Noise-reduction metrology models

• US 7,742,177 B2
• Filed: 01/22/2008
• Issued: 06/22/2010
• Est. Priority Date: 01/22/2008
• Status: Expired due to Fees

• Current Assignee: Tokyo Electron Limited
• Original Assignee: Tokyo Electron Limited
• Inventors: Liu, Yu, Li, Shifang
• Primary Examiner(s): PUNNOOSE, ROY M
• Application Number:
  US12/018,028
• Publication Number:
  US 20090187383A1
• Time in Patent Office: 882 Days
• Field of Search: 356/614, 356/625
• US Class Current: 356/625
• CPC Class Codes: G01B 11/06 for measuring thickness ; e.g. of sheet material thickness measurement by thermal means G01B21/085
• CPC Class Codes: G01B 11/24 for measuring contours or curvatures
• CPC Class Codes: G01N 21/4788 Diffraction for sizing particles G01N15/0205
• CPC Class Codes: G01N 21/9501 Semiconductor wafers manufacturing processes per se of semiconductor devices implementing a measuring step H01L22/10
• CPC Class Codes: G01N 21/956 Inspecting patterns on the surface of objects contactless testing of electronic circuits G01R31/308; testing currency G07D; manufacturing processes per se of semiconductor devices implementing a measuring step H01L22/10

Abstract

The invention can provide apparatus and methods for processing wafers using Noise-Reduction (N-R) metrology models that can be used in Double-Patterning (D-P) processing sequences, Double-Exposure (D-E) processing sequences, or other processing sequences.

Claims

1. A method of processing a wafer comprising:

aligning a reference wafer in a first metrology tool, the reference wafer having a first target structure at a first site;

obtaining a first set of measured diffraction signals, each measured diffraction signal being measured from the first site on the reference wafer using a first set of sensor pixels in a first sensor;

performing a first noise-reduction procedure to generate first noise-reduced data for the first sensor based on the first set of measured diffraction signals, the first noise-reduced data having first accuracy data associated therewith;

obtaining simulation data for the reference wafer, the simulation data having simulated accuracy data associated therewith;

identifying the first target structure using the first noise-reduced data when the first accuracy data is more accurate than the simulated accuracy data; and

identifying the first target structure using the simulation data when the first accuracy data is not more accurate than the simulated accuracy data.

2. The method of claim 1, wherein performing the first noise-reduction procedure comprises:

selecting a first number of sensor pixels;

defining first signals at the first number of sensor pixels using a basis function, wherein each first signal is defined as;

s(i)=a(i)·

b(i,p)+r(i)+n(i)

wherein a(i) and p are parameters to be determined, b(i, p) is a pre-selected basis function, r(i) is a residual signal from the basis function, and n(i) is a noise signal to be suppressed.

3. The method of claim 2, wherein the basis function is defined using a polynomial and the polynomial is:

b(i,p)=p(0)+p(1)·

i+p(2)·

i²+ . . . .

4. The method of claim 2, wherein the parameters a(i) and p are selected so that the basis function is substantially close to a measured diffraction signal, and r(i) is minimized.

5. The method of claim 4, wherein r(i) is minimized using a regression method wherein one or more parameters of the basis function are adjusted.

6. The method of claim 5, wherein the regression method is performed at a center pixel.

7. The method of claim 1, wherein the first noise-reduction procedure is performed in real time.

8. The method of claim 1, wherein the first noise-reduction procedure is performed using Optical Digital Profilometry (ODP) procedures, spectral resolution averaging procedures, Moving Box averaging procedures, or Frequency filtering procedures, or any combination thereof.

9. The method of claim 1, further comprising:

identifying the first noise-reduced data as low risk data and storing the first noise-reduced data, when the first accuracy data is more accurate than the simulated accuracy data; and

performing a first corrective action, when the first accuracy data is not more accurate than the simulated accuracy data.

10. The method of claim 9, wherein the first corrective action includes selecting a new site, re-aligning the reference wafer, selecting a new limit, selecting a new noise-reduction procedure, selecting a new wafer, selecting a new set of pixels, selecting a new sensor, or selecting a new metrology tool, or any combination thereof.

11. The method of claim 1, further comprising:

creating a noise-reduced (N-R) metrology model for the first target structure, the N-R metrology model comprising one or more enhanced profile parameters, one or more enhanced process parameters, and a noise-reduction procedure, wherein a noise component is minimized to improve accuracy for the enhanced profile parameters, or the enhanced process parameters, or a combination thereof; and

storing the N-R metrology model.

12. The method of claim 1, further comprising:

changing an angle of incidence, an intensity, a wavelength, a pixel pattern, or a number of pixels, or any combination thereof; and

obtaining a second set of measured diffraction signals.

13. The method of claim 1, further comprising:

identifying the reference wafer and wafers associated with the reference wafer as low risk wafers and storing the first noise-reduced data, when the first accuracy data is more accurate than the simulated accuracy data; and

performing a first corrective action, when the first accuracy data is not more accurate than the simulated accuracy data.

14. The method of claim 1, further comprising:

identifying the reference wafer and wafers associated with the reference wafer as low risk wafers and storing the first noise-reduced data, when the first noise-reduced data is more accurate than the simulation data in a first set of wavelengths; and

performing a first corrective action, when the first noise-reduced data is not more accurate than the simulation data in the first set of wavelengths.

15. The method as claimed in claim 1, wherein the first target structure is associated with a gate structure, a drain structure, a source structure, a two-dimensional transistor structure, a three-dimensional transistor structure a capacitor structure, a via structure, a trench structure, a two-dimensional memory structure, a three-dimensional memory structure, a sidewall angle, a critical dimension (CD), an array, a periodic structure, an alignment feature, a doping feature, a strain feature, a damaged-structure, or a sensor-related reference structure, or any combination thereof.

16. A system for using Noise-Reduced (N-R) data to examine a semiconductor wafer, the system comprising:

a wafer alignment means configured for aligning a reference wafer in a first metrology tool, the reference wafer having a first target structure at a first site, the wafer alignment means comprising a platform subsystem, an alignment subsystem coupled to the platform subsystem, and an alignment sensor coupled to the alignment subsystem;

one or more collection subsystems configured for obtaining a first set of measured diffraction signals, each measured diffraction signal being measured from the first site on the reference wafer using a first set of sensor pixels in a first sensor; and

a controller configured for performing a first noise-reduction procedure to generate first noise-reduced data for the first sensor based on the first set of measured diffraction signals, the first noise-reduced data having first accuracy data associated therewith, for obtaining simulation data for the reference wafer, the simulation data having simulated accuracy data associated therewith, for identifying the first target structure using the first noise-reduced data when the first accuracy data is more accurate than the simulated accuracy data, and for identifying the first target structure using the simulation data when the first accuracy data is not more accurate than the simulated accuracy data.

17. The system of claim 16, wherein the controller is further configured for selecting a first number of sensor pixels, for defining first signals at the first number of sensor pixels using a basis function, wherein each first signal is defined as:

s(i)=a(i)·

b(i,p)+r(i)+n(i)

wherein a(i) and p are parameters to be determined, b(i, p) is a pre-selected basis function, r(i) is a residual signal from the basis function, and n(i) is a noise signal to be suppressed.

18. The system of claim 17, wherein the controller is further configured for defining the basis function using a polynomial, wherein the polynomial is:

b(i,p)=p(0)+p(1)·

i+p(2)·

i²+. . . .

19. A non-transitory computer-readable medium containing computer-executable instructions for performing a Noise-Reduction (N-R) procedure, the method comprising instructions for:

aligning a reference wafer in a first metrology tool, the reference wafer comprising a first target structure at a first site;

obtaining a first set of measured diffraction signals, each measured diffraction signal being measured from the first site on the reference wafer using a first set of sensor pixels in a first sensor;

performing a first noise-reduction procedure to generate first noise-reduced data for the first sensor based on the first set of measured diffraction signals, the first noise-reduced data having first accuracy data associated therewith;

obtaining simulation data for the reference wafer, the simulation data having simulated accuracy data associated therewith;

identifying the first target structure using the first noise-reduced data when the first accuracy data is more accurate than the simulated accuracy data; and

identifying the first target structure using the simulation data when the first accuracy data is not more accurate than the simulated accuracy data.