Monday, July 28, 2014

Effect of Etching Process

Effects Of Etching Process

(Introduction and Effect in Width Variation)


Chapter 3: Manufacturing Effects and Their Modeling
3.1 3.2a 3.2b 3.3a 3.3b 3.3c 3.4
Introduction Effect Of Etching Process Effect Of Etching Process Chemical Mechanical Planarization Importance Of CMP process Dishing & Erosion (CMP effects) Lithography
3.5a 3.5b 3.5c 3.5d 3.5e 3.5f 3.5g
Metal Width Variation (Type:1-2) Metal Width Variation (Type3) Metal Width Variation (Type:4-5) Metal Width Variation (Type6) Metal Width Variation (Type7) Metal Width Variation (Type8) Metal Width Variation (Summary)

I have discussed in the last post about the manufacturing effects/defects which are responsible for the variations in the layout. Now I will refer the final outcome of that process and then we will discuss the following points one by one



  1. The source/Reason of these variations and
  2. How can we minimize these? What are the different steps/precautions we should follow, so that impact should be minimal?
  3. If we can’t ignore these variations, can we model these somehow and take care during designing the layout.

Note: We can’t ignore these variations. I have tried to explain the reason of this in last post also. So, only 2 questions are open now.

  1. Source of Variations and
  2. How to model them.

There are several reasons of these variations. I will explain them one by one as per the requirement and if possible I will try to list down in the last.

One of the major source of these variations are Etching process and here we are going to discuss about that process ( not in detail but upto a certain limit, so that you can easily figure out how and where it will impact the variation- even which parameter it will impact). And after that I will mention the way we can model different effected parameters.

Manufacturing process involves a series of formation and removal of layers. But these layers are not formed / manufactured directly on the wafer; these are first defined in a polymeric resist film and then transferred to the layer. This (Pattern Transfer) can be done by one of the 2 Techniques as described below (“Lift Off” and “Etching”).


For liftoff technique, pattern resist is formed first and followed by layer deposition. Later we remove the resist which also remove the portion of layer above the. Final outcome is the required Layer pattern. But this pattern is of round shape at the top (as shown in figure).  This rounding top is because of shadowing effect of Resist.

This is the one of disadvantage of lift off technique. Another disadvantage is the temperature limitation. This method is limited to temperature below 200-300 degreeC , at which point resist begins to degrade.

In the etching technique – first layer is deposited and then Resist is formed as per the layer pattern.  After this we etched the layer portion which is not required and Resist work as a masking layer for the underneath layer. Finally we strip the Resist above the Patterned layer.


Etch process is of 2 types.


Back

Isotropic Etch Process:

In the Isotropic etch process etchant attacks the layer surfaces equally in all the directions.  Most of the liquid etchant are isotropic. Since etchant attacks in all the direction (horizontal and vertical both direction), the resultant layer shape is not as per the expectation. Below figure help you to understand.


You can see that the resultant layer become narrow in size. When the layer width and spacing is significantly larger than the thickness of the layer, Isotropic results a minor problem but when layer width and spacing is comparable to thickness of the layer then plasma etching technique which is capable of anisotropic etching is used (I will discuss this later in post).

As per the etchant solution, if amount of narrowing the feature in horizontal direction can be figured out, we can use compensation method to make it as ideal as possible. In the mask compensation method, the mask size is increased from the desired size of layer (which increase the width of deposited Resist) in such a way that after etching and stripping of Resist, resultant layer width becomes equal to (or approximately equal to) required size. But again this method can’t be used if thickness of layer is comparable to width/size of layer.

Below diagram helps you to understand the final result of isotropic etching in more detail.



Now correlate this with the actual process which we have discussed in the previous part. I have just changed the color so that you can understand this clearly.




Back

Anisotropic Etching Process:


As I have mentioned above, when layer width and spacing is comparable to thickness of the layer then plasma etching technique which is capable of anisotropic etching is used.  You can understand this with the following diagram.


The degree of Anisotropy of an etch process can be expressed as:

A = 1 – Vh/Vv

Where

A = Degree of Anisotropy
Vh = horizontal Etch rate
Vv = Vertical Etch Rate

When A=0, it represent Isotropic etching and A=1 represent Anisotropic etching.

Remember, we always try to get A=1 but that’s the ideal scenario. :) In the above diagram, I have mentioned Directional Etch and that’s the thing we can get practically. :)

You might be confused but below figures can remove your confusion (I captured these figures from Internet and then just modified little bit)


From above you can see that in case of Isotropic – Etch rate is same in all the direction, Anisotropic – highly directional etching with different etch rate in different directions.

Now you may ask how this (isotropic and anisotropic) is going to affect differently our process, I will explain this again with the help of following fig (again from Internet).


Here you can see that metal “BIAS” is different in different cases.

In case of Isotropic, we don’t have control in the horizontal direction. It’s equal in both direction and that’s the reason I have mentioned previously that if width is larger than thickness – we can use Isotropic but when with is comparable to height, then if we use this process, then we will get very small (or almost zero) width of the layer (I hope you can imagine this).

In case of Anisotropic, we can control horizontal biasing/etching. And we always like those things which are in our control. J

From the Above discussion, It's clear that In the Width Variation of Metal or Dielectric - Etching has an Important role.

I hope this much is sufficient to build a base of etching process effects. But before I will write my final words, I would like to touch few more things which are related to Etch process just in the form of Summary.

What is ETCHING?

As the name suggest it’s the process of removing something. In the semiconductor world – It’s the process of chemically removing layers from the surface of wafer during manufacturing.

Few things we have to remember:

  • Every wafer has to undergo with several etching steps.
  • You can’t skip the Etching steps and none of the etching process is ideal.
  • Depends on the type of Etching process, you are using – type of defects are introduces in the chip and also depend the intensity of those defects.


There are 2 type of etch process (apart of Isotropic and Anisotropic – which we have discussed above):

  1. Wet Etch
    • Liquid Etchant
    • Chemical process Only
    • Advantage:
      • Low Cost and easy to implement
      • High etching rate
      • Good selectivity for most material.
    • Disadvantage:
      • Very hard to control critical feature Dimension
      • Eliminates Toxic fumes.
  2. Dry etch
    • Gas Phase etchants in a plasma
    • Chemical and physical (sputtering) process
    • Advantage:
      • Both isotropic and anisotropic process can be done.
      • High resolution and cleanliness
      • Better process control
      • Ease of automation.

In the next part we will discuss one more effect of Etching and then the way foundry model these variation’s effects.

4 comments:

  1. This comment has been removed by a blog administrator.

    ReplyDelete
  2. This comment has been removed by a blog administrator.

    ReplyDelete
  3. That is a fact about VLSI industry. Thanks for upload this information

    ReplyDelete
  4. I read your web journal every now and again and I just thought I'd say keep up the astounding work! J Walter Miller Co

    ReplyDelete