Paul Morantz

Ultra-precision: enabling our future.

This paper provides a perspective on the development of ultra-precision technologies: What drove their evolution and what do they now promise for the future as we face the consequences of consumption of the Earth’s finite resources? Improved application of measurement is introduced as a major enabler of mass production, and its resultant impact on wealth generation is considered. This paper identifies the ambitions of the defence, automotive and microelectronics sectors as important drivers of improved manufacturing accuracy capability and ever smaller feature creation. It then describes how science fields such as astronomy have presented significant precision engineering challenges, illustrating how these fields of science have achieved unprecedented levels of accuracy, sensitivity and sheer scale. Notwithstanding their importance to science understanding, many science-driven ultra-precision technologies became key enablers for wealth generation and other well-being issues. Specific ultra-precision machine tools important to major astronomy programmes are discussed, as well as the way in which subsequently evolved machine tools made at the beginning of the twenty-first century, now provide much wider benefits.

Subsurface damage in precision ground ULE(R) and Zerodur(R) surfaces.

The total process cycle time for large ULE® and Zerodur® optics can be improved using a precise and rapid grinding process, with low levels of surface waviness and subsurface damage. In this paper, the amounts of defects beneath ULE® and Zerodur® surfaces ground using a selected grinding mode were compared. The grinding response was characterised by measuring: surface roughness, surface profile and subsurface damage. The observed subsurface damage can be separated into two distinct depth zones, which are: ‘process’ and ‘machine dynamics’ related.

Sub-surface damage issues for effective fabrication of large optics

A new ultra precision large optics grinding machine, BoX® has been developed at Cranfield University. BoX® is located at the UKs Ultra Precision Surfaces laboratory at the OpTIC Technium. This machine offers a rapid and economic solution for grinding large off-axis aspherical and free-form optical components. This paper presents an analysis of subsurface damage assessments of optical ground materials produced using diamond resin bonded grinding wheels. The specific materials used, Zerodur® and ULE® are currently under study for making extremely large telescope (ELT) segmented mirrors such as in the E-ELT project. The grinding experiments have been conducted on the BoX® grinding machine using wheels with grits sizes of 76 μm, 46 μm and 25 μm. Grinding process data was collected using a Kistler dynamometer platform. The highest material removal rate (187.5 mm3/s) used ensures that a 1 metre diameter optic can be ground in less than 10 hours. The surface roughness and surface profile were measured using a Form Talysurf. The subsurface damage was revealed using a sub aperture polishing process in combination with an etching technique. These results are compared with the targeted form accuracy of 1 μm p-v over a 1 metre part, surface roughness of 50-150 nm RMS and subsurface damage in the range of 2-5 μm. This process stage was validated on a 400 mm ULE® blank and a 1 metre hexagonal Zerodur® part.

Precision grinding for rapid fabrication of segments for extremely large telescopes using the Cranfield BoX

An ultra precision large optics grinding machine, BoX®, was developed and produced at Cranfield University. BoX® offers a rapid and economic solution for grinding large off-axis aspherical and free-form optical components. Grinding high accuracy surfaces with low subsurface damage reduces subsequent polishing time. This efficient grinding process provides the capacity to grind 1.5 m parts. This paper presents an analysis of Astrositall® optical ground parts: a hexagonal 84 m radius of curvature mirror of 1 m across corners and an off-axis 350 mm diameter mirror. The 1 m hexagonal part is representative of segments under study for making extremely large telescope (ELT) segmented mirrors. The second part was machined off-axis to demonstrate free-form fabrication capability. These operations demonstrate the scalability of the rapid grinding process developed for large free-form optics. The use of an error compensation procedure improved an initial ground form accuracy to +/- 1 μm p-v over 1 metre surface. The results highlighted the effect of grinding parameters and machine dynamics on form accuracy and fabrication time.

Wheel wear and surface/subsurface qualities when precision grinding optical materials

An ultra precision large optics grinder, which will provide a rapid and economic solution for grinding large off-axis aspherical and free-form optical components, has been developed at Cranfield University. This paper presents representative grinding experiments performed on another machine - a 5 axes Edgetek - in order to verify the proposed BoX(r) grinding cycle. The optical materials assessed included; Zerodur(r), SIC and ULE(r), all three being materials are candidates for extreme large telescope (ELT) mirror segments. Investigated removal rates ranged from 2mm3/s to 200mm3/s. The higher removal rate ensures that a 1 metre size optic could be ground in less than 10 hours. These experiments point out the effect of diamond grit size on the surface quality and wheel wear. The power and forces for each material type at differing removal rates are presented, together with subsurface damage.

Estimates of the difference between thermodynamic temperature and the International Temperature Scale of 1990 in the range 118 K to 303 K

Using exceptionally accurate measurements of the speed of sound in argon, we have made estimates of the difference between thermodynamic temperature, T, and the temperature estimated using the International Temperature Scale of 1990, T90, in the range 118 K to 303 K. Thermodynamic temperature was estimated using the technique of relative primary acoustic thermometry in the NPL-Cranfield combined microwave and acoustic resonator. Our values of (T−T90) agree well with most recent estimates, but because we have taken data at closely spaced temperature intervals, the data reveal previously unseen detail. Most strikingly, we see undulations in (T−T90) below 273.16 K, and the discontinuity in the slope of (T−T90) at 273.16 K appears to have the opposite sign to that previously reported.

Plasma surface figuring of large optical components

Fast figuring of large optical components is well known as a highly challenging manufacturing issue. Different manufacturing technologies including: magnetorheological finishing, loose abrasive polishing, ion beam figuring are presently employed. Yet, these technologies are slow and lead to expensive optics. This explains why plasma-based processes operating at atmospheric pressure have been researched as a cost effective means for figure correction of metre scale optical surfaces. In this paper, fast figure correction of a large optical surface is reported using the Reactive Atom Plasma (RAP) process. Achievements are shown following the scaling-up of the RAP figuring process to a 400 mm diameter area of a substrate made of Corning ULE®. The pre-processing spherical surface is characterized by a 3 metres radius of curvature, 2.3 μm PVr (373nm RMS), and 1.2 nm Sq nanometre roughness. The nanometre scale correction figuring system used for this research work is named the HELIOS 1200, and it is equipped with a unique plasma torch which is driven by a dedicated tool path algorithm. Topography map measurements were carried out using a vertical work station instrumented by a Zygo DynaFiz interferometer. Figuring results, together with the processing times, convergence levels and number of iterations, are reported. The results illustrate the significant potential and advantage of plasma processing for figuring correction of large silicon based optical components.

Manufacturing and measurement of the miri spectrometer optics for the james webb space telescope

The Mid-Infrared Instrument (MIRI) on the JWST, Hubble space telescope replacement, contains an integral field spectrometer utilising four integral field units (IFU). This paper introduces manufacturing and measurement techniques developed to produce the monolithic multi-mirror arrays which perform spatial splitting and reformatting of light within each IFU: the image slicers and re-imagers. Typical slicer components have 18-22 mirrors of 50 nm RMS form accuracy, 5 nm Rq roughness and relative spatial positioning between mirror facets to 20 μm. The paper provides data for the Verification Model slicers produced by Cranfield University for the JWST MIRI consortium.

Pyknometric volume measurement of a quasispherical resonator

We have measured the internal volume of a 1 litre, diamond-turned copper quasispherical resonator with a fractional uncertainty of approximately 1 part in 106 using two independent techniques. This is in response to the need for a uniquely accurate measurement of resonator volume, for the purpose of measuring the Boltzmann constant in pursuit of the redefinition of the kelvin. The first technique is a pyknometric measurement using water as a liquid of known density. We describe the development of a procedure that results in stable, reproducible volume measurements. We provide a detailed discussion of the factors that affect the water density, such as dissolved gases. The second technique is microwave resonance spectroscopy. Here, we measure the resonant frequencies of the TM1n modes and relate them to the dimensions of the resonator. We evaluate the frequency perturbations that arise from the coupling waveguides and the electrical resistivity of the copper surface. The results of the microwave measurements show evidence of a dielectric coating on the surface. We propose that this is an oxide layer and estimate its thickness from the microwave data. Finally, we compare the volume estimates from the two methods, and find that the difference is within the combined uncertainty.

Reactive Atom Plasma for Rapid Figure Correction of Optical Surfaces

Nanometre-scale figuring technique at atmospheric pressure for large optical surfaces is a high profile research topic which attracts numerous competing state-of-the-art technologies. In this context, a dry chemical process, called Reactive Atom Plasma (RAP), was developed as a prospectively ideal alternative to CNC polishing or Ion Beam Figuring. The RAP process combines high material removal rates, nanometre level repeatability and absence of subsurface damage. A RAP figuring facility with metre-scale processing capability, Helios 1200, was then established in the Precision Engineering Centre at Cranfield University. The work presented in this paper is carried out using Helios 1200 and demonstrates the rapid figuring capability of the RAP process. First experimental tests of figure correction are performed on fused silica substrates over 100 mm diameter areas. A 500 nm deep spherical hollow shape is etched onto the central region of 200x200 mm polished surfaces. The test is carried out twice for reproducibility purposes. After two iterative steps, a residual figure error of ~16 nm rms is achieved. Subsequently, the process is scaled up to 140 mm diameter areas and two tests are carried out. First, the developed algorithm for 500 nm deep spherical hollow test is confirmed. Residual deviation over processed area is ~18 nm rms after three iterations. Finally, a surface characterised by random topography (79 nm rms initial figure error) is smoothed down to ~ 16 nm rms within three iteration steps. All results presented in this paper are achieved by means of an in-house developed tool-path algorithm. This can be described as a staggered meander-type tool motion path specifically designed to reduce heat transfer and consequently temperature gradient on the surface. Contiguously, classical de-convolution methods are adapted to non-linear etching rates for the derivation of the surface scanning speed maps. The figuring procedure is carried out iteratively. It is noteworthy that iteration steps never exceed ~7 minutes mean processing time.