UPDATE — Light Sabers have Earthly Origin!
24 September 2014
New investigations of these artifacts have determined the root cause. They are not due to internal light scattering. These so called light saber are the result of bright earth flat fields that left persistence, which was then imprinted on our Frontier Fields data. Some, but not all of the persistence was flagged, thereby resulting in the remaining persistence appearing in our final image products.
Now that we have identified the root cause, we are DOING SOMETHING BETTER…..
16 September 2014
A new undocumented artifact appeared in the Frontier Fields exposures of the MACSJ0416.1-2403 cluster on Visit 94. As seen in Figure 1, the artifact appears almost as a string of satellite trails that appear strongly on the left side of the images and diminish greatly towards the right side of the images.
Figure 1: Scattered light in the MACSJ0416 field of view, thought to be from a bright star just outside the field of view.
Due to their unique appearance, we call these artifacts “light sabers,” following the theme of the WFC3 “Death Star” artifact. Light sabers are most apparent in the first exposure of the F105w filter, and less intensely so in exposures 2-4 for F105w and in exposures 1-4 of the F160w filter. The suspect image is the first exposure in F105w, with the presence of the artifact in subsequent exposures and in the F160w images regarded as persistence of the initial light scattering.
The main cause of the light sabers artifact is considered to be internal scattering of the light from the bright star located at 04:16:6.9752, -24:05:43.01 (RA,DEC). The main light saber is marked in a purple (Figure 2) with the rest of sabers (thought to be harmonics of the primary) marked in other colors. The dashed green line seems more likely to be due to a satellite trail.
Figure 2: Highlighted light saber artifacts, believed to be caused by the bright star just outside the field of view.
To determine if a unique position on the detector would cause the effect, the centroid of the bright star was mapped onto WFC3 detector coordinates. The ACS drizzled image of MACS0416 from 2014-Jan-10 in F435W was used to determine the centroid of the bright star and obtain the precise RA/DEC coordinates. Drizzlepac’s SKYTOPIX was used to map the coordinates of the star to WFC3 and the results are shown in the plot.
Figure 3: The location of the bright star as determined in separate exposures in different filters, shown in WFC3 sky coordinates.
The light saber effect is caused by the exposure indicated with the red circle in Figure 3, with the gold points representing images showing the artifact via persistence. A well-documented and associated glint caused by the bright starlight is also indicated in green. The results seem to indicate that the light saber effect is primarily seen when the bright star is positioned very close to vertical center on the detector and -55 pixels off of WFC3’s frame. This position matches closely with the noted position of the instrument baffle as indicated in Tom Brown’s 2007 ISR 2007-16.
An example of the glint can be seen in Figure 4 inside the red circle. Additional information and examples are available in Brown’s ISR 2008-06. The associated glint is thought to be caused by bright starlight scattering off the knife-edge of the baffle, and we suspect the light saber effect to be caused by the same.
Figure 4: An example of the glint. See ISR 2008-06 for more information.
A mask has been created for light sabers and any images affected have been appropriately reduced. All pixels impacted by this effect in the current epoch were fully masked, so that the final combined mosaics of MACSJ0416.1-2403 do not include any of the affected pixels. Upcoming observations of MACS0717 are found to not be affected by this particular artifact, and future Frontier Fields observations will be positioned appropriately and taken with the effect of light sabers in mind.
Harish Khandrika – Frontier Fields Science Data Products Team Member – WFC3/IR