Effect of atmospheric pressure on optics
Moderators: rjlittlefield, ChrisR, Chris S., Pau
Effect of atmospheric pressure on optics
I was studying patents related to stepper lenses today, and finally understood why many of the most modern stepper lenses have gas hose attachments. It seems that changes in atmospheric pressure are enough to wreck the precision of these amazing lenses. The effect is huge! Maybe this could affect our own work under unusual circumstances, so I thought I'd share it with you.
The patent suggests to fill a stepper lens (typically m=5 or m=10 when reversed) with helium because:
"... a barometric pressure change of 50 mm of mercury, i.e., from 710 mm to 760 mm, results in a back focus change of -5.97 micrometers, whereas the change with helium filling in the lens is only -1.06 micrometers. Similarly, the change in magnification referenced to lateral movement of a reference point at 11 mm from the objective, is reduced from 0.42 micrometers to 0.15 micrometers."
The patent suggests to fill a stepper lens (typically m=5 or m=10 when reversed) with helium because:
"... a barometric pressure change of 50 mm of mercury, i.e., from 710 mm to 760 mm, results in a back focus change of -5.97 micrometers, whereas the change with helium filling in the lens is only -1.06 micrometers. Similarly, the change in magnification referenced to lateral movement of a reference point at 11 mm from the objective, is reduced from 0.42 micrometers to 0.15 micrometers."
- rjlittlefield
- Site Admin
- Posts: 23626
- Joined: Tue Aug 01, 2006 8:34 am
- Location: Richland, Washington State, USA
- Contact:
Thanks for raising this issue. I agree that it's very interesting, but I feel compelled to inject a note of calm.
I believe what's being discussed is that as atmospheric pressure changes, so does the density of a gas, the refractive index of that gas, and thus the total refractive power of gas gaps between lens elements. The changes are proportional to the difference between the gas refractive index and 1.0. The R.I. of helium is a lot closer to 1.0 than air is (about 35 versus 292 parts per million, using values from http://www.kayelaby.npl.co.uk/general_p ... 2_5_7.html ). So, the deviations in lens behavior versus pressure will be a lot smaller with helium also. (Also of interest: the ratio of changes in delta magnification is different from delta focus is different from delta R.I. I assume that somehow has to do with details of the lens design, but I don't have a clue beyond that.)
The numbers quoted in the patent would indeed be huge for the intended purpose of making integrated circuits, where layers must precisely align from one day to the next. The change in back focus distance could be corrected easily by dynamic focusing, but correcting for an image size change of 0.42 micrometer (420 nm) or even 0.15 micrometer would be more challenging.
On the other hand, a change of 0.42 micrometers in 11 mm is less than 1 part in 25,000 change in magnification.
I'm hard pressed to think of any applications other than chip making where a change of that magnitude would have much ill effect.
--Rik
I believe what's being discussed is that as atmospheric pressure changes, so does the density of a gas, the refractive index of that gas, and thus the total refractive power of gas gaps between lens elements. The changes are proportional to the difference between the gas refractive index and 1.0. The R.I. of helium is a lot closer to 1.0 than air is (about 35 versus 292 parts per million, using values from http://www.kayelaby.npl.co.uk/general_p ... 2_5_7.html ). So, the deviations in lens behavior versus pressure will be a lot smaller with helium also. (Also of interest: the ratio of changes in delta magnification is different from delta focus is different from delta R.I. I assume that somehow has to do with details of the lens design, but I don't have a clue beyond that.)
The numbers quoted in the patent would indeed be huge for the intended purpose of making integrated circuits, where layers must precisely align from one day to the next. The change in back focus distance could be corrected easily by dynamic focusing, but correcting for an image size change of 0.42 micrometer (420 nm) or even 0.15 micrometer would be more challenging.
On the other hand, a change of 0.42 micrometers in 11 mm is less than 1 part in 25,000 change in magnification.
I'm hard pressed to think of any applications other than chip making where a change of that magnitude would have much ill effect.
--Rik
- rjlittlefield
- Site Admin
- Posts: 23626
- Joined: Tue Aug 01, 2006 8:34 am
- Location: Richland, Washington State, USA
- Contact:
This lens is typically used to project a large pattern mask onto a smaller wafer covered with a photoresist. So I imagine the back-focus refers to the wafer side.
Here is the link:
https://www.google.com/patents/US4616908
Note the lens diagram. From that diagram, with its asymmetric rays, can you tell whether this is a 1/5x or 1/10x lens?
Here is the link:
https://www.google.com/patents/US4616908
Note the lens diagram. From that diagram, with its asymmetric rays, can you tell whether this is a 1/5x or 1/10x lens?
The semiconductor technology from chip design to chip manufacture is an amazing journey thru science like no other field. Pushing the boundaries (sometimes surpassing them) of physics, chemistry, electronics, CAD and robotics at every corner!
And we (USA) are letting this technology slip thru our fingers daily, but that's another story!
Best,
Mike
And we (USA) are letting this technology slip thru our fingers daily, but that's another story!
Best,
Mike
It really is fascinating. There are lenses with resolutions of a few nanometers over a field size of tens of millimeters!! The lenses have to be polished to an accuracy of three or four atom-widths!!! Each chip-making machine costs many millions of dollars.
Mike, if the lenses are hermetically sealed, maybe the pressure difference between inside and outside could deform the lens elements by a few atom-widths, spoiling the resolution.
Mike, if the lenses are hermetically sealed, maybe the pressure difference between inside and outside could deform the lens elements by a few atom-widths, spoiling the resolution.
Lou,
With todays technology a single sodium atom can cause a chip to malfunction!
A new fab can cost upwards of $15B and the chip processing machines cost many 100s of millions each. It takes a bunch of these machines to create a line to fabricate chips!
We are routinely dealing with 14nm now, soon 10nm and 7nm is in development.
Chip development cost can be as high as $100M and are continually rising. Intel spends way more than this on their processor development I've been told.
However this technology isn't decades away, or years away. It's in your iPhone and iPad now, and it's affordable to the average person!!!
Amazing indeed!!
And yes those lenses are incredible!!
Best,
Mike
With todays technology a single sodium atom can cause a chip to malfunction!
A new fab can cost upwards of $15B and the chip processing machines cost many 100s of millions each. It takes a bunch of these machines to create a line to fabricate chips!
We are routinely dealing with 14nm now, soon 10nm and 7nm is in development.
Chip development cost can be as high as $100M and are continually rising. Intel spends way more than this on their processor development I've been told.
However this technology isn't decades away, or years away. It's in your iPhone and iPad now, and it's affordable to the average person!!!
Amazing indeed!!
And yes those lenses are incredible!!
Best,
Mike
Pau, SEM and x-rays are not easier for a home operation. SEM is expensive, complex, and specimen preparation is challenging and semi-destructive for many soft biological tissues. X rays are also difficult.
UV is a simple extension of normal optical tools that are easy to do at home, and very cheap compared to a SEM set-up, and many of us already do UV photography. Lenses are often less costly than normal macro lenses, though cameras need some modifications. These too can now be had for around $1000 compared to many tens of thousands for SEMS. Extreme UV does require complex light sources, but most of the microlithography lenses on the market are optimized for visible light or near UV.
UV is a simple extension of normal optical tools that are easy to do at home, and very cheap compared to a SEM set-up, and many of us already do UV photography. Lenses are often less costly than normal macro lenses, though cameras need some modifications. These too can now be had for around $1000 compared to many tens of thousands for SEMS. Extreme UV does require complex light sources, but most of the microlithography lenses on the market are optimized for visible light or near UV.
Lou, as the thread is about chip making extreme optics that use very short wavelengths, and neither the optics or sources are trivial, I just wanted to point alternatives.
SEM is not difficult IF you have the right equipment, in special the ion sputter, I've worked few times with it at the University. You can buy an old used SEM at Ebay, although... I concur, I wouldn't put one at home.
There are even tabletop small low resolutions systems able to operate without high vacuum
SEM is not difficult IF you have the right equipment, in special the ion sputter, I've worked few times with it at the University. You can buy an old used SEM at Ebay, although... I concur, I wouldn't put one at home.
There are even tabletop small low resolutions systems able to operate without high vacuum
Pau
Pau, yes, modern chip-making optics are crazy, even worse than SEMs. Even a few years ago, microlithography lenses had reached the 500kg weight range!! But the lens mentioned in the patent referred to in my post is relatively small and is available on ebay right now for under $400. The light source might cost a few dollars more but not much. It is eminently accessible.
Lou
Lou