Estimation Of Enthalpy Of Formation Of Nitrogen - Enriched Materials By Density Functional Theory

Density functional theory (DFT) as one of high effective methods dealing with many-electrons system is widely used and developed in chemistry and physics community for recent decades. Some DFT approaches such as double-hybrid DFT already achieved chemistry accuracy in theoretically estimating thermodynamics properties based on G2 small molecular database at expense of MP2-like costs. However, the calculated errors are usually increase when those molecules have larger atoms or more complex chemical bond types or structures such as ring/cluster. Thus, looking for the suitable methods to balance the accuracy and cost for larger systems play a key role in molecular modelling. In this thesis, I separately applied Gaussian09 package and Dmol3 to calculate the heat of formation for polynitrogen molecules using various DFT method coupled with Pople-type orbitals or numerical basis sets. The main work of this dissertation is following:● The reliability of numerical and Pople basis sets in predictingatomization energy for small moleculesThe calculation errors mainly come from the electron correlation and basis set’s cutoff. It is well known that numerical basis sets showes more rapidly convergence than Pople type orbitals in DFT SCF calculations. However, the global orbital cutoff values for numerical basis set is well dependant to the computational accuracies and efficiencies. In the first study, the effects of global orbital cutoff values and numerical basis set sizes on the theoretical atomization energies (Do) were investigated using density functional theory with the generalized gradient approximation. The cutoff radius-dependant total energies for seven atoms and 44 small molecules demonstrate that the numerical orbital cutoff value should be larger than 6.5A to ensure the converged energetic properties. Comparing the experimental Do of these 44 molecules with the calulated results using different numerical basis sets, DN, DND, DNP, and TNP, We can found that DNP basis set is good enough to predict accurate Do with affordable computational cost. We also compared the ability of numerical and Pople-type basis sets 6-31G,6-31G(d),6-31G(d,p),6-311G(d,p) to accurately describe the atomization energies (Do). The numerical basis sets show similialr behavior compared with their respective near equivalent Pople-type basis sets.● The performance of different DFT on predicting heat of formation in nitrogen-rich many atoms energetic moleculesThe heat of formation for 59 nitrogen-rich molecules were calculated by nine differents functionals (B3LYP, B3PW91, X3LYP, O3LYP, M062X, M052X, M06HF, B3P86, B2PLYP) combined with basis sets cc-PVTZ. The results demonstrated that B3LYP,M062X and B2PLYP methods presented the least deviations from experimental gas phase heat of formations. Among of those three good performance functionals, B2PLYP method was more centralized than others. The mean absolut deviation(MAD) for heat of formation was 7.3 kcal/mol. However, "pure" functionals B3P86 method showed the biggest errors, MAD arrivals 64.2kcal/mol. The calculated deviation of medium-size molecules was larger than that of small-size ones. In addition, G4 model also is applied to calculated the heat of formation of those nitrogen-rich many-atom molecules. Comparing to G4 and density functional theory (DFT) of calculation accuracy for 52 molecules we can found that all MADs for B2PLYP (7.2 kcal/mol), B3LYP (8.5 kcal/mol) and M062X(7.8kcal/mol) are better than that of G4 10.6 kcal/mol. Double hybrid functional B2PLYP is the best model to predice the heat of formation for this many-atom polynitrigon molecules data set.Finally, B2PLYP model are used to calculate heat of formation for some full-nitrigon potential energetic material by atomization energy approch.