Precision engineering for astronomy and gravity science

CIRP Annals - Manufacturing Technology 59 (2010) 694–716

Contents lists available at ScienceDirect

CIRP Annals - Manufacturing Technology jou rnal homep age : ht t p: // ees .e lse vi er. com/ci rp/ def a ult . asp

Precision engineering for astronomy and gravity science P. Shore (2)a,*, C. Cunningham b, D. DeBra (1)c, C. Evans (1)d, J. Hough e, R. Gilmozzi f, H. Kunzmann (1)g, P. Morantz a, X. Tonnellier a a

CUPE, Cranfield University Precision Engineering, Bedford, England, UK UK ATC, Royal Observatory, Edinburgh, Scotland, UK Department of Aeronautics and Astronautics, Stanford University, USA d Zygo Corporation, Middlefield, CT, USA e University of Glasgow, Glasgow, Scotland, UK f European Southern Observatory, Garching, Germany g PTB, Braunschweig, Germany b c

A R T I C L E I N F O

A B S T R A C T

Keywords: Telescopes Precision Machines

The fields of astronomy and gravitational science have presented significant precision engineering challenges. Numerous solutions for these fields of science have achieved unprecedented levels of accuracy, sensitivity and sheer scale. Notwithstanding of their importance to science understanding, many of these precision engineering developments have become key enabling technologies for wealth generation and other human well-being issues. This paper provides a brief historical overview of astronomy and gravitational instruments. Later, details of critical precision engineering developments that supported the establishment of leading astronomical and gravitational instruments are illustrated. Details of specific developments having wider application to the benefit of mankind are provided. Finally, significant precision engineering demands to enable future science programmes are introduced. ß 2010 CIRP.

1. Introduction It has long been the dream of humankind to understand its place within the universe. This aspiration has led to detailed measurements which accurately described to movement and interactions of celestial bodies. From the advent of astronomical instruments it has been clear that these should have high effective resolving power, which is essentially a function of both; scale (e.g. diameter of telescope) and quality (e.g. accuracy of mirror surfaces). From experience these simultaneous demands are known to be difficult to reconcile. This dual requirement is at the centre of ‘‘Precision Engineering’’ and fundamentally coupled to being able to measure with high accuracy and with minimal levels of uncertainty. The maxim ascribed to Galileo tells us to ‘‘measure what is measurable, make measurable what is not’’. 1.1. Objective Scientific progress made in the fields of astronomy and gravitational research has been achieved through significant precision engineering developments. Numerous engineering solutions, that have advanced these fields of science, have demanded unprecedented levels of accuracy, sensitivity and sheer scale. Notwithstanding of their importance to science understanding many of these precision engineering developments have

* Corresponding author. E-mail address: paul.shore@cranfield.ac.uk (P. Shore). 0007-8506/$ – see front matter ß 2010 CIRP. doi:10.1016/j.cirp.2010.05.003

become key enabling technologies for wealth generation and other human well-being issues. In this paper we provide a brief historical overview of astronomical and gravity wave instrument development. Instrument developments (those advancing sensitivity) and some related key scientific findings are introduced. Thereafter critical precision engineering developments that have progressed astronomical and gravitational wave instruments are presented. The paper subsequently provides details of specific developments having found wider application to the benefit of mankind. In summary the authors propose some necessary key precision engineering capabilities demanded for future planned science programmes. 1.2. Astronomical instrument development From the beginning of astronomy as a science, discoveries have been made following the introduction of new manufacturing technology. Considering Galileo’s telescope, built over 400 years ago in 1609, the key manufacturing technology was the newly developed ability to fabricate glass lenses precisely, together with some critical innovations from Holland such as ‘‘stopping down the aperture’’ for enhanced seeing performance [43,201]. Using his telescope Galileo discovered moons around Jupiter and consequently reinforced support for the Copernician model (Fig. 1). James Gregory and Isaac Newton’s reflecting telescope concepts were enabled in the 1660s through the combined techniques of grinding and polishing metal mirrors in speculum [41,83,138,153,184]. These precision machining techniques

[(Fig._1)TD$IG]

[(Fig._2)TD$IG]

P. Shore et al. / CIRP Annals - Manufacturing Technology 59 (2010) 694–716

Fig. 1. Galileo’s refraction telescope and Newton’s reflecting telescope.

enabled William Herschel to build his telescope which discovered Uranus and numerous nebulae [19,184]. Le´on Foucault invented the metallised-glass reflecting telescope, producing a 80 cm telescope in Marseille in 1864 [182]. By employing a ‘‘knife edge’’ measurement test during the mirror manufacturing (in-process metrology) a more controlled machining process was attained [61]. The silvered glass reflecting telescope became the preferred technology for large ground based telescopes, particularly after the limits of refractors were reached. Following a suggestion of Newton nearly 200 years earlier Charles Piazzi Smyth in 1880 carried out observations from the top of Mount Teide, Tenerife [157]. Smyth’s observations demonstrated the benefits of observation made through a reduced atmospheric layer. From that time onwards major ground based telescopes were positioned on mountain tops; the first notable example was the Lick Observatory [111]. During the first half of the 20th century the major development of telescopes was that of size scale up. Large reflecting telescopes by George Ritchey and George Ellery Hale led to reflectors of up to 200 in. (5 m) in size [1,151]. The 200 in. Hale telescope was based on low expansion Pyrex mirrors, ship building structural engineering and passive flexure compensators. In 1929, using observations from the first of these large Californian telescopes, Edwin Hubble confirmed the expansion of the Universe [158]. For the majority of the latter half of the 20th century telescope performance was most significantly advanced through implementation of electronic array detectors replacing photographic plates in visible and IR. Through the greater sensitivity and linearity of charge coupled devices (CCDs) and IR arrays, it was possible to improve the performance of existing 2–4 m scale telescopes. Better detectors achieved an order of magnitude improvement of sensitivity during this time [102,116,117]. Computer control of a telescope was pioneered on the Anglo-Australian Observatory in 1975 [43]. The implementation of computer control paved the way to actively controlled telescopes and ultimately the 8–10 m telescope era of today. The Russian 6 m BTA telescope was one of the first to rely solely on computer control. In 1923 Hermann Oberth proposed the benefits of space based astronomy. Lyman Spitzer in 1946 presented a seminal paper that defined the wider attributes and capabilities of space based telescopes [167,168]. The launch of Sputnik in 1957 ‘‘changed everything’’ [106]. It led to the birth of NASA, the race to the moon and the consequent availability of space technology that led to the great NASA Observatories such as the Chandra X-ray telescope and the 2.4 m Hubble optical/near IR telescope. These NASA space telescopes, together with ESA’s ISO and XMM-Newton telescopes, have produced a broad range of discoveries and provided beautiful images that have captured wide public interest [11,21,92,193] (Fig. 2). By the late 1980s it was generally recognised that the scope for improved sensitivity of ground based telescopes was again only to be afforded through an increase in size. This situation was brought

695

Fig. 2. Hubble space telescope.

about through spectacular improvement of detector sensitivity during the 1970s and 1980s together with improved positioning and motion control. The late 1980s saw astronomy entering the era of the 8–10 m telescopes [193]. Numerous enabling technologies that led to the very large telescope era (VLT, 8–10 m scale telescopes) will be covered in more detail later in this paper. Those considered having historical significance include: segmented primary mirror technology, applied to the Keck and Grantecan telescopes, as proposed and established by Jerry Nelson [136]; and 8 m scale monolithic honey comb spun cast mirror moulding technique developed by Roger Angel and applied to the Magellan Telescope and the Large Binocular telescope [9,81,119,131] (Fig. 3). In parallel, an 8 m mirror capability was also established in France during the 1980s. It is based around a meniscus mirror fabrication capability and use of active optics technology. This technology, developed by REOSC, was first employed by ESO on the 3.6 m New Technology Telescope and subsequently the four 8 m very large telescopes [48,53,67]. Similar technology was subsequently employed for the Japanese Subaru telescope [97]. During the late 1990s and the first decade of the 21st century the 8–10 m scale ground based telescope technology matured. Adaptive Optics, discussed later on, significantly advanced the ‘‘seeing’’ performance of Keck, Gemini and the VLTs producing revolutionary high resolution images of gas giant planetary systems (much like Jupiter) and provided evidence of a massive black hole near the centre of our galaxy [55]. However, as was seen during the first half of the 20th century the main demand of 21st century astronomers is increase in telescope size for ground and space based systems [5,47,71,174]. And as will be shown later this also applies to gravitational wave

[(Fig._3)TD$IG]

Fig. 3. Keck telescopes (10.2 m segmented primary mirror).

[(Fig._4)TD$IG]

696

[(Fig._6)TD$IG]

P. Shore et al. / CIRP Annals - Manufacturing Technology 59 (2010) 694–716

Fig. 6. James Webb space telescope [91]. Fig. 4. Historical size increase of telescopes [72].

detection. The historical increase of telescope size is illustrated in Fig. 4. The detection of so-called earth-like planets, presently a prime objective, is a central justification for the proposed US/Japanese Thirty Metre Telescope and the European-ELT 42 m telescope projects [73,123,134,172] (Fig. 5). The so-called extremely large telescopes (ELTs) are presently in the design and preliminary manufacturing assessment stages [52]. These telescopes are targeted for ‘‘first light’’ by 2017–2020, at which time the pioneering NASA led segmented mirror-based James Webb Space Telescope (JWST) having a 6.5 m diameter will be operational. Used in combination the proposed ground based ELTs and the JWST will yield a new era of astronomical observation [91] (Fig. 6). Pivotal Precision Engineering developments leading up to and enabling this new era of astronomical observation is discussed in Section 2. 1.3. Gravitational wave instrument development Gravitational waves were indirectly confirmed to exist by Hulse and Taylor through radio emission frequency observations of a binary pulsar in 1973 [87]; both were awarded a Nobel Prize in 1993. However, a direct means to detect gravitational waves has not yet been achieved. Such an achievement would open up a new branch of astronomy, complementing electromagnetic based observation. A short historical overview of gravitational wave detection shows that it also demands unprecedented achievements in precision engineering, requiring measurement sensitivity levels quite unfamiliar to most Precision Engineers.

[(Fig._5)TD$IG]

Fig. 5. European extremely large telescope (42 m segmented primary mirror).

Newton presented his universal law of gravitation in 1687 [137]. Later in 1798 Cavendish [33] performed experiments to obtain an accurate value for the Gravitational constant. It was not until Einstein provided his general theory of relativity however that the cause of gravity became better appreciated – in 1915. Einstein’s general theory of relativity showed that the effects of gravity were caused as a response to distortion of space and time [50,51]. Stars and planets effectively warp space-time. This phenomenon is known as the ‘geodetic effect’ as illustrated in Fig. 7. Consequently large masses such as stars and galaxies deform space-time around them. As other objects move through these areas, they are diverted from their original path; apparently they are attracted towards the large mass. Moving masses give rise to perturbations in space-time, these moving space-time disturbances are called gravitational waves (radiation). The distortion of space-time through the rotation of bodies is also considered a twisting around the body, it is referred to as ‘‘Frame dragging’’ [109,180]. Nevertheless it was 45 years after their prediction by Einstein that significant interest in the experimental detection of gravitational radiation arose. In the late 1950s Weber of the University of Maryland suggested the design of some relatively simple apparatus for their detection (Fig. 8). This apparatus in its later stages consisted of an aluminium bar of mass approximately one ton with piezoelectric transducers bonded around its centre line and suspended from anti-vibration mountings in a vacuum tank. By means of the amplified electrical signals from the transducers, Weber monitored the amplitude of oscillation of the fundamental mode of the bar. A gravitational [(Fig._7)TD$IG]

Fig. 7. Representation of the geodetic and frame dragging effects [95].

[(Fig._8)TD$IG]

[(Fig._9)TD$IG]

P. Shore et al. / CIRP Annals - Manufacturing Technology 59 (2010) 694–716

697

Fig. 9. Early Weiss laser interferometer GW detector.

Fig. 8. Weber bar [113].

wave signal of suitable strength would be expected to change the amplitude or phase of the oscillations in the bar. By the late 1960s Weber operated two such systems: one at the University of Maryland and one at the Argonne National Laboratory and reported that he observed coincident excitations of the bars at a rate of one event per day [189]. These events he claimed to be gravitational wave signals. However other experiments – at Moscow State University, IBM Yorktown heights, Bell Labs, Munich and University of Glasgow failed to confirm Weber’s detections. A detailed analysis of the Weber bar detector design suggested that the instruments were unlikely to be sensitive enough to detect any possible sources and that the coincident events were not statistically significant. The field of gravitational wave detection had to find a new way forward, the driving force being the need to improve detector sensitivity. As with electromagnetic telescopes, the two fundamental ways to improve sensitivity were by reducing background noise and by increasing the signal size (scale up). Weber’s early work on developing methods to detect gravitational waves encouraged others to devise potentially more sensitive detector designs. Among these were Gertsenshtein and Pustovoit from Moscow who proposed in 1962 that one could look for small shifts in the fringe pattern in a Michelson interferometer formed between freely hung mirrors [65]. Since gravitational wave signals induce differential strains along perpendicular transverse dimensions in space, the signal size is essentially proportional to the arm length of the Michelson interferometer, provided the arm length is significantly smaller than the wavelength of the gravitational wave. Such an arrangement with arm lengths of 2 m and with rubber isolating stacks was constructed by Robert Forward at the Hughes aircraft corporation in the late 1960s [60]. Illumination was by a few milliwatts of HeNe laser light but with such short arms and low laser power, sensitivity was limited (Fig. 9). Folding of the optical path by means of a Herriott delay line was suggested by Weiss. At MIT he constructed a short interferometric prototype detector using such a delay line in the 1970s. Thereafter, as higher power argon-ion lasers become available, the concept of separated mass detectors employing laser interferometry between the masses separated by significant distances began to look a realistic way forward. A period of experimental system scale up began [85]. A three metre prototype was built in Germany in 1975 using optical delay lines [20]. It was followed by a 30 m instrument

developed in the early 1980s at the Max Planck Institute for Astrophysics in Garching. A 1 m instrument built in the UK in 1977 employed a different form of multibeam optical system, it was followed during the early 1980s by a 10 m interferometer using Fabry–Perot cavities in the arms [152]. A 40 m instrument was then developed in Caltech as a spin off from the 10 m Glasgow instrument. Two prototypes, one using delay lines and the other Fabry–Perot cavities, were also built in Japan. All of these instruments used originally used multi watt argon-ion lasers and the majority achieved displacement sensitivities of better than 10 18 m/HHz over a frequency range of a few hundred Hz to a kHz (Fig. 10). At this stage many countries considered the technology sufficiently mature for construction of detectors of much longer baseline, detectors that should be capable of having a real possibility of detecting gravitational waves. Thus an international network of gravitational wave detectors came into being. In the US, the Laser Interferometer Gravitational-Wave Observatory (LIGO) consists of three multi-kilometer scale interferometers, two located in Hanford, Washington State (of 4 km and 2 km arm length), and one located in Livingston, Louisiana (4 km arms) [112] (Fig. 11). In Europe, Virgo is a 3 km scale interferometer located near Pisa, Italy, and GEO600, a 600 m interferometer, is located near Hannover, Germany. The 300 m TAMA detector is located near Tokyo, Japan. It is hoped that the first direct observation of gravitational waves will be made in the next few years by this international network of ground based detectors. Informed opinion now suggests a sensitivity level of 10 23 m/ HHz over a frequency range of a few hundred Hz to a kHz is necessary for detection of gravitational waves. Consequently major upgrades to LIGO, Virgo and GEO600 will be completed in the next

[(Fig._10)TD$IG]

Fig. 10. LIGO Observatory, Louisiana, USA [112].

[(Fig._1)TD$IG]

698

[(Fig._13)TD$IG]

P. Shore et al. / CIRP Annals - Manufacturing Technology 59 (2010) 694–716

Fig. 11. Basis of LIGO Observatory, Louisiana, USA [112].

5 years [62,63,185]. New gravitational wave detectors in Australia (AIGO) and Japan (LCGT) are also planned (Fig. 12). In combination these new facilities will provide a network with the sensitivity and directionality needed to observe gravitational wave signals at a monthly or even weekly rate and to open the field of gravitational wave astronomy. It will be shown in Section 3 that sensitivity improvements depend on developing mirrors of lower mechanical and optical loss, and improved isolating suspension systems together with non-classical opto-mechanical techniques. Gravitational wave signals resulting from the formation and coalescence of black holes having 103 to 106 solar masses will lie in the frequency range of 10 4 Hz to 10 1 Hz. The strain measurement sensitivity necessary for their detection is also considered to be in the region of 10 23/HHz over the relevant timescales. The most promising way of looking for such signals is to fly a laser interferometer in space. Following work in the early 1980s by Bender and Faller at JILA in Boulder, Colorado, the project LISA (Laser Interferometer Space Antenna) evolved over many years and is being developed as a joint ESA/NASA mission [105] (Fig. 13). LISA will consist of an array of three drag-free spacecraft at the vertices of an equilateral triangle of length of side 5 million km. Drag-free technology relies on sensing the position of the space craft with respect to an internal test mass and servo positioning the craft by means of thrusters to maintain its position essentially constant, with respect to the test mass. This space based positioning technology was originally demonstrated on the TRIAD spacecraft, flown by the US Navy in 1972. It has undergone continuous development, most notably for the NASA Gravity probe B mission flown in 2004/5. A number of notable Precision Engineering achievements of the Gravity probe B project are discussed in Section 3.

[(Fig._12)TD$IG]

Fig. 13. Laser Interferometer Space Antennae [105].

The proposed LISA cluster will be placed in an Earth-like orbit (1 AU) and positioned 208 behind the Earth. Proof masses inside the spacecraft (two in each spacecraft, one of which is the drag-free reference) will form the end points of three separate but not independent heterodyne interferometers illuminated by Nd:YAG lasers of a few watts power. Telescopes of 0.3 m size will be mounted on each of the three spacecraft for transmission and reception of the laser light. Necessary Precision Engineering developments to support operation of LISA are discussed in Section 5. 1.4. Scope The scope of this paper covers some of the major precision engineering developments made to advance astronomy and gravity research since the formation of CIRP. Clearly the choice of the ‘‘major’’ developments is a judgement made by the authors, consequently it is open to debate. In order to reduce the likelihood of major errors, the technical scope has been limited to that with which the authors have confidence. In respect of astronomy the technical scope includes: Mirror (lens) manufacturing. Optical surface measurement. Active optics support. Adaptive optics techniques. In respect of gravity the technical scope includes: Test mass artefact manufacture and calibration. Distance and direction measurement. Surface and strain measurement. 2. Achievements driven through astronomy 2.1. Mirror (lens) manufacturing Ultra precision manufacturing technology for producing advanced optical surfaces has been, and continues to be, a fundamental ‘‘technology enabler’’ for telescope development. In this section we present developments considered to have made significant contribution.

Fig. 12. Initial LIGO displacement sensitivities (2007) [112].

2.1.1. Spun cast moulding A continuing challenge for astronomers has been the need for larger and more precise mirrors and lens. Prior to the grinding and polishing stages it is necessary to produce highly stable optical blank substrates. In this regard the spun cast honeycomb technology developed in Arizona by the Steward Mirror Laboratory

[(Fig._14)TD$IG]

[(Fig._16)TD$IG]

P. Shore et al. / CIRP Annals - Manufacturing Technology 59 (2010) 694–716

699

Fig. 14. Steward Mirror Laboratory’s Spun cast moulding furnace of 8.4 m capacity [173].

should be considered a major breakthrough in large scale precision mirror blank fabrication [121]. The Arizona spun cast technique produces 8 m scale honeycomb structure lightweight mirror blanks. They are produced to a near net shape having fast (steep) focal ratios allowing for compact overall telescope designs. Spun cast mirror blanks are created by melting borosilicate glass blocks into regions between an array of hexagonal ceramic columns held in a mould. This mould is placed in an enormous rotating oven, see Fig. 14, the oven is rotated as the glass melts. The rotational motion of the mould results in a concave parabolic top surface of the molten glass. The hexagonal columns are made from alumina silica such that they retain their strength at 1180 8C. The molten glass, which typically weighs 20 tonnes for an 8 m scale mirror, takes up the void space between the ceramic inserts. Subsequently there is a critically controlled cooling period (typically 12 weeks) during which the rotational speed of the furnace is reduced. After removal from the furnace, the alumina silica hexagonal segments are ‘‘washed out’’ from the blank using a high pressure water jet technique. The resulting mirror blanks are thus light-weighted and possess high stiffness and so make excellent mirror substrates. When located in the telescope, cool air can be pumped into the hollow regions of the light-weighted mirror rapidly gaining thermal stability. As will be shown later this substrate technology is being employed within the first of the extremely large telescopes: the Giant Magellan Telescope, GMT [94] (Fig. 15). 2.1.2. Large mirror grinding/polishing Grinding and polishing of 8 m scale mirrors has also been [(Fig._15)TD$IG]developed at the Steward laboratory through the development of

Fig. 15. Large Optical Generator 8.4 m capacity [164].

Fig. 16. Steward developed active stressed lap for large aspheric mirror polishing [160].

the so-called Large Optical Generator or LOG [146]. LOG carries out both grinding and polishing. In its grinding mode, material removal rates are up to 20 mm3/s [160]. In its polishing mode, parabolic and off-axis aspheric shape surfaces are produced using a ‘‘stressedlap’’ polishing technique. The stressed-lap technique changes the lap’s shape dynamically as it moves across the mirror surface [163] (Fig. 16). The proven form accuracy achieved using LOG is 15–25 nm RMS on 8 m scale mirrors [6,120,194]. This accuracy achievement represents two parts in 1010 expressed as an error to size ratio. To attain this extreme form accuracy level, post-process measurement and subsequent error correcting/converging re-polishing is employed. A similar capacity mirror processing facility has been established in France by REOSC [67]. However, unlike the Steward approach, REOSC processes much thinner 175 mm thick so-called ‘‘meniscus’’ mirror blanks. The meniscus blanks are held on active supports during polishing as employed within the telescope itself (Fig. 17). The ESO VLT and Gemini telescopes employed Zerodur based meniscus substrates made by Schott. They were subsequently ground, polished and measured in situ. At REOSC a large 20 m high optical test tower is provided above the grinding/polishing system itself. The form accuracy achieved by Sagem-REOSC in producing the four VLT primary mirrors was 17.6–33 nm RMS [34,156]. The Japanese Subaru telescope 8.2 m primary mirror is made in Corning ultra-low expansion glass. It was polished by the US company Brashear having a figure accuracy of 14 nm RMS [96]. The form accuracy achieved on the 8 m scale mirrors by Steward Laboratory, Sagem-ROSC and by Brashear corresponds to an optical resolution (relating to image sharpness) of 0.03 arcsecond in the visible spectrum. It provides a theoretical capability of distinguishing two objects separated by only 150 mm at a distance of 1000 km [169]. Brinksmeier et al, has presented a detailed review of the development of Ultra Precision Grinding [25] which includes details relevant to astronomy surfaces.

[(Fig._17)TD$IG]

700

[(Fig._19)TD$IG]

P. Shore et al. / CIRP Annals - Manufacturing Technology 59 (2010) 694–716

Fig. 19. Stressed mirror polishing (courtesy of Goodrich).

Fig. 17. VLT Mirror (8 m) undergoing optical measurement using in situ measuring tower [187].

2.1.3. Segmented mirror manufacturing Nelson pioneered an alternative to large 8 m scale monolithic mirrors: so-called segmented mirror technology [122]. The idea was originally conceived by Horn d’Arturo in the 1950s [84]. Nelson applied it successfully to the Keck telescopes. The two Keck telescopes have 10.2 m primary mirrors formed from 36 hexagonal off-axis aspheric shape mirror segments. Each segment is 1.8 m wide and laid out as shown in Fig. 18. This segment mirror technology is pivotal, Nelson realised that astronomers would not be content with 10 m telescopes. He also observed that beyond 10 m scale monolithic mirrors were becoming unrealistic; even considering their transportation. Importantly; he also understood that more complex shape primary mirrors (i.e. deep aspheric) would enable more sophisticated and elegant telescope designs. By making the primary telescope mirror from a number of smaller interlocking mirror segments, a scalable means for producing any size of telescope would be enabled [199]. Clearly, producing the off-axis aspheric shaped segments would be critical manufacturing issue. Significantly these complex shape mirror segments would need to demonstrate very limited so-called edge ‘‘roll-off’’. Any shape error at the edges of the segments would reduce the optical performance of the telescope and generate highly problematic stray light features.

[(Fig._18)TD$IG]

Fig. 18. Keck 10.2 m segmented primary mirror showing six different 1.8 m segment types [136].

It was therefore important to develop a manufacturing process chain able to realise the 36 mirror segments of 2 m scale. The developed fabrication route included a novel full-aperture polishing technique called ‘‘stressed’’ mirror polishing [115,135]. In summary, Nelson took best fit spherical substrates, mounted them into an instrumented ‘‘stressing harnesses’’ see Fig. 19 prior to polishing. He subsequently polished the ‘‘stressed’’ (deformed) mirror segments using a traditional full-aperture planetary polishing technique. After polishing, and release from their stressing harnesses the mirror segments would relax into the demanded off-axis aspheric shape. This stressed part polishing approach was applied to circular shape mirrors so that edge roll-off could be avoided. Polished blanks were subsequently cut out into hexagonal segment shapes. 2.1.4. Ion beam figuring of mirror segments Although Nelson’s stressed substrate polishing process was effective it appears not to have achieved the final figure accuracy demanded for the Keck telescopes. The overall shape distorted, by an amount thought to be 0.5–1 mm [46], during the hexagonal shape cutting operation. Nelson consequently employed a postpolishing figuring process in order to improve the shape of the stressed polished segments. This final figuring process, an ion beam figuring (IBF) technology causes ionised argon atoms to bombard the mirror surface causing atom to be ‘‘knocked out’’. A large scale broad ion beam figuring facility was developed and operated by Eastman Kodak [4]. This facility followed the research of Wilson and McNeill [202]. Kodak’s IBF facility, see Fig. 20, employed 2.5 m capacity vacuum chambers and was shown to be capable of improving the [(Fig._20)TD$IG]

Fig. 20. View inside the Kodak (ITT) 2.5 m capacity ion beam figuring facility [90].

[(Fig._21)TD$IG]

[(Fig._2)TD$IG]

P. Shore et al. / CIRP Annals - Manufacturing Technology 59 (2010) 694–716

Fig. 21. Grantecan mirror segments being readied for optical testing using the REOSC test tower [156].

form accuracy of the stressed polished Keck segments from 500 nm RMS to 10–30 nm RMS through a number of converging figuring runs [4]. The average figuring improvement cycle time for the Keck segments was reported to be approximately 80 h with a material removal rate in the region of 0.1 mm3/min. Since little material was removed using the IBF process, only minor surface roughness degradation was incurred; segments were produced having 1–2 nm RMS surface texture (Fig. 21). A similar ion beam processing facility was subsequently developed at Sagem-REOSC [69] who produced the mirror segments for the Grantecan telescopes [66]. A notable difference in the production of the Grantecan segments was that they were polished as hexagonal segments prior to IBF. Sub-aperture polishing was achieved using industrial robots. The achieved form accuracy of the Grantecan segments was 12–30 nm RMS. Importantly, Sagem-REOSC developed the ability to polish/ion beam figure the edges of the segments without inducing edge rolloff, ensuring low levels of consequent stray light in the telescope. 2.1.5. Space mirror fabrication The Hubble Space Telescope is perhaps the most iconic and famous telescope. It has undoubtedly been a major scientific success, operating in the UV to the near infrared, 200–2400 nm. Unfortunately a measurement error made during the fabrication of its primary mirror led to significant amounts of poor publicity. Hubble’s initial optical quality was relatively poor; the telescope was however designed from its outset for space based service missions [86]. It was a measurement error made during the final polishing of the primary mirror that led to a highly complex servicing mission. In 1993 NASA astronauts fitted a correcting optical instrument called COSTAR [42]. This service mission established Hubble’s ultimate performance, unrivalled by contemporary instruments. The Hubble telescope primary mirror is a 2.4 m ULE1 based light weight (820 kg) substrate, see Fig. 22. The flight mirror was polished by Perkin Elmer in 1979 using pre-CNC lapping and polishing technologies. Interestingly, a second (spare) Hubble mirror was produced by Kodak. Whilst it took Kodak 16 months to produce it is claimed the actual polishing time itself was limited to 74 h [89]. It was also made using traditional pre-CNC controlled grinding and polishing machines. In regard to developing advanced manufacturing technology it is perhaps the X-ray space mirrors have led to the widest range of advancements. X-ray telescopes are themselves a relatively new development, the first being launched in 1977 [126]. A highly detailed account of their development has recently been provided by Aschebach [11]. X-ray space telescopes are most significantly based on the application of Wolter type mirror designs employing grazing incidence optics [206]. The primary mirror optics are tubelike in shape typically having internal mirror surfaces of parabolic and hyperbolic shape, see Fig. 23.

701

Fig. 22. Hubble space telescope primary mirror [149].

A number of the Wolter type mirrors are nested. The greater the number of nested mirrors, the greater the X-ray light collecting capacity [12]. Consequently the manufacture of thinner section mirrors has been a significant X-ray telescope technology driver. This has led to a notable ultra precision manufacturing demand. As might be expected, these shorter wavelength mirrors demand ultra smooth surfaces typically 0.3–0.8 nm RMS, having form accuracy demands below the 30 nm RMS region. Pioneering X-ray telescopes, such as Exosat launched in 1983, used mirrors produced by a replication process. This replication process used solid glass mandrels that were ground and polished to the inverse shape and to the form accuracy of the demanded mirror. Mandrels were subsequently gold coated. Carrier mirror shells machined to an accuracy of 0.5 mm were produced. The gold coated glass mandrels were placed inside the carriers and epoxy injected between them. After curing, mandrels were removed, leaving behind the gold mirror surface on the carrier [178]. The EXOSAT mirror carriers were made from beryllium. The ultra precision mandrels were said to be reusable up to six times. Subsequent to EXOSAT, the Rosat X-ray telescope launched in 1998, employed an eight mirror X-ray system. Four nested parabolic mirrors and four nested hyperbolic mirrors, see Fig. 24. The largest of the X-ray mirrors was 0.84 m. Co-aligned with the X-ray system was an extreme ultra violent optical system called the Wide Field Camera (not shown in Fig. 24) having a three nested mirror arrangement, the largest being 0.6 m. Rosat’s X-ray wavelength mirrors were made of Zerodur and manufactured in Germany by Zeiss [150]. These Zerodur mirrors were ground on a vertical grinding machine employing air bearings. During grinding, the Rosat mirrors were supported in aluminium fixtures having many supporting pins. The numerous support pins ensured grinding forces did not induce significant distortion. Subsequent to grinding, polishing was carried out using [(Fig._23)TD$IG]

Fig. 23. Cross section of optical layout of the XMM-Newton X-ray space telescope [209].

[(Fig._24)TD$IG]

702

[(Fig._26)TD$IG]

P. Shore et al. / CIRP Annals - Manufacturing Technology 59 (2010) 694–716

Fig. 24. Cross section of Rosat X-ray space telescope showing four nested Wolter type mirrors [154].

a sub-aperture free abrasive polishing technique. Final polishing runs were undertaken with the mirror in a constrained free state, as suggested by Young [210]. The Rosat X-ray mirrors had a 3 A˚ RMS roughness. Critical geometrical errors of X-ray telescope mirrors have been discussed by Zombeck in 1981 [214]. The two most critical errors were stated as; circularity error about axial direction (akin to roundness/ cylindricity) and profile error again in an axial direction. Rosat Xray mirrors typically had circularity quality and axial direction profile accuracy at 0.5 mm and 0.2 mm respectively. Data to confirm this is presented by Aschenbach [11] and measured using in-house constructed profilometers at Zeiss [16] (Fig. 25). Rosat’s EUV mirrors were electro-less nickel coated aluminium and manufactured by CUPE and Ferranti in the UK. The aluminium substrates were diamond turned by CUPE in 1981 using an inhouse developed large diamond turning machine (Fig. 26) [125]. The so-called LDTM employed a post-process profilometer which allowed form errors to be measured in situ. The profilometer was based on an air bearing supported contacting probe, referenced using a HeNe laser interferometer to optical straight-edges mounted within the machine envelope. Circularity and axial profile accuracies were better than 1 mm. Post-diamond turning, the EUV mirrors were processed by Ferranti Astron. Ferranti carried out a pre-polish prior to coating with electro-less nickel (60 mm thickness). This was followed by a final super polishing employing alumina oxide abrasive and a conventional pitch polishing sub-aperture technique. Achieved roughness was claimed at 0.3–0.5 nm RMS. However, form accuracy was degraded by the polishing process. Each of the aluminium substrates contained both parabolic and hyperbolic surfaces. These could easily be processed using the CNC controlled diamond turning process. However the sub-aperture polishing process induced edge roll-up at the intersection regions of the two optical forms [195]. This polishing edge effect issue, as previously mentioned, is a recurring one in the fabrication of complex shape optical elements. The NASA Chandra X-ray space telescope was launched in 1999 and is the most sensitive X-ray space telescope ever deployed. It provides a 800 cm2 collecting area (equivalent to a 0.3 m diameter

[(Fig._25)TD$IG]

Fig. 25. Rosat mirror blank being loaded into grinding/polishing fixture showing numerous support pins [11].

Fig. 26. Rosat EUV mirror being measured on CUPE LDTM prior to NC error correct diamond turning.

optical telescope) and provides a resolution of 0.5 arcsecond (compared with Rosat at 3 arcsecond) [192]. The largest Chandra (AXAF) mirror element is 1.2 m in diameter and 0.83 m long. Fig. 27 shows a man inside a Chandra mirror cleaning its surface. It appears the Chandra telescope mirror fabrication technology was rather similar to that employed by Zeiss for the Rosat X-ray mirrors. The Chandra mirrors were also made in Zerodur. However they were produced by Hughes Danbury in the US. As with the Rosat mirrors, Chandra’s were processed using small tool (subaperture) grinding and polishing techniques. Fig. 28 shows the polishing of the largest of the Chandra mirrors. Note that the mirror is supported by aluminium rings fitted at each end of the mirror. These were removed during post-polishing measurements so that their distortion influence did not affect the measurements. However, Hughes Danbury achieved significantly greater accuracy and at a larger scale on the Chandra mirror as compared with those made for Rosat. This higher accuracy capability was principally achieved through an improved measuring capability [74]. This improved measuring capability required the development of three special purpose measuring machines. Data from these measuring machines ensured subsequent polishing runs were effective in converging towards required final shape. [(Fig._27)TD$IG]

Fig. 27. Chandra X-ray mirrors (Zerodur) being cleaned by man standing inside the mirror element [13].

[(Fig._28)TD$IG]

[(Fig._30)TD$IG]

P. Shore et al. / CIRP Annals - Manufacturing Technology 59 (2010) 694–716

703

Fig. 30. Size and wavelength comparison of Space Telescopes [80].

Fig. 28. Largest of the Chandra X-ray mirrors (Zerodur) being polished by Hughes Danbury [192].

In summary, the three special purpose measuring machines included: Precision Metrology Station (PMS), Circularity and Inner Diameter Station (CIDS) and Micro-Phase Measuring Interferometer (MPMI). The PMS provided axial figure errors, at specific azimuthal positions, by interferometrically measuring the separation between the mirror surface and a calibrated reference surface. CIDS was built by CUPE in the UK, see Fig. 29. It determined the inner diameters and roundness at the near end regions of the mirrors [75]. CIDS like the PMS used interferometers to collect data from the mirror surface with respect to known reference artefacts. In operation, the Chandra mirrors were loaded over the top of CIDS. CIDS rotated inside the Chandra mirror carrying two sets of opposed measuring probes. The MPMI, was a modified Wyko microscope-based phase shift interferometer and was used to provide the necessary roughness data. [(Fig._29)TD$IG]

By fusing the data from the PMS and CIDS instruments Hughes Danbury could create low frequency error maps to drive their time-dwell sub-aperture polishing process. Typically, it took four polishing cycles per element to achieve an axial form error of 5 nm RMS. It was necessary to use polishing tools of differing size to reduce profile error features of differing wavelength. Roughness of the Chandra mirrors was 1.8–3.4 A˚ RMS as measured over distances of 0.01–1 mm [213]. In 2009 the Herschel Space Telescope was launched. It is a far infrared region telescope operating in the 60–670 mm wavelength range. As seen in Fig. 30 the Herschel space telescope complements the Hubble and the soon to be launched James Webb Space Telescope. As might be expected, the form accuracy requirement for its primary mirror is less demanding than for shorter wavelength telescope mirrors. Its 3.5 m mirror has form accuracy around 6 mm RMS with a 30 nm RMS roughness. Nonetheless the Herschel telescope represents a significant precision engineering achievement. It is structurally and optically fabricated from silicon carbide. Its primary mirror has been constructed by brazing together 12 large mirror segments produced by the French company Boostec [15,24]. The light-weighted mirror segments are moulded using an isostatic pressed sintered silicon carbide. The segments were accurately machined and subsequently joined using a silicon braze technology establishing the 3.5 m mirror substrate. Fig. 31 shows silicon carbide segments being assembled prior to brazing. Since the Herschel space telescope operates at cryogenic temperatures it is clearly critical the silicon braze technology is employed so that a ‘‘stress-free’’ bond across the substrate in achieved.

[(Fig._31)TD$IG]

Fig. 29. The Hughes Danbury Circularity and inner Diameter Station (CIDS) made by CUPE UK.

Fig. 31. Herschel space telescope SiC mirror segments being assembled prior to brazing [80].

[(Fig._32)TD$IG]

704

P. Shore et al. / CIRP Annals - Manufacturing Technology 59 (2010) 694–716

Fig. 32. Herschel Space telescope 3.5 m silicon carbide based primary mirror [80].

The application of silicon carbide for the primary and secondary mirrors and the strut supports of the Herschel telescope ensure its total mass is less than 300 kg, see Fig. 31. The use of hard silicon carbide requires adaptation to conventional polishing. The polishing of the Hershcel primary mirror was carried out by Opteon in Finland who developed and employed a novel diamond based polishing technology [101]. The polished mirror was subsequently coated with an adhesive nickel chrome layer, a reflective aluminium layer and finally a protective silicon based polymer (Fig. 32). The James Webb Space Telescope, scheduled to launch in 2014, will be the first operationally segmented mirror telescope to be deployed in space [171]. It employs ultra-lightweight Beryllium optics, folded into position and then adjusted to shape whilst in orbit, using a wavefront sensing and control subsystem. There are 18 mirror segments, each of 1.3 m size weighing approximately 20 kg, making total mass comparable to Herschel’s, whereas its 6.5 m diameter gives it around four times the area. The mirror segments are currently in production (see Fig. 33) – final figure accuracy is expected to be below 30 nm RMS [170]. 2.2. Mirror surface metrology Testing of optics is a wide ranging subject having been adequately covered elsewhere including [82]. Here we present [(Fig._3)TD$IG]

Fig. 33. JWST mirror segments under test at Ball Aerospace prior to final polishing (Courtesy of Ball Aerospace).

some specific metrology systems that are considered as being critical in delivering recent astronomy programmes. The challenge of testing telescope mirrors is presented by their size and high levels of required accuracy. Large telescope mirrors demand large scale optical test structures. With large structures come significant challenges such as: vibration isolation, thermal management and systematic error identification [56,147,190]. It is not surprising that the testing of telescope mirror surfaces has been dominated by laser-based optical interferometry. It is worth noting though, that these optical tests are generally measures of departure of a surface with respect to a generated wavefront. In large scale optical test facilities it is critical to ensure qualification of tests deal with issues such as: potential irregularities of the generated wavefront, distance and orientation of the test surface and calibration. The sizeable separation distances found in large optical test towers make system identification of the test towers themselves a significant demand [27]. Furthermore testing such precise surfaces becomes an integral part of the manufacturing process, demanding either highly deterministic handling or in situ measuring capability. Bely reports that ‘‘Optical testing is delicate and prone to set up or interpretation errors’’ [17]. He goes on to suggest that opticians have a golden rule: ‘‘Always use two methods based on different physical principles’’. He highlights the Hubble mirror ‘‘fiasco’’ as a painful reminder of this rule. The inclusion of supporting optical tests which rely on contacting measuring principles is consequently a recurring theme. However, optical testing is crucial to making a successful telescope. Large scale telescopes demand large scale optical test facilities. The epitome of this point is in Sagem-REOSC’s 35 m optical test tower at their large optics factory, see Fig. 34. REOSC took particular care in designing their testing tower. It features an external sun shield, a main wall made of boards with a thick lagging, a totally independent internal tower made from a welded metal structure which itself bears a double walled plastic ‘‘tent’’. Thermally controlled air is introduced between the two walls of this ‘‘tent’’. The 8 m capacity final polishing machine was positioned directly under the test tower itself [68]. Even with significant care in devising the REOSC test tower, the long measurement paths meant vibration effects prevented the use of conventional phase shift interferometry as it took too long to collect measurement data. REOSC’s approach in the late 1980s was to develop an algorithm which assessed numerous single ‘‘frame’’ interferograms (batches of 50). These interferograms were averaged in order to minimise vibrational and thermal effects [132]. It is perhaps worthy of note that development of instantaneous phase shift interferometers has reduced some of the vibration difficulties encountered when optical testing large mirrors [103,127]. The overall measurement approach taken by REOSC in the production of the ESO VLT mirrors has been presented in detail by Dierickx [46]. In summary, REOSC applied a combination of metrology techniques ensuring independent verification of

[(Fig._34)TD$IG]

Fig. 34. Sagem-REOSC’s large optics factory with enclosure for 35 m optical test tower [155].

[(Fig._35)TD$IG]

[(Fig._36)TD$IG]

P. Shore et al. / CIRP Annals - Manufacturing Technology 59 (2010) 694–716

Fig. 36. Schematic showing influence of lenslet array employed in Shack–Hartmann sensors [159].

Fig. 35. REOSC’s approach to testing the ESO VLT primary mirrors [46].

measurement techniques was provided. Fig. 35 gives an overview of the VLT measurement methods. An IR based interferometer was used to achieve greater dynamic range during the convergence process from grinding, rough polishing towards fine polishing. The use of a contacting type spherometer [48] provided cross-check data for radius of curvature. A measurement certainty offered by the 1 m long spherometer was claimed to be 10 mm measuring the 30 m radius of curvature surfaces. The spherometer used for the VLT fabrication was moved under CNC control using REOSC’s grinding machine’s motions. Null-lenses were used to modify the wavefront from the interferometer in order to match it to that of the VLT primary mirror. This is a very well documented technique and is therefore not detailed in this paper [48]. Along with the interferometric and contact probe methods, REOSC employed Hartmann and Shack–Hartmann tests. The Hartmann test [78], named after the German inventor Johannes Hartmann, is a classical optical test that provides point details of surface slope. It is robust technique and is insensitive to vibrations and air turbulence. Classically the Hartmann test is based on a mask with small holes being placed over the surface to be measured. It is then illuminated by a parallel light beam. Photographs are taken at two positions: ahead and behind the point of focus. By measuring the positional movements of the dots across the two photographs it is possible to construct the ray paths near the point of focus and hence determine the slope of the mirrors at the location of the mask holes. A more modern development of the Hartmann test is the Shack–Hartmann test (sensor) devised in 1971 by the US physicist Roland Shack [148]. The Shack–Hartmann sensor was developed for astronomy programmes in Europe by Wilson [200]. It was applied by REOSC in the fabrication of the ESO VLT mirrors. Shack– Hartmann sensors employ lenslet arrays in front of their detector (CCD) as opposed to the mask holes of the Hartmann technique. The lenslet arrays positioned in front of the camera/sensor (see Fig. 36) provide the slope information necessary to precisely construct wavefronts [159]. In the fabrication of the VLT mirrors the Shack–Hartman sensors were used as control sensors for the 150 point active support system employed during the final polishing processes. This is discussed in Section 2.3. The use of laser trackers to precisely determine the positioning and measurement of the key elements of optical test towers is widely employed [29]. The metrology techniques applied at the Steward Mirror laboratory in fabricating the Gemini telescope primary mirrors, and by Brashear in fabrication the Subaru mirrors, have a numbers of parallels to those employed by REOSC for the VLT mirrors. Detail differences between the three manufacturers are beyond the scope of this paper.

705

Measurement of off-axis aspheric mirror segments as employed within Keck and Grantecan telescopes presented specific challenges. Both Keck and Grantecan telescopes were formed using six different mirror segment types as indicated in Fig. 18. Each mirror segment was measured using an optical test tower. Specific wavefronts for each segment type were created through the use of a computer generated hologram (CGH) diffraction plate [28]. The CGH is a diffractive element, typically made by electron beam lithography (see Fig. 37). The level of departure from spherical shape varied across the six Grantecan segments from 30.6–215 mm. The measured form accuracy of a Gantecan mirror segment is claimed at 14 nm RMS with a lateral measuring resolution of approximately 4 mm. An interferogram is shown in Fig. 38. Ensuring all mirror segments had the same base radius of curvature was achieved by employing a ‘‘master’’ or ‘‘reference’’ segment approach. Six segments (1 of each type) where measured with respect to a known ‘‘master’’ segment, see Fig. 39. The Giant Magellan Telescope (GMT) is perhaps the furthest along in the ‘‘new’’ generation of telescopes that move well beyond the 8–10 m scale. Importantly its design uses a more conventional optical test tower measurement approach. GMT has a primary mirror built as a close packed array of seven 8.4 m scale segments, see Fig. 40. The six identical outer segments have an aspheric departure of approximately 14.5 mm, see Fig. 41, with a large astigmatic component.

[(Fig._37)TD$IG]

Fig. 37. Photograph diffractive CGH null lens (courtesy of Diffraction International).

[(Fig._38)TD$IG]

706

[(Fig._40)TD$IG]

P. Shore et al. / CIRP Annals - Manufacturing Technology 59 (2010) 694–716

Fig. 38. HeNe interferogram of a REOSC produced Grantecan telescope 1.8 mirror segment [156].

The approach for metrology of the GMT primary [70] takes advantage of the existing infrastructure at the Steward Laboratory. The existing (25 m) test tower under the football stadium is being adapted. A 3.75 m aperture light-weighted spherical mirror at the top of the tower, illuminated off axis, folds the beam, changes the f/ #, and adds much of the astigmatism required for a null test. The remaining aspheric departure is corrected using a second folding spherical mirror and a CGH (Fig. 42 omits the CGH). The proposed TMT and E-ELT segmented mirror-based telescopes will employ many hundreds of mirror segments. The number of segment prescriptions will be 82 and 164 respectively for the primary mirrors. The difficulty level of their effective measurement, whereby the base radius of curvature is fully ensured, presents a significant metrology task. For example the TMT primary mirror design has a paraxial radius of curvature of 60 m – suggesting a very large test tower if a conventional approach was adopted. The TMT design, with around a hundred segment prescriptions, and target production rates >2 segments per week strongly suggest that easily reconfigurable tests are required. So far little has been published describing an effective test tower design. In principle a conventional tower could be made adaptable, however the large optical path suggests vibration and turbulence problems may be significant. An approach which uses relatively large ‘‘test plates’’ to shorten the non-common path length had been advocated by Pan and Burge [144]. The technique uses an illumination leg hologram (see Fig. 43). The zero order (spherical) wavefront reflects from the

Fig. 40. Schematic of the Giant Magellan telescope based on seven 8.4 m mirror ‘‘segments’’ [70].

spherical reference surface while the (aspheric) first order reflects from the segment. This test uses a relatively small gap, minimising turbulent noise, and has a common mode imaging leg. A separate CGH would be needed for each segment type. Cross calibration of the many CGHs becomes an important measurement demand. In regard to the contacting measurement techniques, Burge has presented numerous metrology developments supporting the fabrication of large scale convex mirrors, typically used as secondary mirrors in large telescopes [30]. One such instrument is a swing arm profilometer [7] as in Fig. 44. It provides measurement capability that is especially useful between the finish grinding and polishing operations. This contacting profilometer has a claimed scan accuracy of 50 nm RMS over 1 m scale

[(Fig._41)TD$IG] [(Fig._39)TD$IG]

Fig. 39. Photograph showing six Grantecan primary segments under the REOSC optical test tower [156].

Fig. 41. Aspheric departure of an outer element of GMT [70].

[(Fig._42)TD$IG]

[(Fig._45)TD$IG]

P. Shore et al. / CIRP Annals - Manufacturing Technology 59 (2010) 694–716

707

[(Fig._43)TD$IG]

Fig. 45. Photograph of an image slicer made by Cranfield University Precision Engineering for the JWST MIRI instrument [162].

Fig. 42. GMT test tower schematic [70].

relative to the workpiece table such that the motion of the probe is a spherical path through the origin of the mirror under test. An overall surface measurement accuracy of better than 1 mm is claimed. Other swing arm profilometers have been developed for future telescope programmes [100]. Astronomy mirror demands are not limited to large mirrors. Complex shape multi-mirror arrays are also demanded, especially within telescope instrument systems [108]. A significant example is image ‘‘slicers’’ and ‘‘re-imagers’’ as used within spectrometers see Fig. 45. Development of multi axes diamond machining techniques for producing these mirror arrays has been undertaken [22,49]. Shore reported the diamond machining technique employed to produce such optics for the James Webb Space Telescopes MIRI instrument [162]. Morantz reported their measurement using a miniature Twyman–Green interferometer mounted onto a stability enhanced ultra precision CMM [130]. 2.3. Active support techniques

Fig. 43. Illumination Leg Hologram approach to large mirror testing [144].

surfaces [8]. The configuration of the Steward swing arm is innovative and simple in its mechanical construction. An LVDT contacting probe is mounted to an arm attached to an air bearing table. The air bearing table is inclined and positioned

[(Fig._4)TD$IG]

Fig. 44. Schematic showing a swing arm profilometer mounted to a polishing machine [27].

The primary mirrors in the Keck, Grantecan, VLT and Subaru telescopes can be described as active optics. In the case of the Keck and Grantecan telescopes each of their primary mirror segments are actively positioned in regard to their piston, tip and tilt relative to their neighbouring segments [10,93]. The Keck active optics (AO) system controls the 36 segments forming its primary mirror [204]. This AO system consists of 168 position sensors, 108 position actuators (3 per segment). The resolution of the capacitance type sensors is 4 nm and a relative segment positioning accuracy of 25 nm is claimed at 2 Hz frequency. The ‘‘passive’’ support for each segment, employed to restrain the three remaining degrees of freedom (i.e. the two lateral motions and rotation of the hexagon element), is actuated by a motor/rotary encoder system. Restraint is achieved by a thin, stainless-steel ‘‘flex-disk’’ attached to a central ‘‘pocket back’’ in the segment and three whiffletrees with articulating arms that float the segment at 36 evenly distributed points, Fig. 46. In operation a Shack–Hartmann camera is used to provide slope information to align the 36 mirror segments [37]. Although the primary mirror technology is quite different on the VLT and Subaru telescopes, they do employ a similar active optics control philosophy [177,198]. Their meniscus type monolithic mirrors are also brought into the required shape by active supports. In the case of the Subaru telescope the ULE-based primary mirror is supported by 261 forcecontrolled actuators having a load capacity of 1500 N and a 0.15 N control resolution, Fig. 47. The mirror is brought into a floating or zero-gravity condition prior to being corrected into its optimum form [175]. The active control approach of the ESO VLTs is shown in Fig. 48. The wavefront sensor is a Shack–Hartmann derived device [186] whose data collecting is referenced directly to the locations of the

[(Fig._46)TD$IG]

708

[(Fig._48)TD$IG]

P. Shore et al. / CIRP Annals - Manufacturing Technology 59 (2010) 694–716

Fig. 46. photograph of a Keck primary mirror segment fitted with its supporting mechanism [203].

150 mirror supports. The VLT active optics system constantly monitors the optical quality of the image, using an offset reference star as it is picked up in the field by the wavefront sensor CCD in the adapter sensor arm. The same offset star is also used by the acquisition and autoguiding CCD. The system controls the relative position and the shape of the optical elements. The primary mirror shape can be actively controlled by varying the force pattern applied by means of its support system. The latter consists of 150 computer controlled axial actuators, applying a distribution of forces at the back of the mirror. 2.4. Adaptive optics techniques Adaptive optics are optics which can change their surface shape in response to varying control signals. Typically the term is taken to describe optics which are capable of minor dynamic changes to compensate for real-time wavefront errors; the dynamic changes at a small scale may in some cases be made at frequencies measured in kHz [139]. The normal application is to correct atmospheric distortion of light. In order to improve observation quality of distant or faint objects through the atmosphere, guide stars are used to adjust the adaptive optics by monitoring the guide star’s image in real time and compensating accordingly. Natural guide star (NGS) and Laser guide star (LGS) systems are employed [142]. Laser Guide Stars are a form of artificial star created to use as a reference for astronomical adaptive optics. The use of a bright NGS limits the observation areas of the sky to within 30 arcsecond of the NGS [99] whereas a LGS increases the observable sky area. A laser guide star can be located in the lower atmosphere (10 km) using visible or ultraviolet light to reflect off air molecules

Fig. 48. Control system diagram of the ESO VLT [186].

or in the upper part of the Earth’s mesosphere (90  10 km altitude) rich in Na atoms, using a yellow laser light (589 nm) to excite those sodium atoms [205]. The target is to obtain a LGS and for light from the observed object to pass through the same part of the atmosphere. The Keck II telescope laser guide star (LGS) adaptive optics (AO) system [183] uses a sodium layer laser guide star in combination with a tip-tilt sensor stage (TSS), an M3 laser pointing mirror and field steering mirrors (FSMs) to position the LGS on the Wavefront sensor (WFS). For E-ELT projects, as the telescope diameter increases, a multiple LGS is considered [38] to mitigate the wavefront error effect (cone effect) observed in single LGS used on current adaptive optics [(Fig._49)TD$IG]on 8–10 m class telescopes (Fig. 49).

[(Fig._47)TD$IG]

Fig. 47. Photograph of the Subaru primary mirror force-controlled support system [175].

Fig. 49. Laser Guide Star Facility at Paranal Observatory in Chile [104].

[(Fig._51)TD$IG]

P. Shore et al. / CIRP Annals - Manufacturing Technology 59 (2010) 694–716

709

3. Achievements of gravitational research In this section we detail some precision engineering achievements that have advanced gravitational research. In doing so we cover programmes that have more precisely tested the ‘‘geodetic effect’’ and ‘‘frame dragging’’ as depicted in Fig. 7. These advances will be made through reference to the NASA Gravity Probe B programme. In addition we cover advancements made to enable the unequivocal detection of gravity waves, those which will enable a new astronomical form of observation. Vibration isolation development in support of LIGO is also referenced. Future capability offered by the LISA programme is provided later within Section 5. 3.1. Gravity Probe B Gravity Probe B (GP-B) is a NASA funded physics mission designed to test through a direct, controlled experiment, two predictions of Einstein’s general theory of relativity [143]. Firstly ‘‘the geodetic effect’’ – the amount by which the Earth warps the local space-time, and secondly the ‘‘frame-dragging’’ effect – the amount by which the rotating Earth drags its local space-time around with it [208]. In order to measure the minuscule angles predicted by Einstein’s theory, it was necessary to build unprecedented gyroscopes and incorporate them into a spacecraft having a reference telescope. The gyroscopes needed a so-called spin axis drift of less than one hundred-billionth of a degree per hour during rotation. In order to achieve this demand a number of key specifications were defined. These included: imbalance of mass (density distribution), uneven gyroscope surfaces and near zero friction between the bearings and the gyroscope [118] (Fig. 50). The rotors developed for the GP-B were fashioned from 1.5 in. spherical fused quartz, see Fig. 51. A novel four pad polishing technology with ‘‘paired pad’’ and reversal every 10–15 s was developed. It produced mechanical spherical quality of <10 nm RMS. The data for spherical compliance were taken on a Talynova73 with 16 polar great circles tied together with one equatorial trace and fitted with spherical harmonics [114]. A Guinness world record was awarded for this achievement. Each of the four spherical rotors was made to be homogenous. Density homogeneity of three parts 107 assured the mass centre coincides with the geometry centre and the moments of inertia are equal to avoid torques due to the gravity gradient. The measurement technique to assure the homogeneity in fused silica was developed at Aberdeen [44] using the refractive index of quartz as the indication of density, with an immersion technique in a liquid

[(Fig._50)TD$IG]

Fig. 50. Niobium coated spherical rotor and gyroscope housing each having three levitating bearings [58].

Fig. 51. Schematic layout of the four Gravity Probe B gyroscopes [58].

of matched index of refraction to suppress the small variations in physical path length in a cube of the candidate material. The rotors are levitated using a gyro suspension system (GSS) having an operation gap of 32 mm. The GSS had a remarkable operational demand in so much as it needed to both levitate on Earth and whilst performing at 10 8 g with minimal disturbance. Levitation on Earth demanded four orders of magnitude more voltage than that needed in space [26]. The rotor pole axis motion was measured using Superconducting QUantum Interference Devices (SQUID) which detect the movement of the self-generated magnetic field of the superconducting niobium coated rotors. The GP-B SQUID devices provide a 0.1 milliarcsecond sensitivity level [58]. By employing four gyroscopes as shown in Fig. 51, where two of the gyros rotate in one direction and two the other, a highly sensitive system providing robust data and cross checking capability was realised. Reference point setting of the GP-B space craft was provided by a Cassegrain telescope viewing a ‘‘guide’’ star. To satisfy the precision pointing requirements of the GP-B experiment, it was necessary to locate the optical centre of the guide star image in the telescope to an accuracy of 0.1 milliarcsecond (3  10 88). However, the diffraction limit size of the GP-B telescope was limited to 1.4 arcsecond. Consequently a sensitive directional point detector was developed [145]. This directional sensor ensured that once the telescope was locked onto the guide star, the telescope pointing signals were used to compensate for the small pointing deviations to an accuracy of better than 0.1 milliarcsecond [18]. Helium thrusters to steer the GP-B spacecraft were developed by DeBra [45]. In September 2009 the GP-B programme reported frame dragging of 5 milliarcsecond/year with a statistical uncertainty level of 14%. 3.2. LIGO The Laser Interferometer Gravitational-Wave Observatory (LIGO) consists of two widely separated installations within the United States [112]. Each is an L-shaped interferometer with evacuated interferometer arms 4 km long. The interferometers, sensing gravity wave disturbances in suspended test masses, are widely separated in the US – operated in unison as a single observatory. LIGO requires extraordinary levels of optical performance in high vacuum and vibration isolation. The natural seismic noise of the selected ‘quiet sites’ must be reduced by ten orders of magnitude at frequencies near 10 Hz. This is done in a series of stages, the three outer ones are active in all six degrees of freedom (dof). The outermost stage is hydraulic [77]. It provides intrinsic natural damping for some of the elastic modes of the inner stages simplifying the stabilisation of the active loops while providing mm range compensation of ground motion. This hydraulic laminar flow system evolved from precision engineering developments for machines tools [207] (Fig. 52). Mounted on the inner stage is a four stage pendulum system. The passive pendulums operate at the fundamental limit of thermally induced noise and use quartz fibres for their suspension. On the 2nd and 3rd stages one of the limitations is measuring tilt which couples a component of gravity into the horizontal

[(Fig._52)TD$IG]

[(Fig._54)TD$IG]

P. Shore et al. / CIRP Annals - Manufacturing Technology 59 (2010) 694–716

710

Fig. 52. LIGO schematic.

seismometers – the sensors for horizontal isolation. A measurement of the ground rotation at very low rates (<3e 10 radians at 30 mHz) is required. Very significant technical challenges were surmounted in achieving successful operation. A number of these have commercial spin offs, covered in Section 4.

Fig. 54. Large scale ultra precision diamond turning machine at UPS2 IKC.

[(Fig._5)TD$IG]

4. Commercial spin offs This section identifies a number of example developments that were enabled, or at the very least progressed, through astronomy and gravity wave programmes. 4.1. Ultra precision optical fabrication The constant demand of astronomers to see further and with greater sensitivity has progressed capability of optical fabrication and measurement. The ultra precision machine technology developed at CUPE was significantly progressed through astronomy missions. The CUPE LDTM (seen in Fig. 26) was funded by the SERC funded Rosat EUV mirror programme [196]. The CUPE developed CIDS instrument (Fig. 29) was funded by Hughes Danbury for the Chandra telescope. A 2.5 m off-axis grinding machine was built at CUPE for Eastman Kodak’s segment mirror programme [107,197], see Fig. 53. These astronomy programmes established unique large scale ultra precision machine tool technologies at CUPE. CPE Ltd/ Cranfield Precision, and subsequently others, applied this knowhow in producing large drum roll diamond turning machines in series for major blue chip commercial companies, see Fig. 54. Manufacture of large scale micro-textured drum moulds is a major enabling technology for next generation displays. Fabrication of such drums remains a near market research – commercial [(Fig._53)TD$IG]activity operated by Cranfield University at UPS2 IKC [124]. An

Fig. 53. Schematic layout of the laser interferometer controlled CUPE/CPE Ltd OAGM made for Kodak.

Fig. 55. Diamond turned drum mould for reel-to-reel film fabrication.

example diamond turned micro-structured drum is shown in Fig. 55. These diamond turned drum moulds are also enabling cost effective manufacture of performance enhancing solar concentrators for PV systems [188]. The fabrication of the lenses used in current lithography systems, is achieved using ion beam figuring [191]. Lithography systems are the machines which predominantly enable adherence of Moore’s law [129] to be maintained as semiconductor device feature dimensions are reduced (Fig. 56). This ion beam figuring process technology was certainly advanced through the development of the Kodak large scale ion beam figuring system. Kodak results on the Keck mirror segments [(Fig._56)TD$IG]were compelling.

Fig. 56. Commercial ion beam figuring facility for the fabrication of lithographic optics NTGL GmbH [140].

[(Fig._58)TD$IG]

P. Shore et al. / CIRP Annals - Manufacturing Technology 59 (2010) 694–716

711

Carl Zeiss GmbH, having been significantly involved in X-ray and EUV optics fabrication for telescope programmes, are a leading supplier of optical towers for advanced lithographic wafer steppers [98]. Other leading suppliers of EUV optics, such as Nikon employ similar energy beam technologies as finalising process for high end optics. 4.2. Wavefront guided LASIK The Shack–Hartmann wavefront sensor has seen a significant increase of applications. It has found powerful uses not only in astronomy and adaptive optics, but also in ophthalmology, optical testing, laser beam analysis, and semiconductor manufacturing. Part of the reason for this growth in application is the advancement of the various technological components, such as the lenslet array, the CCD and the computer. The Shack–Hartmann sensor has clearly enabled a number of new techniques and significantly wavefront guided laser–assisted in situ keratomileusis (LASIK) [179,78]. Wavefront guided LASIK is a laser-based treatment to correct eyesight. It uses data from a Shack–Hartmann derived wavefront sensor, typically having >1500 data points (micro-lenslets) [31]. The measurement data is then used to control a laser in a timedwell error correcting process. The Hartmann measurement technique, devised to improve the quality of the ‘‘The Great Refractor’’ of Potsdam, has progressed into a significant widely applied sensor that enables modern eye surgery, see Fig. 57 [110].

Fig. 58. Principle of EUV all-reflective normal incidence lithography projection optics [32].

4.3. EUV coating technologies In the early 1970s, Spiller proposed that the problem of achieving normal incidence imaging in the Extreme Ultraviolet (EUV) could be addressed with high reflection multilayer coatings. The technique was to use stacks of quarter-wave thick layers of scattering or absorbing materials, of differing refractive index. These were created originally by thermal source vapour deposition [165,166]. This pioneering work was driven by applications in Space Optics and others [14]. Later developments introduced sputter and other deposition refinements reducing layer thickness and thereby extending applicability to shorter (ultimately X-ray) wavelengths, finding much wider imaging application, including soft x-ray astronomy, microscopy and significantly microlithography [35,79,128]. EUV imaging presents unique technical difficulties. As semiconductor feature dimensions push below 30 nm, consequently required lithography wavelengths push well below 20 nm, at which wavelengths all materials (including gases) are heavy absorbers. Imaging system designs for lithography are driven by the requirement not just for imaging quality but also for optical efficiency. The systems, operating in high vacuum, therefore must use only reflective optics, and even the coated mirror surfaces each absorb a significant fraction (around 30%) of the incident

illumination, so whereas conventional deep UV lithography illuminators, based on refractive principles, have >100 optical surfaces, EUV all-reflective designs have typically six (see Fig. 58). The coatings moreover, made of multiple alternating layers of molybdenum and silicon, must retain coating thickness uniformity of much better than 0.1 nm on steeply curved optical surfaces in order to ensure adequate performance.

[(Fig._59)TD$IG]

4.4. LIGO inspired developments In the development of LIGO a number of commercial spin off can be highlighted. Ultra smooth silica substrates having sub-angstrom RMS roughness and figure accuracy <1 nm RMS have been demanded. It led to the development and first application of the N-positional areal 3-flat test [57]. Variations of that test now provide the basis for NIST calibration of optical flats [76]. It has also been used commercially for flat calibration systems and in subtracting mount induced deformations from measurements of EUVL mask blanks (Fig. 59).

[(Fig._57)TD$IG]

Fig. 57. Modern eye measuring system derived from the Shack–Hartmann wavefront sensor (courtesy of Carl Zeiss).

Fig. 59. end Test Mass measurement at NIST [76].

712

[(Fig._61)TD$IG]

P. Shore et al. / CIRP Annals - Manufacturing Technology 59 (2010) 694–716

The development of ultra strong silica suspension fibres was a specific demand of Gravitational Wave detectors that has been widely applied in the oil industry for gravity gradiometry for bore hole applications [133]. A quasi monolithic bonding technology for oxide materials such as silica, originally developed at Stanford for the GP-B experiment, and further developed in the UK for gravity wave applications, span off into precision optical components and laser application. 5. Future demands 5.1. Extremely large ground based telescopes (ELTs) Extremely large telescopes are classed as telescopes possessing an aperture diameter greater than 20 m for observation in the UV, visible and near IR wavelengths. Monolithic mirrors are limited to around 10 m diameter for reasons of manufacturing, measurement and transportation difficulty. Larger effective apertures are possible through the aperture synthesis achieved by interferometry of combined smaller aperture telescopes, but there are significant advantages in an optically filled aperture, not least light-gathering power. Consequently, segmented mirror technology, as applied to the Keck telescopes, has become the main basis for telescope scale up. Funded ELT programmes scheduled for first light around or before 2020 include GMT at 24.5 m, TMT at 30 m and E-ELT at 42 m apertures, requiring 7, 492 and 984 segments respectively for their primary mirrors. TMT and E-ELT designs have (deliberately) adopted the same segment size (at 1.45 m) for reasons including economy of supply. E-ELT having twice the segment count has twice the light-gathering area. The supply issue is non-trivial: all 20 of the world’s operational very large (>5 m diameter) telescope primary mirrors, manufactured over a period of more than 50 years, have a combined area of around 1000 m2; just TMT and E-ELT between them will require in excess of twice that for their primaries, manufactured within a period of only 10 years. This represents a requirement for a rapid increase in worldwide supply capacity, currently underway, perhaps approaching an order of magnitude. Other science (and potentially industrial) programmes outside the astronomical field pose similarly increasing demands for large optics fabrication (Fig. 60). The stress-lap and stressed mirror polish techniques employed for some existing large mirrors and segments, and described in Section 2.1, utilise modified spherical surface generation techniques with subsequent sub-aperture figure correction. An alternative approach has emerged, which is geared towards higher volumes. This utilises CNC sub-aperture techniques at every stage of surface generation, with an essentially unlimited freeform capability [161]. Targeting a 20 h cycle time for each stage of surface generation for a 1.5 m class mirror, it specifically addresses the requirement for worldwide supply ramp-up, where [(Fig._60)TD$IG]

Fig. 60. Scale comparison of Hale (5.08 m) and E-ELT (42 m) primary mirrors.

Fig. 61. Rapid production line for large scale mirror segment fabrication.

existing techniques take at least an order of magnitude longer (Fig. 61). The approach is based on a high speed, high accuracy ultra-low damage freeform CNC grinding stage, followed by a relatively short CNC figure corrective polishing stage. These technologies have been demonstrated; grinding achieving cycle time and form accuracy targets below 1 mm RMS. An additional CNC sub-aperture Reactive Atom Plasma figure correction technology is at an advanced stage of development [59,176]. Mass production of 1 m scale mirrors has become a commercial demand defined by future astronomy programmes providing a clear potential benefit for optic fabrication. 5.2. Laser Interferometer Space Antenna (LISA) Detection of Gravity Waves is the considered goal of LISA. As introduced in Section 1.3 LISA is a space mission designed to measure gravitational waves over a frequency range of 0.00003– 0.1 Hz with a maximum sensitivity of 1e 23 strain/HHz at 0.004 Hz. Its specified sensitivity will enable the measurement from several gravitational wave sources such as massive black holes merging at the centre of galaxies, binaries of compact stars, and relic radiation from the early phase of the Big Bang. LISA consists of three identical spacecraft flying in a triangular constellation, with equal arms of 5 million km each. As gravitational waves from distant sources reach LISA, they warp spacetime, stretching and compressing the triangle. By accurately measuring the separation between the spacecraft, the effect of gravitational radiation is enabled. LISA employs interferometer measurements between the space craft. The measurement technique itself is derived from radio wave technology, Doppler tracking [181]. In LISA it is based on infrared laser beams. The laser light going out from one spacecraft to the other is not reflected back directly, because diffraction losses over such long distances would be too great. Instead, the phase of the incoming laser is measured, and used to set the phase of the outgoing laser, which is transmitted back at full intensity: this process is known as ‘‘transponding’’. When the transponded laser light arrives back at the original spacecraft, it is superposed with a portion of the original laser beam, and their phases compared. This relative phase measurement, which is referenced to the position of the two test masses, gives information about the separation between the spacecraft. The difference between the phase measurements for the two arms gives information about the relative changes in the two arms, which are induced by gravitational waves (Fig. 62). At the heart of each assembly is a vacuum enclosure containing a free-falling polished platinum-gold 40 mm cube – the proof mass – that serves as the optical reference (mirror) for the light beams. A passing gravitational wave will change the length of the optical path between the proof masses of one arm of the interferometer relative to the other arm. The distance fluctuations are measured to a precision of 4  10 11 m (averaged over 1 s) which, when combined with the large separation between the spacecraft, will allow LISA to detect relative displacements caused by gravitational

[(Fig._62)TD$IG]

[(Fig._63)TD$IG]

P. Shore et al. / CIRP Annals - Manufacturing Technology 59 (2010) 694–716

713

Fig. 62. LISA 40 mm cube test mass whose position is to be capacitance sensed to 4  10 11 m [40].

waves down to an expected level in the order of Dl/l = 10 year of observation with a signal-to-noise ratio of 5 [64].

23

in 1

5.3. International X-ray Observatory The International X-ray Observatory (IXO), a large deployable X-ray mirror structure with 20 m focal length, is a joint effort (NASA/ESA/JAXA) currently planned to be launched for 2021 [88,54]. IXO will be a Wolter type X-ray telescope as described in Chapter 2.1. However, IXO employs segmented mirror technology similar to that employed in JSWT and the proposed ELTs. IXO requires approximately 14000 mirror segments of 200–400 mm size and 0.4 mm thick [36]. A highly efficient manufacturing capability for producing these mirror segments is required. Prior to IXO, a smaller X-ray space telescope will be launched in 2011, called NuSTAR. NuSTAR also employs thin shell X-ray mirror. Segment mass production in support of NuSTAR (which requires approximately 8000 segments), is reported at a rate of 450 per week [23]. Mirror manufacturing technologies for the IXO mirror shell segments include ‘‘slumped glass’’ and ‘‘silicon pore optic’’. These processes are being used commercially: glass slumping is used in liquid crystal flat panel display industry, ‘‘silicon pore’’ optics are standard semiconductor wafer technology processes [2,3]. The IXO demand is creating a higher accuracy capability from these manufacturing technologies [211]. The slump technique is being employed by Goddard Space Flight Centre to fulfil the NuSTAR mission. In this method glass sheets are slumped onto precisely polished form mandrels (see Fig. 63). The substrates are positioned over the precise mandrels in ovens. As the ovens reach 600 8C the substrates deform under their own weight forming to the mould mandrels. The formed glass substrates are subsequently cut and coated. Approaches to reduce the harmful aspects of possible debris have been proposed, including textured and porous forming mandrels. The coated slump glass segments are mounted on transfer modules. These modules are assembled circumferentially in order to obtain the final Wolter mirror assembly. In contrast; silicon pore optics are rectangular silicon plates produced having thin ribs on one side and thin membranes between the ribs. The sandwich construction is bent into the required shape and stacked to obtain a stiff pore structure. The stacks are then mounted into ‘‘tandem’’ structures which are combined to form the X-ray mirror [39]. IXO requires 6948 parabolic and 6948 hyperbolic mirror segments. The surface area demanded for the polished forming mandrels is 60 m2 (721 mandrels). Total mirror segment area over

Fig. 63. Slumped glass segment [141]. Credit: NASA.

800 m2. The form quality requirements for individual mirror segments equates to 100 nm RMS with tighter short wavelength demands [212]. The IXO demand represents a new ultra precision production engineering challenge and a major commercial contract. 6. Conclusions This paper has highlighted unprecedented precision engineering achievements borne of the demands of large scale astronomy and gravity science. These have included:  Ultra Stable and Lightweight Materials.  Ultra Precision Surface Fabrication & Measurement (especially at large scale).  Ultra Precision Machine Tool Techniques.  Ultra Sensitive Vibration Isolation.  Ultra Precision Sensor Techniques. Notable examples of how these techniques have translated into critical manufacturing technologies have been given and important contributions by CIRP members acknowledged. New products that have evolved from science programme solutions have been highlighted. The paper has also identified key precision engineering demands of future astronomy and gravity programmes and the even greater demands that these future programmes place on ultra precision technologies. In addition, the scale of future programmes, such as the Extra large telescopes, demand greater attention to production engineering capabilities in order for them to be realised in acceptable timescale and cost. In closing, this paper has shown important engineering developments emergent from Astronomy and Gravity science. In order for mankind to fully understand its place within the universe continued advancement and innovation of precision engineering technologies will be required. Acknowledgements The authors would like to thank the following for their contributions: Klaus-Friedrich Beckstette, Pat McKeown, David Allen, Bill Wills-Moren and Robert Hocken. Acknowledgement is given to the EPSRC funded UPS2 IKC, ESA, ESO, NASA, STFC and the McKeown Foundation.

714

P. Shore et al. / CIRP Annals - Manufacturing Technology 59 (2010) 694–716

References [1] Adams WS (1938) George Ellery Hale, 1868–1938. Astrophysical Journal 87:369–388. [2] Akilian M (2008) Methods of Improving the Surface Flatness of Thin Glass Sheets and Silicon Wafers. PhD Thesis, Massachusetts Institute of Technology, Dept. of Mechanical Engineering. [3] Akilian M, Foresta CR, Slocum AH, Trumper DL, Schattenburg ML (2007) Thin Optic Constraint. Precision Engineering 31(2):130–138. [4] Allen LN, Keim RE, Lewis TS, Ullom JR (1992) Surface Error Correction of a Keck 10-m Telescope Primary Mirror Segment by Ion Figuring. Proceedings of the SPIE 1531:195–204. [5] Andersen T, Ardeberg A, Owner-Petersen M (2003) Euro50: Design Study of a 50 m Adaptive Optics Telescope. Technical Report. Lund University, Sweden. [6] Anderson DS, Angel JRP, Burge JH, Davison WB, DeRigne ST, Hille BB, Ketelsen DA, Kittrell WC, Martin HM, Nagel RH, Trebisky TJ, West SC, Young RS (1991) Stressed-Lap Polishing of 3.5-m f/1.5 and 1.8-m f/1.0 Mirrors. Proceedings of the SPIE 1531:260–269. [7] Anderson DS, Burge JH (1995) Swing-Arm Profilometry of Aspherics. Proceedings of the SPIE 2536:169–179. [8] Anderson DS, Parks RE, Shao T (1990) A Versatile Profilometer for the Measurement of Aspherics. Optical Fabrication and Testing OSA Technical Digest 11(1):119–122. [9] Angel JRP, Davison WB, Hill JM, Mannery EJ, Martin HM (1990) Progress Toward Making Lightweight 8-m Minors of Short Focal Length. Proceedings of the SPIE 1236:636–640. [10] AO Keck Telescope (2010) http://www2.keck.hawaii.edu/optics/ao/index. html. [11] Aschenbach B (1985) X-ray Telescopes. Progress in Physics 48:579–629. [12] Aschenbach B (2009) Realization of X-ray Telescopes-from Design to Performance. Experimental Astronomy 26:95–109. [13] AXAF High Resolution Mirror (2010) http://hea-www.harvard.edu/HRC/ AXAF/hrma.html. [14] Bajt S, Chapman HN, Spiller E, Hau-Riege S, Alameda J, Nelson J, Walton CC, Kjornrattanawanich B, Aquila A, Dollar F, Gullikson E, Tarrio C, Grantham S (2007) Multilayers for Next Generation X-ray Sources. Proceedings of the SPIE 6586:18–24. [15] Bath DA, Spain D, Ness E, Williams S, Bougoin M (2005) Evaluation of Segmented and Brazed Mirror Assemblies. Proceedings of the SPIE 5868:34–41. [16] Beckstette K, Heynacher E (1983) Contour Line Measurements on ROSAT Xray Mirrors. Proceedings of the SPIE 429:126–129. [17] Bely P (2003) The Design and Construction of Large Optical Telescopes Series Astronomy and Astrophysics Library, vol. XXIV. 978-0-387-95512-4. [18] Bencze W, DeBra D, Herman L, Holmes T, Adams M, Keiser G, Everitt F (2006) On-Orbit Performance of the Gravity Probe B Drag-Free Translation Control System. Advances in the Astronautical Sciences 125:425–440. [19] Bennett JA (1976) On the Power of Penetrating into Space: The Telescopes of William Herschel. Journal for the History of Astronomy 7:75–108. [20] Billing H, Maischberger K, Ruediger A, Schilling R, Schnupp L, Winkler W (1979) An Argon Laser Interferometer for the Detection of Gravitational Radiation. Journal of Physics E Scientific Instruments 12:1043–1050. [21] Bleeker J, Brinkman AC, Culhane J, Koch-Miramond L, Pounds K, Schnopper HW, Silver EH, Spada G, Taylor B, Tru¨mper J, Westergaard NJ (1984) XMM: A Long-Lived Orbiting X-ray Multi-Mirror Observatory. Physica Scripta T7:224– 234. [22] Bonneville C, Cagnat JF, Laurent F, Prieto E, Ancourt G (2004) Image Slicer Manufacturing: from Space Application to Mass Production. Proceedings of the SPIE 5494:188–195. [23] Bookbinder J (2009) The International X-ray Observatory. IXO Astro2010 Decadal 1–19. [24] Breysse J, Castel D, Laviron B, Logut D, Bougoin M (2004) All-SiC Telescope Technology: Recent Progress and Achievements. Proceedings of the ICSO, 659– 671. [25] Brinksmeier E, Klocke F, Ohmori H, Aurich C, Shore P, Mutlugu´es Y (2010) Ultra Precision Grinding. Annals of CIRP 2. [26] Buchman S, Bencze W, Brumley R, Clarke B, Keiser GM, Hansen WW (1998) The Design and Testing of the Gravity Probe B Suspension and Charge Control Systems. The American Institute of Physics 178–187. [27] Burge JH (1996) Measurement of Large Convex Aspheres. Proceedings of the SPIE 2871:362–373. [28] Burge JH (1999) Efficient Testing of Off-Axis Aspheres with Test Plates and Computer-Generated Holograms. Proceedings of the SPIE 3782:348–357. [29] Burge JH, Kott LB, Martin HM, Zhender R, Zhao C (2006) Design and Analysis for Interferometric Testing of the GMT Primary Mirror Segments. Proceedings of the SPIE 6273:6273GM. [30] Burge JH, Su P, Zhao C (2008) Optical Metrology for Very Large Convex Aspheres. Proceedings of the SPIE 7018. 701818–701818-12. [31] Carl Zeiss Meditec (2010) http://www.meditec.zeiss.com. [32] Carl Zeiss SMT (2010) http://www.smt.zeiss.com. [33] Cavendish H (1798) Experiments to Determine the Density of the Earth. Philosophical Transactions of the Royal Society of London (Part II) 88: 469–526. [34] Cayrel M (2000) Completion of VLT and Gemini Primary Mirrors at REOSC. Proceedings of the SPIE 4003:14–23. [35] Ceglio NM, Hawryluk AM (1991) Soft-X-ray Projection Lithography System Design. OSA Proceedings 12:5–10. [36] Chan KW, Zhang W, Saha T, Robinson D, Olsen L, McClelland R, Mazzarella J, Lozipone L, Lehan J, Hong M, Fleetwood C, Evans T, Byron G, Larimore J (2009) An Approach for Alignment, Mounting, and Integration of IXO Mirror Segments. Proceedings of the SPIE 7437. 74371D–74371D-11.

[37] Chanan G, Troy M, Dekens F, Michaels S, Nelson J, Mast T, Kirkman D (1998) Phasing the Mirror Segments of the Keck Telescopes: The Broadband Phasing Algorithm. Applied Optics 37(1):140–155. [38] Clare R, le Louarn M (2009) Simulations of Laser Guide Star Adaptive Optics Systems for the E-ELT. Presentation at Adaptive Optics for Extremely Large Telescopes, Paris. [39] Collon MJ, Gu¨nther R, Ackermann M, Partapsing R, Kelly C, Beijersbergen MW, Bavdaz M, Wallace K, Riekerink MO, Mu¨ller P, Krumrey M (2009) Stacking of Silicon Pore Optics for IXO. Proceedings of the SPIE 7437. 74371A–74371A-7. [40] Core Technologies for LISA (2010) http://lisa.nasa.gov/technology. [41] Court T, Von Rohr M (1929) A History of the Development of the Telescope from about 1675 to 1830 Based on Documents in the Court Collection. Transactions of the Optical Society 30(5):207–260. [42] Crocker JH, Ford HC, Hartig GF, Jedrzejewski RI (1994) Optical Performance of the Corrective Optics Space Telescope Axial Replacement (COSTAR). Proceedings of the SPIE 2198:1170–1180. [43] Cunningham C (2009) Future Technologies for Optical and Infrared Telescopes and Instruments. Experimental Astronomy 26:179–199. [44] De Freitas JM, Player MA (1995) Ultrahigh Precision Measurements of Optical Heterogeneity of High Quality Fused Silica. Applied Physics Letters 66(26):3552–3554. [45] DeBra D, Bencze W, Everitt F, VandenBeukel J, Kirschenbaum J (2008) Proportional Helium Thrusters for Gravity Probe B. Advances in the Astronautical Sciences 131:209–216. [46] Dierickx P (2000) Optical Fabrication in the Large. Proceedings of the SPIE 5382: 224–236. [47] Dierickx P, Brunetto ET, Comeron F, Gilmozzi R, Gonte´ FYJ, Koch F, le Louarn M, Monnet GJ, Spyromilio J, Surdej I, Verinaud C, Yaitskova N (2004) OWL Phase – A Status Report. Proceedings of the SPIE 5489:391–406. [48] Dierickx P, Enard D, Geyl R, Paseri J, Cayrel M, Beraud P (1997) VLT Primary Mirrors: Mirror Production and Measured Performance. Proceedings of the SPIE 2871:385–392. [49] Dubbeldam M, Robertson D, Preuss W (2004) Free Form Diamond Machining of Complex Monolithic Optics for Integral Field Systems. Proceedings of the SPIE 5494:164–175. [50] Einstein A (1905) Zur Elektrodynamik bewegter Ko¨rper. Annalen der Physik 17:891–921. [51] Einstein A (1915) Die Feldgleichungen der Gravitation. Ko¨niglich Preussische Akademie der Wissenschaften 844–847. [52] Enard D, Noethe L, Alvarez P, Atad E, Andersen T, Castro J, Goncharov A, Iye M, Korhonen T, Lemaitre G, Salinari P, Brunetto E, Delabre B, Dierickx P, Dimmler M, Koch F, Schneerman M, Spyromilio J, Swat A, Tamai R (2006) ELT Telescope Design Working Group Report – Issue 3. Technical Report, Working Group on Telescope Design (WG4). . [53] Enard D, Swings JP (1983) The Very Large Telescope Project. Messenger 34: 1–3. [54] ESA Science and Technology (2010) http://sci.esa.int. [55] European Southern Observatory. (1998) The VLT White Book. European Southern Observatory (ESO), Garching near Munich . Electronic access: http:// www.eso.org/public/astronomy/teles-instr/whitebook. [56] Evans CJ, Hocken RJ, Estler TW (1996) Self-Calibration: Reversal, Redundancy, Error Separation, and ‘Absolute Testing’. Annals of CIRP 45(2):617–634. [57] Evans CJ, Kestner RN (1996) Test Optics Error Removal. Applied Optics 35:1015–1021. [58] Extraordinary Technologies of GP-B (2010) http://einstein.stanford.edu/ TECH/technology1.htmlhttp://einstein.stanford.edu/TECH/technology1.html. [59] Fanara C, Shore P, Nicholls JR, Lyford N, Kelley J, Carr J, Sommer P (2006) A New Reactive Atoma Plasma Technology for Precision Machining. Advanced Engineering Materials 8(10):933–939. [60] Forward RL (1978) Wideband Laser-Interferometer Gravitational-Radiation Experiment. Physical Review D 17(2):379–390. [61] Foucault LM (1859) Memoire sur la construction des telescopes en verre argente. Annales de l’Observatoire imperial de Paris 5:197–237. (in French). [62] Fritschel P (2009) LIGO-T010075-v2: Advanced LIGO Systems Design Technical notes: System Engineering. http://dcc.ligo.org, sign in as ‘public user’, ‘number search’ LIGO#‘T010075’. [63] GEO600 Project (2010) http://www.geo600.uni-hannover.de. [64] Gerardi D, Allen G, Conklin JW, Sun KX, DeBra D, Buchman S, Gath P, Fichter W, Byer RL, Johann U (2009) Advanced Drag-Free Concepts for Future SpaceBased Interferometers: Acceleration Noise Performance. PACS:95.55.Ym95.85.Sz. [65] Gertsenshtein ME, Pustovoit VI (1963) On the Detection of Low Frequency Gravitational Waves. Soviet Physics JETP 16:433–435. [66] Geyl R (2005) From VLT to GTC and the ELTs. Proceedings of the SPIE 5965:235–240. [67] Geyl R, Cayrel M (1999) REOSC Contribution to VLT and Gemini. Proceedings of the SPIE 3739:40–46. [68] Geyl R, Cayrel M (2000) REOSC Approach to ELT’s and Segmented Optics. Proceedings of the SPIE 4003:59–64. [69] Geyl R, Rinchet A, Rolland E (1999) Large Optics ion Figuring. Proceedings of the SPIE 3739:161–166. [70] Giant Magellan Telescope (2010) http://www.gmto.org. [71] Gilmozzi R (2004) Science and Technology Drivers for Future Giant Telescopes. Proceedings of the SPIE 5489:1–10. [72] Gilmozzi R, Dierickx P (2000) OWL Concept Study. ESO Messenger 100: 1–10. [73] Gilmozzi R, Spyromilio J (2007) The European Extremely Large Telescope (EELT). ESO Messenger 127:11–19. [74] Glenn P, Slomba A, Babish R (1986) Technology Mirror Assembly Mirror Quality Requirements & Achievements. Proceedings of the SPIE 640:45–58.

P. Shore et al. / CIRP Annals - Manufacturing Technology 59 (2010) 694–716 [75] Gordon T (1990) Circumferential and Inner-Diameter Metrology for the Advanced X-ray Astrophysical Facility (AXAF) Optics. Proceedings of the SPIE 1333:239–247. [76] Griesmann U (2006) Three-Flat Test Solutions Based on Simple Mirror Symmetry. Applied Optics 45(23):5856–5865. [77] Hardham C, DeBra D, et al, (2006) Quiet Hydraulic Actuators for LIGO. IFAC Mechactronics Conference, Darmstadt, Germany, . [78] Harvey JE, Hooker RB (2005) Robert Shannon and Roland Shack: Legends in Applied Optics. SPIE Press978-0-819-45844-5. [79] Hawryluk AM, Seppala LG (1988) Soft-X-ray Projection Lithography Using an-X-ray Reduction Cameras. Journal of Vacuum Science & Technology 6:2162. [80] Herschel (2010) http://www.esa.int/esaMI/Herschel. [81] Hill JM, Salinari P (1998) The Large Binocular Telescope Project. Proceedings of the SPIE 3352:23–33. [82] Hocken RJ, Chakraborty N, Brown C (2005) Optical Metrology of Surfaces. Annals of CIRP 54(2):705–717. [83] Hogg AR (1959) The Last of the Specula. Astronomical Society of the Pacific Leaflets 8:105–112. [84] Horn d’Aturo G (1950) Altri esperimenti con lo specchio a tasselli, Vol V. . p. 11. [85] Hough J, Rowan S, Sathyaprakash BS (2005) The Search for Gravitational Waves. Journal of Physics B Atomic Molecular and Optical Physics 38:S497– S519. [86] Hubble Space Telescope (2010) http://hubble.nasa.gov. [87] Hulse RA, Taylor JH (1975) Discovery of a Pulsar in a Binary System. Astrophysical Journal 195(2):L51–L53. [88] International X-ray Observatory (2010) http://ixo.gsfc.nasa.gov. [89] ITT Hubble Space Telescope (2010) http://www.ssd.itt.com/heritage/hubble. shtml. [90] ITT Ion Figuring (2010) http://www.ssd.itt.com/precisionoptics. [91] James Webb Space Telescope (2010) http://www.jwst.nasa.gov. [92] Jansen F, Lumb D, Altieri B, Clavel J, Ehle M, Erd C, Gabriel C, Guainazzi M, Gondoin P, Much R, Munoz R, Santos M, Schartel N, Texier D, Vacanti G (2001) XMM-Newton Observatory: The Spacecraft and Operations. Astronomy and Astrophysics 365:L1–L6. [93] Jochum L, Castro J, Devaney N (1998) Gran Telescopio Canarias: Current Status of its Optical Design and Optomechanical Support System. Proceedings of the SPIE 3352:621–631. [94] Johns M (2008) Progress on the GMT. Proceedings of the SPIE 7012. 70121B– 70121B-15. [95] Kahn B (2005) A Year in the Life of Gravity Probe B. Stanford Report. . http:// news.stanford.edu/news/2005/may11/gpb-051105.html. [96] Kaifu N, et al, (2000) The First Light of the Subaru Telescope: A New Infrared Image of the Orion Nebula. Publication of the Astronomical Society of Japan 52:1–8. [97] Kaifu N (1998) Subaru Telescope. Proceedings of the SPIE 3352:14–22. [98] Kaiser W, Kuerz P (2008) EUVL – Extreme Ultraviolet Lithography. Optik & Photonik 2:35–38. [99] KECK LGSAO (2010) http://www2.keck.hawaii.edu/optics/lgsao/lgsbasics.html. [100] King CW, Callender M, Efstathiou A, Gee AE, Walker D, Lewis AJ, Oldfield S, Steel R (2006) A 1 m Swing-arm Profilometer for Large Telescope Mirror Elements. Proceedings of the ASPE 1913:1–4. [101] Korhonen T, Keinanen P, Pasanen M, Sillanpaa A (2008) Polishing and Testing of the 3.5 m SiC M1 Mirror of the Herschel Space Observatory of ESA. 7102: 710218-710218-7. [102] Kristian J, Blouke M (1982) Charge-Coupled Devices in Astronomy. Scientific American 247:66–74. [103] Ku¨chel M (1990) The New Zeiss Interferometer. Proceedings of the SPIE 1332:655–663. [104] Laser Guide Star Facility (2010) http://www.eso.org/public/images/vltlgsf_mar2009-1122. [105] Laser Interferometer Space Antenna Mission (2010) http://sci.esa.int/ lisahttp://sci.esa.int/lisa. [106] Launius RD (2005) Sputnik and the Origins of the Space Age. . http:// www.hq.nasa.gov/office/pao/History/sputnik/sputorig.html. [107] Leadbeater PB, Clarke M, Wills-Moren WJ, Wilson TJ (1989) A Unique Machine for Grinding Large Off-Axis Optical Components: The OAGM 2500. Precision Engineering 11(4):191–196. [108] Lee D, Dickson C, Hastings P, Well M, Leclerc M (2004) Image Slicers – Design for Manufacturability. Proceedings of the SPIE 5494:176–187. [109] Levine JL (2004) Early Gravity-Wave Detection Experiments 1960–1975. Physics in Perspective 6:42–75. [110] Liang J (1991) A New Method to Precisely Measure the Wave Aberrations of the Human Eye with a Hartmann–Shack Wavefront-Sensor. PhD Dissertation, University of Heidelberg. [111] Lick Observatory (2010) http://mthamilton.ucolick.org/public/history/. [112] LIGO (2010) http://www.ligo.caltech.edu. [113] Lindley D (2005) A Fleeting Detection of Gravitational Waves. Physical Review Focus. http://focus.aps.org/story/v16/st19. [114] Lipa J, Siddall GJ (1980) High Precision Measurement of Gyro Rotor Sphericity. Precision Engineering 2(3):123–128. [115] Lubliner J, Nelson JE (1980) Stressed Mirror Polishing. 1: A Technique for Producing Nonaxisymmetric Mirrors. Applied Optics 19(14):2332–2340. [116] Lumb DH, Berthiaume GD, Burrows DN, Garmire GP, Nousek JA (1991) Charge Coupled Devices (CCDs) in X-ray Astronomy. Experimental Astronomy 2(3):179–201. [117] Mackay C (1986) Charge-Coupled Devices in Astronomy. Annual Review of Astronomy and Astrophysics 24:255–283. [118] Marcelja F, Keiser GM (1998) Silicon Spheres for Gravity Probe B Experiment. Proceedings of the ASPE Spring Topical Meeting on Silicon Machining 1998:74– 76.

715

[119] Martin HM, Allen RG, Burge JH, Dettmann LR, Ketelsen DA, Miller III SM, Sasian JM (2003) Fabrication of Mirrors for the Magellan Telescopes and the Large Binocular Telescope. Proceedings of the SPIE 4837:609–618. [120] Martin HM, Anderson DS, Angel JRP, Nagel RH, West SC, Young RS (1990) Progress in the Stressed-Lap Polishing of a 1.8-m f/1 Mirror. Proceedings of the SPIE 1236:682–690. [121] Martin HM, Burge JH, Cuerden B, Miller SM, Smith B, Zhao C (2004) Manufacture of 8.4-m Off-Axis Segments: A 1/5-Scale Demonstration. Proceedings of the SPIE 5494:62–70. [122] Mast TS, Nelson JE (1982) Figure Control for a Fully Segmented Telescope Mirror. Applied Optics 21(14):2631–2641. [123] Mast TS, Nelson JE, Sommargren GE (2000) Primary Mirror Segment Fabrication for CELT. Proceedings of the SPIE 4003:43–58. [124] Mateboer AJ, Morantz P, Shore P, Amos S (2009) Development of a Large Scale Micro Surface Structuring Facility. EOS Conference Presentation, . [125] McKeown PA, Read RFJ, Wills-Moren WJ (1986) Experiences in the Precision Machining of Grazing Incidence X-ray Mirror Substrates. Proceedings of the SPIE 571:42–50. [126] Miller R, Austin G, Koch D, Jagoda N, Kirchner T, Dias R (1978) The High Energy Astronomy Observatory X-ray Telescope. IEEE Transactions on Nuclear Science NS-25:422–429. [127] Millerd JE, Brock NJ, Hayes JB, Wyant JC (2004) Instantaneous Phase-Shift Point-Diffraction Interferometer. Proceedings of the SPIE 5531:264–272. [128] Montcalm C, Grabner F, Hudyma R, Schmidt M, Spiller E, Walton C, Wedowski M, Folta J (2002) Atomic-Precision Multilayer Coating of the First Set of Optics for an Extreme-Ultraviolet Lithography Prototype System. Applied Optics 41:3262–3269. [129] Moore GE (1965) Cramming More Components onto Integrated Circuits. Electronics 38:114–117. [130] Morantz P, Shore P, Lee D (2005) Metrology of Imaging Mirror Arrays for Space Telescope Spectrometer Optics. Proceeings of the Lamdamap, Cranfield, 513–522. [131] Mountain M, Kurz R, Oschmann J (1994) Gemini 8 m Telescopes Project. Proceedings of the SPIE 2199:41–55. [132] Mueller R, Hoeness H, Espiard J, Paseri J, Dierickx P (1993) The 8.2-m Primary Mirrors of the VLT. The Messenger 73:1–8. [133] Nabighian MN, Ander ME, Grauch VJS, Hansen RO, LaFehr TR, Li Y, Pearson WC, Peirce JW, Phillips JD, Ruder ME (2005) Historical Development of the Gravity Method in Exploration. Geophysics 70(6):63ND–89ND. [134] Nelson JE (2000) Design Concepts for the California Extremely Large Telescope (CELT). Proceedings of the SPIE 4004:282–289. [135] Nelson JE, Gabor G, Hunt LK, Lubliner J, Mast TS (1980) Stressed Mirror Polishing. 2: Fabrication of an Off-Axis Section of a Paraboloid. Applied Optics 19(14):2341–2352. [136] Nelson JE, Mast TS, Faber SM (1985) The Design of the Keck Observatory and Telescope. Keck Observatory Report 90:5.1–5.44. [137] Newton I (1687) Philosophiae Naturalis Principia Mathematica London. . [138] Newton I (1704) Optics. Royal Society, London. [139] Noethe L, Franza F, Giordano P, Wilson RN, Citterio O, Conti G, Mattaini E (1988) Active Optics II. Results of an Experiment with a Thin 1 m Test Mirror. Journal of Modern Optics 35(9):1427–1457. [140] NTGL Nanotechnologie Leipzig GmbH (2010) http://www.ntgl.de/daten/ daten/ntgl.pdf. [141] NuSTAR Mission (2010) http://www.nustar.caltech.edu. [142] Olivier S (1996) Crucial Steps Taken in Laser Guide Star System. Science & Technology Review 22–23. [143] Overview of the GP-B Mission (2010) http://einstein.stanford.edu/MISSION/ mission1.html. [144] Pan FY, Burge JH (2004) Efficient Testing of Segmented Aspherical Mirrors by Use of Reference Plate and Computer-Generated Holograms. I. Theory and System Optimization. Applied Optics 43(28):5303–5312. [145] Parkinson BW, Kasdin NJ (1988) Twenty Milliarcsec Pointing System for the Rolling GP-B Spacecraft. Advances in the Astronautical Sciences 66:65–89. [146] Parks RE, Lam P, Kuhn W (1985) The Large Optical Generator – a Progress Report. Proceedings of the SPIE 542:28–31. [147] Peters J, Bryan J, Estler TW, Evans C, Kunzmann H, Lucca DA, Sartori S, Sato H, Thwaite EG, Vanherck P (2001) Contribution of CIRP to the Development of Metrology and Surface Quality Evaluation during the last 50 years. Annals of CIRP 50(2):471–488. [148] Platt BC, Shack R (2001) History and Principles of Shack–Hartmann Wavefront Sensing. Journal of Refractive Surgery 17:573–577. [149] Primary mirror of the Hubble (2010) http://pds.jpl.nasa.gov/planets/ captions/hubble/hst01.htm. [150] Reinhardt D (1987) Features of Manufacturing and Qualification Tests of a High-Resolution Wolter I Mirrors. Proceedings of the SPIE 733:145–148. [151] Ritchey GW (1928) The Modern Photographic Telescope and the New Astronomical Photography – Part II, The Ritchey-Chre´tien Reflector. Journal of the Royal Astronomical Society of Canada 22:207–230. [152] Robertson DI, Morrison E, Hough J, Killbourn S, Meers BJ, Newton P, Robertson NA, Strain KA, Ward H (1995) The Glasgow 10 M Prototype Laser Interferometric Gravitational Wave Detector. Review of Scientific Instruments 66(9):4447–4452. [153] Ronan CA (1991) The Origins of the Reflecting Telescope. Journal of the British Astronomical Association 101(6):335–342. [154] ROSAT Mission (2010) http://www.mpe.mpg.de/xray/wave/rosat. [155] Roux R (2000) Telescope Optics: Polishing Takes VLT Mirrors Deep into Space. Laser Focus World 36(5). (Available Online). [156] Ruch E (2009) Polissage et me´trologie des grands composants optiques dans le domaine de l’astronomie et des lasers. Presentation at Polissage Optique pour les Grands Instruments de la Physique et de l’ Astronomie June 2009, Bordeaux

716

P. Shore et al. / CIRP Annals - Manufacturing Technology 59 (2010) 694–716

[157] Sanchez F (1986) The Canary Islands’ Astrophysical Observatories – A European Community Project. Societa Astronomica Italiana Memorie 57(4):655– 672. [158] Sandage A (1989) Edwin Hubble 1889–1953. Journal of the Royal Astronomical Society of Canada 83(6):351–362. [159] Shack–Hartmann (2010) http://www.optocraft.de/products/SHSLab/index.en.php. [160] Shannon RR, Parks RE (1984) Large Optical Generator. Proceedings of the SPIE 444:271–273. [161] Shore P (2008) Ultra Precision Surfaces. Proceedings of the ASPE 75–78. [162] Shore P, Morantz P, Lee D (2006) Manufacturing and Measurement of the MIRI Spectrometer Optics for the James Webb Space Telescope. Annals of the CIRP 55(1):543–547. [163] SOML Lapping Machine (2010) http://www.mmto.org/pr_images/30cmlap. jpg. [164] SOML Mirror Polishing Techniques (2010) http://mirrorlab.as.arizona.edu/ TECH.php?navi=polihttp://mirrorlab.as.arizona.edu/TECH.php?navi=poli. [165] Spiller E (1972) Low-Loss Reflection Coatings Using Absorbing Materials. Applied Physics Letters 20:365–367. [166] Spiller E (1974) Multilayer Interference Coatings for the Vacuum Ultraviolet in Space Optics. National Academy of Sciences – National Research Council. pp. 581–597. [167] Spitzer LJ (1946) Astronomical Advantages of an Extra-Terrestrial Observatory. Project RAND Report. Douglas Aircraft Co.. September 1. [168] Spitzer LJ (1979) History of the Space Telescope. Quarterly Journal of the Royal Astronomical Society 20:29–36. [169] Spyromilio J, Wallander A, Tarenghi M (1998) Commissioning of the Unit Telescopes of the VLT. The Messenger 98:21–24. [170] Stahl HP (2007) JWST Lightweight Mirror TRL-6 Results. Proceedings of IEEE Aerospace Conference, 1–12. [171] Stahl HP (2007) JWST Mirror Technology Development Results. Proceedings of the SPIE 6671:667102–667104. [172] Stepp LM, Strom SE (2004) The Thirty-Meter Telescope Project Design and Development Phase. Proceedings of the SPIE 5382:67–75. [173] Steward Observatory Mirror Lab Brochure (2010) http://mirrorlab.as.arizona.edu/. [174] Strom SE, Stepp LM, Gregory B (2003) Giant Segmented Mirror Telescope: A Point Design Based on Science Drivers. Proceedings of the SPIE 4840:116–128. [175] Subaru Telescope Advanced Technologies (2010) http://www.naoj.org/Introduction/tech.html. [176] Subrahmanyan P, Gardopee G, Verma Y, Li N, Yu T, Kyler T, Fiske P, Sommer P (2007) Rapid Fabrication of Lightweight SiC Aspheres Using Reactive Atom Plasma (RAP) Processing. Proceedings of the SPIE 6666. 66660A.1–66660A.4. [177] Tabata M, Itoh N, Iye M (1993) Active Mirror Support System for SUBARU Telescope. The 7th International Precision Engineering Seminar, 233–244. [178] Taylor BG, Andresen RD, Peacock A, Zobl R (1981) The EXOSAT Mission. Space Science Reviews 30(1–4):479–494. [179] Thibos LN, Hong X (1999) Clinical Applications of the Shack–. [180] Thorne KS (1980) Gravitational-Wave Research: Current Status and Future Prospects. Reviews of Modern Physics 52(2):285–299. [181] Thornton CL, Border JS (2003) Radiometric Tracking Techniques for Deep-Space Navigation. 978-0-471-44534-0. [182] Tobin W (2003) The Life and Science of Leon Foucault. Cambridge University Press0521808553. [183] van Dam MA, Bouchez AH, Le Mignant D, Johansson EM, Wizinowich PL, Campbell RD, Chin JCY, Hartman SK, Lafon RE, Stomski Jr PJ, Summers DM (2006) The W. M. Keck Observatory Laser Guide Star Adaptive Optics System: Performance Characterization. Publications of the Astronomical Society of the Pacific 118:310–318. [184] Van Helden A (1977) The Invention of the Telescope. Transactions of the American Philosophical Society New Series 67(4):1–67. [185] Virgo (2010) http://www.pg.infn.it/virgo. [186] VLT Active Optics System (2010) http://www.eso.org/public/images/ eso9940b. [187] VLT Main Mirror (2010) http://www.eso.org/public/news/eso9952.

[188] Walker N (2010) Micro-Structured Surfaces for Next Generation Optical Films. Proceedings of the SIG, Euspen, Aachen, . [189] Weber J (1969) Evidence for Discovery of Gravitational Radiation. Physical Review Letters 22(24):1320–1324. [190] Weck M, McKeown P, Bonse R, Herbst U (1995) Reduction and Compensation of Thermal Errors in Machine Tools. Annals of the CIRP 44(2):589–598. [191] Weiser M (2009) Ion Beam Figuring for Lithography Optics. Proceedings of 16th Conference on Ion Beam Modification of Materials 267(8–9):1390–1393. [192] Weisskopf MC, O’Dell SL, Elsner RF, Van Speybroeck LP (1995) Advanced Xray Astrophysics Facility AXAF – An Overview. Proceedings of the SPIE 2515:312–329. [193] Weisskopf MC, O’Dell SL, Van Speybroeck LP (1996) Advanced X-ray Astrophysics Facility (AXAF). Proceedings of the SPIE 2805:2–7. [194] West SC, Martin HM, Nagel RH, Young RS, Davison WB, Trebisky TJ, Derigne ST, Hille BB (1994) Practical Design and Performance of the Stressed-Lap Polishing Tool. Applied Optics 33:8094–8100. [195] Willingale R (1988) ROSAT Wide Field Camera Mirrors. Applied Optics 27:1423–1429. [196] Wills-Moren WJ, Modjarrad H, Read RFJ, McKeown PA (1982) Some Aspects of the Design and Development of a Large High Precision CNC Diamond Turning Machine. CIRP Annals – Manufacturing Technology 31(1):409–414. [197] Wills-Moren WJ, Wilson T (1989) The Design and Manufacture of a Large CNC Grinding Machine for Off-Axis Mirror Segments. CIRP Annals 38(1):529–532. [198] Wilson R, Franza F, Noethe L (1987) Active Optics I. A System for Optimizing the Optical Quality and Reducing the Costs of Large Telescopes. Journal of Modern Optics 34(4):485–509. [199] Wilson RN (1999) Reflecting Telescope Optics II: Manufacture, Testing, Alignment, Modern Techniques. Series Astronomy and Astrophysics Library, vol. XVIII. 978-3-540-60356-6. [200] Wilson RN (2003) The History and Development of the ESO Active Optics System. The Messenger 113:2–9. [201] Wilson RN (2004) Reflecting Telescope Optics I: Basic Design Theory and its Historical Development. Series Astronomy and Astrophysics Library, 2nd ed., vol. XVI. 978-3-540-40106-3. [202] Wilson SR, McNeil JR (1987) Neutral Ion Beam Figuring of Large Optical Surfaces. Proceedings of the SPIE 818:320–324. [203] Wizinowich P (2006) Optical Engineering for High Angular Resolution at Keck Observatory Keck Support. SPIE Newsroom 1–3. spie.org/x8797.xml. [204] Wizinowich PL (2005) Adaptive Optics and the Keck Observatory. IEEE Instrumentation and Measurement Magazine 8:12–19. [205] Wizinowich PL, Le Mignant D, Bouchez AH, Campbell RD, Chin JCY, Contos AR, van Dam MA, Hartman SK, Johansson EM, Lafon RE, Lewis H, Stomski PJ, Summers DM (2006) The W. M. Keck Observatory Laser Guide Star Adaptive Optics System: Overview. Publications of the Astronomical Society of the Pacific 118:297–309. [206] Wolter H (1952) Mirror Systems with Glancing Incidence as Image-Producing Optics for X-rays. Annals of Physics 445:94–114. [207] Wu H, DeBra D (2001) Laminar Flow Actuation on Ultraprecision Stage Positioning. ASPE Topical Meeting, Philadelphia, PA, . [208] Xiao YM, Bardas D, Buchman S, Cohen C, Everitt CWF, Gill D, Keiser GM, van Patten RA, Taber M, Turneaure JP, van Hooydonk T, Walter T, Zhou P (1991) Gravity Probe B: III. The Precision Gyroscope. Proceedings 6th Marcel Grossmann Meeting on General Relativity, 394–398. [209] XMM-Newton Science Operations Centre (2010) http://xmm.esac.esa.int. [210] Young PS (1979) Fabrication of the High-Resolution Mirror Assembly for the HEAO-2 X-ray Telescope. Proceedings of the SPIE 184:131–138. [211] Zhang W (2008) Fabrication of Mirror Segments for the IXO Mission. Presentation in MIT Workshop on Lightweight X-Ray Optics, MIT, USA, . [212] Zhang W (2009) Mirror Segment Fabrication and Metrology for the International X-ray Observatory. Presentation at International Workshop on Astronomical X-Ray Optics, Prague. [213] Zhao P, van Speybroeck LP (1995) AXAF VETA-I mirror X-ray Test Results Cross-Check with the HDOS Metrology Data. Proceedings of the SPIE 2515:391–409. [214] Zombeck MV (1981) Advanced X-ray Astrophysics Facility (AXAF). Optical Engineering 20:297–309.