 |
 |
 |
Analysis of a Newly Identified Variable Star In Aquarius
Heath Gibson
4. Methods and Observations
In
order to get observation time at the Mt. Wilson and Table Mountain Observatories,
a proposal was submitted to TIE to participate in their variable star research
project; later a proposal was submitted to JPL, through the California State
University of Los Angeles, requesting research time on the 24” telescope
at their Table Mountain facility.
To prepare for observation, the first step was to establish the star field for the suspected variable and then select reference stars within the field. Star fields were necessary because they helped ensure the variable candidate was in the center of the telescope’s focus. In order to establish the star fields, I used TheSky Astronomy Software to create a field of view indicator, in other words, to create lines representing the boundaries of the field. The border entitled, “ST-6 on Mt. Wilson 24” ” in Figure 1 is the field of view indicator. After I entered the right ascension and declination of the suspected variable, the program displayed a sample image of the variable candidate in its star field, bounded by the field of view indicator. After establishing the star field for a suspected variable, I selected reference stars within the field according to the similarity of their red magnitudes (obtained from the USNO-SA2.0 database) and that of the variable. Figure 1 shows the star field for suspected variable N and its reference stars, created by TheSky.
Figure 1. Star field for suspected variable N (VAR N in
figure). REF stands for reference
star, followed by the star’s designation: A, B, or C.
The procedure for using the telescope at Mt. Wilson and Table Mountain was the same. First, I lowered the temperature of the CCD cameras, either thermoelectrically or cryogenically, and then focused them. Next, I selected the red filter and began imaging my stars, cycling from N to Q in order to detect a change in brightness over time. At the end of each session, I took three to five dark frames with the telescope cover closed, and three to five flat fields with the cover open and the telescope pointing at a uniformly illuminated white projection extending from the wall.
5. Data Analysis
If the stars in question were variables, their change in brightness
over time would be reflected in their photon counts recorded in each image.
Using this reasoning, I analyzed the photon counts of each image taken
over six weeks of observation in order to check for any signs of variability. The star that varied the most in the shortest period of time
was the N star, indicating it was a variable. The next step was to plot a light curve for N to discern what
type of variable it was. In order
to get as many data points as possible, and thereby a pronounced light curve,
all future research was focused on star N.
The next step was to calculate the magnitude of the variable in
each image taken and plot it against Julian time to see how it varied. Then I analyzed that data using the following
algorithm.
The first step was to rectify an image. Using CCDSoft and/or Mira, all of the
dark frames were combined into one image and then that combined image was
subtracted from the original image to eliminate the noise that comes from
the CCD. When Mira or CCDSoft
“subtracts” the combined dark frame image from the original image,
it subtracts the actual photon value on each pixel in the dark frame from
each pixel value in the original image.
After performing a dark subtract, the next step was to divide the
adjusted original image by a combined flat field. After every session, a flat field must
be taken to “remove [the] non-flatness and pixel response from CCD frames”
(MIRA AP User’s Guide 41) and remove artifacts such as “cosmic
ray events, stars, and transient specks of dust on the detector or filter” (MIRA 42). Some irregularities that a
flat field can correct are dust particles or a mark or scratch on the mirror
or CCD chip. The effect of such
artifacts on a CCD frame can
be seen in Figure 2. Using CCDSoft
and/or Mira, the flat fields taken were combined into one and then the adjusted
original image was divided by the combined flat field.
Figure 2. When images are taken using a CCD, abnormalities called donuts can be seen in the image. Each donut is an out-of-focus dust particle on the CCD chip itself. These "donuts" can be identified by the shape of a black ellipses with a dot in the middle.
After performing a dark subtract and flat field divide, the image was fully rectified and ready for photometry. The equation for rectifying images can be seen in Equation 1.
The second step in this process was the photometry. In order to find the magnitude of the suspected variable, I used the following procedure. First, I counted every photon value present in my suspected variable and all three reference stars by opening up an image in either CCDSoft and/or Mira, magnifying the star five to six times so the pixels could be seen, and then running the pointer over every pixel to read the value. Every pixel value was recorded for each star.
Figure 3. The squares in figure 6 are the pixels in the image of a magnified star.
After recording the pixels for all four stars (my suspected variable and three reference stars) in every image, I added up all the pixel values for each star to get a gross photon count. Then, by running my pointer over the background of the image (the part without stars) I obtained a median background value. In order to obtain a net photon count (the value needed for photometry), I multiplied the median background value by the number of pixels in each star and subtracted that product from the gross photon count for each star. Table 5 is an example of the data taken for a single image. After obtaining the net photon count for each star, I calculated the magnitude of variable N in each image. For the first twenty images, I checked the accuracy of the images by calculating the magnitudes of the reference stars as well, verifying that the values of the reference stars were constant. The following equation was used for calculating each staržs magnitude:
Table 5. Hand Photometry Chart
Note. The value 4510 is the median background photon value;
NC is the net photon count value for each star. VAR N is the column for the photon counts on suspected
variable N. REF A-C are the columns
for the photon counts of the reference stars.
