Influence Of The Particle Size Distribution And Morphology Of

Preparing to load PDF file. please wait...

0 of 0
Influence Of The Particle Size Distribution And Morphology Of

Transcript Of Influence Of The Particle Size Distribution And Morphology Of

Ian Bailon-Poujol, Julie Campbell-Tremblay, Lydia Aguirre Rio Tinto Metal Powders
Sorel-Tracy, Québec, Canada
ABSTRACT Dimensional change is one of the most important properties of PM structural parts and part producers are demanding tight tolerances. Copper and nickel elemental powders are commonly used as admixed additives to tailor targeted properties but both result in a significant dimensional change. Both elements generally increase the mechanical properties but copper is known to promote growth while the use of nickel leads to shrinkage.
This paper focuses on the impact of characteristics of copper and nickel additives on the properties of the resulting PM parts and particularly on dimensional change. The particle size distribution (PSD) and the morphology of different types of Cu and Ni additives were characterized and the link between these characteristics and the dimensional change was investigated under slow cooling rates.
INTRODUCTION As powder metallurgy expands to the production of high-performance components with tight tolerances and superior mechanical properties, it becomes of critical importance to better predict and control the dimensional change of sintered parts since secondary operations like sizing and machining are generally costly and difficult to perform. Copper and nickel are commonly used as alloying elements in PM steels. Copper is generally added as an elemental powder to premixes while nickel is suitable to be pre-alloyed or premixed. PM materials containing a combination of these alloying elements are commonly used for highperformance application (e.g. FLNC-4405).
During typical sintering cycles, admixed copper melts at around 1083°C and is redistributed throughout the steel compact by liquid phase displacement, grain boundary diffusion and ultimately by volume diffusion in the iron lattice. The redistribution mechanisms of copper during sintering lead to swelling [1, 2, 3]. The presence of carbon in the Fe-C-Cu system tends to decrease the swelling mainly because carbon decreases the wettability of liquid copper which has a direct impact on the spreading efficiency of the liquid phase and, hence, grain boundary penetration [3]. Elemental copper powders commonly used in PM

have an average particle size of 50-70 µm. Previous studies showed that the size of copper particles has a significant effect on the extent of the swelling of PM parts [2, 4]. This amplitude of the swelling seems to vary considerably with, among other things, the nature of the base powder, the type of formulation and the sintering profile characteristics.
On the other hand, fine nickel particles remain fully solid at typical sintering temperatures. Carbonyl nickel powder with d50 around 7-10 µm is the most commonly used nickel additive in PM steels. Different mechanisms responsible for the redistribution of nickel during sintering have been described in numerous publications but do not seem to be unanimous [1, 5, 6, 7, 8]. It seems that volume diffusion of nickel in iron is limited, so that surface and grain boundary diffusion mechanisms are dominant [7, 8]. Hence, it is not unusual to witness the presence of nickel rich areas in sintered parts, caused by the slow redistribution rate of the nickel in steel particles [6, 7, 8]. In terms of dimensional change, the presence of fine nickel particles increases the total specific surface area of the compact causing larger shrinkage at holding temperature [1]. In that sense, the particle size distribution of nickel has a direct effect on the dimensional change of PM steels.
The dimensional change response of PM steels containing both admixed copper and nickel is not straightforward. It has been demonstrated that there is an interaction between both elements during the course of sintering [4, 5, 9, 10]. This interaction has a significant impact on the dimensional response and depends on the sintering profile characteristics (heating rate, sintering temperature and time, etc). Nabeel et al. suggests that the formation of Cu-Ni liquid phase, which has a lower wetting angle than liquid copper, enhances the redistribution of both species in the compact [5]. In another study, Lindsley et al. found that the use of fine copper with nickel diminishes the extent of growth in FLNC-4408 PM steels when compared with regular copper [4]. This is also in line with Strobl et al. who used copper coated iron particles in their work, an extreme case which can be considered as infinitely fine copper particles, and found that swelling was greatly reduced or even suppressed [11].
Dilatometry is a powerful characterization method to appreciate the in-situ dimensional change of a sample during sintering cycle [12, 13, 14]. Dilatometry provides useful information on thermal expansion, phase transformations, swelling and densification.
In this work, a dilatometry study was carried out by the authors to evaluate the dimensional response of PM steels containing elemental copper and nickel additives. The main objective of this study was to evaluate the effect of specific characteristics of copper and nickel additives on the sintering kinetics, particularly the particle size distribution of copper additives and the morphology of nickel particles. The dilatometry results shown in this study contribute to a better understanding of the interaction between copper and nickel during sintering.
A commercially available water-atomized ferrous powder, pre-alloyed with 0.85% Mo (ATOMET 4401), was used in this study. All mixes contained 0.60% TIMREX® KS15 synthetic graphite (Timcal). Two elemental copper powders, both produced by water-atomization and sieving, but having different particle size distributions, were selected: regular copper (RegCu) and fine copper (FineCu), see Table 1. It can be appreciated that the morphology of copper particles is qualitatively similar for both copper powders (Figure 1). Two types of nickel with different particle morphologies were used: a nickel powder produced

by a carbonyl process (CarbNi) and a nickel produced via a reduction, milling and sieving process (ReduNi). As seen in Figure 1, the production process had a significant impact on the morphology of the nickel particles. The surfaces of particles produced by the carbonyl process have a spiky appearance which is typical of this type of production process, while particle surface looks smoother in the case of reduced nickel.





Figure 1 : SEM images of powder additives. (A) Regular copper. (B) Fine copper. (C) Carbonyl nickel. (D) Reduced nickel.
The particle size characteristics of the different additives were determined by laser diffraction using a HELOS particle size analyzer (SYMPATEC) and are listed in Table 1. All the additives used in this study, including copper and nickel, are commercially available products.

Table 1: Particle size properties of copper and nickel additives used

Cu/Ni additives RegCu FineCu CarbNi ReduNi

d10 (µm)
21.6 5.7 1.7 2.7

d50 (µm)
54.2 11.7 7.0 10.4

d90 (µm)
104.5 20.3 22.9 21.9

d99 (µm)
150.0 28.2 55.8 34.8

Mixes were prepared with various amounts of the different types of copper (0% and 2%) and nickel (0% and 2%) without any lubricant in order to be used for dilatometry. Dilatometry specimens (20.00 mm x 6.35 mm x 6.35 mm) were compacted at 6.80 g/cm³ using die-wall lubrication (manually applied zincstearate spray). Thermal expansion was examined by a vertical push-rod dilatometer (Linseis L70) in an atmosphere composed of 90 vol% of nitrogen and 10 vol% of hydrogen. The sintering profile used for dilatometry consisted of a heating ramp of 25°C/min until 1100°C. The heating rate was slowed down to 10°C/min from 1100°C to 1120°C, and 5°C/min from 1120°C to 1130°C. Temperature was held at 1130°C for 30 minutes. The actual temperature overshot to 1138°C at the beginning of the plateau and progressively settled down to the set-point temperature. Cooling rate was set to 20°C/min from 1130°C to room temperature. The dilatometer failed to maintain this cooling rate below 650°C. The effective cooling rate observed in the dilatometer between 650°C and 400°C is approximately 17°C/min. Prior to the sintering cycle, samples were held in the dilatometer at 125°C for 30 minutes to stabilize the system. The temperature profiles used for dilatometry trials are illustrated in Figure 2. As a comparison, a typical sintering profile without accelerated cooling is also shown in the same figure.
Figure 2: Temperature profiles used for sintering and dilatometry The dilatometry profiles presented in this paper illustrate the percentage of variation of the specimen expansion during the sintering cycle, as compared with the measured length of the initial green specimen at room temperature. Every dilatometry profile was repeated twice to validate the consistency of the measurements. Sintered specimens were cross-sectioned, mounted and polished according to standard practices and microstructural characteristics were revealed using Nital (2%). All the micrographs shown in this paper were obtained using an optical microscope (Olympus GX71).

From each dilatometry profile, it was possible to extract dimensional variations associated with specific events happening during the sintering cycle. More particularly, the amplitude and the rate of the growth caused by copper as well as the shrinkages observed at sintering temperature were evaluated in each case and are listed in Table 2.

Table 2: Dimensional variations related to different events during the sintering cycle

Ref. 1 2 3 4 5 6

MPIF designation
FL-4405 FLN2-4405 FLN2-4405
FLNC-4405 FLNC-4405

2% RegCu 2% FineCu 2% RegCu 2% FineCu

2% CarbNi 2% ReduNi
2% CarbNi 2% CarbNi

Growth (Cu) (%)
0.32 0.37 0.42 0.23

Rate of growth (Cu)
(%/°C) -
0.007 0.007 0.011 0.006

High-temp. shrinkage
(%) 0.17 0.28 0.26 0.10 0.09 0.25 0.22

The reader can relate these values to the individual dilatometry profiles presented in the next pages. Figure 3 shows a typical dilatometry profile, illustrating the key events and their amplitudes. Shrinkages observed at sintering temperature (high-temp. shrinkage) are not perfectly vertical on the graphs due to the temperature overshoot mentioned previously. Nevertheless, the shrinkage was calculated by evaluating the difference between the maximum expansion reached at the beginning of the plateau (1130-1138°C) and the length of the specimen right before cooling began (1130°C).


Cu growth

High. temp. shrinkage

Figure 3: Typical dilatometry profile of PM steel specimen containing copper with key events.

Impact of the morphology of nickel particles First, the impact of the morphology of nickel particles was studied. Figure 4 shows the dilatometry profiles of copper-free samples containing 2% of the two types of nickel (CarbNi and ReduNi). The shaded curve represents the dilatometry profile of a reference sample free of copper and nickel. Thermal expansion of ferrite during the heating phase is similar in both cases and is consistent with past studies [12, 14].
Figure 4: Dilatometry profiles of copper-free samples containing 2% of the different nickel (FLN2-4405). As expected, large shrinkages are observed at holding temperature in both cases (0.26% and 0.28%). This shrinkage is larger than the reference sample (0.17%) which confirms that the presence of fine nickel particles increases the total surface energy of the specimen, resulting in more shrinkage. Nickel produced by the carbonyl process causes a slightly larger shrinkage than reduced nickel. It is difficult to correlate this tiny difference to the morphology of the nickel particles since there are also small differences in the particle size distribution of both powders. Nickel produced by the carbonyl process is slightly finer than reduced nickel, which could have contributed to the larger shrinkage observed (see Table 1). The cooling response is also very comparable in both cases. The beginning of the transformation of austenite is delayed in time and the amplitude of the growth generated by this transformation is smaller when compared to the reference sample. This is due to the presence of residual austenite (nickel rich area) which as a higher density than equilibrium phases ( and Fe3C), as seen in Figure 5. The final lengths of

the bars containing nickel are shorter than for the nickel-free formulation, confirming that the presence of nickel amplifies the overall shrinkage of PM steels.
The dilatometry profiles shown here suggest that the type of nickel, and to a further extent, the morphology of the nickel particles does not have a significant influence on the overall dimensional behaviour, in the range of particle size distribution tested here.



Residual austenite

50 µm

50 µm

50 µm
Figure 5 : Microstructures of sintered specimens, etched with Nital (500x): (A) 2% carbonyl nickel / 0%Cu. (B) 2% reduced nickel / 0%Cu. (C) 0%Ni / 0%Cu Impact of the particle size distribution of copper additives

The dilatometry curves of nickel free samples containing 2% copper are presented in Figure 6. Both

specimens show similar expansion profiles until

Following this transformation,

expansion is accelerated by the diffusion of graphite in the austenite phase. The two samples expand in a

similar way until 1080°C where copper melts. In fact, the growth due to copper occurs earlier in the case

of fine copper (~1083°C) versus regular copper (~1095°C) in the specific sintering conditions of this

study. The amplitude of the growth is slightly larger in the case of fine copper (0.37% vs 0.32% for

regular copper). It is important to mention that the rate of expansion during this growth appears to be

nearly the same for two types of copper.

2.0 1.8 1.6 1.4 1.2





Figure 6 : Dilatometry curves of nickel-free samples containing 2% of the different copper
In the following paragraph, a possible sequence of events resulting in the difference observed between fine and regular copper will be suggested. Figure 7 show the initial size and distribution of copper particles in green compacts (mounted in epoxy). As a consequence of their size, fine copper particles are more homogenously distributed throughout the compact. Liquid copper has to travel smaller distances before penetrating grain boundaries and eventually copper atoms go quicker in solid solution into the iron lattice. Moreover, a better spatial distribution allows more grain boundaries to be influenced by copper diffusion, resulting in more growth. Direct volume diffusion from the copper particles into the iron lattice from the periphery of the iron particles is also enhanced by a better spatial distribution. These are some reasons that can explain the earlier and slightly larger growth seen with fine copper, during this specific sintering cycle.
In the case of regular copper, it is suggested that this growth is postponed in time since liquid copper has to travel as a liquid phase in pore channels first before generating swelling by penetrating grain boundaries or diffusing in volume.



200 µm

200 µm

Figure 7: Microstructure of green specimens, as-polished (200x). (A) regular copper. (B) fine copper
Following copper growth, equivalent high-temperature shrinkages are observed (around 0.10%) and the cooling profiles are practically identical in both cases. The total high-temperature shrinkage of samples containing copper is less than the copper-free reference sample. This can be explained by the fact that, at the beginning of the plateau, shrinkage is probably still in competition with the growth caused by copper penetration.


seems to be modified by the presence of copper. Final microstructures are

presented in Figure 8. Final dimensions observed are larger than green dimensions and confirm that

copper caused a tangible growth.

In the current case, the particle size distribution of copper seems to have a significant influence on the extent of the growth caused by copper penetration, fine copper promoting more growth.



50 µm

50 µm

Figure 8: Microstructures of sintered specimens, etched with Nital (500x): (A) 2% regular copper / 0%Ni. (B) 2% fine copper / 0%Ni.
Interaction between nickel and different copper additives

For this part of the study, samples with 2% copper and 2% nickel were prepared. Since the type of nickel (carbonyl vs reduced) appears to have no impact on the sintering behaviour, the dilatometry profiles shown in Figure 9 relate to samples containing only carbonyl nickel with the different types of copper.

2.0 1.8 1.6 1.4 1.2





Figure 9 : Dilatometry curves of samples containing 2% of different types of copper and 2% of carbonyl nickel (FLNC-4405)
Up to the melting temperature of copper, the expansion behaviour is similar to previous profiles shown in this study and earlier comments still apply here. However, the growth associated with copper adopts a singular shape in FLNC-4405 formulations when comparing with samples containing only copper or only nickel.
In the case of fine copper, clear differences are noticeable when comparing with the nickel-free sample (Figure 6). First, the beginning of the growth caused by copper appears at a higher temperature (~1093°C vs ~1083°C). Second, the total amplitude of this growth (0.23%) is significantly lower than what was observed with the nickel-free specimen (0.37%). The rate of this growth is similar with and without the presence of nickel (~0.007%/°C).
A very different kinetic is observed for regular copper. Strikingly, in the presence of nickel, the total growth is larger (0.42%) than the nickel-free specimen (0.32%) and the rate of growth is approximately 50% more pronounced (0.011%/°C vs 0.007%/°C). The temperature where swelling appears is similar with and without nickel (1095-1096°C).
Moreover, it is important to mention that the end of the growth regime appears to flatten at around 1130°C, just before reaching the sintering temperature, in both cases whereas in the case of nickel-free formulation the growth was still noticeable when reaching the maximum temperature (1138°C). It was postulated that the swelling effect of copper is probably still active during the early stages of the fixed