Sensitivity of neutrino oscillations to the Earth's interior properties

TL;DR

This work investigates how atmospheric neutrino oscillations respond to radial variations in Earth's electron density, proposing neutrino oscillation tomography (NOTE) as a complementary approach to seismology. Using the EarthProbe framework, the authors model atmospheric fluxes, Earth traversal with a PREM-informed electron-density profile, and detector responses to predict energy–angle event distributions for current and next-generation detectors. A focused sensitivity study shows that core electron-density changes of order 1% could be detectable within a few decades with KM3NeT/ORCA–like or HK/DUNE–like detectors, while Mantle Transition Zone variations remain substantially harder to resolve; combining neutrino data from multiple observatories could improve constraints. The work lays groundwork for joint neutrino–seismic inversions to achieve improved 3D imaging of the Earth’s interior and outlines detector-design directions to maximize sensitivity to deep Earth regions.

Abstract

Understanding the Earth s internal structure remains a major challenge, as traditional geophysical methods face ambiguities in linking seismic observations to temperature, composition, or mass density variations. Atmospheric neutrinos offer a complementary probe: while traversing the Earth, they undergo flavor oscillations that depend on the local electron density, which reflects both mass density and composition.

Figures (6)

• Figure 1: Overview of the data-taking configurations and data visualizations produced in NOTE and in seismology: (a) NOTE uses neutrinos that are produced almost isotropically in the atmosphere (stars) and registered at a single neutrino detector (rectangle). The baseline corresponding to the incidence angle $\theta_z$ is shown as a bold black line. (b) In seismology, one earthquake source (star) is recorded at several seismic stations positioned at various distances from the event (triangles). The epicentral distance $\Delta$ (defined as the angular distance between the source and the receiver) is shown in white. (c) Oscillogram representing the $\nu_\mu \rightarrow \nu_\tau$ oscillation probability for a neutrino crossing the Earth (assuming PREM), as a function of neutrino energy $E_\nu$ and the cosine of the incidence angle $\theta_z$. The corresponding epicentral distance is displayed on the vertical axis on the right. (d) Time–distance plot obtained from the envelope of synthetic seismograms computed at a dominant period of 5 using the PREM model for a double-couple source located at a depth of 30k. The envelope of each seismogram represents the energy as a function of time. Each horizontal line in the plot corresponds to one seismogram and is normalized to its maximum.
• Figure 2: Differential flux of atmospheric electron neutrinos (blue shades) and muon neutrinos (red shades) from honda2015atmospheric, averaged over all azimuthal angles. Further details on the neutrino flux are provided in the text. Horizontal neutrinos (solid lines) correspond to $\cos\theta_z$ values between $-0.1$ and $0.0$, while up-going neutrinos (dashed lines) span the range from $-1.0$ to $-0.9$.
• Figure 3: (a) Electron density as defined in Equation \ref{['eq:n_e']} (solid blue line), based on the density model from PREM and the chemical composition from GERM, along with the effects of increasing or decreasing the density by 3% (blue shaded band). The red dash-dotted curve represents an alternative Earth model in which only the density of layer 30 (covering the depths between $\sim$3200k and $\sim$3300k) is increased by 3% relative to the reference Earth model. (b) Transition probability P$_{\nu_\mu \rightarrow \nu_e}$ for PREM (solid blue line) and for the alternative Earth model in (a) (red dash-dotted line), together with the difference between the resulting probability functions in the lower panel. Only core-crossing neutrinos with $\cos\theta_z = -0.89$ are considered.
• Figure 4: (a) Simulated rate of interacting events in the $(\nu_\mu + \bar{\nu}_\mu)$ CC channels, as described by Equation \ref{['eq:interactingrate']} (equivalent to the number of $(\nu_\mu + \bar{\nu}_\mu)$ CC interactions, as defined in Equation \ref{['eq:interactingnumber']}, corresponding to 1 Mton$\cdot$years of exposure of a detector), and multiplied by $E_\nu^2$ to visually compensate for the rapid decrease of the neutrino flux with energy. (b) Expected number of detected track-like events in a KM3NeT/ORCA-like detector with $10$ years of data-taking time.
• Figure 5: a) Sensitivity of two hypothetical next-generation detectors to a 3% density increase in a $\sim$100km-thick layer as a function of layer depth, for an exposure of 200 Mton$\cdot$years. Results are shown for a detector capable of identifying all interaction channels (dashed line) and for one that can only, but perfectly, distinguish track-like and shower-like events (solid line). Further details of these hypothetical detectors are provided in the text. Horizontal lines indicate the upper boundaries of selected Earth structures of interest. b) Sensitivity of the DUNE-, HK-, and KM3NeT/ORCA-like detectors described in Section \ref{['subsec:reconstructed']} to a $3\%$ increase in density within a 100km-thick layer, shown as a function of layer depth. A data-taking period of $20$ years is assumed for all cases, corresponding to exposures of $0.8$, $8$, and $160$ Mton$\cdot$years, respectively, reflecting the different detector sizes ($0.04$, $0.4$, and $8$ Mton, respectively).