A primary goal of optics is to focus, collimate (parallellize), or otherwise shape a beam so it can be used in its intended applications. One particular optical system, conventionally called a beam shaper, converts an input beam with a known intensity profile to a beam with a different defined output intensity distribution. These systems normally change a typical Gaussian (collimated) beam from a laser into a circular or square flat-top beam.
Beam shapers are especially useful in laser machining, laser eye surgery, and automated wafer inspection. Many laser-machining applications, for example, require illuminating a work piece with a uniform line of light. Although this line can be generated with a cylindrical fly's eye array, the result often shows diffraction ringing and speckle within the area of interest on the work piece. Similarly, a diffractive line generator (or diffuser) also exhibits speckle. Because a beam shaper maps one point in the input field to one point in the output field, it avoids these annoyances in the image plane. Also, unlike diffusers and fly's eye elements, beam shapers do not notably increase the etendue (similar to entropy in a thermodynamic system) of an optical system. These advantages suggest that beam shapers are the best method for creating a uniform source for applications where speckle is undesired. However, system designers often resist using beam shapers, because they are sensitive to misalignment and need to be designed for an input source with a specific intensity profile and collimation.
We have found that designing with these constraints in mind makes it possible to mitigate the disadvantages.1 Indeed, often a system can be made to meet requirements with reasonable tolerances in alignment and beam profile (see Figure 1). Designing the phase function needed to map the beam is relatively straightforward. The critical issue is understanding how various fabrication and alignment issues will affect the beam shaper's performance. The intensity and phase of the beam can be accurately simulated with beam-propagation software tools. Further, the effects of various errors, such as misalignment and beam variations, can be simulated to give the system designer an accurate impression of how the beam shaper will perform.
Figure 1. Measured input and output from a beam-shaper system (in arbitrary units). A Gaussian input is converted to a one-dimensional flat-top beam (flat in one dimension and Gaussian in the other).
The idealized continuous-phase profile of the beam shaper can be closely approximated by a discrete, quantized profile using computer optimization. We then employ a microlithographic process that consists of electronic-beam mask writing and advanced fabrication techniques to create diffractive structures with positional placement on the order of nanometers and size control on the order of hundreds of nanometers. Although these discrete levels cause higher-order (stray) light in the system, this process allows for much greater control of the design than with conventional fabrication methods, thereby allowing diffractive optics to perform functions previously not possible.
The results from simulations show that there are certain critical tolerances that must be examined carefully (most other tolerances are of low importance). These tolerances typically include the incoming beam size and divergence, as well as the positional alignment between the beam and the beam-shaper optic. Assuming the required criteria can be met, the beam shaper should easily perform within the target specifications. For example, Table 1 shows a comparison between simulated and measured values of a fabricated diffractive-optic beam-shaper system.
Comparison of simulated and measured parameters of a beam-shaper system.
|Uniformity||7%||8%||Intensity error in optical system|
|Wavefront error||0.01−0.03 rms||< 0.05 rms||Phase error in optical system|
The microlithographic process allows a designer to place multiple optical functions on a single glass substrate, such as collimating or focusing elements, alignment features, and apertures. Combined with the flexibility of beam shapers to control the light profile, complete systems can be created with very few low-cost elements. We are currently working on more advanced systems that could combine these optical elements with active components (such as lasers and detectors) to provide a complete, self-contained module.
Adam Fedor, Marc Himel
Tessera North America
Marc Himel is senior principal engineer.