Quenching Media for Tool Steel, Cooling Rates and Quenching Behaviour
A quench is judged by two numbers rather than one. The first is how fast the medium pulls heat out of the part in the range where pearlite and carbide would otherwise form, and the second is how fast it keeps cooling below the martensite start, where extra speed turns into distortion and cracking risk. The tables below collect the published cooling rates of the common quenching media, the cracking sensitivity and the distortion measured for a polymer quenchant at three concentrations, and the hardness the same media produce through the section of a test bar. Read together they say which medium a given section can be quenched in.
Reference only. The figures below are a reference summary of published Chinese handbook practice, not an Aobo Steel specification. Confirm the medium and the cycle against the mill certificate and the governing drawing before an order or a production run is set against them.
What the quench has to do
Inside the section being hardened the steel has to pass the pearlite and carbide range quickly enough that those transformations never start, and it has to reach the martensite range without tearing itself apart. A quench that is too slow leaves soft spots in a thick section. A quench that is faster than the grade needs adds distortion and cracking risk for nothing. The cooling rate of a medium is therefore read in two separate temperature windows instead of as a single number, and every table on this page is built that way. The upper window, 650 to 500 degrees Celsius, is where soft spots are won or lost. The lower window, 300 to 200 degrees Celsius, is where cracking and distortion are won or lost.
Cooling rate of the common quenching media
The rates in the first table were measured on a 20 mm steel ball and estimated from the published cooling curve of each medium. The first column is the temperature at which the medium cools fastest, the second is that peak rate, and the last two are the average rate across the upper and the lower window.
| Quenching medium | Peak rate at, °C | Peak rate, °C/s | Average 650–500 °C, °C/s | Average 300–200 °C, °C/s |
|---|---|---|---|---|
| Still tap water, 20 °C | 340 | 775 | 135 | 450 |
| Still tap water, 40 °C | 285 | 545 | 110 | 410 |
| Still tap water, 60 °C | 220 | 275 | 80 | 185 |
| 10% NaCl solution, 20 °C | 580 | 2000 | 1900 | 1000 |
| 15% NaOH solution, 20 °C | 560 | 2830 | 2750 | 775 |
| 5% Na2CO3 solution, 20 °C | 430 | 1640 | 1140 | 820 |
| L-AN15 neat oil, 20 °C | 430 | 230 | 60 | 65 |
| L-AN15 neat oil, 80 °C | 430 | 230 | 70 | 55 |
| L-AN32 neat oil, 20 °C | 500 | 120 | 100 | 50 |
Source, Table 8-9, printed page 464. Cooling rate measured on a 20 mm steel ball and estimated from the cooling curve of each medium. L-AN grades are Chinese full loss system oils, L-AN15 near an ISO VG15 and L-AN32 near an ISO VG32. Percentages are mass fractions.
The pattern is the one a tool shop works with. Water is fast and gets slower as the bath warms, so a hot water bath is not a substitute for a controlled quench. A ten per cent salt brine or a caustic bath is roughly three times as fast as still water in the upper window, which is what removes soft spots on a low hardenability carbon steel, and both stay fast below 300 degrees Celsius as well, which is exactly why they crack thin sections. Neat oil is an order of magnitude slower than water in the upper window and just as slow in the lower one, which is why oil suits the grades whose austenite is stable enough to survive a slow pass through the pearlite range.
Salt and caustic solutions against water and oil
A saturated calcium chloride solution was developed as a medium sitting between brine and oil. Its cracking behaviour is closer to water than to oil, but its rate in the upper window is high enough for carbon steel. The second table compares it with the other common media measured in the same two windows.
| Quenching medium | Average 650–550 °C, °C/s | Average 300–200 °C, °C/s |
|---|---|---|
| Saturated CaCl2 solution | 1000 | 150 |
| Water, 18 °C | 600 | 270 |
| 10% NaCl solution, 20 °C | 1100 | 300 |
| 15% NaOH solution, 20 °C | 1200 | 300 |
| Rapeseed oil, 50 °C | 200 | 35 |
| L-AN15 neat oil, 50 °C | 150 | 30 |
| Transformer oil, 50 °C | 120 | 25 |
Source, Table 8-10, printed page 465. Average cooling rate across each range.
Polymer quenchants, cracking and distortion
A polymer quenchant is a water solution whose cooling speed is set by its concentration, so one bath can cover the range from brine down to oil by changing how much polymer it holds. The three tables below come from one test programme on a Chinese NQ-1 polymer, run at five, fifteen and twenty five per cent by mass against water and neat oil as the reference media.
Cracking sensitivity
The cracking test used a 30 mm by 30 mm T8 carbon tool steel specimen, austenitized at 840 degrees Celsius, held for thirty minutes and checked by magnetic particle inspection after quenching. Three specimens were run for each medium.
| Quenching medium | Specimens | Cracked |
|---|---|---|
| Water | 3 | 3 |
| 5% NQ-1 polymer | 3 | 1 |
| 15% NQ-1 polymer | 3 | 0 |
| 25% NQ-1 polymer | 3 | 0 |
| L-AN15 neat oil | 3 | 0 |
Source, Table 8-11, printed page 466. T8 carbon tool steel, 30 mm by 30 mm, austenitized at 840 degrees Celsius and held 30 minutes, checked by magnetic particle inspection.
Hardness through the section
The hardenability test used 45 steel, a plain carbon grade, in two bar sizes. Hardness was read at the surface, at the half radius and at the centre of a bar sectioned along its axis.
| Specimen | Quenching medium | Surface, HRC | Half radius, HRC | Centre, HRC |
|---|---|---|---|---|
| 45 steel, 20 mm × 60 mm | Water | 60 | 54 | 52 |
| 45 steel, 20 mm × 60 mm | 5% NQ-1 polymer | 61 | 57 | 53 |
| 45 steel, 20 mm × 60 mm | 15% NQ-1 polymer | 61 | 57 | 54 |
| 45 steel, 20 mm × 60 mm | 25% NQ-1 polymer | 58 | 52 | 50 |
| 45 steel, 20 mm × 60 mm | L-AN15 neat oil | 54 | 48 | 45 |
| 45 steel, 30 mm × 90 mm | Water | 56 | 33 | 31 |
| 45 steel, 30 mm × 90 mm | 5% NQ-1 polymer | 59 | 40 | 34 |
| 45 steel, 30 mm × 90 mm | 15% NQ-1 polymer | 58 | 38 | 33 |
| 45 steel, 30 mm × 90 mm | 25% NQ-1 polymer | 54 | 33 | 30 |
| 45 steel, 30 mm × 90 mm | L-AN15 neat oil | 50 | 25 | 23 |
Source, Table 8-12, printed page 466. 45 steel, austenitized at 840 degrees Celsius and held 30 minutes, hardness in HRC across the section. The 30 mm bar is short of full hardness at the half radius and at the centre in every medium.
Distortion and surface hardness
The distortion test used a 50 mm by 10 mm 45 steel disc, austenitized at 840 degrees Celsius and held for fifteen minutes, and the measured growth of the disc is the number in the table.
| Quenching medium | Average distortion, mm | Surface hardness, HRC |
|---|---|---|
| Water | +0.80 | 60 |
| 5% NQ-1 polymer | +0.61 | 61 |
| 15% NQ-1 polymer | +0.12 | 60 |
| 25% NQ-1 polymer | +0.26 | 58 |
| L-AN15 neat oil | +0.09 | 54 |
Source, Table 8-13, printed page 466. 45 steel disc, 50 mm by 10 mm, austenitized at 840 degrees Celsius and held 15 minutes. Distortion is the average measured change in size, positive meaning growth.
Reading the tables together
Three things stand out across the five tables. Water cracked every specimen in the cracking test and grew the distortion disc by 0.80 mm, and it is the fastest medium in both windows, so it is a medium of last resort on a grade that distorts. A fifteen per cent polymer bath cracked none of its specimens, held the same surface and centre hardness as water, and grew the disc by only 0.12 mm, which is the narrowest result in the set. A twenty five per cent bath is the mildest on distortion for the harder bars and the softest at the surface, and its hardness through the section falls off first, so its advantage is on shape rather than on hardness. Neat oil is the slowest and the softest, and it is the medium whose upper window rate is low enough to need a fully alloyed grade behind it.
Where the trade between hardness and distortion cannot be made in a liquid at all, an air hardening grade takes the question away, and a steel such as D2 reaches full hardness in still air or in a forced air quench rather than in a bath. Where water is still required, the note in the source is that a fast oil can often replace it on a thin or partly ground part, which is the route W1 is usually taken on and the reason O1 is built for oil in the first place.
Related reference data
The quench speed each grade needs is grouped on the quenching chart for water, oil and air hardening steels, the measured cooling rates for one hot work grade are on the H13 quenching cooling rates page, and full hardening and tempering cycles are on the tool steel heat treatment guide.
Notes on the data
The medium names follow the source. The oil grades are Chinese L-AN full loss system oils, and L-AN15 sits near an ISO VG15 viscosity grade and L-AN32 near an ISO VG32, so the number is a viscosity class rather than a brand. The polymer figures are for the NQ-1 polymer used in the source test programme and the concentrations are mass fractions, so a bath mixed to the same percentage from a different polymer will not necessarily behave identically. The cracking, hardenability and distortion tests were run on T8 and 45 steel, which are plain carbon grades rather than die steels, so those readings measure the medium and are not a prediction for a particular tool steel section.
Compiled from a Chinese mould and tool steel heat treatment handbook, Tables 8-9 to 8-13, printed pages 464 to 466. Tables 8-9, 8-11, 8-12 and 8-13 were read from the page image at 500 dpi and Table 8-10 from the page image and the text layer. The tables are a reference summary of published practice rather than an Aobo Steel specification, so confirm the medium and the cycle against the mill certificate before an order is placed against it.