E-beam lithography is not something you will learn in a day. You will have to know about SEM use, structure design, resist chemistry and processes, etching processes, proximity effect, alignment, EBPG use etc. etc.

The first step is to get experienced with SEM operation. Read the SEM manual here. The SEM becomes an EBPG with the Elphy Quantum system, beam blanker and motor control. The Elphy Quantum program has been set up to divide each writing stepdivides the process in 4 phases (desktops):

1. Design
2. UV coordinate system calibration
3. Alignment or writefield calibration
4. Writing

Very much can be said about the design phase, this manual will just touch the surface. Take into account the proximity effect caused by forwardscattering of electrons in the resist and backscattering in the substrate. Ask help if you need to place features very close together.

Create a new gdsii database or open an existing one (desktop 1). gdsii is a hierarchical format where you can design elements that can be used in elements… and so forth.

Create a new structure in the gdsii database. Structures are elementary pieces of design that can be reused, put in a matrix etc. Alignment markers are typically things that you would put in a structure. Referenced structures are displayed as red rectangles, if you push <alt-i> you see the contents (increase hierarchy level, reverse with <alt-d>.

Layers and working areas form the unit of exposure. What gets exposed in a single step is exactly the set-intersection of design elements in a layer and a working area. Layers are part of the gdsii database and all structures in the database will inherit the layers. Layers are a logical/functional way of grouping design elements. Working area's are a different way of grouping elements, those that fall completely in a rectangle. Typical layer definitions would be your one for each processing stage of your geometry, large contacts and markers (they require different writing parameters / procedures). The working area —the logical writing area— must be smaller than or equal to the writing field (the physical beam deflection field). The placement of the working area in the writing field is set in the pos settings in a position list. The center of a writing field is the UV coordinate at the sample set by the stage, but not that to reproduce this between lithographic steps the electronic beam shift must be the same, so if in your sample ana alignment is going to be done, always set the beam shift to zero!

If a layer needs to be aligned to an already present pattern (a previous layer) alignment markers have to be present in this previous layer. To get an accurate placement relation between parts of the design and these markers, write critical, small (=fast) elements first to avoid drift of the machine to cause placement errors. Order the exposure sequence by pushing <o>.

The layer that has to be aligned needs to contain mark scans at the places of the markers of the previous layer. These mark scans must be large enough to be certain to 'hit' the marker and far enough away from critical parts of the design such that the starting error will not cause an e-beam scan over the critical region. Take into account an initial error of at least 2 um (the stage error plus the rotation error) and write large markers at 1,1 and 14,1 (for a 15×15 mm substrate) for UV calibration.

In a lithographic step, try to avoid changes of the write field size that require a change in the working distance, because then recalibration of writing field and UV system are needed.

When the layers and working areas of your sample have been designed, it is time to think about the beam current. See the calibration plot at the machine for spot sizes. Use 20-100 pA for small features and 5 nA for large ones as a start, make adjustments if this leads to excessive writing times (the higher the current the larger the spot, the smaller the spot the smaller the current and therefore the larger the exposure time fore a given pattern).

Position lists are Elphy's abstraction for any scanning job, whether writing a single pattern, a matrix, doing an alignment to an existing structure or calibrating the writing field size. Here, all information about your design comes together: what layer and working area to write of what pattern in what gdsii database at what place at what beam current with what resist sensitivity, beam step size, dwell time etc. The matrix copy function is handy for matrix exposures and dose tests/scans. Due to a bug in the software it is critical to select the right writefield size in microscope control and click set before dragging items into a position list (or doing anything eles for that matter).

If your positionlist is ready, it's time for the hands-on part: the exposure. Mount the sample aligned with the edge of the sample holder to minimize rotation with respect to the sample holder and while taking into account everything written in JEOL 820 SEM usage click the sample holder onto the stage. Be careful not to accidentally expose the resist by moving the stage!

1. Set the working distance to the desired value (10 mm, 8 for exceptional situations, 3×3 or 4×4 marker for large fields)
2. Select the required aperture
3. Drive the stage to the chessy and switch on the machine.
4. Switch on the SEM at 20 kV, increase HV slowly to 30 kV
5. Set the current to about 1 nA
6. Let the machine warm up for half an hour
7. Adjust filament saturation

1. Focus on the chessy
2. Zoom in
3. Correct astigmatism
4. You should see 50 nm details easily at 1 nA

Alignment is similar to writefield calibration but done on the sample markers. It is usually a good idea to still calibrate the write field (and set shift to zero!) to get an accurate starting position for alignment.

1. Go to desktop 3
2. Select your writing field in microscope control and click set
3. Open a new image and scan it
4. The image should have the right size and not be rotated

If this is not the case a write field alignment has to be performed (normally not necessary unless something weird happens like the JEOL service people recalibrating the scan field of the SEM itself )

1. Drive the stage to the center of a chessy block (point at it and click <crtl+mouse-right>. repeat if necessary
2. Open a writefieldcalibration positionlist called Align<size> or align<size>
3. Markers appear in the image
4. Drag all 4 markers to the right place. They have to be moved!
5. Click calculate and send

1. Click scan all in the writefield calibration position list
2. A macro will execute and ask you to click the marker centers. The marker has to be moved!
3. When the macro finishes and all positionlist entries have a blue marker click calculate and send
4. Set shift to zero

The UV coordinate system is the sample coordinate system. Do not perform a 3-point correction unless necessary (motor step size and stage orthogonality), I do this when I set up your account; 3-point correction requires a trick.

1. Drive the stage to the lower left corner and perform an origin correction (0,0) or drive the stage to the lower left marker and perform an origin alignment (1,1)
2. Wait 3 seconds
3. Use this point as the first point for the angle correction
4. Drive to the lower right corner or marker
5. Wait 3 seconds
6. Use this point as the second point for the angle correction
7. Perform the correction

1. Go to the middle of the sample (beam off!) or to the middle of the x-axis
2. focus on a particle
3. Burn spots at 1 nA, should be round and about 200 nm in size
4. Focus on them
5. repeat

For now, see write field calibration. I will normally teach you this procedure because it is situation dependent.

1. PCD in
2. TV scan
3. Select magnification entered in microscope control (you did click set i hope )
4. Slow 2
5. Beam on (click 'un'blank icon)
6. Set beam current electrometer to desired scale
7. Set beam current to desired value (current knobs: left=up, right=down)
8. Blank (click 'blank' icon)
9. Switch to EBPG
10. Switch blanker to External
11. PCD out
12. Click scan or scan all
13. The current you read is lower than the set value because of backscattering

NEW!! (March 2006) This takes over the fine adjustment control of the beam current using a PI controller. The dial on the controller is set such that 0-10 represents the full scale 0-1 range on the TOP beam current meter, (e.g. for 10nA on a 30nA range you choose something near to 3.20 to get the correct setting on the lower scale). The controller will actively control ~30 seconds after the beam is unblanked, and continue to do so until the beam is blanked for more than a few seconds. If this happens then the controller remembers the last value, and will use that for the first 30 seconds after the beam is unblanked again, (after which it will again start to stabilize).

1. Zoom in somewhere harmless on your sample
2. Take the PCD out, go to EBPG mode and external blanker (as if you are going to write: see above)
3. Set the beam scale and coarse control to give the approximately correct beam current (with fine control somewhere in the middle)
4. Switch on the stabilizer (switch up - above the PCD button)
5. In the software unblank the beam
6. Wait
7. After ~30seconds the PI controller will kick in, and you start to see the voltage on the display change, and the beam current approach the set point. This may take up to one minute. The display on the display is like the variable output that would come from the fine beam control. You want to have the meter in the centre of the display so it can prevent drift in both directions.
8. When you're happy the beam current is correct (adjust dial as necessary - but remember long time constant!) blank the beam
9. Zoom out to you correct magnification for your pattern
10. Write!

Important Note: The only problem facing the stabilizer is significant drift between blanking the beam, and starting to write (steps 8,9,10). If the beam is too unstable the first 30 seconds may be exposed with the 'saved' (and wrong) value before the stabilizing starts again, (you may see this by looking at the voltage on the multimeter - the beam current may be constant, but if there is a lot of drift the voltage will not be). If possible therefore, the first 30s should not be crucial.

Other Note: In the description above, the current you set using the stabilizer is the current you get during exposure, which is less than the current you measure with the PCD on. If you prefer setting the desired current with the PCD on, you have to slightly change the above procedure. Set the correct current with PCD on at a harmless spot, PCD off to see the effect of backscattering, and then use the stabilizer to stabilize this current.

Switch off the machine and remove the sample holder as described in JEOL 820 SEM operation.