TiPolar: Microstructure-Informed Polarized Light Microscopy for Crystallographic Characterization of Additively Manufactured Titanium Alloys

Additive manufacturing of titanium alloys results in fine alpha laths in a basketweave microstructure within alpha variants that belong to elongated prior beta grains. The alpha laths’ average thickness is less than 1 micrometer, while the elongated prior beta grains can be as large as 25 mm. Statistically significant characterization of additively manufactured Ti6Al4V microstructures requires the quantification of size, shape, and crystallographic orientation of alpha laths and alpha variants. Process optimization requires the characterization of prior beta phase using reconstruction from measured alpha phase and the Burger orientation relationship. Thus, large area EBSD has been used to characterize additively manufactured Ti6Al4V to capture spatial and crystallographic orientations with requirements to cover with estimated scanned area of 20 mm x 10 mm at a 1 micrometer resolution. This could take up to 11 days on current EBSD systems. Local microstructure in additively manufactured Ti6Al4V has appreciable variance due to the changing thermal conditions based on local geometries that would require several scans be done on small samples (due to the size limitation of SEM chambers) to representatively measure the microstructure. Alternatively, the MRL team has developed a practical solution for additively manufactured Ti alloys that is based on microstructure-informed polarized-light microscopy—TiPolar^TM. At an effective 200,000 pixels per second with 0.5 micrometer resolution (1000 times faster than EBSD), a 20 mm x 10 mm will be scanned in about 60 mins. In addition, a full slice of an additive part averaging 6”x 6” can be scanned without the need for vacuum. The technical limitation to only structures with optical anisotropy such as hexagonal close packed crystals and the ability to calculate only two of the Euler angles will be discussed in the presentation. Additionally, the data analytics processes to generate texture information, as well as larger scans will be discussed.