Magnetorheological Jet Finishing of Conformal, Freeform and Steep Concave Optics

M. Tricard1 (3), W.I. Kordonski1, A.B. Shorey1

QED Technologies, Inc., Rochester, USA

Submitted by C. Evans (1), Middleeld, USA

Abstract Conformal, freeform and steep concave optics represent important shapes that are difcult or impossible to nish using conventional techniques due to mechanical interferences and steep local slopes. One way to polish these optics is by using a jet of abrasive/uid mixture. Widely used abrasive water jet machining is not applicable for precision polishing because of natural jet instability, which gives an unstable removal function. Theoretical and experimental results in this paper show how this problem can be addressed with a magnetically stabilized jet of magnetorheological uid. Polishing results demonstrate the suitability for this technique for precision nishing of complex shapes. Keywords: Finishing; Precision; Optical

INTRODUCTION

precision nishing had not been reported. To be successful in deterministic, high-precision nishing, a stable, relatively high-speed, low-viscosity uid jet, which remains collimated and coherent before it impinges the surface, is required. Such a unique tool may also resolve the challenging problem of high precision nishing of steep concave surfaces and cavities. 2 MAGNETORHEOLOGICAL JET FINISHING

Evans et al. recently gave an overview of the mechanisms of removal for a variety of nishing techniques [1] including methods based on uid ow such as elastic emission machining (EEM), which can be thought of as a uid jet polishing with grazing incidence, and magnetorheological nishing (MRF) [2]. Successful implementation of such methods for deterministic nishing depends on a stable material removal footprint, and hence, on stable uid ow over the polishing zone. Here we discuss a new polishing process also based on a uid ow: MR Jet nishing. Conventional polishing jets contain abrasive particles with kinetic energy high enough to spall (chip) material away from the surface in the impact zone (see Hoogstrate et al. for instance) [3]. At the same time, a fundamental property of a uid jet is that it begins to lose its coherence as the jet exits a nozzle, due to a combination of abruptly imposed longitudinal and lateral pressure gradients, surface tension forces, and aerodynamic disturbance. The aerodynamic disturbance is dramatically increased with the jet velocity, resulting in instability. A high-speed liquid jet, which is the prime interest for nishing, breaks into droplets and progressively spreads out as the distance from the nozzle increases. All of this results in instability of the ow over the impact zone and, consequently, polishing spot instability, which is unacceptable for deterministic, high-precision nishing. Increasing jet stability by increasing uid viscosity will proportionally increase the resistance to uid ow in the delivery system and the pumping power required to deliver the uid to the nozzle. It makes impractical a high-speed, high-viscosity jet for polishing. Erosion has been widely studied by tribologists, e.g., Finnie [4]. It has been shown that abrasive water jets can be used to work materials such as glass, diamond, ceramics, stainless steel and alloys. The surface roughness in these processes has been reported at > 100 nm Ra [5]. Proper adjustment of process parameters such as jet velocity, abrasive size and concentration makes some reduction of surface roughness possible. At the same time, the use of water jets for high

2.1 Jet stabilization In contrast to conventional abrasive jet polishing methods where material removal relies on the kinetic energy of impacting particles, the technique discussed in this paper is based on an assumption that the energy required for polishing may be supplied by the radial spread of a liquid jet over a surface to be polished [6-11]. Such uid ow may generate sufcient surface stress to provide the regime of material removal, which is characteristic of polishing. As mentioned above, a typical liquid jet breaks down at a very short distance from the nozzle (a few diameters), making regular, jet-based polishing techniques highly sensitive to the nozzle-offset distance. This limits polishing congurations to those in which the removal function is stable, signicantly restricting the nishing of complex shapes. A method of jet stabilization has been proposed, developed, and demonstrated whereby the round jet of magnetorheological uid is magnetized by an axial magnetic eld when it ows out of the nozzle [6-11]. This local magnetic eld induces a structure within the uid [12] and high apparent viscosity [13] within the portion of the jet that is adjacent to the nozzle, resulting in the suppression of all of the most dangerous initial disturbances. As a result, the MR uid ejected from the nozzle denes a highly collimated, coherent jet. The stabilizing structure induced by the magnetic eld within the jet gradually begins to decay as the jet passes beyond the eld. However, the remnant structure still suppresses disturbances and, thus, consequent stabilization of the jet can persist for a sufcient

Annals of the CIRP Vol. 55/1/2006

ow direction

Figure 3: Example of a spot taken on fused silica glass. 0.01

depth (m)

-0.01 -0.03 -0.05 -0.07 -5 -2.5 0 2.5 5

(a) Water

(b) MR Fluid Magnet Off

(c) MR Fluid Magnet On

position (mm)

Figure 4: MR Jet spot proles of spots taken with 50 mm (solid line) and 150 mm (markers) stand-off distance. spindle of the CNC platform and stabilizes the jet. With the magnet activated, a collimated jet is directed vertically upwards to the workpiece held by the spindle. Used uid is collected and recirculated. 2.3 Removal function A typical MR Jet removal function (or polishing spot) is shown in Figure 3. This spot was obtained by dwelling a 1.5 mm diameter jet on a stationary, at, fused silica surface. The distance between the nozzle exit and the part surface (the stand-off distancesee Figure 2) was 50 mm. The stability of the removal function is demonstrated in Figures 4 and 5. Figure 4 shows removal function proles typical for MR Jet polishing spots. (The scan is a prole through the spot center, as shown by the dashed line in Figure 3.) The two spot proles given in this gure were taken at two different stand-off distances, 50 mm (solid line) and 150 mm (markers), but have the same removal function prole, as shown in this gure. The box represents the jet position and diameter directed toward the horizontal axis. This insensitivity to part stand-off is further emphasized in Figure 5, where the two-dimensional maps of the material removal distribution are subtracted. The maximum variation in these removal maps from spots taken with a 100 mm difference in stand-off distance is only 6.5%. This insensitivity to large stand-off distances is particularly important when considering precision nishing of steep concave optics. Figures 3-5 demonstrate important features of the removal mechanism in MR Jet nishing. The prole in Figure 4

Figure 1: Jet snapshot images (velocity = 30 m/s, nozzle diameter = 2 mm). time that it may travel up to several meters (depending on the jet diameter) without signicant spreading or loss of structure. For nozzle sizes on the order of 1 mm in diameter, coherent jets of 0.5 m are routinely achieved. This is illustrated in Figure 1. In the case of water, the jet remains stable only for ~ 2 nozzle diameters. (Note the transparent section inside dashed ovals at the outlet of Figure 1a.) MR uid has higher viscosity and therefore, the coherent part of the jet extends on ~ 78 diameters. Initial disturbances (visible in the form of ripples on the surface of the coherent part of the jet in Figure 1b) eventually result in the jet breaking down and rapid spreading. However, when magnetized at the outlet, the jet of MR uid remains coherent for more than 200 diameters (i.e., ~ 0.4 meters for the case shown in Figure 1c). 2.2 Experimental set up An MR Jet nishing system, a portion of which is shown in Figure 2, has been constructed using a 5-axis CNC platform and polishing control software. The delivery system is comprised of a mixing vessel to disperse the solids in the MR uid, a pump, means to maintain temperature and viscosity of the uid and pressure and ow sensors to monitor the system conditions. A magnetic eld optimized for the properties of the MR uid is located beneath the

Spindle MR Jet Stand-off MR Shaper (a) 50 mm offset (b) 150 mm offset (c) Subtraction

Figure 2: Experimental set-up.

Figure 5: Subtraction of the full 2-D material removal maps (spots) from Figure 4 showing insensitivity to stand-off distance. The difference error is dominated by metrology noise.

shows that there is almost no material removal in the center of jet impingement. (See the boxed region in Figure 4.) Detailed ow analysis shows that this removal prole is consistent with the distribution of the rate of work done (the scalar product of the local shear stress and velocity) as predicted by Prestons model [9]. This is further illustrated in Figure 6. Experimental data for removal rate (square markers) were taken at a jet velocity of 20 m/s with a jet diameter of 2.4 mm. The solid line represents the radial distribution of the computed rate of the work done by uid at the surface under the same conditions. The open triangles show the calculated pressure prole. Very good correlation is observed between the removal rate prole and distribution of the rate of work done at the surface. It is also worth noting that the position of the peak removal rate corresponds to the position of minimum pressure whereas no removal occurs at maximum pressure. This means that the normal stress does not contribute to material removal as the model suggests. The collimated jet provides a stable, laminar shear ow over the impact zone, resulting in a stable material removal function that is independent of stand-off distance. This differs from the jet removal mode where the material removal results from the bombardment (kinetic energy) of particles.

Figure 7: PCA dome polished with MR Jet.

1.0 Normalized Rate of Work, Removal Rate, Pressure 0.8 0.6 0.4 0.2 0.0 0.000

(a) Initial gure error PV = 1740 nm, rms = 255 nm.

(b) Final gure error after MR Jet nishing PV = 133 nm, rms = 12.1 nm.

Figure 8: The initial gure error (a) and the nal gure error (b) after polishing the inside of the glass dome with MR Jet. a stitching approach as implemented in the Subaperture Stitching Interferometer (SSI) [14,15] can provide this fullaperture metrology straightforwardly. The SSI provides a means to collect a number of sub-aperture regions, and stitch them together to provide continuous characterization of the entire surface.

0.005 Radial Distance (m)

0.010

Figure 6: Theoretical and experimental removal rate and rate of work proles. 2.4 Polishing Performance Given the stable spot, appropriate control software, and CNC control, MR Jet has been shown to efciently correct surface gure error to high precision, with less than 30 nm peak-to-valley (PV) error (2 nm rms), in few iterations, with microroughness of < 1 nm rms on a number of materials [6,9,10]. Most of our experience to date with MR Jet has been for optical applications with glass materials, single crystals and some glass mold materials such as tungsten carbide (WC). More recently, this technique has also been successfully used on a stainless steel component. MR Jet provides a particular advantage in nishing complex shapes such as concave surfaces that are difcult to reach using conventional techniques. One such application is the nishing of hemispherical and conformal domes on optical seekers [10]. Figure 7 shows one such dome made from polycrystalline alumina (PCA) where the concave surface (62 mm diameter, 32 mm radius of curvature) was polished with MR Jet. Since MR Jet is a deterministic polishing process, it requires that the gure error be characterized over the full aperture of the surface to be polished as input to the control software. While this is typically a challenge for hemispherical domes (or even steeply convex surfaces),

The gure error of the concave surface of the dome in Figure 7 was characterized in the SSI, and then polished using MR Jet. Figure 8a shows the initial surface error, and Figure 8b shows the error after polishing with MR Jet. The gure was improved more than 20xfrom 255 nm rms to 12.1 nm rms in only 2 iterations. MR Jet, in conjunction with stitching interferometry, is also well-suited to correct lobing errors caused by the crystal structure of sapphire domes, or to improve thickness uniformity, for example. Another example of a unique geometry that has been nished with MR Jet is shown in Figure 9. In this case, an optical at surface at the bottom of a ~10 mm deep cavity was polished. This is nearly impossible to accomplish with most nishing techniques. As shown in Figures 10a and 10b, MR Jet was able to nish this surface to improve the gure from >73.4 nm rms to ~6 nm rms.

Figure 9: Schematic of a cavity that was nished with MR Jet.

5 [1]

REFERENCES Evans, C., Paul, E., Dorneld, D., Lucca, D., Byrne, G., Tricard, M., Klocke, F., Dambon, O., Mullany, B. A., 2003, Material Removal Mechanisms in Lapping and Polishing, Annals of the CIRP, 52/2:611-633. Golini, D., Jacobs, S., Kordonski, W., Dumas, P., 1997, Precision Optics Finishing Using Magnetorheological Finishing, Proc. SPIE, CR67:251-274. Hoogstrate, A., Karpuschewski, B., van Luttervelt, C., Kals, H., 2005, Modeling of High Velocity, Loose Abrasive Machining Processes, Annals of the CIRP, 51(1):263-266. Finnie, I., 1995, Some Reections on the Past and Future of Erosion, Wear, 186:1-10. Momber, A., Kovacevic, R., 1998, Principles of Abrasive Water Jet Machining, 9.6, Springer, New York. Kordonski, W., Shorey, A.B., Sekeres, A, 2004, New Magnetically Assisted Finishing Method: Material Removal with Magnetorheological Fluid Jet, Proc. SPIE, 5180:107-114. Kordonski, W., Golini, D., Hogan, S., Sekeres, A., 1999, System for Abrasive Jet Shaping and Polishing of Surface Using Magnetorheological Fluid, US Patent #5,971,835. Kordonski, W., 2003, Apparatus and Method for Abrasive Jet Finishing of Deeply Concave Surfaces Using Magnetorheological Fluid, US Patent #6,561,874. Kordonski, W. Shorey, A., Tricard M., 2006, Magnetorheological Jet (MR Jet) Finishing Technology, Journal of Fluids Engineering 128(1):2026.

[2] (a) Initial gure error: PV = 304 nm, rms = 73.4 nm.

[3]

[4] [5]

(b) Final gure error: PV = 47 nm, rms = 6 nm. Figure 10: The initial gure error (a) and nal gure error (b) after polishing the bottom of the cavity in Figure 9 with MR Jet. 3 SUMMARY It has been demonstrated that impingement of a magnetically stabilized, collimated jet of MR uid induces radial surface ow, which results in generation of a polishing spot. It was also established that MR Jet nishing can produce highprecision surfaces on the order of tens of nanometers peak-to-valley with roughness < 1 nm rms on a number of materials such as glasses, single crystals, advanced ceramics and metals. Due to the insensitivity to the offset distance, this technique provides a versatile tool for nishing complex shapes, including those used for molds, tooling, prosthetics and other challenging shapes for the medical industry, as well as conformal optics in next generation imaging systems. As an example, the ability to nish the concave surface of a spherical dome and the bottom of a at cavity to extremely high precision was demonstrated. 4 ACKNOWLEDGEMENTS

[6]

[7]

[8]

[9]

[10] Kordonski, W., Shorey, A., Tricard, M., 2005 JetInduced High-Precision Finishing of Challenging Optics, Proc. SPIE, 5869:44-51. [11] Shorey, A., Kordonski, W., Tricard, M., 2005, Deterministic Precision Finishing of Domes and Conformal Optics, Proc. SPIE, 5786:310-318. [12] Graselli, Y., Bossis, G., Lemaire, E., 1993, FieldInduced Structure in Magnetorheological Suspensions, Prog. Colloid Polym. Sci. 93:175-177. [13] Shorey, A.B., Kordonski, W.I., Gorodkin, S.R., Jacobs, S.D., Gans, R.F., Kwong, K.M., Farny, C.H., 1999, Design and Testing of a New Magnetorheometer, Review of Scientic Instruments, 70(11):4200-4206. [14] Murphy, P., Forbes, G., Fleig, J., Dumas, P., Tricard, M., 2003, Stitching Interferometry: a Flexible Solution for Surface Metrology, Optics and Photonics News 14/5:38-43. [15] Fleig, J., Dumas, P., Murphy, P., Forbes, G., 2003, An Automated Subaperture Stitching Interferometer Workstation for Spherical and Aspherical surfaces, Proc. SPIE, 5188: 296-307.

Authors appreciate valuable inputs from Don Golini, Arpad Sekeres, Justin Tracy and other colleagues at QED. In addition, we acknowledge the contributions of Professor S. Gracewski from University of Rochester Department of Mechanical Engineering and the Center for Optics Manufacturing. Portions of the research reported in this presentation were performed in connection with contract DAAD17-03-C0055 with the U.S. Army Research Laboratory (Aberdeen, MD) and contract No. N68936-05-C-0003 from Naval Air Warfare Center Weapons Division (China Lake, CA). The views and conclusions contained in this document are those of the authors and should not be interpreted as presenting the ofcial policies or position, either expressed or implied, of the U.S. ARL, the NAWCW Division, or the U.S. Government unless so designated by other authorized documents. Citation of manufacturers or trade names does not constitute an ofcial endorsement or approval of the use thereof. The U.S. Government is authorized to reproduce and distribute reprints for Government purposes notwithstanding any copyright notation hereon.