Beam shaping technology can greatly improve laser process efficiency by enabling parallel
processes and increasing precision, quality and process stability. This paper outlines a system
design, optical code analysis and the bench testing of a patented [1,2] laser beam homogenization
and imaging system using prism beam splitting to produce a three spot array. The system uses a
beam integrator to produce a rectangular spot that is split into three beams by two prisms. A
second set of prisms directs the two outer beams onto an imaging lens and sets the pitch of the
virtual spots. These beams, with the central beam, are imaged to form three spots with the
required pitch. A prototype system design was developed for two approaches based on the first
principles. The prototype system parameters were adjusted to minimize the requirements of the
elements such as the imaging lens and prisms. Since the two systems require a relatively fast
imaging lens, and there are aberrations associated with the prisms, a detailed optical design was
conducted to determine the performance of the two approaches and to assess the complexity of
the imaging lens. This paper will present the various positive and negative attributes of the two
beam shaper designs within an optical system and how the best design was selected for
prototyping and bench testing. Various data will be presented at each stage of design evaluation
to the final bench test.
Applications requiring injection of a high-power multimode laser into multiple fibers with equal energies, or specific energy ratios, provide unique design challenges. As with most all systems, engineering trades must balance competing requirements to obtain an optimal overall design. This is particularly true when fabrication issues are considered in the design process. A few of these competing design requirements are discussed in this conceptually simple system. This fiber injection system consists of three components; a refractive beam homogenizer, a diffractive beamsplitter, and a fiber array. We show the design process, starting with first-order design, for an example fiber injection system that couples a high-power YAG laser into seven fibers. Design goals include high efficiency, good beamsplitting uniformity, compact overall size, maximum mode filling of the fibers, and low cost of fabrication and assembly.
Manufacturing processes have been developed to produce high performance wafer form microoptics in both bulk zinc selenide and bulk multispectral zinc sulfide. Gray scale photolithography techniques have been used to pattern aspheric refractive lenses and beam shaping diffractive structures in wafer form for both of the zinc based II-VI group materials. High density plasma etching recipes have been refined to etch gray scale photoresist patterns into the bulk II-VI wafer materials with controllable selectivity. These IR materials have the advantage over other IR materials of transmitting broadband radiation, including visible band radiation. This very wide transmission band capability (visible to LWIR) permits dual band applications to use the same optical path. The high index of refraction of these materials permits production of higher numerical aperture lenses that have reduced lens sag requirements.
There are numerous laser applications that require the laser irradiance to possess a controlled shape and uniformity across a specific area at some distance from the source. Such applications include material processing (cutting, welding, drilling, heat-treating), scribing, marking, and microfabrication. One method of achieving this goal is to employ a double-sided micro lens array (MLA) in an imaging multi-aperture beam integrator configuration as discussed by Brown, Dickey, and Weichman. Such a configuration consists of two elements-the-double-sided MLA and a focusing lens. The micro lens array serves to segment the incoming beam at each subaperture into multiple beams, which are then overlapped at the image plane by the focusing lens. The resultant image will have the same shape as the subapertures of the micro lens array, which means that almost any shape can be generated. Lenslet subapertures having square, rectangular, or hexagonal shapes that can be stacked with 100% fill factor are typically used in order to reduce the amount of energy lost at the image plane. The operation of microlens arrays as diffusers are examined in the following paper from a ray optics and a physical optics point of view. Modeling examples/techniques are discussed for both approaches as well.
An innovative method of optically combining the multiple emitters of high power stripe laser diodes is presented. A multi-beam integrator approach is used along with a means of optically rotating the emitters by 90 degrees. The 90-degree rotation allows better balancing of the optical invariants in the slow and fast axis directions, making it easier to "circularize" the image for coupling into fibers or to increase the irradiance for laser machining applications.
Circular gaussian-to-supergaussian beam shapers can easily be designed with ray tracing lens design software that provides access to the merit function. Noncircular top hats or supergaussian have traditionally used iterative back-propagation methods, such as Gershberg-Saxon routines for generating the phase function of the shaper. This paper presents a simple method to design rectangular top-hat beam shapers with ray tracing software by proper generation of the ray targets and selection of the diffractive element coefficients to vary under optimization. The design method is much quicker than previously used back-propagation algorithms.
An innovative multi-aperture beam integrator system producing a polygon pattern or other complex pattern on an image plane, which can be continuously varied in size, aspect ratio, or shape, is described in this paper. The multi-aperture function of the integrator system consists of three array elements, one stationary and the other two movable laterally. The first array element forms an array of Galilean telescopes which collect the light with 100% fill factor, segments the incoming beam into multiple beamlets, and reduces the diameter of each beamlet in order to prevent vignetting at the following movable arrays. The second and third array elements each contain a plurality of the anamorphic lenslets, not all the same, each lenslet pair of which forms one of the sides of the polygon or complex image. These array elements have a surface profile or phase function, which can be described as a general polynomial of the pupil coordinates. An aberration-free integrator lens overlaps the beamlets in the image plane to accomplish beam homogenization and produce a uniform intensity for all the sides of the polygon or segments of the complex image. Small lateral translations of the second and third array elements cause the image characteristics to change continuously.
A common method of shaping laser beams into patterns for product marking and machining involves the use of masks. These masks result in inefficient use of the available laser energy, causing increased fabrication time and cost. Simple patterns, such as rings and tophat profiles can be easily formed with multi-aperture beam integrators, thus utilizing nearly 100% of the laser energy. This paper discusses multi-aperture beam integrators for forming a continuously variable ring pattern on the target plane. Both on-axis and off-axis systems are discussed.
Multi-aperture beam integrators are used to homogenize high power multi- mode lasers and arc lamps for various industrial applications. Some applications, such as product marking and laser machining, require specific odd-shape irradiance patterns on the target. Current masking techniques waste laser power. We explore here the techniques of aperture flipping and highly aberrated lenslets in the beams segmenter element of multi-aperture beam integrators.
This paper reports on work to determine the feasibility of fabricating a liquid crystal display (LCD) backlight using an array of LEDs. The purpose of this backlight is to overcome the efficiency loss of the absorptive color filters in the LCD. Two types of arrays were fabricated and tested. An array of white LEDs was designed for use with an interference color separation filter. An array of red, green, and blue LEDs was designed for use with a cylindrical lenslet array which focuses the three colors onto the appropriate color apertures of the LCD. Although promising results were achieved with standard off-the-shelf LEDs, significant improvements in surface radiance uniformity and efficiency could be obtained with special beam-shaping lenses on the LEDs.
The design, fabrication, and testing of a new anamorphic microlens for laser diode circularization is presented. The microlens is fabricated using photolithography, gray scale masks, and reactive ion etching. A front-to-back mask aligner is used to precisely align two anamorphic aspheric microlenses on opposite sides of a single wafer, creating the ability to circularize or collimate a laser diode with only a single monolithic element. The entire fabrication process is highly nonlinear. This requires that accurate surface metrology methods be incorporated into the process as a feedback loop for iterative corrections to the gray scale mask. We discuss the measurement process and the effects of surface errors for circularizers and circularizer/collimators. This device has been under development for about two years at MEMS Optical. The most recent fabrication and test results of an actual device are presented.
Using two micro lens arrays and a MEMS micro shutter array, an intensity modulating Spatial Light Modulator is being developed at MEMS Optical, Inc. (patent pending) for high speed printing applications. The micro lens arrays are used to focus incident light to a point and then expand it back to its original size. At the focus point, a Foucault micro shutter array is used to modulate the amount of light allowed to pass through the aperture. The purpose for this device is for exposure control for high-speed electronic printing applications. The drive mechanism is based on an electrostatic lateral comb interdigitated drive. Design analysis shows a rise time of 1 - 2 microseconds for high voltage systems. This array of shutters is being implemented in a CMOS compatible process, and is capable of being integrated with on chip circuitry for opening and closing the shutters. The apertures are made using deep RIE etching, and the shutters are released using plasma etching. The result is an electronically controlled method of exposing a photosensitive surface at high speeds for the printing industry, with or without lasers.
Technologies generally used for fabrication of kinoform diffractive optics include; direct writing, plastic molding, diamond turning, and photolithography. Photolithographic methods (either contact or projection) are generally suitable for mass production in glass. Two types of masks are used with photolithographic methods; binary chrome masks and gray scale masks. Contact lithography with binary chrome masks generally limits minimum features sizes, and thus minimum zone widths, due to performance degradation from alignment errors between multiple masks. For example, the minimum zone width of a high efficiency eight-level kinoform is eight times the minimum feature size. Alternatively, gray scale mask technology uses a single mask which eliminates the alignment error problems. Smooth profile (not stair- stepped) high efficiency kinoform zones as well as three microns have been fabricated with this technology. In this paper we report on a direct experimental comparison of costs and performance for a blazed grating with 6-micron zones fabricated with multiple binary chrome masks and a single gray scale mask.
This paper discusses the application of MOEM technology to adaptive optics. An experiment is described in which a micromachined mirror array is used in a closed loop adaptive optic demonstration. An interferometer wavefront sensor is used for wavefront sensing. Parallel analog electronics are used for the wavefront reconstruction. Parallel operational amplifiers are used to drive the micromirrors. The actuators utilize a novel silicon design developed by SY Technology, Inc. The actuators have a measured frequency response of 15kHz, and a maximum usable stroke of 4 microns. The entire adaptive optic demonstration has a bandwidth exceeding 10kHz. Measured performance is described. The experiments conducted are designed to explore the feasibility of creating a single chip adaptive optic system, also described in this paper. This chip would combine all on a single VLSI chip aspects of a complete adaptive optics system, wavefront sensing, wavefront reconstruction, and wavefront correction. The wavefront sensing would be accomplished with a novel compact shearing interferometer design. The analog refractive and diffractive micro optics will be fabricated using a new single step analog mask technology. The reconstruction circuit would use an analog resistive grid solver. The resistive grid would be fabricated in polysilicon. The drive circuits and micromirror actuators would use standard CMOS silicon fabrication methods.
In this paper, cascaded diffractive optical elements are investigated using a design strategy combining genetic algorithms with beam propagation methods. Results are presented for a two element cascaded system for multiple wavelength performance.
Diffractive optical elements (DOEs) have demonstrated applicability to a wide range of optical problems which cannot be solved with conventional optical elements. DOEs are increasingly becoming solutions for numerous nonconventional imaging tasks such as fan-out gratings and extended imagers. Applications include product marking, machining, robotics, medical diagnostic instruments, alignment instruments, optical computers, and fiber optic switches, to list only a few. However, the phase functions provided by most commercial lens design codes for DOE design lack the generality needed for these nonconventional imaging tasks. The solution taken in the past, of pixelating the aperture and independently varying the phase of every pixel, requires writing a specialized wavefront propagator code since such a phase function is unsuitable for ray tracing codes. We show here new phase functions which are more general than those currently provided in commercial lens design codes but which remain easily adaptable to these codes. Example designs, demonstrating the increased flexibility of these new phase functions, are also shown.
A high speed Hartmann wavefront sensor was designed and built for measuring refractive index variations in supersonic air flows. The device contained a lenslet array which formed an array of spots on the focal plane of a high speed camera. Spot motion at the focal plane is directly related to fluctuation of wavefront tilt in the corresponding subaperture. Both refractive and diffractive (binary optic) lenslet arrays were fabricated for the Hartmann sensor. The long focal length needed to meet resolution requirements placed tight tolerances on the wedge error in the refractive lenslets and the cement interface mounting the lenslets to the substrate lens. In spite of the use of state-of-the-art interferometric alignment techniques for assembling the refractive lenslet array, the diffractive lenslet array demonstrated superior alignment and performance. In this application binary optics demonstrated significant advantages over conventional optics. In addition to performance issues, binary optics allows reduced weight and reduced number of elements. For airborne and space applications these advantages translate into significant cost savings.
The focal plane of an infrared seeker was plagued with ghost images and nonuniform stray light irradiance. Teledyne Brown Engineering was tasked to determine the irradiance source and propose inexpensive solutions to the problems. First order analysis approximately modeled the focal plane irradiance and showed a serious flaw in the design. A design flaw allowed normal internally emitted thermal radiation to develop into a high level, nonuniform, focal plane irradiance. Exact ray tracing software, developed by the author, computed focal plane irradiance distributions which closely matched measured distributions. The software performs a non-sequential surface ray trace, splitting rays at partially reflecting surfaces (using a recursive algorithm), and computes internal thermal emission. The stray light problems could have been avoided in a design with the cold stop as the system aperture stop. This paper shows the method of analysis, results, and proposed solutions to the problem. This work demonstrates how infrared optical design requires precautions and considerations. Methods and tools which work well in visible optical design may not work in infrared optical design.
Software is under development at Teledyne Brown Engineering to represent a lens configuration as a y-ybar or Delano diagram. The program determines third-order Seidel and chromatic aberrations for each configuration. It performs a global search through all valid permutations of configuration space and determines, to within a step increment of the space, the configuration with smallest third-order aberrations. The program was developed to generate first-order optical layouts which promised to reach global minima during subsequent conventional optimization. Other operations allowed by the program are: add or delete surfaces, couple surfaces (for Mangin mirrors), shift the stop position, and display first-order properties and the optical layout (surface radii and thicknesses) for subsequent entry into a conventional lens-design program with automatic optimization. Algorithms for performing some of the key functions, not covered by previous authors, are discussed in this paper.
Teledyne Brown Engineering designed, fabricated and tested an infrared telescope using only spherical mirror elements. Aberrations were corrected with a binary optic pattern etched onto a germanium lens. The telescope is an F/3, off-axis Gregorian design with no obscuration. The field-of-view (FOV) is 4x8 degrees and it operates in the 8 to 12 micron waveband, with an entrance pupil of 5 cm. The telescope demonstrates that a single binary optical element can correct a significant amount of both pupil- and field-dependent aberrations introduced by tilted spherical mirrors, while maintaining a broad wavelength band of operation. The line spread functions, measured at 10 microns on the telescope, coincided very well with theoretical line spread functions generated by a commercial lens design code.
The authors analyzed techniques for designing lenses consisting of conventional reflective or refractive elements used in conjunction with binary optical elements (BOEs), and reviewed the use of BOE correctors in systems with both rotational and bilateral symmetry. These systems add a large number of degrees of freedom, such that almost any arbitrary wavefront shape can be obtained. For this reason, automatic optimization with BOEs requires special considerations; especially, tilted or decentered systems, wide-waveband systems, or wide field-of-view systems. If caution is not exercised, one can easily consume large amounts of expensive cpu time only to converge on nonoptimal dead-end solutions. A typical situation that occurs is the convergence upon a solution where the performance is diffraction limited at the finite number of defined field points, but strongly aberrated at field points in between. The key to converging on an optimal solution is the proper use of user defined constraints. The authors studied how to select which BOE coefficients to vary and how to choose field points in order to increase the probability of converging on optimal solutions.
The Advanced X-ray Astrophysical Facility (AXAF) telescope consists of six concentric paraboloid-hyperboloid pairs of mirrors operating near grazing incidence. Because of the substantial polarization effects at large angles of incidence, there is concern about the feasibility of polarimetry near the focal plane. The primary mirror acts as a tangentially directed half-wave linear retarder and nearly completely depolarizes the linearly polarized component of the light. The secondary mirror introduces an additional half wave of linear retardance. The tangentially directed one-wave linear retarder leaves the transmitted beam in the incident polarization state. The net instrumental polarization effects are small, and polarimetry is feasible with the AXAF.