In recent years the application of membrane technology to molecular separation processes
has stimulated interest and showed great potential in a number of industrial fields. Ultrafiltration membranes have been successfully applied to downstream separation of therapeutically
active peptides, to overcome some of the limitations of the conventional techniques in
terms of costs, scale-up, selectivity and solvent recovery. In this research project, Organic
Solvent Nanofiltration of peptide solutions is studied, and this understanding is applied to
the development of innovative membrane-based purification strategies for industrial case
studies. Basic understanding of transport mechanisms was approached by investigating
solvent transport through ceramic nano- and ultrafiltration membranes, and developing
a predictive phenomenological model for the transport of solvents and solvent mixtures. Effects of solvent-membrane interactions strongly affected the solvent permeation through nanofiltration membranes, while they were found to be negligible in the ultrafiltration range.
The effect of the organic solvent on the permeation of neutral and charged solutes (monovalent
salts, a small molecule and peptides) in organic/water mixtures was studied, with
particular attention to the role of preferential solvation in the solvent mixture. It was found
that the solvent composition and the complex association of counter-ions and buffers highly affect membrane permeation and rejection of organic molecules. It is proposed that all these
components change the relative solute-membrane affinity. Since permeation of peptides in
organic/water mixtures is affected by complicated matrices of input parameters, a Design
of Experiment approach was proposed to efficiently investigate the nanofiltration of model
peptides in acetonitrile/water solutions. Statistical models for solvent flux, peptide and
ion rejection were obtained by Analysis of Variance and interpreted from a phenomenological point of view. The statistical models were used to asisst process development for
two industrial case studies: (1) concentration and salt/solvent exchange of a first therapeutic
peptide were optimised, based on the integration of the statistical DoE models with
the process simulation for concentration and diafiltration; (2) the nanofiltration-assisted
synthesis of a second therapeutic peptide, based on the coupling between nanofiltration and reaction in one unique process, was developed and compared to the established process
by techno-economical analysis. The so-called "Reactive Peptide Nanofiltration" was
found to be advantageous in terms of economics, efficacy, impact on the market, and on
the environment.
In conclusion, nano ltration was found to be a solid and competitive technique for
application to peptide processes. On the basis of the results of this research, Lonza decided
to invest in a new nano ltration plant for the downstream of peptides with ceramic
membranes. The advantages of nanofiltration technology, in terms of development of more efficient materials (stable in critical solvents and harsh acid/basic conditions), improvement
of membrane performances (selectivity, lifetime) and integration of nanofiltration with other
techniques in hybrid processes seem therefore promising in overcoming the hesitancy of industries
to modify the established processes and invest in new nano ltration plants, by
making the payback period for the return of investment more attractive. It is plausible
to think that this technology will shortly become a primary choice for new separation and purification processes.