Class 5 Class 5 Secondary Bonds Secondary Bonds Class 5 Class 5 Secondary Bonds Secondary Bonds Class 5 Class 5 Secondary

dary Bonds Secondary Bonds Class 5 Class 5 Secondary Bonds Secondary Bonds
READING
Ibid.
Prof. M.L. Weaver
SECONDARY BONDING SECONDARY BONDING SECONDARY BONDING SECONDARY BONDING
Arisesfrominteractionbetweendipoles
Fluctuatingdipoles
asymmetric electron
ex: liquid H
2
asymmetricelectron
clouds
+ + H H H H
H
2
H
2
ex:liquidH
2
Permanent dipolesmolecule induced
AdaptedfromFig.2.13,
Callister&Rethwisch8e.
secondary
bonding
H H H H
secondary
bonding
Permanentdipolesmoleculeinduced
generalcase:
AdaptedfromFig.2.15,
Callister & Rethwisch 8e
secondary
bonding
+ +
ex:liquidHCl
Callister&Rethwisch8e.
H Cl H Cl
secondary
bonding
Prof. M.L. Weaver
102
ex:polymer
secondarybonding
DipoleForces DipoleForces
Electrostaticinteractions
betweenadjacentmolecules
arecalledDipoleForces p
Themostextremeformof
di l f ltf th dipoleforcesresultfromthe
interactionofahydrogen
atomwithahighly
l i electronegativeatom
resultinginhydrogen
bonding.
DispersionForces(orVanderWaalsforces) DispersionForces(orVanderWaalsforces)
Interactionsresultingfrom
momentaryconcentration
variationsintheelectron
cloudsofadjacentatoms
Th i t ti b i f d Theinteractionsarebriefand
weak,butlargemolecules
(suchaspolymers)have
i i f opportunitiesformany
simultaneousinteractions
LL--J Potential Energy Model (continued) J Potential Energy Model (continued)
To calculate the equilibrium separation (r
o
) between two atoms, or lattice constant, we
differentiate equation (1) to find the minimum of V
o
(r):
(2)
7
6
13
12
1
6 4
1
12 4 0 |
.
|

\
|
- |
.
|

\
|
- = = =
r r dr
dV
F
o
co co
Solving for r allows r
o
to be determined: thus its the size parameter, o,
that determines r
o
.
E i l t t i th f f ti F( ) i th ti f th d i ti t f E (1)
o o 12 . 1 2
6 / 1
= =
o
r
( )
(3)
. \ . \
r r dr
Equivalent to saying the force function, F(r), is the negative of the derivative w.r.t. r of Eq. (1).
Figure on left for inert gases shows this relationship.
The first derivative of the interatomic potential energy, with
di i h f di l Th respect to distance, gives the force-displacement curve. The
positive forces are attractive and the negative forces are
repulsive, and the zero force point is the equilibrium
separation (r
o
). p (
o
)
Good agreement between calculated and measured r
o
s.
The L-J potential is a relatively good approximation and due
to its simplicity often used to describe the properties of gases,
andtomodel dispersionandoverlapinteractionsinmolecular
Prof. M.L. Weaver
and to model dispersion and overlap interactions in molecular
models. It is particularly accurate for noble gas atoms and is a
good approximation at long and short distances for neutral
atoms and molecules, layered solids (graphite) and polymers.
LL--J Potential Energy Model (continued) J Potential Energy Model (continued)
To calculate the cohesive energy per atom, U, in eV/atom, we sum the potential energy (V)
between each atom and all the other atoms in a crystal. The interaction energy, U
j
, of the j-th
atom with all the other N-1 atoms in the crystal is: where r
i
is distance to the i-th
N
atom with all the other N 1 atoms in the crystal is: where r
i
is distance to the i th
atom.
There is a similar term for every atom in the crystal, so the complete sum, U, is:
F t f i i l d dt id ti hi t ti t i
) (
, 1

= =
=
N
j i i
i j
r V U

=
N
j
U U
2
1
(4)
(5)
Factor of is included to avoid counting each interaction twice.
Assuming every atom is indistinguishable, we can rewrite the complete sum:
Then cohesive energy per atom, U is total energy in eq. (6), divided by N:

= j
j
1
2
j
NU U
2
1
=
( )
(6)
1
gy p , gy q ( ) y
How do we actually compute the sum, U
j
, in eq. (4) or eq. (7)?
We use simplest approximation first, i.e., since VDW forces are very weak, the nearest neighbor
shellswill significantlycontributetocohesiveenergy Thus weonlyneedtocomputethesum
j
U U
2
1
'=
(7)
shells will significantly contribute to cohesive energy. Thus, we only need to compute the sum
over the nearest neighbors. For our previous example, all inert gases have face-centered cubic
(FCC) structure, so cohesive energy per atom is:
) (
2
1
'
12
1

=
=
i
i
r V U (8) or ) ( . .
2
1
' r V N N U - =
Prof. M.L. Weaver
For all 12 neighbors, V(r
i
)=V(r
o
)=-c. Substitute into eq.(8), cohesive energy per atom:
We can also calculate this U
This approximation slightly underestimates the measured cohesive energy, since we have ignored all
additional interactions (long range) with atoms outside of the nearest neighbor shell.
i
c = 6 ' U
(9)
FCC crystal structure FCC crystal structure
representative of inert gases representative of inert gases ep ese ve o e g ses ep ese ve o e g ses
#nearest neighbors or
coordination number is 12:
FCC structure: FCC structure :
a 2
The extended
cubic F lattice
Close-packed directions:
length =4R =
2 a
a
1 2
g
Pair energy (each atoms contribution),
Each atom is partitioned
1 2
Prof. M.L. Weaver
This is a cubic face (F) center Bravais lattice.
Lattice +basis vectors =crystal structure.
12-fold coordination of each lattice point (same atom) is identical.
LL--J Potential Energy Model (continued) J Potential Energy Model (continued)
To correct errors from our neglect of long range interactions, we start by expressing all of the
interatomic distances, r
i
, as multiples of the shortest one, r
o
.
(
Thus eq. (8) becomes: therefore determining
cohesive energy/atom:
) (
2
1
'
1

=
=
N
i
o i
r V U
(10)
( ) ( )
(
(

=
'

= =
N
i
r
N
i
r
i i
U
1
6
1
12
2

o

o
c
(11)
is a dimensionless
parameter dependent on crystal structure
amounts to evaluating the sums:
Minimizing eq. (11) to find equilibrium separation (r
o
) results:
( )
12
1
12
/ 1

=
=
N
i
i
A ( )
6
1
6
/ 1

=
=
N
i
i
A
(12)
o
6 / 1
12
2
|
|
.
|

\
|
=
A
A
r
o
(13)
g q ( ) q p (
o
)
For FCC structure, A
12
=12.13 and A
6
=14.45, thus r
o
=1.09o, which differs from the nearest
neighbor estimate we calculated in eq. (3), r
o
=1.12o, by only 3%.
6
|
.

\
A
o
When the corrected value for U is computed using new value for r
o
& eq.(11):
c =
'
6 . 8 U
(14)
Prof. M.L. Weaver
LL--J Potential Energy Model (continued) J Potential Energy Model (continued)
The differences between the values computed using the N.N. and L.R. interaction models are
summarized:
parameter nearest
neighbor
long range difference
neighbor
r
o
1.12o 1.09o 3%
U -6c -8.6c 30%
What this means for FCC structures in terms of U is 12 N.N. only contribute 70% of the total
cohesive energy, and final 30% is supplied by the rest of crystal (next N.N. and so on..).
Relatively good agreement between measured and calculated parameters for inert gases.
The L-J pair potential model has been successfully used to describe bonding in other systems.
For example theinteractionbetweendifferent segmentsonpolymers: oneLJ particlemay For example, the interaction between different segments on polymers: one LJ particle may
represent a single atom on the chain (explicit atom model), a CH
2
segment (united atom model),
or segment consisting of several CH
2
units (coarse-grained model). The united atom model has
been shown to successfully reproduce explicit atom results for polymer melts.
Prof. M.L. Weaver
Significance of L Significance of L- -J Potential Model J Potential Model
WhiletheL-J pair potential eq (1) isintendedasaphysical model for VDWbonds it isoften While the L-J pair potential, eq. (1), is intended as a physical model for VDW bonds, it is often
extended empirically to model other types of crystals by adjusting the constants o and c.
Assuming that regardless of physical mechanism governing interactions among atoms, the total
energy of the system can be treated as sum of attractive and repulsive pair-wise contributions.
Constants o and c are chosen such that pair-wise potentials reproduce known properties with
accuracy.
Empirical models constructed in this way can then be used to compute physical properties that
aredifficult, tediousor impossibletomeasurebyexperiment. Whilequantitativeaccuracycan are difficult, tedious or impossible to measure by experiment. While quantitative accuracy can
not be expected, the relative energies computed from such models are often qualitatively
meaningful and instructive.
For example, a L-J model for Cu (FCC metal) was used to compute the surface energy/unit area
( ) f ti f f l Th f
Anisotropic isproportional to#bondsbroken/area:
() as a function of surface normal. The surface energy
can be defined as the work required for the creation of
a unit area of surface. Creating new surfaces requires
bond breaking. For a surface with a unit cell that contains
Anisotropic is proportional to #bonds broken/area:
(111)<(001)<(110)
g
N atoms in an area A, the is:
where U is bulk potential
energy/atom, V
o
is the L-J
potential definedineq (1) andr isdistancebetweeni
th
A
r V U N
i j
ij o
(
(

'
=

>1
) (

(15)
Prof. M.L. Weaver
potential defined in eq. (1) and r
ij
is distance between i
th
and j
th
atom. Creating a surface breaks symmetry of the
lattice so positions do not have same environment, thus need to compute sum over atom pairs.
Return to Ionic Bonding Model Return to Ionic Bonding Model
(general model) (general model)
r
Z Z ke
c
r V
2 1
2
) (

=
The pair potential for an ionic bond is to add the attractive Coulomb potential energy:
to L-J potential, eq.(1) from last class, to get:
(
(

|
.
|

\
|
|
.
|

\
|
+

=
6 12
2 1
2
4 ) (
r r r
Z Z ke
r V
o o
c
(1)
(g ) (g )
The total lattice energy is written by modifying eq.(1) to include the Madelung constant (o), n
1
and n
2
stoichiometry, and the number of nearest neighbor (NN) contributions to L-J portion of the
energy [long range attractive contributions can be added but in most cases are negligible for ionic bonding].
(

. \ . \
r r r
(1)
New expression for lattice energy as a function of separation for binary crystals is:
Crystal structuresensitive parameters are
o and #of NN.
Compound chemistrysensitive parameters
are Z Z c and
(
(

|
.
|

\
|
|
.
|

\
|
+
+
=
6 12
2 1 2 1
2
4
2
) (
) (
r r
NN
r
Z Z n n ke
r V
o o
c
o
(2)
are r, Z
1
, Z
2
, n
1
, n
2
, c and o.
Since o varies over a small range, variations in the chemistry of the compound affect total lattice energy much more
than variations in the structural configuration.
Appropriate pair potentials for ionic materials must obviously contain an electrostatic
component ashort rangerepulsionandanattractiveVDWcomponent

component, a short-range repulsion and an attractive VDW component.
Ex.: Compute the equilibrium spacing (r
o
) of the Na and Cl atoms in NaCl and the lattice energy,
V(r), of the crystal. Compare the computed values with the known values (next page).
Prof. M.L. Weaver
Return to Ionic Bonding Model Return to Ionic Bonding Model
(practical model: Born (practical model: Born--Mayer Mayer- -Huggins form of potential energy) Huggins form of potential energy)
In the physical description of the ionic bond we used a term from the L-J model to account for
short-range repulsions. This was chosen mainly for consistency. However, in practice, it is
common to use a repulsive parameter of the form: similar in form to
( )
ij ij
/ r
ij ij
e A ) r ( V

=
(3)
our previously defined repulsive force:
A
ij
and
ij
are empirically derived parameters for specific atom pairs
(Table 7.17 below).
The parameters are chosen so that a model using these values reproduces known properties, with
th ti th t th d l ill th b bl f ti k ti
ij ij rep
e A ) r ( V
( )

/ r
R
e F

=
CRC Handbook of Chemistry and Physics
the assumption that the same model will then be capable of computing unknown properties.
When combined with an electrostatic attractive
term, this is the Born-Mayer-Huggins
(BMH) form of the potential energy.
=r
o
() =r
o
=V =V
( ) f f p gy
Simple calculations can be carried out by hand,
while calculations involving more complex
structures or defects are usually carried out by
computersusingtheBMHpotential computers using the BMH potential.
[39]=G.V. Lewis, Physica 131B (1985) 114.
[40]=J.R. Walker and C.R.A. Catlow, J.Phys.C, 15 (1982) 6151.
Prof. M.L. Weaver
attractive
VDW
component
Return to Ionic Bonding Model Return to Ionic Bonding Model
(BMH form continued) (BMH form continued) ( ) ( )
If we combine the repulsive parameter, eq.(3), with the Madelung attractive energy (ignoring
VDW attractive part for the moment), we get the
BMH form of the potential energy:
( )
ij
r
ij
e NNA
r
Z Z n n ke
r V

o
/
2 1 2 1
2
2
) (
) (

+
+
=
(4)
We can compute the equilibrium lattice constant (separation), (r
o
) and lattice energy V(r
o
).
For example, from Table 7.17, a) calculate r
o
andV(r
o
) for MgO using only nearest neighbor
repulsions howdothevaluescomparewithTable71inRohrer
r 2
repulsions, how do the values compare with Table 7.1 in Rohrer.
O-O repulsions and the attractive VDW component (-C
ij
/r
6
) from L-J model [secondary effects]
havelittleinfluenceontheequilibriumlatticeconstant (separation) (r ) andlatticeenergyV(r ) have little influence on the equilibrium lattice constant (separation), (r
o
) and lattice energy V(r
o
).
Ionic pair potential calculations can be used to determine defects (point and area) in ionic
compounds, e.g., NiO grain/tilt boundaries have different surface energies than inside
i l i f i di d i l f FCC C
Prof. M.L. Weaver
grains; analogous to computing surface energies we discussed previously for FCC Cu.
Cohesive Energy (U) and Cohesive Energy (U) and Interatomic Interatomic
Separation Distance (d Separation Distance (d
oo
) in Metallic Bonding ) in Metallic Bonding Separation Distance (d Separation Distance (d
oo
) in Metallic Bonding ) in Metallic Bonding
=r
o
=r
o
Prof. M.L. Weaver
Summary: Bonding Summary: Bonding
Type
Ionic
Bond Energy
Large!
Comments
Nondirectional (ceramics)
Ionic
Covalent
Large!
Variable
Nondirectional (ceramics)
Directional
large-Diamond
small-Bismuth
(semiconductors, ceramics
polymer chains)
Metallic
Variable
large-Tungsten
small Merc r
Nondirectional (metals)
Secondary
small-Mercury
smallest Directional
inter chain (polymer)
Chapter 2 - 115
inter-chain (polymer)
inter-molecular
Summary: Primary Bonds Summary: Primary Bonds
Ceramics
(Ionic & covalent bonding):
Large bond energy
large T
m
large E
ll small o
Metals Variable bond energy
moderate T
(Metallic bonding):
moderate T
m
moderate E
moderate o
Polymers
(Covalent & Secondary):
Directional Properties
Secondary bonding dominates
small T
( y)
small T
m
small E
large o
Chapter 2 - 116