A.C. Madison, K.-T. Liao, J. Schumacher, J.S. Villarrubia, K. Siebein, B.R. Ilic J.A. Liddle, S.M. Stavis
National Institute of Standards and Technology,
Keywords: focused ion beam, nanofabrication, super-resolution
Summary:Focused-ion-beam milling is one of the most powerful and practical methods for prototyping and manufacturing complex devices at the nanometer scale, with applications as diverse as electronics, optics, and fluidics. In most focused-ion-beam systems, the ion-beam profile is approximately Gaussian, and its central peak delivers most of the ion dose to the substrate. However, non-uniform exposure of the substrate by the diffuse periphery, or tails, of the beam degrades the lateral fidelity of pattern transfer through a variety of processes, such as swelling and sputtering, at the edges of features. Reducing the ion-beam current generally improves lateral resolution, but increases the time to mill through the same vertical range, which can inhibit rapid prototyping and prolong drift of the fabrication system to the detriment of lateral resolution. This issue is particularly problematic for dielectric materials, such as silica, which are more difficult to mill than metals or semiconductors due to substrate charging, which affects both vertical and lateral resolution. To solve this problem, we introduce a sacrificial chromia film for electron imaging and ion-beam-tail masking of silica, enabling nanometer vertical resolution and lateral super-resolution of complex nanostructures. To begin, we focus a gallium-ion beam onto the chromia film and mill test pits into the underlying silica by dwelling the beam in discrete positions. We measure the resulting nanostructures by scanning electron microscopy and atomic force microscopy. We empirically model the image data, propagating relevant uncertainties of dimensions by Monte-Carlo methods, to correlate the two microscopy methods. In this way, we achieve an accurate and efficient method of in-line metrology of ion-beam focus, which we demonstrate over a decade of ion-beam currents (Figure 1). This result facilitates the use of scanning electron microscopy in an electron–ion beam system to quantitatively optimize focus by measuring and minimizing radii of test pits with angular sensitivity in near real time. With the capability of reproducibly focusing the ion beam, we mill checkerboard patterns through the chromia and into the underlying silica with ion doses up to 1,020 pC∙µm-2, forming complex nanostructures as deep as 200 nm before removal of the sacrificial chromia film and 130 nm after removal (Figure 2). These structures enable the first systematic study of the focused-ion-beam milling response of chromia on silica, which manifests an initial milling rate as low as 0.04 µm3∙nC-1, while gallium bombardment and chemical etching near the chromia–silica interface yields nanostructures that rise above the silica surface. The sacrificial chromia mask allows vertical resolution of 1 nm to 2 nm down to depths of 130 nm. By inputting milling rates into a geometric model of lateral super-resolution, we simulate values of lateral super-resolution of up to a factor of four, which we achieve experimentally (Figure 3). In conclusion, our study facilitates efficient and accurate in-line metrology during use of electron–ion beam systems and advances the super-resolution milling of complex nanostructures in dielectric materials. We expect that both aspects of our work will broadly impact the field of focused-ion-beam nanofabrication.