Thermodynamics of the hydrodesulfurization (HDS) process

A new method of performing the hydrodesulfurization (HDS) process utilizes a homogeneous catalyst, contrary to the current use of heterogeneous varieties, to extract the sulfur-containing molecules from petroleum feedstocks. The prototype catalyst, shown to the right, is a RuII(NH3)5OH2 complex that is soluble in water.

Environmental concerns regarding the amount of pollution contributed to the atmosphere by combustion engines requires the more efficient removal of organosulfur compounds from petroleum feedstocks. Currently, the desulfurization process utilizes heterogeneous catalyst systems, such as Co-Mo/Al₂O₃ or Ni-Mo/Al₂O₃, to remove the sulfur-containing compounds. Although these catalysts have been successful in removing thiols, thioesters, thiophenes, and benzothiophenes, in a process called hydrodesulfurization (HDS), the catalyst-based processes have been unsuccessful in performing HDS on hindered dibenzothiophenes (4,6-R₂DBT, see Figure 1). The presence of alkyl groups substituted at the 4 and/or 6 positions greatly impedes the interaction between the catalyst and the sulfur. To adhere to the EPA's sulfur-content guidelines⁴, the "deep hydrodesulfurization" of DBTs from fuels is imperative.

Figure 1.	Schematics (left to right) of thiophene, benzothiophene (BT), and 4,6-dibenzothiophene (4,6-R2DBT). R = -H, -CH3, -CH2CH3, etc.

An alternative method utilizes a ruthenium complex, in the absence of catalytic support, in a homogeneous aqueous solution to selectively bind and extract the DBT molecules from petroleum feedstocks.

The simplistic description of the experimental process is to mix the aqueous phase (extracting complex) with the organic phase (sulfur species). This procedure would establish an equilibrium (shown below) that can be used to quantify the affinity that

Figure 2.	Equilibrium geometry for DBT, 2,8-Me2DBT, and 4,6-Me2DBT.

As can be clearly seen, the DBT and the 2,8 complex are very similar in Ru-S bond length and the angle with which the ligand makes with the plane of the NH₃ groups. However, the 4,6 complex clearly exhibits a much sharper angle (approaching 90^o) as well as a slightly longer Ru-S bond length. It is the steric hindrance introduced by substitution at the 4 and 6 positions that causes the heterogeneous catalysts to become ineffective, whereas the homogeneous Ru^II complexes further reduce (deep-HDS) the sulfur content in petroleum feedstocks by about 50%.

The computational methodology that we employ determines the equilibrium geometry using density functional theory (DFT), followed by the vibrational analysis of each specie to insure that it occupies a stationary point. The raw data obtained from theory is then used to calculate the BDE between the metal complex and the ligand, as shown in the steps below.

The focus of this project is to correlate experimentally determined equilibrium ligand affinities to theoretical bond dissociation energies, followed by the extrapolation of our theoretical technique to other metal-centered complexes. An initial study of the DBT derivatives in Figure 2 show that the theoretical BDE values follow the same trend (2,8-Me₂DBT > DBT > 4,6-Me₂DBT) as that of the experimental ligand affinities. Also, owing to the stabilization of the M(II)-S σ bond by the presence of the –CH₃ groups, the BDE of 2,8-DBT is larger than that of the unsubstituted DBT ligand. These trends are consistent when other sulfur-containing organics (thiophene, benzothiophene, etc.) are added to the series. We have seen similar general trends when the identity of the metal has been changed to Fe or Cr, for example. However, when we try to incorporate the BDE values of nitrogen-based organics, the trend begins to fail (i.e. the magnitude of the N-based theoretical BDE within the series is not consistent with its position within the ligand affinity series). This discrepancy will be addressed in our future considerations.

1. Complete vibrational analysis of Co, Rh, Ir, Mo, W, and Os metal complexes.
2. Subsequent BDE comparison of these three complexes to Ru, Fe, Cr.
3. Incorporate solvent effects into theoretical computations.
4. Calculate ΔGrxn values for evaluation/comparison.
5. Perform BDE calculations for these porphyrinic complexes.
6. Compare to M(II)(NH₃)₅(H₂O)²⁺ results.
7. Seek out structural and substitution modifications to further expose/improve predicted binding strength.

UTPB Chemistry 4901 East University Blvd. Odessa, TX 79762 Updated September 26, 2006

www.utpb.edu

chemistry office (432) 552-2235 facsimile (432) 552-2236 email chemistry@utpb.edu College of Arts and Sciences