Other: Metal Optics (1976)



The Introduction of Metal Optics

Question asked in 1976: Would metal optics be beneficial for use in infrared systems, and, if so, how do you make such optics (mirrors)?

An example of metal optics
An example of metal optics

Specific advantages of metal optics and mirror systems

  1. Completely achromatic; apart from wavelength-dependent diffraction, the mirror does the same for all wavelengths. For very short wavelengths (ultraviolet) the reflection coefficient may decrease.
  2. The image errors of spherical mirrors are smaller than those of any refracting spherical elements. Sometimes, a small refractive correction component may be needed to improve quality.
  3. Large mirrors are cheaper than large refractive components. This especially applies to expensive infrared materials.
  4. Mirror systems are shorter in construction. On the other hand, in many cases, part of the pupil has been removed due to obscuration.
  5. Metal mirrors are easier to install and their expansion coefficients are better adapted to the other construction.
  6. With intense beams, heat can be dissipated more easily. This plays an important role, especially with CO2 lasers.
  7. Metal-turned or milled mirrors can be manufactured quickly. While a glass mirror takes 40 hours, a rotated mirror can take approximately 8 hours. No further tasting glasses are needed.
  8. It is easy to give any desired rotationally symmetrical shape to the mirrors. Parabolas, hyperbolas, etc. can be made as easily as spherical surfaces. This is optics. Important in the manufacture of error-free optics and special optics.
  9. The quality of turned mirrors, especially in the infrared, is better adjusted in terms of precision than the hand-polished mirrors which are often far too good. Hairs and pits matter relatively little.

General consideration

We can weigh four manufacturing methods of metal mirror optics:

  1. hand polishing of hard layer of chrome on aluminium; checking with tasting glass, etc.
  2. rough roughening with the machine; polish by hand with diamond paste.
  3. replica methods and pressing methods, such as flashlight reflectors, car headlights, etc.
  4. diamond turning and milling.

The method (a) is laborious and time-consuming. Method (b) only works with plane mirrors to reasonable precision, but not top accuracy (0.5 mrad image error). Method (c) is qualitatively an order too poor for imaging systems. For detection systems, it would still work here and there. Diamond-turning provides high precision (with a good lathe) just like method (g) for all shapes and is a fast manufacturing method.
The problems during implementation focus on the following components:

  1. A good bearing of the main spindle on which the workpiece is clamped and rotates.
  2. A good diamond chisel.
  3. A good measuring system to measure the displacement of the slides or the workpiece.
  4. A control system to make corrections for the “swaying” of the X-Y carriages during movement.

We are dependent on others for the availability of the spindle, the diamond chisel and the slides. Suitable measuring methods cannot be purchased, but with some research, you can build them yourself sufficiently accurately and in a sufficiently short time for relatively low costs. The principle of these measuring methods is interferometric and a laser is used as the light source.

Mirror accuracy

The accuracy of the mirror involves two aspects:

  1. The slow fluctuations over the surface

    If the depth d extends over the distance s, then the resulting angular deviation of the parallel light beam on that part of the plane is 2d/s radians. For example, if s is 10 mm and d = 0.5 µm, the angular deviation is 0.1 mrad. For infrared optics, this is already very good, even the 0.2 mrad total error that occurs if, over the next 10 mm, the deviation goes 0.5 µm in the other direction.

  2. The ripples and rapid fluctuations

    A diamond-turned surface consists of hollow grooves with peaks and valleys, the dimensions of which are determined by the size of the chisel tip and the start. For example, if the infeed is 1 µm and the chisel roundness is 20 µm, then the difference between trough and top is less than 0.01 µm and negligible. If the infeed is 4 µm, then the difference is 0.1 µm and begins to approach the limit of what is acceptable. The wavelength is 0.5 µm for visible light and 3-10 µm for infrared light.

Scattering by surface ridges is approximately 104 times smaller in the infrared region than in the visible region (Rayleigh scattering proportional to λ-4). Sudden buds can therefore be neglected provided they are not too large.
The conclusion is that the variations of the tool position concerning the ideal position should not exceed 0.1 µm over 2 mm if the desired precision is 0.1 mrad. Scratches, nicks, pits and the like are of relatively little concern, especially as the wavelength increases and their size is less than a few tenths of a µm.


Numerous optical systems exist in the field of electro-optical measures, countermeasures and counter-countermeasures. These may include missile warheads, fuzes, lasers, detection equipment, search systems, warning sensors, observation equipment and all this on land, at sea and in the air.
In many cases the system must be cheap, often have a large field of view, be robust, bright, as small as possible, have a long lifespan, etc. This means parabolic secondary mirrors with hyperbolas as secondary mirrors. Or toroidal surfaces for the warning sensors.
Multi-faceted surfaces are also common in “forward-looking infrared” systems and also in surveillance and warning equipment. Here and there one encounters cylinders, mirror drums, etc., in observation equipment.
Systems with passive observation at long distances require large secondary mirrors (30 to 40 cm). Diamond-turned mirrors would enormously reduce the costs of such mirrors (the cost of a “current” parabola of 40 cm, f/1.5 is approximately Dfl. 100,000).
Parabolas with large diameters are also desirable for test equipment but have hitherto been very expensive. The best [example] here is an off-axis parabola (collimators).
Finally, mention should be made of its use in infrared line scanners containing rapidly rotating mirror blocks, as well as in some measuring devices and cameras.

Manufacturing metal optics

Two machines can be considered for making the various desired optical surfaces, each of which has its advantages and limitations. These machines are a diamond milling machine and a numerically controlled diamond lathe respectively.

The diamond milling machine

The machine is currently being built in the Instrument Factory and is in an advanced stage of completion. The machine is intended for making flat and slightly cylindrical mirrors, both round and rectangular. The dimensions of the mirrors are limited to approximately 200 x 150 mm. The machine is not equipped with numerical control. The manufacturing costs of this machine are low compared to the diamond lathe.

The numerically controlled diamond lathe

This machine must be able to produce flat, concave, convex, spherical and parabolic surfaces and combinations thereof. The shape of the mirrors is round, with a diameter of up to 400 to 500 mm. The production of such a machine will require a major effort from several groups and departments in the laboratory. The construction of this machine will be expensive, both in materials and man hours.

The method of construction

There are currently only ideas for the construction of a numerically controlled diamond lathe that certainly deserve further consideration. This further consideration will require a lot of money and effort. The construction of such a machine can therefore only be regarded as a research project. Based on this research, results will certainly be obtained that will be important for the laboratory; however, the final result is uncertain. Based on current knowledge, no one can give any guarantee of a satisfacting project result.

The phasing of construction

Initially, a good measuring system must be developed. Such a system is not commercially available. However, based on some preliminary experiments, the Far-Infrared Research Group has good hope that an accurate measuring system (0.1 µm) will be available. Further development of the measuring system has yet to be started. Perhaps two H.T.S. students, who will soon graduate from the laboratory, can be used for this. Such a suggestion will be submitted to [the management]. The actual construction will then begin, without using the computer as control for the time being. This will have to be achieved in the final phase.


Some machine components must be purchased for construction; Furthermore, depending on the stage of construction, the cooperation of multiple groups and departments must be obtained.
The list of “requirements” can be specified as follows:

  1. Flat table
  2. Hydrostatic bearing spindle
  3. Oil pressure generator
  4. A drive motor (direct current)
  5. A good quality X-Y carriages
  6. The calculation of the hydraulic drive
  7. The manufacture of the hydraulic drive
  8. The unit for the hydraulic drive
  9. Research on the measuring system (optical and electronic)
  10. The production of the measuring systems
  11. The mini or microcomputer
  12. The interface equipment
  13. The actual calculation program
  14. The connection of the computer to the machine
  15. Materials miscellaneous
Dfl.  4,000.-
“ 16,000.-
“   8,000.-

“  7,000.-

“  8,000 .-
“   4,000.-

“ 12,000.-

“ 10,000.-

[…]  There is only a chance of a well-functioning result if we can count on sufficient financial resources and the cooperation of the groups and departments mentioned. If this cannot be achieved: don’t start!

The final result: the Paganini

The research project was initiated at the end of 1976. The final project result was the Paganini metal optics workbench for making metal optics for infrared and Lidar optics in particular.

Paganini metal optics bench
Paganini metal optics workbench


Paganini metal optics bench
Paganini metal optics workbench


 An example of produced metal mirrors
An example of produced metal mirrors


Source: An internal report from October 1976