Spectrochimica Acta Part A 80 (2011) 138–147
Contents lists available at ScienceDirect
Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy journal homepage: www.elsevier.com/locate/saa
Raman spectroscopy as a tool to characterize heterogenite (CoO·OH) (Katanga Province, Democratic Republic of Congo) C. Burlet ∗ , Y. Vanbrabant, H. Goethals, T. Thys, L. Dupin Royal Belgian Institute for Natural Sciences, Geological Survey of Belgium, Jenner Street, 13, BE-1000 Brussels, Belgium
a r t i c l e
i n f o
Article history: Received 13 September 2010 Received in revised form 8 February 2011 Accepted 2 March 2011 Keywords: EBSD Heterogenite Laser-induced transformation Lufilian fold-and-thrust belt Raman microspectroscopy
a b s t r a c t Natural heterogenite (CoO·OH) samples were studied by Raman microspectroscopy, electronic microprobe and Electronic BackScattered Diffraction (EBSD). Raw samples and polished sections were made from 10 mines covering the Katanga copperbelt (Katanga Province, Democratic Republic of Congo). Four typical Raman responses have been obtained leading to investigate the laser-induced dehydroxylation of heterogenite into a Co-spinel structure. The results are also compared with EBSD patterns from oven heated heterogenite samples. A close relationship was established between the chemical substitutions of Co by mainly Cu, Ni, Mn and Al and their impact on the mineral Raman response. © 2011 Elsevier B.V. All rights reserved.
1. Introduction Heterogenite is a natural oxyhydroxide variety of cobalt with the following ideal chemical composition: CoO·OH. Copper, nickel and manganese represent the main substitutes of cobalt, which commonly disrupt the crystallographic lattice leading to amorphous phases [1]. This mineral is encountered in the upper part of the oxidized horizon of Cu–Co bearing formations of the Neoproterozoic Lufilian fold-and-thrust belt in the Katanga Province (Democratic Republic of Congo – DRC) [1]. Heterogenite represents an important commodity operated into formal and informal mining sites where it is extracted in association with Cu-minerals. As heterogenite is considered as a potential candidate for an initiative of traceability and certification process [2–4], similar to what has been done with Coltan (Melcher et al. [5]), or diamond by means of the Kimberley process. Any initiative taken toward the potential traceability of heterogenite should be based on solid and extensive mineralogical and geochemical characterization of the mineral throughout the copperbelt region. Synthetic forms of cobalt oxyhydroxide, hydroxide and oxides are also the subject of numerous material science researches by Raman and IR microspectroscopy, X-ray methods and magnetization analyses. Subjects covered by these studies include the synthesis of films, nanocrystals and coatings for battery technology [6,7], catalysts [8], nanomaterials [9,10] and supraconductors [11].
∗ Corresponding author. Tel.: +32 02788 76 53. E-mail address:
[email protected] (C. Burlet). 1386-1425/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.saa.2011.03.007
In this paper, the Raman microspectroscopy signature of natural heterogenite is investigated according to cobalt cations chemical substitutions. The thermal and laser-induced transformation of heterogenite into the Co-spinel structure (Co3 O4 ) is also linked to Electron BackScattered Diffraction (EBSD) patterns. The geological setting of Lufilian fold-and-thrust belt and its copper–cobalt metallogeny is described with by the main mineralogical characteristics of heterogenite. The Raman microspectroscopy and the EBSD methodology are then explained in order to determine the Raman signature of heterogenite and the chemical and thermal conditions (e.g. laser-induced) that trigger the transformation of this mineral into a Co-spinel structure. 2. Geological setting The Lufilian fold-and-thrust belt (Fig. 1) results from the accretion of mainly Neoproterozoic sedimentary rocks during the pan-African orogeny with a main phase of thrusting and associated metamorphism, which occurs at 566–550 Ma [12]. These sedimentary units lay on the older Kibarian basement (Mesoproterozoic). The lowermost stratigraphical unit deposited on the Kibarian basement is called Roan (880–740 Ma) and is differentiated into a carbonate platform (the Kambove Dolomite Formation), a lagoon basin (Dolomitic Shale Formation) with mudflats (R.A.T. Subgroup) and a siliciclastic margin. The lagoon-basin was subsequently filled by clastics materials (Dipeta Subgroup). The mineralised horizons (mainly Cu-Co sulphides) of the Roan, called ‘Series des Mines’ has a southwest-northeast extent of about 400 km [12–15]. The Lufilian fold-and-thrust belt is also subdivided on the metallogenic point of view into the Zambian copperbelt where the
C. Burlet et al. / Spectrochimica Acta Part A 80 (2011) 138–147
139
Fig. 1. Simplified geological map of the Lufilian fold-and-thrust belt. The studied samples were collected into various mining sites through the arc. Figure inlet shows the extends of the Pan-african belt [15].
deposits are mainly composed of sulphide primary minerals (e.g. carrollite, chalcopyrite, bornite, chalcocite) and the Katanga copperbelt. The Katanga copperbelt shows higher Cobalt content than the Zambian region and thicker oxidized horizons. The superficial weathering of the primary sulphides resulted in the formation of an extensive set of newly formed minerals including Cu and Co hydroxides, oxides, silicates and carbonates [16]. This secondary enrichment also resulted in an upgrade of the Cu and Co content. Copper Belt includes a thick (tens of meters up to 100 m) oxidized horizon above the primary sulphide ore bodies [12,14]. In addition, it is commonly considered that the Katanga Province hold between one third and half world recognized cobalt reserves and heterogenite and asbolane represents the main mineral of the oxidized horizon [16]. Metallogenic debates mainly focus on the mechanisms and the potential sources of copper and cobalt of the primary sulphide ore bodies [15,16]. Surprisingly few researches were conducted on the main mineral sources which is the oxidized part. Deliens et al. [1,17] studied the mineralogy of heterogenite and the main results of their works are presented in the following section.
and heterogenite-3R. Their crystallographic parameters are summarized in Table 1. Heterogenite is mainly found in Katanga, but few occurrences were also reported in Germany, Morroco, Chili, USA, Japan and Australia [24]. The mineral can hold various amounts of Cu, Ni, Mn, Fe and Al. The presence of these components tends to disrupt the crystallographic lattice leading to difficulties to studies the mineral with classical X-ray techniques [1]. 4. Materials and methods Heterogenites samples of various compositions and from 10 mining sites throughout the Katanga Province were investigated by the combination of Raman microspectroscopy with geochemical analyses (electron microprobe and EDS). In addition, the Electron BackScattered Diffraction (EBSD) technique was applied on a well-crystallized sample in order to study the transformation of heterogenite into a Co-spinel structure during a heating test at 420 ◦ C. 4.1. Raman microspectroscopy
3. Mineralogy and crystallography of heterogenite Heterogenite was first described by Frenzel [18] in 1872, but it was rapidly regarded to be part of a larger group of oxides and hydroxides of cobalt and different names, including Stainerite, Mindigite, Transvaalite, Trieuite, Boodtite, etc. and different chemical formulae were suggested by the various authors (for a review see Deliens [1]). In 1962, Hey [19] came to the conclusion that most of these minerals share the same formula, CoO·OH, analogous to that of manganite (MnO·OH) and goethite (FeO·OH). This was later supported by the work of Deliens [1,17], who studied the southern Katanga heterogenites and clarified this nomenclature by suggesting the use of only the name heterogenite as the general name for the oxyhydroxide of cobalt. Later, Deliens and Goethals [17] identified two polymorphs for this mineral, namely heterogenite-2H
The Raman microspectroscopy is part of the large category of vibrational spectroscopic methods based on the emission of phonon during relaxation of energy of molecular bonds. This technique can be applied on various materials including organic and inorganic compounds, fluid and gases. Since the technique investigated the structural organisation of molecules of the studied sample, it does not necessary required the material to be crystallized which is suitable in the case of X-ray amorphous heterogenite Table 1 Crystallographic parameters of heterogenite-3R and heterogenite-2H [1,17,20–23].
System Space group ˚ Dimensions (A)
Heterogenite 3R
Heterogenite 2H
Co-spinel
Trigonal R-3m a: 2.855; c: 13.157
Hexagonal P63 /mmc a: 2.855; c: 8.805
Cubic Fd3m a: 8.1279
140
C. Burlet et al. / Spectrochimica Acta Part A 80 (2011) 138–147
Laser wavelength Laser power Acquisition time Spectrometer slit Spectrometer range Spectrum resolution Objective
532 nm 2 mW 3 × 10 s (1 s pause between exposures) 1000 × 50 m 0–4500 cm−1 9 cm−1 50× long focusing distance
samples. The other advantage of Raman microspectroscopy during the investigation of mineralized materials is that it provides fast results (commonly few seconds to few tens minutes) and can be applied directly on fragments (raw samples). A common set of spectrometer, laser and microscope parameters was defined for all our measurements on raw samples and polished sections. This procedure allowed us to get hundreds of comparable and reproducible measurements on our samples. Otherwise stated, all spectra presented in this paper have been acquired using Senterra BX31 Raman microscope manufactured by Bruker Corporation with the parameters described in Table 2. Calculations were carried out using Bruker Opus software. A series of iterative least-squares curve fittings (Levenberg–Marquant iterative procedure) were made on selected Raman spectra in order to characterize heterogenite response in terms of peaks position, surface and relative intensity. Depending on the noise level of the collected spectra, residual RMS error obtained after curve-fitting is typically comprised between 0.5 and 2. 4.2. Microprobe analysis To identify a relationship between the chemical composition of heterogenite samples and their corresponding response in Raman microspectroscopy, several hundreds of quantitative chemical analysis were carried out on an electronic microprobe (University of Louvain). These analyses were performed on polished sections from 10 mines covering the whole Arc and each measurement was spatially referenced on the sections. In most cases, microprobe analyses were made before the acquisition of Raman spectra. This procedure permitted to obtain an early overview of the chemical variability within the collected heterogenite samples, which helped to guide our Raman analysis. Raman measurements have been typically made 2–3 m away from the previous microprobe analysis in order to avoid any mineral surface alteration due to the microprobe electron beam.
Intensity (arbitrary units)
Table 2 Raman microscope default settings applied for raw sample and polished sections analysis.
a
b c d 1900
1700
1300
1500
1100
900
700
The EBSD technology is a powerful method in Earth and Material Sciences for characterizing the crystallographic composition and the microtextural content of samples. It allows amongst other things the identification of different crystallographic phases, the determination of the lattice orientation of individual crystal and the mapping and morphology study of grain/particles (e.g. size, shape and orientation). EBSD is based on the diffraction of electrons (elastic scattering) on the atomic planes of a tilted sample (usually 70◦ ) located in a SEM chamber. The backscattered electrons are collected on a phosphor screen where they form a pattern composed by the so-called Kikuchi bands. The latter correspond the intersection of the cones of diffracted electrons, also known as Kossel cones, from a lattice plane with the phosphor screen. The diffraction occurs only in the directions satisfying the Bragg’s law. The identification of phases by EBSD method is based on the correct indexation of the Kikuchi bands with respect to the potential reflectors of the possible phases. This correspondence is now fully computer-aided by a voting method described by Wright and Adams [25]. In summary, this procedure is based on the matching between the interplanar lattice angles determined by Kikuchi band triplets and the angles between potential reflectors. These last angles are reported into a look-up table. Different indexing solutions are frequently found and each of them is compared with the other possible combinations of detected bands. A vote is attributed to a
Intensity (arbitrary units)
1202 1133
1500
1300
1100
100
4.3. Electronic BackScatter Diffraction
670 626 572
1700
300
Fig. 2. Different types of Raman response of natural heterogenite samples.
495
1900
500
Raman shift (cm -1)
900
700
500
Raman shift (cm-1) Fig. 3. Curve fitting on a Raman spectra of heterogenite (Mindigi mine).
300
100
C. Burlet et al. / Spectrochimica Acta Part A 80 (2011) 138–147
141
Intensity (arbitrary units)
681 A
601 499 F F
1900
1700
1500
1300
1100
900
700
469 E
500
181 F
300
100
Raman shift (cm-1) Fig. 4. Curve fitting on a Raman spectra of laser-induced Co-spinel (Kabolela mine).
solution for each combination with a positive matching and their respective score can in turn be sorted. The confidence index (CI) is a parameter to determine the uniqueness of a solution and is defined as follows [26]: CI =
V1 − V2 Videal
where V1 , V2 represents the number of votes for the first and the second solution, respectively. Videal is computed as next: Videal =
n! (n − 3)! 3!
with n equals to the number of Kikuchi bands applied for the identification. If the first and second solutions get the same number of votes (CI = 0), both orientations are possible. By contrast, if CI equals to 1 the first solution is unique. The EBSD equipment manufactured by EDAX-AMETEK company uses a Hikari camera mounted on the SEM of the Royal Belgian Institute for Natural Sciences. The results were acquired using OIM Data Collection software and process with OIM Data Analysis.
This heterogenite spectrum mainly shows a major peak centred at 495 cm−1 and a broader peak centred at 623 cm−1 . The latter probably results from the overlapping of three minor peaks at 572 cm−1 , 626 cm−1 and 670 cm−1 . In some spectra, a very minor peak is present in the form of a “shoulder” at around 680 cm−1 (not visible in Fig. 3). These positions are in agreement with those reported by Yang et al. [27] and Pauporté et al. [28] on synthesised CoO(OH) films with a slight different distribution in the broader peak. Two minor peaks are also observed on some samples at 1133 cm−1 and 1202 cm−1 and could not been yet assigned. 661 (Sp)
494 (Het ->Sp) 182 (Sp)
466 (Sp)
8
5. Results 5.1. Heterogenite Raman responses
7
177 (Sp)
About 1100 individual Raman spectra were acquired on the selected heterogenite raw samples and polished sections. Fig. 2 compares the 4 main types of Raman responses acquired on heterogenite samples.
6 5
5.1.1. Type 1: heterogenite s.s. The first response (Fig. 3) is regarded as the real Raman response of heterogenite under the laser beam without major structural transformation due to the heat produced by the laser.
4
658 (Sp)
% mass loss
24
3 2
a
20
b
614 (Het) 491 (Het ->Sp)
16
1
12 8 4 100
200
300
400
500
600
700
800
900
1000
Temperature (°C) Fig. 5. Thermal analysis of pure and impure heterogenite (Reproduction with the authorization of Deliens [1]).
Raman shift(cm−1) Fig. 6. Progressive transformation of heterogenite into a Co-spinel, following 8 successive Raman shoots.
142
C. Burlet et al. / Spectrochimica Acta Part A 80 (2011) 138–147
Fig. 7. Heterogenite unheated sample: (a) SEM-picture, (b) example of EBSD pattern with interpreted reflectors, (c) orientation of corresponding crystal; heated sample (420 ◦ C): (d) SEM-picture, (e) example of EBSD pattern with interpreted reflectors, and (f) orientation of corresponding crystal.
5.1.2. Type 2: dehydroxylated heterogenite (Co3 O4 ) The second most common response (Fig. 4) corresponds to the transformation heterogenite locally adopting a cubic normal spinel crystallographic structure by dehydroxylation of the heterogenite to form a cobalt oxide. Five Raman peaks were determined at 658 cm−1 (major peak), 601 cm−1 (shoulder), and at 499, 469 and 181 cm−1 . These peaks have similar but slightly downshifted positions to those calculated and observed by Shirai et al. [21] and Hadjiev et al. [20] for normal Co3 O4 spinel (691.0, 618.4, 521.6, 482.4, 194.4 cm−1 ). This can be attributed to the very small size of the formed spinel crystals (nanometric scale), like it has been reported by Ai and Jiang [10] on Co3 O4 nanopowders.
(in particular the mass variation with temperature) of several heterogenite samples They observed two mass loss at around 250–300 ◦ C and 900 ◦ C. X-ray diagrams made from these samples before and after heating showed strong differences, as the 250 ◦ C heated sample X-ray diagram presented most of the characteristics of a crystallographic spinel comparable to magnetite (Mn3 O4 ). They concluded that sample heating induced a progressive dehydroxylation of the heterogenite, resulting into the transformation of its original crystallographic structure to a normal cobalt spinel (Co-spinel or Co3 O4 ) structure. The transformation correspond to the following formula [1]:
5.1.3. Type 3: “Ni-bearing” heterogenite A particular Raman response was obtained on an heterogenite sample collected in the Shinkolobwe mine (Fig. 2c). Only a very broad peak centred peak around 500 cm−1 is visible. This heterogenite sample has the particularity to be relatively richer in Ni (above 8% Ni equivalent oxide, microprobe analysis) than samples from other mines. 5.1.4. Type 4: “Burned” heterogenite spectra Some spectra acquired on our heterogenite samples (Fig. 2d) did not give any visible Raman peak but a noisy highly fluorescent spectrum attributed to the hole-burning of the material by the Raman microspectrometer laser.
In 1974, Deliens [1] refined this experiment with heterogenite samples from various Katangan mines and confirmed the Co-spinel transformation near 300 ◦ C and its reversible reduction into Cobalt monoxide at about 900 ◦ C (Fig. 5). More recently, similar experiences were also carried out by Yang et al. [27] on a synthesised CoO·OH film, with assimilable results. The variations between the termo-ponderal curves were attributed to crystallinity differences linked to the presence of other cations in the heterogenite lattice (Cu, Fe, Al, Mn, Ni, etc.) [1,29]. Well-crystallized heterogenite samples (Fig. 5a) show an abrupt dehydroxylation, while poorly crystallized heterogenite exhibits a more progressive transformation, starting at ∼100 ◦ C till slightly above 400 ◦ C [1].
6. Dehydroxylation of heterogenite into Co-spinel
6.1. Laser-induced heterogenite dehydroxylation
In 1958, Orcel et al. [29] published an paper on the heterogenite’s properties They studied the thermodynamic behaviour
The Co3 O4 spinel spectra acquired on some of our heterogenite natural samples suggest that the Raman laser is therefore able to
6CoO · OH → 2Co3 O4 + 3H2 O +
1 O2 2
C. Burlet et al. / Spectrochimica Acta Part A 80 (2011) 138–147
143
Fig. 8. EBSD mapping of (a) heterogenite-2H, unheated (inlet: colour-coded inverse pole figure for a hexagonal system); (b) Co-spinel, heated sample (inlet: colour-coded inverse pole figure for a cubic system).
produce the necessary heat to trigger the structural transformation of the mineral into Co-spinel, as Pauporté et al. [28] have reported on synthetic powders of CoO·OH. For that matter, Raman analyses performed on polished sections often reveal optically visible damages induced by the laser irradiation. The laser-induced transformation of heterogenite into Co-spinel can be further studied on some specific samples during successive Raman tests. Fig. 6 presents the Raman spectra resulting of 8 successive laser shots on the same spot using the default test parameters (see Table 2). The first spectrum exhibits a characteristic heterogenite signature. The 491 cm−1 peak has however already migrated with respect to its equivalent in the reference spectrum (see Fig. 3). A peak attributed to the development of the Co-spinel structure can be observed at 658 cm−1 during the second shot. Its amplitude is equivalent to those of the heterogenite structure and it becomes the dominant peak during the following shots. By contrast the 614 cm−1 peak of heterogenite is overwhelmed by the 661 cm−1 peak and it represents a gentle slope from shot 5. The intensity of the 494 cm−1 peak also decreases during the successive shots and an additional shoulder is observed at 466 cm−1 . Finally, a small peak (177 cm−1 ) appears clearly during the shot 6. It is more likely present since shot 4 but the signal at this stage is almost equivalent to the noise. 6.2. Comparison with EBSD patterns EBSD represents a complementary approach to Raman spectroscopy for studying the transformation of heterogenite into a
Co-spinel structure due to a laser-induced thermal treatment or any heat sources (e.g. oven, furnace). In fact, this transformation implies a significant modification of the lattice structure from a hexagonal system for the heterogenite-2H polymorph or tetragonal system for heterogenite-3R into the cubic organisation of a spinel. However, there is a major limitation on the application of EBSD since preliminary tests show that heterogenite samples are commonly amorphous to electronic-diffraction method. Consequently, only a very restricted number of samples including heterogenite crystals can be applied for this study. They correspond to the purest phases with up to 1% of Cu. The preparation for good-quality EBSD patterns of crystallized heterogenite includes an initial grinding of the surface to be analysed in order to eliminate the major asperities, then a gradual polishing during of a few minutes is carried out using soft cloths with diamond powder and alumina of decreasing grain sizes (from 9 m to 0.3 m). The final surface (Fig. 7a) is obtained by a longduration polishing (1 h) using colloidal silica (0.04 m). In this study, three crystallographic phases (heterogenite-2H & -3R, Co-spinel) are compared by EBSD method during a transformation test of heterogenite. A look-up table for each phase was also computed thanks to OIM Data Collection software (EDAX NV) based on their lattice parameters and space group (see Table 1). The transformation test includes three steps: firstly, the polished surface of an aggregate of fibrous crystals of heterogenite (few tens m thin by hundreds m long) from the Kalabi mining site (see Fig. 1 for location) was studied by Raman microspectroscopy and EBSD. The experimental parameters of Raman spectra corre-
C. Burlet et al. / Spectrochimica Acta Part A 80 (2011) 138–147
spond to those defined earlier in this paper. The sample was then heated into a conventional laboratory oven under normal atmospheric conditions during 1 H at 420 ◦ C. Finally, the heated sample was again investigated by Raman and EBSD methods with the same parameters and conditions. The results of this test show that:
1400 1200 1000
191
614 496
Intensity (a.u.)
Before heating
634
1215 1142
EBSD mappings of the sample were conducted before and after the heating test with the three above-mentioned crystallographic phases (heterogenite-2H & -3R and Co-spinel). Points showing a confidence index (CI) lower than 0.1 and grains of less than 2 pixels were withdrawn from the selection since they are regarded as misindexed. Before heating, heterogenite-2H represents 84.8% of the total fraction of indexed patterns, while heterogenite-3R and Co-spinel constitute only 0.1 and 0.7%, respectively. The remaining area (e.g. 14.4% of the surface) corresponds to points where no solution was found or withdrawn by the CI and grain size filtering. These points match with small cavities and areas of poor polishing and they are represented in black in Fig. 8a. The detection of few Co-spinel and heterogenite-3R patterns is also associated with small irregularities of the polished section. Fig. 8a shows the Inverse Pole Figure (IPF) of the normal direction of the sample surface with respect to the heterogenite-2H axes. This representation can also be regarded as an indication of which crystal axis is pointing toward the viewer. For example, if 2 −1 −1 0-axis is normal to the section surface, the viewer will see a lateral edge of a hexagonal crystal. The areas corresponding with this orientation are represented by green colours in Fig. 8a. The 1 0 −1 0 axis corresponds to a lateral face of the crystal and 0 0 0 1-axis its c-direction. In this case, heterogenite grew as a series of narrow bands C-axis subparallel to the section surface with showing a crystal edge (green) or face (blue). Co-spinel (Fig. 8b) is clearly detected during the EBSD mapping after the thermal treatment, but it represents only 38.8% of the total fraction. Patterns of heterogenite-2H and -3R are also found in 2.3 and 0.2% of cases, respectively. The rest (58.7%) represents misindexed points (CI < 0.1 and grain size < 2 pixels) or without solution. The main problem during this indexing results from the major modification of the polished section surface due to the thermal stress and the consecutive mineralogical transformation. The presence of numerous cracks is moreover highlighted during this EBSD-mapping. Interestingly, the development of Co-spinel crystals during the heating test clearly follows the boundaries of the heterogenite-2H crystals. Fig. 9 shows a comparison of typical Raman spectra of heterogenite before and after the thermal treatment in the oven and its consecutive transformation into a Co-spinel structure. The positions of the Raman peaks of the spinel form clearly show a significant upward shift (10–28 cm−1 ) with respect to those recorded during a laser-induced transformation (see Table 3 for a comparison). The recorded peaks on the oven-heated sample are also closer to those calculated and observed by Shirai et al. [21] on synthetic
520 479
• before heating, EBSD patterns corresponding to heterogenite-2H are recorded through the polished section. Fig. 7b presents an EBSD-pattern of heterogenite-2H with interpreted reflectors. In this case the C-axis (Fig. 7c) dips gently with respect to the sample surface and a crystal face points towards the viewer; • during the heating period, the sample suffered significant modification as indicated by the development of micro-fractures observed on the sample surface (Fig. 7d). Nevertheless the grains still exhibit surface allowing EBSD-patterns to be directly recorded without requiring a new polishing; • after heating, EBSD-patterns corresponding to those of the Cospinel are recorded in numerous points of the section. An EBSD pattern of Co-spinel with the indexed reflectors is shown in Fig. 7e. One edge of cubic crystal is mainly pointing towards the viewer (Fig. 7f).
686
After heating (420 °C)
591
144
800
600
400
200
Raman shift (cm- 1) Fig. 9. Comparison of Raman spectra on a sample before heating (heterogenite-2H) and after heating (Co-spinel) in an oven at 420 ◦ C.
spinels. Difference between the laser-induced and oven-induced spinel is interpreted as a result of a difference in the crystallinity. In fact, the volume of transformed material by the laser influence is small since its related to the laser spot (∼2 m), while the size of crystals of the sample from the oven is significant as shown in Fig. 8b. In addition, the structure of Co-spinel lattice is probably not very well developed during a laser-induced transformation due to its proximity with heterogenite material unaffected by the laser. 7. Influence of heterogenite geochemistry on its Raman signature Analyses performed on polished sections made from our heterogenite samples permitted to couple Raman response of the material with petrographical observations and chemical composition. Table 3 Raman shift comparison of laser induced Co3 O4 spectra (Kabolela mine) with an oven-heated heterogenite sample (Kalabi mine). All units are in cm−1 . Vibratory modes
Raman shift of 2 mW laser induced Co3 O4
Raman shift of a sample heated @ 420 ◦ C
Raman peaks shifts
F2g5 Eg2 F2g4 F2g3 A1g
181 469 499 601 658
191 479 520 614 686
10 10 21 13 28
C. Burlet et al. / Spectrochimica Acta Part A 80 (2011) 138–147
CuO + NiO + Al2O3 + MnO (%)
25.00
Hetergogenite response
145
(a)
20.00
15.00
10.00
5.00
Spinel response Shinkolobwe response
50.00
60.00
70.00
(b)
14.00
12.00
12.00
10.00
10.00
NiO (%)
CuO (%)
14.00
0.00 40.00
8.00
6.00
4.00
4.00
2.00
2.00
50.00
60.00
70.00
80.00
(c)
0.00 40.00
90.00
50.00
CoO (%) 14.00
12.00
12.00
10.00
10.00
8.00 6.00
2.00
2.00
0.00 60.00
70.00
80.00
90.00
80.00
90.00
CoO (%)
80.00
90.00
(e)
6.00 4.00
50.00
70.00
8.00
4.00
40.00
60.00
CoO (%)
(d)
MnO (%)
Al2O3 (%)
14.00
90.00
8.00
6.00
0.00 40.00
80.00
0.00 40.00
50.00
60.00
70.00
CoO (%)
Fig. 10. Results of the electronic microprobes analysis made on the heterogenite samples. The main secondary cations (Cu, Ni, Mn and Al) are plotted against Co content.
Fig. 10 presents the distributions of the main heterogenite secondary cations (Cu, Ni, Mn and Al) plotted against Co concentration (190 measurements). Other cations including Fe, P, V and Ca were found in small quantities (sum of these cations represent less than 1% equivalent oxide) are not plotted on this figure. On each plot, the response type (heterogenite, Co-spinel, and Ni-bearing) has also been attributed to its corresponding microprobe measurement. A quite clear segregation between Raman signatures distinguishes relatively high cation bearing heterogenite (in terms of total Cu, Ni, Mn and Al cation concentration) from more pure samples (see Fig. 10a). Given the default Raman measurement parameters, the transition between heterogenite-type spectra and spinel-type spectra is found at about 8% of secondary cations present in heterogenite. This trend in mainly driven by Cu cation content which represents the main secondary cation found in our
heterogenite samples (with up to 14% Cu equivalent oxide) (see Fig. 10b). However, another trend is visible in Fig. 10b on some points bearing 2% Cu equivalent oxide. These show no correlation with the Co Content. Those points actually correspond to the measurements presenting high-Ni content (between 6 and 8% Ni equivalent oxide) and plotted in Fig. 10c. One can note that the points bearing more than 8% of equiv. NiO tends give a particular Raman signature with our default parameters. All these measurements were acquired on our Shinkolobwe sample and was described as the Ni-bearing type Raman response. 0–6% of Al oxide equivalent has been found in sample of Mindigi, Kalabi and Etoile. Manganese is also present in variable quantities (0–8% Mn equivalent oxide) in Shinkolobwe, Kalabi and Kabolela samples (Fig. 10d and e).
146
C. Burlet et al. / Spectrochimica Acta Part A 80 (2011) 138–147
625
of synthesised CoO·OH films (then measured at 503 cm−1 for the major peak and 653 cm−1 for the broader one) varie in relative intensity as a function of the synthetic film crystallinity. Fig. 11 illustrates a similar evolution in the spectrum shape of heterogenite bearing various amounts of Cu, reflecting the effect of a secondary cation integrated into the mineral lattice organisation. Finally, Fig. 12 presents a clear relationship between Raman response on heterogenite type spectra and chemistry. Relative peak intensity calculations were carried out on about 100 measurements, covering samples from 7 mines and giving heterogenite type Raman spectra. Every spectra has undergone a baseline correction and a curve-fitting based on the two main heterogenite response peaks at 495 cm−1 and 626 cm−1 . The peak intensity ratios were then plotted against the sum of the main secondary cations content (Cu + Ni + Mn + Al).
495
Intensity (arbitrary untis)
10.16% CuO
6.81% CuO
5.7% CuO
1.20% CuO
8. Conclusion 0.51% CuO
1100
900
700
300
500 -1
Raman shift (cm ) Fig. 11. Five Raman measurements made on polished sections with increasing Cucontent (from bottom to top).
These results can be easily correlated with the work of Deliens [1] establishing a close relationship between Cu cation integration in the heterogenite lattice, mineral crystallinity and its tendency to start its Co-spinel transformation in relatively low temperatures. These considerations, supported by optical microscopy observations during measurements leads to conclude that well-crystallized heterogenite tends to give an heterogenite type spectra, while less-crystallized heterogenite phases tends to adopt the spinel crystallographic structure when exposed the laser. 7.1. Heterogenite response peaks relative intensities and chemistry Another way to look at the influence of heterogenite secondary cations on its Raman response is to focus the observation only on the measurements that give Raman heterogenite type responses. Indeed, Pauporté et al. [28] already showed that the Raman peaks
Cu0+MnO+NiO+FeO+Al2O3 vs v01/env
(Cu0+MnO+NiO+FeO+Al2O3)%
12.00 10.00 8.00 6.00 4.00
y = 6.4332x-1.2573 R2 = 0.8463
2.00 0.00 0
2
4
6
8
10
12
v01/env Fig. 12. Plot of the heterogenite Raman main peaks ratios as a function of the cation content.
Heterogenite (CoO·OH), which is operated in numerous formal and informal mining sites through the Lufilian Arc in the Katanga Province – Democratic Republic of Congo (DRC), results from the oxidation of primary sulphide minerals. It also represents a potential important source of tax incomes for the DRC and its Katanga Province. The illegal extraction and trade of the commodity lead however to a squandering of this natural mineral resource. Initiatives of traceability and certification of minerals (diamond, coltan) are already implemented or under consideration. Heterogenite represents a potential candidate for similar initiatives, but before the implementation of a similar process heterogenite from the Katanga Province needs to be fully characterized. In this paper natural heterogenite is investigated by Raman microspectroscopy and results show that: • Raman spectra of heterogenite unaffected by heat are characterized by a major peak centred at 495 cm−1 and a broader peak centred at 623 cm−1 . The latter is more likely the result of the overlapping of at least three thinner peaks. These positions are in agreement with those reported in studies made on synthetics powders and films of CoO·OH; • heterogenite is very sensitive to heat either due to an artificial heating (oven, furnace) or induced by the laser during Raman microspectroscopy; • laser-induced Co-Spinel Raman responses have been obtained and shows five visible Raman peaks determined at 658 cm−1 (major peak), 601 cm−1 (shoulder), and at 499, 469 and 181 cm−1 . These positions are slightly downshifted regarding the Raman response of an oven/furnace heated heterogenite samples; • the transformation of heterogenite into a Co-spinel structure by artificial heating (oven) is further confirmed by EBSD patterns. The oven-heated sample presented a very close response the Raman peaks calculated and observed for synthetic Co-spinel. The conclusion is that the laser-induced dehydroxylation transforms the mineral in a Co-spinel very locally, preventing a good development of the Co-spinel lattice and inducing a shift in the measured Raman peaks; • Raman spectra and geochemical measurements showed that integration of significant amount of Cu, Mn, Al and Ni in the heterogenite lattice, affect the mineral crystallinity and its sensitivity to heat induced dehydroxylation. This result is in good agreement with the observations of Deliens [1]. A relationship between heterogenite Raman peak ratios and its secondary cations content as been established and its mainly driven by the Cu content for the analysed samples.
C. Burlet et al. / Spectrochimica Acta Part A 80 (2011) 138–147
Acknowledgements The authors thank Sophie Decrée, Thierry De Putter and Florias Mees for providing the accessibility to the mineralogical collection of the Royal Museum for Central Africa. The help of Jacques Longang (ULB) and Estelle Petitclerc (GSB) was strongly appreciated during the preparation and heating tests of heterogenite samples. References [1] M. Deliens, Les Oxydes hydraté de cobalt du Shaba Méridional, Doctorate Thesis, University of Louvain, Belgium, 1972. [2] M. Malzato, Environmental liability in the mining sector: prospects for sustainable development in the Democratic Republic of the Congo, in: J.P. Richards (Ed.), Mining, Society and Sustainable World, Springer XXV, 2010, pp. 289–321. [3] C. Lubaba Nkulu Banza, T. Nawrot, V. Haufroid, S. Decree, T. De Putter, E. Smolders, B. Ilunga Kabyla, O. Numbi Luboya, A. NdalaIlunga, A. Mwanza Mutombo, B. Nemery, High human exposure to cobalt and other metals in Katanga, a mining area of the Democratic Republic of Congo, Environmental Research 109 (2009) 745–752. [4] Panel of Experts, Report of the Panel of Experts on the Ilegal Exploitation of Natural Resources and Other Forms of Wealth on the Democratic Republic of the Congo, S/2001/357, United Nations, New York, 2001. [5] F. Melcher, M.A. Sitnikova, T. Graupner, N. Martin, T. Oberthür, F. Henjes-Kunst, E. Gäbler, A. Gerdes, H. Brätz, D.W. Davis, S. Dewaele, Fingerprinting of conflict minerals: columbite-tantalite (“coltan”) ores, SGA News 23 (2008). [6] H.-C. Liu, S.-K. Yen, Characterization of electrolytic Co3O4 thin films as anodes for lithium-ion batteries, Journal of Power Sources 166 (2007) 478–484. [7] Y. Ding, P. Zhang, Z. Long, Y. Jiang, J. Huang, W. Yan, G. Liu, Synthesis and electrochemical properties of Co3 O4 nanofibers as anode materials for lithium-ion batteries, Materials Letters 62 (2008) 3410–3412. [8] G. Marbán, I. López, T. Valdés-Solís, A.B. Fuertes, Highly active structured catalyst made up of mesoporous Co3 O4 nanowires supported on a metal wire mesh for the preferential oxidation of CO International, Journal of Hydrogen Energy 33 (2008) 6687–6695. [9] F. Gu, C. Li, Y. Hu, L. Zhang, Synthesis and optical characterization of Co3 O4 nanocrystals, Journal of Crystal Growth 304 (2007) 369–373. [10] L.H. Ai, J. Jiang, Rapid synthesis of nanocrystalline Co3 O4 by a microwaveassisted combustion method, Powder Technology 195 (2009) 11–14. [11] M. Oda, H. Ikuta, Site depending impurity effect on the properties of meltprocessed bulk superconductors, Physica C: Superconductivity 460–462 (2007) 301–304.
147
[12] H. Porada, V. Berhorst, Towards a new understanding of the NeoproterozoicEarly Palaeozoic Lufilian and northern Zambezi Belts in Zambia and the Democratic Republic of Congo, Journal of African Earth Sciences 30 (2000) 727–771. [13] J.L.H. Cailteux, A.B. Kampunzu, C. Lerouge, A.K. Kaputo, J.P. Milesi, Genesis of sediment-hosted stratiform copper–cobalt mineralization at Luiswishi and Kamoto, Katanga Copperbelt (Democratic Republic of Congo), Journal of African Earth Sciences 42 (2005) 134–158. [14] A. Francois, Synthèse géologique sur l’Arc cuprifère du Shaba (Republique du Zaïre), Société Belge de Géologie—Centenaire, 1987, pp. 15–65. [15] H.A. El Desouky, P. Muchez, A.J. Boyce, J. Schneider, J.L.H. Cailteux, S. Dewaele, A. von Quadt, Genesis of sediment-hosted stratiform copper–cobalt mineralization at Luiswishi and Kamoto, Katanga Copperbelt (Democratic Republic of Congo), Mineralium Deposita (2010). [16] D. Stijn, M. Philippe, J. Vets, M. Fernandez-Alonzo, L. Tack, Multiphase origin of the Cu–Co ore deposits in the western part of the Lufilian fold-and-thrust belt, Katanga (Democratic Republic of Congo), Journal of African Earth Sciences 46 (2006) 455–469. [17] M. Deliens, H. Goethals, Polytypism of heterogenite, Mineralogical Magazine 39 (1973) 152–157. [18] A. Frenzel, Journal für praktische Chemie 5 (1872) 404. [19] M.H. Hey, Cobaltic hydroxide in nature, Mineralogy Magazine 33 (1972) 253–259. [20] V.G. Hadjiev, M.N. Iliev, I.V. Verligov, The raman spectra of CO3 O4 , Journal of Solid state Physics 21 (1988) L199–L201. [21] H. Shirai, Y. Morioka, I. Nakagawa, Infrared and Raman spectra and lattice vibrations of some oxide spinels, Journal of the Physical Society of Japan 51 (1982) 592–597. [22] W.B. White, B.A. Deangelis, Interpretation of the vibrational spectra of spinels, Spectrochemica Acta 23A (1967) 985–995. [23] D. Kondrashev, N. Fedorova, Comptes rendus de l’Académie des sciences de l’URSS 94 (1954) 229–231. [24] J.W. Anthony, R.A. Bideaux, K.W. Bladh, M.C. Nichols, Handbook of Mineralogy III, Mineral Data Publishing, 1997. [25] S.I. Wright, B.L. Adams, Automatic analysis of electron backscatter diffraction patterns, Metallurgical Transactions A 23A (1992) 759–767. [26] D.P. Field, Recent advances in the application of orientation imaging, Ultramicroscopy 67 (1997) 1–9. [27] J. Yang, H. Liu, W.N. Martens, R.L. Frost, Synthesis and characterization of cobalt hydroxide, cobalt oxyhydroxide, and cobalt oxide nanodiscs, Journal of Physical Chemistry C 114 (2010) 111–119. [28] T. Pauporté, L. Mendoza, M. Cassir, M.C. Bernard, J. Chivotc, Direct lowtemperature deposition of crystallized CoOOH films by potentiostatic electrolysis, Journal of Electrochemical Society 152 (2005) C49–C53. [29] J. Orcel, S. Henin, S. Caillère, Sur les propriétés de la Stainérite, Bulletin de la société franc¸aise de minéralogie 81 (1958) 189.
