Skip to main content
U.S. flag

An official website of the United States government

Official websites use .gov
A .gov website belongs to an official government organization in the United States.

Secure .gov websites use HTTPS
A lock ( ) or https:// means you’ve safely connected to the .gov website. Share sensitive information only on official, secure websites.


Patent Number: 11,181,489


Provided herein are methods and apparatus for characterizing high aspect ratio (HAR) structures of fabricated or partially fabricated semiconductor devices. The methods involve using small angle X-ray scattering (SAXS) to determine average parameters of an array of HAR structures. In some implementations, SAXS is used to analyze symmetry of HAR structures in a sample and may be referred to as tilted structural symmetry analysis-SAXS (TSSA-SAXS) or TSSA. Analysis of parameters such as tilt, sidewall angle, bowing, and the presence of multiple tilts in HAR structures may be performed.

patent description

The proposed invention utilizes an X-ray source, X-ray detector, goniometer, and an X­ ray scattering pattern to determine the average tilt (direction and magnitude) in patterned and etched arrays of high aspect ratio (HAR) structures (>20:1) being developed for next generation flash memory (3D-NAND). The technique is a small angle X-ray scattering technique (SAXS) and is called tilted structure symmetry analysis - SAXS (TSSA­ SAXS) or TSSA for short.

The necessary components for the measurement are an X-ray source, an X-ray detector, a goniometer with at least 1 rotation axis and x-y-z translation axes. The sample target for measurement should be a periodic patterned array of HAR structures. A simple representation of the measurement configuration is shown in Figure I. The technique is a transmission mode scattering measurement where the X-ray beam passes through the sample substrate and target pattern. The X-ray scattering pattern is recorded on the detector which is downstream from the sample and at a distance that provides sufficient resolution of the scattering peaks which are divergent from the main X-ray beam.

As the sample is rotated about the measurement rotation axis (#1 in Figure l, which is the y-axis), the scattering pattern changes and provides information regarding the tilt of the structures in the patterned array. With an understanding of the scattering patterns that result from tilted HAR structures, the average tilt of the structures can be determined. When the sample target is at normal incidence to the X-ray beam, the resulting scattering pattern can immediately indicate if the structures are tilted and the direction of the tilt but not its magnitude. When the sample is rotated about #1, the scattering pattern will change to reflect an increase or decrease of the component of tilt in the structures that is in the x­ z plane. The sample is rotated about #1 until the scattering pattern is symmetric about the measurement rotation axis, signifying that the X-ray beam is aligned with the component of tilt in the structures that is in the x-z plane and the only potential component of tilt remaining must be in the y-z plane. The amount by which the sample was rotated is equal and opposite to the component of the HAR structure tilt in the x-z plane. Simulated examples of scattering patterns generated by the technique are given in Figure 2 where:

2.a) shows a normal incidence scattering pattern for a sample having tilted HAR structures and 2.b) shows that the scattering pattern has become symmetric about the vertical axis when the sample is rotated using axis #1 to -1.2°, signifying that the component of tilt in the structures in the x-z plane is 1.2°.The component of tilt in the y-z plane can be determined either by: i) rotating the sample 90 degrees about the sample surface normal (the z-axis in Figure 1) and repeating the measurement with rotation axis

#1 ii) adding an additional rotation axis to the goniometer (#2 in Figure 1) that is orthogonal to the first measurement rotation axis and repeating the measurement with this additional axis.

With additional considerations, it is also possible to significantly reduce the time needed for X-ray scattering pattern collection at each measurement angle. If there is only 1 component of tilt in the structure and it is in the x-z plane, or if the component of tilt in the in the y-z plane is sufficiently small, it can be more difficult to determine the point at which the scattering pattern is symmetric about the measurement rotation axis, #1. The evolution of the scattering pattern with sample rotation angle is such that it may be easier to identify the symmetry in the scattering pattern and thus the component of tilt in the x-z plane, if there is a significant (-1 °) component of tilt in the structures in the y-z plane.

The component of tilt in the y-z plane can be present in the structure itself from the fabrication process and/or it can be induced using an additional rotation axis of the goniometer that is orthogonal to the measurement rotation axis. In Figure 1, where #1 is the measurement rotation axis, this offset would be applied to #2. The total tilt of the structures in the y-z plane is then a combination of the intrinsic tilt in the structure from the fabrication process and the rotation offset, #2. The total value from these two contributions should be -1 °.

As described, the proposed invention is able to determine tilt angle in HAR structures. This parameter is crucial to a number of applications which involve etching of HAR structures. Measuring tilt in HAR structures would be crucial in product development and during process/quality control. An industry relevant 0.1° sensitivity to tilt in HAR holes has been demonstrated using TSSA-SAXS on a non-optimized SAXS instrument.

The proposed invention can use currently available compact X-ray sources and X-ray optics along with an understanding of the scattering pattern that results from tilted HAR structures in a patterned array to determine the average tilt in less than a minute. The advantages over cross sectioning techniques are that it is fast, is non-destructive, does not require sample preparation, samples a statistically significant area (that of the X-ray beam spot size on the sample target), and would therefore be preferential for both process/product development as well as process/quality control.

Created September 12, 2022, Updated November 22, 2022