3D characterization of block copolymer films for lithography
The ever-shrinking length scales of semiconductor patterning have reached the sub-20nm level in the last few years, and multiple patterning steps are now needed to resolve such small dimensions. The semiconductor industry has therefore been pursuing alternative processes that enable even smaller features, as well as a high degree of pattering perfection, at lower complexity and cost. The use of block copolymers (BCPs) in the lithographic process is considered to be one of the most promising strategies for augmenting and enhancing the capabilities of current photolithography processes.1, 2 Indeed, patterns with sub-10nm features have been demonstrated through the use of directed self-assembly (DSA) of BCPs. Although much research in both academic institutions and industry has led to tremendous progress in the DSA field, some challenges still remain. For instance, defect density and precise control over the assembled structure need to be addressed so that DSA technology can be implemented in nanomanufacturing.
In the DSA process, guiding patterns are used to direct the assembly of the BCPs into highly aligned morphologies, with registration to the underlying layer.3, 4 In addition, self-assembly is a bottom-up, 3D process. This means that conventional characterization of patterns, with the use of 2D imaging, is not sufficient for fully probing the assembled structures. To advance the understanding of DSA processing and to meet the challenges in DSA manufacturing, 3D characterization of the assembled structures is therefore necessary.
In our work we use transmission electron microscope (TEM) tomography to explore the 3D morphology in BCP DSA. We can also examine the role of the guiding pattern and the processing conditions on the assembled structure. In a typical TEM tomography characterization, the sample is tilted in the microscope at angles between –70° and +70°, and a series of about 50 (2D) images is acquired. We then use the 2D image series to reconstruct the 3D structure of the specimen.
To image DSA images in a TEM and a scanning TEM (STEM), we have developed an artifact-free sample preparation method. Our method is based on wet back-etching of the silicon wafer and allows us to obtain a BCP film together with its guiding pattern on a silicon nitride window (see Figure 1).5 In addition, we use selective growth of aluminum oxide in the polar domains of the BCP as a new staining method. With this approach we enable high-contrast and stable imaging, which are both necessary for tomography.6, 7
We have used STEM tomography to probe the 3D structure of poly(styrene-block-methyl methacrylate)—PS-b-PMMA—which is one of the most common polymers in DSA patterning. Tomographic data for lamellar PS-b-PMMA DSA on a chemical pattern is shown in Figure 2. This pattern has a 3× density multiplication (i.e., the periodicity of the guiding pattern is three times the periodicity of the BCP). With our visualization of the reconstructed volume—Figure 2(a)—we demonstrate the rich 3D data that is obtained in our tomography process. This information includes the through-film structure of the BCP and the fluctuation (roughness) of the assembled lines. The average cross section of this volume (calculated by averaging all the x–z cross sections) shows how the morphology of the lamellae changes through the film thickness, as well as with their location relative to the guiding pattern. Although we see that all the lamellae are perpendicular at the top of the volume, we find that the lamellae adjacent to the guiding stripes in the guiding pattern are curved near to the interface. This curvature can affect the pattern transfer of the BCP into the underlying layer.8
We have also used STEM tomography to shed light on the 3D structure of defects and thus help understand their origin and behavior. In Figure 3 we show a 3D characterization of a dislocation defect. We compared experimental tomography data—Figure 3(a) and (c)—with the 3D structure obtained from molecular simulations. The simulations in Figure 3(b) illustrate the predicted defect's metastable state. We find an excellent agreement between the experimental results and the molecular simulations and have therefore validated the existence of the dislocation metastable state. This shows that kinetics—not just thermodynamics—play a key role in allowing a polymeric material to self-assemble into a perfect, defect-free, ordered state.9
In summary, we have demonstrated that TEM tomography is a powerful technique for investigating the 3D morphology of block copolymers in thin films. Our characterization approach is important because BCP DSA is a 3D process and the 3D structure of the BCPs must be accounted for when designing DSA for lithography. We have developed high-contrast staining and artifact-free sample preparation for our tomographic characterizations and are thus able to probe the through-film structure, roughness, and defects of our samples at high resolution, and subsequently gain a better understanding of BCP self-assembly processes. In our future work we will focus on the quantitative analysis of roughness in 3D data and on tomographic characterization of contact hole shrink DSA processes.
University of Chicago
Materials Science Division (MSD)
Argonne National Laboratory
Tamar Segal-Peretz obtained her PhD in nanoscience and nanotechnology from the Technion-Israel Institute of Technology, Israel. She is currently a director's postdoctoral fellow at the Argonne National Laboratory and at the IME. She will be joining the Faculty of Chemical Engineering at the Technion in 2017.
Paul Nealey received his PhD in 1994 from the Massachusetts Institute of Technology. He was a professor in the Department of Chemical and Biological Engineering at the University of Wisconsin before joining the IME in 2013. He is currently a Brady W. Dougan Professor of molecular engineering at the University of Chicago and a senior scientist in the MSD.
University of Chicago
Jiaxing Ren is a graduate student at the IME, where he studies the 3D assembly of block copolymers.