“Squirm Mark Array” Alignment: Preliminary Test Process

Roger Robbins                                                                                                                                                                      5/28/2015
Alexandra Joshi-Imre

Purpose

This is a description of a preliminary test to establish a process to measure the suspected random morphological movement of an insulating polymer film due to thermal cycling.  The procedure will be conducted on the Raith 150 TWO e-beam lithography system at UTD.

Introduction

The polymer film is an insulator and will be deposited on a glass slide for study.  The measurement concept is to pattern metal alignment markers on top of the film and then measure the location with the Raith metrology system.  Then the film and substrate is to be sent through a thermal cycle which is anticipated to cause the film to squirm via chemical compression and stretching so that the original rectangular array of markers will be moved into new locations following the movement of the film.  A second measurement of the marker locations should show a map of the film movement.

One problem with this concept is that the film and substrate are insulators, and the Raith e-beam deposits electrons into the insulator which causes the polymer to charge and deflect the electron beam so that imaging or position measurement cannot be accomplished.

Approach

We have already found that a thin coat of Copper over the insulative film will dissipate the electric charge deposited by the beam and allow proper imaging and measurement of marker locations.  Therefore if we pattern metal markers on top of the polymer film, we can measure their initial locations and then re-measure the locations after the thermal cycle and determine the “random” squirm of the film.  This will overcome the native charging effects and allow patterning and marker location measurement.

However, there is a kink in the measurement mathematics because after the thermal cycle, there will be no reference coordinate related to the coordinate system of the initial uniform array of markers for performing the second measurement.  We do have the xy coordinate system of the stage which was related to the initial sample coordinate system, but the sample must be removed from the Raith and patterns developed, thermal cycle processing, etc.  This means that the mechanical position of the sample for the after-thermal-cycle measurement is different from the initial placement so that the stage coordinates now have no relation to the sample coordinates of the initial marker patterning step.

Solution

While one cannot change the rules of geometry, a clever experimentalist can change the boundary conditions of the physical system, which will establish a stable link between the initial and after-annealing coordinate systems, thus leading to a simple mathematical comparison of marker positions which follows all the proper mathematical rules.

The trick is to isolate three of the original markers on top of the film from the body of the film which will squirm.  This can be done by physical removal of a ring of film around each of the three corner markers from the original marker matrix – (i.e. laser cutting a ring around the key markers, for example).  This will enable the reference markers from the original marker array to remain in position during the thermal cycles and serve as a reference framework by which all the “squirmed” markers on the body of the film can be re-measured in their new locations.  Then a simple mathematical comparison of ‘before” and ‘after” marker locations will show the “random” movement of the polymer as a result of the thermal cycling.

Preliminary Experiment

Before embarking on the actual measurement, we propose a test run with a ‘dummy’ sample to evaluate the marker location measurement on an insulator.  The “dummy” sample will be a 2×3 inch rectangular microscope glass slide with PMMA e-beam resist coated on it.  The film thickness will be about 2000 A produced by a spin coater using an acceleration of 3000 rpm/s and a 60 sec spin at a constant speed of 2000 rpm.  The soft bake will be 2 minutes at 180 C.

To drain the e-beam charge, the resist will be coated with a 7 nm film of Copper.  This metal topping will allow fine line exposure of the PMMA resist with no placement distortion.  After exposure, the Copper will be etched away in “Copper Etch” and the PMMA resist will be developed in the standard “Cold” PMMA developer (3:1 ration of IPA and MIBK with and added 1.5 % MEK and at 5 deg C) for 30 seconds and rinsed in IPA for 15 seconds.  After inspection, we will deposit 7 nm of Copper again over the PMMA.  This will allow establishment and testing of the Raith alignment procedure.

Conceptual Overview of Procedure

The concept of this procedure is to work out the Raith procedure for capturing equivalent data whereby two data sets will be obtained by capturing the positions of an array of markers on a glass slide and then recapturing the marker locations with the slide positioned at 180 degrees to the first position.  This gives us two sets of data which may be different so that we can calculate and present the position error vectors, which would simulate calculation of the relative movement of a squirmy film after annealing.

Test Measurement Procedure

  1. Clean and coat a 2×3” glass microscope slide with PMMA (2k rpm). This simulates an insulative substrate and allows room for a convenient set of alignment markers for the test.
  2. Coat the PMMA with 7 nm of Cu to dissipate the electron charge from the e-beam. This produces a very thin film of conducting Cu that will enable stable imaging and printing of e-beam patterns.  It also allows us to conveniently remove all of the Cu in order to develop the exposed patterns, using the standard Cu etch solution from the Cleanroom acid cabinet.
  3. Expose a 7×5 matrix of e-beam markers aligned to the LL corner and lower edge of the glass slide.
  4. After exposure, remove the Cu layer with the Cu etch solution and develop the exposed pattern in “cold” PMMA developer.
  5. Recoat the exposed pattern with 7 nm of Cu for charge dissipation.
  6. Re-insert the slide into the Rath e-beam tool and align to the LL corner and bottom edge.
  7. Focus on the surface by burning a spot near the LL alignment marker.
  8. Run a 3-Point alignment on the LL, LR and UR alignment markers in the array.
    1. Also activate the auto focus system and re-focus at each marker to account for sample tilt.
    2. Perform a write field alignment on a spot near P1 to align the electron beam scan to the P1, P2, and P3 markers.
  9. Open a new Position List
  10. In the right hand panel, select “Writefield” control icon to open “Scan Manager” and click to expand the “Linescans” menu.
  11. Choose or make an appropriate scan length linescan file and then left click, hold and drag it into the Position List; one for the u direction and one for the v direction.
    1. Design or make sure that the linescan is positioned to cross the marker at its center and position it about 80% of the way to the edge of the writefield.
    2. For noise reduction, make the width of the scan ~0.1 um to average the edge roughness of the mark.
  12. Set the u,v coordinates for both linescans to cross the legs of the first marker in the array.
  13. Create a matrix appropriate for the size of the substrate using the main command line “Filter” , “Matrix Copy” command to place scan pairs at all marker locations.
    1. Highlight both of the initial Position List before executing the matrix command so that both the u and v scans will be placed at each marker location.
  14. Highlight the first line in the position List and execute for a trial run.
  15. After the scan completes, double click or “Open” the Position List line just executed. This will open a graph of the scan intensity vs. position and show the profile of the scan.
    1. Click the Blue Square in the header line of the graph and open the Threshold Algorithm parameter window (Scan Properties).
    2. Adjust the parameters to allow proper identification of the edges of the marker. This may also require noise reduction – Expand the filter list and choose “Noise Reduction” then enter the number of iterations and number of points for smoothing the noise out.
    3. Repeat the data collection scan and adjust parameters as needed.
    4. A check of other markers might be necessary to help insure that all the markers are recognized.
  16. For capturing position data for all elements of the array, highlight all lines in the Position List then click on the Position List blue square to run them all.
    1. Save the Position List file to calculate the data on a remote computer having the Raith software.
  17. Perform the post analysis where the actual position of a marker is calculated from the linescan data, by opening the recorded Position List data file, clicking on “Filter’ in the top command line and opening “Linescan Analysis” in the dropdown menu.
    1. This step should calculate the marker position and fill in the blank spaces under Pos1, Pos2, and Pos3 in the Position List.
      1. Pos1 and Pos2 list the edges of the marker.
      2. Pos3 denotes the center of the marker limb.
  • Adjust the trigger level and other parameters to insure that all marker centers can be identified. (Must check several scan lines to determine error liability.)
  1. To calculate all the marker locations, highlight all lines in the position list and click on the top command line “Filter”, “Linescan Analysis” and “Apply” and “OK”.
  2. Verify data quality by checking the edge locations and consequent line center in a number of individual linescan lines.
  3. This above procedure represents the “Original” location of markers on a virgin film. For this test run, we will invert the same slide and repeat the above procedure, capturing data whose Lower Left origin is actually the upper right marker of the same slide.  The difference in locations between these two data sets would be the test case for a “Before” and “After” anneal for the “Squirmy” film of the real experiment.

Calculation of Error Vectors

Extraction of data from Raith software and insertion into EXCEL file.

  1. Open the saved Position List containing one of the measured marker arrays and click “Filter” on the top command line and then click on “Export” in the dropdown list.  This will bring up a small windows screen which allows you to select which of the title columns you want to export as a text field.  Choose u, v, and Pos3 as the only data columns you want.  Then save them into an appropriately named text file.
  2. Open EXCEL and have EXCEL open the extracted data file. Choose a comma delimited text file source in EXCEL and have it open the file.  Now you have numeric data to play with in EXCEL.

Calculation and plotting of error vectors

  1. Combine the two data sets onto one XCEL page and plot both data sets as individual pairs using their full coordinate value (distance from u,v = (0,0)). Color one red and one Black and label them “before” and “after” anneal.
  2. This is a bit tricky in EXCEL, so do it this way:
    1. After importing the data, set up the calculation columns to fix the data points in millimeters for the layout grid and then add 1000 times the error to match the mm scale units and make the error visible.
    2. Layout the two data sets in parallel columns with the v axes labeled and nothing over the u-axis column.
    3. Create the data presentation columns with a blank line between each coordinate pair.
    4. Highlight the first data set and select xy scatter plot, and then hit enter. This establishes the graph and places the base data set. Add graph labels.
    5. To add the second data set, highlight the second data set, copy it, and then paste it into the previous graph. The two data sets will have two colors so the direction of the error vector can be defined as “blue to red” for example.  (Cannot make EXCEL draw a line between the two different data set points.)

Conclusion

We have established and demonstrated a measurement procedure by which the location of a set of markers on top of an experimental polymer film can be measured before and after a thermal anneal so that the lateral polymer “squirm” can be determined to a rather high degree of accuracy using the measurement capabilities of the Raith 150 TWO e-beam lithography system in the UTD cleanroom.  A mathematical formula has been demonstrated to show a visualization of the squirm pattern.  Various degrees of pixel density can be utilized to reveal additional details of the squirm characteristics.

Appendix A  EXCEL Calculation

6/1/2015                                                                                                                                                                 Roger Robbins

EXCEL format after extracting selected data from Raith software and rearranging:

180 deg Mark Location (mm) 180 deg Raw Error (um) 0 deg Raw Error (um) 0 deg Mark Location (mm)
B C D E F G
U V Pos3′ Pos3 U V
0 0.04 0.24000 -0.13 0 0.03
0.04 0 -0.22000 -0.45 0.03 0
10 0.04 10000.07000 10002.34 10 0.03
10.04 0 -0.12000 -1.88 10.03 0
20 0.04 20000.13000 20001.37 20 0.03
20.04 0 -0.24000 -2.18 20.03 0
30 0.04 29999.72000 30000.26 30 0.03
30.04 0 -0.44000 -1.57 30.03 0
40 0.04 39999.44000 39999.97 40 0.03
40.04 0 -0.48000 -1.33 40.03 0
50 0.04 49999.35000 49999.91 50 0.03
50.04 0 -0.56000 -0.92 50.03 0
60 0.04 60000.08000 59999.88 60 0.03
60.04 0 -0.28000 -0.36 60.03 0
0 10.04 0.03000 -0.99 0 10.03
0.04 10 10000.26000 9998.67 0.03 10
10 10.04 10000.34000 10000.12 10 10.03
10.04 10 10000.22000 9999.08 10.03 10
20 10.04 20000.30000 20000.1 20 10.03
20.04 10 10000.02000 9999.59 20.03 10
30 10.04 30000.14000 29999.97 30 10.03
30.04 10 9999.97000 9999.86 30.03 10
40 10.04 39999.80000 39999.48 40 10.03
40.04 10 9999.60000 9999.69 40.03 10
50 10.04 49999.88000 49999.83 50 10.03
50.04 10 9999.56000 9999.86 50.03 10
60 10.04 60000.31000 60000.07 60 10.03
60.04 10 9999.77000 10000.13 60.03 10
0 20.04 -0.01000 -0.13 0 20.03
0.04 20 20000.82000 19997.69 0.03 20
10 20.04 10000.18000 9999.55 10 20.03
10.04 20 20000.39000 19998.34 10.03 20
20 20.04 20000.48000 20000.28 20 20.03
20.04 20 20000.25000 19999.44 20.03 20
30 20.04 30000.25000 29999.99 30 20.03
30.04 20 20000.16000 19999.13 30.03 20
40 20.04 39999.98000 39999.73 40 20.03
40.04 20 19999.94000 19999.48 40.03 20
50 20.04 49999.37000 49999.8 50 20.03
50.04 20 19999.89000 19999.83 50.03 20
60 20.04 60000.46000 59999.96 60 20.03
60.04 20 20000.11000 20000.02 60.03 20
0 30.04 -0.14000 0.11 0 30.03
0.04 30 30001.75000 29997.39 0.03 30
10 30.04 10000.21000 10000.44 10 30.03
10.04 30 30001.17000 29997.93 10.03 30
20 30.04 20000.54000 20000.45 20 30.03
20.04 30 30000.86000 29998.37 20.03 30
30 30.04 29999.26000 29999.96 30 30.03
30.04 30 30000.71000 29998.69 30.03 30
40 30.04 39999.87000 39999.67 40 30.03
40.04 30 30000.56000 29999.16 40.03 30
50 30.04 49999.98000 49999.68 50 30.03
50.04 30 29999.72000 29999.4 50.03 30
60 30.04 60001.35000 59999.89 60 30.03
60.04 30 29999.70000 29999.87 60.03 30
0 40.04 -0.26000 0.12 0 40.03
0.04 40 40002.86000 39996.68 0.03 40
10 40.04 9999.72000 10000.96 10 40.03
10.04 40 40002.89000 39997.26 10.03 40
20 40.04 19999.85000 20000.73 20 40.03
20.04 40 40002.65000 39997.74 20.03 40
30 40.04 29999.57000 30000.07 30 40.03
30.04 40 40002.42000 39998.38 30.03 40
40 40.04 39998.44000 39999.12 40 40.03
40.04 40 40002.26000 39999.2 40.03 40
50 40.04 49997.62000 49999.88 50 40.03
50.04 40 40001.70000 39998.81 50.03 40
60 40.04 60000.08000 59999.71 60 40.03
60.04 40 39999.63000 39999.66 60.03 40

Formulas for Calculation of presentation data from above EXCEL format

The key in presenting a small error from a design location on a large-scale design measurement grid is to exaggerate the error size so that is visible on the graph and shows the pattern of the Squirm for visualization.  For the EXCEL file calculation with the design grid u,v data listed in two columns in units of (mm), and the tiny error position listed in a third column in units of microns with the design grid  u or v position included, the following formula concept is needed to produce the desired visual display of “Squirm”.

u = D11 – (1000*B11) + B11      (EXCEL nomenclature)

Where D11 is the measured error u position with the design location included (um), and B11 is the design location (mm)

Formula description:  “subtract 1000 times the designed marker location in (mm) from the error location in (um) then add back the design location in (mm) à error location in units of (x 1000 um err) from design location in (mm)”.

For comparing the “before annealing” data to the “after annealing” data, two measurement data sets must be calculated and compared on the same graph, using the above formula.

Resulting Presentation Data calculated by above formulas:

Error vector from 0 deg Columns (mm + 1000*(um)) Error vector from 180 deg Columns (mm + 1000*(um))
u v u v
0 deg 180 deg
0.24 -0.22 -0.13 -0.45
10.07 -0.12 12.34 -1.88
20.13 -0.24 21.37 -2.18
29.72 -0.44 30.26 -1.57
39.44 -0.48 39.97 -1.33
49.35 -0.56 49.91 -0.92
60.08 -0.28 59.88 -0.36
0.03 10.26 -0.99 8.67
10.34 10.22 10.12 9.08
20.30 10.02 20.10 9.59
30.14 9.97 29.97 9.86
39.80 9.60 39.48 9.69
49.88 9.56 49.83 9.86
60.31 9.77 60.07 10.13
-0.01 20.82 -0.13 17.69
10.18 20.39 9.55 18.34
20.48 20.25 20.28 19.44
30.25 20.16 29.99 19.13
39.98 19.94 39.73 19.48
49.37 19.89 49.80 19.83
60.46 20.11 59.96 20.02
-0.14 31.75 0.11 27.39
10.21 31.17 10.44 27.93
20.54 30.86 20.45 28.37
29.26 30.71 29.96 28.69
39.87 30.56 39.67 29.16
49.98 29.72 49.68 29.40
61.35 29.70 59.89 29.87
-0.26 42.86 0.12 36.68
9.72 42.89 10.96 37.26
19.85 42.65 20.73 37.74
29.57 42.42 30.07 38.38
38.44 42.26 39.12 39.20
47.62 41.70 49.88 38.81
60.08 39.63 59.71 39.66

Graph of vector errors plotted on Design Grid:

Simulated Film movement from original position (blue) to annealed position (red).