We present serendipitous observations of radio sources at 28.5 GHz ( 1 cm ) , which resulted from our program to image thermal Sunyaev-Zeldovich ( SZ ) effect in 56 galaxy clusters .
In a total area of \sim 0.8 ^ { \circ } sq. , we find 64 radio sources with fluxes down to \sim 0.4 mJy ( > 4 \sigma ) , and within 250 ^ { \prime \prime } from the pointing centers .
The spectral indices ( S \propto \nu ^ { - \alpha } ) of 54 sources with published low frequency flux densities range from -0.6 \lesssim \alpha \lesssim 2 with a mean of 0.77 \pm 0.06 , and a median of 0.84 .
Extending low frequency surveys of radio sources towards galaxy clusters CL 0016+16 , Abell 665 , and Abell 2218 to 28.5 GHz , and selecting sources with S _ { 1.4 { GHz } } \geq 7 mJy to form an unbiased sample , we find a mean spectral index of 0.71 \pm 0.08 and a median of 0.71 .
We find 4 to 7 times more sources predicted from a low frequency survey in areas without galaxy clusters .
This excess can not be accounted for by gravitational lensing of a background radio population by cluster potentials , indicating most of the detected sources are associated with galaxy clusters .
The differential source count slope , \gamma \sim 1.96 ( dN / dS \propto S ^ { - \gamma } ) , is flatter than what is expected for a nonevolving Euclidean population ( \gamma = 2.5 ) .
For the cluster Abell 2218 , the presence of unsubtracted radio sources with S _ { 28.5 { GHz } } \leq 0.5 mJy ( \sim 5 \sigma ) , can only contribute to temperature fluctuations at a level of \Delta T \sim 10 to 25 \mu K. The corresponding error due to radio point source contamination in the Hubble constant derived through a combined analysis of 28.5 GHz SZ images and X-ray emission observations ranges from 1 % to 6 % .