42
3 Metal Cluster Reacting with Oxygen
from singlet to triplet multiplicity in clusters with filled shells. This spin excitation
is linked to the HOMO–LUMO gap that creates an energy barrier when reacting
with oxygen. In practice, clusters with HOMO–LUMO gaps exceeding 1.2 eV are
found to be generally non-reactive to oxygen [18]. Subsequent investigations of Al
clusters reacting with ground-state molecular oxygen and singlet atomic oxygen have
further explained the reason why the mass abundance of selective clusters could be
enhanced, [19] simply being caused by successive fragmentation of larger clusters
along with the formation of a very stable molecule Al 2 O, written as,
Al
−
n+4 + O 2 → Al
−
n + 2Al 2 O
(3.1)
This reactivity rationalizes that individual clusters could resist to oxygen etching
as dominant peaks and even exhibit increased abundance (such as Al 13
– ) in mass
spectrum due to fragments of larger unstable species. The inertness of magic Al
clusters could be used as models to study the chemisorption of O 2 on Al(111) surface,
which is recognized as a model reaction for surface oxidation and catalysis [20].
Among others, Roach et al. [21] reported a study of oxygen etching on alloy metal
clusters Al n Cu
− , and found that a few Al–Cu clusters with fully-filled subshells
and large HOMO–LUMO gaps were also resistant to oxygen etching. For example,
Al 22 Cu
− was observed as the largest peak in the products (Fig. 3.3), which was
interpreted by its relatively large vertical spin excitation (VSE) energy and HOMO–
LUMO gap, resulting from the geometry distortion which leads to a spliting of the
shells in a spherical jellium, named as a crystal-field-like splitting of the electronic
shells.
Further insights of Al-based cluster reactivity have been attained when examining aluminum-magnesium alloys. Since magnesium is divalent while aluminum
is trivalent, the Al–Mg alloy clusters offer larger variations over the electron counts
than pure aluminum clusters and hence provide more rigorous grounds to investigate the underlying principles of oxygen etching effect. As expected, Al 5 Mg 2
− and
Al 11 Mg 3
− which correspond to magic numbers of 20 and 40 electrons are found to
exhibit reasonable stability (Fig. 3.4); but Al 7 Mg 3
− , Al 11 Mg
− and Al 11 Mg 2
− with
electron counts of 28, 36, and 38 respectively displayed unexpected stability. The
stabilities of these non-magic numbers of Al n Mg m
− clusters can be understood via
a crystal-field-like splitting of degenerated shells due to the geometrical distortions
of the clusters. These studies reinforced the importance of near-free electron gas
(NFEG) model (with closed shell n = 2, 8, 20, 40…) in rationalizing electronic
structure and stability of metal clusters. At the same time, also raised is a pending
question regarding the development of a comprehensive model for metal clusters,
that could account for “magic numbers” according to the shell model (like noble gas),
but also can successfully predict stability of all metal clusters where shell model fails.
The study of magic numbers in Al–Mg clusters and their inertness towards oxygen
is also important in industrial anticorrosion and aerospace manufacturing [22].
Similar to the drastic etching effect observed for aluminum and Al–Mg clusters,
the chemical reactivity of cobalt cluster anions Co n
− (n = 2–8) toward O 2 by flow
tube reactor was also found to have rapid rate coefficients, leading to fragmentation
3 Metal Cluster Reacting with Oxygen
from singlet to triplet multiplicity in clusters with filled shells. This spin excitation
is linked to the HOMO–LUMO gap that creates an energy barrier when reacting
with oxygen. In practice, clusters with HOMO–LUMO gaps exceeding 1.2 eV are
found to be generally non-reactive to oxygen [18]. Subsequent investigations of Al
clusters reacting with ground-state molecular oxygen and singlet atomic oxygen have
further explained the reason why the mass abundance of selective clusters could be
enhanced, [19] simply being caused by successive fragmentation of larger clusters
along with the formation of a very stable molecule Al 2 O, written as,
Al
−
n+4 + O 2 → Al
−
n + 2Al 2 O
(3.1)
This reactivity rationalizes that individual clusters could resist to oxygen etching
as dominant peaks and even exhibit increased abundance (such as Al 13
– ) in mass
spectrum due to fragments of larger unstable species. The inertness of magic Al
clusters could be used as models to study the chemisorption of O 2 on Al(111) surface,
which is recognized as a model reaction for surface oxidation and catalysis [20].
Among others, Roach et al. [21] reported a study of oxygen etching on alloy metal
clusters Al n Cu
− , and found that a few Al–Cu clusters with fully-filled subshells
and large HOMO–LUMO gaps were also resistant to oxygen etching. For example,
Al 22 Cu
− was observed as the largest peak in the products (Fig. 3.3), which was
interpreted by its relatively large vertical spin excitation (VSE) energy and HOMO–
LUMO gap, resulting from the geometry distortion which leads to a spliting of the
shells in a spherical jellium, named as a crystal-field-like splitting of the electronic
shells.
Further insights of Al-based cluster reactivity have been attained when examining aluminum-magnesium alloys. Since magnesium is divalent while aluminum
is trivalent, the Al–Mg alloy clusters offer larger variations over the electron counts
than pure aluminum clusters and hence provide more rigorous grounds to investigate the underlying principles of oxygen etching effect. As expected, Al 5 Mg 2
− and
Al 11 Mg 3
− which correspond to magic numbers of 20 and 40 electrons are found to
exhibit reasonable stability (Fig. 3.4); but Al 7 Mg 3
− , Al 11 Mg
− and Al 11 Mg 2
− with
electron counts of 28, 36, and 38 respectively displayed unexpected stability. The
stabilities of these non-magic numbers of Al n Mg m
− clusters can be understood via
a crystal-field-like splitting of degenerated shells due to the geometrical distortions
of the clusters. These studies reinforced the importance of near-free electron gas
(NFEG) model (with closed shell n = 2, 8, 20, 40…) in rationalizing electronic
structure and stability of metal clusters. At the same time, also raised is a pending
question regarding the development of a comprehensive model for metal clusters,
that could account for “magic numbers” according to the shell model (like noble gas),
but also can successfully predict stability of all metal clusters where shell model fails.
The study of magic numbers in Al–Mg clusters and their inertness towards oxygen
is also important in industrial anticorrosion and aerospace manufacturing [22].
Similar to the drastic etching effect observed for aluminum and Al–Mg clusters,
the chemical reactivity of cobalt cluster anions Co n
− (n = 2–8) toward O 2 by flow
tube reactor was also found to have rapid rate coefficients, leading to fragmentation