The material used in the tests was AISI 6061-T6. This aluminum alloy is a medium to high strength heat-treata- ble alloy with a very good corrosion resistance and weldability. It has medium fatigue strength. T6 means solution heat treated and artificially aged. This alloy has applications in different industries such as helicopter rotor skins, tubing, truck frames, boilermaking and rivets. The alloy specimens were ground flat with 180-grit silicon carbide paper. The initial R_a roughness average value was 6.35 nm (R_a) with a standard deviation of 0.15, after 10 measurements. Figure 1 shows two 3D profiles of the initial roughness obtained using an AFM (Microscope diMultimode V, Vecco, Controller di Nanoscope V) on random locations of the specimen surface. The projected surface areas were 10 × 10 µm² and 5 × 5 µm², respectively.

Figure 2(a) and Figure 2(b) shows a micrograph of the specimen before erosion tests with different magnifications. The EDS analysis was performed to obtain the chemical composition of the tested material and the results presented the following elements, carbon (C) 0.023 wt%, silicon (Si) 0.80 wt%, copper (Cu) 0.40 wt%, oxygen (O) 5.94 wt%, magnesium (Mg) 1.19 wt% and aluminum (Al) 69.3 wt%. The Al samples had a Berkovich hardness of 126 with a standard deviation of 3.66 and a Young Modulus of 90 GPa with a standard deviation of 2.33 after 15 measurements. The applied load was 100mN. These results were obtained using a Nano indentation Tester (TTX-NHT, CSM Instruments).

The samples had a rectangular shape with dimensions of 50 mm × 25 mm and 3 mm in thickness. The abrasive particles used were aluminum oxide (alumina) with an angular shape as shown in Figure 3(a) and Figure 3(b). The chemical composition of the abrasive particles presented elements such as oxygen (O) 10.31 wt%, fluorine (F) 4.48 wt%, aluminum (Al) 22.33 wt%, silicon (Si) 3.47 wt%, chromium (Cr) 9.63 wt%, iron (Fe) 45.11 wt% and nickel (Ni) 4.67 wt%. Particle size distribution was obtained using an Analysette 28 image sizer. The particle size grain was between 300 and 400 µm (Figure 3(c)). Most particles were found in this particular range. The hardness and density of the abrasive particle is 1800 HV and 3.99 g/cm³ [8] .

The erosion tests were performed by using a rig, which is based on that in the ASTM G76-95 [9] . The schematic diagram of the rig developed is presented in Figure 4. In this rig, the abrasive particles are accelerated from a nozzle by using a compressed air stream that caused them to impact the surface of the material. The average particle velocity was measured with an opto-electronic flight-timer similar to that described by Kosel and Anand [10] . This system offers the possibility to measure the particle velocity in an accurate mode and the design does not request high costs.

The material was eroded for 20 min although each sample was removed every 2 min to determine the mass loss. The impact angles used for the tests were 30˚, 45˚, 60˚ and 90˚. These angles were selected to evaluate the material at low and high impact angles. A particle velocity of 24 ± 2 m/s and an abrasive flow rate of 63 ± 0.5 g/min were used to conduct the tests. In all of the tests, the specimens were located 10 mm from the end of the stainless steel nozzle. The nozzle had the following dimensions: 4.7 mm internal diameter, 6.3 mm external diameter and a length of 260 mm. The room temperature was between 35˚C and 40˚C. The specimens were weighed using an analytical balance (with an accuracy of ±0.0001 g) before the start of each test and removed every 2 min, cleaned by using acetone and weighed again to determine the amount of mass loss. Micrographs of the eroded surfaces were obtained using a Scanning Electron Microscope (SEM) Quanta 3D FEG (FEI) to analyze the specimens and to identify the possible wear mechanisms.

The erosion damage at 30˚ was characterized by high plastic deformation on the surfaces as observed in Figure 5 and Figure 6. In this particular case, micro-cutting and micro-ploughing (grooves on the surface) actions were

normally seen as observed in other erosion studies with pure aluminum or aluminium alloys [11] - [15] . Some particle fragments were embedded on the surfaces. The alumina particles were probably fragmented during the impact action, due to work-hardening of the aluminum surface inflicted by the permanent impacts and other aspect to consider could be that some particles made contact with embedded fragments and then were fractured. In Figure 6(a) and Figure 6(b), it is possible to observe a particle embedded on the surface with material piled-up in front of it. In addition, Figure 6(c) and d shows EDS spectra in two different points of the damaged surface. Here, the elements that appear in the chemical analysis are carbon (C), oxygen (O) and in the embedded particle a new element can be identified, Al, corresponding to the aluminum oxide. It is possible that a considerable amount of particles could be inserted on the surfaces due to these have high hardness combined with the soft surface of the aluminum alloy.

In respect to the wear damage at 90˚, the wear mechanism was high plastic deformationas that observed at 30˚. Here, it was possible to identify micro-ploughing and micro-cutting actions, embedded particles and wear debris deposited on the damaged surfaces (see Figures 7(a)-(c)). The wear mechanisms are quite consistent because of the soft structure that this aluminum alloy exhibited. The EDS analysis of Figure 7(d) confirmed the presence of another embedded abrasive particle by the elements found. These were carbon (C), oxygen (O) and aluminum (Al).

The volume loss was obtained dividing the mass loss (g) by the density of AISI 6061 (2.70 g/cm³). Two tests were conducted for each incident angle and an average volume loss and standard deviation was obtained. The maximum volume loss (mm³) was reached at 30˚, where the micro-cutting and micro-ploughing actions caused an increase in the material removal. On the other hand, the least volume loss was seen at 90˚. All the results were really close in this particular case. Figure 8 shows the graph of the volume loss against the elapsed time. In all the incident angles, the volume loss was increased in relation to the time.

The total erosion value was also calculated and the results are presented in Figure 9(a) and Figure 9(b). It

was obtained dividing the volume loss (mm³) by the total mass of particles (g) that impacted the surface after 20 min as indicated in ASTM G76-95 [9] . As mentioned previously, the maximum erosion rate was at 30˚ reducing progresively up to 90˚. Here, it is possible to conclude that this material in these particular testing conditions exhibited a ductile type behaviour based on the erosion literature graph presented in different erosion studies [14] [15] . This is a typical behavior of an aluminium alloy because of its high ductility, tenacity and high capacity to be deformed.