R.H. Wijffels

Nitrification by immobilized cells

Abstract Nitrification has been studied extensively for its significance in the nitrogen cycle and within the treatment of wastewater. Often nitrogen removal from wastewater is problematic because of the low growth rate of the bacteria involved. Immobilization is an efficient method to retain slow-growing organisms in continuous-flow reactors. Immobilized cells can be divided into “naturally” attached cells (biofilms) and “artificially” immobilized cells. Biofilm systems are widely used in practice. Immobilized-cell reactor cells entrapped in gel beads form an alternative to these attached-cell systems. An overview is given of studies with immobilized nitrifying cells, with special attention to model development and evaluation. Dynamic models, including growth, diffusion limitation in the support material and in the microcolonies, and external mass transfer, are discussed. Experimental evaluation of the models is essential for their control if the models provide insight in the complex processes and for their use as a tool in reactor design. Several techniques for evaluation are given, including macroscopic substrate consumption rates and local analyses in the support material.

Relevance of rheological properties of gel beads for their mechanical stability in bioreactors.

The mechanical stability of biocatalyst particles in bioreactors is of crucial importance for applications of immobilized-cell technology in bioconversions. The common methods for evaluation of the strength of polymer beads (mostly force-to-fracture or tensile tests) are, however, not yet proven to be relevant for the assessment of their mechanical stability in bioreactors. Therefore, we tested fracture properties of gel materials and investigated their relevance for abrasion in bioreactors. Abrasion of gel beads was assumed to be a continuous fracturing of the bead surface. At first, three rheological properties were considered: stress at fracture; strain at fracture; and the total fracture energy. If stress at fracture is the most important property, beads having a similar fracture energy, but a smaller stress at fracture, would abrade faster in a bioreactor than beads with a larger stress at fracture; if fracture energy the determining factor, beads that require less energy to fracture would abrade faster than those having a larger fracture energy for the same fracture stress. To determine this, beads of kappa-carrageenan and agar (at two different polymer concentrations) were tested for abrasion in four identical bubble columns under the same operating conditions. Agar beads were expected to abrade faster than those of carrageenan because agar had either a lower stress at fracture or a lower fracture energy. However, no correlation between fracture properties and abrasion rate was found in any of the combinations tested. Carrageenan beads abraded faster than those of agar in all combinations. Furthermore, both the stress and strain at fracture of agar and carrageenan beads decreased during the run and those of carrageenan decreased faster, suggesting that the gels are liable to fatigue in different ways. This hypothesis was confirmed by oscillating experiments in which gel samples were subjected to repeated compressions below their fracture levels. Their resistance to compression clearly decreased with the number of oscillations. Fatigue is probably related to the development of microcracks and microfracture propagation within the material. We concluded that: (a) the use of tests based on bead rupture do not provide relevant information on the mechanical stability of gel beads to abrasion; and (b) abrasion of polymer beads is likely to be related to fatigue of the gel materials. (c) 1997 John Wiley & Sons, Inc. Biotechnol Bioeng 56: 517-529, 1997.

Dynamic modeling of immobilized Nitrosomonas europaea: Implementation of diffusion limitation over expanding microcolonies

Abstract A dynamic model (pseudohomogeneous growth model) describing growth and substrate consumption of immobilized Nitrosomonas europaea resulted in predictions for consumption rates that were higher than experimentally obtained data. This overestimation was caused by effects of diffusion limitation over microcolonies. Diffusion limitation over microcolonies has been integrated in the colony expansion model. In this model growth is implemented by expansion of colonies. Colonies breaking through the gel surface cause release of bacteria into the surrounding medium. Two options were considered for this release: either growth resulting in continuous leakage of single bacteria or eruption of entire colonies at once. In the latter case both macroscopic consumption rates and transient-state biomass concentration profiles are described satisfactorily by the model. It is also shown that the results of the previous experiments with Nitrobacter agilis can be described by the model as well.

Dynamics of artificially immobilized Nitrosomonas europaea: Effect of biomass decay.

The dynamics of growth and death of immobilized Nitrosomonas europaea were studied. For this, the death rate of suspended cells was determined in the absence of ammonium or oxygen by following the loss of respiration activity and by fluorescein-diacetate (FDA)/lissamine-green staining techniques. The death rates obtained (1.06 x 10(-6) s(-1) or 4.97 x 10(-6) s(-1) in the absence of oxygen or ammonium, respectively) were incorporated in a dynamic growth model and the effects on the performance of the immobilized-cell process illustrated by model simulations.These model simulations and experimental validation show that if decay of biomass occurs the biomass concentration in the center of the bead decreases. As a result, the systems react slower to changes in substrate concentrations than if all cells remain viable.To show that cells in the center of the bead died, the FDA and lissamine-green staining techniques were adapted for immobilized cells. It was shown that biomass decay occurred, especially in the center of the bead; the amount of cells decreased there, and the remaining cells were all stained with lissamine green indicating cell death. After the substrate availability was decreased, also cells near the surface of the bead lost their viability. The number of viable cells increased again after increasing the substrate concentration as the result of cell multiplication. At low substrate concentrations and low hydraulic retention times, as for example in the treatment of domestic wastewater, the death rate of cells is thus an important parameter for the performance of the immobilized-cell system. (c) 1997 John Wiley & Sons, Inc. Biotechnol Bioeng 55: 630-641, 1997.

Dynamic modeling of the growth of immobilized nitrifying bacteria: biofilm development.

The use of immobilized growing cells has become of increasing interest in the past few years. Traditionally, biofilms are used in wastewater treatment (Canovas-Diaz and Howell 1988, Saez and Rittmann 1988, Suidan et al. 1987, Siegrist and Gujer 1987). In this field the use of entrapped nitrifying bacteria is gaining importance, as this will lead to a more stable and well-defined system (Wijffels and Tramper 1989). The overall volumetric productivity of bioreactors can be largely enhanced by the application of immobilized cells. The dilution rate can be very high since cells will not be washed out of the reactor. The growth rate of nitrifying bacteria is always low and in particular the application of immobilized cells is advantageous.

External Mass Transfer

Numerous papers have been devoted to modelling simultaneous diffusion and conversion of substrate by immobilized biocatalysts. In such systems the substrate is transferred from a liquid phase to a solid phase in which the reaction occurs. In many of these studies it was assumed that the rate-limiting step in the process is the diffusion of substrate through the solid phase, because the thickness of the boundary layer surrounding the solid particles was supposed to be much smaller than the radius of the gel bead used as a support for the immobilized biocatalyst. Furthermore, the diffusion coefficient in the carrier is usually smaller than in water. In the case of immobilized- cell processes, however, external mass transfer needs to be incorporated in the model in order to obtain realistic predictions (De Gooijer et al. 1991, Wijffels et al. 1991). With immobilized growing bacteria a relatively thin layer of biomass will be formed just beneath the support surface. If for example oxygen is the limiting substrate, this active layer will be in the order of 100 µm thick, which is of the same order of magnitude as the thickness of the boundary layer.