2.1.

Components

There were a total of nine separate subassemblies that composed the combined ApA and BF subsystem. The name of each was derived from the dewar interface to which it attached mechanically and thermally. These interfaces are listed and defined in Table 1.

Table 1

Dewar interfaces of the SXS ApA and BFs.

The ApA was integrated into the dewar one subassembly at a time and never constituted a complete assembly by itself outside of the dewar. Starting from the inside and progressing outward, these subassemblies and their interfaces were as follows:

Figure 1(a) is a cross-sectional schematic of the SXS ApA. Note that the CTSL-F, DAL, JTB, IVCS-M, and the ApC-SA were not in physical contact but were held in position by the different temperature stages. Figure 1(b) is the same view without the CTS, with the filter-carrier assemblies removed and shown separately to the side. The four filter carriers (FC) were installed after the dewar main shell was attached.

Fig. 1

Schematic cross section showing the arrangement of the Hitomi/SXS ApA and BFs. (a) With FCs installed and position of the fixed CTS filter shown. The thin film of the CTS filter is not shown to instead show the CTS aperture more clearly. (b) With the FCs removed, and the CTS lid not shown. The tops of the dewar shields are also shown, for reference, but the multilayer insulation has been omitted.

2.2.

Thermal Design, Analysis, and Performance

Thermal design drivers for the ApA included minimizing heat transfer between various dewar stages, meeting filter temperature requirements during normal operation, and providing for decontamination of the filters through heaters and interface design. The thermal design was complicated by the steep temperature gradients, the need to establish a seal around each filter to block contamination and radiation, and the accommodations for decontamination heaters on the filters. Key features of the mechanical/structure and thermal design, shown in Fig. 1, were:

The DMS and shield temperatures were projected to vary considerably during prelaunch testing and throughout the mission. The minimum shield temperatures occurred on the ground during liquid helium transfers; thus, subsystem testing included these minima, with margin, as survival temperatures. The full range of flight and ground-operation temperatures for the integrated SXS shown in Table 2, including margin (10 K on DMS and 10% on cryogenic components), was used to verify the thermal design by analysis and to establish test ranges. Note that the ground minimum for the DMS filter shown in Table 2 is the actual minimum temperature of the DMS filter recorded during thermal vacuum testing. The nominal hot case temperatures and the resultant heat flows are also shown. This case was important as it was used to determine the He lifetime for the SXS mission. The heat flows shown represent the results from a thermal model (developed using the C&R Technologies, Inc. analysis packages SINDA/FLUINT and C&R Thermal Desktop^®). This model was integrated into the dewar thermal model maintained by JAXA and Sumitomo Heavy Industries, the dewar developer. Radiative heat transfer and reflected energy along the exterior of the aperture cylinder were included in the ApA thermal model.

Table 2

Heat flows and temperatures for the SXS ApA.

Various thermal properties measured during subsystem thermal-balance testing are summarized in Table 3. The ApC-SA, DMS-FC, and OVCS-FC were tested together in a nonflight apparatus designed to simulate the flight-dewar thermal interfaces and radiation environment. The IVCS-M and IVCS-FC were qualified down to 4 K (using the engineering model parts) and characterized for flight from a liquid nitrogen interface.

Table 3

Significant results of subsystem thermal balance testing.

2.3.

Resonant Frequencies

The structural resonant frequencies of the ApA were assessed for each of the major components. A detailed finite element model and closed-form equations were used to identify minimum eigenvalue frequencies. Table 4 summarizes the minimum values, showing the analytical prediction and corresponding requirement. All predicted frequencies are axial (spacecraft $Z$ axis) except the ApC-SA, which is lateral. All results met project requirements.

Table 4

Fundamental frequencies of the SXS ApA subassemblies.

Due to the size and fragility of the filter and carrier assemblies, frequency measurements were not performed on these components. Results from cold vibration testing of the IVCS-M and FC assembly showed a minimum frequency of 1360 Hz. This was scaled from the $\sim 80\mathrm{K}$ measurement temperature to the operational temperature of $\sim 20\mathrm{K}$. In ApC-SA vibration tests, the minimum frequency was 831 Hz.

There was adequate margin on all clearances under conservative launch loads. The radiation and contamination baffles were designed to meet functional requirements with nominal lateral and axial clearances of $\sim 5\mathrm{mm}$.

2.4.

Thermomechanical Testing

A series of tests qualified the basic design features and verified the integrity of the flight units. Thermal and mechanical testing of the three silicon-mesh filters was done in their carriers to provide the appropriate mechanical and thermal interface. Operation of the heaters and thermometers required that the carriers be installed in mounts. These tests are summarized in Table 5. Note that vibration qualification of the ApC-SA with OVCS-FC and DMS-FC was done at room temperature since the nitrogen dewar used for vibration testing would have cooled the cylinder interface to $\sim 80\mathrm{K}$, introducing more thermal mismatch at the mount/tube interface than would be present at launch. Vibrating in this condition would be an over test. The under testing of a warm vibration was mitigated by separately measuring the strain at nominal operating temperatures via strain gauges affixed to a prototype cylinder. The measured strain values confirmed the assumptions for thermal strain that had been input to the analysis of the design.

Table 5

Thermo-mechanical tests of the SXS ApA. FM, flight model; EM, engineering model.

All ApA components that were launched on Hitomi passed their thermomechanical acceptance tests. Spare ApC-SA and IVCS-M units and DA and CTS filters also passed, but some damage was experienced on the IVCS, OVCS, and DMS filters, which resulted in changing their ranking. These minor failures were mostly attributable to factors incidental to the tests themselves, such as from debris, pre-existing defects, or installation tools, as will be discussed along with other lessons learned in Sec. 7.

3.1.

Thin-Film Filters and Their Support Structures

Low temperature x-ray microcalorimeters require very thin filters for rejecting non-x-ray radiation while providing the highest possible transmission in the x-ray band (0.1 to 10 keV). Obtaining high x-ray transmission requires films that are only about 1000 Å thick, composed of polymer film coated with aluminum. These filters have diameters that can be as large as several centimeters (see Table 6) and are extremely delicate. The use of these filters in the ultralow-temperature environment required by the SXS is even more challenging since the filters can become tight when under vacuum (due to moisture loss) and very brittle when cold.

Table 6

Locations and aperture diameters of the SXS BFs.

The BFs were designed to prevent radiation with wavelengths longer than x-rays from reaching the detector, where it would heat the pixels or induce photon shot noise. The filters also closed out various thermal shields to minimize heat loads on the cryogens. For the entire stack of filters, the required maximum out-of-band transmission of the filters was 0.006 at 40.8 eV (geo-coronal He II), 0.02 at 21.2 eV, 0.002 at 10.2 eV, and $5\times {10}^{-9}$ in the IR range (3 to $30\mu \mathrm{m}$). The requirement for attenuation of the geocoronal line was satisfied by the telescope thermal shield, with no additional requirement on the filters. The required minimum in-band transmission for the stack was 16% at 0.6 keV, 52% at 1.0 keV, 70% at 6 keV, and 70% at 10 keV. The attenuation requirements were met by design and measurements from the XRS programs, and the white-light transmission of each filter was measured to assess pinhole transmission. The in-band transmission requirements were verified with x-ray measurements.

The SXS filters used aluminized polyimide and were highly reflective in the visible and IR. The UV attenuation was determined by both the aluminum and polyimide thicknesses. The chosen thicknesses came from trading off the high-throughput at low energies, the margin in the thermal design and detector noise budget, and the mechanical strength. Strength is required to survive numerous thermal cycles of low temperatures, mechanical vibration, and pressure gradients of a few hundred Pa. The outer filter was also required to limit the integrated diffusion of helium into the dewar to $<0.5\mu \mathrm{g}$, given a required helium pressure in the spacecraft of $<{10}^{-6}\mathrm{Pa}$. The nominal thicknesses of the films used in the SXS filters are tabulated in Table 7. The composition of each filter was not tailored individually. Instead, two compositions, one for the CTS/DA filters and a thicker one for the IVCS/OVCS/DMS filters, were chosen to meet requirements in aggregate, even in the case of minor damage, such as the development of pinholes. Results of the calibration of the filter transmissions are reported in Ref. 9.

Table 7

Specified thicknesses of the thin films used in the SXS BFs.

The CTS and DA filters had Al frames and were procured from Luxel Corporation as integrated units. For the DMS, OVCS, and IVCS filters, the silicon-mesh frames were provided by the University of Wisconsin, the aluminized polyethylene terephthalate (PET) close-outs (discussed in Sec. 3.3) were produced at Goddard, and the thin-film filter material was supplied by Luxel. Luxel integrated their thin films with the support frames and close-outs at their facility. Specialized tooling and bake-out procedures were developed by Luxel to optimize contact between the filter film and the silicon grid. The development of the mesh filters was achieved via the collaboration of these three institutions.

The University of Wisconsin developed the monolithic two-level silicon support meshes for the three larger filters. Each mesh contained a coarse hexagonal grid that was 0.2 mm thick. Within each 5.28-mm-wide coarse hexagonal cell was a fine mesh that was either 0.025 mm (IVCS, OVCS) or 0.008 mm (DMS) thick. The layout of the 0.33 mm fine-mesh hexagons was the same in each complete coarse-mesh hexagon, independent of filter type. Figure 2 shows the coarse and fine mesh designs. A backing ring of 0.4-mm-thick Si with the kinematic-mount contact features discussed in Sec. 3.3 was glued to each mesh. Figure 7 is a photo of the OVCS filter that clearly shows the two-level mesh structure. The use of 0.025-mm-thick fine mesh had been initially intended for all three of the mesh-supported filters, but stress fractures that would develop during their fabrication prevented any of the DMS meshes from being completed without damage; thus, the thinner mesh was adopted for these larger filters.

Fig. 2

Dimensions (in microns) of (a) the coarse-mesh and (b) fine-mesh cells of the DMS, OVCS, and IVCS BFs of SXS.

If the pressure limit of the mesh-supported filters were dominated by the burst strength of the polyimide, then mesh filters with the even thinner 45-nm polyimide films of the sounding-rocket payload for which they were originally designed¹⁰ should withstand more than 6000 Pa pressure differences. Such a filter was subjected to a pressure differential of 665 Pa applied in both directions (both pushing the film into and away from the mesh) and survived, but no program was undertaken to test an ensemble of filters to failure. In practice, the increased strength was not used to allow increased pressure differentials (as from increased pumping or venting rates) but only to reduce risk.

The open-area fraction of each two-level structure was over 95%, and the Si became transparent at the top of the SXS bandpass. The meshes were constructed from standard silicon-on-insulator wafers that were produced commercially by bonding two wafers together with an oxide layer and then grinding each side down to the desired thickness. These wafers were procured with the desired thicknesses for the front and backing meshes, and one mesh was patterned and etched from each side with a deep reactive ion plasma etch that produced a vertical profile and stopped on the oxide layer between the wafers. The oxide was then removed from the mesh openings with an acid etch. The silicon structure was completed using epoxy to affix the silicon backing ring containing the kinematic-mount interfaces.

3.2.

Filter Heaters and Thermometers

The ApA electrical accommodations included heaters and thermometers on the DMS, OVCS, and IVCS filters. Each of these filters had a thermometer and two separate heater circuits for redundancy. Each heater circuit consisted of several $600\text{-}\mathrm{\Omega }$ microelectronic resistors in parallel. The thermometer and heaters were bonded to the filter frame via epoxy prior to application of the thin film. Gold wire bonds connected the components to each other and to a printed wiring board. The wire-bonding configuration also provided redundancy for both the heater and thermometer circuits, allowing full functionality even if one bond on the positive side and one bond on the negative were to fail. The DMS resistors also featured double bonds between the series components. Connections between each FC and its mount were provided by spring-loaded pins, which were engaged when the FC was installed into its mount. All connections were made through two spring-loaded pins in parallel, for redundancy. Figure 3 shows the schematic of the redundant circuitry and a photo of the heaters and thermometers on the frame of the IVCS filter.

Fig. 3

(a) Electrical schematic of the filter thermometer and heater wiring of a mesh filter and (b) photograph of heater and thermometer components on the IVCS filter.

3.3.

Filter Carriers and Close-Outs

All filters except the CTS were installed after the dewar main shell was attached. Each of these filters required a carrier to provide the mechanical, thermal, and electrical interface (for IVCS, OVCS, and DMS) between the filter and its mount and the mechanical interface to the installation tool.

The DAL-FC was made from aluminum and coated in gold. The DA BF did not have any electrical or thermal features. The DA filter was simply sandwiched between a set of ball plungers and an o-ring and did not have any rotational alignment requirements. Unlike the other filters, the DA filter and its carrier did not become permanently integrated with each other. Figure 4 shows the CTS and DA filters.

Fig. 4

(a) CTS lid with filter (CTSL-F) and (b) DAL filter and carrier (DAL-FC) in nonflight holder.

The IVCS, OVCS, and DMS carrier rings were also made from aluminum and coated in gold. Each carrier had a circumferential o-ring that secured it into its respective mount. Additionally, six ball plungers in the mount supported the carrier and adjusted it to its proper height. Each BF was supported in its carrier via a kinematic-mount system that consisted of three ball plungers and three ball-end Ti set screws. Each pair of a ball-end screw and a ball-headed spring plunger supported the filter and engaged with its kinematic-mount features to constrain motion in 3 (triangular notch), 2 (oblong notch), or 1 (flat contact) degrees of freedom. The kinematic mounting allowed for free movement upon differential thermal expansion and contraction. Figure 5 contains a drawing showing the kinematic mount points on the back of the IVCS filter and a photograph of a triangular notch resting on a ball-end screw. The kinematic-mount notches were laser-cut through-holes that were chemically smoothed.

Fig. 5

(a) Kinematic mount points on the back of the IVCS filter. (b) Photograph of a triangular notch resting on a ball-end screw. The reflection of the ball-end screw is seen in the metalized surface of the silicon frame.

On the IVCS and OVCS-FCs, the structures that held the ball plungers in position also had T-shaped features that were grabbed by the installation and extraction tools. For the DMS carrier, which was inserted with the backing-ring side toward the tool, these pick-up features were on the back of the carrier ring. The orientation of the DMS filter was reversed so that the OVCS and DMS filters could both be installed with the lowest emissivity sides facing the inside of the ApC-SA. The DMS ball-plunger structures also had pick-up features but only for easier handling during assembly.

The small gap between the filter frame and the carrier was covered by a moisture shield (MS) that was also referred to as the close-out. The MS was bonded to the filter frame at its inner diameter and to the carrier at its outer diameter, as shown if Figs. 6 and 7. The MS was made of aluminized PET and incorporated a flexible feature to allow for thermal expansion and contraction of the silicon frame with respect to the aluminum carrier ring. This strain-release feature was formed into the aluminized PET sheet by a heated stainless-steel mold. The MS blocked the passage of light, water vapor, helium, and other contaminants around each filter. Acceptable white-light transmission and helium permeation were verified on samples of the MS material.

Fig. 6

(a) The IVCS, (b) OVCS, and (c) DMS-FCs for SXS. The three triangular structures in each assembly hold the spring-loaded ball plungers.

Fig. 7

(a) Photograph of FM OVCS-FC from the back, revealing the coarse and fine silicon-mesh cells, the moisture-shield interface to the silicon frame, the kinematic mounting, and the spring-loaded electrical contacts. (b) Schematic showing how the MS spanned the gap between the filter and carrier ring, attaching to the back of the filter (shown in the photograph) and the top of the carrier (as seen in Fig. 6).

The IVCS and OVCS carriers included a printed wiring board (PWB) with spring-loaded contact pins. Wire bonds electrically connected thermometers and heaters to the PWB. The spring pins contacted the electrical contact pads in the mount upon installation. The DMS was similar, but the spring-loaded pins were on the mount side of the interface.

Figure 6 shows the DMS, OVCS, and IVCS-FCs. Figure 7 shows the back of the OVCS-FC while being held in the installation tool. This view clearly shows the two-layer mesh, the backing ring, the kinematic-mounting points, the electrical contacts, and the MS interfaces. The equivalent features of the DMS and IVCS filters were conceptually similar.

3.4.

DMS Filter Heating

As a result of projected outgassing rates on the spacecraft, various contamination-mitigation schemes were pursued. The solution chosen was to keep the DMS filter temperature continuously controlled at an elevated temperature after the dewar gate valve was opened. The choice of control temperature for the DMS filter depended on an assessment of the most likely contaminants and what the constituents of the filter could safely tolerate. From analysis of spacecraft outgassing, DEHP (${\mathrm{C}}_{24}{\mathrm{H}}_{38}{\mathrm{O}}_4$) was assumed to be the dominant contaminant, based on the energy-dependent absorption associated with the contamination that accumulated on the Suzaku XIS filters. Measurements of accumulated contamination during baking of the Suzaku spacecraft indicated that the sticking coefficient was zero for surface temperatures $>279\mathrm{K}$. To allow for uncertainty in the nature of the major contaminant, the dependence on surface properties, and the uncertainty in thermal modeling, the plan established for DMS-filter heating was to maintain all points on the DMS filter at a temperature $>300\mathrm{K}$.

The filter temperature was limited by the adhesive used in the construction of the filters. Luxel originally advised limiting the filter temperature to $<65°\mathrm{C}$ (338 K), but that was a standard limit based on an adhesive that was no longer in use at the time of the construction of the SXS filters. Later, Luxel revised the recommended limit to 100°C (373 K). A test of the spare EM DMS filter showed that the filter could be heated over 420 K without introducing any permanent change to the appearance on microscopic visual inspection. Specifically, there was no damage to the fine mesh, no new tears or pinholes in the thin film, and no apparent migration of the epoxy; however, neither out-of-band nor x-ray transmission measurements were made.

The operating plan was to control the temperature of the DMS filter frame at $320\pm 2\mathrm{K}$. The expected required heater power ranged from 0.7 W at the coldest main-shell temperature of 233 K to 0.3 W at a 295 K main-shell temperature, based on subsystem-level testing.

Due to the low thermal conductivity of the silicon mesh and radiative losses from the thin film, the DMS filter was expected to have a gradient from the control thermometer to the center of the mesh in addition to a gradient around the filter frame from the heater/thermometer to 180 deg opposite. SINDA/FLUINT modeling showed that a cold-case DMS temperature of 233 K would result in the highest heater power (850 mW) and therefore the highest gradient, $\sim 23\mathrm{K}$. The coolest area was slightly offset from the center. A lower gradient was projected for the hot case due to the reduced heater power (400 mW).

Given that the model predicted a higher power to control the filter temperature than experimentally determined, we sought a secondary method to increase our confidence that the actual temperature gradient across the filter would be acceptable. A simple analytical model was developed to check the SINDA/FLUINT results and fold in actual measurements. This model addressed the radial gradient in the thin film from radiative losses separately from the lateral gradient in the thicker silicon frame of the filter. This model yielded a value for the conductance of the filter to the carrier that was consistent with the heater power needed for both a 295 and 265 K main shell. This conductance in turn showed that the gradient in the filter frame for the cold case would be $\sim 12\mathrm{K}$. The gradient within the thin film depended on its emissivity. Using a conservative emissivity of 0.2, the coldest point in the filter would be 16 K lower than its perimeter, assuming a uniform temperature at the boundary. Using a more realistic emissivity, the gradient in the frame would dominate the gradient in the film. This simple model shows that conservative assumptions reproduce the scale of the gradients determined in the SINDA/FLUINT modeling.

There are additional layers of conservatism. The cold case represented an extreme DMS temperature that was lower than the expected minimum. In addition, the plan to keep all points of the DMS filter above 300 K included margin over the minimum temperature of 279 K. Thus, we had confidence that the filter could be maintained at an adequate temperature.

3.5.

OVCS and IVCS Filter Heating

The OVCS and IVCS filters were cold enough to condense water vapor that was trapped and released slowly by the multilayer-insulation thermal blankets within the cryostat. The DA filter and the CTS filter were also cold enough to condense water vapor, but they were fully enclosed within cold shields and baffles that do not allow the penetration of water vapor. The primary mitigation for water condensation on the OVCS and IVCS filters was the water vapor barrier in the aperture cylinder and baffling. The secondary mitigation was a heater and a corresponding thermometer to monitor the filter frame temperature during operation.

The desorption rate of water is effectively zero below 140 K and increases rapidly above 150 K. The sticking probability of water on a cold surface is unity in this temperature range. According to Speedy et al.,¹¹ increasing the temperature to 170 K results in a desorption rate of 6 to 12 monolayers per second, depending on the morphology of the ice. Therefore, if ice accumulation was to have become evident in orbit, the IVCS and OVCS filters would be heated to 170 K.

4.1.

Mounts and Baffles

As can be seen in Fig. 1, the ApA components from the OVCS to the DA were arranged with interleaving baffles. These baffles were designed to ensure that no molecular contamination would have a direct line of sight to a filter, but must first encounter a cold baffle surface to which it would stick.

The main bodies of all of the aperture components were made of gold-plated aluminum. Surfaces engaging the installation tool were anodized. The filter receptacles in the DA lid and the IVCS, OVCS, and DMS-M each supported their respective filter via an o-ring groove and six axial ball plungers. Rotational alignment was not required for the DA filter. For the others, a radial slot in the mount provided a visual reference for the rotational alignment, and alignment was actually achieved via the installation tool, which made use of an alignment tab on the carrier. Circumferential tabs on the installation tool engaged slots around each filter receptacle. As the tool pushed the carrier into place, the force was reacted at the tabs.

The IVCS-M and OVCS-M each supported a flat trace harness that terminated on a connector at one end and on a PWB with gold-plated contacts on the other; for the DMS, the PWB contained spring pins. The PWB was located radially and below the o-ring groove in the mount. When the carrier was pressed into position, the spring-loaded pins aligned with the contacts and provided electrical continuity from the filter thermometer and heaters to the connector.

4.2.

Aperture Cylinder

The aperture cylinder consisted of a two-ply gamma-alumina fiber-epoxy composite with a Ti foil liner that was co-cured into the tube. Gamma-alumina fiber-epoxy was chosen because of its very low thermal conductivity combined with good material strength, which allowed the practical limit of only two diagonal plies to provide adequate mechanical strength. Titanium foil (alloy 15-3-3-3, 0.0127 mm thick) provided a barrier to water. Due to design heritage from XRS, a water permeation test was not performed on either a qualification unit or flight unit of the aperture cylinder. Since water will pass through the Ti only through cracks or pinholes, and not diffusion, a light-leak test was performed as a means of verifying the water permeation requirement. Figure 8 is a photograph of the ApC-SA.

Fig. 8

SXS aperture cylinder subassembly, OVCS side up. The harness seen in this view is for the DMS.

To minimize the pressure differential across the filters to 1 Torr (133 Pa) or less during (i) dewar pump down and (ii) filter installation, a set of venting holes was placed in the OVCS-M. Without these, the ApC-SA would have been a completely enclosed volume. Three 1-mm-diameter holes were positioned 120 deg apart. These had a collective cross-sectional area of $2.36{\mathrm{mm}}^2$ against a total enclosed ApC air volume of $4.73\times {10}^5{\mathrm{mm}}^3$. For a $20\text{-}\mathrm{Torr}/\min$ dewar pump down, a 1-Torr differential across the filters only required a vent area of $0.005{\mathrm{mm}}^2$, which is equivalent to a factor of 174 area margin. As for filter insertion, a $20\text{-}\mathrm{mm}/\mathrm{s}$ installation velocity (much faster than the procedure) would require a $0.006{\text{-}\mathrm{mm}}^2$ vent area, which has an equivalent margin of 135 times.