Selection of pre-treatment technology for desalination in seawater reverse osmosis (SWRO) applications tends to polarise opinion. Conventional technology represents the status quo while membrane filtration adopts the role of the challenger, with a rapidly developing track record. Two issues limit the wider adoption of membrane pre-treatment. Firstly it is considered expensive in terms of capital cost, though it does provide operational cost savings. Secondly there are misgivings about whether membrane filtration alone provides sufficient pre-treatment, since preparing a feed suitable for RO requires more than the removal of fine particulates.
This series of articles has examined the issues involved in the selection of pre-treatment technology for SWRO, and has considered the case for conventional and membrane pre-treatment options. The previous articles looked at the issues of the treated water quality for the RO feed, integrity and disinfection, and cost and sustainability. In terms of treated water quality, the barrier process of membrane filtration using ultrafiltration (UF) or microfiltration (MF) provides improved particle removal and less physical fouling. Also, UF and MF membranes provide better integrity than RO due to a more consistent membrane separating layer, and a module construction method which eliminates leak paths. In addition, UF and MF products are designed to be testable and repairable, unlike RO elements. Finally, with regard to cost and sustainability, membrane filtration has become competitive with conventional pre-treatment in many cases, and can have lower operating cost and less environmental impact.
The fourth and final article in the series will examine membrane filtration markets in SWRO pre-treatment, and the experiences of applying this technology in pilots and main plants. The key questions that the article will address are what is the uptake of membrane pre-treatment technology, how well does it work, and does it provide a robust solution?
Since the late 1990s, desalination markets have grown strongly, with seawater reverse osmosis (SWRO) becoming established as the main technology employed. According to the Water Desalination Report , compound growth between 1997 and 2008 has been 17%, exceeding the expectations of the RO industry.
Likewise, growth rates for UF and MF products in water treatment have also been exceptionally strong in the last few years at around 20% . These high growth rates are predicted to continue due to opportunities in the drinking water market, the emergence of wastewater reuse, and the adoption of membrane filtration as a commonplace pre-treatment technology for the rapidly expanding SWRO market.
Though drinking water became established as a key application for UF/MF a decade ago, the use of UF/MF for SWRO pre-treatment was the exception rather than the rule up until as recently as 2006. In early 2007, NWRI carried out a survey of UF and MF installations for the water and wastewater market . The survey relied on data from the membrane companies, including all of the significant international players, but did not include activity in China. Of the total of 13,000 mld of installed cumulative membrane filtration capacity, as at December 2006, only 3% had been supplied to SWRO pre-treatment, as illustrated in Figure 1.
However, a more recent survey published in GWI illustrates the rapid uptake of membrane pre-treatment since then shown in Figure 2, based on data from CH2M Hill . Thus whereas UF/MF formed just 4% of contracted capacity in 2005, GWI estimated that it formed 30% of capacity under bid in 2008.
A more recent survey by CH2M Hill  has shown that the uptake of UF/MF has continued to increase with cumulative installed capacity doubling between 2008 and 2009 from 1,100 mld to 2,200 mld.
Pilot plant experience
A large number of pilot studies have been published on the use of membrane pre-treatment prior to SWRO. Many early studies focused on how the UF or MF products themselves performed in the application, but more recently, there has been more focus on using an integrated membrane system with both UF/MF and RO stages to evaluate the performance of the complete system. Most studies have found that the UF/MF performs well [ and ], but flux levels may be somewhat lower than for a surface water feed of similar quality due to the possibility of occasional contamination by algal cells or other marine organisms.
None of the evaluations question UF/MF performance in terms of the main filtered water quality determinants, i.e. particulates and microbial cells. However, though these parameters are key for establishing fouling and bio-fouling propensity, other factors also have an influence, as discussed in the previous articles. For example, bio-fouling is strongly influenced by dissolved organic carbon, especially if present as Assimilable Organic Carbon (AOC), since it acts as a nutrient source for bio-fouling, and can cause problems even though microbial activity itself is low. Therefore, operating an RO stage after the UF/MF is vital to characterise the stability of the complete system.
Trials at Santa Cruz, California  followed a typical evaluation pattern in that for 70% of the time, the feed was of a consistent high quality with low turbidity (< 5 NTU) and low organics (< 1.3 ppm TOC). During the other 30% of the time, quality could be poor due to storms, which increased turbidity to 40-50 NTU, or due to algal bloom/red tide events, which could take TOC to 10 or even 15ppm.
Coagulant use is often considered for pilot operation since it allows UF/MF flux to be increased and cleaning frequency to be reduced. Also, it is much more likely that performance can be maintained during a poor feed quality episode. In addition to improving stability of the UF/MF operation, it improves dissolved organic removal, and thereby reduces the chance of bio-fouling. Many studies have concluded that coagulant use is beneficial, both improving UF/MF performance , and reducing organics arriving at the RO, and hence bio-fouling .
At Santa Cruz, coagulant dosing was found to provide a beneficial effect to the stability of UF operation, and in appearing to reduce bio-fouling during episodes of algal bloom. However, it was felt that UF possibly increased chances of shearing algal cells, and hence Dissolved Air Flotation (DAF) was being considered as a membrane pre-treatment in the next phase of piloting.
A downside of coagulant dosing is that the RO membrane can suffer from coagulant fouling [eg: 10, 11]. This may be a result of weak floc formation, especially during good feed water quality episodes, or due to shear from pumps, or due to disruption of the floc structure caused by too high a pressure drop in the membrane system design. It is possible that floc breakdown problems could be reduced by longer contact times.
When membrane pre-treatment was first considered for the SWRO pre-treatment application, the hope was that it would allow chemical free or low chemical operation. In reality, this has not been possible, since the UF/MF membranes require either feed dosing chemicals or cleaning chemicals, and sometimes both. Two types of UF/MF operation can be used. One option is to dose the feed with coagulant, operate with medium to high flux, and use a low chemical cleaning frequency. Alternatively, a low flux operation is used on an uncoagulated feed, with a relatively high cleaning frequency. Cleaning for a low fouling situation may just consist of frequent Maintenance Wash (MW) or Chemical Enhance Backwash (CEB), but if fouling is more pronounced, Clean In Place (CIP) will be required on a fairly regular basis.
A survey of pilot and main plant operations  has shown that coagulation is used slightly more frequently than a no coagulant operation, as illustrated in Figure 3. However, the choice is not only influenced by the perceived benefits for stable UF/MF operation, but also by concerns about chemical waste disposal and the ability to obtain chemical use permits in some locations.
It appears that the polyethersulphone (PES) membranes are more likely to realise the benefits of coagulant dosing than polyvinylidene difluoride (PVDF). As a result, the use of coagulant dosing in PES pilots and main plant designs is more widespread than for PVDF. Thus, nearly all PES plants use coagulant, whereas just under half of PVDF plants do so. Reviewing pilot data indicates that PES gains a greater permeability advantage by using coagulant, whereas PVDF, normally operates at lower flux and permeability, but with reasonable stability whether or not coagulant is used.
One of the earliest large scale membrane pre-treatment installations was carried out at Kindasa, situated in Jeddah, Saudi Arabia . Extensive piloting had been carried out with a pressure driven inside feed (PDI) system, utilising PES membranes. Several different flowsheet options were investigated, both with and without coagulant, all of which provided good feed quality to the downstream RO. For direct feeds, in-line coagulation provided excellent stable performance of the UF, but there were occasional signs of partial carry over of the ferric floc. The uncoagulated feed gave less stable UF performance, with more frequent cleaning.
For the actual plant, it was decided to use a rapid rate granular media filter to provide pre-treatment to the UF, and avoid coagulation. During the first six months of operation, excellent RO performance was achieved, with stable permeability, low salt passage, and no requirement for RO cleaning.
Following this period, a bio-fouling incident occurred , which was attributed to the effect of chlorine use, both as an occasional shock dose to control growth in the intake, and in the Chemical Enhanced Backwash (CEB) for the UF system. Accordingly, the shock dose was changed to a non oxidising biocide. Also, CEB frequency was reduced from four times a day to twice a day, and post CEB rinse up efficiency was improved. These actions controlled the bio-fouling and returned the plant to stable performance.
Extensive trials have been carried out on the SWRO plant at Adelaide using different membranes and different operating design procedures. Initial piloting with a PDI system, again utilising PES membranes, showed that a high stable flux could be achieved. However, a relatively frequent CEB was essential, with three CEB's per day providing much better stability than two a day . Filtrate quality was excellent in terms of particulates and microbial cells, and importantly eliminated the variability of conventional treatment that occurred during the ripening period. However, some Fe fouling occurred due to floc carryover, and occasional specialised CIP procedures were required. It was felt that general feed water quality was too good, so that the coagulant dose concentration might be too high, or the flocs created were weak and susceptible to shear.
Subsequent trials were carried out with a submerged system, utilising a coarse UF PVDF membrane , and this system has been used in the main plant. Trials were conducted both with and without coagulant. Excellent treated water quality was obtained with both options, but fouling rates were nearly twice as high with the coagulated feed. Both options gave stable RO operation. It has been decided that for the main plant coagulation will be provided but only utlised under poor feed water quality conditions.
Following further observation of Fe fouling , variable rate in-line dosing is now being proposed more generally  on other UF pre-treatment plants to ensure that coagulant is only used when absolutely required, and that a minimal level of coagulant is used consistent with maintaining stable permeability on the UF stage.
The two previous examples have shown that although coagulation can be desirable, both in terms of ensuring stable UF performance and in removing dissolved organics, it can also cause problems to the RO due to floc carryover and Fe fouling. Recent plant designs have looked at the use of Dissolved Air Flotation (DAF) as a pre-treatment to UF in order to gain the benefits of coagulation without suffering the downsides.
One of the first main plants to use this option in a SWRO plant is located at Escombreras in Spain . Here a system based on a submerged UF PVDF membrane takes the output of a DAF unit. Although the system has not been in operation long, initial indications are encouraging, with excellent stable UF performance with good levels of permeability, and apparently stable performance of the RO. DAF pre-treatment to UF is also being used for one of the world's largest SWRO plants at Hamriyah in the UAE, but this plant has not yet started up.
• In the past 10 years both desalination and membrane filtration (UF and MF) markets have grown strongly.
• The application of membrane filtration in seawater reverse osmosis pre-treatment is relatively recent, and it is only since 2006 that there has been a significant uptake in this technology.
• Initial results have shown that UF/MF gives excellent feed quality to the RO in terms of particulate and microbial cell removal, and that UF/MF can provide a cost effective design with stable performance.
• The question of whether to use coagulation is still open.
• Coagulation improves the flux and stability of the UF/MF and reduces the dissolved organics and therefore bio-fouling potential.
• The coagulant itself may cause ferric fouling especially during periods of good feed quality.
• The current trend is to use a variable coagulation regime, reducing the concentration from a minimal concentration during poor feed quality episodes to zero when quality is good.
• Another alternative which appears promising is the use of Dissolved Air Flotation (DAF) as a UF/MF pre-treatment.
1 SWRO market exceeds Forecast, T Pankratz, Water Desalination Report, Vol 44, no. 22, p1, 21/06/08.
2 Global Water Market 2008: Opportunities in Scarcity and Environmental Regulation, Global Water Intelligence, October 2007.
3 A Global Perspective of Low Pressure Membranes, D Furukawa, NWRI Final Project Report, March 2008.
4 UF in pre-treatment for seawater desalination: annual capacity by supplier, Global Water Intelligence Monthly, p7, Volume 9, Issue 7, July 2008.
5 MF/UF Pre-treatment in Seawater Desalination: Applications and Trends, Huehmer, R P, IDA Conf Proceedings, Dubai, DB09-253 (2009).
6 Membrane Pre-treatment in SWRO: Global Applications and Membrane Type Considerations, Dietrich, J, IDA Conf Proceedings, Dubai, DB09-010 (2009).
7 Results from Nine Investigations Assessing Pacific Ocean Seawater desalination in Santa Cruz California, Desormeaux, E D, et al, IDA Conf Proceedings, Dubai, DB09-291 (2009).
8 Yang Hyun-Jin and Kim Han-Seung, Effect of coagulation on MF/UF for removal of particles as a pre-treatment in SWRO, Desalination 247 (1–3) (2009), pp. 45–52.
9 P-J Remize, J-F Laroche, J Leparc and J-C Schrotter, A pilot scale comparison of granular media filtration and low pressure membrane filtration for seawater pre-treatment, Desalination & Water Treatment 5 (2009), pp. 6–11. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (4)
10 Membrane MF & UF Pre-treatment Design & Operational Experience from Three Seawater RO Plants, Ben Boudinar, M, et al, IDA Conf Proceedings, Dubai, DB09-230 (2009).
11 Long Term Operating Experience of Seaguard UF as Pre-treatment to SWRO in the Mediterranean Region, Knops, F, te Lintelo, R, EDS Conf Proceedings, Baden Baden, (2009).
12 First Successful Operation of SWRO Plants in Saudi Arabia with UF Pre-treatment, Amir Basha, S K, et al, R P, IDA Conf Proceedings, Gran Canaria, MP07-073 (2007).
13 Adelaide Desalination Project Pilot Experience, Blaikie, M, Pelekani, C, Water: Desalination & Membrane Technology, June 2010.
14 Influence of Chemical Treatment on Membrane Pre-treatment for SWRO: Adelaide Case Study, Acciona, Proceedings of the Ozwater Conference, Brisbane, March 2010.
15 UF used as Pre-treatment for SWRO Desalination: Dynamic Coagulant Control and Optimization, Futselaar, H, et al, IDA Conf Proceedings, Dubai, DB09-093 (2009).
16 Escombreras: Large-Scale Spanish Seawater Desalination Utilizing Innovative Membrane Pre-treatment Technology, Vonghia, E, Hagmeyer, G, Fernandez, M G, Sanchez, J, IDA Conf Proceedings, Dubai, DB09-157 (2009).