Five Billion Years of Solitude

by Lee Billings

  Spurred by Goldin’s enthusiasm and the tacit support of the Clinton administration, NASA quickly greenlighted SIM and convened working groups to solidify plans for TPF-I. Both projects eventually ran into major difficulties. Riding a strong initial pulse of funding, SIM met or surpassed all of its key developmental milestones, but by the mid-2000s the ballooning costs of JWST and of the Bush administration’s new Constellation program had reduced the project’s funding to a dribble. Most astronomers were unconcerned—SIM’s hyperspecialization seemed to offer little to the broader community. Even many planet hunters thought it superfluous, and hoped to simply skip over it to build a much more capable TPF. The mission was repeatedly downgraded and its launch continually delayed, piling on empty expenses until, after consuming more than half a billion dollars, in 2010 SIM was quietly cancelled and its nearly complete flight hardware junked or repurposed.

  TPF-I faced a different problem: As its working groups delved deeper into the related technological hurdles, they realized the initial estimates for the mission’s cost and launch date were hopelessly optimistic. Cryogenically cooling all the separate mirrors would be costly and difficult. Reaction wheels required to rotate and point the mirrors on their long boom would cause the entire assembly to vibrate, potentially ruining observations. New designs emerged, including one from the European Space Agency’s own TPF-I project, code-named “Darwin.” Darwin and related concepts would eliminate vibrations by discarding the long boom in favor of a free-flying array of several mirrors that would gather light and direct it into a central beam-combining hub. Instead of one cryo-cooled spacecraft, the project would now require five or six, each needing to fly in formation with centimeter-scale precision in deep space, drastically increasing mission complexity and the amount of propellant required. The runaway cost growth of the complex, cryogenic JWST suggested that, if anything, TPF-I’s early cost estimates of $1.5 billion would balloon to make it even more ruinously expensive than its predecessor. By 2001, JPL’s notional launch date for a TPF-I had slipped to no earlier than 2014, and mission planners were looking for cheaper alternatives, ideally a single non-cryogenic telescope.

  Conventional wisdom held that the very same slippery wavelike behavior of light that enabled interferometry would prevent any single filled-aperture telescope from ever imaging Earth-like exoplanets. To capture the 10-billion-to-one photons emanating from an alien Earth at optical wavelengths, the light must be strictly controlled, the star’s overwhelming glare removed. Yet when starlight falls upon a single mirror it flows in liquescent wavelets, pooling and puddling in frozen ripples and coruscating speckles around the most minuscule surface imperfections. Even a mathematically perfect mirror, of the sort that only exists in computer simulations and the late-night dreams of theorists, would not be immune: light from a point-like distant star striking an ideal circular mirror would still diffract off the mirror edges, forming a central bright disk surrounded by a concentric series of rings. A good number of the disks, ripples, rings, and speckles tended to manifest precisely in the part of a star’s image where one would expect to find any lurking habitable planets. Each aberration would typically only be about a hundredth as bright as a target star, but would still be some eight orders of magnitude brighter than the faint light of any accompanying small, rocky worlds, rendering planetary detections improbable, if not entirely impossible. This was the scientific consensus as presented by any up-to-date optics textbook at the turn of the twenty-first century. And it was totally wrong.

  The key to a single-telescope TPF solution was a device called a coronagraph that could, in theory, blot out a star’s diffraction disks and rings. Invented by the French astronomer Bernard Lyot in 1930 to observe the hot, nebulous corona that surrounds the Sun, a coronagraph is any occulting object placed in front of a telescope’s mirror to block out the unwanted light of a target star. To see a coronagraph in action, make one of your own. Hold your right thumb over the Sun’s disk in the sky to prevent most of its glare from reaching your eyes—the principle is the same. You may notice, however, that even if the Sun is entirely blocked, a small amount of sunlight still diffracts around your thumb’s edge. You can dampen some of that extra glare by placing your left thumb a short distance directly behind your right thumb as an extra barrier to block the Sun from your line of sight. In his coronagraphs, Lyot did something similar, crafting a series of “pupil” lenses, partially transparent “masks,” and disk-shaped opaque “stops” that progressively stripped out the residual light scattered off the edges of the initial occulter. Lyot’s instruments were suitable for imaging the Sun’s corona, which is a million times fainter than the Sun itself. But they leaked far too much stray light into a telescope’s optics to allow for the crucial 10-billion-to-one starlight suppression required to image an exo-Earth in visible light.

  In 2001, while pondering the mounting intricacies of TPF-I, the Harvard-Smithsonian astronomers Wesley Traub and Marc Kuchner struck upon a concept for a new class of coronagraphs, ones that more explicitly relied upon interferometric principles to suppress starlight. Traub and Kuchner found that by superimposing interferometric nulling patterns of spirals or bars on a coronagraphic mask and carefully tweaking the shape of a coronagraph’s stop, they could simultaneously increase the total amount of starlight suppression to block 99.999999999 percent of a star’s light while also channeling the residual starlight to a thin outer region, away from the coronagraph’s central dark shadow. A star’s light would be blocked, nulled out, and finally swept away to the edge of a detector, while the faint light of a nearby planet would pass through unimpeded to form an image within the shadow. The scheme worked almost flawlessly in tightly controlled laboratory tests. Traub and Kuchner’s coronagraphs were straightforward to manufacture, but each mask typically worked best only for a few wavelengths of light rather than a star’s full spectrum. Around the same time that Traub and Kuchner were working on their coronagraphs, another astronomer, David Spergel of Princeton University, independently devised a completely different arrangement of shaped coronagraphic masks and pupils that also achieved extreme starlight suppression.

  JPL and NASA took note, and began funding research into a coronagraphic TPF, TPF-C, a planet-finding telescope meant to operate in optical rather than infrared light. Soon, a rough architecture had emerged: TPF-C would use a large 8-meter monolithic mirror mated to one or more specialized starlight-suppressing coronagraphs inside the telescope. The monolithic primary mirror would be oval-shaped rather than circular, so it could fit into a rocket fairing; segmented foldable mirrors like JWST’s produced too many wavefront aberrations to be compatible with the ultrasensitive coronagraphs. In the aftermath of the breakthroughs by Traub, Kuchner, and Spergel, even more coronagraph designs emerged. Starlight could be suppressed in myriad ways, weakened in mazes of grooves, twisted in networks of spiraling vortices, or phased out and dispersed in labyrinths of masks and lenses. But all still leaked some fraction of unwanted light. The degree to which each design sprayed splashes of photons elsewhere in the telescope was called “the order of the null.”

  Strengthening the order of the null required piping perfect, flawlessly symmetrical wavefronts of light to and through the coronagraphs. The barest fraction of a guidance error in the telescope’s pointing would cause the collimated beam of starlight to almost imperceptibly “walk” across new paths, wandering over different patterns of imperfections in the mirrors, weakening the order of the null. TPF-C would need to point with an accuracy more than five times that of the Hubble. A surface deviation of less than the diameter of a single silicon atom anywhere in the telescope’s reflective surfaces would send imperfect wavefronts cascading down the optical train, weakening the order of the null. Compared with Hubble’s, TPF-C’s mirrors would need to be some hundred times smoother. Producing and maintaining such precise pointing and figuring would require significant and costly breakthroughs in vibrational control, active optics, and mirror manufacturing, but those tasks st
ill seemed cheaper than any conceivable TPF-I mission. TPF-C would be less capable and sensitive than a TPF-I, and would survey fewer stars for potentially habitable worlds, but its probable lower costs made it a winner to budget planners at NASA and JPL. Year by year, TPF-C became ascendant, and TPF-I’s fortunes declined.

  Sensing the change in the winds, in 2005 JPL offered Traub a job, both as TPF-C’s project scientist and as chief scientist overseeing NASA’s exoplanet science programs. He would lead a team of approximately fifty people working to the tune of $50 million per year to do whatever it took to get TPF launched on time. Accepting the position would mean pulling up his stakes in Massachusetts and moving across the country. He was approaching seventy years of age, and knew if he left he would be uprooting old friendships as well. Still, Traub quickly made up his mind and took JPL’s offer. He reasoned the new opportunity would be worth the sacrifice—within perhaps a decade, if all went according to plan, he would be leading TPF’s team as observations streamed in, poring over the gathered light from other, distant habitable worlds. Of all the people to ever live—past, present, and future—perhaps Traub would be one of the lucky few to first find life beyond Earth, beyond the entire solar system. He could play a crucial role in the most profound development for humanity since the discovery of fire. Traub soon arrived in sunny Pasadena and settled into a small rented apartment.

  As the fates of TPF-C and TPF-I diverged, a third method for suppressing starlight emerged, largely based on the work of Spergel, his Princeton colleague Jeremy Kasdin, and Webster Cash of the University of Colorado. All three researchers were concerned by the extreme tolerances required for TPF-C’s mirrors. Rather than putting a coronagraph inside the telescope, inviting in all that corrupting, contaminating starlight, they proposed placing a coronagraph outside the telescope, as a separate free-flying spacecraft that would prevent any stray starlight from ever entering the optical train. They called the free-flying coronagraph a “starshade,” and their simulations of its performance revealed that its ideal shape for diffracting and nullifying starlight would closely resemble a many-petaled sunflower. Unlike the starlight-suppression techniques of TPF-I or TPF-C, which required reams of custom-built kit and only operated in a limited number of wavelengths, a starshade would simply cast a deep shadow onto a telescope, any telescope, allowing more broadband spectroscopy to widen the search for biosignatures. A starshade’s telescope would not require cryogenic cooling like TPF-I, or ultra-smooth monolithic mirrors like TPF-C—any space observatory with a sufficiently large general-purpose mirror would do, including NASA’s planned JWST.

  But building and operating a starshade would be no easy task—many of the extreme tolerances associated with the TPF-C telescope would instead just be exported to a separate spacecraft. Most designs envisioned a starshade between 50 and 100 meters in diameter, gossamer-thin and razor-sharp at its edges, sheathed in dark anti-reflective coatings, floating anywhere between 50,000 and 150,000 kilometers in front of a space telescope. For comparison, the distance between the Earth and the Moon averages some 380,000 kilometers; properly aligning a starshade’s shadow with a telescope would require exquisite orbital control. The starshade would have to autonomously unfurl in space and preserve its vast shape to submillimeter-scale precision, all while using small high-impulse thrusters to linger on or slew between targets. Where an agile TPF-C could switch targets in seconds or minutes, for a starshade the task would take days or weeks. A starshade would survey fewer stars than TPF-C, but at a potentially lower total cost. The starshade concept came to be called TPF-O (“O” stands for “occulter”), but its relatively late entry into serious consideration would relegate it for years to also-ran status among NASA’s mission planners.

  TPF-I, TPF-C, and TPF-O. After nearly a decade of heavy study, NASA, JPL, and other affiliated institutions had identified three broad technologies, each of which could produce images of habitable planets around other stars. All they lacked was a mandate and funding from NASA’s political taskmasters to select an architecture and move forward. The mandate came, to much rejoicing, in a single statement buried in the supporting literature for President George W. Bush’s 2004 vision for NASA, the vision that created the Constellation program to return astronauts to the Moon and onward to Mars.

  Traub wistfully recalled the spirit of the moment when I spoke with him in the summer of 2012, long after the early promise of TPF had become a frozen dream. Traub is a tall, quiet man, with kind blue eyes contrasted by a blond coif and goatee progressively whitening with age. He was in the midst of moving to a new office at JPL, where he still served as head of NASA’s Exoplanet Exploration Program. His desk was surrounded by blue filing boxes that, in aggregate, contained 285 linear feet of books, articles, and communications Traub had accumulated during his half-century scientific career. The bulk had accrued during the past seven years, during his JPL tenure, and much of it was related to the TPFs. He was cutting his stockpile down to 140 linear feet, filling trash bins with a good portion of the recent corpus on the telescopic search for extraterrestrial life. He pulled a folded piece of paper from one of the boxes and examined it through golden wire-rim glasses perched high on his nose. It was a memo from Charles Beichman, a Caltech astronomer who had been a key contributor to Elachi’s Road Map and who was, in years past, the project scientist for TPF initiatives.

  “This is from April of 2004, a year and two months before I came here,” Traub said. “It’s an unusually cheerful letter, considering Chas. He sent it to the members of the TPF science working group, of which I was a member.” He cleared his throat and began to read: “‘I want to inform you of exciting new developments for TPF. As part of the President’s new vision for NASA, the agency has been directed by the President to,’ quote, ‘ “conduct advanced telescope searches for Earth-like planets and habitable environments around other stars.” ’ End quote.” Traub sighed softly and dropped the letter on his desk. “We’ve been living off that statement in the President’s vision for NASA for more than eight years now.”

  Emboldened by Bush’s apparent support, NASA and JPL had come to an audacious decision: Rather than choose between the infrared TPF-I and the optical TPF-C, the agency would fly them both, and soon. NASA and JPL would build and launch TPF-C as soon as 2014, then work with the European Space Agency to build and launch TPF-I before 2020. Scientifically, the case for synergy was solid: spectroscopic observations in both the optical and infrared would allow a far more reliable determination of a planet’s habitability and possible biosphere. Beichman’s 2004 memo served as the unofficial announcement, explaining that this was “the opportunity to move TPF forward as part of the new NASA vision,” and that “in the estimation of NASA HQ and the project, the science, the technology, the political will, and the budgetary resources are in place to support this plan.”

  The planet hunters and the public were ecstatic, but many other astronomers were resentful. NASA had chosen to build not one but two expensive space telescopes devoted almost entirely to exoplanets, all without ever officially consulting the various high-level committees and study groups that tried to orchestrate national plans for space science. Building both TPFs, the critics argued, would leave no money for other more worthy priorities, such as clarifying the nature of dark energy, detecting gravitational waves, and observing active galactic nuclei in high-energy X-rays. In public, the pushback was muted, but privately it seethed. By the time Traub arrived at JPL in 2005, the howls of acrimony had already begun to pull the Laboratory’s lofty aspirations back down to Earth.

  “There is not a great deal of happiness among classical astronomers with planets,” Traub said with typical understatement. “Exoplanets are even worse. Planets generally seem to be frowned upon by the astronomers who only look at stars and galaxies. There are a lot of people with the attitude that it’s fine to study astrophysics, the Big Bang, the evolution of galaxies, and the evolution of dust disks around stars. But don’t ask whether the disks make plan
ets. Don’t dare wonder whether the planets make things that can hop and crawl around. Because that’s somehow beneath our dignity, to think about things that might have anything to do with subjects as complicated as biology and life.”

  When Traub came to JPL, he encountered a stoic acceptance that it would not be easy to muster sufficient community support to fly one TPF mission in the very near future, let alone two. “There was a sense that, well, we certainly weren’t going to be able to have two missions at the same time, and it wouldn’t be happening as fast as this letter said,” he explained. “But that was okay. We’d just work hard and get the science definition reports written, get the technology studies nailed down, and bring everyone onboard. It would take maybe a few extra years, but that was all. In retrospect I’ve learned, to my horror, that the science isn’t everything. In fact, it’s probably the last thing. People don’t support what they think is best for all of science, they support what directly benefits them. These days, what the astronomy community pursues is full employment for astronomers.”