Reverse osmosis biofouling: Impact of feed channel spacer and biofilm development in spacer-filled channels – MyronLMeters.com
Water desalination via reverse osmosis (RO) technology provides a solution to the world’s water shortage problem. Until now, the production of fresh water from seawater has reached 21-million cubic meter per day all around the world (Wangnick, 2005). However, the success of RO technology is subject to improvement as the technology is challenged by a biofouling problem –a problem related to biological material development which forms a sticky layer on the membrane surface (Flemming, 1997; Baker and Dudley, 1998).
Continuous biofouling problems in RO lead to higher energy input requirement as an effect of increased biofilm resistance (Rf) and biofilm enhanced osmotic pressure (BEOP), lower quality of product water due to concentration polarization (CP) – increased concentration due to solutes accumulation on the membrane surface, (Herzberg and Elimelech, 2007), and thus significant increase in both operating and maintenance costs.
Recent studies and objectives
Recent studies show the importance of the operating conditions (e.g. flux and cross flow velocities) in RO biofouling. The presence of feed channel spacers has also been getting more attention as it may have adverse effects. A previous study (Chong et al., 2008) without feed channel spacers showed that RO biofouling was a flux driven process where higher flux increased fouling rate. It was also shown that biofouling caused a BEOP effect due to elevated CP of solutes at the membrane surface, thus resulted in loss of driving force. The BEOP effect was more severe at high flux and low crossflow operation.
In another recent study (Vrouwenvelder et al., 2009a) involving feed channel spacers suggested that flux did not affect fouling and biofouling was more severe when the crossflow velocity was higher. However, these studies were conducted on river water at low level of salinity and under no/very low flux conditions, which may suggested that BEOP effect was not observed in the above studies. These contradictory observations relating to the biofouling process in RO need to be systematically addressed as it is critical to understand the mechanism for sustainable operation of RO technology.
The objective of this study was to observe the impact of spacer towards RO biofouling as well as to investigate the development of biofilm in a spacer filled channel. The experiments were conducted at constant flux and biofouling was observed by the increase of transmembrane pressure (TMP). Observation with confocal light scanning microscope (CLSM) method was conducted to the fouled membrane and spacers to provide information of biofilm development inside the membrane module.
Materials and methods
A lab-scale set-up was arranged to resemble the real RO operation where experiments were performed with elevated salinity, high pressure, imposed flux, and permeation. The schematic diagram of the set-up is depicted in Figure 1. It is a fully-recycled system with two identical RO modules running in series. Feed solution contained constant amounts NaCl and nutrient broth (NB) to provide sufficient TOC level.
The study was conducted in the constant flux mode and biofouling was measured via the rise in TMP. A mass-flow controller was installed at the permeate side to maintain the amount of permeate withdrawn. A bacteria solution was injected into the system before the feed solution entered the RO modules and a set of microfilters (5 μm and 0.2 μm) were installed at downstream to prevent excess bacteria from entering the feed tank and turning the feed tank into an “active bioreactor”.
Model bacteria Pseudomonas aeruginosa (PAO1) was used in the experiment. Bacteria stock solution used in the biofouling tests was prepared in batch and the stock solution was replenished every 24 hours. Bacteria were grown in mixture of NB and NaCl solution where they were harvested after 24 hours and diluted into autoclaved salt solution. The concentration of bacteria was controlled and measured by optical density (OD) using UV spectrophotometer at 600 nm. Batch prepared bacteria stock solution has some advantages over using continuous feed from a chemostat (Chong et al., 2008). A more consistent and fresh bacteria load and without excess nutrient was introduced into the system as nutrient content was completely removed in the harvesting step.
Prior to every experiment, cut RO membranes (DOW Filmtec, BW-30) were soaked in Milli-Q water and sterilized in 70% ethanol solution. Similar pretreatment procedures were applied to membrane support layers and feed channel spacers prior every experiment. The spacers used in the experiments are obtained from unused Hydranautics LFC-1 spiral wound module (Figure 2).
The membranes were compacted at a maximum flux (~65 L/m2.h) overnight with Milli-Q water until a stable flux was achieved. Following compaction, the flux was set to the desired values and NaCl solution was added into the feed tank until the desired concentration was achieved. The system was let to mix for 1.5 hours. NB solution was then added into the feed tank to provide an average background nutrient concentration of 6.5 mg/L TOC. The system was allowed to well-mix for 1.5 hours.
The biofouling test was initiated by continuous injection of bacteria stock solution into the flow line at a dilution rate of 1:500 based on RO cross-flow rate. Biofilm was allowed to grow on the RO membranes. TMP rise due to biofouling was measured over time. The solution in the feed tank was removed and replaced with a fresh solution at the same NaCl and NB concentration twice per day in order to maintain the freshness level of the feed solution.
Upon completion of the fouling test, the RO system was cleaned with:
Tap water adjusted to pH 2 with HNO3 for 1.5 hours
Tap water adjusted to pH 11 with NaOH for 1.5 hours
Flowing tap water for rinsing for 1.5 hours
Final rinsing with Milli-Q water at unadjusted pH
The fouled membranes were removed from the RO cells for membrane autopsy. In this analysis, fluorescence staining methods and confocal laser scanning microscope (CLSM) were used to detect the biofilm.
Biofilms were prepared for CLSM by staining with the LIVE/DEAD BacLight Bacterial Viability Kits (Molecular Probes, L7012). It consists of SYTO 9 green-fluorescent nucleic acid stain and the red-fluorescent nucleic acid stain, propidium iodide (PI). These stains possess different spectral characteristics and different ability to penetrate healthy bacterial cells. When used alone, the SYTO 9 stain generally labels all bacteria in a population — those with intact and damaged membranes. In contrast, propidium iodide penetrates only bacteria with damaged membranes, causing a reduction in the SYTO 9 stain fluorescence when both dyes are present. Thus, with an appropriate mixture of the SYTO 9 and propidium iodide stains, bacteria with intact cell membranes stain fluorescent green, whereas bacteria with damaged membranes stain fluorescent red.
Microscopic observation and image acquisition of biofilms were performed using a confocal laser scanning microscope (ZEISS, model LSM710), equipped with Argon laser at 488 nm and DPSS561-10 laser at 561 nm. Images were captured using confocal microscope bundled program ZEN 2009.
Results and discussions
The cross-flow velocity (CFV) in RO membrane operations is known to affect fouling rate. At higher CFV, the flow causes scouring effects which results in slower fouling (Koltuniewicz et al., 1995). On the other hand, experiments of RO modules without the presence of flux shows that a higher cross-flow velocity may increase biofouling due to more nutrients supply (Vrouwenvelder et al., 2009b).
In our study, the investigation was carried out by varying the cross-flow velocity (CFV) from
0.1, 0.17, to 0.34 m/s. The NaCl concentration used was constant at 2000 mg/L and the applied flux was constant at 35 LMH. TMP values were measured overtime and normalized to the initial TMP.
Figure 3 shows the normalized TMP profiles. Faster TMP rise was observed at lower CFV and both operation with and without spacer show similar profiles. The delay of TMP rise caused by spacer was quantified by measuring the time needed for the TMP to increase by 10 % (Table 1). The effect of spacer was higher at higher CFV where the percentage of the delay was 21.21 % and 42.87 % at 0.10 m/s and 0.17 m/s respectively. An interesting phenomenon was observed during the earlier TMP rise (0-3 days) where change in CFV gives little effect on TMP profiles. Similar phenomenon was observed for operation with and without spacer. A possible explanation for this phenomenon is that during this period bacterial attachment was dominant and therefore operation at constant flux gives similar initial TMP rise. Previous studies (Chong et al., 2008) have shown previously that membrane biofouling is a flux driven process where higher flux increases the TMP rise. However, their study did not include spacers and did not focus on initial TMP rise.
Table 1. The delay of biofouling rate caused by spacer at different CFV
The effect of different salt concentrations was also investigated. In this experiment the flux and CFV were fixed at 35 LMH and 0.17 m/s respectively. Figure 4 shows the normalized TMP profile of three different NaCl concentrations in the feed solution. When the feed channel spacer was absent it was very obvious that faster TMP rise was observed at higher salt concentration. This suggests that the effect of concentration polarization (CP) increases with the salt concentration and confirms the presence of the biofilm enhanced osmotic pressure (BEOP) effect (Herzberg and Elimelech, 2007; Chong et al., 2008). This phenomenon however, was less obvious when the spacer was present on the membrane. The spacer appears to provide flow eddies thus reducing the effect of CP and to be useful to prevent biofouling on the membrane which was indicated with slower TMP rise. The spacer gives bigger effect at higher salt concentration where the time to reach 10 % TMP rise was delayed by 30 % at 100 mg/L and 2000 mg/L NaCl, and 95.7 % at 4000 mg/L (Table 2).
4.2 Biofilm development in spacer-filled RO membrane channel The development of biofilm in spacer-filled channel was observed via microscopic and microscopic method. Macroscopic images are to show overall uniformity of biofilm distribution, while the microscopic images are able to show a more detailed biofilm patterns. All of the images in this study were taken from separate experiments as the samples were unable to be reused after analysis, however all the conditions for the experiments were maintained the same.
Figure 5 shows the macroscopic images of biofilm development. The biofilm sample on the membranes and spacers were stained with 5-cyano-2,3-ditolyl tetrazolium chloride (CTC) dye. CTC stains bacteria with respiration activity and stained cells appear in red colour. Analysis was done after 0, 3, 6, and 10 days, the condition was 35 LMH flux, 0.17 m/s CFV, and 4000 mg/L NaCl concentration. Longer experiment duration gives thicker and denser biofilm, which can be seen from higher red colour intensity. The biofilms have also shown overall uniformity across the membrane area where similar patterns were observed among each spacer squares.
Figure 5. Macroscopic images of biofilm development on membranes and spacers. (A) 0-day, (B) 3-day, (C) 6-day, (D) 10-day. Biofilms stained with CTC dye and images taken with SONY NEX-5 digital camera.
Confocal laser scanning microscope (CLSM) provides a more detailed analysis of biofilm development (Figure 6). Based on the images, it appears that biofilm was initiated on the membrane; it later covered more areas and started to appear on the spacer. Areas behind the attached filaments of the spacer fiber seem to be suitable for the initial bacterial attachments rather than the centre of the spacer. Biofilm build-up observed on areas under the detached filaments was caused by higher shear due to accelerated CFV. Our experiments confirmed that biofouling in RO is a flux driven process. A lower TMP rise was observed at lower flux, which means slower biofouling rate. This is also supported with the biofilm coverage data where less coverage was observed at lower flux.
From the findings above, several conclusions can be drawn. The hydrodynamic condition of the flow is affecting the biofouling process. Cross flow velocity (CFV) is an important parameter and lower fouling can be achieved at higher CFV. Having feed channel spacers on the membrane is advantageous as it provides a more well-mixed flow, reduces concentration polarization and reduces TMP increase. Biofilm enhanced osmotic pressure (BEOP) was another phenomenon observed in this study. Due to the BEOP effect, a faster TMP rise was achieved at higher salinity. However, with the presence of the spacer the BEOP effect was reduced significantly.
From our microscopic analysis of biofilm shows that initial bacterial deposition and biofilm development was started on the membrane especially on areas behind the attached spacer filaments. Biofilm develops over time to cover more areas and starts to grow on the spacer at the later stages. Imposed flux also influences the biofilm development where lower biofouling is achieved at lower flux.
Baker, J. S. and Dudley, L. Y. (1998), “Biofouling in membrane systems – a review”, Desalination, Vol. 118, No. 1-3, pp. 81-90.
Chong, T. H., Wong, F. S. and Fane, A. G. (2008), “The effect of imposed flux on biofouling in reverse osmosis: Role of concentration polarisation and biofilm enhanced osmotic pressure phenomena”, Journal of Membrane Science, Vol. 325, No. 2, pp. 840-850.
Flemming, H. C. (1997), “Reverse osmosis membrane biofouling”, Experimental Thermal and Fluid Science, Vol. 14, No. 4, pp. 382-391.
Herzberg, M. and Elimelech, M. (2007), “Biofouling of reverse osmosis membranes: Role of biofilm-enhanced osmotic pressure”, Journal of Membrane Science, Vol. 295, No. 1-2, pp.
Koltuniewicz, A. B., Field, R. W. and Arnot, T. C. (1995), “Cross-flow and dead-end microfiltration of oily-water emulsion. Part I: Experimental study and analysis of flux decline”, Journal of Membrane Science, Vol. 102, No. 1-3, pp. 193-207.
Suwarno, S. R., Puspitasari, V. L., Chong, T. H., Fane, A. G., Chen, X., Rice, S. A., Mcdougald, D. and Cohen, Y. (2010) “The hydrodynamic effect on biofouling in reverse osmosis membrane processes”, IWA International Young Water Professionals Conference, Sydney,
Vrouwenvelder, J. S., Hinrichs, C., Van Der Meer, W. G., Van Loosdrecht, M. C. and Kruithof, J. C. (2009b), “Pressure drop increase by biofilm accumulation in spiral wound RO and NF membrane systems: role of substrate concentration, flow velocity, substrate load and flow direction”, Biofouling, Vol. 25, No. 6, pp. 543-555.
Wangnick (2005), 2004 Worldwide Desalting Plants Directory, Global Water Intelligence, Oxford, England.
Experimental Methods in Wastewater Treatment – M.C.M. van Loosdrecht, J. Keller, P.H. Nielsen, C.M. Lopez-Vazquez and D. Brdjanovic
Publication Date: Feb 2014 – ISBN – 9781780404745