Stencil Mask Methodology for Parallelized Production of Microscale Mechanical Test Samples

This methodology parallelizes the production of micromechanical test samples from bulk materials.

June 1, 2013

Air Force Research Laboratory, Wright-Patterson Air Force Base, Ohio

Over the past decade, there has been considerable progress in the development of new mechanical testing methods to characterize the properties of materials at the micro and nano scales. One common application of these testing methodologies is the measurement of mechanical properties of structures that are physically small in scale, such as the strength of nanowhiskers and MEMS devices. Another common application is the use of small sample testing to gain insight into plastic deformation processes through systematic alteration of the sample dimensions in order to help isolate selected aspects of material behavior. Examples of these studies include the exploration of size-scale strengthening effects, the quantitative measurement and analysis of dislocation avalanches, and the measurement of local property variations in engineering alloys.

(A) Scanning electron microscope (SEM) image of the **Mask and Substrate** on a SEM pin mount, showing the positioning of the region of interest over the 5-mm thru hole (image taken at a stage tilt of 52 degrees). (B) Optical image of a 2 x 5-mm rectangular aperture positioned above the mask and substrate shown in (A).

For the latter application, focused ion beam (FIB) milling has been widely used to fabricate small samples from bulk crystals for micro-compression, micro-tension, and micro-bending studies. This process can be applied to a wide range of inorganic materials and is relatively low-damage. Importantly, ion beam milling methods enable the extraction of microsamples from modern engineering materials in bulk form that possess complex microstructures (e.g. polycrystalline alloys), whereas it is difficult or impossible to use microelectronic processing methods to directly synthesize microsamples with the same chemistry and distribution of microstructural features. The damage layer from these coarser-scale machining methods can usually be removed via subsequent FIB milling, resulting in a two-step sample preparation method that requires significantly less fabrication time compared to using only FIB milling.

The present work describes a new sample preparation methodology, using ion beam sputtering for material removal, which provides a parallelized production process to enable statistical measurements of microscale mechanical properties. The methodology is inspired by an ion sputtering process termed slope-cutting, which uses a physical mask with a well-defined edge, such as a knifeedge blade used for machine tooling, and a broad ion beam source to prepare a cross-section surface. The physical mask protects one portion of the sample while the edge defines the location of a cross-section surface that is created when the remainder of the sample is eroded by the ion beam. Slope-cutting maintains some of the advantages of FIB cross-section preparation, such as the ability to cleanly section multiphase materials while retaining features such as porosity, and the ability to apply the method to a wide variety of materials. Yet the dramatic increase in total ion current from 50 nA to hundreds of A or higher results in a much shorter time to sputter mesoscopic volumes of material. In addition, the equipment used to perform a slope-cutting experiment is simpler and usually much less expensive than a FIB microscope.

Instead of using a simple rectilinear mask as in slope cutting, the methodology uses microelectronics processing techniques to create physical masks with complex twodimensional shapes, i.e. stencil masks. The stencil mask methodology consists of three primary tasks. The first task is to create a freestanding stencil mask, which defines the pattern that is to be transferred to the substrate. The second step is to ensure that the substrate is in a suitable form for creating a threedimensional test sample via ion sputtering, where the region of interest is a thinned section with uniform thickness on the order of 20 to 100 μm. The final step is to use the stencil mask in conjunction with broad ion beam milling to transfer the pattern into the substrate.

Compared with conventional FIBbased fabrication, the methodology provides a fast and relatively low-cost processing route to manufacture an array of test structures with dimensions that range from 20 to 200 μm in scale. The methodology has been successfully demonstrated using stencil masks made from Si wafers, and pattern transfer to a Ni foil was demonstrated using a commercial broad ion beam milling system.

This work was done by Paul A. Shade and Michael D. Uchic of the Air Force Research Laboratory, and Sang-Lan Kim and Robert Wheeler of UES, Inc. AFRL-0221