Magnetic Resonance Imaging of Membrane Fouling Dr Einar Fridjonsson - PowerPoint PPT Presentation

Magnetic Resonance Imaging of Membrane Fouling Dr Einar Fridjonsson Fluid Science & Resources School of Mechanical and Chemical Engineering University of Western Australia

Mobile NMR technology Research Areas: Low field NMR (Remote Operations): (1) Emulsion & oil discharge monitoring Oil & Gas industry (2) Multi-phase flow metering Oil & Gas industry Mining & Coal seam gas industries (3) Well logging (4) Membrane fouling (Desalination) Desalination industry

Reverse Osmosis Membranes: NEED • 87 million m 3 /day desalination capacity (2015). • 18,426 desalination plants worldwide. • Globally more than 300million people rely on desalination. Sources: UNESCO, IFPRI (Source: IDA - International Desalination Association)

Local motivation 47% of Perth‘s water comes from desalination! Fig. 1. (a) Kwinana desalination plant in Perth, Western Australia; (b) an example of a heavily biofouled desalination membrane module, the dark regions are due to biofilm. 4

Reverse Osmosis Membranes: Construction Feed Feed spacer Feed water Concentrate Permeate Permeate RO Core

Bio-fouling is a major limitation for ROMs

NMR/MRI Studies Bench-top Mobile High-field (Permanent Magnet) (Permanent or No Magnet) (Superconducting) (Cost > $100k) (Cost < $10k) (Cost > $1M) Research aims: • Direct evidence that spacers host biofilm growth and loss of membrane performance • Direct measurement of ROM cleaning potential • Early detection of membrane bio-fouling • Development of low-cost MRI solution for monitoring membrane fouling.

Schematic: Flow loop for spiral wound membrane fouling Tap water Carbon Pressure filter regulator Pump Nutrients RO module Discharge Flow controller Δ P Differential pressure transmitter

Imaging Biomass Accumulation (High-field) Unfouled Fouled

Imaging Biomass Accumulation (High-field) Velocimetry Graf von der Schulenburg, D.A., Vrouwenvelder , J.S., Creber, S.A., van Loosdrecht , M.C.M and Johns, M.L. (2008), Nuclear Magnetic Resonance Microscopy Studies of Membrane Biofouling, J. Memb. Sci. , 323 (1), 37-44.

Imaging Biomass Accumulation – Model System x y z 16 mm 37 mm

Imaging Biofouling cleaning processes - Example Structural pH 12 NaOH at 45°C, 100 mL/min for 1.5 h 0.05 0.05 m/s m/s Velocity -0.01 -0.01 m/s m/s A variety of cleaning protocols assessed and effectiveness related • to original fouling structure Creber, S.A., Vrouwenvelder, J.S., van Loosdrecht, M.C.M and Johns, M.L. (2010), Chemical cleaning of biofouling in reverse osmosis membranes evaluated using magnetic resonance imaging, J. Memb. Sci. 362 (1-2), 202-210.

Clean Fouled Front Middle End

On-line Analysis? (b) (a) (c) (d) 55 mm 55 mm On-line NMR/MRI tool should be simple, robust and low cost. Superconducting Magnets Permanent Magnets

Even Simpler System: Mobile NMR/MRI Nuclear Magnetic Resonance (NMR) measurements conducted using Earth’s magnetic field as the external ( B 0 ) magnetic field.

NMR experiments conducted at end of each fouling stage (indicated by arrows): Fridjonsson et al. J. Memb. Sci. 489 (2015): 227-236.

High field MRI - Observations Spiral Wound Membrane: High Field MRI (400MHz) After Fouling Before Fouling Flat Sheet Membrane: Before Fouling After Fouling Observations: Fouling causes a backbone (Channeling) flow occurs within membrane system: Results in stagnant (slow) flow regions & Flowing regions to flow at higher velocity.

Low field NMR - Observations No Fouling: Linear decrease in NMR signal with increasing velocity:   T U − − T / T ฀  E  0 1 S S e = E 2 d   L d “Outflow” effect Fouling Stage 3: Negligible decrease in NMR signal as function of increasing velocity. NMR signal measured has increased. Results consistent with high field NMR observations: Fouling causes stagnation (low flow) regions to form, resulting in increased total signal, and independence of increasing flow rate. Fridjonsson et al. J. Memb. Sci. 489 (2015): 227-236.

Acquire only the moments of the signal distribution - Test Frequency domain, S Frequency domain, φ Spatial domain Fourier transform k y k y y x k x k x S(k) 1 ≈ − 2 σ 2 2 ln k S max Da Fit Σ y φ ln( S / S max ) x k x k x Fridjonsson et al ., J. Magn. Reson. 252 (2015): 145-150.

Magnetic Resonance Signal Moment Determination using the Earth’s Magnetic Field S(k) 1 ≈ − 2 σ 2 2 ln k S max 1.1 140 1 2 nd Moment - σ 2 -(cm 2 ) 120 2 nd Moment Pressure Drop (kPa) 0.9 100 Pressure Drop 0.8 80 0.7 60 0.6 0.5 40 0 10 20 30 40 Fouling Time (Days)

Future Work: Modelling of Outflow (EF NMR) 400MHz MRI EF NMR EF MRI Figure 1: Typical model output with model prediction, (solid blue line) and NMR output (black crosses). It can be seen that there is good agreement between the model prediction and the NMR signal measured. 26

Future Work - Signal Enhancement & Customisation Signal Enhancement : (i) Dynamic Nuclear Polarization (DNP) (ii) Compressed Sensing Miniaturizing Hardware (iii) Bayesian Analysis (NMR Spectrometer) Custom Built NMR coils : NMR-CUFF (Windt et al. 2011) CUFF – Cut-open, Uniform, Force Free

Measuring Concentration Polarisation A phenomenon whereby the flux through the membrane is controlled by the film mass transfer resistance on the feed-side rather than purely the resistance of the membrane itself. Permeate solute molecules Feed boundary layer Permeate

Sodium ( 23 Na) MRI (High-field) Flat sheet membrane system: - Monitor interplay of fouling and concentration polarisation using sodium MRI. Spiral wound membrane module: - Use 23 Na MRI techniques to monitor concentration polarisation and fouling. (a) 1H image and (b) 23Na MRI images of a flat sheet membrane module (resolution 0.01 by 1mm2). (c) Shows a sodium profile of the operating membrane module (b), with concentration polarisation evident at interface . 29

Membrane module geometries: (i) Spiral wound (ii) Hollow fiber 30

Hollow Fibre Membranes (HFM): Optical MRI Non-invasive performance measurement of membrane distillation hollow fibre modules – Four different arrangements tested. 19mm Collaboration with: Singapore Membrane Technology Centre. Yang et al. J. Memb. Sci., 451, 46-54 (2014). 10mm Bench-top NMR 10mL/min 20mL/min 30mL/min 40mL/min 50mL/min 1000mL/min 1500mL/min 10 0 mL/min 400mL/min 2500mL/min

Ultrafiltration (UF) HF membranes Module type: SIP-1013 Material (membrane & housing): polysulfone (C 27 H 22 O 4 S) n Membranes no.: 400 ID: 0.8 mm; Length : 205 mm 32

2-D MRI (Bench-top) In-plane resolution: 180µm x 180µm Slice thickness: 1.42cm Acquisition time: 2.3hrs Aim: Monitoring effect of fouling on membrane performance using velocity images. 46 mm 33

2-D MRI (High-field MRI) Capillaries Outer shell Permeate Feed water Concentrate (a) 13 mm 0.06 m/s (b) 13 mm -0.02 m/s Flow

Biofouled HFM – impact on flow distribution Clean Fouled (a) (b) 13 mm 0.06 (a) (b) 13 mm -0.02 Flow 35

Acknowledgements Mike Johns Sarah Creber Daniel Graf von der Schulenberg Wiktor Balinski Ryuta Ujihara Nicholas Bristow Andrew Sederman Dan Holland Szilard Bucs Hans Vrouwenvelder Mark von Loosdrecht 36

Funding/Support from 37

Mobile NMR and MRI 38

THANK YOU http://www.fsr.ecm.uwa.edu.au

NMR Measurements: T 1 & T 2 Relaxation: Proton density: Free Water Water (surface interacting) Velocity: Diffusion/DSD: Chemical Shift: 0.20 0.4 Oil 0.15 0.3 amplitude 0.10 0.2 Water 0.05 0.1 0.00 0 -100 0 100 200 300 0.0 5.0 10.0 frequency / Hz Droplet size ( µ m)