How Sedigraph Particle Size Analysis Actually Works in the Lab
The Sedigraph is a gravity sedimentation particle size analyzer. It measures how fast particles settle in a liquid and uses Stokes' law to convert those settling velocities into particle diameters. The machine has been around since the 1960s, originally from Mitchell Instruments, and is still used in ceramics, paints, and foundry work. I'm not going to walk you through every button press. I'll tell you what actually matters when you're running one, and where people mess up. The basic principle is straightforward. You take a sample, disperse it in a liquid medium, pump the suspension into the instrument's chamber, and let gravity do the work. As particles settle, they pass through an X-ray beam. The beam detects how much material is at each depth over time. That gives you a density profile, which translates into a size distribution. Most labs run this on samples between 0.5 microns and around 100 microns. Anything smaller and Brownian motion becomes a problem. Anything larger and the run takes forever. Here's the part nobody puts in the manual. The dispersion step is where most results go wrong. A standard Sedigraph run takes about 45 minutes to an hour depending on your sample. But if your particles are flocculated before they even enter the chamber, you're measuring agglomerates, not individual particles. I spent three days once trying to figure out why a silica batch showed bimodal distribution with peaks at 3 microns and 40 microns when the source material was known to be monomodal around 3 microns. Turns out the sample was just sitting in the dispersant too long before injection. The ultrasound broke up the loose flocs initially, but over 20 minutes of standing, they re-aggregated. Shortened the post-dispersion hold time to under five minutes and the bimodal artifact disappeared.
Another thing: the density of your sample has to be accurate. Stokes' law depends on the difference between particle density and fluid density. If you're using a nominal density from a datasheet and it's off by even 0.1 grams per cubic centimeter, your entire size distribution shifts. I learned this the hard way with a titania sample. The catalog listed 4.23 g/cm³, but my pycnometer reading came back at 3.98. That 6 percent error compressed the whole curve toward smaller sizes. Retuned the calculation with the measured density and the d50 jumped from 8.2 microns to 10.1 microns. That's a meaningful difference when you're formulating a glaze. Let me walk through the actual process. You need a wetting agent in your dispersant. Water alone won't penetrate fine powders properly. I use a solution of sodium hexametaphosphate at about 0.5 percent by weight. Add the sample to the dispersant gradually while stirring. Then sonicate. Most labs use 5 to 10 minutes of ultrasonic dispersion. Don't overdo it. Some materials, especially fragile or layered ones like clays, will break down under prolonged sonication and give you a falsely fine distribution. I once sonicated a kaolin sample for 30 minutes thinking more energy meant better dispersion. The D90 dropped from 12 microns to 4 microns compared to a 5-minute sonication. The XRD confirmed structural breakdown. Five minutes was the right call. After dispersion, degas the slurry. This means running the mixture under vacuum for a few minutes to pull out air bubbles. Air bubbles are the enemy. They rise instead of settle and create noise in the X-ray signal. I've seen whole runs ruined because someone skipped degassing and the bubbles passed through the beam at irregular intervals, showing up as spurious fine particles in the distribution.
Now for the actual run. Load the slurry into the instrument. Set your sampling interval. The Sedigraph typically takes measurements every 30 seconds to a couple of minutes depending on your expected particle size range. Finer particles need shorter intervals. Set the density values correctly. Start the run and wait. While it's running, you should periodically check the stability of your baseline. The empty chamber reading should stay consistent. If it drifts, you've got a temperature issue or a detector problem. Temperature control matters more than most people realize. The viscosity of the dispersant changes with temperature, and viscosity is a direct input to Stokes' law. A 2-degree shift can move your results by a measurable amount. Make sure the lab temperature is stable. Don't run a batch at 9 AM and another at 4 PM if the HVAC cycles between those times. I keep the instrument in a room with a thermostat set to 22 degrees Celsius and let it equilibrate for at least two hours before starting any analysis. This usually stabilizes the results within 0.2 microns on the d50, which is acceptable for most applications. When the run finishes, export the data. The Sedigraph software gives you cumulative and differential distributions, along with summary statistics like d10, d50, d90, and the span value. Save everything. Not just the results, but the raw data and the instrument parameters you used. Two years later, when someone asks why your ceramic body cracked in firing, you'll want to be able to pull up that exact run and verify the particle size profile.
Get the Full Details
There are limitations worth acknowledging. The Sedigraph cannot handle non-settling particles. If your material is colloidal or doesn't have a sufficient density difference from the dispersant, it just won't work. You'd need a laser diffraction instrument for that. The method also assumes spherical particles. Your result is a Stokes diameter, which is the diameter of a sphere that would settle at the same velocity as your particle. Irregularly shaped particles will report as smaller than their longest dimension. A plate-like clay flake might show up as 2 microns when its actual width is 8 microns. This isn't a flaw in the instrument. It's just what the method measures. You need to understand what you're actually getting. Maintenance is relatively simple but shouldn't be ignored. Clean the chamber after every use. Residual sample dried inside the tubing or chamber will throw off future readings. Run distilled water through the system, then a cleaning solution if you've been analyzing dense or sticky materials. Check the X-ray tube hours. Most tubes last 2,000 to 3,000 hours before output degrades noticeably. Track your hours and replace the tube proactively rather than waiting for your results to drift. If you need the software or method documentation for Sedigraph Particle Size Analysis, the original Mitchell Instruments documentation is archived and available through industrial instrument repositories. Some third-party labs still offer calibration services that include method files and validation protocols. Look for sources that are current with ISO 13318, which covers gravity sedimentation methods. That's the relevant standard, even though many older Sedigraph methods predate it.
The bottom line is that the Sedigraph gives reliable results when you respect its assumptions and control your variables. It's not a set-and-forget instrument. The dispersion quality, sample density, temperature stability, and degassing all directly affect your data. Get those right and you'll have size distribution results that are consistent and meaningful. Miss any of them and you'll be chasing ghosts in your data for hours.
