Struvite Recovery From Source-Separated Urine Utilizing Fluidized Bed Technology


TR Number



Journal Title

Journal ISSN

Volume Title


Virginia Tech


Source-separating urine for nutrient recovery may provide multiple benefits with regards to wastewater management, water conservation, and an impending phosphorus fertilizer shortage. Municipal wastewater systems are designed to treat the combination of urine, feces and graywater produced in household applications. Urine accounts for 1% of wastewater by volume, but provides 70-90% of nitrogen, 35-70% of phosphorus and 50% of the contaminants of emerging concern entering municipal wastewater treatment (Larsen and Gujer 1996). Research has shown managing source-separated urine for nutrient recovery is a more cost effective and less treatment intensive method than using traditional systems found in municipal wastewater plants.

Phosphorus fertilizer shortages are projected as current sources diminish and become increasingly difficult to extract and refine. Phosphorus based-fertilizer recovery, in the form of 99.9% pure struvite (MgNH4PO4•6H2O), has been demonstrated successfully in full-scale sidestream treatment using dewatering liquor from anaerobically digested solids (centrate) processed through upflow fluidized bed reactor technologies (Britton et al. 2005). Prior research determined the influence of pH, magnesium to phosphorus (Mg:P) molar ratio, and age of urine on purity, pharmaceutical content and pathogen inclusion in struvite precipitated from source-separated urine. This is the first known example of an attempt to produce a commercially viable struvite product from source-separated urine in a fluidized bed reactor of a design that has been used successfully for struvite recovery in conventional wastewater applications.

In order to assess the feasibility of nutrient recovery of phosphorus-based fertilizer recovery from source-separated urine, the first office-based urine separation and collection building was implemented in the U.S. Urine was collected, in a 400 gallon capacity underground sealed manhole, from HRSD's Main office building beginning in March 2015 from 5 men's waterless urinals and one women's separating toilet. Urine was collected from the manhole on a monthly basis in 275 and 330 gallon plastic totes stored at the HRSD Nansemond WWTP in Suffolk, VA. Collected urine was allowed to age while in storage to encourage the precipitation of excess multivalent cations that may interfere with struvite precipitation and inactivation of pathogens that may be present.

An upflow fluidized bed reactor (UFBR) was used to recover struvite as a slow-release phosphorus based fertilizer (prill), the reactor was loaned to HRSD by the University of British Columbia. A magnesium solution was injected at the bottom of the reactor to facilitate precipitation along with the recycle urine stream and feed urine as shown. Prill production design for the reactor was 0.5 kilograms per day, but while using centrate to determine best operations practices, under loading the reactor to 0.25 kilograms per day maximized struvite recovery while minimizing particulate phosphorus loss. Urine was fed into the reactor for struvite removal based on phosphorus loading with recovery determined through removal of orthophosphate and harvesting of the struvite product. Consistency, size and quality of product including compactness, crystal structure, purity and presence of pharmaceuticals and pathogens were assessed.

The UFBR was run for 50 days total; 10 days for a short term run to compare to operation of the reactor under the same conditions with centrate from anaerobically digested solids as a feed source, 30 days to assess consistency of operations over long term with respect to struvite recovery, and a 10 day test with urine spiked with pharmaceuticals and bacteriophage to evaluate inclusion of trace organics and viruses in recovered struvite. In total 2,040 gallons of urine were fed to the reactor targeting 12.45 kilograms of struvite recovery, a mass of 7.54 kilograms of prills were harvested from the reactor with 1.90 kilograms of phosphorus lost as particulate struvite (representing an recovery efficiency of 60.5%). Overall reactor operation using urine as a feed solution behaved similar to centrate, with slightly less removal of phosphorus. Urine-derived prills were lower in quality due to the lack of compact density seen in struvite recovered during full scale operation but had a visible orthombic pattern seen in precipitated struvite.

Pharmaceuticals that were present in urine feed solution were found in struvite but at less than 1% of the feed mass. Some of this inclusion may have occurred due to porous characteristics of the small-scale UFBR recovered struvite rather than through actual inclusion in the mineral crystal itself. Spiking of caffeine and ibuprofen to high concentrations in the urine yielded no statistical difference from the non-spiked tote. Urine was non-detect for bacteriophage pathogen indicators leading to the assumption that no pathogens were present in urine-derived struvite. Spiking the urine with double-stranded DNA (T3) and single-stranded RNA (MS2) bacteriophage capable of infecting bacterial cells such as Escherichia coli yielded 10^6 plaque forming units per milliliter in source separated urine.

Creating urine-derived struvite prills with minimal inclusion of pharmaceuticals using upflow fluidized bed technology is feasible on a small scale. Large-scale application, recovering 500 kilograms per day of struvite or more, will most likely create a higher quality prill with regards to compactness and diminished presence of pharmaceuticals and virus inclusion. Pretreatment of urine and post-treatment of prills with heat will aid in inactivation of virus that may be present. '



Urine diversion, struvite precipitation, nutrient recovery, pharmaceutical inclusion, pathogen inclusion