| Composition HF : HNO3 : HAc | Temp. oC | Etch rate µm/min |
Anisotropy <100> : <111> |
Masks for selective etching | Remarks |
|---|---|---|---|---|---|
| 1 : 3 : 8 | 20 | 3 | 1 : 1 | none | Etch rate strongly reduced for doping n: <3 · 1017 , p: <3 · 1015 |
| 1 : 2 : 1 | 20 | 40 | 1 : 1 | none |
Isotropic Etching of Silicon
| Module is unfinished | |||||||||||||||||||
|
If we neglect exotic mixtures of chemicals, the universal isotropic Si etchant is a mixture of HNO3 + HF + CH3COOH . | ||||||||||||||||||
|
In other words: Mix nitric acid, hydrofluoric acid and acidic acid (= HAc)! The resultant mixture is known as CPx; the x being 1 ...4, depending on the exact composition - and this abbreviation was chosen for good reasons! | ||||||||||||||||||
|
The etching characteristics you obtain depend very much on the exact composition and the temperature, possibly on the doping of the Si, and to some extent on some small amounts of other chemicals that might be added to the mixture. | ||||||||||||||||||
|
Some data for two typical mixtures are presented below. | ||||||||||||||||||
| |||||||||||||||||||
|
CP etches attack about everything - don't store them in a glass bottle, it will dissolve too! | |
|
Only some polymers, most notably PVC and Teflon, are CP prove | |
|
They generally work by oxidizing the Si (thats what the is doing) and dissolving the oxide (the job of the HF). | |
|
HAc is mostly just for diluting the mixture | |
|
However, HAc is also.... | |
|
A big problem with the CP etches is that they also dissolve all possible masks - usually SiO2 or Si3N4 layers - so they cannot be used for selective etching | |
|
At best, Si3N4 may last for some time - if you hurry up selective etching becomes possible in a confined way. | |
| |
For a strong imbalance between HNO3 and HF, the etchant may change its character: | |
|
Most notably (and not very suprisingly), the etch rate comes way down | |
|
More surprisingly (but not really, if you think about it): it may now be a defect etch, i.e. it attacks Si much faster at the place of defects. | |
|
While only HF dissolves SiO2 , all strong oxidizing agents can oxidize Si. It thus is possible to replace the HNO3 by some other oxidant. Essentially, two oxidizers are used: | |
|
First, H2SO4 can be used instead of HNO3, typically in a ratio HF : H2SO4 : HAc = 1 : 1: 5 | |
|
|
While inferior in over-all "quality" to the CP etches, it does not attack Si3N4 masks very strongly and thus can be used for selective (isoptopic) etching. | |
|
Etch rates are around (2...5) µm/min, again depending on someother factors too. | |
|
Second, CrO3 is used, a relatively weak oxidizer for Si. It only works on "soft spots", i.e. at surface areas were the bonds are weakened because of defects. | |
|
Hf + CrO3 + HAc + many other chemicals (with no clear role) is the base of most defect etchants - a very important techique in semiconductor development | |
 |
 |
 |
 |
© H. Föll (Semiconductors - Script)