Silicon carbide (SiC) is used as the main diffusion barrier to prevent the fission products (FPs) from escaping in high temperature reactors (HTRs). It retains most of the FPs quite effectively with the exception of silver, strontium and europium. There have also been reports on the reactions between some FPs and SiC, raising some concerns on the integrity of SiC as a coating layer and questioning the ability of SiC as the main diffusion barrier. An additional protective layer of tungsten (W) is proposed to cover the SiC and probably reduce the interaction of FPs with SiC. Coating of SiC layer with W will assist in improving the shielding effect, which will allow for high burn up and enrichment without degrading the SiC. W coatings on SiC are also used for device fabrication. (Thus this study will benefit both semiconductor and nuclear application.) The study was conducted by sputter depositing W metal thin films on 6H-SiC at room temperature (RT). The effect of thermal annealing in vacuum, hydrogen (H2) and argon (Ar) of the W thin film deposited on 6H-SiC was investigated as a function of annealing temperature. The resulting solid-state reactions, phase composition and surface structural modification were investigated using Rutherford backscattering spectrometry (RBS), grazing incidence X-ray diffraction (GIXRD), scanning electron microscopy (SEM) and atomic force microscopy (AFM). The thickness of the as-deposited layer obtained from RUMP simulations was about 73.8 nm and was composed of about 63.4 at.% W and 36.6 at.% O. The oxygen was in a form of tungsten oxide (WO3) mixed in the W thin film. The SEM and AFM images of the as-deposited samples showed that the W thin film had a uniform surface with small grains. The surface roughness (Rrms value) of the as-deposited sample was 0.4 nm. The samples where annealed from 700 °C to 1000 °C for 1h in vacuum, hydrogen (H2), and argon (Ar). From the RBS results, the initial reaction for vacuum annealed samples occurred at 850 °C, for H2 annealed samples it was 700 °C and the Ar annealed had an initial reaction at a temperature lower than 700 °C. In all the annealing atmospheres carbon (C) was found to diffuse faster than Si into the W metal. After this C diffusion reached equilibrium, Si also migrated into the W metal. A reduction of oxygen upon annealing was observed for the vacuum annealed samples. Removal of oxygen was observed for the H2 annealed samples, while oxygen was seen to diffuse to the reaction zone (RZ) for the Ar annealed samples. The phases observed from GIXRD at 700 °C for vacuum annealed samples were CW3 and WC, for H2 samples W5Si3 and WC and for Ar annealed samples W5Si3, WC, SiO2 and W2C. The formation of WSi2 and W2C was observed at 800 °C for H2 samples and 900 °C for vacuum samples. The segregation of Si towards the surface at 1000 °C for H2 samples resulted in the formation of SiO2. The results showed that annealing in different atmospheres reduces the initial reactions and phases formed. SEM and AFM revealed that the samples annealed in Ar were rougher than the vacuum and H2 samples, while the vacuum annealed samples were rougher than the H2 annealed samples. The Rrms of the samples annealed in different atmosphere followed the order: Ar ˃ Vacuum˃ H2. From the SEM and AFM images, the H2 annealed samples at 700 °C were composed of small granules which increased with annealing temperature resulting in the formation of distinct grain boundaries. The samples annealed in Ar at 700 °C were composed of big crystals which were randomly orientated. Increase in annealing temperature for the Ar samples resulted in the parasitic growth of the crystals, which is in line with Wulf’s law. The samples annealed in vacuum at 700 °C formed tungsten oxide nanowires on the W metal surface, with the W metal in a form of granules. Annealing at high temperatures resulted in the removal of the tungsten oxide nanowires on the W metal surface and parasitic growth of the crystals. The difference in the crystal growth observed during the vacuum, H2 and Ar is explained by a crystal growth model.