Cryogenic optical measurements of 12-segment-bonded carbon-fiber-reinforced silicon carbidecomposite mirror with support mechanism

Hidehiro Kaneda,1,* Takao Nakagawa,1 Takashi Onaka,2 Keigo Enya,1

Sin’itirou Makiuti,1 Junji Takaki,3 Masaki Haruna,4 Masami Kume,5 and Tsuyoshi Ozaki5

1Institute of Space and Astronautical Science, Japan Aerospace Exploration Agency, Kanagawa 229-8510, Japan2Department of Astronomy, University of Tokyo, Tokyo 113-0033, Japan

3Mitsubishi Electric Corporation, Hyogo 661-8661, Japan4Advanced Technology R&D Center, Mitsubishi Electric Corporation, Hyogo 661-8661, Japan

5Advanced Technology R&D Center, Mitsubishi Electric Corporation, Kanagawa 229-1195, Japan

*Corresponding author: kaneda@ir.isas.jaxa.jp

Received 9 October 2007; accepted 6 January 2008;posted 17 January 2008 (Doc. ID 88410); published 10 March 2008

A 720mm diameter 12-segment-bonded carbon-fiber-reinforced silicon carbide (C/SiC) composite mirrorhas been fabricated and tested at cryogenic temperatures. Interferometric measurements show signifi-cant cryogenic deformation of the C/SiC composite mirror, which is well reproduced by a model analysiswith measured properties of the bonded segments. It is concluded that the deformation is due mostly tovariation in coefficients of thermal expansion among segments. In parallel, a 4-degree-of-freedom ball-bearing support mechanism has been developed for cryogenic applications. The C/SiC composite mirrorwas mounted on an aluminum base plate with the support mechanism and tested again. Cryogenic de-formation of the mirror attributed to thermal contraction of the aluminum base plate via the supportmechanism is highly reduced by the support, confirming that the newly developed support mechanismis promising for its future application to large-aperture cooled space telescopes. © 2008 Optical Societyof America

OCIS codes: 120.6650, 220.4840, 230.4040, 350.1260.

1. Introduction

High thermal stability and good thermal conductionare major characteristics of silicon carbide (SiC)material, which makes SiC mirrors most suitablefor applications under thermally severe conditions,including cryogenic applications. The high specificstiffness of SiC is advantageous for lightweightmirror applications. Furthermore, SiC is very toughmaterial against high-energy particle collision,which is also suitable for space applications. HenceSiCmirrors are our primary candidates for the Space

Infrared Telescope for Cosmology and Astrophysics(SPICA) telescope [1–3], a Japanese infrared astro-nomical satellite project with a 3:5m telescopeplanned to be launched around 2017. The telescopewill be cooled to 4:5K in space by a combination ofmechanical coolers with an efficient radiative cool-ing system. The SPICA telescope has requirementsfor its total weight to be lighter than 700kg andfor the imaging performance to be diffraction limitedat 5 μmat 4:5K. The latter requirement is equivalentto a total wavefront error less than 350nm rms,which implies that the surface figure error of a pri-mary mirror must be no more than 175nm rms or0:28λ rms (here and hereafter, λ is the He–Ne laserwavelength of 632:8nm).

0003-6935/08/081122-07$15.00/0© 2008 Optical Society of America

1122 APPLIED OPTICS / Vol. 47, No. 8 / 10 March 2008

AKARI [4,5] (formerly known as ASTRO-F), aJapanese infrared astronomical satellite that waslaunched in February of 2006 and is currently inoperation, adopts porous-core chemical vapordeposition-coated SiC mirrors [6] for the telescope.Extension of our heritage from the AKARI SiC mir-rors to the SPICA telescope mirrors encounters dif-ficulties: the large chemical-vapor-deposition furnacerequired for 3:5m diameter mirrors is currentlyunavailable. The brittle nature of the SiC materialmight become more serious as the size of the mirrorbecomes larger. One possibility to overcome these dif-ficulties is a segment-bonded carbon-fiber-reinforcedSiC (C/SiC) composite mirror. C/SiC composite ismade from carbon–carbon (C/C) composite matricesthrough silicon infiltration, in which silicon reactswith amorphous carbon and becomes SiC; carbonfibers are incorporated into SiC to reinforce andimprove the fracture toughness significantly. In thesilicon infiltration process, bonding of segmentscan also be accomplished, which facilitates the fabri-cation of large-size mirror blanks. C/SiC compositehas advantages over conventional SiC such as highstrength for machining, which is particularly benefi-cial to optimizing support structures.We have performed critical-component studies

for the SPICA telescope through cryogenic opticaltesting of C/SiC mirrors. The C/SiC composite mirrormaterial was made by Mitsubishi Electric Corpora-tion (MELCO) [7] and is different from the currentHBCesic [8] being developed by MELCO and ECM.The former aims at increasing stiffness, the latterat improving uniformity and homogeneity. Sincecryogenic performance of the C/SiC composite waspoorly understood, four pieces of small-size C/SiCcomposite sample mirrors (160mm in diameter)were fabricated and tested with step-by-step optimi-zations of the manufacturing process [7,9,10] prior tothe studies presented in this paper. Then we fabri-cated a 720mm diameter bonded C/SiC compositespherical mirror; 12 segments consisting of 6 petalson the front and 6 on the rear surface were bondedtogether at the stage of silicon infiltration into C/Ccomposite matrices. The composite nature of C/SiCrequired careful polishing processes [9]; the mirrorsurface was coated with SiSiC slurry before grinding[7]. The specifications of the mirror are listed inTable 1. The mirror has a 5mm thickness for the sur-face and a 2:5mm thickness for the back-side ribs.

The AKARI telescope exhibited considerablechanges in the wavefront figure at cryogenic tem-peratures; the total wavefront error of the tele-scope was dominated by the cryogenic deformationof the primary mirror due to its support structures[11–13]. One of the lessons learned from AKARI isthat, even if the cryogenic performance of a mirroralone, as well as the mechanical performance of sup-port structures of a mirror, is perfect, support struc-tures can seriously degrade the optical performanceof a telescope at cryogenic temperatures. Thereforecareful consideration is definitely needed for mirrorsupport mechanisms for the SPICA telescope, andthus we have also developed an innovative supportmechanism for cryogenic applications.

In this paper, we report results of cryogenic opti-cal testing of the 720mm diameter segment-bondedC/SiC composite mirror in combination with thenewly developed mirror support mechanism. Thetechnologies presented here and potentially applic-able to the SPICA telescope have been developedby MELCO in collaboration with the Institute ofSpace and Astronautical Science, Japan AerospaceExploration Agency (ISAS/JAXA). The mirror itselfis a breadboard model for the SPICA telescope.

2. Measurement

A schematic view of the mirror unit that consists ofthe C/SiC composite mirror and the support mechan-ism is presented in Fig. 1. The C/SiC composite mir-ror without the support mechanism was first testedfrom September to October in 2005, and then themirror unit was tested from April to May in 2006,both by using the liquid-helium chamber at ISAS/JAXA with the optical axis being vertical [14] asschematically shown in Fig. 2. From the mirror-alonetest, we obtained changes in themirror surface figureas a function of temperature, which were comparedwith model prediction based on measured propertiesof the bonded segments. Then we mounted the mir-ror on an aluminum base plate by using the supportmechanism and evaluated cryogenic performance ofthe support mechanism by subtracting the surfacefigure data obtained in the mirror-alone test.

The setup of the cryogenic optical measurementsis basically the same as that explained in [14]. Asis shown in Figs. 2(a) and 2(b), a ZYGO GPI-XPHRinterferometer is placed on top of the chamber withthree-axis shift and two two-axis tilt adjustment

Table 1. Specifications of Tested 720mm Diameter C/SiC Composite Mirrora

Item Requirement Results

Surface figure error (at 300K) 63nm rms (λ/10) 93nm rms (λ/6.4)Clear aperture 690mm Outer, 610mm

Inner, 200mmSurface roughness < 20nm Ra 17:6 − 19:9nm RaRadius of curvature 1300� 5mm 1299:915mmWeight NA 10:574kg

aRa, roughness average.

10 March 2008 / Vol. 47, No. 8 / APPLIED OPTICS 1123

stages. Changes in the mirror surface figure withtemperature are measured from outside the chamberby the interferometer with a spherical Fizeau lens.Inside the chamber, there is a two-axis tilt adjust-ment stage upon which the tested mirror is placed.During the interferometric measurement, the wholesystem was being sustained by four vibration-free supports placed in a square on the bottom ofthe chamber. The fine adjustment of the optical

alignment of the interferometer with respect to themirror was performed by shifting the interferometeralong the z axis and tilting the mirror in the x andy axes. Measurements of the surface figure were per-formed at various temperatures while themirror wasbeing warmed up, where stable measurements werepossible.

We cooled the mirror by using six copper thermalstraps, which were attached to the rear surface of

Fig. 1. (Color online) Illustrative view of the tested mirror unit that consists of the C/SiC composite mirror, Invar flexures, and 4-degree-of-freedom ball-bearing support mechanism.

Fig. 2. (Color online) (a) Schematic view of the measurement configuration of the mirror. (b) Liquid-helium cryochamber at ISAS/JAXAthat was used for optical testing of the mirror. (c) Front surface of the 720mm diameter bonded C/SiC composite mirror mounted on theadjustment stage in the ISAS/JAXA cryochamber. (d), (e) Same C/SiC composite mirror as shown in Fig. 4 below, but with the newlydeveloped support mechanism fixing the mirror onto the aluminum base plate.

1124 APPLIED OPTICS / Vol. 47, No. 8 / 10 March 2008

the mirror; the other end of each thermal strap wasanchored to the copper bar that stands uprightaround the circumference of the 4K stage. Two ther-mometers were attached at the back-side ribs of themirror; one was attached at the place close to thethermal anchors, and the other was at the placefar from them. Except for the initial cooling phase,the difference in temperature between the two ther-mometers attached to the mirror was no morethan 2K, indicating a good thermal conductivity ofthe C/SiC material even at cryogenic temperatures.

3. Results

A. Evaluation of Coefficient of Thermal ExpansionNonuniformity Effects at Cryogenic Temperatures

Nonuniformity and inhomogeneity in the coefficientof thermal expansion (CTE) are potential problemsinherent in composite materials such as the C/SiCmaterial; the segments were prepared from differentC/SiC composite lots, for each of which the CTE of asample was measured. We deliberately combined thesegments with different CTEs so that the measure-ments might enable us to evaluate the effects ofsmall CTE differences among segments on the cryo-genic deformation of the mirror. On the other hand,the CTE inhomogeneity, i.e., difference between in-plane and out-of-plane CTEs, causes the change inradius of curvature of a mirror. The effect is, however,verified to be well correctable within the wavefronterror requirement of the SPICA telescope by introdu-cing a focus adjustment mechanism as adopted in thesecondary mirror of the AKARI telescope that hasbeen operated successfully in orbit [11].Figure 2(c) shows the front surface of the C/SiC

composite mirror mounted on the tilt adjustmentstage in the chamber. We set three fiducial pointson the mirror surface as shown in the figure for sub-traction of surface figure data sets measured at dif-ferent temperatures. On the rear surface of themirror, there are three annular C/SiC plates, whichwere also bonded to the mirror at the silicon infiltra-tion stage; they are used as an interface fitting to thesupport mechanism described below. In this mea-surement of the mirror without the support mechan-ism, the mirror was just put on the stage through ametal ball contact at each of three supporting points.In Figs. 3(a)–3(d), the surface figures thus mea-

sured for the C/SiC composite mirror are shownafter subtraction of the initial surface figure mea-sured at room temperature before cooling and aregiven as half a wavefront error measured by theinterferometer. The joint interface lines betweenthe segments are somewhat obscured, as can be re-cognized in Figs. 3(a)–3(d), where the data are lostdue to vibration disturbances during interfero-metric measurements; nevertheless, there seems tobe no local deformation associated solely with thebonding.Figures 3(a)–3(d) show large astigmatism at low

temperatures; the cryogenic deformation of the

mirror is found to be unacceptably large for theSPICA mirrors. As is shown in Fig. 3(e), however,a finite-element model analysis of the mirror surfacefigure along with properties of the material mea-sured for 12 segment lots, such as CTEs, densities,and Young’s modulus, reproduces the pattern of themeasured cryogenic deformation very well, whichalso shows an excellent quantitative agreement withthe measurement. Table 2 and Fig. 4 show the prop-erties of the samples from the bonded segments; toobtain the CTEs averaged over a temperature rangeof 300–4:5K for all the segments, the measurementof a single sample at 4:5K was utilized by apply-ing a different scaling factor from sample to samplethat reflects CTE variations measured at 300K. Theabove consistency between the model prediction andthe measurement proves such evaluation of CTEsto be accurate enough. The measured CTE has avariation of 1.9% among lots at 300K, for whichwe have deliberately selected those with differentCTEs. We have found that such a CTE nonuniformitylevel of mirror material causes serious deformationof the mirror as shown in Figs. 3(a)–3(d). To meet thewavefront error requirement of the SPICA telescope,we have to reduce the CTE nonuniformity level by atleast a factor of 4–5, which could be achieved byselecting segments with CTEs more similar to eachother and carefully arranging them to form a mirror.

In addition, as shown in Fig. 5, where the rms sur-face figure errors are plotted as a function of tem-perature, there is almost no significant change inthe surface figure below 100K; thanks to the natureof nearly zero CTEs at low temperatures, the testtemperature will not necessarily be well below 100K,which will considerably facilitate evaluation ofthe cryogenic optical performance of large-aperturecooled space telescopes.

B. Cryogenic Performance of Newly DevelopedSupport Mechanism

By considering the importance in designing supportstructures for the SPICA mirror as stressed above,we have also developed a 4-degree-of-freedom cera-mic ball-bearing support mechanism for cryogenicapplications, which is based on a similar techniquedeveloped for the Japanese satellite project, Hinode[15] (Solar-B) at nearly ambient temperatures. TheC/SiC composite mirror was supported by using thismechanism [Figs. 1, 2(d), and 2(e)] and tested againat cryogenic temperatures. Invar flexures were usedto connect the C/SiC interface fitting to the supportmechanism. From separate component tests, wehave confirmed that there is no significant increasein a friction torque of the ball-bearing support me-chanism at 4:5Kwith good reproducibility amongcooling cycles [16].

Details of the design of the support mechanismwillbe reported in the future. The essence is that mate-rials with different CTEs are carefully combined sothat they may prevent a significant increase in fric-tion between a ball bearing and its casing at any

10 March 2008 / Vol. 47, No. 8 / APPLIED OPTICS 1125

temperature from 300 to 4:5K. We used 12 titaniumM5 screws per supporting point to connect the mirrorto the support mechanism via Invar flexures asshown in Fig. 1. We applied a torque in tighteningthe screws to secure a safety factor of ∼40 againstslipping due to thermal stress at 4:5K and ∼1 dueto mechanical stress for launching vibration at thequalification test level.

As seen in Figs. 2(d) and 2(e) the C/SiC compositemirror was firmly fixed on the aluminum base plateby using the support mechanism and was tested atcryogenic temperatures. Figure 6 shows the cryo-genic deformation of the mirror with the supportmechanism, shown after subtraction of the cryo-genic deformation of the mirror without the supportmechanism measured at similar temperatures; both

Fig. 3. (Color online) Surface figures of the C/SiC composite mirror measured at (a) 16K, (b) 92K, (c) 192K, and (d) 263K after sub-traction of the figure at 290K. (e) Mirror surface figure at 4:5K predicted by a finite-element model analysis with the measured propertiesof the materials from the 12 segment lots. All values concerning the surface figure are given in units of the He–Ne laser wavelength of632:8nm.

Table 2. Properties of the C/SiC Samples from Constituent Segments of the Mirrora

Properties Measurement (averaged over samples)

Density 2:718 g=cm3

Young’s modulus 298GPaBending strength 156MPaCTE (at 300K) 2:39ppm=K in plane, 2:75ppm=K out of plane

Variation: σ ∼ 1:95%Averaged CTE (4:5– 300K) 0:70ppm=K in plane, 0:84ppm=K out of planeSpecific heat (at 300K) 7:22J=gK in plane, 7:18J=gK out of planeThermal conductivity (at 300K) 160W=mK in plane, 125W=mK out of planeThermal conductivity (at 30K) 13:5W=mK in plane, 14:0W=mK out of plane

appm, parts in 106.

1126 APPLIED OPTICS / Vol. 47, No. 8 / 10 March 2008

cryogenic deformations are derived by subtractingthe surface figures at room temperature. The changeof the mirror surface figure caused by attaching thesupport mechanism is plotted in Fig. 5. We have ob-served a significant change in cryogenic deformationof the mirror after implementing the support me-chanism, which consists of local humps at support-ing points due to CTE mismatch between C/SiC andInvar as well as very small trefoil deformation due to

thermal contraction of the aluminum base platethrough bearing friction. The former local deforma-tion is quite reasonable because we have not opti-mized the design of the C/SiC interface fitting, i.e.,to increase the stress propagation length to the mir-ror; C/SiC composite material would enable us tomake more complex structures. We evaluate the tre-foil deformation by three-sided astigmatism throughZernike polynomial fitting to the surface change; thetrefoil deformation is inevitable as long as there isCTE mismatch between the mirror and the opticalbench. The trefoil deformation is then estimated tobe 0:05 μmP–V. Considering that the trefoil deforma-tion of the AKARI primary mirror is 0:41 μm P–V bythe same evaluation method, the newly developed

Fig. 4. (Color online) CTEs measured at 300K for the 12 seg-ments of the mirror, 6 on the front (left) and 6 on the rear surface(right) given in units of parts in 106, with the in-plane CTEs in theupper panels and the out-of-plane CTEs in the lower panels. Thevalues in parentheses are lot numbers.

Fig. 5. Change of the surface figure of the C/SiC compositemirrorwith temperature measured during the warm-up process (Fig. 3).For comparison, the change of the mirror surface figure caused byattaching the support mechanism is also indicated by the dottedline (see Fig. 6).

Fig. 6. (Color online) Cryogenic deformation of the mirror with the support mechanism, shown after subtraction of the cryogenicdeformation of the mirror without the support mechanism measured at similar temperatures.

10 March 2008 / Vol. 47, No. 8 / APPLIED OPTICS 1127

support mechanism absorbs the CTE difference be-tween the C/SiCmirror and the aluminum base platefar better than the usual titanium bipod flexuresadopted for the AKARI telescope; this will in turn im-prove the flexibility in designing back-side structuresof a telescope mirror, including structures for focal-plane instruments. Hence the newly developed sup-port mechanism is promising for its application tothe SPICA telescope, whereas it has to be provedin a vibration test that the present mechanism hassufficient strength to retain optical alignment.

4. Conclusion

One of the primary candidates for the mirror mate-rial for the SPICA telescope is C/SiC composite. Wehave fabricated a 720mm diameter bonded C/SiCcomposite mirror and tested it at cryogenic tempera-tures. Twelve segments consisting of six petals onfront and rear surfaces were bonded together at thestage of silicon infiltration into C/C composite ma-trices. The segments were prepared from differentlots, for each of which the properties of C/SiC compo-site samples including CTE were measured. Hencethe measurements have enabled us to evaluate theeffects of possible small CTE differences among seg-ments on the cryogenic deformation. The measure-ment shows significant cryogenic deformation ofthe C/SiC composite mirror, which is well reproducedby a model analysis with measured properties of theconstituent samples. We have found that a 1.9% var-iation in CTE among segments causes serious cryo-genic deformation of a mirror with a surface figureerror of about 1:2 λ rms that is more than 4 times lar-ger than that required for the SPICAmirror; we haveto select segments with CTEs more similar to eachother for applications to the SPICA telescope. In par-allel, we have also developed a 4-degree-of-freedomball-bearing support mechanism for cryogenic appli-cations, which is based on a similar technique devel-oped for the Japanese satellite project, Hinode. Thenthe C/SiC composite mirror was mounted on an alu-minum base plate by using the support mechanismand was again tested at cryogenic temperatures. As aresult, we found that cryogenic deformation of themirror attributed to thermal contraction of the alu-minum base plate via the support mechanism is verysmall, and thus the newly developed support me-chanism is promising for its future application tolarge-aperture cooled space telescopes such as theSPICA telescope.

The SPICA project is managed by ISAS/JAXA incollaboration with JAXA, ISTA/JAXA, National As-tronomical Observatory of Japan, universities, andother research institutes in Japan; international col-laboration with Europe, Korea, and the U.S. is alsounder consideration. The authors are grateful toall the members of the SPICA project for their helpand support. We are also indebted to the staff of therelevant companies who are working on the SPICAproject for their great effort in designing the tele-scope system and developing mirror technologies.

References

1. T. Nakagawa and H. Murakami, “SPICA, cooled telescope op-timized for mid- and far-infrared astronomy,” Adv. Space Res.40, 679–683 (2007).

2. T. Onaka, H. Kaneda, K. Enya, T. Nakagawa, H. Murakami,H.Matsuhara, and H. Kataza, “Development of large aperturecooled telescopes for the space infrared telescope for cosmol-ogy and astrophysics (SPICA) mission,” Proc. SPIE 5962, 448–462 (2005).

3. T. Onaka and T. Nakagawa, “SPICA: a 3:5m space infraredtelescope for mid- and far-infrared astronomy,” Adv. SpaceRes. 36, 1123–1127 (2005).

4. H. Murakami, “The infrared astronomical mission AKARI,”Publ. Astron. Soc. Jpn. 59, 369–376 (2007).

5. T. Onaka, Y. Sugiyama, and S. Miura, “Telescope system of theInfrared Imaging Surveyor (IRIS),” Proc. SPIE 3354, 900–904 (1998).

6. H. Kaneda, T. Onaka, M. Kawada, and H. Murakami, “Cryo-genic optical testing of sandwich-type silicon carbide mirrors,”Appl. Opt. 42, 708–714 (2003).

7. T. Ozaki, M. Kume, T. Oshima, T. Nakagawa, T. Matsumoto,H. Kaneda, H. Murakami, H. Kataza, K. Enya, Y. Yui,T. Onaka, and M. Kroedel, “Mechanical and thermal perfor-mance of C/SiC composites for SPICA mirror,” Proc. SPIE5868, 132–141 (2005).

8. M. R. Krödel and T. Ozaki, “HBCesic composites forspace optics and structures,” Proc. SPIE 6666, 66660E(2007).

9. K. Enya, T. Nakagawa, H. Kaneda, T. Onaka, T. Ozaki, andM. Kume, “Microscopic surface structure of C/SiC compositemirrors for space cryogenic telescope,” Appl. Opt. 46, 2049–2056 (2007).

10. H. Kaneda, T. Nakagawa, T. Onaka, T. Matsumoto, H. Mura-kami, K. Enya, H. Kataza, H. Matsuhara, and Y. Y. Yui, "De-velopment of space infrared telescope for the SPICA mission,"Proc. SPIE 5487, 991–1000 (2004).

11. H. Kaneda, W. Kim, T. Onaka, T. Wada, Y. Ita, I. Sakon, andT. Takagi, “In-orbit focal adjustment of the AKARI telescopewith Infrared Camera (IRC) images,” Publ. Astron. Soc. Jpn.59, 423–427 (2007).

12. H. Kaneda, T. Onaka, T. Nakagawa, K. Enya, H. Murakami,R. Yamashiro, T. Ezaki, Y. Numao, and Y. Sugiyama, “Cryo-genic optical performance of the ASTRO-F SiC telescope,”Appl. Opt. 44, 6823–6832 (2005).

13. H. Kaneda, T. Onaka, T. Nakagawa, K. Enya, H.Murakami, R. Yamashiro, T. Ezaki, Y. Numao, andY. Sugiyama, “Wavefront measurement of space infraredtelescopes at cryogenic temperature,” Proc. SPIE 5965,281–295 (2005).

14. H. Kaneda, T. Onaka, R. Yamashiro, and T. Nakagawa, “Op-tical performance of the ASTRO-F telescope at cryogenic tem-peratures,” Proc. SPIE 4850, 230–240 (2003).

15. K. Ichimoto, S. Tsuneta, Y. Suematsu, T. Shimizu, M. Otsubo,Y. Kato, M. Noguchi, M. Nakagiri, T. Tamura, Y. Katsukawa,M. Kubo, Y. Sakamoto, H. Hara, K. Minesugi, A. Ohnishi,H. Saito, N. Kawaguchi, T. Matsushita, T. Nakaoji, K. Nagae,J. Sakamoto, Y. Hasuyama, I. Mikami, K. Miyawaki, Y.Sakurai, N. Kaido, T. Horiuchi, S. Shimada, T. Inoue, M.Mitsutake, N. Yoshida, O. Takahara, N. Takeyama,M. Suzuki,and S. Abe, “The Solar Optical Telescope onboard the Solar-B,”Proc. SPIE 5487, 1142–1151 (2004).

16. M. Haruna, M. Tabata, T. Oshima, T. Onaka, K. Enya, H.Kaneda, and T. Nakagawa, “Bearing test for mirror supportmechanism of space infrared telescopes—evaluation of hybridball bearing characteristics in cryogenic environment,” in 25thInternational Symposium on Space Technology and Science(2006).

1128 APPLIED OPTICS / Vol. 47, No. 8 / 10 March 2008