Download introduction

Modelling Biological Cr(VI) Reduction in Aquifer Microcosm
Column Systems
Pulane E. Molokwane* and Evans M. N. Chirwa
Environmental Engineering Group, Department of Chemical Engineering, University of Pretoria,
Pretoria 0002. Tel. +27(12)420-5894, Fax +27(12)362-5089, Email: *[email protected]
Abstract
Several chrome processing facilities in South Africa release Cr(VI) into groundwater resources. Pumpand-treat remediation processes have been implemented at some of the sites but have not been
successful in reducing contamination levels. The current study is aimed at developing an
environmentally friendly, cost effective and self-sustained biological method to curb the spread of
chromium at the contaminated sites. An indigenous Cr(VI) reducing mixed-culture of bacteria was
demonstrated to reduce high levels of Cr(VI) in laboratory samples. The effect of Cr(VI) on the
removal rate was evaluated at concentrations up to 400 mg/L. Following the detailed evaluation of
fundamental processes for biological Cr (VI) reduction, a predictive model for Cr(VI) breakthrough
through aquifer microcosm reactors was developed. The reaction rate in batch followed noncompetitive rate kinetics with a Cr(VI) inhibition threshold concentration of approximately 99 mg/L.
This study evaluates the application of the kinetic parameters determined in the batch reactors to the
continuous flow process. The model developed from advection-reaction rate kinetics in a porous media
fitted best the effluent Cr(VI) concentration. The model was also used to elucidate the logistic nature
of biomass growth in the reactor systems.
Keywords: microcosm reactor, aquifer media, reaction inhibition, advection-reaction, in situ
bioremediation.
INTRODUCTION
Chromium is a common contaminant in surface water and groundwater resulting from numerous
industrial, agricultural and military activities (Ehrlich, 1996; Reid et al., 1994; Nriagu and Nieboer,
1988). Hexavalent chromium is known to be both carcinogenic and mutagenic to a wide range of
living organisms including mammals (U.S.EPA, 1978). The common treatment method for Cr(VI)
contaminated groundwater is usually the pump-and-treat method usually accompanied by
precipitation of Cr(VI) using chemical agents. This method proves costly and slow in removing
Cr(VI) from the aquifer media. At heavily contaminated sites, it may take up to 30 years to flash
out single spill pollution in the order of 3000 kg (Molokwane et al., 2008).
Alternatively, biological treatment methods using aquatic biomass and/or Cr(VI) reducing bacteria
have been proposed (Wang et al., 1989; Ohtake et al., 1990; Ganguli and Tripathi, 2002 and
others). These methods may be applied ex situ (Colica et al., 2010; Zakaria et al., 2007) or in situ in
biological barriers (Molokwane and Chirwa, 2009). When applied in situ, it is desirable to use
bacteria from the same or nearby environments to avoid the problem of transfer of alien organisms
across country or regional boundaries.
Permeable reactive barriers have been used mainly for removal of toxic organic compounds using
specialised species of bateria. A typical design for the latter purpose comprises of a double-zone
barrier system with an aeration zone followed by the bioremediation zone. One such system was
evaluated against the removal of methyl-tert-butyl-ether (MTBE) contaminated groundwater (Liu et
al, 2006). Another documented application is the treatment of petrochemical pollutants (i.e.,
benzene, toluene, ethylbenzene, xylene and polyaromatic hydrocarbons), heavy metals (i.e., lead,
arsenic etc), and cyanide in the system designed by Doherty et al (2006) using a modified ash
system. The latter biological permeable reactive barriers (BPRB) system was implemented at an
abandoned gas manufacturing plant.
Specific application of the biological permeable reactive barrier (BPRB) system for the removal of
Cr(VI) in groundwater has only been attempted lately (Jeyasingh at al., 2011). This slow
advancement towards full application has been both due to the unavailability of microorganisms
capable of growing under nutrient deficient conditions and lack of information on the speciation
and mobility of chromium species in the soil.
In this study, operation of an inoculated permeable barrier is mathematically simulated using a
dispersion-reaction model using Cr(VI) reducing microorganisms as biocatalysts. The general
performance of the system and the analysis of the behaviour or microorganisms is presented in the
earlier articles (Molokwane and Chirwa, 2009).
MATERIALS AND METHODS
Microbial Cultures
The mixed culture of bacteria was obtained from dried sludge collected from sand drying beds at
the Brits Wastewater Treatment Works (North West Province, South Africa). The treatment plant
receives periodic flows from a nearby abandoned sodium dichromate processing facility reported to
discharge high levels of Cr(VI) in the sewerage works, thus the bacteria at the treatment plant is
expected to be acclimated to Cr(VI) toxicity. Microbial cultures from the sludge were used to
inoculate the columns directly by feeding source organisms before running the sterile experimental
feed into the reactors.
Microcosm Reactor Systems
Soil columns extracted from the aquifer (below the water table) were installed in a continuous dose
experiment. Eight microcosm columns were installed with the first acting as the control, Reactor
HR2 was used to evaluate the sludge bacteria acting alone, and Reactors HR3 and HR4 were used
to evaluate the native soil bacteria acting alone (in duplicate). The main experiments comprised (in
duplicate) HR5 and HR6 with both sludge bacteria and native soil bacteria but operated without
carbon source, and HR7 and HR8 with soil bacteria and sludge bacteria operated with added carbon
source. The carbon source in HR7 and HR8 consisted of a natural matrix of organics leached from
saw dust. This was intended to simulate humic organics leaching from stems of dead plants in the
veldt. The experimental plan for the detailed evaluation of the performance of aquifer microcosm
reactors is summarised in Table 1.
Table 1: Conditions for the aquifer microcosm range of experiments.
Reactor(s)
Experiment
Reactor HR1
Sterilised (killed native bacteria) + no inoculation
Reactor HR2
Sterilised (killed native bacteria) + inoculated (sludge bacteria)
Reactors HR3 and HR4
Live native soil bacteria (non-sterile) + no inoculation
Reactors HR5 and HR6
Live native soil bacteria (non-sterile) + inoculated (sludge bacteria)
Reactors HR7 and HR8
Live native soil bacteria (non-sterile) +inoculated + carbon source
Cr(VI) and Total Cr Measurement
Cr(VI) was measured using a UV/Vis spectrophotometer (WPA, Light Wave II, Labotech, South
Africa). The measurement was carried out at a wavelength of 540 nm (10 mm light path) after
acidification of 0.2ml samples with 1N H2SO4 and reaction with 1,5-diphenyl carbazide to produce
a purple colour (APHA, 2005). Total Cr was measured at a wavelength of 359.9 nm using a Varian
AA-1275 Series Atomic Adsorption Spectrophotometer (AAS) (Varian, Palo Alto, CA (USA))
equipped with a 3 mA chromium hollow cathode lamp. Cr(III) was determined as the difference
between total Cr and Cr(VI) concentration.
Cr(VI) Reduction Kinetics
The following simple Monod-type enzymatically based kinetic model derived earlier by Shen and
Wang (1994) was utilised in this study to simulate Cr(VI) reduction in batch systems:
r
 d (C ) k mc  C  X

dt
C  Kc
(1)
where kmc = maximum specific Cr(VI) reduction rate coefficient (T-1), Kc = half velocity
concentration (ML-3), C = Cr(VI) concentration (ML-3), and X = concentration of viable cells (ML-3)
at any time t (T).
When non-competitive inhibition, maximum Cr(VI) reduction capacity of cells, and toxicity
threshold is taken into consideration, the model is revised as shown by Molokwane (2010) as
follows:

k C
C C 
dC
 X 0  0

  1-Cr C0mc
dt
Rc 
K
 K c  C  
(2)
where Xo = initial concentration of viable cells (ML-3), Co = initial concentration of Cr(VI) (ML-3),
Rc = Cr(VI) reduction capacity of the cells (g Cr(VI) reduced/g cells inactivated), K = noncompetitive inhibition rate coefficient (ML-3), and Cr = toxicity threshold concentration of Cr(VI)
(ML-3).
Simulation of Cr(VI) Removal Across Column
A dispersion-reaction model was used to determine kinetic and dynamic parameters for the
microbial barrier system. The microcosm reactors were modelled as plug flow reactors with the
Cr(VI) removal influenced by the following internal processes: (1) advection influenced by the
interstitual velocity u (LT-1), (2) mass transport into media particles governed by mass transport rate
coefficient kL (LT-1), (3) adsorption rate governed by mass transport and surface reaction, (4)
Cr(VI) reduction governed by the kinetics described by preliminary batch modelling and (5) cell
replacement rate with the cells acting as the catalyst in the Cr(VI) reduction process.
The predominant processes in the reactor are thus limited to advection, reduction, mass transport
and cell growth. These processes were used in the mass balance for Cr(VI) removal across the bulk
liquid phase in the microcosm reactor:
l  Li
d (C  V )
 A uCin  C   jc .ai  qc  rc   V
dt
l 0
(3)
in which for each segment of reactor of length L (L), V = change in reactor volume (L3), the
interstitial velocity u (LT-1) is assumed to be constant throughout the entire reactor, ai = surface area
in the segment (L2). The adsorption rate, qc (ML-3T-1)  approaches zero as the reactor is operated
for 6 to 10 hours (this study). The Cr(VI) reduction rate, rc (ML-3T-1), is a function of viable
biomass in the reactor (Equations 1 & 2). The above fundamental processes in the reactor during
transient state operation were represented by the Equations 4 to 8 below:
dV
uA
dt
(4)
dC
 k L aC  C s    jc
dt
(5)
dC
 k ad C eq  C    q c
dt
(6)

k C
C C 
dC
 X 0  0
  rc
  1-Cr C0mc
dt
Rc 
K
 K c  C  
(7)
 k S 
dX
 Y  ms   k d X
dt
 Ks  S 
(8)
where Cs and Ceq = are the particle surface and equilibrium surface concentration for the adsorptive
process (ML-3), the coefficient kL = mass transport rate coefficient (LT-1), a = total surface area in
the reactor (L2), kad = adsorption rate coefficient (T-1), Y = cell yield coefficient (MM-1), kms =
specific substrate utilisation rate coefficient (T-1), Ks = half velocity substrate concentration (ML-3),
S = the concentration of carbon sources (ML-3) at time t, and kd = cell death rate coefficient (T-1).
The interstitial space A in the microcosm was estimated as the volume of the mobile phase (bulk
liquid) minus entrained water determined as the difference between the weight of the wet nonflowing reactor and a dry reactor. And process terms are defined as follows: jc = mass transport rate
(ML-2T-1), qc = adsorption rate (ML-3T-1), and rc = Cr(VI) reduction rate (ML-3T-1).
The substrates S and viable biomass could not be measured accurately with current methods. Thus,
for the time being, these are generated as model outputs from an empirical logistic biomass term
(Equation 9 below):
X  X0 
X max
t
1  
 t0



b
(9)
with Xo = the initial viable biomass (ML-3) in the reactor, Xmax = maximum achievable biomass
(ML-3) due to toxicity and space limitations, t0 = x-intercept for the logistic curve, and b = pitch
factor (dimensionless) determining the amplitude of the biomass curve. The predicted
concentrations were calculated using the Computer Programme for Simulation of Aquatic Systems
(AQUASIM 2.0) (Reichert, 1998).
RESULTS AND DISCUSSION
Cr(VI) Reduction in a Control Column
The operation of the sterilised reactor without added biomass showed a characteristic rise following
the effluent tracer line for the clean-bed reactor system (Figure 1). The characteristic exponential
rise suggests that adsorption processes were insignificant, i.e., the column reached equilibrium with
respect to adsorption during the 45 days of operation. The extended controls Reactors (HR2, HR3
and HR4) show that soil bacteria and activated sludge bacteria acting alone in these conditions did
not reduce Cr(VI). The reason why the soil bacteria could not reduce Cr(VI) is most certainly
because there were no Cr(VI) reducing bacterial species in the soil. However, it is not known why
the sludge bacteria did not reduce Cr(VI) after inoculating a sterile column. One suggestion is that
the Cr(VI) reducing bacteria from the sludge required some biochemical metabolites or cofactors
produced by the soil bacteria.
Cr(VI) concentration, mg/L
60
50
40
30
Tracer line
Influent Cr(VI)
Effluent Cr(VI)
20
10
0
0
10
20
30
40
50
Time, days
Figure 1: Performance of a non-inoculated sterile column showing a characteristic exponential
rise in effluent Cr(VI) comparable to the tracer. In the abiotic system, the effluent concentration
will reach influent levels after 4.5 bed volumes.
Cr(VI) Reduction in a Column without Carbon Source
The performance of live cultures from both sources (soil and sludge) working together was
evaluated in system HR5 (Figure 2). The Reactor HR5 represents the operation on only inorganic
carbon sources from soil. Simulation of the system showed that cell growth is possible within the
constraints of the model, from an approximate initial concentration of 8.5 g/m3 to a maximum value
of approximately 45.5 g/m3. The biomass was simulated empirically at this point by the logistic
function (Equation 9). The cell viability was reflected in the increased Cr(VI) reduction removal
rate after 6 bed volumes (day 6). The characteristic flattening of the curve suggests an approach to
the maximum cell growth and Cr(VI) reduction capacity of the reactor.
Cr(VI) Reduction in a Column with Carbon Source
The best performance was observed in Reactor HR7 operated under carbon source (Figure 3). In
this reactor, the simulated viable biomass continually increased with time and this was reflected in
the ever increasing rate of Cr(VI) reduction until the termination of the experiment at day 45. The
improved performance under carbon source was also demonstrated by the ever-increasing
cumulative Cr(VI) reduction in the reactor HR7 (data not shown). The trend in biomass from day 6
to 45 shows no sign of levelling off. This indicates a relaxation in the pitch factor, b (Equation 9),
showing that the full capacity was not reached in the reactor operated under carbon sources.
70
60
50
50
40
40
30
30
20
20
Measured influent Cr(VI)
Measured effluent Cr(VI)
Simulated effluent Cr(VI)
Simulated biomass Xin
10
Viable cell conc. X, g/m3
Cr(VI) concentration, mg/L
60
10
0
0
0
10
20
30
40
50
Time, days
Figure 2: Performance and simulation of the live soil culture microcosm inoculated with
live cultures from sludge and operated without carbon source (Reactor HR5).
70
60
50
50
40
40
30
Measured Influent BR7
Measured Effluent BR7
Simulated effluent Cr(VI)
Simulated biomass Xin
20
30
20
10
Viable cell conc. X, g/m3
Cr(VI) concentration, mg/L
60
10
0
0
0
10
20
30
40
50
Time, days
Figure 3: Model simulation of the live soil culture microcosm inoculated with live cultures from
sludge and operated with carbon source (Reactor HR7).
In this simulation, the system performed well regardless of having started with the lowest initial
biomass value, Xin = 4.5 g/m3. The biomass in the reactor increased to approximately 65.5 mg/L, a
value more than 1.5 times higher than the maximum achievable cell concentration under no carbon
source (Reactor HR5). The logistic time constant t0 almost doubled to 10.4 days as X continued to
increase. The pitch factor, b, and interstitial velocity, u, remained the same as under the other
simulations conditions at 5.03 and 6.24  10-4 m/d, respectively. The increase in biomass with time
in the reactor was accompanied by increasing Cr(VI) reduction which continued until the
experiment was terminated at 45 days.
Effect of Toxicity Inhibition
The Cr(VI) reducing culture used in this experiment is more sensitive to Cr(VI) toxicity compared
to batch cultures grown aerobically. This is due to low access to the organic carbon sources
generally, and specifically (Reactor HR5) due to growth under autotrophic conditions. The model
was modified for toxicity sensitivity by introducing the Cr(VI) inhibition term and the Cr(VI)
toxicity threshold and the cell inactivation term X0Rc in the Monod-like equation (NkhalambayausiChirwa and Wang, 2004). A reasonable accuracy was achieved in the final optimised parameters as
shown in the Figures 2 and 3  simulated effluent model line. The optimum parameters in the final
simulation are shown in Table 2.
Table 2: Final parameter values from the application and optimisation in the microcosm
reactors.
Parameters
Associated Process
Optimum Value
Units
kmc
Kc
Rc
K
Cr
t0
X0
Xmax
D
A
Cr(VI) reduction rate
Cr(VI) reduction rate
Cr(VI) reduction rate
Cr(VI) reduction rate
Cr(VI) reduction rate
Cell growth
Cell growth
Cell growth
Column properties
Column properties
1.385
2.450
0.533
0.50
99
10.4-20.4
4.5-23.5 (HR7-HR4)
3.5-65.3 (HR4-HR7)
98.4
1/d
mg/L
mg/mg
mg/L
mg/L
d
g/m3
g/m3
m2/h
m2
3.65-4.65  10-4
Conclusion
Model simulation confirms the effect of inhibition (inhibition threshold Cr) and direct toxicity (cell
inactivation coefficient Rc) as the dual mechanisms limiting the extent of Cr(VI) reduction in the
aquifer media. The accurate determination of these parameters is critical in the future development
of a pilot scale system for Cr(VI) reduction and immobilization using a microbial permeable barrier
system. Other parameters of importance are the interstitial velocity u and the diffusivity of Cr(VI)
in the soil media which was determined to be slower than the value in water. Overall, the model
successfully captured the effluent trend and Cr(VI) reduction capacity of the columns. However,
only an empirical evaluation of biomass growth was possible at this point. A more accurate model
for the growth term is under investigation based on improved methods for determination of viable
biomass in soil media.
Acknowledgements
The research was funded through the National Research Foundation (NRF) of South Africa through
the Focus Areas Programme Grant No. FA2006031900007 and the Incentive Funding for Rated
Researchers Grant No. IFR2010042900080 awarded to Evans M.N. Chirwa of the University of
Pretoria.
References
1. APHA (2005). Standard Methods for the Examination of Water and Wastewater - Centennial Edition.
American Public Health Association, American Water Works Association, Water Environment
Federation, USA.
2. Colica G., Mecarozzi P.C., De Philippis R. (2010). Treatment of Cr(VI)-containing wastewaters with
exopolysaccharide-producing cyanobacteria in pilot flow through and batch systems, Applied
Microbiology and Biotechnology, 87 (5), 1953-1961.
3. Doherty, R., Phillips, D.H., McGeough, K.L., Walsh, K.P., Kalin R.M. (2006). Development of
modified flyash as a permeable reactive barrier medium for a former manufactured gas plant site,
Northern Ireland. Environmental Geology, 50, 37-46.
4. Ehrlich, H.L. (1996). GeoMicrobiology, 3rd ed.; Marcel Dekker: Inc.: New York, USA.
5. Ganguli, A., Tripathi, A.K. (2002). Bioremediation of toxic chromium from electroplating effluent by
chromate-reducing Pseudomonas aeruginosa A2 Chr in two bioreactors, Applied Microbiology and
Biotechnology, 58 (3), 416-420.
6. Jeyasingh, J., Somasundaram, V., Philip, L., Bhallamudi, M.S. (2011). Pilot scale studies on the
remediation of chromium contaminated aquifer using bio-barrier and reactive zone technologies.
Chemical Engineering Journal, 167 (1), 206-214.
7. Liu, S.-J., Jiang, B., Huang, G.-Q., Li, X.-G. (2006). Laboratory column study for remediation of
MTBE-contaminated groundwater using a biological two-layer permeable barrier. Water Research, 40
(18), 3401-3408.
8. Molokwane P.E., Nkhalambayausi-Chirwa E.M., Meli K.C. (2008). Chromium (VI) reduction in
activated sludge bacteria exposed to high chromium loading: Brits culture (South Africa). Water
Research, 42 (17), 4538-4548.
9. Molokwane P.E, Chirwa E.M.N. (2009). Microbial culture dynamics and chromium (VI) removal in
packed-column microcosm reactors. Water Science and Technology, 60 (2), 381-388.
10. Molokwane, P.E. (2010). Simulation of in situ bioremediation of Cr(VI)) in groundwater aquifer
environments suing a microbial culture barrier. Ph.D. Thesis, University of Pretoria, Pretoria, South
Africa, pp. 50-62.
11. Nkhalambayausi-Chirwa E.M., Wang Y.-T. (2004). Modeling Cr(VI) reduction in a Bacillus sp. pure
culture biofilm reactor. Biotechnology and Bioengineering, 87 (7), 874-883.
12. Nriagu, J. O., Nieboer, E., (1988). Chromium in the Natural and Human Environments; John Wiley and
Sons: New York, NY.
13. Ohtake H., Fujii E., Toda K. (1990). Reduction of toxic chromate in an industrial effluent by use of a
chromate-reducing strain of Enterobacter cloacae. Environmental Technology Letters, 11 (43), 663-668.
14. Reichert, P. (1998). Swiss Federal Institute for Environmental Science and Technology (EAWAG),
Switzerland, ISBN: 3-906484-16-5.
15. Reid, V.M., Wyatt, K.W., Horn, J.A. (1994). A new angle on groundwater remediation. Civil
Engineering, 64 (4), 56-58.
16. Shen, H., Wang, Y.-T. (1994). Modeling hexavalent chromium reduction in Escherichia coli ATCC
33456. Biotechnology and Bioengineering, 43 (4), 293-300.
17. U.S.EPA (1978). Reviews of the environmental effects of pollutants - III. Chromium, EPA 600/1-78023, Washington, District of Columbia.
18. Wang, Y.T., Shen, H. (1997). Modelling Cr(VI) reduction by pure bacterial cultures. Water Research,
41 (4), 727-732.
19. Wang, P., Mori, T., Komori, K., Sasatsu, M., Toda, K., Ohtake, H. (1989). Isolation and characterization
of an Enterobacter cloacae strain that reduces hexavalent chromium under anaerobic conditions.
Applied and Environmental Microbiology, 55 (7), 1665-1669.
20. Zakaria, Z.A, Zakaria, Z, Surif, S., Ahmad, W.A. (2007). Biological detoxification of Cr(VI) using
wood-husk immobilized Acinetobacter haemolyticus. Journal of Hazardous Materials, 148 (1-2), 164171.