*2.1. Sample Preparation*

The samples were manufactured at Irepa Laser using a CLAD®machine (Magic 800, BEAM company, Strasbourg, France) under a neutral argon atmosphere using a Ti6Al4V support plate. In this technology, the powder is injected into a laser beam and melts before falling onto the substrate. The machine can be fed simultaneously by different powder feeders, so it becomes possible to real-time control the chemical composition of the powder mixture. The size distribution of the powder particles measured by laser diffraction analysis is given in Table 1. As sketched in Figure 1, the manufactured gradient materials contain respectively 0, 25, 50, 75 and 100% of Nb or Mo and, conversely, 100, 75, 50, 25 and 0% of Ti or Ti6Al4V. The size of the parts is also given in Figure 1a. The powders and manufacturing data (power, speed, pitch, etc.) of the samples have been detailed in the works of Schneider et al. [34,35].


**Table 1.** Size distribution of the different powders in μm [34,35].

The FGM walls were cut into 1.5 mm thick slices. To take into account the effects of the surface roughness, the samples have been divided into three categories: polished (P), SMATed (S) and SMATed then polished (S + P). The P samples were polished using OPS (Oxide Polishing Suspension, Struers, Champigny-sur-Marne, France) to obtain a mirror-like surface. Samples S and S + P were SMATed for 5 min (frequency 20 kHz, amplitude 60 microns) with 2 mm spherical shots made of 100C6 steel. The shot blasting time was deliberately chosen to be short enough to minimize any possible transfer of chemical species coming from the shot-balls or tooling that are known to pollute the surface under SMAT [21,36–40]. This type of pollution, which is known to develop for long term peening, has been demonstrated to affect the corrosion properties of some parts [39,40] can indeed also have an effect on biological tests. After SMAT, the S + P samples were polished with OP-S over 2 or 3 μm to modify the roughness issued from the SMAT process while remaining in the area where the microstructure was modified. This experimental procedure using FGMs coupled with surface SPD as well as polishing has several advantages. While the comparison between the S and S + P samples makes it possible to study the effect of the roughness created by the SMAT for a given type of refined structure, the comparison between the P and S + P samples authorizes to analyze the effect of the microstructure modification by SPD for a mirror-like surface. In addition, the comparison of the response along the same sample can be used to depict the effect of the chemistry as well as the comparison between Ti-Nb and Ti6Al4V-Mo can be used to find the most biocompatible alloying element.

**Figure 1.** (**a**) Size and chemical composition of the samples for the Ti/Nb and Ti6Al4V/Mo couples; (**b**) complete map of the Ti6Al4V-Mo FGM after reconstruction from Schneider et al. [35]; (**c**) example of the interface Ti6Al4V/Ti6Al4V + 25% Mo before (**c1**) and after beta reconstruction (**c2**).
