Urine Could Feed The World: Scientists Just Need To Keep The Filters Clean

The Next Source of Crop Fertilizer Might Come From a Toilet

Every flush of a toilet sends a surprising amount of fertilizer down the drain. Human urine carries roughly 80% of the nitrogen, 50% of the phosphorus, and more than 60% of the potassium found in household wastewater, nutrients that farmers pay dearly for and that a growing world will need more of. Phosphorus comes from finite rock reserves that some researchers believe could run critically low within a century, and nitrogen production is among the most energy-intensive industrial processes on the planet.

A technology called forward osmosis, which uses a semi-permeable membrane and a concentrated saltwater solution to pull water out of urine and leave behind a liquid dense with recoverable nutrients, has shown real promise for capturing both. But one persistent problem keeps slowing progress: the filters keep getting dirty.

A new study published in the Journal of Environmental Chemical Engineering is one of the first systematic, multi-cycle investigations of how that fouling happens and what it takes to clean it up. Researchers at the University of Surrey and the University of KwaZulu-Natal ran real human urine through a forward osmosis membrane over three back-to-back 24-hour filtration cycles, testing three different urine preparations head-to-head.

Unlike the pressure-driven filtration used in home water filters or desalination plants, forward osmosis draws water across a membrane using a concentration gradient rather than force, a gentler mechanism that is theoretically less prone to clogging. Human urine, though, contains more than 3,000 compounds, from salts and organic molecules to pharmaceutical residues and microorganisms, and all of them are potential trouble for a thin filtration membrane.

Testing Forward Osmosis Urine Nutrient Recovery Across Three Urine Types

Urine was collected from anonymous donors via dry-diversion toilets in Durban, South Africa, then prepared three ways. One batch was passed through fine filter paper to remove large particles, called filtered hydrolyzed urine, or FHU. A second was stored unfiltered, allowing urea to break down into ammonia and raise the pH above 9, producing unfiltered hydrolyzed urine, or UHU. A third had citric acid added before storage, dropping the pH to around 5 to lock nitrogen in its original chemical form, known as stabilized urine, or SU. Each type ran through the same system for three consecutive cycles, with fresh solutions loaded at the end of each round but no cleaning done on the membrane in between.

Why Particle Load Drives Forward Osmosis Membrane Fouling

By the third cycle, the gaps were stark. Within that final cycle alone in this test, unfiltered hydrolyzed urine had caused roughly an 85% drop in the membrane’s water flow rate; filtered urine caused around 70%; stabilized urine came in at about 68%. Water recovery told a similar story, with filtered urine hitting around 48% by the final cycle while unfiltered hydrolyzed urine fell to roughly 20%. Particle count appeared to be the deciding factor: unfiltered hydrolyzed urine carried around 800 milligrams of suspended solids per liter, while filtered urine had fewer than 100. Those particles accumulated cycle after cycle, building a thickening layer that choked water flow. Stabilized urine’s lower particle load, around 300 milligrams per liter, reflects the citric acid dissolving many mineral deposits that form during natural urine storage.

Despite the fouling, the membrane performed well on nutrient retention, rejecting roughly 90% or more of phosphate, calcium, magnesium, and organic compounds across all three urine types. Potassium and ammonium passed through at somewhat higher rates, roughly 80 to 90% rejection, because their smaller size lets them slip through the membrane’s pores more readily.

What Was Actually Growing on the Membrane

When researchers examined the membranes under an electron microscope after the final cycle, they found very little mineral scaling, the kind of hard crust that typically plagues industrial filters. What was there instead was biological. Both alkaline urine types had seeded the membrane surface with rod-shaped and spherical bacteria. Stabilized urine, being acidic, kept bacteria at bay but introduced a different problem: large oval cells with budding structures consistent with yeast colonization. Chemical analysis identified proteins, uric acid, and creatinine as the dominant compounds in the fouling layers across all three types.

Restoring the membranes required different approaches depending on how dirty they had gotten. Filtered urine membranes recovered nearly all of their original flow rate through a high-speed water flush alone. Membranes fouled by unfiltered and stabilized urine needed a follow-up 20-minute soak in dilute bleach, recovering 91% and roughly 98% of original flow respectively. Even so, electron microscope images after cleaning showed residual deposits clinging to every membrane. Some fouling bonds permanently, and recovery numbers alone do not tell the whole story.

Forward Osmosis Urine Nutrient Recovery and the Fertilizer Gap

Forward osmosis achieved up to roughly 65% water recovery with filtered urine in the first cycle, nearly tripling the nutrient concentration in the remaining liquid and making downstream fertilizer production far more practical. What stands between that result and real-world scale, this study makes plain, is preparation. Simple prefiltration made the single largest difference in membrane performance and cleaning ease across every cycle tested. Getting the urine chemistry right before it ever touches a membrane, whether by filtering out particles or stabilizing the pH with acid, is not a secondary consideration. It appears to be the deciding factor in whether this technology can move from lab to field.

---