl2O4 (1 wt% ).
in terms of the ꢀ bond strength and the molecular orbital

ACID CATALYSTS BASED ON Cu/Ru ALUMINA
267
TABLE 2
EPR Features
A
k
Catalyst
gka
g⊥a
(cmꢅ1
)
ꢂ2
1ꢅꢂ2
ꢂ
14.6% Cu/Al2O3
1% Cu/Al2O3
Cu–Al2O3–Ru
Ru–Al2O3–Cu
Ru–Cu/Al2O3
Spectral
2.3504
2.3504
2.3500
2.3504
2.3500
2.3500
2.0800
2.0800
2.0750
2.0650
2.0750
2.0750
0.0153
0.0153
0.0153
0.0153
0.0153
0.0040
0.85
0.85
0.85
0.84
0.85
0.53
0.15
0.15
0.15
0.16
0.15
0.47
0.92
0.92
0.92
0.91
0.92
0.73
subtraction
a
These values are constant within each catalyst series/do not change
with % Cu.
coefficient showed that for the dominant magnetic species
there were no significant differences between the highly
loaded, low loaded, and mixed metal catalysts. Evaluation
of the ꢄ bonding, which is related to g showed there was
⊥
approximately a 10% change in the ꢄ bonding character be-
tween the different catalysts. ThisEPR data showed that the
majority of the electron density was localized on the copper.
The dominant magnetic species for each catalyst seems
to be similar; however, the spectra envelope for the
Ru–Al2O3–Cu showed one distinct difference when com-
pared to the other spectra. Consequently, a spectral sub-
traction was performed between the Ru–Al2O3–Cu and the
Cu–Al2O3–Ru catalysts. The catalyst in which the copper
was deposited last, Ru–Al2O3–Cu, was more catalytically
active than the other materials. When the ruthenium was
deposited last the catalyst was relatively inactive. The re-
sults of the subtraction are shown in Fig. 5. The spectrum
remaining after subtraction gives insight on the catalytically
active species. This subtraction revealed that the magnetic
state was more highly delocalized with the spin density be-
ing distributed over the copper(II) oxide. The loss of re-
solvable hyperfine suggests a more mobile electron species.
This is shown in Table 2.
FIG. 5. The EPR spectral subtraction.
the conversion should be significantly less; however, this
was not the case.
For the mixed metal catalysts of Ru–Cu/Al2O3, Ru–
Al2O3–Cu, and Cu–Al2O3–Ru, the addition of ruthenium
changed both the activity and the selectivity of the reduc-
tion of butyraldehyde. In addition the preparation method
of these materials played a critical role in defining the ac-
tivity and selectivity of the Cu–Ru/Al2O3 catalysts. In the
catalysts where ruthenium was added last in the deposition
order (Cu–Al2O3–Ru), the catalytic activity was lost. This
is probably due to a simple site blockage by the ruthenium,
since the butyraldehyde cannot adsorb onto the catalyst
surface, where the adsorption sites are “masked” by ruthe-
nium. This begs the question as to aggregation. At this load-
ing less than 1/10th of the surface is covered. Site blockage
Catalysis
The first catalyst studied was the standard high coverage may occur if the copper serves as a “nucleation” site for the
(14.6 wt% ) Cu/Al2O3 which converted about 86.4 mole% secondary metal. Since copper appears to be the adsorp-
of the butyraldehyde, giving butanol as 99% of the total tion site of the reactant species, coordination with ruthe-
products. The low loaded pure 1 wt% Cu/Al2O3 was stud- nium would block the active site. Other work supports this
ied and showed significant activity. The total conversion to idea (37).
product was 22 mole% with a significant selectivity toward
Furthermore, it has also been determined that ruthenium
1-butanol. The ratio of 1-butanol to dibutyl ether (minor IV (RuO2) is reduced to ruthenium(0), during many cata-
product) was 4.4:1, where dibutyl ether was 18% of the lytic processes, when the temperature approaches or ex-
products. The 1% by wt Ru/Al2O3 was studied and showed ceeds 150ꢃC. This is due to the decomposition of the organic
no measurable activity. Of significant importance is that material (reactant), which deposits carbon onto the surface,
even though the 1% pure copper catalyst had 14.6 times causing reduction of the ruthenium. When the ruthenium
less metal content than the highly loaded sample, the total is in its reduced form (Ru0) the activity of the catalyst is
conversion was only 4 times smaller. If the conversion was lost (37). In the butyraldehyde reaction the temperature of
based solely on the metal content defining the metal sites, the catalytic reaction was at 150ꢃC, so some of the reactant

268
JANSEN ET AL.
may have decomposed and deposited on the catalyst, caus-
ing the ruthenium to be reduced. Evidence of this reduction
was seen in the XRD spectra of the used catalyst (37). This
is another process implicated in the loss in activity for the
Cu–Al2O3–Ru systems.
TABLE 4
Relative Activity of Copper Catalyst vs Mixed Metal Catalyst
1 wt%
Cu/Al2O3
(mole% )
1.5 mole ratio
Ru–Al2O3–Cu
(mole% )
Ru–Cu/Al2O3/
Time (min)
Cu/Al2O3 ꢇ 100a
When the reverse deposition order is used, where cop-
per is deposited last (Ru–Al2O3–Cu) in the preparation,
the catalyst showed significant activity relative to the pure
Cu/Al2O3 material. The total conversion to product ranged
from 5 to 15 mole% depending on the relative mole ratio
of copper to ruthenium. The selectivity was changed rela-
tive to the pure copper catalyst; here the ratio of 1-butanol
to dibutyl ether was 3:1 and the dibutyl ether was 25% of
the product formation. In the simultaneous deposition cata-
lysts (Cu–Ru/Al2O3) the activity was reasonable, where the
total conversion of product ranged from 4 to 14 mole% de-
pending on the mole ratio of copper to ruthenium. Here the
selectivity was significantly different than any of the other
catalysts, the ratio of 1-butanol to dibutyl ether was 1:1 and
the dibutyl ether was now 50% of the product formation.
These results are summarized in Table 3.
Always of interest is the lifetime or surface stability of
these systems. The lifetime of the pure Cu/Al2O3 catalyst is
extended when ruthenium is added to the material. The cat-
alytic activity of the 1% by wt Cu/Al2O3 decreased by 50%
after 95 min, and after 2 h the activity decreases to 25% ,
where it remained constant for days. Whereas the activity of
the mixed 1.5 mole ratio Ru–Cu/Al2O3 catalysts remained
constant over the same period of time as the pure copper
catalyst. Initially the pure copper catalyst has a higher mole
percent conversion than the mixed metal catalyst, but this
occurs only in the first hour. After 2 h and for the next 20 h
the mixed metal catalyst was approximately 260% more ef-
fective relative to the low loaded pure copper catalyst. In
addition the selectivity of the mixed metal material toward
the ether product was approximately 35% more favorable
than the pure copper sample. Table 4 shows the relative ac-
tivityofthe catalystsover 4–5h:1wt% Cu/Al2O3 vs1.5mole
ratio Ru–Cu/Al2O3.
60
95
120b
21.6
10.8
5.4
14.0
14.0
14.0
65%
130%
260%
a Ratio shows the effectiveness of the mixed metal to the pure cata-
lyst over time.
b
The 120 min point is cosntant for the two catalysts up to and
including 20 h (1200 min).
Reaction Chemistry and Proposed Mechanisms
For this catalytic reaction to be useful there must be good
adsorption of the butyraldehyde onto the catalyst. In order
to produce dibutyl ether, two conditions must be present.
(i) 1-butanol must be present at the surface of the cat-
alyst and (ii) the surface of the catalyst must be acidic.
Table 5 shows the Hammett acidities and production of
dibutyl ether.
The primary reaction is the reduction/hydrogenation of
the butyraldehyde to butanol, where hydrogen dissociates
across the double bond of the carbonyl group. The sec-
ondary reaction can be the bimolecular dehydration of
an unhindered primary alcohol (1-butanol) to the dialkyl
ether. Under acidic dehydration conditions, two reactions
can compete: substitution to give the ether (bimolecular)
and elimination (unimolecular) to give the alkene.
In Table 5 it is shown that the alumina support and the
ruthenium on the alumina support have similar acidities,
suggesting that the ruthenium had no significant effect on
the acidity of the support. When the copper was added to
the alumina and to the ruthenium on alumina the acidity
was decreased by a factor of 10 for the low loaded catalysts
and negligible (or effectively zero) for the highly loaded
copper catalyst. Pure CuO isknown to be considerablybasic
TABLE 3
Initial Catalytic Activity and Selectivity
Ru
(wt% )
Cu
(wt% )
ꢅmoles/
m2 ꢂ min
Mol% /
wt% a
Catalyst
Mol% b
% Butanolc
% Etherc
Cu/Al2O3
Cu/Al2O3
Ru–Al2O3–Cu
Ru–Al2O3–Cu
Ru–Cu/Al2O3
Ru–Cu/Al2O3
—
14.6
1.0
0.15
1.1
0.15
1.1
4.05 ꢇ 10ꢅ2
1.03 ꢇ 10ꢅ2
1.03 ꢇ 10ꢅ3
6.94 ꢇ 10ꢅ3
8.82 ꢇ 10ꢅ4
6.84 ꢇ 10ꢅ3
5.92
21.6
4.7
6.6
4.2
86.4
21.6
5.4
14.0
4.8
99
82
75
75
50
48
<1%
18
25
25
50
1.0
1.0
1.0
1.0
6.5
13.7
52
a Percentage by weight of total metal.
b Mole percentage of total product.
c Percentage distribution of products.

ACID CATALYSTS BASED ON Cu/Ru ALUMINA
269
(42). This suggests that the copper oxide contributes basic
sites upon formation on the alumina support. Also, it is im-
portant to note that the selectivity toward the ether was in-
creased when ruthenium was present in the catalysts. This is
probably due to the ability of ruthenium to transfer protons
and increase the mobility of the hydrogen species. It has
been shown that ruthenium chemisorbs hydrogen rather
easily in both homogeneous and heterogeneous reactions.
With the addition of copper the amount of chemisorbed hy-
drogen decreases (43). These mixed metal catalysts seem
to be acting as multifunctional catalysts where the alu-
mina surface may be the source of the protons (acid). The
basic copper oxide functions as a Lewis base in the ad-
sorption mechanism and is the site of attachment for the
aldehyde. Ruthenium sites are implicated in the hydrogen
processing.
Our proposed mechanism suggests that the aldehyde
must be adsorbed onto the surface of the catalyst at the
cupric site of the copper oxide islands. Once the aldehyde
is bound to the surface of the catalyst, it is reduced to its
corresponding alcohol. Then a secondary reaction, initiated
by acidic protons distributed over the exposed alumina,
occurs. This promotes the formation of the dibutyl ether.
The present mechanism suggests that some of the alcohol,
while present at the surface, is protonated (H+). The H+
which protonates the alcohol is the proton associated with
the support (AlOOH). The alcohol molecules present at
the surface of the catalyst which have not been protonated
may act as the nucleophiles, in particular, the -OH group.
The alcohol attacks the protonated alcohol forming the di-
alkyl ether through a substitution, bimolecular dehydra-
tion. The smaller alcohols such as methanol and ethanol
are known to deprotonate when adsorbed onto CuO (act-
ing as a Brønsted base) but the proton associated with bu-
tanol is less acidic (44), so the possibility of deprotonation
is less likely to occur. A schematic drawing of the proposed
mechanisms is shown in Fig. 6.
FIG. 6. Schematic of the proposed mechanisms.
so there was the possibility of the elimination, unimolecular
dehydration reaction occurring to produce the alkene. The
alkene was not detected; therefore, the secondary reaction
is likely restricted to the bimolecular dehydration.
The catalytic results showed that only two products were
formed, 1-butanol and the dibutyl ether. These catalysts
were moderately acidic as shown by the Hammett titration,
CONCLUSIONS
The EPR data showed that for the dominant magnetic
species there was no significant difference between the
different catalysts. Further analysis of EPR spectral data
showed the presence of a secondary magnetic species which
is thought to be the catalytically active species. This sec-
ondary species appeared to be the more delocalized/mobile
electronic species associated with surface cupric sites.
Since H + on the alumina surface appears necessary for
the reaction, it is not surprising that the low loaded mate-
rials were better catalysts than the high loaded materials in
the production of ether. The important differences shown
are the chagnes in selectivity from the pure copper catalyst
to the mixed Ru–Cu/Al2O3 materials, in which the selec-
tivity is dependent on the preparation method. Here the
TABLE 5
Relationship between the Hammett Acidities and Dibutyl Ether
Catalyst
Al2O3
Hammett acidities
% Dibutyl ethera
ꢄ + 1
ꢄ + 1
ꢄ + 2
ꢄNA
ꢄ + 2
ꢄ + 2
ꢄ + 2
—
—
—
—
18
25
52
Ru/Al2O3
Cu–Al2O3–Ru
b
Cu/Al2O3
Cu/Al2O3
Ru–Al2O3–Cu
Ru–Cu/Al2O3
a Percentage of total products.
b Highly loaded sample.

270
JANSEN ET AL.
selectivity increases for dibutyl ether when ruthenium is 12. Papp, J., Soled, S., Dwight, K., and Wold, A., Chem. Mater. 6, 496
(1994).
present compared to when the catalyst is without ruthe-
nium.
Another important factor is the usable lifetime of these
materials. The lifetime appears to be extended by the ad-
13. Voltz, S. E., Hirschler, A. E., and Smith, A., J. Phys. Chem. 64, 1594
(1960).
14. British patent 718, 723 (1954).
15. Maxwell, I. E., and Naber, J. E., Catal. Lett. 12, 105 (1992).
dition of ruthenium to the Cu/Al2O3 materials. Ruthenium 16. Biswas, J., and Maxwell, I. E., Appl. Catal. 58, 1 (1990).
17. Weitkamp, J., in “American Chemical Society, ACS Symposium Series
20, 1975” (J. W. Ward, Ed.), p. 1–27.
18. Burgt, M. J., Leeuwen, C. J., Dell’Amico, J. J., and Sie, S. T., in
dioxide (RuO2) seems to stabilize the catalyst allowing it
to be more efficient for a significantly longer time than
that seen in the pure copper material. The total percent-
age conversion over a 20-h time interval was much greater.
“Methane Conversion” (D. M. Bibby, C. D. Chang, R. F. Howe, and
S. Yurchak, Eds.), Elsevier Science, Amsterdam, 1988.
This could be due to the ability of ruthenium to transfer 19. Sinfelt, J. H., J. Catal. 29, 308 (1973).
20. Tanabe, K., Sumiyoshi, T., Shibata, K., Kiyoura, T., and Kitagawa, J.,
hydrogen, allowing for the more efficient formation of the
products when compared to the pure copper catalyst. It also
decreases the reductive poisoning of the surface by coke
formation. There is evidence of poisoning in the pure cop-
Bull. Chem. Soc. Jpn. 47, 1064 (1974).
21. Yamaguchi, T., Tanaka, Y., and Tanabe, K., J. Catal. 65, 442 (1980).
22. Shibata, K., Kiyoura, T., Kitagawa, J., Sumiyoshi, T., and Tanabe, K.,
Bull. Chem. Soc. Jpn. 46, 2985 (1973).
per catalyst. Initially the fresh catalyst is a very light color 23. Tanabe, K., Misono, M., Ono, Y., and Hattori, H., “New Solid Acids
and Bases: Their Catalytic Properties.” Elsevier Science, New York,
1989.
24. Tanabe, K., “Solid Acids and Bases.” Kodansha, Tokyo, and Academic
compared to the used catalyst which exhibits a dark brown
color. EPR analysis showed reduced signal intensity as well.
The following reaction mechanisms for these mixed
Press, New York, 1970.
metal catalysts are proposed. the primary reaction is the
reduction/hydrogenation of the butyraldehyde to 1-butanol
and the secondary reaction has been determined to be only
the bimolecular dehydration (substitution) reaction under
acidic conditions. Even though this type of system was not
thought of as an acid catalyst, we have shown in this pa-
per that it can be used as an acid catalyst in some appli-
cations. The results presented on the formation of dibutyl
ether could lead to a new inexpensive process for the pro-
duction of ethers.
25. Jobson, E., Baiker, A., and Wokaun, A., J. Chem. Soc. Faraday Trans.
86(7), 1131 (1990).
26. Kung, M., and Kung, H., Catal. Rev. Sci. Eng. 27, 425 (1985).
27. Peri, J., J. Phys. Chem. 69, 231 (1965).
28. Asakura, C., Bando, K.-K., and Iwasawa, Y., J. Chem. Soc. Faraday
Trans. 86(14), 2645 (1990).
29. Black, J. B., Scott, J. D., Serwicka, E. M., and Goodenough, J. B., J.
Catal. 106, 16 (1987).
30. Jolly, W. L., “Modern Inorganic Chemistry.” McGraw–Hill, New York,
1984.
31. Marques, E. C., Friedman, R. M., and Dahm, D., Appl. Catal. 19, 387
(1985).
32. Pierron, E. D., Rashkin, J. A., and Roth, J. F., J. Catal. 9, 38 (1967).
33. Matsunaga, Y., Bull. Chem. Soc. Jpn. 34, 1291 (1961).
34. Berger, P. A., and Roth, J. F., J. Phys. Chem. 71, 4307 (1967).
35. Fujiwara, S., Katsumata, S., and Seki, T., J. Phys. Chem. 71, 115
(1967).
REFERENCES
1. Crisafulli, C., Galvagno, S., Maggiore, R., Scire, S., and Saeli, A., Catal.
Lett. 6, 77 (1990).
2. Crisafulli, C., Maggiore, R., Schembari, G., Scire, S., and Galvagno, S., 36. Gomez, M., “Analysis of Mixed Metal Materials as Catalysts and
J. Mol. Catal. 50, 67 (1989).
3. Galvagno, S., Crisafulli, C., Maggiore, R., Giannetto, A., and Schwank,
J., J. Therm. Anal. 32, 471 (1987).
4. Niwa, S., Mizukami, F., Kuno, M., Takeshita, K., Nakamura, H.,
Tsuchiya, T., Shimizu, K., and Imamura, J., J. Mol. Catal. 34, 247
(1986).
5. Enomoto, T., Okuhara, T., and Misono, M., Bull. Chem. Soc. Jpn. 58,
1490 (1985).
6. Christmann, K., Ertl, G., and Shimizu, H., J. Catal. 61, 397 (1980).
7. Friedman, R. M., Freeman, J. J., and Lytle, F. W., J. Catal. 55, 10
(1978).
8. Sinfelt, J. H., Rev. Mod. Phys. 51, 569 (1979).
9. Voge, H. H., and Adams, C. R., Adv. Catal. 17, 151 (1967).
10. Ponec, V., and Sachtler, N. M. H., J. Catal. 28, 376 (1973).
11. Wade, L. G., Jr., “Organic Chemistry.” Prentice Hall, New Jersey,
1995.
Electronic Ceramics,” Master thesis, Temple University, Philadelphia,
1993.
37. Palmieri, M., Gomez, M. L., Vadel, N., Singh, D., and Jansen, S. A.,
“Materials Research Society, Symposim Proceedings, Boston, 1993,”
Vol. 323, p. 359. 1993.
38. Maki, A. H., and McGarvey, B. R., J. Chem. Phys. 29, 31 (1958).
39. Nilges, M. J., “Electron Paramagnetic Resonance Studies of Low
Symmetry Nickel(II) and Molybdenum(V) Complexes,” Ph.D. thesis,
University of Illinois, Urbana, 1974.
40. Kivelson, D., and Nerman, R., J. Chem. Phys. 35, 149 (1961).
41. Naccache, C., and Taarit, Y. B., Chem. Phys. Lett. 11, 11 (1971).
42. Niwa, M., Katsuhiro, S., Kishida, M., and Murakami, Y., Appl. Catal.
67, 297–305 (1991).
43. Sinfelt, J. H., Lan, Y. L., Cusumano, J. A., and Barnett, A. E., J. Catal.
42, 227 (1977).
44. McMurray, J., “Organic Chemistry.” Brooks/Cole, CA, 1984.