THE CATALYTIC DECOMPOSITION OF METHANOL INTO SYNGAS FOR USE AS AN AUTOMOTIVE FUEL

David T. Wickham, Boyce W. Logsdon, and Scott W. Cowley

Department of Chemistry and Geochemistry

Colorado School o f Mines, Golden, CO 80401.

Abstract

Although methanol is thought to be an excellent automotive fuel, it has a smaller volumetric fuel value than gasoline. The catalytic decomposition of methanol into syngas, prior to its combustion in the engine, improves its fuel value by approximately 14%. Palladium and platinum on modified alumina supports demonstrate the necessary qualities for this process. The catalytic activity and thermal stability of palladium is strongly affected by the nature of the catalyst support used. A gamma-alumina support was modified with the oxides of Li, Mg, Cs and La. The effect of the modified support on the activity, selectivity and thermal stability of the palladium metal was studied during exposure to a thermal cycle of 300, 500 and 300℃. The profound difference in catalyst behavior may be due to strong metal-support interactions, to variations in metal dispersion, or to chemical alteration of the palladium. In order to determine the chemical state of the palladium metal and the modified support before and after testing, the catalysts were characterized by TPD, XRD, XPS, and volumetric chemisorptions techniques.

Introduction

Methanol is available from renewable sources, such as biomass, or from non-petroleum sources, such as coal. For this reason  it is considered an attractive alternative to gasoline as an automotive fuel. However, it suffers the disadvantage of having a lower enthalpy content than gasoline, which translates into less mileage per gallon of fuel⁽¹⁾. The potential fuel value of methanol can be improved by catalytically decomposing it into syngas prior to it’s combustion in an internal combustion engine. Syngas, which is also referred to as dissociated methanol, has approximately a 14% higher enthalpy content than methanol. The combustion enthalpies for methanol. Syngas (2CO + 4 H₂), and dimethyl ether ( a side product of methanol decomposition) is given in Table 1. A Chevrolet Citation and a Ford Escort have been retrofitted with catalytic converters in order to demonstrate the feasibility of this process. A schematic of the automotive system is shown in Figure 1. The design and testing of this system was carried out at the Solar Energy Research Institute(SERI). A more detailed description of .the automotive system employed in these studies is available in the literature⁽²⁾.

At the time this study was initiated, no catalyst had been designed specifically for the decomposition of methanol into syngas in an automotive converter. I n order to find an effective catalyst for this process, the following catalyst requirements were defined. First, the catalyst must demonstrate good selectivity for syngas over a large temperature range. Second, the catalyst should be able to withstand high temperatures without deactivating. Third, the catalyst should show significant activity at low temperatures. Forth, the catalyst should be mechanically strong. Fifth, the catalyst cost should be reasonable. Several catalysts were prepared by placing active metals (Cu, Ni, Pd, and Pd) on a variety, of supports (magnesia, silica, alumina) were tested and the results were compared with those obtained from several commercial formulations⁽³⁾. All of the commercial formulations failed either the first, second o r fourth requirements or a combination of those requirements. Palladium and platinum were found to be the best metals for the decomposition of methanol and alumina was found to be the best support, however, the alumina support is also active for the dehydration of methanol into dimethyl ether, which is an undesirable side reaction. Therefore, several catalysts were prepared in our laboratory in which active metals were impregnated onto a variety of modified gamma-alumina supports. The modified supports have a marked effect on the activity, selectivity, and thermal stability of some metals and little effect on others.

The objective of this study is to establish the function of these modified supports in controlling the behavior of the active metal. Many of the initial test and characterization results have been reported in the literature^(3.4). However, some of the more interesting catalysts have been studied and characterized in more detail and are discussed below.

Experimental

The catalysts were tested under the same temperature and flow conditions that would be experienced in the automotive system. The tests were conducted in a microcatalytic plug-flow reactor. Methanol was pumped into a heated inlet and vaporized. The catalyst was mounted on a fritted disk in a quartz tube. The methanol vapor was passed over the catalyst bed and the decomposition products were introduced by a gas

sampling valve into a gas chromatograph. The gas chromatograph was equipped with a thermal conductivity detector and a 9 ft. x 1/8 inch stainless steel column packed with Poropack Q. Hydrogen analysis was accomplished by using a 5% hydrogen in

helium carrier gas and a subambient program.

The modified supports were prepared by impregnating a gamma-alumina support with a solution containing the nitrate salt of the desired modifying agent. Subsequently, the modified supports were dried at 150℃, calcined at 550℃, and impregnated with a solution containing either the nitrate or chloride salts of palladium or platinum. The during and calcining steps were repeated for the finished catalyst, the catalysts were activated in the reactor under a flow of hydrogen at 300 ℃ for 1 hour and at 400 ℃ for 1hour prior to testing.

Approximately 0.4 grams o f 14-20 mesh catalyst particals were tested in each run , using a methanol flow rate of 0.19 g MeOH/g cat . - hr . The methanol was distilled over magnesium turnings and sroted over 5A molecular sieve before use. Ultra high purity hydrogen was passed over a Matheson Model 450 purifer.

The total catalyst surface areas were obtained using conventional BET methods. The active metal surface areas were obtained with a volumetric apparatus using hydrogen chemisorption, carbon monoxide chemisorption, and hydrogen-oxygen titration techniques.

For the temperature programmed desorption studies, 0.4 grams of the catalyst were reduced in hydrogen at 400 or 500℃ for 2 hours, the hydrogen was stripped from the catalyst surface in a helium for 15 min. and then cooled in helium to 0 ℃. Subsequently, several pulses of carbon monoxide were introduced intil the surface was saturated. The temperature desorption profile was obtained by increasing the temperature at a rate of 25℃/min. and analyzing the desorption products in a quadrapole mass spectrometer. The temperature programmed desorption apparatus is very similar to that used by other researchers⁽⁵⁾. All the gassed used during the characterization studies were of Ultra High Purity grade and they were further purified to remove traced of oxygen, water and other impurities.

The catalysts were also characterized before and after testing using a Rigaku x-ray diffractometer, a Perkin Elmer 5000 atomic absorption spectrometer, and a Surface Science SSX-100 x-ray photoelectron spectrometer.

Results and discussion

Catalyst testing

All catalysts were tested at 300 ℃, the average operating temperature of the converter, and at 500℃ to simulate a high temperature excursion. Subsequently, the temperature was lowered to 300℃ to check for possible catalyst deactivation. The results of these tests are given in Table II. Each catalyst contains approximately 5 wt.% of the metal oxide modifying agent and a 0.5 wt.% Pd or Pt. the modifying agents Li₂O, MgO, Cs₂O, and La₂O₃ are represented in Table II as Li, Mg and La. The gamma-Al₂O₃ support is represented as Al. the products observed in these tests were hydrogen carbon monoxide, methane, carbon dioxide, water, dimethyl ether, and unconverted methanol. The mole % products in Table II are reported on a hydrogen free basis.

To begin with we will only consider the initial activity of the catalysts at 3000℃ prior to a high temperature excursion at 550℃. The alumina support (Al) shows almost no activity for the formation of the desired CO decomposition product, but is very active for the formation of dimethyl ether. The formation of dimethyl ether is an exothermic reaction and should be avoided. In the case of the Pd-A1 and Pt-A1 catalysts, it is obvious that the metal plays the primary role in the production of CO, however, the dehydration activity of the support is still apparent. For those catalysts using the modified supports, Pd-Li-Al, Pd-Mg-Al, Pd-La-Al, Pt-Li-Al, Pt-Mg-Al, and Pt-La-A1 the dehydration activity of the catalysts has been eliminated or reduced to an acceptable level. It is interesting to note that the modified supports have a pronounced effect on the initial activity o f the Pd metal as indicated by the production of CO, with the Pd-Li-A1 showing the lowest activity and the Pd-La-A1 showing the highest activity. On the other hand, the initial activity of the Pt metal does not appear to be influenced by the nature of the modifier used.

At 550°C, the modified supports have a pronounced effect n the catalyst selectivity. All the Pt catalysts and the Pd-La-A1 catalyst produce substantial amounts of methane, carbon dioxide, and water, while the Pd-A1, Pd-Li-Al, and d-Mg-A1 catalysts are quite selective for the desired CO product. These results suggest that the latter catalysts have thermally deactivated and the resultant activity and selectivity is really due to the support and not the metal.

The final activity at 300°C, after testing at 550°C, shows all of the Pd catalysts (with exception of the 3% Pd catalyst) have thermally deactivated, while the Pt catalysts show no deactivation or an increase in activity.

The catalyst testing was extended to modified catalysts containing approximately 3 wt% Pd, and the results are shown in Table III. In these tests the hydrogen analysis was included. The effects of the modified supports on the initial and final activities of the catalysts is obvious, with the Pd-La-A1 catalyst representing the best case and the Pd-Li-A1 catalyst the worst case. The Pd-La-A1 catalyst appears to experience a slight loss in activity after a thermal cycle. However, exposing the Pd-La-A1 catalyst to repeated thermal cycling over a three day period resulted in no further deactivation. This catalyst was selected for use in the “dissociated methanol car” and has fulfilled all the catalyst requirements with exception of the cost of the material. In any case, it has allowed the feasibility of the process to be demonstrated in an automobile.

Catalyst Characterization

The modified supports have a marked influence on the catalyst activity, selectivity and thermal stability. Modified alumina supports may have an important impact on the behavior of metals other than Pd and Pt. The effects of these modified supports may extend to other important chemical reactions which involve hydrogenolysis, dissociation, or hydrogenation steps, such as, Fischer Tropsch, and methanol synthesis. In order to understand the role of the modified support in controlling the behavior of the active metal component, selected catalysts were characterized using both bulk and surface techniques. Since the Pd-Li-Al and the Pd-La-Al catalysts represent the most extreme cases in activity, selectivity and themal stability, they were selected for futher detailed studies.

The complete thermal deactivation of the Pd-Li-Al catalyst could result from one of the following processes:

1) Loss of total surface area due to support sinteing .

2 ) Loss of total surface area due to pore blockage.

3) Loss of the active Pd metal from the catalyst .

4 ) Sintering of the metal crystallites.

5) Migration of the metal into the support .

6) Covering the active metal with an inactive material .

7) formation for a new inactive Pd compound.

The first two possibilities were eliminated by measuring the total surface area of the fresh and reduced Pd-Li-Al catalyst using the T\BET method. There was no significant change in the total surface area and the results are given in Table IV.

The metal content of the fresh, reduced and tested catalysts were determined using atomic adsorption and the results ate also given in Table IV. It is apparent that the active metal does indeed remain with the catalyst.

Broadening of the Pd lines in x-ray diffraction would suggest sintering of the m etal crystallites was involved. The XRD spectra for both the reduced and tested Pd-Li-Al catalyst is shown in Figure 2. There is no evidence of line broadening, however, all the palladium lined of the tested catalyst haven shifted to lower 2θ values, this suggests that wither a new compound as been formed or the Pd lattice has expanded, possibly due to the palladium forming a solution with another element. These results would also arque against the migration of the Pd metal into the support as a cause for the themal deactivation.

The catalyst deactivation was also found to be reversible, as shown in Figure 3. Although the catalyst was deactivated at high temperatures, its activity could be restored by recalcining the catalyst in air and reducing in hydrogen. The catalyst demonstrated an activity almost identical to its initial activity. However, the catalyst was again deactivated after a thermal cycle. This result strongly arques against migration of the Pd metal into the support , which should be an irreversible process. It also suggests that the element with oxygen, an example of such an element would be carbon.

XPS analysis of the fresh, reduced, and tested Pd-Li-Al and Pd-La-Al catalysts is shown in Figure 4. These survey scans provided information about the elemental composition of the catalyst surfaces before and after testing. The peaks of most interest are the Pd 3d peaks. For the Pd-La-Al catalyst the Pd is present at the surface of the fresh, reduced, an tested catalyst. In fact, the Pd signal is enhanced after testing. On the other hand, the Pd-Li-Al catalyst shows that the Pd signal has almost disappeared on the tested catalyst. This would suggest that the catalyst is being covered by some other element. The carbon signal, just to the right of the Pd peak, appears to increase on the Pd-Li-Al catalyst but nor for the Pd-La-Al catalyst. A depth profile of the surface was obtained by sputtering the surface with argon ions. The results are shown in Figures 6 and 7. The Pd signal for the Pd-Li-Al catalyst increased as the surface was sputtered away, while the carbon signal disappeared at the same time, the Pd-La-Al catalyst showed little effect with sputtering of the surface.

Finally, temperature program desorption/reaction of carbon monoxide from the surface was conducted to study the ability of the two catalysts to disproportionate carbon monoxide into carbon dioxide and carbon. The disproportionation reaction may be the primary source of the carbon for the deactivation of the Pd-Li-Al catalyst, see Figure 5. It is apparent that much more carbon dioxide of produced over the Pd-Li-Al catalyst and the reaction takes place at a much lower temperature. This suggests that the modified supports have a pronounced effect on the surface activity of the Pd metal.

Summary

The activity, selectivity, and thermal stability of Pd metal is strongly influenced by the use of modified supports. This suggests that either the modified support is influencing the behavior of the metal by either controlling the dispersion of the metal crystallites or by exhibiting a strong metal-support interaction.

REFERENCES

1) J. Finegold , J.T. McKinnon, and M. Karpuk, “Reformed Methanol”, Non-petroleum Fuel Symposium 111, October1980.

2 ) J. Finegold , G.P. Glinsky , and G.E. Voecks, " Dissociated Methanol Citation: Final Report", SERI/TR-235-2083, August 1984.

3 ) S.W. Cowley and S.C. Gebhard, “The Catalytic Decomposition of Methanol into Synthesis Gas for Use as a n Automotive Fuel”, CSM Quarterly , 1( 3 ) , 41(1983).

4 ) S.C. Gebhard, B.W. Logsdon, and S.W. Cowley, “The Decomposition of Methanol Over Supported Palladium and Platinum Catalysts”, manuscript in preparation .

5) J.M. Zowtiak and C.H. Bartholomew, J. Catal., 83, 107(1983.

Table I Combustion Enthalpies for Methanol, Dimethyl Ether, Methane, and Synthesis Gas

Figure 1 Automotive system for methanol decomposition. The unhatched pathway delineates the transport of fuel toward and into the engine; the hatched pathway illustrates the discharge of exhaust gases from the engine.

Table II Methanol Decomposition Over Pd and Pt Catalysts

*catalyst Wt.=0.400g, approx. 5% metal oxide modifier present

 Space Velocity=1.9g MeOH/g cat.-hr               

 Pressure=608.0 mm Hg

** After testing at 500℃

Figure 2 The XRD sprcta of a reduced(top) and a tested(bottom) Pd/alumina catalyst  modified with lithia.