1.

Introduction

The far-infrared (FIR) foreground level in space is very low,¹ leaving significant room for future space-based missions to achieve more than three orders of magnitude improvement in sensitivity over former missions, such as AKARI,² Infrared Space Observatory,³ Spitzer,⁴ and Herschel.⁵ This foreground is dominated by zodiacal and galactic emission. Only at wavelengths of $>300\mu \mathrm{m}$ does the cosmic microwave background radiation begin to contribute significantly as a background source. To be sky-noise limited, space-based FIR telescopes need to be cooled to cryogenic temperatures of $\sim 4\mathrm{K}$. Furthermore, low noise detectors with a noise equivalent power (NEP) of $3\times {10}^{-19}\mathrm{W}/\surd \mathrm{Hz}$ for space-based continuum observations are needed to be photon noise limited.¹ The Origins Space Telescope (Origins) consists of a 5.9-m on-axis telescope cooled to 4.5 degrees Kelvin. This feeds an instrument suite consisting of a mid-infrared spectrometer, specialized for the characterization of transiting exoplanets around M-stars [Mid-Infrared Spectrometer (Camera) Transit Spectrometer⁶], the far-infrared survey spectrometer OSS,⁷ and the Far-IR imager/polarimeter (FIP), described here. FIP provides imaging and polarimetric measurement capabilities at 50 and $250\mu \mathrm{m}$. FIP’s large surveys take advantage of Origins's⁸ agility: The observatory can scan-map the sky very fast at 60″ per second, which is essential for efficient mapping, as the FIP $250\text{-}\mu \mathrm{m}$ channel reaches the extragalactic source confusion limit in a few milliseconds. The fast-mapping mode also enables rapid follow-up of transient or variable sources and enables efficient monitoring campaigns. FIP will enable wide area ($\ge 1000{\mathrm{deg}}^2$) photometric surveys, leading to large statistical multiwavelength studies of populations of astronomical objects, complementing ground-based and space-based surveys obtained, e.g., with the Vera Rubin Observatory (VRO) Legacy Survey of Space and Time (LSST)⁹ and the Nancy Grace Roman Space Telescope, a 2.5-m optical survey telescope (formerly WFIRST).¹⁰ A combination of deep and wide unbiased spectroscopic surveys obtained with the onboard FIR spectrometer OSS with wide-area FIP continuum surveys will enable Origins to measure star formation and AGN growth, the rise of metals, and feedback in galaxies over cosmic time and across the cosmic web.¹¹

2.

Enabling Technology

FIP’s major enabling technology is provided by the availability of superconducting detectors, which, on a single pixel basis, have been demonstrated to provide a noise performance that will enable background limited observations with Origins. There are currently two potential technologies considered for the detectors: transition edge sensors (TES) or microwave kinetic inductance devices (KIDs), see also Ref. 18. Examples of existing arrays using these technologies are shown in Fig. 2. TES are currently the baseline detector technology for FIP. FIP’s mapping speed will be enabled by arrays with a pixel count of 8000. The efficient readout of the detectors can be achieved by cryogenic microwave superconducting quantum-interference devices ($\mu$ Wave SQUID) multiplexers that have been developed at National Institute of Standards Boulder.^19,20 Those multiplexers allow for a multiplexing factor of 2000, which results in only 4 RF readout lines for the array.

Fig. 1

Origins/FIP can rapidly produce continuum maps over a very large area. This sky projection shows the footprint of the two widest surveys, LSST and WFIRST-HLS, and the time estimate to cover those areas with FIP. The proposed FIP surveys would include a wide, $\mathrm{10,000}{\mathrm{deg}}^2$ at $250\mu \mathrm{m}$ and a medium, $500{\mathrm{deg}}^2$ area at 50 and $250\mu \mathrm{m}$ that would provide SFRs and dust masses to complement LSST and WFIRST-HLS stellar masses and detect $>99\%$ of the galaxies that will be seen with the Nancy Roman Space telescope in the near infrared.

Fig. 2

The Origins Technology Development Plan recommends parallel investment in two detector technologies for FIP: TES bolometers and KIDS.

Sub-Kelvin coolers are the other main enabling technology for FIP. The FIP detectors, like the OSS detectors, are cooled to 0.05 K by a continuous adiabatic demagnetization refrigerator (CADR²¹). The path to maturation of these technologies is described in the Origins’ Technology Development Plan, a supplement to the full report.²²

3.

Operational Principles

The FIP functional block diagram (Fig. 3) shows the optical, cryogenic, and electronics components in the FIP design. The path of the incoming radiation to the detector is shown in pink. FIP has a pick-off-mirror, fore-optics, half-wave plates (HWP) on mechanisms, a barrel mechanism with two sets of optics, leading through a filter to the detector plane. The CADR assembly controls the required mK temperatures for the detectors. Launch locks, mechanical structure, and pre-amplifier electronics and harnessing complete the design.

Fig. 3

FIP has a simple design with one detector selectable by filter to 50 or $250\mu \mathrm{m}$. This functional block diagram of FIP shows how the power dissipation in the cold 4.5 K section is minimized by placing the readout electronics in the warm part of the observatory.

In pointed or survey imaging and polarimetry mode, FIP observes in one of the two bands, either 50 or $250\mu \mathrm{m}$. A single detector with a $109\times 73\mathrm{pixel}$ format ($\sim 8000\mathrm{pixels}$) is used for all measurements. The wavelength is selected by the optics/filter combination that can be switched with the barrel mechanism. In polarization mode, a HWP modulator and an analyzer (polarizing grid) is brought into the beam. In this setup, the relevant stokes parameters $I,Q$, and $U$ are being measured. These measurements fully characterize the linearly polarized emission from aligned dust grains, enabling the targeted polarimetry science goals.

FIP provides diffraction-limited images with an angular resolution of [$\mathrm{l}/23\times$ wavelength ($\mu \mathrm{m}$)] (arc sec) and covers an instantaneous FOV of ${3.6}^{\prime }\times 2.5^{\prime }$ at $50\mu \mathrm{m}$ and ${13.5}^{\prime }\times 9^{\prime }$ at $250\mu \mathrm{m}$. FIP has an optimized design to meet the maximum achievable optical efficiency in each band while requiring only a moderate FOV from the telescope. Either the field steering mirror (FSM) for pointed observations, or the telescope for survey mapping moves at all times during FIP observations to link the detector response of all pixels within the $1/f$ time interval (i.e., at a frequency higher than that of the $1/f$ noise knee, typically $<0.1\mathrm{Hz}$ for TES). For single pointing observations, only the FSM moves. Several observing patterns can be chosen, for example, a Lissajous Pattern for small maps, or a chessboard pattern for large surveys.²³ For survey mapping, the observatory sweeps the telescope pointing in raster patterns to map significant areas of the sky, up to hundreds or more square degrees. This operational principle has the significant benefit that the telescope does not need to settle at a certain position on the sky. For reconstruction of an image, only the knowledge of where the telescope has been pointed at any moment in time is needed. For Origins, the star trackers, combined with the gyroscopes deliver this knowledge.

4.

Optical Design

The FIP optical system design allows for a compact configuration (Figs. 4 and 5). Light is directed into the FIP instrument box by a pickoff mirror located near the telescope focal surface.

Fig. 4

The FIP mechanical structure and components. The labels point out individual components described in the text.

Fig. 5

The FIP optical system minimizes the number of optical components. This configuration shows the layout for a polarization measurement because the HWP is in the beam.

Inside the instrument box, a Dragone²⁴ collimator comprising a pair of free-form (polynomial surface) mirrors (CM1 and CM2), is used to collimate the light and correct aberrations (Fig. 6). While free-form mirrors would be more difficult to construct than traditional conic shaped ones at the same operating wavelength, we do not think they would be a technical challenge to manufacture given that the long wavelength affords much looser surface figure and roughness tolerances relative to optical surfaces.

Fig. 6

Detector readout scheme for FIP: FIP and RFSoC samples and processes signals simultaneously

Depending on the mode of operation, the collimated beam either traverses or bypasses a HWP before passing through an open window or a polarizer. For each of the two bands, the rays are imaged onto the detector array by a dedicated lens system, which provides the same beam-sampling factor for each, 50 and $250\mu \mathrm{m}$. Like the HWP and polarizer, the lens system switches in and out during operation depending on the waveband. The system uses an $f/2$ or $f/8$ lens for the short and long bands, respectively. The HWP, polarizer, and lenses are all switched in and out through mechanisms that are described in Sec. 8.

5.

Observing Modes

FIP observing modes (Table 1) include continuum and polarimetry observations. Either type of observation can be done in pointed or survey mapping modes. Pointed observations are performed when the observatory points at the source target, whereas the FSM performs a small Lissajous scan that moves the target sources on the central part of the detector array to provide the required pixel to pixel crosslinks. The second observing mode is survey mapping: the observatory moves in a raster form over the sky, whereas the FSM can be used to create more crosslinks between pixels in one $1/f$ time, if the raster scan cannot accommodate this criterion.

Table 1

Instrument parameters.

6.

Detector Subsystems

The FIP requirements could be met with multiple detector approaches,¹⁶ and all of them require some development. Our path to create FIP detectors is described in detail in the Origins Technology Development Plan.²² However, for our studied design, we have adopted superconducting TES bolometers as the baseline.

The TES are membrane-suspended and close-packed with the individual pixels directly coupling to the observed photons. The imaging array for FIP consists of $\sim 8000$ TES elements (pixels), configured in a rectangular shape ($109\times 73$). The TES detectors are densely multiplexed by $\mu$ Wave SQUIDs multiplexers:^18,19 About 2000 channels can be read out over the 4 GHz readout bandwidth of the $\mu$ Wave SQUID. Cryogenic high-electron-mobility transistor (HEMT) amplifiers are used in combination with room temperature readout electronics to facilitate readout and signal processing.

Each TES has its own microwave resonator (built into the $\mu$ Wave SQUID). Since roughly 2000 resonators fit within a 4 GHz processing bandwidth, and 8000 pixels need to be read out, a total of four radio frequency (RF) signals need to be connected to corresponding HEMT amplifiers and transmitted downstream for signal processing.

7.

Readout Electronics

The FIP RF signals will be sampled and processed using the commercially available Xilinx Radio Frequency System-on-Chip (RFSoC), which is not yet space certified. The detector feedback and other current signals needed for the detector readout can also be controlled by the RFSoC in combination with digital electronics on the same board (Fig. 6).

8.

Thermal and Mechanical Design

8.1.

Thermal Architecture

The FIP thermal system shares design features with the OSS instrument thermal design described in Ref. 7. Both instruments benefit from significant overlap and synergy in technology maturation and design methods. The FIP optics and mechanisms operate at 4.5 K with cooling provided by mechanical cryocoolers that are managed at the observatory level. The cooling power of the four cryocoolers can then be shared among the three instruments and the telescope. The dissipation at 4.5 K comes from three main sources within FIP: the HEMTs; the ADR, which provides continuous cooling at 50 mK; and the mechanisms. The design includes thermal straps made of pure annealed copper for all stages, sized for the load and distance required. These copper straps, along with the indium used for thermal connections, totals approximately 4 kg for FIP. The majority of this mass is for conducting the heat dissipated at 4.5 K to a central heat sink on the instrument structure that is, in turn, connected to a heat sink on the instrument mounting plate to which the 4.5 K cryocooler heat exchanger is also connected. This scheme leads to temperature gradients that are less than 5 mK across the 4.5 K portions of the instrument. A five-stage CADR is planned for detector cooling (Fig. 7). Heat to the detectors is intercepted at two locations: an outer guard that includes a cooled filter window and the array itself, which also includes the microwave SQUID multiplexer (Fig. 8).

Fig. 7

The FIP CADR provides continuous cooling at two stages: 50 mK (with $0.4\mu \mathrm{K}$ stability) and 0.7 K (with 0.1 mK stability), and deposits its heat at a 4.5 K heat sink.

Fig. 8

The FIP focal plane array is cooled to 0.05 K and suspended within a 0.7 K enclosure, providing thermal shielding from the 4.5 K stage. The Kevlar suspension, which is common to most instruments with low temperature detectors, provides for excellent stiffness while having a very low heat conductivity.

The five-stage CADR is currently being advanced from TRL 4 to TRL 6 through a grant from the Strategic Astrophysics Technology Program. This CADR will be capable of providing $6\mu \mathrm{W}$ of cooling continuously at 50 mK, which is more than an order of magnitude better than the unit flown on Hitomi.²⁵ Table 2 summarizes the calculated heat load and available lift with the cooler for each of the three FIP cooling stages. Heat loads at 50 mK are roughly evenly divided between harnesses, a rigid strut suspension, and internal dissipation within the detectors and SQUID multiplexer. The first stage HEMT amplifiers are located at the 4.5 K stage. The first stage HEMT heat dissipation shown in the table corresponds to that of commercially available amplifiers operating at up to 8 GHz, with low noise and 10 dB of gain (at this relatively low gain the heat dissipation is being kept minimal). We are currently performing studies to verify this amplifier noise is as advertised. Parasitic conduction in harnesses is book-kept at the observatory level but is much smaller than the amplifier dissipation.

Table 2

Calculations of the heat loads and available lift on the FIP cryogenic stages show margins of at least 200% over the calculated heat load maintained at each temperature stage.

T (K)	Item	Heat load	Total	Capabilitya	Margin (%)
	Harnesses	$0.16\mu \mathrm{W}$	—	—	—
0.05	Dissipation	$0.17\mu \mathrm{W}$	$0.56\mu \mathrm{W}$	$6\mu \mathrm{W}$	971
	Suspension	$0.23\mu \mathrm{W}$	—	—	—
	Harnesses	$7\mu \mathrm{W}$	—	—	—
0.7	Suspension	$76\mu \mathrm{W}$	$83\mu \mathrm{W}$	$292\mu \mathrm{W}$	252
	HEMTs	$3.1\mu \mathrm{W}$	—	—	—
4.5	Mechanisms	3 mW	10.1 mW	64 mW	533
	CADRs	4 mW	—	—	—

^aExcess capability enables a FIP upscope without need for resizing the CADR

9.

Mechanical Architecture

The FIP structure is shown in Fig. 4. Mechanical parts include exterior flexure mounts, a housing frame, thin enclosure, optical bench, and brackets that hold the three main mirrors. The design also includes mechanical components associated with the internal optics and detector assemblies. The material for all the mechanical components is beryllium for its light-weight and thermal conductivity properties.

9.2.

FIP Instrument Control

Figure 9 shows the electrical instrument control system block diagram. The main electronics box includes specialized control boards and a radiation hardened LEON3 CPU. System features include:

• • Mode management (standby, calibration, and execution of observing modes stare and mapping in total power and polarimetry modes)

• • Instrument support (command processing, data collection, and support for firmware updates)

• • Mechanism control

• • Power distribution, including detector biasing

• • Some degree of on-board autonomy (limit checking, some processing command sequences, and failure corrections)

Fig. 9

This FIP block diagram shows how the instrument is controlled.

10.

FIP Alignment, Integration, Test, and Calibration

The FIP instrument will require its own unique facilities for testing and calibration at the instrument level before it is delivered for integration with the observatory. With its outer space background limited FIR detectors, the major requirement for the optical tests will be to provide an extreme dark environment for the tests with test sources (black body source and FIR laser) being attenuated to provide power levels suitable for the detectors, which will allow for gain and optical performance measurements.

Aside from the normal vibration, acoustic, and electromagnetic interference (EMI) environmental testing, the FIP instrument will need a thermal vacuum test chamber that can be cooled to liquid helium temperatures and is large enough to enclose the complete cold section of the instrument. The chamber needs to be FIR light-tight enough so that stray light within the detectors’ sensitivity wavelengths is no greater than the infrared background that would be expected on-orbit. This test chamber should allow the introduction of FIR signals at the expected levels of the on-orbit target sources and the ability to verify the optical design of the instrument. The facility will need to allow connection between the cold instrument assembly in the test chamber and the warm readout electronics over flight-like cables.

The warm electronics for FIP will not require any particular test facilities other than the usual facilities used to conduct vibration, acoustic, EMI, and thermal vacuum tests. A spacecraft simulator will be required to test command and data transfers between FIP and the spacecraft electronics. It will be necessary to develop specialized test hardware, but although the scope of this development will be significant, it is well defined, with no new technology development required. FIP instrument I&T is expected to take 12 months, assuming all the external test support hardware and software is ready in time by the start of testing.

Although the vast majority of the FIP instrument calibration and verification/validation activities will take place before instrument delivery for observatory integration, verification of its performance within the integrated observatory will be necessary to ensure that it still performs as expected in end-to-end observatory testing before launch.

11.

FIP Heritage and Maturity

FIP is a mature instrument design except for the detectors. The design, in particular its optical layout, is similar to the SOFIA/HAWC+ instrument,^26,27 which has operated routinely for science observations since December 2016. HAWC+ on SOFIA uses three of GSFC’s $40\times 32\mathrm{pixel}$ TES-based BUG detector arrays, a potential design for FIP. With 8000 pixels, the required detector array is only roughly a factor of three larger than one of the two focal plane arrays in HAWC+. In light of the fact that the targeted array architectures are tileable, this factor of three in pixel count is achievable now. In terms of detector sensitivity on a single TES pixel level, the required sensitivity has been demonstrated in the Lab.²⁸ The dynamic range requirement might force using two TES per pixel, which could either be read out in parallel or series, the latter has been demonstrated at GSFC.²⁹

12.

FIP Predicted Performance

12.1.

Photometric Survey Time

In survey mapping or pointed modes, FIP quickly reaches the confusion limit in the long wavelength band (in 32 ms at $250\mu \mathrm{m}$), see also Table 1, but allows integrations of unprecedented depth and takes hours to reach the confusion limit in the $50\mu \mathrm{m}$ band. With a detector $\mathrm{NEP}=3\times {10}^{-19}\mathrm{W}/\surd \mathrm{Hz}$, FIP will be dominated by astronomical backgrounds.¹ The corresponding point source sensitivity is $0.9\mu \mathrm{Jy}$ ($5\sigma$, 1 h) at $50\mu \mathrm{m}$, $2.5\mu \mathrm{Jy}$ ($5\sigma$, 1 h) at $250\mu \mathrm{m}$. It is worth noting that the confusion limit at $250\mu \mathrm{m}$ does not allow an integration to $2.5\mu \mathrm{Jy}$. The proposed surveys with FIP at this wavelength take the confusion limit into account. Additional performance parameters are provided in Table 1, whereas optical parameters for the telescope can be found in Ref. 30. Figure 10 shows the unprecedented mapping efficiency FIP will provide. Combined with the fast mapping capability of Origins, FIP can survey thousands of square degrees efficiently with significant advantages in continuum mapping speed over OSS, as shown in the same figure.

Fig. 10

A comparison of photometric survey times for a number of space-based FIR instruments. It is worthy to note that Origins/FIP will be the fastest mapping far-IR survey instrument ever.