The DUNE Near Detector
5.2 Role of the ND in the DUNE Oscillation Program
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 (GeV) νµ
Energy 0
10 20 30 40 50
−9
×10
/POT2 ) at 574m/GeV/cm µν(Φ On-axis6m
12m18m24m30m36m
-mode ν
-mode ν
-mode 33m Off-axis ν
Figure 5.4: The variation in the neutrino energy spectrum shown as a function of detector off-axis position, assuming the nominal ND location 574 m downstream from the production target.
teraction systematic errors that are different from ArgonCube, is an important point of comparison and a systematic cross-check for the flux as determined by ArgonCube.
SAND provides very fast timing and can isolate small energy depositions from neutrons in three dimensions. This provides the capability to incorporate neutrons in the event reconstruction using energy determination via time-of-flight with a high efficiency. This capability should be useful for the low-ν flux determination1 because it either allows events to be tagged with a significant neutron energy component or provides a way to include that energy in the calculation. Including neutrons in detailed studies of neutrino interactions in SAND using single transverse variables may prove useful in motivating improvements in the neutrino interaction model. Although the target for this device is carbon, not argon, basic insights into components of the interaction model may extend to argon. For example, the multi-nucleon component of the interaction model will be used for argon although it was developed in response to observations made on plastic targets.
(GeV) Eν
0 1 2 3 4 5 6
per POT)-2 cm-1 (GeVνΦ
0 5 10 15 20 25 30 35 40
−9
×10
eV2
10-3
× = 0.0022
23
m2
∆ ) = 0.5;
θ23 2( sin
Oscillated FD Flux Composite ND Flux Fit region
σ
± Decay pipe radius
σ
± Horn current
σ
± Water layer
[GeV]
Eν
0 1 2 3 4 5 6
FD (unosc.)ND - FD (osc.)
−0.4
−0.2 0 0.2 0.4
(GeV) Eν
0 1 2 3 4 5 6
per POT)-2 cm-1 (GeVνΦ
0 5 10 15 20 25 30 35 40
−9
×10
eV2 10-3
× = 0.0025 23 m2
∆ ) = 0.65;
θ23 2( sin
Oscillated FD Flux Composite ND Flux Fit region
σ
± Decay pipe radius
σ
± Horn current
σ
± Water layer
[GeV]
Eν
0 1 2 3 4 5 6
FD (unosc.)ND - FD (osc.)
−0.4
−0.2 0 0.2 0.4
Figure 5.5: Linear combinations of off-axis fluxes giving FD oscillated spectra for a range of oscillation parameters. The FD oscillated flux is shown in black, the target flux is shown in green, and the linearly combined flux obtained with the nominal beam Monte Carlo (MC) is shown in red. Systematic effects due to 1σ variations of the decay pipe radius (green), horn current (magenta), and horn cooling water layer thickness (teal) are also shown.
Figure 5.6: The SAND detector configuration with the 3DST inside the KLOE magnet. The drawing shows the 3DST in the center (white), TPCs (magenta), ECAL (green), magnet coil (yellow), and the return yoke (gray).
constraints from data observed in the ND. That comparison and how it varies with the oscillation parameters allows oscillation parameters to be measured.
The connection between the observations in the ND and the FD is made using a simulation that convolves models of the neutrino flux, neutrino interactions, nuclear effects, and detector response. This gives rise to a host of complicating effects that muddy the simple picture. These complications come from two main sources. First, the identification efficiency is not 100 %, and there are some background events (for example, NC interactions with a π0 present a background to νe CC interactions). Both the efficiency and background are imperfectly known. Because the background level tends to be similar in both the FD and ND, it helps if the ND can characterize backgrounds better than the FD.
The second aspect that complicates the simple picture is that the FD (and the similar ND) must use a target material composed of heavy nuclei. The target nucleus affects neutrino interactions in ways that ultimately drive the design of the ND complex. In particular, in heavy nuclei, the nucleons interact with each other and exhibit Fermi motion, providing moving targets for neutrino interactions. The wavelength of an interaction depends on momentum transfer but is often long enough to simultaneously probe multiple nucleons.
Another complication is that neutrino-nucleus scattering models rely on neutrino-nucleus cross sections, but neutrino cross sections onfree nucleons are not generally well known in the kinematic range of interest to DUNE. Since the ND will enable high-statistics measurements on liquid and gaseous argon, rather than another nucleus, it will reduce nuclear model dependence. A final complication comes about because neutrinos produce hadrons within the nucleus. After production the hadrons undergo final-state interactions (FSI) and are thereby attenuated as they leave the target nucleus. Section A.2 of Appendix A discusses neutrino-nucleus scattering in more detail.
Neutrons can be produced from the struck nucleus, as well as from follow-on interactions of the neutrino’s reaction-products with other nuclei. The energy carried away by neutrons is difficult to detect and can bias the reconstructed neutrino energy. The SAND and MPD detectors have capabilities that allow neutron energy to be directly measured. The DUNE-PRISM program constrains the true-to-reconstructed energy relation and is thus also sensitive to energy carried by neutrons.
Heavy nuclei in the detector offer additional complications for particles that have left the struck nucleus, especially in the case where the detector is dense, e.g., in ArgonCube. Particles produced in a neutrino interaction may reinteract inside the detector, creating electromagnetic and hadronic cascades. These cascades, particularly the hadronic ones, confuse the reconstruction program due to overlapping energy and event features. They also cause a degradation of the energy resolution and result in additional energy carried by neutrons that may go missing. Particle identification by dE/dx is less effective for early showering particles, and low-energy particle tracks in a dense detector may be too short to detect. The HPgTPC in the MPD allows us to measure neutrino interactions on argon, but with significantly fewer secondary interactions and much lower-energy tracking thresholds.
Finally, setting aside complications due to heavy nuclei and dense detectors, we note that a sig- nificant fraction of the neutrino interactions in DUNE will come from inelastic processes, not the
simpler quasi-elastic (QE) scattering. This typically leads to a more complex morphology for events and greater challenges for the detector and the modeling. The DUNE ND acts as a control for the FD and is designed to be more capable than the FD at measuring complicated inelastic events.
These complexities are incorporated imperfectly into the neutrino interaction model. The predicted signal in the ND is a convolution of this interaction model with the beam model and the detector response model. The critical role of the ND is to supply the observations used to tune, or calibrate, this convolved model, thereby reducing the overall uncertainty in the expected signal at the FD, which is used for extracting the oscillation parameters via comparison with the observed signal.
And with its high statistics and very capable subsystems, the ND will produce data sets that will provide the raw material for improving the models beyond simple tuning.