Some highlights of my 48 year (so far) career in optical design.

1. My 48 years of optical design – some highlights
Dave Shafer
David Shafer Optical Design

2. As a young boy I was always
fascinated by magnifying glasses
Optics is kind of like magic

3. Some kinds of flashlight
bulbs have a very small glass
lens on their tips.
I used to carefully break the end off
with a hammer and use the tiny lens as
a high power magnifying glass – about
50X magnification

4. I also made water
drop microscopes. A
small drop of water
can very easily give
100X
magnification, but it
has to be held up
extremely close to
your eye for you to
see through it.
The first single lens
microscope, from 300 years
ago, had a tiny glass lens
and was about 250X, but a
water drop works well too.

5. I lived on a small
farm, until I went to
college. We had
5,000 chickens

6. We also had one cow, and I did not drink
pasteurized milk until I went to college.

7. Our farm was
very far from
city lights and
the night skies
were very dark –
perfect for
astronomy.
Many people
have never seen
a really dark sky.

8. When I was 13
years old I got a
mail-order kit for
grinding and
polishing a 150
mm aperture
telescope mirror.

9. I bought a small
star spectroscope
(100 mm long) and
drew charts of the
solar spectrum, with
its many absorption
lines. Now, over 50
years later, that exact
same spectroscope
costs about 10X more
money.

10. I was hooked on optics! When I was 16 years old
I got Conrady’s two books on lens design.

11. I also got a book that was
full of complicated diagrams
like this one. It made optics
look pretty difficult.

12. When I was in high school there were no personal computers yet
and no large main frame computers that were available to the
general public. I traced a few light rays through an achromatic
doublet lens, with trigonometric ray tracing using tables of 6
decimal place logarithms. After you do that once you never want
to do it again! But I still knew that I wanted to be a lens designer.

13. When I went to college in
1961 big universities had a
main frame computer.
Data was input using
punched cards. At the
University of Rochester,
where I went, the Optics
department was able to use
this computer and students
like me were able to do
some simple lens design
problems.

14. At that time, in 1961, there were only two places
in the Western hemisphere where you could get an
undergraduate degree in optics - The University of
Rochester, where I went, and Imperial College in
London. In addition to learning about optics, I
also met my wife there at the University. We have
been married 47 years, have two children and five
grandchildren.
My son worked at ASML for 20 years and was a
vice-president there when he changed jobs a few
years ago. My daughter is a university professor
of Art History.

15. In the 1950’s electro-mechanical calculators (electricity
powered the calculating gears) were used to do optical
raytracing. To trace one light ray through one optical surface
took about 3 minutes. In the early 1960’s true digital
computers (main frames) were developed and they could
trace one ray-surface per second. Today an ordinary PC can
trace about 30 million ray-surfaces per second.

16. Optimization of optical
systems requires matrix
inversion. Hand calculations or
electromechanical calculators in
the 1950’s did 2 X 2 matrix
inversions – two variables and
two aberrations. Very many of
them, in sequence. Today, with
my PC, I optimize complex
lithographic lenses with many
high-order aspherics. I optimize
several thousand rays using
about 100 variables and there is
an enormous matrix inversion –
in just a few seconds.

17. Although computers
have revolutionized
optical design, there is
still a big need for
creative thinking by
the designer, using
your own mental PC

18. Submarine with its periscope
above the water surface
My first job after college, in
1966, was at a small high-tech
company that did military
optics – mostly very high
resolution reconnaissance
cameras for the U-2 spy plane
and for early space satellite
cameras.
I worked on a top-secret
project there that was a new
way to detect Russian
submarines. Some years ago
this secret technology was
declassified and today you can
read all about it on the
internet.

19. In World War I and World War II submarines would be
found by looking for their periscopes sticking up above the
water. Sometimes the sun would reflect off the front surface
of the periscope optics, but there was also sun glint off of the
water waves and it was very hard to tell them apart.

20. From an airplane the water
wake left by the moving
periscope could be seen.
But if the submarine
was moving slowly or
not at all then the
wake was very hard to
see, like in this case
here.

21. What was needed was a
new and highly sensitive
way to spot submarine
periscopes, when they were
above the surface of the
water.
The solution was to use
optics and lasers in a new,
top-secret way.
This new technology was given, back in
1966, the code name “Optical Augmentation”
and it is still called that today. You can look it
up on the internet.

22. We all know
about red eye
from camera
flash photos.

23. Eye retina
The eye retina reflects back the
focused light and then it is collimated
by the eye lens. It can then travel long
distances without spreading very
much. That is why flash camera “red
eyes” are so bright, like this cat.

24. Near IR
laser beam
Water level
Periscope
optics
A low power near-IR laser beam
was sent out over the water
surface, from a ship, and scanned
around by 360 degrees. If there is a
periscope above the water then the
laser light goes down the periscope
optics tube and is focused on the eye
retina of the person who is looking
through the periscope. That light
then reflects off the retina, is
collimated by the eye’s lens, and
reverses its path back up the tube and
out. It travels back over the water to
the ship where the laser is located and
a very bright “red eye” can be seen.
Eye retina

25. The energy collection area of the periscope optics
is very much larger than that of the eye by itself, so
the retro-reflected signal is orders of magnitude
larger and gives a huge “red eye” effect.

26. You may find this hard to believe but
with this relatively simple technology a
submarine periscope can be detected
that is many kilometers away. The laser
used is near IR instead of a visible
wavelength so that the person looking
through the periscope will not know
that they have been detected.
This same technology can be used in other ways. Airplanes can
detect the eyes of soldiers looking through the sights of
camouflaged anti-aircraft guns. Film or a detector array at the
focus of a camera also reflects back light and that is then
collimated by the camera lens on the way back out. Hidden
cameras can be found this way. From the ground level a laser can
detect space satellite camera optics. A pulsed laser can actually
measure the distance to a hidden camera, telescope, or periscope.

27. Today you can buy several versions of
this declassified technology on the
internet for less than $100 and find
hidden cameras in your hotel room or
other places, especially those tiny pinhole sized cameras - like on cellphones.

28. In 1971 I changed jobs and worked for a company
that specialized in infra-red military optics. One
project was this ----
Early warning missile defense system
(Work I did in 1972, 42 years ago).

29. If a missile from behind
the earth comes over the
rim of the earth it will be
seen here by a satellite
against a black sky, but it
will be very close to an
extremely bright earth,
which gives an unwanted
signal that vastly exceeds
the missile’s infra-red heat
signal. But that is the easy
case. Much worse is when
the satellite is on the night
side and the missile is seen
against a sun-lit earth’s
limb.

30. With the sun behind the horizon, the earth’s limb is
ten orders of magnitude brighter than the missile’s
infra-red heat signal.

31. Astronomers use a special kind of telescope, a
coronoscope, to look at the sun’s corona. They need
to block out the light from the body of the sun and
just look at the sun’s edge. This is possible using a
“Lyot stop” and this very old technology was used in
missile defense satellite optics.
It can block out very bright light that is just
outside the field of view of the telescope and which
is being diffracted into that field of view. That
unwanted diffracted light can be many orders of
magnitude brighter than the dim signal that the
telescope wants to see, in its field of view.

32. Rim of aperture stop is source of diffracted light
Field of view rays
Light
from
earth
limb
Two confocal
parabolic mirrors
give well-corrected
afocal imagery
Diffracted light is focused unto second aperture stop
Lyot stop
principle
Second aperture stop is smaller than image of first
stop, and it blocks out-of-field diffracted light from earth
limb.

33. The use of the Lyot stop
principle, plus super-polished
optics, makes it possible to
reject almost all of the
extremely bright unwanted
signal from the sun and the
earth’s limb and to just see
the missile signal.
I worked on some space optics systems to make
accurate measurements of the earth limb signal profile,
as well as some wide angle reflective space-based
telescopes for reconnaissance.

34. Display shows temperature
as different colors.
I also designed optics for medical infra-red imaging
systems. The infra-red heat temperature map of a person’s
face or other parts of the body can often show different kinds
of illness, including cancer. There is no physical contact with
the patient, just infra-red optical imaging.

35. In 1975 I changed
companies again and
went to work for PerkinElmer Corp., a maker of
laboratory instruments.
They were just starting
to get into making some
lithographic equipment.
Their “Micralign” optical system made it possible to make 1.0u
circuit feature sizes on 75 mm diameter silicon wafers, using mercury
i-line light from a lamp. This was a 1.0 X magnification system. I
designed a next generation 5X system that was able to make .50u
feature sizes. The 5X magnification made the mask easier to make.

36. It is hard
for us to
imagine
how small
one micron
really is.
An ant holding 1.0 mm square chip, with tiny
circuit features. What plans does the ant have for this chip?

37. A guitar made the same size as a red blood
cell, using nanotechnology

38. nanotechnology

39. One
micron
30 years ago computer chips had circuit features one micron in size

40. Today’s chips have about .03u circuit features

41. In 1976 I also worked on early experiments in
Laser Fusion

42. Laser fusion, if it ever works, will be about as cost
effective a way to produce energy as it is to go to the moon
in order to get some sand for your children’s sandbox. It’s
main use, if it works, will probably be to test the physics of
new nuclear bomb designs. My work was in the very early
days of laser fusion, around 1976.

43. Conic mirror
Laser beam
Conic mirror
Target pellet
Laser beam
Highly aspheric lens
Very high power laser beams enter from opposite sides
and are focused onto the tiny target pellet.

44. Target pellet
filled with
tritium gas

45. Laser fusion ignition
at 100 million degrees

46. Conic mirror
Laser beam
Conic mirror
Target pellet
Laser beam
Highly aspheric lens
The highly aspheric lens was made of the highest possible
purity glass but it would still absorb enough of the very high
power laser energy so that it would often explode!

47. An identical ray
path is not shown
here for this side
of system
I thought of a new type of design where there are two reflections from the
mirrors instead of one, before focusing on the target pellet. The result is that
the focusing lens is much thinner, with very much less asphericity and it does
not explode. It is also much less expensive to make.

48. One of my first
patents, in
1977, was for
an unusual kind
of telescope
that only has
spherical
mirrors.
Many years later one of these unusual telescopes
was sent on the Cassini space craft to Saturn. Later
another one went to the asteroid Vesta, and it was
there just a few months ago, taking photographs.

49. This is the Cassini space
craft before being launched.
Another of my telescopes is
on a space mission to visit a
comet and fly up close to it.
Close up of
asteroid Vesta,
taken recently
from space with
my telescope.

50. My early design work
back then was done on
an Apple computer,
using the OSLO design
program.
In 1980 I started my own
one-person optical design
business. This was very
unusual, back in 1980 and
is still not very common
today in the USA. In most
other countries it is very
rare. It was possible for
me because I had lots of
business right away in
lithography optics
design, with some
companies like
Tropel, Ultra-Tech, and
Perkin-Elmer.

51. Salvador Dali
Spanish Surrealist artist
One very interesting short
project I had, in 1980, was
for the artist Salvador Dali.
51

52. Salvador Dali had managed to
paint a stereo pair of paintings,
which is an amazingly difficult
thing to do. He wanted a new
type of stereo viewer to go with
his unusual painting pair. The
paintings would be on a wall
and then a person would look at
them with a stereo viewer that
could be adjusted for the
viewer’s distance from the
paintings.

53. Deviating prism
wedges can make a
stereo viewer but they
have a lot of
dispersive color and
mapping distortion
and are not adjustable.
I realized that a different ray path
through a prism can have no color, no
distortion, and be adjustable.

54. The final viewer was just
two 45-90-45 degree
prisms with a flexible
hinge that joined them
along one prism edge.
They could be folded
up, when not being
used, into a larger size
triangle.
I did not have to go
anywhere near a computer
to do this design project!

55. Even a very simple optical element, like
a simple prism, can show some surprising
features. Suppose you place two photos
against a right-angle prism. Viewing
angle #1 sees the photo on the bottom of
the prism. Viewing angle #2 has total
internal reflection off the bottom side and
sees the other photo. Both images seem
to be in the same place – on or below the
bottom side of the prism.
In this different arrangement there are 4
different views. View # 1 is a photo that
you see directly. View #2 is a photo on the
back side of that first photo and you see it
by two internal reflections inside the prism.
View #3 sees a photo on the bottom, after
one internal reflection. View #4 is a photo
that is seen through the prism with no
reflections.

56. After I started my company in 1980 my optical design work has
included camera lenses, medical optics, telescopes, microscopes,
and many other systems. Since 1996 almost all of my work has
been lithographic designs for Zeiss, in Germany, and wafer
inspection designs for KLA-Tencor, in California.
A typical lithographic 4X stepper lens design, from 2004. It is .80
NA, 1000mm long, has 27 lenses and 3 aspherics. The 27 mm field
diameter on the fast speed end has distortion of about 1.0
nanometer, telecentricity of about 2 milliradians, and better than .005
waves r.m.s. over the field at .248u. More modern designs have more
aspherics and fewer lenses.

57. These
lithographic
stepper
lenses are
made by
Zeiss and
then put into
ASML chipmaking
machines.

58. These state of the art
stepper lenses cost about
$20 million each and
many hundreds have
been made by Zeiss and
sent to ASML. In 2006
I invented a new type of
design that combines
mirrors and lenses and it
is now the leading-edge
Zeiss product, making
today’s state of the art
computer chips.

59. Aspheric mirror
wafer
mask
Aspheric mirror
I have several patents on this new kind of lithographic system, that
combines lenses and mirrors. Many of the lenses are aspheric, to reduce the
amount of surfaces and glass volume. Some of these designs have 4 mirrors
and some have 2 mirrors. One important characteristic of these designs is
that there are two images inside the design, while conventional stepper lenses
have no images inside the design. These are immersion designs, with a thin
layer of water between the last lens surface and the silicon wafer that is being
exposed. The design being made today by Zeiss is 1.35 NA and works with
.193u laser light. They will not say, and I won’t either, if it looks like this
design here or one of my other patents.

60. Red blood
cells, 8u
across
With my latest version of this lens/mirror design and double-patterning
exposures it would be possible to write a 300 X 300 spot image onto an
area the size of single red blood cell – more than enough to etch a good
photo of yourself, or to write an office memo, onto that surface.

61. For some years I have been working for Zeiss on EUV
(X-ray) lithography, which will be the next generation of
lithography systems. This only uses mirrors.

62. The aspheric mirrors made for these high-performance
optical systems are aligned to a precision of a few
millionths of a millimeter (i.e. nanometers). Their surface
figure quality (admissible deviation from the exact
mathematically required surface) and the surface roughness
are approximately three or four times the diameter of a
hydrogen atom. (!!!!!!!)

63. wafer
All-silica broadband design
.266u through .800u
For KLA-Tencor I have developed new designs for wafer inspection
that cover an enormous spectral region with only a single glass type.
63

64. Prototype, made by Olympus, .90 NA, wavelength = .266u - .800u

65. Some inspection situations require a long working distance
Long working distance design for deep UV.
Color correction with all silica elements

66. Any questions?