I need to etch 50 nm of SiO2 with 150 nm of PMMA as the mask layer. As I used Freon 23 to etch it, it looks like it takes around 50 nm PMMA out with 50 nm SiO2. Is there any way that I can reduce the etching of PMMA while dry etching SiO2?
May I know do you have other gases in the process?
From my experience, O2 should be avoided to improve the etch selectivity. Also, RF power should be lowered to reduce the physical reaction.
There is no other gases in the system. Only CHF3/CF4. RF power is 90 W. Do you know whether post baking after Electron beam lithography helps preventing etching PMMA? THanks.
Additional inert gas such as Ar and low process pressure will improve selectivity.
Hi Abdul,
Concerning your case (gases in your system), the only way to increase selectivity towards PMMA when etching SiO2 is to:
1-use CHF3 chemistry instead of Freon23
2-Increase physical part by increasing Bias polarisation (you can also do it by reducing pressure) (in fact adding Ar is best solution but nevermind)
3-and finally the 2nd point has no positif consequence on selectivity if there is no control of the wafer or sample temperature. This last point is very important. If you do not have He back-side cooling systen then I will suggest you to etch yopur SiO2 layer by short steps with cooling steps in-betwwen (based on gas flow circulation in the chamber, without plasma).
You can also find an interesting paper overthere for more information.
Regards
Mohamed
to answer your question: "Do you know whether post baking after Electron beam lithography helps preventing etching PMMA?"
IT SHOULD HELP! will it be significant for this small layer thicknesses?, that's the question!
So if you can do it: it can be interesting! (keep us in touch)
Be aware that you can use many chemically amplified resists in electron beam lithography. These resists have much tougher etch resistance, higher sensitivity, and better resist profiles. Of course, if you tried e-beam chemically amplified resists you need to keep airborne contaminants cleaned up, e.g. amines, because they can effect the chemically amplified resist (puts a surface skin or inhibition layer on the resist) The handling of chemically amplified resist is standard in the semiconductor industry, where among other things it is used to dry etch SiO2 and in advanced bi-level and tri-level resist processing.
Thanks a lot Mohamed and Harry. Sorry for the late notice as I was busy with something else. But I will keep in touch with you guys as soon as I get some interesting results.
If you continue to use PMMA, you need good heat-sinking, and you need to keep the dry etch power down and no oxygen. PMMA is Plexiglas plastic and has a low deformation temperature. It flows if it gets heated up. First thing you would need is a graph of etch rates for PMMA and for SiO2 over a range of dry etch powers, particularly the low end.
Be careful if you try a prebake. You need to know the temperature at which it begins to flow and ruin the profile. It is low. This means low and long.
In general, PMMA resist is poor etching selectivity. If you use another e-beam resist like ZEP520-A, you can basically obtain higher etch selectivity. And If you must certainly use a PMMA resist when SiO2 dry etching, you need to select fluorine-series gases like CHF2, C2F6, CF4, C4F8, Ar while excluding O2 gas. In case of using fluorine-series gases you can easily etch SiO2 layer using e-beam resist. As you mentioned, you have to bake finally after development, namely, Hard-bake. And add Ar gas (less than 10 sccm) to CHF3/CF4 gas. To improve more selectivity and to get fine pattern for e-beam lithography, you have to use ICP-RIE tool type than simple RIE type.
HI Abdul,
I am also trying kind of similar thing. Did you finally found any optimized recipe?
Thanks in advance,
Salahuddin
I have had similar problems and it seems that the pre-bake of the PMMA matters. What is you protocol before the lithography? How is your prebaking? Try baking at 155°C for at least 30 min. You can play with the times, it seems that the longer the better, but it could affect your ebeam resolution. You need to characterize it.
Hello,
Actually your reported 1:1 etch selectivity of SiO2/PMMA is one of best reported in the literature/internet. Typically people have less than 0.6:1 for this material:mask combination.
Also from your description it looks like that this selectivity is enough for your final goal, why you need more PMMA left?
I am keeping same parameter from resist to dry etching. I only change the baking temperature at 180 oC from 2 min to 15 min for PMMA A4 resist. Features are improving but not perfect.
Today I found out nearly same selectivity (1:1) while etching SiNx with C9 (PMMA) mask. I used CHF3 15 sccm and SF6 30 sccm.... I want to etch 200 nm of SiNx. Any suggestion on which gas should I use to increase the selectivity?
Dear Avijit,
For SiNx etching the selectivity could be much better. Please try a mixture of CHF3:O2:Ar with flow ratios 9:1:1.5 . Hope they are all available.
Low power levels of both ICP (~100 Watts) and RIE (~10W) are favorable, as well as low chamber pressures (~3 mTorr).
Please keep updated about your results.
Good luck.
- Badges
- Science topic
- Similar topics
- Chemistry
- Nanotechnology
- Nanostructures
- Thin Films