We obtained diamond bands and G and D bands of graphite from five ureilites. The diamond band is obtained either solely or together with graphite peaks. Figure 1 shows the typical shape of the Raman spectrum for each ureilite. NWA 3140 and Y-791538 have the lowest intensity diamond bands. Other than the diamond and graphite bands, we occasionally obtained a small peak around 1315 cm−1. It is most likely that this peak is due to lonsdaleite that is hexagonal diamond and one of the polymorphs of diamond and graphite. It has been reported that the FWHM of the lonsdaleite peak is about 5 times wider than that of the diamond band; also, the intensity of the peak is about 500 times less than that of the diamond band and is located around 1324 cm−1 (Smith and Godard 2009). Nevertheless, peaks similar to ours have been detected also by Gogotsi et al. (1998) and Hu et al. (2009), who determined that the peak comes from lonsdaleite. Therefore, we believe that the peak we obtained around 1315 cm−1 indeed originates from lonsdaleite.
Figure 1. Typical Raman spectra (after baseline correction) for thin sections of five ureilites: Alan Hills A77257 (ALHA77257), Shişr 007, Northwest Africa 3140 (NWA 3140), Yamato 790981 (Y-790981), and Yamato 791538 (Y-791538). The x-axis shows wave numbers and the y-axis shows spectrum intensity. All five spectra have a diamond band around 1332 cm−1, a graphite G band around 1580 cm−1, and a graphite D band around 1350 cm−1.
Download figure to PowerPoint
The data for the G and D bands of graphite are shown in Fig. 2 and summarized in Table 1. Many G band positions are around 1580 cm−1 (Table 1), although some scatter toward higher positions is observed (Fig. 2a). This is very different from the data observed for carbonaceous meteorites (Bonal et al. 2006; Busemann et al. 2007; Matsuda et al. 2010), for which the peak positions of the G bands at the higher wave numbers range from 1580 to 1605 cm−1 and are inversely correlated with FWHMG (FWHM of G band) (Busemann et al. 2007). A beautiful inverse correlation was obtained even in a single meteorite, the C3V Allende chondrite (Matsuda et al. 2010b). In addition, FWHMG of carbonaceous chondrites ranges from 50 to 100 cm−1 and is much larger than that of ureilite, for which FWHMG ranges from 20 to 50 cm−1 (Table 1). The higher shift in the G band in carbonaceous chondrites compares to the normal G band position of graphite (1580 cm−1), and an inverse correlation with FWHMG shows that the carbon material in carbonaceous chondrite is poorly graphitized carbon to amorphous carbon (Matsuda et al. 2010b). According to Ferrari and Robertson (2000), the G band position of graphite at about 1580 cm−1 shifts higher to 1600 cm−1 when it evolves to nanocrystalline graphite, and then shifts even lower to 1520 cm−1 when it changes to amorphous carbon. The G band position at around 1580 cm−1 and the small values of FWHMG for ureilites suggest that the graphite in ureilites is well ordered. The downward shift trend for the G band position from 1580 cm−1 with increasing FWHMG of carbonaceous materials in carbonaceous chondrites indicates that they are in a more amorphous state (Matsuda et al. 2010b).
Figure 2. Raman spectra data of G and D band of graphite. (a) FWHMG (full width at half maximum for G band) versus peak position of the G band, (b) FWHMD (full width at half maximum for D band) versus peak position of the D band, and (c) ID/IG (peak intensity ratio of D band to G band) versus peak position of the G band.
Download figure to PowerPoint
Table 1. Average values of Raman data for G and D bands of graphite in ureilites.
|Sample||Peak center (G band) (cm−1)||FWHMG (cm−1)||Peak center (D band) (cm−1)||FWHMD (cm−1)|| ID/IG|
|ALHA77257||1585 ± 5||25 ± 9||1358 ± 14||78 ± 45||1.1 ± 1.2|
| ||(27*)|| ||(27*)|| || |
|NWA 3140||1584 ± 4||24 ± 9||1355 ± 8||49 ± 32||0.50 ± 0.14|
| ||(71*)|| ||(67*)|| || |
|Shişr 007||1583 ± 3||38 ± 14||1356 ± 10||60 ± 38||0.66 ± 0.27|
| ||(47*)|| ||(28*)|| || |
|Y-790981||1587 ± 6||30 ± 14||1362 ± 14||77 ± 54||0.80 ± 0.81|
| ||(54*)|| ||(38*)|| || |
|Y-791538||1581 ± 1||23 ± 6||1355 ± 7||58 ± 22||0.44 ± 0.26|
| ||(41*)|| ||(17*)|| || |
The diagram of FWHMD (FWHM of D band) against the peak position of the D band (Fig. 2b) shows a very rough positive correlation. The D band positions of ureilites are high compared with those in carbonaceous chondrites (Bonal et al. 2006; Busemann et al. 2007; Matsuda et al. 2010), but the FWHMD in ureilites are small (20–130 cm−1 from Table 1) compared with those in carbonaceous chondrites (60–350 cm−1). Again, this indicates that the graphitic structure is more developed in ureilites.
The intensity ratio of the D and G bands (ID/IG) is plotted against G band position in Fig. 2c. It is known that the D and G band intensity ratio is a good indicator of the degree of graphitization (Tuistra and Koenig 1970). Note, however, that the behavior of the D band is complex. The D band intensity is very localized at the graphite surface, especially where the crystalline structure is not perfect even though the G band intensity is uniform over the entire graphite (Pimenta et al. 2007). In line with this, Sadezky et al. (2005) states that the D band intensity may not be an indicator for graphitization, especially in soot, while Zickler et al. (2006) showed that the relationship between the crystallite size and the ID/IG ratio breaks down for crystallite sizes below 2 nm. However, we consider the relationship to still be useful as the graphite in ureilites is well developed judging from Raman spectroscopy.
La = CL(λL) × (ID/IG)−1 (1)
where λL is the wavelength of the excitation laser and CL(λL) is a factor depending on the excitation laser wavelength, given as follows.
CL(λL) = C0 + λL × C1 (2)
Dresselhaus et al. (2000) gave the values C0 = −126 Å and C1 = 0.033 for the VIS region 400 nm < λL < 700 nm. The obtained ID/IG ratios for ureilites are 0.44–1.1 (Table 1), which then corresponds to sizes between 45 to 110 Å (45 ± 49, 99 ± 28, 75 ± 31, 62 ± 63, and 113 ± 67 Å, respectively, for five ureilites in Table 1). The obtained values are surely larger than the lower limit of the crystallite size (20 Å) for the relationship indicated by Zickler et al. (2006). Kagi et al. (1991) reported 70–180 Å for the domain size of graphite, and our values are roughly identical considering the large variations (about 100%) in the ID/IG ratios (Table 1).
The diamond peak positions and their FWHM are shown in Fig. 3 and are summarized in Table 2. Most peak positions are located at 1332 cm−1 (normal diamond), the exception being NWA 3140, for which some peak positions are shifted to higher wave numbers (Fig. 3). The data positions for Y-791538 and Y-790981 are very similar to those obtained by Miyamoto et al. (1993). Shişr 007, Y-791538, and Y-790981 seem to have a narrow range of peak positions and small FWHM (Fig. 3; Table 2). Interestingly, NWA 3140 has peak positions at higher wave numbers with large FWHM. For NWA 3140, Ross et al. (2011) reported an average peak position at 1332.2 cm−1 with an FWHM of 8.0 cm−1. These values are quite different from our data (1336.1 cm−1 for the peak position and an FWHM of 33 cm−1) and show that NWA 3140 is normal. Karczemska et al. (2009) also conducted a Raman study of NWA 3140, and reported the presence of diamond having a high wave number around 1337 cm−1 and large FWHM in some places, although normal peak positions were obtained in other portions. Thus, it appears that NW 3140 is very inhomogeneous regarding diamond. A shift toward a higher wave number seems to be related to the presence of internal stresses in diamond (Knight and White 1989). We will discuss this below in relation to the origin of diamond.
Figure 3. Raman spectra data of diamond. FWHM (full width at half maximum of diamond band) is plotted against the peak position.
Download figure to PowerPoint
Table 2. Average values of Raman data of diamond bands in ureilites.
|Sample||Peak center (cm−1)||FWHM (cm−1)|
|ALHA77257||1332.8 ± 1.3||15.5 ± 8.2|
| ||(47*)|| |
|NWA 3140||1336.1 ± 8.6||33 ± 24|
| ||(17*)|| |
|Shişr 007||1330.3 ± 0.8||9.7 ± 4.0|
| ||(54*)|| |
|Y-790981||1332.0 ± 1.0||11.9 ± 3.2|
| ||(63*)|| |
|Y-791538||1330.6 ± 0.7||9.6 ± 1.9|
| ||(40*)|| |
On the Origin of Diamond in Ureilites
Miyamoto et al. (1993) measured the Raman spectra of diamonds synthesized under shock-induced high pressure and by CVD, as well as the Raman spectra of two ureilites, Y-790981 and Y-791538. In Fig. 4, we compared their measured peak positions and FWHM of synthesized diamonds with our data of diamonds in ureilites. As stated above, our data for the two Yamato ureilites agree well with those obtained by Miyamoto et al. (1993). One datum of ALHA77257 obtained by Miyamoto et al. (1988) also agrees with our data. Our obtained ureilite data are located at the right-bottom corner of the shock-produced diamonds, but they fit much more nicely in the area for CVD diamonds (Fig. 4). This suggests that the origin of diamonds in ureilites is CVD. Similar agreement between FWHM of ureillite and CVD diamonds is also found for other ureilites (Hezel et al. 2008; Le Guillou et al. 2010; Ross et al. 2011). However, Heymann (1989) reported that the FWHM of a diamond in the Canyon Diablo iron meteorite is only 7 + 2 cm−1. Furthermore, El Goresy et al. (2001) reported that the diamond in the Ries crater has a FWHM of only 4.5 cm−1. From this, the above authors concluded that the FWHM is not a diagnostic tool for determining the origin of diamond. The problem could be related to the duration time of the meteorite impact shock being 103–106 times longer than that in the laboratory experiments, and makes the Raman band of meteorite diamond narrower compared with the shock-produced diamond in the laboratory (Hezel et al. 2008).
While the FWHM of the Raman diamond peak may not be a crucial tool for determining the origin of diamond, the peak position could be such a tool. It is apparent in Fig. 3 that for NWA 3140 it is shifted toward higher wave numbers for large FWHM. In general, it is easy to understand the downshift of the peak position with larger FWHM. This is due to the effect of small grain size (e.g., Yoshikawa et al. 1993), the presence of lonsdaleite (e.g., Miyamoto et al. 1993), and the heating effect of laser power (e.g., El Goresy et al. 2001). Surely, the shock-produced diamonds are located in the area of lower wave numbers and larger FWHM (the left-upper part in Fig. 4).
Meanwhile, our ureilite diamond data show the trend that the peak position shifts to a higher wave number with larger FWHM (Fig. 4). This trend is also observed even in a single ureilite such as ALHA77257 and NWA 3140 (Fig. 3). The shift toward higher wave number indicates the presence of internal stress in diamond caused by a mismatch between diamond and substrate and is often observed in CVD diamonds (Knight and White 1989; Miyamoto et al. 1993). Knight and White (1989) reported that the effect is most noticeable in diamond films deposited on hard substrates such as alumina or carbides. Matsuda et al. (1991) proposed that formation of diamond in ureilites occurred on high-temperature condensates in the nebula. This model can well explain the chemical feature of carbon vein material, e.g., the enrichment in refractory siderophiles such as Re, Ir, W, etc. (high-temperature condensates) and carbon and noble gasses (Janssens et al. 1987; Matsuda et al. 1991). Thus, the shift of the peak center to a higher wave number is easily explained with this CVD model, but would be difficult using the impact shock model.
The main reason to insist on the shock-origin hypothesis has been that diamond in ureilites is always found in conjunction with graphite (Hezel et al. 2008; Ross et al. 2011). However, this is not strong evidence for the shock-origin hypothesis. It is not easy to produce pure CVD diamond under a plasma condition. We produced the CVD diamonds from a gaseous mixture of H2 and CH4 (Fukunaga et al. 1987; Matsuda et al. 1991). Graphite is easily produced together with the diamond when the proportions of H2 and CH4 are slightly changed. Especially, high CH4 content seems to favor the growth of graphite (Fukunaga et al. 1987). Thus, the presence of diamond in conjunction with graphite is not a unique indicator for the shock-origin hypothesis. Even in the primitive solar nebula, diamond and graphite could have been easily produced together with a slight change in nebula condition.
Nakamuta and Aoki (2000) reported that the basal spacing for part of the graphite coexisting with diamond is slightly smaller compared with normal spacing, which has also been used as evidence for shock origin (e.g., Hezel et al. 2008). However, if the shock occurred for a mixed material of graphite and diamond, the graphite in the vicinity of diamond should be compressed because of the hardness of diamond. Thus, this again does not strongly support the shock-origin hypothesis. Nakamuta and Aoki (2000) wrote in their abstract that the intensity of diamond to graphite is correlated with the shock level of the meteorites. However, the shock level of ALHA77257, which had highest diamond-to-graphite ratio, is less than that of Y-791538, and the correlation is not perfect. As written in Matsuda et al. (1991), the diamond content is about the same (approximately 50% of the total carbon) in highly shocked ureilite (Goalpara) and in moderately shocked ureilite (Novo Urei). Takeda et al. (2001) reported the presence of diamond even in DaG 868 that is supposed one of the most weakly shocked ureilite and that the amount of diamond is comparable to that of ALHA77257. Takeda et al. (2001) proposed a catalytic transformation of graphite to diamond at relatively low pressure for the origin of diamond in DaG 868, but the diamond yield is surely proportional to the shock pressure in the laboratory experiments (Matsuda et al. 1995) and it is curious that the diamond-to-graphite ratio has no correlation with the shock level. To support the shock origin for ureilite diamonds it is often cited that there is no diamond in the ureilite ALHA78019 (very low-shocked), but the detection of diamond is not so simple (Matsuda et al. 1991). Ott et al. (1984) reported the presence of diamond in lightly shocked ureilite Nilpena although Jaques and Fitzgerald (1982) reported that there was no diamond in it.
The most severe constraints for the origin of diamond in ureilites are the noble gasses and nitrogen. Shock traps the noble gasses in diamond within a closed system, but it is difficult to explain the fractionation and the trapping efficiency of noble gasses using the shock model (Matsuda et al. 1995). Most telling, the different nitrogen isotopic compositions of diamond and graphite (and amorphous carbon) are very difficult to explain by the shock model (Rai et al. 2002, 2003a). Thus, Le Guillou et al. (2010) have admitted that there are CVD diamonds in a condensation process. Meanwhile, they consider that amorphous carbon could be the product of diamond post-shock annealing, but again, it is difficult to explain the difference in nitrogen isotopes between amorphous carbon and diamond. Amorphous carbon is also produced by CVD, and the elemental abundance patterns of noble gasses are similar to those in CVD diamond (Fukunaga and Matsuda 1997). Thus, we suggest that graphite, amorphous carbon, and diamond are directly formed on the early condensates in the primitive solar nebula. This concept well explains the nitrogen, and noble gas feature of the carbon vein in ureilites. It is likely that the chemical conditions and ionization mechanism necessary to form CVD diamond and other carbon are present in the primitive solar nebula (Matsuda et al. 1988, 1991). The CVD diamonds (and other carbon phases, too) are produced by the decomposition of hydrocarbons of gas phases under the thermodynamically metastable condition. It is interesting that the decomposition requires the presence of a large amount of hydrogen. The major chemical component of the primitive solar nebula is hydrogen, which is very suitable to produce CVD carbon phases. The formation of diamond, amorphous carbon, and graphite depends on the sp3/sp2 bond ratio of the raw material carbon phase. The carbon having sp3 hybrid orbital favors the diamond formation but that having sp2 bond favors the graphite formation. In the primitive solar nebula, CO having sp2 bond is dominant at high-temperature conditions, but it gradually transforms to CH4 having sp3 bond as the temperature decreases. The reaction of CO to CH4 depends also on the total pressure of the primitive solar nebula. The CO begins to transform to CH4 at about 1000–1300 K when the total pressure is 10−3 to 10−2 atm (Matsuda et al. 1991). Laboratory-synthesized CVD diamonds are produced at these temperatures. If the total pressure of the nebula is lower, the transformation starts at lower temperatures. The ionization mechanism in the primitive solar nebula is possible such as a hot plasma conditions (Arrhenius and Alfvén 1971), lightning, solar wind, cosmic rays, etc. (Matsuda et al. 1991). After this deposition of carbonaceous materials, a shock event with silicate minerals occurred in some later stage on a parent body and the compressed feature of graphite in the vicinity of diamond formed. It is also likely that the boundaries of silicate minerals were filled with the CVD carbonaceous material at this shock event.