Maximizing Pathogen Recovery and Flux in Tangential Flow Filtration Processes to Enable Rapid Detection

Jessica Zuponcic, Purdue University

Abstract

Bacteria which enter a viable but non-culturable state cannot be concentrated by enrichment. This means they may not reach a detectable concentration for PCR methods - especially in the presence of sample compounds which may act as PCR reaction inhibitors. An alternative strategy for concentration of bacteria from aqueous samples is explored in this work using tangential flow filtration. The effectiveness of this technology to concentrate pathogens from food-derived samples was previously demonstrated; however, losses of bacteria to the filtration system can still be high (i.e. recovery of bacteria is low). The goal of this research was to maximize recovery of pathogenic microorganisms from hollow fiber filtration processes while also maximizing flux. In this way, high recovery filtration conditions could be selected while keeping filtration time low. It was hypothesized that flux would have relatively lower impact on final recovery of bacteria at high shear rates (27,000 1/s) which are sufficient to remove attached bacteria on surfaces. It was hypothesized that these high shear rates would not cause loss of bacterial viability, and the main cause of bacterial losses during filtration would be accumulation on the membrane surface. To test these hypotheses, single fiber filter modules (both microfilters and ultrafilters with 0.5 mm inner diameter), were constructed and used to concentrate GFP-producing Escherichia coliat a wide range of flux conditions. Post-concentration, fluorescence micrographs of bisected hollow fibers illustrated patterns of bacterial accumulation along the length of the fiber. A simple recovery model was constructed to predict recovery as a function of flux and shear rate, and predictions were compared against the experimental data. Both in the experiments and in the simple recovery model developed in this dissertation, recoveries near 90% were achievable at high shear rates when flux was ≤0.5 mL min-1 cm-2 . This amounted to a 3-hour filtration time for a 225 mL sample. Compared to a filtration with only 30% recovery, detectable bacteria concentrations could be achieved with lower starting concentrations – ~5 CFU/mL starting concentration versus at least 15 CFU/mL. Given these high recoveries (determined with plating methods on agar) occurred at high pressure and shear conditions, it was determined the filtration did not affect bacterial viability. In addition to using the model to predict recovery at various shear and flux conditions, it would be helpful to predict module designs or concentration strategies which could improve bacterial recoveries from the filter. One strategy, explored with preliminary data, was to predevelop a layer of bacteria on the filter surface prior to concentrating samples. Understanding and reducing the losses of bacteria during tangential flow filtration could enable detection of dilute levels of viable but non-culturable microorganisms; in addition, sensitivity of detection could be improved for quickly concentration culturable microorganisms in food and water samples.

Degree

Ph.D.

Advisors

Ladisch, Purdue University.

Subject Area

Fluid mechanics|Hydraulic engineering|Mechanics|Microbiology

COinS