SURFACE SCIENCE 25 (1971) 379-384 o North-Holland Publishing Co.

CALCULATED LAMELLAR BINDING

I. VAN DER WAALS BONDING IN TALC AND PYROPHYLLITE*

WILLIAM WARD and JAMES M. PHILLIPS

University of Missouri-Kansas City, Kansas City, Missouri 64110, U.S.A.

Received 4 March 1970

The preferred inter-layer binding energy (twice the specific surface energy) of talc and
pyrophyllite has been calculated to be 980 erg/cmx. Van der Waals forces are summed over
the surface atoms of two adjacent oxygen sheets for four possible equilibrium stacking
arrangements. The force constants are determined from bulk modulus measurements
made by Bridgman.

1. Introduction

The purpose of these calculations is to provide a theoretical model for the

binding energy of Van der Waals bound lamellar solids. The complex struc-
ture of many engineering materials has prevented rigorous calculations from
being performed. This work is an attempt to determine the relevant features
of bonding computations. A computational model for calculating binding
energies would hopefully be available for use in a variety of applied problems
for which experimental measurements are impractical. To illustrate, such in-
formation regarding the binding energy is valuable in adhesion theory,
friction, and cold welding problems.
This study is the first of a series of calculational approaches to the problem
of inter-layer binding energy in lamellar solids. The key to the series is the
selection of a set of lamellar materials with large sheet size and mechanical
properties suitable for experimental measurements; however, the set of
materials must allow for a clear separation of the types of bonding, i.e.
Van der Waals, ionic, etc. For a clear distinction of the interactions, the
computations require that one material in a family of minerals be purely
van der Waals bound between lamella.
The set of materials chosen for the first part of the study phyllosilicate
subclass of the silicate class of minerals. The phyllosilicate subclassl) is
divided into two sets of materials, dioctahedral and trioctahedral. The

* This research received financial support from the National Aeronautics and Space
Administration, Grant No. NGR 26-001-006.

379
380 WILLIAMWARDANDJAMESM.PHILLIPS

TABLE 1

Alisting of the phyllosilicate minerals of interest to this program - each has the tetrahedral-
octahedral-tetrahedral lamellar structure ideal for cleavage experiments

Dioctahedral phyllosilicate Trioctahedral phyllosilicate

T-O-T Pyrophyllite T-O-T Talc.

AlsSi40ro(OH)a MgsSi401o(OH)a
(T-0-T)-K+ Muscovite (T-0-T)-K+ Phlogopite
KAla [AlSiaOro (OH)21 KMga(AlSis010) (OH)2
(T-O-T)2-Ca2+ Margarite (T-O-T)2-Ca2+ Xanthophyllite
CaAla(Al&isOlo) (OH)2 CaMgs (AlzSiaOlo) (OH)2

members of interest to this program, along with their chemical composition

and structure type are given in table 1. All of the minerals listed in table 1
have sheet-like structure for easy cleavage. All have a basic structure within
a given sheet of a central section of an octahedral [pyrophyllite octahedral,
A140s (OH),] layer linked to two tetrahedral [pyrophyllite tetrahedral,
Si,O,,] sheets (see fig. 1). Hence, the structure type T-O-T.

c axis

60
4 SI

a, OH

: 4Al
2(Oli)*40

2(OH)+40

4 SI
60

b axis

Fig. 1. Structure of pyrophyllite, Al4(Si4010) (OH)4.

The aluminum set of minerals has a dioctahedral structure for the central
section of a single sheet where the magnesium set has a trioctahedral struc-
ture. The oxygen layer on the top and bottom of a single sheet has a structure
made up of the triangular bases of the SiO, tetrahedra arranged in nearly a
hexagonal form (see fig. 2).
 CALCULATED LAMELLAR BINDING. I 381

- _ - - -
CLOSE PACKED S”lFl ED CLOSE PACKED

Fig. 2. Four equilibrium stacking arrangements.

2. Calculation

The primary interaction between the layers in pyrophyllite and talc should
be the oxygen to oxygen interaction across the cleavage plane.
A Lennard-Jones (6-12) potential is used to describe the induced dipole
(Van der Waals) binding. The energy of binding between two oxygen layers
is written as

(1)

Where Rij is the distance from the ith reference atom to thejth atom in the
adjacent layer. The parameters o and E must be determined experimentally.
Defining rij = Rfj /a’, with a the lattice constant, gives

With
u=&(~~cC[(~~(~~-(~~].
C2)
1 j

es = (llrij) (3)
cc
i j
382 WILLIAM WARD AND JAMES M.PHILLIPS

then

(4)

To eliminate the parameters ts and E, the equilibrium spacing z0 and the bulk
modulus B are used.
The distance between layers is z and the atoms within a layer are described
by xi, Yj coordinates. Defining xij=xi-xi and Yij=Yi-Yj gives

Rij = (~2 + Yz + Z2)3

and
ri) = (Xij/U)'+(yijla)'+ (z/a)"* (5)
Then
86, 2sz
-=- 8s+l* (6)
aZ
( a2 >
At equilibrium, z=zO and
au

This yields

(7)

The bulk modulus B= - V(dP/dV). At O”K,

d2U,
dU,=-PdV, dp
( dV > r=-WY
and
d2U,
B=V--- (8)
dV2 *
It can be shown that V=z’Na’ (2/J3), wh ere z’ = 9.3 A is the thickness of one
layer unit (see fig. 1). Here it is assumed that the internal structure of each
layer contributes little to the bulk modulus compared to the layer separation,
since the binding of the former is principally covalent. N is the number of
oxygen atoms in the layer. Now,

and U, = NU. Differentiating

 CALCULATED LAMELLAR BINDING. I 383

The bulk modulus is found to be

Solving for E gives

In terms of z,, and B, using (4), (7), and (1 l),

1*
648.5 - 20387
(12)
7e,es - 4e5e, _
Let
e4e6 - 2e3e7
u(4 = (13)
7e4e8 - 4e,e, z=zo *

Here u(zO) depends on the stacking structure of one layer upon the other.
Early work in crystallography’) suggests four different stacking arrange-
ments from one sheet to another. Because of the different symmetries, the
binding energy is expected to differ from one stacking arrangement to an-
other. Using drawings and models, the four types of stacking shown in fig. 2
appear to be the four equilibrium configurations. It is necessary to develop
sums for all these types of stacking. Natural samples would be expected to
have some of each type of stacking with the lowest energy configuration being
dominant. In the following sums for each of these stacking arrangements are
derived and evaluated. The summing methods are far to lengthy to present
here. Detailed methods of computing these sums are available from the
authors (J.M.P.).
3. Results

For purposes of identification the four stacking arrangements are named:

column, shifted, close packed, and shifted close packed. The value 2.70&
reported by Gruner3), for talc was used as the layer separation. The lattice
constant a was 2.87 A. These values were substituted in the derived series and
a direct summation was performed on an IBM 360 computer. In all cases,
convergence to fourteen decimal places was achieved within eighteen rings of
neighbors.
The value of the sums for each stacking arrangement are
pu (z,,) = - 0.849 (close packed),
,pu (zo) = - 0.836 (shifted close packed),
scu(4 = - 0.579 (shifted column),
cu (zo) = - 0.528 (column).
384 WILLIAM WARD AND JAMES M. PHILLIPS

With

it is found that
,U=-0.331BA3,
,$J = - 0.332 B ti3,
s,U = - 0.230 B A3,
,U=-0.209BA3,
where B is the bulk modulus.
The bulk modulus for pyrophyllite was taken to be 1.722 x lOlo dyne/cm2
(ref. 4). U, the interlayer binding energy per atom, is then multiplied by the
number of atoms per unit area (1.69 x 101’ atoms/cm2) to give the total
interlayer binding energy E. The values for the four different stacking struc-
tures are:
pE = - 980 erg/cm’,
s,E = - 965 erg/cm’,
s,E = - 668 erg/cm’,
J = - 609 erg/cm’.

The favored stacking (980 erg/cm2) is the one named close packed in this
report. The more negative interlayer binding energy would be preferred in an
equilibrium structure. It is possible that a Boltzmann weighting of the areas
having all four such structures would be found in a real sample. The result
found in this work is quite similar to a Van der Waals calculation on graph-
ite5) (660 erg/cm2).
References

1) B. Mason and L. G. Berry, Elements of Minerals (Freeman, San Francisco, California,

1969), p. 436.
2) S. B. Hendricks, Z. Krist. 99A (1938) 264;
B. E. Brown and S. W. Bailey, Am. Mineralogist 47 (1962) 819.
3) J. W. Gruner, Z. Krist. fBA (1934) 412.
4) P. W. Bridgman, Am. J. Sci. 7 (1924) 81.
5) L. A. Girifalco and R. A. Lad, J. Chem. Phys. 25 (1956) 693.