Specific Process Knowledge/Etch/IBE⁄IBSD Ionfab 300/IBE magnetic stack etch

Results of Design of Experiments optimization of magnetic stack etching (Kristian Hagsted Rasmussen spring 2011)

Process parameters

To find the optimal parameters, Design of Experiments was applied with the intervals shown on the figure below. The final optimized recipe is listed in the table.

Parameter	Ti etch acceptance	Design of Experiments setup used to find optimal etch of the multi layers stack. Centerpoints are not shown on the figure. The parameters varied were Beam current, Beam Voltage, Accelerator Voltage, and Incident Angle.
Neutalizer current [mA]	450
RF Power [W]	800
Beam current [mA]	400
Beam voltage [V]	500
Beam accelerator voltage	500
Ar flow to neutralizer [sccm]	5.0
Ar flow to beam [sccm]	10.0
Rotation speed [rpm]	20
Stage angle [degrees]	5
Helium backside cooling [Torr]	37.5

Some SEM profile images of the etched stacks

Magnetic stack of Ta/MnIr/NiFe/Ta/Tao/Ta/MnIr/NiFe/Ta etched with the optimal parameters shown above. Redeposition of etched material at photo resist. Profile ~78°

Etched magnetic stack with poor structure transfer, the parameters chosen was clearly not optimal. Chosen etch parameters result in severe footing

Etched magnetic stack with poor structure transfer, the parameters chosen was clearly not optimal. Chosen etch parameters result in severe trenching

For a good pattern transfer with IBE a large amount of redeposition on the resist edge has been observed. To reduce this redeposition the best option seems to use a photoresist just thick enough for the wanted etch depth. An alternative is sidewall sputtering at low incident angle. However experiments has shown poor cleaning rate and for narrow structures low angles will give rise to shadowing effects.

Endpoint detection with SIMS

The endpoint for etching is determined by the integrated SIMS

End point detection is achieved by SIMS, and the etch rate is approximately 25 Nm/min. Resist stripping can be hard due to burned resist, to remedy this try to lower the current. Changing the currect will chance the sidewall angle and new studies of etch profiles will be necessary. For help, discussion and further info please contact Kristian Hagsted Rasmussen.

Conclusions

• Etch rate around 20 nm/min with optimized recipe.
• Good feature definition after etch.
• Good end point detection by SIMS.
• Burned resist: Lower current, 300 mA can limit resist burning but will affect pattern transfer.
• Hare ears due to redeposition, thinner resist can limit the size.