Miocene age of the Huanan basalt lava flow (NE China) inferred by reset of zircon (U–Th)/He thermochronometer in the underlying sand

Mafic lavas of Cenozoic age are widely distributed in northeast China and received much attention as an important part of the Circum‐Pacific volcanic belt. The age constraints for the volcanic activity were determined mostly by K/Ar and 40Ar/39Ar methods. We present zircon (U–Th)/He ages obtained on the thermally overprinted sands directly underlying a basaltic lava. This thermochronometer is insensitive to weathering and not biased by excess argon, thus it can express accurately the age of thermal effect of the lava flow. As a regional cooling age reference, three granite samples were dated from basement units that have not been thermally influenced by the basalt eruptions. The reference granite samples revealed well‐defined Cretaceous (U–Th)/He‐ages, while 20 zircon crystals from the sand below the basalt lava revealed a prominent Miocene (U–Th)/He age component of 9.33 ± 0.24 Ma. Raman spectroscopy of these zircon crystals supports their thermally overprinted character. We infer that the sand sample has experienced significant thermal overprint by the overlying basalt lava leading to thermal reset of the majority of the detrital zircon crystals. The obtained age is thus interpreted as the eruption age of the basalt lava. The Huanan basalt flow thus belongs to volcanics of the Laoyeling episode in NE China.


| INTRODUCTION
Mafic lavas of Cenozoic age are widely distributed in northeast China.
In the last two decades, new geochronological techniques were introduced for dating young mafic eruptions such as the U-Th disequilibrium method (Zou, Zindler, Xu, & Qi, 2000), indirect dating of volcanics from the surrounding fallout organic material deposits by the 14 C method (Xu, Zhang, Qiu, Ge, & Wu, 2012;Yin et al., 2012), fission track dating of volcanic glasses (Renne, 2000), and magnetite or zircon (U-Th)/He (ZHe) geo-thermochronology (e.g., Blackburn, Stockli, & Walker, 2007;Blondes, Reiners, Edwards, & Biscontini, 2007;Cooper, van Soest, & Hodges, 2011;Farley, 2002). The modern 40 Ar/ 39 Ar approaches may yield precise ages of young volcanic rocks, but typically the age of young and/or low-K lava samples have high errors due to minor proportions of radiogenic Ar (Blondes et al., 2007;McDougall & Harrison, 1999). The magnetite (U-Th)/He method is F I G U R E 1 (a) Simplified geological map of NE China, modified after Ren et al. (2013) and HBGMR (1993). The occurrences of basalt volcanoes and their age (in Ma) are taken from Fan, Sun, Li, and Wang (2006), Fan et al. (2011), Fan, Zhao, Sui, Li, and, Liu (1987), Liu, Chen, Zhong, Lin, and Wang (2017), , Qiu, Liao, and Liu (1991), and Zhang, Xu, Ge, and Ma (2006). The digital elevation model is from the U.S. Geological Survey, 2017. The two faults marked with (1) and (2) Liu, Chen, et al., 2017;Liu, Li, et al., 2017) [Colour figure can be viewed at wileyonlinelibrary.com] also introduced to date mafic volcanic rocks (Blackburn et al., 2007;Fanale & Kulp, 1962). However, this mineral is not suitable for a wide range of applications due to its disadvantages. For example, (a) the Fe-oxide minerals in mafic volcanic formations have frequently irregular external morphology, thus the ejection (FT) correction is hardly feasible and it would generate significant bias (Hernandez Goldstein, Stockli, Ketcham, & Seman, 2014). (b) The interior of the magnetite grains in lavas are highly heterogeneous, often penetrated by ilmenite and haematite lamellae and they contain apatite inclusions. (c) The U content is usually very low. The studies for example, Fanale and Kulp (1962) and Blackburn et al. (2007) were dealing with pre-Cenozoic ages, with a few ppm or even sub-ppm uranium content. In the case of Miocene-Pliocene lavas the uncertainties would be much over the expectations for stratigraphical purposes. Furthermore, the Blackburn et al. (2007) study was made on kimberlites, which has atypical actinide contents and distributions. Zircon analysis has been proven a versatile tool for examining a wide range of geological processes because zircon crystals have a lot of important features for geochronology and thermochronology including high actinide concentrations, occurrence in variable lithologies and resistance to physical and chemical weathering (Reiners, 2005). Like many other minerals, zircon can also be dated by the (U-Th)/He method to reveal the low temperature (180-130 C) thermal history (e.g., Farley, 2002;Reiners, Spell, Nicolescu, & Zanetti, 2004). Comparing to the K/Ar and 40 Ar/ 39 Ar methods, zircon (U-Th)/ He method has the advantage of performing relatively rapidly on selected zircon crystals without neutron irradiation, and high accuracy on young volcanic rocks (Blondes et al., 2007). Even though mafic to intermediate volcanic rocks rarely contain zircon crystals, the strata below lava flows or the host rocks in contact with basaltic dykes, sills, or necks are often rich in zircon crystals. These zircons may become thermally reset upon significant heating (temperature and time) and the (U-Th)/He age obtained on these crystals then indicates cooling after the heating event. Assuming usual fast cooling of lava flows, this age should reflect the eruption age (Blondes et al., 2007;Cooper et al., 2011).
In this study, we report for first time zircon (U-Th)/He ages from a thermally overprinted basal layer of a lava flow from the Huanan region in NE China. Additionally, Raman spectroscopy was used to describe the crystalline state and confirm the thermal reset of the dated zircon crystals.  Jahn et al., 2000;Sengör et al., 1993;Windley et al., 2007; Figure 1). This area was mainly dominated by the Palaeo-Asian Ocean tectonic domain in the Pre-Mesozoic period, and strongly transformed by the circum-Pacific tectonic domain since the Mesozoic (Liu et al., 2010. Since the Late Mesozoic a large continental rift system developed in NE China, related to the subduction of the Pacific Plate and back-arc extension of the Japan Sea (Liu, 1988;Xu & Fan, 2015). This rift system includes the Songliao Basin, Jiamusi-Yitong Fault Zone, Dunhua-Mishan Fault Zone, and other adjacent basins ( Figure 1). Contemporaneously, about 690 volcanic cones and craters and 50,000 km 2 of basaltic lavas with small amounts of alkali trachyte were formed in this area. The Cenozoic volcanism is mainly distributed alongside a series of NE to NNEoriented rift basins and adjacent mountain ranges and on both sides of the Songliao Basin, but major volcanic activity occurred to the east (Liu, 1988; Figure 1). From west to east, the distribution of the volcanic rocks can be divided into several zones, these are the Great  (Bai, Tian, Wu, Xu, & Li, 2005;Bai, Wang, Xu, Liu, & Xu, 2008;Fan & Hooper, 1991;Fan, Liu, Zhang, & Sui, 1998;Fan et al., 1999Fan et al., , 2006Fan et al., , 2007Fan et al., , 2011Fan et al., , 2012Liu, 1987;Liu et al., 1998;Qiu et al., 1991;Zhang et al., 2000;Zhao et al., 2008; Figure 1). The Cenozoic basalts in NE China are considered products of partial melting of the upper mantle, and mixing of depleted mantle and enriched mantle Type I components (Xu et al., 2015;Zou et al., 2000, Zhou, 2006.

| GEOLOGICAL SETTING
Even though numerous geochronological studies have been published from many occurrences of mafic volcanic formations in NE China, high-precision and weathering-insensitive geochronology such as zircon U-Pb or (U-Th)/He dating has not yet been performed on the young volcanic formations of the Huanan area. Previous studies in this area mainly rely on constraints from lithostratigraphic and paleontological evidences (HBGMR, 1993).

| SAMPLE AND ANALYTICAL METHODS
A sand sample (JB40) was collected in an active basalt quarry close to Qunli village (Figure 2;46.2983 N, 130.7182 E). A 2-3 m thick horizontal lava flow is exposed along the excavation walls and the contact to the underlying sand is well preserved and accessible. In the surroundings of the quarry, the sand forms only a few metres thick layer; this young, alluvial sediment covers the granitoid basement. The basal layer of the lava is amygdaloid, but the lava shows a low degree of alteration. We collected a loose sand sample from the topmost 3-5 cm, immediately below the base of the basalt lava  Figure 4). The grains were wrapped in F I G U R E 2 Simplified geological map of the study area, modified after HBGMR (1993). The digital elevation model is taken from the U.S. Geological Survey, 2017. Pt, Palaeoproterozoic strata; J, Jurassic strata; K1, lower Cretaceous strata; N1, Miocene strata; Q2-3, Middle to Upper Quaternary strata; Q4, Holocene strata; γ: Permian granite; β: Cenozoic basalt; yellow star: sample locations in this article and the measured (U-Th)/He age; black star: zircon U-Pb age of granitoids (Dong et al., 2017) The remaining inert gas was measured by a Hidden triple-filter quadrupole mass spectrometer equipped with a positive ion-counting detector.
Following degassing, the capsules were retrieved from the gas extraction line the zircon crystals were extracted from the capsules and spiked with calibrated 230 Th and 233 U solutions in 0.4 ml teflon vials. The crystals were dissolved for 5 days at 220 C in pressurized bombs using a mixture of double distilled 48% HF and 65% HNO 3 .
Each sample batch was prepared with a series of procedural blanks and spiked normals to check the purity and calibration of the reagents and spikes. Spiked solutions were analysed by a Thermo iCAP Q ICP-MS. Procedural U and Th blanks by this method are usually very stable in a measurement session and below 1.5 pg. The ejection correction factors (Ft) were determined for the single crystals by a modified algorithm of Farley et al. (1996) using an in-house spread sheet.
Raman spectroscopy was applied to all zircon samples to identify the thermal influence on the lattice of the zircon crystals as additional information to interpret the (U-Th)/He chronological data. Details of the laboratory procedure can be found in Lünsdorf and Lünsdorf (2016). The IFORS software was used to evaluate the Raman spectra. Fitted peak widths were corrected for the apparatus function after Irmer (1985) and Nasdala et al. (2001).  Table 1). The crystal sizes with c-axis parallel and perpendicular dimensions range from 120 to 319 μm and 55 to 98 μm, F I G U R E 4 Microphotographs of the dated zircon crystals along with the effective U concentration (eU, where eU is calculated as U + 0.235 * Th; Gordon Gastil, DeLisle, & Morgan, 1967)
The ZHe age distribution is visualized as KDE plot by the DensityPlotter v8.4 software (Figure 7; Vermeesch, 2012). The KDE age spectrum shows a typical left-hand asymmetry and the mean of the dominating (about 75%) youngest age component is 9.33 ± 0.24 Ma ( Figure 7). To further corroborate the result, we also use the SIMPLEX method (Cserepes, 1989) to perform a best-fit model to identify the age components by the Popshare software (Dunkl & Székely, 2002). This approach results in a similar best-fit model age at 9.2 ± 0.8 Ma.

| Zircon reset analysis
In the study area, most of the basalt lava overlies the basement dominated by granitoid rocks. In our study site, the lava covers alluvial   F I G U R E 7 Kernel density plot of the measured zircon (U-Th)/He ages and the best fitted model between the measured ages and calculated ages. Grey curve: kernel density plot of the 20 measured zircon crystals (calculated by DensityPlotter, Vermeesch, 2012); cycles: single zircon crystals; inset shows cumulative plot of ZHe ages; horizontal line in the insert: the real measured single detrital zircon crystals' He-ages; curve in the insert: the best fit line between the real data and the calculated model; K-S test: the Kolmogorov-Smirnov test (method after Press, Flannery, Teukolsky, & Vetterling, 1996); RMS: the goodness of fit between the calculated model and the measured data, the lower the value the better (method after Cserepes, 1989); bins in the insert: the error of the model Ulonska, Schleicher, Pidgeon, and van Bronswijk (2001) and Nasdala, Irmer, and Jonckheere (2002) have found some miscorrelation between the Raman bandwidths and positions. These annealed zircon crystals mostly plot above the peak position-peak width trend established for zircons derived from unheated or slowly cooled geological settings (Nasdala et al., 2001(Nasdala et al., , 2002. In our case, the Raman parameters obtained on sample JB40 plot somewhat off the trend constrained by the three granite samples reflecting the regional cooling history (Figure 8). This property of the lattice of the zircons from the sand sample below the lava flow supports their shock-like thermal reset.
In summary, we can conclude that the detrital zircon crystals have been heated and their ZHe clock became fully reset at the contact with the basalt. The ZHe age of 9.33 ± 0.24 Ma of the sand sample is thus interpreted to represent the eruption age of the overlying basalt lava.

| Relation to other Miocene basalt lava occurrences
Liu distinguished 10 Cenozoic volcanic episodes in NE China, which are listed in Figure 9. According to the measured age, the Huanan basalt lava in this study belongs to the Laoyeling volcanic episode (β N 1 3 , 11-7 Ma), which is characterized by alkali olivine basalt, basanite, and basalt with ultramafic xenoliths. The magma of this volcanic episode mainly originated from partial melting of the upper mantle caused by extension of the East Asian continent, driven by the slab F I G U R E 9 Age and major rock types of the 10 Cenozoic volcanic formations in Northeast China (modified after Liu, 1988). Green bar indicates the age of the basalt eruption dated by the JB40 sample of this study [Colour figure can be viewed at wileyonlinelibrary.com] rollback of the Pacific plate's westward subduction (Xu et al., 2012;Xu & Fan, 2015).

| CONCLUSIONS
1 (U-Th)/He dating of detrital zircon grains from a sand layer directly below a basalt lava flow in the Huanan region reveals a dominant age component of 9.33 ± 0.24 Ma. This implies, together with the Raman data that the reset of the ZHe thermochronometer was caused by the thermal effect of the basalt lava, which erupted at this time.
2 The result also implies that the basalt in the Huanan area belongs to the Laoyeling volcanic episode.