196 Ind. Eng. Chem. Prod. Res. Dev.

1904, 23, 196-202

growth of the smallest particles. The result is a size dis- Registry No. FezO3, 1309-37-1; Fe304,1317-61-9; Crz03,
tribution that produces a much smaller average particle 1308-38-9;potassium, 7440-09-7; ethylbenzene,100-41-4; styrene,
radius yielding a higher surface area and more even dis- 100-42-5.
tribution. Literature Cited
Since the center of the pellets has the greatest concen-
Courty, P.; LePage, J. F. "Preparation of Catalysts 11: Scientific Bases for
tration of agglomerates, it will be at a lower temperature the Preparation of Heterogeneous Catalysts: Proceedings of the Second
than the low-concentrationintermediate region during the International Symposium, Louvain-La-Neuve, Sept 4-7, 1978"; Eisevier:
New York, 1979; pp 293-305.
steam treatment. Since the equilibrium partial pressure Davles, E. P.; Eggerson, F. T. (to Shell Development Co.) U S . Patent
of the KOH is larger at lower temperature for this exo- 2461 147, 1949a.
thermic reaction, the transport will occur from the center Davies, E. P.; Eggerson. F. T. (to Shell Development Co.) U.S. Patent
2460311, 1949b.
of the pellets to the intermediate region, thus effecting Gutzeii, C. L. (to Shell Development Co.) US. Patent 2408 140, 1946.
redispersion. Gutzeil, C. L. (to Shell Development Co.) U.S. Patent 2 449 295, 1948.
The movement of promoter along the catalyst bed from Habeshaw, J. "The Manufacture of Styrene Monomer", I n "Benzene and Its
Industrial Derivatives", Hancock, E. G., Ed.; Halstead Press: New York,
a region of bed having higher promoter concentration to 1975.
a region with a lower value will also occur due to a partial Hackerman, N.; Lee, E. H. J. Phys. Chem. 1963, 67, 947.
Harrison, D. P.; Hall, J. W.; Rase, H. F. Ind. Eng. Chem. 1965, 57(1), 18.
pressure gradient of the mobile species produced in much Hattori, T.; Murakami, Y.; Masazumi, 1.; Uchida. H. Kogyo Kagaku Zasshi
the same manner as described for the individual pellets. I Q M , 72(10), 2188-2194.
In the case of the bed, the high concentration region Kearby, K. K. (to Standard Oil Development Company) U.S. Patent 2426829,
Sept 2, 1947.
corresponded to particles with large agglomerates. Kearby. K. K. "Catalytic Dehydrogenation", I n "Catalysis", Vol. I11 Emmett,
P. H., Ed.; Reinhold New York, 1955.
Conclusions Kearby, K. K.; Thom, J. P.; Hinlicky, J. A. (to Exxon Research and Engineer-
The steam treatment of the used catalyst increased the ing Co.) US. Patent 3 314 732, May 26, 1984.
activity and selectivity of the catalyst by two means. Most Lee, E. H.; Holmes, L. H., Jr. J. Phys. Chem. 1963, 6 7 , 947.
Lee, E. H. J. Catal. 1966, 6 , 137-139.
of the increase in the conversion to styrene was attributed Lee, E. H. (to Monsanto Co.) US. Patent 3306942, Aug 15, 1968a.
to coke removal while the increase in selectivity was caused Lee, E. H. (to Monsanto Co.) U.S. Patent 3 387 053, June 4, 1968b.
Lee, E. H. Catal. Rev. 1973, 8(2), 285-305.
by the redistribution of the potassium promoter in the Mulier, J. C.; Gilbert, R . J . Chim. Phys. 1969, 66, 348.
form of potassium hydroxide. Although chemical vapor Pitzer, E. W. (to Phillips Petroleum Co.) U S . Patent 3300942, Aug 14, 1963.
transport appears to be a reasonable model for explaining Sanderson, R. T. "Chemical Periodicity"; Reinhold: New York, 1960,
Schafer, H. "Chemical Transport Reactions"; Academic Press: New York,
redispersion, a priori selection of the ideal gaseous reactant 1964.
and conditions may be difficult because of the inability Shibata, K.; Kiyoura, T. Bull. Chem. Soc. Jpn. 1989, 42(4), 871-874.
Soderquist, F. J.; Wazbinski, T. T.; WaMman, N. (to Dow Chemical Co.); U S .
to predict partial pressures and other thermodynamically Patent 3 907 916, Sept 23, 1975.
related properties of microscopic and submicroscopic Vljh, A. K. J. Chim. Phys. 1975, 72(1). 5-8.
systems. But the task can be more orderly and rational
by applying principles of equilibrium thermodynamics and Received for review September 19, 1983
adiabatic reaction temperatures. Accepted October 31, 1983

Reactions of Ethanol and Acetaldehyde over Noble Metal and Metal

Oxide Catalysts

Robert W. McCabe" and Patrlcla J. Mltchell

Physical Chemistry Department, General Motors Research Laboratories, Warren, Michigan 48090

Laboratory catalyst studies, utiliiing simplified feedstreams simulating exhaust from ethanol-fueled vehicles, were
undertaken to evaluate the effectiveness of catalytic control of aldehyde and unburned ethanol emissions. Two
classes of catalysts were identified-those that promoted ethanol oxidation to acetaldehyde and COPand those
that promoted ethanol dehydration to diethyl ether and ethylene. Characterization studies of similar catalysts,
reported in the literature, suggest that ethanol oxidation is promoted by basic catalysts while ethanol dehydration
Is promoted by acidic catalysts. A commercial hopcalite catalyst and 0.1 wt % Pi on Al2O3 were the most active
catalysts for ethanol and acetaldehyde oxidation. However, the hopcalite was Irreversibly deactivated by heating
in an oxidizing feed at 775 K. Detailed studies with Pt/AI,O, catalysts showed that ethanol oxidation follows a
series-parallel mechanism and that conversions are influenced by rates of intrapellet mass transfer.

Introduction catalysts containing 4 wt % Cu/2 wt % Cr, 4 wt % Mn,

This study was undertaken as part of a program to and 0.1 wt % Pt, all supported on alumina. Although
identify catalysts for potential application to ethanol-fueled complete oxidation of unburned ethanol and aldehydes is
vehicles in Brazil for the control of exhaust emissions of desired, those catalysts promoted the partial oxidation of
unburned ethanol and aldehydes (principally form- ethanol to acetaldehyde to various extents.
aldehyde and acetaldehyde). In a previous study (McCabe The present study represents an attempt to identify
and Mitchell, 1983), we evaluated the effectiveness of catalysts with greater activity and selectivity for ethanol
0196-4321/84/1223-0196$01.50/0 0 1984 American Chemical Society
 Ind. Eng. Chem. Prod. Res. Dev., Vol. 23, No. 2, 1984 197

Table I. Catalysts and Supports

description source metal source solvent comments
3' Grace Chemical Co. 110 mz g - ' BET surface area;
3 mm diam spherical
pellets
ZrO, (98%)-A1,0, (2%) Alfa Products 41 mz g-' BET surface area;
3 mm diam x 3.5 mm long
extrudate
Alfa Products 18 m2 g-' surface area; 3 mm
diam x 3.5 mm long
extrudate
4 wt % Cu-2 wt % Cr on Al,O, Laboratory
3.8 wt % Cu-1.4 wt % Cr Laboratory
o n ZrO, (98%)
1.7 wt % Cu-0.4 wt % Cr Laboratory poor penetration of metal
on MgO salts into catalyst pores
during impregnation
0.1 wt % Pt o n A1,03 Laboratory
2 wt % Ag o n A1,03 Laboratory
10 wt % W o n Al,O, Laboratory
10 wt % W o n ZrO, (98%) Laboratory
4 wt % Fe-4 wt % Ce on A1,0, Laboratory
4 wt % Mn o n A1,0, Laboratory
4 wt % Sn on A1,0, Laboratory
hopcalite (80% Mn0,-20% Mine Safety Appl. co. unsupported pellets of
CUO) irregular size and shape
0.083 i 0.005 wt % Pt o n Laboratory Pt(NH,)4Cl,-H,0 CO chemisorption capacity
Al, 0, (uniformly of 3.1 pmol/g of catalyst
impregnated)"
0.088 f 0.005 wt % Pt on Laboratory H,PtCl, CO chemisorption capacity
A1,0, (surface- of 3.8 bmol/g of catalyst;
impregnated) Pt confined t o 103-pm
band at outer edge of
alumina pellet (band width
defined as the depth which
contains 90% of the P t )
Used in tests t o determine the effects of differences in noble metal impregnation profile on product selectivity and
catalyst activity in ethanol oxidation.
and acetaldehyde oxidation to C02 than those studied lyzers. Details of the analytical procedure were reported
previously. To this end, a survey of catalysts was under- previously (McCabe and Mitchell, 1983). In order to
taken utilizing a laboratory flow reactor to characterize prevent losses of ethanol and acetaldehyde by either con-
catalyst activities and selectivities under simulated exhaust densation or reaction at the walls of the transfer lines
conditions. leading from the reactor to the gas chromatograph, TFE
For the most part, catalysts were selected which had tubing was used and a box furnace was placed between the
been shown in this laboratory or elsewhere to have good reactor outlet and the inlet to the GC to maintain the
activity and/or selectivity for complete hydrocarbon oxi- transfer lines at 373 K. Ethanol and acetaldehyde con-
dation. Additional catalysts were selected to provide centrations were identical for a feedstream passed to the
diverse catalytic properties in order to identify the types GC either through a reactor bypass line or through the
of reaction behavior which ethanol and acetaldehyde can reactor containing quartz beads in place of catalyst.
display in vehicle exhaust. However, whenever supported catalysts or bare support
pellets were contained in the reactor, apparent conversions
Experimental Section of ethanol and acetaldehyde were observed by GC at low
Reactor System. The experimental apparatus and temperatures although no product species (Le., CO, COz,
methods used in this study were similar to those employed and hydrocarbons) were observed. The apparent con-
in the previous study (McCabe and Mitchell, 1983). The versions resulted from ethanol and acetaldehyde adsorp-
reactor was a 2.5 cm 0.d. quartz tube which was contained tion on the catalysts and supports and decreased from high
in a tube furnace with a single heating zone. A 15-cm3 initial levels to levels between 10 and 20% for acetaldehyde
catalyst charge was used in all experiments. Temperatures and 2-5% for ethanol over most catalysts during 1to 2 h
were measured with a chromel-alumel thermocouple exposure to the standard feeds at room temperature pre-
(contained in a 3 mm diameter stainless steel sheath) ceding each experiment. The presence of adsorbed acet-
positioned a few millimeters below the top of the catalyst aldehyde and ethanol was also evidenced at low temper-
bed (the feed was passed through the catalyst bed in a atures on most catalysts by desorption of large quantities
downflow configuration). The axial temperature gradient of ethanol and acetaldehyde as the catalyst was heated
through the catalyst bed was at most 25 K and the tem- from one steady-state temperature to the next. Since
perature increased monotonically from the top to the overshoot invariably occurred in changing from one
bottom of the bed. The feed was introduced into the steady-state temperature to the next, slow adsorption (and
reactor by the method described previously-ethanol was hence, apparent conversion) of ethanol and acetaldehyde
vaporized in a stream of flowing nitrogen; acetaldehyde, persisted above room temperature.
oxygen, CO, and nitrogen were supplied from gas cylinders. Catalysts. Table I gives a description of catalysts and
Hydrocarbons and oxygenated hydrocarbon species were support materials used in the study. All laboratory-pre-
analyzed by gas chromatography. Concentrations of CO pared catalysts were impregnated by incipient wetness
and C02were measured with nondispersive infrared ana- using the salts and solvents reported in Table I. The
198 Ind. Eng. Chem. Prod. Res. Dev., Vol. 23,No. 2, 1984
Table 11. Comparison of Activities and Selectivities of Catalysts for the Oxidation of Ethanol
_____
major carbon-containing productd temp range of max fractional
max acetaldehyde conv of ethanol
“A”-type cat. T50Wa 375 K 500 K 700 K formation, K to acetaldehyde
hopcalite (CuO-MnO,) 405 CH,CHO CO? co2 400-4 25 0.2 0-0.2 5
0.1% Pt/A120, 405 CH,CHO co, COZ 400-425 0.19-0.21
4% Cu-2% Cr/A1,0, 492 CH,CHO CH,CHO co2 505-515 0.28-0.30
3.8% Cu-1.4% Cr/ZrO,
2% Ag/Al,O,
493
510
CH, CHO
CH,CHO
CH,CHO
CH,CHO
co:
COZ
-500
540-560
-0.32
-0.70
1.7% Cu-0.4% CriMgO 531 CH,CHO co2 575 -0.36
-
C
4% Fe-4% Ce/Al,O, 532 C CH,CHO co2 545 -0.24
4% Mn/AI,O, 543 C CH,CHO CO, 5 7 0-600 0.28-0.32
MgO 680 C CH,CHO CH,CHO a773 -0.48 (at 773 K )
“B”-type cat.
10% W/AI,O, 512 C -0.20
10% W/ZrO,
ZrO,
515
57 5
C - 0.07
C -0.10
. ~.

Al,O, 588 C CH,’cHO 575-6 50 -0.15

“A t B”-type cat.
4% Sn/A120, 57 0 C (CH,CH,),O CH,CHO 650-700 0.45 -
O1 T50% is the temperature at the top of the catalyst bed at which 50% conversion of ethanol was obtained with a feed-

stream containing 0.1% ethanol and 1%0, in nitrogen at a space velocity of 52 000 h-l (STP). Fresh catalyst activity.
2’50% of hopcalite increases with extended exposure to temperatures above 700 K. Negligible conversion of ethanol at
375 K. Based o n comparison of the fraction of ethanol in the feed converted t o the product listed relative t o the
fractions converted to other products.

catalysts were dried in air at 373 K for 12 h and then 1 .o a----- -0 1.0
heated in air for 4 h at temperatures between 675 and 775 .- 2 wt% Ag on A1203
5:
K prior to use in the reactor. This pretreatment resulted 0
in catalyst activity which was reproducible in successive
ethanol and acetaldehyde oxidation experiments at tem-
peratures between 300 and 775 K. The commercial hop-
calite catalyst (Mine Safety Appliance Co.) was tested as
received.
A series of Pt/A120, catalysts were formulated to
evaluate the effects of noble metal impregnation profile
on activity and selectivity for the oxidation of ethanol. A
surface-impregnated catalyst was made using H2PtC1,in
0 1 I 0
acetone. Ninety percent of the Pt was contained in a 103 300 400 500 600 700
pm wide band at the outer edge of the catalyst pellet. A Temperature ( K )
uniformly impregnated catalyst was made by using an Figure 1. Ethanol conversion and yields of carbon-containing
aqueous solution of platinum amine salt. The two catalyst products in the oxidation of ethanol over a 2 wt % Ag on A1203
samples contained 0.084.09 wt % Pt and differed in CO catalyst as a function of catalyst temperature.
chemisorption capacity by about 20% (3.1 pmol/g for the
uniformly impregnated catalyst and 3.8 pmol/g for the tivities. However, for the purpose of identifying exhaust
surface-impregnated catalyst). emission catalysts for ethanol-fueled vehicles, we are in-
terested in overall catalyst performance and therefore have
Results and Discussion compared the catalysts on an equal volume basis as they
Catalyst Comparison for Ethanol Oxidation. Table would be deployed in converters.
I1 contains a comparison of catalyst activities and selec- Table I1 contains a listing of the major carbon-containing
tivities for the steady-state oxidation of ethanol in a feed product at 375,500, and 700 K for each catalyst. Over the
containing 0.1 vol % ethanol and 1 vol % O2 in N2 at a group A catalysts, acetaldehyde and COz were the major
space velocity of 52000 (vol feed) (vol cat.)-’ h-’ (STP) products, and acetaldehyde was formed at lower temper-
(hereafter designated simply as h-l). The catalysts are atures (i.e., lower ethanol conversions) than COz. As an
divided into two major groups: (1) those that promoted example, Figure 1 shows the conversion of ethanol and
ethanol oxidation to acetaldehyde and C02nearly exclu- yields of carbon-containingproducts for the 2 wt % Ag on
sively (group A) and (2) those that promoted ethanol Al2O, catalyst as a function of temperature. Acetaldehyde
dehydration to diethyl ether and ethylene nearly exclu- production reached a maximum near 550 K and decreased
sively (group B). In addition, we examined a 4 wt % Sn sharply at higher temperatures as COz production in-
on A120, catalyst which showed a combination of group creased. The maximum in acetaldehyde yield was char-
A and group B behavior. acteristic of all the group A catalysts except MgO, which
Within the two groupings, catalysts are listed in order showed a continuous increase in acetaldehyde yield up to
of decreasing activity based on temperatures of 50% eth- 775 K, the highest temperature of the study.
anol conversion. We note that the physical properties of Although the group A catalysts all showed qualitatively
these catalysts (i.e., metal loading and dispersion, metal similar behavior, acetaldehyde yields and temperatures of
impregnation profile, bulk density, and porosity of the maximum acetaldehyde yield varied widely as shown in
supports, etc.) are diverse and the activities and selectiv- Table 11. Yields of acetaldehyde were greatest over the
ities reported in Table I1 undoubtedly reflect differences Ag catalyst and lowest over the hopcalite and Pt Catalysts.
in their physical properties as well as their intrinsic ac- Temperatures of maximum acetaldehyde production were
 Ind. Eng. Chem. Prod. Res. Dev., Vol. 23, No. 2, 1984 199

1 .o p'.-u 1.0
4 wt% S n on A1203 /

,,
I
. 52.000 h.' Space Velocity
0 1 vol% Ethanol /
1 vol% Oxygen / .-1
G
0.6
Conversion I
$2
r;
0.4
.-c5 g
E
0.2

OL 0
300 400 500 600 700
Temperature ( K ) Temperature (K)
Figure 2. Ethanol conversion and yields of carbon-containing Figure 3. Ethanol conversion and yields of carbon-containing
producta in the reaction of a feed containing 0.11 vol % ethanol and products in the reaction of a feed containing 0.1 vol % ethanol and
1 vol % oxygen over a 10 w t % W on A1203catalyst aa a function of 1 vol ?& oxygen over a 4 wt % Sn on A1203 catalyst as a function of
catalyst temperature. catalyst temperature.

also lowest over the hopcalite and Pt catalysts. Thus, the of ethanol on the supports. Table I1 shows that for the
hopcalite and Pt catalysts showed the greatest ethanol group A Cu-Cr catalysts supported on A1203,ZrG2,and
oxidation activity and selectivity for C02 of all the group MgO, the effect of the support was small; maximum ac-
A catalysts. Acetaldehyde production over those catalysts etaldehyde yields ranged from -0.29 to -0.36. In con-
was limited to temperatures below 575 K which would be trast, the group B W/A1203and W/Zr02 catalysts showed
encountered only during warmup of a catalytic converter. roughly a threefold difference in maximum acetaldehyde
However, we note that ethanol-fueled vehicles emit large yield. The lower maximum yield of acetaldehyde observed
quantities of unburned fuel during cold-starting and over W/Zr02 (0.07) than W/A1203(0.20) may result from
warmup (Chui et al., 1979; Bechtold and Pullman, 1980; the lower selectivity of bare ZrOz than bare Al2O3for ox-
Bailey and Edwards, 1980) and the hopcalite and Pt/Al2O3 idizing ethanol to acetaldehyde (see Table 11).
catalysts could generate significant acetaldehyde emissions Correlation of Ethanol Reaction Selectivity with
during that period. Catalyst Acid-Base Characteristics. Alcohols can react
Diethyl ether and ethylene were the major carbon-con- either as acids or bases in solution phase chemistry de-
taining products over the group B catalysts at most tem- pending on the acid-base characteristics of the reaction
peratures as shown in Table 11. Acetaldehyde was also medium (Morrison and Boyd, 1966). We have demon-
produced over the group B catalysts but in smaller yields strated, in the previous section, that vapor phase oxidation
than generally observed over the group A catalysts. As an of ethanol over a variety of catalysts under identical re-
example, we show ethanol conversion and product yields action conditions can lead to large differences in conver-
in Figure 2 for 10 wt % W on A1203as a function of tem- sions and products. These differences suggest that acid-
perature. At low ethanol conversions diethyl ether was the base properties of the catalysts play an important role in
principal product and its yield reached a maximum near determining activity and selectivity. Catalyst acid-base
500 K as the yield of ethylene increased sharply. The properties were not examined in this study. However, the
ethylene yield decreased above 600 K where the yields of literature contains many examples of catalyst acid-base
acetaldehyde, CO, and C02increased. Similar results were characterization which provide a basis for correlating the
obtained over the other group B catalysts although yields behavior of our A- and B-type catalysts with known
of diethyl ether were lower. The group B catalysts, al- acid-base characteristics of similar materials.
though showing lower acetaldehyde yields than the group As noted previously, the group A catalysts produced
A catalysts, would not make effective emissions catalysts acetaldehyde at low ethanol conversions and C02 at high
for a number of reasons: (1)activity is low compared to ethanol conversions. The oxidative dehydrogenation of
the most active group A catalysts (which indicates that the ethanol to acetaldehydeis an acid reaction and occurs most
catalysts would be ineffective during warmup where eth- readily over catalysts containing strong base sites. Deh-
anol emissions are high); (2) organic products are produced ydrogenation of 2-propanol has been reported to occur at
under all reaction conditions; and (3) acetaldehyde pro- 02-or HOO- sites on MnOz (Hasegawa et al., 1977). On
duction, although small, extends to temperatures charac- MgO, active sites for 2-propanol dehydrogenation have
teristic of the warmed up catalytic converter. been attributed to 02-(Szabo et al., 1975). Ethanol oxi-
Figure 3 shows ethanol conversion and yields over the dation was examined by Takezawa et al. (1980) on MgO
4 wt % Sn on A1203catalyst. The Sn catalyst produced using infrared spectroscopy to identify adsorbed reaction
both oxidation products and dehydration products. intermediates. At 563 K, acetaldehyde and COz were
Maxima were observed in both acetaldehyde yield (char- major carbon-containing reaction products in agreement
acteristic of group A behavior) and diethyl ether yield with our data. Production of both acetaldehyde and C02
(characteristic of group B behavior). T o a lesser extent, was found to correlate with the surface concentration of
the bare alumina and zirconia supports also promoted the an ethoxide intermediate. We conclude, therefore, that
oxidation of ethanol to acetaldehyde and C02 in addition the ethanol oxidation behavior of the group A catalysts is
to ethanol dehydration to diethyl ether and ethylene. For consistent with the presence of strong base sites such as
example, Table I1 shows that acetaldehyde was the major 02-,02-, or HOO-, which promote the dehydrogenation of
product over the alumina at 500 K, where ethanol con- ethanol.
version was low, and that at most 15% of the ethanol was The group B catalysts promoted base reactions of
converted to acetaldehyde between 575 and 650 K. The ethanol-dehydration to diethyl ether and ethylene. Thus,
production of acetaldehyde over all supported catalysts the group B catalysts are expected to demonstrate acid
in this study may have been caused, in part, by oxidation properties. Alumina contains Lewis acid sites (A13+ cat-
200 Ind. Eng. Chem. Prod. Res. Dev., Vol. 23, No. 2, 1984

Table 111. Summary of Catalyst Activities and

Selectivities for Reaction of Acetaldehydea
~ ~~~ ~

T5p carbon-containing
catalyst K reaction products
______ ___ ~ _____
“A”-type catalysts
hopcalite (fresh)c 415
0.1 wt % Pt on predominantly CO,
4 0 3
with small amounts of CO
4 wt 7c c u - 2 wt ’% 510 and CH,
Cr on AI,O,
2 wt % Ag on
A’, 0 3
“B”-type catalysts
10 wt % w on 695 CO and CO, in nearly equal 300 400 500 600 700
A1,0, amounts plus small Temperature ( K )
amounts of ethylene and Figure 4. Conversion of acetaldehyde as a function of temperature
unidentified hydro- over various catalysts for a feed containing 0.025 vol % acetaldehyde
carbons and 1 vol % oxygen at a space velocity of 52000 h-l. Solid curves
a Feedstream containing 0.025 vol 5% acetaldehyde and
indicate acetaldehyde conversions measured by disappearance of
acetaldehyde. Dashed curves indicate acetaldehyde conversions
1 vol R 0, in N , at a space velocity of 52 000 h - ’ (STP).
Temperatures measured by a thermocouple positioned measured by product yields. Differences between solid and dashed
-1 mm below the top of the catalyst bed. Prior t o curves at low temperatures are due to adsorption of acetaldehyde.
heating the catalyst above 500 K . a space velocity of 52000 h-l. Temperatures of 50%
conversion are shown in Table I11 along with a summary
ions) which are involved in reactions of alcohols to ethers of the carbon-containing products over each catalyst. C 0 2
(Jain and Pillai, 1967; Parera and Figoli, 1969; Figueras was the principal carbon-containingproduct over the group
Roca et al., 1968; Knozinger and Stolz, 1970). Alumina also A catalysts; small amounts of CO and CH4 were also
contains Bronsted acid sites which promote dissociation formed. The group B catalyst, 10 wt % W on Al,03,
of the alcohol hydroxyl group during olefin formation produced primarily CO and C02, in nearly equal yields,
(Knozinger, 1972). and also formed small amounts of ethylene and uniden-
Shibata et al. (1973) have shown that the acid amount tified hydrocarbon species.
and acid strength of ZrO, is similar to that of Al,03. Figure 4 shows acetaldehyde conversions as a function
Murrell et al. (1983) measured acid amounts of W03 of temperature for the catalysts in Table 111. As noted in
supported on silica at various loadings and found a one- the Experimental Section, acetaldehyde adsorption on the
to-one correspondence between the number of acid sites alumina support resulted in apparent conversion at low
and W03 content for W03 loadings up to 5 wt %. The acid temperatures. Therefore, we have shown, for each alu-
strengths of the acid sites associated with W03 on silica mina-supported catalyst in Figure 4, both a solid curve,
were much greater than the strength of intrinsic acid sites representing conversions based on the disappearance of
on the silica support. Murrell et al. (1983) also reported acetaldehyde, and a dashed curve, representing conversions
a stronger interaction between W03 and alumina than based on product yields. Product yields remained constant
W03 and silica. They suggested that WO, on alumina a t each steady-state temperature whereas the apparent
forms isolated W03 units that are “locked” into the hy- acetaldehyde conversions decreased slowly as adsorption-
droxylated surface layer of the A1203. Strong acid sites desorption equilibrium was approached. Thus, the dashed
(pK, I -3.0) were reported for a 16% W03-84% A1203 curves more accurately indicate the steady-state acet-
catalyst by Walvekar and Halgeri (1973) using the n-bu- aldehyde conversion than the solid curves in the low-tem-
tylamine titration method. The literature thus provides perature range over each catalyst. In contrast to the alu-
evidence that the group B catalysts contain strong acid mina-supported catalysts, product yields over the unsup-
sites which promote the dehydration of ethanol. ported hopcalite catalyst closely matched acetaldehyde
As a final example of catalyst acid-base characteristics, disappearance, indicating the absence of significant ad-
we note that McAteer (1979) reported both Lewis acid sites sorption.
and basic sites on SnOz by adsorption of pyridine and Catalyst activities for acetaldehyde oxidation (Table 111)
acetic acid, respectively. Our observation of both acid and followed the same order as shown in Table I1 for ethanol
base reactions of ethanol over 4 wt % Sn on A1203thus oxidation; the hopcalite and Pt/A1203catalysts were most
appears consistent with McAteer’s data. active. Moreover, comparison of 50% conversion tem-
The ability to correlate catalyst activity and selectivity peratures for ethanol (Table 11) and acetaldehyde (Table
with acid-base properties should prove useful for identi- 111) shows that ethanol reacted more readily than acet-
fymg promising catalysts for ethanol oxidation. Also, these aldehyde over all of the catalysts. Large amounts of ac-
results suggest that adding acid or base compounds (e.g., etaldehyde desorbed from the alumina-supported catalysts
halide or alkali species) to catalysts may provide a method as they were heated from one steady-state temperature to
of modifying ethanol oxidation activity and selectivity. the next (in the temperature range of low conversions).
Catalyst Comparison for Acetaldehyde Oxidation. Although we have focused on steady-state acetaldehyde
Control of aldehydes, which are irritants (Sittig, 1974) and oxidation in this work, the transient adsorption and de-
photochemically active compounds (Bailey et al., 1978), sorption processes identified here could greatly influence
is a major concern with ethanol-fueled vehicles. Since the emissions characteristics of an ethanol-fueled vehicle
aldehyde emissions are particularly high during engine during warmup of the engine and catalytic converter. In
warmup, catalysts are required which promote acet- particular, the catalyst might store and suddenly release
aldehyde oxidation to C 0 2 at low temperatures. Therefore, large concentrations of ethanol and acetaldehyde during
the temperature dependence of acetaldehyde oxidation was warmup. Additional experiments involving catalyst tem-
measured over some of the catalysts in a feed containing perature-programming and utilizing more realistic feeds
0.025 vol % acetaldehyde, 1 vol % 02, and N, balance at (including H,O) would be required to quantify the effects
 0.8

o,6
1
Hopcalite

Eh
taJ
:+
:/
Conversion
/

9
0.11 vol% Ethanol

/After 7 7 5 K Oxidation
1 .o

0.8

0.6
,E
E
$5
os
::
5
Ind. Eng. Chem. Prod. Res. Dev., Vol. 23, No. 2, 1984

o , 6 ,2.fS
--
Space Velocity x 10-5 (K')

1.f4 0+52 52
,::

CH3CH0,
, 201

'CHJCHO
0.4 'E0 0 CE~OH~
2
LL 01

0.2 cco2
0
,
\I I 1 0
' EtOHO

300 400 500 600 700

Temperature (K)
Figure 5. Conversion of ethanol and yield of acetaldehyde over 0 0.5 1 .o
fresh (solid curves) and thermally treated (dashed curves) samples Fractional Conversion
of hopcalite as a function of catalyst temperature. of Ethanol
Figure 6. Yields of acetaldehyde and C02 in the oxidation of eth-
of adsorption and desorption processes on the transient anol over a 0.1 wt % Pt/AI,OS catalyst at a catalyst temperature of
emissions of ethanol and acetaldehyde. 400 K. The ethanol conversion was varied by changing the space
Deactivation of the Hopcalite Catalyst. The ethanol velocity. The feed contained 0.1 vol % ethanol and 1vol % O2 in
and acetaldehyde oxidation experiments established that all experiments.
a base metal catalyst (hopcalite) can show both fresh ac-
tivity (equal volume basis) and selectivity comparable to markedly from the linear dependence upon ethanol con-
a normally loaded noble metal catalyst (0.1 wt % Pt/ version established at the higher space velocities.
A1203). However, the hopcalite catalyst, which was used The linear relationships between the product yields and
as received from Mine Safety Appliances Co., showed a the conversion of ethanol at the high space velocities are
marked loss in activity after the first ethanol oxidation indicative of irreversible, first-order, parallel reactions for
experiment in which the catalyst was heated to 775 K. the production of C02 and acetaldehyde, i.e.
Figure 5 shows ethanol conversions and acetaldehyde
yields as a function of temperature for both the fresh
hopcalite (solid curves) and the hopcalite after it had been
used to catalyze ethanol oxidation at 775 K for 2 h (dashed k2
curves). The high-temperature oxidation affected the CH3CH20H CH3CHO (2)
activity and selectivity of the catalyst in the subsequent
experiment, causing about a 30 K increase in the 50% where we have shown only the carbon-containing reaction
conversion temperature, an approximate doubling of the species. From the slopes of the lines in Figure 6, taking
maximum yield of acetaldehyde, and a shift in the tem-
perature of maximum acetaldehyde production from -400 -
into account the stoichiometry of the reactions producing
C02and CH3CH0,we find k 2 / k l 5 for ethanol oxidation
at 400 K. The deviation from the linear relationship be-
K to -450 K.
Thermal decomposition of Mn02-CuO catalysts has tween product yields and ethanol conversion at high con-
been analyzed by Kanungo (1979). Compositional water version (low space velocity) is expected and is due to the
(i.e., OH groups) were lost between 473 and 673 K and oxidation of acetaldehyde (produced by reaction 2) to COP.
Mn02 decomposed to Mn203 between 773 and 993 K. This can be seen from Figure 4 which shows the conversion
Similar changes may have occurred in the hopcalite cat- of acetaldehyde over the Pt catalyst during the reaction
alyst of this study during ethanol oxidation at 775 K. of a feedstream containing 0.025 vol % acetaldehyde and
Additional Experiments with Pt/Alz03Catalysts. 1vol 5% oxygen at a space velocity of 52 000 h-l (STP). At
Given the thermal deactivation of the hopcalite catalyst, 400 K, the conversion of acetaldehyde to C02 (dashed line)
0.1 wt % Pt on A1203was the most active catalyst of this is approximately 10%. Thus, in Figure 6, the data ob-
study for oxidizing ethanol and acetaldehyde under con- tained at 52 000 h-l space velocity correspond to roughly
ditions simulating ethanol vehicle exhaust. Therefore, the lowest space velocity (i.e., highest ethanol conversion)
additional experiments were carried out with Pt/A1203 at which acetaldehyde, produced from ethanol oxidation
catalysts to better understand factors controlling ethanol at 400 K, will not be significantly reacted to COP. A more
oxidation selectivity. complete description of the reaction pathways of ethanol
Ethanol was oxidized over the 0.1 wt '3 Pt/A120, cat- over 0.1 wt 5% Pt on alumina is, therefore, given by the
alyst at various space velocities to determine the depen- combined series-parallel mechanism
dence of the product selectivity on the conversion of eth- 0 2 0 2

anol at 400 K. The space velocity was varied between C2H50H(g) (a)r CH,CHO(g) 2C02(g) (3)
26000 h-' and 208000 h-l by changing the amount of
catalyst in the reactor. The results of the experiments at C&WH(g) -% 2COz(g) (4)
the various space velocities are shown in Figure 6 where
both acetaldehyde and C02yields (reactor outlet concen- which we proposed previously (McCabe and Mitchell,
trations normalized against the concentration of ethanol 1983) and which has also been proposed by Ismagilov et
in the feed) are plotted as a function of ethanol conversion. al. (1979).
At ethanol conversionscorresponding to the three highest The identification of a direct oxidation pathway (4) over
space velocities (208000,104 OOO, and 52 000 h-l) both the the Pt catalyst suggests the possibility of modifying the
acetaldehyde and C02 yields increased linearly with the catalyst to increase the selectivity for direct oxidation. One
conversion of ethanol. However, at 26 000 h-' space ve- method of modifying selectivity in reaction networks
locity, the ethanol and acetaldehyde yields deviated containing sequential and/or parallel pathways (with pore
202 Ind. Eng. Chem. Prod. Res. Dev., Vol. 23, No. 2, 1984

100
0 . 0 8 5 t 0 . 0 0 5 wt% Pt on A1203 basis) were commercial hopcalite and 0.1 w t 70Pt/A1203.
However, the hopcalite was irreversibly deactivated at
I-

o a
1-

temperatures reached in exhaust catalytic converters (ca.

Uniform Impregnation
775 K) and, therefore, is not suitable for controlling ex-
Surface Impregnation haust emissions from ethanol-fueled vehicles.
I
(4) Ethanol oxidation over 0.1 wt % Pt/A1203follows
a series-parallel mechanism in which some of the ethanol
is reacted to C02via a gaseous acetaldehyde intermediate
(series mechanism) and some of the ethanol reacts directly
to COz without the formation of gaseous intermediates
300 400 500 600 700 800 (parallel mechanism).
Temperature (K)
(5) There is a pore diffusion limitation to ethanol oxi-
F i g u r e 7. Comparison of ethanol oxidation a c t i v i t y (0) a n d acet- dation over Pt/A1203catalysts demonstrated by differences
aldehyde y i e l d (A)in t h e o x i d a t i o n o f ethanol over surface impreg- in activity of catalysts with nearly equivalent surface area
nated (- - -) and u n i f o r m l y impregnated (-) samples o f 0.085 0.005* but different impregnation profiles. However, only slight
w t % Pt o n A1,03 catalyst as a f u n c t i o n o f catalyst temperature.
differences in selectivity were observed between surface-
diffusion limitations) is to vary the metal impregnation and uniformly impregnated catalysts indicating the ab-
profile (Froment and Bischoff, 1979). To test this, ex- sence of strong diffusion-reaction coupling.
periments were carried out comparing selectivities of Acknowledgment
Pt/A1203catalysts having either uniform or surface im- We wish to thank M. J. D'Aniello Jr., for suggesting
pregnation profiles. Details of the catalyst properties were
preparative methods for the laboratory-prepared catalysts
given in the Experimental Section. and Kim Dang for preparing most of the catalysts. The
Figure 7 compares ethanol conversion and acetaldehyde
surface- and uniformly impregnated Pt catalysts were
yield over the surface- and uniformly impregnated Pt/
prepared and characterized by Joyce E. Carpenter. We
A1203catalysts a t 208 000 h-' space velocity as a function also appreciate helpful discussions with Se H. Oh regarding
of temperature. Ethanol oxidation activity (as measured
the experiments involving the catalysts with different im-
by the temperature of 50% ethanol conversion) was greater pregnation profiles.
for the surface-impregnated catalyst than the uniformly
impregnated catalyst. Thus, a pore diffusion limitation Registry No. ZrO,, 1314-23-4; MgO, 1309-48-4; Cu, 7440-50-8;
to the reaction rate is indicated between 400 and 600 K. Cr, 7440-47-3; Pt, 7440-06-4; Ag, 7440-22-4; W,7440-33-7; Fe,
At higher temperatures the uniformly impregnated catalyst 7439-89-6; Ce, 7440-45-1; Mn, 7439-96-5; Sn, 7440-31-5; C02,
produced slightly greater ethanol conversion than the 124-38-9; CHSCHZOH, 64-17-5; CHSCHO, 75-07-0; (CH&Hz)zO,
surface-impregnated catalyst. Although the reason for the 60-29-7; CzH4,74-85-1; hopcalite, 12612-41-8.
cross-over near 600 K is not known, we note that mass
transfer resistances in the external phase may dominate Literature Cited
Bailey, R. A.; Clark, E. M.; Ferris, J. P.; Krause, S.;Strong, R. L. I n
pore diffusion resistances a t high temperatures. "Chemistry and the Environment"; Academic Press: New York. 1978;
The surface- and uniformly impregnated catalysts also Chapter on "Atmospheric Chemistry".
showed different yields of acetaldehyde as a function of Bailey, W. H.; Edwards, C. F. Proceedings of the Fourth International Sympc-
sium on Alcohol Fuels Technology, Guaruja. Sao Paulo, Brazil, Oct 1980;
temperature. The maximum acetaldehyde yield was nearly Paper B-6 1.
identical but the maximum occurred 40-50 K lower over Bechtoid, R.; Pullman, J. B. Soclety of Automotive Engineers, 1980; Paper
No. 800260.
the surface-impregnated catalyst. Also, more acetaldehyde Chui, G. K.; Anderson, R. D.; Baker, R. E.; Pinto, F. 6. Proceedings of the
was formed a t temperatures above 475 K with the uni- Third International Symposlum on Alcohol Fuel Technology, Asilomar, CA,
formly impregnated catalyst than with the surface-im- May 1979; Paper 11-18,
Figueras Roca, F.; Nohl, A.; de Mourges, L.; Trambouze, Y. C . R . Acad.
pregnated catalyst. Sci., Ser. C . 1988, 266, 1123.
Although overall ethanol oxidation rates (and thus, Froment, G. F.; Bischoff, K. B. "Chemical Reactor Analysis and Design";
Wiiey: New York, 1979; p 214.
ethanol conversions and product yields) were influenced Hasegawa, S.;Yasuda, K.; Mase, T.; Kawaguchi, T. J . Catal. 1977, 46, 125.
to some extent by rates of pore diffusion, coupling between Ismagiiov, Z. R.; Dobrynkin, N. M.; Popovskii, V. V. React. Kinet. Catal.
pore diffusion and surface reaction kinetics did not lead Lett. 1979, 10, 55.
Jain, J. R.: Pilai. C. N. J . Cafal. 1987, 9. 322.
to large differences in acetaldehyde yield for equal con- Kanungo, S. B. J . Catal. 1979, 58, 419.
version of ethanol over the uniformly and surface-im- Knozlnger, H.; Stolz, H. Ber. Bunsenges. Phys. Chem. 1970, 74, 1056.
Knozinger, H.; Buhl, H.; Kochioefl, K. J . Catal. 1972, 2 4 , 57.
pregnated catalysts. These data indicate that significant McAteer, J. C. J . Chem. SOC.,Faraday Trans. 1 1979, 75, 2762.
improvement in the selectivity of ethanol oxidation to C02 McCabe, R. W.; Mitchell, P. J. IndEng. Chem. Prod. Res. Dev. 1983, 22,
cannot be obtained by varying the impregnation profile 212.
Morrison, R. T.; Boyd, R. N. I n "Organic Chemistry", 2nd ed.; Allyn and Ba-
of alumina-supported Pt catalysts. con: Boston, 1966; Chapter 16.
Murrell, L. L.; Grenoble, D.C.; Baker, R. T. K.; Prestridge, E. B.; Fung, S. C.:
Summary Chianelli, R. R.; Cramer, S. P. J . Catal. 1983, 79, 203.
The following conclusions were reached in this study. Parera, J. M.; Figoli, N. S.J . Catal. 1989, 14, 303.
Shibata, K.; Tademitsu, K.; Jun, K.; Takashl, S.; Tanabe, K. Bull. Chem. SOC.
(1) Two types of catalyst behavior were observed for the Jpn. 1973, 46( lo), 2985.
reactions of ethanol in feedstreams containing a large ex- Sittig, M. I n "Pollution Detection and Monitoring Handbook"; Noyes Data
Corporation: Park Ridge, NJ, 1974; Chapter on "Aldehydes".
cess of oxygen-(a) oxidation to acetaldehyde and COz, Szabo. 2 . G.; Jover, B.; Ohmacht, R. J . Catal. 1975, 3 9 , 225.
and (b) dehydration to diethyl ether and ethylene. Takezawa. N.: Hanamaki, C.: Kobavashi, H. J . Res. Inst. Catal. Hokkaido
Univ. 1980, 28(3), 347.
(2) From known acid-base properties of catalysts, those Walvekar, S. P.; Halgeri, A. B. J . Indian Chem. SOC. 1973, L. 387.
which promoted ethanol oxidation have strong base
character, while those which promoted dehydration have Received for review July 5, 1983
strong acid character. Reuised manuscript received D e c e m b e r 2, 1983
(3) The most active catalysts of this study (equal volume Accepted D e c e m b e r 28, 1983