Cobalt oxides (CoO x ), despite its wide applications, are notoriously complex on structure, in particular under experimental reduction/oxidation conditions. Here, using the newly developed machine learning method, stochastic surface walking global optimization in combination with global neural network potential, we are able to, for the first time, explore the global potential energy surface (PES) of CoO x at different Co/O ratios. Rich information on the thermodynamics and kinetics of CoO x is thus gleaned from more than 10 7 PES data, which helps to resolve the long-standing puzzles on CoO x chemistry. We show that (i) only CoO and Co 3 O 4 are thermodynamically stable compositions in CoO x , whereas Co 3 O 4 is the most stable phase. The trivalent Co 2 O 3 , although having a well-defined global minimum, tends to decompose to Co 3 O 4 and O 2 at finite temperatures. (ii) The solid phase transition between wurtzite CoO (h-CoO) and rock salt CoO (c-CoO) follows the reconstructive phase transition mechanism with a high barrier. Because a high temperature is required for transition, the strong preference of structural defects inside c-CoO instead of h-CoO contributes to the one-way solid-phase transition from h-CoO to c-CoO. (iii) It is c-CoO that can achieve a coherent interface with Co 3 O 4 in forming a biphasic junction, which implies the reversibility of Co 3 O 4 and c-CoO transition under reduction/oxidation conditions. Our results demonstrate the power of global neural network potential in material discovery for fast exploration of polymorphism and transition kinetics and lay the structural foundation for understanding CoO x applications.
Heteronuclear transition metal carbonyl cluster cations FeM(CO)8(+) (M = Co, Ni and Cu) and MCu(CO)7(+) (M = Co and Ni) are produced via a laser vaporization supersonic cluster ion source in the gas phase, which are each mass-selected and studied by infrared photodissociation spectroscopy in the carbonyl stretching frequency region. Their geometric and electronic structures are established by comparison of the experimental spectra with those derived from density functional theoretical calculations. The FeM(CO)8(+) (M = Co, Ni, Cu) complexes are determined to have eclipsed (CO)5Fe-M(CO)3(+) structures, and the MCu(CO)7(+) (M = Co, Ni) ions are characterized to have staggered (CO)4M-Cu(CO)3(+) structures. Natural bonding orbital analysis indicates that the positive charge is mainly distributed on the M(CO)3 fragment. The metal-metal interaction involves an σ-type (OC)4,5M→M(CO)3(+) dative bonding.
Fe-Zn and Co-Zn heteronuclear carbonyl cation complexes are produced via a laser vaporization supersonic cluster source in the gas phase. The dinuclear FeZn(CO) and CoZn(CO) cation complexes are observed to be the most intense heterodinuclear carbonyl cation species in the mass spectra. The infrared spectra are obtained via mass selection and infrared photodissociation spectroscopy in the carbonyl stretching frequency region. Their geometric and electronic structures are assigned with the support of density functional calculations. The FeZn(CO) complex is determined to have a (OC)Fe-Zn structure with a Fe-Zn half bond. The CoZn(CO) ion is established to have a staggered (OC)Co-Zn(CO) structure involving a Co-Zn σ single bond.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.