Intensity of the lunar dynamo field at 3 . 0

Introduction: Recent paleomagnetic studies of Apollo samples have indicated that the Moon generated a core dynamo between at least 4.25 and 2.5 Ga and that a high field period with surface field intensities ≈40-110 μT occurred between 3.85 and 3.56 Ga [1]. The high field period was followed by an apparent decline in surface field intensity to below ≈4 μT at or after 3.19 Ga [2], with evidence for a ≈5 ± 2 μT surface field at ≈1-2.5 Ga [3], after which the field further declined to modern intensities < 0.1 μT [4]. The particular mechanisms of lunar field generation have not yet been conclusively identified. By improving our understanding of the timing of the initial period of field decline, we can better assess possible dynamo sources. The poor magnetic recording properties of most lunar igneous rocks have hindered prior attempts to retrieve paleomagnetic records from samples < 3.56 Ga. Reliable paleointensities generally may not be retrieved from mare basalts if they were initially magnetized in paleofields < 10-50 μT, with few exceptions [2, 3, 5]. Because vitrophyric basalts initially cooled below their magnetic blocking temperatures at relatively fast rates, they may contain smaller, higher-fidelity ferromagnetic grains. In previous work, we showed that the rock magnetic properties of two Apollo 12 vitrophyres constrain lunar dynamo fields to < ≈4-7 μT at the time of their formation [6]. We expand upon that work by analyzing the paleomagnetic behavior of a third Apollo 12 vitrophyre and acquiring new radiometric ages for all three samples. This permits improved statistical assessment of field decline after 3.56 Ga. We further investigate the connection between mineralogy and paleomagnetic fidelity through magnetic and petrographic characterization. Samples: Fine-grained olivine vitrophyre mare basalts 12008, 12009, and 12015 were collected in the southeastern region of Oceanus Procellarum during the Apollo 12 mission. We prepared six mutually oriented subsamples from each parent chip studied (12008,72, 12009,156, and 12015,40) using a wire saw. Five subsamples were used for magnetic analyses, while the sixth was reserved for compositional analyses and dating. Initial petrographic analyses were conducted on thin sections 12008,41, 12009,11, and 12015,43, provided by JSC. Characterization and geochemistry: One slice of each parent chip was subdivided and mounted as a thick section. Both thick and thin sections were subject to electron microprobe analyses, which were conducted using the JEOL JXA-8200 Superprobe at Rutgers University (RU). Quantitative point analyses and qualitative (EDS) X-ray maps were performed on both mesostasis glass and larger metallic grains. All three samples contain large (up to 5 mm) olivine, (ortho)pyroxene, and chromite phenocrysts in a glassy matrix, as well as small (≤ 20 μm) Fe-Ni alloy grains with a range of compositions (4.5 < wt.% Ni < 20.5) indicative of kamacite and martensite. Magnetic hysteresis loops and backfield curves were collected using the Princeton MicroMag 2900 AGM at RU and corrected using the IRMDB software provided by the Institute for Rock Magnetism at the University of Minnesota. Radiometric dating. Ages for samples 12008 and 12009 were previously reported as 3.10 and 3.19 Ga respectively, although radiometric dating was complicated by anomalous age spectra [7, 8], while 12015 has never before been radiometrically dated. K-enriched mesostasis glass fragments were micro-milled from each thick section for Ar/Ar dating. Samples were irradiated at the USGS TRIGA reactor in Denver, CO. Argon isotopes were then analyzed using a Mass Analyzer Products (MAP) 215-50 noble gas mass spectrometer at the Rutgers Noble Gas Lab. We report new ages of 3.023 ± 0.066 Ga for 12008, 2.909 ± 0.071 Ga for 12009, and 3.025 ± 0.024 Ga for 12015.