|Abstract: ||When brown dwarfs were first discovered two decades ago, they were introduced as the missing link between stars and giant planets. This led to the extension of the stellar classification sequence down to the planetary regime by introducing the spectral types L,T and Y. We are now in the era of brown dwarf science where large homogeneous samples can be produced, allowing us to learn about their physical properties in a statistically meaningful way.
In this thesis we present a new method, dubbed photo-type, which makes the classification of L0-T8 dwarfs significantly faster as it only relies on already available multi-wavelength photometry. It is based on χ2 template fitting and can be adapted to use on any photometric data. photo-type classifies L and T dwarfs to an accuracy better than 1 spectral type, making it comparable to spectroscopy. This is tested by a range of different methods, ranging from simulations to spectroscopic follow up.
By searching the SDSS+UKIDSS+WISE for L and T dwarfs with 13.0 < J < 17.5 we produce the largest homogeneous sample of these dwarfs to date. These 1157 L and T dwarfs are used to find peculiar objects and populations (e.g. blue/red outliers), L/T binaries, and benchmark systems (by crossmatching with a proper motion catalogue). It also allows us to calculate the best estimates of the space densities of L and T dwarfs to date. By following up several of the peculiar ultra cool dwarfs in our sample, we found a very young object (10Myrs), a number of dusty sources, and some that remain unexplained.
Finally, we conducted a general search for rare objects in UKIDSS. Here we were interested in finding high redshift quasars, very cold white dwarfs or a yet unknown population by analysing objects with large χ2 values. While none of the above were found, we did find a moderate temperature H and He atmosphere white dwarf, a carbon star and some medium redshift quasars.|