Optimizing metal powders for Isostatic Pressing

Introduction

Isostatic pressing is a component-forming process where pressure is applied uniformly (using gas or liquid) to a hermetically sealed container filled with compacted metal powder. Isostatic pressing can be performed at elevated temperatures, known as hot isostatic pressing (HIP), or at ambient temperatures, known as cold isostatic pressing (CIP).

Figure 1. Schematic of a Cold Isostatic Pressing (CIP) process

Isostatic pressing has several benefits over the commonly used press and sinter method, including equal compaction in all directions and a more uniform final component density. Nevertheless, as with other powder metallurgy processes, manufacturers must carefully characterize the metal powder’s properties for the process to be successful. Commonly characterized physical properties include powder flow, density, hardness, particle size, and particle shape of the respective powders.

Other key characteristics include chemistry and microstructure. Chemistry is paramount as the powder needs to comply with the alloy composition of the material specified, while phase composition and grain size can affect powder hardness and melt behavior which can affect the forming and sintering process.

In this study we are focusing on the complementary use of X-ray powder diffraction (XRPD), X-ray fluorescence (XRF), and Automated Imaging to evaluate the performance of various iron-based alloys for use in a CIP manufacturing process. X-ray diffraction is a powerful tool for material characterization as it provides information about the phase/mineralogical composition (ferrite, austenite, cementite, and other carbides) as well as crystallite size and microstrain in the sample powders. X-ray fluorescence provides accurate information about the elemental composition of the sample, and Automated Imaging can provide particle specific size and shape information.

X-ray fluorescence

The basic principle of XRF analysis is relatively straightforward. If a sample is irradiated by X-ray radiation of a given energy it can cause constituent atoms to become ionized. In response, the sample emits fluorescent X-ray radiation. These X-rays have energies (or wavelengths) that represent the elements present in the sample. In other words, by measuring the energies Ei (or wavelengths λi) of these fluorescent X-rays, we can tell what elements are present. These X-rays have energies (or wavelengths) that represent the elements present in the sample. In other words, by measuring the energies Ei (or wavelengths λi) of these fluorescent X-rays, we can tell what elements are present. The characteristic X-ray line energies for each element in the periodic table are well documented. For example, if a 6.92 keV X-ray photon comes out of the sample, then it is likely to be Co K-alpha, and a 7.47 keV photon would be Ni K-alpha, and so on.

Figure 2. Illustration showing the basic principles of X-ray fluorescence embedded in a typical XRF spectrum

Under carefully controlled conditions, we can count the number of X-ray photons coming from each element over a period, for example, one minute, and use this to calculate the proportions of each element in the sample. A powder sample can be presented as a loose powder, pressed powder, fused glass bead, or melt casted disc, while formed parts can be measured directly or cut to a suitable size.

Automated Imaging

Suitable for particles from 0.5 µm to >1 mm, automated imaging offers particle size and shape measurement on a statistically rich ensemble of particles – either as dry powder dispersion or as dispersion in a liquid medium. Automated imaging systems capture individual images of tens of thousands of particles in a dispersed sample within minutes. Multiple size and shape parameters are calculated for each particle and used to build up statistically significant number-based distributions. The most commonly used shape parameters are shown in Figure 3.

Figure 3. Particle morphology parameters revealed through automated image analysis

X-ray diffraction

Phase Analysis

X-ray diffraction-based phase analysis is a powerful tool for material characterization as it provides information about the type of structural components forming certain materials (e.g. ferrite, martensite, austenite, cementite…).
This in turn allows us to predict and/or explain the physical and chemical properties of these materials during subsequent processes.

Figure 4. Illustrations of the crystal structures of Austenite, Ferrite and Martensite and their corresponding diffraction patterns

Figure 5 shows the crystal structures and corresponding diffraction patterns of Austenite, Ferrite, and Martensite and demonstrates the power of XRD for elucidating and quantifying the different phases present in a material. These three different phases differ in the arrangement of the atoms and this can result in very different properties. For example, austenite can absorb more carbon than ferrite and so it has better corrosion resistance. Martensite is a harder form of iron that is formed at high temperatures but can be retained if the sample is rapidly cooled. Further phases are also possible when additional alloying elements are added to the mix. The relative amounts of the different phases will determine the physical properties of the material.

Figure 5. Quantification of the phases present in a material using the Rietveld method

Quantification of the phases present can be achieved using the Rietveld method. Rietveld analysis is based on a least-square fitting approach (“refinement”) where the whole diffraction pattern is modeled based on expected phases and several physical parameters. One of the results of this refinement is a simultaneous quantification of all phases present in the sample. In addition, preparation effects as well as the decrease in tube efficiency are taken into account (no monitoring required!).

Grain size and strain analysis

Peak broadening (β) due to crystallite size (D) can be very effectively expressed in terms of the Scherrer formula. Peak broadening is greatest for smaller crystallite size, and for the same crystallite size, the peak broadening is more for the higher angle peaks as illustrated in Figure 6. The peak broadening caused by this micro-strain is adequately represented by the Tangent Formula.

Figure 6. Illustration showing how peak shape correlates with crystallite size and micro-strain

Larger micro-strain causes larger peak broadening and for a given micro-strain the higher angle Bragg peaks would experience larger broadening. Thus, both the crystallite size as well as the micro-strain broaden the higher angle Bragg peaks more severely. Yet, their relative contributions can be separated due to different dependence on the angle the peak occurs at. Crystallite size has a 1/cosθ dependence and micro-strain has a tanθ dependence. These two formulas are at the heart of all line profile analysis methods.

The single-line method uses the integral breadth as a measure of the peak width and the universal shape factor (K ) to describe the peak shape variation. These numbers can be determined for single peaks as well as for a whole range of peaks. The algorithm then empirically deconvolutes the profile into a Gaussian and a Lorentzian part. The instrumental influence is determined in the same manner and is subtracted from the two profile parts.

The net Gaussian broadening is used to calculate the micro-strain value by the tangent formula. The net Lorentzian broadening enters the Scherrer formula to determine the mean crystallite size.

Experimental

Metal powders were produced from four iron-based alloys; A, B, and C. Some of these powders were formed using gas atomization (AGN), whilst others were formed using water atomization processes (AWN, AWY, BWN, CWN). These initial samples underwent the CIP process and following this, some were subjected to an annealing process to try to improve their process performance (BWNA, CWNA,).

Particle size and shape analysis were performed on a Morphologi 4 Automated Imaging system. X-ray diffraction measurements were performed on a Empyrean Multicore with the PIXcel^3D detector and a Co source. X-ray fluorescence measurements were performed on an Epsilon 3 Spectrometer.

Results and discussion

Table 1 shows the elemental composition of the various powders as measured with XRF. The Chemistry within the different sample series is consistent, but they also show significant differences in chemical composition between the various types of alloys.

Table 2 summarises the XRD results along with particle size, particle shape, and hardness for each alloy including whether that particular powder formed a CIP part successfully. Alloy A formed a successful part when it was water-atomized but not when it was gas-atomized. For these two materials, the phase composition and crystallite size are similar, and they also exhibit similar hardness values, with the gas atomized material being a little softer than water atomized. The particle size of the two materials is also similar but the particle morphology between the two is found to be very different, as Figure 7 illustrates, with the water atomized material having a much lower circularity than the gas atomized one. It is likely that the more irregular particle shape enables particle interlocking during the manufacturing process, conferring stability during the formation of the green body through improved cold welding.

For samples B and C which were both water atomized, the atomized powder did not form stable green parts although the annealed powders did. As particle shape size and shape were similar between annealed and as atomized powders it could not be attributed to morphology but can instead be attributed to a significant reduction in hardness on annealing. This reduction in hardness corresponds to an increase in crystallite size which corresponds to a decreased number of grain boundaries and enhanced deformation.

Figure 7. Comparing circularity for the water and gas-atomized Alloy A, inset: representative particle images, from the mean circularity, for the two samples.

Conclusions

In order to ensure good quality consistent manufacture of metal parts during Cold Isostatic Pressing (CIP) a thorough understanding of raw the metal powders is required including particle size distribution, particle shape, elemental composition, and microstructure of the material. A tool kit of analytical techniques including Image Analysis, X-ray diffraction, and X-ray fluorescence enables such parameters to be assessed and correlated with final product performance.