Interactions between laponite and microbial biofilms in porous media: implications for colloid transport and biofilm stability

Download Interactions between laponite and microbial biofilms in porous media: implications for colloid transport and biofilm stability

Post on 30-Oct-2016




2 download


<ul><li><p>Water Research 38 (2004) 361</p><p>ndt</p><p>wHans-Curt Flemming</p><p>Institute for Interface Biotechnology, University of Duisburg-Essen, Geibelstrasse 41, 47057 Duisburg, Germany</p><p>of colloids. Since these events are likely to occur in subsurface environments, our results suggest that colloidbiolm</p><p>interactions will have implications for colloidbound contaminant transport and the remobilisation of pathogens.</p><p>size range, i.e. approximately between 109 and 106m, mund et al., 1996; Roy and Dzomak, 1997). Under</p><p>ARTICLE IN PRESSconditions which stimulate the retention and aggrega-</p><p>tion of colloids, e.g. high ionic strength, the colloids</p><p>form an integral part of the subsurface structural matrix.</p><p>The importance of colloid-facilitated transport is em-</p><p>phasised by the fact that changes in environmental</p><p>*Corresponding author. Tel.: +49-0-203-379-3827; fax:</p><p>+49-0-203-379-1941.</p><p>E-mail address:</p><p>(M. Strathmann).</p><p>0043-1354/$ - see front matter r 2004 Elsevier Ltd. All rights reserved.doi:10.1016/j.wr 2004 Elsevier Ltd. All rights reserved.</p><p>Keywords: Biolms; Clay; Colloid; Contaminant transport; Laponite; Porous media; Subsurface</p><p>1. Introduction</p><p>The movement of particulates through the subsurface</p><p>has been studied and modelled in detail, and extensive</p><p>reviews are available (Ryan and Elimelech, 1996).</p><p>Movement of particulates which fall in the colloidal</p><p>have proven implications for contaminant transport</p><p>(McCarthy and Zachara, 1989), either when such</p><p>colloids represent contaminants (e.g. heavy metals, some</p><p>organics, and microbial pathogens) (Kersting et al.,</p><p>1999), or when relatively innocuous colloids such as clay</p><p>minerals enhance the transport of pollutants (Groli-Received 14 November 2003; received in revised form 20 April 2004; accepted 12 May 2004</p><p>Abstract</p><p>Quartz sand columns and sand-lled microscope ow cells were used to investigate the transport characteristics of the</p><p>clay colloid laponite, and a biolm-forming bacterium, Pseudomonas aeruginosa SG81. Separate experiments were</p><p>performed with each particle to determine their individual transport characteristics in clean sand columns. In a second</p><p>set of experiments, bacterial biolms were formed prior to introduction of the clay colloids. In the independent</p><p>transport experiments, bacteria and laponite each conformed to known physicochemical principles. A sodium chloride</p><p>concentration of 7 102 M caused complete retention of the laponite within the sand columns. P. aeruginosa SG81was generally less inuenced by ionic strength effects; it showed relatively low mobility at all ionic strengths tested and</p><p>some (albeit reduced) mobility when introduced to the columns in 1M NaCl, the highest concentration tested, but</p><p>nevertheless showed reproducible trends. Under conditions favourable to laponite retention and biolm stability</p><p>(7 102 MNaCl), laponite suspensions were able to remobilise a portion of the attached bacterial biomass. At lowionic strength, the prole of laponite elution was also altered in the presence of a P. aeruginosa biolm. These</p><p>observations suggest that while a reduction in ionic strength has a dominant inuence on the mobilisation of biological</p><p>and inorganic colloids, the presence of laponite and biomass can have a distinct inuence on the mobility of both typesInteractions between laponite amedia: implications for colloid</p><p>C. Felipe Leon-Morales, Andreatres.2004.05.00943626</p><p>microbial biolms in porousransport and biolm stability</p><p>P. Leis, Martin Strathmann*,</p></li><li><p>the model assumes a Fickian type of movement (rate of</p><p>ARTICLE IN PRESSC.F. Leon-Morales et al. / Water Research 38 (2004) 36143626 3615conditions (e.g. sudden reductions in ionic strength) can</p><p>cause remobilisation in a process known as colloid</p><p>generation. Colloids generated in this way can be an</p><p>important mechanism for the transport of strongly</p><p>sorbing pollutants (Grolimund et al., 1996), and other</p><p>colloids.</p><p>Little is known about the in situ mobility of</p><p>microorganisms under extreme conditions such as</p><p>pronounced changes in ionic strength resulting from</p><p>rainfall events or de-icing of streets with salt, although</p><p>laboratory simulations would suggest that these condi-</p><p>tions will have an impact on the mobility of biocolloids</p><p>(Deshpande and Shonnard, 1999). Microbial deposition</p><p>in the subsurface raises an important aspect of microbial</p><p>transport through porous materials: the ability of</p><p>microorganisms to colonise a porous medium, produce</p><p>extracellular polymeric substances (EPS) and grow as</p><p>biolms (Cunningham et al., 1991). Biolm formation</p><p>and release can be considered analogous to the process</p><p>of colloid generation or release described earlier. As a</p><p>consequence, these microbial formations are expected to</p><p>inuence colloid transport in several different ways.</p><p>These inuences can be direct as is the case of particle</p><p>retention within biolm compartments (Okabe et al.,</p><p>1997) or indirect as is the case for changes in</p><p>hydrodynamic conditions caused by extensive biolm</p><p>growth. Changes in hydrodynamic conditions within the</p><p>porous media caused by thick biolm formation have</p><p>been recognised (Sharp et al., 1999). Thick, conuent</p><p>biolms are common in engineered systems or in</p><p>nutrient rich environments (i.e. heavily contaminated</p><p>environments) whereas thin and patchy biolms are</p><p>commonly found in pristine subsurface environments.</p><p>The transport of microorganisms through porous</p><p>media has received considerable attention (Camesano</p><p>et al., 1999; Deshpande and Shonnard, 1999; Jewett</p><p>et al., 1999; Smets et al., 1999), although the focus of</p><p>these studies has been varied. These include studies on</p><p>preventing bacterial migration through aquifer systems</p><p>(e.g. ltration of pathogens) (El-Masry et al., 1995;</p><p>Tufenkji et al., 2002) and studies centred on enhancing</p><p>microbial transport to a contaminated site for bioreme-</p><p>diation (Li and Logan, 1999). Sand columns are</p><p>commonly used to obtain colloid transport parameters,</p><p>such as collision efciencies. As will be expanded upon</p><p>in the theoretical background section, deposition para-</p><p>meters in sand columns rely on the quantication of</p><p>inuent/efuent concentrations. The quantication of</p><p>retained colloids or biocolloids typically involves</p><p>destructive methods such as extrusion of the porous</p><p>medium and slicing to determine cell numbers by</p><p>labelling methods like microbe and radiolabel kinesis</p><p>(MARK) (Gross et al., 1995; Li and Logan, 1999). The</p><p>confocal microscopy approach reported in the present</p><p>study takes advantage of the ability to acquire informa-</p><p>tion on the retention of uorescently-labelled bacteriasolute spread grows linearly with time). As will be seen</p><p>later, these conditions are met in our experimental</p><p>setting.</p><p>Column parameters such as pore volumes, tracer</p><p>dispersivities and average travel times can be obtained</p><p>from experimental tracer breakthrough curves (BTC).</p><p>One can either t experimental data to an analytical</p><p>solution of the transport equation (1) using a standard</p><p>nonlinear least-squares procedure or they can be</p><p>obtained by moment analysis of the BTC (Dyson,</p><p>1990). In sand columns with high Peclet numbers both</p><p>procedures yield similar results (Grolimund et al., 1998).</p><p>For either method it is necessary to obtain a normal-</p><p>isation constant, n0; which represents the total amountusing online, non-destructive technology which can be</p><p>independent of inuent/efuent bacterial concentrations</p><p>and which at the present state of development can serve</p><p>as a semi-quantitative tool for the determination of</p><p>colloidal particle deposition. The main aim of the</p><p>present report is to highlight the differences between</p><p>the transport of the synthetic clay mineral laponite, and</p><p>a microorganism, Pseudomonas aeruginosa SG81, within</p><p>sand-packed laboratory columns. Both particles were</p><p>studied separately, and the transport of laponite and its</p><p>effects on biolm stability were studied subsequent to</p><p>the establishment of microbial biolms within the</p><p>porous medium.</p><p>2. Theoretical background</p><p>According to classical groundwater hydrology, the</p><p>transport of solutes through porous materials can be</p><p>described by accounting for advection (or mechanical</p><p>movement) and hydrodynamic dispersion (as a sum of</p><p>molecular diffusion and mechanical dispersion). The</p><p>equation describing these relationships is known as the</p><p>advectiondispersion equation (ADE). In the case of the</p><p>transport of colloidal particles, several researchers</p><p>(Kretzschmar et al., 1997; Grolimund et al., 1998,</p><p>2001) have commonly used a modied version of the</p><p>ADE that includes a term accounting for particle</p><p>deposition (ltration). The concentration of suspended</p><p>particles at a determined column depth and time, cx; tcan then be written as</p><p>@C</p><p>@t</p><p>@2C</p><p>D@x2 v</p><p>@C</p><p>@x kc; 1</p><p>where v is interstitial colloid particle velocity, D is the</p><p>hydrodynamic dispersion coefcient and k is the particle</p><p>deposition rate coefcient. For this equation to be valid,</p><p>the column should be initially free of colloidal particles</p><p>(clean bed), colloidal release should be minimal com-</p><p>pared with deposition, and colloid concentration should</p><p>be small (no ripening or blocking effects). Additionally,</p></li><li><p>where kdf is the deposition rate constant for fast</p><p>ARTICLE IN PRESSC.F. Leon-Morales et al. / Water Research 38 (2004) 361436263616deposition conditions.</p><p>3. Materials and methods</p><p>3.1. Inorganic colloids</p><p>Laponite RD, a synthetic hectorite clay was chosen as</p><p>a model colloid due to its high purity, monodispersity,</p><p>and homogeneous dispersion properties in deionised</p><p>water. Laponite consists of 30 nm diameter plate-like</p><p>particles with a thickness of 12 nm when hydrated</p><p>(Nicolai and Cocard, 2001). Laponite suspensions were</p><p>used at the following concentrations: 2000mgL1</p><p>suspended in 0M and 7 102M NaCl; 200mgL1</p><p>suspended in 6.25 104, 1.25 103, 2.5 103,5 103, 1 102 and 3.5 102M NaCl. Otherlaponite concentrations and ionic strengths of the</p><p>suspending media were assessed (data not shown),</p><p>which resulted in different laponite aggregation states.</p><p>Aggregation was generally favoured by high ionic</p><p>strengths but was dependent on laponite concentration.</p><p>For instance, 2000mgL1 laponite at 7 102M NaClwas highly aggregated with aggregate sizes of several</p><p>micrometres, but 20mgL1 laponite at 7 102M wasless aggregated and contained smaller-sized aggregates</p><p>(as evidenced by epiuorescence microscopy of rhoda-</p><p>mine-stained laponite preparations). Additionally, the</p><p>laponite dispersions were stained with a cationic dye,of particles injected divided by the total volumetric ux.</p><p>This constant can be obtained experimentally from a</p><p>bypass experiment in which the column is replaced by</p><p>standard tubing. We used the method of moments to</p><p>obtain column parameters and the other colloidal</p><p>transport characteristics. All integration routines were</p><p>made either in Microsoft Excel or using open source</p><p>plotting programs.</p><p>Deposition rate constant kd for colloid transport</p><p>experiments can then be obtained by comparing the</p><p>integrated amount of particles in the efuent C with the</p><p>total injected amount C0 (Grolimund et al., 2001):</p><p>kd 1</p><p>tpln</p><p>C</p><p>C0</p><p> ; 2</p><p>where tp is the average travel time of particles through</p><p>the sand column. At sufciently low salt concentrations</p><p>the deposition rate constant is proportional to the salt</p><p>concentration but at high salt levels this constant</p><p>becomes independent of the salt concentration. Condi-</p><p>tions at high salt levels are known as fast (favourable)</p><p>deposition conditions. The collision efciency, a; canthen be written as</p><p>a kd</p><p>kdf; 3rhodamine 6G (Fluka, Switzerland) at concentrations</p><p>ranging from 5 107 M to 5 106M depending onlaponite concentration, according to the protocol</p><p>described in Tapia Estevez et al. (1993). The adsorption</p><p>of the rhodamine 6G monomer onto laponite was</p><p>conrmed by a shift in the absorption maximum of the</p><p>dye from 526 nm to 535 nm when adsorbed to laponite,</p><p>as reported in detailed spectroscopic studies of rhoda-</p><p>mine 6G adsorption onto other clay minerals (Lopez</p><p>Arbeloa et al., 1996). When dispersed in deionised water,</p><p>laponite dispersions were clear and apparently non-</p><p>aggregated after approximately 40min of stirring. The</p><p>pH of the dispersions was dependent on the laponite</p><p>concentrations and ranged from pH 10 for the highest</p><p>laponite concentration used in this study (2000mgL1)</p><p>to pH 6.6 for less concentrated dispersions (20mgL1).</p><p>The pH of the working laponite suspensions was</p><p>adjusted when necessary to 67 where indicated by</p><p>using 0.2M HCl or 0.2M NaOH. Absorbance measure-</p><p>ments to detect laponite were done at 535 nm in all sand</p><p>column experiments and uorescence detection of the</p><p>laponiterhodamine complex was used for the same</p><p>purpose in the sand-packed ow cell experiments.</p><p>3.2. Biocolloid suspensions</p><p>A well characterised biolm-forming, EPS-producing</p><p>microorganism, P. aeruginosa SG81, was chosen as a</p><p>model biocolloid. P. aeruginosa SG81 is a mucoid,</p><p>Gram-negative, motile bacterium which was originally</p><p>isolated from a technical water distribution system</p><p>(Grobe et al., 1995). The organism was grown in batch</p><p>culture on 15strength tryptic soy broth (TSB) (Merck,</p><p>Darmstadt). According to conventional growth curves</p><p>(data not shown) the organism was harvested during or</p><p>at the end of the logarithmic phase for use in the biolm</p><p>formation or transport experiments, respectively. The</p><p>harvesting at the end of the logarithmic phase was done</p><p>to minimise the potential for increase in cell numbers</p><p>during the transport experiments. For the transport</p><p>experiments, the organisms were concentrated in reac-</p><p>tion tubes by centrifugation and then washed with</p><p>6 104M NaCl to remove traces of nutrient mediumby centrifuging at least 2 times at 3000 g for 5min at5C. The cells were washed with low NaCl concentra-</p><p>tions i.e. 6 104M NaCl, in order to facilitate cell re-suspension by vortexing. Salt concentration in the cell</p><p>suspensions was nally adjusted to 1 103, 1 102,1.4 101 and 1M NaCl. Cell suspensions prepared inthis way typically had cell concentrations of</p><p>1.7 109 cellsmL1 , as determined by total cell counts(TCC) using a standard Thoma cell counting chamber in</p><p>combination with phase contrast microscopy. Calibra-</p><p>tion curves were constructed to determine numbers of</p><p>bacteria in suspension at a given absorbance by plotting</p><p>TCC against absorbance. On some occasions, colony</p></li><li><p>ARTICLE IN PRESSC.F. Leon-Morales et al. / Water Research 38 (2004) 36143626 3617forming units (CFU) were determined to evaluate the</p><p>effectiveness of the acid/base treatment for the sand</p><p>columns.</p><p>For ow cell experiments, cells were stained with the</p><p>nucleic acid-specic uorochrome SYTO 9 (Molecular</p><p>Probes). The staining procedure for bacterial suspen-</p><p>sions varied depending on the type of experiment. In one</p><p>case, 1.5 mL of the SYTO 9 staining solution was addedper millilitre of bacterial suspension. In the other case in</p><p>which bacteria were growing inside the sand-packed</p><p>ow cells, two to three pore volumes of the 1.5 mLmL1</p><p>SYTO 9 solution were pumped through the biolm-</p><p>growing ow cell using a peristaltic pump.</p><p>Detection of bacteria was done either by measuring</p><p>absorbance of unstained suspensions (sand column</p><p>experiments) at 2...</p></li></ul>


View more >