CELL IMMOBILIZATION IN HOLLOW FIBER MEMBRANES

Cell immobilization for biochemical production or in waste degradation has gained increasing interest in recent years owing to its many advantages. Immobilization of cells and enzymes onto different support matrices enables realization of high cell densities that are otherwise not achievable by conventional means, resulting in higher reactor productivity. In addition, the immobilization support can be chosen in the desired geometry that provide greater ease of handling, better mass transfer characteristics, etc. General immobilization techniques include the covalent attachment or adsorption to solid supports, entrapment within a gel lattice, microencapsulation within semipermeable membranes and cross linking to insoluble aggregates. These procedures, in spite of improving catalyst stability, unfortunately often result in loss of catalytic activity due to unfavourable conditions encountered during the immobilization procedure.

 

Immobilization by entrapment of cells in hollow fiber membranes is being studied with great interest due to the distinct advantages offered by hollow fiber membranes as an ideal geometry for immobilization. Since they are fabricated in a process separate from the inoculation procedure, there is greater flexibility in altering the structural and transport properties of the immobilization support without compromising cell viability and productivity. Also, the selective skin layers in the hollow fiber membranes provide in-situ separation of the biocatalysts from the nutrients and product stream. Such an immobilization procedure is relatively easy and possess less mass transfer limitation.

 

Hollow fiber membranes used for immobilization are generally of the asymmetric type consisting of two skins, one of which is dense and the other porous. In between, there is a porous cross sectional matrix with macrovoids having a void volume of between 60 and 90%. The hollow fibers are housed in a tube as in a traditional shell and tube configuration. Nutrients and air may be supplied through the shell side and the desired clarified product derived from the tube side. In general, the few cells which gain entry into the fiber matrix through the porous external skin will subsequently grow within the matrix and propagate to high cell densities.

 

Preparation of membranes of suitable morphology for immobilization depends primarily on formulating the proper polymer solution used for spinning, as well as the conditions for fiber spinning by the phase inversion processes.

 

Current work aims at obtaining the optimal conditions required to prepare membranes with morphological features ideally suited for efficient immobilization applications. Further, the work also involves study of immobilized system of bacterial strain E. coli in the cross sectional matrix of hollow fiber membranes.

 

The polymer solutions are prepared by dissolving the polymer, (polyethersulfone), in a solvent mixture containing N-methyl pyrrolidone as a solvent and diethylene glycol as a nonsolvent additive. Experiments were initially conducted to study the various effects such as polymer concentration, length of air gap and temperature of external coagulation bath on the structure and separation characteristics of the hollow fiber

 

membranes produced. A typical asymmetric structure of the fiber produced is shown in Figure 1. To bring about specific alterations to the fiber morphology to suit immobilization applications, fibers were prepared by adding a surfactant (sodium dodecyl sulfate) to the polymer solution and to the external coagulation bath in the subsequent experiments. The fibers were evaluated for their performance characteristics by studying permeation flux and solute rejection capabilities, and morphological characteristics by scanning electron microscopy analysis.

Figure 1. Asymmetric structure of membrane cross section

Growth rate of immobilized cells in the fiber matrix was monitored in a fed-batch system for various batches of nutrient solutions. Fiber samples were withdrawn after every batch for the physical examination of growth using a scanning electron microscope (SEM) and for the determination of the cell density using dry weight and DNA estimation methods. Progressive growth of cells could be observed from SEM micrographs and the fibers were found to be almost saturated after four batches of growth. Figure 2 depicts a photomicrograph of a fiber void space after four batches of growth. Beyond this, due to excessive cell density, cells were found to leak to the fiber lumen forcing their way through the dense inner skin.

Figure 2. Cells filled fiber void space

A cell density of 90 g/i (based on fiber volume) was measured by dry weight method which is nearly 40 times higher than the cell density that can be achieved in a single batch of suspension cell growth.

 

From the study, it was found that fibers prepared form polymer solutions of 16% polymer concentration, spun at an air gap of 1 cm and coagulated at 24oC produced the hollow fibers best suited for cell immobilization as well as good water flux. The precipitation kinetics of the phase inversion process of membrane preparation was altered by using a surfactant in the polymer solution to modify the external skin to required structural features.

 

A study of the growth kinetics of both the immobilized system and the suspension cell is being made to address the differences between the two.

 

Contact Persons:

 

 

A/P WK Teo, (Tel.- 7726614, Fax: 7791936, Email: cheteowk@nus.sg)

Dr YP Ting, (Tel: 7722190, Fax.- 779 1936, Email: chetyp@nus.sg)