Silt Density Index (SDI): How the Test Works and the Number Your RO Needs

What the Silt Density Index Measures

The silt density index (SDI) tells you how quickly the fine particles in a water supply will clog a reverse osmosis (RO) membrane. You push the water through a very fine test filter at a steady pressure, then time how much slower it flows after a set period. Clean water barely slows the filter down. Water carrying silt, clay, rust and other colloids plugs it fast, and a faster plug means a higher SDI.

For most RO systems the number you want is below 5, and below 3 is the common target. This guide covers how the test is run, how to do the math yourself, how to read the result against your membrane's limit, and what the test can't tell you. If you're still working out what's fouling a membrane in the first place, start with our guide to the four types of membrane fouling. SDI deals with just one of them: colloidal and particulate fouling.

The test is standardized. ASTM International publishes it as test method D4189, and the summary of its 2007 edition says the SDI "has been empirically correlated with the fouling tendency of some water treatment equipment such as reverse osmosis devices." The same summary scopes it to relatively clear water, around 1.0 NTU turbidity, "such as well water, filtered water, or clarified effluent samples." Muddy raw surface water is outside what the test was built for.

How the SDI Test Works

The setup is simple, and the settings are fixed so that results can be compared. A review of RO fouling prediction methods in the journal Membranes lays them out:

Running the Test, Step by Step

1. Pick the sampling point. To check whether your pretreatment is doing its job, test the water the membrane actually sees, right after the last cartridge filter. To size new pretreatment, test the raw supply too.
2. Fit a fresh test filter in the holder and set the regulator so it holds 30 psi through the whole test.
3. Time the first 500 mL. Open the flow and time how long the first 500 mL takes to collect. Call that t_i.
4. Keep the water running through the same filter at 30 psi for the test interval, 15 minutes for the standard test.
5. Time a second 500 mL. At the end of the interval, time another 500 mL. Call that t_f. It's usually longer than the first, because the filter has been loading up with particles the whole time.
6. Write down the water temperature and the filter brand next to the result. You'll see why in the limits section below.

How to Calculate SDI

The calculation compares the two collection times. The more the second one has stretched out, the more the filter plugged:

Here's a worked example. Say the first 500 mL takes 30 seconds and the second, 15 minutes later, takes 50 seconds:

• 30 ÷ 50 = 0.6
• 1 - 0.6 = 0.4, so the filter plugged 40 percent
• 100 × 0.4 = 40
• 40 ÷ 15 = an SDI₁₅ of about 2.7

That water is under 3. Now say the second sample takes 150 seconds instead. The filter plugged 80 percent (1 - 30 ÷ 150 = 0.8), and the math gives about 5.3. That's over a limit of 5, and it trips a second rule as well.

When a 15-Minute Reading Isn't Valid

A 15-minute result only counts if the filter didn't plug too far. A study of pretreatment at two full-scale seawater RO plants, citing the ASTM method, puts the cap at 75 percent of the maximum possible value, which works out to an SDI₁₅ of 5.0. Above that, you shorten the test to 10 or 5 minutes. If even the 5-minute version plugs past its cap, the authors say "another test should be used," such as the modified fouling index.

There's a practical takeaway hiding in that rule. If a 15-minute test on your RO feed plugs past 75 percent, you don't need the shortened test to make the call. The water is already over an SDI limit of 5, and it needs more pretreatment before it reaches the membrane. You'd still run the shorter test, or the modified fouling index, to see how far over it is when you size that pretreatment.

What SDI Does Your RO Membrane Need?

There's no single number for every membrane. The study above notes that the SDI₁₅ membrane manufacturers recommend "differs depending on the water source and pretreatment." Your membrane's own specification is the number that counts. In general, though, the published guidance lines up like this:

For our own equipment, the operating guide for Crystal Quest® point-of-entry RO systems rated 500 to 2,500 gallons per day sets the feed water limit at an SDI under 5. That's one of several feed limits in the guide. Chlorine, hardness and iron each have their own, and our membrane fouling guide explains what each one protects against.

What SDI Can't Tell You

SDI is a standard check in RO pretreatment for good reason. It's simple, quick and standardized. But it's an empirical index, not a direct count of what's in the water, and the sources behind it are candid about its blind spots:

• It isn't an absolute measurement. ASTM's own summary says so, because the size, shape and nature of particles vary from water to water.
• Temperature changes the result. Readings taken at different water temperatures "may not necessarily be comparable," per the ASTM summary.
• The test filter brand changes the result. ASTM says values obtained with filters from different manufacturers can't be compared. Stick with one brand if you're tracking a trend.
• It misses the smallest particles. The 2022 study points out that SDI doesn't cover most colloids much smaller than 0.45 micron, and that the reading doesn't rise in a straight line with particle concentration.
• It only sees particles. Scale, organic fouling and biofilm need their own checks. The plant study notes there's no standard method yet for biological and organic fouling potential in seawater RO.
• It isn't for already-purified water. ASTM excludes effluent from most RO and ultrafiltration systems.

When SDI isn't enough, the next step is usually the modified fouling index (MFI_0.45), published by ASTM as D8002. The plant study describes SDI and MFI as the only ASTM standard methods for estimating how fast particulate fouling may occur in an RO system.

How to Bring a High SDI Down

ASTM notes that SDI "can be used to determine effectiveness of various processes such as filtration or clarification." That's the real power of the test. Run it before and after each pretreatment step and you'll see exactly which step is or isn't pulling its weight.

The plant study shows what that looks like at scale. At one of the plants, raw seawater tested past the 5-minute cap. After a coagulant dose and the first stage of dual media filtration, the SDI₁₅ averaged 5. A second media stage brought it to 3.4, and the final cartridge filters to 3.0, under the membrane maker's guarantee level of 4 at that plant. Each stage took a smaller bite, and the study reports more than 80 percent SDI removal across the whole train.

The same idea works on a smaller system, with tools sized to match:

• Finer cartridge filtration after the sediment filter is the first lever on reasonably clear well or city water.
• Media filtration ahead of the cartridges handles dirtier supplies, so the cartridges aren't asked to do all the work.
• Iron and manganese removal matters if your water has them. Dissolved iron can slip through a sediment filter, then oxidize into fine particles before it reaches the membrane. Our guide to removing iron from well water covers the options.
• Ultrafiltration is another option on difficult supplies. The plant study lists membrane-based pretreatment such as microfiltration and ultrafiltration alongside conventional media.

One timing tip if you run media filters. In the plant study, the filtrate right after a backwash tested far worse than normal, with SDI₅ readings above 15 during a maturation period of about 30 minutes. A sample pulled right after a backwash can make a media filter look far worse than it normally runs. Readings in that study stayed high even after the maturation period, and the authors suggested the backwash routine or the maturation time needed adjusting. Test at a steady state, note when the last backwash ran, and if the numbers stay high, look at the backwash itself.

If the membrane has already taken a hit from particles, our RO membrane cleaning guide covers what cleaning can and can't recover. For larger systems, the commercial RO sizing guide walks through building a pretreatment train around a full water analysis, with SDI as one input.