As featured in the January/February 2017 issue of

Thermal Processing magazine

Figure 1: The mechanism of quenching of oil: (a) The moment of immersion showing the presence of a vapor film around the component; (b) After five seconds, the boiling
phase commences at the corners of the component; (c) After 120 seconds, the boiling front moves along the component; (d) After 15 seconds, all three phases are present;
(e) After 30 seconds, the convection phase is the dominant heat transfer mechanism; (f) After 60 seconds, the convection phase is nearly complete [1].

Understanding the Cooling

Curve Test
By D. Scott MacKenzie

In heat treating, the cooling curve test is often used as a tool to compare
quenchants or as a method to ensure that the quenchant being used is suitable
for continued use and will satisfy current requirements.

W
When a hot component comes in contact with • Vapor stage (stage A or vapor blanket The vapor stage is encountered when the hot
the liquid quenchant, there are normally three stage) surface of the heated component first comes
stages of quenching. There are exceptions to • Boiling stage (stage B or nucleate boiling in contact with the liquid quenchant. The
this, which will be explained in each stage. stage) component becomes surrounded with a blan-
The three stages of quenching are: • Convection stage (stage C) ket of vapor. In this stage, heat transfer is slow
28 | Thermal Processing
 Figure 2: Example of cooling curves and cooling rate curves of fast, medium, and slow quench oils
fluid. The boiling stage stops when the tem- There are many methods to evaluate
perature of the component’s surface reaches the extraction of heat from a quenchant.
a temperature below the boiling point of the Examples include the GM Quenchometer
liquid. For many distortion-prone compo- test, ASTM D3520 [2]; and the hot-wire test.
nents, high boiling temperature oils or liquid In the GM Quenchometer test, the time to
salts are used if the media is fast enough to cool a 12-mm nickel ball to the Curie tem-
harden the steel, but both of these quenchants perature (352°C) is measured. In the hot-wire
see relatively little use in induction hardening. test, a Nichrome wire of a standard gauge
The final stage of quenching is the convec- and electrical resistance is immersed into
tion stage. This occurs when the component the quenchant. A current is passed through
has reached a point below that of the quen- the wire. The current is gradually increased
chant’s boiling temperature. Heat is removed until the burnout temperature (where the wire
by convection and is controlled by the quen- melts) is reached. The cooling power is rep-
chant’s specific heat and thermal conductiv- resented by the maximum current sustained
ity and the temperature differential between by the wire. In each of these cases, only one
and occurs primarily by radiation through the the component’s temperature and that of the value (either the time to cool or the burnout
vapor blanket. Some conduction also occurs quenchant. The convection stage is usually current) is the sole measurand available to
through the vapor phase. This blanket is the slowest of the three stages. Typically, it is compare quenchants. No information is pro-
stable, and its removal can only be enhanced this stage where most distortion occurs. An vided regarding the quench path.
by agitation or speed-improving additives. example showing the three stages of quench- The cooling curve test is the best procedure
This stage is responsible for many of the ing is shown in Figure 1. for characterizing the ability of a quenchant
surface soft spots encountered in quenching.
Strong agitation eliminates this stage. If the
Method
vapor phase persists, then non-martensitic
Parameter ISO 9950 AFNOR JIS K2242 Z8 E 45003 ASTM D6200
transformation products can occur. NFT-60778
The second stage encountered in quench-
ing is the boiling stage. This is where the Country International France Japan China USA

vapor stage starts to collapse and all liquid Probe Alloy Inconel 600 Silver Silver Silver Inconel 600
99.999% 99.999% 99.999%
in contact with the component surface erupts Pure Pure Pure
into boiling bubbles. This is the fastest stage Probe Dimensions 12.5 x 60 16 x 48 10 x 30 10 x 30 12.5 x 60
of quenching. The high heat extraction rates Vessel Dimensions, mm 115±5 dia. 138 dia. X 99 300 ml 300 ml 115±5 dia.
are due to carrying away heat from the hot high beaker beaker
surface and transferring it further into the liq-
Oil Volume, ml 2000 800 250 250 2000
uid quenchant, which allows cooled liquid to
Oil Temperature ºC 40 ± 2 50 ± 2 80, 120, 160 80 ± 2 40 ± 2
replace it at the surface. In many quenchants,
Probe Temperature, ºC 850 ± 5 800 ± 5 810 ± 5 810 ± 5 850 ± 5
additives have been added to enhance the
maximum cooling rates obtained by a given Table 1: Comparison of international cooling curve standards

thermalprocessing.com | 29
 Temp.
Max. Cooling
at Max. Time to Time to Time to
Property Cooling Rate at
Cooling 600°C 400°C 200°C
Rate 300°C
Rate

Variation 8.7% of
2.1 12.7 0.4 0.5 1.3
(2σ) Mean

Variation 25% of
8.6 25.3 1.4 2.1 10.1
(2σ) Mean

Table 2: Bias and precision (95-percent confidence) of ASTM D6200 for single
operator (top row) and interlaboratory testing (different operators with different
equipment, testing the same sample, last row) [3]

to extract heat. Cooling curves provide a complete picture of the heat

extraction and cooling performance of a quenchant, as a function of
surface temperature or center temperature of a probe. In this test, an
instrumented probe of some material is heated to the desired tem-
perature (usually in the austenite range) and then removed from the
furnace. It is then transferred to the quenchant, recording the time
verses temperature of the probe. Additional information is obtained Figure 3: Effect of oxidation on the shape of the cooling curve for a simple slow oil.
by taking the instantaneous slope of the time-temperature curve to The total acid number, in mg KOH/g is shown for each curve. Temperature of the
obtain the cooling rate curve. An example of the typical cooling rate quenchant was 40°C with no agitation
curves for a selection of fast to slow oils is shown in Figure 2.
Standardized test procedures are absolutely necessary to evaluate
quenchant performance. This fact is recognized by ASTM [3], SAE
[4], NADCAP [5], CQI-9 [6], ISO [7], and other auditing bodies for
aerospace, automotive, and other industries. The use of standardized
procedures for quenchant evaluation, in particular heat extraction
capabilities, allows reproducible historical data collection. It allows
evaluation to determine if a quenchant is suitable for a particular
application or enables quality checks on current processes.
One of the first documented references on cooling curve apparatus
was published by Pilling and Lynch [8]. This was an extension of
the work by Le Chatelier [9]. In this study, Pilling and Lynch used a
platinum-platinum rhodium thermocouple welded to the geometrical
center of a 6.4 mm diameter by 50 mm probe fabricated from nickel
+ 5% silicon. This alloy was found to be free from the transforma-
tion effects of steel and avoided any transformations of nickel. The
addition of silicon afforded oxidation resistance. They systematically Figure 4: Cooling curve response of a slow oil with different contents of water added;
examined the quenching characteristics of water, brine, soap solu- oil was tested at 40°C, with no agitation
tions, and three different oils. They observed three different modes
of cooling: type A, which is now known as the vapor blanket; type
B, or nucleate boiling; and type C, or convection.
Instrumented probes of many different types, shapes, and alloys
are presently being used for cooling curve analysis. Presently, most
standards specify either a pure silver or nickel-base Inconel 600 (UNS
N06600) alloy. An early standard adopted by Japan, JIS K 2242 [10],
is the Tamura probe, which measures surface temperature variation
by a thermocouple located at the probe surface. Two other national
standards utilize a silver probe: the French AFNOR NFT-60778 [11]
and China’s ZB E 45003-88 [12]. A reusable Inconel 600 probe was
developed at the Wolfson Heat Treatment Center, Birmingham, U.K.
[13]. This probe is the basis for ISO 9950 [7] and ASTM D6200 [3].
A comparison of the test methods is shown in Table 1.
In the United States, the most commonly used method for deter-
mining cooling curves is ISO 9950 [7] or ASTM D6200 [3]. This
standard is used extensively for the characterization of cooling curve
behavior of oil quenchants. It is used to determine the suitability of Figure 5: Comparison of cooling curves for the same oil from four different
customers, compared to new oil (control). All customers are using oil identical to the
a quenchant for a particular application and is increasingly used as a control. Oil was tested at 60°C, with no agitation
quality control check of used oil to ensure proper quenching and to
observe any oil deterioration. without agitation. The temperature of the oil can vary, but is typi-
Typically, the Inconel 600 probe is heated to 871°C and quenched cally either 60°C for “cold” oils or 121°C for martempering oils. No
into the desired quenchant. In the case of oil, 1-2 liters of oil is used agitation is used because of the difficulty in quantifying agitation.
30 | Thermal Processing
 Property Unit Control A B C D Max Min

Maximum Cooling Rate °C/s 94.6 102.4 99.6 86.1 98.7 103.2 86.0

Probe Initial Temperature °C 850.0 850.0 850.0 850.0 850.0    

Temp at Start of Boiling °C 725.0 713.4 720.7 736.8 730.1    

Temp. at Max. Cooling Rate °C 596.2 596.1 608.1 593.2 621.6 621.5 570.9

Temp at Start of Convection °C 347.1 366.2 355.2 350.5 334.1    

Cooling Rate at 300ºC °C/s 5.9 6.0 5.7 5.5 5.7 7.4 4.4

Time to 850ºC s 0.0 0.0 0.0 0.0 0.0    

Time to 600ºC s 6.8 7.0 6.8 6.6 6.7 8.2 5.4

Time to 400ºC s 10.5 10.7 10.7 10.6 10.2 12.6 8.4

Time to 200ºC s 45.7 45.6 47.3 46.6 43.2 55.8 35.6

Table 3: Extracted data from the customer cooling curves in Figure 5; all oil is tested at 60°C and no agitation
The influence of test conditions on ASTM D6200 has been studied Experience with poorly functioning quenching oils indicates that
by Moore and Guisbert [14]. Guisbert [15] examined the precision in at least half the cases, water gets into the oil, causing cracking,
and bias of cooling curve testing during the round robin evaluation uneven hardness, and soft spots. Water can enter the oil in a number
of ASTM D6200 prior to establishment as a standard. He found that of ways, and the sources must be tracked down and eliminated.
the cooling curve test of ASTM D6200 showed high repeatability and A leaky cooler might bring water into the oil or water may drip
reproducibility, provided that the probes were properly calibrated. The into the tank from the roof. A tiny amount of water (about 0.1
published bias and precision of ASTM D6200 is shown in Table 2. percent) can cause a bath to foam excessively and greatly increase
the danger of fire.
APPLICATION OF THE COOLING CURVE The cooling curve response of oils containing water shows an
Typical data that can be extracted from the ASTM D6200 cooling increased maximum cooling rate with an extended vapor phase.
curve test includes: Further, and probably the most telltale sign that water is present, is
• Maximum cooling rate, °C/s the rounding of the nucleate boiling to convection transition. It is this
• Temperature at maximum cooling rate, °C change in cooling curve response that is responsible for cracking of
• Cooling rate at 300°C, °C/s parts when water is present. A schematic showing the effect of water
• Time to cool to 600°C, 400°C, and 200°C/s on the cooling curve response is shown in Figure 4.
The cooling curve test is also used as a quality control check to
Other data can be extracted such as the temperature at the start of ensure that the oil is operating as it should. Deviations from process
boiling and the temperature at the start of convection. This data can be parameters indicate that corrective action should be taken to ensure
used to compare oils or to evaluate oils for a new process or application. that parts are quenched properly. For example, four cooling curves
When examining in-use oils, it is important to understand the from customers using the same oil (but different applications) were
effects of “real-world” contaminates to the cooling curve. Oxidation, selected and compared to new oil. The representative cooling curves
particulate, and water can have a considerable effect on the shape of are shown in Figure 5. All the cooling curves show similar responses.
the cooling curve. Data from the cooling curve can be used to determine if the oils meet
Oxidation and fine particulate have the effect of suppressing new manufacturing specifications for the quenchant.
the vapor phase and increasing the maximum cooling rate. It will Extracting data from the cooling curve and applying the bias and
also increase the temperature at which the maximum cooling rate precision from ASTM D6200, it can be determined if the customer
occurs. This is because the fine particulate or precursors to oxida- cooling curves meet specifications for new oil. The extracted data
tion that are soluble in the oil act as nucleation sites for the initia- is shown in Table 3. Inspection of the data from the cooling curves
tion of nucleate boiling. A schematic representation of the effect show that all the customer cooling curve data meets the requirements
of fine particulate or oxidation on the shape of the cooling curve for new oil. However, customer 3 shows that the maximum cooling
is shown in Figure 3. rate is a bit slower than the control, but still within the limits of
While oxidation and fine particulate such as soot can yield similar error for the test. If the customer is achieving good hardness during
cooling curve behavior, it is possible to differentiate between the quenching and not seeing a gradual decline in hardness over time,
causes with testing. Determining the total acid number per ASTM then the oil is acceptable for continued use. However, if the customer
D664 [16] or ASTM D974 [17] will tell if the oil is oxidized. If the is observing a gradual decrease in hardness over time, then it might
oil is not oxidized and has a low TAN, then the likely cause of the be appropriate to add a speed improver to his oil to recover the oil
change in cooling curve response is due to fine particulate or soot. back to original specifications.
Fine filtration (less than 3 µm) can reduce the effect of fine par- It is not unexpected that different customers would exhibit dif-
ticulate. Additions of antioxidant can reduce the oxidation of the ferent cooling curves for the same oil. These customers use this
quench oil and return the curve close to normal. oil (Houghto-Quench G) in a variety of applications, for different
thermalprocessing.com | 31
lengths of time, and under different mainte- plier can only specify whether the used oil 9. H. L. Le Chatelier, “Etudes sur la
nance conditions. None of these customers satisfies the manufacturing limits for new Trempe de l'Acier,” Rev. Met., vol. 1,
are reporting problems achieving hardness oil. Working with the customer, these limits p. 473, 1904.
or properties. can be modified for each application.  10. Japan Metal Working Fluids Association,
There are multiple reasons why the used Heat Treating Oil, Japan Standards
customer oils could show a greater scatter. REFERENCES Association.
Greater levels of dirt, including soot and scale, 1. Houghton Internationa l, Inc., 11. A FNOR, Industrial Quenching
could cause differences in the cooling curve “Houghton on Quenching,” 1991. Oils — Determination of Cooling
behavior of the oils. Further, recycling or 2. ASTM, “Standard Method for Characteristics — Silver Probe Test
recovery of the oil from washers could change Quenching Time of Heat Treating Method, Saint-Denis: AFNOR.
the cooling curve. However, fine soot and Fluids (Magnetic Quenchometer 12. Heat Treating Oils — Determination
residual washer residue would likely increase Method),” American Society for of Cooling Ability, Z8 E 45003,
the maximum cooling rate. Standards and Materials, Conshocken, Beijing.
This same methodology can be used to PA, 88. 13. L aboratory Test for Assessing the Cooling
select oil for a new application or to replace 3. ASTM International, Standard Test Curve Chacteristics of Industrial
an existing oil. The cooling curves are com- Method for Determination of Cooling Quenching Media, Birmingham:
pared, and the extracted data from the curves Characteristics of Quench Oils by Wolfson Heat Treatment Center
is further compared. Additional information Cooling Curve Analysis, Conshocken: Engineering Group Specification, 1982.
such as oxidation resistance and cost also ASTM, 2012. 14. D. L. Moore and D. A. Guisbert,
enter into the decision. 4. SAE International, Heat Treatment “Inf luence of Test Conditions on
of Steel Parts, General Requirements, Cooling Curve Data,” in Proc. 14th
CONCLUSION Warrendale, PA: SAE International, ASM HTS Conference & Exposition,
The cooling curve test is a powerful method 2014. Indianapolis, IN, 1993.
of examining the entire quench path of 5. SAE International, National Aerospace 15. D. A. Guisbert, “Precision and Accuracy
the quenchant. It can be used to examine and Defense Contractors Accreditation of the Continuous Cooling Curve Test
the condition of an oil to ensure that the Program (NADCAP) Requirements for Method,” in Proc. 16th ASM HTS
quenching characteristics are the same as Heat Treating, Warrendale, PA: SAE Conference & Exposition, Cincinnati,
new oil and whether corrective action must International, 2008. OH, 1996.
be taken. 6. AIAG, Special Process: Heat Treat 16. A STM, “ASTM D664: Standard
Many of the latest revisions of specifica- System Assessment, AIAG, 2011. Test Method for Acid Number of
tions, such as AMS 2759 [4] and auditing 7. Internationa l Organization for Petroleum Products by Potentiometric
agencies such as NADCAP [5], require the Standardization, Industrial quench- Titration.,” American Society of Testing
heat treating quenchant supplier to specify ing oils — Determination of cooling and Materials International, West
in the report whether the oil is “good” characteristics —Nickel-alloy probe test Conshohocken, PA.
or “bad.” However, the supplier does not method, Geneva, Switzerland. 17. ASTM, “ASTM D974: Standard
completely know the parts processed or 8. N. B. Pilling and T. D. Lynch, “Cooling Test Method for Acid Number and
the processes used nor does the heat treat Properties of Technical Quenching Base Number by Color-Indicator
quenchant supplier control any of the pro- Liquids,” Trans. American Ins. Mining Titration.,” American Society of Testing
cesses or parameters associated with the and Metallurgical Engineers, vol. 62, and Materials International, West
heat treating process. The quenchant sup- pp. 665-688, 1919. Conshohocken, PA.

ABOUT THE AUTHOR: D. Scott MacKenzie, Ph.D., FASM, is a research scientist of metallurgy at Houghton International, a global metalworking
fluids supplier. Previous to this position, MacKenzie was associate technical fellow at Boeing — performing failure analysis. At McDonnell Douglas, he
was the manufacturing engineer responsible for all heat treating activities at McDonnell Douglas – St. Louis. He obtained his B.S. from The Ohio State
University in 1981 and his Ph.D. from the University of Missouri-Rolla in 2000. He is the author of several books and over 100 papers, articles, and
chapters. He is a member of ASM International.

32 | Thermal Processing