Non-destructive acoustic metrology for void detection

• US 9,991,176 B2
• Filed: 09/29/2015
• Issued: 06/05/2018
• Est. Priority Date: 09/29/2014
• Status: Active Grant

• Current Assignee: Onto Innovation Inc., The Board of Regents of the University of Colorado (University of Colorado System)
• Original Assignee: Rudolph Technologies Incorporated (Onto Innovation Inc.), The Board of Regents of the University of Colorado (University of Colorado System)
• Inventors: Mehendale, Manjusha, Kotelyanskii, Michael, Murray, Todd W., Mair, Robin, Mukundhan, Priya, Dove, Jacob D., Ru, Xueping, Cohen, Jonathan, Kryman, Timothy
• Primary Examiner(s): ZARNEKE, DAVID A
• Application Number:
  US15/515,126
• Publication Number:
  US 20170221778A1
• Time in Patent Office: 980 Days
• CPC Class Codes: G01N 2291/0289 Internal structure, e.g. defects, grain size, texture
• CPC Class Codes: G01N 29/04 Analysing solids using acoustic emission techniques G01N29/14
• CPC Class Codes: G01N 29/06 Visualisation of the interior, e.g. acoustic microscopy medical or veterinary diagnosis using sonic waves A61B8/00; representation of acoustic wave distribution G01H3/125, G01H9/002; short-range imaging systems using reflection of acoustic waves G01S15/8906
• CPC Class Codes: G01N 29/0681 by acoustic microscopy, e.g. scanning acoustic microscopy
• CPC Class Codes: G01N 29/069 Defect imaging, localisation and sizing using, e.g. time of flight diffraction [TOFD], synthetic aperture focusing technique [SAFT], Amplituden-Laufzeit-Ortskurven [ALOK] technique
• CPC Class Codes: G01N 29/07 by measuring propagation velocity or propagation time of acoustic waves
• CPC Class Codes: G01N 29/2418 using optoacoustic interaction with the material, e.g. laser radiation, photoacoustics photoacoustic cells G01N21/1702; measuring characteristics of vibrations by using radiation-sensitive means G01H9/00; acousto-optical conversion techniques for short-range imaging G01S15/8965; sound-producing devices using laser bundle G10K15/046
• CPC Class Codes: G01N 29/46 by spectral analysis, e.g. Fourier analysis or wavelet analysis spectral signal processing per se G06F17/14
• CPC Class Codes: H01L 21/76898 formed through a semiconductor substrate
• CPC Class Codes: H01L 22/12 for structural parameters, e.g. thickness, line width, refractive index, temperature, warp, bond strength, defects, optical inspection, electrical measurement of structural dimensions, metallurgic measurement of diffusions electrical measurement of diffusions H01L22/14
• CPC Class Codes: H01L 22/20 Sequence of activities consisting of a plurality of measurements, corrections, marking or sorting steps

Abstract

Advanced interconnect technologies such as Through Silicon Vias (TSVs) have become an integral part of 3-D integration. Methods and systems and provided for laser-based acoustic techniques in which a short laser pulse generates broadband acoustic waves that propagate in the TSV structure. An optical interferometer detects the surface displacement caused by the acoustic waves reflecting within the structure as well as other acoustic waves traveling near the surface that has information about the structure dimensions and irregularities, such as voids. Features of voids, such as their location, are also identified based on the characteristics of the acoustic wave as it propagates through the via. Measurements typically take few seconds per site and can be easily adopted for in-line process monitoring.

Claims

1. A method of non-destructive acoustic metrology for void detection in a via in a sample, the method comprising:

generating a model for determining the existence of a void within the via, comprising;

generating a model signal representative of a signal that would be received from performing acoustic metrology on the via;

removing a signal component due to temperature rise and subsequent heat dissipation from the model signal;

generating a model periodogram by performing a Fourier transform over a sliding window;

applying a modified image processing segmentation algorithm to locate centroid locations of echo patterns in the model periodogram; and

generating a correlation of the centroid locations to a depth of where a model acoustic pulse was reflected within the via; and

directing an excitation laser beam at the via, such that an acoustic wave propagates through the via;

directing a probe laser beam at the via, such that the probe laser beam reflects off the sample, creating a reflected beam encoding data about propagation of the acoustic wave in the via;

analyzing the reflected beam, using a detection periodogram, to determine one or more detection centroids;

applying the generated correlation to determine the depth of an acoustic wave reflection associated with the one or more detection centroids;

comparing the determined depth with an expected depth for the via; and

based on the comparison of the determined depth with the expected depth, generating an indication that there is the void in the via.

2. The method of claim 1, further comprising, based on the indication that there is a void in the via, diverting the sample from production for additional inspection.

3. The method of claim 1, wherein the model periodogram is represented as P(μ

f)=|FT[u(t),μ

)]| and is defined as an amplitude of a Fourier transform of a top displacement u(t) signal calculated with a window centered at delay time τ

.

4. The method of claim 1, further comprising displaying the depth of the void in the via associated with a first detection centroid and a depth of the via.

5. The method of claim 1, further comprising determining whether the void is in a body of the via or at a bottom of the via.

6. The method of claim 5, wherein determining whether the void is in the body of the via or at a bottom of the via comprises identifying a first detection centroid having a first associated time shorter than an expected time and a second detection centroid having a second associated time of approximately double the first associated time.

7. The method of claim 6, wherein the first detection centroid has a first associated frequency and the second detection centroid has a second associated frequency, the first and second frequencies being approximately equal.

8. The method of claim 7, further comprising, based on the first and second associated times, determining the void is at the bottom of the via.

9. The method of claim 5, wherein determining whether the void is in the body of the via or at a bottom of the via comprises identifying a first detection centroid having a first associated time shorter than an expected time and a second detection centroid having a second associated time approximately equal to the expected time.

10. The method of claim 9, further comprising, based on the first and second associated times, determining that the void is in the body of the via.

11. The method of claim 5, wherein determining whether the void is in the body of the via or at a bottom of the via comprises:

identifying a first detection centroid having a first associated time shorter than an expected time and a first associated frequency approximately equal to an expected frequency; and

identifying a second detection centroid having a second associated frequency higher than the first associated frequency.

12. The method of claim 11, further comprising, based on the first and second associated frequencies, determining that the void is at the bottom of the via.

13. A method of non-destructive acoustic metrology for void detection in a via in a sample, the method comprising:

directing an excitation laser beam at the via, such that an acoustic wave propagates through the via;

directing a probe laser beam at the via, such that the probe laser beam reflects off the sample, creating a reflected beam encoding data about propagation of the acoustic wave in the via;

analyzing the reflected beam, using a periodogram, to identify a first centroid and a second centroid, wherein the first centroid has a first associated delay time and a first associated frequency and the second centroid has a second associated delay time and a second associated frequency;

comparing the first and second associated delay times to an expected delay time for an ideally filled via;

comparing the first and second associated frequencies to an expected frequency range;

based on the comparison of the first and second delay times and the comparison of the first and second associated frequencies, generating an indication that the via contains one or more of the group selected from;

a void in a body of the via, a void at the bottom of the via, and no voids.

14. The method of claim 13, wherein an indication that the via contains no voids is generated if the first associated delay time is approximately equal to the expected delay time and the second associated delay time is approximately double the expected delay time.

15. The method of claim 13, wherein an indication that the via contains the void in the body of the via is generated if the first associated delay time is shorter than the expected delay time and the second associated delay time is approximately equal to the expected delay time.

16. The method of claim 13, wherein an indication that the via contains the void at the bottom of the via is generated if the first associated delay time is shorter than the expected delay time and the second associated delay time is approximately equal to double the first associated delay time.

17. The method of claim 13, further comprising, based on an indication that the via contains a void in a body of the via or a void at the bottom of the via, modifying the operation of one or more process tools to produce subsequent integrated circuit devices that have a lower likelihood of having voids.

18. A system for non-destructive acoustic metrology for void detection in a via in a sample, the system comprising:

a first source for generating a excitation beam directed towards the via in the sample;

a second source for generating a probe beam directed towards the via in the sample;

an optical sensor for detecting a reflected probe beam reflected from the sample;

at least one processor; and

a memory storing instructions, that when executed by the at least one processor, perform actions comprising;

generate a model for determining the existence of a void within the via, comprising;

generate a model signal representative of a signal that would be received from performing acoustic metrology on the via;

remove a signal component due to temperature rise and subsequent heat dissipation from the model signal;

generate a model periodogram by performing a Fourier transform over a sliding window;

apply a modified image processing segmentation algorithm to locate centroid locations of echo patterns in the model periodogram; and

generate a correlation of the centroid locations to a depth of where a model acoustic pulse was reflected within the via; and

analyze data associated with the reflected beam, using a detection periodogram, to determine one or more detection centroids;

compare features of the one or more detection centroids to expected features of the via to determine if there is a void in the via; and

if there is the void in the via, generating an indication that there is the void in the via.

19. The system of claim 18, wherein the via is a through silicon via (TSV) having a height to width ratio of approximately 10:

1 and a depth of approximately 10-20 μ

m.

20. The system of claim 18, wherein the first source is a short pulse laser configured to generate a 532 nm wavelength laser beam, and the second source is a continuous wave laser configured to generate a 660 nm wavelength laser.

21. The system of claim 18, wherein the instructions, that when executed by the at least one processor, perform actions further comprising:

applying the generated correlation to determine the depth of an acoustic wave reflection associated with the one or more detection centroids; and

displaying the depth associated with the one or more detection centroids.

22. The system of claim 18, wherein the instructions, that when executed by the at least one processor, perform actions further comprising:

if there is the void in the via, diverting the sample from production for additional inspection; and

based on the additional inspection, modifying the operation of one or more process tools to produce subsequent integrated circuit devices that have a lower likelihood of having voids.

23. A substantially void-free integrated circuit device prepared by a process comprising the steps of:

manufacturing at least a portion of a first wafer using one or more process tools;

receiving the first wafer for inspection, wherein the first wafer includes multiple vias;

directing an excitation laser beam at a via of the multiple vias, such that an acoustic wave propagates through the via;

directing a probe laser beam at the via, such that the probe laser beam reflects off the sample, creating a reflected beam encoding data about propagation of the acoustic wave in the via;

analyzing, using a periodogram, the reflected beam to identify a first centroid and a second centroid, wherein the first centroid has a first associated delay time and a first associated frequency and the second centroid has a second associated delay time and a second associated frequency;

comparing the first and second associated delay times to an expected delay time for an ideally filled via;

comparing the first and second associated frequencies to an expected frequency range;

based on the comparison of the first and second delay times and the comparison of the first and second associated frequencies, determining that the wafer contains at least one void in one or more of the multiple vias;

diverting the first wafer from production for additional inspection;

modifying the operation of the one or more process tools to produce subsequent wafers that have a lower likelihood of having voids; and

manufacturing a second wafer using the modified operation of the one or more process tools to produce the substantially void-free integrated circuit device.