Asteroid of the Half-Month: 25143 Itokawa Part II

Soichiro Honda (yes, that Honda) once got asked the secret of his success. He reportedly said ‘One word: lucky.’ During the rise of Honda and similar manufacturers, Japan was dismissed as a land of cheap knockoffs. That era is clearly done. Japan has done the world’s first asteroid sample return, Hayabusa, though luck (good and bad) was clearly part of its story.

Japan has long wanted asteroid materials. In Jun 1985, nearly two decades before the launch of Hayabusa, a sample return meeting¹ was gathered. This itself was after NASA and the NSF led an oversubscribed, 1971 international conference² on asteroid exploration. In that and the following years, Peak Oil and the embargos added even more urgency if not funding. Through Hayabusa, we found something even more enabling: water in space.

Quick recap: the Hayabusa mission launched in May 2003, reached asteroid 25143 Itokawa, and gathered a sample. It returned in Jun 2010, dropping a reentry pack.

But back to 1985. That year, a Japanese team was assembled to analyze sample returns. They devised a mission to 1943 Anteros and back. Its trajectory was “surprisingly but accidentally very closely identical to the orbit of Hayabusa.” However, given 1985 technology (like chemical rockets) the mission was far too big and heavy (Kawaguchi et al. 2006). But by 1994, the Clementine probe had tested miniaturization, in flight. Crude forms of electric propulsion were flying, and ion thrusters were about to (gradually) take over communications satellites (such as Japan’s own ETS-6). Before Clementine even disbanded, a new Japanese team had formed, to study electric rockets for Mars and Near-Earth Objects (NEOs). They moved fast (for space), publishing a paper the next year on electric-thrust sample return (Kawaguchi et al. 1995). That mission got approved by Apr 1996, the start of Japan’s next fiscal year.

You may be thinking ‘April 1996 approval to May 2003 launch? What took seven years?’ There’s the story- the story of Japanese organizations trying to be first, not second at some feat. It was obvious Japan had no ground infrastructure comparable to NASA’s Deep Space Network (DSN), for communicating with distant probes. Nor had Japan gotten experience in tracking such deep space probes. While two probes (Suisei and Sakigake) had been launched as part of the ‘Halley Armada’, Japan’s contribution flew through free space, not targeting Halley’s Comet with any precision. Navigation and tracking were not critical at all. Japan’s responsible bureau (ISAS- Institute of Space and Astronautical Science) got a partnership agreement with NASA to handle such requirements (via the DSN) within their project’s first year.

That year (1995), the planned launch vehicle program (Mu-V, the fifth in the series given the Greek letter Mu) was delayed. First flight would not be until 1996. Even then, a small solid rocket would be added, as a final stage to add capacity. This is a common US practice.

Besides electric propulsion, one enabling technology that would allow a mission was proximal operations. Asteroids have low gravity, so one does not really land on them; ‘landing’ is more like docking. Still, the speed of light is not instantaneous. Radio signals take 18 minutes to make the trip (‘light time’), so ground crews cannot control- or even watch- these events ‘live’. A probe would need an ability to ‘dock’ and depart autonomously. The study team determined that all this could take place with existing cameras and electronics- no magic necessary. Their 1995 reports are surprisingly close to the actual Hayabusa. Still, a probe would have to drop landing aids- reflective markers- onto the asteroid. In 1995, the markers were shiny disks, which would be visible to cameras if either side faced up. At launch, the markers had become cubes. There was a chance a disk would hit the ‘soil’ sideways and fail to appear on camera.

The module for Earth return was not advanced- NASA flew similar heat shields on the Pioneer Venus mission, and the atmospheric probe of Galileo. The asteroid cache was tiny, making the problem easier. Still, ISAS had NASA test the heat shield in a NASA facility, under NASA review. If there was an issue, it was that the module’s interior design was a tight squeeze.

An original Japanese technology was electrodeless thrusters. Ion engines use electrodes; these are vulnerable to arcing and wear. Hayabusa thrusters heat the propellant (xenon) via microwaves (electron-cyclotron resonance), reducing their vulnerability to hot gas.

While the project development (MUSES-C, for Mu Space Engineering Spacecraft, third in its series) was underway, so was the search for asteroids. For over a human lifetime, mankind searched for them using photographic plates. The introduction of CCDs- digital cameras- would revolutionize sky searches. Once CCD detectors got large enough, they were mounted to telescopes to find faint objects. The LINEAR program (Lincoln Near-Earth Asteroid Research), from MIT’s Lincoln Laboratory, wasn’t the first to use CCDs. But until just a few years ago (with the Pan-STARRS program), it was the best. LINEAR had discovered more asteroids than any other group. This includes asteroid 1998 SF₃₆, spotted in late Sep 1998. By this time, discovery rates were high enough that the subscript numbers (the sequence of discoveries within a reporting period) were reaching the thirties and forties.

The spacecraft was proceeding, though not without incident. Most technologies were not that new, having flown on e. g., Clementine, NEAR, Mars Global Surveyor, etc. The main camera (AMICA- Asteroid Multi-band Imaging CAmera) did had a custom, low-noise CCD. The chemical sensor (XRS- X-Ray Spectrometer) flew a solid-state CCD, not a gas or crystal. Batteries went solid-state too- lithium-ion, not nickel-hydrogen. What was an issue were programmatics. Development was not on schedule, and the craft was overweight- 510 kg, versus 480 specified. As an exclamation point, the third flight of the Mu-V rocket exploded in 2000. Even if the probe was ready, its launcher would not be. The mission slipped from its primary target (4660 Nereus), to the backup (1989 ML), then the backup backup, 1998 SF₃₆.

Provisional designations just aren’t memorable. Hideo Itokawa, the ‘father of Japanese rocketry’, had died in 1999. The team asked that their target, 1998 SF₃₆ be named Itokawa. LINEAR, who, as the discoverers, got the naming rights, agreed. Others were not so feted- about a year before launch, AMICA’s staff got sacked and replaced. XRS also had issues.

To be fair, NASA had its misstep. In return for flight support, NASA was given some mass for a payload. A rover was chosen, dubbed MUSES-CN (for MUSES-C plus NASA). Even smaller than the Sojourner rover on Mars, the “nanorover” was as big as a hardbound book. It had a camera, needed to steer it, and an infrared instrument. It was then felt that an alpha-particle instrument should be added; this would take mineral data, besides the infrared readings. After all, Sojourner had an alpha-instrument. But this book didn’t sell; the project spiraled out of control and got cancelled. Let this be a lesson to the backseat drivers who say ‘why don’t you add ____?’ It should certainly be a lesson to those saying ‘how can you not add ____?’ Everything is discretionary spending, as the missions themselves are discretionary.

The craft did launch, on May 9 2003 instead of Jan 2002. That’s the beauty of asteroid missions- take your pick, there’ll be one. The name officially changed from MUSES-C to Hayabusa (“falcon”) after the launch was a success. The ion engines lit over May 26 and 27, firmly establishing that technology. But a solar flare hit the probe in November. Solar activity isn’t rare, but this one did cause radiation damage to the solar cells, cutting power slightly. Still, a pass by Earth for a gravity assist, in May 2004, went successfully. Instead of two trajectory corrections with the chemical (hydrazine) thrusters, the mission simply rescheduled ion thrusting. The high efficiency of electric propulsion gives projects lots of flexibility.

During this time, the asteroid made a close pass by Earth. Just as in 2001, scientists on and off the project used the opportunity to observe Itokawa. Besides retiring risks for the probe, it’s a valuable exercise. Instruments and analyses can be checked versus later probe results.

The mission reached 25143 Itokawa on Sep 12 2005, again not without incident. On Jul 31, a reaction wheel failed. Reaction wheels turn a spacecraft, by adding or subtracting rotational energy. Most probes have three such wheels, to turn in three dimensions. Wheels have failed before, and ISAS had switched to two-wheel operations as needed. But a second reaction wheel failed Oct 2. Some operations would need Hayabusa to expend thruster fuel.

After about three months of scans (including gravity tracking), the project would try sampling. Nov 4, a trial descent was run, without contact. One of the landing markers was ejected as a test, but it escaped to space. Nov 12 saw a second trial. A micro-lander, MINERVA, was released. Weighing only 600 grams, it could hop on Itokawa by torquing its reaction wheels. This would allow crude but close imaging, and temperature data at each site. But deployment occurred just as Hayabusa’s onboard guidance was firing a thruster. The mother craft was ascending slightly, not descending, and MINERVA, too, drifted off instead of landing.

On Nov 19 2005, a descent begun; the onboard controls triggered an abort. But Hayabusa did, in fact, contact Itokawa, even bouncing and resettling. By the time ground staff regained situational awareness, 30 minutes was spent on the surface. Luckily, no damage was found; again, gravity is so low, it wasn’t much of a fall. Nov 25, an actual sampling was planned. Hayabusa would then thrust back towards Earth in December, for arrival in Jun 2007.

To sample Itokawa, Hayabusa had a collector horn or funnel. Instead of a feed or scoop, the horn had a ‘pellet gun’. A gas drive would fire a tiny bullet, kicking surface materials. The shot speed, as fast as a rifle, could even knock bits off solid rock faces. (Itokawa, fortunately, had lots of gravelly areas.) Some rock would drift up the horn, and into two collection chambers in a sample canister. The can would then be sealed, and driven into the reentry module. On returning to Earth, the module and sample would be ejected, for reentry and landing.

When the Nov 19 attempt aborted, the computer’s fault mode included safing the sampler gun. No surface material was actively driven up the horn. For the Nov 25 run, some sensors and modes were disabled, to prevent another fault and ensure pellet firing. But when Hayabusa made contact, and tried to fire a pellet, software conflicts still blocked it. The probe ascended again with all pellets aboard but, some were fearing, no asteroid samples.

Granted, this was a moderate setback. The mission still demonstrated most technologies, and did scan an asteroid at close range- as low as 3 km. And there was always a chance that dust did enter the collector, pellet or not, just by brushing the horn rim on the surface. But that wouldn’t matter if the craft didn’t make it to Earth. On the ascent from Itokawa, a thruster leaked, depleting the chemical propellant and sending Hayabusa into a spin. Not only would engines not work while spinning, but radios would not send any clear signal.

For over a month, Hayabusa was spinning in silence. The Itokawa departure window had been missed. Finally, Jan 23 2006, a weak signal was received; Hayabusa was alive if not healthy. By Feb 25, the dish antenna was in line for a strong signal. Readout indicated the battery had been compromised (loss of capacity) by months of inactivity, but there was no critical damage. No eclipses would occur in free space, so battery capacity isn’t really an issue.

Despinning and thrusting were needed. With no chemical thrusters, xenon propellant was emitted through the neutralizers of the ion engines. Though the neutralizers weren’t intended as propulsion, they just needed to brake the tumble. Hayabusa was steadied by May 8.

Since the departure window was missed, and by five months, Earth and Itokawa were out of phase in their orbits. A new window would open in Spring 2007. On Apr 25, Hayabusa then tried to make it to Earth. One reaction wheel worked, and the steering thrusters had no propellant left. But the ion engines were on a gimbal platform. A limited amount of steering could be had by tilting the ion exhaust slightly off-center, as had always been planned.

A bigger issue was engine life. The ion thrusters were designed for 10,000-13,000 hours, and tested to 18,000. As feared, the engines eventually ‘failed.’ Hayabusa has four ion thrusters, named A through D. To continue to Earth, Hayabusa was reprogrammed to cross-fire the last working thruster, B, with the last working neutralizer, A. One of the mission goals was in-flight demonstration of minimum engine life; it adapted its way to maximum life.

The craft thus managed to limp back to Earth. The mission had been to aim for Earth, release the reentry module, then have the mother craft divert back to space. With no hydrazine and one xenon thruster, that was out. The probe aimed for the Woomera Range in Australia, a giant, secure military testing ground. The reentry pack landed on Jun 13, 2010, three years behind schedule, or two full orbits of Itokawa. The main Hayabusa craft, having no heat shield, burned up behind the sample container, creating a dual fireworks display.

So ends the space tale, but not the science investigation. Japanese recovery staff immediately found the sample pack, and returned with local soil for comparison. With such a tiny sample- if any- it remained to be demonstrated that the canister contents weren’t construction debris, samples of Australia, etc. The destruction of the rest of Hayabusa was itself monitored as a reentry experiment. Consider it an engineering trial, or an artificial and known meteor.

The cache was taken to Japan’s Sample Curation Facility. (The second in the world- Europe doesn’t have one.) Part of the mission was founding a secure, contamination-controlled lab. Project staff opened the can- in a clean room, under nitrogen gas- and got particles. Incidental contact had stirred Itokawa’s dust. More recently, it is postulated that static charging had lofted fine particles. Also recently: some particles are seen to be hydrated (Jin et al. 2018)…

Late or not late, limping or not limping, Hayabusa made it home. We now have samples, kept in a can, of a known body, from known spots on that body, with instrument observations for both global and local context. No meteorite gives that. Late and limping, we’re still lucky.

1. Asteroid Sample and Return Workshop, ISAS, Jun 29 1985
2. Physical Studies of Minor Planets, 12th Colloquium of the IAU, Mar 6 1971, Tucson AZ

Kawaguchi, J. Scientific Satellites Prospect. ISAS Report no. 43, Dec 1986
Kawaguchi, J. Kuninaka, H. Toki, K. Matsuo, H. Mizutani, H. 1995 Acta Astronautica Supplement vol. 35 p. 193
Kawaguchi, J. Kuninaka, H. Fujiwara, A. Uesugi, T. 2006 Acta Astronautica vol. 59 p. 669
Jin, Z. Bose, M. Peeters, 2018 49th Lunar and Planetary Science Conference, abstract 2083