Full text of DOI:10.1557/JMR.2002.0147

Microstructures and mechanical behavior of bulk
nanocrystalline ␥–Ni–Fe produced by a
mechanochemical method
X.Y. Qina)
Laboratory of Internal Friction and defects in Solids, Institute of Solid State Physics Academia
Sinica, 230031 Hefei, People’s Republic of China
J.S. Lee and C.S. Lee
Department of Metallurgy and Material Science, Hanyang University, 425-791 Ansan, Korea
(Received 5 January 2001; accepted 5 February 2002)
The microstructures and mechanical behavior of bulk nanocrystalline ␥–Ni–xFe
(n-Ni–Fe) with x ס 19–21 wt%, synthesized by a mechanochemical method plus
hot-isostatic pressing, were investigated using microstructural analysis [x-ray
diffraction, energy-dispersive spectroscopy, light emission spectrum, atomic force
microscopy (AFM), and optical microscopy (OM)], and mechanical (indentation and
compression) tests, respectively. The results indicated that the yield strength (␴0.2) of
n-Ni–Fe (d 33 nm) is about 13 times greater than that of conventional counterpart.
The change of yield strength with grain size was basically in agreement with
Hall–Petch relation in the size range (33–100 nm) investigated. OM observations
demonstrated the existence of two sets of macroscopic bandlike deformation traces
mostly orienting at 45–55° to the compression axis, while AFM observations revealed
that these bandlike traces consist of ultrafine lines. The cause for high strength and the
possible deformation mechanisms were discussed based on the characteristics of
microstructures and deformation morphology of n-Ni–Fe.
that the main cause of conflicting results or uncertainty
I. INTRODUCTION
(Refs. 5, 13, 17, and 18) in experimental phenomenon
lies in difficulties in manufacturing large, dense, and
(geometrically) standard specimens for conventional me-
chanical tests. Hence, to extract intrinsic mechanical
behavior and acquire deformation mechanisms of n-
materials, it is significant for one to prepare fully dense
n-material and to perform mechanical tests with conven-
tional test methods on standard specimens.
Nanocrystalline materials (n-material)—i.e., polycrys-
talline materials with grain size smaller than 100 nm—
are of considerable scientific and technological interest at
present.1,2 Since the early days of the research in this
field, mechanical properties of n-materials have aroused
much attention due to their extremely fine grain sizes and
large volume fraction (5–50%) of interfaces, which may
lead to different and/or improved mechanical properties,
such as high strength in n-metals3 and high ductility in
n-ceramics.4 In the meantime, some contradictory results
in the mechanical behavior have been reported. For in-
stance, some investigators reported that normal Hall–
Petch (H-P) relation existed5–9; while others reported that
inverse H-P relations (hardness decreased with decreas-
ing grain size) were found in some n-materials.10–12 Even
hardening and softening behavior on decreasing grain
size were found to coexist in some n-materials.13–15 In
recent work, the relationship between Vickers hardness
and yield stress in n-Cu was found to be different from
that in conventional polycrystalline materials.16 In view
of most earlier work (for example, Ref. 15), we believe
In this paper, we report our investigations on micro-
structures and mechanical behavior of bulk nanocrystal-
line ␥–Ni–Fe (n-Ni–Fe) with x
19 to 21 wt%,
synthesized by a mechanochemical process plus hot iso-
static pressing (HIPing). ␥–Ni–xFe with x ס 10 to
65 wt%, so called Permalloy, is an important material
with wide applications in industry.19 Many techniques
have been used to prepare nanocrystalline Ni–xFe,
mostly in powder state.14,20–24 However, microstructures
and mechanical properties of bulk n-Ni–Fe with nearly
full density were rarely reported.14 The synthetic method
used here has been shown to be an effective method that
has the advantage of producing nano-powders with large
quantity and light agglomeration.24 HIPing is one effec-
tive route for consolidating powder into bulk material in
relatively low temperature and short time and thus is
beneficial to retaining nanophase microstructures.25
Microstructures and deformation features were characterized
a)Address all correspondence to this author.
e-mail: xyqine@mail.issp.ac.cn
J. Mater. Res., Vol. 17, No. 5, May 2002
© 2002 Materials Research Society
991
http://journals.cambridge.org
Downloaded: 18 Mar 2015
IP address: 128.255.6.125

X.Y. Qin et al.: Microstructures and mechanical behavior of bulk nanocrystalline ␥–Ni–Fe produced by a mechanochemical method
C. Processing of specimens for mechanical
testing and surface observation
using various methods including x-ray diffraction
(XRD), dispersive x-ray spectroscopy (EDS), light emis-
sion spectrum (ES), atomic force microscopy (AFM),
and optical microscopy (OM). Both conventional Vick-
ers hardness and compression tests were used to charac-
terize mechanical behavior.
To carry out compression tests, strip shaped specimens
were cut from HIPed bars (K series) in the transverse
direction with a spark erosion machine. To avoid the
effect of surface damage by the spark erosion and to
reduce surface flaws, after cutting, a layer of more
than approximately 0.1 mm in thickness was polished
away from all sides of the as-machined specimens
first with grinding papers (the final was No. 800# paper)
and then with diamond pastes (the final had a particle
size of 1 ␮m). Unless otherwise mentioned, specimen
length is usually 3 times larger than the largest the
cross dimension (typical dimensions were approximately
6.2 × 2 × 2 mm3 after final polishing) to maintain a free
deformation section (to avoid the frictional constraints
from both ends) in the specimen tested. To observe de-
fects and deformation traces on specimens with OM and
AFM, one lateral surface was further specially polished
using diamond paste of particle size ¼ ␮m and then
etched with Nital (3%HNO3 + 97% alcohol) for 15 s.
There was no surface processing after deformation.
II. EXPERIMENTAL
A. Synthesis of n-Ni–Fe powder
The powders of n-Ni–Fe alloy were synthesized by a
mechanochemical process.26,28 Experimental details
have been given elsewhere.26 Briefly, the dry mixed ox-
ide powder was obtained by blending ␣–Fe2O3 (30 ␮m,
99.99%) and NiO (4 ␮m, 99.99%). Then these dry mixed
oxides were ball milled in methyl alcohol at a speed of
300 rpm for 10 h. The ball media and impeller used were
made of stainless steel. After ball milling they were dried
and sieved. The milled oxides were then reduced at
500 °C for 1 h and then at 550 °C for 0.5 h in hydrogen
to create Ni–Fe powder.
B. Preparation of the bulk material
D. Microstructural characterization
The obtained Ni–Fe powder was then compacted un-
der uniaxial pressure of approximately 900 MPa into
rectangular bars with dimensions of 24 × 8 × approxi-
mately 3 mm3 (labeled K) and 24 × 8 × approximately
1.5 mm3 (labeled T). They had a green (fraction) density
of 60–65%. These raw bars were then presintered at
650 °C for 1.5 h in H2 to remove any oxide present.
High-density bars were prepared by HIPing in Bodycote
IMT GmbH, Essen, Germany. To do this, the presintered
bars were embedded in high-purity Al2O3 powder sealed
in an evacuated stainless can (see Fig. 1) and then
pressed at 750 °C for 1 h under the pressure of 190 MPa
in Ar atmosphere.
The crystalline structure of the bulk bars was checked
using XRD. Also, the mean grain size was deter-
mined with XRD based on the Scherrer formula. Well-
annealed Si powder was used for the calibration of in-
strumental broadening. The bar densities were measured
based on the Archimedes principle with an accuracy of
±0.2%. A piece of polycrystalline Ni (with purity of 99.99%)
annealed at 800 °C for 18 h was used for checking the
reliability of density measurements.
Main constituents (Ni, Fe) and their line distributions
and spot analysis (including Cr) were determined with
EDS. Impurities were analyzed with ES. AFM was used
for the evaluation of grain size and its distribution on the
specially polished surfaces. In addition, OM and AFM
were used to observe the specially polished specimen
surfaces before and after deformation (or indentation) to
reveal any defects and deformation traces.
E. Mechanical tests
Compression tests were conducted at room tempera-
ture using an Instron type testing machine (Model:
Instron 1195) operating at constant rate of anvil displace-
ment. To avoid indenting anvil heads and increase the
accuracy in reading yield strength, two well-machined
hard-metal plates were placed between the specimen
ends and the anvil heads. MoS2 powder was sprayed in
the plates for lubrication. The initial strain rate was ap-
proximately 1.4 × 10−4 s−1. To investigate the effect of
grain size of the material on mechanical behavior (yield
FIG. 1. Schematic drawing showing the placement of the bars in the
sealed stainless steel can to be used for hot-isostatic pressing.
992
J. Mater. Res., Vol. 17, No. 5, May 2002
Downloaded: 18 Mar 2015
http://journals.cambridge.org
IP address: 128.255.6.125

X.Y. Qin et al.: Microstructures and mechanical behavior of bulk nanocrystalline ␥–Ni–Fe produced by a mechanochemical method
TABLE II. Mean grain size d, density ␳, fraction density, Dr, Vickers
strength and hardness) a series of specimens cut from the
hardness Hv, and yield stress ␴
for several n-Ni–Fe bars
0.2
same bulk bar were isothermally annealed at 850 °C for
different time in sealed quartz tubes filled with argon.
Vickers hardness tests were performed with a conven-
tional hardness tester on the specially polished surfaces
at room temperature. The applied loads ranged from 3 to
30 kg. An indentation time of 20 s was used in all cases.
after HIPing.
Bar
d (nm)
␳ (g/cm3)
Dr (%)
Hv (GPa)
␴
(GPa)
0.2
K1
K2
K3
K5
K6
T1
33
36
34
34
35
38
8.34
8.49
8.49
8.31
8.31
8.33
98.1
99.2
99.2
97.8
97.8
98.0
5.0 ± 0.1
4.2 ± 0.1
4.2 ± 0.1
4.8 ± 0.1
4.9 ± 0.1
4.6 ± 0.2
1.96 ± 0.06
1.59 ± 0.03
1.64 ± 0.03
1.85 ± 0.05
1.94 ± 0.05
1.81 ± 0.05a
III. RESULTS
aDatum obtained by tensile test (details are given in Ref. 27).
A. Microstructures and defects
Before HIPing, the crystal structure of all the presin-
tered bars was determined by XRD. The results indicated
that a complete gamma phase has already formed similar
to the early reports24 for n-Ni–Fe alloys prepared with
the same method (in similar condition) used here. With the
line profile analysis in XRD, mean grain size was found
to be around 25 nm. Fraction densities Dr were deter-
mined to be in the range of 65–70% (assuming the theo-
retical density is 8.6 g/cm3)21.
Table I lists the analytical results of chemical elements
for several bars after HIPing. The Fe content was in the
range of 19–21 wt%, changing slightly from bar to bar.
Apart from the main components (Ni, Fe), there were
some impurities, i.e., Cr, Mn, Al, and Si. The Cr content
was around 1 wt% (except for K2 and K3 bars), and
contents for the other impurity elements were ഛ0.1 wt%.
From Table I, one can see that the contents of Fe, Cr, and
Mn in bar K2 and K3 were obviously lower than in the
other bars. The Al content was nearly same for all
the bars. It is worthwhile to note that Si content obtained
here with ES had relatively large error since the samples
for ES analysis were easily contaminated by dust (silica)
from the air when ES measurements were taken, and
it is inappropriate to compare the obtained Si content
quantitatively.
Figure 2 gives a typical XRD pattern for n-Ni–Fe after
HIPing, where the XRD pattern for pure Ni is also given
for the sake of comparison. It can be seen that in addition
to the face-centered-cubic (fcc) structure, no other phase
was detected. The mean grain size was found from XRD
to be around 30–40 nm; the concrete value changed
slightly from bar to bar presumably due to different lo-
cations within the can used in HIPing experiments.
Table II lists some detailed results of grain size for
each bar.
Grain size and morphology were investigated using
AFM. As an example, Fig. 3 shows an AFM image of
grain morphology for a specimen cut from the bar labeled
T1. It can be a seen from this figure that the grains are
fairly homogeneous and largely display spherical shape.
Figure 4 gives the histogram of the size distribution ob-
tained by counting the grains according to their sizes. It
shows that most of the grains have the size of 30–40 nm,
which is in good agreement with XRD results (see
Table II).
Figure 5 shows the surface OM micrograph of a speci-
men cut from bar K5. As expected there was some po-
rosity (labeled A in the graph) in the specimens,
consistent with the results of density measurements.
Moreover, besides normal porosity some microcracklike
defects were occasionally found on the surface as indi-
cated by B in the Fig. 5. However, these microcracklike
defects were usually very shallow because they usually
disappeared after further polishing. Hence, they can be
ascribed to unfinished bonding of the particles (agglom-
erates of the grains) in HIPing. In addition, from Fig. 5
one can observe some bright regions on the surface,
whose typical size is around 50–100 ␮m. They are invis-
ible under OM if they are not etched. Similar patterns
existed on the other lateral surfaces, and their main fea-
ture (shown in Fig. 5) did not change with different pol-
ishing depth, indicating its bulk and particulate character.
It is worth noting that grain morphology (size, distribu-
tion, or shape) in these bright regions, observed using
AFM, has no detectable difference from that in dark re-
gions. That is, they consist of nano-grains features simi-
lar to the one in Fig. 3. However, these bright regions
The density for the bulk bars was greater than 8.3 g/cm3,
as shown in Table II. Since the materials contained some
lighter impurity elements (Al, Si, and Cr, as compared to
Ni), the theoretical density ␳0, neglecting the possible
change in lattice parameter, was estimated to be 8.50–
8.56 g/cm. Using this value for ␳0 one can obtain fraction
density of the obtained bulk material to be around 98–
99%, as listed in Table II.
TABLE I. Composition of n-Ni–Fe bulk material after HIPing.
Bar
K1
K2
K3
K5
K6
T1
Ni (wt%)
Fe (wt%)
Cr (wt%)
Mn (wt%)
Al (wt%)
Si (wt%)
78.4
20.6
0.8
0.08
0.03
0.1
80.3
19.3
0.1
0.05
0.03
0.05
79.6
19.7
0.3
0.05
0.05
0.1
78.2
20.6
1
0.08
0.03
0.05
77.6
21.1
1
0.1
0.05
0.05
78.8
20.2
0.8
0.08
0.03
0.05
J. Mater. Res., Vol. 17, No. 5, May 2002
Downloaded: 18 Mar 2015
993
http://journals.cambridge.org
IP address: 128.255.6.125

X.Y. Qin et al.: Microstructures and mechanical behavior of bulk nanocrystalline ␥–Ni–Fe produced by a mechanochemical method
were found to exhibit better resistance to chemical ero-
sion during etching. Since the specimens had only single
phase, these regions cannot be explained as second phase.
One possible explanation is that the bright regions could
possess higher density than the adjacent area. If this is
true, then these denser regions would presumably be
related to the hard agglomerates in the powder before
compaction.24 However, experimentally there were dif-
ficulties in determining local density changes at such a
small scale to justify this explanation. Another explana-
tion is that the regions with bright contrast would have
different chemical compositions from that in the sur-
rounding area. The line distributions of three main com-
ponents for bar K5, for instance, were analyzed with EDS
in an electron microprobe, and shown in Fig. 6. It can be
seen that the concentrations of the two main elements
Ni and Fe, as well as Cr, fluctuate with distance and
have a semi-periodicity of approximately 100 ␮m in
their changes. Qualitatively, this concentration change
with distance was consistent with the characteristics of
dark-bright contrasts (Fig. 5; it should be noted that the
area for the line analysis was not just the same area
viewed in Fig. 5). In addition, spot analysis with EDS in
SEM revealed that bright regions had slightly higher Cr
(also higher Fe) content than the dark regions. This
analysis seemed to support the assumption that the dif-
ferent contrast under OM would be related to chemical
inhomogeneity.
FIG. 2. XRD patterns (Cu K␣ radiation) for (a) a nickel standard and (b) n-Ni–Fe.
FIG. 3. AFM images for the surface of a specimen cut from bar T1: (a) height image, (b) phase image [it shows the same viewing field as in (a)].
994
J. Mater. Res., Vol. 17, No. 5, May 2002
http://journals.cambridge.org
Downloaded: 18 Mar 2015
IP address: 128.255.6.125

X.Y. Qin et al.: Microstructures and mechanical behavior of bulk nanocrystalline ␥–Ni–Fe produced by a mechanochemical method
B. Mechanical properties
1. Indentation results
Indentation experiments showed that Vickers hardness
for the HIPed bulk material ranged from 4.2 to 5.0 GPa,
which changed slightly from bar to bar (as given in
Table II). The difference in hardness values for different
specimens cut from a same bar is usually smaller than
3%. The influences of both measurement number and the
applied load on the hardness values were examined. As
an example, Fig. 7 gives the plot of the hardness for a
specimen cut from K3, obtained under the load of 20 kg,
as a function of measurement number. It can be seen that
under this load, the average value of its hardness was
4234 ± 158 MPa. The amplitude of the datum fluctuation
around its average value was usually smaller than 5%. In
addition, experiments showed that hardness did not
change obviously with the load as load P > 10 kg. When
FIG. 4. Distribution of grain size obtained from the AFM images
shown in Fig. 3.
FIG. 6. Line distribution for the three main components Ni, Fe, and
Cr. The other impurity elements were neglected in this analysis.
FIG. 5. Micrograph of the surface of a specimen cut from bar K5
observed using an optical microscope. A shows porosity; B shows a
microcracklike defect.
FIG. 7. Vickers hardness for a specimen cut from bar K3 versus
measurement number (applied load, 20 kg).
J. Mater. Res., Vol. 17, No. 5, May 2002
Downloaded: 18 Mar 2015
995
http://journals.cambridge.org
IP address: 128.255.6.125

X.Y. Qin et al.: Microstructures and mechanical behavior of bulk nanocrystalline ␥–Ni–Fe produced by a mechanochemical method
the load was smaller than 10 kg (3 kg, for instance), how-
ever, there was a clear tendency that the hardness de-
creased with decreasing load (see Fig. 8), suggesting that
too small load (here <10 kg) would cause large error in
determining the hardness.
(the peak stress) there was hardly strain hardening. This
result is similar to the early reports on n-Fe–28Al–2Cr.28
Besides, after the specimen deformed plastically to the
strain >15%, no fracture was found. The yield stress
(␴0.2) obtained from this nonstandard specimen [see
curves (b) and (c)] was 1.7 GPa, that is, about 6% higher
than that (1.6 GPa) from the standard specimen [see
curve (a)], indicating the influence of the constraining
effect of the anvil heads.
The effect of annealing on the mechanical property of
the n-Ni–Fe can be clearly seen in Fig. 9. Figure 11 gives
yield strength (␴0.2) and hardness as a function of an-
nealing time, which shows that yield strength or hardness
2. Compression results
Figure 9 gives engineering stress–strain curves for K1
series specimens in the as-prepared and annealed states.
It can be seen that the flow stress for the as-HIPed speci-
men decreased after plastic deformation of approxi-
mately 0.3%. For the annealed specimens, the strain was
slightly larger upon drop of the stress. This decrease in
flow stress was caused by bend of the specimens due to
their great aspect ratio (length/cross dimension). Yield
strength (␴0.2) for the as-HIPed specimen was obtained
from curve (a) to be 1.96 GPa. This value was used to
represent the yield stress of the bulk material from which
the specimens were cut. The corresponding values for
other bulk material (bars) are listed in Table II. It can
be seen that the yield strength ranged from 1.59 to
1.96 GPa. Comparing hardness values with the corre-
sponding yield strength, one could find that the yield
strength value is about 12% greater than the hardness
value divided by 3, indicating that basically a three-times
relationship between the hardness value and yield
strength held in the n-Ni–Fe. Usually Vickers hardness
HV is about three times greater than yield stress ␴
.
0.2
That is, there is relation HV ס 3␴
.
0.2
To observe the flow behavior in the large strain range,
nonstandard specimens with smaller aspect ratio had to
be used to avoid bending too soon. Figure 10 gives the
stress–strain curves for two K3 specimens, in which
curve (a) is the engineering stress–strain curve for a stan-
dard specimen, and curves (b) and (c) were stress–strain
curves for a nonstandard one (dimension 2 × 2 × 3 mm3).
It can be seen from curve (c) that after complete yielding
FIG. 9. Engineering stress–strain curves for K1 series specimens in
the as-HIPed state (a), and annealed at 850 °C for (b) 0.5 h, (c) 3 h, and
(d) 5 h.
FIG. 10. Stress–strain curves for two K3 specimens: (a) engineering
stress–strain curve for a standard specimen, (b) engineering stress–
strain curve for a nonstandard specimen, (c) true stress–strain curve for
the nonstandard specimen.
FIG. 8. Vickers hardness of a specimen cut from bar K3 as a function
of applied load in indentation.
996
J. Mater. Res., Vol. 17, No. 5, May 2002
Downloaded: 18 Mar 2015
http://journals.cambridge.org
IP address: 128.255.6.125

X.Y. Qin et al.: Microstructures and mechanical behavior of bulk nanocrystalline ␥–Ni–Fe produced by a mechanochemical method
decreased monotonously with increasing annealing time.
This decrease in strength can be reasonably ascribed to
growth of its grain size. Figure 12 shows the plot of the
yield stress versus d−1/2. For comparison, the hardness
values divided by three is also given in the figure. It
can be seen that basically a straight line between yield
stress and d−1/2 can be drawn. By best fit of the data to
load (30 kg) applied, no cracks were found near the in-
dentation corners where the indenter produced con-
siderable tensile stresses, as shown in Fig. 14. In this
case, the deformed area (pileup) extended to over
100 ␮m away from the indentation. In addition, the plas-
tically deformed area did not appear to be smooth; rather
its surface looked very rugged, implying deformation
localization.
After compression tests were stopped (usually with
nominal strain of 1–3%), OM was used to check the
deformation traces. Observations indicated that usually
relatively few deformed traces could be found in the
central parts of the specimens. In contrast, a large number
of deformation traces existed near both ends of the
compressed specimens. Figure 15 shows a typical
OM micrograph of surface morphology. It can be seen
that the deformation has the characteristic of high het-
erogeneity. However, the striking feature of the morphol-
ogy was the formation of two sets of macroscopic
bandlike traces (BLTs) that were symmetric to load
axis and mostly oriented at 45–55° to the compression axis.
H-P relation, friction stress ␴ ס 0.70 GPa and slope
0
k ס 7.0 GPa (nm1/2) were obtained. There was a similar
linear relation between hardness and d−1/2
.
C. Deformation morphology
Light OM observations on the specimen surface after
indentation indicated that there were some plastic areas
(pileups) around indentation traces similar to the case of
n-Fe with grain size d > 20 nm.29 Figure 13 shows
this observation for a specimen cut from K3. In all
the indentation tests no crack was observed around the
indentation traces (see Fig. 13). Even for the largest
FIG. 11. Plot of yield strength and the hardness (divided by 3) versus
time of isothermal annealing at 850 °C.
FIG. 12. Plot of yield strength and hardness (divided by 3) versus
d−1/2. Solid and open circles are experimental data, and the straight
lines (solid and dot) are the corresponding linear fits to the data.
FIG. 13. Optical micrograph of Vickers indentations: (a) indentation
load, 10 kg; (b) and (c) indentation load, 20 kg; (d) indentation load,
30 kg.
J. Mater. Res., Vol. 17, No. 5, May 2002
Downloaded: 18 Mar 2015
997
http://journals.cambridge.org
IP address: 128.255.6.125

X.Y. Qin et al.: Microstructures and mechanical behavior of bulk nanocrystalline ␥–Ni–Fe produced by a mechanochemical method
Figure 16 shows an AFM image for the same deforma-
tion traces. In this magnification (image size 40 × 40 ␮m),
the BLTs constituted parallelogramic networks on the
specimen surface. Figure 17 gives an AFM image with
higher magnification. In this image, grain contours are
faint, but they are resolvable. Specially, one can see
clearly a great number of sharp (with most of their trans-
verse dimensions being clearly smaller than grain sizes)
and long nanochannels or grooves, demonstrating ultra-
fine structures of BLTs (Figs. 15 and 16). Due to limi-
tations of the spatial resolution (approximately 15 nm) of
AFM, it is difficult to resolve if deformation happened
within the individual nano-grains (approximately
30 nm). Nevertheless, from close inspection of the image
it could be seen that most of the sharp channels seemed
to run between the grains, implying deformation took
place mainly in grain boundary regions.
strength are about 13 times greater than the literature
values (150 MPa) for the permalloy with similar Ni con-
tent,30 and even greater than that (approximately
1.4 GPa) of high-strength steels.31 This great yield
strength can be explained reasonably in terms of the re-
finement of the grains in the specimens because the
IV. DISCUSSIONS
A. Yield strength and hardness of n-Ni–Fe
In this study, very high values of yield strength and
hardness (␴0.2 ס 1.54 to 1.96 GPa and Hv ס 4.2 to 5.0 GPa)
were obtained for n-Ni–Fe alloy. These values of yield
FIG. 15. Optical micrograph of a lateral surface for a specimen cut
from bar K3 after compression. The direction of compression axis is
vertical.
FIG. 14. Magnified micrograph of Vickers indentations under a load
of 30 kg.
FIG. 16. AFM image of the compressed specimen surface (image
size: 40 × 40 ␮m). The direction of compression axis is vertical.
998
J. Mater. Res., Vol. 17, No. 5, May 2002
Downloaded: 18 Mar 2015
http://journals.cambridge.org
IP address: 128.255.6.125

X.Y. Qin et al.: Microstructures and mechanical behavior of bulk nanocrystalline ␥–Ni–Fe produced by a mechanochemical method
in densification among different bars would be smaller
than 1%. The effect of this small densification difference
on strength could be negligibly small as compared to
other factors (see following context).
(ii) Chemical composition: As compared to bar K2 and
K3 with lower yield strength, the contents of Fe, Cr,
and Mn in the other bars were clearly higher. The higher
contents of these alloying (impurity) elements would
cause the material to become harder through solid solu-
tion strengthening mechanism.
(iii) Local inhomogeneity: As shown in Fig. 5, there
was microstructural inhomogeneity in n-Ni–Fe. Al-
though present experiments did not justify if it came
from density fluctuation, chemical fluctuation (Fig. 6)
or from both, one thing seemed to be certain was that the
regions with bright contrast deformed less (Fig. 15) than
other areas during compression, indicating their greater
hardness. Their contribution to the total strength of the
material would be similar to harder phases within com-
FIG. 17. AFM image with high magnification (image size: 3 × 3 ␮m).
posites. The more its content, the higher the strength of
The viewing field is shown by the square in Fig. 16. The direction of
the material as a whole would be. Experimentally there
compression axis is vertical.
was difficulty in determining the amount and spatial or/
hardening effect of grain refinement has been demon-
strated clearly in Fig. 12. It is worthwhile to note that the
hardness of n-Ni–Fe obtained here is higher than early
results.15 In the n-Ni–Fe alloy synthesized by electro-
deposition, a hardness value of about 3.2 GPa can be
obtained if their results are linearly extrapolated to the
grain size of approximately 30 nm.15 The lower strength
for their n-Ni–Fe would originate from the existence of
the texture in the material15 because as their experiments
indicated, microtexture formed in specimens during elec-
trodeposition. The existence of this texture would lead to
softening of the material due to easy transfer of disloca-
tions from one grain to another with a similar orientation.
In contrast, no evidence of texture in our specimens was
revealed by XRD (see Fig. 2).
and size distributions of the harder aggregates in respec-
tive bars. However, if one assumes that the impurity
content, such as Cr, was related to (or reflected) the
amount of the harder aggregates, the bars with higher
alloying (impurity) contents would be harder. Then the
different strength (or hardness) among different bars
(Table II) can be qualitatively explained.
B. Deformation mechanism
As shown in Fig. 12, the stress increased with decreas-
ing grain size, basically consistent with a normal
H-P relation. In other words, this means that the strength-
ening effect due to grain size refinement functioned in
this grain size range (33–100 nm). Similar strengthening
effect by grain size was obtained recently in n-Cu32 and
n-Fe.29 According to dislocation theory33 and Taylor
theory,34 the critical normal stress ␴ for dislocation gen-
eration is ␴ ≈ m␶ ≈ 3 Gb/L (here ␶ is the critical shear
stress, G shear modulus, b the Burgers vector, L the
length of dislocation segment, and m ≈ 3.06 orientation
factor). In a n-material, the length of a dislocation should
be constrained by grain size. That is, there is a relation
L ഛ d (here d is grain size). As a qualitative evalua-
tion, substituting b ס 0.25 nm for Ni,35 G ס 78 GPa,30
and d ס 33 nm into formula (2), ␴ ס 1.81 GPa is ob-
tained. This value coincides well with the yield stress of
n-Ni–Fe in the as-HIPed state [see curve (a) in Fig. 9 and
Table II]. This evaluation suggests that multiplication of
dislocations with the length of the grain sizes could occur
in n-Ni–Fe. The existence of a large number of sharp,
long, and fairly straight nano-channels revealed with
AFM (Fig. 17) might be explained by traces of dis-
location motion, i.e., “slip lines.” However, since no
The difference in strength (or hardness) among differ-
ent bars should be related to the difference in their mi-
crostructures. This may include the following:
(i) Densification: The fraction density (99.2%) of bar
K2 and K3 was about 1% higher than that of the others,
while both yield stresses were obviously smaller than
those of the others. Normally the higher the densification
of n-material, the greater its strength would be.9 Hence,
the difference in strength could hardly be explained
based on density difference. One point that should be
noted is that there was some error in the calculations of
the fraction densities listed in Table II since the actual theo-
retic density will change with chemical compositions of
an alloy. Concretely speaking, the Ni content in bar K2
and K3 was obviously higher than that in the others.
Correspondingly, the theoretic density for K2 and K3
would be higher than that of others due to the greater
atomic weight of nickel. In other words, the actual difference
J. Mater. Res., Vol. 17, No. 5, May 2002
Downloaded: 18 Mar 2015
999
http://journals.cambridge.org
IP address: 128.255.6.125

X.Y. Qin et al.: Microstructures and mechanical behavior of bulk nanocrystalline ␥–Ni–Fe produced by a mechanochemical method
deformation within individual nano-grains was resolved
(perhaps due to the limitation of AFM resolution), the
dislocation activity, if any, could more likely be related
to grain boundary (GB) dislocations because most of
channels seemed to run through in between the grains
(Fig. 17). Very recently, the present authors found that
the strain-rate sensitivity exponent for n-Ni–Fe was
n ס 0.01 at room temperature,36 which excluded possi-
bility of normal GB sliding mechanism, for n is usually
near 0.5 as GB sliding prevails. Other investigations
showed that GB sliding occurred as homologous tem-
perature (T/Tm) > 0.36 on n-Ni37 and n-Mg.38 Hence, it
seems unlikely that GB sliding would occur here at room
temperature [approximately 0.18 Tm; here Tm is melting
point (approximately 1452 °C) of Ni–Fe]. Therefore, de-
formation mechanism in n-Ni–Fe at room temperature
could still be some kind of dislocation mode.
Prof. Q.P. Kong for their helpful discussions, and
Dr. L.F. Chi, Dr. S. Gao, and Mr. X.G. Zhu for their help
in AFM observations as well as Bodycote IMT GmbH,
Essen, Germany for HIPing experiments. This work was
financially supported partly by Academic Sinica through
the Hundred Person Program and the Korean Ministry of
Science and Technology through the 2001 National Re-
search Laboratory Program.
REFERENCES
1. H. Gleiter, Prog. Mater. Sci. 33, 223 (1989).
2. C. Suryanarayanan, Int. Met. Rev. 40, 41 (1995).
3. V.Y. Gertsman, R. Birringer, R.Z. Valiev, and H. Gleiter, Scripta
Metall. Mater. 30, 229 (1994).
4. J. Karch, R. Birringer, and H. Gleiter, Nature 330, 556 (1987).
5. G.W. Nieman, J.R. Weertman, and R.W. Siegel, Scripta Metall.
Mater. 23, 2013 (1989).
6. J.S.C. Jang and C.C. Koch, Scripta Metall. Mater. 24, 1599
(1990).
7. S.K. Ganapathi and D.A. Rigney, Scripta Metall. Mater. 24, 1675
(1990).
8. H.J. Ho¨fler and R.S. Averback, Scripta Metall. Mater. 24, 2401
(1990).
9. X.Y. Qin, X.J. Wu, and L.D. Zhang, Nanostruct. Mater. 5, 101
(1995).
10. A.H. Chokshi, A. Rosen, J. Karch, and H. Gleiter, Scripta Metall.
Mater. 23, 1679 (1989).
11. K. Lu, W.D. Wei, and J.T. Wang, Scripta Metall. Mater. 24, 2319
(1990).
12. T. Christman and M. Jain, Scripta Metall. Mater. 25, 767 (1991).
13. G.E. Fougere, J.R. Weertman, R.W. Siegel, and S. Kim, Scripta
Metall. Mater. 26, 1879 (1992).
V. CONCLUSION
Nano ␥–Ni–xFe (x 19 to 21 wt%) with grain size as
small as 33 nm and with fraction density D > 98% was
synthesized by a mechanochemical method plus hot iso-
static pressing. Compression tests revealed that in the
grain size range (33–100 nm) investigated, yield stress
increased with decreasing grain size, basically in agree-
ment with a normal Hall–Petch relation. Experiments
also indicated that the yield strength (␴0.2) for n-Ni–Fe
(d 33 nm) is about 13 times greater than that for the
conventional counterpart. OM observations demon-
strated the existence of two sets of macroscopic bandlike
deformation traces mostly orienting at 45–55° to the
compression axis, while AFM observations revealed that
these traces consist of nanochannels or ultrafine lines.
The great strength or hardness of n-Ni–Fe could be
explained as strengthening effect caused by grain bound-
aries and grain refinement. Microstructural inhomogene-
ity would mainly be responsible for additional strength
difference among respective bulk materials. However,
further work is needed to elucidate the deformation
mechanism in n-Ni–Fe.
14. C. Cheung, F. Djuanda, U. Erb, and G. Palumbo, Nanostruct.
Mater. 5, 513 (1995).
15. J.R. Weertman, M. Niedzielka, and C. Youngdahl, Mechanical
Properties and Deformation Behavior of Materials Having Ultra-
fine Microstructures, edited by M. Nastasi, D.M. Parkin, and
H. Gleiter (Kluwer, Dordrecht, Germany, 1993), p. 241.
16. P.G. Sanders, J.A. Eastman, and J.R. Weertman, Acta Mater. 45,
4019 (1997).
17. G.W. Nieman, J.R. Weertman, and R.W. Siegel, J. Mater. Res. 6,
1012 (1991).
18. G.W. Nieman, J.R. Weertman, and R.W. Siegel, Scripta Metall.
Mater. 24, 145 (1990).
19. Ferromagnetic Materials, edited by E.P. Wohlfarth (North-
Holland, Amsterdam, The Netherlands, 1980), Vol. 2, p. 123.
20. C. Kuhrt and L. Schultz, J. Appl. Phys. 73, 1975 (1993).
21. F.D. Ge, L.M. Chen, W.J. Ku, and J. Zhu, Nanostruct. Mater. 8,
703 (1997).
22. S. Eroglu, S.C. Zhang, and G.L. Messing, J. Mater. Res. 11, 1231
(1996).
23. Y.H. Zhou, M. Harmelin, and J. Bigot, Scripta Metall. 23, 1391
(1989).
24. X.Y. Qin, J.S. Lee, J.G. Nam, and B.S. Kim, Nanostruct. Mater.
11, 383 (1999).
25. A.S. Helle, K.E. Easterling, and M.F. Ashby, Acta Metall. 33,
2163 (1985).
26. J.S. Lee, T.H. Kim, J.H. Yu, and S.W. Chung, Nanostruct. Mater.
9, 153 (1997).
27. X.Y. Qin, J.S. Lee, and C.S. Lee (unpublished).
28. M. Tain and T. Christman, Acta. Metall. Mater. 42, 1901
(1994).
ACKNOWLEDGMENTS
One author (Qin) gratefully acknowledges the
Alexander von Humboldt Foundation for providing a
fellowship. He especially wishes to acknowledge
Prof. E. Nembach who, as a host professor, contributed
much in both arranging facilities for this work and in-
structing the author to carry out this research. The au-
thors would like to thank the Institut fu¨r Metallforschung,
Universita¨t Mu¨nster for providing all the necessary fa-
cilities. In addition, the authors are indebted to Mr. T. Krol
for his assistance in mechanical tests, Dr. D. Baither and
1000
J. Mater. Res., Vol. 17, No. 5, May 2002
Downloaded: 18 Mar 2015
http://journals.cambridge.org
IP address: 128.255.6.125

X.Y. Qin et al.: Microstructures and mechanical behavior of bulk nanocrystalline ␥–Ni–Fe produced by a mechanochemical method
29. T.R. Malow and C.C. Koch, Metall. Mater. Trans. 29A, 2285
(1998).
34. R.W.K. Honeycombe, The Plastic Deformation of Metals, 2nd ed.
(Edward Arnold, London, United Kingdom, 1984), p. 227.
35. J. Friedel, Dislocations (Pergamon Press, London, United King-
dom, 1964), Appendix B.
36. X.Y. Qin, X.G. Zhu, G. Song, and J.S. Lee, J. Phys. Condens.
Matter 14, 2605 (2002).
30. Ferromagnetic Materials, edited by E.P. Wohlfarth (North-
Holland, Amsterdam, The Netherlands, 1980), Vol. 2, p. 123.
31. D. Wiliam and J.R. Callister, Materials Science and Engineering,
2nd ed. (John Wiley & Sons, New York, 1991), p. 112.
32. V.Y. Gertsman, M. Hoffmann, H. Gleiter, and R. Birringer, Acta
Metall. Mater. 42, 3539 (1994).
37. L.I. Trusov, T.P. Khvostantseva, V.A. Solov’ev, and V.A. Mel’nikov,
Nanostruct. Mater. 4, 803 (1994).
33. J. Friedel, Dislocations (Pergamon Press, London, United King-
dom, 1964), Chap. 8.
38. S. Hwang, C. Nishimura, and P.G. McCormick, Scripta Metall.
44, 1507 (2001).
J. Mater. Res., Vol. 17, No. 5, May 2002
Downloaded: 18 Mar 2015
1001
http://journals.cambridge.org
IP address: 128.255.6.125