Advanced Net Flux Radiometer for the Ice Giants

Abstract

The design of an advanced Net Flux Radiometer (NFR), for inclusion as a payload on a future Ice Giants probe mission, is given. The Ice Giants NFR (IG-NFR) will measure the upward and downward radiation flux (hence net radiation flux), in seven spectral bands, spanning the range from solar to far infra-red wavelengths, each with a $5^{\circ}$ Field-Of-View (FOV) and in five sequential view angles ($\pm 80^{\circ}$, $\pm 45^{\circ}$, and $0^{\circ}$) as a function of altitude. IG-NFR measurements within either Uranus or Neptune’s atmospheres, using dedicated spectral filter bands will help derive radiative heating and cooling profiles, and will significantly contribute to our understanding of the planet’s atmospheric heat balance and structure, tropospheric 3-D flow, and compositions and opacities of the cloud layers. The IG-NFR uses an array of non-imaging Winston cones integrated to a matched thermopile detector Focal Plane Assembly (FPA), with individual bandpass filters, housed in a diamond windowed vacuum micro-vessel. The FPA thermopile detector signals are read out in parallel mode, amplified and processed by a multi-channel digitizer application specific integrated circuit (MCD ASIC) under field programmable gate array (FPGA) control. The vacuum micro-vessel rotates providing chopping between FOV’s of upward and downward radiation fluxes. This unique design allows for small net flux measurements in the presence of large ambient fluxes and rapidly changing ambient temperatures during the probe descent to ${\geq} 10$ bar pressure.

1 Introduction

The ice giants, Uranus and Neptune, are of significant scientific importance and the Ice Giants Pre-Decadal Survey Mission Study Report (IGPDS, 2017) has identified the high priority of an exploration mission with an orbiter and a probe to one of the ice giants with preferential launch dates in the 2029–2034 timeframe. Such a mission will advance our understanding of the Solar System, exoplanetary systems, planetary formation and evolution. Twelve science objectives were identified by the IGPDS report and the Vision and Voyages Planetary Science Decadal Survey for ice giant exploration; hereafter referred to as V&V (NRC 2011). Two main questions of equal scientific importance that relate to the atmospheric (thermal) structure of these icy worlds arise: (i) what is the planet’s atmospheric heat balance? and (ii) what determines the planet’s tropospheric 3-D flow? The IG-NFR concept, Fig. 1, onboard a probe descending deep into the atmosphere will contribute greatly to answering these questions by measuring the upward and downward radiation flux, in seven spectral channels, each with a $5^{\circ}$ Field-Of-View (FOV) and in five sequential view (sky) angles as a function of altitude/pressure. For both Uranus and Neptune, $5^{\circ}$ FOV observations at one scale height above the 1 bar level captures the radiation field from 1 bar to ∼7 mb, a region of solar heating and predicted location of the CH₄ ice cloud.

Fig. 1

Top: IG-NFR concept responds to the science objectives of a future Ice Giants probe mission to either Uranus or Neptune. The IG-NFR will be capable of measuring energy flux in seven spectral bands (channels) covering the 0.2–300 μm spectral range, each with a $5^{\circ}$ FOV projected into the sky. Bottom: Graphic showing accommodation of the IG-NFR on a 1-m diameter probe deck

The IG-NFR is designed to exceed baseline instrument requirements discussed in the IGPDS study Science Traceability Matrix (STM). The IGPDS study baselines the Galileo probe NFR (Sromovsky et al. 1992, 1998). The science objectives and the features for both instruments are given in Table 1. The IG-NFR will have more capability with extra viewing angles and narrower FOV’s to look for hemispheric dichotomies. The IG-NFR achieves all the science goals, with the necessary signal-to-noise ratio (SNR) in filter bands spanning the spectral range 0.2 to 300 μm. At the lowest anticipated temperature, for example 50 K at Neptune at a pressure of 0.1 bar, the long-wavelength cut-off at 300 μm, in the bandpass filter set, amply captures the tail of the Planck function which peaks at ∼58 μm. Additionally, the IG-NFR will serve to provide an in situ reference for the remote sensing from visible-wavelength cameras and visible-near infrared mapping spectrometers’ radiative transfer analysis of cloud structure, removing significant ambiguity that necessarily remains in the remote sensing retrievals.

Table 1 Science goals and comparison of IGPDS baseline NFR [1] with the IG-NFR

The IG-NFR FPA is comprised of integrated detector and Winston cone sub-assemblies, housed inside a vacuum micro-vessel, Fig. 1. The vacuum micro-vessel helps mitigate the effects of rapid changes in temperature of the FPA that the instrument will experience during a probe descent into either a Uranus or a Neptune atmosphere.

The nominal measurement regime for the IG-NFR extends from about 0.1 bar to at least 10 bars. In the following we first briefly describe the scientific objectives, then describe the IG-NFR instrument design and how it functions with particular emphasis on the FPA housed in the vacuum micro-vessel, give an account of the front end-electronic readout, give a radiometric calibration scheme and finally give the expected radiometric performance.

2 Scientific Objectives

Ice giant meteorology regimes depend on internal heat flux levels. Both incident solar insolation and thermal energy from the planetary interior, can have altitude and location dependent variations. Such radiative energy differences cause atmospheric heating and cooling, and result in buoyancy differences that are the primary driving force for Uranus and Neptune’s atmospheric motions (Allison et al. 1991; Bishop et al. 1995). The three-dimensional, planetary-scale circulation pattern, as well as smaller-scale storms and convection, are the primary mechanisms for energy and mass transport in the ice giant atmospheres, and are important for understanding planetary structure, circulation, and evolution (Lissauer 2005; Dodson-Robinson and Bodenheimer 2010; Turrini et al. 2014). These processes couple different vertical regions of the atmosphere, and must be understood to infer properties of the deeper atmosphere and cloud decks (Fig. 2).

Fig. 2

Cartoon of a Probe descent through either Uranus or Neptune’s poorly understood atmospheres. The NFR will reveal thermal structure, opacity sources, and help provide a global radiative balance. As the probe descends, seven boresighted spectral channels measure energy flux, sequentially and repetitively (clockwise and anti-clockwise) at five viewing angles. The view angles in the cartoon (1-to-5) are shown at 1.5 s time intervals (1 s integration and 0.5 s slew to next position). The sequence repeats anti-clockwise (5-to-1). Each spectral channel samples different processes. Short Wave radiation—SW; Long Wave radiation—LW

The atmospheres of Uranus and Neptune are expected to resemble Jupiter and Saturn’s in the broad sense, with convective tropospheres giving way to radiatively regulated stratospheres at pressure levels of around 0.1 bar (∼50–100 km altitude above the cloud tops). However, there are likely to be many differences in detail due to variations in composition and distance from the Sun, Table 2. The solar constant drops from ${\sim }51~\mbox{W}/\mbox{m}^{2}$ at Jupiter to $15~\mbox{W}/\mbox{m}^{2}$ at Saturn, $3.7~\mbox{W}/\mbox{m}^{2}$ at Uranus, and $1.5~\mbox{W}/\mbox{m}^{2}$ at Neptune, leading to progressively colder atmospheric temperatures. While on Jupiter and Saturn ammonia, ammonium hydrosulfide (NH₄SH) and water are predicted to be key cloud-forming species in the troposphere, the colder atmospheres of Uranus and Neptune are expected to form clouds from methane ice and hydrogen sulfide ice at observable pressure levels.

Table 2 Mean temperature for the gas and ice giants at 1 bar, showing the thermal emission peak wavelengths

Many authors have given thermochemical models and predictions of the main cloud layers as a function of compound abundances (e.g., Baines and Hammel 1994; Baines et al. 1995; Atreya and Wong 2004, 2005; Irwin 2009; Mousis et al. 2018).

Observations at microwave wavelengths with the Very Large Array found a missing component of continuum absorption (de Pater and Massie 1985; de Pater et al. 1989, 1991) that was concluded to most likely arise from the pressure-broadened wings of H₂S lines and recent observations at near-infrared wavelengths (Irwin et al. 2017, 2018, 2019) have directly detected absorption lines of H₂S above the main observable cloud deck on Uranus and Neptune. Hence, it now appears that the abundance of H₂S exceeds that of NH₃ in the observable atmosphere of Uranus and Neptune, which react/condense together to form a cloud of NH₄SH at a pressure of ∼40 bars, leaving the left-over H₂S to condense alone at 3–5 bar. Figure 3(a) shows the distribution of NH₃ and H₂S absorption lines as a function of wavelength. Figures 3(b) and 3(c) show the results of a first order, cloud structure of Neptune and Uranus calculated using an Equilibrium Cloud Condensation Model (ECCM) that employs basic principles of thermodynamics (Irwin 2009). The cloud densities represent upper limits, as cloud microphysical processes (precipitation) would almost certainly reduce the density by factors of $10^{2}$/$10^{3}$ or more.

Fig. 3

(a) Distribution of NH₃ (black) and H₂S (red) absorption lines as a function of wavelength (1–300 μm), courtesy of Irwin, private communication. (b) ECCM model of Uranus’ atmosphere, calculated cloud layers: H₂O cloud (water, then ice) at $p > 1000$ bar, NH₄SH at ∼40 bar, H₂S ice at ∼5 bar, and CH₄ ice at ∼1.2 bar. Assumed composition: O/H = $100\times$ the solar value, N/H = the solar value, S/H = $11\times$ the solar value, C/H = $40\times$ the solar value. (c) ECCM model of Neptune’s atmosphere, calculated cloud layers: H₂O cloud (water, then ice) at $p > 1000$ bar, NH₄SH at ∼50 bar, H₂S ice at ∼8 bar, and CH₄ ice at ∼2 bar. Assumed composition: O/H = $10\times$ the solar value, N/H = the solar value, S/H = $11\times$ the solar value, C/H = $50\times$ the solar value. An IG-NFR would provide crucial ground-truth measurements to test current models

It is not known in detail how the energy inputs to the atmosphere-solar insolation from above and the remnant heat-of-formation from below interact to create the planetary-scale patterns seen on these ice giants (IGPDS 2017). An understanding of circulation in these planetary systems requires knowledge of the vertical profile of radiative heating and cooling and also its horizontal distribution. In situ measurements with the IG-NFR, using judiciously chosen filter channels, will contribute to our understanding of the balance between the upward and downward radiation flux’s to evaluate effects due to primary opacity sources, and to establish the extent of solar heating e.g., below 1 bar pressure for the ice giants.

Given the ice giants’ importance to our understanding of the formation of our solar system and the processes that have shaped it, as well as how common they are in exoplanetary systems, the V&V made these poorly studied worlds a priority. An ice giant mission was identified as among the top 3 Flagship-class missions. The Science Definition Team (SDT) of the IGPDS reaffirmed the importance of a Flagship mission to an ice giant and the science goals and objectives that were defined by the SDT in their STM were consistent with those in V&V.

Twelve science objectives were identified in the STM of which two are highly relevant to the IG-NFR. The two science objectives to (i) determine the planet’s atmospheric heat balance and (ii) determine the planet’s tropospheric 3-D flow are intricately linked. It is not known how the energy inputs to the atmosphere are distributed and how these inputs interact to create the planetary-scale patterns. Important questions arise: What are the altitudes/pressures and compositions of the cloud layers? How do the cloud layers interact with solar visible and planetary thermal radiation to influence the atmospheric energy balance? How does the energy balance contribute to atmospheric dynamics? Is the intrinsic flux spatially inhomogeneous? Is there a hemispheric dichotomy? Can the intrinsic flux of Uranus or Neptune be detected? What is the nature of convection and circulation on the planets and how does it couple to the temperature field? To answer these questions a IG-NFR will provide unique in situ measurements, as the probe descends deep into the atmosphere, of solar insolation and thermal emission in five viewing angles with narrow fields of view and will provide complementary data to science instruments measuring visible albedo and thermal emission at a range of solar phase angles and incidence angles from an orbiter. Currently, three baseline spectral channels have been identified, listed in Table 3, that include a broadband solar channel (0.2–3.5 μm) for total deposition of solar radiation and hot spot detection, a broadband thermal channel (2.5–300 μm) for deposition/loss of thermal radiation and a broadband channel (0.6–3.5 μm) for solar deposition in the methane absorption region and cloud particles. The remaining four science spectral channels will be optimized using radiative transfer modeling studies, using NEMISIS (Irwin et al. 2008) and the Planetary Spectrum Generator (Villanueva et al. 2017) codes, that are currently underway. Results of this modeling will identify optimal spectral channels for sensitivity and discrimination between cloud and H₂S, NH₃, H₂O and CH₄ gaseous abundances, and will be reported in a future publication.

Table 3 Three of the seven IG-NFR science spectral filter channels that have been identified

2.1 In Situ Net Energy Flux Measurement

The net energy flux, the difference between upward and downward radiative energy crossing a horizontal surface per unit area is directly related to the radiative heating or cooling of the local atmosphere: the radiative power per unit area absorbed by an infinitesimal thin atmospheric layer is equal to the difference in net fluxes at the boundaries of the layer. At any point in the atmosphere, radiative power absorbed per unit volume is thus given by the vertical derivative of net flux ($dF/dz$) in the plane-parallel approximation where the flux is horizontally uniform; the corresponding heating rate is then $(dF/dz)/(\rho C _{p})$, where $\rho$ is the local atmospheric density and $C _{p}$ is the local atmospheric specific heat at constant pressure.

For the ice giants, the thermal structure and the nominal IG-NFR measurement regime extends from ∼0.1 bar (near the tropopause which coincides with the temperature minimum) to 10 bar (ideally ≥50 bar). The 1986 Voyager encounter at Uranus (in the vicinity of the summer solstice) revealed very little details on atmospheric phenomena. However, distinctive bright high-altitude CH₄ ice clouds have been observed more and more frequently moving toward the northern spring equinox in 2008 and in subsequent years (de Pater et al. 2015). During the Voyager encounter Neptune displayed an ice cloud layer at ∼1 bar. More recent observations show these CH₄ clouds to be transitory and patchy in the 400 mbar to 1 bar level believed to be associated with convective upwelling (Karkoschka and Tomasko 2011; Irwin et al. 2019). Atreya and Wong (2004) predicted the base of the water-ice cloud for solar composition O/H to be at ∼40 bar level, and the NH₃H₂O solution clouds ∼80 bar. However, the ice giants are likely to be significantly enriched in heavy elements compared to the Sun. This pushes the water cloud down to much higher pressures. Using the heavy-element enrichment suggested by a number of authors it is expected that NH₄SH clouds are at ∼30 bar, HS clouds at ∼4–5 bar and CH₄ clouds at <1.5 bar pressure (Irwin 2009). So far, for Uranus, only an upper limit is known for its heat flow based on Voyager 2 (Pearl et al. 1990). In situ probe measurements will help to define sources and sinks of planetary radiation, regions of solar energy deposition, and provide constraints on atmospheric composition and cloud layers.

2.2 State-of-the-Art

Jupiter’s Galileo probe made its first measurements on December 7th, 1995 (Sromovsky et al. 1998). Cassini’s Huygens probe included the Descent Imager/Spectral Radiometer (DISR) (Tomasko et al. 2002) and landed on Titan on January 14th, 2005. DISR only looked at upward and downward direct and diffuse solar flux between 0.35 and 1.7 μm spectral range during its descent. To our knowledge, since the days of the Galileo probe no net flux radiometer, that included solar and thermal channels, has been on-board an atmospheric entry probe. Table 4 compares the Galileo probe NFR with the IG-NFR.

Table 4 Galileo NFR parameters compared to the IG-NFR

3 Instrument Description

Since the days of the Galileo probe NFR, there have been substantial advancements in synthetic diamond windows, filters (interference and mesh), uncooled detectors, insulating materials, and radiation hardened FPGA’s and ASIC electronic readout technologies, which allow for a more capable and robust NFR. In order to meet the science objectives, the NFR was designed to have (i) seven spectral channels; (ii) a clear unobstructed $5^{\circ}$ FOV for each spectral channel; (iv) thermopile detectors that can measure a change of flux of at least 0.5 W/m² per decade of pressure; (iii) five distinct view angles ($\pm 80^{\circ}$, $\pm 45^{\circ}$, and $0^{\circ}$); (iv) a detector response that can be predicted with changing temperature environment; (v) an FPGA controlled MCD ASIC technology for detector readout; (vi) signal processing that can integrate radiance for 1 s or longer, and (vii) sampling that can view each angle including calibration targets every 10 s (assuming $100~\mbox{m}/\mbox{s}$ descent rate).

The NFR concept (Aslam et al. 2015, 2018, 2019; Mousis et al. 2018) is shown in Fig. 4. Figure 5 shows the Winston cone non-imaging optics, detectors, filters and fold mirror all housed in a vacuum micro-vessel so as to mitigate rapid excursions of temperature during the probe descent. A close hexagonal-packing array of Winston cones gives seven channels, with each Winston cone designed to give a $5^{\circ}$ clear FOV. The Winston cone design can easily be changed to accommodate a larger FOV, resulting in a larger detector signal and a more compact vacuum micro-vessel design, since the Winston cone lengths will be smaller. Uncooled single pixel thermopile detectors are chosen for good detection sensitivity of radiation flux. A stepper motor with the aid of a gearbox rotates the vacuum micro-vessel, to each of the five view angles, so that the diamond window on the micro-vessel has an unobstructed view through apertures in the outer housing into the atmosphere. The synthetic diamond window has a strong absorption feature centered near 5 μm, this will be accounted for in calibration schemes by measuring the NFR relative spectral response as a function of wavelength for each spectral channel.

Fig. 4

(a) Side view section of the IG-NFR concept. (b) Quarter view section of the vacuum micro-vessel with a diamond window, which houses the fold mirror, filters, Winston cones, and detector FPA, to minimize rapid temperature excursions during probe descent

Fig. 5

Top: Internals of the vacuum micro-vessel. Bottom: Fold mirror sub-assembly that includes the filter bank

The outer housing (not shown in Fig. 5 for clarity) accommodates hot and cold targets and a light emitting diode for radiometric calibration for each sequence of measurements (5-views). The technical specifications of the NFR are given in Table 5.

Table 5 IG-NFR technical specifications designed to meet the science objectives for net energy flux measurements in the atmospheres of Uranus or Neptune

The system block diagram is shown in Fig. 6 and highlights the major subsystems in the FPA, comprising the thermopile detector sub-assembly and the Winston cone sub-assembly. The FPA is integrated to the fold mirror sub-assembly and aligned and assembled into a vacuum micro-vessel as shown in Fig. 5. The FPA detector and thermistor signals are read out via a connector to the Front End Electronics (FEE) i.e., a MCD ASIC that communicates to a FPGA.

Fig. 6

IG-NFR system block diagram. WC—Winston cone; a-analog

3.1 Focal Plane Assembly

The FPA consists of a Winston cone sub-assembly integrated to a detector sub-assembly. The Winston cone sub-assembly is shown in Fig. 5 with a half-section view and specifications shown in Fig. 7. The center-to-center distance of entrance diameters of two cones is 10 mm. The detector sub-assembly is also shown in Fig. 5. The detector fanout board is shown in Fig. 8, which accommodates seven thermopile detectors packaged in TO-39 housings manufactured by the Institute of Photonics Technology (IPHT) in Jenna, Germany and available commercially from Micro-Hybrid Electronic GmbH with a silver black absorber. For flight, custom thermopile detectors with gold black coatings will be manufactured (Aslam et al. 2016). One dark reference detector is mounted on the backside of the fanout board. Thermistors mounted in the TO-39 housings and on the fanout board allow for precise temperature measurement of the detectors and the environment ensuring thermal compensation corrections can be applied to the measurements. Terminal pins on the fanout board are connected to the MCD ASIC readout circuit by a shielded cable to reduce interference noise (not shown). Thermopile detector parameters are listed in Table 6. For the IG-NFR, eight thermopile pixels (7 science + 1 blind) and 9 thermistor signal outputs will be captured simultaneously by the MCD ASIC and processed in parallel under FPGA control. The DAC integrated on the MCD is used to generate a pixel reference to position the output signals of the thermopiles.

Fig. 7

Half sectional view of Winston cone and thermopile detector geometry. The Winston cone is mounted in an end plate. (1) Winston cone; (2) output end plate; (3) thermopile chip; (4) cap with gold internal coating to act as a reflective cavity; (5) header. Note: $2.5^{\circ}$ defines half FOV in rotational symmetry (light rays perpendicular to the inclination are the only rays that are accepted into the Winston Cone)

Fig. 8

Thermopile detector fanout board with thermistor temperature readout. Each detector package also has a thermistor temperature sensor. Board dimension: ∼ 4 cm diameter

Table 6 Thermopile detector specifications. The custom thermopiles with gold black absorbers, manufactured by Institute of Photonics Technology (IPHT), have essentially a flat spectral response from 0.2 to 300 μm. The detectors survive ${>}38~\mbox{W}/\mbox{m}^{2}$ (cf.$3.7~\mbox{W}/\mbox{m}^{2}$ solar constant at Uranus) so inadvertent FOV pointing at the sun during the probe decent will pose no risk

3.2 MCD ASIC Readout

The MCD ASIC (5 mm × 5 mm chip) is a fully custom radiation-hardened-by-design (RHBD) integrated circuit that was designed using Towerjazz Semiconductor’s 180 nm CMOS CA18HD process node for near-DC signal parallel readout, Fig. 9 (Quilligan et al. 2014, 2015; Aslam et al. 2012). The MCD ASIC has twenty readout channels each comprising a low noise variable gain amplifier driving a dedicated high-resolution sigma-delta A/D converter ($\varSigma \Delta$ ADC). The channels are designed to interface directly to thermopile outputs and amplify/digitize the signals with variable gain/resolution. Effective 20-bit digitization of the signals can be attained by using the analog front-end gain and oversampling ratio. The original design targeted a proposed Europa Clipper thermal instrument (Aslam et al. 2016) readout for operation in the harsh Jovian orbital environment, i.e., immunity to $> 3$ Mrad (Si) total ionizing dose (TID). The MCD ASIC also has a multiplexed buffered $\varSigma \Delta$ ADC for general purpose (housekeeping) tasks, 10 DACs, a serial port and a die temperature sensor. The Towerjazz process node CA18HD was selected because it has inherent radiation tolerance and immunity to latch-up. With the addition of RHBD techniques, the MCD was hardened to levels that far exceed the original design requirements.

Fig. 9

(a) MCD ASIC chip micrograph and simplified system block diagram; (b) 20-channel MCD ASIC board. This radiation hard MCD ASIC enables the IG-NFR thermopile readout electronics to be compact and to have small volume, mass and power

3.2.1 MCD ASIC Resolution

The $\varSigma \Delta$ ADCs utilize oversampling to quantize their inputs. $\varSigma \Delta$ conversion is used because the thermopile signal bandwidths are very low and the architecture is inherently linear. The effective resolution of the MCD digitizers is increased by a scheme which is a function of (1) the analog front end (AFE) gain (variable) and (2) the oversampling ratio (OSR, also variable). 20-bits of resolution is achieved by choosing an appropriate AFE gain and OSR. The main barrier to achieving 20-bit resolution is noise, particularly $1/f$, Johnson and noise pickup from the environment. The MCD channels employ frequency translation, correlated double sampling and auto-zero to reduce $1/f$ noise before being digitized by the $\varSigma \Delta$ ADCs. Active filtering within each channel reduces the Johnson noise. The signal paths from thermopile to ADC are pseudo-differential which reduces common-mode noise (Quilligan et al. 2017; Quilligan et al. 2018a; Quilligan et al. 2018b). Each ADC outputs its data as a single bit stream for off-chip decimation and filtering, allowing the user to change the OSR and thus the ADC resolution.

3.2.2 MCD ASIC Noise

The thermopile detector outputs are directly connected to on-chip instrumentation amplifiers so as to minimize signal loss due to the high output impedances of the thermopiles. Since the thermopiles output very low frequency signals, the $1/f$ noise of the readout electronics can be significant and thus must be mitigated along with the thermal noise contribution. To reduce readout $1/f$ noise, the output signal of each detector is modulated at 64 kHz, amplified and then demodulated back to the baseband frequency region. Since signal amplification happens at high frequency, the input referred $1/f$ noise of the detected signal is dramatically reduced. Figure 10 shows the input analog path of one typical channel. Moreover, the variable gain amplifier is reconfigurable to avoid saturation.

Fig. 10

Analog input signal path of one typical MCD ASIC channel

The $\Sigma \Delta$ ADCs on the MCD ASIC utilize oversampling and noise shaping to achieve high resolution. As a result, most of the quantization noise is shifted to the high frequency region which can easily be filtered out with a digital filter (Fig. 11).

Fig. 11

Measured noise spectral density of 1-bit stream from delta-sigma ADCs (average in black and $\pm 1\sigma$ deviation in red from 20 measurements). Decimation filter on FPGA will remove high frequencies noise

3.2.3 MCD ASIC Radiation Qualification

All of the circuitry in the MCD uses RHBD techniques including (a) enclosed layout transistors (ELTs) and (b) double guard ring isolation. The ELTs harden the circuits against total ionizing dose (TID) while the isolation rings help prevent single event latch-up (SEL). The first generation MCD ASIC was tested for TID at GSFC’s radiation effects facility up to 53 Mrad (Si) TID with no increase in the supply current and very small parametric shifts in gain and offset (see Figs. 12 and 13). The ASIC was also tested with heavy ions up to $174~\mbox{MeV-cm}^{2}/\mbox{mg}$ at Texas A&M University’s without latching up. Figure 14 is a plot of MCD supply current and transient effects over time with the heavy ion beam both off and on. Supply current for the entire chip increases slightly due to charge deposition from the beam but this is bled off once the beam is deactivated. Figure 14 also shows a plot of single event transients (SETs) at a high LET value. In this plot the MCD is digitizing a mid-scale signal from one channel at a channel gain of 25. When the beam turns on, the charge transients cause the channel auto-zero circuit to vary its correction voltage to the amplifier. Due to the pipeline nature of the channels, any single event upsets would get flushed out or corrected by the periodic refresh of the channel settings.

Fig. 12

MCD ASIC supply current vs. ionizing dose (dots indicate measurement points)

Fig. 13

MCD ASIC transfer function for non-irradiated (green) vs. post 50 Mrad (red) parts

Fig. 14

MCD ASIC supply current/SETs during heavy ion testing at TAMU at $174.8~\mbox{MeV-cm}^{2}/\mbox{mg}$

3.2.4 MCD ASIC—FPGA Control

An FPGA controls the MCD ASIC through a SPI serial protocol. During run time, the MCD ASIC channels output up to twenty 1-bit data streams from the $\Sigma \Delta$ ADCs to the FPGA where they get decimated back to useful multi-bit words. In this case, multiple parallel Sinc3 filters with a decimation ratio of 1024 were implemented on the FPGA. Figure 15 shows the block diagram of the data acquisition module. Figure 16(a) shows the MCD ASIC under FPGA control configuration to establish readout noise and Fig. 16(b) shows a typical noise spectrum when the output data rate is at 239 Hz with a modulating frequency of 64 kHz, the input was shorted to the pixel reference to avoid interference from any external noise source. A typical readout channel exhibits an average spot noise of ∼12 nV/$\sqrt{\vphantom{a}}$Hz. For a signal frequency that is lower than the output data rate, a moving average algorithm can be used to further reduce thermal noise and random noise from the signal. To verify the operation of the parallel readout circuit an infra-red broadband source was pulsed at 0.5 Hz, collimated and focused onto a quad thermopile detector array with distinct filter bands, centered at wavelengths 2.7 μm, 4.26 μm, 3.95 μm and 6.58 μm (for demonstration purpose only). The parallel readout ADC codes for the four spectral channels are shown in Fig. 17.

Fig. 15

Block system diagram of the data acquisition system

Fig. 16

(a) Thermopile detector readout electronics; (b) MCD ASIC noise spectrum with an ADC sample rate of 240 Hz, channel gain of 250 and modulating frequency of 64 kHz; the input was shorted to the pixel reference to avoid interference from external noise sources

Fig. 17

Parallel ADC code readout from a quad thermopile detector array with narrow band filters centered at wavelengths 2.7 μm, 4.26 μm, 3.95 μm and 6.58 μm

4 Radiometric Calibration

The basic equation relating digital count output in a given channel to the external radiation flux is essentially (Sromovsky et al. 1992):

$$ C=Kf( {T} / {T_{0}} ) ( L_{u} - L_{d} ) $$

(1)

where $K$ is an absolute calibration constant, $f( {T} / {T_{0}} )$ is a relative response function describing the instrument response dependence on detector temperature, and where $L_{u}$, $L_{d}$ are the hemispherical integrals of radiance, i.e., the up- and down average radiances within the FPA FOV respectively and are given by:

$$\begin{aligned} L_{u} =& \int _{0}^{\infty } s_{\lambda } d\lambda \int _{0}^{\pi /2} \sin ( \theta ) d\theta \int _{0}^{2\pi } d\phi a_{u} ( \theta ,\phi ) L ( \theta ,\phi ,\lambda ) \end{aligned}$$

(2)

$$\begin{aligned} L_{d} =& \int _{0}^{\infty } s_{\lambda } d\lambda \int _{\pi /2}^{ \pi } \sin ( \theta ) d\theta \int _{0}^{2\pi } d\phi a _{d} ( \theta ,\phi ) L ( \theta ,\phi ,\lambda ) \end{aligned}$$

(3)

where $s_{\lambda }$ is the relative spectral response at wavelength $\lambda$; $a(\theta ,\phi )$ relative angular response at $\theta$ (angle from vertical), $\phi$ (azimuth angle); $L ( \theta ,\phi ,\lambda )$, spectral radiance at wavelength $\lambda$ angles $\theta ,\phi$, where $s_{\lambda }$ and $a(\theta ,\phi )$ are both normalized to have unit integrals. Because the thermopile detectors are hermetically sealed, no response to external pressure is expected. Because the Winston cones have fixed FOVs, no crosstalk between the detectors is expected.

Consider a flat radiation field, as an extended area blackbody, then $L ( \theta ,\phi ,\lambda )$ is given by the Planck radiance function $B_{\lambda } (T)$ and Eqs. (2) and (3) become:

$$\begin{aligned} L_{u} =& \int _{0}^{\infty } s_{\lambda } B_{\lambda } ( T_{u} )d\lambda \int _{0}^{\pi /2} \sin ( \theta ) d\theta \int _{0}^{2 \pi } d\phi a_{u} ( \theta ,\phi ) \end{aligned}$$

(4)

$$\begin{aligned} L_{d} =& \int _{0}^{\infty } s_{\lambda } B_{\lambda } ( T_{d} )d\lambda \int _{\pi /2}^{\pi } \sin ( \theta ) d\theta \int _{0} ^{2\pi } d\phi a_{d} ( \theta ,\phi ) \end{aligned}$$

(5)

Which reduces to the spectrally weighted average of the net Planck radiance. For a spectrally flat broadband channel, the average net radiation, $L_{n}$, is simply $L_{n} =( L_{u} - L_{d} )$.

During test and validation in a simulated thermal-vacuum environment, the instrument will measure (i) relative angular response for each Winston cone-detector assembly using a blackbody source operating at 500 K; (ii) relative spectral response for each spectral band (using a quartz halogen lamp between 0.2 and 3 μm and 1000 K SiC globar for wavelength 3 μm to about 40 μm); (iii) relative crosstalk coefficients between neighboring channels (ensure that this is less than 0.2%); (iv) relative response versus temperature for each channel and (v) absolute responsivity measurements for each channel using a source that fills the FPA FOV. The testing will also provide sufficient measurements to ensure that FPA digital count output as a function of input radiation flux and FPA temperature can be calculated and predicted.

5 Expected Radiometric Performance

The calculated SNR’s for three of the defined baseline filter channels on an IG-NFR descending through Uranus’ atmosphere are given in Table 7. The conservative calculations are based on the Winston cone étendue of $1.54\times 10^{-7}~\mbox{sr}\,\mbox{m}^{2}$, window transmission efficiency of 0.5, fold mirror reflection efficiency of 0.5, short wavelength bandpass filter transmission efficiency of 0.5, long wavelength bandpass filter transmission efficiency of 0.2, bandpass spectral radiances for both solar and thermal contributions taking into account solar reflection losses from the top of the atmosphere (0.3), the scene (0.7), emissivity of the scene (0.5), emissivity of optics (0.5), emissivity of housing (0.7) and averaging ${(1} / {2\tau )}$ based on the time constant ($\tau$) of the detector (36 ms) over a 1 s integration time. The system Noise Equivalent Power (NEP) at 298 K in a 1 Hz electrical bandwidth is derived from the detector responsivity ($295~\mbox{V}/\mbox{W}$) divided by the summation in quadrature of all primary noise voltage sources, i.e., detector Johnson noise (18 nV) and readout noise (12 nV), giving an overall system noise voltage of 22 nV, hence a system NEP of 73 pW (cf. detector NEP of 60 pW for 1 Hz bandwidth, see Table 6).

Table 7 Calculated SNRs at Uranus for 1 s integrations for three of the IG-NFR spectral channels defined in Table 3