Water search 169 (2020) 115212 Contents lists available at ScienceDirect ey Water Research journal homepage: www. evier.com/locate/watre: ELSEVIER Membrane ageing in full-scale water treatment plants Shona Robinson, Pierre R. Bérubé" Deparment of Cv Engineering, 2002 ~ 6250 Appi Since Lane Univesity of Bish Columbia Vancurer BC Cano, VET 124 te ARTICLE INFO ABSTRACT rd Received 12 August 2019, Received in reed form ecepted 16 etber 2019 Fale onine 8 Oereer 2019, ‘Membrane filtration i a rapidly expanding choice for drinking water treatment. Unfortunately, there is limited data on long-term changes in the membranes’ performance as they age. The present research investigated changes in performance factors as well as chemical characteristics for hollow-fbee ulta- ‘tration membranes that ranged in age from 8 full-scale drinking water treatment plants. Membranes ‘were harvested by plant operators regularly and analyzed using standardized laboratory tests. APPrOX- imately half of the membranes were a new PVDF-based chemistry. These were abserved to have insignificant changes in performance factors and chemical characteristice since their beginning of ele ‘operation. However, because these membranes were newer, only data for the first 5 years of operation Memberoe aging ‘was available. The other half ofthe membranes, with an aider PYDF-based chemistry, were observed to Ueaieaton have stable behaviour until approximately 5 years of operation; ater this time, performance factors and Ferformanc factors Pulsed ageing chemical characteristics ofthe membranes began to change significant. or these membranes, the clean ‘water resistance and fouling rate increase after 5 years of operation The mechanical properties ofthese membranes also deteriorated ater 5 years of operation, suggesting that thei susceptibility to breach is higher after prolonged use. Tse changes in performance factors paralleled, and were possibly caused by, the removal of hydrophilic additives from the membrane material. Clean water resistance was identiied asa good benchmark forall the parameters studied finding that is useful for water treatment facies in quickly assessing the status oftheir membranes Finally although cumulative exposure dose (Ce) was not used as a metic of membrane age, we observed that when higher doses of hypochlorite ‘were applied, all metrics changed faster than expected based only on years of operation. Therefore, limiting the magnitude of the cumulative hypochlorite dose is essential in managing membrane dete riortion, This research iluminates the knowledge gap between benci-scale ageing stales and opera- tional water treatment plants {© 2019 The Authors. Published by Elsevier Lx This san open acess article under the CC BY-NC-ND Ticense (hep |ereativecommons.orgiicenses[by-ne-nd40)). 1. Introduction 11, Background & research gap Ultrafiltration membrane filtration, hereafter simply referred to {as membrane filtration is a popular technology for drinking water treatment. However, there is limited data surrounding long.term changes that can occur as membranes age. Ageing results from operational stresses induced on membranes, during repeated permeation and cleaning. Eventually membrane replacement is required (Robinson et al, 2016). An improved understanding of ‘membrane ageing in full-scale water treatment plants would allow for more predictable long-term operation and capital planning. * Conesponding author Erma addres: erubeocvLaece (PR Bleue) np: doiorgyi01016j. waues2018215212 ‘Membrane fouling results ftom the accumulation of retained ‘material, both inside the pores and on the surface of membranes ‘during filtration, and increases the resistance to permeate flow. Various hydraulic cleaning strategies are employed to minimize andjor reverse membrane fouling (Crittenden, 2005). A common strategy employed to periodically remove foulants is backwashing, ‘of membranes with clean water. Ths allows some of the foulants within membrane pores and on the membrane surface to be transported away ftom the membrane. Sheat is typically intro- duced via air sparging or cross-flow during backwashing to tenhance the transport of material away from the membrane sur- face. However, after repeated filtration and backwash cycles, chemical cleaning is required to remove material which has adsorbed onto the membrane and which cannot be removed hy- 100%, most 1000As),as per ASTM D3822. Eight replicates were performed per harvesting event (4 from each test module). “The maximum stress and Young's modulus were plotted as relevant toan individual fibre, rather than per cross-sectional area. Example stress-strain curves are presented in the supplementary material. 24, Data analysis Linear regression was used to determine the relationship be- tween performance factors or characteristics and membrane ‘operational age, expressed in years. All 100A membranes had been operational for ess than 5 years, or half oftheir recommended life, so were analyzed together. The 10008 membranes ranged from 3 t0 83 years, and were analyzed as newer 10008 membranes (<3 years of operation) and older 10008 membranes (>5 years of operation). The p-values presented in this manuscript relate to the ‘null hypothesis that the slope of each trend equals zero for a linear relationship between a measured variable and the membrane age. Where p-values suggested a linear relationship between variables, the slope and 95% confidence interval of the trend-line are pre- sented. A similar statistical approach was performed to link the surface HA content to the fouling rate, testing for a non-zero slope. Detailed statistical analysis forthe different conditions considered {s presented in the supplementary material. Both the extent of cleaning during a given period and the cleaning rate correlated closely with the fouling rate for 1000A and 10008 membranes (see supplementary material). This was ex- pected because cleaning is greatly impacted by the amount of foulant present on the membrane prior to cleaning. Because the extent and rate of cleaning do not provide additional insight into ‘our understanding of membrane ageing, it not discussed further. Details ofthe analysis relating to the extent and rate of cleaning are provided in the supplementary material. 3. Results & discussion 41, Effect of ageing on performance factors Four performance factors, as previously described, were assessed for fibres harvested from full-scale membrane facilites clean membrane resistance, fouling rate, cleaning rate, suscepti- bility to breach (Robinson etal, 2016), Clean membrane resistance and fouling rate were assessed via clean water resistance tests and permeation tests, respectively, Susceptibility to breach was assessed indirectly via tensile testing. Clean membrane resistance, fouling rate and tensile testing observations are presented in ‘igs. 3-5, respectively. Data relating to each performance factor have been segregated into 1000A membranes (left side of each figure) and 10008 membranes (right side of each figure), All three membrane performance factors - the clean membrane resistance (Fig. 3a), fouling rate (Fig. 4a) and all three indicators of susceptibility to breach, percent elongation at break, maximum sess, and Young's modulus (Fig 5a , and Se respectively) - were ‘observed to be remain unchanged for 100A membranes with less ‘than 5 years of operation. Unfortunately, no data is available for 1000A membranes with more than approximately 5 years of ‘operation. Further research is required to assess the impact of age ‘on performance factors for this newer market product, and to ‘compare to the results for the 10008 membranes (below). Note that the magnitudes of performance factors should not be compared directly for the two membrane types, as they are different products with slightly different chemical and physical characteristics (see Table 2) All three membrane performance factors -the clean membrane resistance (Fig. 3b, fouling rate (Fig. 4b) and all three indicators of susceptibility to breach, percent elongation at break, maximum stress, and Young's modulus (Fig. 5b, d, and 5f respectively) ~ were also observed to be remain unchanged for 10008 membranes with less than 5 years of operation. However, all three membrane performance factors were ‘observed to change significantly for 10008 membranes with more than 5 years of operation. As illustrated in Fig. 3b, the clean ‘membrane resistance was higher and more variable for older 10008, ‘membranes than for newer 10008 membranes; the magnitude of the clean membrane resistance increased with membrane age ‘beyond 5 years. The fouling rate was also higher forthe older 10008, ‘membranes, and increased in magnitude with membrane age beyond 5 years, a illustrated in Fig. 4b, For 10008 membranes greater than 5 years of age, the percent elongation at break increased, while the maximum stress and Young's modulus decreased with membrane age Fig. 5b 5d, and Sf respectively). ‘Most bench-scale single-soak accelerated ageing studies of PVDF membranes have reported a decrease in membrane resis- tance with age (Abdullah and Bérubé, 2013; Hajbabania et a, 2012; Le-Clech, 2014). These studies attributed the reduction in resistance to an increase in membrane pore size caused by ageing. ‘The general explanation is that prolonged exposure to NaOCl (cleaning agent) causes removal of additives (eg. HA), resulting in an opening of the membrane pore structure (Abdulla and Bérubs, 2013; Arkhangelsky etal, 2007; Hajibabania etal, 2012; ng eta ‘2004; Pellegrin etal, 2015). However, in bench-scale studies where cycled accelerated ageing was applied (i.e. repeated fouling and cleaning), an increase in resistance with age has been reported (Hajbabania et al, 2012; He et al, 2014; Wang et al, 2010). The increase in resistance observed in the present study was atributed to an accumulation of irreversible foulants on the membrane. The increase in the magnitude and variability of the clean water resis- tance observed in the present study after 5 years of operation for the 1000B membranes at full-scale is consistent with the results reported by others for cycled ageing studies, suggesting that the resistance increase at full-scale is due to irreversible fouling. Similar fouling rate results have been reported by others for both single-soak ageing and cycled ageing studies, These studies hhave reported an increase in fouling rate with age (Abdullah and Bérubé, 2013; Gaudichet-Maurin and Thominette, 2006; Qin al, 2004, 2003). This increase in fouling rate is usually atrit ‘uted to an decrease in membrane surface hydrophilicity with age, as HA’s are removed; the impact of age on HA content is further ‘iscussed in Section 32 (Abdullah and Bérubé, 2013; Arkhangelsky et al, 2008; Gaudichet-Maurin and Thominette, 2006; Levitsky et al, 2012, 2011, Qin et al, 2004, 2003). The increase in fouling,6 5 Robinson PR, Bre / Water Research 1692020) 15212 @ Plant a n x ma neo + Aba a0 (lean Membrane Releance (10% m" Memerane Age (years) ° T ze i g i L “ i Ladys é ete ° o 2 4 6 6 Martane fos ea) Fig.3. Gran membrane resstance with respect to membrane age for 10004) a 10008 (8) means. The age grey atoms arene israte perl en. ror bars indicat the maxim and minim fora ven sampling event The regression fr 10008 membranes older than 3 yas has alge of0.738 + 030910" yan asope 10° Lpmmin, Fost Ray Membrane Age (ee) Fouling Rate10m' (ppm) . 2 . 6 8 © Momtrane Ago (ears) "ig. 4. Mernrane oling ate wit spect to membrane ag fo 10004) an 1008 (5) membranes The ge grey ars are nce osteo ens. Ero ars Indicate the maximum and mani fora en sampling event The regesson fo 1008 meranes ker than years has lope of 2039 = O07"10"m ‘pp yr" anda Slope pale af 0036 Nee that olin ates are nralzed Yo olan conceatrton, rate observed in the present study after 5 years of for the 10008 ‘membranes at full-scale is consistent withthe increase in fouling rate reported by others for single-soak and cycled ageing studies, though magnitudes cannot be compared due to variation in foulant type and testing protocol. ‘The tensile testing results from the present study indicate that aged 10008 membranes are more ductile. Similar results have been reported previously (Wang et al, 2010), though others have re- ported decreasing elongation with age (ajibabania et al, 2012; Le ‘Cech, 2014), The observed maximum stress at break indicates that ‘aged 10008 membranes are weaker. Similar results have been re- ported for bench-scale aged membranes in two prior studies (Hajibabania et al, 2012; Wang et al, 2010). Te elasticity results indicate that aged 10008 membranes ate more readily deformed with age. A decreasing Young's modulus has also been observed by ‘others for bench-scale aged membranes (Wang et al. 2010}. Me- ‘chanical deterioration has been attributed to change in crystallinity ‘of the PVDF as additives are removed around the PVDF molecules (Hashim etal, 2011). However, the mechanism behind the deteri- ‘ration of mechanical properties is not well understood, and was beyond the scope of the present research. Nonetheless, the ‘observed changes in mechanical properties suggest that aged ‘membranes may be more susceptible to breach than newer ‘membranes. 3.2, Bec of ageing on membrane characteristics Similar trends to those observed for the performance factors were also observed for the membrane chemical characteristics. ‘Two analytical techniques, ATR-FTIR spectroscopy and NMR spec troscopy, were used to quantify the amount of hydrophilic additive (HA) present on the membrane surfaces and in the membrane bulk, respectively. ATR-FTIR. and NMR observations are presented Figs. 6 and 7, Data relating to each characteristic have been segre- gated into 10008 membranes (left side of each figure) and 10008 ‘membranes (right side ofeach figure). The membrane surface HA content, assessed via ATR-FTIR, decreased slightly (Fig. 6a). and the bulk HA content, assessed via NMR, was observed to be remain unchanged! (Fig. 7a) for 100A, ‘membranes, ranging from 0 to $ years of operation. Though‘Robinson PR, Bras Water Resear 169 (202015212 = o Prcont Elongation at Break (8) amet ae oz Jag a _ . tiaaemmnbeicinyemy : : : a {~ a _ i: : . ewtwonle a a © os ® a =| gooe os | ree be=|, jddiayftbeimiep fo tee cn fea et 7 = 7 i Ss 1 2 3 4 5 oe 2 4 6 8 10 Lata ont Fg 5, Mechanical properties wih respect to membrane ag fr 10008 (a) and 1008 (bf mera The are ae rows anced istate overall ends Ee bars inate the stander er fom eight epee The repression fr 10008 membranes ater years have slopes of0.366. 027%) elongation, -OOT x 004hgit (xan sues) and L007 = 0003 ay (Yunds maduls an slope p-values 00100017 and GO especie. The nso Young's ads ad Makin Sues ae wh espet 02 ingle fee. notnemalzed to rsssectinal a2) decrease in HA surface content was statistically significant, amounted to less than 10% surface HA being removed over 5 s0 was a small effect. As previously discussed, no data is available for 100A membranes with more than approximately 5 years of ‘operation. To gain insight into the stability of HA for 1000A mem branes, beyond! a 5-year period, an ad hoc single-soak bench-scale accelerated ageing test was performed. Newi000A membranes ‘were soaked in a 5000 ppm solution of NaOC, at pH 10, for 16 days: {equivalent to a dose of 2 000 000 ppmh. Based on this test, 35% of the HA content was removed from 100A membrane surface (ATR- FTIR analysis) and 22% from the bulk (NMR analysis). These results indicate that the HA content of full-scale 10008 membranes is ‘expected t0 decrease in a manner similar to that observed for the 10008 membranes. This suggests that the performance factors are® 5 Robinson PR, Bre / Water Research 1692020) 15212 ? wee doo i a ost 3 02 Mertrane Age eer) Fig Sc HA content (ATIF with reipec to membrane ag fr 1000 (and 10008) membranes. The age rey srows ae ince to strate vel trends Eo Sars ect the standard erro four replat speca The regen for 10008 memanes le than 5 years has a ope of 0135x044 yr anda slope pale of 007 Pts are noalzed to vign (000A) or ewes (10008) HA content, and within-sample HA was measued flit PVD. las ta) ‘| o 1 2 38 @ Mesmtrane Ao (vss) Momivane Ag (yours) Fig 7. Bak HA content NMR) with respect to membrane age or 100A) and 1008 (b) membranes. The rege arows are included co state oer ends. De 0 Time instrament avait fur repeats were peered for ol 2 samples selected at random Represent Bars showing anda ero are stated or these two samples. The repesion ater years ope -01502 0058 andaslope p-value of 018 Pts ae nome ova (10008 or newest (10008 HA conten and ‘witinsimple HA was measured rate to NDE «also expected to change in a manner similar to those observed for the 10008 membranes beyond a 5 year period, Continued charac- terization of the performance factors and chemical characteristics ‘of membranes aged at full-scale is recommended for both the 100A and 10008 membra Both chemical characteristics - the surface HA content (Fie. 6b) ‘and the bulk HA content (Fiz. 7b) ~ were constant with respect t0 ‘membrane age for 10008 membranes with less than 5 years of ‘operation. However, both chemical characteristics were observed to ‘change significantly for 10008 membranes with more than 5 years ‘of operation, For 10008 membranes greater than 5 years of age the surface HA content was lower than for those less than 5 years of, age; the magnitude of the surface HA content decreased with membrane age (P= 0.007), as illustrated in 6b. For 10008 mem- branes greater than 5 years of age, the bulk HA content was lower than for those less than 5 years of age; the magnitude of the bulk HA content decreased with membrane age (P~0.018), as illus- trated in Fig. 7. Statistical analyses are presented in the supple- ‘mentary material To better understand membrane degradation, bulk and surface HA:PVDF ratios were compared as presented in Fig. 8, HA content ‘was compared to assess whether removal is preferentially from the ‘membrane surface of the bulk. “The bulk and surface HA content was generally lower for aged 10008 membranes (Figs. 6b and 7}, respectively) As illustrated in Fig, 8b, the results indicate that although the removal of HA ‘occurred throughout the membrane matrix surface HA may have been initially preferentially removed. but that bulk HA was removed {0 a greater extent as membranes aged, This trend is illustrated with a curved arrow (Fig 8b), though it was beyond the scope of the present work to establish a model for this phenomenon, ‘though the mechanism of HA removal was not investigated in the present work, others have observed that HA in membranes with 10008 type chemistry is oxidized by NaOCI (Abdulla and ‘Berube, 2013). The oxidation products are soluble, and are there- fore released from the membrane matrix. For 1000A membranes, bulk and surface HA content remained relatively constant (Figs. 6a and 7a, respectively) and, as jlustrated‘Robinson PR, Bras Water Resear 169 (202015212 ° ® ‘Bulk HAContet (rormalized) ° 00 02 04 06 08 10 12 ‘Surface HA Content (normalized) 14 2 ® ‘Buk HA Content (normalized) o 02 04 of o8 4 ‘Surface HA Content (nermalized) 12 Fe. Sure an balk composian empiri for 10008 () an 10008 () membranes. Te ag sey ar i ince to aseate the apparent over end ata are oc vii (10008) er newest (10008) HA comes apd wihin-sample HA was measured ceive to PDE. The sea the ynbolscoesponds Wid ears of oprah. in Fig. 8a, the bulk and surface HA content are clustered together, regardless of membrane age. However, as previously discussed, the HA content of the 10008 membranes is expected to change in a ‘manner similar to that observed for the 10008 membranes as the ‘numberof years of operation increases. 3.3, Linking performance factors to characteristics Potential links between changes in performance factors and ‘membrane characteristics were assessed to better understand the ageing phenomena, Because there was no significant change in the performance factors or chemical characteristics of 1000A mem- branes for the range of membrane ages available (Le. less than 5 years old), the following analysis is based only on data for the 10008 membranes ofall ages Many studies have suggested that the HA content of PVDF ‘membranes directly impacts the membrane fouling rate (Robinson cecal, 2016). The fouling rate with respect tothe surface HA content Js presented in Fig. 9. The fouling rate was significantly higher for 110008 membranes with lower HA content on the membrane sur- face. These results support the concept that as the surface HA content decreases, more hydrophobic membranes have a higher affinity for organic foulants. 34, Identifying benchmarks for operational water treatment plants ‘To adapt the findings ofthe present study for use by operational facilities, we sought to identify a convenient metric of membrane ageing for full-scale facilities. Changes were observed in all metrics (performance factors and chemical characteristics) after about 5 years for 10008 membranes. Continuous monitoring ofall perfor- ‘mance factors and characteristics to monitor ageing would not be feasible at full-scale facilities; the time, personnel, analysis, and harvesting requirements would be onerous. However, continuous ‘monitoring of resistance and, in particular, resistance following extensive chemical cleaning is performed at full-scale. Resistance following extensive chemical cleaning is analogous to the clean water resistance performance factor presented here. ‘As presented in the supplementary material, the clean mem- brane resistance correlated well with all other metrics as 04 8 Fouling Rete(*10"? mJ [ppm*min)) 8 ° 00 02 04 08 = 08 ‘Surface HA Content (normalized) wo 12 ig 9, Fong ate ets HA content (FI for 10008 membeanes. The eres as 2 Slope of 024120083, (er) anda slope pevaue of 00000, Symbal ze portant meme ae ‘membranes age. Therefore, high resistance following extensive cleaning could be used as an indicator of membrane ageing at full scale facilites. The observed change in the various metrics considered with respect to age, and with respect to observed clean membrane resistance, are summarized in Table 3. ‘Based on these observations, membrane facilities with sus- tained and irreversible increases in resistance following, extensive cleaning should be vigilant of membrane ageing. Note that all resistance values should be normalized to a standard temperature (eg. 20°C) to account for seasonal changes in temperature0 5 Robinson PR, Bre / Water Research 1692020) 15212 ‘ables ‘Change in ageing metic with membrane age and ean membrane rstance, for 10008 membranes Comelaions with ge were acute in Section 21 and 52: Daa fom corelatons with resistance af povided i the supplementary mata Conan th Membrane Age Correa wth ian Membrane Resitance oaing tte Srae HA alka Percent Bongation ‘Young's Moduls 35. Relationship between age and exposure dose ‘The estimates of the exposure dose (C't) to NaOCl compared with operational ages forthe 5 facilities for which such data was available are presented in Fig. 10. Note that most of the partner facilities assume that the concentration of NaOCl in the cleaning solution remains constant during storage and use. It should be noted that the NaOCl concentration is expected to decrease as it reacts with foulants and other material in the membrane tank. A strong correlation (R2 =0.95) was observed between membrane age and reported NaOCl exposure dose for most facilities. However, for one facility the cumulative NaOCl exposure dase was much higher than the others fora given operational age. Data from this facility are identified by the symbol %' in Fig, 10, as well as in Figs. 3-7. Further investigation revealed that this facility nad imple- ‘mented an atypical (ie. frequent and aggressive) chemical cleaning protocol in an attempt to address operational problems. Consid- ‘ering the linear relationship between the cumulative exposure ‘dose and operational age, it may be possible to estimate an effective age for this atypical facility, by identifying the operational age for ‘which the estimated exposure dose intersects this linear relation- ship (Fig. 10), For the atypical facility, the effective age was esti- ‘mated tobe approximately 13 years. Unfortunately, the membranes at the atypical facility were replaced soon after our study ‘commenced, so further temporal data were unavailable for this, 400000 | (= “Typical Atypical ‘as0000 = ‘Effective! x 300000 250000, 200000 ! Oy = 28696x-31143 RP=0.951 150000 Bost Estimate of * (ppm 100000 50000 o 5 10 18 Membrane Age (Years) ig 1. Bes estimates of mate ifetime NOC fr al 10008 membrane in {he study Te membranes inated with XC ae Hoan a0 Spel la sd were ‘lod fom the eres analy particular train. Recalling Sections 31 and 3.2, the clean membrane resistance for membranes from the atypical facility was more than double that from all other facilites (Fig. 3b), and fouling rates from this facility were among the greatest observed (Fig, 4b). The mechanical properties of membranes from this facility were also poorer than the other facilities in the study (Fig. 5b, d, and had the lowest ‘observed HA content (both surface and bulk) of all (Figs. 6b and 7b, respectively). The ageing trends in Fig. 3 through Fig.7 included the ‘membranes from the atypical plant at their operational age (approximately 8 years). However, when the performance factors and characteristics for membranes from this atypical facility are plotted with respect to effective age, rather than actual age, al Correlations considered improve significantly, as presented in the supplementary material, ‘These result indicate that if extensive and frequent chemical cleaning is required for operation, beyond what is typical, mem- thus, changes in perfor- expected to occur earlier and to a greater magnitude. In such cases. operational protocols should be reviewed to reduce the extent and frequency of chemical cleaning - for example, by improving or implementing. pre- treatment. 4. Conclusions The present research investigated changes in performance fac- tors as well as chemical characteristics of full-scale aged ultrafil twation membranes. Our work illuminates the knowledge gap between results from previous bench-scale ageing studies and performance factors and chemical characteristics relevant to operational membrane facilities. For 000A membranes, no significant changes in performance or chemical characteristics were observed over the course of the study. Older 1000A membranes would be necessary to determine whether this stable operation persists, but such full-scale aged ‘membranes are not yet available. Results from a single-soaked accelerated ageing test of 1000A membranes indicated that HA Content of fullscale aged 1000A membranes is expected to decrease with sufficient NaOCl exposure; thus, we anticipate that performance may also change after long-term full-scale ageing, as ‘observed for the 10008 membranes. We observed that the clean membrane resistance of 10008 ‘membranes increased after approximately 5 years of operation. Increasing resistance is consistent with some bench-scale aged ‘membranes that have been cycled with foulant and cleaning agent. However, our results contradict bench-scale ageing using a pro- longed single-soak in NaOCl. The discrepancy likely stems from the ‘opposing effects of pore enlargement and irreversible fouling on the clean membrane resistance. The membrane fouling rate also increased for older membranes, consistent with the literature on bench-scale ageing. Though the susceptibility to breach cannot be ‘measured directly, tatee metrics (elongation at break, maximum‘Robinson PR, Bras Water Resear 169 (202015212 " stress, and Young's modulus) were all observed to deteriorate with ‘membrane age, also in agreement with bench-scale ageing Both bulk and surface chemistry of the 10008 membranes changed with age after 5 years of operation. The hydrophilic ad- itive was observed to decrease in the bulk and on the membrane surface, The reduction in HA content likely increases the fouling rate for older membranes, as foulants tend to sorb to more hy- {ophobie surfaces. ‘We observed that the clean membrane resistance correlated well with all other performance factors and chemical characteris- tics studied. Ifa sustained irreversible increase in clean membrane resistance is observed over time, plant operators should be cogni- sant that fouling and achieving sufficient flux may become dificult, and that membrane breaches may occur more frequently. This finding is useful to water treatment plant operators or managers seeking a metric to assess their infrastructure, and for membrane ‘manufacturers in better understanding their products Finally although C*t was not used as a metric of membrane age, we observed that for one plant that applied an atypically high dose ‘of NaOCl the effective age ofthe membranes appeared to be greater than their operational age. Thus, as indicated by the results of bench-scae studies, voiding excessive cumulative NaOCl exposure is advised, Furthermore, NaOCl C°t would likely provide a suitable ageing metric for ullscale ageing if dose fluctuates over years of operation. Declaration of Competing Interest ‘The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. ‘Acknowledgements ‘The authors would like to acknowledge the National Sciences & Engineering Research Council of Canada (NSERC) for funding the research presented inthis article. ‘Appendix A. Supplementary data Supplementary data to this article can be found online at baps:/idoi.org/10.1016)}.watres.2019.115212. References, Andula, 52, Béubé PR, 201. Assessing the effets of som hypochonte ‘exposure on the characteristics of PVDF Based menbranes. Water Res. 7, $S3o2~s300hrps'/douorg/ 100). watres 20130808. ArkhangetsyB, Goren U. Gs, V. 2008 Retin of orgaic matter by elluose ‘etate membranes ceaned with typocont. Desaiaton 223, 7-106, replete) 101016 desal200704070 ‘chanel. £, Kuzmenko, D- Gt NV. Vbograda, M, Ki 8, Gti, 2007, Hypochlorite leaning causes degradation of polymer membranes Tol: 24, 109-116 hep dovorg/101007/t248-007-9253- ASTM 2014 Standard Test Method for Tensile Properties of Single Texte bers. ns oor 10132n/D9522 D3822/D3822M ‘causetand, C. Kowa 5 Lal J-P. Aimar. P, 2008. Ageing of pohsullone ‘embrane none wih each ston: role of ral Oidtion and Sime’ dsaved metal ane Chem. Eng, acess 48°56, hp )oo ‘0016092007080, Ccetenden Cy a 2005, Water Teatmen: Pipes and Design Wie. Feu A. De Wide. W. Gaeriner M, Weertaes M. de Gueve,C. Van De stene A. 201 aborting he membrane fe cance in fal ale bers MOR. Memb Se. azinae 329-354. paar 10101 Jinemse 2012.08.01, Ccaudicet Maurin E, Thominet, F, 2005. Ageing of polysulfone uration ‘membranes ih contact with beach salons.) Meni. Se 282, 198-204, ‘epydetor/ 101016} memes 200603023, ‘aww 20 reste Vancouver Wate Distt 2016 QultyContrel Annual Report, wo 1 (Vancouver. Hajbabuni, , tony. A, Les, Lele, P2012 Reative pact of fol ‘ahd cleaning on PVDF membrane hyrale performances Separ Put Tehol fn, 204-212 eps dry 1106) seppur 201203001 Hashum NAL Lu YU, K 2011 Stblity of PVDE hollow fe membranes ino ‘hum yond squeais soko. Cher en. Se 66 1365155. htbe | {dotorg/ 1010165 cez01012018 He, ¥ sharma, J. Boga. Use, RQ. Goodwin C, Marshall K, 204 impacts of gens and chemislGeaning onthe popes and peforance of uta {ation membranes in potable water Weatment. Sepa Sct Techol. 49, 1517135. hs tog 0 100/0196395.2014882389. Jong 8. Yoon. Ks Kir, Ree. HW. 2008, fet of molar weigh of ‘netic addithce om oration, ericaton ropes and ypc et ‘eat of esmmetnepolacvontle membranes. Membr. Sc 243, 43-57 Fp:iauera 101016] mens 2004.60. evcken A, Wang Tg KH. teste. Persson, K. Kole. SJ. Spans, T2016 alton of novel hallow Mer membranes Tor NOM renal by anced ‘membrane autopey Water Se Tech, Water Supply {Es 628-40 pe) oiorg( 10.2168 js.2015.170 each F 2014 rice Tools fr Membrane Ageing WA Publishing London evs, {) Duck A, Anoangeshy. €. chen, D. Kadoshan, Ty Sheu "Na Cis V-201-Understanding the dative deaning af UF membranes |. Men. Seb 377, 206-213. hp doen 101015 mes. 201104040 tevily 1 Due, A. aim, Arbangelsty.E, GV 2012 Cleaning UF mem ates with angle and ormulated salts. Chen Eng. Se 68, S72" ep odor Toi)cx201L10.080, bearce, G_200), Memtraneflaton markt trends heed on comprehensive elrece Ist Survey. In: Bh IWA MembrabeTechology Conference & Exh bition for Waer and Wastewater Treatment and Rese, P Sngapore Pellegrin Mezza Hana Y, Szymeay A. Remy}, causeandC, 2015. tran perfocnance and poe sae dlstibton hypochlorite aged PESIPVE Uigtlraton membranes: Membe. Sa 474175186 hips sor 101018) Jimensei 201409028. (Qin) Co, YoML. L,Y. UY, Oo, MA ee, 2008, Hollow ber ultra ‘ation membeaes made fom blends of PAN and PVP Sear, Pun Technol 36, "4d" 155 hep ogi 10 1015389 5665(0)00210 7 (in JUV Wee LS, ee 2003, else acetate hallow fiber uration ‘erbranes made fom CAPVP 360 KINMP ater j Memb, S218. 173-183. eps:ouerg/1010163037-7388(03)00170, obs, 5. Abllh SZ. Berube P. Pete L-C. 2016, Aging of membranes for ater eaten ining changes opesformance.} Mem Se 503,177 18 ps:ioiag/101016h meme 201312033 ‘Robinson, Sf. 2018 racial Conieratons fer Aging of Drinking Water Mem rans Universi of Bish Colusa Wing. Wang. Wa, Z. Zou. 0. Yan. D. 2010. Et of hypochlorite cesning ‘oh the pysochemicl characterise of plywayldene forge membranes hem: Eg J 162, 1050-1056 toon ovo ee.201007018,