Technical Deep Dive · Engineering Perspective · Published: September 28, 2026
Table of Contents
Key Takeaways
- Seawater reverse osmosis typically operates near 50% recovery, which makes the concentrate roughly twice as salty as the feed. Salinity measurement has to follow the process from inlet to outlet, not sit at one convenient point [1]
- High-recovery brackish systems push brine to 15–30 g/L TDS at 70–85% recovery, and a full zero liquid discharge (ZLD) train feeds its crystallizer at 20–25% TDS [2]
- Inductive (toroidal) conductivity is the preferred measurement physics for desalination and brine service; modern instruments span up to 2,000 mS/cm, territory where contacting cells polarize and foul [3][4]
- Sodium chloride changes conductivity by roughly 2.26% per °C, so an uncompensated 10 °C swing moves the reading by more than a fifth of its value. Temperature compensation is not optional in hot brine lines [5]
- Membrane-based brine concentration can reach about 210,000 mg/L TDS before evaporation takes over, which compresses the entire monitoring chain into the top of the sensor range [6]
- Salinity data should drive recovery, specific energy and scaling decisions. That requires sensors built for the range, plus digital outputs your control system and AI water models can consume.
The Salinity Ladder: One Train, Four Regimes
Brine concentration is not one duty but a sequence of increasingly extreme ones, and the salinity band shifts by orders of magnitude between the first and the last. Understanding the ladder is the first step in specifying sensors that survive it.
Segment 1 — Feed water. A seawater RO plant draws feed at ocean salinity and, at the recovery rates such plants typically run, converts about half of it to permeate. At 50% recovery the concentrate flow is roughly half the feed flow and the salt concentration is about twice the feed concentration [1]. Brackish feeds run lower in absolute terms, but the same doubling logic applies: every point of recovery you add piles the rejected salts into a shrinking liquid stream.
Segment 2 — RO concentrate. Industrial and brackish-water systems routinely run harder than seawater plants. A primary brackish RO stage operating at 70–85% recovery produces a brine at roughly 15–30 g/L TDS [2]. This is where many plants first discover that their feed-water instruments were never specified for the concentrate side: conductivity can be ten to fifty times the permeate value, and the scaling species — calcium carbonate, calcium sulfate, silica — concentrate right alongside the sodium chloride.
Segment 3 — Membrane brine concentration. Downstream of the primary RO, dedicated brine concentration and volume-reduction systems use pressure-driven membranes to squeeze the stream further. Suppliers of such systems report concentration up to about 210,000 mg/L TDS before thermal evaporation becomes the economical next step [6]. The measurement problem changes character here. The solution is corrosive, near the saturation point of several salts, and viscous enough that flow and temperature gradients distort readings.
Segment 4 — Evaporator and crystallizer liquor. True zero liquid discharge ends in a crystallizer, and that machine expects brine arriving at 20–25% TDS [2]. In the evaporator itself the liquor approaches sodium chloride saturation; control hinges on holding supersaturation, temperature and circulation rate inside a narrow window so that salt grows as usable crystals rather than as scale on the heat-transfer surfaces [2].
| Process segment | Typical salinity band | What the sensor must survive | Preferred measurement |
|---|---|---|---|
| RO feed (seawater or brackish) | site-specific; baseline of the ladder | biofouling, marine growth, suspended solids | contacting 2- or 4-electrode, temperature-compensated |
| SWRO concentrate (~50% recovery) | roughly 2× feed salinity [1] | high ionic strength, pressure lines | toroidal (inductive) conductivity |
| Brackish RO concentrate (70–85% recovery) | 15–30 g/L TDS [2] | scaling species: CaCO3, CaSO4, silica | toroidal (inductive) conductivity |
| Membrane brine concentration | up to ~210,000 mg/L TDS [6] | corrosion, high viscosity, near-saturation | toroidal (inductive) conductivity |
| Evaporator / crystallizer liquor | 20–25% TDS feed to crystallizer [2] | crystal nucleation, fouling, heat | toroidal (inductive), cross-checked with density |
Contacting versus Inductive: Picking the Right Physics
Two families of sensors dominate industrial conductivity work, and the salinity ladder mostly decides between them for you.
Contacting (electrode) sensors measure the resistance of the solution between metal electrodes — a classical ohmic measurement. Two-electrode cells are unmatched at the low end: they are the standard for pure and ultrapure water, where sensitivity and low noise matter more than fouling tolerance [3]. Four-electrode contacting designs extend the usable range dramatically, on the order of 1 µS/cm to 2,000 mS/cm, and are a solid choice for wide-range process water and mildly contaminated streams, though the electrodes still contact the medium [3].
Inductive (toroidal, electrodeless) sensors take a different approach entirely. Two wire-wound toroids are encapsulated in a corrosion-resistant body; the analyzer drives one coil, the liquid completes the transformer circuit, and the second coil reads the induced current, which is proportional to conductivity [4]. Because no electrode ever touches the process, coating and scaling cause far less error than on contacting cells. That is why desalination practice has converged on inductive measurement: in brine service the sample is dirty, corrosive or scaling-prone by definition, and the preferred method in desalination processes is inductive conductivity [4].
The trade-offs are symmetric and worth stating plainly. Toroidal sensors resolve poorly below about 50 µS/cm, so they are the wrong tool on the permeate side of a train [3]. Their range ceiling, however, sits comfortably above every segment of the salinity ladder — instruments are commonly specified to 2,000 mS/cm, and nothing in a membrane brine concentrator approaches that limit [3]. A practical architecture therefore pairs contacting cells on the clean, low-conductivity side (permeate, product water) with toroidal sensors on everything at or beyond the concentrate.
There is also a packaging argument. A brine train already needs temperature, pH and often ORP alongside conductivity at each control point. A 4-in-1 multi-parameter probe that carries pH, conductivity, ORP and temperature in a single body removes the alignment and flow-velocity mismatches that come from bolting four separate sensors into one short spool piece. Four parameters. One probe. Zero alignment error.
Fouling and Temperature: The Two Errors That Corrupt Brine Data
Fouling. Every surface in a brine line fouls; the only questions are how fast and how unevenly. On a contacting sensor, a scale layer on the electrodes changes the effective cell constant and silently biases the reading — the instrument still shows a confident number. An inductive sensor degrades far more gracefully, because the measurement field extends through the sample volume rather than across a thin surface layer [4]. Even so, toroidal sensors are not immune: heavy crystal blankets or oil films around the body alter the sampled volume, and open-geometry designs may need on-site verification against a grab sample. The engineering response is placement and access. Install the sensor in a continuously flushed side stream rather than a dead leg, orient it so crystals cannot settle on it, and make it reachable for inspection without scaffolding.
Temperature. Conductivity is strongly temperature-dependent, and sodium chloride is typical of the salt solutions that dominate brine service: its conductivity coefficient is about 2.26% per °C [5]. A brine line that runs 15 °C warmer in summer than in winter will show more than a 30% apparent conductivity change from temperature alone if compensation is naive. Two practical rules follow. First, the temperature element must be fast and co-located — a lagging temperature sensor inside a slug of hot brine produces compensated readings that trail reality by the transport delay of the line. Second, linear compensation that is perfectly adequate at low concentration becomes visibly nonlinear near saturation; at the evaporator end of the ladder, use sample-specific compensation curves verified against laboratory titration rather than a generic 2% per °C setting.
From Measurement to Control: Energy, Recovery, and Scaling
Salinity data earns its keep only when it closes a control loop. Three loops matter most in a concentration train.
Recovery control. Concentrate conductivity is the fastest available proxy for the instant recovery ratio: push recovery up and the concentrate salinity rises almost immediately, long before a lab sample would tell you. Holding the concentrate at its design conductivity — rather than its design flow — lets the train follow changes in feed salinity automatically. The same measurement protects the membranes: a sudden drop in concentrate conductivity relative to feed often signals a membrane integrity problem, while a steady climb at constant recovery signals concentration polarization or the early stage of scaling.
Energy control. In the thermal section, the crystallizer feed concentration is the dominant energy variable. Feed it too dilute and you boil water you should have removed upstream; push it too concentrated and viscosity, boiling-point elevation and fouling all climb together, attacking the heat-transfer coefficient that sets specific energy consumption. Holding the feed in its design band — the 20–25% TDS region cited for crystallizer inlets [2] — is precisely the duty a toroidal sensor in the feed line was built for.
Scaling defense. Between the segments, conductivity trend lines are the earliest cheap warning of scaling onset. A brackish RO concentrate whose conductivity drifts upward at constant feed and recovery is concentrating something extra; combined with pH and temperature from the same multi-parameter probe, the trend gives operators hours of lead time to adjust antiscalant dosing or back off recovery. The alternative — discovering scale on the next pressure-drop trend — costs far more than the sensor that would have caught it.
One honest caveat belongs here: conductivity is a proxy, not a salt assay. Converting mS/cm to TDS uses a factor that depends on the ionic mix, and mixed-industrial brines with sulfates or organics deviate from any sodium-chloride-based table. Treat the conductivity number as the control variable, and calibrate the proxy relationship with periodic laboratory analysis of the actual brine.
What Shanghai ChiMay Brings to Brine Service
Shanghai ChiMay builds the sensing layer for exactly this ladder: in-line conductivity electrodes and toroidal-style inductive sensors rated for high-conductivity, high-fouling service, 4-in-1 multi-parameter probes that add pH, ORP and temperature in the same body, and transmitters with Modbus RTU/TCP digital output so the salinity stream lands in your SCADA, historian, or AI water model without a signal-conditioning detour. See the conductivity electrode product page for the current specification set: Conductivity Electrode.
We take the same position on brine sensors that we take on every analyzer we ship: buy the sensor, own the outcome. We quote 5-year cost, not just unit price — because in brine service, the replacement cycle, the coating behavior and the calibration burden over five years dwarf the sticker on the box.
References
- Salton Sea Long-Range Plan, Appendix G: Investigation of Desalination Methods — California state planning document; seawater RO recovery in the 40–50% range and the resulting concentration of the reject stream (a 50% recovery stage roughly doubles concentrate salinity) — https://saltonsea.ca.gov/wp-content/uploads/2024/03/LRP-Appendix-G-Investigation-of-Desalination-Methods.pdf
- Aguato Insider — “Zero Liquid Discharge: When ZLD Makes Sense and When It Doesn’t” — ZLD stage salinities: primary RO brine at 15–30 g/L TDS (70–85% recovery), crystallizer feed at 20–25% TDS — https://aguato.com/resources/zero-liquid-discharge-zld-industrial
- SpecForge Encyclopedia — “Conductivity Meter” — comparative measurement ranges and fouling behavior of two-electrode, four-electrode contacting, and toroidal sensors (up to 2,000 mS/cm; toroidal low-end limit ~50 µS/cm) — https://www.sourcebyspec.com/encyclopedia/conductivity-meter.html
- Novaspect — “Improved Conductivity” — application note on contacting versus inductive conductivity, including why desalination processes prefer inductive measurement — https://www.novaspect.com/asset/72592
- Yokogawa Electric — Conductivity Analyzer Guide (TI12D08A01-01E) — temperature coefficients of electrolyte solutions; NaCl α ≈ 226 × 10⁻⁴ per °C (≈ 2.26%/°C) — https://web-material3.yokogawa.com/2/144/overview/TI12D08A01-01E.pdf
- STJ EPT — “Brine Concentration & Volume Reduction Systems” — membrane-based brine concentration achieving up to 210,000 mg/L TDS — https://www.stjept.com/brine-concentration-volume-reduction-systems-product
About the author: Written by the Shanghai ChiMay Technical Editorial Team — the engineering group behind Shanghai ChiMay’s in-line conductivity and multi-parameter water quality analyzers, with application work spanning desalination, brine concentration, and industrial water reuse.
