ELSEVIER Journal of Non-Crystalline Solids 202 (1996) 290-296
]OURNAL OF
Avrami exponents for transformations producing anisotropic
particles
Michael C. Weinberg *, Dunbar P. Birnie III
Arizona Materials Laboratory, Department of Materials Science and Engineering, University of Arizona, Tucson, AZ 85712, USA
Received 18 December 1995
Abstract
The behavior of Avrami exponents (AE) for systems which produce anisotropic particles (for which shielding effects and
phantoms occur) is examined. Probabilistic arguments used in previous studies are applied to derive expressions for the AE
as a function of time or the extent of transformation. Numerical calculations for the one dimensional (1D) cases of site
saturation and continuous nucleation show that blocking effects produce a minimum in plots of AE versus extent of
transformation. The influence of growth rate anisotropy upon the magnitude of the reduction in the AE is also examined and
certain features of the behavior of the AE for 2D and 3D systems are discussed.
I. Introduction
The goal of the 'formal theory of transformation
kinetics' [1] for nucleation and growth transforma-
tions is the description of the fraction of material
transformed at a given time in terms of the rates of
the latter processes. The Johnson-Mehl-Avrami-
Kolmogorov (JMAK) theory [2-6] is generally used
for this purpose. Furthermore, JMAK theory has
been used extensively for the interpretation of crys-
tallization experiments and for inferring crystalliza-
tion mechanisms from the results of such experi-
ments. Quite often these analyses have relied upon
the use of the Avrami exponent (AE) described
below.
* Corresponding author. Tel.: + 1-520 621 6070; fax: + 1-520
621 8059; e-mail: migs@ccit.arizona.edu.
For a class of isothermal crystallization processes
the fraction of material crystallized after a time t,
X(t), can be written as
X(t) = 1 - exp{-kt"}, (1)
where k is a constant and n is the AE. The value of
n depends upon the crystallization mechanism. For
example, for continuous nucleation and 3D crystal
growth n = 4. Other cases are discussed in Ref. [1].
Thus, if one determines X(t) experimentally and
uses the data to construct a plot of In{ - In[1 - X(t)]}
versus In t, then from Eq. (1) it should be linear and
possess a slope given by the AE. From the determi-
nation of the AE the crystallization mechanism is
inferred.
Eq. (1) is only valid for isothermal crystallization
although it has been invoked frequently for the inter-
pretation of non-isothermal DTA/DSC experiments.
Furthermore, it is known that Eq. (1) must be modi-
fied (or is totally erroneous) for certain types of
isothermal crystallization processes, too. Examples
0022-3093/96/$15.00 Copyright © 1996 Elsevier Science B.V. All rights reserved.
PH S0022-3093(96)003 89-4