GRAVITY+: reducing non-common-path aberrations for sub-10 {\mu}as astrometric accuracy

Imagine the universe as a giant, dark ocean, and the stars as tiny, distant lighthouses. For centuries, astronomers have tried to measure the exact distance between these lighthouses to understand how they dance around each other. But looking at them from Earth is like trying to read a newspaper through a wavy, dirty window; the air above us shimmers, and our telescopes have tiny imperfections that blur the view. To fix this, scientists built a super-powered "eye" called an interferometer. Instead of one giant mirror, it links several telescopes together, acting like a single, massive lens to see details that are incredibly small—so small we measure them in microarcseconds (a tiny fraction of the width of a human hair seen from a kilometer away). This is the realm of the GRAVITY instrument, a high-tech machine that has already helped us weigh the black hole at the center of our galaxy. But even the sharpest eyes have a blind spot: tiny, invisible ripples in the light's path that the machine can't quite see or fix on its own.

This paper is a detective story about finding and fixing those invisible ripples. The team behind the GRAVITY instrument realized that while their machine was amazing, it wasn't quite perfect yet. They suspected that tiny, high-order distortions in the light's journey—specifically in a part of the machine called the "fiber coupler"—were messing up their measurements. Think of the fiber coupler as the machine's internal plumbing, guiding light from the telescope into the sensors. The scientists wanted to know: Is the plumbing itself slightly bent or bumpy, causing the light to arrive at the wrong time? They set out to measure these tiny errors, figure out exactly where they come from, and design a new, smoother plumbing system to make the instrument even sharper.

The Detective Work: Finding the "Ghost" in the Machine

The researchers, led by Quentin Fournier and the GRAVITY+ Collaboration, started by looking at the light coming out of the machine in their lab. They used a special technique called phase-shifting interferometry, which is like shining a laser through a foggy window to see exactly how the fog distorts the beam. They found that the light wasn't perfectly smooth; it had "high-order wavefront errors" with a roughness of about 75 nanometers in one mode and 50 nanometers in another. To put that in perspective, a nanometer is a billionth of a meter—so these errors are incredibly tiny, but for an instrument trying to measure the universe with such extreme precision, they are huge.

To figure out if these errors came from the whole telescope system or just the fiber coupler, they built a 1:1 replica of the fiber coupler in their lab. It was like building a perfect model of a car engine to test if the engine itself was the problem, without the rest of the car getting in the way. When they tested this replica, they found two main culprits causing the distortions, and the culprit changed depending on how the machine was looking at the stars.

The Two Villains: Bumpy Mirrors and Shiny Ghosts

The first villain appeared when the machine was looking at two different stars at once (called "off-axis" mode). The scientists found that the mirrors inside the fiber coupler, specifically the off-axis parabolas (OAPs), had tiny, repeating scratches left over from their manufacturing process. These are called "diamond-turning signatures." Imagine a mirror that looks smooth to the naked eye, but under a microscope, it has a pattern of tiny, rhythmic bumps, like the grooves on a vinyl record. These bumps were causing the light to wobble, creating an error of about 37 nanometers in the off-axis mode.

The second villain showed up when the machine was looking at a single bright star (the "on-axis" mode). Here, the problem wasn't the bumps on the mirrors, but something called "optical fringing." This happens when light bounces back and forth between two glass surfaces, like a whisper echoing in a narrow hallway, creating a pattern of bright and dark stripes. The scientists discovered that the special coating on a prism inside the machine was acting like a tiny echo chamber, creating a massive error of about 90 nanometers. This was the dominant problem for the on-axis mode.

Testing the Theory on the Real Sky

To prove that these lab findings were actually affecting their real observations, the team turned their gaze to a famous double-star system called GJ65. They knew exactly how far apart these two stars were supposed to be. They then watched the stars using the real telescope in both the "off-axis" and "on-axis" modes.

The results were a perfect match. When they looked at the data from the on-axis mode, they saw a rhythmic, wiggly pattern in the measurements that matched the "echo" effect they found in the lab. It was like hearing a specific song playing in the background of a recording and realizing, "Ah, that's the echo from the hallway!" They found that this wiggly pattern caused the measurements to be off by about 300 nanometers in the optical path. In the off-axis mode, the wiggles were much smaller, but the "bumpy mirror" scratches were still there, causing a smaller but noticeable error.

The team also noticed that these errors were tricky. If you watch the stars for a long time, the Earth's rotation moves the telescope, and the errors might average out. But if you need to take a quick snapshot (just a few minutes), these errors stick around and mess up the measurement. This is a big deal for studying things like exoplanets, where you often need quick, precise measurements.

The Fix: Smoother Mirrors and Better Tricks

So, how do we fix a machine that is already one of the best in the world? The team has a two-part plan. First, they are using a clever trick called "pupil modulation." Imagine spinning the window you are looking through while you take a picture; the blurry spots move around and cancel each other out. They are already doing this for some observations, and it helps smooth out the "echo" errors.

Second, and more importantly, they are building a brand new set of mirrors for the fiber coupler. They are working with a company called TNO to manufacture new off-axis parabolas using a special polishing method. They tested a prototype mirror, and the results were stunning: the new mirror was ten times smoother than the old ones, with surface errors dropping from 84 nanometers down to just 8 nanometers. It's like replacing a bumpy, old road with a perfectly paved highway.

These new mirrors are scheduled to be installed by the end of 2027 as part of the "GRAVITY+" upgrade. Once this happens, the scientists expect to push the accuracy of their measurements down to less than 10 microarcseconds. That's like being able to spot a coin on the Moon from Earth, or seeing a single human hair from 10 kilometers away. By fixing these tiny, invisible ripples in the light, the GRAVITY+ team is preparing to unlock even deeper secrets of the universe, from the spin of black holes to the orbits of planets around other stars.