The fact that both outer Cr2N layer and inner Cr-carbide layer were present in the specimen surface at 0 h coating time suggests that these two layers started to form simultaneously during heating stage and grew in thickness as coating time increased (Fig. 5). Detailed analysis showed that the total thickness of these two layers dT, i.e., dT = Cr2N layer thickness + Cr-carbide layer thickness, increased linearly with square root of coating time t or t1/2, although the slope of this linear relationship changed at approximately t = 2.7 h (Fig. 6a), suggesting that the growth of this total layer thickness was controlled by a diffusion mechanism. This is consistent with the coating growth mechanism observed in general for coatings formed in pack cementation process
However, the growth behaviour of individual layers differed as shown in Fig. 6b, which plots the layer thickness versus coating time t for individual Cr2N and Cr-carbide layers. The growth of the outer Cr2N layer continued until coating time reached 4 h, after which the growth of this layer stopped. For the inner Cr-carbide layer, its thickness decreased abruptly at a coating time between 2 h and 4 h. Both below and above this coating time, the inner Cr-carbide layer thickness increased continuously with coating time. It can be seen that the inner Cr-carbide layer continued to grow at coating times longer than 4 h, which were the times when the outer Cr2N layer had stopped growing. Therefore, it seems that Cr atoms deposited from the vapour phase in the pack can diffuse across the outer Cr2N layer to enable the continued growth of the inner Cr-carbide layer even when the growth of the outer Cr2N layer has stopped.
It is of importance to take notice of another feature of individual layer growth in the coatings, i.e., at coating times shorter than 2 h, the thickness of the outer Cr2N layer was smaller than that of the inner Cr-carbide layer, but at coating times longer than 4 h, the order was reversed (Fig. 6b). This change in layer thickness order happened at a coating time between 2 h and 4 h, which coincided with the coating time interval in which the inner Cr-carbide layer thickness showed a sudden decrease as described previously. It is thus highly likely that both of these two phenomena occurred at the same coating time. These phenomena may be accounted for by the same process, i.e., the change in the phase of the inner Cr-carbide layer as coating time increased. As discussed previously, the phase of inner Cr-carbide layer was (Cr, Fe)7C3 in the specimen coated for 2 h, but it changed to (Cr, Fe)23C6 in the specimen coated for 8 h as shown in Fig. 7, which was measured from a surface after manually grinding away the outer Cr2N layer (however, a strong peak at 40.506° from (002) of Cr2N was still present, indicating that the outer Cr2N layer was not totally removed after grinding in this case). It is possible that this phase change in the inner Cr-carbide layer took place at a coating time between 2 h and 4 h, which not only caused the two phenomena discussed above but also the change in the slope of the thickness dT versus t1/2