Metalworking coolant sumps: solidifying tramp oil and emulsion

A 1,200 litre sump on a machining centre, a 6 mm layer of tramp oil on top, and a waste contractor quoting per tonne. The fitter had been told a scoop of powder would turn the whole thing into something he could shovel out. He tried it in a bucket first. What came back was a swollen lump of gel sitting in a puddle of oil, plus a mess that took longer to clean than the sump would have.

He was not doing it wrong. He was using a water absorber on a mixture where the interesting part is not water. Coolant sumps are one of the few jobs where SAP works well, but only after you accept what it can and cannot touch, and only after you stop expecting the volume to disappear.

What is actually in a sump

A typical water-miscible coolant at 6 percent concentration is around 92 percent water, 4 to 6 percent oil, and the rest emulsifiers, corrosion inhibitors, biocides and a suspension of metal fines. The oil is not floating free in the way it looks; much of it is bound in emulsion, and the tramp oil on top is the part that has already broken out, usually from hydraulic leaks and way lube.

SAP takes the water. It does not take the oil, and it will not break an emulsion on its own. Dose powder into a full sump and you get gel with oil trapped in the pockets and a film on the outside of every lump, which then leaks oil into the skip, the drum and eventually the leachate test at the landfill gate.

Skim first, then gel

Take the free oil off before any powder goes in. A belt skimmer running overnight, a coalescer, or a simple weir and decant will remove most of the tramp layer; a dipstick with water-finding paste tells you when the layer is down to a couple of millimetres. That oil goes to the waste oil tank where it belongs, and it may even have a recovery value.

Then dose the water phase. Starting point is 1.5 to 2.5 percent by weight of the liquid you are treating, adjusted for conductivity and surfactant load. A sump at 2,000 microsiemens will hold far less water per kilo than the same product in deionised water, so the datasheet number is a ceiling, not a promise. Dose in two passes, mix for five minutes, and check with a tilt test on a sample: no free liquid running when the container goes past 45 degrees.

The number nobody believes: volume does not shrink

Here is the part that derails budget conversations. Gelling does not reduce the volume much. Water that becomes gel still occupies roughly the space it occupied as water. A 1,200 litre sump still leaves you with something in the order of a cubic metre of solid, sometimes a little more because of the air the gel traps.

What changes is the classification and the handling. You are no longer shipping free liquid, so you pass a paint filter or a tilt test, you can use a standard skip instead of a vacuum tanker, and you stop paying for water at tanker rates. On a site generating 4 tonnes a month of sump fluid, that shift in handling is usually where the saving sits — not in the tonnage, which barely moves.

The honest exception is when a plant already has an on-site evaporator or a permitted discharge. In that case gelling adds a step and a cost, and the right answer is to decline the order. I have watched a plant buy 500 kg for exactly that reason and then store it for two years; the powder was fine, the decision was not.

Run the bucket trial before the sump

Five litres of real sump fluid, a jug, a scale and forty minutes will tell you more than any specification. Dose at 0.5 percent steps, stir each time for a full minute, and log what happens at each step. Watch for three things: how fast the gel forms, whether oil separates to the surface as the water is taken up, and whether the gel stays cohesive when you cut it with a trowel.

If oil rises to the top during the trial, you have a skimming problem, not a dosing problem, and no amount of extra powder will fix it. If the gel crumbles instead of holding together, your conductivity or your biocide package is fighting the polymer, and a coarser grade with a slower swell often does better than a finer one.

Send the trial numbers along with the request when you are ready to specify — pH, conductivity, oil percentage, fines load, and where the drum is going. The grade selection notes for industrial liquid solidification set out which of those four variables changes the recommendation first.