Yes. Reverse osmosis removes 85% to 95% of nitrate from drinking water under typical residential conditions, and properly maintained systems can exceed that range. The mechanism is physical rejection at a membrane, not chemical neutralization, so the nitrate exits as concentrate rather than disappearing. Performance depends on membrane condition, pressure, pH, and how much nitrate enters the system.
How the membrane rejects nitrate
Reverse osmosis works by forcing pressurized water against a semipermeable membrane. Water molecules pass through; dissolved ions largely do not. The membrane is a thin polyamide film with a dense active layer roughly 0.1 micrometers or less in functional thickness, and rejection happens through two combined effects. Size exclusion blocks larger hydrated ions, while charge repulsion pushes back ions that carry the same negative charge as the membrane surface.
Nitrate carries a single negative charge and is relatively small, which makes it harder to reject than sulfate or calcium. Rejection improves when the membrane surface is more negatively charged, which happens at higher pH. That is why nitrate passage rises at low pH: the membrane loses some repulsive force against an anion that is already small. A plain-language version: the membrane is a charged filter, and nitrate is a small negative ion that can slip through more easily than bigger or more charged ions.
Rejection also depends on recovery, the fraction of feed water converted to permeate. Higher recovery concentrates the feed stream, raising the driving force for ions to cross. Residential systems typically run at 20% to 35% recovery, which keeps concentrate concentration moderate. Pressure matters too: nitrate passage falls as applied pressure rises relative to osmotic pressure, because water flux increases faster than ion flux. A system running at low pressure with a partially clogged prefilter can drop below its rated rejection without any visible failure.
Quantified context: what the numbers mean
The EPA maximum contaminant level for nitrate as nitrogen is 10 mg/L, and the equivalent for nitrate ion is 45 mg/L. These are the same limit expressed two ways, because nitrate nitrogen is about 22.6% of nitrate ion by mass. The EPA also sets a separate limit of 1 mg/L for nitrite as nitrogen. Private wells are not regulated, so testing is the only way to know what enters the system.
Regulatory limits describe finished water, not source water. A system that removes 90% of nitrate from feed water at 20 mg/L as nitrogen leaves 2 mg/L in the permeate, which is comfortably below the limit. The same system fed 100 mg/L leaves 10 mg/L, exactly at the limit, and small performance losses push it over. Rejection percentage alone does not tell you whether the treated water is safe; you need the feed concentration and the measured permeate concentration together.
Rejection rates also vary by membrane type and operating conditions. Thin-film composite membranes generally outperform older cellulose acetate designs for nitrate, but manufacturer-rated rejection is measured under standardized conditions that residential systems rarely match. Rated rejection of 90% at specified pressure, temperature, and recovery can fall to 80% or lower in a cold, low-pressure installation. Temperature matters because cold water is more viscous, reducing flux and shifting the balance between water and ion passage.
Where nitrate comes from and why the margin is thin
Nitrate in drinking water comes mainly from agricultural fertilizer, animal waste, septic systems, and to a lesser extent atmospheric deposition. Shallow wells near cropland or feedlots are the highest-risk sources, and contamination often follows seasonal patterns tied to rainfall and application timing. Nitrate is highly mobile in soil because it does not bind to soil particles, so it moves with groundwater rather than staying near the surface.
The health concern is methemoglobinemia in infants under six months, a condition in which nitrate converts to nitrite and binds hemoglobin, reducing oxygen transport. This is why the limit is set where it is, and why infant formula made with well water is the classic exposure pathway. A counterintuitive detail: boiling nitrate-contaminated water makes it worse, because water evaporates and nitrate concentrates. Boiling is the standard emergency instruction for pathogens, and it is exactly wrong for nitrate.
Municipal systems test and treat, so regulated tap water generally meets the limit at the point of compliance. Private wells have no such oversight, and testing frequency depends entirely on the owner. A single test is a snapshot; nitrate levels can change seasonally and after land-use changes near the well.
Comparison of nitrate removal options
| Method | Typical nitrate removal | Key limitation |
|---|---|---|
| Reverse osmosis | 85% to 95% | Wastes concentrate water; needs pressure and maintenance |
| Anion exchange | 90% to 99% | Regeneration produces high-nitrate brine; sulfate competes |
| Distillation | Greater than 95% | High energy use; slow output |
| Boiling | None, and it concentrates nitrate | Unsafe for nitrate; only for pathogens |
| Carbon filtration | Essentially none | Removes taste and odor compounds, not nitrate |
| Water softener | None | Exchanges hardness ions; adds sodium, leaves nitrate |
The table shows why carbon pitchers and standard softeners are not nitrate solutions. They address different contaminants. Anyone relying on a carbon filter for taste should not assume it touches nitrate.
Edge cases and failure modes
System performance is only as good as the whole chain. A degraded prefilter lets sediment reach the membrane and foul it, which reduces flux and can change rejection. A saturated post-filter or a storage tank can reintroduce contamination if not sanitized. A system that has sat unused for weeks can develop biofilm on the membrane surface, and the first water drawn after stagnation may carry elevated nitrate from the concentrate side.
Red flag: any RO system without a functioning prefilter, since membrane fouling degrades rejection silently. Red flag: a system rated only by a generic "removes contaminants" claim with no nitrate-specific rejection data. Red flag: a permeate pump or booster that has failed, since low pressure raises nitrate passage. Red flag: a storage tank that has never been sanitized, which can harbor microbial growth. Red flag: relying on a single startup test rather than periodic permeate testing on a private well.
Blending is another edge case. Some under-sink systems blend a small fraction of feed water back into the permeate to improve taste and stabilize pH. That blend carries nitrate with it, so a system marketed for mineral retention may have lower nitrate rejection than a pure permeate design. The tradeoff is real and usually undocumented in consumer materials.
Common misconceptions
The biggest misconception is that any filter removes nitrate. Carbon filters and pitcher systems remove chlorine, taste, and some organics but leave nitrate essentially untouched. Another is that RO removes everything, which is false: small uncharged molecules and some volatile organics pass through more readily, and nitrate rejection is good but not absolute.
A third misconception is that rejection percentage alone proves safety. As the quantified section shows, a high percentage applied to a high feed concentration can still leave permeate above the limit. The number that matters is the permeate concentration, measured.
A professional caveat worth stating plainly: certification claims and lab rejection rates are measured on new membranes under controlled conditions. Field performance drifts with fouling, temperature, pressure, and prefilter condition. Treat any single rejection figure as a starting point, not a guarantee, and verify with a permeate test if the source is a private well.
Practical actions and thresholds
Test the source water first. For a private well, test nitrate annually and after any flooding, land-use change, or new septic installation. Use a certified lab, and record results as nitrate as nitrogen so you can compare directly to the 10 mg/L limit. If the result is above 5 mg/L as nitrogen, treat it as a signal to act, because you want margin before you reach the limit.
Size and maintain the system deliberately. Confirm the membrane is rated for nitrate rejection, not just general contaminant reduction. Replace prefilters on the manufacturer's schedule, typically every 6 to 12 months depending on sediment load. Replace the membrane when permeate quality drops or output falls, not only on a calendar. Sanitize storage tanks per manufacturer instructions.
Verify performance, not marketing. Test permeate nitrate after installation and then periodically, especially on private wells. If permeate exceeds 10 mg/L as nitrogen, stop using it for infant formula and investigate the system. For infants, the safe course is to use a tested low-nitrate source until permeate is confirmed below the limit.
Reconcile the thresholds explicitly. The 10 mg/L nitrate-as-nitrogen limit and the 45 mg/L nitrate-ion limit are the same standard in different units. The 1 mg/L nitrite limit is separate and stricter because nitrite is more directly toxic. When comparing a lab report to a regulation, check which unit the lab used; mixing them up produces errors of more than fourfold. A result reported as 30 mg/L nitrate ion is about 6.6 mg/L as nitrogen, which is below the limit, while 30 mg/L as nitrogen is three times the limit.
Limitations and what RO cannot promise
RO does not make nitrate vanish. It concentrates it in reject water, typically 2 to 4 gallons of concentrate per gallon of permeate in residential systems. That reject stream carries elevated nitrate and must be managed; discharging it to a septic system or garden can reintroduce nitrate to groundwater over time. Whole-house RO is uncommon because of water waste and cost, so under-sink point-of-use treatment is the usual approach, which means only the drinking water tap is protected.
RO also removes beneficial minerals and lowers pH, which some users address with a remineralization stage. That stage does not add nitrate, but any blending or bypass does. System cost varies; a residential under-sink unit is a modest upfront purchase with recurring filter and membrane costs, and the concentrate waste is a real operating cost in water-scarce regions.
Finally, RO is a treatment barrier, not a source solution. If a well is contaminated by agricultural nitrate, treatment protects the drinking water but does not fix the aquifer. The most durable fix is source protection, and treatment is the interim or long-term barrier depending on the situation. RO reliably reduces nitrate when the system is specified, installed, and maintained for that purpose, and it reliably fails to do so when treated as a generic filter.