Magneto-structure

We conducted a series of combined AFM/MFM measurements to study the magnetic properties of the martensitic twins in as-built and heat-treated Ni-Mn-Ga samples in publications III and IV. To allow a comparison, a heat-treated and electropolished single-crystalline sample with a 10M martensite structure (40° Ni50Mn28.5Ga21.5 alloy by AdaptaMat Ltd.) was heated to ~80 °C and then cooled to ambient temperature to create a random twin structure.

Figure 4.7a shows the MFM image of the 10M sample (two-variant structure) with the a- and c-axes perpendicular to the plane of view. A labyrinth-like magnetic domain structure is observable in the variants, with the c-axis perpendicular to the image plane. On this scale, the TBs appear as clearly distinguishable straight lines, and the twin variants with widths as low as ~1-2 µm exhibit a clear MFM contrast between the different twin variants. Figure 4.7b shows the MFM image of the 10M sample with the c-axis oriented approximately to the plane of view. A slight relief between the two variants visible in the corresponding AFM image – see publication III – indicates that the sample was not cut precisely along {001} lattice plane.

Figure 4.7c-d shows the MFM scans obtained from two sections of an as-built sample in publication III. The MSM scans show very narrow stripes, below 1 µm in width, exhibiting magnetic anisotropy (MFM contrast) that is considerably weaker compared to the 10M reference sample. Additionally, there is minor variation between the MFM contrasts between measurements, possibly due to minor compositional variation effects, L-PBF-induced internal stresses, or the differences in MFM contrast between different variants at different angles depending on the localized crystalline orientation of each grain.

Figure 4.7e shows a polarized light image of a top surface section of the sample homogenized at 1080 ℃ for 24 h in publication IV. A few areas of the sample with the largest features attributed to the twins were chosen for MFM scanning. The specified locations corresponding to the MFM images A1 and A2 are outlined with white squares.

The optical image shows martensitic twins with widths at the limit of the optical resolution and their orientation varying from grain to grain. This parallel stripe-like surface relief with contrasting cannot be observed in some areas, possibly due to the twins’ thickness lying under the optical resolution limit. Overall, the heat-treated sample exhibits drastically improved MFM contrast in comparison to the as-built sample. The MFM images reveal thick bands that correspond to the optical image and thinner bands with thicknesses down to the nanoscale. Hence, the thicker bands are composed of very narrow bands, which, in combination with the differences in MFM contrast between different variants at different angles, explains the unusually weak contrast obtained for the optical images; see Chapter 3 for reference. Additionally, the MFM image A1 shows visible signs of branching in the central band. Overall, the observed twin structure is consistent with polycrystalline Ni-Mn-Ga 14M martensites (Li et al., 2016), as also confirmed by the XRD measurement.

4.2 Characterization of as-built and heat-treated samples 53

Figure 4.7: MFM images obtained from: (a) a section of a single-crystalline 10M martensite reference sample showing a two-variant microstructure with the a- and c-axes perpendicular to the plane of view; (b) a section of the 10M sample with the c-axis oriented in the plane of view;

and (c-d) two sections of an as-built L-PBF sample (III-OPT) perpendicular to the build direction.

(e) Optical image (left) with polarized light contrast of an L-PBF sample homogenized at 1080℃

for 24 h (IV-11). The image was taken perpendicular to the build direction. Areas A1 and A2 correspond to the locations of the two MFM scans shown on the right. (Modified from publications III and IV.)

Table 4.5: Critical phase transformation temperatures and martensite lattice parameters, corresponding to the 14M, 10M and NM martensites, obtained for each sample in publications

III-V. The corresponding X-ray diffractograms were collected at ambient temperature. The lattice parameters are presented in the cubic coordinate system of the parent austenite unit cell

with an approximate accuracy of ± 0.01 Å and ± 0.01º. (Modified from publications III-V.) Martensite lattice parameters at 22 ℃ Transformation temperatures Publication

The approximate martensite lattice parameters obtained for the samples manufactured via L-PBF in publications III-V are summarized in Table 4.5. All lattice parameters are presented in the cubic coordinate system of the parent austenite unit cell. For the corresponding L-PBF process parameters and other properties, the reader is referred to Table 3.3 and Table 4.1. Figure 4.8 shows the X-ray diffractograms obtained for each sample in publication IV at ambient temperature. The identified peaks, corresponding to the 10M and 14M martensites, are indexed relative to the coordinate system of the parent austenite unit cell. The unindexed diffraction peaks originate from the modulated superstructure. Overall, the samples produced in publication IV showed very little lattice parameter variation from sample to sample, which was expected because the average compositions of these samples were nearly the same; see Table 4.1 for reference.

4.2 Characterization of as-built and heat-treated samples 55

Figure 4.8: XRD patterns obtained for each sample at ambient temperature in publication IV.

Intensities have been scaled and the baseline is offset. Recognized peaks of the 14M and 10M martensites are marked with grey dashed lines and indexed relative to the coordinate system of the parent austenite unit cell. (Modified from publication IV.)

The as-built sample IV-1 and the heat-treated sample IV-2 (atomic ordering treatment, no prior homogenization) showed a mixed crystal structure of 14M and 10M martensites with moderately high intensities of the 10M peaks at the angles of ~62.0° and ~62.5°.

This was expected as the as-built sample (III-OPT) in publication III also exhibited a mixed structure of 14M and NM martensites at ambient temperature. Additionally, the diffraction lines obtained for the as-built samples in both publications were generally broader and exhibited lower diffraction intensities in comparison to the heat-treated

samples. In publication III, it was suggested that these observations were likely caused by a localized variation of the lattice parameters, possibly due to the inhomogeneity of the chemical composition typically present in the anisotropic microstructures developed during L-PBF combined with L-PBF-induced internal stresses. This is supported by the fact that the samples’ chemical compositions do not correspond to those (Chernenko, 1999) that would typically exhibit co-existing martensite phases at ambient temperature.

Homogenization at temperatures above 1040 °C leads to chemical homogenization, which increases the stability of the 14M martensite at the cost of the 10M martensite.

Consequently, the intensity of the 10M lines is considerably reduced for all homogenized samples, and thus, the lattice parameters of the 10M martensite could only be obtained for samples IV-1 → IV-5. Finally, the samples homogenized at 1080 °C for 12–24 h showed only the diffraction peaks originating from 14M martensite at ambient temperature. Moreover, the variation and distribution of peak intensities within the measured samples suggested the presence of a crystallographic texture. However, the complex nature of the diffraction patterns produced by modulated martensites in multivariant polycrystalline samples, as produced by L-PBF in these publications, made a more thorough analysis of the crystallographic texture in publications III-IV unfeasible.