Hi, I used to use FeCl etchant or Cu. 53g of FeCl in 100ml H2O at 45°C is more than enough :) The etch rate is high. Can I ask you what kind of resist did you used? maybe the problem is with your resist/lithography process.

Hi Jose,

I have had the same problem in the past. This is particularly evident on large substrates, where the wet etch time must be prolonged to ensure all areas of the wafer are completely etched. This results in areas of the wafer exhibiting high undercuts.

One important piece of information missing in your question is what sort of area are you working on. Also, what sort of undercut (average, max and min) do you get at the moment?

There is no final solution to this problem (except switching to dry anisotropic etch), but you can nevertheless try to mitigate the effect.

You can certainly try to improve the etch uniformity across your wafer. If you reach an ideally uniform process, you can stop the etch exactly when your underlying layer is exposed. The etch will be the same across your wafer and your under-cut will be compaprable to your etched thickness (1um per side in your case, so you will get about 18um lines). To achieve this you can try for instance agitating the solution during etching. Diluting the solution also will give you better uniformity control, but you will have to increase the process time.

There are also studies that show the effect of the mask used for etching on the etch profile. I do not have experience with specific regards to copper, but you can maybe get some hints in the literature. Changing your mask (for instance from a photoresist to a hard-mask material) might help you obtain more vertical slopes, hence less under-cut.

I hope you find this helpful.

Giuseppe

A group developed a special wet etching method for copper that resulted a good patterns with no undercut (Fig. 9 in the research), that is what is said without details(in Fine Pitch Copper Etching paragraph). Contact the others for more information. 

title: FINE LINE PHOTOLITHOGRAPHY AND ULTRA HIGH DENSITY PACKAGE SUBSTRATE FOR NEXT GENERATION SYSTEM ON-PACKAGE (SOP)

others: : Fuhan Liu, Venky Sundaram, Boyd Wiedenman and Rao Tummala, Georgia institute of Technology, USA

published in: IEEE Conference Publications, Electronic Packaging Technology Conference 2007

Hi Jose. As you mentioned, undercut is inevitable in almost all wet etching processes. However the reason for  a large amount of undercut might be something else. I suggest that if you have access to SEM, take your sample before etching and observe it under the microscope. It could as well be that the edges of your photo resist might be delaminated and thus etchant can creep in and produce a large amount of undercut. Check that out. Best of luck

Thank you all for the replies. I will explain few doubts that were raised in the replies.

I am using Copper as a sacrificial layer to create MEMS structures. I need the thickness to be 1 um. My minimum feature size in lateral dimension is 20 um. I am using positive photoresist S1813 as the masking material for my etch. When I etch, ideally I am expecting a rectangular cross-section, but I am ending up with a dome shaped cross-section. The undercut is not equal if I inspect the cross-section of the etched shape. It is more at the top and as it comes down it is feeble (it is obvious). 

@ Marian : I never tired using Ferric Chloride. I was always relying on HCl and HNO3 based etchants

@ Giuseppe : Yeah...I will try using a hard mask for this purpose..It may help me..

@ Hacene, : I have gone through the paper that you have suggested, but they didnt disclose the etchant that they used to achieve the projected result.

@ Alireza: I will see the photoresist patterns in SEM.. probably this may also be a reason for the large undercut

Thanks all

Jose

And I found a solution for the problem. When we are using HCl+H2O2 etchant for copper, use H2O2 in higher concentration than HCl. That gives reasonably less undercut while etching copper. What I have used is a ratio of 10:1:2 (DI water: 30% HCl : 29% H2O2). This worked quite well and a 1um thick film got etched in less than 20 sec.