The effect of composition on marker movement and kirkendall porosity in ternary alloys
- PDF / 368,871 Bytes
- 5 Pages / 612 x 792 pts (letter) Page_size
- 26 Downloads / 228 Views
I.
INTRODUCTION
THE "Kirkendall effect"
states that unbalanced intrinsic fluxes during interdiffusion lead to marker movement and porosity in the reaction zone of a diffusion couple. 1~'2,31 The effect is illustrated in Figure 1 for a couple with concentration differences of AC ~ between the initial alloys. The marker moves a distance X K in time t~, and its instantaneous velocity is related to the intrinsic fluxes and partial molar volumes of all elements at the marker plane. Recently, it was shown t4Tthat the distance of marker movement can be expressed as a linear function of initial concentration differences for a given alloy system. The general expression of marker movement for an ncomponent alloy system is as follows: X x = (a,AC~
+ a2AC~ + ...
+ a,_~AC,~
Ill
In Eq. [1 I, each ai is a coefficient which is assumed to be constant for small AC,.~ and can be determined from intrinsic diffusivities and inverse square root diffusivities of the system. ~41 However, the ai coefficients are expected to vary with alloy composition in principle, because the diffusivities are functions of composition. 121 Therefore, Eq. [1] needs to be tested experimentally in order to determine the composition range over which it is valid. The present work is a report of such tests. Experiments were performed on marker movement during isothermal annealing of Ni-Cr-AI y-phase alloys at 1200 ~ The linearity of the experimental results was examined, and then the conditions when there was no marker movement and when there was maximum marker movement were determined. The agreement found suggests that Eq. [1] can be a useful basis for modifying alloy systems in order to reduce or eliminate the Kirkendall effect.
II.
EXPERIMENTAL PROCEDURE
A number of y-phase Ni-Cr-AI alloys were made, and their chemical compositions are plotted as circles in Figure 2. The compositions in the figure which are conYOON-HO SON, Graduate Assistant, and J.E. MORRAL, Professor, are with the Department of Metallurgy and Institute of Materials Science, the University of Connecticut, U-136, Rm. 111, 97 N. Eagleville Rd.. Storrs, CT 06269-3136. Manuscript submitted January 24, 1989. METAI.LURGICAL TRANSACTIONS A
nected by a line were formed into diffusion couples. Alloy 1 had a composition of (Ni-8.8 at. pct Cr-8.8 at. pet AI), while the other alloys had compositions that differed from this by the amount given in Table I. The alloys were prepared from high-purity Ni (99.99 pct), Cr (99.99 pct), and A1 (99.99 pct) by arc melting in a high-purity argon atmosphere. Repeated melting in the arc melter provided a uniform composition within the alloy button. Subsequently, each button was dropcast into a water-cooled copper mold to form a rod-shaped specimen of about 6 - m m diameter. The specimens were vacuum encapsulated in a quartz tube and then homogenized at 1100 ~ for 100 hours. After cooling, each rod specimen was sliced into disks having a thickness of 2 mm. The disk faces were ground and polished to insure that both were parallel. In order to make a diffusio
Data Loading...
