Systems for parsing material properties from within SHG signals
Semiconductor metrology systems based on directing radiation on a wafer, detecting second harmonic generated (SHG) radiation from the wafer and correlating the second harmonic generated (SHG) signal to one or more electrical properties of the wafer are disclosed. The disclosure also includes parsing the SHG signal to remove contribution to the SHG signal from one or more material properties of the sample such as thickness. Systems and methods described herein include machine learning methodologies to automatically classify obtained SHG signal
1. An automated method of characterizing electrical properties of a sample, the method comprising:
directing a beam of electro-magnetic radiation to a sample using an optical source thereby producing Second Harmonic Generation (SHG);
receiving a signal comprising an SHG signal; and
under the control of a hardware computing device:
processing the received signal to extract features from the SHG signal related to the electrical properties of the sample, wherein features from the SHG signal are extracted at least in part by applying a transform to the SHG signal; and
correlating the extracted features to one or more electrical properties of the sample.
2. The automated method of claim 1 , wherein the extracted features include spatio-temporal intensity of the SHG signal.
3. The automated method of claim 1 , wherein the transform comprises at least one of: a Fourier transform, a wavelet or a machine learning kernel.
4. The automated method of claim 1 , wherein correlating the extracted features comprises:
under the control of the hardware computing device: decoding the extracted features using a decoder;
mapping the decoded extracted features onto a decision; and classifying the SHG signal based on the decision.
5. The automated method of claim 1 , further comprising: under the control of the hardware computing device:
removing a portion of the SHG signal attributed to one or more material properties of the sample.
6. The automated method of claim 1 , wherein the transform comprises a Fourier transform.
7. The automated method of claim 1 , wherein the transform comprises a wavelet kernel.
8. The automated method of claim 1 , wherein the transform comprises a machine learning kernel.
9. The automated method of claim 4 , wherein the decision includes presence or absence of metal contaminant.
10. The automated method of claim 4 , wherein the decision includes type of contaminant.
11. The automated method of claim 4 , wherein the decoder is a linear decoder.
12. The automated method of claim 4 , wherein the decoder is a nonlinear decoder.
13. The automated method of claim 4 , wherein mapping the decoded extracted features onto a decision comprises projecting the extracted features onto a decision boundary.
14. The automated method of claim 4 , wherein the decision includes amount of metal contaminant.
15. The automated method of claim 5 , wherein the portion of the SHG signal attributed to one or more material properties of the sample is removed prior to extracting features from the SHG signal.
16. The automated method of claim 13 , wherein the decision boundary is obtained during a training phase of an automated system.
17. The automated method of claim 15 , wherein removing the portion of the SHG signal attributed to one or more material properties of the sample comprises:
receiving data associated with one or more material properties of the sample;
determining a quantitative relationship between the received data associated with one or more material properties of the sample and the received signal; and
normalizing the received signal to remove the portion of the SHG signal attributed to one or more material properties of the sample.
18. The automated method of claim 17 , wherein data associated with one or more material properties of the sample is received from using a secondary semiconductor analysis device.