Dating young basalt eruptions by (U-Th)/He on xenolithic zircons Madalyn S. Blondes* Department of Geology and Geophysics, Yale University, 210 Whitney Avenue, Peter W. Reiners† New Haven, Connecticut 06511, USA Benjamin R. Edwards* Department of Geology, Dickinson College, James Center, Carlisle, Pennsylvania 17013, USA Adrian Biscontini* ABSTRACT young and/or low-K samples often have large Accurate ages for young (e.g., Pleistocene) volcanic eruptions are important for geo morphic, errors due to the small proportion of radiogenic tectonic, climatic, and hazard studies. Existing techniques can be time-consuming and expen- Ar to total Ar (McDougall and Harrison, 1999). sive when many ages are needed, and in the case of K/Ar and 40Ar/ 39Ar dating, extraneous Ar In contrast, ZHe can be performed relatively often can limit precision, especially for continental basalts erupted through old lithosphere. rapidly on replicate single grains, requires no We present a new technique for dating young basaltic eruptions by (U-Th)/He dating of zir- irradiation, and the greater diffusivity and lower cons (ZHe) from crustal xenoliths. Single-crystal ZHe dates generally have lower precision atmospheric abundance of He may lead to a than typical 40Ar/ 39Ar dates, but can be determined relatively easily on multiple replicate grain lower inherited daughter abundance. Because aliquots. We dated zircons from xenoliths from four volcanic centers in western North Amer- inherited He is likely to be less of a problem ica: Little Bear Mountain, British Columbia (157 ± 3.5 [2.2%] ka weighted 95% confi dence than inherited Ar, ZHe ages should be younger interval [CI], mean square of weighted deviates [MSWD] = 1.7) and Prindle Volcano, Alaska than Ar ages in these samples. (176 ± 16 [8.9%] ka, MSWD = 13), in the northern Cordilleran volcanic province, and Fish ZHe dating is typically applied to thermo- Springs (273 ± 23 [8.6%], MSWD = 43) and Oak Creek (179 ± 8.1 [4.5%] ka, MSWD = 12), in chronologic studies of tectonic and erosional the Big Pine Volcanic Field, California. All ZHe ages are either equivalent to or younger than exhumation, where it has a closure temperature previously determined K/Ar or 40Ar/ 39Ar ages, indicating the possibility of inherited 40Ar in (for typical crystal sizes and assuming a cooling some of the previous measurements. Zircons from upper crustal xenoliths in the Oak Creek rate of ~10 °C/m.y.) of ~180 °C (Reiners et al., and Fish Springs vents show poorer reproducibility and multiple apparent age distribution 2004; Reiners, 2005; Stockli, 2005). It has also peaks, consistent with either intracrystalline U-Th zonation or <99.99% He degassing (assum- been used for geochronology, to date volcanic ing ca. 100 Ma pre-entrainment ZHe ages) of some zircons during magmatic entrainment. rocks and resetting from intrusions and sub surface Removal of clear outliers in the older age-distribution peaks of the upper crustal xenoliths, coal fi res (Farley et al., 2002; Tagami et al., 2003; most of which have extremely high U compared to other zircons of the same xenolith, improve Heffern et al., 2007). In rare cases, reproducibil- the reproducibilities of Fish Springs to 4.7% (95% CI, MSWD = 4.8) and Oak Creek to 3.4% ity of ZHe ages on replicate single-grain analyses (95% CI, MSWD = 6.2). Coupled thermal and He diffusion modeling using appropriate xeno- approaches that of analytical precision (<4%, 2σ) lith sizes and magma temperatures and assuming published diffusion kinetics for zircon indi- (e.g., Nasdala et al., 2004; one example in this cate that incomplete He degassing would require entrainment times <1 h. However, the obser- study), but in most cases, reproducibility of typi- vation of extremely high U in most zircons with older ages raises the possibility that zircons cal zircons requiring alpha-ejection corrections with high radiation dosages may have more retentive He diffusion characteristics. such as those of the Fish Canyon Tuff (FCT) is ~8%–9% (2σ) (Reiners, 2005). In the case of Keywords: (U-Th)/He, geochronology, thermochronology, Big Pine Volcanic Field, northern Cor- zircons reset by short-duration reheating, such dilleran volcanic province, xenoliths. as in this study, incomplete resetting is another potential source of scatter. Zircons retaining some INTRODUCTION and Cornette, 1986; Duncan and Hogan, 1994; fraction of their pre-xenolithic He after magmatic Dating young volcanic rocks is important for Guillou et al., 1997; Heizler et al., 1999; Singer entrainment and eruption may yield signifi cantly a variety of purposes, including understanding et al., 2000, 2004). In a different approach, Gil- older ages (though this would still provide a geomorphic, tectonic, and climatic chronolo- lespie et al. (1982, 1983, 1984) circumvented maximum eruption age estimate). U and Th gies as well as hazard assessments. Pleistocene the lack of U-, Th-, or K-bearing phenocrysts zonation may also affect ZHe ages by as much as eruptions have typically been dated using K/Ar, in Pleistocene basalts in the Big Pine Volcanic 35% in extreme cases, though age scatter arising 40Ar/ 39Ar, U-series disequilibria, 14C, fi ssion Field by using the 40Ar/ 39Ar step-heating method from zonation is typically smaller than this (e.g., track of volcanic glasses (Renne, 2000, and ref- of K-feldspars extracted from granitic xenoliths. ~± 5% for FCT; Hourigan et al., 2005). erences therein), and (U-Th)/He (Farley et al., Early-released gas yielded precise apparent 2002). Young basaltic eruptions can be particu- 40Ar/ 39Ar ages interpreted as eruption age, and SAMPLE LOCATIONS larly problematic due to the rarity of U-, Th-, or later steps revealed inherited Ar in more reten- AND PETROLOGY K-bearing phenocrysts such as zircon, apatite, tive domains that was not completely outgassed We measured ZHe ages from felsic xenoliths or K-feldspar within them, though precise tech- during heating. This approach was an important from four basaltic centers: two in the Big Pine niques have been developed for whole-rock K/Ar step in the development of multidomain dif- Volcanic Field (BPVF) of eastern California, and 40Ar/ 39Ar analyses of young basalts (Gillot fusion theory in 40Ar/ 39Ar thermochronology and two from the northern Cordilleran volcanic (McDougall and Harrison, 1999). province (NCVP) in British Columbia, Canada, Although in certain circumstances modern and eastern Alaska. Units from each locality *E-mails: [email protected]; edwardsb@ 40Ar/ 39Ar approaches can yield precise ages have been dated previously, though in some dickinson.edu; [email protected]. †Current address: Department of Geosciences, of young mafi c eruptions, in many cases zir- cases data are only reported as average ages in University of Arizona, Tucson, Arizona 85721, USA; con (U-Th)/He (ZHe) may be more expedient. abstracts, and in at least one case multiple erup- [email protected]. The 40Ar/ 39Ar method requires irradiation, and tions of different age may have come from a ©2007 Geological Society of America. For permission to copy, contact Copyright Permissions, GSA, or [email protected]. GEOLOGY,Geology, January January 2007; 2007 v. 35; no. 1; p. 17–20; doi: 10.1130/G22956A.1; 2 fi gures; 1 table; Data Repository item 2007013. 17 single vent complex (Table 1). (See Data Repos- but are signifi cantly younger than the Gillespie 16 itory Table DR11 for detailed sample locations et al. (1983) 40Ar/ 39Ar ages (Table 1). Because Oak Creek 14 12 and descriptions.) Oak Creek consists of many separate outcrops 176 ± 6.1 [3.4%] ka of basalt fl ows that are not clearly related, this MSWD = 6.2 10 METHODS suggests that fl ows in the Oak Creek area repre- 8 Single zircon crystals with c-axis-perpendicu- sent several distinct eruptions with widely vary- 6 μ lar dimensions ranging from 61 to 210 m were ing ages. The fl ow remnants dated in this study 4 179 ± 8.1 [4.5%] ka photographed, measured, and wrapped in Nb and those dated by K/Ar all may be the same 2 foil using methods outlined in Reiners (2005). fl ow, but the 40Ar/ 39Ar ages from the fl ow farther MSWD = 12 0 Alpha-ejection corrections were made using the west that was studied by Gillespie et al. (1983) 14 approach of Farley (2002) and Hourigan et al. indicate at least one additional earlier eruption. Fish Springs (2005). Analytical uncertainties (1σ) propagated Recent whole-rock 40Ar/ 39Ar ages of the same 12 243 ± 11 [4.7%] ka 4 from uncertainties on U, Th, and He measure- three Gillespie et al. (1983) Oak Creek outcrops MSWD = 4.8 10 ments (including original manometric calibra- confi rm the presence of two distinct fl ows with 8 tions) are typically <2.5%, but a few crystals, ages of ca. 1.2 and 0.20 Ma (Blondes et al., 6 4 σ 273 ± 23 [8.6%] ka most with He contents <3 fmol, have 1 uncer- unpublished data). MSWD = 43 4 tainties as high as 4.5%. We analyzed 15 zircons from 1 xenolith from 2 Using the statistics defi ned in Isoplot (Lud- the Fish Springs cone. The ZHe ages range wig, 2003), we calculated weighted mean ages, from 233 to 381 ka and the entire population 0 95% confi dence intervals, and mean square of has a weighted mean age of 273 ± 23 ka (8.6%, 8 weighted deviates (MSWD) values using 1σ weighted 95% CI, MSWD = 43) (Fig. 1, Table 1; Little Bear 7 157 ± 3.5 [2.2%] ka 6 analytical error as weights. These weights rep- Table DR1 [see footnote 1]). The weighted mean probability Relative MSWD = 1.7 resent only formal analytical uncertainties and age with the eight oldest measurements of the 5 do not take into account uncertainties on alpha- two distinct older age peaks removed is 243 ± 4 ejection corrections and the effects of U-Th 11 ka (4.7%, weighted 95% CI, MSWD = 4.8).