Bohman, John
A Photometric Approach to the Redshift of Galaxies
Faculty Mentor: Joseph Moody, Astronomy
It is necessary to study the distribution of matter to better understand the universe. There are many difficulties associated with this task however, one of the most basic being that that the universe is an extremely large space, and it takes a lot of time and effort to observe faint objects. Multi fiber spectrometers have made this task much easier, and over 1.4 million spectroscopic redshifts have been obtained. While this is certainly impressive, much remains to be done. This project attempts to develop a survey method capable of detecting and finding the redshift of relatively close (meaning a redshift of less than 10,000 km/sec) dwarf galaxies with strong emission lines using photometric methods instead of spectroscopic methods.
This method will allow us to more easily study the distribution of dwarf galaxies inside of the galactic voids. The overall structure of the universe has been studied, and we have a good idea of the galactic superstructure. We know that our own Milky Way galaxy is located within a group of galaxies known as the Local Group, which is itself contained within a larger cluster of groups called the Virgo Supercluster. Superclusters form a vast web where most galaxies occur, with spaces in between called voids. There are several different theories about galaxy formation, and conflicting ideas on whether they allow the formation of galaxies within the voids. This new method, developed by my mentor and I, aims to effectively study dwarf galaxies that fall within the voids, specifically FN2 and FN8. In order to tell the distance to the galaxy, and consequently, if the galaxy is in a void, it is vital to determine the redshift.
We obtained a set of specially crafted “ramp” filters to image through. These filters vary linearly in transmission with opposite slope over the same bandpass. The Blue Slope, or “BS” filter peaks at 660 nm, while the Red Slope, or “RS”, filter peaks at 681 nm. We used a standard filter centered at 697 nm to measure the continuum (see figure). We took data on 19 different Seyfert type I and II galaxies. Galaxies of known redshift with a strong hydrogen alpha emission line were chosen in order to test the reliability of this method. The ramp filters were designed to measure the hydrogen alpha line at redshifts occurring between 3,500 and 10,000 km/sec. Data was taken using BYU’s West Mountain Observatory (WMO) 0.9m telescope, and BYU’s ROVOR 16” telescope. Exposures were between two and three minutes for the RS and BS filters, and about ten minutes for the 697 filter.
Data was reduced using MIRA Pro UE. We used the data from the continuum filters to adjust the data for the RS and BS filters. This was done so that galaxies with high continuum would be
result in comparable measured redshift compared to galaxies with low continuum. We took the BS/RS continuum corrected ratio and plotted it against the known redshift values. We used this to fit a third order polynomial of the relation between the calculated and known values. We found the relationship between the calculated redshift, and the published redshift. The standard deviation of the data is 449 km/sec.
An accuracy of 450 km/sec is well within acceptable bounds. The voids FN2 and FN8 have diameters of about 4000 km/sec, so we will still be able to tell if the observed dwarf galaxy is within the void or not.
There are several ways to improve the accuracy of this data. We found that the accuracy is dependent upon several factors, the strength of the hydrogen alpha emission line, the redshift of the galaxy, and signal to noise. We found that in general errors are minimized at the crossover point of the ramp filters, at 668 nm. A galaxy that has a hydrogen alpha line at 668 nm will have a redshift of about 5300 km/sec. Perfectly photometric data is required, as bad signal to noise makes the data practically unusable. In fact, we originally gathered data on 48 different galaxies, but had to drop much of the data due to interference from clouds and other photometric obfuscations.
The continuum turned out to be more difficult to work with than anticipated. When we were developing the method to determine the redshift from the data, we hypothesized that the continuum would behave much more predictably. We took data using a Stromgren Y filter, which peaks at 547 nm, and attempted to use this to create a linear estimate of the continuum at each wavelength. However; this created much higher errors, more on the order of 1000 km/sec. We learned that the continuum is much better predicted by a single filter close to the ramp filters. The signal to noise of the continuum is strongly correlated with the final error in the redshift.
We also discovered that there is a sulphur emission line that often occurs next to the hydrogen alpha line. In the future, it would be helpful to obtain the spectra of the galaxies we have taken data on, and see if we could further improve the accuracy of our predictions taking this into account.
A good way to further improve the accuracy of this method would be to gather many more data points. As already discussed, most our original data had to be thrown out due to photometric conditions on the night it was taken. More data would allow us to further refine the accuracy of the final predicted redshift from the BS/RS ratio.
We have developed a new method to predict the redshift of dwarf galaxies with strong emission lines. While we have already worked out several improvements and refinements, there are several more things that could be done to further improve this method. However; we have shown that this method is a viable way to obtain the redshift of dwarf galaxies with strong emission. This is a method that could be used in a wide scale survey of the voids. These results have been presented to the American Astronomical Society (AAS), and have already resulted in a scientific publication. We anticipate that we will be able to further refine this method, and publish the results in a peer reviewed scientific journal.