The membranes are based on a three-layer composite structure that is manufactured in several production stages.
The membranes are based on a three-layer composite structure that is manufactured in several production stages.

An efficient RO process relies upon stability and durability for operational reliability. One of the targets in the Lewabrane development program was to improve the membrane polymerization chemistry to provide a more robust separation performance over the lifetime of the RO membrane element. This case history illustrates the operational results over a one year timeline upon installation of new Lewabrane RO membrane elements.

Since the introduction of the new Lewabrane RO element brand during early 2012, there have been several thousand RO elements installed in more than 24 countries for both industrial and potable water application. Most users agree that process stability and durability are the key parameters of an excellent RO process, and that the technology and quality of the RO membrane elements are a fundamental factor in providing operational reliability.

Improved RO membrane chemistry

The development focus for this new Lewabrane RO membrane chemistry resulted in a polymerization reaction that provides a higher degree of crosslinkage than the routine commercial RO elements available in the market. One of the key targets during the development of the new membrane chemistry was to improve the polymerization reaction which forms the polyamide layer.

The new state-of-the-art production line in Germany employs a process that increases this polymerization degree. This is achieved by the usage of high purity chemicals, and optimized process conditions, which reduce secondary reactions often causing reduced quality and yield of the polymerization process.

In RO membrane preparation, the polyamide layer, which is responsible for the separation, has a thickness of around 0.1 µm. The high polymerization degree is essential for the chemical and mechanical durability of the membrane. Further, the high degree of polymerization also results in a lower negative surface charge of the membrane. This improves the fouling tendency during the treatment of difficult surface water supplies [1,2,3] and the salt rejection.

The higher salt rejection is achieved because the rejection is not caused by repulsion effects of the negative charged surface and the dissolved anions, but rather by apparent pore size and diffusion. Therefore, the concentration polarization effects of ions on the membrane surface, which decrease the rejection of strong charged membranes, influences the salt rejection less, and helps to maintain high rejection [4].

The improved Lewabrane salt rejection performance was proven by laboratory tests with mixed salt solutions, including boron, nitrate and silica, using Lewabrane RO elements (HR types; 15.5 bar test pressure) and known industry countertypes. In these tests, it could be shown that the salt rejection was higher than competitive RO membranes of similar chemistry. The largest observed difference was the deviation was the salt passage coefficient for Nitrate, which was one half of the SP (salt passage) value compared to the best of the competitive RO membrane elements in the same test.

Also, the Boron rejection at different pH is a strong indication that the salt rejection is caused by pore size or diffusion, and less influenced by anionic repulsion.

At a pH value of 9.2, the boric acid salt starts to dissociate, so that the rejection in the case of a strong negative surface charge, increases dramatically. At lower pH values, the rejection is much lower compared to an RO membrane with ambient surface charge, but with higher cross-linkage (as in the case with Lewabrane RO membrane).

The superior rejection performance for Boron and Nitrate could also be demonstrated with the Low Pressure (10.3 bar test pressure; also known as Low Energy) membrane type, which has an even thinner polyamide layer to increase the performance of the membrane at lower pressure.

Case history Lewabrane in Spain

One RO plant where this theoretical approach was confirmed in actual practice on a real industrial feed water was for an RO plant in Spain, operating for more than one year. The installation is a two stage system with a capacity of 18 m3/h in the textile industry, and located in Blanes (Girona), Spain. The feed water is brackish well with a TDS between 350-400 ppm. The Lewabrane elements with 37m2 and an average salt rejection of 99.7% were installed in August, 2012.

The plant has two RO lines which are alternating. In the other line, standard test pressure RO elements from another manufacturer were installed within two weeks of the Lewabrane installation, allowing a direct performance comparison to a commonly used RO membrane product. This RO plant was originally installed in 1998 as a pretreatment step in front of a standard two bed ion exchange.

As expected, the initial flux from the high cross-linked membrane was slightly lower than the competitive one, but also a constant higher rejection could be observed. This is not unexpected as the higher crosslinkage membrane is slightly less permeable than the competitive membrane, but also has slightly better salt passage characteristics. This is supported by the below graphs for flow and rejection between the two side by side RO plants over the past year.

The higher rejection has an significant influence on performance of the downstream ion exchange process as the chemical regeneration cycles could be extended. This leads to a lower consumption of chemicals for the regeneration in relation to the net produced water from the plant.

The actual data show the challenge of working with real world data, and the difficulty in making direct comparisons. Unfortunately, the performance of the two lines decreased in the first hundreds days of service, drastically, or at least more than would be considered a routine productivity decline. The plant was designed to produce 18 m3/h at 11.3 bar, but after 100 days in service, 14.3 bar was required to produce 15 m3/h.

Autopsy – scope and findings

As a result of this poor productivity performance, the customer was requested to send the lead element and last Lewabrane element from the plant to Bitterfeld (Germany) for autopsy. This autopsy included weight measurement, performance testing at standard test conditions, and surface analysis of the membrane by destructive opening of the element.

The autopsy revealed that the weight of the lead element was already 20% higher compare to a new element, and the standard tests showed a significant decrease of permeate flow with an increase of the pressure drop. The last element showed constant performance compared to the initial values of the new elements. In both cases, the salt rejection of the used RO element was as high as the as new, original performance of the RO elements.

Also, the autopsy inspection on the outside of the lead element showed that the feed spacer had started to move, which is an indication that the feed spacer is clogged. The visual inspection after opening the element showed heavy biofouling in this area, which also explained the increase of the pressure drop. In total, 3kg of biomass were collected from this element.

The results (fouled elements) were quite unexpected since the plant continuously adds biocides in the feed water to prevent biological fouling. It was recommended to switch from a continuous dosing to a shock dosing to avoid the genetic adaptation of the bacteria to the specific biocide.

As the RO elements which were sent to Bitterfeld for autopsy were replace by new RO elements, the permeate flux of the Lewabrane trains was slightly higher than the competitive train. Another important point was to determine if the cleaning of the RO train after 230 days in operation lead to a return of the initial water flux of the RO train. This data show that even this heavy fouled RO train could be successfully cleaned.

LewaPlus performance modeling

The performance projection of the plant was prepared using the LewaPlus design software, which has the option to model the plant performance under different conditions. The design program projected a water flux of 18m3/h and a pressure 15.7 bar, which is reasonably close to the observed reality after cleaning. The observed flux was in the range of 16.4- 17.5m3/h at a pressure in the range of 14.7-16.0 bar. Therefore, the promised performance of the plant could be held even though the RO plant experienced a severe biofouling.

After one year of successful performance, it can now be said that the system builder (OEM) of this RO plant was so convinced of the Lewabrane membrane reliability, technical performance, and vendor support, that six other plants were subsequently installed with Lewabrane elements.

Summary

Since the early 2012 Lewabrane product launch, this brand of RO elements are now installed in over 24 countries, worldwide. The product portfolio now covers a wide range of brackish water RO membranes, including Low Pressure (Low Energy) and Fouling Resistant membrane types. Additionally, since market entry, the Lewabrane brand as received the approval of the NSF for drinking water applications. And, looking to the future, Lanxess expects to launch RO membrane elements for seawater desalination by mid-2014.