Partial Radiogenic Heat Model for Earth Revealed by Geoneutrino Measurements
Total Page:16
File Type:pdf, Size:1020Kb
ARTICLES PUBLISHED ONLINE: 17 JULY 2011 | DOI: 10.1038/NGEO1205 Partial radiogenic heat model for Earth revealed by geoneutrino measurements The KamLAND Collaboration* The Earth has cooled since its formation, yet the decay of radiogenic isotopes, and in particular uranium, thorium and potassium, in the planet’s interior provides a continuing heat source. The current total heat flux from the Earth to space is 44:2 ± 1:0 TW, but the relative contributions from residual primordial heat and radiogenic decay remain uncertain. However, radiogenic decay can be estimated from the flux of geoneutrinos, electrically neutral particles that are emitted during radioactive decay and can pass through the Earth virtually unaffected. Here we combine precise measurements of the geoneutrino flux from the Kamioka Liquid-Scintillator Antineutrino Detector, Japan, with existing measurements from the C8:8 Borexino detector, Italy. We find that decay of uranium-238 and thorium-232 together contribute 20:0−8:6 TW to Earth’s heat flux. The neutrinos emitted from the decay of potassium-40 are below the limits of detection in our experiments, but are known to contribute 4 TW. Taken together, our observations indicate that heat from radioactive decay contributes about half of Earth’s total heat flux. We therefore conclude that Earth’s primordial heat supply has not yet been exhausted. he Kamioka Liquid-Scintillator Antineutrino Detector mantle, outer core and inner core. Although the mechanical (KamLAND) Collaboration reported the results of the first properties and bulk composition of the shells are well established, Tstudy of electron antineutrinos (νe s) produced within the their detailed composition, including the abundances of radiogenic Earth in 2005 (ref. 1). The KamLAND data indicated an excess species, remains uncertain. of the νe flux at energies in the range consistent with the decay Since the model assumptions are not well grounded, the chains of 238U and 232Th. The rate was consistent with geophysical predictions have large uncertainties. The bulk silicate Earth (BSE) expectations for heavy-element concentrations throughout the model of ref. 5, for example, assumes that the primordial Earth Earth's interior. The statistical significance improved as more data was formed from accretion of matter from the same nebula were obtained2. Recently, the Borexino Collaboration at Gran Sasso that formed the sun and the rest of the planets, and that the reported an excess of events they attribute to geoneutrinos3. While BSE abundances of refractory lithophile elements such as U and Borexino convincingly confirms the KamLAND excess, neither Th can be determined by a combination of measured elemental result is precise enough to provide much guidance for geophysical abundances of chondritic meteorites and mantle peridotites. The models. In this paper we present more recent KamLAND results composition deduced from this model results in a radiogenic heat that begin to constrain the models. production of 8 TW from the 238U decay chain, 8 TW from the The present analysis is possible because of the recently improved 232Th decay chain and 4 TW from 40K (ref. 6). In this model, sensitivity to geoneutrinos in the KamLAND experiment. The the radiogenic heat contribution is nearly half of the Earth's KamLAND geoneutrino background is dominated by νe s from total heat flow (44:2 ± 1:0 TW, ref. 7; see also Supplementary commercial nuclear reactors and the 13C(α;n)16O reaction initiated Note S1). Clearly, quantitative information about the radiogenic, by the decay of radioactive contaminants in the detector. In recent heat-producing elements is essential for establishing the energy years the reactor νe flux, which is outside the control of the budget, which in turn is key to understanding the Earth's experiment, was significantly reduced2 because of an extended formation and evolution. shutdown of the Kashiwazaki–Kariwa nuclear power station following an earthquake in July 2007. In the meantime, at great Geoneutrino flux effort, the purity of the KamLAND scintillator was improved, The Earth is nearly transparent to neutrinos because they interact eliminating most of the 210Pb that feeds the decay chain responsible only through the weak force. Geoneutrinos are a unique, direct for the production of α-particles from 210Po decay. The background window into the interior of the planet. Accurately mapping from the 13C(α;n)16O reaction went down by a factor of ∼20. neutrino sources inside the Earth by measuring geoneutrino fluxes The measurement presented here is based on data collected at the surface is an inverse problem requiring multiple detection between 9 March 2002 and 4 November 2009, and includes the sites. The prospects for multisite measurements have been discussed 1,2 data used in our previous publications . The total exposure to νe in refs 4,8, and an analysis of the likely sensitivity of such is (3:49±0:07)×1032 target proton years, a fivefold increase from measurements was discussed in ref. 9. For the present work, which the first KamLAND report1. The expected signal in one geophysical has limited statistics, a simple BSE-inspired reference model4 for the model4 increases from 19 events to 106 events. radiogenic material distribution is employed. The model assumes that U or Th is present in the crust and mantle but not in the Earth composition model core. The effects of radiogenic material in the local geology of Analyses of seismic waves indicate a shell structure for the Earth's Kamioka are carefully assessed but account for less than 10% of interior, conventionally denoted as crust, upper mantle, lower the total expected flux. *A full list of authors and their affiliations appears at the end of the paper. NATURE GEOSCIENCE j VOL 4 j SEPTEMBER 2011 j www.nature.com/naturegeoscience 647 © 2011 Macmillan Publishers Limited. All rights reserved. ARTICLES NATURE GEOSCIENCE DOI: 10.1038/NGEO1205 a The differential geoneutrino flux at a position r is determined 160 0 0 Best–fit reactor ν from the isotopic abundances ai(r ) at the location of the sources, r , e 140 KamLAND data Accidental isotopes Z 0 0 0 120 d8(Eν ;r) X dni(Eν ) 0 ai(r )ρ(r )P(Eν ;jr−r j) D A d3r (1) 100 i j − 0j2 dEν i dEν ⊕ 4π r r 80 60 13C(α, n)16O where the integration extends over the Earth's volume, Ai is the Best–fit geo ν Events/0.2 MeV e 40 ν decay rate per unit mass, dni(Eν )=dEν is the νe energy spectrum for Best–fit reactor e+ background ν 0 20 + best–fit geo e each mode of decay, ai(r ) is in units of isotope mass per unit rock 0 0 0 mass, ρ(r ) is the rock density and P(Eν ;jr−r j) is the νe `survival' 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 probability due to the phenomenon of oscillation after travelling a E (MeV) 0 p distance jr−r j. For the present purpose, the νe survival probability b Data–background–best–fit reactor ν is well approximated by the two-flavour oscillation formula, 40 e ν Reference geo e 2 T 2U T U 20 2 2 1:271m21 eV L m P(Eν ;L) ' 1−sin 2θ12 sin (2) 0 Eν TMeVU Events/0.2 MeV 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 where L D jr − r0j. `Matter effects' on neutrino oscillations10 E (MeV) are expected to change equation (2) by about 1%, which is p negligible compared with the statistical uncertainty. The oscillation c 100 2 2 80 parameters 1m21 and sin 2θ12 are determined with substantial Selection efficiency ν accuracy by a combined statistical analysis with KamLAND's 60 for geo e measurement of ν s produced at nuclear reactors and data from e Efficiency (%) 40 solar-neutrino experiments (assuming charge–parity–time (CPT) 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 10 symmetry ), and are given in the next section. Given the size of the Ep (MeV) Earth and the values of the neutrino oscillation parameters, for the energy range of detectable geoneutrinos the second sine function Figure 1 j Prompt energy spectrum and event selection efficiency. in equation (2) is well averaged over the volume of the Earth, giving a, Prompt energy spectrum of low-energy νe s in KamLAND. The 2 P(Eν ;L)'1−0:5sin 2θ12 to an excellent approximation. histograms indicate the backgrounds, whereas the best fit (including geoneutrinos) is shown in blue. b, Background-subtracted energy spectrum. Geoneutrino detection The blue shaded spectrum is the expectation from the reference model, KamLAND is located under Mount Ikenoyama (36:42◦ N, consisting of contributions from U (dashed curve) and Th (dotted curve). c, 137:31◦ E), near the town of Kamioka, Japan. The underground Energy dependence of the geoneutrino event selection efficiency averaged site provides an effective overburden of 2,700 m water equivalent, over the data-taking period. Statistical uncertainties are shown for the data reducing the cosmic-ray-induced atmospheric muon flux to in a, and uncertainties on the background estimation are added in b. −2 −1 5:37 ± 0:41 m h (ref. 11). The νe s are detected in 1 kt of 2 D liquid scintillator (LS) through the inverse β-decay reaction, Taking the neutrino oscillation parameter values 1m21 C ! C C C0:19 −5 2 2 νe p e n, with a 1.8 MeV neutrino energy threshold. This 7:50−0:20 × 10 eV and sin 2θ12 D 0:84 ± 0:03 from the fit to the 238 threshold cuts off much of the geoneutrino signal from the U data discussed below, the expected number of reactor νe events 232 40 and Th decay chains and renders the detector insensitive to K in the geoneutrino energy region (defined as 0:9 MeV < Ep < (other unobserved isotopes such as 235U contribute negligibly to 2:6 MeV) is 484:7 ± 26:5, including a small contribution from the the heating).