Interdiffusion of Sn and Pb in liquid Pb-Sn alloys

Citation
M. Klassen et Jr. Cahoon, Interdiffusion of Sn and Pb in liquid Pb-Sn alloys, MET MAT T A, 31(5), 2000, pp. 1343-1352
Citations number
35
Categorie Soggetti
Apllied Physucs/Condensed Matter/Materiales Science",Metallurgy
Journal title
METALLURGICAL AND MATERIALS TRANSACTIONS A-PHYSICAL METALLURGY AND MATERIALS SCIENCE
ISSN journal
10735623 → ACNP
Volume
31
Issue
5
Year of publication
2000
Pages
1343 - 1352
Database
ISI
SICI code
1073-5623(200005)31:5<1343:IOSAPI>2.0.ZU;2-I
Abstract
Coefficients for the interdiffusion of Sn in Pb-rich alloys and Pb in Sn-ri ch alloys were established using 1.5-mm-diameter capillaries and the semi-i nfinite rod technique. Interdiffusion coefficients are presented for the en tire concentration range from pure Ph to pure Sn, for temperatures from 668 to 1031 K. The concentration dependence of the interdiffusion coefficients was determined by establishing the concentration along the length of the c apillaries and calculating the coefficients using a finite-difference techn ique. The interdiffusion of Sn in Pb, extrapolated to 0 at. pet Sn, is give n by D = 8.8 X 10(-8) exp - (22,600/RT) m(2)/s and that for Pb in Sn, extrapolated to 0 at. pet Ph, by D = 2.4 X 10(-8) exp - (19,300/RT) m(2)/s The "average" value for the interdiffusion of Sn in Pb, for the concentrati on range from 0 to 74 at. pet Sn, is given by D = 1.1 X 10(-7) exp - (25,200/RT) m(2)/s and the average value for the interdiffusion of Pb in Sn, for the concentra tion range from 0 to 26 at. pet ph, is given by D = 1.3 X 10-8 exp - (22,600/RT) m(2)/s The values obtained for the coefficients agree reasonably well with previou s results for the diffusion of Sn in Pb-rich alloys and are consistent with solvent self-diffusion coefficients for pure Pb and pure Sn. However, whil e the diffusion coefficients obtained from these Arrhenius equations are li kely of the right order of magnitude, it is concluded that the results are affected by fluid flow in the capillaries, resulting in higher than actual activation energies. It is suggested that, for the capillary-reservoir tech nique, convective flow in the reservoir across the open end of the capillar ies induces "lid-driven" flow in the upper portions of the capillaries, res ulting in higher than actual diffusion coefficients, particularly for the S n-rich alloys, since the Sn-rich end of the capillaries was open to the res ervoir. Because of fluid motion induced in the capillaries, all of the resu lts for solute and self-diffusion in Pb, both present and previous, are lik ely erroneous because they were obtained using the capillary-reservoir tech nique. Some previous results for solvent self-diffusion in liquid Sn were obtained using either the thin disk or the semi-infinite rod technique and, since t hese results agree with results obtained in microgravity, it is concluded t hat the nonreservoir methods may provide a means of obtaining more accurate liquid diffusion data.