Journal of Materials Processing Technology 114 (2001) 57±70

In¯uence of quenching process parameters on residual stresses in steel

Janez Grum*, Slavko BozÏicÏ, Martin ZupancÏicÏ
Faculty of Mechanical Engineering, University of Ljubljana, AsÏkercÏeva 6, 1000 Ljubljana, Slovenia
Received 28 October 1999; received in revised form 21 August 2000; accepted 6 February 2001

Abstract

The paper presents results regarding cooling intensities of quenching oil of domestic production and of polymeric water solutions.
Numerous residual stresses measurements on heat-treatment steel specimens with different masses were made. To measure the residual
stresses, the relaxation method was used, involving gradual mechanical removal of the hardened layer in which the deformation of the
specimen was measured by resistance strain gauges. On the basis of the measured strain at a speci®c depth of material removal, it is possible
to calculate the size or variation of residual stresses depending on the depth of the hardened layer. Numerous residual stresses
measurements provided very important data on various steels when using different quenching agents. # 2001 Elsevier Science B.V. All
rights reserved.

Keywords: Quenching; Quenching agents; Cooling intensity; Residual stresses; Microhardness

1. Introduction consequently on quality of a machine part. In practice, it is of

extreme importance that in steel hardening suitable cooling
Mechanical properties of steels are a function of their rates are ensured at individual points within a machine part.
chemical composition and microstructure, which are them- In hardening, a cooling rate should be equal to or higher than
selves a result of various physicochemical processes in the critical cooling rate required for formation of martensite
manufacturing of steels. For manufacturing of semi-pro- microstructure. The actual cooling rate in steel hardening
ducts having suitable size and properties, various forming depends on numerous more or less in¯uencing factors. The
and heat-treatment processes are required. In mass produc- actual cooling rate is affected, on one hand, by the machine-
tion, forming of blanks prepared from steelworker's semi- part geometry and, on the other hand, by the quenching
products may be used. Forming may be cold or hot and may agent used. Temperature and time thus seem to be the most
be followed by various machining processes with removing important in¯uences on ef®ciency of hardening. In practice,
the material, which may be followed by various kinds of heat several standardised methods of description of steel proper-
treatment. Heat treatment includes various ways of heating ties and their suitability for heat treatment are known. For
workpieces to a speci®ed temperature, requirements for determination of through-hardenability of steels, Jominy
overheating at a certain temperature and ®nally quenching and Grossman methods with determination of an ideal
of machine parts. In heat treatment of machine parts, it is of critical diameter, are well known. For determination of a
major importance that heating be adjusted in terms of the suitable quenching agent ensuring the desired properties of
selection of a proper temperature and heating time and the machine part after hardening, however, Lamont's and
soaking time, respectively, in order to ensure desired micro- Asimow's diagrams are used. Various methods for determi-
structure changes. nation of cooling ability of different quenching agents are
Heat treatment of steels, with regard to the present state of known as well. A fundamental question seems to be how to
development of technical engineering and technology, is combine data on materials and quenching agents in order to
easy to control and often highly automated. All heat-treat- predict proper hardening conditions.
ment processes which are performed at high cooling rates The most important in¯uences on ef®cient hardening of
require a suitable quenching agent. Cooling, i.e. quenching, machine parts are as follows:
of steel in heat treatment is a technological phase which
has an important in¯uence on microstructure changes and 1. In¯uence of material:
 chemical composition of steel;
*
Corresponding author.  specific heat of steel;

0924-0136/01/$ ± see front matter # 2001 Elsevier Science B.V. All rights reserved.
PII: S 0 9 2 4 - 0 1 3 6 ( 0 1 ) 0 0 5 6 0 - X
58 J. Grum et al. / Journal of Materials Processing Technology 114 (2001) 57±70

 thermal conductivity of steel; designations of this steel in various national standards. A

 heat transfer coefficient at the steel/quenchant chemical composition of the steel is given in Table 2. The
interface; designation indicates that this is an alloyed heat-treatable
 austenitizing temperature; CrMo steel with a medium carbon content. The steel is
 heating-up time to attain equal temperatures at the designed for dynamically loaded structural parts as well as
surface and in the core; for speci®c machine parts such as gear wheels, travelling
 soaking time at the austenitizing temperature to ensure wheels, crankshafts and connecting Ð rods, stub axles and
homogeneous austenitic microstructure; similar.
 mass of workpiece, i.e. machine part; The steel contains between 0.38 and 0.45% of carbon
 shape of workpiece, i.e. machine part; which makes it a typical heat-treatable steel and, with regard
 relation between the surface and volume of a machine to the other alloying elements, also ideal for ¯ame hardening
part; or surface induction hardening. The steel is not susceptible
 surface condition of a machine part; to brittleness after tempering at normal temperatures and
 movement of a machine part in the quenching agent. duration of heat treatment; therefore, there are no special
2. In¯uence of quenching agent: manufacturer's requirements regarding steel quenching and
 type of quenching agent; tempering. The steel shows very high toughness values at
 temperature of the quenching agent; ambient temperature and keeps them also in operating
 boiling temperature, latent heat, specific heat, thermal conditions, even considerably below an average ambient
conductivity, viscosity and other properties of the temperature. The strength of the steel concerned as well as
quenching agent; its surface hardness and consequently wear resistance may
 quenching methods such as immersion quenching, be increased by heat treatment and thermochemical treat-
immersion quenching with agitation or spray ment. Mechanical properties of steel having diameters of up
quenching. to 16 mm and between 16 and 40 mm are given in Table 3.
Tensile strength of the steel concerned varies between 980
The levels of in¯uence of individual machine-part char- and 1280 N/mm2 and a minimum toughness value r3
acteristics differ very much; therefore, selection of proper amounts to around 41 J.
heat-treatment conditions to ensure the desired mechanical The steel is very sensitive to notch and transitions on a
properties is very dif®cult. But in hardening, a particular machine parts subjected to fatigue loading. Table 4 gives
machine part made of a particular steel can be affected only values which are a basis of evaluation of fatigue strength for
by the selection of a quenching agent and by movement of various loading conditions. Fatigue strength of a material sD
the machine part and/or the quenching agent. is lowest under torsional load and varies, for the diameters
mentioned, i.e. 16±40 mm, from 285 to 325 N/mm2. Fatigue
strength under torsional load is three or four times lower than
2. Experimental the tensile strength of steel under static load.

2.1. Selection of steel 2.2. Selection of quenching agents

For an investigation of the in¯uence of steel mass and of a Development and testing of various quenching agents is
quenching agent on mechanical properties of steel after heat of great importance to steel quenching in terms of determi-
treatment we used, alloyed heat-treatable steel 42CrMo4 nation of cooling rates of machine parts during the hardening
according to DIN standard. Table 1 gives a comparison of process in individual temperature intervals. Knowledge of

Table 1
Designation of alloyed heat-treatable steel 42CrMo4 according to some national standards

Germany England USA Russia

W. No. DIN 17007 Designation DIN 17006 BS AISI GOCT

1.7225 42CrMo4 En 19 4140 ±

Table 2
Chemical composition (%) of the heat-treatable Cr±Mo steel 42CrMo4

C Si Mn P S Cr Mo

0.38±0.45 0.15±0.40 0.50±0.80 Max. 0.035 Max. 0.035 0.90±1.20 0.15±0.30

 J. Grum et al. / Journal of Materials Processing Technology 114 (2001) 57±70 59

Table 3
Mechanical properties of the alloyed heat-treatable steel concerned

Diameter, Min. yield strength, Min. tensile strength, Min. elongation, Min. reduction, Min. toughness,
D (mm) Rp0.2 (N/mm2) RM (N/mm2) A5 (%) Z (%) r3 (J)

Up to 16 885 1080±1280 10 40 34
16±40 765 980±1180 11 45 41
40±100 635 880±1080 12 50 41
100±160 560 780±930 13 55 41
160±250 510 740±890 14 55 41

physical processes in the workpiece/quenching-agent between Slovenian manufacturer ``Olma'' and the world-
boundary layer permits to avoid defects after hardening, known manufacturer ``Petrofer''. Aquatensid is a synthetic
which affects product quality and operational life. In the concentrate with the following properties:
investigation of the in¯uence of steel mass and quenching
agent on mechanical properties of steel and size of residual  cooling rate for machine parts can be adapted by different
stresses after heat treatment, a quenching oil Olmakal Rapid concentrations of the water solution;
Extra and a 10% polymeric water solution of Slovenian  the liquid is inflammable;
manufacturer ``Olma'', Ljubljana, were used.  it is insensitive to micro-organisms;
The oil Olmakal Rapid Extra is a high-quality quenching  it has long operational life;
oil suitable for use in closed or open quenching systems. It is  machine parts after quenching do not corrode;
recommended for hardening structural or tool steels. The oil  polymeric water solutions are insensitive to hardness of
Olmakal Rapid Extra has the following properties: water;
 it does not contain any harmful substances;
 very stable cooling properties after a longer operational
 polymeric water solutions of this concentrate pollute the
life;
environment very little;
 the components added increase its thermal conductivity,
 the concentrate and its polymeric water solutions are
reduce its susceptibility to cooking and formation of pitch
physiologically neutral.
components;
 it is not necessary to change oil often since it suffices to
2.3. Distortion and residual stresses after quenching
repour some fresh oil at times.
The oil Olmakal Rapid Extra is made of selected raw Residual internal stresses in a machine part are self-
materials having high ¯ash points and favourable cooling balanced internal stresses which persist in a material due
properties. The action of the oil in quenching is very mild so to different thermo-mechanical processes during its machin-
that the risk of crack formation is reduced. Variations of ing. Residual stresses are present already in semi-products,
internal stresses during quenching are favourable so that which are later treated by various machining processes to
susceptibility of machine parts to distortions during hard- make a ®nal machine part. Residual stresses in a workpiece
ening is small. change after each treatment. Changes in residual stresses in
Olmakal Rapid Extra is a high-quality oil for quenching of the workpiece after individual machining processes depend
steels in open quenching systems. The properties of the oil on the kind of process applied and the conditions under
Olmakal Rapid Extra are very similar to those of the which the process is performed. After each machining
quenching oil Olmakal Rapid 90. process, a new residual internal stress is obtained as a result
The polymeric water solution with a 10% concentration of previous residual stresses before machining and the new
was made of a concentrate of Aquatensid. The concentrate is stresses introduced. The resulting variation of residual inter-
designed for steel hardening as well as for heat treatment of nal stresses depends on the sign and values of the previous
various non-ferrous alloys. It is a result of co-operation residual stresses and the new stresses. In the machining

Table 4
Fatigue strength of the heat-treatable Cr±Mo steel 42CrMo4

Diameter, D (mm) Fatigue strength

Bending, sD (N/mm2) Compression±tension, sC±T (N/mm2) Torsion, sT (N/mm2)

Up to 16 530 420 325
16±40 470 375 285
40±100 430 345 255
100±250 380 305 225
60 J. Grum et al. / Journal of Materials Processing Technology 114 (2001) 57±70

processes such as grinding, boring, milling, etc., new addi- Grinberg et al. [6] treated various undercooling conditions of
tional stresses in the thin surface layer are produced while tool steel M7 between the ®rst and the second tempering
their in¯uence in greater depths is almost negligible. The which ensured a minimum reduction of hardness and by
circumstances are, however, different in working of blanks/ almost 50% increased toughness with dimensional devia-
workpieces where residual stresses are found across the tions within the limits expected.
whole cross-section of a product. Residual stresses occur-
ring after various heat-treatment processes or even thermo- 2.4. Determination of cooling curves and cooling rates
mechanical treatment of blanks/workpieces, where residual
stresses are a result of volume changes which are themselves Knowledge of quenching agents in terms of cooling
due to temperature differences and to phase transformations intensity and knowledge of physical phenomena in the
between the core and the surface, are very important. boundary layer permit improvement of monitoring and heat
In heating and subsequent quenching of blanks/work- treatment technology. At present, various methods of
pieces, it is very important that internal stresses of work- description of the phenomena in the workpiece/quench-
pieces are always lower than yield stress of the material at a ing-agent boundary layer are used. The Technical Commit-
given temperature in order to avoid distortion of the work- tee ``Scienti®c and Technological Aspects of Quenching'' of
piece. Moreaux et al. [1] treated in their discussion the size the IFHT (International Federation for Heat Treatment) has
of distortions after quenching of specimens made of alumi- prepared a standardised proposal, i.e. laboratory test for
nium alloys with 4 mm in thickness in water (20 and 608C) quenching oils which is available as international standard
as well as in polymeric water solutions and air. Account has ISO/DIS 9950. Classi®cation of various quenching agents
to be taken of which of the quenching conditions of the with regard to their physicochemical properties in the
specimen ful®lled the requirements related to speci®ed quenching process can be made by using a probe made of
mechanical properties. This indicates that only certain cool- a certain material, which is adapted in size and shape. A
ing rates of the specimen ensure the required mechanical probe is thus a test specimen made of materials having no
properties. With these requirements also other conditions, phase transformations in the solid state. Thus, errors in
i.e. minimum distortions of the specimen, should be ensured. determination of cooling intensity of a certain quenching
The tests performed on specimens often give good informa- agent due to an additional heat source produced by phase
tion on effects of individual quenching agents. In particular, transformations in the temperature interval of specimen
cases involving workpieces having exacting shapes, the quenching may be avoided [7,8]. The probe heated to the
conditions most often change and the quenching conditions temperature speci®ed is then immersed into a quenching
have to be adapted to the workpiece shape itself. agent with an automatic shift to ensure equal conditions
Filice and Pritchard [2] discussed two quenching methods during immersion, and, consequently, comparable results
for exacting aircraft components made of high-strength after quenching.
steels in order to reduce distortions. Heat treatment was The test results obtained in laboratory testing of quench-
conducted in vacuum ovens and by oil quenching in the ®rst ing agents are shown in Figs. 1 and 2. They are shown as
case and by combined gas/oil quenching in the second case. cooling curves in a temperature/time diagram (T±t) and in
With the combined gas/oil quenching to an intermediate
phase, i.e. to a temperature of 5408C, and then by oil
quenching to ambient temperature, temperature differences
between the surface and the core reduced and more favour-
able conditions in transformation of an undercooled auste-
nite were produced. The quenching method selected ensured
lower internal stresses during heat treatment and a lower
level of distortion after heat treatment was obtained. For the
selected components, two to four times lower level of
distortion was obtained by the combined gas/oil quenching.
Thuvander [3] calculated cooling rates and determined
changes in tool dimensions after quenching of tools of
various shapes by means of a selected physical model using
the ®nite element method. The distortions calculated in
accordance with the model were then compared to measured
dimensional variations in the size of the tool concerned after
hardening test. A similar study was made also by Tszeng
et al. [4] on a hollow cylindrical part.
Linggoputro [5] made a study on changes in dimensions Fig. 1. Comparison of temperature±time curves for immersion quenching
on adapted ``Navy-C'' rings after quenching of steel AISI of an Inconel 600 alloy specimen 112:5  600 mm2 in different quench-
1045 in water, and in 10 and 30% polymeric water solutions. ing agents, tested at 258C without agitation.
 J. Grum et al. / Journal of Materials Processing Technology 114 (2001) 57±70 61

Fig. 3. CCT diagram for the alloyed heat-treatable steel 42CrMo4 with a
Fig. 2. Comparison of cooling rate curves for immersion quenching of an
diameter of 95 mm; cooling curves are plotted for water quenching, oil
Inconel 600 alloy specimen 112:5  600 mm2 in different quenching
quenching and air quenching for the surface ``S'' and the core ``C''.
agents, tested at 258C without agitation.

diagrams showing the dependence of temperature on the parts with low masses as far as mechanical properties after
cooling rate of the quenching agent (T±v). Fig. 1 shows heat treatment and unwanted side effects (soft spots, dis-
cooling curves as recorded with a device QUENCHOTEST, tortions, cracks) are concerned.
developed by ``The Swedish Institute of Production Engi- Prediction of microstructure changes after hardening of
neering Research'' (IVF) in Gothenburg and indicating specimens having different masses being required and
quenching of a probe made of INCONEL from a tempera- quenching being performed in various quenching agents,
ture of 8508C in various quenching agents. The same ®gure a continuous cooling transformation (CCT) diagram was
(right) symbolically shows a characteristic cooling curve used for the steel concerned [10,11]. Fig. 3 shows a CCT
showing individual steps of heat transfer from the heated diagram for 42CrMo4 steel. The diagram shows the cool-
probe to the quenching agent [9]. The cooling curves show ing curves for quenching a steel bar with a diameter of
differences in heat transfer, in quenching, from the heated 95 mm in different quenching agents such as water, oil, air.
probe to the quenching agent in a period of 60 s. An average or expected hardness value at the surface was
Fig. 2 shows changes in the cooling rate of the probe in the 58 HRC for water quenching which ensures, in the surface
temperature interval from 850 to 208C for individual layer, a complete transformation into martensite. The
quenching agents treated. Water quenching of machine parts average hardness obtained in the surface layer after oil
with the highest quenching rate, i.e. around 2008C/s, pro- quenching ensures a martensite±bainite microstructure
ceeds in the temperature interval between 650 and 5008C with hardness of 30 HRC. In air quenching, which is
(curve 1). When using a polymeric water solution with a slowest, a perlite±ferrite microstructure with hardness of
10% Aquatensid concentration, the highest cooling rates are about 220 HB is ensured in the surface layer as well as in
achieved in the temperature interval between 750 and 6508C the core.
(curve 2), the cooling curve, however, being very similar to In our case quenching was limited to cylindrical speci-
that obtained in water quenching. mens made of the same heat-treatment steel having the same
Another important difference is the difference in the diameter …D ˆ 35 mm† but different heights, i.e. 5 and
cooling rates below a temperature of 3508C. It indicates 15 mm. Cold specimens were put in an oven heated up to
that in quenching in the polymeric water solution in the ®nal 6008C. When this temperature was reached by the surface as
stage of martensite transformation somewhat lower cooling well as by the core, they were gradually heated to a hard-
rates are obtained than in water quenching. Such quenching ening temperature of 8508C.
is more favourable since internal stresses in quenching and The cylindrical specimens had four thermocouples
consequently the size of distortions and/or residual stresses inserted at selected measuring points. The latter permitted
are reduced. The lowest cooling rates are obtained in to determine two cooling rates at the characteristic points of
quenching a probe in the oil Rapid Extra (curve 3). The the surface and those of the core. Fig. 4 shows a cylindrical
cooling rate gets considerably slowed down already before specimen with the thermocouples inserted. The heated
martensite transformation and becomes nearly constant cylindrical specimens were carefully transmitted to a cool-
from a temperature of 4008C. It is to be expected for the ing bath, i.e. immersed into a quenching agent. The experi-
oil to be exceptionally favourable for quenching of machine mental set-up permitted us to record cooling curves during
62 J. Grum et al. / Journal of Materials Processing Technology 114 (2001) 57±70

Fig. 4. System of designation of individual specimens and measuring points.

quenching of the cylindrical specimens in: The experimental set-up is composed of:
 10% Aquatensid polymeric water solution;  specimens having different masses;
 oil Olmakal Rapid Extra.  thermocouples connected at four different measuring
points;
For measurement of the cooling curves during the quench-  connecting terminal SC2071;
ing process, four thermocouples were inserted in the centre  link cable for signal routing to the PC;
of specimens having different heights but the same diameter  programme card LAB-PC-1200/AI;
of 3 mm. The thermocouples were staggered by 908 each; in  software package LabVIEW.
the core they were located at r ˆ 0 and r ˆ R=4 and at the
surface at r ˆ 34 R and r ˆ R. The measuring system comprises the connecting terminal
This experimental procedure was to indicate the in¯uence SC2071 which permits connection of analogous and digital
of cooling intensity of a quenching agent on the cylindrical signals via the link cable NB1 to a compatible PC. Measur-
specimens having different masses. In order to be able to ing of temperature variations in specimen quenching is
predict microstructure, hardness and strength variations in stored by the PC by means of the card LAB-PC-1200/AI
various specimen masses, a computer-aided experimental and the software package LabVIEW. The software package
set-up for temperature measurement and data processing ensures data input and plotting of the cooling rates in a CCT
was elaborated in our laboratory of heat treatment as shown diagram for individual quenching agents and individual
in Fig. 5. The experimental set-up was composed of speci- measuring points of the specimens.
mens of various masses in which, at selected measuring The data on time-dependent temperature variations, i.e.
points, thermocouples for measurement of cooling curves cooling curves, were then processed by a software package
were inserted. The cooling curves were, via a multi-channel EXCEL 5.0. The results obtained were displayed as dia-
temperature converter, transferred to a PC for subsequent grams, tables and column charts. The cooling curves mea-
data processing. Thus, it is possible for us to monitor sured at different measuring points of the specimens having
physical phenomena in the boundary layer via cooling different masses permit several displays in diagrams of:
curves which were recorded for an individual quenching  cooling curves;
agent [12,13].  cooling rates as a function of temperature;

Fig. 5. Experimental set-up for capturing cooling curves on specimens having different masses and being cooled in different quenching agents.
 J. Grum et al. / Journal of Materials Processing Technology 114 (2001) 57±70 63

 cooling curves in a CCT diagram; In the CCT diagrams in Fig. 7, cooling curves are plotted
 predicting of portions of individual microstructure for the surface and the core, the specimen heights being
phases; 5 mm (Fig. 7A) and 15 mm (Fig. 7B) after quenching in the
 predicting hardness after quenching of the specimens oil Rapid Extra.
having different masses in different quenching agents. The variations in the cooling rates in the CCT diagram
permit drawing the following conclusions:

3. Experimental results  the cooling curves recorded at the specimen surfaces

differ considerably with different specimen heights;
3.1. Cooling curves and cooling rates  the cooling curves recorded in the specimen cores also
differ considerably with different specimen heights;
By temperature measurement at individual measuring  in the specimen with a smaller height, a predominantly
points cooling rates were obtained for the selected cylind- martensite microstructure with an increased portion of
rical specimens. bainite in the specimen centre is obtained;
The number of measured and calculated data for indivi-  in the specimen with a bigger height there is also a
dual specimens being very high, only the results, to be predominantly martensite microstructure with an
looked over with more ease, obtained at the measuring increased bainite portion than in the specimen with a
points no. 1, i.e. at the specimen surface, and no. 4, i.e. smaller height as well as a smaller portion of ferrite are
in the specimen core, are shown. The ®rst CCT diagram in obtained.
Fig. 6 shows the cooling rates for the surface and the core of
the specimen with 5 mm in height while the second diagram A comparison of the results obtained in quenching the
shows the cooling rates for the surface and the core of the specimens having different height in the 10% polymeric
specimen with 15 mm in height after quenching in the 10% water solution and in oil shows that in these cases a more
polymeric water solution. intensive quenching agent should be selected for ef®cient
The variations of the cooling curves in the CCT diagrams quenching. Such a quenching agent is the 10% polymeric
permit drawing the following conclusions: water solution, which proved very ef®cient and most sui-
table as regards microstructure changes and the hardness
 The cooling curves for the surface are very similar values attained in the specimen cross-sections.
regardless of different height of the specimens. The cool- Fig. 8 shows the cooling rates attained at the specimen
ing rate is higher than the critical cooling rate so that a surface and in the core in the case of quenching in the 10%
martensite microstructure was obtained at the specimen polymeric water solution. The upper ®gure shows the cool-
surface. ing rates attained with the specimen having 5 mm in height
 The cooling curves for the core differ a little which and the lower one the cooling rates attained with the speci-
indicates that the cooling rate is a little higher with a men with 15 mm in height. A comparison of both diagrams
smaller mass of the specimen. In both cases a smaller permits drawing the following conclusions:
portion of bainite with martensite is obtained in the  The cooling rates in the specimen with a bigger height are
specimen centre, which affects the hardness variation more favourable. The maximum cooling rate achieved at
in the specimen cross-section. the surface is the same for both masses and amounts to

Fig. 6. CCT diagram showing cooling curves for surface and core of specimens with 5 mm in height (A) and 15 mm in height (B) after quenching in 10%
polymeric water solution.
64 J. Grum et al. / Journal of Materials Processing Technology 114 (2001) 57±70

Fig. 7. CCT diagram showing cooling curves for surface and core of the specimens with 5 mm in height (A) and 15 mm in height (B) after quenching in oil
Rapid Extra.

around 2808C/s, but with the bigger mass it is attained in a Extra. The upper ®gure shows the cooling rates achieved in
somewhat narrower temperature interval. The cooling the specimen with 5 mm in height and the lower one those
rates at the surface are very similar in the temperature achieved in the specimen with 15 mm in height.
interval of martensite transformation and in spite of A comparison of the two diagrams permits drawing the
different masses, quite favourable as regards internal following conclusions:
stresses during quenching and residual stresses after
quenching, respectively.  With the smaller specimen mass the maximum cooling rate
 The cooling rates achieved in core quenching, however, achieved at the surface amounts to around 2308C/s and with
differ more strongly in the case of different masses. Thus, the greater specimen mass to around 2208C/s. The max-
in the specimen with a smaller mass the maximum cool- imum cooling rate with the bigger specimen mass as well as
ing rate amounts to around 958C/s and in the specimen with the smaller specimen mass was achieved at a tem-
with a bigger mass to only 608C/s. It is interesting that the perature of 6008C. A stronger deviation of the cooling rate
variation of cooling rate in the temperature interval of at the surface occurs in the temperature interval between
martensite transformation is more emphasised in the 600 and 3008C. The data on the cooling rates at the surface
specimen with the greater mass, which in a certain way indicate that in the greater specimen mass somewhat higher
causes formation of microstructural internal stresses dur- cooling rates are also achieved at the surface.
ing quenching.  With the smaller specimen mass the maximum cooling
rate achieved in the core is higher and amounts to around
Fig. 9 shows the cooling rates achieved at the specimen 458C/s while with the greater specimen mass it is con-
surface and in the core in quenching with the oil Rapid siderably lower and amounts to around 288C/s. Another

Fig. 8. Dependence of cooling rate on specimen temperature in quenching in 10% polymeric water solution.
 J. Grum et al. / Journal of Materials Processing Technology 114 (2001) 57±70 65

Fig. 9. Dependence of cooling rate on specimen temperature in quenching in oil Rapid Extra.

very important finding is that in the smaller specimen procedure, it should be realised that a machine part is
mass the cooling rate changes from 45 to 408C/s in the damaged and is, therefore, very often unsuitable to be
temperature interval between 730 and 4508C. On the built-in. This method is most suitable for measurement of
contrary, with the specimen having a greater mass, the internal stresses in various loaded structures. A damage due
average cooling rate of around 208C/s is achieved in the to a small drill-hole in the surface of a structure is usually of
same temperature interval, i.e. between 730 and 4508C. no importance since it does not jeopardise its further opera-
This average cooling rate in the core of the specimen tion. Only in the cases in which the drill-hole does not impair
having a smaller mass is only a half of that obtained in ®tness for purpose of a structure may even this method for
quenching the specimen with a greater mass. We will try measurement of residual stresses or stresses in a particular
to relate the circumstances achieved in various quenching element of the loaded structure be considered non-destruc-
conditions of the specimens with the residual stresses tive. It may be stated generally that the method is very
produced and measured in the specimens. practical, quick, cost-effective and reliable enough; there-
fore, it is used for various applications nowadays. An
3.2. Measurement of residual stresses advantage of the method is its simplicity and adaptability
to different applications since measurements may be per-
Measurement of residual stresses was carried out by formed in a laboratory, in production or even on site. This
means of a relaxation method called hole drilling strain relaxation method is, therefore, very often used for measure-
gage method, which is, due to the small diameter and depth ment of residual stresses.
of the drill-hole, classi®ed among semi-destructive testing For measurement of residual stresses at a selected point
methods. The method is suitable for measurement of resi- satisfactory conditions for placing a three-legged resistance
dual stresses, i.e. measuring rosette should be ful®lled. The method is based
on measurement of speci®c deformations in individual
 measurement on specimens having adapted size and
directions with regard to the location of the three-legged
shape;
resistance measuring rosette, which allows subsequent
 direct measurement on machine parts;
calculation of main stresses and determination of their
 measurement on different structures.
location.
Measurement of residual stresses in the specimens The procedure of measuring residual stresses is the follo-
machined in different machining conditions may offer good wing:
insight into in¯uence of individual parameters of heat
treatment. Findings on in¯uences of individual parameters  Selection and preparation of the surface of a measuring
in heat treatment and ®ndings based on the specimens can point.
very successfully be applied to machine parts of simple or  Adhesive bonding of the resistance measuring three-
exacting shapes as well. Measurement of residual stresses in legged rosette on the surface of the measuring point.
machine parts seems to be most appropriate since it permits  Connection of the resistance measuring three-legged
a simpler and more accurate analysis of the in¯uence rosette via electrical connection and electronic computer
of individual parameters of heat treatment. With this card for signal capturing and transmission to a PC.
66 J. Grum et al. / Journal of Materials Processing Technology 114 (2001) 57±70

 Optical placing of a drilling machine at the measuring are observed in absolute stress values. Thus, in both cases
point of the specimen, which is at the same time the centre residual tensile stresses are observed to the depth analysed
part of the resistance measuring rosette. but they are highest at the surface and then gradually reduce
 Calibration of the experimental set-up. after a depth of 0.1 mm. A further change in residual stresses
 Drilling of the specimen material at the centroid of the to the depth analysed is negligible, i.e. very slow. The
resistance measuring three-legged rosette and measure- diagram shows the two main residual stresses, i.e. the
ment of specific deformations in individual directions of highest one s1 and the lowest one s2. At the surface of
legs 1, 2 and 3. the specimen with a higher mass a maximum residual stress
 Data processing, calculation of the main residual stresses of the tensile character amounting to 750 N/mm2 was
and angle orientation of the main stresses. achieved. It is considerably higher than the one measured
 For drilling, a diamond drill bit with a diameter of 1.8 mm in the specimen with a smaller mass. In the specimen with a
was used. Drilling was carried out to the depth which smaller mass the maximum residual stress was also tensile
equals the diameter, i.e. 1.8 mm, with an increment of stress amounting to 300 N/mm2.
drilling Dz ˆ 0:1 mm. In the depth exceeding 0.1 mm, the average value of the
 Data on the resistance measuring rosette selected, i.e. type highest main stress in the specimen with a smaller mass
TEA-06-062RK-120 manufactured by Vishay Measure- amounts to around 160 N/mm2. Both main stresses are
ment Group, are contained in software package of the tensile character to the analysed depth of both
RESTRESS. specimens.
 Measurements were carried out by means of the software Fig. 11 shows variations of main residual stresses in the
package LABVIEW. specimen with 5 mm in height (Fig. 11A) and for the
 Calculated input data, their displays in tables and graphic specimen with 15 mm in height (Fig. 11B) after quenching
charts, i.e. results of measurement of residual stresses, in the oil Rapid Extra. A maximum main residual stress in
were carried out by means of the software package the specimen with a smaller mass is a tensile residual stress
RESTRESS. throughout while the minimum main residual stress to a
 Selected input data on the steel: depth of 0.37 mm is a tensile stress and then passes over to a
 Rp0:2 ˆ 780 N=mm2 : yield stress. compressive stress. The maximum tensile residual stress is
 E ˆ 210:000 N=mm2 : Young's modulus of elasticity. also found at the surface and amounts to around 85 N/mm2
 n ˆ 0:33: Poisson's ratio. and then gradually reduces to 35 N/mm2 in a depth of
0.9 mm. The variation of residual stresses in the specimen
Residual stresses were measured by means of the afore- with a greater mass, however, is quite different. The max-
mentioned relaxation method by drilling holes and simulta- imum main residual stress at the surface is barely noticeable
neous measurement of deformations with resistance and is of the tensile character and amounts to around
measuring rosettes. Fig. 10 shows the through-thickness 20 N/mm2. The maximum main stress is of the tensile
variation of residual stresses for the specimen with 5 mm character and amounts to 90 N/mm2 in a depth of 0.2 mm
in height (Fig. 10A) and for the specimen with 15 mm in and then gradually reduces to around 72 N/mm2 in a depth of
height (Fig. 10B) after quenching of the specimens in the 0.9 mm. The minimum main stress is mainly of a compres-
10% polymeric water solution. The variations of stresses in sive character; it gradually increases from a depth of
both specimens are very similar but important differences 0.13 mm to a depth of 0.9 mm and ®nally amounts to

Fig. 10. Variation of residual stresses for the specimen with 5 mm in height (A) and for the specimen with 15 mm in height (B) after quenching in 10%
polymeric water solution.
 J. Grum et al. / Journal of Materials Processing Technology 114 (2001) 57±70 67

Fig. 11. Variation of residual stresses for the specimen with 5 mm in height (A) and for the specimen with 15 mm in height (B) after quenching in oil Rapid
Extra.

65 N/mm2. The results, in terms of values and variations of differ in their signs, the condition obtained is a result of
the main residual stresses, discussed above indicate that: the conditions in heat removal from the heated specimen
to the ambient quenching agent.
 when quenching in a more intensive quenching agent, i.e.
the 10% polymeric water solution, the maximum main 3.3. Microhardness
residual stress obtained is of the tensile character;
 the maximum values of the main residual stresses in In order to be able to explain in¯uences of quenching of
quenching of a greater mass are higher than in quenching the specimens concerned having different masses in differ-
of a smaller mass; stresses in a greater mass are two or ent quenching agents, we decided to measure microhard-
three times higher than in a smaller mass; ness. Microhardness was measured in the transverse
 after quenching in a less intensive quenching agent, i.e. oil direction in the specimen centre and in the axial direction
Rapid Extra, the maximum main residual stresses are throughout the specimen height. For microhardness mea-
three or seven times lower than in quenching of the same surements, a load of 3 N was selected which permitted a very
specimens in the 10% polymeric water solution; accurate determination of the variation of microhardness.
 taking into account that in the specimen with a smaller Fig. 12 shows a variation of microhardness in the transverse
mass after quenching in the oil Rapid Extra, the variations (^) and axial (^) directions of the cylindrical specimen
of the maximum and the minimum main residual stresses having 5 mm in height after quenching in the 10% polymeric

Fig. 12. Variation of microhardness in transverse and axial directions of the cylindrical specimen with 5 mm in height after quenching in 10% polymeric
water solution (A) and after quenching in oil Rapid Extra (B).
68 J. Grum et al. / Journal of Materials Processing Technology 114 (2001) 57±70

water solution (Fig. 12A) and after quenching in the oil 10% polymeric water solution and amounts to 820 HV0.3
Rapid Extra (Fig. 12B). The results of measurements show while it is lower, at the same point, after oil quenching and
that: amounts to 720 HV0.3.
 The lowest microhardness value in the axial direction is
 A microhardness value in the transverse direction of the obtained in quenching in the 10% polymeric water solu-
specimen at its surface after quenching in the 10% poly- tion and amounts to 671 HV0.3. It is higher than the lowest
meric water solution amounts to 835 HV0.3 and is some- microhardness value obtained after quenching in the oil
what higher than the microhardness value measured after Rapid Extra which is 616 HV0.3.
oil quenching of another specimen which amounts to 710
Fig. 13 shows the variation of microhardness in the
HV0.3. Data on the microhardness values measured con-
transverse and axial directions of the cylindrical specimen
firm assessment of the microhardness values expected on
with 15 mm in height after quenching in the 10% polymeric
the basis of variations of cooling curves in a CCT dia-
water solution (Fig. 13A) and after quenching in the oil
gram. The results of microhardness values measured in
Rapid Extra (Fig. 13B). The results of microhardness mea-
the transverse direction of the specimen confirm that the
surements show:
microstructure obtained throughout its volume is mainly
martensite with a smaller portion of very fine bainite.  The microhardness measured in the transverse direction
 The microhardness value measured throughout the speci- of the cylindrical specimen with 15 mm in height is
men height, i.e. in its axial direction, is somewhat higher considerably lower than the one measured in the specimen
at the auxiliary side of the specimen but then gradually with 5 mm in height. The highest microhardness value
reduces with the height of the specimen. Immersion of the was obtained at the surface and amounts to 712 HV0.3
specimen during quenching was carried out by moving after quenching in the 10% polymeric water solution and
the specimen slowly in the axial direction into the quench- to 680 HV0.3 after quenching of another sample in the oil
ing agent. The point having the highest microhardness Rapid Extra. A comparison of the results of microhard-
value was the first to be immersed into the quenching ness measurements in the transverse direction of the
agent. It is thought that the influence on the difference specimen permits a conclusion that the influence of the
obtained in the microhardness value measured in the axial specimen mass as well as that of the type of quenching
direction was smallest and negligible. A more important agent used are very distinctive.
influence on the difference in the microhardness values  The microhardness measured in the axial direction, i.e.
measured is attributed to different cooling rates obtained throughout the specimen height, is almost the same at the
throughout the specimen height. Variations of microhard- inferior as well as the upper front surface. The highest
ness and the differences obtained in the microhardness microhardness value measured at the upper as well as at
values measured are very similar in both cases, i.e. in the inferior surface is 723 HV0.3. A somewhat slower
quenching in the 10% polymeric water solution and in oil reduction of microhardness at the inferior side of the
quenching. The highest microhardness value is obtained specimen is observed after quenching in the 10% poly-
at the auxiliary side of the specimen after quenching in the meric water solution.

Fig. 13. Variation in microhardness in transverse and axial directions of the cylindrical specimen with 15 mm in height after quenching in 10% polymeric
water solution (A) and after quenching in oil Rapid Extra (B).
 J. Grum et al. / Journal of Materials Processing Technology 114 (2001) 57±70 69

 The lowest microhardness value measured in the axial the axial direction. In the cases of quenching in which a
direction, i.e. 590 HV0.3, is obtained in quenching in the predominantly martensite microstructure was achieved
10% polymeric water solution while the lowest micro- throughout the specimen volume, very high tensile strength
hardness value obtained in oil quenching is 631 HV0.3. values were obtained at the specimen surface. On the
contrary, in the cases in which a martensite±bainite micro-
The variation of microhardness in the axial direction structure was achieved, considerably lower tensile or even
differs considerably from the expected variation of micro- compressive residual stresses were obtained. The investiga-
hardness in quenching in the 10% polymeric water solution tion showed that knowledge of the microstructure and/or the
as well as in oil quenching. It is estimated that the differ- microhardness variation permits a successful assessment of
ences in variations of microhardness throughout the height changes in the size of specimen or workpiece as well as of
of the cylindrical specimen are due mainly to workpiece/ the size and variation of the expected and/or residual internal
quenching-agent boundary layer in¯uences. stresses.

4. Conclusions References

In steel hardening, it is necessary to select a proper [1] F. Moreaux, G. Beck, B. Clement, J. Olivier, New agueous solution
for metallic alloys quenching Ð importance of agitation, in: G.
austenitizing temperature and duration considering its
Krauss (Ed.), Heat Treatment and Surface Engineering: New
initial microstructure in order to ensure a ®ne and homo- Technology and Practical Application, Proceedings of the Sixth
geneous austenite microstructure which consequently, after International Congress on the Heat Treatment of Materials, Vol. 167,
quenching, ensures formation of ®ne martensite. Quench- American Society of Metals, Chicago, IL, 1988.
ing of the steel concerned should be such that the actual [2] C. Filice, J.E. Pritchard, Techniques for distortion control during
vacuum oil quenching of high strength steel aircraft components, in:
cooling rate approximates to the critical cooling rate on
D.L. Milan, D.A. Poteet, G.D. Pfaffmann, V. Rudner, A. Muehlbauer,
microstructure. A proper cooling rate of the steel concerned W.A. Albert (Eds.), 17th ASM Heat Treating Society Conference
during quenching, however, may be ensured only if the Proceedings Including the First International Induction Heat Treating
cooling capacity of the quenching agent and the in¯uence of Symposium, Vol. 133, American Society of Metals/Heat Treating
mass, which have a synergetic effect on microstructure Society, Indianapolis, IN, 1997.
[3] A. Thuvander, Calculation of distortion of the tool steel dies during
formation after quenching, are well known. Specimen
hardening, in: Proceedings of the Second International Conference on
quenching should proceed in a way to ensure the smallest Quenching and Control of Distortion, Cleveland, OH, Vol. 297, 1997.
temperature differences possible between the surface and [4] T.C. Tszeng, W.T. Wu, L. Semiatin, A sensitivity study of the process
the core during quenching and an optimum microstructure model for predicting the distortion during heat treating, in: G.E.
after quenching. The temperature differences between the Totten, M.A.H. Howes, S.J. SjoÈstrom, K. Funatani (Eds.), Proceed-
surface and the core in the zone of martensite transforma- ings of the Second International Conference on Quenching and
Control of Distortion, Vol. 321, American Society of Metals/The
tion and the rate of rising of martensite transformation are of Materials Information Society, Cleveland, OH, 1996.
the main importance. [5] R. Linggoputro, Cooling characteristics of a PAG polymer quenchant
The investigation conducted provides a comprehensive and its application on medium carbon steel, in: G.E. Totten, M.A.H.
insight into the cooling curves for the 10% polymeric Howes, S.J. SjoÈstrom, K. Funatani (Eds.), Proceedings of the Second
International Conference on Quenching and Control of Distortion,
water solution and for the oil Rapid Extra, which were
Vol. 571, American Society of Metals/The Materials Information
recorded by the device QUENCHOTEST, and into appro- Society, Cleveland, OH, 1996.
priate cooling rates. By means of temperature measure- [6] D.M.K. Grinberg, A. Grinberg, I.V. Carrasco, The M-7 steel: sub-
ments in the specimens made of the heat-treatment steel zero treatments, mechanical properties and temperatures, in: G.E.
42CrMo4 microstructure and hardness after quenching Totten (Ed.), Proceedings of the First International Conference on
Quenching and Control of Distortion, Vol. 51, American Society of
were determined and found to be in accordance with a
Metals, Chicago, IL, 1992.
CCT diagram. The variations of microhardness in the [7] B. LisÏcÏicÂ, Steel heat treatment, in: G.E. Totten, M.A.H. Howes
radial and axial directions permit assessment of hardening (Eds.), Steel Heat Treatment Handbook, Vol. 527, Marcel Dekker,
ef®ciency. New York, 1997 (Chapter 8).
The results of microhardness variation in the axial direc- [8] B. LisÏcÏicÂ, Hardenability, in: G.E. Totten, M.A.H. Howes (Eds.), Steel
tion warned us of the extreme importance of quenching. The Heat Treatment Handbook, Vol. 93, Marcel Dekker, New York, 1997
(Chapter 3).
cooling rates in the specimens with a smaller mass are [9] S. Segerberg, J. Bodin, Correction between quenching characteristics
extremely high; therefore, it is necessary to reduce the of quenching media and hardness distribution in steel, in: T. Bell
duration of specimen immersion during quenching. The (Ed.), Proceedings of the Heat Treatment '84, London, The Metals
results obtained con®rmed that irregularity in the micro- Society, London, 1984.
[10] J. Bodin, S. Segerberg, Measurement and evaluation of the quenching
structure obtained after quenching may be avoided by the
power of quenching media for hardening, in: G.E. Totten (Ed.),
method of quenching used. The conditions after immersion Proceedings of the First International Conference on Quenching and
quenching were described, in addition to microhardness Control of Distortion, American Society of Metals, Chicago, IL,
measurement, also by measurement of residual stresses in 1992, pp. 1±12.
70 J. Grum et al. / Journal of Materials Processing Technology 114 (2001) 57±70

[11] C.E. Bates, G.E. Totten, Application of quench factor analysis to Including the Liu Dai Memorial Symposium, Vol. 645, American
predict harness under laboratory and production condition, in: G.E. Society of Metals/The Materials Information Society, Cleveland,
Totten (Ed.), Proceedings of the First International Conference on OH, 1998.
Quenching and Control of Distortion, Vol. 33, American Society of [13] J. Grum, S. BozÏicÏ, M. ZupancÏicÏ, In¯uence of quenching process
Metals, Chicago, IL, 1992. parameters on residual stresses in steels, in: G.E. Totten, B. LisÏcÏicÂ,
[12] J. Grum, S. BozÏicÏ, R. LavricÏ, In¯uence of mass of steel and a H.M. Tensi (Eds.), Proceedings of the Third International Conference
quenching agent on mechanical properties of steel, heat treating, in: on Quenching and Control of Distortion, Vol. 436, ASM/Heat
R.A. Wallis, H.V. Walton (Eds.), Proceedings of the 18th Conference, Treating Society, Prague, 1999.