Asphere Fabrication Goes Commercial, Part I: Grinding and Polishing
By: Charles Klinger, OptiPro
Aspheric optics offer designers a number of advantages over conventional spherical optics. In many cases multiple surfaces can be replaced by a single aspheric surface, reducing the number of elements, weight, assembly complexity, and overall size of optical systems. Despite the power of aspheric components, optical engineers have been slow to use them. The complexity of the optical prescription has historically made aspheres costly and difficult to produce, so the production of quality aspheres has traditionally been the niche of a few companies.
This situation is rapidly changing as optical equipment manufacturers bring automated or specialty equipment to market. To understand the significance of the changes taking place, it is necessary to have a general understanding of how aspheres have been traditionally produced and the very specific challenges that are faced by those who attempt the task.
Grinding
What actually takes place during the grinding process? Grinding is a very effective technique for producing shapes in a glass blank. Although it will create the basic optic prescription or desired form, grinding will not produce a specular high-quality optical surface. The process of shaping a hard, brittle material like glass is more like controlled cracking than the conventional machining performed on metals. When metal is machined, a sharp tool penetrates the material and cuts away a layer of metal. In glass, diamond bound onto or embedded into a tool presses against the surface, causing a micro-indentation. After the diamond has been removed, cracks propagate in the surface, releasing a chip. Cracks and microfissures remain in the material left behind, however; these cracks are known as subsurface damage (SSD).
This grinding operation takes place on a computer-controlled platform using a series of diamond-impregnated tools with successively finer diamond particles. The finer the diamond, the smaller the chips released from the optic, and the shallower the cracks. Subsurface damage becomes progressively more shallow with each successive level of grinding, leaving a smoother surface. After the last of the diamond tools, the surface is "polish ready", though it will still exhibit some irregularities and predictable amounts of subsurface damage.
Grinding spherical optics is relatively straightforward process, but aspheric surfaces are far more difficult. If a convex and a concave spherical surface are rubbed together, both will become more spherical. In an aspheric surface, the radius of curvature varies as a function of axial height. Thus, small tools must be used, and dwell times must vary to create the desired optical prescription.
Polishing
Unlike grinding, which is purely mechanical and relatively deterministic, glass polishing is chemo-mechanical process. Cerium oxide particles suspended in a water-based medium are rubbed across the ground surface with a pad or tool. The cerium oxide causes a chemical change in the glass, softening the surface layer. The mechanical rubbing action of the pad removes the softened layer from any asperities or projections above the local surface, reducing the height of the feature. Repeated applications of the pad remove more and more material.
The objective of the polishing process is twofold:
- Remove enough glass to get below the subsurface damage into the virgin, undisturbed material
- Leave a surface that matches the optical prescription to within the tolerance specified by the customer
Optical polishing has traditionally been considered more of an art than a science. The polishing pad must be rubbed over the entire surface of the lens in such a manner that the correct figure is produced while preventing unwanted local deviations. In the case of a spherical optical surface, the local curve is the same throughout the entire surface, which allows a skilled operator to use a pad that is larger than the optic and sweep it across the part.
Although in concept, this is easy, in practice it is not. Conventional machines require much skill from the operator. When trying to obtain optical surfaces that are within l/20 regularity, small effects can have major influences.
The process becomes further complicated with aspheres. Because the local curvature varies as a function of radial position, opticians cannot use oversized pads. An oversized pad would always contact on the high zones and force them to be lower, even if that was not desired. A sub-aperture pad, smaller than the lens in production, must be used. The tool must also be sufficiently flexible to follow the local curvature and yet still be able to blend to all areas. The challenge is to achieve a polished surface that matches the desired asphere to within the tolerances required by the specification.
Measurement
In optics there is a well known phrase that goes something like this: "You can't make it if you can't measure it." To make aspheres, opticians have to measure them, and asphere metrology is not trivial. Whereas spherical surfaces can be measured readily with test plates or digitizing interferometry, in general aspheres cannot. Some mild aspheres have a small enough departure from spherical that careful analysis and examination of interferograms can yield acceptable results.
For others having more departure, null lenses can be fabricated and used as a testing device. These have been found to be very acceptable but the high cost of fabrication and time involved virtually eliminate their use for rapid prototype or limited-production aspheres. Often manufacturing of the aspheric null lens is as complex as fabrication of production optic. Currently, the only option for rapid prototype or limited production aspheres is the use of a device that actually measures the profile of the lens. The measured shape can be quickly compared to that of the theoretical shape and a chart of errors produced.
In part II of this article, Charles Klinger will describe how aspheres are produced with computer-automated equipment.
About the author…
Charles Klinger is a senior applications engineer at OptiPro Systems, 6368 Dean Parkway, Ontario, NY 144519. Phone: 716-265-0160; fax: 716-265-9416.