Work around a changing detector

The BRITE team encountered more radiation-related CCD damage than expected. Added shielding on later spacecraft, including Heweliusz, and changes in observing and readout methods helped reduce its effects. These were responses to measured problems, not proof that a small satellite can ignore its radiation environment.

A hot detector pixel can add an unwanted signal at a fixed detector location. Starlight also creates bright pixels, so brightness alone cannot distinguish the two. Information about position, time, background and repeated exposures helps determine whether a feature belongs to the sky or the detector.

Hardware changes are constrained by the stage of a mission. A detector shield may be altered before launch if testing, volume and mass permit; that does not mean the same modification can be installed on a spacecraft already in orbit. Operational workarounds address a different set of possibilities.

BRITE adopted a method called chopping: successive exposures place a target at different detector positions. Subtracting the images suppresses stable detector defects while leaving positive and negative images of the moved star. The published paper describes a benefit in photometric quality and a cost in the amount of image data required.

Use an original three-position toy detector. Exposure A contains values [100, 30, 0]; exposure B contains [0, 30, 100]. The star moves from the first position to the third, while the fixed defect contributes 30 at the middle position in both. Subtracting B from A gives [100, 0, −100].

The negative value is a difference, not negative physical light. It marks where the star appeared in the subtracted exposure. Both original exposures contain nonnegative counts. The signs in the difference image retain the direction of subtraction and must be interpreted accordingly.

Move the target and subtract the fixed defect

The ideal cancellation above assumes that the defect is stable between exposures. If the middle value becomes 40 in exposure B, subtraction leaves −10 at that position. A changing defect or altered background can therefore leave a residual. Subtraction improves evidence only within the assumptions tested.

The method also does not magically repair every way a damaged CCD can distort charge. The team treated charge-transfer effects separately through readout changes. Different failure mechanisms require different evidence; a cleaner-looking background does not demonstrate perfect recovery of all stellar signals.

In an invented data-budget example, a small image region has 24 × 24 = 576 pixels. Extending it to 48 × 24 = 1,152 pixels doubles the pixel count. At unchanged bits per pixel and no compression change, the same link can carry half as many such image regions. Better measurements can be worth the cost, but the trade should be explicit.