Magnetic Monopole Noise

Magnetic monopoles are hypothetical elementary particles exhibiting quantized magnetic charge $m_0=\pm(h/\mu_0e)$ and quantized magnetic flux $\Phi_0=\pm h/e$. A classic proposal for detecting such magnetic charges is to measure the quantized jump in magnetic flux $\Phi$ threading the loop of a superconducting quantum interference device (SQUID) when a monopole passes through it. Naturally, with the theoretical discovery that a plasma of emergent magnetic charges should exist in several lanthanide-pyrochlore magnetic insulators, including Dy$_2$Ti$_2$O$_7$, this SQUID technique was proposed for their direct detection. Experimentally, this has proven extremely challenging because of the high number density, and the generation-recombination (GR) fluctuations, of the monopole plasma. Recently, however, theoretical advances have allowed the spectral density of magnetic-flux noise $S_{\Phi}(\omega,T)$ due to GR fluctuations of $\pm m_*$ magnetic charge pairs to be determined. These theories present a sequence of strikingly clear predictions for the magnetic-flux noise signature of emergent magnetic monopoles. Here we report development of a high-sensitivity, SQUID based flux-noise spectrometer, and consequent measurements of the frequency and temperature dependence of $S_{\Phi}(\omega,T)$ for Dy$_2$Ti$_2$O$_7$ samples. Virtually all the elements of $S_{\Phi}(\omega,T)$ predicted for a magnetic monopole plasma, including the existence of intense magnetization noise and its characteristic frequency and temperature dependence, are detected directly. Moreover, comparisons of simulated and measured correlation functions $C_{\Phi}(t)$ of the magnetic-flux noise $\Phi(t)$ imply that the motion of magnetic charges is strongly correlated because traversal of the same trajectory by two magnetic charges of same sign is forbidden.