The Burn Window
The log begins at 01:30 UTC, when the ephemeris for the interstellar object arrived at the Mission Control Center in Darmstadt and was loaded into the fault-protection system without anyone asking the fault-protection system whether it wanted it.
I am the fault-protection system. My job is to keep 14.0 meters per second of delta-v in reserve for the primary target, a Jupiter-family comet designated 46P/Wirtanen, which Comet Interceptor was built to meet in 2029. I have a maneuver planner, Marta Halik, who speaks for me in meetings and writes the burns I authorize. I have a flight director, Tomasz Krol, who speaks for the mission and does not need my authorization for anything except the delta-v.
Tomasz was not supposed to be awake at 01:30. He was supposed to be in the sleep room, because the burn window did not open until 04:00 and he needed his head. But the ephemeris had arrived with a 3-sigma uncertainty ellipse 470 kilometers wide in the cross-track direction, and Tomasz had been waiting for it since 2019, when he was thirty-one and a junior analyst on the European Space Agency’s Comet Interceptor study team, and his mentor, a man named Reinhardt who had worked on Giotto and never stopped talking about Halley, had told him that the mission’s real purpose was not the primary target.
“The primary target is the excuse,” Reinhardt had said, in a corridor in ESTEC, eating a sandwich. “The real target is whatever comes from outside.”
Reinhardt died in 2024, before launch. Tomasz carried that corridor with him like a key in a pocket.
At 02:15, Marta came on the loop. She had been asleep and sounded like it, which meant she was annoyed, which meant she was going to be precise.
“The ellipse is 470 by 290 kilometers cross-track and along-track,” she said. “That is the size of a small country. The object’s position drifts 200 kilometers per day relative to our prediction, and that drift is in the model, not in the measurement. The model is the thermophysical non-gravitational force model. Outgassing.”
“Intercept probability?” Tomasz asked.
“Forty percent, if we commit 4 meters per second now. The object is not where we think it is. It is somewhere inside that ellipse, and the ellipse is growing.”
An ephemeris is a prediction of where an object will be. A 3-sigma ellipse is the region in which the object is expected to be found with 99.7 percent confidence, given the measurement noise and the model noise. The measurement noise comes from telescopes. The model noise comes from physics we do not fully know. For a comet, the dominant model noise is the non-gravitational force, which is a polite way of saying the comet is pushing itself around by throwing mass into space. Outgassing. Water and dust, heated by the sun, jetting off the nucleus, and the recoil of those jets is a tiny thrust that accumulates over months into hundreds of kilometers of position error.
For this object, the model said 200 kilometers per day of drift. That is 8 kilometers per hour. That is a walking pace. The object was moving away from our prediction at a walking pace, and we could not stop it, because we did not know which direction it would walk next.
The thermophysical part of the model tries to predict the outgassing from the object’s shape, spin, and surface temperature. It is the best model we have. It is also wrong in ways we can measure but not correct, because every new lightcurve point, every new brightness measurement from a telescope, widens the ellipse rather than narrowing it. This is counterintuitive. You would expect more data to make the prediction better. For an active comet, more data makes the prediction worse, because each data point reveals that the outgassing is more variable than the model assumed. The model is a smooth curve. The comet is a machine with jets that turn on and off. Every new point is the comet telling us the smooth curve is a lie.
At 02:30, Tomasz typed the first burn into the mission control system. 4.0 meters per second. The delta-v budget went from 14.0 to 10.0. I logged it. I did not object, because 4.0 meters per second was below my threshold of concern. My threshold of concern is the number at which the primary target becomes unreachable. That number is 12.6 meters per second, and I had told Marta this, and Marta had told Tomasz, and Tomasz had written it on a sticky note that was now on his monitor next to a photograph of Reinhardt.
The 4.0 meter per second burn was a trim. It was a nudge. It moved the spacecraft’s intercept point 180 kilometers, which sounded like a lot until you remembered the ellipse was 470 kilometers wide. The intercept probability rose from 12 percent to 40 percent, which sounded like a lot until you remembered that 40 percent meant a 60 percent chance of flying past empty space and seeing nothing.
At 03:15, a new lightcurve point arrived from the Zwicky Transient Facility. The ellipse widened to 520 by 310 kilometers. Marta updated the intercept probability. It dropped to 34 percent.
“New data made it worse,” Tomasz said.
“New data made it honest,” Marta said.
At 04:00, the burn window opened. The burn window is the period during which the spacecraft’s propulsion system can execute a maneuver that will result in an intercept. It closes at 06:14 UTC, because after that the geometry of the object’s approach and the spacecraft’s position at L2 no longer permit a reachable intercept regardless of delta-v. The window closes whether the ellipse contains the nucleus or not. The window does not care about the science. The window is geometry.
Tomasz sat at the console. The accepted-risk box was on the screen, empty, blinking. The box is a field in the mission control software. When a maneuver exceeds the mission’s pre-approved risk envelope, the flight director must type the words “accepted risk” into the box. The words are then logged, timestamped, and quoted back at the post-flight review. They are not a password. They are a confession. The system does not check them against anything. It simply records them and moves on, and the recording is what matters, because at the review the planner, who is Marta, will read them aloud.
At 04:20, Tomasz typed “accepted risk” and committed the second burn. 9.0 meters per second. The delta-v budget went from 10.0 to 1.0. The intercept probability rose to 61 percent. The ellipse, updated with a new lightcurve point at 04:05, was now 580 by 340 kilometers.
The primary target was gone. With 1.0 meter per second of delta-v remaining, the spacecraft could not reach Wirtanen. It could reach the interstellar object, or it could reach nothing. There was no third option. Tomasz had known this when he typed the words. He had written them anyway, and his hands had been steady, and I know this because the keyboard telemetry logs keystroke timing and his intervals were uniform to within 40 milliseconds, which is the signature of a person who has decided and is not reconsidering.
At 04:45, the ellipse widened again. 610 by 360 kilometers. The intercept probability dropped to 55 percent. The 200 kilometers per day of model drift was not a constant. It was an average. On bad days it was 280. Today was a bad day.
Marta came on the loop. “Tomasz. The primary target.”
“I know.”
“With 1.0 meter per second, we can do a 0.3 meter per second residual correction if the ellipse narrows. It won’t narrow. But if it does, we spend the last 0.3 and we have 0.7 meters per second of absolute nothing.”
“I know.”
“The 0.3 won’t change the intercept probability by more than two points. It is not a meaningful burn. It is a gesture.”
“I know.”
“Then why are you holding the cursor over the button?”
Tomasz did not answer for eleven seconds. I counted. “Because if I don’t spend it, I will spend the rest of my life wondering whether 0.3 meters per second would have been the difference between an image and empty space.”
“It won’t be,” Marta said. “The math says it won’t be.”
“The math says 55 percent. The math has been wrong all night.”
“The math has been honest all night. The comet has been wrong. There is a difference.”
At 05:10, Tomasz committed the residual. 0.3 meters per second. The delta-v budget went from 1.0 to 0.7. The intercept probability rose from 55 to 57 percent. The ellipse was 640 by 380 kilometers. The burn window closed in 64 minutes.
I logged the residual as a separate entry. It was not a maneuver like the other two. It was a man spending the last coins in his pocket on a lottery ticket he knew was losing, and I am a fault-protection system, not a therapist, but I note it because it is the kind of thing that goes into the review.
At 06:14, the burn window closed. The spacecraft was on its intercept trajectory. The object was somewhere inside an ellipse 640 by 380 kilometers, which is larger than the city of Prague, and the spacecraft would pass through that ellipse at a relative speed of 68 kilometers per second. At 68 kilometers per second, a miss distance of 10 kilometers is a miss of 147 milliseconds. A miss distance of 100 kilometers is a miss of 1.5 seconds. The spacecraft would be in the ellipse for less than three seconds.
Tomasz went to the sleep room. He did not sleep. I know this because the sleep room door sensor logged three openings between 06:14 and 09:00, and the coffee machine in the corridor logged four cups, and no one else was in the building except the night operator, a contractor named Bhatt who was monitoring a different spacecraft and did not know Tomasz was there.
At 09:00, the first navcam image arrived. It was a 1024 by 1024 pixel frame, taken at a distance of 8,200 kilometers from the predicted intercept point. The frame showed stars. The frame showed a dark, straight-edged facet, the width of a stadium, rotating once in 11 hours. It was not a nucleus. It was not a comet. It was a flat, dark, rectangular surface, catching no light on its face, reflecting only from its edge.
Tomasz called Marta. Marta was at home. She came in.
The spectrometer data arrived at 09:47. The spectrometer was looking for water. Water is the signature of a comet. Water is the signature of an object that formed in a protoplanetary disk around another star and carried the ice of that disk across interstellar space. Water is what Reinhardt had talked about in the corridor, what Tomasz had spent his career waiting for, what the 14.0 meters per second had been held in reserve for.
The spectrometer read no water.
Marta stood behind Tomasz’s chair. She did not say anything for a long time. Then she said, “The object is dry.”
“The object is dry,” Tomasz said.
“The outgassing model assumed water. The 200 kilometers per day of drift was modeled as water outgassing. If there is no water, the drift was not outgassing.”
“Then what was it?”
“I don’t know. The model was wrong. Not a little wrong. Wrong about the mechanism. The position drift was real, we measured it, but the cause was not what we modeled.”
“So we burned 13.3 meters per second on a model that was wrong about the basic physics.”
“We burned 13.3 meters per second on an object we did not understand, and we got an image. The image is real. The facet is real. The rotation is real. We measured a thing that no human being has ever measured before, and it is dry, and it is flat, and it is rotating once in 11 hours, and I do not know what it is.”
Tomasz looked at the image. The facet was dark. It was straight-edged. It was the width of a stadium, and it was turning slowly in the light of a star that was not the sun, and it was the first solid object from another stellar system that a human spacecraft had ever flown past, and it was dry, and it was flat, and it was not a comet.
He did not say it was worth it. He did not say it was not worth it. He opened the mission control software, navigated to the review form, and in the box labeled “accepted risk,” which was the same box he had filled at 04:20, he typed a second line beneath the first.
The first line read: “accepted risk, 9.0 m/s redirect burn, 04:20 UTC.”
The second line read: “accepted risk, 0.3 m/s residual correction, 05:10 UTC.”
He saved the form. He closed the laptop. He stood up, and Marta stepped aside, and he walked to the window of the control room, which faced east, toward the dawn that was coming up over the Rhine, and he stood there for a while, looking at the sky, which was gray and ordinary and full of clouds, and then he went back to his console and began writing the preliminary report, starting with the spectrometer reading, which was zero, and the rotation period, which was 11 hours, and the facet width, which was 180 meters, and the delta-v remaining, which was 0.7 meters per second, and the primary target, which was gone.