A global rain of micrometeorites following breakup of the L-chondrite parent body—Evidence from solar wind-implanted Ne in fossil extraterrestrial chromite grains from China

Authors


E-mail: carl.alwmark@geol.lu.se

Abstract

Abstract– Previous studies of limestone beds of mid-Ordovician age from both Sweden and China show that the Earth saw an at least two orders of magnitude increase in the influx of extraterrestrial material approximately 470 Ma, following the disruption of an L-chondrite parent body in the asteroid belt. Recovered extraterrestrial material consists of fossil meteorites and sediment-dispersed extraterrestrial chromite (SEC) grains, both with L-chondritic origin. Ne isotope analysis of SEC grains from one of the Swedish limestone sections revealed that the vast majority of the grains were delivered to Earth as micrometeorites. In this study, we extend the previous work, both in time and geographically, by measuring concentrations and isotopic ratios of Ne in individual SEC grains (60–120 μm in diameter) from three different beds from a contemporary Middle Ordovician limestone section in China. All of the Chinese SEC grains, 44 in total, contain surface-implanted Ne of fractionated solar wind composition, implying that these grains were, as in the case of the Swedish SEC grains, delivered to Earth as micrometeorites. This gives further compelling evidence that the two to three orders of magnitude increase in the influx of micrometeoritic material following the breakup of the L-chondrite parent body was indeed a global event. The rain of micrometeorites prevailed for at least 2 Myr (the estimated time of the deposition of the topmost Chinese bed) after the breakup event.

Introduction

The largest documented asteroid breakup event in the history of our solar system occurred at approximately 470 Ma, when an L-chondrite parent body was disrupted after a collision in the asteroid belt (Anders 1964; Heymann 1967; Keil et al. 1994; Korochantseva et al. 2007). The finding of more than 100 fossil L-chondritic meteorites in a Middle Ordovician limestone section, deposited over a period of approximately 2 Myr, in the Thorsberg quarry at Kinnekulle in southern Sweden shows that the meteorite flux to Earth was enhanced by two orders of magnitude for at least a few million years after the disruption event (Schmitz et al. 1997, 2001; Bridges et al. 2007). The meteorites are heavily altered; all original minerals have been replaced by pseudomorphs, with the exception of chromite. The L-chondritic origin of the fossil meteorites was confirmed by elemental and oxygen isotopic analyses of the relict chromite grains as well as elemental analyses of silicate inclusions enclosed in the chromite grains (Schmitz et al. 2001; Alwmark and Schmitz 2009; Heck et al. 2010). Furthermore, the cosmic-ray exposure (CRE) ages based on 21Ne in the chromite grains correlate inversely with the time of deposition of the sediments in which they lie, indicating a common origin, i.e., the parent body breakup event, as well as very rapid transport (<1 Myr) to Earth (Heck et al. 2004).

The two orders of magnitude enhancement in extra terrestrial material during that period was substantiated by the finding of anomalously high concentrations of sediment-dispersed extraterrestrial chromite (SEC) grains of L-chondritic composition in the same Swedish sediment beds as the fossil meteorites (Schmitz et al. 2003; Schmitz and Häggström 2006; Häggström and Schmitz 2007). This was later further corroborated by additional findings of equally high concentrations of L-chondritic SEC grains in both Chinese and Russian contemporary limestones (Korochantsev et al. 2009; Cronholm and Schmitz 2010; Lindskog et al. 2012). Measurements of He and Ne concentrations and isotopic ratios in batches of 5–6 SEC grains from the fossil meteorite-bearing limestone beds in the Thorsberg quarry showed that at least part of the grains were delivered to Earth as micrometeorites or parts thereof, as the batches contained high amounts of noble gases implanted by the solar wind (SW; Heck et al. 2008). The results were further constrained in a recent study by Meier et al. (2010), in which Ne and He concentrations and isotopic ratios were measured on 37 individual SEC grains from the Thorsberg quarry. Essentially all (approximately 95%) of the SEC grains contained He and Ne of SW composition, i.e., the vast majority were micrometeorites. An alternative explanation for the SW-implanted noble gases in the SEC grains is that they were derived from decomposed SW-rich macroscopic regolith-breccia meteorites. This hypothesis was excluded based on abundance and probability arguments (see Meier et al. [2010] and below).

The aim of this study was to extend the previous work by Meier et al. (2010), both geographically and in time, by measuring concentrations and isotopic ratios of Ne in individual SEC grains from three different beds from a contemporary Middle Ordovician Chinese limestone section. The section, which in the mid-Ordovician was situated several thousand kilometers to the east of Kinnekulle, also contains an extraordinarily high concentration of L-chondritic SEC grains (Cronholm and Schmitz 2010). Furthermore, two of the Chinese beds are situated higher and one is situated lower, in the stratigraphy, relative to the sediment beds in the Thorsberg quarry, in which the majority of the SEC grains in the previous study (Meier et al. 2010) were found. This means that, within this study, SEC grains from the L-chondrite breakup event that arrived earlier and later, relative to the SEC grains in the previous study, will be investigated. The analyses of Chinese SEC grains from different intervals will answer the question whether there is any spatial or temporal difference in the amount of SEC grains with a micrometeoritic signature following the breakup event. This will help us to better understand the delivery mechanisms of cosmic material to Earth as well as the L-chondrite parent body breakup event and its terrestrial implications.

Material and Methods

In this study, SEC grains (approximately 60–120 μm in diameter) extracted from Middle Ordovician limestone from the Puxi river section in China (Cronholm and Schmitz 2010), were analyzed for Ne isotopes (Fig. 1a). A total of 44 SEC grains from three different limestone beds (P6, Y10, and P1; Table 1), with increasing deposition times in relation to the breakup event, were analyzed (Fig. 1b); nine SEC grains from the lowermost sample P6, taken at the top of the Lenodus variabilis conodont zone, 20 grains from Y10, which resides in the lower/middle of the conodont zone Microzarkodina hagetiana, and 15 grains from P1, located in the lower part of the Microzarkodina ozarkodella conodont zone (Zhang 1998; Schmitz et al. 2008; Cronholm and Schmitz 2010). In addition to the SEC grains, three terrestrial chromite grains, one from each bed, were analyzed to control for any possible Ne contamination from the sediments. Only one terrestrial chromite grain from each bed could be studied because of the scarcity (1–2 grains per sample) of terrestrial heavy minerals.

Figure 1.

 a) Map showing the location of the sampled sections in China. b) Stratigraphic profile showing the relative order of the Chinese samples and also how it correlates with the Swedish section at Kinnekulle. The correlation of the two profiles is based on conodont biostratigraphy (Zhang 1998; Schmitz et al. 2008; Cronholm and Schmitz 2010). The CRE ages for the two fossil meteorites are from Heck et al. (2008). Both figures are modified after Cronholm and Schmitz (2010).

Table 1.   The grain masses, 20Ne concentrations, and isotopic ratios for all the grains within the study.
Sediment bedSample nameMass (μg)Ne-20 (10−5ccSTP g−1)Ne-20/22Ne-21/22
  1. aDue to the small gas amounts, errors are very large, thus only the upper limits of the Ne-20 concentrations are given and the ratios not determined (n.d.), although they are compatible with atmospheric composition.

  2. Errors are 1σ.

P6P6-Cr21.7 ± 0.17.94 ± 0.4711.14 ± 0.050.0297 ± 0.0001
 P6-Cr31.75 ± 0.150.94 ± 0.0811.67 ± 0.090.0305 ± 0.0004
 P6-Cr42.3 ± 0.11.13 ± 0.0511.57 ± 0.080.0300 ± 0.0003
 P6-Cr51.0 ± 0.37.85 ± 2.3611.54 ± 0.040.0299 ± 0.0002
 P6-Cr62.15 ± 0.154.12 ± 0.2911.07 ± 0.050.0295 ± 0.0003
 P6-Cr70.85 ± 0.153.66 ± 0.6511.18 ± 0.080.0302 ± 0.0005
 P6-Cr80.6 ± 0.113.4 ± 2.211.55 ± 0.050.0304 ± 0.0002
 P6-Cr91.6 ± 0.20.422 ± 0.05311.81 ± 0.200.0306 ± 0.0013
 P6-Cr100.35 ± 0.0519.4 ± 2.811.20 ± 0.050.0302 ± 0.0003
Y10Y10-Cr91.25 ± 0.050.160 ± 0.00711.49 ± 0.490.0309 ± 0.0019
 Y10-Cr70.85 ± 0.0511.4 ± 0.711.38 ± 0.040.0301 ± 0.0002
 Y10-Cr122.35 ± 0.154.59 ± 0.2911.29 ± 0.040.0299 ± 0.0001
 Y10-Cr182.9 ± 0.31.58 ± 0.1611.44 ± 0.050.0302 ± 0.0003
 Y10-Cr191.95 ± 0.250.0089 ± 0.002216.10 ± 7.570.031 ± 0.026
 Y10-Cr201.45 ± 0.254.32 ± 0.7511.38 ± 0.040.0298 ± 0.0002
 Y10-Cr220.85 ± 0.0512.1 ± 0.711.29 ± 0.040.0297 ± 0.0002
 Y10-Cr30.8 ± 0.10.519 ± 0.06512.09 ± 0.270.0316 ± 0.0013
 Y10-Cr50.45 ± 0.050.0212 ± 0.004412.59 ± 7.910.032 ± 0.037
 Y10-Cr112.55 ± 0.057.12 ± 0.1411.51 ± 0.040.0303 ± 0.0001
 Y10-Cr131.95 ± 0.152.51 ± 0.1910.82 ± 0.060.0291 ± 0.0003
 Y10-Cr160.85 ± 0.1512.5 ± 2.211.28 ± 0.040.0298 ± 0.0001
 Y10-Cr171.55 ± 0.259.17 ± 1.4811.52 ± 0.030.0304 ± 0.0001
 Y10-Cr14.1 ± 0.51.79 ± 0.2211.38 ± 0.050.0300 ± 0.0002
 Y10-Cr23.25 ± 0.153.63 ± 0.1711.56 ± 0.030.0304 ± 0.0001
 Y10-Cr41.5 ± 0.23.54 ± 0.4711.20 ± 0.050.0296 ± 0.0002
 Y10-Cr82.85 ± 0.257.28 ± 0.6411.59 ± 0.040.0305 ± 0.0001
 Y10-Cr141.75 ± 0.250.0319 ± 0.004912.80 ± 1.800.0332 ± 0.0091
 Y10-Cr151.85 ± 0.055.40 ± 0.1511.48 ± 0.030.0302 ± 0.0001
 Y10-Cr211.4 ± 0.10.0106 ± 0.001715.04 ± 7.620.036 ± 0.035
P1P1-Cr12.2 ± 0.44.27 ± 0.7811.09 ± 0.060.0300 ± 0.0001
 P1-Cr31.6 ± 0.10.0705 ± 0.004711.75 ± 0.760.0334 ± 0.0045
 P1-Cr62.2 ± 0.20.364 ± 0.03311.44 ± 0.140.0312 ± 0.0008
 P1-Cr81.55 ± 0.150.297 ± 0.02911.31 ± 0.230.0333 ± 0.0027
 P1-Cr131.25 ± 0.052.58 ± 0.1011.49 ± 0.060.0301 ± 0.0002
 P1-Cr174.05 ± 0.150.771 ± 0.02911.09 ± 0.060.0295 ± 0.0002
 P1-Cr190.65 ± 0.151.93 ± 0.4511.69 ± 0.100.0306 ± 0.0006
 P1-Cr111.75 ± 0.051.39 ± 0.0412.06 ± 0.070.0313 ± 0.0003
 P1-Cr42.3 ± 0.20.756 ± 0.06611.42 ± 0.110.0298 ± 0.0005
 P1-Cr52.5 ± 0.31.75 ± 0.2111.75 ± 0.060.0306 ± 0.0003
 P1-Cr70.65 ± 0.152.42 ± 0.5611.06 ± 0.240.0296 ± 0.0005
 P1-Cr95.25 ± 0.050.0613 ± 0.000911.65 ± 0.300.0320 ± 0.0016
 P1-Cr101.75 ± 0.350.918 ± 0.1811.30 ± 0.090.0293 ± 0.0004
 P1-Cr142.25 ± 0.154.66 ± 0.3111.21 ± 0.040.0299 ± 0.0002
 P1-Cr161.75 ± 0.051.62 ± 0.0511.10 ± 0.070.0302 ± 0.0003
OCa grainsOC-P612.05 ± 0.15<0.0021n.d.n.d.
 OC-Y1011.3 ± 0.1<0.0042n.d.n.d.
 OC-P110.25 ± 0.05<0.0073n.d.n.d.

The chromite grains were recovered from limestone samples, weighing between 9 and 21 kg, by dissolving the limestone in hydrochloric and hydrofluoric acid (see Schmitz and Häggström [2006] for details on the dissolution method). The acid-insoluble residue was studied under an optical microscope where suspected chromite grains were picked. The major and minor elemental composition of each grain was determined using an energy dispersive spectrometer (EDX; Inca X-sight from Oxford Instruments) with a Si-detector linked to a scanning electron microscope (SEM; Hitachi S-3400N). This was done to separate the grains with an elemental composition indicating an ordinary chondritic origin from the ones displaying a terrestrial composition. The division is based on the definition of Schmitz and Häggström (2006) for major elemental composition of ordinary chondritic chromite. Although these analyses are performed on unpolished grains, resulting in semi-quantitative data, they are adequate to clearly discern between SEC and terrestrial chromite grains. Grain masses were determined with a microbalance (Table 1). Tare measurements of the empty balance were performed before and after the weighing of each grain. The accuracy was generally between 5 and 20% and was defined as half the difference between the two tare measurements.

We used a low-blank extraction line and an ultra-high-sensitivity mass spectrometer for the gas analysis (Baur 1999; Heck 2005; Meier et al. 2012). The mass spectrometer is equipped with a molecular drag pump (compressor), which concentrates the sample gas into the ion source, giving approximately two orders of magnitude higher sensitivity for Ne than the same instrument would have without a compressor ion source (Baur 1999). The grains were mounted individually in a pitted Al sample holder. The gases were then extracted by melting the grains using an infrared (1064 nm) Nd:YAG-laser in continuous wave mode, for typically 60–90 s. Chemically active gases and the heavy noble gases Ar, Kr, and Xe were removed using metal-oxide getters and N2-cooled activated-charcoal traps. Contributions of species interfering on the three Ne isotope peaks were monitored by measuring also masses 18 (H216O to control for H218O), 40 (40Ar+ to control for 40Ar++), and 44 (CO2+ to control for CO2++). However, interference corrections were always considerably smaller than the statistical uncertainties of the measured ion count rates, and thus neglected. All ion peaks of interest were measured in sequence in several cycles according to the method developed by Heck (2005), where the data points measured prior to extraction, i.e., the memory gas signals, are extrapolated forward to the moment of extraction and the data points measured after extraction, i.e., the signals from sample plus memory gases, are extrapolated backward to the same moment. The difference between the two extrapolated values then constitutes our sample signal. This allows for a better correction for blank contributions than the classical method with separate blank runs, especially in the case of grains with very low gas amounts. Also, the spectrometer memory was accurately monitored. Nevertheless, regular blank runs, by targeting the laser on the empty sample holder, were carried out to control for possible contributions from the sample holder or extraction chamber. These blank runs showed that the contribution was negligible compared with sample Ne amounts, on average 3–4 orders of magnitude lower than the signal of a typical SEC grain. Spectrometer sensitivity and mass discrimination were determined by analyses of known amounts of calibration gas with essentially atmospheric isotopic composition of Ne (Heber et al. 2009), at regular intervals. For further details on the instrument and the analytical procedures, see Heck et al. (2004, 2008), and Meier et al. (2010).

Results

The grain masses, 20Ne concentrations, and isotopic ratios (20Ne/22Ne; 21Ne/22Ne) for both SEC and terrestrial chromite grains are listed in Table 1.

The masses of the SEC grains range from 0.4 to 5.3 μg with a median value of 1.8 μg. There is no clear difference in the grain mass distribution between the different beds, given the limited statistics. All 44 analyzed SEC grains contained detectable amounts of Ne with 20Ne concentrations ranging between 8.9 × 10−8 and 1.9 × 10−4 ccSTP g−1. The Ne three-isotope diagram (Fig. 2) shows that all of the grains contain trapped Ne of solar-like composition, as most data points fall along the line connecting the SW component (20Ne/22Ne = 13.78 ± 0.03, 21Ne/22Ne = 0.0329 ± 0.0001; Heber et al. 2009) with the fractionated SW (fSW; former SEP) component (20Ne/22Ne ≡ 11.2, 21Ne/22Ne = 0.0295 ± 0.0001). The 21Ne/22Ne ratio of fSW was calculated using a mass-dependent fractionation from the SW ratios given by Heber et al. (2009), assuming a 20Ne/22Ne ratio of 11.2. About 65% of the SEC grains plot significantly (more than 1σ) to the right of the line connecting the SW with the fSW points, indicating a discernible contribution of cosmogenic 21Ne. The rest of the data points are within errors, on the mixing line, although the data points themselves fall to the right of the line. This suggests that most of these grains probably contain measurable amounts of cosmogenic Ne, although in many cases this amount cannot be resolved from zero. From the cosmogenic Ne, CRE ages can be calculated. These will be dealt with in a separate forthcoming publication. The data points of a few grains, predominantly from the Y10 bed, plot above or to the right of the SW-fSW line, but they still encompass this line within their large 1σ error bars. These are grains with very small amounts of gas, close to the detection limit.

Figure 2.

 Three-isotope diagrams for Ne for all individual SEC grains from the three sediment beds. All grains plot near the line connecting the fractionated solar wind (fSW) with the solar wind component (SW). The galactic cosmic ray component (GCR) plots outside of the graph (20Ne/22Ne ≈ 0.95, 21Ne/22Ne ≈ 0.85; Wieler 2002). Note that in the lower diagram, five grains are not plotted because of their low gas concentrations, close to the detection limit, resulting in very large error bars. All error bars are 1σ.

That the trapped Ne in the majority of the SEC grains was implanted by the solar wind, while residing in space, is further corroborated by data shown in Fig. 3a, which show for many grains an inverse correlation between mass and 20Ne concentration. Smaller grains tend to have larger concentrations of trapped Ne, which implies that the gases reside near the grain surface, as is the case for the implanted SW Ne (<30 nm; Tamhane and Agrawal 1979). However, some of the grains do not show such a correlation between mass and 20Ne concentrations, i.e., they plot close to the x-axis regardless of mass. This could either be due to partial loss of surface-implanted gas during atmospheric travel or that only part of the surface of the SEC grain was exposed to the SW during its journey to Earth, i.e., the grain was part of a “larger” micrometeorite.

Figure 3.

 a) Ne-20 concentrations versus grain mass. b) Comparison of histograms showing the distribution in mass between the Chinese SEC grains within this study and Swedish SEC grains from Meier et al. (2010).

The size of the three terrestrial chromite grains is similar to that of the SEC grains, with a mass varying between 0.3 and 2 μg. The terrestrial grains all contained typically 3–4 orders of magnitude lower Ne concentrations than the SEC grains, the values for the terrestrial chromite grains being close to or below detection limits, resulting in huge error bars. This shows that Ne contamination from the surrounding sediments is negligible.

Discussion

All of the 44 SEC grains from the Chinese beds contain trapped Ne of solar wind composition. As the solar wind only penetrates a few tens of nm into solid matter, this means that at least parts of the grains’ surfaces were directly exposed to the SW, implying that the grains were delivered to Earth as micrometeorites or surface-exposed parts thereof. A putative alternative explanation for the presence of SW Ne in the grains is that it was implanted while the SEC grains resided in the topmost layer of the parent body regolith and that they then were transported to Earth as parts of macroscopic regolith-breccia meteorites. The SEC grains would then be sediment-dispersed remnants of these decomposed solar gas-rich macroscopic meteorites. This explanation has, however, already been discarded by Meier et al. (2010) for the Swedish grains based on the fact that only approximately 3% of all L-chondrites falling today are asteroidal regolith breccias (Bischoff and Schultz 2004), while approximately 100% of the SEC grains contain SW gases. Also, only one of nine (11%) fossil meteorites analyzed for their noble gas content, from the limestone section in the Thorsberg quarry at Kinnekulle, has so far been identified as a regolith breccia (meteorite Ark 002 [official name: Österplana 002] in Heck et al. 2004, 2008). This finding also indicates that fossil meteorites of regolithic origin are not necessarily more prone to decomposition compared with the non-regolithic fossil meteorites. Finally, many of the grains display concentrations of 20Ne in the typical range observed for recent micrometeorites (Olinger et al. 1990; Osawa and Nagao 2002). Thus, it is likely that the vast majority of the grains were delivered to Earth as micrometeorites. The fact that all of the Chinese SEC grains contained Ne of SW composition also corroborates the interpretation in the study by Cronholm and Schmitz (2010) that the SEC grains are extraterrestrial.

Compositional analyses by Cronholm and Schmitz (2010) on 291 SEC grains from the Chinese section gave an average composition in major elements of (in weight %): Mg: 1.7 ± 0.5, Al: 3.2 ± 0.2, Fe: 21.3 ± 1.4, Cr: 39.6 ± 0.8. This is identical within error (1σ) with compositional data from Schmitz and Häggström (2006) of 276 SEC grains from the Swedish section at Kinnekulle, showing an average of (in weight %): Mg: 1.6 ± 0.5, Al: 3.2 ± 0.4, Fe: 21.3 ± 2.0, Cr: 39.4 ± 1.1. This is evidence that the Chinese SEC grains studied here and the Swedish grains investigated by Meier et al. (2010) belong to the same population, a conclusion that is also supported by the almost identical mass distribution (Fig. 3b) of the grains in the two studies as well as the simultaneous increase in the concentration of extraterrestrial material at the two localities. This corroborates previous results, based on elemental and oxygen isotopic analyses (Schmitz et al. 2008; Heck et al. 2010), that the micrometeorites have a common origin; i.e., the L-chondrite parent body. The high number of SEC grains in the studied samples (1.1, 0.5, and 2.3 SEC grains kg−1, respectively), compared with the background value of 1 SEC grain in 110 kg (Cronholm and Schmitz 2010), together with the fact that all of the SEC grains were delivered to Earth as micrometeorites, demonstrates that the influx of micrometeoritic material to the sections in China was 2–3 orders of magnitude higher during the investigated time interval. As the Puxi river section in China, situated on a continent that in the mid-Ordovician was located several thousands of kilometers to the east of the Swedish Kinnekulle section (Cocks and Torsvik 2002, 2005), also saw this increase, proves that this was indeed a global event.

The majority of the Swedish SEC grains in the previous study by Meier et al. (2010) were from a limestone bed named Sextummen, which is situated in the conodont zone Yangtzeplacognathus crassus (Fig. 1b). The three Chinese beds in this study all derive from different conodont zones compared with Sextummen, i.e., clearly both younger (P6) and older (Y10 and P1) with respect to the breakup event. We can accordingly extend our knowledge of the duration, both backward and forward in time, of the micrometeoritic flux. The exposure age difference between the lowermost (0.2 Ma) and uppermost (0.9 Ma) fossil meteorite in the sediment column in the Thorsberg quarry is 0.7 Ma (Heck et al. 2008). As these two meteorites are separated by about 3 m (Fig. 1b), this yields an average sedimentation rate during that time interval of approximately 4 mm ka−1. On the basis of correlation in conodont biostratigraphy between China and Sweden (Schmitz et al. 2008; Cronholm and Schmitz 2010), we can estimate that the youngest Chinese bed P1 (which is in the lower part of Microzarkodina ozarkodella conodont zone) corresponds to a position approximately 7 m above the initial breakup event as recorded in the Thorsberg quarry. If we then assume that the sedimentation rate of 4 mm ka−1 is constant, the post breakup age for the youngest Chinese bed (P1) is approximately 2 Myr. In other words, the 2–3 orders of magnitude increase in the global influx of micrometeorites to Earth prevailed for at least 2 Myr following the breakup event.

Conclusion

All of the 44 SEC grains, from three different Chinese Middle Ordovician limestone beds, contain Ne of solar wind composition. This means that the grains were delivered to Earth as micrometeorites, similar to the Swedish grains in the study by Meier et al. (2010). This fact confirms that the two to three orders of magnitude increase in the influx of micrometeoritic material following the breakup of the L-chondrite parent body was a global event. The results also reveal that this increased global influx of micrometeorites prevailed for at least 2 Myr (the estimated time of the deposition of the topmost bed P1) after the breakup event.

Acknowledgments— A. Cronholm and S. Holm are thanked for their help with sample preparation. C. A. was funded by a grant from the Swedish Research Council. B. S. was funded by the European Research Council and the Swedish Research Council. M. M. was supported by the Swiss National Science Foundation. R. W. thanks for the hospitality during his stay in Lund. Helpful reviews were provided by P. R. Heck and one anonymous reviewer.

Editorial Handling— Dr. Marc Caffee

Ancillary