Russell got interested in ion engines in 1992, when he met Scott Benson, an engineer at NASA’s Lewis Research Center in Cleveland (now the Glenn Research Center), who had recently begun experimenting with ion propulsion. In fact, NASA had explored the technology as far back as the 1960s but lost interest as the agency’s focus shifted to the space shuttle; ion engines had been developed only to make minor adjustments in the paths of Earth-orbiting satellites. When NASA started its New Millennium program in the 1990s to develop innovative spaceflight technologies, research on ion thrusters began again, this time in earnest. “One of the features of ion propulsion is that it essentially allows you to fly on a smaller launch vehicle, at lower cost, to destinations that would require a larger vehicle with chemical propulsion,” says Benson. At first Russell’s instinct was to use ion propulsion to go back to the moon. As a postgraduate researcher on the Apollo program, he had developed instrumentation for the command module that measured the lunar magnetic field. With Benson he spent two years on a sequel of sorts, a lunar orbiter that used an ion engine, but the idea was passed over. He next worked up a proposal to go to Vesta but again failed to win backing from NASA. Russell suspects that ion propulsion was deemed too risky—it had never been used on a space probe. He tried to be philosophical: “Each time you lose,” he says, “you learn something.”

The challenge was to turn an engine intended for occasional use on a satellite into a trustworthy interplanetary thruster. Deep Space 1, an engineering test mission launched by NASA in 1998, demonstrated that an ion engine could be used to move around the solar system. “That excited people,” Russell says. “That was a winner.” In December 2001, NASA gave Dawn a green light.

“Dawn really reflects a big departure from what we used to do in planetary exploration,” Russell adds. “The way we’re probing these bodies is very cost-effective.” NASA considered the cost of exploring both Vesta and Ceres with chemical rockets and concluded that it would have required two missions at $750 million each, as opposed to Dawn’s sub-$500 million price tag. “We’re saving a billion dollars compared with what it would have cost us to do it any other way,” Russell says.

In Russell’s proposal, Dawn used the same basic engine design as Deep Space 1 but needed a larger xenon fuel tank and other changes to ensure the system would survive its eight-year mission. Making these alterations nearly doomed the project, forcing it way over its $373 million budget. “The design parameters of Dawn were ambitious,” says Tom Jones, a former shuttle astronaut and now a consultant to NASA. “No probe had ever gone to one body, slowed down and achieved orbit, and then turned around and gone to a second body. That puts a lot of stress on an engine, and you have to make it reliable.”

By October 2005 Dawn was $73 million over budget. That, combined with concerns over the fuel tank’s design and the mission’s management, prompted NASA to pull the plug, canceling the project altogether in March 2006. NASA was also scrambling for funds to cover President George W. Bush’s moon program. Despite having already spent hundreds of millions of dollars, administrators may have been willing to scrap Dawn to avoid spending any more. Russell insists the project’s technical troubles were nothing out of the ordinary for such a complicated mission, and that NASA’s decision to cancel the project was foolhardy. “I don’t have any logical reason for why they did that,” he says. “To throw away the roughly $300 million that had been invested was crazy. Why not just finish off the project and get a return on this investment?” Fortunately, NASA’s chief administrator, Michael Griffin, allowed an appeal, and the mission was reinstated.

Now journeying outward, Dawn is following a flight plan unlike that of a conventional spacecraft. To set course for Vesta, a chemical rocket would burn for a few minutes near Earth, putting it on a path that intersected Vesta’s orbit, and then burn again to enter that orbit. Dawn’s ion engine, by contrast, has to accelerate the spacecraft continuously for months on end, spiraling outward until its trajectory matches Vesta’s orbit. The thrust from each of Dawn’s three ion engines is minuscule, a force equivalent to that of the weight of a piece of paper resting on the palm of your hand. But an engine will be firing during 90 percent of the trip, building up a speed as high as that attained by any chemical rocket.

A Mars flyby in February 2009 will help things along, giving Dawn a gravitational kick. In August 2011 it will begin slowing down as it approaches and then settles into orbit about Vesta. The craft will fire up its engines again in May 2012 to set course for Ceres. It will arrive in February 2015, once again slowing down to enter orbit and snap photos.

Taking snapshots will be a major part of its mission, because Dawn is not exactly a flying lab bristling with instruments. It has only three—part of the trade-off necessary to keep its weight and cost under control. A camera will create detailed maps of the two asteroids, with a resolution of about 225 feet for Vesta and 400 feet for Ceres. A spectrometer will measure the light absorbed by the asteroids’ surfaces, which will tell much about their composition. And a gamma ray and neutron detector will measure cosmic radiation bouncing off the surface of the asteroids. (It will be able to scan several yards below Ceres’ surface, searching for ice or liquid water.) In addition, variations in Dawn’s radio signal will be monitored to provide information about the gravitational pull—hence the internal structure—of the asteroids.

During the years of proposals and rejections, Russell had plenty of time to think about what Dawn might find when it finally reached its mystery worlds. His interests naturally led him to McCord, another asteroid hunter, who had gotten into the business indirectly. At Caltech in the 1960s, McCord helped develop instruments for remote spectrometry—analyzing the light coming off planets and stars. The first thing McCord and his colleagues trained their new instruments on was the moon, but soon they began measuring everything in sight. They worked their way through the planets and down to the asteroids, and eventually Vesta found itself in their crosshairs.

“It doesn’t sound like exploration, but that’s the way it really works,” McCord says. “You’ve got an instrument and you just go out and do everything you can with it. New data are power in science, and if you can measure something 10 times better than somebody else can, you’re going to learn a lot of exciting things.”

McCord didn’t get around to looking at Vesta until the early 1970s, and even then he didn’t give it much thought until his team got around to processing the data. “I was in the lab one day and one of the guys pulled the Vesta spectrum out of the computer, which we had observed a week or a month before,” he says. “And my God, it had one of the most beautiful absorption features you ever saw on a planetary object.” The data indicated that Vesta was basaltic, which suggested that Vesta’s rocks had been heated to melting at some point and then cooled. The discovery also established that the HED meteorites and Vesta shared the same composition.

McCord and his researchers also looked at Ceres but didn’t get far. Ceres was darker and murkier, and it didn’t have the clearly identifiable spectrum of Vesta. McCord’s grad students set to work on the data and came up with some preliminary findings: Ceres was a carbonaceous chondrite (a type of asteroid composed of water locked in minerals and carbon-based materials), and it had not been thermally altered. In other words, it had never melted and cooled, as Vesta had. This posed more questions than it answered. How did a large asteroid evolve and retain significant amounts of water? Nobody had any theories to explain it, and the researchers dropped the subject.

When the Dawn mission was approved, much of the focus was on Vesta. “You’re human, so you’re generally interested in things you know about,” Russell admits. “If you don’t have any information, you don’t have that thing to grab your interest.” That attitude began to shift in 2002, when McCord took a sabbatical to Nantes, on France’s west coast. “I got to thinking about Ceres, and I learned that the people who had been doing the most careful orbit and mass determinations were at the University of Bordeaux, a two-hour drive to the south.” McCord went down and learned that researchers there had been able to make accurate estimates of Ceres’ density. Pure water has a density of 1 (measured in grams per cubic centimeter). A conventional dry asteroid, made of silicates with some iron mixed in, would have a density of 3 or 3.5; Vesta’s is thought to be in this range. Ceres has a density only slightly higher than 2. That means there is a lot of water in the mix.

McCord found the work of Christophe Sotin and his graduate students at the University of Nantes even more intriguing. Sotin had developed a computer model of how Saturn’s biggest moon, Titan, could have formed without its liquids boiling off. Although Titan is chemically very different from Ceres, it too contains a lot of water. Perhaps, thought McCord, some version of Sotin’s model could explain how Ceres could have formed with its water intact. “We began to see that it was easy for Ceres in the early, early history to have created a liquid ocean,” McCord says.

Here’s how the theory goes: In the early solar system, dust particles glommed together to form bigger dust particles, which formed pebbles, then rocks, and so forth, until they combined into an object up to several hundred miles in diameter. The original dust particles were made largely of silicates mixed with other materials, including water and aluminum 26, a radioactive isotope with a half-life of about 700,000 years. That’s just long enough to make a big difference in how an asteroid evolves. Vesta and most other asteroids, the theory goes, accreted quickly and accumulated a lot of aluminum 26 that had not yet decayed. The aluminum 26 produced so much heat inside the asteroid that any water evaporated into space. Ceres, by contrast, accreted more slowly, so by the time it formed, the aluminum 26 had already mostly burned itself out. As a result, Ceres retained most of its water—and a memory of the solar system’s original composition.

These findings ignited McCord’s interest in Ceres, to the point where “I kept demanding we go to Ceres first,” he says. Russell sympathizes. “If we had to pick which was the most interesting, Ceres or Vesta, it’s not clear which one would win,” he says.

The argument is moot: Vesta is closer than Ceres, and therefore it must be Dawn’s first stop. But Ceres may make the bigger headlines. Vesta seems like Mercury or the moon, writ on a smaller scale. Ceres is unique. Imaging of the surface may reveal whether there is indeed an ocean beneath an icy crust. Observing the surface should allow scientists to glean some idea of how the interior behaves—if there’s volcanic activity that could provide the heat to sustain life, for instance. Dawn’s spectrometer will be able to detect the presence of organic molecules.

Unfortunately, Dawn isn’t equipped to search for past or (dare we dream?) present life on Ceres. That would require penetrating the surface and taking and analyzing samples. “To detect life, you need a pretty sophisticated lab on the surface or in the interior or wherever the environment is,” McCord says. “That’s technically a major challenge and virtually impossible—nobody’s willing to spend the amount of money to do that.”

For now, at least. After Dawn’s visit, attitudes might change.