Specific Process Knowledge/Etch/ICP Metal Etcher: Difference between revisions

Revision as of 15:18, 23 February 2011

Etching of nanostructures in silicon using the ICP Metal Etcher

**The starting point**
Break	Gas	Cl₂ 20 sccm
	Pressure	2 mTorr, Strike 3 secs @ 5 mTorr
	Power	600 W CP, 200 W PP
	Temperature	20 degs
	Hardware	100 mm Spacers
	Time	15 secs
Main	Gas	HBr 20 sccm
	Pressure	2 mTorr, Strike 3 secs @ 5 mTorr
	Power	900 W CP, 50 W PP
	Temperature	20 degs
	Hardware	100 mm Spacers
	Time	? secs

ER 200 nm/min, 3:1 over resist. Vertical profile. To improve selectivity to oxide under-layers you can add a small amount of O2 ( e.g 2 sccm if the MFC is small enough). This should not give an undercut.

Standardization images of the shallolr recipe
The 30 nm trenches are somewhat wider due to overexposure of E-beam resist

The 30 nm trenches are somewhat wider due to overexposure of E-beam resist
The 30 nm trenches are somewhat wider due to overexposure of E-beam resist

The 30 nm trenches are somewhat wider due to overexposure of E-beam resist

It looks like the trenches are closing up as the etch goes deeper, consistent with too much polymer deposition on the sidewall.

The options I would look at are

       1) to reduce the coil power to 700W so that the dc bias of the etch increases & you sputter more of the polymer off.
       2) increase the platen temperature to 40 C so as to reduce polymer condensation on the wafer.
       3) substitute Cl2 for some of the HBr - try a 50: 50 mix of Cl2 / HBr for the main etch with the same total gas flow.  This will reduce the amount      of polymerising species in the plasma & therefore help reduce the amount of sidewall deposition.   I would try these separately.

Just as a sanity check, have you got any pre-etch SEM images of the ZEP? There's a contradiction in the SEM images - in that you are seeing what looks to be undercut, whilst the trenches are also closing up. One effect is consistent with too little polymerisation whilst the other is caused by too much. The breakthrough step is actually quite a physical & directional etch, which usually doesn't cause an undercut. What exposed area do you have on the samples?

@@ Line 48: / Line 48: @@
 <gallery caption="Standardization images of the shallolr recipe" widths="300px" heights="300px" perrow="2">
-Image:jmlshal070921 pos1 2mu_09.jpg|The profile of a 2 <math>\mu</math>m trench
+Image:WF 2F-6 dec092010-30 nm.jpg.jpg|The 30 nm trenches are somewhat wider due to overexposure of E-beam resist
-image:jmlshal070921 pos1 50mu_08.jpg|The profile of a 50 <math>\mu</math>m trench
+image:WF 2F-6 dec092010-120 nm.jpg.jpg|The 30 nm trenches are somewhat wider due to overexposure of E-beam resist
 </gallery>
+It looks like the trenches are closing up as the etch goes deeper, consistent with too much polymer deposition on the sidewall.
+The options I would look at are
+) to reduce the coil power to 700W so that the dc bias of the etch increases & you sputter more of the polymer off.
+) increase the platen temperature to 40 C so as to reduce polymer condensation on the wafer.
+) substitute Cl2 for some of the HBr - try a 50: 50 mix of Cl2 / HBr for the main etch with the same total gas flow.  This will reduce the amount      of polymerising species in the plasma & therefore help reduce the amount of sidewall deposition.   I would try these separately.
+Just as a sanity check, have you got any pre-etch SEM images of the ZEP?  There's a contradiction in the SEM images  - in that you are seeing what looks to be undercut, whilst the trenches are also closing up.  One effect is consistent with too little polymerisation whilst the other is caused by too much.   The breakthrough step is actually quite a physical & directional etch, which usually doesn't cause an undercut.  What exposed area do you have on the samples?