interactions between laponite and microbial biofilms in porous media: implications for colloid...

Water Research 38 (2004) 3614–3626

ARTICLE IN PRESS

*Correspond

+49-0-203-379

E-mail addr

(M. Strathman

0043-1354/$ - se

doi:10.1016/j.w

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

C. Felipe Leon-Morales, Andrew P. Leis, Martin Strathmann*,Hans-Curt Flemming

Institute for Interface Biotechnology, University of Duisburg-Essen, Geibelstrasse 41, 47057 Duisburg, Germany

Received 14 November 2003; received in revised form 20 April 2004; accepted 12 May 2004

Abstract

Quartz sand columns and sand-filled microscope flow cells were used to investigate the transport characteristics of the

clay colloid laponite, and a biofilm-forming bacterium, Pseudomonas aeruginosa SG81. Separate experiments were

performed with each particle to determine their individual transport characteristics in clean sand columns. In a second

set of experiments, bacterial biofilms were formed prior to introduction of the clay colloids. In the independent

transport experiments, bacteria and laponite each conformed to known physicochemical principles. A sodium chloride

concentration of 7� 10�2 M caused complete retention of the laponite within the sand columns. P. aeruginosa SG81

was generally less influenced by ionic strength effects; it showed relatively low mobility at all ionic strengths tested and

some (albeit reduced) mobility when introduced to the columns in 1 M NaCl, the highest concentration tested, but

nevertheless showed reproducible trends. Under conditions favourable to laponite retention and biofilm stability

(7� 10�2 M NaCl), laponite suspensions were able to remobilise a portion of the attached bacterial biomass. At low

ionic strength, the profile of laponite elution was also altered in the presence of a P. aeruginosa biofilm. These

observations suggest that while a reduction in ionic strength has a dominant influence on the mobilisation of biological

and inorganic colloids, the presence of laponite and biomass can have a distinct influence on the mobility of both types

of colloids. Since these events are likely to occur in subsurface environments, our results suggest that colloid–biofilm

interactions will have implications for colloid–bound contaminant transport and the remobilisation of pathogens.

r 2004 Elsevier Ltd. All rights reserved.

Keywords: Biofilms; Clay; Colloid; Contaminant transport; Laponite; Porous media; Subsurface

1. Introduction

The movement of particulates through the subsurface

has been studied and modelled in detail, and extensive

reviews are available (Ryan and Elimelech, 1996).

Movement of particulates which fall in the colloidal

size range, i.e. approximately between 10�9 and 10�6 m,

ing author. Tel.: +49-0-203-379-3827; fax:

-1941.

ess: [email protected]

n).

e front matter r 2004 Elsevier Ltd. All rights reserve

atres.2004.05.009

have proven implications for contaminant transport

(McCarthy and Zachara, 1989), either when such

colloids represent contaminants (e.g. heavy metals, some

organics, and microbial pathogens) (Kersting et al.,

1999), or when relatively innocuous colloids such as clay

minerals enhance the transport of pollutants (Groli-

mund et al., 1996; Roy and Dzomak, 1997). Under

conditions which stimulate the retention and aggrega-

tion of colloids, e.g. high ionic strength, the colloids

form an integral part of the subsurface structural matrix.

The importance of colloid-facilitated transport is em-

phasised by the fact that changes in environmental

d.

ARTICLE IN PRESSC.F. Leon-Morales et al. / Water Research 38 (2004) 3614–3626 3615

conditions (e.g. sudden reductions in ionic strength) can

cause remobilisation in a process known as colloid

generation. Colloids ‘‘generated’’ in this way can be an

important mechanism for the transport of strongly

sorbing pollutants (Grolimund et al., 1996), and other

colloids.

Little is known about the in situ mobility of

microorganisms under extreme conditions such as

pronounced changes in ionic strength resulting from

rainfall events or de-icing of streets with salt, although

laboratory simulations would suggest that these condi-

tions will have an impact on the mobility of biocolloids

(Deshpande and Shonnard, 1999). Microbial deposition

in the subsurface raises an important aspect of microbial

transport through porous materials: the ability of

microorganisms to colonise a porous medium, produce

extracellular polymeric substances (EPS) and grow as

biofilms (Cunningham et al., 1991). Biofilm formation

and release can be considered analogous to the process

of colloid generation or release described earlier. As a

consequence, these microbial formations are expected to

influence colloid transport in several different ways.

These influences can be direct as is the case of particle

retention within biofilm compartments (Okabe et al.,

1997) or indirect as is the case for changes in

hydrodynamic conditions caused by extensive biofilm

growth. Changes in hydrodynamic conditions within the

porous media caused by thick biofilm formation have

been recognised (Sharp et al., 1999). Thick, confluent

biofilms are common in engineered systems or in

nutrient rich environments (i.e. heavily contaminated

environments) whereas thin and patchy biofilms are

commonly found in pristine subsurface environments.

The transport of microorganisms through porous

media has received considerable attention (Camesano

et al., 1999; Deshpande and Shonnard, 1999; Jewett

et al., 1999; Smets et al., 1999), although the focus of

these studies has been varied. These include studies on

preventing bacterial migration through aquifer systems

(e.g. filtration of pathogens) (El-Masry et al., 1995;

Tufenkji et al., 2002) and studies centred on enhancing

microbial transport to a contaminated site for bioreme-

diation (Li and Logan, 1999). Sand columns are

commonly used to obtain colloid transport parameters,

such as collision efficiencies. As will be expanded upon

in the theoretical background section, deposition para-

meters in sand columns rely on the quantification of

influent/effluent concentrations. The quantification of

retained colloids or biocolloids typically involves

destructive methods such as extrusion of the porous

medium and slicing to determine cell numbers by

labelling methods like microbe and radiolabel kinesis

(MARK) (Gross et al., 1995; Li and Logan, 1999). The

confocal microscopy approach reported in the present

study takes advantage of the ability to acquire informa-

tion on the retention of fluorescently-labelled bacteria

using online, non-destructive technology which can be

independent of influent/effluent bacterial concentrations

and which at the present state of development can serve

as a semi-quantitative tool for the determination of

colloidal particle deposition. The main aim of the

present report is to highlight the differences between

the transport of the synthetic clay mineral laponite, and

a microorganism, Pseudomonas aeruginosa SG81, within

sand-packed laboratory columns. Both particles were

studied separately, and the transport of laponite and its

effects on biofilm stability were studied subsequent to

the establishment of microbial biofilms within the

porous medium.

2. Theoretical background

According to classical groundwater hydrology, the

transport of solutes through porous materials can be

described by accounting for advection (or mechanical

movement) and hydrodynamic dispersion (as a sum of

molecular diffusion and mechanical dispersion). The

equation describing these relationships is known as the

advection–dispersion equation (ADE). In the case of the

transport of colloidal particles, several researchers

(Kretzschmar et al., 1997; Grolimund et al., 1998,

2001) have commonly used a modified version of the

ADE that includes a term accounting for particle

deposition (filtration). The concentration of suspended

particles at a determined column depth and time, cðx; tÞcan then be written as

@C

@t¼

@2C

D@x2� v

@C

@x� kc; ð1Þ

where v is interstitial colloid particle velocity, D is the

hydrodynamic dispersion coefficient and k is the particle

deposition rate coefficient. For this equation to be valid,

the column should be initially free of colloidal particles

(clean bed), colloidal release should be minimal com-

pared with deposition, and colloid concentration should

be small (no ripening or blocking effects). Additionally,

the model assumes a Fickian type of movement (rate of

solute spread grows linearly with time). As will be seen

later, these conditions are met in our experimental

setting.

Column parameters such as pore volumes, tracer

dispersivities and average travel times can be obtained

from experimental tracer breakthrough curves (BTC).

One can either fit experimental data to an analytical

solution of the transport equation (1) using a standard

nonlinear least-squares procedure or they can be

obtained by moment analysis of the BTC (Dyson,

1990). In sand columns with high Peclet numbers both

procedures yield similar results (Grolimund et al., 1998).

For either method it is necessary to obtain a normal-

isation constant, n0; which represents the total amount

ARTICLE IN PRESSC.F. Leon-Morales et al. / Water Research 38 (2004) 3614–36263616

of particles injected divided by the total volumetric flux.

This constant can be obtained experimentally from a

bypass experiment in which the column is replaced by

standard tubing. We used the method of moments to

obtain column parameters and the other colloidal

transport characteristics. All integration routines were

made either in Microsoft Excel or using open source

plotting programs.

Deposition rate constant kd for colloid transport

experiments can then be obtained by comparing the

integrated amount of particles in the effluent C with the

total injected amount C0 (Grolimund et al., 2001):

kd ¼1

tpln

C

C0

� �; ð2Þ

where tp is the average travel time of particles through

the sand column. At sufficiently low salt concentrations

the deposition rate constant is proportional to the salt

concentration but at high salt levels this constant

becomes independent of the salt concentration. Condi-

tions at high salt levels are known as fast (favourable)

deposition conditions. The collision efficiency, a; can

then be written as

a ¼kd

kdf; ð3Þ

where kdf is the deposition rate constant for fast

deposition conditions.

3. Materials and methods

3.1. Inorganic colloids

Laponite RD, a synthetic hectorite clay was chosen as

a model colloid due to its high purity, monodispersity,

and homogeneous dispersion properties in deionised

water. Laponite consists of 30 nm diameter plate-like

particles with a thickness of 1–2 nm when hydrated

(Nicolai and Cocard, 2001). Laponite suspensions were

used at the following concentrations: 2000 mg L�1

suspended in 0 M and 7� 10�2 M NaCl; 200 mg L�1

suspended in 6.25� 10�4, 1.25� 10�3, 2.5� 10�3,

5� 10�3, 1� 10�2 and 3.5� 10�2 M NaCl. Other

laponite concentrations and ionic strengths of the

suspending media were assessed (data not shown),

which resulted in different laponite aggregation states.

Aggregation was generally favoured by high ionic

strengths but was dependent on laponite concentration.

For instance, 2000 mg L�1 laponite at 7� 10�2 M NaCl

was highly aggregated with aggregate sizes of several

micrometres, but 20 mg L�1 laponite at 7� 10�2 M was

less aggregated and contained smaller-sized aggregates

(as evidenced by epifluorescence microscopy of rhoda-

mine-stained laponite preparations). Additionally, the

laponite dispersions were stained with a cationic dye,

rhodamine 6G (Fluka, Switzerland) at concentrations

ranging from 5� 10�7 M to 5� 10�6 M depending on

laponite concentration, according to the protocol

described in Tapia Estevez et al. (1993). The adsorption

of the rhodamine 6G monomer onto laponite was

confirmed by a shift in the absorption maximum of the

dye from 526 nm to 535 nm when adsorbed to laponite,

as reported in detailed spectroscopic studies of rhoda-

mine 6G adsorption onto other clay minerals (Lopez

Arbeloa et al., 1996). When dispersed in deionised water,

laponite dispersions were clear and apparently non-

aggregated after approximately 40 min of stirring. The

pH of the dispersions was dependent on the laponite

concentrations and ranged from pH 10 for the highest

laponite concentration used in this study (2000 mg L�1)

to pH 6.6 for less concentrated dispersions (20 mg L�1).

The pH of the working laponite suspensions was

adjusted when necessary to 6–7 where indicated by

using 0.2 M HCl or 0.2 M NaOH. Absorbance measure-

ments to detect laponite were done at 535 nm in all sand

column experiments and fluorescence detection of the

laponite–rhodamine complex was used for the same

purpose in the sand-packed flow cell experiments.

3.2. Biocolloid suspensions

A well characterised biofilm-forming, EPS-producing

microorganism, P. aeruginosa SG81, was chosen as a

model biocolloid. P. aeruginosa SG81 is a mucoid,

Gram-negative, motile bacterium which was originally

isolated from a technical water distribution system

(Grobe et al., 1995). The organism was grown in batch

culture on 15

strength tryptic soy broth (TSB) (Merck,

Darmstadt). According to conventional growth curves

(data not shown) the organism was harvested during or

at the end of the logarithmic phase for use in the biofilm

formation or transport experiments, respectively. The

harvesting at the end of the logarithmic phase was done

to minimise the potential for increase in cell numbers

during the transport experiments. For the transport

experiments, the organisms were concentrated in reac-

tion tubes by centrifugation and then washed with

6� 10�4 M NaCl to remove traces of nutrient medium

by centrifuging at least 2 times at 3000� g for 5 min at

5�C. The cells were washed with low NaCl concentra-

tions i.e. 6� 10�4 M NaCl, in order to facilitate cell re-

suspension by vortexing. Salt concentration in the cell

suspensions was finally adjusted to 1� 10�3, 1� 10�2,

1.4� 10�1 and 1 M NaCl. Cell suspensions prepared in

this way typically had cell concentrations of

1.7� 109 cells mL�1 , as determined by total cell counts

(TCC) using a standard Thoma cell counting chamber in

combination with phase contrast microscopy. Calibra-

tion curves were constructed to determine numbers of

bacteria in suspension at a given absorbance by plotting

TCC against absorbance. On some occasions, colony


forming units (CFU) were determined to evaluate the

effectiveness of the acid/base treatment for the sand

columns.

For flow cell experiments, cells were stained with the

nucleic acid-specific fluorochrome SYTO 9 (Molecular

Probes). The staining procedure for bacterial suspen-

sions varied depending on the type of experiment. In one

case, 1.5 mL of the SYTO 9 staining solution was added

per millilitre of bacterial suspension. In the other case in

which bacteria were growing inside the sand-packed

flow cells, two to three pore volumes of the 1.5 mL mL�1

SYTO 9 solution were pumped through the biofilm-

growing flow cell using a peristaltic pump.

Detection of bacteria was done either by measuring

absorbance of unstained suspensions (sand column

experiments) at 240 nm, or by measuring the fluores-

cence emitted by the SYTO 9 stain when combined with

bacterial DNA using confocal scanning laser micro-

scopy (CLSM, as described in Section 3.4). Absorbance

was measured with a UV/Visible spectrophotometer

(Cary 50, Varian Inc.). Other measurements i.e. total cell

counts were performed off-line by collecting samples

using a fraction collector.

3.3. Column experiments

Glass chromatography columns (Amersham Bios-

ciences, Freiburg, Germany) with a diameter of 1 cm

and a length of 10 cm were used. Tygon LFL tubing

(Novodirect, Kehl, Germany) was used to connect the

columns to a peristaltic pump (Ismatec SA, Switzer-

land). The columns were packed with quartz sand type

F34 with an average size of 0.2 mm. The sand was first

washed by means of an acid/base treatment which

consisted of the introduction of 10 pore volumes of

0.2 M NaOH and 10 pore volumes of 0.2 M HCl, each

followed by the introduction of sterile deionised water.

This treatment served to standardise the chemical

homogeneity of the columns, remove mineral and

organic colloidal residues, and minimise contamination

by microorganisms. Columns were not autoclaved

because this treatment alters the water saturation

conditions under which the columns were packed.

Regular bacteriological plate counts and microscopic

inspections of column effluent did not show any

evidence for microbial contamination.

Each column was packed with 11 g of sand (dry

weight), minimising air introduction as described in

Deshpande and Shonnard (1999). General column

characteristics were extracted from the BTC of nigrosine

tracers as illustrated in the theoretical background

section. Peclet numbers (Pe) were calculated from the

formula, Pe ¼ VxL=D; where Vx represents interstitial

particle velocity, L; length of the column and D,

dispersion. Flow speed was confirmed by measuring

precise effluent volumes over time. The column influent

was always introduced in an up-stream mode in an effort

to minimise non-fickian types of fluid movement. UV/

Visible spectrophotometry was performed online using

quartz flow-through cuvettes (Starna, Germany).

In one type of experiment, pulses of rhodamine-

stained laponite (600mL, 200 mg L�1 at ionic strength as

described above) or P. aeruginosa SG81 (concentrations

described above) were introduced to clean columns.

Prior to pulse injection, the pH of the column effluent

was stable and the appropriate NaCl solution was used

to pre-condition the column. After pulse injection, the

same NaCl solution was re-introduced. In the cases

where the injected colloids were retained inside the

columns, their re-mobilisation after depletion of ionic

strength was also recorded using spectrophotometry.

BTCs were baseline corrected and the first three

moments (i.e. peak area, peak centre of gravity and

peak variance) were used to obtain transport parameters

using equations as given above.

In biofilm formation experiments, sterile columns

were pre-conditioned with 15

strength TSB before the

introduction of approximately one pore volume of

bacterial inoculum. The cells were not washed for these

experiments and were introduced suspended in the

original growth medium (15

TSB). The inoculation was

performed when the cells were in the logarithmic phase

(typically 5 h). The inoculated columns were left

stagnant for 2–3 days, after which 15

TSB was introduced

at a flow rate of 17 mL h�1. The flow was introduced for

up to 10 h a day for 1–2 weeks. Under these conditions,

the biofilm-containing columns were subjected to

sudden changes in ionic strength (depletion) or to the

introduction of aggregated or un-aggregated (low/high

ionic strength) laponite suspensions. Prior to the

application of these disturbances, the column influent

was changed from 15

TSB to a higher ionic strength NaCl

solution, i.e. 7� 10�2 M NaCl. After several pore

volumes, when absorbance of the column effluent and

bacterial cell numbers were constant (i.e. at

1� 107 cells mL�1), the disturbances were introduced.

In the first case, the NaCl solution was changed to

deionised water. In the second case, the NaCl solution

was substituted with laponite dispersed in 6.2� 10�4 M

NaCl (low ionic strength) or laponite dispersed in

7� 10�2 M NaCl (high ionic strength). After elution of

laponite was complete, the influent continued with the

same ionic strength as that of the laponite suspension

injected previously. In the same way, absorbance and

total cell counts were monitored ca. every 8 min, and pH

was monitored at regular intervals.

3.4. Flow cell experiments

Stainless steel flow cells normally used for growing

biofilms on the surfaces of glass slides were adapted for

porous medium experiments by packing with quartz


sand F34. The flow cells consisted of two independent

parallel chambers with individual channel dimensions

of 8 mm� 3 mm� 54 mm corresponding to a volume

of 1.32 mL. Microscope glass cover slips (dimensions

60 mm� 24 mm� 0.17 mm) were fixed onto both sides

of the flow cell, using additive-free silicone rubber

(Fig. 1). The basic parameters determined as for the

chromatography columns were also determined for the

flow cells (see Table 1). The same tubing, used for the

sand columns, was used to connect the flow cells to the

peristaltic pump, to influent solutions, and to a waste

reservoir. Packed flow cells received the same cleaning

treatment as the chromatography columns (commensu-

rate with size). The inoculation of the flow cells was

performed with sterile syringes through ports located

close to the flow cell inlet.

For transport experiments in flow cells, 100 mL of

bacterial suspensions stained with SYTO 9 (lex;max:

480 nm, lem;max 500 nm) or 100 mL of rhodamine-stained

UV/VIS

purge

fractioncollector

pH

Off-line:total cell counts(TCC) using phase contrast microscopy

influent

peristalticpump

effluent

Fig. 1. General exper

Table 1

Sand column and sand-packed flow cell characteristics. Results were

Basic parameters From mo

Sand columns Flow cells

Flow rate (mL min�1) 0.8 0.3 Velocity (

Flow rate (cm min�1) 2.17 2.66 Travel tim

Bulk density (g cm�3) 1.41 1.41 Pe

Porosity (%) 47 47

Pore volume (mL) 3.69 0.62

laponite suspensions (lex;max: 500 nm, lem;max 551 nm)

were introduced at a constant flow of 0.3 mL min�1

into flow cells which were pre-conditioned with the

appropriate NaCl solution. NaCl concentrations for

bacteria or laponite were as indicated above. Flow cells

were placed horizontally on the microscope stage.

Observation sites were located at inlet, middle and

outlet positions. Breakthrough data were recorded

at outlet positions (still within the porous medium).

BTCs were obtained by image analysis of CLSM cross-

sectional images recorded at 10 s intervals. Stained

areas were equated to the entire observation area

(230.30 mm� 230.30 mm) minus grain areas as measured

from a rhodamine-stained laponite suspension run

previously through the same observational spot at

extremely low ionic strength (0 M NaCl). Rhodamine–

laponite suspensions at very low ionic strength were

shown to be transported very efficiently (almost no

retention at inlet, middle and outlet positions) through

inoculation port

CLSM

sand-packedflow cell

imental setting.

obtained from triplicate tracer experiments

ment analysis of tracer BTCs

Sand columns Flow cells

m s�1) 5.7� 10�4 (74.2� 10�5) 3.2� 10�4 (72� 10�5)

e (s) 175 (78.6) 171.8 (711.6)

1040 (734) 1612 (7177)

ARTICLE IN PRESS

-3 -2 -1 0

-2

-1

0

Log

colli

sion

effi

cien

cy

Log ionic strength [M]

Laponite

Bacteria

-3 -2 -1 0

-2

-1

0Lo

g co

llisi

on e

ffici

ency

Log ionic strength [M]

Bacteria

Laponite

(A)

(B)

Fig. 2. (A) Collision efficiencies for P. aeruginosa SG81 and

laponite RD obtained from sand column experiments as a

function of ionic strength (sodium and chloride ions), (B)

Collision efficiencies obtained from sand-packed flow cell

experiments. All data shown represent at least duplicate

experiments.

C.F. Leon-Morales et al. / Water Research 38 (2004) 3614–3626 3619

the sand-packed flow cells. Prior to the experiments the

intensities of both the low ionic strength ‘‘tracer’’

rhodamine–laponite suspension and the colloidal sus-

pension to be assessed were calibrated to match very

closely. This was achieved by using microscope settings

(excitation wavelength of 488 nm and a 505 nm long

pass-filter) that allowed for the simultaneous detection

of both types of dyes at a single detection channel and

by modifying the microscope detector gain.

Both in the case of laponite and bacterial transport

experiments the maximum covered area achieved by the

rhodamine–laponite suspension was taken as 100%

coverage (C0) and the elution of subsequent colloidal

dispersions was taken as a fraction of this measure (C).

The method also allowed for the quantification of

retained cells at the different observation sites in a non-

destructive manner. Transport parameters were ob-

tained by moment analysis of the BTCs generated in

this way.

For biofilm growth experiments, approximately one

pore volume of P. aeruginosa SG81 suspension as

described earlier was introduced to each chamber. The

inoculated flow cell was left stagnant for one day. A flow

of 0.1 mL min�1 was introduced initially and increased

incrementally with several hours or days in between, up

to 0.5 mL min�1; this allowed the biofilms to adapt

gradually to the increased flow without considerable loss

of biomass. For the execution of the experiments, except

where an increase of flow was monitored as a

disturbance, the flow rate was kept constant. Before,

during and after the SYTO 9 staining procedure, a field

of view near the flow cell outlet was monitored

constantly by CLSM. An increase of cell detachment

due to this treatment could not be detected. This

indicated that biofilm instability was not caused by the

SYTO 9 staining operation at least for the duration of

the experiments. Following this period, the disturbance

(changes in ionic strength, flow, or laponite introduc-

tion) was applied and the event was recorded in real time

using CLSM.

4. Results

4.1. Mobility of P. aeruginosa SG81 and laponite in sand

columns and flow cells

Collision efficiencies for bacteria and laponite were

calculated using deposition rate constants extracted

from the BTCs obtained from both the CLSM method

and the sand columns method (Fig. 2). Relevant column

characteristics are summarised in Table 1. In the absence

of electrolyte or at low concentrations i.e. 6� 10�4 M

NaCl, laponite was highly mobile (collision efficiencies

of 0.0370.01). As electrolyte concentration increased,

the transport of laponite through sand columns

and sand-packed flow cells was greatly affected

(Figs. 2 and 3A). At high laponite (2000 mg L�1) and

electrolyte concentrations (7� 10�2 M NaCl) the clay

was highly aggregated and gelation occurred on the

surfaces of sand grains (Fig. 3C). At low laponite

(20 mg L�1) concentrations but high electrolyte concen-

trations laponite was still retained inside the porous

material.

Bacteria were in general less mobile with lowest

collision efficiency of 0.07 (70.03) at 1� 10�3 M NaCl

but still showed mobility in sand columns at concentra-

tions up to 1 M NaCl (Fig. 3B). In flow cells (Fig. 2B),

despite a similar trend, bacteria were less mobile. Most

of the retained bacteria could be found in the portion of

the sand-packed flow cell closest to the inlet for all three

ARTICLE IN PRESS

0 21 3 40

0.2

0.4

0.6

0.8

1.0 Laponite + 6.2 x 10-4 M NaCl Laponite + 2.5 x 10-3 M NaCl Laponite + 1.0 x 10-2 M NaCl

C/C

0

Pore volumes

Pore volumes

0 21 3

0.1

0.2

0.3

Bacteria + 1M NaCl

Bacteria + 1.4 x 10-1 M NaCl

0

C/C

0

(A)

(B)

(C)

Fig. 3. (A) Laponite and (B) bacterial transport profiles through non-inoculated sand columns as a function of salt concentration, (C)

clean F34 quartz sand (left) and sand coated with laponite (2000 mg L�1). The laponite was retained as a gel-like substance at elevated

ionic strength (7� 10�2 M).

C.F. Leon-Morales et al. / Water Research 38 (2004) 3614–36263620

ionic strengths, again shown by confocal microscopy;

this was especially apparent in the case of the 1 M NaCl

solution (data not shown).

Following deposition within the porous medium, both

the bacterium and the clay colloid could be partially re-

mobilised by depleting the ionic strength (Figs. 4A, 4B

and 5) or by sudden increases in flow rate. After

attachment to the substratum, the cells grew to form

patchy, localised biofilms (Fig. 6). The response of the P.

aeruginosa SG81 biofilms to physicochemical alterations

such as an incremental increase in flow rate or a

lowering of ionic strength was characterised by increased

cell detachment which could be visualised using confocal

microscopy (Fig. 5).

4.2. Interactions between laponite and P. aeruginosa

SG81: enhanced colloidal mobility

A pulse of laponite through biofilm-colonised sand

columns at low ionic strength (6.2� 10�4 M) followed

by a corresponding ionic strength reduction of the

influent resulted in an altered pattern of elution (Fig. 7B)

as compared with the elution through sterile columns

(Fig. 3A). Elution of laponite through biofilm-colonised

sand columns at an ionic strength which caused its

retention in sterile columns resulted in the enhanced

detachment of bacterial cells (Fig. 8). This detachment

event was proportional to laponite concentration and

therefore to laponite aggregation state. This effect was

ARTICLE IN PRESS

0 2 3 40

1x107

2x107

3x107

4x107

5x107

6x107

7x107

8x107

9x107

1x108

cells

mL-1

Pore volumes

TCC

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0

A24

0

Absorbance units

0 4 8 10 12

0

0.02

0.04

0.06

0.08

0.10

0.12

0.14

A24

0

Pore volumes

1

2 6

(A)

(B)

Fig. 4. (A) Release of biofilm cells after a sudden change in ionic strength from 7� 10�2 M NaCl to deionised water, (B) re-

mobilisation of the laponite trapped in the sand columns after a change in ionic strength similar to ‘‘A’’. The arrow indicates the time

point at which the ionic strength was reduced.


similar to effects observed with other types of dis-

turbances, such as a depletion in ionic strength or a

sudden increase in flow rate.

5. Discussion

5.1. Transport characteristics of P. aeruginosa SG81 and

laponite RD through quartz sand

The study of the movement of colloids in isolation

shows that regardless of the nature of the colloidal

particle (bacteria, laponite) transport parameters follow

predictable patterns in terms of aggregation and

deposition at different ionic strengths. Even though

bacterial collision efficiencies were generally higher than

those of laponite, the bacterium was mobile over a wider

range of ionic strengths as compared to laponite, the

latter of which was more drastically affected by

relatively small changes in ionic strength. Highly

aggregated laponite suspensions (2000 mg L�1) were

retained in the porous medium with the influence of

clogging effects as observed by CLSM. Smaller aggre-

gates (20 mg L�1) were also retained inside clean bed

sand columns at high ionic strengths. These facts and the

mobility patterns observed indicate that the transport of

laponite is determined almost exclusively by electrostatic

interactions. In contrast, the transport of the bacterium

ARTICLE IN PRESS

Fig. 5. Confocal laser micrographs showing direct evidence for the detachment event observed indirectly in Fig. 4. The cells were

labelled with the DNA-binding stain SYTO 9. The micrographs belong to a 30 s interval time-series begun when a high ionic strength

flow cell influent (7� 10�2 M NaCl) was changed to deionised water. Light areas correspond to fluorescent SYTO 9 –labelled bacteria;

dark regions include the non-fluorescent regions of the sand grains and the surrounding interstitial solution.

Fig. 6. (A) Scanning electron micrograph showing the surface

of a sand grain (200 mm approx. diameter) with generally sparse

bacterial colonisation as suggested by EDX analysis, (B)

Region of bacterial colonisation on quartz sand. A bacterial

microcolony (enlargement from (A)) composed of several cells

and their extracellular polymer matrix is evident.


probably depends on other factors as well including

surface heterogeneities and the production of extracel-

lular polymeric substances.

Transport parameters obtained using the CLSM

method were equivalent to those obtained using sand

columns. This demonstrates that the method, even

though still under active development in our laboratory,

can be adequate for colloid transport studies. Although

the information provided by 2D image stacks was

sufficient to semi-quantitatively determine transport

parameters under carefully controlled conditions, the

method is expected to provide more detailed informa-

tion when more complete 3D image stacks are available.

The impediments to obtaining these types of images at

the moment are the impossibility to penetrate through

an opaque substratum such as sand grains and the

higher processing speed needed. This limits observations

to the first layer of sand grains and 100–200 mm from the

interface between sand grains and the flow cell wall to

the centre of the packed flow cell. One way of addressing

these limitations is the use of transparent materials in

ARTICLE IN PRESS

0 1 2 3 4 5 6 7 8

0.05

0.10

0.15

0.20

0.25

0.30

Abs

orba

nce

units

Pore volumes

0 1 2 3 4 5 6 7 8

Pore volumes

A535

5x108

4x108

3x108

2x108

1x108

0

5x108

4x108

3x108

2x108

1x108

0

cells

A535

cells

cel

ls.m

L-1 c

ells

.mL-1

0.05

0.10

0.15

0.20

0.25

0.30

Abs

orba

nce

units

(A)

(B)Fig. 7. Absorbance measurements and total cell counts of the effluents from inoculated sand columns after injection of 600 mL of a

200 mg L�1 laponite suspension at low ionic strength (6.2� 10�4 M NaCl). (A) Unstained laponite, (B) rhodamine-stained laponite.

After pulse, influent continued as 6.2� 10�4 M NaCl. Graphs are means of duplicate experiments. Bacteria are apparently remobilised

in two fronts the first of which was not sufficient to allow for bacterial detection by absorbance at 535 nm. As rhodamine–laponite

complex maximum absorption is at 535 nm probably, the first peak of ‘‘B’’ is due to the elution of laponite. There is an apparent early

elution and altered transport pattern due to hydrodynamic changes in the porous media caused by biofilm growth and to laponite/

bacteria co-elution.


conjunction with 2-photon microscopy which could

offer the possibility to assess both biofilm formation

and colloid transport at a higher range of distances and

even through several layers of particular substrata.

Nevertheless, CLSM has provided important insights

into the spatial distribution of biofilms growing inside

porous materials by allowing the determination of both

elution of colloids (outlet BTCs) and the direct

observation of colloid retention at different distances

from the inlet position. This is not achievable with sand

column experiments.

5.2. Colloid mobilisation phenomena in the presence of

microbial biofilms in porous media

In natural environments, the transport of colloids

rarely occurs in well-defined conditions. Variations

in ionic strength caused by for instance sea water

ARTICLE IN PRESS

1 3 50

0.05

0.10

0.15

0.20

0.25

0.30

A24

0

0

8.0x107

7.0x107

6.0x107

5.0x107

4.0x107

3.0x107

2.0x107

1.0x107

TCC

cel

ls m

L-1

sterile column (A240)

Pore volumes

biofilm growing column (A240)

0

1

2

3

4

5

6

7

8

pH

pH

2 4 6 7

Fig. 8. Column effluent after the introduction of 2000 mg L�1 laponite dispersed in 7� 10�2 M NaCl. Note the retention of a similar

laponite suspension in control (biofilm-free) columns (empty triangles). Results were obtained from duplicate experiments.


infiltration, torrential rain or de-icing will cause the

mobilisation or retention of a great variety of colloidal

particles. Clean bed transport studies are useful for

determining fundamental transport parameters in con-

trolled laboratory conditions. They fail, however to fully

describe transport in more complex environments. As

observed, bacteria have the ability to attach to surfaces,

stimulated by physicochemical factors which will also

stimulate the attachment of other colloidal particles

including inorganic colloids (Deshpande and Shonnard,

1999). They also are remobilised by similar physicho-

chemical factors causing detachment or release of other

colloids. As a consequence, bacterial biofilms growing in

porous materials are extremely dynamic (Figs. 4A, 5 and

8). Their influence on colloid and therefore contaminant

transport is expected to be composed of several additive

effects including changes in porous medium hydrody-

namics, sorption to biofilm components and co-elution

with re-mobilised biofilm elements. In this study, the

main challenge in studying colloid–colloid and biofilm–

colloid influences on their mutual transport through

sand columns under stimulated mobility conditions (i.e.

low ionic strength) was to setup an appropriate

detection system. One would expect that all species will

contribute to the absorption measurements; however, as

observed in Fig. 7 and as evidenced in other experiments

(data not shown), bacterial absorbance at 535 nm would

be negligible until their concentration reaches more than

1.5� 108 cells mL�1. As conditions changed, i.e. laponite

pulse and reduction in subsequent influent ionic

strength, bacteria were remobilised mainly after the

elution of laponite (Fig. 7A, B) was complete. The first

front of bacterial detachment in this case was not

sufficient to produce a measurable alteration in absor-

bance at 535 nm. In fact, the spectrometric monitoring

of bacteria is usually done at 240–280 nm. According to

Fig. 7B there is an alteration of laponite elution profile

(early breakthrough and tailing) due probably to a

reduction in the porous medium permeability (increased

particle velocity), bacterial co-elution and NaCl hydro-

dynamic dispersion. Laponite/bacteria co-elution could

be the result of direct interactions between laponite and

cells or cellular products such as polysaccharides. Clays

are known to interact strongly with polysaccharides as

reviewed by Chenu and Stotzky (2002), and polysac-

charides are quantitatively important components of the

extracellular biofilm matrix.

The introduction of laponite at an ionic strength and

concentration which would stimulate its aggregation and

retention in sand columns (Fig. 8) was able to stimulate

remobilisation of P. aeruginosa SG81 under conditions

in which retention would generally be favourable. The

influence of laponite on cellular detachment could be

due to indirect interactions caused by changes of the

hydrodynamic conditions inside the porous medium,

leading to increased shear forces—and therefore to cell

detachment. Even though pH variations were apparent

as the laponite was being eluted (Fig. 8), it is very

unlikely that changes of this magnitude (less than 2 pH

units) could have influenced the detachment of cells

(Jewett et al., 1995).

6. Conclusions

* Confocal microscopy was shown to be suitable

for online monitoring of fluorescently labelled

microorganisms and inorganic colloidal particles in


a sand-packed flow cell. This allowed the determina-

tion of deposition and transport characteristics of P.

aeruginosa SG81 and laponite RD as a function of

ionic strength and the qualitative determination of

colloidal retention inside the porous medium in a

non-destructive manner.* In clean bed columns, the mobility of laponite was

strongly influenced by ionic strength. Mobility was

inversely correlated to NaCl concentration; NaCl

concentrations of 1� 10�2 M resulted in almost

complete laponite retention within the sand columns.* Mobility of P. aeruginosa SG81 was reduced

predictably by NaCl concentrations up to 1 M. This

effect was not as pronounced as for laponite, and

some bacteria were still eluted under these condi-

tions.* At high relative ionic strength conditions

(7� 10�2 M), introduction of laponite to a sand

column containing a primary biofilm of P. aeruginosa

SG81 resulted in remobilisation of a portion of the

attached cells, as indicated by effluent cell numbers.

Low ionic strength (6.2� 10�4 M NaCl) caused

detachment of biofilm cells, but also an altered

laponite elution profile due to co-elution or hydro-

dynamic effects.* Environmental events leading to significant changes

in ionic strength of the subsurface matrix will cause

significant changes to the mobility of microorgan-

isms, inorganic colloids, and colloid-borne contami-

nants. These events include rainfall, melting snow or

ice containing dispersing salt, seawater infiltration,

and tidal processes at estuaries.

Acknowledgements

This work forms part of the cooperative research

project ‘‘KORESI’’-Kolloide Stofftransport bei der

Regenwassersickerung-financed by the German Re-

search Society (DFG). We thank Professors U. For-

stner, F.H. Frimmel and P.A. Wilderer, and their staff,

for constructive input and criticisms. The assistance of

Mariel Grooters is gratefully acknowledged.

References

Camesano, T., Unice, K., Logan, B., 1999. Blocking and

ripening of colloids in porous media and their implications

for bacterial transport. Coll. Surf. A: Physicochem. Eng.

Aspects 160, 291–308.

Chenu, C., Stotzky, G., 2002. Interactions between micro-

organisms and soil particles: an overview. In: Huang, P.M.,

Bollag, J.M., Senesi, N. (Eds.), Interactions Between Soil

Particles and Microorganisms. Wiley, New York, pp. 3–40.

Cunningham, A., Characklis, W., Abedeen, F., Crawford, D.,

1991. Influence of biofilm accumulation on porous media

hydrodynamics. Environ. Sci. Technol. 25, 1305–1311.

Deshpande, P., Shonnard, D., 1999. Modeling the effects of

systematic variation in ionic strength on the attachment

kinetics of Pseudomonas fluorescens UPER-1 in saturated

sand columns. Water Resour. Res. 35, 1619–1627.

Dyson, N., 1990. Statistical moments of a chromatographic

peak. In: Smith, R. (Ed.), Chromatographic Integration

Methods. The Royal Society of Chemistry, Cambridge,

pp. 20–25.

El-Masry, M., Hassouna, M., El-Rakshy, N., Mousa, I., 1995.

Bacterial populations in the biofilm and non-biofilm

components of a sand filter used in water treatment. FEMS

Microbiol. Lett. 131, 263–269.

Grobe, S., Wingender, J., Truper, H., 1995. Characterization of

mucoid Pseudomonas aeruginosa strains isolated from

technical water systems. J. Appl. Bacteriol. 79, 94–102.

Grolimund, D., Borkovec, M., Bartmettler, K., Sticher, H.,

1996. Colloid-facilitated transport of strongly sorbing

contaminants in natural porous media: a laboratory column

study. Environ. Sci. Technol. 30, 3118–3123.

Grolimund, D., Elimelech, M., Borkovec, M., Barmettler, K.,

Kretzschmar, R., Sticher, H., 1998. Transport of in situ

mobilized colloidal particles in packed soil columns.

Environ. Sci. Technol. 32, 3562–3569.

Grolimund, D., Elimelech, M., Borkovec, M., 2001. Aggrega-

tion and deposition kinetics of mobile colloidal particles in

natural porous media. Coll. Surf. A: Physicochem. Eng.

Aspects 191, 179–188.

Gross, M., Albinger, O., Jewett, D., Logan, B., Bales, R.,

Arnold, R., 1995. Measurement of bacterial collision

efficiencies in porous media. Water Res. 29, 1151–1158.

Jewett, D., Hilbert, T., Logan, B., Arnold, R., Bales, R., 1995.

Bacterial transport in laboratory columns and filters:

influence of ionic strength and pH on collision efficiency.

Water Res. 29, 1673–1680.

Jewett, D., Logan, B., Arnold, R., Bales, R., 1999. Transport of

Pseudomonas fluorescens strain P 17 through quartz sand

columns as a function of water content. J. Contam. Hydrol.

36, 73–89.

Kersting, A., Efurd, D., Finnegan, D., Rokop, D., Smith, D.,

Thompson, J.L., 1999. Migration of plutonium in ground

water at the Nevada test site. Nature 397, 56–59.

Kretzschmar, R., Barmettler, K., Grolimund, D., Yan, Y.,

Borkovec, M., Sticher, H., 1997. Experimental deter-

mination of colloid deposition rates and collision efficien-

cies in natural porous media. Water Resour. Res. 33,

1129–1137.

Li, Q., Logan, B., 1999. Enhancing bacterial transport for

bioaugmentation of aquifers using low ionic strength

solutions and surfactants. Water Res. 33, 1090–1100.

Lopez Arbeloa, F., Tapia Estevez, M., Lopez Arbeloa, T.,

Lopez Arbeloa, I., 1996. Spectroscopic study of the

adsorption of rhodamine 6G on clay minerals in aqueous

suspensions. Clay Miner. 32, 97–106.

McCarthy, J., Zachara, J., 1989. Subsurface transport of

contaminants. Environ. Sci. Technol. 23, 496–502.

Nicolai, T., Cocard, S., 2001. Dynamic light-scattering study of

aggregating and gelling colloidal disks. J. Coll. Int. Sci. 244,

51–57.


Okabe, S., Yasuda, T., Watanabe, Y., 1997. Uptake and release

of inert fluorescent particles by mixed populations biofilms.

Biotechnol. Bioeng. 53, 459–469.

Roy, S., Dzomak, D., 1997. Chemical factors influencing

colloid-facilitated transport of contaminants in porous

media. Environ. Sci. Technol. 31, 656–664.

Ryan, J., Elimelech, M., 1996. Colloid mobilization and

transport in groundwater. Coll. Surf. A: Physicochem.

Eng. Aspects 107, 1–56.

Sharp, R., Cunningham, A., Komlos, J., Billmayer, J., 1999.

Observation of thick biofilm accumulation and struc-

ture in porous media and corresponding hydrodynamic

and mass transfer effects. Water Sci. Technol. 39,

195–201.

Smets, B., Grasso, D., Engwall, M., Machinist, B.J., 1999.

Surface physicochemical properties of Pseudomonas fluor-

escens and impact on adhesion and transport through

porous media. Coll. Surf. B: Biointerfaces 14, 121–139.

Tapia Estevez, M., Lopez Arbeloa, F., Lopez Arbeloa, T.,

Lopez Arbeloa, I., Schoonheydt, R., 1993. Spectroscopic

study of the adsorption of rhodamine 6G on laponite B for

low loadings. Clay Miner. 29, 105–113.

Tufenkji, N., Ryan, J., Elimelech, M., 2002. Bank filtration.

Environ. Sci. Technol. 1, 423A–428A.

interactions between laponite and microbial biofilms in porous media: implications for colloid...

Documents