09.12.2012 Views

NAMS 2002 Workshop - ICOM 2008

NAMS 2002 Workshop - ICOM 2008

NAMS 2002 Workshop - ICOM 2008

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

membranes were then challenged with different protein formulations (pH, ionic<br />

strength, protein mass) for 5 to 25h depending on the fouling response. During all<br />

tests, the in-situ detection of proteins associating with the membranes used<br />

ultrasonic frequency- domain reflectometry (UFDR) integrated with a fast Fourier<br />

transform protocol to process the signals. Time-domain signals of acoustic scans<br />

from ultrasonic transducers mounted on cross-flow cells were transformed into<br />

amplitude versus frequency distributions. From these distributions, reflected<br />

power was obtained, and standard statistical indices were used to report and<br />

characterize the distributions. Following cross-flow cell tests, membrane samples<br />

were analyzed using ESEM, and the proteins associated with the membranes<br />

were determined using a bicinchoninic protein assay.<br />

Depending on the fouling challenge conditions, permeate flow-rate decreased in<br />

a range between 40-90%. Permeate flow-rate responses, and patterns of<br />

corresponding acoustic reflection power changes, indicated that both internal<br />

membrane and surface deposition occurred, and could be identified as separate<br />

fouling mechanisms during some challenge conditions. The permeate-flow rate<br />

decline data can be described using the combined pore blockage, pore<br />

constriction, and cake filtration model. The best fit model parameters can be<br />

independently obtained based on the property of the feed and membrane<br />

characteristics. In some instances however, transducers were unable to detect<br />

reflected power changes even after significant permeate-flow rate decline<br />

occurred. This phenomenon could be explained by the fact that permeate flowrate<br />

observations are derived from overall membrane behavior, whereas UFDR<br />

is applied in a sentinel format, which reports acoustic responses of small area<br />

(point) observations on a very short timescales. In addition, membrane-<br />

associated protein deposits are visco-elastic, and can (and likely do) reposition<br />

on or through a membrane during the course of a fouling challenge; this can<br />

manifest in a wide variability of reflected power changes as protein density<br />

changes on a local scale during a test. Biochemical assays of protein<br />

concentrations associated with membranes and ESEM micrographs confirmed<br />

that a significant heterogeneity of protein deposition was in part responsible for<br />

the fouling behavior and local density changes observed by UFDR. Where<br />

protein concentration on the membrane varied between 5 to 100µg/cm 2 , ESEM<br />

observations showed non- uniformity of protein deposition in a capricious patchy<br />

array, with a significant amount of clean membrane area exposed; beyond this<br />

range however (100µg/cm 2 ), continuous protein layers were observed on<br />

challenged membrane surfaces.<br />

The use of fouling models in combination with non-invasive, real-time monitoring<br />

provides a unique capability to improve the fundamental understanding and<br />

control of MF membrane fouling by commercially significant proteinaceous<br />

biopolymers.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!