Reverse Osmosis Desalination: What Changes When the Feed Is Seawater

Every so often a contractor who runs brackish RO gets asked to quote a seawater job — a coastal processing plant, a resort on well water that turned out to be tidal, an aquaculture operation that wants to stop trucking in fresh water. The instinct is that it's the same machine with a bigger pump. Same vessels, same 8040 elements, crank the pressure, done.

It is not the same machine. The pump is different, the elements are different, the piping schedule is different, the recovery you can hope for is roughly half, and the electric bill becomes the dominant line item in the operating budget. A brackish skid pushed to seawater duty will destroy itself, and a seawater system designed like a brackish one will produce water you can't use.

The physics driving all of that fits in one number: osmotic pressure. Everything about a seawater system is a consequence of it.

Osmotic pressure is the whole game

Open-ocean seawater runs 32,000 to 40,000 mg/L TDS. The quick rule — about 1 psi of osmotic pressure per 100 ppm of TDS — pencils seawater at 35,000 ppm out near 350 psi; the measured value is closer to 390 psi (27 bar) before the membrane has produced a drop. A brackish feed at 2,000 ppm starts at about 20 psi. That gap is the whole story.

And the feed number is the easy end. As permeate leaves, the concentrate gets saltier. Run 45% recovery on a 35,000 ppm feed and the tail-end concentrate is around 64,000 ppm — roughly 700 psi of osmotic pressure. Your feed pressure has to clear that with enough left over to actually drive water through the membrane. Which is why seawater systems run 800 to 1,000 psi feed pressure on hardware rated to 1,200 psi, while a comparable brackish system cruises at 150 to 400 psi.

That same math caps recovery. Brackish systems routinely run 70 to 85% recovery and hit their ceiling on scaling — sparingly soluble salts crashing out of the concentrate, the problem covered in the brackish recovery article. Seawater systems almost never get that far. Single-pass SWRO typically runs 35 to 50% recovery, most commonly 40 to 45%, because past that point the osmotic pressure at the tail end eats your entire driving force. You're pressure-limited, not scaling-limited. Half the seawater you pump goes back to the ocean as concentrate, and the intake, pretreatment, and pumps all have to be sized for that.

Seawater elements are a different build

A seawater element looks identical to a brackish one on the loading dock — same 8-inch by 40-inch spiral-wound package, same vessel. The differences are in the membrane chemistry and the pressure rating, and they show up in the numbers on the data sheet.

Rejection is the one to stare at. On a 300 ppm city feed, the difference between 99.5% and 99.8% rejection is a rounding error. On a 32,000 ppm feed it is the difference between water you can use and water you can't. At standard seawater test conditions, 99.7% rejection leaves about 96 ppm in the permeate; 99.85% leaves about 48 ppm. Same feed, double the salt passage, and every downstream spec — boiler makeup, product water blending, boron — inherits it.

Test conditions matter when you read these: 32,000 ppm NaCl at 800 psi and 25°C, with recovery at 8% for the NanoH2O and FilmTec elements and 10% for the Hydranautics. Element-level test numbers, not system predictions — individual elements vary ±15% on flow. But they're measured on the same yardstick, and on that yardstick the NanoH2O SW 400 R (formerly LG Chem / LG NanoH2O) posts the highest salt rejection of the three at 99.85%, with 12.5% more permeate flow than the FilmTec workhorse. That's the thin-film nanocomposite chemistry doing its job: nanoparticles embedded in the polyamide layer buy flux without giving back rejection.

Street price for the SW 400 R is running about $784 an element from US distributor stock right now — the same neighborhood as premium brackish elements, which surprises people who expect seawater membranes to cost double.

Energy is the bill that never stops

Membranes are a line item every five years or so. Power is a line item every month, and in seawater service it is 40 to 60% of operating cost. Pressurizing every gallon of feed to 800-plus psi and then throwing half of it away as concentrate is expensive by construction.

A well-designed SWRO plant lands somewhere between 2.5 and 4.5 kWh per cubic meter of permeate — call it 9.5 to 17 kWh per 1,000 gallons. At $0.12/kWh that's $1.14 to $2.04 per 1,000 gallons in electricity alone, before chemicals, membranes, or labor. A brackish system making the same water typically burns a fraction of that, simply because it starts 300-plus psi lower.

The single biggest lever is the energy recovery device. The concentrate leaves the membrane array still carrying nearly all its pressure — 900 psi of feed becomes roughly 870 psi of brine — and an isobaric pressure exchanger hands 95 to 98% of that hydraulic energy directly to incoming feed. Without energy recovery the same plant runs 6 kWh/m³ or worse. On anything bigger than a small skid, an ERD is not an upgrade, it's the difference between a viable operating budget and a science project. If you're comparing quotes and one is meaningfully cheaper, check whether the ERD is missing. That's usually where the money went.

Boron: the spec you read after the first permeate test

Seawater carries 4.5 to 6.5 mg/L of boron, and at seawater pH it exists mostly as boric acid — a small, uncharged molecule that slips through a polyamide membrane far more easily than a salt ion. A membrane rejecting 99.85% of NaCl might reject only 90-something percent of boron on the test stand, and less in a warm, high-recovery system.

Run the numbers on the data sheet: 93% boron rejection against a 5 mg/L feed leaves 0.35 mg/L in the permeate at test conditions. Real systems run higher recovery and often warmer water than the test stand, and boron passage climbs with both. The WHO drinking-water guideline is 2.4 mg/L, the EU limit is 1.5 mg/L — but agriculture is the spec that bites. Boron-sensitive crops like citrus and avocado start running into trouble once irrigation water climbs past roughly 0.5 to 1 mg/L. Plenty of SWRO permeate that passes drinking-water specs will quietly damage a citrus grove.

Where the boron spec is tight, single-pass SWRO usually can't get there at neutral pH, because boron rejection is pH-driven: boric acid's pKa sits near 9.2, so raise the pH past 9.5 and it ionizes to borate, which the membrane rejects above 95%. That's why boron-critical plants run a partial second pass — a brackish stage on part of the permeate, dosed with caustic to pH 9.5 or higher. It's also why first-pass boron rejection matters commercially: the higher it is, the smaller and cheaper the second pass, or the better your odds of not needing one.

The intake decides how long your membranes live

The failure pattern specific to seawater work isn't in the membrane array at all. It's upstream. An open seawater intake delivers a live biological soup — algae, bacteria, organics that spike with every bloom and storm — and it never takes a day off. Brackish wells are relatively stable; the ocean is not. Pretreatment that was adequate in March gets overrun in August, differential pressure climbs across the first stage, and the operator responds by cleaning more often until the cleanings stop holding.

The trap is what happens next. The obvious response to biology is chlorine, and plenty of seawater intakes chlorinate to keep the intake line and pretreatment clean. But polyamide seawater membranes carry the same hard limit as every other polyamide element: free chlorine under 0.1 ppm, effectively zero — the chemistry covered in the chlorine article. So the design chlorinates upstream and dechlorinates with sodium bisulfite just ahead of the membranes, and now membrane life depends on a dosing pump and a control loop. When bisulfite runs out on a weekend, or an ORP probe drifts out of calibration, chlorinated seawater reaches the array. Oxidation damage reads as rising permeate TDS with flow intact or improving — the membrane is getting leakier, not dirtier — and no cleaning brings rejection back. On an 800 psi feed, a rejection loss that would be an annoyance on brackish water is fatal, because salt passage you can absorb at 2,000 ppm you cannot absorb at 35,000.

Feedwater discipline is the same as on any RO system, just with the volume turned up: SDI under 5 ahead of the membranes — measured, not assumed — a dechlorination system with an alarm on it, and normalized data tracked from day one so a biofouling trend and an oxidation event don't get mistaken for each other. They present differently in the data and they get treated in opposite ways: one gets a cleaning, the other gets a warranty claim and a changeout.

Practical Bottom Line

Seawater RO is a pressure problem before it is anything else: 800 to 1,000 psi feed, recovery capped near 45%, an energy recovery device as standard equipment, and pretreatment engineered for an intake that changes with the weather. Element choice is where you claw margin back, and on the data sheets the 8040 seawater field has a clear leader: the NanoH2O SW 400 R posts the highest salt rejection of the majors at 99.85% and 93% boron rejection, at street pricing in line with what the market pays for premium elements. FilmTec's SW30 line is the incumbent — the reasons it stays specced are contractor familiarity and installed-base inertia, not the numbers on the test stand. Spec the membrane on rejection, flow, and boron passage per element, and the SW 400 R is where that comparison lands.

Enviro Membranes stocks NanoH2O seawater and brackish elements in the 8040 format — SW 400 R, BW 400 R G2, BW 440 R G2 — ready to ship from US inventory. Planning a seawater system or replacing elements in one? Call us or reach out through enviromembranes.com and we'll help you spec it right the first time.