Posted by Les Szepesi (63.135.212.115) on July 03, 2012 at 18:03:58:
In Reply to: Re: Is oxynitride a problem? posted by Dan on July 03, 2012 at 16:50:43:
: : I know you don't like having nitride passivation on the SRA samples, but is oxynitride a problem? I'm designing some test structures for SRA, but I'm afraid there won't be adequate selectivity between the oxynitride and underlying oxides and we'd either break through to the silicon and possibly erode the silicon or we'd make the physical definition of the SRA boxes degraded as they will be defined by variations in oxide thickness.
: First of all, thank you for designing SRA structures. We appreciate it when our clients incorporate structures we can use into their masks.
: I believe the problem would depend on the thickness and hardness of the oxynitride. Our main issue with passivation nitride is that it is usually thick - several microns thick in some cases. And since it is not easily susceptible to HF stripping we have no way of removing it. This means we have to bull our way through it while it is still on the sample. The hardness and the thickness of the material creates problems getting good clean cross section bevels. Especially when our depth of interest is a fraction of the nitride thickness.
: I don't know how hard oxynitride is. We use tungsten carbide probes to make our measurements. If the oxynitride is harder than the probes it could blunt them and knock them out of calibration. I know some forms of oxynitride can form a ceramic layer. I imagine those forms would be too hard for us to leave on the samples for fear of damaging the probes.
: If the oxynitride isn't too hard, you may be able to keep the oxynitride to a few tens of nanometers thick and we could bevel through it without stripping it. If it is an HF strippable material we could strip it when necessary. Unfortunately we do not have a plasma etcher.
: I'm sorry I don't know more about oxynitride. If I learn anything useful I will append more to this thread.
In our particular case, we're using 1.5um of PECVD oxynitride, which is etchable in HF. My biggest concern was that if we used HF to remove the oxynitride, the various oxide windows that define the SRA structures would be destroyed. However in three of the five cases, there will be an oxidation in the structure that should give a large enough step in the silicon to make them visible.
What I will do is put a ring around the other two structures that will receive the same oxidation.
In general, this deposited SION film is nowhere near as "robust" as LPCVD nitride. It isn't as much of a barrier nor is it as scratch resistant, but it does do better than plain oxide in these regards.