The quantum-chemical method selection for modeling of photochemical oxidation of sulfides by organic nitrocompounds

The search for the optimal level of theory has been performed to achieve the best correspondence of calculated values and experimental results for further quantum-chemical modeling of systems composed from sulfides and organic nitrocompounds in the triplet state. The reaction: HS + NO2 → HSO + NO was selected as the model reaction. The standard enthalpy of this reaction calculated theoretically was compared with its experimental value known from literature. Such theories as DFT, UHF and ROHF and the following basis sets: 631++G**, 6-311++G**, aug-cc-pVDZ, were used for computations. XC functionals for DFT calculations are the following: b3lyp, bhlyp, pbe0, xpbe96_cpbe96.


Introduction
Hydrogen sulfide is one of the most common waste products of industrial chemical synthesis.At the moment there are no ways of useful utilization of hydrogen sulfide and its derivatives.The reason of this situation is that there are no experimental base that can be used to produce useful industrial ways of sulfur wastes utilization for its future reuse.
According to some literature information [1] it can be shown that organic nitrocompounds can interact to oxidize sulfur compounds in the triplet state as a result of a light irradiation.It is very difficult to conduct such experiments to study this phenomenon more directly, but we can computationally simulate all these reactions and make some conclusions.
The aim of this research is to find the most suitable level of theory for future quantumchemical calculations of the reactions described above.

Results and Discussion
To achieve our aim we should find a reaction in which sulfur-containing molecule will interact with nitrogroup in an organic compound.Fortunately, we found information concerning the following reaction [2] : The standard enthalpy of this reaction is ΔH f °298 = -26 kcal•mol -1 .Then we calculated the standard enthalpy of the initial and the final states of this system and compared its difference with the standard enthalpy of the reaction.Results are shown in Table 1.All computations were performed by using North-West Chemistry software package [3] .
First we should pay attention to the opposite signs of theoretical deviation values of DFT and HF methods calculation results.We can suppose that it is occurred because of difference in computation mechanism of these methods.But someone would have a question: if atoms of the system are located correctly according to the reaction formula.To verify this we can look at the Figure 1: Figure 1.The initial and the final states of the system.Visualizations are made by using Avogadro software [4] .
No doubts, system geometry is correct.According to the information in the last row of Table 1, computations based on ROHF method are expectedly much more useful than any other computations performed during this research.Its theoretical error value is about 4-5 kcal•mol -1 .But its accuracy is not enough for complete quantum-chemical modeling of our system and all similar systems.
Now, let us examine the results in more details.The half of DFT (bhlyp, xpbe96_cpbe96) method calculation results are extremely unacceptable.For xpbe96_cpbe96 XC functional the error is about 17 kcal•mol -1 , for bhlyp the 2/3 of results are failed to converge but despite this the 6-311++G** basis set brought us one of the least theoretical error in this research: -5.735 kcal•mol -1 which is almost equal to the same values of ROHF's.Another part of DFT functionals (b3lyp, pbe0) is more acceptable than results described above but still is not eligible for our aim.Also we can notice very interesting trend: all the unrestricted methods calculations in this research (this part of DFT's and UHF) have close in absolute value theoretical errors (~12-13 kcal•mol -1 ), this fact is a strong argument for ineligibility of unrestricted methods using in modeling of such open-shell systems in the triplet state.