Separation-membrane procurement in 2026 is being reshaped by two parallel tracks: a stable polyamide RO and nanofiltration (NF) commodity layer that still carries the bulk of seawater and brackish desalination capacity, and a thinner, higher-margin track of electroactive membranes (EMs) and membrane-electrode modules aimed at fouling control and contaminant destruction in wastewater reuse [S1].
The shift is incremental but measurable. Nanofiltration (NF) membrane development began approximately 20 years after reverse osmosis, with Filmtec in the United States coining the term "nanofiltration" and membrane components being commercialized in the mid-1980s. A review of electrofiltration notes that membrane-electrode architectures can dramatically lower the cell potentials required to drive electrophoresis, electrostatic repulsion, and redox-based fouling control, relative to plate-electrode cells [S1].
RO and NF stock: where 2026 procurement volumes actually sit
The polyamide RO/NF layer remains the high-volume procurement category. NF operates at a trans-membrane pressure of 0.5-2.0 MPa, typically 0.5-3.0 MPa below the pressure required by RO for an equivalent permeate flux, and that pressure delta is the single largest driver of NF substitution in multi-stage reuse trains [S2]. NF materials are not a single chemistry: cellulose acetate, cellulose triacetate, sulfonated polysulfone, sulfonated polyethersulfone, aromatic polyamide composites, and several inorganic compositions are all qualified for industrial elements, which broadens the supplier base compared with the near-monopoly of aromatic polyamide in seawater RO [S2].
For procurement specs, the practical gate is molecular weight cut-off (MWCO). NF elements are commonly specified at 200-1000 Daltons MWCO, with a strongly negative surface charge that drives selective divalent-ion rejection while letting monovalent ions pass, the so-called "loose RO" behaviour first commercialised by the Israeli hybrid-filtration programmes and later rebranded as NF by Filmtec [S2]. Buyers running side-by-side RO/NF cost models in 2026 typically use the 0.5-2.0 MPa operating window as the boundary between "RO-skid replacement" and "NF retrofit", because pump, energy-recovery device, and pipe-class selection all change at that break. The baseline chemistry and operational envelope for RO/NF systems are catalogued on the reverse osmosis reference page and the modified bitumen membrane reference which also tracks polymeric sheet-form membrane classifications used in adjacent civil-engineering applications.
Electroactive and membrane-electrode modules: the 2026 spec front
Electroactive membranes (EMs), both polymeric and inorganic, are now specified where the buyer needs simultaneous separation and electrochemical reaction: biofouling suppression, mineral-scaling control, organic-fouling oxidation, and in-situ destruction of small uncharged contaminants such as NDMA, phenol, acetone, and boron that slip past pressure-driven RO/NF [S1]. The energy case is specific: by using the membrane itself as an electrode in the electrochemical circuit, the cell potential required to drive electrophoresis, dielectrophoresis, electrostatic repulsion, and redox at the membrane/water interface is reduced relative to a plate-electrode cell, which also suppresses electrocorrosion of the metal hardware and limits unwanted byproduct formation [S1].
The trade-off is fabrication cost. EM materials, especially inorganic conducting frameworks, sit at a higher unit cost than commodity TFC polyamide, and most 2026 EM deployments are retrofit kits on existing RO/NF skids or pilot-scale modules in water-reuse and brackish-desalination trains rather than greenfield full-scale plants [S1]. For online monitoring of feed and permeate quality around these modules, procurement teams typically pair the membrane skid with an online water analyzer on the permeate and concentrate streams, since conductivity, pH, and oxidation-reduction potential (ORP) drift are the leading indicators of electrode passivation and membrane fouling. Power conditioning for the EM rectifier and the cleaning-in-place (CIP) loop is normally drawn from a dedicated DC power supply sized for the cell's continuous duty, with the rectifier harmonic profile specified separately from the main plant power supply bus.
Comparison of the three membrane classes specifiers see in 2026 RFQs
The decision matrix for 2026 RFQs lines up along four axes: operating pressure, salt-rejection profile, contaminant-class coverage, and module cost per m^2 of active area. Commodity aromatic polyamide TFC RO is the baseline at 1.0-8.0 MPa operating pressure, very high monovalent rejection, and the lowest module cost per square metre, but it cannot, on its own, destroy small uncharged organics and it fouls on biofilm and scale without chemical pretreatment [S1][S2]. NF elements, including sulfonated polyethersulfone and cellulose-acetate variants, sit at 0.5-2.0 MPa, 90-95% NaCl rejection, strong divalent selectivity, and intermediate module cost, and they let buyers trade monovalent rejection for a 0.5-3.0 MPa pump-energy saving [S2].
Electroactive membranes (polymeric EMs and inorganic conducting composites) add a fourth axis, electrochemical function, that the other two cannot deliver: in-situ oxidation of fouling precursors, electrostatic rejection of charged colloids, and pH perturbation at the membrane/water interface, all at higher module cost and with a rectifier load that has to be specified alongside the membrane itself [S1]. For specifiers writing 2026 tenders, the short-list logic is: choose RO if the feed is seawater and the permeate spec is total dissolved solids (TDS)-driven, choose NF if divalent hardness, sulphate, or trace organics are the target and the plant can absorb partial monovalent passage, and choose EMs only when fouling or recalcitrant organics are the binding constraint and the buyer will accept a pilot-to-full-scale ramp [S1][S2].
Supply chain structure: where the 2026 risk concentrates
The TFC chemistry is concentrated in a small number of Asian and US membrane-element converters, so a single-line upset at a major producer typically moves global spot RO element lead times by weeks; the auxiliary layer, including non-woven polyester or polypropylene substrate, feed spacers, permeate carriers, and the polyurethane or epoxy potting at the element's anti-telescoping device, is sourced from a wider supplier base and tends to recover faster from logistics shocks [S1][S2].
For EM modules the risk profile is different. The bottleneck is the conducting substrate, carbon nanotube coatings, metal-oxide frameworks, doped conducting polymers, and the membrane-electrode contact, and these chemistries are typically sourced from specialty chemical houses rather than the high-volume TFC converters, so scale-up is gated by pilot runs rather than by line capacity [S1]. A short procurement note from 9 August 2026 flagged an on-call controls bid at a Mid-Atlantic water plant that pulled in maintenance spend across the membrane-s skid instrumentation chain, a signal that operations and maintenance (O&M) budgets are the first place 2026 water-sector capex is being reallocated, and a comparable shift is visible in the switchgear procurement signal where the loudest supply-chain chatter is running ahead of the thinnest demand pockets. Adjacent 2026 industrial procurement signals, including the lighting compliance pressure and theatrical fixtures note, show a similar O&M-led pattern.
Selection criteria and the 2026 spec checklist
Specifiers writing membrane purchase orders in 2026 should lock four parameters before opening the supplier short-list: feed-water characterisation (TDS, hardness as CaCO3, sulphate, free chlorine, total organic carbon), permeate target (TDS, specific ion limits, microbiological barriers), operating envelope (trans-membrane pressure 0.5-2.0 MPa for NF, 1.0-8.0 MPa for RO), and module geometry (4040, 8040, or custom spiral-wound; capillary; hollow-fibre; flat-sheet cassette) [S2]. For EM retrofits, the checklist adds a fifth parameter: rectifier sizing, because the cell's continuous-duty current draw has to be specified against the plant's available power supply capacity, and a sixth: monitoring, because EM modules need continuous ORP and pH data from the online water analyzer on the concentrate loop to detect electrode passivation early [S1].
Material choice is the second filter. Cellulose acetate and cellulose triacetate are still specified for low-fouling, low-pressure duties and for plants that want to avoid free-chlorine tolerance issues, even though they trade thermal and chemical stability against the polyamide family [S2]. Sulfonated polysulfone and sulfonated polyethersulfone are the default for high-pH CIP tolerance, and aromatic polyamide composites remain the workhorse for seawater and high-rejection brackish duty. The supporting polymeric sheet membranes used in plant infrastructure, including the civil waterproofing layers, are tracked on the waterproof membrane reference and the modified bitumen membrane reference, and although these are not process membranes, specifiers writing full plant packages are increasingly asked to harmonise polymer sourcing across process and civil scopes. A separate track on mining roller bearing selection and on AGV selection for port logistics shows the same dual-track pattern, with a high-volume commodity layer and a specialty layer running in parallel.
Standards, verification, and 2026 compliance hooks
There is no single ISO or ASTM standard that names "electroactive membrane" as a discrete product class; instead, EM modules inherit the RO/NF element testing regime (typically expressed as rejection and flow at standard conditions) and add an electrochemical verification step that is currently left to OEM datasheets and third-party pilots [S1][S2]. Buyers should require, at minimum, a published NaCl rejection at standard test conditions, a published permeate flow at standard test conditions, a stated active-area in square metres, and a stated maximum operating pressure, with deviations from standard test conditions tied back to the correction curves on the OEM datasheet.
For polyamide RO/NF elements, the long-standing industry convention, first codified in the 1980s commercialisation of Filmtec-style NF, treats 90% NaCl rejection as the boundary between "NF" and "loose RO" or "ultra-low-pressure RO", and 95% as the boundary between "NF" and "tight RO" for procurement classification [S2]. That 90% / 95% gate is the single most useful number a 2026 specifier can write into a tender, because it lines up the chemistry, the operating-pressure window (0.5-2.0 MPa for NF versus higher for RO), and the supplier short-list in a single line item, and it travels cleanly across the cellulose-acetate, sulfonated-polyethersulfone, and aromatic-polyamide families without re-specifying the membrane class [S2].