GEOPHYSICAL RESEARCH LETTERS, VOL. 12, NO. 1, PAGES 17-20, JANUARY1985 ERUPTION OF BASALT AND ANDESITE LAVA DEGASSES 222Rn and210po J. Gill, R. Williams, and K. Bruland Earth Sciences Board, University of California at Santa Cruz Abstract. Activities of 222Rn and 210po . were provide opportunities to study these phenomena measured in a September, 1983, basic andesite lava For those nuclides, use of parentheses denotes from Arenal and a November, 1983, basalt from activities (i.e. concentration times decay rate). Kilauea, starting 3 and 1 days after eruption, re- The nuclides include 222Rn, a rare gas with a 3.8d spectively. In both cases, in-growthpatterns half-life, and 210po, a Te homologue which vola- show that all Rn volatalized during eruption. Po degassing also was complete at Kilauea but only 84*_10% at Arenal. Introduct ion As the result of vapor saturation in magma, both near-surface vesiculation and volcanic erup- tions themselves result in magmatic differentia- tion and the transfer of volatile elements from magma to the atmosphere. However, because ele- ments partition between silicate liquids and the exsolved gas phase to varying extents, degassing efficiency is variable and difficult to measure. We will define a degassing efficiency factor (DEF) to be (C i - Cf)/Ci, where C i is the initial concentration of an element (e.g., at the time of phenocryst precipitation), and Cf is its concen- tration after eruption. Presently, subaerial eruptions of basalt are thought to have DEFs of ~1.0 for S and CO2, 0.7 or higher for H20 , and talizes at low temperatures and has a 138d half- life. For example from the in-growth pattern of its (214Bi)/(208Tli ratio,,dacitic pumice erupted in 1976 from Sakurajima volcano, Japan, was inter- preted to have lost 40 to 100% of its initial Rn, but 1977 rhyolitic pumice from Usu volcano to have lost none (Sato and Sato, 1977; Sato et al., 1979). Experimental melting of basalt at 1300øC in atmosphere for an hour by Sato et al. (1980) released only about 60% of its radon although all of the lighter rare gases were lost. In addition, Po has been found to be greatly enriched in some volcanic plumes such that 50 to 75% of atmospheric 210po probably hasa volcanic origin (Lambert et al., 1979; Lambert, 1983). Here, we present data for (222Rn) and (210po) in 1983 lavas from Arenal and Kilauea volcanoes which had the same pre-eruption Rn and Po concen- trations. The results show that, despite •ignifi- cant differences in eruption style, both nuclides were completely degassed (DEF = 1.0) or almost much lower values for C1 and F, with the DEF being completely degassed (DEFpo = 0.84 at Arenal) during a function of eruption style and duration (Harris eruption. and Anderson, 1983; Roseet al., 1983; Swanson and Analytical Methods Fabbi, 1973;Greenland et al., 1984). A DEF>0.9 for H20 has been estimated for rhyolite lava For the radon measurements, lavas were wet flows, decreasing to 0.2 in airfall tephra (Taylor sawed to make a flat surface for efficient contact et al., 1983), and DEFs for K and Rbmay be >0.3 with the gamma-detector and dried in a 40øC oven at Kilauea (Zoller et al., 1983). However, quan- for 30 minutes. About 200 grams werewrapped in tification of DEFs is difficult due to inherent parafilm and sealed inside two polyethylene bags. uncertainties in C i. Counting began immediately after sealing. Follow- If the DEF of gas transfer is known, the volume ing completeingrowth of Rn, the sample from Are- of vesiculated magma at depth can be computed from nal was unwrapped, washed,dried, resealed and elementfluxes observed in plumes aboveeruption placed on the detector as before. The observed volcanoes. Such estimates using sulfur and radon fluxes typically exceed the erupted volume of magma, leading to the important realization that several times more vesiculated magma resides at depth than erupts (e.g. Rose et al., 1983; Lam- bert, 1983). However, how much more depends on (equilibrium) activities did not change, indicat- ing that insignificant Rn was lost during sample preparat ion. The samples were then counted with Ge(Li) gamma-detector shielded with 4" of Pb. Samples were counted continuously and each data point the poorly-known DEF. represents approximately two days of counting. We Highly efficient loss of Ar gas also is impor- monitored the 609 Kev214Bi gamma-ray and the two tant to K-Ar geochronology. Discovery of "extran- 214pb gamma-rays at 295and 352Kev; both are eous Ar" in subaerial lavas (Dalrymple, 1969; short-lived daughters of 222Rn in the 238U decay McDougall et al., 1969) indicates that there can series. Counts in eachpeak werecorrected for be incomplete loss evenof rare gas under some both Compton and detector background. Both circumstances. These circumstances are important (214Bi) and (214pb) were observed to grow into to predict now that K-Ar methods are being used to plateau values from initial activities only date very young rocks. Thus, degassing efficiencies are important to know and of variable magnitude, but difficult to determine. Short-lived, volatile radionuclides Copyright 1985 by the American Geophysical Union. Paper number 4L6355. 009 4-82 76/8 5/004L-63555 03.00 slightly above background. As an indicator of system stability, both the 583 Kev 208T1 and the 212 s 239 Kev Pb gamma-ray also were monitored. These nuclides are daughters of very short-lived 220Rn of the Th decay series and should be in .... •• ...... • 224Ra No significant variation was observed, which means that 224Ra (tl/2 -- 3.64d) was neither enriched nor depleted with respect to its parent 228Th,and that the detector efficiency remained constant. 17