**1. Introduction**

618 Mass Transfer - Advanced Aspects

Heim, A. & Solecki, M., (1999). Disintegration of microorganisms in bead mill with a multi-

Hetherington, P. J., Follows, M., Dunnill, P. & Lilly, M. D. (1971). Release of protein from

Limon-Lason, J., Hoare, M., Orsborn, C.B., Doyle, D.J. & Dunnill, P. (1979). Reactor

Marffy, F. & Kula, M.R. (1974). Enzyme yields from cells of brewer's yeast disrupted by

Mashmoushy, H., Zhang, Z. & Thomas, C.R. (1998). Micromanipulation measurement of

Melendres, A.V., Honda, H., Shiragam, N. & Unno, H. (1991). A kinetic analysis of cell disruption by bead mill. *Bioseparation,* Vol.2, pp. 231-236, ISSN 0923-179X Melendres, A.V., Honda, H., Shiragami, N. & Unno, H. (1993). Enzyme release kinetics in a

Middelberg, A.P.J., O'Neill, B.K. & Bogle D.L. (1991). A novel technique for the

Shiu, C. Zhang, Z. & Thomas, C.R. (1999). A novel technique for the study of bacterial cell mechanical properties, *Biotechnology Techniques,* Vol.13, pp. 707–713, ISSN 0951-208X Smith, A.E., Moxham, K.E., Middelberg, A.P.J. (1998). On uniquely determining cell-wall

Smith, A.E., Zhang, Z. & Thomas, C.R. (2000). Wall material properties of yeast cells: Part 1.

Solecki, M. (2007). Yeast disintegration in a bead mill. *Chemical and Process Engineering*,

Solecki, M. (2009). Modeling of microorganism disintegration in spherical packing of a bead

*Particulate Solids*, ISBN 978-0858 259 065, Brisbane Australia, August 2009 Solecki, M. (2011). The theory of random transformation of dispersed matter. (unpublished) Stenson, J.D., Ren, Y., Donald, A.M. & Zhang, Z. (2010). Compression testing by

Svaldo-Lanero, T., Krol, S., Magrassi, R., Diaspro, A., Rolandi, R. Gliozzi, A. & Cavalleri, O.

Whitworth, D.A. (1974). Hydrocarbon fermentation: protein and enzyme solubilization from

baker's yeast (*Saccharomyces cerevisiae*) by disruption in an industrial homogeniser. *Transactions of the Institution of Chemical Engineers,* Vol.49, pp. 142–148, ISSN 0046-

properties of a high-speed bead mill for microbial cell rupture. *Biotechnology and* 

treatment in a horizontal disintegrator. *Biotechnology and Bioengineering*, Vol.16, pp.

the mechanical properties of baker's yeast cells. *Biotechnology Techniques,* Vol.12,

cell disruption chamber of a bead mill. *Journal of Chemical Engineering of Japan,*

measurement of disruption in high-pressure homogenization: studies on *E. coli* containing recombinant inclusion bodies. *Biotechnology and Bioengineering,* Vol.38,

material properties with the compression experiment, *Chemical Engineering Science,*

Cell measurements and compression experiments, *Chemical Engineering Science,*

mill. *Proceedings of 5th International Conference for Conveying and Handlings of* 

nanomanipulation in environmental scanning electron microscope. *Experimental* 

(2007). Morphology, mechanical properties and viability of encapsulated cells.

*C. lipolytica* using an industrial homogenizert. *Biotechnology and Bioengineering,* Vol.

disc impeller. *Powder Technology,* 105, 390-396, ISSN 0032-5910

*Bioengineering,* Vol.21, pp. 745-774, ISSN 0006-3592

9858

623-634, ISSN 0006-3592

No.12, pp. 925–929, ISSN 0951-208X

Vol.26 No.2, pp. 148-152, ISSN 0021-9592

Vol.53, pp. 3913-3922, ISSN 0009-2509

Vol.55, pp. 2031-2041, ISSN 0009-2509

*Techniques*, Vol.34, No.2, pp. 60-62, ISSN 1747-1567

*Ultramicroscopy,* Vol.107, pp. 913–921, ISSN 0304-3991

Vol.28, 649-660, ISSN 02508-64

16, pp. 1399-1406, ISSN 0006-3592

pp. 363–370, ISSN 0006-3592

Membrane Contactors (MC) make possible to accomplish gas-liquid or liquid-liquid mass transfer operations without dispersion of one phase within another. The membrane acts as a mere physical support for the interface and does not contribute to the separation through its selectivity, the separation being primarily based on the principle of phase equilibrium.

Porous membranes with narrow pore size (typically in the range 0.02-0.2 μm) are used. One of the two phases enters the membrane pores and contacts the other phase at the pore mouth on the opposite side. Generally MC operations involve an aqueous phase and the membrane has hydrophobic character, however hydrophilic membranes can be used too [Kosaraju & Sirkar, 2007]. The key factor is that one of the two phases enters the pores whereas the other phase is kept outside. In the case of hydrophobic materials (Fig. 1), the membrane pores are filled by the non-polar phase or by the gas while the aqueous phase can not penetrate into the pores.

Fig. 1. The membrane contactor concept based on porous hydrophobic membrane in contact with an aqueous (non wetting) liquid at one side and a gas or organic (wetting) liquid at the other side

Recovery of Biosynthetic Products Using Membrane Contactors 621

Polymeric membranes were usually employed in membrane contactor studies and applications, because they are very competitive in performance and economics. Membranes with high hydrophobicity were preferred in most applications. Polypropylene (PP), polytetrafluoroethylene (PTFE), and polyvinylidenefluoride (PVDF) have been extensively used. Among these, microporous PTFE membrane shows excellent performance and stability, however, the application on industrial scale is limited, since PTFE membranes are available in flat sheet form rather than as hollow fibres. At present hollow fibre PP membranes exhibit the widest applications due to the good thermal and chemical stability, well-controlled porosity, and low cost. Recently increasing efforts have been devoted to develop ceramic membranes [Koonaphapdeelert et al., 2007, 2009; Li, 2007.], or hybrid membranes to get better chemical and thermal stability as well as higher mechanical strength. Membrane surface modification techniques to improve the hydrophobicity has

In principle both flat sheet membranes and hollow fiber can be used to make MC apparatuses. Fig. 2 shows the concept of a plate and frame module: it contains parallel membrane sheets with interposed spacers. Problems can arise from the susceptibility of the membrane to mechanical abrasion that required a careful design of the spacers and/or the use of supported membrane with an active layer, made for example of PTFE, and a woven or unwoven fabric. In this case the spacer in the active layer side can be substituted by a simple frame. Co-current flow is to be preferred to counter-current flow, in order to have a nearly constant pressure difference along the membrane, avoiding pressure inversion and

The membrane area, a, per unit volume of these apparatuses is quite limited; assuming for example 1 mm the thickness of the fluid channels and 2 mm the thickness of the frames, the membrane area is a = 250 m2/m3 . Taking into account the borders and the manifolds,

**2. Membrane and modules** 

also been investigated [Lv et al., 2011].

Fig. 2. The concept of plate & frame membrane contactors

membrane movement.

It is worth noting that the hydrophobicity of the materials is not a warranty for keeping the aqueous phase outside of the pores, indeed if a critical pressure value, called *breakthrough pressure*, is exceeded, the aqueous phase enters the membrane pores. The *breakthrough pressure* depends on the maximum pore size, dP, the interfacial tension between the two phases, γ, the contact angle between the membrane and the two fluid, θ, according to the Young-Laplace equation (strictly valid for cylindrical pores):

$$
\Delta P\_c = \frac{4\chi \left| \cos \theta \right|}{d\_P} \tag{1}
$$

As a consequence it is important to carefully control the operating pressures: the pressure of the aqueous phases has to be a bit higher than the pressure of the organic/gas phase, but lower than the breakthrough pressure. Selecting appropriate membrane materials and pore size it is possible to assure a pretty wide range of pressure for safe operations, for example PTFE or polypropylene membranes with nominal pore size of 0.2 μm exhibit a breakthrough pressure of nearly 3 bars for the water-air system. The discussion above can be extended to hydrophilic membranes; in that case the pores are filled by the aqueous phase, whose pressure has to be taken a bit lower than the gas/organic phase pressure, which in turn has to not overcome the breakthrough pressure, given again by Eq. (1). The choice between hydrophobic and hydrophilic membranes is dictated by the need to reduce the membrane resistance. As a rule the membrane pores should be filled by the phase in which the transferred species is most soluble. For example, if the species has higher affinity with the non polar or gas phase hydrophobic membrane will be preferred. If there is higher affinity with the polar phases the membrane will be hydrophilic.

MC have many advantages with respect to conventional mass transfer apparatuses: membrane modules, in the form hollow fibres, provide interface areas per unit volume significantly greater than traditional devices, leading to more compact systems, - the interfacial area is well defined and remains constant regardless of the flow rates, whereas the design of the conventional devices is restricted by limitations in the relative flows of the two streams. - there is no mix of the two phases separate by the membrane and thus no need to separate the two phases downstream the process and no need of difference of density. - In addition MCs are easy in scale up and control, modular in design, flexible in operation.

Membrane contactors can be used for carrying out the recovery of bioconversion products from aqueous solutions by organic solvents, usually conduct in centrifugal devices, mixersettler or columns. An additional advantage is relevant in this case: the operation can be performed in the presence of the living cells or enzymes that indeed do not get in direct contact with the solvent. The use of membrane based, dispersion free, solvent extraction for the recovery of bio-products, pioneered by Sirkar [Frank & Sirkar, 1987] has been well documented [Gawronski, 2000; Lazarova et al., 2002; Schlosser al., 2005].

Of course MCs present also some disadvantages: - no doubt the membrane represents an additional mass transfer resistance, and indeed overall mass transfer coefficients lower than those of conventional devices have been reported. However this drawback is well offset by the larger specific area, as a consequence the volumetric mass transfer coefficient for MCs is well above the values achievable in conventional devices. As in other membrane processes, problems my be the limited life time and fouling. However fouling my be less severe than in pressure driven processes; suspended solids and solutes are indeed not forced into the pores by convective flow, the species being transferred from one phase to the other by only diffusion.
