Latest results from Daya Bay

Transcription

1 Journal of Physics: Conference Series PAPER OPEN ACCESS Latest results from Daya Bay To cite this article: Vit Vorobel and Daya Bay Collaboration 2017 J. Phys.: Conf. Ser View the article online for updates and enhancements. Related content - Recent Results from the Daya Bay Experiment Z Y Yu and Daya Bay collaboration - Energy response model of the Daya Bay experiment Nicolás Viaux and Daya Bay Collaboration - Calibration and Reconstruction of the Daya Bay Antineutrino Detector Bei-Zhen Hu and Daya Bay Collaboration This content was downloaded from IP address on 09/02/2018 at 13:11

2 Latest results from Daya Bay Vit Vorobel on behalf of Daya Bay Collaboration Charles University, Faculty of Mathematics and Physics, Prague, Czech Republic Abstract. The Daya Bay Reactor Neutrino Experiment was designed to measure θ 13, the smallest mixing angle in the three-neutrino mixing framework, with unprecedented precision. The experiment consists of eight functionally identical detectors placed underground at different baselines from three pairs of nuclear reactors in South China. Since Dec. 2011, the experiment has been running stably for more than 4 years, and has collected the largest reactor anti-neutrino sample to date. Daya Bay is able to greatly improve the precision on θ 13 and to make an independent measurement of the effective mass splitting in the electron antineutrino disappearance channel. Daya Bay can also perform a number of other precise measurements, such as a high-statistics determination of the absolute reactor antineutrino flux and spectrum, as well as a search for sterile neutrino mixing, among others. The most recent results from Daya Bay are discussed in this paper, as well as the current status and future prospects of the experiment. 1. Introduction The big progress in neutrino research during the last two decades brought a lot of experimental data which strongly imply that neutrinos undergo mixing between flavor and mass eigenstates and hence carry non-zero mass. It also represents the first evidence of physics beyond the Standard Model of particle physics. Since its discovery, the neutrino flavor oscillations have been confirmed and precisely measured with data from natural (atmospheric and solar) and man-made (reactor and accelerator) neutrino sources. The phenomena of neutrino flavor oscillation arise from two remarkable characteristics of neutrinos: small differences between different masses of the three neutrino states, m 1 m 2 m 3, and an inequivalence between neutrino flavor and mass eigenstates. Produced in a flavor eigenstate by the weak interaction, a neutrino state evolves as a coherent superposition of mass eigenstates. Interference between the phases of each mass component results in the periodic oscillation of the neutrino flavor. The flavor oscillates with phases given by m 2 ji L/4E ν, where L is the propagation distance, E ν is the neutrino energy, and m 2 ji = m 2 j m 2 i is the difference of the squared masses. The amplitude of flavor oscillation is determined by the amount of mixing between the flavor and mass eigenstates. Using the unitary Pontecorvo-Maki-Nakagawa-Sakata (PMNS) matrix, U PMNS, a neutrino with flavor α can be expressed as a combination of mass states, ν α = i U αi ν i. In the three-flavor model, U PMNS is commonly parameterized using three mixing angles, θ 12, θ 23, θ 13 and an off-diagonal CP-violating phase δ CP. With sensitivity to the small neutrino mass separations, oscillation experiments have provided strong evidence for three distinct neutrino mass states. Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. Published under licence by Ltd 1

3 2. The Daya Bay Neutrino Experiment The Daya Bay Reactor Neutrino Experiment was designed to measure θ 13, the smallest mixing angle in the three-neutrino mixing framework, with unprecedented precision [1]. The experiment profits from rare constellation of a nuclear power station complex situated near Hong Kong and adjacent mountains. The reactors serve as the neutrinos source while the mountains provide a sufficient overburden suppressing cosmic muons the strongest background source (see Fig. 1). The Daya Bay and Ling Ao nuclear power plant (NPP) reactors (red circles) were situated on a narrow coastal shelf between the Daya Bay coastline and inland mountains. At the time of the measurement the facility consisted of six pressurized water reactors (PWRs). The electron antineutrinos are emitted in beta-decay of fission fragments released in the chain reaction. The antineutrino flux and the energy spectrum is determined by the total thermal power of the reactor, the fraction of each fissile isotope in the fuel, the fission rate of each isotope and the energy spectrum of neutrinos from each isotope. All the reactors have the same thermal power 2.9 GW th each and all together produced roughly ν e /s with energies up to 8 MeV making it one of the most intense ν e sources on Earth. Two antineutrino detectors installed in each underground experimental hall near to the reactors (Hall 1 and Hall 2) measured the ν e flux emitted by the reactors, while four detectors in the far experimental hall (Hall 3) measured a deficit in the ν e flux due to oscillation in the location where the neutrino oscillation effect is expected to be the strongest. Such configuration allows to suppress reactor related uncertainty in the measured neutrino flux. The disappearance signal is most pronounced at the first oscillation minimum. Based on existing accelerator and atmospheric neutrino oscillation measurements this corresponded to a distance L f 1.6 km for reactor ν e with a mean energy of ~4 MeV. The detectors were built and initially tested in a surface assembly building (SAB), transported to a Liquid Scintillator hall for filling, and then installed in an experimental hall. Fig. 1. Layout of the Daya Bay Neutrino Experiment. 2

4 Detection of antineutrinos is based on the same principle as in the famous experiment of Reines and Cowan [2] who registered reactor antineutrinos in A sensitive detector part consist of a hydrogen rich liquid scintillator doped with gadolinium (Reines and Cowan used Cd instead as the dopant). Antineutrino interacts via inverse beta-decay (IBD) ν e + p e + + n with a proton (hydrogen) giving rise to neutron and positron. The positron deposits its kinetic energy to the scintillator then annihilates on electron and generates two gamma quanta 511 kev which together with the deposited positron kinetic energy cause a prompt signal within a few nanoseconds. The neutron is first moderated then is captured on Gd typically ~30 ns after the prompt signal. Consequently, a cascade of gamma quanta with total energy of 8 MeV is emitted and generates a delayed signal. Appearance of the two signals prompt and delayed is a signature of the antineutrino registration. Fig. 2. Scheme of the antineutrino detector (AD) left, and the near site detection view right. The Daya Bay antineutrino detector modules have an onion like structure (see Fig. 2, left). The innermost volume is filled with 20 tons of the Gd-loaded liquid scintillator (GdLS), serving as the antineutrino target. The second layer the gamma catcher - is filled with additional 20 tons of normal liquid scintillator (LS) which can register most of the gamma energies from neutron capture or positron annihilation. Neutrino interactions in the gamma catcher will not satisfy the trigger since only the signal of neutron-capture on Gd will trigger a neutrino event. The outer-most layer is normal mineral oil (MO) that shields radiation from the PMT glass from entering the fiducial volume. The two inner vessels are fabricated of PMMA which is transparent for optical photons and chemically resistant against the used liquids, the outer-most 5 m by 5 m tank is made of stainless steel and is equipped with inch PMTs. Specular reflectors are located above and below the outer PMMA vessel to improve the light collection uniformity while the vertical wall of the detector is black. Three automated calibration units are used to deploy radioactive sources ( 60 Co, 68 Ge, and 241 Am- 13 C) and light-emitting diodes through narrow teflonbellow penetrations into the GdLS and LS regions. After filling, the antineutrino detectors were installed in a 10 m deep water pool in each underground experimental hall, as shown in Fig. 2, right. The water shielded the detectors from γ-rays arising from natural radioactivity and muon-induced neutrons, which were primarily emanating from the cavern rock walls. The pool was optically separated into two independent regions, the inner (IWS) and outer water 3

5 shield (OWS). Both regions were instrumented with PMTs to detect the Cherenkov light produced by cosmogenic muons. A 4-layer resistive plate chamber (RPC) system was installed over the pool, which served in studies of muons and muon-induced backgrounds. Identification of muons which passed through the IWS, OWS, and RPC system enhanced the rejection of background from neutrons generated by muon interactions in the immediate vicinity of the antineutrino detectors. Each detector (ADs, IWS, OWS) operated as an independently triggered system. 3. Detector response The purpose of calibration was to refine the recorded digital times (TDCs) and charges (ADCs) so that they provided accurate, consistent, and stable estimates of the time, energy, and position of particle interactions within the detectors. The time at which each detector triggered was recorded with 25 ns precision. A GPS-synchronized trigger time stamp (25 ns resolution) provided a measure of the absolute time for each triggered event. Fast-pulsed calibration LEDs were used to measure the relative time responses of the PMTs within a single detector. The time delays observed in each channel were corrected for LED-to-PMT distances and were fitted as a function of light intensity. The results were recorded to a database and used to correct TDC values during data analysis. The timing calibration was repeated whenever a modification was made to a detector system. In order to provide proper information about the deposited energy, the registered signal amplitude was corrected for the individual PMT gain, number of active PMTs, spatial non-uniformity of the scintillation light collection and light yield. The energy resolution of the detectors were adequately modelled using the expression σ E = a E 2 + b2 + c2 rec E rec E2. rec Modeling of the detector resolution gave a = 0.016, b = MeV 1/2, and c = MeV Nonlinearity While uncertainty in absolute calibration had negligible impact on the measurement of θ 13, it influenced the estimate of the neutrino mass-squared difference. This can be seen as a direct consequence of the oscillation phase dependence on the antineutrino energy. The most significant bias arose from the nonlinear response of the detector. Nonlinearity of the reconstructed positron energy relative to the true reconstruction energy originated from two sources: particle-dependent nonlinear light yield of the scintillator, and charge dependent nonlinearity associated with the electronic readout of the PMT charge signal. Positron interactions with the scintillator are predominantly identical to electrons, except for their annihilation. Therefore, the visible energy for a positron was effectively modeled as E vis,e + = E vis,e + 2 E vis,γ (0.511 MeV). The scintillation light output for low-energy electrons was suppressed due to ionization quenching. A semi-empirical analytic approach based on Birks law was used to model this mechanism. The energydependent contribution from Cherenkov light, predicted to be at the level of a few percent relative to scintillation light, induced an additional nonlinearity. The charge nonlinearity induced by the electronics arose from a complex interplay of the time profile of scintillation light and the response of the readout electronics, which could not be easily calibrated at the single PMT channel level. Instead, the resulting nonlinearity was effectively modeled at the detector level. 4

6 Fig. 3. The ratio of the observed energy to the true energy. Comparison of the model with -ray reference data left, estimate of the detector response to positrons right. The complete energy model, relating the reconstructed energy to true particle kinetic energy, contained four free parameters: Birks constant, the contribution from Cherenkov radiation, and the two parameters of the electronics model. The parameter values were obtained from fit to various calibration data, consisting of twelve gamma lines from both deployed and naturally-occurring sources, as well as the continuous electron spectrum from the decays of 12 B produced by cosmogenic muon. Fig. 3 left, compares the resulting energy response model with the calibration data. The resulting estimate of the detector response to positrons is shown in Fig. 3 right. This model of the detector response to positrons was validated using independent calibration reference data. These included the 53 MeV endpoint in the Michel electron spectrum from muon decays, and the continuous β+γ spectra from bismuth and thallium decay. 4. Selection of the antineutrino detection candidates As discussed in Sec. 2, antineutrino inverse beta decay interactions provide a very characteristic pattern of two time-correlated signals of specific energies. Within the Gd-loaded target region, 84% of the captures occurred on either 157 Gd ( 81.5%, 7.95 MeV) or 155 Gd ( 18.5%, 8.54 MeV). The multiple γ- rays emitted upon capture could be distinguished from natural radioactive backgrounds with energies predominately below 5 MeV. The remaining 16% of neutrons captured almost entirely on 1 H, releasing a single 2.2 MeV γ-ray. For capture on 1 H, it was not possible to clearly discriminate whether the ν e interacted within the target region or in the gamma-catcher region. Analysis of the n- 1 H data therefore suffered from larger uncertainties in target volume and detector response, as well as a much more significant background contamination. Despite these obstacles, independent measurements of neutrino oscillation have been obtained using these interactions, with results that were consistent with the analysis of signals identified by n-gd capture, albeit with less precision. Fig. 4 left, shows the distribution of prompt versus delayed energy for signal pairs which satisfied the ν e inverse beta decay selection criteria. Interactions with prompt energy of 0.7 MeV < E p < 12 MeV and delayed energy of 6 MeV < E d < 12 MeV were used for the n-gd neutrino oscillation analysis (red dotted box). A few-percent contamination from accidental backgrounds (symmetric under interchange of prompt and delayed energy) and 9 Li decay and fast neutron backgrounds (high prompt and 8 MeV delayed energy) are visible within the selected region. Inverse beta decay interactions where the neutron was captured on 1 H provided an additional signal region with E d 2.2 MeV, albeit with much higher background. 5

7 Fig. 4. The distribution of prompt versus delayed energy left, and the observed reconstructed positron energy spectra observed in the far experimental hall right. 5. Background The vast majority of triggered signals in each detector were caused by natural radioactive backgrounds, with less than one in 10 5 resulting from reactor ν e interactions. Aside from natural radioactivity, a minor PMT instrumentation-related background was discovered during detector assembly and commissioning. Selecting pairs of prompt plus delayed signals with the proper energies and time separations rejected almost all backgrounds caused by natural radioactivity. Occasionally two such uncorrelated interactions would accidentally satisfy the antineutrino selection criteria. Fig. 4 right, shows reconstructed positron energy spectrum in far detectors as well as contribution from different background origins. The majority of prompt-like uncorrelated signals were from natural radioactivity in the detector components and the surrounding environment, and had E rec < 3 MeV. Delayed-like uncorrelated signals were primarily from two sources. The first were beta decays of unstable isotopes, an unavoidable byproduct of cosmogenic muon interactions with the scintillator. 12 B was by far the most prominent of such isotopes, although others such as 8 Li, 8 B, and 9 C are also produced. The second were high-energy γ-rays produced by the capture of neutrons emitted by the 241 Am- 13 C calibration sources located in the calibration unit on the detector lid. The rate, energy spectrum, and other characteristics of these backgrounds were precisely modeled from studies of individual uncorrelated signals. Each day, only a few of the 10 7 uncorrelated signals were estimated to form a pair which satisfied the antineutrino selection. Detailed studies of all uncorrelated interactions measured this background contamination to be from 1% to 2% depending on detector, with negligible uncertainty. All remaining backgrounds are from physical processes that produce a pair of correlated interactions, which potentially mimic inverse beta decay. The vast majority of such correlated background results from cosmogenic muon interactions with the detector or nearby environment. Vetoing signals that occur soon after muon interactions with the detector or muon systems effectively reduced the contamination caused by these backgrounds to less than 0.5%. Lastly, three or more signals would occasionally occur close in time, giving no chance to recognize which pair was most likely the result of an antineutrino interaction. To avoid this ambiguity, sets of signals with multiplicity >2 were rejected. Six potential sources of correlated backgrounds were identified: Cosmogenic muons: Cosmogenic muons fragmented stable nuclei in or near the detectors, creating free neutrons and unstable nuclei. The signals from the initial muon interaction, subsequent neutron captures, or beta decays of unstable nuclei could potentially form a pair which satisfied the antineutrino selection. Muons were easily identified by their large scintillation light production or by Cherenkov 6

8 light in the water shield. Vetoing concurrent or subsequent signals rendered this background negligible, at the cost of 2% of the far and 14% 18% of the near hall live times. Fast neutrons: Cosmogenic muon interactions in the environment near the detector generated fast neutrons. A nuclear collision of the neutron within the scintillator could mimic a prompt signal, while the subsequent neutron capture was identical to a true IBD delayed signal. β-n decays: Cosmogenic muon interactions occasionally produced the rare unstable isotopes 9 Li and 8 He, which β-decay with a chance of simultaneously emitting a neutron. 241 Am- 13 C neutron sources: During detector operation it was found that neutrons from the 241 Am- 13 C calibration sources within the ACUs occasionally introduced several γ-rays, correlated in time, into the detector. (α,n) interactions: α s emitted by natural radioactivity within the detector could eject neutrons from stable nuclei, with 13 C(α,n) 16 O being the most prevalent interaction. Protons scattered by the neutron or 16 O de-excitation γ-rays could mimic a prompt signal, while the eventual neutron capture provided a delayed signal. High-multiplicity signals: The pile-up of correlated ν e signal pairs with uncorrelated radioactive backgrounds resulted in three or more signals within the correlation time for IBD candidates. Such sets of signals with multiplicity 3 resulted in ambiguity in determination of the actual prompt and delayed ν e signals. Complete rejection of these high-multiplicity combinations resolved this ambiguity, but introduced a 2.5% loss of ν e efficiency. 6. Results The presented results are from analysis of data collected in the Daya Bay experiment with 6 detectors in 217 days (Dec/2011 Jul/2012) and with 8 detectors in 1013 days (Oct/2012 Jul/2015). During 1230 days of operation, the Daya Bay experiment collected more than million inverse beta decays in near halls and more than 150 thousand IBD have been detected in far hall. Daily rate is ~2500 IBD events in near halls and ~300 IBD in far hall Oscillation analysis based on n-gd [21] A much simplified rate only analysis found sin 2 2θ 13 = ± (stat. ) ± (syst. ) The distortion of the energy spectrum at the far hall relative to near halls was consistent with oscillation, and allowed measurement of Δm 2 ee. The parameters of the three-flavor model in best agreement with the observed rate and energy spectra were sin 2 2θ 13 = ± (stat. ) ± (syst. ), Δm 2 ee = [2.50 ± 0.06(stat. ) ± 0.06(syst. )] 10 3 ev 2, 2 (NH) = [2.45 ± 0.06(stat. ) ± 0.06(syst. )] 10 3 ev 2, 2 (IH) = [ 2.55 ± 0.06(stat. ) ± 0.06(syst. )] 10 3 ev 2. Δm 32 Δm 32 2 The Δm 32 values were obtained under the assumptions of normal (NH) and inverted (IH) mass ordering. Fig. 5 - left, shows the observed electron survival probability as a function of effective baseline L eff divided by the average antineutrino energy E ν. Almost one full oscillation disappearance and reappearance cycle was sampled, given the range of L/E ν values which were measured. 2 The confidence intervals for Δm ee versus sin 2 2θ 13 are shown in Fig. 5 right. The 1σ, 2σ, and 3σ 2-D confidence intervals are estimated using χ 2 values of 2.30 (red), 6.18 (green), and (blue) relative to the best fit. The upper panel provides the 1-D χ 2 for sin 2 2θ 13 obtained by profiling Δm 2 ee (blue line), and dash lines mark the corresponding 1σ, 2σ, and 3σ intervals. The right panel is 7

9 Fig. 5. Oscillation survival probability versus antineutrino proper time left. Confidence intervals for sin 2 2θ 13 and Δm 2 ee right. Fig. 6. Comparison of sin 2 2θ 13 provided by RENO [13], D-CHOOZ [14], T2K [15] and MINOS [16], and Δm 2 32 values provided by RENO [13], T2K[15], MINOS [17], NOvA [18], Super-K [19] and IceCube [20]. the same, but for Δm 2 ee, with sin 2 2θ 13 profiled. The point marks the best estimates, and the error bars display their 1-D 1σ confidence intervals. The Fig. 6 compares the values provided by different experiments Oscillation analysis based on n-h [22] Alternative data analysis of data taken in 621 days, based on the events in which the neutron from IBD is captured on hydrogen results with sin 2 2θ 13 = ± Combination of the n-h and n-gd results gives 8% improvement in precision. sin 2 2θ 13 = ± Search for Light Sterile Neutrino The large statistics collected with full configuration of eight detectors in the Daya Bay experiment allowed new precise analysis with aim to search for a light sterile neutrino [23]. A relative comparison of the rate and energy spectrum of reactor antineutrinos in the three experimental halls yields no evidence of sterile neutrino mixing in the < Δm 2 41 < 0.3 ev 2 mass range. The resulting limits on sin 2 2θ 14 shown on the Fig. 7 left, constitute the most stringent constraints to date in the Δm 2 41 < 0.2 ev 2 region. 8

10 Searches for a light sterile neutrino have been independently performed by the MINOS and the Daya Bay experiments using the muon (anti)neutrino and electron antineutrino disappearance channels, respectively. Results from both experiments are combined with those from the Bugey-3 reactor neutrino experiment to constrain oscillations into light sterile neutrinos [24]. The three experiments are sensitive to complementary regions of parameter space, enabling the combined analysis to probe regions allowed by the LSND and MiniBooNE experiments in a minimally extended four-neutrino flavor framework. Stringent limits on sin 2 2θ μe are set over six orders of magnitude in the sterile mass-squared splitting Δm The sterile-neutrino mixing phase space allowed by the LSND and MiniBooNE experiments is excluded for Δm 2 41 < 0.8 ev 2 at 95% CL s, see the Fig. 7 right. Fig. 7. Constraints for sterile light neutrino provided by Daya Bay [23] left, and combined analysis of data from MINOS and Daya Bay/Bugey-3 [24] right Reactor antineutrino flux and spectrum [25] Data collected in 621 days were used to measure the IBD yield in the eight detectors, and the ratio of measured to predicted flux was found to be ± (0.992 ± 0.021) for the Huber+Mueller (ILL+Vogel) model, see Fig. 8 left. A 2.9σ deviation was found in the measured IBD positron energy spectrum compared to the predictions. In particular, an excess of events in the region of 4 6 MeV was found in the measured spectrum, with a local significance of 4.4σ, see Fig. 8 right. Fig. 8. Yield Y for the IBD events in the 6-AD only (top) and 8-AD only (bottom) left. Comparison of predicted and measured prompt energy spectra right. 9

11 References [1] An F P et al. (Daya Bay Collaboration) 2012 Phys. Rev. Lett [2] Cowan C L, Reines F, Harrison F B, Kruse H W, McGuire A D 1956 Science [3] Ahn J K et al. (RENO Collaboration) 2012 Phys. Rev. Lett [4] Abe Y et al. (Double Chooz Collaboration) 2015 JHEP [err: 2015 JHEP ] [5] Abe K et al. (T2K Collaboration) 2014 Phys. Rev. Lett [6] An F P et al. (Daya Bay Collaboration) 2014 Phys. Rev. Lett [7] An F P et al. (Daya Bay Collaboration) 2015 Phys. Rev. Lett [8] An F P et al. (Daya Bay Collaboration) arxiv: [9] Seo S (RENO Collaboration) th Int. Conf. on Neutrino [10] Beringer J et al. (Particle Data Group) 2012 Phys. Rev. D see Sec. 13 [11] Adamson P et al. (MINOS Collaboration) 2014 Phys. Rev. Lett [12] Abe K et al. (T2K Collaboration) 2014 Phys. Rev. Lett [13] Choi J H et al. (RENO Collaboration) 2016 Phys. Rev. Lett. 116, (arxiv: [hepex]) [14] Ishitsuka M (Double Chooz Collaboration) st Rencontres de Moriond, Electroweak Interactions and United Theories [15] Iwamoto K (T2K Collaboration) th Int. Conf. on High Energy Physics [16] Adamson P et al. (MINOS Collaboration) 2013 Phys. Rev. Lett [17] Evans J (MINOS and MINOS+ Collaborations) th Int. Conf. on Neutrino Physics and Astrophysics [18] Vahle P (NOvA Collaboration) th Int. Conf. on Neutrino Physics and Astrophysics [19] Moriyama S (Super-Kamiokande Collaboration) th Int. Conf. on Neutrino Physics and Astrophysics [20] Koskinen J (Ice Cube Collaboration) th Int. Conf. on Neutrino Physics and Astrophysics [21] An F P et al. (Daya Bay Collaboration) 16 Oct 2016 arxiv: v1 [hep-ex] [22] An F P et al. (Daya Bay Collaboration) 2016 Phys. Rev. D [23] An F P et al. (Daya Bay Collaboration) 2016 Phys. Rev. Lett. 117 no [24] An F P et al. (Daya Bay Collaboration, MINOS Collaboration) 2016 Phys. Rev. Lett. 117 no Addendum: 2016 Phys. Rev. Lett. 117 no [25] An F P et al. (Daya Bay Collaboration) 2016 Phys. Rev. Lett. 116 no