Radio frequency (RF) non-thermal (NT) bio-effects have been a subject of debate and attracted significant interests due to potential health risks or beneficial applications. In this work, we report a systematic method for broadband investigation of RF NT effects on Saccharomyces cerevisiae yeast growth. The method includes a transverse electro-magnetic (TEM) device and a dielectric spectroscopy technique for RF frequency selection. A stripline-based TEM device has two 240-μL chambers 3D printed for cell cultures. The fabricated device operates up to a few GHz and produces uniform RF fields for cell exposure testing. A vector network analyzer (VNA) was used to provide -20 dBm continuous-wave (CW) RF power. The heating effects on cell growth were estimated to be negligible. Frequency regions, where large permittivity differences between the medium and yeast cultures were obtained and used to select RF testing frequencies, e.g., 1.0 MHz, 3.162 MHz, 10 MHz. These differences may indicate RF field gradients near cell membrane, and the gradients may affect local nutrient transport. Additionally, RF at 905 MHz is tested for comparison purpose. Yeast cells in the exponential growth phase were examined at four RF frequencies and compared with two controls. One control device held at the same temperature as the test device, while the other control was held at a temperature 1 °C higher. The results showed that the RF fields at 3.162 MHz reduced yeast growth rates by 15.1%; however, the RF fields at 1.0 MHz enhanced cell growth by 13.7%, while the observed 4.3% growth rate increase at 10 MHz is insignificant and the RF fields at 905 MHz had no effects on the cell growth. These results showed a clear RF NT effects on S. cerevisiae growth that was frequency dependent. The hypothesized mechanisms of these effects, i.e., non-uniform RF fields near cell membranes and fluidic diodes in cell membrane ion channels may play important roles in nutrient transport, need to be further investigated.