Literature data on the effect of calcium on biofilm structures induced a preliminary study. The effect of calcium removal by water softening (<1.0 mg Ca2+.L−1) under real-world drinking water conditions on biofilm formation was studied in a pilot plant with reverse osmosis (RO) membranes and in a laboratory-scale biofilm production unit (BPU) with plasticized polyvinyl chloride (PVC-P) and glass as substratum. The results showed a halving of the exponential biofouling rate in the RO membrane and also a halving of the exponential carbohydrate (CH) production rate in the biofilm on PVC-P and glass in the BPU by softening of the water. In PVC-P biofilms, softening did not affect adenosine tri-phosphate (ATP) production and bacterial species composition (terminal restriction fragment length polymorphism analysis). At low substrate concentrations in glass and RO membrane biofilms softening reduced significantly ATP and CH production and changed the species composition on the membrane. The importance of the two hypothesized physical or physiological mechanisms as causes for the observed Ca2+ effect on biofilm formation and the effect of Ca2+ concentration on those, needs further studies.

INTRODUCTION

Biofilm formation or biofouling in water systems can cause health related or aesthetic water quality problems or operational problems in treatment processes (filtration, recharge well infiltration, reversed osmosis (RO)/nanofiltration (NF) membranes; production loss and increased energy consumption). In the Netherlands, drinking water is distributed without residual chlorine and for biofouling control in RO/NF membranes chlorine cannot be applied. To control biofouling problems: (i) microbial growth potential of water (assimilable organic carbon (AOC)) must be reduced to pre-defined levels assessed with well-defined methods, (ii) nutrient-free chemicals (anti-scalants) and piping materials should be applied and (iii) for cleaning most effective cleaning agents must be selected (Rittmann & Snoeyink 1984; Van der Kooij & Van der Wielen 2014; Hijnen & van der Kooij 1992; Hijnen et al. 2009, 2012). Despite these preventive strategies, some extent of biofilm formation will occur in water systems. Exopolymeric substances (EPS) composed of a sticky negatively charged network of polysaccharides, proteins and ribonucleic acids largely determine the stability of biofilms (Flemming & Wingender 2010). It is known from the literature that the presence of multivalent cations, especially Ca2+ affects the structure of biofilms (Flemming & Wingender 2001; Wloka et al. 2004; de Kerchove & Elimelech 2008; Al-Amoudi 2010; Zhao et al. 2015). Increased adhesion and aggregation of organic matter, bacteria and their produced biopolymers have been documented. Calcium enhances the resistance of these biofouling networks to mechanical and chemical stress endured during cleaning and is most likely one of the reasons for the relatively low cleaning efficiencies of commonly used reversed osmosis and nanofiltration (RO/NF)-membrane cleaning procedures (Cornelissen et al. 2007; Hijnen et al. 2012). Many of the above-mentioned studies on the effect of Ca2+, however, are short running laboratory-scale studies using model foulants (i.e. alginate) and pre-cultured bacterial cells (i.e. Pseudomonas aeruginosa) under wastewater water conditions with high nutrient levels. In our previous studies on membrane fouling it was demonstrated that two biomass parameters adenosine tri-phosphate (ATP) and carbohydrate (CH) production were significantly inter-correlated and correlated with pressure drop in the feed channel. To investigate the effect of removal of multivalent cations under ‘real-world’ drinking water conditions on naturally grown biofilms and membrane biofouling a preliminary study was conducted at laboratory and pilot plant scale. The objective was to investigate the effect of Ca2+ removal by softening on the pressure drop development in the feed channel of a RO-membrane module and on the biomass production (ATP/CH) and composition (DNA based) on the surface of plasticized polyvinyl chloride (PVC-P) at high nutrient levels, and on the surfaces of glass and the RO-membrane at low nutrient levels.

MATERIAL AND METHODS

Experimental set-up

The effect of calcium removal by softening on biofilm formation was studied in two small scale biofilm production units (40 L.h−1; BPU) and in a spiral-wound RO membrane pilot (350 L.h−1; spiral wound membrane (SWM)) with two separate RO-modules (Figure 1). Both systems were supplied with pre-filtered tap water (TW) without or with softening (soft water (SW)). Prior to this, TW was pre-treated by cartridge filters (07PP001-09NNP, Borselen Ing.; 1 μm) to eliminate particulate matter. The cartridge filters were replaced every 7 days. The water softening unit consisted of two cation ion exchange columns which were operated in parallel and regenerated with NaCl every 7,000 L (approximately 10 hours). The average water quality of the TW and SW is presented in Table 1 and clearly showed removal of calcium and magnesium after softening. The membrane setup was equipped with two 2.5-inch Filmtec RO membranes (TW30-2540 DOW). The membranes were operated at a feed pressure of approximately 10 bar, a flow rate of 350 L.h-1 and a permeate flow of 75 L.h−1. To initiate biofilm formation in the RO-membranes and on the glass samples in the BPU, sodium acetate was dosed at 10 μg C.L−1 to the TW and SW by a dosing unit consisting of a pump (ProMinent gamma/L, 0.5 L.h−1), acetate feed tank (50 L) and a measuring scale (Satex SA 250).
Figure 1

The experimental scheme of the study with TW and softened tap water.

Figure 1

The experimental scheme of the study with TW and softened tap water.

Table 1

The pH (±SD, n = 26), conductivity (± SD, μS.cm−1, n = 29) and the ion concentrations in the feed TW and SW (±SD, mg.L−1, n = 6) determined by ICP-MS (inductively coupled plasma mass spectrometry)

 TW (mg.L−1)SW (mg.L−1)
pH 7.72 ± 0.18 7.88 ± 0.19 
Conductivity 248 ± 13 243 ± 13 
Ca 36.1 ± 2.2 0.3 ± 0.3 
Mg 6.9 ± 0.2 0.02 ± 0.01 
Fe 0.027 ± 0.02 0.017 ± 0.007 
Al 0.011 ± 0.02 0.011 ± 0.01 
Na 14.7 ± 0.8 77.9 ± 3.1 
Mn 0.001 ± 0.001 0.0005 ± 0.001 
Sr 0.149 ± 0.005 0.001 ± 0.0005 
1.3 ± 0.1 0.4 ± 0.2 
Ba 0.013 ± 0.001 0.000 ± 0.0001 
 TW (mg.L−1)SW (mg.L−1)
pH 7.72 ± 0.18 7.88 ± 0.19 
Conductivity 248 ± 13 243 ± 13 
Ca 36.1 ± 2.2 0.3 ± 0.3 
Mg 6.9 ± 0.2 0.02 ± 0.01 
Fe 0.027 ± 0.02 0.017 ± 0.007 
Al 0.011 ± 0.02 0.011 ± 0.01 
Na 14.7 ± 0.8 77.9 ± 3.1 
Mn 0.001 ± 0.001 0.0005 ± 0.001 
Sr 0.149 ± 0.005 0.001 ± 0.0005 
1.3 ± 0.1 0.4 ± 0.2 
Ba 0.013 ± 0.001 0.000 ± 0.0001 

Biofilm sample production

The BPU consists of a hard PVC-chamber with five separate parallel rows of PVC-P samples (30 × 20 × 3 mm W × L × H) placed vertically and thus creating six flow channels of 4.5 mm in width and 25 mm in height with a removable transparent perspex cover darkened during operation to avoid light exposure. A centrifugal pump (Verder International B.V.) was used to recirculate the water inside the system with a steady flow of 350 L/h creating laminar flow conditions (Re = 936) with a flow rate of 0.14 m.s−1 similar to the flow conditions in spiral-wound membranes. The chamber was continuously supplied with fresh TW or SW at 20 L.h−1 The biofilm production in the BPU was established by the internal substrate in the PVC-P samples which generates a thick biofilm mimicking severe biofouling in SW membranes (Hijnen et al. 2012). To investigate the softening effect in the BPU at the same biofouling conditions as realized in the SW-membranes (nutrient free material and low nutrient supply, see below), glass-samples (25 × 25 × 1 mm) were introduced in the chamber. To prevent nutrient limitation nitrogen (0.9 mg NO3-N.L−1) and phosphorus (0.35 mg PO4-P.L−1) was dosed to the system using a pulse pump (Masterflex Easyload II) with a flow rate of 0.5 mL.min−1 from concentrated nutrient solutions prepared in sterilized MilliQ water (4.331 g KNO3.L−1 and 1.045 g K2HPO4.L−1). A second run with new PVC-P samples was started to verify the role of Ca2+ in the biofouling process. Calcium (40 mg Ca2+.L−1; Table 1) was dosed to the SW supplied of the BPU from a concentrated and acidified CaCl2 solution (73.845 g.L−1; pH 1.5).

Biofouling rate and biofilm production monitoring

The exponential biofouling rate Rf in the RO membranes was calculated as described in Hijnen et al. (2011a). Rf and CH-production rate were calculated using Microsoft-Excel©. Biofilm formation on the PVC-P and glass samples was monitored during the operational time and on the RO-membrane after destructive autopsy of the membrane modules at the end of the experimental period, based on a relative normalized pressure drop increase (NPDi) maximum of 1,000%. Biofilm concentrations were monitored with ATP (active biomass) and CH (carbohydrates) at the inlet (0–10 cm), in the middle (45–55 cm) and the outlet (90–100 cm) of the membrane by the methods described by Hijnen et al. (2012).

DNA extraction and analysis

DNA of the biofilm from PVC-P (6 × 20 × 3 mm) and membrane pieces (10 × 6 mm) that were sampled after day 1, 7 and 29 days was extracted using a commercially available DNA extraction kit (MO-BIO PowerBiofilm© DNA Isolation kit, Cat.# 24000–50). After extraction, DNA was eluated in buffer and kept at −18°C until further analyses. The stored DNA samples were analysed for the microbial population profile using a 16S rRNA gene-targeted terminal restriction fragment length polymorphism (t-RFLP). Amplification of the 16S rRNA gene was performed by polymerase chain reaction (PCR) using the bacteria-specific primers 8F-FAM labelled with a fluorochrome (5′-FAM-AGAGTTTGATC(A/C)TGGCTCAG-BHQ-3′) and 1392R (5′-ACGGGCGGTGTGTACA-3′). The generated fragments were treated with the Clean & Concentrator Kit 5 from Zymo Research and subsequently digested with the restriction enzyme HhaI. The Clean and Concentrator Kit was again used to purify the restriction products according to the suppliers’ manual. Separation of the digestion products and the added base-pair marker and detection of labelled fragments were achieved by capillary electrophoresis with an ABI Prism 310 genetic sequencer (Applied Biosystems, USA). Analysis of t-RFLP profiles was conducted with the supplied software (BioNumerics).

RESULTS AND DISCUSSION

Biofouling of the RO membranes

Biofilm growth in the RO membrane was monitored with pressure drop increase. The biofouling rate in the RO membranes supplied with TW and SW was described by an exponential function as also observed for the biofouling rate in a previously presented study (Hijnen et al. 2009; 2011a) in a membrane fouling simulator (MFS) (Vrouwenvelder et al. 2007). The exponential fouling rate (Rf) calculated for the RO element supplied with TW was 0.199 ± 0.012 d−1 (Table 2). It is noteworthy to mention that this biofouling rate in this ‘real-world’ RO-membrane supplied with TW enriched with 10 μg acetate-C L−1 was almost similar to the biofouling rate presented for the MFS study with the same acetate concentration (Hijnen et al. 2011a; Rf = 0.205 d−1). This observation clearly demonstrates that the biofouling process in membranes induced by easily biodegradable compounds occurs predominantly in the first 0.2 m of a RO element with a total length of 1.0 m and is well simulated by the MFS with a length of 0.2 cm. The authors do not know of more membrane biofouling studies presenting the exponential biofouling rate at similar low acetate-C concentrations.

Table 2

The exponential biofouling rate Rf (d−1; regression statistics), final relative pressure drop increase at the membrane autopsy and the biomass concentrations at different locations of the membrane supplied with TW and SW enriched with acetate

ParameterAutopsy locationTW (10 μg ac.-C.L−1)SW (10 μg ac.-C.L−1)
Rf (LN[ΔP] d−1 ±SD; n, r2 0.199 ± 0.012 (23; 0.93) 0.091 ± 0.005 (24; 0.95) 
NPD increase (%)  710% in 34 days 360% in 51 days 
ATP (ng ATP.cm−2Inlet 38.2 ± 0.7 (n = 3) 13.4 ± 0.4 (n = 3) 
Middle 5.0 ± 1.2 (n = 3) 2.1 ± 0.3 (n = 3) 
Outlet 2.3 ± 1.3 (n = 3) 0.95 ± 0.06 (n = 3) 
CH (μg CH.cm−2Inlet 92.6 ± 11.6 (n = 3) 30.6 ± 0.4 (n = 3) 
Middle 16.4 ± 3.3 (n = 3) 14.1 ± 0.1 (n = 3) 
Outlet 13.2 ± 2.0 (n = 3) 8.1 ± 0.3 (n = 3) 
ParameterAutopsy locationTW (10 μg ac.-C.L−1)SW (10 μg ac.-C.L−1)
Rf (LN[ΔP] d−1 ±SD; n, r2 0.199 ± 0.012 (23; 0.93) 0.091 ± 0.005 (24; 0.95) 
NPD increase (%)  710% in 34 days 360% in 51 days 
ATP (ng ATP.cm−2Inlet 38.2 ± 0.7 (n = 3) 13.4 ± 0.4 (n = 3) 
Middle 5.0 ± 1.2 (n = 3) 2.1 ± 0.3 (n = 3) 
Outlet 2.3 ± 1.3 (n = 3) 0.95 ± 0.06 (n = 3) 
CH (μg CH.cm−2Inlet 92.6 ± 11.6 (n = 3) 30.6 ± 0.4 (n = 3) 
Middle 16.4 ± 3.3 (n = 3) 14.1 ± 0.1 (n = 3) 
Outlet 13.2 ± 2.0 (n = 3) 8.1 ± 0.3 (n = 3) 

Biofilm formation on the membrane material

After 34 and 51 days for, respectively, TW and SW the relative pressure drop increase (normalized pressure drop increase (NPDi)) of these membranes was 710% and 360% and the membranes were sampled by destructive autopsy. In accordance with the difference in NPDi at the time of autopsy (Table 2), the biomass concentration (ATP and CH) in the membrane element supplied with TW was higher than in the element supplied with SW. This corroborates with the difference in biofouling rates observed in the TW and SW elements. ATP concentrations at the inlet of both membranes were a factor of seven higher than in the middle section (Table 2). CH concentrations in the inlet of the elements were a factor of 5.6 and 2.2 higher than in the middle for, respectively, the element supplied with TW and SW. The distribution of biomass over the flow direction in the membranes corresponds with the above-mentioned conclusion that fast and severe biofouling of spiral-wound membranes caused by low concentration of easily biodegradable compounds occurs in the first 20 cm of the membrane. Biomass accumulation in the feed-channel of spiral-wound membranes simulated by the MFS and determined with the ATP and CH methods, have been significantly correlated with the pressure drop increase (Hijnen et al. 2011b). From the presented equations in this previous study and the NPDi observed in the RO elements of the current study (710 and 360%) ATP concentrations (25.3 and 12.8 ng ATP cm−2) and CH concentrations (17.6 and 12.1 μg CH cm−2) have been calculated. Comparison with the inlet concentrations observed in the RO elements (Table 2) revealed that the calculated and observed ATP concentrations were in the same order of magnitude. The observed CH concentrations in the RO membranes (Table 2) were a factor of 2–6 higher than the calculated CH concentrations. The latter concentrations were calculated from a correlation with highly variable CH concentrations observed in the membrane feed channels (Hijnen et al. 2011b).

Biofilm formation on the PVC-P and glass samples

Biofilm growth in the BPU loaded with PVC-P and glass samples is demonstrated by the steady increase of both ATP and CH concentrations on the surface (Figure 2).
Figure 2

The biofilm growth (error bar ±SD) determined with ATP and CH on PVC-P and glass samples supplied with TW, SW or SW + Ca.

Figure 2

The biofilm growth (error bar ±SD) determined with ATP and CH on PVC-P and glass samples supplied with TW, SW or SW + Ca.

This biofilm growth on the PVC-P samples was clearly higher than on the glass samples, thus showing the difference in nutrient conditions of the AOC leaching PVC-P material and of the glass surface supplied with TW and SW supplemented with 10 μg acetate-C.L−1. Important to notify is the comparison of the ATP and CH concentrations on the glass (Figure 2) and in the membrane material (Table 2), which were in the same order of magnitude and much lower than on the PVC-C.

Under a high nutrient level on PVC-P, no difference in ATP concentrations supplied with TW and SW was observed for either run 1 or 2. Differently from the ATP production, however, the CH production on the PVC-P samples showed a lower production rate for the samples supplied with SW-1 compared to the samples supplied with TW-1. On the glass samples, both ATP and CH production on the samples supplied with TW was higher than ATP and CH production on the samples supplied with SW. In the second run with PVC-P samples, Ca2+ addition to the SW (SW-2 + Ca) resulted in a similar CH production rate as observed for the TW-2, thus demonstrating the effect of Ca2+ on CH production in the biofilm.

The CH production rates on the different surfaces were described by an exponential function and for the different tests summarized in Table 3. The exponential CH production rates (RCH; μg CH.cm−2 d−1) were calculated from the first part of the production curves (<60 days) with the highest regression coefficients. For the tests TW-1, TW-2 and SW-2 + Ca the RCH was 0.079, 0.075 and 0.071 μg CH cm−2.d−1, whereas for SW-1 the RCH was 0.039 μg CH cm−2.d−1. The RCH for the glass samples was 0.026 for the TW and 0.010 μg CH cm−2.d−1 for the SW. Thus, on PVC-P and glass the CH production rate in the biofilms was approximately a factor of two lower when the samples were exposed to SW with a decreased Ca2+ concentration of 0.3 mg.L−1 (Table 1).

Table 3

The exponential CH production rate (±SD) of the different tests with different substrates (n = number of data points during the production period (Figure 1))

TestSubstrateRCH CH-production rate (LN[CH] d−1)Period (d)nR2
TW-1 PVC-P 0.079 ± 0.005 21–56 0.98 
SW-1 PVC-P 0.039 ± 0.003 21–56 0.98 
TW-2 PVC-P 0.075 ± 0.008 7–48 0.95 
SW-2* PVC-P 0.071 ± 0.009 7–48 0.93 
TW-1 Glass + 10 μg acet.-C L−1 0.026 ± 0.003 13–30 0.99 
SW-1 Glass + 10 μg acet.-C L−1 0.010 ± 0.014 17–37 0.33 
TestSubstrateRCH CH-production rate (LN[CH] d−1)Period (d)nR2
TW-1 PVC-P 0.079 ± 0.005 21–56 0.98 
SW-1 PVC-P 0.039 ± 0.003 21–56 0.98 
TW-2 PVC-P 0.075 ± 0.008 7–48 0.95 
SW-2* PVC-P 0.071 ± 0.009 7–48 0.93 
TW-1 Glass + 10 μg acet.-C L−1 0.026 ± 0.003 13–30 0.99 
SW-1 Glass + 10 μg acet.-C L−1 0.010 ± 0.014 17–37 0.33 

*Calcium added.

The biofouling rate in the RO element supplied with SW was a factor of two lower, resulting in an Rf-value of 0.091 ± 0.005 (d−1; Table 2). This effect of softening on biofouling in RO elements was in the order of magnitude similar to the effect observed for the CH production rates in the BPU on both PVC-P and glass.

t-RFLP analysis of the biofilms

16S rRNA gene-targeted t-RFLP resulted in peak patterns (fingerprints). These fingerprints are compared and clustered based on similarity of bacterial community composition with a percentage of relationship as a quantitative measure which is presented in a dendrogram (Figure 3). The species composition of the cluster is not significantly different at a level of ≥94% (unpublished data). DNA isolates of PVC-P biofilm samples produced in the BPU after 7, 14 and 29 days of the first experiment (TW-1 and SW-1) and the membrane autopsy biofilm samples supplied with TW and SW were analysed.
Figure 3

Dendrogram of the t-RFLP patterns of the 16S rRNA gene amplified from DNA samples that were isolated from biofilm samples from the BPU (PVC-P) and RO-membranes supplied with TW and SW.

Figure 3

Dendrogram of the t-RFLP patterns of the 16S rRNA gene amplified from DNA samples that were isolated from biofilm samples from the BPU (PVC-P) and RO-membranes supplied with TW and SW.

Each sample had a relatively high number of peaks with t-RFLP (shown as bands in the Supplementary Material, Figure S1, available with the online version of this paper), indicating a relatively high diversity of the dominant bacterial population in each sample. The samples Membrane TW Middle and Membrane TW Outlet had more peaks (and thus a higher bacterial diversity) than the other samples. The t-RFLP patterns of each sample were also compared and the dendrogram of this comparison is shown in Figure 3. The results from this comparison demonstrate that the bacterial species composition on the PVC-P samples in the BPU is changing from 7 to 29 days of operation, although the communities formed on PVC-P supplied with TW and SW are very similar at a given day. Furthermore, the bacterial species composition on PVC-P samples in the BPU seems to be more different between TW and SW supplied biofilm when the incubation time increases. Still, after 29 days of incubation the population structures on the PVC-P supplied with both waters is very similar (>90% similarity), which coincides with similar active biomass (ATP) concentrations on the PVC-P supplied with TW and SW (Figure 2).

The t-RFLP results from the PVC-P samples in the BPU are in contrast to the t-RFLP patterns obtained with the RO-membranes. There, it was observed that the bacterial species composition in TW supplied membranes was less than 50% similar to the bacterial species composition in SW supplied membranes. In addition, the bacterial population at the inlet of both membranes (TW and SW) was significantly different from the bacterial populations sampled in the middle and outlet of the membranes (<86% and <92% similarity for SW and TW membrane, respectively). These results show clearly that the species composition of the bacterial population in the biofilm of those membranes is also affected by the softening of the water, which coincides with differences in biomass concentration (ATP and CH; Table 2).

Hypothesis on the role of Ca2+

Removal of Ca2+ is one of the major effects of softening and softening of the water affected (a) the CH production rate in the biofilm on PVC-P and glass and the biofouling rate in the RO elements to the same extent, (b) the biomass concentrations on the PVC-P, glass and membrane surface and (c) the composition of the bacterial population in the membrane. It can be hypothesized that Ca2+ concentration in the water (i) affects the physical characteristics of biofilms with lower resistance to shear and increased detachment of biofilm at low Ca2+ concentrations and/or (ii) affects the bacterial population by selecting for biofilm-associated bacteria synthesizing more and/or different exo-polysaccharides at higher Ca2+ concentrations as presented above. The high effect of Ca2+ on the EPS-matrix stability due to its strong physical association with polysaccharides has been recognized in literature (i.e. Flemming & Wingender 2001; Wloka et al. 2004; de Kerchove & Elimelech 2008; Al-Amoudi 2010; Zhao et al. 2015). In the current study, Ca2+ dosing to the SW resulted in a similar CH production rate on the PVC-P samples as observed for TW. Moreover, the other multivalent cation, magnesium, which was also reduced by softening (Table 1), plays a minor role in biofilm stability compared to calcium (Lattner et al. 2003).

There are also studies which indicate a biotic and physiological role of Ca2+ in biofilm-associated bacteria. A brief literature review on the ecological function of Ca2+ in biofilms demonstrates that Ca2+ is involved in the gene expression of the polysaccharide synthesizing activity in pathogenic biofilm-associated bacteria (Kierek and Watnick 2003; Sarkisova et al. 2005; Patrauchan et al. 2007). This may select for biofilm-associated bacterial species with different exopolysaccharide synthesis characteristics. The results of the PVC-P biofilms (no differences in species and ATP concentration for TW and SW) indicate that the physical role of Ca2+ is dominant under these conditions. The results of the membranes (differences in species as well as in ATP concentration for TW and SW), however, suggest that the influence of Ca2+ on the exopolysaccharide synthesis by biofilm-associated bacteria is also occurring in fresh water biofilms at low nutritional conditions. Future microbiological studies with, for instance, next generation sequencing and physiological studies with isolated pure cultures are required to elucidate which of the above-mentioned roles of Ca2+ has the largest impact on biofilm growth and stability: the physical effect of Ca2+ on biofilm stability or the physiological effect on polysaccharide synthesis in biofilm-associated bacteria. The outcome of such studies may also reveal that both the physical and physiological mechanisms are occurring simultaneously and the significance of one or the other depends on other yet unknown environmental factors.

CONCLUSIONS

Based on the results of this preliminary study it was concluded that deep softening of water and Ca2+ removal to concentrations <1.0 mg L−1 affects naturally grown biofilms in terms of biomass concentration measured with ATP and CH and in terms of composition of the microbial population determined with DNA based t-RFLP. Under ‘real-world’ drinking water conditions in an RO membrane module at pilot plant scale, pressure drop increase caused by biofouling and biomass accumulation was reduced. Under low nutrient conditions, bacterial population composition was changed significantly by Ca2+ removal. Evidence for the exclusive role of Ca2+ was derived from an additional experiment with Ca2+ to the SW which increased CH production rate to the same rate as observed in the non-SW. Thus, softening of the feed water of water systems may result in reduced biofilm growth which additionally may be more susceptible to common cleaning procedures (flushing, filter back washing and RO/NF membrane cleaning). Further research is required to elucidate if besides the physical effect of Ca2+ on the biofilm-matrix, the physiological effect on CH synthesis in biofilm-associated bacteria play a role in the biofouling process and to verify the effect of different Ca2+ concentrations.

ACKNOWLEDGEMENTS

The research was conducted as part of the preliminary Research Fund of KWR Watercycle Research Institute. The authors want to acknowledge H. van Wegen, A. van der Veen, A. J. Brouwer-Hanzens and L. Bereschenko for their excellent technical assistance.

REFERENCES

REFERENCES
Cornelissen
E. R.
Vrouwenvelder
J. S.
Heijman
S.
Viallefont
X.
van der Kooij
D.
Wessels
L.
2007
Periodic air/water cleaning for control of biofouling in spiral wound membrane elements
.
J. Membr. Sci.
287
(
II
),
94
101
.
Flemming
H.-C.
Wingender
J.
2001
Relevance of microbial extracellular polymeric substances (EPSs) – Part 1: structural and ecological aspects
.
Wat. Sci. Tech.
43
(
6
),
1
8
.
Flemming
H.-C.
Wingender
J.
2010
The biofilm matrix
.
Nat. Rev. Microbiol.
8
,
623
633
.
Hijnen
W. A. M.
Biraud
D.
Cornelissen
E. R.
van der Kooij
D.
2009
Threshold concentration of easily assimilable organic carbon in feedwater for biofouling of spiral-wound membranes
.
Environ. Sci. Technol.
43
(
II
),
4890
4895
.
Hijnen
W. A. M.
Biraud
D.
Cornelissen
E. R.
van der Kooij
D.
2011a
Addition/correction: threshold concentration of easily assimilable organic carbon in feedwater for biofouling of spiral-wound membranes
.
Environ. Sci. Technol.
45
(
II
),
1744
1745
.
Hijnen
W. A. M.
Cornelissen
E. R.
van der Kooij
D.
2011b
Threshold concentrations of biomass and iron for pressure drop increase in spiral-wound membrane elements
.
Water Res.
45
(
4
),
1607
1616
.
Hijnen
W. A. M.
Castillo
C.
Brouwer-Hanzens
A. H.
Harmsen
D. J.
Cornelissen
E. R.
van der Kooij
D.
2012
Quantitative assessment of the efficacy of spiral-wound membrane cleaning procedures to remove biofilms
.
Water Res.
46
(
19
),
6369
6381
.
Kierek
K.
Watnick
P. I.
2003
Environmental determinants of Vibrio cholerae biofilm development
.
Appl. Environ. Microbiol.
69
(
9
),
5079
5088
.
Lattner
D.
Flemming
H.-C.
Mayer
C.
2003
13C-NMR Study of the interaction of bacterial alginate with bivalent cations
.
Int. J. Biol. Macromol.
33
,
81
88
.
Patrauchan
M. A.
Sarkisova
S. A.
Franklin
M. J.
2007
Strain-specific proteome responses of Pseudomonas aeruginosa to biofilm-associated growth and to calcium
.
Microbiology
153
,
3838
3851
.
Rittmann
B. E.
Snoeyink
V. L.
1984
Achieving biologically stable drinking water
.
J. Am. Water Works
76
(
10
),
106
114
.
Sarkisova
S.
Patrauchan
M. A.
Berglund
D.
Nivens
D. E.
Franklin
M. J.
2005
Calcium-induced virulence factors associated with the extracellular matrix of mucoid Pseudomonas aeruginosa biofilms
.
J. Bacteriol.
187
,
4327
4337
.
Van der Kooij
D.
van der Wielen
P. W. J. J.
2014
Microbial Growth in Drinking-Water Supplies
.
IWA Publishing
,
London
,
UK
.
Vrouwenvelder
J. S.
Bakker
S. M.
Wessels
L. P.
van Paassen
J. A. M.
2007
The membrane fouling simulator as a new tool for biofouling control of spiral-wound membranes
.
Desalination
204
(
II
),
170
174
.
Zhao
F.
Xu
K.
Ren
H.
Ding
L.
Geng
J.
Zhang
Y.
2015
Combined effects of organic matter and calcium on biofouling
.
J. Membr. Sci.
486
(
II
),
177
188
.

Supplementary data