Here Be Dragons:
Characterization of ACS/WFC Scattered Light Anomalies


We present a study characterizing scattered light anomalies that occur near the edges of the Advanced Camera for Surveys (ACS) Wide Field Channel (WFC) images. The study is based on all full-frame WFC raw images ever produced by ACS. Using the 2MASS catalog, we identified stars that cause two particular scattered light artifacts known as ”dragon’s breath” and edge glow. These artifacts are caused by stars located in narrow bands outside the ACS/WFC field of view. We have completed this study for the ACS F606W and F814W filters. The results for both filters are similar when expressed in total fluence, or flux multiplied by exposure time. We provide a map of risky areas around the ACS chips and an upper limit of magnitudes to be concerned about. We will use these results to develop interactive tools that will aid the astronomical community in the proposal process for ACS/WFC.


ACS/WFC images can suffer from a number of optical and scattered light anomalies. Most of the optical anomalies that effect ACS have been well characterized. Hardware, software, and optical anomalies are discussed in ISR 2008-01. This is not the case for the scattered light anomalies known as ”dragon’s breath” and edge glow. Dragon’s breath is caused by reflections being scattered back to the detector. There is a knife-edged mask in front of the CCD that scatters light back to the detector when its back side is illuminated by reflections from the CCD surface. These phenomena were discovered in early testing of ACS and were mitigated by sharpening the knife edges and coating them black. However, when point sources fall on the edge of the mask, scattering still occurs (Hartig et. al.).

Figure 1a: Dragon’s Breath

Figure 1b: Edge Glow

Although ACS was designed with a requirement limiting the amount of energy that may be contained in an anomalous feature relative to the object producing it, this scattering exceeds that limit by an order of magnitude (Hartig et. al.). These anomalies can have potentially severe effects on data but can be avoided when designing observations.

In this report, we identify the upper right and lower left corners of the detector as the most severely affected regions of ACS. A range of magnitudes has also been determined in which stars may cause scattering. This information has been worked into an interactive tool to aid the community in the proposal process.


Creating Guide Star Files

In November 2013, MAST deposited all full-frame WFC raw images into directories sorted by cycle and anneal date on ACS servers. For each full-array broad-band ACS/WFC exposure, we generated a catalog of stars from the Hubble Guide Star Catalog II (GSC II) within 3’ of the pointing. This catalog contains guide star ID, position, several photographic magnitudes, image file location, gain, and exposure time.

Identifying Dragon’s Breath

We inspected all full-frame ACS/WFC images with exposure times longer than 350 seconds for inspection. We created distortion corrected FITS mosaics with an overlay of 2MASS objects for each frame that contained anomalies. Using these images we matched the ”offending stars” off the field of view (FOV) with the scattered light on the detector. We also measured the size of each artifact by observing its linear extent in pixels. In more crowded fields it was important that features, such as glint or diffraction spikes, were not mistakenly identified as dragon’s breath. Edge glow required less nuance to identify, and was typically accompanied by a central spike which made selecting the correct offending star simple.

Figure 2: 2MASS stars overlayed on ACS/WFC full frame image. The stars are shown as solid white circles with sizes corresponding to their brightness.

Figure 3: A portion of an ACS/WFC image with the 2MASS overlay. The offending star associated with the larger dragon’s breath feature is marked in pink. The star marked in yellow is an example of a star on the edge of the chip, which is causing glint.

Once we created this catalog containing offending stars, their coordinates, and the size of the scattering, they were matched with guide stars in the previously created guide star files. The physical coordinates recorded from the ACS/WFC images were converted into RA and Dec, then matched to guide star coordinates in the corresponding guide star files with a 5 arcsecond threshold to account for possible alignment offsets. Mismatches are possible, but cases with two very close stars were discarded to mitigate this possibility.

With the guide stars identified by name, we were able to use the information from the guide star catalogs to determine the total ACS filter magnitude for each star. The guide star catalog uses F and J magnitudes. Due to the lack of J magnitude information for many of the stars estimates were made based only on the F magnitude. The photographic F filter has an effective wavelength around 6700Å. Magnitudes in F606W or F814W are similar (to within less than a magnitude). We also normalized the magnitude of each star, based on the exposure time in seconds, to 500 seconds

\begin{equation} mag500=F_{mag}-2.5log(\frac{exptime}{500})\nonumber \\ \end{equation}

The resulting mag500 is a measure of the total fluence, or flux multiplied by exposure time:

\begin{equation} flux500=10^{-0.4\times mag500}=flux\times(\frac{exptime}{500})\nonumber \\ \end{equation}

The greater the fluence, the more charge deposited onto the CCD, and the greater the scattered light artifact for a given star position relative to the ACS/WFC detectors.


How Dragon’s Breath Can Manifest

Through the cataloging process we discovered that dragon’s breath can appear in different forms. The name ”dragon’s breath” comes from classic examples as seen in figure 4a, which look like fire shooting onto the frame. The scattering can also take on more irregular forms as seen in Figure 4b. Location can affect the size and shape of the anomalies.

Figure 4a: Two cases of classic Dragon’s Breath. This HST ”preview” image was generated by MAST and is available at

Figure 4b: Elongated scattering on the upper left side stretches to the center of the detector. This HST ”preview” image was generated by MAST and is available at

By inspecting individual stars that caused scattering in multiple exposures, we determined that small changes in star location can affect the size and shape of scattering. Take the particularly egregious example of a star in the corner of the chip in Figure 5. These two instances of scatter are caused by the same star in slightly different positions with respect to the detector. Figure 6 shows another instance of corner scattering which changes with the position of the star. This location dependency is not limited to the corners and can occur anywhere around the detector.

Figure 5b: Corner Scattering. Frame j8de59y0q.

Figure 5a: Corner Scattering. Frame j8de59xwq.

Figure 6: Corner Scattering four observations: j8zq07zvq, zyq, zwq, and a1q.

We also inspected cases where observations were made in both filters during the same visit. Differences in dragon's breath between the two filters studied are due to the stars being different magnitudes in the different filters. In the two examples below, images from the same visit are compared and small differences in the size and shape of the dragon's breath are apparent. The sizes of the circles in the plots below are different in each filter due to the plotting, not necessarily the magnitudes.

Figure 7a: In the F606W image on the left the normalized magnitude is 13.7 with a 510s exposure time. In the F814W image on the right the normalized magnitude is 14.1 with a 350s exposure time. The scattering in the F814W image is slightly smaller with differences in structure.

Figure 7b: In the F606W image on the left the normalized magnitude is 15.0 with a 707s exposure time. In the F814W image on the right the normalized magnitude is 15.8 with a 357s exposure time. The scattering in the F814W image has a slightly larger spread.

Anomaly Map

We plotted the locations of all scattered light anomalies in both F606W and F814W around a footprint of the detector. The anomaly map in figure 8 shows that the scattered light occurs in a very thin band in the upper right and lower left corners of the detector. This positioning could be due to the fact that the WFC detectors are a rhombus shape, or it could be related to the positioning of the knife-edge mask above the detector. There are two clear loci. The outer locus is made up of dragon’s breath whereas the inner locus is predominantly edge glow.

Figure 8: Positions relative to the ACS/WFC detectors of stars which caused dragon’s breath or edge glow.


The anomalies we identified were due to stars with corrected magnitudes between 10 and 20. A histogram of the the magnitudes of offending stars are plotted along with the magnitudes of all guide stars is shown in figure 9.

We also compared the linear extent of dragon's breath in pixels to the magnitude of the star causing it. Figure 10 shows that there is correlation between magnitude and scatter length.
Replace this text with your caption

Web Interface

Using the catalog we created two interactive plots which allows users to explore our results. The plot is similar in appearance to the anomaly map in figure 8. Each point represents a star in or near an ACS/WFC observation and the black lines represent the ACS/WFC chips. Users can hover over each point and see the image in which it caused scattering. Above the image there is also information about the star including the guide star name, the name of the image, the filter used, the filter magnitude, and the magnitude corrected to 500 seconds. There is also a slider which selects stars based on their exposure time corrected detector magnitudes. This slider allows the user to view stars withing 1 magnitude of the selected magnitude. This is a great tool for exploring examples of scattered light appearing in multiple dithered observations, in multiple filters, and how magnitude can determine the length of the scattering.

Conclusions and Future Work

Our anomaly map clearly shows that small displacements (\(\sim\)1 arcsecond) can prevent potentially severe scattered light in your observations. The locations prone to severe scattered light are apparently filter-independent. Two clear loci of scattering are evident in figure 8. The STScI ACS instrument team is making available the interactive web interface described above for further exploration of these results.

[Someone else is editing this]

You are editing this file