63027 03

35

CHAPTER 3

Soil Mineralogy

Figure 3.1 Particle size ranges in soils.

3.1 IMPORTANCE OF SOIL MINERALOGY INGEOTECHNICAL ENGINEERING

Soil is composed of solid particles, liquid, and gas andranges from very soft, organic deposits through lesscompressible clays and sands to soft rock. The solidparticles vary in size from large boulders to minuteparticles that are visible only with the aid of the elec-tron microscope. Particle shapes range from nearlyspherical, bulky grains to thin, flat plates and long,slender needles. Some organic material and noncrys-talline inorganic components are found in most naturalfine-grained soils. A soil may contain virtually any el-ement contained in Earth’s crust; however, by far themost abundant are oxygen, silicon, hydrogen, and alu-minum. These elements, along with calcium, sodium,potassium, magnesium, and carbon, comprise over 99percent of the solid mass of soils worldwide. Atomsof these elements are organized into various crystallineforms to yield the common minerals found in soil.Crystalline minerals comprise the greatest proportionof most soils encountered in engineering practice, andthe amount of nonclay material usually exceeds theamount of clay. Nonetheless, clay and organic matterin a soil usually influence properties in a manner fargreater than their abundance.

Mineralogy is the primary factor controlling the size,shape, and properties of soil particles. These same fac-tors determine the possible ranges of physical andchemical properties of any given soil; therefore, apriori knowledge of what minerals are in a soil pro-vides intuitive insight as to its behavior. Commonlydefined particle size ranges are shown in Fig. 3.1. Thedivisions between gravel, sand, silt, and clay sizes arearbitrary but convenient. Particles smaller than about200 mesh sieve size (0.074 mm), which is the bound-ary between sand and silt sizes, cannot be seen by the

naked eye. Clay can refer both to a size and to a classof minerals. As a size term, it refers to all constituentsof a soil smaller than a particular size, usually 0.002mm (2 �m) in engineering classifications. As a mineralterm, it refers to specific clay minerals that are distin-guished by (1) small particle size, (2) a net negativeelectrical charge, (3) plasticity when mixed with water,and (4) high weathering resistance. Clay minerals areprimarily hydrous aluminum silicates. Not all clay par-ticles are smaller than 2 �m, and not all nonclay par-ticles are coarser than 2 �m; however, the amount ofclay mineral in a soil is often closely approximated bythe amount of material finer than 2 �m. Thus, it isuseful to use the terms clay size and clay mineral con-tent to avoid confusion. A further important differencebetween clay and nonclay minerals is that the nonclaysare composed primarily of bulky particles; whereas,the particles of most of the clay minerals are platy, andin a few cases they are needle shaped or tubular.

The great range in soil particle sizes in relation toother particulate materials, electromagnetic wavelengths, and other size-dependent factors can be seenin Fig. 3.2. The liquid phase of most soil systems iscomposed of water containing various types andamounts of dissolved electrolytes. Organic compounds,both soluble and immiscible, are found in soils at sites

Copy

right

ed M

ater

ial

Copyright © 2005 John Wiley & Sons Retrieved from: www.knovel.com

36

Copy

right

ed M

ater

ial


37

Figure 3.2 Characteristics of particles and particle dispersoids (adapted from Stanford Re-search Institute Journal, Third Quarter, 1961).

Copy

right

ed M

ater

ial


38 3 SOIL MINERALOGY

Figure 3.3 Simplified representation of an atom.

that have been affected by chemical spills, leakingwastes, and contaminated groundwater. The gas phase,in partially saturated soils, is usually air, although or-ganic gases may be present in zones of high biologicalactivity or in chemically contaminated soils.

The mechanical properties of soils depend directlyon interactions of these phases with each other andwith applied potentials (e.g., stress, hydraulic head,electrical potential, and temperature). Because of theseinteractions, we cannot understand soil behavior interms of the solid particles alone. Nonetheless, thestructure of these particles tells us a great deal abouttheir surface characteristics and their potential inter-actions with adjacent phases.

Interatomic and intermolecular bonding forces holdmatter together. Unbalanced forces exist at phaseboundaries. The nature and magnitude of these forcesinfluence the formation of soil minerals, the structure,size, and shape of soil particles, and the physicochem-ical phenomena that determine engineering propertiesand behavior. In this chapter some aspects of atomicand intermolecular forces, crystal structure, structurestability, and characteristics of surfaces that are perti-nent to the understanding of soil behavior are sum-marized simply and briefly. This is followed by asomewhat more detailed treatment of soil minerals andtheir characteristics.

3.2 ATOMIC STRUCTURE

Current concepts of atomic structure and interparticlebonding forces are based on quantum mechanics. Anelectron can have only certain values of energy. Elec-

tronic energy can jump to a higher level by the ab-sorption of radiant energy or drop to a lower level bythe emission of radiant energy. No more than two elec-trons in an atom can have the same energy level, andthe spins of these two electrons must be in oppositedirections. Different bonding characteristics for differ-ent elements exist because of the combined effects ofelectronic energy quantization and the limitation on thenumber of electrons at each energy level.

An atom may be represented in simplified form bya small nucleus surrounded by diffuse concentric‘‘clouds’’ of electrons (Fig. 3.3). The maximum num-ber of electrons that may be located in each diffuseshell is determined by quantum theory. The numberand arrangement of electrons in the outermost shell areof prime importance for the development of differenttypes of interatomic bonding and crystal structure.

Interatomic bonds form when electrons in adjacentatoms interact in such a way that their energy levelsare lowered. If the energy reduction is large, then astrong, primary bond develops. The way in which thebonding electrons are localized in space determineswhether or not the bonds are directional. The strengthand directionality of interatomic bonds, together withthe relative sizes of the bonded atoms, determine thetype of crystal structure assumed by a given compo-sition.

3.3 INTERATOMIC BONDING

Primary Bonds

Only the outer shell or valence electrons participate inthe formation of primary interatomic bonds. There are

Copy

right

ed M

ater

ial


SECONDARY BONDS 39

three limiting types: covalent, ionic, and metallic. Theydiffer because of how the bonding electrons are local-ized in space. The energy of these bonds per mole ofbonded atoms is from 60 � 103 to more than 400 �103 joules (J; 15 to 100 kcal). As there are 6.023 �1023 molecules per mole, it might be argued that suchbonds are weak; however, relative to the weight of anatom they are very large.

Covalent Bonds In the covalent bond, one or morebonding electrons are shared by two atomic nuclei tocomplete the outer shell for each atom. Covalent bondsare common in gases. If outer shell electrons are rep-resented by dots, then examples for (1) hydrogen gas,(2) methane, and (3) chlorine gas are:

1. H � � �H � H:H

H2.

��C � � 4H � � H:C:H��

H

3.� � � �

:Cl � � �Cl: � :Cl:Cl:� � � �

In the solid state, covalent bonds form primarily be-tween nonmetallic atoms such as oxygen, chlorine,nitrogen, and fluorine. Since only certain electronsparticipate in the bonding, covalent bonds are direc-tional. As a result, atoms bonded covalently pack insuch a way that there are fixed bond angles.

Ionic Bonds Ionic bonds form between positivelyand negatively charged free ions that acquire theircharge through gain or loss of electrons. Cations (pos-itively charged atoms that are attracted by the cathodein an electric field) form by atoms giving up one ormore loosely held electrons that lie outside a com-pleted electron shell and have a high energy level. Met-als, alkalies (e.g., sodium, potassium), and alkalineearths (e.g., calcium, magnesium) form cations. Anions(negatively charged atoms that are attracted to the an-ode) are those atoms requiring only a few electrons tocomplete their outer shell. Because the outer shells ofions are complete, structures cannot form by electronsharing as in the case of the covalent bond. Since ionsare electrically charged, however, strong electrical at-tractions (and repulsions) can develop between them.

The ionic bond is nondirectional. Each cation at-tracts all neighboring anions. In sodium chloride,which is one of the best examples of ionic bonding, asodium cation attracts as many chlorine anions as willfit around it. Geometric considerations and electricalneutrality determine the actual arrangement of ioni-cally bonded atoms.

As ionic bonding causes a separation between thecenters of positive and negative charge in a molecule,the molecule will orient in an electrical field forminga dipole. The strength of this dipole is expressed in

terms of the dipole moment �. If two electrical chargesof magnitude ��e, where e is the electronic charge,are separated by a distance d, then

� � d � �e (3.1)

Covalently bonded atoms may also produce dipolarmolecules.

Metallic Bonds Metals contain loosely held val-ence electrons that hold the positive metal ions to-gether but are free to travel through the solid material.Metallic bonds are nondirectional and can exist onlyamong a large group of atoms. It is the large group ofelectrons and their freedom to move that make metalssuch good conductors of electricity and heat. The me-tallic bond is of little importance in most soils.

Bonding in Soil Minerals

A combination of ionic and covalent bonding is typicalin most nonmetallic solids. Purely ionic or covalentbonding is a limiting condition that is the exceptionrather than the rule in most cases. Silicate minerals arethe most abundant constituents of most soils. The in-teratomic bond in silica (SiO2) is about half covalentand half ionic.

3.4 SECONDARY BONDS

Secondary bonds that are weak relative to ionic andcovalent bonds also form between units of matter. Theymay be strong enough to determine the final arrange-ments of atoms in solids, and they may be sources ofattraction between very small particles and betweenliquids and solid particles.

The Hydrogen Bond

If a hydrogen ion forms the positive end of a dipole,then its attraction to the negative end of an adjacentmolecule is termed a hydrogen bond. Hydrogen bondsform only between strongly electronegative atoms suchas oxygen and fluorine because these atoms producethe strongest dipoles. When the electron is detachedfrom a hydrogen atom, such as when it combines withoxygen to form water, only a proton remains. As theelectrons shared between the oxygen and hydrogen at-oms spend most of their time between the atoms, theoxygens act as the negative ends of dipoles, and thehydrogen protons act as the positive ends. The positiveand negative ends of adjacent water molecules tie themtogether forming water and ice.

The strength of the hydrogen bond is much greaterthan that of other secondary bonds because of the smallsize of the hydrogen ion. Hydrogen bonds are impor-

Copy

right

ed M

ater

ial



Figure 3.4 Examples of some common crystals. (hkl) arecleavage plane indices. From Dana’s Manual of Mineralogy,by C. S. Hurlbut, 16th Edition. Copyright � 1957 by JohnWiley & Sons. Reprinted with permission from John Wiley& Sons.

tant in determining some of the characteristics of theclay minerals and in the interaction between soil par-ticle surfaces and water.

van der Waals Bonds

Permanent dipole bonds such as the hydrogen bond aredirectional. Fluctuating dipole bonds, commonlytermed van der Waals bonds, also exist because at anyone time there may be more electrons on one side ofthe atomic nucleus than on the other. This creates weakinstantaneous dipoles whose oppositely charged endsattract each other.

Although individual van der Waals bonds are weak,typically an order of magnitude weaker than a hydro-gen bond, they are nondirectional and additive betweenatoms. Consequently, they decrease less rapidly withdistance than primary valence and hydrogen bondswhen there are large groups of atoms. They are strongenough to determine the final arrangements of groupsof atoms in some solids (e.g., many polymers), andthey may be responsible for small cohesions in fine-grained soils. Van der Waals forces are described fur-ther in Chapter 7.

3.5 CRYSTALS AND THEIR PROPERTIES

Particles composed of mineral crystals form thegreatest proportion of the solid phase of a soil. A crys-tal is a homogeneous body bounded by smooth planesurfaces that are the external expression of an orderlyinternal atomic arrangement. A solid without internalatomic order is termed amorphous.

Crystal Formation

Crystals may form in three ways:

1. From Solution Ions combine as they separatefrom solution and gradually build up a solid ofdefinite structure and shape. Halite (sodium chlo-ride) and other evaporites are examples.

2. By Fusion Crystals form directly from a liquidas a result of cooling. Examples are igneous rockminerals solidified from molten rock magma andice from water.

3. From Vapor Although not of particular impor-tance in the formation of soil minerals, crystalscan form directly from cooling vapors. Examplesinclude snowflakes and flowers of sulfur.

Examples of some common crystals are shown inFig. 3.4.

Characteristics of Crystals

Certain crystal characteristics are used to distinguishdifferent classes or groups of minerals. Variations inthese characteristics result in different properties.

1. Structure The atoms in a crystal are arrangedin a definite orderly manner to form a three-dimensional network termed a lattice. Positionswithin the lattice where atoms or atomic groups

Copy

right

ed M

ater

ial


CRYSTALS AND THEIR PROPERTIES 41

Figure 3.5 Unit cells of the 14 Bravais space lattices. The capital letters refer to the typeof cell: P, primitive cell; C, cell with a lattice point in the center of two parallel faces; F,cell with a lattice point in the center of each face; I, cell with a lattice point in the centerof the interior; R, rhombohedral primitive cell. All points indicated are lattice points. Thereis no general agreement on the unit cell to use for the hexagonal Bravais lattice; some preferthe P cell shown with solid lines, and others prefer the C cell shown in dashed lines (modifiedfrom Moffatt et al., 1965).

are located are termed lattice points. Only 14 dif-ferent arrangements of lattice points in space arepossible. These are the Bravais space lattices,and they are illustrated in Fig. 3.5.

The smallest subdivision of a crystal that stillpossesses the characteristic composition and spa-tial arrangement of atoms in the crystal is the unit

cell. The unit cell is the basic repeating unit ofthe space lattice.

2. Cleavage and Outward Form The angles be-tween corresponding faces on crystals of thesame substance are constant. Crystals breakalong smooth cleavage planes. Cleavage planeslie between planes in which the atoms are most

Copy

right

ed M

ater

ial



Figure 3.6 The six crystal systems.

densely packed. This is because the center-to-center distance between atoms on opposite sidesof the plane is greater than along other planesthrough the crystal. As a result, the strength alongcleavage planes is less than in other directions.

3. Optical Properties The specific atomic arrange-ments within crystals allow light diffraction andpolarization. These properties are useful for iden-tification and classification. Identification of rockminerals by optical means is common. Opticalstudies in soil are less useful because of the smallsizes of most soil particles.

4. X-ray and Electron Diffraction The orderlyatomic arrangements in crystals cause them tobehave with respect to X-ray and electron beamsin much the same way as does a diffraction grat-ing with respect to visible light. Different crystalsyield different diffraction patterns. This makes X-ray diffraction a powerful tool for the study andidentification of very small particles, such as claythat cannot be seen using optical means.

5. Symmetry There are 32 distinct crystal classesbased on symmetry considerations involving thearrangement and orientation of crystal faces.These 32 classes may be grouped into 6 crystalsystems with the classes within each system bear-ing close relationships to each other.

The six crystal systems are illustrated in Fig. 3.6.Crystallographic axes parallel to the intersection edgesof prominent crystal faces are established for each ofthe six crystal systems. In most crystals, these axes willalso be symmetry axes or axes normal to symmetryplanes. In five of the six systems, the crystals are re-ferred to three crystallographic axes. In the sixth (thehexagonal system), four axes are used. The axes aredenoted by a, b, c (a1, a2, a3, and c in the hexagonalsystem) and the angles between the axes by �, �,and .

Isometric or Cubic System There are three mutu-ally perpendicular axes of equal length. Mineralexamples are galena, halite, magnetite, and pyrite.

Hexagonal System Three equal horizontal axes ly-ing in the same plane intersect at 60� with a fourthaxis perpendicular to the other three and of dif-ferent length. Examples are quartz, brucite, cal-cite, and beryl.

Tetragonal System There are three mutually per-pendicular axes, with two horizontal of equallength, but different than that of the vertical axis.Zircon is an example.

Orthorhombic System There are three mutuallyperpendicular axes, each of different length. Ex-amples include sulfur, anhydrite, barite, diaspore,and topaz.

Monoclinic System There are three unequal axes,two inclined to each other at an oblique angle,with the third perpendicular to the other two. Ex-amples are orthoclase feldspar, gypsum, musco-vite, biotite, gibbsite, and chlorite.

Triclinic System Three unequal axes intersect atoblique angles. Examples are plagioclase feldspar,kaolinite, albite, microcline, and turquoise.

3.6 CRYSTAL NOTATION

Miller indices are used to describe plane orientationsand directions in a crystal. This information, alongwith the distances that separate parallel planes is im-portant for the identification and classification of dif-ferent minerals. All lengths are expressed in terms ofunit cell lengths. Any plane through a crystal may bedescribed by intercepts, in terms of unit cell lengths,on the three or four crystallographic axes for the sys-tem in which the crystal falls. The reciprocals of theseintercepts are used to index the plane. Reciprocals areused to avoid fractions and to account for planes par-allel to an axis (an intercept of infinity equals an indexvalue of 0).

Copy

right

ed M

ater

ial


CRYSTAL NOTATION 43

Figure 3.7 Miller indices: (a) Unit cell of muscovite, (b) (002) plane for muscovite, (c)(014) plane for muscovite, and (d) plane for muscovite.(623)

An example illustrates the determination and mean-ing of Miller indices. Consider the mineral muscovite,a member of the monoclinic system. It has unit celldimensions of a � 0.52 nanometers (nm), b � 0.90nm, c � 2.0 nm, and � � 95� 30�. Both the compo-sition and crystal structure of muscovite are similar tothose of some of the important clay minerals.

The muscovite unit cell dimensions and interceptsare shown in Fig. 3.7a. The intercepts for any plane ofinterest are first determined in terms of unit cell

lengths. Take plane mnp in Fig. 3.7a as an example.The intercepts of this plane are a � 1, b � 1, andc � 1. The Miller indices of this plane are found bytaking the reciprocals of these intercepts and clearingof fractions. Thus,

Reciprocals are 1/1, 1/1, 1/1Miller indices are (111)

The indices are always enclosed within parenthesesand indicated in the order abc without commas. Paren-

Copy

right

ed M

ater

ial



Table 3.1 Atomic Packing, Structure, and StructuralStability

RadiusRatioa Nb Geometry Example Stability

0–0.155 2 Line — —0.155–

0.2253 Triangle (CO3)2� Very high

0.225–0.414

4 Tetrahedron (SiO4)4� Moderatelyhigh

0.414–0.732

6 Octahedron [Al(OH)6]3� High

0.732–1.0

8 Body-cen-tered cube

Iron Low

1.0 12 Sheet K–O bondin mica

Very low

aRange of cation to anion diameter ratios over whichstable coordination is expected.

bCoordination number.

Table 3.2 Relative Stabilities of Some Soil MineralStructural Units

Structural Unit

ApproximateRelative Bond

Strength(Valence/N)

Silicon tetrahedron, (SiO4)4� 4/4 � 1Aluminum tetrahedron, [Al(OH)4]1� 3/4Aluminum octahedron, [Al(OH)6]3� 3/6 � 1/2Magnesium octahedron, [Mg(OH)6]4� 2/6 � 1/3K–O12

�23 1/12

theses are always used to indicate crystallographicplanes, whereas brackets are used to indicate direc-tions. For example, [111] designates line oq in Fig.3.7a. Additional examples of Miller indices for planesthrough the muscovite crystal are shown in Figs. 3.7b,3.7c, and 3.7d. A plane that cuts a negative axis isdesignated by placing a bar over the index that pertainsto the negative intercept (Fig. 3.7d). The general index(hkl) is used to refer to any plane that cuts all threeaxes. Similarly (h00) designates a plane cutting onlythe a axis, (h0l) designates a plane parallel to the baxis, and so on. For crystals in the hexagonal system,the Miller index contains four numbers. The (001)planes of soil minerals are of particular interest be-cause they are indicative of specific clay mineral types.

3.7 FACTORS CONTROLLING CRYSTALSTRUCTURES

Organized crystal structures do not develop by chance.The most stable arrangement of atoms in a crystal isthat which minimizes the energy per unit volume. Thisis achieved by preserving electrical neutrality, satisfy-ing bond directionality, minimizing strong ion repul-sions, and packing atoms closely together.

If the interatomic bonding is nondirectional, then therelative atomic sizes have a controlling influence onpacking. The closest possible packing will maximizethe number of bonds per unit volume and minimize thebonding energy. If interatomic bonds are directional,as is the case for covalent bonds, then both bond anglesand atomic size are important.

Anions are usually larger than cations because ofelectron transfer from cations to anions. The numberof nearest neighbor anions that a cation possesses in astructure is termed the coordination number (N) or li-gancy. Possible values of coordination number in solidstructures are 1 (trivial), 2, 3, 4, 6, 8, and 12. Therelationships between atomic sizes, expressed as theratio of cationic to anionic radii, coordination number,and the geometry formed by the anions are indicatedin Table 3.1.

Most solids do not have bonds that are completelynondirectional, and the second nearest neighbors mayinfluence packing as well as the nearest neighbors.Even so, the predicted and observed coordination num-bers are in quite good agreement for many materials.The valence of the cation divided by the number ofcoordinated anions is an approximate indication of therelative bond strength, which, in turn, is related to thestructural stability of the unit. Some of the structuralunits common in soil minerals and their relative bondstrengths are listed in Table 3.2.

The basic coordination polyhedra are seldom elec-trically neutral. In crystals formed by ionic bonded pol-yhedra, the packing maintains electrical neutrality andminimizes strong repulsions between ions with likecharge. In such cases, the valence of the central cationequals the total charge of the coordinated anions, andthe unit is really a molecule. Units of this type are heldtogether by weaker, secondary bonds. An example isbrucite, a mineral that has the composition Mg(OH)2.The Mg2� ions are in octahedral coordination with six(OH)� ions forming a sheet structure in such a waythat each (OH)� is shared by 3Mg2�. In a sheet con-taining N Mg2� ions, therefore, there must be 6N/3 �2N (OH)� ions. Thus, electrical neutrality results, andthe sheet is in reality a large molecule. Successive oc-

Copy

right

ed M

ater

ial


SURFACES 45

tahedral sheets are loosely bonded by van der Waalsforces. Because of this, brucite has perfect basal cleav-age parallel to the sheets.

Cations concentrate their charge in a smaller volumethan do anions, so the repulsion between cations isgreater than between anions. Cationic repulsions areminimized when the anions are located at the centersof coordination polyhedra. If the cations have a lowvalence, then the anion polyhedra pack as closely aspossible to minimize energy per unit volume. If, on theother hand, the cations are small and highly charged,then the units arrange in a variety of ways in responseto the repulsions. The silicon cation is in this category.

3.8 SILICATE CRYSTALS

Small cations form structures with coordination num-bers of 3 and 4 (Table 3.1). These cations are oftenhighly charged and generate strong repulsions betweenadjacent triangles or tetrahedra. As a result, such struc-tures share only corners and possibly edges, but neverfaces, since to do so would bring the cations too closetogether. The radius of silicon is only 0.039 nm,whereas that of oxygen is 0.132 nm. Thus silicon andoxygen combine in tetrahedral coordination, with thesilicon occupying the space at the center of the tetra-hedron formed by the four oxygens. The tetrahedralarrangement satisfies both the directionality of thebonds (the Si–O bond is about half covalent and halfionic) and the geometry imposed by the radius ratio.Silicon is very abundant in Earth’s crust, amounting toabout 25 percent by weight, but only 0.8 percent byvolume. Almost half of igneous rock by weight and91.8 percent by volume is oxygen.

Silica tetrahedra join only at their corners, andsometimes not at all. Thus many crystal structures arepossible, and there is a large number of silicate min-erals. Silicate minerals are classified according to howthe silica tetrahedra (SiO4)

4� associate with each other,as shown in Fig. 3.8. The tetrahedral combinations in-crease in complexity from the beginning to the end ofthe figure. The structural stability increases in the samedirection.

Island (independent) silicates are those in which thetetrahedra are not joined to each other. Instead, the fourexcess oxygen electrons are bonded to other positiveions in the crystal structure. In the olivine group, theminerals have the composition R2

2� � SiO44�. Garnets

contain cations of different valences and coordinationnumbers R3

2� � R23�(SiO4)3. The negative charge of the

SiO4 group in zircon is all balanced by the single Zr4�.Ring and chain silicates are formed when corners of

tetrahedra are shared. The formulas for these structures

contain (SiO3)2�. The pyroxene minerals are in this

class. Enstatite, MgSiO3, is a simple member of thisgroup. Some of the positions normally occupied bySi4� in single-chain structures may be filled by Al3�.

Substitution of ions of one kind by ions of anothertype, having either the same or different valence, butthe same crystal structure, is termed isomorphous sub-stitution. The term substitution implies a replacementwhereby a cation in the structure is replaced at sometime by a cation of another type. In reality, however,the replaced cations were never there, and the mineralwas formed with its present proportions of the differentcations in the structure.

Double chains of indefinite length may form with(Si4O11)

6� as part of the structure. The amphiboles fallinto this group (Fig. 3.8). Hornblendes have the samebasic structure, but some of the Si4� positions are filledby Al3�. The cations Na� and K� can be incorporatedinto the structure to satisfy electrical neutrality; Al3�,Fe3�, Fe2�, and Mn2� can replace part of the Mg2� insixfold coordination, and the (OH)� group can be re-placed by F�.

In sheet silicates three of the four oxygens of eachtetrahedron are shared to give structures containing(Si2O5)

2�. The micas, chlorites, and many of the clayminerals contain silica in a sheet structure. Frameworksilicates result when all four of the oxygens are sharedwith other tetrahedra. The most common example isquartz. In quartz, the silica tetrahedra are grouped toform spirals. The feldspars also have three-dimensionalframework structures. Some of the silicon positions arefilled by aluminum, and the excess negative chargethus created is balanced by cations of high coordina-tion such as potassium, calcium, sodium, and barium.Differences in the amounts of this isomorphous sub-stitution are responsible for the different members ofthe feldspar family.

3.9 SURFACES

All liquids and solids terminate at a surface, or phaseboundary, on the other side of which is matter of adifferent composition or state. In solids, atoms arebonded into a three-dimensional structure, and the ter-mination of this structure at a surface, or phase bound-ary, produces unsatisfied force fields. In a fine-grainedparticulate material such as clay soil the surface areamay be very large relative to the mass of the material,and, as is emphasized throughout this book, the influ-ences of the surface forces on properties and behaviormay be very large.

Unsatisfied forces at solid surfaces may be balancedin any of the following ways:

Copy

right

ed M

ater

ial



Figure 3.8 Silica tetrahedral arrangements in different silicate mineral structures. ReprintedGillott (1968) with permission from Elsevier Science Publishers BV.

Copy

right

ed M

ater

ial


SURFACES 47

Figure 3.8 (Continued )

1. Attraction and adsorption of molecules from theadjacent phase

2. Cohesion with the surface of another mass of thesame substance

3. Solid-state adjustments of the structure beneaththe surface.

Each unsatisfied bond force is significant relative tothe weight of atoms and molecules. The actual mag-nitude of 10�11 N or less, however, is infinitesimalcompared to the weight of a piece of gravel or a grainof sand. On the other hand, consider the effect of re-ducing particle size. A cube 10 mm on an edge has a

Copy

right

ed M

ater

ial



surface area of 6.0 � 10�4 m2. If it is cut in half inthe three directions, eight cubes result, each 5 mm onan edge. The surface area now is 12.0 � 10�4 m2. Ifthe cubes are further divided to 1 �m on an edge, thesurface becomes 6.0 m2 for the same 1000 mm3 ofmaterial. Thus, as a solid is subdivided into smallerand smaller units, the proportion of surface area toweight becomes larger and larger. For a given particleshape, the ratio of surface area to volume is inverselyproportional to some effective particle diameter.

For many materials when particle size is reduced to1 or 2 �m or less the surface forces begin to exert adistinct influence on the behavior. Study of the behav-ior of particles of this size and less requires consider-ations of colloidal and surface chemistry. Most clayparticles behave as colloids, both because of theirsmall size and because they have unbalanced surfaceelectrical forces as a result of isomorphous substitu-tions within their structure.

Montmorillonite, which is one of the members ofthe smectite clay mineral group (see Section 3.17),may break down into particles that are only 1 unit cellthick (1.0 nm) when in a dispersed state and have aspecific surface area of 800 m2/g. If all particles con-tained in about 10 g of this clay could be spread outside by side, they would cover a football field.

3.10 GRAVEL, SAND, AND SILT PARTICLES

The physical characteristics of cohesionless soils, thatis, gravel, sand, and nonplastic silts, are determinedprimarily by particle size, shape, surface texture, andsize distribution. The mineral composition determineshardness, cleavage, and resistance to physical andchemical breakdown. Some carbonate and sulfate min-erals, such as calcite and gypsum, are sufficiently sol-uble that their decomposition may be significant withinthe time frame of many projects. In many cases, how-ever, the nonclay particles may be treated as relativelyinert, with interactions that are predominantly physicalin nature. Evidence of this is provided by the soils onthe Moon. Lunar soils have a silty, fine sand gradation;however, their compositions are totally different thanthose of terrestrial soils of the same gradation. Theengineering properties of the two materials are sur-prisingly similar, however.

The gravel, sand, and most of the silt fraction in asoil are composed of bulky, nonclay particles. As mostsoils are the products of the breakdown of preexistingrocks and soils, they are weathering products. Thus,the predominant mineral constituents of any soil arethose that are one or more of the following:

1. Very abundant in the source material2. Highly resistant to weathering, abrasion, and im-

pact3. Weathering products

The nonclays are predominantly rock fragments ormineral grains of the common rock-forming minerals.In igneous rocks, which are the original source mate-rial for many soils, the most prevalent minerals are thefeldspars (about 60 percent) and the pyroxenes andamphiboles (about 17 percent). Quartz accounts forabout 12 percent of these rocks, micas for 4 percent,and other minerals for about 8 percent.

However, in most soils, quartz is by far the mostabundant mineral, with small amounts of feldspar andmica also present. Pyroxenes and amphiboles are sel-dom found in significant amounts. Carbonate minerals,mainly calcite and dolomite, are also found in somesoils and can occur as bulky particles, shells, precipi-tates, or in solution. Carbonates dominate the compo-sition of some deep-sea sediments. Sulfates, in variousforms, are found primarily in soils of semiarid and aridregions, with gypsum (CaSO4 � 2H2O) being the mostcommon. Iron and aluminum oxides are abundant inresidual soils of tropical regions.

Quartz is composed of silica tetrahedra grouped toform spirals, with all tetrahedral oxygens bonded tosilicon. The tetrahedral structure has a high stability.In addition, the spiral grouping of tetrahedra producesa structure without cleavage planes, quartz is alreadyan oxide, there are no weakly bonded ions in the struc-ture, and the mineral has high hardness. Collectively,these factors account for the high persistence of quartzin soils.

Feldspars are silicate minerals with a three-dimensional framework structure in which part of thesilicon is replaced by aluminum. The excess negativecharge resulting from this replacement is balanced bycations such as potassium, calcium, sodium, strontium,and barium. As these cations are relatively large, theircoordination number is also large. This results in anopen structure with low bond strengths between units.Consequently, there are cleavage planes, the hardnessis only moderate, and feldspars are relatively easilybroken down. This accounts for their lack of abun-dance in soils compared to their abundance in igneousrocks.

Mica has a sheet structure composed of tetrahedraland octahedral units. Sheets are stacked one on theother and held together primarily by potassium ions in12-fold coordination that provide an electrostatic bondof moderate strength. In comparison with the intralayerbonds, however, this bond is weak, which accounts forthe perfect basal cleavage of mica. As a result of the

Copy

right

ed M

ater

ial


STRUCTURAL UNITS OF THE LAYER SILICATES 49

Figure 3.9 Swelling index as a function of mica content forcoarse-grained mixtures (data from Terzaghi, 1931).

thin-plate morphology of mica flakes, sand and siltscontaining only a few percent mica may exhibit highcompressibility when loaded and large swelling whenunloaded, as may be seen in Fig. 3.9. The amphiboles,pyroxenes, and olivine have crystal structures that arerapidly broken down by weathering; hence they areabsent from most soils.

Some examples of silt and sand particles from dif-ferent soils are shown in Fig. 3.10. Angularity androundness can be used to describe particle shapes, asshown in Fig. 3.11. Elongated and platy particles candevelop preferred orientations, which can be respon-sible for anisotropic properties within a soil mass. Thesurface texture of the grains influences the stress–deformation and strength properties.

3.11 SOIL MINERALS AND MATERIALSFORMED BY BIOGENIC AND GEOCHEMICALPROCESSES

Evaporite deposits formed by precipitation of saltsfrom salt lakes and seas as a result of the evaporationof water are sometimes found in layers that are severalmeters thick. The major constituents of seawater andtheir relative proportions are listed in Table 3.3. Alsolisted are some of the more important evaporite de-

posits. In some areas alternating layers of evaporite andclay or other fine-grained sediments are formed duringcyclic wet and dry periods.

Many limestones, as well as coral, have been formedby precipitation or from the remains of various organ-isms. Because of the much greater solubility of lime-stone than most other rock types, it may be the sourceof special problems caused by solution channels andcavities under foundations.

Chemical sediments and rocks in freshwater lakes,ponds, swamps, and bays are occasionally encounteredin civil engineering projects. Biochemical processesform marl, which ranges from relatively pure calciumcarbonate to mixtures with mud and organic matter.Iron oxide is formed in some lakes. Diatomite or dia-tomaceous earth is essentially pure silica formed fromthe skeletal remains of small (up to a few tenths of amillimeter) freshwater and saltwater organisms. Owingto their solubility limestone, calcite, gypsum, and othersalts may cause special geotechnical problems.

Oxidation and reduction of pyrite-bearing earth ma-terials, that is, soils and rocks containing FeS2, can bethe source of many types of geotechnical problems,including ground heave, high swell pressures, forma-tion of acid drainage, damage to concrete, and corro-sion of steel (Bryant et al., 2003). The chemical andbiological processes and consequences of pyritic re-actions are covered in Sections 8.3, 8.11, and 8.16.

More than 12 percent of Canada is covered by apeaty material, termed muskeg, composed almost en-tirely of decaying vegetation. Peat and muskeg mayhave water contents of 1000 percent or more; they arevery compressible, and they have low strength. Thespecial properties of these materials and methods foranalysis of geotechnical problems associated withthem are given by MacFarlane (1969), Dhowian andEdil (1980), and Edil and Mochtar (1984).

3.12 SUMMARY OF NONCLAY MINERALCHARACTERISTICS

Important compositional, structural, and morphologicalcharacteristics of the important nonclay minerals foundin soils are summarized in Table 3.4. Of these miner-als, quartz is by far the most common, both in termsof the number of soils in which it is found and itsabundance in a typical soil. Feldspar and mica are fre-quently present in small percentages.

3.13 STRUCTURAL UNITS OF THE LAYERSILICATES

Clay minerals in soils belong to the mineral familytermed phyllosilicates, which also contains other layer

Copy

right

ed M

ater

ial



Figure 3.10 Photomicrographs of sand and silt particles from several soils: (a) Ottawa stan-dard sand, (b) Monterey sand, (c) Sacramento River sand, (d) Eliot sand, and (e) lunar soilmineral grains (photo courtesy Johnson Space Center). Squares in background area are 1�1mm. (ƒ) Recrystallized breccia particles from lunar soil (photo courtesy of NASA JohnsonSpace Center). Squares in background grid are 1�1 mm.

silicates such as serpentine, pyrophyllite, talc, mica,and chlorite. Clay minerals occur in small particlesizes, and their unit cells ordinarily have a residualnegative charge that is balanced by the adsorption ofcations from solution.

The structures of the common layer silicates aremade up of combinations of two simple structuralunits, the silicon tetrahedron (Fig. 3.12) and the alu-minum or magnesium octahedron (Fig. 3.13). Differentclay mineral groups are characterized by the stackingarrangements of sheets1 (sometimes chains) of these

1 In conformity with the nomenclature of the Clay Minerals Society(Bailey et al., 1971), the following terms are used: a plane of atoms,a sheet of basic structural units, and a layer of unit cells composedof two, three, or four sheets.

units and the manner in which two successive two- orthree-sheet layers are held together.

Differences among minerals within clay mineralgroups result primarily from differences in the type andamount of isomorphous substitution within the crystalstructure. Possible substitutions are nearly endless innumber, and the crystal structure arrangement mayrange from very poor to nearly perfect. Fortunately forengineering purposes, knowledge of the structural andcompositional characteristics of each group, withoutdetailed study of the subtleties of each specific mineral,is adequate.

Silica Sheet

In most clay mineral structures, the silica tetrahedraare interconnected in a sheet structure. Three of the

Copy

right

ed M

ater

ial


STRUCTURAL UNITS OF THE LAYER SILICATES 51

Figure 3.10 (Continued )

Figure 3.11 Sand and silt size particle shapes as seen insilhouette.

four oxygens in each tetrahedron are shared to form ahexagonal net, as shown in Figs. 3.12b and 3.14. Thebases of the tetrahedra are all in the same plane, andthe tips all point in the same direction. The structurehas the composition (Si4O10)

4� and can repeat indefi-nitely. Electrical neutrality can be obtained by replace-ment of four oxygens by hydroxyls or by union witha sheet of different composition that is positivelycharged. The oxygen-to-oxygen distance is 2.55 ang-stroms (A),2 the space available for the silicon ion is0.55 A, and the thickness of the sheet in clay mineralstructures is 4.63 A (Grim, 1968).

2 In conformity with the SI system of units, lengths should be givenin nanometers. For convenience, however, the angstrom unit is re-tained for atomic dimensions, where 1 A � 0.1 nm.

Silica Chains

In some of the less common clay minerals, silica tet-rahedra are arranged in bands made of double chainsof composition (Si4O11)

6�. Electrical neutrality isachieved and the bands are bound together by alumi-num and/or magnesium ions. A diagrammatic sketchof this structure is shown in Fig. 3.8. Minerals in thisgroup resemble the amphiboles in structure.

Octahedral Sheet

This sheet structure is composed of magnesium or alu-minum in octahedral coordination with oxygens or hy-droxyls. In some cases, other cations are present inplace of Al3� and Mg2�, such as Fe2�, Fe3�, Mn2�,Ti4�, Ni2�, Cr3�, and Li�. Figure 3.13b is a schematicdiagram of such a sheet structure. The oxygen-to-oxygen distance is 2.60 A, and the space available forthe octahedrally coordinated cation is 0.61 A. Thethickness of the sheet is 5.05 A in clays (Grim, 1968).

If the cation is trivalent, then normally only two-thirds of the possible cationic spaces are filled, and thestructure is termed dioctahedral. In the case of alu-minum, the composition is Al2(OH)6. This compositionand structure form the mineral gibbsite. When com-bined with silica sheets, as is the case in clay mineral

Copy

right

ed M

ater

ial



Table 3.3 Major Constituents of Seawater and Evaporite Deposits

Ion Grams per LiterPercent by Weight

of Total Solids Important Evaporite Deposits

Sodium, Na� 10.56 30.61 Anhydrite CaSO4

Magnesium, Mg2� 1.27 3.69 Barite BaSO4

Calcium, Ca2� 0.40 1.16 Celesite SrSO4

Potassium, K� 0.38 1.10 Kieserite MgSO4 � H2OStrontium, Sr2� 0.013 0.04 Gypsum CaSO4 � 2H2OChloride, Cl� 18.98 55.04 Polyhalite Ca2K2Mg(SO4) � 2H2OSulfate, SO4

2� 2.65 7.68 Bloedite Ma2Mg(SO4)2 � 4H2OBicarbonate, HCO3

� 0.14 0.41 Hexahydrite MgSO4 � 6H2OBromide, Br� 0.065 0.19 Epsomite MgSO4 � 7H2OFluoride, F� 0.001 — Kainite K4Mg4(Cl/SO4) � 1 1H2OBoric Acid, H3BO3 0.026 0.08 Halite NaCl

34.485 100.00 Sylvite KClFlourite CaF2

Bischofite MgCl2 � 6H2OCarnallite KMgCl3 � 6H2O

Adapted from data by Degens (1965).

structures, an aluminum octahedral sheet is referred toas a gibbsite sheet.

If the octahedrally coordinated cation is divalent,then normally all possible cation sites are occupied andthe structure is trioctahedral. In the case of magne-sium, the composition is Mg3(OH)6, giving the mineralbrucite. In clay mineral structures, a sheet of magne-sium octahedra is termed a brucite sheet.

Schematic representations of the sheets are usefulfor simplified diagrams of the structures of the differ-ent clay minerals:

Silica sheet or

Octahedral sheet (Various cations in octahedral coordination)

Gibbsite sheet (Octahedral sheet cations are mainly aluminum)

Brucite sheet (Octahedral sheet cations are mainly magnesium)

Water layers are found in some structures and maybe represented by �� for each molecular layer.Atoms of a specific type, for example, potassium, arerepresented thus: .�K

The diagrams are indicative of the clay mineral layerstructure. They do not indicate the correct width-to-length ratios for the actual particles. The structuresshown are idealized; in actual minerals, irregular sub-stitutions and interlayering or mixed-layer structuresare common. Furthermore, direct assembly of the basic

units does not necessarily form the naturally occurringminerals. The ‘‘building block’’ approach is useful,however, for the development of conceptual models.

3.14 SYNTHESIS PATTERN ANDCLASSIFICATION OF THE CLAY MINERALS

The manner in which atoms are assembled into tetra-hedral and octahedral units, followed by the formation

of sheets and their stacking to form layers that combineto produce the different clay mineral groups is illus-trated in Fig. 3.15. The basic structures shown in thebottom row of Fig. 3.15 comprise the great prepon-derance of the clay mineral types that are found insoils.

Grouping the clay minerals according to crystalstructure and stacking sequence of the layers is con-venient since members of the same group have gen-

Copy

right

ed M

ater

ial


SYNTHESIS PATTERN AND CLASSIFICATION OF THE CLAY MINERALS 53

Table 3.4 Properties and Characteristics of Nonclay Minerals in Soils

Mineral FormulaCrystalSystem Cleavage

ParticleShape

SpecificGravity Hardness

Occurrencein Soils of

EngineeringInterest

Quartz SiO2 Hexagonal None Bulky 2.65 7 Veryabundant

Orthoclasefeldspar

KalSi3O8 Monoclinic 2 planes Elongate 2.57 6 Common

Plagioclasefeldspar

NaAlSi3O8

CaAl2Si3O8 (variable)Triclinic 2 planes Bulky—

elongate2.62–2.76 6 Common

Muscovitemica

Kal3Si3O10(OH)2 Monoclinic Perfect basal Thin plates 2.76–3.1 2–21⁄2 Common

Biotite mica K(Mg,FE)3AlSi3O10(OH)2 Monoclinic Perfect basal Thin plates 2.8–3.2 21⁄2–3 CommonHornblende Na,Ca,Mg,Fe,Al silicate Monoclinic Perfect

prismaticPrismatic 3.2 5–6 Uncommon

Augite(pyroxene)

Ca(Mg,Fe,Al)(Al,Si)2O6 Monoclinic Good prismatic Prismatic 3.2–3.4 5–6 Uncommon

Olivine (Mg,Fe)2SiO4 Orthorhombic Conchoidalfracture

Bulky 3.27–3.37 61⁄2–7 Uncommon

Calcite CaCO3 Hexagonal Perfect Bulky 2.72 21⁄2–3 May beabundantlocally

Dolomite CaMg(CO3)2 Hexagonal Perfectrhombohedral

Bulky 2.85 31⁄2–4 May beabundantlocally

Gypsum CaSO4 � 2H2O Monoclinic 4 planes Elongate 2.32 2 May beabundantlocally

Pyrite FeS2 Isometric Cubical Bulky cubic 5.02 6–61⁄2

Data from Hurlbut (1957).

Figure 3.12 Silicon tetrahedron and silica tetrahedra arranged in a hexagonal network.

Copy

right

ed M

ater

ial



Figure 3.13 Octahedral unit and sheet structure of octahedral units.

Figure 3.14 Silica sheet in plan view.

erally similar engineering properties. The mineralshave unit cells consisting of two, three, or four sheets.The two-sheet minerals are made up of a silica sheetand an octahedral sheet. The unit layer of the three-sheet minerals is composed of either a dioctahedral ortrioctahedral sheet sandwiched between two silicasheets. Unit layers may be stacked closely together orwater layers may intervene. The four-sheet structure ofchlorite is composed of a 2�1 layer plus an interlayerhydroxide sheet. In some soils, inorganic, claylike ma-terial is found that has no clearly identifiable crystal

structure. Such material is referred to as allophane ornoncrystalline clay.

The bottom row of Fig. 3.15 shows that the 2�1minerals differ from each other mainly in the type andamount of ‘‘glue’’ that holds the successive layers to-gether. For example, smectite has loosely held cationsbetween the layers, illite contains firmly fixed potas-sium ions, and vermiculite has somewhat organizedlayers of water and cations. The chlorite group repre-sents an end member that has 2�1 layers bonded by anorganized hydroxide sheet. The charge per formula

Copy

right

ed M

ater

ial


INTERSHEET AND INTERLAYER BONDING IN THE CLAY MINERALS 55

Figure 3.15 Synthesis pattern for the clay minerals.

unit is variable both within and among groups, andreflects the fact that the range of compositions is greatowing to varying amounts of isomorphous substitution.Accordingly, the boundaries between groups are some-what arbitrary.

Isomorphous Substitution

The concept of isomorphous substitution was intro-duced in Section 3.13 in connection with some of thesilicate crystals. It is very important in the structureand properties of the clay minerals. In an ideal gibbsitesheet, only two-thirds of the octahedral positions arefilled, and all of the cations are aluminum. In an idealbrucite sheet, all the octahedral spaces are filled bymagnesium. In an ideal silica sheet, silicons occupy alltetrahedral spaces. In clay minerals, however, some ofthe tetrahedral and octahedral spaces are occupied bycations other than those in the ideal structure. Commonexamples are aluminum in place of silicon, magnesiuminstead of aluminum, and ferrous iron (Fe2�) for mag-nesium. This presence in an octahedral or tetrahedralposition of a cation other than that normally found,without change in crystal structure, is isomorphoussubstitution. The actual tetrahedral and octahedral cat-ion distributions may develop during initial formationor subsequent alteration of the mineral.

3.15 INTERSHEET AND INTERLAYERBONDING IN THE CLAY MINERALS

A single plane of atoms that are common to both thetetrahedral and octahedral sheets forms a part of the

clay mineral layers. Bonding between these sheets isof the primary valence type and is very strong. How-ever, the bonds holding the unit layers together maybe of several types, and they may be sufficiently weakthat the physical and chemical behavior of the clay isinfluenced by the response of these bonds to changesin environmental conditions.

Isomorphous substitution in all of the clay minerals,with the possible exception of those in the kaolinitegroup, gives clay particles a net negative charge. Topreserve electrical neutrality, cations are attracted andheld between the layers and on the surfaces and edgesof the particles. Many of these cations are exchange-able cations because they may be replaced by cationsof another type. The quantity of exchangeable cationsis termed the cation exchange capacity (cec) and isusually expressed as milliequivalents (meq)3 per 100 gof dry clay.

Five types of interlayer bonding are possible in thelayer silicates (Marshall, 1964).

1. Neutral parallel layers are held by van der Waalsforces. Bonding is weak; however, stable crystalsof appreciable thickness such as the nonclay min-

3 Equivalent weight � combining weight of an element � (atomicweight /valence). Number of equivalents � (weight of element /atomic weight) � valence. The number of ions in an equivalent �Avogardro’s number /valence. Avogadro’s number � 6.02 � 1023. Anequivalent contains 6.02 � 1023 electron charges or 96,500 coulombs,which is 1 faraday.

Copy

right

ed M

ater

ial



erals of pyrophyllite and talc may form. Theseminerals cleave parallel to the layers.

2. In some minerals (e.g., kaolinite, brucite, gibb-site), there are opposing layers of oxygens andhydroxyls or hydroxyls and hydroxyls. Hydrogenbonding then develops between the layers as wellas van der Waals bonding. Hydrogen bonds re-main stable in the presence of water.

3. Neutral silicate layers that are separated byhighly polar water molecules may be held to-gether by hydrogen bonds.

4. Cations needed for electrical neutrality may be inpositions that control interlayer bonding. In mi-cas, some of the silicon is replaced by aluminumin the silica sheets. The resulting charge defi-ciency is partly balanced by potassium ions be-tween the unit cell layers. The potassium ion justfits into the holes formed by the bases of thesilica tetrahedra (Fig. 3.12). As a result, it gen-erates a strong bond between the layers. In thechlorites, the charge deficiencies from substitu-tions in the octahedral sheet of the 2�1 sandwichare balanced by excess charge on the single-sheetlayer interleaved between the three-sheet layers.This provides a strongly bonded structure thatwhile exhibiting cleavage will not separate in thepresence of water or other polar liquids.

5. When the surface charge density is moderate, asin smectite and vermiculite, the silicate layersreadily adsorb polar molecules, and also the ad-sorbed cations may hydrate, resulting in layerseparation and expansion. The strength of the in-terlayer bond is low and is a strong function ofcharge distribution, ion hydration energy, surfaceion configuration, and structure of the polar mol-ecule.

Smectite and vermiculite particles adsorb water be-tween the unit layers and swell, whereas particles ofthe nonclay minerals, pyrophyllite and talc, which havecomparable structures, do not. There are two possiblereasons (van Olphen, 1977):

1. The interlayer cations in smectite hydrate, andthe hydration energy overcomes the attractiveforces between the unit layers. There are no in-terlayer cations in pyrophyllite; hence, no swell-ing.

2. Water does not hydrate the cations but is ad-sorbed on oxygen surfaces by hydrogen bonds.There is no swelling in pyrophyllite and talc be-cause the surface hydration energy is too smallto overcome the van der Waals forces between

layers, which are greater in these minerals be-cause of a smaller interlayer distance.

Whatever the reason, the smectite minerals are thedominant source of swelling in the expansive soils thatare so prevalent throughout the world.

3.16 THE 1�1 MINERALS

The kaolinite–serpentine minerals are composed of al-ternating silica and octahedral sheets as shown sche-matically in Fig. 3.16. The tips of the silica tetrahedraand one of the planes of atoms in the octahedral sheetare common. The tips of the tetrahedra all point in thesame direction, toward the center of the unit layer. Inthe plane of atoms common to both sheets, two-thirdsof the atoms are oxygens and are shared by both sili-con and the octahedral cations. The remaining atomsin this plane are (OH) located so that each is directlybelow the hole in the hexagonal net formed by thebases of the silica tetrahedra. If the octahedral layer isbrucite, then a mineral of the serpentine subgroup re-sults, whereas dioctahedral gibbsite layers give clayminerals in the kaolinite subgroup. Trioctahedral 1�1minerals are relatively rare, usually occur mixed withkaolinite or illite, and are hard to identify. A diagram-matic sketch of the kaolinite structure is shown in Fig.3.17. The structural formula is (OH)8Si4Al4O10, and thecharge distribution is indicated in Fig. 3.18.

Mineral particles of the kaolinite subgroup consistof the basic units stacked in the c direction. The bond-ing between successive layers is by both van der Waalsforces and hydrogen bonds. The bonding is sufficientlystrong that there is no interlayer swelling in the pres-ence of water.

Because of slight differences in the oxygen-to-oxygen distances in the tetrahedral and octahedral lay-ers, there is some distortion of the ideal tetrahedralnetwork. As a result, kaolinite, which is the most abun-dant member of the subgroup and a common soil min-eral, is triclinic instead of monoclinic. The unit celldimensions are a � 5.16 A, b � 8.94 A, c � 7.37 A,� � 91.8�, � � 104.5�, and � 90�.

Variations in stacking of layers above each other,and possibly in the position of aluminum ions withinthe available sites in the octahedral sheet, produce dif-ferent members of the kaolinite subgroup. The dickiteunit cell is made up of two unit layers, and the nacriteunit cell contains six. Both appear to be formed byhydrothermal processes. Dickite is fairly common assecondary clay in the pores of sandstone and in coalbeds. Neither dickite nor nacrite is common in soils.

Copy

right

ed M

ater

ial


THE 1�1 MINERALS 57

Figure 3.16 Schematic diagrams of the structures of kaolinite and serpentine: (a) kaoliniteand (b) serpentine.

Figure 3.17 Diagrammatic sketch of the kaolinite structure.

Figure 3.18 Charge distribution on kaolinite.

Halloysite

Halloysite is a particularly interesting mineral of thekaolinite subgroup. Two distinct endpoint forms of thismineral exist, as shown in Fig. 3.19; one, a hydratedform consisting of unit kaolinite layers separated fromeach other by a single layer of water molecules andhaving the composition (OH)8Si4Al4O10 � 4H2O, andthe other, a nonhydrated form having the same unitlayer structure and chemical composition as kaolinite.The basal spacing in the c direction d(001) for the non-hydrated form is about 7.2 A, as for kaolinite. Becauseof the interleaved water layer, d(001) for hydrated hal-loysite is about 10.1 A. The difference between these

Copy

right

ed M

ater

ial



Figure 3.19 Schematic diagrams of the structure of halloysite: (a) halloysite (10 A) and (b)halloysite (7 A).

Figure 3.20 Electron photomicrograph of well-crystallizedkaolinite from St. Austell, Cornwall, England. Picture widthis 17 �m (Tovey, 1971).

values, 2.9 A, is the approximate thickness of a singlelayer of water molecules.

The recommended terms for the two forms of hal-loysite are halloysite (7 A) and halloysite (10 A).Transformation from halloysite (10 A) to halloysite (7A) by dehydration can occur at relatively low temper-atures and is irreversible. Halloysite is often found insoils formed from volcanic parent materials in wet en-vironments. It can be responsible for special propertiesand problems in earthwork construction, as discussedlater in this book.

Isomorphous Substitution and Exchange Capacity

Whether or not measurable isomorphous substitutionexists within the structure of the kaolinite minerals isuncertain. Nevertheless, values of cation exchange ca-pacity in the range of 3 to 15 meq/100 g for kaoliniteand from 5 to 40 meq/100 g for halloysite have beenmeasured. Thus, kaolinite particles possess a net neg-ative charge. Possible sources are:

1. Substitution of Al3� for Si4� in the silica sheet ora divalent ion for Al3� in the octahedral sheet.Replacement of only 1 Si in every 400 would beadequate to account for the exchange capacity.

2. The hydrogen of exposed hydroxyls may be re-placed by exchangeable cations. According toGrim (1968), however, this mechanism is notlikely because the hydrogen would probably notbe replaceable under the conditions of mostexchange reactions.

3. Broken bonds around particle edges may give un-satisfied charges that are balanced by adsorbedcations.

Kaolinite particles are charged positively on theiredges when in a low pH (acid) environment, but neg-atively charged in a high pH (basic) environment. Lowexchange capacities are measured under low pH con-ditions and high exchange capacities are obtained for

determinations at high pH. This suggests that brokenbonds are at least a partial source of exchange capacity.That a positive cation exchange capacity is measuredunder low pH conditions when edges are positivelycharged indicates that some isomorphous substitutionmust exist also.

As interlayer separation does not occur in kaolinite,balancing cations must adsorb on the exterior surfacesand edges of the particles.

Morphology and Surface Area

Well-crystallized particles of kaolinite (Fig. 3.20), na-crite, and dickite occur as well-formed six-sided plates.The lateral dimensions of these plates range fromabout 0.1 to 4 �m, and their thicknesses are from about0.05 to 2 �m. Poorly crystallized kaolinite generallyoccurs as less distinct hexagonal plates, and the parti-cle size is usually smaller than for the well-crystallizedvarieties.

Copy

right

ed M

ater

ial


SMECTITE MINERALS 59

Figure 3.21 Electron photomicrograph of halloysite fromBedford, Indiana. Picture width is 2 �m (Tovey, 1971).

Halloysite (10 A) occurs as cylindrical tubes ofoverlapping sheets of the kaolinite type (Fig. 3.21).The c axis at any point nearly coincides with the tuberadius. The formation of tubes has been attributed to amisfit in the b direction of the silica and gibbsite sheets(Bates et al., 1950). The b dimension in kaolinite is8.93 A; in gibbsite it is only 8.62 A. This means thatthe (OH) spacing in gibbsite sheets is stretched in orderto obtain a fit with the silica sheet. Evidently, in hal-loysite (10 A), the reduced interlayer bond, caused bythe intervening layer of water molecules, enables the(OH) layer to revert to 8.62 A, resulting in a curvaturewith the hydroxyls on the inside and the bases of thesilica tetrahedra on the outside. The outside diametersof the tubular particles range from about 0.05 to 0.20�m, with a median value of 0.07 �m. The wall thick-ness is about 0.02 �m. The tubes range in length froma fraction of a micrometer to several micrometers. Dry-ing of halloysite (10 A) may result in splitting or un-rolling of the tubes. The specific surface area ofkaolinite is about 10 to 20 m2/g of dry clay; that ofhalloysite (10 A) is 35 to 70 m2/g.

3.17 SMECTITE MINERALS

Structure

The minerals of the smectite group have a prototypestructure similar to that of pyrophyllite, consisting ofan octahedral sheet sandwiched between two silicasheets, as shown schematically in Fig. 3.22 and dia-grammatically in three dimensions in Fig. 3.23. All thetips of the tetrahedra point toward the center of theunit cell. The oxygens forming the tips of the tetra-hedra are common to the octahedral sheet as well. Theanions in the octahedral sheet that fall directly above

and below the hexagonal holes formed by the bases ofthe silica tetrahedra are hydroxyls.

The layers formed in this way are continuous in thea and b directions and stacked one above the other inthe c direction. Bonding between successive layers isby van der Waals forces and by cations that balancecharge deficiencies in the structure. These bonds areweak and easily separated by cleavage or adsorptionof water or other polar liquids. The basal spacing inthe c direction, d(001), is variable, ranging from about9.6 A to complete separation.

The theoretical composition in the absence ofisomorphous substitutions is (OH)4Si8Al4O20 �n(interlayer)H2O. The structural configuration and cor-responding charge distribution are shown in Fig. 3.24.The structure shown is electrically neutral, and theatomic configuration is essentially the same as that inthe nonclay mineral pyrophyllite.

Isomorphous Substitution in the Smectite Minerals

Smectite minerals differ from pyrophyllite in that thereis extensive isomorphous substitution for silicon andaluminum by other cations. Aluminum in the octahe-dral sheet may be replaced by magnesium, iron, zinc,nickel, lithium, or other cations. Aluminum may re-place up to 15 percent of the silicon ions in the tetra-hedral sheet. Possibly some of the silicon positions canbe occupied by phosphorous (Grim, 1968).

Substitutions for aluminum in the octahedral sheetmay be one-for-one or three-for-two (aluminum oc-cupies only two-thirds of the available octahedral sites)in any combination from a few to complete replace-ment. The resulting structure, however, is either almostexactly dioctahedral (montmorillonite subgroup) ortrioctahedral (saponite subgroup). The charge defi-ciency resulting from these substitutions ranges from0.5 to 1.2 per unit cell. Usually, it is close to 0.66 perunit cell. A charge deficiency of this amount wouldresult from replacement of every sixth aluminum by amagnesium ion. Montmorillonite, the most commonmineral of the group, has this composition. Charge de-ficiencies that result from isomorphous substitution arebalanced by exchangeable cations located between theunit cell layers and on the surfaces of particles.

Some minerals of the smectite group and their com-positions are listed in Table 3.5. An arrow indicates thesource of the charge deficiency, which has been as-sumed to be 0.66 per unit cell in each case. Sodium isindicated as the balancing cation. The formulas shouldbe considered indicative of the general character of themineral, but not as absolute, because a variety of com-positions can exist within the same basic crystal struc-ture. Because of the large amount of unbalanced

Copy

right

ed M

ater

ial



Figure 3.22 Schematic diagrams of the structures of the smectite minerals: (a) montmoril-lonite and (b) saponite.

Figure 3.23 Diagrammatic sketch of the montmorillonitestructure.

Figure 3.24 Charge distribution in pyrophyllite (type struc-ture for montmorillonite).

substitution in the smectite minerals, they have highcation exchange capacities, generally in the range of80 to 150 meq/100 g.


Montmorillonite may occur as equidimensional flakesthat are so thin as to appear more like films, as shown

in Fig. 3.25. Particles range in thickness from 1-nmunit layers upward to about 1/100 of the width. Thelong axis of the particle is usually less than 1 or 2 �m.When there is a large amount of substitution of ironand/or magnesium for aluminum, the particles may belath or needle shaped because the larger Mg2� and Fe3�

ions cause a directional strain in the structure.The specific surface area of smectite can be very

large. The primary surface area, that is, the surface areaexclusive of interlayer zones, ranges from 50 to 120m2/g. The secondary specific surface that is exposed

Copy

right

ed M

ater

ial


SMECTITE MINERALS 61

Table 3.5 Some Minerals of the Smectite Group

MineralTetrahedral Sheet

SubstitutionsOctahedral Sheet

Substitutions Formula/Unit Cella

Dioctahedral, Smectites orMontmorillonites

Montmorillonite None 1Mg2� for every sixth Al3� (OH)4Si8(Al3.34Mg0.66) O20

↓Na0.66

Beidellite Al for Si None (OH)4(Si6.34Al1.66) Al4.34O20

↓Na0.66

Nontronite Al for Si Fe3� for Al (OH)4(Si7.34Al0.66) Fe43�O20

↓Na0.66

Trioctahedral, Smectites,or Saponites

Hectorite None Li for Mg (OH)4Si8(Mg5.34Li0.66) O20

↓Na0.66

Saponite Al for Si Fe3� for Mg (OH)4(Si7.34Al0.66) Mg6O20

↓Na0.66

Sauconite Al for Si Zn for Mg (OH)4(Si8�yAly)(Zn6�xMgx) O20

↓Na0.66

aTwo formula units are needed to give one unit cell.After Ross and Hendricks (1945); Marshall (1964); and Warshaw and Roy (1961).

Figure 3.25 Electron photomicrograph of montmorillonite(bentonite) from Clay Spur, Wyoming. Picture width is 7.5�m (Tovey, 1971).

by expanding the lattice so that polar molecules canpenetrate between layers can be up to 840 m2/g.

Bentonite

A very highly plastic, swelling clay material known asbentonite is very widely used for a variety of purposes,ranging from drilling mud and slurry walls to clarifi-cation of beer and wine. The bentonite familiar to mostgeoengineers is a highly colloidal, expansive alterationproduct of volcanic ash. It has a liquid limit of 500percent or more. It is widely used as a backfill duringthe construction of slurry trench walls, as a soil ad-mixture for construction of seepage barriers, as a groutmaterial, as a sealant for piezometer installations, andfor other special applications.

When present as a major constituent in soft shale oras a seam in rock formations, bentonite may be a causeof continuing slope stability problems. Slide problemsat Portuguese Bend along the Pacific Ocean in southernCalifornia, in the Bearpaw shale in Saskatchewan, andin the Pierre shale in South Dakota are in large mea-

Copy

right

ed M

ater

ial



Figure 3.26 Schematic diagram of the structures of muscovite, illite, and vermiculite: (a)muscovite and illite and (b) vermiculite.

sure due to the high content of bentonite. Stabilityproblems in underground construction may be causedby the presence of montmorillonite in joints and faults(Brekke and Selmer-Olsen, 1965).

3.18 MICALIKE CLAY MINERALS

Illite is the most commonly found clay mineral in soilsencountered in engineering practice. Its structure isquite similar to that of muscovite mica, and it is some-times referred to as hydrous mica. Vermiculite is alsooften found as a clay phase constituent of soils. Itsstructure is related to that of biotite mica.

Structure

The basic structural unit for the muscovite (white mica)is shown schematically in Fig. 3.26a. It is the three-layer silica–gibbsite–silica sandwich that forms pyro-phyllite, with the tips of all the tetrahedra pointingtoward the center and common with octahedral sheetions. Muscovite differs from pyrophyllite, however, inthat about one-fourth of the silicon positions are filledby aluminum, and the resulting charge deficiency isbalanced by potassium between the layers. The layersare continuous in the a and b directions and stacked inthe c direction. The radius of the potassium ion, 1.33A, is such that it fits snugly in the 1.32 A radius holeformed by the bases of the silica tetrahedra. It is in 12-fold coordination with the 6 oxygens in each layer.

A diagrammatic three-dimensional sketch of themuscovite structure is shown in Fig. 3.27. The struc-tural configuration and charge distribution are shown

in Fig. 3.28. The unit cell is electrically neutral andhas the formula (OH)4K2(Si6Al2)Al4O20. Muscovite isthe dioctahedral end member of the micas and containsonly Al3� in the octahedral layer. Phlogopite (brownmica) is the trioctahedral end member, with the octa-hedral positions filled entirely by magnesium. It hasthe formula (OH)4K2(Si6Al2)Mg6O20. Biotite (blackmica) is trioctahedral, with the octahedral positionsfilled mostly by magnesium and iron. It has the generalformula (OH)4K2(Si6Al2)(MgFe)6O20. The relative pro-portions of magnesium and iron may vary widely.

Illite differs from mica in the following ways (Grim,1968):

1. Fewer of the Si4� positions are filled by Al3� inillite.

2. There is some randomness in the stacking of lay-ers in illite.

3. There is less potassium in illite. Well-organizedillite contains 9 to 10 percent K2O (Weaver andPollard, 1973).

4. Illite particles are much smaller than mica parti-cles.

Some illite may contain magnesium and iron in theoctahedral sheet as well as aluminum (Marshall, 1964).Iron-rich illite, usually occurring as earthy green pel-lets, is termed glauconite.

The vermiculite structure consists of regular inter-stratification of biotite mica layers and double molec-ular layers of water, as shown schematically in Fig.3.26b. The actual thickness of the water layer dependson the cations that balance the charge deficiencies in

Copy

right

ed M

ater

ial


MICALIKE CLAY MINERALS 63

Figure 3.27 Diagrammatic sketch of the structure of muscovite.

Figure 3.28 Charge distribution in muscovite.

the biotitelike layers. With magnesium or calciumpresent, which is the usual case in nature, there aretwo water layers, giving a basal spacing of 14 A. Ageneral formula for vermiculite is

(OH)4(MgCa)x(Si8xAlx)(MgFe)6O20yH2O

x � 1 to 1.4 y � 8

Isomorphous Substitution and Exchange Capacity

There is extensive isomorphous substitution in illiteand vermiculite. The charge deficiency in illite is 1.3to 1.5 per unit cell. It is located primarily in the silicasheets and is balanced partly by the nonexchangeablepotassium between layers. Thus, the cation exchangecapacity of illite is less than that of smectite, amount-ing to 10 to 40 meq/100 g. Values greater than 10 to15 meq/100 g may be indicative of some expandinglayers (Weaver and Pollard, 1973). In the absence offixed potassium the exchange capacity would be about150 meq/100 g. Interlayer bonding by potassium is sostrong that the basal spacing of illite remains fixed at10 A in the presence of polar liquids.

The charge deficiency in vermiculite is 1 to 1.4 perunit cell. Since the interlayer cations are exchangeable,the exchange capacity of vermiculite is high, amount-

Copy

right

ed M

ater

ial



Figure 3.30 Schematic diagram of the structure of chlorite.

Figure 3.29 Electron photomicrograph of illite from Morris,Illinois. Picture width is 7.5 �m (Tovey, 1971).

ing to 100 to 150 meq/100 g. The basal spacing, d(001),is influenced by both the type of cation and the hy-dration state. With potassium or ammonium in theexchange positions, the basal spacing is only 10.5 to11 A. Lithium gives 12.2 A. Interlayer water can bedriven off by heating to temperatures above 100�C.This dehydration is accompanied by a reduction inbasal spacing to about 10 A. The mineral quickly re-hydrates and expands again to 14 A when exposed tomoist air at room temperature.


Illite usually occurs as very small, flaky particlesmixed with other clay and nonclay materials. High-purity deposits of illite are uncommon. The flaky par-ticles may have a hexagonal outline if well crystallized.The long axis dimension ranges from 0.1 �m or lessto several micrometers, and the plate thickness may beas small as 3 nm. An electron photomicrograph of illiteis shown in Fig. 3.29. Vermiculite may occur in natureas large crystalline masses having a sheet structuresomewhat similar in appearance to mica. In soils, ver-miculite occurs as small particles mixed with otherclay minerals.

The specific surface area of illite is about 65 to 100m2/g. The primary surface of vermiculites is 40 to 80m2/g, and the secondary (interlayer) surface may be ashigh as 870 m2/g.

3.19 OTHER CLAY MINERALS

Chlorite Minerals

Structure The chlorite structure consists of alter-nating micalike and brucitelike layers as shown sche-matically in Fig. 3.30. The structure is similar to that

of vermiculite, except that an organized octahedralsheet replaces the double water layer between micalayers. The layers are continuous in the a and b direc-tions and stacked in the c direction. The basal spacingis fixed at 14 A.

Isomorphous Substitution The central sheet of themica layer is trioctahedral, with magnesium as the pre-dominant cation. There is often partial replacement ofMg2� by Al3�, Fe2� and Fe3�. There is substitution ofAl3� for Mg2� in the brucitelike layer. The variousmembers of the chlorite group differ in the kind andamounts of substitution and in the stacking of succes-sive layers. The cation exchange capacity of chloritesis in the range of 10 to 40 meq/100 g.

Morphology Chlorite minerals occur as micro-scopic grains of platy morphology and poorly definedcrystal edges in altered igneous and metamorphic rocksand their derived soils. In soils, chlorites always appearto occur in mixtures with other clay minerals.

Chain Structure Clay Minerals

A few clay minerals are formed from bands (doublechains) of silica tetrahedra. These include attapulgiteand imogolite. They have lathlike or fine threadlikemorphologies, with particle diameters of 5 to 10 nmand lengths up to 4 to 5 �m. An electron photomicro-graph of bundles of attapulgite particles is shown inFig. 3.31.

Although these minerals are not frequently encoun-tered, attapulgite is commercially mined and is used asa drilling mud in saline and other special environmentsbecause of its high stability in suspensions.

Copy

right

ed M

ater

ial


DETERMINATION OF SOIL COMPOSITION 65

Figure 3.31 Electron photomicrograph of attapulgite fromAttapulgis, Georgia. Picture width is 4.7 �m (Tovey, 1971).

Mixed-Layer Clays

More than one type of clay mineral is usually foundin most soils. Because of the great similarity in crystalstructure among the different minerals, interstratifica-tion of two or more layer types often occurs within asingle particle. Interstratification may be regular, witha definite repetition of the different layers in sequence,or it may be random. According to Weaver and Pollard(1973), randomly interstratified clay minerals are sec-ond only to illite in abundance.

The most abundant mixed-layer material is com-posed of expanded water-bearing layers and contractednon-water-bearing layers. Montmorillonite–illite ismost common, and chlorite–vermiculite and chlorite–montmorillonite are often found. Rectorite is an inter-stratified clay with high charge, micalike layers withfixed interlayer cations alternating in a regular mannerwith low-charge montmorillonite-like layers containingexchangeable cations capable of hydration.

Noncrystalline Clay Materials

Allophane Clay materials that are so poorly crys-talline that a definite structure cannot be determinedare termed allophane. Such material is amorphous toX-rays because there is insufficient long-range order ofthe octahedral and tetrahedral units to produce sharpdiffraction effects, although in some cases there maybe diffraction bands. Allophane has no definite com-position or shape and may exhibit a wide range ofphysical properties. Some noncrystalline clay materialis probably contained in all fine-grained soils. It iscommon in volcanic soils because of the abundance ofglass particles.

Oxides All soils probably contain some amount ofcolloidal oxides and hydrous oxides (Marshall, 1964).The oxides and hydroxides of aluminum, silicon, andiron are most frequently found. These materials mayoccur as gels or precipitates and coat mineral particles,or they may cement particles together. They may alsooccur as distinct crystalline units; for example, gibb-site, boehmite, hematite, and magnetite. Limonite andbauxite, which are noncrystalline mixtures of iron andaluminum hydroxides, are also sometimes found.

Oxides are particularly common in soils formedfrom volcanic ash and in tropical residual soils. Somesoils rich in allophane and oxides may exhibit signif-icant irreversible decreases in plasticity and increasesin strength when dried. Many are susceptible to break-down and strength loss when subjected to traffic ormanipulation during earthwork construction (Mitchelland Sitar, 1982; Mitchell and Coutinho, 1991).

3.20 SUMMARY OF CLAY MINERALCHARACTERISTICS

The important structural, compositional, and morpho-logical characteristics of the important clay mineralsare summarized in Table 3.6. Data on the structuralcharacteristics of the tetrahedral and octahedral sheetstructures are included.

3.21 DETERMINATION OF SOILCOMPOSITION

Introduction

Identification of the fine-grained minerals in a soil isusually done by X-ray diffraction. Simple chemicaltests can be used to indicate the presence of organicmatter and other constituents. The microscope may beused to identify the constituents of the nonclay frac-tion. Accurate determination of the proportions of dif-ferent mineral, organic, and amorphous solid materialin a soil, while probably possible with the expenditureof great time and at great cost, is unlikely to be worth-while owing to our inability to make exact quantitativelinks from composition to properties. Accordingly,from knowledge of grain size distribution, the relativeintensities of different X-ray diffraction peaks, and afew other simple tests a semiquantitative analysis maybe made that is usually adequate for most purposes.

A general approach is given in this section for thedetermination of soil composition, some of the tech-niques are described briefly, and criteria for identifi-cation of important soil constituents are stated.

Copy

right

ed M

ater

ial



Table 3.6 Summary of Clay Mineral Characteristics

Structural1. Silica Tetrahedron: Si atom at center. Tetrahedron units form hexagonal network � Si4O8(OH)4

2. Gibbsite Sheet: Aluminum in octahedral coordination. Two-thirds of possible positions filled. Al2(OH)—O—O � 2.60 A.3. Brucite Sheet: Magnesium in octahedral coordination. All possible positions filled. Mg2(OH)—O—O � 2.60 A.

TypeSubgroup and

Schematic Structure Mineral Complete Formula / Unit CellaOctahedral Layer

CationsTetrahedral Layer

Cations

Structure

Isomorphous Substitution Interlayer Bond

Allophane Allophanes Amorphous — —

Kaolinite Kaolinite (OH)8Si4Al4O11 Al4 Si4 Little O—OHHydrogen Strong

1�1

Dickite

Nacrite

Halloysite(dehydrated)

Halloysite(hydrated)

(OH)8Si4Al4O10

(OH)8Si4Al4O10

(OH)8Si4Al4O10

(OH)8Si4Al4O10 � 4H2O

Al4

Al4

Al4

Al4

Si4

Si4

Si4

Si4

Little

Little

Little

Little

O—OHHydrogen Strong

O—OHHydrogen StrongO—OHHydrogen StrongO—OHHydrogen Strong

Montmorillonite(OH)4Si8Al4O20 � NH2O(Theoretical

Unsubsitituted)

Montmorillonite (OH)4Si8(Al3.34Mg.66O20nH2O↓ *Na.66

Al3.34Mg.66 Si8 Mg for Al, Net chargealways � 0.66- / unitcell

O—OVery weak

expanding lattice

Beidellite

Nontronite

(OH)4(Si7.34Al66)(Al4)O20nH2O↓

Na.66

(OH)4(Si7.34Al.66)Fe43�O20nH2O

↓Na.66

Al4

Fe4

Si7.34Al.66

Si7.34Al.66

Al for Si, Net chargealways � 0.66- / forunit cell

Fe for Al, Al for Si, Netcharge always � 0.66-/ for unit cell

O—OVery weak

expanding latticeO—OVery weak

expanding lattice

2�1 Saponite Hectorite

Saponite

Sauconite

(OH)4Si8(Mg5.34Li.66)P20nH2O↓

Na.66

(OH)4(Si7.34Al.66)Mg6O20nH2O↓Na.66

(Si6.94Al1.06)Al.66Fe.34Mg.36Zn4.80O20(OH)4

↓ � nH2ONa.66

Mg5.34Li.66

Mg, Fe3�

Al.44Fe.34Mg.36Zn4.80

Si8

Si7.34Al.66

Si6.94Al1.06

Mg, Li for Al, Netcharge always � 0.66-/ unit cell

Mg for Al, Al for Si,Net charge always �0.66- / for unit cell

Zn for Al

O—OVery weak



expanding lattice

Hydrous Mica (Illite) Illites (K, H2O)2(Si)8(Al,Mg,Fe)4,6O20(OH)4 (Al,Mg,Fe)4-6 (Al,Si)8 Some Si always replacedby Al, Balanced by Kbetween layers.

K ions; strong

Vermiculite Vermiculite (OH)4(Mg,Ca)x(Si8�xAlx)(Mg.Fe)6O20.yH2Ox � 1 to 1.4, y � 8

(Mg,Fe)6 (Si,Al)8 Al for Si not charge of 1to 1.4 / unit cell

Weak

2�1�1 Chlorite Chlorite(Several varieties

known)

(OH)4(SiAl)8(Mg.Fe)6O20 (2�1 layer)(MgAl)6(OH)12 interlayer

(Mg,Fe)6(2�1 layer)(Mg,Al)6 interlayer

(Si,Al)8 Al for Si in 2�1 layerAl for Mg in interlayer

ChainStructure

Sepiolite

Attapulgite

Si4O11(Mg.H2)3H2O2(H2O)

(OH2)4(OH)2Mg5Si8O20.4H2O

Fe or Al for Mg

Some for Al for Si Weak � chainslinked by 0

a Arrows indicate source of charge deficiency. Equivalent Na listed as balancing cation. Two formula units (Table 3.4) are required per unit cell.b Electron microscope data.

Copy

right

ed M

ater

ial



Table 3.6 (Continued )

UnitsAll bases in same plane. O—O � 2.55 A—Space for Si � 0.55 A—Thickness8 4.93 A. C—C height � 2.1 A.OH—OH � 2.94 A. Space for ion � 0.61 A. Thickness of unit � 5.05 A. Dioctahedral.OH—OH � 2.94 A. Space for ion � 0.61 A. Thickness of unit � 5.05 A. Trioctahedral.

Structure

Crystal Structure Basal Spacing Shape SizebCation ExchangeCap.(meq / 100 g)

SpecificGravity

Specific Surfacem2 / g

Occurrence in Soilsof Engineering

Interest

Irregular, some-what rounded

0.05–1 �Common

Triclinica � 5.14, b � 8.93, c � 7.37� � 91.6�, � � 104.8�, � 89.9�

7.2 A 6-sided flakes 0.1–4 � �single�0.05–2 �

to 3000 � 4000(stacks)

3–15 2.60–2.68 10–20 Very common

Monoclinica � 5.15, b � 8.95, c � 14.42� � 96�48�

14.4 A Unit cellcontains 2unit layers

6-sided flakes 0.07–300 � 2.5–1000 �

1–30 Rare

Almost Orthorhombica � 5.15, b � 8.96, c � 43� � 90�20�a � 5.14 in O Planea � 5.06 in OH Planeb � 8.93 in O Planeb � 8.62 in OH Plane� layers curve

43 A

7.2 A

10.1 A

Unit cellcontains 6unit layers

Randomstacking ofunit cells

Water layerbetween unitcells

Rounded flakes

Tubes

Tubes

1 � � 0.025–0.15 �

0.07 � O.D.0.04 � I.D.1 � long.

5–10

5–40

2.55–2.56

2.0–2.2 35–70

Rare

Occasional

Occasional

9.6A—Completeseparation

Dioctahedral Flakes (equi-dimensional)

�10 A � up to10 �

80–150 2.35–2.7 50–120 Primary700–840 Secondary

Very common


Dioctahedral Rare


Dioctahedral Laths Breadth � 1 / 5length toseveral � �unit cell

110–150 2.2–2.7 Rare


Trioctahedral To 1 � � unitcell breadth �0.02 � 0.1�

17.5 Rare

Trioctahedral Similar tomont.

Similar to mont. 70–90 2.24–2.30 Rare

Trioctahedral Brand laths 50 A Thick Rare

10 A Bothdioctrahedralandtrioctahedral

Flakes 0.003–0.1 � �up to 10 �

10–40 2.6–3.0 65–100 Very common

a � 5.34, b � 9.20c � 28.91, � � 93�15�

10.5–14 A AlternatingMica anddouble H2Olayers

Similar to illite 100–150 40–80 Primary870 Secondary

Fairly common

Monoclinic (Mainly)a � 5.3, b � 9.3c � 28.52, � � 97�8�

14 A Similar to illite 1 � 10–40 2.6–2.96 Common

Monoclinica � 2 � 11.6, b � 2 � 7.86c � 5.33a0 Sin � � 12.9 b0 � 18c0 � 5.2

Chain

Double silicachains

Flakes or fibers

Laths Max, 4–5 � �50–100 A

Width � 2t

20–30

20–30

2.08 Rare

Occasional

From Grim, R. E. (1968) Clay Mineralogy, 2d edition, McGraw-Hill, New York. Brown, G. (editor) (1961) The X-ray Identification and Crystal Structure of Clay Materials, MineralogicalSociety (Clay Minerals Group), London.

Copy

right

ed M

ater

ial



Methods for Compositional Analysis

Methods and techniques that may be employed for de-termination of soil composition and study of soil grainsinclude:

1. Particle size analysis and separation2. Various pretreatments prior to mineralogical

analysis3. Chemical analyses for free oxides, hydroxides,

amorphous constituents, and organic matter4. Petrographic microscope study of silt and sand

grains5. Electron microscope study6. X-ray diffraction for identification of crystalline

minerals7. Thermal analysis8. Determination of specific surface area9. Chemical analysis for layer charge, cation

exchange capacity, exchangeable cations, pH,and soluble salts

10. Staining tests for identification of clays

Procedures for determination of soil composition aredescribed in detail in publications of the American So-ciety of Agronomy. Part 1—Physical and Mineralog-ical Methods provides a set of procedures formineralogical analyses for use by soil scientists andengineers. Part 2—Microbiological and BiochemicalProperties, published in 1994, is useful for determi-nations needed for bioremediation and other geoen-vironmental purposes. Part 3—Chemical Methods,published in 1996 contains methods for characterizingsoil chemical properties as well as several methods forcharacterizing soil chemical processes. Part 4—Physical Methods, published in 2002, is an updatedversion of the physical methods covered in Part 1. Foreach method, principles are presented as well as thedetails of the method. In addition, the interpretation ofresults is discussed, and extensive bibliographies aregiven.

Accuracy of Compositional Analysis

Techniques for chemical analysis are generally of ahigh order of accuracy. However, this accuracy doesnot extend to the overall compositional analysis of asoil in terms of components of interest in understand-ing and quantifying behavior. This is because knowl-edge of the chemical composition of a soil is of limitedvalue by itself. Chemical analysis of the solid phase ofa soil does not indicate the organization of the ele-ments into crystalline and noncrystalline components.

For quantitative mineralogical analysis of the clayfraction, it is usually necessary to assume that the

properties of the mineral in the soil are the same asthose of a reference mineral. However, different sam-ples of any given clay mineral may exhibit significantdifferences in composition, surface area, particle sizeand shape, and cation exchange capacity. Thus, selec-tion of ‘‘standard’’ minerals for reference is arbitrary.Quantitative clay mineral determinations cannot bemade to an accuracy of more than about plus or minusa few percent without exhaustive chemical and min-eralogical tests.

General Scheme for Compositional Analysis

A general scheme for determination of the componentsof a soil is given in Fig. 3.32. Techniques of the mostvalue for qualitative and semiquantitative analysis areindicated by a double asterisk, and those of particularuse for explaining unusual properties are indicated bya single asterisk. The scheme shown is by no meansthe only one that could be used; a feedback approachis desirable wherein the results of each test are used toplan subsequent tests. Brief discussions of the varioustechniques listed in Fig. 3.32 are given below. X-raydiffraction analysis is treated in more detail in the nextsection because of its particular usefulness for theidentification of fine-grained soil minerals.

Grain Size Analysis Determination of particle sizeand size distribution is usually done using sieve anal-ysis for the coarse fraction [sizes greater than 74 �m(i.e., 200 mesh sieve)] and by sedimentation methodsfor the fine fraction. Details of these methods are pre-sented in standard soil mechanics texts and in the stan-dards of the American Society for Testing andMaterials (ASTM). Determination of sizes by sedi-mentation is based on the application of Stokes’s lawfor the settling velocity of spherical particles:

� s w 2v � D (3.2)18�

where s � unit weight of particle, w � unit weightof liquid, � � viscosity of liquid, and D � diameterof sphere. Sizes determined by Stoke’s law are not ac-tual particle diameters but, rather, equivalent sphericaldiameters. Gravity sedimentation is limited to particlesizes in the range of about 0.2 mm to 0.2 �m, theupper bound reflecting the size limit where flow aroundthe particles is no longer laminar, and the lower boundrepresenting a size where Brownian motion keeps par-ticles in suspension indefinitely.

The times for particles of 2, 5, and 20 �m equivalentspherical diameter to fall through water a distance of10 cm are about 8 h, 1.25 h, and 5 min, respectively,at 20�C. At 30�C the required times are about 6.5 h, 1

Copy

right

ed M

ater

ial



Figure 3.32 Flow sheet for compositional analysis of soils (adapted from Lambe and Martin,1954).

h, and 4 min. A centrifuge can be used for acceleratingthe settlement of small particles and is the most prac-tical means for extracting particles smaller than abouta micrometer in size.

Sedimentation methods call for treatment of a soil–water suspension with a dispersing agent and thoroughmixing prior to the start of the test. This causes break-down of aggregates of soil particles, and the degree ofbreakdown may vary greatly with the method of prep-aration. For example, the ASTM standard method oftest permits the use of either an air dispersion cup ora blender-type mixer. The amount of material less than2 �m equivalent spherical diameter may vary by asmuch as a factor of 2 by the two techniques. The re-lationship between the size distribution that results

from laboratory preparation of the sample to that ofthe particles and aggregates in the natural soil is un-known.

Optical and electron microscopes are sometimesused to study particle sizes and size distributions andto provide information on particle shape, aggregation,angularity, weathering, and surface texture.

Pore Fluid Electrolyte The total concentration ofsoluble salts may be determined from the electricalconductivity of extracted pore fluid. Chemical or pho-tometric techniques may be used to determine the el-emental constituents of the extract (Rhoades, 1982).Removal of excess soluble salts by washing the samplewith water or alcohol may be necessary before pro-ceeding with subsequent analysis. If they are not re-

Copy

right

ed M

ater

ial



moved, the soil may be difficult to disperse, it may bedifficult to remove organic matter, reliable cationexchange capacity determinations will be impossible,and mineralogical analyses will be complicated (Kunzeand Dixon, 1986).

pH Determination of the acidity or alkalinity of asoil in terms of the pH is a relatively simple measure-ment that can be made using a pH meter or specialindicators (American Society for Testing and Materi-als, 1970; McLean, 1982). The value obtained dependson the ratio of soil to water, so it is usual to standardizethe measurement using a 1�1 ratio of soil to water byweight. For highly plastic soils a lower soil-to-waterratio may be required to produce a suspension suitablefor pH measurement. The pH decreases with increas-ing concentration of neutral salts in solution and withincreasing amounts of dissolved CO2.

Carbonates Carbonates, in the form of calcite(CaCO3), dolomite [CaMg(CO3)2], marl, and shells arefrequently found in soils, and they can be readily de-tected by effervescence when the soil is treated withdilute HCl. Many methods for determining inorganiccarbonates, calcite, and dolomite in soils are available(Nelson, 1982). These include dissolution in acid, dif-ferential thermal analysis, X-ray diffraction, and chem-ical analyses.

Gypsum Gypsum (CaSO4 � 2H2O) can be deter-mined by a simple heating test. Visible grains will turnwhite when heated on a metal plate as a result of de-hydration to form ‘‘dead-burnt gypsum’’ (Shearman,1979). Quantitative determinations can be made usingprocedures described by Nelson (1982).

Organic Matter Organic matter can be readily de-tected by treatment of the soil with a 15 percent hy-drogen peroxide solution. H2O2 reacts with organicmatter to give vigorous effervescence. As organic mat-ter has an aggregating effect, and because its presencemay interfere with other mineralogical analyses, it isdesirable to remove most of it by digestion with H2O2

(Kunze and Dixon, 1986). Quantitative analysis meth-ods for soil organic matter are given by the AmericanSociety for Testing and Materials (1970), Nelson andSommers (1982), and Schnitzer (1982).

Oxides and Hydroxides Free oxides and hydrox-ides that may be present in soils include crystalline andnoncrystalline (amorphous) compounds of silicon, alu-minum, and iron. These materials may occur as dis-crete particles, as coatings on particles, and ascementing agents between particles. They may makesoil dispersion difficult, and they may interfere withother analysis procedures. Methods for oxide and hy-droxide detection, quantitative analysis, and removalare given by Jackson et al. (1986).

Exchange Complex Determination of the cationexchange capacity (expressed in milliequivalents perhundred grams of dry soil) is made after first freeingthe soil of excess soluble salts. The adsorbed cationsare then replaced by a known cation species, and theamount of the known cation needed to saturate theexchange sites is determined analytically (Rhoades,1982). The composition of the original cation complexcan be determined by chemical analysis of the originalextract (Thomas, 1982).

Potash The hydrous mica minerals (illites) are theonly minerals commonly found in the clay size fractionof soils that contain potassium in their crystal structure.Thus, knowledge of the K2O content is useful for quan-titative determination of their abundance. A method forpotassium determination is given by Knudsen et al.(1986). Well-organized 10-A illite layers contain 9 to10 percent K2O (Weaver and Pollard, 1973).

Specific Surface Area Ethylene glycol and glyceroladsorb on clay surfaces. As different clay mineralshave different values of specific surface, the amount ofglycol or glycerol retained under controlled conditionscan be used to aid in the quantitative determinationsof clay minerals and for estimation of specific surfacearea (Martin, 1955; Diamond and Kinter, 1956; andAmerican Society for Testing and Materials, 1970).

Use of ethylene glycol monoethyl ether (EGME) asthe polar molecule for determining surface area offersthe advantages of the attainment of adsorption equilib-rium more rapidly and with greater precision (Carteret al., 1982). A monomolecular layer of EGME is as-sumed to form in vacuum on a predried clay sample.The weight of EGME adsorbed after equilibrium isreached is converted to specific surface using a factorof 0.000286 g EGME per square meter of surface.

3.22 X-RAY DIFFRACTION ANALYSIS

X-Rays and Their Generation

X-ray diffraction is the most widely used method foridentification of fine-grained soil minerals and thestudy of their crystal structure. X-rays are one of sev-eral types of waves in the electromagnetic spectrum(Fig. 3.2). X-rays have wavelengths in the range of0.01 to 100 A. When high-speed electrons impinge ona target material, one of two phenomena may occur:

1. The high-speed electron strikes and displaces anelectron from an inner shell of one of the atomsof the target material. An electron from one ofthe outer shells then falls into the vacancy tolower the energy state of the atom. An X-rayphoton of wavelength and intensity characteristic

Copy

right

ed M

ater

ial


X-RAY DIFFRACTION ANALYSIS 71

Figure 3.35 Composite relationship for X-ray intensity as afunction of wavelength.

Figure 3.34 X-ray generation by deceleration of electronsin an electric field.

Figure 3.33 X-ray generation by electron displacement. Let-ters designate shells in which electron transfer takes place.

of the target atom and of the particular electronicpositions is emitted. Because electronic transfersmay take place in several shells and each has acharacteristic frequency, the result is a relation-ship between radiation intensity and wavelengthas shown in Fig. 3.33.

2. The high-speed electron does not strike an elec-tron in the target material but slows down in theintense electric fields near atomic nuclei. The de-crease in energy is converted to heat and to X-ray photons. X-rays produced in this way areindependent of the nature of the bombarded at-oms and appear as a band of continuously vary-ing wavelength as shown in Fig. 3.34.

The resulting output of X-rays from these two ef-fects acting together is shown in Fig. 3.35. X-rays aregenerated using a tube in which electrons stream froma filament to a target material across a voltage drop of20 to 50 kV. Curved crystal monochrometers can beused to give X-rays of a single wavelength. Alterna-tively, certain materials are able to absorb X-rays ofdifferent wavelengths, so it is possible to filter the out-put of an X-ray tube to give rays of only one wave-

length. The wavelengths of monochromatic radiation(usually K�, Fig. 3.33) produced from commonly usedtarget materials range from 0.71 A for molybdenum to2.29 A for chromium. Copper radiation, which is mostfrequently used for mineral identification, has a wave-length of 1.54 A.

Diffraction of X-rays

Because wavelengths of about 1 A are of the sameorder as the spacing of atomic planes in crystallinematerials, X-rays are useful for analysis of crystalstructures. When X-rays strike a crystal, they penetrateto a depth of several million layers before being ab-sorbed. At each atomic plane a minute portion of thebeam is absorbed by individual atoms that then oscil-late as dipoles and radiate waves in all directions. Ra-diated waves in certain directions will be in phase andcan be interpreted in simplistic fashion as a wave re-sulting from a reflection of the incident beam. In-phaseradiations emerge as a coherent beam that can be de-tected on film or by a radiation counting device. Theorientation of parallel atomic planes, relative to the di-rection of the incident beam, at which radiations are inphase depends on the wave length of the X-rays andthe spacing between atomic planes.

Figure 3.36 shows a parallel beam of X-rays ofwavelength � striking a crystal at an angle � to parallelatomic planes spaced at distance d. If the reflectedwave from C is to reinforce the wave reflected fromA, then the path length difference between the twowaves must be an integral number of wave lengths n�.From Fig. 3.36, this difference is distance BC � CD.Thus,

BC � CD � n�

Copy

right

ed M

ater

ial



Figure 3.36 Geometrical conditions for X-ray diffraction according to Bragg’s law.

From symmetry, BC � CD, and by trigonometry,CD � d sin �. Thus the necessary condition is givenby

n� � 2d sin � (3.3)

This is Bragg’s law. It forms the basis for identificationof crystals using X-ray diffraction. Since no two min-erals have the same spacings of interatomic planes inthree dimensions, the angles at which diffractions oc-cur (and the atomic spacings calculated from them) canbe used for identification. X-ray diffraction is partic-ularly well suited for identification of clay mineralsbecause the (001) spacing is characteristic for eachclay mineral group. The basal planes generally give themost intense reflections of any planes in the crystalsbecause of the close packing of atoms in these planes.The common nonclay minerals occurring in soils arealso detectable by X-ray diffraction.

Detection of Diffracted X-rays

Because the small size of most soil particles preventsthe study of single crystals, use is made of the powdermethod and of oriented aggregates of particles. In thepowder method, a small sample containing particles atall possible orientations is placed in a collimated beamof parallel X-rays, and diffracted beams of various in-tensities are scanned by a Geiger, proportional, or scin-tillation tube and recorded automatically to produce achart showing the intensity of diffracted beam as afunction of angle 2�. As an example, the diffractionpattern for quartz is shown in Fig. 3.37. The powdermethod works because the very large number of par-ticles in a sample ensures that some will always beproperly oriented to produce a reflection.

All prominent atomic planes in a crystal will pro-duce a reflection if properly positioned with respect to

the X-ray beam. Thus, each mineral will produce acharacteristic set of reflections at values of � corre-sponding to the interatomic spacings between theprominent planes. The intensities of the different re-flections vary according to the density of atomic pack-ing and other factors.

When the oriented aggregate method is used, platyclay particles are precipitated onto a glass slide, usu-ally by drying from a deflocculated suspension or sep-arated from a suspension on a porous ceramic plate.With most particles oriented parallel to the slide, the(001) reflections are intensified, whereas reflectionsfrom (hk0) planes are minimized.

In the Bragg equation, n may be any whole number.The reflection corresponding to n � 1 is termed thefirst-order reflection. If the first-order reflection for amineral gives d(001) � 10 A, then for n � 2 there canbe a reflection at 5 A, for n � 3 there can be a reflec-tion at 3.33 A, and so on. It is common to refer tothese as higher-order reflections due to the (002) plane,the (003) plane, and so on, even though atomic planesdo not exist at these spacings. They are, in reality, val-ues of d /n � � / (2 sin �) for integer values of n � 1.

Analysis of X-ray Patterns

A complete X-ray diffraction pattern consists of a se-ries of reflections of different intensities at differentvalues of 2�. Each reflection must be assigned to somecomponent of the sample. The first step in the analysisis to determine all values of d/n for the particular typeof radiation (which determines �) using Eq. (3.3). Thetest pattern may be compared directly with patterns forknown materials. The American Society for Testingand Materials maintains a file of patterns for manymaterials indexed on the basis of the strongest lines inthe pattern. X-ray diffraction data for the clay mineralsand other common soil minerals are given in Grim

Copy

right

ed M

ater

ial


X-RAY DIFFRACTION ANALYSIS 73

Figure 3.37 X-ray diffractometer chart for quartz. Peaks occur at specific 2� angles, whichcan be converted to d spacings by Bragg’s law. Numbers in parentheses are the Miller indicesfor the crystal planes responsible for the indicated peak.

Copy

right

ed M

ater

ial



(1968), Carroll (1970), Brindley and Brown (1980),Whittig and Allardice (1986), and Moore and Reyn-olds (1997). The most intense reflections for mineralscommonly found in powder samples of soils are listedin Table 3.7. Basal spacings for different clay mineralsassociated with different pretreatments are listed in Ta-ble 3.8 and shown pictorially in Fig. 3.38.

Criteria for Clay Minerals

The different clay minerals are characterized by first-order basal reflections at 7, 10, or 14 A. Positive iden-tification of specific mineral groups ordinarily requiresspecific pretreatments. Separation of size fractions re-quires thorough dispersion of the sample. As cement-ing compounds may both inhibit dispersion andadversely affect the quality of the diffraction patterns,their removal may be necessary. To ensure uniform ex-pansion due to hydration for all crystals of a particularmineral, the clay should be made homoionic. Magne-sium and potassium are most frequently used for sat-uration of the exchange sites. Detailed procedures forpretreatments useful in X-ray diffraction analysis ofclay soils are given by Whittig and Allardice (1986)and Moore and Reynolds (1997).

Kaolinite Minerals The kaolinite basal spacing ofabout 7.2 A is insensitive to drying or moderate heat-ing. Heating to 500�C destroys kaolinite minerals, butnot the other clay minerals. Hydrated halloysite has abasal spacing of 10 A, which collapses irreversibly to7 A on drying at 110�C. Organic chemical treatmentsare sometimes used to distinguish dehydrated halloy-site from kaolinite (MacEwan and Wilson, 1980). Theelectron microscope can also be used to distinguishdehydrated halloysite with its tubular morphology fromkaolinite.

Hydrous Mica (Illite) Minerals Illite is character-ized by d(001) of about 10 A, which remains fixed bothin the presence of polar liquids and after drying.

Smectite (Montmorillonite) Minerals The expan-sive character of this group of minerals provides thebasis for their positive identification. When air dried,these minerals may have basal spacings of 12 to 15 A.After treatment with ethylene glycol or glycerol, thesmectites expand to a d(001) value of 17 to 18 A. Whenoven dried, d(001) drops to about 10 A as a result of theremoval of interlayer water.

Vermiculite Although an expansive mineral, thegreater interlayer ordering in vermiculite results in lessvariability in basal spacing than occurs in the smectiteminerals. When Mg saturated, the hydration states ofvermiculite yield a discrete set of basal spacings, re-sulting from a changing but ordered arrangement ofMg cations and water in the interlayer complex. Whenfully saturated, the d spacing is 14.8 A, which reduces

to 11.6 A when heated at 70�C. All interlayer watercan be expelled at 500�C, but rehydration is rapid oncooling. Permanent dehydration and collapse to 9.02A can be achieved by heating to 700�C.

Chlorite Minerals The basal spacing of chloriteminerals is fixed at 14 A because of the strong orderingof the interlayer complex. Chlorites often have a clearsequence of four or five basal reflections. The third-order reflection at 4.7 A is often strong. Iron-rich chlo-rites have a weak first-order reflection but strongsecond-order reflections and, thus, may be confusedwith kaolinite. The facts that chlorite is destroyedwhen treated with 1 N HCl at 60�C while kaolinite isunaffected, and that kaolinite is destroyed but chloritemay not be affected on heating to 600�C, are usefulfor distinguishing the two clay mineral types.

Criteria for Nonclay Minerals

Strong X-ray diffraction reflections for some of thenonclay minerals are listed in Table 3.7. These includefeldspar, quartz, and carbonates. More detailed listingsof X-ray powder data for specific iron oxide minerals,silica minerals, feldspars, carbonates, and calcium sul-fate minerals are given in Brindley and Brown (1980)as well as in standard reference files.

Quantitative Analysis by X-ray Diffraction

Quantitative determination of the amounts of differentminerals in a soil on the basis of simple comparisonof diffraction peak heights or areas are uncertainbecause of differences in mass absorption coefficientsof different minerals, particle orientations, sampleweights, surface texture of the sample, mineral crys-tallinity, hydration, and other factors. Estimates basedon X-ray data alone are usually at best semiquantita-tive; however, in some cases techniques that accountfor differences in mass absorption characteristics andutilize comparisons with known mixtures or internalstandards may give good results. Soils containing onlytwo or three well-crystallized mineral components aremore easily analyzed than those with multimineralcompositions and mixed layering. For more detailedtreatment of X-ray diffraction theory, identification cri-teria, and techniques, particularly as related to thestudy of clays, see Klug and Alexander (1974), Carroll(1970), Brindley and Brown (1980), Whittig and Al-lardice (1986), and especially Moore and Reynolds(1997).

3.23 OTHER METHODS FOR COMPOSITIONALANALYSIS

Thermal Analysis

Principle Differential thermal analysis (DTA) con-sists of simultaneously heating a test sample and athermally inert substance at constant rate (usually

Copy

right

ed M

ater

ial


OTHER METHODS FOR COMPOSITIONAL ANALYSIS 75

Table 3.7 X-ray Diffraction Data for Clay Minerals and Common Nonclay Minerals

d (A) Minerala d (A) Minerala

14 Mont. (VS) Chl. Verm. (VS)b 2.93–3.00 Felds.12 Sepiolite, heated corrensite 2.89–2.90 Carb.10 Illite, Mica (S), Halloysite 2.86 Felds.

9.23 Heated Verm. 2.84 Carb. Chl.7 Kaol. (S). Chl. 2.84–2.87 Chl.

6.90 Chl. 2.73 Carb.6.44 Attapulgite 2.61 Attapulgite6.39 Felds. 2.60 Verm., Sepiol.

4.90–5.00 Illite, Mica, Halloysite 2.56 Illite (VS), Kaol.4.70–4.79 Chlor. (S) 2.53–2.56 Chlor., Felds., Mont.

4.60 Verm. (S) 2.49 Kaol. (VS)4.45–4.50 Illite (VS), Sepiolite 2.46 Quartz, heated Verm.

4.46 Kaol. 2.43–2.46 Chlorite4.36 Kaol. 2.39 Verm., Illite4.26 Quartz (S) 2.38 Kaol.4.18 Kaol. 2.34 Kaol. (VS)

4.02–4.04 Felds. (S) 2.29 Kaol. (VS)3.85–3.90 Felds. 2.28 Quartz, Sepiol.

3.82 Sepiol. 2.23 Illite, Chl.3.78 Felds. 2.13 Quartz, Mica3.67 Felds. 2.05–2.06 Kaol. (WK)3.58 Carbonate, Chl. 1.99–2.00 Mica, Illite (S), Kaol. Chl.3.57 Kaol. (VS), Chl. 1.90 Kaol.

3.54–3.56 Verm. 1.83 Carb.3.50 Felds., Chlor. 1.82 Quartz3.40 Carb. 1.79 Kaol.3.34 Quartz (VS) 1.68 Quartz

3.32–3.35 Illite (VS) 1.66 Kaolin3.30 Carb. 1.62 Kaolin3.23 Attapulgite 1.54B Verm. (S), Quartz3.21 Felds. 1.55 Quartz3.20 Mica 1.58 Chl.3.19 Felds. (VS) 1.53 Verm., Illite3.05 Mont. 1.50 Ill. (S), Kaol.3.04 Carb. (VS) 1.48–1.50 Kaol. (VS), Mont.3.02 Felds. 1.45B Kaol.3.00 Heated Verm. 1.38 Quartz, Chl.2.98 Mica (S) 1.31, 1.34, 1.36 Kaol. (B)

a(B) � broad; (S) � strong; (VS) � very strong; (WK) � weak; Mont. � montmorillonite; Ch1. � chlorite; Verm. �vermiculite; Kaol. � kaolinite; Carb. � carbonate; Felds. � feldspar; Sepiol. � sepiolite.

b Italics indicates (001) spacing.

about 10�C/min) to over 1000�C and continuouslymeasuring differences in temperature between the sam-ple and the inert material. Differences in temperaturebetween the sample and the inert substance reflect re-actions in the sample brought about by the heating.Thermogravimetric analyses, based on changes inweight caused by loss of water or CO2 or gain in ox-

ygen, are also used to some extent. Thermal analysistechniques are described in detail by Tan et al. (1986).

The results of differential thermal analysis are pre-sented as a plot of the difference in temperaturebetween sample and inert material ( T) versus tem-perature (T) as indicated in Fig. 3.39. Endothermic re-actions are those wherein the sample takes up heat,

Copy

right

ed M

ater

ial



Table 3.8 X-ray Identification of the Principal Clay Minerals (� 2 �m) in an Oriented Mount of a ClayFraction Separated from Sedimentary Material

Mineral Basal d Spacings (001)Glycolation Effect

(1 h, 60�C) Heating Effect (1 h)

Kaolinite 7.15 A (001); 3.75 A (002) No change Becomes amorphous 550–600�C

Kaolinite, disordered 7.15 A (001) broad; 3.75 Abroad

No change Becomes amorphous at lowertemperatures than kaolinite

Halloysite, 4H2O(hydrated)

10 A (001) broad No change Dehydrates to 2H2O at 110�C

Halloysite, 2H2O(dehydrated)

7.2 A (001) broad No change Dehydrates at 125–150�C;becomes amorphous 560–590�C

Mica 10 A (002); 5 A (004)generally referred to as(001) and (002)

No change (001) becomes more intense onheating but structure ismaintained to 700�C

Illite 10 A (002), broad, otherbasal spacings presentbut small

No change (001) noticeably more intenseon heating as water layersare removed; at highertemperatures like mica

Montmorillonite group 15 A (001) and integralseries of basal spacings

(001) expands to 17A with rationalsequence ofhigher orders

At 300�C (001) becomes 9 A

Vermiculite 14 A (001) and integralseries of basal spacings

No change Dehydrates in steps

Chlorite, Mg-form 14 A (001) and integralseries of basal spacings

No change (001) increases in intensity;�800�C shows weight lossbut no structural change

Chlorite, Fe-form 14 A (001) less intensethan in Mg-form;integral series of basalspacings

No change (001) scarcely increases;structure collapses below800�C

Mixed-layer minerals Regular, one (001) andintegral series of basalspacings

No change unlessan expandablecomponent ispresent

Various, see descriptions ofindividual minerals

Random, (001) is additionof individual mineralsand depends on amountof those present

Expands ifmontmorillonite isa constituent

Depends on minerals present ininterlayered mineral

Attapulgite(palygorskite)

High intensity d reflectionsat 10.5, 4.5, 3.23, and2.62 A

No change Dehydrates stepwise (seedescription)

Sepiolite High intensity reflections at12.6, 4.31, and 2.61 A

No change Dehydrates stepwise (seedescription)

Amorphous clay,allophane

No d reflections No change Dehydrates and loses weight

Compiled by Carroll (1970).

Copy

right

ed M

ater

ial


OTHER METHODS FOR COMPOSITIONAL ANALYSIS 77

Figure 3.38 Pictorial representation of response of phyllosilicates to differentiating treat-ments. Approximate spacings in nm (1 nm � 10 A) (from Whittig and Allardice, 1986).Reproduced with permission from The American Society of Agronomy, Inc., Madison, WI.

and in exothermic reactions, heat is liberated. Analysisof test results consists of comparing the sample curvewith those for known materials so that each deflectioncan be accounted for.

Apparatus Apparatus for DTA consists of a sampleholder, usually ceramic, nickel, or platinum; a furnace;a temperature controller to provide a constant rate ofheating; thermocouples for measurement of tempera-ture and the difference in temperature between thesample and inert reference material; and a recorder forthe thermocouple output. The amount of sample re-

quired is about 1 g. Although the temperatures atwhich thermal reactions take place are a function onlyof the sample, the size and shape of the reaction peaksdepend also on the thermal characteristics of the ap-paratus and the heating rate.

Reactions Producing Thermal Peaks The impor-tant thermal reactions that generate peaks on the ther-mogram are:

1. Dehydration Water in a soil may be present inthree forms in addition to free pore water: (1)

Copy

right

ed M

ater

ial



Figure 3.39 Thermogram of a sandy clay soil.

adsorbed water or water of hydration, which isdriven off at 100 to 300�C, (2) interlayer watersuch as in halloysite and expanded smectite, and(3) crystal lattice water in the form of (OH) ions,the removal of which is termed dehydroxylation.Dehydroxylation destroys mineral structures. Thetemperature at which the major amount of crystallattice water is lost is the most indicative propertyfor identification of minerals. Dehydration reac-tions are endothermic and occur in the range of500 to 1000�C.

2. Crystallization New crystals form from amor-phous materials or from old crystals destroyed ata lower temperature. Crystallization reactionsusually are accompanied by an energy loss and,thus, are exothermic, occurring between 800 and1000�C.

3. Phase Changes Some crystal structures changefrom one form to another at a specific tempera-ture, and the energy of transformation shows upas a peak on the thermogram. For example,quartz changes from the � to � form reversiblyat 573�C. The peak for the quartz phase changeis sharp, and its amplitude is nearly in direct pro-portion to the amount of quartz present. Thequartz peak is frequently masked within the peakfor some other reacting material, but may bereadily identified by determining the thermogramduring cooling of the sample or by letting it coolfirst and then rerunning it. The other minerals aredestroyed during the initial run while the quartzreaction is reversible.

4. Oxidation Exothermic oxidation reactions in-clude the combustion of organic matter and theoxidation of Fe2� to Fe3�. Organic matter oxi-dizes in the 250 to 450�C temperature range.

Beside quartz, the only common nonclay mineralsin soils that give thermal reactions with large peaks arecarbonates and free oxides such as gibbsite, brucite,and goethite. The carbonates give very large endother-mic peaks between about 800 and 1000�C, and the ox-ides have an endothermic peak between about 250 and450�C. Thermograms for many clay and nonclay min-erals are presented by Lambe (1952).

Quantitative Analysis Theoretically, the area of thereaction peak is a measure of the amount of mineralpresent in the sample. For sharp, large amplitude peakssuch as the quartz inversion at 573�C and the kaoliniteendotherm at 650�C, the amplitude can be used forquantitative analysis. In either case, calibration of theapparatus is necessary, and the overall accuracy is ofthe order of plus or minus 5 percent.

Optical Microscope

Both binocular and petrographic microscopes can beused to study the identity, size, shape, texture, and con-dition of single grains and aggregates in the silt andsand size range; for study in the thin section of thefabric, that is, the spatial distribution and interrelation-ships of the constituents; and for study of the orien-tations of groups of clay particles. Because the in-focusdepth of field decreases sharply as magnification in-creases, study of soil thin sections is impractical atmagnifications greater than a few hundred. Thus, in-dividual clay particles cannot usually be distinguishedusing an optical microscope.

Useful information about the shape, texture, size,and size distribution of silt and sand grains may beobtained directly without formal previous training inpetrographic techniques. Some background is neededto identify the various minerals; however, relativelysimple diagnostic criteria that can be used for identi-fication of over 80 percent of the coarse grains in mostsoils are given by Cady et al. (1986). These criteriaare based on such factors as color, refractive index,birefringence, cleavage, and particle morphology. Thenature of surface textures, the presence of coatings,layers of decomposition, and so on are useful both forinterpretation of the history of a soil and as a guide tothe soundness and durability of the particles.

Electron Microscope

With modern electron microscopes it is possible to re-solve distances to less than 100 A, thus making studyof small clay particles feasible. Electron diffractionstudy of single particles may also be useful. Electrondiffraction is similar to X-ray diffraction except anelectron beam instead of an X-ray beam is used.

Copy

right

ed M

ater

ial


QUANTITATIVE ESTIMATION OF SOIL COMPONENTS 79

Magnetic lenses that refract an electron beam formthe basis of the transmission electron microscope(TEM) optical system. An electron beam is focused onthe specimen, which is usually a replica of the surfacestructure of the material under study. Some of the elec-trons are scattered from the specimen, and differentparts of the specimen appear light or dark in proportionto the amount of scattering. After passing through aseries of lenses, the image is displayed on a fluorescentscreen for viewing. Probably the most critical aspectof successful transmission electron microscopy is spec-imen preparation.

In the scanning electron microscope (SEM), second-ary electrons emitted from a sample surface form whatappear to be three-dimensional images. The SEM hasa �20 to �150,000 magnification range and a depthof field some 300 times greater than that of the lightmicroscope. These characteristics, coupled with thefact that clay particles themselves and fracture surfacesthrough soil masses may be viewed directly, have ledto extensive use of the SEM for study of clays. Ex-amples of electron photomicrographs of clays and soilsare given earlier in this chapter and in Chapter 5. Prin-ciples of electron microscopy techniques and addi-tional examples are presented in McCrone and Delly(1973) and Sudo et al. (1981).

3.24 QUANTITATIVE ESTIMATION OF SOILCOMPONENTS

Qualitative X-ray diffraction and a few simple testswill generally indicate the minerals present in a soil.More data are needed, however, for more precise quan-titative estimates. As a rule, the number of differentanalyses needed is equal to the number of mineral spe-cies present. The results of glycol adsorption, cationexchange capacity, X-ray diffraction, differential ther-mal analysis, and chemical tests all give data that maybe used for quantitative estimations. Some pertinentidentification criteria and reference values for the clayminerals are given in Table 3.9.

After the quantities of organic matter, carbonates,free oxides, and nonclay minerals have been deter-mined, the percentages of clay minerals are estimatedusing the appropriate glycol adsorption, cation ex-change capacity, K2O, and DTA data. The nonclayscan be identified, and their abundance determined, us-ing the microscope, grain size distribution analysis, X-ray diffraction, and DTA. The amount of illite isestimated from the K2O content since this is the onlyclay mineral containing potassium. The amount of ka-olinite is most reliably determined from the 600�C

DTA endotherm amplitude. If X-ray has indicatedmontmorillonite, chlorite, and/or vermiculite, thenquantitative estimates are made based on the glycoladsorption and exchange capacity data. The totalexchange capacity and glycol retention are ascribed tothe clay minerals, and the measured values must beaccounted for in terms of proportionate contributionsby the different clay minerals present.

As a simple example, assume that quartz, illite, andsmectite are identified in the �2 �m fraction of a soil.Additional data indicate 4.0 percent K2O, ethylene gly-col retention of 100 mg/g, and a cation exchange ca-pacity of 35 meq/100 g. Then, assuming 9 percent asan average value of for pure illite (Table 3.9), the con-tent of illite is estimated at 4.0/9.0, or 44 percent. Be-cause only the illite and smectite will contribute to theglycol adsorption, the amount of smectite may be es-timated:

0.44 � 60 � S � 300 � 100

100 � 26.4� S � � 25%

300

The remaining 31 percent can be ascribed to quartzand other nonclay components. For this clay mineralcomposition, the theoretical cation exchange capacityshould be, based on the reference values in Table 3.9:

0.44 � 25 � 0.25 � 85 � 11 � 21 � 33 meq/100 g

This compares favorably with the measured quantityof 35 meq/100 g. Thus, the composition of the claysize fraction is

Illite 44%Smectite 25%Quartz and other nonclays 31%

The main difficulty in this method for quantitative min-eralogical analysis is the uncertainty in the referencevalues for the different clay minerals.

A semiquantitative analysis is sufficient for most ap-plications. This may be done as follows. The silt andsand fraction can be examined by microscope and theapproximate proportion of nonclay minerals deter-mined. The amount of clay size material �2 �m canbe estimated by grain size distribution analysis. As afirst approximation, it may be assumed that the amountof clay mineral equals at least the amount of clay size.This assumption is justified for the following reasons.Nonclay minerals, principally quartz, are found in theclay size fraction. On the other hand, for most soils,

Copy

right

ed M

ater

ial



Table 3.9 Summary of Clay Mineral Identification Criteria—Reference Data for Clay Mineral Identification(�2-�m fraction)

ClayX-rayd(001)

Glycol(mg/g)

CEC(meq/100 g)

K2O(%) DTAa

Kaolinite 7 16 3 0 End. 500–660� � Sharpb

Exo. 900–975� SharpDehydrated halloysite 7 35 12 0 Same as kaolinite but 600 peak

slope ratio � 2.5Hydrated halloysite 10 60 12 0 Same as kaolinite but 600� peak

slope ratio � 2.5Illite 10 60 25 8–10 End. 500–650� Broad

End. 800–900� BroadExo. 950�

Vermiculite 10–14 200 150 0Smectite 10–18 300 85 End. 600–750�

End. 900�

Exo. 950�

Chlorite 14c 30 40 0 End. 610 � 10� or 720 � 20�

aFor clays prepared at same relative humidity the size of the 100–300�C endotherm (adsorbed water removal) increasesin the order kaolinite–illite–smectite.

bFor samples started at 50% RH the amplitude of 600� peak/amplitude of adsorbed water peak ��1.cHeat treatment will accentuate 14 A line and weaken 7 A line.

the amount of clay mineral exceeds the amount of claysize. This most probably results from cementation ofsmall clay particles into aggregates larger than 2 �min diameter. Approximate proportions of the differentclay minerals in the clay fraction can be estimatedfrom the relative intensities of the X-ray diffractionreflections for each mineral. The presence of organicmatter and carbonates can be easily detected using thetests listed in Section 3.21.

3.25 CONCLUDING COMMENTS

The sizes, shapes, and surface characteristics of theparticles in a soil are determined in large measure bytheir mineralogy. Mineralogy also determines interac-tions with fluid phases. Together, these factors deter-mine plasticity, swelling, compression, strength, andfluid conductivity behavior. Thus, mineralogy is fun-damental to the understanding of geotechnical prop-erties, even though mineralogical determinations arenot made for many geotechnical investigations. In-stead, other characteristics that reflect both composi-tion and engineering properties, such as Atterberglimits and grain size distribution, are determined.

Interatomic bonding, crystal structure, and surfacecharacteristics determine the size, shape, and stabilityof soil particles and the interactions of soil particleswith liquids and gases. The structural stability of thedifferent minerals controls their resistance to weath-ering and hence accounts in part for the relative abun-dance of different minerals in different soils.

Because interatomic bonds in soil particles arestrong, primary valence bonds, whereas usual interpar-ticle bonds are of the secondary valence or hydrogenbond type, individual particles are strong compared togroups of particles. Thus, most soil masses behave asassemblages of particles in which deformation proc-esses are dominated by displacements between parti-cles and not by deformations of particles themselves,although grain crushing becomes important in coarse-grained soils such as sands and gravels when they areunder very high stresses.

The type of bonding between the unit layers of theclay minerals, coupled with the adsorption propertiesof the particle surfaces, controls soil swelling. Ad-sorption and desorption processes are important ininteractions between chemicals and soils. These inter-actions in turn determine the flow and attenuation ofvarious substances through soil. Changes in surface

Copy

right

ed M

ater

ial


QUESTIONS AND PROBLEMS 81

forces owing to changes in chemical environment mayalter the structural state of a soil.

Mineralogy is related to soil properties in much thesame way as the composition and structure of cementand aggregates are to concrete, or as the compositionand crystal structure of steel relate to its strength anddeformability. With these engineering materials—soil,concrete, and steel—mechanical properties can bemeasured directly; however, they cannot be explainedwithout consideration of the composition and structureof their components.

Since about 1980, environmental problems, espe-cially those related to the safe disposal and contain-ment of municipal, hazardous, and nuclear waste andto the clean up of contaminated sites and the protectionof groundwater, have assumed a major role in geo-technical engineering practice. This has required agreatly increased focus on the compositional charac-teristics of soils and their relation to the long-termphysical and chemical properties that control soil be-havior under changed and extreme environmental con-ditions.

QUESTIONS AND PROBLEMS

1. A montmorillonite has a cation exchange capacityof 130 meq/100 g and a total external and internalsurface area of 800 m2/g.a. How many calcium ions will there be on a par-

ticle that is 0.4 �m � 0.2 �m � one unit cellin thickness?

b. What percentage of the dry weight of the clayis composed of calcium?

2. An orthorhombic crystal has axial ratios of 0.6,0.3, and 1.0. The (500) plane is 2.0 A horizontallyfrom the origin. This crystal is irradiated withCuK� X-rays (wave length of 1.54 A). At whatvalue of � does the second-order (010) reflectionoccur?

3. Sketch the following planes relative to crystallo-graphic axes: (001), (243), (hk0), (hkl), (111),(060), (010).

4. Consider an orthorhombic crystal of dimensionsa � 6A, b � 12 A, c � 8 A. With the aid ofsketches determine the angle of intersection be-tween the planes of each pair indicated below. Ifthe planes do not intersect, then so indicate.a. (002) and (020)b. (001) and (002)c. (111) and (222)

d. (111) and ( )111e. (112) and (001)

5. A clay has a surface density of charge of onecharge per 150 A2. Its cation exchange capacity is10 meq/100 g. Determine the specific surfacearea.

6. Why are soils containing smectite often expansive,whereas soils containing illite and/or kaolinite arenot?

7. As the geotechnical engineer on a project, you findan inorganic soil containing 15 percent by weightof particles finer than 100 �m, as measured byhydrometer analysis. What soil components doyou expect? Why?How could you confirm this expectation? Be spe-cific in terms of tests and diagnostic criteria.

8. What is the smallest interplanar spacing that canbe measured by X-ray diffraction using copper K�

radiation?

9. You suspect that a fine-grained soil sample con-tains kaolinite, illite, and smectite minerals. De-scribe in logical sequence the tests you would doto verify that these clay minerals are present. In-dicate the reasons why you choose these tests andthe criteria for distinguishing among the minerals.

10. An inorganic clay has a liquid limit of 350 percent.a. What is the most probable predominant clay

mineral in this soil?b. Explain the high liquid limit in terms of the

crystal structure of this mineral.c. Would you recommend founding light struc-

tures on shallow footings above this soil? Why?

11. A soil sample has a cation exchange capacity of30 meq/100 g and a specific surface area of 50m2/g. You wish to determine the type of clay min-eral in this soil. Based on your general knowledgeof the area from which it came, including the ge-ology, you suspect the possibility of hydrated hal-loysite, illite, and smectite. State specifically howyou would determine which mineral is present.

12. An X-ray diffraction pattern for a soil sample froma site where light structures (houses, a shoppingcenter) are to be located shows peaks at 2� � 5�,10�, 12.2�, 20.8�, 24.7�, and 26.7�. Copper K� ra-diation was used.a. What minerals are present in the sample?b. If the measure cation capacity is 40 meq/100

g, what is the approximate minimum amount of

Copy

right

ed M

ater

ial



Figure 3.40 Gradation curve for a sandy clay soil.

clay mineral in the sample by weight percent-age?

c. What concerns would you have about this soilas a foundation material?

d. How could you minimize any problems iden-tified in part (c)?

13. In general the average clay particle size as repre-sented by some effective diameter D, for smectiteparticles (S) is less than that of hydrous mica (il-lite) (HM) particles, which, in turn, is less thanthat of kaolinite (K) particles. In addition, the av-erage particle thicknesses are in the order

tS � tHM � tK

and values of the thickness-to-diameter ratio (t /D)are in the order

(t /D)S � (t /D)HM � (t /D)K

What are some implications of these relationshipswith respect to the relative values of plasticity, hy-draulic conductivity, compression–swell behavior,and strength characteristics of three soils: one con-taining a large amount of smectite, one containinga large amount of hydrous mica (illite), and onecontaining a large amount of kaolinite?

14. The gradation curve for a sandy clay soil is shownin Fig. 3.40.a. What are the percentages by weight of sand,

silt, and clay size material?b. Consider a 100-g sample of the soil and assume

that all sand particles are of a size equal to theaverage particle size in the sand size range, thesilt particles are of a size equal to the averageparticle size in the silt size range, and all clayparticles are of a size equal to the average par-ticle size in the clay size range. Base your de-termination of average particle size in eachrange on equal weights of particles coarser andfiner than the average for each size range.

Estimate the number of sand, silt, and clayparticle in the sample. For purposes of thisestimate, the sand and silt particles can beassumed to be spherical. Assume the clayparticles to be flat disks having a diameter-to-thickness ratio of 10. Assume the average sizeof clay particles on the gradation curve to rep-resent the disk diameter.

c. Estimate the specific surface area of this soil insquare meters per gram. Determine the per-centages of this total that are contributed by thesand, silt, and clay fractions.

d. Are the estimates of the numbers of particlesand specific surface area made in this way toohigh, too low, or correct? Why?

Copy

right

ed M

ater

ial


63027 03

Documents

Transcript of 63027 03