Improved Method for Measuring and Assessing Reticle Pinhole Defects for the 100nm Lithography Node

 

Darren Taylor

Photronics, Allen TX 75013

 

Anthony Vacca, Larry Zurbrick

KLA-Tencor, 160 Rio Robles, San Jose, CA 95161

 

 

 

 

ABSTRACT

 

 

With the approach of the 100nm-lithography node, an accurate and reliable method of measuring reticle pinhole defects becomes necessary to assess the capabilities of high-end reticle inspection equipment. The current measurement method of programmed defect pinholes consists of using a SEM.  While this method is repeatable, it does not reliably represent the true nature of a pinhole.

 

Earlier studies have suggested that since the SEM images only a top down view of the pinhole, the measurement does not accurately account for edge wall angle and partial filling which both reduce pinhole transmission and subsequent printability.  Since wafer lithography and reticle inspection tools use transmitted illumination, pinhole detection performance based on SEM measurements is often erroneous. 

 

In this study, a pinhole test reticle was manufactured to further characterize the capabilities of a transmission method to measure pinholes.

 

Keywords: Inspection, Pinhole, Printability

 

 

 

1.      PROBLEM DEFINITION

 

 

Because of the inability to reliably manufacture and measure programmed pinhole defects on test masks, some reticle inspection tool vendors have stoppedshied away from setting a specification on these types of defects. This has left a hole in setting specifications from the customer of the mask houses.

 

From 1979 to the present, optical shearing microscopes were used to measure these defects. The resulting measurements were used in multiple printability studies in the industry. This method was adequate when the size of the defects were relatively large, however, issues began to arise when the when the defects fell below 500nm. The repeatability of the measurements became poor below 500nm and very poor below 200nm. In 2000 KLA-Tencor migrated to using a KLA-Tencor 8100XPR CD-SEM to measure the defects. This was done primarily for a high repeatability. The sizing method is the inscribed circle method. This involves placing a circle inside the defect and reporting the size of the defect. This methodology was chosen due to thebecause it has pretty good correlation back to the historical optical sizing1..

 

While SEM measurement are highly repeatable there are issues with this type of measurement. It cannot take into account poor sidewall or the presence of debris in the bottom of the pinhole. An example of this is shown in Iimage 1.

 

 

Image 1.

CD-SEM Circle Inscribed Methodology

 

The green circle is where the measurement was taken. As shown in the image there are very poor sidewalls on the defect. Image 2 shows an example of a "good" defect. This defect has a relatively high sidewall angle and is completely cleared.

 

 

 

Image 2.

Example of a "good" pinhole

 

 

 

As the test mask geometry and defect sizes shrink s, and therefore the defects present, have become increasingly aggressive it has become necessary to find a new measurement protocol to overcome the issues seen with the SEM measurements. Several methods were investigated such as probe type metrology, SEM improvements and transmission based metrology. There are issues associated with each of these. Probe type metrology is interesting but the size of the tip approaches or is larger than the defect of interest. SEM metrology could be improved to capture a better point in the sidewall, however, it does not address the issue of foreign material in the defect. Transmission based metrology appears to be the most comprehensive of the three and it moreis closely resembles r to the actual function of the mask in optical lithography.

 

Two transmission based measurement tools were investigated in this, an ongoingthis ongoing study of the issue, the KLA-Tencor Defect Energy Meter and the AVIÔ Flux-Area Measurement tool. The KLA-Tencor DEM is an engineering mode tool that uses images from an SLF27 inspection tool to measure the defects. A special noncommercial extraction program was written that enabled the The AVI tool to uses the same images. but obtains them from a results file. Both measurement techniques were shown to be repeatable within a single tool and from tool to tool in an earlier study on this problem2..

 

The decision was made to use the AVI tool primarily because it has the ability to create a reference within the defect image and the KLA-Tencor DEM tool needs a separate reference image. KLA-Tencor has the ability to capture images of the defects manually on the SLF27, however, only the defect image is captured, the reference image is not fully aligned which is necessary for the needed by the DEM to functionis not stored.

 

 

 

2.      EXPERIMENT

 

 

A programmed defect test mask based on the KLA-Tencor UTM was built using a 50KeV-lithography tool. The plate was designed for the 100nm node. The main feature size was 400nm and the defects were in increments of 20nm. The basis for the test plate was the KLA-Tencor UTM, this is the mask design that KLA-Tencor intends to use for setting the detection rate specifications on the 5XX generation tool.

 

This test mask was inspectedran multiple times on several KLA-Tencor SLF27 inspection tools and measurements made usingon the AVI tool on the individual runs.  Chart 1 shows the repeatability of the AVI pinhole measurements from system to system.  The X axis shows the average defect size measured with the AVI software while the X axis shows the variation from the average for each individual system.  The data show less than ±3nm variation all the way down to the 116nm sized pinhole.

 

Sensitivity measurements were taken from the inspection tools and the results are shown in image 5. The gray area in the sensitivity chart represents 100% capture rate. The individual boxes show the actual capture rate for that particular defect.

 

            Chart 1Image 3.

          AVI System to System Measurement Repeatability

 

 

 

 

Chart 2 shows the relationship between the AVI measured sizes and the CD-SEM sizes along with the delta between them. The delta has a range of 13nm. The yellow bar in the chart shows the 100% capture point for the SLF27 tool.

 

 

 

         Chart 2Image 4.

      SEM vs AVI measurements

 

 

Image 4 shows the relationship between the AVI measured sizes and the CD-SEM sizes along with the delta between them. The delta has a range of 13nm. The yellow bar in the chart shows the 100% capture point for the SLF27 tool.

 

Sensitivity measurements were taken from numerous inspection tools and the results from one system are shown in Chart 3. The gray area in the sensitivity chart represents 100% capture rate. The individual boxes show the actual capture rate for that particular defect.  The average defect sizes as measured by the AVI are shown along the right side vertical axis

 

 

 

 

Chart 3Image 5.

SLF Pinhole Sensitivity from Test Mask

 

 

 

3.      PRINTABILITY

 

In earlier phases of this study printability was performed to get a sense of how transmission measurements compared to printability. A test mask designed for the 130nm node (520nm lines) was printed with aggressive low k1 lithography (~ .42).

 

 

              

 

           Defect 12           Defect 12-wafer                    Defect 11              Defect 11-wafer

 

Images 36 through 69.

Printed Images for 130nm node

 

 

Images 36 through 69 show two consecutive defects on a programmed defect mask. There are two pair of images, one SEM image of the mask and the other is the SEM image of the printed wafer. The SEM circle inscribed method measurement on defect 12 was 215nm and the SEM measurement for defect 11 was 212nm. The AVI measurements for the same defects were 183nm and 114nm respectively. As shown in on the wafer images, defect 12 on the mask printed on the wafer causing the line to pinch. Defect 11 on the other hand did not print although the SEM measured only 3nm difference between them. The AVI measurement showed a difference of 69nm, which would be in better agreement with the printing results.

 

In the case of isolated pinholes the largest pinhole on the mask, which measured 279nm by SEM and 155nm by energy measurement, did not print.

 

The printability results suggest that the capture rate of the KLA-Tencor SLF27 tool is well below the printing threshold. The SLF captured the defects in the 116nm range 100% of the time. It also shows that the repeatability of the AVI tool is stable well below printing levels.

 

 

 

 

4.      SUMMARY

 

Large strides have been made in overcoming the obstaclesability in to repeatablebly measurement of small programmed pinhole reticle defects. The AVI transmission based measurement method, AVI, that was chosen in this study showed good repeatability on very small defects far below what actually prints. The KLA-Tencor SLF27 system also demonstrated the ability toshowed a capture rate on pinholes far below the printing threshold.

 

 

 

5.      REFERENCES

 

 

1    L. Zurbrick, et al, “Reticle defect size calibration using low-voltage SEM and pattern recognition techniques for sub-200 nm defects”, Proc. SPIE Vol. 3873, p.651-658, 19th Annual Symposium on Photomask Technology, F. E. Abboud; B. J. Grenon; Eds. (1999)

 

 

2   D. Taylor, et al, "Improved Method for Measuring and Assessing Reticle Pinhole Defects", Proc. SPIE Vol. 4562, p. 272-278, 21st Annual BACUS Symposium on Photomask Technology, G. T. Dao; B. J. Grenon, Eds. (2002)

 

L. Zurbrick, et al, “Reticle defect size calibration using low-voltage SEM and pattern recognition techniques for sub-200 nm defects”, Proc. SPIE Vol. 3873, p.651-658, 19th Annual Symposium on Photomask Technology, F. E. Abboud; B. J. Grenon; Eds. (1999)

 

6.      ACKNOWLEDGEMENTS

 

 

 

Matthew Lassiter, Ben Eynon - Photronics

 

Mohsen Ahmadian, Ed Longboy, Lantian Wang - KLA-Tencor

 

Peter Fiekowsky - AVI

 

IMEC for graciously printing wafers for this project