Usually a combination of RF and DC sputtering is preferred compared to PLD for fabricating CoFeB/MgO/CoFeB MTJ's. Has this got something to do with the composition of boron? Are we not able to maintain the stoichiometry for CoFeB using PLD?
This may not be the case for all PLD systems, however for the one my group are currently using, we cannot control the thickness to the fine level needed in the layers of an MTJ, where the tunnelling barrier is usually no more than ~ 1 nm. Our level of PLD thickness control amounts to ~ 10 nm. PLD ablates a blob of material, and the deposited amount per pulse may not be as easily tunable as it is with sputtering, where sub-Angstrom per second growth can be achieved.
However, stoichiometry may not be an issue, since PLD generally maintains the composition of the target in the deposited film. One final consideration is that MTJs are very much a device that is intended for industrial use rather than purely for research aims. Sputtering is something that can be scaled up to large volumes in order to produce commercial quantities of devices, whereas PLD is more of a research tool for smaller batches of samples. Because of this, if you have already optimised growth of your MTJs by sputtering, that recipe can more easily be transformed into a viable commercial operation. This is related to the reason that GMR took off properly when Stuart Parkin at IBM managed to grow GMR multilayer structures by sputtering, even though the original GMR multilayers of Albert Fert and Peter Grunberg were made by MBE.
- Badges
- Science topic
- Similar topics
- Chemistry
- Nanotechnology
- Nanostructures
- Thin Films