Decoupled bulk and surface crystallization in Pd 85 Si 15 glassy metallic alloys: Description of isothermal crystallizat
- PDF / 564,297 Bytes
- 8 Pages / 593.28 x 841.68 pts Page_size
- 41 Downloads / 133 Views
Isothermal devitrification of Pdg5Si15 amorphous alloys has been analyzed using differential scanning calorimetry (DSC) and x-ray diffractometry. Both as-quenched and aged amorphous ribbons were investigated. Crystallization of aged samples starts from the surface and proceeds several micrometers into the bulk. The product of this process is a layer of strongly textured palladium (111) followed by a mixture of Pd2Si, Masumoto MSI phase, and untextured palladium. Next, the crystallization occurs via a different (bulk) mechanism, resulting in a mixture of Masumoto MSII phase and untextured palladium. The bulk mechanism is the only one observed in as-quenched samples. The surface and bulk crystallization mechanisms are spatially decoupled and, therefore, the corresponding DSC data can be analyzed separately. This has been done according to the Kolmogorov-Johnson-Mehl-Avrami model and also using the recently developed concept of local value of Avrami exponent n. For both the surface and bulk crystallization the phase transition process cannot be characterized by a single value of n. Observed variation of n with the crystallized fraction* is explained by a considerable variation of the nucleation rate that takes place during devitrification.
I. INTRODUCTION The quantitative description of isothermal devitrification by nucleation and growth is based on the concept first developed by Kolmogorov1 and later independently by Avrami,2 in accordance with empiricalfindingsby Johnson and Mehl.3 A convenient general form of the transformation equation is given below4: In(l-je)=i4(#n)
\rmIv{t-r)mdT,
(1)
where x is the volume fraction crystallized after time t, y is the growth rate, lv is the nucleation rate per unit volume, r is the nucleation induction period (or time lag), A(m) is a constant that depends on the growth topology, and m reflects the growth controlling mechanism as well as the dimensionality of this process. Usually at least part of the information necessary for evaluation of the integral in Eq. (1) is not easily available and, therefore, certain approximations are introduced. Frequently it is assumed that (i) the growth rate y is proportional to either first power or square root of time, depending whether the devitrification is interface or diffusion controlled, respectively, and (ii) the nucleation rate can be classified as either zero (crystallization on preexisting nuclei) or constant. The short but detectable nucleation transients can be effectively approximated by offsetting the time scale by the time lag r.5 Under " Also at the Metals Physics Unit.
J. Mater. Res. 3 (1), Jan/Feb 1988
http://journals.cambridge.org
these assumptions Eq. (1) simplifies to the Kolmogorov-Johnson-Mehl-Avrami (KJMA) form: x(O = l - e x p [ - A - ( f - r ) " ] ,
where A" is a thermally activated rate constant and n is the Avrami coefficient. Equation (2) has been widely used for analysis of isothermal crystallization data. By plotting ln[ — ln( 1 — x) ] vs ln(? — r) (the Avrami plot) one expects to obtain a straight line of t
Data Loading...
