Using Liqui-Cel® Membrane Contactors to Solve Resistivity Problems in a High Purity Water Loop

News from 3M Separation and Purification Sciences Division

09/09/2008

Pureflow, Inc. is a Southeastern-based expert in designing complete water treatment systems as well as providing value-added solutions for fixing operational issues in existing systems. Pureflow teamed up with Membrana to help solve an operational issue at one of their customer’s facilities. The existing high purity system consisted of a pre-treatment skid that supplied the feed stream for a 5 gpm reverse osmosis (RO) system. Product water from this RO system flowed to a 5 gpm electrodeionization (EDI) unit. Effluent from the EDI system went to a storage tank before moving to an ultra-violet light, and then to the resin-based deionization tank. The effluent then passed through final filtration before entering the facility’s water loop. This system design is shown schematically in Figure 1. Water fed to the facility was expected to be high quality ultra pure water with consistent effluent properties. One of the desired parameters of the product water was to have a resistivity reading of 18.0 M-ohm/cm or better. Measurements taken of the post-EDI water stream, however, showed resistivity ranging from a low of 1.0 M-ohm/cm to only 13.7 M-ohm/cm. Additionally, the readings were very inconsistent and erratic. Figure 2 shows actual readings taken over a 10 week period of time. Further measurements taken upstream of the EDI equipment showed an abundance of carbon dioxide in the water. The amount of CO2 in the water fluctuated from 8 to 20 ppm. It is well documented in technical papers that the presence of CO2 and bicarbonate (HCO3 -) reduces the performance of EDI systems. The presence of CO2 and HCO3 - in essence occupies excess “capacity” in the EDI unit leaving little to no capacity to remove silica and other components from the water. The result is that while the EDI unit may be functioning properly, chemistry within the unit dictates that compounds such as silica will “breakthrough” into the effluent. These contaminants contribute to declines in the water quality as measured by resistivity. As the CO2 in the EDI feed increases, more capacity in the EDI unit will be lost causing increased drops in water quality. In order to improve the quality and consistency of the water at this facility, it was clear that additional EDI capacity was required or that the dissolved CO2 needed to be reduced before the EDI.