Z-profiling of wafers based on X-ray measurements

Patent metadata

Field	Value
Publication number	US-12480898-B2
Application number	US-202217901705-A
Country	US
Kind code	B2
Filing date	Sep 1, 2022
Priority date	Sep 1, 2022
Publication date	Nov 25, 2025
Grant date	Nov 25, 2025

Abstract

A computer-based method for non-destructive z-profiling of samples. The method includes: a measurement operation and a data analysis operation. The measurement operation includes, for each of a plurality of landing energies: (i) projecting an electron beam on a sample at a respective landing energy, such that light-emitting interactions between electrons from the electron beam and the sample occur within a respective probed region of the sample, which is centered about a respective depth; and (ii) measuring the emitted light to obtain an optical emission data set of the sample. The data analysis operation includes obtaining from the measured optical emission data sets a concentration map quantifying a dependence of a concentration of a material, which the sample comprises, on at least the depth.

First claim

What is claimed is:

1 . A system for non-destructive depth-resolved profiling of patterned wafers, the system comprising:
an electron beam (e-beam) source configured to project e-beams on a patterned wafer at a plurality of landing energies and at a plurality of lateral locations within a profiled region of the patterned wafer that includes lateral non-uniformity, the e-beams inducing X-ray-emitting interactions within respective probed volumes, whose depth varies according to the landing energy; an X-ray sensing module configured to detect X-rays emitted from the patterned wafer and to generate X-ray emission data sets for the probed volumes, each associated with the landing energy and the lateral location of the inducing e-beam; and a computational module configured to generate, based on the X-ray emission data sets, a depth-resolved concentration map that quantifies a dependence of a concentration of a profiled material within the profiled region on depth and one or more lateral coordinates, wherein the computational module is configured to execute an algorithm that employs or is at least partially derived from simulated X-ray emission data generated by computer simulation of X-ray emissions predicted for e-beams of varying landing energies incident at a plurality of locations according to intended-design data for the profiled region.

2 . The system of claim 1 , wherein the X-ray sensing module is configured to measure an intensity of at least a portion of the respectively emitted X-rays, which have a frequency equal to, or within a frequency range about, a peak characteristic X-ray emission frequency of the profiled material.

3 . The system of claim 2 , wherein the X-ray sensing module comprises an energy-dispersive X-ray spectrometer or a wavelength-dispersive X-ray spectrometer.

4 . The system of claim 1 , wherein the computational module is configured to execute a machine-learning derived algorithm trained at least in part using the simulated X-ray emission data, whose output is the concentration map and whose inputs comprise the X-ray emission data sets, each labeled by the respective landing energy.

5 . The system of claim 4 , wherein the machine-learning derived algorithm is a classification neural network and the concentration map specifies the density of the profiled material to each of a plurality of density ranges.

6 . The system of claim 1 , wherein the concentration map is three-dimensional and quantifies the dependence of the concentration of the profiled material on depth and two lateral coordinates.

7 . The system of claim 1 , wherein a lateral resolution of the concentration map corresponds to a spacing between adjacent lateral locations.

8 . The system of claim 1 , wherein the spacing between adjacent lateral locations is smaller than a lateral width of segments of differing composition in the profiled region of the patterned wafer.

9 . The system of claim 1 , wherein values of the concentration map quantify a concentration of the profiled material within a corresponding voxel localized at the corresponding depth and lateral coordinates in the profiled region, and wherein the X-ray emission data sets for some or all of the landing energies correspond to X-ray-emitting interactions occurring within a probed volume of lateral dimensions that overlap a plurality of the voxels.

10 . The system of claim 1 , wherein the profiled region exhibits non-flat topology comprising areas at different elevations and/or includes at least one cavity or hole.

11 . The system of claim 1 , wherein sets of landing energies employed at different lateral locations are selected to differ from one another to achieve comparable depth coverage across materials of differing elevation and/or composition.

12 . The system of claim 1 , wherein the profiled material is an element combined with a semiconductor whose concentration modifies an electrical property of the combination.

13 . The system of claim 1 , wherein the concentration map is generated for two or more profiled materials.

14 . The system of claim 1 , wherein the simulated X-ray emission data is generated by a computer simulation calibrated using measured X-ray emission data sets and corresponding measured concentration maps for one or more reference samples.

15 . The system of claim 1 , wherein the computational module is further configured to use intended-design data of the profiled region as an input when generating the concentration map.

16 . The system of claim 1 , wherein the system is configured to resolve variations in concentration of the profiled material to within 1-5% of a nominal concentration of the profiled material.

17 . A computer-implemented method for non-destructive depth-resolved profiling of patterned wafers, the method comprising:
for each of a plurality of landing energies and for each of a plurality of lateral locations within a profiled region of the patterned wafer that includes lateral non-uniformity:
projecting an electron beam (e-beam) to induce X-ray-emitting interactions within a respective probed volume whose depth varies according to the landing energy; and
detecting emitted X-rays to obtain an X-ray emission data set labeled by the landing energy and the lateral location; and
generating, based at least on the X-ray emission data sets, a depth-resolved concentration map that quantifies a dependence of a concentration of a profiled material within the profiled region on depth and one or more lateral coordinates, wherein generating the concentration map comprises executing an algorithm that employs or is at least partially derived from simulated X-ray emission data generated by computer simulation of X-ray emissions predicted for e-beams of varying landing energies incident at a plurality of locations according to intended-design data for the profiled region.

18 . The method of claim 17 , wherein intended-design data of the profiled region is used as an input in generating the depth-resolved concentration map.

19 . The method of claim 17 , wherein detecting X-rays emitted from the profiled region comprises measuring an intensity of a portion of the respectively emitted X-rays which have a frequency equal to, or within a frequency range about, a peak characteristic X-ray emission frequency of the profiled material.

20 . The method of claim 17 , wherein the concentration map is an output of a machine learning derived algorithm trained at least in part using the simulated X-ray emission data, whose inputs comprise the X-ray emission data sets, each labeled by the landing energy of the respective inducing e-beam.

Assignees

• Applied Materials Israel Ltd

Inventors

• Girmonsky Doron
• Eilon Michal
• Shemesh Dror
• Hadar Uri

Classifications

• G06N3/08

  Learning methods · CPC title

• G01N23/2252 Primary

  Measuring emitted X-rays, e.g. electron probe microanalysis [EPMA] · CPC title