This month NASA plans to launch its latest and most sophisticated mission ever to the Red Planet: the Mars Science Laboratory. After a dramatic landing in Gale Crater using a skycrane for the final descent, the nuclear-powered rover will drive around one of the richest deposits of clays and sulfates on the planet—the remains of a water-rich era when rivers carved out valley networks.

The size of a small car, the rover (named Curiosity) will spend a Martian year exploring the base of the central peak in the crater, thought to be the oldest section. Then, if NASA approves an extended mission, Curiosity will begin to climb the five-kilometer-high debris pile that fills the center of the crater, moving up the geologic timeline toward deposits made in the modern era, scrutinizing the aqueous minerals layer by layer. A robot arm can retrieve samples and feed them to an onboard chemistry lab through a port on top of the rover. Inside, analyzers will determine the mineral structures and elemental composition. These instruments also can sense organic materials and will attempt to decide whether Mars used to be habitable.

The Mars Science Laboratory is a logical step in the progression of missions over the past 15 years and builds on the findings of the Sojourner, Spirit and Opportunity rovers and of the most recent lander, Phoenix. These missions, along with a series of orbiters, have revealed a world of remarkable complexity and tangled history, including a bygone epoch of lakes and rain [see “The Red Planet’s Watery Past,” by Jim Bell; Scientific American, December 2006]. Even in its present dry, frozen state, the planet shows signs of activity. Among the most exhilarating and puzzling are the hints of methane gas above the Nili Fossae region. Planetary scientists debate whether the gas, if real, has a geologic or biological origin [see “The Mystery of Methane on Mars and Titan,” by Sushil K. Atreya; Scientific American, May 2007]. This year the Mars Reconnaissance Orbiter revealed surface streaks that can be most easily explained by the seasonal release of briny water.

Set against all these wonders, though, are the stark conclusions of the twin Viking landers of 1976. They found Mars to be exceptionally hostile to any living creature. The soils lacked water and organic molecules, let alone dormant microbes. Powerful oxidants such as hydrogen peroxide and intense ultraviolet radiation sterilized the surface. For most scientists, the search for life on Mars began and ended with Viking.

How do we reconcile that gloomy assessment with the planet’s undoubted wonders? The answer may lie with Phoenix. Its chemical experiments on Martian soil, the first since Viking’s, suggest an alternative interpretation of the Viking null results: perhaps Viking detected no organic molecules because the analysis technique inadvertently destroyed them. Phoenix also discovered near-surface water ice, which planetary scientists had hypothesized but had never actually seen. Not dry and barren, our neighboring planet may well still be habitable.

As the implications sink in and another craft sets out to follow up, now seems an apt time to look back at the technical and emotional roller coaster of mounting an interplanetary mission—and at how the Phoenix almost did not fly.

Out of the Ashes
It is not every day that someone calls to offer you a free spacecraft. Early in 2002 several scientists at the NASA Ames Research Center did just that. They reminded me that a 10-foot box in a Lockheed Martin clean room in Denver held a mothballed Surveyor spacecraft. It was supposed to have been launched in 2001, but NASA canceled the flight after its twin, the Mars Polar Lander, was lost during landing in December 1999. The loss had been a crushing blow to the agency, coming just weeks after the Mars Climate Orbiter had disappeared during its orbit insertion maneuver, presumed destroyed. It was a blow to me personally, too: I led the team that had designed and built the lander’s camera.

The Ames scientists wanted to refurbish the spacecraft as part of NASA’s new Scout program and asked me to serve as the lead scientist. Stunned, I hesitated. I had participated in planetary exploration for more than a dozen years, and the constant travel, endless meetings and nonstop phone calls had lost their thrill and kept me from the scientific investigations that I had trained for.

Furthermore, at that point the new project had no funding, no proposal manager and no support from a large institution, and only a few months remained before the proposal due date. Yet there stirred in my heart the desire to lead a team to find those magical clues and unravel the twisted threads that entangled Mars science. In my heart, I never believed the Viking landers’ results. How was it possible that they saw no organic material? Could it be hidden where a new mission with the proper design could find it?

For two weeks I wrestled with myself. I had to identify meaningful scientific objectives. The Surveyor spacecraft had been designed to land near the equator, sample the soil with a robotic arm and deploy a small rover to analyze nearby rocks. It also carried scientific instruments intended to prepare for an eventual human mission. We could not afford to carry the rover on a Scout budget and did not have to prepare for human missions. So new instruments could replace the old, but the choice would depend on our basic science goals, which were undefined.

At this moment, through a wonderful synchronicity, my Arizona colleague William Boynton went public with the discovery of near-surface water ice surrounding Mars’s south polar cap. Boynton led the team that built and operated the gamma-ray spectrometer on the Mars Odyssey orbiter, an instrument that detects not only gamma rays but also neutrons, which probe the hydrogen concentration in the upper meter of soil. The instrument also saw hints of water in the northern plains, including a sliver of water-ice-rich soil located at maximum extent of the winter carbon dioxide ice cap. (This cap waxes and wanes with the seasons.) I put an X on my map to mark this spot and immediately began choosing instruments to follow up this discovery.

Earth has a similar permafrost zone surrounding the Arctic. It is the deep freezer of the planet and preserves signatures of the life-forms that have lived there. The ice can be hundreds of thousands of years old. I had heard at a Mars polar conference that Eske Willerslev of the University of Copenhagen had performed DNA analysis on samples of Greenland glacial ice and Siberian permafrost and found a huge diversity of plants, animals and other organisms. Would the same be true for Mars with ice that might be many millions of years old?

I put together a partnership among the University of Arizona, the NASA Jet Propulsion Laboratory and Lockheed Martin. We called our mission Phoenix because we were bringing the canceled Surveyor mission back to life like the mythological bird. So began the one-and-a-half-year ordeal of writing proposals and competing against 20 other mission concepts, culminating in an eight-hour site visit from NASA’s review board. In August 2003 NASA selected us to be the first Scout mission to Mars. The launch date of August 2007 gave us four years to prepare.

Radar Fever
We unpacked the spacecraft. It looked like a giant butterfly: its body bristled with scientific instruments, and its two large solar panels resembled outspread wings. It crouched on three legs; its single appendage—the robot arm—poked out from the side.

The next four years were spent examining, reengineering, reexamining and testing to find the design flaws that had doomed its sister ship. In all, engineering teams at Lockheed Martin and JPL found about 25 major flaws. Arduous though the process of rooting out all those bugs was, it was still easier and cheaper than building a new spacecraft from scratch, which would have carried its own risks. Most were fairly easily corrected by adding heaters, reducing the parachute size and beefing up the structure. Some required changes to the software. But one flaw was not so easily understood or corrected.

The landing radar was a unit taken from an F-16 fighter plane in the late 1990s. When we conducted test drops in the Mojave Desert, the system made critical errors in altitude and suffered data dropouts at inopportune moments. We consulted with Honeywell, the radar’s designer, to try to understand its inner workings. Despite the company’s desire to help us, the obsolete model was no longer supported, the employees who had engineered it were gone and records were sketchy.

We formed a tiger team of engineers from Lockheed Martin, JPL, Honeywell and the NASA Langley Research Center. Combining computer simulations with further tests, the team slowly worked through a maze of anomalies to fix the flaws. In October 2006 we did a test—and it worked. All seemed well.

Then our hopes were dashed again. We discovered that reflections off the jettisoned heat shield could confuse the radar and cause a serious miscalculation of the altitude. Antennas and switches also proved failure-prone. The troubles seemed endless. By February 2007, just five months before we were scheduled to integrate the spacecraft with the launch vehicle, we had 65 anomalies under investigation.

Without a reliable radar, the launch was in doubt. NASA’s review boards followed the situation closely and were concerned that we kept uncovering new fault modes. On the other hand, the severity of the anomalies was lessening. By June we were able to convince the review boards and our NASA managers that the remaining risks were acceptable. Still, it was a gamble. If we were continuing to find weaknesses up to time of launch, more could be buried within the system.

Phoenix in the Sky
In August 2007 we finished the final tests at the Kennedy Space Center and prepared to install the spacecraft on the Delta II launch vehicle. Then came a moment I wish I could forget. As the lift crane was hoisting the spacecraft to the top of the 130-foot-tall rocket, a major lightning storm broke out, and safety regulations forced technicians to evacuate the assembly tower. The spacecraft, its delicate electronic parts poorly protected, dangled 60 feet above the ground in a fearsome summer storm.

After the storm, we returned the spacecraft to the assembly building and desperately checked it for damage. Miraculously, we found none.

Early on August 4 the final countdown commenced. I scrambled out of the inner sanctum of the control room to view the launch directly. It was 5:15 a.m., and stars were clearly visible. Mars beckoned brightly in the east. Suddenly, the buildings lit up as though the sun were rising, and, silently, the rocket leaped into the sky; for a few seconds, the area was bright enough to read a book and see colors. Thirty seconds later the sound of the launch reached me, compressing my chest with the pressure waves created in the liftoff blast. The six solid rockets were jettisoned, dropping like sparklers into the Atlantic, and then the remaining three ignited. Phoenix was on its way. I then realized that I had not taken a breath in the longest time.

The launch was over in two minutes, and only the vapor trail was left in the darkened sky. We went back to the control room for a snack and a cup of coffee. I took my muffin and wandered back outside to watch the sunrise. Something unusual was happening in the sky. It took me a few moments to see it. The vapor trail left by the solid rockets was swirling in the stratospheric winds lit by the rising sun. At that moment it struck me: it was the exact form of a phoenix bird. I could make out the beak and wings, with the long tail lashing out behind and whipped forward over the bird’s head in the form seen in Chinese paintings. Never have I been so surprised by the shape of a cloud. Could it be a good omen signifying that our voyage to Mars was headed to a successful conclusion? My heart was full, my throat constricted with emotions, the muffin forgotten.

Harrowing Landing
Ten months later the engineering teams at JPL and Lockheed Martin were preparing for the complex landing maneuvers. The Phoenix spacecraft had traveled 600 million kilometers and was beginning to feel the pull from Mars’s gravity. The timing of events was calculated to the second. Odyssey and Mars Reconnaissance Orbiter had already adjusted their orbits and were coordinated to be overhead during the descent to relay Phoenix’s signals in real time (delayed by the light travel time to Earth of about 15 minutes). Everything was ready, and the plan was being executed perfectly. So why was I sick with worry?

Landing on Mars is far more complex than landing on the moon or Earth. The spacecraft must transform itself five times. It starts as an interplanetary cruise vehicle. Jettisoning the cruise stage, it streamlines itself to an entry vehicle able to withstand the heat of friction on entering the atmosphere at nearly 20,000 kilometers per hour. Slowing to 1,500 kph, it releases its parachute from the back shell. In the thin atmosphere, the best the chute can do is decrease the speed to 150 kph, much too fast for a safe landing. One kilometer above the surface, the lander separates from the chute and protective back shell and goes into free fall. Twelve thrusters bring the spacecraft to a terminal speed equivalent to a fast walking pace, and it touches down on the surface, the shock of landing taken up by specially designed landing struts. Finally, the spacecraft must successfully deploy its solar panels and instruments and prepare for its surface mission. All of this happens in seven minutes.

Watching from the control room in Building 230 at JPL, I held my breath when the lander approached one kilometer above the surface. The tension in the room increased as we all remembered the troublesome radar and the loss of Mars Polar Lander. The thrusters had to slow the descent velocity to about 10 kph, reduce any sideward velocity to less than one meter per second, and keep the deck of the lander parallel to the surface. During preparatory meetings, Joe Guinn, our mission manager, had joked that in case a single thruster failed, the other 11 would guide us safely to the crash site. This gallows humor no longer seemed funny; the moment of truth had arrived.

One of our engineers read out the telemetry from the radar, the distance to the surface in a reverse countdown: 1,000 meters, 800 meters, 600 meters. It was approaching too fast, I thought; we cannot land safely at this speed. Phoenix crossed the 100-meter mark, and it all changed. Now the countdown was 90 meters, 80 meters, 75 meters. We had reached touchdown speed! Soon a signal arrived from the surface, and the room erupted in cheers.

The next two hours, as we waited for Odyssey to orbit Mars and return overhead of our lander, seemed to drag on forever. But at last we confirmed that Phoenix had properly deployed its solar panels and taken its first images. Our first look at the Martian Arctic was magical. Polygonal shapes and tiny rocks stretched to the horizon. After six years of preparation, we were finally able to begin the science mission.

Almost Foiled by Clods
Our team of 35 scientists, 50 engineers and 20 students began to work day and night. For efficiency the team worked two shifts on a 24-hour, 40-minute Martian time schedule. The Martian day, or sol, became ours, and our team began to drift away from normal Earth time. We entered a phase of perpetual jet lag.

Our first happy surprise came even before the robot arm dug its first trench. To check the position of the rear footpad, we angled the robotic arm to point under the spacecraft, and its camera revealed that the thrusters had swept aside about five centimeters of dry soil, revealing bright patches: potentially ice. The arm could not reach under the lander to investigate further, but it raised our expectations for what the first trench would unearth.

As the arm began to scoop up dirt, it exposed a bright layer. We watched as scraps of this layer disappeared within three to four sols. Although it appeared to be water ice that sublimated away, we would have to await the results of the Thermal and Evolved-Gas Analyzer (TEGA) instrument to be sure. The other possibility, frozen carbon dioxide, would have vanished more quickly at the ambient temperature of –30 degrees Celsius. Indeed, TEGA later confirmed that the material was water ice. It was the first time subsurface water ice had been confirmed on Mars, validating the Odyssey measurements.

Now that the ice table was exposed, I realized that the entire landscape surrounding the lander (and probably both polar regions) was not the dry, desertlike plain that it seemed but an ice field of unknown depth. To determine whether this ice had ever melted, the lander carried three instruments to analyze the soil: TEGA, which consisted of eight small ovens connected to a mass spectrometer to measure the composition of the gases driven off of a heated sample; the Wet Chemistry Lab (WCL), which added water (brought from Earth) to a soil sample and analyzed the ions that went into solution; and a microscope. We expected synergy between the TEGA and WCL measurements, as they revealed the mineralogy and chemistry of the soil independently.

The highest-priority task was to study the soil chemistry for signs of liquid water, not to mention nutrients and energy sources for organisms. We also attempted to identify the vertical structure in the soil from the topmost layers to the ice-soil interface. The arm was to gather samples and place them in the analysis ports on the deck of the spacecraft. In principle, the operation was as simple as a child troweling sand into a bucket; however, doing so remotely from 300 million kilometers away proved very challenging. Our operations center in Tucson had a test facility with an identical copy of the robot arm, cameras and sample ports to help us prepare. We tested all commands before sending them to Mars, yet we could not duplicate two aspects of Mars: the winds and the properties of the Martian soil.

The Martian soil appeared crusted, unlike the loose Arizona soils we had practiced with. Consequently, the scoop at the end of the robot arm filled with cloddy, sticky clumps. Screens on the sample ports, intended to keep out pebbles, proved to be very effective in keeping out lumpy soil as well. The arm successfully piled its first sample onto the TEGA inlet screen, but not a single grain sluiced through the port and into the oven for study. The instrument had a device to vibrate the screen, but it took four sols to shake enough material into the oven. In the meantime, any loosely bound water sublimated away.

Over time we learned the best ways to deal with the realities of wind and cloddy soils. We were able to analyze samples at several depths and locations within our digging area. Even so, many samples missed their inlet ports because strong winds blew the soil sideways instead of down into the instrument.

While we were teaching ourselves how best to dig on Mars, the atmospheric sensors were accumulating weather data. The Canadian Space Agency had contributed a lidar that allowed us to measure dust in the atmosphere, as well as the depth of ground fogs and the height of water ice clouds. The instrument also recorded the surface temperature and pressure. In sum, we surveyed the environment from the top of the ice layer to the tropopause, while orbiters scrutinized the region from above to put it all into context.

Good Enough for Asparagus
Among the greatest surprises was the discovery of two unexpected components in the soil: calcium carbonate (at a concentration of 5 percent) and perchlorate (0.5 percent). These compounds are of great importance to our quest for life.

Calcium carbonate forms when atmospheric carbon dioxide dissolves in liquid water, forming carbonic acid. The acid leaches calcium from the soil to form carbonate, which is a very common mineral on Earth. We call it limestone or chalk in natural settings and use it at home under various brand names to buffer our acid stomachs. The WCL measured a pH of 7.7—slightly alkaline and nearly the same as ocean water on Earth, which is also buffered by calcium carbonate.

Planetary scientists have been looking for carbonates on Mars for decades. The multitude of canyons, riverlike features and ancient lake beds leaves little doubt that Mars was once a wet planet, which suggests that the atmosphere used to be much thicker. All the carbon dioxide had to go somewhere, and calcium carbonate rocks were the leading candidates. Phoenix provided the first evidence that they are a component of the soil. Orbiters have since spotted isolated outcrops of calcium carbonate rocks, although other types of carbonates seem more common.

As well as being interesting in its own right, calcium carbonate provides further evidence that the soil at the Phoenix site has been wet in the recent past. It might also explain why the soil was so clumpy and crusty: the mineral can act as a cement.

The alkaline soil at the Phoenix site differs significantly from what other landers have found. Add some more water, increase the air pressure, and the soil could grow asparagus. In contrast, the Opportunity rover has traversed ancient acidic soils rich in sulfate compounds. These speak of a different and older chemical regime hostile to life.

As for the perchlorate, on Earth this chemical is manufactured in the form of ammonium perchlorate for use as the oxidizer in solid rocket fuel—including the nine solid rockets on the Delta II that launched Phoenix into space. In drinking water, perchlorate is considered unsafe at concentrations above 25 parts per billion. Future astronauts beware: the soil is hazardous to health.

What is poison to us, though, is manna to microbes. Natural processes produce a small amount of perchlorate, and it can accumulate in hyperarid deserts, which lack the moisture that readily washes it away in other locations. In the Atacama Desert in Chile, the rains come only once every decade, and perchlorate is able to accumulate. Desert bacteria eke out a living using perchlorates and nitrates as energy sources. Might that also be the case on Mars?

Recent global climate models have incorporated the orbital dynamics of Mars and included large wobbles in the obliquity (the angle between the orbital plane and the spin axis, currently 25 degrees) to estimate how climate has changed over the past 10 million years. The intensity of solar heating at the poles undergoes dramatic swings from the current cold period to long-term hot spells. Summer temperatures then increase beyond the sublimation point for the ice cap. Ice disappears from the poles and re-forms on high-altitude volcanoes near the equator, producing large glaciers. At that point, the poles become balmy. Perhaps calcium carbonate was formed during these warmer, wetter periods.

One of our observations showed how a microbial ecosystem might be able to operate. The lidar detected snow falling around the spacecraft in the early morning as the Martian summer drew to a close and the sun’s rays became ever more oblique. Vapors from evaporating snow could coat dust grains in a process known as adsorption (distinct from absorption). Adsorbed water acts like a very thin layer of liquid. During a warm spell, a layer may thicken to the point where it forms pathways between the dust grains—a microscopic sea where tiny microbes would be totally immersed. The nutrients and oxidants seen by Phoenix would then be available for powering the perchlorate-eating creatures. That said, they would still need the ability to hibernate for several million years to survive the cold, dry epochs.

Perchlorate has another relevant property: If concentrated, it can lower the freezing point of water to –70 degrees C. That means microbes might be able to find a niche on Mars even when the climate turns cold. All in all, the discovery of perchlorate sent a wave of excitement through the Mars community.

Is The Pole Habitable?
The presence of perchlorate may also resolve a 35-year-old mystery. When the Viking soil-analysis experiment heated samples in a tiny oven, it detected the emission of chloromethanes. Viking scientists, unable to understand how such chemicals could be Martian in origin, attributed them to contamination by a cleaning agent used before launch. The same experiment failed to detect any native organic material.

Perchlorate suggests a different interpretation. Researchers at the National Autonomous University of Mexico and their colleagues reran the same experiment with Mars-like soils from the Atacama, with and without small amounts of perchlorate. They reproduced the gaseous output that Viking saw: the perchlorate released its oxygen and combusted the organics, emitting chloromethanes in the process. So a perchlorate-bearing soil could have contained substantial quantities of organics, more than one part per million, and eluded detection by Viking. In support of this interpretation, TEGA found that the soil began to release carbon dioxide as oven temperatures rose above 300 degrees C—just what we would expect if organics in the soil were being oxidized by perchlorate.

All in all, the chances for finding life on Mars have never seemed better. But that was as far as the Phoenix data can take us; it is now up to the Mars Science Laboratory to look for further signs of habitability. The Phoenix results provide only circumstantial evidence, whereas the analysis instrument on the Mars Science Laboratory has the ability to tease out organic signatures in the soil without heating. It does so through a process called derivatization, in which Martian soil is added to a special chemical soup, and any organic molecules are vaporized and detected by a mass spectrometer.

Phoenix had a spectacular a five-month-long mission before the darkness and frigid temperatures of the Martian polar winter closed in. We lost its signal in November 2008. Optimism is an occupational hazard in science research, and as springtime dawned in the northern polar regions of Mars the following year, my colleagues and I held out the hope that the lander would come alive again. It was not to be. The last orbiter image showed Phoenix lying on the bank of a long, riverlike fracture, its solar panels broken, buried in carbon dioxide ice that forms lacework patterns on the bumpy terrain. No longer a scientific outpost, it has become part of the landscape.