Specific Process Knowledge/Etch/Etching of Silicon/Si etch using RIE1 or RIE2

Feedback to this page: click here

RIE (Reactive Ion Etch) can be used for etching silicon - both crystalline and polycrystalline silicon. The etch can be isotropic or anisotropic with vertical sidewalls depending on the process recipe and the masking material and geometry. The Si is etched by flour radicals that are created from SF₆(g) in a RF generated plasma.

Most used recipe (this recipe is being watched in the in the quality assurance program, see the QC procedure here):

Quality Controle (QC) for RIE1 and RIE2

QC Recipe:	OH_POLYA
SF₆ flow	32 sccm
O₂ flow	8 sccm
Pressure	80 mTorr
RF-power	30 W
Etch Load	50%

QC limits	RIE1	RIE2
Etch rate in Si	390 - 620 nm/min	358 - 554 nm/min
Non-uniformity	<8.6 %	<11.0 %

Hint: When using resist as masking material then do a 15 sec. buffer dip before the etch to remove any residual of resist which is normally left after the photo lithography step. A short O₂ plasma will also do.

Recipe:	OH_POLYA
SF₆ flow	32 sccm
O₂ flow	8 sccm
Pressure	80 mTorr
RF-power	30 W

	Expected results in RIE1	Expected results in RIE2
Etch rate in Si	~0.40 µm/min	~0.45 µm/min
Etch rate in Si₃N₄	.	~0.05 µm/min
Etch rate in SiO₂	.	~0.02 µm/min
Etch rate in photoresist	~0.06 µm/min	~0.05 µm/min

The etch load affects the results

The etching results are dependent on the mask geometry, primarily the etch load (meaning the area fraction of the silicon substrate that is exposed to the plasma).

To illustrate this, you can see on the images below the effect of etch load on the results of the OH_PolyA recipe. The first one has a very low load of 1-10% and the second has a load of 50%. It can be seen that the sidewalls are closer to vertical with 50% load than with 1-10 % load. Also the etch rate is effected by the load.

Etch load <50%:

The etch rate is approximately the same.
The etch profile is more isotropic

Etch load >50%:

The etch rate decreases almost linear with the exposed area. The etch rate of a 100% load has once been seen to be reduced to about 65% of the etch rate of a 60% load.
The etch profile is not expected to change significantly.

RIE Travka results - Silicon etch rate and profile as function of etch load

More information about the etch rate and profile as function of the etch load for RIE1 and RIE2 can be found here:

RIE1 Travka results

RIE2 Travka results

A description of the Travka mask set can be found here:

Travka mask set

Etch of small structures

RIE1:

A Design Of Experiment (DOE) has been performed. Here is a short summary:

Recipe:
SF₆ flow	30 sccm
O₂ flow	20-29 sccm
CHF₃ flow	5-15 scmm
Pressure	31-41 mTorr
RF-power	20 W

More details on the DOE can be seen here: "DOE RIE1 silicon etch of small structures": Media:RIE1_of_Silicon_20070306.pdf

Here are a few images of some of the single experiments in the DOE. The width of the trenches are roughly 1.5µm:

O2:24.5sccm CHF3:10sccm Pressure:36mTorr	O2:29sccm CHF3:15sccm Pressure:31mTorr	O2:20sccm CHF3:5sccm Pressure:31mTorr

RIE2:

A Design Of Experiment (DOE) has been performed. Here is a short summary:

Recipe:
SF₆ flow	30 sccm
O₂ flow	20-29 sccm
CHF₃ flow	5-15 scmm
Pressure	36 mTorr
RF-power	20 W

More details on the DOE can be seen here: "DOE RIE2 silicon etch of small structures" (not available yet).

Here are a few images of some of the single experiments in the DOE. The width of the trenches are roughly 1.5µm:

O2:24.5sccm CHF3:10sccm	O2:29sccm CHF3:15sccm	O2:20sccm CHF3:5sccm