An experimental approach for determining thermodynamic surface complexation descriptors of weak-acid oxyanions onto metal (hydr)oxides: Case study of arsenic and titanium dioxide

The extensive literature review suggests that there are two main reasons for contradictory thermodynamic parameter values obtained via sorption experiments: (1) many of the studies are conducted under unrealistic conditions where the sorbate/sorbent ratios are so high that physisorption is artificially induced; or (2) many of the studies incorrectly calculate the equilibrium constants. The goal of this study is to demonstrate a methodology that describes how to properly determine and verify theoretically predicted thermodynamic descriptors. The study employs arsenate and titanium dioxide as a model sorbate-sorbent pair, which is equilibrated under realistic conditions for a period of 3 days at two different pH conditions (~6.5 and ~8.5) and three different temperatures (7 °C, 25 °C and 35 °C) in 10 mM NaHCO ₃ . At pH ≈ 8.55, ΔG ^o values were −83.38 ± 1.62 kJ/mol, −88.13 ± 0.66 kJ/mol, and −90.78 ± 0.61 kJ/mol for sorption performed at 7 °C, 25 °C and 35 °C, respectively. Decreasing the pH to about 6.65 resulted in slightly less negative values of ΔG ^o to −73.38 ± 1.58 kJ/mol, −77.14 ± 1.52 kJ/mol, and −78.75 ± 1.53 kJ/mol for sorption conducted at the same respective temperature conditions. These values overlap with the ΔG ^o ranges reported for sorption of arsenate on metal oxides. Change in enthalpy values of ΔH ^o = −19.04 kJ/mol at pH ≈ 6.65 and ΔH ^o =−9.35 kJ/mol at pH ≈ 8.55 were observed. Based on reports, which suggest that at lower pH more bidentate ligands are being formed, these values are expected. The change in entropy values ranged from ΔS ^o = 0.19 kJ/mol K at pH ≈ 6.55 to ΔS ^o = 0.26 kJ/mol K at pH ≈ 8.55, which suggests lower level of disorder among the created complexes at lower pH and it is in line with the rationale that bidentate complexes are better organized on the surface of the sorbent and less susceptible with desorption. These findings clearly demonstrate that experimentally obtained ΔG ⁰ and other thermodynamic values and trends could be obtained to reflect and confirm model predictions when the existing sorption theory is properly translated into experimental practice. The sorbate-sorbent bond in chemisorption has covalent character, characterized with short bond length and higher bond energy, which makes it less reversible when compared to physisorption, and therefore highly significant from a sorbent remediation-performance practical point of view and long-term waste sorbents disposal. While thermodynamic parameter modeling represents a good first step in determining the suitability of an initial design, experimental techniques potentially have the ability to provide far more superior description of the thermodynamic sorbent/sorbate interactions under realistic conditions.