Experimental Mechanics DOI 10.

1007/s11340-010-9373-z

Determination of Mechanical Properties of Sand Grains by Nanoindentation

N.P. Daphalapurkar & F. Wang & B. Fu & H. Lu & R. Komanduri

Received: 3 August 2009 / Accepted: 24 May 2010 # Society for Experimental Mechanics 2010

Abstract Determination of the mechanical properties of individual sand grains by conventional material testing methods at the macroscale is somewhat difficult due to the sizes of the individual sand particles (a few m to mm). In this paper, we used the nanoindentation technique with a Berkovich tip to measure the Youngs modulus, hardness, and fracture toughness. An inverse problem solving approach was adopted to determine the stress-strain relationship of sand at the granular level using the finite element method. A cube-corner indenter tip was used to generate radial cracks, the lengths of which were used to determine the fracture toughness. Scatter in the data was observed, as is common with most brittle materials. In order to consider the overall mechanical behavior of the sand grains, statistical analysis of the mechanical properties data (including the variability in the properties) was conducted
N.P. Daphalapurkar Department of Mechanical Engineering, Johns Hopkins University, Baltimore, MD 21218, USA e-mail: nitin@jhu.edu F. Wang : H. Lu (SEM member) Department of Mechanical Engineering, The University of Texas at Dallas, Richardson, TX 75080, USA H. Lu e-mail: hongbing.lu@utdallas.edu B. Fu : R. Komanduri (*) School of Mechanical and Aerospace Engineering, Oklahoma State University, 218 Engineering North, Stillwater, OK 74078, USA e-mail: ranga.komanduri@okstate.edu B. Fu e-mail: boshen.fu@okstate.edu

using the Weibull distribution function. This data can be used in the mesoscale simulations. Keywords Sand . Mechanical properties . Nanoindentation . Youngs modulus . Hardness . Stress-strain . Fracture toughness . Particulate mechanics

Introduction Granular materials, such as sand, are conglomerates of discrete particles held together (but not bonded) with significant void space (3565%). They are unique in that they behave in some respects similar to the other familiar forms of matter, namely, solids, liquids, and gases and in other respects in a dissimilar form. For example, they pack like solids but flow like liquids. Like liquids, they take the shape of the container but unlike them they can adopt to a variety of shapes when they are free standing. Similarly, like gases, they are made of discrete particles with negligible cohesive forces between them. Like solid, they can support load, but unlike a solid, they hardly support any tensile load. In view of their unique behavior, some consider granular materials as the fifth state of matter, alongside, solids, liquids, gases, and plasma. They cannot strictly be modeled as a continuum, yet it is done by considering the movement of particles in the void space under load, akin to the deformation and flow in solids. Sand is formed largely by erosion and disintegration of larger rocks into particles by natural forces, such as wind, pressure, water, ice, friction, and heat. Over millions of years, such processes have led to the formation of sand of various grain sizes ranging from a fraction of a micrometer to several millimeters. Investigation of the mechanical behavior of sand from granular (mesoscale) to macro

Exp Mech

(continuum) scale can contribute towards a fundamental understanding of the underlying mechanisms of deformation, flow, and fracture of granular materials under load and ultimately realize the underlying science-base for large scale phenomena, such as the motion of sand particles by an impacting solid, and the size (typically a few mm) of the influential shearing zones in tectonic plates in sliding during earthquakes and landslides. Mechanical properties, such as the Youngs modulus, hardness, fracture toughness, and stress-strain relations of sand grains, are of particular importance since their individual behavior dictates the overall behavior of sand at the macroscale. Moreover, the mechanical properties of the grains can vary with the mineral composition, size, defect structure, and crystal orientation of sand. Grain size plays an important role due to the fact that larger grains would contain higher number of defects in the form of voids, ridges, and microcracks which could adversely effect the yield properties. Hence, mechanical properties of larger grains cannot be directly used to imply the behavior of finer grains. As far as larger (on the order of a centimeter or more) sand particles (rocks) are concerned, measurement of mechanical properties can be carried out using tensile/ compression testing machine. However, difficulties arise in the preparation of the sample and/or holding smaller grains in the material testing machine prior to subjecting them to compression or pulled in tension for the measurements of mechanical properties. For the past two decades, nanoindentation has been used for the measurement of mechanical properties of different types of materials (e.g. elastic-plastic, viscoelastic). The instrumented nanoindentation is widely accepted as a standardized testing method for the characterization of mechanical properties for elastic-plastic materials. The resolution in such a test can reach a fraction of a nanometer in displacement and N in load. The added advantage of such testing is that in some cases, wherein the materials exhibit length-scale dependent behavior, it becomes possible to characterize such materials and extract the moduli and hardness values at different depths of indentation. The material properties extracted will be a function of the maximum indentation depth reached or the deformation zone beneath the indenter tip. Such an ease in the characterization of length-scale dependent material properties using nanoindentation is in contrast with the tedious procedure involved in using conventional tensile testing on small samples. Thus, extracting the material properties by indentation at depths < 100 nm would essentially provide mechanical properties at nanoscale. As a result, nanoindentation technique is being widely used to extract mechanical properties of very small volume of materials, such as thin films, wires, and coatings as well as for MEMS and NEMS components.

In this investigation, we used the nanoindentation technique to determine the force-displacement relation and extracted the mechanical properties, such as the Youngs modulus, hardness, fracture toughness, and stress-strain relationship of sand granules at microscale. We conducted nanoindentations on several grains to assess the variance in the mechanical properties. Furthermore, by adopting an inverse methodology approach, i.e., by using the experimental load-depth information in combination with Finite Element (FE) numerical simulations, we predicted the stress-strain response. By using the cube-corner of an indenter tip to induce cracks in the sand grain, we characterized the fracture behavior of sand particles.

Brief Review of Literature It is well known that dense/highly compacted sand possess high compressive strength and high energy absorption capacity. Sand is often used to provide ballistic protection for military and national security applications, in sandbags to prevent soil erosion at the banks and act as temporary dams against floods, and as a construction material in civil engineering structures. Consequently, the mechanical behavior of granular or particulate materials, such as sand, has been an area of increasing interest especially in the past decade or so. This is partially due to the emergence of new technological instruments, such as nanoindentation, atomic force microscopy (AFM), X-ray microtomography as well as the availability of rapidly increasing computational power that has facilitated in addressing this complex problem. However, this field yet remains to be relatively insufficiently understood. Considerable research had been reported in the literature on the mechanical behavior of sand, especially in the soil mechanics area. Poorooshasb et al. [1] and Lade and Duncan [2] conducted triaxial compression tests on sand to determine the nature of deformation and attempted to describe it using a constitutive law. Triaxial compression tests were also carried out by Arthur and Menzies [3] to investigate anisotropy, softening and preshearing effects by Lade and Duncan [4], and shear band formation by Wang and Lade [5]. Additionally, simple shear tests were conducted by Haythornthwaite [6] to determine shear locus of sand. However, for smaller systems at the granular level, the stresses are considered to propagate in a manner described by the wave-like (hyperbolic) equations [79] rather than by the elliptic equations of static elasticity. Also, for small-scale systems, the localization of deformation [10], such as the formation of force-chains is more pronounced and deformation mechanisms at meso/micro scales need to be considered [11].

Exp Mech

Most of the literature on characterization of mechanical properties of sand or rocks has focused on either compressive behavior of sand as bulk or macroscale properties of rocks by pulling them in a tensile testing machine. For geomechanics applications and simulations, it is necessary to know the mechanical properties of individual grains. There have been many efforts on characterizing mechanical behavior, including, failure behavior of rocks. For example, Liao et al. [12] conducted direct tensile tests on transversely isotropic cylindrical argillite specimens and reported the stress-strain relationships. They determined the material properties and found significant variations depending on the amount and orientation of the micro-fissures. However, the data for larger sizes cannot be extended directly for individual sand grains, mainly due to the difference in the length scales. The mechanisms of deformation and fracture in the individual sand grains can be different from their larger counterparts. Ulm and Abousleiman [13] conducted nanoindentation measurements on shales, a type of sedimentary rock, and concluded that shales are nanogranular composite materials with their mechanical properties dictated by particle-to-particle contact and packing density. Nanoindentation provides an effective technique for the measurement of local material response in terms of hardness and Youngs modulus at micro- and nano- length scales (see for e.g. Refs. [1417]). Methods for measuring the elastic-plastic properties, such as dynamic Youngs modulus and hardness have been well established [18, 19]. The wide acceptance of nanoindentation technique stems from the improvements incorporated in the technique. For example, the elimination of the need for direct observation of the residual indentation impressions. Instead, the unloading curve is used to extract the elastic properties of an elastic-plastic material, namely, elastic modulus and hardness by calculating the area of indent impression from the loading/unloading curves. This method is very well established and has been implemented in commercially instrumented nanoindenters for use on elastic-plastic mateFig. 1 (a) SEM image of sand grains, (b) (color online) Magnified optical image of polished sand grains in an epoxy matrix

rials. Inverse methodologies, using numerical techniques, in combination with experimental results have been developed effectively to measure the material properties and also the stress-strain relationships [20, 21]. In addition to the measurement of the elastic modulus and hardness, nanoindentation has also been used to determine the fracture toughness of brittle materials [22, 23]. This procedure was found to be adequate for estimating the fracture toughness in homogeneous materials. In this investigation, we conducted nanoindentation tests using a Berkovich indenter tip on several (250) individual sand grains of 1 mm size and recorded the loaddisplacement data. Youngs modulus and hardness were obtained directly from the MTS TestWorks software output based on the analysis of unloading segment using the method developed by Oliver and Pharr [19]. A representative stress-strain curve was obtained from which the corresponding average modulus was obtained. This was used for solving the inverse problem to determine the stress-strain curve using FEM. Nanoindents were also made using a cube-corner indenter to generate cracks for the estimation of fracture toughness. In all the cases, the indent impressions were obtained using MTS NanoVision setup. The lengths of the cracks developed were used to predict the range of fracture toughness values for the sand grains. Statistical analysis was conducted to determine the variability in the mechanical properties of sand grains.

Nanoindentation Measurements A sample of sand, available near the lakes in Stillwater, OK was used in this investigation. The sand was washed and subsequently dried in an oven at 55C. It maybe noted that the sand sample used for nanoindentation has widely distributed sizes. Figure 1(a) shows the SEM image of sand grains. They exhibit different shapes, sizes, contrast, and porosity under an optical microscope. Additionally, an EDS

Exp Mech

spectrum was obtained in the SEM showing the presence of Si, Al, Ca, K and Fe in the sand sample. The sand particle size distributions express the percentage (by mass) of individual size ranges [24]. Multidirectional shaker was used to determine the size distribution of the sand grains. The sieves use metal wire cloth with an ASTME-11 standard sieve series. To separate different sizes of the sand, the sand agglomerate was vibrated through a series of progressively smaller sieves that were stacked one on top of another. After 10 min of shaking, the sand that falls through a mesh was given the designation of passed weight and the sand that remains on top of that mesh was designated as remaining weight. The grain size distribution used in this investigation was 0.1 to 2 mm with an average grain size of 1 mm. The sand grains were embedded in an epoxy matrix and mounted in a sample holder. The samples were cured in an oven resulting in a composite of sand grains in a hardened epoxy matrix. The sand grains exhibited different colors due to differences in the density, crystal orientation, defect structure, and mineral content. No particular analysis was carried out to determine the type of sand within this sample size, as the main objective of this investigation was to determine the mechanical properties of individual grains by nanoindentation. The samples were polished using an alumina abrasive powder (from Buehler Inc.) in a water slurry. To obtain a smooth surface suitable for nanoindentation, the finest abrasive size used in the final polishing was 50 nm. Figure 1(b) shows an optical image of the polished sand surface with grains oriented in different directions. Within the sand grains, defects such as small pits and ridges were observed. An MTS Nano Indenter XP system was used for nanoindentation measurements. This indenter can reach a maximum indentation depth of 500 m and a maximum load of 500 mN. The displacement and load resolutions are 0.2 nm and 50 nN, respectively. Both Berkovich and cubecorner indenter tips, made of single crystal diamond, were used in this investigation. Nanoindentations were made on flat, polished sand grain surfaces under a constant rate of loading. The applied load on the indenter tip was increased until it reached a user-defined value. At this point, the load is held constant for a period or removed. Since, no information is extracted from the hold segment, the holding time in our experiments was zero. The load-displacement curves obtained thus are characteristic of that particular sand grain. Analysis is carried out on the load-displacement output to determine the mechanical properties of the sand grains based on contact mechanics of nanoindentation. Hardness (H) is obtained using, H Pmax Ac 1

where Pmax is the maximum indentation force and Ac is the contact area corresponding to the contact depth (hc) at maximum load, and is calculated based on the tip area function. The reduced modulus Er is obtained using 1 1 n2 1 n2 s i ; Er Es Ei 2

where Es and s are the Youngs modulus and Poissons ratio of the specimen, respectively, while Ei and i are the Youngs modulus and Poissons ratio of the indenter tip (made of diamond), respectively. Due to finite stiffness of the indenter tip, its modulus has to be considered in the expression for the calculation of specimen modulus from the contact stiffness. The contact stiffness (S) is calculated from the slope of the initial unloading curve. Thus, S p dP 2 p Er Ac hc dh p 3

Equations (2) and (3) along with the known values of the area function, slope of the unloading curve, and modulus and Poissons ratio values for the indenter tip can be used to determine the elastic modulus for a specimen corresponding to its Poissons ratio.

Results and Discussion Youngs Modulus and Hardness For the measurement of elastic modulus and hardness, a given force was applied by the nanoindenter using a Berkovich tip on the sand grain. Figure 2(a) shows an inverted image (3D) and a typical nanoindentation residual impression obtained using the NanoVision, a module of the MTS Nanoindenter system. The inverted image enables the determination of the depth of the indent more accurately and delineates its topographical features. Figure 2(b) shows a typical nanoindentation load-displacement curve. Nanoindentations were carried out under load-control, in which the indenter-tip was pressed onto the sample and the load on it was increased linearly. The initial part of the loaddisplacement curve, characterized by an increase in the load from zero to maximum, is termed the loading segment. Once the user-defined maximum load was reached, the load was removed, again in a controlled manner (linearly) and is termed, the unloading segment. It can be seen that no cracks were formed when indented with a Berkovich indenter-tip. Hence, the continuum approximation can be applied and equations (1) to (3) can be used to determine the modulus and hardness of the samples. To determine whether the sand grains are isotropic or not, nanoindentation tests were carried out on a polished, relatively large sand grain (size 1.5 mm) in three

Exp Mech Fig. 2 (color online) Nanoindentation test on sand grain using Berkovich indenter, (a) Residual indent impression and 3D inverse image; (b) Loaddisplacement curve

orthogonal directions on the cube corner of a sand grain (1 mm0.8 mm0.7 mm) embedded in an epoxy matrix. The three faces were designated as: face XY, face YZ, and face ZX of the sand grain. Nanoindentation tests were made on all three faces. On each face, 8 to 10 tests were carried out and the distances between the neighboring test locations were kept apart by at least 50 m. The Youngs modulus values for face XY, YZ, and ZX were 43.63.85 GPa, 45.74.30 GPa, and 44.085.21 GPa, respectively. These values are very close to each other, indicating that the sand grain can be modeled as an isotropic material. Youngs modulus values, however, were found to be lower, compared to the modulus of quartz, which has modulus values of 79 GPa, 79 GPa, and 103 GPa when indentations were made on an X-cut, Y-cut, and Z-cut surfaces. This could be induced by different material constituents in the sand grain. In order to investigate the variability in the properties over a single grain (size 0.7 mm), nanoindentations were made on a sand grain with the same maximum load at different locations. Figure 3(a) and (b) show Weibull plots for the data obtained on Youngs modulus and hardness, respectively. The median value (corresponding to P50 probability) for Youngs modulus and hardness are 66.9 GPa and 10.74 GPa, respectively. The variation in

the measured modulus and hardness at different locations is attributed here to the heterogeneity in the sand grain. Once an estimate was obtained for a single grain (of representative size 0.7 mm), nanoindentations were carried out on 250 different grains with two indentations per grain. The results from the two tests were averaged and taken as the mean value for that grain. Youngs modulus and hardness values were obtained using the slope of the unloading curve and the values of load and contact area at maximum indentation depth, respectively. These were direct outputs from the nanoindentation software. The Weibull distributions of the modulus and the hardness values were plotted in [Fig. 4(a) and (b)], respectively. Using Weibull distribution function, we determined the median value (corresponding to P50 value) for Youngs modulus of the sand to be 90.1 GPa (range 41.4 to 115.8 GPa) and hardness to be 10.7 GPa (range 5.4 to 13.7 GPa). Scatter in the data is attributed to different types of sand grains due primarily to the variations in the material constituents, defect structure, and crystal orientations. Stress-Strain Response Solution for the indentation problem has been well established [2530]. Inverse methodologies, which use

Exp Mech

Fig. 3 (a) Weibull plot of Youngs modulus from nanoindentations on a single sand grain, (b) Weibull plot of hardness from nanoindentations on a single sand grain

experimentation in combination with numerical techniques to aid in characterizing the material properties, are used in situations where it is difficult to extract mechanical properties using analytical solutions alone due to nonlinearities or complexities in the material, geometry and loading conditions. In such cases, it is easier to simulate the equivalent model using a numerical method, such as Finite Element Method (FEM). Initially, we can start with certain assumed values of material parameters for e.g. plastic flow. The values are adjusted so that numerical results obtained can be comparable with the experimental values, for e.g. by comparing the load-displacement curves, in the case of nanoindentation. It maybe noted that FEM has been applied successfully in the simulation of nanoindentation problem [3140].

For determining the stress-strain relationship of granular sand material, nanoindentation was modeled using FEM. To predict the elastic-plastic properties, von Mises yield criterion was used along with isotropic hardening to simulate the deformation characteristics of a sand grain. The plastic behavior under compression was assumed to satisfy the RambergOsgood relationship between the true stress and true strain as  s 1=n " for s ! s y 4 K where n is the work hardening exponent, K is the reference stress value, and is the equivalent strain, p " "ij "ij "y "p , where y is the yield strain and p is the plastic q strain. The von Mises equivalent stress is given as, s 3 s ij s ij . 2

Fig. 4 (a) Weibull plot for Youngs modulus from nanoindentations on different sand grains, (b) Weibull plot of hardness from nanoindentations on different sand grains

Exp Mech

Fig. 5 Comparison of load-displacement relationship curves from nanoindentation and FE simulation

Within the elastic limit, s E "; for s sy 5

where E is the Youngs modulus. ABAQUS V6.8-4 standard [41] software program was used to perform the calculations, assuming finite deformation characteristics. The von Mises yield criteria can be written as s 1 s 2 2 s 2 s 3 2 s 3 s 1 2 2s 2 y 6

where, 1, 2, and 3 are the principal stresses and y is the yield strength measured in a uniaxial tension test. Finite Element (FE) method was used in 3D simulations of nanoindentation. The Berkovich indenter was simulated based on its three sided pyramidal geometry. Because of the Berkovich indenters axisymmetric geometry, only one sixth of the entire model was used in this simulation to reduce the computational time. The displacement history

from the experiment was given as input for the FE analysis. Poissons ratio of 0.17 and Youngs Modulus of 75 GPa were used in this simulation. The output of the FE analysis was the resulting reaction force or load. This numerical load was plotted versus the displacement into the surface to give load-displacement curve from the simulation. The mesh size selected was tested for convergence of the loaddisplacement curve. The numerical values of the loaddisplacement curve were compared with the experimental values and the measure of fit was carried out by minimizing the least squares correlation coefficient. In order to obtain a better fit, the initially assumed values were so adjusted as to minimize the least squares correlation coefficient. Thus, the material parameters were adjusted until a good agreement was reached between experimental and numerical data. The best-fit parameters were then used to determine the stressstrain relationship for the sand grain sample. It should be noted that using this approach, the maximum strain up to which the stress-strain curve is valid, is limited by the strain produced by the nanoindentation test [42]. No cracks were observed by examining the indent impressions (for the Berkovich indenter tip) obtained from NanoVision. Thus, the continuum plasticity material model was justified. As stated earlier, we used an inverse problem solving approach to determine the stressstrain relationship of sand at granular level by correlating the FEM simulated nanoindentation load-displacement data with the measured results. Figure 5 shows the fit obtained using the FEM simulation (obtained using ABAQUSStandard) to the representative experimental data of P50. In this simulation, the tip of the Berkovich indenter was assumed to be perfectly sharp but in practice, the indenter has a tip with a finite radius of tens to a few hundred nanometers. The perfectly sharp tip model in the simu-

Fig. 6 (a) Compressive strain distribution along radial direction of the model, (b) Stress-strain curve (in compression) from FE simulation

Exp Mech Fig. 7 (color online) Surface profile showing cracks generated after nanoindentation using cube corner tip, (a) at maximum load of 80 mN, (b) at maximum load of 70 mN

lations will cause the stress near the tip to be much higher. In [Fig. 6(a)], the compressive strain distribution along the radial direction, starting from the tip contact area is shown. At a depth less than 200 nm, the strains are unreasonably high (90% compressive strain) at the tip due to singularity. The strains in the region at a radial distance of more than

200 nm from the tip, decrease slowly from 0.6. Since an artificially sharp Berkovich indenter model was used in the simulation, it will result in unreasonable values near the tip. We consider a strain of 0.6, which is in the region 200 nm away from the tip, as the highest strain. The predicted stress-strain response [Fig. 6(b)] is represented in the form

Fig. 8 Weibull plot of fracture toughness from nanoindentations on different sand grains

Fig. 9 (color online) Inverse image of nanoindentation on a sand grain at 5 mN load using a cube corner nanoindenter tip

Exp Mech

of Ramberg-Osgood model and the parameters obtained were n=0.25 and K=14.27 GPa. The yield stress obtained from the simulation results is y =6.1 GPa. Fracture Toughness When a brittle material is loaded by a sharp indenter under an appropriate load, radial cracks propagate out of the indenter corners. In such a case, the fracture toughness can be estimated by measuring the lengths of the radial cracks produced at a given indentation load. In order to investigate the failure behavior of the sand grains, a cube corner indenter was used to indent into a polished sand grain to initiate cracks at the corners of three edges of the tip. It is well known that sand grains would behave in a brittle manner under load. In order to estimate the fracture toughness, a formula derived by Antis et al. [22] was used, KC a  0:5   E P H c3=2 7

defects. Figure 9 shows a 3D inverse image of a nanoindent on a sand grain using a cube corner tip at 5 mN load. No cracks were seen and indentation impression clearly showed flow lines from the surface down to the bottom of the indentation indicating ductile flow in a nominally brittle material. This may be the minimal load below which the sand grains would behave in a ductile manner with a possible size effect implications. Further studies are needed to confirm this hypothesis.

Conclusions In order to assess the granular-level mechanical behavior of sand, nanoindentation tests were conducted on individual sand grains to characterize their mechanical properties, namely, Youngs modulus, hardness, fracture toughness, and stress-strain relationship. As can be expected of a brittle material, a wide variation in the granular behavior of sand was observed. Representative values (50 percentile) of Youngs modulus (P50) for the sand grains was found to be 90.1 GPa (range 41.4 to 115.8 GPa), hardness to be 10.7 GPa (range 5.4 to 13.7 GPa), and fracture toughness to be 1.77 MPa-m0.5 (range 0.8 to 3.6 MPa-m0.5). Power-law was used to represent the homogenized and isotropic stressstrain behavior of sand at the granular level. However, the variation in the property value indicates that sand at granular level is very inhomogeneous due to different material constituents, defect structure, and crystal orientations. Need exists to establish the link between nanoindentation measured properties of sand grains with the observed macro properties for accurate prediction of the bulk behavior of sand. The data reported here can be used for mesoscale (granular) simulations of sand in which the individual sand grains would have different properties along with a range of distributions obtained in this study.
Acknowledgements This work is supported by an AFOSR DEPSCoR grant (FA9550-08-1-0328) and the authors acknowledge the strong support and interest of Dr. William L Cooper of the Air Force Research Laboratory, Eglin AFB, and Dr. Victor Giurgiutiu of the Air Force Office of Scientific Research (AFOSR). Additionally, H.L. acknowledges the support of NSF (CMMI-0555902 and CMS-9985060) and R.K. acknowledges the A.H. Nelson, Jr. Endowed Chair in Engineering for additional financial support. The authors also thank Mrs. Sharon Green for her highly skilled editorial assistance.

where c is the crack length, is an empirical constant that takes into account the geometry of the indenter tip (for a cube-corner tip =0.032 [23]). Before proceeding with the nanoindentation on sand grains, a sample test was run on a standard fused silica sample (results not included here). The average value for fracture toughness obtained was 0.6 MPa-m0.5 which is in reasonable agreement with the value reported for fused silica by Harding et al. [43]. For sand, the values of the elastic modulus (E) and hardness (H) were obtained initially using the Berkovich nanoindenter tip. This was followed by indentations with a cube-corner tip (see Youngs Modulus and Hardness). Since we are interested in the ratio of E/H from the nanoindentation tests for calculation of fracture toughness, we obtained an average value of the ratio as 8.5. In equation (7), the crack length c is determined from the surface scans (two of the scans along with inverse images) as shown in [Fig. 7(a) and (b)]. Both 3D inverse images and 2D nanoindentation residual images are shown for details on the crack formation and fracture. It maybe noted that due to slight errors in the alignment and inhomogeniety of the sand grains, it is somewhat difficult to induce three cracks with equal length. When there are differences between the crack lengths from the same indent, we have taken the average crack length for calculation of fracture toughness. The Weibull distribution of the fracture toughness values obtained is shown in Fig. 8. From it, the median value for the fracture toughness was obtained as 1.77 MPa-m0.5 (range 0.8 to 3.6 MPa-m0.5). It maybe noted that cracks observed on some sand grains were not exactly straight. This is attributed to the possible inhomogeneity within a single sand grain, such as presence of

References
1. Poorooshasb HB, Holubec I, Sherbourne AN (1966) Yielding and flow of sand in triaxial compression Part I. Can Geotech J 3 (4):179190 2. Lade PV, Duncan JM (1973) Cubical triaxial tests on cohesionless soil. J Soil Mech Found Div 99(10):793812

Exp Mech 3. Arthur JRF, Menzies BK (1972) Inherent anisotropy in a sand. Geotech 22(1):115128 4. Lade PV, Prabucki MJ (1995) Softening and preshearing effects in sand. Soils Found 35(4):93104 5. Wang Q, Lade PV (2001) Shear banding in true triaxial tests and its effect on failure in sand. J Eng Mech 127(8):754761 6. Haythornthwaite R (1997) Simple shear in granular media. App Mech Rev 50(11):8186 7. Goldenberg C, Goldhirsch I (2005) Friction enhances elasticity in granular solids. Nature 435:188191 8. Geng J, Howell E, Behringer R, Reydellet G, Vanel L, Clement E, Luding S (2001) Footprints in sand: the response of a granular material to local perturbations. Phys Rev Lett 87(3):35506 9. Oda M, Konishi J, Nemat-Nasser S (1980) Some experimentally based fundamental results on the mechanical behavior of granular materials. Geotech 30:479 10. Desrues J, Lanier J, Stutz P (1985) Localization of the deformation in tests on sand sample. Eng Fract Mech 21 (4):909921 11. Oda M, Konishi J (1974) Microscopic deformation mechanism of granular material in simple shear. Soils Found 14(4):2538 12. Liao JJ, Yang M-T, Hsieh H-Y (1996) Direct tensile behavior of a transversely isotropic rock. Int J Rock Mech Min Sci 34(5):837 849 13. Ulm F-J, Abousleiman Y (2006) The nanogranular nature of shale. Acta Geotech 1:7788 14. Pethica JB, Hutchings R, Oliver WC (1983) Hardness measurement at penetration displacements as small as 20 nm. Philos Mag A 48:593606 15. Oliver WC, Hutchings R, Pethica JB (1986) Measurement of hardness at indentation displacements as low as 20 nanometers. Microindentation Techniques in Materials Science and Engineering, ASTM STP 889, American Society for Testing and Materials. Philadelphia, 90108 16. Li X, Bhushan B (2002) A review of nanoindentation continuous stiffness measurement technique and its applications. Mater Charact 48:1136 17. Pethica JB, Oliver WC (1987) Tip surface interactions in STM and AFM. Phys Scripta T19:A61A66 18. Doerner MF, Nix WD (1986) A method for interpreting the data from depth-sensing indentation instruments. J Mater Res 1:601 613 19. Oliver WC, Pharr GM (1992) An improved technique for determining hardness and elastic modulus using load and displacement sensing indentation experiments. J Mater Res 7:15641583 20. Liu Y, Wang B, Yoshino M, Roy S, Lu H, Komanduri R (2005) Combined numerical simulation and nanoindentation for determining mechanical properties of single crystal copper at mesoscale. J Mech Phys Solids 53:27182741 21. Liu Y, Varghese S, Ma J, Yoshino M, Lu H, Komanduri R (2008) Orientation effects in nanoindentation of single crystal copper. Intl J Plasticity 24:19902015 22. Antis GR, Chantikul P, Lawn BR, Marshall DB (1981) A critical evaluation of indentation techniques for measuring fracture toughness: I. Direct crack measurement. J Amer Ceram Soc 64:533538 23. Pharr GM, Harding DS, Oliver WC (1993) Measurement of fracture toughness in thin films and small volumes using nanoindentation method. In: Nastasi M, Parkin DM, Gleiter H (eds) Mechanical properties and deformation behavior of materials having ultra-fine microstructures. Springer-Verlag, New York, pp 449461 Jillavenkatesa A, Dapkunas SJ, Lin-Sien Lum (2001) Particle size characterization. NIST Special Publication 9601 Prandtl L (1920) Ueber die haerte plasticher koerper. Goettinger Nachr, Math Phys Kl 7485 Hill R, Lee EH, Tupper SJ (1947) The theory of combined plastic and elastic deformation with particular reference to a thick tube under internal pressure. Proc R Soc Lond A188:273 Ishlinskii AJ (1944) The axi-symmetrical problem in plasticity and the Brinell test. J Appl Math Mech 8:201 Shield RT (1955) Plastic flow of metals under conditions of axial symmetry. Proc R Soc Lond A233:267287 Tabor D (1970) The hardness of solids. Rev Phys Technol 1:145179 Johnson KL (1970) The correlation of indentation experiments. J Mech Phys Solids 18(2):115126 Bhattacharya AK, Nix WD (1988) Finite element simulation of indentation experiments. Int J Solids Struct 24(9):881891 Bhattacharya AK, Nix WD (1991) Finite element analysis of cone indentation. Int J Solids Struct 27(8):10471058 Wang HF, Bangert H (1993) Three-dimensional finite element simulation of Vickers indentation on coated systems. Mater Sci Eng A 163(1):4350 Sun Y, Bell T, Zheng S (1995) Finite element analysis of the critical ratio of coating thickness to indentation depth for coating property measurements by nanoindentation. Thin Solid Films 258:198204 Lichinchi M, Lenardi C, Haupt J, Vitali R (1998) Simulation of Berkovich nanoindentation experiments on thin films using finite element method. Thin Solid Films 12:240248 Fivel MC, Robertson CF, Canova GR, Boulanger L (1998) Threedimensional modeling of indent-induced plastic zone at a mesoscale. Acta Mater 46(17):67836194 Knapp JA, Follstaedt DM, Myers SM, Barbour JC, Friedmann TA, Ager JW, Monteiro OR, Brown IG (1998) Finite-element modeling of nanoindentation for evaluating mechanical properties of MEMS materials. Surf Coat Technol 103104:268275 Huang G, Lu H (2007) Measurements of two independent viscoelastic functions by nanoindentation. Special Issue- Nanoscale Measurements in Mechanics. Exp Mech 47:8798 Huang G, Daphalapurkar N, Gan R, Lu H (2008) A method for measuring linearly viscoelastic properties of human tympanic membrane using nanoindentation. J Biomech Eng Trans ASME 130:014501-1 Daphalapurkar N, Dai C, Gan R, Lu H (2009) Characterization of the Linearly Viscoelastic Behavior of Human Tympanic Membrane by Nanoindentation. J Mech Behav Biomed Mater 2:8292 ABAQUS version 6.8 documentation (2008) Dassault Systmes, Providence, RI Dao M, Chollacoop N, Van Vliet KJ, Venkatesh TA, Suresh S (2001) Computational modeling of the forward and reverse problems in instrumented sharp indentation. Acta Mater 49:38993918 Harding DS, Oliver WC, Pharr GM (1995) Cracking during nanoindentation and its use in the measurement of fracture toughness. Mater Res Soc Symp Proc 356:663668

24. 25. 26.

27. 28. 29. 30. 31. 32. 33.

34.

35.

36.

37.

38.

39.

40.

41. 42.

43.