To enhance the capabilities of high-precision optics, the ion beam figuring (IBF) technique has been used for nearly 20 years by a few laboratories and companies over the world. First demonstrated by Wilson et al.1, this type of figuring would later be used on a wider scale to treat various optical materials. The process derives from the creation of local physical etchings made through the collision of accelerated ions with target atoms (the so-called sputtering process), and is accomplished through rastering an ion beam with appropriate computed velocities across the workpiece. As a result, a determined profile is etched.
Following other mechanical polishing methods, IBF is usually performed as the final step to remove the last long spatial wavelength surface errors (hundreds of nanometers). The main advantages of figuring in this manner are the following: it is a deterministic method (non-iterative and therefore timesaving), it does not require contact when performed (useful for lightweight or ultra-thin substrates), and it allows for figuring exotic shapes. The main constraint is that it requires a vacuum environment for operation. Surface roughening2 and heating3 can also restrict the applicability or performance of the technique.
At the Centre Spatial de Liège (CSL), we have been using and developing the IBF technique for more than a decade in order to correct small to medium optics (ranging from 25 to 200 mm in diameter). The facility includes a 1.5 m3 vacuum chamber located in a 10,000-grade clean room.4 The five axes driving the ion source and the substrate mount are software controlled in such a way that the ion beam always scans the workpiece at normal incidence, constant distance, and appropriate velocity. In addition, equipment has been upscaled for a local company to treat optics up to 1-meter diameter. More recently, our ion beam system has also been used to texture surfaces in a deterministic way.
Figure 1. Typical etching profiles measured on chemical vapor deposited (CVD) silicon carbide for two different ion sources (Kaufman and end-Hall). For the Kaufman ion source, three different profiles are given: unmasked and masked beam with 11 mm and 4 mm aperture (mask 1 and 2, respectively).
Efficient IBF correction of optics requires a beam removal function (the etching profile of the ion beam in the material scaled to etching time) that is suitable for the mirror size and the spatial wavelength of the surface errors. Accordingly, we tested numerous approaches to get different beam sizes from our 3-cm Kaufman ion source, including different sized ion grid optics (1, 2, 3 cm), electrostatic focalization, and masking the ion beam. An ion source without a grid (end-Hall type) has also been tested for IBF.5 Typical etching profiles measured on chemical vapor deposited (CVD) silicon carbide for both ion sources and different masks are shown in Figure 1. A combination of different beam profiles often reduces the time to reach the targeted surface quality.
Figure 2. 3D plot of surface errors shows more errors (a) before the IBF technique where rms=243 nm and peak to valley (PV)=1049 nm, than (b) after two IBF runs where the rms=13 nm and PV=86 nm.
Figure 3. Rq, or rms roughness, measured in function of the etching depth for a sample of optical materials (optical profiler, X40 objective), where the roughness of the materials used in optics is shown to be very stable.
An IBF correction on a CVD silicon carbide mirror is shown in Figure 2. After two sweeps, the initial 243 nm rms surface error decreased to 13 nm rms on the 88-mm diameter useful optical aperture. For the first run (∼6h30′), the 3-cm Kaufman ion source full beam was used. For the second run (∼8h30′), a narrower beam obtained with mask 1 was used (see Figure 1). In this case, these steps reduced the process time by nearly a factor of two (with the ion beam set on) and improved by nearly a factor of three the rms surface figure compared to a single run with the narrow beam (mask 1). Figure 2b shows that the large low-frequency errors were removed while initial high-frequency errors (like the cross displayed at the center of the 3D plot) were beyond the beam profile resolution used. Another interesting feature of the figuring technique was the capacity to etch any arbitrary profile into a surface. We applied this principle for the global correction of an optical set-up consisting of several mirrors. Corrections were applied on only one mirror placed at the exit of the set-up.
Another concern for IBF is the material roughening under ion sputtering depending on the material structure and the ion-material interaction. Figure 3 shows the roughness (Rq) evolution measured for different optical materials as a function of the etching depth (650 eV Ar+). The roughness is quite stable for several materials used in optics (glass, CVD silicon carbide, electroless nickel), while it increases very rapidly and strongly for others, such as electroplated metals. Usually polycrystalline materials exhibit a great roughness increase because the different grain orientations have different etching rates. Our work demonstrated that ion beam parameters (energy, gas) can also influence the roughness evolution, notably on CVD silicon carbide.6 IBF can sometimes be applied on unfavorable materials by limiting the correction to a low etching depth or by coating the substrate with a better material. For example, we used three coating/IBF sequences to correct more than 2 microns of a physical vapor deposition (PVD) nickel layer coat on an aluminum mirror.
CSL has worked to improve the IBF technique, notably through developments in beam profile and roughness aspects. Our study of material roughening is the object of on-going research.
Patrick Gailly, Karl Fleury-Frenette, Jean-Hervé Lecat, Jean-Paul Collette, Jean-Marc Defise
Centre Spatial de Liége (CSL)
Université de Liége
Patrick Gailly received a BS in physics from University of Louvain-La-Neuve (1995) and an MS in physics (Opto-electronics) from University of Liège (1996). He joined CSL in 1996 where he has worked mainly on ion beam figuring and other etching techniques. He is currently preparing a thesis on material roughening under ion milling.
Karl Fleury-Frenette graduated from McGill University in physics, obtained a master's degree from Laval University in Qué‚bec, and a PhD after working on magneto-optical thin films at LiéŠge University. He has been leading the Advanced Surfaces Division at CSL since 2000. His main research interests include sputtering processes, optical characterization of surfaces, and the generation of nano-objects and nanostructures for optical applications.