Multi-elemental surface mapping and analysis of carbonaceous shale by laser-induced breakdown spectroscopy

Multi-elemental surface mapping and analysis of carbonaceous shale by laser-induced breakdown spectroscopy

Spectrochimica Acta Part B 115 (2016) 31–39 Contents lists available at ScienceDirect Spectrochimica Acta Part B journal homepage: www.elsevier.com/...

1MB Sizes 2 Downloads 91 Views

Spectrochimica Acta Part B 115 (2016) 31–39

Contents lists available at ScienceDirect

Spectrochimica Acta Part B journal homepage: www.elsevier.com/locate/sab

Multi-elemental surface mapping and analysis of carbonaceous shale by laser-induced breakdown spectroscopy Tao Xu a,b, Jie Liu c, Qi Shi a, Yi He d, Guanghui Niu a, Yixiang Duan b,⁎ a

College of Chemistry, Sichuan University, Chengdu 610064, China Research Centre of Analytical Instrumentation, Key Laboratory of Bio-resource and Eco-environment, Ministry of Education, College of Life Science, Sichuan University, Chengdu 610064, China Institute of Hydrogeology and Environmental Geology, Chinese Academy of Geological Sciences, 92 Zhongshan East Road, Zhengding, Hebei Province 050803, China d Analytical & Testing Center, Sichuan University, Chengdu 610064, China b c

a r t i c l e

i n f o

Article history: Received 26 May 2015 Accepted 21 October 2015 Available online 28 October 2015 Keywords: Laser-induced breakdown spectroscopy Shale Mapping Surface analysis Geochemical proxy

a b s t r a c t Gas shale is one of the important unconventional hydrocarbon source rocks, whose composition, such as mineral components and redox sensitive trace elements, has been proved as important geochemical proxies playing essential roles in indicating the gas potential and gas productivity in recent geological researches. Fast and accurate measurements for the shale composition, especially those with spatial resolution, will reveal rich information for the understanding and evaluation of gas shale reservoirs. In this paper, we demonstrated the potentiality as well as feasibility of laser-induced breakdown spectroscopy as an effective technique to perform spectrochemical analysis for shale samples. In case of the bulk analysis of pressed shale pellet, spectral analysis of the plasma emission revealed high sensitivity of LIBS for major, minor and even trace elements. More than 356 lines emitted by 19 different elements can be found. Among these species, redox sensitive trace elements such as V, Cr, and Ni were detected with high signal-to-ratios. Two-dimensional surface micro-analysis for the concerned major or minor elements with strong emissions was then applied to the smoothed shale slab. Local thermodynamic equilibrium for the plasma was first verified with a line profile point-by-point on the sample surface, the matrix effect was then assessed as negligible by the extracted electron density and temperature of the plasmas induced at each position on the same profile. Concentration mappings for the major elements of Si, Al, Fe, Ca, Mg, Na and K were finally constructed with their measured relative variations of line emission intensities. The distribution and correlations of these elements in concentration may reflect changes of shale mineral components with respected to the variations of the depositional environments and provide an important clue in identifying sedimentary processes when combined with other geological or geochemical evidences. These results well demonstrated the potential of LIBS technique for shale studies. © 2015 Published by Elsevier B.V.

1. Introduction Laser-induced breakdown spectroscopy (LIBS) technique has been increasingly utilized in recent years for the analysis of different mineralogy and rock samples in earth scientific research fields [1,2]. More recently, in response to globally rising consumption and demand for hydrocarbon-based products [3,4], researches in unconventional resources, such as gas shale, tight sandstone and oil shale, have been accelerated with the increased hydrocarbon production from shale sequences throughout the world [5–9]. Shales are characteristic of low-permeability and always composed of clay mineral, silicon, and carbonate minerals, etc. These compositions play essential roles in the understanding of gas producibility and gas adsorption capacity in shale [3,9]. Minerals influence the conversion of organic matter during maturation and the release of hydrocarbons during ⁎ Corresponding author. Tel./fax: +86 28 85418180. E-mail address: [email protected] (Y. Duan).

http://dx.doi.org/10.1016/j.sab.2015.10.008 0584-8547/© 2015 Published by Elsevier B.V.

shale processing [10]. It was suggested that smectite, kaolinite, calcite, pyrite, and siderite greatly affect the processing of oil shale [11]. Inhibitory effects of silicate minerals and catalytic effects of carbonate minerals can be observed on pyrolysis reaction of the oil shales [12]. Lithological compositions also play important roles in pore structure and gas storage potential of shale. Permeability is some times higher in shales with high clay content compared with quartz-rich shales due to the open porosity associated with the aluminosilicate fraction [13], and the mutual proportions of clay and quartz appear to be one of the critical factors characterizing the brittleness of shales, which is essential to fracture stimulation [14]. In addition, a certain number of specific redox sensitive trace elements presenting in shales, such as V, Cr, Ni, Co and Mo, etc., have been proved to be important geochemical proxies indicating productive intervals of gas in shale [7,15,16]. Their abundance has been linked to shale strata with increased organic paleoproductivity indicators and gas potential in recent studies [7,16]. Therefore, fast and accurate quantification of shale composition may provide scientific, practical and economic information indispensable

32

T. Xu et al. / Spectrochimica Acta Part B 115 (2016) 31–39

for understanding the genetic and historical features of shales. As a result, highly sensitive detection technique will be involved for the analysis of elements in shale substances. Among available elemental analysis techniques, LIBS offers attractive features such as capacity of multi-elemental detection, rapid response, high spatially-resolved chemical analysis and depth profiling, high sensitivity, and easy sample preparation without restrictions on the shape or size [17,18]. Importantly, compared with the usually employed analytical techniques, such as Inductively Coupled PlasmaOptical Emission Spectrometry (ICP-OES) or Inductively Coupled Plasma-Mass Spectrometry (ICP-MS) [19,20], X-ray microprobes and electron microprobe [21–23], LIBS technique always possesses advantages of relative simplicity and robustness of instrumentation, which allows on-line and in situ measurements. These promising features enable LIBS to be an efficient tool for elemental analysis with proper calibration and validation. Such a tool would be not only suitable for laboratory analysis with high figure of merit, but also applicable for real time monitoring and measurements in geological field work. So far, rare researches have been reported about employing LIBS for qualitative or quantitative characterization of gas shale to the best of our knowledge. However, to apply LIBS technique to achieve an accurate quantitative determination of certain elements in solid geomaterials is always challenging owing to the complex ‘matrix effects’, especially when it comes to two-dimensional elemental mapping analysis, where the elements are measured at different sites in a quantitative or semiquantitative way, the matrix effect resulted from the dependence of the properties of the plasma on the physical and chemical proprieties of the sample surface is accentuated and becomes a major concern [17]. In general, this is because of the significant variation of the induced plasma caused by the inhomogeneity of sample while being scanned by the ablation laser pulses even with constant ablation conditions. As a result, in practice, LIBS measurements for large area of the inhomogeneous sample surface always render it difficult to use matrix-matched laboratory standards. This leads the calibration more difficult. In this work, we demonstrate an exploratory study on the feasibility of LIBS for highly sensitive and rapid elemental measurements as well as multi-elemental surface mapping of shale elemental compositions. Following the description of our experimental design, we present our results of elemental detection and multiple-elemental concentration mappings of the shale sample using LIBS. Nineteen elements were detected in a typical carbonaceous shale collected in the formation of Longmaxi in Sichuan Basin, China. Experimental results are presented by a list of elements with spectral emission lines identified in LIBS spectra. Detailed spectra emitted by the identified trace elements are also presented with the aim to show their LIBS measurement sensitivities. Prior to the multi-elemental surface mapping of the polished shale slab section, a linear profile point by point measurement across the sample was first carried out. Local thermodynamic equilibrium (LTE) state was verified by the calculated spectral parameters of the plasma. The matrix effect was then evaluated with the extracted electron densities and plasma temperatures for the same linear profile point-by-point measurements. The results showed negligible matrix effect in our case over the analyzed surface area despite the different sedimentary laminae areas. Given this condition, the relative elemental concentrations were then correlated to the corresponding line emission intensities, leading to the construction of Si, Al, Fe, Ca, Mg, K and Na concentration mappings for this shale slab sample. Correlations of the relative concentration variations of the detected elements with the spatial distribution of mineral components in shale sample could be clearly observed. The detailed information obtained using LIBS can contribute to better understanding of the properties of gas shales. 2. Experimental For this experiment, a Nd:YAG laser at the fundamental wavelength of 1064 nm providing pulses with duration in 4–7 ns FWHM was used

for the shale materials ablation with a repetition rate of 2 Hz. Pulse energy was fixed at 60 mJ measured by a calorimeter in order to get high detection sensitivity. The laser beam is focused on the sample surface using a bi-convex quartz lens with an 80 mm focal length. The focused spot size was estimated to be 250 μm in diameter, leading to a laser radiation of about 2.04 × 1010 w cm−2. An echelle spectrometer equipped with an ICCD camera (LTB ARYELLE 200 and iStar from Andor Technology, spectral range: 220 nm to 800 nm, spectral resolution: λ/Δλ = 9000) was employed to disperse and record the spectra. Emission from the induced plasma was collected with an off-axis parabolic mirror at an incidence angle of 45° and delivered by a one-meter long coupling fiber optic system (400 μm core diameter) onto the echelle spectrometer entrance slit. The object distance of the off-axis parabolic mirror is set 20 cm. Prior to the experiments, the spectrometer calibration of wavelength and the spectral response of the whole detection system were performed with its mercury argon lamp and a deuterium tungsten halogen light source (AvaLight-DH-BAL-CAL from Avantes) respectively. For our experiment, the typical shale core drilling sample was collected from the formation of Longmaxi in Sichuan Basin, China. One segment of the shale bore core was split using a rock slicing machine with one half polished for LIBS scanning and the other divided into small sections. About 150 g of the small shale sections was crushed and milled to fine powder (b74 μm particle size) with agate ball machine grinder and was sampled for assay by both XRF and ICP-OES analysis. Pellets were then prepared for LIBS measurement. The shale sample probed in this study was analyzed in the XRF and ICP-OES labs at the National Research Center of Geoanalysis, China (under the direction of Yingchun Li) using standard operating procedures. For the XRF analysis of major elements, samples were prepared as fused lithium borate glass discs using a modification of the method of Claisse and Samaon [24]. The shale power sample was first prepared with flux fusion method in high frequency induction melting furnace at 1150 °C for 10 min in order to oxidize the iron to Fe3+ and remove volatiles. All major elements were measured simultaneously using a PANalytical AXIOS XRF spectrometer. Intensity correction for nonlinear background, inter-element interferences, and calibrations were carried out with SuperQ version 4.0 J software containing empirical coefficient method for matrix correction from PANalytical Company. Verified by the national standard sample GBW07107, the results of analysis by WDXRF can coincide with the calibration values. Pretreatment method of alkali fusion modified from those of Morgan [25] and Markey [26] was conducted for trace elements. The starting shale powder was mixed with fusion mixture(the weight ratio of shale: NaOH:Na2O2 = 1:15:10)and then heated at 500 °C in corundum crucible for 7 min, which could assist in dissolving any oxide form of the elements present in shale. The fused material was then added to distilled water and further dissolved for trace elements determination by ICP-OES. To form the pellets for bulk elemental LIBS measurements, 1.5 g of the milled fine powder shale sample was poured into a die with 20 mm inside diameter and 2 tons of a hydraulic pressure was applied for 20 s. The resultant pellet sample depth is about 2 mm, which was properly thick for LIBS experiments. As shown in Fig. 1, the retained half of the split shale bore core with fine laminated structure was then polished with diamond powder suspension and cleaned with ultrapure water for two-dimensional surface spatially-resolved analysis. To avoid the strong continuum radiation immediately after the laser impact on the sample due to the mechanisms involving free electrons, as well as the consideration of the deviations in the temporal profile of the emission lines resulted from their difference in the spontaneous emission coefficients, a delay time of 1.0 μs and a detection window of 5.0 μs were applied for the plasma emission measurements. This optimized detection window allowed achieving good signal-to-noise ratios (SNRs) and enhanced detection limit. For the pressed shale pellet, 10 spectra of 20 laser shots for different fresh sites were obtained to get an averaged emission intensities in order to deduce the effect of the

T. Xu et al. / Spectrochimica Acta Part B 115 (2016) 31–39

33

Table 1 Laboratory assay of the milled shale fine powder. The major elements weight percent (wt.%) were determined by WDXRF spectroscopy, and the trace element contents (μg/g) were analyzed by ICP-OES.

pyrite-bearing shale laminations 3 mm

Fig. 1. Photo of the studied shale section. The rectangle in the figure (3 mm × 9 mm) delimits the measured area of multi-elemental surface mappings.

inhomogeneity from direct LIBS measurement of solid samples. In the case of the smoothed slab sample, each spectrum was accumulated over 10 laser shots at the same site in order for a sensitive acquisition analysis. The polished shale slab was moved quickly on a X–Y–Z motorized stage point by point by 300 μm step after each emission acquisition.

Concentrations (μg/g)

Na Mg Al Si P K Ca Ti Mn Fe (total) S

0.03 1.27 4.77 26.70 0.05 1.71 1.77 0.28 0.02 6.01 2.19

V Cr Ni Rb Sr Y Zr Cu Zn Ba Nb Co Mo

50.41 41.60 220.02 36.58 143.75 11.81 199.88 47.93 168.71 4164.77 6.99 45.30 90.20

KI 766.49

BaI 705.99

LiI 670.79

CaI 649.38

NaI 589.00 BaII 614.17 CaI 616.22

BaII 493.41

BaII 553.55

SrI 460.73

MgI 383.83

AlI 396.15 FeI 404.58 CaII 396.85 CaI 422.67 FeI 440.48 BaII 455.42

A typical LIBS spectrum of the pressed shale sample is shown in Fig. 2 with a large number of lines over an acceptable background. The plasma was induced on the pressed shale pellet surface. The spectrum calibrated for the spectral response ranges from 220 to 800 nm. Via meticulous identification line-by-line based on NIST spectral database [27], a list of more than 356 characteristic lines was identified, including the lines emitted by atomic or ionic species as well as a certain number of ghost lines. The spectrum is clearly dominated by emissions from the major elements of the shale material, such as Si, Al, Fe, Ca, Mg and Ti, etc., with pretty good SNRs, indicating that these typical major elements can be easily detected because of their relatively higher concentrations. No apparent self-absorption effect such as the self-reversal effects is observed for most of the strong emission lines under this present experimental conditions. In addition, emissions from minor or trace elements such as Li, Rb, Sr, and Ba can also be identified simultaneously, with high SNRs owing to their relative simple electronic configurations, illustrating that their concentrations are also above the corresponding limits of LIBS detection in shale samples.

AlI 309.28 TiII 334.92 FeI 344.07 FeI 357.02

Trace elements

MgI 518.36

3.1. Typical shale spectrum

MgII 279.55 SiI 288.16

Concentrations (wt.%)

For emission line identification, the wavelengths of the lines as well as the comparison of their intensities to the transition probabilities indicated in the NIST database were considered first. Especially in cases of the elements with many observed emission lines, to further verify false line designations, correlations for their different lines were also established between the experimental intensities and the relative intensities indicated in the NIST database. In addition, in order to deduce possible matrix effect and self-absorption for strong experimental lines, the abundances of the elements in shale sample listed in Table 1 were also taken into account and were compared with the spectral line intensities. These check procedures allow us to eventually identify most of the different elements with emission lines in very close wavelengths in shale sample. A list of 19 elements identified in the spectrum of the pressed shale sample is provided in Table 2. For the elements emitting large numbers of lines, only the most intense lines in the observed spectral region are listed. Besides the major elements (Si, Al, Fe, Ca), a large number of minor or trace elements including Na, K, Li, Sr, Ba, Cu, V, Cr, and Ni, as well as non-metallic elements C, H, and O are also included. As indicated in Table 1, most of the minor elements always have a low concentration in common carbonaceous shale samples. However, emissions from some of the trace elements like V, Cr, and Ni were even distinctly visible at this intensity scale, and were measured with good SNRs. To illustrate the sensitivity of LIBS detection for these metallic trace elements in shale, the detailed spectra of the shale samples are shown in Fig. 3. From these spectra, we can see that emissions from certain numbers of trace elements like V (Fig. 3 c, d), Cr (Fig. 3 d, e), Ni (Fig. 3 b) and Cu

3. Results and discussion

SiI 251.61

Major elements

CaI 527.03

9 mm

continuous black shale laminations

Fig. 2. Typical LIBS spectrum of a carbonaceous shale. The plasma is induced on the pressed shale sample surface.

34

T. Xu et al. / Spectrochimica Acta Part B 115 (2016) 31–39

Table 2 List of identified spectral lines observed in the LIBS spectrum of carbonaceous shale. Species

Wavelength (nm)

HI CI Li I Na I Mg I Mg II Al I Si I KI Ca I

656.28 247.85 460.28 588.99 277.98 279.08 308.22 250.69 766.49 422.67 526.56 616.22 315.89 363.55 398.98 307.87 337.28 403.08 248.33 346.59 374.83 406.36 234.35 260.71 460.73 407.77 780.03 599.71 728.03 455.4 410.52 425.43 344.63 324.75

Ca II Ti I Ti II Mn I Fe I

Fe II Sr I Sr II Rb I Ba I Ba II VI Cr I Ni I Cu I

460.29 589.59 285.21 279.55 309.27 251.43 769.9 430.25 558.2 616.96 317.93 364.27 399.86 308.8 338.38 403.31 248.81 353.31 374.95 407.17 238.20 261.19 640.85 421.55 794.76 611.08 767.21 493.41 411.18 520.84 349.3 327.4

670.78

670.79

383.23 279.8 309.28 251.61

383.83 280.27 394.4 251.92

517.27

518.36

396.15 252.41

252.85

288.16

390.55

430.77 558.88 643.91 393.37 365.35 498.17 323.45 375.93 403.45 249.06 355.85 375.82 411.85 239.56 269.26

442.54 559.45 646.26 396.85 375.29 499.11 323.66 376.13

443.5 559.85 647.17

445.48 585.75 649.38

487.81 610.27 671.77

518.88 612.22 714.82

394.87 500.72 323.9

395.63 506.47 334.9

395.82

398.18

334.94

336.12

252.28 358.12 382.04 413.47 240.49 273.95

271.9 371.99 383.422 438.35 249.33 274.93

278.81 373.49 385.99

299.44 373.71 388.63

344.06 374.56 404.58

258.59 275.57

259.84

259.94

649.88

652.73

659.53

669.38

705.99

712.03

553.55 437.92

585.37 438.99

614.17

649.69

351.51

The boldfaced data are used for 2-D multi-elemental spatially-resolved analysis.

(Fig. 3 a) are still observable in these spectra with satisfying SNRs of ~ 50 – ~ 1500. This means that the concentrations for these trace elements in the shale sample are still higher than their LIBS detection limits using our LIBS experimental setup. These trace elements are important geochemical proxies for reservoir property analysis as discussed in the introduction section of this paper. Their abundance has been linked to the paleo-productivity and gas potential in recent studies as important indicators [7,16]. To the best of our knowledge, no research has been reported about the detection of such geochemical proxy elements employing LIBS technique as used in our experiment. Nevertheless, emissions from other trace elements such as Zn are not yet observed at this intensity scale, a spectrometer with broader UV spectrum range may be needed. According to the certified concentration data of elements listed in Table 1, we may have a general idea of what orders of magnitude of the trace elements concentration can be detected by LIBS. A good detection limit lower than tens μg/g can be obtained for these trace elements such as V, Cr, Ni and Cu under our LIBS experimental conditions. 3.2. Multi-elemental spatially-resolved analysis 3.2.1. Evaluation of LTE for a linear profile on the sample surface It is critical to determine the establishment of LTE condition especially for the spatially-integrated detection for the inhomogeneous samples [28,29]. When LTE conditions are established, the plasma can be described by a unique temperature and therefore exploited for elemental detection and quantification usually with lines of high intensities [28, 30]. In this work, linear scan with 30 equi-step sampling points was executed on the shale sample surface perpendicular to the shale sedimentary laminae prior to 2-D mapping. The linear profile crossed over the distinct colored laminar regions with the colors ranging from black to the pyrite-bearing laminations, representing the whole sample surface.

The electron density (Ne), which determines the necessary condition for LTE according to the McWhirter's criterion [31], was first measured. The electron density was calculated from the Stark broadening of the atomic emission line of Si (I) at 288.16 nm with the needed spectral data extracted from Griem [32]. This line is well isolated, since its transition is not bound to the ground state of the silicon atomic, the contributions of resonance broadening were insignificant under our present experimental conditions. For the calculation of the electron density, the Stark broadenings were extracted using the Voigt profile fitting functions containing Gaussian broadening contribution from the spectrometer resolution [33]. The calculated values of the electron densities of the laser ablation transient plasma were within the range of 6.0 × 1017 cm−3–7.0 × 1017 cm−3. Such order of magnitude of Ne fulfills the critical electron densities condition estimated by the McWhirter's criterion, which is necessary to ensure the LTE conditions in our present experiments. 3.2.2. Assessment of the matrix effect Prior to the performance of the elemental mapping for the shale slab, the matrix effect in this experiment was assessed by measuring the electron densities and temperatures of the plasmas induced on each different positions of the linear profile. The electron density was calculated for these 30 equi-step measurement points with the same protocol as described above. In further checks related to the temperatures, Boltzmann plots [34] were performed to extract the excitation temperatures. As shown in Table 3, atomic Fe lines of weak self-absorption, which have been successfully used in LIBS analysis [35], were selected to extract the excitation temperatures with the spectral parameters taken from the NIST database [27]. The results shown in Fig. 4 illustrate random dispersions of the electron densities and the temperatures along the profile. This suggests no existence of correlation between these observables and the colored laminations. In terms of the values of the relative standard deviation (RSD), the measured electron density and the temperature show fluctuations along the linear profile of 3.4% and 4.8% respectively. Such values of RSDs can be well compatible with those of typical LIBS measurements due to experimental fluctuations, indicating that the insignificant matrix effect can be reasonably ignored for the LIBS analysis of the shale sample. With the negligible matrix effect, lines emission intensities of elements can thus be linearly related to their concentrations in the shale sample, which contributes to the achievement of the reconstruction of the elemental concentration profiles. 3.2.3. Two-dimensional spatially-resolved analysis on the shale surface Two-dimensional spatially-resolved analysis of relative variations of elemental concentrations was performed by scanning over the rectangular sample surface of 3 mm × 9 mm as shown in Fig. 1. The scanning was preformed point by point with an equal step-length of 300 μm in the form of regular grid. Each spectrum was accumulated for 10 laser pulses at the same point with the laser repetition of 2 Hz. Microscope imaging allowed the size of the craters caused by laser ablation to be measured at equal step-length of 300 μm without mutual influence. The two-dimensional spatial resolution was therefore 300 × 300 μm2. One or more emission lines for each concerned element were therefore selected to indicate the corresponding elemental concentration with their intensities. Besides visual inspection of the line shape, emission lines often with a lower state other than the ground state were carefully chosen to ensure the reduction of self-absorption. With the method employed in Ref. [26], the line emission intensities of the boldface lines listed in Table 2 were used to evaluate the relative variation of elemental concentration over the surface profile. Specifically, for a given element M, one of its emission lines was first integrated to get a mean intensity for all of the measured (the 30 × 10 sampling) points, ĪM λ = M M 1/300 × ∑300 i = 1Iλ,i, where Iλ,i denotes the neat peak height intensity of the emission line at a wavelength λ of the given element M measured at the sampling point i in the profile. Then, the relative variation of the

T. Xu et al. / Spectrochimica Acta Part B 115 (2016) 31–39

35

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

Wavelength (nm) Fig. 3. Detailed LIBS spectra of trace elements from the carbonaceous shale.

line intensity, which is considered to be positively correlated with the relative variation of the elemental concentration over the surface profile,

Table 3 The parameters used for plasma temperature calculation.

M

Lines

λ (nm)

Aki

Ei (eV)

Ek (eV)

Fe I Fe I Fe I Fe I Fe I Fe I Fe I Fe I Fe I Fe I Fe I Fe I Fe I

370.56 370.92 372.26 372.76 373.49 373.71 374.34 374.56 374.83 374.95 375.82 376.38 376.72

3.21E + 06 1.56E + 07 4.97E + 06 2.24E + 07 9.01E + 07 1.41E + 07 2.60E + 07 1.15E + 07 9.15E + 06 7.63E + 07 6.34E + 07 5.44E + 07 6.39E + 07

0.0516 0.9146 0.0873 0.9582 0.8590 0.0516 0.9901 0.0873 0.1101 0.9146 0.9582 0.9901 1.0111

3.3965 4.2562 3.4170 4.2833 4.1777 3.3683 4.3013 3.3965 3.4170 4.2204 4.2562 4.2833 4.3013

M

M M can be calculated for each sampling point i, ΔC M λ;i =C λ ¼ ΔI λ;i =I λ ¼ ðI λ;i − M

Iλ Þ=IM λ;i . In the case of the considered element with several lines availM

able, an averaged relative variation of concentration, ΔC M i =C , could still be obtained for each sampling point with further averaging over these relative variations of concentration. Surface mappings of the relative variations of concentration in Fig. 5 are presented with a natural light photo of the sample for elements of Si, Al, Fe, Ca, Mg, Na, and K. The spatially-resolved distributions of the elements shown in Fig. 5 are considerably different. As for the profiles of relative variations in the Si and Al concentration, our LIBS measurements show a moderate fluctuation with observed relative concentration variation levels of 0.8 and 0.6 respectively. Since Si and Al are the

36

T. Xu et al. / Spectrochimica Acta Part B 115 (2016) 31–39

(a)

(b)

Fig. 4. Profiles of the electron density Ne (a) and the excitation temperature T(K) (b) on the sample surface. Error bars for Ne were estimated by considering the 20% accuracy of the stark broadening parameter (w, electron collision parameter) as well as the errors from the Voigt-profile fitting for the atomic emission line of Si(I) at 288.16 nm, and the statistical errors associated to the T(K) retrieved from the Boltzmann linear fitting.

matrix elements of the analyzed shale, such fluctuations could be considered due to their relative stable distribution over the shale surface. Observations of the polished slab section, as is shown in Fig. 1, indicate that the black shale is mainly composed of organic-rich fine laminae without any evidence of sedimentary rate changes, such as gravity-generated coarse-grained layers. The lithofacies and sedimentary features may suggest a constant deposition in quiet environment condition that dominate the development of the shale deposition. Nevertheless, we also notify pronounced changes of the relative variation of the concentration of Si in Fig. 5 at some detected sites, this may be associated with different mineralogical compositions characterized by aluminosilicate minerals or silty material. While for the other elements, such as Fe, this fluctuation presents significantly larger values which corresponds to the pyrite scattering in the shale. As is shown in Figs. 1 and 5, submillimeter-sized pyrite clusters occur randomly and individually in the black shales, especially in the middle part of which, pyrite crystals are concentrated and displace laminations in the matrix, suggesting that they were formed before the shale sediments were compacted. Pyrite is one common authigenic mineral which occurs as euhedral crystals or framboids (closely packed, spherical aggregates of uniform-sized microcrysts) in both modern anoxic sediments and ancient sedimentary rocks. Pyrites generally form just below the oxic/euxinic interface, normally below or within centimeters of the sediment–water interface underlying oxic bottom water columns [36], and the maximum framboid diameter of pyrite can be used to indicate water column redox conditions [36]. Pyrite framboid size distributions have therefore been used to indicate whether fine-grained sedimentary black shales were deposited under oxic or anoxic conditions [36] and trace marine redox condition

alternations through geological time [37,38]. In our present work, the apparent fluctuations of the values of pyrite cluster size throughout the whole section may represent microenvironmental changes during deposition of the continuous laminations of the black shale. As a consequence, the relative variations of Fe concentration obtained with LIBS may be well interpreted as reflecting such variations of the deposition visible in the natural light photos. Noticeable changes of the concentration of Na and K can also be observed over the surface profile. While in the cases of other elements of Ca and Mg, the fluctuations present smaller relative concentration variations. Taking into account the experimental fluctuations observed from the calculated Ne and T(K) in Fig. 4, the observed concentration variations of the typical digenetic elements over the sample surface can therefore be determined with good confidence and considered as due to the variation of the concentration of the elements associated to mineralogy types in shale. In order to reveal the possible spatial correlation among groups of elements, the linear correlation coefficient R was further calculated using the relative concentration variations measured over the analyzed area for different pairs of elements and listed in Table 4. As for the different groups of the correlated elements, such as Si and Al as well as Na and K, etc., these correlation coefficients and the surface mappings shown in Fig. 5 could be considered to be related to a structural correlation. A threshold of R = 0.46 was then chosen to allow the distinction of different mineralogical composition distributions in shale. In general, the gas shales from the Longmaxi formation, Lower Silurian in Sichuan Basin, China, are usually characteristic of laminated structure and mainly composed of the minerals such as quartz, feldspar and pyrite, clay minerals, calcite and dolomite minerals [39]. The correlated elements

T. Xu et al. / Spectrochimica Acta Part B 115 (2016) 31–39

37

Fig. 5. Two-dimensional mappings of relative variation of concentrations for typical major elements. The analyzed areas on the natural light photo are presented as covering the shale laminations and pyrite layer. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article).

of Si and Al may correspond to the aluminosilicate minerals in the silty and clay phases in shale, such as the potassium feldspar (K(AlSi3O8)), albite (Na(AlSi3O8)), as well as the clay minerals of illite (KAl3Si3O10(OH)2) and montmorillonite ((Na, Ca)0.33(Al, Mg)2Si4O10(OH)2), etc. Another correlation of the group concerning Na and K probably corresponds the presence of the feldspar minerals as well as the sodium-bearing and potassium-bearing clay mineral components in the shale. Lamination is one common feature of shale rocks. The lamina of shale are developed with different mineral compositions. Although it Table 4 Linear correlation coefficients between the different detected elements. R

Si

Al

Fe

Ca

Mg

Na

K

Si Al Fe Ca Mg Na K

1

0.46 1

0.39 −0.32 1

0.67 0.71 0.21 1

0.37 0.46 0.28 0.34 1

0.41 0.71 −0.069 0.25 0.21 1

0.32 0.66 0.025 0.21 0.29 0.62 1

is really not easy to ascertain the different mineral types owing to the fact of the complexity of shale components, the elemental distribution acquired from our LIBS measurements point by point may reveal the sedimentary source provenance-specific features of shales. This will provide an important clue in identifying sedimentary processes when combined with other geological and geochemical evidences. For other cases, a well correlation is also obtained between the concerned Al and Ca elements, as well as an average one for Al and Mg over the analyzed area. This accords with the quasiuniform distributions of Ca and Mg observed in Fig. 5 with their smaller relative concentration variation levels of ~0.7 and ~0.8 respectively. In view of their low concentration shown in Table 1, this may demonstrate to some extent the yielding of small amounts of carbonate and dolomite components during the shale development. 4. Conclusions We have performed multi-elemental detection and twodimensional surface mapping of shale samples benefited from the satisfying spatial resolution of LIBS technique. Our results obtained in this

38

T. Xu et al. / Spectrochimica Acta Part B 115 (2016) 31–39

work have shown the potentiality of LIBS for elemental measurement and spatially-resolved elemental analysis of shale samples. The use of fundamental wavelength of Nd:YAG nano-second laser for shale ablation provides sensitive detection under our experimental conditions. A certain number of major, minor or trace elements can be identified in the LIBS spectra. In addition to the alkali trace elements Li, Rb, Sr and Ba, etc., transition elements, such as V, Cr, and Ni were also observed. These elements are important geochemical proxies that have been linked to shale strata with increased organic paleoproductivity and gas potential in recent studies. For the best of our knowledge, rare related researches have been reported for their detection with LIBS so far. By using an optimized dedicated spectroscopic detection system or double-pulsed LIBS technique, the limit of detection can be significantly decreased for these concerned trace elements. Accumulation of larger numbers of spectra can be expected to reduce statistical dispersions of the bulk shale measurements with LIBS method. For spatially-resolved measurement, with the verified LTE for the induced plasma point-by-point for a line profile under the experimental condition, the matrix effect was further checked and presented as negligible. Relative variations of elemental concentration were therefore positively related to the emission intensities of the analyzed elements. The relative variations of concentration for typical major elements were measured over a delimited rectangle area through a twodimensional point-by-point analysis in regular grid. Elemental concentration mappings were reconstructed for the concerned elements using their LIBS emission line intensities. The elemental concentration mappings may to some extent show the concentration variations of these elements in shale lamination, which assists in indicating the spatial distribution features of different mineral assemblages. The spatial resolution of LIBS should be further improved by using a microscopic objective to focus the laser beam for ablation in order to obtain more information of the fine laminar structures of the shales. Furthermore, elaborate works are still needed for a full diagnostic of the plasma generation for complex shale substances. When an ideal plasma is obtained, the matrix effects inherent in the LIBS measurement could be better overcome. This will be very useful especially for trace elements determination in the form of ratios [28,40], and multi-elemental mapping for samples of strong inhomogeneities. Acknowledgments Authors are thankful to the financial support from the National Major Scientific Instruments and Equipment Development Projects of China (No. 2011YQ030113), the National Recruitment Program of Global Experts (NRPGE), and the start-up funding of Sichuan University for setting up the Research Centre of Analytical Instrumentation. References [1] G.S. Senesi, Laser-Induced Breakdown Spectroscopy (LIBS) applied to terrestrial and extraterrestrial analogue geomaterials with emphasis to minerals and rocks, EarthSci. Rev. 139 (2014) 231–267, http://dx.doi.org/10.1016/j.earscirev.2014.09.008. [2] R.S. Harmon, R.E. Russo, R.R. Hark, Applications of laser-induced breakdown spectroscopy for geochemical and environmental analysis: a comprehensive review, Spectrochim. Acta B 87 (2013) 11–26, http://dx.doi.org/10.1016/j.sab.2013.05.017. [3] D.J. Ross, R.M. Bustin, Characterizing the shale gas resource potential of Devonian– Mississippian strata in the Western Canada sedimentary basin: application of an integrated formation evaluation, AAPG Bull. 92 (2008) 87–125, http://dx.doi.org/10. 1306/09040707048. [4] D. Zhang, J. Zhang, Y. Wang, Y. Tang, W. Yu, China's unconventional oil and gas exploration and development: progress and prospects, Resour. Sci. 37 (2015) (http://dx.doi.org/1007-7588(2015)05-1068-08). [5] Z. Xingang, K. Jiaoli, L. Bei, Focus on the development of shale gas in China—based on SWOT analysis, Renew. Sust. Energ. Rev. 21 (2013) 603–613, http://dx.doi.org/10. 1016/j.rser.2012.12.044. [6] G.R. Chalmers, R.M. Bustin, I.M. Power, Characterization of gas shale pore systems by porosimetry, pycnometry, surface area, and field emission scanning electron microscopy/transmission electron microscopy image analyses: examples from the Barnett, Woodford, Haynesville, Marcellus, and Doig units, AAPG Bull. 96 (2012) 1099–1119, http://dx.doi.org/10.1306/10171111052.

[7] D.J. Ross, R.M. Bustin, Investigating the use of sedimentary geochemical proxies for paleo-environment interpretation of thermally mature organic-rich strata: examples from the Devonian–Mississippian shales, Western Canadian Sedimentary Basin, Chem. Geol. 260 (2009) 1–19, http://dx.doi.org/10.1016/j.chemgeo.2008.10. 027. [8] R. Lahann, M. Mastalerz, J.A. Rupp, A. Drobniak, Influence of CO2 on New Albany Shale composition and pore structure, Int. J. Coal Geol. 108 (2013) 2–9, http://dx. doi.org/10.1016/j.coal.2011.05.004. [9] Y. Chen, A. Furmann, M. Mastalerz, A. Schimmelmann, Quantitative analysis of shales by KBr-FTIR and micro-FTIR, Fuel 116 (2014) 538–549, http://dx.doi.org/10. 1016/j.fuel.2013.08.052. [10] S. Bhargava, F. Awaja, N.D. Subasinghe, Characterisation of some Australian oil shale using thermal, X-ray and IR techniques, Fuel 84 (2005) 707–715, http://dx.doi.org/ 10.1016/j.fuel.2004.11.013. [11] J. Patterson, D. Henstridge, Comparison of the mineralogy and geochemistry of the Kerosene Creek Member, Rundle and Stuart oil shale deposits, Queensland, Australia, Chem. Geol. 82 (1990) 319–339, http://dx.doi.org/10.1016/0009-2541(90)90088-O. [12] A. Karabakan, Y. Yürüm, Effect of the mineral matrix in the reactions of shales. Part 2. Oxidation reactions of Turkish Göynük and US Western Reference oil shales, Fuel 79 (2000) 785–792, http://dx.doi.org/10.1016/S0016-2361(99)00200-8. [13] D.J.K. Ross, R. Marc Bustin, The importance of shale composition and pore structure upon gas storage potential of shale gas reservoirs, Mar. Pet. Geol. 26 (2009) 916–927, http://dx.doi.org/10.1016/j.marpetgeo.2008.06.004. [14] D.M. Jarvie, R.J. Hill, T.E. Ruble, R.M. Pollastro, Unconventional shale–gas systems: the Mississippian Barnett Shale of north-central Texas as one model for thermogenic shale–gas assessment, AAPG Bull. 91 (2007) 475–499, http://dx.doi.org/10.1306/ 12190606068. [15] T.J. Algeo, H. Rowe, Paleoceanographic applications of trace-metal concentration data, Chem. Geol. 324 (2012) 6–18, http://dx.doi.org/10.1016/j.chemgeo.2011.09. 002. [16] B.B. Sageman, A.E. Murphy, J.P. Werne, C.A. Ver Straeten, D.J. Hollander, T.W. Lyons, A tale of shales: the relative roles of production, decomposition, and dilution in the accumulation of organic-rich strata, Middle–Upper Devonian, Appalachian basin, Chem. Geol. 195 (2003) 229–273, http://dx.doi.org/10.1016/S0009-2541(02)00397-2. [17] S. Musazzi, U. Perini, Laser-Induced Breakdown Spectroscopy: Theory and Applications, Springer, 2014. [18] V. Piñon, M. Mateo, G. Nicolas, Laser-induced breakdown spectroscopy for chemical mapping of materials, Appl. Spectrosc. Rev. 48 (2013) 357–383, http://dx.doi.org/ 10.1080/05704928.2012.717569. [19] S.N. Madjid, I. Kitazima, T.J. Lie, H. Kurniawan, K. Kagawa, K. Ikezawa, T. Maruyama, Spectrochemical analysis using low-background laser plasma induced by Nd-YAG laser at low pressure, Jpn. J. Appl. Phys. 42 (2003) 3452, http://dx.doi.org/10.1143/ JJAP.42.3452. [20] C. Storey, M. Smith, T. Jeffries, In situ LA-ICP-MS U–Pb dating of metavolcanics of Norrbotten, Sweden: records of extended geological histories in complex titanite grains, Chem. Geol. 240 (2007) 163–181, http://dx.doi.org/10.1016/j.chemgeo. 2007.02.004. [21] A. Kuczumow, B. Vekemans, O. Schalm, K. Gysels, C.-U. Ro, R. Van Grieken, Analysis of speleothems by electron and X-ray microprobes, J. Anal. At. Spectrom. 16 (2001) 90–95, http://dx.doi.org/10.1039/B007725I. [22] A. Kuczumow, D. Genty, P. Chevallier, J. Nowak, M. Florek, A. Buczyńska, X-ray and electron microprobe investigation of the speleothems from Godarville tunnel, X-Ray Spectrom. 34 (2005) 502–508, http://dx.doi.org/10.1002/xrs.877. [23] G. Chu, Q. Sun, S. Li, Y. Lin, X. Wang, M. Xie, W. Shang, A. Li, K. Yang, Minor element variations during the past 1300 years in the varved sediments of Lake Xiaolongwan, north-eastern China, GFF 135 (2013) 265–272, http://dx.doi.org/10.1080/11035897. 2013.833548. [24] F. Claisse, C. Samson, Heterogeneity Effects in X-ray Analysis, Springer, 1962. [25] J.W. Morgan, R.J. Walker, Isotopic determinations of rhenium and osmium in meteorites by using fusion, distillation and ion-exchange separations, Anal. Chim. Acta 222 (1989) 291–300, http://dx.doi.org/10.1016/S0003-2670(00)81904-2. [26] R. Markey, H. Stein, J. Morgan, Highly precise Re–Os dating for molybdenite using alkaline fusion and NTIMS, Talanta 45 (1998) 935–946, http://dx.doi.org/10.1016/ S0039-9140(97)00198-7. [27] NIST, http://physics.nist.gov/PhysRefData/ASD/lines_form.html. [28] W. Lei, V. Motto-Ros, M. Boueri, Q. Ma, D. Zhang, L. Zheng, H. Zeng, J. Yu, Timeresolved characterization of laser-induced plasma from fresh potatoes, Spectrochim. Acta B 64 (2009) 891–898, http://dx.doi.org/10.1016/j.sab.2009.07.015. [29] Q. Ma, V. Motto-Ros, W. Lei, M. Boueri, L. Zheng, H. Zeng, M. Bar-Matthews, A. Ayalon, G. Panczer, J. Yu, Multi-elemental mapping of a speleothem using laserinduced breakdown spectroscopy, Spectrochim. Acta B 65 (2010) 707–714, http:// dx.doi.org/10.1016/j.sab.2010.03.004. [30] M. Sabsabi, P. Cielo, Quantitative analysis of aluminum alloys by laser-induced breakdown spectroscopy and plasma characterization, Appl. Spectrosc. 49 (1995) 499–507. [31] G. Cristoforetti, A. De Giacomo, M. Dell'Aglio, S. Legnaioli, E. Tognoni, V. Palleschi, N. Omenetto, Local thermodynamic equilibrium in laser-induced breakdown spectroscopy: beyond the McWhirter criterion, Spectrochim. Acta B 65 (2010) 86–95, http://dx.doi.org/10.1016/j.sab.2009.11.005. [32] H. Griem, Spectral Line Broadening by Plasmas, Elsevier, 2012. [33] M. Gigosos, S. Mar, C. Pérez, I. De la Rosa, Experimental Stark widths and shifts and transition probabilities of several Xe II lines, Phys. Rev. E 49 (1994) 1575, http://dx. doi.org/10.1103/PhysRevE.49.1575. [34] W.T.Y. Mohamed, Study of the matrix effect on the plasma characterization of six elements in aluminum alloys using LIBS with a portable echelle spectrometer, Prog. Phys. 42 (2007).

T. Xu et al. / Spectrochimica Acta Part B 115 (2016) 31–39 [35] M. Dong, X. Mao, J.J. Gonzalez, J. Lu, R.E. Russo, Time-resolved LIBS of atomic and molecular carbon from coal in air, argon and helium, J. Anal. At. Spectrom. 27 (2012) 2066–2075, http://dx.doi.org/10.1039/C2JA30222E. [36] R.T. Wilkin, H.L. Barnes, S.L. Brantley, The size distribution of framboidal pyrite in modern sediments: an indicator of redox conditions, Geochim. Cosmochim. Acta 60 (1996) 3897–3912, http://dx.doi.org/10.1016/0016-7037(96)00209-8. [37] R.T. Wilkin, M.A. Arthur, W.E. Dean, History of water-column anoxia in the Black Sea indicated by pyrite framboid size distributions, Earth Planet. Sci. Lett. 148 (1997) 517–525, http://dx.doi.org/10.1016/S0012-821X(97)00053-8. [38] C. Guan, C. Zhou, W. Wang, B. Wan, X. Yuan, Z. Chen, Fluctuation of shelf basin redox conditions in the early Ediacaran: evidence from Lantian Formation black shales in

39

South China, Precambrian Res. 245 (2014) 1–12, http://dx.doi.org/10.1016/j. precamres.2014.01.003. [39] W. Chen, W. Zhou, P. Luo, H. Deng, Q. Li, R. Shan, M. Qi, Analysis of the shale gas reservoir in the Lower Silurian Longmaxi formation, Changxin 1 well, Southeast Sichuan Basin, China, Acta Petrol. Sin. 29 (2013) 1073–1086 (http://dx.doi.org/ 1000-0569/2013/029(03)-1073-86). [40] C. Aragón, J.A. Aguilera, Characterization of laser induced plasmas by optical emission spectroscopy: a review of experiments and methods, Spectrochim. Acta B 63 (2008) 893–916, http://dx.doi.org/10.1016/j.sab.2008.05.010.