Coolant Selection Basics: Flood, Mist, Air, or Dry

By Bradley Taylor · August 2026

Every shop has a guy who floods everything and a guy who cuts everything dry, and both of them will tell you the other one is ruining tools. The truth is less exciting. Coolant is a tool like any other, and picking it starts with knowing what job you are actually asking it to do at the cut.

The three jobs coolant does

Coolant earns its keep three ways. It pulls heat out of the cut and the part, it lubricates the interface between the tool and the chip, and it moves chips out of the way so they do not get recut. Every operation needs all three to some degree, but the weighting changes completely from job to job. Drilling deep holes lives and dies on chip evacuation. Threading and tapping care about lubrication more than anything else. High speed turning of steel mostly wants heat carried away. Once you think of it as three separate jobs instead of one magic fluid, the rest of coolant selection gets a lot easier.

The fluid families in plain terms

Straight oils are cutting oil with no water in them. They are the best lubricant you can put on a cut and the worst coolant, because oil simply cannot carry heat away the way water can. Screw machines, heavy thread cutting, and broaching love straight oil for exactly that reason. The cuts are slow and the loads are high, so lubrication is the whole game and heat is manageable.

Water based fluids trade lube for cooling in steps, and the percentages are worth getting straight. A soluble oil concentrate is typically 40 to 70 percent mineral oil, a semisynthetic concentrate runs maybe 5 to 30 percent, and a full synthetic has none. But every one of them gets diluted 5 to 10 percent into water at the sump, so the working fluid in the machine is over 90 percent water no matter which family you buy, and their raw ability to carry heat is nearly identical. What actually separates them at the cut is the oil film. Soluble oils keep the most lubricity, semisynthetics sit in the middle, which is why so many general job shops end up running one, and synthetics are the weakest lubricant but the cleanest runner. They wet and rinse better, they stay transparent so you can see the cut, and they resist growing things. That makes synthetics a fine match for grinding and high speed work, and a poor match for heavy tapping or gummy low speed cuts. Two more synthetic traits worth knowing: they tend to reject tramp oil and float it to the surface where a skimmer can get it, which makes sump upkeep easier, and their higher pH is harder on skin, so gloves and rinsing stop being optional.

Where dry is the right call

Cast iron is the classic dry material. The graphite in gray iron acts as its own built in lubricant, the chips come off as powder rather than long curls, and adding flood coolant just turns that powder into an abrasive gray mud that gets into every way cover and slide on the machine. Gray iron work mostly runs dry with a vacuum on the chips, and everyone is happier for it, with two footnotes. Ductile iron and compacted graphite iron get run wet all the time, for dust control and thermal stability, so do not read gray iron habits onto the whole family. And dry iron dust is not just a housekeeping issue. Casting skin carries sand, which means respirable silica, so the vacuum and the dust collection are doing health work, not just keeping the floor clean.

Ceramic inserts get their own rule. Ceramics and whisker reinforced grades hate thermal shock even more than carbide does, so they run either fully dry or under constant uninterrupted flood, never coolant that comes and goes, and never a switch from dry to wet in the middle of a cut. Most CBN hard turning follows the same dry or steady logic.

The other big case is interrupted cutting with carbide, which mostly means milling. Every time an insert leaves the cut it cools, and every time it reenters it heats back up. Flood coolant makes that swing far more violent, and carbide handles steady heat much better than it handles thermal cycling. The cracks that form from that cycling run perpendicular to the edge and they will take out an insert well before normal wear would. This is why intermittent or marginal coolant in milling can genuinely be worse than none at all. A weak stream that splashes the insert on some passes and misses it on others gives you all of the thermal shock and none of the cooling. If you cannot flood the cut properly and keep it flooded, running the cutter dry and adjusting your surface footage down is often the better trade. The speeds and feeds calculator is the place to work out where those numbers land.

Air blast has its own lane

An air blast does one of the three jobs, chip clearing, and skips the other two. That turns out to be exactly right for a surprising number of cuts, but be honest about why. Plastics run under air mostly for chemical and dimensional reasons, not thermal ones. Coolant chemistry and oils craze and stress crack polycarbonate, acrylic, and other amorphous grades, sometimes weeks after the part shipped, and nylon and PEEK drink moisture and drift on size. The heat story actually argues the other way: a plastic conducts heat so poorly that the heat concentrates right at the cut, which is exactly why plastics melt, gum, and grow oversize. So air on plastics is a compatibility decision that leaves you managing heat with sharp tools and honest feeds, and plenty of shops run acrylic and PTFE wet on purpose when the material allows it. Aluminum finishing under plain air needs the same honesty. Aluminum's failure mode is adhesion, chips welding to the edge and building up, which is a lubrication problem, and bone dry air often gives you built up edge and a smeared finish in 6061. Air works there with polished flute tooling and high surface speed, and it works far better with a trace of lubricant in it, which is the next section.

Mist and MQL, the middle path

Between flood and dry sits minimum quantity lubrication, MQL to the catalogs and mist to everyone at the machine. A metered aerosol of oil rides the air blast to the cut, typically a few milliliters per hour, little enough that parts and chips come off essentially dry. You get the chip clearing of air plus exactly the lubricity that dry aluminum work was missing, which is why MQL is the modern answer for aluminum and for shops trying to get out of the sump business entirely. No coolant to test, no biology to manage, no tramp oil, dry chips worth more at the recycler. The costs are real too: an MQL system is plumbing and metering hardware, not a nozzle on a compressor, it does very little bulk cooling, so it is wrong for deep holes and heavy stainless work, and an oil aerosol is exactly the kind of thing your lungs want no part of, so enclosure and mist collection matter. Which is a good place to say plainly that mist exposure is a regulated health topic. Fluid mist has occupational exposure limits, dermatitis is the most common complaint in wet shops, and long term bioaerosol exposure from dirty sumps is linked to real respiratory disease. Run the mist collector, keep the enclosure shut, and take the skin cream seriously.

Two materials with their own rules

Magnesium first, because this one is a safety line, not a preference. Never put water based coolant on magnesium. Hot magnesium and water make hydrogen gas, magnesium chips burn as a Class D metal fire, and water on that fire makes it worse. Magnesium runs dry or with the specific mineral oils rated for it, the chips go in covered steel cans, and the extinguisher on the wall needs to be a Class D unit, because the red one does nothing good. If magnesium work is new to your shop, read up before the first chip comes off.

Titanium sits at the other pole. Its terrible thermal conductivity concentrates heat at the edge, so it wants generous flood or high pressure coolant, and running it dry is both a tool killer and a fire risk once fine chips pile up. One more titanium specific wrinkle: aerospace work commonly restricts chlorine bearing cutting fluids on titanium over stress corrosion concerns, so check the customer's spec before the fluid hits the part. And while aluminum in bulk is not the fire hazard those two are, fine aluminum dust from sanding and polishing is a legitimate combustible dust, and wet aluminum fines sitting in a sump or a collector generate hydrogen, so treat the fines differently than the chips.

Tapping wants lube, specifically

Thread cutting is a low speed, high pressure, high contact operation, which is the exact profile where lubrication dominates and cooling barely matters. This is why a hand tapping job goes so much better with a dab of tapping compound than it does under a flood nozzle. The compound is a dedicated extreme pressure lubricant sitting right where the tap needs it, while flood coolant is mostly water arriving at high volume to solve a heat problem you do not have. Rigid tapping under power in a machine full of soluble oil works fine, but if a tap is squealing or tearing threads, the fix is almost always more lubricity, not more flow.

Concentration, or why the refractometer exists

Water based coolant is a mix, and the mix drifts. Water evaporates out of the sump all day while the oil and additives stay behind, so concentration creeps up over time unless you are topping off with lean makeup. A refractometer is the cheap handheld tool that reads that concentration by how much the fluid bends light. You put a drop on the window, hold it up to the light, and read a number that you multiply by the fluid's factor to get percent. The factor comes off the data sheet and usually lands somewhere between 1 and 2.5. One catch with aged sumps: tramp oil and dissolved junk blur the reading line and push it high, so a tired sump that reads 8 percent may really be sitting at 5. When the line gets fuzzy, cross check against a fresh mixed sample or have the fluid titrated instead of trusting the number.

Both directions of drift cost you. Run too lean and the corrosion inhibitors are diluted, so parts and machine surfaces start to rust, and the biocides thin out too, which is how a sump starts growing things. Run too rich and the coolant foams, smokes at the cut, leaves sticky residue on everything, and irritates skin. Typical working ranges for general purpose water based fluids are roughly 5 to 10 percent, with the fluid maker's data sheet as the real authority. Material matters here too, since aluminum and steel do not always want the same chemistry, and the material library covers those differences.

Keep the sump honest

Way lube and hydraulic oil leak into the sump on every machine, and that tramp oil floats on top and seals the coolant off from air. Bacteria that thrive without oxygen take over underneath, and that is the famous Monday morning smell after a weekend of still coolant. The smell is the polite symptom. A biologically dirty sump is also what drives the dermatitis and the respiratory complaints in wet shops, so sump upkeep is a health practice wearing a maintenance costume. Skim or wheel the tramp oil off regularly, keep the coolant circulating when you can, and check concentration weekly, and a sump will last months instead of weeks.

High pressure coolant in one paragraph

High pressure systems do more than cool. In turning, where the headline figure of about 1000 psi comes from, a jet aimed into the cut zone gets under the chip and breaks it, which is a big deal in stainless, titanium, and nickel alloys where stringy chips wrap the tool and ruin finishes. Through spindle drilling systems usually run lower, a few hundred to a thousand psi, and in deep holes the flow volume matters as much as the pressure, because the job there is flushing chips up and out of holes that would otherwise pack and snap the drill. If a shop quotes a lot of nickel alloy or titanium work, high pressure coolant usually shows up in the same building.

As always, this is general practice, not a spec. Fluid maker data sheets, prints, and customer requirements win every argument.