Journal of Manufacturing Processes 22 (2016) 115119

Contents lists available at ScienceDirect

Journal of Manufacturing Processes

journal homepage: www.elsevier.com/locate/manpro

Inuence of prior cold deformation on microstructure evolution of

AISI D2 tool steel after hardening heat treatment
Hadi Ghasemi Nanesa , Julien Boulgakoff, Mohammad Jahazi
Department of Mechanical Engineering, cole de Technologie Suprieure, 1100 rue Notre-Dame Ouest, Montral, QC, Canada H3C 1K3

a r t i c l e

i n f o

Article history:
Received 26 October 2015
Received in revised form 6 February 2016
Accepted 12 February 2016
Available online 31 March 2016
Keywords:
Electron microscopy
Hardness measurements
Steel
Bulk deformation
Grain renement
Phase transformation

a b s t r a c t
4 mm thick plates of AISI D2 steel were subjected to two different percentage of cold rolling prior to
quench hardening. Prior austenite grain size and volume fraction of carbides were reduced after hardening heat treatment for both 10% and 20% cold-rolled samples in comparison with non-deformed one.
Hardness measurements revealed that pre-deformation resulted in a 3 to 4 HRC decrease. Scanning electron microscopy images revealed more uniform mean plate size of martensite for the cold rolled samples.
The obtained results are analyzed in the framework of the existing theories on martensitic transformation and the inuence of strain on second phase dissolution kinetics. The drop in hardness is related
to the effect of pre-deformation on the dissolution rate of carbides. The volume fraction of martensite
compared to non-deformed condition didnt show dominant reduction.
2016 The Society of Manufacturing Engineers. Published by Elsevier Ltd. All rights reserved.

1. Introduction
AISI D2 tool steel is widely used in die making industry and
different type of cutting tools [1,2]. High hardness and ultimate
strength combined with good wear resistance are characteristics of
this steel. The superior mechanical properties are due to its chemical composition (high carbon and alloying elements) and specic
processing conditions that allow the formation of various types of
strong phases (e.g. martensite and bainite) and hard second phase
particles. The processing of the alloy consists of ingot casting followed by forging and/or hot rolling [3]. The as-deformed material is
then subjected to quench and temper operations [3]. Austenitization temperatures between 1233 K and 1353 K are used to allow for
higher dissolution of alloying elements in austenite and partial carbide dissolution before quenching [14]. The higher dissolution of
alloying elements increases the propensity for solid solution hardening upon quenching and the undissolved carbides halt extreme
austenite grain growth [5].
Phase transformation during conventional quenching of this
steel has received much interest because of the high hardness
(close to 63 HRC) and hence brittleness of the material after
quenching. Most reports are focused on the application of a cryogenic treatment (below 200 K) and subsequent tempering process
in the temperature range of 423773 K to reduce brittleness by

Corresponding author. Tel.: +1 514 396 8974; fax: +1 514 396 8530.
E-mail address: hadi.ghasemi-nanesa.1@ens.etsmtl.ca (H.G. Nanesa).

transforming more austenite to martensite, relieving internal

stresses, and precipitating temper carbides [69].
Cold deformation is widely used in metallic materials for grain
rening and strengthening of the initial microstructure [5]. The
inuence of cold deformation of martensite on recrystallization
behavior of low carbon steels was studied by Tokizane et al.
[10]. Chojnowski and Tegart [11], and Joo et al. [12] reported
that prior cold deformation accelerated the spherodiziation process in hypereutectoid steels. In addition, it is well established
that reduction in prior austenite grain size (PAGS) affects the start
temperature of martensitic transformation (Ms ) as discerned by
Guimares [13,14], Lee and Lee [15], and Yang and Bhadeshia
[16]. The smaller mean PAGS pushes down the Ms temperature
and increases the stability of retained austenite instead of martensite [1315]. However, examination of the literature indicated
that no work has been reported on the inuence of prior cold
deformation on the initial microstructure of AISI D2 steel and its
impact on microstructure evolution during quenching and subsequently the nal mechanical properties. The present work has
therefore objective to investigate the effect of pre-deformation
on microstructure evolution, carbide dissolution behavior, mean
PAGS reduction, hardness changes, and the kinetics of martensitic
transformation after hardening cycle in the investigated AISI D2
alloy. Adding cold deformation route before hardening heat treatment of tool steels is expected to improve the strength and wear
properties of these alloys and therefore result in higher quality
products and longer service life for cutting tools made of these
alloys.

http://dx.doi.org/10.1016/j.jmapro.2016.02.002
1526-6125/ 2016 The Society of Manufacturing Engineers. Published by Elsevier Ltd. All rights reserved.

116

H.G. Nanesa et al. / Journal of Manufacturing Processes 22 (2016) 115119

Fig. 1. Microstructure exhibits martensite plates, second phase, highlighted prior austenite grain boundaries, and primary carbide (PC) with other ner carbides either M7 C3
or M23 C6 : (a) and (b): non-deformed, (c) and (d):10% cold-rolled, (e) and (f): 20% cold-rolled and hardened samples, highlighted prior austenite grain boundaries are shown
in part (b), (d), and (f) in a consecutive order corresponding to conditions (a), (c), and (e), respectively; smaller mean martensite plate size in (c) and (e) compared to (a).
Black highlighted particles in parts (b), (d), and (f) correspond to intergranular carbides which were eliminated during grain size measurements.

2. Experimental material and procedures

AISI D2 tool steel plates with nominal composition of C 1.54Si
0.33Mn 0.32Cr 11.88Mo 0.76V 0.75P 0.008S 0.008 (wt%)
and 4 mm initial thickness were given 10% or 20% reductions in
thickness by cold rolling. The rolling machine was a 2-high laboratory milling mills fabricated by Fenn Model: 4-046 with anvil
diameters of 100 mm and maximum rolling pressure of 136 kN. In
order to avoid internal cracking because of high work-hardening
susceptibility of the alloy, a rolling speed of 1 m min1 was used.
Samples were cut from the cold rolled bars and then heated to
1303 K for austenitizing and maintained for 1200 s followed by
water quenching (hardening cycle). For microstructural investigations and hardness measurements, cold rolled samples were sliced
in half along the width and their central faces parallel to top surface
were polished and etched.
Samples from the as-received bar were also given similar treatment for comparison purposes. An etchant with the following

composition 40 g NaOH + 60 g H2 O + 15 g NaNO3 initially proposed

by Goun et al. [17] was modied and successfully used to reveal
martensite as well as prior austenite grain boundaries. HITACHI
TM3030 scanning electron microscope (SEM), PANalytical X-ray
diffraction machine model XPert Pro, MIP image analysis software
[18], and Rockwell C (150 kg, 10 s) macrohardness were employed
to evaluate microstructure evolution after cold rolling and hardening cycle. In order to measure the volume fraction of carbides, back
scattered electron (BSE) images showing phase contrast (carbides
as black regions and matrix bright) were taken from 3 different
areas in each sample and then were analyzed by MIP software.
The limitation for the size of carbides is considered to be 0.2 m
to eliminate the noise effect from BSE images [19,20]. In order to
measure the volume fraction of retained austenite, ASTM E975-13
standard was used. It is worth mentioning that the constant values in this standard are just for Cr and Mo targets. For this reason,
the constants for Cu-target were calculated by the present authors.
For metallographic studies, after conventional surface preparation

H.G. Nanesa et al. / Journal of Manufacturing Processes 22 (2016) 115119

Fig. 2. X-ray diffraction line proles of (a) non-deformed, (b) 10% cold-rolled, and (c)
20% cold-rolled samples. Each prole represents the set of (hkl) indicating different
diffraction planes of martensite (M) and austenite (A), superimposition of double
M7 C3 Cr7 C3 peaks, triple M7 C3 M23 C6 M peaks, double M7 C3 M23 C6 peaks at 2
close to 39 , 44.5 , and 82 , respectively.

using abrasive papers, polishing was carried out with 1 m diamond suspension. For nal polishing, sample was hold for 24 h in
an automatic vibrometer using 0.05 m diamond suspension. Carbides were classied into three different categories based on their
sizes and variation in the aspect ratio of carbides reported by Das
et al. [20]: (a) Mean diameter 1 m, (b) Mean diameter >1 m
and <5 m, and (c) Mean diameter 5 m. Mean PAGS for each
condition was calculated using 3 different images by MIP software
equipped with ASTM E112 standard. Prior austenite grain boundaries were highlighted for the accuracy of measurements. Error bar
of mean PAGS is the standard deviation of three measurements for
each condition.
3. Results
Fig. 1ae shows the microstructures of non-deformed and hardened (a and b), 10% cold-rolled and hardened (c and d), and 20%
cold-rolled and hardened (e and f) samples, respectively. Different
types of carbides, martensite and prior austenite grain boundaries were observed for the three conditions: (1) Primary carbides
formed at the austenite grain boundaries and then dispersed as
a result of hot working, (2) Other carbides (smaller in size and
distributed mostly within the grains) are the result of secondary
precipitation during normalizing heat treatment. In agreement
with the results reported by Bombac et al. [3] and Das et al. [7],
the primary carbides are probably M7 C3 and the secondary carbides as M2 C and M23 C6 . (3) Comparison between both parts (c)
and (e) with (a) shows also that the mean plate sizes of martensite
are smaller in the cold rolled samples than the non-deformed one.
XRD experiments were conducted with the view to conrm the
presence of various phases after each heat treatment and also to
study the dissolution behavior of carbides as illustrated in Fig. 2.
Four peaks of martensite (M), two small peaks of austenite (A),
and carbide peaks were identied by XPert HighScore software as
well as by searching in the published literature [68]. For 2 angles
between 35 and 55 , detected peaks correspond to M7 C3 , Cr7 C3 ,
and M23 C6 carbides, M, and A. More detailed analysis revealed that
the double M7 C3 Cr7 C3 peaks, triple M7 C3 M23 C6 M peaks, and
double M7 C3 M23 C6 peaks are superimposed at 2 values close to
39 , 44.5 , and 82 , respectively. The results reveal also that the
relative peak intensity of carbides is lower in cold deformed samples compared to non-deformed one. It is worth noting that lower
intensity in the XRD diagrams is indication of a reduction in the volume fraction of carbides. Therefore, suggesting that the prior cold

117

deformation could inuence the dissolution kinetics of carbides.

This aspect will be discussed in the following sections. Finally, the
presence of very small peaks of retained austenite is discerned at
2 values close to 43 , 50.5 ; however no signicant changes in the
relative peak intensity were found indicating that some amount of
retained austenite is present in the material independently of the
prior cold rolling. The calculations based on ASTM E975-13 showed
10%, 11%, 12% retained austenite for non-deformed, 10% cold rolled,
and 20% cold rolled samples, respectively. Cold rolling of D2 steel
could increase the volume fraction of retained austenite by maximum 2%. However, by 2% more retained austenite, no signicant
change in hardness value can be expected as reported by Beswick
[5] in SAE 52100 after cold deformation up to 64% and hardening.
The effect of cold rolling on the variation of the volume fraction of carbides was studied by image analysis of the unetched
microstructure by the method described in Ref. [6]. In fact, etching
could cause higher volume fraction measurement due to the bigger
perceptible area for carbides and lower amount for small carbides
(as they may be washed out during the etching process). To this end,
3 pictures with the same magnication were taken from randomly
selected areas showing different sizes of carbides. Back scattered
electron (BSE) detector was used to remove topographical contrasts
and use the phase contrast between the matrix and the precipitates
for better accuracy. Fig. 3 shows the image analysis results of the
size and volume fraction of carbides for the three testing conditions.
The absolute volume fraction of carbides in both cold-rolled samples is about 2% lower than the non-deformed one. Although this
amount is relatively low, however considering the repeatability of
the measurements, the obtained value indicates that cold rolling
appears to accelerate carbide dissolution process during austenitizing. Macrohardness measurements, depicted in Fig. 3d, show
continuous reduction in the hardness value from non-deformed
alloy to 10%, and then to 20% cold-rolled one. It is worth noting
that, in spite of the higher volume fraction of carbides for 20% coldrolled sample compared to 10% cold-rolled one, lower hardness is
measured. More in depth analysis of carbides by classifying them
into three different categories based on their sizes (mean diameter
below 1 m, between 1 m and 5 m, and above 5 m) showed
that carbide growth occurs signicantly in the medium-size ones
after 20% cold rolling. Therefore, the volume fraction of large size
carbides increases at the expense of the medium sized ones. As the
main contribution to hardness is mostly due to medium-size carbides, therefore the observed reduction in hardness for the 20% cold
rolled sample can be related to the lower volume fraction of these
carbides (Fig. 3e).
The mean PAGS was measured using three images with the same
magnication from three randomly selected areas taken with the
optical microscope at 500 magnication (as shown in Fig. 1b, d
and f). The results showed continuous reduction from 9 0.3 m
(for non-deformed) to 8 0.1 m (for 10% cold-rolled) and then to
7 0.2 m (for 20% cold-rolled) (Fig. 3d), indicating that prior cold
deformation results in the reduction of the PAGS.

4. Discussion
4.1. The effect of grain renement on the maximum attainable
hardness
The mean PAGS inuences martensite hardness, the Ms temperature, and the extent of transformation [1316]. Inverse relation
between grain size and yield stress is well established by HallPetch
equation [21]. Also, direct relation between yield strength and
hardness has been reported by Busby et al. [22]. It is generally
accepted that smaller grain sizes result in higher hardness values. Fig. 4 shows the inuence of processing conditions on the

118

H.G. Nanesa et al. / Journal of Manufacturing Processes 22 (2016) 115119

Fig. 3. (a)(c) BSE images from the unetched microstructure of (a) non-deformed, (b) 10% cold rolled, and (c) 20% cold rolled samples showing carbides as black spots with
different sizes and matrix as brighter background, (d) Inuence of prior cold rolling on the PAGS, volume fraction of carbides, and hardness value (e) distribution of carbides
volume fraction-carbide size as a function of processing conditions: mean diameter below 1 m and above 0.2 m, between 1 m and 5 m, and above 5 m showing
signicant growth of medium-size carbides from 10% to 20% cold rolling.

mean PAGS and hardness. It can be seen that both hardness and
PAGS decrease as the amount of prior deformation increases. In the
present work, the microstructure of the investigated alloy contains
both martensite and second phase carbides. Therefore, carbides
in the AISI D2 steel, by their characteristics, inuence the hardness value considerably. The faster dissolution process of carbides,
observed after cold deformation, results in lower volume fraction
values after quenching as observed in Fig. 3ae. This loss in the
size, quantity, and volume fraction of hard carbides reduces the
maximum attainable hardness after quenching.

in several research studies [15,21,23,24]. Specically, Mason and

Prevy [24] reported that the presence of retained austenite in the
hardened microstructure decreases the hardness value [24]. In the
present work, the effect of the PAGS on the value of Ms is measured
using a general linear relationship between ASTM grain size number (G), and the Ms temperature introduced by Lee and Lee [15] as
follows:

4.2. Relationship between grain renement and the extent of

martensitic transformation

The effect of changes in Ms temperature on the extent of the

resulting martensitic transformation was studied using an experimental equation proposed by the present authors (derived from
dilatometry diagrams of rapidly quenched D2 tool steel the rates of
10 and 50 K s1 [25] as follows:

The effect of smaller mean PAGS on the reduction of the Ms temperature and the stability of retained austenite have been reported

Ms ( C) = 542.3 30 G

(1)

H.G. Nanesa et al. / Journal of Manufacturing Processes 22 (2016) 115119

61

Hardness (HRC)
Prior austenite grins size (PAGS-m)

90.3

9.5

Hardness (HRC)

59.90.2

8.5

59

80.1

8
58
7.5
57

70.2
56.80.6

56

Average PAGS (m)

60

6.5

119

accordingly the volume fraction of retained austenite as a result of

prior deformation (Fig. 2). The above ndings are also in agreement
with those reported by Beswick [5] who also found a maximum of
1% and 2% increment in the volume fraction of retained austenite after 30% and 64% cold deformation, respectively in SAE 52100
hypereutectoid steel with no signicant impact on hardness value.
The lower hardness levels after cold rolling and hardening, in
spite of the smaller PAGS, appear to be rather related to the lower
amount of carbides in comparison with the non-deformed sample
(Fig. 3) than to the amount of retained austenite.

55.30.4

55
Non-deformed and
hardened sample

10% cold-rolled and

hardened sample

5. Conclusions

20% cold-rolled and

hardened sample

Heat treatment condition

Fig. 4. Hardness and the mean PAGS of (a) non-deformed, (b) 10% cold-rolled, and
(c) 20% cold-rolled samples.

230

fm (volume fraction fo martensite-%)

100

Ms (C) temperature

226
99

225

98

220
215

97

214
96

96
95

210

200

95
94

205
Ms (C) temperature
fm (volume fraction of martensite-%)

(a) 10.55

(b) 10.93

94

200
93
(c)11.41

The inuence of pre-deformation on microstructure evolution

and hardness of AISI D2 tool steel was studied and compared to
non-deformed and hardened alloy. Increasing the amount of cold
rolling resulted in lower PAGS as well as reduction in the hardness
values in spite of grain renement. Reduction in the volume fraction
of carbides is proposed as the root cause for lower hardness values.
Acknowledgements
The authors would like to thank the Natural Sciences and Engineering Research Council of Canada (NSERC) grant number 461604
and MITACS Canada grant number IT04882 for their nancial supports. The authors also appreciate the collaboration of DK SPEC
Inc. for providing experimental materials, support from CanmetMATERIALS in the framework of RIEM program for dilatometry
experiments, Hitachi Canada for privileged access to advanced
microscopy facilities.

ASTM grain size (G)

References
Fig. 5. Relationship between the Ms , G (ASTM grain size), and fm (volume fraction
of martensite-%): (a) non-deformed, (b) 10% cold-rolled, and (c) 20% cold-rolled
samples.

E = E0 3.83 106 + m T 4.235 103 exp (b (Ms T ))

(2)
where E is the total dilatation during transformation, E0 is the
dilatation at the Ms temperature, T is cooling temperature, b is the
rate parameter (0.0128 at 10 K s1 and 0.0156 at 50 K s1 ), and m
is the thermal expansion coefcient of martensite. In the present
study, a quenching rate of 30 K s1 was obtained for the three conditions. As reported in [26], the critical cooling rate to achieve fully
martensitic transformation was 3 K s1 . In order to have an estimation of parameter b for cooling at 30 K s1 , the average value of
0.0142 [(0.0128 + 0.0156)/2] is considered. The volume fraction of
martensite (fm ) was therefore calculated by using b = 0.0142 in the
Koistinen and Marburger (KM) equation [27].
fm = 1 exp [b (Ms T )]

(3)

Fig. 5 illustrates the relationship between G, Ms [from Eq. (1)]

and fm [from Eq. (3)] with the nal quenching temperature equal
to 273 K for the three samples. The results show that the ner
is the grain size (i.e. larger G), the lower will be the Ms temperature. However, it is interesting to note that fm calculations
didnt show signicant reduction in the volume fraction of martensite. This nding is in agreement with the XRD diagrams which
didnt show major changes in the intensity of austenite peaks and

[1] ASM International. ASM handbook. Heat treating. USA: ASM International;
1991.
[2] Roberts G, Krauss G, Kennedy R. Tool steels. 5th ed. USA: ASM International ;
1998.
[3] Bombac D, Fazarinc M, Saha Podder A, Kugler G. JMEPEG 2013;22(3):7427.
[4] Vecko Pirtovsek T, Kugler G, Tercelj M. Mater Charact 2013;83:97108.
[5] Beswick J. Metall Trans A 1984;15:299306.
[6] Ghasemi-Nanesa H, Jahazi M. Mater Sci Eng A 2014;598:4139.
[7] Das D, Dutta AK, Ray KK. Philos Mag Lett 2008;88:80111.
[8] Das D, Dutta AK, Ray KK. Mater Sci Eng A 2010;527(9):218293.
[9] Singh SG, Singh J, Singh R, Singh H. J Adv Manuf Technol 2011;54:5982.
[10] Tokizane M, Matsumura N, Tsuzoki K, Maki T, Tamura I. Metall Trans A
1982;133:137988.
[11] Chojnowski EA, Tegart WJMcG. Met Sci J 1968;2:148.
[12] Joo HS, Hwang SK, Baek HM, Im YT, Son lH, Bae CM. J Mater Process Technol
2015;216:34856.
[13] Guimares JRC. Scr Mater 2007;57:2379.
[14] Guimares JRC, Gomes JC. Acta Metall 1978;26:15916.
[15] Lee S-J, Lee Y-K. Mater Sci Forum 2005;475479:316972.
[16] Yang H-S, Bhadeshia HKDH. Scr Mater 2009;60:4935.
[17] Goun M, Bouaziz O, Allain S, Zhu K, Takahashi M. Mater Lett 2012;671:1879.
[18] Nahamin Pardazan Asia, MIP (Metallographical Image Processing), July, 2014,
Iran, http://en.metsofts.ir/.
[19] da Silva Farina PF, Barbosa CA, Goldenste H. ASTM Spec Tech Publ
2012;1532:5770.
[20] Das D, Dutta AK, Ray KK. Philos Mag 2009;891:5576.
[21] Bacon D, Dieter GE, editors. Mechanical metallurgy: SI metric. London:
McGraw-Hill Book Company; 1988.
[22] Busby JT, Hash MC, Was GS. J Nucl Mater 2005;336(23):26778.
[23] Huang J, Xu Z. Mater Sci Eng A 2006;438440:2547.
[24] Mason PW, Prevy PS. JMEPEG 2001;10:1421.
[25] Ghasemi-Nanesa H, Jahazi M, Naraghi R. J Mater Sci 2015;50(17):575868.
[26] Metal Ravne, May 2015 Slovenia http://www.metalravne.com/steelselector/
steels/OCR12VM.html.
[27] Koistinen DP, Marburger RE. Acta Metall 1959;7:5960.