Long-Baseline Horizontal Radio-Frequency Transmission Through Polar Ice

We report on analysis of englacial radio-frequency (RF) pulser data received over horizontal baselines of 1-5 km, based on broadcasts from two sets of transmitters deployed to depths of up to 1500 meters at the South Pole. First, we analyze data collected using two RF bicone transmitters 1400 meters below the ice surface, and frozen into boreholes drilled for the IceCube experiment in 2011. Additionally, in Dec., 2018, a fat-dipole antenna, fed by one of three high-voltage (O(1 kV)), fast (O(1-5 ns risetime)) signal generators was lowered into the 1700-m deep icehole drilled for the South Pole Ice Core Experiment (SPICE), approximately 3 km from the geographic South Pole. Signals from transmitters were recorded on the five englacial multi-receiver ARA stations, with receiver depths between 60-200 m. From analysis of deep transmitter data, we estimate: i) the range of refractive index profiles of Antarctic ice with depth allowed by our measurements, ii) due to birefringence, a time difference between arrival times for vertically polarized vs. horizontally polarized signals (per km) for horizontally propagating signal, and iii) for the first time, the attenuation length for electromagnetic signals in the radio-frequency regime broadcast horizontally (rather than reflected vertically from bedrock). We additionally present data suggesting anomalous ice propagation effects, and contrary to expectations for a transport medium with a smoothly varying refractive index profile.Our results imply negligible uncertainty in overall neutrino detection volume due to refractive index uncertainties. Our birefringence time-difference measurements are fit to the functional form delta(t)(H - V)[ns=km] = a cos theta + b, with H/V the signal arrival times for the horizontally/vertically polarized EM signal components, and theta the opening angle in the horizontal plane between the signal propagation direction and the local ice flow direction, extracting a=8.3 +/- 1.3 ns/km, and b=-8.6 +/- 0.9 ns/km (errors combined statistical and systematic), allowing a similar to 15% range estimate for future measurements of in-ice neutrino interactions. Finally, we find attenuation length values clustering around 1.5 km, with measurements from the bicone transmitters yielding L-atten = 1.43 +/- 0.25 +/- 0.37 km. Taken together, these measurements support cold polar ice as a near-optimal platform for ultra-high energy neutrino detectors.