A complex experimental set-up was built to study the impact of liquid alumina droplets on different substrates (stainless steel 304 L, sintered alumina, carbon carbon) kept at temperatures up to 2100 K. The impact behavior: rebound, deposition, splashing, spattering was systematically studied as well as the resulting splat shapes. The set-up consists in a controlled atmosphere chamber where molten alumina particles with diameters between 10 and 90 mum, are produced by a d.c. plasma torch, substrates being heated by a second d.c. plasma torch. In such conditions, it was possible to achieve particle temperatures between 2300 and 4200 K with velocities in the range 50 to 300 m/s. The particle behavior at impact was characterized by the Sommerfeld parameter K (K=We(1/2) Re-1/4 We and Re being respectively the Weber and Reynolds numbers of impacting particles). It was possible to vary K between 3 and 1300. Low K values were obtained by tilting the substrate up to 60degrees. The parameters of a single particle at impact were measured: its velocity v(p) and diameter d(p) by Phase Doppler Anemometry (Deltav(p) = 5%, Deltad(p) = 10%) and its temperature T-p by fast (100 ns) two color pyrometry (DeltaT(p) = 15%). The particle impact was visualized by a fast camera coupled to a microscope (exposure delay time 50 ns...100 ms) with complex synchronization and light intensity problems. To solve the latter, the impacting particle had to be illuminated with a 2 W c.w. Ar+ laser at 488 nm. Unfortunately, the controlled atmosphere chamber did not allow to change the substrate after each particle impact. Starting from a smooth surface for the first impact, due to the successively deposited splats, rapidly droplets impacted on a rough surface (Rasimilar to5 mum). For splats collected on a hot alumina substrate (2100 K), where flattening is completed before solidification starts (case similar to that of ethanol droplets on cold copper) deposition occurs for K between 4 and 90 while splashing occurs for K as low as 30. These results are slightly different from those related to the ethanol droplet for which deposition occurs for 357.7. This could be due to the precision of measured values and the rough surface. For splats collected in spraying conditions splashing is always the rule K values up to 1400) especially on rough surfaces. However the particle impact velocity and temperature, the substrate temperature and tilting plays an important role on the resulting splat diameters, distortion and elongation rates. The question which is still pending is which quantity of splashed material is incorporated within the constructing coating and how does it affect its thermophysical properties.