Catalyst Poisoning
Catalyst poisoning is the loss of catalytic activity caused by a species, often a feed impurity such as sulphur, that adsorbs strongly on the active surface. The simplest picture is site blocking: the poison occupies sites that reactants would otherwise use. A poison can also change the electronic structure of nearby, unoccupied surface atoms, lowering their intrinsic reactivity. Density functional theory makes it possible to separate these effects by computing how poison binding varies with facet, site and coverage, and how the surface d states respond. See also d-Band model and Coverage-dependent adsorption.
Sulphur on Cu water-gas shift catalysts (2009)
İnoğlu and Kitchin computed S adsorption (referenced to H2S and H2) on Cu(111), (100) and (110) at 0.25 to 1.00 ML in 43 structures with DFT (PW91). S binds most strongly on (110) and most weakly on (111), prefers the most highly coordinated sites, and binds more weakly as coverage increases; on Cu(111) adsorption becomes endothermic above 0.50 ML. Because the surface d-band filling stays near 9.9 electrons per Cu atom, adsorbate-induced band broadening shifts the d-band center down, and the S adsorption energy weakens nearly linearly with that shift for hollow and bridge sites (atop sites and dimer or trimer structures excepted). The authors conclude that sulphur poisoning of Cu combines site blocking with an intrinsic electronic deactivation of the surface (İnoğlu 2009). Summary: inoglu2009a.
Thermodynamics of sulfur poisoning of Cu under water-gas shift conditions (2009)
İnoğlu and Kitchin used DFT-based atomistic thermodynamics to ask when trace H2S poisons Cu(111), (100) and (110) in a WGS environment (INOGLU 2009). Clean surfaces are stable only below a S chemical potential (referenced to H2S/H2) of about -1.25 eV; above that, 0.25 to 0.50 ML S phases form in hollow sites until bulk CuS (formation energy -0.40 eV) becomes favored. On Cu(111) at 400 K and 1 ppm H2S, roughly 10³ atm of H2 would be needed to keep the surface clean, so poisoning is unavoidable at low temperature, whereas at 650 K and 25 atm H2 the clean surface is again most stable, suggesting hydrogen regeneration at high temperature (with a sintering risk). Summary: inoglu2009.
Screening sulfur-tolerant bimetallic surfaces (2011)
İnoğlu and Kitchin screened 370 binary alloy surfaces built from ten late transition metals for sulfur tolerance, defining tolerance as the S chemical potential (set by the H2S/H2 ratio) at which a 0.25 ML S-covered surface first becomes more stable than the clean one. S adsorption energies were estimated from a DFT-calibrated linear relation between shifts in S binding and shifts in surface d-band width (slope 1.59), reproducing 92% of DFT values within 20%. Alloying that broadens the d band, through compressive strain or partners with larger d orbitals, weakens S binding: for Cu, pure Cu poisons near -0.93 eV, a Cu-Ir-Cu structure is more tolerant, and Cu on Ru poisons earlier (near -1.30 eV). Predictions agreed qualitatively with known S-tolerant Pd-Cu and Pd-Au membranes and Au-modified Pt and Rh (İnoğlu 2011). Summary: inoglu2011.
SO2 poisoning of an amine CO2 sorbent (2013)
The same idea of a strongly bound impurity deactivating active sites applies to amine sorbents for CO2 capture. Hallenbeck and Kitchin exposed a primary-amine ion-exchange resin (OC 1065) to simulated flue gas containing 431 ppm SO2 and found that CO2 capacity fell nearly linearly to about 0.2 mol/kg within seven cycles, with 0.98 mol/kg of sulfur retained (Hallenbeck 2013). Surface area and pore structure were unchanged, so the loss was attributed to reaction of the acid gas SO2 with the amine sites rather than physical blocking of pores. The adsorbed sulfur survived heating to 208 C, and NaOH treatment removed only 43% of it, restoring about 35% of capacity, whereas resin poisoned by sulfuric acid was fully restored by NaOH. Summary: hallenbeck2013.