Model Pumices Supported Metal Catalysts: II. Liquid Phase Selective Hydrogenation of 1,3-Cyclooctadiene

Article abstract of DOI:10.1006/jcat.1997.1764

The catalyzed, selective hydrogenation, in liquid phase, of 1,3-cyclooctadiene was studied on a series of Pd catalysts supported on natural pumice, model pumices (with variable content of alkali metal ions), silica, and sodium-doped silica. At constant pressure of H₂ (1 atm.) the reaction follows a zero-order kinetic for all the Pd catalysts. At low metal dispersion (D_x < 20%), Pd/pumice catalysts exhibit higher activity as compared to Pd/silica catalysts; the turnover frequency is maintained even at high metal dispersion in Pd/pumice but Pd/silica and Pd/model pumices without alkali ions show a decrease in activity. At increasing D_x, the binding energy shift of Pd 3d level is negative in Pd/pumice and positive in Pd/silica with respect to unsupported Pd metal. The different performances of the Pd/pumice catalysts are explained by the presence of alkali metal ions in the framework of the support. Addition of sodium ions to Pd/silica catalysts produces a negative shift of the binding energy, but the activity is not improved because the number of active sites diminished due to decoration of palladium particles by sodium ions. The Pd catalysts with alkali metal ions in the support are resistant to air oxidation. In Pd catalysts containing alkali metal ions the selectivity to cyclooctene is practically 100% and the constant rates ratio k₁/k₂ is more than 1000 with a maximum at dispersion 35-40%, whereas the selectivity in Pd catalysts without alkali metal ions decreases continuously at increasing dispersion.

Full text of DOI:10.1006/jcat.1997.1764

JOURNAL OF CATALYSIS 171, 177–183 (1997)
ARTICLE NO. CA971764
Model Pumices Supported Metal Catalysts
II. Liquid Phase Selective Hydrogenation of 1,3-Cyclooctadiene
,
,
ꢀ
ꢀ
ꢀ
Leonarda F. Liotta, Anna Maria Venezia, † Antonino Martorana,† and Giulio Deganello †
ꢀ
Istituto di Chimica e Tecnologia dei Prodotti Naturali del CNR (ICTPN-CNR), via Ugo La Malfa 153, 90146 Palermo, Italy; and †Dipartimento di
Chimica Inorganica, Universita` di Palermo, via Archirafi 26-28, 90123 Palermo, Italy
Received February 10, 1997; revised April 21, 1997; accepted May 14, 1997
catalytic hydrogenation of a particular substrate. The re-
cent application of the gel technique (14) to the synthesis of
pumice-like materials with different alkali metal ions con-
tent allows one to obtain a series of Pd catalysts with a large
range of alkali metal ions/palladium atomic ratios R aiming
to find selective catalysts for particular reactions. Selectiv-
ity is indeed important for the purification of alkene feed-
stocks for the polymer manufactures from small amounts
of alkadienes and alkynes at ppm levels (15–17). Although
the addition of ionic salts (18) and traces of other met-
als (Si, Ge, Pb, Mn) (16, 19–21) to the Pd/Al2O3 catalyst,
usually employed in industrial processes, improved the se-
lectivity, at present carbon monoxide is most often used
for such purpose (16, 21, 22). This reagent, besides being
toxic, does not inhibit the formation of the so-called “green
oil,” an oligomer fraction which decreases the value of the
monomer. The development of a selective catalyst which
does not need the addition of carbon monoxide appears
extremely important. The good results of selectivity so far
obtained (6) with natural pumice as an alternative support
for a palladium catalyst (1) might be improved by choosing
a particular ratio of the alkali ions in the framework of the
support (14) to the Pd atoms.
The catalyzed, selective hydrogenation, in liquid phase, of 1,3-
cyclooctadiene was studied on a series of Pd catalysts supported
on natural pumice, model pumices (with variable content of alkali
metal ions), silica, and sodium-doped silica. At constant pressure
of H2 (1 atm.) the reaction follows a zero-order kinetic for all the
Pd catalysts. At low metal dispersion (Dx < 20%), Pd/pumice cata-
lysts exhibit higher activity as compared to Pd/silica catalysts; the
turnover frequency is maintained even at high metal dispersion
in Pd/pumice but Pd/silica and Pd/model pumices without alkali
ions show a decrease in activity. At increasing Dx, the binding en-
ergy shift of Pd 3d level is negative in Pd/pumice and positive in
Pd/silica with respect to unsupported Pd metal. The different per-
formances of the Pd/pumice catalysts are explained by the presence
of alkali metal ions in the framework of the support. Addition of
sodium ions to Pd/silica catalysts produces a negative shift of the
binding energy, but the activity is not improved because the number
of active sites diminished due to decoration of palladium particles
by sodium ions. The Pd catalysts with alkali metal ions in the sup-
port are resistant to air oxidation. In Pd catalysts containing alkali
metal ions the selectivity to cyclooctene is practically 100% and the
constant rates ratio k1/k2 is more than 1000 with a maximum at
dispersion 35–40%, whereas the selectivity in Pd catalysts without
alkali metal ions decreases continuously at increasing dispersion.
c
ꢁ 1997 Academic Press
To test this possibility a series of palladium catalysts on
synthetic pumice having a different content of Na+, K+,
and (Na+ + K+) (14), and on silica, are here used in the
catalytic hydrogenation of 1,3-cyclooctadiene (1,3-COD),
INTRODUCTION
Some recent results on pumice-supported metal catalysts in liquid phase. The choice of 1,3-COD as probe substrate,
(1–7) showed that electronic influences by the support can instead of other dienes, is justified by the absence of any
be important and can produce substantial improvement in isomerization process and by the possibility of immediate
the activity and selectivity of supported palladium catalysts comparison with the performances of Pd/pumice (2) and
–
(2–4, 6). It is now established that the alkali metal ions Pd Pt/pumice (23) already tested in our laboratory in the
(Na+ and K+) present in the framework of natural pumice hydrogenation of this substrate.
(8, 9) increase the electron density on the supported metal
(10, 11) depending on metal dispersion (12). The alkali
metal ions in the framework of pumice appear to decrease
the adsorption of CO (13) but so far there have been no
EXPERIMENTAL
All experimental procedures were performed in stan-
studies dealing with a definitive amount of alkali metal ions dard Schlenk glassware under atmosphere of prepurified
producing the best activity/selectivity performances in the nitrogen. Anhydrous tetrahydrofuran (THF) was prepared
177
0021-9517/97 $25.00
c
Copyright ꢁ 1997 by Academic Press
All rights of reproduction in any form reserved.

178
LIOTTA ET AL.
according to literature procedures (24) and distilled from
potassium under nitrogen just before use.
TABLE 1
Structural Characteristics of Supported Palladium Catalysts
1,3-Cyclooctadiene and cyclooctene (COE) (Aldrich)
were purified by fractional distillation from potassium,
passed over activated alumina and stored at ꢂ5ꢃC, under
nitrogen. Purity was tested by GC analysis. All other chem-
icals (Aldrich) were of reagent grade purity and were used
without further purification.
Atomic Atomic
Metal
1E
ratio
ratio dispersionc (Pd 3d5/2
)
Catalysta
Si2
Si1
Si0
Z2
Z1
Z0
Na6
Na5
Na4
Na3
Na2
Supportb Pd% Na/Pd
K/Pd
Dx%
(eV)
SiO2
SiO2
SiO2
Synt0
Synt0
Synt0
SyntNa
SyntNa
SyntNa
SyntNa
SyntNa
SyntNa
SyntNa
SyntNa-K
SyntNa-K
SyntK
1.0
1.0
1.0
1.0
1.0
1.2
0.3
0.28
0.95
0.3
0.95
1.4
0.55
0.27
0.3
0
0
0
0
32
39
47
10
17
39
18d
25d
27
32d
32
39
50
32d
39d
27d
39d
56
14
22
28
32
32
39
51
62
12d
52
—
—
—
—
—
ꢂ0.1
ꢂ0.6
ꢂ0.5
ꢂ0.5
—
0
0
0
0
Supports and Palladium Catalysts
0
0
0
0
The preparation of the supports is reported in the previ-
ous paper (14). The preparations of the “model pumices”
supported palladium catalysts were performed, adapting
literature procedures used for Pd/silica catalysts, with two
different precursors, [Pd(NH3)4](NO3)2, by ion exchange
(25–27) or [Pd(C3H5)2] according to Yermakov (28). Struc- Na1
tural characterization of the catalysts was performed with
X-ray (29) and XPS (10–12) techniques as described pre-
viously (14). The catalysts used in this work and their mi-
crostructural characteristics are listed in Table 1.
6.4
0
1
4
1
1
1
2
3
68
0
2.1
6.4
2.1
3.5
2.9
7.5
20
0
0
0
6.4
18.3
7.9
11.1
17.4
62.8
33.9
136
22.5
6.4
0
0
0
0
—
ꢂ0.4
Na0
Na-K1
Na-K0
K2
K1
K0
W7
W6
0
—
6.2
16
64
88
20
7.0
20
8.6
12
19
68.5
37
148
—
—
—
—
1
2
0.2
+0.1
SyntK
SyntK
0.15
0.55
1.0
0.37
0.86
0.61
0.39
0.1
—
ꢂ0.1
ꢂ0.4
ꢂ0.5
ꢂ0.4
ꢂ0.6
ꢂ0.4
ꢂ0.7
ꢂ0.6
Pumice
Pumice
Pumice
Pumice
Pumice
Pumice
Pumice
Pumice
Kinetic Studies
The hydrogenation reactions were performed in an ap- W5
paratus described previously (2, 4), at 25 ꢄ 0.1ꢃC. The re-
W4
W3
W2
W1
actor was connected by stopcocks to a vacuum line and to
a gas adsorption system operating at atmospheric pressure
of hydrogen. A side arm closed by a silicone septum al-
lowed withdrawal of the samples for GC analyses. All the Pd/Na-synt Na-synt
0.2
0.05
0.22
W0
ꢂ1.0
e
(6.4) Na-Si Na-doped 1.5
ꢂ0.5
ꢂ0.6
reactions were performed at the constant stirring rate of
2.500 rpm. At this stirring rate there is no gas–liquid dif-
fusion (2). The amorphicity and the low specific surface
area of pumices guarantee also limited liquid–solid diffu-
sive effects (30). Before starting the reaction the catalyst–
THF system was stirred for half an hour with H2. In the
case of Pd catalysts without sodium ions in the support this
time must be much longer to have reproducible results (see
Results and Discussion).
In all the experiments the molar ratio substrate/Pd was
5.000 : 1 and substrate concentration was 0.25 M, in THF.
Kinetic analyses were performed by a GC DANI 3800 HR
PTV, equipped with a Supelco capillary column SPB1701
and a DANI 68/10 FID detector. The system was associated
to a Shimadzu C-RIB Chromatopac automatic integrator.
Deviations in repeated experiments were less than 4% .
SiO2
(1.6) Na-Si Na-doped 1.5
SiO2
1.6
—
56
a
Six = Pd/SiO2; Nax = Pd/synthetic pumice with sodium; Kx = Pd/syn-
thetic pumice with potassium; Na-Kx = Pd/synthetic pumice with sodium
and potassium; Zx = Pd/synthetic pumice without sodium and potassium;
Pd/Na-synt = catalyst doped with sodium on the support surface. For the
preparation and characterization of Pd/pumice catalysts (W series) see
Ref. (29).
b
For the preparation and characterization of the synt supports see
preceding paper (14).
c Determined from volume-weighted average Pd crystallite diameters
obtained by LB analysis of the (111) XRD peak. See Ref. (29).
d Determined from SAXS measurements. See Ref. (29).
e The Pd 3d signal was too noisy to obtain a reliable shift.
used to fit the experimental data (t vs C1,3-COD) was
t = [(1 ꢂ Q)(1 ꢂ C1,3-COD ) ꢂ Q ln C1,3-COD ]/k, [1]
RESULTS AND DISCUSSION
The reaction mechanism for the catalyzed hydrogena- where Q is the ratio of the adsorption constants on the
tion of 1,3-COD to COE was obtained, as previously de- catalyst surface KCOE /K
scribed (2), following a procedure similar to that employed COE in the liquid phase is CCOE
by Happel and Sellers (31), neglecting further hydrogena- that depends only on the pressure of H2; and Q and k are
tion of COE since cyclooctane (COA) is not detectable nonlinear parameters. For a detailed description of Eq. [1]
before complete consumption of 1,3-COD. The equation and its derivation see Ref. (2). Some selected experiments
and the concentration of
1,3-COD= 1-C1,3-COD; k is a constant

SELECTIVE HYDROGENATION OF 1,3-COD ON Pd/PUMICE CATALYSTS
179
Zx series, at similar metal dispersion, and the amount of
the potassium ion does not influence greatly the activity.
In all the other Pd/pumice series the activity is high and
practically constant up to Dx ꢅ 30% , then decreases with
increasing dispersion. However, at similar dispersion, cata-
lysts containing sodium in the framework of the support are
always more active than those of the Six and Zx series. The
different behavior can be attributed to electronic effects
played by the supports. Indeed the positive shift of Pd 3d
binding energy in Pd/silica at high dispersion was associated
(33–36) to a decreased electron density on the supported
palladium which caused the drastic decrease of activity in
the hydrogenation of the electron-rich highly unsaturated
hydrocarbons (alkadienes and alkynes) because of a strong
interaction between the metal and the substrate (17, 36–39).
Accordingly, on addition of a strong ꢀ-donor ligand such
FIG. 1. Hydrogenation of 1,3-COD on Pd catalysts: W2 (᭜), Na6 (᭹),
K2 (᭡). Experimental points and fitting with Eq. [1].
TABLE 2
using the fitting procedure of Eq. [1] are reported in Fig. 1,
while data for all the catalysts are listed in Table 2.
The reactions follow a zero-order kinetic behavior up to
high extents of conversion, in agreement with the experi-
ments performed on the analogous reaction with Pd/natural
pumice catalysts (2). As already established (2, 23), the rate
determining step is the surface reaction
Kinetic Data of the Catalyzed Hydrogenation
of 1,3-Cyclooctadiene
Metal
TOF1 TOF2 ꢆ 103
Catalysta dispersionb Dx% (sꢂ1
)
(sꢂ1
)
k1/k2 Q = KE/K1.3
Si2
Si1
Si0
Z2
Z1
Z0
32
39
47
10
17
39
18c
25c
27
32c
32
39
50
32c
39c
27c
39c
56
14
22
28
32
32
39
51
62
12c
52
56
42
34
32
79
78
35
93
88
82
79
68
47
34
88
75
52
37
25
89
91
91
89
88
85
62
32
40
16
12
75
76
76
75
74
75
119
94
84
71
78
65
48
81
77
62
69
73
110
80
77
73
70
62
42
39
43
12
11
560
450
420
1050
1030
470
780
870
940
1030
960
730
710
1070
1000
840
540
350
810
1150
1200
1230
1250
1400
1480
830
940
1360
1090
0.24
0.22
0.20
0.34
0.31
0.21
0.67
0.49
0.24
0.47
0.48
0.86
0.41
0.43
0.85
0.51
0.48
0.44
0.44
0.34
0.26
0.21
0.20
0.18
0.17
0.16
0.15
0.14
0.11
k1
1,3-COD + 2H ⇒COE ,
[2]
ꢀ
ꢀ
ꢀ
where X represents a surface adsorbed species.
ꢀ
Na6
Na5
The dependence of turnover frequencies (TOF1) with
dispersion Dx is shown in Fig. 2, where the results obtained Na4
with Pd/natural pumice (2) and Pd/SiO2 catalysts are also
reported. Figure 2a contains data relative to Pd catalysts
with alkali metal ions in the framework of the support.
Figure 2b contains data relative to Pd catalysts without or
with alkali ions added to the catalyst surface. Although in Na-K0
all the series of Pd catalysts here examined (Table 2 and
Fig. 2) the activity (TOF1) is influenced by a combination
of particle size and support effects, the latter appears more
important. In order to facilitate the interpretation of the
data in Fig. 2, the TOF1 values corresponding to catalysts W5
having similar or the same support have been connected.
Along the same series of catalysts TOF1 does not change
significantly with Dx in the region of low metal dispersion.
For Dx higher than 20–30% the decrease of metal parti-
cle size produces a general decrease in activity. Such a de- Pd/Na-synt
Na3
Na2
Na1
Na0
Na-K1
K2
K1
K0
W7
W6
W4
W3
W2
W1
W0
(6.4) Na-Si
(1.6) Na-Si
crease is influenced by the presence of alkali metal ions in
the framework of the support.
a
In the series of catalysts without alkali metal ions in
Six = Pd/SiO2; Nax = Pd/synthetic pumice with sodium; Kx = Pd/syn-
the support (Six and Zx) the trend (Fig. 2b) is similar to thetic pumice with potassium; Na-Kx = Pd/synthetic pumice with sodium
and potassium; Zx = Pd/synthetic pumice without sodium and potassium;
Pd/Na-synt = catalyst doped with sodium on the support surface. For
Pd/pumice catalysts (W series) see Ref. (29).
that found for the selective hydrogenation of butadiene on
Pd/Al2O3 (32), Pd/SiO2 (33), and Pd/C (33), i.e. a drastic
decrease in activity at dispersion higher than 20% . Pd cata-
lysts with only potassium (K series) (Fig. 2a) do not show a
remarkable increase of activity with respect to the Six and
b
Determined from volume weighted average Pd crystallite diameters
obtained by LB analysis of the (111) XRD peak. See Ref. (29).
c Determined from SAXS measurements. See Ref. (29).

180
LIOTTA ET AL.
does not change the consideration that a positive B.E. shift
in Pd/Al2O3 catalysts is related to a decrease of the activ-
ity in the hydrogenation of unsaturated hydrocarbons on
comparison with Pd/Al2O3 catalysts which do not show the
positive B.E. shift (41).
On the contrary, the negative shift of B.E. in Pd/pumice
(10, 11) is related to an electron density increase on the
supported palladium, as shown by the trend of the Auger
parameter shift (10). Such a parameter, given by the combi-
nation of the B.E. shift and Auger shift, has the advantage of
being independent from possible charging and energy ref-
erence problems (42); it allows one to estimate the charge
distribution on the palladium atoms. The electron density
on the metal in the Pd/pumice series increases with metal
dispersion (10) and with the atomicratio R = (Na)/(Pd) (43)
and the trend of TOF1 appears directly related to the fac-
tors which increase the electron density on the metal. In the
Pd/pumice series the alkali metal ions, in addition to having
a general positive effect on the activity, enlarge the range
of utilization of the Pd catalysts toward higher metal dis-
persion since the decline of activity with increasing metal
dispersion is gradual and much smaller than in other Pd
catalysts on traditional supports (alumina (32), silica (33),
carbon (32, 44)).
In Fig. 3, TOF1 is reported as a function of R for metal
dispersions Dx = 35 ꢄ 5% . TOF1 increases with R up to a
constant maximum value. The data show that the effect of
sodium is more relevant at high dispersion and that the Pd
catalysts supported on silica or on model pumice not con-
taining alkali metal ions have low activity. The potassium
effect is practically absent, as seen by the position of K1
in Fig. 3. Such behavior is in accord with the XPS results
FIG. 2. Change of TOF1 with metal dispersion. Comparison of Pd/
pumice catalysts (W series (᭜)). (a) Pd catalysts containing alkali metal
ions in the support framework: Na (᭹), K (᭡), Na-K (—). (b) Pd catalysts
without alkali metal ions (Si (ꢀ) and Z (ꢁ)) or with alkali metal ions on
the catalyst surface (ꢀ).
as piperidine, reducing the electron deficiency of the sup-
ported palladium enhanced the activity of the catalysts (38).
Recently the positive binding energy (B.E.) shift in XPS
measurements of Pd/Al2O3 catalysts (40, 41) was attributed
to metal particle morphology rather than to metal disper-
sion. Even if morphology of the metal particles could also
FIG. 3. Influence of R(Na + K)/Pd on TOF1 (Dx = 35 ꢄ 5% ). Upper
line (Dx = 32% ) and lower line (Dx = 39% ) connect catalysts with the same
be important in other Pd catalysts, the above interpretation experimental Dx.

SELECTIVE HYDROGENATION OF 1,3-COD ON Pd/PUMICE CATALYSTS
181
of Table 1, where the B.E. shifts of the potassium contain-
ing catalysts are not significant (35), differently from the
sodium containing catalysts. A possible explanation of the
much higher influence of sodium than that of potassium
is that the electrostatic field is much higher for the sodium
than for the potassium ion, because of the ratio Z/r (Z, ionic
charge; r, ionic radius).
Therefore, the progressive decrease of activity in
Pd/pumice catalysts, which ismore evident at high metaldis-
persion (10), isattributed to a lower interaction between the
electron-rich 1,3-COD and the electron density enriched
supported palladium, i.e., to a reason which is the opposite
of that responsible for the decrease in activity of palladium
catalysts on silica or alumina (17, 36–39). Related to this
different trend in the B.E. shift is the easy oxidation in air
of the palladium catalysts not containing alkali metal ions
(Z or Si series), specially when highly dispersed, so that
treatment with hydrogen for long time is required before
adding the diene to avoid very low activity. This treatment is
not necessary for the pumice-supported palladium catalysts
(W, Na, Na-K series), left for months in open atmosphere.
In order to find the Pd catalyst with the best activ-
ity/selectivity performances a detailed study of the second
semi-hydrogenation was also performed. The rate deter-
mining step is again a surface reaction, namely,
FIG. 5. Change of TOF2 with metal dispersion: Pd catalysts without
sodium (K (᭡), Si (ꢀ), and Z (ꢁ)) or with sodium on the catalyst surface
(ꢀ).
the contrary, providing that sufficient stirring with H2 was
performed before starting the kinetics, no change in TOF2
was found with Pd catalysts without sodium ions in the
framework (Fig. 5). Similar results were already reported
by Boitiaux et al. (37) in the liquid phase hydrogenation of
1-butene with Pd/Al2O3, while Hub et al. (39) found an in-
crease of TOF2 in the gas phase hydrogenation of 1-butene.
Again potassium alone in the support structure does not
show a relevant effect on the catalytic performance. Al-
though with Pd/pumice catalysts k1 is about three orders of
magnitude higher than k2 and the selectivity can be consid-
ered fairly good, it is interesting to follow the changes of
k1/k2 with metal dispersion. A volcano curve of k1/k2 versus
metal dispersion is found (Fig. 6) in the sodium-containing
Pd catalysts, in accord with previous results on Pd/pumice
catalysts (2). On the contrary (Fig. 7), a continuous decrease
of k1/k2 is found for all the other catalysts (Si, Z, and K se-
ries).
k2
COE + 2H ⇒COA .
[3]
ꢀ
ꢀ
ꢀ
When alkali metal ions are present in the supports of Pd
catalysts, the rate constant k2 relative to reaction (3) is the
same in both the direct hydrogenation of the pure COE
and in the hydrogenation of the COE obtained from 1,3-
COD; TOF2 decreases with metal dispersion (Fig. 4). On
The high selectivity of palladium catalysts was attributed
to stronger adsorption of the diene as compared to the mo-
noene (45). These results were obtained in a competitive
gas phase hydrogenation of butadiene and 1-butene (which
have similar hydrogenation constant rates) on Pd/SiO2 cata-
lysts at 60% metal dispersion and under high hydrogen
pressure. As shown by several authors (34, 35, 39), the large
positive B.E. shift of the Pd 3d level found at such high
metal dispersion is indicative of a loss of metallic character
which determines an increase of the interaction between
the electron-rich substrate and the electron-deficient metal
centers, with an expected larger effect on the diene as com-
pared to the monoene. In our Pd catalyst without sodium
ions in the structure of the supports (Z, K, and Si series)
FIG. 4. Change of TOF2 with metal dispersion: Pd/pumice catalysts
(W (᭜) series and Na (᭹) series).

182
LIOTTA ET AL.
to synt0 (14), and the other two by addition of NaNO3 to Si1
–
(13). As already reported for similar Na Pd/SiO2 catalysts
(46) and as shown in Table 1, the B.E. shift in the three
catalysts is negative as for all the Pd/pumice catalysts pre-
viously described. The TOF1, however, is quite different
from that found in those Pd/pumice catalysts with similar
R values, due to the different location of the alkali metal
ions. When the sodium ions are added to Pd/silica they dis-
tributed on the surface of the catalysts decorating part of
the supported palladium, as shown by a LEIS (low energy
ion scattering) study (47). In these cases, a decrease of the
active sites occurs which counterbalances the increase of ac-
tivity expected for the augmented electron density on the
supported metal. Therefore it is not surprising that addi-
tion of sodium ions lowers the TOF1 values (Fig. 2b) with
respect to sodium free Pd catalysts of the same dispersion.
Although the hydrogenation of 1,3-COD is just a probe
reaction, the results might be easily extended to substrates
of industrial interest. Knowing the electron donor capac-
ity of the substrate to be hydrogenated, the electron den-
sity on the Pd/pumice catalysts could be modulated by act-
ing properly on metal dispersion and/or on alkali metal
ions/palladium atomic ratio in order to achieve the best
activity/selectivity performances in the selected reaction.
FIG. 6. Volcano trend of k1/k2 ratio of Pd catalysts (Na (᭹) and Na-K
(—) series) versus metal dispersion (Dx).
TOF2 is practically constant on increasing metal dispersion
in accord with the decrease of selectivity (Fig. 7). The con-
tinuous decrease of TOF2 with metal dispersion in the Pd
catalysts of the series W, Na, and Na-K indicates that the
adsorption of the monoene on the metal becomes progres-
sively lower as the electron density on the metal increases.
Finally we consider the effect of the location of al-
kali metal ions when they are not in the support frame-
work. Three catalysts, namely Pd/Na-synt, 6.4 Na-Si, and
1.6 Na-Si, (Table 2), were prepared, the first one by support-
ing palladium on Na-synt obtained by addition of NaOCH3
CONCLUSION
The experimental resultsofthisarticle give a clear picture
of the influence of alkali metal ions on the performances
of supported palladium catalysts and of the importance of
their location. The alkali metal ions in the framework of
the support increase the electron density of the supported
metal, decreasing the interaction of the supported palla-
dium with the electron rich substrate to be hydrogenated.
This effect produces a general increase of activity (TOF)
compared with Pd catalysts on conventional supports (alu-
mina, silica, carbon). The latter show indeed a decreasing
activity at metal dispersion higher than 20–25% , because of
a too strong interaction of the electron rich substrates with
the electron-deficient supported metal. On the contrary the
gradual decrease of activity of the Pd/pumice catalysts at
increasing Dx is due to an insufficient activation of the elec-
tron rich substrate because of its repulsion towards the sup-
ported metal which becomes progressively electron richer.
These arguments are supported by a progressive negative
shift of B.E. of Pd 3d orbital in the Pd/pumice catalysts
compared to the B.E. of the unsupported Pd metal (12, 43).
A similar electronic effect of sodium ions added to Pd/SiO2
catalystsisdocumented byXPS(43, 46). However, the influ-
ence of the increased electron density of these Pd catalysts
on the activity of 1,3-COD hydrogenation is limited by the
contemporary decrease of active sites due to decoration of
Na+ ions of the supported palladium as demonstrated by a
LEIS study (47).
FIG. 7. The decrease of k1/k2 ratio with metal dispersion in Pd cata-
lysts without sodium ions in the support framework: (Si (ꢀ), Z (ꢁ), and
K (᭡) series).

SELECTIVE HYDROGENATION OF 1,3-COD ON Pd/PUMICE CATALYSTS
183
The influence of alkali metal ions in the framework pro- 20. Aduriz, H. R., Bodnariuk, P., Coq., B., and Figueras, F., J. Catal. 129,
47 (1991).
duces a progressive decrease of TOF2 with dispersion. Since
TOF1 starts to decrease at metal dispersion higher than
30–35% , the rate constants ratio k1/k2 has a volcano trend
21. Gutman, H., and Lindlar, H., in “Chemistry of Acetylenes” (H. C.
Viehe, Ed.), 1969, p. 355. Dekker, New York.
22. Sarka´ni, A., Weiss, A. H., and Guczi, L., J. Catal. 98, 550 (1986).
with a maximum of selectivity for the title reaction at about
35% metal dispersion.
23. Deganello, G., Duca, D., Liotta, L. F., Martorana, A., Venezia, A. M.,
Benedetti, A., and Fagherazzi, G., J. Catal. 151, 125 (1995).
24. Wisseberger, A., and Proskauer, E. S., “Organic Solvents,” Vol. VII.
Interscience, New York, 1955.
25. Gubitosa, G., Berton, A., Camia, M., and Pernicone, N., in “Prepara-
tion of Catalysts. III” (G. Poncelet, P. Grange, and P. A. Jacobs, Eds.),
p. 431. Elsevier, Amsterdam, 1983.
ACKNOWLEDGMENTS
This work was supported by CNR (Progetto Strategico “Metodologie
Chimiche Innovative”), Ministero della Universita` e della Ricerca Scien-
tifica e Tecnologica (MURST, 40% ), and Pumex S.p.A.
26. Benesi, H. A., Curtis, R. M., and Studer, H. P., J. Catal. 10, 328
(1968).
27. Fuentes, S., and Figueras, F., J. Chem. Soc., Faraday Trans. 74, 174
(1978).
REFERENCES
28. Yermakov, Yu. I., Catal. Rev. 13, 77 (1986).
1. Deganello, G., Duca, D., Liotta, L. F., Martorana, A., and Venezia, 29. Fagherazzi, G., Benedetti, A., Deganello, G., Duca, D., Martorana, A.,
A. M., Gazz. Chim. It. 124, 229 (1994). and Spoto, G., J. Catal. 150, 117 (1994).
2. Deganello, G., Duca, D., Martorana, A., Fagherazzi, G., and Benedetti, 30. Ruiz, P., Crine´, M., Germain, A., and L’Homme, G., in “Chemical and
A., J. Catal. 150, 127 (1994).
Catalytic Reactor Modelling” (M. P. Dudukovic and P. L. Mills Eds.),
p. 15, ACS Symposium Series 237. Am. Chem. Soc., Washington, DC,
1984.
3. Duca, D., Liotta, L. F., and Deganello, G., J. Catal. 154, 69 (1995).
4. Duca, D., Liotta, L. F., and Deganello, G., Catal. Today 24, 15 (1995).
5. Venezia, A. M., Bertoncello, R., and Deganello, G., SIA Surf. Interface
Anal. 23, 239 (1995).
6. Duca, D., Frusteri, F., Parmaliana, A., and Deganello, G., Appl. Catal.
A 146, 269 (1996).
31. Happel, J., and Sellers, P. H., Adv. Catal. 32, 273 (1983).
32. Boitiaux, J. P., Cosyns, J., and Vasudevan, S., in “Preparation of Cata-
lysts. III” (G. Poncelet, P. Grange, and P. A. Jacobs, Eds.), p. 123.
Elsevier, Amsterdam, 1983.
7. Duca, D., Arena, F., Parmaliana, A., and Deganello, G., Appl. Catal.
A, in press.
33. Bertolini, J. C., Delichere, P., Khandra, B. C., Massardier, J., Noupa,
C., and Tardy, B., Catal. Lett. 6, 215 (1990).
8. Floriano, M. A., Venezia, A. M., Deganello, G., Svensson, E. C., and
Root, J. H., J. Appl. Crystallogr. 27, 271 (1994).
9. Venezia, A. M., Floriano, M. A., Deganello, G., and Rossi, A., SIA
Surf. Interface Anal. 18, 532 (1992).
10. Venezia, A. M., Rossi, A., Duca, D., Martorana, A., and Deganello,
G., Appl. Catal. A 125, 113 (1995).
34. Takasu, Y., Unwin, R., Tesche, B., Bradshaw, A. M., and Grunze, M.,
Surf. Sci. 77, 219 (1978).
35. Mason, M. G., Phys. Rev. B 27, 748 (1983).
36. Ryndin, Yu. A., Nosova, L. V., Boronin, A. I., and Chuvilin, A. L.,
Appl. Catal. 42, 31 (1988).
37. Boitiaux, J. P., Cosyns, J., and Vasudevan, S., Appl. Catal. 6, 41 (1983).
11. Venezia, A. M., Duca, D., Floriano, M. A., Deganello, G., and Rossi, 38. Boitiaux, J. P., Cosyns, J., and Vasudevan, S., Appl. Catal. 15, 317
A., SIA Surf. Interface Anal. 18, 619 (1992). (1985).
12. Venezia, A. M., Duca, D., Floriano, M. A., Deganello, G., and Rossi, 39. Hub, S., Hilaire, L., and Touroude, R., Appl. Catal. 36, 307 (1988).
A., SIA Surf. Interface Anal. 19, 543 (1992).
40. Goetz, J., Volpe, M. A., Sica, A. M., Gigola, C. E., and Touroude, R.,
13. Liotta, L. F., Martin, G. A., and Deganello, G., J. Catal. 164, 322 (1996).
J. Catal. 153, 86 (1995).
14. Liotta, L. F., Venezia, A. M., Martorana, A., Rossi, A., and Deganello, 41. Goetz, J., Volpe, M. A., and Touroude, R., J. Catal. 164, 369 (1996).
G., J. Catal. 171, 169 (1997).
42. Wagner, C. D., and Joshi, A., J. Electron Spectros. Relat. Phenom. 47,
283 (1988).
43. Venezia, A. M., Rossi, A., Liotta, L. F., Martorana, A., and Deganello,
G., Appl. Catal. A 147, 81 (1996).
44. Tardy, B., Noupa, C., Leclercq, C., Bertolini, J. C., Hoareau, A.,
Treilleux, M., Faure, J. P., and Nihoul, G., J. Catal. 129, 1 (1991).
45. Ouchaib, T., Massardier, J., and Renouprez, A., J. Catal. 119, 517
(1989).
15. Weiss, A. H., Gambher, B. S., La Pierre, R. B., and Bell, W. K., Ind.
Eng. Chem. Process, Res. Div. 16, 352 (1977).
16. Boitiaux, J. P., Cosins, J., Derrien, M., and Le´ger, P., Hydrocarbon
Process. 64, 51 (1985).
17. Gigola, C. E., Aduriz, H. R., and Bodnariuk, P., Appl. Catal. 27, 133
(1986).
18. Aduriz, H. R., Bodnariuk, P., Dennehy, M., and Cigola, C. E., Appl.
Catal. 58, 227 (1990).
46. Pitchon, V., Gue´nin, M., and Praliaud, H., Appl. Catal. 63, 333 (1990).
19. Rajaram, J., Narula, A. P. S., Chawla, H. P. S., and Sukh Dev, Tetrahe- 47. Liotta, L. F., Deganello, G., Deliche`re, P., Leclercq, C., and Martin,
dron 13, 2315 (1983). G. A., J. Catal. 164, 334 (1996).