Enhancing the performance and capabilities of the MIRI instrument on JWST

Imagine the universe as a giant, dusty attic filled with secrets. Most telescopes are like flashlights that only see the bright, obvious furniture in the center of the room. But some of the most interesting things—like the warm glow of baby stars being born or the dusty rings around distant planets—are hidden in the shadows, glowing in a special kind of light called "mid-infrared." This light is like a heat signature; it's invisible to our eyes and gets blocked by Earth's atmosphere, so we can't see it from the ground. To catch these heat signatures, we need a super-sensitive, space-based camera that can operate in the freezing cold of space. That's where the James Webb Space Telescope (JWST) comes in, and specifically, its "Mid-Infrared Instrument" (MIRI). Think of MIRI as the only pair of night-vision goggles on the entire telescope that can see past 5 micrometers (a unit of light wavelength). It's the only tool we have to peek into this specific, challenging corner of the cosmic attic.

Now, even the best tools need a little tune-up. Just four years into its mission, the team behind MIRI realized that while the camera was taking amazing pictures, it was occasionally tripping over its own shoelaces. This paper is essentially a "mechanic's report" on how they are tightening those laces. The authors aren't discovering new stars in this document; instead, they are describing a series of software updates and hardware tweaks designed to make the instrument faster, quieter, and less likely to make mistakes. They are fixing a buzzing noise that messes up the data, inventing new ways to read the camera's sensors to save space and reduce static, and creating a safety net to prevent the telescope from losing its target. The goal is simple: to ensure that when MIRI looks at the universe, it sees it clearly, without the fuzz, the buzz, or the blunders.

The Buzzing Noise and the "In-Phase" Fix

Imagine you are trying to listen to a quiet song on the radio, but every time the needle hits a specific beat, a loud, rhythmic buzz interferes. That's what was happening to MIRI's camera. The instrument was picking up a specific electromagnetic interference (EMI) noise at a frequency of 390 Hz. This noise was like a ghost that only appeared in certain parts of the image, and it was especially annoying during "Time Series Observations" (TSOs), where the telescope watches a single object for hours to see how it changes. Because the noise pattern moved around, it looked like a shimmering, distracting artifact in the data.

The engineers realized that the noise was out of sync with the way the camera was reading its pixels. It was like trying to clap in rhythm with a drummer who is slightly off-beat. To fix this, they didn't just try to filter the noise out later; they redesigned the "subarrays" (smaller, faster sections of the camera sensor used for specific tasks). They created new versions of these subarrays, labeled with "IP" for "In-Phase." By slightly changing the shape and size of these reading areas, they shifted the timing of the pixel reads so that they landed exactly in sync with the noise. Now, instead of fighting the buzz, the camera reads the data in a way that the noise doesn't interfere. For example, the old SUB128 subarray is being replaced by a new SUB128 IP, which is just a tiny bit wider (132 columns instead of 128) to keep the timing perfect. This change is so precise that observers can switch between the old and new versions with almost no difference in how long the observation takes.

Swapping the "Slow" for the "Smart Average"

The paper also tackles a problem with how the camera reads data. The standard way MIRI reads its sensors is called FASTR1, where it snaps a picture of every pixel once. For very faint objects, the team previously tried a mode called SLOWR1, which was supposed to be like taking eight photos of the same pixel and averaging them together to cancel out the "static" (read noise). The idea was that averaging eight photos would make the image clearer by a factor of the square root of 8.

However, the team found that SLOWR1 wasn't working as expected. Not only was the noise reduction not as good as predicted, but switching between the fast mode (FASTR1) and the slow mode (SLOWR1) caused the instrument to heat up and cool down unevenly. This thermal settling created weird, arc-shaped artifacts in the images, like a ghostly smear appearing right after the switch. It was like changing gears in a car so abruptly that the engine shuddered and left a skid mark.

The solution was a clever middle ground called FASTGRPAVG8. Instead of changing the camera's internal state to take multiple slow reads, the camera still snaps a single fast photo of each pixel (just like FASTR1). But then, the computer takes 8 of these fast frames and averages them together after they are read, but before they are saved. This achieves the same noise-reduction benefit as the slow mode without the thermal shuddering or the weird artifacts. It's like taking eight quick snapshots with a smartphone and letting the phone's software blend them into one perfect, low-noise image instantly. The team plans to switch to this FASTGRPAVG8 mode for science observations starting in Cycle 6, phasing out the problematic SLOWR1 mode.

The "Dithering" Safety Net

One of the most critical parts of taking a picture with MIRI is "Target Acquisition" (TA). Before the telescope can start its real science, it has to find the exact center of the target star or galaxy and lock onto it. For some modes, the target has to be placed within a tiny slit (0.5 arcseconds wide) or a specific spot on a mask. If the telescope misses this spot, the whole observation is ruined.

The team analyzed hundreds of past attempts and found that while most were successful, about 12% failed. Many of these failures were due to user errors, like picking the wrong brightness settings, but some were caused by unpredictable "rogue" pixels on the camera or sudden hits from cosmic rays (high-energy particles from space). Imagine trying to balance a coin on a table; if a tiny speck of dust (a bad pixel) or a sudden gust of wind (a cosmic ray) hits the table right when you're placing the coin, it might fall.

To fix this, the team is introducing a "dithered" approach. Instead of taking just one picture to find the target, the telescope will take three pictures, shifting its position slightly between each one. If a cosmic ray hits the sensor in the first picture, it won't be in the same spot in the second or third. When the computer combines these three images, the random noise and bad pixels disappear, leaving a clear picture of the target. It's like taking three photos of a moving bird and stacking them; the bird stays in the center, but the random specks of dust on the lens cancel out. This new method is expected to roll out in late 2026 or early 2027, making the telescope much more reliable at finding its targets.

New Ways to Look at the Universe

Finally, the paper highlights a few new "modes" or ways of using the instrument that are being unlocked. One major addition is "Wide-Field Slitless Spectroscopy" (WFSS). Previously, the telescope could only spread out the light of a single object into a rainbow (spectrum) if it was looking through a specific slit. Now, with WFSS, the telescope can spread out the light of every object in a large field of view at once. It's like turning a single-lens camera into a wide-angle prism that captures the rainbows of an entire crowd of stars simultaneously. This mode will be available starting in Cycle 5.

Additionally, the team is testing the use of the narrow slit for Time Series Observations. Previously, they only used the wide, slitless mode for watching planets transit (cross in front of) stars because they were worried the telescope might drift and lose the target in the tiny slit. However, new data shows the telescope is incredibly stable, staying within 0.4% of a pixel over 24 hours. This stability means they can now use the narrow slit, which blocks out more background noise, to get even clearer data on these transiting planets. They are also introducing new, smaller subarrays (like SUBSLIT) that allow for faster reading of data, which is crucial for catching the rapid changes in these planetary events.

In summary, this paper is a celebration of continuous improvement. The MIRI instrument is already a marvel of engineering, but by fixing the buzzing noise, smoothing out the reading process, and adding safety nets for target acquisition, the team is ensuring that the telescope continues to deliver the clearest, most reliable views of the mid-infrared universe possible. These aren't just minor tweaks; they are essential upgrades that will keep JWST at the cutting edge of discovery for years to come.