Cerium oxide cluster cations (Ce(m)O(n)(+), m = 2-16; n = 2m, 2m +/- 1 and 2m +/- 2) are prepared by laser ablation and reacted with carbon monoxide (CO) and small hydrocarbon molecules (CH(4), C(2)H(4), and C(2)H(6)) in a fast flow reactor. A time of flight mass spectrometer is used to detect the cluster distribution before and after the reactions. The observation of oxygen reduction and hydrogen pickup of Ce(m)O(2m)(+) clusters strongly suggests the following reactions: (1) Ce(m)O(2m)(+) + C(2)H(4) --> Ce(m)O(2m-1)(+) + C(2)H(4)O (m = 2-6); (2) Ce(m)O(2m)(+) + CO --> Ce(m)O(2m-1)(+) + CO(2) (m = 4-6); and (3) Ce(m)O(2m)(+) + CH(4)/C(2)H(6) --> Ce(m)O(2m)H(+) + CH(3)/C(2)H(5) (m = 2-4). Density functional theory (DFT) calculations are performed to study reaction mechanisms of Ce(2)O(4)(+) + X (X = CO, CH(4), C(2)H(4), and C(2)H(6)). The calculated results are in good agreement with the experimental observations. The structural and bonding properties of Ce(m)O(2m)(+) (m = 2-5) clusters are also investigated by the DFT calculations. The unpaired electron in each of the clusters is mainly distributed over one Ce atom (4f and 5p orbitals) and two O atoms (2p orbital) in a CeO(2) moiety, which can be considered as the active site in the cluster. To further understand the nature of the active sites in Ce(m)O(2m)(+) clusters, the fast flow reaction experiments are also carried out on zirconium oxide clusters Zr(m)O(n)(+), because both Zr ([Kr]4d(2)5s(2)) and Ce ([Xe]4f(1)5d(1)6s(2)) have the same number of valence electrons while the latter has one more f and one less d electrons. In addition to the oxygen transfer reactions such as Zr(m)O(2m)(+) + C(2)H(4) --> Zr(m)O(2m-1)(+) + C(2)H(4)O (m = 1-4) reported in the literature, hydrogen abstraction reactions Zr(m)O(2m)(+) + CH(4)/C(2)H(6) --> Zr(m)O(2m)H(+) + CH(3)/C(2)H(5) are also identified. The rate constants of CO oxidation as well as hydrogen abstraction by Ce(m)O(2m)(+) and Zr(m)O(2m)(+) are very different. The reactivity and selectivity of Ce(m)O(2m)(+) versus Zr(m)O(2m)(+) can be well rationalized based on the DFT calculations. The oxygen transfer and hydrogen abstraction reactions studied in this work are of widespread importance. The nature of the active sites of Ce(m)O(2m)(+) clusters is unique and may be considered in the use and design of cerium oxide based catalysts.

Saturated hydrocarbons, or alkanes, are major constituents of natural gas and oil. Directly transforming alkanes into more complex organic compounds is a value-adding process, but the task is very difficult to achieve, especially at low temperature. Alkanes can react at high temperature, but these reactions (with oxygen, for example) are difficult to control and usually proceed to carbon dioxide and water, the thermodynamically stable byproducts. Consequently, a great deal of research effort has been focused on generating and studying chemical entities that are able to react with alkanes or efficiently activate C-H bonds at lower temperatures, preferably room temperature. To identify low-temperature methods of C-H bond activation, researchers have investigated free radicals, that is, species with open-shell electronic structures. Oxygen-centered radicals are typical of the open-shell species that naturally occur in atmospheric, chemical, and biological systems. In this Account, we survey atomic clusters that contain oxygen-centered radicals (O(-•)), with an emphasis on radical generation and reaction with alkanes near room temperature. Atomic clusters are an intermediate state of matter, situated between isolated atoms and condensed-phase materials. Atomic clusters containing the O(-•) moiety have generated promising results for low-temperature C-H bond activation. After a brief introduction to the experimental methods and the compositions of atomic clusters that contain O(-•) radicals, we focus on two important factors that can dramatically influence C-H bond activation. The first factor is spin. The O(-•)-containing clusters have unpaired spin density distributions over the oxygen atoms. We show that the nature of the unpaired spin density distribution, such as localization and delocalization within the clusters, heavily influences the reactivity of O(-•) radicals in C-H bond activation. The second factor is charge. The O(-•)-containing clusters can be negatively charged, positively charged, or neutral overall. We discuss how the charge state may influence C-H bond activation. Moreover, for a given charge state, such as the cationic state, it can be demonstrated that local charge distribution around the O(-•) centers can also significantly change the reactivity in C-H bond activation. Through judicious synthetic choices, spin and charge can be readily controllable physical quantities in atomic clusters. The adjustment of these two properties can impact C-H bond activation, thus constituting an important consideration in the rational design of catalysts for practical alkane transformations.