The corpus analysed here contains 2,019 patent families, searched across title, abstract, claims, and description using terms covering electrodialysis, electrodialysis reversal, ED stacks, ion-exchange membranes, bipolar membranes, and selective membranes, further filtered to IPC codes B01D61/44, B01D61/46, and C08J5/22. Patent families — rather than raw document counts — are the appropriate unit because they deduplicate continuation filings and multi-jurisdiction equivalents that would otherwise inflate the count. The 2,019-family figure therefore represents distinct inventive disclosures, not the total number of published documents in the space.

The assignee ranking is rendered from the underlying dataset and reflects family counts across the full 2015–2026 coverage window. The top tier is concentrated — a small group of organisations accounts for a disproportionate share of families — while a long tail of entrants holds one or two families each. Notably, several historically active assignees including Evoqua Water Technologies (Evoqua Water Technologies LLC) recorded zero published families in the most recent year, representing a -100% year-on-year change. Whether this reflects publication lag or deliberate portfolio capping is not determinable from the filing data alone and warrants continued monitoring.

A bipolar membrane (BPM) is a laminated structure comprising a cation-exchange layer and an anion-exchange layer joined at a catalytic interface. Under reverse-bias voltage, water molecules dissociate at the interface and the resulting protons and hydroxide ions migrate outward through their respective layers — effectively splitting water into acid and base without electrolysis gases or added reagents. This capability makes BPMs central to closed-loop acid–base recovery, CO₂ capture, and electrochemical synthesis, all of which are high-priority targets for industrial decarbonisation. The concentration of citation weight around composite solid polymer electrolyte membrane records reflects decades of foundational work on BPM architecture; these documents sit at the heart of the prior-art thicket that any new BPM-focused programme must navigate.

Several sub-areas show materially lower prior-art density relative to the desalination and chlor-alkali core. The fermentation-acid recovery space (mapped to IPC C12P, 106 records) is the clearest example: electrodialysis applied to recovering lactic, succinic, or itaconic acid from fermentation broths sits at the junction of biorefinery economics and membrane engineering, yet citation density is far lower than in the desalination mainstream. Other under-claimed areas include BPM-mediated separation of CO₂ electrolysis products, selective lithium-ion transport membranes for brine concentration, and surface-functionalised antifouling membranes using zwitterionic grafting. In each case, the opportunity lies in application-specific or material-specific claims that are not anticipated by the general polymer composition patents at the top of the citation ranking.

The corpus is genuinely global: the United States leads with 413 records, followed by China (295), the EPO (278), Japan (252), WIPO PCT (157), and Canada (104). This distribution means that a product commercialised in any major industrial market is likely to encounter granted patents in that jurisdiction. An FTO opinion scoped to a single country — typically the US — will structurally miss risk in China, Europe, and Japan, all of which have substantial claim coverage. WIPO PCT filings (157 records) indicate that many assignees are deliberately preserving multi-jurisdictional options before committing to national-phase prosecution, so the risk landscape will continue to expand as those applications enter national phase over the next several years.

Patent applications are typically published 18 months after their priority filing date, which means that filings made in late 2024 and throughout 2025 will not appear in a dataset with a mid-2026 cut-off. The most recent year's published count (14 families) is therefore a structural undercount, not evidence that filing activity has collapsed. The more meaningful signal is the 2024 peak of 126 families: that cohort of applications will shape the claim landscape for the next ten to fifteen years and represents the wave of prosecution activity currently underway at major offices. Analysts should treat the 2024 figure as the current high-water mark and revisit the trend once 2025 filings are fully published.

Yes, and they are easy to underestimate. The C25B electrolytic synthesis sub-cluster contains 313 records filed primarily by chemical-production and energy companies whose primary commercial interest is not water desalination but acid, base, or halogen production. Claims originally written for chlor-alkali or electrochemical synthesis contexts can read on ED membrane configurations used in water treatment or biorefinery applications, creating unexpected infringement exposure. The Solvay–EDF co-assignment pairing (17 jointly assigned families) is an example of exactly this cross-sector dynamic: a specialty-chemicals major and a utility building shared IP around industrial-scale acid–base recovery. Any FTO opinion in the ED membrane space should include a cross-classification search covering both B01D and C25B to capture this category of risk.

Ten co-assignee pairs are documented in the corpus, a relatively sparse network for a 2,019-family field. The two strongest pairings are the POSCO group internal pairing (POSCO Research Institute of Industrialscience with POSCO Holdings, 19 families) and the Solvay–EDF cross-sector pairing (17 families). The POSCO pairing represents coordinated intra-group IP strategy — the families are jointly assigned to channel both research and commercial rights within the corporate group. The Solvay–EDF pairing is more strategically revealing: it signals that at least two major European industrial organisations have already identified BPM-based acid–base recovery as a shared commercial target and have built a joint IP position around it. The overall sparseness of the co-assignee network suggests that most players in this field are prosecuting independently, which typically correlates with more aggressive enforcement postures and fewer pre-existing cross-licensing structures.