• EBeam Home
• News
• Remote Work
• Training
• Documentation
  • Overview
  • Safety
  • CAD
  • PatternPrep
  • JobFiles
  • Exposure
    • EBeam Basics 1
    • EBeam Basics 2
    • Dose
    • Planning
    • PATHs
    • Averaging
  • Alignment
  • JEOL Documentation
  • Operation
• Process
• Reference
• Site Index
• Site Map
• Website Updates

Exposure Averaging Techniques

The JBX-6300FS offers two alternative writing methods that introduce some averaging to the beam writing, which provides a degree of noise reduction, providing smoother line edges and stitch boundaries. Next, each writing mode is illustrated with an animation. Each animation shows a simple pattern consisting of 4 fields of shapes, together forming a simple letter "F". Below the animations, there are details about how to utilize these averaging techniques in your exposures.

Normal Writing

First, this illustrates the "Normal" writing sequence. In this case, the stage is moved in sequence to each of the 4 fields, and the shapes in each field are written scanning from top to bottom.

Overlap Writing

In Overlap Writing, the stage sequence is the same, stepping to each of the 4 fields in sequence, but in this case, the pattern is not written in a single scan. Instead, the pattern is written in multiple scans repeatedly writing over the same area. The number of overlapping scans n can range from 2 to 8 times. Note that in order to write the pattern with the same specified exposure dose, each of the n scans must be n times faster than the normal writing above, so that the dwell time per pixel during each of the scans is 1/n of the dwell time of the normal case. This places some restrictions on the writing conditions then, that the clock rate must still always be below 50 MHz for each of the overlapping scans. Depending on your beam current, exposure grid (shot pitch), and writing dose, you may have to adjust one or more of those parameters in order to use overlap writing. The animation here demonstrates OVERLAP 4, in which the fields are scanned 4 overlapping times.

Field Shift Writing

Field Shift Writing, is similar to overlap writing in that it writes the pattern in multiple, overlapping passes, but in addition, as the name implies, adds additional stage shifts between each writing pass. This has the effect of not only averaging over time as in Overlap writing, but also averaging with position within the exposure field, thus further reducing the effect of deflection-induced aberrations and stitch errors. Consider the animated example here, which shows Field Shift with n = 2, in which the field is shifted by 1/2 of the field size between the 2 passes. If you consider a point on a field boundary in the first pass, that same point is written near the field center on the second overlapping pass. Field Shift writing places the same restrictions on the writing conditions as overlap writing, that the clock rate must still always be below 50 MHz for each of the overlapping scans. Depending on your beam current, exposure grid (shot pitch), and writing dose, you may have to adjust one or more of those parameters in order to use overlap writing.

Using Averaging During Pattern Writing

Using OVERLAP writing

To use OVERLAP writing, in your JDF file, you modify the P(#) pattern definition command by appending ,OVERLAP,n where n is the number of overlapping passes, from 2-8. Note the commas are required punctuation. Here's an example:

P(3)  'uwebeam0008.v30'(0,0), OVERLAP,2 

Using FIELD SHIFT writing

To use FIELD SHIFT writing, in your JDF file, you modify the P(#) pattern definition command by appending ,MULTI,n where n is the number of overlapping passes, from 2-5. Note the commas are required punctuation. Here's an example:

P(2)  'uwebeam0010r3_MU2.v30'(0,0), MULTI,2 

NOTE that to use FIELD SHIFT writing, you must also prepare your pattern with a field size of 1/n, where n is the number of multiple passes required. So for a 4th-lens exposure, with a normal field size of 500 µm, to use Field Shift 2, you must tell LayoutBEAMER that the field size is 250 µm. Likewise, for a Field Shift = 4, the Field Size used in the pattern preparation must be 125 µm, of 1/4 of the full 500 µm field size. If you don't match the Field Size in your .v30 pattern file from LayoutBEAMER to the specified MULTI value, you will get a fatal error at job deck compile. (You may notice that in the example above, I append "_MU2" to the filename of my .v30 file; this is optional, but it helps me keep track of what Field Shift value I intended to expose that pattern with. )