Search

We present a new, to the best of our knowledge, experimental method for assessing sub-micron level subsurface damage (SSD) on optical glass. The method correlates surface characteristics such as the fracture toughness and Young's modulus via nanoindentation with the penetration depth into the tested surfaces at different overall penetration depths, as revealed by magnetorheological finishing spotting techniques. Our results on ground surfaces suggest that low surface roughness does not necessarily imply the absence of SSD. We also compared SSD on surfaces processed by deterministic microgrinding and femtosecond (fs) laser polishing. The fs-laser polished surfaces revealed no detectable SSD, thus establishing the feasibility of fs-laser polishing for precision optical manufacturing.

This paper is the second part of a study of the grinding of three different grades of silicon carbide (SiC) under the same conditions. In this paper, subsurface damage is analyzed using magnetorheological finishing (MRF). The MRF ribbon is brought into contact with the surface and allowed to dwell for different lengths of time to produce dimples or spots at different depths. The roughness parameters are evaluated at the base of the spots. As the spot depth increases the roughness parameters decrease, eventually saturating at a steady-state value. The depth at which saturation occurs is much greater than the initial peak-to-value roughness of the surface and is therefore assumed to be correlated to the depth of subsurface damage in the material. Estimates of the damage depth are comparable to other estimates given in the literature. The validity of this technique is assessed across different grades of SiC under different grinding conditions, and limitations are identified. The study shows that the microstructure of the SiC grade is a major factor that affects the validity of the technique. The technique is suitable for SiC grades that have a more homogeneous microstructure such as chemical vapor deposited or chemical vapor composite grades. The presence of porosity as in the direct sintered grade, or the presence of a secondary phase, for example, silicon in the reaction-bonded and silicon-infiltrated grades, could hinder the technique from providing conclusive results.

This paper presents a study of the grinding of three different grades of silicon carbide (SiC) under the same conditions. Surface topography is analyzed using coherent scanning interferometry and scanning electron microscopy. The study provides a baseline understanding of the process mechanics and targets effective selection of process parameters for grinding SiC optics with near optical level surface roughness, thus reducing the need for post-polishing. Samples are raster and spiral ground on conventional precision machines with metal and copper-resin bonded wheels under rough, medium, and finish grinding conditions. Material microstructure and grinding conditions affect attainable surface roughness. Local surface roughness of less than 3 nm RMS was attained in both chemical vapor deposition (CVD) and chemical vapor composite (CVC) SiC. The tool footprint is suitable for sub-aperture machining of a large freeform optics possibly without the need for surface finish correction by post-polishing. Subsurface damage will be assessed in Part 2 of this paper series.

Freeform optical components enable dramatic advances for optical systems in both performance and packaging. Surface form metrology of manufactured freeform optics remains a challenge and an active area of research. Towards addressing this challenge, we previously reported on a novel architecture, cascade optical coherence tomography (C-OCT), which was validated for its ability of high-precision sag measurement at a given point. Here, we demonstrate freeform surface measurements, enabled by the development of a custom optical-relay-based scanning mechanism and a unique high-speed rotation mechanism. Experimental results on a flat mirror demonstrate an RMS flatness of 14 nm (∼λ/44 at the He-Ne wavelength). Measurement on a freeform mirror is achieved with an RMS residual of 69 nm (∼λ/9). The system-level investigations and validation provide the groundwork for advancing C-OCT as a viable freeform metrology technique.

Grinding or finishing of thin curved optics introduces surface-localized stresses that may result in bending deformations, thus affecting optical performance (Twyman effect). The Twyman effect depends on local thickness of the optic as well as its curvature. We use numerical and analytical techniques to assess the effects of these variables for shallow and deeply curved axisymmetric optics. A useful metric of comparison, valid for shallow optics, is the deformation of a similar flat optic (window). This comparison is inadequate for more curved optics, which demonstrate a smaller deformation that is now localized in a boundary layer near the optic-free edge. The effects of thickness variation are also presented. Normal deflection and slope for a positive optic (thickness at the center greater than the tip) are well approximated by a constant thickness optic. This is not the case for a negative optic, where deformations near the apex are more pronounced, thus extending significant deformations through the full optic aperture.

We investigate the interaction of 355-nm and 1064-nm nanosecond laser pulses with nominally spherical metallic particles dispersed on the input surface of transparent substrates or high-reflectivity (HR) multilayer dielectric coatings, respectively. The objective is to elucidate the interaction mechanisms associated with contaminant-induced degradation and damage of transparent and reflective optical elements for high-power laser systems. The experiments involve time-resolved imaging capturing the dynamics of the interaction pathway, which includes plasma formation, particle ejection, and secondary contamination by droplets originating from the liquefied layer of the particle. The results suggest that HR coatings are more susceptible to secondary contamination by liquid droplets produced by the particles because of the different geometry of excitation and the location of plasma initiation. Modeling results focus on better understanding the melting of the particle surface, leading to ejections of liquid droplets and the pressure applied to the substrate, leading to mechanical damage.

The physical mechanisms and ensuing material modification associated with laser-induced damage in multilayer dielectric high reflectors is investigated for pulses between 0.6 and 100 ps. We explore low-loss multilayer dielectric SiO 2 /HfO 2 mirrors which are commonly employed in petawatt-class laser systems. The spatial features of damage sites are precisely characterized, enabling the direct correlation of the observed damage morphology to the location of energy deposition and the corresponding standing-wave electric-field intensities within the layer structure. The results suggest that there are three discrete damage-initiation morphologies arising from distinctly different mechanisms: the first prevailing at laser pulse lengths shorter than about 2.3 ps, while the other two are observed for longer pulses. Modeling of the thermomechanical response of the material to localized laser-energy deposition was performed for each type of damage morphology to better understand the underlying mechanisms of energy deposition and subsequent material response.

Multilayer dielectric (MLD) gratings used in ultrahigh-intensity laser systems often exhibit a laser-induced damage performance below that of their constituent materials. Reduced performance may arise from fabrication- and/or design-related issues. Finite element models were developed to simulate stress waves in MLD grating structures generated by laser-induced damage events. These models specifically investigate the influence of geometric and material parameters on how stress waves can lead to degradation of material structural integrity that can have adverse effects on its optical performance under subsequent laser irradiation: closer impedance matching of the layer materials reduces maximum interface stresses by ~20% to 30%; increasing sole thickness from 50 nm to 500 nm reduces maximum interface stresses by ~50%.

We demonstrate how a nanomechanical test identifies areas of mechanical field concentration as being comparable to areas where optical fields are known to be concentrated, in the special context of laser-induced damage testing (LIDT) of diffractive gratings of silica deposited on optical multilayers. The nano-indentation response of the diffraction gratings is measured in a new mode that allows for the extraction of a measurable metric characterizing the brittleness of the gratings, as well as their ductility. We show that lower LIDTs are positively correlated with an increased grating brittleness, and therefore identify a nanomechanical approach to describe LIDTs. We present extensive numerical simulations of nano-indentation tests and identify different deformation modes including stretching, shear concentration, and bending as precursors to mechanical failure in the nano-indentation test. The effects of geometrical inhomogeneities on enhanced stress generation in these gratings are specifically examined and addressed, and we show the agreement between nanomechanical testing and analytical interpretation of these inhomogeneities.

Chemical-vapor-deposited (CVD) ZnS is an example of a polycrystalline material that is difficult to polish smoothly via the magnetorheological finishing (MRF) technique. When MRF-polished, the internal infrastructure of the material tends to manifest on the surface as millimeter-sized "pebbles," and the surface roughness observed is considerably high. The fluid's parameters important to developing a magnetorheological (MR) fluid that is capable of polishing CVD ZnS smoothly were previously discussed and presented. These parameters were acidic pH (∼4.5) and low viscosity (∼47  cP). MRF with such a unique MR fluid was shown to reduce surface artifacts in the form of pebbles; however, surface microroughness was still relatively high because of the absence of a polishing abrasive in the formulation. In this study, we examine the effect of two polishing abrasives-alumina and nanodiamond-on the surface finish of several CVD ZnS substrates, and on other important IR polycrystalline materials that were finished with acidic MR fluids containing these two polishing abrasives. Surface microroughness results obtained were as low as ∼28  nm peak-to-valley and ∼6-nm root mean square.

We describe the anisotropy in the material removal rate (MRR) of the polycrystalline, chemical-vapor deposited zinc sulfide (ZnS). We define the polycrystalline anisotropy via microhardness and chemical erosion tests for four crystallographic orientations of ZnS: (100), (110), (111), and (311). Anisotropy in the MRR was studied under magnetorheological finishing (MRF) conditions. Three chemically and mechanically modified magnetorheological (MR) fluids at pH values of 4, 5, and 6 were used to test the MRR variations among the four single-crystal planes. When polishing the single-crystal planes and the polycrystalline with pH 5 and pH 6 MR fluids, variations were found in the MRR among the four single-crystal planes and surface artifacts were observed on the polycrystalline material. When polishing the single-crystal planes and the polycrystalline with the modified MR fluid at pH 4, however, minimal variation was observed in the MRR among the four orientations and a reduction in surface artifacts was achieved on the polycrystalline material.

Mechanical characterization of optical oxide thin films is performed using nano-indentation, and the results are explained based on the deposition conditions used. These oxide films are generally deposited to have a porous microstructure that optimizes laser induced damage thresholds, but changes in deposition conditions lead to varying degrees of porosity, density, and possibly the microstructure of the thin film. This can directly explain the differences in the mechanical properties of the film studied here and those reported in literature. Of the four single-layer thin films tested, alumina was observed to demonstrate the highest values of nano-indentation hardness and elastic modulus. This is likely a result of the dense microstructure of the thin film arising from the particular deposition conditions used.

A thermomechanical model is developed to estimate the stress response of an oxide coating to elevated-temperature chemical cleaning. Using a hafnia-silica multilayer dielectric pulse compressor grating as a case study, we demonstrate that substrate thickness can strongly affect the thermal stress response of the thin-film coating. As a result, coatings on large, thick substrates may be susceptible to modes of stress-induced failure (crazing or delamination) not seen in small parts. We compare the stress response of meter-scale optics to the behavior of small-scale test or "witness" samples, which are expected to be representative of their full-size counterparts. The effects of materials selection, solution temperature, and heating/cooling rates are explored. Extending the model to other situations, thermal stress results are surveyed for various combinations of commonly used materials. Seven oxide coatings (hafnia, silica, tantala, niobia, alumina, and multilayers of hafnia-silica and alumina-silica) and three glass substrates (BK7, borosilicate float glass, and fused silica) are examined to highlight some interesting results.

During the fabrication of multilayer-dielectric (MLD) thin-film-coated optics, such as the diffraction gratings used in OMEGA EP's pulse compressors, acid piranha cleaning can lead to the formation of chemically induced delamination defects. We investigate the causes of these defects and describe a mechanism for the deformation and failure of the MLD coating in response to hydrogen peroxide in the cleaning solution. A fracture mechanics model is developed and used to calculate the crack path that maximizes the energy-release rate, which is found to be consistent with the characteristic fracture pattern observed in MLD coating delamination defects.

Plasma-assisted electron-beam evaporation leads to changes in the crystallinity, density, and stresses of thin films. A dual-source plasma system provides stress control of large-aperture, high-fluence coatings used in vacuum for substrates 1m in aperture.

We use a recent experimental technique to measure in situ shear and normal stresses during magnetorheological finishing (MRF) of a borosilicate glass over a range of magnetic fields. At low fields shear stresses increase with magnetic field, but become field-independent at higher magnetic fields. Micromechanical models of formation of magnetic particle chains suggest a complex behavior of magnetorheological (MR) fluids that combines fluid- and solid-like responses. We discuss the hypothesis that, at higher fields, slip occurs between magnetic particle chains and the immersed glass part, while the normal stress is governed by the MRF ribbon elasticity.

We investigate the effects of processing parameters on material removal for borosilicate glass. Data are collected on a magnetorheological finishing (MRF) spot taking machine (STM) with a standard aqueous magnetorheological (MR) fluid. Normal and shear forces are measured simultaneously, in situ, with a dynamic dual load cell. Shear stress is found to be independent of nanodiamond concentration, penetration depth, magnetic field strength, and the relative velocity between the part and the rotating MR fluid ribbon. Shear stress, determined primarily by the material mechanical properties, dominates removal in MRF. The addition of nanodiamond abrasives greatly enhances the material removal efficiency, with the removal rate saturating at a high abrasive concentration. The volumetric removal rate (VRR) increases with penetration depth but is insensitive to magnetic field strength. The VRR is strongly correlated with the relative velocity between the ribbon and the part, as expected by the Preston equation. A modified removal rate model for MRF offers a better estimation of MRF removal capability by including nanodiamond concentration and penetration depth.

This paper describes the surface modification of micrometer-sized magnetic carbonyl iron particles (CI) with zirconia from zirconium(IV) butoxide using a sol-gel method. Zirconia shells with various thicknesses and different grain sizes and shapes are coated on the surface of CI particles by changing the reaction conditions, such as the amounts of zirconia sol, nitric acid, and CI particles. A silica adhesive layer made from 3-aminopropyl trimethoxysilane (APTMS) can be introduced first onto the surface of CI particles in order to adjust both the size and the shape of zirconia crystals, and thus the roughness of the coating. The microanalyses on these coated particles are studied by field-emission scanning electron microscopy (FE-SEM) and X-ray-diffraction (XRD). Accelerated acid corrosion and air oxidation tests indicate that the coating process dramatically improved oxidation and acid corrosion resistances, which are critical issues in various applications of CI magnetic particles., (Copyright 2009. Published by Elsevier Inc.)

We report on magnetorheological finishing (MRF) spotting experiments performed on glasses and ceramics using a zirconia-coated carbonyl-iron (CI)-particle-based magnetorheological (MR) fluid. The zirconia-coated magnetic CI particles were prepared via sol-gel synthesis in kilogram quantities. The coating layer was approximately 50-100 nm thick, faceted in surface structure, and well adhered. Coated particles showed long-term stability against aqueous corrosion. "Free" nanocrystalline zirconia polishing abrasives were cogenerated in the coating process, resulting in an abrasive-charged powder for MRF. A viable MR fluid was prepared simply by adding water. Spot polishing tests were performed on a variety of optical glasses and ceramics over a period of nearly three weeks with no signs of MR fluid degradation or corrosion. Stable material removal rates and smooth surfaces inside spots were obtained.

We report in situ, simultaneous measurements of both drag and normal forces in magnetorheological finishing (MRF) for what is believed to be the first time, using a spot taking machine (STM) as a test bed to take MRF spots on stationary parts. The measurements are carried out over the entire area where material is being removed, i.e., the projected area of the MRF removal function/spot on the part surface, using a dual force sensor. This approach experimentally addresses the mechanisms governing material removal in MRF for optical glasses in terms of the hydrodynamic pressure and shear stress, applied by the hydrodynamic flow of magnetorheological fluid at the gap between the part surface and the STM wheel. This work demonstrates that the volumetric removal rate shows a positive linear dependence on shear stress. Shear stress exhibits a positive linear dependence on a material figure of merit that depends upon Young's modulus, fracture toughness, and hardness. A modified Preston's equation is proposed that better estimates MRF material removal rate for optical glasses by incorporating mechanical properties, shear stress, and velocity.

Magnetorheological finishing (MRF) is a deterministic subaperture polishing process. The process uses a magnetorheological (MR) fluid that consists of micrometer-sized, spherical, magnetic carbonyl iron (CI) particles, nonmagnetic polishing abrasives, water, and stabilizers. Material removal occurs when the CI and nonmagnetic polishing abrasives shear material off the surface being polished. We introduce a new MRF material removal rate model for glass. This model contains terms for the near surface mechanical properties of glass, drag force, polishing abrasive size and concentration, chemical durability of the glass, MR fluid pH, and the glass composition. We introduce quantitative chemical predictors for the first time, to the best of our knowledge, into an MRF removal rate model. We validate individual terms in our model separately and then combine all of the terms to show the whole MRF material removal model compared with experimental data. All of our experimental data were obtained using nanodiamond MR fluids and a set of six optical glasses.

We demonstrate the use of spots taken with magnetorheological finishing (MRF) for estimating subsurface damage (SSD) depth from deterministic microgrinding for three hard ceramics: aluminum oxynitride (Al(23)O(27)N(5)/ALON), polycrystalline alumina (Al(2)O(3)/PCA), and chemical vapor deposited (CVD) silicon carbide (Si(4)C/SiC). Using various microscopy techniques to characterize the surfaces, we find that the evolution of surface microroughness with the amount of material removed shows two stages. In the first, the damaged layer and SSD induced by microgrinding are removed, and the surface microroughness reaches a low value. Peak-to-valley (p-v) surface microroughness induced from grinding gives a measure of the SSD depth in the first stage. With the removal of additional material, a second stage develops, wherein the interaction of MRF and the material's microstructure is revealed. We study the development of this texture for these hard ceramics with the use of power spectral density to characterize surface features.

We present a nondestructive method for estimating the depth of subsurface damage (SSD) in some single crystalline optical materials (silicon, lithium niobate, calcium fluoride, magnesium fluoride, and sapphire); the method is established by correlating surface microroughness measurements, specifically, the peak-to-valley (p-v) microroughness, to the depth of SSD found by a novel destructive method. Previous methods for directly determining the depth of SSD may be insufficient when applied to single crystals that are very soft or very hard. Our novel destructive technique uses magnetorheological finishing to polish spots onto a ground surface. We find that p-v surface microroughness, appropriately scaled, gives an upper bound to SSD. Our data suggest that SSD in the single crystalline optical materials included in our study (deterministically microground, lapped, and sawed) is always less than 1.4 times the p-v surface microroughness found by white-light interferometry. We also discuss another way of estimating SSD based on the abrasive size used.

Polymer cholesteric liquid-crystal flakes suspended in a fluid with nonnegligible conductivity can exhibit motion in the presence of an ac electric field. The plateletlike particles with a Grandjean texture initially lie parallel to the cell substrates and exhibit a strong selective reflection that is diminished or extinguished as the flakes move. Flake motion was seen within a specific frequency bandwidth in an electric field as low as 5 mV(rms)/microm. Flakes reoriented about their longest axis to align parallel to theapplied field, and a rise time of 560 ms was seen in a 40-mV(rms)/microm field at 100 Hz.

Results of experiments on stress-inhibited laser-driven crack growth and stress-delayed laser-damage initiation thresholds in fused silica, borosilicate glass (BK-7), and cleaved bulk silica are presented. A numerical model is developed to explain the crack arrest in fused silica. Good agreement is obtained between the model and a finite-element code. The crack arrest is demonstrated to be the result of the breaking of a hoop-stress symmetry that is responsible for crack propagation in fused silica.

The 351-nm laser-damage initiation threshold for surface damage in conventionally polished fused silica is demonstrated to be stress dependent. By circumferential application of modest loads to a sample, a controllable stress field can be established within the clear aperture of a fused-silica specimen, in response to which both the damage-initiation fluence and the crack-propagation fluence requirements are increased above those for unstressed conditions.

Measurements of birefringence induced in fused-silica specimens by a crack produced by a 351-nm/500-ps Nd:glass laser as a function of laser fluence F(L) and of number of laser shots N are presented. The varying dimensional parameter is found to be the crack depth a and can be put in the form a(mm) = (0.0096 ? 0.0021)N[(F(L)/F(exit/th)) - 1](2/3) with F(L) >/= F(exit/th)(F(exit/th) is the exit-surface damage threshold). The retardance data are converted into units of stress, thus permitting the estimation of residual stress near the crack. The results of the measured residual stress can be cast in the form varsigma(r)(MPa) approximately (. ? .)[(F(L)/F(exit/th)) - 1](1/2) N(2/3) with F(L)>/= F(exit/th). A theoretical model giving the stress field around a crack is developed for comparison and shows reasonable agreement with the experiment. Good agreement with experimental data of others is also obtained. The effect of residual stresses on fracture strength is pointed out. The results obtained show that the presence of birefringence/residual stress in a fused-silica specimen with a crack on its surface has a strong effect on fracture and should be taken into account in any formulation that involves the failure strength of optical components used in inertial-confinement-fusion experiments.

We present an interpretation of the lapping hardness of commercially available optical glasses in terms of a micromechanics model of material removal by subsurface lateral cracking. We analyze data on loose abrasive microgrinding, or lapping at fixed nominal pressure, for many commercially available optical glasses in terms of this model. The Schott and Hoya data on lapping hardness are correlated with the results of such a model. Lapping hardness is a function of the mechanical properties of the glass: The volume removal rate increases approximately linearly with Young's modulus, and it decreases with fracture toughness and (approximately) the square of the Knoop hardness. The microroughness induced by lapping depends on the plastic and elastic properties of the glass, depending on abrasive shape. This is in contrast to deterministic microgrinding (fixed infeed rate), where it is determined from the plastic and fracture properties of the glass. We also show that Preston's coefficient has a similar dependence as lapping hardness on glass mechanical properties, as well as a linear dependence on abrasive size for the case of brittle material removal. These observations lead to the definition of an augmented Preston coefficient during brittle material removal. The augmented Preston coefficient does not depend on glass material properties or abrasive size and thus describes the interaction of the glass surface with the coolant-immersed abrasive grain and the backing plate. Numerical simulations of indentation are used to locate the origin of subsurface cracks and the distribution of residual surface and subsurface stresses, known to cause surface (radial) and subsurface (median, lateral) cracks.

In the Twyman effect (1905), when one side of a thin plate with both sides polished is ground, the plate bends: The ground side becomes convex and is in a state of compressive residual stress, described in terms of force per unit length (Newtons per meter) induced by grinding, the stress (Newtons per square meter) induced by grinding, and the depth of the compressive layer (micrometers). We describe and correlate experiments on optical glasses from the literature in conditions of loose abrasive grinding (lapping at fixed nominal pressure, with abrasives 4-400 μm in size) and deterministic microgrinding experiments (at a fixed infeed rate) conducted at the Center for Optics Manufacturing with bound diamond abrasive tools (with a diamond size of 3-40 μm, embedded in metallic bond) and loose abrasive microgrinding (abrasives of less than 3 μm in size). In brittle grinding conditions, the grinding force and the depth of the compressive layer correlate well with glass mechanical properties describing the fracture process, such as indentation crack size. The maximum surface residual compressive stress decreases, and the depth of the compressive layer increases with increasing abrasive size. In lapping conditions the depth of the abrasive grain penetration into the glass surface scales with the surface roughness, and both are determined primarily by glass hardness and secondarily by Young's modulus for various abrasive sizes and coolants. In the limit of small abrasive size (ductile-mode grinding), the maximum surface compressive stress achieved is near the yield stress of the glass, in agreement with finite-element simulations of indentation in elastic-plastic solids.

Deterministic microgrinding of precision optical components with rigid, computer-controlled machining centers and high-speed tool spindles is now possible on a commercial scale. Platforms such as the Opticam systems at the Center for Optics Manufacturing produce convex and concave spherical surfaces with radii from 5 mm to ∞, i.e., planar, and work diameters from 10 to 150 mm. Aspherical surfaces are also being manufactured. The resulting specular surfaces have a typical rms microroughness of 20 nm, 1 μm of subsurface damage, and a figure error of less than 1 wave peak to valley. Surface roughness under deterministic microgrinding conditions (fixed infeed rate) with bound abrasive diamond ring tools with various degrees of bond hardness is correlated to a material length scale, identified as a ductility index, involving the hardness and fracture toughness of glasses. This result is in contrast to loose abrasive grinding (fixed nominal pressure), in which surface microroughness is determined by the elastic stiffness and the hardness of the glass. We summarize measurements of fracture toughness and microhardness by microindentation for crown and flint optical glasses, and fused silica. The microindentation fracture toughness in nondensifying optical glasses is in good agreement with bulk fracture toughness measurement methods.

We have designed, constructed, and operated a 3.8-cm clear-aperture Cr:Nd:GSGG active-mirror amplifier. We believe this to be the first active mirror that uses a crystalline host and the largest-aperture Cr:Nd:GSGG amplifier yet reported. We have measured a small-signal gain of 1.6. The wave front has been measured and found to be less than three waves of defocus at repetition rates of up to 10 Hz. Surface displacements were measured and compared with theory. Depolarization was less than 2.5% at maximum power at any location in the clear aperture.