Quantitative analysis of the volume fraction of acicular ö phase
in an Inconel 718 alloy
A. Hazotte, M. Dehmas, P. Weisbecker, E. Gautier, C. Dumont*
Laboratoire de Science et Genie des Materiaux Metalliques (INPL/CNRS n°7584) - Nancy - FRANCE
* Aubert & Duval - Les Ancizes - FRANCE
Introduction
Nickel-based superalloys strengthened by precipitation of the Ni,Nb-based phase, y’’, are widely used to
produce hot parts of turbine engines [1]. One key-point of their manufacturing process is to control the
amount of & phase, another form of the Ni,Nb intermetallic compound, which precipitates during high
temperature thermo-mechanical treatments. While a small amount of § phase is required to hinder the
grain growth during forming, a too high amount is detrimental to the final mechanical properties of the
‚erhalten. von alloy. Thus, 8 phase volume fraction has to be evaluated very precisely at the different stages of the
nterschiedlicher process. The present study deals with the comparison of ö amounts measured by different techniques on
lität. bei‘ der various samples of an Inconel 718 alloy, the most usual y’’-strengthened superalloy [1]. The techniques
erhesserung der used were X-ray diffraction and quantitative analysis of images of light and scanning electron microscopy.
n Kemer. W. Material and techniques
ad an Stabilität The chemical composition of Inconel 718 alloy is reported in Table 1. Note that, with regard to the large
gierung weisen composition range allowed -particularly for niobium- Inconel 718 has to be considered as a generic term
Grobes Kom rather than as a strict alloy specification. Thus, for a given set of heat and mechanical treatments,
zu arößeren significant variations of phase amounts can be encountered as the result of composition differences.
| Mikıns . - = Zo. _ m . - - .
te element | Ni Cr Fe Co Mo Nb Ti Al C Mn | Si |
sat pe hr Hh a dh Toa To]
Table 1: Chemical composition range of Inconel 718 alloys
A typical processing route of this type of alloys is schematised in Figure 1. It is composed by a first
forming stage performed above 1000°C, then a second forming stage between 900°C and 1000°C, and
mstopde finally heat treatments between 600°C and 800°C, which promote the precipitation of the strengthening
intermetallic compounds, Ni,Nb-based y’” phase but also Ni;(Al,Ti)-based ¥’ phase [1,2]. During the first
forming stage, the only stable phases are the fcc solid solution y plus some Nb- and Ti-carbides. However,
the high temperature thermo-mechanical treatment applied during the second forming stage leads to the
precipitation of 8 phase, the high temperature stable form of the Ni,Nb compound. As illustrated in Figure
2, this phase can be found in the form of rather equiaxed intergranular precipitates or of acicular
intragranular precipitates. The § volume fraction, as well as the amount of intra- and intergranular ö
precipitates, strongly depend on the process parameters (temperature, time, strain, etc..). In the industrial
context, 8 volume fraction is usually measured — together with y grain size - by conventional quantitative
analysis of light microscopy images.
Lu In the present work, we measured the § phase volume fraction, Vv(9), in a lot of Inconel 718 samples taken
at different times during the second stage of the industrial process. We also analysed samples heat treated
during different times, at different temperatures ranging between 900°C to 1000°C, with the objective to
characterise the § precipitation kinetics. These thermal and thermo-mechanical cycles promoted the
eri, Band VI precipitation of various amounts of with different ratios of inter- and intragranular particles (ctf. Fig.2).
i H. Riedel) Vv(S) was measured by quantitative X-ray diffraction and by automatic analysis of images from a light
met microscope (LM) or from a scanning microscope in back-scattered electron mode (BSEM). X-ray
diffraction was performed with a D500-Siemens diffractometer, using an anticathode in cobalt. The
working voltage was 37 kV and the probe intensity 20 mA. The diffraction peaks obtained were
quantitatively analysed using the software FullProf [3]. Both LM and BSEM observations require to
INS