Burst Pressure Tests: Rectangular Windows
Simpore produced a wafer of 50 nm NPN with a matrix of (4) widths X (5) lengths
Widths of 100 um, 300 um, 500 um, and 700 um
Lengths of 1 mm, 1.5 mm, 2 mm, 2.5 mm and 3 mm
Each chip has one membrane. These are the dimensions of the membranes.
Using the burst pressure set-up I burst 4 chips of each flavor. Some of the chips broke before the test was performed, most data points are for four chips, some are just three, and a few are just two chips. The first graph groups the chips by length. The second groups by width. You can see that while there is some increase in burst pressure with the shorter lengths, the variation is much higher for decreasing the width. Of course the width increases 7 fold at most while the length only increases 3 fold. But Even increasing the length from 100 µm to 300 µm (3 fold increase) the burst pressure decreases more than the same 3 fold increase in width. (the change in burst pressure with change in width may not even be significant; take a look at the error bars)
Please place this data in the context of Sarah and Josh’s old data for squares. If all the data is correct, then at some point below 1 mm, length does matter to the burst pressure.