The application of thermodynamic models in the development of chromatographic separation processes is discussed. The paper analyses the thermodynamic principles of protein adsorption. It can be modeled either as a reversible association between the adsorbate and the ligands or as a steady-state process where the rate of adsorption is equal to the rate of desorption. The analysis includes the competitive Langmuir isotherm and the exponentially modified Langmuir isotherm. If the adsorbate binds to one ligand only, the different approaches become identical. When the adsorbate acts as a ligand, dimerization takes place and will give rise to a sigmoid isotherm. A model that accounts for dimerization is discussed and a sample calculation shows the behavior of this isotherm. Insulin is known to have a concave isotherm at low concentrations. The calculation of the standard Gibbs energy change of adsorption is discussed. Hydrophobic and reversed phase chromatography are useful techniques for measuring solute activity coefficients at infinite dilution.
Process Development of Chromatographic SeparationsThe fulcrum point in the design of separation processes like distillation, extraction, absorption, or adsorption, is the phase equilibrium properties. Today, we use simulation programmes to develop and design distillation, extraction and absorption units, but when it comes to the development and design of chromatographic separations in the biopharmaceutical industry, the state of the art looks like the chemical industry in the 1950s. The conventional chromatographic process development is at present mostly an experimental trial and error based approach supported by high throughput screening platforms. However, if one takes advantage of the models of adsorption discussed in this paper, the agenda can be updated. The scope will no longer be to develop the chromatographic separations by trial and error, but to employ simulations to assist the process development. This approach will reduce the time and cost of process development. The adsorption model parameters can be estimated from isocratic retention data supplemented with a few capacity measurements. The 'efficiency' is, as usual, determined by the mass-transfer coefficients that can be estimated from the second central moment of the isocratic peaks. This requires little material and time and a limited analytical effort. Limiting factors are only computational effort and the quality of the thermodynamic models. Having these models and the mass transfer coefficients, we can use a simulator to design robust separations that can deliver a predefined quality as regards yield, purity and productivity. The simulated separations must of course be validated experimentally. If necessary, we can tune the model parameters to improve the agreement between the simulated and experimental results.Furthermore, the use of a simulator is PAT compliant because the simulator can also support: (1) identification of critical process parameters and critical sources of variability, (2) a study of h...