K-feldspar is known to be preferentially suitable for 40Ar/39Ar dating (e.g., McIntosh et al. 1990; McDougall and Harrison 1999). The 28.03 ± 0.08 Ma (Jourdan and Renne 2007) Fish Canyon sanidine, for example, is widely used as standard in 40Ar/39Ar geochronology. In impact geochronological studies, K-feldspar (sanidine) has been used as dating material for the approximately 74 Ma Manson impact structure (Iowa, USA; Izett et al. 1998), and the ≥142 Ma Gosses Bluff impact structure (Australia; Milton and Sutter 1987), and recently for the 201 ± 2 Ma Rochechouart impact structure (France; Schmieder et al. 2009b); for a recent discussion and recommendation of the Manson and Gosses Bluff impact ages see Jourdan et al. (2009). The characteristics of the 40Ar/39Ar age plateau for recrystallized K-feldspar in the impact-metamorphosed Ries granite (Fig. 6), together with a representative number of heating steps with concordant step ages (8/10; with an 39Ar release of more than 95%) and a high probability of 99% (Fig. 6; Table 3), provide a robust internal data set. In terms of the plateau shape, the age spectrum is notably similar to the well-defined age plateaux obtained for, e.g., the Pilot (Canada; Bottomley et al. 1990), Rochechouart (France; Schmieder et al. 2009b), or Keurusselkä (Finland; Schmieder et al. 2009c) impact structures. Furthermore, age plateaux showing similar characteristics were obtained, for instance, from Chicxulub melt rocks and microtektites associated with the well-established approximately 65 Ma Cretaceous/Paleogene boundary event (Swisher et al. 1992), as well as the Ries tektites (e.g., Schwarz and Lippolt 2002).
The virtually fresh state of the impact-metamorphosed biotite granite (i.e., there is no indication of stronger alteration of K-feldspar domains in thin section), comparatively large domains of recrystallized K-feldspar (up to 2 cm in protolith crystal length), and the lack of visible fluid inclusions suggest that the internal K-Ar isotopic system is not disturbed by argon loss, recoil effects, or excess argon (e.g., McDougall and Harrison 1999; Kelley 2002; Jourdan et al. 2007a), which is obviously supported by a good fit of the data in inverse isochron plots (Fig. 7). The stable position of the Ries crater on the Swabian-Franconian Alb plateau (S German terrane) suggests that no postimpact thermometamorphism occurred that could have affected the Ries impactites and disturbed the K-feldspar age (or that of suevite glasses). This is underlined by the dating results acquired in the last 50 yrs from mixed and monomineralic melt lithologies from the Nördlinger Ries (see Table 1), as well as for the crystalline rocks of the underlying Paleozoic Moldanubian basement (predominantly about 320 Ma granites, gneisses, amphibolites, and pegmatites; e.g., Jessberger et al. 1978; Staudacher et al. 1982; von Engelhardt and Graup 1984), respectively, ranging in a comparatively narrow frame.
The first 40Ar/39Ar measurement on a single Ries glass sample and a moldavite was performed by Staudacher et al. (1982) yielding ages of 14.98 ± 0.49 (1.04) Ma and 15.21 ± 0.30 (0.50) Ma, respectively, with a mean age of 15.19 ± 0.59 (0.71) Ma. The 40Ar/39Ar ages were calibrated against the fluence monitor Bern4/M with an age of 18.5 ± 0.4 Ma (Jäger 1970); uncertainties of the age monitors are included in external errors, displayed in parentheses in the text. The first Lusatian moldavite age was presented by Lange et al. (1995) with a plateau age value of 14.51 ± 0.08 (0.12) Ma, measured against the HD-B1 monitor. This value is recalculated for the most recent HD-B1 age of 24.18 ± 0.18 Ma (Schwarz and Trieloff 2007a); for recalculation procedure, see Schwarz and Lippolt (2002). A comparison between the age of the Bohemian and a Lusatian moldavite was reported by Schwarz and Lippolt (2002) yielding a weighted mean age of 14.48 ± 0.16 (0.18) Ma for the Bohemian moldavites and 14.36 ± 0.26 (0.28) Ma for the Lusatian moldavite, with a weighted mean of 14.47 ± 0.18 (0.20) Ma for all seven samples (all data recalculated against the new HD-B1 monitor age of 24.18 ± 0.18 Ma). Laurenzi et al. (2003) carried out 4–14 individual measurements on seven individual tektites (six from Bohemia and one from Moravia) that resulted in total fusion ages of 14.51 ± 0.14 (0.16) Ma and 14.46 ± 0.14 (0.16) Ma, respectively; the weighted mean age of all 77 single measurements totals to 14.50 ± 0.17 (0.19) Ma. Sample ages were recalculated by Di Vincenzo and Skála (2009) on the basis of the new age value for the FCTb mineral standard of 28.26 ± 0.18 Ma. Buchner et al. (2003) measured one suevite glass from proximal Ries ejecta (Seelbronn) and two from Graupensand sediments (reworked Ries ejecta in North Alpine Foreland Basin deposits) with a weighted mean age for three individual samples (10 measurements each) of 14.53 ± 0.08 (0.20) Ma, recalculated against the TCs-1 mineral standard (monitor age: 28.34 ± 0.28 Ma, Renne et al. 1998). Abdul Aziz calculated a mean age of 14.88 ± 0.11 (0.15) Ma for 15 suevite glass samples from Otting (replicate measurements) against the FCs monitor of 28.02 ± 0.16 Ma (e.g., Jourdan and Renne 2007). The most recent work on moldavites (Di Vincenzo and Skála 2009) was carried out by both 40Ar/39Ar step-heating (four samples, two from Cheb area and one from Moravia and Bohemia each), and total fusion analyses on five Cheb area moldavites and one from Moravia and Bohemia each; the results of the four step-heating analyses were included in total fusion measurements. In that study, Di Vincenzo and Skála (2009) obtained mean ages of 14.68 ± 0.09 (0.13) Ma for the Cheb area moldavites and of 14.67 ± 0.09 (0.13) and 14.66 ± 0.11 (0.15) Ma for the Bohemia and Moravia moldavites, respectively, with a total weighted mean age of 14.68 ± 0.11 (0.15) Ma, compiled from all seven samples (against FCTb with an age of 28.26 ± 0.18 Ma, calibrated in the study by Di Vincenzo and Skála  against FCs with an age of 28.02 ± 0.16 Ma, see, e.g., Jourdan and Renne 2007). All 40Ar/39Ar dating results are compiled in Table 1.
Figure 8. Weighted mean ages of 40Ar/39Ar dating obtained from mixed-glass samples (suevite), Ries tektites (moldavites), and monomineralic melts (K-feldspar) published since 1995; error bars are 2σ external errors (including monitor uncertainty); see Table 1.
Download figure to PowerPoint
Figure 9. Weighted mean ages of all step-heating plateau ages and total fusion samples obtained by 40Ar/39Ar dating of Ries impact glasses and moldavites, listed in the more recent publications since 1995 (except data of Abdul Aziz et al. 2008; see Discussion); error bars are 2σ external errors (including monitor uncertainty); see Table 1.
Download figure to PowerPoint
For their age calculation, Staudacher et al. (1982) used the monitor Bern4/M that is not directly calibrated against one of the standards used in other 40Ar/39Ar suevite/moldavite age determinations. All other monitors were directly (FCTb and TCs-1 against FCs) or indirectly (HD-B1, BMus/2, and FCs against NL-25, GA1550, Hb3gr) intercalibrated, see Renne et al. (1998), Schwarz and Trieloff (2007a), Jourdan and Renne (2007), and Di Vincenzo and Skála (2009). Due to possible miscalibration of Bern4/M, the results by Staudacher et al. (1982) have not been included in further calculations.
In the study by Abdul Aziz et al. (2008), the 15 total fusion ages of the suevite glass samples exhibit nearly indistinguishable ages with one exception of 17.44 ± 0.26 (0.28) Ma indicating excess argon in this sample, which is commonly observed in impact melt glasses (e.g., von Engelhardt et al. 1969; their fig. 22; Buchner et al. 2003). Likewise, Ries tektites may contain a small portion of extraneous (inherited or excess) argon (see, e.g., Schwarz and Lippolt 2002 and Di Vincenzo and Skála 2009). The presence of excess argon in dating samples tends to increase the apparent age of the sample by up to 2%. Schwarz and Lippolt (2002), for instance, measured a 14.68 Ma plateau age while the (total) age of the same sample is 15.01 Ma. This slightly older apparent age is in the range of the age data presented by Abdul Aziz et al. (2008). The interpretation of the calculated isochron age and the 40Ar/36Ar intercept of total fusion ages used to generate an isochron (as done by Abdul Aziz et al. 2008) is difficult and its validity might be problematic, especially as excess argon in Ries glasses cannot be ruled out (see also the discussion by Di Vincenzo and Skála 2009). Furthermore, when only a single step in a step-heating experiment is contaminated with excess argon (due to sample impurity or absorbed argon) and plotted in an isochron diagram, excess argon as the cause of older apparent ages cannot be clearly identified. Accordingly, the age data of Abdul Aziz et al. (2008) are also not included in the further age calculations (see Figs. 8 and 9).
To compare fission track, K-Ar, and 40Ar/39Ar dating results, the miscalibration between the U and K decay serie(s) has to be taken into account. The K-Ar or 40Ar/39Ar ages have to be considered to be 1% higher compared with the fission track ages (see, e.g., Begemann et al. 2001; Mundil et al. 2006; Schwarz and Trieloff 2007b). Due to relatively large errors of the fission track and conventional K-Ar ages, the data for all measurements agree within uncertainties with the 40Ar/39Ar ages obtained in the last decade. In the case of K-Ar ages, inherited radiogenic 40Ar from microscopic inclusions of older (Paleozoic) basement-derived phases could be considered as a potential source of error (e.g., Staudacher et al. 1982; Jourdan et al. 2007b). Incompletely degassed basement-derived “schlieren” that may have been present in larger glass separates (several 100 mg to g) used for K/Ar analyses of suevite glasses (Buchner et al. 2003) also account for older apparent ages. Therefore, in order to obtain a best value for the Nördlinger Ries impact age, fission track and K-Ar ages were not considered, accordingly.
As discussed above, the (partially recalculated) 40Ar/39Ar age data of this study, by Lange et al. (1995), Schwarz and Lippolt (2002), Laurenzi et al. (2003), Buchner et al. (2003) and Di Vincenzo and Skála (2009) (see Figs. 8 and 9) agree within uncertainties. Thus, we propose a new mean age for the Ries impact event from the largely concordant individual Ries ages published. In a first step, we calculated a mean age for all 13 40Ar/39Ar step-heating plateau ages available, resulting in an age of 14.56 ± 0.22 Ma. Secondly, we calculated a mean age for all 17 total fusion measurements resulting in an age of 14.60 ± 0.18 Ma (see Figs. 8 and 9). Although the numbers of replicate measurements per sample ranged from 4 to 23, each sample was weighted equally. As both mean values (step-heating and total fusion) are in perfect agreement within uncertainties, we calculated a weighted mean value of 14.59 ± 0.20 Ma for the Ries impact (see Figs. 8 and 9) from a total of 30 samples (13 plateau and 17 total fusion ages). The 14.37 ± 0.30 Ma age obtained from the Polsingen granite is about 0.2 Ma younger than this weighted mean age but still within uncertainties at the 2σ level; this effect can be explained by the 14.68 ± 0.11 Ma Ries tektite ages reported by Di Vincenzo and Skála (2009) that slightly increase the mean age for the Nördlinger Ries impact (for summary and references see Table 1, Fig. 8).