Dataset DOI: 10.5061/dryad.xsj3tx9tn

Dataset overview

This study investigated the interactions between martensitic phase transformation and plastic slip in superelastic Nickel-Titanium (NiTi, or nitinol). The experiment was designed to be analogous to the early-stage duty cycle of a nitinol cardiovascular implant. The dataset comprises the following:

1. SEM-DIC (HR-DIC): high-resolution, sub-grain deformation maps obtained via in situ scanning electron microscopy digital image correlation (SEM-DIC, sometimes called HR-DIC)
2. EBSD: orientation maps obtained via electron backscatter diffraction (EBSD)
3. MMICROSTRUCTURE: activity fields for each slip system and martensitic variant reconstructed via the MMICROSTRUCTURE framework, which is developed and presented in the accompanying publication

Full experimental details are available in the linked publications.

File structure

SEM-DIC (HR-DIC)

Nanopatterned tensile samples were loaded in situ in an FEI Teneo FEG-SEM with a Kammrath and Weiss tensile-compressive module. At different macroscopic strain levels, loading was paused for image acquisition. Each file is labeled with a string "e_", which refers to the macroscopic strain:

• Cycle 0: e1 (2.5% strain), e2 (5.5% strain), e3 (8.5% strain)
• Cycle 1: e4 (3.5% strain), e5 (5.5% strain)
• Cycle 2: e6 (3.5% strain), e7 (5.5% strain), e8 (unloaded)

sem_dic_images.zip contains the raw SEM image data from the tensile test.

• Tiled images were acquired in a 7x7 grid at 20 kV and 1.6 nA, with a 3.2-microsecond dwell time.
• SEM images are (4096, 4096) 16-bit grayscale TIF, with a resolution of 19 nm/px. Images were captured with a custom scan controller (DOI: 10.1016/j.ultramic.2018.08.025) to avoid beam artifacts.
• Each of the subfolders (e0-e8) contains the tiled images from one of the macroscopic strain levels described above. Each image is named according to both the strain index and the row/column indices; for example, "e3_r0c1.tif" should be read as "strain level 3, row 0, column 1".

e1.mat, e2.mat, e3.mat, e4.mat, e5.mat, e6.mat, e7.mat, e8.mat: these files contain the stitched displacement and deformation fields formatted as Matlab structs. 

• Displacements are "u" and "v"; positions in the original images are "x" and "y"; Lagrange-Green strains are "exx", "eyy", and "exy"; displacement gradients are "fxx", "fxy", "fyx", "fyy"; the correlation confidence index is "sigma"

• Displacements were obtained by digital image correlation in VIC-2D, using 25-pixel (475 nm) subsets and 5-pixel (95 nm) step size. DIC was performed with first-order shape functions and the normalized SSD error metric. Each image tile was correlated independently.
• Strains and displacement gradients were calculated in VIC-2D with 5-pixel Gaussian prefiltering.
• We have presented all DIC quantities in the coordinate systems defined by VIC-2D. Please note that "exy," "fxy", and "fyx" must be multiplied by -1 to match the EBSD coordinate system.
• After correlation, the fields were stitched together using cross-correlation (DOI: 10.1007/s11340-018-0419-y). The stitching code is available on Github at andrewchristison/dic-stitching.

dds.txt contains the raw force and displacement measured by the load cell (1 kN) and extensometer (LVDT), respectively, as a function of time. Please note that the LVDT extensometer is somewhat unreliable. For the published stress-strain curves, these displacement data were rescaled using the SEM-measured end-to-end displacement of the gauge section.

• Stress data were scaled using the upper plateau stress. In our material, this was 309 MPa (65 N).
• The maximum end-to-end strain measured in the gauge section was 8.54%.

EBSD

Austenite grain orientation maps were obtained prior to the tensile test via EBSD. Diffraction patterns were acquired at 30 kV with an EDAX camera at a 1-micron step size, then indexed with the Hough transform in OIM Analysis. The orientation map was aligned to the DIC data by manually selecting control points and warping with a third-order polynomial.

• oima_orientations.ang is the orientation data exported from OIM Analysis. Diffraction patterns were acquired at 30 kV with an EDAX camera at a 1-micron step size, then indexed with the Hough transform in OIM analysis.
• mtex_orientations.mat is the EBSD orientations after alignment with the DIC data. Note that these two datasets do not have the same resolution; instead, each EBSD data point contains a saved position (x, y), which corresponds to the pixel location in the DIC data. Data are saved in the format of the MTEX library.
• mtex_grains.mat contains the grains calculated in MTEX, with a misorientation threshold of 0.5 degrees.

MMICROSTRUCTURE

The MMICROSTRUCTURE framework allows users to simultaneously reconstruct martensitic transformation and plastic slip from SEM-DIC and EBSD data. mmicrostructure.mat is a Matlab struct containing:

• The activity fields stored as a 3D array (4944, 4873, 18). Here, we include the data for strain levels e1, e2, e3, and e8. Each page of the array contains either the volume fraction of a martensitic variant (1:12) or the activity of a slip system (13:18). This reconstruction was obtained with alpha = 0.05 and epsilon = 0.02, as described in the linked publication.
• The kinematics eta stored as a 3D array (3, 3, 18). Each page contains the full displacement gradient tensor of a martensitic variant (1:12) or slip system (13:18). Note that each tensor must be rotated into the coordinate system of the grain before reconstruction.