Astrometry and Photometry
Astrometry
To properly identify, get accurate RA and Dec, and measure the magnitude of sources on our plates, we need to have a “truth.” In our work, we’ve used Gaia EDR3 data to project the location of stars at particular dates. This requires us to calculate proper motion of the stars on the date of the plate. In a recent example, we worked with a plate taken on March 22, 1905. The equations we applied to the Gaia data to identify the locations of sources on that date are:
RA(1905) = RA(2006) - ((µRA(2016) * 110.8)/(1000 * cos(Dec(2016) * 3600))
Dec(1905) = Dec(2006) - ((µDec(2016) * 110.8)/(1000 * 3600))
We limited our results based on magnitude. This is done because first, results drop off for sources greater than about 16.9 mag because of the limitations of the scanner and the plate. Secondly, Aperture Photometry Tool (APT), the software we use to determine magnitude on our plates, has an operational limit of 100,000 stars, so on a crowded field, it is necessary to make some cuts.
Aperture Photometry Tool (APT)
APT is free, and is compatible with both Windows and Mac OS. To determine magnitudes, APT places a fixed aperture on a source, automatically centered on the light based on predetermined settings, and subtracts the sky background value to produce a quantity. We input the list of positions derived from Gaia into APT and produce a table of magnitudes at those locations on the plate.
Different settings in APT can influence the output. The selected aperture radius, centroid radius, and sky subtraction options all matter in optimizing the plate catalog results. It is important to adjust each for the specific characteristics of particular plates.
Accuracy of the plate catalog
After using the Gaia data to create a plate catalog, we run the whole process again on the rotated plate scan. This provides the opportunity to analyze the astrometric and photometric repeatability of results, as well as allows us to utilize the more accurate results from the two scans in further studies.
Utilizing the Tool for OPerations on Catalogus And Tables (TOPCAT), we match the APT-derived plate catalog with the Gaia catalog. (Note that these catalogs are not independent of each other, as we have utilized Gaia to create the plate catalog. There are, unsurprisingly, errors that occur. Irregularities introduced by the scanner can be mathematically corrected and applied to the original APT tables. Once corrected, the perpendicular scans can be matched again, and the final, corrected catalog can be matched with Gaia. In this final match, we use the coordinates with the least residuals (in our experience, this is RA from non-rotated, and Dec from the rotated scan).
The final match leaves us with a catalog with the highest number of matched sources and a low residual. In a recent study, this process resulted in 66,000 matched sources with a median residual of 1.37 arcsec.
Photometric repeatability and calibration
Utilizing two scans (rotated and non-rotated) makes it critical that our results are repeatable. Testing indicates that this is indeed the case. In a recent study, the average difference between mag measurements was about 0.09 mag.
To calibrate to a photometric zero-point that is native to the photographic plate, we must account for the differences in our effective bandpass (covering ~400-500nm, limited at the blue end by lenses, at the red end by emulsion sensitivity, and impacted by interstellar redding). Our strategy follows the practice of Russell et al (cite), and uses the simple relation pg=bp + α * (bp-rp) where pg is the photographic mag based on Gaia photometry, bp is the Gaia mag, and bp-rp is Gaia color. The value of α minimizes the dependence of pg - mag on bp - rp (and also minimizes the scatter in other relationships.)
Catalog completeness
In a recent study, this process established a source list of 66,000 stars out of 100,000 possible matches. When those 66,000 matches are even more specific, focusing on 3.25º from the focal center, our source list matches 53,600 stars, just over 81%.
The difference is due to:
- Faint stars that are below the threshold because the sky background subtraction is slightly overestimated,
- Faint stars that are near brighter stars, and are therefore missed,
- Optical aberrations towards the edges of the field make stars harder to identify,
- A vignetting effect causes less light to reach the edges of the plate.
A pending publication applies these methods, and will be available in the near future. Data sets used can be accessed through the Knowledge@UChicago repository here.
Ongoing research demonstrates the resulting catalog is usable for contemporary astronomical research purposes, including identification of novae, identification and magnitude measurements of supergiants, and variability study.