Thermal effects of the excitation laser power on carbonaceous meteorite Northwest Africa 6603 by Raman spectroscopy: an undergraduate research project

Abstract Raman spectroscopy has been used extensively on meteoritic samples since it is a nondestructive tool that provides information about their structure and mineralogical composition, which can give important clues about planet formation. However, the power of the excitation laser used in this technique can alter the properties of the samples due to thermal effects. In this undergraduate research work, the laser-induced thermal effects produced on the carbonaceous chondritic meteorite Northwest Africa 6603 were studied in detail by analyzing the low- and high-resolution Raman spectra parameters of the minerals found in the inclusions and matrix of this sample as a function of the excitation power. Olivine (forsterite), graphitic carbon, pyroxene (enstatite), hematite and gehlenite were the minerals identified in the studied regions. The Raman parameters of these minerals were affected by the laser power to a greater or lesser extent, indicating an increase in structural disorder. In general, the alterations observed were permanent (reverse changes were not observed). These thermal effects were correlated with the topography of the irradiated regions by analyzing their changes using optical microscopy. The micrographs of a few regions showed changes on their topography after irradiating the regions with the maximum applied power (e.g., more depressed areas). These findings exhibited strong evidence of the thermal effects induced by the laser power on the materials found in this fragment, which must be considered to avoid alterations of the physical and chemical properties of the meteoritic samples. In addition, this work presents the first study done on the mineralogical composition of Northwest Africa 6603 using Raman spectroscopy. Furthermore, from an educational standpoint, this project exposed the involved undergraduate physics students to numerous research steps (e.g., experiment preparation, data acquisition/analysis, and manuscript preparation) which provided them with a broad spectrum of valuable scientific and technical tools for their future careers.


Introduction
Raman spectroscopy is a nondestructive technique that has been broadly used to investigate the chemical composition and structure of a material. [1][2][3][4] In particular, it has been used extensively on meteoritic samples since it provides information about their mineralogical makeup [5][6][7][8][9][10] which can give valuable clues about mechanisms of planet formation and secondary processes, such as shock event scenarios [11,12] and thermal metamorphisms experienced by their parent bodies (e.g., asteroids and comets), [13][14][15] respectively. Carbonaceous chondritic meteorites are of special interest to understand how planets formed as they are considered some of the most primitive surviving materials from the early Solar System. [16] However, the power of the excitation source used in this technique can change the Raman spectra parameters (peak position, [5,17,18] width, [17,18] intensity and/or area [18] ). This means that the properties of the studied samples (e.g., structure) has been altered due to thermal effects. [19][20][21][22][23][24] Hence, it is important to study how the laser power can affect this kind of samples to avoid any unintentional alterations to obtained reliable results, and it can help to investigate how the properties of these relics could have been affected due to thermal conditions during or after their formation. [25] In this work, the laser-induced thermal effects produced on the carbonaceous chondritic meteorite Northwest Africa 6603 (NWA 6603) are studied in detail by analyzing the low-and highresolution Raman spectra parameters of all the minerals found in the selected inclusions and matrix of the sample as a function of the excitation laser power. These thermal effects are correlated with the topography of the irradiated regions by examining their changes using optical microscopy. The results obtained using these techniques are discussed in the context of the structural alterations produced in the constituent materials of this meteoritic sample. To the best of the authors' knowledge, there are no studies done on NWA 6603 using this spectroscopic technique, and there is only one work done on one meteoritic sample (Allende) in which only carbonaceous matter has been analyzed as a function of the Raman spectroscopy laser power. [17] Therefore, this work not only provides information in detail about how the materials found were affected by the laser power but also presents the first study done on the mineralogical composition of NWA 6603 using Raman spectroscopy. In addition, from a pedagogical point of view, this study was conceived as an undergraduate research project to expose students, in particular physics majors, to all the stages of an experimental scientific work using spectroscopy and microscopy techniques.

Experimental methods
Sample A 1.9961-g fragment of a carbonaceous chondrite from Northwest Africa, NWA 6603, was studied using Raman spectroscopy and optical microscopy as a function of the Raman excitation laser power. This meteorite was found in Morocco in 2006. The fragment of NWA 6603 analyzed in this work is a flat rectangular chip and was purchased from The Meteorite Market. The sample was sawed and ground by the dealer using an isopropyl alcohol cooled-diamond blade and dry diamond lap, respectively, and then it was rinsed in 99% isopropyl alcohol. Figure 1 shows a photograph of the sample revealing numerous inclusions with their surrounding matrix. The red circles marked the four main inclusions (1, 2, 3 and 4) and surrounding matrix examined in this work.

Raman spectroscopy
Raman spectroscopy measurements for this study were carried out at room temperature using a custom-built fiber-coupled micro-Raman system with a backscattering configuration. The equipment had a Raman probe (Renishaw RP20V) with a 10X objective with a numerical aperture of 0.25, and a 300-mm focal length spectrograph (Andor SR-303i-A). Two different gratings were employed for low-and high-resolution measurements: 600 and 1200 grooves/mm gratings. The 532-nm line of a linearly polarized semiconductor laser (Coherent Sapphire SF 532 nm) was employed as the excitation source. A power range of 1-19.3 mW and a focal spot of diameter $3 mm were used on the studied spots, producing a power density range of $141-2730 W mm À2 . The laser power was adjusted using different combinations of neutral density filters. A thermocooled CCD camera (Andor iDus DU401A-BVF) was used to detect the Raman-scattered light from the same objective. Individual spectra were taken from the selected spots of the examined regions on the meteoritic fragments through the Raman system software (Solis) using an integration time of 1 second and 100 accumulations of spectra for each spot to obtain a high signal-tonoise ratio. Some cosmic rays (high energy particles from outer space) were detected in the Raman spectra and were removed from them using the Raman system software (Solis).
The studied regions were selected considering high signal-to-noise ratios of the spectra of their constituent minerals. For each spot, individual spectra were taken starting from the lowest laser power used, and then the power was increased up to the highest value used in increments of $1 mW. Finally, the power was brought back to $4.7 mW, which provided a high signal-to-noise-ratio in the spectra with a low power density, to examine possible permanent alterations produced in the irradiated regions.
The obtained Raman spectra were analyzed using a commercial software package (OriginPro): Gaussian functions were fitted to the spectrum peaks to determine their position (Raman shift or peak position, x), full width at half maximum (FWHM or Raman peak width, C), and integrated intensity or area under the peak (Raman peak area, A). For all the spectra that presented a fluorescence background at a particular laser power, a baseline for background subtraction was created using an adjacent-averaging smoothing algorithm and a spline interpolation method. The fitting parameters were used to identify the mineralogical composition of the selected regions. To recognize the unknown materials from the obtained spectra, the authors created a Raman library written in Python using the Raman spectra database provided by RRUFF. [26] In addition, x, C and the ratio between areas of selected peaks for each material (e.g., doublets) were analyzed as a function of the laser power.

Optical microscopy
Optical micrographs of the studied regions were taken before and after irradiating them with the different laser powers to inspect the surfaces of the inclusions (micro/millimeter-sized structures) and surrounding matrix, and to evaluate if any topological changes were produced. The micrographs were taken using a Nikon Eclipse ME600 microscope at magnifications of 50Â, 100Â, 200Â, 500Â and 1000Â. These optical images were correlated with the corresponding Raman spectra.

Results
Four inclusions and surrounding matrix ( Fig. 1) of an NWA 6603 fragment were studied as a function of the Raman spectroscopy laser power due to the higher signal-to-noise ratios of the spectra of their constituent materials. Two of these inclusions (inclusions 1 and 3) were identified as chondrules due to their well-defined circular edges. [27] Inclusion 2 presented the characteristics of a calcium-aluminum inclusion (CAI) due to its irregular shape and white color, as well as the presence of minerals commonly found in CAIs, [27] such as gehlenite (melilite). [28] The fourth inclusion was identified as an ameboid olivine aggregate (AOA) due to its irregular shape and abundance of olivine. [27] Numerous spots of the studied regions were measured and analyzed. Olivine (forsterite), graphitic carbon, pyroxene (enstatite), hematite and gehlenite were the materials found in these spots. The values of their Raman peak position given below to describe their characteristic spectra correspond to the results obtained at a laser power of $4.7 mW. Figure 2 shows representative high-resolution Raman spectra obtained from one of the spots ("spot 1") of inclusion 4 at some selected laser powers. In this case, the spectra correspond to olivine which is characterized by a doublet at the region of 810-860 cm À1 as the result of the coupling between the symmetric and antisymmetric stretching modes of nonbridging Si-O nb bonds of the constituent SiO 4 tetrahedra. In general, the olivine spectra measured from the studied meteoric fragment had a doublet at $824/855 cm À1 indicating a forsterite content (Mg 2 SiO 4 ) of $100%. [23] From Fig. 2 is not clear to observe changes of the spectra with the laser power. However, after analyzing the Raman peak parameters, it is possible to distinguish that the positions of the doublet peaks and their area ratio decrease while their widths increase as the power increases (Fig. 3). Figure 4 shows representative high-resolution Raman spectra obtained from one of the spots ("spot 1") in the matrix around inclusion 3 at some selected laser powers. These spectra indicate the main presence of graphitic carbon and pyroxene. This figure shows the main characteristic first-order Raman bands of graphitic carbon: D-band at $1347 cm À1 ("D" for "disordered" graphite) and G-band at $1600 cm À1 ("G" for graphitic carbon, corresponding to one of the C-C stretching modes). [29][30][31] From this figure it is possible to distinguish a downshift in the position of the bands. These changes and other ones are shown in Fig. 5 which includes the graphs of the Raman parameters analyzed from the matrix ("spot 2") around inclusion 3: the position of the bands and the width of G-band decreases, while the width of D-band and the area ratio increase as the power increases.
In the case of pyroxene, its characteristic Raman peaks are at 344, 371, 667/688 and 1013/1034 cm À1 , which correspond to enstatite (MgSiO 3 ). The last two groups of peaks are due to symmetrical stretching of the nonbridging Si-O nb and bridging Si-O b -Si bonds, respectively, while the two first peaks correspond to nontetrahedral stretching between cations and oxygen bonds (M1-O and M2-O). [32,33] In Fig. 4 Figure S4 (in Supplementary Material) shows representative high-resolution Raman spectra obtained from one of the measured spots ("spot 2") of inclusion 1 at some selected laser powers. These spectra present the characteristic peaks of hematite (a-Fe 2 O 3 ) at $223, 290 and 405 cm À1 due to transverse optical modes, at $1309 cm À1 because of an overtone associate with a Raman-forbidden longitudinal optical mode near 660 cm À1 , [5] and at $820 cm À1 which is of magnetic origin known as a magnon peak. [24] In Fig.  S4 it is possibly to observe several changes produced in the Raman peaks as the laser power increased, such as their positions shifted to lower values, and the 405 and 1309-cm À1 peak intensity decreased. Figures S5 and S6 ( Material) show in detail the trends of the position, width and area ratio of the main peaks as a function of the laser power. For the four peaks analyzed ($223, 290, 405 and 1309 cm À1 ) as the power increases, the Raman peak positions decrease, the widths increase, except for the 405cm À1 -peak width which decreases, and the ratio between the areas of the two main peaks (223 and 290-cm À1 ) increases. Figure S7 (in Supplementary Material) shows representative high-resolution Raman spectra obtained from spot 1 of inclusion 2 at some selected powers. These spectra have several peaks that could correspond to more than one mineral. The main peaks, located at $305, 625/669, 799/ 846, 911 and 977 cm À1 , match well with those ones corresponding to crystalline gehlenite (Ca 2 Al 2 SiO 7 ). [26,34] The strongest band at $625 cm À1 in the spectrum of this mineral is a symmetrical stretching mode of bridging oxygen of pyrosilicate anions. The shoulders at $669 and 846 cm À1 and the band at $799 cm À1 could be assigned to vibrational modes involving AlO 4 tetrahedra, while the bands at $911 and 977 cm À1 are attributed to symmetrical stretching modes of nonbridging oxygen of the pyrosilicate group. [34] This mineral assignment agrees with the results obtained from CAIs found in NWA 6603 using other experimental techniques (e.g., mXRD and 27 Al 3Q MAS NMR spectroscopy). [28] In Fig. S7 the most observable change is the downshift of the main peak and shoulder ($625/669 cm À1 ). Figure S8 (in Supplementary Material) shows the trends of the Raman parameters of this peak group as a function of the laser power: the positions of the peak and shoulder, width of the 625cm À1 peak and area ratio decrease, while the width of the shoulder increases as the power increases.

in Supplementary
After reaching the maximum set power and coming back to $4.7 mW, most of the analyzed spots presented some permanent alteration in their Raman parameters (represented with a star in Figs. 3, 5, S1-S3, S5, S6 and S8) since these values changed with respect to the first measurements done at that power. These spots were inspected using optical microscopy and, in general, no evident changes in the topography of these regions were found before and after the spots were irradiated. However, it was possible to observe a clear topographical alteration at spot 1 of inclusion 1: a $20-mm structure was removed from one of the irradiated regions. This area is indicated in Fig. 6 and corresponds to the spectra shown in Fig. 7.

Discussion
All the materials found in the studied regions (forsterite olivine, graphitic carbon, enstatite pyroxene, hematite and gehlenite) experienced, to a greater or lesser extent, changes in their Raman parameters with the laser power, in particular for x and C. Table 1 shows the largest Raman peak position and FWHM changes (Dx and DC, respectively) estimated for each material for their most affected peaks, which indicates that hematite was the most sensitive mineral, while olivine (forsterite) was the least affected one. This could be explained considering that hematite is an iron-oxide material since this kind of materials are good absorbers of light in the typical wavelength range used for Raman excitation lasers, and thus they can be more easily affected by increased temperature. [35] In the case of forsterite, on the other hand, its Raman parameters would be more difficult to modify with temperature due to its strong Si-O bonds. [36] In general, the Raman peak positions of all these materials decreased with the laser power. This major result agrees with the findings obtained in other studies done on the Raman parameters as a function of temperature for hematite [5,37] olivine, [36,38] pyroxene, [22,39] graphitic carbon [17,19,21,40] and crystalline gehlenite. [41] These changes can be interpreted in the context of the structural alteration of the materials. For instance, the materials could have experienced some modifications in their molecular structure transforming them into other materials of the same family, such as from high-rich to lower-rich forsterite in the case of olivine ($100% forsterite to $90% forsterite and 10% fayalite using the 100 Mg/(Mg þ Fe) molar ratio), [23] or into other allotropes, such as from graphitic to more disorder structures for carbon materials. [2] In addition, in the case of graphitic carbon and other carbon allotropes (diamond, graphite, carbon nanotubes, graphene, etc.), when the laser power increased, and so the temperature on the irradiated regions, the materials could have experienced an expansion of their C-C bond length resulting in a decrease of their Raman peak position. [19][20][21] Something similar could have happened for pyroxene, olivine and hematite, since their Raman peak positions decrease with the bond distances: for instance, Si-O tetrahedral distances for pyroxene and olivine, [22,23] and bond lengths of the Fe-O molecules for hematite. [24] In the case of crystalline gehlenite, the shift to lower wavenumbers could have indicated that the structure of the material was becoming more polymerized. [34] Regarding the Raman widths of the peaks analyzed in this work, a clear increase of the width with the laser power was observed for most of the peaks of the different materials, which agrees with other studies found in the literature for graphitic carbon, [17] olivine, [36,38] and pyroxene. [39] This trend can be interpreted as an increase in structural disorder resulting in an amorphization behavior of the studied materials when the laser power increases. [3,17,34]    Largest changes of Raman peak positions (Dx) and widths (DC) estimated for the materials found in the studied regions of carbonaceous meteoritic fragment Northwest Africa 6603 for the most affected peaks (ordered from most to least affected materials): hematite, graphitic carbon, gehlenite, pyroxene (enstatite) and olivine (forsterite).
x: Raman shift or peak position; Dx: Raman peak position change; DC: Raman peak full width at half maximum change.
The area ratios of the selected peaks showed, in general, an increasing or decreasing trend with the laser power. This means that one of the peaks were becoming larger in area with respect to the other one, indicating that the coupling between modes was getting weaker at higher temperatures. For instance, in the case of olivine, the area ratio of the 824-cm À1 peak with respect to the 855cm À1 peak decreased with the laser power, as shown in Fig. 3(e), since the area of the first peak decreased more than the area of the second one with increasing temperature. This indicates that the first peak was more sensitive to temperature than the second one, which implies that these two peaks became more independent of each other and thus the coupling between them got weaker. [36] Most of the values of the Raman parameters were not reproduced after using the highest power applied and remeasuring with a power of 4.7 mW (represented with a star in Figs. 3, 5, S1-S3, S5, S6 and S8), indicating that permanent alterations were produced in the studied regions, which agreed with the changes observed in the topography in a few regions using optical microscopy, which became more depressed.

Conclusion
In this work, the thermal effects produced by the laser power on the carbonaceous chondritic meteorite Northwest Africa 6603 were studied in detail using low-and high-resolution Raman spectroscopy. Olivine (forsterite), graphitic carbon, pyroxene (enstatite), hematite and gehlenite were the materials found in the inclusions and/or surrounding matrix of the studied regions. All the analyzed Raman parameters (peak position, width and area ratio) of these minerals were affected by the laser power. In general, the Raman peak positions decreased with the laser power: 23.6 cm À1 for the 1309-cm À1 hematite peak, 6.8 cm À1 for G band of graphitic carbon, 7.1 cm À1 for the 669cm À1 gehlenite peak, 3.8 cm À1 for the 667-cm À1 pyroxene peak, and 0.6 cm À1 for the 824-cm À1 olivine peak. These results indicated an expansion of the bond distances and rearrangements of their structures (e.g., from $100% forsterite to $90% forsterite and 10% fayalite olivine, and from graphitic to more disorder structures for carbon materials). The peak widths increased with the laser power for most of the Raman peaks which can also be interpreted as an increase in structural disorder. The peak area ratios showed an increasing or decreasing trend, suggesting that the coupling between Raman modes (e.g., doublets) was becoming weaker as the temperature increased. These changes in the Raman parameters seemed to be permanent (reverse changes were not observed in the Raman parameters when the laser power was brought back to lower values) which agreed with the observations on the topography of some studied regions using optical microscopy since part of the structure of the irradiated surface became more depressed. Therefore, even though Raman spectroscopy is a powerful technique to study the mineralogical composition of meteoritic samples, in particular carbonaceous chondrites, the thermal effects produced in the materials by the laser power must be considered to avoid undesirable alterations of their properties. In addition, these findings can help provide a better understanding about conditions, scenarios and specific processes (e.g., thermal metamorphism) that might have taken place in the formation and later alteration of chondritic meteorites which contain important clues about how planets formed and later evolved in our Solar System. Furthermore, this work provided the first study done on the mineralogical composition of NWA 6603 using Raman spectroscopy and analyzing the Raman parameters of obtained materials as a function of the laser power. This adds to the research that has been conducted before for the composition of this meteorite using other techniques [28] and for the study of the Raman parameters as a function of the laser power done only on carbonaceous matter of a different carbonaceous chondritic meteorite (Allende). [17] Finally, this project offered the undergraduate students who are coauthors in this work a wide range of experimental and analytical methods for spectroscopy and microscopy experiments done on carbonaceous chondrites, such as, sample and experiment preparation, data acquisition and analysis with commercial software and Python codes, technical and scientific literature comprehension, conference presentations, [42] and manuscript preparation. Thus, these tools helped them develop new valuable skills for their future studies and careers.