Mechanical properties of organic matter in shales mapped at the nanometer scale

Mechanical properties of organic matter in shales mapped at the nanometer scale

Marine and Petroleum Geology 59 (2015) 294e304 Contents lists available at ScienceDirect Marine and Petroleum Geology journal homepage: www.elsevier...

6MB Sizes 1 Downloads 35 Views

Marine and Petroleum Geology 59 (2015) 294e304

Contents lists available at ScienceDirect

Marine and Petroleum Geology journal homepage: www.elsevier.com/locate/marpetgeo

Research paper

Mechanical properties of organic matter in shales mapped at the nanometer scale Moshe Eliyahu a, Simon Emmanuel a, *, Ruarri J. Day-Stirrat b, Calum I. Macaulay b a b

Institute of Earth Sciences, The Hebrew University of Jerusalem, Edmond J. Safra Campus, Givat Ram, Jerusalem, 91904, Israel Shell International Exploration and Production Inc., Shell Westhollow Technology Center, R1004B 3333 Highway 6, South Houston, TX, 77082, USA

a r t i c l e i n f o

a b s t r a c t

Article history: Received 13 March 2014 Received in revised form 29 July 2014 Accepted 8 September 2014 Available online 18 September 2014

The mechanical properties of organic matter strongly affect the way shales deform and fracture. However, the way organic matter responds to mechanical stresses is poorly understood, representing a critical obstacle to assessing oil and gas production in shale formations. Little is known about the mechanical properties of organic matter in fine grained rocks primarily because it often occupies tiny nanometerscale voids between the mineral grains which cannot be accessed using standard mechanical testing techniques. Here, we use a new atomic force microscopy technique (PeakForce QNM™) to map the mechanical properties of organic and inorganic components at the nanometer scale. We find that the method is able to distinguish between different phases such as pyrite, quartz, clays, and organic matter. Moreover, within the organic component Young's modulus values ranged from 0 to 25 GPa; in 3 different samples e all of which come from thermally mature Type II/III source rocks in the dry gas window e a modal value of 15e16 GPa was measured, with additional peaks measured at 10 GPa. In addition, the maps suggest that some porous organic macerals possess a soft core surrounded by a harder outer shell 50e100 nm thick. Thus, our results demonstrate that the method represents a powerful new petrographic tool with which to characterize the mechanical properties of organic-rich sedimentary rocks.

Keywords: Bitumen Kerogen Mudrocks Young's modulus

© 2014 Elsevier Ltd. All rights reserved.

1. Introduction One of the challenges in developing hydrocarbon reservoirs hosted by mudrock is the accurate prediction of the mechanical properties associated with heterogeneous geological material. Properties, such as the elastic moduli and yield strength, are critical in determining borehole stability and the response of the mudrock to fracturing techniques (van Oort et al., 1994; Mody, 1996; Fam et al., 2003; Dewhurst et al., 2011). Consequently, mechanical properties affect both the efficiency of fracturing methods and the flow of hydrocarbons to the wellbore. In addition, in source rocks containing high levels of total organic carbon (TOC), the elastic properties of organic components affect acoustic velocities and impact seismic expression. However, predicting the mechanical properties of mudrocks is a non-trivial task. Mudrocks often comprise a diverse assemblage of minerals, including clays, quartz,

* Corresponding author. E-mail addresses: (S. Emmanuel).

[email protected],

http://dx.doi.org/10.1016/j.marpetgeo.2014.09.007 0264-8172/© 2014 Elsevier Ltd. All rights reserved.

[email protected]

carbonates, and sulfides, as well as varying degrees of cementation and organic matter (OM) content. Further complicating characterization, individual mineral grains are often submicron in size so that mudrocks are in effect complex natural nano-composite materials (Ulm and Abousleiman, 2006), which exhibit a range of anisotropic textures that reflect the shape and orientation of grains. In mudrocks, uniaxial compressive strength is highly dependent on porosity, varying from around 250 MPa in low porosity (~1%) mudrocks to <10 MPa in high porosity (~35%) dry samples (e.g., Hoshino, 1993; Vernik et al., 1993; Lashkaripour, 2002). In addition to porosity, both small scale intergranular interactions and the mechanical properties e such as the elastic modulus e of the inorganic and organic components can also strongly influence rock characteristics (Benveniste, 1987; Sheng, 1990; Hudson, 1991; Eseme et al., 2007; Scaffaro et al., 2011; Emmanuel and DayStirrat, 2012). Although mechanical properties are reasonably well known for many of the minerals commonly present in mudrocks, the properties of organic matter are less well constrained. Further complicating matters, mudrocks typically include a variety of different organic components that undergo considerable changes upon burial: thermal maturation transforms kerogen

M. Eliyahu et al. / Marine and Petroleum Geology 59 (2015) 294e304

295

Table 1 Summary of sample characteristics. Sample

Clay mineralsa [%]

Silicate mineralsb [%]

Carbonate, pyritec [%]

TOCd [%]

Porositye [%]

Sample 1 Sample 2 Sample 3

16 18 33

48 39 49

36 43 18

4.56 0.52 3.01

10.9 4.7 10.9

a b c d e

Illite, Fe-chlorite (estimated % mass of mineral phases). Quartz, K-feldspar, albite (estimated % mass of mineral phases). Calcite, ankerite, siderite, pyrite (estimated % mass of mineral phases). Total organic content, (% of total mass). % of bulk volume.

Figure 1. (a) Schematic diagram of an atomic force microscope. (b) Schematic representation of a force distance curve derived using the PeakForce QNM™ imaging mode. Such a curve is derived for each pixel, and parameters including the reduced modulus, surface adhesion, and deformation can be calculated.

Figure 2. A 20 mm  20 mm region of Sample 1 scanned using different imaging methods. (a) SEM image (BSE mode); (b) AFM topographic mode; (c) AFM PeakForce error mode used to provide a pseudo-3D image of the surface; (d) Young's modulus map; green indicates organic matter (low stiffness), blue indicates clays, while light-blue and pink indicates quartz and calcite (high stiffness). Modulus values are cut-off at 100 GPa. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

296

M. Eliyahu et al. / Marine and Petroleum Geology 59 (2015) 294e304

into new hydrocarbon phases, a process that can change the composition and properties, such as density, of the residual kerogen (e.g., Ujiie, 1978; Okiongbo et al., 2005). However, as the organic matter in mudrocks often occupies the submicron-scale intergranular pore space, determining the mechanical properties requires the use of ultra-high resolution techniques. So far, most of the attempts to characterize the in situ mechanical properties of shale organic matter have relied on nanoindentation techniques (e.g., Zeszotarski et al., 2004; Ulm and Abousleiman, 2006; Bobko and Ulm, 2008; Ahmadov et al., 2009; Ahmadov, 2011; Kumar et al., 2012; Shukla, 2013; Zargari et al., 2013). While these studies have provided important constraints concerning the Young modulus of organic matter, nano-indentation has the disadvantage of being a destructive technique that provides micron scale e rather than true nanometer scale e resolution. In this study, we use a recently developed non-destructive technique, based on atomic force microscopy (AFM), to create high resolution quantitative moduli maps of the organic and inorganic components in selected shale samples. We explore the implications of our results for predicting the mechanical properties of shales in general, and discuss how the technique may best be used to characterize other organic-rich rocks. 2. Methodology 2.1. Sample preparation and characterization We selected three organic-rich shale samples for analysis. The samples are from a single well from an Upper Jurassic source rock in the continental US that was deposited in relatively shallow water in a restricted basin. Total organic content varied from 0.5 to 4.5 wt.% (Table 1), and thermal maturity in the well is high (vitrinite reflectance [Ro], ~2.1%). Geochemical analysis indicates dry gas production. Original kerogen types were mainly gas-prone Type III and some oil-and gas-prone Type II/III materials. Mineralogy in the samples was determined using XRD (Table 1), and the samples contain varying amounts of illite, Fe-chlorite, carbonate, pyrite and detrital quartz and feldspar. To prepare the samples for imaging, they were first polished by hand with a diamond impregnated cloth, and then milled using argon ion milling (Fischione Instruments SEM Mill model 1060; accelerating voltage 4 kV); this technique produces an extremely smooth surface free of artifacts e such as grain plucking e which are often associated with mechanical polishing of shales (Loucks et al., 2009; Emmanuel and Day-Stirrat, 2012). Scanning electron microscopy (SEM) was used to image the surface (Quanta 200FEG Environmental Scanning Electron Microscope, FEI; Sirion high resolution scanning electron microscope, FEI), and chemical element maps were obtained using energy dispersive spectroscopy (EDS; Silicon Drift Detectors X-Max 20 SDD, Oxford Instruments) operated at an accelerating voltage of 10 kV. 2.2. Nano-scale mechanical mapping PeakForce quantitative nano-mechanical mapping (or PeakForce QNM™) is a recent AFM mode derived from the PeakForce Tapping™ method, which oscillates the Z piezo at a rate far below the probe resonance (Su, 2010; Pittenger et al., 2014). The sample surface is measured by the deflection of the cantilever, and a force curve is produced for each image pixel; the peak force of each tap is used as a control signal to maintain a constant imaging force (Fig. 1a). The PeakForce Tapping™ mode modulates the Z piezo at about 2 kHz with a default peak force amplitude of 150 nm. Analysis of force curve data (Fig. 1b) is carried out during the scan, and the reduced elastic modulus (E*) can be derived using the

Figure 3. (a) Young's modulus map for Sample 1; the red horizontal line indicates the transect along which (b) modulus and (c) height values are compared. There is no apparent correlation between the 2 parameters. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

DerjaguineMullereToporov (DMT) model (Derjaguin et al., 1975; Trtik et al., 2012):

4 Finteraction ¼ E* 3

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Rðd  d0 Þ3 þ Fadh ;

(1)

where Finteraction is the tip e sample force, R is the tip radius, d0 is the surface rest position, (dd0) is the sample deformation, and Fadh is the adhesion force during the contact. The Young modulus of the sample, Es, is related to the reduced modulus by

E* ¼

2

1  y2s 1  ytip þ Es Etip

!1 ;

(2)

where ns is Poisson's ratio of the sample, ntip is Poisson's ratio of the probe, and Etip is the Young modulus of the tip material. In this

M. Eliyahu et al. / Marine and Petroleum Geology 59 (2015) 294e304

297

Figure 4. Comparison of EDS elemental mapping and Young's modulus map for Sample 1. (a) Oxygen; (b) aluminum; (c) silicon; (d) calcium; (e) carbon; (f) potassium; (g) sulfur; (h) iron; (i) Young's modulus.

study, the reported Young's modulus values were calculated according to a Poisson's ratio of 0.3. The effect of different ratios is discussed in Section 3.2. In addition to the extremely high spatial resolution that can be attained using the technique (on the order of several nm's), each scan in the PeakForce QNM™ mode can yield over 260,000 individual mechanical measurements, providing an excellent statistical characterization of heterogeneous samples. The technique has already been used to characterize biological materials (e.g., Sweers €n et al., 2011), and cement minerals et al., 2011), polymers (Scho (Trtik et al., 2012), although as far as we are aware this paper represents the first time this method has been applied to the characterization of rocks. To obtain maps of Young's modulus as well as topography, the ion-milled samples were scanned using a Veeco Multimode 8 AFM with a NanoScope V controller and NanoScope version 8.15 software. For each sample, suitable scan areas were identified using the AFM's optical microscope; surface steps caused by the ion milling and reflective pyrite grains served as reference points that facilitated the comparison of AFM and electron microscopy images obtained after the mechanical characterization.

As is the case in many AFM applications, choosing a cantilever with a suitable spring constant is crucial for accurate measurements in the PeakForce QNM™ mode; for a given cantilever spring constant, there is a limited range of moduli that can be measured. For shale samples e which often contain hard minerals such as quartz, apatite, and pyrite e we found that stiff diamond tips (DNISP; Bruker; spring constant 650 N m1) yielded an estimated uncertainty of less than 15% in the range 10 GPae70 GPa, consistent with the results reported for concrete by Trtik et al. (2012). Below 10 GPa, comparison with nano-indentation measurements indicated that the uncertainty is on the order of ±2 GPa. In our protocol, the deflection sensitivity of the probe was determined by ramping onto a stiff sapphire sample (E ¼ 345 N m1) after every laser alignment. A polycrystalline titanium standard was used to determine the tip radius (40 nm). This value was verified by scanning a highly-ordered pyrolitic graphite standard (HOPG-15 M, with a reported elastic modulus of 18 GPa); a constant peak force set point was applied producing a mean deformation of around 1e2 nm, which was suitable for fitting the DMT model. By adopting the calculated tip radius of curvature, a scanned area of 1 mm  1 mm with 512  512 measured points yielded an average modulus value

298

M. Eliyahu et al. / Marine and Petroleum Geology 59 (2015) 294e304

Figure 5. Comparison of three different AFM scan sizes. (a) Topographic image; (b) Young's modulus map; and (c) Young's modulus histogram for 50 mm  50 mm scan. (d) Topographic image; (e) Young's modulus map; and (f) Young's modulus histogram for 20 mm  20 mm region (Fig. 2). (h) Topographic image; (i) Young's modulus map; and (g) Young modulus's histogram for 5 mm  5 mm region. Modulus values are cut-off at 100 GPa.

17.7GPa ± 1.7 GPa. After the calibration procedure, the samples were scanned at 0.5 Hz while applying the lowest possible constant peak force set point while retaining tracking of surface. Every 2-3 scans, the diamond tip was cleaned by ramping on a standard gold sample and the radius of curvature of the tip was checked for wear. During the course of our measurements, wear caused the radius of curvature to increase from 40 nm to ~60 nm. Data analysis was carried out using the Nanoscope Analysis software package; smoothing algorithms were not used on the data reported here. 3. Results and discussion 3.1. Mechanical measurements in shale samples From BSE images of Sample 1 (Fig. 2a), different phases including pyrite framboids, calcite, clay minerals, and larger (5e10 mm) quartz grains can be identified. Organic matter and pore space appear dark grey or black, and the contrast between them is poor. Although the topographic channel in the AFM scans can be used to identify a number of individual pores (Fig. 2b), the streaks

left by ion milling make the identification of grains and phases difficult; interestingly, the topographic highs located on the pyrite framboids are a result of the growth of a secondary oxidized phase due to exposure to air. The PeakForce Error mode (Fig. 2c) is far less affected by milling lines so that this mode can be used to identify the outlines of individual grains, although differences between phases are not readily apparent. By contrast, the map of Young's modulus reveals the presence of 4 distinct phases: (i) a stiff minor mineral phase (>90 GPa); (ii) isolated grains with an average Young's modulus of 58 ± 9 (1s) GPa; (iii) an intergranular matrix with a mean value of 29 ± 1 (1s) GPa; and (iv) a softer intergranular phase (0e25 GPa) dispersed throughout the sample. Importantly, the modulus mode is not strongly affected by milling ridges and other topographic features, so that the overall effect of topography on modulus values is relatively minor (Fig. 3). Importantly, EDS elemental maps can be used to infer the mineralogy of the phases identified by mechanical mapping (Fig. 4). As expected, the stiffest phase corresponds to pyrite, which has a reported Young's modulus of 250e312 GPa (Mavko et al., 2009) that is beyond the reliable maximal range of the PeakForce-QNM™

M. Eliyahu et al. / Marine and Petroleum Geology 59 (2015) 294e304

299

Figure 6. SEM and AFM images from an additional region in Sample 1. (a) SEM image (BSE mode); (b) Young's modulus map of the same region; (c) high resolution AFM scan (topographic mode) of the region bounded by the dashed line in (a) and (b); (d) corresponding Young's modulus map of the bounded region. Modulus values are cut-off at 100 GPa.

method. Of the isolated grains, many are quartz with a mean modulus value of 63 ± 8 (1s) GPa, while others are calcite with an average value of 53 ± 6 (1s) GPa; both these values are lower than the typically reported ranges for quartz (77e96 GPa; Mavko et al.,

2009) and calcite (74e83 GPa; Mavko et al., 2009). Such differences may be due to the difficulties in accurately measuring the moduli of stiff mineral components; however, the elastic properties of both quartz and calcite are highly anisotropic, and modulus

Figure 7. AFM scans of Sample 2. (a) 12 mm  12 mm topographic map and (b) corresponding Young's modulus map; (c) high resolution topographic map (5 mm  5 mm) of the region in (a) and (b) bounded by the dashed line; (d) Young's modulus map of the same region. In (a) and (b), low stiffness specks appear on some of the grains and this is thought to be an artifact of the milling process caused by the re-deposition of organic matter. Modulus values are cut-off at 100 GPa.

300

M. Eliyahu et al. / Marine and Petroleum Geology 59 (2015) 294e304

Figure 8. A 20 mm  20 mm region in Sample 3 scanned using different imaging methods. (a) SEM image (BSE mode); (b) AFM topographic mode; (c) AFM PeakForce error mode used to provide pseudo-3D image of the surface; and (d) Young's modulus map. The green crystalline grain to the right of the image is in fact apatite, although is seems to have been contaminated with an organic coating which causes an anomalously low Young's modulus measurement. The slightly higher modulus values in the center of the apatite grain could indicate that some of the organic coating was been removed by the AFM tip during the scan. Modulus values are cut-off at 100 GPa. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

values can vary significantly depending on crystal orientation (Timms et al., 2010). In between the individual grains, the matrix e with a Young's modulus of 29 ± 1 (1s) GPa e appears to comprise aggregates of clay minerals (mainly illite and Fe-chlorite). The measured modulus falls within the range of 21e55 GPa predicted for clays in a number of previous studies (Tosaya, 1982; Castagna et al., 1985; Han et al., 1986; Katahara, 1996; Wang et al., 2001), although it is significantly higher than the value of 6.2 GPa reported for dickite (a polymorph of kaolinite) in a study that employed atomic force acoustic microscopy (Prasad et al., 2002). Finally, the softest phase we observe (0e25 GPa) corresponds to carbon-rich areas which make up the organic phase. Scans at different resolutions (Fig. 5) e ranging from 50 mm  50 mme5 mm  5 mm e show overall similar patterns and distributions of modulus values. At all scales, quartz grains and clay minerals exhibit values of around 60 GPa and 30 GPa, respectively. However, despite the overall similarity, additional detail is revealed in the highest resolution scan (Fig. 5g and h): although modulus values for organic matter remains in the range (0e25 GPa), two peaks in the distribution are apparent at around 6 GPa and 15 GPa, suggesting that there are at least two distinct types of organic matter in the shale. Comparison of the topographic data with the modulus map (Fig. 5g and h) suggests that the softer material comprises porous macerals, while the stiffer non-porous material is wedged between clay particles in the matrix. Mapping of different regions of the Sample 1 (Fig. 6) yielded similar results, confirming that the PeakForce QNM™ method produces consistent results. Moreover, analyses of Samples 2 and 3

reveal organic matter with overall similar features. In the high resolution scan of Sample 2 (Fig. 7), the organic matter is again porous, with void sizes ranging from microns to tens of nanometers; intragranular pores also appear to be lined with softer material. Sample 3 also exhibits overall similar moduli in the organic and mineral components (Fig. 8). Interestingly, an apatite crystal, shown in Figure 8, is assigned an anomalously low Young's modulus value (~20 GPa versus 120 GPa; Gilmore and Katz, 1982). Imaging with EDS indicates the presence of carbonaceous material (Fig. 9), possibly reflecting the presence of a thin organic coating on the apatite grain. Such a coating could have accumulated during milling, although the reason for preferential deposition on apatite is unclear. Thus, at least some mineral surfaces could be contaminated as a result of the sample preparation procedure, and this result demonstrates that EDS imaging is an important stage in quality control during the measurements. A comparison of the estimated probability density functions (PDF) shows a number of common features in the statistical properties of the moduli distributions associated with the different samples (Fig. 10). In all three samples, there is a strong peak at around 30e35 GPa corresponding to clay minerals; however, the varying proportions of the various components yields quite different overall mean values for the Young's moduli: 35.4 GPa for Sample 1; 56.6 GPa for Sample 2; and 41.7 GPa for Sample 3. However, it should be noted that the way in which these average values relate to the effective elastic modulus of the rock is not yet clear. In future work, the measured moduli of bulk samples could be compared with predicted values obtained from realistic 3D grainscale models that incorporate AFM mechanical data. Such a

M. Eliyahu et al. / Marine and Petroleum Geology 59 (2015) 294e304

301

Figure 9. Comparison of EDS elemental mapping and Young's modulus map for Sample 3. (a) Oxygen; (b) aluminum; (c) silicon; (d) calcium; (e) carbon; (f) phosphorous; (g) sulfur; (h) iron; (i) Young's modulus.

comparison could provide an important way forward in understanding the impact of nanometer scale heterogeneities on the behavior of shales at much larger scales in the subsurface. 3.2. Young's modulus of organic matter in shale Our results indicate that organic matter has a distribution of values in the range 0e25 GPa. As the organic matter in all the samples is a relatively minor phase, its signature is not always clear in the probability density functions associated with the entire 20 mm  20 mm scans. However, by focusing on regions of the scan containing organic matter, a number of important features are enhanced (Fig. 10). In Sample 1, the PDF of the cropped region shows two distinct peaks at 2 GPa and 5 GPa (Fig. 10b), which are lower than the modal values of 6 GPa and 15 GPa apparent in the same sample in Figure 4. In Sample 2 (Fig. 10d), two peaks are also seen (10 GPa and 16 GPa), while in Sample 3 (Fig. 10f), two peaks are measured at 8 GPa and 15 GPa; in that sample, an additional minor peak also appears at ~5 GPa. Thus, despite the variability between samples and even within different regions of the same sample, a stiff organic element is present in all samples at around 15 GPa,

which is followed by weaker peaks with lower values. While it is possible that the stiffer material could correspond to kerogen, with the softer regions corresponding to bituminous material, such an interpretation may be an over-simplification as pyrobitumen has also been observed in these high maturity samples. Moreover, Kumar (2012) found values of 15e16 GPa for high maturity kerogen and values of 6e9 GPa for low maturity kerogen. However, in the same study, highly porous kerogen exhibited Young's modulus values as low as 2 GPa, which is the same as that reported by Zargari et al. (2013) for extruded bitumen. At present, one of the main limitations of using AFM based methods to determine the exact values of Young's moduli in shales is that little is known about Poisson's ratios in the organic component (e.g., Ahmadov, 2011). Although we chose a representative Poisson's ratio of 0.3 for organic matter, the expected range (0.05e0.45) introduces an uncertainty of up to 12% in the calculated modulus values (Fig. 11). Clearly, better constraints on Poisson's ratios for organic matter will help to reduce such errors in future studies. In addition to shedding light on the ranges of elastic moduli in the organic matter, our analyses indicate that the mechanical properties within organic phases may vary systematically within

Figure 10. Probability density functions (PDF's) of Young's moduli and corresponding modulus maps in different samples. (a)e(b) Sample 1; (c)e(d) Sample 2; and (e)e(f) Sample 3. For all samples, the PDF of the region outlined in red is marked referred to as the cropped region. PDF's were calculated using the kernel density estimation method (for details see Vermeesch (2012)). Green arrows indicate major peaks associated with the organic matter in each scan. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Figure 11. Dependence of Young's modulus on Poisson's ratio for different components in the shale. The dashed line indicates the assumed value for the analyzed surfaces. The moduli at a Poisson's ratio of 0.3 are representative of values measured in the present study.

individual macerals. Close inspection of some regions suggests that some macerals may have a harder outer shell surrounding a softer core (Fig. 12), similar to the structure inferred by Curtis et al. (2012) from BSE images of organic matter in shales. This exterior shell appears to be approximately 50e100 nm thick, although this may be an artifact of spatial averaging caused by the relatively large radius of the diamond tip. However, the internal structure of the maceral also seems to contain both hard and soft regions, and there is a good correspondence between stiffer regions in the AFM maps and brighter areas in the BSE images that indicate higher atomic number. While the current study does not provide a complete description of the mechanical properties of organic matter in shales, the large datasets provided by the AFM scans represent an important step towards describing the mechanical properties of organic matter in shales using a probabilistic, rather than deterministic, approach. In future studies, the mechanical properties of different organic components could be determined by analyzing shale samples that have been treated to remove the bituminous component. Further description of the structure of the organic

M. Eliyahu et al. / Marine and Petroleum Geology 59 (2015) 294e304

303

Figure 12. Comparison of SEM back-scattered electron image and modulus map of an organic maceral in Sample 1. Note that the maceral appears to be surrounded by a hard shell; its internal structure also shows regions of different stiffness.

material (soft core e versus hard shell) could also shed important light on the process of petroleum generation. 4. Concluding remarks In this paper, we used a novel AFM technique to map for the first time the mechanical properties of shales with nano-scale resolution. We show that organic matter in organic-rich shales possesses Young's modulus values in the range 0e25 GPa. Moreover, in the 3 samples tested here the organic matter exhibited an apparent bimodal distribution within this range, which could correspond to a stiffer kerogen component and a more compliant bituminous component. Furthermore, the structure of the organic macerals suggests that they could consist of a softer core surrounded by a harder shell. Thus, the high resolution provided by the PeakForce QNM™ technique represents a powerful new petrographic tool for relating structure and composition to the mechanical properties geological materials at the nanometer-scale. Clearly, the results shown here represent only a limited characterization of organic matter in shales, and much remains to be learned concerning the link between different types of organic matter, maturation, and the evolution of mechanical properties. In addition, it is still unclear how the structure and distribution of organic matter revealed by high resolution imaging can be incorporated into models used to describe mechanical properties of rocks at the continuum scale. Some of these limitations will be addressed in future work that will focus on characterizing organic matter in shales exhibiting maturation gradients, as well as shales that have been subjected to pyrolysis experiments. One important conclusion of the current work is that the elastic modulus of organic matter is heterogeneous at the nanometerscale. As a result, methods with poor spatial resolution will never be able to fully characterize organic matter in fine grained sedimentary rocks, such as shales. Furthermore, our results indicate that, for one maturity window at least, it may be more appropriate to assign a distribution of values to represent different organic components, rather than attempting to assign specific singular values. Relating these different distributions to the degree of maturation of the organic matter could provide an important window into the process of oil and gas generation, and should help improve the prediction of mechanical properties in shales. Acknowledgments We thank three anonymous reviewers for their constructive comments. This work was supported by the Israeli Ministry of National Infrastructures, Energy, and Water Resources (213-170123-10643).

References Ahmadov, R.S.O., 2011. Microtextural, Elastic and Transport Properties of Source Rocks (Ph.D. thesis). Stanford University. Ahmadov, R., Vanorio, T., Mavko, G., 2009. Confocal laser scanning and atomic-force microscopy in estimation of elastic properties of the organic-rich Bazhenov Formation. Lead. Edge 28, 18e23. Benveniste, Y., 1987. A new approach to the application of Mori-Tanaka's theory in composite materials. Mech. Mater. 6, 147e157. Bobko, C., Ulm, F.J., 2008. The nano-mechanical morphology of shale. Mech. Mater. 40, 318e337. Castagna, J.P., Batzle, M.L., Eastwood, R.L., 1985. Relationships between compressional-wave and shear-wave velocities in clastic silicate rocks. Geophysics 50, 571e581. Curtis, M.E., Cardott, B.J., Sondergeld, C.H., Rai, C.S., 2012. Development of organic porosity in the woodford shale with increasing thermal maturity. Int. J. Coal Geol. 103, 26e31. Derjaguin, B., Muller, V., Toporov, Y., 1975. Effect of contact deformations on adhesion of particles. J. Colloid Interface Sci. 53, 314e326. Dewhurst, D.N., Siggins, A.F., Sarout, J., Raven, M.D., Nordgard-Bolas, H.M., 2011. Geomechanical and ultrasonic characterization of a Norwegian Sea shale. Geophysics 76, WA101eWA111. Emmanuel, S., Day-Stirrat, R.J., 2012. A framework for quantifying size dependent deformation of nano-scale pores in mudrocks. J. Appl. Geophys. 86, 29e35. Eseme, E., Urai, J.L., Kroos, B.M., Littke, R., 2007. Review of mechanical properties of oil shales: implications for exploitation and basin modelling. Oil Shale 24, 159e174. Fam, M.A., Dusseault, M.B., Fooks, J.C., 2003. Drilling in mudrocks: rock behavior issues. J. Petrol. Sci. Eng. 38, 155e166. Gilmore, R.S., Katz, J.L., 1982. Elastic properties of apatites. J. Mater. Sci. 17, 1131e1141. Han, D.-H., Nur, A., Morgan, D., 1986. Effects of porosity and clay content on wave velocities in sandstones. Geophysics 51, 2093e2107. Hoshino, K., 1993. Geological evaluation from the soil to the rock: mechanism of lithification and change of mechanical properties. In: Proceedings of the International Symposium on Geotechnical Engineering of Hard Soils-soft Rocks, Athens, Greece, Balkema, Rotterdam, vol. 2, pp. 121e128. Hudson, J.A., 1991. Overall properties of heterogeneous material. Geophys. J. Int. 107, 505e511. Katahara, K.W., 1996. Clay mineral elastic properties. In: Society of Exploration Geophysicists Annual Meeting, Denver. Kumar, V., 2012. Geomechanical Characterization of Shale Using Nano-indentation (Diss). University of Oklahoma. Kumar, V., Sondergeld, C., Rai, C., 2012. Nano to macro mechanical characterization of shale. In: SPE Annual Technical Conference and Exhibition. Lashkaripour, G.R., 2002. Predicting mechanical properties of mudrock from index parameters. Bull. Eng. Geol. Environ. 61, 73e77. Loucks, R.G., Reed, R.M., Ruppel, S.C., Jarvie, D.M., 2009. Morphology, genesis, and distribution of nanometer-scale pores in siliceous mudstones of the Mississippian Barnett Shale. J. Sediment. Res. 79, 848e861. Mavko, G., Mukerji, T., Dvorkin, J., 2009. The Rock Physics Handbook. Cambridge University Press, p. 511. Mody, F.K., 1996. Borehole stability in shales: a scientific and practical approach to improving water-base mud performance. In: AADE-Houston Chapter Drilling Fluids Conference, Houston. Okiongbo, K.S., Aplin, A.C., Larter, S.R., 2005. Changes in type II kerogen density as a function of maturity: evidence from the Kimmeridge Clay Formation. Energy Fuels 19, 2495e2499. Pittenger, B., Erina, N., Su, C., 2014. Mechanical Property Mapping at the Nanoscale Using PeakForce QNM Scanning Probe Technique. In: Nanomechanical Analysis of High Performance Materials. Springer, Netherlands, pp. 31e51. Prasad, M., Kopycinska, M., Rabe, U., Arnold, W., 2002. Measurement of Young's modulus of clay minerals using atomic force acoustic microscopy. Geophys. Res. Lett. 29, 1e4.

304

M. Eliyahu et al. / Marine and Petroleum Geology 59 (2015) 294e304

Scaffaro, R., Botta, L., Ceraulo, M., La Mantia, F.P., 2011. Effect of kind and content of organo-modified clay on properties of pet nanocomposites. J. Appl. Polym. Sci. 122, 384e392. €n, P., Bagdi, K., Moln nszky, B., Vancso, G.J., 2011. QuanScho ar, K., Markus, P., Puka titative mapping of elastic moduli at the nanoscale in phase separated polyurethanes by AFM. Eur. Polym. J. 47, 692e698. Sheng, P., 1990. Effective-medium theory of sedimentary rocks. Phys. Rev. B 41, 4507e4512. Shukla, P., 2013. Nanoindentation Studies on Shales (Ph.D disseration). University of Oklahoma, p. 2013. Su, C., 2010. Mapping quantitative mechanical properties at molecular scale using peak force tapping AFM. Microsc. Microanal. 16 (Suppl. 1), 364e365. Sweers, K., van der Werf, K., Bennink, M., Subramaniam, V., 2011. Nanomechanical properties of a-synuclein amyloid fibrils: a comparative study by nanoindentation, harmonic force microscopy, and PeakForce QNM. Nanoscale Res. Lett. 6, 270. Timms, N.E., Healy, D., Reyes-Montes, J.M., Collins, D.S., Prior, D.J., Young, R.P., 2010. Effects of crystallographic anisotropy on fracture development and acoustic emission in quartz. J. Geophys. Res. 115 (B7). Tosaya, C.A., 1982. Acoustical Properties of Clay-bearing Rocks (Ph.D. dissertation). Stanford University, Stanford.

Trtik, P., Kaufmann, J., Volz, U., 2012. On the use of peak-force tapping atomic force microscopy for quantification of the local elastic modulus in hardened cement paste. Cem. Concr. Res. 42, 215e221. Ujiie, Y., 1978. Kerogen maturation and petroleum genesis. Nature 272, 438e439. Ulm, F.J., Abousleiman, Y., 2006. The nanogranular nature of shale. Acta Geotech. 1, 77e88. van Oort, E., Hale, A., Mody, F., Roy, S., 1994. Critical parameters in modelling the chemical aspects of borehole stability in shales and in designing improved water based shale drilling fluids. In: 69th SPE Annual Technical Conference and Exhibition, New Orleans. Vermeesch, P., 2012. On the visualisation of detrital age distributions. Chem. Geol. 190e194. Vernik, L., Bruno, M., Bovberg, C., 1993. Empirical relations between compressive strength and porosity of siliciclastic rocks. Int. J. Rock Mech. Min. 30, 677e680. Wang, Z., Wang, H., Cates, M.E., 2001. Effective elastic properties of solid clays. Geophysics 66, 428e440. Zargari, S., Prasad, M., Mba, K.C., Mattson, E.D., 2013. Organic maturity, elastic properties, and textural characteristics of self resourcing reservoirs. Geophysics 78, D223eD235. Zeszotarski, J.C., Chromik, R.R., Vinci, R.P., Messmer, M.C., Michels, R., Larsen, J.W., 2004. Imaging and mechanical property measurements of kerogen via nanoindentation. Geochim. Cosmochim. Acta 68, 4113e4119.