Nanofiltration is a pressure driven membrane process that sits between reverse osmosis and ultrafiltration in the filtration spectrum, rejecting most multivalent ions such as calcium, magnesium, and sulfate along with a large share of dissolved organics and smaller molecules, while passing a meaningfully larger fraction of monovalent salts such as sodium chloride than a reverse osmosis membrane would at the same feed pressure. An industrial nanofiltration skid is a rack of spiral wound elements in fiberglass pressure vessels, a high pressure pump, an antiscalant and cleaning package, and controls that together deliver partial softening, sulfate and organics cut, or color removal at pressures typically near half of what brackish reverse osmosis demands. DuPont and Hydranautics both catalog industrial nanofiltration element families built on this principle, and this page treats the complete skid as the purchasing object.
This page is written for the engineer specifying a nanofiltration system for boiler and cooling makeup polishing, food and dairy concentration, dye and textile process water, or produced water sulfate removal. It works through what the membrane does and does not reject, how molecular weight cutoff and divalent selectivity govern sizing, which element classes and skid configurations fit which duties, the parameters that appear on vendor data sheets, and the selection factors and RFQ language that decide whether the installed skid holds its rejection and flux for years.
The decision concentrates in five places: the feed water's hardness, sulfate, and organics profile, the element class whose rejection profile matches the target water, the design flux and recovery the duty will sustain without scaling, the pretreatment that keeps the feed compatible with a spiral wound element, and the cleaning regime that holds flux for a stated membrane life. Get those right and the vessels, pump, and chemical sizes follow.
Chapter 1 / 06
What Nanofiltration Is and What Is Being Decided
The industrial machine this page describes is a nanofiltration skid. Its working element is the spiral wound module: a flat sheet membrane wrapped around a permeate collector tube with feed spacer mesh between the leaves, rolled into a cylinder, and loaded six or seven deep in series inside a fiberglass reinforced plastic pressure vessel. DuPont's FilmTec lines including the NF270 and NF90 classes and Hydranautics' ESNA and HYDRACoRe families are representative catalogs, with element diameters of 4 and 8 inches and active areas published per element for design. Feed enters the vessel end, crosses the element stack axially, and permeate spirals inward to the central tube while the retained hardness, sulfate, and organics leave in a concentrate stream at a controlled recovery.
What is being decided when a nanofiltration skid is purchased is a rejection profile plus a productivity guarantee. The buyer is purchasing permeate with most hardness and sulfate removed, often above 90 percent for divalent ions, with a defined and usually partial passage of sodium chloride that can run from roughly 10 to 70 percent depending on element class, and with organics above the molecular weight cutoff largely removed, all at a design flux the elements sustain between cleanings for a stated membrane life of three to seven years. Everything on the vendor data sheet, salt passage, magnesium sulfate rejection, flux, maximum pressure, area per element, exists to support those guarantees, and the honest vendor quotes rejection at stated feed concentration and pH rather than a single flattering number.
Nanofiltration occupies a defined band between its neighbors and the boundaries are the first thing to get right. It is not reverse osmosis: an RO membrane is tighter, rejects nearly all monovalent ions as well, and is the desalination workhorse, while nanofiltration deliberately passes more sodium chloride and runs at lower pressure. It is not ultrafiltration, which rejects particles and microorganisms through a porous barrier but passes dissolved ions almost completely. It is not ion exchange softening, which exchanges hardness for sodium on a resin until regeneration, and it is not electrodialysis, which moves ions through membranes under a direct current field rather than pressure. For engineers specifying ultrapure chains, nanofiltration also is not electrodeionization or a mixed bed polisher, which polish already desalted water to trace ionic levels.
Different buyers buy the same membrane for different reasons. A power station buys nanofiltration to cut hardness and organics ahead of a reverse osmosis train on a fouling prone surface water, dropping the scaling load on the tighter membrane downstream. An offshore platform buys it for sulfate removal from injection seawater, where keeping barium and strontium sulfate out of the reservoir brine protects injectivity. A dairy buys nanofiltration to concentrate lactose while letting salts pass, a duty closer to process separation than water treatment. A textile dyer buys it to decolorize and reuse rinse water. A municipal reuse plant buys it as an organics and trace contaminant barrier. The specification language differs, but the element physics is shared.
The economics deserve their own paragraph because the skid's first cost is the minority of its lifetime bill. Operating costs are dominated by membrane replacement on the multi year cycle, high pressure pump energy scaled by feed pressure and recovery, antiscalant and cleaning chemicals scaled by the scaling and fouling rate, and the labor of trend watching; against those stand the value of softened water without brine regenerant, a protected downstream RO, or reuse water that avoids discharge. A buyer who compares offers on first cost alone has compared the smallest number in the least important column, and the flux and recovery basis behind each offer matters more than the price on the front page, because an overextended design flux quietly converts into frequent cleanings and early replacement.
A short tour of named duties makes the skid concrete. At a groundwater plant, hard well water passes cartridge guards, a nanofiltration rack, and degassing on its way to boiler makeup with hardness cut more than 90 percent. At a seawater sulfate removal package on a floating platform, open sea water feeds nanofiltration vessels operating near 40 bar with sulfate rejection above 99 percent protecting the reservoir. At a whey processing hall, nanofiltration concentrates lactose while mineral salts pass to the permeate, tailoring the mineral profile of the product. Each duty names the same knobs, element class, flux, recovery, scaling control, and cleaning, with numbers that differ while the spiral wound element barely changes.
Chapter 2 / 06
How the Membrane Does the Work
The separation mechanism is a blend of size exclusion, Donnan charge exclusion, and solution diffusion through a thin interfacial polyamide film a fraction of a micrometer thick on a porous polysulfone support. The membrane carries a weak negative surface charge, so multivalent cations such as calcium and magnesium are repelled more strongly than sodium, and multivalent anions such as sulfate are rejected far more strongly than chloride. That charge selectivity, not pore size alone, is why a nanofiltration element can pass a large share of sodium chloride while holding sulfate rejection near 99 percent, and it is why feed ionic strength and pH shift the rejection numbers: charge screening rises with salinity, and rejection falls with it.
Molecular weight cutoff gives the second lens. Nanofiltration membranes reject neutral organic solutes above a molecular weight roughly between 150 and 300 daltons with high probability, which places sucrose, lactose, dyes, humic substances, and most disinfection byproduct precursors on the retained side while small alcohols and truly dissolved gases pass. Vendor data sheets express this as percent rejection of magnesium sulfate and of a reference organic such as sucrose or a dye challenge, and a specifying engineer should read those as anchors on a rejection curve rather than absolute barriers. Between the ionic rejection anchor and the organics anchor sits the practical promise of the technology: partial desalination with strong softening and color removal in one step.
Pressure drives the flow and the numbers sit clearly below reverse osmosis practice. Feed pressures for brackish nanofiltration commonly run between 3 and 10 bar, specialty sulfate removal duties on seawater reach 40 bar or more, and permeate flux follows the net driving pressure minus the osmotic pressure difference across the membrane. Because a large share of monovalent salts passes, the osmotic burden is smaller than in RO, and that is the energy argument for choosing nanofiltration when the duty only needs divalent and organics rejection. Flux, expressed as liters per square meter of membrane per hour, is set by design at values that respect the feed's fouling tendency, commonly between 12 and 25 for clean well waters and lower for colored or organics laden surface feeds.
Recovery and scaling interact more tightly in nanofiltration than in many membrane duties because the very ions being concentrated, calcium and sulfate, are the classic scale formers. As recovery rises, the concentrate stream enriches in hardness and sulfate until calcium carbonate, calcium sulfate, or barium sulfate exceeds solubility and precipitates on the membrane surface. Design therefore couples a recovery target, commonly 75 to 85 percent for brackish skids, with an antiscalant dose, an acid feed for carbonate control, or both, and with a concentrate staging arrangement that keeps crossflow velocity high enough to sweep the boundary layer. A skid specified without a scaling projection for the actual water analysis is a skid designed for a water it will never see.
Fouling is the operating reality that shapes every cleaning program. Organic color adsorbs onto the polyamide surface, colloidal matter and biofilm compact into the feed spacer, and where iron or manganese is present in the feed they oxidize and stain the lead elements. Skids run with 5 to 20 micrometer cartridge filtration ahead of the high pressure pump as the last line of defense, frequently preceded by a multimedia filter or coagulation on surface waters, and sometimes by ultrafiltration outright on difficult feeds. The cleaning regime then restores what pretreatment misses: a warmed acid circulation for carbonate and iron sulfate scale followed by an alkaline cleaner with detergent for organics and biofilm, typically monthly to quarterly, with membrane life tied directly to how faithfully that cycle is run.
Monitoring closes the loop between design and reality. Normalized permeate flow, normalized salt passage, and differential pressure across each stage, all trended against start of life baselines, tell the operator whether the skid is clean, fouling, scaling, or chemically damaged before the permeate quality visibly moves. A conductivity meter on the permeate header gives the continuous proxy for salt passage, online hardness analysis or periodic laboratory titration tracks the softening guarantee on critical duties, and a turbidity meter on the feed guards the cartridge filters. A nanofiltration train that is not normalized and trended is a skid whose end of life will arrive as a surprise.
Chapter 3 / 06
Element Types, Skid Configurations, and Duty Boundaries
Element classes divide by rejection character rather than by hardware, and choosing among them is the first design act. Loose nanofiltration classes, represented by elements near the NF270 type, emphasize high flux and high organics rejection with moderate divalent rejection and substantial monovalent passage, fitting color removal, organics cut, and partial softening. Tight classes near the NF90 type reject a far larger share of monovalent salts alongside near complete sulfate rejection, approaching loose reverse osmosis behavior and fitting duties where partial desalination plus softening is wanted in one step. Specialty sulfate removal elements built for oilfield injection water push sulfate rejection above 99 percent at seawater pressures. The wrong class for the duty either wastes energy pushing salt the buyer wanted kept or fails the softening guarantee outright.
Element class
Typical MgSO4 rejection
NaCl passage
Typical duty
Loose NF, high flux
95-97%
40-70%
Color and organics removal, partial softening
Medium NF softening
97-99%
20-45%
Boiler and cooling makeup softening
Tight NF
99%+
5-20%
Partial desalination plus softening
Sulfate removal specialty
99%+ SO4
seawater duty
Offshore injection water sulfate control
Dairy and process NF
lactose retained
salts pass
Lactose concentration, mineral tailoring
Skid configuration follows standard brackish reverse osmosis practice because the hardware is shared. A two stage array with vessels arranged 2 to 1 concentrates feed to design recovery while keeping crossflow velocity in the tail vessels above the fouling threshold, and larger duties stack parallel trains behind a common feed header and shared cleaning connections. Vessels hold six or seven elements of 40 inches length, interconnecting through standard couplings, and the rack carries feed, concentrate, and permeate headers with sampling ports at every vessel so lead and tail performance can be separated. Skids ship as factory assembled blocks on stainless or coated steel frames with the high pressure pump, motor starter, instruments, and cleaning connections pretested, which shortens site work to plumbing and power.
Duty boundaries keep the technology honest. Nanofiltration does not disinfect as a stand alone barrier the way ultrafiltration does, so microbiologically unsafe feeds still need downstream ultraviolet treatment or a uv sterilizer step where the product demands it. It does not remove low molecular weight organics below roughly 150 daltons reliably, nor dissolved gases, nor silica beyond a modest share. It cannot soften a brine whose osmotic pressure approaches the available pump pressure, and it is the wrong tool when total dissolved solids removal, not selective rejection, is the goal, which is the reverse osmosis case. Within those boundaries, it replaces lime softening, cuts chemical softening sludge, and pairs naturally with downstream polishing such as electrodeionization when higher purity is needed.
Where nanofiltration sits in a treatment chain matters as much as the elements. Ahead of a two pass reverse osmosis train on a colored surface water, a loose NF stage lifts the organic load off the RO membranes at lower pressure. Ahead of an ion exchange demineralizer or an electrodialysis reversal package on hard well water, nanofiltration cuts the regenerant load and the sludge. Behind a clarifier and multimedia filter on a river intake, it polishes turbidity precursors and organics before the high pressure pump. And where a water softener already exists for trace polish, nanofiltration reduces the salt the softener consumes rather than replacing it outright, an arrangement procurement should price as a chain rather than as isolated boxes.
Chapter 4 / 06
Key Specification Parameters
The data sheet parameters fall into three groups: rejection anchors, productivity anchors, and mechanical limits. Rejection anchors are the magnesium sulfate and sodium chloride rejection figures quoted at a stated feed concentration, pressure, temperature, and recovery, together with an organics anchor such as sucrose rejection for some classes. Productivity anchors are tested permeate flow per element at the reference condition and active membrane area. Mechanical limits are maximum feed pressure, maximum feed temperature, pH range for continuous operation and for cleaning, free chlorine tolerance which for polyamide membranes is effectively zero, and maximum turbidity or silt density index of the feed. A quote that omits the reference conditions behind its rejection numbers is not comparable to one that states them.
Parameter
Typical brackish NF range
Design note
Element diameter
4 or 8 inch
8 inch dominates large skids
Active area per 8 inch element
37-42 m2
Sets element count per vessel
Test flux reference
per catalog PDS
Compare like for like only
Feed pressure, brackish
3-10 bar
Seawater sulfate duty far higher
MgSO4 rejection
95-99%+
Anchor of the softening guarantee
NaCl passage
5-70%
Class dependent, salinity dependent
Design recovery
75-85%
Coupled to scaling projection
Feed pH continuous
3-10
Cleaning pH wider, time limited
Free chlorine tolerance
0
Dechlorinate ahead of elements
Design flux and recovery deserve explicit numbers in every specification because they size the skid and predict its cleaning cycle. Clean well water feeds sustain design fluxes toward the upper end of the 12 to 25 liters per square meter hour band with cleanings a few times a year; colored surface feeds and food streams run lower flux or accept more frequent cleaning. Recovery is bounded on the concentrate side by the scaling projection, with calcium carbonate, calcium sulfate, and barium sulfate saturation indices evaluated at the tail element conditions, and on the hydraulics side by the minimum concentrate flow per vessel that keeps crossflow scour alive. Specifying a flux and recovery without the water analysis that supports them transfers the scaling risk from designer to operator.
Materials and mechanical details round out the sheet. Element construction is polyamide thin film on polysulfone support with a feed spacer, product tube, and outer wrap rated for the stated pressure; vessels are fiberglass reinforced plastic with end caps rated to the skid maximum; the high pressure pump is typically a multistage centrifugal or positive displacement unit on small skids with an inverter drive for soft start and turndown. Piping is uPVC or stainless steel below the high pressure boundary and stainless at pressure. Instruments at minimum are feed and permeate conductivity, feed pressure, interstage pressure, concentrate flow, permeate flow, and temperature, because normalization without temperature and flow is theater.
Standards and certifications appear selectively and the honest specification cites them where they apply. Drinking water component certifications govern elements and vessels on potable duties in many jurisdictions, and food contact regulations govern dairy and process elements. NSF/ANSI 61 style listings for materials in contact with potable water and NSF/ANSI 419 style integrity and rejection claims are the relevant named references for municipal potable reuse contexts. Performance test methods follow ASTM D4194 for water and salt permeation of RO and NF elements and ASTM D6908 style practice for verifying field performance by normalization, and an RFQ should demand rejection verification by a stated method at a stated condition rather than by brochure.
Chapter 5 / 06
Selection Decision Factors and RFQ Language
Selection starts from the water analysis and the target water, written as ions rather than as a vague hardness number. The designer needs calcium, magnesium, sodium, potassium, bicarbonate, carbonate, sulfate, chloride, fluoride, nitrate, silica, iron, manganese, total organic carbon, conductivity, pH, and temperature across seasons, because rejection projections shift with ionic strength and the scaling projection shifts with every number. The target water is then expressed the same way: permeate hardness below a stated value for boiler makeup, sulfate below a stated value for injection water, color or organics removal by percentage. Between those two statements, element class choice, staging, recovery, and pump pressure follow with little freedom left, which is the sign of a well posed problem.
Duty driver
Favored configuration
Principal trade
Partial softening, no brine plant
Medium NF, 2-stage 2:1
5-10 bar pump, modest pass of NaCl
Organics and color cut ahead of RO
Loose NF single pass
Extra stage cost vs RO protection
Sulfate removal, offshore injection
Specialty NF, high pressure
Energy and materials at 40 bar class
Lactose or dye concentration
Sanitary NF, CIP built in
Cleanability vs plain industrial skid
Trace polish after NF
EDI or mixed bed downstream
Chain cost vs single step purity
Pretreatment selection is the second decision and it is decided by the feed, not by preference. Well waters with iron above the element limit need oxidation and filtration or sequestration chemistry evaluated against the membrane; surface waters with turbidity and organics need coagulation plus a multimedia filter at minimum, often ultrafiltration ahead of the cartridge guards; seawater sulfate packages need robust backwashable filtration plus cartridge polishing because cleaning access offshore is expensive. Cartridge filtration at 5 to 20 micrometers immediately ahead of the high pressure pump is universal, with a activated carbon filter or chemical dechlorination wherever the upstream disinfection leaves oxidant, because polyamide nanofiltration membranes tolerate essentially no free chlorine and oxidation damage is the most common irreversible failure.
Commercial comparison is only fair on a normalized basis. Quotes should be forced to state element class and manufacturer, number of elements and vessels, design flux, design recovery, projected permeate quality ion by ion, specific energy at design point, antiscalant and cleaning chemical consumption, and the membrane replacement allowance over a stated life. The cheapest skid on the front page is very often the one running the highest flux on the fewest elements, and its life cycle cost lands higher through cleaning frequency and early replacement. Where energy dominates, a energy recovery device rarely pays on brackish pressure but the pump selection and inverter turndown do; where chemicals dominate, the antiscalant dose rate projection belongs in the comparison table next to the price.
The RFQ language that returns comparable offers states the feed and target explicitly. State feed TDS and the full ion analysis with seasonal range, hardness and sulfate as ions, temperature range, target permeate quality naming which ions or organics must be rejected and to what level, required recovery or acceptable concentrate volume, pretreatment already in place or to be included, turndown requirements, and the standards regime that applies, drinking water, food contact, or industrial only. A complete RFQ paragraph reads: hard well water at 620 ppm hardness as calcium carbonate, sulfate at 310 ppm, 18 degrees C, permeate hardness below 40 ppm as calcium carbonate for boiler makeup, 80 percent recovery target, cartridge guards and dechlorination included, antiscalant dosing to be quoted with consumption.
Integration questions close the selection. Where does the concentrate go, and does the discharge permit accept it; nanofiltration concentrate is a concentrated hard brine and on small sites it can overload a septic field. Where does the permeate go, and does downstream demineralization, an electrodeionization train, a mixed bed polisher, or a water softener for trace polish, need to see stable or variable permeate. Who owns the normalization data and the cleaning schedule, and is the control system set up to trend normalized flow and salt passage with alarms. These questions cost nothing to ask before purchase and everything to ask after startup.
Chapter 6 / 06
Installation, Standards, Failure Modes, and Maintenance
Installation follows brackish membrane practice. Skids arrive factory tested and need a level pad, feed and drain connections, three phase power to the pump and controls, and a commissioning sequence that flushes preservative, loads and soak starts elements per manufacturer instruction, verifies conductivity instrument calibration, and records the first normalized baseline within the first days of operation. Feed water should reach the skid within the stated temperature window, and the cartridge housings must be upstream of the high pressure pump with differential pressure gauges across them so loading is visible. A commissioning report that records start of life normalized flow, salt passage, and differential pressure per stage is the anchor for every warranty claim that follows.
Failure modes arrive in a known order and each has a signature. Scaling, usually calcium carbonate in the tail stage, shows as rising differential pressure and falling normalized flow with permeate conductivity rising late; organic fouling shows as falling normalized flow at stable differential pressure; biofouling shows as both rising together with a feed pressure creep; oxidation damage from chlorine exposure shows as a step jump in salt passage that no cleaning reverses; and telescoped or crushed elements from poor start up or pressure spikes show as sudden flow maldistribution and O ring damage signatures. Diagnosing from the trend data before pulling elements is the difference between a cleaning event and a replacement event, and the failing skid almost always tells the operator which it is.
Symptom
Likely cause
First response
Rising dP, tail stage first
mineral scale
Acid clean, review antiscalant
Falling normalized flow, stable dP
organic fouling
Alkaline detergent clean
Flow and dP both creeping
biofilm
Biocide compatible clean, check pretreatment
Salt passage step jump
oxidative damage
Find chlorine source, replace elements
Permeate hardness drift up
element or O ring bypass
Probe vessel, replace seals or element
Cleaning practice is planned, not improvised. The standard regime is a low pH circulation with citric or hydrochloric acid based cleaner at warm temperature for carbonate and iron scale, followed by a high pH cleaner with detergent and, where biofilm is present, an approved biocide contact, each followed by flush to drain until permeate conductivity returns. Cleaning frequency is dictated by the normalized trend, commonly triggered at a 10 percent flow loss or 15 percent differential pressure rise from baseline, and overcleaning with excessive temperature or pH shortens element life nearly as fast as undercleaning. Elements should be replaced on projection rather than failure, staged lead to tail or in matched sets per vessel, and the replacement should reset the baseline documentation rather than inherit an unrecorded history.
Long term stewardship is what the purchase actually bought. Quarterly normalization review against the commissioning baseline, annual projection of membrane replacement on the observed fouling rate, calibration of the conductivity and flow instruments on a schedule, and a freeze on uncontrolled chemical changes upstream, particularly oxidants and new antiscalant brands, keep the guarantees alive. Sites that treat the skid as a sealed box drift into higher energy at fixed output and blame the vendor; sites that trend it catch a fouling feed change in weeks and spend a cleaning instead of a rack. The difference is procedure, not hardware.
Monitoring and consumables deserve one closing paragraph because the polish chain the skid feeds sets its own discipline. A cartridge guard in the precision filter class after the rack protects sensitive downstream steps from any fines shed on start up, a bag filter or industrial cartridge stage catches coarser debris in the feed line where pretreatment is minimal, and a ph meter on the feed and permeate guards both the scaling projection and the membrane's cleaning chemistry, since rejection shifts measurably with pH. These are small instruments and small filters, but they are the witnesses that make the guarantees arguable: the hardness analyzer proves the softening claim, the conductivity meter proves the salt split, and the trended pH explains both when they move. Plants that instrument the whole chain settle warranty questions in an afternoon; plants that instrument only the permeate header settle them in a lawsuit.
FAQ
What is the difference between nanofiltration and reverse osmosis?
Nanofiltration uses a looser thin film membrane that rejects most multivalent ions such as hardness and sulfate and much of the organics above roughly 150 to 300 daltons, while passing a larger share of monovalent salts, and it runs at lower feed pressure, commonly 3 to 10 bar on brackish water. Reverse osmosis rejects nearly all dissolved ions and is the choice for true desalination. Choose nanofiltration when the duty is softening, sulfate or organics removal rather than total dissolved solids reduction.
What rejection should I expect for hardness and sulfate?
Typical industrial nanofiltration elements reject magnesium sulfate in the 95 to beyond 99 percent band depending on element class, feed salinity, and pH, with hardness as calcium and magnesium removed accordingly. Sodium chloride passage is class dependent and can run from under 10 percent on tight classes to above half on loose high flux classes. Honest quotes state rejection at a named feed concentration, pressure, temperature, and recovery rather than as a single uncoupled number.
What feed pressure does a nanofiltration skid need?
Brackish softening and organics duties commonly run between 3 and 10 bar feed pressure, well below the 10 to 20 bar typical of brackish reverse osmosis, because much of the monovalent salt passes the membrane and osmotic back pressure is smaller. Offshore injection water sulfate removal is the exception, with specialty elements operating at seawater pressures near 40 bar and above to hold sulfate rejection above 99 percent.
How much pretreatment does nanofiltration require?
Every skid needs cartridge filtration at 5 to 20 micrometers immediately ahead of the high pressure pump, plus oxidant removal where disinfectant carries over, since polyamide membranes tolerate essentially no free chlorine. Surface waters typically need coagulation and a multimedia filter and often ultrafiltration ahead of the cartridges; well waters with iron or manganese need oxidation plus filtration or sequestration review against the membrane.
What recovery can I design for?
Brackish nanofiltration skids commonly recover 75 to 85 percent of feed as permeate, but the ceiling is set by a scaling projection on the actual water analysis, evaluating calcium carbonate, calcium sulfate, and barium sulfate at tail element conditions with the proposed antiscalant and pH control. Designing recovery without that projection is the most common route to an early scaling failure.
How often do nanofiltration membranes need cleaning and replacement?
Cleaning is trend triggered, commonly when normalized flow falls about 10 percent or stage differential pressure rises about 15 percent from baseline, which lands between monthly and quarterly on most duties using an acid cycle for scale and an alkaline detergent cycle for organics and biofilm. With competent pretreatment and faithful cleaning, elements last three to seven years; chlorine exposure and neglected scaling are the two things that cut that life short.
Can nanofiltration replace a water softener?
It replaces most of the softening load, cutting hardness typically above 90 percent without brine regenerant or salt deliveries, and it adds organics and sulfate removal a softener cannot do. Trace hardness polish, for example high pressure boiler feed, is sometimes still finished by ion exchange downstream, but at a fraction of the regenerant consumption. Sites should price the chain, nanofiltration with or without a small downstream water softener, against the traditional softener with its salt handling and brine discharge.