Specific Process Knowledge/Thin film deposition/ALD2 (PEALD)/AlN deposition using ALD2
General Information
The AlN process should never run above 350 °C due to the decomposition of the trimethylaluminium. As already mentioned in the general ALD2 (PEALD) page oxides and nitrides cannot be deposited at the same time. Since the aluminium nitride requires the plasma source we only run all nitride processes without the thermal lid. Hence the chamber volume is bigger than with the thermal lid which causes that the substrate temperature is around 50 °C lower than the setpoint.
Result from acceptance test
AlN with N2 plasma
The test was done with 300 cycles at 350 °C where the growth rate was measured to be 0.0625 nm/cycle.
| TMA | N2 plasma | |
|---|---|---|
| Nitrogen flow | 150 sccm | 100 sccm |
| Pulse time | 0.1 s | 21.0 s |
| Purge time | 4.0 s | 4.0 s |
Uniformity profile across 150 mm Si wafer based on 49 measurement points measured with ellipsometer. The values for the grown oxide thickness can be seen in the table below.
| Parameter | Average | Min. | Max. | Std.Dev | %Range |
|---|---|---|---|---|---|
| Thickness (nm) | 18.77 | 18.02 | 19.73 | 0.41 | 9.0619 |
AlN with N2 plasma on trenches
The test was done with 1000 cycles at 350 °C.
| TMA | N2 plasma | |
|---|---|---|
| Nitrogen flow | 150 sccm | 100 sccm |
| Pulse time | 0.1 s | 26.0 s |
| Purge time | 4.0 s | 15.0 s |
Additionally the picoflow was used for the TMA precursor.
As the SEM images above show only a thin layer of AlN could be deposited at the bottom of the trenches. On the top the growth was significantly higher which lead to a tree like growth. Whereas in the images below no picoflow was used with 500 cycles and the same other parameters as in the run shown above. Without the picoflow a more homogeneous coverage could be obtained and an average growth rate of 0,05585 nm/cycle was measured.
