a b s t r a c tAnalysis of recent experimental investigations, in particular by transmission electron microscopy, suggests spheroidal graphite grows by 2-D nucleation of new graphite layers at the outer surface of the nodules. These layers spread over the surface along the prismatic direction of graphite which is the energetically preferred growth direction of graphite when the apparent growth direction of the nodules is along the basal direction of graphite. 2-D nucleation-growth models first developed for precipitation of pure substances are then adapted to graphite growth from the liquid in spheroidal graphite cast irons. Lateral extension of the new graphite layers is controlled by carbon diffusion in the liquid. This allows describing quantitatively previous experimental results giving strong support to this approach.
IntroductionGraphite spheroids in spheroidal graphite cast iron are known to consist of piling up of graphite layers having their c axis oriented parallel to the spheroid radius. To accommodate for the change in orientation in the tangential direction, the spheroids appear to be divided in adjacent sectors which are easily observed by optical microscopy as illustrated in Fig. 1. The change in orientation of the c axis at the boundary between two sectors may amount from 10 to several tens of degrees, while the orientation changes within sectors are more limited [1,2].It has long been recognized that graphite layers within spheroids are arranged in growth blocks which are elongated along the prismatic a direction of the graphite structure [3]. This would imply that graphite grows along the prismatic (tangential) direction during spheroidal growth even though the overall (apparent) growth direction is the radial one. Two types of models have been suggested in the past to account for this tangential growth: i) those based on screw dislocations [4] or cone-helix growth [5e7]; and ii) those based on the continuous growth of a graphite layer folding around the spheroid [8,9]. In the first type, tangential growth proceeds around screw dislocations or cones emanating from the spheroid's centre (Fig. 2-a), giving features that would agree with the observation of sectors. However, transmission electron microscopy (TEM) observations have shown that the orientation of the c axis along a sector tilts at random and in either ways [10] which implies that continuous growth around a screw dislocation did not occur. The second type of models (Fig. 2-b) would hardly explain the formation of sectors as already stressed by Gruzleski [9]. However, a slight modification where nucleation of new layers proceeds at the step between two neighbouring sectors which was suggested by Double and Hellawell [11] would (Fig. 2-c).There has been a renewed interest these last years for investigating the growth mechanism of spheroidal graphite [12e14]. Qing et al. [14] report observation of defects and dislocations in graphite, but do not seem to have observed long range ordered arrangements of these defects that would support the ...