for the activation of the CaH bond [74]. Consequently, the cyclotrimerization becomes again more efficient than the hydrogenation of the C 4 H 4 intermediate.
Going to even larger particles or to Pd(111) single crystals the cyclotrimerization to
benzene is selectively catalyzed.
17.3.3
Turn-Over Frequencies of Catalytic Reactions on Supported Clusters
The examples presented above showed that size-selected clusters on surfaces reveal
remarkable size-effects, however, they have been studied so far only by one-cycle
experiments. Thus an experimental proof that these systems are active for catalytic
processes, e.g. several cycles of a catalytic reaction are promoted without destruction of the catalyst, is still missing. In the following we present an experimental
scheme to study catalytic processes of clusters on surfaces with high sensitivity and
we report turn-over frequencies TOFs for the oxidation of CO on size-distributed
supported Pd clusters.
17.3.3.1 A Newly Designed Pulsed Valve for Molecular Beam Experiments
To study catalytic processes of clusters on surfaces with high sensitivity, a newly
designed pulsed valve with excellent pulse-to-pulse stability, in combination with
absolutely calibrated mass spectrometry, was used to determine turn-over frequencies [75].
A systematic view of the relevant elements is depicted in Figure 17.10. The deposited clusters can be exposed to different reactant gases by two kinds of valves.
First, they can be exposed isotropically to e.g. O 2 by a commercial, ultra-high vacuum (UHV) compatible, variable leak valve. Second, reactant molecules (e.g. CO)
can be introduced via a pulsed molecular beam produced by a piezo-electric driven,
pulsed valve. This pulsed valve has a high pulse-to-pulse stability (time profile),
and allows the study of catalytic processes on supported clusters at relatively high
pressures (up to 10
À2 mbar). Furthermore, a stainless steel tube is attached to the
pulsed nozzle in order to collimate the molecular beam and to expose the reactant
molecules to the substrate only. The pulse duration at the position of the sample
can, in principle, be varied from 1 ms up to continuous operation. For the experiments described below a constant pulse duration of about 100 ms was used. The
repetition rate of the pulsed valve can be up to 100 Hz. The experiments were carried out at 0.1 Hz; the 10 s interlude allows the reactant gas to be pumped completely.
After ejecting the CO molecule to the substrate at a constant isotropic pressure
of O 2 (5 Â 10
À7 mbar), the resulting product molecules (CO 2 ) were detected by an
absolutely calibrated quadrupole mass spectrometer. The differentially pumped,
computer-controlled mass spectrometer (Balzers QMG 421) is mounted in line-ofsight with the sample. At the entrance, near the ionizer of the quadrupole mass
spectrometer, a skimmer with an opening of 3 mm is mounted. The skimmer is
biased to À150 V in order to prevent electron-stimulated desorption induced by
emitting electrons from the ionizer of the mass spectrometer. After amplifying
17 Nanocatalysis
578
Going to even larger particles or to Pd(111) single crystals the cyclotrimerization to
benzene is selectively catalyzed.
17.3.3
Turn-Over Frequencies of Catalytic Reactions on Supported Clusters
The examples presented above showed that size-selected clusters on surfaces reveal
remarkable size-effects, however, they have been studied so far only by one-cycle
experiments. Thus an experimental proof that these systems are active for catalytic
processes, e.g. several cycles of a catalytic reaction are promoted without destruction of the catalyst, is still missing. In the following we present an experimental
scheme to study catalytic processes of clusters on surfaces with high sensitivity and
we report turn-over frequencies TOFs for the oxidation of CO on size-distributed
supported Pd clusters.
17.3.3.1 A Newly Designed Pulsed Valve for Molecular Beam Experiments
To study catalytic processes of clusters on surfaces with high sensitivity, a newly
designed pulsed valve with excellent pulse-to-pulse stability, in combination with
absolutely calibrated mass spectrometry, was used to determine turn-over frequencies [75].
A systematic view of the relevant elements is depicted in Figure 17.10. The deposited clusters can be exposed to different reactant gases by two kinds of valves.
First, they can be exposed isotropically to e.g. O 2 by a commercial, ultra-high vacuum (UHV) compatible, variable leak valve. Second, reactant molecules (e.g. CO)
can be introduced via a pulsed molecular beam produced by a piezo-electric driven,
pulsed valve. This pulsed valve has a high pulse-to-pulse stability (time profile),
and allows the study of catalytic processes on supported clusters at relatively high
pressures (up to 10
À2 mbar). Furthermore, a stainless steel tube is attached to the
pulsed nozzle in order to collimate the molecular beam and to expose the reactant
molecules to the substrate only. The pulse duration at the position of the sample
can, in principle, be varied from 1 ms up to continuous operation. For the experiments described below a constant pulse duration of about 100 ms was used. The
repetition rate of the pulsed valve can be up to 100 Hz. The experiments were carried out at 0.1 Hz; the 10 s interlude allows the reactant gas to be pumped completely.
After ejecting the CO molecule to the substrate at a constant isotropic pressure
of O 2 (5 Â 10
À7 mbar), the resulting product molecules (CO 2 ) were detected by an
absolutely calibrated quadrupole mass spectrometer. The differentially pumped,
computer-controlled mass spectrometer (Balzers QMG 421) is mounted in line-ofsight with the sample. At the entrance, near the ionizer of the quadrupole mass
spectrometer, a skimmer with an opening of 3 mm is mounted. The skimmer is
biased to À150 V in order to prevent electron-stimulated desorption induced by
emitting electrons from the ionizer of the mass spectrometer. After amplifying
17 Nanocatalysis
578