By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Wiley Online Library will be unavailable on Saturday 7th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 08.00 EDT / 13.00 BST / 17:30 IST / 20.00 SGT and Sunday 8th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 06.00 EDT / 11.00 BST / 15:30 IST / 18.00 SGT for essential maintenance. Apologies for the inconvenience.
The Vargeão impact structure was formed in the Serra Geral basaltic and rhyodacitic to rhyolitic lava flows of southern Brazil, that belong to the Paraná-Etendeka large igneous province. The Chapecó-type rhyodacites contain small baddeleyite crystals recently dated at 134.3 (±0.8) Ma, which is regarded as the age of this acid volcanism coeval to the flood basalt eruption. Inside the impact structure, a brecciated rhyodacitic sample displays fine veinlets containing numerous lithic fragments in a former melt. This impact breccia contains newly formed zircons, either in the veins or at the contact between a vein and the volcanic matrix. The zircons are 10–50 μm in length, clear and nearly unzoned. In situ laser-ablation dating of the zircons provides a concordant Early Aptian age of 123.0 ± 1.4 Ma that is regarded as the age of the impact event. As in situ age determination ensures the best possible selection of the analyzed mineral grains, the methodology employed in this study also represents a promising method for dating other impact structures.
Large impact events can modify significantly the Earth's surface, producing extreme disturbances in the geosphere and the biosphere (e.g., French 2004). The age of an impact crater is a critical parameter to correlate the impact event with its effects on natural systems. However, precise and accurate dating of terrestrial impact events has often proved difficult (e.g., Jourdan et al. 2009, 2012). According to Jourdan et al. (2009), only 11 of the 174 then confirmed impact structures on Earth had radiometric ages that can be considered statistically and geologically robust. In South America, only the Araguainha impact structure has been reliably dated, first by Hammerschmidt and Engelhardt (1995) and then by Tohver et al. (2012). The ages of the remaining eight confirmed impact structures on the continent remain poorly constrained.
The Vargeão Dome is one of the biggest impact structures in Brazil (http://www.passc.net/EarthImpactDatabase). Since this impact affects a basaltic province, it is particularly interesting as an analog to other impact structures on rocky planets and satellites. The Vargeão structure was recently described by Crósta et al. (2012). Although most shocked volcanic rocks in Vargeão are basalts, the subordinate Chapecó-type rhyodacites were also locally affected by the impact. This felsic volcanism was recently dated by Janasi et al. (2011), who obtained a precise U-Pb age of 134.3 ± 0.8 Ma on tiny baddeleyite crystals separated from the Ourinhos rhyodacites in the northern Paraná basin. This age overlaps the 40Ar-39Ar ages (134.6 ± 0.6 Ma) reported by Thiele and Vasconcelos (2010) for the flood basalts, pointing to a very rapid eruption of the Etendeka-Paraná bimodal igneous province, a common observation for large igneous provinces worldwide (e.g., Larsen and Tegner 2006; Chenet et al. 2007).
Because they are already well-dated and due to their higher Zr contents with respect to basalts (Table 1), the rhyodacites were the preferred target to look for impact-produced zircons, which would provide a way to date the impact event with accuracy. Impact dating of newly formed zircons has been performed only in a few cases, e.g., Vredefort (Gibson et al. 1997), and Manicouagan, Morokweng, and Sudbury (reported by Wielicki et al. 2012), and appears as a very promising method. Here, we have focused particularly on impact-related brecciated veinlets with newly formed zircons in a shocked rhyodacite, thus avoiding inherited ages. In situ U/Pb dating of impact-produced zircons is presented here for the first time, a method that provides both precise and reliable ages and ensures the best possible selection of the analyzed minerals, thus avoiding any contamination by inherited crystals. Thermometry of associated quartz is then performed to investigate the high-temperature processes related to the impact event.
Table 1. Composition of reference rhyodacite (VG4) and selected in situ analyses of melt (glass) in VG5 pseudotachylitic veins
The so-called Vargeão Dome is centered on 26°49′S and 52°10′W in Santa Catarina state, in the southern-central part of the Paraná Basin (Fig. 1a). This impact structure is 12.4 km in diameter and appears as a circular depression with a concentric multiring pattern recognized in remote sensing images (Crósta et al. 2012). It was formed in the Cretaceous Serra Geral Formation, the stratigraphic unit corresponding to the Paraná-Etendeka large igneous province. Volcanics include mainly tholeiitic flood basalts and minor felsic rocks emplaced on top of the Jurassic-Cretaceous Botucatu sandstones (Piccirillo and Melfi 1988; Fig. 1b). The maximum thickness of the lava pile is estimated at approximately 1500 m. Tholeiitic basalts represent approximately 90% of the whole pile. Acid volcanics (rhyodacites and rhyolites) are only abundant in the southern part of the Basin. They are easily classified into two types, namely the Palmas type that is usually aphyric, and the Chapecó type that contains conspicuous plagioclase phenocrysts. Some prevolcanic sandstones were uplifted in the center of the Vargeão structure as a consequence of the impact (Crósta et al. 2012); Fig. 1c). Thin intercalations of these sedimentary rocks are also found among the basaltic flows of the Paraná province and of its prerifting Etendeka counterpart, indicating that the conditions were still arid at the time of basalt flooding (Jerram and Stollhofen 2002).
Crósta et al. (2012) presented diagnostic features of shock metamorphism, such as shatter cones and planar deformation features in quartz. They identified brecciated basaltic rocks in the inner part of the structure. These rocks contain red oxidized veinlets (1 mm to 1 cm in width) with numerous lithic fragments. Melt portions were recognized in these brecciated basalts. Chapecó-type volcanics occur undeformed to the north of the impact structure. Large tilted or collapsed blocks of rhyodacite were observed at the rim of the impact structure.
Material and Methods
Brecciated Chapecó rhyodacite (VG5) was sampled inside the Vargeão impact structure at its southern border (Fig. 1c). The GPS coordinates of this sampling site are 7027410 and 382381 (UTM). The impact melt veins can easily be missed in the field. However, they are recognized in hand sample by their conspicuous purple red color (Fig. 1d). This sample was sliced repeatedly to generate a sufficient number of polished thin sections for microscopic observations and analyses. Five thin sections (VG5-B0 to B4) were used for dating. Another sample of nonbrecciated Chapecó-type rhyodacite (VG4) was collected for comparison along the Chapecozinho River, i.e., at a topographically and stratigraphically lower level than the brecciated sample (Figs. 1b and 1c).
The composition of the nonbrecciated rhyodacite reference sample (VG4) was obtained at SARM (Service d'Analyse des Roches et Minéraux) in the CRPG laboratory (Nancy, France). Major element contents are obtained with a precision better than 1% and Zr content with a precision better than 5%.
Phenocryst and melt compositions were determined with the CAMECA SX50 electron microprobe of the GET laboratory (Toulouse) operating at the usual conditions (beam current 20 nA for anhydrous minerals and 10 nA for hydrous minerals and glasses). Precision on the data is about 1%.
Baddeleyite and zircon were identified with the help of backscattered electron imaging (BSE) and EDS analyses using the JEOL 6360 scanning electron microscope of the GET laboratory. Operating conditions were: an accelerating voltage of 20 kV, magnifications between 40 and 1000×, and a spot size of 50 μm.
The U–Pb geochronology of zircon was conducted using LA-ICP-MS at the Magmas & Volcans laboratory (Clermont-Ferrand, France). Analytical conditions are described in detail in Table S1. The method used for the zircon isotope dating with LA-ICP-MS is basically similar to that developed for monazite (Paquette and Tiepolo 2007). Laser ablation was directly operated on polished petrographic thin sections to preserve textural relationships. A spot diameter of 20 μm associated with a repetition rate of 3 Hz and a laser energy of 4 mJ producing a fluence of 9.5 J cm−2 were used for zircon dating. Concordia ages and diagrams were generated using the Isoplot/Ex v.2.49 software package (Ludwig 2001). Uranium, Th, and Pb concentrations were calibrated relative to the certified GJ-1 zircon standard (Jackson et al. 2004). The zircon analyses were projected on 207Pb/206Pb versus 238U/206Pb diagrams (Tera and Wasserburg 1972), where the analytical points plot along a mixing line between the common Pb composition at the upper intercept and the zircon age at the lower intercept. This method is commonly used to date Phanerozoic zircons using in situ techniques (Claoué-Long et al. 1995; Jackson et al. 2004).
The estimation of minimum melt crystallization temperatures was performed indirectly using the Ti-in-quartz (TitaniQ) thermometer (Wark and Watson 2006) on the quartz crystallized in the brecciated veins. Titanium contents in quartz were analyzed with the electron microprobe of the GET laboratory operating at an accelerating voltage of 15 kV, a beam current of 200 nA, and 300 s counting time. The detection limit for Ti is 50 ppm; the precision interval for each analysis is given in Table 5. Temperatures were determined graphically using Ti-isopleths in the diagram of Thomas et al. (2010).
Petrography and Major-Element Chemistry
The nonbrecciated Chapecó-type rhyodacite (VG4) is a grayish to reddish rock with a porphyritic texture. The phenocryst assemblage comprises plagioclase (3–12 mm of length), light green augite (2–5 mm), and iron oxides (Ti-magnetite and ilmenite, locally with skeletal and stellar-like shapes). Apatite is present as a microphenocryst. The groundmass is felsitic and contains numerous small iron oxide grains as well as a few clinopyroxene crystals. This mineral assemblage is typical of a water-undersaturated high-temperature magma. Baddeleyite ZrO2 (and not zircon ZrSiO4) is the main Zr-bearing mineral in the groundmass, as reported by Janasi et al. (2011), who used baddeleyite crystals to date the Chapecó-type rhyodacite. The presence of baddeleyite in a high-silica rock is unusual and had been reported only once so far, namely from the Yellowstone rhyolites (Bindeman and Valley 2001). Indeed, baddelyite typically occurs in silica-undersaturated rocks, but can be present also in high-temperature hybrid melts or in silicic differentiates from mafic magmas. This is probably the case for the Chapecó-type and Yellowstone high-temperature silicic magmas that are genetically linked to mafic magmas of mantle plume origin (Cordani et al. 1988; Hildreth et al. 1991). The whole-rock composition of sample VG4 is presented in Table 1. With 64.34 wt% silica, this reference rock is in the typical composition range of the Chapecó-type rhyodacites, i.e., 63–69 wt% SiO2 (with a peak at 64–65 wt%) after Piccirillo et al. (1988). Pyroxene compositions are presented in Table 2 and feldspar compositions in Table 3. Pyroxene phenocrysts are poorly zoned augite with XMg = 0.67–0.63, in agreement with the data of Secco et al. (1988). The plagioclase phenocrysts are made of calcic andesine, generally homogeneous in composition (An43–42). These An contents are in the usual range for the Chapecó-type rhyodacites (An47–40 after Janasi et al. 2007).
Table 2. Representative pyroxene analyses and structural formulae on the basis of 6 O
Table 3. Representative feldspar analyses and structural formulae on the basis of 8 O
Fragments in brecciated veinlet
The nonbrecciated part of sample VG5 is very similar to the reference sample in terms of mineral composition. The main difference is the existence of some irregular magmatic banding. Pyroxene phenocrysts are augite with XMg = 0.60–0.59 (Table 2) and plagioclase phenocrysts are andesine of An44–42 composition (Table 3). The rock contains numerous baddeleyite crystals, that are always small (10–20 μm of length) and blade-shaped as described in the samples dated by Janasi et al. (2011), but no early magmatic zircon could be identified.
The brecciated part contains veins a few mm to 1 cm in width. The larger veins are rich in rock and phenocryst fragments (Fig. 2a). In thin section, the lithic fragments are often darkened (Fig. 2a). The plagioclase fragments seem mostly unchanged (Table 3), whereas the pyroxene fragments locally appear brownish and destabilized i.e., decomposed with formation of new iron oxides. Compositions of the pyroxene fragments are variable. Some of them appear unmodified, whereas others contain a variable proportion of sodium (Na = 0.1–0.6 atom per formula unit) and are actually aegerine-augite (Table 2). Finer veins branch off the larger veins. These small veins generally do not contain any lithic fragments. They are filled with newly crystallized quartz or other phases replacing the former melt.
In the veins, the melt comprises a noticeable proportion of glass besides newly formed minerals. Microprobe analyses of glass are highly variable. They range from 46 to 94 wt% silica (Table 1). Such a high variability has been observed in several impact structures (e.g., Zhamanshin in Kazakhstan, Dressler and Reimold 2001) and suggests in situ melting rather than tapping a large pool of melted and mixed rocks. Moreover, two of the glass compositions (P20 and P21, Table 1) resemble a granite in composition, i.e., they are consistent with partial melting of the rhyodacite, rather than melting of basalt or sandstone, the two other target lithologies. In addition, the studied sample is located rather far from the center of the impact structure, i.e., quite far from the zone where shock melting could have been effective. Although the distinction between shock melting and friction melting may be difficult in thin section (Reimold 1998), based on our observations, we consider these veins as pseudotachylitic breccias resulting from friction melting formed immediately following the impact event.
Quartz is an abundant new phase, often crystallized as lath-shaped crystals, locally grown perpendicularly to the vein border (Figs. 2b and 2c). Zircon grains were observed inside or along the impact veins, but never in the unbrecciated part of the rhyodacite. This observation demonstrates their direct relation to the impact event. They are often associated with quartz (Figs. 2b and 2c). These new crystals are either euhedral (Fig. 2d) to subhedral in shape (Figs. 2b, 2c, and 2f). Their sizes range from 10 to 50 μm. Zircon grain in Figs. 2f and 2g display elbow twinning (Jocelyn and Pidgeon 1974). The zircon crystal of Figs. 2b and 2c is rooted in the contact zone between the vein and the rhyodacite groundmass. In other places, anhedral zircon crystals occur in the groundmass, always close to a vein border. As the thin section provides only a 2-D observation, it is not possible to appreciate what their exact relations to the melt vein are. SEM observations show that the zircons are unzoned or poorly zoned (Fig. 2g).
Results of in Situ U/PB Zircon Datinga
Twelve impact-related zircon grains in five different thin sections have been selected for analysis on the basis of location, texture, and size criteria. Eight analyses are duplicates on the same grain (e.g., Z2-1 and Z2-2 in Table 4). Hence, a total of 20 spot analyses were plotted into a 207Pb/206Pb versus 238U/206Pb diagram (Tera and Wasserburg 1972). They plot along a Discordia line indicating a lower intercept with the Concordia at 122.0 ± 1.3 Ma (Table 4 and Fig. 3). The upper intercept is anchored to the isotopic composition of common Pb at 123 Ma (207Pb/206Pb = 0.84 ± 0.08). The lower intercept age is lowered by 0.1 Ma if a present-day isotopic composition for common Pb is considered. When excluding the three points containing a significant amount of common Pb (white ellipses in Fig. 3), the mean 206Pb/238U isotope ratios of the 17 remaining analyses yield a precise value of 123.1 ± 0.8 Ma (MSWD = 1.04). Among the 20 analyses, six analyses performed on five different zircon grains are concordant and yield a Concordia age at 123.0 ± 1.4 Ma (Fig. 3). Although lower intercept, weighted mean 207Pb/238U isotope ratios, and concordant points produce similar ages within error limits, we consider the 123.0 ± 1.4 Ma age as the more robust crystallization age, because it was calculated from the concordant zircon grains.
Table 4. U-Pb data from impact-related zircons by in situ laser-ablation ICP-MS
Decay constants determined by Jaffrey et al. (1971) and recommended by Steiger and Jäger (1977).
Interestingly, Th and U concentrations are particularly high (Table 4), ranging from 290 to 3910 ppm for Th, with a mean value of 2080 ± 1150 ppm (1 standard deviation), and from 120 to 1850 ppm for U with a mean value of 910 ± 490 ppm (1 standard deviation). These contents allow calculating a Th/U mean ratio of 2.3 ± 0.5, which is also unusually high for a silica-rich magmatic rock. According to Wang et al. (2011), increasing magma temperature should promote higher Th content relative to U content, which suggests that all these zircon grains are crystallized at particularly high temperature during a single event at 123.0 ± 1.4 Ma.
Titanium is one of the elements that substitute for Si in quartz. Titanium contents of quartz crystals spatially associated or enclosing the impact-produced zircons in the veins are given in Table 5. Twenty-one analyses were obtained with 1–4 analyses per quartz crystal. Ti contents range from 1033 to 9 ppm, with an average value of 336 ± 210 ppm (1 standard deviation). The distribution of Ti contents is presented in Fig 4a. It is typically unimodal with a well-defined, although asymmetrical, peak corresponding to values in the range 301–350 pm.
Table 5. Analyses of Ti in impact-related quartz from VG5
Weight Ti (ppm)
Detection limit (ppm)
Using fig. 8b of Thomas et al. (2010) at atmospheric pressure.
Wark and Watson (2006) calibrated the temperature dependence of this substitution as a “TitaniQ” thermometer at 1.0 GPa. The pressure dependence of this thermometer was then checked by Thomas et al. (2010). These authors established that, if pressure can be constrained to approximately ±120 MPa, temperature can be constrained to approximately ±20 °C. Their isopleths are reproduced in Fig. 4b for a TiO2 activity of 0.5, recommended by the authors for cases where no rutile is present. We considered that the melt veins crystallized during postimpact deformation, i.e., during uplift and collapse of the impact structure. The corresponding average temperature is 710°C (−120°, +70°, 1 standard deviation).
Source of Zr for Crystallization of Impact-Related Zircon
The location and textural relationships of zircon in the brecciated veins are compelling evidence for their syn- to immediately postimpact nature. No early magmatic (i.e., coeval to lava emplacement) zircon grains were observed in the studied rhyodacite samples, where the Zr-bearing phase is essentially baddeleyite. Janasi et al. (2011) reported the existence of rare late, interstitial zircon grains in the Chapecó-type rhyodacites from Ourinhos, despite the fact that baddeleyite is the main Zr-bearing mineral. We stress that the euhedral to subhedral zircon crystals observed in the studied sample VG5 from Vargeão impact structure are not inherited, but crystallized freely in the melt vein as demonstrated by textural relationships. For instance, the observed delicately elbow-twinned zircons could not have survived the impact deformation and obviously crystallized from the melt in the vein. Such rare twinned zircons are indeed regarded as magmatic and not metamorphic (Jocelyn and Pidgeon 1974). As baddeleyite is a very high-temperature mineral (Heaman and LeCheminant 1993), pre-existing baddeleyite crystals may have survived impact melting. Indeed a few baddeleyite crystals were observed in the brecciated veins without any sign of their transformation in zircon. Hence, baddeleyite may not have released Zr necessary for zircon crystallization. However, pyroxene often contains Zr as a trace element (Bea et al. 2006; Janasi et al. 2011) and likely constituted the former Zr reservoir for new zircon growth.
Temperature of Zircon Crystallization
The TitaniQ temperature estimate of approximately 710 °C is consistent with quartz crystallization from a felsic impact-generated melt. Nevertheless, it must be noticed that this may be only a minimum estimate, because equilibrium conditions may not have been reached if the postimpact undercooling had been both large and fast. Frequently, zircon and quartz crystals grew directionally from the margin of the vein toward the center, a typical feature of rapid crystallization in an undercooled magma body (London 2008). In addition, from textural relationships, zircon crystallized earlier than quartz in these silica-rich melt veins. Therefore, the TitaniQ thermometry provides only a minimum estimate for the zircon crystallization temperature. A higher temperature for crystallization of the new zircons is also suggested by their unusually high Th/U ratios.
None of the 20 analyzed zircon crystals yielded an age corresponding to the formation of the Chapecó-type rhyodacites (i.e., 134.3 ± 0.8 Ma). Among these analyses, six analyses on five different zircon grains are concordant and yield a Concordia age of 123.0 ± 1.4 Ma, which is therefore the crystallization age of these zircons. This age is significantly younger than the age of the Serra Geral basalts and rhyodacites. As the analyzed zircons obviously crystallized from a frictional melt that was generated immediately after the impact, their age can also be regarded as a robust age for the Vargeão impact. No additional rejuvenation event occurred to have modified this age.
The U-Pb zircon dating is the most powerful geochronometer, but its application to impact dating is restricted because of the scarcity of authigenic zircons. In some cases, preimpact zircon grains are only reset. In addition, this method was often tentatively applied to zircon separates from impactites that contain a mixture of inherited zircon grains from various sources as well as a few newly grown zircons, therefore dramatically reducing the chance to obtain a precise age (Ferrière et al. 2010). We stress that accurate dating of impact-produced zircon can be greatly improved by using the in situ dating technique after detailed (optical and SEM) microscopic study, a method that enables us to check the authigenic character of the analyzed grains.
A sample of Chapecó-type rhyodacite from the Vargeão impact structure displays breccia veinlets filled with an impact-generated melt and lithic fragments. The melt, now partly crystallized as quartz, alkali feldspar, and aegerine-augite, also contains newly formed zircons, whose textural relationships unambiguously prove their impact-related nature. The high Th/U ratios of the zircon grains, as well as additional Ti-in Q thermometric estimates on quartz grown from the melt, confirm the high-temperature crystallization of the impact-related zircons. In situ laser-ablation dating of zircon grains provides a concordant Early Aptian age of 123.0 ± 1.4 Ma, which is regarded as a robust age for the Vargeão impact event. Because it enables us to avoid inherited crystals, the in situ dating method appears to be appropriate to provide robust ages for other confirmed impact structures in the future.
A. Nédélec and D. Baratoux received funding from the PNP (Programme National de Planétologie) of the INSU (Institut National des Sciences de l'Univers, Paris, France). E. Yokoyama thanks the CAPES (Coordenação de Aperfeiçoamento de Pessoal de Nível Superior, Brazil) for support. C. Cavaré-Hester, S. Gouy, and L. Menjot are warmly thanked for their technical assistance. Reviews of the manuscript by A. Crósta and E. Tohver and careful editing by W. U. Reimold are acknowledged.