Murdoch University Research Repository

Welcome to the Murdoch University Research Repository

The Murdoch University Research Repository is an open access digital collection of research
created by Murdoch University staff, researchers and postgraduate students.

Learn more

Investigations of a continuous Poiseuille flow struvite seed crystallizer – Mixer performance and aggregate disruption by sonication

Burns, M., Natividad-Marin, L. and Schneider, P.A. (2016) Investigations of a continuous Poiseuille flow struvite seed crystallizer – Mixer performance and aggregate disruption by sonication. Chemical Engineering Journal, 295 . pp. 552-562.

Link to Published Version:
*Subscription may be required


A Poiseuille flow reactor (PFR) is examined for the continuous production of struvite seed crystals. Feed phosphorus concentration was 0.02 M and Mg:N:P molar ratio was 1.5:1:1. Non-equilibrium feed saturation index (SI) was varied from 0.8 to 1.4. A Roughton (R) style vortex mixer was compared to the more common impinging jet (IJ) mixer to assess differences in mixing performance. Particle size distribution (PSD) was measured either in-line (without sonication) or after filtration and sonication. All particles produced in this work were within the range of those used for previous struvite seeding studies. Increased feed SI led to an increased in-line volume median particle diameter (D[50]) and PSD width, while sonicated samples produced much smaller PSD variation with increasing SI. At a feed condition of SI = 0.8, the R mixer produced similar PSDs for in-line and sonicated samples, whereas the IJ mixer produced D[50]s and distribution widths which were larger in-line than after sonication. At SI = 1.0 and 1.4, PSD deviation between in-line and sonicated samples was caused by weakly-bound aggregates that were easily disrupted by sonication. Scatter in PSD data at higher SIs indicated incomplete mixing in both mixers. This calls into doubt the applicability of the induction time and mixing models utilized in this study. Improved mixer design models should incorporate mixer aspect ratios. A viscous sub-layer of 0.2 mm of fluid spent longer than crystal induction time in the R mixer, which is expected to have caused scale formation. Discrepancies between particle mass measurements and predictions at the reactor outlet, suggest that between 5% and 13% of particle mass was likely retained within the reactor.

Item Type: Journal Article
Publisher: Elsevier
Copyright: © 2016 Elsevier B.V.
Item Control Page Item Control Page