Reverse osmosis membrane failure rarely begins with the membrane itself. In many cases, the real cause is feed water that was not adequately tested, treated or monitored.For conventional aromatic polyamide RO membranes, the most important baseline conditions are straightforward: keep turbidity below 1 NTU, maintain SDI15 below 5—and preferably below 3—remove free chlorine and other oxidants, control metals and organics, and prevent scale formation at the design recovery. If even one critical parameter is neglected, the result may be rapid fouling, frequent clean-in-place (CIP) cycles, declining permeate flow, higher salt passage or irreversible membrane damage.
This guide explains 13 critical RO membrane feed water quality limits, what happens when they are out of control, and which pretreatment measures are normally used to manage them.
Engineering note: The values below are practical screening targets for standard industrial polyamide RO systems. They are not universal warranty limits. The membrane manufacturer’s current product data sheet, projection software and project-specific water analysis must always take priority.
RO Membrane Feed Water Quality Limits at a Glance
- Feed water temperature: Normally 5–45°C (41–113°F). Low temperature reduces permeate flow, while excessive temperature may damage the membrane element.
- pH during continuous operation: A conservative operating target is pH 3–10; many modern elements allow approximately pH 2–11. Always follow the product-specific data sheet.
- Free chlorine and other oxidants: Non-detectable is the safest target. Oxidants can irreversibly damage the polyamide membrane layer.
- Turbidity: Keep below 1 NTU; ≤0.5 NTU is preferred for reliable long-term operation.
- Silt Density Index, SDI15: Keep below 5 as a maximum and preferably below 3 to reduce particulate and colloidal fouling.
- Total iron: Preferably below 0.05 mg/L at the RO inlet to prevent oxidation and precipitation deposits.
- Manganese: Preferably below 0.02–0.05 mg/L at the RO inlet to minimize persistent manganese oxide fouling.
- Total Organic Carbon, TOC: A conventional screening target is below 3 mg/L to reduce organic fouling and biofilm growth.
- Chemical Oxygen Demand, CODCr: A conventional screening target is below 15 mg/L.
- Biochemical Oxygen Demand, BOD5: A conventional screening target is below 10 mg/L to limit microbial growth potential.
- Oil and grease: Non-detectable is preferred; contamination above 0.1 mg/L should be removed before the RO system.
- Surfactants: Keep non-detectable or within a concentration specifically approved by the membrane supplier.
- Scaling potential: The projected concentrate composition and system recovery must not cause uncontrolled carbonate, sulfate, silica or other scale precipitation.
1. Feed Water Temperature: Normally 5–45°C
Water temperature directly affects viscosity and membrane permeability. As the temperature falls, water becomes more viscous and permeate flow decreases. This is a normal physical effect and should not automatically be diagnosed as membrane fouling.
At higher temperatures, permeate flow generally increases, but salt rejection may decrease. Excessive temperature can also weaken membrane and element materials. Many industrial RO elements specify a maximum feed temperature of 45°C (113°F), but the allowable maximum may change with pH, pressure and membrane type.
The system should therefore be designed at the expected minimum and maximum water temperatures—not only at the standard test temperature of 25°C.
2. pH: Follow the Continuous-Operation Limit
Many modern polyamide RO elements allow a continuous operating pH range of approximately 2–11, while a conservative project target of pH 3–10 is common. The exact range depends on the membrane series, operating temperature and exposure time.
Chemical cleaning may use a wider pH range for a limited time, but the permitted cleaning pH narrows as temperature increases. A cleaning limit must never be treated as a continuous operating limit.
pH also affects scale formation. Raising pH can increase calcium carbonate scaling potential, while changing pH may alter the solubility of iron, manganese, silica and other species. pH control must therefore be coordinated with the antiscalant program and the projected recovery.
3. Free Chlorine and Oxidants: Non-Detectable Is the Safest Target
Standard aromatic polyamide membranes have limited resistance to oxidants. Free chlorine attacks the selective polyamide layer and causes irreversible salt-rejection loss. Ozone, permanganate and other oxidizing chemicals may cause similar damage.
Some membrane data sheets state a feed chlorine concentration below 0.1 mg/L, while others require chlorine to be non-detectable. For that reason, “below 0.1 mg/L” should not be presented as a universal safe limit.
Common dechlorination methods include granular activated carbon and sodium metabisulfite or sodium bisulfite dosing. Online free-chlorine or oxidation-reduction potential monitoring should be considered where oxidant breakthrough is possible. Dechlorination also removes biological protection, so downstream pipework, cartridge filters and stagnant sections must be managed to prevent biofilm growth.
4. Turbidity: Keep It Below 1 NTU
Turbidity is an operational indicator of suspended particles and colloids. RO feed water should generally remain below 1 NTU. For stable long-term operation, a target of 0.5 NTU or lower provides a better safety margin.
High turbidity can rapidly load security cartridge filters, deposit solids on the lead elements and increase first-stage differential pressure. However, low turbidity alone does not prove that the water has low fouling potential, which is why SDI must also be checked.
5. SDI15: Below 5, Preferably Below 3
The Silt Density Index measured over 15 minutes, or SDI15, estimates how quickly suspended and colloidal matter will plug a standard test filter. Most membrane suppliers specify SDI15 below 5 as an operating maximum. In practical industrial operation, especially with surface water or wastewater reuse, SDI15 below 3 is a better pretreatment target.
An SDI value close to 5 may technically meet a maximum limit but still result in frequent cartridge filter replacement, rapid pressure-drop increase and shortened CIP intervals. Trend stability is as important as a single reading.
6–7. Iron and Manganese: Control Them Before They Oxidize
Iron and manganese can remain dissolved under reducing conditions, then oxidize and precipitate when they encounter oxygen, chlorine or a pH change. The resulting hydroxides and oxides can coat membrane surfaces and block feed spacers.
A practical target is total iron below 0.05 mg/L and manganese below approximately 0.02–0.05 mg/L at the RO inlet. Different projects may use different limits depending on metal speciation, dissolved oxygen, pH and the presence of oxidants.
When oxidation and filtration are used for iron or manganese removal, the precipitated solids must be fully removed upstream. Residual oxidant must then be quenched before the water reaches a polyamide RO membrane. Simply measuring dissolved Fe²⁺ in the raw water is not enough; total iron at the RO inlet should be monitored.
8–10. TOC, COD and BOD: Three Different Views of Organic Risk
No single laboratory test fully describes organic fouling potential.
- TOC measures the carbon content of organic compounds. A conventional screening target below 3 mg/L is often used for relatively clean RO feed water.
- CODCr estimates the chemically oxidizable load. A value below 15 mg/L is a useful conventional target, but wastewater-reuse RO systems may operate at higher values when specifically designed for them.
- BOD5 estimates the biodegradable fraction of the organic load. Keeping it below 10 mg/L reduces risk, but even low BOD does not guarantee freedom from biofouling.
These figures are screening targets, not universal membrane limits. The type of organic matter, molecular weight, charge, hydrophobicity and biodegradability can matter more than the total concentration. A sudden increase from the established baseline may be more significant than a single result that remains below a generic number.
11. Oil and Grease: Remove Them Before the RO
Oil and grease are strongly hydrophobic and can adsorb onto the membrane and feed-spacer surfaces. The result may be severe flux loss and a deposit that standard CIP procedures cannot completely remove.
Oil and grease should be non-detectable whenever possible. Where the analytical method provides a numerical result, contamination above approximately 0.1 mg/L should be removed through suitable source control and pretreatment. Depending on the wastewater, treatment may require oil separation, dissolved air flotation, coagulation, activated carbon, ultrafiltration or a combination of processes.
12. Surfactants: Compatibility Matters More Than a Generic Number
Surfactants are not all equally harmful. Their behavior depends on whether they are anionic, nonionic or cationic, as well as on concentration, membrane chemistry and other contaminants in the water.
Cationic surfactants can be particularly difficult because they may strongly adsorb to negatively charged membrane surfaces. Surfactants may also stabilize emulsified oil, making upstream separation more difficult. Feed water containing process detergents should therefore be tested, and the membrane supplier should confirm compatibility. “Non-detectable” is the safest general target when the surfactant composition is unknown.
13. Scaling Potential: There Is No Single Hardness Limit
Hardness, alkalinity, sulfate, silica, phosphate, fluoride, barium and strontium must be evaluated together at the concentrate-side conditions. A feed concentration that appears acceptable can still form scale after the RO system concentrates it several times.
The correct engineering check is not a single hardness number. The complete ionic analysis should be entered into membrane projection and antiscalant software at the proposed recovery, pH and temperature. The design must demonstrate adequate control of calcium carbonate, calcium sulfate, barium sulfate, strontium sulfate, calcium fluoride, silica and any other relevant precipitate.
Pretreatment options include acid dosing, softening, ion exchange, scale-selective removal, antiscalant dosing and reduced system recovery. Antiscalant should support a verified projection; it should not be used to compensate for incomplete water analysis.
What the Failure Pattern Can Tell You
Different changes in RO performance often point to different feed water problems:
| Operating Change | Likely Causes to Investigate |
| Rapid increase in first-stage differential pressure | Suspended solids, colloids, biofilm, iron or manganese deposits |
| Gradual normalized permeate-flow decline | Mineral scale, organic fouling, biofouling or feed-channel blockage |
| Increase in normalized salt passage | Oxidation, chemical damage, membrane leakage, O-ring failure or severe scaling |
| Lower flow only during cold weather | Normal temperature effect; normalize the data before diagnosing fouling |
| Shorter time between CIP cycles | Pretreatment instability, excessive flux or recovery, poor cleaning, or mixed foulants not fully removed |
Membrane condition should be judged with normalized operating data. A change in raw flow or pressure alone can be misleading when temperature, salinity, recovery or permeate backpressure has also changed.
Pretreatment Response by Contaminant
| Feed Water Problem | Common Pretreatment Approach |
| High turbidity or SDI | Coagulation/clarification, multimedia filtration, microfiltration or ultrafiltration, followed by a security cartridge filter |
| Free chlorine or oxidants | Activated carbon or reducing-agent dosing, with verified contact time and online monitoring where required |
| Iron and manganese | Controlled oxidation and filtration, catalytic media or another validated removal process, followed by oxidant removal before RO |
| High TOC, COD or BOD | Source segregation, biological treatment, coagulation, activated carbon, advanced oxidation where appropriate, and membrane pretreatment |
| Oil and surfactants | Source control, oil separation, DAF, coagulation, carbon adsorption and/or UF after treatability testing |
| Scaling potential | pH adjustment, softening, ion exchange, validated antiscalant dosing and recovery optimization |
Frequently Asked Questions（FAQs)
What SDI value is acceptable for an RO system?
SDI15 below 5 is a common membrane operating maximum, but below 3 is a better target for reliable long-term operation. Surface-water and wastewater-reuse systems may require more robust pretreatment and tighter control.
Is free chlorine below 0.1 mg/L always safe for an RO membrane?
No. Some membrane data sheets allow less than 0.1 mg/L, while others require free chlorine and other oxidants to be non-detectable. Always follow the current product-specific limit.
Can an RO system operate when feed water pH is above 10?
Some modern RO elements allow continuous operation up to approximately pH 11, but the permitted range is product- and temperature-dependent. High pH can also increase scaling risk and must be included in the system projection.
Do low TOC and COD values guarantee that biofouling will not occur?
No. Biofouling depends on the type and biodegradability of the organics, microbial activity, nutrient availability, temperature, hydraulic conditions and system hygiene. Trend monitoring is more useful than relying on one organic indicator.
How often should RO feed water quality be tested?
Temperature, pH, conductivity, turbidity, pressure and flow should normally be monitored continuously or during every operating shift. Free chlorine or ORP should be continuously monitored where oxidant breakthrough is possible. SDI should be measured at a frequency appropriate to feed variability—often daily during commissioning and unstable operation. Metals, organic indicators, ions and microbiological parameters should be tested on a scheduled basis and whenever the water source or pretreatment performance changes.
Do these limits apply to every membrane material?
No. This article focuses on conventional aromatic polyamide RO membranes used in industrial water treatment. Cellulose acetate, nanofiltration and specialty sanitary or high-temperature membranes have different chemical and operating limits.
Final Takeaway
RO pretreatment is not complete when the water merely looks clear. It is complete only when the feed water remains within the membrane’s approved operating limits and the system projection confirms that fouling, oxidation and scaling risks are controlled at the design recovery.
The most reliable approach is to establish a feed water baseline, monitor trends instead of isolated test results, and investigate any change before membrane performance deteriorates. Protecting the membrane upstream is always less expensive than repeated cleaning, lost production or premature element replacement.
Need help evaluating RO pretreatment filters or membrane filtration components? Send us your complete raw-water analysis, required permeate capacity, recovery target and operating temperature. Our team can help identify the key pretreatment risks before equipment selection.
For element specifications and industrial desalination applications, explore our RO membrane solutions.
Technical References
- DuPont Water Solutions, FilmTec™ Reverse Osmosis / Nanofiltration Elements Operation Excellence and Limiting Conditions.
- DuPont Water Solutions, FilmTec™ Reverse Osmosis Membranes Technical Manual and cleaning guidance.
- Toray Industries, current RO membrane product data sheets and operating limits.
- Hydranautics, Chemical Pretreatment for RO and NF and RO technical service bulletins.
- CSM, Reverse Osmosis Membrane Technical Manual.
