Fluorescence as an online tool for monitoring membrane integrity

Download files
Access & Terms of Use
open access
Embargoed until 2015-02-28
Copyright: Singh, Sachin
Altmetric
Abstract
Membranes play an increasingly important role in water recycling, with the proven ability to produce high purity water from a highly contaminated source (wastewater). The effectiveness of a membrane system depends on the intactness of that system. Integrity monitoring is thus an important part of managing membrane systems. Currently used integrity monitoring techniques, while diverse, have inherent weaknesses that limit the sensitivity of detecting integrity breaches in a membrane. A fluorescence excitation-emission matrix (EEM) method was optimised in order to produce consistent and reliable measurements from fluorescent organics in RO permeate samples. A benchtop fluorescence spectrophotometer was used for this purpose and operational parameters adjusted to optimise fluorescence EEMs. Adjusting slit width settings act as a coarse adjustment for sensitivity. Further refinement in resolution and quality was achieved by adjusting scan speed, photomultiplier tube voltage and introducing signal averaging. Results indicated optimum EEM spectra could be achieved at excitation and emission monochromator slits at 10 nm, PMT voltage setting producing an intensity of Raman peak (λEx/Em = 348/393 nm) in the blank at 270 ± 8 a.f.u (Typically this was between 840-900 V), scan speed of 9600 nm min-1 and 3 signal averages. Treatment processes from five advanced water recycling plants in Australia were studied using fluorescence and other established water quality parameters which included dissolved organic carbon, turbidity, electrical conductivity and ultraviolet absorption (UV254). Treatment processes scrutinised at these plants were media filtration using either sand or anthracite, microfiltration, ultrafiltration, reverse osmosis, ultraviolet disinfection and chlorine disinfection. Fluorescence intensities were reduced by anthracite filters, ultrafiltration in one plant only, and reverse osmosis. Correlations with other water quality parameters failed to show any trends; illustrating the complexity of fluorescent organics in wastewater and the difficulty in predicting its behaviour based on moderate changes to other water quality parameters. From the three membrane processes studied, it was established that reverse osmosis was the only membrane process that consistently reduced fluorescence intensities and was thus the only practically suitable membrane system for fluorescence based membrane integrity monitoring. The fluorescence region of λEx/Em = 320-340/410-430 nm (Peak C) was selected as the best indicator for integrity loss after further scrutiny of results using probabilistic (Monte-Carlo simulation) analysis showed this peak was consistently rejected (> 99 %) at all operational stages of RO systems compared to other fluorescence peaks. An impaired membrane was also identified in this study using fluorescence and conductivity. A flow-rig was designed to blend reverse osmosis feedwater and permeate to simulate integrity loss. Cyclops 7 CDOM fluorescence sensor (Turner Designs) was used to measure fluorescence. Several experiments were conducted to identify the impacts of flowrate, temperature, pH, antiscalant usage and dosage, free chlorine and dechlorination on fluorescence-based detection of integrity loss. Integrity loss was simulated by blending RO feed with permeate at defined ratios. The sensor was capable of detecting approximately 0.2% of the RO feed depending on RO feed CDOM concentration. It was found that the sensitivity of the sensor was not affected by any of the tested variables. A 1% RO feed intrusion in the permeate was shown to still contain very low level of fluorescence, and any changes to fluorescence intensities due to tested variables were too small to be determined. Trials were undertaken using the fluorescence-based sensor to conduct real-time, online monitoring of reverse osmosis membranes from two Australian municipal advanced water recycling plants. The sensor measured at λEx/Em= 350/430 nm and was able to detect differences between permeates from staged reverse osmosis systems and responded to changes in membrane performance. The measured membrane rejections for fluorescence were below 99%. This was because the group of chemicals measured was rejected by less than 99% and was not a limitation of the sensor's measurement range. The results also revealed an increase in permeate fluorescence due to an underperforming membrane that was suspected to have an integrity breach. Another important observation was the opposing trend of permeate electrical conductivity and fluorescence from the underperforming membrane. Decreasing feed electrical conductivity resulted in an increase in flux when operated under constant pressure conditions. While the permeate electrical conductivity displayed the same trend as the feed electrical conductivity, the fluorescence of the RO permeate was observed to increase in response to flux increase and RO feed electrical conductivity decrease, giving an impression of improving water quality (low conductivity) while organic content (as fluorescence) increased in the permeate. This illustrated the limitations of relying solely on electrical conductivity as an online monitoring tool for RO underperformance. In the second plant a relationship between transmembrane pressure and permeate fluorescence was established in stage 1 membranes, with an observed increase in permeate fluorescence coinciding with transmembrane pressure increase. This was believed to be related to membrane fouling and flags another potential application of the fluorescence sensor for detecting the onset of membrane fouling. The overall results show fluorescence spectroscopy is sensitive to subtle changes in reverse osmosis permeate water quality and can thus effectively detect integrity loss in reverse osmosis membrane systems, specifically in regard to organic chemical substances.
Persistent link to this record
Link to Publisher Version
Link to Open Access Version
Additional Link
Author(s)
Singh, Sachin
Supervisor(s)
Khan, Stuart
Stuetz, Richard
Henderson, Rita
Creator(s)
Editor(s)
Translator(s)
Curator(s)
Designer(s)
Arranger(s)
Composer(s)
Recordist(s)
Conference Proceedings Editor(s)
Other Contributor(s)
Corporate/Industry Contributor(s)
Publication Year
2013
Resource Type
Thesis
Degree Type
PhD Doctorate
UNSW Faculty
Files
download whole.pdf 5.7 MB Adobe Portable Document Format
Related dataset(s)