The motivation for the development of hard-x-ray focusing optics, those operating at energies approaching 100 keV, is clear. Grazing incidence telescopes, such as the one at the heart of the Chandra X-Ray Observatory, have brought about spectacular advances at soft-x-ray wavelengths (below 10 keV). The hard-x-ray region, where such optics are yet to be routinely used, remains relatively unexplored at high sensitivity and fine angular resolution, however.1 The power of focusing, which concentrates source flux but not background, is such that even modest collecting areas can give a large increase in sensitivity over non-focusing devices.
The critical angle δc below which x-rays can be efficiently scattered from smooth surfaces is quite small, hence the term grazing incidence. Away from absorption edges, δc can be approximated as
where E is the x-ray energy in keV and ρ is the density of the reflecting material (typically gold or iridium). For soft x-rays (around 1 keV), the critical angle is a few degrees. It scales approximately inversely with energy, however, and therein lies the challenge for hard-x-ray optics: As the graze angle decreases, the projected area of a mirror shell becomes very small. To overcome this, we must use a large number of mirror shells and as long a focal length as possible. For convenience, we can nest the shells in a mirror module, then use multiple modules, each with its own focal plane x-ray detector.
German physicist Hans Wolter proposed several mirror shell configurations, including the Wolter I (see figure 1).2 In a Wolter I design, like that used for Chandra, incoming photons undergo two reflections, the first from a parabolic surface and the second from a hyperbolic surface, to give an image that is essentially coma free.
Figure 1. This mirror configuration contains four nested Wolter I shells, each consisting of a paired hyperboloid of revolution and paraboloid of revolution.
There has been much discussion of coating x-ray optics with graded multilayers, which, unlike conventional coatings, can exhibit moderate reflectivity at grazing angles several times the critical value. Although much progress has been made with these coatings, they place stringent requirements on surface micro-roughness and require the precise deposition of hundreds of very thin layers of material on the inside of each mirror shell. These factors make multilayer-coated x-ray optics unattractive, at least for budget-limited projects like the balloon-payload component of our NASA research program. Hard-x-ray telescopes with conventional coatings offer certain advantages over a multilayer-coated telescope, including greater effective area per unit mass, less diffractive scattering by surface micro-roughness, and less stringent manufacturing requirements.3Electroformed Nickel Replication
The mirror fabrication process that we are developing is that of electroformed nickel replication (ENR), in which nickel mirror shells are electroformed onto a figured and superpolished aluminum mandrel from which they are later released by differential thermal contraction. This process was pioneered in Italy for x-ray mirror fabrication and has been used for soft-x-ray astronomy in such missions as XMM-Newton, which featured three mirror modules, each with 68 electroformed nickel shells.4
A distinct advantage of the electroforming process is that the resulting mirror shells are full circles of revolution and thus are inherently very stable. This stability permits good figure accuracy, and hence very good angular resolution. A second advantage is that multiple identical copies can be made from a single mandrel and this permits the easy fabrication of multiple mirror modules. One drawback of the approach for future space applications, however, is the high density of nickel, which necessitates very thin shells to achieve the lightweight optics necessary to keep costs reasonable. These shells must be strong enough to withstand the stresses of fabrication and subsequent handling without permanent deformation. They must also be electroformed in an ultra-low-stress environment to prevent stress-induced distortions once they are released.
To ensure high-quality optics, we have therefore made developments in material strength, adhesion and release, and plating-bath stress control.5,6 The High Energy Replicated Optics (HERO) evolutionary balloon program provides us with a testbed for our techniques. The program involves in-house-fabricated hard-x-ray mirrors plus detectors, and a gondola and pointing system. The design incorporates a large number of shallow-graze-angle, iridium-coated nickel shells, moderately nested in multiple mirror modules.
Figure 2. The nickel/cobalt alloy electroforming bath with the multipart plater produces replicated mirror shells.
The use of a large length-to-diameter ratio for the shells reduces the number of mandrels. To keep costs appropriate to a balloon program, we grind the mandrels to sub-micron-accuracy figure and then polish them to 3 to 4 Å rms surface roughness with simple, in-house-designed machines. A multipart plater then allows us to electroform several shells simultaneously (see figure 2). HERO in Flight
The May 2001 flight of a HERO prototype produced the first hard-x-ray images of astronomical objects and demonstrated sub-arcminute angular resolution with six 3-m-focal-length mirror shells.7 Since that time, we've refined our mirror fabrication techniques and a much larger payload that will eventually feature 240 6-m-focal-length mirrors is under construction. These mirrors will be arranged in 16 housings each containing 15 nested shells ranging from 50 to 94 mm in diameter. At the time of this writing, a HERO partial-payload flight for spring was slated to carry eight mirror modules, each with eight mirror shells for a total effective area of 60 cm2 at 40 keV. The shells are currently undergoing x-ray testing at Marshall Space Flight Center (MSFC; Huntsville, AL) using the Center's 100-m beam tube.
One goal with the new balloon payload was to improve the mirror angular resolution substantially over those originally flown in 2001. Typical metrology for current mandrels gives a performance prediction for the shells of around 8 to 10 arcsec half-power diameter (HPD), meaning half the reflected flux from a point source falls within this angular range. X-ray tests reveal shell performances in the 13- to 15-arcsec range with modules running around 17 arcsec HPD. This module performance is about a factor of three better than our original units, but further improvements are possible.
The current HERO mandrels represent conical approximations to Wolter-I geometry. This approach simplifies fabrication and reduces the cost to a level appropriate for a balloon program, but limits the performance to a theoretical 6 to 10 arcsec HPD for the range of shell diameters, assuming perfect replication. The difference between the predicted performance and the actual shell measurements is due to minute figure distortions in the shell fabrication process. We are investigating annealing techniques that may help minimize these effects. Finally, errors in the mounting of concentric shells can be overcome through an optical monitoring system that ensures that all shells are co-aligned and circular before bonding to the support spider takes place. The combination of these techniques should permit the production of even higher-resolution optic modules. Moving Forward
Another component of our ENR development work, in conjunction with the Smithsonian Astrophysical Observatory (Cambridge, MA) and the Astronomical Observatory (Brera, Italy), concerns building and testing a prototype hard-x-ray telescope for possible use on NASA's Constellation-X mission, a follow-on to Chandra planned for launch in about 2015.8 Our role in the project is the production of two demonstration shells, one 150 mm in diameter and coated with sputtered iridium, and the other 230 mm in diameter to be coated with graded multilayer thin films. The challenging aspects of this work are the very thin mirror shell walls (100-µm thick) dictated by the tight Constellation-X weight budget, and the high-surface-quality finish (better than 3 Å rms) on the interior of the shells designated for multilayer coating.
Figure 3. A testbed for the Constellation-X project consists of a 100-µm-thick shell, shown here mounted in a support ring.
We have fabricated test shells to demonstrate the viability of 100-µm-thick shell production and to evaluate the effects of coating stresses on the shell figure (see figure 3). Measurements on these shells indicate that multilayer stresses do not produce measurable figure distortions and that simple support rings can permit handling and mounting of ultra-thin mirror shells while maintaining their shape.
Both Constellation-X mandrels have been completed and are ready for electroforming. Interferometer data indicates surface roughnesses of 2.7 Å rms for the larger mandrel and a figure error of less than 0.1 µm rms. The performance prediction for shells from this mandrel, a conical approximation, is 10 arcsec HPD. We will fabricate test shells from these mandrels later this year.
Electroformed nickel replication offers an extremely attractive and economical solution to the problem of hard-x-ray optics production. The process lends itself readily to the multiple-mirror-module approach that small graze angles necessitate and the resulting shells provide excellent angular resolution that results in high sensitivity observations. Finally, with the use of high-strength alloys one can achieve the stringent weight budgets that future missions require. oe
Eyes on the Sky
When you love your work, time passes quickly, even when you've been very busy for 29 years.
Martin C. Weisskopf, project scientist at NASA in charge of developing NASA's Chandra X-Ray Observatory, and chief scientist for x-ray astronomy in the space sciences department at NASA's Marshall Space Flight Center (Huntsville, AL), has been working on the Chandra X-Ray Observatory since 1977.
After receiving his bachelor's degree with honors in physics from Oberlin College (Oberlin, OH) in 1967 and his doctorate in physics from Brandeis University (Waltham, MA) in 1969, Weisskopf began his post-graduate career at Columbia University (New York, NY), where he became an assistant professor and performed many pioneering experiments in x-ray astronomy. In 1977, Weisskopf left Columbia to become senior x-ray astronomer at Marshall Space Flight Center and Chandra X-Ray Observatory project scientist.
Weisskopf has held numerous special appointments and earned many awards during his career, including his appointment as senior co-investigator of the European Space Agency's international x-ray imaging experiment, called IBIS.
Weisskopf's award list includes NASA Medals for Exceptional Service in 1992 and for Scientific Achievement in 1999, as well as Fellow of the American Physical Society in 1994. In 2001, he was selected as an SPIE Fellow for his significant scientific and technical contributions to the Society and the optics community, followed by the Rossi Prize of the High Energy Astrophysics Division of the American Astronomical Society in 2004, which he shared with H. Tananbaum. Weisskopf has also authored or co-authored 225 publications, including refereed journal articles, articles in books, monographs, and papers in conference proceedings.
2. H. Wolter, Annalen der Physik 10 (1952).
3. M. Weisskopf, et al., "Graded multilayers not required for hard x-ray imaging!" Proceedings of The Next Generation of X-Ray Observatories, University of Leicester, Leicester, England, July 10-12 (1996).
7. B. Ramsey, et al., Astrophysical Journal 568, p. 432 (2002).
Brian Ramsey, Martin Weisskopf
Brian Ramsey is an experimentalist and Martin Weisskopf is chief scientist for x-ray astronomy in the space sciences department, NASA/Marshall Space Flight Center, Huntsville, AL.