EPMA Proceedings 2009 Sintering PDF

Euro PM2009 Field Assisted Sintering
Manuscript refereed by Professor Bernd Kieback, Fraunhofer IFAM/TU Dresden,

Germany

Sintering of WC-12Co Powders with Different Particle Sizes by
Pulsed Electric Current Sintering and Effect of Powder
Characteristics on Microstructure and Mechanical Properties

M. E. Cura
a
, J. Lagerbom
b
, J. Lotta
a
, X.W. Liu
a
, J. Syren
a
, T. Ritvonen
b
, U. Kanerva
a
, Y. Ge
a
,
R. Ritasalo
a
, O. Sderberg
a
, E. Turunen
b
, S-P. Hannula
a
a
Helsinki University of Technology, Department of Materials Science and Engineering,
POB6200, FI-02015, Espoo, Finland;
b
VTT, Technical Research Center of Finland,
POB1300, FI-33101, Tampere Finland

Abstract. WC-12Co powders were prepared from three different WC carbide and metallic Co powder
grades (nano, fine and coarse) by vibrator milling for 5, 15, or 30 min in argon. Powders were
characterised by laser diffractometer, SEM/EDS, and XRD. Nano-Co was cubic in the milled mixtures,
while the other Co grades transformed to hexagonal. All the powders were pulsed electric current
sintered by using the processing parameters 1453 K, 3 min, 50 MPa, and 100 K/min. Densification
was studied by Archimedes method. Microstructure of the compacted samples was analysed by
SEM/EDS, and XRD. The WC grain size and porosity were measured from SEM pictures by applying
the image analysis. The mechanical properties of the compacts were studied with HV1 microhardness
test and by nanoindentation for one sample. Nearly full density was obtained in compacts with fine-
WC/nano-Co-powder, while the highest hardness 23 GPa was obtained with powders containing
nano-WC.
1. Introduction
Hard metals and their composites are commonly used in harsh environments such as drilling,
metal cutting, wire drawing, mining, metal extrusion due to their high hardness and wear resistance,
good chemical stability and high melting point. In cemented carbide hard metals carbide particles are
embedded in a continuous metal binder skeleton [1-3]. These materials are produced by powder
metallurgy methods including liquid phase sintering, hot pressing and hot isostatic pressing. Relatively
new sintering methods such as pulsed electric current sintering, high-frequency induction heating
sintering, electric discharge sintering and microwave sintering are also used for consolidation [3-8].
The mechanical properties of WC-Co cermets can be controlled by the characteristics of starting
powders and cobalt content [9]. Size of WC particles has a great influence on hardness, strength and
wear-resistance of the final material [2,4,6]. Lately considerable efforts in research have been
demonstrated on producing, processing and consolidating of ultrafine and nanocrystalline WC-Co
powders [3,4,6,10]. During sintering grain growth occurs and it is usually not possible to observe
nanocrystalline structure afterwards. This is more severe in conventional techniques, and more rapid
for finer powders [11]. Coarsening of grains is believed to be result of grain boundary migration and
Ostwald ripening process [11,12]. Grain growth can be controlled by addition of different carbide
phases (VC, Cr
3
C
2
, NbC, Mo
2
C, etc.) as grain growth inhibitors and by controlling sintering parameters
[13-15]. In the recent years sintering techniques exhibiting rapid heating and shorter sintering periods
have been of interest i.e. pulsed electric current sintering (PECS) also known as spark plasma
sintering (SPS) or field assisted sintering technique (FAST). It is a relatively new method for sintering
of wide variety of materials and has attracted a lot of attention in the last decade. It uses Joule
heating, provided by electric current flowing through two electrodes and powders in graphite moulds
between them. It is fast and requires lower sintering temperatures and durations than conventional
methods. Uniaxial pressure is applied for compressing the powder and to provide a continuous contact
between the two electrodes [2-10]. In the present study, WC 12 % wt Co powders with different
particle sizes were consolidated by PECS. Effect of powder characteristics on densification and
hardness has been studied. Microstructure examination and grain size analysis were conducted for
the sintered pieces.
2. Experimental
The nano WC powder was milled from the fine powder having nanofractioned structure with a
particle size of 200 nm (Zhuzhou Cemented Carbide Works Import & Export Company). The other
start powders used were the fine size WC DS80 powder (0.8-0.9 m, H.C. Starck), coarse technical
grade (2-5 m, H.C. Starck), nano cobalt with average particle size of ~28 nm (variation 0-60 nm;
specific surface area 40-60 m
2
/g, NaBond Technologies Co., Ltd), fine Co was S-80 (4 m) and
coarse Co S-320 (7.9 m, OMG Kokkola Chemicals Ltd) were used.The WC-12Co powder mixtures
were prepared by using Spex 8000 high energy vibratory ball mill. Three different mixing times, 5, 15
and 30 min were applied, and mixing was carried out in argon. The powder size distribution of the
powders was studied by laser diffractometer (Lecotrac LT100). Powder morphology was examined
with the scanning electron microscope. Consolidation of the powders was carried out with pulsed
electric current sintering equipment (FCT HP D 25-2) sintering was carried out at 1453 K for 3 min,
under the vacuum of 7 Pa. Powders were pressed with a pressure of 50 MPa, heating rate was 100
K/min and pulse/pause ratio was 10:5. The samples were compacted in 20 mm diameter graphite
moulds and 18 g of powder was used in every experiment. After sintering all the samples were ground
and polished with the diamond discs (120-1200 grits) and diamond paste (6, 3 and 1 m) respectively,
and a final polishing was made by colloidal silica (0.16 m, pH 9.8). The densities of the sintered
compacts were measured by applying the Archimedes method (Sartorius CPA224S, 0.1 mg). While
evaluating the theoretical density, the measured density was compared to the value of 14.46 g/cm
3
[5].
Grain size measurements were made by image analysis from SEM images of sintered compacts [16-
18].The Vickers microhardness measurements with 1 kg load were measured with the Zwick & Co
Z323 apparatus. Nanoindentation was also performed to the compact sintered from 30 min milled
nano WC and nano Co grade powder. Nine indents were made with Berkovich tip and the applied load
was 2000 N. The XRDs of the materials were measured with Philips PW 3830 powder diffractometer
(Cu
K
radiation).
3. Results and Discussion
3.1 Powder characteristics
The powder grades for WC and cobalt used in the experiments were nano, fine and coarse. The
particle size distributions after grinding are given in Table 1. The lowest size measured was 0.183
0.894 2.765 for nano grade and highest was 0.443 3.847 15.91 for the coarse grade as in d
10

d
50
d
90
. Three different milling times (5 min, 15 min and 30 min.) apparently did not have any
significant effect on particle size, the values being close to each other. However, longer milling time
helped breaking the agglomerates which improved the densification. Especially in nano grade
powders agglomeration was in large scale before milling.
Table 1. Powder grades, their particle size distributions and properties of sintered compacts.
Sample
Powder Compact
Raw material
powders (type)
Milling
time
(min)
Particle size distribution
d10-d50-d90 (1. peak)
(m)
Density
Hardness
(HV1)
Measured
(g/cm
3
)
T.D.
1nn
200 nm WC
28 nm Co
5 0.773 2.268 14.61 (0.315) 13.9215 0.0429 96.3 0.3 2304.0 57.0
2nn 15 0.183 0.894 2.765 (0.223) 14.0719 0.0307 97.3 0.2 2225.1 54.0
3nn 30 0.256 1.040 1.619 (0.289) 14.1400 0.0395 97.8 0.3 2294.8 79.3
1nf
200 nm WC
4 m Co
5 0.303 1.562 12.73 (0.243) 14.0777 0.0827 97.4 0.6 2105.8 65.9
2nf 15 0.298 1.215 7.003 (0.289) 14.0477 0.0232 97.1 0.2 2123.1 45.9
3nf 30 0.180 0.747 3.829 (0.223) 14.0445 0.0408 97.1 0.3 2196.4 48.2
1nc
200 nm WC
7.9 m Co
5 0.193 0.961 12.99 (0.223) 14.0133 0.0553 96.9 0.4 2295.3 88.4
2nc 15 0.185 0.857 8.489 (0.223) 14.0402 0.0380 97.1 0.3 2274.3 66.6
3nc 30 0.159 0.696 6.335 (0.204) 14.0303 0.0349 97.0 0.2 2285.0 83.8
1fn 0.8-0.9 m
WC
28 nm Co
5 0.228 1.195 12.73 (0.315) 14.3777 0.0418 99.4 0.3 1996.6 89.0
2fn 15 0.256 1.040 1.619 (0.289) 14.3902 0.0684 99.5 0.5 1951.6 129.6
3fn 30 0.216 0.999 2.017 (0.265) 14.3955 0.0267 99.6 0.2 2045.9 77.9
1ff 0.8-0.9 m
WC
4 m Co
5 0.215 0.935 2.630 (0.289) 14.2091 0.0366 98.3 0.3 1877.0 38.3
2ff 15 0.260 1.027 2.548 (0.315) 14.2122 0.0360 98.3 0.2 1832.8 30.0
3ff 30 0.250 1.001 2.328 (0.315) 14.2336 0.0235 98.4 0.2 1862.5 49.6
1fc 0.8-0.9 m
WC
7.9 m Co
5 0.243 1.037 4.952 (0.289) 14.1662 0.0353 98.0 0.2 1862.5 49.6
2fc 15 0.119 0.409 1.838 (0.172) 14.1433 0.0492 97.8 0.2 1878.5 74.1
3fc 30 0.210 0.973 5.279 (0.243) 14.0759 0.0604 97.3 0.4 1847.6 44.0
1cn
2-5 m WC
28 nm Co
5 0.756 1.690 10.42 (0.289) 13.9157 0.0629 96.2 0.4 1384.1 139.8
2cn 15 0.498 1.553 11.13 (0.289) 13.8975 0.0488 96.1 0.3 1320.5 108.4
3cn 30 0.606 1.763 11.63 (0.289) 14.0665 0.0344 97.3 0.2 1397.4 31.8
1cf
2-5 m WC
4 m Co
5 0.374 2.623 11.48 (0.243) 13.8649 0.0342 95.9 0.2 1395.8 87.7
2cf 15 0.306 2.382 - 10.64 (0.243) 13.9379 0.0731 96.4 0.5 1366.1 65.9
3cf 30 0.384 2.113 - 9.071 (0.315) 13.9447 0.0477 96.4 0.3 1432.3 52.1
1cc
2-5 m WC
7.9 m Co
5 0.443 3.847 15.91 (0.315) 13.7535 0.0450 95.1 0.3 1369.9 51.9
2cc 15 0.322 3.102 14.73 (0.243) 13.8941 0.0553 96.1 0.4 1417.7 54.6
3cc 30 0.278 2.321 11.77 (0.243) 13.8328 0.0325 95.7 0.2 1442.6 59.5

According to Figure 1, the powders deagglomerate as a function of milling time. Thus, 30 min milling
time was accepted as optimum duration in this study and microstructural evaluation was carried out
with compacts of 30 min milled powders. The morphologies of selected powders are given in Figure 2.

Figure 1. Particle size distributions after 5, 15 and 30 minutes of milling of nano WC and Co powder.

Figure 2. SEM images of powders with different WC and Co grades, milled for 30 min:
a) 3nn b) 3nf, c) 3nc, d) 3fn, e) 3ff, f) 3fc, g) 3cc.
The XRD study did not show significant change of the structure after different milling times, and
therefore, only the patterns for the fine-WC with the different cobalt grades and for the fine-Co mixed
with different WC grades are shown in Figure 3. Cubic cobalt was observed only in the nano-sized
cobalt containing powders after all milling times, while in all the other powders the cobalt structure was
hexagonal independently of the milling time. This is in accordance with the fact that the smaller grain
size favours the cubic structure in cobalt [1]. However, the original cobalt powder was in the cubic
form, and it transformed to the hexagonal form due to the mechanically induced transformation;
according to [19] even nano-cobalt would transform to hexagonal after 10 hours milling. When
comparing the different WC-grades in mixture with fine-Co (Figure 3b), no remarkable difference in the
WC peaks of XRD could be observed. In coarse WC grade powders metallic tungsten and W
2
C
observed in the XRD, are probably due to poor carburisation during production process.

Figure 3. XRD diffractograms for (a) for the fine-WC mixed with different cobalt grades and
(b) for the fine-Co mixed with different WC grades. All the powders were milled for 30 min.
3.2 Densification and Microstructure
Sintering temperature, time and pressure were kept constant (1453 K, 3 min, 50 MPa,
respectively) during all compactions in order to observe the effect of initial powder characteristics. The
heating rate was 100 K/min in order to keep thermal cycle relatively short. In the conventional sintering
densification starts already at heating stage and this contributes the grain growth. As faster heating
rates are applicable in pulsed electric current sintering, therefore both the fast heating and the short
processing time enables the retaining of the original grain size.
The density comparison of sintered compacts from different grade powders is given in Figure 4.
Fine WC containing powders were observed to result in the highest density. In this grade the density
values were increased for the smaller particle size of cobalt. Compacts made from powders milled for
30 min showed either clearly better or nearly the same densities as those milled for 15 min. Breaking
of agglomerates can be the reason for this. The microstructures of the bulk materials for the powders
shown in Figure 2 are given in Figure 5. Despite of the high theoretical density values of the sintered
materials, discontinuity of WC structure is observed in SEM images as the WC particles on the very
surface may be pulled out during sample preparation (due to the long polishing stage). Those samples
having these Co lakes in the microstructure showed higher standard deviation during hardness
measurements (Table 1, Figure 4). Also, grain growth is observed in the sintered compacts.
Coarsening of grains was more severe in those samples compacted from nano-powders. The average
grain sizes of compacts sintered from powders 3nn, 3nf, 3nc, 3fn, 3ff, 3fc, 3cc were 0.4, 0.37, 0.41,
0.50, 0.53, 0.52, and 0.56 m, respectively.
Nano WC Fine WC Coarse WC
92,0
93,0
94,0
95,0
96,0
97,0
98,0
99,0
100,0
D
e
n
s
i
t
y
,
T
.
D
.
%
5 min
15 min
30 min

Figure 4. Densities of sintered compacts.

Figure 5. SEM images of bulk materials compacted of the powders shown in Figure 2:
a) 3nn b) 3nf, c) 3nc, d) 3fn, e) 3ff, f) 3fc, g) 3cc.
In Figure 6, the XRD patterns of the powders are compared to those of the respective bulk
materials. The hexagonal Co of the powders is transformed to cubic above 683 K which is clearly
exceeded in the sintering process [1-19], and since no reverse transformation exists during cooling,
the observed cobalt in the bulks is cubic. In the sintered compacts of fine WC and nano Co, -carbide
phase of Co
3
W
3
C is observed (Fig. 6b). During high temperature processing WC particles can partially
dissolve in the binder phase and form W
2
C and carbon rich -phases (Co
x
W
y
C) [8]. Formation of first
is normally due to low C content. The amount is related with particles size of the components and
heating rate.

Figure 6. XRD diffractograms for the powders and bulks of (a) the nano-WC grade combined to different Co
grades, and (b) the fine-WC mixed with different Co grades. All the powders were milled for 30 min, and the bulks
were pulsed electric current sintered with 1453 K, 3 min, 50 MPa, and 100 K/min.
3.3 Mechanical Properties
Measured Vickers microhardness values are presented in Figure 7. They are in good
accordance with earlier reported values [5-7]. Microhardness values close to 2300 HV1 (e.g. 23 GPa)
were measured for nano-grade materials which is higher than published results for WC 12%wt Co
material made with pulsed electric current sintering [5,6,13,14]. On the other hand, rather low
hardness values in the range of 1300-1400 HV1 (~14 GPa) were measured for those samples
containing coarse WC grade. Even though the fine grade materials had the highest densification, their
hardness values were lower than nano-grade ones. Bulk materials prepared from 30 min milled
powders had slightly higher hardness values probably due to deagglomeration during milling and more
homogenous distribution of carbide and metal powders. Nanoindentation of the 3nn compact was
mostly made on WC particles resulting in values between 16.6-34.3 GPa range (Figure 8). These
values were comparable with values from literature [20].
0
500
1000
1500
2000
2500
3000
H
a
r
d
n
e
s
s
,

H
V
1

5 min
15 min
30 min

Figure 7. Vickers microhardness (HV1) of the sintered compacts.
SPS_WC-12Co: Berkovich tip / P
max
= 2000N / Load control
1.6 m
1
2
4 3
5
6
7
8
9
Point H(GPa) Er(GPa) E(GPa)
1 21.0 376.6 528.7
2 16.6 286.6 360.2
3 19.1 292.0 369.3
4 30.0 439.9 673.0
5 29.1 414.5 612.1
6 21.4 365.8 506.4
7 22.0 359.9 494.5
8 34.3 400.3 579.9
9 30.3 381.0 538.0
Literature values of *
WC: 12-24 / 700 GPa
Co: 2 / 200 GPa

Figure 8. Nanoindentation results of the 3nn bulk material. (a) SPM images of the measuring spot before
indentations; (b) the same after the test.

4. Conclusions
WC-12Co powders prepared from the nano-Co exhibited the cubic cobalt still after the longest
milling time of 30 min, while in the other powders cobalt was transformed to hexagonal phase even
with the shortest milling time (5 min). Pulsed electric current sintering produced fully dense compacts
in three minutes at 1453 K from the mixed fine grade WC and the nano grade Co powders. In the
other Co grade mixtures, hexagonal cobalt was transformed into cubic cobalt during sintering process.
The highest microhardness (23 GPa) value was achieved for compacts sintered from the nano-WC
and the nano-Co powder mixtures. Hardness value measured by nanoindentation of the WC particles
in the same material was as high as 34.4 GPa. The grain sizes of final compacts were between 0.37
and 0.56 m which indicates little grain growth during sintering.
Acknowledgements
The authors thank acknowledge also the Finnish Funding Agency for Technology and Innovation
(Tekes), and the consortium of Finnish Companies, for funding the study. Also, the Center of New
Materials at TKK is acknowledged for the partial funding of the SPS equipment.
References
1. Upadhyaya, G.S. Cemented Tungsten Carbides - Production, Properties, and Testing. USA:
William Andrew Publishing/Noyes, 1998. 436 s.
2. Zhao J.F., Holland T., Unuvar C., Munir Z.A., Sparking plasma sintering of nanometric tungsten
carbide. Int. J. Refract. Metal. Hard Mater. 27, 130-139.
3. Fang Z.Z., Wang X. Ryu T.G., Hwang K.S., Sohn H.Y., Synthesis, sintering, and mechanical
properties of nanocrystalline cemented tungsten carbide A review. Int. J. Refract. Metal. Hard
Mater 2009, 27, 288-299.
4. Z. A. Munir, U. Anselmi-Tamburini, M. Ohyanagi, The effect of electric field and pressure on the
synthesis and consolidation of materials: A review of the spark plasma sintering method,
J.Material Science, 2006, 41, 763-777
5. Sivaprahasam D, Chandrasekar SB, Sundaresan R. Microstructure and mechanical properties
of nanocrystalline WC12Co consolidated by spark plasma sintering. Int. J. Refract. Metal. Hard
Mater. 2007, 25, 144152.
6. Kim H-C., Shon I-J., Yoon J-K., Doh J-M., Consolidation of ultra fine WC and WC-Co hard
materials by pulsed current activated sintering and its mechanical properties. Int. J. Refract.
Metal. Hard Mater. 2007, 25, 46-52.
7. Kim H.C., Jeong I.K., Shon I.J., Ko I.Y., Doh J.M., Fabrication of WC-8 wt.%Co hard materials
by two rapid sintering processes. Int. J. Refract. Metal. Hard Mater. 2007, 25, 336-340.
8. Picas J.A., Xiong Y., Punset M., Ajdelsztajn L., Forn A., Schoenung J.M., Microstructure and
wear resistance of WC-Co by three consolidation processing techniques. Int. J. Refract. Metal.
Hard Mater. 2009, 27, 344-349.
9. Huang S.G., Vanmeensel K., Li L., Van der Biest O., Vleugels J., Influence of starting powder
on the microstructure of WCCo hardmetals obtained by spark plasma sintering. Mat. Sci. Eng.
A, 2008, 475, 87-91.
10. Zhao S.X., Song X.Y., Zhang J.X., Liu X.M., Effects of scale combination and contact condition
of raw powders on SPS sintered near-nanocrystalline WC-Co alloy, Mat. Sci. Eng. A 2008, 473,
323-329.
11. Wang X., Fang Z.Z., Sohn H.Y., Grain growth during the early stage of sintering of nanosized
WC-Co powder. Int. J. Refract. Metal. Hard Mater. 2008, 26, 232-241.
12. Weidow J., Norgren S., Andrn H-O., Effect of V, Cr and Mn additions on the microstructure of
WC-Co. Int. J. Refract. Metal. Hard Mater (2009) doi:10.1016/j.ijrmhm.2009.02.002.
13. Huang S.G., Liu R.L., Li L., Van der Biest O., Vleugels J., NbC as grain growth inhibitor and
carbide in WC-Co hardmetals. Int. J. Refract. Metal. Hard Mater. 2008, 26, 389-395.
14. Huang S.G., Li L., Vanmeensel K., Van der Biest O., Vleugels J., VC, Cr
3
C
2
and NbC doped
WC-Co cemented carbides prepared by pulsed electric current sintering. Int. J. Refract. Metal.
Hard Mater. 2007, 25, 417-422.
15. Huang S.G., Vanmeensel K., Li L., Van der Biest O., Vleugels J., Tailored sintering of VC-doped
WC-Co cemented carbides by pulsed electric current sintering. Int. J. Refract. Metal. Hard
Mater. 2008, 26, 256262.
16. Park S.J., Cowan K., Johnson J.L., German R.M., Grain size measurement methods and
models for nanograined WC-Co. Int. J. Refract. Metal. Hard Mater. 2008, 26, 152-163.
17. Mannesson K., Elfwing M., Kusoffsky A., Norgren S., gren J., Analysis of WC grain growth
during sintering using electron backscatter diffraction and image analysis. Int. J. Refract. Metal.
Hard Mater.2008, 26, 449-455.
18. Savard M., The characterization of cemented carbides by automated image analysis, JOM
2008, 60, 23-28.
19. Zhang F.L., Wang C.Y., Zhu M., Nanostructured WC/Co composite powder prepared by high
energy ball milling, Scripta Mat. 2003, 49, 1123-1128.
20. H. Engqvist, U. Wiklund, Mapping of mechanical properties of WCCo using nanoindentation,
Tribology Letters 2000, 8, 147152
Manuscript refereed by Dr J rgen Schmidt, Fraunhofer Institute, Germany

Electro Discharge Sintering as a Process for Rapid Compaction in
PM-Technology

P. Schtte
1
, J. Garcia
1,2
, W. Theisen
1

1
Chair of Materials Technology, Ruhr University Bochum, D- 44780 Bochum, Germany
2
Helmholtz-Zentrum Berlin fr Materialien und Energie GmbH, D-14109 Berlin, Germany

Abstract
Electro Discharge Sintering (EDS) is used as a rapid compaction process for metal powders.
The application of a constant pressure and a current pulse, in the amount of 3 x 10
3
ampere,
yields to a dense microstructure. The process takes place within milliseconds. The new
sintering technique discloses the potential to avoid diffusion-controlled changes in the
microstructure, which take place in conventional sintering techniques. In this work first
investigations of the physical densification on metal powders, produced with the EDS
technique, are presented. Correlations between sintering parameters (energy, pressure),
powders characteristics (grain size, alloying elements) and physical properties (density) are
shown.

Introduction
The technique of EDS has been introduced in 1976 by Clyens, Al-Hassani and Johnson [1].
In the beginning of the development stage porous bodies were consolidated. This was
related to low energy sintering machines using capacitors that provided 80 F [2]. Recent
papers describe the current assisted consolidation of metal powders to highly dense parts.
Densification of aluminium [3], WC-Co [4], AISI M2 high speed steel [5] as well as titanium,
tin and zinc [6] are presented. First investigations regarding process theory and technology
are carried out by [7, 8]. Although the Field Activated Sintering Technology (FAST)
processes have been extensively investigated, a fully understanding of the process
phenomena is still lacking.

Materials
In this paper atomized steel powders are investigated with a main focus on the cold work
steel 1.2380. This steel is widely used for wear resistant applications and usually
consolidated by hot isostatic pressing [9]. After heat treatment 1.2380 presents small
carbides in a martensitic matrix. To investigate the influence of the powder size on the EDS
consolidation behaviour, different grain size fractions were sintered. Furthermore the steel
powders 1.4404 and 1.8905, with grain size as-atomized, were sintered. They where chosen
to investigate the consolidation behaviour as layer assembly or intermixture of dissimilar
materials. The investigation of the movement of different alloying elements (C, Cr, V) allow
conclusions of the diffusion process during sintering. The chemical compositions of the
materials investigated are shown in Table 1.

Table 1: Materials and their composition in wt %
C Cr V Mo Ni Mn
1.2380 2,30 13,00 4,00 1,00 0 0,30
1.4404 0,02 18,00 0 2,30 12,00 2,00
1.8905 0,13 0,20 0 0 1,00 0,95

Experimental Procedure
The setup for EDS experiments is given in a schematic drawing in Fig. 1 a). A non
conductive forming die made from silicon nitride is filled with powder. At top and bottom
highly conductive copper pressure dies allow to apply a compressive force (via hydraulic)
and an electrical current of some hundred thousand amperes. The current is discharged by
capacitors and converted by transformers. In Fig. 1 b) shows the pressure and current
regime versus time. At first the powder is compressed with constant pressure that is limited
by the pressure strength of the copper dies. Then the DC current is discharged. The
temperature can be observed on a macroscopic and microscopic scale. The samples for the
results presented in this paper were produced using an industrial Capacitor Discharge
Welding unit provided by Manfred Schlemmer GmbH. 7500 F capacitors provide 40000 Ws
at 24 V output voltage. Energy and pressure can be varied in 1 % and 1 kN steps. The
forming die used for the experiments has a 16 mm inner diameter. For all experiments 10 g
of powder were used for the consolidation experiments resulting in a sample height of about
5 mm. The whole process takes place at atmospheric condition without any shielding gas or
vacuum atmosphere.

Figure 1: a) Schematic assembly of the EDS machine, b) Characteristics of EDS process parameters

Experimental Methods
Compacted samples were cut, embedded in resin and polished using diamond and
aluminium oxide suspension up to 0.25 m grain size. The polished samples were etched
with V2A pickle to develop the microstructure. The light optical photographs were taken
using an Olympus BX-60M microscope. For higher magnifications a Scanning Electron
Microscope LEO 1530 VP is used. X-Ray diffraction measurements are carried out using a
Siemens D-500. Vickers Microhardness is measured with a Paar Physica MHT-10 micro
hardness tester under a load of 490 mN.

Results and Discussion
Process Parameters and Consolidation Behaviour
The initial experiments were designed to determine the optimal densification parameters. For
a better detection of microstructural changes the powder grain size was chosen to be lager
than 250 m. In Fig. 2 it is shown that the density of the microstructure correlates with the
increase of the discharge energy reaching full dense microstructure at 70 % of discharge
energy. Additionally the discharge current and the displacement of the pressure die into the
forming die disclose a nearly linear performance with increasing discharge energy. Further
increase of the discharge energy or multiple discharge pulses did not change the results at
that stage of experiments. As seen in Fig. 2 the interparticle pores between the powder
grains are subsequently reduced with increased discharge energy. As a main densification
process neck formation is observed. With increasing discharge energy the necks become
larger until all the pores are filled up. Although a constant pressure is applied the spherical
shapes of the powder grains remain unchanged, which indicates that no plastic deformation
occurs to the powder grains by the applied pressure under the conditions investigated. Fig. 3
shows the microstructures near the surface of the specimen using different grain size
fractions at 50 and 70 % densification energy. The low discharge energy leads to an
enhanced densification especially near the sample surface. This tendency is observed for all
powder grain sizes. For that reason a discharge energy of 70 % was chosen for the following
experiments. The higher porosity near the surface indicates that densification proceeds from
the core to the case of the sample. The same behaviour is described in former literature [10]
as a result of the electromagnetic pinch effect. It is specified that the pinch effect aids to fill
the adjacent interparticle pores which has the highest influence in the middle of the sample
and decreases towards the surface. This was observed on materials consolidated by using a
EDS technique where the current is directly discharged from the capacitors, which leads to a
high discharge voltage. In contrast to this, the setup used for these investigations generates
low discharge voltages but significantly higher current, so that joule heating is the initial
activator for densification. Once the discharge energy is sufficiently high the surface near
microstructure will be densified by filling up the interparticle pores with molten material which
has been produced at the surface of the powder grains.

Figure 2: Correlation of discharged energy, electrical current, indentation of the die and the sample
density (1.2380)

Figure 3: Coherence of
discharged energy, grain size
and micro structure at the
surface (1.2380)

Microstructure formation and evolution
The dense consolidated microstructure of 1.2380 cold work steel is shown in a small
magnification in Fig. 4a). Fig. 4b) shows the region between two consolidated grains (former
interparticle pore). On the left and right side the microstructure of the cold work steel in
atomized condition is present. It consists of a eutectic carbide network surrounding metal
cells. Depending on the high solidification rate of a powder grain during atomization the
network is fairly small. In the middle of Fig. 4b) the arrow points to an area where a liquid
phase has been present during EDS showing a much finer carbide network. This is due to a
smaller volume of liquid metal in EDS than in atomization and a higher cooling rate brought
about by a self-quenching effect. X-Ray diffraction measurements were carried out on
1.2380 cold work steel in order to clarify if a microstructural change other than a finer shape
of the carbide network occurs due to consolidation by EDS. For reasons of comparison
powders in the state of as-atomized and heat-treated (HT=hardening and tempering) are
investigated after consolidation with EDS. The results are shown in Fig. 5 where as-atomized
powder (fast solidification from liquid phase) (1) presents MC and M
7
C
3
carbides, numerous
austenite and very few martensite. After heat-treatment (2) all retained austenite has
transformed to martensite and MC plus M
7
C
3
carbides. During the EDS process the material
that filled up the interparticle pores experienced fast melting and solidification. Therefore it
was assumed that the phases should be equal to the state as-atomized. This is generally the
case as proven by a consolidated sample (3) consisting of as-atomized powder. By
comparing the heat-treated powder before (2) and after EDS consolidation (4) a clear
abnormality is shown. Again austenite peaks are present, which proves that changes in the
microstructure occur during EDS. These results correspond with the change of
microstructure shown in Fig. 4.

Figure 4: a) Powder
grains with filled up
interparticle pores b)
Change of microstructure
due to liquid phases
(1.2380)

Figure 5: X-Ray
diffraction of powder as-
atomized (1) powder
heat-treated (2) EDS
consolidated powder as-
atomized (3) and EDS
consolidated powder
heat-treated (4); all
1.2380

Microhardness Measurements
The atomized powder grains and the consolidated microstructures were characterised by
microhardness measurements (Table 2). The microhardness of 1.2380 powder is the lowest
(517 HV 0.05) principally due to the high amount of retained austenite that is present after
atomisation. The filled-up interparticle pores show a slight increase of the microhardness,
which might result from the finer carbide network. After hardening and tempering the
microhardness increases up to 840 HV 0.05 because of the martensitic metal matrix without
retained austenite. If this heat treated powder is consolidated with comparable EDS
parameters, a density of only 95 % is reached. The hardness measured at the filled-up
interparticle pores of the consolidated heat-treated powder increases to 978 HV 0.05. The
reason for this has not been fully investigated up to now. One possible explanation might be
the difference in thermal conductivity. The powder in the hardened and tempered state is
fully martensitic and due to that it shows a higher thermal conductivity than the powder in the
as atomized state which is predominantly austenitic. This leads to a higher cooling rate of the
liquid interparticle material due to a stronger self-quenching effect.

Table 2: Microhardness of embedded powder (as-atomized and heat-treated), filled-up interparticle
pores (after EDS using as-atomized and heat-treated powder)
Hardness HV 0.05
1.2380 Powder 517
1.2380 Filled-Up Interparticle Pores 568
1.2380 Powder HT 840
1.2380 HT Filled-Up Interparticle Pores 978

Bonding and Diffusion Behaviour
Fig. 6 shows both the consolidated microstructures and the fracture surfaces of a
1.2380/1.8905 intermixture and a pure 1.2380 sample. Because of the high density achieved
(> 99 %) a homogeneous fracture surface is expected, but can not be found at the
intermixture of 1.2380/1.8905. Due to the enormous differences of the material strength, the
fracture surface shows a ductile fracture behaviour of 1.8905 and no failure of 1.2380
powder grains. It can also be observed that no interconnection of the different powder
materials has developed during EDS consolidation. The fracture surface of pure 1.2380
shows a transcrystalline fracture as it is demanded from satisfactory consolidated powders.
In order to investigate the diffusion and bonding behaviour, especially of the powder
intermixtures, EDX mappings are carried out. One example of consolidated intermixture from
1.8905/1.4404 is shown in Fig. 7. The arrow in Fig. 7 points at a filled-up interparticle pore.
As shown in the element allocations, a mixture of elements from the powder components is
only present in areas that were liquid during consolidation. Diffusion of elements along grain
boundaries seems not to be noticeable with EDX measurement. Future measurements with
high resolution microscopy will be carried out in order to investigate the diffusion behaviour
and element distribution.

Figure 6: Comparison of the
microstructure and the fracture
surface of 1.2380/1.8905
intermixture and a sample of
pure 1.2380


Figure 7: EDX- Mapping of
1.4404 (top) and 1.8905
(bottom) intermixture

Conclusions
It has been shown that Electro-Discharge-Sintering is a suitable technique to consolidate
cold work steel powder within a very short time (5 10 ms). Optimal parameters were found
to sinter dense samples of the steel powders used. Independent from powder grain size,
densification increases with increasing discharge energy from the core towards the surface
of the samples. It can be assumed that the interparticle pores are filled up with molten
material from the powder grain boundaries during the consolidation process. The material
solidifies rapidly due to self-quenching and shows a finer carbide network. Using X-Ray
diffraction it was observed that changes in the microstructure occur after consolidation.
Samples showed retained austenite after consolidation, whereas none was present in the
initial heat-treated powders. Although the microstructure of the powder intermixtures was
dense, fracture surfaces showed weak adhesion in between the powder grains for the
dissimilar powders as achieved.

Acknowledgements
The authors thank for the financial support of the German Federal Ministry for Education and
Research (BMBF), sponsoring the consortium project 03X3515.

References
1. Alp, T., Al-Hassani, S.T.S., Johnson, W., J. Mat. Techn. 107 (1985) 186-194.
2. Clyens, S., Al-Hassani, S.T.S., Johnson, W., Int. J. Mech. Sci. 18 1 (1976) 37-44.
3. Qiu, J., Shibata, T., Rock, C., Okazaki, K., Mat. Trans., JIM 38 3 (1997) 226 - 231
4. Grigoriev, E.G., Rosliakov, A.V., J. Mat. Proc. Techn. 191 (2007) 182-184
5. Fais, A., Maizza, G., J. Mat. Proc. Techn. 202 (2008) 70-75
6. Rajagopalan, P.K., Desai, S.V., Kalghatgi, R.S., Krishnan, T.S., Bose, D.K., Mat. Sci.Eng.,
A280 (2000) 289-293
7. Yavuz, N., Etemoglu, A.B., Gullu, E., Can, M., Metall, 58 (2004) 295-300
8. Ryabinina, O.N., Raichenko, A.I., Burenkov, G.L., Translated from Poroshkovaya
Metallurgiya, 11 167 (1976) 16-21
9. Berns, H., Theisen, W., Ferrous Materials Steel and cast iron, Springer Verlag Berlin
(2008) 242-243
10. Alp, T., Can, M., Al-Hassani, S.T.S., Powder Met. 30 1 (1987) 29-36
Germany

Induction Sintering of Fe-2Cu PM Compacts

Uur AVDAR*, Enver ATK*
* Celal Bayar University, Department of Mechanical Engineering, Muradiye Campus
Manisa/TURKEY

Abstract:
Fe powders mixed with 2% Cu by weight were compacted with 600 MPa to form 3-point
bending samples having dimensions of 10x10x55 mm. Compacted powders were sintered by
using medium-low frequency (30 kHz, 12kW) induction energy in conveyor system. Sintering
process was completed in several sintering durations from 400 to 1300 seconds at 1120
o
C
under atmospheric environment (open air). Micro structural and mechanical properties of
sintered samples were investigated. Maximum stress for 3-point bending was achieved at
1000 seconds of sintering duration at 1120 C. The mechanical properties of induction
sintered samples were compared with conventionally sintered samples at 1120 C for 30
minutes of sintering duration under argon atmosphere. The maximum stress values for
induction sintered sample were nearly similar to conventionally sintered samples.
Keywords: Induction, Sintering, Powder Metal.
1. INTRODUCTION
The sintering process was generally done in the sintering furnaces. The sintering furnace
controls heat and time during the sintering loop. Additionally, it keeps the atmosphere and
provides the possibility of heat treatment after the sintering. The sintering process was done
in a batch furnace or continuous furnace [1,2].
The sintering process of an iron-based powder metal sample in a classical furnace was
carried out at 1120C in 30 minutes. The reason for the induction sintering was to decrease
sintering time.
As an alternative to classical sintering methods, novel rapid sintering processes have started
being used. These are; induction heating, microwave heating, plasma heating and laser
heating [3].
Based on the literature induction sintering was studied in 1980s by W. Hermel et al. about
induction sintering fundamentals, applications [23, 24] and A. Salak et al. about induction
sintered iron based samples [25]. Sintering by induction was patented in 1988 in the USA
under the name of Induction sintering process and apparatus [4]. The most significant
studies done on sintering by induction after that date have been done under the name of high
frequency induction heat sintering which are called HFIHS. Parallel to our study, M.
Nakamura et al. [5] took out the patent of rapid sintering of iron and steel by induction heat
under hydrogen and nitrogen gases in 2003.

HFIHS (High Frequency Induction Heat Sintering) [7 - 18]:
HFIHS method is a method of rapid sintering like SPS (Spark Plasma Sintering) method. As
compared to SPS, HFIHS is a new method of sintering [6].
This method is generally applied to ceramics, composites and bio materials. For example;
are used in materials such as WC-Co [7], uranium oxide (UO) [8], composite like WC-Ni,
Fe, Co [9, 10], WC-Mo
2
C [11], WC-TiC [12], WC-8%Co [13], 8YSZ-Fe
2
O
3
[14], TiB
2
-WB
2
[15],
NbSi
2
-Si
3
N
4
[16], bio ceramics like Hydroxyapatite (Hap) [17] and Al
2
O
3
-(ZrO
2
+3%Y
2
O
3
)
[18]. When the powders which are generally placed between Al
2
O
3
double axis blocks in the
plate are pressed as double axis, they are sintered by high frequency induction coil.
Cavdar et al. studied iron and iron based powder metals for sintering samples with low-
medium frequency induction sintering [19, 20, 21].

2. MATERIALS AND METHOD
The samples used in tests were produced as Fe + 2% Cu (weight) of ratio %2 copper (0.74
gr.) and iron powder (36.26 gr.). Zinc stearate of the ratio 0.8% (0.296 gr.) by weight was
used as the lubricant.
The powder metal used in the samples are ASC10029 (The characteristics are shown at
Table 1) produced by Hgens. The powder metal was thoroughly mixed for 20 minutes in
25 cycle / min to have a homogeneous mixture. The powder metal samples having the size
of 10x10x55 mm and weight of 37 g, were produced by pressing at a pressure of 600 MPa
using one axis press.
Table 1. The Characteristics of Iron Powder [26].
SPECIFICATION RESULT %
MN MAX
GRANULOMETRY
B.S.S. MESH MICRONS
85 180 0.0 0.0
100 150 0.5 0.3
150 106 12.5
200 74 19.1
300 53 19.8
350 45 11.6
-350 45 30.0 40.0 36.7
PHYSICAL PROPERTIES
Apparent Density g/cc 2.55 2.75 2.69
Flow Secs 30 26
CHEMICAL ANALYSIS
Copper % 99.00 99.74
Oxygen % 0.15 0.08

All sintering temperatures were applied at 1120C. A group of pressed metal powder
samples were sintered in classical sintering furnace in argon gas environment and
atmospheric environment (open air) in 30 minutes. The other group was sintered by 12 kW,
30 kHz (medium-low frequency) induction generator and atmospheric environment (open
air). Sintering process was completed in several sintering durations from 400 to 1300
seconds. The samples temperature was measured by infrared thermometer during the
induction sintering process. While samples temperature was measured by thermocouple,
temperature was affected by magnetic flux. Therefore temperature was measured by infrared
pyrometer.

Figure 1. Induction Machine and Sintering Process (Celal Bayar University, research
laboratory, Muradiye Campus-Manisa-Turkey)
3. RESULTS AND DISCUSSION
5 different samples were used for all induction sintering experiments, average of the
experiments results were shown in the figures. According to the classic sintering experiments
results were used in average of 3 samples results in the figures.
3.1. 3 point bending results for 2% Cu based iron powder samples.
3 point bending tests was performed in Autograph Shimadzu AG-IS 100 kN universal test
machine to compare the sintered samples by the classical sintering process and sintered
samples by induction sintering. Bending strength values of powder metal samples are shown
Figure 2. The percentage of the break strain values of Powder metal samples is shown in
Figure 3.

Figure 2. Bending strength of %2 Cu based iron samples.

Figure 3. The percentage of the break strain values of %2 Cu based iron samples.
3.2. Vickers hardness results
Micro hardness measurement was performed with Future Tech FM-700 for using 10 sec., 50
grf .The results are given in Table 2.

Table 2. Average hardness values of the Powder metal samples.
Samples Names Ratio hardness values (HV)
Classical Sintering (without argon gases) 136
Classical Sintering (with argon gases) 142
400 Sec. Induction sintering (without argon gases) 152

3.3. Micro structure photographs of the samples.
Induction sintering samples micro structure photos are shown in figure 4.

A B

C D
Figure 4. A) 400 second induction sintering sample (sintering in open air), B) 700 second induction
sintering sample (sintering in open air), C) 1000 second induction sintering sample (sintering in open
air), D) 1300 second induction sintering sample (sintering in open air) photo

4. CONCLUSION
Iron based samples included %2 Cu were sintered with induction sintering method.
During the active induction heating system, the more sample strength is increased the more sintering
time is. The maximum stress value is observed in 1000 seconds induction sintering process. When
sintering time passes 1000 second, maximum stress value is decreased.
A maximum stress value of classic sintering in atmospheric environment (open air) is reached in 700
second induction sintering process.
A maximum stress value of classic sintering in argon environment is reached in 1000 second induction
sintering process.
The percentage of the highest break strain value obtained induction sintering process in atmospheric
environment (open air) is reached in consequence of 1300 second induction sintering. With this value,
it can be achieved only to the percentage of the break strain value obtained by classical sintering in
atmospheric environment (open air).
According to the Average Vickers hardness values of induction sintering are harder than classic
sintering in both argon and atmosphere environments hardness. The more induction sintering time is
increased, the hardness value is increased.
According to the mechanical properties expected from the sample induction sintering parameters must
be able to be optimized.
Induction sintering parameters could be optimized according to mechanical properties of samples.

SPECIAL THANKS
To Prof. Dr. Cevdet Meri (cmeric@fatih.edu.tr) and Prof. Dr. Halun KARACA
(haldun.karaca@deu.edu.tr ) for their academic helps,
To Tozmetal Inc. (http://www.tozmetal.com/english.htm ), Ayta Ata (Aytac@tozmetal.com) for their
help as the provider of the metal powders,

4. REFERENCES
1. German, R.M., Powder Metallurgy Science, MPIF, USA, 1984.
2. Randall M. German; translator editors: S. Sarta, M. Trker, N. Durlu, Toz Metalrjisi ve
Paralkl Malzemeler lemleri TTMD, Ankara, Trkiye, 2007.
3. Randal M. German, Sintering theory and practice The Pennsylvania State University Park,
Pennsylvania, A willey interscience publication, Jon Wiley & Sons, INC. 1996 USA, pp. 313-
362, 373-400, 403-420.
4. United States Patent, Patent Number: 4,720,615. Inventor: Jerry R. Dunn, Boaz, Ala;
Assignee: Tocco, Inc., Boaz, Ala; Appl. No: 770,768; Date of patent: Jan.19.1988, USA.
5. M. Nakamura et al., I. Japan Soc., Iron and steel rapid sintering of steels by induction
heating in hydrogen-nitrogen Patents, Metal-powder.net, p. 33, March 2003, MPR.
6. H.C. Kim, D.K. Kim, K.D.Woo, I.Y.Ko, I.J. Shon; Consolidation of binderless WC-TiC by high
frequency induction heating sintering, Elsevier, International Journal of Refractory Maters &
Hard Materials 26 (2008) 48-54, Republic of Korea, accepted 18 January 2007.
7. Hwan-Cheol Kim, Dong-Young Oh, In-Jin Shon, Sintering of nanophase WC - 15 vol. %Co
hard metals by rapid sintering process, Elsevier, Refractory Metals & Hard Materials 22(2004)
197-203, South Korea, accepted 22 june 2004.
8. J.H. Yang, Y.W. Kim, J.H. Kim, D.J. Kim, K.W.Kang, Y.W. Rhee, K.S. Kim, K.W. Song;
Pressure less rapid sintering of UO
2
assisted by high frequency induction heating process,
The American Ceramic Society, 91 [10] 3202-3206, 2008.
9. In-Jin Shon, In-Kyoon Jeong, In-Yong Ko, Jung-Mann Doh, Kee-Do Woo; Sintering behavior
and mechanical properties of WC-10Co, WC-10Ni and WC10Fe hard materials produced by
high-frequency induction heated sintering Elsevier, Ceramic International xxx(2008) xxx-xxx,
Republic of Korea, accepted 3 November 2007.
10. H.C. Kim, I. J. Shon, Z. A. Munir, Rapid sintering of ultra - fine WC - 10 wt % Co by high -
frequency induction heating, Elsevier, International Journal of Refractory metals & Hard
Materials 24 (2006) 427-431, accepted 5 July 2005.
11. Hwan-Cheol Kim, Hyun-Kuk Park, In-Kyonn Jeong, In-Yong Ko, In-Jin Shon; Sintering of
binderless WC-Mo
2
C hard materials by rapid sintering process, Elsevier, Ceramic
International xxx (2007) xxx-xxx, Republic of Korea, accepted 27 march 2007.
12. H.C. Kim, D.K. Kim, K.D.Woo, I.Y.Ko, I.J. Shon; Consolidation of binderless WC-TiC by high
frequency induction heating sintering, Elsevier, International Journal of Refractory Maters &
Hard Materials 26 (2008) 48-54, Republic of Korea, accepted 18 January 2007.
13. H.C.Kim, I.K.Jeong, I.J.Shon, I.Y. Ko, J.M.Doh; Fabrication of WC-8wt.%Co hard materials
by two rapid sintering processes Elsevier, International Journal of Refractory Metals & Hard
Materials 25 (2007) 336-340, Republic of Korea, accepted 4 September 2006.
14. I.J. Shon, I. K. Jeong, J.H. Park, B.R. Kim, K.T. Lee; Effect of Fe
2
O
3
addition on
consolidation and properties of 8 mol% yttria- stabilized zirconia by high-frequency induction
heated sintering (HFIHS), Elsevier, Ceramics International xxx(2008)xxx-xxx, accepted 6
November 2007.
15. M.Shibuya, M.Ohanagi Effect of nickel boride additive on simultaneous densification and
phase decomposition of TiB
2
-WB
2
solid solutions by pressure less sintering using induction
heating, Elsevier, Journal of the European Ceramic Society 27 (2007) 301-306, Japan,
accepted 6 May 2006.
16. H.K. Park, I.J. Shon, J.K. Yoon, J.M. Doh, I.Y. Ko, Z.A. Munir; Simultaneous synthesis and
consolidation of nanostructured NbSi
2
-Si
3
N
4
composite from mechanically activated powders
by high frequency induction-heated combustion Elsevier, Journal of Alloys and Compounds
461 (2008) 560-564, Available online 25 June 2007.
17. K.A. Khalil, S.W. Kim, N.Darmaraj, K.W.Kim, H.Y. Kim; Novel mechanism to improve
toughness of the hyroxyapatite bio ceramics using high-frequency induction heat sintering,
Elsevier, Journal of Materials Processing Technology 187-188 (2007) 417-420.
18. K.A. Khalil, S.W.Kim; Mechanical wet-milling and subsequent consolidation of ultra-fine Al
2
O
3
-
(ZrO
2
+Y
2
O
3
) bio ceramics by using high-frequency induction heat sintering, Science Press,
Trans. Nonferrous Met. Soc. China 17(2007) 21-26, accepted 13 September 2006.
19. U. Cavdar, E. Atik; Sintering with induction, International Powder metal Congress &
Exhibition, Rostengarten congress center, Euro PM2008, Mannheim/Germany, September
29th - 1st October, Vol.3, p.p: 33-38.
20. U. Cavdar, E. Atik; The effects of powders size in Induction sintering, in Turkish, 12.
International Materials Symposium, October 15-17 2008, Pamukkale University Congress &
Cultural Center, Denizli, Turkey, vol.2 p.p: 1286-1290.
21. U. Cavdar, E. Atik; Induction sintering of %3 cu contented iron based powder metal parts5th
International Powder Metallurgy Conference Turkey, TOBB University, Ankara/Turkey, 8-12
October 2008.
22. U. Cavdar, E. Atik; Iron based powder bushings sintering via induction generator in Turkish,
5th International Powder Metallurgy Conference Turkey, TOBB University, Ankara/Turkey, 8-
12 October 2008.
23. W. Hermel, W. Foerster, G. Leitner, K. Voigt, Short time induction sintering of TiC- and WC-
based hard metals, Metal Powder Report, Shrewsbury, England, p 10, v 1. 110. 13.
24. W. Hermel, G. Leitner, and R. Krumphold, "Review of Induction Sintering: Fundamentals and
Applications," Powder Metal., 23 [3] 1305 (1980).
25. A. Salak, G. Leitner, W. Hermel. Properties of Induction-sintered Fe-Mn-C and Fe-Mn-Cu-C
steel in the sintered and forged states. Powder Metallurgy International. 1981. Vol 13. N 1.
Pp:21-4-45.
26. http://www.hoganas.com/
Euro PM 2009 Field Assisted Sintering
Manuscript refereed by Dr J rgen Schmidt, Fraunhofer IFAM, Germany

Spark Plasma Sintering Behavior of Ceramic Biocomposites Based
on Hydroxyapatite Nanopowders

Gingu O
1
., Lupu N
2
., Tanasescu S
3
., Rotaru P
1
., Harabor A
1
., Mangra M.
1
, Ciupitu I.
1
,
Sima G.
1
, Olei A.
1
, Bucse I.
1

1
University of Craiova, 13 A.I.Cuza, 200585, Craiova, Romania
2
Institute of Technical Physics, Blv. Mangeron, Iasi, Romania
3
Institute of Chemical Physics I.G. Murgulescu, Splaiul Independentei, Bucharest, Romania

Abstract: The advanced materials used for biomedical applications are characterized by the
nanostructure feature. The implant biocomposites used for bone tissue reconstruction are
recommended to be nanostructured improving the functionalisation of the interface between
the implant and the bone tissue. The paper presents the experimental results concerning the
elaboration of biocomposite materials processed by Spark Plasma Sintering (SPS).
Micrometric Ti powders and nanometric hydroxyapatite powders have been mixed and cold
compacted at 120 MPa. Nanostructured sintered biocomposites are processed by SPS in
vacuum. The elaborated biocomposites are analyzed from the point of view of structural
(SEM) and mechanical (microhardness) properties. The experimental data are compared to
the similar ones obtained for the same materials processed by other technology routes.

1. Introduction
The hydroxyapatite, with chemical formula:Ca
5
(PO
4
)
3
(OH) (acronym HA), is one of
the well-known biomaterial resembling to human hard tissue [1], being used for bone
repairing by implants and grafts. But the excellent HAs biocompatibility is counterbalanced
by its poor mechanical properties: elastic modulus (40-117) GPa, compressive strength
294MPa, bending strength 147 MPa, Vickers hardness 3,43 GPa. These properties are
different from the hard tissues ones in comparison with the main mechanical properties of the
cortical bone (the denser one among the bones) which are: elastic modulus (4-27) GPa,
compressive strength (100-160) MPa, tensile strength (45-175) MPa, shear modulus (2-9)
GPa, shear strength (50-70) MPa [2].
Because of this behavior, HA is usually used as a matrix for biocomposite materials.
The reinforcing component could be:
polymeric, i.e. collagen type I (a natural polymer found in bones) providing the following
advantages: an accelerated osteogenesis process, high stability, slow 3D swelling and
high mechanical wet properties [3], [4], [5];
ceramic, i.e. zirconia, providing high fatigue resistance at high pressure [6] as well as high
bending strength, microhardness and Young modulus [7];
metallic alloys based on Ti, Ni, Co-Cr-Mo, stainless steel, amalgam [8] also for increasing
the mechanical properties [9];
non-metallic, as carbon nanotubes (SWCNT or MWCNT) up to 7% vol. with advantageous
effects regarding the improvement of the fracture toughness and flexural strength by 50%
and 28% respectively because of MWCNT capacity to absorb some fracture energy and
prevent propagation of cracks [10].

In the same time, due to recent advances in nanostructured materials synthesis and
characterization techniques, metallic and ceramic nanostructured biomaterials are the key to
the future of the biomedical industry [11]. The nanostructure characteristic of these
biocomposites is necessary to be fulfilled because of the superior properties provided in
comparison with the micro-structured similar materials [12]. Thus, HA nanocrystalline paste
is used as autogeneous bone graft with a good success rate in maxillofacial and stomatology
surgery and orthopaedics, providing an accelerated fracture healing and bone density
increase at a high degree of tolerance [13, 14]. Despite of this, HA nanocrystalline is
recommended always to be combined with some form of stable osteosynthesis, preferably
with an angularly stable plate, due to its lack of dimensional stability [15].
In this paper the research focuses the PM processing and characterisation of the
biocomposite material based on HA nanopowders reinforced by micrometric Ti powder
particles. The Spark Plasma Sintering (SPS) process is applied to process the ceramic
matrix of HA nanopowders that will provide the main properties of the composite from the
point of view of the biocompatibility with the human hard tissues as well as of the
nanostructured feature. The experimental results are compared with similar materials
processed by other technological routes.

2. Materials and experimental procedure
The raw materials used for the research are: HA nanopowders having less than 200
nm particle size (BET), type Aldrich, and micrometric Ti powder particles (~ 100 microns
particle size).
The PM route to process biocomposites type HA/Ti includes the following steps:
2.1 the calcination treatment of HA powder particles (in air, at 900
0
C for 1 hour) in order to
release the H2O content. Beyond this main effect, an undesirable one is the
agglomeration of the nanopowders. Fig. 1 presents the particle size distribution of HA
nanopowders before (a) and after (b) the calcination treatment. The analysis has been
developed on BROOKHAVEN 90 PLUS BI-MAS equipment, which measures by laser
scattering the powders particle size between 2 nm 2 m.

Fig.1a Particle size distribution of HA nanopowders before calcination (initial)


Fig.1b (Continuation) Particle size distribution of HA nanopowders

2.2 the wet mixing of HA nanopowders and Ti micrometric ones as 3:1 ratio in a planetary
ball mill type Fritsch / Pulverisette 6. The mixing parameters are: balls/powder mixture
ratio = 2:1; rotation = 200 rpm, mixing atmosphere = air; mixing medium = wet (ethanol).
The mixing stage is a critical one in order to provide the dispersion of calcinated HA
nanopowders that show the conglomeration behavior during the calcination treatment,
Fig.1.b. The effect of the mixing process is visible from the optic and electronic
microscopy images (see below).
2.3 the mixture drying in air at 200
0
C, overnight, for alcohol evaporation;
2.4 the dry mixing of HA/Ti mixture for 1 minute in Fritsch / Pulverisette 6, balls/powder
mixture ratio = 2:1; rotation = 200 rpm, mixing atmosphere = air;
2.5 the vacuum SPS process for biocomposites sintering. The sintering parameters are
presented in Tab. 1 and one of the thermal cycles is presented in Fig.2.

Tab.1 SPS parameters for HA/Ti processing
Nr.crt. Sample type Sintering conditions
atmosphere temperature [
0
C] time [min.]
1
HA/Ti (3/1) vacuum
1.000 10
2 20
3 1.100 10
4 20


Fig.2 Thermal cycle for HA/Ti biocomposite SPS-ed at 1000
0
C/10 minutes

After biocomposite processing, optical and electronic microscopy analysis underline
the morphological aspects of the sintered samples. Also, microhardness tests provide
information regarding this mechanical characteristic of HA/Ti processed by SPS.

3. Experimental results and discussions

The aim of calcination process is to decrease as much as possible the water content
from the HA nanopowders in order to improve its behavior during sintering process that can
be negatively influenced by H2O content. Because of the hydroxyapatites hygroscopicity, it
shows some agglomerates even in initial stage (< 200 nm, as elaborated powder), the mean
diameter of HA nanopowders being 401,8 nm (Fig. 1a). Once the calcination treatment
develops, the agglomerates grow up to clusters of 992,7 nm (Fig. 1b).
Thus, the mixing process plays an important role in dispersing of HA nanopowders
clusters among the micrometric Ti powder particles. The homogeneous structure of HA/Ti
powder mixture can be observed in Fig.2 which presents the biocomposite processed by
SPS at 1000
0
C/10 min.
The samples porosity is quite low and Ti reinforcing component (grey spots) is
homogeneously dispersed in HA matrix (white spots), Fig. 2.
From the picture above it can be drawn that SPS process is able to keep the initial HA
nanopowders in the submicronic range. In comparison, Fig. 4 presents the microstructure of
the same biocomposite processed by classic sintering at 1200
0
C/2 hours.


Fig.2 SEM (BSE) for HA/Ti processed by SPS at 1000
0
C/10 min, providing a homogeneous
structure regarding the reinforcing component (Ti) in the ceramic matrix (HA)

Fig.3 SEM (BSE) of HA/Ti elaborated at 1000
0
C/10 min showing the presence of submicronic
HA powder particles after SPS processing

HA matrix
Ti reinforcement
Submicronic HA
powder particles
pore

Fig. 4 SEM (BSE) of HA/Ti processed by classic sintering at 1200
0
C/2h

The next research will provide more information regarding the effect of SPS on wear
behavior as well as the biocompatibility tests results, in comparison with similar materials
processed by other PM technologies.

4. Conclusions
PM biocomposite materials HA/Ti are able to be processed by SPS technique,
especially in the case of HA nanostructured matrix reinforced by metallic (Ti) micrometric
powders. The main advantages of spark plasma sintering consist in low porosity and
nanostructured feature of the biocomposite structure that could have a great impact on its
mechanical and biocompatible properties.

Acknowledgements
The authors would like to thank to Prof. Mario Rosso and his collaborators from
University Politecnico of Turin (www.polito.it), DISMIC (Department of Materials Science and
Chemical Engineering), Italy, supporting the SEM analysis.

References
1. L.L. Hench, Bioceramics Journal of the American Ceramic Society 81 (1998) 1705-1728
2. www.eng.tau.ac.il/~gefen/BB_lec1.pdf
3. J. Xie, M.J. Baumann, L.R. McCabe, Osteoblasts respond to hydroxyapatite surfaces with
intermediate changes in gene expression, J. Biomed.Mater.Res. A 71 (2004) 108-117
4. A.C. Lawson, Y.T. Czernuszka, Collagen-calcium phosphate composites, Proc. Int. Mech. Eng.
Part H, J. Eng. Med. 212 (1998) 413-425
5. A. Scabbia, L. Trombelli, A comparative study on the use of HA/collagen/chrondroitin sulphate
biomaterial (Biostite&reg.) in the treatment of deep intraosseous defects, J. Clin. Periodontal 31
(2004) 348-355
6. E. Chang, W.J. Chang, B.C. Wang, C.Y. Yang, Journal of Materials Science Medicine 8 (1997)
193-200
7. X. Miao, Y. Chen, H. Guo, K.A. Khor, Ceramics International 30 (2004) 1793-1796
8. M. Niinomi, Science and technology for Advanced Materials 4 (2003) 445-454
9. R. McPherson et al., J.Mater.Sci.Med. 1995, 327-334
10. Y.H. Meng et al., J.Mater.Sci.Mater.Med., 2008, 19, 75-81
11. K. Jurczyk et al., European J.Med.Res., 11, Suppl.II, 2006, 133], [B.C. Ward, T.J. Webster,
Mat.Sc.Eng., C27, 2007, 575
12. K. Niespodziana et al., Rev. Adv. Mater. Sci. 18 (2008), 236-240], [Y.H. Meng et al.,
J.Mater.Sci.Med. (2008), 19:75-81
13. F.D. Burstein, Cleft Palate Craniofac.J., 37, 2000, 1],
14. R. Schnettler et al., Eur.J.Trauma, 30, (4), 2004, 219
15. F.-X. Huber et al., J.Mater.Sci.Mater.Med., 2008, 19, 33-38
Germany

Consolidation Process of Tungsten Carbide - Cobalt Powder by
Electro-Discharge Compaction

E.G. Grigoryev
Moscow Engineering Physics Institute
115409, Kashirskoe sh., 31, Moscow, Russia

ABSTRACT:

The consolidation of high strength structure of WC-Co composite material is investigated and
optimal operating parameters are defined. Tungsten carbide cobalt composites were
produced by the method of high voltage electrical discharge together with application of
mechanical pressure to powder compact. Pulse current parameters were measured by
Rogovsky coil. The temperature evolution during electro-discharge compaction process was
measured by means of thermocouple method. It was studied densification process of powder
material during electro-discharge compaction by means of high-velocity filming. The
microstructure of WC Co cemented carbides was investigated by optical microscopy.
Densification to near theoretical density in a relatively short time can be accomplished with
insignificant change in grain size. We have found that the powder densification process has
wave nature in electro-discharge compaction. I defined the wave front velocity of
densification process and the pressure amplitude in wave front to be subject to parameters of
electro-discharge compaction.

INTRODUCTION

The methods of consolidation of powder materials, based on various techniques of
transmission of electric current pulses through a powder under mechanical pressure, are
widely studied in many research centers. Nowadays these methods are especially important
because it makes possible to obtain the bulk nanostructure materials. These methods
include field-assisted sintering technique (FAST), plasma assisted sintering (PAS), spark
plasma sintering (SPS) [1], electric-discharge sintering [2], electro-discharge compaction
(EDC) [3], etc. The large number of these methods is related to the wide range of variation in
the electrical parameters of the action on a powder. The efficiency of electric-pulse methods
is determined by the multifactor effect on consolidated materials [4]. To obtain materials with
required properties, one has to know the macroscopic processes occurring in the bulk of a
consolidated sample. For example, the kinetics of consolidation of powder materials in these
methods is significantly different, and their duration changes from several tens of minutes for
electric-discharge sintering and spark plasma sintering [1, 2] to several milliseconds for
electro-discharge compaction [3]. In this paper, we report the results of studying the
macroscopic phenomena occurring under electro-discharge compaction of conducting WC
Co powders. Today WCCo composites are extensively used to enhance the wear
resistance of various engineering components, e.g. cutting tools and dies. Cemented
carbides are used throughout industry for high wear, abrasive applications as a result of their
extreme hardness. Apart from the high hardness, WC has other unique properties such as
high melting point, high wear resistance, good thermal shock resistance, thermal conductivity
and good oxidation resistance. Matrices of ductile metals, such as cobalt (Co), greatly
improve its toughness so that brittle fracture can be effectively avoided during operation.
Electro-discharge compaction (EDC) is advanced method which can produce near-net-shape
compacts of high relative density much more rapidly than others conventional processes
such as pressure-less sintering, hot press and HIP. This method has demonstrated a
potential to provide distinct technological and economical benefits in the consolidation of
difficult-to-sinter powders, including short processing times, fewer processing steps,
elimination of the need for sintering aids, and near net shape capabilities.
The principle of EDC is to discharge a high-voltage (up to 30 kV), high-density current (~100
kA/cm
2
) pulse (for less than 300 s) from a capacitor bank through the powders under
external pressure, resulting in a temperature rise of more than 2500 K, instantaneously to
weld grains of powders together. This discharge time is long enough for densification, yet too
short for extensive grain growth as exhibited in conventional technique. In this way full or
near full densification may be achieved with minimal undesirable microstructural changes
due to short consolidation time. WCCo powders could be consolidated into solid bulks by
electro-discharge compaction (EDC) with densities close to theoretical.

EXPERIMENTAL PROCEDURES

Dense WCCo composites were fabricated by an EDC method. The schematic of the
Electro-Discharge Compaction (EDC) system is shown in Fig. 1.

Fig. 1. Schematic of EDC apparatus
1 charging unit, 2 capacitor bank, 3 trigatron switch, 4 control system, 5 electrical
discharge ignition system, 6 pulse electrical discharge registration system, 7 punch -
electrode, 8 die , 9 powder.

Electro-discharge compaction (EDC) apparatus for powder consolidation consists basically of
charging unit (1); a bank of capacitors (2) and trigatron switch (3) to connect a powder
column (9) suddenly across the charged capacitor bank. The capacitor bank consists of thirty
200 F capacitors that can store up to 6 kV. EDC uses the pulse current generated from the
capacitor bank to quickly heat a powder column subjected to constant pressure during the
process. Powder column was a circular cross-section rod of diameter ~ 10 mm and length
from 10 to 15 mm. In this process the WCCo powder is poured into an electrically non-
conducting ceramic die (8). The ceramic die is plugged at two ends with molybdenum
electrodes-punches (7) and an external pressure up to 400 MPa is applied to the powder on
air-operated press. A high voltage capacitor bank is discharged through the powder. We
used commercial WCCo powders (grain size WC < 5m) as starting material for electro-
discharge compaction. Characteristics of the chemical composition of these powders are
resulted in Table 1.

Chemical element WC Co free carbon total oxygen
mass % ~ 80 20 0.101 0.13

Table 1. The chemical composition of WC - Co powders

Table 1 gives the general values before the electric discharge compaction.
The discharge current is measured by a toroidal Rogowsky coil (6) around the powder
column. An oscillograph showing a typical output from the Rogowski coil is shown in Fig. 2.
EDC applies a high-voltage, high-density current pulse to the powder column under external
pressure for a very short period of time. This method uses the passage of the pulse electric
current to provide the resistive heating of the powder by the Joule effect. Joule heating
occurs at the inter-particle contact to instantaneously weld powder particles, resulting in
densification.

Fig. 2. Typical pulse current traces from registration system (Rogowski coil)
(Peak currents:1 50 kA , 2 80 kA , 3 110 kA)

The achieved WC-Co powder compact density as a result of EDC process depends on
applied external air-operated pressure, magnitude and waveform of pulse current that
depends on RLC parameters of the electrical discharge circuit.
Density measurements after EDC process were performed using the Archimedes principle in
distilled water.
The temperature evolution during electro discharge compaction process was measured by
means of thermocouple method. A standard temperature curves at electro discharge
compaction are shown in Figure 3.

Fig. 3. Standard temperature curves of powder column side surface during EDC process
(at constant pressure P = 200 MPa)

Temperature variations 1, 2, 3 correspond of powder column side surface during EDC
process. Temperature dependence 1 was got at amplitude of discharge pulse current equal
95 kA /cm
2
, 2 90 kA /cm
2
, 3 85 kA/cm
2
. The powder material densification process during
of EDC takes place at approximately constant temperature. It follows from the measured
temperature curves (Fig. 3).

RESULTS AND DISCUSSION

The most important factor which determines the success of the EDC process is the
instantaneous current density in the powder column [5]. Fig. 4 shows the effect magnitude of
pulse current on the resultant WCCo composite densities after EDC (external pressure
200 MPa). This experimental dependence has a maximum on fixed peak current density
(Fig.4).

Fig. 4. WCCo compact density dependence from current density

The resultant WCCo composite density increases within the current density region: from 75
kA/cm
2
to 95 kA/cm
2
. The resultant density reaches the maximum value at ~95 kA/cm
2
and
drastically decreases beyond 100 kA/cm
2
.
Representative micrographs of WCCo compact samples were observed on the polished
cross-sections using an optical microscopy. There is a difference between microstructures at
the edge and inside the sample depending on the distribution of magnetic pressure induced
by the pulse current. The distribution of magnetic pressure (pinch effect) is defined by the
distribution of a current density in the powder compact. The distribution of magnetic pressure
has a parabolic profile with maximum inside sample (~ 18 MPa) when a current density has
the homogeneous distribution in the powder column [5]. The magnetic pressure is more
homogeneous in powder compact volume when the skin effect is strong. Fig. 5 shows the
typical WC-Co compact structure with the strong skin effect. In this case, the external
compressed-air pressure was 200 MPa, and the peak pulse current density was 95 kA/cm
2
.


Fig. 5. WC-Co compact structure after EDC with a strong skin effect

We studied the influence of pulse current density and external pressure on the kinetics of
powder material densification at electro discharge compaction. Consolidation of powder
material takes place at constant pressure P during of the electro discharge compaction
process. Dependences of WC-Co compact density versus the time during EDC process are
resulted in Fig. 6.

Fig. 6. Variations of powder column relative density () as a function of time in EDC process
at constant external pressure.

On Fig. 6 curves: 1 (75 kA/cm
2
), 2 (90 kA /cm
2
) were got at constant pressure P = 200 MPa.
The densification time t
d
of powder material is 6 ms < t
d
< 16 ms for all our experiments. The
results of experiments show that motion of punches in the process of the electro discharge
compaction takes place with steady speed. The value of speed depends on amplitude of
pulse current and external pressure. The magnitude of external pressure determines initial
specific resistance of powder column and, accordingly, amount of heat, selected in powder
material. With the increase of pressure the specific resistance of powder column goes down
sharply, that results in the less heating of powder material. The densification of the
compacted material takes place due to an intensive plastic strain which depends on external
pressure P and yield stress of powder matter
(T) (T temperature).
Therefore speed of plastic flow of the compacted powder material, and, consequently, speed
of change of length of the compacted powder column is determined by the temperature at
EDC process. The speed of densification depends on a dimensionless parameter
=
(T)/P.
It should be noted that kinetics of process of densification of powder material at EDC
substantially differs from kinetics at PAS and SPS [1]. It is related to distinction in speed of
input energy and in the parameters of influence by external pressure on powder material.

CONCLUSIONS

The response of WC-Co powder column (loaded external pressure) to high energy electrical
discharge has been described and understood in terms of peak current density and external
pressure. It was found that the density of WC-Co powder column reached its maximum
values at applied pressure P = 200 MPa and high voltage electrical discharge parameters
(the peak pulse current density = 95 kA/cm
2
). There is an upper limit of the discharge voltage
beyond which the powder material disintegrates like an exploding wire. Attempts of
compacting WC-Co powder composites by EDC method would give future fruitful results.

REFERENCES

1. Zhang J., Zavaliangos A., Groza J. (2003) P/M Sci. Technol. Briefs 5 (3): 17
2. Raichenko A.I. Fundamentals of Powder Sintering by Passing an Electric Current (1987)
Metallurgiya, Moscow
3. Grigoryev E. G., Rosliakov A.V. (2007) J Mater. Proc. Techn. 191 (1-3): 182
4. Baranov Yu.V., Troitskii O.A., Avraamov, Yu.S., Shlyapin A.D. Physical Bases of Electric-
Pulse and Electroplastic Treatments and New Materials (2001): Publisher MGIU, Moscow
5. Shakery M., Al-Hassani S. T. S., T. J. Davies: (1979) Powder Met. Int. 11: 120


Germany

Microstructure and Mechanical Properties of Fe
3
Al Bonded WC
Composites Obtained by Pulsed Electric Current Sintering

S.G. Huang, O. Van der Biest, J. Vleugels
Department of Metallurgy and Materials Engineering (MTM), Katholieke Universiteit Leuven,
Kasteelpark Arenberg 44, B-3001 Heverlee, Belgium

Abstract
In the present study, pulsed electric current sintering (PECS) or spark plasma sintering
(SPS), was applied to make 5-20 vol% Fe
3
Al bonded submicrometer sized WC composites.
This technique offers a unique combination of rapid heating/cooling and short processing
time, which makes it an interesting tool to consolidate WC-FeAl composites with limited WC
grain growth. The WC based composites with 5-20 vol% Fe
3
Al binder were consolidated by
PECS in the solid state at 1200C for 4 min. The microstructures and basic mechanical
properties were assessed. Microstructural analysis revealed a homogeneous Fe
3
Al binder
distributions and limited WC grain growth, together with some residual Al
2
O
3
clusters. An
excellent Vickers hardness (HV
10
) of 25.6 GPa and flexural strength of 1 GPa were achieved
for the WC-5 vol% Fe
3
Al composite. With 20 vol% binder phase, the hardness decreased to
19.9 GPa, while the strength slightly increased to 1.2 GPa.

1. Introduction
Intermetallic aluminides are attractive for applications as high-temperature structural material
due to their good strength at intermediate temperatures and excellent corrosion resistance at
elevated temperatures in oxidizing, carburising and sulfidising atmospheres. Iron aluminide
has been investigated for high-temperature structural applications [1] and binder phase for
ceramic particles such as WC, TiC, TiB
2
and ZrB
2
[2,3].
So far, ceramic particles reinforced iron aluminides are commonly produced by liquid
phase sintering or melt infiltrating of a porous ceramic skeleton [4]. By liquid phase sintering,
iron aluminide bonded TiC and WC composites with up to 60 vol% carbide were produced
and the fracture toughness of WC-40% Fe
60
Al
40
composites was in the range of conventional
WC-Co cemented carbides [2]. The wear resistance of the 40 vol% Fe
60
Al
40
bonded WC
composite was superior [2] and the oxidation resistance is 4 times higher [4] than for WC-40
vol% Co cemented carbides. Residual porosity however was unavoidable at higher ceramic
contents, mainly due to the low solubility of TiC and WC in liquid iron aluminide and the lack
of substantial dissolution-reprecipitation [4]. Nearly fully dense composites with 70-90 vol%
WC could be produced by the pressureless melt infiltration process [5].
In the present study, the possibility of solid state densification of 5 - 20 vol% Fe
3
Al bonded
submicrometer WC composites by means of pulsed electric current sintering (PECS), also
known as spark plasma sintering (SPS), was assessed. This technique offers a unique
combination of rapid heating/cooling and shorter processing times, which makes it an
interesting tool to consolidate WC-Fe
3
Al composites with limited WC grain growth. The
influence of the binder content on the microstructure development and mechanical properties
was investigated.

2. Experimental procedure
Starting powder mixtures composed of submicrometer sized WC (grade CRC015, Wolfram
Bergbau und Htten GmbH, Austria, FSSS = 0.5 m) with 5, 10 or 20 vol% N
2
-atomized
Fe
3
Al powder (AGH University of Science and Technology, Poland, -325 mesh) or 5, 10 and
20 vol% Co (Umicore grade Co-HMP, Belgium, FSSS = 0.55 m,) were prepared by wet-
mixing in ethanol on a multi-directional mixer (type T2A, Basel, Switzerland) for 24 h using
WC-7.5 wt% Co milling balls (grade MG15, Ceratizit). The suspension was dried in a rotating

evaporator at 65C. Pulsed electric current sintering (Type HP D25/1, FCT Systeme,
Rauenstein, Germany) was performed in a dynamic vacuum of 4 Pa. A pulsed electric
current was applied with a pulse/pause duration of 10/5 ms throughout all the experiments.
Around 32 g of powder mixture was poured into a cylindrical graphite die with an inner and
outer diameter of 30 and 56 mm and sintered for 4 min at 1200C under a pressure of 90
MPa, with a constant heating rate of 200C/min and an initial cooling rate of 200C/min.
Graphite paper inserts were used to separate the graphite die/punch set-up and the powder
mixture. A two-colour pyrometer (400-2300C, Impac, Chesterfield, UK) was focused at the
bottom of a central core hole in the upper punch, just 2 mm away from the top surface of the
sample. The die/punch/powder assembly is described elsewhere [6].
The bulk density of the sintered samples was measured in ethanol. The microstructure of
polished surfaces was examined by scanning electron microscopy (SEM, XL30-FEG, FEI,
Eindhoven, the Netherlands). The average WC grain size was determined from SEM
micrographs using Image-pro Plus software [7] to measure the linear intercept length of at
least 200 WC grains for each material grade. The reported values are the actually measured
average intercept length. Phase identification was conducted by a - X-ray diffractometer
(XRD, Seifert, Ahrensburg, Germany) using Cu K radiation (40 kV, 40 mA). The Vickers
hardness (HV
10
) was measured (Model FV-700, Future-Tech Corp., Tokyo, Japan) with an
indentation load of 98.1 N. The fracture toughness, K
IC
, was calculated from the length of the
radial cracks originating from the corners of these indentations according to the formula
proposed by Shetty et al. [8]. The reported values are the mean and standard deviation of
five indentations. The flexural strength at room temperature was measured in a 3-point
bending test (Instron 4467, PA, USA) on rectangular bard (25.0 3.0 2.0 mm), with a span
length of 20 mm and a cross-head displacement of 0.1 mm/min. The reported values are the
mean and standard deviation of 5 measurements. All sample surfaces were ground with a
diamond grinding wheel (type D46SW-50-X2, Technodiamant, The Netherlands) on a Jung
grinding machine (JF415DS, Gppingen, Germany).

3. Results and discussion
3.1. Phase diagram information of the WC-Fe
3
Al system
In order to select a proper processing temperature and estimate the solid solubility of Fe
3
Al in
WC, Thermo-Calc software [9] was used to calculate the phase diagram of the WC-Fe
3
Al
system. The solubility of WC in the solid bcc structured Fe
3
Al is 3.5 wt% at 1200C and
increases to 8.6 wt% at 1440C at the three phases equilibrium of WC, bcc and liquid. The
solubility of WC in the Fe
3
Al binder phase however is substantially lower than in the
conventional WC-Co material. At 1200C, solid fcc-Co dissolves nearly 8.5 wt% WC and the
solubility reaches 12 wt% at 1310C in the solid binder. Based on the phase diagrams
information, a substantial difference in WC grain growth is to be expected in the Fe
3
Al and
Co bonded WC composites sintered both in the solid and liquid states. In the present study,
a solid state densification temperature of 1200C was selected to consolidate the Fe
3
Al and
Co binder composites.

3.2. WC and Fe
3
Al starting powders
The morphology of the WC and Fe
3
Al starting powders, as well as the WC-10 vol% Fe
3
Al
powder mixture is presented in Fig. 1. The WC starting powder is in the form of soft
agglomerates of 200 nm sized WC grains. The N
2
gas-atomised Fe
3
Al starting powder
exhibits a round shape and the diameter varies from 10-30 m. Due to the inherent room
temperature brittleness of Fe
3
Al powder, the low energy mixing process caused the break up
of the Fe
3
Al agglomerates into sub-micrometer size grains and even dispersion in the WC
matrix, as shown in Fig. 2.c. A small amount of coarse and oxygen-rich grains were also
observed in the powder mixtures, as confirmed by EDS point analysis (see Fig. 1.d). The
oxygen content in the Fe
3
Al powder was reported to be 0.07 wt% [10]. Aluminides easily form
an impervious Al
2
O
3
layer that can provide excellent corrosion resistance in a wide range of
corrosive environments [11,12]. This alumina layer however was experienced as an obstacle
during liquid phase processing of Fe
3
Al with 70 vol% WC/TiC, resulting in an inhomogeneous
low density microstructure [4]. In this study, the influence of the Al
2
O
3
particles on the

densification behaviour of the powder mixtures is less profound due to the relatively low
content.
(a)
(b)
(c) (d)
Fig. 1. Morphology of WC (a), Fe
3
Al (b) and mixed WC-10 vol% Fe
3
Al (c) starting powders,
together with an EDS point analysis of the labelled grain in (d).

3.3. Microstructure and phase constitution
The bulk density of the sintered composites is summarized in Table 1. According to ASTM
B276-91 [13], the apparent porosity of the composites is better than the A02B00C00 level,
indicating near full density can be obtained by solid state PECS at 1200C. Plastic
deformation of the Fe
3
Al binder above 900C could assist densification of the WC-Fe
3
Al
composites [14,15]. Fig. 2 shows the low magnification images of the composites with 5 and
20 vol% Fe
3
Al, PECS at 1200C for 4 min. Beside the bright WC-Fe
3
Al composite matrix, a
third phase forming dark contrast clusters can be clearly differentiated. The dark clusters are
aligned perpendicularly to the PECS pressing direction. Compositional analysis using energy
dispersive spectrometry allowed to identify the black phase as pure Al
2
O
3
. Besides the Al
2
O
3
in the Fe
3
Al powder, oxygen pick-up during powder processing could be an additional source
for Al
2
O
3
particle formation during PECS. The observation of the aligned Al
2
O
3
strings is
consistent with the work of Maziasz et al. [16], illustrating a nearly continuous film of Al
2
O
3

particles at the Fe-Al particle boundaries after extrusion below 1100C.

(a) (b)
Fig. 2. SEM micrographs of polished 5 (a) and 20 (b) vol% Fe
3
Al composites.

1

Although the WC-Fe
3
Al and WC-Co composites were consolidated in the solid state, the
WC crystal morphology changed from equiaxed in the starting powder to a faceted angular
shape after PECS for 4 min at 1200C, regardless of the binder content as presented in Fig.
3. The darker contrast Fe
3
Al binder phase was found to be homogeneously located around
the brighter WC grains, indicating good wetting between WC and Fe
3
Al. The average size of
WC grains in the fully dense WC-5, 10 and 20 vol% Fe
3
Al composites, as determined by the
linear intercept method, was 218, 224 and 206 nm, while the mean free path length of the
Fe
3
Al binder phase is less than 30 nm.
Compared to the PECS WC-Fe
3
Al composites, the solid state sintered WC-Co cemented
carbides had a much coarser WC grain structure, as shown in Fig. 3.d. An average WC grain
size of 380 nm was measured in the WC-20 vol% Co grade, PECS for 4 min at 1200C. The
finer WC grain size in the WC-Fe
3
Al composites, processed under identical conditions, can
therefore be attributed to the lower WC solubility, leading to a slower WC grain growth. For
WC-Co cemented carbides, rapid WC grain growth takes place when sintering nanometer
sized WC-Co starting powders even in the solid state [17]. The maximum densification rate in
the WC-Co system is reported in the 1200-1280C range corresponding to the rapid
dissolution and significant solubility of WC in the Co binder [18].

(a) (b)
(c) (d)
Fig. 3. Microstructure of the PECS composites on polished surfaces, WC-5 vol% Fe
3
Al (a),
WC-10 vol% Fe
3
Al (b), WC-20 vol% Fe
3
Al (c) and WC-20 vol% Co (d).

3.4. Mechanical properties
The Youngs modulus (E), Vickers hardness (HV
10
), fracture toughness (K
IC
) and flexural
strength (TRS) of the WC-Fe
3
Al materials are compared with those of the WC-Co cemented
carbides in Table 1. The Young's modulus of the WC-Fe
3
Al composites (see Table 1) is
substantially higher than for the WC-Co materials, due to the higher stiffness of the
intermetallic binder. The stiffness decreased from 693 GPa for the 5 vol% Fe
3
Al composite to
539 GPa for the 20 vol% Fe
3
Al composite. The hardness of the Fe
3
Al and Co binder
materials decreases with increasing binder content due to the lower intrinsic hardness of the
binder phase. The hardness of the WC-Fe
3
Al composites was measured to decrease from an
excellent 25.65 0.32 GPa at 5 vol% Fe
3
Al to 19.90 0.29 GPa at 20 vol% Fe
3
Al. These
values are about 4 GPa higher than for the corresponding WC-Co materials. Since both
composites have the same binder content, the hardness difference can be directly correlated

to the WC grain size and binder type. For a given Co binder content, it is well known that the
hardness of cemented carbides increases with decreasing WC grain size, follows actually a
Hall-Petch type relationship with 1/(d
WC
)
0.5
[19]. The much finer WC grain size in the WC-
Fe
3
Al composites mainly attributed the higher hardness of the intermetallic binder grades.
Combined with a hardness of 27.46 GPa for fully dense binderless WC [20], the influence of
the Fe
3
Al binder content on the hardness of the solid state sintered WC-Fe
3
Al composites
can be linearly fitted as HV
10
(GPa) = 27.46 - 0.36 vol% Fe
3
Al, while the hardness of the
WC-Co materials can be expressed as HV
10
(GPa) = 23.65 - 0.35 vol% Co.

Table 1. Physical and mechanical properties of the PECS WC-Fe
3
Al composites
Composition
(vol %)

(g/cm
3
)
E
(GPa)
HV
10

(GPa)
K
IC

(MPa m
1/2
)
TRS
(MPa)
WC - 5 Fe
3
Al 14.85 693 25.65 0.32 7.6 0.3 1009 89
WC - 10 Fe
3
Al 14.29 630 24.13 0.17 7.5 0.1 1191 151
WC - 20 Fe
3
Al 13.31 539 19.90 0.29 7.6 0.1 1245 140
WC - 5 Co 14.77 585 21.79 0.24 7.6 0.2 1923 234
WC - 10 Co 14.85 566 20.29 0.50 8.7 0.1 2298 101
WC - 20 Co 14.30 507 16.58 0.13 9.4 0.2 2166 252

With respect to the fracture toughness, there exists an antagonistic correlation between the
hardness and fracture toughness of WC-Co materials, i.e., the hardness increases and
fracture toughness decreases with decreasing WC grain size for a given Co content [21]. The
Co binder plays a crucial role in shielding the stress field in front of a crack tip to improve the
toughness [22]. Ductile failure of the Co binder is usually found on fractured WC-Co
surfaces. The toughness of the WC-Fe
3
Al composites however is independent on the Fe
3
Al
content and is substantially lower than for the WC-Co grades with 10 and 20 vol% binder.
Microscopy analysis of the crack propagation pattern in WC-Fe
3
Al indicated an intergranular
fracture mechanism, similar as in WC-Co. The modest fracture toughness of the WC-Fe
3
Al
composites can therefore be attributed to the much finer WC grains and especially the
intrinsic brittle nature of the intermetallic Fe
3
Al binder. A higher fracture toughness of 12-21
MPa m
1/2
, but lower hardness of 10-14 GPa, was reported for WC reinforced FeAl
composites with a high binder volume fraction and coarse WC grains [2,4,23].
The 3-point flexural strength, listed in Table 1, ranges from 1000 to 1200 MPa for the WC-
Fe
3
Al composites, suggesting that the binder content has no significant influence on the
mechanical strength. Generally, the flexural strength of conventional WC-Co hardmetals is in
the range of 2000 to 4000 MPa, which is functions of Co content, WC grain size, carbon
balance, and other chemical composition and microstructure factors [24]. In this study, the
PECS WC-Co materials have a flexural strength of 1900 to 2200 MPa. It is known that the
flexural strength correlates basically with the fracture toughness and the critical flaw size,
which is determined by the grain size, the defects in the microstructure such as pores,
agglomerates, large grains and micro-cracks [25]. Since all WC-Fe
3
Al composites were fully
densified, exhibit a similar microstructural homogeneity and comparable WC grain size, it is
not impossible that the relatively larger Al
2
O
3
crystal clusters are strength determining.

4. Conclusions
WC-5, 10 and 20 vol% Fe
3
Al composite powders could be fully densified in the solid state by
means of pulsed electric current sintering within 4 min at 1200C under a pressure of 90
MPa. Due to the small solid solubility of WC in the Fe
3
Al binder, WC grain growth could be
limited to 200 nm independent on the binder content. Aligned Al
2
O
3
particle clusters were
formed during PECS due to the alumina contamination in the Fe
3
Al starting powder.
The hardness of the WC-Fe
3
Al composites linearly decreased from 25.65 GPa for the 5
vol% Fe
3
Al composite to 19.90 GPa for the 20 vol% Fe
3
Al composite. A very high Young's
modulus of 693 GPa was measured for the 5 vol% Fe
3
Al composite. The fracture toughness,
7.6 MPa.m
1/2
, and flexural strength, 1000-1200 MPa, were hardly influenced by the Fe
3
Al
content.

Compared to the WC-Co materials processed under exactly the same conditions, the WC-
Fe
3
Al composites exhibit a substantially higher hardness and Young's modulus of about 4
and 100 GPa respectively. Although the toughness of the 5 vol % binder materials is
comparable, the toughness of the higher binder content WC-Fe
3
Al composites is lower than
for the WC-Co materials. The bending strength of the WC-Co grades is superior.

5. Acknowledgements
This work was performed within the framework of the Research Fund of K.U.Leuven under
project GOA/08/007. The authors also acknowledge the Belgian Federal Science Policy
Office (BELSPO) for financing the NACER project (contract P2/00/07) and thank the AGH
University of Science and Technology, Poland, for the N
2
-atomized Fe
3
Al powder.

6. References
[1] P.R. Munroe, Mater. Forum 1997;21:79-99.
[2] J.H. Schneibel, C.A. Carmichael, E.D. Specht and R. Subramanian. Intermetallics
1997;5:61-67.
[3] T.N. Tiegs, K.B. Alexander, K.P. Plucknett, P.A. Menchhofer, P.F. Becher and S.B.
Waters. Mater. Sci. Eng. A 1996;209:243-247.
[4] R. Subramanian and J. H. Schneibel, Materials Science and Engineering A
1998;244:103-112.
[5] R. Subramanian and J.H. Schneibel,Intermetallics 1997;5:401-408.
[6] K. Vanmeensel, A. Laptev, J. Hennicke, J. Vleugels and O. Van der Biest, Acta Materialia
2005;53:4379-4388.
[7] Image-Pro Plus, version 4, Media Cybernetics, Silver Spring, Maryland.
[8] D.K. Shetty, I.G. Wright, P.N. Mincer, A.H. Clauer, J. Mater. Res. 1985;20:1873-1882.
[9] B. Sundman, CALPHAD 1985;9:153-169.
[10] T. Sleboda, K. Doniec, Archives of Materials Science and Engineering 2007;28:613-616.
[11] P.F. Tortorelli, J.H. DeVan, in: J.H. Schneibel, M.A. Crimp (Eds.), TMS, Warrendale, PA,
1994, pp. 257.
[12] C.G. McKamey, J.H. DeVan, P.F. Tortorelli and V.K. Sikka. J. Mater. Res. 1991;6:1779-
1805.
[13] ASTM standard B 276-91, Standard test method for apparent porosity in cemented
carbides, ASTM Annual Book of Standards, Philadelphia, PA, 1996.
[14] D.Lin and Y.Liu, Mater. Sci. Eng. A 1999;268:83-89.
[15] R. yszkowski and J. Bystrzycki, Intermetallics 2006;14:1231-1237.
[16] P.J. Maziasz, D.J. Alexander and J.L. Wright, Intermetallics 1997;5:547-562.
[17] Huang SG, Li L, Vanmeensel K, Van der Biest O, Vleugels J, Int J Refract Met Hard
Mater 2007;25:417-422.
[18] Cha SI, Hong SH, Kim BK, Mater Sci Eng A 2003;351:31-38.
[19] H.C. Lee, J. Gurland Jr. Mater. Sci. Eng 1978;33: 125-133.
[20] S.G. Huang, K. Vanmeensel, O. Van der Biest, J. Vleugels, Int J Refract Met Hard Mater
2008;26: 41-47.
[21] Schubert WD, Neumeister H, Kinger G, Lux B. Int J Refract Met Hard Mater 1998; 16:
133-142.
[22] M. Kotoul, Acta Mater 45 (1997), 3363-3376.
[23] R. Subramanian and J. H. Schneibel, J. Metals 1997;49:50-54.
[24] Zhigang Zak Fang, Int J Refract Met Hard Mater, 2005;23:119-127.
[25] Michel W. Barsoum, International Editions 1997, The McGraw-Hill Companies, Inc. pp
391-438.


Sintering the Easy Way A Fast Heating Process by Dr. Fritsch for
Metal and Ceramic Powders

Author: Pascal Schreyer, Dr. Fritsch GmbH
Co-Author: Jens Huber, Dr. Fritsch GmbH
Dr. Fritsch Sondermaschinen GmbH
Powder Shaping Technologies
Dieselstrasse 8
D-70736 Fellbach
Phone +49-711-51832-0
pst@dr-fritsch.de

ABSTRACT
Direct Hot-Pressing, developed by Dr. Fritsch, is nothing less than a little revolution in the
field of sintering technologies. Uni-axial hot pressing is widely known and established in the
industry. Most of the established systems heat up the furnace chamber which results in long
durations for heating-up. Additionally, a lot of the energy is lost since it is used for heating up
the atmosphere instead of the sinter part. DIRECT Hot-Pressing is different. By connecting
the mould directly with the electrical power, sinter cycle times of only a few minutes can be
achieved . Additionally, the energy efficiency very high, resulting in comparably low operating
cost. At the same time, the near-net-shape precision achieved is very high and reduces
mechanical re-working to an absolute minimum. DIRECT Hot Pressing Technology from Dr.
Fritsch has been sold more than 900 times worldwide and many different kind of industries
benefit from it.

MAIN PART
Hot pressing is a well known technique in the metal powder field. It is one of the sintering
methods to solidify cold pressed (metal) powders. However, for mass production, free
sintering is preferred. With free sintering, the manual work to fill the pressing moulds can be
saved, resulting in cheaper final products.
Another competing technology for traditional hot pressing is the hot isostatic pressing or HIP
process. This process puts Pascal's Law - "pressure is transmitted equally and in all directions
through a liquid or a gas" to work to produce high quality parts.
While HIPping also needs preparation of the prepressed powder parts, it gives high final
density and good structural proportion to the parts. The costs of a HIP press, is however,
substantial.
Uni-axial hot pressing is mainly used in the diamond tool business although there are also
some other applications. Here, a green density of only approximately 65 per cent is used
because of the abrasive character of the powders. Higher green
density would inevitably result in poor tool life of the cold press die. To achieve higher green
density, a higher pressing force would be needed. This would mean more friction and - due to
the use of diamond grit - more wear, and therefore lower tool life. This low green density
needs substantial stroke length for sintering. Sintering temperatures of 650-1200C
are normally used. Within hot pressing technology, three distinctly different types of heating
can be found in use:

Inductive heating
Indirect resistance heating
Direct resistance heating


1. Inductive heating

Figure 1. Inductive heating
(1) Ring mould
(2) Top / bottom punch
(3) Powder part
(4) Induction coil

It works by heating with high-frequency coils that can induce temperature rises. The mould is
made out of graphite or steel, and pressure is applied by one or two cylinders onto the
punches. The mould is positioned in an induction coil. During sintering a high frequency
generator and the induction coil generates heat in the mould.
The advantage is that the pressure and the inductive power are completely independent.
Powders with a liquid phase are suitable and low pressures are possible.
Among the disadvantages are the expense of a high-frequency generator and
the need for proper alignment. If the mould is placed off centre, the heat distribution is
uneven. But the main disadvantage is the dependence of the process on good heat
conductibility. The magnetic field can penetrate the mould only 0.5mm to 3mm.
From there on, the heat has to be "transported" into the mould by the thermal conductivity of
the mould material. Even heating is much more difficult if the mould is off-centre. Another
potential problem is heating rate. Too high a heat up rate will therefore result in high
temperature differences that can destroy the mould.

2. Indirect resistance heating

Figure 2. Indirect resistance heating
(1) Ring mould
(3) Powder part
(4) Graphite heating element

With indirect resistance heating technology, the mould is placed in a heating chamber. The
chamber is heated by graphite heating elements. These elements are heated by electrical
current. The heat is then transferred into the mould by convection. As the electrical energy
heats the heating elements that then heat the mould in a second state, the process is called
indirect resistance heating. Advantages are high achievable temperatures, independent from
the conductibility of the mould and independent from heat and pressure. Main disadvantage
is the time that it takes to heat up the mould. It takes relatively long to bring the heat from
the outside of the mould to the centre and to distribute the heat evenly.

3. Dr. Fritsch direct resistance heating technology

Figure 3. Direct resistance heating
(1) Ring mould
(3) Powder part
(4) Graphite electrode
(5) Brass electrode
(6) Copper plate
(7) Transformer

With direct resistance heating, the mould is directly connected to electrical power. The
resistance of the mould and the powder part leads to heat that is generated directly in the
mould. This results in very high heating speed. The previous two methods are
both somehow related to the physical properties of thermal conductivity.
However, with direct resistance heating, the heat is generated where it is needed.
Up to now, this technology has been used mainly in the metal powder field with typical
maximum temperatures of 1,400C.
Using metal powder, the conductivity of the mould is ideal for fast heating of the work-piece.
Moulds that have a big diameter and relatively small height can be heated up very fast.
Since 2005, Dr. Fritsch has been offering a high temperature sintering press. This machine
has reached 2,400C and combines high heating rates with high temperatures and precise
temperature control. The workpiece is directly measured by means of a pyroscope.
As standard, maximum pressing power is up to 600kN (60 tonnes), heating power is 150kVA
at 4.5V to 6V secondary voltage. The machine can operate at vacuum or under protective
atmosphere; overpressure can be realised as an option. The machine has been successfully
tested for metal and ceramic work pieces such as boron carbide at temperatures up to
2400C. An optional magazine is available which can load the workpieces automatically. Up
to 54 pieces can be stocked. Process control and off-line programming via PC is
also an option. The machine is especially suitable for processes that need high heating rates,
e.g. for materials that should not be kept at high temperatures too long or for processes that
require fast heating rates for high productivity. For further information go to
www.directhotpressing.com.

Manuscript refereed by Dr Jrgen Schmidt, Fraunhofer IFAM, Germany

Comparative Wear Behavior of the Antifriction Al-Based Composite
Materials Sintered by Classic and Spark Plasma Sintering Route

SIMA, G., GNGU, O., MANGRA, M.

University of Craiova, Faculty of Engineering and Management of Technological Systems,
1
st
, Calugareni street, 220037, Drobeta Turnu Severin, ROMANIA

Abstract. The new materials have challenged the automotive industry since they have been
showed a lot of improved properties than the classic materials. Considering the large section
of the applications for the automobiles, the important request for these materials is the
optimal ratio high strength/low density in economical financial terms. The Al based matrix
composites are one of the advanced materials, exhibiting good and stable properties at
ambient and high temperatures; these are the reasons for their application in the automotive
industry. It is well know that Powder Metallurgy technology is the most used for antifriction
composite materials manufacturing. Depending on the Powder Metallurgy route and the
components nature, the antifriction composite materials present a different behavior for the
same tests. The paper is focusing on the wear behavior of antifriction composite materials
elaborated by classic sintering and Spark Plasma Sintering.

1. Introduction
In many industrial applications, the most important parameter in material selection is
specific strength. Regarding the characteristic of wear resistance of the antifriction composite
materials (ACM
s
), references offer many experimental data. Significant developments have
been achieved in the system Al/SiC metal matrix composites.
The aluminium matrix composites reinforced with dispersed particles (Graphite, SiC)
appeared as an alternative for components from automotive and aircraft industries. These
materials are characterized by a high mechanical strength/density ratio, high thermal
conductivity, high wear strength and damping capacity.
The improvements in the tribological properties in the presence of Graphite generally
have been attributed to the formation of a lubricant third-body film on the contact surface.
Gibson et al. reported that ACM
s
based Al-Si containing 2 pct Graphite had better wear
properties than those with 8 pct Graphite [1, 2]. Biswas found that wear resistance of
unreinforced Al-Si was superior to those with 3 to 6 pct Graphite content [3]. Both authors
concluded that the degradation of the strength and ductility of the composites whit Graphite
outweighed the beneficial effects of reduced friction.
Taking into account these considerations, the authors proposed study of wear
behavior of the ACM
s
reinforced with SiC and Graphite and elaborated by classic sintering
(CS) and Spark Plasma Sintering (SPS). The research programme has been carried out in
cooperation between Faculty of Engineering and Management of Technological Systems,
University of Craiova (Ro), Max-Planck-Institute for chemical physics of solids (Dresden,
Germany) and Politehnica University of Bucharest (Ro).

2. Materials
Different routes to obtain ACM
s
have an important influence especially on the
morphology of the powder particles and, as result, on the microstructural properties of the
composite. The hybrid composites studied in this paper represent the class of the Al-matrix
composites reinforced by SiC and Graphite. The chemical composition is mentioned in
Tab.1.

Tab. 1 Chemical composition of composite powders and their code
No. Chemical composition of composite powders Sample code
1 Al+20(vol.%)SiC+ 5(vol.%)Graphite ACM-5
2 Al+20 (vol. %)SiC+10 (vol. %)Graphite ACM-10
The ACM
s
are processed by Reactive Mechanical Alloying (RMA). The
Al/(SiC+Graphite) composites were successfully processed from composite powder particles
that have been elaborated by RMA. The particle size distribution as well as the compaction
behavior of these powders were studied elsewhere [4-7].
For both types of tribological composites based on Al, the great advantage induced
by RMA process is the occurring of Al
4
C
3
,
having an important role on the composite
hardening phenomenon. Phases
microhardness of the Al/SiCGr
composites processed by different PM
routes are shown in Fig 1. Should be
noted that the real values of Al matrix and
Al
4
C
3
microhardness are different
represented from the SiC microhardness-
this one is much higher that 300HV
0,1
, but
this representation way was adopted
considering the visibility scale. Due to the
presence of the aluminium carbide, the
cold compaction and sintering in inert
atmosphere) are enough to reach the as
sintered stage of the composite at hight
hardness values (105120) HV [8].

Al
4
C
3
phase is the result of the diffusion between Al matrix and SiC reinforcements.
Beside the hardness improvement by aluminium carbide presence in the microstructure,
Figure 2 (Al/20SiC) and Figure 2(Al/20SiC+5Gr), the most important effect of Al
4
C
3
is the
increasing of the wear resistance during pin (the composite sample) on disc (carbon steel)
tests in dry conditions.

Afterwards, the sintering behaviour of those compacts has been researched through
two sintering routes [9]:
- classic, in inert atmosphere, at (650-750)
0
C, two values for the dwell time (3; 6 h) and
two extreme cooling rates (5; 1200)
0
/min at Laboratory of University of Craiova
(Romania)
- advanced meaning SPS, in the same condition, except the single cooling rate (by the
furnace) - at Max-Planck Institute for Chemical Physics of Solids, (Dresden, Germany)
The materials were researched from the point of view of the physical and microstructural
characteristics obtained after the CS and SPS sintering routes. The main properties of the
processed ACM
s
elaborated by CS and SPS are mentioned in Tab.2, respectively Tab 3.

0
100
200
300
complex PM
steps
classic PM
steps
classic PM
steps
Al matrix
SiCp
Al4C3
M
i
c
r
o
h
a
r
d
n
e
s
s

H
V
0
,
1

[
M
P
a
]

RMA+
+cold compaction+sintering
3h maintaining - 6h maintaining
complex
PM steps
Fig. 1. Phases microhardness of the Al/SiCGr
composites processed by different PM routes [8]

50 m
SiC
pores
Al
4
C
3
Al
matrix
a)
Figure 2 Microstructure of Al/20SiC composite: RMA+
+cold compaction + sintering (700C/6h maintaining)

2 m
SiC
Gr
cracks
Figure 3 Microstructure of Al/20SiC+5Gr composite:
RMA + cold compaction + sintering (750C/3h
maintaining)


Table. 2. The best physical characteristics of ACM
s
elaborate by CS

Table. 3.. The best physical characteristics of ACM
s
processed by SPS

The experimental results were compared. It is obviously that in the case of SPS,
because the compaction and sintering developed simultaneously and the thermal effects is
higher due to the generated plasma, the experimental values are the best, fact confirmed by
other authors in their papers [10, 11].
The respective microstructures elaborated by CS, respectively SPS related to the
processing parameters from the Tab. 2, 3 are shown in Fig 4-7

3. Experimental procedures. Tribological tests
Dry friction tests were performed. The examined materials were ACM-5 and ACM-10
elaborated by CS and SPS process. Tribological experiments were carried out using a pin-
on-disc apparatus, developed at the Laboratory of Machine Components and Tribology, from
Politehnica University of Bucharest, Romania, Fig 8.

Simple code
Technological parameters Dimensional characteristics Sintered density
Sintering
temperature
[
0
C]
Maintaining
sintering
time [h]
Cooling
rate
[
0
C/min]

[mm]
Height
[mm]
Volume
[mm]

[g/cm
3
]
[%]
from t
ACM-5/CS 650 6 5 10,54 8,15 0,710 2,42 89,1
ACM-10/CS 650 3 5 10,78 9,77 0,891 2,39 90,8

Simple code
Technological parameters Dimensional
characteristics

Weight
[g]
Sintered density
Heating
rate
[
0
C/min]
Sintering
temperature
[
0
C]
Maintaining
sintering
time [h]
Cooling
rate
[
0
C/min]

[mm]
Height
[mm]
Volume
[mm]

[g/cm
3
]

[%] from t
ACM- 5/SPS 20 600 0,5 5 10 5,20 0,408 1,106 2,71 99,81
ACM-10/SPS 20 600 0,5 5 10 6,80 0,534 1,155 2,16 82,12
Fig. 6. Microstructure of ACM-5 with
= 2,71 g/cm
3
elaborate by SPS

= 2,16 g/cm
3
elaborate by SPS

Fig.4 Microstructure of ACM-5 with
= 2,42 g/cm
3
elaborate by CS
= 2,39 g/cm
3
elaborate by CS

The wear tests were performed in following experimental conditions monitored in Tab 4.

Tab.4. Technological and constructive parameters of the test pin-on-disc wear

Parameters of the tests pin on disc wear
Contact force [N] 322,6
Contact pressure p [MPa] 5,81
Sliding velocity v [m/s] 0,754
Sliding distance L [mm] 100

In the present investigation, the authors realized a comparative analysis on the wear
behaviour of the researched ACM processed by the both routes. There have been
determined the friction coefficient values. The experimental results are mentioned in Tab 5.

Table.5. Wear test condition

Simple code
Experimental results
Lamellar elastic force T N Friction coefficient
iniial final mediun iniial final medium
ACM-5
classic
650
0
C / 3h / 5
0
C min 15,254 35,593 25,423 0,205 0,479 0,342
650
0
C / 6h / 5
0
C min 13,559 36,441 25 0,182 0,49 0,336
700
0
C / 3h / 5
0
C min 23,729 39,831 31,780 0,319 0,536 0,427
700
0
C / 6h / 5
0
C min 24,576 39,831 32,203 0,33 0,536 0,433

ACM-10
classic
sintering
650
0
C / 3h / 5
0
C min 11,864 38,136 25 0,16 0,536 0,348
650
0
C / 6h / 5
0
C min 13,559 43,22 28,389 0,182 0,581 0,381
700
0
C / 3h / 5
0
C min 10,214 38, 733 24,473 0,121 0,509 0,315
700
0
C / 6h / 5
0
C min 8,475 37,288 22,881 0,114 0,501 0,307
ACM-5
SPS
20
0
C min/600
0
C / 0,5h 6,78 11,864 9,322 0,09 0,158 0,124
20
0
Cmin/ 500
0
C / 0,5h 8,475 15,254 11,864 0,113 0,203 0,158
20
0
Cmin/ 500
0
C / 0,5h broken sample
100
0
Cmin/500
0
C/ 0,5h 8,475 11,864 10,169 0,113 0,158 0,135

ACM-10
SPS
20
0
C min/600
0
C / 0,5h 5,085 8,475 6,78 0,068 0,113 0,09
100
0
Cmin/500
0
C/ 0,5h 13,559 16,102 14,830 0,181 0,215 0,198
20
0
Cmin/ 500
0
C / 0,5h 5,932 11,864 8,898 0,079 0,158 0,118
100
0
Cmin/500
0
C/ 0,5h broken sample

4. Results and discussions
The influence of the graphite and the sintering technological parameters on the
composites wear behavior was studied both for the classic and advanced sintering
procedure.
The most advantageous processing parameters from the point of view of the wear
behavior are presented in table 6.
1 - weight; 2 - gradual rule; 3 - test tubes; 4 - slider fixing
tubes; 5 - disc; 6 - radial-axial level; 7 - rigid support;
8 - countershaft; 9 - level; 10 - trapezoidal transmission;
11 - reduction; 12 - electric motor; 13 - gauge force; 14 -
deck force; 15 - elastic laminae; 16 dowel; 17 - binder
Fig.8. Pin-on-disc equipment for friction test
Table.6. The most advantageous processing parameters, from the point of view of the wear behavior

The best wear behavior, in the case of the CS composite, was obtained for the 10%
graphite reinforcement that had the lowest friction coefficient (=0,307) [12]. The
experimental results highlight the conclusions drawn by the researches which mentioned that
the maximum 10% reinforcing level in graphite can provide the best antifriction properties.
Also, in the case of SPS processed composite materials, the same 10% graphite
reinforcement material showed the best wear behavior (=0,09) [12].
Given the results, the author realized a comparative analysis on the wear behaviour
of the researched ACM processed by the mentioned two sintering routes, Tab 7.

Tab.7. The most advantageous processing parameters in both routes,
from the point of view of the wear behaviour
Technological parameters
Sintering temperature/Maintaining sintering time/Cooling rate
Composites
type
friction
coefficient
classic 700
0
/ 6h / 5
0
C/min MCA-10 0,307
SPS 600
0
/ 0,5h / 5
0
C/min MCA-10 0,09

The best processing route to elaborate antifriction composite materials with Al matrix
reinforced by 20%SiC and 10%Gr is the SPS in the following technological parameters
(heate rate=20
0
C/min, sinterint temperature =600
0
C, dwell time=0,5 h, cooling rate=5
0
C/min).
The processed material has high characteristics (density=2,16 g/cm
3
) and self-lubricating
properties (=0,09).
On the experimental results (friction coefficient values), there are drawn the
conclusions regarding the influence, in classic and SPS sintering, of the sintering
temperature and time respectively the heating rate and sintering time as well as the graphite
content on the friction coefficient that is a important parameters able to recommend these
materials as antifriction materials.

5. Conclusions
The applied technologies are the CS as well as the advanced technique of SPS. It is
well known that the SPS systems offer many advantages: lower sintering temperatures,
shorter sintering time and higher properties values of the sintered material vs. the
corresponding ones obtained by the classical sintering route [10, 11].
The tribological composite materials type Al/SiC+Graphite is able have a large
applications in the automotive industry asa a performing friction material due to:
- the favourable associantion betwwn a hight hardness (105120) HV and a low density
(8395)%TD.
- the PM route
- the good wear resistance, a property pointed out by the friction coefficient.
For the researched materials, in the case of SPS, there were underlined the decreasing
of the sintering temperatures with 100
0
C and the dwell time of less than (33%...66%) vs. the
corresponding values to the technological parameters from the classic sintering case (Tab 7)
The experimental results confirm the researches data in the field:
Simple code Technological parameters of sintering
Sintering temp/Maintaining sintering
time/Cooling rate
physical characteristics
= max[g/cm
3
]
self-lubricating
properties

Classic sintering
ACM-5/CS 650
0
C/6h / 5
0
C/min 2,42 0,336
ACM-10/CS 700
0
C /6h/ 5
0
C/min 2,29 0,307
SPS
ACM-5/SPS 600
0
C/ 0,5

h / 5
0
C/min 2,71 0,124
ACM-10/SPS 600
0
C/ 0,5

h / 5
0
C/min 2,16 0.09
the most advantageous processing route to elaborate the antifriction composite materials
is by SPS because thus the obtained properties are superior than for the same materials
processed by the classic technologies;
in order to obtain a good antifriction composite material, the highest reinforcing level in
graphite in 10% (vol.). It is already know that the exceeding of this reinforcing level
beyond 10% leads to the decreasing of the coefficient friction value, due to the occurring
of the phenomena specific to the third body.

Acknowledgments
The authors of this paper would like to give special thanks to Prof Dr. Yuri Grin from Max-
Planck Institute for Chemical Physics of Solids (Dresden, Germany) for his support
regarding the SPS research. Also the authors would like to thank to Prof. Dr. Andrei Tudor
from Politehnica University of Bucharest (Ro), Laboratory of Machine Elements and
Tribology, for the collaboration to carry out experiments on study the wear behavior of
materials under research

References
[1] Gibson, P.R., Clegg, A.J., Das, S., Material Science and Techology, vol I, 1985, p.559-
567;
[2] Das, S., Prasad, S.V., Ramachandram, T.R., Wear, 1989, vol.133, p.173-187;
[3] Biswas, S.K., Pramila Bai, B.N., Wear, 1984, vol.68, p.347-358;
[4] Mangra, M., Sima, G., Gingu, O., Teisanu, C., Ciupitu, I. Gheorghe, St., Tribological
composite materials for the automotive industry. Physical and mechanical
characteristics, Proc. of the World Tribology Congress WTC 2001, Vienna, Austria,
2001, ISBN 973-8043-324-4, p.491-496;
[5] Sima, G., Mangra, M., Gingu, O., Critical study on the dimensional change during solid
state sintering of PM Al/SiC+graphite composites, Proc. of the 3rd International
Conference on Materials and Technologies MATEHN02, Cluj-Napoca, Romania, 2002,
ISSN 1221-587, p.545-550;
[6] Mangra, M., Sima, G., Gingu, O., Ciupitu, I., Influence on the cooling rate on Al/SiC+Gr
composite microstructure, University Carlos the III Madrid, Spain, Proc. of the 1st
International Conference on Advanced Materials and Processing Technologies
AMPT01, Madrid, Spania, 18-21 Sept. 2001, ISBN 84-95821-08-7, p.1547-1551;
[7] Sima, G., Mangra, M., Gingu, O., Ciolacu, F., On microstructural evolution of hybrid
composites Al/SiC+Graphite during solid state sintering, Second International
Conference -Advanced Materials and Structures, Timisoara, Romania, 2002, p.329-332;
[8] Mangra M., Sima G., Gingu O., Orban R.: Research Concerning Al/(SiC+Graphite)
Nanocomposites Powder Particles Elaboration Having Self-Lubricant Properties, 7
th

ICCE, July 9-11, Colorado, 577-578
[9] Sima, G., Gingu, o., Criveanu, M., Comparative the physical and mechanical properties
of the antifriction Al-based composite materials sintered by classic and SPS route,
International Powder Metallurgy Congress EURO PM2007, Toulouse, France, vol.3,
p.221-223;
[10] Tokita, M., Mechanism of Spark Plasma Sintering (SPS), 2000 Powder Metallurgy World
Congress, Kyoto, vol II, p.252-255;
[11] Tokita, M., Trends in Advanced (SPS) Spark Plasma Sintering Systems and
Technology, Journal of Society of Powder Technology Japan, 1993, vol 30, p.790-804;
[12] Sima, G., Gingu, O., Tribological Properties on the Antifriction Composite Materials
(ACM) Al-based Sintered by Classic route, RECENT Journal, International Conference
on Economic Engineering and Manufacturing Systems, Iceems, Brasov, Brasov, 25-26
october 2007;
[13] Sima G., Gingu, O., Mangra, M., Influence of the reinforcing elements on the wear
behavior of Al/(SiC+Graphite) composites elaborated by SPS technology, 4
th
International Conference on Powder Metallurgy RoPM 2009, Craiova, Romania 8-11
iulie 2009 (will be published).

Euro PM2009 Microwave Sintering
Manuscript refereed by Professor Oana Gingu, Craiova University, Romania

Attempts to Sinter a Fine Nickel Powder in a Microwave Single-
Mode Cavity

S. Charmond, D. Bouvard, C.P. Carry
Grenoble INP/ CNRS / Universit Joseph Fourier
Laboratoire SIMaP
1130

rue de la piscine - BP75
F-38402 Saint Martin d'Hres

Abstract

Some authors claimed the feasibility to sinter miscellaneous metal powders by microwave
heating with or without susceptor. But few works reported the microwave sintering of nickel
powders, especially in a single-mode cavity. The present study therefore aims at
investigating the heating and sintering behaviour of porous nickel powder metal compacts by
direct microwave sintering. An instrumented TE
10p
single-mode microwave cavity was used
to position the specimens in a separated E- or H-field. Direct microwave heating experiments
were performed in controlled atmosphere (argon, helium/4%-hydrogene, nitrogen, air). Main
difficulties encountered during microwave heating were arcing and plasma, which arose in
any case. Finally, near fully dense compacts have been obtained in a few minutes under E-
field and in He-H
2
atmosphere. However the heating did not result from direct coupling of the
material with the microwaves but it occurs by transfer from the plasma formed around the
compact.

Introduction

Microwave sintering has proved its applicability on ceramics [1] [2] [3] and hardmetals
[4] [5] [6]. An unexpected use of microwaves is the sintering of metal powders. Bulk metals
reflect microwaves but Cheng and al. [7] [8] showed in a single-mode how various materials
including metallic compacts can be heated up in electric field or magnetic field. These
authors proved that powdered metal compacts could absorb microwaves, especially though
the magnetic field, at room temperature and above. However the reached temperature was
too low (< 800C) to lead to full sintering of the metal compacts. Others authors used
multimode or single-mode cavity including susceptors, so that the sintering can be
considered as indirect or hybrid. These studies have demonstrated that in most cases the
mechanical properties of microwave sintered parts are improved in comparison with
conventionally sintered ones [9] [10]. Noteworthy is the study by Saitou [11] about indirect
microwave sintering of metal powders (including nickel) in a single-mode cavity. Saitou
concluded that microwave radiation promoted the sintering without changing the main
sintering mechanism (lattice diffusion in nickel).
Anyway the effects of microwaves on the sintering of metal powders are poorly understood.
As far as we know, nobody reported any result about direct microwave sintering of nickel
powders in a single-mode cavity. The aim of this paper is thus to investigate the possibility of
sintering a micronic nickel powder in a microwave single-mode cavity by direct heating i.e.
without any susceptor in E-field or H-field.

Experiments

The experimental setup is schematized in Figure 1. It includes a high voltage power
supply linked to a magnetron that delivers a variable forward power up to 2 kW at 2.45 GHz.
A rectangular wave-guide of section 86.36 x 43.18 mm allows for the transport of the
microwave radiation to a rectangular TE
10p
cavity. This resonant cavity is ended by a
coupling iris (a vertical slot in a copper sheet) on the magnetron side and by a reflector, also
called sliding short-circuit piston, on the other side.
Samples are set upon a quartz plate inside a vertical 40 mm-diameter quartz tube that goes
across the cavity through two holes. To reduce microwaves escape, a 93 mm-high chimney
is fixed to the cavity at each hole. The quartz tube can be filled with various sintering
atmospheres. During processing, the temperature of the upper surface of the specimen is
continuously measured by a thermal imaging camera located above the center of the cavity
and protected by a ZnSe window.
During microwave heating, the forward electromagnetic wave is reflected against the
conductive wall of the short-circuit piston, leading to a standing wave made of a single
stacking of forward and reflected waves. In the empty cavity, the points where the magnetic
and electric fields are respectively maximum and minimum (near to zero) are exactly distant
from each other by a quarter of the guided wave-length (43 mm). Besides, when the iris-
piston distance (cavity length) is equal to 4 or 5 times the half guided wave-length, the iris
reflects almost entirely the wave in the cavity (TE
104
or TE
105
resonant mode). The resonance
phenomenon consists in the stacking of multiple waves reflected against the piston and the
iris, which results in an increase of the electromagnetic energy in the cavity. In our cavity, the
TE
104
and TE
105
resonant modes are used for materials interacting respectively with the
magnetic field or the electric field. Ideally, the iris-specimen distance should be chosen to be
about equal to half of the cavity length leading to a resonant mode. That allows the best
coupling between the material and the electric field at resonance (maximum dissipated
power). However we are aware that the simple electric and magnetic patterns expected in an
ideal close and empty cavity are significantly perturbated by the chimneys and the
sintering compact. In particular it is most likely that the compact never undergoes pure
magnetic field nor electric field. Anyway, we assume that in TE
104
(respectively TE
105
)
configuration the magnetic field (respectively electric) is predominant.

Figure 1. Schematic of the microwave furnace
The raw material is an Inco type 110 PM nickel powder made of extra fine spherical particles
produced by Inco carbonyl technology. This powder is tightly sized in the 1 - 2 m range
(Fisher Sub-Sieve Sizer). Green samples were prepared by uniaxial pressing at 100 or 600
MPa in a cylindrical die. The green densities obtained with these pressures were respectively
around 4.1 g/cm
3
and 5.5 g/cm
3
, i.e. 46 % and 62 % of theoretical density (TD : 8.908
g/cm
3
)). Green parts were sintered by direct microwave heating and by conventional sintering
in a vertical dilatometer. The bulk density of the sintered samples was measured by an
immersion technique in ethanol based on Archimede's principle. Sintered specimens were
cut in half cylinder with a diamond saw. The microstructure of polished and chemical etched
(Marble's reagent) sections were observed by scanning electron microscopy (SEM Back
Scattered Electron detector).

Results

- Conventional sintering:
Two nickel compacts uniaxially pressed at 100 MPa were sintered in a dilatometer at
10C/min up to 800C and 1000C in hydrogenated atmosphere (He-4vol%H
2
). The relative
density and the densification rate are plotted vs. the temperature in Figure 2. It appears that
the densification started at 240C and the maximum densification rate was at about 640C.
The final relative density were 92.3 % and 95.9% at respectively 800C and 1000C. Note
that the total processing time was about 150 minutes with this conventional route.

Figure 2. Changes of relative density and densification rate
during conventional sintering at 10C/min

- Microwave heating under magnetic field:
Green compacts were positionned first under magnetic field (TE
104
mode) in the
previously described resonant single-mode cavity. Thus the distance between the coupling
iris and the specimen was fixed to 165 mm The forward power was 1000 W and the position
of the piston was adjusted to be at the resonance, i.e. the specimen-piston distance was in
the 175 to 165 mm range. This allowed the best coupling with microwaves and the maximum
power dissipated in the cavity. Four different atmospheres (air, nitrogene, argon, helium-4%
hydrogen) have been tested. Table I presents miscellaneous data on these experiments. The
temperature has been measured on the upper surface of the specimen with the IR camera
assuming an emissivity factor = 0.9.

Test
label
Compaction
stress
(MPa)
Atmosphere Forward
power (W)
Maximum
temperature
(C)
Observations
M1 600 He + 4% H
2
1000 155 arcing
M2 600 Ar 1000 195 arcing
M3 600 N
2
1000 180 arcing
M4 600 Air 1000 190 arcing
Table I. Miscellaneous data on microwave heating experiments in magnetic field

Despite their ferromagnetism properties, nickel compacts did not not significantly heat in
magnetic field (maximum temperature less than 200C) and no densification was observed.
Besides, arcing phenomena occurred near to the coupling iris and in the wave-guide. Thus, it
is assumed that the microwave power was mainly dissipated in the cavity by arc discharging.

- Microwave heating under electric field:
Compacts were next heated under electric field (TE
105
mode). The positions of the
coupling iris and the piston were adjusted to get the maximum of the electric field in the
specimen region. The distance between the coupling iris and the specimen was fixed to 218
mm and the position of the piston was adjusted to be at the resonance, i.e. in the 222 to 218
mm range. The forward power was 1000 W or 500 W. Table II shows miscellaneous data on
the performed experiments. Three values of temperature are given: the average temperature
calculated by the IR camera on the upper surface of the specimen and the minimum and
maximum temperatures in this surface.

Test
label
Compaction
stress
(MPa)
Atmosphere Forward
power
(W)
Observations Ttemperature
(C)
Average
Min-Max
Final
relative
density
(%TD)
E1 600 Air 1000 arcing/plasma 840
730-1010
92.1
E2 600 N
2
1000 plasma 790
640-1170
92.3
E3 600 Ar 1000 plasma 350
250-490
63
E4 600 He + H
2

(4%)
1000 plasma 500
410-1000
98.4
E5 100 He + H
2

(4%)
1000 plasma 600
370-990
97.4
E6 100 He + H
2

(4%)
500 plasma 540
450-710
96.3
Table II. Miscellaneous data on microwave heating experiments in electric field

Finite element modelling of the microwave sintering of ceramic material [12] showed that the
electric field inside the compact is much lower than the electric field applied (forward power
delivered by the magnetron) and it partially gets around the part. That led to strong electric
field gradients localized above and below the specimen. This phenomena should be even
stronger with a metallic part. Such high electric field gradients could ignite plasma below
and/or above the metal part. In any case a plasma occurs around the compact when the
temperature was around 100C.
The nickel compact did not sinter in argon. We suppose that the plasma did not have enough
energy to heat up the specimen. Nevertheless, the colour of the compact changed locally:
the light-grey and matt colour of the surface turned into iridescent where plasma occurred.
Both in air and nitrogene, the material densified up to 92% and the surface of the specimens
turned into green. Moreover, the lower corner of the sintered parts was damaged by localized
plasma (Figure 3 and Figure 4). Besides, in air arcing occurs simultaneously with plasma and
slightly spoiled the upper surface of the specimen (spots).


Figure 3. Picture of the compact microwave sintered in
air (E1). The colour turns into green during sintering.
Surface damage and spots are caused
by plasma and arc discharging.

Figure 4. Image provided by the
thermal IR camera that illustrates
the plasma in the lower corner
of E1 specimen.

Finally, the compacts sintered in He-H
2
atmosphere led to the highest density (more than 96
%TD) and exhibited a metallic appearance. Plasma occurred again close to the upper
surface and local melting of nickel was observed during E4 and E5 tests with a forward
power of 1000 W, as shown in Figure 5 and Figure 6 . Apparently no melting happened
during E6 test with a reduced forward power of 500 W. In this case microwave plasma-
assisted sintering allowed for near full densification in less than 5 minutes.

Figure 5. Image provided by the IR camera showing
the upper surface of E5 compact
before (left) and during (right) local melting

Figure 6. Picture of the E5
green part (left) and E5
compact after microwave
sintering (right)

The microstructure was found to be homogeneous in spite of plasma except in melting region
where large grains appeared (Figure 7). The average grain grain size was in the 100-200 m
range (Figure 8).

Figure 7. SEM micrograph
of E5 polished section

Figure 8. Magnified SEM micrograph
of E5 polished section

Conclusions

Direct microwave heating of a micronic nickel powder was performed in a single-
mode cavity. Nickel compacts were heated under preponderant magnetic field or electric
field. Influence of the sintering atmosphere was also studied. The heating behaviour of the
compacts was unexpected. Despite its ferromagnetism properties, attempts to sinter nickel
parts under magnetic field failed. Estimated temperature with a thermal imaging camera
never passed 200C. For each test under electric field, plasma occured in the specimen
environment at low temperature (< 100C) due to high and localized electric field, especially
below and above the specimens. Thermal energy was transferred rapidly through the surface
facing plasma to the volume of the compacts. Although direct microwave heating could not
be performed because of plasma occurring, ultra-rapid heating and near full densification
took place by this way in a few minutes. Moreover, the forward power was adjusted to
prevent nickel parts from surface damages and local melting induced by plasma or arcing
phenomena. Nickel parts were sintered in He-H
2
atmosphere at density > 96%TD with good
surface appearance in shorter times than conventional sintering.

References

[1] T.T. MEEK, C.E. HOLCOMBE, N. DYKES, Microwave Sintering of Some Oxide Materials
using Sintering Aids, Journal of Materials Science Letters 6, 1060-1062 (1987).
[2] W.H. SUTTON, Microwave Processing of Ceramic Materials, Ceramic Bulletin, 68 [2],
376-386 (1989).
[3] D.E. CLARK, W.H. SUTTON, Microwave Processing of Materials, Annu. Rev. Mater. Sci.
26, 299-331 (1996).
[4] J.P. CHENG, D.K. AGRAWAL, S. KOMARNENI, M. MATHIS and R. ROY, Microwave
Processing of WC-Co Composites and Ferroic Titanates, Materials Research Innovation 1,
44 (1997).
[5] K. RDIGER, K. DREYER, T. GERDES, M. WILLERT-PORADA, Microwave Sintering of
Hardmetals, International Journal of Refractory Metals & Hard Materials 16, 409-416 (1998).
[6] E. BREVAL, J.P. CHENG, D.K. AGRAWAL, P. GIGL, M. DENNIS, R. ROY, A.J.
PAPWORTH, Comparison between Microwave and Conventional Sintering of WC/Co
Composites, Materials Science and Engineering A391 (1-2), 285-295 (2005).
[7] J. CHENG, R. ROY, D. AGRAWAL, Experimental Proof of Major Role of Magnetic Field
Losses in Microwave Heating of Metal and Metallic Composites, Journal of Materials Science
Letters 20, 1561-1563 (2001).
[8] J. CHENG, R. ROY, D. AGRAWAL, Radically Different Effects on Materials by Separated
Microwave Electric and Magnetic Fields, Mat. Res. Innovat. 5, 170-177 (2002).
[9] R. ROY, D. AGRAWAL, J. CHENG, S. GEDEVANISHVILI, Full Sintering of Powdered-
Metal Bodies in a Microwave Field, Nature, Vol. 399, 668-670 (1999).
[10] D.K. AGRAWAL, Microwave Sintering of Metallic Materials, Powd. Metall. 1, 797-800
(2000)
[11] K. SAITOU, Microwave Sintering of Iron, Cobalt, Nickel, Copper and Stainless Steel
Powders, Scripta Materialia 54, 875-879 (2006).
[12] D. BOUVARD, S. CHARMOND, P.C. CARRY, Finite Element Modelling of Microwave
Sintering, Ceram. Trans. vol. XX, xxx-xxx (2009) to be published.

Comparison of Thermal and Microwave-Assisted Plasma
Sintering of Nickel-Diamond Composites

Barry Twomey, Aidan Breen, Greg Byrne, Alan Hynes and Denis P. Dowling*

School of Electrical, Electronic and Mechanical Engineering, University College
Dublin, Belfield, Dublin 4, Ireland *(denis.dowling@ucd.ie)

Abstract
There is considerable interest in processing technologies which can lead to more
energy efficient sintering of metal powders. Microwave sintering has recently been
shown to reduce energy usage as the volumetric heating process is considerably
more efficient than resistance heating. RF plasma sintering meanwhile has been
shown to deliver heat via uniform excitation of the processing gas resulting in ion
bombardment of the workpiece. In this study the use of a rapid, novel microwave-
assisted plasma sintering (MaPS) technology for processing of nickel-diamond metal
matrix composites is evaluated. Nickel powder and polycrystalline diamond were
mixed to prepare 20 mm discs under uniaxial compaction pressures of 100, 200 and
300 MPa. The discs were fired in a low pressure microwave plasma under a
hydrogen atmosphere. For comparison, discs were also sintered using conventional
tube furnace firing. The MaPS sintering is very rapid with full disc strength of
>1000N, based on 3-point bend tests, being achieved within 10 minutes compared
with 8 hours for furnace treatment. This study demonstrates that the microwave-
assisted plasma sintered discs produced similar or superior performance to discs
fired using furnace firing conditions but with sintering cycle time reduced by up to
95%.
Introduction
Plasma sintering has been demonstrated as a rapid and uniform sintering technique
since its first reported studies by Bennett et al. [1]. Sintering has been carried out in
plasmas generated using microwave, rf or DC power supplies [1-3] and the dominant
heating mechanism is associated with ion and fast neutral bombardment. Diatomic
gases such as hydrogen or nitrogen are typically used for plasma sintering. Plasma
sintering is carried out under vacuum although unlike spark plasma sintering (SPS),
there is no external force exerted on the workpiece [4]. There have also been a large
number of reports in the literature on the microwave interaction with materials
including the type of processing microwave heating can offer [5, 6]. The ability to
penetrate the surface of the workpiece enables rapid volumetric heating in
microwave processing, reducing the need for external heat sources [6]. Advantages
of non-plasma microwave processing over furnace treatment include finer grain
sizes, rounded porosity and higher ductility and toughness [7]. Increased shrinkage
rates and decreased grain sizes can be achieved with microwave and plasma
processing compared with furnace sintering [1, 8, 9]. With these advantages and
significantly reduced cycle times, microwave-assisted plasma sintering (MaPS) offers
an alternative route to production and reduced energy consumption per process.
This study compares, for the first time, furnace and microwave assisted plasma
sintering of diamond metal matrix composites (MMCs). Nickel powder was chosen as
the test material as it is used in a wide range of engineering components due to its
corrosion resistance, wear resistance, mechanical strength, thermal expansion,
electrical conductivity and magnetic permeability [10]. Nickel powder is also used as
a metal binder for many bonding applications in diamond and carbide machining
tools [11].
Experimental Methods
Nickel-diamond MMCs were pressed in a uniaxial 20 mm diameter die at pressures
of 100, 200 and 300 MPa (average green densities were 52, 58 and 62 %
respectively). INCO nickel (T110) and nickel coated diamond (6-12 m) from
Element Six were mixed by weights of 80 and 20 % respectively in a T2F-Turbula for
100 minutes prior to pressing. Density was determined using the Archimedes
principle in mercury and confirmed by measuring volume based on sample
dimensions.
The microwave assisted plasma sintering process was carried out using a
Circumferencial Antenna Plasma (CAP) microwave system described in more detail
elsewhere [12]. The plasma was formed at a pressure of 20 mbar in a hydrogen
atmosphere. Three disc samples were rotated in a plasma located at the centre of
the chamber per run (see Figure 1). Input powers of 2.4 kW were supplied from a
Mugge microwave power supply operating at 2.45 GHz. Sample temperatures were
measured using a LASCON QP003 two-colour pyrometer from Dr. Merganthaler
GmBH & Co. The use of a two-colour pyrometer is expected to eliminate the
interference effect of the plasma on the emissivity of the sample. Typical heating and
cooling rates of 7
o
C/s were observed in the microwave plasma fired samples.
Emission spectroscopy was also carried out on the plasma using an Ocean Optics
USB4000 spectrometer. Measurements were carried out to investigate changes in
the electron density and total light emission of the plasma with increasing applied
power. Furnace sintered samples were treated in tube furnace at 850
o
C in a flowing
argon atmosphere. Heating rates of 4
o
C/min were used with the same dwell time of
10 minutes at the maximum temperature. The cooling rate was typically 2-3
o
C/min.

Figure 1: Samples (two visible) being sintered in a hydrogen microwave
plasma
Flexural stress tests were carried out using a three point bend test. The span
between the bottom pins (11 mm) and the cross sectional area of the fractured
samples were used to calculate approximate flexural stress with an average value of
three samples taken. Hardness testing was carried out using a Rockwell diamond
indenter (HR 15 N scale), with the average value of 12 measurements taken.
Rockwell indentation was chosen as the large indent area ( 1 mm in diameter)
should limit the effect of porosity or diamond reinforcement ( 20 m). Pin-on-disc
testing was undertaken using a TEER coatings POD-2 tester, with a 5 mm tungsten
carbide ball as the pin, to evaluate the wear and tribological behaviour of the sintered
samples [13]. The pin was loaded against the treated samples at 10 N at a linear
speed of 5 cm/s for 1200 revolutions.
Results and discussion
Sintering conditions
Nickel-diamond samples pressed at 100, 200 and 300 MPa were sintered at 2.4 kW
applied power in the microwave assisted hydrogen plasma. Following a two minute
ramp up to an applied power of 2.4 kW, a maximum firing temperature of 850
o
C was
achieved. Samples were sintered in sets of three with a 10 minute dwell time. A total
cycle time of 20 minutes including pump down, firing and cooling was required. An
increase in applied power resulted in an increase in plasma emission intensity
(Figure 2). The peak shift and full width half maximum (FWHM) of the H
and H
, at
656.3 and 486.1 nm respectively, indicate the electron density in the plasma [14]. As
the applied power increases from 1.2 to 2.4 kW, no increase in the normalised
FWHM was observed and the ratio of the two peaks (H
/ H
) remained almost
constant at 5.3. This indicates there is little or no increase in electron density with
increasing applied power although the overall emission doubled as a result of a
volume increase of the plasma. Furnace sintered samples were treated in a tube
furnace at 850
o
C in batches of three in a flowing argon atmosphere. A total cycle
time of over 8 hours, including heat up and cool down times, was observed.

Figure 2: Emission intensity observed in hydrogen plasma
Mechanical behaviour of sintered MMCs
Density measurements were carried out using the Archimedes principle and also
calculated based on the sample dimensions. As illustrated in Figure 3, there is little
difference in sample density from the two sintering techniques despite a significant
difference in total treatment times (20 minutes versus 8 hours). Sintering at these
temperatures resulted in a 15 % density increase for both plasma and furnace
treated samples.

Figure 3: Density of pressed and sintered samples
Flexural stress values (see Figure 4) were determined using a three-point bend test.
Due to the similarity in the cross-sectional area and densities achieved between the
both sintering techniques, it was not unexpected that there was little difference
observed between the two techniques. Samples pressed at higher pressures (300
MPa) performed better when microwave plasma treated and samples pressed at
lower pressures (100 and 200 MPa) performing better when furnace treated.
As shown in Figure 4 however, microwave plasma fired samples exhibited
significantly better hardness (34 %) then the furnace fired samples. As previously
reported for microwave treated samples versus furnace treated samples, thermal
equilibrium occurs almost instantaneously enabling sintering to occur in minutes [15].

Figure 4: Flexural stress (left) and hardness values (right) of microwave
assisted plasma and furnace sintered samples pressed at 100, 200 and 300
MPa

Examining the cross-sectional area of the nickel-diamond composites (Figure 5), it
can be seen that the MaPS treated samples exhibit a much finer microstructure with
increased nickel bonding between diamond particles (black areas) rather then
increased flow into larger masses. The formation of larger masses of nickel may lead
to larger pores in the furnace treated samples resulting in reduced hardness. This is
most likely due to the heating rate of 420
o
C/min versus 4
o
C/min for MaPS over
furnace treatment. These heating and equivalent cooling rates can be achieved as
the high energy plasma discharge is isolated from the colder surroundings ( 30
o
C)
of the discharge chamber. The finer microstructure of the MaPS treated samples may
explain the increased hardness values observed [7].

Figure 5: Cross-sectional microstructure of 300MPa pressed MaPS (left) and
furnace (right) sintered
Pin on disc wear testing was undertaken on each of the samples to determine the
wear performance of samples pressed at increasing pressures. The wear track
generated was examined using SEM. The wear track itself appeared to be masked in
a foreign material, as shown in Figure 6. Using EDX elemental analysis a large
presence of tungsten in the wear track was detected with a maximum concentration
of 67 % in wear track centre. This result demonstrated that there was significant wear
of the tungsten carbide pin. Due to this wear of the pin, direct comparison of the wear
tracks was not possible.

Figure 6: EDX analysis (left) and SEM image (right) of pin-on-disc wear track
The wear rate of the pin was measured using an optical microscope and ImageJ
analysis software [16]. For each of the sample compaction pressures (100, 200 and
300 MPa), it was observed that microwave plasma sintered samples caused an
increased level of pin wear compared to furnace sintered samples (3, 19 and 41 %
increase respectively). This is not unexpected given the increased surface hardness
of plasma sintered samples. However, samples pressed at highest pressures caused
lower levels of pin wear compared with those pressed at the lower pressures for both
firing techniques. This may be due to increased diamond pull out due to the
decreased density of the 100 MPa pressed samples. As the diamond is removed, an
increased level of abrasion may occur due to the abrasive debris rather then the
diamond containing MMC disk. To examine this requires further testing using a
different pin materials or disc-on-disc wear testing.
Conclusions
Nickel-diamond MMCs have been produced using microwave assisted plasma
sintering (MaPS) and compared with composites produced using furnace sintering.
As the plasma is isolated from the colder walls of the vacuum chamber, rapid heat
up and quenching rates can be achieved resulting in a treatment time reduction of
95%. Despite the significantly reduced treatment times comparable sample densities
were observed in samples from the two techniques. Similar breaking strengths were
also observed for both furnace and plasma fired MMCs although plasma sintered
samples exhibited increased hardness values, most likely due to the finer
microstructure of the MaPS samples. MaPS treated MMCs caused greater tungsten
carbide pin wear during pin-on-disc testing, while samples pressed at lower
pressures tended to exhibit higher pin wear rates.
References
1. C.E.G. Bennett, N.A. McKinnon, and L.S. Williams, Nature, 1968, 217(5135),
p.1287-1288.
2. D. Lynn Johnson, W. Sanderson, J. Knowlton, E. Kemer, and M.-Y. Chen,
Science of Sintering, 1988, 20(2/3), p.109-113.
3. J.M. Loureno, A.M. Maliska, A.N. Klein, and J.L.R. Muzart, Materials
Research, 2004, 7(2), p.269-275.
4. P. Alvaredo, E. Gordo, K. Vanmeensel, and O. Van der Biest, 17th Plansee
Seminar, 2009, Reutte, Austria, p.GT 41/1
5. Y.V. Bykov, K.I. Rybakov, and V.E. Semenov, Journal of Physics D: Applied
Physics, 2001, (13), p.R55.
6. D.K. Agrawal, Current Opinion in Solid State and Materials Science, 1998,
3(5), p.480-485.
7. R. Roy, D. Agrawal, J. Cheng, and S. Gedevanishvili, Nature, 1999,
399(6737), p.668-670.
8. R. Wroe, Metal Powder Report, 1999, 54((7/8)), p.24-28.
9. K. Saitou, Scripta Materialia, 2006, 54, p.875-79.
10. A.D. Jatkar, V.A. Tracey, D.O. Gothard, and J.J. Fischer, ASM Handbook, 7:
Powder Metallurgy, 1993, ASM International, p.395-400.
11. D. Egan and J.A. Engels, Industrial Diamond Review, 2004, 04(4), p.36-38.
12. M.L. McConnell, D.P. Dowling, C. Pope, K. Donnelly, A.G. Ryder, and G.M.
OConnor, Diamond and related materials, 2002, 11, p.1036-1040.
13. E. Arslan, Y. Totik, E. Demirci, and A. Alsaran, "Influence of Surface
Roughness on Corrosion and Tribological Behavior of CP-Ti After Thermal
Oxidation Treatment", Journal of Materials Engineering and Performance,
2009, DOI: 10.1007/s11665-009-9504-9
14. J. Ashkenazy, R. Kipper, and M. Caner, Physical Review A, 1991, 43(10),
p.5568.
15. E. Breval, J.P. Cheng, D.K. Agrawal, P. Gigl, M. Dennis, R. Roy, and A.J.
Papworth, Materials Science and Engineering A, 2005, 391(1-2), p.285-295.
16. M.D. Abramoff, P.J. Magelhaes, and S.J. Ram, Biophotonics International,
2004, 11(7), p.36-42.


Microwave Rapid Sintering of Nanostructured Stainless Steel Metal
Powders

P. Veronesi
*
, C.Leonelli
*
, G. Poli
*
, E. Bassoli, L. Denti, A. Gatto

*
Dept. of Materials and Environmental Engineering, University of Modena, Via Vignolese 905,
41100, Italy
Dept. of Mechanical Engineering, University of Modena, Via Vignolese 905, 41100, Italy

Abstract
Nanostructured AISI430 stainless steel powders, obtained by high energy milling have been
uniaxially pressed in order to obtain 8 mm diameter cylindrical compacts. Microwave sintering of
the green compacts has been performed in a single mode microwave applicator, at 2.45 GHz.
Maximum sintering temperature of each sample, monitored using a sapphire optical fibre, was
varied in the 1000-1200C range, with heating rates in the 20-30 K/s range, applying 1200 W of
microwave forward power. Dense nanostructured samples, with less than 1% porosity were
obtained, presenting small local variation of Cr content in proximity of the necking region. Rapid
microwave sintering helps maintaining the nanostructure, despite a non homogenous densification
due to the uneven electromagnetic field distribution in the single-mode applicator

KEYWORDS: Microwave sintering, nanostructure, AISI430

Introduction
Maintaining the nanostructure in materials subjected to thermal treatments, like sintering, is a
demanding task, since grain growth has to be prevented. As a matter of fact, the large surface area
of nanopowders and nanostructered materials provides a strong driving force for both sintering and
grain growth processes. Different sintering techniques have been studied to preserve the
nanoscale structure in the fully sintered product [1]. With this respect, Field Assisted Sintering
Techniques (usually indicated by the evocative acronym FAST), like Spark Plasma Sintering-SPS
and Microwave sintering, present the advantage of rapid processing time, which inhibits diffusion
phenomena [2].
As a mater of fact, FAST is an emerging technology for the fabrication of metals, ceramics and
their composites starting from powders. It has the potential of densifying powders with nanosize or
nanostructure while avoiding the coarsening which accompanies standard densification routes.
Recently [3], ultra fine grained and nanostructured iron alloys have been produced by SPS,
showing the effects of pre-milling powders, sintering temperature and pressure. However, SPS
presents also some drawback, as shown by Kodash et al. [4], regarding Ni nanopowders and
studying the influence of heating rate on densification and final grain structure. Results
demonstrated that there seem to be a upper limit to the applicable heating rate, since very high
heating rates resulted in non-uniform densification of the samples and formation of cracks during
sintering. This was ascribed to the intense current application, which hinders particles gliding, thus
decreasing the rate of densification. Moreover, the powders subjected to SPS experience a
different thermal history and pressure levels, which are detrimental for homogenization of density
and final properties [5]. The nature of thermal and electrical gradients during the SPS version of
FAST was investigated by numerical simulation by Zavaliangos [6], showing that the contact of the
powders with the die and the applied pressure determine the lack of homogeneity. Anselmi-
Tamburini et al. [7] developed an extensive model demonstrating the role of the graphite die in
determining sintering unhomogeneities, also confirmed by experimental evidence.
The microwave sintering version of FAST does not necessarily require the physical contact with
the green powders to generate heat, thus some of the lacks of homogeneities could be avoided by
this technique. Nevertheless, even during pressureless microwave sintering, usually powder
compacts are supported by refractory materials and are exposed to the furnace ambient, which
can alter the temperature distribution in the material, as shown by the authors in a different
application of microwave heating to synthesise intermetallic compounds [8]. However, a proper
choice of supporting refractory materials and the use of microwave auxiliary absorbers can help
reducing thermal gradients, leading to a much more homogenous dense material [9]. Moreover,
previous studies by some of the authors demonstrated also that breakdown phenomena can occur
during microwave sintering of metals at high electric field intensity [10]. This implies that there can
be also a transient lack of homogeneity at the micrometric scale, but this phenomenon could be
positively exploited if sintering involves micrometric nanostructured powders instead of nanosized
ones.
Thus, aim of the present work is the investigation of the microwave assisted sintering of
nanostructured AISI 430 powders under applied high electric field intensity, in order to verify if the
nanostructure can be preserved and if a satisfactory sample homogeneity can be achieved.

Experimental
Nanostructured powders of AISI 430 stainless steel (MBN Nanomaterialia SpA, Italy) obtained by a
proprietary high energy milling process (Mechanomade) were used as starting materials.
Average size of the nano-domains is 20nm (figure 1a), while sieving allowed to select only the
powders with particle size is below 250 m. Powders are irregularly shaped, as a result of the high
energy milling process, as shown in figure 1b.

a) b)

Figure 1 Starting AISI430 stainless steel powders: a) FIB micrograph showing nanocrystalline domains in
a single particle ; b) SEM micrograph of the starting nanostructured powders

The powders status (strained by repeated deformation induced by the collision with the milling
media) and shape is expected to pose relevant problems during forming. Thus, uniaxial pressing at
550MPa in lubricated die was used to prepare 1 g cylindrical samples having 8 mm diameter. The
green parts were placed, one per run, into an alumina fibre lining (Ultraboard
, M.E. Schupp,
Germany) presenting a 2 mm diameter hole in order to perform temperature measurement by
means of a sapphire fibre connected to a Mikron signal conditioner (temperature range: 600-
1800C). This setup allows to have a minimum electromagnetic field perturbation when inserted in
the microwave applicator, since the materials are non conductive and present negligible loss
factors and dielectric constants close to 2.
A TE
10n
(n= 3 or 4 depending on the short circuit position) microwave single mode applicator was
used to perform the sintering tests. The applicator is fed by a TM030 3 kW magnetron operating at
2.45 GHz, connected to a SM1150 switching power supply (Alter, Italy) in order to set the emitted
power at 1200 W level. Microwave generator and applicator are connected by means of a
rectangular WR340 waveguide, along which impedance matching (3-stub tuner) and reflected
power measurement devices (directional coupler) are installed. The microwave source is protected
by reflected power by means of a 3-ports isolator, water cooled. The microwave applicator is
terminated by a movable short circuit, which allows to alter the electromagnetic field distribution in
the applicator (position of maximum electric field intensity with respect to the sample). The electric

(E) and magnetic (H) field envelopes qualitative distributions in the empty microwave applicator,
depending on the short circuit position, are shown in Figure 2. Samples were placed, one per run,
in the region intermediate between maximum electric field (centre of the applicator) and maximum
magnetic field (side of the applicator) intensity and subjected to a 20 Nml/min N
2
-2%H
2
flux to
prevent oxidation.

a) b)

Figure 2 Electric (E) and magnetic (H) field envelope in the microwave single mode applicator depending
on the short circuit position: a) TE
103
mode and b) TE
104
mode

The single mode applicator and the sample position, intermediate between the two maxima of field
intensity, was chosen so that not only direct microwave heating occurs, but also breakdown
phenomena triggered by the high electric field can take place between the powders. This is
expected to lead to early necks formation and overheating limited preferentially to the particles
surface, in order to try to prevent the bulk nanostructure coarsening.
After sintering, the samples were cut and selected cross sections were metallographically prepared
for optical microscope observation. SEM-EDS observations (QUANTA 200, FEI Co.) were
conducted in order to evidence possible compositional changes. Focused Ion Beam (DB Strata
235 M, FEI Co.) was used to nano-machine the samples surfaces and assess the nanostructure
retention after sintering. Vickers micro-hardness using 1 kg load was measured using a Wolpert
Microhardnes Tester, in order to verify the expected hardness increase due to grain size
refinement.

Figure 3 shows the typical heating curve measured in case of samples sintered at 1150C. The
signal of the measured temperature seems to indicate that the heating rate is almost constant
throughout the whole duration of microwave exposure (from time= 3 s to time=48 s). However,
since the optical pyrometer readings start from 600C, it should be considered that during the first
5 seconds (from time= 3 s to time= 8 s) a temperature increase of almost 600C is experienced by
the material, which then starts to heat less rapidly. A possible explanation of this phenomenon
could be that the powders reached the Curie temperature (which, for AISI 430 steel is
approximately 610C [11]), thus losing their ferromagnetic properties. This plays an important role
in the microwave-matter interaction, since the power density generated in a material can be
expressed as:

P
d
(x,y,z) "
eff
E
rms
2
+ "
eff
H
rms

2
(1)
where:
P
d
= power density in the material (W/m
3
), at the position (x,y,z)
f (Hz), f = frequency of the incident microwaves
eff
= effective loss factor, including conductivity losses,
"
eff
= imaginary part of the effective magnetic permeability
E
rms
= local (x,y,z) electric field intensity (V/m),
H
rms
= local (x,y,z) magnetic field intensity (A/m),

Thus, the rapid decrease of the magnetic permeability above the Curie point leads to a much lower
microwave power density in the material, which thus experiences a lower heating rate.

Figure 3 Heating and cooling curve of the samples sintered at 1150C

Optical microscopy of polished cross sections, coupled with image analysis, allowed to measure
the samples porosity, which for temperatures in excess of 1150C is lower than 1%, as shown in
Figure 4. For all the studied sintering temperatures, porosity is always round shaped, as shown in
Figure 5a, with roundness values in excess of 0.9

0
100
200
300
400
500
600
700
1050 1100 1150 1200 1250
T (C)

H
V
1
0
0,5
1
1,5
2
2,5
3
3,5
4
4,5
5
P
o
r
o
s
i
t
y

(
%
)
HV1
P%

Figure 4 Porosity and Vickers microhardness of the microwave sintered samples as a function of
maximum temperature

Vickers micro-hardness values of the sintered pieces resulted much higher than the typical
annealed hardness of AISI 430 steels, indicating that a possible grain size refinement
strengthening occurred. Average porosity tends to decrease as sintering temperature increases
from 1100 to 1150C, then it stabilizes at values lower than 1%, within the experimental error. A
possible explanation is that due to the very low sintering time and no holding time nor applied
pressure, the temperature increase is expected to lead to an increase of diffusion phenomena,
thus favouring densification. According to the data of Figure 4, hardness increases as temperature
T (C)
increases, indicating that probably some concurrent reaction takes place as sintering proceeds. As
a matter of fact, the samples obtained at high temperature should present a larger grain size, and
thus a lower micro-hardness, for a given porosity level.

a) b)

Figure 5 - a) SEM micrograph (BSE) of a sample sintered at 1150C, showing round shaped porosity and
composition heterogeneities; b) EDS spectra and quantitative weight analysis of Cr content in the region
indicated by 1 and 2.

Figure 5b shows that the sample is not homogenous and that there is the presence of Cr-richer
regions, which correspond roughly to the former particles borders. This, and the fact that no
debinding has been performed on the green parts prior to sintering, despite the use of lubricated
die, could be ascribed to the formation of chromium carbide at the grain boundaries, where
localised heating occurred and the lubricant acted as a source of carbon, leading to a slight
sensitisation of the stainless steel, as confirmed by previous studies [12]. Moreover, considering
both the microstructural features and the micro-hardness values reported in Figure 4, which
evidence a pronounced scattering, it can be deduced that the sample homogeneity is not
completely achieved. However, this result is somehow expected, since the samples are lying on a
much colder alumina fibre lining and the electromagnetic field distribution in the applicator region
where the samples are placed is not constant. Thus, power generation occurs at different levels,
according to equation (1), and the presence of the colder lining induces higher heat losses from the
surfaces than from the bulk material. Macroscopically, this led to a lower densification of the outer
layers of the sample, which could be probably avoided if microwave heating homogeneity is
increased. Micro-hardness testing on polished cross sections is currently being applied to quantify
the extent of heating unhomogeneity
Possible solutions, currently under investigation, are the introduction of auxiliary microwave
absorbers, in order to eliminate temperature gradients due to the colder environment, and the
movement of the sample inside the applicator, to statistically homogenize the electromagnetic field
intensity. However, the experimental setup allowed to process only one sample at the time, having
small dimensions, while it is more industrially relevant having the possibility of sintering many
larger parts in a single run. This can be achieved changing applicator, i.e. using a multi-mode
applicator. A first series of tests conducted with the same power levels of the single mode
applicator, but in a multi-mode applicator, however, caused a significant increase of processing
time, which led to an unacceptably high grain growth, as shown in Figure 6. This could be ascribed
to the much lower electromagnetic field intensity in the multimode applicator, which can be of
orders of magnitude lower than in the single mode one.

Conclusion
Nanostructured AISI430 stainless steel powders, obtained by high energy milling have been
successfully sintered in less than a minute in a single mode microwave applicator, at 2.45 GHz.
Heating rate resulted significantly higher in the first seconds of microwave processing, with values
Spectrum 1: Cr= 16.89%
Spectrum 2: Cr= 15.45%
1
2
up to 150 K/s, then drastically lowering when the temperature reached the Curie point of the
material. Samples sintered at 1150C or above, applying 1200 W of forward microwave power,
have less than 1% porosity and a microhardness much higher than microcrystalline AISI 430
Stainless steel. Chromium-richer zones are formed in correspondence of the necking zones,
probably due to sensitization and to localised overheating triggered by the high intensity
electromagnetic field. Rapid microwave sintering, when conducted in the more suitable microwave-
matter interaction conditions helps maintaining the nanostructure, despite a non homogenous
densification due to the uneven electromagnetic field and temperature distribution in the single-
mode applicator. A first series of tests conducted in a multimode microwave applicator allowed to
reduce heterogeneities, but grain coarsening occurred to a much larger extent.

Figure 6 FIB image of the sample sintered in multi-mode applicator, showing sub-micrometric or
micrometric grains. A round-shaped pore (black) is also present, 3 m below the sample surface.

References
[1] J.R. Groza, Nanocrystalline powder consolidation methods, in: C.C.Koch (Ed.), Nanostructured
Materials, Noyes Publications, William Andrew Publishers, NY, 2002, pp. 115178.
[2] P. Angerer, L. G. Yu, K. A. Khor, and G. Krumpel, Spark-plasma-sintering (SPS) of nanostructured
and submicron titanium oxide powders, Materials Science and Engineering A
Volume 381, Issues 1-2, 2004, pp. 16-19
[3] S. Libardi, M. Zadra, F. Casari, A. Molinari, Mechanical properteis of nanostructured and ultrafine-
grained iron alloys produced by spark plasma sintering of ball milled powders, Materials Science and
Engineering A, 478, 2008, 243-250.
[4] V.Y. Kodash, J.R. Groza, K.C. Cho, B.R. Klotz, R.J. Dowding, Field-assisted sintering of Ni
nanopowders. Materials Science and Engineering A (385), 2007, pp. 367 - 371
[5] B.McWilliams, A. Zavaliangos, Multi-phenomena simulation of electric field assisted sintering,
Journal of Materials Science Volume 43, Issue 14, 2008, pp. 5031-5035
[6] A. Zavaliangos, J. Zhang, M. Krammer, and J. Groza, Temperature Evolution During Field Activated
Sintering, Mat. Sci. and Eng. A, 2004. pp. 379.
[7] U. Anselmi-Tamburini, S. Gennari, J.E. Garay, and Z.A. Munir, Fundamental investigations on the
spark plasma sintering/synthesis process II. Modeling of current and temperature distributions. Mat. Sci.
and Eng. A, vol 394, 2005, pp.139-148.
[8] G. Poli, R. Sola, P. Veronesi, Microwave-assisted combustion synthesis of NiAl intermetallics in a
single mode applicator: Modeling and optimisation, Materials Science and Engineering: A
Volume 441, Issues 1-2, 2006, pp. 149-156
[9] C. Leonelli, P. Veronesi, L. Denti, A. Gatto, L. Iuliano, Microwave assisted sintering of green metal
parts, Journal of Materials Processing Technology, Volume 205, 1-3, 2008, pp. 489-496
[10] P. Veronesi, C. Leonelli, G. Poli, Microwave enhancement of the early stages of sintering of
metallic powder compacts and metal-containing composites, Microwaves in the engineering and
applied science, E. Caponetti Editor, ENEA, Italy, 2007, pp. 125-138
[11] CRC Handbook of Chemistry and Ph sics,63rd edition, edited by R. C. Weast (CRC Press, 1982)
[12] P. Veronesi, C. Leonelli, L. Denti, A. Gatto, L. Iuliano, Preliminary studies of the rapid microwave
sintering of green parts made of 420L stainless steel, EUROPM2006 Proceedings vol2 - powder
injection moulding, 2006, pp.183-188

Cu Dispersion in Microwave Sintered Iron Compacts

I.M.Kusoglu
(a,b)
, K.Onel
(a)
, M.Willert-Porada
(b)

(a) Department of Metallurgical and Materials Engineering, University of Dokuz Eyll,
Izmir,Turkey
(b) Chair of Materials Processing, University of Bayreuth, Bayreuth, Germany

Abstract
Gas atomized Cu powder is added to Iron powder in varying amounts and mixed in a V-type
mixer for 30 minutes at 15 rpm. As lubricant 1 wt% Acrawax C is used. The mixtures are
compacted at 600 MPa to form TRS (Transverse Rupture Strength) samples with dimensions
6,35*12,8*31,8 mm. The samples are sintered using conventional and microwave furnaces for 30
minutes at 1150
o
C under 90% N
2
-10% H
2
atmosphere. The microstructure of sintered
specimens and Cu dispersion in Fe are observed by SEM and EDS, and the strength is
determined by TRS tests.

Introduction
Powder Metallurgy (PM) is a processing technique that involves the production of metal powders
and conversion of these powders into useful engineering components. Iron is one of the
fundamental elements to build these engineered structures. The material is useful for relatively
low tensile strength, but high ductility requirements. Sintering of compacted powders is one of the
most critical and complicated process in PM so it has to be well understood for densification
behaviour of iron compacts. Several studies have been carried out to understand the mechanism
of sintering iron compacts [1]. Sintering plain iron compacts can be carried out in a high
temperature furnace under inert atmosphere depending on whether the temperature is less than
or greater than 1150 C. Reducing the cost of sintering and saving energy is important in
sintering process because of its greatest portion in production cycle. Nowadays, microwave (MW)
energy is a popular research topic in sintering process of PM parts. Research on MW energy has
shown that it can be used in sintering of PM products [1-7]. It reduces sintering time and saves
energy. On the other hand many reports claim that MW sintered materials have more favorable
microstructures and mechanical properties [6-9]. The aim of this work is to investigate the effect
of microwave energy on 2, 3, 4, 8, 10 and 12 wt. % Cu dispersion in Fe compacts. Microstructure
and mechanical properties of compacted specimens are investigated by scanning electron
microscopy (SEM), energy dispersive spectroscopy (EDS) and TRS tests.

Experimental procedure
Iron powders (99,9%, d
50
= 87m, Hgenas) used in the experiments are mixed with 2, 3, 4, 8, 10,
12 wt. % gas atomized Cu powders (d
50
= 32m, Sentes-Bir, Izmir, Turkey) and 1 wt.% Acrawax C
as a lubricant in a V-type mixer at 15 rpm for 30 min. Mixed powders are compacted at 600 MPa
to form Transverse Rupture Strength (TRS) specimens with dimensions 6.35*12,8*31,8 mm. The
green compacts have been sintered by using microwave heating and conventional electric
furnace heating. The heating rate was 10C/min for conventional sintering (CS) and 20-30 C/min
for microwave sintering (MWS). CS has been carried out in an quartz tube and MWS has been
carried out in a batch type Hot Wall Resonator Molybdenum furnace. The furnace is equipped
with 9 kW output power microwave heating at a frequency of 2.45 GHz, fed from two rectangular
wave guides. The specimens were kept at 150 C for drying the thermal insulation materials prior
to sintering. The specimens were placed in a transparent thermal insulation casket inside the 0,2
m
3
Mo chamber as shown in Figure 1. AlN which is placed on a SiC plate is used as a specimen
holder. A pyrometer (Keller PZ20 AF1, 200-2000C) is used for temperature measurements. In
each sintering cycle 15 samples are placed in the furnace. Temperature correction of pyrometer
was carried out for each sintering cycle by using silver and copper pieces. The specimens
sintered at 1150C, for conventional sintering (CS) and microwave sintering (MWS). A dwell time
of 15 minutes applied for delubrication in CS. Sintering of specimens was carried out under 90%
N
2
- 10% H
2
atmosphere with a dwell time of 30 minutes for both processes. The morphologies
and copper dispersions in iron were investigated by SEM and EDS. Porosity levels of sintered
samples were measured by mercury porosimeter. Transverse rupture strength (TRS)
measurements were carried out according to ASTM B 528. The diameter of supports and upper
punch is 3 mm. During TRS measurements a compressive force is applied parallel to the
pressing direction of the specimen at 2,54 mm/min until rupture occurs. TRS values are
calculated according to the equation given below (1), where P is max load when rupture occurred;
L is the length between supports which is 25.4 mm; t is the thickness of the specimen and w is
the width of the specimen. All measurements are the average of 3 specimens for each sintering
parameter.

TRS = 3 x P x L / 2 x w x t
2
(1)

Fig. 1. Schematic description of the microwave sintering set up.

Results and discussions
Morphologies of each composition have been examined using SEM and EDS (Fig 3). According
to Fig. 3, pore structures and distributions are different for each composition. It is seen that up to
8 wt. % Cu addition, EDS analysis for copper dispersion is illustrated under SEM images for each
composition and sintering processes. White dots in EDS images show copper dispersion in the
cross-section of specimens. It has been observed that copper is homogenously distributed in iron
up to solubility limit of copper. Around 9.2 wt.% Cu can be soluble in Fe at 1150
o
C according to
Fe-Cu binary phase diagram shown in Fig. 2. Above this solubility limit of Cu in Fe at 1150
o
C will
drive to transient liquid phase sintering. Excess copper above solubility limit in iron caused partial
liquid phase sintering. Liquid copper dissolves small iron grains that decrease the number of
grains in the structure which results in grain growth. Reprecipitation of solution results in better
packing of bigger grains and let the liquid copper to fill small pores and also solidify between
rounded Cu-rich iron grains in the regions between large pores as seen in Fig. 4. This grain
growth decreases rupture strength of compacts as shown in Fig. 5. During this solution-
reprecipitation phenomena smaller pores filled with copper melt. The rest of the pores come
closer by grain growth and reprecipitation that generates bigger and more rounded pores in the
structure. It is seen from the Fig. 3 that microwave sintered specimens have smaller and rounded
pores which are homogeneously disturbed according to conventionally sintered ones that improve
rupture strength.

Figure 2. Fe-Cu binary phase diagram [ASM HandbookVol. 3 Alloy phase diagrams]

(a) (b)

(c) (d)

(e) (f)

(g) (h)

(i) (j)

(k) (l)
Figure 3. SEM morphologies and EDS analysis of cross-section of sintered specimens. Left column
corresponds to CS and right column corresponds to MWS arranged in order alloyed with 2 wt. % (a,b),
3 wt. % (c,d), 4 wt. % (e,f), 8 wt. % (g,h), 10 wt. (i,j) and 12 wt. % (k,l) copper.

(a)

(b)
Figure 4. Back scattered mode SEM morphologies of Iron samples alloyed with 8 wt.%, 10 wt.% and
12 wt.% Cu, compacted via CS (a) and MWS (b).

280
305
330
355
380
405
430
0 2 3 4 8 10 12
wt. % Cu
T
R
S

(
M
P
a
)
Fe-Cu alloy MWS Fe-Cu alloy CS

Figure 5. TRS values of sintered compacts by CS and MWS
14
15
16
17
18
19
20
21
22
23
0 2 3 4 8 10 12
wt. % Cu
%

P
o
r
o
s
i
t
y
MWS CS

Figure 6. Porosity levels of sintered compacts by CS and MWS

It is clear from Fig. 5 that, microwave sintering yields higher values of rupture strength than
conventional sintering in Fe and Fe-Cu alloys for the same amount of copper addition. Mercury
porosimeter results are shown in Fig 6. Microwave sintered samples have less porosity amounts than
conventionally sintered ones according to Fig 6. Decrease in porosity level means more densed
compacts that improves rupture strengths as seen in Fig 5. In most liquid-phase sintering systems the
amount of liquid is insufficient to fill all the pore space on melt formation and rearrangement. Solution-
reprecipitation allows the growing solid grains to deviate from a spherical (minimum energy) shape for
better space filling. This releases liquid into the remaining pores. However, large pores have a low
capillary pressure and are stable through a considerable portion of the sintering cycle. The remaining
large pores are difficult to fill. Maximum strength for MWS was 438 MPa for 3 wt. % Cu addition which
is approximately 20% higher than CS. It was found that maximum rupture strength was 380 MPa for 2
wt. % Cu addition for CS. The strength was decreased upto 10 wt % Cu addition for CS which
corresponds to the solubility limit of copper in iron. The rupture strength of 10 wt. % Cu addition was
lower than unalloyed Fe compacts. In addition to the overall pore volume, the morphology of the
porosity also influences mechanical properties. This includes the size distribution as well as the shape
of pores. For the same total volume, many small pores are less detrimental than a few large ones.
Similarly, well-rounded porosity, which typically develops in high-temperature sintering, is less
damaging than sharp, angular porosity [9].

Conclusions
Unalloyed and alloyed Fe compacts are sintered succesfully via MWS by using maximum 1000 Watts of energy
with a heating rate of 20-30
o
C/min. Microwave sintered specimens show higher TRS values than conventionally
sintered ones. Maximum TRS value is 438 MPa for 3 wt. % copper addition for MWS. Cu is distributed
homogenously in the structure upto 8 wt.% Cu addition. Microstructural observations suggest that the liquid-phase
sintering mechanism is operative for the samples containing 8 to 12 wt. % Cu. The microstructures also indicate
that the solution and reprecipitation of iron results in the formation of rounded grains and accelerated grain
growth. Excess liquid copper is solidified at the boundaries of rounded Cu-rich iron grains, in regions between the
large pores.

References
[1] F. Petzoldt, B. Scholz, H. S. Park and M. Willert-Porada,Microwave sintering of PM steels, Advances in
Microwave and Radio Frequency Processing Report from the 8th International Conference on Microwave and
High Frequency Heating held in Bayreuth, Germany, September 37, 2001
[2] K. Saitou, Microwave sintering of iron, cobalt, nickel, copper and stainless steel powders, Scripta Materialia
54 (2006) 875879
[3] M. Willert-Porada, H.-S. Park, Heating and sintering of steel powders with microwaves at 2.45 GHz
Frequency Relation between heating behaviour and electrical conductivity, Ceram. Trans., Vol. 111, 459-471,
2001
[4] C. Leonelli, P. Veronesi, L. Denti, A. Gattob, L. Iuliano, Microwave assisted sintering of green metal parts,
Journal of materials processing technology 205 (2008) 489496
[5] A. Taskinen, M.H. Tikkanen and G. B. Bockstiegel, Scandinavian Journal for Powder Metallurgy, 10, 1981, 55
[6] I.M.Kusoglu, K.Onel, M.Willert-Porada, Microwave assisted sintering of Iron compacts, 5th International
Powder Metallurgy Conference, 8-10 October 2008 Ankara, Turkey
[7] A. N. M. Negari, R. S. Mamoory, A. Simchi, N. Ehsani, Determination of the physical and mechanical
properties of iron-based powder materials produced by microwave sintering, Powder Metallurgy and Metal
Ceramics, Vol. 46, Nos. 9-10, 2007
[8] Y. Peng, J. Yi, S. Luo, Y. Guo, L. Li,Microstructure analysis of microwave sintered ferrous PM alloys, Journal
of Wuhan University of Technology-Matter Sci. Ed. Vol. 24, No.2, April 2009
[9] H. Danninger, G. Jangg, B. Weiss, and R. Stickler. Microstructure and Mechanical Properties of Sintered Iron,
Part I, Powder Metal. Int., Vol 25 (No. 3), 1993, p. 111-117


The Effect of Microwave Sintering on In-Situ Synthesis of
Cordierite Powder

S. AKPINAR, I.M.KUSOGLU, I.A.ALTUN and K.ONEL

University of Dokuz Eylul, Department of Metallurgical & Materials Engineering, Campus
of Tinaztepe, 35160, Buca, Izmir, TURKEY

Abstract

Cordierite (2MgO.2Al
2
O
3
.5SiO
2
) is commonly used phase for high temperature and
thermal shock resistant applications in ceramic industry due to its low thermal expansion
and excellent thermal shock resistance. In this study microwave and conventional
energy sources were used to form cordierite phase. Sintered magnesia, kaolinite and
quartz were used as starting materials. The starting materials were mixed according to
stoichiometric ratio of 2MgO.2Al
2
O
3
.5SiO
2
and compacted as coin shaped test samples
by uniaxial pressing. Test samples were sintered for different sintering durations and
temperatures by using both energy sources. Obtained products were characterized by
XRD analysis and density tests. The results were discussed.

Keywords: cordierite, microwave sintering, conventional sintering

1. Introduction

Cordierite (2MgO.2Al
2
O
3
.5SiO
2
) is one of the important phases of the MgOSiO
2
Al
2
O
3

system. The cordierite ceramics due to their low thermal expansion coefficient, high
chemical durability, low dielectric constant, high resistance to thermal shock and high
refractoriness are promising candidate in many applications. Some examples of the
applications are electrical porcelains, catalytic converter substrates for exhaust gas
control in automobiles, heat exchanger for gas turbine engines, industrial furnaces,
packing materials in electronic packing, refractory coating on metals, integrated circuit
substrates, etc. [12]. There are several methods such as solid-state reaction, sol gel
and crystallization of glass to synthesize cordierite phase. Among these methods, the
conventional methods for the synthesis of cordierite phase which includes the solid-state
sintering of individual oxides of magnesium, aluminum and silicon corresponding to the
chemical composition of cordierite, crystallizing the glasses or the sintering of the natural
raw materials such as talc and china clay with low silicon content [1]. Chemical routes
such as the solgel technique generally yield high purity cordierite. Natural raw materials
are often used in industrial production of cordierite ceramics for refractory application.
Some of the starting raw materials to form cordierite phase reported in literature include
(i) a mixture of magnesium compounds and kaolinite [2], (ii) alkaline-earth-
aluminosilicate glass, kaolin, alumina and magnesite [3], (iii) talc, calcined alumina and
fly ash [4], (iv) kaolin, talc, silica and alumina [5], (v) talc, kaolinitic clay and gibbsite [6],
(vi) kaolin, talc and magnesium oxide [7], (vii) talc, kaolin, silica, sepiolite and feldspar
[8], and (viii) kaolin and talc [9]. Cordierite phase has three polymorphic forms: (i) the
stable high-temperature disorder form, also known as indialite ( - or hexagonal-
cordierite) which is stable below 1450
o
C, (ii) -cordierite, called orthorhombic cordierite,
stable between 1450
o
C and its melting point (1460
o
C), and (iii) -cordierite, called
metastable cordierite phase, the low-temperature ordered orthorhombic form [1], which
can be prepared only under special conditions.

Using microwaves for firing or sintering ceramics, has reduced the processing time and
energy consumption, and also improved the microstructural and mechanical properties.
Microwave energy supports uniform and volumetric heating in drying, calcining and
sintering processes. The focus of this study was to produce cordierite phase by
microwave and conventional sintering processes. Moreover, the influence of the
sintering parameters such as time and temperature on the formation of cordierite phase
by using starting materials that contain magnesia, kaolin and quartz were investigated.

2. Experimental Studies

Kaolin (d
50
=9.822 m, specific surface area 11.804 m
2
/g), magnesia (d
50
=23.78 m,
specific surface area 1.98 m
2
/g) and quartz (d
50
=25.69 m, specific surface area 0.588
m
2
/g) were used as the starting materials for in-situ formation of cordierite phase. The
composition of the Cordierite phase consists of MgO:Al
2
O
3
:SiO
2
in weight ratios
13.7:34.9: 51.4. Calculated stoichiometric ratios for the formation of cordierite phase
were 82.91 wt. % Kaolin, 12.20 wt. % Magnesia and 4.89 wt. % Quartz.

The stoichiometric cordierite batch was dry mixed in a ball milling unit with alumina balls
for 5 minutes. Green compacts were compacted by using unaxial press at 10 MPa
pressure. The shaped samples were sintered via microwave and conventional energy
sources for different duration and temperatures. Conventionally sintering processes
were carried under air atmosphere at 1300, 1350 and 1400
o
C with a heating rate of 10
o
C min
-1
for 1 h and at 1350
o
C for 3 h in order to determine phase evaluation.
Microwave sintering processes were carried out in air at 1250, 1300, 1350, 1400 and
1425
o
C for 1 h and at 1400
o
C for 2 h. The samples were heated to 600
o
C in 30 min
and held for 10 min using a power of 320 W to evaporate chemically bound water of
kaolin, and heated to given sintering temperatures in 60 min using a power of 960 W and
sintered with this power for given temperatures and durations.

A microwave transparent ceramic based board was used as sample holder. SiC
susceptors are used to assist microwave heating of samples as shown in Fig.2. The
temperature of the sample was monitored using an infrared pyrometer with circular
crosswire focused on the surface of samples. The emmisivity of pyrometer was set to
0,65 for all microwave heating processes. The samples were cooled in 3 h to 300
o
C
after sintering is completed.

Figure 1. Linn High Therm Dual-Frequency 2.45 GHz/2.4 kW Microwave
Sintering Furnace.

Figure 2. Experimental set-up, including refractory housing and four susceptors.

Differential thermal analysis (DTA; Shimadzu, DTG-60H) was performed in order to
determine the thermal cycle of cordierite composition. Sintered samples were ground
into powder form for pycnometer density measurements. The phase composition of the
sintered samples was determined by standard powder X-ray diffractometer (RIGAKU,
Japan).

3. Result and discussion

The thermal behaviour of starting material mixtures was characterized by differential
thermal analyzer in air up to 1350
o
C with a heating rate of 10
o
C min
-1
(Fig. 3). The
endothermic peak at 557
o
C might be attributed to the transformation of kaolinite to
aluminasilicate (meta-kaolinite) and evaporation of chemically bound water in kaolinite.
The exothermic peak around 980
o
C corresponds to the transformation of primer mullit
from meta-kaolinite. As a result of the mullitization of the kaolinite and the solid-state
reactions up to 1280
o
C, MgAl
2
O
4
spinel formed with the interaction of [AlO
6
] and [MgO
6
]
octahedral, enstatite turned into protoenstatite, amorphous silica was crystallized to
cristobalite, and the transformation of to -cordierite had occurred. The endothermic
peak at 1280
o
C is due to decomposition and dissolution of protoenstatite in liquid
silicate that forms cordierite crystallization.
-70
-60
-50
-40
-30
-20
-10
0
10
20
0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300
Temperature (
o
C)
D
T
A

(
v
)
557
o
C
983
o
C
1280
o
C
E
x
o
E
n
d
o

Figure 3. The DTA patterns of the mixed starting materials.

1300
o
C-1h
1350
o
C-1h
1350
o
C-3h
1400
o
C-1h
0 10 20 30 40 50 60 70 80 90
Cordierite Mullite : Ensteatite : Cristobalite : Corundum : Spinel

Figure 4. The XRD patterns of the samples sintered by conventional heating at different
sintering temperatures and durations.

The X-ray diffraction patterns of conventionally sintered samples at 1300
o
C for 1 h
indicated the presence of cordierite, mullite, cristobalite, spinel, ensteatite and corundum
phases as shown in Fig. 4. Most of the spinel, ensteatite, corundum and mullite phases
transformed to cordierite phase which are sintered at 1350
o
C for 1h. Intensities of the
peaks of cordierite phase increased for the samples sintered at 1350
o
C for 3 h.
Intensities of all peaks of cordierite phase decreased after sintering at 1400
o
C for 1 h.
This is a result of glassification of cordierite phase for this temperature.

Figure 5. The XRD patterns of the samples sintered by microwave heating at different
sintering temperatures and durations.

The X-ray diffraction patterns of samples sintered by microwave heating at 1250
o
C and
1300
o
C for 1 h indicated the presence of cordierite, mullite, quartz, spinel, forsterite and
ensteatite phases as given in Fig. 5. XRD patterns of the samples sintered at 1350
o
C for
1h indicate that the ratio of transformation of cordierite phase increases with increased
sintering temperature. All of the characteristic peaks of cordierite phase are observed for
the samples sintered at 1400
o
C for 1h by transformation of secondary crystalline phases
such as forsterite, spinel, ensteatite phaseses. There is still mullite phase present at this
sintering temperature. Intensity of peaks of cordierite phase has increased for the
samples sintered at 1400
o
C for 2 h. The peak intensities of cordierite phase have
decreased by raising temperature from 1400
o
C to 1425
o
C.
Table 1. Densities of cordierite powders
Sintering
temperature
and duration
b (g/cm
3
)
Conventional
Sintering
Microwave
Sintering
1250-1h - 2.55
1300-1h 2.23 2.60
1350-1h 2.47 2.71
1350-3h 2.49 -
1400-1h 2.45 2.57
1400-2h - 2.47

Densities of samples sintered at different temperatures and durations are shown in
Table 1. The density of cordierite phase, conventionally sintered at 1350
o
C for 3h, is
2.49 g/cm
3
and for microwave sintering at 1400
o
C for 2 h, is 2.47 g/cm
3
.These values
are very close to the theoretical density of cordierite phase (2.51 g/cm
3
).

4. Conclusions

Cordierite phases in powder form were successfully synthesized by using raw materials
including kaolin, magnesia and quartz which are sintered by using microwave heating
and conventional heating. Cordierite phase transformation was fully completed at 1350
o
C by conventional sintering for 3 h and mostly achieved with minor mullite phase by
microwave sintering at 1400
o
C for 2 h. It is also found that densities of synthesized
cordierite phase were very close to the theoretical one for sintering temperatures and
durations given above. Formation of cordierite phase from the same starting materials
proved that microwave sintering temperature is higher than conventional sintering
temperature. This is an unexpected result for microwave sintering process. It is thought
that emmisivity value of the compacted starting material mixtures could be greater than
0,65. Further studies will be focus on different emmisivity values and determination of
true values of emmisivity for the used starting materials.

References
[1] A.V. Ganesha, B. Basavalings, J.A.K. Tareen, M.A. Pasha, Curr. Sci. 87 (2004) 104
108.
[2] Y. Kobayashi, K. Sumi, E. Kato, Ceram. Int. 26 (2000) 739743.
[3] D.U. Tulyaganov, M.E. Tukhtaev, J.I. Escalante, M.J. Ribeiro, J.A. Labrincha, J. Eur.
Ceram. Soc. 22 (2002) 17751782.
[4] S. Kumar, K.K. Singh, P. Ramachadrarao, J. Mater. Sci. Lett. 19 (2000) 12631265.
[5] J.R. Gonzalez-Velasco, M.A. Gutierrez-Ortiz, R. Ferret, J. Mater. Sci. 34 (1999)
19992002.
[6] S. Tamborenea, A.D. Mazzoni, E.F. Aglietti, Thermochim. Acta 411 (2003) 224219.
[7] A. Yamuna, R. Johnson, Y.R. Mahajan, M. Lalithambika, J. Eur. Ceram. Soc. 24
(2004) 6573.
[8] Z. Acimovic, L. Pavlovic, L. Trumbulovic, L. Andric, M. Stamatovic, Mater. Lett. 4217
(2002) 16.
[9] L. Trumbulovic, Z. Acimovic, S. Panic, L. Andric, FME Trans. 31 (2003) 4347.

Oxidation Behavior Resistance of a Duplex NiAl/Ti-Ni-Al Coating by
Microwave Assisted SHS on Ti Substrate

I.Boromei, A.Casagrande-SMETEC; University of Bologna, Italy
G.Poli, P.Veronesi, R. RosaDepartment of Materials and Environmental
Engineering;.University of Modena and Reggio Emilia, Italy

Abstract
Microwaves have been used to ignite the Self-propagating-High-temperature Synthesis
(SHS) of Ni and Al powder mixtures to produce a duplex intermetallic coating on Ti
substrates. Due to the high exothermic nature of the reaction, the newly formed NiAl is in the
liquid phase and can react with the underlying Ti to form a tough ternary intermediate layer in
a one step process. The obtained thick intermediate layer, Ti-Ni-Al ternary compound, with
improved mechanical properties could replace the functionality of hard and brittle NiAl
aluminide coatings in service. Moreover, the thickness of the intermediate layer , presenting
an interlocking microstructure, can be controllably altered, up to 500 m, by annealing at
950C for times ranging in the 3-20 minutes.
The present work intends to assess the high temperature performances of the Ti-Ni-Al
ternary coating, compared to pure Ti, Ti-6Al-4V and NiAl. In particular, cyclic oxidation
response in the 750-900C range and wear resistance at 750C, measured using a ball-on-
disk tribometer, were evaluated. The tendency for scale spallation, adherence and oxidation
rates for NiAl and Ni-Al-Ti intermediate layer was evaluated as well.

Keywords: coating, microwaves, Ni-aluminides, oxidation and wear.

Introduction
The role of coatings on Ti and Ti alloys at high operating temperature (above 500C) is
simply to limit environmental attack of the metallic substrate. Nowadays, owing to an
increasing demand of high temperature performances, it is no longer sufficient for the coating
to just minimize the corrosion-oxidation rate; the protective coating must form a slow-growing
external protective layer, with the best adhesion to the substrate to limit spallation and they
should possess good mechanical properties also at high temperature.
Titanium and its alloys are suitable candidates for fabrication of light weight components to
be used in aerospace and automotive sector, owing to their high-temperature strength in
combination with low density [1]. However, due to severe oxidation problems at temperature
above 520C and reduced wear resistance, for many applications it becomes necessary a
modification of the surface properties of the metallic substrate [2]. In this context, the powder
metallurgy processing, and in particular the Self-High Temperature-Synthesis (SHS) using
microwaves to ignite the reaction, has proved to an interesting route to form Ni-aluminide
protective duplex coating on Ti [3,4]; forming a NiAl outer layer and an inner ipoeutectic Ni-
Ti-Al ternary layer.
In this work, in order to simulate real operating conditions of parts in e.g. jet engines, cycling
oxidation periods and mechanical behaviours on NiAl and Ti-Ni-Al ternary layer formed at
Ti/interface ( which could replace the functionality of hard and brittle NiAl in service are
performed at the same isothermal treatments and compared to the behaviour of uncoated Ti
and Ti alloys.
Therefore, a set of compositional guidelines for the formation of protective Ni-aluminide
coatings is proposed to maximize the coating lifetimes at temperature up to 900C.

Experimental Procedures
Discs of 30 mm diameter, 3 mm thick, made of grade 2 titanium, kindly provided by
Akrapovic, have been polished up to grit 1000 sand paper and cleaned in acetone. Ni
carbonyl powders (99% pure, 5-7 m in diameter) and Al gas atomized powders (99.8%
pure, 10-50 m in diam.) were mixed under vacuum in a 1:1 atomic ratio. Uniaxially pressed
30 mm diameter discs, 7g weight, were obtained (applied pressure: 140 MPa) and placed
under 20 Nml/min argon flux, between a substrate (a titanium disc) and a sacrificial covering
(AISI 304 disc, 3 mm thick), in a single mode microwave applicator (TE10n, n=3 or 4, based
on WR340 geometry). The microwave generator (2.45GHz) output power was set to 1200 W,
and exposure of samples to the electromagnetic field was prolonged for 1 s after SHS
ignition. During synthesis, a moderate 0.45 MPa uniaxial pressure was applied in order to
reduce porosity formation. The microwave applicator and the sample arrangement are
reported in detail elsewhere [5].
The samples, after synthesis occurred, were then ground in order to remove the sacrificial
stainless steel covering, and subsequently ground and polished up to 800 grit sand paper.
Depending on the intermetallic thickness removed by polishing, as prepared in a previous
work [4] after annealing at 950C for 20 minutes to increase the thickness of the Ni-Ti-Al
ternary layer, a flat surface made of NiAl, or of Ni-Al-Ti ternary compound formed at the Ti
interface, was prepared for subsequent oxidation resistance characterization in the range
750-900C and tribological tests at 750C.
Oxidation treatments of all samples (grade 2 Ti/NiAl coated, grade 2 Ti/Ti-Ni-Al ternary
coated, grade 2 Ti uncoated and grade 5 Ti uncoated)) were performed in air at atmospheric
pressure at two range temperatures: the first at 750C for 100 h, the second at 900C for
0,75, 1,5 and 100 h. Prior to oxidation, specimens were polished to a 3- m finish with
diamond paste and cleaned ultrasonically in alcohol.
After the defined oxidation times, the surfaces of each sample were examined by SEM and
EDX microanalyser in order to obtain information on the morphological growth and the
composition of the formed superficial oxide. Cross-section of samples were investigated
using optical and SEM microscope and EDS microanalyser in order to characterize the
morphological changes occurring in oxides of the intermetallic coatings during oxidation
treatments as well as chemical evolution of the superficial oxides. Meanwhile, non oxidized
samples of discs of 30 mm diameter and 3 mm thick, ( grade 2 Ti/NiAl coated, grade 2 Ti/Ti-
Ni-Al coated, grade 2 Ti uncoated and grade 5 Ti uncoated), were used for tribological tests
at 750C isothermal treatment.
Rotating unidirectional ball-on-disk tests were performed using a pin-on-disk tribometer
(CSM Instruments, Peseux, Switzerland, each test repeated at least twice), according to the
test method ASTM G99-04. The samples (discs) were cleaned in acetone before each tests,
conducted at a measured sample temperature of 750C and ~50% RH in air. Surfaces made
of grade 2 titanium (indicated by the G2 acronym in the text) , grade 5 titanium (G5), NiAl
intermetallic (NA) and ternary Ni-Al-Ti intermetallic compound (TNA) were used as disk
materials. The counterpart material (ball) was an alumina sphere, 6 mm of diameter, on
which a 5 N normal load was applied troughout the whole test. Overall sliding distance was
set to 1000 m, at a speed of 0.2 m/s. Friction coefficient as a function of sliding distance was
measured during the whole test. The obtained wear tracks on each sample were measured
by optical confocal profilometry (ConScan profilometer, CSM Instruments) and the volume
loss was determined, allowing the calculation of the wear rate.
A comparison of the oxidation behaviour of coated and uncoated Ti and Ti alloy correlated to
wear resistance and friction coefficient at the same isothermal treatment was made.


SCALE MORPHOLOGY AND COMPOSITION AT 750C ISOTHERMAL TREATMENT FOR
100 h

The average atomic composition of NiAl and ternary Ni-Ti-Al coating considered in the
oxidation process, as reported by EDS semiquantitative analysis [4], was respectively Ni
52% and Al 48% and Ni 20%, Ti 65%, Al 15%.
Figure 1 shows SEM micrographs of the oxide formed on NiAl, on Ni-Ti-Al ternary compound
and grade 5 Ti. It can be seen that a dense oxide, rich in Al, is growing on the intermetallic
NiAl surface (Fig 1a); it is the main oxide on the coating and its atomic composition by EDX
analysis, suggesting the formation of Al
2
O
3
. The oxide is probably made of -Al
2
O
3
because
was formed at relatively low temperature in oxidizing treatment and because -Al
2
O
3
has the
fastest growth rate among all alumina phase [6].. Figure 1 b shows SEM micrograph of the
oxide scales formed on the Ni-Ti-Al ternary coating. The surface morphology shows a
relative dense oxides, which EDS analysis indicates as made of Ti, Al and O contents,
suggesting that -alumina and TiO
2
mixture or a Ti-Al-O spinel are the main products. The
SEM micrograph of Fig. 1c, rapresentative of both grade 2 Ti similar to grade 5 Ti, shows a
typical non protective powdery TiO
2
scale.

Figure 1 SEM micrographs of oxide morphology on: a) NiAl; b) Ni-Ti-Al, c) grade 5 Ti

The Figure 2 shows the optical cross-section of the scale thickness on tested samples. The
best result, as far as oxidation resistence is concerned, is obtained on the NiAl coating where
the thickness of the -alumina layer, not continuous on the surface, measured approximately
1 m. The ternary coating seem to be efficient in producing a reduction in the oxidation rate,
compared to uncoated Ti: 15 m against 60 m,.after 100 hours, as shown in Fig 2 b,c.

Figure 2. Optical cross-sectional micrographs showing the oxidelayer thickness and the electrolytic
Cu deposited to protect the oxide during and polishing: a) NiAl, b) Ni-Ti-Al, c) grade 5 Ti

SCALE MORPHOLOGY AND COMPOSITION AT 900C ISOTHERMAL TREATMENT

Figure 3. shows typical micrographs of cross-sections of oxide scales formed on NiAl, Ni-Ti-
Al ternary compound and uncoated Ti grade 5 under the same oxidation conditions, 100 h
isothermal exposure.

Figure 3. Optical cross-sectional micrographs: a) NiAl, b) Ni-Ti-Al, c) grade 5 Ti.

a) b) c)
a) b) c)
Electrolytic Cu
Ni-Ti-Al
NiAl
Oxyde
Ti
a)
b) c)
The alumina scale (dark grey layer, confirmed by EDS analysis) thickness of the NiAl was
the thinnest among the three samples, i.e. 3-5 m. The oxide grown on the ternary Ni-Ti-Al
coating, Fig. 3b, presented a double layer; the outer one more dense of about 120 m thick
and the inner one , more powdery, about 130 m thick, entirely composed by TiO
2
on metallic
substrate. The uncoated metallic substrate presents a thick powdery oxide coating, about
550 m Thick, Fig. 3c.
As shown in Fig.3b, the ternary Ni-Ti-Al layer, originally about 100 m thick, was entirely
destroyed and transformed in complex non protective oxide structures in air at 900C for 100
h exposure. Moreover, it was experimentally observed that the oxidation rate of the ternary
layer increases substantially after only 0,75 h and probably grows in exponential way after
1,5 h. After oxidation for 100 h, there was a difference in the superficial appearance of the
oxide on the ternary Ni-Ti-Al compound, NiAl, Ti and Ti alloy as shown in Fig 4

Figure 4. SEM micrographs of oxide morphology on: a) NiAl, b) Ni-Ti-Al, c) grade 5 Ti.

A more detailed explanation on the origin of the duble layer on yhe ternary compound and to
understand the correct mechanism of the internal oxidation of the ternary Ti-Ni-Al compound
and the growth of an external oxide scale, at expense of the original ternary Ni-Ti-Al coating,
in this situation, requires a knowledge of many parameters including the gas-phase oxygen
pressure and its effect on the rate constant for the grow of the external scale, he diffusion
coefficients and the solubility of oxygen in the pure metals forming the matrix and the
diffusion coefficient of the metal components in the phase [7] and work in currently in
progress.

WEAR TESTS AT 750C AT ISOTHERMAL TREATMENT
Friction coefficient against alumina, as a function of sliding distance, was continuously
measured during the tribological tests. Results are shown in figure 5, comparing the different
behaviour of the Grade 2 Titanium (G2), Grade 5 titanium (G5), NiAl coating (NA) and the
ternary layer coating (TNA).

0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 100 200 300 400 500 600 700 800 900 1000
sliding distance (m)
f
r
i
c
t
i
o
n

Figure 5.: Friction coefficient against alumina as a function of sliding distance for: Grade 2 Titanium
(G2), Grade 5 titanium (G5), NiAl coating (NA) and the ternary layer coating (TNA).

G2 G5
TNA NA
a) b) c)
0
10
20
30
40
50
60
70
80
G2 G5 NA TNA
W
.
R
.

(
E
-
0
5

m
m
3
/

N

m
)

Figure 6. wear rate after 1000 m sliding against alumina at 750C

The friction coefficient of the grade 2 titanium is higher throughout the duration of the whole
test, while the ternary compound coating seems to provide the lower friction. Rather
unexpectedly the NA sample present a friction coefficient higher than grade 5 titanium,
probably due to some diffused residual open porosity which interacts with the alumina pin,
locally increasing friction.
The optical microscope micrographs of wear tracks, after 1000 m sliding against alumina at
750C, are shown in figure 7.
It is evident the presence of oxidised products in case of the G2 and NA sample, while
oxidation is mucgh less pronounced in case of G5 and TNA samples. Considering the cycling
oxidation results, this apparent contradiction can be due to the continous removal of non
protective oxides during sliding, thus only wear rate will be considered a significant result.
Optical microscopy also evidenced the presence of multiple fine cracks (a longitudinal crack
is visible also in the NA micrograph of figure.7 c.) in the whole surface of the NA samples,
while the TNA sample retained its integrity. In case of TNA samples, the prolonged exposure
to the 750C temperature evidenced the dendritic microstructure of TNA samples, already
detectable at low magnification (TNA micrograph of figure.7 d).
A possible explanation of NiAl damaging could be different values of the coefficient of
thermal expansion (CTE) mismatch between the substrate, Ni-aluminide coatings and oxide
scales[8]. Thus, tensile stresses are expected to raise, and to be higher in case of the NiAl
coating.compared to ternary Ni-Ti-Al compound. On the other side, the slightly lower CTE of
the ternary compound, similar to the Ti one, and its superior toughness [4], are probably
responsible for the absence of cracks on the ternary compound coating.
Wear track width is much higher in case of uncoated samples, while the presence of
intermetallic coating seems to lead to a lower mass removal during testing.
In order to confirm the higher wear resistance of intermetallic-coated samples, optical
profilometry measurements of the wear tracks was used to quantify the effective wear rate of
the titanium and intermetallic-coated discs, expressed in terms of volume removed during the
test normalised for the sliding distance and the applied load.
Wear rate results are shown in figure 6. Due to the unevenness of the wear track profiles,
multiple measurements were taken at 30 different positions in the wear tracks.
Scattering, in percentage, is more pronounced in case of the ternary layer coating, probably
due to its heterogeneous microstructure, while the monophasic material tends to lead to
more reproducible results. It should be noted, however, that similar tests conducted on more


G2 1mm G5 1mm

NA 1mm TNA 1mm
Figure.7. Wear tracks of the G2, G5, Na, TNA samples after 1000m sliding against alumina at 750C

porous coatings, obtained applying a lower pressure during synthesis, lead to higher values
of the wear rate, namely 3.85 E-05 mm
3
/N m for the NA samples and 5.554 E-05 mm
3
/N m
for the TNA ones. In both cases, wear rates of the intermetallic coatings resulted one order of
magnitude lower than uncoated samples, confirming that the presence of aluminide coatings
improves high temperature wear resistance in air at temperature up to 750C.

CONCLUSIONS
The oxidation rate of NiAl coating on Titanium was found to be slower than that of Ni-Ti-Al
ternary compound because the formation of a protective thin Al
2
O
3
layer both at 750 and
900C;
The ternary Ni-Ti-Al coating can be considered as a protective layer only up to 750C. On the
ternary compound, a continous aluminium oxide layer is not formed, and this can be ascribed
to a lower and insufficient Al content.
On the contrary, the Ni-Ti-Al ternary coating presents the higher wear resistance up to 750C
in air (conditions in which it is still protective), due to its superior toughness and hardness
and can replace the functionality of hard and brittle NiAl aluminide coatings in service at
medium temperatures.

References
[1] M.J.Donachie Jr, Titanium and Titanium Alloys, ASM, Metals Park, OH, 1982, p.3;
[2] R.R. Boyer, Adv. Perform Mater, 2 (1995) 349;
[3] A.Casagrande, GP.Cammarota, C.Leonelli, G.Poli, P.Veronesi, EURO PM2005,
Proceedings,Vol.I, October 2-5 2005 p.107, Prague;
[[4] GP:Cammarota, A.Casagrande, G.Poli, P.Veronesi, Surf. Coat.Technol.203 (209) 1429-
1437;
[5] P. Veronesi, R. Sola, G. Poli, Microwave Assisted Combustion Synthesis of NiAl
intermetallics in a single mode applicator: modeling and optimisation, Materials Science and
Engineering A, 441, 149-156, 2006;
[6] E.Schumann, Oxid.Met. 43, 157 (1995)
[7] F.Gesmondo and Y.Niu, Oxid.Met.62, 516 December (2004);
[8] C.I.Hsieh and W.H. Tuan, Mat.Sci.Eng. A, Vol.460-461 15/07/2007 pp.453-458].


In-Situ Synthesis of Mullite Powder from Seydisehir Alumina and
Gibbsite via Microwave and Conventional Energy Sources

O.ERTUGRUL
(*)
, I.M.KUSOGLU, S.AKPINAR, I.A.ALTUN and K.ONEL

University of Dokuz Eylul, Department of Metallurgical & Materials Engineering,
Campus of Tinaztepe, 35160, Buca, Izmir, TURKEY

Abstract

Mullite (3Al
2
O
3
.2SiO
2
) is commonly used phase for high temperature and thermal shock resistant
applications in ceramic industry. In the present investigation, two different formation energies as
microwave and conventional sources have been used to form mullite phase. Seydisehir alumina,
gibbsite, kaolinite and quartz were used as starting materials. Seydisehir alumina and gibbsite
raw materials have been transformed into -alumina phase by heating at elevated temperature to
improve mullite synthesis. The raw materials have been mixed according to stoichiometric ratio of
3Al
2
O
3
.2SiO
2
and compacted as coin shaped test samples by uniaxial pressing. Test samples
have been sintered using microwave and conventional energy sources with different duration and
temperatures. The products have been characterized by SEM/EDS, XRD analysis.

Keywords: mullite, microwave sintering, conventional sintering

I. Introduction

Mullite, ideally 3Al
2
O
3
2SiO
2
, is a high melting crystalline alumino-silicate material which has long
been used in heavy-duty refractories. Using cheap raw materials such as kaolinite, kaolinite
alumina and sillimanite minerals, it is possible to produce high purity mullite [1]. The sintering of
mullite powders requires high temperature and long thermal treatment due to mullite powders
have poor solid state sinterability because of the low interdiffusion rates of Si
4+
and Al
3+
within the
mullite lattice [1,2]. There are many possible additives, which act as sintering aids and grain
growth inhibitors. Among these additives, magnesia (MgO) is the most consumed and effective
one in obtaining mullite products with high density using relatively low temperatures and shorter
sintering periods. Addition of 1 wt% magnesia improves the densification of mullite powder, giving
rise to more compact and well developed microstructures. A lower amount of additive does not
produce any significant density improvement, whereas larger quantities lead to glassy phase
formation which can be prejudicial for the performance of the component [2].

Mullite is becoming increasingly important in electronic, optical, and high temperature structural
applications, because of its low dielectric constant, good transparency for mid-infrared light and
excellent creep resistance. Classical uses of mullite include refractories in the metallurgical
industries for electric-furnace roofs, hot metal mixers and lowfrequency
induction furnaces [3].

Using microwaves for firing or sintering ceramics has reduced the processing time and energy
consumption, and also improved the microstructural and mechanical properties. Microwave
energy provides uniform and volumetric heating in drying, calcining and sintering processes.

The aim of this study was to produce mullite phase by conventional and microwave sintering
processes. Furthermore, the influence of the sintering parameters such as time and temperature
on the formation of mullite phase from a raw material mixture containing Seydisehir alumina,
gibbsite and kaolinite was investigated.

2. Experimental Studies

Seydiehir Alumina (15 wt.% -Al
2
O
3
, d
50
=60.207 m and specific surface area 56.629 m
2
/g) was
used as source of -Al
2
O
3
and transformed fully to -Al
2
O
3
by rapid heating with 25
o
C/min to
1250
o
C with a soaking time of 1 h. Ground -Al
2
O
3
for 3 hours by ball milling (d
50
=3.388m and
specific surface area 9.130 m
2
/g) and Grolleg Kaolin (d
50
=9.822 m, specific surface area 11.804
m
2
/g) have been used as the starting materials for in-situ mullite formation. Gibsitte and MgO are
also added as sintering aid.

The composition of the mullite phase consists of SiO
2
/ Al
2
O
3
in weight ratios 28.17 / 71.83. To
provide the formation of this phase 42.54 wt.% -Al
2
O
3
and 52.46 wt.% Kaolin as starting
powders are used. Gibsitte(5 wt.%) and MgO(1 wt.%) are added to this mixture.

The thermal behavior of the mullite composition is characterized by differential thermal analyzer
(DTA; Shimadzu, DTG-60H) in air up to 1350
o
C at a heating rate of 10
o
C min
-1
(Fig. 3). As
mentioned in the literature [4], the endothermic peak at 545
o
C might be attributed to the
transformation of kaolinite to aluminasilicate (meta-kaolinite) and vaporization of chemically
bound water of kaolinite (as shown in TG curve). The exothermic peak around 997
o
C
corresponds to the transformation of primer mullite from meta-kaolinite. As a result of the
mullitization of the kaolinite and the solid-state reactions up to 1200
o
C, Al-Si spinel is formed to
allow mullite crystallization.
-50
-40
-30
-20
-10
0
10
0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400
Temperature (
o
C)
D
T
A

(
m
v
)
22,5
23,0
23,5
24,0
24,5
25,0
T
G

(
m
g
)
DTA TGA
545
o
C 997
o
C
1200
o
C
E
x
o
E
n
d
o

Figure 3. The DTA patterns of the physical raw materials mixtures.
(Seydiehir Alumina + Kaolin).

The stoichiometric mullite batch have been dry mixed in a ball milling unit with alumina balls for 5
minutes and then green compacts are prepared by using unaxial pressure of 10MPa. The shaped
samples have been sintered via microwave and conventional energy sources for different
duration and temperatures. Conventional sintering processes is carried out in air at 1400, 1500
and 1600
o
C with a rate of 10
o
C min
-1
for 1 h and at 1600
o
C for 3 h in order to determine phase
formation. Microwave sintering processes has been carried out in air at 1300, 1350 and 1400
o
C
for 1 h. The samples have been heated to 600
o
C in 30 min and held for 10 min using a power of
320W in order to evaporate chemically bound water from kaolin. The samples have been heated
to given sintering temperatures in 60 min using a power of 960 W and sintered using this power
for given temperatures and durations.

Fig.1 shows the microwave furnace used for the present investigation. A microwave transparent
ceramic board has been used as sample holder. SiC susceptors are used to assist microwave
heating of samples as shown in Fig.2. The temperature of the sample is monitored using an
infrared pyrometer with the circular crosswire focused on the surface of samples. The emissivity
of pyrometer has been set to 0,65 for all microwave heating processes. The samples have been
cooled in 3 h to 300
o
C after sintering is completed.

Figure 1. Linn High Therm Dual-Frequency 2.45 GHz/2.4 kW Sintering Furnace.

Figure 2. Experimental set-up, including refractory housing and four susceptors.

The phase composition of the sintered samples has been determined by standard powder X-ray
diffractometer (RIGAKU, Japan). Microstructural analysis of sintered samples was observed by
scanning electron microscopy (JEOL 6060). As for the real densities, the sintered samples have
been first ground to powder and then tested with a pycnometer.

3. Result and discussion

The X-ray diffraction patterns of conventionally sintered samples at 1400
o
C for 1 h indicate the
presence of mullite, corundum and cristobalite phases (Fig.4). When the sintering temperature is
increased from 1400
o
C to 1500
o
C, the peak intensities of mullite phase have increased,
meanwhile the intensities of corundum phase peaks have decreased. There is no observed
difference between intensity of mullite peaks in the samples sintered at 1500 and 1600
o
C for 1h,
but the intensity of corundum peaks is increased. There is no significant difference between
diffraction patterns of samples sintered at 1600
o
C for 1 h and 3 h. As a result, it was found that at
the sintering temperatures given above it was not possible to obtain pure mullite phase due to the
presence of unreacted corundum phase. 15 wt. % -Al
2
O
3
in the composition of Seydisehir
alumina contributes mullite transformation while transition aluminas do not react with SiO
2
and
exist as corundum in the system during sintering processes [1]. For this reason, Seydisehir
alumina and gibbsite raw materials which contain transition alumina have been treated for -
Al
2
O
3
transformation by rapid heating to 1250
o
C with a soaking time of 1 h. The treated materials
have been used as the source of Al
2
O
3
for in-situ synthesis of mullite.


Figure 4. The XRD patterns of the samples (Seydiehir Alumina + Kaolin mixture) sintered by
conventional heating at different temperatures and durations.

Figure 5 shows the X-ray diffraction patterns of alumina based raw materials in the as received
and treated forms. In the patterns of as received form of Seydisehir alumina, there are some
peaks belonging to the transition alumina phases between 2=35-70
o
, and most of them have
disappeared after the treatment. But there are still small peaks in the patterns of treated gibbsite
that refer to transition alumina. For this reason, Seydisehir alumina has been selected as source
of -Al
2
O
3
.

Figure 5. The XRD patterns of the alumina raw materials as received and after heated
at elevated temperatures.

Figure 6 shows the X-ray diffraction patterns of new mullite batch which has been prepared by
using treated Seydisehir alumina (after complete transformation into -Al
2
O
3
. Using treated
material the mixture has been sintered 1600
o
C for 4 hours. The observations indicate that a
complete mullite phase transformation has been achieved.


Figure 6. The XRD patterns of the conventionally sintered samples
(Calcined Seydiehir Alumina + Kaolin mixture)

Figure 7 shows the X-ray diffraction patterns of calcined Seydisehir alumina and kaolin mixture
sintered by microwave heating at different temperatures and durations. It was observed that there
was mullite and corundum phase in the diffraction pattern of sintered sample at 1300
o
C. Some
intensities of corundum peaks decreased by increasing sintering temperature from 1300
o
C to
1350
o
C but there was no difference between mullite phase peaks observed in the patterns of
sintered samples at both temperatures. The intensity of corundum phase peaks increased by
increasing sintering temperature from 1350
o
C to 1400
o
C but there was no change in the peak
intensity of mullite phase. As a result, it was found that the increase in the sintering temperature
was not enough in the transformation of mullite phase completely but the phase transformation of
the samples sintered at 1350
o
C is better than the others. In addition, it was mentioned that the
mullite phase transformation was successfully completed in a literature with kaolin and alumina
mixture by microwave heating at 1312
o
C [5]. For this reason, it was considered that the sintering
temperature should be proper between 1300-1350
o
C for mullite phase transformation but it was
also dependant on the duration for given temperature range.
0 10 20 30 40 50 60 70 80 90
1300 1h 1350 1h 1400 1h
: Mullite
: Corundum

Figure 7. The XRD patterns of the samples (Seydiehir Alumina + Kaolin mixture) sintered by
microwave heating at different temperatures and durations.

Figure 8 and 9 shows the SEM images of mullite samples sintered using conventional and
microwave heating processes. The major difference in the microstructures of sintered samples is
the shape of grains. The structure consists of larger elongated primary mullite grains and smaller
equiaxed secondary mullite grains in the conventionally sintered samples. However the grains are
of only spherical shape in the samples undergone microwave sintering process. The grain size is
in the range from 1 to 5 microns both in the conventionally and microwave sintered samples.

Figure 8. SEM micrograph of the samples sintered by conventional heating at 1600
o
C for 4 h.

Figure 9. SEM micrograph of the samples sintered by microwave heating at 1400
o
C for 1 h.

4. Conclusions

Mullite phase in powder form has been successfully synthesized using raw materials such as
calcined Seydisehir alumina, kaolin, gibbsite and magnesia which are sintered by using
conventional and microwave heating. Sintering experiments carried out up to 1400
o
C for
conventional sintering has shown that Seydisehir alumina has to be transformed into -Al
2
O
3
to
contribute mullite transformation. Compacted raw materials transform to mullite phase with a
small amount of corundum phase after microwave sintering at 1350
o
C for 1h. It is found that
unreacted corundum grains grow at 1400
o
C during microwave sintering. According to
conventional sintering results, mullite phase transformation was fully completed at 1600
o
C for 4h.
Further studies will be focussed on several stoichiometric composition of raw materials with
longer sintering durations to complete mullite transformation at 1350
o
C by microwave sintering.

References
[1] V. Viswabaskaran, F.D. Gnanama, M. Balasubramanian, Effect of MgO, Y
2
O
3
and boehmite
additives on the sintering behaviour of mullite formed from kaolinite-reactive alumina, Journal of
Materials Processing Technology 142 (2003) 275281
[2] E. Roncari, C. Galassi and C. Bassarello, Mullite Suspensions for Reticulate Ceramic
Preparation, J. Am.Ceram Soc. 83, [12], 2993-98, (2000)
[3] J. Anggono, Mullite Ceramics: Its Properties, Structure, and Synthesis,JURNAL TEKNIK
MESIN Vol. 7, No. 1, April 2005: 1 10
[4] A. K. Chakraborty, DTA study of preheated kaolinite in the mullite formation region,
Thermochimica Acta 398 (2003) 203209
[5] M. Panneerselvam and K. J. Rao, Novel Microwave Method for the Synthesis and Sintering of
Mullite from Kaolinite, Chem. Mater. 2003, 15, 2247-2252

Microwave Ignited Combustion Synthesis of Intermetallic-Based
Functionally Graded Materials

R. Rosa, P. Veronesi, C. Leonelli, A. B. Corradi
Department of Materials and Environmental Engineering, University of Modena and Reggio
Emilia, via Vignolese 905, 41100 Modena, Italy

Abstract
Combustion synthesis (CS) of different functionally graded materials (FGMs) based on
intermetallic alloys were ignited by 2.45 GHz frequency microwaves, in order to obtain
candidate materials for high temperature structural and coating applications. Combining CS
with a moderate pressure, to promote porosity reduction, it was possible to obtain both
freestanding samples and hard coatings on non-ferrous alloys substrates. The high
temperature generated due to the high exothermicity of CS, and the selective microwave
absorption by the reacting species, led to the formation of complex binary and ternary
compounds and microstructures, at the interface between the reacted powders and the
substrate, presenting, in some cases, an increased micro-hardness value with respect to
parent materials.

Introduction
A great number of intermetallic compounds and alloys can be synthesized directly starting
from the pure components suitably mixed together. One of the most promising methods in
terms of energy and time savings, purity of the products and low-cost equipments is the so-
called combustion synthesis (CS).
CS involves highly exothermic reactions, usually between powder-reactants, which, after
being ignited by a proper external ignition source, become self-sustaining and yield the final
product without requiring additional heat from the outside.
Two different modes can be used to conduct CS: the self-propagating high-temperature
synthesis (SHS) mode and the thermal explosion (TE) mode. In the SHS the reaction is
ignited at one end of the sample and it self-propagates trough the reactants in the form of a
combustion wave, while in a typical TE experiment the whole volume of the sample is
uniformly heated in a controlled manner until reaction takes place simultaneously throughout
the volume [1].
Among the large number of methods used to ignite CS processes, microwaves (MW)
appeared as a very attracting one [2] due to the possibility of providing rapid, volumetric and
energy efficient way of ignition of loose or compacted powders. Microwave assisted
combustion synthesis (MACS) can also been combined with the application of pressure
during synthesis in order to reduce the products porosity and to obtain well adhered coatings
on metallic substrates [3].
The aim of this work is to present the possibility to extend our microwave assisted
combustion synthesis and compaction (MACSC) approach to the synthesis of FGMs, both as
freestanding samples or as protective coatings on non-ferrous alloys. In fact although some
papers appeared in literature in the last years dealing either with MACS [3-6] or with the CS
of FGMs [7-10], no studies were conducted up to now, to the best of authors knowledge,
regarding microwave assisted combustion synthesis of functionally graded materials.
In this work, MACSC is applied to the preparation of NiAl-CoAl based FGMs, starting from
the precursor powders, as freestanding samples or as coatings on grade 5 titanium.

Experimental
Materials
Starting materials were
Ni powders, <400 mesh, 99.7% purity (Sigma Aldrich, Milan, Italy)
Al powders, -200 mesh, 99% purity (Sigma Aldrich, Milan, Italy)
Ti powders, -325 mesh, 99.5% purity (Alfa Aesar, Karlsruhe, Germany)
Co powders, <400 mesh, 99.8% purity (Alfa Aesar, Karlsruhe, Germany)
Formation of graded powder compacts: general procedure
The obtainment of the FGMs started from the forming step, using powder metallurgy
techniques, where graded green compacts of powders were prepared by stacking different
powder layers, whose composition is reported in Table 1.

Table 1. Composition and arrangement of each graded powders compact prepared.

Sample Description
Layer composition
(wt%)
Layer mass
(g)
layer
NC2F
2 layers
freestanding
FGM
Ni+Al (50:50 at%) Co+Al (50:50 at%)
0 100 3 1
st

100 0 3 2
nd

NC6F
6 layers
freestanding
FGM

Ni+Al (50:50 at%) Co+Al (50:50 at%)
0 100 1 1
st

20 80 1 2
nd

40 60 1 3
rd

60 40 1 4
th

80 20 1 5
th

100 0 1 6
th

NC2CT
2 layers FGM
coating on Ti
grade 5
Ni+Al (50:50 at%) Co+Al (50:50 at%)
0 100 3 1
st

100 0 3 2
nd

NC6CT
6 layers FGM
coating on Ti
grade 5
Ni+Al (50:50 at%) Co+Al (50:50 at%)
0 100 1 1
st

20 80 1 2
nd

40 60 1 3
rd

60 40 1 4
th

80 20 1 5
th

100 0 1 6
th

Each base powder mixture, pertaining to each layer, was separately mixed in an Al
2
O
3

ceramic jar for 30 minutes under vacuum, in the indicated molar ratio of the constituents. The
powders, in the sequence indicated in the last column of table 1, were placed in a cylindrical
metal die and cylindrical pellets of 20 mm diameter were obtained by uniaxial pressing at 150
MPa.

MACS Experimental apparatus
All the microwave assisted combustion syntheses were performed in a single mode
microwave applicator operating at 2.45 GHz, presenting 2 facing cylindrical ports under cut-
off to allow pressure application to the reacting species and to avoid microwave leakage. The
applicator is fed by a generator (magnetron TM030, Alter, Reggio Emilia, Italy), via a
waveguide (WR340) and a three-stubs tuner, and is terminated by a shorting plunger. The
water-cooled magnetron output power can be varied from 300 to 3000 W. A Three ports
circulator with a water load was used to protect the magnetron from reflected power. As
reported in Figure 1 a), the application of a slight pressure on the reacting sample was made
possible by some loads positioned on the top of two refractory cylinders fitting perfectly with
the cavity dimensions and perturbing as less as possible the electromagnetic field
distribution.

MACSC: general procedure
Figure 1 b) represents a cross sectional view of the samples arrangement inside the MW
applicator. A MW co-absorber made of silicon carbide (SiC) was positioned on the top of the
bottom refractory plate in order to increase the temperature of the load and to improve its
microwave coupling efficiency [5], leading to a sort of hybrid heating system.
On the top of the SiC disc, the reacting green compact was placed between two AISI304
stainless steel discs, 20 mm diameter, 3 mm thick, which in our experiments have showed
their relative inertness toward the reacting species used. Specimens CT (i.e Coating on
Titanium, namely NC2CT and NC6CT), instead, were realized replacing the bottom steel disc
with a grade 5 titanium disc having the same dimensions. Under continuous 20 Nml/min Ar
flux, a MW output power of 1.2 kW was applied to ignite the combustion synthesis of each
sample, and simultaneously, a moderate pressure of 0.8 MPa was applied.

Figure 1. a) the microwave single mode applicator used for all MACSC experiments. b) cross-section
view of how the samples were positioned inside the MW applicator during the combustion synthesis.

Characterization
The cross sections of reacted specimens were subjected to scanning electron microscopy
(ESEM Quanta-200 Fei, Oxford Instruments) to analyze the microstructures and the interface
between Ti and FGM coatings. Energy dispersive X-ray microanalysis spectroscopy (X-EDS,
Oxford, INCA-350) was used to perform semi-quantitative elemental analysis in order to
investigate the atomic ratio of constituent phases, as well as the composition of the
interfaces and to obtain elements distribution maps along the cross sections. Vickers micro-
hardness variations along the cross section were measured applying a constant load of 0.98
N for 15 seconds (Wolpert group, 402 MVD).

In all the performed CS experiments, the ignition of the compacted powders pellets was
reached in few seconds (generally less than 30 s, depending on the sample mass and
composition). After ignition occurs, the system control unit, connected to a light sensor
positioned along the waveguide, immediately turns off MW power generation. This allows
reducing the uncertainties, which could rise in case of different microwave exposure times
after reaction occurs. As a matter of fact, microwaves can continue to provide energy to the
reacting specie, altering the final microstructure and extending the time of existence of the
liquid phase [11]. According to the applied experimental setup, in all experiments the first
layer of all samples from Table 1 was the one closer to the SiC co-absorber.
SEM micrographs and elements distribution maps of freestanding FGMs of intermetallic
alloys CoAl-NiAl, samples NC2F and NC6F, are reported in figures 2 and 3 respectively.
Elemental distribution and EDS spectra results confirmed the obtainment of the desired
atomic ratios (1:1) in the as synthesized NiAl and CoAl intermetallics, corresponding to the
two outer layers. Moreover, it is evident the presence of a compositional gradient of Ni and
Co, while the Al distribution, as expected, remains almost constant (relative elemental maps
shown in Figure 2 must take into account the higher porosity near the CoAl layer,
progressively decreasing towards the NiAl one). Thus, despite the formation of a liquid phase
during CS, and the application of pressure during synthesis, the original FGM structure of the
green compact is retained, indicating that no mixing occurs. This could be explained
considering the high cooling rates occurring when microwave power is turned off, which
numerical simulation, for a similar configuration, estimated in more than 600 Ks
-1
[11].
Therefore, the liquid phase formed as CS proceeds is rapidly quenched and no significant
mixing occurs.

NC2F section Co map

Ni map Al map

Figure 2. SEM micrograph of NC2F cross section and elemental points maps.

NC6F section Co linescan

Ni linescan Al linescan

Figure 3. SEM micrograph of the NC6F cross section and elements linescan along the indicated line.

Scanning electron micrographs of the cross sections of FGMs coatings on grade 5 titanium,
samples NC2CT and NC6CT, with some higher magnification micrographs of the Ti/CoAl
interface are shown in figure 4, comprehensive of the values of HV
0.1
measured at the
interface and the EDS results. The CoAl-richer part of the FGM presents a much higher
porosity with respect to the outer NiAl richer region, despite the application of pressure
during synthesis. The higher porosity of cobalt mono-aluminide, however, is in full agreement
with literature results [12]. Porosity of the 6-layer sample (NC6CT) is higher with respect to
1 mm
NC2CT sample. This could be probably ascribed to the higher thickness of NC6CT sample,
which hinders the removal from the transient liquid phase of gases formed during CS.

Cross section:
left: Ti - right: FGM
Ti/CoAl interface
left: Ti right: CoAl
EDS results
at%
N
C
2
C
T

zone Ti Al V Co
A 86.6 10.3 3.1 -
B 51.5 26.5 2.0 20.0
C - 51.1 - 48.9
D 0.6 39.6 - 59.8
HV
0.1
= 528.957.8
N
C
6
C
T

zone Ti Al V Co
E 79.2 13.5 3.4 3.9
F 55.4 25.3 2.6 16.7
G - 50.9 - 49.1
HV
0.1
= 591.524.7
Figure 4. SEM micrographs, EDS results and Vickers micro-hardness values for FGMs coating on Ti-
6Al-4V.

The higher porosity tends to lead to a less homogenous interface with the underlying Ti
substrate, but in both samples the measured HV
0.1
values at the interface are similar and
much higher than the measured average values of grade 5 titanium (HV
0.1
=283.933.4), pure
CoAl (HV
0.1
=257.451.8) and pure NiAl (HV
0.1
=287.789.9). The micro-hardness increase
could be explained considering the formation of more complex ternary intermetallic systems
at the interfaces between Ti substrate and CoAl, as emerged by EDS analyses, and in
agreement with literature results [13].
The NiAl/CoAl interfaces, far from the Ti substrate, appear continuous and interpenetrated,
as shown in figure 5.

NC2CT
CoAl/NiAl interface
EDS results, at%
NC6CT
CoAl/NiAl interface

zone Ni Al Co

H 50.1 49.6 0.3
I - 48.0 52.0
J 13.7 49.7 36.6
K 49.9 50.1 -
HV
0.1
= 361.038.7 HV
0.1
= 440.4112.5
Figure 5.SEM micrographs and EDS results of CoAl/NiAl interfaces.

Figure 5 shows also that the micro-hardness values are significantly lower than the Ti-
containing intermetallics formed at the interface with the substrate, but still higher than the
parent NiAl and CoAl. This suggests that compositional or microstructural changes occurred
and probably different phases have been formed at the interface. The possible formation of
NiAl(Co) alloys [14] should be more accurately investigated in a future work, by XRD.
Further investigation will be conducted to assess possible applications of such systems as
aluminium rich intermetallic bond coat and source of thermally grown oxide (TGO) in thermal
barrier coatings. Also the resulting porous microstructure could be useful to increase the
radiative surface and the efficiency of thermal protection systems. The possibility of further
improving our method applying higher pressure during synthesis is currently under study, as
well as the extension of MACSC to FGMs with a compositional gradient from intermetallic to
cermets and to pure ceramics.

Conclusions
The possibility of extending our consolidated MACSC approach to the synthesis of FGMs
based on intermetallic alloys was demonstrated. Compositional gradients of different reactive
powders layers were preserved after CS, due to the rapid quenching of the transient liquid
phase originated by the reaction. Both freestanding samples and coating on grade 5 titanium,
belonging to the system NiAl-CoAl were synthesized in less than 30 seconds applying
1200W of microwave power. In case of coatings, at the interface with the titanium substrate
there is the formation of Ti-containing intermetallics, with an interlocking microstructure,
which presents a much higher hardness with respect to parent materials. Interface between
NiAl and CoAl is continuous and Vickers microhardness is higher than NiAl and CoAl one,
probably because of the formation of NiAl(Co) alloys.

References
[1] A. Varma, J.P. Lebrat, Combustion synthesis of advanced materials, Chem. Eng. Sci. 47,
1992, pp. 2179-2194.
[2] S. Gedevanishvili, D. Agrawal, R. Roy, Microwave combustion synthesis and sintering of
intermetallics and alloys, J. Mater. Sci. Letts. 18,1999, pp. 665-668.
[3] G. P. Cammarota, A. Casagrande, G. Poli, P. Veronesi, Ni-Al-Ti coatings obtained by
microwave assisted SHS: effect of annealing on microstructural and mechanical properties,
Surf. Coat. Technol. 203, 2009, pp. 1429-1437.
[4] K. Naplocha, K. Granat, Combustion synthesis of Al-Cr performs activated in microwave
field, J. Alloys Compds., 2008, doi: 10.1016/j.jallcom.2009.02.085.
[5] G. Poli, R. Sola, P. Veronesi, Microwave-assisted combustion synthesis of NiAl
intermetallics in a single mode applicator: modeling and optimisation, Mater. Sci. Eng. A 441,
2006, pp. 149-156.
[6] J. R. Jokisaari, S. Bhaduri, S. B. Bhaduri, Microwave activated combustion synthesis of
titanium aluminides, Mater. Sci. Eng. A 394, 2005, 385-392.
[7] D. E. Burkes, J. J. Moore, Microstructure and kinetics of a functionally graded NiTi-TiC
x

composite produced by combustion synthesis, J. Alloys Compds. 430, 2007, 274-281.
[8] N. Bertolino, M. Monagheddu, A. Tacca, P. Giuliani, C. Zanotti, F. Maglia, U. Anselmi
Tamburini, Self-propagating high-temperature synthesis of functionally graded materials as
thermal protection systems for high-temperature applications, J. Mater. Res. 18, 2003, pp.
448-455.
[9] M. Cirakoglu, S. Bhaduri, S. B. Bhaduri, Combustion synthesis processing of functionally
graded materials in the Ti-B binary system, J. Alloys Compds. 347, 2002, pp. 259-265.
[10] Y. Zhang, J. Han, X. Zhang, X. He, Z. Li, S. Du, Rapid prototyping and combustion
synthesis of TiC/Ni functionally gradient materials, Mat. Sci. Eng. A 299, 2001, pp. 218-224.
[11] P. Veronesi, C. Leonelli, G. Poli, A. Casagrande, Enhanced reactive NiAl coatings by
microwave assisted SHS, Compel 27, 2008, pp. 491-499.
[12] C. L. Yeh, C. C. Yeh, Preparation of CoAl intermetallic compound by combustion
synthesis in self-propagating mode, J. Alloys Compds. 388, 2005, pp. 241-249.
[13] K. Ishikawa, Y. Himuro, I. Ohnuma, R. Kainuma, K. Aoki, K. Ishida, Phase equilibria in
the Co-Ti portion of the Co-Al-Ti ternary system, J. of Phase Equil. 22, 2001, pp. 219-226.
[14] L. Chen, Y. Han, The microstructure and compressive properties in NiAl(Co) alloys by
HPXD technique, Mater. Sci. Eng. A 329-331, 2002, pp.725-728.
Euro PM2009 Novel Sintering Approaches
Manuscript refereed by Professor Detlev Stoever, Forschungszentrum J uelich GmbH,
Germany
Additive Manufacturing of Dense Metal Parts by Electron Beam
Melting

U. Ackelid and M. Svensson, Arcam AB, Krokslatts Fabriker 27A, 431 37 Molndal, Sweden

Abstract
In recent years, Electron Beam Melting (EBM) has matured as a technology for additive
manufacturing of dense metal parts. The parts are built by additive consolidation of thin layers of
metal powder using an electron beam. With EBM, it is possible to create parts with geometries
too complex to be fabricated by other methods, e.g. fine network structures and internal cavities.
The process is run in a vacuum chamber, which makes it well suited for materials with a high
affinity to oxygen, e. g. titanium alloys. A recent study of EBM built Ti6Al4V, including chemical,
microstructural, tensile and fatigue data, shows that the material is comparable to wrought
Ti6Al4V.

Introduction
In the past two decades, several technologies for additive fabrication of metal parts from
metal powder have emerged [1]. The basic idea is to build up parts in thin layers of powder, each
of which is consolidated by an energy source to the precise geometry defined by a computer
model. The energy source is usually a laser beam or an electron beam. This layer-based
fabrication technique gives a geometrical freedom larger than conventional fabrication methods. It
is possible to create metal parts with very complex shapes, including e.g. fine network structures,
internal cavities and engineered surface porosity. Figure 1 shows an example of a product
currently in series production.
Electron Beam Melting (EBM) is an additive manufacturing technology developed by Arcam
AB in Sweden [2]. A schematic drawing of an EBM machine is shown in Figure 2. The principle is
very similar to that of a scanning electron microscope. A heated tungsten filament emits electrons
which are collimated and accelerated to a kinetic energy of ~60 keV. The electron beam is
controlled by two magnetic coils. The first one is a magnetic lens which focuses the beam to the
desired diameter, and the second one deflects the focused beam to the desired point on the
building table. The electron beam gun itself is fixed; thus there are no moving mechanical parts
involved in the beam deflection. The beam current is controlled in the range 1- 50 mA and the
beam diameter can be focused down to ~0.1 mm.

Figure 1. Orthopedic hip implant (acetabular cup) in titanium,
built with EBM. The implant has a tailored surface design for
optimized bone ingrowth. Two European manufacturers are
currently producing similar implants with Arcam EBM
technology.
Figure 2. Schematic drawing of an Arcam EBM machine.

In the chamber in the lower part of Figure 2, metal powder layers are formed by a powder
raking mechanism. The typical layer thickness is 0.1 mm. The computer-controlled electron beam
scans over the powder layer in a predefined pattern and consolidates the desired areas into solid,
dense metal. A new powder layer is created on top and the melt process is repeated. In this
manner, the parts are built up layer-by-layer. Melting one layer takes typically 1-2 minutes,
corresponding to a speed of 3-6 mm build height per hour. The maximum powder area within
reach of the electron beam is a circle of diameter 300 mm.
The whole process takes place in high vacuum. The typical pressures in an Arcam EBM
machine are 10
-3
Pa in the vacuum chamber and 10
-5
Pa in the electron gun. A high vacuum is
necessary for many reasons: (i) it maintains a high quality electron beam since collisions between
electrons and gas molecules are avoided, (ii) it provides a clean and safe environment for molten
metals, especially highly reactive metals like titanium, (iii) it helps to outgas impurities
incorporated in the metal powder, and (iv) it gives good thermal insulation.
When all layers have been melted, the built parts are allowed to cool inside the process
chamber, which is filled up with helium gas to speed up the cooling. When the parts are taken out
from the EBM machine, they are surrounded by a shell of softly agglomerated metal powder. This
powder can easily be disintegrated, sieved and reused in the next build. Thus the material waste
is kept to a minimum. All powder that is not melted by the electron beam is recycled.
The Arcam EBM technology has been available commercially for about six years. So far, the
main focus has been on titanium parts for medical implants and aerospace applications. One
successful example is the series production of titanium acetabular cups conducted by two
medical implant manufacturers in Europe (Figure 1). The energy density of the electron beam is
high enough, however, to melt almost any metal or alloy. Laboratory tests by Arcam and other
EBM users have indicated that the EBM process has potential to work with many material
classes, for example stainless steel (17-4), tool steel (H13), Ni-base superalloys (IN625 and
IN718), Co-base superalloys (Stellite 21), Low-expansion alloys (Invar), hardmetals (NiWC),
intermetallic compounds (TiAl), aluminum, copper, beryllium and niobium.

Material Properties of Ti6Al4V and Ti6Al4V-ELI Manufactured by EBM
The alloy Ti6Al4V (Ti grade 5) and its extra-low interstitial counterpart Ti6Al4V-ELI (Ti grade
23) are among the most widely used titanium alloys. Recently, a comprehensive study of material
properties of EBM manufactured Ti6Al4V and Ti6Al4V-ELI was conducted at Arcam. The
investigation comprised chemical analyses, microstructure examination, porosity measurement,
static tensile testing, axial high cycle fatigue testing (HCF) and rotating beam high-cycle fatigue
testing (RBF). The purpose was to establish the quality of material built with the latest version of
the EBM software (EBM Control 3.1) and the latest set of beam control parameters. Test
specimens were built with Arcams standard powder, which is an argon gas atomized powder of
size 45-105 m (-140/+325 mesh). In order to check the isotropy, the specimens were tested in
two directions:
the horisontal direction, parallel to the powder layers, denominated XY
the vertical direction, perpendicular to the powder layers, denominated Z
The materials were also tested in two conditions, to investigate the influence of porosity:
untreated, such as they came out from the EBM process, denominated As-built or RAW
after hot isostatic pressing (HIP), denominated HIP
The HIP treatment was performed in argon gas at a pressure of 100 MPa and a temperature of
920 C for 120 minutes.

Chemical Composition
Table 1 shows the results of the chemical analyses. For comparison, the requirements in two
common ASTM standards for surgical implants, ASTM F1472 (non-ELI) and ASTM F136 (ELI),
are also given. It is seen that the composition of the EBM materials comply with the standards.
No contamination by foreign elements could be seen. Note that the composition of as-built and
HIP materials are the same, except for the oxygen content which appears to be higher for the HIP
materials. This oxygen increase is not caused by the HIP treatment. The reason for the higher
oxygen in the HIP materials is that the HIP specimens were built at a later stage, when the metal
powder had been recycled several times through the process. It is known that powder recycling in
the EBM process gives a small oxygen pickup.

Table 1. Chemical composition of the EBM manufactured specimens.
Others include the elements Cu, Mn, Mo, Sn and Zr. All concentrations in weight-%.
The requirements in ASTM F1472 and ASTM F136 have been added for comparison.
C Al Fe H N O V Ti Y Others
Ti6Al4V As-built 0,010 5,80 0,04
0,000
4
0,014 0,16 3,87 Bal.
<0,00
1
<0,01
Ti6Al4V HIP 0,010 5,89 0,05
0,001
1
0,014 0,19 3,91 Bal.
<0,00
1
<0,01
ASTM F1472 <0,08
5,5-
6,75
<0,30
<0,01
5
<0,05 <0,20
3,5-
4,5
Bal.
<0,00
5
-
Ti6AL4V-ELI As-
built
0,009 5,82 0,04
0,000
5
0,020 0,098 4,14 Bal.
<0,00
1
<0,01
Ti6Al4V-ELI HIP 0,011 5,79 0,04
0,001
2
0,022 0,12 4,14 Bal.
<0,00
1
<0,01
ASTM F136
(ELI material)
<0,08 5,5-6,5 <0,25
<0,01
2
<0,05 <0,13
3,5-
4,5
Bal. - -

Microstructure
Figure 3 shows the microstructures found for the eight possible combinations of material,
post-treatment and direction. The general microstructure consists of fine elongated epitaxial
grown -grains, with a secondary -phase grown from former -boundaries. The shape of the -
grains is acicular or plate-like with a Widmansttten structure. The colony size of Ti6Al4V is
measured as the size of the grouping of equally aligned acicular alpha plates within one former
beta grain. The colony size is an important indication of fatigue properties. In EBM manufactured
material some variation of colony size exists. In general it is extremely fine with an almost
singular, and evenly distributed Widmansttten structure, whereas some areas are more aligned
with much larger clusters of -plates, originating from former -boundaries.
Figure 4a shows spherical pores of size 50 m which were occasionally seen in as-built
material. Previous tests at Arcam have shown that this type of porosity originates from the
powder itself. The powder usually contains some entrapped microporosity, i.e. argon gas bubbles
from the gas atomization process (Figure 4b). Due to the fast melting and resolidification in the
EBM process, some spherical pores from the powder actually survives into the final as-built
material. If a powder without entrapped porosity had been used, e.g. a PREP powder, then there
would be no spherical pores in the final material. These spherical pores are so small that they are
safely eliminated by the HIP treatment.

Porosity Measurement
The total porosity in the powders and in the EBM manufactured specimens was quantified by
helium pycnometry. The results presented in Table 2 confirm that
the powders contain a small quantity of entrapped gas
some gas bubbles remain also in the EBM built material
the porosity is removed by hot isostatic pressing

Table 2. Total internal porosity in powders and solid specimens
Porosity [%]
Powder Ti6Al4V 0.27
Powder Ti6Al4V-ELI 0.17
Solid Ti6Al4V, As-built 0.19
Solid Ti6Al4V, HIP 0.01
Solid Ti6Al4V-ELI, As-built 0.11
Solid Ti6Al4V-ELI, HIP 0.02

Static Tensile Properties
The results of the static tensile test are shown in Table 3. The tensile requirements in ASTM
F1472 and ASTM F136 are given for comparison. Observe that:
All measured data comply with the standards.
The ELI material has lower strength than the high oxygen material. This is expected since
oxygen is known to strengthen titanium.

Figure 3. Microstructures of EBM manufactured Ti6Al4V and Ti6Al4V-ELI, before and after HIP. The bars in the
lower right corners correspond to 20 m.

Figure 4a. An example of spherical microporosity in as-built Ti6Al4V. Such pores originate from internal pores in
gas atomized metal powder (see Figure 4b). No spherical porosity was seen in the HIP material.
Figure 4b. A polished cross-section of gas atomized Ti6Al4V particles with internal spherical pores.
4a
4b
Ti6Al4V-ELI, HIP, Z Ti6Al4V-ELI, HIP, XY
Ti6Al4V-ELI, as-built, Z Ti6Al4V-ELI, as-built, XY
Ti6Al4V, HIP, Z
Ti6Al4V, HIP, XY
Ti6Al4V, as-built, Z Ti6Al4V, as-built, XY
The tensile strength appears to be 1-2% higher in the XY direction. Thus for most practical
purposes the tensile properties can be considered as isotropic.
The difference between the as-built and the HIP specimens is small, indicating that there is no
gross porosity in the materials.

Table 3. Static tensile test data of the EBM manufactured Ti6Al4V and Ti6Al4V-ELI
Four tensile bars were tested in each direction. S.D. = Standard Deviation
YTS
R
p0.2
[MPa]
UTS
R
m
[MPa]
Elongation
A [%]
Area Reduction
Z [%]
Dir
.
Mean (S.D.) Mean (S.D.) Mean (S.D.) Mean (S.D.)
Ti6Al4V As-built
Z 879 (12,5) 953 (8,8) 13,8 (0,9) 46,0 (0,0)
XY 870 (8,1) 971 (3,1) 12,1 (0,9) 34,5 (2,6)
Ti6Al4V HIP
Z 868 (2,9) 942 (2,6) 12,9 (0,8) 43,5 (2,1)
XY 866 (6,4) 959 (8,2) 13,6 (0,6) 36,8 (2,2)
ASTM F1472 - >860 - >930 - >10 - >25 -
Ti6Al4V-ELI As-
built
Z 802 (7,9) 904 (6,0) 13,8 (0,9) 47,0 (2,4)
XY 817 (4,3) 918 (1,0) 12,6 (0,8) 42,5 (2,5)
Ti6Al4V-ELI HIP
Z 807 (8,4) 902 (8,7) 14,8 (0,5) 46,8 (2,9)
XY 814 (2,4) 916 (2,5) 13,6 (1,2) 35,3 (3,9)
ASTM F136 - >795 - >860 - >10 - >25 -

High Cycle Fatigue Properties
Figure 5 shows S/N data from the axial HCF test. The HCF test was performed with non-ELI
material only. Ten million cycles was chosen as run-out limit in this test, in accordance with
common practice within the medical implant industry. For as-built Ti6Al4V, the fatigue limits are
390 MPa in the Z direction and 450 MPa in the XY direction. For the HIP specimens, the fatigue
limits increase to 540 MPa in the Z direction and 550 MPa in the XY direction. However, note that
the HIP S/N curve in Figure 5 has not leveled off at 10
7
cycles.
There are no HCF requirements in the common ASTM standards for Ti6Al4V. The following
high-cycle fatigue limits can be found for Ti6Al4V in the literature [3]:
Cast material: 200-360 MPa at 10
7
cycles
Cast plus HIP: 200-520 MPa at 10
7
cycles
Wrought annealed: 400-680 MPa at 10
7
cycles
We thus conclude that the HIP EBM material has a HCF performance fully comparable to
wrought material, whereas the as-built EBM material is comparable to low-end wrought or cast
plus HIP material. The fact that HIP improves HCF performance indicates that there are some
remaining defects in the as-built material, e.g. microporosity. This was confirmed by the
micrographs (Figure 4).

300
400
500
600
700
800
900
10 000 100 000 1 000 000 10 000 000
M
a
x
-
s
t
r
e
s
s

(
M
P
a
)
N Cycles to failure
RAW XY
RAW XY Run-outs
RAW Z
RAW Z Run-outs
HIP XY
HIP XY Run-outs
HIP Z
HIP Z Run-outs

Figure 5. High-cycle fatigue data of EBM built Ti6Al4V, before and after HIP. Axial load, R = +0.1, frequency 40
Hz, run-out limit 10
7
cycles.
Rotating Beam Fatigue Properties
Figure 6 shows S/N-curves from the RBF test. Ten million cycles was chosen as run-out limit
also in this test. The following fatigue limits were found:
As-built Ti6Al4V: 296 MPa at 10
7
cycles
HIP TI6Al4V: 518 MPa at 10
7
cycles
As-built Ti6Al4V-ELI: 257 MPa at 10
7
cycles
HIP Ti6Al4V-ELI: 518 MPa at 10
7
cycles
The fact that HIP improves RBF performance indicates that there are some remaining
defects in the as-built material, e.g. microporosity. This was confirmed by the micrographs (Figure
4). After HIP, the EBM manufactured material has a rotating-beam fatigue limit comparable to
literature data: 585 MPa [4] and 528 MPa [5].

200
300
400
500
600
700
10 000 100 000 1 000 000 10 000 000
L
o
a
d

(
M
P
a
)
Cycles to failure
Ti6Al4V RAW
Ti6AL4V HIP
Ti6AL4V ELI
Ti6Al4V ELI HIP

Figure 6. Rotating beam fatigue (RBF) data of EBM built Ti6Al4V and Ti6Al4V-ELI, before and after HIP. R = -1,
frequency 133 Hz, run-out limit 10
7
cycles

Conclusions
Some basic material properties of solid Ti6Al4V and Ti6Al4V-ELI manufactured from gas-
atomized metal powder by Electron Beam Melting have been studied. The results show that
The chemical composition fully complies with ASTM standards.
The microstructure is a fine acicular secondary -phase grown in prior -grains.
There is some spherical microporosity originating from argon gas entrapped in the powder.
Hot isostatic pressing eliminates all microporosity.
The static tensile test data also comply with ASTM standards.
The fatigue performance (HCF and RBF) of as-built material is somewhat impeded by the
microporosity. After hot isostatic pressing, however, the fatigue data is similar to wrought
Ti6Al4V.

References
[1] www.wohlersassociates.com
[2] www.arcam.com
[3] M. J. Donachie, Jr., Titanium, A Technical Guide, Second Edition, ASM International, 2000.
[4] L. D. Zardiackas, M. J. Kraay and H. L. Freese, Titanium, Niobium, Zirconium, and Tantalum
for Medical and Surgical Applications, ASTM International, 2006.
[5] S. A. Brown and J. E. Lemons, Medical Applications of Titanium and Its Alloys: The Material
and Biological Issues, ASTM International, 1996.
Manuscript refereed by Professor Detlev Stoever, Forschungszentrum Juelich GmbH,
Germany

Influence of Compaction on Sintering of a Two Layer Component

A. Thomazic
a
, Y. Le Guennec
b
, Y. Kamdem
a
, C. Pascal
a
, J.M. Chaix
a
, P. Doremus
a
,
D. Imbault
b
, D. Bouvard
a
, F. Dor
c

a
SIMaP, CNRS/Grenoble INP/UJF, BP 75, 38402 Saint-Martin-d'Hres, France
b
3S-R, CNRS/Grenoble INP/UJF, BP 53, 38041 Saint-Martin-d'Hres, France
c
CETIM, BP802, 42952 Saint-Etienne Cedex 9, France
Abstract
The association of a hard and a tough material in one component can be achieved by
powder metallurgy, but many difficulties must be overcome for compaction and sintering.
This paper deals with a two layer component, a WC-Fe cemented carbide and a Fe-W-C
steel, both elaborated from tungsten carbide, iron and graphite powders. The two layers are
co-compacted then co-sintered.
In this paper, the experimental and numerical approaches of compaction and sintering are
associated to understand the involved phenomena and improve the process.
1 Introduction
Two layer components allow the combination of properties (mechanical, thermal, electrical...)
that could not be achieved with a single material. Multi-layer components are conventionally
obtained by processes which imply intermediate steps of alloy manufacturing, assembling
and machining. The powder metallurgy process is an appropriate route to obtain directly this
product with the required dimensions and shapes. Post-sintering machining is reduced or
even useless.
In order to shape multi-layer components, the compaction step is critical to obtain the
expected geometry and a good initial density repartition and to minimize mechanical strains
during sintering. Moreover, compositional difference between the two materials could induce
chemical phenomena at the interface between layers [1, 2]. This study aims to understand
the influence of compaction and thermal treatment parameters on sintering of a two layer
component, which associates these two materials in order to combine hardness and
toughness properties.
In a previous investigation [3], the composition of each layer was selected from physico-
chemical data and sintering compatibility. In order to understand the influence of compaction,
tests have been performed on single layer components, then on two layer components.
Different thermal cycles have been tested. Experimental results were compared to numerical
simulations to describe the process and highlight the differences between experimental and
numerical results.
2 Experimental procedure
In this study, to ensure chemical compatibility and to limit the number of elements, the hard
material (WC base material) consists of tungsten carbide hard particles cemented by an iron
base tough binder whereas cobalt is employed in classic cemented carbides [4]. The second
material (Fe base material) is mainly composed of iron, with a small amount of tungsten
carbide and graphite powders.
The materials were prepared from WC, Fe (Eurotungstene Poudres) and C
graphite
(Hgans)
powders (table 1). Similar average WC and Fe particle sizes are used for the WC base
powder mixture while for the Fe base one, finer WC particles are chosen to improve carbide
distribution in iron rich matrix and to reduce residual porosity after sintering [3]. The
compositions of powder mixtures are reported in table 1.
The mixtures were processed through attritor-milling with 2wt.% polyethylene glycol (PEG
3400) dissolved in acetone media. Organic binder is needed to promote compaction and to
improve green strength in order to enable eventual green machining. Powder mixtures were
Material Fe base layer WC base layer
Composition [wt.%] Fe-7 %WC-0.6 %C WC-6.6 %Fe-0.14 %C
Powder designation Fe2010 CW5200 C-KS44 Fe2010 CW5722 C-KS44
Typical laser D50 (m) 6.8 1.6 18.6 6.8 6.1 18.6
Table 1: Compositions and characteristics of the powders
dried and granulated through a 500 m sieve.
To compare the results, we used relative density evaluated from the ratio between
experimental density and theoretical final density. Experimental densities were calculated
from mass and volume. Theoretical densities were estimated from initial constitutive material
densities (14.6 for WC base material and 8.0 for Fe base material).
3 Results and discussion
3.1 Single layer components
3.1.1 Compaction
In order to predict the density distribution in a two layer materials, a numerical model was
fitted with experimental data obtained on single materials. Simulations have been run with
the finite element software ABAQUS [5] which included a Drucker-Prager constitutive
expression [6]. The parameters of this constitutive law for each material were fitted using
three different tests: diametral compression, uniaxial compression and an instrumented die
test [7].
A hydraulic press which can apply a force up to 10 tons was used. Radial stress applied to
the die by the powder was measured by a strain gauge located on the outer surface of the
die. Top and bottom punches were also equipped with strain gauges to evaluate axial
stresses during compaction. The punches diameter was 12.5 mm and their radial clearance
with the die was 0.2500.005 mm. The friction coefficient between powder and die, which
has a strong influence on the density distribution of the compact, was also determined on
both powders. We obtained a friction coefficient of 0.17 for the Fe base material and 0.22 for
the WC base one. To check the fitting of the Drucker-Prager model and friction coefficients,
numerical results are compared to the experiments after uniaxial die compaction (figure 1).
As tests as well as sample preparation were realized using a single-action die (mobile upper
punch), a device independent half height stress was used. Considering a linear evolution of
the density along the compact height, the mean density of the compact is related to the half-
height stress:
low up 2 / h
= [7]. For the two materials the experimental evolution of the
final density versus the half-height stress is correctly reproduced by the model.

3
4
5
6
0 200 400 600
Half-height stress (MPa)
D
e
n
s
i
t
y

(
g
/
c
m
3
)
Simulation
Experiment
6
7
8
9
10
11
0 200 400 600
D
e
n
s
i
t
y

(
g
/
c
m
3
)
Simulation
Experiment

Figure 1: Experimental evolution of the final density versus the half-height stress for
a) Fe base material and b) WC base material
3.1.2 Sintering
The powder mixtures were uniaxially compacted into cylindrical samples (=8 mm, H=6-
8 mm) under P=600 MPa. After compaction, the average values of green densities are 10.0
(69 %) for WC base compacts and 6.0 (74 %) for Fe base compacts. The thermal treatment
is carried out in a dilatometer to measure axial length changes during heating ( L/L
0
). Before
the sintering step, the compacts were held for one hour under He-4%H
2
flow at 360C to
remove organic binder. After the debinding stage, Fe base and WC base powder compacts
are heated in argon atmosphere at 5C/min and isothermally held at 1300C during 60
minutes (figure 2).
The experimental curves were discussed in a previous work [8]. The comparison of sintering
behaviours of Fe and WC base single materials indicates that dimensional changes occur at
different temperatures with different magnitudes. These differences could originate stresses
a) Fe base b) WC base
and cracks at the interface of the two material components during sintering process. After
cooling to room temperature, the experimental densities of the Fe base and WC base
materials are 7.7 (96%) and 13.8 (94%) respectively.
The numerical simulation is a helpful tool to anticipate the behaviour of bi-materials during
sintering. The constitutive laws of the two materials were defined with the help of dilatometry
experiments based on Gillias procedure [9]. Figure 2 shows that the numerical simulations
predict reasonably well the experimental data. The deviation observed at the beginning of
each thermal treatment (below 1000C) is due to phenomena such as debinding step or
phase transformation. These phenomena are not taken into consideration in this model.

a)

b)
Figure 2: Experimental dilatometric curves compared to numerical simulations for the
sintering of a) Fe base and b) WC base single materials (1300C-60 min)
3.2 Two layer components
3.2.1 Influence of sintering
Several sintering parameters are expected to monitor the properties of the sintered two layer
material: graphite content in Fe base layer, co-compaction pressure, heating rate, sintering
temperature and duration. Design of experiments (DOE) approach was used to analyse the
effect of these parameters on sintering with a limited number of experiments [10]. This study
highlights the formation of liquid in the Fe base layer at 1300C. Part of this liquid migrates
towards the WC base layer by capillary infiltration. This phenomenon has consequences on
both materials: geometrical deformation, lower densification, new brittle phase formation
(M
6
C) at the interface between layers...
For the two layer compacts, WC base mixture was poured first and slightly compacted in the
three part matrix (=8 mm). Then Fe base mixture was added and pressed at 400 MPa to
achieve a compact with 4 mm thick layers (figure 3). On figure 4, experimental result on the
1300C-60 min cycle is compared to numerical simulation. Model is in agreement with
experimental results until 1300C. During the isothermal stage at 1300C, the shrinkage rate
decreases in the first 10 min as the numerical simulation, then, shrinkage starts again. After

Figure 3: Schematic representation
of a two layer component
Figure 4: Experimental dilatometric curves compared to
numerical simulations of the bi-material sintered at 1300C
cooling to room temperature, the experimental densities of the Fe base and WC base layers
are 7.6 (95%) and 13.3 (91%) respectively. The axial shrinkage is twice higher (22%) than
the shrinkage reached in the numerical simulation (10%). This deviation is due to the liquid
formation and then liquid migration which cannot be predicted by numerical simulation. The
analysis of DOE combined to microstructural observations and numerical simulations
indicates that a sintering at 1290C during 30 min is a suitable cycle. This cycle was applied
for subsequent experiments.
3.2.2 Influence of compaction
3.2.2.1 Compaction of two layer components
Compaction determines the density distribution inside the green material. For a two layer
component formed by single-action pressing, this gradient of density depends on the order of
die filling, i.e. whether the WC base layer is below or above the Fe base layer. If we assume
they reach full density after sintering, the two layers must have the same porosity or relative
density in the green compact to have an identical shrinkage.
Numerical simulation of the compaction of a two layers cylinder was used to optimize
pressing conditions and search for the best distribution of relative density in each layer.
Figure 5 shows the relative density distribution calculated for two orders of the layers in a
component with a mean density of 70%. The density when the WC base material is above
the Fe base one is more homogeneous and is expected to give better results after sintering.
This simulation was confirmed experimentally regarding the differential radial shrinkage
between the two layers with
Fe WC
Fe WC
r r
r r
2 ; r
wc
radial shrinkage of WC base and r
Fe
radial
shrinkage of Fe base. When the Fe base material is on the top =0.14 and =0.09 when it is
the WC base one.
The compaction step for a two layer component impacts the density repartition but also
influences the interface geometry. It should be noticed that the uniaxial compaction from the
top of the sample induces a downward positive curvature of the interface between the two
materials (figure 5), which is also observed on real samples.
After sintering, the sample showed cracks initiated at the sample surface at the two layers
interface, and propagated in the WC base layer (figure 6). The crack seems to have
propagated inwards from the sample outer surface at (or close to) the triple line with the
interface. The first question is when this crack appeared: after compaction or during
sintering?
3.2.2.2 Origin of crack: Compaction/sintering
To search for possible cracks inside two layer green components, compacts were axially cut
and observed. No crack could be observed but the granular nature of green components
could hide such cracks. Tensile tests were done to observe where the two layers samples
would break. During the tensile test, fracture always appeared in the material with the lowest
density, whatever its chemical nature, and never at the interface (figure 7). We can suppose
that if there was a crack initiation at the interface the break would be at this interface. This
argument reinforces the hypothesis that cracks are not present in the green samples, but
appear during sintering (or cooling). However compaction parameters have a strong impact
WC
Fe WC
Fe
Relative density
%

Figure 5: Distribution of relative density in a two
layer component compacted at 570MPa
Figure 6: Fracture inside a sintered two
layer material observed on an axial
cross section
on the amplitude of this crack, because two related parameters influence fracture during
sintering: the order of the layers (Fe base under or above WC base) and the interface
curvature.
3.2.2.3 Influence of layers order: interface geometry and density repartition
The order of the powder during the uniaxial die compaction impacts the density repartition as
well as the interface curvature. Experimental results show that WC base on top position
leads to fracture despite a more uniform relative density (figure 8). Using a curved mobile
punch, samples with the same density repartition and with a planar interface (figure 8c) were
produced. In this case, no crack was observed after sintering, demonstrating the primordial
role of interface curvature.
WC
Fe

Figure 7: Tensile fracture test on green two layer components. In this experiment, the lowest
density material was the Fe base one.
a)
b)
c)

WC on the top:
sample fracture

WC in the bottom:
No fracture (no or
limited crack)

Fe
WC

WC on the top with
planar interface
geometry:
No fracture
Figure 8: Influence of layer order and interface curvature on fracture after sintering
-2E-02
-1E-02
-5E-03
0E+00
5E-03
1E-02
2E-02
2E-02
0 5000 10000 15000 20000 25000
Time (s)
A
x
i
a
l

s
t
r
e
s
s

(
M
P
a
)
0
200
400
600
800
1000
1200
1400
T
e
m
p
e
r
a
t
u
r
e

(
C
)
a case
b case
c case

WC

Fe

WC
Fe

Fe
WC

Figure 9: Axial stress at 500 m of the interface in WC base material in different cases:
a) with downwards curvature when WC base is above; b) with downwards curvature when
WC base is below; c) with upwards curvature when WC is above - The black dots on the
schemes indicate the point positions

WC
Fe

WC

Fe
In order to understand the origin of the fracture, a numerical simulation of the three different
geometries was performed. Stresses developed near the interface of the components were
observed (figure 9). The numerical simulation shows that sintering generates tensile stresses
in WC base near the interface when WC base is in the concave side of interface, whatever
the top or bottom position of the layer. On the contrary, compression stresses are generated
when WC base is in the convex part. The Fe powder is in opposite situations. When the
interface is flat, only small stresses, alternatively tensile and compressive, are generated. If,
in addition, we assume that the WC base material has a lower tensile fracture limit than the
more ductile Fe base one, these stresses may explain the fracture behaviours observed in
the different curvature situations.
4 Conclusion
Cemented carbides and steel are compatible materials for co-sintering with a potential of
application in hardness/toughness couples. Successful achievement of these materials
implies to control all steps, from compaction to sintering. Knowledge of the behaviour of
single material is essential but not sufficient. The sintering of bi-materials also results from
the initial density in each layer, and from specific parameters of the interface between them,
such as curvature. In this paper, the association of experiments and modelling on one hand,
and, of compaction and sintering on the other hand, were used. This association has proved
to be fruitful to understand the observed phenomena and to define guidelines for optimal
powder metallurgy process of the studied two layer material.
5 Acknowledgements
The authors wish to thank Eurotungstene Poudres for raw materials. They acknowledge
Direction Gnrale des Entreprises, Conseil Gnral de la Loire and Saint Etienne Mtropole
for their financial support of the MULTIMAT Project, within the Competitiveness Cluster
ViaMca and CETIM Foundation for its financial support of the Multimaterial and
Multifunctional Products Project.
References
[1] German, R.M., Heaney, D.F. and Johnson, J.L. Bi-material components using powder
injection molding: densification, shape complexity, and performance attributes, Advances in
Powder Metallurgy and Particulate Materials, 2005, 41-52.
[2] Green, D.J., Guillon, O. and Rdel, J. Constrained sintering: A delicate balance of
scales, Journal of the European Ceramic Society, 2008, 28 (7), 1451-1466.
[3] Pascal, C., Chaix, J.M., Dor, F. and Allibert, C.H. Design of multimaterial processed
by powder metallurgy: Processing of a (steel/cemented carbides) bilayer material, Journal of
Materials Processing Technology, 2009, 209, 1254-1261.
[4] Gabriel, A., Pastor, H., Deo, D.M., Basu, S. and Allibert, C.H. New experimental data
in the C-Co-W, C-Fe-W, C-Ni-W systems and their applications to sintering conditions,
Proceedings of the 11th International Plansee Seminar, Reutte, AUSTRIA, 2, 509-525.
[5] ABAQUS "Analysis User's Manual" (ABAQUS Inc., 2007).
[6] Drucker, D.C. and Prager, W. Solid mechanics and plastic analysis for limit design,
Quarterly of Applied Mathematics, 1952, 10 (2), 157-165.
[7] Doremus, P., Toussaint, F. and Alvain, O. Simple tests and standard procedure for the
characterisation of green compacted powder, In Zavaliangos, A. and Laptev, A., eds. Recent
Developments in Computer Modelling of Powder Metallurgy Processes, 29-41 (IOS Press,
2001).
[8] Thomazic, A., Pascal, C. and Chaix, J.M. Optimization of density, microstructure and
interface region in a co-sintered (steel/cemented) carbide bi-layered material, Proceedings of
Sintering 2008, La Jolla, USA, November 16-20, 2008, in press.
[9] Gillia, O. and Bouvard, D. Phenomenological analysis of densification kinetics during
sintering: application to WC-Co mixture, Materials Science and Engineering a-Structural
Materials Properties Microstructure and Processing, 2000, 279 (1-2), 185-191.
[10] Thomazic, A., Pascal, C. and Chaix, J.M. Fabrication of (cemented carbides/steel)
bilayered materials by Powder Metallurgy, Proceedings of FGM'08, Sendai, Japan,
September 22-25, 2008, in press.
Germany
Influence of Compaction on Sintering of a Two Layer Component

A. Thomazic
a
, Y. Le Guennec
b
, Y. Kamdem
a
, C. Pascal
a
, J.M. Chaix
a
, P. Doremus
a
,
D. Imbault
b
, D. Bouvard
a
, F. Dor
c

a
SIMaP, CNRS/Grenoble INP/UJF, BP 75, 38402 Saint-Martin-d'Hres, France
b
3S-R, CNRS/Grenoble INP/UJF, BP 53, 38041 Saint-Martin-d'Hres, France
c
CETIM, BP802, 42952 Saint-Etienne Cedex 9, France
Abstract
The association of a hard and a tough material in one component can be achieved by
powder metallurgy, but many difficulties must be overcome for compaction and sintering.
This paper deals with a two layer component, a WC-Fe cemented carbide and a Fe-W-C
steel, both elaborated from tungsten carbide, iron and graphite powders. The two layers are
co-compacted then co-sintered.
In this paper, the experimental and numerical approaches of compaction and sintering are
associated to understand the involved phenomena and improve the process.
1 Introduction
Two layer components allow the combination of properties (mechanical, thermal, electrical...)
that could not be achieved with a single material. Multi-layer components are conventionally
obtained by processes which imply intermediate steps of alloy manufacturing, assembling
and machining. The powder metallurgy process is an appropriate route to obtain directly this
product with the required dimensions and shapes. Post-sintering machining is reduced or
even useless.
In order to shape multi-layer components, the compaction step is critical to obtain the
expected geometry and a good initial density repartition and to minimize mechanical strains
during sintering. Moreover, compositional difference between the two materials could induce
chemical phenomena at the interface between layers [1, 2]. This study aims to understand
the influence of compaction and thermal treatment parameters on sintering of a two layer
component, which associates these two materials in order to combine hardness and
toughness properties.
In a previous investigation [3], the composition of each layer was selected from physico-
chemical data and sintering compatibility. In order to understand the influence of compaction,
tests have been performed on single layer components, then on two layer components.
Different thermal cycles have been tested. Experimental results were compared to numerical
simulations to describe the process and highlight the differences between experimental and
numerical results.
2 Experimental procedure
In this study, to ensure chemical compatibility and to limit the number of elements, the hard
material (WC base material) consists of tungsten carbide hard particles cemented by an iron
base tough binder whereas cobalt is employed in classic cemented carbides [4]. The second
material (Fe base material) is mainly composed of iron, with a small amount of tungsten
carbide and graphite powders.
The materials were prepared from WC, Fe (Eurotungstene Poudres) and C
graphite
(Hgans)
powders (table 1). Similar average WC and Fe particle sizes are used for the WC base
powder mixture while for the Fe base one, finer WC particles are chosen to improve carbide
distribution in iron rich matrix and to reduce residual porosity after sintering [3]. The
compositions of powder mixtures are reported in table 1.
The mixtures were processed through attritor-milling with 2wt.% polyethylene glycol (PEG
3400) dissolved in acetone media. Organic binder is needed to promote compaction and to
improve green strength in order to enable eventual green machining. Powder mixtures were
Material Fe base layer WC base layer
Composition [wt.%] Fe-7 %WC-0.6 %C WC-6.6 %Fe-0.14 %C
Powder designation Fe2010 CW5200 C-KS44 Fe2010 CW5722 C-KS44
Typical laser D50 (m) 6.8 1.6 18.6 6.8 6.1 18.6
Table 1: Compositions and characteristics of the powders
dried and granulated through a 500 m sieve.
To compare the results, we used relative density evaluated from the ratio between
experimental density and theoretical final density. Experimental densities were calculated
from mass and volume. Theoretical densities were estimated from initial constitutive material
densities (14.6 for WC base material and 8.0 for Fe base material).
3 Results and discussion
3.1 Single layer components
3.1.1 Compaction
In order to predict the density distribution in a two layer materials, a numerical model was
fitted with experimental data obtained on single materials. Simulations have been run with
the finite element software ABAQUS [5] which included a Drucker-Prager constitutive
expression [6]. The parameters of this constitutive law for each material were fitted using
three different tests: diametral compression, uniaxial compression and an instrumented die
test [7].
A hydraulic press which can apply a force up to 10 tons was used. Radial stress applied to
the die by the powder was measured by a strain gauge located on the outer surface of the
die. Top and bottom punches were also equipped with strain gauges to evaluate axial
stresses during compaction. The punches diameter was 12.5 mm and their radial clearance
with the die was 0.2500.005 mm. The friction coefficient between powder and die, which
has a strong influence on the density distribution of the compact, was also determined on
both powders. We obtained a friction coefficient of 0.17 for the Fe base material and 0.22 for
the WC base one. To check the fitting of the Drucker-Prager model and friction coefficients,
numerical results are compared to the experiments after uniaxial die compaction (figure 1).
As tests as well as sample preparation were realized using a single-action die (mobile upper
punch), a device independent half height stress was used. Considering a linear evolution of
the density along the compact height, the mean density of the compact is related to the half-
height stress:
low up 2 / h
= [7]. For the two materials the experimental evolution of the
final density versus the half-height stress is correctly reproduced by the model.

3
4
5
6
0 200 400 600
D
e
n
s
i
t
y

(
g
/
c
m
3
)
Simulation
Experiment
6
7
8
9
10
11
0 200 400 600
D
e
n
s
i
t
y

(
g
/
c
m
3
)
Simulation
Experiment

Figure 1: Experimental evolution of the final density versus the half-height stress for
a) Fe base material and b) WC base material
3.1.2 Sintering
The powder mixtures were uniaxially compacted into cylindrical samples (=8 mm, H=6-
8 mm) under P=600 MPa. After compaction, the average values of green densities are 10.0
(69 %) for WC base compacts and 6.0 (74 %) for Fe base compacts. The thermal treatment
is carried out in a dilatometer to measure axial length changes during heating ( L/L
0
). Before
the sintering step, the compacts were held for one hour under He-4%H
2
flow at 360C to
remove organic binder. After the debinding stage, Fe base and WC base powder compacts
are heated in argon atmosphere at 5C/min and isothermally held at 1300C during 60
minutes (figure 2).
The experimental curves were discussed in a previous work [8]. The comparison of sintering
behaviours of Fe and WC base single materials indicates that dimensional changes occur at
different temperatures with different magnitudes. These differences could originate stresses
a) Fe base b) WC base
and cracks at the interface of the two material components during sintering process. After
cooling to room temperature, the experimental densities of the Fe base and WC base
materials are 7.7 (96%) and 13.8 (94%) respectively.
The numerical simulation is a helpful tool to anticipate the behaviour of bi-materials during
sintering. The constitutive laws of the two materials were defined with the help of dilatometry
experiments based on Gillias procedure [9]. Figure 2 shows that the numerical simulations
predict reasonably well the experimental data. The deviation observed at the beginning of
each thermal treatment (below 1000C) is due to phenomena such as debinding step or
phase transformation. These phenomena are not taken into consideration in this model.

a)

b)
Figure 2: Experimental dilatometric curves compared to numerical simulations for the
sintering of a) Fe base and b) WC base single materials (1300C-60 min)
3.2 Two layer components
3.2.1 Influence of sintering
Several sintering parameters are expected to monitor the properties of the sintered two layer
material: graphite content in Fe base layer, co-compaction pressure, heating rate, sintering
temperature and duration. Design of experiments (DOE) approach was used to analyse the
effect of these parameters on sintering with a limited number of experiments [10]. This study
highlights the formation of liquid in the Fe base layer at 1300C. Part of this liquid migrates
towards the WC base layer by capillary infiltration. This phenomenon has consequences on
both materials: geometrical deformation, lower densification, new brittle phase formation
(M
6
C) at the interface between layers...
For the two layer compacts, WC base mixture was poured first and slightly compacted in the
three part matrix (=8 mm). Then Fe base mixture was added and pressed at 400 MPa to
achieve a compact with 4 mm thick layers (figure 3). On figure 4, experimental result on the
1300C-60 min cycle is compared to numerical simulation. Model is in agreement with
experimental results until 1300C. During the isothermal stage at 1300C, the shrinkage rate
decreases in the first 10 min as the numerical simulation, then, shrinkage starts again. After
cooling to room temperature, the experimental densities of the Fe base and WC base layers

Figure 3: Schematic representation
of a two layer component
Figure 4: Experimental dilatometric curves compared to
numerical simulations of the bi-material sintered at 1300C
are 7.6 (95%) and 13.3 (91%) respectively. The axial shrinkage is twice higher (22%) than
the shrinkage reached in the numerical simulation (10%). This deviation is due to the liquid
formation and then liquid migration which cannot be predicted by numerical simulation. The
analysis of DOE combined to microstructural observations and numerical simulations
indicates that a sintering at 1290C during 30 min is a suitable cycle. This cycle was applied
for subsequent experiments.
3.2.2 Influence of compaction
3.2.2.1 Compaction of two layer components
Compaction determines the density distribution inside the green material. For a two layer
component formed by single-action pressing, this gradient of density depends on the order of
die filling, i.e. whether the WC base layer is below or above the Fe base layer. If we assume
they reach full density after sintering, the two layers must have the same porosity or relative
density in the green compact to have an identical shrinkage.
Numerical simulation of the compaction of a two layers cylinder was used to optimize
pressing conditions and search for the best distribution of relative density in each layer.
Figure 5 shows the relative density distribution calculated for two orders of the layers in a
component with a mean density of 70%. The density when the WC base material is above
the Fe base one is more homogeneous and is expected to give better results after sintering.
This simulation was confirmed experimentally regarding the differential radial shrinkage
between the two layers with
Fe WC
Fe WC
r r
r r
2 ; r
wc
radial shrinkage of WC base and r
Fe
radial
shrinkage of Fe base. When the Fe base material is on the top =0.14 and =0.09 when it is
the WC base one.
The compaction step for a two layer component impacts the density repartition but also
influences the interface geometry. It should be noticed that the uniaxial compaction from the
top of the sample induces a downward positive curvature of the interface between the two
materials (figure 5), which is also observed on real samples.
After sintering, the sample showed cracks initiated at the sample surface at the two layers
interface, and propagated in the WC base layer (figure 6). The crack seems to have
propagated inwards from the sample outer surface at (or close to) the triple line with the
interface. The first question is when this crack appeared: after compaction or during
sintering?
3.2.2.2 Origin of crack: Compaction/sintering
To search for possible cracks inside two layer green components, compacts were axially cut
and observed. No crack could be observed but the granular nature of green components
could hide such cracks. Tensile tests were done to observe where the two layers samples
would break. During the tensile test, fracture always appeared in the material with the lowest
density, whatever its chemical nature, and never at the interface (figure 7). We can suppose
that if there was a crack initiation at the interface the break would be at this interface. This
argument reinforces the hypothesis that cracks are not present in the green samples, but
appear during sintering (or cooling). However compaction parameters have a strong impact
on the amplitude of this crack, because two related parameters influence fracture during
WC
Fe WC
Fe
Relative density
%

Figure 5: Distribution of relative density in a two
layer component compacted at 570MPa
Figure 6: Fracture inside a sintered two
layer material observed on an axial
cross section
sintering: the order of the layers (Fe base under or above WC base) and the interface
curvature.

3.2.2.3 Influence of layers order: interface geometry and density repartition
The order of the powder during the uniaxial die compaction impacts the density repartition as
well as the interface curvature. Experimental results show that WC base on top position
leads to fracture despite a more uniform relative density (figure 8). Using a curved mobile
punch, samples with the same density repartition and with a planar interface (figure 8c) were
produced. In this case, no crack was observed after sintering, demonstrating the primordial
role of interface curvature.
WC
Fe

Figure 7: Tensile fracture test on green two layer components. In this experiment, the lowest
density material was the Fe base one.
a)
b)
c)

WC on the top:
sample fracture

WC in the bottom:
No fracture (no or
limited crack)

Fe
WC

WC on the top with
planar interface
geometry:
No fracture
Figure 8: Influence of layer order and interface curvature on fracture after sintering
-2E-02
-1E-02
-5E-03
0E+00
5E-03
1E-02
2E-02
2E-02
0 5000 10000 15000 20000 25000
Time (s)
A
x
i
a
l

s
t
r
e
s
s

(
M
P
a
)
0
200
400
600
800
1000
1200
1400
T
e
m
p
e
r
a
t
u
r
e

(
C
)
a case
b case
c case

WC

Fe

WC
Fe

Fe
WC

Figure 9: Axial stress at 500 m of the interface in WC base material in different cases:
a) with downwards curvature when WC base is above; b) with downwards curvature when
WC base is below; c) with upwards curvature when WC is above - The black dots on the
schemes indicate the point positions

WC
Fe

WC

Fe
In order to understand the origin of the fracture, a numerical simulation of the three different
geometries was performed. Stresses developed near the interface of the components were
observed (figure 9). The numerical simulation shows that sintering generates tensile stresses
in WC base near the interface when WC base is in the concave side of interface, whatever
the top or bottom position of the layer. On the contrary, compression stresses are generated
when WC base is in the convex part. The Fe powder is in opposite situations. When the
interface is flat, only small stresses, alternatively tensile and compressive, are generated. If,
in addition, we assume that the WC base material has a lower tensile fracture limit than the
more ductile Fe base one, these stresses may explain the fracture behaviours observed in
the different curvature situations.
4 Conclusion
Cemented carbides and steel are compatible materials for co-sintering with a potential of
application in hardness/toughness couples. Successful achievement of these materials
implies to control all steps, from compaction to sintering. Knowledge of the behaviour of
single material is essential but not sufficient. The sintering of bi-materials also results from
the initial density in each layer, and from specific parameters of the interface between them,
such as curvature. In this paper, the association of experiments and modelling on one hand,
and, of compaction and sintering on the other hand, were used. This association has proved
to be fruitful to understand the observed phenomena and to define guidelines for optimal
powder metallurgy process of the studied two layer material.
5 Acknowledgements
The authors wish to thank Eurotungstene Poudres for raw materials. They acknowledge
Direction Gnrale des Entreprises, Conseil Gnral de la Loire and Saint Etienne Mtropole
for their financial support of the MULTIMAT Project, within the Competitiveness Cluster
ViaMca and CETIM Foundation for its financial support of the Multimaterial and
Multifunctional Products Project.
References
[1] German, R.M., Heaney, D.F. and Johnson, J.L. Bi-material components using powder
injection molding: densification, shape complexity, and performance attributes, Advances in
Powder Metallurgy and Particulate Materials, 2005, 41-52.
[2] Green, D.J., Guillon, O. and Rdel, J. Constrained sintering: A delicate balance of
scales, Journal of the European Ceramic Society, 2008, 28 (7), 1451-1466.
[3] Pascal, C., Chaix, J.M., Dor, F. and Allibert, C.H. Design of multimaterial processed
by powder metallurgy: Processing of a (steel/cemented carbides) bilayer material, Journal of
Materials Processing Technology, 2009, 209, 1254-1261.
[4] Gabriel, A., Pastor, H., Deo, D.M., Basu, S. and Allibert, C.H. New experimental data
in the C-Co-W, C-Fe-W, C-Ni-W systems and their applications to sintering conditions,
Proceedings of the 11th International Plansee Seminar, Reutte, AUSTRIA, 2, 509-525.
[5] ABAQUS "Analysis User's Manual" (ABAQUS Inc., 2007).
[6] Drucker, D.C. and Prager, W. Solid mechanics and plastic analysis for limit design,
Quarterly of Applied Mathematics, 1952, 10 (2), 157-165.
[7] Doremus, P., Toussaint, F. and Alvain, O. Simple tests and standard procedure for the
characterisation of green compacted powder, In Zavaliangos, A. and Laptev, A., eds. Recent
Developments in Computer Modelling of Powder Metallurgy Processes, 29-41 (IOS Press,
2001).
[8] Thomazic, A., Pascal, C. and Chaix, J.M. Optimization of density, microstructure and
interface region in a co-sintered (steel/cemented) carbide bi-layered material, Proceedings of
Sintering 2008, La Jolla, USA, November 16-20, 2008, in press.
[9] Gillia, O. and Bouvard, D. Phenomenological analysis of densification kinetics during
sintering: application to WC-Co mixture, Materials Science and Engineering a-Structural
Materials Properties Microstructure and Processing, 2000, 279 (1-2), 185-191.
[10] Thomazic, A., Pascal, C. and Chaix, J.M. Fabrication of (cemented carbides/steel)
bilayered materials by Powder Metallurgy, Proceedings of FGM'08, Sendai, Japan,
September 22-25, 2008, in press.
Manuscript refereed by Professor Didier Bouvard, PHELMA, France

Liquid Phase Sintering of High-Alloyed Plastic Mould Steels:
Calculations and Experimental Findings on Atmospheric Influences

H. Hill
1
, S. Weber
1;2
, S. Siebert
1
, W. Theisen
1

1
Ruhr-University Bochum
Institute for Materials
Chair of Materials Technology
D-44780 Bochum, Germany
Tel. 0049-234-3225964
Fax 0049-234-3214104
2
Helmholtz-Zentrum Berlin fr Materialien und Energie GmbH
D-14109 Berlin, Germany

1 Abstract

Many tool devices are exposed to a combined load of corrosion and wear. Plastic mould
steels exhibit a good corrosion resistance provided by a sufficient content of chromium and a
good wear resistance due to carbides dispersed in a martensitic matrix. A further gain in
wear resistance is achieved by the addition of hard particles to the steel for obtaining a
particulate reinforced metal matrix composite (MMC). The common processing route for
these MMC is hot isostatic pressing, but they can also be consolidated by solid state or liquid
phase sintering.
In this work, detailed investigations on the sintering behaviour and resulting microstructures
of two plastic mould steels were performed. The experiments were supported by
thermodynamic calculations to explain the influence of the sintering atmosphere. The results
show an improved corrosion and wear resistance compared to HIP materials, being related
to microstructural findings.

2 Introduction

Besides the effect of wear in technical applications, e.g. in plastic processing there is also a
huge material damage due to corrosion. For this reason and especially for this kind of
applications, corrosion and wear resistant cold work tool steels were developed, the so called
plastic mould steels. To withstand the first one, a sufficient amount of chromium in the metal
matrix is needed and for the latter a martensitic matrix with embedded carbides if they are
harder than the affecting abrasive particles and of similar size [1].
Normally the production process for this kind of alloys is the hot isostatic pressing (HIP) for
consolidation of the PM-steel. But in recent years several works were published about the
liquid phase sintering of pre-alloyed tool steels using the so called super solidus liquid phase
sintering (SLPS) [2, 3, 4]. Compared to HIP, the SLPS process offers advantages due to
lower costs because of a faster consolidation time and nonessential capsulation of the
powder. Furthermore, a larger flexibility concerning the geometry is possible.
SLPS takes place in the fusion interval between solidus and liquidus at a temperature which
is needed to obtain a liquid phase fraction of 20-30% [5]. The first step in the densification
process is the formation of liquid phase along the grain boundaries leading to a viscous flow
[6]. The wetting of the interfaces and a separation to smaller particles occurs with an
increasing amount of liquid phase. The last step of densification is related to capillary forces
which ensure a rearrangement of the powder particles.
Besides the sintering in vacuum there is also the possibility of using nitrogen as a sintering
atmosphere [3, 7, 8, 9]. This affects the sintering behaviour and consequently the micro-
structure. Hence, the scope of this work is to investigate the difference between the two
atmospheres. Furthermore, the experimental findings on atmospheric influences were
supported by thermodynamic calculations to achieve a better understanding.
3 Experimental

A gas-atomized tool steel powder, namely X190CrVMo20-4 (X190), was used for the
investigations on the influences of the sintering atmosphere. It belongs to the group of plastic
mould steels. The chemical composition, which was used for the thermodynamic
calculations, was measured by spark spectroscopy of samples sintered in vacuum and is
given in Table 1. The sintering experiments with the powder were carried out without
fractionation.

Using Thermo-Calc version R and the TCFe4 database (Thermo-Calc AB, Stockholm,
Sweden) the element levels for this steel were calculated in dependence of the sinter
atmosphere. In case of nitrogen sintering the consideration of the gas pressure took place.
The calculations were performed with a reduced phase-set of LIQUID, BCC-A2, FCC-A1,
CEMENTITE, M6C, M7C3, M23C6 and HCP-A3.

For the sintering, a tubular furnace (Haereus Company) was used with a maximum
temperature of 1300C. It is possible to operate the furnace at minimum pressure of 5*10
-1

Pa or with pure nitrogen by means of a subsequent backfilling up to 3*10
5
Pa. Therefore it is
attainable to sinter either in vacuum or in nitrogen using the same experimental setup.
Sintering experiments were performed in small alumina crucibles with a pre-compaction of
the powders only to tap density.
For the final heat treatment, samples were protected from oxidation by heat treatment foil,
austenitized in a muffle type furnace at 1070C in argon atmosphere for 30 min and then
quenched in oil. The hardening was immediately followed by a double tempering for two
hours performed at a temperature of 450C with subsequent air cooling.

For the microstructural investigations using optical and scanning electron microscopy (OM
and SEM) small samples were cut of the sintered parts and embedded in a conductive
polymer resin. Subsequently they were ground on SiC paper and polished with diamond
suspension. Etching was performed using theV2A" etchant [10].

For the investigation of the corrosion behaviour current potential curves were measured. The
configuration of this test consists of a working electrode, a platinum counter electrode and a
reference electrode with a potential shift of +244 mV in comparison to the standard hydrogen
electrode.
The working electrode represents the sample for the corrosion test and it is connected with a
spot welded wire to the potentiostat. In the next step embedding in a resin and grinding on
SiC paper followed. This was followed by sealing of the gap between the sample surface, the
resin and the rest of the resin body with a lacquer to avoid an undesirable electrical contact
and crevice corrosion. As electrolyte an 0.5 molar sulphuric acid was used with a potential
change of 600 mV/h. Due to the fact that the measured current will relate to the surface, this
area was determined previously with the help of an image analysis software.

4 Results and Discussion

As mentioned before, during sintering in nitrogen atmospheres an uptake of this element
must be considered whereby a change in the chemical composition occurs. Therefore the
equilibrium state for a typical hardening temperature of 1070C was calculated for two
different gas pressures of nitrogen, namely 0.01 MPa and 0.02 MPa (Tab. 2).
The phase equilibria at this temperature for the vacuum and the nitrogen sintered samples
are given. The overall nitrogen content was set according to the amount soluble at a sintering
temperature of 1280C. It is obvious that a significant change in the vanadium, nitrogen and
carbon values occurs. To validate these findings, calculations of the mono carbide volume
fraction over the nitrogen content were performed. Figure 1 shows that the amount of these
carbides increases with nitrogen content, what is in accordance to the calculations at the
hardening temperature.
With a closer look it is evident that the vanadium rich mono carbides are more and more a
carbonitride with increasing nitrogen content. The carbon is crowded out from the MC which
is virtually transformed from a carbide to a nitride. This is in agreement with [8] where the
authors describe the process with the following equation:
MC + N = M(C;N) + C
Furthermore, alloying with nitrogen reduces the vanadium content in the M
7
C
3
carbide and
increases it in the M(C,N) carbonitride.

The sintering tests were performed at the optimal sintering temperature of 1280C for the
X190 in vacuum atmosphere. To receive appropriate results for the nitrogen sintered state,
the sintering tests were also performed at 1280C with two different amounts of nitrogen. In
both case a dense microstructure is achieved.
Figure 2 a) shows the microstructure of X190 after sintering in vacuum in the QT condition. A
tempered martensite microstructure is visible with dispersed carbides which can be identified
as chromium rich carbides of the type M
7
C
3
. With the help of the backscattered electron
mode at the SEM there is also a small amount of vanadium rich MC visible (see Fig. 2 b)
what is in agreement with the Thermo-Calc calculations.
In Figure 3 a) and b) the microstructures in the nitrogen sintered conditions at 0.01 and 0.02
MPa gas pressure are depicted. It is evident that there is an increase of the amount of
vanadium rich carbonitrides and this confirms the calculations performed previously.

In Figure 4 the results of the current potential measurements are shown for the three vacuum
and nitrogen sintered conditions. The open circuit potential was detected at a level of
483 mV. Based on this potential for the vacuum sintered sample a strong increase in the
current density is evident up to a value of 2800 A/cm
2
. Afterwards there is a drop of the
current density with increasing potential to approximately -165 mV were the current
increases a second time. With further progress of the potential the passive area is reached
and therefore the current achieve a minimum due to the formation of the passive layer [11].
Below this potential the transpassive area is reached and the current increases because of
the breakdown of the passive layer.
Based on the equilibrium rest potential the measured curves for the nitrogen sintered
samples show the same characteristics but each with a higher current level. For a better
understanding of the corrosion mechanisms, microstructural investigations were performed
after 30 minutes in 0.5 molar sulphuric acid. The result is depicted in Figure 5. It is clear that
especially the phase boundary between the carbides and the matrix is attacked which is in
accordance to findings in the literature [12]. This occurs due to a potential difference between
the more noble carbides and the surrounding metallic matrix. Furthermore, the corrosive
attack is intensified when this area is depleted in chromium in case of chromium rich
carbides [13, 14, 15]. With an increasing amount of boundary surface because of more
carbides this effect is boosted. Therefore the results of the comparison between the nitrogen
sintered sampled is plausible.

5 Summary

It is possible to consolidate pre-alloyed powders of plastic mould steels with SLPS in vacuum
atmosphere or nitrogen. This leads to a change of the existing MC to M(C,N) and to a further
precipitation of vanadium rich mono carbonitrides. The corrosion resistance is deteriorating
due to the higher amount of carbides and the fact that there is a potential difference
regarding the less noble metal matrix.
Furthermore, the investigation has shown that nitrogen uptake can be calculated by means
of Thermo-Calc being in accordance with microstructural findings.

6 Tables

Table 1: Content of the main alloying elements of the investigated plastic mould steel; values in wt.%
C N Si Mn Cr Nb Mo V W
X190CrVMo20-4 2.079 0.09 0.597 0.31 20.09 0.1 0.95 3.97 0.50

Table 2: Composition of equilibrium phases in X190 at T=1070C corresponding to vacuum and
nitrogen sintering according to Thermo-Calc; values in wt.%
C N Si Mn Cr Nb Mo V W
X190 vacuum
Austenite 0.33 0.006 0.62 0.35 12.25 0.001 0.85 0.76 0.41
M(C,N) 11.97 4.28 - - 8.51 5.00 2.91 65.81 0.52
M
7
C
3
8.79 - - 0.16 59.11 - 1.20 10.18 0.91
X190 nitrogen 0.01 MPa
Global 2.08 0.90 0.60 0.31 20.09 0.1 0.95 3.97 0.50
Austenite 0.49 0.10 0.65 0.35 11.63 0.001 0.85 0.11 0.37
M(C,N) 1.53 18.68 - - 1.65 2.27 0.16 74.27 0.02
M
7
C
3
8.72 - - 0.23 61.49 - 1.53 2.04 1.11
X190 nitrogen 0.02 MPa
Gloabal 2.08 0.99 0.60 0.31 20.09 0.1 0.95 3.97 0.50
Austenite 0.51 0.14 0.65 0.34 11.72 0.001 0.85 0.08 0.02
M(C,N) 1.21 19.17 - - 1.92 2.18 0.18 73.72 0.02
M
7
C
3
8.71 - - 0.24 61.30 - 1.55 1.55 1.12

7 Figures

Fig. 1: Calculated carbide content in dependence of nitrogen derived with Thermo-Calc

Fig. 2: Cross sections of X190 after sintering in vacuum

Fig. 3: Cross sections of X190 after sintering in nitrogen at different gas pressures
a = 0.01MPa
b = 0.02MPa

Fig. 4: Current potential curves of X190 after vacuum and nitrogen sintering in the quenched and
tempered (QT) condition

Fig. 5: SEM micrograph of X190 sintered in nitrogen (0.01 MPa) after 30 minutes in 0.5 molar
sulphuric acid

8 References

[1] H. Berns, V. G. Gavriljuk, B. Shanina, Intensive interstitial strengthening of stainless
steels, Advanced Engineering Materials 10 (12) (2008) 1083-1093.
[2] W. Theisen, Herstellung verschleibestndiger Metallmatrix-Verbunde auf Fe-Basis,
Materialwissenschaft und Werkstofftechnik 36 (2005) 360-364.
[3] S. Weber, W. Theisen, Optimised sintering route for cold work tool steels, Powder
Metallurgy 47 (4) (2006) 355-362.
[4] S. Weber, W. Theisen, Grundlagen, Herstellung und Anwendungsaspekte des
Supersolidus Flssigphasensinterns von hochlegierten Werkzeugsthlen und Metallmatrix-
Verbundwerkstoffen, Materialwissenschaft und Werkstofftechnik 38 (6) (2007) 442-452.
[5] R. M. German, Supersolidus liquid-phase sintering of prealloyed powders, Metallurgical
And Materials Transactions A-Physical Metallurgy And Materials Science 28 (7) (1997) 1553-
1567.
[6] R. M. German, Liquid Phase Sintering, Plenum Press, 1985.
[7] S. Weber, Gezielte Ausnutzung des Stofftransportes zur Herstellung neuartiger Pm-
Hartverbundwerkstoffe auf Eisenbasis, Ph.D. thesis, Ruhr-Universitt Bochum, Institut fr
Werkstoffe (2005).
[8] I. Aguirre, S. Gimenez, S. Talacchia, T. Gomez-Acebo, I. Iturriza, Effect of nitrogen on
supersolidus sintering of modified m35m high speed steel, Powder Metallurgy 42 (4) (1999)
353-357.
[9] S. Gimenez, I. Iturriza, Study of the sintering behaviour of pm hss under nitrogen
atmospheres: Application to alloy design, European powder metallurgy association,
20.-22.10., Valencia (2003) 251-258.
[10] G. Petzow, Metallographisches, Keramographisches, Plastographisches tzen, 6.
berarbeitete Auflage, Stuttgart, 1994.
[11] H. Kaesche, Corrosion of Metals. Physicochemical Principles and Current Problems,
Springer Verlag, 2003.
[12] L. C. Herzbach, H. Schmidt, M. Windler, Gefge und Korrosionsverhalten eines
stickstofflegierten rostfreien Stahls, Materialwissenschaft und Werkstofftechnik (29) (1998)
329-355.
[13] H. Berns, G. Wang, Verschleibestndige PM-Hartlegierungen mit verbessertem
Korrosionswiderstand, in: Tagungsband Surtec91, Berlin, 1991, 637-644.
[14] S. Peissl, G. Mori, H. Leitner, R. Ebner, S. Eglsaer, Influence of chromium, molybdenum
and cobalt on the corrosion behaviour of high carbon steels in dependence of heat treatment,
Materials and Corrsion, 57 (10) (2006) 759-765.
[15] A. Kajinic, A. L. Wojcieszynski, M. K. Sawford, Microstructure and mechanical
properties of a new cpm corrosion and wear resistant tool steel: 7th international
tooling conference - Tooling materials and their applications from research to market -
Volume I (2006) 283-290.
Manuscript refereed by Professor Didier Bouvard, PHELMA, SIMAP Laboratory,
France

Microstructural Evolution in Solar Sintered M2 HSS

G. Herranz, R. Alonso, J. Arvalo, G. Rodrguez.

Universidad de Castilla-La Mancha. Metallic Materials Group.
ETSII-UCLM. Avda. Camilo Jos Cela s/n. E-13071 Ciudad Real, SPAIN
Keywords: solar sintering, high speed steel, microstructure.

Abstract: We have checked the feasibility of concentrated solar energy for sintering-
consolidation of green parts previously obtained by compaction of irregular shape M2 HSS
powder. The use of solar installations allows materials processing at much faster heating
and/or cooling rates than in conventional furnaces. The experiments were carried out using
two facilities; a parabolic solar furnace of the Solar Plant of Almeria (PSA) and a Fresnel lens
placed in the Castilla-La Mancha University (ETSII-UCLM). The research focused in the
microstructural evolution of the samples treated in both facilities. In this work we compared
the results of the sintering process in N
2
-H
2
atmosphere under concentrated solar energy
with the conventional furnace results. A near full density was achieved when the solar
furnace has been used and the mechanical evaluation through microhardness
measurements (900 HV) indicates that this new solar sintering process has been
successfully developed obtaining better mechanical properties with a very important
reduction in the total time of the cycle.

Introduction

The AISI M2 HSS is a conventional cutting tool material with a desired combination of hot
hardness, wear resistance and toughness [1]. Powder processing of this alloy can overcome
the complicate processing to produce the finished product and offers a cost effective method
for producing near net shape components. The complexity of their composition makes them
extremely sensitive to sintering parameters. The sintering process has been widely studied
by several researches describing a supersolidus liquid phase sintering (SLPS) mechanism
[2]. This process involves heating prealloyed powders to an intermediate temperature
between the liquidus and solidus, resulting in a partial melting and densification through a
particle rearrangement and solution-reprecipitation mechanisms [2]. The sintering
temperatures and the carbon content are the most important process variables since these
determine the volume fraction of liquid phase [3]. Jauregi et al demonstrated that the use of
gas atmosphere of N
2
-H
2
reduces the sintering temperature and improves the microstructure
and the mechanical properties of the tool parts [4]. In all the cases, the HSS sintering takes
place in a narrow temperature range (the sintering window). Also, it is well known that the
best compromise between toughness and hardness is achieved after heat treatments.
The demand for faster and less expensive products processing has resulted in the
development of high energy techniques that allows the fabrication of prototypes and short
series tooling as laser sintering [5] or capacitor discharge sintering (CDS) [6]. The use of this
kind of faster technology usually has achieved finer microstructure, has improved the
mechanical properties and the sintering window. The main disadvantage of this processes
are their low energy efficiency. In the search for new sintering systems that overcome this
issue, the use of concentrated solar energy (CSE) appears to be an interesting technique.
The CSE is clean, renewable and pollutant free. In spite of its apparent limitation for
industrial application, due to the unpredictable availability of solar radiation, the CSE shows a
clear activator effect in different processes [7, 8]. The use of solar furnaces allows materials
processing at much higher heating and/or cooling rates resulting in improved mechanical
properties and a very important reduction in the total time of the cycle. The feasibility of solar
furnace for sintering-consolidation of green parts previously obtained by compaction has
been demonstrated in ceramic systems [9-11] and also a preliminary study of a copper
system has been published [12]. Therefore, it is a challenge to check the feasibility of the
solar sintering of M2 HSS. The research has been developed in two different solar
installations: the Solar Furnace of the PSA (Plataforma Solar de Almera) and the Fresnel
lens in the UCLM (Universidad de Castilla La Mancha). Although the solar influence in the
sintering mechanisms has not been yet fully elucidated, the present research has the
purpose of studying the microstructural evolution and to check the influence of the high
density energy in the mechanical properties, comparing the solar process with the
conventional method.

Experimental Procedure

The metal powder used in the experiments was water atomized M2 High Speed Steel sold by
Hgans. As shown in Figure 1, the particle shape is irregular and 80 % of the particles
were sub 150 m. The chemical composition is
given in Table I.
The samples were cold uniaxially compacted at
650 MPa to fabricate bars (31.6 x 12.8 x 2.4
mm). After that, some of the bars were sintered
in the solar furnace of the PSA, some in the
Fresnel lens in the same conditions of N
2
-5%H
2
flowing gas at temperatures in the range
between about 1115 and 1200 C. For
comparison of the results, the sintering
experiments were carried out in the conventional
tubular furnace in the same atmosphere and in
the range between 1260 and 1310C for 30 min
(dwell time).
Densities of the sintered parts were measured
using the Archimedes method. Microhardness was measured on polished surfaces with a
Future Tech FM-7 microdurometer. The microstructure evolution was observed by a Philips
XL 30 scanning electron microscope equipped with BSE and EDS detectors. All the EDS
spectra were collected at the some optimum conditions.

Table I.- Chemical composition (%wt) as received M2 powder together AISI specification.

C W V Mo Cr Mn
Present study 0.79 6.39 1.78 5.11 4.24 0.28
AISI specification 0.80-0.9 6.00-6.75 1.75-2.05 4.75-5.50 3.75-4.50 0.10

Experiments were performed in two
facilities, Solar Plant of Almeria (PSA)
and in a Fresnel lens placed in the
Castilla-La Mancha University (ETSII-
UCLM). The solar furnace of the PSA
consists of a flat heliostat of 160 m
2

that follows the sun and it reflects the
solar beam towards a parabolic
concentrator. The maximum
concentration at the focal point (23
cm of diameter) is 3000 times the
solar radiation. In the focal area a
reactor with a quartz window is
placed. The tests were done in a
horizontal position using a cooled
mirror placed in the optical axis of the
Figure 1. M2 HSS powder as observed by SEM
Figure 2. Typical heating/cooling curves in the solar
sintering processing obtained from the thermocouple
register.

0 1000 2000 3000 4000 5000
0
200
400
600
800
1000
1200
T
e
m
p
e
r
a
t
u
r
e

(
C
)
Time (s)
parabolic concentrator. The samples were heated up until the sintering temperature at rate of
50C/min, kept them for 30 min at the maximum temperature and cooled down in the N
2
/H
2

atmosphere at the same speed. Figure. 2 show the thermocouple display.
On the other hand the Fresnel lens at the University of Castilla-La Mancha (Ciudad Real,
Spain) is an acrylic lens with a 1m diameter which concentrates solar energy up to 300
W/cm
2
(concentration factor: 2644) in a 0.8 cm
2
focal area [12]. This concentration is
achieved at 757 mm from the lens in its
optical axis (focal distance). The density
power obtained in its focus is high
enough to increase the sample
temperature to the sintering ones. To
process the samples at different sintering
temperatures the tests were performed
changing the distance to the focus. The
samples were then heated up to the
maximum temperature for 30 minutes.
The total time of the sintering process
(including the heating and the cooling)
was less than 70 minutes in all the cases.
In both solar installations the samples
were situated in a reactor with a quartz
window and the temperature was
measured through a thermocouple placed
in the back of each sample.

Results and Discussion

The microstructures obtained after sintering in the conventional tubular furnace were in
agreement with the results described elsewhere [4]. The microstructure of M2 sintered in the
atmosphere N
2
-5%H
2
at the optimum sintering temperature, around 1290C, presents a
ferrite matrix with some retained austenite, figure 4. An important amount of bright rounded
M
6
C carbides (identified rich in W and Mo) were homogeneously distributed. The sintering
atmosphere has an important influence on the microstructure development because the
nitrogen content of the steel has increased slightly the optimum sintering temperature. The
increase in nitrogen content has no effect on the M
6
C carbides but, as pointed out before by
Jauregi, the MC carbides (rich in V) transform into MX carbonitrides that appear with grey
contrast inside the grains and at the grain boundaries. At 1300C an oversintered
microstructure was detected. It could be described as a continuous M
6
C carbides film around
the grain boundaries. The EDX analysis revealed that at high temperatures the MX
Figure 4. SEM micrograph of M2 sample sintered at
1290C in nitrogen atmosphere in a conventional
furnace.
1300C in nitrogen atmosphere in a conventional
furnace.
VN
M(C, N)
Figure 3. Fresnel lens at the University of Castilla-La
Mancha (Ciudad Real, Spain).
carbonitrides were able to transform to square VN nitrides detected in black contrast in the
grain boundaries.
In the case of the samples treated in the PSA solar furnace, a completed densification has
been found. Figure 6 shows the densification process and the microhardness evaluation. The
feasibility of the sintering process under CSE was demonstrated as the parts have developed
a well-defined neck and the porosity was highly reduced. At 1115 C, figure 7, the
microstructure was close to the optimum but the presence of some porosity in the surface
reduces the total density (90%). In any case, at this temperature, only some rounded and
isolated porosity was observed and a homogeneous distribution of carbides indicates that the
sintering process really took place. DRX and EDX analysis indicated the presence of M
6
C
carbides and probably MC and MX, although, they cannot be well detected up to now, likely
due to their small size. Microhardness measurements indicate that, at this temperature, the
distribution of the carbides was homogeneous and the porosity had disappeared in most of
the sample. The microhardness achieved values similar to those of the samples sintered in
the conventional furnace (580 HV). The faster heating rate could explain that the
densification process occurred so fast. At 1125 C, figure 8, full density was achieved and an
abrupt growing of the grains took place. Massive bright carbide segregation is observed as
well as the formation of the eutectic phase but only in some areas. Most of the grain
boundaries dont show carbides indicating a partial dissolution of them during the sintering
process. The increase of nitrogen in the matrix, thanks to the sintering gas, allowed the
1100 1150 1200 1250 1300
88
90
92
94
96
98
100
Relative Densification
Microhardness
Sintering Temperature (C)
R
e
l
a
t
i
v
e

D
e
n
s
i
f
i
c
a
t
i
o
n

(
%
)
550
600
650
700
750
800
850
900
950
M
i
c
r
o
h
a
r
d
n
e
s
s

(
H
V
)
Figure 6. Densification process of the samples treated in the solar furnace of PSA.
1115C in nitrogen atmosphere in the solar furnace.

1125C in nitrogen atmosphere in the solar furnace.

VN
M
6
C
formation of VN nitrides both inside the particles and along the grain boundaries, these facts
were identified by EDX technique. In spite of the grain growth this sample had more than 900
HV. The rapid heating could be producing a local melting of the grains dissolving partially the
carbides. On the other hand, the rapid cooling could be the cause of the small and well-
distributed nitrides responsible of the high hardness measured. The rapid heating rate of
sintering is not clearly described in the literature so far, making more difficult to explain the
obtained microstructures. The fast solidification of high speed steel was described by
Boccalini and Goldenstein [14] and the studies about selective laser sintering of M2 made by
Niu et al [15] could clarify some results. Increasing of the cooling rate leads to a substantial
bigger amount of retained austenite. This is due to the higher solubility of carbon and
carbide-forming elements in the austenite. The high heating and cooling rate provides
insufficient time for carbon to segregate and to form carbide. The microstructure and the
hardness measurements were similar to those found in the M2 system treated by selective
laser sintering [15]. Concentrated solar energy get to decrease the sintering temperature
around 175C and reduce the processing time from 500 to 90 minutes, and permits to obtain
microstructures with higher hardness without any additional thermal treatment.
Similar results were obtained when the sintering process was carried out in the Fresnel lens.
In this case, the sintering temperature is changed by moving closer and further the samples
from the focus area in the lens axis. The focus area has the maximum concentration of the
flux density. The sintered samples have been expected to receive from 15% to 55% of the
maximum concentration resulting in parts with relative densities that vary from 85 to 99%
respectively. Figure 9 shows the microstructure of a part with 85% of relative density. This
micrograph corroborates the feasibility of this installation to sinter with CSE. Well-defined
necks between the particles have been developed and there are connected and isolated
porosity indicating that the sintering process has started. At the optimum temperature a
similar results than in the PSA were obtained, an abrupt growing of the grains took place
although the hardness increases up to 760 HV. The microstructure at this temperature, figure
10, also shows an important dissolution of the carbides in the matrix, a formation of some
eutectic carbide along the grain boundaries and a homogeneous distribution of VN nitrides
inside the particles and along the grain boundaries identified by EDX technique.

The main difference between the two solar installations used in this research is that in the
Fresnel lens, due to the smaller dimensions of the reaction chamber used, a better control of
the atmosphere and temperature is achieved. Due to this, the microstructures of the samples
sintered in Fresnel lens were more uniform and a layer of porosity in the surfaces was not
observed. The dwell time was the same, 30 minutes, while the heating and the cooling was
faster in the Fresnel lens. This fact permits to sinter the samples in 50 minutes. The Fresnel
Figure 9. SEM micrograph of M2 sample sintered
with 12% of maximum concentration during 30 min
in nitrogen atmosphere in the Fresnel lens.
Figure 10. SEM micrograph of M2 sample sintered
with 45% of maximum concentration during 30 min in
nitrogen atmosphere in the Fresnel lens.

lens equipment reduces the processing time from near to 8 hours to less than one hour
obtaining microstructures with higher hardness without any additional thermal treatment.
As it was discussed before, the fast heating and cooling rate could justify the microstructure
evolution. With the use of a high energy beam, free and pollutant clean, it has been
demonstrated the feasibility of a sintering process. The results indicate that the solar
experimental conditions, quite far away from the equilibrium ones, originating a new
microstructure evolution. It is necessary to go more deeply into the sintering mechanism.
Probably, as Simchi [5] proposed for the laser-sintering technique, the system is changing
between a supersolidus liquid phase sintering mechanism (SLPS) and a localized
melting/solidification mechanism depending on the solar energy provided. This could explain
the abrupt changes observed in the different samples.

Conclusions

A new solar sintering process has been developed for steels. The feasibility of concentrated
solar energy (CSE) to obtain sintered parts of compacted metal powder has been
demonstrated.
The sintering of M2 HSS parts under concentrated solar energy takes place at a lower
temperature compared with the conventional furnace sintering.
The microstructures found have a different appearance than the obtained by conventional
methods. They can be compared with the microstructures developed on the laser sintering of
powder.
Almost full density has been achieved in 75 minutes in the PSA and in 50 minutes in the
Fresnel lens installation with higher hardness than the conventional microstructures (760-900
HV). The higher heating and cooling rates could explain these facts as the equilibrium phase
diagram changes with the heating rate and in these new heating conditions new sintering
mechanisms have occurred.

Acknowledgements
The authors thank JCCM (PBI08-0168-1040) for the financial support. Also they acknowledge
PSA staff for their collaboration in some measurements and for the use of their installations
through the program: Acceso a Grandes Instalaciones.

References
[1] Hoyle, G. Butterworth and Co. Cambridge. (1988)
[2] German, R. M. International Journal of Powder Metallurgy 26, 1. 23-34 (1990)
[3] Zhang, H. Materials and Manufacturing Processes 12, 4, 673-679. (1997)
[4] Jauregi, S., Fernndez, F., Palma, R.H., Martnez, V., Urcola, J.J.. Metallurgical
Transactions A, 23A, 389-400. (1992)
[5] Asgharzadeh, H., Simchi A. Materials Science and Engineering A 403, 290-298. (2005).
[6] Fais A., Maizza G. Journal of Materials Processing Technology 202, 70-75. (2008)
[7] Herranz, G., Rodrguez, G. Proceedings of Euro PM2007. Vol. 3, 425-428. (2007)
[8] Herranz, G., Rodrguez, G. Metal Powder Report, April 28-31. (2008)
[9] Zhilinska, N. Zalite, I., Rodrguez, J., Martinez, D., Caadas, I. Proceedings of Euro
PM2003. Vol. 3, 423-428. (2003).
[10] Almeida Costa Oliveira, F., Shohoji, N., Cruz Fernandes, J., Guerra Rosa, L. Solar
Energy 78, 351-361. (2005).
[11] Roman, R., Caadas, I., Rodrguez, J., Hernndez, M.T., Gonzlez, M. Solar Energy 82,
893-902. (2008)
[12] Ferriere, A., Rodriguez, G.P, Sobrino, J.A. Journal of Solar Energy Engineering 126,
654-660. (2004)
[13] Caadas, I., Martnez, D., Rodrguez, J., Gallardo, J.M. Revista Metalurgia SPEC2005,
165-169. (2005)
[14] Boccalini, M., Goldenstein, H. International Materials Reviews 46, 92-107. (2001)
[15] Niu, H. J., Chang I.T.H. Journal of Materials Science 35, 31-38. (2000)
Manuscript refereed by Professor Detlev Stoever, Forschungszentrum J lich GmbH,
Germany

Investigations of Intermetallides, Obtained in Ti-Al-B System in
Dynamic Conditions

Mikheil Chikhradze
1,2
, Akaki Gigineishvili
1
, George Oniashvili
1,2
and Nikoloz Chikhradze
1,3

1
Georgian Technical University, 77, Kostava St., 0175, Tbilisi, Georgia, e-
mail:chikhradzem@yahoo.com
2
F. Tavadze Institute of Metallurgy and Materials Science, 15, Kazbegi St., 0111, Tbilisi,
Georgia
3
G. Tsulukidze Mining Institute 7, E. Mindeli St., 0186, Tbilisi, Georgia

Keywords: Intermetallide, shock wave, compaction, SHS

Abstract

In the paper are presented the results of investigations of the reactions in Ti-Al-B system
induced by shock wave created extreme conditions. Extreme conditions were formed by
explosive loadings in the cylindrical scheme. Theoretically calculated were the energetic
characteristics and stress tensor components in steel container containing the Ti-Al-B initial
powder composition. The phase constitution of obtained samples, microstructure and
mechanisms of intermetallides formation are investigated depending on the initial ratio of
components in reaction mixture. Welding zone of intermetallides and cylindrical surface of steel
ampoule were studied and the micro-hardness distribution in the structure was determined.

INTRODUCTION

Alloys obtained in Ti-Al and Ti-Al-B ternary systems and technologies of their fabrication are the
subject of continuous intensive investigations due their practical applications. Intermetallide
compounds obtained in Ti-Al system are attractive materials for aerospace, machine and
chemical applications [1, 2, 3]. They have high specific strength under tensile and compression
conditions. At the same time they have good high temperature corrosion, oxidation and wear
resistant properties and widely used in chemical technologies. There are known several
conventional methods of preparing alloys in Ti-Al system: mechanical alloying, self propagating
high temperature synthesis (SHS), Laser engineered net shaping, etc.
Each method has advantages and disadvantages. The one of main difficulties of all-mentioned
technologies are connected with control of phase formation and fabrication of materials with
desired homogenous structure. The next main problem is limitation in shape and sizes of
processed materials. Preparation of layered multifunctional materials in situ is limited as well.
Synthesis of Ti-Al intermetallics using shock wave compaction technology was the interest of
several investigations as well.
The shock wave compaction technology (SC) has advantages, which becomes attractive in
combination with SHS for common application. During the last years, numbers of investigations
by a group of Professor N. Thadani were carried out for understanding of mechanisms of SHS
processes in shock loaded reaction mixtures. The main purpose of investigations was the
establishment to assist or induce the SHS by Shock waves. But on the base of authors
conclusions can be said that the researches in this direction are not completed yet, especially
for multi component systems.
This method is very attractive, because of the simplicity and low cost of technology and good
potential for obtaining different dimensional materials. The following mechanical treatment may
be used for obtaining materials in different shapes. But in spite of above mentioned very
important advantages, this method is too difficult to control, because of the rapidly propagated
very high strain rate conditions, formed under shock wave. Initiation of shock waves using
explosives is dangerous as well, if using sensitive and powerful explosive materials, which
0
5
10
15
20
25
0 0,5 1 1,5 2 2,5 3 3,5 4 4,5 5
Distance, y, mm
T
a
n
g
e
n
t
i
a
l

S
t
r
e
s
s
,

G
P
a
P=5GPa
P=10GPa
P=20GPa
generate necessary high pressure on the shock wave front. Accordingly, the correct modeling of
the shock wave induced reaction in Ti-Al-X system, experimental testing and determination of
optimal technological characteristics is very actual for prediction of results and controlling of
phase compositions of intermetallics.
In this research has been made the preliminary theoretical calculations of stress tensor
components, estimation of energetic conditions using the classical and empiric relations and
experimental adiabatic conditions of Ti and Al/B under the shock wave loadings. The calculated
results were tested experimentally, correlated and confirmed. The hardness distribution and
structure of compacted Ti/Al/B intermetallic in different zone was investigated.

THEORETICAL INVESTIGATIONS

Stress components in the Ti, Al and B under explosive loadings, were calculated using the
solutions of a dynamic model [4].
It was taking into account that the powder blend is situated into the steel ampoule. Pressure
impulses generated by explosives (and propagated along the treated surface at the boundary
explosive material-ampoule with the constant rate V=V
detonation
) were estimated using the shock
adiabatic. It is clear, that stresses will be determined in multitude of real numbers if the
conditions V C
1
, where, C is the longitudinal sound speed in material.
Some characteristics of explosives, used for calculations are presented in table 1.

Table I. Composition and Detonation Characteristics of Explosive Materials [4, 5]

Explosive Material Density, 10
3
g/m
3
Detonation rate, V m /s Pressure on St.3 P 10
9

N/m
2

79%NH
4
NO
3
+
21%C
6
H
2
(NO
2
)
3
CH
3

1.0-1.2 3600-4800 10
95%NH
4
NO
3
+
5%C
6
H
2
(NO
2
)
3
CH
3

1,00 1700 5
C
3
H
6
O
6
H
6
1.0 6050 20

Distribution of normal stresses
y
=f(y, x
0,
t
0
), in loaded assembly of cylindrical geometry are
presented in fig.1 and fig 2.

Fig.1. Distribution of normal stresses in the shock loaded Ti /Al composition.
Distribution of Normal Streses in Shock Wave Loaded Boron
0
2
4
6
8
10
12
0 0,5 1 1,5 2 2,5 3 3,5 4
Y, mm
S
i
g
m
a

Y
,

G
P
a
P=5GPa
P=10GPa

Fig.2. Distribution of normal stresses in shock loaded B.

Taking into account the above presented calculations and Ti-Al and Ti-B phase diagrams, we
can determine the desired stress-deformed state of composition for obtaining optimal results
during explosive loadings.

EXPERIMENTAL PROCEDURE

Starting elemental powders used were coarse Ti (99.0% purity), Al (99.0% purity) and
amorphous B (purity - 99.6%, containing: C-0.13%, Cl-0.05%, metallic impurity totals-0.5%;
specific surface 14m
2
/g). Three initial compositions consisting of Ti/Al and Ti/Al/B elemental
molar ratios of 1.38:1.25, 0.5:0.5:2.0 and 0.2:7.0:0.2 were prepared for compaction under shock
wave, generated by chemical explosions.

Fig. 3 Scheme of EC/SHS setup: (1) electrical detonator, (2) explosives container, (3) explosive, (4) steel
tube, (5) reaction mixture, (6) steel plugs, (7) base table, (8) central rod, (9) detonator seizer.

The ratio of Ti, Al and B were selected according phase diagrams and focused on formation of
intermetallics TiAl
3
and TiB
2
phases. The precursors were processed in ball mill for activation of
surfaces before compaction. Particle size was in range of 40-80m. For powder compaction the
axial-symmetric experimental set-up of cylindrical geometry was chosen. Scheme of shock
wave loading of assembly is shown in Fig. 3.
A cylindrical steel rod (4), 8mm in diameter, was placed along the powder container - tube axis.
The powder was packed in a low carbon steel (St.3) tube around the central steel rod and pre-
densification up to 50% of theoretical mass density. The powder container closed in top and
bottom sides by steel plugs. The length of powder container was 120mm, diameter - 22mm, wall
tikness-4mm. The assembly was placed in cardboard pipe (explosive container), 44mm in
diameter and 164mm in length and surrounded with powdered explosive materials. The
ammonite, 50 wt.% ammonite - 50 wt.% ammonium nitride and Hexogene were used as a
explosives.

On the top side of explosive container placed the drilled disk detonator seizer and after
charging the set-up were detonated electrically. Pressure and high temperature on the front of
shock wave configuration, in the Al-Ti/B system induce combustion syntheses and realized the
in situ process of shock compaction + SHS (EC/SHS).

RESULTS AND DISCUSSIONS

Compacts of cylindrical geometry were obtained after EC/SHS consolidation and some of them
followed by rolling 50%deformation at 800 C. The photographs of the compacts, cut from the
long dimensional cylinders and rolled one of ellipse form are shown on Fig. 4.

Fig.4. Photograph of SC/SHS compacted Ti/Al/B samples of different shape

It must be mentioned, that the central rods (4 on Fig.3) prevents and decreases Mach stream
effect influence in assembly, loaded by shock wave, induced ammonium nitride containing
explosives. At the same time, in such conditions is formed special configuration of
induced/reflected/unloading waves in ampoule preservation and increase impact time (as result,
increase in the workability of explosive energy). Using the more powerful explosive-hexogene
gives destructive effects in the bottom of the assembly.

Following SC/SHS, the compacts were recovered in different shapes and prepared for
investigations. The density of specimens were determined (cut from different part of samples),
by the Archimedean method. Investigations on hardness distributions along the diameters were
performed on Samples 1 and 2 using a diamond Berkovich indenter probe. The line ran along
the radius from the edge to the center across the diameters (long and short in case of ellipse
shape). Spacing of the indents was decreased near the interfaces of the intermetalides and
steel. The loading function consisted of a five-second linear loading segment to a peak load of
10 mN. Character of microhardness distribution along the diameters of ellipse shape cross
section of SC/SHS + 50% rolled Ti/Al/B compact is shown on Fig.5.
0
1
2
3
4
5
6
7
8
0 5 10 15 20 25 30 35
Location (mm)
H
a
r
d
n
e
s
s

(
G
P
a
)
Long Diameter
Short Diameter

Fig. 5: Hardness versus position from the line of indents on ellipse shape, obtained after 50% rolling of
compacted sample obtained by SC/SHS from Ti/Al/B composition

As we can see the curve of distribution of above properties (between intermetallics and Steel
surfaces in the welding zones) is close to sinusoidal shape, the change is quite smooth, which
is very important for stability and reliability of functioning during its usage in extreme conditions.

Fig.6 SEM and ESD images of EC/SHS 0.5Ti/0.5Al/2B compact
The type and sizes of pores in compacts were determined. The pore sizes are in range of (25-
75) m and total porosity of compacted samples is up to 2.25%.
The recovered SW/SHS compacts were investigated using scanning electron microscopy (SEM)
SEM-800. The Shock induced SHS assisted with plastic deformation process in grains by SEM
images was confirmed (Fig. 6).

CONCLUSIONS

The Investigations were carried out to explain formation of Titanium aluminade intermetallics via
reaction under shock wave compaction. Ti-Al based powders of three different compositions
were explosively loaded in cylindrical mode and successfully realized shock-wave induced SHS
reactions. The stress components, generated by pressure impulse, propagated along the
surface by constant rate velocity have been theoretically calculated. The calculated results were
compared with experimental result and were confirmed. Microhardness distribution along the
diameter was determined of synthesized samples. The SEM analyses were carried out for
investigation of microstructure and phase composition. The porosity was estimated using the
optical microscope and possibility of obtaining high dense compacts was confirmed.

ACKNOWLEDGMENT

The designate project has been fulfilled by the support of Georgia National Science
Foundation (Grant #GNSF/PRES08/346). Any idea in this publication is possessed by the
author and may not represent the opinion of Georgia National Foundation.

REFERENCES

1. A. M. Limarga, D. S. Wilkinson, Creep-driven nitride scale growth in -TiAl, Acta
Materiala Inc. actamat. 2006.07.024
2. H. B. Lee, S. H. Kim, S. W. Kang and Y. H. Han. Characterization of mechanically
alloyed Ti-Al-B nanocomposite consolidated by spark plasma sinterng, British Ceramic
Transactions, 2003 Vol. 102, # 6, 231
3. D. Horvitz, I. Gotman, Pressure-assisted SHS syntheses of MgAl[2]O[4]-TiAl in situ
composites with interpenetrating networks, Acta materiala, ISSN 1359-6454
4. N.M. Chikhradze, L.J. Kecskes, I. Lomidze, L. A. Japaridze and A.B. Peikrishvili,
Calculation of Stress Deformed Conditions of Explosive Loaded Powder Materials, PM
2

TEC-2002, Orlando, USA.
5. F.A. Baum, K.P. Ctaniukovich, B.I. Shekhter, Explosion Physics, Gosudarstvennoe izd.
Piziko-matematicheskoi literaturi, Moscow, 1959.
France

Surface and Mechanical Properties of Laser Melted Maraging Steel

Marc Jobin, Raphael Foschia, Franois de Mestral, Eric Vittecoq
hepia, University of Applied Science HES-SO, 4 rue de la Prairie, 1202 Genve, Switzerland

Denis Boisseleau, Alexis Saitta, Pierre Escourou
Turbomeca , SG/FE - BP Piano 2, 64511 BORDES Cedex, France

Guillaume Vansteenkiste
Ple Europen de Plasturgie (PEP), 2 rue P. et M. Curie - 01100 Bellignat, France

Cyril Gouverd, Laurent Cataldo
BESTinCLASS SA, Chemin de la Caisserie, 1272 Genolier, Switzerland

Carmen Sanz, Oscar Gonzalo
Fundacion TEKNIKER, Avda.Otaola 20, Eibar 20600, Spain

Abstract

In order to drastically reduce the cycle and cost production of engine parts in the transport
industry, we have developed laser powder melting processes to produce several key engine
parts in maraging steel, Co-Cr and Inconel 718. Results presented here concerns maraging
steel only. Laser melted samples were polished with three different treatments: Abrasive
Flow Machining (from Extrudehone), Vibratory finishing (from Rsler) and MMP (Micro
Machining Process from BESTinCLASS). The dimensional tolerance after polishing was
5m.
The resulting surface topographies were measured with Stylus Profilometry and Atomic
Force Microscopy. The MMP process allows reducing the rms roughness (Rq) to 12 nm only
for a 20m x 20m field of view. In order to investigate the surface treatments in terms of
what characteristic length (i.e spatial frequency) have been filtered, we present the results
with the help of power spectral density. Concerning the mechanical properties, the anisotropy
of fatigue and tensile strength is not significant as a consequence of the direct melting
process compared to sintering.

Introduction
This work is part of our RC2 European project (Reduction of cycle and cost
www.rc2project.eu), which aimed to reduce both the time-to-market and the production cost
of nine engine parts (turbine wheel, straightener, etc...) by 50%. Each of those 9 parts was
produced by rapid prototyping techniques in 3 materials: Maraging Steel (MS), Cobalt-
Chromium (Co-Cr) and Inconel 718. Although all of them have specific and interesting
applications in various industries (mechanical, medical, aeronautic), will present in this
paper the results for MS only for the sake of conciseness.
We have used a commercial Selective Laser Melting (SLM, or commercial denomination
DMLS for Direct Metal Laser Sintering) machine (EOS M270) which allows the actual
melting of the powder to a depth of 30-40 microns. We will then refer to the produced
materials as laser melted rather than sintered.
The surface finishes of the pieces are potentially important for both the mechanical
properties and the visual aspect. Indeed, as standard sintered materials have a porosity
which very likely will destroy tensile strength or fatigue resistance, we first expected a strong
correlation between the polishing treatments and the mechanical properties. However, we
found that laser melted materials had very few surface defects or porosity. The analysis of
the microstructure after fracture revealed that the cracks mainly originated from non-metallic
inclusions close to the surface. We therefore investigate the surface topography in term of
roughness rather than in term of defects or porosity.

Sample preparation
All samples have been produced as 5cm x 5cm x 2cm parallelepiped with a commercial
selective laser machine from EOS (model M270). The MS powder had a typical diameter of
20 m and was bought from EOS directly. Microstructure investigations revealed the melting
process very nicely, as crystallites were clearly extended over the thickness of the powder
film.
Fig 1a) is a SEM micrograph (50x, 2.3mm x 1.6 mm) showing the surface of the laser melted
MS without any additional treatment (raw). We clearly see the laser path, each track
typically 100 um wide. To a large extent, this is the structure that has to be polished.

a) Raw laser melted b) Abrasive Flow Machining

c) Vibratory finishing d) MMP

Fig. 1 SEM pictures of Maraging Steel

Three different polishing treatments have been used on raw MS: Abrasive Flow Machining,
Vibratory finishing and MMP (Micro Machining Process) from the company BESTinCLASS
(BinC). The parameters of the polishing processes have been chosen by the industrial
provider of the process directly. The target was the smallest roughness under the following
constraint: 5m dimensional tolerance on a standard sample with various geometrical
structures (including a cylinder of 2mm diameter and 3mm high, and a corresponding hole).
Fig 1b), 1c) and 1d) are SEM micrographs (same magnification than Fig1a) of the polished
samples. In all cases, we see a complete removal of structure due to the laser path. For
Abrasive Flow Machining, the direction of the abrasive flow is quite visible, as expected,
while for Vibratory finishing and MMP, the polishing is isotropic (no preferential polishing
direction). In the case of Vibratory finishing, we have a structure which can be seen as 150-
400um spot in the backscattered SEM image.

Surface Profilometry

Both surface profilometry (SP) and Atomic Force Microscopy (AFM) have been use for
surface topography analysis. For SP, we choose a standard ISO scan length of 5.7 mm
acquired on 8000 points. Data were taken with a 1m tip radius probe on a Mahr Perthomer
PGK instrument. Although any 5.7mm scan is reasonably representative (i.e. without any
particular attention paid to the location), we measured 5 different profiles for FFT analysis
As an illustration, the top (colored) profiles of Fig. 2 have been measured on the raw MS and
the bottom (black profiles) has been measured on the MMP.

Fig. 2 : Actual SP profiles on MS raw and MS MMP.

Topographic profiles such as those of Fig. 2 are usually characterized with amplitude
(including standard Ra and Rq roughness), spacing or hybrid parameters [1]. In this study,
we choose to analyze the SP profiles in terms of Fourier power spectrum:

where z(x) is the height profile acquired by the profilometer. Each wave vector k corresponds
to a spatial wavelength given by:

k / 2

The motivations for representing the topographic profiles as F(k) were:
- the various polishing treatment do not operate uniformly over the spatial wavelength
of the topography. A key aspect of this work was to analyze and to understand what
wavelengths are filtered by the polishing process and what wavelengths are not;
Raw
MMP
1
0
2
) (
2
sin( ) (
2
cos( ) ( ) (
N
x
x x
x k
N
i x k
N
x z
N
L
k F
- we found a direct correlation between the optical reflectivity (and thus the visual
aspect) and the partially integrated Fourier spectrum.

1.E+00
1.E+01
1.E+02
1.E+03
1.E+04
1.E+05
1.E+06
1.E+07
1.E+08
1.E+09
0.01 0.1 1 10
k [um
-1
]
F
(
k
)
Raw
Abrasive Flow Machining
Vibratory finishing
MMP

Fig. 3 : F(k) of the SP profiles for the raw and polished MS samples

Fig. 3 is a log-log plot of F(k) for the raw and polished Maraging Steel samples. The MMP
treatment provides with the most efficient polishing over the entire spectrum, reducing each
Fourier component by at least two orders of magnitude. By contrast, the Abrasive Flow
Machining and Vibratory finishing treatment have no effect in the k [0.3; 1] (m
-1
) range. The
Vibratory finishing process is however quite efficient for k>2 (m
-1
).

Atomic Force Microscopy

For AFM, we decided to perform 20m x 20m scan on a location which exhibited no
particular structure a long as optical microscopy is concerned. All images were done in
contact mode with high aspect ratio conical silicon probes
(PointProbe from NanoWorld). We have used a commercial AFM from Park Systems (model
XE-100).
High definition (1024 x 1024 pixels) images were acquired in order to perform Power
Spectral Density (PSD) analysis, the 2D counterpart of the 1D FFT of the surface profile
reported above.

Abrasive Flow Machining (Rq
= 1.23 m)
Vibratory finishing (Rq =
4.74 m)
MMP (Rq = 12 nm)

Fig. 4: AFM images and Rq roughness on the polished maraging steel
Typical AFM images on the polished Maraging Steel samples (raw sample was too rough for
AFM) are shown in Fig. 4. The directionality of the Abrasive Flow Machining process is
clearly observable, even at the scale (20um x 20um).

As for SP profiles, analysis in the wavelength space has been performed with AFM images.
While SP provides results only on the direction of the profile, AFM, as providing with a 2D
images, gives information in all directions. We decided to choose to represent the result in
terms of Power Spectral Density (PSD), which is the integration of the Fourier power
spectrum over the k-space:

2
2
2 2
) , (
1
) (
f f f
y x
x
f f F
L
f PSD

NB : For PSD in AFM, the spatial frequency f=1/ is usually used instead of the wavevector
k=2 / .

As the Rq (i.e rms roughness) is given by the integration of the PSD

f
q
df f PSD R ) (
each PSD(f) point corresponds to the contribution to the roughness of the structure having a
characteristic spatial frequency of =1/k, referred to as superstructure.

PSD analysis on AFM images is very powerful [2, 3] provided one take care of the following
possible artifact: during AFM imaging, a topographic line can be shifted because of
instability due to a surface impurity (dust). If such a line scan is not properly interpolate by
image processing, a strong oscillation will appear in the PSD.

Fig. 5 shows the PSD analysis for the Maraging Steel polished samples. At such high spatial
frequencies, we see that MMP is more efficient at any frequencies than Abrasive Flow
Machining, itself more efficient than Vibratory finishing (again at any frequencies). The k
values in the [7; 20] interval correspond to the optical wavelength and is well correlated with
optical reflectivity results.

1.E-10
1.E-09
1.E-08
1.E-07
1.E-06
1.E-05
1.E-04
1.E-03
1.E-02
1.E-01
1.E+00
0.1 1 10 100
Spatial frequency [1/um]
P
S
D

(
n
m
4
)
MMP
AbFlMa
VibFi

Fig. 5: PSD calculated from 20m x 20m AFM images on polished MS samples.

Mechanical properties

Although the laser melting process is very anisotropic by nature (layer by layer fabrication),
we did not notice large differences of the mechanical properties depending on the direction:
both fatigue and tensile strength were rather isotropic. However, we get significant difference
depending on the height (the first melted layer is at height zero) on which the mechanical
properties have been measured. This is due to thermal diffusion process (which is faster at
lower height), which results in a differentiated thermal treatment depending on the height (the
temperature of the part rises with increasing height). Extensive results on the heterogeneity
mechanical properties of laser melted samples will be published elsewhere [4].

References

[1] E.S Gadelmawla et al., J.Mat. Proc.Tech. 123 (2002) 133-145
[2] S. J. Fang, S. Haplepete, W. Chen, and C. R. Helms, J. Appl. Phys. 82 (1997) 5891
[3] M. Flemming, K. Roder and A. Duparre, Proc. SPIE 5965, (2005) 59650A
[4] E. Vittecoq, F. de Mestral, et al, to be published.
Germany

Selective Laser Melting (SLM) Optimisation Using Design of
Experiment (DoE)

T. Kurzynowski, M. Baranowicz, E.Chlebus
Wroclaw University of Technology
CAMT - Centre for Advanced Manufacturing Technologies
ul. Lukasiewicza 5, 50-371 Wroclaw, POLAND

ABSTRACT
The field of Rapid Prototyping (LBMM) is quite young. It started when 3D Systems Inc.
presented the first stereolithography machine. Since that time many RP machines based on
plastics were developed. Nowadays, LBMMs have transformed into generative technologies,
capable of manufacturing objects from selected plastics and almost any metal (stainless and
tool steel, titanium alloys, Co-Cr, aluminium alloys). For metal technologies the objective is
the capability to manufacture fully functional products (e.g. medical implants) with
customized mechanical properties. Design of Experiments is one of the most effective
methods of investigating various input variables on result of a process. Applying statistically
designed experiments allows the study of the effect of several variables at one time
including interactions. In this paper authors will present the results of the application of this
method in Selective Laser Melting Process. Fractional Factorial Design were used to
examine the influence of five potential factors on density.
INTRODUCTION
The search for new products with individual properties, cost reduction and higher quality has
forced engineers and scientists to develop new materials, technologies and to optimize
existing ones. One of these technologies is Selective Laser Melting (SLM) which uses a wide
range of metal powders. Metallic powders of single composition are successively melted in a
micro zone by a laser and solidify with controlled porosity or almost full density. Building of
3D models from tool steel with different parameters were investigated and results are
presented in this paper.
SELECTIVE LASER MELTING TECHNOLOGY
The SLM technology is one of the additive methods that create models by selectively melting
layers of powder and are capable of creating geometrically complex parts, including internal
structures.
It is based on local melting of a thin layer of metal powder by a focused beam of Nd:YAG
laser with the maximum power of 100W.

Fig. 1. Principle of the SLM technology

The layers are created by depositing thin layers of powder (depending on the grain size it
may be from 30 to 100 m thick), which are leveled by a wiper and then selectively melted by
a laser, in areas within the cross-section through the model of the part at the current level.
The metal powder is dosed from a build powder container, where it is supplied by a system
of valves from the main powder container [1, 2, 3]. After a pattern is melted on a layer, the
building platform moves down by a thickness of a layer, new powder is deposited and leveled
and the process is repeated until the complete part is built (Fig. 2).

Fig. 2. Model of the SLM technology

DESIGN OF EXPERIMENTS METHODOLOGY
Design of Experiments (DoE) is a powerful approach to product and process development,
and serves to improve the yield and stability of an ongoing manufacturing process [4]. It is a
methodology that applies statistics to the experimental process. It was first introduced and
developed by Sir Ronald Fisher in the early 1920s to study the effect of several variables
simultaneously on the outcome. In his early applications, Fisher wanted to determine the
effect of factors such as rain, water, fertilizer, sunshine, etc., on the final condition of the crop
[5]. Methods developed by Fisher were a fundamental break from traditional one-factor-at-a-
time (OFAT) or hit or miss approach to experimentation. Design of Experiments is
essentially a strategy of industrial experimentation whereby one may vary a number of
factors in a process simultaneously to study their effect on the process output.
EXPERIMENT
In powder metallurgy it is known that the most important parameter that has the highest
impact on mechanical properties is a density of produced parts.
A lot of parameters have an influence on the process. There are parameters that we can pre-
set (e.g. process parameters, powder properties) and parameters that cannot, or they not be
changed (e.g. the environment).
Environment - the atmosphere in which the process takes place, is the first thing to control in
SLM. Non-inert particles, like oxygen, can react with the molten metal. This will influence the
final properties of the component and should be avoided [6]. The controlled atmosphere is
achieved with the argon gas, and the oxygen level should be kept below 2%.
Temperature the temperature in the building chamber can be set between 22 (room
temp.) and 200C. The higher the temperature the better for processing. The properties of
metal powder are better (flowability, creating of layers) and it is easier to melt powder layers
with laser (the powder is preheated).
Powder the next group of properties is the material that is used to produce the part. The
choice is made by the engineer to obtain suitable properties for an application. The selected
powder must be spherical. The shape determines the possibility of powder processing. In
SLM technology the powder is deposited with gravity through a system of valves. If the
powder was irregular it would stuck in the valves and smooth layer would not be created.
Process parameters there is more than 100 process parameters. This parameters
influence the heat balance (laser power, layer thickness), the building speed (distance
between scanning points, scanning lines, time of scanning one point):
laser power
layer thickness

distance between scanning lines

distance between scanning points

scanning time of one point

The goal of the experiment carried out was to determine the influence of five factors on the
density of produced parts (Tab. 1).

Tab. 1. List of factors for the experiment

In this situation, with five input variables, conducting replicated full factorial experiment would
require 64 experimental trials. It was decided to run fractional factorial design rather than full
factorial design. The best possible design matrix for this experiment was fraction 2
(5-1)

fractional factorial design. The selection of design matrix with resolution V resulted with an
alias structure which allowed for analyzing the influence of every factor and interaction
independently: A + BCDE, B + ACDE, C + ABDE, D + ABCE, E + ABCD, AB + CDE, AC +
BDE, AD + BCE, AE + BCD, BC + ADE, BD + ACE, BE + ACD, CD + ABE, CE + ABD, DE +
ABC. Each experimental design point was replicated to increase the precision of the
experiment. To minimize the effect of bias induced into the experiment, the trial conditions
were also randomized.
In order to identify the most important main and interaction effects, it was decided to
generate Pareto Chart of the Standardized Effects (Fig. 3) and Normal Plot of the
Standardized Effects (Fig. 4).

Fig. 3. Pareto Chart of the Standardized Effects

Fig. 4. Normal Plot of the Standardized Effects

Figures 3 and 4 show that four main effects (B, C, D and E) and five interactions (BD, AE,
BC, CD and DE) are found statistically at five per cent significance level. The Pareto Chart
and Normal Plot indicate that the distance between scanning points and scanning time of
one point, followed by interaction between those two factors, have the greatest influence on
the density of produced parts.
In order to determine the optimal selective laser melting factor settings which yield the
maximum density of produced parts, main effects plot was produced (Fig. 5).

Fig. 5. Main effects plot for density

The presented main effects plot confirms that the distance between scanning points and the
scanning time, have the greatest influence on the considered process output. The graph also
shows that the maximum density was achieved when the scanning time was on level +1 and
the distance between scanning points was on level -1. In this experiment, Pareto Chart
presents that beside those two factors, interaction between them should be considered as
statistically important. The interaction between the scanning points and the scanning time of
one point is shown in Figure 6. The interaction plot presents that the density increases when
the distance between scanning points is on the low level and scanning time is on the high
level.

Fig. 6. Interaction plot - scanning points x scanning time of one point

CONCLUSION
The paper shows the importance of setting proper parameters to obtain preferable
(customized) material density, which determines mechanical properties (end-user
requirements) of a part being built. It is important for parts like inserts for injection moulding
tools with conformal cooling channels (high density to prevent leakage).
Tab. 2. Results of the experiment

Also after these experiments we can define 5 irregularities that have the biggest impact on
the density in the selective laser melting (laser micrometallurgy) process:
Effect of the start point (Fig. 7a) is connected mainly with line scans. It occurs when
the process starts. At the beginning of a scan track, more powder particles are
available in the neighbourhood, and these particles are attracted to the melt pool,
creating a larger ball [6]
Balling (Fig. 7b) occurs when the energy delivered to the melt pool is too small or
the time of the laser melting is too short [7]
Satellites (Fig. 7c) some solid powder grains are connected with the melt pool. They
do not melt. Powder grains are surrounded by liquid metal.
Pores (Fig. 7d) this irregularity is connected with balling (the main cause of
porosity). Porosity is the main problem of the laser powder micrometallurgy that has
the highest impact on mechanical properties of produced parts.
Effect of the first layer (Fig. 7e) the melted powder on the first layer on top of the
support structure sinks down into the free powder below.

Fig. 7. Five irregularities in the selective laser melting (laser micrometallurgy) process: a) the
effect of the start point; b) balling; c) satellities; d) pores; e) effect of the first layer

REFERENCES
[1] T. Kurzynowski, A. Lipowicz, B. Dybaa Customised medical model building with
Selective laser Melting, Production Engineering 2006, pp. 105-110
[2] A. Lipowicz, T. Kurzynowski, E. Chlebus Rapid Manufacturing in medicine, Journal of
Material Testers 2009/1, pp. 27-36
[3] T. Kurzynowski, E. Chlebus, B. Dybaa Manufacturing of customised products using
selective laser melting technology 7
th
Youth Symposium on Experimental Solid
Mechanics 2008
[4] Montgomery D.C. The use of statistical process control and design of experiments in
product and process improvement, IEE Transactions, Vol. 24 No. 5, 1992,
[5] Sirvanci, M.B., Durmaz, M. , Variation reduction by the use of designed experiments,
Quality Engineering, Vol. 5 No. 4, 1993,
[6] M. Van Elsen, Complexity of Selective Laser Melting: a new optimisation approach,
PhD Thesis, Katholieke Universiteit Leuven, 2007
[7] H. J. Niu, T. H. Chang , Instability of scan tracks of selective laser sintering of high
speed steel powder Scripta Materialia, 41, No. 11, 1999, pp. 1229 1234
Germany

SHS Technology Coupled with Thermal Explosion for Production of
Single-Phase Nanocrystalline Materials of Ti-Al System

G.Oniashvili
1
, G. Zakharovi
1
, Z. Aslamazashvili
1
, A.Peikrishvili
2

1
F. Tavadze Institute of Metallurgy and Materials Science
2
G. Tsulukidze Mining Institute

Abstract

The paper describes the processes of obtaining Ti-Al in SHS and thermal explosion mode.
As this stage there no adequate explanation of the mechanisms of microstructure formation in
this system. However, study of grain size influence on mechanical properties of the
intermetallics at room temperature, revealed an unexpectedly strong dependence of plasticity
on the above microstructure parameter. Thus, reduction of grain size to submicron level (d<1 )
in leads to drastic shift of brittle-viscous transition temperature range to the low temperatures,
and also increases plasticity at room temperature. The increase in plasticity, at room as well as
at elevated temperatures, after grain sizes in Ti
3
Al and TiAl reach a definite value, is related
with the improvement of relaxation ability of grain boundaries.

Introduction

Wide application of the Single-phase Ti-Al compositions is restrained by the lack in effective and
inexpensive production technologies.
Development of relatively simple and reasonably priced technology for production single phase
and nanostructured intermetallics in Ti-Al system was an important and unsolved task. For now
the problem can be solved according to the works described in this paper.
In Ti-Al system the synthesis of TiAl is possible by Self-propagating High-temperature
Synthesis, but the synthesis of Ti
3
Al and TiAl
3
is possible by thermal explosion due to their low
exothermal ability.
In order to solve the above tasks we developed an original and universal equipment for
performance of SHS in the mode of thermal explosion. The equipment is capable to fabricate
TiAl, Ti
3
Al and TiAl
3
single phase materials and components. The equipment of thermal
explosion is presented on picture 1. The experiments which were carried out to obtain single-
phase compositions enabled us to determine the conditions and optimal characteristics to
obtain desired single-phase composition.
Setchiometric compositions TiAl (Ti-63,95% weight, Al-36,05% weight) (Ti-84,18 % weight, Al-
15,82% weight) and (Ti-37,16% weight, Al-15,62 % weight) , Ti and Al powders are mixed in ball
mill for 8-10 hours. The obtained mixture is dense and formed under 0,09-0.2 Kbar pressure for
TiAl, 0.5-0.7 Kbar for Ti
3
Al and under 0.25-0.4 Kbar for TiAl
3
, the obtained sample is heated
150-200 C/minute speed for TiAl, 380-400 C/minute for Ti
3
Al, and 2-4 C/minute for TiAl
3.
During
the TiAl synthesis 1280-1320
0
C temperature is developed, 1050-1100
0
C temperature during
Ti
3
Al synthesis and 1250 - 1300 C during TiAl
3
synthesis.
After the synthesis the compaction under pressure takes place in order to have for non porous
conditions: for TiAl, 0.05-0.09 Kbar, for Ti
3
Al 0.07-0.09 Kbar and 0.04-0.08 Kbar for TiAl
3
.

Picture #1 Thermal explosion equipment

Experimental Procedures

The idea of obtaining nanostructure materials in Ti-Al system proposed by us is that during
mixing of components high quality deformation is performed, after that pre-compaction and
plastic deformation takes place under high pressure. At the next stage the high speed heating
process is realized in thermal explosion regime, which is possible during realization of SHS for
high exothermic chasm. As soon as the process is finished the sample is finally compacted and
plastically deformed under P 1,2 kbar pressure.
As far as we are using high exothermic chasm for obtaining TiAl and Ti
3
Al in thermal explosion
mode, during which the synthesis and emission of energy happens in SHS mode, it is possible to
obtain the high quality deformed and compacted sample of Ti-Al system in 4-8 seconds.
Heating up to high temperature, which enables to obtain nanostructured admixture of Ti-Al
system, and during the immediate compacting the temperature is sharply reduced and the grains
could not increase in that short time and the nanostructured pattern is obtained.
It must be mentioned that according to the authors opinion, the mechanisms of formation of
nanostructure in Ti-Al system is connected to the following: multi high quality plastic deformation,
fast speed heating up to high temperatures and relativli fasrt cooling which prevents the increase
of grains.
The influence of the process on grain formation can be described as follow: at first during the
mixing the Ti grains break into small pieces, also Al grains are partly broken, and the small
particles of Ti and Al are plastered to one another. The process can be discussed as a result of
destruction of defective dislocation stricture formed during the plastic deformation. The first stage
can be supposed as mixing of chasm components with high quality plastic deformation.
At the second stage the plastic deformation, dislocation changes and following breaking process
of defective structure continues while forming in special press forms of various size (44x44 mm,
70x70 mm,90 mm,25 mm) in under high pressure at room temperature. This stage can
be discussed as pre-compaction process. In the samples partly broke Ti and Al particles are
under cold ductility effect, because in the particles of defective dislocation structure, formed
under platic deformation during the cold ductility, there is no conditions for dislocations to move
easily due to lack of energy.
In our experiments carried out in Ti-Al system for synthesis in thermal explosion mode, high
exothermic chasm (such as: Ti -B, Ti C- Me, Ti-Cr-C-Me, Ti-C-B-Me , etc) were used. They are
sensitive during SHS and for 70x70mm dimension sample in 2-7 seconds is generated enough
energy for heating TiAl or Ti
3
Al samples up to T 2000C. Because of high heat effect
dislocations are moved easily and fast, the dislocation structure is destroyed new nano size (60-
90 nanometer) TiAl or Ti
3
Al grains are formed.
The third stage can be discussed as high-temperature action and synthesis of aimed mixture.
Just as soon as the TiAl or Ti
3
Al are synthesized in thermal explosion mode, the immediate
compaction of melted final product takes place for obtaining non porous material under P 1,2
Kbar pressure. This the final compaction stage. During the compaction the temperature is
reduced below 800
0
C which ensures the maintenance of nanostructure in the product.
Metallographic and phase analysis were conducted with scanning electron microscope (SEM)
and x-ray microanalyzer (EPMA). In order to determine phase and elemental compositions of
the synthesized samples, SEM and EPMA studies were performed. The SEM/EPMA was
conducted on a Quanta200 (FEI, USA) scanning electron microscope equipped with INCA
(Oxford Instrument, UK) energy dispersive spectrometer (EDS). SEM images were captured
using secondary electron detector as well as backscattered electron detector operating in
compositional contrast mode. The results are presented on the microstructures and X-ray
diagrams.

Investigation results

On pictures #3 and #4 is presented the border of TiB
0,6
and TiAl phases by secondary and
scattering electrons. It is clear that there is no sharp border on micro level and we have smooth
transient zone from one phase to another. According to authors this provides firm connection
between the phases, also strength of the material under thermal and mechanical loadings.
On pictures #5 and #6 are presented the microstructure of TiAl samples. It is clear that on
nanostructure level this phase consists of relatively dark nano areas which corresponds to TiAl
and TiAl
2
phases and light areas which corresponds to TiAl
2
phase. But it must be mentioned
that average content corresponds to TiAl phase.
On pictures #7 and #8 is presented the border of Ti
3
Al and TiB
0,6
phases by secondary and
scattering electrons. It is clear that there is no sharp border on micro level and we have smooth
transient zone from one phase to another. According to authors this provides firm connection
between the phases, also strength of the material under thermal and mechanical loadings.

Picture 3. Border shown with scattered electrons
between TiB0,6 and TiAl phases
.
point1 Al-23,22 weight %; Ti-76,78 weight .%; Point 2: Al-23,89w weight .%; Ti-76,11 weight .%
Point 3: Al-39,47 weight .%; Ti-60,53 weight .% Point 4: Al-48,92 weight .%; Ti-51,08 weight .%
Picture 5. Microstructure of the samples.
Picture 4. The border between TiB0,6 and
TiAl phases with secondary electrons

Al-34,37weight %; Ti-65,63 weight .%
Picture 6. Microstructure of TiAl samples.

Picture 7. Border shown with scattered electrons
between Ti3Al and TiB0,6 phases

Point 1:Al-6,31 weight %; Ti-93,69 weight %; Point 2: Al-8,84 weight.%; Ti-91,16weight %;
Point 3: Al-60,75 weight %; Ti-39,25 weight.%; Point 4: Al-62,44 weight.%; Ti-37,56 weight.%;
Picture 9. Microstructore of Ti3Al samples.

Picture 8. Border between TiB0,6 and TiAl
phases with secondary electrons

Al-35,8 weight.%; Ti-64,2 weight.%
Picture 10. Microstructure of Ti3Al samples.

Picture11. Microstructure of Ti3Al samples.

On pictures 9, 10, 11 is presented the microstructure of Ti
3
Al, it is shown that on nanostructure
level this phase consists of dark nano areas reach with Ti and the ratio wit Al is between 8:1
and 6:1 and also consists of light areas which is relevant to TiAl
3
. It must be mentioned that the
average content corresponds to Ti
3
Al phase.

Conclusion

In TiAl and Ti
3
Al cases during SHS in thermal explosion mode when the synthesis is provoked
by energy of
10
B containing high exothermic chasm, the light and nonporous high-effective
functionally gradient materials are obtained. In this product one layer is TiAl or Ti
3
Al
nanostructured material with 70-90 nanometer grain size, and the other layer is
10
B containing

hard alloy.
These gradient materials can be successfully used in space technologies for protection from
radiation and meteorites.

References

1. .., .., .., ...

. .
- 8-9 2002 ,
., - , 2002 . 110-114.
2. . .., .., ...
. . 1998,
.85, 1, . 140-146
3. G. Varsghalomidze, Z. Aslamazashvili, G. Zakharov, A. Berner, G.Oniashvili Development
of Innovative SHS Technology Coupled with Thermal Explosion for Production of Single-
Phase Nanocrystalline Materials of Ti-Al System, Bulletin of the Georgian National
Academy of Sciences, vol.2, no.3, 2008

Euro PM2009 Sintering Atmospheres
Manuscript refereed by Dr Mikael Larsen, Aalborg University, Denmark

Atmosphere Effects on Sintering of Carbon-free Powder Compacts
from Different Ferrous Powders

Herbert Danninger, Vassilka Vassileva, Christian Gierl
Technische Universitt Wien, Institut fr Chemische Technologien und Analytik,
Getreidemarkt 9/164, A-1060 Wien/Vienna, Austria

Abstract: Powder compacts were prepared from plain iron powders atomized, sponge,
and carbonyl grades and Fe-Mo prealloyed grades and sintered at 1200C in flowing Ar,
N
2
, and H
2
, respectively. It showed that despite markedly different green density levels, the
sintered density did not differ too much between the materials and was almost independent
of the atmosphere. The impact energy, in contrast, was very much affected by the atmo-
sphere in case of atomized and sponge iron grades, sintering in Ar resulting in very low IE
levels, but not for carbonyl iron which yielded the highest values nor for Fe-Mo materials.
The reason was that excessive grain growth and resulting intergranular failure occurred with
the coarser Fe powder grades but not with carbonyl Fe despite its higher sintering activity
or with Fe-4Mo. This indicates that grain growth during isothermal sintering is rather ineffect-
ive compared to the excessive grain growth occurring in high purity Fe during the austenite-
ferrite transformation.
Keywords: PM ferrous materials, sintering atmosphere, grain growth, intergranular failure

Introduction
The sintering atmosphere is one of the most crucial parameters in preparation of powder
metallurgy products obtained by pressing and sintering [1, 2], both positive and negative
effects being possible depending on the right choice of the atmosphere. On one hand, the
atmosphere can contaminate the green bodies, which aspect is particularly critical with
oxygen-sensitive materials such as Ti, but also with reactive sintered steels such as Cr
and/or Mn alloyed grades; oxygen and particularly water vapour also are strongly
decarburizing agents. Generally, pressed green compacts are particularly sensitive to
reactions with the atmo-sphere, due to their large specific surface. On the other hand, well
selected and reasonably pure sintering atmospheres not only prevent undesirable reactions
but also remove impur-ities such as oxygen or carbon, resulting in products with desired
mechanical, electrical or magnetic properties.

In [3] it has been shown that during sintering of high purity iron powder compacts in some
atmospheres, excessive grain growth may occur, associated with expansion of the speci-
mens, which effect is clearly visible when sintering in the dilatometer. The reason for the
latter phenomenon is asymmetrical phase transformation, the expansion during the -
transformation in the cooling section of the sintering run being markedly more pronounced
than the shrinkage caused by the - change in the heating section. These effects have been
observed very early in wrought iron [4], and the mechanism has been intensely discussed
e.g. by Kuroki [5, 6]. Both phenomena have been observed to occur during sintering in inert
atmospheres such as Ar or vacuum, resulting in pronounced intergranular embrittlement,
while being absent in N
2
containing atmospheres such as N
2
or N
2
-H
2
[7]. In plain H
2
, grain
growth was observed, but in this case it was not linked to embrittlement.

In the present work, the effect of both the sintering atmosphere and the starting powder
grade was studied, different plain iron powder grades as well as Fe-Mo prealloyed steel
powder being employed.

Experimental
The starting powders used were standard atomized and sponge iron powder grades ASC
100.29 and SC 100.26, respectively, both from Hgans AB, Sweden. Furthermore, carbonyl
iron powder (BASF grade CN) was employed as well as steel powders Fe-1.5%Mo (Astaloy
Mo, Hgans AB) and Fe-4%Mo (MSP3.5Mo, QMP). The powders were mixed with 0.5
mass% of lubricant (Microwax C) and compacted in a tool with floating die to impact test bars
55 x 10 x 8 mm
3
(ISO 5754). The pressure was uniformly 600 MPa.

All specimens were dewaxed in flowing high purity nitrogen at 600C. Sintering was done for
60 min isothermally at 1200C, at which temperature in the Fe-4%Mo grade some activat-
ion but not full sintering could be expected. For sintering in H
2
, a cold-wall pushtype
furnace with Mo heating elements, Degussa type Baby was used, while sintering in N
2
and
Ar was done in a hot-wall furnace heated by SiC rods and equipped with a gas-tight super-
alloy retort (Kanthal APM), to safely prevent contaminants from the air outside entering the
furnace. The sintered specimens were then characterized using the common techniques, i.e.
sintered density through water displacement, dimensional change, impact energy and
apparent hardness. Metallographic and fractographic studies were done to reveal the as-
sintered grain size and especially the fracture mode in impact loading.

Dimensional and mechanical properties
The properties obtained after sintering in different atmospheres are listed in Table 1 and
selectively for dimensional change and impact energy shown graphically in Fig.1. %
parallel specimens were tested each; the standard deviation levels for the impact energy
were <15 rel% except for the very brittle specimens (I.E. < 10 J.cm
-2
)

Table 1: Properties of carbon-free specimens prepared from different ferrous powders and
sintered in different atmospheres. Compacted at 600 MPa, sintered 60 min at 1200C

Powder
grade
Sintering
Atmosphere
Green
Density
g.cm
-3

Sintered
density
g.cm
-3

Dim.
Change
% linear
Impact
energy
J.cm
-2

Apparent
Hardness
HV10
ASC100.29 Ar 7.06 7.14 -0.33 3 55
H
2
7.04 7.15 -0.38 46 53
N
2
7.08 7.12 -0.09 15 62
SC100.26 Ar 6.98 7.07 -0.24 2 56
H
2
6.98 7.09 -0.28 40 57
N
2
6.98 7.06 -0.05 6 69
Astaloy Mo Ar 7.01 7.06 -0.24 46 59
H
2
7.01 7.09 -0.25 48 60
N
2
7.00 7.04 -0.16 33 75
MSP3.5Mo Ar 6.81 7.19 -1.72 71 80
H
2
6.81 7.25 -1.96 97 81
N
2
6.80 7.21 -1.03 56 86
Carbonyl Ar 6.10 6.91 -3.25 96 61
BASF CN H
2
6.20 6.88 -3.03 90 52
N
2
6.25 6.91 -2.99 79 68

As can be clearly seen, there is no expansion during sintering but shrinkage for all materials,
which supports Kurokis explanation [6] that the swelling observed in dilatometric runs e.g. for
vacuum and Ar [3] is a consequence of temperature gradients in the dilatometer, resulting in
gradual phase transformation. Here, also sintering in Ar of plain iron compacts from atomized
powders did not result in expansion, unlike in dilatometric sintering [3, 7], but in even more
shrinkage than in N
2
. Of course, the most pronounced shrinkage was observed for the
compacts from Carbonyl Fe, due to the low green density and high sintering activity. Also
MSP3.5Mo revealed significant shrinkage; this material resulted in the highest sintered
density levels, as a consequence of at least partial -sintering (full -sintering should result in
density levels >7.5 g.cm
-3
). Generally, the density levels attained under the sintering
conditions selected here are not too different despite the significant differences in powder
morphology and chemistry.

Fig.1: Dimensional change and impact energy of specimens as in Table 1

With regard to the impact energy data, much larger differences are visible. As described also
in [7], compacts from atomized iron powder fail in a very brittle manner when sintered in Ar,
while sintering in H
2
results in quite reasonable IE data. N
2
is not too effective, too, since the
reduction effects caused by H
2
are not effected here.

Sponge iron powder results in very similar IE data as does the atomized grade, while
carbonyl Fe yields very high impact energy in all atmospheres, indicating the ductile failure
that can be expected from plain iron. The Mo alloyed steel grades are also highly ductile, the
difference between these materials and those prepared from plain iron powders with similar
morphology being particularly visible after sintering in Ar.

The hardness levels are fairly similar for all the materials, only specimens prepared from
powder Fe-4%Mo exhibiting slightly higher values, as a consequence of the solid solution
hardening by Mo. Generally, sintering in N
2
slightly increases the hardness compared to Ar
or H
2
, which can be attributed to the slight nitrogen pickup that has been shown by thermo-
analytical studies to occur also in high purity N
2
atmosphere [8].

Microstructures and fracture surfaces

Fig.2a: Fe atomized

Fig.2b: Fe sponge

Fig.2c: Fe Carbonyl

Fig.2: Metallographic sections of plain iron and Fe-Mo compacts sintered in different atmo-
spheres. Compacted at 600 MPa, sintered 60 min at 1200C

The metallographic sections confirmed the grain growth phenomena occurring during sinter-
ing in inert atmosphere, as shown also in [3, 7]. However, this effect cannot be observed
after sintering in N
2
and with Mo alloyed steels, nor is it visible with specimens from carbonyl
iron. This is at first surprising: if the grain growth would occur during isothermal sintering
which is the standard temperature range in which grain growth would be expected the
highly active carbonyl iron (with its large specific surface) and Fe-4%Mo (with at least partial
sintering would be the first candidates for grain coarsening.

However, it is the high purity materials that exhibit the coarsest microstructure. This confirms
that the excessive grain growth occurs during the - transformation and is a consequence of
transformation within an extremely narrow temperature interval; here, only a few nuclei trig-
ger off the transformation, which results in large grains. In the case of further elements being
present, such as the interstitials N (or C, as shown in [3]) as well as metallic alloy elements
such as Mo (also Cr and Mn being effective [7]), the temperature range of transformation is
extended, and significantly more nuclei are generated, which results in the fine-grained
microstructure commonly attained during sintering of plain iron.

Abb.3a: Fe atomized, Ar

Abb.3b: Fe atomized, H
2

Abb.3c: Fe atomized, N
2

Abb.3d: Fe sponge, Ar

Abb.3e: Fe sponge, H
2

Abb.3f: Fe-1.5Mo, Ar

Fig.3a-f: Fracture surfaces of iron and Fe-Mo compacts sintered in different atmospheres.
Compacted at 600 MPa, sintered 60 min at 1200C


Abb.3g: Fe-4Mo, Ar

Abb.3h: MSP, H
2

Abb.3i: Fe Carbonyl, Ar

Fig.3g-i: Fracture surfaces of iron and Fe-Mo compacts sintered in different atmospheres.

The fracture surfaces shown in Fig.3 clearly indicate that the fracture modes observed can
be directly related to the impact energy data obtained. For plain iron from ASC 100.29, after
sintering in Ar pronounced intergranular failure is evident, as are the very large grains as
visible from the smooth fracture surfaces. After sintering in N
2
, in part intergranular fracture,
in part ductile rupture is visible, but the grains are significantly smaller than when sintering in
Ar. In hydrogen, finally, the classical ductile rupture as expected from sintered plain iron has
occurred.

Quite similar results have been obtained when using the sponge iron powder grade SC
100.26: also here, sintering in Ar results in consistently intergranular failure along the bound-
aries of the very large grains, while N
2
sintering yields mixed failure modes but with finer
grains, and the specimens sintered in H
2
show uniformly ductile rupture.

The alloyed steel grades exhibit a much lower effect of the atmosphere on the fracture mor-
phology, which is once more in agreement with the impact energy data: in all atmospheres,
also in Ar, ductile rupture has occurred, only in N
2
, also some intergranular failure being
observed. Carbonyl iron, finally, shows a very nice example of ductile rupture in all atmo-
spheres, also in the case of Ar no indication of intergranular failure being observed.

Chemical analysis
In order to obtain some information about the interstitial contents C, O, N, S - of the sin-
tered materials, specimens prepared from ASC 100.29 as the most representative material
and sintered in different atmospheres were analyzed using the standard combustion / car-
rier gas hot extraction (LECO) techniques. The data are given in Table 2, both as mass%
and as atomic% since for the relatively light interstitials mass% tend to indicate too low
levels.

The results clearly show that the carbon content is hardly affected by the atmosphere while
the oxygen level and also the S content are significantly lowered by sintering in H
2
. As
shown also in [8], sintering of plain Fe in N
2
results in at least detectable nitrogen pickup
during sintering while in the other atmosphere the N content was below the detection limit.
Generally, only oxygen (and, if present, nitrogen) are those interstitials that are contained in
levels that can be expected to affect the mechanical behaviour while both the carbon and
sulphur contents can be regarded as insignificant.

Table 2: Interstitial content of plain Fe from ASC 100.29 sintered in different atmospheres.

Sintering
Atmosphere
C content
Mass% / at%
O content
Mass% / at%
N content
Mass% / at%
S content
Mass% / at%
Ar 0.0045 / 0.021 0.0943 / 0.329 << 0.0081 / 0.014
H
2
0.0053 / 0.025 0.0132 / 0.046 << 0.0031 / 0.005
N
2
0.0040 / 0.019 0.0897 / 0.313 0.0245 / 0.098 0.0086 / 0.015

The bulk chemical analysis thus gives clear information about the presence of the interstit-
ials. In principle, intergranular enrichment of interstitials can be detected from the fracture
surfaces by Auger electron spectroscopy (AES). However, the significance of AES studies is
very limited in cases as the one presented here, since the reference material which here
would be the material sintered in H
2
simply does not expose the grain boundaries during
fracture, and thus, detection of interstitials in the fracture surfaces of specimens sintered in
Ar is not a definitive proof that these are not present also in the ductile materials since they
are not accessible to AES [9].

Summary
It has been shown that sintering of plain iron prepared from atomized and sponge iron pow-
ders respectively, results in grain coarsening and embrittlement when sintering in inert atmo-
spheres such as Ar, while both N
2
and H
2
prevent these effects at least partially, as does Mo
as alloy element. Mo alloyed steels exhibit relatively fine-grained microstructures and ductile
rupture, as do specimens prepared from carbonyl iron.

The reasons for the different mechanical behaviour caused by the atmospheres can be de-
rived from the different grain coarsening effects occurring during cooling, and on the other
hand from the interstitial contents: if pronounced coarsening has occurred, as in the case of
Ar sintering of ASC and SC based compacts, the remaining oxygen tends to concentrate on
the relatively small grain boundary area. When sintering in N
2
, the oxygen content is not
markedly lower, but the microstructure is finer and the grain boundary area is much larger,
resulting in less O concentration there. Sintering in H
2
, finally, lowers the oxygen content to
very low levels, resulting in clean grain boundaries, and grain boundary decohesion does not
occur even if grain coarsening has taken place during sintering.

Acknowledgement
This work was financially supported by the Austrian Fonds zur Frderung der wissenschaft-
lichen Forschung (Project no. 14889).

References
[1] F.V.Lenel: Powder Metallurgy Principles & Applications. MPIF, Princeton NJ (1980)
[2] A.Salak: Ferrous Powder Metallurgy. Cambridge Int. Sci. Publ., Cambridge UK (1995)
[3] H.Danninger: Powder Metall. Progress 3 No.2 (2003) 75
[4] P.Lehr: C.R.Acad. Sci. of France 242 (1956) 1172
[5] H.Kuroki, Y.Tokunaga; H.Y.Suzuki: Proc. PM2004 Vienna, H.Danninger, R.Ratzi eds.,
EPMA, Shrewsbury (2004) Vol.3, 17
[6] H.Kuroki, H.Y.Suzuki: Mat. Trans. (Japan) 47 No.10 (2006) 2449
[7] M.Jaliliziyaeian et al.: Proc. EuroPM2007 Toulouse, EPMA, Shrewsbury (2007) Vol.3, 131
[8] C.Gierl, H.Danninger : Proc. EuroPM2005 Prague, EPMA, Shrewsbury (2005) Vol.3, 3
[9] H.Danninger, A. Atari, B.Lux, E.Kny, A.Tschulik : Fresenius Z. Anal. Chemie 1989, 333,
417


Advanced Carbon Control in Sintering Atmospheres

Christoph Laumen*, Akin Malas*, Sren Wiberg** and Sigurd Berg***,
*Linde Gas Application Development, **AGA Gas AB, ***Hgans AB

ABSTRACT

As the major driver for the Powder Metallurgy (PM) industry is the automotive industry, the sintering
process now has to be more cost effective than ever. One way to success here is to combine multiple
processes and to control the processes by newer technological development as well as the deployment of
processes with less energy and lower consumption of resources.
The quality of PM parts is directly related to the control of process parameters of the individual production
steps. The metallurgical properties and especially the microstructural changes in the case of a sintering
component can only be regulated by controlling the furnace atmosphere accurately. The carbon control
of the sintering process has however always been somewhat out of control since no suitable control
concept was available. It is therefore a primary objective of this paper to give detailed results of
successful control of carbon potential in sintering processes and leveraging the advantage of the control
to produce higher quality product and at the same time to facilitate a lower production cost.
INTRODUCTION

Chromium has become more popular as an alloying element in the sintering industry in recent years
because of its advantages in increasing hardenability and resistance to softening during tempering. Nickel
is a more common alloying element but it is starting to be substituted by chromium because of its
significantly higher price. One of the main reasons why this shift is slow is that chromium has properties
that affect its performance in the sintering process, particularly its high affinity for oxygen.
In addition, the powder metal (PM) sintering industry is shifting towards sinter hardening in which two
processes are combined in an effort to achieve faster, cheaper and more reliable production.

High quality sinter production should translate into a process in which all parameters are controlled to
achieve all the product specifications. A failure to control the furnace atmosphere would jeopardize the
total quality control. Nevertheless, it has been common practice not to control furnace atmosphere.
Sintering of chromium alloys has now increased the need to control not only the carbon potential (C-
Potential) but also the oxygen partial pressure. A patented furnace atmosphere control system,
SINTERFLEX
, for controlling the C-Potential in sintering furnaces specifically designed through

cooperation between Hgans AB, and Linde Gas is described in this paper. Details include the results of
an experimental study.
The second most important part of the sinter hardening process is to harden the alloys that have been
sintered under a C-Potential controlled atmosphere without damage to their properties, completing the
sinter production. The relevant Linde Gas technology FRIOFLEX is still under development and is
reviewed in the final chapter of this paper.

FURNACE ATMOSPHERES IN THE SINTERING PROCESS

As with all furnace atmospheres used in the heat treatment of ferrous and non-ferrous metals, the type
and function of the furnace atmosphere is determined by its constituents. An atmosphere can be neutral,
carburizing, decarburizing, oxidizing or reducing, each of which can be a required function for different
purposes in the manufacturing industry. In a sintering furnace however different functions are expected in
different zones of the furnace, making sintering furnace atmospheres more challenging in terms of control
and optimization.

Surface decarburization is the most common problem associated with the use of present furnace
atmospheres for chromium alloys such as Astaloy CrM. The composition of Astaloy CrM is given in Table-
1 below.

Table-1: The composition of the component prepared using Astaloy CrM in weight percent
Alloy C Cr Mo Fe
Astaloy CrM 0.45 3.0 0.5 Bal.

The decarburization problem is therefore that the use of PM sintered parts in manufacturing is restricted
because of their increased sensitivity to fatigue cracks
*4 & *3
. Malas et al
*1
discussed the problems
associated with existing furnace atmospheres, including endothermic atmospheres and nitrogen-
hydrogen mixtures.

Increasing the C-potential at sintering temperature by adding fixed flows of hydrocarbons to conventional
atmosphere types has been shown to be impractical, as well as producing a high carbon potential at
lower temperatures, leading to sooting and excessive carbide formation. This experimental study and
development have therefore focused on a new gas mixture as the furnace atmosphere and a new
measurement method to control carbon in the mixture. The real problem for C-Control in fact is the lack of
a satisfactory on-line system for measuring the exact conditions inside the furnace. Therefore monitoring
of, and intervention in, the furnace atmosphere for any undesired condition is rather slow and too late in
the process timeframe. Current industrial practice in monitoring furnace atmospheres gives insufficient
indication of the carbon potential. A new approach is needed to address the issue.

DEVELOPMENT OF A NEW SINTERING CONTROL SYSTEM: SINTERFLEX

1. What is SINTERFLEX

SINTERFLEX

is a new and patented C-Potential control technology developed by Linde Gas and
Hgans AB for sintering furnaces. The system uses the simple gas sampling principle of the
sintering furnace and a newly designed oxygen probe and internal analysis furnace arrangement to
give exact C-Potential readings in the sintering zone of the furnace. The system controls the gas
composition, particularly its carbon enrichment, to maintain a healthy C-Potential to produce sintered
products of chromium alloys such as Astaloy CrM that are free of decarburization.

2. EXPERIMENTAL PROCEDURE

Knowledge of the oxygen partial pressure in the sintering zone tells the operator the actual
conditions during sintering. It can be measured by an oxygen sensor, which determines whether the
processed metal is oxidized or whether a metal oxide is reduced
*2
.

The development study has focused on two areas: a new furnace atmosphere to introduce carbon
into the sintering zone that can be controllable and responds to any changes; and a new
measurement and monitoring system with an oxygen probe to control this carbon containing
atmosphere. The new furnace atmospheres used are mainly mixtures of CO, hydrogen and nitrogen
with addition of propane, in proportions controlled by the new sintering atmosphere control system
SINTERFLEX
.

2.1 INITIAL TESTS
First tests were carried out in a lab unit consisting of a heating chamber to generate the sintering
furnace temperatures.
Synthetic gas mixtures similar to sintering furnace atmospheres were introduced into the system
through the flange arrangement at the top of the probe unit. This work focused on the control of
oxygen via an oxygen probe. The maximum limit of pO
2
is 5.5xE
-18
for Astaloy CrM. A new zirconia
probe was developed that could measure such very low oxygen partial pressures and meet C-
potential control requirements. This probe worked well for low levels of oxygen measurements in the
lab scale furnace in the first set of trials. It was then decided to demonstrate that the system could

work in an industrial furnace up to 1160 C (2100 F). This temperature is limited only by the
maximum heating capability of the internal heating element inside the heating chamber of the
analyzer.

Direct addition of CO to the furnace atmosphere and CO produced by the decomposition of
hydrocarbons such as methane and propane provided the required conditions to control the carbon
content via reaction [1] and this CO value is measured by a CO analyzer.

CO = C + O
2
[1]

2.2 FURTHER DEVELOPMENTS
A new approach was needed to adapt the idea for industrial scale furnaces leading to further
developments of the probe in the more convenient and compact design.
The newly designed system is shown in Figure-1, with the oxygen probe and its Al
2
O
3
cover
surrounded by a small electrically heated chamber, which would be heated to sintering temperatures.
Gas sampled from the sintering zone passes through the encapsulated area where it interacts with
the zirconium probe. The gas sample is also analyzed for its CO and H
2
content. These readings are
used in the C-potential calculations. The probe readings are continuously monitored and recorded.
The system can be set to the C-Potential required by the process and the SINTERFLEX system
then maintains this set value by means of regulating the flow of the enrichment gases.

Figure-1: Schematic illustration of SINTERFLEX ACS mini.

The complete SINTERFLEX
system as shown in Figure-2 comprises an analyzer and a

FLOWTRAIN
allowing operators to control the sintering furnace atmosphere completely.


Figure-2: SINTERFLEX
System

The oxygen sensor works on the same principle as a regular oxygen probe to make its measurements but
within a small scale furnace. The sensor chamber would be filled with fresh furnace atmosphere providing
an on-line measurement of oxygen in an mV- signal. The sample gas also provides on-line analysis of
the furnace atmosphere. Input data is processed by a EUROTHERM 3504 controller and the C-Potential
is calculated as an on-line measurement. Operators can control the furnace atmosphere parameters and
correct any undesired condition immediately it occurs.

ATMOSPHERE CONSIDERATIONS AND RESULTS OF FINAL TESTS

1. Measurements during the recent tests have shown that the C-Potential could be regulated with
propane (C
3
H
8
) added to the carrier gas as the controlling parameter. The trials were performed in
Hgans ABs pilot scale furnace.
2. As stated above, sintering furnace atmospheres consist of mixtures of nitrogen, hydrogen, carbon
monoxide and propane. Trials compared results with atmospheres with and without hydrogen. The
trials aimed to provide an atmosphere with a neutral C-Potential to the parts at sintering temperatures
thereby ensuring that carburization would be achieved during cooling. The importance of preventing
decarburization is evident especially when considering the dynamic properties as shown in figure 6.
Decarburization causes residual tensile stress at the surface. In figure 6 the residual stress obtained is
150MPa giving a bending fatigue strength of 200 MPa. A carburized surface removes tensile stresses
and creates residual compressive stresses. The positive carbon gradient gives bending fatigue
strengths of 450 to 520MPa, depending on the achieved stress levels.

Figure 3 a. CCT diagram surface core microstructure b. Residual stress effect on bending fatigue
strength.
Core
LB
Tensile residual
stress
Time
Temp
Ms
decarburized surface core
Ms
0
100
200
300
400
500
600
-150 -100 -50 0 50 100 150 200
Residual stress [MPa]
S
t
r
e
s
s

A
m
p
l
i
t
u
d
e

[
M
P
a
]
50%, PBF, R=-1

3. The carbon potential setting of the furnace atmosphere is decided from calculations of the amount of
oxygen and CO in the furnace. The carbon content of the components will differ depending on the
chemical composition of the material being sintered. Their carbon content can be calculated by using
Gunnarssons formula [2] and the carbon potential of the atmosphere.

Log Cp / C = 0.055 (% Si) 0.013 (% Mn) 0.040 (% Cr) + 0.014 (% Ni) 0.013 (% Mo) C. [2]

The new oxygen probe successfully maintained the required atmosphere conditions and C-Potential.
Figure-4 shows a 3-hour measurement window of a test where the C-potential was set to 0.38 at 2% H
2

atmosphere.

Figure 4 Record of SINTERFLEX-controlled atmosphere parameters over a 3h production period

4. In a further investigation, an atmosphere containing N
2
- 2% H
2
- 2% CO and additions of C
3
H
8
, was
tested to carburize the surface to 0.45% carbon where the carbon content in the centre is 0.38%.
The results showed a positive carburizing gradient approximately ~150m thick for CrM material.
The same material however showed no carbon profile in a dry atmosphere of N
2
10% H
2.
Figure- 5
illustrates a comparison of these two studies. It has also been observed that at a constant
atmosphere mixture, the carbon potential increases as the temperature falls.

Astaloy CrM treated with CO atmosphere Astaloy CrM treated with N2/H2 Atmosphere
Figure-5: Comparison of CrM Microstructures in different atmosphere conditions

EFFICIENT COOLING FOR SINTER HARDENING
Malas et al *1 have achieved cooling rates of 2.5 4.5 C/sec in an industrial furnace where many
operational and constructional limitations affected the results, including the impingement time, distance
and flow rates. It was therefore decided to go one step back and undertake pilot tests to optimize the
process for industrial application. Linde Gas FRIOFLEX technology, which is being developed for sinter
hardening furnaces, is now under full scale test as this paper is being published.

The recently updated system now consists of a series of top and bottom nozzles that will allow the parts
to cool down until they reach the temperature where transformations stop at the hard structures required
by the specifications. The system will be available as a separate or a retrofit unit.

CONCLUSIONS
The newly developed oxygen probe offers the following advantages:

1. Use of and quantitative control of carbon neutral or carburizing atmospheres to avoid surface
decarburization giving higher quality parts and increased product performance.
2. Potential to replace a subsequent case hardening operation by means of tailored surface carbon
gradients during sintering, leading to lower total cost and better dimensional tolerances.
3. Provision of valuable information on the quality of sintering along with traceability through the optional
process data storage system which will help avoid scrap and increase overall equipment
effectiveness.
4. Any disturbances to the process and potential atmosphere problems (gas flow/-rate and draughts) are
detected quickly and adjusted easily through on-line control which will again reduce the costs of lost
production and scrap.

REFERENCES
1. A New approach to sintering furnace atmosphere control and sinter hardening by gas impingement
cooling, Akin Malas*, Sren Wiberg*, Daniel Nilsson**, and Sigurd Berg**, *Linde Gas Application
Development, **Hgans AB, PM2008 World Congress - Washington D.C.
2. Considerations when Sintering Oxidation Sensitive PM Steels, Bjorn Lindqvist, Koki Kanno, Hgans
AB, Sweden, Hgans Japan K.K., Japan
3. Influence of decarburized layers and surface defects on the stress state of a specimen strained in
bending. A. G. Rakhshtadt, V. M. Semenov, S. M. Serebrin, I. E. Stepanov and N. N. Shaposhnikov.
Scientific-Research Institute for Automobile and Tractor Materials, Moscow. Translated from
Problemy Prochnosti, No. 5, pp. 6872, May, 1974.
4. Impact of Decarburization on the Fatigue Life of Powder Metal-Forged Connecting Rods, Document
Number: 2001-01-0403, Date Published: March 2001, Author(s): Edmond Ilia - Mascotech Sintered
Components, Russell A. Chernenkoff - Ford Research Laboratory.

Analysis of Oxide Reduction during Sintering of Cr-alloyed Steel
Powder through Photoacoustic Spectroscopy Measurements

Ola Bergman, Hgans AB, SE-263 83 Hgans, Sweden
Karin Frisk, Swerea KIMAB AB, SE-102 16 Stockholm, Sweden
Lars Nyborg, Chalmers University of Technology, SE-412 96 Gteborg, Sweden

Abstract
The usage of steel powders prealloyed with chromium in PM components for high
performance applications is steadily increasing. Excellent mechanical properties are
achieved with these powder grades, provided that the processing into components is
accurately done. Photoacoustic spectroscopy (PAS) technique has been used here to study
oxide reduction during sintering in 90N
2
/10H
2
atmosphere of test specimens based on 3%Cr-
0.5%Mo prealloyed powders with 0.5% C added. Results from performed reduction trials
show that Fe oxides on the powder surfaces are removed by H
2
reduction early in the
heating stage of the process. Carbothermal reduction is the main mechanism for reduction of
stable Cr-containing oxides from 900C and above. Efficient Cr oxide reduction requires a
sintering temperature of at least around 1200C. Therefore, small oxide particles remain in
the material after sintering at 1120C for 30 minutes. However, such oxides have earlier
been shown to have no significant influence on mechanical properties.

Introduction
The development of Cr-alloyed steel powder grades for structural PM components has been
driven by the high and volatile prices in recent years of the commonly used alloying elements
Mo and Ni. Water-atomized powders prealloyed with 1.5-3% Cr and 0.2-0.5% Mo have now
been available for some years and using these materials is a cost effective way to reach the
mechanical properties needed for high performance applications [1]. Well-controlled sintering
is a prerequisite for successful processing of Cr-alloyed powders due to the oxidation
sensitivity of Cr. Hence, sintering atmospheres with low oxygen content must be used in
order to have reducing conditions [2]. Furthermore, an appropriate temperature cycle should
be applied to ensure that sufficient reduction of oxides on the powder surfaces occurs in the
process. Components with high mechanical performance can be obtained after sintering at a
temperature (1120C) typically used in conventional mesh belt production furnaces, but to
exploit the full potential of the Cr-alloyed materials higher sintering temperatures must be
used [3]. The aim of the investigation presented in this article was to study oxide reduction
mechanisms during sintering of Cr-alloyed steel powder and, thereby, gain better knowledge
on how to optimize a sintering process for this type of material.

Surface oxides on the Cr-alloyed steel powder grades have a heterogeneous structure
consisting of Cr-rich particulate compounds (size up to 0.2 m) surrounded by a thin (6-7 nm)
continuous Fe oxide layer [4]. The possibility to reduce these oxides is determined by their
thermodynamic stability, which depends on alloy composition of the steel, temperature, and
oxygen pressure in the surrounding atmosphere. Figure 1 shows oxide phases appearing at
thermodynamic equilibrium in a Fe-3%Cr alloy for different temperatures and fixed oxygen
partial pressure. Accordingly, Cr
2
O
3
is the most stable oxide phase in such an alloy, followed
by the spinel oxide FeCr
2
O
4
, and relatively high temperatures are required to have reducing
conditions for these oxides at the given oxygen pressure. The iron oxide phases (FeO, Fe
3
O
4
,
Fe
2
O
3
) are less thermodynamically stable and may be reduced at lower temperatures.


Figure 1. Mole fraction of phases versus temperature for Fe-3%Cr alloy in equilibrium with
atmosphere where p
O2
= 10
-19
atm (from Thermo-Calc).

Besides favourable thermodynamic conditions, reduction media have to be available for
removal of the oxides. In the sintering process, metal oxide (M
x
O
y
) reduction occurs either
through interaction with H
2
in the atmosphere according to reaction (1), or via carbothermal
reduction through interaction with C in the material according to reactions (2) and (3).

M
x
O
y
(s) + yH
2
(g) xM(s) + yH
2
O(g) (1)

M
x
O
y
(s) + yC(g) xM(s) + yCO(g) (2)

M
x
O
y
(s) + yCO(g) xM(s) + yCO
2
(g) (3)

The carbothermal reduction starts in accordance with reaction (2) when the temperature is
high enough to activate the carbon in the powder compact. However, as soon as CO
becomes available reaction (3) will be the dominating reduction mechanism, and
regeneration of CO is ensured at high temperatures by the Boudouard reaction:

C(s) + CO
2
(g) 2CO(g) (4)

The reactions above show that oxide reduction during sintering may be studied by monitoring
concentrations of the gaseous reaction species H
2
O, CO and CO
2
in the process atmosphere.
In earlier investigations of oxide reduction in Cr-alloyed PM grades, mass spectrometry and
dilatometry have been used to study deoxidation during slow heating of compacts in vacuum
or He atmosphere [5-7]. In the present investigation, PAS technique for high precision gas
analysis was applied in slow heating as well as simulated sintering trials in N
2
/H
2
atmosphere
with pressed specimens based on Cr-alloyed steel powder.

Experimental procedure
The water-atomized powder grade Astaloy CrM, which is prealloyed with 3% Cr and 0.5%
Mo, was used as test material. Oxygen content of the utilized powder sample was 0.15%.
This powder was mixed with 0.5% graphite (Kropfmhl UF4) and 0.6% lubricant (Kenolube
P11). Cylindrical test specimens (diameter 25 mm, height 20 mm) were produced by uniaxial
compaction of the powder mix with a compaction pressure of 600 MPa. Dewaxing was done
by heat treating the test specimens for 30 minutes at 700C in 75H
2
/25N
2
atmosphere.

Density (g/cm
3
) C (%) O (%)
6.94 0.50 0.20
Table 1. Test specimen properties after dewaxing.
An experimental set-up consisting of a tube furnace connected to a PAS instrument for gas
analysis was used for oxide reduction trials. Three different reduction trials were performed
with one specimen in each trial (see Table 2). A very low heating rate was applied in the first
trial, while faster heating was used in the other two trials in order to simulate realistic
sintering cycles. In all trials, a constant flow of high purity (dew point ~ -60C) 90N
2
/10H
2

atmosphere was running through the furnace tube (inner diameter 35 mm) at a rate of 5 litres
per minute. Concentrations (>1 ppm) of H
2
O, CO, CO
2
and CH
4
in the outgoing gas from the
furnace were analysed by the PAS instrument throughout the experiments. The time
between sample collections by the PAS instrument was about 60 seconds.

Trial # Temperature cycle Atmosphere
1 Slow heating (5C/min) to 1300C + 10 min isotherm 90N
2
/10H
2

2 Simulated sintering cycle: 1120C for 30 min 90N
2
/10H
2

3 Simulated sintering cycle: 1250C for 30 min 90N
2
/10H
2

Table 2. Temperature cycles applied in the oxide reduction trials. Sintering temperature was
reached after about 68 minutes in both simulated sintering cycles.

After the reduction trials, cross sections of the test specimens were examined by light optical
microscopy (LOM). Bulk analyses of oxygen and carbon contents were done in LECO
instruments.

Results
Results from the PAS analysis of gas concentrations in the reduction trials are presented in
Figure 2. The CH
4
concentration was low (<10 ppm) throughout all three trials and is not
included in the graphs.

In the slow heating trial, there is an H
2
O peak starting at around 300C and with a maximum
value at 370C, which corresponds to H
2
reduction of easily reduced iron oxides. After this
peak has declined the H
2
O concentration stays below 10 ppm the rest of the trial. The CO
concentration starts to increase at about 700C, which marks the onset of carbothermal
reduction of more stable oxides. There is a small CO peak just below 800C and at around
850C the CO concentration increases sharply to form a large peak at 1230C. This peak
also has a shoulder at about 1030C. The CO
2
concentration increases steadily from about
850C to form a peak at around 1230C, in conjunction with the much larger CO peak.

The H
2
O concentration curves from the two simulated sintering trials are almost identical.
There is a single peak with maximum value at about 550C, corresponding to hydrogen
reduction of iron oxides, where after the H
2
O concentration slowly decreases from 20 ppm to
10 ppm at the end of the trial. A small CO peak appears in both trials in connection with the
H
2
O peak. In the 1120C sintering trial, CO formation due to carbothermal oxide reduction
starts at around 750C and there is a relatively small CO peak at 815C. After that, the CO
concentration increases rapidly and peaks at 1010C followed by a second peak as the
sintering temperature 1120C is reached. The CO content decreases continuously during the
hold time at 1120C but the concentration is still relatively high (~190 ppm) at the end of the
isotherm. Also in the 1250C sintering trial, CO formation due to carbothermal reduction
appears at around 750C. The CO concentration increases slowly until 900C after which the
concentration increases rapidly to a peak value at 1160C, with a peak shoulder at around
1010C. When the sintering temperature 1250C is reached, the CO content starts to
decrease rapidly reaching a relatively low level (~50 ppm) at the end of the isotherm.
Increase in CO
2
concentration starts in both sintering trials between 800C and 850C. In the
1120C trial, CO
2
increases steadily until the sintering temperature is reached, whereupon
the concentration stays more or less constant at ~70 ppm during the isotherm. In the 1250C
trial, CO
2
increases until 1160C where the concentration levels out at ~85 ppm until 1220C,
and then the concentration decreases slowly until the end of the isotherm.

Trial 1: Slow heating to 1300C.

Trial 2: Sintering at 1120C for 30 min. Trial 3: Sintering at 1250C for 30 min.
Figure 2. Results from gas analysis by PAS in oxide reduction trials on test specimens of
Astaloy CrM + 0.5% C in 90N
2
/10H
2
atmosphere.

Bulk analyses of oxygen contents in the test specimens show that the material contains
0.099% O after the 1120C sintering trial and 0.014% O after the 1250C sintering trial. This
large difference in oxygen content is also visible in the microstructures of the specimens (see
Figure 3). There are small oxide particles (up to a few m in size) throughout the structure of
the 1120C sintered material, both along former powder particle boundaries and inside the
former powder particles, while the structure of the 1250C sintered material is very clean and
only contains few traces of oxides. The specimen from the slow heating trial has similar
structure as the 1250C sintered material and also low bulk oxygen content (0.021% O).

Trial 2: Sintering at 1120C for 30 min. Trial 3: Sintering at 1250C for 30 min.
Figure 3. As-polished microstructures in test specimens of Astaloy CrM + 0.5% C after oxide
reduction trials in 90N
2
/10H
2
atmosphere. Oxides in the structures are marked by arrows.

The test specimen from the 1120C sintering trial has highest carbon content (0.43% C) and
lowest density (7.02 g/cm
3
). The specimens from the other two trials are comparable in both
carbon content and density, with 0.36% C and 7.08 g/cm
3
for the 1250C sintered material,
plus 0.38% C and 7.09 g/cm
3
for the material from the slow heating trial.

Discussion
The thermodynamic calculation results presented in Figure 1 give a stability ranking for oxide
phases in the investigated material. This stability ranking may be used to determine which
oxides the different reduction peaks correspond to in the gas concentration curves from the
reduction trials. However, agreement between the results in Figure 1 and the experiments
cannot be expected regarding reduction temperatures, since the oxygen partial pressure is
not constant in the reduction trials. In a process gas with fixed dew point, p
O2
will increase
with temperature in accordance with the gas equilibrium reaction H
2
+ O
2
= H
2
O. The N
2
/H
2

atmosphere used in the trials has a H
2
O level of ~10 ppm, which means that p
O2
is below the
value (10
-19
atm) used in the calculations up to temperatures over 1250C.

In the slow heating trial, the H
2
O peak at 370C should represent reduction of the easily
reduced iron oxides Fe
2
O
3
and Fe
3
O
4
. The small CO peak at around 800C should
correspond to FeO reduction, which is supported by earlier reports on the necessary
temperature for carbothermal reduction of this oxide [8]. Remaining is reduction of the Cr-
containing oxides, which is represented by the large CO peak starting at about 850C. The
peak shoulder at 1030C is attributed to transition from Fe-Cr oxide reduction to Cr oxide
reduction. Reduction maximum at 1230C for the most stable oxide

is in line with results from
previous studies on oxide reduction of Cr-alloyed steel powder compacts [5,6].

Corresponding reduction peaks to those in the slow heating trial are found in the two
simulated sintering trials. Accordingly, hydrogen reduction of Fe
2
O
3
/Fe
3
O
4
occurs in the initial
heating stage, represented by the H
2
O peaks at around 550C. The carbothermal reduction
is initiated at about 750C and there is some FeO reduction before reduction of the Cr-
containing oxides starts at around 900C, as marked by the sharp increases in CO
concentration at that temperature. The change in the CO curves just above 1000C indicates
that the Fe-Cr oxide reduction declines and that the Cr oxide

reduction starts, although the
decrease in heating rate at that temperature should also influence. In the 1120C sintering
trial, there is a drop off in CO concentration when the sintering temperature is reached,
showing that the Cr oxide

reduction progresses rather slowly during the isotherm. In the
1250C sintering trial, the peak in CO concentration at 1160C should be a consequence of
the decrease in heating rate at that temperature. However, there are still high CO
concentrations up to the sintering temperature indicating effective Cr oxide

reduction, and the
rapid decline in CO during the isotherm should be due to that most oxides are reduced at this
point.

The fact that efficient oxide reduction is achieved in the 1250C sintering trial is verified by
the low oxygen content in the material afterwards and the basically oxide free microstructure
of the sintered specimen. The specimen from the 1120C sintering trial, on the other hand,
has about ~0.1% residual oxygen and many small oxide particles are visible in the
microstructure. These should be Cr-rich oxides that originate either from particulate
compounds on the powder surfaces or from small oxides inside the powder particles. Hence,
the kinetics for Cr oxide

reduction is not good enough at 1120C. As the slow heating trial
shows, the Cr oxide

reduction becomes much more efficient at temperatures around 1200C
and above. It has also been demonstrated in a previous investigation that nearly complete
oxide reduction is achieved when sintering Cr-alloyed PM steels at 1200C [3].

Although some oxides remain in the Cr-alloyed material after sintering at 1120C, the
microstructure demonstrates that there are well-developed sinter necks between powder
particles. This is ensured by the removal of the continuous Fe oxide layer on the powder
particles early in the heating stage of the sintering process, as shown in the reduction trials.
Consequently, the Cr oxide particles do not have any significant effect on the bonding
between particles during sintering, and it is clear that high mechanical performance is
attainable with the Cr-alloyed powder grades after 1120C sintering [1]. However, by
sintering at temperatures where complete oxide reduction occurs, the mechanical properties
of Cr-alloyed PM steel can be boosted to even higher levels [1,3].

It is evident from the reduction trials that carbothermal reduction, represented by the large
CO peaks, is the main mechanism for oxide reduction in the sintering process. Hydrogen-
assisted reduction is detected during heating, but there is no H
2
O formation above around
700C. Yet, the increases in CO
2
concentration obtained in the trials at higher temperatures
might be an indication of oxide reduction by H
2
, since any H
2
O formed according to reaction
(1) is likely to react with CO and form CO
2
according to the water-gas shift reaction:

CO(g) + H
2
O(g) CO
2
(g) + H
2
(g) (5)

The test specimens used in the investigation were somewhat oxidized in the dewaxing
process prior to the reduction trials. In spite of this, the specimen characteristics after the
simulated sintering trials are comparable to what is usually obtained in corresponding
sinterings of Cr-alloyed powder grades. Hence, this oxidation did not have any significant
influence on the overall reduction processes in the performed experiments.

Conclusions
Conclusions regarding sintering of pressed compacts (density ~7.0 g/cm
3
), based on 3%Cr-
0.5%Mo prealloyed steel powder with 0.5% C added, in 90N
2
/10H
2
atmosphere:
Iron oxides on the powder surfaces are reduced by H
2
early in the heating stage of the
process. Carbothermal reduction by C in the material starts at around 750C, and is the
dominating mechanism for removal of stable Cr-containing oxides at higher temperatures.
Cr oxide reduction is initiated at around 1000C, but for efficient reduction of this oxide a
temperature of about 1200C or above is required.
Small oxide

particles, typically of m size, remain in the material after sintering at 1120C
for 30 minutes due to insufficient kinetics for Cr oxide

reduction. Such particles are not
detrimental for the mechanical properties of the sintered component.
Sintering at 1250C for 30 minutes leads to practically complete reduction of oxides,
leaving only traces of oxides in the material.

References
[1] U. Engstrm, D. Milligan, A. Klekovkin: Advances in Powder Metallurgy & Particulate
Materials, 2006, Vol. 2, Part 7, pp. 21-32.
[2] O. Bergman: Powder Metallurgy, 2007, Vol. 50, No. 3, pp. 243-249.
[3] O. Bergman, B. Lindqvist, S. Bengtsson: Materials Science Forum, 2007, Vols. 534-
536, pp. 545-548.
[4] H. Karlsson, L. Nyborg, S. Berg: Powder Metallurgy, 2005, Vol. 48, pp. 51-58.
[5] H. Danninger, C. Gierl, S. Kremel, G. Leitner, K. Jaenicke-Roessler, Y. Yu: Powder
Metallurgy Progress, 2002, Vol. 2, pp. 125-140.
[6] S. Kremel, H. Danninger, Y. Yu: Powder Metallurgy Progress, 2002, Vol. 2, pp. 211-
221.
[7] H. Danninger, C. Xu: Proc. of Euro PM2003, Valencia, Spain, October 2003, EPMA,
Vol. 1, pp. 269-274.
[8] S. C. Mitchell, A. Cias: Powder Metallurgy Progress, 2004, Vol. 4, pp. 132-142.

Studies of Surface Oxides on Steel Powders using Photo Acoustic
Spectroscopy coupled with Thermodynamic Calculations

Karin Frisk, Hans Sderberg, Sophie Caddeo

Swerea KIMAB AB, P.O. Box 55970, SE-102 16 Stockholm, Sweden

The properties of PM steel are affected by the oxygen content of the material. The surface
oxides on the powder grains are an important source of oxygen in steel produced from
powder, and it is therefore interesting to investigate under what conditions the surface oxides
are formed, and can be reduced. In the present work Photo Acoustic Spectroscopy (PAS)
studies were used to study surface oxides on powders. The measurements were coupled
with thermodynamic calculations, to identify the type of oxides. The use of the PAS
instrument is a novel technique for gas analysis at ppm levels under controlled atmospheres
and heat treatments. Examples are shown from studies of oxide reduction, furnace
atmosphere control, and moisture uptake. Sintered steel, stainless steel, and low alloyed
steels were investigated. The accuracy and reproducibility of the method for different
applications is assessed.

Introduction

The oxygen content of powder metallurgical steels has an important limiting effect on
mechanical properties, and it is desirable to keep oxygen contents low. A significant amount
of the total oxygen content of a compacted material derives from adsorbed oxygen, moisture,
or oxides on the surface of the powder grains. To limit the oxygen content it is therefore
important to understand how to avoid oxygen uptake on the surface of powder grains, and to
investigate methods to reduce surface oxides.
In the present work experimental studies using a photo acoustic spectrometer (PAS)
are used to analyse surface reactions on steel powder. The analysis is performed by
exposing steel powder to a reducing atmosphere, at varying temperatures, while the
composition of the gas is registered. When oxygen is reduced from the surface of the powder
grains, the reaction products that are released to the gas are detected with high accuracy
using the photo acoustic spectrometer.
The temperature for reduction of different oxides in a specified gas can be determined
by this method. The reduction temperature of an oxide is given by the thermodynamic
stability of the oxide. By coupling the results of PAS measurements with thermodynamic
calculations it is possible to identify the type of oxide that is reduced at a specific
temperature. The combined analysis makes it possible to determine the types of oxides that
are present on the surface of the powder grains, and the temperatures that are required to
reduce these oxides. An indication of the amount of reduced oxides can also be obtained.

Method Description

Photo Acoustic Spectroscopy (PAS) is a highly sensitive technique by which ppm
levels of specific gas species within a carrier gas can be measured. In this work an Innova
1314 instrument from Lumasense Technologies has been used. In Fig. 1 a schematic view of
the experimental setup is shown. A carrier gas flows through the tube furnace and on the
exhaust side the PAS unit is connected through a junction allowing it to sample the exhaust
gas.


Fig. 1. Schematic view of the experimental setup (not to scale). The carrier gas flows from
the left to right over the sample. Intermittently the exhaust gas is sampled by the PAS.

Once the gas has entered the PAS unit it is sealed within a compartment in which the
gas is illuminated by pulses of infrared light. The gas molecules absorb certain wavelengths
corresponding to the resonant vibration frequency of the gas molecule. Upon relaxation a
pressure wave will result which in turn is picked up by two highly sensitive microphones in
the test chamber. To optimize this process several optical filters are mounted inside which
are used to filter the infrared light so that only the wavelength necessary for excitation of a
specific molecule passes. PAS is an absolute zero-level technique, i.e. if no water molecules
exist then no water signal can be detected, in addition, the measured signal is directly
proportional to the amount of the measured gas species within the carrier gas. The PAS unit
is calibrated to measure the following gases and concentrations, CO (0.2-2000 ppm), CO
2

(1.5-2000 ppm), CH
4
(0.4-2000 ppm), H
2
O (0.1-1000 ppm), and dew point. Nitrogen, 90/10 %
N
2
/H
2
, and hydrogen are usually used as carrier gases.
When measuring the mentioned gases the sampling time is slightly below 1 min.
Decreasing the number of gases will reduce the sampling time, down to a minimum of
approximately 30 s.
Prior to analysis the tube furnace and the Al
2
O
3
crucible is degassed. This treatment
consists of a 3 h heating to 1300 C in a flow of hydrogen. After degassing the sample is
loaded and thereafter, before the heating commences, the furnace is further flushed in order
to achieve a low water signal background (~10-15 ppm).
The PAS is a stable instrument producing reproducible results. Figure 2 shows a
comparison of two different PAS analyses. The measurement was performed on powder
samples. Typically 10-20 g of powder is placed in an Al
2
O
3
crucible, however other sample
forms and geometries have also been examined throughout the years, e.g. solids and cold
pressed powder compacts. It is of vital importance to obtain a powder sample that is
representative of the fractions available in the as-received powder since e.g. finer fractions
contain more surface oxides compared to coarser fractions for a specific weight. As a
consequence, sample splitters are used. The results shown in Fig. 2 were registered on two
different powder samples. The composition of the powder was the same, but they were taken
from different batches. The three species measured during the two trials are more or less
overlapping, the slight displacement that is seen is due to the sample time. The result of
Fig. 2 validates the reproducibility of the method.
The thermodynamic calculations were performed using a dedicated thermodynamic
database for oxides. The calculations were set up by specifying the composition of the alloy,
the composition of the reducing gas including the oxygen content. From the calculations the
onset temperature for reduction of the various stable oxides can be evaluated. The rate of
reduction, that often is time limiting in the case of oxides, can not be obtained from these
calculations.
Good indications of the types of oxides that can be expected on the surface of powder
for a specified alloy composition are obtained. The composition of the respective surface
oxide is determined by the conditions under which they were formed (atmosphere,
temperature). Some surface oxides originate from the atomisation process (high
temperatures), and others from handling of the powder (low temperatures). Calculations of
the reduction temperatures should take the true composition of the oxides into account.
However, the conditions of formation of the oxides are not known, and in the present work
we therefore approximated the compositions. The compositions of the oxides were given the
stable composition at the reduction temperature, which is believed to be a reasonably correct
approach, although this gives an uncertainty in the reduction temperatures.

0
0 20 40 60 80 100 120
Time [a.u.]
C
o
n
c
e
n
t
r
a
t
i
o
n

[
p
p
m
]
-200
T
e
m
p
e
r
a
t
u
r
e

[
a
.
u
.
]
Species 1
Species 2
Species 3
Temperature

Fig. 2. Comparison of two PAS measurements on identical steel powder, but from different
batches, showing the reproducibility of the method. Three species were registered. The slight
displacement seen is due to the sampling time.

Reduction of surface oxides on sintered steel

Examples of the output from PAS measurements on Astaloy CrM + 0.5%C steel in
90/10 N
2
/H
2
are presented by Bergman et al. [1]. Different heating rates and sintering
temperatures were applied. The PAS measurements clearly show that by increasing the
sintering temperature from 1120 to 1250
o
C a more complete reduction of oxides is
accomplished. As discussed in [1] it is possible to compare the registration from PAS with
thermodynamic calculations to identify the type of oxide that is reduced in different stages of
the sintering. By this method it was shown that a more complete reduction of Cr
2
O
3
oxide is
obtained by sintering at 1250
o
C than at 1120
o
C.
However, when a heating rate is applied in the PAS measurement, the comparison with
thermodynamic calculations is difficult since the rate of oxide reduction, which is a kinetic
effect, is not predicted by the calculations. To eliminate this uncertainty, a stepwise heating
was used. In the measurements presented by Bergman et al. samples of Astaloy CrM with
0.5%C additions were analysed. For a more theoretical evaluation of the oxide stability
powder without carbon additions were reduced in H
2
gas, see Fig. 3. The reducing
atmosphere is 99.995 % pure. In this gas the oxide reduction is registered through the H
2
O
concentration of the gas.
The result shown in Fig. 3 indicates that the reduction proceeds rapidly as the
temperature is increased, and is nearly completed after the 30 minutes holding time at
constant temperature for each step. By comparing the results with thermodynamic
calculations we can conclude that the high peak at 1200
o
C mainly consists of Cr
2
O
3
. The
reduction of this oxide is nearly completed during 30 minutes holding time at 1200
o
C. At the
lower temperatures the spinel phase, (Cr,Fe)
3
O
4
, is reduced.
The amount of oxide that is reduced at each temperature can be evaluated from the
area under the PAS curves at each temperature. By this method we see that the largest
amount of oxide is of Cr
2
O
3
which is reduced at high temperatures.
From the PAS curves the total oxygen content that was released during the heat
treatment was evaluated to 0.14%O. The chemical analysis before and after PAS
measurement showed a difference of 0.137%O. The amount of reduced oxygen evaluated
from the PAS measurements is close to the value from chemical analysis, and can be used
with confidence.

0
10
20
30
40
50
60
70
0 50 100 150 200
Time [min]
T
e
m
p
.
/
2
0

[
C
]

/

H
2
O

c
o
n
c
.

[
p
p
m
]
107 ppm
900 C
1000 C
1100 C
1200 C
1250 C

Fig. 3. A PAS registration of the oxide reduction of Astaloy CrM powder in H
2
in a stepwise
heating cycle. The water content in the gas is registered at the different temperature steps.

Atmosphere control during sintering

The effect of varying the sintering atmosphere of Astaloy CrM + 0.6%C samples was
measured using PAS. The results are shown in Figs. 4 and 5. In Fig. 4 a 90/10 N
2
/H
2

atmosphere was used. The registered curves are similar to those presented in [1] on the
same material. The low temperature reduction of Fe-oxides is shown by the water peak. At
higher temperatures CO and CO
2
are released as the high temperature oxides are reduced
by a reaction involving C from the sample.
In Fig. 5 the same sintering is shown, with a different sintering atmosphere. In this case
an addition of 2.5%CO was made to the sintering temperature, substituting a part of the H
2

with CO. The effect of such a change can be understood by comparing the PAS curves. The
addition of CO to the sintering gas leads to a change in the reaction products. At low
temperatures, the Fe-oxides are reduced by hydrogen, producing water, but CO also
contributes, which is seen by the CO
2
peak at low temperatures. The reduction of oxides at
higher temperatures is shown by a small water peak and a large CO
2
peak. The result
indicates that less C from the sample will be consumed to reduce oxides. This was also
confirmed by the chemical analysis of the samples (0.48%C after sintering in 90/10 N
2
/H
2
gas
and 0.54%C after sintering in 90N
2
/7.5H
2
/2.5CO gas).

Powder handling

Incorrect handling of powder can deteriorate the properties by oxygen contamination.
The low detection limits and the high accuracy of the PAS measurements make it possible to
quantify the effect of powder handling on the surface oxides.
Figs. 6 and 7 show two examples from such studies. In Fig. 6 results are shown for
three samples of a stainless steel powder previously exposed to a humid atmosphere at
20
o
C (65% relative humidity). The released oxygen from the three different samples was
registered by PAS, as shown in Fig. 6. The initial temperature of the powder samples was 15
or 20
o
C. The lowest moisture uptake was found for the powder sample where the
temperature of the powder was 20
o
C. The samples that were 15
o
C at the start of the
exposure were more contaminated, independent of the time for exposure (1 hour or 24
hours). It should be noted that the effect is clear, despite of the low contents of detected
water.

0
50
100
150
200
250
0 20 40 60 80 100
Time [min]
C
o
n
c
e
n
t
r
a
t
i
o
n

[
p
p
m
]

0
100
200
300
400
500
600
700
800
900
1000
1100
1200
1300
T
e
m
p
e
r
a
t
u
r
e

[
C
]
H
2
O
CH
4
Temperature
CO divided by 10
CO
2

Fig. 4. PAS registration of a simulated sintering of Astaloy CrM+0.6%C in a 90/10 N
2
/H
2
gas.
0
50
100
150
200
250
0 20 40 60 80 100
Time [min]
C
o
n
c
e
n
t
r
a
t
i
o
n

[
%
,

p
p
m
]
0
100
200
300
400
500
600
700
800
900
1000
1100
1200
1300
T
e
m
p
e
r
a
t
u
r
e

[
C
]
H
2
O [ppm]
CO
2
[ppm]
CH
4
[ppm] CO multiplied by 10 [%]
Temperature

Fig. 5. PAS registration of a simulated sintering of Astaloy CrM+0.6%C in a
90N
2
/7.5H
2
/2.5CO gas.

Fig. 7 shows a similar study on a low alloyed steel. The PAS measurement was
conducted up to high temperatures. This study shows that only low temperature oxides are
affected by moisture uptake. The climate chamber treated powders show a significantly
higher water peak at low temperatures compared to the as received powder. The reduction
peaks from higher temperatures overlap for the three samples. The high temperature oxides
originate from the atomisation process, and are not significantly affected by an incorrect
powder handling at room temperature.

Conclusions

PAS measurements have been applied to study the type of oxides on sinter steel
powders, and their reduction temperatures. Combined with calculations the method has
aided in the understanding of what types of oxides that are present on the surface of the
powders, and of the effects governing the oxide reduction during sintering.
The effects of changes of sintering gas have been studied using PAS, showing the
effect of atmosphere on carbon control during sintering. This is a powerful method that gives
information on the reactions between powder and atmosphere in the course of the sintering.
Incorrect handling of powders affects the low temperature oxides. PAS measurements
can accurately characterize such effects. Small effects can be detected with such accuracy
that different handling procedures can be evaluated.

0
5
10
15
20
25
30
0 25 50 75 100 125 150
Time [min]
C
o
n
c
e
n
t
r
a
t
i
o
n

[
p
p
m
]
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
T
e
m
p
e
r
a
t
u
r
e

[
C
]
H
2
O (T
p
20, t24)
H
2
O (T
p
15, t1)
H
2
O (T
p
15, t24 )
Temperature

Fig. 6. Comparison between three climate chamber treated stainless steel powders. Powder
temperature was 15 or 20
o
C (T
p
15 or T
p
20) and the exposure time was 1h (t1) and 24 h (t24)
respectively. Heat treated in pure hydrogen gas.
0
20
40
60
80
100
120
140
160
180
200
0 30 60 90 120 150 180 210 240 270
Time [min]
C
o
n
c
e
n
t
r
a
t
i
o
n

[
p
p
m
]
0
200
400
600
800
1000
1200
1400
T
e
m
p
e
r
a
t
u
r
e

[
C
]
CH
4
(A, t16, t30)
H
2
O(t30)
H
2
O(t16)
H
2
O(A)
CH
4
(t30)
CH
4
(t16)
CH
4
(A)
CO (A, t16, t30)
Temperature

Fig. 7. Comparison between as-received (A) and climate chamber treated steel powder, 16 h
(t16) and 30 h (t30) respectively. Heat treated in pure hydrogen gas.

References

[1] Ola Bergman, Karin Frisk and Lars Nyborg, Analysis of oxide reduction during sintering
of Cr-alloyed steel powder through photoacoustic spectroscopy measurements,
Proceedings of the Euro PM2009 conference, Kpenhamn, 2009.

Manuscript refereed by Dr Mikael Larsen, Aalborg University, Sweden

Analysis of Delubrication from Iron Powder Compacts Based on a
Direct Observation of Decomposing Lubricant during Sintering

Takashi KAWANO, Tomoshige ONO and Yukiko OZAKI

JFE Steel Corporation
Steel Research Laboratory
1 Kawasaki-cho, Chuo-ku, Chiba, 260-0835, JAPAN

ABSTRACT

Delubrication behavior of iron powder compacts containing zinc stearate (ZnSt) or ethylene-
bisstearamide (EBS) was directly observed during sintering. First, small droplets of lubricant
melt appeared scattered on the surface. Then, they gradually formed large pools, and finally
vaporized. In addition, ZnSt melt more remarkably bubbled before vaporization than EBS
melt. After sintering the compact containing ZnSt in nitrogen, spot stains of rough surface
and concentrated zinc with the diameter of 1 mm were found on the top of their surface. It is
probable that the lubricant melt inhomogeneously flew out through the easier paths in the
compacts to the surface accompanied by the exhausting gas of the decomposed lubricants.

1. INTRODUCTION

Organic compounds such as zinc stearate (ZnSt) or ethylenebisstearamide (EBS) are
practically used in order to improve flowability of iron powder, increase green density and
suppress ejection force in compaction process. Though the lubricants are almost eliminated
during sintering, they sometimes generate soot or stains on the surface of the sintered
compacts, which deteriorate the products and result in the yield loss.
Many studies have revealed the decomposing mechanism of ZnSt and EBS [1-3]. As for the
delublication, German has analyzed the debinding from MIM products [4]. Nayar et. al. has
summarized the typical defects on the sintered iron powder compacts and proposed
countermeasures on production [5]. Kameoka et. al. have analyzed delubrication resulting in
the rough surface in the endothermic gas [6]. However, more precise study on delubrication
is required in order to improve the quality of the product by reducing the surface defects.
In addition to the above studies, visualizing the phenomena greatly helps understand what
really occurs particularly in the powder metallurgy. Kondo et. al. have visually captured the
powder flowing into the cavity and proposed a new method for uniform filling [6].
In this study, delubrication from iron powder compacts was directly observed during sintering
in the furnace. Also, the critical factors for avoiding the stains on the sintered products are
proposed based on the delubrication model derived from the morphological and elemental
analysis of the stains.

2. EXPERIMENTAL PROCEDURE

2.1 Raw materials

Water-atomized steel powder (JFE Steel Corporation, JIP-300A) was mixed or segregation-
free treated with 2 mass% of atomized Cu powder, 0.8 mass% of a natural graphite powder
and 0.5 mass% of zinc stearate (ZnSt) powder or ethylenebisstearamide powder (EBS).

2.2 Processing conditions

The mixed powders were compacted into the cylinder shape with the diameter of 25 mm, the
thickness of 5 mm or 25 mm and the green density of 7.0 and 7.2 Mg/m3. The green
compacts were delubricated at 600 C for 20 min, and then sintered at 1130 C for 20 min in
various atmospheres; endothermic gas (RX gas), mixed gas composed of 80 vol % nitrogen
and 20 vol% hydrogen (N2-H2 gas), and pure nitrogen (N2 gas). Surface defects on the
specimen were analyzed. Also, delubrication from the compacts were directly observed
utilizing electric furnace with silica glass windows, heating at 25 C/min up to 1130 C
without delubrication step, where the atmosphere was pure nitrogen in order to suppress
delubrication intensively.

2.3 Property evaluation methods

Melting and decomposing temperature of the lubricants were identified by thermal analysis in
He gas. Morphological shape of the stains on the sintered compacts was analyzed with
optical microscope and scanning electron microscope (SEM). Spot X-ray diffraction analysis
(XRD) was used to identify the crystallographic phase of the stains. In addition, electron
prove micro analysis (EPMA) was applied to obtain the elemental distribution around the
stains.

3. RESULTS

3.1 Effect of sintering atmosphere on surface
stains

Green compacts of 25 mm thickness and 7.0
Mg/m3 density using powders mixed with ZnSt or
EBS were sintered. Photographs of the sintered
compacts surface are shown in Fig.1. The
surfaces of the sintered compacts having
contained EBS were clean independent of the
sintering atmosphere. On the other hand, small
amount of black stains were observed on those
having contained ZnSt and sintered in RX gas or
N2-H2 gas. Also, black fine particles like carbon
powder were found on those having contained
ZnSt and sintered in N2 gas.

3.2 Effect of lubricants on delubrication behavior

Delubrication behavior of the green compacts of 5 mm thickness and 7.0 Mg/m3 density are
shown in Fig.2 Lubricant melt started leaching above 300 C on the green compacts
containing ZnSt through mixing. This melt gradually gathered and formed larger droplets on
heating. Also, small droplets of several mm in diameter were observed flowing on the side
Fig.1 The photographs of the sintered compacts
surface dependent on the sintering atmosphere
and mixed lubricant.
RX
80%N
2
-20%H
2
N
2
ZnSt EBS Lubricant
Fig.1 The photographs of the sintered compacts
surface dependent on the sintering atmosphere
and mixed lubricant.
RX
80%N
2
-20%H
2
N
2
ZnSt EBS Lubricant
Fig.2 Directly obeserved delubrication behavior of the compacts containing ZnSt or EBS.
Mixed
Segregation
-free treated
250
o
C 300
o
C 350
o
C 400
o
C 425
o
C
EBS
ZnSt
Temperature
Fig.2 Directly obeserved delubrication behavior of the compacts containing ZnSt or EBS.
Mixed
Segregation
-free treated
250
o
C 300
o
C 350
o
C 400
o
C 425
o
C
EBS
ZnSt
Temperature
Mixed
Segregation
-free treated
250
o
C 300
o
C 350
o
C 400
o
C 425
o
C
EBS
ZnSt
Temperature
surface down to the setting plate above 350 C. When the temperature was higher than 400
C, the lubricant melt started bubbling. The bubbling was observed both on the top surface
and at the bottom edge except on the side surface of the compacts at specific points
(nucleus). It was weakened and the lubricant melt gradually disappeared above 450 C.
Almost the same phenomenon as above
result was observed for the green
compacts containing ZnSt added by
segregation-free treatment. However,
starting temperature of leaching and
bubbling of the lubricant melt dropped by
about 20 C respectively.
In case of the green compacts containing
EBS, leaching and bubbling of the
lubricant melt started at temperature 50
C lower and the degree of bubbling was
weaker than that observed for the
specimen containing ZnSt.

Melting point of ZnSt and EBS was 125
C and 148 C respectively as shown in Fig. 3. Decomposition temperature was ranged from
400 C to 450 C for ZnSt, and from 350 C to 400 C for EBS. The observed temperature
where the lubricant melt started bubbling coincided with that of decomposition derived by the
thermal analysis.

3.3 Effect of thickness and density of the green compacts on the delubrication behavior

Delubrication behavior of the green compacts containing ZnSt with the thickness of 25 mm
and the density of 7.0 or 7.2 Mg/m3 are shown in Fig.4 Leaching and bubbling started at the
same temperature as that for the specimen with the thickness of 5 mm. However, the
amount of the lubricant melt pooled at the top surface was larger and the bubbling of the
melt was more remarkable compared to those of the
thinner specimen. In addition, grayish white ash was
found on the whole surface of the specimen at 800
C, which gradually shrunk as the temperature rose
and finally disappeared at 1130 C. On the other
hand, no ash was observed in case of the green
compacts containing EBS. Therefore, it is probable
that the ash was the residue of decomposed ZnSt.
No difference was found in the delubrication
behaviors comparing the specimen with the green
density of 7.0 and that of 7.2 Mg/m3.

-40
-20
0
20
40
0 200 400 600 800 1000
D
T
A
ZnSt
EB S
melting
decomposition
Temperature (
o
C)
D
T
A
(
V
)
-40
-20
0
20
40
0 200 400 600 800 1000
D
T
A
ZnSt
EB S
melting
decomposition
Temperature (
o
C)
D
T
A
(
V
)
Fig.3 Result of the the differntial thermal analysis in
He gas for ZnSt and EBS.
EBS
ZnSt
-40
-20
0
20
40
0 200 400 600 800 1000
D
T
A
ZnSt
EB S
melting
decomposition
Temperature (
o
C)
D
T
A
(
V
)
-40
-20
0
20
40
0 200 400 600 800 1000
D
T
A
ZnSt
EB S
melting
decomposition
Temperature (
o
C)
D
T
A
(
V
)
Fig.3 Result of the the differntial thermal analysis in
He gas for ZnSt and EBS.
EBS
ZnSt
Fig.4 Effect of the density of the green compacts with 25 mm thicknessson the delubrication behavior of ZnSt.
250
o
C 300
o
C 350
o
C 800
o
C
(Side View of the Specimen)
7.0 Mg/m
3
7.2 Mg/m
3
Green Density
mixed segregation-free
treated
Temperature
Fig.4 Effect of the density of the green compacts with 25 mm thicknessson the delubrication behavior of ZnSt.
250
o
C 300
o
C 350
o
C 800
o
C
7.0 Mg/m
3
7.2 Mg/m
3
Green Density
treated
Temperature 250
o
C 300
o
C 350
o
C 800
o
C
7.0 Mg/m
3
7.2 Mg/m
3
Green Density
treated
Temperature
Fig.5 Effect of thickness and lubricant
on the stains on the top surface of the
compact sintered in N
2
gas without
holding at 600
o
C for debinding.
5 mm
25 mm
Scale
ZnSt EBS Lubricant
Thickness
Fig.5 Effect of thickness and lubricant
on the stains on the top surface of the
compact sintered in N
2
gas without
holding at 600
o
C for debinding.
5 mm
25 mm
Scale
ZnSt EBS Lubricant
Thickness
5 mm
25 mm
Scale
ZnSt EBS Lubricant
Thickness
3.4 Characterization of the stains on the specimen sintered in N2 gas

The top views of the sintered specimens with the
green density of 7.0 Mg/m3 are shown in Fig.5.
There was no stain found of the specimen having
contained EBS. On the other hand, the sintered
compacts having contained ZnSt showed gray or
black stains adhered on the surface. Also, a gray
scale, which was easily peeled off, was found on the
top surface. Moreover, ring-shaped tarnished stains
with the diameter of about 1 mm, which could not be
rubbed off, were found under those scales on the
surface of the specimen as shown in Fig.6
SEM images of the stains are shown in Fig.7.
Spngeform structure was found at the center of the
ring stain, which was surrounded by three bands with
different morphology of 100 m width each. The
morphological shapes of the inner band, the center
band and the outer band were fine needle, bypiramid
and rectangle respectively. Crystallographic structure, identified by XRD, of the sponge form
was pure iron oxide, which decreased as the analysis point went from the inner ring to the
outer one, and was finally substituted by metallic iron at the outside of the stains. This
change in the degree of oxidation was also identified by the oxygen distribution analyzed by
EPMA as shown in Fig.8. In addition, Zn was concentrated at the center of the stain, which
suggests that ZnSt melt preferentially flew out
from this point.

4. DISCUSSION

4.1 The model of delubrication

Fig.9 indicates the model of delubrication and
stain formation for the compact of iron
powder mixed with lubricant, ZnSt. The solid
lubricant forms a liquid above the melting
point at the early stage of sintering. The
lubricant liquid starts leaching to the surface
of the compact as it swells on heating. Then,
when the temperature rises above
decomposition point of the lubricant, the melt
is pushed out of the compact preferentially
through the easiest path (open pore) because
of the high pressure of the vaporized
lubricant. Also it bubbles at this spouting point
in appropriate conditions. Though, the
lubricant melt almost vaporized by additional
heating, inorganic material including metal
and metal oxide is left concentrated around
the spouting point. Thus, this inorganic
residue reacted with the compact [6] can
result in the stains composed of rough
surface on the surface of the sintered
product.

Fig.6 Ring-shaped stains
found adhered on the top
surface of the sintered
compact under a scale shown
in Fig.5.
1mm
Fig.6 Ring-shaped stains
found adhered on the top
surface of the sintered
compact under a scale shown
in Fig.5.
1mm 1mm
Fig.7 SEM photographs of themorphological
shape of the stain shown in Fig.6
Bypiramid
Fine needle
Rectangle
50 m
0.2 mm
Spongeform
Fig.7 SEM photographs of themorphological
shape of the stain shown in Fig.6
Bypiramid
Fine needle
Rectangle
50 m
0.2 mm
Spongeform
Bypiramid
Fine needle
Rectangle
50 m
0.2 mm
Spongeform
Zn O
Fig.8 Distribution of zinc and oxygen at the stain
shown in Fig.6
Zn O Zn O
Fig.8 Distribution of zinc and oxygen at the stain
shown in Fig.6
4.2 The controlling factor of delubrication and stain formation

In the light of the above delubrication
model, it is important to focus on
three factors in the delubrication
process in order to avoid stains on
the sintered compacts. First, the
characteristics of the starting material
including lubricants, the iron powder,
and the other additives. Particularly,
decomposing temperature should be
lower in order to realize the complete
vaporization of the lubricant before
the impurities arisen from the
lubricant react with the iron powder to
form stains on the surface. Also,
impurities originated from the iron
powder and the other additives
should be get rid of. Second,
migration of the lubricant melt through the open pore to the surface. Therefore, thinner
compact with much open pore is preferable. Though the green density did not affect the
delubrication behavior in this study, it is expected that the local impurity segregation be
reduced if the green density of the compact is smaller than 7.0 Mg/m3. Finally, the reaction
among the lubricant, the compact and sintering atmosphere. In addition to composition, due
point, direction and mass flow of the atmospheric gas, impurities contained in the gas such
as sulfur or carbon must be controlled in order to avoid undesirable reaction with the
lubricant at the compacts surface during sintering.

5. CONCLUSIONS

Iron powder compacts containing ZnSt or EBS were sintered, where the delubrication
behavior was directly observed up to 1130 C. The delubrication model and the critical factor
for avoiding the stains on the sintered body have been proposed based on the
morphological and elemental analysis of those stains. The major conclusions of this study
are as follows:
(1) Stains on the sintered compacts having contained ZnSt were remarkable when
sintered in nitrogen. The ring shaped stain composed of rough surface with Zn segregation
at its center was formed.
(2) Leaching of the ZnSt melt to the surface of the green compact occurred at lower
temperature than EBS melt.
(3) Thicker compact generated larger pools of lubricant melt.
(4) Green density of 7.0 and 7.2 Mg/m3 did not affect the delubrication behavior.
(5) Segregation-free treatment lowered both leaching and bubbling temperature by 20
C.
(6) Controlling the characteristics of the starting material, the migration of the lubricant
melt and the reaction of the lubricant with atmospheric gas are crucial in order to avoid the
stains on the sintered compacts.

Fig.9 Model of delubrication and stain formation for the
compact of the iron powder mixed with the lubricant, ZnSt.
Dew point, PO
2
Gas flow
Surface
Iron powder
Zn segregation
ZnSt
ZnSt melt
Decomposed gas
Flowing out
through
open pores
Fig.9 Model of delubrication and stain formation for the
compact of the iron powder mixed with the lubricant, ZnSt.
Dew point, PO
2
Gas flow
Surface
Iron powder
Zn segregation
ZnSt
ZnSt melt
Decomposed gas
Flowing out
through
open pores
Dew point, PO
2
Gas flow
Surface
Iron powder
Zn segregation
ZnSt
ZnSt melt
Decomposed gas
Flowing out
through
open pores
REFERENCES

[1] C.F.Legzdins, I. V. Samarasekera and T. Troczinski: Proc. 1998 PM World Congress,
266(1998)
[2] J. Dwyer, H. Nayar, W. Gerristead and B. Wasicziko: Adv. Powder Metall. & Patticul.
Mater. 1992, 3-186(1992)
[3] G. White and H. Nayar: Adv. Powder Metall. & Particul. Mater. 1996, 3-27(1996)
[4] R. M. German; Inter. J. Powder Met., 23, 4 ,237(1987)
[5] H. S. Nayar,: Metals Handbook Ninth Edition (Vol.7 Powder Metallurgy), American
Society for Metals, Ohio,1984; 339-350.
[6] T. Kameoka, K. Ishihara, H. Hamamoto and M. Obayashi; Proc. Jpn. Soc. Powder and
Powder Metall., 78(1983)
[7] M. Kondo, S. Takemoto and I. Urata; J. Jpn. Soc. Powder and Powder Metall., 45,
5(1998)
Euro PM2009 Sintering Processes & Materials
Germany

Characterization of Cu-Al
2
o
3
Composites Manufactured by Powder
Metallurgy Method

I. Altnsoy, G. F. elebi Efe, T. Yener, M. Ipek, S. Zeytin, C. Bindal

Sakarya University, Engineering Faculty, Department of Metallurgy and Materials
Engineering, Esentepe Campus, 54187 Sakarya-Turkey

ABSTRACT

In the present study, we characterized copper matrix composite reinforced with alumina at ratios
of 1, 3, 5 and 7 % by weight manufactured by powder metallurgy method. Sintering process was
performed at 875, 925 and 975C for 2h in graphite powder. The relative densities of test
materials for sintered at 875, 925 and 975C were 95.4, 97.4, 97.4 and 97.2; 96.2, 98.6, 98.5 and
98.4; 94.7, 94.3, 93.8 and 92.7; for 1, 3, 5 and 7%, respectively. Optical and SEM studies
revealed that alumina particles were located around Cu grains. The presence of alumina and
copper were confirmed by EDS and XRD analysis. Depending on the amount of alumina the
electrical conductivities of test materials ranged from 84 to 40%IACS which lower than that of Cu
with a electrical conductivity of 98%IACS. The microhardness of test materials ranged from 107-
139 HV for composites with a particle size of 10 m copper. The microhardness of test materials
increased as amount of reinforcing materials increased.
Keywords: Powder metallurgy, alumina, Copper, Conductivity, Composite.
1. INTRODUCTION

Copper is the best electrical conductor metal after the silver in such applications which the
electrical conductivity being main demand (1). But, copper can not protect its mechanical
properties at high temperatures (2). Therefore, in order to increase mechanical properties of
copper for high temperature applications, ceramic particles such as SiC,Al
2
O
3
are added to
copper by producing metal matrix composites reinforced with ceramic particles (3). Alumina is
one of the most used materials which has high elastic modulus (390 GPa) and hardness (16
GPa), also protects its properties at high temperatures. Thus, in copper-alumina composites,
hardness of copper matrix reinforced with fine Al
2
O
3
particules increases by dispersion
hardening (4). On the other hand, increasing of amount of the reinforcement particles decreases
the electrical conductivity of composites. However, the electrical conductivity and hardness of the
composites can be adjusted by controlling of amount of reinforcement materials in the
composites (5).
Copper-alumina composites can be produced by different techniques such as casting, powder
metallurgy, inner oxidation of Cu-Al alloy and chemical precipitation followed by hydrogen
reduction (6). In the present study, Cu- Al
2
O
3
composites were manufactured by powder
metallurgy method. For characterization of mechanical and electrical properties of test materials,
XRD,SEM-EDS and Vickers Indentation Technique were utilized.
2. EXPERIMENTAL DETAILS
In this study, commercial copper powder with particle size of 10 m and Al
2
O
3
powder with the
particle size of 0.3 m were used as starting materials. Graphite was used as a protective
medium for obtaining oxide-free composites. The production route of test materials is given in
Figure 1.

Figure 1. Chematic presentation of production route of test materials.
Copper powder
(10 ve 40 m.)
Al
2
O
3
powder
(0.3 m.)

+
Mixing
Pressing (with a load of
50 kN)
Sintering (875,925 ve
975 C; 2h)

Hot pressing
(load:125kN)

Metallographical
studies(Grinding,
polishing,etching)
Characterization
(Optical, SEM-EDS,
XRD, density,
hardness, electrical
conductivity

3. RESULTS
The optical micrographs of the test materials were shown in Figure 2.
%wt,
Al
2
O
3
875C 925C 975C
1

3

5

7

Figure 2 Optical micrographs of Cu (10 m)-Al
2
O
3
composites sintered at different temperatures
for 2h (Magnification: 500X).

Figure 2 showed that the reinforcing particles are homogenously dispersed into the matrix and
take placed around the copper grains. White colored and the dominant phase indicates copper
matrix and black-grey colored areas indicate alumina as a reinforcement material.
The presence of phases in composites was confirmed by XRD analysis technique. The XRD
diffraction patterns were given in Figure 3. As it can be seen in Figure 3, test materials is
dominantly consisting of Cu, Al
2
O
3
with small amount of Cu
2
O.

% wt, Al
2
O
3

% 7
875

925

975

Figure 3 X-ray diffraction patterns of test materials sintered at different temperatures for 2h.

Temperatu
re(C)
% wt, Al
2
O
3

7
875

SEM micrograph No:1 No:2

No:3 No:4 No:5
925


No:3 No:4

975


No:3 No:4
Figure 4 SEM-EDS dot analysis of test materials sintered at different temperatures for 2h.

The SEM-EDS results showed that, white fields indicated copper matrix and black fiels indicated
Al
2
O
3
particules as a reinforcing material (Figure 4). The results confirm the optical micrographs
in which Al
2
O
3
particules take placed in the grain boundaries. Also, SEM-EDS results are in good
agreement with the XRD results. Although, SEM-EDS results show no copper-oxide (Cu
2
O) on
the copper grains indicated by white fields, composites contain oxygen according to EDS marks
obtained from dark areas. These oxygen peaks point out the Al
2
O
3
(dominantly) and Cu
2
O (small
amount) phases as verified with XRD results. Also, copper-oxide phases probably take place
between the surface of copper grains and alumina particles. Because, oxygen peaks were just
obtained from interfaces.
The relative densities, microhardness and electrical conductivities of test materials were given in
Table 1. As it can be seen in Table 1, with increasing of amount of reinforcing materials, the
microhardness of test materials increased, whereas the electrical conductivity decreased
because of the hard and insulator nature of Al
2
O
3.
There is no remarkable effect the sintering
temperatures and amount of reinforcement materials on the relative density of the composites.
Table 1 The relative densities, microhardness and electrical conductivities of test materials with a
matrix size of 10 m.

Cu-% wt Al
2
O
3

875 925 975
Relative Density (%)
1 95.39 96.21 94.72
3 97.44 98.61 94.31
5 97.41 98.47 93.77
7 97.20 98.42 92.70
Microhardness (HV)
1 107.6 109.3 107.7
3 125.8 128 129.2
5 131.2 132.8 132.7
7 137 139 137.3
Electrical Conductivity (% IACS)
1 83.60 84.91 82.76
3 73.71 68.43 63.79
5 61.21 59.91 51.72
7 43.10 50.86 43.10

4. DISCUSSION
The above results revealed that reinforcing ceramic based and insulator nature Al
2
O
3
particles
were located around grains of copper matrix which confirmed by OM and SEM examinations. The
presence of phases in composites confirmed by x-ray diffraction analysis technique revealed that
the dominant phases are Cu and Al
2
O
3
in addition to trace amount of Cu
2
O. It was found that
there is a good agreement between x-ray diffraction analysis and SEM-EDS results. SEM-EDS
analysis verified the location of oxygen at interface being between copper grain and alumina
particles. The relative densities of test materials for sintered at 875, 925 and 975C for 2 h in
embedded graphite powders were 95.4, 97.4, 97.4 and 97.2; 96.2, 98.6, 98.5 and 98.4; 94.7,
94.3, 93.8 and 92.7; for 1, 3, 5 and 7%, respectively. Depending on the amount of alumina the
electrical conductivities of test materials ranged from 84 to 40%IACS which lower than that of Cu
with a electrical conductivity of 98%IACS. The microhardness of test materials ranged from 107-
139 HV for composites with a particle size of 10 m copper. The microhardness of test materials
increased as amount of reinforcing materials increased.
5. CONCLUSIONS
Al
2
O
3
particles were located around grains of copper matrix. .Depending on the addition amount
of alumina the hardness of composites increased and electrical conductivity decreased due to
effect of lattice defects. As it is known-well lattice defects scatter electrons providing electrical
conductivity and thus reduce the mobility, mean free path of electrons and conductivity of metal.
There is no remarkable effect the sintering temperatures and amount of reinforcement materials
on the relative density of the composites.

ACKNOWLEDGEMENTS
Authors thank to Fuat Kayis for performing XRD and SEM studies and special appreciations are
extended to technician Ersan Demir for experimental assistance at Sakarya University.

REFERENCES
[1] www.bpc.edu/mathscience/chemistry/electrical_conductivity.html
[2] www.cda.org.uk
[3] www.magnet.fsu.edu/magtech/facilities/materials/highstrength.html
[4] D.W. Lee, B. K. Kim, Nanostructured Cu-Al
2
O
3
composite produced by thermochemical
process for elctrode application, Materials Letters,58 (2004), 378-383.
[5] P.K. Jena, E.A. Brocchi, M. S. Motta, In-situ formation of Cu-Al
2
O
3
nano-scale composites by
chemical routes and studies on their microstructures, Materials Science & Engineering A 313
(2001) 180-186.
[6] Hyun-Ki Kang, Microstructure and electrical conductivity of high volume Al
2
O
3
-reinforced
copper matrix composites produced by plasma spray, Surface and Coatings Technology,
190, 2005, 448-452.

France

The Development of Low Silica Containing Kiln Furniture for the
Sintering of Powder Metals

Dipl.-Ing. Jens Sperber*
Head of R&D and Application Technology
STEULER Industrieller Korrosionsschutz GmbH, Hhr-Grenzhausen, Germany

Dipl.-Ing. Martin Enders, Manager Metal Injection Moulding
SCHUNK Sintermetalltechnik GmbH, Heuchelheim, Germany

Abstract

The aim of the work is the development of low silica containing kiln furniture for the
sintering of powder metals. In the past a lot of different materials were used, in the meantime
based on cordierite containing ceramics due to the need of a high thermal shock resistance.
The second phase of this kind of material was more or less mullite, which can not really
decrease the amount of silica in the product. This kind of ceramic has meanwhile two
disadvantages. The first one is the silica content which causes contact reactions with the
metal powder parts in the sintering process with the result of surface defects on the sintered
parts, especially a decrease in the hardness of the surface. The second disadvantage is the
lower ability of positive reactions through the sintering process, especially in the case of
sintering parts which are produced by the MIM process. The reason for that was found in the
porosity itself and the pore size distribution.
The new developed kiln furniture based on corundum with a silica content of less than 5 %
has decreased the contact reaction down to zero. Due to the special composition the positive
reactions through the sinter process is also possible. And with the combination of corundum
and mullite or other phases which enhance the thermal shock resistance the new kiln
furniture can be used in all kinds of fast firing processes.

Contact reactions and hardness

Cordierite and mullite containing slabs are used for the sintering of powder metals.
Depending on the firing temperature the content of cordierite is varying between 10 and 60
%. For more thermal stability the content of cordierite should be decreased whereas the
content of mullite should be increased. For most of the sinter parts produced without the MIM
process cordierite based kiln furniture is suitable.
Depending on the firing temperature and the kind of kiln furniture especially in the MIM
process there is the possibility to get contact reactions between the sinter parts and the kiln
furniture. Some investigations have shown a relationship between the silica content of the
ceramic slab and the hardening problems of the surface of the sinter parts. Figure 1 and 2
are showing the surface contamination of a sinter part.


Figure 1: White surface layer low cordierite Figure 2: White surface layer high cordierite

Additionally to that the REM-EDX analysis of sinter parts is showing a silica distribution within
the sinter parts shown on the following figures 3-6.

Figure 3: Marked area for EDX analysis Figure 4: Result for area A

Figure 5: Result for area B Figure 6: Result for area C

Figure 3 shows the white surface layer. The upper part, abt. 60 % of the figure, shows this
layer and the different contamination with silica. We can say the sinter metal has absorbed
silica depending on the closeness to the surface.
So as a result we can see that silica reacts with the sinter parts. This happens only on the
surface of the parts in a layer which is app. 20 to 60 m thick depending on the kind of the
ceramic kiln furniture. More inside the parts we do not find a silicon peak. So it is obvious that
the differences in hardness between surface and core of the sinter parts go back to the
silicon content of the surface. The following table shows the distribution of the hardness (HV
0.1) as a function of the distance from the surface of the sinter part.
A
B
C

Surface layer (white) Surface 0.15 mm 0.30 mm Core
198 363 343 334 338
Table 1: Hardness in relation to the distance from surface

Kiln furniture

For the firing and also in some cases for the transportation within the plant some different
materials are used as kiln furniture. Due to the conditions to withstand the first material for
the use in this application is kiln furniture based on cordierite. If there is a need to go up to
higher firing temperatures the content of cordierite is reduced and the amount of mullite is
heightened as described above. Both materials have no problems with the gas atmosphere
in the firing kilns and with the thermal shock resistance. But this kind of material can cause
contact reactions as described above.
To avoid reactions we have to minimize the silica content of the ceramic material.
Therefore we take ceramic material with higher alumina content and lower silica content. At
first we tested a common used mullite material with the result that this kind of material
withstands the conditions caused by atmosphere and temperature but creates also contact
reaction between sinter parts and ceramic slab. As well we get good thermal shock
behaviour. The properties of the different ceramics are listed in table 2.

Cordierit /
Mullit
Andalusit /
Cordierit
Mullite
Corundum
A
Corundum
B
Chemical Analysis Unit
Al
2
O
3
[wt. %] 36.15 63.64 76.94 97.54 94.36
SiO
2
[wt. %] 51.02 33.15 22.45 1.79 5.01
Fe
2
O
3
[wt. %] 1.45 0.73 0.14 0.17 0.20
TiO
2
[wt. %] 0.65 0.34 0.18 0.18 0.10
MgO [wt. %] 8.80 1.61 0.01 0.02 0.01
Na
2
O+K
2
O [wt. %] 1.21 0.48 0.19 0.21 0.27
Physical properties
Bulk Density
[g/cm
3
]
2.07 2.48 2.51 3.00 2.64
Aparent Porosity [vol. %] 19.68 19.04 20.61 20.69 28.63
Cold Crushing Strength [MPa] 101 n.n. n.n. n.n. n.n.
Modulus of Rupture [MPa] 13.94 10.12 12.58 33.96 18.84
Youngs Modulus [MPa] 18,000 10,000 18.000 73,000 30,000
Thermal Expansion
Coefficient
[*10
-6
/K]
2.62 4.90 5.20 8.35 8.17
Poissons Modulus
(estimated)
0.22 0.22 0.22 0.22 0.22
Thermal Shock
Parameter R 1
[K] 230 K 160 K 105 K 45 K 60 K

Table 2: Properties of different kiln furniture

Based on the knowledge that pure alumina material can not cause contact reactions we
tested a high alumina product (Corundum A) with relatively low silica content. Indeed this
material caused no contact reactions like we see it in case of pure alumina. On the following
figures 7 and 8 we see that there is no contact reaction which can be seen on the surface
layer of the sinter parts. Nevertheless we see that the structure of the sinter part is nearly
homogenous in the area from the surface layer to the core. So this kind of material will cause
no problems during the hardening process.

Figure 7: Surface of parts sintered on low Figure 8: Surface of parts sintered on pure
silica containing corundum alumina

As shown in table 2 the thermal shock parameter is not good as this of previously tested
ceramic slabs. Obviously the low thermal shock resistance causes some difficulties in
practical use of the kiln furniture. For the use as real kiln furniture the thermal shock
resistance has to be improved.

Pore size distribution and specific surface area

Additionally we have to focus on another problem. Based on our experience especially in
the MIM process we can have production steps where the structure of the pores is important.
That is caused by the process itself in which the injection moulding process is a very special
step. Therefore the kiln furniture can affect the process in a positive way.
Some investigations show that the pore size distribution and the specific surface area is a
relevant parameter for the prediction of the performance of the ceramic kiln furniture. In
figure 9 the specific surface area of the different ceramic material is shown.
Cordierite /
Mullite
Andalusite /
Cordierite
Mullite
Corundum A
Corundum B
0
0,05
0,1
0,15
0,2
0,25
S
p
e
c
i
f
i
c

s
u
r
f
a
c
e

a
r
e
a

i
n

[
m
/
g
]
Specific surface area of different kiln furniture

Figure 9: Specific surface area of different ceramic material
As we know ceramics based on Cordierite/Mullite and Andalusite/Cordierite have a
sufficient function as buffer storage, whereas Corundum A works insufficiently. So there is
another problem to solve if the non reacting Corundum A material should be applicable.

Corundum B

Based on the experience with Corundum A and the commonly used cordierite containing
kiln furniture we have designed a new corundum material that will achieve the necessary
properties. Therefore it was essential to change the grain size distribution of the ceramic to
get more porosity. Accompanied with a lower firing temperature in the primary sinter process
we get a tremendous increase in specific surface area. The result is shown in figure 9. The
chemical analysis and physical properties are shown in table 1. The higher porosity in
combination with a suitable strength and a little enhanced thermal shock resistance make it
possible to use this kind of ceramic. To get this strength it was necessary to increase the
content of mullite and glassy phase in the ceramic structure which implicates higher silica
content. As we have seen during the tests the higher silica content of Corundum B does not
cause any contact reactions.
The fact that there is only a slight improvement of the thermal shock resistance proved by
the better thermal shock parameter indicates that this is a critical point for the performance
and extended use as kiln furniture.

Further development

To obtain a ceramic material with better thermal shock resistance it is also possible to
use another silica free crystalline phase in the structure. One possibility is to design new kiln
furniture based on corundum but with the addition of titanium oxide to obtain aluminium
titanate during the primary firing process. The data of this kind of material is shown in table 3.

Chemical Analysis Unit Physical properties
Al
2
O
3
[wt. %] 82.56 Bulk Density
[g/cm
3
]
2.70
SiO
2
[wt. %] 2.57 Aparent Porosity [vol. %] 28.40
Fe
2
O
3
[wt. %] 0.09 Cold Crushing Strength [MPa] 53
TiO
2
[wt. %] 14.29 Modulus of Rupture [MPa] 8.4
Na
2
O+K
2
O [wt. %] 0.24 Youngs Modulus [MPa] 2.800
Thermal Expansion
Coefficient
[*10
-6
/K] 4.65
Poissons Modulus
(estimated)
0.22
Thermal Shock
Parameter R 1
[K] 479 K
Data of Corundum / Aluminium Titanate
Table 3: Data of Corundum / Aluminium Titanate kiln furniture

As we see in the table there is one point that could be critical for a future use. The lower
MOR could cause some problems depending on the use of the new kiln furniture even when
it is additionally used for transportation.
The content of aluminium titanate reduces the thermal expansion and because of the
expansion difference between corundum and titanate the Youngs modulus is also decreased
dramatically. As a result we get new designed kiln furniture for an universal use in the case
of sintering metal powder even in the metal injection moulding process.


Fracture Toughness of Al
2
o
3
-Ni Composites Prepared by Powder
Metallurgy Method
T. Yener, I. Altinsoy, G. F. elebi Efe, M. Ipek, S. Zeytin, C. Bindal
Sakarya University, Engineering Faculty, Department of Metallurgy and Materials
Engineering, Esentepe Campus, 54187 Sakarya-Turkey
ABSTRACT
In the present study, we investigated the fracture toughness of alumina matrix composite reinforced
with nickel at ratios of 1%, 3%, 5% and 7% by weight prepared by powder metallurgy method.
Sintering process was performed at 1425 and 1440C for 1h in graphite powder. The fracture
toughness of test materials for 1425C and 1440C were 5.1, 5.2, 6.0, 6.7 MPa.m
1/2
and 5.4, 5.7, 6.1,
7.1 MPa.m
1/2
respectively. It was found that the more Ni content leads the higher fracture toughness
because of crack arresting/energy absorbing effect of metallic ductile Ni particles. The distribution of
reinforced nickel particles in the matrix was characterized by optical and SEM and the presence of
nickel and alumina phases were confirmed with EDS and XRD analysis. XRD analysis showed that it
has not been detected any oxidized Ni body in the composite. Optical and SEM studies revealed that
the distribution of nickel particles is uniform in the alumina matrix.
Keywords: Fracture toughness, Al
2
O
3
-Ni composite, hardness, sintering, powder metallurgy.
1. INTRODUCTION
The limitations imposed on using ceramics as structural components are mostly due to their low
toughness. Thus, the toughening of ceramic materials is very necessary, and has been investigated
for the years. Enhancing the latter can be done by different ways. The use of ductile phases is one of
the most applicable techniques known. The toughness increment achieved in these composites is
due to the, bridging, of ductile phase across the crack so exerting a closure stress behind its tip.
Especially in Al
2
O
3
system, the incorporation of a ductile phase has been considered an effective
toughening mechanism because its dispersion can lead to a remarkable absorption of the energy of
crack propagation due to plastic deformation. The dispersion of a ductile metallic phase (Ni, W, Mo,
Cu etc.) in a brittle ceramic matrix is found to be a promising way [1-5]. Besides, the distribution of
ductile phase in ceramic is very important for optimizing mechanical and functional properties of
composites. The main aim of present study is to investigate the preparation process and resultant
fracture toughness properties of alumina/Ni composites manufactured by powder metallurgy method.
2. EXPERIMENTAL DETAILS
2.1 Powder Processing
In this study, Al
2
(SO
4
)
3
was used as raw material to prepare pure alumina powders with an average
particle size of 0.3 m and purity of 99.9% and doped with metallic nickel powders with a mean
particle size of 3-7 m and purity of 99.9%. Al
2
O
3
-Ni composites with 1%, 3%, 5% and 7% Ni by
weight respectively firstly were mixed and ball milled for 48 h with alumina balls, then the composites
obtained were uniaxially pressed into pellets of dimension 15 mm x 3 mm with a pressure of 170
MPa which has a weight of 2 g. The compacts were pressed then sintered at 1425 and 1440C. To
avoid melting of Ni in an atmosphere controlled furnace at a heating rate of 5C min
-1
for 1 h and
after sintering they were cooled with a rate of 5C min
-1
up to room temperature.
2.2 Characterization
The presence of phases formed in sintered samples was determined by x-ray diffraction by using
CuK radiation with a wavelength of 1.5418 A over a 2 range of 10 to 80
o
. The morphology of
sintered alumina-nickel ceramic composites was confirmed by means of scanning electron
microscope (SEM). In order to verify the presence of Al
2
O
3
and Ni powders energy-dispersive x-ray
spectrometric analysis (EDS) was used. The hardness of samples was determined by using Vickers
indentation technique and fracture toughness of test materials was measured via indentation fracture
technique using Vickers diamond indenter (Fig.1) with a load of 98 N.


136
0
between
opposite faces

Fig.1 Vickers diamond indenter.
The equation used for calculating fracture toughness was: Kc= X P/c
3/2

where X is the residual-indentation coefficient [6, 7] which depends on hardness-to- modulus ratio
(E/H) of ceramic composites. The constant X is 0.016 (E/H)
1/2
, where H and E hardness and Youngs
modulus of test materials respectively. P is the applied load and c is the indentation half crack length.

3. RESULTS
3.1. X-Ray Characterization
X-ray diffraction analysis revealed that alumina has phase and nickel has metallic form in samples
sintered at 1425 and 1440C (Fig.2, Fig.3). As it can be seen in Fig.2 and Fig.3 oxidation-free nickel
was obtained and the more Ni content the higher the intensity.
0 20 40 60 80 100
C
o
u
n
t
s
0
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
*
.
Ni
Al
2
O
3
.
.
.
.
.
.
.
.
.
.. . .
*
*
.
.
T: 1440C
t: 1h
T: 1425C
t: 1h

7 wt.%Ni
5 wt.%Ni
3 wt. %Ni
1 wt. %Ni

Figure 2. X-ray diffraction patterns of test materials sintered at 1425C.


.
0 20 40 60 80 100
C
o
u
n
t
s
0
2000
4000
6000
8000
10000
12000
14000
16000
18000
*
.
Ni
Al
2
O
3
*
*
.
.
.
.
.
.
.
.
.
.
5 wt.%Ni
3 wt. %Ni
1 wt. %Ni
.
..
.
T: 1440C
t: 1h
7 wt.%Ni
Figure 3. X-ray diffraction patterns of test materials sintered at 1440C.
3.2. Microstructure
Fig. 4 shows the optical cross section view of the alumina nickel powder mixtures sintered at 1425
o
C
and 1440
o
C. Fig.5 and Fig.6 reveals SEM images with EDS spectrum of the alumina nickel powder
mixtures sintered at 1425
o
C and 1440
o
C relatively. Grey areas are Al
2
O
3
and white ones are Ni. The
distribution of Ni in composite is homogenous. SEM examinations revealed that Ni particles have a
close spherical morphology with micrometer particle size scale and does not contain significant
agglomeration.
In order to verify the presence of Al
2
O
3
and Ni powders energy-dispersive x-ray diffraction analysis
(EDS) was used and it was found that the dominant constituents are alumina and nickel. It was
observed that Ni particles were well-dispersed in alumina matrix. As a result it is possible to claim
that Al
2
O
3
-Ni composites can be obtained using powder metallurgy method.
Corresponding values of sintering temperature and composition of composite are given in Table 1 for
each compound.
Table 1. Description of symbols
Symbol Sintering Temperature
(C)
Composition
A1

1425

Al
2
O
3
- 1 wt. %Ni
A3 Al
2
O
3
- 3 wt. %Ni
A5 Al
2
O
3
- 5 wt. %Ni
A7 Al
2
O
3
- 7 wt. %Ni
B1

1440

Al
2
O
3
- 1 wt. %Ni
B3 Al
2
O
3
- 3 wt. %Ni
B5 Al
2
O
3
- 5 wt. %Ni
B7 Al
2
O
3
- 7 wt. %Ni


(A1) (A3) (A5) (A7)

(B1) (B3) (B5) (B7)
Figure 4. Optical cross section views of alumina nickel composites

A1 Mark 1 Mark 2

A3 Mark 1 Mark 2

A5 Mark 1 Mark 2
200m 200m
200m 200m
200m 200m
200m 200m


A7 Mark 1 Mark 2
Figure 5. SEM images of Al
2
O
3
-Ni composites including EDS spectra for 1425
o
C sintering
temperature

B1 Mark 1 Mark 2

B3 Mark 1 Mark 2

B5 Mark 1 Mark 2

B7 Mark 1 Mark 2

Figure 6. SEM images of Al
2
O
3
-Ni composites including EDS spectra for 1440
o
C sintering
temperature
3.3. Mechanical Properties
The fracture toughness values of alumina-nickel ceramic composites calculated by Vickers
indentation technique. The toughness values of the composites sintered at 1425
o
C increased from
5.06 to 6.7 MPa.m
1/2
with addition of nickel particles (Fig 7). The fracture toughness of samples
sintered at 1440
o
C increased from 5.39 to 7.05 MPa.m
1/2
being second ductile phase (Fig.8). When it
compared with the pure alumina, the toughness of Al
2
O
3
-Ni composites has been improved by 50%
with addition of nickel particles.
Besides, increment in fracture toughness, hardness values of Al
2
O
3
-Ni composites as a function of
nickel amount showed that as fracture toughness increased, hardness values of test materials
decreased because of ductile nickel particles (Fig 7, Fig 8).
0 2 4 6 8
Ni Content, wt.%
4.8
5.2
5.6
6
6.4
6.8
F
r
a
c
t
u
r
e

T
o
u
g
h
n
e
s
s

(
M
P
a
.
m
1
/
2
)
1320
1350
1380
1410
1440
1470
1500
1530
H
a
r
d
n
e
s
s

(
H
V
)
Fracture Toughness
Hardness

Figure 7. The variation of fracture toughness and hardness of Al
2
O
3
-Ni composites sintered at 1425

o
C as a function of nickel amount.
Hardness values of samples sintered at1425
o
C decreased from 1510 to 1327 HVN, and the samples
sintered at 1440
o
C decreased from 1559 to 1393 HVN.

0 2 4 6 8
Ni Content, wt.%
5.2
5.6
6
6.4
6.8
7.2
F
r
a
c
t
u
r
e

T
o
u
g
h
n
e
s
s

(
M
P
a
.
m
1
/
2
)
1380
1410
1440
1470
1500
1530
1560
1590
H
a
r
d
n
e
s
s

(
H
V
)
Fracture Toughness
Hardness

Figure 8. The variation of fracture toughness and hardness of Al
2
O
3
-Ni composites sintered at 1440

o
C as a function of nickel amount.
As it can be seen in Fig.9, the second ductile phase nickel particles arrest the crack propagation.

Figure 9. Microstructure of Al
2
O
3
-Ni composite including indentation mark and cracks which formed
by Vickers indenter.
4. DISCUSSION
As it is well known that the main problem for producing of ductile phase-reinforced ceramic matrix
composites is the oxidation of ductile phase. In order to overcome this problem at present study
production of test materials was carried out in graphite powder. Thus we were able to prevent
oxidation of nickel being ductile phase in composite. The presence of alumina and nickel were
confirmed by x-ray diffraction (XRD) analysis and no oxidation of nickel was observed. SEM
examinations of cross-sections of test materials showed that Ni particles are distributed in alumina
matrix uniformly. For 1425
o
C sintering temperature the fracture toughness values of test materials
were 5, 5.2, 6 and 6,7 MPa.m
1/2
and corresponding values of hardness measured by Vickers
indenter were 1510, 1445, 1428 and 1327 HVN. As for 1440
o
C sintering temperature, the fracture
toughness of test materials were 5.4, 5.7, 6.1 and 7.1 MPa.m
1/2
and corresponding values of
hardness measured by Vickers indenter were 1560, 1506, 1467 and 1394 HVN. As the addition of
nickel increased from 1% to 7%, hardness for 1440
o
C decreased from 1560 HVN to 1394 HVN and
indentation fracture toughness increased from 5.4 MPa.m
1/2
to 7.1 MPa.m
1/2
, respectively. According

to this result, the toughness of Al
2
O
3
-Ni composites has been improved between 35 and 50%
compared to that of pure alumina which has a fracture toughness of 3.9 MPa.m
1/2
. This increment in
toughness can be attributed to crack deflection and absorbing energy of crack by second ductile
nickel particle.
5. CONCLUSIONS
The fracture toughness of Al
2
O
3
can be enhanced by adding a second ductile Ni phase particles that
dispersed into matrix grains or grain boundaries. Alumina-nickel composites were produced
successfully from alumina and nickel powders by powder metallurgy method with oxidation-free Ni
particles. It was observed that during sintering Ni particles distributed in the alumina matrix uniformly.
Nickel addition to alumina increased fracture toughness by providing crack deflection and absorbing.
ACKNOWLEDGEMENTS
Authors thank to Fuat Kayis for performing XRD and SEM studies and special appreciations are
extended to technician Ersan Demir for experimental assistance at Sakarya University.

REFERENCES

1. Mekky, W., Nicholson, P.S., R-Curve Modelling for Ni/Al
2
O
3
Laminates, Composites Part B:
Engineering, Vol. 38, pp. 35-43, 2007.
2. Mekky, W., Nicholson, P.S., Fracture Toughness of Ni/Al
2
O
3
Laminates by digital image
correlation II: Bridging-stresses and R-curve models, Engineering Fracture Mechanics, Vol.73,
pp. 583-592, 2006.
3. Oh, I., Lee, J., Han, J., Microstructural Characterization of Alumina/Ni Composites Prepared By
Electroless Deposition, Surface&Coating Technology, Vol.192, pp. 39-42, 2005.
4. Li, G., Huang, X., Guo, J., Fabrication of Ni-coated Al
2
O
3
powders by the heterogeneous
precipitation method, Materials Science and Engineering, Vol.36, pp. 1307-1315, 2001.
5. Chen, R.Z., Chiu, Y.T., Tuan, W.H., Toughening alumina with both nickel and zirconia inclusions,
Journal of the European Ceramic Society, Vol. 20, pp. 1901-1906, 2000.
6. Celebi, G., Ipek, M., Bindal, C., Ucisik, A. H., Some Mechanical Properties of Borides Formed
on AISI 8620 Steel, Materials Forum, Vol. 29, pp. 456-460, 2005.
7. Ipek, M., Investigation of Sintering and Fracture Behaviour of Alumina- Zirconia Composite
Produced With Two Different Routes, Ph.D. Thesis, Sakarya University, Institute of Science
and technology, 2005.

EPMA Proceedings 2009 Sintering PDF

Hochgeladen von

Dokumentinformationen

Originalbeschreibung:

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

EPMA Proceedings 2009 Sintering PDF

Hochgeladen von

Copyright:

Verfügbare Formate

Euro PM2009 Field Assisted Sintering

Manuscript refereed by Professor Bernd Kieback, Fraunhofer IFAM/TU Dresden,

, for controlling the C-Potential in sintering furnaces specifically designed through

system as shown in Figure-2 comprises an analyzer and a

allowing operators to control the sintering furnace atmosphere completely.

Das könnte Ihnen auch gefallen