The Mid-Infrared Instrument for the James Webb Space Telescope, II: Design and Build

The OBA consists of the optics module (the OM, shown in Fig. 1), the electrical control and data handling boxes associated with MIRI, which are maintained at 300 K in the separate ISIM Electronics Compartment (IEC; ISIM = Integrated Science Instrument Module), and the necessary interconnecting harnesses.

To combine the science functions into a single package and facilitate an easier assembly, integration and verification program, a modular optical design was chosen where lower level assemblies could be manufactured and their performance verified prior to being brought together in the complete instrument. This approach also enables parallelism and flexibility in the build, test and qualification flow but places stringent requirements on the "systems engineering" component of the project, with interfaces between subsystems needing to be defined, controlled, and monitored carefully at all stages. The design solution resulted in the OM being split among four main optical modules (subsystems), as shown in Figure 2a and listed as follows:

• 1.  
  Interface Optics and Calibration (IOC)
• 2.  
  Mid-infrared imager (MIRIM), interfacing to one Focal Plane Module (FPM) with its detector array. MIRIM encompasses the imager, low-resolution spectroscopy, and coronagraph modes of the instrument.
• 3.  
  Spectrometer Pre-Optics (SPO)
• 4.  
  Spectrometer Main Optics (SMO) interfaced to two FPMs with detector arrays; the SPO and SMO constitute the Medium Resolution Spectrometer (MRS).

These modules were integrated onto a single structure, the Deck. The completed OM is mounted to the JWST ISIM via a carbon fiber reinforced polymer (CFRP) hexapod mounting system (the black rods in Fig. 1). This hexapod thermally isolates the OM from the ISIM, which is passively cooled to about 40 K, while supporting it against the mechanical loads encountered during launch (Jessen et al. 2004).

Zoom In Zoom Out Reset image size

The MIRI optics take full advantage of state-of-the-art large-format mid-infrared detector arrays. Three focal plane modules (FPMs) with 1024 × 1024 pixel Si:As IBC detector arrays (Rieke et al. 2015b, hereafter Paper VII; Ressler et al. 2015, hereafter Paper VIII) interface to the OM, with one array dedicated to imaging, coronagraphy, and low-resolution spectroscopy, and the other two used in the medium resolution spectrometer. The FPMs attach to the outside of the optics modules, mating two flat surfaces (with locating fixtures) to provide robust and accurate alignment onto the outputs of the instrument optics.

MIRI is the only instrument that must be cooled below the temperature achieved by passive cooling of the ISIM to optimize the detector performance and reduce the thermal background below the detector dark current. The design was developed with thermal constraints as a key driver. The optics module is maintained at a temperature below 7 K by the cooler system. Because of its well understood structural and thermal behavior, aluminum alloy was used to make the supporting structure of the deck and the four optical subsystem modules. The reflective optical surfaces are also of aluminum to simplify the thermo-mechanical design and for the stability of alignment during cool-down. This approach had been proven for other flight and ground based instruments in the mid and far-IR (e.g., IRS and MIPS on Spitzer, SPIRE and PACS on Herschel, VISIR on ESO-VLT, Michelle on Gemini/UKIRT) but has been taken to higher levels of precision in MIRI. The instrument optical subsystems and the FPMs are built and aligned at room temperature, and remain aligned when cooled. The designs of both the imaging and spectrometer channels were implemented using the minimum number of low power cryogenic mechanisms (§ 6) to minimize the heat load to the cooler.

Inside MIRIM, the field-of-view (FOV) is partitioned into three functional areas; imager, coronagraph, and low-resolution spectrometer as indicated in Figure 3, enabling all science functions to be supported by a single detector array and a single wheel mechanism. The light is collimated and, at the pupil image formed by the collimator, the single wheel holds the filters for the imager and coronagraphs, a prism assembly for the low-resolution spectrometer, a blank for dark current measurements and a pupil imaging lens. This entrance focal plane is imaged onto the detector using a 3-mirror anastigmat camera with separate areas of the detector being dedicated to the imaging, coronagraphy, and spectroscopy functions. A full description of MIRIM can be found in Paper III.

The MRS (Paper VI) provides diffraction limited integral field spectroscopy over the whole wavelength range from 5 to 28.5 μm. It consists of two modules shown in Figure 2b—the SPO, which splits the incoming light both spatially, to form an entrance slit for the grating spectrometer, and spectrally into four channels, each of three sub-bands, that are dispersed and imaged onto the detectors in the SMO. The wavelength range is divided into four channels using dichroic mirrors in the SPO; the channels have separate dedicated integral field units and the spectra from each of the four channels occupy half of one of the two MRS detectors. Each channel is split, by a dichroic chain, into three sub-bands that are observed sequentially by rotation of just two mechanisms that carry both the wavelength sorting dichroics and the dispersion gratings in a very compact and efficient configuration.

The limited space allocated to MIRI, plus the need to keep the instrument overall as compact as possible to minimize the radiative heat load on the outer envelope, resulted in the use of relatively fast optical beams. These optics are designed to operate without vignetting and to meet image quality requirements in the presence of up to 4% pupil shear (i.e., the misalignment of the telescope exit pupil and the instrument entrance pupil in units of their diameter) and 2 mm of focus offset, tolerances that became requirements for the optical alignment strategy.

A tolerance analysis showed that MIRI would not need a focus mechanism, so long as tight alignment tolerances were maintained to place the focus position onto the detector and to position MIRI onto the telescope. The design solution to ensure the detector surface was placed correctly was to measure both sides of the flat interface plane and then machine a dedicated "shim surface." The detector surface position was measured both warm and at cryogenic temperatures to support this approach. To place the output of the optics correctly, an alignment budget was created that gave pupil shear and focus allocations to the subsystems (IOC, MIRIM, SPO, SMO); to each of the interfaces between these subsystems and the deck; and to the ISIM-MIRI hexapod mount. These budgets were set with the intention of achieving an overall focus within 1 mm and pupil shear of no more than 2% for MIRI. The imager and spectrometer were also required to be confocal, which was achieved via the mechanical alignment of the subsystems to the deck. The all aluminum structure means that all subsystem interfaces are direct between mounting pads on each surface, fixing one lateral and two angular degrees-of-freedom, with dowels defining the other two lateral and one angular degrees-of-freedom. The tolerances on these mechanical interfaces and Monte Carlo analysis showed that alignment within budget was possible without recourse to a measure-adjust-measure cycle.

The optical alignment of MIRI with respect to the telescope requires that the position and orientation of the entrance focal plane and the entrance pupil coincide with the focal surface and exit pupil of the telescope, respectively. Achieving this accurately is key to the scientific performance of the instrument. The positions of the telescope focal surface and exit pupil are well-defined with respect to the telescope optical elements and the mechanical interface between MIRI and the ISIM. However, these definitions are for the in-orbit environment of cryogenic temperature and zero gravity. The design of MIRI needed to take into account the offsets that will occur to the system from the warm as-built conditions found in a terrestrial lab.

On cooling, the distances between the CFRP leg to deck mounting points decrease by the integrated CTE (coefficient of thermal expansion) of aluminum, and the legs shorten by the integrated CTE of CFRP. Analysis showed that the leg/deck interface points would move toward the MIRI/ISIM interface by 0.49 mm on cooling and by only ∼20 μm when gravity is reduced to zero. At the same time, cooling of the optical bench causes the pick-off mirror (POM) to move toward the leg/deck interface.

The optical design model of MIRI was used to find by analysis the warm position and orientation of the POM that simultaneously placed the telescope focal plane at the MIRI entrance focal plane and the MIRI entrance pupil and telescope exit pupil at the same location when the system is cooled to its operating temperature. This warm position of the POM was used to inform the design of the IOC and to define the nominal warm positions of the MIRI entrance focal plane and entrance pupil. These were used for alignment verification during the room temperature construction of the optical subsystems and their integration into the OM.

Prior to delivery to NASA the overall alignment of MIRI was checked at room temperature using a NASA supplied reference system called the ASMIF which reproduced both the mechanical and optical interfaces within the ISIM and hence ultimately to the telescope. A series of measurements of pupil shear and focus were made, before and after vibration and cooling to operating temperature, using the references built into MIRIM. The data show that, with measurement uncertainties of 0.35%, the MIRI contribution to pupil shear is 1% and there is no discernable change with changing gravity vector. Focus measurements demonstrated that the MIRI focus is within 0.5 mm of the nominal position. The relative alignment between the entrance pupils of the MRS and MIRIM was measured at cryogenic operating temperature by scanning a point source across one quadrant of the MIRI pupils and correlating the resulting pupil maps to the as-built optical design models of the MRS and MIRIM. No measurable offset between the imager and spectrometer pupils was found. The warm pupil measurements were repeated after delivery using the same fixture to verify that there had been no unexpected issues arising from the transfer. Pupil shear and focus of MIRI relative to the nominal position within ISIM have subsequently been measured at NASA Goddard at cryogenic operating temperature, confirming the warm measurements. All of these results comfortably meet the targets set in the alignment budgets (§ 3.2) to have no significant impact on the performance of MIRI.

The excellent alignment of MIRI determined during test, and the end-end performance discussed in Papers III, IV, V, and VI demonstrate the success of the opto-mechanical approach to the MIRI optical design and alignment.

Careful attention has been paid to stray-light control. The fine steering mirror (FSM) within the telescope optics is surrounded by a cold stop that provides the defining cold baffle around the primary mirror. Cold pupil stops are provided within each of the instrument modules. They are slightly oversized to avoid vignetting at the FSM stop even in the presence of a small level of pupil shear, so they provide an additional level of stray-light rejection without affecting the optical path. Papers III and VI describe the stray-light suppression features within the MIRIM and the MRS.

Stable sources of illumination are needed on-board MIRI for calibration of the instrument's response close in time to an astronomical observation. The requirement is to achieve high signal-to-noise ratio in a short exposure time to derive high spatial frequency flat fields (pixel-pixel gain matrix) and for the source to be sufficiently stable that it can be used to monitor relative detector gain between observations of standard stars. The illumination should therefore be smooth on a spatial scale larger than one pixel and stable for timescales of tens of days.

Identical calibration sources are provided for both the imager and the spectrometer (Glasse et al. 2006). One source is mounted in the spectrometer preoptics and light is injected into the spectrometer optics via a hole in a folding flat mirror. For the imager the source is mounted in the IOC and light injected via a small relay mirror.

Figure 4 shows the source design. Pseudo blackbody radiation is produced by miniature tungsten filament lamps and is rendered uniform by a diffusing surface within an integrating sphere. There are two filaments in each sphere for redundancy. To avoid the steep fall-off at short wavelengths in the blackbody spectrum, the filament must emit with an effective temperature of at least 500 K. In practice, the operating temperature is restricted to less than 1000 K to maximize the filament lifetime.

Zoom In Zoom Out Reset image size

The long-term stability of the MIRI internal calibration sources was verified during the Flight Model (FM) campaign. The relative flux and repeatability of the source current were measured on eight occasions spread over a period of 46 days. Prior to each measurement the detector was annealed to ensure the results were not affected by its previous history (e.g., latents, image persistency, or other detector issues; Ressler et al. 2015). Signals through the flight filters called F560W, F1130W, F1800W, and F2100W (respectively at 5.6, 11.3, 18, and 21 μm, Bouchet et al. 2015) were measured, to provide a good representation of the wavelength range covered by the MIRI imager.

The relative flux from the source was defined as

where t is the measurement number, and φ is the measured average flux in DN/s within a single photometry region (three boxes of 100 × 100 pixels² and one of 200 × 200 pixels² using clean areas of the detector). The standard deviation of the relative flux value over time ranged from 0.411% to 0.270 % depending on the filter. These values were afterward corrected to account for variations in the MIRI Instrument Control Electronics (ICE) calibration source drive current, where the corrections were derived from the accurately sampled current values recorded in telemetry. This correction has been implemented in the Flight Software as an autonomous adjustment, to be applied to the source current once after every switch on.

The final, corrected, calibration source relative flux stability was found to range from 0.039% to 0.203% on a per filter basis, over a 46-day period. This compares very favorably with the accuracy of absolute flux calibration using standard stars which is estimated to be about 1%.

The MIRI OM is of isothermal construction with the Deck and the optical subsystems all constructed in aluminum alloy and thermally coupled together by bolted interfaces. The combination of this all-aluminum assembly with a simple and efficient CFRP hexapod provided a well understood structure, which could be specified and built to warm dimensions and whose offsets at cryogenic temperatures and under zero-gravity could be accurately predicted.

The main driving requirements for the structure as defined at the outset of the program are listed in Table 1. Sizing of the Deck and the Hexapod elements was driven by these requirements.

The deck is a ribbed and pocketed structure designed to support the various elements with the least expenditure of mass, maximum stiffness, maximum stability and lowest technological risk. It is machined in four parts from aluminum alloy 6061. The support for the spectrometer is made as a single part, carrying interfaces for the two Spectrometer Main Optics modules, the Spectrometer Pre Optics, and the Hexapod. It is bolted onto the "lower deck" structure that carries interfaces for the Cooler 6 K Heat Exchanger, the Imager and the Input Optics and Calibration assembly plus some ancillary items. Two struts bolted to this lower deck help to support the wider side extensions to the spectrometer part. The Deck is highly lightweighted and the pocketing is such that there is significant mass only in the regions where the subsystem interface pads locate. The Deck provides the stiffness between the Hexapod apices and was qualified early in the program using the Structural Thermal Model (STM).

The hexapod struts were manufactured from Carbon Fiber Reinforced Plastic (CFRP), which has a favorable combination of strength and low thermal conductivity at cryogenic temperatures. In practice, the hexapod design is stiffness limited, with the driving goal being to minimize the thermal cross section whilst maintaining a margin on the first frequency requirement. The second important design driver for the hexapod struts is buckling. High stiffness and high buckling resistance is achieved by having a high Young's modulus in the axial direction of the strut. This implies a lower cross section to achieve the first eigenfrequency, a key requirement in Table 1. The chosen design solution for the hexapod is six struts of length 405 mm, diameter 35.5 mm, and 1.2 mm wall thickness. This sizing avoids buckling under the design load and maintains a damage-tolerant wall thickness (Jessen et al. 2004).

The characteristics of the (T300) fiber used in the hexapod are well documented and their use has significant space heritage. This high strength fiber was preferred over a high modulus fiber due to its lower sensitivity to microcracking when cycled to cryogenic temperatures. The resin system is L20/SG. The hexapod strut end fittings and brackets are made from invar for thermo-elastic compatibility with the CFRP.

The Hexapod struts went through an extensive test campaign at strut level to qualify the manufacturing process and confirm strut performance. The performance of the Hexapod and Deck to provide the required alignment stability and withstand expected launch loads was verified by the Flight Model test campaign.

As described in §§ 3.2 and 3.3, the internal alignment of MIRI and the alignment into JWST are achieved by mechanical design and tolerances. To meet the required alignment performance the majority of the structural parts and the mirrors are constructed of two compositionally similar, heat treatable aluminum alloys, 6061 and 6082. Structural and optical components were thermally aged to ensure adequate dimensional stability through subsequent temperature cycling. The optical subsystem modules were mounted on the Deck using bolted and dowel-pinned interfaces. The deck was machined to a surface flatness of 20 μm throughout and the dowel hole location tolerances ranged from 25 μm for the IOC to Deck interface to 90 μm for the SMO to Deck interface. We note that no optical misalignments have been seen during testing of the MIRI Flight Model.

Repeatable mechanical location of MIRI with respect to test and flight interfaces was achieved by means of 6 quarter inch dowel pins, 2 at each hexapod "foot." For the flight interface to ISIM, these pins locate in a hole and a slot per foot on the ISIM side of the interface.

Conventional practices dictate that ground support equipment to handle the instrument should not occupy flight interfaces. However, transportation (which is the most severe environment seen) was carried out using appropriately protected flight interfaces. This was because of the need to control mass and the presence of an assembled, alignment critical and unconstrained friction locked hexapod. (For all other purposes, there are lifting brackets on the Deck conveniently close to the center of mass.) To avoid accidental damage to the hexapods, a system of "tie rods" was employed to support their feet when the instrument was not mounted at its mechanical interface. The tie rods incorporate length adjustment such that the foot positions could be microadjusted to accommodate manufacturing tolerance differences among the various test, flight, and transport equipment interfaces that the instrument would be mounted to. This system resulted in the overall pupil shear and focus measurements reported in § 3.3.

The sine loads dictated the design case for the semikinematic MIRI OM, as there are structure resonances in the range 50–100 Hz. Notching of these input loads during test was essential to protect the flexures in the end fittings from damage. The vibration test approach and results are described in detail in Sykes et al. (2012).

The random vibration levels specified at the MIRI instrument interface are relatively low, but nevertheless, attention must be paid to the critical subassemblies during test to ensure that subassembly specifications are not locally exceeded during instrument test. In particular, the mechanism vibration tests were heavily notched. The notched subassembly test inputs became constraints on the instrument level test as secondary notch limits.

As a basis for interface design and primary notching, limit loads were specified. The MIRI vibration test was a force limited test, meaning that acceleration input was controlled such that the measured force at the interface would not exceed a predetermined maximum. The maximum was set by direct reference to the design load for sine vibration, or by reference to the NASA semi-empirical method (Scharton 1997), in the case of random vibration. By this method, the launch loads were verified to be enveloped without the excessive over testing that would result from applying nominal vibration inputs through the main structural resonances.

MIRI is the coldest instrument on the observatory. The detectors themselves (Papers VII and VIII) must be held at a nominal temperature of 6.7 K with a temperature stability range of 20 mK over a 1000-s exposure. The deck and optics must be held at a temperature below 15.5 K with a stability within a 1 K band to avoid background radiation at long wavelengths that would impact the system sensitivity. It was decided to cool the whole deck and modules attached to it to ∼7 K, near the detector temperature, to remove temperature gradients and therefore possible sources of misalignment on cooling.

The total heat load from the OM to the cooler heat-exchanger stage during nominal operation must not exceed 46.5 mW. The nominal time-averaged dissipations internal to the OM are determined to be 10.46 mW by correlation of measurements with the MIRI thermal and operating models, leaving about 36 mW maximum for conductive and radiative loads.

To isolate the OM from heat generated in the ISIM that might undermine its thermal design, the OM is enclosed within a cooled shield at a temperature of around 23 K (Fig. 5). The OM is conductively isolated from the ISIM by the hexapod struts visible in Figure 1, which are attached to the ISIM conductive interface, having a temperature of about 40 K. The OM and shield are cooled actively by couplings to the 6 K stage and the 23 K Heat exchanger stage assembly of the MIRI-dedicated cryocooler, which is described below.

Zoom In Zoom Out Reset image size

The cooling to the detectors, OM, and thermal shield is provided by a ∼6 K/18 K hybrid mechanical cooler, provided by Northrop–Grumman Aerospace Systems in collaboration with JPL. The system is a further development from the NASA Advanced Cryocooler Technology Development Program (ACTDP), which achieved a breakthrough in cooler efficiency while achieving heat lifts of 30 mW from 6 K and 150 mW from 18 K (Ross 2004).

The MIRI Cooler System uses helium as the working fluid and consists of a three stage Pulse-Tube (PT) Precooler that reaches ∼18 K and a fourth ∼6 K stage, which is a Joule–Thomson (JT) cooler.³⁸ These two temperatures are made available to the instrument through heat exchangers. It is characteristic of JT devices that their heat lift decreases substantially with increasing precooling temperature. Therefore, a valve bypasses the JT expander to allow the initial cool-down of the instrument by the pulse tube stages. At ∼18 K, the bypass valve is closed and cooling by the JT expander continues to 6 K. This crossover is termed the "pinch-point" because it is the temperature where there is a minimum in overall heat lift capability.

The cooler system architecture is made particularly challenging since the cooler spans the length of the JWST observatory (Fig. 6). The Cooler Compressor Assembly (CCA) is in the JWST Spacecraft Bus and is at room temperature, while the Cold Head Assembly (CHA) is mounted on the ISIM structure near the MIRI OM. Both the compressors (PT and JT) are driven by the block-redundant Cooler Control Electronics Assembly (CCEA) which is located in JWST Spacecraft Bus.

Zoom In Zoom Out Reset image size

The CCA consists of the PT-Pre-cooler, the PT, and JT Compressors (see Fig. 7), plus a radiation shield for the various stages of the precooler. It also supports the stowage structure for the Refrigerant Line Deployable Assembly (RLDA) that runs through the observatory to carry the working fluid from the CCA to the CHA and back. Finally, it includes the structural element that mounts to the spacecraft bus for launch and provides thermal interface to the JWST heat rejection system. The refrigerant lines are supported by thermally isolating line supports. The CCEA provides the drive to the compressors and also implements the functions of thermal control, compressor vibration reduction, telemetry generation, and heater and valve control. The CHA contains the Joule–Thomson constriction and has the cryogenic valve that affords switching between the precooler mode and the JT-cooling mode during the MIRI optical system cool-down. Another valve bypasses more of the cold assembly to allow warming the cooler lines for decontamination.

Zoom In Zoom Out Reset image size

A flight-like CHA was delivered in the 2013 spring. It, along with a ground support equipment pre-cooler, successfully supported the cool-down and operations of the MIRI OS during the ISIM Cryovac1 Test (CV1-RR). The flight CCA and the CCEA are currently in development.

The MIRI cooler system operations will be verified and validated through a series of acceptance test programs followed by a MIRI end-to-end test where a full complement of flight-like cooler hardware will be tested along with thermally representative MIRI hardware (the STM, summarized in § 1). The end-to-end test will include a flight-like MIRI Thermal Shield, which is also cooled by the MIRI precooler.

The CFRP hexapod is one of the two main conductive paths between the MIRI OM and the ISIM. The conductance of each strut is about 0.02 mW/K at a mean temperature of 10 K increasing to about 0.06 mW/K at a mean temperature of 40 K (Shaughnessy et al. 2007). Table 2 summarizes the main conductive and radiative heatloads between the MIRI OM and the ISIM or Shield calculated from the correlated thermal model.

Electronic harnesses and a purge pipe also provide paths for conductive heat loads. Harness loads are managed by (1) minimizing of the number of wires required, (2) definition and control of the effective thermal length of the harness between the ISIM heat-sink and the OM deck, and (3) selection of low thermal conductivity materials for construction of the harness (manganin, phosphor bronze, and stainless steel). The purge pipe (which is primarily needed to maintain a clean and dry environment for the wheel lubricants to allow their operation during nonvacuum test activities) is constructed from stainless steel, sized to minimize the conducted load, while allowing the required volumetric flow rate to provide a suitable positive pressure in the instrument.

With the exception of the optical aperture and access for the cooler heat exchanger, the outer surface of the OM is covered with Single and Multi-Layer Insulation (SLI and MLI) blankets. The SLI is used around the six struts to minimize the conducted heat load from the warmer ISIM. The SLI also encloses harnesses and the purge-pipe that are attached to the struts.

There are two issues regarding MLI for MIRI. First, at low temperatures the thermal isolation achievable is small (e.g., Spradley et al. 1990). Nonetheless, it was decided to encase the OM in MLI both for the thermal gain and as a protective measure for its low-emissivity surfaces. Second, the skin depth of aluminum at a wavelength of 400 μm (by Wien's Displacement Law, corresponding to a temperature of ∼7 K) is about 0.1 μm. Therefore, the 1000 Å (0.1 μm) aluminum coating on typical blanket materials can be somewhat transmissive to thermal radiation originating from MIRI and the cold ISIM environment. Consequently, the MLI blankets were constructed using Kapton having a 5000 Å (0.5 μm) deposition of aluminum on one side. Eight layers are interleaved with crinkled double aluminized Mylar to inhibit conductive heat transfer (net spacers are not used as they are a source of particulate contamination).

The detectors are thermally isolated within the FPMs and cooled using a thermal strap of high purity copper from the mount for the detector arrays to a thermal connector on the FPM housings. From this connector, another strap runs approximately 0.5 m from the FPMs to the interface point near the 6 K heat exchanger. This section of strap is constructed from two 2 mm diameter pure (99.999%) aluminum wires that are clamped into specially designed end-fittings to maximize interface conductance. The FPM strap has been sized to allow heating the detectors to keep them warmer than the OM during cool-down (for contamination control), and to permit annealing cycles where the detector temperature is raised briefly to ∼15 K.

As a consequence of the MIRI being colder than the rest of JWST, particular attention was paid in the design to contamination control. There is natural protection from the thermal isolation until the MIRI cooler is activated. Once that occurs, the instrument cools below its surroundings and contaminants from them can collect. Models of the expected JWST outgassing indicate that in the worst case, without protection, a 1.5 μm layer of water ice along with various organic contaminants that are still volatile at 40 K could accumulate on exposed MIRI optical surfaces that are at 7 K. These considerations led to a design with a contamination control cover (CCC) just inside the optical train after the pick-off mirror, which is thermally isolated from the rest of the optics and can be decontaminated by heating. The first cold surface is most vulnerable; the long path of the IOC protects optics further down from transported contaminants when the CCC is open. The CCC design is discussed in § 6.

The Pick-Off Mirror (POM) is the coldest exposed optical surface within the ISIM. It is thermally isolated within the OM structure and fitted with redundant heaters to allow it to be warmed if necessary to drive off contamination (solid N₂, O₂, H₂O, CO₂) that may stick to the mirror. Before the POM heater is activated the CCC is closed to ensure contaminants do not freeze out onto the sensitive internal surfaces.

A cryogenic test facility was developed at the Science and Technology Research Council's Rutherford Appleton Laboratory to simulate the environment of the ISIM, including the 40 K radiative environment specified at that time (i.e., excluding the Thermal Shield described above) and the conductive interface with the OM and the 6 K heat-exchanger (Shaughnessy & Eccleston 2009). A 3-month cryogenic test was undertaken on the Flight Model OM to verify and calibrate its performance and to assess the thermal subsystem. Two phases of dedicated thermal tests provided steady-state and transient data for validating thermal models. A close correlation of heat load and temperatures was made to the steady-state data. Heat loads were correlated to within 0.5 mW of measurements and temperatures were correlated to well within 100 mK of measurements.

The nominal steady-state heat load predicted with the correlated model is 33.8 ± 6 mW. This shows a margin of 6.7 mW from the requirement of 46.5 mW, demonstrating that the cooler subsystem will be able to cool and maintain the OM at the required operating temperature.

The in-orbit cool-down prediction using this correlated model is presented in Figure 8. In the model, the ISIM boundary temperatures follow specified profiles, shown in the figure. The temperature of the MIRI shield shown is also interface data for the cool-down prediction and includes the response of the shield to the precooler. The OM is cooled passively until it reaches approximately 100 K, at which point the cooler is activated. A temperature-dependent cooler heat lift was provided for analysis. To demonstrate margin on the requirement, the effective heat lift in the model was reduced by 25%.

Zoom In Zoom Out Reset image size

The analysis predicts that it takes about 110 days for the OM to reach operational temperatures. The cooler is activated after about 80 days. The inflection in the OM temperature just past 100 days marks the transition through the pinch-point. For contamination control, the OM critical optical elements are required to remain above 165 K until the ISIM is 140 K or below. The analysis confirmed that the OM cool-down lags that of the ISIM and cools below 165 K about 15 days after the ISIM passes 140 K.

The MIRI Telescope Simulator (MTS) was the cryogenic optical system developed to generate the illumination sources for MIRI performance measurements. The MTS detailed design is described in Belenguer et al. (2008). In this section, we summarize its major functions and describe the computational model (MTSSim) that was produced to predict its photometric output. At the heart of the MTS was a laboratory standard blackbody whose temperature could be selected in the range 100–800 K. The collimated beam from this source passed through an adjustable iris diaphragm (to set the flux level) before reaching the MTS filter wheel. The wheel included a closed position to block the MTS hot source for background estimation; a clear position for broad band illumination; one long-wave pass filter and one short-wave pass filter for measurement of spectral leaks; and four solid state Fabry–Perot etalons which provided a comb of spectral lines for the wavelength characterization and calibration of the MRS. The output beam from the filter wheel was then presented to an integrating sphere whose spatially uniform output was matched to the input of a Cassegrain telescope (the Main Optical System [MOS]). By inserting a pinhole (one of two mounted on a three axis moveable stage) at this input, the MOS was designed to reproduce the point spread function delivered by the JWST. The point sources could be moved to any point within the MIRI field-of-view with an absolute accuracy equivalent to 1 imager pixel, and a relative accuracy of better than 0.1 pixel for small displacements. With the pinhole mechanism driven out of the beam, flood illumination across the full MIRI entrance focal plane was obtained. An infrared LED source was included at the exit pupil of the MOS which could be scanned across one quadrant of the MIRI pupil to measure the relative concentrations of the pupils associated with each of the MIRI optical subsystems.

The absolute flux calibration of the MTS was determined by modeling its optical throughput. This throughput estimate was embodied in a computer simulation, MTSSim, a program written in IDL to calculate the irradiance provided by the MTS at the MIRI input plane. MTSSim implemented a radiometric model of the MTS, with the hot source treated as a gray body with an emissivity of 95%. No diffraction effects were considered, and the system losses were limited to those caused by nonunity transmittance of the optical elements, as determined from subsystem measurements. Crucial to the accuracy of the end-to-end transmission budget was the error in estimating the transmission of the integrating sphere, since this could only be determined by geometrical modeling. As discussed in Glasse et al. (2015, hereafter Paper IX), this uncertainty in the estimated efficiency of the MTS was regarded as consistent with the 55% difference seen when using it as a flux standard for measuring the throughput of MIRI, as compared with measurements of MIRI's subsystems.

To provide a convenient reduction environment that was strictly configuration controlled and available to all of the international team, we developed the Data Handling and Analysis System (DHAS) (Morrison 2011).

The DHAS was based on a C++ analysis section, with a flexible IDL user interface. It first converts the raw integration ramps to slopes, subtracting the dark signals and correcting for nonlinearity. It also incorporates the best known algorithms to correct nonideal detector behavior, such as the reset anomaly (commonly seen in infrared arrays; the first samples after a reset are offset from the rest) (see Paper VIII for more discussion of non-ideal array behavior). It provides functions to display the resulting images and manipulate them, and also to organize the output of the MRS into a data cube.

The DHAS essentially implements the prototype for a MIRI data reduction pipeline. In addition to the continued use for instrument test data (e.g., at ISIM level), it is also being used to test and validate algorithms for the more sophisticated data reduction pipeline under development at STScI. The DHAS is also the means by which ongoing experiments on latent images, subarrays, annealing optimization, and other aspects of MIRI operations are evaluated.

We would like to thank the following national and international funding agencies for their support of the MIRI development: NASA; ESA; Belgian Science Policy Office; Centre Nationale D'Etudes Spatiales (CNES); Danish National Space Centre; Deutsches Zentrum fur Luft-und Raumfahrt (DLR); Enterprise Ireland; Ministerio De Economiá y Competividad; Netherlands Research School for Astronomy (NOVA); Netherlands Organisation for Scientific Research (NWO); Science and Technology Facilities Council; Swiss Space Office; Swedish National Space Board; and UK Space Agency.

We take this opportunity to thank the ESA JWST Project team and the NASA Goddard ISIM team for their capable technical support in the development of MIRI, its delivery, and successful integration.

We are grateful for the comments of the external referee which helped us to improve the clarity of high-level description of the instrument in this paper.