A novel approach for fabricating thick cubic boron nitride coatings on silicon and tungsten carbide-based cutting tools

Abstract

The major aim of this project is to examine the effectiveness of several deposition schemes for the fabrication of thick boron nitride films with high volume fraction of the cubic phase (cBN) and acceptable adhesion to substrates. The project importance is not only in the fact that cBN-rich coatings are more superior to diamond films in wear resistance coatings on cutting tools, but also in problems associated with the delamination of cBN-rich coatings thicker than 200 nm. Firstly, cBN-rich film has high internal stress. Secondly, a soft sp2-bonded layer is usually formed at the early stage of the growth, which forms a weak linkage between the cBN-rich layer and its substrate. Thirdly, the cBN-rich may contain some excessive boron atoms, which accelerate the oxidation of the film material. We fabricated a series of cubic boron nitride films based on dual ion beam technique. In this technique, one ion gun was used to sputter a boron (B) target, while another ion gun was used to generate an assist ion beam containing nitrogen (N) and argon (Ar) ions to bombard the growing film surface. Two types of substrates were used, i.e. (100) single crystal silicon and tungsten carbide (WC). Four groups of film samples were prepared with the following schemes: Group 1: Single-step process to fabricate BN films on a Si substrate Group 2: Low temperature deposition of cBN-rich layer on gradient BN buffer layer / Si substrate. Group 3: Multi-layered films with a layered structure of cBN-rich layer / gradient BN buffer layer / Si substrate grown at 680C. Group 4: Multi-layered films with a layered structure of cBN-rich layer / gradient BN buffer layer / metal (Zr or Ti) / WC plates deposited at 680C, followed by post-annealing at 850C for 1 hr. and irradiation by N2+ions for 10 min. Firstly, the structure of the samples of Group 1 and 3 were compared to illustrate the importance of the addition of a gradient BN buffer layer to the adhesion of cBN-rich films on Si substrate. Secondly, the process for fabricating the samples of Group 2 was designed aiming at growing cBN-rich films at low temperature as the first trial of multi-layered structure. Thirdly, samples of Group 3 were fabricated to examine how thicker cBN-rich layers with satisfactory adhesion can be achieved by the modification of gradient BN buffer layers and the use of appropriate assist ion energy. The process for fabricating samples of Group 4 was designed aiming at further improving the adhesion of cBN layers on WC substrates. In particular, the post-annealing process at 850C was supposed to initiate the diffusion of Zr or Ti atoms into the soft gradient BN buffer layer, and reacted with B and N atoms in the layer to create boride or nitride. This layer, if formed, is expected to replace the original softer sp2-bonded buffer layer, and thus enhances the cross-linking between the top cBN-rich layer and the substrate. Results show that the addition of a buffer layer is crucially important in obtaining acceptable adhesion of cBN-rich layer on Si or WC. Otherwise the top cBN-rich layer would detach from the substrate once the thickness of the layer exceeds 183 nm. For the low temperature deposited BN layers of the samples of Group 2, sp2-bonded structure dominates. The cBN phase cannot grow without substrate heating. However, nano-sized sp3-bonded clusters were found to be formed in these layers. The presence of the nano-sized sp3-bonded clusters is supposed to be the reason for the layer to have a hardness of 23 GPa, which is much higher than that of the pure sp2-bonded structure. After adding substrate heating and a gradient BN buffer layer with a thickness of 492 nm, a top cBN-rich layer can be fabricated to reach a thickness of 643 nm. This top layer does not peel off since the day of deposition. The volume fraction of the cBN phase in the top layer can reach a level of 87 %. The hardness and elastic modulus of the top cBN-rich layer can be as high as 57 GPa and 612 GPa respectively. The internal compressive stress of this sample is 2.8 GPa. More interestingly, it is demonstrated clearly that after post-annealing, Zr atoms can diffuse into the gradient BN buffer layer and react with B and N atoms.