Professional Documents
Culture Documents
ENGINEERING
JOURNAL
ELSEVIER The Chemical EngineeringJournal 65 ( 1997) 55-61
Abstract
Realistic mathematical models are essential for the scaling-up or the design of biofilters. However, the mathematical models which describe
the steady-state and the transient operation of biofilters are very limited. Although some of these models are based on restrictive assumptions,
they have been widely used in industry for designing actual biofilter units. This study gives detailed analysis and comparison between these
models. The asymptotic behavior of recent models are presented. The results show that diffusion is an important phenomenon which should
not be neglected in developing biofilter models, and that neglecting oxygen transport and its effects on growth kinetics will give unrealistic
values for the effective as well as the actual film thickness.
In a recent study, Zarook et al. [ 31 extended the work of 2. Theory and analysis
Ottengraf [ 21 and presented a detailed steady-state biofiltra-
tion model for single VOCs. In this model: (i) The actual A steady-state biofiltration model constitutes a set of mass
kinetic expressions resulting from shake-flask experiments balances within the biofilm as shown in Fig. 1 and in the gas
were used instead of limiting cases such as zero- or first-order phase. The assumptions on which the model equations are
kinetics. (ii) Oxygen was also considered in the growth rate based are given in detail elsewhere [ 3,6]. Here, for compar-
expression, i.e. oxygen was not assumed to be in excess. (iii) ison purposes, we present the model equations for removal
The film thickness was not assumed to be constant throughout of a single VOC. The mass balance equations in the biofilm
the column and it was determined as the thickness of the are:
biolayer where one of the limiting substrates (either VOC or
oxygen) depletes before it reaches the biolayers/solid sup-
K xv
-y-/-4C,?GO) (1)
port interface. Model equations based on these improvements
Dz=
constituted a two point non-linear boundary value problem
for which analytical solutions were not possible. Since both D -~CIO= xv
u P(C,,Go) (2)
the models of Zarook et al. [ 31 and Ottengraf [ 1] and Otten- o dx’ o
graf and van den Oever [ 21 are valid for, steady-state, single with boundary conditions:
VOC removal, a detailed analysis and comparison between
these models are necessary and important so that the appro- c,= 2 and C,, = 59 at x=0 (3)
priate model can be used for the experimental validation of m m0
ug s
dh
=A,D, -
Go
I 1 dx X=0
one can clearly identify two regimes of operation: diffusion
limited and reaction limited regimes. In the diffusion limited
regime, the rate of diffusion is slow compared to the VOC
utilization rate in the biofilm [ 21. Ottengraf and van den
with initial conditions Oever [ 21 showed that the effective biofilm thickness for this
C, = C,, and C,o = Cso, at h = 0 region is given by
(11)
2Dc,
S= (17)
2.1 Asymptotic behavior of the model km
Here, one should be aware that equation 17 is based on the
Ottengraf and van den Oever [ 21 implicitly assume that assumption that the VOC gets completely consumed in a
oxygen is in excess and it does not exert any limitation on portion of the actual biofilm thickness. However, if oxygen
the biodegradation rate. In general, the biodegradation rate in is limiting then the effective biofilm thickness calculated
the biofilm can be expressed as using Eq. (17) may be invalid. The concentration profile
X” along the biofiltei for zero-order kinetics for the diffusion
--r= -P (12) limited regime is given by [ 21
Y
Although the authors [ 21 acknowledge that based on shake- (18)
flask experiments the biodegradation kinetics of single VOCs
follow the Monod model (Eq. (5) ) , they only consider two
In the reaction limited regime, the total biofilm is fully
limiting cases. At high concentrations they assumezero-order
active. For this case, the concentration profile in the gas phase
kinetics while at low concentration they assume first-order
is given by [ 21
kinetics. Thus, in the excess oxygen limitation (C,o B Ko)
for a large value of C,o, Eq. (7) gives a zero-order reaction
rate as,
(C,BK)
where 6* is the actual biofilm thickness
Go@p=) Cl UP-) values are very large and unrealistic. Also this model is based
107 J- 2000 on the assumption of a constant film thickness along the
biofilter. Furthermore, from Eq. (19), one can easily show
that the removal rate is given by A,6*k,. Thus, despite the
changes in the inlet concentration and flow rates, one gets
6.
constant removal rates (R,cm.RL) as shown in column 10 of
Table 1. The maximum percentage difference ( ERLin column
-1000
11) is as high as 67.7%.
4-
When the kinetics are assumed to be first-order, then one
can not clearly distinguish between different regimes. For
- 500
2- this case, an analytical solution was possible and it is given
in Eq. (20). The first-order reaction rate constant was esti-
01
mated through Eq. (14). As in the previous case, we have
0 5 10 ls 20 25 30 two parameters As and 6* which need to be found. However,
5 (w) as seen from Eq. (20)) here we can not lump (ASS*) simply
Fig. 2. Concentration profiles of oxygen and VOC in the biofilm at a pattic- as in the previous case. In order to find this parameter, Eq.
ular location in the biofilter. (20) is expanded using the Taylor series as follows
I.5 -
l-
0.5 -
^ ^_ I 0,.
U.U5 U5 5
0 0.25
0 dimensianleYGheight (hG5
Fig. 4. Effectiveness factor versus Thiele modulus alongthe biofiltercolumn. Fig. 5 Effectiveness factor versus dimensionless height along the biofilter
Curves 1, 2, 3 and 4 are for first four sets of data given m Table I. column. The curve is for the data set 7 given in Table 1.
Jersey Institute of Technology (NJIT), USA, for invaluable X” biofilm density defined as the dry weight of cell
advice on biofiltration research and his excellent teaching. per volume of biofilm
Y amount of biomass produced per amount of VOC
consumed
Appendix A: Nomenclature yo amount of biomass produced per amount of oxy-
gen consumed
AS biolayer surface area per unit volume of biofilter
C6 concentration of VOC in the gas phase at a height, Greek letters
h, along the column
C6, concentration of VOC at the entrance of the active biofilm thickness as shown in Fig. 1
biofilter actual biofilm thickness as shown in Fig. 1
C60 concentration of oxygen in the gas phase at a effectiveness factor as defined by Eq. (25)
height, h, along the column specific growth rate of the biomass on VOC
C601 concentration of oxygen at the entrance of the the maximum specific growth rate in Eqs. (5)
biofilter and (7)
Cl concentration of VOC at a position x in the kinetic constant in Eq. (6)
biofilm Thiele modulus as defined by Eq. (2 I )
CIO concentrations of oxygen at a position x in the
biofilm Subscripts
D effective diffusion coefficient of VOC in the
biofilm DL diffusion limited
Do effective diffusion coefficient of oxygen in the i entrance conditions at h = 0
biofilm g gas phase
E percentage error between model predicted and 1 liquid phase
experimentally evaluated removal rates RL reaction limited
h height of the biofilter at any point
ko zero-order reaction rate constant
k, first-order reaction rate constant References
km ratio of mass of VOC in the biofilm to mass of
[ I] S.P.P. Ottengraf, in W. Shonbom (ed.), Biotechnology, Vol. 8 VCH,
VOC in the gas phase as defined in Ref. [ 91 Weinheim. 1986, p. 425.
K kinetic constant in Eqs. (5) and (6) [2] S.P.P. Ottengraf and A.H.C. van den Oever. Biofechnd Bioeng., 25
Ko kinetic constant in Eqs. (7) and (8) (1983) 3089.
KI kinetic constant in Eq. (6) [ 31 S.M. Zarook, B.C. Bahzis, Y.-S. Oh and R. Bartha, Biorechnol. Bioeng.,
m air/biofilm distribution coefficient for the VOC 41 (1993) 512.
[4] C. van Lith, S.L. David and R. Marsh, Truns. IChemE., 68 ( 1990) 127.
as dictated by Henry’s law [ 51 S. Dharmavaram, in Proceedings of the 84th Annual A&WMA Meeting,
m0 air/biofilm distribution coefficient for the oxy- Paper No. 91-103.2, Vancouver, B.C., June 1621. 1991.
gen as dictated by Henry’s law [6] B.C. Baltzis and SM. Zarook, in Proceedings of the 86th Annual
-r biodegradation reaction rate A&WMA Meeting, Paper No. 93-TP-52A.03, Denver, CO, 13-18 June,
R removal rate defined as the mass of VOC 1993.
[7] M.A. Deshusses and I.J. Dunn, in Proceedings of the 6th European
removed per volume of packing per time Congress on Biotechnology, Florence. 13-17 June, 1993.
% superficial velocity of the air stream [ 8 ] SM. Zarook and B.C. Bahzis, Chem. Eng. Sci.. 49 (24A) ( 1994) 4347.
X distance in the biofilm [9] D.S. Hodgeand J.S. Devinny.J. Env. Eng., 12 (1) (1995) 21.