Ageing of drinking water production membranes

During cleaning steps in water drinking production, ultrafiltration membranes are mechanically and chemically stressed. This can result in membrane degradations and failures: separation properties of the membrane can change causing an increase in the permeation of large molecules through the membrane and/ or cracking of the membrane layer takes place leading to leakage of feed stream into the permeate stream. Failure of the filtration system requires a shut down for membrane replacement, which directly affects the production rate of the process. So, this raises the problem of the necessary optimization of the cleaning procedures which have to take into account the best cleaning efficiency as well as the less detrimental procedures for the membranes. That is why ageing is becoming the new bone of contention between end users, membrane manufacturers and chemical product suppliers. Nevertheless, in spite of numerous scientific researches realized on the subject, several gaps can be underlined: - First, the commonly used parameter “concentration × time of exposure” (c × t) is used to make ageing experiments easier and to shorten the experimental times involved by the ageing mechanisms. This way of thinking would suppose that strong concentrations for short times of exposure could be considered equivalent to low concentrations for long times of exposure. This approach remains questionable and the major issue of the current membrane ageing researches would be to define consistent and representative ageing protocols.- Secondly, the studies performed until now are mainly focused on hypochlorite, while the latter is not universally applied and the use of chlorine-based reagents is expected to be restricted in the next years. Some new membranes are not chlorine tolerant and the use of chlorine was shown to generate chlorinated organics with detrimental effects on health and environmental impacts. So, alternative detergents could significantly be used in the future. However very few work have been focused on the membrane ageing resulting from the use of formulated detergents.In this paper, some static ageing results have been detailed. Polysulfone hollow fibers membranes were used to simulate industrial ageing in static conditions by NaOCl and by four commercial detergents: P3-Ultrasil 75 (acidic), P3-Ultrasil 110 (alkaline), P3-Ultrasil 67/P3-Ultrasil 69New (Enzymatic/buffer) and P3-Oxysan ZS (free chlorine oxidant). The macroscopic changes were monitored by permeability measurement and mechanical strength tests coupled with a microscopic characterization by ATR-FTIR (to identify the functional groups affected by ageing) and SEM (to visualize qualitatively the surface modifications). The aim is to achieve a better understanding of changes in membrane properties, or at least an identification of the combination of cleaning parameters (concentration, time and temperature) which should be avoid in cleaning procedures.-First, preliminary experiments were carried out (only with NaOCl) for a same value of the c × t parameter, but with a simultaneous variation of concentration and time of exposure. For a same value of the c × t parameter, it was shown that the macroscopic as well as the microscopic ageing were not equivalent for a high concentration/short time of exposure and low concentration/long time of exposure. For the ultrafiltration membranes and for the experimental conditions tested, the permeability deterioration in function of c × t parameter has been shown to be not homogeneous according to the temperature. -Secondly, further experiments were realized using an approach based on experimental designs in order to achieve a relevant and predictive ageing pattern without using an accelerated ageing protocol. For that, a wide concentrations range around the industrial concentrations commonly used (200 ppm for NaOCl, 0.3% w/v for P3-Ultrasil 75, 0.7%w/v for P3-Ultrasil 110, 0.3%/0.8% w/v for P3-Ultrasil 67 / 69New, 0.1% w/v for P3-Oxysan ZS), several temperatures (from 5°C to 40°C, except for P3-Oxysan ZS, only tested at 20°C) and several soaking times (from 7 days to 180 days) were tested. For all the tested detergents, permeability and elongation at break values obtained with the statistical model are simulated with an error included in the measurement accuracy. So, the pattern established for macroscopic membrane deteriorations can be considered satisfying and reliable to simulate membrane ageing in function of the tested parameters. These results show high deteriorations involved by NaOCl (increase of permeability, decrease of strength at break, modifications of surface properties, apparition of crackles, and disappearance of specific peaks from ATR-FTIR spectra). The results at 20°C of fibers aged in NaOCl and P3-Oxysan ZS at their respective industrial concentrations show that degradations of membrane are far more detrimental with NaOCl than with P3-Oxysan ZS. Concerning the other detergents tested (and considering the long ageing duration applied in this study) it could be concluded that the ECOLAB formulated detergents are quite harmless for the membrane for a time corresponding to the effective cumulated industrial time of contact. For the first time in the chemical membranes ageing researches, designs of experiments were used to organize at best the experiments in order to achieve a relevant establishment of an ageing pattern without using an accelerated ageing. Results showed that this scientific approach could provide a satisfying and reliable pattern established to simulate membrane ageing in function of the chemical parameters. Then, results show that NaOCl causes the worst ageing deteriorations on membranes after extended exposure: changes in membrane chemical groups, hydraulic performances, mechanical properties and physical structure. Only static ageing is presented in this abstract. To complete this work, dynamic ageing experiments have also been realized (in the same conditions of concentration, time and temperature as the ones studied during static ageing). The aim is to be able to compare the ageing due to the chemical contact and the ageing due to the mechanical stress (pressure, scratches, abrasions…) induced by process parameters such as cycling during operation or extreme pressure