Use the oil that cools the part at the rate its steel asks for: fast enough to harden it, but not so abruptly that the part warps or cracks. Steel that hardens easily calls for conventional quenching oil; steel that is hard to harden, or a thick part, calls for a faster oil, medium-speed or accelerated. Where fire in the tank is a concern, a polymer-in-water fluid is the option, and a test of how the fluid cools a test piece confirms the choice.
What does quenching do to steel?
It hardens the steel. This is the moment when the part comes out of the furnace glowing and hisses as it enters the tank: within seconds, a steel that could be machined becomes hard enough to serve as a gear or a shaft.
In the furnace, the iron atoms rearrange themselves and open up room for carbon to dissolve between them, just as hot water takes in more sugar than cold. This arrangement is called austenite, and heating the steel up to it is called austenitizing.
From there on, everything depends on the cooling rate:
Slowly, the carbon has time to gather into thin layers inside the steel itself, without leaving the part, and the steel goes soft again.
Quickly, the carbon stays trapped between the iron atoms, the arrangement stretches in one direction, and martensite is born: the hard structure that quenching is after.
Figure 1. Cooled slowly, the carbon gathers and the steel goes soft; cooled quickly, it stays trapped and martensite forms. Qualitative drawing.
The speed of the bath matters because quenching comes at the end of manufacturing. The part that reaches the tank has already been through the lathe and through the cutting fluid for machining. A machining mistake can still be corrected, but a wrong quench often cannot: a cracked part becomes scrap.
If you need the numbers: austenitizing typically runs from 800 to 900 °C in carbon steels (unalloyed) and low-alloy steels (with only a small amount of added elements). Slow cooling forms the soft structures called ferrite and pearlite.
But how fast is "quickly"?
Why does each steel need a different cooling rate?
Because each steel has a minimum cooling rate for turning into martensite, the critical cooling rate.
Speed counts most in a temperature range in the middle of the path: if the part lingers there, some of the steel turns into the soft structure of slow cooling, and the hardness falls short of what was specified. In this article, we call that range "the range where the steel softens".
Carbon decides how hard the martensite gets, and alloying elements (chromium, manganese, molybdenum, nickel and, in a tiny dose, boron) let the steel harden with gentler cooling. Steel that receives these elements on purpose, above the usual content, is alloy steel.
Thickness matters too. Put a boiled potato in cold water: the skin cools in seconds, while the middle stays hot for minutes. In a thick part, the core may never reach the critical rate, and hardness drops gradually from the surface toward the center.
The thickness at the point that matters is called the section. How easily the steel hardens all the way to the core is called hardenability.
Figure 2. Same steel, same bath: the thin bar hardens to the center; in the thick one, hardness drops from the surface to the core. Qualitative drawing.
The result is a simple rule: a steel with little alloy, or a thick part, needs a bath that removes heat faster.
If you need the numbers: the range where the steel softens lies around 700 to 500 °C, depending on the steel. SAE 1045, an unalloyed medium-carbon steel common in shafts, hardens well in conventional oil only in thin parts.
So is it enough to cool as fast as possible?
Why does cooling too fast crack the part?
Because the surface and the core of the part transform at different moments, and a strong bath pushes those moments further apart. Two facts explain the rest.
First, martensite starts to form at a temperature that is specific to each steel, the Ms (from martensite start).
Second, martensite takes up a little more space than austenite: the steel "swells" slightly as it hardens, while the rest of the part shrinks as it cools.
The surface always cools before the core. A strong bath widens that gap, and that is where cracking comes from:
The stronger the bath around the Ms, the larger the temperature difference between the surface and the core.
With that difference, the surface passes the Ms and turns into martensite well before the core.
When the core finally transforms, it swells and pushes against the surface, which is already hard and brittle.
This internal force is called stress. If it exceeds what the steel can take, the part warps (bends) or cracks, sometimes hours after the bath. The crack may only show up at grinding, the finishing operation that wears the surface with a grinding wheel.
That is why the rule is to use the gentlest bath that still gives the hardness specified, at the depth specified. A gear, for example, needs a very hard surface and a tough core, but not as hard as the surface.
The steel, the shape of the part and the furnace also weigh on cracking; here the focus is the bath, the part that the oil controls.
If you need the numbers: the Ms generally lies between about 200 and 400 °C and falls with more carbon and more alloy. In medium-carbon steels, Andrews' empirical formula (1965, valid up to about 0.6% carbon) gives around 300 to 350 °C. High-carbon steels sit below 300 °C.
The ideal bath would be fast in the range where the steel softens and slow in the range of the Ms. What controls this is the quenching fluid.
What role does oil play in quenching?
Oil is one of the quenching media, the fluid that removes heat from the part. The strength with which a medium removes heat is called severity.
Among cold liquids, with the bath still, the most severe is brine (salt water); next come water, polymer dissolved in water and oil. (Polymer, here, is a kind of plastic that dissolves in water and makes it thicker.) Depending on the concentration, the polymer is gentler, close to oil, or more severe, close to water.
Milder media, such as molten salt, air and gas, serve mainly high-alloy steels and special processes, such as marquenching in salt. Here they appear only in passing.
Water and brine harden carbon steel (unalloyed steel) well, with a higher risk of cracking. Of the four cold liquids, oil is the most gradual, the standard for alloy steel and precision parts.
If you need the numbers: on Grossmann's H factor, a classic measure of severity, brine sits near 2, water near 1 and still oil from 0.25 to 0.30.
If water and oil both go into the tank cold, why is oil gentler?
How does oil quenching work, and why is it gentler than water?
The part loses heat in three phases, each with its own speed. Anyone who has seen a drop of water jump on a hot plate knows the first one.
Vapor blanket. The drop "dances" because its underside turns to vapor and holds it up. In the tank, the oil that touches the glowing metal also turns to vapor and forms a blanket that insulates the part: cooling is slow.
Boiling. The blanket breaks, the oil touches the metal and boils in bubbles, like water in a pot. Each bubble carries heat away, and this is the fastest phase.
Convection. When the surface gets colder than the temperature at which the oil boils, the bubbles stop, and heat leaves only with the oil circulating around the part. It is the slowest phase.
Figure 3. The part cools slowly in the vapor, fast in boiling and slowly in convection, when it crosses the Ms. Qualitative drawing.
A good part of the hardness is decided as the vapor leaves: it has to break before the part enters the range where the steel softens. If it still covers the part in that range, hardness drops; if it covers only a stretch, a soft spot appears.
The strength of oil is that it boils at a much higher temperature than water. So, in general, the bubbles stop and the slow phase begins before the surface reaches the Ms. From there on, heat leaves slowly, and the core has time to catch up with the surface.
When the surface crosses the Ms, the temperature difference to the core is smaller than in water, and so is the stress. In water, the bubble phase, the fastest one, only ends when the part's surface gets close to 100 °C, well after crossing the Ms.
If you need the numbers: quenching oil generally boils between about 230 and 480 °C, depending on the formulation, and the Ms lies between about 200 and 400 °C. Since the ranges overlap, the advantage holds for many steels, not for all of them.
Further on, the cooling curve test shows how much heat an oil still removes in the Ms range.
And does every quenching oil do this the same way?
What is the best oil for quenching steel?
The best is the one that delivers the speed the steel and the section of the part ask for. By speed, quenching oils form three groups:
Conventional oil (or normal): the slowest, for steel that hardens easily.
Medium-speed oil: the middle ground, for steel of intermediate hardenability or a part of medium section.
Accelerated oil (or fast): carries additives, chemicals in a small dose, that break the vapor blanket earlier.
Figure 4. Accelerated oil breaks the vapor earlier and passes faster through the range where the steel softens. Qualitative drawing.
Besides these three groups, marquenching oil works hot, close to the steel's Ms or slightly below it: the part waits in the bath until surface and core reach the same temperature, and only then cools in air. That is why it warps less.
After any quench, the part goes on to tempering, reheating to a temperature well below the austenitizing temperature, which relieves the stress and the brittleness of the martensite.
Beyond oil, the table below also brings in polymer in water. For the choice between oil and water by steel, read quench in oil or water.
| If in your plant… | Use | Why |
|---|---|---|
| The steel hardens easily, or the part is thin | Conventional oil | Lower risk of cracking and warping |
| The steel has intermediate hardenability | Medium-speed oil | Faster than conventional |
| The steel has low hardenability, or the section is too thick for conventional oil | Accelerated oil | Breaks the vapor earlier and hardens deeper |
| The part warps a lot, or has very different sections (thin and thick wall in the same part) | Marquenching in hot oil (or in molten salt) | Surface and core equalize in temperature in the bath and transform almost together |
| The tank must be non-flammable, or the severity must fall between that of water and that of oil | Polymer in water | It does not catch fire; the concentration adjusts the severity |
Tank agitation always increases severity; so agitate the same way in every load, so the result does not change.
At Kelpen Oil, we formulate quenching oils in two lines: the Micro Temp line, with options for steels that are hard to harden, and the Austenity line, with conventional and accelerated quenching options for ferrous metals (those based on iron, such as steel). We also formulate a non-flammable polymer-in-water quenching fluid.
For medium speed and marquenching, talk to the Kelpen Oil Technical Team. The quenching products are listed under heat treatment oils, and the Technical Team recommends the oil based on the steel and the section of the part.
The bath temperature matters too. In water and polymer, a hotter bath lengthens the vapor blanket. In oil, up to a point, it makes the liquid less viscous (it flows more easily) and speeds up cooling; beyond that point, the vapor blanket also becomes more stable and the oil slows down again.
If you need the numbers: in classic marquenching, in molten salt, the bath sits just above the Ms. In oil, it is typically between 100 and 200 °C, often below the steel's Ms (modified marquenching, which cools faster). Commercial marquenching oils work from about 60 to 230 °C.
What remains is whether an oil is really fast, and not only on the label.
What is a cooling curve?
It is the plot of how a test piece cools in the oil. Imagine a thermometer inside it noting the temperature at every instant, from the furnace to the end of the bath: connect the points and you have the cooling curve.
The test piece is called a probe. It is made of a nickel alloy that does not change structure as it cools, so the curve shows only the oil, without the transformation of a steel in the middle.
The test gives two curves: temperature against time, and cooling rate against temperature. The second one, the cooling rate at each temperature, shows the three phases in the oil under test: when the vapor breaks, how fast the boiling is and how mild the convection is.
Figure 5. The probe's cooling rate at each temperature: low in the vapor, highest in boiling, low again in convection. Qualitative drawing.
Reading the curve, a vapor that breaks early and a fast boiling harden more. The rate at 300 °C, which the test records, shows the risk of cracking, because it lies in the Ms range.
Your part has another section, another shape, other agitation and another load in the basket. That is why the curve serves to compare oils with each other and the same oil over time. "Fast oil" without a curve is a promise: the curve is what belongs in the purchase specification.
If you need the numbers: the oil test is ASTM D6200, equivalent to ISO 9950, with an Inconel 600 probe 12.5 mm in diameter by 60 mm long, a temperature sensor (thermocouple) in the center and still oil.
The test records the maximum rate and the temperature at which it occurs, the rate at 300 °C and the times to 600, 400 and 200 °C.
A usual classification in the technical literature (Gear Solutions magazine, the series Oil Quenchants: Understanding the Chemistry, part I) separates oils by maximum rate: normal below about 60 °C/s, medium from 60 to 90 °C/s and fast above 90 °C/s.
For polymer, ASTM D6482 applies, with bath agitation (the Tensi method), because polymer changes a lot with agitation.
And an oil's curve does not stay the same forever.
How do you know if quenching oil is still good?
By comparing the curve of a sample from the tank with that of the new oil, or with that of the time when the process was approved. In service, oil ages:
it reacts with oxygen in the air, faster when hot: this is oxidation;
its molecules are broken by heat: this is thermal cracking;
it forms sludge, a mud that does not dissolve, settles at the bottom of the tank and leaves varnish on the part;
it takes in water and soot.
The drift can go either way. The oil gets slower when the accelerator additive is used up or when oxidation thickens it, and faster when the oxidation products break the vapor earlier. To the naked eye, both cases look the same; the curve tells them apart.
How this shows up in your plant: low hardness or soft spots, cracking and warping that force you to straighten the part or to leave more stock allowance (the extra material that grinding removes afterwards).
Water is the worst contaminant, because it acts in both directions. First, it makes the vapor blanket longer, which gives a soft spot. Then the boiling becomes more violent and extends closer to the Ms, where fast cooling cracks and warps the part.
Water can also make the oil foam and overflow (see the question about fire).
Figure 6. With water, the vapor lasts longer (soft spot) and the rate peak is higher and at a lower temperature (cracking). Qualitative drawing.
In polymer, the concentration is also monitored, with a refractometer, an instrument that measures how much light bends as it passes through the liquid. The reading multiplied by the correction factor that the polymer manufacturer provides gives the concentration.
Part number 10,000 only quenches like part number 1 if someone is measuring.
If you need the numbers: with water, the maximum rate rises and the temperature at which it occurs falls. The ASTM D6710 guide sets no limits. As a reference, it cites less than 0.1% water (1,000 ppm, parts per million), a value that the technical literature gives as the most that quenching oil usually dissolves.
Above that, free water is left in the tank. In steel of low hardenability, even below 0.05% (500 ppm) water can already cause a soft spot, so the target is as close to zero as possible.
That leaves the practical side: who does this measurement, and how.
How does Kelpen Oil measure the cooling curve?
The Kelpen Oil in-house laboratory in Arujá (SP) plots the curve of quenching oils by ASTM D6200 and that of polymers by ASTM D6482, on the IVF Smart Quench equipment.
We use the test to control our own batches and we run the analysis for other companies, including oil from another manufacturer, as a paid service. The curve tells whether the oil has changed; the other lubricant laboratory tests tell why:
| What we measure | What it reveals | Standard |
|---|---|---|
| Oil cooling curve | Whether the oil still quenches as it did on the day of approval | ASTM D6200, on the IVF Smart Quench |
| Polymer cooling curve, with agitation | The same reading for polymer in water | ASTM D6482, on the IVF Smart Quench |
| Water in oil (Karl Fischer method, a chemical reaction that counts the water) | Water content, in ppm: it changes the curve and brings a risk of foam and fire | Coulometric Karl Fischer |
| Acid number (TAN, for Total Acid Number) | How much acid oxidation has already formed in the oil | NBR 14448 |
| Kinematic viscosity at 40 °C (the oil's resistance to flow) | Whether the oil has thinned (thermal cracking or mixing with a thinner oil) or thickened through oxidation | ASTM D7042, on the SVM 4001 viscometer |
| Flash point (the lowest temperature at which the oil's vapor flashes when it meets a flame, without sustaining the fire) | Safety margin against fire in the tank; a drop indicates contamination or thermal cracking | NBR 11341 |
With these results in hand, ask your tank two questions. Does this oil deliver the severity the part needs? And does it still behave as it did on the day the process was approved?
Frequently asked questions
What is the difference between quenching and tempering? Quenching hardens the steel. Tempering comes right after: it reheats the part well below the austenitizing temperature to relieve the stress and the brittleness of the martensite, trading a little hardness for toughness (resistance to breaking under impact).
What is the quenching temperature of steel? In carbon and low-alloy steels, austenitizing typically lies between 800 and 900 °C, depending on the alloy. High-alloy tool steels and martensitic stainless steels (the ones that harden on quenching) use higher temperatures; the exact value comes from the steel specification.
How do you soften quenched steel? By annealing: the part is reheated to the austenitizing temperature and cooled slowly, usually inside the furnace itself, which brings back the soft, machinable structure.
Does quenching oil catch fire? It can, because oil is combustible. Fire usually comes from water in the oil (it turns to vapor all at once, and the oil foams and overflows). It can also come from a load stuck half out of the bath (the glowing metal sits at the surface of the oil) or from too little oil for the size of the load. The NFPA 86 standard requires the tank to be sized so that the oil works, at maximum load, at least 28 °C (50 °F) below the flash point.
Does quenching oil need periodic analysis? Yes. The curve changes with oxidation and contamination, and only the ASTM D6200 test on a sample from the tank shows whether the oil still quenches as it did on the day of approval.
Glossary
Austenitizing: heating the steel up to austenite, the arrangement in which carbon dissolves in iron.
Martensite: the hard structure that forms when steel cools fast and the carbon stays trapped between the iron atoms.
Hardenability: how easily the steel hardens all the way to the core of the part.
Ms: the temperature at which martensite starts to form; it varies with carbon and alloy.
Severity: the strength with which the quenching medium (water, polymer, oil) removes heat from the part.
Vapor blanket: blanket of vapor that wraps the part at the start of the bath and insulates the surface.
Cooling curve: record of the temperature of a standardized probe in the bath over time.
Marquenching: quenching in a hot bath (oil or salt), close to the Ms or slightly below it, in which surface and core equalize in temperature before the part cools in air.
Talk to the Technical Team
The Kelpen Oil Technical Team evaluates the type of steel, the geometry and section of the part, the process temperature, the hardness required and the tank volume, and confirms the right product (quenching oil or polymer), with the cooling curve measured in our in-house laboratory. If the process already runs with another product, the Arujá (SP) laboratory performs the analysis of the oil in use with a sample from the tank, as a paid service. Save this article or send it to whoever is in charge of heat treatment at your company.
