However, fixed biomass does not necessarily have higher biological activity per unit
of organic dry mass than the suspended cells or flocs, in terms of substrate consumption
rate, partly on account of the internal diffusional limitations caused by the polymeric
matrix. For example, in nitrification experiments (Wiesmann, 1994), values around 0.2 g
of per gram of dry biomass and per day were obtained for both activated sludges
and biofilms. Since the total biomass concentration was significantly higher in the fixed
biomass systems, it seems that either there were less active nitrifying bacteria in the
biofilm than in the activated sludge, or each cell in the biofilm had, in average, a lower
biological activity than one cell in the suspended biomass.
This raises a very important question about the metabolic state of the cells in biofilms:
are their yield coefficients, maximum specific growth rates, saturation constants, etc., the
same as when they are in suspension? Biofilm modelling has been developed by
considering that the main differences in biological activity result from the fact that the
substrate diffusional limitations are more severe in a biofilm than in a suspended biomass
system. However, it does not seem correct to assume that the cells in the biofilm act in
every other aspect as if they were freely dispersed, because the microenvironment around
them can be totally different; not only the substrate concentrations can be lower than in
suspended cultures, but also the distances between cells are much smaller. Additionally,
there are increasing evidences of phenotypic changes in cells when they go from a
planktonic growth mode to biofilms (Costerton and Lappin-Scott, 1995).
BIOFILM REACTOR MODELLING
Biofilm reactors are usually calculated on the basis of lumped empirical parameters, the
values of which are assumed to be known from previous experience. An example of such
parameters is the so-called “eliminated load” (mass of substrate consumed per unit time
and unit volume of the reactor), frequently referred to in the design of wastewater
treatment plants (Harremöes and Henze, 1995). Hence, there are no generalised
relationships between the lump parameters and operational or design variables like
substrate concentration, liquid velocity, hydraulic residence time, biofilm thickness,
support characteristics, etc.
As yet, not even the existing unstructured mathematical models based on a
phenomenological approach of mass transfer and biological reaction rates (e.g.,
AQUASIM, see: Reichert 1994, 1995; Wanner and Reichert, 1996) are commonly used
by the practical designers, on account of the lack of sound estimates of the kinetic and
diffusion coefficients. In this section, a simple diffusion-reaction model, applied to a
biofilm system, will be presented with the aim of estimating the bio-reactor volume and
of offering the reader a basis for a quantitative analysis of the underlying mechanisms
that affect the reactor performance. This can also be helpful in terms of the understanding
of further developments of more sophisticated mathematical tools available for the design
of bio-reactors.
Biofilm reactors 301