Back to JEOL JBX-8100FS E-Beam Writer page
Videos are for demonstration and learning assistance purposes but are not frequently updated. You must refer to the text of the SOP for full, proper, up-to-date information on operation, information, and important considerations. If there is a conflict between the content in the videos and the SOP text, the SOP text should be followed.
Text in the SOP is somewhat color coded to assist the reader:
"Normal" text is black (or grey, in dark mode)
Important reminders and the names of buttons to press are in bold
Links are blue.
⛔ Safety, points of potential tool damage, or critical reminders are red with the “no entry” sign emoji to the left.
Background informational notes are purple.
Use the drop down menus to get more information on each step of the SOP.
0 Software Update Notes - 20221215
Click to expand for detailed step-by-step and video links
Helpfully, the machine will now ignore presses of the LOAD/UNLOAD button when it is not green.
Pressing LOAD/UNLOAD at the wrong time (when there is not a cassette on the stage at the ORG position, or when no cassette is on the stage and cassette is in the load lock) will no longer trigger an error that requires Bill or I reset via Manager mode.
HOWEVER, you can still cause vacuum problems if you press the EVAC/VENT button when a cassette is on the stage.
ONLY press EVAC/VENT if the “Cassette on Stage” light is off!
Due to changes in the underlying software, any “old” MGN files that you would like to re-expose will need to be remade “fresh” for future exposures.
This can be done as normal with the purple SCHD square (or by re-creating in Job Maker, if you do it the old way).
Contact me if you have any questions.
There is now an “Origin” button to move the stage+cassette to the ORG/Origin position.
This is instead of the old method of selecting ORG from the Fixed Position drop down, and clicking the MOVE button.
Annoyingly, it is very close to, and looks very similar to, the MOVE button. Be sure to tell them apart:
There is no longer an ORG position under the ‘Fixed Position’ drop down:
I have added an “Origin (ORG)” entry to the ‘User Def. Position’ drop down in case it’s needed…but you should just use the new Origin button:
The new “Origin” button has a major bug.
If the HEI button is clicked and height measurement fails, the Origin button will be stuck greyed out.
To un-grey it:
In the Fixed Position drop down, select BE.
Click the MOVE button.
Click the HEI button to successfully measure the height of the BE mark.
Unclick the HEI button to stop height measurement.
The Origin button will be un-greyed and is then usable again.
There is now a “Pause” button in the Exposure page, but it’s not useful to us as far as I can tell.
The Pause button pauses exposure after an entire Layer is complete, which means it will not pause in the middle of a V30, in the middle of an array of the same V30, or in the middle of an array of different V30s. We usually put all the files in the same Layer (unless maybe if you’re exposing on different windows within the same MGN?).
If you do use it, I also can’t tell the difference between the “Do not unload” and “Unload” options.
When loading/unloading the cassette with LOAD/UNLOAD, previously both IV2 and IV3 were closed. Now only IV3 closes.
Per JEOL, this is intentional and will not cause problems on the system.
The Manual Loader Viewer window will look different when you’re loading/unloading a cassette:
These changes have not yet been updated in the SOP, I will integrate them as soon as I'm able.
1 Training Materials
2 Data Preparation
2.1 Layout
Create a layout in GDS or OAS formats (can also be others) using something like KLayout, LayoutEditor, L-Edit, or GdsPy.
Click to expand for detailed step-by-step and video links
Have a layout in GDS, OAS, DXF, CIF, etc (OAS is far superior to other other formats but it is underutilized). These can be created with a variety of programs:
2.2 8100calculator.xlsx
Use 8100calculator.xlsx to help determine your Shot Pitch and other write parameters.
Click to expand for detailed step-by-step and video links
The 8100calculator.xlsx file is meant to help you select your current, shot pitch, beam step size, ensure your required clock speed is under the maximum clock speed of the tool, and estimate the write time of your pattern.
8100 Calculator video notes - Since the video was recorded, the file has been updated to:
Fix a bug in the calculation of the max. clock speed. The predicted max. clock speed from 8100calculator now matches that on the 8100 software.
The number of points that the beam steps has been clarified to be the "shot pitch" (as in Beamer) and the size of the beam step in nanometers has been clarified to be the "beam step size". This is now consistent with the rest of the materials.
Video 2.2: 8100calculator.xlsx
Download and open 8100calculator.xlsx or use a previously downloaded copy.
If it opens in Protected View, click Enable Editing.
White boxes can/should be edited, grey boxes are locked (and can be unlocked with the password specified in the file), and the Shot Pitch and Req. Clock Speed boxes will be green if acceptable values are input, or red if unacceptable values are input.
Use the file to help pick a good Shot Pitch and ensure your exposure parameters will not exceed the maximum clock speed of 125 MHz.
In the Beam Current box, use the drop down to select the write current
The theoretical minimum beam diameter is displayed directly below the current.
In the Minimum Dose box, input your planned minimum dose for your exposure.
For PEC exposures, this will be the smallest dose multiplier times your base dose. For dose arrays, this will be your smallest dose (accounting for PEC adjustments, if necessary), and for non-PEC and non-dose array exposures, this will just be your base dose.
In the Pattern Unit box, by default leave this at 0.5 nm (needs to equal the value used in the Beamer Export module used to create your V30 file).
This can be changed if necessary, but typically should be left at 0.5 nm and the Beam Scan Step/Shot Pitch should be varied.
In the Beam Scan Step box, input your Scan Step (in points or pixels) to match your selected value in the Beamer Export module used to create your V30 file.
The resulting Shot Pitch (Pattern Unit (nm) x Beam Scan Step (pts) = Shot Pitch (nm)) is shown, and is highlighted in green if less than the minimum beam diameter, and red if greater than the minimum beam diameter.
Ideally, the shot pitch should be ~3-4x smaller than your minimum feature size.
For small features: The larger the shot pitch, the larger the edge roughness and feature size non-uniformity.
A shot pitch larger than the beam diameter will usually give poor features, hence the red highlighting. It would be good to use a shot pitch at least 1/2 your beam diameter, but doesn't need to be smaller than 1/4 your beam diameter.
The required clock speed will also be shown, and needs to be ≤125 MHz.
This will be highlighted in green when ≤125 MHz, and red if > 125 MHz.
This absolutely needs to be < 125 MHz or the machine will not be able to expose the pattern, and will throw an error.
You can also use the file to very roughly estimate your write time
In Pattern area per chip, enter your pattern area from Beamer.
Write time is proportional to pattern area.
In Average Dose, enter your base dose, or a more appropriate dose estimate if available.
Write time is proportional to dose.
In Fields in each chip, enter the number of write fields in each chip (as shown by Beamer for your V30 file)
Motion between fields takes ~0.25 s on the 8100.
In Number of chips, enter the total number of chips in the exposure.
For exposure of a single pattern/chip, this will be 1.
In the lower section, you will see estimates for E-beam write time and stage motion time, and the total write time.
Here, the estimated write time is based only on beam on time (the necessary time for the beam to be on to achieve the average dose on your pattern area). In reality, there are additional time components due to subfield settling, polygon placement settling, beam blanking, etc., which are not included in this estimate. PEC may increase the write time due to these effects.
The stage motion time is just 0.25 s x Total Fields. This may be large for large chip area but small exposure area (sparse) patterns and small field sizes (e.g. 100 μm by 100 μm).
Total write time is the addition of the E-beam write time and Stage motion time.
If desired, save the file for your own later use.
2.3 Tracer (Optional) - Create PSF for PEC
If performing PEC on a new materials stack, use Tracer to export a PSF for import into Beamer. Tracer is available on the PC next to the JEOL itself.
Click to expand for detailed step-by-step and video links
Tracer can optionally be used to create a point spread function of your exposure for use in Beamer with Proximity Effect Correction. This is not mandatory, but is a good idea when you're starting out to get a feel for scattering and start using PEC on your patterns. You will have a much easier time finding doses when your patterns change if you have been using PEC and know the base dose for your patterns/substrate/etc. If you're interested in learning about advanced PEC, refer to the Genisys Beamer PEC Webinars.
Video 2.3a: Tracer - Scattering vs. Vacc
Video 2.3b: Tracer - PSF Creation
2.4 Beamer - V30
In Beamer, Import → Heal → Export. In export, set shot pitch as appropriate. Then Run or Run to last box. For info on more advanced operation, watch the Beamer Webinar Material. Beamer is available on the PC next to the JEOL itself, and a remote computer (optional instructions for accessing the remote computer here).
Initial Setup for Beamer that may be needed to see custom export modules
Open Beamer
Go to File → Properties
In the Properties window that pops up, go to the Directories tab
In the bottom section titled Custom Module Configuration Directories, locate the Global parameter.
Set the Global parameter to: C:\Users\Public\Documents\.GenISys\Repository
Click OK, then restart Beamer.
After restarting, click the black triangle in the bottom right of the Export Module. Click and drag the 8100 Export module to your Beamer flow and use as normal.
Click to expand for detailed step-by-step and video links
We use Genisys Beamer to convert files from standard formats (GDS, OAS, etc.) into a format that can be understood by the 8100. Beamer is incredibly powerful, and the steps below just scratch the surface of what it's capable of.
If you're interested, it's suggested that you read the manual for more information (in Beamer, via Help > Download Documents, and Help > Formatter Notes > JEOL) or refer to the Beamer Training Videos accessible on the Beamer PC.
If you try to open Beamer and get an error that "All licenses are in use", some else is logged into the PC and still has Beamer open. You can restart the computer to clear their open session.
A video of simple V30 conversion can be seen in the 3.7 VisualJob (Beamer) and Magazine File section. A more extensive video will be added in the section in the future.
Convert your layout file into a V30 file (converted design usable by the JEOL) with Beamer. This is just an example of a beginner's flow.
Open Beamer on the JEOL Beamer PC or the Remote Beamer PC by searching for "Beamer" in Windows.
Usually the most up to date version of Beamer is the only one installed. This is not always the case though.
Import your layout
Drag the Import module into the flow space
Most modules we'll use will be in the Layout Operation section.
Select and open your layout file.
In the Import Layout window, you may want to specify which layers to import using the format #(#) (e.g. 0(0), 1(0) ).
Note that this will redefine the boundaries of your file to the maximum extent of your imported layers. In general, it's a good idea to put 4 very small (~10 nm) squares at the edges of all of your layout layers so that extent changes are not a problem.
Click OK.
View your imported layout with VIEWER
Click the Run to Icon (Play button with a Line) to execute the flow until that step.
Then click the View icon (Box in a box).
This will open up VIEWER, Beamer's built in layout viewer.
Inspect your pattern, an ensure it was properly imported.
Some VIEWER tips:
Recenter the view: Left click.
Zoom In: Click and hold, and then drag right to specify an area. Or mouse scroll wheel, which zooms based on the pointer position.
Zoom Out: Click and hold, and then drag left to specify an area. Or mouse scroll wheel, which zooms based on the pointer position.
Measure a distance: Right click, move the specified distance, and then right click again. By default, this measures on horizontal, vertical, and 45 degree lines, and will try to snap to your shape edges.
When finished, click the X in the upper right to close VIEWER.
Heal any layout overlaps
Drag the Heal module under your Import module, and get close enough that the two automatically connect with a grey line. Once, they're connected, release the mouse.
Heal will eliminate overlaps in the design (which would otherwise be double exposed) and closes boundaries prior to fracturing. It's usually a good idea to do this on all of your patterns, unless you have specific reasons to not.
Double click on the Heal module to open up its property window.
The default options are good for now, but check the Beamer manual for more information on what these mean and what they do.
Once satisfied, click the OK button.
On the Heal module, click the Run to Icon, and then the View icon.
Inspect the healed pattern as needed in Viewer, and exit with the X in the upper right when finished.
Optional: Run PEC
In the Process Correction section, drag in the PEC module, and connect it to your Heal module.
PEC isn't necessary to use with every pattern, but it's a good idea to use it when you're getting started to see the effect of scattering and proximity effects. These are an ever present fact of life at high kV EBL.
Under the PSF Representation, click the Numerical PSF radio button.
Under PSF File Name, click the Browse... button and open your lpsf file from the Tracer.
There are a number of important parameters in PEC, but for now we'll just do some long range dose correction, which account for backscattered electrons.
Click the OK button.
On the PEC module, click the Run To icon, and then the View icon.
In the VIEWER window that pops up:
You'll see that your pattern is now sliced into a bunch of pieces.
Click the Doses tab (two left of the Layer tab in the upper right). This will display the relative doses of your pattern compared to a base dose.
Click the "Color by Dose button" in the second layer of icons (Looks like 3 nested blue U's). This will color your pattern by the dose received by each feature.
Explore your pattern and note the effects of backscattered doses. Typically, isolated features will receive ~1.5x the base dose, and features with a lot of nearby features may receive a dose of <1.
Exit with the X in the upper right when finished.
Insert the export module
In the Layout Operation section, drag in the Export module, and connect it to your PEC module.
In the Save File popup window, enter a name for your V30 file.
In general, I suggest names that begin with a letter, don't have too many special characters, and aren't too long, because sometime the Linux machine can have issues.
After you've entered a name, click the Save button.
Set Export JEOL window parameters
Ensure that Machine Type is set to JBX-8100FS
Ensure EOS mode is set to 3.
Ensure Pattern Units (nm) are 0.5 nm.
You can change this to 1 nm if you'd prefer (or something bigger, though there's no real need), though we'll assume 0.5 for this document.
Set the Beam Scan Step (a integer number of points or pixels) from 8100calculator.xlsx. For this training example, we'll use 10.
The Beam Scan Step (called Shot Pitch integer in Beamer) x Pattern Unit = Shot Pitch (nm). The Shot Pitch is the distance in nm that each shot will be placed apart, and which ideally your pattern will be on an evenly divisible grid of.
Leave Size (um) at 1000 and 1000 (this is the field size)
Size (um) is the field size, by default it's set to the maximum of 1000 um by 1000 um. You can make this smaller to fit your pattern and your needs. It does not need to be square.
In Fracturing mode, leave LRFT.
LRFT is a good default. Curved may be helpful if you have a lot of curved structures.
Leave unchecked Shot Pitch Fracturing, Slim Trapezoids to Single Lines, and Symmetric Fracturing.
These can be very useful in relevant situations, read up on them in the Beamer manual.
Leave Feature Order at ArrayCompaction (for this example)
This section is very important, it determines how your features are exposed within the write field. Different Feature Order methods will be ideal for different situations. You'll want to read about the different options for Feature Order in the Beamer Manual.
NoCompaction doesn't change the feature sorting, and isn't usually the most useful.
ArrayCompaction is good for the reducing the size of the V30 file for large array.
WritingOrder tries to reduce beam jumps by exposing nearest neighbors.
FollowGeometry tries to expose contiguous patterns at once, which is great for waveguides. Regional Traversal is not applicable here.
Layer order allows you to specify a specific order exposure to the layers. If none is selected, it is first in, first out.
Region Size: From the Beamer Manual (v5.8): Region Size [μm] reduces the area in which features are exposed successively. Selecting this will preserve existing Arrays and also run an additional compression step to detect arrays. This command allows to significantly reduce the file size and fracturing time of large, homogenous arrays
Click the OK button.
Export V30
On the Export module, click the Run To icon. Your V30 is now created and saved.
Save your Beamer flow by going to the File menu, then Save As. It will be saved as a .ftxt file. This can be reused for future writes.
Optional: View V30
We can get a good amount of information from viewing the V30 file.
On the Export module, click the View icon.
Click the Doses tab, and Color by Dose button. This will show the patterns by their relative dose multipliers.
Under the E-Beam drop down, click Show Traversal path. This will show the order of the fields to be written.
If the fields or the field traversal plan looks problematic or sub-optimal, you can go back into the Export module, under the Advanced tab, and use more advanced Field Ordering. Selecting Floating and checking Center to Field may be useful for many patterns.
Under the E-Beam drop down, click Show Writing order. This will show the order of the fractured primitives will be written.
If the beam is going all over the place in this, you may want to go back and change the selected Feature Order in the Export Module, and then re-export the V30.
Under the E-Beam drop down, click Subfields.
This will show individual writing subfields where the beam will settle. These generally add overhead time, but are necessary for PEC.
Zoom far into a feature. Then under the E-Beam drop down, click Show shots.
This shows how Beamer thinks the individual shots will be distributed within each primitive.
Exit with the X in the upper right when finished.
Optional: Re-export V30
If inspection of the V30 file after export encouraged you to make changes to your setup, double click the Export module (now re-named Out JEOL52).
Make the needed changes (including changing the File Name, if desired) and click OK when finished.
Exit Beamer
To close Beamer, exit with the X in the upper right when finished.
⛔ This is critical to do, as other users will not be able to use Beamer while you have it open.
2.5 Beamer - VisualJob creation of SDF and JDF files
Setup write parameters for the particular V30 file(s) to be written (Substrate → Global Mark → Layers → Arrays → Data to Place). Save, then Generate Jobdeck to make JDF and SDF files.
Click to expand for detailed step-by-step and video links
A video of simple visual-Job use for both aligned and unaligned writes can be seen in the 3.7 VisualJob (Beamer) and Magazine File section. A more extensive video will be added in the section in the future.
While DAILYCAL, runs (or whenever), on the Beamer PC, launch Beamer and open visual-Job.
It is recommended that you use visual-Job to create your SDF and JDF files while the machine completes the DAILYCAL batch calibration, but this may be done prior to this, or any time before needing a magazine file for exposure.
In general, the mouse-over tool tips are very useful in VisualJob.
After Beamer is loaded, Open visual-Job with either the visual-Job button (right of the Detach button) or with the visual-Job module under Layout Operation.
With the visual-Job window open, create a New Jobdeck.
This can be done in multiple ways: Ctrl+N on the keyboard, the New Jobdeck button, or File > New Jobdeck.
Save the new Jobdeck file in a new folder with a name that begins with a letter, and contains only letters and numbers.
It's recommended that you make a new folder for your Jobdeck, JDF, and SDF files. Copy your V30 and JDI files into this folder. Using this method will keep all of your information together.
By default, the name of the Jobdeck file will be the name of the schedule file and the schedule name. There are restrictions on the types of names that can be used on the Linux PC for SDF files. Because of these, please save the Jobdeck project with a name that begins with a letter, and contains only numbers and letters. If you do not do this, after saving, rename your Schedule File and Schedule Name appropriately.
The Jobdeck Project (.Jobdeck) is a file that visual-Job will use to save your information. It is not needed on the JEOL Linux PC.
Once you've saved the new Jobdeck file, ensure the names of the Schedule File and Schedule Name are appropriate.
On the left menu, under visual-Job Project and under your Jobdeck project name on the left, click the Substrate listing to bring up the Holder and Substrate page.
Set the Type, Size (in), and Window for the holder to be used for the exposure. Refer to the 4.1 Cassette Window Identifiers table for more information.
By default, leave the 'When Creating Jobs' drop down at "Write all chips / Ignore substrate boundaries"
This may need to be changed to either "Only write chips all or partially within the substrate" or "Only write chips entirely within the substrate" if you are executing an aligned write. For aligned writes, if the machine cannot measure the height at a mark, or cannot detect the mark, it may cause the exposure to fail, or may write the chip with significant misalignment.
Click the Global mark listing in the menu on the left to bring up the Global Mark page.
If not using physical marks for alignment, set Mode to NONE(C).
When NONE(C) is selected, the status of the Measure Height box is ignored.
If using physical global marks, select Semi-Automatic(S) and check Measure Height.
Input the design mark positions of your marks into the table. Note that you may either use 1 mark (P), 2 marks (P and Q), or 4 marks (P, Q, R, and S), but 4 marks are recommended.
Then set the Width (um), Length (um), Type, and Rotation (deg) of the P mark and the Q mark.
Type will be + for a cross mark, or L for an L-shaped mark. Cross marks are recommended.
Rotation (deg) really only applies to L-shaped marks. This should be left at 0 for cross marks.
Click the Layers listing in the left menu.
Each Layer will become a unique JDF file.
Layers may share patterns with the same EOS (Current), PATH (exposure calibration), Shot Pitch, Base Dose, type of Local Alignment marks, and substrate origin offset.
If you have patterns in which one of these parameters changes between layouts, you will need to create a new layer. Use the Insert and Delete buttons to add/remove layers, and use the Enable box to enable/disable layers if needed. You can right click a layer name in the left menu and select "Duplicate Layer" if needed.
Click the layer name (e.g. Layer_1) in the left menu
Layer properties should be edited from the Layer Parameters window, found by clicking the layer name (e.g. Layer_1) in the left menu.
The JDF File Name and JOB Name default to the name of the SDF file, with the Layer ID changing. You may edit the JDF File Name or JOB Name if you desire, but this is not required.
Set the EOS drop down to your desired current (Condition File).
⛔ Warning: The drop down here can be changed with the mouse scroll wheel. It is easy to change the EOS, then click, and then try to scroll down the page with the scroll wheel. This will change the EOS (e.g. scrolling up from the Purdue_30nA that you selected up to Purdue_2nA). So be sure to double check this setting after you're finished in the Layer Parameters window.
Set the PATH, which should almost always be DIRE20. If you are doing a write that will take less than 20 minutes, you can select DIRE00 to omit drift correction.
Set the Shot Pitch to the same value you used to create your V30. (JCW note - I don't know what happens if you select a different shot pitch than the V30 used). The Beam Step Size in nm is displayed to the right, and should also match what you used for your V30.
Set the Area Base Dose to your chosen exposure dose, or the base dose if you're using PEC.
If not using single lines, you may leave the Line Base Dose at 1000 nC/cm (it will not cause problems). If using single lines, set this to your desired dosage in either nC/cm or µC/cm2.
Set Local Align to correspond with your plan for chip alignment marks:
1-Mark(1) for a single M1 physical mark.
4-Mark(4) for four physical chip marks: M1, M2, M3, and M4. This is recommended if you want the best possible alignment.
Height Only -1(v1) for a single virtual mark. This will be the default for unaligned writes.
Height Only-4(V4) for four virtual marks. This is good for larger V30 files on substrates that are potentially bowed. It will measure the height at 4 points (usually we'll specify the corners), take the average, and apply it to the entire V30.
⛔ You should never select NONE(0) or SEM(S). You MUST do some form of local mark to perform height detection and get proper focus of the beam. Unaligned writes should always use V1 or V4. If you have physical global marks but no physical chip marks, you can use the global marks as chip marks (mode 1 or 4) to get a better alignment than using V1 or V4. Or just use V1 or V4 to get reasonable good alignment accuracy and also have proper focus.
Substrate Origin Offset X/Y corresponds with the OFFSET line in the SDF. When the magazine file is made, the SDF OFFSET will be where the center of the pattern is written, in substrate coordinates. It must be updated either after you are done with VisualJob via the SDF file itself, or can be set here in VisualJob if you already know it. Either method is fine.
Advanced settings usually do not need to be changed, but have useful settings for those doing aligned writes.
The first drop down will select the behavior of the exposure if a chip alignment mark location fails. It can be set to "Always write this chip" or "Skip this chip".
The lower box sets when to abort the exposure of the layer entirely, after the set number of consecutive chips have failed detection.
The Measure Height box should always be checked.
If desired, you can right click on the Layer_# identifier and select Duplicate Layer.
Typically, you'll only use one Layer per SDF file. The exception would be if you had different shot pitches in V30 files in the same SDF (or other weird things, which probably aren't good ideas but could be if you've thought it through). While you could create different layer with different base doses, it's much easier to modify this in the Arrays listing to come.
If you're interested in a dual current write, talk to BNC Staff. There is currently not a standard way to do this, and it's usually just easier to pick an intermediate current. But there could be good reasons to do a dual current write.
Click on the Arrays listing in the left menu.
Click the Add Array button in the top right.
Under the Data to place drop down, select Chip.
Click the Browse... button. Select your V30 and click Open.
The array will default to a 1 by 1 array centered at the Origin with a Pitch equal to the V30 dimensions.
Leave the array settings as they are for exposing a single chip.
To expose an array here, modify the position, repetition, and pitch as desired.
If performing a dose array, leave enough room for back scattering between chips (3*Beta, usually 100 µm will be enough).
The Position Mode sets which chip in the array will be placed at the Position location: the upper left chip, the center chip, or the lower left chip.
If you would like to do a dose array, click the Assign Dose... button:
This is by far the best way to perform a dose array.
Chose an array based on entering absolute doses (e.g. 800 µC/cm2, 1300 µC/cm2, etc) or relative doses (multiples of the base dose previously specified in the Layer Parameters window, e.g., x0.8, x1.3, etc.).
Fixed Dose will set the same dose to every chip in the array.
Individual Dose will let you enter in the dose for each chip.
But what you should use for a dose array is the Dose Series. Click Dose Series, and :
Assign the starting and ending dose values (or starting and step, if you'll use a Linear Step Step Mode).
Set the Step Mode: Linear Range or Logarithmic, as preferred if a start and end dose will be specified. Or pick Linear Step and assign the starting dose and dose step.
Pick whichever deploy mode makes the most sense to you.
You should take a picture or screen shot of the dose array display to later remember what you assigned.
Click OK.
Add in a Comment to remember significant info if you desire. This will only be stored for viewing in VisualJob.
Click OK
Inspect the position of the array and of your patterns in the Viewer window in the bottom half of the screen.
You can measure distances here by right clicking at the start and end of the measurement.
The assigned dose will be displayed on each chip.
If your pattern is very complex, you can press the "Show Chip Outline Only" button to turn off viewing the pattern in each chip.
You can press the Toggle Mode button to toggle between Pick/Measuring in the Layout and Edit Jobdeck mode.
In the Edit mode, you can left click a chip to select it, and then right click to view useful options.
Assign Dose... will let you change the dose of that chip.
To remove a chip from the array, select Clear Chip(s). Cleared chips will not be exposed.
You can re-insert a cleared chip with the Insert Chip: Filename option.
You may want to edit the Array Positions to remove a subset of chips or adjust doses, if that is easier than in Toggle Mode.
You can adjust the position, repetition, pitch, and position mode in the Arrays listing. You can also delete an array, or add additional arrays.
Next we'll need to assign chip mark locations to each chip. Click the Data to Place expand button in the left menu.
Click the name of a chip.
Enter in the design location (relative to the chip coordinate system) of the chip marks. If using virtual marks:
In V1, it's customary to place M1 at the chip origin (0,0)
In V4, it's customary to place the marks at the edges of the chip. M1: -X/+Y, M2: +X / +Y, M3: +X / -Y, M4: -X / -Y.
Also enter the width and lengths of the marks.
If using virtual marks, these values are place holders. Just pick reasonable values (e.g. Width = 1 or 10, Length = 10 or 100).
Chip mark (and global mark) locations will be visible in the Viewer.
Double check the viewer to ensure the alignments are set as expected, the marks are where they should be, etc. Make any needed adjustments to positions before proceeding.
Remember: The patterns will be offset in the exposure by the OFFSET you'll put in, or what you placed in Substrate Origin Offset. The Substrate Origin Offset is not reflected in the Viewer position, but will be reflected after the MGN is made and loaded into the exposure tab.
Save the Jobdeck with Save icon, or File → Save.
Jobdeck files may be modified and reused for later writes. If changing too much, it's recommended to just start fresh though.
Press the Generate Jobdeck button (red play arrow into a vertical line, right of the Save icon) to create the job files.
You should have the jdi file for each V30 file or an error may occur.
The JDF and SDF files will be created in the folder the Jobdeck file was saved in. The location of the V30 (if not in the same place, which is recommended) will be displayed in the Generate Jobdeck Log.
You can view the JDF and SDF files if desired, they are text files.
3 Machine Operation
3.1 Starting
Enable the JBX-8100FS in the BRK Lithography Core iLab Kiosk
Bring up the jbxwriter software. It may be minimized or on a different desktop tab (check the boxes in the lower right).
Check if a cassette is already on the stage, or if one is in the load lock. Proceed accordingly.
Check the set condition file; change the if necessary (Ensure Restore and DEMAG are checked)
Click to expand for detailed step-by-step and video links
Video 3.1: Starting
Enable the JEOL JBX-8100FS in the BRK Lithography Core iLab Kiosk.
Open jbxwriter software.
It may be minimized or on a different desktop tab (check the boxes in the lower right).
⛔ Never try to open a new instance of jbxwriter, this may cause tool communication problems.
Bring up the jbxwriter software, which should already be running. If you can't find it or it's not running, submit a support ticket.
⛔ When you arrive at the tool, there should not be a cassette on the stage or in the load lock (LL), and the load lock should be evacuated. But always check if a cassette is already on the stage
In jbxwriter, in the upper right corner, go to the Manual Loader Viewer tab.
Observe the image. The IV values should be closed (X's) and the stage and load lock positions should both be white.
If there is a cassette in the machine, either the stage or LL position will be will be a medium blue
Confusingly, the 'Cassette Present' medium blue is not labeled. It is a blue in between the labeled Evacuation/Venting light blue and Atmosphere dark blue.
If a cassette is on the stage or in the LL, proceed to the 3.9 Unload Cassette from Machine section.
Check the currently set condition file
In jbxwriter, under Information > Cond. Name, check the loaded Condition File
The condition file sets the write current, and automatically loads the appropriate column settings and objective aperture.
The currently offered currents are 2, 10, 30, and 100 for HT/4th Lens/EOS 3. A 0.5 nA file is offered for HR/5th Lens/EOS 6, but you this mode requires significant settling time to change to from EOS 3, and you should email jcwirth@purdue.edu if you are interested in using HR/5th Lens/EOS 6.
These are listed as Purdue_2nA, for example.
Only use condition files that begin with Purdue. Other files are not maintained.
Lower currents (2, 10 nA) should be loaded at least ~20 minutes before calibrating. ~40 minutes of settling time is required before the machine fully stabilizes. Higher currents (100 nA) may not require quite so much time.
Change the condition file if necessary before proceeding
Under Condition page > Condition Setting tab > Condition File Loading section, press the Select… button.
Choose the current for your exposure (remember, only use condition files that begin with Purdue). Click OK, and then OK.
Ensure ‘Restore’ and ‘Demag’ are checked, click the Execute button, and OK to execute EOSSET.
'Demag' executes a degaussing on the column lenses. Changing currents usually changes setting on Lenses 2&3 (Zoom lenses, which set the current), the objective aperture (which also helps to set the current), and Lens 4 (the focusing lens).
If changing between 2 nA and 10 nA, or vice versa, you may leave 'Demag' unchecked because these files are setup to use the same Lens 2/3 values and similar Lens 4 values, only changing the aperture. But if you do this, be sure to recheck it after loading the current so it is checked for the next user.
This will take ~2 minutes, wait until software is un-greyed and proceed.
3.2 Mount Sample onto Cassette
3.2.1 General Cassette Reminders
Click to expand for detailed step-by-step and video links
Video 3.2.1: Cassette Reminders
Put on a clean pair of latex gloves.
Determine appropriate cassette for loading in the cassette stack:
Piece cassettes: There are two piece cassettes, these are identical.
For pieces, you can load onto any of the three mounting areas. While you can put multiple pieces on the same mounting area (if sample thicknesses are the same), you won't be able to adjust their rotation independently. It's best to put each piece on its own window.
Wafer cassettes:
3"
4"
6"
⛔ Continue with care! The cassettes have many sensitive pieces, and are themselves sensitive. They are precision machined, costing up to $30,000 each. Be gentle and careful with them, and be sure not to roughly bump or drop them.
3.2.2 Wafer Cassette Mounting
Load wafer into cassette, ensuring wafer is flat and well aligned
Click to expand for detailed step-by-step and video links
Video 3.2.2: Wafer Cassette Mounting
Place cassette upside down on cleanroom wipe
Remove backside holder, and place it top side up to ensure grounding pins are not damaged
For 4" and 6" cassettes: First rotate the lock at the top of the holder ~120° counter clockwise (CCW).
With two fingers, press down towards the edges of the V wing spring and rotate until it's free of the cassette overhang.
Carefully lift the backside holder up and place it top side up on a cleanroom wipe.
⛔ There are two grounding pins on the backside holder which ensure the wafer backside has an electrical connection to ground. These are necessary for charge dissipation and are also very sensitive, and will be damaged if the backside holder is placed face down on a hard surface.
With the cassette wafer grounding clip held slightly back, place the wafer in holder with the topside towards the table
⛔ The wafer will drop slightly when placing it. Fragile substrates may require the use of plastic tipped wafer tweezers to let the wafer down gently.
Ensure the wafer flats align with the guide pins of the holder so rotation is minimal
Release pressure on the wafer clip, and check to ensure good contact is maintained. If not, hold the clip slightly back and adjust the wafer and clip as necessary.
Ensure the wafer is flush with guide pins and flat around edge
Replace backside holder, ensuring that guide pins align with holder holes
Check that backside holder is flush against wafer.
With two fingers, press down towards the edges of the V wing spring and rotate until it is appropriately centered into the cassette overhang.
For 4" and 6" cassettes: Rotate the lock at the top of the holder ~120° clockwise (CW) to secure the backside holder to the cassette.
Place cassette right side up on the table.
Inspect wafer to cassette contact.
Ensure no loose particles are present on the wafer or cassette. Remove these with N2 gun next to table if necessary.
3.2.3 Piece Cassette Mounting
Use correct spacer ring and ensure clips won’t block height laser
Mount sample
Adjust rotation as necessary (<1 degree for aligned writes, ideally <0.2 degrees, use 8100calculator.xlsx and verniers for easy fine adjustment).
Zero stage on LL Zero Landmark (pic, video).
Find chip edges/alignment marks with microscope.
Record location of chip center (unaligned write) or P mark (aligned write).
Click to expand for detailed step-by-step and video links
Video 3.2.3a: UPDATE - Better 0/0 Reference: Lower Left Zero Landmark
Video 3.2.3b: Alignment Microscope Overview
Video 3.2.3c: Piece Cassette Use
Note: This video describes the previous method of mounting, which required the center screw be visible to the alignment microscope. A better, newer improved method uses the Lower Left Zero Landmark as described in the video above. Using the Lower Left Zero Landmark, the center screw does not need to be visible to the microscope. You may still use the old method, but the new method is better in every way.
Use of pieces on the 8100 is more complicated than wafers due to the way pieces are loaded. If you can use wafers rather than pieces, it will likely be much cheaper for you due to the time saved. If you are using pieces, you should not use pieces smaller than 10 mm by 10 mm, or you will likely have a hard time with a number of things. Be aware that using pieces smaller than this is usually a bad idea.
With pieces, we will need to:
Use a spacer ring to offset the holder by an appropriate amount for your chip. You will need to check the spacer ring before each write to ensure the proper one for your chip is present, and will need to change the rings if the one you need is not currently loaded.
Affix your chip to the top of the holder using at least two clips (and screws) on the piece. You'll generally want to mount the pieces in the X direction so the machine's laser height sensor is not blocked. The height laser being blocked by the clip is a common problem for both new and experienced users, and is due to either the clips not being placed in mostly the X direction, or due to the piece being too small, or due to the global or chip marks being too close to the clip.
Adjust rotation of the piece by loosening the rotation adjustment screws, manually changing the rotation, and then tightening the screws once the rotation is set. Since we need rotation accuracy of <1 degree for aligned writes, and want accuracy of <0.2 degrees, we will need to do this with the help of the alignment microscope. Their is a Vernier scale on each holder which will assist with this.
After rotation is set, use the alignment find the center of the piece or the P mark, in either case using the alignment microscope. We'll calculate the offset of the piece from the holder origin, which is roughly the position of the center screw in each holder.
Use appropriate spacer rings for sample thickness.
⛔ A spacer ring is required for pieces! You must check if a ring is already on the holder, and if it’s the proper thickness for your sample. If not, change it.
⛔ Spacer rings are precision machined aluminum, and may be damaged (scratched, bent) very easily! They are critical to achieving good flatness with pieces, and they are expensive (~$1k/ea) to replace, so be very careful with them and store them carefully!
Spacer rings come in two diameters (82mm OD x 65mm ID for the 3” holder, 56mm OD x 46mm ID for the 2” holders) at 3 different thicknesses:
2”/3” = 325 µm
4” = 525 µm
6”/8”/12” = 700 µm
Combinations of these can be stacked up to 1.5 mm
Determine a suitable holder and mounting location.
Check/place the appropriate spacer ring
Place cassette upside down on cleanroom wipe
Remove holder: With two fingers, press down towards the edges of the V wing spring and rotate until it's free of the cassette overhang.
Lift the holder out by holding each side of the wing with two finger (requires two hands to do) and gently lift out, being mindful of the set pin keeping the rotation in place.
Then place the holder top side up on a cleanroom wipe to ensure holder surface is not damaged (this not as critical as with the wafer cassette holder, which ABSOLUTELY must be placed top side up).
Check the thickness of the current spacer ring (if any), and replace with proper thickness ring if needed. Use the provided non-metal tipped tweezers and delicately handle the ring, if needed.
Replace holder, ensuring that guide pins align with holder holes.
⛔ Ensure the clips on the front do not get stick between the holder and cassette.
Check that holder is flush against cassette, then clamp holder into the cassette and flip cassette right side up. on a cleanroom wipe
Place the piece on the holder.
⛔ Be sure the piece clips do not block the height laser, which comes from the upper right of the holder (corner between hook and 3” area) and leaves towards the opposite corner (the corner between the end without the hook and the 2” area) at 56° from the hook, and 18° above the XY plane. If height cannot be measured at your sample, the electron beam will write your pattern out of focus. If height cannot be measured at your marks, the detected mark positions will be wrong, the detected rotation will be wrong, and the electron beam may be out of focus during the write. None of these are good from your write.
For most pieces, place the holder clips in the X axis to avoid blocking the laser
To ensure this for very small pieces (which don’t need alignment), place them diagonally, use the clips to secure them. Then adjust the rotation of the holder to align them with the clips on the sides
Very small pieces that do need alignment should have the clips places upside down on the top left and lower right corners.
Tighten one piece clip, slide chip against it, then tighten the other.
To adjust angular offset, loosen the two screws (inner, centered screw in the cluster of 3) around edges, adjust rotation as needed, then tighten.
Pieces for aligned writes need to be within 1°, and ideally within 0.2°. While the machine compensates for rotation, this is imperfect, and also can't compensate for stitching between fields. A smaller angle will give you a better quality alignment.
Each wafer holder has a vernier scale. This is extremely useful for precision adjustment of the alignment.
Place the holder on the alignment microscope:
First ensure the microscope base plate is centered on the two back pins, and flush against them.
Then gently slide the cassette into the holder place.
Prepare the microscope:
Turn on the LED, select a different source pattern if needed.
Adjust the interpupillary distance adjustment and objective fine focus so you can see well.
Focus on your sample position.
For determining mark/chip positions with high accuracy (repeatability <5 µm, offset < 100 µm)
Adjust rotation:
Loosen the side screws
Find the location of an alignment mark or your chip corner. Zero Xo and Yo.
Move the stage to another mark (or corner).
Note the Xo and Yo values. Calculate the rotation (Google on WolframAlpha the Beamer PC, or your phone) by taking the tangent of Yo/Xo, in degrees. Example for a 70 micron Yo for marks 20000 microns apart in Xo.
Manually adjust the rotation of the ring. Use the vernier scale for high accuracy changes!
Refind the initial mark/corner, zero Xo and Yo, and continue.
Repeat until the proper rotation is achieved.
Tighten the screws
Remeasure the rotation angle to ensure it was not changed by the screws being tightened.
Offset determination, aligned write:
At maximum zoom, move to and focus on the Lower Left Zero Landmark.
Zero the Xo and Yo.
Zoom out, find P mark on chip
At maximum zoom, focus on the P mark
Record the P mark position, and input it into 8100calculator.xlsx to calculate the offset.
Offset determination, unaligned write:
At maximum zoom, move to and focus on the Lower Left Zero Landmark.
Zero the Xo and Yo.
Zoom out, find top left chip corner.
At maximum zoom, focus on the top left chip corner.
Record the top left chip corner X and Y locations (X1, Y1).
Repeat for the lower right chip corner, with X and Y locations as (X2, Y2).
Input the chip corner locations as =AVERAGE(X1, X2) and =AVERAGE(Y1, Y2) into 8100calculator.xlsx to calculate the offset.
Please turn off the microscope light when done. You can leave the profiler on.
3.3 Load Cassette into Machine
Open Exchange Chamber Access door. At load lock, ensure "Cassette on stage" LED is off, then EVAC/VENT.
Open LL, place cassette into pouch, insert pouch into LL, then LOAD/UNLOAD. Close door.
Click to expand for detailed step-by-step and video links
Video 3.3a: Loading - Load Lock View
Video 3.3b: Loading - Jbxwriter View
Open the Exchange Chamber Access door.
⛔ Ensure the CASSETTE ON STAGE LED is OFF, if a cassette is already on the stage, proceed to the 3.9 Unload Cassette from Machine section
If no cassette is on the stage, check the vacuum gauge:
If it reads close to -97.5 in red letters, the load lock is under vacuum. Hold the EVAC/VENT button to vent the LL. This should take ~2 minutes.
⛔ You should only ever press the EVAC/VENT button if the CASSETTE ON STAGE LED is OFF.
If it reads 0.0 in green letters, the load lock is at atmosphere.
This would occur if the previous user forgot to evacuate the LL before leaving. The system will take longer than normal to load the cassette, and you may need to wait some time for the system to reach thermal equilibrium.
After the load lock is vented/at atmosphere and the EVAC/BUTTON button has stopped flashing, turn the lock knob 90 degrees CCW to disengage the lock.
Open the load lock chamber door fully.
⛔ There should not be a cassette in the pouch, but visually check the LL viewport to see if a cassette is already in the loading pouch.
⛔ If there is a cassette present
⛔ Remove the pouch from the LL, holding it with the handle end higher than the hook end. If it is held improperly, the will fall out!
⛔ When sliding the pouch out of the LL, be sure also to not scrape the Teflon holding block with the pouch.
⛔ Remove the cassette from the pouch.
If as expected, no cassette is present, remove the pouch from the LL, ensuring not to scrape the Teflon holding block with the pouch.
⛔ As good practice when handling the pouch, always hold it with the handle end higher than the hook end. If it is held improperly, a cassette (if present) will fall out!
Insert the cassette into the pouch, and insert the pouch into the load lock.
⛔ Ensure at all times that the pouch with cassette is kept level or with the handle end slightly higher than the hook end.
⛔ When sliding the pouch into the LL, be sure also to not scrape the Teflon holding block with the pouch.
Close the load lock chamber door.
Close the lock knob, rotating it 90 degrees CW to lock.
Before proceeding, ensure that the LOAD/UNLOAD button is lit green, and only 2 LEDs (Stage Initial, Exp Ready) are lit green.
⛔ If the Cassette on Stage LED is green, there is already a cassette on the stage, and you should proceed to the 3.9 Unload Cassette from Machine section.
Having ensured the LOAD/UNLOAD button is green, open the cover of the LOAD/UNLOAD button, press and hold until it beeps and begins flashing (about 3-5 seconds). The cassette will be loaded into the chamber (~2.5 min).
Close the cover of the LOAD/UNLOAD button.
⛔ You should only ever press the LOAD/UNLOAD button if it is green.
Close the Exchange Chamber Access door.
Leaving this open will change the temperature of the column, adversely affecting your write and the writes of later users.
In jbxwriter, observe the image in the Manual Loader Viewer tab. When loading is complete:
Both boxes on the column (the IV valves) will be white, indicating they are open.
The stage position will be medium blue, indicating the presence of a wafer.
The box between the column and stage (LL valve) will be white, indicating it is open.
The LL position will be white, indicating there is no cassette in the LL.
3.4 Check Sample Height
In jbxwriter, set cassette in Cassette Type… button in Stage Control tab
Once cassette is loaded, go to sample center or P mark, check sample height with HEI button.
Click to expand for detailed step-by-step and video links
Video 3.4a: Height Detection - Basic
Video 3.4b: Height Detection - Advanced
Tell jbxwriter which cassette you just loaded, and which wafer position you'll be using
In jbxwriter, in the upper right corner, go to the 'Stage Control' tab.
click the Cassette Type… button in Stage Control tab.
Cassette Types for Stage Control Tab
3” Wafer | Multi Wafer / / 3A |
Piece Holder - 3” | Multi Wafer / / 3G |
Piece Holder - 2” A | Multi Wafer / / 2A |
Piece Holder - 2” B | Multi Wafer / / 2B |
4” Wafer | Wafer / 4.0 / |
6” Wafer | Wafer / 6.0 / |
Determine a location to measure height: If doing an aligned write, do this at the P Mark. For an unaligned write, do this at this chip center.
If you measured the location of chip edges or your mark locations, you can calculate the center of the chip using 8100calculator.xlsx.
In the Stage Control tab, select the Substrate coordinate system
The 8100 has two sets of coordinates: Stage coordinates, and Substrate coordinates (also referred to as material coordinates).
Stage coordinates are useful in determining if the cassette is at the origin (ORG) position (~100/100). It's also used extensively by BNC staff and JEOL for calibrations. This coordinate system has the origin at the upper left of the stage, with X towards the LL, and Y towards the Beamer PC (down, on the stage diagram). This makes it somewhat confusing to use.
Substrate coordinates have their origin at the center of each wafer holder. The holder is set with the Cassette Type... button in the stage control tab, as we did before. Having the wrong set cassette or window will give you the "wrong" origin. Substrate coordinates have their X towards the LL, and their Y towards the back of the machine (up, on the stage diagram), and they're most useful when working with samples and doing exposures. Generally, you'll use substrate coordinates at all times, unless ensuring that the stage is at the origin (ORG) position (~100/100 in Stage Coordinates), as there isn't a sensible way to talk about the origin position in substrate coordinates.
Set the Target Position radio button to Absolute.
Relative can be useful when moving between marks, chips, etc., if you know their offset, but we'll mostly use Absolute mode (except when finding the P Mark Offset, which uses Relative).
Type in your height measurement location into the X and Y Target Position (um) boxes.
Almost all coordinates and distances in jbxwriter are in µm. The exceptions are the SEM and Spot tabs in the Image Control area, which are both in nm.
You cannot paste into the X and Y Target Position boxes. This is very annoying. However, you can copy from them to paste elsewhere in the software.
As you enter a coordinate, you'll see the blue coordinate indicator snap to your input position. You can also right click in the Stage Position Window to set a target position with the track ball.
Click the MOVE button to move the stage to the input target position.
The red coordinate indicator will snap towards your Target Position, and the X and Y Current Position boxes will be very close (within 5 µm) to your Target Position. The machine uses laser based correction (LBC) of the stage position to make up the difference. This modifies the beam position electrostatically rather than moving the stage mechanically.
You can also use the Stage Arrows to the left of the MOVE button to move by the distance in the um/step window. Note that these move the STAGE in the given direction, which is counter intuitive. If we're at substrate position of 0/0 with a 1000 um/step setting, the following clicks will give the following moves:
→ moves to -1000/0
← moves to 1000/0
↑ moves to 0/-1000
↓ moves to 0/ 1000
Positions and um/step can be set to large (10s of mm or 10000s of µm) or small (1 nm or 0.001 µm) values. The stage position is accurate/controlled to within 0.6 nm with the help of LBC, even though the mechanical accuracy of the stage is only /- 5 µm.
To start measuring the height, hit the HEI button. The height (in µm) will be displayed to the right of the button. Click the HEI button again to stop measuring the height.
The height is displayed relative to the zero position (0 µm) of the stage.
The stage cannot be moved while the height is being measured. If you're interested is measuring the height across your sample, go to the section on performing a Height Map.
The height should absolutely be within +/- 100 µm, and ideally within +/- 50 µm. Closer to 0 µm will give better write quality, with anything being within +/- 10 µm being great.
You can open the small push door (below the monitor, above the keyboard and EOS operation panel (which you should never touch), and left of the Emergency Off (which you should also never touch, unless there's an emergency)), and see a 4x1 VGA Switch. It should almost always be on INPUT 2. If you're having problems measuring the height, you can change it to INPUT 1, which will display the height measurement raw output on the left monitor. The white bar is the laser, the machine determines the center of this and assigns that as the height (indicated with the green arrow/cross). If the white bar doesn't span the window or looks blocked, it may give you an erroneous height reading, or indicate that you're by a piece clip or the edge of a wafer holder. Be sure to go back to INPUT 2 after you do this, and close up the small push door.
If the height of the chip isn't within +/- 100 µm, or the height can't be measured at one of your marks, remove the cassette (proceed to the 3.9 Unload Cassette from Machine section) and fix the issue.
Optional: If the height can't be measured, it's possible the location you measured in the alignment microscope is off. You could pull the cassette back out to check, but it may be good to first try to find a landmark with the SEM to avoid needing to do that,.
You can use the SEM to try to find a chip edge, an alignment mark, a piece clip, wafer holder edge, etc., to get your bearings.
Under the Image Control section of jbxwriter, click the SEM tab (not the SEM button, which turns on the SEM).
Using the SEM will heavily expose resist that is imaged. For most substrates, you should not image anything within 100 µm of your pattern to be exposed since backscattered electrons from the SEM will contribute to exposing the resist. Directly exposing part of your sample to be patterned will drastically affect your developed pattern. This is also true for the automated mark detection algorithm, to a lesser degree.
Under the SEM tab, you'll set appropriate values for Rapid, Grid, and Scan Width [nm]
The Rapid check box will scale the Scan Width by a factor of 50% in both X and Y, scanning just the center 25%. This is useful to have checked when moving around with the SEM on, because it shows images at 4x the speed.
The Grid check box will display a 10 by 10 green dotted grid over the image (division size = Scan Width / 10). The center divisions are colored red. This is useful when precisely aligning to a mark/feature, but should generally be left off when hunting for features because it may obscure them.
The Scan Width [nm] is the X and Y width in nms that will be imaged by the SEM. This is set in 1000 nm intervals, and it's useful to click the blue slider and then use the scroll dial on the track ball to quickly increase or decrease this. Be aware that Rapid will decrease this by a factor of 50% in both the X and Y, showing only the center 25%. Also be aware that the Grid interval = (Scan Width)/10, so picking a nice round scan width can line up the Grid with your mark edges. For example, with Grid and Rapid checked, a centered 10 µm wide mark will hit the green grid lines at a Scan Width of 100000 nm, or 50000 nm if Rapid is off.
For the other settings in the SEM, use the following settings (and contact BNC Staff if you have questions about changing these, though you won't hurt anything by changing the settings in the SEM tab):
Signal: BE Radio Button
Integration: 1
Scan clock [nsec]: 5000
Main Def. Offset [nm] X & Y: 0
Sub Def. Offset [nm] X & Y: 0
Doublecheck that Rapid and Scan Width [nm] are appropriately set (these control the area of the SEM that will expose your sample), and hit the SEM Button to turn on the SEM.
The SEM image will be displayed on the left monitor. The image can be saved with the Capture... button in the jbxwriter Image Control section.
The SEM will be on until the SEM Button is pressed again. After turning it off, the last scan will persist on the left monitor, allowing you in inspect it carefully without the SEM actively scanning, if you desire.
If you need to change the above SEM settings, click the Apply button.
The SEM settings (aside from BE Brightness, Contrast, and Contrast Multiplier) will not be applied until the Apply button is pressed.
Adjust the BE Brightness, Contrast, and Contrast Multiplier based on reasonable values for your sample.
Good values for these depend heavily on the current, as well as on your substrate and any features on it.
The set values for these in jbxwriter will generally change each time you to come to the tool because they will be what the last user used.
Once you find good values for a particular current, substrate, and substrate features, you can use those values in the future if you remember them/take a picture of them/write them down.
Some good starting values are:
Current - Brightness - Contrast - Contrast Multiplier
2 nA 1225 1635 x56
10 nA 1975 3590 x1
30 nA 1865 2785 x1
100 nA 1332 1805 x1
Adjust the Scan Width, BE Brightness/Contrast/Multiplier, and move positions until you're able to find a landmark that help guide you to your feature of interest.
Once it is verified that chip height at the necessary positions can be measured, proceed to the next section.
3.5 Material Corrections
3.5.1 Unaligned Exposure Material Corrections
In the Calibration Page, Material Corr. tab, click Load button, load 0_currentnA.mcorrprm. In the Beam Drift sub-tab, click Set BE Mark, then Execute, then LR Update, then LR Save.
Click to expand for detailed step-by-step and video links
Video 3.5.1: Unaligned Material Corrections
In jbxwriter, go to the Calibration page → Material Corr. tab.
In the Correction File section, next to File Name, click the Load button.
Select and open the 0_currentnA.mcorrprm file that corresponds with your current/condition file.
You can click the Name bar above the listed files to list in A/Z or Z/A. List in A/Z, and these files will be at the top.
In the checkboxes under file name, ensure only Beam Drift Correction is checked.
This is important, because the Execute button at the lower right of the Calibration page will execute all checked subprograms.
In the Beam Drift sub-tab.
Ensure the BE Mark radio button is selected.
Press the Set BE Mark button
In the lower right hand corner, press the Execute button.
In the log, the Mark Scan X, Y reported by DRIFT is the location the mark was found, in stage coordinates. The Position Shift X, Y = Newly Measured Position - Previous Mark Position.
Use the buttons in the lower right to perform an Update/Save:
Press the lower right Update button.
Press the lower right Save button.
This is necessary to update the saved position of the BE mark, which is used by the batch file and the exposure file.
The lower right Save button saves the displayed information to the subprogram file in the condition file. This will need to be pressed for each setting tab in the Material Corr. tab in order to use those settings during an exposure. After an exposure completes, whether successfully or not, a Restore is performed, which reloads all of the saved subprograms files back into jbxwriter. This will overwrite your previous values if you have not used the lower right Save button. For this reason, you should generally not use the upper middle Save button to save the mcorrprm file after an exposure has been attempted, as some important settings may have been changed, and the change may not be obvious. Do not save the 0_currentnA.mcorrprm files, these are updated by BNC staff. However, you may load them and then use the upper middle Save As... to save them under a new file name for your modification and later use. Please use the following convention: alias_YYYYMMDD_currentnA_windowID_markpattern.mpcorrprm, where windowID is 2A, 2B, or 3G for the piece holder windows, or 3A, 4W, or 6W for the wafer windows.
The Update and Save buttons in the lower left hand corner no longer need to be pressed, but doing a lower left hand Update and then lower left hand Save will not hurt anything.
Proceed to Run DAILYCAL.
3.5.2 Aligned Exposure Material Corrections
3.5.2.1 Find the P Mark
Find the P Mark with the SEM
Register the Substrate position in UsrMayRegist: P Mark.
Click to expand for detailed step-by-step and video links
Overview
Mark detection can be complicated, and the software allows you to do many things that our users do not typically need to do. If you have questions about more advanced mark detection, contact jcwirth@purdue.edu. The following sections assume you have 4 global marks, all of which have the same width and length (though P may be longer with no changes to the procedure necessary) and either have 4 physical chip marks, or are using the global marks as chip marks, or are using virtual marks.
See the Run Card Supplements for suggestions for robust example marks and mark parameters. There may be good reasons to use other values, these only serve as a robust starting point.
Video 3.5.2.1: Find P Mark
In the Stage Control tab, select the Substrate coordinate system
Set the Target Position radio button to Absolute.
Ensure the Cassette Type… button is set with your proper Type/Size/Window.
Type the following location into the into the X and Y Target Position (um) boxes:
If using a piece window: measured P Mark location from the microscope.
If using a wafer window: design P mark location (the offset with the wafer holders is small, on the order of <100 μm).
Click the MOVE button.
Under the Image Control section of jbxwriter, click the SEM tab (not the SEM button, which turns on the SEM).
Under the SEM tab, set appropriate values for Rapid, Grid, and Scan Width [nm].
You could start by leaving Rapid unchecked, Grid checked, and a Scan Width of 200,000 nm (you may later need to increase the scan width).
For other values: Signal = BE Radio Button, Integration = 1, Scan clock [nsec]= 5000, Main Def. Offset [nm] X & Y = 0, Sub Def. Offset [nm] X & Y: 0
Press the SEM Button to turn on the SEM.
⛔ Remember, we're exposing any resist we see in the SEM, and backscattering can expose resist as far as ~100 um from the beam. So be aware of when you turn the SEM on, and be sure to turn it off before moving across your write area.
Adjust the BE Brightness, Contrast, and Contrast Multiplier based on reasonable values for your sample, refering to the 4.2 Example SEM Values if needed.
If you're confident you're using appropriate BE Brightness, Contrast, and Contrast Multiplier values, but cannot see the P mark, it's likely outside your scan area.
Increase the Scan Width [nm] a reasonable amount (e.g. 100,000 nm), then click the Apply button.
It's useful to click the blue slider and then use the scroll dial on the track ball to quickly increase or decrease the Scan Width.
Remember that the Rapid check box will scale the Scan Width by a factor of 50% in both X and Y, scanning just the center 25%.
Continue to increase the Scan Width [nm] (and turn off Rapid, if necessary) until you've found part of the P mark.
If you're unable to find the P mark even at the maximum Scan Width and with Rapid off, it's likely your BE Brightness, Contrast, and Contrast Multiplier values are off.
If you adjust these and still cannot find your mark, your position from the alignment microscope was very far off. Try to move to one of the chip corners (or off the chip, to one of the clips, etc.), find a landmark that is useful for reference, and navigate to your P mark from there.
If you still can't find it this way, you will need to unload the cassette to visually check what happened.
Once you've located the P mark, find the center and center your cross in the SEM grid using the Stage Arrows to the left of the MOVE button to move by the distance in the um/step window.
After it is centered, turn off the SEM by pressing the SEM button.
Locate the User Def. Position drop down, and select UsrMayRegist: P Mark,
Under UsrMayRegist: P Mark, locate the Regist. button
⛔ NEVER click the Regist. button under Fixed Position, only ever click the Regist. button under User Def. Position!
Click the Regist. button under User Def. Position
It will ask for confirmation, click OK.
Again, in the User Def. Position drop down, select UsrMayRegist: P Mark
This is done because, although we have save the P Mark location into UsrMayRegist: P Mark, our target position is still the old value of UsrMayRegist: P Mark. Reselecting UsrMayRegist: P Mark will make our P Mark location both the target and current position.
3.5.2.2 Load your MCORRPRM file
Load previously created MCORRPRM file, or create a new one. If creating, load 0_currentnA.mcorrprm, save as alias_YYYYMMDD_currentnA_windowID_markpattern.mpcorrprm
Click to expand for detailed step-by-step and video links
Video 3.5.2.2: Load MCORRPRM File
In jbxwriter, go to the Calibration page → Material Corr. tab.
In the Correction File section, next to File Name, click the Load button.
If you have previously saved a mcorrprm file, select and open it.
If you do not have a previously saved mcorrprm file, or are using different enough mark parameters that you would like to create a new one:
Load the 0_currentnA.mcorrprm file that corresponds with your current/condition file.
⛔ Do not save over/overwrite the 0_currentnA.mcorrprm file
Click the Save As... button.
Please save a new file using the following convention: alias_YYYYMMDD_currentnA_windowID_markpattern.mpcorrprm, where windowID is 2A, 2B, or 3G for the piece holder windows, or 3A, 4W, or 6W for the wafer windows. The markpattern portion could be an identifier of your mark composition, thickness/depth, type, size, location, substrate, etc. This just needs to make sense to each user so they can distinguish which mpcorrprm file is useful as a starting point for future writes.
⛔ !!!!! Very important note !!!!!: Loading a mcorrprm file only loads the values into the jbxwriter software. It is not until you click the lower right Save button on each sub tab of the Material Corr. window that each sub program's values will be saved into the machine settings, which are used by the exposure. After an exposure, the machine will execute a Restore (without a Demag) which will load the saved values back into jbxwriter, which will be different from the values you had unless you've saved each subprogram. For this reason, you should generally be careful about when you save your mcorrprm file, and should not save it after an exposure has been attempted. Additionally, when you open the RG Detection Condition or BE Detection Condition button in the sub tabs, you should press the OK button to close it. Pressing the Cancel button after initially loading the mcorrprm file and opening the RG Detection Condition window will change the values in the RG Detection Condition window back to what they were before your load. For more information review the "Notes on the Material Correction Tab" section of the .
3.5.2.3 Determine the P Mark Offset and Update Global Mark Detection Offset
Move to P Mark Design Position, select UsrMayRegist: P Mark, select Relative, type Target X/Y into Offset in the Global Mark Detection sub-tab, press LR Save, select back to Absolute.
Click to expand for detailed step-by-step and video links
Video 3.5.2.3: Determine P Mark Offset
The Offset is the relative offset vector from our Design P Mark Position to our Actual P Mark Position. To find this, we'll move to our Design P Mark Position (in substrate coordinates). Then select our saved Actual P Mark Position (previously saved in UsrMayRegist: P Mark), display the relative offset between them, and save that to our Global Mark Detection sub-tab in the Offset X/Y boxes.
The Offset will later need to be added to your exposure data, either by manually changing the SDF's OFFSET (0, 0) line (and saving) before converting it into a magazine (MGN) file, or in Beamer → VisualJob → Layer_#→ Substrate Origin Offset X/Y. If you do not do either of these later, during the exposure the machine will look for your P Mark at substrate coordinate (0/0), and it's probably not going to be there.
Type in your P mark design position into the Target Position X and Y values (ensuring Absolute is checked and the proper Type/Size/Window are set in the Cassette Type button).
You cannot paste into the Target Position boxes.
Click the Move button.
In the User Def. Position drop down, select UsrMayRegist: P Mark.
Under Target Position (um), select Relative.
In the Material Corr. tab, check the Global Mark Detection box.
Type the displayed Target Position X and Y values in the Offset X and Y values in the Global Mark Detection subtab
You can also use the mouse to select a displayed Target Position value, right click, Copy, then go to the Offset box, right click, Paste.
In the Global Mark Detection sub-tab, press the lower right (LR) Save button.
Under Target Position (um), select Absolute.
Go to the Global Mark Detection sub-tab
Check only Global Mark Detection.
Set Global Mark Parameters: Material Type/Size/Window; Check Height Measurement; Enter P/Q/R/S design positions; Semiauto; 1 degree; press LR Save.
Setup for AGC: Check AGC, uncheck Q, R, S.
Set P Rough RG Detect Conditions: Scan Position X/Y = Scan Width X/Y = 3 * Wg, rest from 4.7 Example RG Detect Condition. Applies to another subprogram..., Mark Type All select, Subprogram that can be set, close, OK. Execute.
Once AGC completes: All select, OK.
Set P Fine RG Detect Conditions: Scan Position X/Y = 50% Lg, Scan Width X/Y = 3 * Wg, Applies to another subprogram..., Mark Type All select, deselect P rough, deselect Q rough, Subprogram that can be set, close, OK.
Set Q Rough RG Detect Conditions: Scan Position X/Y = 3 * Wg, Scan Width X/Y = 25 * Wg, OK.
Setup to run Global Mark Detection: Uncheck AGC, recheck Q, R, S. Press LR Save, Save MCORRPRM.
Click to expand for detailed step-by-step and video links
Video 3.5.2.4: AGC and Setup MCORRPRM
⛔ If you're working from a saved mcorrprm file that is already properly setup for your sample, you may skip to the next step.
In jbxwriter, go to the Calibration page → Material Corr. tab. → Global Mark subtab.
Click the subprogram check boxes so that only the Global Mark Detection check box is checked.
Setup basic parameters in the Global Mark subtab:
Appropriately set the Material Type, Material Size, and Material Window with the help of the supplement 4.1 Cassette Window Identifiers.
Ensure Height Detection is checked.
If this is unchecked, the machine will not correct for the measured mark shift and rotation that come with height differences. That will not be good for your alignment.
Ensure Detection Mode is Semiauto.
Auto mode will fail mark detection if a mark cannot be found. Semiauto is like auto, but will give you the option of using the SEM to find a mark if it cannot be found during the rough scan. Manual will just let you tell the machine where the mark center is, but shouldn't be used because it will not correct for rotation. Discuss with BNC staff if you have any questions or think you need to use the other modes.
Ensure allowable rotation angle (degree) is set to 1.0000
You can set this arbitrarily high but you will not get good exposures if your marks are more than 1 degree off. This just sets whether the machine will stop executing mark detection or not. We set it to 2 so that you can verify the rotation in the log and decide what to do with this information.
Enter the design positions of your marks (i.e. what your layout expects, not the actual position on the fabricated wafer) into the P, Q, R, and S boxes.
See the supplement for more information on which mark should be called what.
Setup the Global Mark subtab to run AGC:
Running AGC in the Global Mark Detection subtab will use the P rough scan at the Offset X/Y to scan for your P Mark. Then it will execute automatic gain correction using the P Fine parameters. It will then execute a P Fine scan on the mark, and proceed with detecting the other marks that are checked. We will uncheck the Q, R, and S marks to save time in case the AGC doesn't go well the first time.
Check the AGC box.
Uncheck the Q, R, and S boxes.
Click the P Rough Scan RG Detect Condition... button.
Set the values to the following (Copied from except for the Scan Position and Scan Width). Wg being your global mark width.
Scan: PDEF, X→ Y, 50%*Lg (For X & Y), 3*Wg (For X & Y), 1, 3, 19000
Offsets: All 0
Scan Type: Fine, Box, Cross, Wg, Lg
Gain: Don't check available, Add signal (first derivative), SEM Multiplier, SEM Contrast, SEM Brightness
Correlate: Check Available, 100, 100, 100, End of Data Addition, 0, 50
Allowance: Check Available, 2.5, 2.5, 2.5, 2.5, 50, 50, 50, 50
Raster: 0.000, 1
Retry: 1, 0.5, 0.5
Click the Applies to another subprogram button.
Under the Set Item section, leave all of the items selected.
Under the Mark Type: Registration Mark (RG) section, click the All select button.
Click the Subprogram that can be set button. (This is an awkward way of saying "Apply to the selected").
Then click the Close button.
Click the OK button.
Start AGC by clicking the Execute button
Watch the left monitor as AGC executes.
The WaveMax and WaveMin values are the top and bottom DAC values of the waveform. Sub is the difference between these.
The level is a measure of the signal to noise ratio (JCW: I think?). A minimum acceptable value is set deep in the machine for this, and your mark may fail AGC if this value is not high enough (not enough contrast).
The CGAIN is the BE Contrast Multiplier, the MGAIN is the BE Contract, and OFFSET is the BE Brightness. Contrast is a multiplicative adjustment (i.e. a gain), and brightness is an additive adjustment (i.e. an offset).
You should be able to eventually see a step in the scan that represents your mark. If this does not become apparent, AGC is not finding your mark. It may suggest garbage values, so ensure the waveform is reasonable.
If AGC fails or cannot find your waveform:
Ensure your AGC location is actually in the location where your mark is located.
If you're sure it's executing on your mark, recheck that the Scan Width and Scan Position values in RG Detection Condition are appropriate for your mark.
If you're sure it's executing on your mark, and are confident in your Scan Width and Scan Position values are appropriate, AGC may not work for your mark.
In this case, use the SEM to adjust the BE Contrast Multiplier, BE Contrast, and BE Brightness to maximize viewing of the mark.
Then use these values to manually set the Gain parameters for the P Rough subprogram, and apply to all other subprograms:
Click the P Rough Scan RG Detect Condition button.
Go to Gain tab, to: Don't check available, Add signal (first derivative), SEM Multiplier, SEM Contrast, SEM Brightness
Click the Applies to another subprogram button.
Under the Set Item section, click only Gain. You could leave all selected by if you've changed something it else inadvertently it could cause problems, so best to just apply what we need.
Under the Mark Type: Registration Mark (RG) section, click the All select button.
Click the Subprogram that can be set button.
Then click the Close button.
Click the OK button.
Click OK, skip the step below (since AGC was not successful), and continue at the step below that.
If AGC is successful, the Updating inquiry of preamp constant popup window will appear:
The values under preamp constant are the AGC values: Coarse Gain = BE Contrast Multiplier, Middle Gain = BE Contrast, Offset = BE Brightness.
You can later use these values in the SEM to see what the machine is seeing, if you want.
Click the All select button.
This will select all subprograms within the Material Corr. tab to apply the AGC values to.
AGCRG = AGC
CHIPAL = Chip Mark Detection
DRIFT (rough) = Not actually stored even if it's selected. This is the BE Detection Condition... of the Beam Drift Window. However, we don't want to use the AGC of the marks for this, we want to use the AGC of the BE mark, so the software knows to ignore your selection of this. That's why we can use the All select button without causing any issues.
DRIFT (fine) = BE Mark Detection, RG Detection Condition
MDRG = User Defined Mark Detection
SETWFR (P rough) = Global Mark Detection, P Rough Scan
SETWFR (P fine) = Global Mark Detection, P Fine Scan
SETWFR (Q rough) = Global Mark Detection, Q Rough Scan. Despite the name, this actually applies to detection of the Q, R, and S marks.
SETWFR (Q fine) = Global Mark Detection, Q Fine Scan. Despite the name, this actually applies to detection of the Q, R, and S marks.
Click OK.
This will apply the AGC values to all of the selected subprograms in their RG Detection Condition... windows, under the Gain tab.
Setup the Global Mark Detection tab for Global Mark Detection:
Uncheck the AGC box.
Check the Q, R, and S boxes.
Click the P Fine Scan RG Detect Condition... button.
Set the Scan Position X/Y to 50%*Lg.
Ensure Scan Width X/Y is set to 3*Wg.
Click the Applies to another subprogram button.
Under the Set Item section, leave all of the items selected.
Under the Mark Type: Registration Mark (RG) section, click the All select button.
⛔ Importantly, the next step is different from what we did previously!
Under the Mark Type: Registration Mark (RG) section, deselect P rough and deselect Q rough.
Click the Subprogram that can be set button.
Then click the Close button
Click the OK button.
Click the Q Rough Scan RG Detect Condition... button.
Ensure the Scan Position X/Y is set to 3*Wg.
Set the Scan Width X/Y to 25*Wg. In words, Twenty Five times Wg.
Click the OK button.
Click the lower right Save button.
(Optional): If using physical chip marks, setup parameters in the Chip Mark subtab:
Ensure AGC is unchecked
Ensure the Mode radio button is set to 4.
Ensure the Height measurement for Mode 1 and Mode 4 box is checked.
Enter the design center of the chip you wish to scan in the Center X and Y boxes.
Since we've already done global mark detection and set the Offset, the machine can automatically go to the design center of the chip.
Enter the design positions of your marks (i.e. what your layout expects, not the actual position on the fabricated wafer) into the M1, M2, M3, and M4 boxes.
Click the RG Detection Condition... button:
In the Scan Tab, set the Scan Position X/Y to 40%*Lc and the Scan Width X/Y to 3*Wc.
In the Scan Type tab, set the Mark Width to Wc and Mark Length to Lc.
In the Retry tab, set Maximum number of retries (times) to be 10 and set the Amount of shift at a retry (um) to be 0.5 for X and Y
Click OK.
Click the lower right Save button.
In the Correction File section, Click the Save button to save your file.
Go to the Global Mark subtab, ensure only the the Global Mark Detection check box is checked.
Click the Execute button.
If a mark isn't found: Continue, use SEM and move stage to unfound mark centers, OK.
If necessary, adjust scan position/scan widths to find all marks automatically. When successful, press LR Save, Save MCORRPRM.
Click to expand for detailed step-by-step and video links
Video 3.5.2.5a: Global Mark Detection
Video 3.5.2.5b: Potential Issues in Global Mark Detection
In jbxwriter, go to the Calibration page → Material Corr. tab. → Global Mark subtab.
Click the subprogram check boxes so that only the Global Mark Detection check box is checked.
Perform a scan:
Click the Execute button.
Observe the left screen as detection proceeds to ensure the machine finds your marks, and nothing odd is going on.
We're using the Semiauto mode, so if a mark can't be found, the machine will display the Confirm popup window:
The text says "While operating the cancel/continue buttons That appear in the Global Mark Tab to search the mark."
This means "Click Okay to have the SEM display where the machine thinks the mark is. Use the SEM and the arrow buttons to find the center, and then click OK so the machine can rescan at the location you specify".
Click OK.
Adjust the SEM parameters and move the stage so the mark is found and centered in the SEM image.
When you're satisfied with the position, click the Continue button under Mark detection execution check.
If detection fails, continue to fail to find marks despite being on top of them in the SEM, or it seems something odd is going on, it's likely that the Scan Width and Scan Positions need to be adjusted. Adjust these as appropriate until mark detection can proceed without needing to open up the SEM and manually find any of the marks. If you can't find your marks, email BNC Staff for support (which may need to be at a different time).
Inspect the log:
Once detection proceeds successfully, click the Log button and go to the Calibration tab.
Scroll up slightly to the [P-point mark measurement result] section.
The machine notes its measured Offset here (this is the machine's measured offset of the offset that goes in the Offset X and Y values of the Global Mark Detection subtab). This should be very close (< 1 µm) to your Offset. If it is more than a few microns different from the Offset you currently have, adjust your offset to match the machine's measured offset. This only needs to be accurate enough to find the P mark, so small differences and anything < 1 um don't matter.
Inspect the mark rotation angle of each mark. These should be reasonably close, and should ideally all be <0.2°, with <0.5° being "good", <1° being "okay" and <2° being "not good". Better angular alignment will mean better alignment accuracy and smaller field stitching issues.
Inspect the Height measurement success/fail and height of each mark. Height measurement should succeed at each mark, and the heights across your chip should be within +/- 10 um. Ideally, all marks should be within +/-50 um, but definitely need to be within +/- 100 um.
Note the Height (um) under the [P-point correction data] section. This is the machine's calculated average height for your chip. Ensure that one of the measured mark heights is not throwing off the average substantially.
When satisfied, click the Close button.
Press the lower right Save button and then deselect the Global Mark Detection check box.
If needed, in the Correction File section, Click the Save button to save your MCORRPRM file.
If using a new MCORRPRM, first set parameters:
AGC unchecked, 4, Check Height Detection, Set design M1/M2/M3/M4 position, enter a design chip center X/Y.
RG Detect Conditions: Scan Tab: Scan Position X/Y = 40%*Lc, Scan Width X/Y = 3*Wc. Scan Type tab: Mark Width = Wc, Mark Length = Lc. Retry tab: 10, 0.5µm, 0.5µm. OK.
Execute. Adjust scan position/width if failing. When successful, Press LR Save, Save MCORRPRM.
Click to expand for detailed step-by-step and video links
Video 3.5.2.6: Chip Mark Detection
⛔ This section may be omitted if you are using global marks, but opting to use virtual marks for height detection.
⛔ Even if you do not have physical chip marks distinct from your global marks, you can actually use the global marks as chip marks. Simply go to the Chip Mark Detection subtab, make sure AGC is unchecked, choose Mode 4, ensure Height measurement for modes 1 and 4 is checked, put the center X and Y as 0 / 0, and assign your global marks as M1, M2, M3, and M4, such that M1 is -X / +Y, M2 is +X / +Y, M3 is +X / -Y, and M4 is -X / -Y.
In jbxwriter, go to the Calibration page → Material Corr. tab. → Chip Mark subtab.
Click the subprogram check boxes so that only the Chip Mark Detection check box is checked.
⛔ If you're working from a saved mcorrprm file that is already properly setup for your sample, skip to the "Execute" step below.
Setup parameters in the Chip Mark subtab:
Ensure AGC is unchecked
Ensure the Mode radio button is set to 4. Assuming you're using 4 chip marks. You could instead use only 1.
Ensure the Height measurement for Mode 1 and Mode 4 box is checked.
Enter the design center of the chip you wish to scan in the Center X and Y boxes.
Since we've already done global mark detection and set the Offset, the machine can automatically go to the design center of the chip.
Enter the design positions of your marks (i.e. what your layout expects, not the actual position on the fabricated wafer) into the M1, M2, M3, and M4 boxes.
Click the RG Detection Condition... button:
In the Scan Tab, set the Scan Position X/Y to 40%*Lc and the Scan Width X/Y to 3*Wc.
In the Scan Type tab, set the Mark Width to Wc and Mark Length to Lc.
In the Retry tab, set Maximum number of retries (times) to be 10 and set the Amount of shift at a retry (um) to be 0.5 for X and Y
Click OK.
Click the lower right Save button.
Click the Execute button.
Observe the left screen as detection proceeds to ensure the machine finds your marks, and nothing odd is going on.
Chip marks only execute in 'Auto' mode, meaning if a mark can't be found, it is not used for alignment.
Once execution completes, press the Log button.
In chip mark detection, the given offsets and mark rotation are relative to where they were predicted to be with your chip mark detection. Heights are absolute.
Ensure the height is measured for each mark, and the offsets and mark rotations are reasonable.
When satisfied, click the Close button.
Click the lower right Save button and then deselect the Global Mark Detection check box.
If needed, in the Correction File section, Click the Save button to save your MCORRPRM file.
3.5.2.7 Save the Beam Drift parameters
Update Beam Drift RG Detection Conditions
If using Global Marks but virtual chip marks. Global Mark sub-tab, P Fine RG Detect Condition, Applies to another subprogram..., select Drift (fine), Subprogram that can be set, close, OK.
OR
If using Global Marks AND physical Chip Marks. Chip Mark sub-tab, RG Detect Condition, Applies to another subprogram..., select Drift (fine), Subprogram that can be set, close, OK.
Click to expand for detailed step-by-step and video links
Video 3.5.2.7: Aligned Beam Drift
⛔ The Beam Drift settings need to be set every time you do an aligned write, because the MCORRPRM file does not save the RG Detect Conditions for the Beam Drift tab (or the AGC or User Defined Mark, but we're not using those here).
When executing drift correction during the exposure (which you should do via DIRE20 if your write will take 20 minutes or longer), the machine will update drift on the best mark it has.
For an unaligned write that's the BE mark, and will use the preset BE Detect Condition settings in the loaded 0_currentnA.mcorrprm file.
For an aligned write with global marks but no physical chip marks, it will use the P mark.
For an aligned with with both global marks and physical chip marks, it will use the M1 mark on the particular chip.
So we need to think what we're doing, and apply the RG Detect Conditions settings of either the P Mark or the Chip Marks to the Beam Drift (called the DRIFT (fine) program).
In jbxwriter, go to the Calibration page → Material Corr. tab.
Decide which set of mark detection parameters to use, and execute only one of the following:
If you are using Global Marks but do not have physical chip marks, use the P fine settings:
Ensure the Global Mark Detection check box is checked.
In the Global Mark subtab, click the P Fine Scan RG Detect Condition... button.
Click the Applies to another subprogram button.
⛔ Importantly, the next step is different from previous sections!
Under the Mark Type: Registration Mark (RG) section, select only DRIFT (fine).
Click the Subprogram that can be set button.
Then click the Close button
Click the OK button.
If you are using physical Chip Marks as well as Global Marks, use the Chip Mark settings:
Ensure the Chip Mark Detection check box is checked.
In the Chip Mark subtab, click the RG Detect Condition... button.
Click the Applies to another subprogram button.
⛔ Importantly, the next step is different from previous sections!
Under the Mark Type: Registration Mark (RG) section, select only DRIFT (fine).
Click the Subprogram that can be set button.
Then click the Close button
Click the OK button.
If you prefer, you can also choose settings directly in the Beam Drift subtab.
In jbxwriter, go to the Calibration page → Material Corr. tab. → Beam Drift subtab.
Click the subprogram check boxes so that only the Beam Drift Correction check box is checked.
Click the lower right Save button.
You may proceed.
3.6 Run DAILYCAL
15 to 40 minutes after changing currents, go to Batch tab.
Ensure DAILYCAL is loaded, with list on right of: Beam Current Measurement, AE Mark Detection, AE Mark Detection, Static Focus Correction, BE Mark Detection, BE Mark Detection, Main DEF Correction, Sub DEF Correction, Distor Correction, Save.
Execute DAILYCAL batch calibration file.
Click to expand for detailed step-by-step and video links
Video 3.6: Run DAILYCAL
In jbxwriter, go to the Calibration page > Batch tab.
The DAILYCAL batch calibrations need to be done for each current every day. They should be run after ~20 to 40 minutes of settling time after changing the current.
If you know that another user you trust has:
used the same write current within the last 24 hours
waited for the appropriate settling time before running DAILYCAL
and had completely performed DAILYCAL without issues, you can omit running them. However, we recommend that all users run these each time they use the machine for optimum write quality, and it could be a costly mistake to omit them.
Ensure that Batch Name shows DAILYCAL.
If the Batch Name is not DAILYCAL, click the Open button, select the DAILYCAL file, and click Open.
Ensure the list on the left is unchanged: If the list does not correspond with that below, click the Open button, select the DAILYCAL file, and click Open.
Beam Current Measurement
AE Mark Detection
AE Mark Detection
Static Focus Correction
BE Mark Detection
BE Mark Detection
Main DEF Correction
Sub DEF Correction
Distor Correction
Save
Click Execute in the lower right corner to run.
This will take ~ 8 minutes.
You will see the execution of the batch calibrations on the left screen.
During the calibration, you may click the Log button in the lower right, and go to the Log window Calibration tab. This will display the results of the calibrations as they run.
The calibrations that execute are:
Beam Current Measurement (CURRNT): The tool measures the beam current (averaged over 5 measurements) to determine the current for dose calculations. This is done to ensure the current can be measured successfully before exposure.
AE Mark Detection (INITAE): The tool scans the beam over a gold knife edge, and measures the current during this sweep. This is used to determine the position of the AE Mark for focus correction. This is run twice in order to allow the user to check the difference between the two scans in the log file, if they so desire. A difference of more than ~5 nm between scans indicates the beam may not be properly stable. The log also displays the measured size of the beam, which depends on both the quality of the AE mark (the scan position needs to be periodically moved by BNC staff) and the quality of the beam. This measured size is always larger than the true beam size due to electron scattering.
Static Focus Correction (SFOCUS): The tool changes the focus lens (Lens 4) setting under and over focus to determine the optimal focus value. On the left monitor, lateral movement of the peak during this scan is indicative of slight objective aperture misalignment ("wobble"), which is corrected weekly by BNC staff to minimize this motion. The STIG is the difference between the optimal focus point for the X size and the Y size, and should be less than 20 points.
BE Mark Detection (INITBE): The tool scans the beam over a 1 µm wide gold cross on a silicon substrate at z = 0 µm (the BE Mark), and another identical mark at z = -50 µm (the Under BE Mark). These are used to determine the position of the beam during further calibrations. This is run twice in order to allow the user to check the difference between the two scans in the log file, if they so desire. A difference of more than ~5 nm between scans indicates the beam may not be properly stable.
Main DEF. Correction (PDEFBE): Main deflector correction. The position of the BE mark is measured as the stage moves to the edges of the maximum main field size (1000 µm). Then the edge points are retraced, and the positions measured again. The Gain and Rotation calibration values for the main field are adjusted. Then the same thing happens for the Under BE Mark. This allows the tool to calibration how the beam position changes with height, setting the height correction coefficient gain and rotation.
Sub DEF. Correction (SUBDEFBE): Sub deflector correction. The position of the BE mark is measured as the stage moves to the corners and edges of the maximum main sub size (8 µm). Then the corner and edge points are retraced, and the positions measured again. The Shift, Gain and Rotation calibration values for the sub field are adjusted. Then the same thing happens for the Under BE Mark. This allows the tool to calibration how the beam position changes with height, setting the height correction coefficient shift, gain and rotation.
Distor Correction (DISTBE): Distortion correction. The position of the BE mark is measured as the stage moves to a 7 x 7 grid throughout the write field. Then the grid points are retraced, and the positions measured again. Placement corrections are calculated and applied for each point. This correction take the most time (~4 minutes), especially if the DAILYCAL has not been done for a few days.
Save: This saves the values to the restore file for this current. It is the equivalent of clicking the lower left Update button followed by the lower left Save button.
Once the buttons are ungreyed, execution is complete.
3.7 VisualJob (Beamer) and Magazine File
If not done before, while JOBCAL runs (or whenever), on Beamer PC, launch Beamer and run visual-Job as described in Data Prep - 5: Beamer- VisualJob above.
Use USB stick to transfer all relevant JDF, SDF, and V30 files from Beamer PC F:/alias/thisexposure folder to Linux PC within job/user/alias/thisexposure.
Within job/user/alias/thisexposure, open (display) SDF text file, update Offset to Global Mark Detection Offset. or measured chip center in alignment microscope, Save, close text editor.
Launch SCHD GUI (SCHD Purple Square), drag SDF, compile job, ensure no errors, close.
Click to expand for detailed step-by-step and video links
Video 3.7.1: v30
Video 3.7.2a: VisualJob Unaligned
Video 3.7.2b: VisualJob Aligned
Video 3.7.3: Copy Files from Beamer PC
Video 3.7.4: Make Magazine File
While DAILYCAL, runs (or whenever), create your necessary exposure files (V30s, JDFs, SDFs).
To convert your layout (GDS, OAS, etc.) into V30 files, follow the instructions in 2.4 Beamer - V30
Then, follow the instructions in 2.5 Beamer - VisualJob creation of SDF and JDF files to create your JDF and SDF file(s).
Save all of the relevant V30(s), JDF(s), and SDF(s) into your F:/alias/thisexposure folder, and copy that folder to your alias folder on the JCWLoaner USB stick.
The JCWLoaner stick is not backed up and will be wiped as needed for more space, so do not store write files on it. It is only for transferring files.
Copy your files to the Linux PC
With your V30(s), JDF(s), and SDF(s) copied to the JCWLoaner USB stick in your alias/thisexposure folder, remove JCWLoaner from the Beamer PC and plug it into the USB extension connector on the Linux PC.
Do not plug JCWLoaner into the Linux PC keyboard USB port. It doesn't fit well, and this port is a USB1.0 port, making file transfer extremely slow.
On the Linux PC, the JCWLoaner folder should automatically pop-up.
If needed, select it at the top of the Linux toolbar.
if this is full screen, click the box (Restore Window/Maximize Window) in the upper right of the window.
In JCWLoaner, open your alias folder and find your thisexposure folder.
Also open the folder to be copied to: /home/eb0/jeoleb/job/user/alias/thisexposure
Click the file drawer icon (File Browser) on the Linux toolbar (between the terminal button and camera button).
if this is full screen, click the box (Restore Window/Maximize Window) in the upper right of the window.
Under the Places listing, click job/user
Find and double click your alias folder
To quickly find your alias folder, you can click the Name or Date Modified headings to sort in alphabetical order or most recently changed.
Drag and drop your alias/thisexposure folder from the JCWLoaner/alias window to the job/user/alias window.
This will leave the original files on the JCWLoaner USB stick.
Close the JCWLoaner/alias window with the X in the upper right corner of the window, but leave open the job/user/alias window.
Update your substrate offset
⛔ Updating the substrate offset is necessary to ensure you're exposing the pattern where you want it to be. Omitting this step will write the pattern in the center of the window. This step may be omitted if using wafers, as their offset is usually small enough (~100 to 200 µm max.) to be negligible.
Open your copied thisexposure folder.
Double click the SDF file to open it with gedit.
If it asks 'Do you want to run "yoursdf.sdf", or display its contents?' click the Display button.
Alternately, right click the SDF file and choose 'Open with g-edit'
Locate the OFFSET (0, 0) line.
Change (0, 0) to the X and Y positions measured in the microscope of either your chip center (if doing an unaligned write) or your P mark (if doing an aligned write).
Repeat this as appropriate for all OFFSET lines in the file.
Each JDF file called by an SDF file will have an OFFSET line.
Click the Save button then close g-edit.
Convert your SDF to a magazine (MGN) file with SCHD GUI
With your /home/eb0/jeoleb/job/user/alias/thisexposure folder still open, click the SCHD Purple Square (called SCHD GUI) in the Linux toolbar, left of the sound, network, and time/date.
⛔ Do not include spaces in the name of the thisexposure folder, or any folders in the path to your sdf file, or SCHD GUI will fail.
This will open a blank Terminal, and the Job Complier window.
Drag and drop your sdf file from you thisexposure folder to the 'Drop SDF or MGN here' area in the Job Compiler window.
The SDF file location should show up in the white box.
Click the Compile Job button.
Wait until 'Compiling Done.' is shown at the bottom of the terminal window.
This will be instantaneous for small files, or may take 30 - 60 s for very large jobs.
With compiling done, scroll up in the Terminal window to see if any errors have occurred.
⛔ If you attempt to use a scan speed >125 MHz due to your current, dose, and shot pitch choices, this is the first time the machine will throw an error.
Some common errors result from file names not matching or trying to expose with too high of a scan speed (>125 MHz).
Errors will show up with "schd: Error" preceding an explanation of the issue. The terminal will also show '*** Abnormal End ***' above the 'Compiling Done.' message.
Close the Job Compiler window with the Close button.
The result of running SCHD GUI should be an MGN file in your thisexposure folder, as well as scanner files (.scn).
The magazine file will be selected in jbxwriter to expose the pattern.
The scanner files don't need to be interacted with, but are necessary for the exposure to occur, as they specify the beam scanning commands.
3.8 Expose
In jbxwriter Exposure page, remove any magazine files in Queue, open your MGN(s), Start Exposure.
Watch to ensure exposure completes DIRE20 and mark detection (if doing an aligned write), and exposure of pattern begins.
Click to expand for detailed step-by-step and video links
Video 3.8a: Unaligned Exposure
Video 3.8b: Aligned Exposure
In jbxwriter, go to the Exposure page.
Remove any magazine files left in the Queue by selecting them and then clicking the Remove button.
When you press the Start Exposure button, Jbxwriter will expose all magazine files listed in the queue. This can be used if you actually want to expose multiple MGNs, but will mean previous users patterns will be exposed if they are not removed from the Queue.
Load your MGN file(s) into the Queue
Click the Open button.
Navigate to your .../jeoleb/job/user/alias/thisexposure folder
Double click your MGN file to load it into the Queue (or single click it, and press the Open button in the bottom right of the window).
Repeat as necessary for all MGN files for the exposure.
Most exposures will only be of one MGN file.
Double check the Queue to ensure the proper MGN files are loaded.
Check the Job Progress area to ensure your pattern looks right, and will exposed in the right area of the window.
Zoom in by left clicking, dragging the cursor down and right, and then releasing.
Zoom out by left clicking and dragging any direction that's not down and right. Double click to return to the original zoom.
The selected JDF will be highlighted green. Select other by double clicking them.
The fields of the selected JDF/chip will be shown in chip progress, and the pattern associated with the chip will be shown in the 'Pattern' area.
In the bottom left, select the appropriate setting for 'Behavior at the time of the exposure error'
This controls whether the tool will continue exposing if an exposure error is encountered (and exposing can continue) or if it will stop. This is most relevant to aligned writes on multiple chips with chip marks, as mark detection may fail on some chips but not others. It's also relevant to exposures of multiple chips if height detection fails on some chips but not others.
'Continue' is selected by default. This is a good setting if you're exposing a single chip, or have previously encountered errors that you know may occur. But be sure to check the log file after the exposure to see if any errors occurred.
You should select 'Stop' if you're first starting to use multiple chips in a single MGN, or if you do not anticipate any errors.
Click the Start Exposure button. Watch to ensure exposure completes mark detection and PATH, and exposure of the pattern begins.
The machine will begin by executing the selected PATH (DIRE20 usually). This will take ~2 minutes.
If doing an aligned write: global mark detection will then occur. In semiauto mode, if you did not properly set your global alignment parameters, you may need to manually find some marks here.
If doing DIRE20 or DIRE05, drift measurement will then occur. This will measure the current then the position of the best mark the machine has for drift:
For unaligned writes: BE Mark
Global marks, no chip marks: P Mark
Global marks and chip marks: M1 Mark.
If this fails, you forgot to save the proper beam drift parameters.
For unaligned writes: See 3.5.1 Unaligned Exposure Material Corrections
For aligned writes, see 3.5.2.7 Save the Beam Drift parameters
After DRIFT completes, the machine will begin exposing your file. The scanning of the beam can be seen on the left monitor.
jbxwriter will indicate progress in multiple ways:
The Exposure Progress bar will show completion of chips
The elapsed time and remaining time will show times estimated for your exposure by SCHD GUI. These are usually pretty accurate.
PATH will show the PATH name (DIRE20) if it is executing. This will execute at set intervals (20 minutes for DIRE20) and before each new JDF file.
Completed and current chips will be indicated by blue and yellow overlays, respectively, in the job progress and chip progress areas.
If there are problems:
The status tower will flash yellow or red, depending on the severity.
Inspect the issues as reported in the software. If the errors are minor, you should clear them with the Confirm button, and by opening the Alarm window, reviewing the message, and then closing it.
Submit a support ticket with any questions about errors.
When exposing is finished:
The stage will automatically move to the origin (ORG) position.
The status tower will flash green instead of the usual solid green.
A popup will show 'Exposing Complete'. Click the OK button to acknowledge this.
3.9 Unload Cassette from Machine
Ensure cassette is at origin (Stage Coordinate System ~100/100). Cassette should move here automatically after a successful exposure.
Open Exchange Chamber Access Door. At LL, ensure LOAD/UNLOAD button is green and all 3 side LEDs are lit, hold LOAD/UNLOAD to unload cassette to LL and automatically vent LL.
Open LL, remove cassette from pouch, insert pouch into LL, hit the EVAC/VENT button. Close door.
Click to expand for detailed step-by-step and video links
Video 3.9a: Unloading - Load Lock View
Video 3.9b: Unloading - Jbxwriter View
In the Stage Control tab → Fixed Position drop down, select the ORG position, then click the MOVE button.
This step is unnecessary after an exposure has just completed, as the machine will automatically move the stage to ORG after a successful exposure.
⛔ NEVER press the Regist. button under the Fixed Position drop down! This would change the position of important system locations and cause problems for you and subsequent users. If you do inadvertently press the Regist. button under the Fixed Position drop down. a confirmation pop-up will appear, select Cancel. If instead you click OK, the previous system position will be overwritten. If you have inadvertently overwritten this position, please submit a support ticket immediately so the position can be remapped.
Open the Exchange Chamber Access door.
Ensure the:
LOAD/UNLOAD Button is GREEN
STAGE INITIAL LED is GREEN
CASSETTE ON STAGE LED is GREEN
EXP READY LED is GREEN
⛔ If either the LOAD/UNLOAD Button or the STAGE INITIAL LED is off, DO NOT PROCEED. The cassette is not at the origin (or the machine does not believe it is at the origin). It must be moved there first or an error requiring staff intervention will be thrown. If this error occurs, submit a support ticket immediately.
Open the cover of the LOAD/UNLOAD Button, press and hold until it beeps and begins flashing (3-5 seconds). Wait for the unload, and for the load lock to vent.
⛔ You should only ever press the LOAD/UNLOAD button if it is green.
Once the LL is at atmosphere and the LOAD/UNLOAD button has stopped blinking, turn the LL knob 90 degrees CCW to disengage the lock.
Open the LL handle fully.
⛔ Remove the pouch, holding it with the handle end higher than the hook end. If it is held improperly, the cassette will fall out!
⛔ Ensure at all times that the pouch with cassette is kept level or with the handle end slightly higher than the hook end.
Remove the cassette from the pouch and place it on a cleanroom wipe on the table.
Insert the pouch into the LL.
⛔ When sliding the pouch into the LL, be sure also to not scrape the Teflon holding block with the pouch.
Close the LL handle.
Close the LL knob, rotating it 90 degrees CW to engage the load lock.
Open the cover of the EVAC/VENT button, press and hold until it beeps and begins flashing.
⛔ You should only ever press the EVAC/VENT button if the CASSETTE ON STAGE LED is OFF.
Close the Exchange Chamber Access door.
Leaving this open will change the temperature of the column, adversely affecting your write and the writes of later users.
3.10 Finishing
Load current for next user, if they have indicated a current in iLab.
Dismount your sample(s) from cassette, place cassette back in plastic box.
Ensure the height rings are properly arranged.
Close Beamer, log out of Beamer PC.
Disable 8100 in iLab Kiosk
Click to expand for detailed step-by-step and video links
Video 3.10: Finishing
Load current for next user, if they have indicated a current in iLab.
Dismount your sample(s) from the cassette.
Place cassette back in plastic box.
Ensure the height rings are properly arranged.
Close Beamer, log out of the Beamer PC.
End the session in the BRK Lithography Core iLab Kiosk.
4 Supplements
4.1 Cassette Window Identifiers
Program | jbxwriter | jbxwriter | visualJob | jobmaker | |
|---|
Cassette Identifier | Cassette Type | Type / Size / Window | Type / Window | Cassette Name |  |
Location | “Cassette Type” Button | Material Coor. Window | Substrate | Job Settings Window |
Cassette | | | | |
3” Wafer | Multi Wafer 3A | Wafer / 3.0 / 3A | MULTI / 3A/W | 3multi A |
Piece Holder - 3” | Multi Wafer 3G | Wafer / 3.0 / 3G | MULTI / 3G/W | Piece3 |
Piece Holder - 2” A | Multi Wafer 2A | Wafer / 2.0 / 2A | MULTI / 2A/W | Piece2 A |
Piece Holder - 2” B | Multi Wafer 2B | Wafer / 2.0 / 2B | MULTI / 2B/W | Piece2 B |
4” Wafer | Wafer 4.0 | Wafer / 4.0 / None | WAFER / 4 | 100mm Wafer |
6” Wafer | Wafer 6.0 | Wafer / 6.0 / None | WAFER / 6 | 150mm Wafer |
4.2 Example SEM Values
Current (nA) | Brightness | Contrast | Contrast Multiplier |
|---|
2 | 1225 | 1635 | x56 |
10 | 1975 | 3590 | x1 |
30 | 1865 | 2785 | x1 |
100 | 1332 | 1805 | x1 |
4.3 Example Mark Location
Mark Type | Substrate Type | Mark | X | Y | |
|---|
Global Marks | Wafer | P | - | 0 |  |
|---|
Q | + | 0 |
R | 0 | + |
S | 0 | - |
Piece | P | - | + |
Q | + | - |
R | + | + |
S | - | - |
Chip Marks | Either | M1 | - | + |
|---|
M2 | + | + |
M3 | + | - |
M4 | - | - |
4.4 Example Mark Composition
Mark Tone | Safe Parameters for Silicon Substrates | General Rules |
|---|
Liftoff | For samples that will never be placed in an ICP RIE chamber If etching in ICP RIE systems, do not use gold. Instead of Au (gold), use either: 200 nm Ni 80 nm Pd use etched marks.
| For non-silicon substrates or other mark materials, follow Rooks' Rule:
Tmin (nm) = 2600 / (Zmark - Zsub) where: Tmin (nm): the minimum mark thickness for easy to detect marks Zmark: Atomic number of mark material Zmark: Atomic number of substrate material
|
|---|
Etched | 1 μm deep (can go thinner, you'll need to experiment with how thin if this matters). | Width must be < 5 μm. Depths can be between 200 - 1000 nm and work, but you may need to employ more advanced mark detection methods. |
|---|
4.5 Example Mark Dimension
Mark Type | Mark Width, X/Y | Mark Length, X/Y |
|---|
Global Marks (P, Q, R, S) | Wg = 5 µm | Lg = 250 µm |
|---|
Chip Marks (M1, M2, M3, M4) | Wc = 2 µm | Lc = 150 µm |
|---|
Wg = Global Mark Width. Lg = Global Mark Length. Wc = Chip Mark Width. Lc = Chip Mark Length. |
|---|
4.6 Example Mark Scan Parameters
Mark | Scan Position, X/Y | Scan Width, X/Y |  |
|---|
P Rough | 3 * Wg | 3 * Wg |
P Fine (& all else) | 50% * Lg | 3 * Wg |
Q Rough | 3 * Wg | 25 * Wg |
Chip marks | 40% * Lc | 3 * Wc |
4.7 Example RG Detect Condition
Tab | Suggested Basic Values (Note: Italicized text is a variable/placeholder) |
|---|
Scan | PDEF, X→ Y, 50%*ThisMarkLength (For X & Y), 3*ThisMarkWidth (For X & Y), 1, 3, 19000 |
|---|
Offsets | All 0 |
|---|
Scan Type | Fine, Box, Cross, ThisMarkWidth, ThisMarkLength |
|---|
Gain | Don't check available, Add signal (first derivative), SEM Multiplier, SEM Contrast, SEM Brightness |
|---|
Correlate | Check Available, 100, 100, 100, End of Data Addition, 0, 50 |
|---|
Allowance | Check Available, 2.5, 2.5, 2.5, 2.5, 50, 50, 50, 50 |
|---|
Raster | 0.000, 1 |
|---|
Retry | Global Marks: 1, 0.5, 0.5
Chip Marks: 10, 0.5, 0.5 |
|---|
Return to 3.5.2.4
5 Frequently Asked Questions
5.1 What if jbxwriter software is not up when I come to the machine?
Click to expand for a detailed answer
First verify that jbxwriter isn't just running on another desktop, check the bottom right to switch between all three desktops.
If the software has indeed been closed, open jbxwriter with the icon on the desktop. When it opens, ensure the "on" icon in the bottom left is green. If so, the software is ready to use. If the "on" icon is blue, contact BNC staff.
5.2 How can I check if the beam is on?
Click to expand for a detailed answer
In the current version of jbxwriter, the beam should usually be off and should not turn on unexpectedly. However, the software does not have a reliable beam on/off indication, and you can ensure the beam is off by measuring current at the Faraday Cup (FC).
To verify the beam is off
In jbxwriter, under Stage Control > Fixed Position, select ‘FC’ and click MOVE
Under Monitor > General, ensure current is 0 pA ( /- 1 pA).
If the current is non-zero, the beam is on. Turn it off:
Under Image Control, press the Spot button (above the Spot tab)
This will turn the beam “on”, even if it is already on. A white dot will appear on the left hand monitor.
Click Spot again to turn the beam off. The white dot will disappear.
Ensure the current is 0 pA ( /- 1 pA).
5.3 How can I execute a Height Map?
Click to expand for a detailed answer
The Height Map function (Located in the Calibration → Material Corr. → Height Map tab) can be used to quickly measure heights across the samples. The machine will step through a grid, and measure the height at each point
Basically, in the tab: After selecting the appropriate holder, input the sample center in material coordinates as the Offset, then give the number of points in X and Y and their spacing. Click Execute. Once it's complete, click the "Log" button, and see the output of the subprogram.
For measuring height across the sample, use a height map:
Go to Calibration > Material Corr. > Height Map
Select the proper Material Type / Size / Window
Input the chip center (in material coordinates) into Offset [um]
Define the measurement array. A height map will measure height across an array, centered at the Offset position.
Meas. Points [point] is the number of X and Y points in the array (E.g. Values of 5 and 5 will give a 5 x 5 array). The total number of points measured is Xpoints * Ypoints.
Meas. Pitch [um] is the point separation (E.g. Values of 500 and 500 will give 500 um in X and Y between each point).
The total array size is (Xpoints * Xpitch) * (Ypoints * Ypitch)
Note: The array will continue to run even if no height is measured at a point, meaning the array can be larger than your chip
After setting up the parameters, in Calibration > Material Corr. > Correction File, click the “Height Map” box, uncheck all others.
Ensure the beam is off! See above for details.
Click the Execute button in the lower right, wait for the measurement to finish.
Click the Log button, results will appear in the Calibration tab after execution is complete.
5.4 What do I do if I see a Load Error (ebloaderd)? How can I avoid it?
Click to expand for a detailed answer
This error seems to have started with the installation of a new VAC CONT board back in 2018 (maybe). This board is one of the three computers that controls the system, and is responsible for load/unload operations.
With the old board, if there was no cassette on the stage and no cassette in the load lock, and the LOAD/UNLOAD button was pressed, the system apparently ignored it. With the new board, there seems to be a change in behavior: now, pressing the LOAD/UNLOAD button without a cassette on the stage or in the load lock will cause an error in the software: "ebloaderd - Failed to start LOAD_EVAC or LOAD_VENT. Abnormality occurred in the program for monitoring for autoloader unit status."
Clearing this error requires staff intervention, as the process, if done improperly, could vent the column, cause a loss in column vacuum pressure, or change critical software parameters. For this reason, if you accidentally encounter this error, please email Justin or Bill Rowe immediately, and we’ll get it resolved as soon as we can. Users are also encouraged to report the problem on the 8100 wiki page (preferably via the Equipment Status section of the page, though the comments section works as well) so other users know about it.
To avoid seeing this error:
5.5 What do I do if AE or BE Mark detection fails?
Click to expand for a detailed answer
NOTE: This does not apply to finding the BE mark for DRIFT/Beam Drift correction. In that case, see answer 14.
If you attempt this, please submit a support ticket to let us know you had the issue and had to take these steps.
DO NOT change anything beyond what is described below in the AE/BE detect condition tab, as this could negatively and unpredictability affect your write and the writes of other users.
If the AE or BE mark detection fails:
Go to the Calibration page → Std. Mark Detection tab. Ensure both AE Mark Detection and BE Mark Detection boxes are checked.
Go to the AE Mark subtab, click the AE detect condition button. Change the Scan Width (NOT Scan Position) from 4 to 24 for both X and Y.
Go to the BE Mark subtab, click the BE detect condition button. Change the Scan Width (NOT Scan Position) from 4 to 24 for both X and Y.
Click Execute in the lower right corner. If detection is successful, click Save. If it fails, contact BNC Staff.
Then, go to the AE Mark subtab, click the AE detect condition button. Change the Scan Width (NOT Scan Position) back to 4 for both X and Y.
Go to the BE Mark subtab, click the BE detect condition button. Change the Scan Width (NOT Scan Position) back to 4 for both X and Y.
Click Execute in the lower right corner. If detection is successful, click Save. If it fails, contact BNC Staff.
Then, rerun the DAILYCAL batch file.
Remember: Do not change anything else in the AE/BE detect condition tabs, as changes could negatively and unpredictability affect your write and the writes of other users.
5.6 What do I do if Beamer has an "all licenses are in use error"?
Click to expand for a detailed answer
Beamer has two licenses, one on the Raith/BRK 1239 PC and one on the Beamer PC next to the JEOL. If a previous user has not shut down Beamer, you will not be able to use it. If you encounter this error and have reserved the Raith/BRK 1239 or JEOL in Beamer, you are within your rights to restart the PC. This will cause the other user to lose their work, but the should have been courteous enough to save and exit Beamer in the first place. Once the PC is restarted, the Beamer license will be free again.
5.7 Why am I getting a CHIPAL error?
Click to expand for a detailed answer
A CHIPAL error has been seen when software setting for virtual marks are incorrect, or when the sample height is out of range due to mis-mounting.
5.8 Why is the beam drift correction not executing properly?
Click to expand for a detailed answer
Beam Drift correction (the DRIFT subprogram) executes automatically as part of the DIRE20 and DIRE05 calibration routines. When used during an exposure, it automatically uses the most accurate drift mark available to it:
For unaligned writes: BE Mark
Global marks, no chip marks: P Mark
Global marks and chip marks: M1 Mark.
The DRIFT scan settings for the BE mark are contained in the 0_currentnA.mcorrprm file, and the user double checks the location of the mark as part of 3.5.1 Unaligned Exposure Material Corrections.
The location of the P global mark or M1 chip mark will be used automatically, the scan settings must be applied by the user as part of 3.5.2.7 Save the Beam Drift parameters.
If a DRIFT error like the one below appears, the SOP has not been properly followed:
S/N ratio during X-direction mark detection exceeded allowable value
An abnormality occured while the execution was being requested to the signal processing function (BCS)
Failed to measure the beam positional deviation (DRIFT)
DRIFT: An error occured when the subprogram was executed.
To resolve this:
Older, likely fixed issues
Q: I'm seeing an issue that may be due to lack of current measurement: The current is not being measured after DAILYCAL, or AGC won't run properly even when I can clearly see my mark?
A: The current measurement subprogram may have been corrupted. In jbxwriter, under the Calibration tab, go to the Curr. Meas. sub-tab. Ensure only Current Measurement is checked, click the Execute button. Then Save (lower right), then Update, then Save (lower left).
