1255 (100nm NPSN) Partially Etched Super Thin TEMs

ByGreg Madejski

I’ve been trying to recover a pore distribution from my 1085 MgF2 SERS wafer that had a ~60 nm pore size on average. This has been difficult with thinner, 50 nm NPN because there is less material to shape; it’s harder to stop at a certain point to get the right pore distribution. So, using a thicker material should allow some more process flexibility to arrive at the correct distribution. I’m using some of the 100 nm 1255 NPSN material to test out to what degree I can etch and achieve the correct distribution. 1255 has one large rectangular window.

 

1255 100nm NPN pore Processor TEM and Histogram. At 13500 and 6300x, the processing software produces a reasonable estimate of pore size.

 

1255 100nm NPN pores have some isotropy to them. We have to be careful not to count the full edges of the pore.

I used a SF6/CHF3 etch as described in this post to remove material from the backside of the nanomembrane (etched the trench). This preserves the front surface more (etch will still attack sidewalls) and allows for a more fair comparison. However, it seems that I overetched considerably.

13500x image of 1255 NPSN etched for 50s with 100W 30 sccm SF6, 15 sccm CHF3, 10 sccm O2. Pores are very blown out, though the siloxynitride scum remains.

 

Tilt stack of etched 1255, -14 to +36 degrees. The membrane is very thin but intact. Hardly any sidewalls at all. The siloxynitride scum appears thicker by comparison.

 

8900x TEM image. Lot of merged pores. Average is 98nm.

Similar Posts