SiO2 catalysts

SiO2 catalysts

Journal of Molecular Catalysis, 59 (1990) 61-73 61 EFFECTS OF ALI(ALINE EARTH CATIONS ON THE MECHANISMS OF CO HYDROGENATION OVER PROMOTED Pd/Si02 CA...

817KB Sizes 1 Downloads 25 Views

Journal of Molecular Catalysis, 59 (1990) 61-73

61

EFFECTS OF ALI(ALINE EARTH CATIONS ON THE MECHANISMS OF CO HYDROGENATION OVER PROMOTED Pd/Si02 CATALYSTS YANFEI SHEN,* J’ITAO LI, LIFEN CAI and KAIHUEI HUANG Department of Chemistry and Institute of Physical Chemistry, Xiamen University, Xiumen (China)

(Received April 4,1989; revised August 22,1989)

Summary The mechanisms of CO hydrogenation over alkaline-earth-promoted Pd/SiOz catalysts have been investigated by means of TPD-MS. The spectra, which were assigned carefully, showed that formate species was the most abundant intermediate on all the catalysts, but its surface states, its concentrations and its stability largely depend on the nature of the promoters. In two main HCOO states, PdHC(0)OM2’ (M y Mg, Ca, Sr and Ba) was found to be much more easily converted to methanol than Pd’HCOOSi, and its concentration was inversely proportional to the turnover frequencies for methanol formation. Thus, the formate species was considered to be the active intermediate, and its conversion to HCO or other species was suggested to be more important for methanol synthesis than its formation. Two reaction routes were claimed to occur simultaneously on two different Pd-containing centers, but one of them may be dominant for a specific catalyst.

Introduction The mechanisms of CO hydrogenation over supported Pd catalysts have been extensively investigated since the early 1980s. Much evidence has been reported, but there still exist some controversies concerning the active intermediate [l-91, the rate-controlling step [l-2, 4, 7, 91, and the active adsorbed CO species [g-11 I for methanol formation and so on. For example, the formate species and its formation were found to be important for methanol formation by Tamaru’s group [1,21, although the formyl (or other species such as HCXOH)) was claimed to be the active intermediate and its formation was thought to be the rate-controlling step by other researchers 13, 71, etc. On the other hand, promoters or supports seem to affect the mechanisms [l-3, 121, but little attention has been paid to this aspect. *Presentaddress: Groupe de Physico-Chimie Min&aIe et CataIyse, Universiti Catholique de Louvain, Place Croix du Sud 1,1348 Louvain-la-Neuve, Belgium. 0304-5102/90/$3.50

@ Elsevier Sequoia/Printed in The Netherlands

62

Recently, we have found [13-X1 that alkaline earth cation-promoted Pd/SiOx catalysts, with almost the same Pd dispersions, can catalyze CO hydrogenation at atmospheric pressure with cu. 100% methanol selectivity and the following activity order: Pd-Mgz+/SiOz > Pd-Ca2’/Si02 > PdMoreover, the hydrogen sorption W’/SiO, > Pd/Si02 > Pd-Ba2+/Si02. capacity of the catalysts is affected as follows: Pd-Mg+/Si02 > PdW’/SiO, > Pd/SiOz > Pd-Ba2+/Si02 > Pd-Ca2+/Si02. Furthermore, the Mgz+, S9’ and Ba2+-promoted catalysts exhibit much more positive Pd valencies than the other two catalysts. Based on the above, the present study is aimed at dealing with the effects of alkaline earth cations on the intermediates and mechanisms of CO hydrogenation over Pd/Si02 catalysts. A TPD-MS technique was used for this purpose. Experimental Catalyst preparation The Pd-M2+/Si02 (M = Mg, Ca, Sr and Ba) and Pd/Si02 catalysts (Pd/M = 1: 1 mole ratio and Pd loading = 5 wt.%) were prepared by impregnating an aqueous solution of a PdCl-MC12 mixture onto pretreated silica gel (1.2 ml g-l and 370 m2 g-l). The impregnated samples were dried at room temperature and at 110 “C for a few hours. TPD-MS measurements A special reactor illustrated in Fig. 1 was designed for catalyst pretreatments or reactions. After a catalyst was pretreated or reacted, it was transferred to the side quartz tube (5), which was connected to the reactor by a Cajon fitting (3) and a Teflon valve (4) in an in situ atmosphere for TPD-MS measurements. The catalyst in the quartz tube was first cooled in a liquid nitrogenethanol trap and then evacuated until the base pressure in the analyzing

Fig. 1. Illustration of the TF’D-MS reactor; (1) reactor for pretreatmenta or reactions; (2) Nupro valves; (3) Cajon valve; (4) Teflon valve; (5) quartz tube for TPD-MS measurements.

63

chamber of a quadrupole mass spectrometer dropped to -2 x 10e5 torr. After the trap was removed slowly to keep the temperature increasing constantly to room temperature, heating was begun at a rate of 10 “C min-‘. Four MS signals could be simultaneously recorded by a data processor. To assign reaction TPD-MS peaks, pure chemicals, such as HCOOH, COP, HCHO and CO, were preadsorbed on the pretreated catalysts and the corresponding TPD-MS spectra were recorded under the same conditions as those used for reaction spectra.

Results and discussion TPD-MS spectra of CO hydrogenation The spectra of CO hydrogenation at 210 “C, 1 atm and Hz/CO = 2 are shown in Figs. 2-6, where four MS signals at m/e = 45, 44, 31 and 29 are recorded. These four signals respectively correspond to the desorbed species of HCOO, COz, CH30 and HCO. It is shown from these figures that COz-desorbed species are the most abundant for all five catalysts, where three kinds of CO,-desorbed peaks are observable in the following temperature ranges: 85-190 “C, 300-320 “C and 400-600 “C, but the peak numbers and desorption temperatures depend on the catalyst used. For example, no CO2 peak at the 85-190 “C range appears for the Pd/SiOa catalyst, but a strong COz peak is measured at about 320 “C. For the other four catalysts, the intensity of the COz peak in the 85-190 “C range decreases in the following order: Pd-Ba2+/SiOz > Pd-S3+/Si02 > Pd-Ca2+/Si02 > Pd-Mg2+/SiO,; moreover, the CO, peak at 300-320 “C is absent on both the Pd-Ca2’/Si02 and the Pd-Ba2’/Si02 catalysts. As for the other three desorbed their desorption temperatures

600

T

1

500

iI ii

Desorption temperaturel”C)

Fig. 2. TPD-MS spectra of CO hydrogenation over the Pd-M$+/SiOz HJCO = 2.

catalyst at 21O”C, 1 atm,

Fig. 3. TPD-MS spectra of CO hydrogenation over Pd-Ca2+/SiOz H&O = 2.

catalyst at 210 “C, 1 atm,

64

700-

12-

6W-

IO-

SW-

_ 3 6-

400-

? 5 2 ‘3. 6 .E kQ a” 20 0

300-

44

L

, lipk_&f\% I loll

I 2ml

, 300

, 400

I 500

I 600

I 700

-1W 0

Desorption temperature(T)

Desoption temperature(T)

Fig. 4. TPD-MS spectra of CO hydrogenation over Pd-S?/SiO, H&O = 2. Fig. 5. TPD-MS H&O = 2.

IO0 200 3w 400 wo 800

spectra of CO hydrogenation over Pd/SiO,

catalyst at 210 “C, 1 atm, catalyst at 21O”C, 1 atm,

and intensities are also dependent on the catalysts. For instance, the Pd-M$+/Si02 and the Pd-Ca2+/Si02 catalysts have a stronger CH30desorbed peak at the temperature range from 80 to 130 “C, as compared with the other catalysts. In addition, a remarkable CH30 peak at 15 “C appears only on the Pd-Ca2+/Si02 catalyst. In the case of the HCO-desorbed species, more complicated peaks appear. For example, a low-temperature HCO peak at -60 to 15"C is observable with all the catalysts (no detection for the PdS?‘/SiO, catalyst). A HCO peak at 45-70 “C appears only on the Pd-M2+/Si02, Pd/Si02 and Pd-Ba2+/Si02 catalysts, but the PdCa2+/Si02 and Pd/SiO,, whose palladium shows almost metallic valence states [Xl, have another HCO peak at llO-130°C. Moreover, a high-

wo44t.101

2

wo-

E

x 8

,L_ ! \ ,._.,, *o

.Z

;

loo-

?” ! /y,.,_,,

2

o / -ml

/

0

4

l.\.\.

L______

IW 200 3w

400

‘.A

500

600

Desolption temperature(T)

Fig. 6. TPD-MS spectra of CO hydrogenation over Pd-Ba2+/SiOz H&O = 2.

catalyst at 210 “C, 1 atqn.

65

temperature HCO peak at 300-330°C is detected on all the catalysts except Pd-M$+/SiOx, which gives a broad HCO peak in the range from 350430 “C instead. The Pd-Ba2’/Si02 catalyst also gives a HCO peak at 180-260°C. As for HCOO species, only some of the catalysts, such as Pd/Si02 and Pd-Ba2+/Si02, exhibit this desorbed species. In conclusion, the TPD-MS spectra of CO hydrogenation show a remarkable difference in both peak intensities and desorption temperatures among the above-mentioned catalysts. Assignments of the reaction TPD-MS peaks Because the above spectra are complicated, we should carefully determine the peaks. Here, we try to do this on the basis of the TPD-MS spectra of pure chemicals mentioned above and on other facts such as the distribution of palladium valence states [151. First, in order to identity the CO2 and HCOO desorbed peaks, the spectrum of formic acid preadsorbed on the Pd-Mg+/Si02 catalyst was taken after formic acid was introduced in the stream of hydrogen onto the reduced catalyst. The result in Fig. 7 shows that two main CO2 desorbed peaks appear at about 180 and 310 “C. These two CO2 peaks should correspond to two adsorbed formate states, because it has been reported that a HCOOH sample is adsorbed in the form of formate and hydrogen species [161. Furthermore, the peak at 180°C can be assigned to a Pd+HC(O)OMg2+ state, and the other to a Pd’?EIC(O)O4% state because of the following facts: (1) the Pd-M3+/Si02 catalyst has more abundant Pd’ ions, detected by ESR, than the Pd/Si02, which shows almost a Pd metallic valence state [Xl and has only a main CO2 peak at 320 “C, and (2) the HCOOH-modified Pd-Mti+/SiO, catalyst, which should have more abundant Pd+ ions than the above-reacted Pd-Mg+/Si02, shows a much stronger CO2 peak at 180 “C

-106

0

133 303 333 430 500 633 Dempticm temperarure(“C)

Fig. ‘7. TPD-MS 210 “C.

spectra of HCOOH adsorption in H, stream on the Pd-_M%+/SiO,

catalyst at

(Fig. 7) than that of th e reacted catalyst (Fig. 2). This assignment should be also true for the other catalysts except the Pd-Ca2+/Si02 catalyst, where the CO2 peak at 190°C should correspond to the Pd”HC(0)OCa2+ state instead of the Pd+HC(0)OCa2+ state, because of the few Pd’ ions in the catalyst. Figure 7 also shows three HCO peaks at about 50-60, 170 and 310 “C. The latter two peaks should be the fragments of the two adsorbed HCOO states. However, the first peak probably results from the hydrogenation of the HCOO species. In order to identify this peak and the HCO peaks in Figs. 2-6, a TPD-MS spectrum of HCHO adsorption in the stream of Hz on the Pd-M$+/Si02 was taken, as shown in Fig. 8. It tells us that only one HCO peak at 50-60 “C appears and no adsorbed HCOO species is detectable. Compared with Fig. 7, this result indicates that the first HCO peak in Fig. 7 comes from an adsorbed formaldehyde and that the adsorbed HCOO species can be hydrogenated to an HCHO species over the Pd-Mg”+/Si02 catalyst, but the reverse process does not happen. For determining the high-temperature CO2 peaks in Figs. 2-6, a TPD-MS spectrum of adsorbed CO2 on the Pd-Mg2+/Si02 catalyst was also taken, as illustrated in Fig. 9. The result reveals that the high temperature CO2 peaks should come from the desorption of adsorbed CO2 or from the decomposition of carbonates formed during the adsorption of C02. Such adsorbed CO2 or carbonate species have also been observed over Cu-ZnA120s catalysts by our group 1181. They are probably formed from adsorbed HCOO or from CO disproportionation. It is well known that CO2 is easily adsorbed over alkaline earth oxides, forming carbonates such as monodentate and bidentates. Therefore, we suggest that the high temperature CO2 peaks

29

-I-

I6

6z

5

s x ,x z 6 .E t4 8

4-

i

5elo

_.'.-._.-.-.-.-., 0

I

I

loo 2m

I , I I 300 400 500 600

Desmption tempemNle(“C)

Desmption temuerature(‘C)

Fig. 8. TPD-MS spectra of HCHO adsorption in H, stream on the Pd-M2+/SiO, 210°C.

catalyst at

Fig. 9. TPD-MS spectra of CO, adsorption in H, stream on the Pd-M$+/SiO, 210 “C.

catalyst at

67

are from adsorbed COz or decomposed carbonates on the promoters or supports. As for the determination of the other peaks, the results of other researchers provide evidence that the CHBO peak at llO-130°C probably results from adsorbed methoxides 13, 71, which is also confirmed by our chemical trapping results 1171. In the case of the HCO peak at 110-130 “C, a part of it might be the fragments of the adsorbed CH30 species, but most parts probably come from adsorbed HCO species. Such species may be more stable on metallic Pd atoms because more abundant such species are present on the Pd-Ca2+/Si02 and Pd/Si02 catalysts. As for the low temperature HCO peak at -60-15 “C, no direct evidence can be used to assign it. But we prefer to assign it to an adsorbed HCOH species, because such a species is unstable and has been reported elsewhere. 191 It appears likely from the desorption temperature of this species that it is more stable on metallic Pd atoms than on Pd”+ ions. This agrees with the results of Hicks and Bell [9] about the formation of HCOH on Pd” or Pds-. Based on the above assignments, the intensities, desorption temperatures and corresponding adsorbed states or intermediates of the peaks in Figs. 2-6 are summarized in Table 1. Discussion of the reaction spectra It is shown from Table 1 that the Pd+HC(0)OM2’ intermediate state is only present on those catalysts which have abundant Pd+ ions, namely the M$+-, S?‘- and Ba2+-promoted catalysts. Moreover, it reveals that the reaction intermediate, including its stability, conversion and concentration, largely depends on the promoter or Pd valence state. This conclusion is further confirmed by the work of both Driessen et al. [20], who observed a good correlation between the number of extractable Pd”+ ions and the ratio of Mg/Pd, and Naito et al. [ll, who reported that a highly active methanol Pd-Na+/SiO, catalyst showed abundant surface formate ions, while a less active methanol Pd/Si02 catalyst had no such formate species. In addition, Hindermann et al. [7] observed a good relationship between the concentration of a formyl species and the ratio of Mg/Pd. All these results prove that promoters or supports can greatly affect the formation and conversion of an intermediate. The following three kinds of such promoter or support effects may occur: (1) indirect influence on an intermediate through their effects on Pd active centers, as in our case; (2) direct participation in the formation of an intermediate, as observed here and elsewhere 11, 2, 4, 51; and (3) effect on the stability and conversion of an intermediate via the dipole induction of their cations to the oxygen end of the intermediate, as stated elsewhere 121, 221. The results in Table 1 provide new evidence for the effects of promoters on intermediates. The table shows that the Pd+HC(0)OM2+ species can be desorbed as either CO2 or HCOO species, depending on the hydrogenation or dehydrogenation ability of the catalysts. In the case of the Pd-M$+/Si02

a CO hy~~tion b Peak in&u&y.

Pd-HCHO PdHcxOH)

M-0-M(Si)

(7”““;

fL.ri

Pd%XIOSi or Pd?KXKK!a2’

Pd+HCOOMz+

intermediate

or

45 130 HCO

WC0

CH20

co2

co2

HCO

co2

HCO

was conducted at 1 atm and 210 “C.

67 -60

225

72105

CH,O

NC0 KC0

105

640

l&5

520

co2

co2

100

400646

WC0

1250

6300

165

Pd-Ca2’/Si0,

o-15

llO130 llO130

500

430

456

290 180

300

165

130

43

65

30

2ooo 145

I (mV>

of CO hy~ge~tion

I MS T CmWb fragment CC)

320

co2

co2

MS T fragment WI

Pd-Mg’/SiOz

TPWMS peaks

TPD-MS peaks and their corresponding ~~~~a~

TABLE 1

HCO n.d.

HCO

HCO

HCO

(302

CO!2

80

loo200

605

512

406

312 317

175

x MS fragment (%I

Pd--S?+/SiOa

50

loo

760

9150

60

9300 300

3600

HCO

HCO

HCO

co2

CQ2

HCO HCOO HCO

45 0

120

310 320 325

130 200

a0

196 205

195

7100 196 160 50

MS T I fragment t*C) (mV1

Pd/SiOz

catalystsa

I (mV>

over [email protected]“/SiO,

HCO NC0

co,

HCO

HCOO HCO

co2

60 -60

650

400

85 82 305

MS T fragment (“6)

Pd-Ba’+/SiO

2

310 166

125

60

5450 100 250

z (mV)

8

69

catalyst, the palladium can have sufhcient Pd4d5s vacant bands to accommodate adsorbed hydrogen, due to its strong Mg2+3s-Pd4d1’5s direct orbital interactions, as confirmed elsewhere [13, 231. For this reason, both Pd’HC(G>OM2’ and Pd”HCCO)OSi species are thermally desorbed in the form of CO2 instead of HCOO species; as for the Pd-Ba2’/Si02 catalyst, however, few direct Ba2+6s-Pd4d1°5s interactions can be found, based on a SCC-DV-Xa calculation [231, so that only a small amount of Pd vacant bands can adsorb or absorb hydrogen, as observed by us [13, 151. Therefore, there is still some HCOO desorption at -80-160 “C for the Pd’HC(0)OBa2’ species. In addition, the Pd’HC(0)OBa2+ species is desorbed at -90 “C lower than that of the other two Pd’HC(0)OM2’ species. This can be ascribed to the weakest dipole induction of the Ba2+ cations to the oxygen of the HCOO species; as far as the Pd-S?/SiO, catalyst is concerned, the desorption temperature of the Pd+HC(O)OW+ species is almost the same as that of the Pd+HC(O)OM$+ species, but its intensity is between those of the MS+ and Ba2+-promoted catalysts, as in the case of the desorbed amount of hydrogen 1131. It can be thus concluded that the promoters greatly influence the formation, the stability and hence the reactivity of the Pd’HC(0)OM2+ species, probably through either their orbital interactions with Pd atoms or their dipole induction to the formate. Moreover, compared with the results of both the hydrogen sorption capacity and the activity for methanol 1131, the above TPD-MS results indicate that the intensity of Pd+HC(0)OM2+ species of the three abovementioned catalysts is inversely proportional to both the H2 desorbed amount and the activity for methanol. Thus, it is reasonable to conclude that the conversion of the Pd’HC(0)OM2’ species may play a more important role in methanol synthesis than the formation of this intermediate. In order to identify whether or not the Pd’HC(0)OM2’ species can be hydrogenated to methanol, the reacted Pd-M$+/Si02 catalyst, which had been covered with the Pd+HC(0)OM2+ and Pd’HC(O)OSi species, was exposed to the stream of hydrogen at 260 “C for 2 h, and the effluent gas was passed through a liquid nitrogen trap. After that, the catalyst was characterized by TPD-MS. The result (Fig. 10) shows that the Pd’HC(O)OM$’ species was completely hydrogenated to methanol, which was detected by GC as a main product of the Pd+HC(O)OM$+ hydrogenation, while the Pd’HC(O)OSi species can only partially be converted, indicating that the Pd+HC(O)OM$’ species can be converted to methanol much more easily than the Pd’HC(O)OSi species. This result leads us to the conclusion that the Pd+HC(O)OM$+ species may be the active intermediate for methanol formation. The above conclusion is similar to that of Naito et al. [ll,who observed a concurrent change of CH30H with HCOO species over a Pd-Na+/Si02 catalyst. Moreover, they found that the formation of both methanol and HCOO species increased with reduction temperature. Thus, we suggest that the HCOO species can be formed either on Pd+ ions or on Pd atoms, but the HCOO species desorbed at below 200 “C is probably formed on the promoter

Fig. 10. TPD-MS spectra of the Pd-w’/SiO, catalyst: (a) asker CO hydrogenation at 210 “C ’ and 1 atm, and (b) after (a) was exposed to the H, stream at 210 “C and 1 atm.

cation-containing Pd centers such as Pd(X, Cl)Na’ or Pd(X, Cl)M$+ (X = Cl or 0), because no such species is detected over Pd/SiOz catalysts here and elsewhere [ 1, 23. Therefore, we believe that Pd valencies and structures as well as promoters or supports have a significant effect on the formation and stability of the HCOO species. It is also shown from Table 1 that the Pd’HC(O)OSi is desorbed to CO2 at about 150 “C higher than the Pd’HC(0)OM2’ species. Moreover, it appears that the PdOHC(O)OSi species is more easily decomposed to the HCO species than the Pd+HC(0)OM2+ species. This probably implies that the Pd’HC(O)O--Si bond is stronger than the HC(O)-O and Pd+HC(O)OM2’ bonds. The result of bonding function calculations, which represent relative bonding strengths [241, can certify this speculation, as shown in Table 2. It points out that Pd+ ions can activate the Si-0 bond more easily than they do the C-O bond in the HC(O)OSi species. Also, Pd+ ions have much stronger ability to activate the Si-0 bond than Pd” atoms. But Pd” can activate the C-O bond in the HC(O)OSi species much more strongly than it activates the Si-0 bond. Based on these results and those in Fig. 10, we suggest that the Pd’HC(0)OM2’ species is hydrogenated faster than the TABLE 2 Effecta of Pd valence states on the bonding functions of the SC-0 and HC( 0)-O Pd valence state

Pd+ Pd”

bonds

Bonding function Sk-0 bond ( X lo-‘)

HC(OJ-0

3.056 0.858

2.384 2.118

bond (x10-s)

71

Pd’HC(O)OSi species and that the Pd”HC(0)OSi or Pd”HC(0)OCa2’ species are probably inclined to decompose to the HCO species. As demonstrated before [ 14-151, adsorbed hydrogen plays an important role in HCOO conversion and in methanol formation. In our opinion, adsorbed H6+ species, which was detected by the chemical trapping of a CH,O-Na+-CH,OH solution and proven to exist more abundantly on Pd”+ ions than on Pd0[141, can participate in the conversions of both the Pd+HCOOM2+ species to the Pd+HCO (or HCHO) ...M2+-OH species and the CH&-M2+ species to methanol; while the H”- species, which was trapped by a (CHB)2S04 agent and found to be more favorable on Pd” 1141, can nucleophilically attack the carbon atom of CO or HCOO or other intermediates such as HCO species. Based on the above, therefore, we suggest that during methanol formation the Pd’HC(0)OM2’ species may prefer to react with the H”+ species, giving Pd+HCO (or HCHO) species, while the Pd”HC(0)OCa2+ and Pd’HC(O)OSi species are probably decomposed to the Pd’HCO species. Table 1 also gives the HCHO and HC(OH) species. In our opinion, the HCHO species results from the hydrogenation of either the HCOO species or the HCO species, as observed above and elsewhere 1181; while the HC(OH) species may be formed from the reaction of the HCO species with Ha+, rather than from the reaction of adsorbed CO with molecularly adsorbed H2 because it is almost impossible to have molecular hydrogen on the catalysts under the reaction conditions. In addition to the above, it is also possible that the HCO species can be formed on the Pd-Ca2+/Si02 catalyst through the direct hydrogenation of bridge adsorbed CO because the catalyst has stronger Pd back-donation abilities and more abundant H”- species [151.

Conclusions The reaction TPD-MS spectra show large dependence of intermediates and mechanisms on the nature of alkaline earth cations. The following conclusions can be drawn: Formate species, as the most abundant intermediate on the all catalysts, is not formed from adsorbed formyl or formaldehyde species, but probably from the reaction of bridge-adsorbed CO, Ha-, and surface oxygen species. Normally, there are two kinds of formate states: PdHC(0)OM2+ and Pd’HC(O)OSi. The concentration of the first state is found to decrease with alkaline earth cations in the following order: Pd-Ba2+/Si02>PdS?‘/SiO, > Pd-Ca2+/Si02 > Pd-M$+/Si02 which is opposite to that of the TOF for methanol [ 131. Moreover, the first state can be completely hydrogenated to methanol much faster than the second. The formation, conversion and stability of the for-mate or other intermediates largely depend on the nature of the cations and Pd valencies. An

72

interpretation has been given on the basis of the so-called bonding function calculation. Two reaction mechanisms are claimed to occur simultaneously on two different active centers. The first one is assumed to undergo hydrogenolysis of Pd’HC(O)OM’+ to HCO or HCHO species as a rate-determining step (rds) on a P#Pd’bridge(Cl, O)M2+ center, while the rds of the second one is asserted to be the hydrogenolysis or decomposition of the Pd”HC(0)OCa2+ or PdOHC(O)OSi species to HCO or HC(OH) species over a Pd” center. The first one prevails on the M$+-, S?‘- and Ba2+-promoted catalysts; while the second is dominant on the other two catalysts.

Acknowledgements The authors wish to thank the Chinese Natural Science Foundation for its financial support. Thanks are also due to Prof. K. R. Tsai for his guidance and encouragement as well as to Prof. B. Delmon for his providing the opportunity for us to type the paper.

References 1 S. Naito, H. Yoshika, H. Orita aud K. Temaru, Proc. &h Znt. Congr. Cutal, Berlin, 1984, Vol. 3, p. 207. 2 Y. Kikuzono, S. Kagami, S. Naito, T. Onshi and K. Tamaru, Faraday Discuss. Chem. Sot., 72 (1982) 135. 3 R. Kieffer, R. Stupfler and A Deluzarche, Tetmhedron Lett., (1980) 3371. 4 A Palazov, G. Kadinov, Ch. Bonev and D. Shopov, J. Cutal., 74 (1982) 44. 5 C. Sudhakar and M. A. Van&e, J. Catal., 95 (1985) 227. 6 M. Ichikawa, Nippon Kagaku fiishi, 2 (1982) 213. 7 J. P. Hindermanu, A Kiennemann, A. Chakor-AIami and R. Kieffer, Prwc. &h Znt. Congr. Catal., Berlin, 1984, Vol. 2, p. 163. 8 M. L. Poutsma, L. F. Elek, P. A Ibarbia, A P. Riech and J. A Raabo, J. Cutal., 52 (1978) 157. 9 R. F. Hicks aud A T. Bell, J. Catal., 91 (1985) 104. 10 C. S. Kellner and A T. Bell, J. Cutal., 71 (1981) 288. 11 J. P. Hindermann, A Deluzarche, R. Kieffer and A Kiennemann, Can. Chem. Eng., 61 (1983) 21. 12 R. Kieffer, E. Remameon, A Deluzarche and Y. Tramvou, React. K&t. Catal. L.&t., 16 (1981) 207. 13 Y. F. Shen, J. T. Li, L. F. Cai and K.. H. Huang. Prw. 3rd China-Japan-USA Symp. Catal., Xiamen, China, 1987, C-15. 14 Y. F. Shen, L. F. Cai, J. T. Li, S. J. Wang and K. H. Huang, Catal. Today, in press. 15 Y. F. Shen and K. H. Huang, Appl. Cat&., submitted. 16 K. Tamaru, Z+vc. ad China-Japan-USA Symp. Catal., Xiemen, China, 1987, A-2. 17 Y. F. Shen and K. H. Huang, J. Mol. Cat& submitted. 18 L. F. Chen and K. H. Huang, CO, Hydrogenation ouer &-Based Catalyst, to be published. 19 E. K. Poels, E. H. van Broekhoven, W. A A van Bameveld and V. Ponec, React.Kinet. Cat& L.&t., 18 (1981) 223.

20 J. M. Driessen, E. K. Poels, J. P. Hindermann and V. Ponec, J. Catal., 82 (1963) 26. 21 F. Fajula, R. G. Anthony and J. H. Lunsford, J. Cutul., 73 (1982) 237. 22 A. Kiennemann, J. P. Hindermann, R. Breault and H. Idriss, in Symp. Chemicals from Syngas and MeOH, New York, 1966. 23 W. J. Lai and K. H. Huang, 4th China Symp. Cutal., Tianjin, 1988, Abstr. 1-E-13. 24 K H. Huang, Appl. Cutal., 15 (1985) 175