Available online at [Link].

com

Scripta Materialia 58 (2008) 445–448

[Link]/locate/scriptamat

Grain size effect on wear resistance of a nanostructured

AISI52100 steel
L. Zhou, G. Liu, Z. Han and K. Lu*
Shenyang National Laboratory for Materials Science, Institute of Metal Research, Chinese Academy of Sciences,
Shenyang 110016, China
Received 31 July 2007; revised 22 October 2007; accepted 22 October 2007
Available online 26 November 2007

We report that the sliding wear resistance of a nanostructured spheroidal pearlite steel varies with grain size: it increases when the
grain size of the ferrite phase (Da) increases from 8 to 32 nm, and decreases with a further increase in Da. A maximum wear resis-
tance at an average grain size of 32 nm, corresponding to the most significant material transfer on the mating ball, was analyzed in
terms of the combination of hardness and plasticity.
 2007 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

Keywords: Nanostructured materials; Spheroidal pearlite steel; Grain size; Wear resistance

Extensive investigations on nanostructured metals such as CG metals and alloys with comparable ductility,
have indicated that reducing grain sizes to the nanome- wear volume is reversely proportional to the hardness of
ter regime leads to substantial hardening. The strength the material. However, for brittle materials, K becomes
and hardness increase monotonically with decreasing very large due to enhanced material removal via brittle
grain size following the empirical Hall–Petch relation- fracture during wear, leading to a reduced wear resis-
ship even for grains as small as 10 nm for Cu [1]. Mean- tance. Consequently, a high wear resistance is always
while, metals become rather brittle when grains are obtained in materials with a high strength together with
refined into the nanometer regime [2–4]. Most nano- a considerable ductility. As strength and ductility vary
structured metals fracture at an elongation typically of monotonically with a reduction in grain size, one may
a few per cent in tensile tests, even for those very ductile anticipate a maximum wear resistance at a critical grain
in coarse-grained (CG) form such as pure Cu and Al. size where the strength–ductility combination is opti-
This much reduced ductility limits the technological mized. In other words, a strongest grain size for wear
applications of nanostructured metals. resistance is expected for a polycrystalline material when
Wear resistance is one of the most important proper- its grains are refined from the macro- to the nanoscale.
ties for material applications, especially for those involv- To identify the strongest grain size for wear resis-
ing contact deformation and damage. Wear resistance of tance, we carried out an experimental investigation on
a material is determined by both strength and ductility, grain size effect on wear resistance in a nanostructured
as described by the modified empirical Archard equation steel.
[5–7]: The sample used is an AISI52100 steel plate (sphero-
idizing annealed: 1.0% C, 1.5% Cr, 0.31% Mn, 0.24% Si,
P 0.08% Ni, 0.15% W, 0.03% Mo and balance Fe). The as-
W ¼K ; ð1Þ
H received microstructure consists of spheroidal cementite
where W is the wear rate with an applied pressure P, H particles of sizes ranging from 1 lm to 200 nm embed-
is the material’s hardness, and K is a pre-factor relative ded in a ferrite matrix, of which the average grain size
to the material’s ductility. For conventional materials is about 15 lm. Cementite particles are uniformly
distributed with a total volume fraction of about 16%.
A steel plate 100 · 100 · 8 mm3 was subjected to the sur-
* Corresponding author. Tel.: +86 24 2390 6826; fax: +86 24 2399 face mechanical attrition treatment (SMAT), the set-up
8660; e-mail: lu@[Link] and procedure of which can be found in Refs. [8,9].

1359-6462/$ - see front matter  2007 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/[Link].2007.10.034
446 L. Zhou et al. / Scripta Materialia 58 (2008) 445–448

The SMAT process was carried out in vacuum

(6 · 102 Pa) at ambient temperature for 60 min with a a
vibrating frequency of 50 Hz. Steel balls of 8 mm in
diameter, with exactly the same chemical composition
as the to-be-treated steel plate, were used to eliminate
contamination from the treated sample. Plastic defor-
mation in the surface layer at a high strain rate results
in a progressive refinement of coarse grains into the
100 nm
nanometer scale. The thickness of the nanostructure
layer (the average grain size is below 100 nm) is about
60 lm, and evidence of plastic deformation was ob-
served in the surface layer of about 250 lm thick [10]. b
An annealing treatment to the nanocrystalline surface
layer was conducted in the temperature range 723–
1023 K for 60 min in a vacuum of 1 · 106 Pa. The
surface microstructure features of the samples were
characterized using a JEM 2010 transmission electron
microscope (TEM) operated at an accelerating voltage
of 200 kV. Thin foils for TEM observations were cut 100 nm
from the treated surface layer, polished and ion milled
at low temperatures.
An Optimol SRV III tester in a ball-on-plate contact c
configuration was used to evaluate dry sliding wear
resistance of the steel samples under ambient conditions.
Sliding wear tests were carried out against WC–Co balls
(10 mm in diameter, with a hardness of Hv  1750) un-
der normal loads of 50 and 70 N, respectively, with an
oscillating stroke of 2 mm, a frequency of 5 Hz and a
sliding duration of 30 min. The profiles of the worn 500 nm
surfaces were measured using a surface profilometer to
determine the wear volume loss V = AL (where A is
the worn area determined by its profiles, and L is the
the oscillating stroke). The worn morphologies were 800 d
characterized using a Leica DMRX optical microscope
600
(OM) and a LEO SUPRA 35 scanning electron micro-
scope (SEM).
D (nm)

400
Microhardness of the top surface layer was measured
on an MVK-H3 hardness tester equipped with a Vickers 200 Ferrite
indenter. The applied loads are in a range from 25 to Cementite

300 g with durations of 5 and 10 s. 0

TEM observations showed that the as-SMAT sample
exhibits a homogeneous nanostructure with an average 200 400 600 800 1000
size of about 8 nm for the ferrite phase and no cementite Ta (K)
phase could be detected from the electron diffraction
(Fig. 1a). The linear intercept method was used to quan- Figure 1. Bright-field TEM images for the as-SMAT sample (a) and
the SMAT samples annealed at 923 K (b) and 1023 K (c), respectively.
tify the grain sizes from a large number of TEM images.
The corresponding selected electron diffraction patterns are shown in
As discussed in our previous work [10], a substantial insets of (a) and (b). (d) Variations of the average grain sizes for ferrite
refinement (or fragmentation) and dissolution of and cementite with annealing temperature.
cementite were induced by the SMAT process, so that
the size and the amount of cementite in the as-SMAT
steel sample are too small to be detected. Upon anneal- Nanosized cementite particles are precipitated in the
ing, microstructures are changed with two distinct ferrite matrix upon annealing at temperatures lower
features: precipitation of cementite phase and grain than 973 K. The size of cementite precipitates (Dh) is
coarsening of both phases (ferrite and cementite). The close to that of the ferrite matrix (Da). Dh seems to in-
grain coarsening of ferrite is less pronounced with a crease slightly with annealing temperature, analogous
slight increment in grain sizes when the annealing tem- to that of the ferrite phase. When Ta > 973 K, Da in-
perature (Ta) is below 973 K (see Fig. 1b). The average creases more rapidly with Ta than Dh, and cementite
grain sizes increase to about 13 nm at 723 K and particles are obviously smaller than ferrite grains, as
32 nm at 923 K. When Ta > 973 K, significant grain seen in Figure 1d. Annealing at 1023 K yields a micro-
coarsening takes place. The average grain size reaches structure consisting of coarse ferrite grains with cement-
about 577 nm when Ta = 1023 K. Figure 1c shows an ite particles precipitated along the ferrite grain boundary
obvious recrystallization structure with rather coarse (GB) or occasionally in the ferrite grain interiors
ferrite grains. (Fig. 1c). The high thermal stability of the nanostruc-
 L. Zhou et al. / Scripta Materialia 58 (2008) 445–448 447

tured steel sample might be attributed to the GB segre- 0.8

Friction cofficient
gation of carbon atoms as well as pinning effect of nano- 200 0.7
CG
sized cementite particles at the GB of the ferrite phase 0.6
160
[11,12]. 0.5

W-1 (mm-3)
Figure 2a-1 shows the variation of the micro-hard- 120 0 20 40 60 500 1000
Dα (nm)
ness (H) with Da for the SMAT specimen annealed at
different temperatures. It is found that H decreases mod- 80
CG
erately from 11.4 to about 7.8 GPa as Da increases up to
40 50 N
57 nm. In accordance with the intensive grain growth, H 70 N

drops sharply to about 4.9 GPa when Da = 577 nm. Mi- 0

cro-hardness indentation analysis was also employed to 0 20 40 60 500 1000
determine the critical load (P*) for cracking around the Dα (nm)
indent, which can be qualitatively used to evaluate the
plasticity of different specimens. For the sample with Figure 3. Variations of reciprocal wear volume (W1) with the ferrite
grain size for the nanostructured steel samples under loads of 50 and
an average size of 32 nm, cracks are formed around
70 N. W1 values of the CG steel sample are included for comparison.
the indent with a normal load of 200 g, while no crack Insert shows variation of the measured steady-state friction coefficient
is found at 100 g (Fig. 2b-2 and b-3). This means that under a load of 70 N with ferrite grain size.
the critical load (P*) for crack formation falls in between
100 and 200 g. Similarly, P* values for samples with var-
ious grain sizes were determined, as depicted in Figure tity representing the wear resistance, of the as-SMAT
2b-1, which represents the relative plasticity of the in- steel sample is comparable to that of the CG sample
dented surface layers with different ferrite grain sizes. at loads of 50 and 70 N, respectively. It seems that no
Clearly the plasticity increases gradually with an increas- obvious improvement in the wear resistance is achieved
ing Da. for the nanostructured sample fabricated by means of
The friction and wear characteristics of the nano- the SMAT process relative to the CG sample.
structured samples were summarized in comparison However, it is interesting to note that the wear resis-
with those of the CG sample (Fig. 3). The steady-state tance is markedly enhanced after thermal annealing
friction coefficient of the as-SMAT sample remains un- which induces grain coarsening (Fig. 3). The W1 value
changed under a load of 70 N after annealing at different increases with an increasing Da, reaching a maximum at
temperatures, which is slightly higher than that of the about Da = 32 nm. When Da > 32 nm, the W1 value
original CG sample. It means the friction coefficient of decreases at larger ferrite grain size. The same variation
the nanostructured steel samples is independent of grain tendency of W1 can be obtained for two different loads
size. The reciprocal of wear volume (W1), i.e., a quan- (50 and 70 N).
The variation of the wear scar morphologies (Fig. 4)
coincides with the results of wear volume measurements.
a-1 12
For the as-SMAT sample the worn surface is rather
a-2
12 10 rough (Fig. 4a-1). After annealing, the worn surface be-
H(GPa)

8 comes smooth with a decreased roughness for Da < 32

10 6 nm (Fig. 4b-1). With a further increment of Da above
4
32 nm, the surface becomes rough again (Fig. 4c-1).
H (GPa)

0.0 0.1 0.2 0.3

8 -1/2
Dα (nm
-1/2
) Clearly, the sample with Da = 32 nm exhibits the
smoothest wear scar (Fig. 4b-1), corresponding to the
6 maximum W1 value in Figure 3.

4
b-1
No crack
300 Cracks
Cracks/no crack
250

200
P* (g)

b-2
150

100
b-3

50
10 μm

0 20 40 60 560 580
Dα (nm)

Figure 2. Variations of hardness (a-1) and the critical load (P*) for
cracking around the indent (b-1) with the ferrite grain size. (a-2) Hall–
Petch plot of the measured hardness vs. Da1=2 . The morphologies of Figure 4. Worn morphologies of the specimen and the mating WC–Co
Vickers hardness indents under loads of 200 and 100 g for the annealed balls for the as-SMAT sample (a-1, a-2), the samples with Da = 32 nm
SMAT sample with Da = 32 nm are shown in (b-2) and (b-3), (b-1, b-2) and with Da = 577 nm (c-1, c-2) under a load of 70 N. (The
respectively. arrows indicate the sliding direction.)
448 L. Zhou et al. / Scripta Materialia 58 (2008) 445–448

The surfaces of the mating WC–Co balls sliding in the electron diffraction patterns, we did not notice
against steel samples were examined by means of OM an obvious change in the relative amount of cementite
observations. Consistent with the wear scar morphol- with annealing temperature. This is also verified by the
ogy, for the sample with Da = 32 nm, the amount of Hall–Petch plot of the measured hardness vs. D1=2a , as
material transfer on the WC–Co ball (of dark contrast shown in the inset of Figure 2a-1. The variation of hard-
in Figure 4b-2, which is mainly iron oxide particles as ness follows nicely the Hall–Petch line for the annealed
indicated by energy-dispersive spectroscopy results) is SMAT samples, indicating that the amount of cementite
the largest of all the samples. The amount of material in the annealed samples at different annealing tempera-
transfer on the WC–Co ball decreases as Da decreases tures remains approximately unchanged. Therefore, the
below 32 nm (Fig. 4a-2) or as Da increases above observed variation in wear resistance in the present steel
32 nm (Fig. 4c-2). This observation implies that the samples can mainly be attributed to the grain size effect,
transferred material on the mating ball serves as a dis- while the cementite quantity effect can be neglected.
continuous protecting layer to reduce direct contact be- In summary, a nanostructured surface layer on a
tween the WC–Co asperities and the sample, thereby spheroidal pearlite steel plate was synthesized by
diminishing the material removal rate of the steel sam- means of SMAT, and exhibits a wear resistance com-
ples during wear [13,14]. Therefore, the minimum wear parable to that of the original CG steel. The grain size
volume at Da = 32 nm corresponds well with the maxi- of the nanostructured steel sample is varied by anneal-
mum amount of material transfer. ing at different temperatures, and ranges from 8 to
As seen in Figure 2, hardness and plasticity of the 577 nm. The measured wear resistance of the samples
nanostructured samples vary monotonically with grain with different grain sizes shows a maximum wear resis-
size, while the sliding wear resistance does not (Fig. 3). tance at Da = 32 nm. Grain sizes smaller and larger
The maximum wear resistance does not correspond to than this size may lead to an obvious decrease in wear
the highest hardness (the as-SMAT sample), or to the resistance. The highest wear resistance in the sample
highest plasticity (Da = 577 nm). Instead, the sample of with Da = 32 nm, which corresponds to the most sig-
Da = 32 nm with a medium hardness and a medium nificant material transfer on the mating ball, originates
plasticity exhibits the highest wear resistance. This from the optimized combination of high hardness and
clearly indicates that wear resistance is determined by moderate plasticity.
a combination of hardness and plasticity, as related in
Eq. (1). This work was financially supported by NSF of
The sample with the highest hardness, i.e., the as- China (Grants No. 50621091, 50431010, 50325101,
SMAT sample with the smallest ferrite grain size, has 50571095) and the Ministry of Science & Technology
the lowest plasticity, which greatly facilitates material of China (2005CB623604). The authors are grateful to
removal via brittle fracture during sliding wear. There- Y.S. Zhang for stimulating discussions and to H.W.
fore, the wear resistance is not high. In fact, the wear Song and J.B. Zhang for providing the material.
resistance of the as-SMAT sample is very close to that
of the untreated CG sample, of which the hardness is
much smaller than that for the as-SMAT sample. An- [1] J. Chen, L. Lu, K. Lu, Scripta Mater. 54 (2006) 1913.
other extreme case is the SMAT sample annealed at [2] C.C. Koch, D.G. Morris, K. Lu, A. Inoue, MRS Bull. 24
1023 K with the largest ferrite grain size, which pos- (2) (1999) 54.
sesses the lowest hardness and the highest plasticity of [3] M.A. Meyers, A. Mishra, D.J. Benson, Prog. Mater. Sci.
all the SMAT samples. Its low wear resistance might 51 (2006) 427.
be attributed to the softening effect. Obviously, the sam- [4] X.H. Chen, J. Lu, L. Lu, K. Lu, Scripta Mater. 52 (2005)
1039.
ple with Da = 32 nm possesses the highest wear resis-
[5] J.F. Archard, J. Appl. Phys. 24 (8) (1953) 981.
tance of all the samples, implying an optimized [6] E. Hornbogen, Wear 33 (1975) 251.
combination of a high hardness with a moderate plastic- [7] I. Sevim, I. Eryurek, Mater. Des. 27 (2006) 911.
ity which is efficient in impeding formation and propa- [8] K. Lu, J. Lu, J. Mater. Sci. Technol. 15 (1999) 193.
gation of cracks during sliding wear. [9] N.R. Tao, Z.B. Wang, W.P. Tong, M.L. Sui, J. Lu, K.
The effect of amount of cementite in the samples with Lu, Acta Mater. 50 (2002) 4603.
different ferrite grain sizes on the wear resistance should [10] L. Zhou, G. Liu, X.L. Ma, K. Lu, Acta Mater. (2007),
be taken into account. However, due to the small sizes doi:10.1016/[Link].2007.09.003.
and/or the small volume fraction of cementite phase in [11] J.W. Cahn, Acta Metall. 10 (1962) 789.
the as-SMAT and the SMAT-annealed samples, accu- [12] S.C. Mehta, D.A. Smith, U. Erb, Mater. Sci. Eng. A 204
(1995) 227.
rate determination of the volume fraction of cementite
[13] Y.S. Zhang, Z. Han, Tribol. Lett. 27 (2007) 53.
and its variation with grain size is difficult experimen- [14] K. Kato, Wear 241 (2000) 151.
tally. From a rough estimation of the relative intensity