Hi, all
I am currently working on SiO2 ecthing by using RIE (gas: CF4 or SF6).
I used PMMA A4 as the mask by developing pattern with e-beam lithography. However, after RIE and liftoff process, I only saw some residues left and nothing were etched away. (RIE etching: Rate: 20sccm, pressure:0.2torr, power:40W, time: 6mins)
I tried to figure out why and I found there is an note on the Internet says the presence of carbon on the surface will reduce the amount of fluorine available to etch silicon. And C-F polymers would form on the surface by combing with F radicals. (Please see p.22 of the link)
So I guess PMMA (lots of carbon) can not served as the mask for etching SiO2, am I right? Or is there some RIE parameter I can try with PMMA mask? (because I have to use e-beam lithography to design the pattern)
Thank you,
http://web.utk.edu/~prack/Thin%20films/Etching.pdf
maybe you have some pmma residues within your exposed/developed structure. Try a short oxygen plasma step prior to your SiO2 etching.
Thank you, Martin Hetzl. I will tried a short O2 plasma step prior to SiO2 etching as you suggested in the next experiment.
So if there is no residue left, is PMMA an good mask for SiO2 etching? Because the carbon issue as the note says still concerns me. Thanks again!
dear Wu,
all polymer based masks, i.e. all standard photoresists used in IC, have lots of carbon in them, but this does not interfere with plasma etching as plasma is tuned to cope with the (small) amount of etched polymer.
regards,
Alvise
dear Wu,
We always do some O2-plasma before any dry etching of SiO2 or SiNx, but in case of SF6/O2 recipes actually it is not neccessary, because the selectivity SiO2 to PMMA is so bad, that we always have to coat very thick layers of PMMA to form a good mask for etching, otherwise we etch PMMA completely thus thinning the whole SiO2 film... Maybe you have that case? Are you sure, that you have full coverage by the PMMA after etching step?
Another thing is your recipe - the pressure is rather high. Conventionally it is a good practice to use pressures much lower -
Dear Bagolini and Galiev:
Thank you very much for your kind reply!
In my case,after RIE, I could still see the whole PMMA mask under optical microscope inspection, so I guess PMMA was still there although the thickness has been reduced by RIE.
I also have tried another recipe of RIE by adding some O2 during CF4 etching (CF4: 20sccm, O2: 5sccm, Pressure:0.2torr, power:40W, duration:6mins) And In this case, the PMMA mask was completely etched away, but I still could not see any etched pattern but only some residue (because I would expect although PMMA was etched away, there should be some etching height difference)
Thank you, Galiev, for your suggestion about the etching pressure, I will try low pressure conditions (0.02 torr or lower) this week. Hope this time I will make it.
Hi Galiev,
Thank you so much for asking. Sorry for being late to have feedback on this problem. The SiO2 etching is now okay with the pressure as low as 0.02torr and 0.01torr. But the PMMA A4 mask (thickness~230nm) seems to be etched away completely after RIE (only CF4 were used). So I am still working on this and trying to use thicker PMMA (A6) or different recipe for RIE. Do you have any suggestion? Thank you very much! :)
Hi Wu,
I'm sorry, but I have no suggestions yet, because we ended up with using another mask for etching - for us it is much more convenient to use thin ALD-formed Al2O3 coat as mask, which we successfully etch through PMMA by BCl3 mix and then etch SiNx or SiO2 via this mask with C3F8/O2 or SF6/O2 mix after PMMA removal. It seems to me that there is not much ways to harden PMMA to improve the selectivity. The best choice is to use another mask in that case (ZEP520, metal, Al2O3 ?). Changing the recipe probably will not help.
Hi Wu,
What etch selectivity do you need to reach your desired results? Among the reported recipes the best selectivity is achieved with CHF3/Ar mixture.
With a choice of CF4 and SF6 only I would prefer CF4 of course. If possible, add some argon (10-20%).
What are the current RF powers and etch time that lead to the complete removal of PMMA mask? Also, what is the tool you use?
Dear Victor
I am facing same problem. Can you provide detail parameters used in CHF3/Ar recipe.
Thank you.
Dear Anant Singh ,
You can use these parameters for a start:
CHF3 75 sscm, Ar 15 sccm; ICP power 100W, RF power 10W;
chamber pressure 3 mT.
These are for an ICP-RIE tool. In case your system doesn't have an ICP ability, try to some 12-15W of RF only. Also, if your MFCs allow lower gas flows only, reduce the values from above and just keep the flows ratio 5:1.
Please note that the etch result will be highly dependant on the specific tool you use, and the above values were used with the SPTS system. In case your tool shows actual DC bias values, it's better to tune the process parameters to the highest possible voltage values.
Also please be advised to keep the wafer/sample temperature as low as possible. Backside helium cooling is highly advised, and for small pieces you want to bond them to the carrier wafer either by a high-vacuum oil or substances like Crystalbond (in case your process flow allows using them; anyway proper oil and Crystalbond can be removed by acetone after etch).
Hi Anant Singh,
the pressure I had used was too high, so after I went to mtorr level (like what Victor suggested in his post), the etching was working. But in my result, we could roughly get 30nm etching on SiO2 at most because PMMA A4 is not thick enough to survive the process. But I guess you can use different resist material to get better result than I did.
Best,
Victor V Solovyev , Yueh-Chun Wu thanks...I'll update once I get any fruitful results.
- Badges
- Science topic
- Similar topics
- Engineering
- Materials Engineering