Developing laser powder bed fusion ti6al4v-reinforced nanoparticulate composites for aerospace applications

Abstract

The aim of this research study was to investigate the reinforcement of titanium alloy Ti6Al4V(ELI) using nanoparticulate materials. Initially, titanium diboride (TiB2) and carbon nanotubes (CNTs) were selected as suitable materials for reinforcement. However, due to the unavailability of titanium diboride in the local market, the focus was turned to CNTs due to their availability locally and their excellent properties of strength, hardness, and Young’s modulus. The original intention was to have individual CNTs attached to Ti6Al4V(ELI) particles. However, agglomeration of the CNTs prevented them from successfully attaching onto Ti6Al4V(ELI) particles. Therefore, reinforcement was carried out using CNT agglomerates dispersed in Ti6Al4V(ELI). Mixing of the two powders was carried out in three separate stages using different mixing techniques for six different volume fractions of CNTs, including 3 %, 8 %, 15 %, 20 %, 25 %, and 30 %. Analysis of the mixed powder was then carried out using scanning electron microscopy (SEM) and optical microscopy used for the built tracks. The first stage of preliminary mixing of the Ti6Al4V powder with the CNTs was carried out at two separate levels. In the first level, the two powders were mixed mechanically. In the second level, acetone and ethanol were used as dispersion agents. Dispersion of CNTs in Ti6Al4V(ELI) at both levels proved to be insufficient. However, the use of the two dispersion agents provided better dispersion compared to the dry mixing of powders. Furthermore, a comparison of the effectiveness of the two dispersion agents showed that ethanol provided better dispersion compared to acetone. Determining the effectiveness of dispersion based on the size (area and perimeter) and number of the agglomerates in each sample, it was concluded that while using acetone as a dispersant lowered the average area of agglomerates, the use of ethanol led to a wider dispersion of CNTs with a marginal increase in the average areas of CNT agglomerates at increased % volume CNTs and was thus preferable. Thus, the effectiveness of dispersion was achieved where ethanol was used compared to acetone and manual mixing. These results were quantified using a continuous statistical approach in which the agglomerates in each sample were measured. Using this approach, the agglomerates in the ethanol samples generally had smaller sizes with respect to area and perimeter as opposed to the agglomerates found in the acetone and mechanically mixed samples. The second stage of mixing was more rigorous and included manual mixing, turbula mixing, and mechanical-turbula-further manual mixing (hereafter referred to as combined mixing), which was done separately in this order. Combined mixing produced better dispersion and separation of CNTs. However, it also brought about defects in Ti6l4V(ELI) particles occurring from the increased manual mixing process applied in an effort to separate the CNT agglomerates. Turbula mixing was the least effective and had the greatest separation of the two powders due to the rotary mixing process inherent in it and the difference in densities between the two materials. Manual mixing was much improved from the one used in the preliminary stage due to the use of a different tool. This tool had a larger surface area than the initial tool used, providing better contact between powder particles and thus improved dispersion. The results showed that the effectiveness of mixing was better achieved at the low vol.% of CNTs in Ti6Al4V(ELI) of 3 % and 8 %. Increasing the vol.% of CNTs in Ti6Al4V(ELI) above these values led to increased agglomeration of CNTs in the mixture. The presence of single CNT fibres was observed to be more prevalent at low volume fractions of 3 % and 8 %. Although single CNTs were observed at a lower volume of CNTs, this was not at a large scale. This showed that a much-needed technique for the separation of CNTs is required. Sodium dodecyl sulphate (SDS) was used as a surfactant in the final mixing stage. The same techniques used in the second stage of mixing were employed in this stage. In manual mixing, SDS achieved good dispersion of CNTs in Ti6l4V(ELI). However, significant agglomeration of CNTs occurred at 8 % and 30 % vol. of CNTs. The case of agglomeration at 8 % vol. of CNTs in Ti6Al4V(ELI) is an anomaly as it is expected only at higher content of CNTs. This mixing technique produced defects in the particles of Ti6Al4V(ELI) at 8 % and 25 % vol. of CNTs in Ti6Al4V(ELI). The defects occurred on the surfaces of Ti6Al4V(ELI) particles at 8 % and 25 % vol. of CNTs in Ti6Al4V(ELI). When turbula mixing was used, agglomeration of CNTs increased from 15 % to 30 % vol. of CNTs in Ti6Al4V(ELI). The number of clusters of CNTs and Ti6Al4V(ELI) particles increased at 25 % and 30 % vol. of CNTs in Ti6Al4V(ELI), likely due to the combination of SDS and increased rotational speeds of the mixer which provided more particle-particle and CNT interaction. Combined mixing produced clusters of tightly packed Ti6Al4V(ELI) particles and CNTs, which were most prominent at 8 % and 15 % vol. of CNTs in Ti6Al4V(ELI). At 20 %, 25 %, and 30 % vol. of CNTs in Ti6Al4V(ELI), increased agglomeration of CNTs was observed. The combined mixing technique produced the most particle defects of the three techniques used at this stage, which were observed in different forms at 3 %, 8 %, 20 %, 25 %, and 30 % vol. of CNTs in Ti6Al4V(ELI). The use of SDS produced samples with enhanced CNT attachments to Ti6Al4V(ELI) particles. The known optimum laser scanning speed and scanning power of 0.6 ms-1 and 100 W, respectively, of Ti6Al4V(ELI) printed on an M2 LaserCusing® machine from Concept Laser GmbH, were used as a starting point to develop a matrix of laser scanning speeds and scanning powers for the nanocomposite. This was done based on the thermophysical properties of the two constituents and the Reuss rule for composites. Printed single tracks based on this matrix were analysed to determine the best process parameters of laser power and scanning speed from the best ones amongst them. Stable tracks were obtained at a laser power of 60 W and 80 W. However, the geometrical features produced track builds with aspect ratios above 0.5. The formation of keyholes became dominant in tracks built at laser powers of 100 W to 180 W, and the formation of pores increased for values of laser power of 160 W to 180 W. Linear energy density showed an inverse linear relationship with the scanning speed, the aspect ratio showed an initial decrease followed by an increase for the highest scanning speed. This is a behaviour similar in trend to those observed at a laser power of 60 W somewhat, and 100 W, 140 W and 160 W. The curves obtained at laser powers of 80 W and 120 W were inverted. This difference in behaviour warrants further investigation. Nonlinearity of laser energy density (LED) was observed between laser powers of 60–140 W; however, linearity occurred at laser powers of 160–180 W, which warrants further research. For all the other values of laser power used in this work, apart from 60 W, the concave upwards curves of aspect ratio and the fact of the minimum values of aspect ratio being higher than the threshold value of 0.5 imply that optimum process parameter sets of laser power and scanning speed do not exist. Further research is thus required to establish the best processing parameter sets for processing.