RSCM method (O. Beyssac)

Estimating metamorphic temperature using the RSCM method

(RSCM : Raman Spectroscopy of Carbonaceous Material)

Olivier Beyssac - Laboratoire de Géologie, Ecole Normale Supérieure, Paris.

(This page intends to present briefly the main technical and theoretical advices concerning the RSCM method. More details can be found in the references given through the text.)

During diagenesis and metamorphism, CM present in the initial sedimentary rock progressively transforms into graphite (graphitization). The corresponding progressive evolution of degree of organization of the CM is considered to be a reliable indicator of metamorphic grade, especially of temperature [Rietmeijer and Mackinnon 1985, Wopenka and Pasteris 1993]. Because of the irreversible character of graphitization (CM is tending toward the thermodynamic stable phase which is graphite), CM structure is not sensitive to retrograde metamorphism and therefore primarily depends on the maximum temperature reached during metamorphism, whatever the retrograde history of the sample. It also has been observed that samples collected from neighboring outcrops with clearly different strain have the same degree of graphitization, indicating that deformation does not significantly affect the structural organization of the CM [Beyssac et al, 2002a] .
The first-order Raman spectrum of disordered CM exhibits a graphite “G” band at 1580 cm-1, E2g2 mode corresponding to in-plane vibration of aromatic carbons, and several defect bands, corresponding to "physico-chemical defects" (see Beyssac et al. [2003] and references therein). The structural organization of CM can be quantified through the R2 parameter defined as the relative area of the main defect band D1 (R2= D1/(G+D1+D2) peak area ratio). A linear correlation between this R2 parameter and metamorphic temperature was calibrated using samples from different regional metamorphic belts with well-known P-T conditions in the range 330-650°C (RSCM method). RSCM can be applied to metasediments of pelitic lithology in which CM precursor is mainly a kerogen mixed with minor hydrocarbons trapped during diagenesis. The uncertainty on temperature is ± 50°C due to uncertainties on petrologic data used for the calibration (Beyssac et al., 2002b). The relative uncertainties on temperature are much smaller, perhaps 10°C [Beyssac et al, 2004].

Related references:

Beyssac, O., Rouzaud J.N., Goffé B., Brunet F. and C. Chopin Graphitization in high-pressure, low-temperature metamorphic gradient: a HRTEM and Raman microspectroscopy study. Contributions to Mineralogy and Petrology, 143, 19-31, 2002a.

Beyssac O., Goffé B., Chopin, C. and J.N. Rouzaud, Raman spectra of carbonaceous material from metasediments: a new geothermometer. Journal of Metamorphic Geology, 20, 859-871, 2002b.

Beyssac O., Goffé, B., Petitet, J.P., Froigneux, E., Moreau, M. and Rouzaud, J.N. (2003) On the characterization of disordered and heterogeneous carbonaceous materials using Raman spectroscopy. Spectrochimica Acta A, 59, 2267-2276.

Beyssac O., Bollinger L., Avouac J.P. and Goffé B, Peak metamorphism temperatures in the lesser Himalaya (Nepal) from the study of carbonaceous material by Raman spectroscopy, Earth and Planetary Science Letters, 225, 233-241.

Rietmeijer, F.J.M., and I.D.R. Mackinnon, Poorly graphitized carbon as a new cosmothermometer for primitive extraterrestrial materials. Nature, 316, 733-736, 1985.

Wopenka B. and J.D. Pasteris, Structural characterization of kerogens to granulite-facies graphite: Applicability of Raman microprobe spectroscopy. American Mineralogist 78, 533-557, 1993.

Measuring the Raman spectrum of CM

1. The Raman spectrum of CM

Raman spectroscopy is used from the early seventies for the study of carbon materials. The Raman spectrum of CM can be divided in first- and second-order regions as illustrated by Fig. 1 [15-17]. In the first-order region (1100-1800 cm-1), the E2g vibration modes of graphite with D6h4 crystal symmetry are attributed to the vibration of carbons within the polyaromatic structure. The first one, E2g1 mode, is a shear mode and corresponds to the relative vibration of the atoms perpendicularly to the aromatic layers. Because interactions between atoms of different layers are very weak (Van der Waals interactions), this mode occurs at very low frequency (~ 42 cm-1). However, it is very difficult to separate practically this mode from the Rayleigh band and it is therefore rarely studied. The second mode, E2g2 mode, corresponds to the stretching vibration in the aromatic layers. Because the aromatic bond involves very high energy, this mode occurs at an unusually high frequency (1580 cm-1). In practice, it is the studied mode and it is called the G band. In perfect crystalline CM (graphite s.s.), there is only the G band in the first order region (Fig. 1a).
For poorly organized CM or very small crystallite dimensions, additional bands appear in the first-order region around 1150, 1350, 1500 and 1620 cm-1 (Fig. 1b). The 1150 cm-1 component appears only in very poorly organized CM, but its attribution is still strongly debated. The 1350 cm-1 band (D1 band), commonly called the defect band, is the most important feature. This band is attributed to a A1g mode, due to a change in the selection rules for the Raman effect concerning some phonons between the K and M points of the Brillouin zone [15-18]. By comparing the thermal evolution of different carbon materials using HRTEM and Raman spectroscopy, Bény-Bassez and Rouzaud [12] attributed (1) the D1 band to in-plane defects and heteroatoms and (2) the 1500 cm-1 band (D3 band), present only as a very wide band in poorly crystallized CM, to defects outside the plane of aromatic layers like tetrahedral carbons. Lastly, the 1620 cm-1 band (D2 band) makes a shoulder on the G band, but its signification is not yet well understood. It is however important to underline that (1) this band is always present when the D1 band is present, and (2) its intensity decreases with increasing degree of organization.
The second-order region (2200-3400 cm-1) shows several features about 2400, 2700, 2900 and 3300 cm-1 (Fig. 1), attributed to overtone or combination scattering [16-19]. The most important one, near 2700 cm-1 (S1 band), splits into two bands in well crystallized graphite. After Lespade et al. [19], this splitting occurs when CM acquires a triperiodic order.
Different parameters are then useful to estimate the CM degree of organization like the bands position, bands FWHM, D1 / G intensity ratio (R1 ratio), and D1 / (G+D1+D2) area ratio (R2 ratio).

2. Some cautionary advices

Several authors have raised the problem of analytical mismatches when analyzing CM by Raman microspectroscopy [11,20,21]. Being opaque material, CM is characterized by a high extinction coefficient of visible-light. Therefore, the visible-light from the laser beam only excites Raman scattering in the uppermost several tens to one hundred nm of the CM sample [19]. The first problem is thus the sensitivity of CM to the sample preparation, because the structural organization of this superficial part of CM can be affected by the thin section fabrication, especially by the polishing stage. In order to avoid any mechanical disruption of CM, the best solution is to focus the laser on CM situated beneath the surface of a transparent adjacent grain [20], as shown by Fig. 2 and illustrated by Fig. 3. Such CM situated in the section thickness was therefore not exposed during the thin section fabrication.
Figure 4 compares the degree of organization (R2 corresponding to the D1/(G+D1+D2) area ratio, see discussion in the following) of several metamorphic CM measured at the surface of thin section (i.e. polished CM) and below a transparent adjacent grain (i.e. unpolished CM), within the same thin section. For poorly organized CM (R2 higher than 0.5), there is no significant difference but it is difficult to conclude if polishing did really not affect the CM structure, or if there is a strong overlap with the pre-existing defects within CM. On the contrary, in well organized CM (R2 lower than 0.5) the spectra measured at the surface exhibit a higher contribution of the defect bands (D1 and D2). This indicates the formation of new defects created by the thin section fabrication. Therefore, in order to obtain homogeneous results, all measurements of CM using polished sections should be performed below a transparent adjacent grain. In case of experimental products, grinding in an agathe mortar has no effect on CM structure as demonstrated by Salver-Disma et al. [22].
Another consequence of the opaque character of CM is the extreme sensitivity of CM to laser-induced heating [23,24]. It is then necessary to use low laser power (< 5 mW). However, measuring CM within the thin section thickness as described above, allows to use higher laser power, because of the strong heating removal by the surrounding mineral matrix. The laser-induced heating can have spectacular consequences in case of poorly organized CM, which can be locally altered as attested by the observation of crater at the sample surface in reflected light.
Because of a strong structural anisotropy, graphite does not exhibit the same Raman spectrum when the measurement is performed parallel or perpendicularly to the c axis [21,25,26]. In the latter case, which corresponds to a measure in the graphite edge plane, the contribution of the defect bands is higher. This problem of orientation can not be easily solved in case of powder samples. Metamorphic as experimental CM in this study have generally coherent domains with diameters lower than a few microns [4,14]. As the grains size is at least several tens of microns, if not hundreds, it is therefore difficult to assume that the coherent domains lie perpendicularly to the c axis on the glass slide, because they are randomly oriented within the powder grains. As discussed in the following, this point constitutes a severe limit to the characterization of CM powder using Raman microspectroscopy. In case of metamorphic CM, it is possible to avoid this problem of orientation by using oriented thin sections, cut perpendicular to the foliation. It is then possible to obtain an optical axis of the laser beam set perpendicular to the mean CM c axis [27].

Corresponding References and Figures can be found in:

Some related Figures:

	Comparison between the Raman spectrum of CM measured at the surface of the thin section (case 1) and the spectrum measured within the section thickness (case 2). Note the higher relative intensity of the defect band D1 in case (1). To avoid any effect of polishing on the CM structure, Raman spectra of CM must be recorded below a transparent adjacent mineral (quartz...).
	Selection of representative first-order and corresponding second-order regions of Raman spectra from various samples in order of increasing metamorphicgrade. In the first-order region, the narrow band around 1050–1100 cm-1 is due to the overlying transparent mineral.
	(a) First- and corresponding second-order regions of the Raman spectrum of perfect graphite. (b) First- and corresponding second-order regions of the Raman spectrum of disordered CM. To quantify the CM degree of organization, we use the R2 ratio which is equal to the D1/(G+D1+D2) area ratio.
	R2 ratio versus peak metamorphic temperature.

Back to main