ABSTRACTThermodynamic modeling of the MOCVD process, using the standard free energy minimization algorithm, cannot always explain the deposition of hybrid films that occurs. The present investigation explores a modification of the procedure to account for the observed simultaneous deposition of metallic iron, Fe3O4, and carbon nanotubes from a single precursor. Such composite films have potential application in various device architectures and sensors, and are being studied as electrode material in energy storage devices such as lithium ion batteries and supercapacitors.With ferric acetylacetonate [Fe(acac)3] as the precursor, MOCVD in argon ambient results in a nanocomposite of CNT, Fe, and Fe3O4 (characterized by XRD and Raman spectroscopy) when growth temperature T and total reactor pressure P are in the range from 600°C-800°C and 5-30 torr, respectively. No previous report could be found on the single-step formation of a CNT-metal-metal oxide composite. Equilibrium thermodynamic modeling using available software predicts the deposition of only Fe3C and carbon, without any co-deposition of Fe and Fe3O4, in contrast with experimental observations. To reconcile this contradiction, the modeling of the process was approached by taking the molecular structure of the precursor into account, whereas “standard” thermodynamic simulations are restricted to the total number of atoms of each element in the reactant(s) as the input. When Ocon (statistical average of the oxygen atom(s) taken up by each metal atom during CVD) is restricted to lie between 0 and 1, thermodynamic computations predict simultaneous deposition of FeO1-x, Fe3C, Fe3O4 and C in the inert ambient. At high temperature and in a carbon-rich atmosphere, iron carbide decomposes to iron and carbon. Furthermore, FeO1-x yields Fe and Fe3O4 when cooled below 567°C. Therefore, the resulting film would be composed of Fe3O4, Fe and C, in agreement with experiment. The weight percentage of carbon (∼40%) calculated from thermodynamic analysis matches well with experimental data from TG-DTA.