The Design and Engineering of Curiosity

by Emily Lakdawalla

  8.2 RAD: RADIATION ASSESSMENT DETECTOR

  The Radiation Assessment Detector (RAD) is an energetic particle analyzer, performing the first-ever direct radiation measurements on the surface of Mars.1 It is funded by NASA’s Exploration Science Mission Directorate – the human exploration side of NASA. A major goal of the RAD investigation is to assess the hazard that energetic particles pose to future human astronauts. RAD was already detecting them while Curiosity cruised toward Mars, so RAD produced some of the first scientific results of the Curiosity mission. The RAD principal investigator is Donald Hassler of Southwest Research Institute, Boulder, Colorado.

  8.2.1 Scientific background

  Most of the energetic particles that RAD detects on the surface of Mars are galactic cosmic rays, except during solar events. Galactic cosmic rays are high-energy particles that probably originate in supernovae. Most of them (85–90%) are protons, and most of the rest are helium nuclei; electrons and heavier nuclei account for 1% each. The flux of galactic cosmic rays varies with solar activity: there are fewer cosmic rays near solar maximum, when the heliosphere pushes outwards and provides the solar system more protection from cosmic rays.

  Solar energetic particles originate in the solar corona along with flares and coronal mass ejections. Solar events can produce 10,000 times more particles than the normal background rate. During big solar events, solar particles can overwhelm galactic cosmic rays as the primary source of energetic particles at the surface of Mars, but only for brief periods of a few hours to a few days. Mars and Earth will usually not see the same fluxes of solar particles, because they usually see the Sun from different directions.

  Mars’ thin atmosphere shields the surface from lower-energy solar energetic particles but presents almost no barrier to cosmic rays. Mars lacks any significant magnetic field, so can’t deflect incoming charged particles, as Earth can. Almost all of the cosmic rays that hit the atmosphere reach the surface, producing a radiation dose 1000 times that experienced at the surface of Earth. As the incoming particles strike atoms in the Martian atmosphere and surface (and in rovers), the collisions generate secondary particles like neutrons and gamma rays, so the number of energetic particles at the Martian surface is actually greater than it is above the atmosphere. However, this increase in energetic particle flux doesn’t result in an increased radiation dose, because the heaviest, most damaging ions – such as iron nuclei – fragment into smaller nuclei as they experience collisions in the atmosphere, with only 25% of them reaching the surface. So Mars’ atmosphere does provide a small amount of protective shielding, but many protons still reach the surface unimpeded, and the breakup of larger nuclei results in more protons than in space.

  Mars’ surface has presumably been exposed to this bombardment from energetic particles for many millions (perhaps billions) of years. Ionizing radiation has likely changed the chemistry of the Martian surface. It could be one cause of the weathering rinds that have formed on Martian rocks. Mars is the inverse of Mercury in terms of its experience of space weathering: Mars suffers ionizing radiation but not micrometeorite impacts, while Mercury (which has a weak magnetic field but no atmosphere) suffers micrometeorite impacts but little ionizing radiation. The Moon withstands both; Earth, little of either.

  8.2.2 How RAD works

  RAD is located within the body of the rover. It looks upward through a thin Kapton window on the rover deck that allows particles to pass into the instrument (Figure 8.1). The sensor has a view cone 65° wide. The size of the cone represents a trade-off between observing as wide a portion of the sky as possible and keeping the mass and volume of the instrument as small as possible. Its sensor head contains multiple silicon detectors in a vertical stack. Depending on their energy, charged particles may penetrate none, some, or all of the detectors. Besides charged particles, RAD can also detect neutrons and gamma rays. RAD accumulates about 400 kilobytes of data in an ordinary sol. It has 16 megabytes of RAM.

  Figure 8.1. Top: The top of the RAD sensor head is visible as a circular plate on Curiosity’s deck in this Navcam photo (NRA_409403780RADLF0051858) from sol 134. The plate covers the cylindrical instrument (inset). NASA/JPL-Caltech/Emily Lakdawalla. Bottom: schematic diagram of the RAD detectors. Colored lines show possible paths of charged and neutral particles that RAD can detect. Diagram modified from Rafkin ( 2014 ).

  In order to measure the energy, mass, and charge of incoming particles, RAD examines how many of the stacked detectors a particle passes through and how much energy the particle loses as it passes through the detectors. At the top of the stack are three solid-state silicon detectors (labeled A, B, and C in Figure 8.1). Then comes a thick cesium iodide scintillator (D) and a plastic scintillator (E). Finally, another plastic scintillator (F) encloses D and E. Neutral or charged particles interacting with scintillators D, E, and F cause the scintillating material to emit orange light which is, in turn, detected with photodiodes. In order to make sure that all the detectors are looking at the same population of particles, the planar A, B, and C detectors have different widths, and D has a pyramid shape, defining a viewing cone about 65° wide.

  Charged particles with very high energies can penetrate the entire instrument (purple line in Figure 8.1). Lower-energy ones may get stopped part of the way through (blue lines). Charged particles must pass through at least the top (A) detector and register in the B detector in order to be counted as an “event.” If a charged particle strikes only detector A, or any other lower detector without hitting the upper ones, it will be rejected from analysis (red lines). The role of the anticoincidence shield (detector F and outer rings on detectors B and C) is to detect the charged particles coming from the “wrong” directions and allow the RAD team to reject events triggering detections in D and E that did not enter from the direction of the viewing cone. If particles stop within the detector stack, RAD can determine their charge, mass, and energy. Some particles will pass all the way through, but will be dramatically slowed, and will deposit much of their energy in detector E. In this case, RAD can determine the charge and energy of these particles, but not their mass.

  For neutral particles like gamma rays and neutrons, the viewing cone doesn’t matter; RAD detects them coming from all directions as events detected in D and E but not any of the other detectors. Curiosity’s own MMRTG generates lots of gamma rays and neutrons, but most of them are at low energies. RAD can detect these particles, but threshold parameters are set to reject them so as not to saturate the processing electronics, which are optimized to measure naturally occurring particles. The high atomic mass of the cesium iodide composing scintillator D makes it effective at detecting incoming gamma rays. The plastic in scintillator E makes it poor at detecting gamma rays, but good at detecting neutrons. Neutrons striking hydrogen nuclei produce recoil protons that may, in turn, produce an event in detector D. (For more about the interactions between fast-moving neutrons and hydrogen, read section 8.3 about DAN.)

  RAD was originally planned to operate continuously in the background, around the clock. Unfortunately, constraints on the power system were initially quite conservative. As a result, RAD was first proposed to operate on a one-hour cycle, awake part of the time and asleep part of the time. The one-hour cycle was chosen in order to allow RAD to notice the onset of solar particle events soon after they begin, and change its observation cadence without any commands from Earth.

  8.2.3 Using RAD

  RAD works independently of the rover’s other activities. RAD began the mission taking observations once per hour for 16 minutes, and sleeping for the other 44 minutes. Over the course of the mission, RAD operations have become nearly continuous. As of sol 1800, a typical cadence is 16 minutes 10 seconds of observations, followed by only 27 seconds of sleep. During its sleep period, RAD bundles its ongoing observations and resets the instrument.2 RAD operates even when the rover’s main computer is asleep, gathering and storing data until the rover main computer requests it in preparation for downlinking it to Ear
th, originally about twice per week, but now at the start of every UHF communications pass. Each time RAD wakes up, it performs a 10-second “pre-observation” measurement. If it detects a high particle flux during the pre-observation measurement, it automatically shifts into a solar event mode, in which it makes more frequent observations. In solar event mode, RAD flags the data as high priority, so that the next time the rover communicates with Earth, the mission will learn of the solar event. This has happened only about 5 times since landing because of the unexpectedly low activity of the most recent solar maximum.3

  One interesting finding from RAD is that diurnal changes in atmospheric pressure have an effect on the number of energetic particles reaching the surface, and the effect was large enough (and RAD sensitive enough) for RAD to detect it.4

  Figure 8.2 shows some sample RAD data. RAD has been operating almost continuously since sol 9. The only hiccup in RAD operations came early in the mission, with an outage from sol 29 to 34 caused by an unexpected problem in the way that RAD and the main rover computer communicated with each other. Once the problem was understood, the rover engineers developed a workaround, and the problem hasn’t happened since.5 Other gaps in RAD data have causes external to the instrument, such as rover software updates, the sol 200 anomaly, and conjunctions.

  Figure 8.2. Dose rate measured by the RAD E detector during the first 350 sols of the mission. An astronaut would experience daily doses of roughly 200 milligrays. For context, medical X-rays and CT scans typically impart from 0.01 to 10 milligrays, depending on the type. From Rafkin et al. ( 2014 ).

  Unlike any of the rest of Curiosity’s instrument data, RAD data are archived at the Planetary Plasma Interactions node of the Planetary Data System.

  8.3 DAN: DYNAMIC ALBEDO OF NEUTRONS

  Dynamic Albedo of Neutrons (DAN) surveys the ground up to a meter underneath Curiosity’s traverse for chemically unusual spots and for the presence and abundance of subsurface hydrogen. The Federal Space Agency of Russia contributed DAN to the Curiosity project. Its Principal Investigator is Igor Mitrofanov of the Institute for Space Research (IKI). DAN comes from a long line of in-space neutron detectors going back to the Luna and Apollo missions. Its design is based upon the High Energy Neutron Detector (HEND), part of the Gamma Ray Spectrometer on Mars Odyssey. HEND and other neutron detectors flown to Mars, the Moon, Mercury, and asteroids have mapped water ice and mineral-bound water across the inner solar system.6

  8.3.1 Scientific background

  How do neutrons reveal the presence of subsurface hydrogen to these instruments? It begins with the same galactic cosmic rays and high-energy solar particles, that RAD detects (section 8.2.1). Curiosity also constantly emits neutrons from the plutonium decaying in its MMRTG. They bombard Mars’ surface, colliding with atoms in the rocks and soils surrounding the rover (Figure 8.3). The impacts have so much energy that they can excite atomic nuclei into higher-energy states. The nuclei emit neutrons and other nuclear particles as they return to their lower-energy states. This process creates a constant source of high-energy neutrons. The neutrons lose energy with each collision, until they reach an equilibrium (or “thermal”) energy. If neutrons escape before experiencing enough collisions to be “thermalized,” they are “fast” neutrons. Slower neutrons are “epithermal,” and the slowest, “thermal.”

  Figure 8.3. Cartoon showing the various sources of nuclear radiation from a planetary surface. Most result from incoming cosmic rays. NASA/JPL-Caltech/UA.

  Neutrons are low in mass compared to most atomic nuclei; the neutrons bounce off most nuclei with nearly the same energy they began with. But when a neutron collides with the nucleus of a small atom, the atom’s nucleus recoils from the collision, and the incoming neutron loses speed. In the limiting case, an incoming neutron collides with a hydrogen nucleus – a proton – which has the same mass as the neutron. Therefore, the presence of hydrogen in soil dramatically slows neutrons, and there are more thermal and fewer epithermal and fast neutrons emitted from the surface above soil that is hydrogen-rich.

  Another way that neutrons interact with soil atoms is by neutron capture. Nuclei with relatively large cross sections, like iron and chlorine, are more likely to capture neutrons; they’re especially likely to capture the slowest (thermal) neutrons. By measuring the relative abundances of epithermal and thermal neutrons – that is, measuring the static albedo of neutrons – a neutron spectrometer can constrain the abundance of hydrogen and neutron-absorbing elements (not individually, but in bulk) in the subsurface.

  Curiosity’s Dynamic Albedo of Neutrons instrument goes a step further than this static measurement. DAN includes an active neutron source, a Pulsing Neutron Generator (PNG). By actively bombarding the surface with neutrons and then counting how the flux of epithermal and thermal neutrons vary with time, the DAN team can infer the distribution of hydrogen (or other neutron-slowing or -absorbing species) with depth beneath the surface, for the upper meter of soil. DAN is the first spaceborne neutron detector to have an active neutron source. Active neutron experiments are more commonly used in oil exploration geology on Earth, where neutron detectors and generators are lowered into boreholes to scan for the presence of hydrogen-rich hydrocarbons in subsurface rocks. DAN’s neutron source is similar to Russian industrial instruments.

  8.3.2 How DAN works

  DAN has two components, a module with the detector and electronics, and the neutron generator (Figure 8.4). The detector is located in the left rear corner of Curiosity, and the neutron generator is in the rover’s right rear corner (Figure 8.5). The rear Hazcams are mounted to the outside of the corner boxes containing DAN.

  Figure 8.4. Components of the DAN instrument. Photos from IKI Laboratory for Space Gamma Spectroscopy.

  Figure 8.5. Location of the DAN instrument components on the rover. Top photo: NASA/JPL-Caltech release PIA14257. Lower left: PIA15181. Lower right view with belly pan removed: NASA/JPL-Caltech, annotated by Emily Lakdawalla.

  The detector has two counters, both of which are filled with helium-3 at a pressure of 300 kilopascals. Detection happens when a neutron is captured by a helium-3 nucleus, which produces a proton and a triton (a hydrogen-3 nucleus). The two counters differ in their shielding. One is enclosed in a shield made of lead, and the other in a shield made of cadmium. Both lead and cadmium screen out X-rays, but the cadmium also shields out low-energy (thermal) neutrons. The cadmium-shielded one detects only epithermal neutrons and is called the Counter of Epithermal Neutrons (CETN), while the lead-enclosed detector is called the Counter of Thermal Neutrons (CTN). The lead-enclosed detector will always count more neutrons, and subtracting the counts of the cadmium-shielded detector from those of the lead-shielded detector will yield a count of thermal neutrons.

  DAN can perform passive neutron detection continuously, but can also be commanded to operate in an active mode. Curiosity must be sitting still to perform a DAN active observation. The generator is a compact ion accelerator that steers deuterium ions into a tritium-enriched target to generate neutrons with energies of 14.1 MeV. (For comparison, incoming galactic cosmic rays have energies ranging from about 10 to 20 MeV.) It generates 13.4 million neutrons with each pulse, all within a period of about 2 microseconds. It can be operated with a single pulse, but usually generates 10 pulses per second.

  After an active pulse it can take several milliseconds for neutrons of different energies to leak out of the surface. The detectors count up the arriving neutrons over time, producing a “die-away curve,” a graph of the number of counts with respect to time since the pulse for each detector. Stacking die-away curves from many pulses improves the signal-to-noise ratio of the DAN data. The most commonly used DAN active observation includes 20 minutes of 10-hertz pulsing, or about 12,000 total pulses. The rover usually takes rear Hazcam images during a DAN active observation to document the kind of material underneath the DAN instrument at the time.

  The detector is expected to count roughly 10 neutrons returning from each
pulse, or about 100 counts per second during 10-hertz pulsing. The amount of neutrons that leaks varies by a factor of a few, depending upon how much hydrogen and other neutron absorbers are present in the surface. For comparison, the continuous neutron emission from the MMRTG produces about 25 and 10 counts per second in the lead- and cadmium-shielded detectors, respectively. During cruise, DAN detected higher counts of 35 and 15 per second as a result of exposure of the spacecraft to galactic cosmic rays. Once on Mars, the background increased even further because of the response of the surface of Mars to galactic cosmic rays. Thus the background is comparable in magnitude to the dynamic contribution from the neutron generator. The DAN team uses the tenth-of-a-second delay between pulses to measure the background, which can be removed from the results of active surveys.

  Turning DAN active neutron counts into estimates of subsurface hydrogen abundance requires mathematical modeling. The DAN team performs simulations with a large set of models for the subsurface. The models all begin with a typical Martian soil composition (based on APXS measurements from the Mars Exploration Rover mission). They also assume a rate of incoming cosmic radiation, which is dependent upon the density of the atmosphere above the rover at the time of the measurement, so they incorporate REMS data on the atmospheric pressure at the time of the active DAN measurement. They allow other model parameters to vary. Some models are homogeneous ones, in which the total abundance of hydrogen and chlorine are allowed to vary. Other models are two-layer ones, in which chlorine is held constant but the amount of hydrogen is allowed to vary in upper or lower layers, and the layer thickness also allowed to vary. The models spit out die-away curves, and then the DAN team performs a least-squares analysis to find which set of model parameters best fits the observed die-away curve.