A new investigation of electron neutrino appearance oscillations with improved sensitivity in the MiniBooNE+ experiment

MiniBooNE has measured a $3.8\sigma$ excess of oscillation candidate events in the combined $\nu_{\mu}$ and $\bar{\nu}_{\mu}$ data sets collected to date at Fermilab [5]. As can be seen in Fig. 2, the predicted backgrounds in the low energy regions, where the excess is most substantial, are dominated by neutral current backgrounds. These NC backgrounds are from two major sources: misidentification of the $\pi^{0}$ (“$\pi^{0}$ misid”) and the production of $\Delta$ baryons which then radiatively decay (“$\Delta\rightarrow N\gamma$”). A test of these NC backgrounds in a measurement with different systematic errors would be quite valuable to firmly establish the oscillation excess.

MiniBooNE can perform this test by detecting neutrons associated with oscillation candidate events. In brief, at low $E_{\nu}^{QE}$, true CC oscillation events (Fig. 3a) should contain final-state neutrons in less than 10% of the events while the NC backgrounds (Figs. 3b,3c) should contain neutrons in $\approx 50\%$ of the events.

Note that $E_{\nu}^{QE}$ is the reconstructed neutrino energy using the assumption of neutrino quasielastic scattering from a neutron. This quantity should be a good estimate for the true neutrino energy in true CC oscillation events (excepting possible nuclear effects, Sec. 2.4). However, because of the large missing energy in NC events, the true neutrino energy is, on average, much higher than $E_{\nu}^{QE}$ for NC backgrounds events. So, in a first approximation, the NC backgrounds will contain more final state neutrons because the events are from higher true neutrino energy. More energy is transferred to the nucleus which causes more neutron production as compared to the CC signal in which the true neutrino energy is lower and (neglecting final state interactions) produces a single proton.

In practice, we would rerun the MiniBooNE oscillation search in neutrino mode after the addition of scintillator in order to enable neutron detection. Oscillation candidates would be selected with the same strategy as the original search. From this sample, we would search for neutron capture events and measure the neutron fraction which would test the NC background estimates. An important feature of this measurement is that the neutron fraction may be “calibrated” for the oscillation search via separate MiniBooNE $\nu_{\mu}$ CCQE and $\nu_{\mu}$ NC$\pi^{0}$ measurements which greatly reduces errors from any nuclear model uncertainties.

The NC neutrino-nucleon elastic scattering (NC elastic) interaction, $\nu N\rightarrow\nu N$, is sensitive to the isoscalar-axial structure of the nucleon [7], so will be sensitive to the effects of strange-quark contributions to the nucleon spin ($\Delta s$). Therefore, the right type of measurement of NC elastic scattering would contribute substantially to the nucleon spin puzzle, an area of continued interest and effort (e.g. [8]). This measurement of NC elastic scattering has not yet been realized.

MiniBooNE has made the most accurate measurement to date of the differential cross section for $\nu N\rightarrow\nu N$ [9] and the analysis for the $\bar{\nu}N\rightarrow\bar{\nu}N$ process is almost complete [10]. While these are valuable measurements to help with understanding of neutrino-nucleon scattering, they are not sensitive to $\Delta s$ because MiniBooNE is not able currently to distinguish between neutrons and protons. The $\nu p\rightarrow\nu p$ process is sensitive to $\Delta s$ with the opposite sign as $\nu n\rightarrow\nu n$ and any strange quark effects cancel in the existing MiniBooNE measurement.

This situation changes abruptly with the addition of neutron-capture tagging. In that case, the neutrons and protons can be separately identified and the neutron/proton ratio,

is quite sensitive to $\Delta s$ [7]. Based on previous studies [11], a rough estimate is that a 10% measurement of $R(p/n)$ should result in an error of $\approx 0.05$ uncertainty on $\Delta s$. It should be realized that the recent results from MiniBooNE on the unexpectedly large CCQE cross section [12] may call into question the theoretical uncertainty involved in extracting $\Delta s$ from $R(p/n)$. If there are multinucleon correlations contributing substantially to NC elastic scattering, it may not be clear how that affects the extraction of $\Delta s$. Regardless, a 10% measurement of $R(p/n)$ will be a valuable constraint and will further more theoretical investigation.

The reaction $\nu_{\mu}~^{12}C\rightarrow\mu^{-}~{}^{12}N_{\textrm{g.s.}}$ is an interesting reaction to study with a scintillator-enhanced MiniBooNE for several reasons. It comes with a distinctive tag from the $\beta$-decay of the ${}^{12}N_{\textrm{g.s.}}$ with endpoint energy of 16.3 MeV and lifetime of 15.9 ms. This addition of scintillator to MiniBooNE will allow for high efficiency and better reconstruction of the $\beta$-decay. Since it is an exclusive reaction, the theoretical cross section can be calculated to $\approx 2\%$ very near threshold [13]. It was measured by LSND for both $\nu_{\mu}$ and $\nu_{e}$ [13, 14] and by KARMEN for $\nu_{e}$ [15] to agree with theory to within experimental errors. A measurement by MiniBooNE of this theoretically well-known reaction would enable a test of the low-energy neutrino flux which could better constrain the low-energy oscillation excess.

The event signature is quite distinct. The low-energy prompt $\mu^{-}$ and subsequent decay $e^{-}$ would be detected with the usual techniques employed for $\nu_{\mu}$ CCQE events combined with a requirement for a detected $\beta$-decay candidate. With the addition of scintillator to make 2.2 MeV $\gamma$ visible, the efficiency for detecting the 16.3 MeV-endpoint $\beta$ will be quite high.

The challenge is that the fraction of the total $\nu_{\mu}$ scattering events that interact via $\nu_{\mu}~^{12}C\rightarrow\mu^{-}~{}^{12}N_{\textrm{g.s.}}$ is small. In the lowest energy bin at $E_{\nu}\approx 250$ MeV, the cross section is about 4% of the $\nu_{\mu}$ CCQE cross section, falling to about 0.5% by $E_{\nu}\approx 400$ MeV [16]. However, with the data sample proposed here the total $\nu_{\mu}$ event sample will be large, the ${}^{12}N_{\textrm{g.s.}}$ signature quite distinct, and an analysis will be worth the effort.

MiniBooNE has reported absolutely normalized cross sections for various $\nu_{\mu}$-carbon processes including $\nu_{\mu}$ CCQE [12], CC$\pi^{+}$ [17], CC$\pi^{0}$ [18], NC elastic [9], and NC$\pi^{0}$ [19]. They all show a 30-40% larger cross section than predicted in previously existing models (e.g. [20]). One emerging idea is that two-nucleon correlations in carbon are contributing significantly to the interaction cross section [21, 22]. If this is the correct explanation for the extra strength in these neutrino interactions, then it could also have a significant effect on the reconstructed neutrino energy in oscillation events, $E_{\nu}^{QE}$, which assumes quasielastic scattering from single nucleons within carbon [23]. In short, the reconstructed neutrino energy may be incorrect in a large fraction of the oscillation events leading to incorrect conclusions about the resulting fits to oscillation models.

The addition of scintillator will allow this idea to be experimentally tested. With the scintillator addition proposed here, the detector response to final state nucleons in a typical CCQE event will be increased by about a factor of five. This scintillation light is a measure of the total energy in the event ($E_{\nu}^{total}$) as opposed to that reconstructed from the just the lepton track, $E_{\nu}^{QE}$. A comparison of $E_{\nu}^{QE}$ with $E_{\nu}^{total}$ will allow further insight into the two-nucleon correlation issue in general and, specifically, into its relevance to the low-energy oscillation excess.

As of this writing, we plan to add 300 kg of PPO to the $1\times 10^{6}$ liters of MiniBooNE mineral oil (300 mg/l). A preliminary price quote for PPO from the supplier to NOvA is $250/kg or $75k for scintillator. The solubility of PPO will allow us to add the entire 300 kg to the MiniBooNE 10 kl overflow and then introduce that into the main volume by recirculation. However, it would be prudent to do this addition in at least two steps by taking the concentration to about 50% of the desired amount and monitoring detector response with cosmic muons and muon-decay electrons. We may do recirculation without the addition of scintillator as a first step, as the MiniBooNE oil has not been recirculated since commissioning in 2002.

Our base plan for running with scintillator is only to add scintillator with no other changes. New readout electronics could be considered, but are not required for the physics goals set here. The current rate of PMT and failures extrapolated for a 3-year run is not a problem. We estimate that the rate of electronics failures over that time period will be covered with our current supply of spares. There will likely be some changes to the computing infrastructure to keep up with hardware failures and security concerns, but an “as-needed” approach is our current plan.

The neutron fraction measurement for oscillation candidates is statistics limited and $6.5\times 10^{20}$ POT is required for our current desired accuracy. When the MicroBooNE experiment is running, our assumption is that $2\times 10^{20}$ POT/year will be delivered to the Booster Neutrino Beamline (BNB). This sets a 3-year duration for the proposed scintillator phase of MiniBooNE. Preparation for running only requires the addition of scintillator (along with modest detector maintenance), which we estimate will require about 3 months with no beam requirement.