A new mathematical model to evaluate simazine removal in three different immobilized-biomass reactors

ARTICLE IN PRESS WAT E R R E S E A R C H

42 (2008) 1035 – 1042

Available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/watres

A new mathematical model to evaluate simazine removal in three different immobilized-biomass reactors M. Martina,, L. Casasusb, C. Garbia, M. Nandea, R. Vargasa, J.I. Roblac, M. Sancheza, J.L. Allended a

Departamento de Bioquimica y Biologia Molecular IV, Facultad de Veterinaria, Universidad Complutense de Madrid (UCM), Avda. Puerta de Hierro s/n, 28040 Madrid, Spain b Departamento de Matematica Aplicada, E.T.S.I. Industriales, Universidad Politecnica de Madrid (UPM), C/Jose´ Gutie´rrez Abascal, 2, 28006 Madrid, Spain c Centro Nacional Investigaciones Metalurgicas (CENIM), Consejo Superior de Investigaciones Cientificas (CSIC), Avda. Gregorio del Amo, s/n. 28040 Madrid, Spain d Departamento de Fisica Aplicada, Facultad de Veterinaria, Universidad Complutense de Madrid (UCM), Avda. Puerta de Hierro s/n, 28040 Madrid, Spain

art i cle info

ab st rac t

Article history:

A new mathematical model based on the cinetical Langmuir equation is developed to

Received 21 December 2006

interpret and predict the effectiveness of simazine (SZ) removal in immobilized-biomass

Received in revised form

reactor (IBR), to consider herbicide-support affinity (Cx), and herbicide-cell affinity (Cy).

18 July 2007

Three solid supports: sepiolite monolith, granular sepiolite, and alginate were used in pilot-

Accepted 26 September 2007

scale reactors that were inoculated with Klebsiella planticola DSZ. The abiotic process was

Available online 4 October 2007

analysed by measuring the SZ sorption capacity of the reactor supports. Sepiolite monolith

Keywords: Simazine Biomass-reactor Mathematical model Biofilm

showed the maximum value for herbicide-support affinity (28.0270.9%). The effectiveness of the biotic process was estimated considering the formation of biomass and SZ biodegradation. Granular sepiolite showed either higher affinity with SZ and viability rate (0.90) throughout the process, and SZ removal rate was 3.3970.06 mg/h. The mathematical model presented in this paper provides useful insights into the interpretation of experimental data as well as prediction for the implementation of biological reactors. & 2007 Elsevier Ltd. All rights reserved.

1.

Introduction

s-Triazine herbicides have been used in a variety of weedcontrol programs with major crops, but herbicides containing an s-triazine ring are relatively persistent in the environment. Simazine (SZ) is a synthetic chemical that is widely used as a selective triazine herbicide to control the growth of broadleaved weeds and annual grasses in field, berry fruit, nuts, vegetable and ornamental crops, turfgrass, orchards, and

vineyards. At higher concentrations, it is used for nonselective weed control in industrial areas, and before 1992, it was used to control submerged weeds and algae in large aquariums, farm ponds, fish hatcheries, swimming pools, ornamental ponds, and cooling towers. The increasing awareness of the harmful effects of environmental pollution has led to a notable increase in research on various strategies that may be employed to clean up the environment (Vandecasteele et al., 2000; Wackett et al., 2002). It is now accepted

Corresponding author. Tel.: +34 913943911; fax: +34 913943886.

E-mail addresses: [email protected] (M. Martin), [email protected] (L. Casasus), [email protected] (C. Garbi), [email protected] (M. Nande), [email protected] (R. Vargas), [email protected] (J.I. Robla), [email protected] (J.L. Allende). 0043-1354/$ - see front matter & 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2007.09.026

ARTICLE IN PRESS 1036

WAT E R R E S E A R C H

42 (2008) 1035– 1042

that microbial metabolism provides a safer, more efficient, and less expensive alternative to physicochemical methods for pollution abatement (Pandey and Jain, 2002; Boudreau and Daugulis, 2006; Zheng et al., 2006). The natural tendency of many microorganisms to bind to solid surfaces (Chen et al., 2005; Jefferson, 2004; Zacarias et al., 2005) has been utilized in a number of biotechnological processes. For example, immobilized microorganisms are used for the decontamination of wastewater in fixed/film treatment plants with systems such as trickling filters, rotating biological contactors, or fluidised beds (Costerton et al., 1995; Garbi et al., 2006; Nicolella et al., 2000). In all these systems, the bacteria responsible for biodegradation are present in a microbial biofilm (Branda et al., 2005). Physiological cooperativity is a major factor in shaping the structure and in establishing the eventual juxtapositions that make mature biofilms very efficient microbial communities adhering to surfaces (Cogan et al., 2005; Heipieper and Bont, 1994; Panikov, 1995) and immobilized cells are known to show a greater tolerance of various antibiotics and xenobiotic compounds. Bioremediation systems, whether with free or with immobilized cells as in biofilm reactors, must be based on a sound knowledge of the processes and factors involved in the system. For the technology to be effective it must be shown to bring about a higher rate of biodegradation of the contaminant without any harmful environmental effects on the final products. The optimization of the system requires expert knowledge at the cellular level and an understanding of the macroscopic processes that affect the persistence and the possible elimination of the contaminants. The quest for this knowledge which underlies both the design of bioreactors and the modelling of the systems at cellular and/or macroscopic levels can now count on the recent advances in molecular technology and on mathematical modelling (Alonso-Sanz and Martin, 2005a, b, 2006; Laspidou and Rittmann, 2004) to predict the behaviour of a system from such variables as the dynamics of the contaminants, of the fluids, and of the biomass. The control and understanding of processes catalysed by biofilms are important from both industrial and ecological perspectives. Mathematical models represent one end of a spectrum of activities designed to investigate natural phenomena. They attempt to simplify systems to uncover relationships that yield a consistent pattern when compared with in situ behaviour (Garbi et al., 2006; Alonso-Sanz and Martin, 2006). The aim of this work is to develop a mathematical model to evaluate the effectiveness of herbicide removal in three different immobilized-biomass reactors (IBRs) by calculating the coefficients of herbicide-support affinity (Cx), and herbicide-cell affinity (CY). SZ biodegradation data from bioreactor experiments and simulations are compared, and the capacity of the model to predict the process is evaluated in this study.

2.

Materials and methods

2.1.

Pilot-scale immobilized-biomass reactor (IBR)

internal diameter of the reactor glass column of 8 cm, and a total height of 38 cm (Martin et al., 2000). Sterile monolith (500 g), granular sepiolite (500 g), or alginate beads (495 g) were set up in the reactor, and liquid samples were distributed over the packing material through a microsprinkler. Dissolved oxygen, pH, and temperature were monitored each minute by specific sensors connected to a Biocontroller ADI 1030 (Applikon). The experiments were performed at room temperature (2072 1C) by circulation of MB medium (Sanchez et al., 2005) and 0.025 mM SZ. The circulation flow was of 50 ml/min, and the reactor was operated over a cycle of 100 days. Samples were taken periodically to monitor the state of the chlorinated herbicide and microorganisms.

2.2.

Klebsiella planticola strain DSZ was isolated from agricultural fields in Alcala de Henares (SE-Madrid, Spain), and it grows on a wide range of s-triazine and aromatic compounds (MartinMontalvo et al., 1997; Sanchez et al., 2005). Cells were grown aerobically at 30 1C in MB medium and SZ was added to make up 0.025 mM. Cell immobilization by attachment was done by using sepiolite (Tolsa SA, Spain), a porous carrier material, with two different structures: monolith (Ferrer et al., 1996) and granular sepiolite (3–5 mm +) (Martin et al., 2000). The immobilization method is described elsewhere (Ferrer et al., 1996), we used the pure culture of DSZ strain grown on 0.025 mM SZ as the cell source, harvesting cells at the exponential phase for use as the inoculum in the reactor experiments. Cell immobilization by entrapment was carried out with alginate beads (Gibello et al., 2005). DSZ strain cells grown on 0.025 mM SZ were suspended in a 3% alginate suspension by stirring for 20 min, and the mixture was extruded through a needle into a 1% CaCl2 solution. The beads were collected in 0.9% NaCl and washed twice with saline solution. For the immobilized-cell assays an inoculum of 1.15  1010 cells per gram of alginate beads was used. Cell quantification in the immobilized-cell samples from the reactors was carried out by using 40 ,60 -diamidino-2phenylindole (DAPI) for DNA staining. DAPI solution (0.55 mM) in distilled water was added to each sample and left in contact for 10–15 min at room temperature in the dark. The remaining DAPI solution was removed by rinsing twice with distilled water and finally air-dried. Each sample was mounted with drops of Vecta-Shield on a slide and the preparations were examined by confocal laser scanning microscopy (CLSM).

2.3.

Chemicals

SZ (99% purity), and [U-ring 14C] SZ (5mCi mmol1, 95% radiochemical purity) were purchased from Sigma-Aldrich (St. Louis, Mo). All the chemical compounds were of the highest purity commercially available.

2.4.

To evaluate the effectiveness of decontamination of the IBR, the experiments were performed in a bioreactor with an

Biomass immobilization

Chemical analyses

SZ was measured by HPLC as previously reported by Sanchez et al. (2005), and the analyses were performed with a Waters model 616PDA996 photodiode array detector equipped with a

ARTICLE IN PRESS WAT E R R E S E A R C H

42 (2008) 1035 – 1042

Millennium 20/10 for data analysis. Separation was done in a Novapack C-18 (3.9  150 mm2) column, using a mobile phase consisting of 40% acetonitrile in water at a flow rate of 0.5 ml/min. SZ was monitored at 214 nm. The injection volume was 10 ml, and SZ was identified by coelution with the standard in the HPLC analysis. The analytical curves were made using five different concentrations (10 mg/l1–5 mg/l), with three replicates each. Limits of detection and quantification show that the method could detect SZ in concentrations below the maximum residue levels established by the European Union legislation. SZ adsorbed by cell membranes was identified by gas chromatography–mass spectra (GC–MS) as previously reported by Sanchez et al. (2005). The analyses were done with a Hewlett-Packard model 5890 Series II gas chromatograph equipped with a VA-5 capillary column (30 m, 0.25 mm i.d.), programmed from 80 to 290 1C (15 1C/min), and connected to an HP-5989A quadrupole mass detector. Solid-phase extractions were done with Varian C18 and SCX Cartridges, acetone was used as the eluent, and 2 ml aliquots were injected into the column. SZ was identified by comparing their electron impact-MS spectra with a standard sample and by coelution in GC.

2.5.

Sorption/desorption experiments

Sorption isotherms were carried out by the standard batch equilibration method (OECD, 2000). [U-ring 14C] SZ with a specific activity of 3.14 MBq/mg was used in our experiments. A non-radio-labelled solution of SZ was prepared separately and then mixed with [U-ring 14C] SZ to achieve the desired concentrations: 0.001, 0.0025, 0.005, 0.025, and 0.05 mM. The assays were carried out in triplicate, in the reactors sealed with Teflon stoppers, containing sterile 20 mM PBS and sterile solid supports to bring the samples up to water saturation. After 12 h in the dark at room temperature (2171 1C), the supernatants were taken and replaced by labelled solutions. According to previous kinetic studies (Garbi et al., 2006), equilibrium was reached within the 12 h equilibrium period. Samples (2 ml) were taken in duplicate from each reactor, and the concentration of SZ was determined by liquid scintillation counting and verified by HPLC. The amount of solute sorbed was determined by the difference between the amount of 14Cradioactivity added and that recovered in liquid phase. To validate this method, SZ was extracted with methanol from selected samples for each ceramic support and the concentrations of SZ in the extracts were verified by HPLC. Desorption experiments were carried out immediately after the sorption experiments to prevent further degradation. The starting SZ concentration was 0.025 mM, by successive dilution. The supernatants were replaced with sterile 20 mM PBS, and the capped syringes were incubated in the dark at room temperature (2171 1C) for 12 h. The supernatants were periodically sampled in duplicate, and the SZ concentration in the liquid phase was determined by HPLC and liquid scintillation counting, using the above procedure. Once the supernatants were sampled and analysed for SZ, the syringes were refilled with 20 mM PBS to the original volume. This procedure was repeated for a total of five successive desorption periods.

2.6.

1037

Microscopy analyses

Cell viability in immobilized-cell samples was analysed by laser scanning confocal microscopy (LSCM) as previously reported (Martin et al., 2000; Garbi et al., 2006). A Leica TCS SP2 CLSM was set up with the standard configuration. Immobilized cell samples from the surface and from inside the reactor were stained with two fluorescent dyes, 10 mg/ml SYTO-13 and 5 mg/ml propidium iodide (PI) (Molecular Probes, Europe BV). The samples were incubated for 15 min, at RT, and washed twice in PBS to eliminate the remaining dyes. The green emission from SYTO-13 was collected through 509 and 514 nm band-pass filters. The red emission from PI was collected at 610 nm. Viability data were measured in triplicate. The area measured was 0.23 mm2, with an approximate number of cells of 0.2  106. Alginate immobilized cells were stained with the two fluorescent dyes mentioned above, and analysed by LSCM as previously reported by Gibello et al. (2005). Results were expressed as viability rate defined as the number of live cells/total number cells (Garbi et al., 2006).

2.7.

Statistical analysis

The coefficient of determination, the square of the Pearson’s product moment correlation coefficient (R2) was applied to describe the proportion of the total variance in the experimental data that can be explained by the model. (Legates and McCabe, 1999).

3.

Results and discussion

3.1.

Mathematical model

We developed a model for the sorption of substrate in biological decontamination processes, describing the sorbed mass of the substrate x(t) per unit solid support mass and the sorbed/degraded mass of the substrate y(t) per unit cell mass. The equations of this model are dx=dt ¼ ðCx z  xÞðxmax  xÞ,

(1)

dy=dt ¼ ðCy z  yÞðymax  yÞ,

(2)

dz=dt ¼ mx ðdx=dtÞ  my ðdy=dtÞ,

(3)

where z(t) is the concentration of sorbate in the water, Cx and Cy the relative affinity coefficients for the solid support and the biocatalizer, respectively, and mx the total mass of solid support/total mass of water, my the total mass of cellular material/total mass of water. Eqs. (1) and (2) were obtained after a suitable modification of the classical cinetical Langmuir equation, i.e., dx=dt ¼ kðCzðxmax  xÞ  xÞ, where k and C are constants and xmax the saturation value of x. The term x(xmax–x) in Eq. (1) represents the partially irreversible process of desorption in porous media. The sorption term Cx z (xmaxx) is bilinear: (i) z is the concentration of free monomer, and (xmaxx) expresses the residual capacity of sorption. The term for desorption is also bilinear, dependent on the amount of sorbed contaminant (x)

ARTICLE IN PRESS 1038

WAT E R R E S E A R C H

42 (2008) 1035– 1042

the processes. The Cx values of simulation (Eq. (1)) showed that the affinity between the support and the SZ is greatest in the monolith and lowest in the alginate, and this was also observed in the experimental data. In the three IBR the decreasing evolution of Cx agreed with the well-known sorption/desorption process, but these data suggest that sepiolite, in both structures, may be an effective sorbent for the herbicide. We had reported the high affinity of this ceramic support for xenobiotic compounds when it was used in a propachlor-removing bioreactor (Martin et al., 2000). Sepiolite has a large specific surface and high porosity that together with the presence of zeolitic channels enable it to absorb liquids of different polarity. These properties are important in processes that require little mass transfer as is the case in xenobiotics dissolved in an aqueous medium (Garbi et al., 2006; Shreve and Vogel, 1993). In addition, active centres on the surface of sepiolite allow selective adsorption of various types of molecules (Garbi et al., 2006). Alginate, however, has practically no properties of sorption, as is shown in Table 1.

and on the residual capacity of sorption of the support. (ii) Eq. (2) introduces the biotic process, that is, the affinity between the bacteria and the contaminant. We only consider the effect of attached cells to sorbent; their biodegradation and mineralization rates are proportional to the term (Cyzy) in Eq. (2), describing the concentration gradient between membranes and the aqueous phase. (iii) Eq. (3) takes into account the specific characteristics of the immobilized cells as well as those of the sorption/desorption of the support. Our experiments give the sorption equilibrium states, and the model provides the solution of an ‘‘inverse problem’’, determining the relative time-dependent affinities Cx and Cy. We numerically solve Eqs. (1)–(3) for a range of values of the parameters, determined by the equilibrium values [(dx/dt) ¼ (dy/dt) ¼ (dz/dt) ¼ 0]. The effectiveness of herbicide removal by three different immobilized-biomass systems was evaluated by this mathematical model, which gives a prediction of the capacity to remove—(sorption/biodegradation)—a contaminant in an IBR, from the characteristics of the support and of the degradation capacity of the bacteria strain that acts as the biocatalizador. The model allows a distinction to be made between the abiotic process of calculation of the specific coefficient of affinity between the support and the contaminant (Cx) and the biotic process by calculation of the Cy coefficient, that which defines the affinity of the immobilized cell with the contaminant and its biodegradative capability.

3.2.

3.3.

Biotic process

One of the parameters considered in the mathematical model was the SZ affinity and the capacity of biodegradation of the bacteria immobilized on the support. The bacterium used in this study, K. planticola strain DSZ, was able to use different xenobiotic compounds as the carbon source; among them were some herbicides such as propachlor, alachlor, atrazine, and SZ. This strain grows on SZ, which can be used as nitrogen and carbon sources at concentrations between 0.01 and 0.1 mM in liquid cultures (Sanchez et al., 2005). When DSZ cells were inoculated in a liquid medium containing 0.025 mM SZ, in 60 s an important decrease of the herbicide concentration (15% of the initial SZ) was observed when the SZ concentration was measured in the culture medium. The presence of SZ in the cell membranes was assessed by GC–MS, a main compound with m/z 201(M+1), and a retention time of 20.65 min was identified for SZ. Then, DSZ takes soluble SZ in an aqueous phase into its membrane by passive transport, and there it is held by its lipophillic nature. We reported previously (Sanchez et al., 2005) the capacity of DSZ cells to adhere to the crystal surface of the SZ, using it at the interphase for its growth, with an increase of the herbicide-cell

Abiotic process

Initially, the solid supports: sepiolite monolith (IBRM), granular sepiolite (IBRG), or alginate beads (IBRA) were placed in the reactor that operated under aerobic conditions for 100 days, and the SZ sorption capacity of the reactor supports was analysed. Table 1 shows the results of the experiments. The maximum value for sorption of SZ by the monolith support was 28.02(70.9)% higher than that obtained in granular sepiolite 19.0(70.8)%. Alginate beads showed the lowest simazine-sorption capacity (0.5(70.007)%). According to the model, the analyses of the values of Cx (expressed as mg SZ sorbed per gram support) revealed the affinity of the immobilization support with the contaminant. Fig. 1 shows the results of the three experimental systems and those of the proposed model, with their evolution during

Table 1 – Abiotic process: parameters for sorption of simazine of selected supports, and comparison of the simazinesorption capacity calculated by the mathematical model and experimental results from the IBR after equilibrium and at the end of operation of the reactors (%) Sorbed simazine Theoretical

IBRM IBRG IBRA

Experimental

Equilibrium

End operation

Equilibrium

End operation

23.75 19.95 0.76

8.27 7.12 0.71

28.0270.9 19.0070.8 0.570.007

7.1270.08 7.1270.08 0.7170.006

ARTICLE IN PRESS WAT E R R E S E A R C H

42 (2008) 1035 – 1042

IBRM

8

R2=0.92

CX

6 4 2 0 0

25

50

75

100

t (days) IBRG

5

R2=0.9

CX

4 3 2 1 0 0

25

50

75

100

t (days) IBRA

0.4

CX

R2=0.77

0.2

0

0

25

50 t (days)

75

100

Fig. 1 – Evolution of Cx values in the three IBR systems, during the cycle of 100 days. (~) Theoretical values calculated from Eq. (1), and (’) experimental data obtained from pilot-scale reactors. The coefficient of determination (R2) describes the proportion of the total variance in the experimental data that can be explained by the model.

affinity and the consequent degradation of the compound. Moreover, the membrane activity to assimilate SZ even at low temperatures was confirmed by membrane phase transition temperature (FTIR spectroscopy) (Sanchez et al., 2005), so K. planticola strain DSZ offers advantages for its use in decontamination systems on an industrial scale. To study the SZ removal by IBRM and IBRG, the reactors were inoculated with 2.5  107 cells per g ceramic support. The EM analyses of immobilized-cell samples from the operating reactors showed that within 60 min, a monolayer of bacteria formed on the ceramic surface of both systems, IBRM (Fig. 2A) and IBRG (Fig. 2B), and by 6–12 h, this monolayer almost completely covered the ceramic supports and was punctuated

1039

by microcolonies that became more numerous. The scanning electron micrographs (Fig. 2C) showed the biofilm formation, and the maximal thickness (125 mm) was reached after 50 days of reactor operation. In the IBRA, alginate beads containing 2.15  107 cells per gram of alginate were placed in the reactor, and after 4 days the LSCM analyses of the alginate beads showed that immobilized cells were mainly located in the peripheric area of the bead (Fig. 2D). The amount of biomass formed in the three reactors was assessed by DAPI throughout the experiments with SZ treatment. An analysis of the experimental results in Table 2 shows that IBRM increases the formation of biomass by 2.5 times that of IBRG, and that the lowest value was that of IBRA. The immobilized cells of DSZ showed neither morphological change as a consequence of their adhesion to the support nor of the presence of the xenobiotic; this accords with the data of Ferrer et al. (1996) of Pseudonomas PEM1 and of Martin et al. (2000) when using Pseudonomas GCH1 in a propachlor-removing reactor. The biodegradation process started when biomass fixation was concluded (5 days). The reactors operated under aerobic conditions, and the influent containing 0.025 mM SZ was circulated through the column at a flow rate of 50 ml min1. The HPLC quantification of the amount of SZ remaining in the effluent (7.8 mM (IBRG), 15.6 mM (IBRM), and 19.9 mM (IBRA)) during the process allowed the determination of the experimental values of Cy (expressed as mg SZ removed g1 biomass) (Fig. 3). According to the model, the analyses of the values of Cy revealed the effectiveness of the biotic process, and this coefficient was estimated in the three IBR studied by their SZ-removal rates. Fig. 3 presents the Cy evolution obtained for each experimental system and the values from the proposed model (Eq. (2)). In all three cases the downward trend of the experimental data reflected that of the Cy values. The Cy values in IBRM were lower than those obtained in IBRG and quite similar to the values in IBRA (Fig. 3). IBRM showed experimental data slightly higher at the end of the process than was predicted, but both were fairly constant through the process. Table 2 shows the SZ removal rates in the three systems calculated from the Cy values, considering biomass formation. These results showed that IBRM was not a very effective system for use in long SZ-removing processes, so SZ removal rate did not increase significantly during the process. However, data obtained from IBRG showed an increase of effectiveness during the biotic process (Fig. 3 and Table 2), in which the Cy coefficient was more than threefold higher than those obtained in IBRM. The effectiveness of the IBRG rose throughout the process to head the lists of the predicted Cy coefficients and in the experimental data (Table 2). The effectiveness in IBRA showed a decrease throughout the process with very low Cy values that fitted very well with the predicted Cy coefficients (Fig. 3), and agreed with the SZ removal rates obtained (Table 2). We suggested that in the IBR the cells could use the SZ both in solution and when sorbed by the support. Thomas et al. (1987) considered it probable that some of these sorbed compounds may be used directly by the microorganisms adhering to these surfaces. That would mean that the organism would make direct contact with the compound, penetrating the cell without entering the surrounding liquid,

ARTICLE IN PRESS 1040

WAT E R R E S E A R C H

42 (2008) 1035– 1042

Fig. 2 – Scanning electron micrographs of immobilized DSZ cells adsorbed onto the surface of the ceramic support, 60 min after inoculation: IBRM (A), IBRG (B), and biofilm formation (C). Laser scanning confocal micrographs showing cells immobilized in alginate beads (IBRA) (D).

in the way that compounds not soluble in water are used by bacteria that adhere to these substances. The data of the mathematical model show that the Cx values are modified since the cells using the SZ absorbed by the support would be changing the conditions of equilibrium, which means that support acquires greater affinity with the free SZ to restore equilibrium, bringing an increase of the coefficient of affinity (CY). The experimental data in IBRA show that the values of CY remain constant during the experiment. Cell viability was measured during the process in the three IBRs, and Table 2 shows viability rate after biomass fixation and at the end of the experiments. The viability rate after biomass fixation in IBRM was 0.82, lower than those obtained in IBRG and IBRA, which showed similar values, 0.95 and 0.92, respectively. A comparison of viability evolution in the three systems revealed that in IBRA cell viability rose during the first 10 days of the test, but fell by 70% on day 15 (data not shown), and the system kept a low viability rate (0.66) during the process. Both IBRM and IBRG systems showed a slight decrease in the viability rate values throughout the operation cycle.

An analysis of the results (Fig. 3 and Table 2) shows that IBRG is the system with the highest biological ability to eliminate the contaminant. It has a rising coefficient of CY, unlike the evolution of IBRM and IBRA. In IBRG the cells showed a higher affinity with the SZ throughout the process, which suggests that the ability of the cell to adapt to the SZ benefits from this support. This might be a consequence of the greater viability of the cells in the sepiolite, perhaps because of the irregularities of the porous support that promote the formation of micro-colonies accessible to both the absorbed SZ and to the free SZ in the channels of the sepiolite. The IBRG has also shown its effectiveness of herbicide removal working in a vineyard nursery at the concentration range 0.001–0.010 mM SZ. The use of bioreactors is intended to provide optimum conditions for the large-scale use of the cellular activities for the elimination of the contaminant. Biofilm reactors are used in situations wherein the reactor capacity obtained by using freely suspended organisms is limited by the biomass concentration and the hydraulic residence time (Shreve and Vogel, 1993).

ARTICLE IN PRESS WAT E R R E S E A R C H

1041

42 (2008) 1035 – 1042

Table 2 – Biotic process: parameters for immobilized biomass, viability rate, and comparison of the SZ removal rates calculated by the mathematical model and experimental results from the IBR after biomass fixation and at the end of operation of the reactors SZ removal ratea Theoretical

IBRM IBRG IBRA a b c

Viability rateb

Biomassc

Experimental

Initial (10 days)

End operation

Initial (10 days)

End operation

Initial (10 days)

End operation

1.17 1.28 0.34

1.15 1.85 0.24

0.9870.02 1.1270.06 0.3770.04

1.1770.05 3.3970.06 0.2770.01

0.82 0.92 0.95

0.79 0.90 0.66

3.1270.06 2.4570.04 0.9170.06

Simazine removal rate expressed as mg Simazine/h. Viability rate defined as number live cells/total number cells. Immobilized biomass expressed as mg dry weight/g support.

IBRM 0.6

CY

0.4 R2 =0.82 0.2 0 0

25

50 t (days)

75

100

IBRG

4.

2

CY

1. 5 1

R2 =0.9

0. 5 0 0

25

50

Different models have been proposed to explain the behaviour of systems in which the processes of sorption/ desorption occur simultaneously with biodegradation. Park et al. (2001, 2003) assume that there is no biodegradation of sorbed contaminant. Their results showed, however, that the enhanced mineralization rates could not be explained if instantaneous desorption was assumed. Living bacteria could degrade this material directly, without desorption, and this could occur through direct partitioning to the cell membrane or via degradation by extracellular enzymes. Nonetheless, the bioavailability of sorbed contaminant was suggested.

75

100

t (days) IBRA

Conclusions

The experimental results and those obtained with the proposed mathematical model coincide with their designation of granular sepiolite as the most acceptable material in applying IBR on an industrial scale. As a support it offers the highest rate of biological degradation, together with good mechanical stability, a low cost, and it is readily available and easily transported on account of its wide commercial use. The mathematical model was valuable in providing additional data of the biological domain, which were not determined experimentally.

0. 6

Acknowledgements

CY

0. 4 0. 2

Funding for this study was provided by the Ministerio de Ciencia y Tecnologia (Project AGL2002-09226-C03-02) and Ministerio de Medio Ambiente (Project 1.2-080/2005/3-B).

R2=0.88

0 0

25

50 t (days)

75

100

Fig. 3 – Evolution of Cy values in the three IBR systems, during the cycle of 100 days. (~) Theoretical values calculated from Eq. (2), and (’) experimental data obtained from pilot-scale reactors. The coefficient of determination (R2) describes the proportion of the total variance in the experimental data that can be explained by the model.

R E F E R E N C E S

Alonso-Sanz, R., Martin, M., 2005a. One-dimensional cellular automata with memory in cells of the most frequent recent value. Complex Syst. 15, 203–236. Alonso-Sanz, R., Martin, M., 2005b. Phase transitions in an elementary probabilistic cellular automaton with memory. Physica A 347, 383–401.

ARTICLE IN PRESS 1042

WAT E R R E S E A R C H

42 (2008) 1035– 1042

Alonso-Sanz, R., Martin, M., 2006. Elementary cellular automata with elementary memory rules in cells: the case of linear rules. J. Cell. Automata 1, 71–87. Boudreau, N.G., Daugulis, A.J., 2006. Transient performance of two-phase partitioning bioreactors treating a toluene contaminated gas stream. Biotechnol. Bioeng. 94, 448–457. Branda, S.S., Vik, A., Friedman, L., Kolter, R., 2005. Biofilms: the matrix revisited. Trends Microbiol. 13 (1), 134–138. Chen, M.J., Zhang, Z., Bott, T.R., 2005. Effects of operating conditions on the adhesive strength of Pseudomonas fluorescens biofilms in tubes. Colloids Surf. B: Biointerfaces 43, 61–71. Cogan, N.G., Cortez, R., Fauci, L., 2005. Modelling physiological resistance in bacterial biofilms. Bull. Math. Biol. 67, 831–853. Costerton, J.W., Lewandowski, Z., Caldwell, D.E., Korber, D.R., Lappin-Scott, H.M., 1995. Microbial biofilms. Annu. Rev. Microbiol. 49, 711–745. Ferrer, E., Blanco, J., Alonso-Sanz, R., Martin, M., 1996. Propachlor and alachlor degradation by immobilized and suspended Pseudomonas cells. In: Wijfels, R.H. (Ed.), Progress Biotechnology, vol. 11. Elsevier Sci, Amsterdam, pp. 762–769. Garbi, C., Casasus, L., Martinez-Alvarez, R., Robla, J.I., Martin, M., 2006. Biodegradation of oxadiazon by a soil isolated Pseudomonas strain CG5: implementation in an herbicide removal reactor and modelling. Water Res. 40, 1217–1223. Gibello, A., Garbi, C., Allende, J.L., Martin, M., 2005. Improving dioxygenase stability by gene chromosome insertion: implementation in immobilized-cell systems. Curr. Microbiol. 49, 390–395. Heipieper, H.J., Bont, J.A., 1994. Adaptation of Pseudomonas putida S12 to ethanol and toluene at the level of fatty acid composition of membranes. Appl. Environ. Microbiol. 60, 4440–4444. Jefferson, K.K., 2004. What drives bacteria to produce a biofilm? FEMS Microbiol. Lett. 236, 163–173. Laspidou, C.S., Rittmann, B.E., 2004. Modeling the development of biofilm density including active bacteria, inert biomass, and extracellular polymeric substances. Water Res. 38, 3349–3361. Legates, D.V., McCabe, G.J., 1999. Evaluating the use of ‘‘goodness of fit’’ measures in hydrologic and hydroclimatic model validation. Water Resour. Res. 35, 233–241. Martin, M., Mengs, G., Plaza, E., Garbi, C., Sanchez, M., Gibello, A., Gutierrez, F., Ferrer, E., 2000. Propachlor removal by Pseudomonas strain GCH1 in an immobilized-cell system. Appl. Environ. Microbiol. 66, 1190–1194.

Martin-Montalvo, D., Mengs, G., Ferrer, E., Allende, J.L., AlonsoSanz, R., Martin, M., 1997. Simazine degradation by immobilized and suspended soil bacterium. Int. Biodeterior. Biodegrad. 40, 93–99. Nicolella, C., van Loosdrecht, M., Heijnen, S.J., 2000. Particle-based biofilm reactor technology. TIBTECH 18, 312–320. OECD, 2000. OECD guidelines for the testing of chemicals. In: Sorption/desorption using a batch equilibrium method OECD Test Guideline, 106. OECD Publications, Paris. Pandey, G., Jain, R.K., 2002. Bacterial chemotaxis toward environmental pollutants: role in bioremediation. Appl. Environ. Microbiol. 68, 5789–5795. Panikov, N.S., 1995. Microbial Growth Kinetics. Chapman & Hall, London. Park, J.H., Zhao, X., Voice, T.C., 2001. Biodegradation of nondesorbable naphthalene in soils. Environ. Sci. Technol. 35, 2734–2740. Park, J.H., Feng, Y., Ji, P., Voice, T.C., Boyd, S.A., 2003. Assesment of bioavailability of soil-sorbed atrazine. Appl. Environ. Microbiol. 69, 3288–3298. Sanchez, M., Garbi, C., Martinez-Alvarez, R., Ortiz, L.T., Allende, J.L., Martin, M., 2005. Klebsiella planticola strain DSZ mineralizes simazine: physiological adaptations involved in the process. Appl. Microbiol. Biotechnol. 66, 60–67. Shreve, G.S., Vogel, T.M., 1993. Comparison of substrate utilization and growth kinetics between immobilized and suspended Pseudomonas cells. Biotechnol. Bioeng. 41, 370–379. Thomas, J.M., Yordy, J.R., Amador, J.A., Alexander, M., 1987. Rates of dissolution and biodegradation of water-insoluble organic compounds. Appl. Environ. Microbiol. 52, 290–296. Vandecasteele, K., Gaus, I., Debreuck, W., Valraevens, K., 2000. Identification and quantification of 77 pesticides in groundwater using solid phase coupled to liquid-liquid microextraction and reverse-phase liquid chromatography. Anal. Chem. 72, 3093–3101. Wackett, L.P., Sadowsky, M.J., Martinez, B., Shapir, N., 2002. Biodegradation of atrazine and related s-triazine compounds from enzymes to field studies. Appl. Microbiol. Biotechnol. 58, 39–45. Zacarias, G.D., Ferreira, C.P., Velasco-Hernandez, J.X., 2005. Porosity and tortuosity relations as revealed by a mathematical model of biofilm structure. J. Theor. Biol. 233, 245–251. Zheng, D., Angenent, L.T., Raskin, L., 2006. Monitoring granule formation in anaerobic upflow bioreactors using oligonucleotide hybridization probes. Biotechnol. Bioeng. 94, 458–472.