Golden channel at a neutrino factory revisited: Improved sensitivities from a magnetized iron neutrino detector

This paper describes the performance and sensitivity to neutrino mixing parameters of a Magnetised Iron Neutrino Detector at a Neutrino Factory with a neutrino beam created from the decay of 10 GeV muons. Specifically, it is concerned with the ability of such a detector to detect muons of the opposite sign to those stored (wrong-sign muons) while suppressing contamination of the signal from the interactions of other neutrino species in the beam. A new, more realistic simulation and analysis, which improves the efficiency of this detector at low energies, has been developed using the GENIE neutrino event generator and the GEANT4 simulation toolkit. Low-energy neutrino events down to 1 GeV were selected, while reducing backgrounds to the ${10}^{-4}$ level. Signal efficiency plateaus of $\sim 60\%$ for ${\nu }_{\mu }$ and $\sim 70\%$ for ${\overline{\nu }}_{\mu }$ events were achieved starting at $\sim 5\mathrm{GeV}$. Contamination from the ${\nu }_{\mu }\to {\nu }_{\tau }$ oscillation channel was studied for the first time and was found to be at the level between 1% and 4%. Full response matrices are supplied for all the signal and background channels from 1 GeV to 10 GeV. The sensitivity of an experiment involving a Magnetised Iron Neutrino Detector detector of 100 ktons at 2000 km from the Neutrino Factory is calculated for the case of ${sin}^{2}2{\theta }_{13}\sim {10}^{-1}$. For this value of ${\theta }_{13}$, the accuracy in the measurement of the $CP$-violating phase is estimated to be $\Delta {\delta }_{CP}\sim 3°–5°$, depending on the value of ${\delta }_{CP}$, the $CP$ coverage at $5\sigma$ is 85% and the mass hierarchy would be determined with better than $5\sigma$ level for all values of ${\delta }_{CP}$.

Figure 1

Proportion of total number of interactions of different $\nu$ interaction processes for events generated using GENIE and passed to the G4MIND simulation.

Figure 2

The digitization and voxel clustering of an example event: (top left) bending plane view, (top right) nonbending plane, (bottom) an individual scintillator plane. The individual hits are small dots (in red), the blue squares are the voxels and the black asterisks represent the centroid positions of the clusters.

Figure 3

Muon candidate hit purity for ${\nu }_{\mu }$ CC (top) and ${\overline{\nu }}_{\mu }$ CC (bottom) interactions extracted using (left) Kalman filter method and (right) cellular automaton method.

Figure 4

Pull on the reconstructed momentum (the difference between the true and reconstructed momentum divided by the measured error) (left) and momentum resolution (right).

Figure 5

Distribution of the proportion of clusters fitted in the trajectory for ${\nu }_{\mu }$ appearance (left) and ${\overline{\nu }}_{\mu }$ appearance (right), normalized to total remaining events individually for each interaction type.

Figure 6

Log-likelihood distribution (${\mathcal{L}}_{q/p}$) to separate wrong-sign muons for signal and background for ${\nu }_{\mu }$ (left) and ${\overline{\nu }}_{\mu }$ (right) appearance experiments.

Figure 7

Distributions of displacement and momentum with cut levels: (top left) relative displacement in the bending plane to the $z$ direction against candidate hits for signal events, (top right) reconstructed momentum against displacement in $z$ for signal events and (bottom) as top for ${\nu }_{\mu }$ (${\overline{\nu }}_{\mu }$) CC backgrounds. The red lines represent the cuts from Eqs. (5, 6).

Figure 8

Distribution of the $q{p}_{\mathrm{par}}$ variable, with the region where the parameter is $<0$ representing those candidates fitted with charge opposite to the initial Kalman filter for ${\nu }_{\mu }$ (left) and ${\overline{\nu }}_{\mu }$ appearance experiments. The distributions are normalized to the total remaining events, individually, for each interaction type.

Figure 9

Distribution of the number of fit clusters in candidate track used to calculate the log-likelihood-based charged current selection.

Figure 10

Distribution of ${\mathcal{L}}_1$ likelihood ratios used to reject NC and other background signals.

Figure 11

Distributions of kinematic variables: (left) reconstructed muon momentum with reconstructed neutrino energy for ($\mathrm{\text{top}}\to \mathrm{\text{bottom}}$) ${\nu }_{\mu }$ (${\overline{\nu }}_{\mu }$) signal, ${\nu }_{\mu }$ (${\overline{\nu }}_{\mu }$) CC background, NC background, ${\nu }_e$ (${\overline{\nu }}_e$) CC background and (right) $Q_t$ variable (in the same order). The red lines represent the cuts from Eqs. (9, 10).

Figure 12

Background from misidentification of ${\nu }_{\mu }$ (${\overline{\nu }}_{\mu }$) CC interactions as the opposite polarity. (left) ${\overline{\nu }}_{\mu }$ CC reconstructed as ${\nu }_{\mu }$ CC, (right) ${\nu }_{\mu }$ CC reconstructed as ${\overline{\nu }}_{\mu }$ CC as a function of true energy.

Figure 13

Background from misidentification of NC interactions as ${\nu }_{\mu }$ (${\overline{\nu }}_{\mu }$) CC interactions. (left) NC reconstructed as ${\nu }_{\mu }$ CC, (right) NC reconstructed as ${\overline{\nu }}_{\mu }$ CC as a function of true energy.

Figure 14

Background from misidentification of ${\nu }_e$ (${\overline{\nu }}_e$) CC interactions as ${\nu }_{\mu }$ (${\overline{\nu }}_{\mu }$) CC interactions. (left) ${\nu }_e$ CC reconstructed as ${\nu }_{\mu }$ CC, (right) ${\overline{\nu }}_e$ CC reconstructed as ${\overline{\nu }}_{\mu }$ CC as a function of true energy.

Figure 15

Efficiency of reconstruction of ${\nu }_{\mu }$ (${\overline{\nu }}_{\mu }$) CC interactions. (left) ${\nu }_{\mu }$ CC efficiency, (right) ${\overline{\nu }}_{\mu }$ CC efficiency as a function of true energy.

Figure 16

${\nu }_{\mu }$ CC and ${\overline{\nu }}_{\mu }$ CC signal detection efficiency as a function of $y$ (left) and the normalized distribution of all events considered in each polarity as a function of $y$ (right).

Figure 17

Expected level of contamination from ${\nu }_{\tau }$ (${\overline{\nu }}_{\tau }$) CC interactions due to the platinum channel. (left) ${\nu }_{\tau }$ CC reconstructed as ${\nu }_{\mu }$ CC, (right) ${\overline{\nu }}_{\tau }$ CC reconstructed as ${\overline{\nu }}_{\mu }$ CC as a function of true energy.

Figure 18

Expected level of contamination from ${\nu }_{\tau }$ (${\overline{\nu }}_{\tau }$) CC interactions due to the dominant oscillation. (left) ${\overline{\nu }}_{\tau }$ CC reconstructed as ${\nu }_{\mu }$ CC, (right) ${\nu }_{\tau }$ CC reconstructed as ${\overline{\nu }}_{\mu }$ CC as a function of true energy.

Figure 19

Expected interactions in a 100 kton MIND 2000 km from the Neutrino Factory for ${\theta }_{13}=9.0°$. Left column for ${\nu }_{\mu }$ appearance and right column for ${\overline{\nu }}_{\mu }$ appearance.

Figure 20

Efficiencies for a pure DIS sample compared to the nominal case. (top) Signal efficiency, (second line) ${\nu }_{\mu }$ (${\overline{\nu }}_{\mu }$) CC background, (third line) NC background and (bottom) ${\nu }_e$ (${\overline{\nu }}_e$) CC background. ${\nu }_{\mu }$ appearance on the left and ${\overline{\nu }}_{\mu }$ appearance on the right.

Figure 21

Calculated error on signal efficiencies on increasing (top) and decreasing (bottom) the proportion of non-DIS interactions in the data-set. (left) Errors on true energy ${\nu }_{\mu }$ CC efficiency and (right) errors on true energy ${\overline{\nu }}_{\mu }$ CC efficiency

Figure 22

Variation of signal efficiency (top) and NC backgrounds (bottom) due to a 6% variation in the hadron shower energy and direction resolution and a more pessimistic 50% reduction in angular resolution (focused on region of greatest variation).

Figure 23

The Golden channel analysis efficiencies from simulations of charge current neutrino events in a MIND detector generated using GENIE and NUANCE neutrino generators. The differences between the efficiencies as fractions of the GENIE efficiency are also shown.

Figure 24

Examples of $1\sigma$, $3\sigma$ and $5\sigma$ fits to simulated data for the 10 GeV Neutrino Factory, for ${sin}^{2}2{\theta }_{13}=0.096$.

Figure 25

${\delta }_{CP}$ $3\sigma$, $5\sigma$ and $10\sigma$ measurements (top) and ${\delta }_{CP}$ $5\sigma$ coverage to measure $CP$ violation (bottom) as a function of ${sin}^{2}2{\theta }_{13}$ for true normal hierarchy (left) and true inverted hierarchy (right). The vertical lines represent the range of possible values of ${sin}^{2}2{\theta }_{13}$.

Figure 26

Left: precision in the value of ${\delta }_{CP}$ as a function of ${\delta }_{CP}$ under two different assumptions: top line (3.0% normalisation and 2.5% cross-section error), bottom line (1% normalization and cross-section error). Right: ${\delta }_{CP}$ fraction coverage that can be achieved above each value of $\Delta {\delta }_{CP}$ for 3.0% and 2.5% errors (right curve) and for 1% and 1% errors (left curve).