Dislocation Configurations in Nanocrystalline FeMo Sintered
Components
PAOLO SCARDI, MIRCO D’INCAU, MATTEO LEONI, and ALESSANDRO FAIS
Net-shape compaction of a nanocrystalline ball-milled commercial Fe-1.5 wt pct Mo powder
was done via spark plasma sintering (SPS) and capacitor discharge sintering (CDS). A detailed
microstructure analysis, performed by X-ray diffraction–whole powder pattern modeling
(XRD-WPPM), shows that CDS, owing to the faster sintering conditions, retains a much finer
and more uniform microstructure with dislocations uniformly distributed inside the nanocrystalline grains. Conversely, SPS causes dislocations to pile up and extensive grain growth to
occur, especially when high sintering temperatures are employed.
DOI: 10.1007/s11661-009-9987-x
Ó The Minerals, Metals & Materials Society and ASM International 2009
I.
INTRODUCTION
ONE of the challenges in metal powder sintering
technology is how to transfer the peculiar properties of
nanocrystalline materials to compacted components.
This has become of special interest in recent years, in
which extensive research has been devoted to the production of nanocrystalline, heavily deformed metal
powders and their compaction by advanced sintering
techniques.[1–4] The best processing conditions—highest
(i.e. theoretical) density in the sintered component, while
preserving as much as possible the fine grain size and high
lattice defect density of the starting nanopowder—can
hardly be achieved by conventional sintering.[2,5]
Within the large family of field-assisted sintering
techniques (FAST) or electric current activated sintering
(ECAS), spark plasma sintering (SPS) has been studied
in some detail.[6] Despite the good results reported in the
literature,[7,8] sintering time is still on the order of
minutes and the process is controlled in temperature,
with an electronic feedback, thus allowing for a certain
degree of thermal equilibrium to be reached. Despite
undisputed improvements with respect to traditional
techniques, the relatively high sintering temperatures
required to reach high densities tend to anneal the
microstructure, thus losing at least part of the properties
of the starting powders.[1,2] Although suitable additions
may be used to stabilize the nanostructure and to reduce
the damaging effect of high temperature/long process
PAOLO SCARDI, Professor, and MIRCO D’INCAU and
MATTEO LEONI, Doctors, are with the Department of Materials
Engineering and Industrial Technologies, University of Trento, 38100
Trento, Italy. Contact e-mail: paolo.scardi@unitn.it ALESSANDRO
FAIS, Doctor, is with the Department of Materials Science and
Chemical Engineering, Turin Polytechnic, 10129 Torino, Italy.
This article is based on a presentation given in the symposium
‘‘Neutron and X-Ray Studies of Advanced Materials,’’ which occurred
February 15–19, 2009, during the TMS Annual Meeting in San
Francisco, CA, under the auspices of TMS, TMS Structural Materials
Division, TMS/ASM Mechanical Behavior of Materials Committee,
TMS: Advanced Characterization, Testing, and Simulation Committee, and TMS: Titanium Committee.
METALLURGICAL AND MATERIALS TRANSACTIONS A
time,[9] substantial improvements should probably
involve a different approach.
Key factors such as rapidity and effectiveness of the
sintering process seem to be peculiar to another family of
FASTs, known as single-pulse techniques, or electrodischarge sintering (EDS). Electrodischarge sintering
requires a single, short pulse of electromagnetic energy
to sinter the powder compacts while applying mechanical
pressure.[10] A recently developed EDS technique, named
capacitor discharge sintering (CDS), is based on a single
pulse of low voltage, high current electromagnetic
energy, delivered in a short time (tens of milliseconds)
to powders in a conducting mold previously loaded with
mechanical pressure. Capacitor discharge sintering, compared to traditional EDS,[11–14] employs low voltages
(from 5 to 30 V) and is composed of two mutually
coupled freely oscillating circuits with a solid-state switch
on the primary circuit analogous to those used in
capacitor discharge welding. The output voltage closely
resembles another EDS named high energy-high rate
(HEHR),[15,16] but the sintering/discharge time is only 20
to 30 ms in CDS compared to 3 seconds in HEHR.
Preliminary studies have already proved the effectiveness
of CDS in compacting several metal powders.[17–19]
The present article provides a comparison between
SPS and CDS of a typical metal system, a Mo-stabilized
iron alloy. Since the main differences between SPS and
CDS components are in the microstructure, namely, the
crystalline domain size and lattice defects density, an
advanced X-ray diffraction (XRD) line profile analysis
(LPA) method is used. Whole powder pattern modeling
(WPPM)[20–23] provides information on crystalline
domain size and dispersion, and on the density and
distribution of dislocations in the ferritic iron phase.
Correlations with two basic macroscopic properties,
density and microhardness, are discussed.
II.
EXPERIMENTAL
Fe-1.5 wt pct Mo powders were ground at increasing
times, for 8, 32, 64, 96, and 128 hours (Reference 24 for
details). Milled powders were capacitor-discharge sintered into small disks of 2 g inside a graphite mold with
metallic plungers/electrodes, which applied a nominal
pressure of 250 MPa and with different specific energy
inputs (SEIs) from 1.3 to 3.4 kJ/g. The SPS of the
different powders was carried out at 600 °C, 650 °C,
700 °C, 750 °C, and 800 °C, with a heating rate to
temperature of 50 °C/min (respectively 13, 14, 15, 16,
and 17 minutes) in graphite punches and dies and a
nominal pressure of 76 MPa. All samples where disks
with a radius of 10 mm. The densities were evaluated
with Archimede’s principle, while Vickers microhardness was obtained, after polishing, from the mean
value of 8 to 12 hardness measurements done in two
different laboratories with a force of 300 and of 50 gf
and a hold time of 15 seconds.[17,25]
To avoid including contributions from polishing,
before XRD measurements, all specimens were chemically thinned using a solution of nitric acid (4 pct)orthophosphoric acid (80 pct). The XRD patterns were
collected at each layer removal stage (approximately
20 lm), until no changes were observed after further
removal.
III.
WHOLE POWDER PATTERN MODELING
The observed (h) diffraction profile can be thought of
as a convolution of instrumental (g) and specimenrelated (f) profile components.[26] Fourier analysis is
especially useful to deal with this case, as the complex
problem of handling a convolution integral can be
turned into the relatively simpler one of a Fourier
integral:[20,21]
Z
X
Iðs; shkl Þ /
whkl Chkl ðLÞ exp½2piLðs shkl ÞdL ½1
distributed by NIST; e.g., see Reference 23 for details),
so it is in any case a known quantity. In the present case
of study, the phase of interest is a bcc metal (a-Fe) for
which we can expect the main contribution to the f
profile to come from the fine size of crystalline domains
(AS) and dislocations (AD). For the ‘‘size’’ effect, we
can assume that the approximately equiaxial domains
can be represented by spherical crystallites whose
diameters follow a lognormal distribution with mean
l and variance r.[20–22] Concerning dislocations, the
Krivoglaz–Wilkens theory[28–30] is still the most appropriate one to model the effect of a system of dislocations
with
an average density q (assuming equally a populated
110 h111i slip system) and an effective outer cut-off
radius Re.[29,30] Expressions for the FT can be evaluated
both for screw and for edge dislocations, for which it is
possible to calculate the corresponding average ‘‘‘visibility’’ effect, or contrast factor,[28–32] which represents
the anisotropic line broadening effect of the line defect
and of the crystalline medium.
Diffraction peak profiles can then be modeled by
adjusting (by nonlinear least squares) just a few microstructural parameters, which in this case are q, Re, the
dislocation character (fE, edge/screw fraction), l, and r
of the domain size distribution. Additional parameters
are used to reproduce a background, unit cell parameters, and some common aberrations of the powder
geometry. Peak intensities are usually considered as free
parameters, but they can also be constrained according
to a structural model, as in the Rietveld method.[33] If
this is the case, WPPM also provides for a quantitative
phase analysis, of interest in phase mixtures.
For further details and a survey of WPPM applications, see Reference 23 and the references therein.
IV.
RESULTS AND DISCUSSION
hkl
where s is the reciprocal space variable, and shkl is s
in Bragg condition, including possible shifts due to
lattice defects (e.g., faulting). Since the latter may differently affect the various equivalent points in reciprocal space, the observed powder diffraction profile
results from the sum of Fourier integrals for all hkl
equivalent combinations, with a weight whkl. Chkl is
the product of the Fourier transforms (FTs) of each
profile component:
Chkl ¼ TIP AS AF AD AAPB . . .
Figure 1 shows an example of WPPM for a FeMo
sintered component. In this specific case (CDS 64 hour),
½2
where TIP groups all contributions to the g profile, and
the other terms (A) refer to f components. The latter are
known—in many cases, in closed analytical form—for
several contributing effects, such as domain shape and
size distributions (AS), dislocations (AD), planar defects
(AF), antiphase boundaries (AAPB), etc. Virtually any
other contributing effect can be included by adding
(multiplying in Eq. [2]) the corresponding FT.[20–23]
The g component can be handled by a fundamental
parameters approach[27] or, more simply, through a
parameterization of the experimental pattern of a
suitable line profile standard (e.g., LaB6 SRM 660a
Fig. 1—WPPM results for CDS 64-h sintered component: experimental data (dot), model (line), and their difference (residual, line
below). The inset shows a logarithmic plot, with indications of the
most intense lines of the c-Fe minority fraction (arrows).
METALLURGICAL AND MATERIALS TRANSACTIONS A
as in the other cases discussed subsequently, a minor
fraction of c-Fe (arrows) appears together with the main
a-Fe phase. Minor amounts of iron oxide (magnetite)
were also found in several sintered components. As
WPPM supports the analysis of all phases, allowing also
for a quantification of the relative fractions, microstructural results for the main phase of interest, a-Fe, are
nearly unaffected by the presence of other phases. The
main results of WPPM are summarized in Figure 2,
where average dislocation density ((a) and (b)) and mean
domain size ((c) and (d)) are shown as a function of
sintering temperature (T in °C) and SEI (in J/g),
respectively, for SPS and CDS components.
As a general trend, dislocation density tends to
decrease with the SPS temperature, the diminution
being more evident for 96- and 128-hour powders
sintered above 750 °C. Correspondingly, the domain
size increases, with a marked growth above 750 °C. The
domain sizes for SPS at 800 °C are of several hundreds
of nanometers, close to, or possibly above, values
reliably accessible to XRD LPA. The overall result is
that the nanostructure is preserved up to sintering
temperatures of 750 °C, with the powders ball milled for
128 hours preserving a proportionally higher dislocation
density/smaller domain size than the powders ground
for shorter time. Above 750 °C, extensive grain growth
and defect annealing takes place.
Capacitor discharge sintering also causes a marked
decrease in dislocation density, the difference between
different powders decreasing for increasing SEI. The
main difference with respect to SPS seems to be the
domain size, never exceeding 100 to 150 nm, even for
the highest SEI employed. According to this information, CDS seems to better preserve the nanostructure
than SPS. Electron microscopy supports this hypothesis.
As shown in Figure 3, CDS gives a uniform, finegrained (<1 lm) microstructure, whereas SPS clearly
shows a coarser microstructure. The bimodality of the
SPS grain size distribution is also evident, with scattered
regions of fine grains (submicron size) separated by
larger grains (5 to 10 lm). This feature of the SPS has
already been observed[34,35] and can partly be controlled
by varying (increasing) the pressure.
As a potential drawback of the CDS, a minor but
clearly measurable fraction of austenitic phase is
formed. As shown in Figure 4, the amount of austenite
tends to increase for increasing SEI, and the effect is
more visible for the most extensively ball-milled pow-
(a)
(b)
(c)
(d)
Fig. 2—WPPM results. (a) and (b) Average dislocation density and (c) and (d) mean domain size as a function of sintering temperature or SEI,
respectively, for SPS or CDS components obtained from powders ball milled for 128 h (square), 96 h (circle), 64 h (rhomb), and 32 h (triangle).
Lines are drawn just to drive the eye. Data points at T = 0 and SEI = 0 refer to the ball-milled powders, before sintering.
METALLURGICAL AND MATERIALS TRANSACTIONS A
(a)
Fig. 3—ESEM micrographs of (a) SPS 96 h and (b) CDS 96 h
components.
(b)
Fig. 5—Correlation between microhardness and density for (a) SPS
and (b) CDS components: Hv 0.05 (full symbol) and Hv 0.3 (open
symbol). The dashed line highlights the corner region of high microhardness/large density. Refer to the text for discussion.
Fig. 4—Residual austenite in CDS components as a function of SEI.
Trends refer to the four starting powders, ground for different times
(32, 64, 96, and 128 h).
ders, which have a higher excess energy. The SPS
components are made of ferritic phase only, with traces
of iron oxide, with the exception of the components
made of 128-hour powder sintered at 750 °C and
800 °C, where iron oxides reach 3 pct and austenite is
about 2 pct.
It is interesting to integrate the information provided
by XRD-LPA with two basic properties of the sintered
components: density and microhardness. Data are
shown in Figure 5 for (a) SPS and (b) CDS components,
using two indentation loads. It is quite evident that CDS
components reach higher densities and higher microhardness (upper-right corner), the latter being slightly
sensitive to variations in the load. On the contrary, SPS
components show a remarkable difference in microhardness, depending both on the density and on the
applied load. These results are coherent with the WPPM
results and with the morphological observations. Capacitor discharge sintering provides a much finer and more
uniform microstructure than SPS, and the latter causes
extensive annealing when high temperatures (>750 °C)
are employed.
Whole powder pattern modeling provides additional
information, useful for a deeper understanding of the
two different sintering processes. In fact, in addition to
domain size and dislocation density, WPPM also
METALLURGICAL AND MATERIALS TRANSACTIONS A
provides the effective outer cut-off radius, which can be
conveniently combined with the dislocation
pffiffiffi density in
the so-called Wilkens parameter, M ¼ Re q. The latter
is shown in Figure 6 for SPS (open symbols) and CDS
(full symbols) components, as a function of the dislocation density. It is worth noting that, in this plot,
movement from right to left (i.e., for decreasing dislocation density) corresponds to components sintered at
increasing SEI (CDS) and increasing T (SPS). The value
of M tends to decrease after the SPS process, whereas it
increases markedly after CDS. According to the theory
of the line broadening effects of dislocations,[28–30] M
values close to (and below) unity are observed when
dislocations are strongly interacting, as in dipole configurations or in dislocation walls: this is the case for the
SPS components, as the result of near thermal equilibrium process conditions, and could be the result of a
tendency of dislocations to pileup and reach lower
energy configurations, as in all temperature-activated
sintering processes. A Wilkens parameter well above
unity, instead, may result from randomly dispersed
dislocations. This is the case observed in CDS components, where the fast sintering conditions push the
system very far from equilibrium, with the dislocations
remaining ‘‘frozen’’ in a random distribution inside the
crystalline domains.
It is therefore concluded that, although the average
dislocation density after SPS and CDS is similar, the
distribution of the dislocations is probably different, and
might contribute to an explanation of the higher
microhardness observed for CDS components with
respect to SPS components with similar density and
mean domain size. This observation confirms analogous
results obtained in a sensibly different system such as
pure copper[19] and is still the object of ongoing research
activity.
pffiffiffi
Fig. 6—Correlation between the Wilkens parameter (M ¼ Re q)
and average dislocation density (q) for CDS (full symbols) and SPS
(open symbols) components obtained from nanocrystalline FeMo
powders ground for different times: 128 h (square), 96 h (circle), and
64 h (triangle). Full symbols in gray on the right side (after the
x-axis break) refer to the starting powders. Lines are drawn just to
highlight the overall trend.
METALLURGICAL AND MATERIALS TRANSACTIONS A
As a final remark, this article shows that XRD LPA,
following a WPPM approach, provides detailed information on the microstructure of sintered components
from nanocrystalline metal powders. Information is
given on the domain size as well as on the density and
possible arrangement of line defects. The same approach
can be used for any type of crystalline material, as the
applicability of the model has recently been extended to
systems with any crystal symmetry,[31,32] and can also
account for other microstructural features and lattice
defects that produce measurable effects on the diffraction line profile.[23]
ACKNOWLEDGMENT
The authors thank Dr. M. Zadra for the precious
support in the preparation of SPS components.
REFERENCES
1. H.J. Fecht: in Nanostructured Materials and Composites Prepared
by Solid State Processing, C.C. Koch, ed., Noyes Publications,
Norwick, NY, 2002, pp. 73–113.
2. J.R. Groza: in Nanostructured Materials and Composites Prepared
by Solid State Processing, C.C. Koch, ed., Noyes Publications,
Norwick, NY, 2002, pp. 115–78.
3. C. Suryanarayana: Prog. Mater. Sci., 2001, vol. 46, pp. 1–184.
4. R.Z. Valiev, R.K. Islamgaliev, and I.V. Alexandrov: Prog. Mater.
Sci., 2000, vol. 45, p. 103.
5. C.C. Koch: J. Mater. Sci., 2007, vol. 42, pp. 1403–14.
6. J. Rawers: Nanostruct. Mater., 1999, vol. 11 (4), pp. 513–22.
7. K.Y. Sastry, L. Froyen, J. Vleugels, O. Van der Biest,
R. Schattevoy, and J. Hennicke: Rev. Adv. Mater. Sci., 2004,
vol. 8 (1), pp. 27–32.
8. K.Y. Sastry, L. Froyen, J. Vleugels, O. Van der Biest, R. Schattevoy,
and K. Hummert: Mater. Sci. Forum, 2006, vols. 519–521,
pp. 1409–14.
9. S. Libardi, M. Leoni, L. Facchini, M. D’Incau, P. Scardi, and A.
Molinari: Mater. Sci. Eng. A, 2007, vols. 445–446, pp. 244–50.
10. R. Orrù, R. Licheri, A.M. Locci, A. Cincotti, and G. Cao: Mater.
Sci. Eng. R, 2009, vol. 63, pp. 127–288.
11. S. Clyens, S.T.S. Al-Hassani, and W. Johnson: Int. J. Mech. Sci.,
1976, vol. 18 (1), pp. 37–44.
12. T. Alp, S.T.S. Al-Hassani, and W. Johnson: J. Eng. Mater.
Technol., 1985, vol. 107, p. 109.
13. W. Johnson, S. Clyens, and S.T.S. Al-Hassani: Metallurgia Met.
Form., 1976, vol. 43, p. 82.
14. M. Shakery, S.T.S. Al-Hassani, and T.J. Davis: Powder Metall.
Int., 1979, vol. 11 (3), pp. 120–25.
15. D.R. Ervin, D.L. Bourell, C. Persad, and L. Rabenberg: Mater.
Sci. Eng. A, 1988, vol. 102, pp. 25–30.
16. S.L. Raghunathan, C. Persad, D.L. Bourell, and H.L. Marcus:
Mater. Sci. Eng. A, 1991, vol. 131, pp. 243–53.
17. A. Fais: Ph.D. Thesis, Turin Polytechnic, Turin, 2008.
18. A. Fais and G. Maizza: J. Mater. Proc. Technol., 2008, vol. 202,
pp. 70–75.
19. A. Fais and P. Scardi: Z. Kristallogr. Suppl., 2008, vol. 27,
pp. 37–44.
20. P. Scardi and M. Leoni: Acta Cryst. A, 2001, vol. 57, pp. 604–13.
21. P. Scardi and M. Leoni: Acta Cryst. A, 1002, vol. 58, pp. 190–200.
22. M. Leoni and P. Scardi: J. Appl. Cryst., 2004, vol. 37, pp. 629–34.
23. P. Scardi: in Powder Diffraction: Theory and Practice, R.E.
Dinnebier and S.J.L. Billinge, eds., The Royal Society of
Chemistry, Cambridge, United Kingdom, 2008, pp. 376–413.
24. M. D’Incau, M. Leoni, and P. Scardi: J. Mater. Res., 2007, vol. 22,
pp. 1744–53.
25. M. D’Incau: Ph.D. Thesis, University of Trento, Trento, 2008.
26. H.P. Klug and L.E. Alexander: X-ray Diffraction Procedures for
Polycrystalline and Amorphous Materials, Wiley, New York, NY,
1974.
27. R.W. Cheary and A. Coelho: J. Appl. Cryst., 1992, vol. 25,
pp. 109–21.
28. M.A. Krivoglaz and K.P. Ryaboshapka: Fiz. Metall., 1963,
vol. 15, pp. 18–28.
29. M. Wilkens: Phys. Status Solidi A, 1970, vol. 2, pp. 359–70.
30. M. Wilkens: Fundamental Aspects of Dislocation Theory, J.A.
Simmons, R. de Wit, and R. Bullough, eds., National Bureau of
Standards (US) Special Publication No. 317, NBS, Washington,
DC, vol. II, pp. 1195–1221.
31. M. Leoni, J. Martinez-Garcia, and P. Scardi: J. Appl. Cryst., 2007,
vol. 40, pp. 719–24.
32. J. Martinez-Garcia, M. Leoni, and P. Scardi: Acta Cryst. A, 2009,
vol. 65, pp. 109–19.
33. R.A. Young: The Rietveld Method, Oxford University Press,
Oxford, United Kingdom, 1993.
34. S. Kawamura and H. Kuwano: J. Jpn. Soc. Powder Metall., 2003,
vol. 50 (7), pp. 545–50.
35. S. Kawaura, H. Kuwano, Y. Takeda, S. Takahashi, and H. Kaga:
J. Jpn. Soc. Powder Metall., 2003, vol. 50 (12), pp. 1052–56.
METALLURGICAL AND MATERIALS TRANSACTIONS A
Keep reading this paper — and 50 million others — with a free Academia account
Used by leading Academics
Kazuki Morita
The University of Tokyo
Suseendran Jayachandran
IMEC
Estela Blaisten-Barojas
George Mason University
Kenneth Vecchio
University of California, San Diego
