The Mystery of Photoresist Ridges: A Question About Light and its Behavior

[James T Clemens]

Let me start off with a reference: http://www.lithoguru.com/scientist/lithobasics.html This is an article about the field of lithography and microlithography. Before anyone dismisses the importance of this technology , let me point out that it is used to make EVERY integrated circuit and EVERY circuit board that is used in EVERY computer used worldwide. It is a fundamental optical - chemistry technology of the computer age. Now let me post SEM (scanning electron micrograph) of patterned photoresist material ,

called photoresist. Photoresist is a very general term for a whole class of polymers, but it means a polymer that is light sensitive. The light exposure restructures the polymer and allows it to either be dissolved in a solvent or become non-soluble in a solvent. The resolution of photoresist is on the order of 1-10 nanometers. I won't go into the process of integrated microcircuit fabrication, my question has to do with the ridges that appear on the sidewalls of the photoresist material. This effect is well known to those who work in the field. The effect is very pronounced when the substrate is metallic, a strong reflector of light.
It has been determined (general consensus) that this pattern is caused by the exposure of the photoresist by the standing waves (constructive and destructive addition of the E field) of light used to expose the photosensitive polymer ( photoresist ). Classical optics, based on EM field theory, is used to describe this effect. In fact, the problem is easily solved by coating the substrate with an absorbing material, so there is no reflection. If you read the referenced paper, that layer is called BARC ( Bottom Anti-Reflection Coating).
Now while seems like a very simple optics issue, I, personally are very disturbed by it. Let me explain why and then whoever reads this can comment on my thinking.
In classical E-M theory one will obtain these standing waves when they solve for the EM fields at the boundary of a conducting plane or surface. The solution is obtained by assuming that the EM fields go to zero at the boundary. But if you think about this approach, then one is defining the entire system and boundary conditions at time = zero. That's a little troublesome. In the experimental setup that is used to expose the photoresist, There is an incoherent ( sodium vapor lamp) or coherent (exicmer laser) light source. When an excited atom releases a photon at time = 0, it does not know how far the optical path will be (that includes air and the lenses in the system). These systems are very complex with lenses consisting of tens of elements and different atmospheres ( some resists work better with no oxygen present and others with no water in the atmosphere). And in the incoherent system, the light source is usually a mercury vapor source and the atoms are emitting photons at random times.
So how does one explain the observed phenomena in either classical optics or quantum optics.
As was mentioned before, in classical optics the "solution" is arrived at by setting of the boundary conditions at time = 0, but in classical optics we know that the EM wave has to travel a finite distance, when the exposure takes places. The concept of using t=0 and deriving a steady state solution is troubling, when the experiment is not conducted that way.
In a quantum mechanical view, single photons are released randomly in time. We seem to understand a lot about photons, but, at least to me, there are some big gapping holes in our fundamental understanding. Photons are bosons, with spin 1 , they also given a frequency and wavelength, which is dependent on their energy. In the optical and near ultraviolet regions, the wave lengths are between 0.25 to 0.80 microns, which is about 10^3 times larger than the atom that releases them. The above photograph is based upon a mercury source. And the ridge spacing clearly corresponds with an interference pattern of the wavelength.
Now if I am not mistaken, photons do not interact with each other, so that seems to imply that each photon has to create its own interference pattern for energy absorption within the photoresist material and that energy is very puzzling. Is it absorbed at a single chemical bond and breaks it, or is it absorbed within a range of the photoresist and then causes one bond to break, since it cannot break many bonds.
The point I am trying to get at is that there is a lot that we do not understand (maybe just me) about light and how it behaves. We have based our models on macroscopic results of experiements. When I think about the subtle points as mentioned above, I realize that there is a lot we may not understand. Ror example - Is this simple exposure pattern that we see consistent with entanglement?

 

[Spinnor]

If this is a common effect can you point me to similar images or suggest a good Google search for them?

 

[Nugatory]

As was mentioned before, in classical optics the "solution" is arrived at by setting of the boundary conditions at time = 0... The concept of using t=0 and deriving a steady state solution is troubling, when the experiment is not conducted that way.

We'd have to look at the actual solution, math and all, to be able comment sensibly. But as a general principle, standing wave solutions are often determined primarily by the spatial geometry of the problem. That's not changing with time in this problem, so the $t=0$ thing may be a red herring; either the solution may be exact anyway, or it may be an approximation adequate to use here.

In a quantum mechanical view, single photons are released randomly in time.

The first thing to ask is whether the quantum mechanical view is even relevant here. This depends on the number of photons "in the air" at any given moment, and you can calculate that from the intensity of the light, the energy per photon, and the known relationship between frequency and photon energy. If that number is large, you use the classical instead of the quantum model (for about the same reason that the people who design pumps treat water as a fluid obeying the laws of hydrodynamics instead of as a collection of water molecules each individually obeying Newton's laws).

 

[James T Clemens]

If this is a common effect can you point me to similar images or suggest a good Google search for them?

Yes you can Google "standing wave patterns in photoresist" and hundreds of sites will appear. What I suggest you do is Goggle the term and then switch to image mode, there you will find many examples. (SEM micrographs). Here is one such page that illustrates the issue: http://spie.org/newsroom/4198-in-si...terization-of-photoresists-during-development Now if you read my first post, I said that EVERY integrated circuit and circuit board in Every computer in the world is fabricated using photolithography technology. And how do I know this? Well I was the Director of Research of Lithographic Processes and Systems at AT&T Bell Laboratories for 6 years. That was 1984 through1989. I was also the co developer of exicmer laser projection lithography at the Labs and we were the first research and development team to use krypton fluoride lasers (248nm) as light sources. But my formal training is a theoretical nuclear physicist. I hope that these discussions get deeper and more individuals get involved, since there is some wonderful physics problems that are seen everyday in the manufacturing world and never get the attention of the research community. After my PhD, I worked in the Allentown Western Electric Factory researching materials, transistor and device physics and processing techniques used to fabricate silicon integrated circuits. After fifteen years, I was promoted and reassigned to the Murray Hill Central Research Laboratory, where I was the manager of all advanced lithographic development. You can find me by Google
using "James T. Clemens, PhD ".

 

[James T Clemens]

We'd have to look at the actual solution, math and all, to be able comment sensibly. But as a general principle, standing wave solutions are often determined primarily by the spatial geometry of the problem. That's not changing with time in this problem, so the $t=0$ thing may be a red herring; either the solution may be exact anyway, or it may be an approximation adequate to use here.

The first thing to ask is whether the quantum mechanical view is even relevant here. This depends on the number of photons "in the air" at any given moment, and you can calculate that from the intensity of the light, the energy per photon, and the known relationship between frequency and photon energy. If that number is large, you use the classical instead of the quantum model (for about the same reason that the people who design pumps treat water as a fluid obeying the laws of hydrodynamics instead of as a collection of water molecules each individually obeying Newton's laws).

First, the quantum mechanical view is extremely important here. Why because the light sources are not black body radiators. We are dealing with essentially monochromatic light that is emitted by atoms - electrons changing well defined states. Thus we cannot shut out the discussion of the photon. Respectfully, this is a Physics Forum and not and engineering forum, we attempt to bring understanding to the observed phenomena. I know that classical optics will give rise to interference patterns. Whether a math model yields approximately correct answers, does not shed any understanding on the physical processes involved.

With all due respect, I would bet [ :>) ] that if we replaced the filtered mercury light source, with filtered star light, where photons have been traveling for thousands of years, even before photolithography had been developed, we would still get standing waves in the photoresist material. These starlight photons are incoherent and had no way to sense (in the broadest definition of sense) that they would arrive at the apparatus and create standing wave patterns.

I know that this is a very deep question I have raised, because we have not addressed how the photons travel through the lenses, and the photoresist material.
We may have to eventually get into areas of Quantum Chromodynamics, and possibly other fields. But we still don't have a clear understanding of the poton and how it transmits energy. Is the energy spread out over the wavelength(s) extent of the photon of is it contained in a single small section of the spatially extended photon.

 

[Spinnor]

Your search suggestion yielded this,

https://people.rit.edu/deeemc/courses/0305-676/swingcurves/standing_waves.pdf

which seems to account for the effect using only classical physics. But we are told everything is quantum, classical physics is only the proper limit of quantum physics?

Are the ridges a problem?

 

[James T Clemens]

Your search suggestion yielded this,

https://people.rit.edu/deeemc/courses/0305-676/swingcurves/standing_waves.pdf

which seems to account for the effect using only classical physics. But we are told everything is quantum, classical physics is only the proper limit of quantum physics?

Are the ridges a problem?

The ridges can be eliminated by using a underlying coat of material called BARC (Bottom - Anti -reflective Coating) which is applied under the photoresist material.

However, that does not answer the basic physics questions that I have raised. That question is how do we explain these ridges in the material.

Now I will admit that this seemingly simple phenomena is dismissed with classical EM theory, however, that theory does not apply.

Lithography is a multibillion industry from photosensitive chemistry, industrial coating machines, expensive exposure machines (10+ million $ per machine),
to critically controlling pattern integrity. This is a whole field of what one would call applied physics, but I have the feeling that it extends well beyond that limited and overlooked field, and is not what we call "basic research physics" or fundamental physics. Imagine where we would be if this field was not developed by applied physicists and engineers. The world of electronics, from the internet to wifi and everything else that we are now implementing would not exist - period.
So I am looking for somenone one to explain what is the basic physics of tis situation, and what we really know about photon behavior.

 

[Spinnor]

[QUOTE="James T Clemens, post: 5639904, member: 610655"...

Now I will admit that this seemingly simple phenomena is dismissed with classical EM theory, however, that theory does not apply.

...[/QUOTE]

Can you please explain, maybe again, why you think classical theory does not apply?

 

[PeterDonis]

The OP can't respond, so this thread is closed.