The Large Millimeter Telescope in the Mexican state of Puebla

A few months ago I went to Cambridge, Mass. to check in with the Event Horizon Telescope crew and found Shep Doeleman, the project leader, fresh off the completion of a major purchase. He and his colleagues had just closed a deal on two hydrogen masers, among the most precise atomic clocks available. He displayed the weary pride of a guy who had just bought a new house. “We spent a half million dollars on clocks today,” he said. “Ask me what time it is. I dare you.”

Einstein taught us that time is flexible. That doesn’t mean that time is no big deal. If you’re going to take a picture of a black hole, you must become obsessed with precision timekeeping. We’ve already covered the basic design of the Event Horizon Telescope: radio dishes around the world simultaneously observe a black hole, time-stamp the data they collect, and then astronomers later correlate all that data to mimic the performance of a single, planet-size dish. To match up the data from the various telescopes, you have to know to within a fraction of a microsecond when, exactly, specific photons struck specific telescopes. And so, to participate in Event Horizon Telescope observations, a telescope must have on site an atomic clock. Atomic clocks aren’t necessarily standard radio-telescope equipment, though. One of the big tasks now facing the astronomers of the Event Horizon Telescopes is getting suitable clocks to those telescopes that don’t have them.

One of the clocks Doeleman procured that day in Cambridge was bound for Mexico’s Large Millimeter Telescope, a 50-meter dish that should function as a useful stepping stone between the EHT’s North and South American observatories. In late April, Doeleman, Patrick Owings, a technician from the maser manufacturer Microsemi, and I met up with that clock. The three of us had flown separately into Mexico City the day before. Today, we and several others would haul this quarter-million-dollar instrument into the Pueblan hinterlands, drive it to the top of a 15,000-foot extinct volcano, and install it in the LMT, the world’s largest millimeter-wave telescope.

A very expensive clock in a very fancy box

The hydrogen maser had arrived at the campus of the Instituto Nacional de Astrof?sica ?ptica y Electr?nica (INAOE) a few days earlier. Together with the University of Massachusetts, Amherst, INAOE runs the LMT. When we got to INAOE’s resort-like campus, the maser was waiting in the high bay, a hangar-size building on top of the highest hill around. The maser is a black metal box about the size of a gas-station ATM, but it came in a synthetic-wood crate speckled with customs forms, warnings (“DANGEROUS GOODS IN MACHINERY”), and tip-and-tilt indicators, little stick-on levels that rat out any baggage handlers who have tilted the crate more than 30 degrees.

I’m not sure how much the young forklift driver who lifted the maser into the transport truck had been told about his cargo, but this must have been the most closely scrutinized maneuver of his career. A half dozen people recorded the operation on their iPhones, as if he were handling the Ark of the Covenant. But astronomers had been talking about installing an atomic frequency standard in the LMT for years, so this was a big moment. It was a necessary step toward the Event Horizon Telescope’s ultimate goal of imaging Sagittarius A*, the black hole at the center of the Milky Way. It was also a delicate task: Drop the maser more than six inches and it becomes a useless 500-pound hulk of metal.

Like all clocks, a hydrogen maser generates a strictly periodic oscillation–this case, the 1420 megahertz tone produced by the hyperfine transition in hydrogen atoms. Here’s how Doeleman described it: “You get together a bunch of hydrogen atoms which are in the excited hyperfine state. … You collect these atoms in a bulb. Get all these atoms in the same energy state, and then when one of them emits they are all stimulated to emit at the same wavelength, just like a laser.” An antenna in or near the bulb detects that 1420 megahertz tone, amplifies it, and then uses it to control the output of a quartz oscillator. Over time, if the oscillations of the quartz and the hydrogen start to drift apart, the maser applies a voltage to correct the quartz oscillator. “On long time scales, the crystal is corrected to always come back to what the hydrogen is doing. That’s phase locked.”

The hydrogen process is so stable that the maser should drift by only about a second ever 100 million years. That stability is crucial when it comes time to combine data collected at different telescopes. “Typically we synchronize [all the telescopes] to within much better than a microsecond using GPS,” Doeleman says. “The real reason we need masers is that over the course of VLBI integration time we need very high stability. We’re using the maser to record the phase of the incoming waves that hit the dish. If the maser is not stable, then that translates directly into jitter of the phase of the recorded waves. Imagine you have a sine wave coming from the black hole. It’s received in Chile and California. If you have perfect phase recording, then when you correlate those together you get a big correlation. The problem is that when you record one that is jittering–changing frequency of phase slightly–when you go to correlate those you start to wash out your signal. Masers have to be stable enough so that that does not happen.”

The drive up Sierra Negra on a good day. The peak in the distance is Pico de Orizaba, Sierra Negra's sister, which at 18,500 feet is Mexico's highest mountain.

As our atomic-clock caravan left the INAOE campus, the greatest threat to the maser’s stability was the 80 miles of road ahead. Doeleman, Owings, a two-man film crew down from Mexico City and I piled into an SUV and followed the air-ride truck carrying the maser out of the campus gates. Before long we reached a potholed freeway and started crawling. Thirty kilometers per hour. Forty kilometers per hour. We probably should have gotten a ticket for driving too slowly. Someone did the math and figured it would take about eight hours to get to the summit at this rate, so as soon as the road improved we accelerated to a respectable speed. After a couple of hours, we turned off the highway onto a local two-lane road. Sierra Negra was a brown hulk in the distance; the LMT’s giant silver dish sat on top like a metallic cherry.

The road to the summit was a glorified 4X4 trail. We rumbled up this winding lane for what seemed like hours, through stands of pine and herds of sheep, past enormous maguey cacti and farmers hauling bundles of wood on the backs of burros. Just above the tree line, snowflakes began to fall. Pretty soon, snow was lashing against the windows. People have gotten stuck at the summit when enough snow or fog descended on the mountain, so the change in the weather was reason to hurry.

Just above the tree line, the snow started falling.

We clambered on as quickly as we could, and after about two hours of off-roading, we arrived at the summit. As soon as we rolled into the gravel lot at the base of the telescope, workers began roping the maser to a crane that looked like it could have lifted a battleship. The crane operator hoisted the crate above the truck, and the driver backed the truck out of the way. “We have liftoff!” Doeleman said.

Maybe it was just the altitude, but no one seemed to breathe well as the crane operator lowered the maser toward the ground. The guy was good, though, and after a minute or so he set the clock down as if on a feather pillow. The workers swarmed around the crate with a dolly and sheets of plywood to smooth the ride, and within minutes they had rolled the maser into the telescope’s cargo bay.

You could argue that the industrial crane was overkill, but it worked.

It took two and a half days for the maser to warm up. In the meantime, Jason SooHoo of MIT’s Haystack Observatory; Jonathan Leon-Tavares, an astronomer at INAOE; Gisela Ortiz, a grad student at the Centro de Radioastronom?a y Astrof?sica (CRyA) of the Universidad Nacional Aut?noma de M?xico (UNAM); and Doeleman set about wiring the maser into LMT’s nervous system. They strung cable from floor to floor, rummaged through bins of gold-plated connectors, puzzled over oscilloscope readings. Finally, one morning the green flashing light on the maser’s control panel stopped flashing. Large Millimeter Telescope’s maser “locked,” meaning the vibration of the quartz was synchronized to the frequency of the hydrogen atoms in the machine’s copper cavity.

The hydrogen maser in its new home inside the LMT's apex cone.

By then, Gopal Narayanan of the University of Massachusetts at Amherst, who has been working on the LMT since it was little more than a dream, had arrived. Together with Leon-Tavares and Ortiz, he would begin conducting observations in concert with the Very Long Baseline Array (VLBA) back in the U.S. The observations had a few goals. One was to shake down the LMT with the new clock installed. Another was to begin the process of measuring the distance to the galactic center using parallax.

Before the EHT can take a picture of Sagittarius A*, the LMT needs a receiver capable of observing in the 1 mm range–the wavelength of light that should yield an image of the event horizon of the black hole. That’s a project for either later this year or early next. More, as always, as things develop.