Wayne Pantley and David Collier
It is now common practice for systems manufacturers in a broad range of industries to outsource the production of entire finished printed-circuit-board assemblies; indeed, this trend is even being extended to optoelectronic subassemblies, where vendors are routinely being asked to integrate components such as detectors, amplifiers, and signal-processing electronics on a single board (see Laser Focus World, May 2000, p. 215). This situation is also beginning to include optomechanical assemblies for several reasons.
Technician uses an autocollimator alignment system to adjust an optomechanical assembly in a custom jig.
One reason is that laser technology is maturing, and laser-based products are becoming more turnkey. While this is especially true of newer laser types, such as diodes and diode-pumped solid-state lasers, even older technologies, including helium-neon, argon-ion, and excimer lasers, are achieving new levels of reliability and ease of use. As maturing laser technology reduces the need to employ a staff of experts in optics and optomechanical assembly, OEMs must then carefully examine if it makes sense to retain some capability in this area or if it is better to outsource the entire process.
Typically, the most important criteria used in this "make or buy" decision are cost and cycle time. Outsourcing can move assembly from being a fixed part of overhead costs to a variable expense.
Another important trend that is increasing the desirability of outsourcing subassemblies is the development of various reliable, high-power ultraviolet (UV) lasers (especially frequency-converted solid-state devices). Optics for UV lasers are more sensitive than their visible and infrared counterparts. Deep-UV coatings must be cleaned and handled very carefully to avoid the introduction of any contaminants and to prevent damage when used in high-power laser systems.
Furthermore, a given flatness specification is much harder to achieve in the UV than in the visible. A l/10 flatness specification at 532 nm, for example, is only half the actual surface flatness of l/10 at 266 nm. Thus, maintaining a given surface figure throughout the coating and mounting process becomes ever more difficult at shorter wavelengths.
Optomechanical assembly requirements
Actually producing precision optomechanical subassemblies requires a thorough knowledge of the interaction of optical components and mechanical mounts, as well as an understanding of how the assembly will ultimately be used and expected to perform. The vendor must have the ability to build or buy both optical and mechanical components of the requisite quality level.
There are also important issues in obtaining substrates from one vendor and having them coated by another. High-damage-threshold, deep-UV coatings, in particular, place stress on an optic, distorting its surface figure. Thus, to produce very flat, coated optics (such as l/20), it is necessary to understand the stress characteristics of the coating so that these can be taken into account during the substrate polishing process. Depending upon whether the coating is compressive or tensile, the part may have to be contoured so that the coating stress will pull it into the desired shape, rather than away from that shape.
Producing precision optomechanical products also requires the ability to design any necessary assembly tooling, having the appropriate personnel and facilities for performing the mating (such as clean rooms), and the ability to measure total performance (typically autocollimation and interferometry) of the subassembly. Packaging that avoids environmental contamination, such as nitrogen backfilling, may even be required.
Precision assembly techniques
To understand these issues and the type of equipment and expertise that must be applied in precision assembly, it is useful to examine a few examples. One of the most basic and commonly encountered optomechanical assembly tasks is placing a round mirror in a mount that retains the optic using a single setscrew. This may seem simple, but it is actually difficult to perform while routinely maintaining a flatness specification of l/10 or better. The problem is that the setscrew places an asymmetrical stress on the optic, which can push it out of figure.
To perform this type of assembly to high precision, the process must be monitored interferometrically. This allows the effects of screw tightness on surface flatness to be assessed dynamically. One remedy used to counteract or lessen the screw stress is to shim the optic; another solution is to add a large pad to the end of the screw to spread out the force it places on the optic over a larger area.
An alternate approach is to first fasten the optic inside an annular metal collar, typically using cement, and then place this assembly in the mirror mount. The rigidity of the metal collar then largely protects the mirror from the screw force. This approach is valuable for optics that must be replaced in the field, as it allows personnel to put a new optic in the mount and tighten the retaining screw with relative impunity.
An example of a more-complex assembly involves bonding a UV mirror to a galvanometer scanner while maintaining l/20 mirror flatness and arc-second-level alignment of the mirror to the mounting post. This type of assembly is found in various types of semiconductor-process equipment, such as memory-repair units.
One issue is the effect of the adhesive on mirror flatness. As the adhesive cures, it can change shape slightly, putting stress on the optic and deforming it from the specified flatness. One way to counteract this is to thoroughly characterize the effects of the glue on the optic beforehand. Then, the optics can be purposefully shaped during fabrication so that the stress introduced by the adhesive pulls them into greater flatness.
Another possible solution is to put an optical coating on the backside of the mirror. This second coating, which has no optical function, can mechanically prestress the optic so that it is doesn't warp out of shape during bonding.
The accurate mating of the mirror to the mounting post requires custom tooling that allows the alignment to be monitored by laser autocollimation (see photo on p. 131). Active adjustment of the mirror position may then be required to compensate for any shifting that occurs as the cement cures.
WAYNE PANTLEY is sales manager and DAVID COLLIER is president of Alpine Research Optics, 3180 Sterling Circle, Ste. 101, Boulder, CO 80301; e-mail: [email protected].