It is expected that the complex oxides in this study would exhibit an even lower melting point than 1083 ◦C, which is in accordance with the suggestion outlined by Klimenko et al. [21] that the compounds on the wear surface would possess a lower melting point than that of the chip. It is deduced that the oxides are readily melted and thereby lead to the wetting of tool surface. The layer thus generated may result in adhesion of the binder compound [12] and hence it would adhere on the tool simultaneously. It is noted that this layer is semi- liquid rather than in mobile state [14], hence it may not wet the tool face completely. Instead, the oxides are deposited on the tool surface as suggested by Barry and Byrne [9], and hence the layer in Fig. 8(b) is seen not very smooth. This layer could work as diffusion and thermal barriers, and it is beneficial to tool life. The same proposition had been suggested by others as well [9,14], and this protective effect is similar to that of the oxide layer formed on P-type carbide tools in the machining of Ca-treated steels [22]. When the speed is further increased, the adhesive layer could increase the friction force. At very high cut- ting speed such as V = 130 m/min, the layer cannot withstand the high friction force any more, and it is apt to be worn away from the tool face. Since at this moment the temperature increases significantly which weakens the bond between the hard particles of the cutting tool, the hard particles would be plucked out of the tool face via adhesion (Fig. 7c). Similar viewpoint was proposed by Chou et al. [12]. Detachment of hard particles on tool face can be seen more clearly in Fig. 10 where certain part of Fig. 8c is magnified. The rough tool surface could contribute further increase of friction force.