Method for predicting adsorbent performance

Abstract

A method for predicting selective performance of an adsorbent is disclosed. The adsorbent is selected from a list of metals and/or metal cations for use in removing contaminants as thiophene derivatives in hydrocarbon feed. The metals or metal cations are identified from a list having a positive value for E.sub.rel, wherein the metal or metal cation having the largest value for E.sub.rel is the most selective adsorbent.

Claims

1. A method of predicting selective performance of an adsorbent from a list of metals and/or metal cations for use in selectively removing thiophene and thiophene derivatives in a hydrocarbon feed containing aromatics having a similar aromaticity or a same ring number corresponding to the thiophene and thiophene derivatives comprising: identifying the thiophene and thiophene derivatives to be removed from the hydrocarbon feed as contaminants; identifying aromatics in the hydrocarbon feed having said similar aromaticity or the same ring number corresponding to the identified thiophene and thiophene derivatives; computing an interaction of each of the metals and/or metal cations on the list with each of the thiophene and thiophene derivatives as a first binding energy; computing an interaction of the metals and/or metal cations with one or more aromatic compounds having said similar aromaticity or the same ring number as a second binding energy; subtracting the first binding energy from the second binding energy to get a value E.sub.rel for each of the metals and/or metal cations and each identified pair of thiophene and thiophene derivative and corresponding aromatic compound; identifying a selective adsorbent from the metals and/or metal cations having a positive value for E.sub.rel, wherein the metal or metal cation having the largest positive value for E.sub.rel, is the most selective adsorbent, and contacting said selective adsorbent with the hydrocarbon feed containing thiophene, thiophene derivatives, and other arenes, thereby selectively adsorbing said thiophene and thiophene derivatives.

2. A method of tailoring an adsorbent to a hydrocarbon feed for selectively removing thiophene and thiophene derivatives wherein the hydrocarbon feed contains aromatics having a similar aromaticity or a same ring number corresponding to the thiophene and the thiophene derivatives, comprising the steps of: identifying one or more thiophene and thiophene derivatives in the hydrocarbon feed; identifying aromatics in the hydrocarbon feed having said similar aromaticity or the same ring number corresponding to the identified thiophene and thiophene derivatives; identifying a plurality of metals and/or metal cations for consideration as candidates for the adsorbent; computing an interaction of each the identified metals and/or metal cations with each of the identified thiophene and thiophene derivatives as a first binding energy; computing an interaction of each the identified metals and/or metal cations with each of the identified aromatics as a second binding energy; subtracting the first binding energy from the second binding energy to compute a binding energy difference as E.sub.rel; identifying at least a metal or a metal cation having a positive value for E.sub.rel, contacting said adsorbent with the hydrocarbon feed containing thiophene, thiophene derivatives, and other arenes, thereby selectively adsorbing said thiophene and thiophene derivatives.

3. The method of claim 2, further comprising selecting the metal or metal cation having the largest positive value of E.sub.rel as a component for the adsorbent.

4. A method of enhancing selective performance of an adsorbent comprising at least a metal or a metal cation in selectively removing thiophene and thiophene derivatives in a hydrocarbon feed containing aromatics having a similar aromaticity or a same ring number corresponding to the thiophene and thiophene derivatives, comprising the steps of: providing a list of compounds for use as chelating agents for the metal or the metal cation to form chelated metals or chelated metal cations; identifying thiophene and thiophene derivatives to be removed from the hydrocarbon feed as contaminants; identifying aromatics in the hydrocarbon feed having said similar aromaticity or the same ring number corresponding to the identified thiophene and thiophene derivatives; computing an interaction of each of the chelated metals or chelated metal cations on the list of compounds with each of the thiophene and thiophene derivatives as a first binding energy; computing an interaction of each of the chelated metals or chelated metal cations with one or more aromatic compounds having said similar aromaticity or the same ring number as a second binding energy; subtracting the first binding energy from the second binding energy to get a value E.sub.rel for each of the chelated metals or chelated metal cations and each identified pair of thiophene and thiophene derivative and corresponding aromatic compound; identifying a selective adsorbent from the chelated metals or chelated metal cations having a positive value for E.sub.rel, wherein the chelated metal or chelated metal cation having the largest positive value for E.sub.rel is the most selective adsorbent, and contacting said selective adsorbent with a hydrocarbon feed containing thiophene, thiophene derivatives, and other arenes, thereby selectively adsorbing said thiophene and thiophene derivatives.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1 shows the HOMO electron density surfaces for the interactions of Cu.sup.2+ and Hg.sup.2+ with naphthalene.

(2) FIG. 2 shows the electron density surfaces for HOMO for the interactions of Zn.sup.2+ and Pd.sup.2+ with naphthalene and electron density surfaces for the LUMO for the interaction of Pd.sup.2+ with naphthalene.

(3) FIG. 3 shows the electron density surfaces the HOMO for the interactions of Co.sup.2+, Hg.sup.2+, Pd.sup.0, and Au.sup.0 with naphthalene.

(4) FIG. 4 shows the electron density surfaces of the HOMO (0.06 eV/Å.sup.3) for interaction with Au.sup.0, Hg.sup.0, Cu.sup.0, and Ag.sup.0 with naphthene.

(5) FIG. 5 is a plot of E.sub.rel, the difference between the binding energy of arenes and thiophene derivatives, as a function of effective partial charge on Au.

(6) FIG. 6 shows the electron density surfaces of the HOMO for [Au(N(CH.sub.3).sub.3)].sup.+ complexes with benzothiophene and dibenzothiophene.

(7) FIG. 7 shows the electron density surfaces of the HOMO ((0.06 eV/Å.sup.3) for Au.sup.0 interactions with benzothiophene, dibenzothiophene, and 4,6-dimethyl dibenzothiophene.

(8) FIG. 8 shows the electron density surfaces of the HOMO for interactions of Ag.sup.+ with benzothiophene and dibenzothiophene.

DETAILED DESCRIPTION

(9) In one embodiment, the invention relates to a method to predict the selective performance of metals and/or metal cations in the removal of thiophene derivatives from hydrocarbon products, e.g., vacuum gas oils, kerosene and diesel fuels, particularly hydrocarbon fuels containing thiophene derivatives.

(10) In one embodiment, the selective performance of an adsorbent can be predicted by first carrying out quantum chemical calculations of the binding energy of naphthalene, anthracene, benzothiophene, dibenzothiophene, and 4,6-dimethyl dibenzothiophene for a variety of metal atoms and cations. In the next step, the results are analyzed in order to determine the principles by which to select adsorption centers exhibiting preferential binding of the thiopene derivative over arenes in the hydrocarbon feed having similar aromaticity. What was found is that thiophene derivatives are bound preferentially to arenes on metal cations having low Lewis acidity and high chemical hardness. The difference in the binding energies between arene and thiophene derivative can be optimized by controlling the effective charge on the cation.

(11) In one embodiment, the calculations can be performed at the DFT/B3LYP (DFT: density functional theory; B3LYP: Becke 3-Parameter) level using a Gaussian 03 package. Metal atoms and cations can be described using the LANL2DZ (Los Alamos National Laboratory 2-double-z (density functional theory)) basis set, whereas ligands (CO, N(CH.sub.3).sub.3, P(CH.sub.3).sub.3, pyridine) and adsorbates (naphthalene (NPT), anthracene (ANT), benzothiophene (BTP), dibenzothiophene (DBT) 4,6-dimethyl dibenzothiophene (DMDBT)) can be described using cc-pVQZ basis sets (cc-pVQZ: correlation-consistent polarized valence quadruple zeta).

(12) The partial atomic charges of adsorbates can be obtained using RESP (Restrained Electrostatic Potential) in Antechamber tools of Amber 11. The strength of metal atoms or cation interaction with adsorbates can be interpreted in terms of the electron affinity, χ, and hardness, η, of the interacting species, and by the extent of orbital overlap between the interacting species. The electron affinity and chemical hardness can be calculated from the following relationships:

(13) $\chi =-\frac{\left({\mathrm{.Math.}}_{\mathrm{HOMO}}+{\mathrm{.Math.}}_{\mathrm{LUMO}}\right)}{2}$$\eta =\frac{\left({\mathrm{.Math.}}_{\mathrm{HOMO}}-{\mathrm{.Math.}}_{\mathrm{LUMO}}\right)}{2}$

(14) In the above equations, and represent the energy levels of the highest occupied molecular orbital and the lowest unoccupied molecular orbital, respectively. Electron density surfaces of the HOMO (Highest Occupied Molecular Orbital) and LUMO (Lowest Unoccupied Molecular Orbital) states can be obtained using VESTA. All the HOMO and LUMO electron density surfaces are reported for an isovalue 0.02 (electrons/A.sup.3). The blue and brown colors of the isosurfaces represent lobes of positive and negative phase wave function, respectively. The extent of overlap between the electron density isosurfaces associated with the metal and the adsorbate provides information about the character of the metal-adsorbate bonding. Constructive overlap between metal and adsorbate orbitals (overlap of orbital lobes with similar phase) is characteristic of covalent bonds or pi back bonding. For the systems examined, this condition was observed for transition metals with partially filled d orbitals and low chemical hardness (e.g., Rh.sup.3+, Pd.sup.2+, Ni.sup.0(LS)). A negligible overlap between metal and ligand orbitals is characteristic of charge controlled interactions, which can be strong or weak depending on the charge of the interacting species.

EXAMPLES

(15) The following illustrative examples are intended to be non-limiting. In the examples, a number of adsorbates were considered. Values of electronegativity χ and chemical hardness η for each of the adsorbates considered are listed in Table 1.

(16) TABLE-US-00001 TABLE 1 Adsorbate Electronegativity, χ (eV) Chemical Hardness, η (eV) BTP 0.1314 0.0945 NPT 0.1372 0.0879 4,6-DMDBT 0.1309 0.0870 DBT 0.1362 0.0864 ANT 0.1388 0.0651

(17) The χ values or the HOMO-LUMO gap calculated from DFT studies and known semi empirical methods differ. However, the values from semi empirical calculations are in agreement with quantum chemical calculations. The results (NPT and ANT; 4.78 and 3.54 eV) are in agreement with the η values reported in the literature. The values of χ for all of the adsorbates lie between 0.13 and 0.14 eV. Since strong Lewis acids are characterized by χ≧1.0 eV, the arene and thiophenes considered here are relatively weak Lewis acids, and correspondingly relative strong Lewis bases. A greater variation occurs in the values of η, arenes (NPT and ANT) being somewhat softer Lewis bases than thiophenes (BTP, DBT, and DMDBT). As observed, with increasing ring number, both the arenes and thiophenes become softer Lewis bases, reflecting the increase in molecular polarizability with increasing aromaticity of the system.

(18) The quantum calculations also show that the electronic charge on the sulfur atom in the thiophene ring increases with increasing number of benzene rings attached to the thiophene ring and the presence of electron-donating methyl groups on the benzene rings. The charge on the S atom of thiophene (TPH) is positive (qs=0.038|e|) but becomes negative for all benzothiophene and dibenzothiophene derivatives, decreasing in the order BTP (−0.039|e|)>DBT (−0.136|e|)>4,6-DMDBT (−0.158|e|). It is noted that the carbon atoms have larger negative partial charges than the sulfur atom in case of BTP and DBT, but the opposite is true in the case of DMDBT due to presence of electron donating methyl groups.

(19) Table 2 lists the values of electronegativity χ and chemical hardness η for each of the metal atoms or cations investigated and the absolute binding and relative energies for the interactions of these adsorbents with NPT and BTP.

(20) TABLE-US-00002 TABLE 2 Chemical Electro- E.sub.bind E.sub.bind hardness negativity (M-NPT), (M-BTP), E.sub.rel, Metal η (eV) χ (eV) kJ/mol kJ/mol kJ/mol Rh.sup.3+(LS) 0.12 2.96 −2609.5 −2573.4 −36.07 Zn.sup.2+ 0.6 2.06 −2553.5 −2553.6 0.1 Pd.sup.2+(LS) 0.05 2.03 −1229 −1225.7 −3.28 Cu.sup.2+ 0.45 1.93 −1085.4 −1079.8 −5.6 Co.sup.2+(HS) 0.37 1.78 −845.42 −833.39 −12.03 Hg.sup.2+ 0.44 1.94 −715 −699.7 −15.3 Au.sup.+1 0.17 1.06 −294.29 −245.69 −48.6 Cu.sup.+1 0.19 0.95 −233.67 −235.29 1.62 Ni.sup.0(LS) 0.06 0.3 −202.99 −157.13 −45.86 Ag.sup.+1 0.26 1.01 −165.4 −168.4 3 Pd.sup.0 0.09 0.26 −98.6 −76.66 −21.94 Au.sup.0 0.22 0.27 −14.67 −25.04 10.37 Cu.sup.0 0.17 0.21 −21.56 −13.62 −7.94 Ag.sup.0 0.17 0.21 −3.18 −3.12 −0.06 Hg.sup.0 0.25 0.22 −0.0289 −0.032 0.0031

(21) In general, the strength of adsorbent-adsorbate interaction decreases with decreasing strength of Lewis acidicity (χ) and for equivalent, or nearly equivalent, values of χ with a decreasing value of η. Thus, for χ.sub.1≈χ.sub.2 and η.sub.1<η.sub.2, the strength of adsorbate binding will be higher for metal 1 than metal 2, i.e., E.sub.1<E.sub.2. This trend is exactly the trend that would be predicted by Pearson's rules for the interactions of acids and bases. However, a number of anomalies to this trend are observed. For example, when χ.sub.1≈χ.sub.2 and η.sub.1≈η.sub.2, E.sub.1<E.sub.2, as is the case for Cu.sup.2+ compared to Hg.sup.2+. The reason for the higher interaction energy for Cu.sup.2+ can be ascribed to the degree of constructive overlap between orbitals of the metal and the adsorbate. The t.sub.2g-type d orbitals of Cu.sup.2+ interact with a larger fraction of the electron-density surface of NPT than does the s orbital of Hg.sup.2+, as can be seen in FIG. 1.

(22) Anomalies exist, as well, for two other cases. The first is when χ.sub.1≈χ.sub.2 and η.sub.1>η.sub.2, as is the case for Zn.sup.2+ compared to Pd.sup.2+. The second is when χ.sub.1<χ.sub.2 and η.sub.1<η.sub.2, as is the case for Co.sup.2+ compared to Hg.sup.2+, Ni.sup.0 compared to Ag.sup.+, Pd.sup.0 compared to Au.sup.0, and Cu.sup.0 compared to Au.sup.0. In each of these two cases, the higher binding energy occurs for the first member of the pair. The explanations of the anomalies for the two cases are different. For the first case, exemplified by Pd.sup.2+ and Zn.sup.2+, Pd.sup.2+ has a lower chemical hardness than Zn.sup.2+ and hence should have better overlap of metal-ligand orbitals. This is, indeed, observed in the electron density plot of the HOMO of the Pd.sup.2+-NPT complex (see FIG. 2). However, in the electron density plot of the LUMO, this overlap is offset as a consequence of the overlap of orbitals lobes of opposite phases. The overlap causes a reduction in the overall pi-back bonding in this case. For this reason Pd.sup.2+ binding energy is lower than that for Zn.sup.2+. Zn.sup.2+ does not have orbital overlap with NPT, due to its high chemical hardness and consequently only charge-controlled interactions with NPT occur. For the second case, the first metal atom or cation is less electronegative than the second one, but softer.

(23) The higher binding energy for the first metal atom or cation is attributable to the greater orbital overlap between the atomic orbitals of the metal atom (cation), and the molecular orbitals of NPT. This effect is seen in the electron density plot shown in FIG. 3 for the HOMO of Co.sup.2+ compared to Hg.sup.2+, and Pd.sup.0 compared to Au.sup.0. Thus, the extended rules for anticipating strong metal NPT interactions are high Lewis acidity (χ), low chemical hardness (η), and a high degree of overlap between the orbitals of the metal atom (or cation) and the adsorbate. While these rules were inferred from an examination of the binding energies for NPT, a similar pattern can also be deduced for BTP.

(24) Inspection of Table 2 reveals a second interesting feature. While the Lewis acidities and chemical hardnesses of NPT and BTP are very similar, the binding energies for these two molecules can differ considerably. Therefore, the small differences in the values of χ and η for BTP and NPT cannot account for the differences between the metal-adsorbate binding energies for NPT and BTP, as listed in Table 2.

(25) Stronger binding for BTP relative to NPT (e.g. E.sub.rel>0, where E.sub.rel is difference in binding energy of NPT and BTP complex) occurs when the metal atom or cation has a high chemical hardness relative to that of metal atom or cations of comparable Lewis acidity. This point is nicely illustrated by comparing the values of E.sub.rel for Au.sup.0 and Pd.sup.0, Au.sup.0 and Au.sup.−1, Ag.sup.+ and Au.sup.+ or Cu.sup.+ and Au.sup.+ for which the first member of the pair exhibits a higher binding energy for BTP than NPT. The difference between the binding energy of the BTP and NPT also increases when the HOMO of the metal atom (or cation) is similar in shape to the p orbitals of sulfur. The point is best illustrated by comparing Au.sup.0 and Hg.sup.0, for which η.sub.2>η.sub.1. The first member of the pair has a higher value of E.sub.rel, due to presence of e.sub.g-type d orbitals that are shaped similarly to the p orbitals of S (see FIG. 4). The reason why the difference in binding energies for BTP versus NPT is so large for Au.sup.0 relative to that for Cu.sup.0 and Ag.sup.0 is that the d orbitals of these latter metals do not participate in the binding to S, as can be seen from FIG. 4. Interestingly, the main group elements Li.sup.+, K.sup.+ and Mg.sup.2+ in Table 3 have lower binding energies and negative values Erel compared to transition metals of similar χ. These differences are probably due to the absence of d orbitals for these cations. Thus, metals with high η, low χ, and eg-type d orbitals are the best candidates for strong binding to the S atom of thiophenes.

(26) TABLE-US-00003 TABLE 3 Binding energies, kJ/mol of Li.sup.+, K.sup.+ and Mg.sup.2+. Electro- Chemical E.sub.bind E.sub.bind negativity, Hardness, (M-NPT), (M-BTP), E.sub.rel, Metal χ (eV) η (eV) (kJ/mol) (kJ/mol) (kJ/mol) Li.sup.+ 1.30 1.05 −163.69 −148.52 −15.17 K.sup.+ 0.60 0.39 −68.62 −63.09 −5.53 Mg.sup.2+ 0.26 0.26 −531.83 −499.85 −31.98

(27) The relationship between the binding energies of thiophene derivatives and arenes with the same number of aromatic rings was extended to DBT and ANT, and to DMDBT and ANT. The results of these calculations are presented in Table 4, showing the binding energy of Au in various oxidation states with adsorbates.

(28) TABLE-US-00004 TABLE 4 E.sub.bind (4,6- Metal q.sub.M E.sub.bind(NPT) E.sub.bind(BTP) E.sub.bind(ANT) E.sub.bind(DBT) DMDBT) Au.sup.−1 −1 −34.73 −29.86 — — — Au.sup.0 0 −14.67 −25.04 −16.38 −28.47 −32.50 [Au(P(CH.sub.3).sub.3].sup.+ 0.56 −136.67 −149.47 −140.97 — −165.20 [Au(N(CH.sub.3).sub.3].sup.+ 0.68 −192.36 −194.19 −203.59 −224.17 −229.19 [Au(NC.sub.5H.sub.5)].sup.+ 0.70 −201.09 −219.10 −211.98 — −237.19 [Au(CO)].sup.+ 0.88 −252.83 −262.71 −392.09 — −412.86 Au.sup.+ +1 −294.29 −245.69 −319.90 −320.02 −334.90 Cu.sup.+ +1 −233.67 −235.29 −249.64 −220.28 −254.87 Ag.sup.+ +1 −165.40 −168.40 −183.80 −173.70 −195.01 Cu.sup.2+ +2 −1085.40 −1079.80 −1223.80 — −1182.38 Mg.sup.2+ +2 −531.83 −499.85 −535.35 — −514.29 K.sup.+ +1 −68.62 −63.09 −74.66 — −75.44

(29) It is observed first of all that for Au.sup.0 the binding energy for ANT is slightly higher than that of NPT. Since ANT and NPT have very similar Lewis acidities, the observed preference for binding ANT is due to its lower chemical hardness, which enhances the ability of this molecule to have better orbital overlap with Au.sup.0. The same effect accounts for the increasing binding strength as one progresses from BTP to DBT to DMDBT. Similar patterns in the absolute binding energies of ANT compared to NPT and the progressively stronger binding energy of the thiophene derivatives with increasing number of benzene rings and methyl groups can be seen for Au.sup.+, except that the differences in all cases are larger. This is very likely a consequence of the higher Lewis acidity of Au.sup.+ compared to Au.sup.0.

(30) Table 4 also shows that for a given oxidation state of Au, the difference between the binding energies for arene versus the corresponding thiophene derivative increases (becomes more positive) with increasing negative charge on the S atom of the thiophene derivative. This effect is particularly dramatic for interactions with Au.sup.+, for which the difference in binding energies is negative for NPT compared to BTP, is nearly zero for ANT compared to DBT, and is positive for ANT compared to DMDBT. What these trends mean is that Au.sup.0 will preferentially interact with all thiophene derivatives and that Au.sup.+ will interact preferentially with DMDBT relative to NPT, ANT, BTP, or DBT. Similar trends are observed for Cu.sup.+ and Ag.sup.+, which show higher binding energy for BTP and DMDBT than NPT and ANT but lower binding energy for DBT than ANT.

(31) A higher relative binding energy for DMDBT versus ANT, but not for DBT versus ANT or BTP versus NPT, was also observed for other metals with intermediate Lewis acidity Cu.sup.+, Ag.sup.+ and K.sup.+. However, this effect was not seen for more Lewis acidic metals, such as Cu.sup.2+ and Mg.sup.2+. DMDBT is an exception because the methyl side groups cause the negative charge on sulfur to be larger than that on the carbon atoms, resulting in a larger binding energy for DMDBT than ANT with metals of intermediate Lewis acidity. Additional selectivity towards the coordination of thiophene derivatives relative to the corresponding arenes can be achieved by tuning the charge on Au by allowing Au.sup.+ to interact with a single ligand.

(32) Table 4 and FIG. 5 show the difference between the binding energies of arenes and thiophene derivatives, E.sub.rel, having the same number of aromatic rings. E.sub.rel becomes more positive going from Au.sup.0 to Au(P(CH.sub.3).sub.3).sup.+(q.sub.eff (Au)=0.54|e|), less positive for Au(N(CH.sub.3).sub.3).sup.+(q.sub.eff(Au)=0.67|e|), becomes strongly positive for Au(NC.sub.5H.sub.5).sup.+(q.sub.eff(Au)=0.70|e|), becomes slightly less positive for Au(CO).sup.+(q.sub.eff (Au)=0.88|e|), and then becomes strongly negative for Au.sup.+. The decrease in the difference of binding energies between NPT and BTP for Au(N(CH.sub.3).sub.3).sup.+ is due to the interaction of Au(N(CH.sub.3).sub.3).sup.+ with the C.sub.7 of BTP instead of the sulfur atom which reduces selectivity for BTP, as shown in the HOMO LUMO electron density surfaces in FIG. 6. By contrast, the difference in binding energies for ANT compared to DMDBT is always positive but reaches a maximum for Au(P(CH.sub.3).sub.3).sup.+ and Au(N(CH.sub.3).sub.3).sup.+.

(33) It should be noted that the absolute binding energies for all substrates increase monotonically as the charge on Au increases from 0 to +1, consistent with the progressively higher Lewis acidity of the metal. The variation in the difference in binding energies between ANT and DBT follows a pattern very similar to that seen for the difference in binding energies between ANT and DMDBT, but the magnitude of the difference is smaller. An important observation is that Au.sup.0 and [AuL].sup.+ are unique in that these species interact selectively with the S atom of all the thiophenes examined FIG. 7 clearly shows this to be the case for Au.sup.0 complexes.

(34) The results demonstrate that preferential adsorption of thiophene derivatives over arenes is achieved by using a metal cation that is a weak Lewis acid and with a high chemical hardness and atomic projected orbitals, having p character and, hence, good overlap well the p orbitals of sulfur. By contrast, soft Lewis acids increase the degree of i-back bonding, leading to preferred binding of arenes. Zero valent Au meets the specifications for selective preferential binding of thiophene derivatives versus arenes containing the same number of aromatic rings. On the other hand, Au.sup.+ does not meet the specifications, except for the case of DMDBT binding versus ANT binding. However, when an electron-donating ligand is allowed to interact with Au+, the effective charge on the cation is reduced.

(35) At an effective charge of about 0.8|e|, preferential adsorption of benzothiphene over naphthalene is observed, as well as preferential binding of dibenzothiophene and dimethyl diobenzothiophene over anthracene. This can be achieved using pyridine as the ligand. The analysis of the factors controlling the selective adsorption of thiophene derivative compared to arenes can also explain why Ag.sup.+ and Cu.sup.+ have been proposed in the literature as good candidates for the deep desulfurization of diesel based on MNDO (MDNO: (modified neglect of diatomic overlap) calculations, showing preferential thiophene over benzene, have, in fact, not proven to be effective for deep desulfurization of commercial diesel.

(36) Table 5 shows that while the binding of BTP is stronger than that of NPT for Ag.sup.+ and Cu.sup.+ as the number of aromatic rings increases, the selectivity observed for BTP versus NPT reverses and now ANT binds more favorably than DBT. A careful analysis of HOMO electron density surfaces (see FIG. 8) of these complexes shows that Cu.sup.+ and Ag.sup.+ do not interact with the S atom in the DBT and, instead, interact preferentially with the neighboring more basic C atoms. The reversal in the selectivity pattern with increasing number of aromatic rings is attributable to the unpredictable changes in partial atomic charges on the carbon atoms of arenes and thiophenes. By contrast, the negative charge on sulfur increases monotonically with increasing number of arene rings. Since diesel is rich in higher thiophenes, Cu.sup.+ and Ag.sup.+ are not ideal candidates for deep desulfurization of this fuel

(37) TABLE-US-00005 TABLE 5 Binding energy of Cu.sup.+ and Ag.sup.+ with NPT, BTP, ANT and DBT E.sub.bind E.sub.bind E.sub.bind E.sub.bind (M-NPT), (M-BTP), (M-ANT), (M-DBT), Metal kJ/mol kJ/mol kJ/mol kJ/mol Cu.sup.+ −233.67 −235.29 −249.64 −220.28 Ag.sup.+ −165.40 −168.40 −183.80 −173.70

(38) For the purposes of this specification and appended claims, unless otherwise indicated, all numbers expressing quantities, percentages or proportions, and other numerical values used in the specification and claims, are to be understood as being modified in all instances by the term “about.” Accordingly, unless indicated to the contrary, the numerical parameters set forth in the following specification and attached claims are approximations that can vary depending upon the desired properties sought to be obtained by the present invention. It is noted that, as used in this specification and the appended claims, the singular forms “a,” “an,” and “the,” include plural references unless expressly and unequivocally limited to one referent. As used herein, the term “include” and its grammatical variants are intended to be non-limiting, such that recitation of items in a list is not to the exclusion of other like items that can be substituted or added to the listed items.

(39) This written description uses examples to disclose the invention, including the best mode, and also to enable any person skilled in the art to make and use the invention. The patentable scope is defined by the claims, and can include other examples that occur to those skilled in the art. Such other examples are intended to be within the scope of the claims if they have structural elements that do not differ from the literal language of the claims, or if they include equivalent structural elements with insubstantial differences from the literal languages of the claims. All citations referred herein are expressly incorporated herein by reference.